Merge tag 'devicetree-fixes-for-5.16-1' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux

Pull devicetree fixes from Rob Herring:

 - Two fixes due to DT node name changes on Arm, Ltd. boards

 - Treewide rename of Ingenic CGU headers

 - Update ST email addresses

 - Remove Netlogic DT bindings

 - Dropping few more cases of redundant 'maxItems' in schemas

 - Convert toshiba,tc358767 bridge binding to schema

* tag 'devicetree-fixes-for-5.16-1' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:
  dt-bindings: watchdog: sunxi: fix error in schema
  bindings: media: venus: Drop redundant maxItems for power-domain-names
  dt-bindings: Remove Netlogic bindings
  clk: versatile: clk-icst: Ensure clock names are unique
  of: Support using 'mask' in making device bus id
  dt-bindings: treewide: Update @st.com email address to @foss.st.com
  dt-bindings: media: Update maintainers for st,stm32-hwspinlock.yaml
  dt-bindings: media: Update maintainers for st,stm32-cec.yaml
  dt-bindings: mfd: timers: Update maintainers for st,stm32-timers
  dt-bindings: timer: Update maintainers for st,stm32-timer
  dt-bindings: i2c: imx: hardware do not restrict clock-frequency to only 100 and 400 kHz
  dt-bindings: display: bridge: Convert toshiba,tc358767.txt to yaml
  dt-bindings: Rename Ingenic CGU headers to ingenic,*.h
diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs
index f627e70..b268e3e 100644
--- a/Documentation/ABI/testing/sysfs-fs-f2fs
+++ b/Documentation/ABI/testing/sysfs-fs-f2fs
@@ -512,3 +512,19 @@
 Contact:	"Daeho Jeong" <daehojeong@google.com>
 Description:	You can	control the multiplier value of	bdi device readahead window size
 		between 2 (default) and 256 for POSIX_FADV_SEQUENTIAL advise option.
+
+What:		/sys/fs/f2fs/<disk>/max_fragment_chunk
+Date:		August 2021
+Contact:	"Daeho Jeong" <daehojeong@google.com>
+Description:	With "mode=fragment:block" mount options, we can scatter block allocation.
+		f2fs will allocate 1..<max_fragment_chunk> blocks in a chunk and make a hole
+		in the length of 1..<max_fragment_hole> by turns. This value can be set
+		between 1..512 and the default value is 4.
+
+What:		/sys/fs/f2fs/<disk>/max_fragment_hole
+Date:		August 2021
+Contact:	"Daeho Jeong" <daehojeong@google.com>
+Description:	With "mode=fragment:block" mount options, we can scatter block allocation.
+		f2fs will allocate 1..<max_fragment_chunk> blocks in a chunk and make a hole
+		in the length of 1..<max_fragment_hole> by turns. This value can be set
+		between 1..512 and the default value is 4.
diff --git a/Documentation/devicetree/bindings/clock/sifive/fu740-prci.yaml b/Documentation/devicetree/bindings/clock/sifive/fu740-prci.yaml
index e17143c..252085a 100644
--- a/Documentation/devicetree/bindings/clock/sifive/fu740-prci.yaml
+++ b/Documentation/devicetree/bindings/clock/sifive/fu740-prci.yaml
@@ -42,6 +42,9 @@
   "#clock-cells":
     const: 1
 
+  "#reset-cells":
+    const: 1
+
 required:
   - compatible
   - reg
@@ -57,4 +60,5 @@
       reg = <0x10000000 0x1000>;
       clocks = <&hfclk>, <&rtcclk>;
       #clock-cells = <1>;
+      #reset-cells = <1>;
     };
diff --git a/Documentation/devicetree/bindings/clock/silabs,si5351.txt b/Documentation/devicetree/bindings/clock/silabs,si5351.txt
index 8fe6f80..bfda6af 100644
--- a/Documentation/devicetree/bindings/clock/silabs,si5351.txt
+++ b/Documentation/devicetree/bindings/clock/silabs,si5351.txt
@@ -2,7 +2,7 @@
 
 Reference
 [1] Si5351A/B/C Data Sheet
-    https://www.silabs.com/Support%20Documents/TechnicalDocs/Si5351.pdf
+    https://www.skyworksinc.com/-/media/Skyworks/SL/documents/public/data-sheets/Si5351-B.pdf
 
 The Si5351a/b/c are programmable i2c clock generators with up to 8 output
 clocks. Si5351a also has a reduced pin-count package (MSOP10) where only
diff --git a/Documentation/devicetree/bindings/clock/socionext,uniphier-clock.yaml b/Documentation/devicetree/bindings/clock/socionext,uniphier-clock.yaml
index c3930ed..9a0cc73 100644
--- a/Documentation/devicetree/bindings/clock/socionext,uniphier-clock.yaml
+++ b/Documentation/devicetree/bindings/clock/socionext,uniphier-clock.yaml
@@ -23,6 +23,7 @@
           - socionext,uniphier-ld11-clock
           - socionext,uniphier-ld20-clock
           - socionext,uniphier-pxs3-clock
+          - socionext,uniphier-nx1-clock
       - description: Media I/O (MIO) clock, SD clock
         enum:
           - socionext,uniphier-ld4-mio-clock
@@ -33,6 +34,7 @@
           - socionext,uniphier-ld11-mio-clock
           - socionext,uniphier-ld20-sd-clock
           - socionext,uniphier-pxs3-sd-clock
+          - socionext,uniphier-nx1-sd-clock
       - description: Peripheral clock
         enum:
           - socionext,uniphier-ld4-peri-clock
@@ -43,6 +45,10 @@
           - socionext,uniphier-ld11-peri-clock
           - socionext,uniphier-ld20-peri-clock
           - socionext,uniphier-pxs3-peri-clock
+          - socionext,uniphier-nx1-peri-clock
+      - description: SoC-glue clock
+        enum:
+          - socionext,uniphier-pro4-sg-clock
 
   "#clock-cells":
     const: 1
diff --git a/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml b/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml
index 304a136..1faae3e 100644
--- a/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml
+++ b/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml
@@ -49,11 +49,26 @@
 
     properties:
       port@0:
-        $ref: /schemas/graph.yaml#/properties/port
+        $ref: /schemas/graph.yaml#/$defs/port-base
         description: |
           For LVDS encoders, port 0 is the parallel input
           For LVDS decoders, port 0 is the LVDS input
 
+        properties:
+          endpoint:
+            $ref: /schemas/media/video-interfaces.yaml#
+            unevaluatedProperties: false
+
+            properties:
+              data-mapping:
+                enum:
+                  - jeida-18
+                  - jeida-24
+                  - vesa-24
+                description: |
+                  The color signals mapping order. See details in
+                  Documentation/devicetree/bindings/display/panel/lvds.yaml
+
       port@1:
         $ref: /schemas/graph.yaml#/properties/port
         description: |
@@ -71,6 +86,22 @@
 
   power-supply: true
 
+if:
+  not:
+    properties:
+      compatible:
+        contains:
+          const: lvds-decoder
+then:
+  properties:
+    ports:
+      properties:
+        port@0:
+          properties:
+            endpoint:
+              properties:
+                data-mapping: false
+
 required:
   - compatible
   - ports
diff --git a/Documentation/devicetree/bindings/display/bridge/ps8640.yaml b/Documentation/devicetree/bindings/display/bridge/ps8640.yaml
index fce82b6..cdaf7a7 100644
--- a/Documentation/devicetree/bindings/display/bridge/ps8640.yaml
+++ b/Documentation/devicetree/bindings/display/bridge/ps8640.yaml
@@ -40,6 +40,9 @@
   vdd33-supply:
     description: Regulator for 3.3V digital core power.
 
+  aux-bus:
+    $ref: /schemas/display/dp-aux-bus.yaml#
+
   ports:
     $ref: /schemas/graph.yaml#/properties/ports
 
@@ -98,7 +101,21 @@
                     reg = <1>;
                     ps8640_out: endpoint {
                         remote-endpoint = <&panel_in>;
-                   };
+                    };
+                };
+            };
+
+            aux-bus {
+                panel {
+                    compatible = "boe,nv133fhm-n62";
+                    power-supply = <&pp3300_dx_edp>;
+                    backlight = <&backlight>;
+
+                    port {
+                        panel_in: endpoint {
+                            remote-endpoint = <&ps8640_out>;
+                        };
+                    };
                 };
             };
         };
diff --git a/Documentation/devicetree/bindings/display/panel/panel-simple.yaml b/Documentation/devicetree/bindings/display/panel/panel-simple.yaml
index 335776c..f3c9395d2 100644
--- a/Documentation/devicetree/bindings/display/panel/panel-simple.yaml
+++ b/Documentation/devicetree/bindings/display/panel/panel-simple.yaml
@@ -166,6 +166,8 @@
       - innolux,at070tn92
         # Innolux G070Y2-L01 7" WVGA (800x480) TFT LCD panel
       - innolux,g070y2-l01
+        # Innolux G070Y2-T02 7" WVGA (800x480) TFT LCD TTL panel
+      - innolux,g070y2-t02
         # Innolux Corporation 10.1" G101ICE-L01 WXGA (1280x800) LVDS panel
       - innolux,g101ice-l01
         # Innolux Corporation 12.1" WXGA (1280x800) TFT LCD panel
@@ -309,6 +311,8 @@
       - urt,umsh-8596md-11t
       - urt,umsh-8596md-19t
       - urt,umsh-8596md-20t
+        # Vivax TPC-9150 tablet 9.0" WSVGA TFT LCD panel
+      - vivax,tpc9150-panel
         # VXT 800x480 color TFT LCD panel
       - vxt,vl050-8048nt-c01
         # Winstar Display Corporation 3.5" QVGA (320x240) TFT LCD panel
@@ -317,6 +321,7 @@
       - yes-optoelectronics,ytc700tlag-05-201c
 
   backlight: true
+  ddc-i2c-bus: true
   enable-gpios: true
   port: true
   power-supply: true
diff --git a/Documentation/devicetree/bindings/display/panel/sharp,ls060t1sx01.yaml b/Documentation/devicetree/bindings/display/panel/sharp,ls060t1sx01.yaml
new file mode 100644
index 0000000..271c097
--- /dev/null
+++ b/Documentation/devicetree/bindings/display/panel/sharp,ls060t1sx01.yaml
@@ -0,0 +1,56 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/display/panel/sharp,ls060t1sx01.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Sharp Microelectronics 6.0" FullHD TFT LCD panel
+
+maintainers:
+  - Dmitry Baryskov <dmitry.baryshkov@linaro.org>
+
+allOf:
+  - $ref: panel-common.yaml#
+
+properties:
+  compatible:
+    const: sharp,ls060t1sx01
+
+  reg: true
+  backlight: true
+  reset-gpios: true
+  port: true
+
+  avdd-supply:
+    description: handle of the regulator that provides the positive supply voltage
+  avee-supply:
+    description: handle of the regulator that provides the negative supply voltage
+  vddi-supply:
+    description: handle of the regulator that provides the I/O supply voltage
+  vddh-supply:
+    description: handle of the regulator that provides the analog supply voltage
+
+required:
+  - compatible
+  - reg
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+
+    dsi {
+        #address-cells = <1>;
+        #size-cells = <0>;
+
+        panel@0 {
+            compatible = "sharp,ls060t1sx01";
+            reg = <0>;
+            avdd-supply = <&pm8941_l22>;
+            backlight = <&backlight>;
+            reset-gpios = <&pm8916_gpios 25 GPIO_ACTIVE_LOW>;
+        };
+    };
+
+...
diff --git a/Documentation/devicetree/bindings/input/cypress-sf.yaml b/Documentation/devicetree/bindings/input/cypress-sf.yaml
new file mode 100644
index 0000000..c0b0514
--- /dev/null
+++ b/Documentation/devicetree/bindings/input/cypress-sf.yaml
@@ -0,0 +1,61 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/input/cypress-sf.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Cypress StreetFighter touchkey controller
+
+maintainers:
+  - Yassine Oudjana <y.oudjana@protonmail.com>
+
+allOf:
+  - $ref: input.yaml#
+
+properties:
+  compatible:
+    const: cypress,sf3155
+
+  reg:
+    maxItems: 1
+
+  interrupts:
+    maxItems: 1
+
+  avdd-supply:
+    description: Regulator for AVDD analog voltage
+
+  vdd-supply:
+    description: Regulator for VDD digital voltage
+
+  linux,keycodes:
+    minItems: 1
+    maxItems: 8
+
+required:
+  - compatible
+  - reg
+  - interrupts
+  - avdd-supply
+  - vdd-supply
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/input/input.h>
+    #include <dt-bindings/interrupt-controller/irq.h>
+    i2c {
+        #address-cells = <1>;
+        #size-cells = <0>;
+
+        touchkey@28 {
+                compatible = "cypress,sf3155";
+                reg = <0x28>;
+                interrupt-parent = <&msmgpio>;
+                interrupts = <77 IRQ_TYPE_EDGE_FALLING>;
+                avdd-supply = <&vreg_l6a_1p8>;
+                vdd-supply = <&vdd_3v2_tp>;
+                linux,keycodes = <KEY_BACK KEY_MENU>;
+        };
+    };
diff --git a/Documentation/devicetree/bindings/input/microchip,cap11xx.yaml b/Documentation/devicetree/bindings/input/microchip,cap11xx.yaml
index fa0f37a..d5d6bce 100644
--- a/Documentation/devicetree/bindings/input/microchip,cap11xx.yaml
+++ b/Documentation/devicetree/bindings/input/microchip,cap11xx.yaml
@@ -19,6 +19,7 @@
       - microchip,cap1106
       - microchip,cap1126
       - microchip,cap1188
+      - microchip,cap1206
 
   reg:
     maxItems: 1
diff --git a/Documentation/devicetree/bindings/pwm/renesas,tpu-pwm.yaml b/Documentation/devicetree/bindings/pwm/renesas,tpu-pwm.yaml
index 81ccb21..1f5c638 100644
--- a/Documentation/devicetree/bindings/pwm/renesas,tpu-pwm.yaml
+++ b/Documentation/devicetree/bindings/pwm/renesas,tpu-pwm.yaml
@@ -35,9 +35,11 @@
           - renesas,tpu-r8a7794   # R-Car E2
           - renesas,tpu-r8a7795   # R-Car H3
           - renesas,tpu-r8a7796   # R-Car M3-W
+          - renesas,tpu-r8a77961  # R-Car M3-W+
           - renesas,tpu-r8a77965  # R-Car M3-N
           - renesas,tpu-r8a77970  # R-Car V3M
           - renesas,tpu-r8a77980  # R-Car V3H
+          - renesas,tpu-r8a779a0  # R-Car V3U
       - const: renesas,tpu
 
   reg:
diff --git a/Documentation/devicetree/bindings/rtc/mstar,msc313-rtc.yaml b/Documentation/devicetree/bindings/rtc/mstar,msc313-rtc.yaml
new file mode 100644
index 0000000..114199c
--- /dev/null
+++ b/Documentation/devicetree/bindings/rtc/mstar,msc313-rtc.yaml
@@ -0,0 +1,49 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/rtc/mstar,msc313-rtc.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Mstar MSC313e RTC Device Tree Bindings
+
+allOf:
+  - $ref: "rtc.yaml#"
+
+maintainers:
+  - Daniel Palmer <daniel@0x0f.com>
+  - Romain Perier <romain.perier@gmail.com>
+
+properties:
+  compatible:
+    enum:
+      - mstar,msc313-rtc
+
+  reg:
+    maxItems: 1
+
+  interrupts:
+    maxItems: 1
+
+  start-year: true
+
+  clocks:
+    maxItems: 1
+
+required:
+  - compatible
+  - reg
+  - interrupts
+  - clocks
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/interrupt-controller/arm-gic.h>
+    rtc@2400 {
+        compatible = "mstar,msc313-rtc";
+        reg = <0x2400 0x40>;
+        clocks = <&xtal_div2>;
+        interrupts-extended = <&intc_irq GIC_SPI 44 IRQ_TYPE_LEVEL_HIGH>;
+    };
+...
diff --git a/Documentation/devicetree/bindings/rtc/nxp,pcf85063.txt b/Documentation/devicetree/bindings/rtc/nxp,pcf85063.txt
index 627bb53..6439682 100644
--- a/Documentation/devicetree/bindings/rtc/nxp,pcf85063.txt
+++ b/Documentation/devicetree/bindings/rtc/nxp,pcf85063.txt
@@ -13,10 +13,19 @@
   expressed in femto Farad (fF). Valid values are 7000 and 12500.
   Default value (if no value is specified) is 7000fF.
 
+Optional child node:
+- clock: Provide this if the square wave pin is used as boot-enabled fixed clock.
+
 Example:
 
 pcf85063: rtc@51 {
 	compatible = "nxp,pcf85063";
 	reg = <0x51>;
 	quartz-load-femtofarads = <12500>;
+
+		clock {
+			compatible = "fixed-clock";
+			#clock-cells = <0>;
+			clock-frequency = <32768>;
+		};
 };
diff --git a/Documentation/devicetree/bindings/vendor-prefixes.yaml b/Documentation/devicetree/bindings/vendor-prefixes.yaml
index 983f5e4..66d6432 100644
--- a/Documentation/devicetree/bindings/vendor-prefixes.yaml
+++ b/Documentation/devicetree/bindings/vendor-prefixes.yaml
@@ -1286,6 +1286,8 @@
     description: Vitesse Semiconductor Corporation
   "^vivante,.*":
     description: Vivante Corporation
+  "^vivax,.*":
+    description: Vivax brand by M SAN Grupa d.o.o.
   "^vocore,.*":
     description: VoCore Studio
   "^voipac,.*":
diff --git a/Documentation/filesystems/f2fs.rst b/Documentation/filesystems/f2fs.rst
index 6f3c6e9..d7b8469 100644
--- a/Documentation/filesystems/f2fs.rst
+++ b/Documentation/filesystems/f2fs.rst
@@ -197,10 +197,29 @@
 			 FAULT_DISCARD		  0x000002000
 			 FAULT_WRITE_IO		  0x000004000
 			 FAULT_SLAB_ALLOC	  0x000008000
+			 FAULT_DQUOT_INIT	  0x000010000
 			 ===================	  ===========
 mode=%s			 Control block allocation mode which supports "adaptive"
 			 and "lfs". In "lfs" mode, there should be no random
 			 writes towards main area.
+			 "fragment:segment" and "fragment:block" are newly added here.
+			 These are developer options for experiments to simulate filesystem
+			 fragmentation/after-GC situation itself. The developers use these
+			 modes to understand filesystem fragmentation/after-GC condition well,
+			 and eventually get some insights to handle them better.
+			 In "fragment:segment", f2fs allocates a new segment in ramdom
+			 position. With this, we can simulate the after-GC condition.
+			 In "fragment:block", we can scatter block allocation with
+			 "max_fragment_chunk" and "max_fragment_hole" sysfs nodes.
+			 We added some randomness to both chunk and hole size to make
+			 it close to realistic IO pattern. So, in this mode, f2fs will allocate
+			 1..<max_fragment_chunk> blocks in a chunk and make a hole in the
+			 length of 1..<max_fragment_hole> by turns. With this, the newly
+			 allocated blocks will be scattered throughout the whole partition.
+			 Note that "fragment:block" implicitly enables "fragment:segment"
+			 option for more randomness.
+			 Please, use these options for your experiments and we strongly
+			 recommend to re-format the filesystem after using these options.
 io_bits=%u		 Set the bit size of write IO requests. It should be set
 			 with "mode=lfs".
 usrquota		 Enable plain user disk quota accounting.
diff --git a/Documentation/gpu/todo.rst b/Documentation/gpu/todo.rst
index 6613543..60d1d7e 100644
--- a/Documentation/gpu/todo.rst
+++ b/Documentation/gpu/todo.rst
@@ -314,16 +314,19 @@
 Garbage collect fbdev scrolling acceleration
 --------------------------------------------
 
-Scroll acceleration is disabled in fbcon by hard-wiring p->scrollmode =
-SCROLL_REDRAW. There's a ton of code this will allow us to remove:
+Scroll acceleration has been disabled in fbcon. Now it works as the old
+SCROLL_REDRAW mode. A ton of code was removed in fbcon.c and the hook bmove was
+removed from fbcon_ops.
+Remaining tasks:
 
-- lots of code in fbcon.c
-
-- a bunch of the hooks in fbcon_ops, maybe the remaining hooks could be called
+- a bunch of the hooks in fbcon_ops could be removed or simplified by calling
   directly instead of the function table (with a switch on p->rotate)
 
 - fb_copyarea is unused after this, and can be deleted from all drivers
 
+- after that, fb_copyarea can be deleted from fb_ops in include/linux/fb.h as
+  well as cfb_copyarea
+
 Note that not all acceleration code can be deleted, since clearing and cursor
 support is still accelerated, which might be good candidates for further
 deletion projects.
diff --git a/Documentation/security/SCTP.rst b/Documentation/security/SCTP.rst
index 406cc68..d5fd6cc 100644
--- a/Documentation/security/SCTP.rst
+++ b/Documentation/security/SCTP.rst
@@ -15,7 +15,10 @@
     security_sctp_assoc_request()
     security_sctp_bind_connect()
     security_sctp_sk_clone()
-    security_sctp_assoc_established()
+
+Also the following security hook has been utilised::
+
+    security_inet_conn_established()
 
 The usage of these hooks are described below with the SELinux implementation
 described in the `SCTP SELinux Support`_ chapter.
@@ -119,12 +122,11 @@
     @newsk - pointer to new sock structure.
 
 
-security_sctp_assoc_established()
+security_inet_conn_established()
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-Called when a COOKIE ACK is received, and the peer secid will be
-saved into ``@asoc->peer_secid`` for client::
+Called when a COOKIE ACK is received::
 
-    @asoc - pointer to sctp association structure.
+    @sk  - pointer to sock structure.
     @skb - pointer to skbuff of the COOKIE ACK packet.
 
 
@@ -132,7 +134,7 @@
 -------------------------------------------------
 
 The following diagram shows the use of ``security_sctp_bind_connect()``,
-``security_sctp_assoc_request()``, ``security_sctp_assoc_established()`` when
+``security_sctp_assoc_request()``, ``security_inet_conn_established()`` when
 establishing an association.
 ::
 
@@ -170,7 +172,7 @@
           <------------------------------------------- COOKIE ACK
           |                                               |
     sctp_sf_do_5_1E_ca                                    |
- Call security_sctp_assoc_established()                   |
+ Call security_inet_conn_established()                    |
  to set the peer label.                                   |
           |                                               |
           |                               If SCTP_SOCKET_TCP or peeled off
@@ -196,7 +198,7 @@
     security_sctp_assoc_request()
     security_sctp_bind_connect()
     security_sctp_sk_clone()
-    security_sctp_assoc_established()
+    security_inet_conn_established()
 
 
 security_sctp_assoc_request()
@@ -269,12 +271,12 @@
     @newsk - pointer to new sock structure.
 
 
-security_sctp_assoc_established()
+security_inet_conn_established()
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 Called when a COOKIE ACK is received where it sets the connection's peer sid
 to that in ``@skb``::
 
-    @asoc - pointer to sctp association structure.
+    @sk  - pointer to sock structure.
     @skb - pointer to skbuff of the COOKIE ACK packet.
 
 
diff --git a/Documentation/virt/kvm/api.rst b/Documentation/virt/kvm/api.rst
index 3b093d6d..aeeb071 100644
--- a/Documentation/virt/kvm/api.rst
+++ b/Documentation/virt/kvm/api.rst
@@ -6911,6 +6911,20 @@
 When enabled the VMM may make use of the ``KVM_ARM_MTE_COPY_TAGS`` ioctl to
 perform a bulk copy of tags to/from the guest.
 
+7.29 KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM
+-------------------------------------
+
+Architectures: x86 SEV enabled
+Type: vm
+Parameters: args[0] is the fd of the source vm
+Returns: 0 on success
+
+This capability enables userspace to migrate the encryption context from the VM
+indicated by the fd to the VM this is called on.
+
+This is intended to support intra-host migration of VMs between userspace VMMs,
+upgrading the VMM process without interrupting the guest.
+
 8. Other capabilities.
 ======================
 
diff --git a/Documentation/x86/xstate.rst b/Documentation/x86/xstate.rst
index 65de3f0..5cec7fb 100644
--- a/Documentation/x86/xstate.rst
+++ b/Documentation/x86/xstate.rst
@@ -63,3 +63,12 @@
 the handler allocates a larger xstate buffer for the task so the large
 state can be context switched. In the unlikely cases that the allocation
 fails, the kernel sends SIGSEGV.
+
+Dynamic features in signal frames
+---------------------------------
+
+Dynamcally enabled features are not written to the signal frame upon signal
+entry if the feature is in its initial configuration.  This differs from
+non-dynamic features which are always written regardless of their
+configuration.  Signal handlers can examine the XSAVE buffer's XSTATE_BV
+field to determine if a features was written.
diff --git a/MAINTAINERS b/MAINTAINERS
index ceba033..7a2345c 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2283,6 +2283,7 @@
 F:	arch/arm/mach-mstar/
 F:	drivers/clk/mstar/
 F:	drivers/gpio/gpio-msc313.c
+F:	drivers/rtc/rtc-msc313.c
 F:	drivers/watchdog/msc313e_wdt.c
 F:	include/dt-bindings/clock/mstar-*
 F:	include/dt-bindings/gpio/msc313-gpio.h
@@ -4458,7 +4459,7 @@
 M:	Hans de Goede <hdegoede@redhat.com>
 L:	linux-input@vger.kernel.org
 S:	Maintained
-F:	Documentation/devicetree/bindings/input/touchscreen/chipone_icn8318.txt
+F:	Documentation/devicetree/bindings/input/touchscreen/chipone,icn8318.yaml
 F:	drivers/input/touchscreen/chipone_icn8318.c
 
 CHIPONE ICN8505 I2C TOUCHSCREEN DRIVER
@@ -4675,11 +4676,10 @@
 M:	Julia Lawall <Julia.Lawall@inria.fr>
 M:	Gilles Muller <Gilles.Muller@inria.fr>
 M:	Nicolas Palix <nicolas.palix@imag.fr>
-M:	Michal Marek <michal.lkml@markovi.net>
-L:	cocci@systeme.lip6.fr (moderated for non-subscribers)
+L:	cocci@inria.fr (moderated for non-subscribers)
 S:	Supported
-W:	http://coccinelle.lip6.fr/
-T:	git git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild.git misc
+W:	https://coccinelle.gitlabpages.inria.fr/website/
+T:	git git://git.kernel.org/pub/scm/linux/kernel/git/jlawall/linux.git
 F:	Documentation/dev-tools/coccinelle.rst
 F:	scripts/coccicheck
 F:	scripts/coccinelle/
@@ -5204,6 +5204,13 @@
 S:	Maintained
 F:	drivers/input/touchscreen/cy8ctma140.c
 
+CYPRESS STREETFIGHTER TOUCHKEYS DRIVER
+M:	Yassine Oudjana <y.oudjana@protonmail.com>
+L:	linux-input@vger.kernel.org
+S:	Maintained
+F:	Documentation/devicetree/bindings/input/cypress-sf.yaml
+F:	drivers/input/keyboard/cypress-sf.c
+
 CYTTSP TOUCHSCREEN DRIVER
 M:	Linus Walleij <linus.walleij@linaro.org>
 L:	linux-input@vger.kernel.org
@@ -8062,9 +8069,10 @@
 
 GOODIX TOUCHSCREEN
 M:	Bastien Nocera <hadess@hadess.net>
+M:	Hans de Goede <hdegoede@redhat.com>
 L:	linux-input@vger.kernel.org
 S:	Maintained
-F:	drivers/input/touchscreen/goodix.c
+F:	drivers/input/touchscreen/goodix*
 
 GOOGLE ETHERNET DRIVERS
 M:	Jeroen de Borst <jeroendb@google.com>
@@ -10799,11 +10807,6 @@
 F:	include/linux/ata.h
 F:	include/linux/libata.h
 
-LIBLOCKDEP
-M:	Sasha Levin <alexander.levin@microsoft.com>
-S:	Maintained
-F:	tools/lib/lockdep/
-
 LIBNVDIMM BLK: MMIO-APERTURE DRIVER
 M:	Dan Williams <dan.j.williams@intel.com>
 M:	Vishal Verma <vishal.l.verma@intel.com>
@@ -14409,7 +14412,9 @@
 M:	Deep Shah <sdeep@vmware.com>
 M:	"VMware, Inc." <pv-drivers@vmware.com>
 L:	virtualization@lists.linux-foundation.org
+L:	x86@kernel.org
 S:	Supported
+T:	git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86/core
 F:	Documentation/virt/paravirt_ops.rst
 F:	arch/*/include/asm/paravirt*.h
 F:	arch/*/kernel/paravirt*
@@ -21075,6 +21080,18 @@
 F:	include/linux/zsmalloc.h
 F:	mm/zsmalloc.c
 
+ZSTD
+M:	Nick Terrell <terrelln@fb.com>
+S:	Maintained
+B:	https://github.com/facebook/zstd/issues
+T:	git git://github.com/terrelln/linux.git
+F:	include/linux/zstd*
+F:	lib/zstd/
+F:	lib/decompress_unzstd.c
+F:	crypto/zstd.c
+N:	zstd
+K:	zstd
+
 ZSWAP COMPRESSED SWAP CACHING
 M:	Seth Jennings <sjenning@redhat.com>
 M:	Dan Streetman <ddstreet@ieee.org>
diff --git a/arch/arm64/include/asm/esr.h b/arch/arm64/include/asm/esr.h
index a305ce2..d52a0b2 100644
--- a/arch/arm64/include/asm/esr.h
+++ b/arch/arm64/include/asm/esr.h
@@ -68,6 +68,7 @@
 #define ESR_ELx_EC_MAX		(0x3F)
 
 #define ESR_ELx_EC_SHIFT	(26)
+#define ESR_ELx_EC_WIDTH	(6)
 #define ESR_ELx_EC_MASK		(UL(0x3F) << ESR_ELx_EC_SHIFT)
 #define ESR_ELx_EC(esr)		(((esr) & ESR_ELx_EC_MASK) >> ESR_ELx_EC_SHIFT)
 
diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
index 4be8486..2a5f7f3 100644
--- a/arch/arm64/include/asm/kvm_host.h
+++ b/arch/arm64/include/asm/kvm_host.h
@@ -584,7 +584,7 @@ struct kvm_vcpu_stat {
 	u64 exits;
 };
 
-int kvm_vcpu_preferred_target(struct kvm_vcpu_init *init);
+void kvm_vcpu_preferred_target(struct kvm_vcpu_init *init);
 unsigned long kvm_arm_num_regs(struct kvm_vcpu *vcpu);
 int kvm_arm_copy_reg_indices(struct kvm_vcpu *vcpu, u64 __user *indices);
 int kvm_arm_get_reg(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg);
diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
index f5490af..2f03cbf 100644
--- a/arch/arm64/kvm/arm.c
+++ b/arch/arm64/kvm/arm.c
@@ -1389,12 +1389,9 @@ long kvm_arch_vm_ioctl(struct file *filp,
 		return kvm_vm_ioctl_set_device_addr(kvm, &dev_addr);
 	}
 	case KVM_ARM_PREFERRED_TARGET: {
-		int err;
 		struct kvm_vcpu_init init;
 
-		err = kvm_vcpu_preferred_target(&init);
-		if (err)
-			return err;
+		kvm_vcpu_preferred_target(&init);
 
 		if (copy_to_user(argp, &init, sizeof(init)))
 			return -EFAULT;
diff --git a/arch/arm64/kvm/guest.c b/arch/arm64/kvm/guest.c
index 5ce26be..e116c77 100644
--- a/arch/arm64/kvm/guest.c
+++ b/arch/arm64/kvm/guest.c
@@ -869,13 +869,10 @@ u32 __attribute_const__ kvm_target_cpu(void)
 	return KVM_ARM_TARGET_GENERIC_V8;
 }
 
-int kvm_vcpu_preferred_target(struct kvm_vcpu_init *init)
+void kvm_vcpu_preferred_target(struct kvm_vcpu_init *init)
 {
 	u32 target = kvm_target_cpu();
 
-	if (target < 0)
-		return -ENODEV;
-
 	memset(init, 0, sizeof(*init));
 
 	/*
@@ -885,8 +882,6 @@ int kvm_vcpu_preferred_target(struct kvm_vcpu_init *init)
 	 * target type.
 	 */
 	init->target = (__u32)target;
-
-	return 0;
 }
 
 int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
diff --git a/arch/arm64/kvm/hyp/hyp-entry.S b/arch/arm64/kvm/hyp/hyp-entry.S
index 9aa9b734..b6b6801 100644
--- a/arch/arm64/kvm/hyp/hyp-entry.S
+++ b/arch/arm64/kvm/hyp/hyp-entry.S
@@ -44,7 +44,7 @@
 el1_sync:				// Guest trapped into EL2
 
 	mrs	x0, esr_el2
-	lsr	x0, x0, #ESR_ELx_EC_SHIFT
+	ubfx	x0, x0, #ESR_ELx_EC_SHIFT, #ESR_ELx_EC_WIDTH
 	cmp	x0, #ESR_ELx_EC_HVC64
 	ccmp	x0, #ESR_ELx_EC_HVC32, #4, ne
 	b.ne	el1_trap
diff --git a/arch/arm64/kvm/hyp/nvhe/host.S b/arch/arm64/kvm/hyp/nvhe/host.S
index 0c6116d..3d613e7 100644
--- a/arch/arm64/kvm/hyp/nvhe/host.S
+++ b/arch/arm64/kvm/hyp/nvhe/host.S
@@ -141,7 +141,7 @@
 .L__vect_start\@:
 	stp	x0, x1, [sp, #-16]!
 	mrs	x0, esr_el2
-	lsr	x0, x0, #ESR_ELx_EC_SHIFT
+	ubfx	x0, x0, #ESR_ELx_EC_SHIFT, #ESR_ELx_EC_WIDTH
 	cmp	x0, #ESR_ELx_EC_HVC64
 	b.eq	__host_hvc
 	b	__host_exit
diff --git a/arch/arm64/kvm/hyp/nvhe/setup.c b/arch/arm64/kvm/hyp/nvhe/setup.c
index 862c7b5..578f717 100644
--- a/arch/arm64/kvm/hyp/nvhe/setup.c
+++ b/arch/arm64/kvm/hyp/nvhe/setup.c
@@ -178,7 +178,7 @@ static int finalize_host_mappings_walker(u64 addr, u64 end, u32 level,
 
 	phys = kvm_pte_to_phys(pte);
 	if (!addr_is_memory(phys))
-		return 0;
+		return -EINVAL;
 
 	/*
 	 * Adjust the host stage-2 mappings to match the ownership attributes
@@ -207,8 +207,18 @@ static int finalize_host_mappings(void)
 		.cb	= finalize_host_mappings_walker,
 		.flags	= KVM_PGTABLE_WALK_LEAF,
 	};
+	int i, ret;
 
-	return kvm_pgtable_walk(&pkvm_pgtable, 0, BIT(pkvm_pgtable.ia_bits), &walker);
+	for (i = 0; i < hyp_memblock_nr; i++) {
+		struct memblock_region *reg = &hyp_memory[i];
+		u64 start = (u64)hyp_phys_to_virt(reg->base);
+
+		ret = kvm_pgtable_walk(&pkvm_pgtable, start, reg->size, &walker);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
 }
 
 void __noreturn __pkvm_init_finalise(void)
diff --git a/arch/arm64/kvm/hyp/nvhe/sys_regs.c b/arch/arm64/kvm/hyp/nvhe/sys_regs.c
index 3787ee6..792cf6e 100644
--- a/arch/arm64/kvm/hyp/nvhe/sys_regs.c
+++ b/arch/arm64/kvm/hyp/nvhe/sys_regs.c
@@ -474,7 +474,7 @@ bool kvm_handle_pvm_sysreg(struct kvm_vcpu *vcpu, u64 *exit_code)
 	return true;
 }
 
-/**
+/*
  * Handler for protected VM restricted exceptions.
  *
  * Inject an undefined exception into the guest and return true to indicate that
diff --git a/arch/mips/Kbuild.platforms b/arch/mips/Kbuild.platforms
index 2c57994..30193bc 100644
--- a/arch/mips/Kbuild.platforms
+++ b/arch/mips/Kbuild.platforms
@@ -37,4 +37,4 @@
 platform-$(CONFIG_MACH_VR41XX)		+= vr41xx/
 
 # include the platform specific files
-include $(patsubst %, $(srctree)/arch/mips/%/Platform, $(platform-y))
+include $(patsubst %/, $(srctree)/arch/mips/%/Platform, $(platform-y))
diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig
index 8651074..de60ad1 100644
--- a/arch/mips/Kconfig
+++ b/arch/mips/Kconfig
@@ -292,6 +292,8 @@
 	select USB_OHCI_BIG_ENDIAN_DESC if CPU_BIG_ENDIAN
 	select USB_OHCI_BIG_ENDIAN_MMIO if CPU_BIG_ENDIAN
 	select HARDIRQS_SW_RESEND
+	select HAVE_PCI
+	select PCI_DRIVERS_GENERIC
 	help
 	  Build a generic DT-based kernel image that boots on select
 	  BCM33xx cable modem chips, BCM63xx DSL chips, and BCM7xxx set-top
@@ -333,6 +335,9 @@
 	select SYS_SUPPORTS_32BIT_KERNEL
 	select SYS_SUPPORTS_BIG_ENDIAN
 	select SYS_HAS_EARLY_PRINTK
+	select SYS_HAS_CPU_BMIPS32_3300
+	select SYS_HAS_CPU_BMIPS4350
+	select SYS_HAS_CPU_BMIPS4380
 	select SWAP_IO_SPACE
 	select GPIOLIB
 	select MIPS_L1_CACHE_SHIFT_4
diff --git a/arch/mips/Makefile b/arch/mips/Makefile
index e036fc0..ace7f03 100644
--- a/arch/mips/Makefile
+++ b/arch/mips/Makefile
@@ -253,7 +253,9 @@
 #
 # Board-dependent options and extra files
 #
+ifdef need-compiler
 include $(srctree)/arch/mips/Kbuild.platforms
+endif
 
 ifdef CONFIG_PHYSICAL_START
 load-y					= $(CONFIG_PHYSICAL_START)
diff --git a/arch/mips/boot/compressed/.gitignore b/arch/mips/boot/compressed/.gitignore
deleted file mode 100644
index d358395..0000000
--- a/arch/mips/boot/compressed/.gitignore
+++ /dev/null
@@ -1,3 +0,0 @@
-# SPDX-License-Identifier: GPL-2.0-only
-ashldi3.c
-bswapsi.c
diff --git a/arch/mips/boot/compressed/Makefile b/arch/mips/boot/compressed/Makefile
index 3548b3b..2861a05 100644
--- a/arch/mips/boot/compressed/Makefile
+++ b/arch/mips/boot/compressed/Makefile
@@ -50,19 +50,9 @@
 vmlinuzobjs-$(CONFIG_ATH79)			   += $(obj)/uart-ath79.o
 endif
 
-extra-y += uart-ath79.c
-$(obj)/uart-ath79.c: $(srctree)/arch/mips/ath79/early_printk.c
-	$(call cmd,shipped)
-
 vmlinuzobjs-$(CONFIG_KERNEL_XZ) += $(obj)/ashldi3.o
 
-extra-y += ashldi3.c
-$(obj)/ashldi3.c: $(obj)/%.c: $(srctree)/lib/%.c FORCE
-	$(call if_changed,shipped)
-
-extra-y += bswapsi.c
-$(obj)/bswapsi.c: $(obj)/%.c: $(srctree)/arch/mips/lib/%.c FORCE
-	$(call if_changed,shipped)
+vmlinuzobjs-$(CONFIG_KERNEL_ZSTD) += $(obj)/bswapdi.o
 
 targets := $(notdir $(vmlinuzobjs-y))
 
diff --git a/arch/mips/boot/compressed/ashldi3.c b/arch/mips/boot/compressed/ashldi3.c
new file mode 100644
index 0000000..f7bf6a7
--- /dev/null
+++ b/arch/mips/boot/compressed/ashldi3.c
@@ -0,0 +1,2 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include "../../../../lib/ashldi3.c"
diff --git a/arch/mips/boot/compressed/bswapdi.c b/arch/mips/boot/compressed/bswapdi.c
new file mode 100644
index 0000000..acb28ae
--- /dev/null
+++ b/arch/mips/boot/compressed/bswapdi.c
@@ -0,0 +1,2 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include "../../lib/bswapdi.c"
diff --git a/arch/mips/boot/compressed/bswapsi.c b/arch/mips/boot/compressed/bswapsi.c
new file mode 100644
index 0000000..fdb9c64
--- /dev/null
+++ b/arch/mips/boot/compressed/bswapsi.c
@@ -0,0 +1,2 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include "../../lib/bswapsi.c"
diff --git a/arch/mips/boot/compressed/uart-ath79.c b/arch/mips/boot/compressed/uart-ath79.c
new file mode 100644
index 0000000..d686820
--- /dev/null
+++ b/arch/mips/boot/compressed/uart-ath79.c
@@ -0,0 +1,2 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include "../../ath79/early_printk.c"
diff --git a/arch/mips/configs/bmips_stb_defconfig b/arch/mips/configs/bmips_stb_defconfig
index 625bd2d..5956fb9 100644
--- a/arch/mips/configs/bmips_stb_defconfig
+++ b/arch/mips/configs/bmips_stb_defconfig
@@ -1,6 +1,7 @@
 # CONFIG_LOCALVERSION_AUTO is not set
 # CONFIG_SWAP is not set
 CONFIG_NO_HZ=y
+CONFIG_HZ=1000
 CONFIG_BLK_DEV_INITRD=y
 CONFIG_EXPERT=y
 # CONFIG_VM_EVENT_COUNTERS is not set
@@ -8,17 +9,34 @@
 CONFIG_BMIPS_GENERIC=y
 CONFIG_CPU_LITTLE_ENDIAN=y
 CONFIG_HIGHMEM=y
+CONFIG_HIGH_RES_TIMERS=y
 CONFIG_SMP=y
 CONFIG_NR_CPUS=4
+CONFIG_CC_STACKPROTECTOR_STRONG=y
 # CONFIG_SECCOMP is not set
 CONFIG_MIPS_O32_FP64_SUPPORT=y
+# CONFIG_RD_GZIP is not set
+# CONFIG_RD_BZIP2 is not set
+# CONFIG_RD_LZMA is not set
+CONFIG_RD_XZ=y
+# CONFIG_RD_LZO is not set
+# CONFIG_RD_LZ4 is not set
+# CONFIG_IOSCHED_DEADLINE is not set
+# CONFIG_IOSCHED_CFQ is not set
+CONFIG_PCI=y
+CONFIG_PCI_MSI=y
+CONFIG_PCIEASPM_POWERSAVE=y
+CONFIG_PCIEPORTBUS=y
+CONFIG_PCIE_BRCMSTB=y
 CONFIG_CPU_FREQ=y
 CONFIG_CPU_FREQ_STAT=y
+CONFIG_CPU_FREQ_STAT_DETAILS=y
+CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE=y
+CONFIG_CPU_FREQ_GOV_PERFORMANCE=y
 CONFIG_CPU_FREQ_GOV_POWERSAVE=y
 CONFIG_CPU_FREQ_GOV_USERSPACE=y
 CONFIG_CPU_FREQ_GOV_ONDEMAND=y
 CONFIG_CPU_FREQ_GOV_CONSERVATIVE=y
-CONFIG_CPU_FREQ_GOV_SCHEDUTIL=y
 CONFIG_BMIPS_CPUFREQ=y
 # CONFIG_BLK_DEV_BSG is not set
 CONFIG_NET=y
@@ -32,32 +50,99 @@
 # CONFIG_INET_DIAG is not set
 CONFIG_CFG80211=y
 CONFIG_NL80211_TESTMODE=y
+CONFIG_WIRELESS=y
 CONFIG_MAC80211=y
+CONFIG_NL80211=y
 CONFIG_DEVTMPFS=y
 CONFIG_DEVTMPFS_MOUNT=y
 # CONFIG_STANDALONE is not set
 # CONFIG_PREVENT_FIRMWARE_BUILD is not set
+CONFIG_BRCMSTB_GISB_ARB=y
+CONFIG_MODULES=y
+CONFIG_MODULE_FORCE_LOAD=y
+CONFIG_MODULE_UNLOAD=y
+CONFIG_MODVERSIONS=y
+CONFIG_IP_MULTICAST=y
+CONFIG_IP_PNP=y
+CONFIG_IP_PNP_DHCP=y
+CONFIG_IP_PNP_BOOTP=y
+CONFIG_IP_PNP_RARP=y
+CONFIG_IP_MROUTE=y
+CONFIG_IP_PIMSM_V1=y
+CONFIG_IP_PIMSM_V2=y
+# CONFIG_INET_XFRM_MODE_TRANSPORT is not set
+# CONFIG_INET_XFRM_MODE_TUNNEL is not set
+# CONFIG_INET_XFRM_MODE_BEET is not set
+# CONFIG_INET_LRO is not set
+CONFIG_INET_UDP_DIAG=y
+CONFIG_TCP_CONG_ADVANCED=y
+CONFIG_TCP_CONG_BIC=y
+# CONFIG_TCP_CONG_WESTWOOD is not set
+# CONFIG_TCP_CONG_HTCP is not set
+# CONFIG_IPV6 is not set
+CONFIG_IP_NF_IPTABLES=y
+CONFIG_IP_NF_FILTER=y
+CONFIG_NETFILTER=y
+CONFIG_NETFILTER_XTABLES=y
+CONFIG_BRIDGE=y
+CONFIG_BRIDGE_NETFILTER=m
+CONFIG_BRIDGE_NF_EBTABLES=m
+CONFIG_BRIDGE_EBT_BROUTE=m
+CONFIG_NET_DSA=y
+CONFIG_NET_SWITCHDEV=y
+CONFIG_DMA_CMA=y
+CONFIG_CMA_ALIGNMENT=12
+CONFIG_SPI=y
+CONFIG_SPI_BRCMSTB=y
 CONFIG_MTD=y
+CONFIG_MTD_CMDLINE_PARTS=y
+CONFIG_MTD_BLOCK=y
 CONFIG_MTD_CFI=y
+CONFIG_MTD_JEDECPROBE=y
 CONFIG_MTD_CFI_INTELEXT=y
 CONFIG_MTD_CFI_AMDSTD=y
-CONFIG_MTD_PHYSMAP=y
+CONFIG_MTD_CFI_STAA=y
+CONFIG_MTD_ROM=y
+CONFIG_MTD_ABSENT=y
+CONFIG_MTD_PHYSMAP_OF=y
+CONFIG_MTD_M25P80=y
+CONFIG_MTD_NAND=y
+CONFIG_MTD_NAND_BRCMNAND=y
+CONFIG_MTD_SPI_NOR=y
+# CONFIG_MTD_SPI_NOR_USE_4K_SECTORS is not set
+CONFIG_MTD_UBI=y
+CONFIG_MTD_UBI_GLUEBI=y
+CONFIG_PROC_DEVICETREE=y
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_SIZE=8192
 # CONFIG_BLK_DEV is not set
 CONFIG_SCSI=y
 CONFIG_BLK_DEV_SD=y
+CONFIG_CHR_DEV_SG=y
+CONFIG_SCSI_MULTI_LUN=y
 # CONFIG_SCSI_LOWLEVEL is not set
 CONFIG_NETDEVICES=y
+CONFIG_VLAN_8021Q=y
+CONFIG_MACVLAN=y
 CONFIG_BCMGENET=y
 CONFIG_USB_USBNET=y
-# CONFIG_INPUT is not set
+CONFIG_INPUT_EVDEV=y
+# CONFIG_INPUT_KEYBOARD is not set
+# CONFIG_INPUT_MOUSE is not set
+CONFIG_INPUT_MISC=y
+CONFIG_INPUT_UINPUT=y
 # CONFIG_SERIO is not set
-# CONFIG_VT is not set
+CONFIG_VT=y
+CONFIG_VT_HW_CONSOLE_BINDING=y
+# CONFIG_DEVKMEM is not set
 CONFIG_SERIAL_8250=y
 # CONFIG_SERIAL_8250_DEPRECATED_OPTIONS is not set
 CONFIG_SERIAL_8250_CONSOLE=y
 CONFIG_SERIAL_OF_PLATFORM=y
 # CONFIG_HW_RANDOM is not set
 CONFIG_POWER_RESET=y
+CONFIG_POWER_RESET_BRCMSTB=y
 CONFIG_POWER_RESET_SYSCON=y
 CONFIG_POWER_SUPPLY=y
 # CONFIG_HWMON is not set
@@ -69,22 +154,76 @@
 CONFIG_USB_OHCI_HCD_PLATFORM=y
 CONFIG_USB_STORAGE=y
 CONFIG_SOC_BRCMSTB=y
+CONFIG_MMC=y
+CONFIG_MMC_BLOCK_MINORS=16
+CONFIG_MMC_SDHCI=y
+CONFIG_MMC_SDHCI_PLTFM=y
 CONFIG_EXT4_FS=y
 CONFIG_EXT4_FS_POSIX_ACL=y
 CONFIG_EXT4_FS_SECURITY=y
 # CONFIG_DNOTIFY is not set
-CONFIG_FUSE_FS=y
-CONFIG_VFAT_FS=y
 CONFIG_PROC_KCORE=y
-CONFIG_TMPFS=y
-CONFIG_NFS_FS=y
 CONFIG_CIFS=y
+CONFIG_JBD2_DEBUG=y
+CONFIG_FUSE_FS=y
+CONFIG_FHANDLE=y
+CONFIG_CGROUPS=y
+CONFIG_CUSE=y
+CONFIG_ISO9660_FS=y
+CONFIG_JOLIET=y
+CONFIG_ZISOFS=y
+CONFIG_UDF_FS=y
+CONFIG_MSDOS_FS=y
+CONFIG_VFAT_FS=y
+CONFIG_TMPFS=y
+CONFIG_JFFS2_FS=y
+CONFIG_UBIFS_FS=y
+CONFIG_SQUASHFS=y
+CONFIG_SQUASHFS_LZO=y
+CONFIG_SQUASHFS_XZ=y
+CONFIG_NFS_FS=y
+CONFIG_NFS_V3_ACL=y
+CONFIG_NFS_V4=y
+CONFIG_NFS_V4_1=y
+CONFIG_NFS_V4_2=y
+CONFIG_ROOT_NFS=y
 CONFIG_NLS_CODEPAGE_437=y
-CONFIG_NLS_ASCII=y
 CONFIG_NLS_ISO8859_1=y
-# CONFIG_CRYPTO_HW is not set
 CONFIG_PRINTK_TIME=y
+CONFIG_DYNAMIC_DEBUG=y
+# CONFIG_DEBUG_INFO is not set
+# CONFIG_DEBUG_INFO_REDUCED is not set
 CONFIG_DEBUG_FS=y
 CONFIG_MAGIC_SYSRQ=y
+CONFIG_LOCKUP_DETECTOR=y
+CONFIG_DEBUG_USER=y
 CONFIG_CMDLINE_BOOL=y
 CONFIG_CMDLINE="earlycon"
+# CONFIG_MIPS_CMDLINE_FROM_DTB is not set
+CONFIG_MIPS_CMDLINE_DTB_EXTEND=y
+# CONFIG_MIPS_CMDLINE_FROM_BOOTLOADER is not set
+# CONFIG_CRYPTO_HW is not set
+CONFIG_DT_BCM974XX=y
+CONFIG_FW_CFE=y
+CONFIG_ATA=y
+CONFIG_SATA_AHCI_PLATFORM=y
+CONFIG_AHCI_BRCMSTB=y
+CONFIG_GENERIC_PHY=y
+CONFIG_GPIOLIB=y
+CONFIG_GPIO_SYSFS=y
+CONFIG_PHY_BRCM_USB=y
+CONFIG_PHY_BRCM_SATA=y
+CONFIG_PM_RUNTIME=y
+CONFIG_PM_DEBUG=y
+CONFIG_SYSVIPC=y
+CONFIG_FUNCTION_GRAPH_TRACER=y
+CONFIG_DYNAMIC_FTRACE=y
+CONFIG_FUNCTION_TRACER=y
+CONFIG_FUNCTION_PROFILER=y
+CONFIG_IRQSOFF_TRACER=y
+CONFIG_SCHED_TRACER=y
+CONFIG_BLK_DEV_IO_TRACE=y
+CONFIG_FTRACE_SYSCALLS=y
+CONFIG_TRACER_SNAPSHOT=y
+CONFIG_TRACER_SNAPSHOT_PER_CPU_SWAP=y
+CONFIG_STACK_TRACER=y
diff --git a/arch/mips/dec/setup.c b/arch/mips/dec/setup.c
index eaad0ed..a8a30bb 100644
--- a/arch/mips/dec/setup.c
+++ b/arch/mips/dec/setup.c
@@ -117,21 +117,21 @@ static void __init dec_be_init(void)
 {
 	switch (mips_machtype) {
 	case MACH_DS23100:	/* DS2100/DS3100 Pmin/Pmax */
-		board_be_handler = dec_kn01_be_handler;
+		mips_set_be_handler(dec_kn01_be_handler);
 		busirq_handler = dec_kn01_be_interrupt;
 		busirq_flags |= IRQF_SHARED;
 		dec_kn01_be_init();
 		break;
 	case MACH_DS5000_1XX:	/* DS5000/1xx 3min */
 	case MACH_DS5000_XX:	/* DS5000/xx Maxine */
-		board_be_handler = dec_kn02xa_be_handler;
+		mips_set_be_handler(dec_kn02xa_be_handler);
 		busirq_handler = dec_kn02xa_be_interrupt;
 		dec_kn02xa_be_init();
 		break;
 	case MACH_DS5000_200:	/* DS5000/200 3max */
 	case MACH_DS5000_2X0:	/* DS5000/240 3max+ */
 	case MACH_DS5900:	/* DS5900 bigmax */
-		board_be_handler = dec_ecc_be_handler;
+		mips_set_be_handler(dec_ecc_be_handler);
 		busirq_handler = dec_ecc_be_interrupt;
 		dec_ecc_be_init();
 		break;
diff --git a/arch/mips/include/asm/traps.h b/arch/mips/include/asm/traps.h
index b710e76..15cde63 100644
--- a/arch/mips/include/asm/traps.h
+++ b/arch/mips/include/asm/traps.h
@@ -15,7 +15,7 @@
 #define MIPS_BE_FATAL	2		/* treat as an unrecoverable error */
 
 extern void (*board_be_init)(void);
-extern int (*board_be_handler)(struct pt_regs *regs, int is_fixup);
+void mips_set_be_handler(int (*handler)(struct pt_regs *reg, int is_fixup));
 
 extern void (*board_nmi_handler_setup)(void);
 extern void (*board_ejtag_handler_setup)(void);
diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c
index 6f07362..d26b0fb 100644
--- a/arch/mips/kernel/traps.c
+++ b/arch/mips/kernel/traps.c
@@ -103,13 +103,19 @@ extern asmlinkage void handle_reserved(void);
 extern void tlb_do_page_fault_0(void);
 
 void (*board_be_init)(void);
-int (*board_be_handler)(struct pt_regs *regs, int is_fixup);
+static int (*board_be_handler)(struct pt_regs *regs, int is_fixup);
 void (*board_nmi_handler_setup)(void);
 void (*board_ejtag_handler_setup)(void);
 void (*board_bind_eic_interrupt)(int irq, int regset);
 void (*board_ebase_setup)(void);
 void(*board_cache_error_setup)(void);
 
+void mips_set_be_handler(int (*handler)(struct pt_regs *regs, int is_fixup))
+{
+	board_be_handler = handler;
+}
+EXPORT_SYMBOL_GPL(mips_set_be_handler);
+
 static void show_raw_backtrace(unsigned long reg29, const char *loglvl,
 			       bool user)
 {
diff --git a/arch/mips/sgi-ip22/ip22-berr.c b/arch/mips/sgi-ip22/ip22-berr.c
index dc0110a..afe8a61 100644
--- a/arch/mips/sgi-ip22/ip22-berr.c
+++ b/arch/mips/sgi-ip22/ip22-berr.c
@@ -112,5 +112,5 @@ static int ip22_be_handler(struct pt_regs *regs, int is_fixup)
 
 void __init ip22_be_init(void)
 {
-	board_be_handler = ip22_be_handler;
+	mips_set_be_handler(ip22_be_handler);
 }
diff --git a/arch/mips/sgi-ip22/ip28-berr.c b/arch/mips/sgi-ip22/ip28-berr.c
index c61362d..16ca470 100644
--- a/arch/mips/sgi-ip22/ip28-berr.c
+++ b/arch/mips/sgi-ip22/ip28-berr.c
@@ -468,7 +468,7 @@ static int ip28_be_handler(struct pt_regs *regs, int is_fixup)
 
 void __init ip22_be_init(void)
 {
-	board_be_handler = ip28_be_handler;
+	mips_set_be_handler(ip28_be_handler);
 }
 
 int ip28_show_be_info(struct seq_file *m)
diff --git a/arch/mips/sgi-ip27/ip27-berr.c b/arch/mips/sgi-ip27/ip27-berr.c
index 5a38ae6..923a63a 100644
--- a/arch/mips/sgi-ip27/ip27-berr.c
+++ b/arch/mips/sgi-ip27/ip27-berr.c
@@ -85,7 +85,7 @@ void __init ip27_be_init(void)
 	int cpu = LOCAL_HUB_L(PI_CPU_NUM);
 	int cpuoff = cpu << 8;
 
-	board_be_handler = ip27_be_handler;
+	mips_set_be_handler(ip27_be_handler);
 
 	LOCAL_HUB_S(PI_ERR_INT_PEND,
 		    cpu ? PI_ERR_CLEAR_ALL_B : PI_ERR_CLEAR_ALL_A);
diff --git a/arch/mips/sgi-ip32/ip32-berr.c b/arch/mips/sgi-ip32/ip32-berr.c
index c860f95a..478b63b 100644
--- a/arch/mips/sgi-ip32/ip32-berr.c
+++ b/arch/mips/sgi-ip32/ip32-berr.c
@@ -34,5 +34,5 @@ static int ip32_be_handler(struct pt_regs *regs, int is_fixup)
 
 void __init ip32_be_init(void)
 {
-	board_be_handler = ip32_be_handler;
+	mips_set_be_handler(ip32_be_handler);
 }
diff --git a/arch/mips/sibyte/swarm/setup.c b/arch/mips/sibyte/swarm/setup.c
index f07b15d..72a31ee 100644
--- a/arch/mips/sibyte/swarm/setup.c
+++ b/arch/mips/sibyte/swarm/setup.c
@@ -122,7 +122,7 @@ void __init plat_mem_setup(void)
 #error invalid SiByte board configuration
 #endif
 
-	board_be_handler = swarm_be_handler;
+	mips_set_be_handler(swarm_be_handler);
 
 	if (xicor_probe())
 		swarm_rtc_type = RTC_XICOR;
diff --git a/arch/mips/txx9/generic/setup_tx4927.c b/arch/mips/txx9/generic/setup_tx4927.c
index 46e9c41..63f9725 100644
--- a/arch/mips/txx9/generic/setup_tx4927.c
+++ b/arch/mips/txx9/generic/setup_tx4927.c
@@ -80,7 +80,7 @@ static int tx4927_be_handler(struct pt_regs *regs, int is_fixup)
 }
 static void __init tx4927_be_init(void)
 {
-	board_be_handler = tx4927_be_handler;
+	mips_set_be_handler(tx4927_be_handler);
 }
 
 static struct resource tx4927_sdram_resource[4];
diff --git a/arch/mips/txx9/generic/setup_tx4938.c b/arch/mips/txx9/generic/setup_tx4938.c
index 17395d5..ba64654 100644
--- a/arch/mips/txx9/generic/setup_tx4938.c
+++ b/arch/mips/txx9/generic/setup_tx4938.c
@@ -82,7 +82,7 @@ static int tx4938_be_handler(struct pt_regs *regs, int is_fixup)
 }
 static void __init tx4938_be_init(void)
 {
-	board_be_handler = tx4938_be_handler;
+	mips_set_be_handler(tx4938_be_handler);
 }
 
 static struct resource tx4938_sdram_resource[4];
diff --git a/arch/mips/txx9/generic/setup_tx4939.c b/arch/mips/txx9/generic/setup_tx4939.c
index bf8a3cd..f5f59b7 100644
--- a/arch/mips/txx9/generic/setup_tx4939.c
+++ b/arch/mips/txx9/generic/setup_tx4939.c
@@ -86,7 +86,7 @@ static int tx4939_be_handler(struct pt_regs *regs, int is_fixup)
 }
 static void __init tx4939_be_init(void)
 {
-	board_be_handler = tx4939_be_handler;
+	mips_set_be_handler(tx4939_be_handler);
 }
 
 static struct resource tx4939_sdram_resource[4];
diff --git a/arch/mips/vdso/Makefile b/arch/mips/vdso/Makefile
index 1b2ea34..d65f55f 100644
--- a/arch/mips/vdso/Makefile
+++ b/arch/mips/vdso/Makefile
@@ -57,7 +57,7 @@
 
 # VDSO linker flags.
 ldflags-y := -Bsymbolic --no-undefined -soname=linux-vdso.so.1 \
-	$(filter -E%,$(KBUILD_CFLAGS)) -nostdlib -shared \
+	$(filter -E%,$(KBUILD_CFLAGS)) -shared \
 	-G 0 --eh-frame-hdr --hash-style=sysv --build-id=sha1 -T
 
 CFLAGS_REMOVE_vdso.o = $(CC_FLAGS_FTRACE)
diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
index a34c531..821252b 100644
--- a/arch/riscv/Kconfig
+++ b/arch/riscv/Kconfig
@@ -62,6 +62,7 @@
 	select GENERIC_SCHED_CLOCK
 	select GENERIC_SMP_IDLE_THREAD
 	select GENERIC_TIME_VSYSCALL if MMU && 64BIT
+	select GENERIC_VDSO_TIME_NS if HAVE_GENERIC_VDSO
 	select HAVE_ARCH_AUDITSYSCALL
 	select HAVE_ARCH_JUMP_LABEL if !XIP_KERNEL
 	select HAVE_ARCH_JUMP_LABEL_RELATIVE if !XIP_KERNEL
diff --git a/arch/riscv/Makefile b/arch/riscv/Makefile
index 7f19b78..5927c94 100644
--- a/arch/riscv/Makefile
+++ b/arch/riscv/Makefile
@@ -136,3 +136,13 @@
 install zinstall:
 	$(CONFIG_SHELL) $(srctree)/$(boot)/install.sh $(KERNELRELEASE) \
 	$(boot)/$(install-image) System.map "$(INSTALL_PATH)"
+
+PHONY += rv32_randconfig
+rv32_randconfig:
+	$(Q)$(MAKE) KCONFIG_ALLCONFIG=$(srctree)/arch/riscv/configs/32-bit.config \
+		-f $(srctree)/Makefile randconfig
+
+PHONY += rv64_randconfig
+rv64_randconfig:
+	$(Q)$(MAKE) KCONFIG_ALLCONFIG=$(srctree)/arch/riscv/configs/64-bit.config \
+		-f $(srctree)/Makefile randconfig
diff --git a/arch/riscv/boot/dts/microchip/microchip-mpfs-icicle-kit.dts b/arch/riscv/boot/dts/microchip/microchip-mpfs-icicle-kit.dts
index b254c60..fc1e586 100644
--- a/arch/riscv/boot/dts/microchip/microchip-mpfs-icicle-kit.dts
+++ b/arch/riscv/boot/dts/microchip/microchip-mpfs-icicle-kit.dts
@@ -9,10 +9,8 @@
 #define RTCCLK_FREQ		1000000
 
 / {
-	#address-cells = <2>;
-	#size-cells = <2>;
 	model = "Microchip PolarFire-SoC Icicle Kit";
-	compatible = "microchip,mpfs-icicle-kit";
+	compatible = "microchip,mpfs-icicle-kit", "microchip,mpfs";
 
 	aliases {
 		ethernet0 = &emac1;
@@ -35,9 +33,6 @@ memory@80000000 {
 		reg = <0x0 0x80000000 0x0 0x40000000>;
 		clocks = <&clkcfg 26>;
 	};
-
-	soc {
-	};
 };
 
 &serial0 {
@@ -56,8 +51,17 @@ &serial3 {
 	status = "okay";
 };
 
-&sdcard {
+&mmc {
 	status = "okay";
+
+	bus-width = <4>;
+	disable-wp;
+	cap-sd-highspeed;
+	card-detect-delay = <200>;
+	sd-uhs-sdr12;
+	sd-uhs-sdr25;
+	sd-uhs-sdr50;
+	sd-uhs-sdr104;
 };
 
 &emac0 {
diff --git a/arch/riscv/boot/dts/microchip/microchip-mpfs.dtsi b/arch/riscv/boot/dts/microchip/microchip-mpfs.dtsi
index 9d2fbbc..c9f6d20 100644
--- a/arch/riscv/boot/dts/microchip/microchip-mpfs.dtsi
+++ b/arch/riscv/boot/dts/microchip/microchip-mpfs.dtsi
@@ -6,8 +6,8 @@
 / {
 	#address-cells = <2>;
 	#size-cells = <2>;
-	model = "Microchip MPFS Icicle Kit";
-	compatible = "microchip,mpfs-icicle-kit";
+	model = "Microchip PolarFire SoC";
+	compatible = "microchip,mpfs";
 
 	chosen {
 	};
@@ -161,7 +161,7 @@ cache-controller@2010000 {
 		};
 
 		clint@2000000 {
-			compatible = "sifive,clint0";
+			compatible = "sifive,fu540-c000-clint", "sifive,clint0";
 			reg = <0x0 0x2000000 0x0 0xC000>;
 			interrupts-extended = <&cpu0_intc 3 &cpu0_intc 7
 						&cpu1_intc 3 &cpu1_intc 7
@@ -172,7 +172,7 @@ &cpu3_intc 3 &cpu3_intc 7
 
 		plic: interrupt-controller@c000000 {
 			#interrupt-cells = <1>;
-			compatible = "sifive,plic-1.0.0";
+			compatible = "sifive,fu540-c000-plic", "sifive,plic-1.0.0";
 			reg = <0x0 0xc000000 0x0 0x4000000>;
 			riscv,ndev = <186>;
 			interrupt-controller;
@@ -262,39 +262,13 @@ serial3: serial@20104000 {
 			status = "disabled";
 		};
 
-		emmc: mmc@20008000 {
-			compatible = "cdns,sd4hc";
+		/* Common node entry for emmc/sd */
+		mmc: mmc@20008000 {
+			compatible = "microchip,mpfs-sd4hc", "cdns,sd4hc";
 			reg = <0x0 0x20008000 0x0 0x1000>;
 			interrupt-parent = <&plic>;
 			interrupts = <88 89>;
-			pinctrl-names = "default";
 			clocks = <&clkcfg 6>;
-			bus-width = <4>;
-			cap-mmc-highspeed;
-			mmc-ddr-3_3v;
-			max-frequency = <200000000>;
-			non-removable;
-			no-sd;
-			no-sdio;
-			voltage-ranges = <3300 3300>;
-			status = "disabled";
-		};
-
-		sdcard: sdhc@20008000 {
-			compatible = "cdns,sd4hc";
-			reg = <0x0 0x20008000 0x0 0x1000>;
-			interrupt-parent = <&plic>;
-			interrupts = <88>;
-			pinctrl-names = "default";
-			clocks = <&clkcfg 6>;
-			bus-width = <4>;
-			disable-wp;
-			cap-sd-highspeed;
-			card-detect-delay = <200>;
-			sd-uhs-sdr12;
-			sd-uhs-sdr25;
-			sd-uhs-sdr50;
-			sd-uhs-sdr104;
 			max-frequency = <200000000>;
 			status = "disabled";
 		};
diff --git a/arch/riscv/boot/dts/sifive/fu540-c000.dtsi b/arch/riscv/boot/dts/sifive/fu540-c000.dtsi
index 7db8610..0655b5c 100644
--- a/arch/riscv/boot/dts/sifive/fu540-c000.dtsi
+++ b/arch/riscv/boot/dts/sifive/fu540-c000.dtsi
@@ -141,7 +141,7 @@ soc {
 		ranges;
 		plic0: interrupt-controller@c000000 {
 			#interrupt-cells = <1>;
-			compatible = "sifive,plic-1.0.0";
+			compatible = "sifive,fu540-c000-plic", "sifive,plic-1.0.0";
 			reg = <0x0 0xc000000 0x0 0x4000000>;
 			riscv,ndev = <53>;
 			interrupt-controller;
diff --git a/arch/riscv/boot/dts/sifive/hifive-unleashed-a00.dts b/arch/riscv/boot/dts/sifive/hifive-unleashed-a00.dts
index 60846e8..ba304d4 100644
--- a/arch/riscv/boot/dts/sifive/hifive-unleashed-a00.dts
+++ b/arch/riscv/boot/dts/sifive/hifive-unleashed-a00.dts
@@ -8,10 +8,9 @@
 #define RTCCLK_FREQ		1000000
 
 / {
-	#address-cells = <2>;
-	#size-cells = <2>;
 	model = "SiFive HiFive Unleashed A00";
-	compatible = "sifive,hifive-unleashed-a00", "sifive,fu540-c000";
+	compatible = "sifive,hifive-unleashed-a00", "sifive,fu540-c000",
+		     "sifive,fu540";
 
 	chosen {
 		stdout-path = "serial0";
@@ -26,9 +25,6 @@ memory@80000000 {
 		reg = <0x0 0x80000000 0x2 0x00000000>;
 	};
 
-	soc {
-	};
-
 	hfclk: hfclk {
 		#clock-cells = <0>;
 		compatible = "fixed-clock";
@@ -63,7 +59,7 @@ &i2c0 {
 &qspi0 {
 	status = "okay";
 	flash@0 {
-		compatible = "issi,is25wp256", "jedec,spi-nor";
+		compatible = "jedec,spi-nor";
 		reg = <0>;
 		spi-max-frequency = <50000000>;
 		m25p,fast-read;
diff --git a/arch/riscv/boot/dts/sifive/hifive-unmatched-a00.dts b/arch/riscv/boot/dts/sifive/hifive-unmatched-a00.dts
index 2e4ea84..4f66919 100644
--- a/arch/riscv/boot/dts/sifive/hifive-unmatched-a00.dts
+++ b/arch/riscv/boot/dts/sifive/hifive-unmatched-a00.dts
@@ -8,8 +8,6 @@
 #define RTCCLK_FREQ		1000000
 
 / {
-	#address-cells = <2>;
-	#size-cells = <2>;
 	model = "SiFive HiFive Unmatched A00";
 	compatible = "sifive,hifive-unmatched-a00", "sifive,fu740-c000",
 		     "sifive,fu740";
@@ -27,9 +25,6 @@ memory@80000000 {
 		reg = <0x0 0x80000000 0x4 0x00000000>;
 	};
 
-	soc {
-	};
-
 	hfclk: hfclk {
 		#clock-cells = <0>;
 		compatible = "fixed-clock";
@@ -211,7 +206,7 @@ vdd_ldo11: ldo11 {
 &qspi0 {
 	status = "okay";
 	flash@0 {
-		compatible = "issi,is25wp256", "jedec,spi-nor";
+		compatible = "jedec,spi-nor";
 		reg = <0>;
 		spi-max-frequency = <50000000>;
 		m25p,fast-read;
diff --git a/arch/riscv/configs/32-bit.config b/arch/riscv/configs/32-bit.config
new file mode 100644
index 0000000..43f4132
--- /dev/null
+++ b/arch/riscv/configs/32-bit.config
@@ -0,0 +1,2 @@
+CONFIG_ARCH_RV32I=y
+CONFIG_32BIT=y
diff --git a/arch/riscv/configs/64-bit.config b/arch/riscv/configs/64-bit.config
new file mode 100644
index 0000000..313edc5
--- /dev/null
+++ b/arch/riscv/configs/64-bit.config
@@ -0,0 +1,2 @@
+CONFIG_ARCH_RV64I=y
+CONFIG_64BIT=y
diff --git a/arch/riscv/configs/defconfig b/arch/riscv/configs/defconfig
index 4ebc803..c252fd5 100644
--- a/arch/riscv/configs/defconfig
+++ b/arch/riscv/configs/defconfig
@@ -72,9 +72,10 @@
 CONFIG_GPIO_SIFIVE=y
 # CONFIG_PTP_1588_CLOCK is not set
 CONFIG_POWER_RESET=y
-CONFIG_DRM=y
-CONFIG_DRM_RADEON=y
-CONFIG_DRM_VIRTIO_GPU=y
+CONFIG_DRM=m
+CONFIG_DRM_RADEON=m
+CONFIG_DRM_NOUVEAU=m
+CONFIG_DRM_VIRTIO_GPU=m
 CONFIG_FRAMEBUFFER_CONSOLE=y
 CONFIG_USB=y
 CONFIG_USB_XHCI_HCD=y
diff --git a/arch/riscv/include/asm/page.h b/arch/riscv/include/asm/page.h
index 109c97e..b3e5ff0 100644
--- a/arch/riscv/include/asm/page.h
+++ b/arch/riscv/include/asm/page.h
@@ -157,6 +157,8 @@ extern phys_addr_t __phys_addr_symbol(unsigned long x);
 #define page_to_bus(page)	(page_to_phys(page))
 #define phys_to_page(paddr)	(pfn_to_page(phys_to_pfn(paddr)))
 
+#define sym_to_pfn(x)           __phys_to_pfn(__pa_symbol(x))
+
 #ifdef CONFIG_FLATMEM
 #define pfn_valid(pfn) \
 	(((pfn) >= ARCH_PFN_OFFSET) && (((pfn) - ARCH_PFN_OFFSET) < max_mapnr))
diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
index 39b5503..bf204e7 100644
--- a/arch/riscv/include/asm/pgtable.h
+++ b/arch/riscv/include/asm/pgtable.h
@@ -75,7 +75,8 @@
 #endif
 
 #ifdef CONFIG_XIP_KERNEL
-#define XIP_OFFSET		SZ_8M
+#define XIP_OFFSET		SZ_32M
+#define XIP_OFFSET_MASK		(SZ_32M - 1)
 #else
 #define XIP_OFFSET		0
 #endif
@@ -97,7 +98,8 @@
 #ifdef CONFIG_XIP_KERNEL
 #define XIP_FIXUP(addr) ({							\
 	uintptr_t __a = (uintptr_t)(addr);					\
-	(__a >= CONFIG_XIP_PHYS_ADDR && __a < CONFIG_XIP_PHYS_ADDR + SZ_16M) ?	\
+	(__a >= CONFIG_XIP_PHYS_ADDR && \
+	 __a < CONFIG_XIP_PHYS_ADDR + XIP_OFFSET * 2) ?	\
 		__a - CONFIG_XIP_PHYS_ADDR + CONFIG_PHYS_RAM_BASE - XIP_OFFSET :\
 		__a;								\
 	})
diff --git a/arch/riscv/include/asm/vdso.h b/arch/riscv/include/asm/vdso.h
index 208e31b..bc6f75f 100644
--- a/arch/riscv/include/asm/vdso.h
+++ b/arch/riscv/include/asm/vdso.h
@@ -8,30 +8,19 @@
 #ifndef _ASM_RISCV_VDSO_H
 #define _ASM_RISCV_VDSO_H
 
-
 /*
  * All systems with an MMU have a VDSO, but systems without an MMU don't
  * support shared libraries and therefor don't have one.
  */
 #ifdef CONFIG_MMU
 
-#include <linux/types.h>
-/*
- * All systems with an MMU have a VDSO, but systems without an MMU don't
- * support shared libraries and therefor don't have one.
- */
-#ifdef CONFIG_MMU
-
-#define __VVAR_PAGES    1
+#define __VVAR_PAGES    2
 
 #ifndef __ASSEMBLY__
 #include <generated/vdso-offsets.h>
 
 #define VDSO_SYMBOL(base, name)							\
 	(void __user *)((unsigned long)(base) + __vdso_##name##_offset)
-
-#endif /* CONFIG_MMU */
-
 #endif /* !__ASSEMBLY__ */
 
 #endif /* CONFIG_MMU */
diff --git a/arch/riscv/include/asm/vdso/gettimeofday.h b/arch/riscv/include/asm/vdso/gettimeofday.h
index f839f16..77d9c2f 100644
--- a/arch/riscv/include/asm/vdso/gettimeofday.h
+++ b/arch/riscv/include/asm/vdso/gettimeofday.h
@@ -76,6 +76,13 @@ static __always_inline const struct vdso_data *__arch_get_vdso_data(void)
 	return _vdso_data;
 }
 
+#ifdef CONFIG_TIME_NS
+static __always_inline
+const struct vdso_data *__arch_get_timens_vdso_data(const struct vdso_data *vd)
+{
+	return _timens_data;
+}
+#endif
 #endif /* !__ASSEMBLY__ */
 
 #endif /* __ASM_VDSO_GETTIMEOFDAY_H */
diff --git a/arch/riscv/kernel/head.S b/arch/riscv/kernel/head.S
index 25ec505..f52f01e 100644
--- a/arch/riscv/kernel/head.S
+++ b/arch/riscv/kernel/head.S
@@ -20,10 +20,20 @@
 	REG_L t0, _xip_fixup
 	add \reg, \reg, t0
 .endm
+.macro XIP_FIXUP_FLASH_OFFSET reg
+	la t1, __data_loc
+	li t0, XIP_OFFSET_MASK
+	and t1, t1, t0
+	li t1, XIP_OFFSET
+	sub t0, t0, t1
+	sub \reg, \reg, t0
+.endm
 _xip_fixup: .dword CONFIG_PHYS_RAM_BASE - CONFIG_XIP_PHYS_ADDR - XIP_OFFSET
 #else
 .macro XIP_FIXUP_OFFSET reg
 .endm
+.macro XIP_FIXUP_FLASH_OFFSET reg
+.endm
 #endif /* CONFIG_XIP_KERNEL */
 
 __HEAD
@@ -267,6 +277,7 @@
 	la a3, hart_lottery
 	mv a2, a3
 	XIP_FIXUP_OFFSET a2
+	XIP_FIXUP_FLASH_OFFSET a3
 	lw t1, (a3)
 	amoswap.w t0, t1, (a2)
 	/* first time here if hart_lottery in RAM is not set */
@@ -305,6 +316,7 @@
 	XIP_FIXUP_OFFSET sp
 #ifdef CONFIG_BUILTIN_DTB
 	la a0, __dtb_start
+	XIP_FIXUP_OFFSET a0
 #else
 	mv a0, s1
 #endif /* CONFIG_BUILTIN_DTB */
diff --git a/arch/riscv/kernel/reset.c b/arch/riscv/kernel/reset.c
index ee5878d..9c842c4 100644
--- a/arch/riscv/kernel/reset.c
+++ b/arch/riscv/kernel/reset.c
@@ -12,7 +12,7 @@ static void default_power_off(void)
 		wait_for_interrupt();
 }
 
-void (*pm_power_off)(void) = default_power_off;
+void (*pm_power_off)(void) = NULL;
 EXPORT_SYMBOL(pm_power_off);
 
 void machine_restart(char *cmd)
@@ -23,10 +23,16 @@ void machine_restart(char *cmd)
 
 void machine_halt(void)
 {
-	pm_power_off();
+	if (pm_power_off != NULL)
+		pm_power_off();
+	else
+		default_power_off();
 }
 
 void machine_power_off(void)
 {
-	pm_power_off();
+	if (pm_power_off != NULL)
+		pm_power_off();
+	else
+		default_power_off();
 }
diff --git a/arch/riscv/kernel/vdso.c b/arch/riscv/kernel/vdso.c
index b70956d..a9436a6 100644
--- a/arch/riscv/kernel/vdso.c
+++ b/arch/riscv/kernel/vdso.c
@@ -13,6 +13,7 @@
 #include <linux/err.h>
 #include <asm/page.h>
 #include <asm/vdso.h>
+#include <linux/time_namespace.h>
 
 #ifdef CONFIG_GENERIC_TIME_VSYSCALL
 #include <vdso/datapage.h>
@@ -25,14 +26,12 @@ extern char vdso_start[], vdso_end[];
 
 enum vvar_pages {
 	VVAR_DATA_PAGE_OFFSET,
+	VVAR_TIMENS_PAGE_OFFSET,
 	VVAR_NR_PAGES,
 };
 
 #define VVAR_SIZE  (VVAR_NR_PAGES << PAGE_SHIFT)
 
-static unsigned int vdso_pages __ro_after_init;
-static struct page **vdso_pagelist __ro_after_init;
-
 /*
  * The vDSO data page.
  */
@@ -42,83 +41,228 @@ static union {
 } vdso_data_store __page_aligned_data;
 struct vdso_data *vdso_data = &vdso_data_store.data;
 
-static int __init vdso_init(void)
+struct __vdso_info {
+	const char *name;
+	const char *vdso_code_start;
+	const char *vdso_code_end;
+	unsigned long vdso_pages;
+	/* Data Mapping */
+	struct vm_special_mapping *dm;
+	/* Code Mapping */
+	struct vm_special_mapping *cm;
+};
+
+static struct __vdso_info vdso_info __ro_after_init = {
+	.name = "vdso",
+	.vdso_code_start = vdso_start,
+	.vdso_code_end = vdso_end,
+};
+
+static int vdso_mremap(const struct vm_special_mapping *sm,
+		       struct vm_area_struct *new_vma)
 {
-	unsigned int i;
-
-	vdso_pages = (vdso_end - vdso_start) >> PAGE_SHIFT;
-	vdso_pagelist =
-		kcalloc(vdso_pages + VVAR_NR_PAGES, sizeof(struct page *), GFP_KERNEL);
-	if (unlikely(vdso_pagelist == NULL)) {
-		pr_err("vdso: pagelist allocation failed\n");
-		return -ENOMEM;
-	}
-
-	for (i = 0; i < vdso_pages; i++) {
-		struct page *pg;
-
-		pg = virt_to_page(vdso_start + (i << PAGE_SHIFT));
-		vdso_pagelist[i] = pg;
-	}
-	vdso_pagelist[i] = virt_to_page(vdso_data);
+	current->mm->context.vdso = (void *)new_vma->vm_start;
 
 	return 0;
 }
+
+static int __init __vdso_init(void)
+{
+	unsigned int i;
+	struct page **vdso_pagelist;
+	unsigned long pfn;
+
+	if (memcmp(vdso_info.vdso_code_start, "\177ELF", 4)) {
+		pr_err("vDSO is not a valid ELF object!\n");
+		return -EINVAL;
+	}
+
+	vdso_info.vdso_pages = (
+		vdso_info.vdso_code_end -
+		vdso_info.vdso_code_start) >>
+		PAGE_SHIFT;
+
+	vdso_pagelist = kcalloc(vdso_info.vdso_pages,
+				sizeof(struct page *),
+				GFP_KERNEL);
+	if (vdso_pagelist == NULL)
+		return -ENOMEM;
+
+	/* Grab the vDSO code pages. */
+	pfn = sym_to_pfn(vdso_info.vdso_code_start);
+
+	for (i = 0; i < vdso_info.vdso_pages; i++)
+		vdso_pagelist[i] = pfn_to_page(pfn + i);
+
+	vdso_info.cm->pages = vdso_pagelist;
+
+	return 0;
+}
+
+#ifdef CONFIG_TIME_NS
+struct vdso_data *arch_get_vdso_data(void *vvar_page)
+{
+	return (struct vdso_data *)(vvar_page);
+}
+
+/*
+ * The vvar mapping contains data for a specific time namespace, so when a task
+ * changes namespace we must unmap its vvar data for the old namespace.
+ * Subsequent faults will map in data for the new namespace.
+ *
+ * For more details see timens_setup_vdso_data().
+ */
+int vdso_join_timens(struct task_struct *task, struct time_namespace *ns)
+{
+	struct mm_struct *mm = task->mm;
+	struct vm_area_struct *vma;
+
+	mmap_read_lock(mm);
+
+	for (vma = mm->mmap; vma; vma = vma->vm_next) {
+		unsigned long size = vma->vm_end - vma->vm_start;
+
+		if (vma_is_special_mapping(vma, vdso_info.dm))
+			zap_page_range(vma, vma->vm_start, size);
+	}
+
+	mmap_read_unlock(mm);
+	return 0;
+}
+
+static struct page *find_timens_vvar_page(struct vm_area_struct *vma)
+{
+	if (likely(vma->vm_mm == current->mm))
+		return current->nsproxy->time_ns->vvar_page;
+
+	/*
+	 * VM_PFNMAP | VM_IO protect .fault() handler from being called
+	 * through interfaces like /proc/$pid/mem or
+	 * process_vm_{readv,writev}() as long as there's no .access()
+	 * in special_mapping_vmops.
+	 * For more details check_vma_flags() and __access_remote_vm()
+	 */
+	WARN(1, "vvar_page accessed remotely");
+
+	return NULL;
+}
+#else
+static struct page *find_timens_vvar_page(struct vm_area_struct *vma)
+{
+	return NULL;
+}
+#endif
+
+static vm_fault_t vvar_fault(const struct vm_special_mapping *sm,
+			     struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	struct page *timens_page = find_timens_vvar_page(vma);
+	unsigned long pfn;
+
+	switch (vmf->pgoff) {
+	case VVAR_DATA_PAGE_OFFSET:
+		if (timens_page)
+			pfn = page_to_pfn(timens_page);
+		else
+			pfn = sym_to_pfn(vdso_data);
+		break;
+#ifdef CONFIG_TIME_NS
+	case VVAR_TIMENS_PAGE_OFFSET:
+		/*
+		 * If a task belongs to a time namespace then a namespace
+		 * specific VVAR is mapped with the VVAR_DATA_PAGE_OFFSET and
+		 * the real VVAR page is mapped with the VVAR_TIMENS_PAGE_OFFSET
+		 * offset.
+		 * See also the comment near timens_setup_vdso_data().
+		 */
+		if (!timens_page)
+			return VM_FAULT_SIGBUS;
+		pfn = sym_to_pfn(vdso_data);
+		break;
+#endif /* CONFIG_TIME_NS */
+	default:
+		return VM_FAULT_SIGBUS;
+	}
+
+	return vmf_insert_pfn(vma, vmf->address, pfn);
+}
+
+enum rv_vdso_map {
+	RV_VDSO_MAP_VVAR,
+	RV_VDSO_MAP_VDSO,
+};
+
+static struct vm_special_mapping rv_vdso_maps[] __ro_after_init = {
+	[RV_VDSO_MAP_VVAR] = {
+		.name   = "[vvar]",
+		.fault = vvar_fault,
+	},
+	[RV_VDSO_MAP_VDSO] = {
+		.name   = "[vdso]",
+		.mremap = vdso_mremap,
+	},
+};
+
+static int __init vdso_init(void)
+{
+	vdso_info.dm = &rv_vdso_maps[RV_VDSO_MAP_VVAR];
+	vdso_info.cm = &rv_vdso_maps[RV_VDSO_MAP_VDSO];
+
+	return __vdso_init();
+}
 arch_initcall(vdso_init);
 
-int arch_setup_additional_pages(struct linux_binprm *bprm,
-	int uses_interp)
+static int __setup_additional_pages(struct mm_struct *mm,
+				    struct linux_binprm *bprm,
+				    int uses_interp)
 {
-	struct mm_struct *mm = current->mm;
-	unsigned long vdso_base, vdso_len;
-	int ret;
+	unsigned long vdso_base, vdso_text_len, vdso_mapping_len;
+	void *ret;
 
 	BUILD_BUG_ON(VVAR_NR_PAGES != __VVAR_PAGES);
 
-	vdso_len = (vdso_pages + VVAR_NR_PAGES) << PAGE_SHIFT;
+	vdso_text_len = vdso_info.vdso_pages << PAGE_SHIFT;
+	/* Be sure to map the data page */
+	vdso_mapping_len = vdso_text_len + VVAR_SIZE;
+
+	vdso_base = get_unmapped_area(NULL, 0, vdso_mapping_len, 0, 0);
+	if (IS_ERR_VALUE(vdso_base)) {
+		ret = ERR_PTR(vdso_base);
+		goto up_fail;
+	}
+
+	ret = _install_special_mapping(mm, vdso_base, VVAR_SIZE,
+		(VM_READ | VM_MAYREAD | VM_PFNMAP), vdso_info.dm);
+	if (IS_ERR(ret))
+		goto up_fail;
+
+	vdso_base += VVAR_SIZE;
+	mm->context.vdso = (void *)vdso_base;
+	ret =
+	   _install_special_mapping(mm, vdso_base, vdso_text_len,
+		(VM_READ | VM_EXEC | VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC),
+		vdso_info.cm);
+
+	if (IS_ERR(ret))
+		goto up_fail;
+
+	return 0;
+
+up_fail:
+	mm->context.vdso = NULL;
+	return PTR_ERR(ret);
+}
+
+int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
+{
+	struct mm_struct *mm = current->mm;
+	int ret;
 
 	if (mmap_write_lock_killable(mm))
 		return -EINTR;
 
-	vdso_base = get_unmapped_area(NULL, 0, vdso_len, 0, 0);
-	if (IS_ERR_VALUE(vdso_base)) {
-		ret = vdso_base;
-		goto end;
-	}
-
-	mm->context.vdso = NULL;
-	ret = install_special_mapping(mm, vdso_base, VVAR_SIZE,
-		(VM_READ | VM_MAYREAD), &vdso_pagelist[vdso_pages]);
-	if (unlikely(ret))
-		goto end;
-
-	ret =
-	   install_special_mapping(mm, vdso_base + VVAR_SIZE,
-		vdso_pages << PAGE_SHIFT,
-		(VM_READ | VM_EXEC | VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC),
-		vdso_pagelist);
-
-	if (unlikely(ret))
-		goto end;
-
-	/*
-	 * Put vDSO base into mm struct. We need to do this before calling
-	 * install_special_mapping or the perf counter mmap tracking code
-	 * will fail to recognise it as a vDSO (since arch_vma_name fails).
-	 */
-	mm->context.vdso = (void *)vdso_base + VVAR_SIZE;
-
-end:
+	ret = __setup_additional_pages(mm, bprm, uses_interp);
 	mmap_write_unlock(mm);
-	return ret;
-}
 
-const char *arch_vma_name(struct vm_area_struct *vma)
-{
-	if (vma->vm_mm && (vma->vm_start == (long)vma->vm_mm->context.vdso))
-		return "[vdso]";
-	if (vma->vm_mm && (vma->vm_start ==
-			   (long)vma->vm_mm->context.vdso - VVAR_SIZE))
-		return "[vdso_data]";
-	return NULL;
+	return ret;
 }
diff --git a/arch/riscv/kernel/vdso/vdso.lds.S b/arch/riscv/kernel/vdso/vdso.lds.S
index e9111f7..01d94aa 100644
--- a/arch/riscv/kernel/vdso/vdso.lds.S
+++ b/arch/riscv/kernel/vdso/vdso.lds.S
@@ -10,6 +10,9 @@
 SECTIONS
 {
 	PROVIDE(_vdso_data = . - __VVAR_PAGES * PAGE_SIZE);
+#ifdef CONFIG_TIME_NS
+	PROVIDE(_timens_data = _vdso_data + PAGE_SIZE);
+#endif
 	. = SIZEOF_HEADERS;
 
 	.hash		: { *(.hash) }			:text
diff --git a/arch/riscv/kernel/vmlinux-xip.lds.S b/arch/riscv/kernel/vmlinux-xip.lds.S
index 9c9f350..f5ed082 100644
--- a/arch/riscv/kernel/vmlinux-xip.lds.S
+++ b/arch/riscv/kernel/vmlinux-xip.lds.S
@@ -64,8 +64,11 @@
 /*
  * From this point, stuff is considered writable and will be copied to RAM
  */
-	__data_loc = ALIGN(16);		/* location in file */
-	. = LOAD_OFFSET + XIP_OFFSET;	/* location in memory */
+	__data_loc = ALIGN(PAGE_SIZE);		/* location in file */
+	. = KERNEL_LINK_ADDR + XIP_OFFSET;	/* location in memory */
+
+#undef LOAD_OFFSET
+#define LOAD_OFFSET (KERNEL_LINK_ADDR + XIP_OFFSET - (__data_loc & XIP_OFFSET_MASK))
 
 	_sdata = .;			/* Start of data section */
 	_data = .;
@@ -96,7 +99,6 @@
 		KEEP(*(__soc_builtin_dtb_table))
 		__soc_builtin_dtb_table_end = .;
 	}
-	PERCPU_SECTION(L1_CACHE_BYTES)
 
 	. = ALIGN(8);
 	.alternative : {
@@ -122,6 +124,8 @@
 
 	BSS_SECTION(PAGE_SIZE, PAGE_SIZE, 0)
 
+	PERCPU_SECTION(L1_CACHE_BYTES)
+
 	.rel.dyn : AT(ADDR(.rel.dyn) - LOAD_OFFSET) {
 		*(.rel.dyn*)
 	}
diff --git a/arch/riscv/mm/context.c b/arch/riscv/mm/context.c
index ee3459c..ea54cc0 100644
--- a/arch/riscv/mm/context.c
+++ b/arch/riscv/mm/context.c
@@ -233,8 +233,10 @@ static int __init asids_init(void)
 	local_flush_tlb_all();
 
 	/* Pre-compute ASID details */
-	num_asids = 1 << asid_bits;
-	asid_mask = num_asids - 1;
+	if (asid_bits) {
+		num_asids = 1 << asid_bits;
+		asid_mask = num_asids - 1;
+	}
 
 	/*
 	 * Use ASID allocator only if number of HW ASIDs are
@@ -255,7 +257,7 @@ static int __init asids_init(void)
 		pr_info("ASID allocator using %lu bits (%lu entries)\n",
 			asid_bits, num_asids);
 	} else {
-		pr_info("ASID allocator disabled\n");
+		pr_info("ASID allocator disabled (%lu bits)\n", asid_bits);
 	}
 
 	return 0;
diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c
index c0cddf0..24b2b80 100644
--- a/arch/riscv/mm/init.c
+++ b/arch/riscv/mm/init.c
@@ -41,7 +41,7 @@ phys_addr_t phys_ram_base __ro_after_init;
 EXPORT_SYMBOL(phys_ram_base);
 
 #ifdef CONFIG_XIP_KERNEL
-extern char _xiprom[], _exiprom[];
+extern char _xiprom[], _exiprom[], __data_loc;
 #endif
 
 unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)]
@@ -454,10 +454,9 @@ static uintptr_t __init best_map_size(phys_addr_t base, phys_addr_t size)
 /* called from head.S with MMU off */
 asmlinkage void __init __copy_data(void)
 {
-	void *from = (void *)(&_sdata);
-	void *end = (void *)(&_end);
+	void *from = (void *)(&__data_loc);
 	void *to = (void *)CONFIG_PHYS_RAM_BASE;
-	size_t sz = (size_t)(end - from + 1);
+	size_t sz = (size_t)((uintptr_t)(&_end) - (uintptr_t)(&_sdata));
 
 	memcpy(to, from, sz);
 }
diff --git a/arch/s390/include/asm/pci.h b/arch/s390/include/asm/pci.h
index 6b3c366..90824be 100644
--- a/arch/s390/include/asm/pci.h
+++ b/arch/s390/include/asm/pci.h
@@ -210,9 +210,11 @@ int zpci_deconfigure_device(struct zpci_dev *zdev);
 void zpci_device_reserved(struct zpci_dev *zdev);
 bool zpci_is_device_configured(struct zpci_dev *zdev);
 
+int zpci_hot_reset_device(struct zpci_dev *zdev);
 int zpci_register_ioat(struct zpci_dev *, u8, u64, u64, u64);
 int zpci_unregister_ioat(struct zpci_dev *, u8);
 void zpci_remove_reserved_devices(void);
+void zpci_update_fh(struct zpci_dev *zdev, u32 fh);
 
 /* CLP */
 int clp_setup_writeback_mio(void);
@@ -294,8 +296,10 @@ void zpci_debug_exit(void);
 void zpci_debug_init_device(struct zpci_dev *, const char *);
 void zpci_debug_exit_device(struct zpci_dev *);
 
-/* Error reporting */
+/* Error handling */
 int zpci_report_error(struct pci_dev *, struct zpci_report_error_header *);
+int zpci_clear_error_state(struct zpci_dev *zdev);
+int zpci_reset_load_store_blocked(struct zpci_dev *zdev);
 
 #ifdef CONFIG_NUMA
 
diff --git a/arch/s390/kernel/perf_cpum_cf.c b/arch/s390/kernel/perf_cpum_cf.c
index 6f431fa..ee8707a 100644
--- a/arch/s390/kernel/perf_cpum_cf.c
+++ b/arch/s390/kernel/perf_cpum_cf.c
@@ -687,8 +687,10 @@ static void cpumf_pmu_stop(struct perf_event *event, int flags)
 						      false);
 			if (cfdiag_diffctr(cpuhw, event->hw.config_base))
 				cfdiag_push_sample(event, cpuhw);
-		} else
+		} else if (cpuhw->flags & PMU_F_RESERVED) {
+			/* Only update when PMU not hotplugged off */
 			hw_perf_event_update(event);
+		}
 		hwc->state |= PERF_HES_UPTODATE;
 	}
 }
diff --git a/arch/s390/pci/pci.c b/arch/s390/pci/pci.c
index 872d772..2f9b78f 100644
--- a/arch/s390/pci/pci.c
+++ b/arch/s390/pci/pci.c
@@ -481,6 +481,34 @@ static void zpci_free_iomap(struct zpci_dev *zdev, int entry)
 	spin_unlock(&zpci_iomap_lock);
 }
 
+static void zpci_do_update_iomap_fh(struct zpci_dev *zdev, u32 fh)
+{
+	int bar, idx;
+
+	spin_lock(&zpci_iomap_lock);
+	for (bar = 0; bar < PCI_STD_NUM_BARS; bar++) {
+		if (!zdev->bars[bar].size)
+			continue;
+		idx = zdev->bars[bar].map_idx;
+		if (!zpci_iomap_start[idx].count)
+			continue;
+		WRITE_ONCE(zpci_iomap_start[idx].fh, zdev->fh);
+	}
+	spin_unlock(&zpci_iomap_lock);
+}
+
+void zpci_update_fh(struct zpci_dev *zdev, u32 fh)
+{
+	if (!fh || zdev->fh == fh)
+		return;
+
+	zdev->fh = fh;
+	if (zpci_use_mio(zdev))
+		return;
+	if (zdev->has_resources && zdev_enabled(zdev))
+		zpci_do_update_iomap_fh(zdev, fh);
+}
+
 static struct resource *__alloc_res(struct zpci_dev *zdev, unsigned long start,
 				    unsigned long size, unsigned long flags)
 {
@@ -668,7 +696,7 @@ int zpci_enable_device(struct zpci_dev *zdev)
 	if (clp_enable_fh(zdev, &fh, ZPCI_NR_DMA_SPACES))
 		rc = -EIO;
 	else
-		zdev->fh = fh;
+		zpci_update_fh(zdev, fh);
 	return rc;
 }
 
@@ -679,14 +707,14 @@ int zpci_disable_device(struct zpci_dev *zdev)
 
 	cc = clp_disable_fh(zdev, &fh);
 	if (!cc) {
-		zdev->fh = fh;
+		zpci_update_fh(zdev, fh);
 	} else if (cc == CLP_RC_SETPCIFN_ALRDY) {
 		pr_info("Disabling PCI function %08x had no effect as it was already disabled\n",
 			zdev->fid);
 		/* Function is already disabled - update handle */
 		rc = clp_refresh_fh(zdev->fid, &fh);
 		if (!rc) {
-			zdev->fh = fh;
+			zpci_update_fh(zdev, fh);
 			rc = -EINVAL;
 		}
 	} else {
@@ -696,6 +724,65 @@ int zpci_disable_device(struct zpci_dev *zdev)
 }
 
 /**
+ * zpci_hot_reset_device - perform a reset of the given zPCI function
+ * @zdev: the slot which should be reset
+ *
+ * Performs a low level reset of the zPCI function. The reset is low level in
+ * the sense that the zPCI function can be reset without detaching it from the
+ * common PCI subsystem. The reset may be performed while under control of
+ * either DMA or IOMMU APIs in which case the existing DMA/IOMMU translation
+ * table is reinstated at the end of the reset.
+ *
+ * After the reset the functions internal state is reset to an initial state
+ * equivalent to its state during boot when first probing a driver.
+ * Consequently after reset the PCI function requires re-initialization via the
+ * common PCI code including re-enabling IRQs via pci_alloc_irq_vectors()
+ * and enabling the function via e.g.pci_enablde_device_flags().The caller
+ * must guard against concurrent reset attempts.
+ *
+ * In most cases this function should not be called directly but through
+ * pci_reset_function() or pci_reset_bus() which handle the save/restore and
+ * locking.
+ *
+ * Return: 0 on success and an error value otherwise
+ */
+int zpci_hot_reset_device(struct zpci_dev *zdev)
+{
+	int rc;
+
+	zpci_dbg(3, "rst fid:%x, fh:%x\n", zdev->fid, zdev->fh);
+	if (zdev_enabled(zdev)) {
+		/* Disables device access, DMAs and IRQs (reset state) */
+		rc = zpci_disable_device(zdev);
+		/*
+		 * Due to a z/VM vs LPAR inconsistency in the error state the
+		 * FH may indicate an enabled device but disable says the
+		 * device is already disabled don't treat it as an error here.
+		 */
+		if (rc == -EINVAL)
+			rc = 0;
+		if (rc)
+			return rc;
+	}
+
+	rc = zpci_enable_device(zdev);
+	if (rc)
+		return rc;
+
+	if (zdev->dma_table)
+		rc = zpci_register_ioat(zdev, 0, zdev->start_dma, zdev->end_dma,
+					(u64)zdev->dma_table);
+	else
+		rc = zpci_dma_init_device(zdev);
+	if (rc) {
+		zpci_disable_device(zdev);
+		return rc;
+	}
+
+	return 0;
+}
+
+/**
  * zpci_create_device() - Create a new zpci_dev and add it to the zbus
  * @fid: Function ID of the device to be created
  * @fh: Current Function Handle of the device to be created
@@ -776,7 +863,7 @@ int zpci_scan_configured_device(struct zpci_dev *zdev, u32 fh)
 {
 	int rc;
 
-	zdev->fh = fh;
+	zpci_update_fh(zdev, fh);
 	/* the PCI function will be scanned once function 0 appears */
 	if (!zdev->zbus->bus)
 		return 0;
@@ -903,6 +990,59 @@ int zpci_report_error(struct pci_dev *pdev,
 }
 EXPORT_SYMBOL(zpci_report_error);
 
+/**
+ * zpci_clear_error_state() - Clears the zPCI error state of the device
+ * @zdev: The zdev for which the zPCI error state should be reset
+ *
+ * Clear the zPCI error state of the device. If clearing the zPCI error state
+ * fails the device is left in the error state. In this case it may make sense
+ * to call zpci_io_perm_failure() on the associated pdev if it exists.
+ *
+ * Returns: 0 on success, -EIO otherwise
+ */
+int zpci_clear_error_state(struct zpci_dev *zdev)
+{
+	u64 req = ZPCI_CREATE_REQ(zdev->fh, 0, ZPCI_MOD_FC_RESET_ERROR);
+	struct zpci_fib fib = {0};
+	u8 status;
+	int cc;
+
+	cc = zpci_mod_fc(req, &fib, &status);
+	if (cc) {
+		zpci_dbg(3, "ces fid:%x, cc:%d, status:%x\n", zdev->fid, cc, status);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+/**
+ * zpci_reset_load_store_blocked() - Re-enables L/S from error state
+ * @zdev: The zdev for which to unblock load/store access
+ *
+ * Re-enables load/store access for a PCI function in the error state while
+ * keeping DMA blocked. In this state drivers can poke MMIO space to determine
+ * if error recovery is possible while catching any rogue DMA access from the
+ * device.
+ *
+ * Returns: 0 on success, -EIO otherwise
+ */
+int zpci_reset_load_store_blocked(struct zpci_dev *zdev)
+{
+	u64 req = ZPCI_CREATE_REQ(zdev->fh, 0, ZPCI_MOD_FC_RESET_BLOCK);
+	struct zpci_fib fib = {0};
+	u8 status;
+	int cc;
+
+	cc = zpci_mod_fc(req, &fib, &status);
+	if (cc) {
+		zpci_dbg(3, "rls fid:%x, cc:%d, status:%x\n", zdev->fid, cc, status);
+		return -EIO;
+	}
+
+	return 0;
+}
+
 static int zpci_mem_init(void)
 {
 	BUILD_BUG_ON(!is_power_of_2(__alignof__(struct zpci_fmb)) ||
diff --git a/arch/s390/pci/pci_event.c b/arch/s390/pci/pci_event.c
index 6a5bfa9..2e3e5b2 100644
--- a/arch/s390/pci/pci_event.c
+++ b/arch/s390/pci/pci_event.c
@@ -47,18 +47,223 @@ struct zpci_ccdf_avail {
 	u16 pec;			/* PCI event code */
 } __packed;
 
+static inline bool ers_result_indicates_abort(pci_ers_result_t ers_res)
+{
+	switch (ers_res) {
+	case PCI_ERS_RESULT_CAN_RECOVER:
+	case PCI_ERS_RESULT_RECOVERED:
+	case PCI_ERS_RESULT_NEED_RESET:
+		return false;
+	default:
+		return true;
+	}
+}
+
+static bool is_passed_through(struct zpci_dev *zdev)
+{
+	return zdev->s390_domain;
+}
+
+static bool is_driver_supported(struct pci_driver *driver)
+{
+	if (!driver || !driver->err_handler)
+		return false;
+	if (!driver->err_handler->error_detected)
+		return false;
+	if (!driver->err_handler->slot_reset)
+		return false;
+	if (!driver->err_handler->resume)
+		return false;
+	return true;
+}
+
+static pci_ers_result_t zpci_event_notify_error_detected(struct pci_dev *pdev,
+							 struct pci_driver *driver)
+{
+	pci_ers_result_t ers_res = PCI_ERS_RESULT_DISCONNECT;
+
+	ers_res = driver->err_handler->error_detected(pdev,  pdev->error_state);
+	if (ers_result_indicates_abort(ers_res))
+		pr_info("%s: Automatic recovery failed after initial reporting\n", pci_name(pdev));
+	else if (ers_res == PCI_ERS_RESULT_NEED_RESET)
+		pr_debug("%s: Driver needs reset to recover\n", pci_name(pdev));
+
+	return ers_res;
+}
+
+static pci_ers_result_t zpci_event_do_error_state_clear(struct pci_dev *pdev,
+							struct pci_driver *driver)
+{
+	pci_ers_result_t ers_res = PCI_ERS_RESULT_DISCONNECT;
+	struct zpci_dev *zdev = to_zpci(pdev);
+	int rc;
+
+	pr_info("%s: Unblocking device access for examination\n", pci_name(pdev));
+	rc = zpci_reset_load_store_blocked(zdev);
+	if (rc) {
+		pr_err("%s: Unblocking device access failed\n", pci_name(pdev));
+		/* Let's try a full reset instead */
+		return PCI_ERS_RESULT_NEED_RESET;
+	}
+
+	if (driver->err_handler->mmio_enabled) {
+		ers_res = driver->err_handler->mmio_enabled(pdev);
+		if (ers_result_indicates_abort(ers_res)) {
+			pr_info("%s: Automatic recovery failed after MMIO re-enable\n",
+				pci_name(pdev));
+			return ers_res;
+		} else if (ers_res == PCI_ERS_RESULT_NEED_RESET) {
+			pr_debug("%s: Driver needs reset to recover\n", pci_name(pdev));
+			return ers_res;
+		}
+	}
+
+	pr_debug("%s: Unblocking DMA\n", pci_name(pdev));
+	rc = zpci_clear_error_state(zdev);
+	if (!rc) {
+		pdev->error_state = pci_channel_io_normal;
+	} else {
+		pr_err("%s: Unblocking DMA failed\n", pci_name(pdev));
+		/* Let's try a full reset instead */
+		return PCI_ERS_RESULT_NEED_RESET;
+	}
+
+	return ers_res;
+}
+
+static pci_ers_result_t zpci_event_do_reset(struct pci_dev *pdev,
+					    struct pci_driver *driver)
+{
+	pci_ers_result_t ers_res = PCI_ERS_RESULT_DISCONNECT;
+
+	pr_info("%s: Initiating reset\n", pci_name(pdev));
+	if (zpci_hot_reset_device(to_zpci(pdev))) {
+		pr_err("%s: The reset request failed\n", pci_name(pdev));
+		return ers_res;
+	}
+	pdev->error_state = pci_channel_io_normal;
+	ers_res = driver->err_handler->slot_reset(pdev);
+	if (ers_result_indicates_abort(ers_res)) {
+		pr_info("%s: Automatic recovery failed after slot reset\n", pci_name(pdev));
+		return ers_res;
+	}
+
+	return ers_res;
+}
+
+/* zpci_event_attempt_error_recovery - Try to recover the given PCI function
+ * @pdev: PCI function to recover currently in the error state
+ *
+ * We follow the scheme outlined in Documentation/PCI/pci-error-recovery.rst.
+ * With the simplification that recovery always happens per function
+ * and the platform determines which functions are affected for
+ * multi-function devices.
+ */
+static pci_ers_result_t zpci_event_attempt_error_recovery(struct pci_dev *pdev)
+{
+	pci_ers_result_t ers_res = PCI_ERS_RESULT_DISCONNECT;
+	struct pci_driver *driver;
+
+	/*
+	 * Ensure that the PCI function is not removed concurrently, no driver
+	 * is unbound or probed and that userspace can't access its
+	 * configuration space while we perform recovery.
+	 */
+	pci_dev_lock(pdev);
+	if (pdev->error_state == pci_channel_io_perm_failure) {
+		ers_res = PCI_ERS_RESULT_DISCONNECT;
+		goto out_unlock;
+	}
+	pdev->error_state = pci_channel_io_frozen;
+
+	if (is_passed_through(to_zpci(pdev))) {
+		pr_info("%s: Cannot be recovered in the host because it is a pass-through device\n",
+			pci_name(pdev));
+		goto out_unlock;
+	}
+
+	driver = to_pci_driver(pdev->dev.driver);
+	if (!is_driver_supported(driver)) {
+		if (!driver)
+			pr_info("%s: Cannot be recovered because no driver is bound to the device\n",
+				pci_name(pdev));
+		else
+			pr_info("%s: The %s driver bound to the device does not support error recovery\n",
+				pci_name(pdev),
+				driver->name);
+		goto out_unlock;
+	}
+
+	ers_res = zpci_event_notify_error_detected(pdev, driver);
+	if (ers_result_indicates_abort(ers_res))
+		goto out_unlock;
+
+	if (ers_res == PCI_ERS_RESULT_CAN_RECOVER) {
+		ers_res = zpci_event_do_error_state_clear(pdev, driver);
+		if (ers_result_indicates_abort(ers_res))
+			goto out_unlock;
+	}
+
+	if (ers_res == PCI_ERS_RESULT_NEED_RESET)
+		ers_res = zpci_event_do_reset(pdev, driver);
+
+	if (ers_res != PCI_ERS_RESULT_RECOVERED) {
+		pr_err("%s: Automatic recovery failed; operator intervention is required\n",
+		       pci_name(pdev));
+		goto out_unlock;
+	}
+
+	pr_info("%s: The device is ready to resume operations\n", pci_name(pdev));
+	if (driver->err_handler->resume)
+		driver->err_handler->resume(pdev);
+out_unlock:
+	pci_dev_unlock(pdev);
+
+	return ers_res;
+}
+
+/* zpci_event_io_failure - Report PCI channel failure state to driver
+ * @pdev: PCI function for which to report
+ * @es: PCI channel failure state to report
+ */
+static void zpci_event_io_failure(struct pci_dev *pdev, pci_channel_state_t es)
+{
+	struct pci_driver *driver;
+
+	pci_dev_lock(pdev);
+	pdev->error_state = es;
+	/**
+	 * While vfio-pci's error_detected callback notifies user-space QEMU
+	 * reacts to this by freezing the guest. In an s390 environment PCI
+	 * errors are rarely fatal so this is overkill. Instead in the future
+	 * we will inject the error event and let the guest recover the device
+	 * itself.
+	 */
+	if (is_passed_through(to_zpci(pdev)))
+		goto out;
+	driver = to_pci_driver(pdev->dev.driver);
+	if (driver && driver->err_handler && driver->err_handler->error_detected)
+		driver->err_handler->error_detected(pdev, pdev->error_state);
+out:
+	pci_dev_unlock(pdev);
+}
+
 static void __zpci_event_error(struct zpci_ccdf_err *ccdf)
 {
 	struct zpci_dev *zdev = get_zdev_by_fid(ccdf->fid);
 	struct pci_dev *pdev = NULL;
+	pci_ers_result_t ers_res;
 
 	zpci_dbg(3, "err fid:%x, fh:%x, pec:%x\n",
 		 ccdf->fid, ccdf->fh, ccdf->pec);
 	zpci_err("error CCDF:\n");
 	zpci_err_hex(ccdf, sizeof(*ccdf));
 
-	if (zdev)
-		pdev = pci_get_slot(zdev->zbus->bus, zdev->devfn);
+	if (zdev) {
+		zpci_update_fh(zdev, ccdf->fh);
+		if (zdev->zbus->bus)
+			pdev = pci_get_slot(zdev->zbus->bus, zdev->devfn);
+	}
 
 	pr_err("%s: Event 0x%x reports an error for PCI function 0x%x\n",
 	       pdev ? pci_name(pdev) : "n/a", ccdf->pec, ccdf->fid);
@@ -66,7 +271,20 @@ static void __zpci_event_error(struct zpci_ccdf_err *ccdf)
 	if (!pdev)
 		return;
 
-	pdev->error_state = pci_channel_io_perm_failure;
+	switch (ccdf->pec) {
+	case 0x003a: /* Service Action or Error Recovery Successful */
+		ers_res = zpci_event_attempt_error_recovery(pdev);
+		if (ers_res != PCI_ERS_RESULT_RECOVERED)
+			zpci_event_io_failure(pdev, pci_channel_io_perm_failure);
+		break;
+	default:
+		/*
+		 * Mark as frozen not permanently failed because the device
+		 * could be subsequently recovered by the platform.
+		 */
+		zpci_event_io_failure(pdev, pci_channel_io_frozen);
+		break;
+	}
 	pci_dev_put(pdev);
 }
 
@@ -78,7 +296,7 @@ void zpci_event_error(void *data)
 
 static void zpci_event_hard_deconfigured(struct zpci_dev *zdev, u32 fh)
 {
-	zdev->fh = fh;
+	zpci_update_fh(zdev, fh);
 	/* Give the driver a hint that the function is
 	 * already unusable.
 	 */
@@ -121,7 +339,7 @@ static void __zpci_event_availability(struct zpci_ccdf_avail *ccdf)
 		if (!zdev)
 			zpci_create_device(ccdf->fid, ccdf->fh, ZPCI_FN_STATE_STANDBY);
 		else
-			zdev->fh = ccdf->fh;
+			zpci_update_fh(zdev, ccdf->fh);
 		break;
 	case 0x0303: /* Deconfiguration requested */
 		if (zdev) {
@@ -130,7 +348,7 @@ static void __zpci_event_availability(struct zpci_ccdf_avail *ccdf)
 			 */
 			if (zdev->state != ZPCI_FN_STATE_CONFIGURED)
 				break;
-			zdev->fh = ccdf->fh;
+			zpci_update_fh(zdev, ccdf->fh);
 			zpci_deconfigure_device(zdev);
 		}
 		break;
diff --git a/arch/s390/pci/pci_insn.c b/arch/s390/pci/pci_insn.c
index 2e43996..28d863a 100644
--- a/arch/s390/pci/pci_insn.c
+++ b/arch/s390/pci/pci_insn.c
@@ -163,7 +163,7 @@ static inline int zpci_load_fh(u64 *data, const volatile void __iomem *addr,
 			       unsigned long len)
 {
 	struct zpci_iomap_entry *entry = &zpci_iomap_start[ZPCI_IDX(addr)];
-	u64 req = ZPCI_CREATE_REQ(entry->fh, entry->bar, len);
+	u64 req = ZPCI_CREATE_REQ(READ_ONCE(entry->fh), entry->bar, len);
 
 	return __zpci_load(data, req, ZPCI_OFFSET(addr));
 }
@@ -244,7 +244,7 @@ static inline int zpci_store_fh(const volatile void __iomem *addr, u64 data,
 				unsigned long len)
 {
 	struct zpci_iomap_entry *entry = &zpci_iomap_start[ZPCI_IDX(addr)];
-	u64 req = ZPCI_CREATE_REQ(entry->fh, entry->bar, len);
+	u64 req = ZPCI_CREATE_REQ(READ_ONCE(entry->fh), entry->bar, len);
 
 	return __zpci_store(data, req, ZPCI_OFFSET(addr));
 }
diff --git a/arch/s390/pci/pci_irq.c b/arch/s390/pci/pci_irq.c
index 3823e15..954bb7a 100644
--- a/arch/s390/pci/pci_irq.c
+++ b/arch/s390/pci/pci_irq.c
@@ -387,6 +387,15 @@ void arch_teardown_msi_irqs(struct pci_dev *pdev)
 		airq_iv_free(zpci_ibv[0], zdev->msi_first_bit, zdev->msi_nr_irqs);
 }
 
+void arch_restore_msi_irqs(struct pci_dev *pdev)
+{
+	struct zpci_dev *zdev = to_zpci(pdev);
+
+	if (!zdev->irqs_registered)
+		zpci_set_irq(zdev);
+	default_restore_msi_irqs(pdev);
+}
+
 static struct airq_struct zpci_airq = {
 	.handler = zpci_floating_irq_handler,
 	.isc = PCI_ISC,
diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c
index 6039644..42cf01e 100644
--- a/arch/x86/events/intel/core.c
+++ b/arch/x86/events/intel/core.c
@@ -3048,8 +3048,10 @@ intel_vlbr_constraints(struct perf_event *event)
 {
 	struct event_constraint *c = &vlbr_constraint;
 
-	if (unlikely(constraint_match(c, event->hw.config)))
+	if (unlikely(constraint_match(c, event->hw.config))) {
+		event->hw.flags |= c->flags;
 		return c;
+	}
 
 	return NULL;
 }
diff --git a/arch/x86/events/intel/lbr.c b/arch/x86/events/intel/lbr.c
index 6b72e9b..8043213 100644
--- a/arch/x86/events/intel/lbr.c
+++ b/arch/x86/events/intel/lbr.c
@@ -265,6 +265,8 @@ void intel_pmu_lbr_reset(void)
 
 	cpuc->last_task_ctx = NULL;
 	cpuc->last_log_id = 0;
+	if (!static_cpu_has(X86_FEATURE_ARCH_LBR) && cpuc->lbr_select)
+		wrmsrl(MSR_LBR_SELECT, 0);
 }
 
 /*
diff --git a/arch/x86/include/asm/fpu/xcr.h b/arch/x86/include/asm/fpu/xcr.h
index 79f95d3..9656a5b 100644
--- a/arch/x86/include/asm/fpu/xcr.h
+++ b/arch/x86/include/asm/fpu/xcr.h
@@ -3,6 +3,7 @@
 #define _ASM_X86_FPU_XCR_H
 
 #define XCR_XFEATURE_ENABLED_MASK	0x00000000
+#define XCR_XFEATURE_IN_USE_MASK	0x00000001
 
 static inline u64 xgetbv(u32 index)
 {
@@ -20,4 +21,15 @@ static inline void xsetbv(u32 index, u64 value)
 	asm volatile("xsetbv" :: "a" (eax), "d" (edx), "c" (index));
 }
 
+/*
+ * Return a mask of xfeatures which are currently being tracked
+ * by the processor as being in the initial configuration.
+ *
+ * Callers should check X86_FEATURE_XGETBV1.
+ */
+static inline u64 xfeatures_in_use(void)
+{
+	return xgetbv(XCR_XFEATURE_IN_USE_MASK);
+}
+
 #endif /* _ASM_X86_FPU_XCR_H */
diff --git a/arch/x86/include/asm/fpu/xstate.h b/arch/x86/include/asm/fpu/xstate.h
index 0f8b90a..cd3dd17 100644
--- a/arch/x86/include/asm/fpu/xstate.h
+++ b/arch/x86/include/asm/fpu/xstate.h
@@ -92,6 +92,13 @@
 #define XFEATURE_MASK_FPSTATE	(XFEATURE_MASK_USER_RESTORE | \
 				 XFEATURE_MASK_SUPERVISOR_SUPPORTED)
 
+/*
+ * Features in this mask have space allocated in the signal frame, but may not
+ * have that space initialized when the feature is in its init state.
+ */
+#define XFEATURE_MASK_SIGFRAME_INITOPT	(XFEATURE_MASK_XTILE | \
+					 XFEATURE_MASK_USER_DYNAMIC)
+
 extern u64 xstate_fx_sw_bytes[USER_XSTATE_FX_SW_WORDS];
 
 extern void __init update_regset_xstate_info(unsigned int size,
diff --git a/arch/x86/include/asm/intel-family.h b/arch/x86/include/asm/intel-family.h
index 2715843..5a0bcf8 100644
--- a/arch/x86/include/asm/intel-family.h
+++ b/arch/x86/include/asm/intel-family.h
@@ -108,6 +108,8 @@
 #define INTEL_FAM6_ALDERLAKE		0x97	/* Golden Cove / Gracemont */
 #define INTEL_FAM6_ALDERLAKE_L		0x9A	/* Golden Cove / Gracemont */
 
+#define INTEL_FAM6_RAPTOR_LAKE		0xB7
+
 /* "Small Core" Processors (Atom) */
 
 #define INTEL_FAM6_ATOM_BONNELL		0x1C /* Diamondville, Pineview */
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 2acf37c..e5d8700 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -38,7 +38,6 @@
 #define __KVM_HAVE_ARCH_VCPU_DEBUGFS
 
 #define KVM_MAX_VCPUS 1024
-#define KVM_SOFT_MAX_VCPUS 710
 
 /*
  * In x86, the VCPU ID corresponds to the APIC ID, and APIC IDs
@@ -725,6 +724,7 @@ struct kvm_vcpu_arch {
 
 	int cpuid_nent;
 	struct kvm_cpuid_entry2 *cpuid_entries;
+	u32 kvm_cpuid_base;
 
 	u64 reserved_gpa_bits;
 	int maxphyaddr;
@@ -748,7 +748,7 @@ struct kvm_vcpu_arch {
 		u8 preempted;
 		u64 msr_val;
 		u64 last_steal;
-		struct gfn_to_pfn_cache cache;
+		struct gfn_to_hva_cache cache;
 	} st;
 
 	u64 l1_tsc_offset;
@@ -1034,6 +1034,7 @@ struct kvm_x86_msr_filter {
 #define APICV_INHIBIT_REASON_IRQWIN     3
 #define APICV_INHIBIT_REASON_PIT_REINJ  4
 #define APICV_INHIBIT_REASON_X2APIC	5
+#define APICV_INHIBIT_REASON_BLOCKIRQ	6
 
 struct kvm_arch {
 	unsigned long n_used_mmu_pages;
@@ -1476,6 +1477,7 @@ struct kvm_x86_ops {
 	int (*mem_enc_reg_region)(struct kvm *kvm, struct kvm_enc_region *argp);
 	int (*mem_enc_unreg_region)(struct kvm *kvm, struct kvm_enc_region *argp);
 	int (*vm_copy_enc_context_from)(struct kvm *kvm, unsigned int source_fd);
+	int (*vm_move_enc_context_from)(struct kvm *kvm, unsigned int source_fd);
 
 	int (*get_msr_feature)(struct kvm_msr_entry *entry);
 
diff --git a/arch/x86/include/asm/kvm_para.h b/arch/x86/include/asm/kvm_para.h
index 6929987..56935eb 100644
--- a/arch/x86/include/asm/kvm_para.h
+++ b/arch/x86/include/asm/kvm_para.h
@@ -83,6 +83,18 @@ static inline long kvm_hypercall4(unsigned int nr, unsigned long p1,
 	return ret;
 }
 
+static inline long kvm_sev_hypercall3(unsigned int nr, unsigned long p1,
+				      unsigned long p2, unsigned long p3)
+{
+	long ret;
+
+	asm volatile("vmmcall"
+		     : "=a"(ret)
+		     : "a"(nr), "b"(p1), "c"(p2), "d"(p3)
+		     : "memory");
+	return ret;
+}
+
 #ifdef CONFIG_KVM_GUEST
 void kvmclock_init(void);
 void kvmclock_disable(void);
diff --git a/arch/x86/include/asm/mem_encrypt.h b/arch/x86/include/asm/mem_encrypt.h
index 2d4f5c1..e2c6f43 100644
--- a/arch/x86/include/asm/mem_encrypt.h
+++ b/arch/x86/include/asm/mem_encrypt.h
@@ -44,6 +44,8 @@ void __init sme_enable(struct boot_params *bp);
 
 int __init early_set_memory_decrypted(unsigned long vaddr, unsigned long size);
 int __init early_set_memory_encrypted(unsigned long vaddr, unsigned long size);
+void __init early_set_mem_enc_dec_hypercall(unsigned long vaddr, int npages,
+					    bool enc);
 
 void __init mem_encrypt_free_decrypted_mem(void);
 
@@ -78,6 +80,8 @@ static inline int __init
 early_set_memory_decrypted(unsigned long vaddr, unsigned long size) { return 0; }
 static inline int __init
 early_set_memory_encrypted(unsigned long vaddr, unsigned long size) { return 0; }
+static inline void __init
+early_set_mem_enc_dec_hypercall(unsigned long vaddr, int npages, bool enc) {}
 
 static inline void mem_encrypt_free_decrypted_mem(void) { }
 
diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h
index cebec95..21c4a69 100644
--- a/arch/x86/include/asm/paravirt.h
+++ b/arch/x86/include/asm/paravirt.h
@@ -97,6 +97,12 @@ static inline void paravirt_arch_exit_mmap(struct mm_struct *mm)
 	PVOP_VCALL1(mmu.exit_mmap, mm);
 }
 
+static inline void notify_page_enc_status_changed(unsigned long pfn,
+						  int npages, bool enc)
+{
+	PVOP_VCALL3(mmu.notify_page_enc_status_changed, pfn, npages, enc);
+}
+
 #ifdef CONFIG_PARAVIRT_XXL
 static inline void load_sp0(unsigned long sp0)
 {
diff --git a/arch/x86/include/asm/paravirt_types.h b/arch/x86/include/asm/paravirt_types.h
index fc1151e..a69012e 100644
--- a/arch/x86/include/asm/paravirt_types.h
+++ b/arch/x86/include/asm/paravirt_types.h
@@ -168,6 +168,7 @@ struct pv_mmu_ops {
 
 	/* Hook for intercepting the destruction of an mm_struct. */
 	void (*exit_mmap)(struct mm_struct *mm);
+	void (*notify_page_enc_status_changed)(unsigned long pfn, int npages, bool enc);
 
 #ifdef CONFIG_PARAVIRT_XXL
 	struct paravirt_callee_save read_cr2;
diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index 191878a..355d38c 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -806,11 +806,14 @@ static inline u32 amd_get_nodes_per_socket(void)	{ return 0; }
 static inline u32 amd_get_highest_perf(void)		{ return 0; }
 #endif
 
+#define for_each_possible_hypervisor_cpuid_base(function) \
+	for (function = 0x40000000; function < 0x40010000; function += 0x100)
+
 static inline uint32_t hypervisor_cpuid_base(const char *sig, uint32_t leaves)
 {
 	uint32_t base, eax, signature[3];
 
-	for (base = 0x40000000; base < 0x40010000; base += 0x100) {
+	for_each_possible_hypervisor_cpuid_base(base) {
 		cpuid(base, &eax, &signature[0], &signature[1], &signature[2]);
 
 		if (!memcmp(sig, signature, 12) &&
diff --git a/arch/x86/include/asm/set_memory.h b/arch/x86/include/asm/set_memory.h
index 43fa081..8726175 100644
--- a/arch/x86/include/asm/set_memory.h
+++ b/arch/x86/include/asm/set_memory.h
@@ -83,6 +83,7 @@ int set_pages_rw(struct page *page, int numpages);
 int set_direct_map_invalid_noflush(struct page *page);
 int set_direct_map_default_noflush(struct page *page);
 bool kernel_page_present(struct page *page);
+void notify_range_enc_status_changed(unsigned long vaddr, int npages, bool enc);
 
 extern int kernel_set_to_readonly;
 
diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h
index 08b0e90..81a0211 100644
--- a/arch/x86/include/asm/smp.h
+++ b/arch/x86/include/asm/smp.h
@@ -126,6 +126,7 @@ static inline void arch_send_call_function_ipi_mask(const struct cpumask *mask)
 
 void cpu_disable_common(void);
 void native_smp_prepare_boot_cpu(void);
+void smp_prepare_cpus_common(void);
 void native_smp_prepare_cpus(unsigned int max_cpus);
 void calculate_max_logical_packages(void);
 void native_smp_cpus_done(unsigned int max_cpus);
diff --git a/arch/x86/include/asm/static_call.h b/arch/x86/include/asm/static_call.h
index cbb67b6..39ebe05 100644
--- a/arch/x86/include/asm/static_call.h
+++ b/arch/x86/include/asm/static_call.h
@@ -27,6 +27,7 @@
 	    ".globl " STATIC_CALL_TRAMP_STR(name) "		\n"	\
 	    STATIC_CALL_TRAMP_STR(name) ":			\n"	\
 	    insns "						\n"	\
+	    ".byte 0x53, 0x43, 0x54				\n"	\
 	    ".type " STATIC_CALL_TRAMP_STR(name) ", @function	\n"	\
 	    ".size " STATIC_CALL_TRAMP_STR(name) ", . - " STATIC_CALL_TRAMP_STR(name) " \n" \
 	    ".popsection					\n")
diff --git a/arch/x86/include/uapi/asm/kvm_para.h b/arch/x86/include/uapi/asm/kvm_para.h
index 5146bba..6e64b27b 100644
--- a/arch/x86/include/uapi/asm/kvm_para.h
+++ b/arch/x86/include/uapi/asm/kvm_para.h
@@ -8,6 +8,7 @@
  * should be used to determine that a VM is running under KVM.
  */
 #define KVM_CPUID_SIGNATURE	0x40000000
+#define KVM_SIGNATURE "KVMKVMKVM\0\0\0"
 
 /* This CPUID returns two feature bitmaps in eax, edx. Before enabling
  * a particular paravirtualization, the appropriate feature bit should
diff --git a/arch/x86/kernel/cpu/cpuid-deps.c b/arch/x86/kernel/cpu/cpuid-deps.c
index cb2fdd13..c881bca 100644
--- a/arch/x86/kernel/cpu/cpuid-deps.c
+++ b/arch/x86/kernel/cpu/cpuid-deps.c
@@ -76,6 +76,7 @@ static const struct cpuid_dep cpuid_deps[] = {
 	{ X86_FEATURE_SGX1,			X86_FEATURE_SGX       },
 	{ X86_FEATURE_SGX2,			X86_FEATURE_SGX1      },
 	{ X86_FEATURE_XFD,			X86_FEATURE_XSAVES    },
+	{ X86_FEATURE_XFD,			X86_FEATURE_XGETBV1   },
 	{ X86_FEATURE_AMX_TILE,			X86_FEATURE_XFD       },
 	{}
 };
diff --git a/arch/x86/kernel/cpu/mce/intel.c b/arch/x86/kernel/cpu/mce/intel.c
index acfd5d9..bb9a46a 100644
--- a/arch/x86/kernel/cpu/mce/intel.c
+++ b/arch/x86/kernel/cpu/mce/intel.c
@@ -547,12 +547,13 @@ bool intel_filter_mce(struct mce *m)
 {
 	struct cpuinfo_x86 *c = &boot_cpu_data;
 
-	/* MCE errata HSD131, HSM142, HSW131, BDM48, and HSM142 */
+	/* MCE errata HSD131, HSM142, HSW131, BDM48, HSM142 and SKX37 */
 	if ((c->x86 == 6) &&
 	    ((c->x86_model == INTEL_FAM6_HASWELL) ||
 	     (c->x86_model == INTEL_FAM6_HASWELL_L) ||
 	     (c->x86_model == INTEL_FAM6_BROADWELL) ||
-	     (c->x86_model == INTEL_FAM6_HASWELL_G)) &&
+	     (c->x86_model == INTEL_FAM6_HASWELL_G) ||
+	     (c->x86_model == INTEL_FAM6_SKYLAKE_X)) &&
 	    (m->bank == 0) &&
 	    ((m->status & 0xa0000000ffffffff) == 0x80000000000f0005))
 		return true;
diff --git a/arch/x86/kernel/fpu/xstate.h b/arch/x86/kernel/fpu/xstate.h
index e18210d..86ea7c0 100644
--- a/arch/x86/kernel/fpu/xstate.h
+++ b/arch/x86/kernel/fpu/xstate.h
@@ -4,6 +4,7 @@
 
 #include <asm/cpufeature.h>
 #include <asm/fpu/xstate.h>
+#include <asm/fpu/xcr.h>
 
 #ifdef CONFIG_X86_64
 DECLARE_PER_CPU(u64, xfd_state);
@@ -199,6 +200,32 @@ static inline void os_xrstor_supervisor(struct fpstate *fpstate)
 }
 
 /*
+ * XSAVE itself always writes all requested xfeatures.  Removing features
+ * from the request bitmap reduces the features which are written.
+ * Generate a mask of features which must be written to a sigframe.  The
+ * unset features can be optimized away and not written.
+ *
+ * This optimization is user-visible.  Only use for states where
+ * uninitialized sigframe contents are tolerable, like dynamic features.
+ *
+ * Users of buffers produced with this optimization must check XSTATE_BV
+ * to determine which features have been optimized out.
+ */
+static inline u64 xfeatures_need_sigframe_write(void)
+{
+	u64 xfeaures_to_write;
+
+	/* In-use features must be written: */
+	xfeaures_to_write = xfeatures_in_use();
+
+	/* Also write all non-optimizable sigframe features: */
+	xfeaures_to_write |= XFEATURE_MASK_USER_SUPPORTED &
+			     ~XFEATURE_MASK_SIGFRAME_INITOPT;
+
+	return xfeaures_to_write;
+}
+
+/*
  * Save xstate to user space xsave area.
  *
  * We don't use modified optimization because xrstor/xrstors might track
@@ -220,10 +247,16 @@ static inline int xsave_to_user_sigframe(struct xregs_state __user *buf)
 	 */
 	struct fpstate *fpstate = current->thread.fpu.fpstate;
 	u64 mask = fpstate->user_xfeatures;
-	u32 lmask = mask;
-	u32 hmask = mask >> 32;
+	u32 lmask;
+	u32 hmask;
 	int err;
 
+	/* Optimize away writing unnecessary xfeatures: */
+	if (fpu_state_size_dynamic())
+		mask &= xfeatures_need_sigframe_write();
+
+	lmask = mask;
+	hmask = mask >> 32;
 	xfd_validate_state(fpstate, mask, false);
 
 	stac();
diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
index 8863d19..59abbda 100644
--- a/arch/x86/kernel/kvm.c
+++ b/arch/x86/kernel/kvm.c
@@ -28,6 +28,7 @@
 #include <linux/swait.h>
 #include <linux/syscore_ops.h>
 #include <linux/cc_platform.h>
+#include <linux/efi.h>
 #include <asm/timer.h>
 #include <asm/cpu.h>
 #include <asm/traps.h>
@@ -41,6 +42,7 @@
 #include <asm/ptrace.h>
 #include <asm/reboot.h>
 #include <asm/svm.h>
+#include <asm/e820/api.h>
 
 DEFINE_STATIC_KEY_FALSE(kvm_async_pf_enabled);
 
@@ -434,6 +436,8 @@ static void kvm_guest_cpu_offline(bool shutdown)
 	kvm_disable_steal_time();
 	if (kvm_para_has_feature(KVM_FEATURE_PV_EOI))
 		wrmsrl(MSR_KVM_PV_EOI_EN, 0);
+	if (kvm_para_has_feature(KVM_FEATURE_MIGRATION_CONTROL))
+		wrmsrl(MSR_KVM_MIGRATION_CONTROL, 0);
 	kvm_pv_disable_apf();
 	if (!shutdown)
 		apf_task_wake_all();
@@ -548,6 +552,55 @@ static void kvm_send_ipi_mask_allbutself(const struct cpumask *mask, int vector)
 	__send_ipi_mask(local_mask, vector);
 }
 
+static int __init setup_efi_kvm_sev_migration(void)
+{
+	efi_char16_t efi_sev_live_migration_enabled[] = L"SevLiveMigrationEnabled";
+	efi_guid_t efi_variable_guid = AMD_SEV_MEM_ENCRYPT_GUID;
+	efi_status_t status;
+	unsigned long size;
+	bool enabled;
+
+	if (!cc_platform_has(CC_ATTR_GUEST_MEM_ENCRYPT) ||
+	    !kvm_para_has_feature(KVM_FEATURE_MIGRATION_CONTROL))
+		return 0;
+
+	if (!efi_enabled(EFI_BOOT))
+		return 0;
+
+	if (!efi_enabled(EFI_RUNTIME_SERVICES)) {
+		pr_info("%s : EFI runtime services are not enabled\n", __func__);
+		return 0;
+	}
+
+	size = sizeof(enabled);
+
+	/* Get variable contents into buffer */
+	status = efi.get_variable(efi_sev_live_migration_enabled,
+				  &efi_variable_guid, NULL, &size, &enabled);
+
+	if (status == EFI_NOT_FOUND) {
+		pr_info("%s : EFI live migration variable not found\n", __func__);
+		return 0;
+	}
+
+	if (status != EFI_SUCCESS) {
+		pr_info("%s : EFI variable retrieval failed\n", __func__);
+		return 0;
+	}
+
+	if (enabled == 0) {
+		pr_info("%s: live migration disabled in EFI\n", __func__);
+		return 0;
+	}
+
+	pr_info("%s : live migration enabled in EFI\n", __func__);
+	wrmsrl(MSR_KVM_MIGRATION_CONTROL, KVM_MIGRATION_READY);
+
+	return 1;
+}
+
+late_initcall(setup_efi_kvm_sev_migration);
+
 /*
  * Set the IPI entry points
  */
@@ -756,7 +809,7 @@ static noinline uint32_t __kvm_cpuid_base(void)
 		return 0;	/* So we don't blow up on old processors */
 
 	if (boot_cpu_has(X86_FEATURE_HYPERVISOR))
-		return hypervisor_cpuid_base("KVMKVMKVM\0\0\0", 0);
+		return hypervisor_cpuid_base(KVM_SIGNATURE, 0);
 
 	return 0;
 }
@@ -806,8 +859,62 @@ static bool __init kvm_msi_ext_dest_id(void)
 	return kvm_para_has_feature(KVM_FEATURE_MSI_EXT_DEST_ID);
 }
 
+static void kvm_sev_hc_page_enc_status(unsigned long pfn, int npages, bool enc)
+{
+	kvm_sev_hypercall3(KVM_HC_MAP_GPA_RANGE, pfn << PAGE_SHIFT, npages,
+			   KVM_MAP_GPA_RANGE_ENC_STAT(enc) | KVM_MAP_GPA_RANGE_PAGE_SZ_4K);
+}
+
 static void __init kvm_init_platform(void)
 {
+	if (cc_platform_has(CC_ATTR_GUEST_MEM_ENCRYPT) &&
+	    kvm_para_has_feature(KVM_FEATURE_MIGRATION_CONTROL)) {
+		unsigned long nr_pages;
+		int i;
+
+		pv_ops.mmu.notify_page_enc_status_changed =
+			kvm_sev_hc_page_enc_status;
+
+		/*
+		 * Reset the host's shared pages list related to kernel
+		 * specific page encryption status settings before we load a
+		 * new kernel by kexec. Reset the page encryption status
+		 * during early boot intead of just before kexec to avoid SMP
+		 * races during kvm_pv_guest_cpu_reboot().
+		 * NOTE: We cannot reset the complete shared pages list
+		 * here as we need to retain the UEFI/OVMF firmware
+		 * specific settings.
+		 */
+
+		for (i = 0; i < e820_table->nr_entries; i++) {
+			struct e820_entry *entry = &e820_table->entries[i];
+
+			if (entry->type != E820_TYPE_RAM)
+				continue;
+
+			nr_pages = DIV_ROUND_UP(entry->size, PAGE_SIZE);
+
+			kvm_sev_hypercall3(KVM_HC_MAP_GPA_RANGE, entry->addr,
+				       nr_pages,
+				       KVM_MAP_GPA_RANGE_ENCRYPTED | KVM_MAP_GPA_RANGE_PAGE_SZ_4K);
+		}
+
+		/*
+		 * Ensure that _bss_decrypted section is marked as decrypted in the
+		 * shared pages list.
+		 */
+		nr_pages = DIV_ROUND_UP(__end_bss_decrypted - __start_bss_decrypted,
+					PAGE_SIZE);
+		early_set_mem_enc_dec_hypercall((unsigned long)__start_bss_decrypted,
+						nr_pages, 0);
+
+		/*
+		 * If not booted using EFI, enable Live migration support.
+		 */
+		if (!efi_enabled(EFI_BOOT))
+			wrmsrl(MSR_KVM_MIGRATION_CONTROL,
+			       KVM_MIGRATION_READY);
+	}
 	kvmclock_init();
 	x86_platform.apic_post_init = kvm_apic_init;
 }
diff --git a/arch/x86/kernel/paravirt.c b/arch/x86/kernel/paravirt.c
index 7157c2d..7f7636a 100644
--- a/arch/x86/kernel/paravirt.c
+++ b/arch/x86/kernel/paravirt.c
@@ -337,6 +337,7 @@ struct paravirt_patch_template pv_ops = {
 			(void (*)(struct mmu_gather *, void *))tlb_remove_page,
 
 	.mmu.exit_mmap		= paravirt_nop,
+	.mmu.notify_page_enc_status_changed	= paravirt_nop,
 
 #ifdef CONFIG_PARAVIRT_XXL
 	.mmu.read_cr2		= __PV_IS_CALLEE_SAVE(pv_native_read_cr2),
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 8241927..ac2909f 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -1350,12 +1350,7 @@ static void __init smp_get_logical_apicid(void)
 		cpu0_logical_apicid = GET_APIC_LOGICAL_ID(apic_read(APIC_LDR));
 }
 
-/*
- * Prepare for SMP bootup.
- * @max_cpus: configured maximum number of CPUs, It is a legacy parameter
- *            for common interface support.
- */
-void __init native_smp_prepare_cpus(unsigned int max_cpus)
+void __init smp_prepare_cpus_common(void)
 {
 	unsigned int i;
 
@@ -1386,6 +1381,17 @@ void __init native_smp_prepare_cpus(unsigned int max_cpus)
 	set_sched_topology(x86_topology);
 
 	set_cpu_sibling_map(0);
+}
+
+/*
+ * Prepare for SMP bootup.
+ * @max_cpus: configured maximum number of CPUs, It is a legacy parameter
+ *            for common interface support.
+ */
+void __init native_smp_prepare_cpus(unsigned int max_cpus)
+{
+	smp_prepare_cpus_common();
+
 	init_freq_invariance(false, false);
 	smp_sanity_check();
 
diff --git a/arch/x86/kernel/static_call.c b/arch/x86/kernel/static_call.c
index ea028e7..9c407a3 100644
--- a/arch/x86/kernel/static_call.c
+++ b/arch/x86/kernel/static_call.c
@@ -56,10 +56,15 @@ static void __ref __static_call_transform(void *insn, enum insn_type type, void
 	text_poke_bp(insn, code, size, emulate);
 }
 
-static void __static_call_validate(void *insn, bool tail)
+static void __static_call_validate(void *insn, bool tail, bool tramp)
 {
 	u8 opcode = *(u8 *)insn;
 
+	if (tramp && memcmp(insn+5, "SCT", 3)) {
+		pr_err("trampoline signature fail");
+		BUG();
+	}
+
 	if (tail) {
 		if (opcode == JMP32_INSN_OPCODE ||
 		    opcode == RET_INSN_OPCODE)
@@ -74,7 +79,8 @@ static void __static_call_validate(void *insn, bool tail)
 	/*
 	 * If we ever trigger this, our text is corrupt, we'll probably not live long.
 	 */
-	WARN_ONCE(1, "unexpected static_call insn opcode 0x%x at %pS\n", opcode, insn);
+	pr_err("unexpected static_call insn opcode 0x%x at %pS\n", opcode, insn);
+	BUG();
 }
 
 static inline enum insn_type __sc_insn(bool null, bool tail)
@@ -97,12 +103,12 @@ void arch_static_call_transform(void *site, void *tramp, void *func, bool tail)
 	mutex_lock(&text_mutex);
 
 	if (tramp) {
-		__static_call_validate(tramp, true);
+		__static_call_validate(tramp, true, true);
 		__static_call_transform(tramp, __sc_insn(!func, true), func);
 	}
 
 	if (IS_ENABLED(CONFIG_HAVE_STATIC_CALL_INLINE) && site) {
-		__static_call_validate(site, tail);
+		__static_call_validate(site, tail, false);
 		__static_call_transform(site, __sc_insn(!func, tail), func);
 	}
 
diff --git a/arch/x86/kernel/vm86_32.c b/arch/x86/kernel/vm86_32.c
index f14f69d..cce1c89 100644
--- a/arch/x86/kernel/vm86_32.c
+++ b/arch/x86/kernel/vm86_32.c
@@ -106,7 +106,7 @@ void save_v86_state(struct kernel_vm86_regs *regs, int retval)
 	 */
 	local_irq_enable();
 
-	BUG_ON(!vm86 || !vm86->user_vm86);
+	BUG_ON(!vm86);
 
 	set_flags(regs->pt.flags, VEFLAGS, X86_EFLAGS_VIF | vm86->veflags_mask);
 	user = vm86->user_vm86;
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index 2d70edb..e19dabf 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -99,11 +99,45 @@ static int kvm_check_cpuid(struct kvm_cpuid_entry2 *entries, int nent)
 	return 0;
 }
 
+static void kvm_update_kvm_cpuid_base(struct kvm_vcpu *vcpu)
+{
+	u32 function;
+	struct kvm_cpuid_entry2 *entry;
+
+	vcpu->arch.kvm_cpuid_base = 0;
+
+	for_each_possible_hypervisor_cpuid_base(function) {
+		entry = kvm_find_cpuid_entry(vcpu, function, 0);
+
+		if (entry) {
+			u32 signature[3];
+
+			signature[0] = entry->ebx;
+			signature[1] = entry->ecx;
+			signature[2] = entry->edx;
+
+			BUILD_BUG_ON(sizeof(signature) > sizeof(KVM_SIGNATURE));
+			if (!memcmp(signature, KVM_SIGNATURE, sizeof(signature))) {
+				vcpu->arch.kvm_cpuid_base = function;
+				break;
+			}
+		}
+	}
+}
+
+struct kvm_cpuid_entry2 *kvm_find_kvm_cpuid_features(struct kvm_vcpu *vcpu)
+{
+	u32 base = vcpu->arch.kvm_cpuid_base;
+
+	if (!base)
+		return NULL;
+
+	return kvm_find_cpuid_entry(vcpu, base | KVM_CPUID_FEATURES, 0);
+}
+
 void kvm_update_pv_runtime(struct kvm_vcpu *vcpu)
 {
-	struct kvm_cpuid_entry2 *best;
-
-	best = kvm_find_cpuid_entry(vcpu, KVM_CPUID_FEATURES, 0);
+	struct kvm_cpuid_entry2 *best = kvm_find_kvm_cpuid_features(vcpu);
 
 	/*
 	 * save the feature bitmap to avoid cpuid lookup for every PV
@@ -142,7 +176,7 @@ void kvm_update_cpuid_runtime(struct kvm_vcpu *vcpu)
 		     cpuid_entry_has(best, X86_FEATURE_XSAVEC)))
 		best->ebx = xstate_required_size(vcpu->arch.xcr0, true);
 
-	best = kvm_find_cpuid_entry(vcpu, KVM_CPUID_FEATURES, 0);
+	best = kvm_find_kvm_cpuid_features(vcpu);
 	if (kvm_hlt_in_guest(vcpu->kvm) && best &&
 		(best->eax & (1 << KVM_FEATURE_PV_UNHALT)))
 		best->eax &= ~(1 << KVM_FEATURE_PV_UNHALT);
@@ -239,6 +273,26 @@ u64 kvm_vcpu_reserved_gpa_bits_raw(struct kvm_vcpu *vcpu)
 	return rsvd_bits(cpuid_maxphyaddr(vcpu), 63);
 }
 
+static int kvm_set_cpuid(struct kvm_vcpu *vcpu, struct kvm_cpuid_entry2 *e2,
+                        int nent)
+{
+    int r;
+
+    r = kvm_check_cpuid(e2, nent);
+    if (r)
+        return r;
+
+    kvfree(vcpu->arch.cpuid_entries);
+    vcpu->arch.cpuid_entries = e2;
+    vcpu->arch.cpuid_nent = nent;
+
+    kvm_update_kvm_cpuid_base(vcpu);
+    kvm_update_cpuid_runtime(vcpu);
+    kvm_vcpu_after_set_cpuid(vcpu);
+
+    return 0;
+}
+
 /* when an old userspace process fills a new kernel module */
 int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
 			     struct kvm_cpuid *cpuid,
@@ -275,18 +329,9 @@ int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
 		e2[i].padding[2] = 0;
 	}
 
-	r = kvm_check_cpuid(e2, cpuid->nent);
-	if (r) {
+	r = kvm_set_cpuid(vcpu, e2, cpuid->nent);
+	if (r)
 		kvfree(e2);
-		goto out_free_cpuid;
-	}
-
-	kvfree(vcpu->arch.cpuid_entries);
-	vcpu->arch.cpuid_entries = e2;
-	vcpu->arch.cpuid_nent = cpuid->nent;
-
-	kvm_update_cpuid_runtime(vcpu);
-	kvm_vcpu_after_set_cpuid(vcpu);
 
 out_free_cpuid:
 	kvfree(e);
@@ -310,20 +355,11 @@ int kvm_vcpu_ioctl_set_cpuid2(struct kvm_vcpu *vcpu,
 			return PTR_ERR(e2);
 	}
 
-	r = kvm_check_cpuid(e2, cpuid->nent);
-	if (r) {
+	r = kvm_set_cpuid(vcpu, e2, cpuid->nent);
+	if (r)
 		kvfree(e2);
-		return r;
-	}
 
-	kvfree(vcpu->arch.cpuid_entries);
-	vcpu->arch.cpuid_entries = e2;
-	vcpu->arch.cpuid_nent = cpuid->nent;
-
-	kvm_update_cpuid_runtime(vcpu);
-	kvm_vcpu_after_set_cpuid(vcpu);
-
-	return 0;
+	return r;
 }
 
 int kvm_vcpu_ioctl_get_cpuid2(struct kvm_vcpu *vcpu,
@@ -871,8 +907,7 @@ static inline int __do_cpuid_func(struct kvm_cpuid_array *array, u32 function)
 		}
 		break;
 	case KVM_CPUID_SIGNATURE: {
-		static const char signature[12] = "KVMKVMKVM\0\0";
-		const u32 *sigptr = (const u32 *)signature;
+		const u32 *sigptr = (const u32 *)KVM_SIGNATURE;
 		entry->eax = KVM_CPUID_FEATURES;
 		entry->ebx = sigptr[0];
 		entry->ecx = sigptr[1];
diff --git a/arch/x86/kvm/hyperv.c b/arch/x86/kvm/hyperv.c
index 4f15c01..4a555f3 100644
--- a/arch/x86/kvm/hyperv.c
+++ b/arch/x86/kvm/hyperv.c
@@ -1472,7 +1472,7 @@ static int kvm_hv_set_msr(struct kvm_vcpu *vcpu, u32 msr, u64 data, bool host)
 
 		if (!(data & HV_X64_MSR_VP_ASSIST_PAGE_ENABLE)) {
 			hv_vcpu->hv_vapic = data;
-			if (kvm_lapic_enable_pv_eoi(vcpu, 0, 0))
+			if (kvm_lapic_set_pv_eoi(vcpu, 0, 0))
 				return 1;
 			break;
 		}
@@ -1490,7 +1490,7 @@ static int kvm_hv_set_msr(struct kvm_vcpu *vcpu, u32 msr, u64 data, bool host)
 			return 1;
 		hv_vcpu->hv_vapic = data;
 		kvm_vcpu_mark_page_dirty(vcpu, gfn);
-		if (kvm_lapic_enable_pv_eoi(vcpu,
+		if (kvm_lapic_set_pv_eoi(vcpu,
 					    gfn_to_gpa(gfn) | KVM_MSR_ENABLED,
 					    sizeof(struct hv_vp_assist_page)))
 			return 1;
diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index d6ac32f..759952d 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -2856,25 +2856,30 @@ int kvm_hv_vapic_msr_read(struct kvm_vcpu *vcpu, u32 reg, u64 *data)
 	return 0;
 }
 
-int kvm_lapic_enable_pv_eoi(struct kvm_vcpu *vcpu, u64 data, unsigned long len)
+int kvm_lapic_set_pv_eoi(struct kvm_vcpu *vcpu, u64 data, unsigned long len)
 {
 	u64 addr = data & ~KVM_MSR_ENABLED;
 	struct gfn_to_hva_cache *ghc = &vcpu->arch.pv_eoi.data;
 	unsigned long new_len;
+	int ret;
 
 	if (!IS_ALIGNED(addr, 4))
 		return 1;
 
+	if (data & KVM_MSR_ENABLED) {
+		if (addr == ghc->gpa && len <= ghc->len)
+			new_len = ghc->len;
+		else
+			new_len = len;
+
+		ret = kvm_gfn_to_hva_cache_init(vcpu->kvm, ghc, addr, new_len);
+		if (ret)
+			return ret;
+	}
+
 	vcpu->arch.pv_eoi.msr_val = data;
-	if (!pv_eoi_enabled(vcpu))
-		return 0;
 
-	if (addr == ghc->gpa && len <= ghc->len)
-		new_len = ghc->len;
-	else
-		new_len = len;
-
-	return kvm_gfn_to_hva_cache_init(vcpu->kvm, ghc, addr, new_len);
+	return 0;
 }
 
 int kvm_apic_accept_events(struct kvm_vcpu *vcpu)
diff --git a/arch/x86/kvm/lapic.h b/arch/x86/kvm/lapic.h
index d7c25d0..2b44e53 100644
--- a/arch/x86/kvm/lapic.h
+++ b/arch/x86/kvm/lapic.h
@@ -127,7 +127,7 @@ int kvm_x2apic_msr_read(struct kvm_vcpu *vcpu, u32 msr, u64 *data);
 int kvm_hv_vapic_msr_write(struct kvm_vcpu *vcpu, u32 msr, u64 data);
 int kvm_hv_vapic_msr_read(struct kvm_vcpu *vcpu, u32 msr, u64 *data);
 
-int kvm_lapic_enable_pv_eoi(struct kvm_vcpu *vcpu, u64 data, unsigned long len);
+int kvm_lapic_set_pv_eoi(struct kvm_vcpu *vcpu, u64 data, unsigned long len);
 void kvm_lapic_exit(void);
 
 #define VEC_POS(v) ((v) & (32 - 1))
diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 323b505..3379437 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -3191,17 +3191,17 @@ static int fast_page_fault(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault)
 			new_spte |= PT_WRITABLE_MASK;
 
 			/*
-			 * Do not fix write-permission on the large spte.  Since
-			 * we only dirty the first page into the dirty-bitmap in
+			 * Do not fix write-permission on the large spte when
+			 * dirty logging is enabled. Since we only dirty the
+			 * first page into the dirty-bitmap in
 			 * fast_pf_fix_direct_spte(), other pages are missed
 			 * if its slot has dirty logging enabled.
 			 *
 			 * Instead, we let the slow page fault path create a
 			 * normal spte to fix the access.
-			 *
-			 * See the comments in kvm_arch_commit_memory_region().
 			 */
-			if (sp->role.level > PG_LEVEL_4K)
+			if (sp->role.level > PG_LEVEL_4K &&
+			    kvm_slot_dirty_track_enabled(fault->slot))
 				break;
 		}
 
diff --git a/arch/x86/kvm/mmu/tdp_mmu.c b/arch/x86/kvm/mmu/tdp_mmu.c
index 7c5dd83..a54c349 100644
--- a/arch/x86/kvm/mmu/tdp_mmu.c
+++ b/arch/x86/kvm/mmu/tdp_mmu.c
@@ -897,7 +897,7 @@ static int tdp_mmu_map_handle_target_level(struct kvm_vcpu *vcpu,
 					  struct kvm_page_fault *fault,
 					  struct tdp_iter *iter)
 {
-	struct kvm_mmu_page *sp = sptep_to_sp(iter->sptep);
+	struct kvm_mmu_page *sp = sptep_to_sp(rcu_dereference(iter->sptep));
 	u64 new_spte;
 	int ret = RET_PF_FIXED;
 	bool wrprot = false;
diff --git a/arch/x86/kvm/pmu.c b/arch/x86/kvm/pmu.c
index 0772bad..09873f6 100644
--- a/arch/x86/kvm/pmu.c
+++ b/arch/x86/kvm/pmu.c
@@ -319,7 +319,7 @@ void kvm_pmu_handle_event(struct kvm_vcpu *vcpu)
 }
 
 /* check if idx is a valid index to access PMU */
-int kvm_pmu_is_valid_rdpmc_ecx(struct kvm_vcpu *vcpu, unsigned int idx)
+bool kvm_pmu_is_valid_rdpmc_ecx(struct kvm_vcpu *vcpu, unsigned int idx)
 {
 	return kvm_x86_ops.pmu_ops->is_valid_rdpmc_ecx(vcpu, idx);
 }
diff --git a/arch/x86/kvm/pmu.h b/arch/x86/kvm/pmu.h
index 0e4f2b1..59d6b76 100644
--- a/arch/x86/kvm/pmu.h
+++ b/arch/x86/kvm/pmu.h
@@ -32,7 +32,7 @@ struct kvm_pmu_ops {
 	struct kvm_pmc *(*rdpmc_ecx_to_pmc)(struct kvm_vcpu *vcpu,
 		unsigned int idx, u64 *mask);
 	struct kvm_pmc *(*msr_idx_to_pmc)(struct kvm_vcpu *vcpu, u32 msr);
-	int (*is_valid_rdpmc_ecx)(struct kvm_vcpu *vcpu, unsigned int idx);
+	bool (*is_valid_rdpmc_ecx)(struct kvm_vcpu *vcpu, unsigned int idx);
 	bool (*is_valid_msr)(struct kvm_vcpu *vcpu, u32 msr);
 	int (*get_msr)(struct kvm_vcpu *vcpu, struct msr_data *msr_info);
 	int (*set_msr)(struct kvm_vcpu *vcpu, struct msr_data *msr_info);
@@ -149,7 +149,7 @@ void reprogram_counter(struct kvm_pmu *pmu, int pmc_idx);
 void kvm_pmu_deliver_pmi(struct kvm_vcpu *vcpu);
 void kvm_pmu_handle_event(struct kvm_vcpu *vcpu);
 int kvm_pmu_rdpmc(struct kvm_vcpu *vcpu, unsigned pmc, u64 *data);
-int kvm_pmu_is_valid_rdpmc_ecx(struct kvm_vcpu *vcpu, unsigned int idx);
+bool kvm_pmu_is_valid_rdpmc_ecx(struct kvm_vcpu *vcpu, unsigned int idx);
 bool kvm_pmu_is_valid_msr(struct kvm_vcpu *vcpu, u32 msr);
 int kvm_pmu_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info);
 int kvm_pmu_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info);
diff --git a/arch/x86/kvm/svm/avic.c b/arch/x86/kvm/svm/avic.c
index 8052d92..affc0ea 100644
--- a/arch/x86/kvm/svm/avic.c
+++ b/arch/x86/kvm/svm/avic.c
@@ -904,7 +904,8 @@ bool svm_check_apicv_inhibit_reasons(ulong bit)
 			  BIT(APICV_INHIBIT_REASON_NESTED) |
 			  BIT(APICV_INHIBIT_REASON_IRQWIN) |
 			  BIT(APICV_INHIBIT_REASON_PIT_REINJ) |
-			  BIT(APICV_INHIBIT_REASON_X2APIC);
+			  BIT(APICV_INHIBIT_REASON_X2APIC) |
+			  BIT(APICV_INHIBIT_REASON_BLOCKIRQ);
 
 	return supported & BIT(bit);
 }
diff --git a/arch/x86/kvm/svm/pmu.c b/arch/x86/kvm/svm/pmu.c
index fdf587f..871c426 100644
--- a/arch/x86/kvm/svm/pmu.c
+++ b/arch/x86/kvm/svm/pmu.c
@@ -181,14 +181,13 @@ static struct kvm_pmc *amd_pmc_idx_to_pmc(struct kvm_pmu *pmu, int pmc_idx)
 	return get_gp_pmc_amd(pmu, base + pmc_idx, PMU_TYPE_COUNTER);
 }
 
-/* returns 0 if idx's corresponding MSR exists; otherwise returns 1. */
-static int amd_is_valid_rdpmc_ecx(struct kvm_vcpu *vcpu, unsigned int idx)
+static bool amd_is_valid_rdpmc_ecx(struct kvm_vcpu *vcpu, unsigned int idx)
 {
 	struct kvm_pmu *pmu = vcpu_to_pmu(vcpu);
 
 	idx &= ~(3u << 30);
 
-	return (idx >= pmu->nr_arch_gp_counters);
+	return idx < pmu->nr_arch_gp_counters;
 }
 
 /* idx is the ECX register of RDPMC instruction */
diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
index 1964b9a..902c52a 100644
--- a/arch/x86/kvm/svm/sev.c
+++ b/arch/x86/kvm/svm/sev.c
@@ -120,16 +120,26 @@ static bool __sev_recycle_asids(int min_asid, int max_asid)
 	return true;
 }
 
+static int sev_misc_cg_try_charge(struct kvm_sev_info *sev)
+{
+	enum misc_res_type type = sev->es_active ? MISC_CG_RES_SEV_ES : MISC_CG_RES_SEV;
+	return misc_cg_try_charge(type, sev->misc_cg, 1);
+}
+
+static void sev_misc_cg_uncharge(struct kvm_sev_info *sev)
+{
+	enum misc_res_type type = sev->es_active ? MISC_CG_RES_SEV_ES : MISC_CG_RES_SEV;
+	misc_cg_uncharge(type, sev->misc_cg, 1);
+}
+
 static int sev_asid_new(struct kvm_sev_info *sev)
 {
 	int asid, min_asid, max_asid, ret;
 	bool retry = true;
-	enum misc_res_type type;
 
-	type = sev->es_active ? MISC_CG_RES_SEV_ES : MISC_CG_RES_SEV;
 	WARN_ON(sev->misc_cg);
 	sev->misc_cg = get_current_misc_cg();
-	ret = misc_cg_try_charge(type, sev->misc_cg, 1);
+	ret = sev_misc_cg_try_charge(sev);
 	if (ret) {
 		put_misc_cg(sev->misc_cg);
 		sev->misc_cg = NULL;
@@ -162,7 +172,7 @@ static int sev_asid_new(struct kvm_sev_info *sev)
 
 	return asid;
 e_uncharge:
-	misc_cg_uncharge(type, sev->misc_cg, 1);
+	sev_misc_cg_uncharge(sev);
 	put_misc_cg(sev->misc_cg);
 	sev->misc_cg = NULL;
 	return ret;
@@ -179,7 +189,6 @@ static void sev_asid_free(struct kvm_sev_info *sev)
 {
 	struct svm_cpu_data *sd;
 	int cpu;
-	enum misc_res_type type;
 
 	mutex_lock(&sev_bitmap_lock);
 
@@ -192,8 +201,7 @@ static void sev_asid_free(struct kvm_sev_info *sev)
 
 	mutex_unlock(&sev_bitmap_lock);
 
-	type = sev->es_active ? MISC_CG_RES_SEV_ES : MISC_CG_RES_SEV;
-	misc_cg_uncharge(type, sev->misc_cg, 1);
+	sev_misc_cg_uncharge(sev);
 	put_misc_cg(sev->misc_cg);
 	sev->misc_cg = NULL;
 }
@@ -590,7 +598,7 @@ static int sev_es_sync_vmsa(struct vcpu_svm *svm)
 	 * traditional VMSA as it has been built so far (in prep
 	 * for LAUNCH_UPDATE_VMSA) to be the initial SEV-ES state.
 	 */
-	memcpy(svm->vmsa, save, sizeof(*save));
+	memcpy(svm->sev_es.vmsa, save, sizeof(*save));
 
 	return 0;
 }
@@ -612,11 +620,11 @@ static int __sev_launch_update_vmsa(struct kvm *kvm, struct kvm_vcpu *vcpu,
 	 * the VMSA memory content (i.e it will write the same memory region
 	 * with the guest's key), so invalidate it first.
 	 */
-	clflush_cache_range(svm->vmsa, PAGE_SIZE);
+	clflush_cache_range(svm->sev_es.vmsa, PAGE_SIZE);
 
 	vmsa.reserved = 0;
 	vmsa.handle = to_kvm_svm(kvm)->sev_info.handle;
-	vmsa.address = __sme_pa(svm->vmsa);
+	vmsa.address = __sme_pa(svm->sev_es.vmsa);
 	vmsa.len = PAGE_SIZE;
 	ret = sev_issue_cmd(kvm, SEV_CMD_LAUNCH_UPDATE_VMSA, &vmsa, error);
 	if (ret)
@@ -1536,6 +1544,201 @@ static bool cmd_allowed_from_miror(u32 cmd_id)
 	return false;
 }
 
+static int sev_lock_for_migration(struct kvm *kvm)
+{
+	struct kvm_sev_info *sev = &to_kvm_svm(kvm)->sev_info;
+
+	/*
+	 * Bail if this VM is already involved in a migration to avoid deadlock
+	 * between two VMs trying to migrate to/from each other.
+	 */
+	if (atomic_cmpxchg_acquire(&sev->migration_in_progress, 0, 1))
+		return -EBUSY;
+
+	mutex_lock(&kvm->lock);
+
+	return 0;
+}
+
+static void sev_unlock_after_migration(struct kvm *kvm)
+{
+	struct kvm_sev_info *sev = &to_kvm_svm(kvm)->sev_info;
+
+	mutex_unlock(&kvm->lock);
+	atomic_set_release(&sev->migration_in_progress, 0);
+}
+
+
+static int sev_lock_vcpus_for_migration(struct kvm *kvm)
+{
+	struct kvm_vcpu *vcpu;
+	int i, j;
+
+	kvm_for_each_vcpu(i, vcpu, kvm) {
+		if (mutex_lock_killable(&vcpu->mutex))
+			goto out_unlock;
+	}
+
+	return 0;
+
+out_unlock:
+	kvm_for_each_vcpu(j, vcpu, kvm) {
+		if (i == j)
+			break;
+
+		mutex_unlock(&vcpu->mutex);
+	}
+	return -EINTR;
+}
+
+static void sev_unlock_vcpus_for_migration(struct kvm *kvm)
+{
+	struct kvm_vcpu *vcpu;
+	int i;
+
+	kvm_for_each_vcpu(i, vcpu, kvm) {
+		mutex_unlock(&vcpu->mutex);
+	}
+}
+
+static void sev_migrate_from(struct kvm_sev_info *dst,
+			      struct kvm_sev_info *src)
+{
+	dst->active = true;
+	dst->asid = src->asid;
+	dst->handle = src->handle;
+	dst->pages_locked = src->pages_locked;
+
+	src->asid = 0;
+	src->active = false;
+	src->handle = 0;
+	src->pages_locked = 0;
+
+	INIT_LIST_HEAD(&dst->regions_list);
+	list_replace_init(&src->regions_list, &dst->regions_list);
+}
+
+static int sev_es_migrate_from(struct kvm *dst, struct kvm *src)
+{
+	int i;
+	struct kvm_vcpu *dst_vcpu, *src_vcpu;
+	struct vcpu_svm *dst_svm, *src_svm;
+
+	if (atomic_read(&src->online_vcpus) != atomic_read(&dst->online_vcpus))
+		return -EINVAL;
+
+	kvm_for_each_vcpu(i, src_vcpu, src) {
+		if (!src_vcpu->arch.guest_state_protected)
+			return -EINVAL;
+	}
+
+	kvm_for_each_vcpu(i, src_vcpu, src) {
+		src_svm = to_svm(src_vcpu);
+		dst_vcpu = kvm_get_vcpu(dst, i);
+		dst_svm = to_svm(dst_vcpu);
+
+		/*
+		 * Transfer VMSA and GHCB state to the destination.  Nullify and
+		 * clear source fields as appropriate, the state now belongs to
+		 * the destination.
+		 */
+		memcpy(&dst_svm->sev_es, &src_svm->sev_es, sizeof(src_svm->sev_es));
+		dst_svm->vmcb->control.ghcb_gpa = src_svm->vmcb->control.ghcb_gpa;
+		dst_svm->vmcb->control.vmsa_pa = src_svm->vmcb->control.vmsa_pa;
+		dst_vcpu->arch.guest_state_protected = true;
+
+		memset(&src_svm->sev_es, 0, sizeof(src_svm->sev_es));
+		src_svm->vmcb->control.ghcb_gpa = INVALID_PAGE;
+		src_svm->vmcb->control.vmsa_pa = INVALID_PAGE;
+		src_vcpu->arch.guest_state_protected = false;
+	}
+	to_kvm_svm(src)->sev_info.es_active = false;
+	to_kvm_svm(dst)->sev_info.es_active = true;
+
+	return 0;
+}
+
+int svm_vm_migrate_from(struct kvm *kvm, unsigned int source_fd)
+{
+	struct kvm_sev_info *dst_sev = &to_kvm_svm(kvm)->sev_info;
+	struct kvm_sev_info *src_sev, *cg_cleanup_sev;
+	struct file *source_kvm_file;
+	struct kvm *source_kvm;
+	bool charged = false;
+	int ret;
+
+	ret = sev_lock_for_migration(kvm);
+	if (ret)
+		return ret;
+
+	if (sev_guest(kvm)) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+
+	source_kvm_file = fget(source_fd);
+	if (!file_is_kvm(source_kvm_file)) {
+		ret = -EBADF;
+		goto out_fput;
+	}
+
+	source_kvm = source_kvm_file->private_data;
+	ret = sev_lock_for_migration(source_kvm);
+	if (ret)
+		goto out_fput;
+
+	if (!sev_guest(source_kvm)) {
+		ret = -EINVAL;
+		goto out_source;
+	}
+
+	src_sev = &to_kvm_svm(source_kvm)->sev_info;
+	dst_sev->misc_cg = get_current_misc_cg();
+	cg_cleanup_sev = dst_sev;
+	if (dst_sev->misc_cg != src_sev->misc_cg) {
+		ret = sev_misc_cg_try_charge(dst_sev);
+		if (ret)
+			goto out_dst_cgroup;
+		charged = true;
+	}
+
+	ret = sev_lock_vcpus_for_migration(kvm);
+	if (ret)
+		goto out_dst_cgroup;
+	ret = sev_lock_vcpus_for_migration(source_kvm);
+	if (ret)
+		goto out_dst_vcpu;
+
+	if (sev_es_guest(source_kvm)) {
+		ret = sev_es_migrate_from(kvm, source_kvm);
+		if (ret)
+			goto out_source_vcpu;
+	}
+	sev_migrate_from(dst_sev, src_sev);
+	kvm_vm_dead(source_kvm);
+	cg_cleanup_sev = src_sev;
+	ret = 0;
+
+out_source_vcpu:
+	sev_unlock_vcpus_for_migration(source_kvm);
+out_dst_vcpu:
+	sev_unlock_vcpus_for_migration(kvm);
+out_dst_cgroup:
+	/* Operates on the source on success, on the destination on failure.  */
+	if (charged)
+		sev_misc_cg_uncharge(cg_cleanup_sev);
+	put_misc_cg(cg_cleanup_sev->misc_cg);
+	cg_cleanup_sev->misc_cg = NULL;
+out_source:
+	sev_unlock_after_migration(source_kvm);
+out_fput:
+	if (source_kvm_file)
+		fput(source_kvm_file);
+out_unlock:
+	sev_unlock_after_migration(kvm);
+	return ret;
+}
+
 int svm_mem_enc_op(struct kvm *kvm, void __user *argp)
 {
 	struct kvm_sev_cmd sev_cmd;
@@ -2038,16 +2241,16 @@ void sev_free_vcpu(struct kvm_vcpu *vcpu)
 	svm = to_svm(vcpu);
 
 	if (vcpu->arch.guest_state_protected)
-		sev_flush_guest_memory(svm, svm->vmsa, PAGE_SIZE);
-	__free_page(virt_to_page(svm->vmsa));
+		sev_flush_guest_memory(svm, svm->sev_es.vmsa, PAGE_SIZE);
+	__free_page(virt_to_page(svm->sev_es.vmsa));
 
-	if (svm->ghcb_sa_free)
-		kfree(svm->ghcb_sa);
+	if (svm->sev_es.ghcb_sa_free)
+		kfree(svm->sev_es.ghcb_sa);
 }
 
 static void dump_ghcb(struct vcpu_svm *svm)
 {
-	struct ghcb *ghcb = svm->ghcb;
+	struct ghcb *ghcb = svm->sev_es.ghcb;
 	unsigned int nbits;
 
 	/* Re-use the dump_invalid_vmcb module parameter */
@@ -2073,7 +2276,7 @@ static void dump_ghcb(struct vcpu_svm *svm)
 static void sev_es_sync_to_ghcb(struct vcpu_svm *svm)
 {
 	struct kvm_vcpu *vcpu = &svm->vcpu;
-	struct ghcb *ghcb = svm->ghcb;
+	struct ghcb *ghcb = svm->sev_es.ghcb;
 
 	/*
 	 * The GHCB protocol so far allows for the following data
@@ -2093,7 +2296,7 @@ static void sev_es_sync_from_ghcb(struct vcpu_svm *svm)
 {
 	struct vmcb_control_area *control = &svm->vmcb->control;
 	struct kvm_vcpu *vcpu = &svm->vcpu;
-	struct ghcb *ghcb = svm->ghcb;
+	struct ghcb *ghcb = svm->sev_es.ghcb;
 	u64 exit_code;
 
 	/*
@@ -2140,7 +2343,7 @@ static int sev_es_validate_vmgexit(struct vcpu_svm *svm)
 	struct ghcb *ghcb;
 	u64 exit_code = 0;
 
-	ghcb = svm->ghcb;
+	ghcb = svm->sev_es.ghcb;
 
 	/* Only GHCB Usage code 0 is supported */
 	if (ghcb->ghcb_usage)
@@ -2258,33 +2461,34 @@ static int sev_es_validate_vmgexit(struct vcpu_svm *svm)
 
 void sev_es_unmap_ghcb(struct vcpu_svm *svm)
 {
-	if (!svm->ghcb)
+	if (!svm->sev_es.ghcb)
 		return;
 
-	if (svm->ghcb_sa_free) {
+	if (svm->sev_es.ghcb_sa_free) {
 		/*
 		 * The scratch area lives outside the GHCB, so there is a
 		 * buffer that, depending on the operation performed, may
 		 * need to be synced, then freed.
 		 */
-		if (svm->ghcb_sa_sync) {
+		if (svm->sev_es.ghcb_sa_sync) {
 			kvm_write_guest(svm->vcpu.kvm,
-					ghcb_get_sw_scratch(svm->ghcb),
-					svm->ghcb_sa, svm->ghcb_sa_len);
-			svm->ghcb_sa_sync = false;
+					ghcb_get_sw_scratch(svm->sev_es.ghcb),
+					svm->sev_es.ghcb_sa,
+					svm->sev_es.ghcb_sa_len);
+			svm->sev_es.ghcb_sa_sync = false;
 		}
 
-		kfree(svm->ghcb_sa);
-		svm->ghcb_sa = NULL;
-		svm->ghcb_sa_free = false;
+		kfree(svm->sev_es.ghcb_sa);
+		svm->sev_es.ghcb_sa = NULL;
+		svm->sev_es.ghcb_sa_free = false;
 	}
 
-	trace_kvm_vmgexit_exit(svm->vcpu.vcpu_id, svm->ghcb);
+	trace_kvm_vmgexit_exit(svm->vcpu.vcpu_id, svm->sev_es.ghcb);
 
 	sev_es_sync_to_ghcb(svm);
 
-	kvm_vcpu_unmap(&svm->vcpu, &svm->ghcb_map, true);
-	svm->ghcb = NULL;
+	kvm_vcpu_unmap(&svm->vcpu, &svm->sev_es.ghcb_map, true);
+	svm->sev_es.ghcb = NULL;
 }
 
 void pre_sev_run(struct vcpu_svm *svm, int cpu)
@@ -2314,7 +2518,7 @@ void pre_sev_run(struct vcpu_svm *svm, int cpu)
 static bool setup_vmgexit_scratch(struct vcpu_svm *svm, bool sync, u64 len)
 {
 	struct vmcb_control_area *control = &svm->vmcb->control;
-	struct ghcb *ghcb = svm->ghcb;
+	struct ghcb *ghcb = svm->sev_es.ghcb;
 	u64 ghcb_scratch_beg, ghcb_scratch_end;
 	u64 scratch_gpa_beg, scratch_gpa_end;
 	void *scratch_va;
@@ -2350,7 +2554,7 @@ static bool setup_vmgexit_scratch(struct vcpu_svm *svm, bool sync, u64 len)
 			return false;
 		}
 
-		scratch_va = (void *)svm->ghcb;
+		scratch_va = (void *)svm->sev_es.ghcb;
 		scratch_va += (scratch_gpa_beg - control->ghcb_gpa);
 	} else {
 		/*
@@ -2380,12 +2584,12 @@ static bool setup_vmgexit_scratch(struct vcpu_svm *svm, bool sync, u64 len)
 		 * the vCPU next time (i.e. a read was requested so the data
 		 * must be written back to the guest memory).
 		 */
-		svm->ghcb_sa_sync = sync;
-		svm->ghcb_sa_free = true;
+		svm->sev_es.ghcb_sa_sync = sync;
+		svm->sev_es.ghcb_sa_free = true;
 	}
 
-	svm->ghcb_sa = scratch_va;
-	svm->ghcb_sa_len = len;
+	svm->sev_es.ghcb_sa = scratch_va;
+	svm->sev_es.ghcb_sa_len = len;
 
 	return true;
 }
@@ -2504,15 +2708,15 @@ int sev_handle_vmgexit(struct kvm_vcpu *vcpu)
 		return -EINVAL;
 	}
 
-	if (kvm_vcpu_map(vcpu, ghcb_gpa >> PAGE_SHIFT, &svm->ghcb_map)) {
+	if (kvm_vcpu_map(vcpu, ghcb_gpa >> PAGE_SHIFT, &svm->sev_es.ghcb_map)) {
 		/* Unable to map GHCB from guest */
 		vcpu_unimpl(vcpu, "vmgexit: error mapping GHCB [%#llx] from guest\n",
 			    ghcb_gpa);
 		return -EINVAL;
 	}
 
-	svm->ghcb = svm->ghcb_map.hva;
-	ghcb = svm->ghcb_map.hva;
+	svm->sev_es.ghcb = svm->sev_es.ghcb_map.hva;
+	ghcb = svm->sev_es.ghcb_map.hva;
 
 	trace_kvm_vmgexit_enter(vcpu->vcpu_id, ghcb);
 
@@ -2535,7 +2739,7 @@ int sev_handle_vmgexit(struct kvm_vcpu *vcpu)
 		ret = kvm_sev_es_mmio_read(vcpu,
 					   control->exit_info_1,
 					   control->exit_info_2,
-					   svm->ghcb_sa);
+					   svm->sev_es.ghcb_sa);
 		break;
 	case SVM_VMGEXIT_MMIO_WRITE:
 		if (!setup_vmgexit_scratch(svm, false, control->exit_info_2))
@@ -2544,7 +2748,7 @@ int sev_handle_vmgexit(struct kvm_vcpu *vcpu)
 		ret = kvm_sev_es_mmio_write(vcpu,
 					    control->exit_info_1,
 					    control->exit_info_2,
-					    svm->ghcb_sa);
+					    svm->sev_es.ghcb_sa);
 		break;
 	case SVM_VMGEXIT_NMI_COMPLETE:
 		ret = svm_invoke_exit_handler(vcpu, SVM_EXIT_IRET);
@@ -2604,7 +2808,8 @@ int sev_es_string_io(struct vcpu_svm *svm, int size, unsigned int port, int in)
 	if (!setup_vmgexit_scratch(svm, in, bytes))
 		return -EINVAL;
 
-	return kvm_sev_es_string_io(&svm->vcpu, size, port, svm->ghcb_sa, count, in);
+	return kvm_sev_es_string_io(&svm->vcpu, size, port, svm->sev_es.ghcb_sa,
+				    count, in);
 }
 
 void sev_es_init_vmcb(struct vcpu_svm *svm)
@@ -2619,7 +2824,7 @@ void sev_es_init_vmcb(struct vcpu_svm *svm)
 	 * VMCB page. Do not include the encryption mask on the VMSA physical
 	 * address since hardware will access it using the guest key.
 	 */
-	svm->vmcb->control.vmsa_pa = __pa(svm->vmsa);
+	svm->vmcb->control.vmsa_pa = __pa(svm->sev_es.vmsa);
 
 	/* Can't intercept CR register access, HV can't modify CR registers */
 	svm_clr_intercept(svm, INTERCEPT_CR0_READ);
@@ -2691,8 +2896,8 @@ void sev_vcpu_deliver_sipi_vector(struct kvm_vcpu *vcpu, u8 vector)
 	struct vcpu_svm *svm = to_svm(vcpu);
 
 	/* First SIPI: Use the values as initially set by the VMM */
-	if (!svm->received_first_sipi) {
-		svm->received_first_sipi = true;
+	if (!svm->sev_es.received_first_sipi) {
+		svm->sev_es.received_first_sipi = true;
 		return;
 	}
 
@@ -2701,8 +2906,8 @@ void sev_vcpu_deliver_sipi_vector(struct kvm_vcpu *vcpu, u8 vector)
 	 * the guest will set the CS and RIP. Set SW_EXIT_INFO_2 to a
 	 * non-zero value.
 	 */
-	if (!svm->ghcb)
+	if (!svm->sev_es.ghcb)
 		return;
 
-	ghcb_set_sw_exit_info_2(svm->ghcb, 1);
+	ghcb_set_sw_exit_info_2(svm->sev_es.ghcb, 1);
 }
diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index b36ca4e..5630c24 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -1452,7 +1452,7 @@ static int svm_create_vcpu(struct kvm_vcpu *vcpu)
 	svm_switch_vmcb(svm, &svm->vmcb01);
 
 	if (vmsa_page)
-		svm->vmsa = page_address(vmsa_page);
+		svm->sev_es.vmsa = page_address(vmsa_page);
 
 	svm->guest_state_loaded = false;
 
@@ -2835,11 +2835,11 @@ static int svm_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 static int svm_complete_emulated_msr(struct kvm_vcpu *vcpu, int err)
 {
 	struct vcpu_svm *svm = to_svm(vcpu);
-	if (!err || !sev_es_guest(vcpu->kvm) || WARN_ON_ONCE(!svm->ghcb))
+	if (!err || !sev_es_guest(vcpu->kvm) || WARN_ON_ONCE(!svm->sev_es.ghcb))
 		return kvm_complete_insn_gp(vcpu, err);
 
-	ghcb_set_sw_exit_info_1(svm->ghcb, 1);
-	ghcb_set_sw_exit_info_2(svm->ghcb,
+	ghcb_set_sw_exit_info_1(svm->sev_es.ghcb, 1);
+	ghcb_set_sw_exit_info_2(svm->sev_es.ghcb,
 				X86_TRAP_GP |
 				SVM_EVTINJ_TYPE_EXEPT |
 				SVM_EVTINJ_VALID);
@@ -3121,11 +3121,6 @@ static int invpcid_interception(struct kvm_vcpu *vcpu)
 	type = svm->vmcb->control.exit_info_2;
 	gva = svm->vmcb->control.exit_info_1;
 
-	if (type > 3) {
-		kvm_inject_gp(vcpu, 0);
-		return 1;
-	}
-
 	return kvm_handle_invpcid(vcpu, type, gva);
 }
 
@@ -4701,6 +4696,7 @@ static struct kvm_x86_ops svm_x86_ops __initdata = {
 	.mem_enc_unreg_region = svm_unregister_enc_region,
 
 	.vm_copy_enc_context_from = svm_vm_copy_asid_from,
+	.vm_move_enc_context_from = svm_vm_migrate_from,
 
 	.can_emulate_instruction = svm_can_emulate_instruction,
 
diff --git a/arch/x86/kvm/svm/svm.h b/arch/x86/kvm/svm/svm.h
index 5e9510d..437e685 100644
--- a/arch/x86/kvm/svm/svm.h
+++ b/arch/x86/kvm/svm/svm.h
@@ -80,6 +80,7 @@ struct kvm_sev_info {
 	u64 ap_jump_table;	/* SEV-ES AP Jump Table address */
 	struct kvm *enc_context_owner; /* Owner of copied encryption context */
 	struct misc_cg *misc_cg; /* For misc cgroup accounting */
+	atomic_t migration_in_progress;
 };
 
 struct kvm_svm {
@@ -123,6 +124,20 @@ struct svm_nested_state {
 	bool initialized;
 };
 
+struct vcpu_sev_es_state {
+	/* SEV-ES support */
+	struct vmcb_save_area *vmsa;
+	struct ghcb *ghcb;
+	struct kvm_host_map ghcb_map;
+	bool received_first_sipi;
+
+	/* SEV-ES scratch area support */
+	void *ghcb_sa;
+	u32 ghcb_sa_len;
+	bool ghcb_sa_sync;
+	bool ghcb_sa_free;
+};
+
 struct vcpu_svm {
 	struct kvm_vcpu vcpu;
 	/* vmcb always points at current_vmcb->ptr, it's purely a shorthand. */
@@ -186,17 +201,7 @@ struct vcpu_svm {
 		DECLARE_BITMAP(write, MAX_DIRECT_ACCESS_MSRS);
 	} shadow_msr_intercept;
 
-	/* SEV-ES support */
-	struct vmcb_save_area *vmsa;
-	struct ghcb *ghcb;
-	struct kvm_host_map ghcb_map;
-	bool received_first_sipi;
-
-	/* SEV-ES scratch area support */
-	void *ghcb_sa;
-	u32 ghcb_sa_len;
-	bool ghcb_sa_sync;
-	bool ghcb_sa_free;
+	struct vcpu_sev_es_state sev_es;
 
 	bool guest_state_loaded;
 };
@@ -558,6 +563,7 @@ int svm_register_enc_region(struct kvm *kvm,
 int svm_unregister_enc_region(struct kvm *kvm,
 			      struct kvm_enc_region *range);
 int svm_vm_copy_asid_from(struct kvm *kvm, unsigned int source_fd);
+int svm_vm_migrate_from(struct kvm *kvm, unsigned int source_fd);
 void pre_sev_run(struct vcpu_svm *svm, int cpu);
 void __init sev_set_cpu_caps(void);
 void __init sev_hardware_setup(void);
diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c
index b4ee5e9..b213ca9 100644
--- a/arch/x86/kvm/vmx/nested.c
+++ b/arch/x86/kvm/vmx/nested.c
@@ -525,67 +525,19 @@ static int nested_vmx_check_tpr_shadow_controls(struct kvm_vcpu *vcpu,
 }
 
 /*
- * Check if MSR is intercepted for L01 MSR bitmap.
+ * For x2APIC MSRs, ignore the vmcs01 bitmap.  L1 can enable x2APIC without L1
+ * itself utilizing x2APIC.  All MSRs were previously set to be intercepted,
+ * only the "disable intercept" case needs to be handled.
  */
-static bool msr_write_intercepted_l01(struct kvm_vcpu *vcpu, u32 msr)
+static void nested_vmx_disable_intercept_for_x2apic_msr(unsigned long *msr_bitmap_l1,
+							unsigned long *msr_bitmap_l0,
+							u32 msr, int type)
 {
-	unsigned long *msr_bitmap;
-	int f = sizeof(unsigned long);
+	if (type & MSR_TYPE_R && !vmx_test_msr_bitmap_read(msr_bitmap_l1, msr))
+		vmx_clear_msr_bitmap_read(msr_bitmap_l0, msr);
 
-	if (!cpu_has_vmx_msr_bitmap())
-		return true;
-
-	msr_bitmap = to_vmx(vcpu)->vmcs01.msr_bitmap;
-
-	if (msr <= 0x1fff) {
-		return !!test_bit(msr, msr_bitmap + 0x800 / f);
-	} else if ((msr >= 0xc0000000) && (msr <= 0xc0001fff)) {
-		msr &= 0x1fff;
-		return !!test_bit(msr, msr_bitmap + 0xc00 / f);
-	}
-
-	return true;
-}
-
-/*
- * If a msr is allowed by L0, we should check whether it is allowed by L1.
- * The corresponding bit will be cleared unless both of L0 and L1 allow it.
- */
-static void nested_vmx_disable_intercept_for_msr(unsigned long *msr_bitmap_l1,
-					       unsigned long *msr_bitmap_nested,
-					       u32 msr, int type)
-{
-	int f = sizeof(unsigned long);
-
-	/*
-	 * See Intel PRM Vol. 3, 20.6.9 (MSR-Bitmap Address). Early manuals
-	 * have the write-low and read-high bitmap offsets the wrong way round.
-	 * We can control MSRs 0x00000000-0x00001fff and 0xc0000000-0xc0001fff.
-	 */
-	if (msr <= 0x1fff) {
-		if (type & MSR_TYPE_R &&
-		   !test_bit(msr, msr_bitmap_l1 + 0x000 / f))
-			/* read-low */
-			__clear_bit(msr, msr_bitmap_nested + 0x000 / f);
-
-		if (type & MSR_TYPE_W &&
-		   !test_bit(msr, msr_bitmap_l1 + 0x800 / f))
-			/* write-low */
-			__clear_bit(msr, msr_bitmap_nested + 0x800 / f);
-
-	} else if ((msr >= 0xc0000000) && (msr <= 0xc0001fff)) {
-		msr &= 0x1fff;
-		if (type & MSR_TYPE_R &&
-		   !test_bit(msr, msr_bitmap_l1 + 0x400 / f))
-			/* read-high */
-			__clear_bit(msr, msr_bitmap_nested + 0x400 / f);
-
-		if (type & MSR_TYPE_W &&
-		   !test_bit(msr, msr_bitmap_l1 + 0xc00 / f))
-			/* write-high */
-			__clear_bit(msr, msr_bitmap_nested + 0xc00 / f);
-
-	}
+	if (type & MSR_TYPE_W && !vmx_test_msr_bitmap_write(msr_bitmap_l1, msr))
+		vmx_clear_msr_bitmap_write(msr_bitmap_l0, msr);
 }
 
 static inline void enable_x2apic_msr_intercepts(unsigned long *msr_bitmap)
@@ -600,6 +552,34 @@ static inline void enable_x2apic_msr_intercepts(unsigned long *msr_bitmap)
 	}
 }
 
+#define BUILD_NVMX_MSR_INTERCEPT_HELPER(rw)					\
+static inline									\
+void nested_vmx_set_msr_##rw##_intercept(struct vcpu_vmx *vmx,			\
+					 unsigned long *msr_bitmap_l1,		\
+					 unsigned long *msr_bitmap_l0, u32 msr)	\
+{										\
+	if (vmx_test_msr_bitmap_##rw(vmx->vmcs01.msr_bitmap, msr) ||		\
+	    vmx_test_msr_bitmap_##rw(msr_bitmap_l1, msr))			\
+		vmx_set_msr_bitmap_##rw(msr_bitmap_l0, msr);			\
+	else									\
+		vmx_clear_msr_bitmap_##rw(msr_bitmap_l0, msr);			\
+}
+BUILD_NVMX_MSR_INTERCEPT_HELPER(read)
+BUILD_NVMX_MSR_INTERCEPT_HELPER(write)
+
+static inline void nested_vmx_set_intercept_for_msr(struct vcpu_vmx *vmx,
+						    unsigned long *msr_bitmap_l1,
+						    unsigned long *msr_bitmap_l0,
+						    u32 msr, int types)
+{
+	if (types & MSR_TYPE_R)
+		nested_vmx_set_msr_read_intercept(vmx, msr_bitmap_l1,
+						  msr_bitmap_l0, msr);
+	if (types & MSR_TYPE_W)
+		nested_vmx_set_msr_write_intercept(vmx, msr_bitmap_l1,
+						   msr_bitmap_l0, msr);
+}
+
 /*
  * Merge L0's and L1's MSR bitmap, return false to indicate that
  * we do not use the hardware.
@@ -607,10 +587,11 @@ static inline void enable_x2apic_msr_intercepts(unsigned long *msr_bitmap)
 static inline bool nested_vmx_prepare_msr_bitmap(struct kvm_vcpu *vcpu,
 						 struct vmcs12 *vmcs12)
 {
+	struct vcpu_vmx *vmx = to_vmx(vcpu);
 	int msr;
 	unsigned long *msr_bitmap_l1;
-	unsigned long *msr_bitmap_l0 = to_vmx(vcpu)->nested.vmcs02.msr_bitmap;
-	struct kvm_host_map *map = &to_vmx(vcpu)->nested.msr_bitmap_map;
+	unsigned long *msr_bitmap_l0 = vmx->nested.vmcs02.msr_bitmap;
+	struct kvm_host_map *map = &vmx->nested.msr_bitmap_map;
 
 	/* Nothing to do if the MSR bitmap is not in use.  */
 	if (!cpu_has_vmx_msr_bitmap() ||
@@ -625,7 +606,7 @@ static inline bool nested_vmx_prepare_msr_bitmap(struct kvm_vcpu *vcpu,
 	/*
 	 * To keep the control flow simple, pay eight 8-byte writes (sixteen
 	 * 4-byte writes on 32-bit systems) up front to enable intercepts for
-	 * the x2APIC MSR range and selectively disable them below.
+	 * the x2APIC MSR range and selectively toggle those relevant to L2.
 	 */
 	enable_x2apic_msr_intercepts(msr_bitmap_l0);
 
@@ -644,61 +625,44 @@ static inline bool nested_vmx_prepare_msr_bitmap(struct kvm_vcpu *vcpu,
 			}
 		}
 
-		nested_vmx_disable_intercept_for_msr(
+		nested_vmx_disable_intercept_for_x2apic_msr(
 			msr_bitmap_l1, msr_bitmap_l0,
 			X2APIC_MSR(APIC_TASKPRI),
 			MSR_TYPE_R | MSR_TYPE_W);
 
 		if (nested_cpu_has_vid(vmcs12)) {
-			nested_vmx_disable_intercept_for_msr(
+			nested_vmx_disable_intercept_for_x2apic_msr(
 				msr_bitmap_l1, msr_bitmap_l0,
 				X2APIC_MSR(APIC_EOI),
 				MSR_TYPE_W);
-			nested_vmx_disable_intercept_for_msr(
+			nested_vmx_disable_intercept_for_x2apic_msr(
 				msr_bitmap_l1, msr_bitmap_l0,
 				X2APIC_MSR(APIC_SELF_IPI),
 				MSR_TYPE_W);
 		}
 	}
 
-	/* KVM unconditionally exposes the FS/GS base MSRs to L1. */
-#ifdef CONFIG_X86_64
-	nested_vmx_disable_intercept_for_msr(msr_bitmap_l1, msr_bitmap_l0,
-					     MSR_FS_BASE, MSR_TYPE_RW);
-
-	nested_vmx_disable_intercept_for_msr(msr_bitmap_l1, msr_bitmap_l0,
-					     MSR_GS_BASE, MSR_TYPE_RW);
-
-	nested_vmx_disable_intercept_for_msr(msr_bitmap_l1, msr_bitmap_l0,
-					     MSR_KERNEL_GS_BASE, MSR_TYPE_RW);
-#endif
-
 	/*
-	 * Checking the L0->L1 bitmap is trying to verify two things:
-	 *
-	 * 1. L0 gave a permission to L1 to actually passthrough the MSR. This
-	 *    ensures that we do not accidentally generate an L02 MSR bitmap
-	 *    from the L12 MSR bitmap that is too permissive.
-	 * 2. That L1 or L2s have actually used the MSR. This avoids
-	 *    unnecessarily merging of the bitmap if the MSR is unused. This
-	 *    works properly because we only update the L01 MSR bitmap lazily.
-	 *    So even if L0 should pass L1 these MSRs, the L01 bitmap is only
-	 *    updated to reflect this when L1 (or its L2s) actually write to
-	 *    the MSR.
+	 * Always check vmcs01's bitmap to honor userspace MSR filters and any
+	 * other runtime changes to vmcs01's bitmap, e.g. dynamic pass-through.
 	 */
-	if (!msr_write_intercepted_l01(vcpu, MSR_IA32_SPEC_CTRL))
-		nested_vmx_disable_intercept_for_msr(
-					msr_bitmap_l1, msr_bitmap_l0,
-					MSR_IA32_SPEC_CTRL,
-					MSR_TYPE_R | MSR_TYPE_W);
+#ifdef CONFIG_X86_64
+	nested_vmx_set_intercept_for_msr(vmx, msr_bitmap_l1, msr_bitmap_l0,
+					 MSR_FS_BASE, MSR_TYPE_RW);
 
-	if (!msr_write_intercepted_l01(vcpu, MSR_IA32_PRED_CMD))
-		nested_vmx_disable_intercept_for_msr(
-					msr_bitmap_l1, msr_bitmap_l0,
-					MSR_IA32_PRED_CMD,
-					MSR_TYPE_W);
+	nested_vmx_set_intercept_for_msr(vmx, msr_bitmap_l1, msr_bitmap_l0,
+					 MSR_GS_BASE, MSR_TYPE_RW);
 
-	kvm_vcpu_unmap(vcpu, &to_vmx(vcpu)->nested.msr_bitmap_map, false);
+	nested_vmx_set_intercept_for_msr(vmx, msr_bitmap_l1, msr_bitmap_l0,
+					 MSR_KERNEL_GS_BASE, MSR_TYPE_RW);
+#endif
+	nested_vmx_set_intercept_for_msr(vmx, msr_bitmap_l1, msr_bitmap_l0,
+					 MSR_IA32_SPEC_CTRL, MSR_TYPE_RW);
+
+	nested_vmx_set_intercept_for_msr(vmx, msr_bitmap_l1, msr_bitmap_l0,
+					 MSR_IA32_PRED_CMD, MSR_TYPE_W);
+
+	kvm_vcpu_unmap(vcpu, &vmx->nested.msr_bitmap_map, false);
 
 	return true;
 }
@@ -5379,7 +5343,7 @@ static int handle_invept(struct kvm_vcpu *vcpu)
 	struct {
 		u64 eptp, gpa;
 	} operand;
-	int i, r;
+	int i, r, gpr_index;
 
 	if (!(vmx->nested.msrs.secondary_ctls_high &
 	      SECONDARY_EXEC_ENABLE_EPT) ||
@@ -5392,7 +5356,8 @@ static int handle_invept(struct kvm_vcpu *vcpu)
 		return 1;
 
 	vmx_instruction_info = vmcs_read32(VMX_INSTRUCTION_INFO);
-	type = kvm_register_read(vcpu, (vmx_instruction_info >> 28) & 0xf);
+	gpr_index = vmx_get_instr_info_reg2(vmx_instruction_info);
+	type = kvm_register_read(vcpu, gpr_index);
 
 	types = (vmx->nested.msrs.ept_caps >> VMX_EPT_EXTENT_SHIFT) & 6;
 
@@ -5459,7 +5424,7 @@ static int handle_invvpid(struct kvm_vcpu *vcpu)
 		u64 gla;
 	} operand;
 	u16 vpid02;
-	int r;
+	int r, gpr_index;
 
 	if (!(vmx->nested.msrs.secondary_ctls_high &
 	      SECONDARY_EXEC_ENABLE_VPID) ||
@@ -5472,7 +5437,8 @@ static int handle_invvpid(struct kvm_vcpu *vcpu)
 		return 1;
 
 	vmx_instruction_info = vmcs_read32(VMX_INSTRUCTION_INFO);
-	type = kvm_register_read(vcpu, (vmx_instruction_info >> 28) & 0xf);
+	gpr_index = vmx_get_instr_info_reg2(vmx_instruction_info);
+	type = kvm_register_read(vcpu, gpr_index);
 
 	types = (vmx->nested.msrs.vpid_caps &
 			VMX_VPID_EXTENT_SUPPORTED_MASK) >> 8;
diff --git a/arch/x86/kvm/vmx/pmu_intel.c b/arch/x86/kvm/vmx/pmu_intel.c
index b8e0d21..1b7456b 100644
--- a/arch/x86/kvm/vmx/pmu_intel.c
+++ b/arch/x86/kvm/vmx/pmu_intel.c
@@ -118,16 +118,15 @@ static struct kvm_pmc *intel_pmc_idx_to_pmc(struct kvm_pmu *pmu, int pmc_idx)
 	}
 }
 
-/* returns 0 if idx's corresponding MSR exists; otherwise returns 1. */
-static int intel_is_valid_rdpmc_ecx(struct kvm_vcpu *vcpu, unsigned int idx)
+static bool intel_is_valid_rdpmc_ecx(struct kvm_vcpu *vcpu, unsigned int idx)
 {
 	struct kvm_pmu *pmu = vcpu_to_pmu(vcpu);
 	bool fixed = idx & (1u << 30);
 
 	idx &= ~(3u << 30);
 
-	return (!fixed && idx >= pmu->nr_arch_gp_counters) ||
-		(fixed && idx >= pmu->nr_arch_fixed_counters);
+	return fixed ? idx < pmu->nr_arch_fixed_counters
+		     : idx < pmu->nr_arch_gp_counters;
 }
 
 static struct kvm_pmc *intel_rdpmc_ecx_to_pmc(struct kvm_vcpu *vcpu,
diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index 76861b6..ba66c17 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -769,24 +769,13 @@ void vmx_update_exception_bitmap(struct kvm_vcpu *vcpu)
 /*
  * Check if MSR is intercepted for currently loaded MSR bitmap.
  */
-static bool msr_write_intercepted(struct kvm_vcpu *vcpu, u32 msr)
+static bool msr_write_intercepted(struct vcpu_vmx *vmx, u32 msr)
 {
-	unsigned long *msr_bitmap;
-	int f = sizeof(unsigned long);
-
-	if (!cpu_has_vmx_msr_bitmap())
+	if (!(exec_controls_get(vmx) & CPU_BASED_USE_MSR_BITMAPS))
 		return true;
 
-	msr_bitmap = to_vmx(vcpu)->loaded_vmcs->msr_bitmap;
-
-	if (msr <= 0x1fff) {
-		return !!test_bit(msr, msr_bitmap + 0x800 / f);
-	} else if ((msr >= 0xc0000000) && (msr <= 0xc0001fff)) {
-		msr &= 0x1fff;
-		return !!test_bit(msr, msr_bitmap + 0xc00 / f);
-	}
-
-	return true;
+	return vmx_test_msr_bitmap_write(vmx->loaded_vmcs->msr_bitmap,
+					 MSR_IA32_SPEC_CTRL);
 }
 
 static void clear_atomic_switch_msr_special(struct vcpu_vmx *vmx,
@@ -3697,46 +3686,6 @@ void free_vpid(int vpid)
 	spin_unlock(&vmx_vpid_lock);
 }
 
-static void vmx_clear_msr_bitmap_read(ulong *msr_bitmap, u32 msr)
-{
-	int f = sizeof(unsigned long);
-
-	if (msr <= 0x1fff)
-		__clear_bit(msr, msr_bitmap + 0x000 / f);
-	else if ((msr >= 0xc0000000) && (msr <= 0xc0001fff))
-		__clear_bit(msr & 0x1fff, msr_bitmap + 0x400 / f);
-}
-
-static void vmx_clear_msr_bitmap_write(ulong *msr_bitmap, u32 msr)
-{
-	int f = sizeof(unsigned long);
-
-	if (msr <= 0x1fff)
-		__clear_bit(msr, msr_bitmap + 0x800 / f);
-	else if ((msr >= 0xc0000000) && (msr <= 0xc0001fff))
-		__clear_bit(msr & 0x1fff, msr_bitmap + 0xc00 / f);
-}
-
-static void vmx_set_msr_bitmap_read(ulong *msr_bitmap, u32 msr)
-{
-	int f = sizeof(unsigned long);
-
-	if (msr <= 0x1fff)
-		__set_bit(msr, msr_bitmap + 0x000 / f);
-	else if ((msr >= 0xc0000000) && (msr <= 0xc0001fff))
-		__set_bit(msr & 0x1fff, msr_bitmap + 0x400 / f);
-}
-
-static void vmx_set_msr_bitmap_write(ulong *msr_bitmap, u32 msr)
-{
-	int f = sizeof(unsigned long);
-
-	if (msr <= 0x1fff)
-		__set_bit(msr, msr_bitmap + 0x800 / f);
-	else if ((msr >= 0xc0000000) && (msr <= 0xc0001fff))
-		__set_bit(msr & 0x1fff, msr_bitmap + 0xc00 / f);
-}
-
 void vmx_disable_intercept_for_msr(struct kvm_vcpu *vcpu, u32 msr, int type)
 {
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
@@ -5494,6 +5443,7 @@ static int handle_invpcid(struct kvm_vcpu *vcpu)
 		u64 pcid;
 		u64 gla;
 	} operand;
+	int gpr_index;
 
 	if (!guest_cpuid_has(vcpu, X86_FEATURE_INVPCID)) {
 		kvm_queue_exception(vcpu, UD_VECTOR);
@@ -5501,12 +5451,8 @@ static int handle_invpcid(struct kvm_vcpu *vcpu)
 	}
 
 	vmx_instruction_info = vmcs_read32(VMX_INSTRUCTION_INFO);
-	type = kvm_register_read(vcpu, (vmx_instruction_info >> 28) & 0xf);
-
-	if (type > 3) {
-		kvm_inject_gp(vcpu, 0);
-		return 1;
-	}
+	gpr_index = vmx_get_instr_info_reg2(vmx_instruction_info);
+	type = kvm_register_read(vcpu, gpr_index);
 
 	/* According to the Intel instruction reference, the memory operand
 	 * is read even if it isn't needed (e.g., for type==all)
@@ -6749,7 +6695,7 @@ static fastpath_t vmx_vcpu_run(struct kvm_vcpu *vcpu)
 	 * If the L02 MSR bitmap does not intercept the MSR, then we need to
 	 * save it.
 	 */
-	if (unlikely(!msr_write_intercepted(vcpu, MSR_IA32_SPEC_CTRL)))
+	if (unlikely(!msr_write_intercepted(vmx, MSR_IA32_SPEC_CTRL)))
 		vmx->spec_ctrl = native_read_msr(MSR_IA32_SPEC_CTRL);
 
 	x86_spec_ctrl_restore_host(vmx->spec_ctrl, 0);
@@ -7563,7 +7509,8 @@ static void hardware_unsetup(void)
 static bool vmx_check_apicv_inhibit_reasons(ulong bit)
 {
 	ulong supported = BIT(APICV_INHIBIT_REASON_DISABLE) |
-			  BIT(APICV_INHIBIT_REASON_HYPERV);
+			  BIT(APICV_INHIBIT_REASON_HYPERV) |
+			  BIT(APICV_INHIBIT_REASON_BLOCKIRQ);
 
 	return supported & BIT(bit);
 }
diff --git a/arch/x86/kvm/vmx/vmx.h b/arch/x86/kvm/vmx/vmx.h
index e7db42e..a4ead60 100644
--- a/arch/x86/kvm/vmx/vmx.h
+++ b/arch/x86/kvm/vmx/vmx.h
@@ -400,6 +400,34 @@ static inline void vmx_set_intercept_for_msr(struct kvm_vcpu *vcpu, u32 msr,
 
 void vmx_update_cpu_dirty_logging(struct kvm_vcpu *vcpu);
 
+/*
+ * Note, early Intel manuals have the write-low and read-high bitmap offsets
+ * the wrong way round.  The bitmaps control MSRs 0x00000000-0x00001fff and
+ * 0xc0000000-0xc0001fff.  The former (low) uses bytes 0-0x3ff for reads and
+ * 0x800-0xbff for writes.  The latter (high) uses 0x400-0x7ff for reads and
+ * 0xc00-0xfff for writes.  MSRs not covered by either of the ranges always
+ * VM-Exit.
+ */
+#define __BUILD_VMX_MSR_BITMAP_HELPER(rtype, action, bitop, access, base)      \
+static inline rtype vmx_##action##_msr_bitmap_##access(unsigned long *bitmap,  \
+						       u32 msr)		       \
+{									       \
+	int f = sizeof(unsigned long);					       \
+									       \
+	if (msr <= 0x1fff)						       \
+		return bitop##_bit(msr, bitmap + base / f);		       \
+	else if ((msr >= 0xc0000000) && (msr <= 0xc0001fff))		       \
+		return bitop##_bit(msr & 0x1fff, bitmap + (base + 0x400) / f); \
+	return (rtype)true;						       \
+}
+#define BUILD_VMX_MSR_BITMAP_HELPERS(ret_type, action, bitop)		       \
+	__BUILD_VMX_MSR_BITMAP_HELPER(ret_type, action, bitop, read,  0x0)     \
+	__BUILD_VMX_MSR_BITMAP_HELPER(ret_type, action, bitop, write, 0x800)
+
+BUILD_VMX_MSR_BITMAP_HELPERS(bool, test, test)
+BUILD_VMX_MSR_BITMAP_HELPERS(void, clear, __clear)
+BUILD_VMX_MSR_BITMAP_HELPERS(void, set, __set)
+
 static inline u8 vmx_get_rvi(void)
 {
 	return vmcs_read16(GUEST_INTR_STATUS) & 0xff;
@@ -522,4 +550,9 @@ static inline bool vmx_guest_state_valid(struct kvm_vcpu *vcpu)
 
 void dump_vmcs(struct kvm_vcpu *vcpu);
 
+static inline int vmx_get_instr_info_reg2(u32 vmx_instr_info)
+{
+	return (vmx_instr_info >> 28) & 0xf;
+}
+
 #endif /* __KVM_X86_VMX_H */
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index c1c4e2b..dc7eb5f 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -3260,8 +3260,11 @@ static void kvm_vcpu_flush_tlb_guest(struct kvm_vcpu *vcpu)
 
 static void record_steal_time(struct kvm_vcpu *vcpu)
 {
-	struct kvm_host_map map;
-	struct kvm_steal_time *st;
+	struct gfn_to_hva_cache *ghc = &vcpu->arch.st.cache;
+	struct kvm_steal_time __user *st;
+	struct kvm_memslots *slots;
+	u64 steal;
+	u32 version;
 
 	if (kvm_xen_msr_enabled(vcpu->kvm)) {
 		kvm_xen_runstate_set_running(vcpu);
@@ -3271,47 +3274,86 @@ static void record_steal_time(struct kvm_vcpu *vcpu)
 	if (!(vcpu->arch.st.msr_val & KVM_MSR_ENABLED))
 		return;
 
-	/* -EAGAIN is returned in atomic context so we can just return. */
-	if (kvm_map_gfn(vcpu, vcpu->arch.st.msr_val >> PAGE_SHIFT,
-			&map, &vcpu->arch.st.cache, false))
+	if (WARN_ON_ONCE(current->mm != vcpu->kvm->mm))
 		return;
 
-	st = map.hva +
-		offset_in_page(vcpu->arch.st.msr_val & KVM_STEAL_VALID_BITS);
+	slots = kvm_memslots(vcpu->kvm);
 
+	if (unlikely(slots->generation != ghc->generation ||
+		     kvm_is_error_hva(ghc->hva) || !ghc->memslot)) {
+		gfn_t gfn = vcpu->arch.st.msr_val & KVM_STEAL_VALID_BITS;
+
+		/* We rely on the fact that it fits in a single page. */
+		BUILD_BUG_ON((sizeof(*st) - 1) & KVM_STEAL_VALID_BITS);
+
+		if (kvm_gfn_to_hva_cache_init(vcpu->kvm, ghc, gfn, sizeof(*st)) ||
+		    kvm_is_error_hva(ghc->hva) || !ghc->memslot)
+			return;
+	}
+
+	st = (struct kvm_steal_time __user *)ghc->hva;
 	/*
 	 * Doing a TLB flush here, on the guest's behalf, can avoid
 	 * expensive IPIs.
 	 */
 	if (guest_pv_has(vcpu, KVM_FEATURE_PV_TLB_FLUSH)) {
-		u8 st_preempted = xchg(&st->preempted, 0);
+		u8 st_preempted = 0;
+		int err = -EFAULT;
+
+		if (!user_access_begin(st, sizeof(*st)))
+			return;
+
+		asm volatile("1: xchgb %0, %2\n"
+			     "xor %1, %1\n"
+			     "2:\n"
+			     _ASM_EXTABLE_UA(1b, 2b)
+			     : "+r" (st_preempted),
+			       "+&r" (err)
+			     : "m" (st->preempted));
+		if (err)
+			goto out;
+
+		user_access_end();
+
+		vcpu->arch.st.preempted = 0;
 
 		trace_kvm_pv_tlb_flush(vcpu->vcpu_id,
 				       st_preempted & KVM_VCPU_FLUSH_TLB);
 		if (st_preempted & KVM_VCPU_FLUSH_TLB)
 			kvm_vcpu_flush_tlb_guest(vcpu);
+
+		if (!user_access_begin(st, sizeof(*st)))
+			goto dirty;
 	} else {
-		st->preempted = 0;
+		if (!user_access_begin(st, sizeof(*st)))
+			return;
+
+		unsafe_put_user(0, &st->preempted, out);
+		vcpu->arch.st.preempted = 0;
 	}
 
-	vcpu->arch.st.preempted = 0;
+	unsafe_get_user(version, &st->version, out);
+	if (version & 1)
+		version += 1;  /* first time write, random junk */
 
-	if (st->version & 1)
-		st->version += 1;  /* first time write, random junk */
-
-	st->version += 1;
+	version += 1;
+	unsafe_put_user(version, &st->version, out);
 
 	smp_wmb();
 
-	st->steal += current->sched_info.run_delay -
+	unsafe_get_user(steal, &st->steal, out);
+	steal += current->sched_info.run_delay -
 		vcpu->arch.st.last_steal;
 	vcpu->arch.st.last_steal = current->sched_info.run_delay;
+	unsafe_put_user(steal, &st->steal, out);
 
-	smp_wmb();
+	version += 1;
+	unsafe_put_user(version, &st->version, out);
 
-	st->version += 1;
-
-	kvm_unmap_gfn(vcpu, &map, &vcpu->arch.st.cache, true, false);
+ out:
+	user_access_end();
+ dirty:
+	mark_page_dirty_in_slot(vcpu->kvm, ghc->memslot, gpa_to_gfn(ghc->gpa));
 }
 
 int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
@@ -3517,7 +3559,7 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 		if (!guest_pv_has(vcpu, KVM_FEATURE_PV_EOI))
 			return 1;
 
-		if (kvm_lapic_enable_pv_eoi(vcpu, data, sizeof(u8)))
+		if (kvm_lapic_set_pv_eoi(vcpu, data, sizeof(u8)))
 			return 1;
 		break;
 
@@ -4137,7 +4179,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 		r = !static_call(kvm_x86_cpu_has_accelerated_tpr)();
 		break;
 	case KVM_CAP_NR_VCPUS:
-		r = KVM_SOFT_MAX_VCPUS;
+		r = num_online_cpus();
 		break;
 	case KVM_CAP_MAX_VCPUS:
 		r = KVM_MAX_VCPUS;
@@ -4351,8 +4393,10 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
 
 static void kvm_steal_time_set_preempted(struct kvm_vcpu *vcpu)
 {
-	struct kvm_host_map map;
-	struct kvm_steal_time *st;
+	struct gfn_to_hva_cache *ghc = &vcpu->arch.st.cache;
+	struct kvm_steal_time __user *st;
+	struct kvm_memslots *slots;
+	static const u8 preempted = KVM_VCPU_PREEMPTED;
 
 	if (!(vcpu->arch.st.msr_val & KVM_MSR_ENABLED))
 		return;
@@ -4360,16 +4404,23 @@ static void kvm_steal_time_set_preempted(struct kvm_vcpu *vcpu)
 	if (vcpu->arch.st.preempted)
 		return;
 
-	if (kvm_map_gfn(vcpu, vcpu->arch.st.msr_val >> PAGE_SHIFT, &map,
-			&vcpu->arch.st.cache, true))
+	/* This happens on process exit */
+	if (unlikely(current->mm != vcpu->kvm->mm))
 		return;
 
-	st = map.hva +
-		offset_in_page(vcpu->arch.st.msr_val & KVM_STEAL_VALID_BITS);
+	slots = kvm_memslots(vcpu->kvm);
 
-	st->preempted = vcpu->arch.st.preempted = KVM_VCPU_PREEMPTED;
+	if (unlikely(slots->generation != ghc->generation ||
+		     kvm_is_error_hva(ghc->hva) || !ghc->memslot))
+		return;
 
-	kvm_unmap_gfn(vcpu, &map, &vcpu->arch.st.cache, true, true);
+	st = (struct kvm_steal_time __user *)ghc->hva;
+	BUILD_BUG_ON(sizeof(st->preempted) != sizeof(preempted));
+
+	if (!copy_to_user_nofault(&st->preempted, &preempted, sizeof(preempted)))
+		vcpu->arch.st.preempted = KVM_VCPU_PREEMPTED;
+
+	mark_page_dirty_in_slot(vcpu->kvm, ghc->memslot, gpa_to_gfn(ghc->gpa));
 }
 
 void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
@@ -5728,6 +5779,12 @@ int kvm_vm_ioctl_enable_cap(struct kvm *kvm,
 		if (kvm_x86_ops.vm_copy_enc_context_from)
 			r = kvm_x86_ops.vm_copy_enc_context_from(kvm, cap->args[0]);
 		return r;
+	case KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM:
+		r = -EINVAL;
+		if (kvm_x86_ops.vm_move_enc_context_from)
+			r = kvm_x86_ops.vm_move_enc_context_from(
+				kvm, cap->args[0]);
+		return r;
 	case KVM_CAP_EXIT_HYPERCALL:
 		if (cap->args[0] & ~KVM_EXIT_HYPERCALL_VALID_MASK) {
 			r = -EINVAL;
@@ -7328,7 +7385,9 @@ static void emulator_set_smbase(struct x86_emulate_ctxt *ctxt, u64 smbase)
 static int emulator_check_pmc(struct x86_emulate_ctxt *ctxt,
 			      u32 pmc)
 {
-	return kvm_pmu_is_valid_rdpmc_ecx(emul_to_vcpu(ctxt), pmc);
+	if (kvm_pmu_is_valid_rdpmc_ecx(emul_to_vcpu(ctxt), pmc))
+		return 0;
+	return -EINVAL;
 }
 
 static int emulator_read_pmc(struct x86_emulate_ctxt *ctxt,
@@ -9552,7 +9611,7 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
 	}
 
 	if (kvm_request_pending(vcpu)) {
-		if (kvm_check_request(KVM_REQ_VM_BUGGED, vcpu)) {
+		if (kvm_check_request(KVM_REQ_VM_DEAD, vcpu)) {
 			r = -EIO;
 			goto out;
 		}
@@ -10564,6 +10623,24 @@ int kvm_arch_vcpu_ioctl_set_sregs(struct kvm_vcpu *vcpu,
 	return ret;
 }
 
+static void kvm_arch_vcpu_guestdbg_update_apicv_inhibit(struct kvm *kvm)
+{
+	bool inhibit = false;
+	struct kvm_vcpu *vcpu;
+	int i;
+
+	down_write(&kvm->arch.apicv_update_lock);
+
+	kvm_for_each_vcpu(i, vcpu, kvm) {
+		if (vcpu->guest_debug & KVM_GUESTDBG_BLOCKIRQ) {
+			inhibit = true;
+			break;
+		}
+	}
+	__kvm_request_apicv_update(kvm, !inhibit, APICV_INHIBIT_REASON_BLOCKIRQ);
+	up_write(&kvm->arch.apicv_update_lock);
+}
+
 int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
 					struct kvm_guest_debug *dbg)
 {
@@ -10616,6 +10693,8 @@ int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
 
 	static_call(kvm_x86_update_exception_bitmap)(vcpu);
 
+	kvm_arch_vcpu_guestdbg_update_apicv_inhibit(vcpu->kvm);
+
 	r = 0;
 
 out:
@@ -10859,11 +10938,8 @@ void kvm_arch_vcpu_postcreate(struct kvm_vcpu *vcpu)
 
 void kvm_arch_vcpu_destroy(struct kvm_vcpu *vcpu)
 {
-	struct gfn_to_pfn_cache *cache = &vcpu->arch.st.cache;
 	int idx;
 
-	kvm_release_pfn(cache->pfn, cache->dirty, cache);
-
 	kvmclock_reset(vcpu);
 
 	static_call(kvm_x86_vcpu_free)(vcpu);
@@ -12275,7 +12351,8 @@ int kvm_handle_invpcid(struct kvm_vcpu *vcpu, unsigned long type, gva_t gva)
 		return kvm_skip_emulated_instruction(vcpu);
 
 	default:
-		BUG(); /* We have already checked above that type <= 3 */
+		kvm_inject_gp(vcpu, 0);
+		return 1;
 	}
 }
 EXPORT_SYMBOL_GPL(kvm_handle_invpcid);
diff --git a/arch/x86/mm/mem_encrypt.c b/arch/x86/mm/mem_encrypt.c
index 23d54b8..3548730 100644
--- a/arch/x86/mm/mem_encrypt.c
+++ b/arch/x86/mm/mem_encrypt.c
@@ -229,28 +229,75 @@ void __init sev_setup_arch(void)
 	swiotlb_adjust_size(size);
 }
 
+static unsigned long pg_level_to_pfn(int level, pte_t *kpte, pgprot_t *ret_prot)
+{
+	unsigned long pfn = 0;
+	pgprot_t prot;
+
+	switch (level) {
+	case PG_LEVEL_4K:
+		pfn = pte_pfn(*kpte);
+		prot = pte_pgprot(*kpte);
+		break;
+	case PG_LEVEL_2M:
+		pfn = pmd_pfn(*(pmd_t *)kpte);
+		prot = pmd_pgprot(*(pmd_t *)kpte);
+		break;
+	case PG_LEVEL_1G:
+		pfn = pud_pfn(*(pud_t *)kpte);
+		prot = pud_pgprot(*(pud_t *)kpte);
+		break;
+	default:
+		WARN_ONCE(1, "Invalid level for kpte\n");
+		return 0;
+	}
+
+	if (ret_prot)
+		*ret_prot = prot;
+
+	return pfn;
+}
+
+void notify_range_enc_status_changed(unsigned long vaddr, int npages, bool enc)
+{
+#ifdef CONFIG_PARAVIRT
+	unsigned long sz = npages << PAGE_SHIFT;
+	unsigned long vaddr_end = vaddr + sz;
+
+	while (vaddr < vaddr_end) {
+		int psize, pmask, level;
+		unsigned long pfn;
+		pte_t *kpte;
+
+		kpte = lookup_address(vaddr, &level);
+		if (!kpte || pte_none(*kpte)) {
+			WARN_ONCE(1, "kpte lookup for vaddr\n");
+			return;
+		}
+
+		pfn = pg_level_to_pfn(level, kpte, NULL);
+		if (!pfn)
+			continue;
+
+		psize = page_level_size(level);
+		pmask = page_level_mask(level);
+
+		notify_page_enc_status_changed(pfn, psize >> PAGE_SHIFT, enc);
+
+		vaddr = (vaddr & pmask) + psize;
+	}
+#endif
+}
+
 static void __init __set_clr_pte_enc(pte_t *kpte, int level, bool enc)
 {
 	pgprot_t old_prot, new_prot;
 	unsigned long pfn, pa, size;
 	pte_t new_pte;
 
-	switch (level) {
-	case PG_LEVEL_4K:
-		pfn = pte_pfn(*kpte);
-		old_prot = pte_pgprot(*kpte);
-		break;
-	case PG_LEVEL_2M:
-		pfn = pmd_pfn(*(pmd_t *)kpte);
-		old_prot = pmd_pgprot(*(pmd_t *)kpte);
-		break;
-	case PG_LEVEL_1G:
-		pfn = pud_pfn(*(pud_t *)kpte);
-		old_prot = pud_pgprot(*(pud_t *)kpte);
-		break;
-	default:
+	pfn = pg_level_to_pfn(level, kpte, &old_prot);
+	if (!pfn)
 		return;
-	}
 
 	new_prot = old_prot;
 	if (enc)
@@ -286,12 +333,13 @@ static void __init __set_clr_pte_enc(pte_t *kpte, int level, bool enc)
 static int __init early_set_memory_enc_dec(unsigned long vaddr,
 					   unsigned long size, bool enc)
 {
-	unsigned long vaddr_end, vaddr_next;
+	unsigned long vaddr_end, vaddr_next, start;
 	unsigned long psize, pmask;
 	int split_page_size_mask;
 	int level, ret;
 	pte_t *kpte;
 
+	start = vaddr;
 	vaddr_next = vaddr;
 	vaddr_end = vaddr + size;
 
@@ -346,6 +394,7 @@ static int __init early_set_memory_enc_dec(unsigned long vaddr,
 
 	ret = 0;
 
+	notify_range_enc_status_changed(start, PAGE_ALIGN(size) >> PAGE_SHIFT, enc);
 out:
 	__flush_tlb_all();
 	return ret;
@@ -361,6 +410,11 @@ int __init early_set_memory_encrypted(unsigned long vaddr, unsigned long size)
 	return early_set_memory_enc_dec(vaddr, size, true);
 }
 
+void __init early_set_mem_enc_dec_hypercall(unsigned long vaddr, int npages, bool enc)
+{
+	notify_range_enc_status_changed(vaddr, npages, enc);
+}
+
 /* Override for DMA direct allocation check - ARCH_HAS_FORCE_DMA_UNENCRYPTED */
 bool force_dma_unencrypted(struct device *dev)
 {
diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c
index 934dc5b..b407211 100644
--- a/arch/x86/mm/pat/set_memory.c
+++ b/arch/x86/mm/pat/set_memory.c
@@ -2023,6 +2023,12 @@ static int __set_memory_enc_pgtable(unsigned long addr, int numpages, bool enc)
 	 */
 	cpa_flush(&cpa, 0);
 
+	/*
+	 * Notify hypervisor that a given memory range is mapped encrypted
+	 * or decrypted.
+	 */
+	notify_range_enc_status_changed(addr, numpages, enc);
+
 	return ret;
 }
 
diff --git a/arch/x86/xen/smp_pv.c b/arch/x86/xen/smp_pv.c
index 9e55bcb..6a8f3b5 100644
--- a/arch/x86/xen/smp_pv.c
+++ b/arch/x86/xen/smp_pv.c
@@ -225,7 +225,6 @@ static void __init xen_pv_smp_prepare_boot_cpu(void)
 static void __init xen_pv_smp_prepare_cpus(unsigned int max_cpus)
 {
 	unsigned cpu;
-	unsigned int i;
 
 	if (skip_ioapic_setup) {
 		char *m = (max_cpus == 0) ?
@@ -238,16 +237,9 @@ static void __init xen_pv_smp_prepare_cpus(unsigned int max_cpus)
 	}
 	xen_init_lock_cpu(0);
 
-	smp_store_boot_cpu_info();
-	cpu_data(0).x86_max_cores = 1;
+	smp_prepare_cpus_common();
 
-	for_each_possible_cpu(i) {
-		zalloc_cpumask_var(&per_cpu(cpu_sibling_map, i), GFP_KERNEL);
-		zalloc_cpumask_var(&per_cpu(cpu_core_map, i), GFP_KERNEL);
-		zalloc_cpumask_var(&per_cpu(cpu_die_map, i), GFP_KERNEL);
-		zalloc_cpumask_var(&per_cpu(cpu_llc_shared_map, i), GFP_KERNEL);
-	}
-	set_cpu_sibling_map(0);
+	cpu_data(0).x86_max_cores = 1;
 
 	speculative_store_bypass_ht_init();
 
diff --git a/block/blk-core.c b/block/blk-core.c
index b043de2..9ee32f8 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -809,10 +809,8 @@ noinline_for_stack bool submit_bio_checks(struct bio *bio)
 	if (unlikely(!current->io_context))
 		create_task_io_context(current, GFP_ATOMIC, q->node);
 
-	if (blk_throtl_bio(bio)) {
-		blkcg_bio_issue_init(bio);
+	if (blk_throtl_bio(bio))
 		return false;
-	}
 
 	blk_cgroup_bio_start(bio);
 	blkcg_bio_issue_init(bio);
diff --git a/block/blk-ia-ranges.c b/block/blk-ia-ranges.c
index c246c42..b925f3d 100644
--- a/block/blk-ia-ranges.c
+++ b/block/blk-ia-ranges.c
@@ -104,8 +104,8 @@ static struct kobj_type blk_ia_ranges_ktype = {
 };
 
 /**
- * disk_register_ia_ranges - register with sysfs a set of independent
- *			    access ranges
+ * disk_register_independent_access_ranges - register with sysfs a set of
+ *		independent access ranges
  * @disk:	Target disk
  * @new_iars:	New set of independent access ranges
  *
diff --git a/block/blk-mq-sched.c b/block/blk-mq-sched.c
index 4be652f..ba21449 100644
--- a/block/blk-mq-sched.c
+++ b/block/blk-mq-sched.c
@@ -370,9 +370,6 @@ bool blk_mq_sched_bio_merge(struct request_queue *q, struct bio *bio,
 	bool ret = false;
 	enum hctx_type type;
 
-	if (bio_queue_enter(bio))
-		return false;
-
 	if (e && e->type->ops.bio_merge) {
 		ret = e->type->ops.bio_merge(q, bio, nr_segs);
 		goto out_put;
@@ -397,7 +394,6 @@ bool blk_mq_sched_bio_merge(struct request_queue *q, struct bio *bio,
 
 	spin_unlock(&ctx->lock);
 out_put:
-	blk_queue_exit(q);
 	return ret;
 }
 
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 629cf42..3ab34c4 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -2495,8 +2495,9 @@ static inline unsigned short blk_plug_max_rq_count(struct blk_plug *plug)
 	return BLK_MAX_REQUEST_COUNT;
 }
 
-static bool blk_attempt_bio_merge(struct request_queue *q, struct bio *bio,
-				  unsigned int nr_segs, bool *same_queue_rq)
+static bool blk_mq_attempt_bio_merge(struct request_queue *q,
+				     struct bio *bio, unsigned int nr_segs,
+				     bool *same_queue_rq)
 {
 	if (!blk_queue_nomerges(q) && bio_mergeable(bio)) {
 		if (blk_attempt_plug_merge(q, bio, nr_segs, same_queue_rq))
@@ -2520,12 +2521,8 @@ static struct request *blk_mq_get_new_requests(struct request_queue *q,
 	};
 	struct request *rq;
 
-	if (unlikely(bio_queue_enter(bio)))
+	if (blk_mq_attempt_bio_merge(q, bio, nsegs, same_queue_rq))
 		return NULL;
-	if (unlikely(!submit_bio_checks(bio)))
-		goto put_exit;
-	if (blk_attempt_bio_merge(q, bio, nsegs, same_queue_rq))
-		goto put_exit;
 
 	rq_qos_throttle(q, bio);
 
@@ -2542,26 +2539,44 @@ static struct request *blk_mq_get_new_requests(struct request_queue *q,
 	rq_qos_cleanup(q, bio);
 	if (bio->bi_opf & REQ_NOWAIT)
 		bio_wouldblock_error(bio);
-put_exit:
-	blk_queue_exit(q);
+
 	return NULL;
 }
 
+static inline bool blk_mq_can_use_cached_rq(struct request *rq,
+		struct bio *bio)
+{
+	if (blk_mq_get_hctx_type(bio->bi_opf) != rq->mq_hctx->type)
+		return false;
+
+	if (op_is_flush(rq->cmd_flags) != op_is_flush(bio->bi_opf))
+		return false;
+
+	return true;
+}
+
 static inline struct request *blk_mq_get_request(struct request_queue *q,
 						 struct blk_plug *plug,
 						 struct bio *bio,
 						 unsigned int nsegs,
 						 bool *same_queue_rq)
 {
+	struct request *rq;
+	bool checked = false;
+
 	if (plug) {
-		struct request *rq;
 
 		rq = rq_list_peek(&plug->cached_rq);
 		if (rq && rq->q == q) {
 			if (unlikely(!submit_bio_checks(bio)))
 				return NULL;
-			if (blk_attempt_bio_merge(q, bio, nsegs, same_queue_rq))
+			if (blk_mq_attempt_bio_merge(q, bio, nsegs,
+						same_queue_rq))
 				return NULL;
+			checked = true;
+			if (!blk_mq_can_use_cached_rq(rq, bio))
+				goto fallback;
+			rq->cmd_flags = bio->bi_opf;
 			plug->cached_rq = rq_list_next(rq);
 			INIT_LIST_HEAD(&rq->queuelist);
 			rq_qos_throttle(q, bio);
@@ -2569,7 +2584,15 @@ static inline struct request *blk_mq_get_request(struct request_queue *q,
 		}
 	}
 
-	return blk_mq_get_new_requests(q, plug, bio, nsegs, same_queue_rq);
+fallback:
+	if (unlikely(bio_queue_enter(bio)))
+		return NULL;
+	if (!checked && !submit_bio_checks(bio))
+		return NULL;
+	rq = blk_mq_get_new_requests(q, plug, bio, nsegs, same_queue_rq);
+	if (!rq)
+		blk_queue_exit(q);
+	return rq;
 }
 
 /**
diff --git a/block/blk-mq.h b/block/blk-mq.h
index cb0b548..8acfa65 100644
--- a/block/blk-mq.h
+++ b/block/blk-mq.h
@@ -89,15 +89,7 @@ static inline struct blk_mq_hw_ctx *blk_mq_map_queue_type(struct request_queue *
 	return q->queue_hw_ctx[q->tag_set->map[type].mq_map[cpu]];
 }
 
-/*
- * blk_mq_map_queue() - map (cmd_flags,type) to hardware queue
- * @q: request queue
- * @flags: request command flags
- * @ctx: software queue cpu ctx
- */
-static inline struct blk_mq_hw_ctx *blk_mq_map_queue(struct request_queue *q,
-						     unsigned int flags,
-						     struct blk_mq_ctx *ctx)
+static inline enum hctx_type blk_mq_get_hctx_type(unsigned int flags)
 {
 	enum hctx_type type = HCTX_TYPE_DEFAULT;
 
@@ -108,8 +100,20 @@ static inline struct blk_mq_hw_ctx *blk_mq_map_queue(struct request_queue *q,
 		type = HCTX_TYPE_POLL;
 	else if ((flags & REQ_OP_MASK) == REQ_OP_READ)
 		type = HCTX_TYPE_READ;
-	
-	return ctx->hctxs[type];
+	return type;
+}
+
+/*
+ * blk_mq_map_queue() - map (cmd_flags,type) to hardware queue
+ * @q: request queue
+ * @flags: request command flags
+ * @ctx: software queue cpu ctx
+ */
+static inline struct blk_mq_hw_ctx *blk_mq_map_queue(struct request_queue *q,
+						     unsigned int flags,
+						     struct blk_mq_ctx *ctx)
+{
+	return ctx->hctxs[blk_mq_get_hctx_type(flags)];
 }
 
 /*
@@ -149,7 +153,7 @@ struct blk_mq_alloc_data {
 	blk_mq_req_flags_t flags;
 	unsigned int shallow_depth;
 	unsigned int cmd_flags;
-	unsigned int rq_flags;
+	req_flags_t rq_flags;
 
 	/* allocate multiple requests/tags in one go */
 	unsigned int nr_tags;
diff --git a/block/blk-zoned.c b/block/blk-zoned.c
index 1d0c76c..774ecc5 100644
--- a/block/blk-zoned.c
+++ b/block/blk-zoned.c
@@ -429,9 +429,10 @@ int blkdev_zone_mgmt_ioctl(struct block_device *bdev, fmode_t mode,
 		op = REQ_OP_ZONE_RESET;
 
 		/* Invalidate the page cache, including dirty pages. */
+		filemap_invalidate_lock(bdev->bd_inode->i_mapping);
 		ret = blkdev_truncate_zone_range(bdev, mode, &zrange);
 		if (ret)
-			return ret;
+			goto fail;
 		break;
 	case BLKOPENZONE:
 		op = REQ_OP_ZONE_OPEN;
@@ -449,15 +450,9 @@ int blkdev_zone_mgmt_ioctl(struct block_device *bdev, fmode_t mode,
 	ret = blkdev_zone_mgmt(bdev, op, zrange.sector, zrange.nr_sectors,
 			       GFP_KERNEL);
 
-	/*
-	 * Invalidate the page cache again for zone reset: writes can only be
-	 * direct for zoned devices so concurrent writes would not add any page
-	 * to the page cache after/during reset. The page cache may be filled
-	 * again due to concurrent reads though and dropping the pages for
-	 * these is fine.
-	 */
-	if (!ret && cmd == BLKRESETZONE)
-		ret = blkdev_truncate_zone_range(bdev, mode, &zrange);
+fail:
+	if (cmd == BLKRESETZONE)
+		filemap_invalidate_unlock(bdev->bd_inode->i_mapping);
 
 	return ret;
 }
diff --git a/block/genhd.c b/block/genhd.c
index ca2fbab..c5392cc 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -394,8 +394,8 @@ static void disk_scan_partitions(struct gendisk *disk)
  * This function registers the partitioning information in @disk
  * with the kernel.
  */
-int device_add_disk(struct device *parent, struct gendisk *disk,
-		     const struct attribute_group **groups)
+int __must_check device_add_disk(struct device *parent, struct gendisk *disk,
+				 const struct attribute_group **groups)
 
 {
 	struct device *ddev = disk_to_dev(disk);
@@ -544,7 +544,7 @@ int device_add_disk(struct device *parent, struct gendisk *disk,
 out_free_ext_minor:
 	if (disk->major == BLOCK_EXT_MAJOR)
 		blk_free_ext_minor(disk->first_minor);
-	return WARN_ON_ONCE(ret); /* keep until all callers handle errors */
+	return ret;
 }
 EXPORT_SYMBOL(device_add_disk);
 
diff --git a/block/ioctl.c b/block/ioctl.c
index d6af0ac..0a1d10a 100644
--- a/block/ioctl.c
+++ b/block/ioctl.c
@@ -113,6 +113,7 @@ static int blk_ioctl_discard(struct block_device *bdev, fmode_t mode,
 	uint64_t range[2];
 	uint64_t start, len;
 	struct request_queue *q = bdev_get_queue(bdev);
+	struct inode *inode = bdev->bd_inode;
 	int err;
 
 	if (!(mode & FMODE_WRITE))
@@ -135,12 +136,17 @@ static int blk_ioctl_discard(struct block_device *bdev, fmode_t mode,
 	if (start + len > bdev_nr_bytes(bdev))
 		return -EINVAL;
 
+	filemap_invalidate_lock(inode->i_mapping);
 	err = truncate_bdev_range(bdev, mode, start, start + len - 1);
 	if (err)
-		return err;
+		goto fail;
 
-	return blkdev_issue_discard(bdev, start >> 9, len >> 9,
-				    GFP_KERNEL, flags);
+	err = blkdev_issue_discard(bdev, start >> 9, len >> 9,
+				   GFP_KERNEL, flags);
+
+fail:
+	filemap_invalidate_unlock(inode->i_mapping);
+	return err;
 }
 
 static int blk_ioctl_zeroout(struct block_device *bdev, fmode_t mode,
@@ -148,6 +154,7 @@ static int blk_ioctl_zeroout(struct block_device *bdev, fmode_t mode,
 {
 	uint64_t range[2];
 	uint64_t start, end, len;
+	struct inode *inode = bdev->bd_inode;
 	int err;
 
 	if (!(mode & FMODE_WRITE))
@@ -170,12 +177,17 @@ static int blk_ioctl_zeroout(struct block_device *bdev, fmode_t mode,
 		return -EINVAL;
 
 	/* Invalidate the page cache, including dirty pages */
+	filemap_invalidate_lock(inode->i_mapping);
 	err = truncate_bdev_range(bdev, mode, start, end);
 	if (err)
-		return err;
+		goto fail;
 
-	return blkdev_issue_zeroout(bdev, start >> 9, len >> 9, GFP_KERNEL,
-			BLKDEV_ZERO_NOUNMAP);
+	err = blkdev_issue_zeroout(bdev, start >> 9, len >> 9, GFP_KERNEL,
+				   BLKDEV_ZERO_NOUNMAP);
+
+fail:
+	filemap_invalidate_unlock(inode->i_mapping);
+	return err;
 }
 
 static int put_ushort(unsigned short __user *argp, unsigned short val)
diff --git a/crypto/algapi.c b/crypto/algapi.c
index d379fd9..a366cb3 100644
--- a/crypto/algapi.c
+++ b/crypto/algapi.c
@@ -284,6 +284,8 @@ static struct crypto_larval *__crypto_register_alg(struct crypto_alg *alg)
 
 	if (larval)
 		list_add(&larval->alg.cra_list, &crypto_alg_list);
+	else
+		alg->cra_flags |= CRYPTO_ALG_TESTED;
 
 	crypto_stats_init(alg);
 
diff --git a/crypto/zstd.c b/crypto/zstd.c
index 1a3309f..154a969 100644
--- a/crypto/zstd.c
+++ b/crypto/zstd.c
@@ -18,22 +18,22 @@
 #define ZSTD_DEF_LEVEL	3
 
 struct zstd_ctx {
-	ZSTD_CCtx *cctx;
-	ZSTD_DCtx *dctx;
+	zstd_cctx *cctx;
+	zstd_dctx *dctx;
 	void *cwksp;
 	void *dwksp;
 };
 
-static ZSTD_parameters zstd_params(void)
+static zstd_parameters zstd_params(void)
 {
-	return ZSTD_getParams(ZSTD_DEF_LEVEL, 0, 0);
+	return zstd_get_params(ZSTD_DEF_LEVEL, 0);
 }
 
 static int zstd_comp_init(struct zstd_ctx *ctx)
 {
 	int ret = 0;
-	const ZSTD_parameters params = zstd_params();
-	const size_t wksp_size = ZSTD_CCtxWorkspaceBound(params.cParams);
+	const zstd_parameters params = zstd_params();
+	const size_t wksp_size = zstd_cctx_workspace_bound(&params.cParams);
 
 	ctx->cwksp = vzalloc(wksp_size);
 	if (!ctx->cwksp) {
@@ -41,7 +41,7 @@ static int zstd_comp_init(struct zstd_ctx *ctx)
 		goto out;
 	}
 
-	ctx->cctx = ZSTD_initCCtx(ctx->cwksp, wksp_size);
+	ctx->cctx = zstd_init_cctx(ctx->cwksp, wksp_size);
 	if (!ctx->cctx) {
 		ret = -EINVAL;
 		goto out_free;
@@ -56,7 +56,7 @@ static int zstd_comp_init(struct zstd_ctx *ctx)
 static int zstd_decomp_init(struct zstd_ctx *ctx)
 {
 	int ret = 0;
-	const size_t wksp_size = ZSTD_DCtxWorkspaceBound();
+	const size_t wksp_size = zstd_dctx_workspace_bound();
 
 	ctx->dwksp = vzalloc(wksp_size);
 	if (!ctx->dwksp) {
@@ -64,7 +64,7 @@ static int zstd_decomp_init(struct zstd_ctx *ctx)
 		goto out;
 	}
 
-	ctx->dctx = ZSTD_initDCtx(ctx->dwksp, wksp_size);
+	ctx->dctx = zstd_init_dctx(ctx->dwksp, wksp_size);
 	if (!ctx->dctx) {
 		ret = -EINVAL;
 		goto out_free;
@@ -152,10 +152,10 @@ static int __zstd_compress(const u8 *src, unsigned int slen,
 {
 	size_t out_len;
 	struct zstd_ctx *zctx = ctx;
-	const ZSTD_parameters params = zstd_params();
+	const zstd_parameters params = zstd_params();
 
-	out_len = ZSTD_compressCCtx(zctx->cctx, dst, *dlen, src, slen, params);
-	if (ZSTD_isError(out_len))
+	out_len = zstd_compress_cctx(zctx->cctx, dst, *dlen, src, slen, &params);
+	if (zstd_is_error(out_len))
 		return -EINVAL;
 	*dlen = out_len;
 	return 0;
@@ -182,8 +182,8 @@ static int __zstd_decompress(const u8 *src, unsigned int slen,
 	size_t out_len;
 	struct zstd_ctx *zctx = ctx;
 
-	out_len = ZSTD_decompressDCtx(zctx->dctx, dst, *dlen, src, slen);
-	if (ZSTD_isError(out_len))
+	out_len = zstd_decompress_dctx(zctx->dctx, dst, *dlen, src, slen);
+	if (zstd_is_error(out_len))
 		return -EINVAL;
 	*dlen = out_len;
 	return 0;
diff --git a/drivers/ata/libahci.c b/drivers/ata/libahci.c
index 28430c0..8a6835b 100644
--- a/drivers/ata/libahci.c
+++ b/drivers/ata/libahci.c
@@ -131,7 +131,7 @@ const struct attribute_group *ahci_shost_groups[] = {
 };
 EXPORT_SYMBOL_GPL(ahci_shost_groups);
 
-struct attribute *ahci_sdev_attrs[] = {
+static struct attribute *ahci_sdev_attrs[] = {
 	&dev_attr_sw_activity.attr,
 	&dev_attr_unload_heads.attr,
 	&dev_attr_ncq_prio_supported.attr,
diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c
index 3018ca8..8a0ccb1 100644
--- a/drivers/ata/libata-core.c
+++ b/drivers/ata/libata-core.c
@@ -2052,8 +2052,19 @@ static bool ata_identify_page_supported(struct ata_device *dev, u8 page)
 	struct ata_port *ap = dev->link->ap;
 	unsigned int err, i;
 
+	if (dev->horkage & ATA_HORKAGE_NO_ID_DEV_LOG)
+		return false;
+
 	if (!ata_log_supported(dev, ATA_LOG_IDENTIFY_DEVICE)) {
-		ata_dev_warn(dev, "ATA Identify Device Log not supported\n");
+		/*
+		 * IDENTIFY DEVICE data log is defined as mandatory starting
+		 * with ACS-3 (ATA version 10). Warn about the missing log
+		 * for drives which implement this ATA level or above.
+		 */
+		if (ata_id_major_version(dev->id) >= 10)
+			ata_dev_warn(dev,
+				"ATA Identify Device Log not supported\n");
+		dev->horkage |= ATA_HORKAGE_NO_ID_DEV_LOG;
 		return false;
 	}
 
diff --git a/drivers/ata/libata-eh.c b/drivers/ata/libata-eh.c
index bf9c4b6..1d4a6f1 100644
--- a/drivers/ata/libata-eh.c
+++ b/drivers/ata/libata-eh.c
@@ -93,6 +93,12 @@ static const unsigned long ata_eh_identify_timeouts[] = {
 	ULONG_MAX,
 };
 
+static const unsigned long ata_eh_revalidate_timeouts[] = {
+	15000,	/* Some drives are slow to read log pages when waking-up */
+	15000,  /* combined time till here is enough even for media access */
+	ULONG_MAX,
+};
+
 static const unsigned long ata_eh_flush_timeouts[] = {
 	15000,	/* be generous with flush */
 	15000,  /* ditto */
@@ -129,6 +135,8 @@ static const struct ata_eh_cmd_timeout_ent
 ata_eh_cmd_timeout_table[ATA_EH_CMD_TIMEOUT_TABLE_SIZE] = {
 	{ .commands = CMDS(ATA_CMD_ID_ATA, ATA_CMD_ID_ATAPI),
 	  .timeouts = ata_eh_identify_timeouts, },
+	{ .commands = CMDS(ATA_CMD_READ_LOG_EXT, ATA_CMD_READ_LOG_DMA_EXT),
+	  .timeouts = ata_eh_revalidate_timeouts, },
 	{ .commands = CMDS(ATA_CMD_READ_NATIVE_MAX, ATA_CMD_READ_NATIVE_MAX_EXT),
 	  .timeouts = ata_eh_other_timeouts, },
 	{ .commands = CMDS(ATA_CMD_SET_MAX, ATA_CMD_SET_MAX_EXT),
diff --git a/drivers/ata/sata_highbank.c b/drivers/ata/sata_highbank.c
index 8440203..b29d3f1 100644
--- a/drivers/ata/sata_highbank.c
+++ b/drivers/ata/sata_highbank.c
@@ -469,10 +469,8 @@ static int ahci_highbank_probe(struct platform_device *pdev)
 	}
 
 	irq = platform_get_irq(pdev, 0);
-	if (irq < 0) {
-		dev_err(dev, "no irq\n");
+	if (irq < 0)
 		return irq;
-	}
 	if (!irq)
 		return -EINVAL;
 
diff --git a/drivers/base/arch_topology.c b/drivers/base/arch_topology.c
index 981e72a..ff16a36 100644
--- a/drivers/base/arch_topology.c
+++ b/drivers/base/arch_topology.c
@@ -677,6 +677,8 @@ void remove_cpu_topology(unsigned int cpu)
 		cpumask_clear_cpu(cpu, topology_core_cpumask(sibling));
 	for_each_cpu(sibling, topology_sibling_cpumask(cpu))
 		cpumask_clear_cpu(cpu, topology_sibling_cpumask(sibling));
+	for_each_cpu(sibling, topology_cluster_cpumask(cpu))
+		cpumask_clear_cpu(cpu, topology_cluster_cpumask(sibling));
 	for_each_cpu(sibling, topology_llc_cpumask(cpu))
 		cpumask_clear_cpu(cpu, topology_llc_cpumask(sibling));
 
diff --git a/drivers/bus/brcmstb_gisb.c b/drivers/bus/brcmstb_gisb.c
index 4c2f7d6..183d5cc 100644
--- a/drivers/bus/brcmstb_gisb.c
+++ b/drivers/bus/brcmstb_gisb.c
@@ -485,7 +485,7 @@ static int __init brcmstb_gisb_arb_probe(struct platform_device *pdev)
 	list_add_tail(&gdev->next, &brcmstb_gisb_arb_device_list);
 
 #ifdef CONFIG_MIPS
-	board_be_handler = brcmstb_bus_error_handler;
+	mips_set_be_handler(brcmstb_bus_error_handler);
 #endif
 
 	if (list_is_singular(&brcmstb_gisb_arb_device_list)) {
diff --git a/drivers/clk/actions/owl-factor.c b/drivers/clk/actions/owl-factor.c
index f15e262..64f316c 100644
--- a/drivers/clk/actions/owl-factor.c
+++ b/drivers/clk/actions/owl-factor.c
@@ -10,7 +10,6 @@
 
 #include <linux/clk-provider.h>
 #include <linux/regmap.h>
-#include <linux/slab.h>
 
 #include "owl-factor.h"
 
diff --git a/drivers/clk/clk-ast2600.c b/drivers/clk/clk-ast2600.c
index bc3be5f..24dab23 100644
--- a/drivers/clk/clk-ast2600.c
+++ b/drivers/clk/clk-ast2600.c
@@ -51,6 +51,8 @@ static DEFINE_SPINLOCK(aspeed_g6_clk_lock);
 static struct clk_hw_onecell_data *aspeed_g6_clk_data;
 
 static void __iomem *scu_g6_base;
+/* AST2600 revision: A0, A1, A2, etc */
+static u8 soc_rev;
 
 /*
  * Clocks marked with CLK_IS_CRITICAL:
@@ -191,9 +193,8 @@ static struct clk_hw *ast2600_calc_pll(const char *name, u32 val)
 static struct clk_hw *ast2600_calc_apll(const char *name, u32 val)
 {
 	unsigned int mult, div;
-	u32 chip_id = readl(scu_g6_base + ASPEED_G6_SILICON_REV);
 
-	if (((chip_id & CHIP_REVISION_ID) >> 16) >= 2) {
+	if (soc_rev >= 2) {
 		if (val & BIT(24)) {
 			/* Pass through mode */
 			mult = div = 1;
@@ -707,7 +708,7 @@ static const u32 ast2600_a1_axi_ahb200_tbl[] = {
 static void __init aspeed_g6_cc(struct regmap *map)
 {
 	struct clk_hw *hw;
-	u32 val, div, divbits, chip_id, axi_div, ahb_div;
+	u32 val, div, divbits, axi_div, ahb_div;
 
 	clk_hw_register_fixed_rate(NULL, "clkin", NULL, 0, 25000000);
 
@@ -738,8 +739,7 @@ static void __init aspeed_g6_cc(struct regmap *map)
 		axi_div = 2;
 
 	divbits = (val >> 11) & 0x3;
-	regmap_read(map, ASPEED_G6_SILICON_REV, &chip_id);
-	if (chip_id & BIT(16)) {
+	if (soc_rev >= 1) {
 		if (!divbits) {
 			ahb_div = ast2600_a1_axi_ahb200_tbl[(val >> 8) & 0x3];
 			if (val & BIT(16))
@@ -784,6 +784,8 @@ static void __init aspeed_g6_cc_init(struct device_node *np)
 	if (!scu_g6_base)
 		return;
 
+	soc_rev = (readl(scu_g6_base + ASPEED_G6_SILICON_REV) & CHIP_REVISION_ID) >> 16;
+
 	aspeed_g6_clk_data = kzalloc(struct_size(aspeed_g6_clk_data, hws,
 				      ASPEED_G6_NUM_CLKS), GFP_KERNEL);
 	if (!aspeed_g6_clk_data)
diff --git a/drivers/clk/clk-composite.c b/drivers/clk/clk-composite.c
index c04ae0e..b9c5f90 100644
--- a/drivers/clk/clk-composite.c
+++ b/drivers/clk/clk-composite.c
@@ -97,6 +97,7 @@ static int clk_composite_determine_rate(struct clk_hw *hw,
 				return ret;
 
 			req->rate = tmp_req.rate;
+			req->best_parent_hw = tmp_req.best_parent_hw;
 			req->best_parent_rate = tmp_req.best_parent_rate;
 
 			return 0;
diff --git a/drivers/clk/clk-si5351.c b/drivers/clk/clk-si5351.c
index 57e4597..93fa8c9 100644
--- a/drivers/clk/clk-si5351.c
+++ b/drivers/clk/clk-si5351.c
@@ -1,15 +1,15 @@
 // SPDX-License-Identifier: GPL-2.0-or-later
 /*
- * clk-si5351.c: Silicon Laboratories Si5351A/B/C I2C Clock Generator
+ * clk-si5351.c: Skyworks / Silicon Labs Si5351A/B/C I2C Clock Generator
  *
  * Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
  * Rabeeh Khoury <rabeeh@solid-run.com>
  *
  * References:
  * [1] "Si5351A/B/C Data Sheet"
- *     https://www.silabs.com/Support%20Documents/TechnicalDocs/Si5351.pdf
- * [2] "Manually Generating an Si5351 Register Map"
- *     https://www.silabs.com/Support%20Documents/TechnicalDocs/AN619.pdf
+ *     https://www.skyworksinc.com/-/media/Skyworks/SL/documents/public/data-sheets/Si5351-B.pdf
+ * [2] "AN619: Manually Generating an Si5351 Register Map"
+ *     https://www.skyworksinc.com/-/media/Skyworks/SL/documents/public/application-notes/AN619.pdf
  */
 
 #include <linux/module.h>
diff --git a/drivers/clk/clk-si5351.h b/drivers/clk/clk-si5351.h
index 73dc8ef..e9e2bfd 100644
--- a/drivers/clk/clk-si5351.h
+++ b/drivers/clk/clk-si5351.h
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-or-later */
 /*
- * clk-si5351.h: Silicon Laboratories Si5351A/B/C I2C Clock Generator
+ * clk-si5351.h: Skyworks / Silicon Labs Si5351A/B/C I2C Clock Generator
  *
  * Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
  * Rabeeh Khoury <rabeeh@solid-run.com>
diff --git a/drivers/clk/clk-versaclock5.c b/drivers/clk/clk-versaclock5.c
index c6d3b1a..e7be3e5 100644
--- a/drivers/clk/clk-versaclock5.c
+++ b/drivers/clk/clk-versaclock5.c
@@ -905,7 +905,7 @@ static int vc5_get_output_config(struct i2c_client *client,
 
 static const struct of_device_id clk_vc5_of_match[];
 
-static int vc5_probe(struct i2c_client *client, const struct i2c_device_id *id)
+static int vc5_probe(struct i2c_client *client)
 {
 	unsigned int oe, sd, src_mask = 0, src_val = 0;
 	struct vc5_driver_data *vc5;
@@ -1244,7 +1244,7 @@ static struct i2c_driver vc5_driver = {
 		.pm	= &vc5_pm_ops,
 		.of_match_table = clk_vc5_of_match,
 	},
-	.probe		= vc5_probe,
+	.probe_new	= vc5_probe,
 	.remove		= vc5_remove,
 	.id_table	= vc5_id,
 };
diff --git a/drivers/clk/imx/clk.h b/drivers/clk/imx/clk.h
index 8199499..7d220a0 100644
--- a/drivers/clk/imx/clk.h
+++ b/drivers/clk/imx/clk.h
@@ -391,11 +391,11 @@ struct clk_hw *__imx8m_clk_hw_composite(const char *name,
 
 #define imx8m_clk_hw_composite(name, parent_names, reg) \
 	_imx8m_clk_hw_composite(name, parent_names, reg, \
-			IMX_COMPOSITE_CORE, IMX_COMPOSITE_CLK_FLAGS_DEFAULT)
+			0, IMX_COMPOSITE_CLK_FLAGS_DEFAULT)
 
 #define imx8m_clk_hw_composite_critical(name, parent_names, reg) \
 	_imx8m_clk_hw_composite(name, parent_names, reg, \
-			IMX_COMPOSITE_CORE, IMX_COMPOSITE_CLK_FLAGS_CRITICAL)
+			0, IMX_COMPOSITE_CLK_FLAGS_CRITICAL)
 
 #define imx8m_clk_hw_composite_bus(name, parent_names, reg)	\
 	_imx8m_clk_hw_composite(name, parent_names, reg, \
diff --git a/drivers/clk/ingenic/cgu.c b/drivers/clk/ingenic/cgu.c
index 266c759..af31633 100644
--- a/drivers/clk/ingenic/cgu.c
+++ b/drivers/clk/ingenic/cgu.c
@@ -453,15 +453,15 @@ ingenic_clk_calc_div(struct clk_hw *hw,
 	}
 
 	/* Impose hardware constraints */
-	div = min_t(unsigned, div, 1 << clk_info->div.bits);
-	div = max_t(unsigned, div, 1);
+	div = clamp_t(unsigned int, div, clk_info->div.div,
+		      clk_info->div.div << clk_info->div.bits);
 
 	/*
 	 * If the divider value itself must be divided before being written to
 	 * the divider register, we must ensure we don't have any bits set that
 	 * would be lost as a result of doing so.
 	 */
-	div /= clk_info->div.div;
+	div = DIV_ROUND_UP(div, clk_info->div.div);
 	div *= clk_info->div.div;
 
 	return div;
diff --git a/drivers/clk/mediatek/clk-mt8195-imp_iic_wrap.c b/drivers/clk/mediatek/clk-mt8195-imp_iic_wrap.c
index 0e2ac0a..4ab312e 100644
--- a/drivers/clk/mediatek/clk-mt8195-imp_iic_wrap.c
+++ b/drivers/clk/mediatek/clk-mt8195-imp_iic_wrap.c
@@ -10,8 +10,6 @@
 #include <linux/clk-provider.h>
 #include <linux/platform_device.h>
 
-#include <dt-bindings/clock/mt8195-clk.h>
-
 static const struct mtk_gate_regs imp_iic_wrap_cg_regs = {
 	.set_ofs = 0xe08,
 	.clr_ofs = 0xe04,
diff --git a/drivers/clk/qcom/gcc-msm8996.c b/drivers/clk/qcom/gcc-msm8996.c
index 3c3a7ff..9b1674b 100644
--- a/drivers/clk/qcom/gcc-msm8996.c
+++ b/drivers/clk/qcom/gcc-msm8996.c
@@ -2937,20 +2937,6 @@ static struct clk_branch gcc_smmu_aggre0_ahb_clk = {
 	},
 };
 
-static struct clk_branch gcc_aggre1_pnoc_ahb_clk = {
-	.halt_reg = 0x82014,
-	.clkr = {
-		.enable_reg = 0x82014,
-		.enable_mask = BIT(0),
-		.hw.init = &(struct clk_init_data){
-			.name = "gcc_aggre1_pnoc_ahb_clk",
-			.parent_names = (const char *[]){ "periph_noc_clk_src" },
-			.num_parents = 1,
-			.ops = &clk_branch2_ops,
-		},
-	},
-};
-
 static struct clk_branch gcc_aggre2_ufs_axi_clk = {
 	.halt_reg = 0x83014,
 	.clkr = {
@@ -3474,7 +3460,6 @@ static struct clk_regmap *gcc_msm8996_clocks[] = {
 	[GCC_AGGRE0_CNOC_AHB_CLK] = &gcc_aggre0_cnoc_ahb_clk.clkr,
 	[GCC_SMMU_AGGRE0_AXI_CLK] = &gcc_smmu_aggre0_axi_clk.clkr,
 	[GCC_SMMU_AGGRE0_AHB_CLK] = &gcc_smmu_aggre0_ahb_clk.clkr,
-	[GCC_AGGRE1_PNOC_AHB_CLK] = &gcc_aggre1_pnoc_ahb_clk.clkr,
 	[GCC_AGGRE2_UFS_AXI_CLK] = &gcc_aggre2_ufs_axi_clk.clkr,
 	[GCC_AGGRE2_USB3_AXI_CLK] = &gcc_aggre2_usb3_axi_clk.clkr,
 	[GCC_QSPI_AHB_CLK] = &gcc_qspi_ahb_clk.clkr,
diff --git a/drivers/clk/rockchip/Kconfig b/drivers/clk/rockchip/Kconfig
index 2dfd6a3..3067bdb 100644
--- a/drivers/clk/rockchip/Kconfig
+++ b/drivers/clk/rockchip/Kconfig
@@ -80,14 +80,14 @@
 	  Build the driver for RK3368 Clock Driver.
 
 config CLK_RK3399
-	tristate "Rockchip RK3399 clock controller support"
+	bool "Rockchip RK3399 clock controller support"
 	depends on ARM64 || COMPILE_TEST
 	default y
 	help
 	  Build the driver for RK3399 Clock Driver.
 
 config CLK_RK3568
-	tristate "Rockchip RK3568 clock controller support"
+	bool "Rockchip RK3568 clock controller support"
 	depends on ARM64 || COMPILE_TEST
 	default y
 	help
diff --git a/drivers/clk/rockchip/clk-rk3399.c b/drivers/clk/rockchip/clk-rk3399.c
index 7924598..306910a 100644
--- a/drivers/clk/rockchip/clk-rk3399.c
+++ b/drivers/clk/rockchip/clk-rk3399.c
@@ -1630,7 +1630,6 @@ static const struct of_device_id clk_rk3399_match_table[] = {
 	},
 	{ }
 };
-MODULE_DEVICE_TABLE(of, clk_rk3399_match_table);
 
 static int __init clk_rk3399_probe(struct platform_device *pdev)
 {
@@ -1656,7 +1655,4 @@ static struct platform_driver clk_rk3399_driver = {
 		.suppress_bind_attrs = true,
 	},
 };
-module_platform_driver_probe(clk_rk3399_driver, clk_rk3399_probe);
-
-MODULE_DESCRIPTION("Rockchip RK3399 Clock Driver");
-MODULE_LICENSE("GPL");
+builtin_platform_driver_probe(clk_rk3399_driver, clk_rk3399_probe);
diff --git a/drivers/clk/rockchip/clk-rk3568.c b/drivers/clk/rockchip/clk-rk3568.c
index 939e707..69a9e80 100644
--- a/drivers/clk/rockchip/clk-rk3568.c
+++ b/drivers/clk/rockchip/clk-rk3568.c
@@ -1693,7 +1693,6 @@ static const struct of_device_id clk_rk3568_match_table[] = {
 	},
 	{ }
 };
-MODULE_DEVICE_TABLE(of, clk_rk3568_match_table);
 
 static int __init clk_rk3568_probe(struct platform_device *pdev)
 {
@@ -1719,7 +1718,4 @@ static struct platform_driver clk_rk3568_driver = {
 		.suppress_bind_attrs = true,
 	},
 };
-module_platform_driver_probe(clk_rk3568_driver, clk_rk3568_probe);
-
-MODULE_DESCRIPTION("Rockchip RK3568 Clock Driver");
-MODULE_LICENSE("GPL");
+builtin_platform_driver_probe(clk_rk3568_driver, clk_rk3568_probe);
diff --git a/drivers/clk/uniphier/clk-uniphier-core.c b/drivers/clk/uniphier/clk-uniphier-core.c
index 1238023..46c66fa 100644
--- a/drivers/clk/uniphier/clk-uniphier-core.c
+++ b/drivers/clk/uniphier/clk-uniphier-core.c
@@ -132,6 +132,10 @@ static const struct of_device_id uniphier_clk_match[] = {
 		.compatible = "socionext,uniphier-pxs3-clock",
 		.data = uniphier_pxs3_sys_clk_data,
 	},
+	{
+		.compatible = "socionext,uniphier-nx1-clock",
+		.data = uniphier_nx1_sys_clk_data,
+	},
 	/* Media I/O clock, SD clock */
 	{
 		.compatible = "socionext,uniphier-ld4-mio-clock",
@@ -165,6 +169,10 @@ static const struct of_device_id uniphier_clk_match[] = {
 		.compatible = "socionext,uniphier-pxs3-sd-clock",
 		.data = uniphier_pro5_sd_clk_data,
 	},
+	{
+		.compatible = "socionext,uniphier-nx1-sd-clock",
+		.data = uniphier_pro5_sd_clk_data,
+	},
 	/* Peripheral clock */
 	{
 		.compatible = "socionext,uniphier-ld4-peri-clock",
@@ -198,6 +206,15 @@ static const struct of_device_id uniphier_clk_match[] = {
 		.compatible = "socionext,uniphier-pxs3-peri-clock",
 		.data = uniphier_pro4_peri_clk_data,
 	},
+	{
+		.compatible = "socionext,uniphier-nx1-peri-clock",
+		.data = uniphier_pro4_peri_clk_data,
+	},
+	/* SoC-glue clock */
+	{
+		.compatible = "socionext,uniphier-pro4-sg-clock",
+		.data = uniphier_pro4_sg_clk_data,
+	},
 	{ /* sentinel */ }
 };
 
diff --git a/drivers/clk/uniphier/clk-uniphier-sys.c b/drivers/clk/uniphier/clk-uniphier-sys.c
index 32b3017..0180470 100644
--- a/drivers/clk/uniphier/clk-uniphier-sys.c
+++ b/drivers/clk/uniphier/clk-uniphier-sys.c
@@ -20,6 +20,10 @@
 	UNIPHIER_CLK_FACTOR("sd-200m", -1, "spll", 1, 10),		\
 	UNIPHIER_CLK_FACTOR("sd-133m", -1, "spll", 1, 15)
 
+#define UNIPHIER_NX1_SYS_CLK_SD						\
+	UNIPHIER_CLK_FACTOR("sd-200m", -1, "spll", 1, 4),		\
+	UNIPHIER_CLK_FACTOR("sd-133m", -1, "spll", 1, 6)
+
 #define UNIPHIER_LD4_SYS_CLK_NAND(idx)					\
 	UNIPHIER_CLK_FACTOR("nand-50m", -1, "spll", 1, 32),		\
 	UNIPHIER_CLK_GATE("nand", (idx), "nand-50m", 0x2104, 2)
@@ -288,6 +292,8 @@ const struct uniphier_clk_data uniphier_pxs3_sys_clk_data[] = {
 	UNIPHIER_CLK_GATE("sata0", 28, NULL, 0x210c, 7),
 	UNIPHIER_CLK_GATE("sata1", 29, NULL, 0x210c, 8),
 	UNIPHIER_CLK_GATE("sata-phy", 30, NULL, 0x210c, 21),
+	UNIPHIER_LD11_SYS_CLK_AIO(40),
+	UNIPHIER_LD11_SYS_CLK_EXIV(42),
 	/* CPU gears */
 	UNIPHIER_CLK_DIV4("cpll", 2, 3, 4, 8),
 	UNIPHIER_CLK_DIV4("spll", 2, 3, 4, 8),
@@ -300,3 +306,44 @@ const struct uniphier_clk_data uniphier_pxs3_sys_clk_data[] = {
 			     "spll/4", "spll/8", "s2pll/4", "s2pll/8"),
 	{ /* sentinel */ }
 };
+
+const struct uniphier_clk_data uniphier_nx1_sys_clk_data[] = {
+	UNIPHIER_CLK_FACTOR("cpll", -1, "ref", 100, 1),		/* ARM: 2500 MHz */
+	UNIPHIER_CLK_FACTOR("spll", -1, "ref", 32, 1),		/* 800 MHz */
+	UNIPHIER_CLK_FACTOR("uart", 0, "spll", 1, 6),
+	UNIPHIER_CLK_FACTOR("i2c", 1, "spll", 1, 16),
+	UNIPHIER_NX1_SYS_CLK_SD,
+	UNIPHIER_CLK_GATE("emmc", 4, NULL, 0x2108, 8),
+	UNIPHIER_CLK_GATE("ether", 6, NULL, 0x210c, 0),
+	UNIPHIER_CLK_GATE("usb30-0", 12, NULL, 0x210c, 16),	/* =GIO */
+	UNIPHIER_CLK_GATE("usb30-1", 13, NULL, 0x210c, 20),	/* =GIO1P */
+	UNIPHIER_CLK_GATE("usb30-hsphy0", 16, NULL, 0x210c, 24),
+	UNIPHIER_CLK_GATE("usb30-ssphy0", 17, NULL, 0x210c, 25),
+	UNIPHIER_CLK_GATE("usb30-ssphy1", 18, NULL, 0x210c, 26),
+	UNIPHIER_CLK_GATE("pcie", 24, NULL, 0x210c, 8),
+	UNIPHIER_CLK_GATE("voc", 52, NULL, 0x2110, 0),
+	UNIPHIER_CLK_GATE("hdmitx", 58, NULL, 0x2110, 8),
+	/* CPU gears */
+	UNIPHIER_CLK_DIV5("cpll", 2, 4, 8, 16, 32),
+	UNIPHIER_CLK_CPUGEAR("cpu-ca53", 33, 0x8080, 0xf, 5,
+			     "cpll/2", "cpll/4", "cpll/8", "cpll/16",
+			     "cpll/32"),
+	{ /* sentinel */ }
+};
+
+const struct uniphier_clk_data uniphier_pro4_sg_clk_data[] = {
+	UNIPHIER_CLK_DIV("gpll", 4),
+	{
+		.name = "sata-ref",
+		.type = UNIPHIER_CLK_TYPE_MUX,
+		.idx = 0,
+		.data.mux = {
+			.parent_names = { "gpll/4", "ref", },
+			.num_parents = 2,
+			.reg = 0x1a28,
+			.masks = { 0x1, 0x1, },
+			.vals  = { 0x0, 0x1, },
+		},
+	},
+	{ /* sentinel */ }
+};
diff --git a/drivers/clk/uniphier/clk-uniphier.h b/drivers/clk/uniphier/clk-uniphier.h
index 9e30362..dea0c78 100644
--- a/drivers/clk/uniphier/clk-uniphier.h
+++ b/drivers/clk/uniphier/clk-uniphier.h
@@ -119,6 +119,10 @@ struct uniphier_clk_data {
 	UNIPHIER_CLK_DIV2(parent, div0, div1),			\
 	UNIPHIER_CLK_DIV2(parent, div2, div3)
 
+#define UNIPHIER_CLK_DIV5(parent, div0, div1, div2, div3, div4)	\
+	UNIPHIER_CLK_DIV4(parent, div0, div1, div2, div3),	\
+	UNIPHIER_CLK_DIV(parent, div4)
+
 struct clk_hw *uniphier_clk_register_cpugear(struct device *dev,
 					     struct regmap *regmap,
 					     const char *name,
@@ -146,9 +150,11 @@ extern const struct uniphier_clk_data uniphier_pxs2_sys_clk_data[];
 extern const struct uniphier_clk_data uniphier_ld11_sys_clk_data[];
 extern const struct uniphier_clk_data uniphier_ld20_sys_clk_data[];
 extern const struct uniphier_clk_data uniphier_pxs3_sys_clk_data[];
+extern const struct uniphier_clk_data uniphier_nx1_sys_clk_data[];
 extern const struct uniphier_clk_data uniphier_ld4_mio_clk_data[];
 extern const struct uniphier_clk_data uniphier_pro5_sd_clk_data[];
 extern const struct uniphier_clk_data uniphier_ld4_peri_clk_data[];
 extern const struct uniphier_clk_data uniphier_pro4_peri_clk_data[];
+extern const struct uniphier_clk_data uniphier_pro4_sg_clk_data[];
 
 #endif /* __CLK_UNIPHIER_H__ */
diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c
index 2ea59cb..6437b2e 100644
--- a/drivers/dma-buf/dma-buf.c
+++ b/drivers/dma-buf/dma-buf.c
@@ -67,12 +67,9 @@ static void dma_buf_release(struct dentry *dentry)
 	BUG_ON(dmabuf->vmapping_counter);
 
 	/*
-	 * Any fences that a dma-buf poll can wait on should be signaled
-	 * before releasing dma-buf. This is the responsibility of each
-	 * driver that uses the reservation objects.
-	 *
-	 * If you hit this BUG() it means someone dropped their ref to the
-	 * dma-buf while still having pending operation to the buffer.
+	 * If you hit this BUG() it could mean:
+	 * * There's a file reference imbalance in dma_buf_poll / dma_buf_poll_cb or somewhere else
+	 * * dmabuf->cb_in/out.active are non-0 despite no pending fence callback
 	 */
 	BUG_ON(dmabuf->cb_in.active || dmabuf->cb_out.active);
 
@@ -200,6 +197,7 @@ static loff_t dma_buf_llseek(struct file *file, loff_t offset, int whence)
 static void dma_buf_poll_cb(struct dma_fence *fence, struct dma_fence_cb *cb)
 {
 	struct dma_buf_poll_cb_t *dcb = (struct dma_buf_poll_cb_t *)cb;
+	struct dma_buf *dmabuf = container_of(dcb->poll, struct dma_buf, poll);
 	unsigned long flags;
 
 	spin_lock_irqsave(&dcb->poll->lock, flags);
@@ -207,21 +205,18 @@ static void dma_buf_poll_cb(struct dma_fence *fence, struct dma_fence_cb *cb)
 	dcb->active = 0;
 	spin_unlock_irqrestore(&dcb->poll->lock, flags);
 	dma_fence_put(fence);
+	/* Paired with get_file in dma_buf_poll */
+	fput(dmabuf->file);
 }
 
-static bool dma_buf_poll_shared(struct dma_resv *resv,
+static bool dma_buf_poll_add_cb(struct dma_resv *resv, bool write,
 				struct dma_buf_poll_cb_t *dcb)
 {
-	struct dma_resv_list *fobj = dma_resv_shared_list(resv);
+	struct dma_resv_iter cursor;
 	struct dma_fence *fence;
-	int i, r;
+	int r;
 
-	if (!fobj)
-		return false;
-
-	for (i = 0; i < fobj->shared_count; ++i) {
-		fence = rcu_dereference_protected(fobj->shared[i],
-						  dma_resv_held(resv));
+	dma_resv_for_each_fence(&cursor, resv, write, fence) {
 		dma_fence_get(fence);
 		r = dma_fence_add_callback(fence, &dcb->cb, dma_buf_poll_cb);
 		if (!r)
@@ -232,24 +227,6 @@ static bool dma_buf_poll_shared(struct dma_resv *resv,
 	return false;
 }
 
-static bool dma_buf_poll_excl(struct dma_resv *resv,
-			      struct dma_buf_poll_cb_t *dcb)
-{
-	struct dma_fence *fence = dma_resv_excl_fence(resv);
-	int r;
-
-	if (!fence)
-		return false;
-
-	dma_fence_get(fence);
-	r = dma_fence_add_callback(fence, &dcb->cb, dma_buf_poll_cb);
-	if (!r)
-		return true;
-	dma_fence_put(fence);
-
-	return false;
-}
-
 static __poll_t dma_buf_poll(struct file *file, poll_table *poll)
 {
 	struct dma_buf *dmabuf;
@@ -282,8 +259,10 @@ static __poll_t dma_buf_poll(struct file *file, poll_table *poll)
 		spin_unlock_irq(&dmabuf->poll.lock);
 
 		if (events & EPOLLOUT) {
-			if (!dma_buf_poll_shared(resv, dcb) &&
-			    !dma_buf_poll_excl(resv, dcb))
+			/* Paired with fput in dma_buf_poll_cb */
+			get_file(dmabuf->file);
+
+			if (!dma_buf_poll_add_cb(resv, true, dcb))
 				/* No callback queued, wake up any other waiters */
 				dma_buf_poll_cb(NULL, &dcb->cb);
 			else
@@ -303,7 +282,10 @@ static __poll_t dma_buf_poll(struct file *file, poll_table *poll)
 		spin_unlock_irq(&dmabuf->poll.lock);
 
 		if (events & EPOLLIN) {
-			if (!dma_buf_poll_excl(resv, dcb))
+			/* Paired with fput in dma_buf_poll_cb */
+			get_file(dmabuf->file);
+
+			if (!dma_buf_poll_add_cb(resv, false, dcb))
 				/* No callback queued, wake up any other waiters */
 				dma_buf_poll_cb(NULL, &dcb->cb);
 			else
@@ -1356,10 +1338,9 @@ static int dma_buf_debug_show(struct seq_file *s, void *unused)
 {
 	struct dma_buf *buf_obj;
 	struct dma_buf_attachment *attach_obj;
-	struct dma_resv *robj;
-	struct dma_resv_list *fobj;
+	struct dma_resv_iter cursor;
 	struct dma_fence *fence;
-	int count = 0, attach_count, shared_count, i;
+	int count = 0, attach_count;
 	size_t size = 0;
 	int ret;
 
@@ -1378,6 +1359,8 @@ static int dma_buf_debug_show(struct seq_file *s, void *unused)
 		if (ret)
 			goto error_unlock;
 
+
+		spin_lock(&buf_obj->name_lock);
 		seq_printf(s, "%08zu\t%08x\t%08x\t%08ld\t%s\t%08lu\t%s\n",
 				buf_obj->size,
 				buf_obj->file->f_flags, buf_obj->file->f_mode,
@@ -1385,22 +1368,12 @@ static int dma_buf_debug_show(struct seq_file *s, void *unused)
 				buf_obj->exp_name,
 				file_inode(buf_obj->file)->i_ino,
 				buf_obj->name ?: "");
+		spin_unlock(&buf_obj->name_lock);
 
-		robj = buf_obj->resv;
-		fence = dma_resv_excl_fence(robj);
-		if (fence)
-			seq_printf(s, "\tExclusive fence: %s %s %ssignalled\n",
-				   fence->ops->get_driver_name(fence),
-				   fence->ops->get_timeline_name(fence),
-				   dma_fence_is_signaled(fence) ? "" : "un");
-
-		fobj = rcu_dereference_protected(robj->fence,
-						 dma_resv_held(robj));
-		shared_count = fobj ? fobj->shared_count : 0;
-		for (i = 0; i < shared_count; i++) {
-			fence = rcu_dereference_protected(fobj->shared[i],
-							  dma_resv_held(robj));
-			seq_printf(s, "\tShared fence: %s %s %ssignalled\n",
+		dma_resv_for_each_fence(&cursor, buf_obj->resv, true, fence) {
+			seq_printf(s, "\t%s fence: %s %s %ssignalled\n",
+				   dma_resv_iter_is_exclusive(&cursor) ?
+					"Exclusive" : "Shared",
 				   fence->ops->get_driver_name(fence),
 				   fence->ops->get_timeline_name(fence),
 				   dma_fence_is_signaled(fence) ? "" : "un");
diff --git a/drivers/dma-buf/dma-resv.c b/drivers/dma-buf/dma-resv.c
index a480af9..9eb2baa 100644
--- a/drivers/dma-buf/dma-resv.c
+++ b/drivers/dma-buf/dma-resv.c
@@ -333,10 +333,14 @@ static void dma_resv_iter_restart_unlocked(struct dma_resv_iter *cursor)
 {
 	cursor->seq = read_seqcount_begin(&cursor->obj->seq);
 	cursor->index = -1;
-	if (cursor->all_fences)
+	cursor->shared_count = 0;
+	if (cursor->all_fences) {
 		cursor->fences = dma_resv_shared_list(cursor->obj);
-	else
+		if (cursor->fences)
+			cursor->shared_count = cursor->fences->shared_count;
+	} else {
 		cursor->fences = NULL;
+	}
 	cursor->is_restarted = true;
 }
 
@@ -363,7 +367,7 @@ static void dma_resv_iter_walk_unlocked(struct dma_resv_iter *cursor)
 				continue;
 
 		} else if (!cursor->fences ||
-			   cursor->index >= cursor->fences->shared_count) {
+			   cursor->index >= cursor->shared_count) {
 			cursor->fence = NULL;
 			break;
 
@@ -424,6 +428,57 @@ struct dma_fence *dma_resv_iter_next_unlocked(struct dma_resv_iter *cursor)
 EXPORT_SYMBOL(dma_resv_iter_next_unlocked);
 
 /**
+ * dma_resv_iter_first - first fence from a locked dma_resv object
+ * @cursor: cursor to record the current position
+ *
+ * Return the first fence in the dma_resv object while holding the
+ * &dma_resv.lock.
+ */
+struct dma_fence *dma_resv_iter_first(struct dma_resv_iter *cursor)
+{
+	struct dma_fence *fence;
+
+	dma_resv_assert_held(cursor->obj);
+
+	cursor->index = 0;
+	if (cursor->all_fences)
+		cursor->fences = dma_resv_shared_list(cursor->obj);
+	else
+		cursor->fences = NULL;
+
+	fence = dma_resv_excl_fence(cursor->obj);
+	if (!fence)
+		fence = dma_resv_iter_next(cursor);
+
+	cursor->is_restarted = true;
+	return fence;
+}
+EXPORT_SYMBOL_GPL(dma_resv_iter_first);
+
+/**
+ * dma_resv_iter_next - next fence from a locked dma_resv object
+ * @cursor: cursor to record the current position
+ *
+ * Return the next fences from the dma_resv object while holding the
+ * &dma_resv.lock.
+ */
+struct dma_fence *dma_resv_iter_next(struct dma_resv_iter *cursor)
+{
+	unsigned int idx;
+
+	dma_resv_assert_held(cursor->obj);
+
+	cursor->is_restarted = false;
+	if (!cursor->fences || cursor->index >= cursor->fences->shared_count)
+		return NULL;
+
+	idx = cursor->index++;
+	return rcu_dereference_protected(cursor->fences->shared[idx],
+					 dma_resv_held(cursor->obj));
+}
+EXPORT_SYMBOL_GPL(dma_resv_iter_next);
+
+/**
  * dma_resv_copy_fences - Copy all fences from src to dst.
  * @dst: the destination reservation object
  * @src: the source reservation object
@@ -448,10 +503,8 @@ int dma_resv_copy_fences(struct dma_resv *dst, struct dma_resv *src)
 			dma_resv_list_free(list);
 			dma_fence_put(excl);
 
-			if (cursor.fences) {
-				unsigned int cnt = cursor.fences->shared_count;
-
-				list = dma_resv_list_alloc(cnt);
+			if (cursor.shared_count) {
+				list = dma_resv_list_alloc(cursor.shared_count);
 				if (!list) {
 					dma_resv_iter_end(&cursor);
 					return -ENOMEM;
@@ -522,7 +575,7 @@ int dma_resv_get_fences(struct dma_resv *obj, struct dma_fence **fence_excl,
 			if (fence_excl)
 				dma_fence_put(*fence_excl);
 
-			count = cursor.fences ? cursor.fences->shared_count : 0;
+			count = cursor.shared_count;
 			count += fence_excl ? 0 : 1;
 
 			/* Eventually re-allocate the array */
diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index 2a926d0d..0039df2 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -100,11 +100,25 @@
           This has the potential to use a lot of memory and print some very
           large kernel messages. If in doubt, say "N".
 
+config DRM_DEBUG_MODESET_LOCK
+	bool "Enable backtrace history for lock contention"
+	depends on STACKTRACE_SUPPORT
+	depends on DEBUG_KERNEL
+	depends on EXPERT
+	select STACKDEPOT
+	default y if DEBUG_WW_MUTEX_SLOWPATH
+	help
+	  Enable debug tracing of failures to gracefully handle drm modeset lock
+	  contention. A history of each drm modeset lock path hitting -EDEADLK
+	  will be saved until gracefully handled, and the backtrace will be
+	  printed when attempting to lock a contended lock.
+
+	  If in doubt, say "N".
+
 config DRM_FBDEV_EMULATION
 	bool "Enable legacy fbdev support for your modesetting driver"
-	depends on DRM
-	depends on FB=y || FB=DRM
-	select DRM_KMS_HELPER
+	depends on DRM_KMS_HELPER
+	depends on FB=y || FB=DRM_KMS_HELPER
 	select FB_CFB_FILLRECT
 	select FB_CFB_COPYAREA
 	select FB_CFB_IMAGEBLIT
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index 751557a..a15a478 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -297,7 +297,7 @@ void amdgpu_amdkfd_ras_poison_consumption_handler(struct kgd_dev *kgd);
 void amdgpu_amdkfd_gpuvm_init_mem_limits(void);
 void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
 				struct amdgpu_vm *vm);
-void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo);
+void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo);
 void amdgpu_amdkfd_reserve_system_mem(uint64_t size);
 #else
 static inline
@@ -312,7 +312,7 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
 }
 
 static inline
-void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo)
+void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo)
 {
 }
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 0e9cfe9..71acd57 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -207,7 +207,7 @@ static void unreserve_mem_limit(struct amdgpu_device *adev,
 	spin_unlock(&kfd_mem_limit.mem_limit_lock);
 }
 
-void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo)
+void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
 	u32 domain = bo->preferred_domains;
@@ -219,6 +219,8 @@ void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo)
 	}
 
 	unreserve_mem_limit(adev, amdgpu_bo_size(bo), domain, sg);
+
+	kfree(bo->kfd_bo);
 }
 
 
@@ -734,14 +736,19 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
 		}
 
 		/* Add BO to VM internal data structures */
+		ret = amdgpu_bo_reserve(bo[i], false);
+		if (ret) {
+			pr_debug("Unable to reserve BO during memory attach");
+			goto unwind;
+		}
 		attachment[i]->bo_va = amdgpu_vm_bo_add(adev, vm, bo[i]);
+		amdgpu_bo_unreserve(bo[i]);
 		if (unlikely(!attachment[i]->bo_va)) {
 			ret = -ENOMEM;
 			pr_err("Failed to add BO object to VM. ret == %d\n",
 			       ret);
 			goto unwind;
 		}
-
 		attachment[i]->va = va;
 		attachment[i]->pte_flags = get_pte_flags(adev, mem);
 		attachment[i]->adev = adev;
@@ -757,7 +764,9 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
 		if (!attachment[i])
 			continue;
 		if (attachment[i]->bo_va) {
+			amdgpu_bo_reserve(bo[i], true);
 			amdgpu_vm_bo_rmv(adev, attachment[i]->bo_va);
+			amdgpu_bo_unreserve(bo[i]);
 			list_del(&attachment[i]->list);
 		}
 		if (bo[i])
@@ -1568,12 +1577,12 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 	pr_debug("Release VA 0x%llx - 0x%llx\n", mem->va,
 		mem->va + bo_size * (1 + mem->aql_queue));
 
-	ret = unreserve_bo_and_vms(&ctx, false, false);
-
 	/* Remove from VM internal data structures */
 	list_for_each_entry_safe(entry, tmp, &mem->attachments, list)
 		kfd_mem_detach(entry);
 
+	ret = unreserve_bo_and_vms(&ctx, false, false);
+
 	/* Free the sync object */
 	amdgpu_sync_free(&mem->sync);
 
@@ -1600,9 +1609,13 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 	drm_vma_node_revoke(&mem->bo->tbo.base.vma_node, drm_priv);
 	if (mem->dmabuf)
 		dma_buf_put(mem->dmabuf);
-	drm_gem_object_put(&mem->bo->tbo.base);
 	mutex_destroy(&mem->lock);
-	kfree(mem);
+
+	/* If this releases the last reference, it will end up calling
+	 * amdgpu_amdkfd_release_notify and kfree the mem struct. That's why
+	 * this needs to be the last call here.
+	 */
+	drm_gem_object_put(&mem->bo->tbo.base);
 
 	return ret;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 6e40cc1..5625f77 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2398,10 +2398,6 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
 	if (!adev->gmc.xgmi.pending_reset)
 		amdgpu_amdkfd_device_init(adev);
 
-	r = amdgpu_amdkfd_resume_iommu(adev);
-	if (r)
-		goto init_failed;
-
 	amdgpu_fru_get_product_info(adev);
 
 init_failed:
@@ -3171,11 +3167,21 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type)
 {
 	switch (asic_type) {
 #if defined(CONFIG_DRM_AMD_DC)
-#if defined(CONFIG_DRM_AMD_DC_SI)
 	case CHIP_TAHITI:
 	case CHIP_PITCAIRN:
 	case CHIP_VERDE:
 	case CHIP_OLAND:
+		/*
+		 * We have systems in the wild with these ASICs that require
+		 * LVDS and VGA support which is not supported with DC.
+		 *
+		 * Fallback to the non-DC driver here by default so as not to
+		 * cause regressions.
+		 */
+#if defined(CONFIG_DRM_AMD_DC_SI)
+		return amdgpu_dc > 0;
+#else
+		return false;
 #endif
 	case CHIP_BONAIRE:
 	case CHIP_KAVERI:
@@ -4287,8 +4293,6 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
 	if (r)
 		return r;
 
-	amdgpu_amdkfd_pre_reset(adev);
-
 	/* Resume IP prior to SMC */
 	r = amdgpu_device_ip_reinit_early_sriov(adev);
 	if (r)
@@ -4850,6 +4854,9 @@ static void amdgpu_device_recheck_guilty_jobs(
 
 		/* clear job's guilty and depend the folowing step to decide the real one */
 		drm_sched_reset_karma(s_job);
+		/* for the real bad job, it will be resubmitted twice, adding a dma_fence_get
+		 * to make sure fence is balanced */
+		dma_fence_get(s_job->s_fence->parent);
 		drm_sched_resubmit_jobs_ext(&ring->sched, 1);
 
 		ret = dma_fence_wait_timeout(s_job->s_fence->parent, false, ring->sched.timeout);
@@ -4885,6 +4892,7 @@ static void amdgpu_device_recheck_guilty_jobs(
 
 		/* got the hw fence, signal finished fence */
 		atomic_dec(ring->sched.score);
+		dma_fence_put(s_job->s_fence->parent);
 		dma_fence_get(&s_job->s_fence->finished);
 		dma_fence_signal(&s_job->s_fence->finished);
 		dma_fence_put(&s_job->s_fence->finished);
@@ -5020,8 +5028,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 
 		cancel_delayed_work_sync(&tmp_adev->delayed_init_work);
 
-		if (!amdgpu_sriov_vf(tmp_adev))
-			amdgpu_amdkfd_pre_reset(tmp_adev);
+		amdgpu_amdkfd_pre_reset(tmp_adev);
 
 		/*
 		 * Mark these ASICs to be reseted as untracked first
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index d7c8d9e..ff70bc2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -867,7 +867,8 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 		case IP_VERSION(2, 0, 2):
 		case IP_VERSION(2, 2, 0):
 			amdgpu_device_ip_block_add(adev, &vcn_v2_0_ip_block);
-			amdgpu_device_ip_block_add(adev, &jpeg_v2_0_ip_block);
+			if (!amdgpu_sriov_vf(adev))
+				amdgpu_device_ip_block_add(adev, &jpeg_v2_0_ip_block);
 			break;
 		case IP_VERSION(2, 0, 3):
 			break;
@@ -881,6 +882,7 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 			break;
 		case IP_VERSION(3, 0, 0):
 		case IP_VERSION(3, 0, 16):
+		case IP_VERSION(3, 0, 64):
 		case IP_VERSION(3, 1, 1):
 		case IP_VERSION(3, 0, 2):
 			amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index a573424..a1e63ba 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -60,9 +60,10 @@ static vm_fault_t amdgpu_gem_fault(struct vm_fault *vmf)
 			goto unlock;
 		}
 
-		ret = ttm_bo_vm_fault_reserved(vmf, vmf->vma->vm_page_prot,
-					       TTM_BO_VM_NUM_PREFAULT, 1);
-		drm_dev_exit(idx);
+		 ret = ttm_bo_vm_fault_reserved(vmf, vmf->vma->vm_page_prot,
+						TTM_BO_VM_NUM_PREFAULT);
+
+		 drm_dev_exit(idx);
 	} else {
 		ret = ttm_bo_vm_dummy_page(vmf, vmf->vma->vm_page_prot);
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index dfe667e..651c7ab 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -1423,6 +1423,8 @@ static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused)
 	struct drm_amdgpu_info_firmware fw_info;
 	struct drm_amdgpu_query_fw query_fw;
 	struct atom_context *ctx = adev->mode_info.atom_context;
+	uint8_t smu_minor, smu_debug;
+	uint16_t smu_major;
 	int ret, i;
 
 	static const char *ta_fw_name[TA_FW_TYPE_MAX_INDEX] = {
@@ -1568,8 +1570,11 @@ static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused)
 	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
 	if (ret)
 		return ret;
-	seq_printf(m, "SMC feature version: %u, firmware version: 0x%08x\n",
-		   fw_info.feature, fw_info.ver);
+	smu_major = (fw_info.ver >> 16) & 0xffff;
+	smu_minor = (fw_info.ver >> 8) & 0xff;
+	smu_debug = (fw_info.ver >> 0) & 0xff;
+	seq_printf(m, "SMC feature version: %u, firmware version: 0x%08x (%d.%d.%d)\n",
+		   fw_info.feature, fw_info.ver, smu_major, smu_minor, smu_debug);
 
 	/* SDMA */
 	query_fw.fw_type = AMDGPU_INFO_FW_SDMA;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index aeb92e5..4fcfc23 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -1274,7 +1274,7 @@ void amdgpu_bo_release_notify(struct ttm_buffer_object *bo)
 	abo = ttm_to_amdgpu_bo(bo);
 
 	if (abo->kfd_bo)
-		amdgpu_amdkfd_unreserve_memory_limit(abo);
+		amdgpu_amdkfd_release_notify(abo);
 
 	/* We only remove the fence if the resv has individualized. */
 	WARN_ON_ONCE(bo->type == ttm_bo_type_kernel
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index 2658414..4f7c708 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -134,6 +134,7 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
 			adev->vcn.indirect_sram = true;
 		break;
 	case IP_VERSION(3, 0, 0):
+	case IP_VERSION(3, 0, 64):
 		if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0))
 			fw_name = FIRMWARE_SIENNA_CICHLID;
 		else
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
index 978ac92..0fad2bf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
@@ -806,9 +806,9 @@ static void amdgpu_xgmi_reset_ras_error_count(struct amdgpu_device *adev)
 		for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++)
 			pcs_clear_status(adev,
 					 xgmi23_pcs_err_status_reg_aldebaran[i]);
-		for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++)
+		for (i = 0; i < ARRAY_SIZE(xgmi3x16_pcs_err_status_reg_aldebaran); i++)
 			pcs_clear_status(adev,
-					 xgmi23_pcs_err_status_reg_aldebaran[i]);
+					 xgmi3x16_pcs_err_status_reg_aldebaran[i]);
 		for (i = 0; i < ARRAY_SIZE(walf_pcs_err_status_reg_aldebaran); i++)
 			pcs_clear_status(adev,
 					 walf_pcs_err_status_reg_aldebaran[i]);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 90a834d..e7dfeb4 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -8249,6 +8249,9 @@ static int gfx_v10_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 {
 	u32 reg, data;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+
 	/* not for *_SOC15 */
 	reg = SOC15_REG_OFFSET(GC, 0, mmRLC_SPM_MC_CNTL);
 	if (amdgpu_sriov_is_pp_one_vf(adev))
@@ -8263,6 +8266,8 @@ static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 		WREG32_SOC15_NO_KIQ(GC, 0, mmRLC_SPM_MC_CNTL, data);
 	else
 		WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data);
+
+	amdgpu_gfx_off_ctrl(adev, true);
 }
 
 static bool gfx_v10_0_check_rlcg_range(struct amdgpu_device *adev,
@@ -8316,11 +8321,8 @@ static void gfx_v10_cntl_power_gating(struct amdgpu_device *adev, bool enable)
 	if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG)) {
 		switch (adev->ip_versions[GC_HWIP][0]) {
 		case IP_VERSION(10, 3, 1):
-			data = 0x4E20 & RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG_MASK_Vangogh;
-			WREG32_SOC15(GC, 0, mmRLC_PG_DELAY_3, data);
-			break;
 		case IP_VERSION(10, 3, 3):
-			data = 0x1388 & RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG_MASK_Vangogh;
+			data = 0x4E20 & RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG_MASK_Vangogh;
 			WREG32_SOC15(GC, 0, mmRLC_PG_DELAY_3, data);
 			break;
 		default:
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
index 37b4a3d..d17a6f3 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
@@ -3575,12 +3575,16 @@ static void gfx_v7_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 {
 	u32 data;
 
+	amdgpu_gfx_off_ctrl(adev, false);
+
 	data = RREG32(mmRLC_SPM_VMID);
 
 	data &= ~RLC_SPM_VMID__RLC_SPM_VMID_MASK;
 	data |= (vmid & RLC_SPM_VMID__RLC_SPM_VMID_MASK) << RLC_SPM_VMID__RLC_SPM_VMID__SHIFT;
 
 	WREG32(mmRLC_SPM_VMID, data);
+
+	amdgpu_gfx_off_ctrl(adev, true);
 }
 
 static void gfx_v7_0_enable_cgcg(struct amdgpu_device *adev, bool enable)
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index e0302c2..5f112ef 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -5624,6 +5624,8 @@ static void gfx_v8_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 {
 	u32 data;
 
+	amdgpu_gfx_off_ctrl(adev, false);
+
 	if (amdgpu_sriov_is_pp_one_vf(adev))
 		data = RREG32_NO_KIQ(mmRLC_SPM_VMID);
 	else
@@ -5636,6 +5638,8 @@ static void gfx_v8_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 		WREG32_NO_KIQ(mmRLC_SPM_VMID, data);
 	else
 		WREG32(mmRLC_SPM_VMID, data);
+
+	amdgpu_gfx_off_ctrl(adev, true);
 }
 
 static const struct amdgpu_rlc_funcs iceland_rlc_funcs = {
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 7f944bb..b4b80f2 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -2462,7 +2462,9 @@ static int gfx_v9_0_sw_fini(void *handle)
 	amdgpu_gfx_kiq_fini(adev);
 
 	gfx_v9_0_mec_fini(adev);
-	amdgpu_bo_unref(&adev->gfx.rlc.clear_state_obj);
+	amdgpu_bo_free_kernel(&adev->gfx.rlc.clear_state_obj,
+				&adev->gfx.rlc.clear_state_gpu_addr,
+				(void **)&adev->gfx.rlc.cs_ptr);
 	if (adev->flags & AMD_IS_APU) {
 		amdgpu_bo_free_kernel(&adev->gfx.rlc.cp_table_obj,
 				&adev->gfx.rlc.cp_table_gpu_addr,
@@ -5102,6 +5104,8 @@ static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 {
 	u32 reg, data;
 
+	amdgpu_gfx_off_ctrl(adev, false);
+
 	reg = SOC15_REG_OFFSET(GC, 0, mmRLC_SPM_MC_CNTL);
 	if (amdgpu_sriov_is_pp_one_vf(adev))
 		data = RREG32_NO_KIQ(reg);
@@ -5115,6 +5119,8 @@ static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 		WREG32_SOC15_NO_KIQ(GC, 0, mmRLC_SPM_MC_CNTL, data);
 	else
 		WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data);
+
+	amdgpu_gfx_off_ctrl(adev, true);
 }
 
 static bool gfx_v9_0_check_rlcg_range(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
index bda1542..480e418 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
@@ -348,6 +348,10 @@ static void gfxhub_v1_0_gart_disable(struct amdgpu_device *adev)
 		WREG32_SOC15_OFFSET(GC, 0, mmVM_CONTEXT0_CNTL,
 				    i * hub->ctx_distance, 0);
 
+	if (amdgpu_sriov_vf(adev))
+		/* Avoid write to GMC registers */
+		return;
+
 	/* Setup TLB control */
 	tmp = RREG32_SOC15(GC, 0, mmMC_VM_MX_L1_TLB_CNTL);
 	tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_1.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_1.c
index 497b86c..90f0aef 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_1.c
@@ -54,15 +54,17 @@ int gfxhub_v1_1_get_xgmi_info(struct amdgpu_device *adev)
 		seg_size = REG_GET_FIELD(
 			RREG32_SOC15(GC, 0, mmMC_VM_XGMI_LFB_SIZE_ALDE),
 			MC_VM_XGMI_LFB_SIZE, PF_LFB_SIZE) << 24;
+		max_region =
+			REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL_ALDE, PF_MAX_REGION);
 	} else {
 		xgmi_lfb_cntl = RREG32_SOC15(GC, 0, mmMC_VM_XGMI_LFB_CNTL);
 		seg_size = REG_GET_FIELD(
 			RREG32_SOC15(GC, 0, mmMC_VM_XGMI_LFB_SIZE),
 			MC_VM_XGMI_LFB_SIZE, PF_LFB_SIZE) << 24;
+		max_region =
+			REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL, PF_MAX_REGION);
 	}
 
-	max_region =
-		REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL, PF_MAX_REGION);
 
 
 	switch (adev->asic_type) {
@@ -89,9 +91,15 @@ int gfxhub_v1_1_get_xgmi_info(struct amdgpu_device *adev)
 		if (adev->gmc.xgmi.num_physical_nodes > max_num_physical_nodes)
 			return -EINVAL;
 
-		adev->gmc.xgmi.physical_node_id =
-		REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL,
-			      PF_LFB_REGION);
+		if (adev->asic_type == CHIP_ALDEBARAN) {
+			adev->gmc.xgmi.physical_node_id =
+				REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL_ALDE,
+						PF_LFB_REGION);
+		} else {
+			adev->gmc.xgmi.physical_node_id =
+				REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL,
+						PF_LFB_REGION);
+		}
 
 		if (adev->gmc.xgmi.physical_node_id > max_physical_node_id)
 			return -EINVAL;
diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index febc903..59eafa3 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -182,6 +182,7 @@ static int nv_query_video_codecs(struct amdgpu_device *adev, bool encode,
 {
 	switch (adev->ip_versions[UVD_HWIP][0]) {
 	case IP_VERSION(3, 0, 0):
+	case IP_VERSION(3, 0, 64):
 		if (amdgpu_sriov_vf(adev)) {
 			if (encode)
 				*codecs = &sriov_sc_video_codecs_encode;
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
index d5d023a..2d558c2 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
@@ -534,6 +534,19 @@ static int uvd_v6_0_hw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	cancel_delayed_work_sync(&adev->uvd.idle_work);
+
+	if (RREG32(mmUVD_STATUS) != 0)
+		uvd_v6_0_stop(adev);
+
+	return 0;
+}
+
+static int uvd_v6_0_suspend(void *handle)
+{
+	int r;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
 	/*
 	 * Proper cleanups before halting the HW engine:
 	 *   - cancel the delayed idle work
@@ -558,17 +571,6 @@ static int uvd_v6_0_hw_fini(void *handle)
 						       AMD_CG_STATE_GATE);
 	}
 
-	if (RREG32(mmUVD_STATUS) != 0)
-		uvd_v6_0_stop(adev);
-
-	return 0;
-}
-
-static int uvd_v6_0_suspend(void *handle)
-{
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
 	r = uvd_v6_0_hw_fini(adev);
 	if (r)
 		return r;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index 0fffaf8..3b119db1 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -406,7 +406,7 @@ static const struct kfd_device_info aldebaran_device_info = {
 static const struct kfd_device_info renoir_device_info = {
 	.asic_family = CHIP_RENOIR,
 	.asic_name = "renoir",
-	.gfx_target_version = 90002,
+	.gfx_target_version = 90012,
 	.max_pasid_bits = 16,
 	.max_no_of_hqd  = 24,
 	.doorbell_size  = 8,
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
index 533b27b..003ba6a 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -1430,7 +1430,7 @@ static int unmap_queues_cpsch(struct device_queue_manager *dqm,
 
 	if (!dqm->sched_running)
 		return 0;
-	if (dqm->is_hws_hang)
+	if (dqm->is_hws_hang || dqm->is_resetting)
 		return -EIO;
 	if (!dqm->active_runlist)
 		return retval;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
index 2e86692d..d138889 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
@@ -308,7 +308,7 @@
  * 16MB are reserved for kernel use (CWSR trap handler and kernel IB
  * for now).
  */
-#define SVM_USER_BASE 0x1000000ull
+#define SVM_USER_BASE (u64)(KFD_CWSR_TBA_TMA_SIZE + 2*PAGE_SIZE)
 #define SVM_CWSR_BASE (SVM_USER_BASE - KFD_CWSR_TBA_TMA_SIZE)
 #define SVM_IB_BASE   (SVM_CWSR_BASE - PAGE_SIZE)
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
index d43bfd8..9b9c2b9 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
@@ -281,6 +281,19 @@ static unsigned long svm_migrate_successful_pages(struct migrate_vma *migrate)
 	return cpages;
 }
 
+static unsigned long svm_migrate_unsuccessful_pages(struct migrate_vma *migrate)
+{
+	unsigned long upages = 0;
+	unsigned long i;
+
+	for (i = 0; i < migrate->npages; i++) {
+		if (migrate->src[i] & MIGRATE_PFN_VALID &&
+		    !(migrate->src[i] & MIGRATE_PFN_MIGRATE))
+			upages++;
+	}
+	return upages;
+}
+
 static int
 svm_migrate_copy_to_vram(struct amdgpu_device *adev, struct svm_range *prange,
 			 struct migrate_vma *migrate, struct dma_fence **mfence,
@@ -632,10 +645,11 @@ svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange,
 		       struct vm_area_struct *vma, uint64_t start, uint64_t end)
 {
 	uint64_t npages = (end - start) >> PAGE_SHIFT;
+	unsigned long upages = npages;
+	unsigned long cpages = 0;
 	struct kfd_process_device *pdd;
 	struct dma_fence *mfence = NULL;
 	struct migrate_vma migrate;
-	unsigned long cpages = 0;
 	dma_addr_t *scratch;
 	size_t size;
 	void *buf;
@@ -669,6 +683,7 @@ svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange,
 	if (!cpages) {
 		pr_debug("failed collect migrate device pages [0x%lx 0x%lx]\n",
 			 prange->start, prange->last);
+		upages = svm_migrate_unsuccessful_pages(&migrate);
 		goto out_free;
 	}
 	if (cpages != npages)
@@ -681,8 +696,9 @@ svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange,
 				    scratch, npages);
 	migrate_vma_pages(&migrate);
 
-	pr_debug("successful/cpages/npages 0x%lx/0x%lx/0x%lx\n",
-		svm_migrate_successful_pages(&migrate), cpages, migrate.npages);
+	upages = svm_migrate_unsuccessful_pages(&migrate);
+	pr_debug("unsuccessful/cpages/npages 0x%lx/0x%lx/0x%lx\n",
+		 upages, cpages, migrate.npages);
 
 	svm_migrate_copy_done(adev, mfence);
 	migrate_vma_finalize(&migrate);
@@ -696,9 +712,9 @@ svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange,
 		if (pdd)
 			WRITE_ONCE(pdd->page_out, pdd->page_out + cpages);
 
-		return cpages;
+		return upages;
 	}
-	return r;
+	return r ? r : upages;
 }
 
 /**
@@ -718,7 +734,7 @@ int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm)
 	unsigned long addr;
 	unsigned long start;
 	unsigned long end;
-	unsigned long cpages = 0;
+	unsigned long upages = 0;
 	long r = 0;
 
 	if (!prange->actual_loc) {
@@ -754,12 +770,12 @@ int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm)
 			pr_debug("failed %ld to migrate\n", r);
 			break;
 		} else {
-			cpages += r;
+			upages += r;
 		}
 		addr = next;
 	}
 
-	if (cpages) {
+	if (!upages) {
 		svm_range_vram_node_free(prange);
 		prange->actual_loc = 0;
 	}
@@ -782,7 +798,7 @@ static int
 svm_migrate_vram_to_vram(struct svm_range *prange, uint32_t best_loc,
 			 struct mm_struct *mm)
 {
-	int r;
+	int r, retries = 3;
 
 	/*
 	 * TODO: for both devices with PCIe large bar or on same xgmi hive, skip
@@ -791,9 +807,14 @@ svm_migrate_vram_to_vram(struct svm_range *prange, uint32_t best_loc,
 
 	pr_debug("from gpu 0x%x to gpu 0x%x\n", prange->actual_loc, best_loc);
 
-	r = svm_migrate_vram_to_ram(prange, mm);
-	if (r)
-		return r;
+	do {
+		r = svm_migrate_vram_to_ram(prange, mm);
+		if (r)
+			return r;
+	} while (prange->actual_loc && --retries);
+
+	if (prange->actual_loc)
+		return -EDEADLK;
 
 	return svm_migrate_ram_to_vram(prange, best_loc, mm);
 }
@@ -838,6 +859,11 @@ static vm_fault_t svm_migrate_to_ram(struct vm_fault *vmf)
 		pr_debug("failed find process at fault address 0x%lx\n", addr);
 		return VM_FAULT_SIGBUS;
 	}
+	if (READ_ONCE(p->svms.faulting_task) == current) {
+		pr_debug("skipping ram migration\n");
+		kfd_unref_process(p);
+		return 0;
+	}
 	addr >>= PAGE_SHIFT;
 	pr_debug("CPU page fault svms 0x%p address 0x%lx\n", &p->svms, addr);
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
index 4104b16..94e92c0 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
@@ -766,8 +766,10 @@ struct svm_range_list {
 	struct list_head		deferred_range_list;
 	spinlock_t			deferred_list_lock;
 	atomic_t			evicted_ranges;
+	bool				drain_pagefaults;
 	struct delayed_work		restore_work;
 	DECLARE_BITMAP(bitmap_supported, MAX_GPU_INSTANCE);
+	struct task_struct 		*faulting_task;
 };
 
 /* Process data */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index 4578638..b993011 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -1715,7 +1715,11 @@ int kfd_process_evict_queues(struct kfd_process *p)
 
 		r = pdd->dev->dqm->ops.evict_process_queues(pdd->dev->dqm,
 							    &pdd->qpd);
-		if (r) {
+		/* evict return -EIO if HWS is hang or asic is resetting, in this case
+		 * we would like to set all the queues to be in evicted state to prevent
+		 * them been add back since they actually not be saved right now.
+		 */
+		if (r && r != -EIO) {
 			pr_err("Failed to evict process queues\n");
 			goto fail;
 		}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index b691c84..16137c4 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -1496,9 +1496,11 @@ static int svm_range_validate_and_map(struct mm_struct *mm,
 
 		next = min(vma->vm_end, end);
 		npages = (next - addr) >> PAGE_SHIFT;
+		WRITE_ONCE(p->svms.faulting_task, current);
 		r = amdgpu_hmm_range_get_pages(&prange->notifier, mm, NULL,
 					       addr, npages, &hmm_range,
 					       readonly, true, owner);
+		WRITE_ONCE(p->svms.faulting_task, NULL);
 		if (r) {
 			pr_debug("failed %d to get svm range pages\n", r);
 			goto unreserve_out;
@@ -2000,20 +2002,28 @@ static void svm_range_deferred_list_work(struct work_struct *work)
 		pr_debug("prange 0x%p [0x%lx 0x%lx] op %d\n", prange,
 			 prange->start, prange->last, prange->work_item.op);
 
-		/* Make sure no stale retry fault coming after range is freed */
-		if (prange->work_item.op == SVM_OP_UNMAP_RANGE)
-			svm_range_drain_retry_fault(prange->svms);
-
 		mm = prange->work_item.mm;
+retry:
 		mmap_write_lock(mm);
 		mutex_lock(&svms->lock);
 
-		/* Remove from deferred_list must be inside mmap write lock,
+		/* Checking for the need to drain retry faults must be in
+		 * mmap write lock to serialize with munmap notifiers.
+		 *
+		 * Remove from deferred_list must be inside mmap write lock,
 		 * otherwise, svm_range_list_lock_and_flush_work may hold mmap
 		 * write lock, and continue because deferred_list is empty, then
 		 * deferred_list handle is blocked by mmap write lock.
 		 */
 		spin_lock(&svms->deferred_list_lock);
+		if (unlikely(svms->drain_pagefaults)) {
+			svms->drain_pagefaults = false;
+			spin_unlock(&svms->deferred_list_lock);
+			mutex_unlock(&svms->lock);
+			mmap_write_unlock(mm);
+			svm_range_drain_retry_fault(svms);
+			goto retry;
+		}
 		list_del_init(&prange->deferred_list);
 		spin_unlock(&svms->deferred_list_lock);
 
@@ -2046,6 +2056,12 @@ svm_range_add_list_work(struct svm_range_list *svms, struct svm_range *prange,
 			struct mm_struct *mm, enum svm_work_list_ops op)
 {
 	spin_lock(&svms->deferred_list_lock);
+	/* Make sure pending page faults are drained in the deferred worker
+	 * before the range is freed to avoid straggler interrupts on
+	 * unmapped memory causing "phantom faults".
+	 */
+	if (op == SVM_OP_UNMAP_RANGE)
+		svms->drain_pagefaults = true;
 	/* if prange is on the deferred list */
 	if (!list_empty(&prange->deferred_list)) {
 		pr_debug("update exist prange 0x%p work op %d\n", prange, op);
@@ -2261,7 +2277,7 @@ svm_range_from_addr(struct svm_range_list *svms, unsigned long addr,
  * migration if actual loc is not best location, then update GPU page table
  * mapping to the best location.
  *
- * If vm fault gpu is range preferred loc, the best_loc is preferred loc.
+ * If the preferred loc is accessible by faulting GPU, use preferred loc.
  * If vm fault gpu idx is on range ACCESSIBLE bitmap, best_loc is vm fault gpu
  * If vm fault gpu idx is on range ACCESSIBLE_IN_PLACE bitmap, then
  *    if range actual loc is cpu, best_loc is cpu
@@ -2278,7 +2294,7 @@ svm_range_best_restore_location(struct svm_range *prange,
 				struct amdgpu_device *adev,
 				int32_t *gpuidx)
 {
-	struct amdgpu_device *bo_adev;
+	struct amdgpu_device *bo_adev, *preferred_adev;
 	struct kfd_process *p;
 	uint32_t gpuid;
 	int r;
@@ -2291,8 +2307,16 @@ svm_range_best_restore_location(struct svm_range *prange,
 		return -1;
 	}
 
-	if (prange->preferred_loc == gpuid)
+	if (prange->preferred_loc == gpuid ||
+	    prange->preferred_loc == KFD_IOCTL_SVM_LOCATION_SYSMEM) {
 		return prange->preferred_loc;
+	} else if (prange->preferred_loc != KFD_IOCTL_SVM_LOCATION_UNDEFINED) {
+		preferred_adev = svm_range_get_adev_by_id(prange,
+							prange->preferred_loc);
+		if (amdgpu_xgmi_same_hive(adev, preferred_adev))
+			return prange->preferred_loc;
+		/* fall through */
+	}
 
 	if (test_bit(*gpuidx, prange->bitmap_access))
 		return gpuid;
@@ -2313,7 +2337,8 @@ svm_range_best_restore_location(struct svm_range *prange,
 
 static int
 svm_range_get_range_boundaries(struct kfd_process *p, int64_t addr,
-				unsigned long *start, unsigned long *last)
+			       unsigned long *start, unsigned long *last,
+			       bool *is_heap_stack)
 {
 	struct vm_area_struct *vma;
 	struct interval_tree_node *node;
@@ -2324,6 +2349,12 @@ svm_range_get_range_boundaries(struct kfd_process *p, int64_t addr,
 		pr_debug("VMA does not exist in address [0x%llx]\n", addr);
 		return -EFAULT;
 	}
+
+	*is_heap_stack = (vma->vm_start <= vma->vm_mm->brk &&
+			  vma->vm_end >= vma->vm_mm->start_brk) ||
+			 (vma->vm_start <= vma->vm_mm->start_stack &&
+			  vma->vm_end >= vma->vm_mm->start_stack);
+
 	start_limit = max(vma->vm_start >> PAGE_SHIFT,
 		      (unsigned long)ALIGN_DOWN(addr, 2UL << 8));
 	end_limit = min(vma->vm_end >> PAGE_SHIFT,
@@ -2353,9 +2384,9 @@ svm_range_get_range_boundaries(struct kfd_process *p, int64_t addr,
 	*start = start_limit;
 	*last = end_limit - 1;
 
-	pr_debug("vma start: 0x%lx start: 0x%lx vma end: 0x%lx last: 0x%lx\n",
-		  vma->vm_start >> PAGE_SHIFT, *start,
-		  vma->vm_end >> PAGE_SHIFT, *last);
+	pr_debug("vma [0x%lx 0x%lx] range [0x%lx 0x%lx] is_heap_stack %d\n",
+		 vma->vm_start >> PAGE_SHIFT, vma->vm_end >> PAGE_SHIFT,
+		 *start, *last, *is_heap_stack);
 
 	return 0;
 }
@@ -2420,11 +2451,13 @@ svm_range *svm_range_create_unregistered_range(struct amdgpu_device *adev,
 	struct svm_range *prange = NULL;
 	unsigned long start, last;
 	uint32_t gpuid, gpuidx;
+	bool is_heap_stack;
 	uint64_t bo_s = 0;
 	uint64_t bo_l = 0;
 	int r;
 
-	if (svm_range_get_range_boundaries(p, addr, &start, &last))
+	if (svm_range_get_range_boundaries(p, addr, &start, &last,
+					   &is_heap_stack))
 		return NULL;
 
 	r = svm_range_check_vm(p, start, last, &bo_s, &bo_l);
@@ -2451,6 +2484,9 @@ svm_range *svm_range_create_unregistered_range(struct amdgpu_device *adev,
 		return NULL;
 	}
 
+	if (is_heap_stack)
+		prange->preferred_loc = KFD_IOCTL_SVM_LOCATION_SYSMEM;
+
 	svm_range_add_to_svms(prange);
 	svm_range_add_notifier_locked(mm, prange);
 
@@ -3076,6 +3112,8 @@ static void svm_range_evict_svm_bo_worker(struct work_struct *work)
 		struct svm_range *prange =
 				list_first_entry(&svm_bo->range_list,
 						struct svm_range, svm_bo_list);
+		int retries = 3;
+
 		list_del_init(&prange->svm_bo_list);
 		spin_unlock(&svm_bo->list_lock);
 
@@ -3083,7 +3121,11 @@ static void svm_range_evict_svm_bo_worker(struct work_struct *work)
 			 prange->start, prange->last);
 
 		mutex_lock(&prange->migrate_mutex);
-		svm_migrate_vram_to_ram(prange, svm_bo->eviction_fence->mm);
+		do {
+			svm_migrate_vram_to_ram(prange,
+						svm_bo->eviction_fence->mm);
+		} while (prange->actual_loc && --retries);
+		WARN(prange->actual_loc, "Migration failed during eviction");
 
 		mutex_lock(&prange->lock);
 		prange->svm_bo = NULL;
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 43e983e..c911b30 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -217,6 +217,7 @@ static const struct drm_format_info *
 amd_get_format_info(const struct drm_mode_fb_cmd2 *cmd);
 
 static void handle_hpd_irq_helper(struct amdgpu_dm_connector *aconnector);
+static void handle_hpd_rx_irq(void *param);
 
 static bool
 is_timing_unchanged_for_freesync(struct drm_crtc_state *old_crtc_state,
@@ -619,7 +620,7 @@ static void dm_dcn_vertical_interrupt0_high_irq(void *interrupt_params)
 
 	amdgpu_dm_crtc_handle_crc_window_irq(&acrtc->base);
 }
-#endif
+#endif /* CONFIG_DRM_AMD_SECURE_DISPLAY */
 
 /**
  * dmub_aux_setconfig_reply_callback - Callback for AUX or SET_CONFIG command.
@@ -669,10 +670,7 @@ void dmub_hpd_callback(struct amdgpu_device *adev, struct dmub_notification *not
 		return;
 	}
 
-	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
-
 	link_index = notify->link_index;
-
 	link = adev->dm.dc->links[link_index];
 
 	drm_connector_list_iter_begin(dev, &iter);
@@ -685,10 +683,13 @@ void dmub_hpd_callback(struct amdgpu_device *adev, struct dmub_notification *not
 		}
 	}
 	drm_connector_list_iter_end(&iter);
-	drm_modeset_unlock(&dev->mode_config.connection_mutex);
 
-	if (hpd_aconnector)
-		handle_hpd_irq_helper(hpd_aconnector);
+	if (hpd_aconnector) {
+		if (notify->type == DMUB_NOTIFICATION_HPD)
+			handle_hpd_irq_helper(hpd_aconnector);
+		else if (notify->type == DMUB_NOTIFICATION_HPD_IRQ)
+			handle_hpd_rx_irq(hpd_aconnector);
+	}
 }
 
 /**
@@ -764,6 +765,10 @@ static void dm_dmub_outbox1_low_irq(void *interrupt_params)
 				DRM_ERROR("DM: notify type %d invalid!", notify.type);
 				continue;
 			}
+			if (!dm->dmub_callback[notify.type]) {
+				DRM_DEBUG_DRIVER("DMUB notification skipped, no handler: type=%d\n", notify.type);
+				continue;
+			}
 			if (dm->dmub_thread_offload[notify.type] == true) {
 				dmub_hpd_wrk = kzalloc(sizeof(*dmub_hpd_wrk), GFP_ATOMIC);
 				if (!dmub_hpd_wrk) {
@@ -813,7 +818,7 @@ static void dm_dmub_outbox1_low_irq(void *interrupt_params)
 	if (count > DMUB_TRACE_MAX_READ)
 		DRM_DEBUG_DRIVER("Warning : count > DMUB_TRACE_MAX_READ");
 }
-#endif
+#endif /* CONFIG_DRM_AMD_DC_DCN */
 
 static int dm_set_clockgating_state(void *handle,
 		  enum amd_clockgating_state state)
@@ -1410,7 +1415,15 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 		switch (adev->ip_versions[DCE_HWIP][0]) {
 		case IP_VERSION(2, 1, 0):
 			init_data.flags.gpu_vm_support = true;
-			init_data.flags.disable_dmcu = true;
+			switch (adev->dm.dmcub_fw_version) {
+			case 0: /* development */
+			case 0x1: /* linux-firmware.git hash 6d9f399 */
+			case 0x01000000: /* linux-firmware.git hash 9a0b0f4 */
+				init_data.flags.disable_dmcu = false;
+				break;
+			default:
+				init_data.flags.disable_dmcu = true;
+			}
 			break;
 		case IP_VERSION(1, 0, 0):
 		case IP_VERSION(1, 0, 1):
@@ -1556,7 +1569,11 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 			DRM_ERROR("amdgpu: fail to register dmub hpd callback");
 			goto error;
 		}
-#endif
+		if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_HPD_IRQ, dmub_hpd_callback, true)) {
+			DRM_ERROR("amdgpu: fail to register dmub hpd callback");
+			goto error;
+		}
+#endif /* CONFIG_DRM_AMD_DC_DCN */
 	}
 
 	if (amdgpu_dm_initialize_drm_device(adev)) {
@@ -4565,7 +4582,8 @@ static void get_min_max_dc_plane_scaling(struct drm_device *dev,
 }
 
 
-static int fill_dc_scaling_info(const struct drm_plane_state *state,
+static int fill_dc_scaling_info(struct amdgpu_device *adev,
+				const struct drm_plane_state *state,
 				struct dc_scaling_info *scaling_info)
 {
 	int scale_w, scale_h, min_downscale, max_upscale;
@@ -4579,7 +4597,8 @@ static int fill_dc_scaling_info(const struct drm_plane_state *state,
 	/*
 	 * For reasons we don't (yet) fully understand a non-zero
 	 * src_y coordinate into an NV12 buffer can cause a
-	 * system hang. To avoid hangs (and maybe be overly cautious)
+	 * system hang on DCN1x.
+	 * To avoid hangs (and maybe be overly cautious)
 	 * let's reject both non-zero src_x and src_y.
 	 *
 	 * We currently know of only one use-case to reproduce a
@@ -4587,10 +4606,10 @@ static int fill_dc_scaling_info(const struct drm_plane_state *state,
 	 * is to gesture the YouTube Android app into full screen
 	 * on ChromeOS.
 	 */
-	if (state->fb &&
-	    state->fb->format->format == DRM_FORMAT_NV12 &&
-	    (scaling_info->src_rect.x != 0 ||
-	     scaling_info->src_rect.y != 0))
+	if (((adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 0)) ||
+	    (adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 1))) &&
+	    (state->fb && state->fb->format->format == DRM_FORMAT_NV12 &&
+	    (scaling_info->src_rect.x != 0 || scaling_info->src_rect.y != 0)))
 		return -EINVAL;
 
 	scaling_info->src_rect.width = state->src_w >> 16;
@@ -5496,7 +5515,7 @@ static int fill_dc_plane_attributes(struct amdgpu_device *adev,
 	int ret;
 	bool force_disable_dcc = false;
 
-	ret = fill_dc_scaling_info(plane_state, &scaling_info);
+	ret = fill_dc_scaling_info(adev, plane_state, &scaling_info);
 	if (ret)
 		return ret;
 
@@ -6070,7 +6089,7 @@ static void apply_dsc_policy_for_stream(struct amdgpu_dm_connector *aconnector,
 	if (stream->timing.flags.DSC && aconnector->dsc_settings.dsc_bits_per_pixel)
 		stream->timing.dsc_cfg.bits_per_pixel = aconnector->dsc_settings.dsc_bits_per_pixel;
 }
-#endif
+#endif /* CONFIG_DRM_AMD_DC_DCN */
 
 /**
  * DOC: FreeSync Video
@@ -7241,8 +7260,8 @@ static int dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
 	struct drm_connector_state *new_con_state;
 	struct amdgpu_dm_connector *aconnector;
 	struct dm_connector_state *dm_conn_state;
-	int i, j, clock;
-	int vcpi, pbn_div, pbn = 0;
+	int i, j;
+	int vcpi, pbn_div, pbn, slot_num = 0;
 
 	for_each_new_connector_in_state(state, connector, new_con_state, i) {
 
@@ -7270,17 +7289,7 @@ static int dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
 		if (!stream)
 			continue;
 
-		if (stream->timing.flags.DSC != 1) {
-			drm_dp_mst_atomic_enable_dsc(state,
-						     aconnector->port,
-						     dm_conn_state->pbn,
-						     0,
-						     false);
-			continue;
-		}
-
 		pbn_div = dm_mst_get_pbn_divider(stream->link);
-		clock = stream->timing.pix_clk_100hz / 10;
 		/* pbn is calculated by compute_mst_dsc_configs_for_state*/
 		for (j = 0; j < dc_state->stream_count; j++) {
 			if (vars[j].aconnector == aconnector) {
@@ -7289,6 +7298,23 @@ static int dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
 			}
 		}
 
+		if (j == dc_state->stream_count)
+			continue;
+
+		slot_num = DIV_ROUND_UP(pbn, pbn_div);
+
+		if (stream->timing.flags.DSC != 1) {
+			dm_conn_state->pbn = pbn;
+			dm_conn_state->vcpi_slots = slot_num;
+
+			drm_dp_mst_atomic_enable_dsc(state,
+						     aconnector->port,
+						     dm_conn_state->pbn,
+						     0,
+						     false);
+			continue;
+		}
+
 		vcpi = drm_dp_mst_atomic_enable_dsc(state,
 						    aconnector->port,
 						    pbn, pbn_div,
@@ -7552,7 +7578,7 @@ static int dm_plane_atomic_check(struct drm_plane *plane,
 	if (ret)
 		return ret;
 
-	ret = fill_dc_scaling_info(new_plane_state, &scaling_info);
+	ret = fill_dc_scaling_info(adev, new_plane_state, &scaling_info);
 	if (ret)
 		return ret;
 
@@ -9000,7 +9026,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 			bundle->surface_updates[planes_count].gamut_remap_matrix = &dc_plane->gamut_remap_matrix;
 		}
 
-		fill_dc_scaling_info(new_plane_state,
+		fill_dc_scaling_info(dm->adev, new_plane_state,
 				     &bundle->scaling_infos[planes_count]);
 
 		bundle->surface_updates[planes_count].scaling_info =
@@ -10787,7 +10813,7 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 
 		ret = drm_atomic_add_affected_connectors(state, crtc);
 		if (ret)
-			return ret;
+			goto fail;
 
 		ret = drm_atomic_add_affected_planes(state, crtc);
 		if (ret)
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
index 3655663..9d43ecb 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
@@ -78,12 +78,10 @@ static int parse_write_buffer_into_params(char *wr_buf, uint32_t wr_buf_size,
 
 	wr_buf_ptr = wr_buf;
 
-	r = copy_from_user(wr_buf_ptr, buf, wr_buf_size);
-
-		/* r is bytes not be copied */
-	if (r >= wr_buf_size) {
-		DRM_DEBUG_DRIVER("user data not be read\n");
-		return -EINVAL;
+	/* r is bytes not be copied */
+	if (copy_from_user(wr_buf_ptr, buf, wr_buf_size)) {
+		DRM_DEBUG_DRIVER("user data could not be read successfully\n");
+		return -EFAULT;
 	}
 
 	/* check number of parameters. isspace could not differ space and \n */
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index 874a49b..32a5ce0 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -534,13 +534,14 @@ static int kbps_to_peak_pbn(int kbps)
 
 static void set_dsc_configs_from_fairness_vars(struct dsc_mst_fairness_params *params,
 		struct dsc_mst_fairness_vars *vars,
-		int count)
+		int count,
+		int k)
 {
 	int i;
 
 	for (i = 0; i < count; i++) {
 		memset(&params[i].timing->dsc_cfg, 0, sizeof(params[i].timing->dsc_cfg));
-		if (vars[i].dsc_enabled && dc_dsc_compute_config(
+		if (vars[i + k].dsc_enabled && dc_dsc_compute_config(
 					params[i].sink->ctx->dc->res_pool->dscs[0],
 					&params[i].sink->dsc_caps.dsc_dec_caps,
 					params[i].sink->ctx->dc->debug.dsc_min_slice_height_override,
@@ -553,7 +554,7 @@ static void set_dsc_configs_from_fairness_vars(struct dsc_mst_fairness_params *p
 			if (params[i].bpp_overwrite)
 				params[i].timing->dsc_cfg.bits_per_pixel = params[i].bpp_overwrite;
 			else
-				params[i].timing->dsc_cfg.bits_per_pixel = vars[i].bpp_x16;
+				params[i].timing->dsc_cfg.bits_per_pixel = vars[i + k].bpp_x16;
 
 			if (params[i].num_slices_h)
 				params[i].timing->dsc_cfg.num_slices_h = params[i].num_slices_h;
@@ -586,7 +587,8 @@ static void increase_dsc_bpp(struct drm_atomic_state *state,
 			     struct dc_link *dc_link,
 			     struct dsc_mst_fairness_params *params,
 			     struct dsc_mst_fairness_vars *vars,
-			     int count)
+			     int count,
+			     int k)
 {
 	int i;
 	bool bpp_increased[MAX_PIPES];
@@ -601,8 +603,9 @@ static void increase_dsc_bpp(struct drm_atomic_state *state,
 	pbn_per_timeslot = dm_mst_get_pbn_divider(dc_link);
 
 	for (i = 0; i < count; i++) {
-		if (vars[i].dsc_enabled) {
-			initial_slack[i] = kbps_to_peak_pbn(params[i].bw_range.max_kbps) - vars[i].pbn;
+		if (vars[i + k].dsc_enabled) {
+			initial_slack[i] =
+			kbps_to_peak_pbn(params[i].bw_range.max_kbps) - vars[i + k].pbn;
 			bpp_increased[i] = false;
 			remaining_to_increase += 1;
 		} else {
@@ -629,7 +632,7 @@ static void increase_dsc_bpp(struct drm_atomic_state *state,
 		link_timeslots_used = 0;
 
 		for (i = 0; i < count; i++)
-			link_timeslots_used += DIV_ROUND_UP(vars[i].pbn, pbn_per_timeslot);
+			link_timeslots_used += DIV_ROUND_UP(vars[i + k].pbn, pbn_per_timeslot);
 
 		fair_pbn_alloc = (63 - link_timeslots_used) / remaining_to_increase * pbn_per_timeslot;
 
@@ -682,7 +685,8 @@ static void try_disable_dsc(struct drm_atomic_state *state,
 			    struct dc_link *dc_link,
 			    struct dsc_mst_fairness_params *params,
 			    struct dsc_mst_fairness_vars *vars,
-			    int count)
+			    int count,
+			    int k)
 {
 	int i;
 	bool tried[MAX_PIPES];
@@ -692,8 +696,8 @@ static void try_disable_dsc(struct drm_atomic_state *state,
 	int remaining_to_try = 0;
 
 	for (i = 0; i < count; i++) {
-		if (vars[i].dsc_enabled
-				&& vars[i].bpp_x16 == params[i].bw_range.max_target_bpp_x16
+		if (vars[i + k].dsc_enabled
+				&& vars[i + k].bpp_x16 == params[i].bw_range.max_target_bpp_x16
 				&& params[i].clock_force_enable == DSC_CLK_FORCE_DEFAULT) {
 			kbps_increase[i] = params[i].bw_range.stream_kbps - params[i].bw_range.max_kbps;
 			tried[i] = false;
@@ -748,9 +752,10 @@ static void try_disable_dsc(struct drm_atomic_state *state,
 static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 					     struct dc_state *dc_state,
 					     struct dc_link *dc_link,
-					     struct dsc_mst_fairness_vars *vars)
+					     struct dsc_mst_fairness_vars *vars,
+					     int *link_vars_start_index)
 {
-	int i;
+	int i, k;
 	struct dc_stream_state *stream;
 	struct dsc_mst_fairness_params params[MAX_PIPES];
 	struct amdgpu_dm_connector *aconnector;
@@ -768,11 +773,17 @@ static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 		if (stream->link != dc_link)
 			continue;
 
+		aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
+		if (!aconnector)
+			continue;
+
+		if (!aconnector->port)
+			continue;
+
 		stream->timing.flags.DSC = 0;
 
 		params[count].timing = &stream->timing;
 		params[count].sink = stream->sink;
-		aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
 		params[count].aconnector = aconnector;
 		params[count].port = aconnector->port;
 		params[count].clock_force_enable = aconnector->dsc_settings.dsc_force_enable;
@@ -794,44 +805,55 @@ static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 
 		count++;
 	}
+
+	if (count == 0) {
+		ASSERT(0);
+		return true;
+	}
+
+	/* k is start index of vars for current phy link used by mst hub */
+	k = *link_vars_start_index;
+	/* set vars start index for next mst hub phy link */
+	*link_vars_start_index += count;
+
 	/* Try no compression */
 	for (i = 0; i < count; i++) {
-		vars[i].aconnector = params[i].aconnector;
-		vars[i].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps);
-		vars[i].dsc_enabled = false;
-		vars[i].bpp_x16 = 0;
+		vars[i + k].aconnector = params[i].aconnector;
+		vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps);
+		vars[i + k].dsc_enabled = false;
+		vars[i + k].bpp_x16 = 0;
 		if (drm_dp_atomic_find_vcpi_slots(state,
 						 params[i].port->mgr,
 						 params[i].port,
-						 vars[i].pbn,
+						 vars[i + k].pbn,
 						 dm_mst_get_pbn_divider(dc_link)) < 0)
 			return false;
 	}
 	if (!drm_dp_mst_atomic_check(state) && !debugfs_overwrite) {
-		set_dsc_configs_from_fairness_vars(params, vars, count);
+		set_dsc_configs_from_fairness_vars(params, vars, count, k);
 		return true;
 	}
 
 	/* Try max compression */
 	for (i = 0; i < count; i++) {
 		if (params[i].compression_possible && params[i].clock_force_enable != DSC_CLK_FORCE_DISABLE) {
-			vars[i].pbn = kbps_to_peak_pbn(params[i].bw_range.min_kbps);
-			vars[i].dsc_enabled = true;
-			vars[i].bpp_x16 = params[i].bw_range.min_target_bpp_x16;
+			vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.min_kbps);
+			vars[i + k].dsc_enabled = true;
+			vars[i + k].bpp_x16 = params[i].bw_range.min_target_bpp_x16;
 			if (drm_dp_atomic_find_vcpi_slots(state,
 							  params[i].port->mgr,
 							  params[i].port,
-							  vars[i].pbn,
+							  vars[i + k].pbn,
 							  dm_mst_get_pbn_divider(dc_link)) < 0)
 				return false;
 		} else {
-			vars[i].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps);
-			vars[i].dsc_enabled = false;
-			vars[i].bpp_x16 = 0;
+			vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps);
+			vars[i + k].dsc_enabled = false;
+			vars[i + k].bpp_x16 = 0;
 			if (drm_dp_atomic_find_vcpi_slots(state,
 							  params[i].port->mgr,
 							  params[i].port,
-							  vars[i].pbn,
+							  vars[i + k].pbn,
 							  dm_mst_get_pbn_divider(dc_link)) < 0)
 				return false;
 		}
@@ -840,15 +862,76 @@ static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 		return false;
 
 	/* Optimize degree of compression */
-	increase_dsc_bpp(state, dc_link, params, vars, count);
+	increase_dsc_bpp(state, dc_link, params, vars, count, k);
 
-	try_disable_dsc(state, dc_link, params, vars, count);
+	try_disable_dsc(state, dc_link, params, vars, count, k);
 
-	set_dsc_configs_from_fairness_vars(params, vars, count);
+	set_dsc_configs_from_fairness_vars(params, vars, count, k);
 
 	return true;
 }
 
+static bool is_dsc_need_re_compute(
+	struct drm_atomic_state *state,
+	struct dc_state *dc_state,
+	struct dc_link *dc_link)
+{
+	int i;
+	bool is_dsc_need_re_compute = false;
+
+	/* only check phy used by mst branch */
+	if (dc_link->type != dc_connection_mst_branch)
+		return false;
+
+	/* check if there is mode change in new request */
+	for (i = 0; i < dc_state->stream_count; i++) {
+		struct amdgpu_dm_connector *aconnector;
+		struct dc_stream_state *stream;
+		struct drm_crtc_state *new_crtc_state;
+		struct drm_connector_state *new_conn_state;
+
+		stream = dc_state->streams[i];
+
+		if (!stream)
+			continue;
+
+		/* check if stream using the same link for mst */
+		if (stream->link != dc_link)
+			continue;
+
+		aconnector = (struct amdgpu_dm_connector *) stream->dm_stream_context;
+		if (!aconnector)
+			continue;
+
+		new_conn_state = drm_atomic_get_new_connector_state(state, &aconnector->base);
+
+		if (!new_conn_state)
+			continue;
+
+		if (IS_ERR(new_conn_state))
+			continue;
+
+		if (!new_conn_state->crtc)
+			continue;
+
+		new_crtc_state = drm_atomic_get_new_crtc_state(state, new_conn_state->crtc);
+
+		if (!new_crtc_state)
+			continue;
+
+		if (IS_ERR(new_crtc_state))
+			continue;
+
+		if (new_crtc_state->enable && new_crtc_state->active) {
+			if (new_crtc_state->mode_changed || new_crtc_state->active_changed ||
+				new_crtc_state->connectors_changed)
+				is_dsc_need_re_compute = true;
+		}
+	}
+
+	return is_dsc_need_re_compute;
+}
+
 bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
 				       struct dc_state *dc_state,
 				       struct dsc_mst_fairness_vars *vars)
@@ -857,6 +940,7 @@ bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
 	struct dc_stream_state *stream;
 	bool computed_streams[MAX_PIPES];
 	struct amdgpu_dm_connector *aconnector;
+	int link_vars_start_index = 0;
 
 	for (i = 0; i < dc_state->stream_count; i++)
 		computed_streams[i] = false;
@@ -881,8 +965,12 @@ bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
 		if (dcn20_remove_stream_from_ctx(stream->ctx->dc, dc_state, stream) != DC_OK)
 			return false;
 
+		if (!is_dsc_need_re_compute(state, dc_state, stream->link))
+			continue;
+
 		mutex_lock(&aconnector->mst_mgr.lock);
-		if (!compute_mst_dsc_configs_for_link(state, dc_state, stream->link, vars)) {
+		if (!compute_mst_dsc_configs_for_link(state, dc_state, stream->link,
+			vars, &link_vars_start_index)) {
 			mutex_unlock(&aconnector->mst_mgr.lock);
 			return false;
 		}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index 12e5470..0ded4de 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -1085,6 +1085,8 @@ static void disable_dangling_plane(struct dc *dc, struct dc_state *context)
 		struct dc_stream_state *old_stream =
 				dc->current_state->res_ctx.pipe_ctx[i].stream;
 		bool should_disable = true;
+		bool pipe_split_change =
+			context->res_ctx.pipe_ctx[i].top_pipe != dc->current_state->res_ctx.pipe_ctx[i].top_pipe;
 
 		for (j = 0; j < context->stream_count; j++) {
 			if (old_stream == context->streams[j]) {
@@ -1092,6 +1094,9 @@ static void disable_dangling_plane(struct dc *dc, struct dc_state *context)
 				break;
 			}
 		}
+		if (!should_disable && pipe_split_change)
+			should_disable = true;
+
 		if (should_disable && old_stream) {
 			dc_rem_all_planes_for_stream(dc, old_stream, dangling_context);
 			disable_all_writeback_pipes_for_stream(dc, old_stream, dangling_context);
@@ -1887,6 +1892,7 @@ static bool is_flip_pending_in_pipes(struct dc *dc, struct dc_state *context)
 	return false;
 }
 
+#ifdef CONFIG_DRM_AMD_DC_DCN
 /* Perform updates here which need to be deferred until next vupdate
  *
  * i.e. blnd lut, 3dlut, and shaper lut bypass regs are double buffered
@@ -1896,7 +1902,6 @@ static bool is_flip_pending_in_pipes(struct dc *dc, struct dc_state *context)
  */
 static void process_deferred_updates(struct dc *dc)
 {
-#ifdef CONFIG_DRM_AMD_DC_DCN
 	int i = 0;
 
 	if (dc->debug.enable_mem_low_power.bits.cm) {
@@ -1905,8 +1910,8 @@ static void process_deferred_updates(struct dc *dc)
 			if (dc->res_pool->dpps[i]->funcs->dpp_deferred_update)
 				dc->res_pool->dpps[i]->funcs->dpp_deferred_update(dc->res_pool->dpps[i]);
 	}
-#endif
 }
+#endif /* CONFIG_DRM_AMD_DC_DCN */
 
 void dc_post_update_surfaces_to_stream(struct dc *dc)
 {
@@ -1933,7 +1938,9 @@ void dc_post_update_surfaces_to_stream(struct dc *dc)
 			dc->hwss.disable_plane(dc, &context->res_ctx.pipe_ctx[i]);
 		}
 
+#ifdef CONFIG_DRM_AMD_DC_DCN
 	process_deferred_updates(dc);
+#endif
 
 	dc->hwss.optimize_bandwidth(dc, context);
 
@@ -3603,7 +3610,8 @@ bool dc_enable_dmub_notifications(struct dc *dc)
 #if defined(CONFIG_DRM_AMD_DC_DCN)
 	/* YELLOW_CARP B0 USB4 DPIA needs dmub notifications for interrupts */
 	if (dc->ctx->asic_id.chip_family == FAMILY_YELLOW_CARP &&
-	    dc->ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0)
+	    dc->ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0 &&
+	    !dc->debug.dpia_debug.bits.disable_dpia)
 		return true;
 #endif
 	/* dmub aux needs dmub notifications to be enabled */
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link.c b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
index 2796bdd..6054478 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
@@ -4279,6 +4279,8 @@ void core_link_enable_stream(
 			 */
 			if (status != DC_FAIL_DP_LINK_TRAINING ||
 					pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
+				if (false == stream->link->link_status.link_active)
+					disable_link(stream->link, pipe_ctx->stream->signal);
 				BREAK_TO_DEBUGGER();
 				return;
 			}
@@ -4768,7 +4770,7 @@ uint32_t dc_bandwidth_in_kbps_from_timing(
 				timing->dsc_cfg.bits_per_pixel,
 				timing->dsc_cfg.num_slices_h,
 				timing->dsc_cfg.is_dp);
-#endif
+#endif /* CONFIG_DRM_AMD_DC_DCN */
 
 	switch (timing->display_color_depth) {
 	case COLOR_DEPTH_666:
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
index cc25ba0..cb7bf91 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
@@ -5329,6 +5329,14 @@ bool dc_link_dp_set_test_pattern(
 			return false;
 
 		if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_12) {
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+			if (test_pattern == DP_TEST_PATTERN_SQUARE_PULSE)
+				core_link_write_dpcd(link,
+						DP_LINK_SQUARE_PATTERN,
+						p_custom_pattern,
+						1);
+
+#endif
 			/* tell receiver that we are sending qualification
 			 * pattern DP 1.2 or later - DP receiver's link quality
 			 * pattern is set using DPCD LINK_QUAL_LANEx_SET
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_enc_cfg.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_enc_cfg.c
index 72b0f85..25e48a8 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_enc_cfg.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_enc_cfg.c
@@ -236,6 +236,23 @@ static struct link_encoder *get_link_enc_used_by_link(
 
 	return link_enc;
 }
+/* Clear all link encoder assignments. */
+static void clear_enc_assignments(struct dc_state *state)
+{
+	int i;
+	enum engine_id eng_id;
+	struct dc_stream_state *stream;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].valid = false;
+		eng_id = state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].eng_id;
+		stream = state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].stream;
+		if (eng_id != ENGINE_ID_UNKNOWN)
+			state->res_ctx.link_enc_cfg_ctx.link_enc_avail[eng_id - ENGINE_ID_DIGA] = eng_id;
+		if (stream)
+			stream->link_enc = NULL;
+	}
+}
 
 void link_enc_cfg_init(
 		struct dc *dc,
@@ -250,6 +267,8 @@ void link_enc_cfg_init(
 			state->res_ctx.link_enc_cfg_ctx.link_enc_avail[i] = ENGINE_ID_UNKNOWN;
 	}
 
+	clear_enc_assignments(state);
+
 	state->res_ctx.link_enc_cfg_ctx.mode = LINK_ENC_CFG_STEADY;
 }
 
@@ -265,6 +284,9 @@ void link_enc_cfg_link_encs_assign(
 
 	ASSERT(state->stream_count == stream_count);
 
+	if (stream_count == 0)
+		clear_enc_assignments(state);
+
 	/* Release DIG link encoder resources before running assignment algorithm. */
 	for (i = 0; i < stream_count; i++)
 		dc->res_pool->funcs->link_enc_unassign(state, streams[i]);
diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index a533979..3aac3f4 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -47,7 +47,7 @@ struct aux_payload;
 struct set_config_cmd_payload;
 struct dmub_notification;
 
-#define DC_VER "3.2.159"
+#define DC_VER "3.2.160"
 
 #define MAX_SURFACES 3
 #define MAX_PLANES 6
@@ -675,6 +675,7 @@ struct dc_debug_options {
 #endif
 	union mem_low_power_enable_options enable_mem_low_power;
 	union root_clock_optimization_options root_clock_optimization;
+	bool hpo_optimization;
 	bool force_vblank_alignment;
 
 	/* Enable dmub aux for legacy ddc */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
index bc87ea0..e68e9a8 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
@@ -898,6 +898,9 @@ struct dpcd_usb4_dp_tunneling_info {
 #ifndef DP_DFP_CAPABILITY_EXTENSION_SUPPORT
 #define DP_DFP_CAPABILITY_EXTENSION_SUPPORT		0x0A3
 #endif
+#ifndef DP_LINK_SQUARE_PATTERN
+#define DP_LINK_SQUARE_PATTERN				0x10F
+#endif
 #ifndef DP_DSC_CONFIGURATION
 #define DP_DSC_CONFIGURATION				0x161
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h b/drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h
index 989f5b6..a3fee92 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h
@@ -671,6 +671,7 @@ struct dce_hwseq_registers {
 	uint32_t MC_VM_FB_LOCATION_BASE;
 	uint32_t MC_VM_FB_LOCATION_TOP;
 	uint32_t MC_VM_FB_OFFSET;
+	uint32_t HPO_TOP_HW_CONTROL;
 };
  /* set field name */
 #define HWS_SF(blk_name, reg_name, field_name, post_fix)\
@@ -1152,7 +1153,8 @@ struct dce_hwseq_registers {
 	type DOMAIN_PGFSM_PWR_STATUS;\
 	type HPO_HDMISTREAMCLK_G_GATE_DIS;\
 	type DISABLE_HOSTVM_FORCE_ALLOW_PSTATE;\
-	type I2C_LIGHT_SLEEP_FORCE;
+	type I2C_LIGHT_SLEEP_FORCE;\
+	type HPO_IO_EN;
 
 struct dce_hwseq_shift {
 	HWSEQ_REG_FIELD_LIST(uint8_t)
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
index af3e68d..24e47df 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
@@ -1244,6 +1244,12 @@ void dce110_disable_stream(struct pipe_ctx *pipe_ctx)
 #endif
 	if (dc_is_dp_signal(pipe_ctx->stream->signal))
 		dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_DISCONNECT_DIG_FE_BE);
+
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+	if (dc->hwseq->funcs.setup_hpo_hw_control && is_dp_128b_132b_signal(pipe_ctx))
+		dc->hwseq->funcs.setup_hpo_hw_control(dc->hwseq, false);
+#endif
+
 }
 
 void dce110_unblank_stream(struct pipe_ctx *pipe_ctx,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
index a25732d..0b788d7 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
@@ -231,7 +231,7 @@ static void dcn10_log_hubp_states(struct dc *dc, void *log_ctx)
 
 		if (!s->blank_en)
 			DTN_INFO("[%2d]:  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh"
-				"%  8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh"
+				"  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh"
 				"  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh\n",
 				pool->hubps[i]->inst, dlg_regs->refcyc_h_blank_end, dlg_regs->dlg_vblank_end, dlg_regs->min_dst_y_next_start,
 				dlg_regs->refcyc_per_htotal, dlg_regs->refcyc_x_after_scaler, dlg_regs->dst_y_after_scaler,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
index cfee456..4f88376 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
@@ -2397,6 +2397,9 @@ void dcn20_enable_stream(struct pipe_ctx *pipe_ctx)
 	 * BY this, it is logic clean to separate stream and link
 	 */
 	if (is_dp_128b_132b_signal(pipe_ctx)) {
+		if (pipe_ctx->stream->ctx->dc->hwseq->funcs.setup_hpo_hw_control)
+			pipe_ctx->stream->ctx->dc->hwseq->funcs.setup_hpo_hw_control(
+				pipe_ctx->stream->ctx->dc->hwseq, true);
 		setup_dp_hpo_stream(pipe_ctx, true);
 		pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->enable_stream(
 				pipe_ctx->stream_res.hpo_dp_stream_enc);
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mpc.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mpc.c
index a82319f..9514973 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mpc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mpc.c
@@ -1381,13 +1381,11 @@ int mpcc3_release_rmu(struct mpc *mpc, int mpcc_id)
 
 }
 
-static void mpc3_mpc_init(struct mpc *mpc)
+static void mpc3_set_mpc_mem_lp_mode(struct mpc *mpc)
 {
 	struct dcn30_mpc *mpc30 = TO_DCN30_MPC(mpc);
 	int mpcc_id;
 
-	mpc1_mpc_init(mpc);
-
 	if (mpc->ctx->dc->debug.enable_mem_low_power.bits.mpc) {
 		if (mpc30->mpc_mask->MPC_RMU0_MEM_LOW_PWR_MODE && mpc30->mpc_mask->MPC_RMU1_MEM_LOW_PWR_MODE) {
 			REG_UPDATE(MPC_RMU_MEM_PWR_CTRL, MPC_RMU0_MEM_LOW_PWR_MODE, 3);
@@ -1405,7 +1403,7 @@ const struct mpc_funcs dcn30_mpc_funcs = {
 	.read_mpcc_state = mpc1_read_mpcc_state,
 	.insert_plane = mpc1_insert_plane,
 	.remove_mpcc = mpc1_remove_mpcc,
-	.mpc_init = mpc3_mpc_init,
+	.mpc_init = mpc1_mpc_init,
 	.mpc_init_single_inst = mpc1_mpc_init_single_inst,
 	.update_blending = mpc2_update_blending,
 	.cursor_lock = mpc1_cursor_lock,
@@ -1432,6 +1430,7 @@ const struct mpc_funcs dcn30_mpc_funcs = {
 	.power_on_mpc_mem_pwr = mpc3_power_on_ogam_lut,
 	.get_mpc_out_mux = mpc1_get_mpc_out_mux,
 	.set_bg_color = mpc1_set_bg_color,
+	.set_mpc_mem_lp_mode = mpc3_set_mpc_mem_lp_mode,
 };
 
 void dcn30_mpc_construct(struct dcn30_mpc *mpc30,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c
index e50c695..79a66e0 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c
@@ -2128,10 +2128,10 @@ static noinline void dcn30_calculate_wm_and_dlg_fp(
 		int pipe_cnt,
 		int vlevel)
 {
+	int maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb;
 	int i, pipe_idx;
-	double dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
-	bool pstate_en = context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] !=
-			dm_dram_clock_change_unsupported;
+	double dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][maxMpcComb];
+	bool pstate_en = context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] != dm_dram_clock_change_unsupported;
 
 	if (context->bw_ctx.dml.soc.min_dcfclk > dcfclk)
 		dcfclk = context->bw_ctx.dml.soc.min_dcfclk;
@@ -2207,6 +2207,7 @@ static noinline void dcn30_calculate_wm_and_dlg_fp(
 		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_enter_plus_exit_time_us;
 		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_exit_time_us;
 	}
+
 	context->bw_ctx.bw.dcn.watermarks.c.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
 	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
 	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c
index d24ad77..5dd1ce9 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c
@@ -66,6 +66,45 @@
 #define FN(reg_name, field_name) \
 	hws->shifts->field_name, hws->masks->field_name
 
+static void enable_memory_low_power(struct dc *dc)
+{
+	struct dce_hwseq *hws = dc->hwseq;
+	int i;
+
+	if (dc->debug.enable_mem_low_power.bits.dmcu) {
+		// Force ERAM to shutdown if DMCU is not enabled
+		if (dc->debug.disable_dmcu || dc->config.disable_dmcu) {
+			REG_UPDATE(DMU_MEM_PWR_CNTL, DMCU_ERAM_MEM_PWR_FORCE, 3);
+		}
+	}
+
+	// Set default OPTC memory power states
+	if (dc->debug.enable_mem_low_power.bits.optc) {
+		// Shutdown when unassigned and light sleep in VBLANK
+		REG_SET_2(ODM_MEM_PWR_CTRL3, 0, ODM_MEM_UNASSIGNED_PWR_MODE, 3, ODM_MEM_VBLANK_PWR_MODE, 1);
+	}
+
+	if (dc->debug.enable_mem_low_power.bits.vga) {
+		// Power down VGA memory
+		REG_UPDATE(MMHUBBUB_MEM_PWR_CNTL, VGA_MEM_PWR_FORCE, 1);
+	}
+
+	if (dc->debug.enable_mem_low_power.bits.mpc)
+		dc->res_pool->mpc->funcs->set_mpc_mem_lp_mode(dc->res_pool->mpc);
+
+
+	if (dc->debug.enable_mem_low_power.bits.vpg && dc->res_pool->stream_enc[0]->vpg->funcs->vpg_powerdown) {
+		// Power down VPGs
+		for (i = 0; i < dc->res_pool->stream_enc_count; i++)
+			dc->res_pool->stream_enc[i]->vpg->funcs->vpg_powerdown(dc->res_pool->stream_enc[i]->vpg);
+#if defined(CONFIG_DRM_AMD_DC_DCN)
+		for (i = 0; i < dc->res_pool->hpo_dp_stream_enc_count; i++)
+			dc->res_pool->hpo_dp_stream_enc[i]->vpg->funcs->vpg_powerdown(dc->res_pool->hpo_dp_stream_enc[i]->vpg);
+#endif
+	}
+
+}
+
 void dcn31_init_hw(struct dc *dc)
 {
 	struct abm **abms = dc->res_pool->multiple_abms;
@@ -108,35 +147,7 @@ void dcn31_init_hw(struct dc *dc)
 	if (res_pool->dccg->funcs->dccg_init)
 		res_pool->dccg->funcs->dccg_init(res_pool->dccg);
 
-	if (dc->debug.enable_mem_low_power.bits.dmcu) {
-		// Force ERAM to shutdown if DMCU is not enabled
-		if (dc->debug.disable_dmcu || dc->config.disable_dmcu) {
-			REG_UPDATE(DMU_MEM_PWR_CNTL, DMCU_ERAM_MEM_PWR_FORCE, 3);
-		}
-	}
-
-	// Set default OPTC memory power states
-	if (dc->debug.enable_mem_low_power.bits.optc) {
-		// Shutdown when unassigned and light sleep in VBLANK
-		REG_SET_2(ODM_MEM_PWR_CTRL3, 0, ODM_MEM_UNASSIGNED_PWR_MODE, 3, ODM_MEM_VBLANK_PWR_MODE, 1);
-	}
-
-	if (dc->debug.enable_mem_low_power.bits.vga) {
-		// Power down VGA memory
-		REG_UPDATE(MMHUBBUB_MEM_PWR_CNTL, VGA_MEM_PWR_FORCE, 1);
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dc->debug.enable_mem_low_power.bits.vpg && dc->res_pool->stream_enc[0]->vpg->funcs->vpg_powerdown) {
-		// Power down VPGs
-		for (i = 0; i < dc->res_pool->stream_enc_count; i++)
-			dc->res_pool->stream_enc[i]->vpg->funcs->vpg_powerdown(dc->res_pool->stream_enc[i]->vpg);
-#if defined(CONFIG_DRM_AMD_DC_DP2_0)
-		for (i = 0; i < dc->res_pool->hpo_dp_stream_enc_count; i++)
-			dc->res_pool->hpo_dp_stream_enc[i]->vpg->funcs->vpg_powerdown(dc->res_pool->hpo_dp_stream_enc[i]->vpg);
-#endif
-	}
-#endif
+	enable_memory_low_power(dc);
 
 	if (dc->ctx->dc_bios->fw_info_valid) {
 		res_pool->ref_clocks.xtalin_clock_inKhz =
@@ -264,6 +275,9 @@ void dcn31_init_hw(struct dc *dc)
 	if (dc->debug.enable_mem_low_power.bits.i2c)
 		REG_UPDATE(DIO_MEM_PWR_CTRL, I2C_LIGHT_SLEEP_FORCE, 1);
 
+	if (hws->funcs.setup_hpo_hw_control)
+		hws->funcs.setup_hpo_hw_control(hws, false);
+
 	if (!dc->debug.disable_clock_gate) {
 		/* enable all DCN clock gating */
 		REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0);
@@ -597,3 +611,9 @@ void dcn31_reset_hw_ctx_wrap(
 	/* New dc_state in the process of being applied to hardware. */
 	dc->current_state->res_ctx.link_enc_cfg_ctx.mode = LINK_ENC_CFG_TRANSIENT;
 }
+
+void dcn31_setup_hpo_hw_control(const struct dce_hwseq *hws, bool enable)
+{
+	if (hws->ctx->dc->debug.hpo_optimization)
+		REG_UPDATE(HPO_TOP_HW_CONTROL, HPO_IO_EN, !!enable);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.h b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.h
index 7ae45dd..edfc01d 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.h
@@ -54,5 +54,6 @@ void dcn31_reset_hw_ctx_wrap(
 bool dcn31_is_abm_supported(struct dc *dc,
 		struct dc_state *context, struct dc_stream_state *stream);
 void dcn31_init_pipes(struct dc *dc, struct dc_state *context);
+void dcn31_setup_hpo_hw_control(const struct dce_hwseq *hws, bool enable);
 
 #endif /* __DC_HWSS_DCN31_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_init.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_init.c
index c6a73778..05335a8 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_init.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_init.c
@@ -137,6 +137,7 @@ static const struct hwseq_private_funcs dcn31_private_funcs = {
 	.dccg_init = dcn20_dccg_init,
 	.set_blend_lut = dcn30_set_blend_lut,
 	.set_shaper_3dlut = dcn20_set_shaper_3dlut,
+	.setup_hpo_hw_control = dcn31_setup_hpo_hw_control,
 };
 
 void dcn31_hw_sequencer_construct(struct dc *dc)
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c
index 87b2c24..1889629 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c
@@ -860,7 +860,8 @@ static const struct dccg_mask dccg_mask = {
 	SR(D6VGA_CONTROL), \
 	SR(DC_IP_REQUEST_CNTL), \
 	SR(AZALIA_AUDIO_DTO), \
-	SR(AZALIA_CONTROLLER_CLOCK_GATING)
+	SR(AZALIA_CONTROLLER_CLOCK_GATING), \
+	SR(HPO_TOP_HW_CONTROL)
 
 static const struct dce_hwseq_registers hwseq_reg = {
 		HWSEQ_DCN31_REG_LIST()
@@ -898,7 +899,8 @@ static const struct dce_hwseq_registers hwseq_reg = {
 	HWS_SF(, ODM_MEM_PWR_CTRL3, ODM_MEM_UNASSIGNED_PWR_MODE, mask_sh), \
 	HWS_SF(, ODM_MEM_PWR_CTRL3, ODM_MEM_VBLANK_PWR_MODE, mask_sh), \
 	HWS_SF(, MMHUBBUB_MEM_PWR_CNTL, VGA_MEM_PWR_FORCE, mask_sh), \
-	HWS_SF(, DIO_MEM_PWR_CTRL, I2C_LIGHT_SLEEP_FORCE, mask_sh)
+	HWS_SF(, DIO_MEM_PWR_CTRL, I2C_LIGHT_SLEEP_FORCE, mask_sh), \
+	HWS_SF(, HPO_TOP_HW_CONTROL, HPO_IO_EN, mask_sh)
 
 static const struct dce_hwseq_shift hwseq_shift = {
 		HWSEQ_DCN31_MASK_SH_LIST(__SHIFT)
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c
index e3d9f1dec..f47d82d 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c
@@ -3576,16 +3576,9 @@ static double TruncToValidBPP(
 		MinDSCBPP = 8;
 		MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16;
 	} else {
-		if (Output == dm_hdmi) {
-			NonDSCBPP0 = 24;
-			NonDSCBPP1 = 24;
-			NonDSCBPP2 = 24;
-		}
-		else {
-			NonDSCBPP0 = 16;
-			NonDSCBPP1 = 20;
-			NonDSCBPP2 = 24;
-		}
+		NonDSCBPP0 = 16;
+		NonDSCBPP1 = 20;
+		NonDSCBPP2 = 24;
 
 		if (Format == dm_n422) {
 			MinDSCBPP = 7;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c
index d58925c..7e937bd 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c
@@ -3892,15 +3892,11 @@ static double TruncToValidBPP(
 		MinDSCBPP = 8;
 		MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16;
 	} else {
-		if (Output == dm_hdmi) {
-			NonDSCBPP0 = 24;
-			NonDSCBPP1 = 24;
-			NonDSCBPP2 = 24;
-		} else {
-			NonDSCBPP0 = 16;
-			NonDSCBPP1 = 20;
-			NonDSCBPP2 = 24;
-		}
+
+		NonDSCBPP0 = 16;
+		NonDSCBPP1 = 20;
+		NonDSCBPP2 = 24;
+
 		if (Format == dm_n422) {
 			MinDSCBPP = 7;
 			MaxDSCBPP = 2 * DSCInputBitPerComponent - 1.0 / 16.0;
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h b/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h
index 04d6ec3..f5fd2a0 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h
@@ -367,6 +367,7 @@ struct mpc_funcs {
 	void (*set_bg_color)(struct mpc *mpc,
 			struct tg_color *bg_color,
 			int mpcc_id);
+	void (*set_mpc_mem_lp_mode)(struct mpc *mpc);
 };
 
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer_private.h b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer_private.h
index f324285..c200825 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer_private.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer_private.h
@@ -143,6 +143,7 @@ struct hwseq_private_funcs {
 			const struct dc_plane_state *plane_state);
 	void (*PLAT_58856_wa)(struct dc_state *context,
 			struct pipe_ctx *pipe_ctx);
+	void (*setup_hpo_hw_control)(const struct dce_hwseq *hws, bool enable);
 };
 
 struct dce_hwseq {
diff --git a/drivers/gpu/drm/amd/display/dmub/dmub_srv.h b/drivers/gpu/drm/amd/display/dmub/dmub_srv.h
index 717c0e5..cd204ee 100644
--- a/drivers/gpu/drm/amd/display/dmub/dmub_srv.h
+++ b/drivers/gpu/drm/amd/display/dmub/dmub_srv.h
@@ -238,6 +238,7 @@ struct dmub_srv_hw_params {
 	bool load_inst_const;
 	bool skip_panel_power_sequence;
 	bool disable_z10;
+	bool power_optimization;
 	bool dpia_supported;
 	bool disable_dpia;
 };
diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
index 0293c58..c29a67c 100644
--- a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
+++ b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
@@ -46,10 +46,10 @@
 
 /* Firmware versioning. */
 #ifdef DMUB_EXPOSE_VERSION
-#define DMUB_FW_VERSION_GIT_HASH 0x9525efb5
+#define DMUB_FW_VERSION_GIT_HASH 0x1d82d23e
 #define DMUB_FW_VERSION_MAJOR 0
 #define DMUB_FW_VERSION_MINOR 0
-#define DMUB_FW_VERSION_REVISION 90
+#define DMUB_FW_VERSION_REVISION 91
 #define DMUB_FW_VERSION_TEST 0
 #define DMUB_FW_VERSION_VBIOS 0
 #define DMUB_FW_VERSION_HOTFIX 0
diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c
index 10ebf20..fa05691 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c
@@ -340,6 +340,7 @@ void dmub_dcn31_enable_dmub_boot_options(struct dmub_srv *dmub, const struct dmu
 	boot_options.bits.z10_disable = params->disable_z10;
 	boot_options.bits.dpia_supported = params->dpia_supported;
 	boot_options.bits.enable_dpia = params->disable_dpia ? 0 : 1;
+	boot_options.bits.power_optimization = params->power_optimization;
 
 	boot_options.bits.sel_mux_phy_c_d_phy_f_g = (dmub->asic == DMUB_ASIC_DCN31B) ? 1 : 0;
 
diff --git a/drivers/gpu/drm/amd/pm/amdgpu_pm.c b/drivers/gpu/drm/amd/pm/amdgpu_pm.c
index 49fe415..41472ed 100644
--- a/drivers/gpu/drm/amd/pm/amdgpu_pm.c
+++ b/drivers/gpu/drm/amd/pm/amdgpu_pm.c
@@ -2094,6 +2094,10 @@ static int default_attr_update(struct amdgpu_device *adev, struct amdgpu_device_
 	} else if (DEVICE_ATTR_IS(pp_dpm_dclk)) {
 		if (!(asic_type == CHIP_VANGOGH || asic_type == CHIP_SIENNA_CICHLID))
 			*states = ATTR_STATE_UNSUPPORTED;
+	} else if (DEVICE_ATTR_IS(pp_power_profile_mode)) {
+		if (!adev->powerplay.pp_funcs->get_power_profile_mode ||
+		    amdgpu_dpm_get_power_profile_mode(adev, NULL) == -EOPNOTSUPP)
+			*states = ATTR_STATE_UNSUPPORTED;
 	}
 
 	switch (asic_type) {
diff --git a/drivers/gpu/drm/amd/pm/inc/smu_v13_0_1_ppsmc.h b/drivers/gpu/drm/amd/pm/inc/smu_v13_0_1_ppsmc.h
index 1d34479..fc91988 100644
--- a/drivers/gpu/drm/amd/pm/inc/smu_v13_0_1_ppsmc.h
+++ b/drivers/gpu/drm/amd/pm/inc/smu_v13_0_1_ppsmc.h
@@ -51,7 +51,7 @@
 #define PPSMC_MSG_PowerUpVcn                    0x07 ///< Power up VCN; VCN is power gated by default
 #define PPSMC_MSG_SetHardMinVcn                 0x08 ///< For wireless display
 #define PPSMC_MSG_SetSoftMinGfxclk              0x09 ///< Set SoftMin for GFXCLK, argument is frequency in MHz
-#define PPSMC_MSG_ActiveProcessNotify           0x0A ///< Set active work load type
+#define PPSMC_MSG_ActiveProcessNotify           0x0A ///< Deprecated (Not to be used)
 #define PPSMC_MSG_ForcePowerDownGfx             0x0B ///< Force power down GFX, i.e. enter GFXOFF
 #define PPSMC_MSG_PrepareMp1ForUnload           0x0C ///< Prepare PMFW for GFX driver unload
 #define PPSMC_MSG_SetDriverDramAddrHigh         0x0D ///< Set high 32 bits of DRAM address for Driver table transfer
@@ -63,7 +63,7 @@
 #define PPSMC_MSG_SetHardMinSocclkByFreq        0x13 ///< Set hard min for SOC CLK
 #define PPSMC_MSG_SetSoftMinFclk                0x14 ///< Set hard min for FCLK
 #define PPSMC_MSG_SetSoftMinVcn                 0x15 ///< Set soft min for VCN clocks (VCLK and DCLK)
-#define PPSMC_MSG_SPARE0                        0x16 ///< Spared
+#define PPSMC_MSG_SPARE                         0x16 ///< Spare
 #define PPSMC_MSG_GetGfxclkFrequency            0x17 ///< Get GFX clock frequency
 #define PPSMC_MSG_GetFclkFrequency              0x18 ///< Get FCLK frequency
 #define PPSMC_MSG_AllowGfxOff                   0x19 ///< Inform PMFW of allowing GFXOFF entry
diff --git a/drivers/gpu/drm/amd/pm/powerplay/amd_powerplay.c b/drivers/gpu/drm/amd/pm/powerplay/amd_powerplay.c
index 3212150..8d796ed 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/amd_powerplay.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/amd_powerplay.c
@@ -875,34 +875,30 @@ pp_dpm_get_vce_clock_state(void *handle, unsigned idx)
 static int pp_get_power_profile_mode(void *handle, char *buf)
 {
 	struct pp_hwmgr *hwmgr = handle;
+	int ret;
 
-	if (!hwmgr || !hwmgr->pm_en || !buf)
+	if (!hwmgr || !hwmgr->pm_en || !hwmgr->hwmgr_func->get_power_profile_mode)
+		return -EOPNOTSUPP;
+	if (!buf)
 		return -EINVAL;
 
-	if (hwmgr->hwmgr_func->get_power_profile_mode == NULL) {
-		pr_info_ratelimited("%s was not implemented.\n", __func__);
-		return snprintf(buf, PAGE_SIZE, "\n");
-	}
-
-	return hwmgr->hwmgr_func->get_power_profile_mode(hwmgr, buf);
+	mutex_lock(&hwmgr->smu_lock);
+	ret = hwmgr->hwmgr_func->get_power_profile_mode(hwmgr, buf);
+	mutex_unlock(&hwmgr->smu_lock);
+	return ret;
 }
 
 static int pp_set_power_profile_mode(void *handle, long *input, uint32_t size)
 {
 	struct pp_hwmgr *hwmgr = handle;
-	int ret = -EINVAL;
+	int ret = -EOPNOTSUPP;
 
-	if (!hwmgr || !hwmgr->pm_en)
+	if (!hwmgr || !hwmgr->pm_en || !hwmgr->hwmgr_func->set_power_profile_mode)
 		return ret;
 
-	if (hwmgr->hwmgr_func->set_power_profile_mode == NULL) {
-		pr_info_ratelimited("%s was not implemented.\n", __func__);
-		return ret;
-	}
-
 	if (hwmgr->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL) {
 		pr_debug("power profile setting is for manual dpm mode only.\n");
-		return ret;
+		return -EINVAL;
 	}
 
 	mutex_lock(&hwmgr->smu_lock);
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.c
index 1de3ae7..258c573 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu10_hwmgr.c
@@ -1024,6 +1024,8 @@ static int smu10_print_clock_levels(struct pp_hwmgr *hwmgr,
 	uint32_t min_freq, max_freq = 0;
 	uint32_t ret = 0;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	switch (type) {
 	case PP_SCLK:
 		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetGfxclkFrequency, &now);
@@ -1065,7 +1067,7 @@ static int smu10_print_clock_levels(struct pp_hwmgr *hwmgr,
 			if (ret)
 				return ret;
 
-			size = sysfs_emit(buf, "%s:\n", "OD_SCLK");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_SCLK");
 			size += sysfs_emit_at(buf, size, "0: %10uMhz\n",
 			(data->gfx_actual_soft_min_freq > 0) ? data->gfx_actual_soft_min_freq : min_freq);
 			size += sysfs_emit_at(buf, size, "1: %10uMhz\n",
@@ -1081,7 +1083,7 @@ static int smu10_print_clock_levels(struct pp_hwmgr *hwmgr,
 			if (ret)
 				return ret;
 
-			size = sysfs_emit(buf, "%s:\n", "OD_RANGE");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_RANGE");
 			size += sysfs_emit_at(buf, size, "SCLK: %7uMHz %10uMHz\n",
 				min_freq, max_freq);
 		}
@@ -1456,6 +1458,8 @@ static int smu10_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf)
 	if (!buf)
 		return -EINVAL;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	size += sysfs_emit_at(buf, size, "%s %16s %s %s %s %s\n",title[0],
 			title[1], title[2], title[3], title[4], title[5]);
 
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c
index e7803ce..aceebf5 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c
@@ -4914,6 +4914,8 @@ static int smu7_print_clock_levels(struct pp_hwmgr *hwmgr,
 	int size = 0;
 	uint32_t i, now, clock, pcie_speed;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	switch (type) {
 	case PP_SCLK:
 		smum_send_msg_to_smc(hwmgr, PPSMC_MSG_API_GetSclkFrequency, &clock);
@@ -4963,7 +4965,7 @@ static int smu7_print_clock_levels(struct pp_hwmgr *hwmgr,
 		break;
 	case OD_SCLK:
 		if (hwmgr->od_enabled) {
-			size = sysfs_emit(buf, "%s:\n", "OD_SCLK");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_SCLK");
 			for (i = 0; i < odn_sclk_table->num_of_pl; i++)
 				size += sysfs_emit_at(buf, size, "%d: %10uMHz %10umV\n",
 					i, odn_sclk_table->entries[i].clock/100,
@@ -4972,7 +4974,7 @@ static int smu7_print_clock_levels(struct pp_hwmgr *hwmgr,
 		break;
 	case OD_MCLK:
 		if (hwmgr->od_enabled) {
-			size = sysfs_emit(buf, "%s:\n", "OD_MCLK");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_MCLK");
 			for (i = 0; i < odn_mclk_table->num_of_pl; i++)
 				size += sysfs_emit_at(buf, size, "%d: %10uMHz %10umV\n",
 					i, odn_mclk_table->entries[i].clock/100,
@@ -4981,7 +4983,7 @@ static int smu7_print_clock_levels(struct pp_hwmgr *hwmgr,
 		break;
 	case OD_RANGE:
 		if (hwmgr->od_enabled) {
-			size = sysfs_emit(buf, "%s:\n", "OD_RANGE");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_RANGE");
 			size += sysfs_emit_at(buf, size, "SCLK: %7uMHz %10uMHz\n",
 				data->golden_dpm_table.sclk_table.dpm_levels[0].value/100,
 				hwmgr->platform_descriptor.overdriveLimit.engineClock/100);
@@ -5518,6 +5520,8 @@ static int smu7_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf)
 	if (!buf)
 		return -EINVAL;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	size += sysfs_emit_at(buf, size, "%s %16s %16s %16s %16s %16s %16s %16s\n",
 			title[0], title[1], title[2], title[3],
 			title[4], title[5], title[6], title[7]);
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu8_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu8_hwmgr.c
index b94a77e..8e28a8e 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu8_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu8_hwmgr.c
@@ -1550,6 +1550,8 @@ static int smu8_print_clock_levels(struct pp_hwmgr *hwmgr,
 	uint32_t i, now;
 	int size = 0;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	switch (type) {
 	case PP_SCLK:
 		now = PHM_GET_FIELD(cgs_read_ind_register(hwmgr->device,
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu_helper.h b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu_helper.h
index ad33983..2a75da1 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu_helper.h
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu_helper.h
@@ -109,6 +109,19 @@ int phm_irq_process(struct amdgpu_device *adev,
 			   struct amdgpu_irq_src *source,
 			   struct amdgpu_iv_entry *entry);
 
+/*
+ * Helper function to make sysfs_emit_at() happy. Align buf to
+ * the current page boundary and record the offset.
+ */
+static inline void phm_get_sysfs_buf(char **buf, int *offset)
+{
+	if (!*buf || !offset)
+		return;
+
+	*offset = offset_in_page(*buf);
+	*buf -= *offset;
+}
+
 int smu9_register_irq_handlers(struct pp_hwmgr *hwmgr);
 
 void *smu_atom_get_data_table(void *dev, uint32_t table, uint16_t *size,
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c
index c152a61d..c981fc2 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c
@@ -4548,6 +4548,8 @@ static int vega10_get_ppfeature_status(struct pp_hwmgr *hwmgr, char *buf)
 	int ret = 0;
 	int size = 0;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	ret = vega10_get_enabled_smc_features(hwmgr, &features_enabled);
 	PP_ASSERT_WITH_CODE(!ret,
 			"[EnableAllSmuFeatures] Failed to get enabled smc features!",
@@ -4637,6 +4639,8 @@ static int vega10_print_clock_levels(struct pp_hwmgr *hwmgr,
 
 	int i, now, size = 0, count = 0;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	switch (type) {
 	case PP_SCLK:
 		if (data->registry_data.sclk_dpm_key_disabled)
@@ -4717,7 +4721,7 @@ static int vega10_print_clock_levels(struct pp_hwmgr *hwmgr,
 
 	case OD_SCLK:
 		if (hwmgr->od_enabled) {
-			size = sysfs_emit(buf, "%s:\n", "OD_SCLK");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_SCLK");
 			podn_vdd_dep = &data->odn_dpm_table.vdd_dep_on_sclk;
 			for (i = 0; i < podn_vdd_dep->count; i++)
 				size += sysfs_emit_at(buf, size, "%d: %10uMhz %10umV\n",
@@ -4727,7 +4731,7 @@ static int vega10_print_clock_levels(struct pp_hwmgr *hwmgr,
 		break;
 	case OD_MCLK:
 		if (hwmgr->od_enabled) {
-			size = sysfs_emit(buf, "%s:\n", "OD_MCLK");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_MCLK");
 			podn_vdd_dep = &data->odn_dpm_table.vdd_dep_on_mclk;
 			for (i = 0; i < podn_vdd_dep->count; i++)
 				size += sysfs_emit_at(buf, size, "%d: %10uMhz %10umV\n",
@@ -4737,7 +4741,7 @@ static int vega10_print_clock_levels(struct pp_hwmgr *hwmgr,
 		break;
 	case OD_RANGE:
 		if (hwmgr->od_enabled) {
-			size = sysfs_emit(buf, "%s:\n", "OD_RANGE");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_RANGE");
 			size += sysfs_emit_at(buf, size, "SCLK: %7uMHz %10uMHz\n",
 				data->golden_dpm_table.gfx_table.dpm_levels[0].value/100,
 				hwmgr->platform_descriptor.overdriveLimit.engineClock/100);
@@ -5112,6 +5116,8 @@ static int vega10_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf)
 	if (!buf)
 		return -EINVAL;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	size += sysfs_emit_at(buf, size, "%s %16s %s %s %s %s\n",title[0],
 			title[1], title[2], title[3], title[4], title[5]);
 
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega12_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega12_hwmgr.c
index 8558718..f7e783e 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega12_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega12_hwmgr.c
@@ -2141,6 +2141,8 @@ static int vega12_get_ppfeature_status(struct pp_hwmgr *hwmgr, char *buf)
 	int ret = 0;
 	int size = 0;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	ret = vega12_get_enabled_smc_features(hwmgr, &features_enabled);
 	PP_ASSERT_WITH_CODE(!ret,
 		"[EnableAllSmuFeatures] Failed to get enabled smc features!",
@@ -2244,6 +2246,8 @@ static int vega12_print_clock_levels(struct pp_hwmgr *hwmgr,
 	int i, now, size = 0;
 	struct pp_clock_levels_with_latency clocks;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	switch (type) {
 	case PP_SCLK:
 		PP_ASSERT_WITH_CODE(
diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_hwmgr.c
index 0cf39c1..03e63be 100644
--- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_hwmgr.c
+++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_hwmgr.c
@@ -3238,6 +3238,8 @@ static int vega20_get_ppfeature_status(struct pp_hwmgr *hwmgr, char *buf)
 	int ret = 0;
 	int size = 0;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	ret = vega20_get_enabled_smc_features(hwmgr, &features_enabled);
 	PP_ASSERT_WITH_CODE(!ret,
 			"[EnableAllSmuFeatures] Failed to get enabled smc features!",
@@ -3364,6 +3366,8 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr,
 	int ret = 0;
 	uint32_t gen_speed, lane_width, current_gen_speed, current_lane_width;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	switch (type) {
 	case PP_SCLK:
 		ret = vega20_get_current_clk_freq(hwmgr, PPCLK_GFXCLK, &now);
@@ -3479,7 +3483,7 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr,
 	case OD_SCLK:
 		if (od8_settings[OD8_SETTING_GFXCLK_FMIN].feature_id &&
 		    od8_settings[OD8_SETTING_GFXCLK_FMAX].feature_id) {
-			size = sysfs_emit(buf, "%s:\n", "OD_SCLK");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_SCLK");
 			size += sysfs_emit_at(buf, size, "0: %10uMhz\n",
 				od_table->GfxclkFmin);
 			size += sysfs_emit_at(buf, size, "1: %10uMhz\n",
@@ -3489,7 +3493,7 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr,
 
 	case OD_MCLK:
 		if (od8_settings[OD8_SETTING_UCLK_FMAX].feature_id) {
-			size = sysfs_emit(buf, "%s:\n", "OD_MCLK");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_MCLK");
 			size += sysfs_emit_at(buf, size, "1: %10uMhz\n",
 				od_table->UclkFmax);
 		}
@@ -3503,7 +3507,7 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr,
 		    od8_settings[OD8_SETTING_GFXCLK_VOLTAGE1].feature_id &&
 		    od8_settings[OD8_SETTING_GFXCLK_VOLTAGE2].feature_id &&
 		    od8_settings[OD8_SETTING_GFXCLK_VOLTAGE3].feature_id) {
-			size = sysfs_emit(buf, "%s:\n", "OD_VDDC_CURVE");
+			size += sysfs_emit_at(buf, size, "%s:\n", "OD_VDDC_CURVE");
 			size += sysfs_emit_at(buf, size, "0: %10uMhz %10dmV\n",
 				od_table->GfxclkFreq1,
 				od_table->GfxclkVolt1 / VOLTAGE_SCALE);
@@ -3518,7 +3522,7 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr,
 		break;
 
 	case OD_RANGE:
-		size = sysfs_emit(buf, "%s:\n", "OD_RANGE");
+		size += sysfs_emit_at(buf, size, "%s:\n", "OD_RANGE");
 
 		if (od8_settings[OD8_SETTING_GFXCLK_FMIN].feature_id &&
 		    od8_settings[OD8_SETTING_GFXCLK_FMAX].feature_id) {
@@ -4003,6 +4007,8 @@ static int vega20_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf)
 	if (!buf)
 		return -EINVAL;
 
+	phm_get_sysfs_buf(&buf, &size);
+
 	size += sysfs_emit_at(buf, size, "%16s %s %s %s %s %s %s %s %s %s %s\n",
 			title[0], title[1], title[2], title[3], title[4], title[5],
 			title[6], title[7], title[8], title[9], title[10]);
diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
index b06c59d..01168b8 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
@@ -1468,7 +1468,7 @@ static int smu_disable_dpms(struct smu_context *smu)
 			dev_err(adev->dev, "Failed to disable smu features.\n");
 	}
 
-	if (adev->ip_versions[MP1_HWIP][0] >= IP_VERSION(11, 0, 0) &&
+	if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 0, 0) &&
 	    adev->gfx.rlc.funcs->stop)
 		adev->gfx.rlc.funcs->stop(adev);
 
@@ -2534,13 +2534,15 @@ static int smu_get_power_profile_mode(void *handle, char *buf)
 	struct smu_context *smu = handle;
 	int ret = 0;
 
-	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
+	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled ||
+	    !smu->ppt_funcs->get_power_profile_mode)
 		return -EOPNOTSUPP;
+	if (!buf)
+		return -EINVAL;
 
 	mutex_lock(&smu->mutex);
 
-	if (smu->ppt_funcs->get_power_profile_mode)
-		ret = smu->ppt_funcs->get_power_profile_mode(smu, buf);
+	ret = smu->ppt_funcs->get_power_profile_mode(smu, buf);
 
 	mutex_unlock(&smu->mutex);
 
@@ -2554,7 +2556,8 @@ static int smu_set_power_profile_mode(void *handle,
 	struct smu_context *smu = handle;
 	int ret = 0;
 
-	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
+	if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled ||
+	    !smu->ppt_funcs->set_power_profile_mode)
 		return -EOPNOTSUPP;
 
 	mutex_lock(&smu->mutex);
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c
index a403657..8215bbf 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/yellow_carp_ppt.c
@@ -64,7 +64,6 @@ static struct cmn2asic_msg_mapping yellow_carp_message_map[SMU_MSG_MAX_COUNT] =
 	MSG_MAP(PowerDownVcn,                   PPSMC_MSG_PowerDownVcn,			1),
 	MSG_MAP(PowerUpVcn,                     PPSMC_MSG_PowerUpVcn,			1),
 	MSG_MAP(SetHardMinVcn,                  PPSMC_MSG_SetHardMinVcn,		1),
-	MSG_MAP(ActiveProcessNotify,            PPSMC_MSG_ActiveProcessNotify,		1),
 	MSG_MAP(PrepareMp1ForUnload,            PPSMC_MSG_PrepareMp1ForUnload,      1),
 	MSG_MAP(SetDriverDramAddrHigh,          PPSMC_MSG_SetDriverDramAddrHigh,	1),
 	MSG_MAP(SetDriverDramAddrLow,           PPSMC_MSG_SetDriverDramAddrLow,		1),
@@ -135,14 +134,6 @@ static struct cmn2asic_mapping yellow_carp_table_map[SMU_TABLE_COUNT] = {
 	TAB_MAP_VALID(CUSTOM_DPM),
 	TAB_MAP_VALID(DPMCLOCKS),
 };
-
-static struct cmn2asic_mapping yellow_carp_workload_map[PP_SMC_POWER_PROFILE_COUNT] = {
-	WORKLOAD_MAP(PP_SMC_POWER_PROFILE_FULLSCREEN3D,		WORKLOAD_PPLIB_FULL_SCREEN_3D_BIT),
-	WORKLOAD_MAP(PP_SMC_POWER_PROFILE_VIDEO,		WORKLOAD_PPLIB_VIDEO_BIT),
-	WORKLOAD_MAP(PP_SMC_POWER_PROFILE_VR,			WORKLOAD_PPLIB_VR_BIT),
-	WORKLOAD_MAP(PP_SMC_POWER_PROFILE_COMPUTE,		WORKLOAD_PPLIB_COMPUTE_BIT),
-	WORKLOAD_MAP(PP_SMC_POWER_PROFILE_CUSTOM,		WORKLOAD_PPLIB_CUSTOM_BIT),
-};
 	
 static int yellow_carp_init_smc_tables(struct smu_context *smu)
 {
@@ -543,81 +534,6 @@ static int yellow_carp_set_watermarks_table(struct smu_context *smu,
 	return 0;
 }
 
-static int yellow_carp_get_power_profile_mode(struct smu_context *smu,
-						char *buf)
-{
-	static const char *profile_name[] = {
-					"BOOTUP_DEFAULT",
-					"3D_FULL_SCREEN",
-					"POWER_SAVING",
-					"VIDEO",
-					"VR",
-					"COMPUTE",
-					"CUSTOM"};
-	uint32_t i, size = 0;
-	int16_t workload_type = 0;
-
-	if (!buf)
-		return -EINVAL;
-
-	for (i = 0; i <= PP_SMC_POWER_PROFILE_CUSTOM; i++) {
-		/*
-		 * Conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT.
-		 * Not all profile modes are supported on yellow carp.
-		 */
-		workload_type = smu_cmn_to_asic_specific_index(smu,
-							       CMN2ASIC_MAPPING_WORKLOAD,
-							       i);
-
-		if (workload_type < 0)
-			continue;
-
-		size += sysfs_emit_at(buf, size, "%2d %14s%s\n",
-			i, profile_name[i], (i == smu->power_profile_mode) ? "*" : " ");
-	}
-
-	return size;
-}
-
-static int yellow_carp_set_power_profile_mode(struct smu_context *smu,
-						long *input, uint32_t size)
-{
-	int workload_type, ret;
-	uint32_t profile_mode = input[size];
-
-	if (profile_mode > PP_SMC_POWER_PROFILE_CUSTOM) {
-		dev_err(smu->adev->dev, "Invalid power profile mode %d\n", profile_mode);
-		return -EINVAL;
-	}
-
-	if (profile_mode == PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT ||
-			profile_mode == PP_SMC_POWER_PROFILE_POWERSAVING)
-		return 0;
-
-	/* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
-	workload_type = smu_cmn_to_asic_specific_index(smu,
-						       CMN2ASIC_MAPPING_WORKLOAD,
-						       profile_mode);
-	if (workload_type < 0) {
-		dev_dbg(smu->adev->dev, "Unsupported power profile mode %d on YELLOWCARP\n",
-					profile_mode);
-		return -EINVAL;
-	}
-
-	ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_ActiveProcessNotify,
-				    1 << workload_type,
-				    NULL);
-	if (ret) {
-		dev_err_once(smu->adev->dev, "Fail to set workload type %d\n",
-					workload_type);
-		return ret;
-	}
-
-	smu->power_profile_mode = profile_mode;
-
-	return 0;
-}
-
 static ssize_t yellow_carp_get_gpu_metrics(struct smu_context *smu,
 						void **table)
 {
@@ -1238,8 +1154,6 @@ static const struct pptable_funcs yellow_carp_ppt_funcs = {
 	.read_sensor = yellow_carp_read_sensor,
 	.is_dpm_running = yellow_carp_is_dpm_running,
 	.set_watermarks_table = yellow_carp_set_watermarks_table,
-	.get_power_profile_mode = yellow_carp_get_power_profile_mode,
-	.set_power_profile_mode = yellow_carp_set_power_profile_mode,
 	.get_gpu_metrics = yellow_carp_get_gpu_metrics,
 	.get_enabled_mask = smu_cmn_get_enabled_32_bits_mask,
 	.get_pp_feature_mask = smu_cmn_get_pp_feature_mask,
@@ -1261,6 +1175,5 @@ void yellow_carp_set_ppt_funcs(struct smu_context *smu)
 	smu->message_map = yellow_carp_message_map;
 	smu->feature_map = yellow_carp_feature_mask_map;
 	smu->table_map = yellow_carp_table_map;
-	smu->workload_map = yellow_carp_workload_map;
 	smu->is_apu = true;
 }
diff --git a/drivers/gpu/drm/bridge/lontium-lt9611uxc.c b/drivers/gpu/drm/bridge/lontium-lt9611uxc.c
index 3cac16d..010657e 100644
--- a/drivers/gpu/drm/bridge/lontium-lt9611uxc.c
+++ b/drivers/gpu/drm/bridge/lontium-lt9611uxc.c
@@ -167,9 +167,10 @@ static void lt9611uxc_hpd_work(struct work_struct *work)
 	struct lt9611uxc *lt9611uxc = container_of(work, struct lt9611uxc, work);
 	bool connected;
 
-	if (lt9611uxc->connector.dev)
-		drm_kms_helper_hotplug_event(lt9611uxc->connector.dev);
-	else {
+	if (lt9611uxc->connector.dev) {
+		if (lt9611uxc->connector.dev->mode_config.funcs)
+			drm_kms_helper_hotplug_event(lt9611uxc->connector.dev);
+	} else {
 
 		mutex_lock(&lt9611uxc->ocm_lock);
 		connected = lt9611uxc->hdmi_connected;
@@ -339,6 +340,8 @@ static int lt9611uxc_connector_init(struct drm_bridge *bridge, struct lt9611uxc
 		return -ENODEV;
 	}
 
+	lt9611uxc->connector.polled = DRM_CONNECTOR_POLL_HPD;
+
 	drm_connector_helper_add(&lt9611uxc->connector,
 				 &lt9611uxc_bridge_connector_helper_funcs);
 	ret = drm_connector_init(bridge->dev, &lt9611uxc->connector,
diff --git a/drivers/gpu/drm/bridge/lvds-codec.c b/drivers/gpu/drm/bridge/lvds-codec.c
index dcf579a..ad460b9 100644
--- a/drivers/gpu/drm/bridge/lvds-codec.c
+++ b/drivers/gpu/drm/bridge/lvds-codec.c
@@ -12,6 +12,7 @@
 #include <linux/platform_device.h>
 #include <linux/regulator/consumer.h>
 
+#include <drm/drm_atomic_helper.h>
 #include <drm/drm_bridge.h>
 #include <drm/drm_panel.h>
 
@@ -22,6 +23,7 @@ struct lvds_codec {
 	struct regulator *vcc;
 	struct gpio_desc *powerdown_gpio;
 	u32 connector_type;
+	unsigned int bus_format;
 };
 
 static inline struct lvds_codec *to_lvds_codec(struct drm_bridge *bridge)
@@ -74,12 +76,50 @@ static const struct drm_bridge_funcs funcs = {
 	.disable = lvds_codec_disable,
 };
 
+#define MAX_INPUT_SEL_FORMATS 1
+static u32 *
+lvds_codec_atomic_get_input_bus_fmts(struct drm_bridge *bridge,
+				     struct drm_bridge_state *bridge_state,
+				     struct drm_crtc_state *crtc_state,
+				     struct drm_connector_state *conn_state,
+				     u32 output_fmt,
+				     unsigned int *num_input_fmts)
+{
+	struct lvds_codec *lvds_codec = to_lvds_codec(bridge);
+	u32 *input_fmts;
+
+	*num_input_fmts = 0;
+
+	input_fmts = kcalloc(MAX_INPUT_SEL_FORMATS, sizeof(*input_fmts),
+			     GFP_KERNEL);
+	if (!input_fmts)
+		return NULL;
+
+	input_fmts[0] = lvds_codec->bus_format;
+	*num_input_fmts = MAX_INPUT_SEL_FORMATS;
+
+	return input_fmts;
+}
+
+static const struct drm_bridge_funcs funcs_decoder = {
+	.attach = lvds_codec_attach,
+	.enable = lvds_codec_enable,
+	.disable = lvds_codec_disable,
+	.atomic_duplicate_state = drm_atomic_helper_bridge_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_bridge_destroy_state,
+	.atomic_reset = drm_atomic_helper_bridge_reset,
+	.atomic_get_input_bus_fmts = lvds_codec_atomic_get_input_bus_fmts,
+};
+
 static int lvds_codec_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct device_node *panel_node;
+	struct device_node *bus_node;
 	struct drm_panel *panel;
 	struct lvds_codec *lvds_codec;
+	const char *mapping;
+	int ret;
 
 	lvds_codec = devm_kzalloc(dev, sizeof(*lvds_codec), GFP_KERNEL);
 	if (!lvds_codec)
@@ -119,13 +159,47 @@ static int lvds_codec_probe(struct platform_device *pdev)
 	if (IS_ERR(lvds_codec->panel_bridge))
 		return PTR_ERR(lvds_codec->panel_bridge);
 
+	lvds_codec->bridge.funcs = &funcs;
+
+	/*
+	 * Decoder input LVDS format is a property of the decoder chip or even
+	 * its strapping. Handle data-mapping the same way lvds-panel does. In
+	 * case data-mapping is not present, do nothing, since there are still
+	 * legacy bindings which do not specify this property.
+	 */
+	if (lvds_codec->connector_type != DRM_MODE_CONNECTOR_LVDS) {
+		bus_node = of_graph_get_endpoint_by_regs(dev->of_node, 0, 0);
+		if (!bus_node) {
+			dev_dbg(dev, "bus DT node not found\n");
+			return -ENXIO;
+		}
+
+		ret = of_property_read_string(bus_node, "data-mapping",
+					      &mapping);
+		of_node_put(bus_node);
+		if (ret < 0) {
+			dev_warn(dev, "missing 'data-mapping' DT property\n");
+		} else {
+			if (!strcmp(mapping, "jeida-18")) {
+				lvds_codec->bus_format = MEDIA_BUS_FMT_RGB666_1X7X3_SPWG;
+			} else if (!strcmp(mapping, "jeida-24")) {
+				lvds_codec->bus_format = MEDIA_BUS_FMT_RGB888_1X7X4_JEIDA;
+			} else if (!strcmp(mapping, "vesa-24")) {
+				lvds_codec->bus_format = MEDIA_BUS_FMT_RGB888_1X7X4_SPWG;
+			} else {
+				dev_err(dev, "invalid 'data-mapping' DT property\n");
+				return -EINVAL;
+			}
+			lvds_codec->bridge.funcs = &funcs_decoder;
+		}
+	}
+
 	/*
 	 * The panel_bridge bridge is attached to the panel's of_node,
 	 * but we need a bridge attached to our of_node for our user
 	 * to look up.
 	 */
 	lvds_codec->bridge.of_node = dev->of_node;
-	lvds_codec->bridge.funcs = &funcs;
 	drm_bridge_add(&lvds_codec->bridge);
 
 	platform_set_drvdata(pdev, lvds_codec);
diff --git a/drivers/gpu/drm/bridge/nwl-dsi.c b/drivers/gpu/drm/bridge/nwl-dsi.c
index ed8ac50..a7389a0 100644
--- a/drivers/gpu/drm/bridge/nwl-dsi.c
+++ b/drivers/gpu/drm/bridge/nwl-dsi.c
@@ -939,6 +939,40 @@ static void nwl_dsi_bridge_detach(struct drm_bridge *bridge)
 	drm_of_panel_bridge_remove(dsi->dev->of_node, 1, 0);
 }
 
+static u32 *nwl_bridge_atomic_get_input_bus_fmts(struct drm_bridge *bridge,
+						 struct drm_bridge_state *bridge_state,
+						 struct drm_crtc_state *crtc_state,
+						 struct drm_connector_state *conn_state,
+						 u32 output_fmt,
+						 unsigned int *num_input_fmts)
+{
+	u32 *input_fmts, input_fmt;
+
+	*num_input_fmts = 0;
+
+	switch (output_fmt) {
+	/* If MEDIA_BUS_FMT_FIXED is tested, return default bus format */
+	case MEDIA_BUS_FMT_FIXED:
+		input_fmt = MEDIA_BUS_FMT_RGB888_1X24;
+		break;
+	case MEDIA_BUS_FMT_RGB888_1X24:
+	case MEDIA_BUS_FMT_RGB666_1X18:
+	case MEDIA_BUS_FMT_RGB565_1X16:
+		input_fmt = output_fmt;
+		break;
+	default:
+		return NULL;
+	}
+
+	input_fmts = kcalloc(1, sizeof(*input_fmts), GFP_KERNEL);
+	if (!input_fmts)
+		return NULL;
+	input_fmts[0] = input_fmt;
+	*num_input_fmts = 1;
+
+	return input_fmts;
+}
+
 static const struct drm_bridge_funcs nwl_dsi_bridge_funcs = {
 	.atomic_duplicate_state	= drm_atomic_helper_bridge_duplicate_state,
 	.atomic_destroy_state	= drm_atomic_helper_bridge_destroy_state,
@@ -946,6 +980,7 @@ static const struct drm_bridge_funcs nwl_dsi_bridge_funcs = {
 	.atomic_check		= nwl_dsi_bridge_atomic_check,
 	.atomic_enable		= nwl_dsi_bridge_atomic_enable,
 	.atomic_disable		= nwl_dsi_bridge_atomic_disable,
+	.atomic_get_input_bus_fmts = nwl_bridge_atomic_get_input_bus_fmts,
 	.mode_set		= nwl_dsi_bridge_mode_set,
 	.mode_valid		= nwl_dsi_bridge_mode_valid,
 	.attach			= nwl_dsi_bridge_attach,
diff --git a/drivers/gpu/drm/bridge/ti-sn65dsi83.c b/drivers/gpu/drm/bridge/ti-sn65dsi83.c
index a32f70b..ba1160e 100644
--- a/drivers/gpu/drm/bridge/ti-sn65dsi83.c
+++ b/drivers/gpu/drm/bridge/ti-sn65dsi83.c
@@ -288,6 +288,19 @@ static int sn65dsi83_attach(struct drm_bridge *bridge,
 	return ret;
 }
 
+static void sn65dsi83_detach(struct drm_bridge *bridge)
+{
+	struct sn65dsi83 *ctx = bridge_to_sn65dsi83(bridge);
+
+	if (!ctx->dsi)
+		return;
+
+	mipi_dsi_detach(ctx->dsi);
+	mipi_dsi_device_unregister(ctx->dsi);
+	drm_bridge_remove(&ctx->bridge);
+	ctx->dsi = NULL;
+}
+
 static void sn65dsi83_atomic_pre_enable(struct drm_bridge *bridge,
 					struct drm_bridge_state *old_bridge_state)
 {
@@ -583,6 +596,7 @@ sn65dsi83_atomic_get_input_bus_fmts(struct drm_bridge *bridge,
 
 static const struct drm_bridge_funcs sn65dsi83_funcs = {
 	.attach			= sn65dsi83_attach,
+	.detach			= sn65dsi83_detach,
 	.atomic_pre_enable	= sn65dsi83_atomic_pre_enable,
 	.atomic_enable		= sn65dsi83_atomic_enable,
 	.atomic_disable		= sn65dsi83_atomic_disable,
@@ -697,9 +711,6 @@ static int sn65dsi83_remove(struct i2c_client *client)
 {
 	struct sn65dsi83 *ctx = i2c_get_clientdata(client);
 
-	mipi_dsi_detach(ctx->dsi);
-	mipi_dsi_device_unregister(ctx->dsi);
-	drm_bridge_remove(&ctx->bridge);
 	of_node_put(ctx->host_node);
 
 	return 0;
diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c
index 3bc782b..52e20c6 100644
--- a/drivers/gpu/drm/drm_connector.c
+++ b/drivers/gpu/drm/drm_connector.c
@@ -625,6 +625,8 @@ int drm_connector_register_all(struct drm_device *dev)
  *
  * In contrast to the other drm_get_*_name functions this one here returns a
  * const pointer and hence is threadsafe.
+ *
+ * Returns: connector status string
  */
 const char *drm_get_connector_status_name(enum drm_connector_status status)
 {
@@ -707,7 +709,7 @@ __drm_connector_put_safe(struct drm_connector *conn)
  * drm_connector_list_iter_next - return next connector
  * @iter: connector_list iterator
  *
- * Returns the next connector for @iter, or NULL when the list walk has
+ * Returns: the next connector for @iter, or NULL when the list walk has
  * completed.
  */
 struct drm_connector *
@@ -780,6 +782,8 @@ static const struct drm_prop_enum_list drm_subpixel_enum_list[] = {
  *
  * Note you could abuse this and return something out of bounds, but that
  * would be a caller error.  No unscrubbed user data should make it here.
+ *
+ * Returns: string describing an enumerated subpixel property
  */
 const char *drm_get_subpixel_order_name(enum subpixel_order order)
 {
@@ -809,6 +813,9 @@ static const struct drm_prop_enum_list drm_link_status_enum_list[] = {
  * Store the supported bus formats in display info structure.
  * See MEDIA_BUS_FMT_* definitions in include/uapi/linux/media-bus-format.h for
  * a full list of available formats.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure.
  */
 int drm_display_info_set_bus_formats(struct drm_display_info *info,
 				     const u32 *formats,
@@ -1326,6 +1333,8 @@ int drm_connector_create_standard_properties(struct drm_device *dev)
  * @dev: DRM device
  *
  * Called by a driver the first time a DVI-I connector is made.
+ *
+ * Returns: %0
  */
 int drm_mode_create_dvi_i_properties(struct drm_device *dev)
 {
@@ -1397,6 +1406,8 @@ EXPORT_SYMBOL(drm_connector_attach_dp_subconnector_property);
  *	Game:
  *		Content type is game
  *
+ *	The meaning of each content type is defined in CTA-861-G table 15.
+ *
  *	Drivers can set up this property by calling
  *	drm_connector_attach_content_type_property(). Decoding to
  *	infoframe values is done through drm_hdmi_avi_infoframe_content_type().
@@ -1407,6 +1418,8 @@ EXPORT_SYMBOL(drm_connector_attach_dp_subconnector_property);
  * @connector: connector to attach content type property on.
  *
  * Called by a driver the first time a HDMI connector is made.
+ *
+ * Returns: %0
  */
 int drm_connector_attach_content_type_property(struct drm_connector *connector)
 {
@@ -1487,6 +1500,9 @@ EXPORT_SYMBOL(drm_connector_attach_tv_margin_properties);
  * creates the TV margin properties for a given device. No need to call this
  * function for an SDTV connector, it's already called from
  * drm_mode_create_tv_properties().
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure.
  */
 int drm_mode_create_tv_margin_properties(struct drm_device *dev)
 {
@@ -1527,6 +1543,9 @@ EXPORT_SYMBOL(drm_mode_create_tv_margin_properties);
  * the TV specific connector properties for a given device.  Caller is
  * responsible for allocating a list of format names and passing them to
  * this routine.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure.
  */
 int drm_mode_create_tv_properties(struct drm_device *dev,
 				  unsigned int num_modes,
@@ -1622,6 +1641,8 @@ EXPORT_SYMBOL(drm_mode_create_tv_properties);
  * Atomic drivers should use drm_connector_attach_scaling_mode_property()
  * instead to correctly assign &drm_connector_state.scaling_mode
  * in the atomic state.
+ *
+ * Returns: %0
  */
 int drm_mode_create_scaling_mode_property(struct drm_device *dev)
 {
@@ -1939,6 +1960,9 @@ EXPORT_SYMBOL(drm_mode_create_content_type_property);
  * @dev: DRM device
  *
  * Create the suggested x/y offset property for connectors.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure.
  */
 int drm_mode_create_suggested_offset_properties(struct drm_device *dev)
 {
@@ -2312,8 +2336,8 @@ int drm_connector_set_panel_orientation(
 EXPORT_SYMBOL(drm_connector_set_panel_orientation);
 
 /**
- * drm_connector_set_panel_orientation_with_quirk -
- *	set the connector's panel_orientation after checking for quirks
+ * drm_connector_set_panel_orientation_with_quirk - set the
+ *	connector's panel_orientation after checking for quirks
  * @connector: connector for which to init the panel-orientation property.
  * @panel_orientation: drm_panel_orientation value to set
  * @width: width in pixels of the panel, used for panel quirk detection
@@ -2597,7 +2621,7 @@ struct drm_connector *drm_connector_find_by_fwnode(struct fwnode_handle *fwnode)
 
 /**
  * drm_connector_oob_hotplug_event - Report out-of-band hotplug event to connector
- * @connector: connector to report the event on
+ * @connector_fwnode: fwnode_handle to report the event on
  *
  * On some hardware a hotplug event notification may come from outside the display
  * driver / device. An example of this is some USB Type-C setups where the hardware
diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index 09c8200..4dcdec6 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -1340,31 +1340,15 @@ int drm_gem_fence_array_add_implicit(struct xarray *fence_array,
 				     struct drm_gem_object *obj,
 				     bool write)
 {
-	int ret;
-	struct dma_fence **fences;
-	unsigned int i, fence_count;
+	struct dma_resv_iter cursor;
+	struct dma_fence *fence;
+	int ret = 0;
 
-	if (!write) {
-		struct dma_fence *fence =
-			dma_resv_get_excl_unlocked(obj->resv);
-
-		return drm_gem_fence_array_add(fence_array, fence);
-	}
-
-	ret = dma_resv_get_fences(obj->resv, NULL,
-						&fence_count, &fences);
-	if (ret || !fence_count)
-		return ret;
-
-	for (i = 0; i < fence_count; i++) {
-		ret = drm_gem_fence_array_add(fence_array, fences[i]);
+	dma_resv_for_each_fence(&cursor, obj->resv, write, fence) {
+		ret = drm_gem_fence_array_add(fence_array, fence);
 		if (ret)
 			break;
 	}
-
-	for (; i < fence_count; i++)
-		dma_fence_put(fences[i]);
-	kfree(fences);
 	return ret;
 }
 EXPORT_SYMBOL(drm_gem_fence_array_add_implicit);
diff --git a/drivers/gpu/drm/drm_modeset_lock.c b/drivers/gpu/drm/drm_modeset_lock.c
index bf8a6e8..c973233 100644
--- a/drivers/gpu/drm/drm_modeset_lock.c
+++ b/drivers/gpu/drm/drm_modeset_lock.c
@@ -25,6 +25,7 @@
 #include <drm/drm_crtc.h>
 #include <drm/drm_device.h>
 #include <drm/drm_modeset_lock.h>
+#include <drm/drm_print.h>
 
 /**
  * DOC: kms locking
@@ -77,6 +78,45 @@
 
 static DEFINE_WW_CLASS(crtc_ww_class);
 
+#if IS_ENABLED(CONFIG_DRM_DEBUG_MODESET_LOCK)
+static noinline depot_stack_handle_t __drm_stack_depot_save(void)
+{
+	unsigned long entries[8];
+	unsigned int n;
+
+	n = stack_trace_save(entries, ARRAY_SIZE(entries), 1);
+
+	return stack_depot_save(entries, n, GFP_NOWAIT | __GFP_NOWARN);
+}
+
+static void __drm_stack_depot_print(depot_stack_handle_t stack_depot)
+{
+	struct drm_printer p = drm_debug_printer("drm_modeset_lock");
+	unsigned long *entries;
+	unsigned int nr_entries;
+	char *buf;
+
+	buf = kmalloc(PAGE_SIZE, GFP_NOWAIT | __GFP_NOWARN);
+	if (!buf)
+		return;
+
+	nr_entries = stack_depot_fetch(stack_depot, &entries);
+	stack_trace_snprint(buf, PAGE_SIZE, entries, nr_entries, 2);
+
+	drm_printf(&p, "attempting to lock a contended lock without backoff:\n%s", buf);
+
+	kfree(buf);
+}
+#else /* CONFIG_DRM_DEBUG_MODESET_LOCK */
+static depot_stack_handle_t __drm_stack_depot_save(void)
+{
+	return 0;
+}
+static void __drm_stack_depot_print(depot_stack_handle_t stack_depot)
+{
+}
+#endif /* CONFIG_DRM_DEBUG_MODESET_LOCK */
+
 /**
  * drm_modeset_lock_all - take all modeset locks
  * @dev: DRM device
@@ -225,7 +265,9 @@ EXPORT_SYMBOL(drm_modeset_acquire_fini);
  */
 void drm_modeset_drop_locks(struct drm_modeset_acquire_ctx *ctx)
 {
-	WARN_ON(ctx->contended);
+	if (WARN_ON(ctx->contended))
+		__drm_stack_depot_print(ctx->stack_depot);
+
 	while (!list_empty(&ctx->locked)) {
 		struct drm_modeset_lock *lock;
 
@@ -243,7 +285,8 @@ static inline int modeset_lock(struct drm_modeset_lock *lock,
 {
 	int ret;
 
-	WARN_ON(ctx->contended);
+	if (WARN_ON(ctx->contended))
+		__drm_stack_depot_print(ctx->stack_depot);
 
 	if (ctx->trylock_only) {
 		lockdep_assert_held(&ctx->ww_ctx);
@@ -274,6 +317,7 @@ static inline int modeset_lock(struct drm_modeset_lock *lock,
 		ret = 0;
 	} else if (ret == -EDEADLK) {
 		ctx->contended = lock;
+		ctx->stack_depot = __drm_stack_depot_save();
 	}
 
 	return ret;
@@ -296,6 +340,7 @@ int drm_modeset_backoff(struct drm_modeset_acquire_ctx *ctx)
 	struct drm_modeset_lock *contended = ctx->contended;
 
 	ctx->contended = NULL;
+	ctx->stack_depot = 0;
 
 	if (WARN_ON(!contended))
 		return 0;
diff --git a/drivers/gpu/drm/drm_plane_helper.c b/drivers/gpu/drm/drm_plane_helper.c
index 5b2d0ca..838b32b 100644
--- a/drivers/gpu/drm/drm_plane_helper.c
+++ b/drivers/gpu/drm/drm_plane_helper.c
@@ -123,7 +123,6 @@ static int drm_plane_helper_check_update(struct drm_plane *plane,
 		.crtc_w = drm_rect_width(dst),
 		.crtc_h = drm_rect_height(dst),
 		.rotation = rotation,
-		.visible = *visible,
 	};
 	struct drm_crtc_state crtc_state = {
 		.crtc = crtc,
diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
index d8ba957..c773d3d 100644
--- a/drivers/gpu/drm/drm_prime.c
+++ b/drivers/gpu/drm/drm_prime.c
@@ -722,11 +722,13 @@ int drm_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
 	if (obj->funcs && obj->funcs->mmap) {
 		vma->vm_ops = obj->funcs->vm_ops;
 
-		ret = obj->funcs->mmap(obj, vma);
-		if (ret)
-			return ret;
-		vma->vm_private_data = obj;
 		drm_gem_object_get(obj);
+		ret = obj->funcs->mmap(obj, vma);
+		if (ret) {
+			drm_gem_object_put(obj);
+			return ret;
+		}
+		vma->vm_private_data = obj;
 		return 0;
 	}
 
diff --git a/drivers/gpu/drm/i915/display/g4x_hdmi.c b/drivers/gpu/drm/i915/display/g4x_hdmi.c
index 88c427f..f5b4dd5 100644
--- a/drivers/gpu/drm/i915/display/g4x_hdmi.c
+++ b/drivers/gpu/drm/i915/display/g4x_hdmi.c
@@ -584,6 +584,7 @@ void g4x_hdmi_init(struct drm_i915_private *dev_priv,
 		else
 			intel_encoder->enable = g4x_enable_hdmi;
 	}
+	intel_encoder->shutdown = intel_hdmi_encoder_shutdown;
 
 	intel_encoder->type = INTEL_OUTPUT_HDMI;
 	intel_encoder->power_domain = intel_port_to_power_domain(port);
diff --git a/drivers/gpu/drm/i915/display/intel_bios.c b/drivers/gpu/drm/i915/display/intel_bios.c
index b99907c6..2b1423a 100644
--- a/drivers/gpu/drm/i915/display/intel_bios.c
+++ b/drivers/gpu/drm/i915/display/intel_bios.c
@@ -1707,6 +1707,39 @@ static void sanitize_aux_ch(struct intel_bios_encoder_data *devdata,
 	child->aux_channel = 0;
 }
 
+static u8 dvo_port_type(u8 dvo_port)
+{
+	switch (dvo_port) {
+	case DVO_PORT_HDMIA:
+	case DVO_PORT_HDMIB:
+	case DVO_PORT_HDMIC:
+	case DVO_PORT_HDMID:
+	case DVO_PORT_HDMIE:
+	case DVO_PORT_HDMIF:
+	case DVO_PORT_HDMIG:
+	case DVO_PORT_HDMIH:
+	case DVO_PORT_HDMII:
+		return DVO_PORT_HDMIA;
+	case DVO_PORT_DPA:
+	case DVO_PORT_DPB:
+	case DVO_PORT_DPC:
+	case DVO_PORT_DPD:
+	case DVO_PORT_DPE:
+	case DVO_PORT_DPF:
+	case DVO_PORT_DPG:
+	case DVO_PORT_DPH:
+	case DVO_PORT_DPI:
+		return DVO_PORT_DPA;
+	case DVO_PORT_MIPIA:
+	case DVO_PORT_MIPIB:
+	case DVO_PORT_MIPIC:
+	case DVO_PORT_MIPID:
+		return DVO_PORT_MIPIA;
+	default:
+		return dvo_port;
+	}
+}
+
 static enum port __dvo_port_to_port(int n_ports, int n_dvo,
 				    const int port_mapping[][3], u8 dvo_port)
 {
@@ -1930,50 +1963,6 @@ static int _intel_bios_max_tmds_clock(const struct intel_bios_encoder_data *devd
 	}
 }
 
-static enum port get_edp_port(struct drm_i915_private *i915)
-{
-	const struct intel_bios_encoder_data *devdata;
-	enum port port;
-
-	for_each_port(port) {
-		devdata = i915->vbt.ports[port];
-
-		if (devdata && intel_bios_encoder_supports_edp(devdata))
-			return port;
-	}
-
-	return PORT_NONE;
-}
-
-/*
- * FIXME: The power sequencer and backlight code currently do not support more
- * than one set registers, at least not on anything other than VLV/CHV. It will
- * clobber the registers. As a temporary workaround, gracefully prevent more
- * than one eDP from being registered.
- */
-static void sanitize_dual_edp(struct intel_bios_encoder_data *devdata,
-			      enum port port)
-{
-	struct drm_i915_private *i915 = devdata->i915;
-	struct child_device_config *child = &devdata->child;
-	enum port p;
-
-	/* CHV might not clobber PPS registers. */
-	if (IS_CHERRYVIEW(i915))
-		return;
-
-	p = get_edp_port(i915);
-	if (p == PORT_NONE)
-		return;
-
-	drm_dbg_kms(&i915->drm, "both ports %c and %c configured as eDP, "
-		    "disabling port %c eDP\n", port_name(p), port_name(port),
-		    port_name(port));
-
-	child->device_type &= ~DEVICE_TYPE_DISPLAYPORT_OUTPUT;
-	child->device_type &= ~DEVICE_TYPE_INTERNAL_CONNECTOR;
-}
-
 static bool is_port_valid(struct drm_i915_private *i915, enum port port)
 {
 	/*
@@ -2031,9 +2020,6 @@ static void parse_ddi_port(struct drm_i915_private *i915,
 		    supports_typec_usb, supports_tbt,
 		    devdata->dsc != NULL);
 
-	if (is_edp)
-		sanitize_dual_edp(devdata, port);
-
 	if (is_dvi)
 		sanitize_ddc_pin(devdata, port);
 
@@ -2670,8 +2656,25 @@ bool intel_bios_is_port_edp(struct drm_i915_private *i915, enum port port)
 	return false;
 }
 
-static bool child_dev_is_dp_dual_mode(const struct child_device_config *child,
-				      enum port port)
+static bool child_dev_is_dp_dual_mode(const struct child_device_config *child)
+{
+	if ((child->device_type & DEVICE_TYPE_DP_DUAL_MODE_BITS) !=
+	    (DEVICE_TYPE_DP_DUAL_MODE & DEVICE_TYPE_DP_DUAL_MODE_BITS))
+		return false;
+
+	if (dvo_port_type(child->dvo_port) == DVO_PORT_DPA)
+		return true;
+
+	/* Only accept a HDMI dvo_port as DP++ if it has an AUX channel */
+	if (dvo_port_type(child->dvo_port) == DVO_PORT_HDMIA &&
+	    child->aux_channel != 0)
+		return true;
+
+	return false;
+}
+
+bool intel_bios_is_port_dp_dual_mode(struct drm_i915_private *i915,
+				     enum port port)
 {
 	static const struct {
 		u16 dp, hdmi;
@@ -2686,32 +2689,23 @@ static bool child_dev_is_dp_dual_mode(const struct child_device_config *child,
 		[PORT_E] = { DVO_PORT_DPE, DVO_PORT_HDMIE, },
 		[PORT_F] = { DVO_PORT_DPF, DVO_PORT_HDMIF, },
 	};
+	const struct intel_bios_encoder_data *devdata;
+
+	if (HAS_DDI(i915)) {
+		const struct intel_bios_encoder_data *devdata;
+
+		devdata = intel_bios_encoder_data_lookup(i915, port);
+
+		return devdata && child_dev_is_dp_dual_mode(&devdata->child);
+	}
 
 	if (port == PORT_A || port >= ARRAY_SIZE(port_mapping))
 		return false;
 
-	if ((child->device_type & DEVICE_TYPE_DP_DUAL_MODE_BITS) !=
-	    (DEVICE_TYPE_DP_DUAL_MODE & DEVICE_TYPE_DP_DUAL_MODE_BITS))
-		return false;
-
-	if (child->dvo_port == port_mapping[port].dp)
-		return true;
-
-	/* Only accept a HDMI dvo_port as DP++ if it has an AUX channel */
-	if (child->dvo_port == port_mapping[port].hdmi &&
-	    child->aux_channel != 0)
-		return true;
-
-	return false;
-}
-
-bool intel_bios_is_port_dp_dual_mode(struct drm_i915_private *i915,
-				     enum port port)
-{
-	const struct intel_bios_encoder_data *devdata;
-
 	list_for_each_entry(devdata, &i915->vbt.display_devices, node) {
-		if (child_dev_is_dp_dual_mode(&devdata->child, port))
+		if ((devdata->child.dvo_port == port_mapping[port].dp ||
+		     devdata->child.dvo_port == port_mapping[port].hdmi) &&
+		    child_dev_is_dp_dual_mode(&devdata->child))
 			return true;
 	}
 
diff --git a/drivers/gpu/drm/i915/display/intel_cdclk.c b/drivers/gpu/drm/i915/display/intel_cdclk.c
index 9e466d8..868dd43 100644
--- a/drivers/gpu/drm/i915/display/intel_cdclk.c
+++ b/drivers/gpu/drm/i915/display/intel_cdclk.c
@@ -2885,7 +2885,7 @@ u32 intel_read_rawclk(struct drm_i915_private *dev_priv)
 	return freq;
 }
 
-static struct intel_cdclk_funcs tgl_cdclk_funcs = {
+static const struct intel_cdclk_funcs tgl_cdclk_funcs = {
 	.get_cdclk = bxt_get_cdclk,
 	.set_cdclk = bxt_set_cdclk,
 	.bw_calc_min_cdclk = skl_bw_calc_min_cdclk,
@@ -2893,7 +2893,7 @@ static struct intel_cdclk_funcs tgl_cdclk_funcs = {
 	.calc_voltage_level = tgl_calc_voltage_level,
 };
 
-static struct intel_cdclk_funcs ehl_cdclk_funcs = {
+static const struct intel_cdclk_funcs ehl_cdclk_funcs = {
 	.get_cdclk = bxt_get_cdclk,
 	.set_cdclk = bxt_set_cdclk,
 	.bw_calc_min_cdclk = skl_bw_calc_min_cdclk,
@@ -2901,7 +2901,7 @@ static struct intel_cdclk_funcs ehl_cdclk_funcs = {
 	.calc_voltage_level = ehl_calc_voltage_level,
 };
 
-static struct intel_cdclk_funcs icl_cdclk_funcs = {
+static const struct intel_cdclk_funcs icl_cdclk_funcs = {
 	.get_cdclk = bxt_get_cdclk,
 	.set_cdclk = bxt_set_cdclk,
 	.bw_calc_min_cdclk = skl_bw_calc_min_cdclk,
@@ -2909,7 +2909,7 @@ static struct intel_cdclk_funcs icl_cdclk_funcs = {
 	.calc_voltage_level = icl_calc_voltage_level,
 };
 
-static struct intel_cdclk_funcs bxt_cdclk_funcs = {
+static const struct intel_cdclk_funcs bxt_cdclk_funcs = {
 	.get_cdclk = bxt_get_cdclk,
 	.set_cdclk = bxt_set_cdclk,
 	.bw_calc_min_cdclk = skl_bw_calc_min_cdclk,
@@ -2917,54 +2917,54 @@ static struct intel_cdclk_funcs bxt_cdclk_funcs = {
 	.calc_voltage_level = bxt_calc_voltage_level,
 };
 
-static struct intel_cdclk_funcs skl_cdclk_funcs = {
+static const struct intel_cdclk_funcs skl_cdclk_funcs = {
 	.get_cdclk = skl_get_cdclk,
 	.set_cdclk = skl_set_cdclk,
 	.bw_calc_min_cdclk = skl_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = skl_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs bdw_cdclk_funcs = {
+static const struct intel_cdclk_funcs bdw_cdclk_funcs = {
 	.get_cdclk = bdw_get_cdclk,
 	.set_cdclk = bdw_set_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = bdw_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs chv_cdclk_funcs = {
+static const struct intel_cdclk_funcs chv_cdclk_funcs = {
 	.get_cdclk = vlv_get_cdclk,
 	.set_cdclk = chv_set_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = vlv_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs vlv_cdclk_funcs = {
+static const struct intel_cdclk_funcs vlv_cdclk_funcs = {
 	.get_cdclk = vlv_get_cdclk,
 	.set_cdclk = vlv_set_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = vlv_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs hsw_cdclk_funcs = {
+static const struct intel_cdclk_funcs hsw_cdclk_funcs = {
 	.get_cdclk = hsw_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
 };
 
 /* SNB, IVB, 965G, 945G */
-static struct intel_cdclk_funcs fixed_400mhz_cdclk_funcs = {
+static const struct intel_cdclk_funcs fixed_400mhz_cdclk_funcs = {
 	.get_cdclk = fixed_400mhz_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs ilk_cdclk_funcs = {
+static const struct intel_cdclk_funcs ilk_cdclk_funcs = {
 	.get_cdclk = fixed_450mhz_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs gm45_cdclk_funcs = {
+static const struct intel_cdclk_funcs gm45_cdclk_funcs = {
 	.get_cdclk = gm45_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
@@ -2972,7 +2972,7 @@ static struct intel_cdclk_funcs gm45_cdclk_funcs = {
 
 /* G45 uses G33 */
 
-static struct intel_cdclk_funcs i965gm_cdclk_funcs = {
+static const struct intel_cdclk_funcs i965gm_cdclk_funcs = {
 	.get_cdclk = i965gm_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
@@ -2980,19 +2980,19 @@ static struct intel_cdclk_funcs i965gm_cdclk_funcs = {
 
 /* i965G uses fixed 400 */
 
-static struct intel_cdclk_funcs pnv_cdclk_funcs = {
+static const struct intel_cdclk_funcs pnv_cdclk_funcs = {
 	.get_cdclk = pnv_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs g33_cdclk_funcs = {
+static const struct intel_cdclk_funcs g33_cdclk_funcs = {
 	.get_cdclk = g33_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs i945gm_cdclk_funcs = {
+static const struct intel_cdclk_funcs i945gm_cdclk_funcs = {
 	.get_cdclk = i945gm_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
@@ -3000,37 +3000,37 @@ static struct intel_cdclk_funcs i945gm_cdclk_funcs = {
 
 /* i945G uses fixed 400 */
 
-static struct intel_cdclk_funcs i915gm_cdclk_funcs = {
+static const struct intel_cdclk_funcs i915gm_cdclk_funcs = {
 	.get_cdclk = i915gm_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs i915g_cdclk_funcs = {
+static const struct intel_cdclk_funcs i915g_cdclk_funcs = {
 	.get_cdclk = fixed_333mhz_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs i865g_cdclk_funcs = {
+static const struct intel_cdclk_funcs i865g_cdclk_funcs = {
 	.get_cdclk = fixed_266mhz_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs i85x_cdclk_funcs = {
+static const struct intel_cdclk_funcs i85x_cdclk_funcs = {
 	.get_cdclk = i85x_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs i845g_cdclk_funcs = {
+static const struct intel_cdclk_funcs i845g_cdclk_funcs = {
 	.get_cdclk = fixed_200mhz_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
 };
 
-static struct intel_cdclk_funcs i830_cdclk_funcs = {
+static const struct intel_cdclk_funcs i830_cdclk_funcs = {
 	.get_cdclk = fixed_133mhz_get_cdclk,
 	.bw_calc_min_cdclk = intel_bw_calc_min_cdclk,
 	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c
index 1dcfe31..cfb567d 100644
--- a/drivers/gpu/drm/i915/display/intel_ddi.c
+++ b/drivers/gpu/drm/i915/display/intel_ddi.c
@@ -4361,6 +4361,7 @@ static void intel_ddi_encoder_shutdown(struct intel_encoder *encoder)
 	enum phy phy = intel_port_to_phy(i915, encoder->port);
 
 	intel_dp_encoder_shutdown(encoder);
+	intel_hdmi_encoder_shutdown(encoder);
 
 	if (!intel_phy_is_tc(i915, phy))
 		return;
diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
index ff598b6..ec403e4 100644
--- a/drivers/gpu/drm/i915/display/intel_display.c
+++ b/drivers/gpu/drm/i915/display/intel_display.c
@@ -848,9 +848,16 @@ unsigned int intel_remapped_info_size(const struct intel_remapped_info *rem_info
 	int i;
 
 	for (i = 0 ; i < ARRAY_SIZE(rem_info->plane); i++) {
+		unsigned int plane_size;
+
+		plane_size = rem_info->plane[i].dst_stride * rem_info->plane[i].height;
+		if (plane_size == 0)
+			continue;
+
 		if (rem_info->plane_alignment)
 			size = ALIGN(size, rem_info->plane_alignment);
-		size += rem_info->plane[i].dst_stride * rem_info->plane[i].height;
+
+		size += plane_size;
 	}
 
 	return size;
diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
index 23de500..be88346 100644
--- a/drivers/gpu/drm/i915/display/intel_dp.c
+++ b/drivers/gpu/drm/i915/display/intel_dp.c
@@ -120,6 +120,12 @@ bool intel_dp_is_uhbr(const struct intel_crtc_state *crtc_state)
 	return crtc_state->port_clock >= 1000000;
 }
 
+static void intel_dp_set_default_sink_rates(struct intel_dp *intel_dp)
+{
+	intel_dp->sink_rates[0] = 162000;
+	intel_dp->num_sink_rates = 1;
+}
+
 /* update sink rates from dpcd */
 static void intel_dp_set_sink_rates(struct intel_dp *intel_dp)
 {
@@ -281,7 +287,7 @@ intel_dp_max_data_rate(int max_link_rate, int max_lanes)
 		 */
 		int max_link_rate_kbps = max_link_rate * 10;
 
-		max_link_rate_kbps = DIV_ROUND_CLOSEST_ULL(max_link_rate_kbps * 9671, 10000);
+		max_link_rate_kbps = DIV_ROUND_CLOSEST_ULL(mul_u32_u32(max_link_rate_kbps, 9671), 10000);
 		max_link_rate = max_link_rate_kbps / 8;
 	}
 
@@ -1858,6 +1864,12 @@ void intel_dp_set_link_params(struct intel_dp *intel_dp,
 	intel_dp->lane_count = lane_count;
 }
 
+static void intel_dp_reset_max_link_params(struct intel_dp *intel_dp)
+{
+	intel_dp->max_link_lane_count = intel_dp_max_common_lane_count(intel_dp);
+	intel_dp->max_link_rate = intel_dp_max_common_rate(intel_dp);
+}
+
 /* Enable backlight PWM and backlight PP control. */
 void intel_edp_backlight_on(const struct intel_crtc_state *crtc_state,
 			    const struct drm_connector_state *conn_state)
@@ -2017,8 +2029,7 @@ void intel_dp_sync_state(struct intel_encoder *encoder,
 	if (intel_dp->dpcd[DP_DPCD_REV] == 0)
 		intel_dp_get_dpcd(intel_dp);
 
-	intel_dp->max_link_lane_count = intel_dp_max_common_lane_count(intel_dp);
-	intel_dp->max_link_rate = intel_dp_max_common_rate(intel_dp);
+	intel_dp_reset_max_link_params(intel_dp);
 }
 
 bool intel_dp_initial_fastset_check(struct intel_encoder *encoder,
@@ -2556,6 +2567,9 @@ intel_edp_init_dpcd(struct intel_dp *intel_dp)
 	 */
 	intel_psr_init_dpcd(intel_dp);
 
+	/* Clear the default sink rates */
+	intel_dp->num_sink_rates = 0;
+
 	/* Read the eDP 1.4+ supported link rates. */
 	if (intel_dp->edp_dpcd[0] >= DP_EDP_14) {
 		__le16 sink_rates[DP_MAX_SUPPORTED_RATES];
@@ -2591,6 +2605,7 @@ intel_edp_init_dpcd(struct intel_dp *intel_dp)
 		intel_dp_set_sink_rates(intel_dp);
 
 	intel_dp_set_common_rates(intel_dp);
+	intel_dp_reset_max_link_params(intel_dp);
 
 	/* Read the eDP DSC DPCD registers */
 	if (DISPLAY_VER(dev_priv) >= 10)
@@ -4332,12 +4347,7 @@ intel_dp_detect(struct drm_connector *connector,
 	 * supports link training fallback params.
 	 */
 	if (intel_dp->reset_link_params || intel_dp->is_mst) {
-		/* Initial max link lane count */
-		intel_dp->max_link_lane_count = intel_dp_max_common_lane_count(intel_dp);
-
-		/* Initial max link rate */
-		intel_dp->max_link_rate = intel_dp_max_common_rate(intel_dp);
-
+		intel_dp_reset_max_link_params(intel_dp);
 		intel_dp->reset_link_params = false;
 	}
 
@@ -5003,6 +5013,9 @@ intel_dp_init_connector(struct intel_digital_port *dig_port,
 	}
 
 	intel_dp_set_source_rates(intel_dp);
+	intel_dp_set_default_sink_rates(intel_dp);
+	intel_dp_set_common_rates(intel_dp);
+	intel_dp_reset_max_link_params(intel_dp);
 
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
 		intel_dp->pps.active_pipe = vlv_active_pipe(intel_dp);
diff --git a/drivers/gpu/drm/i915/display/intel_fb.c b/drivers/gpu/drm/i915/display/intel_fb.c
index fa1f375..cb511b2 100644
--- a/drivers/gpu/drm/i915/display/intel_fb.c
+++ b/drivers/gpu/drm/i915/display/intel_fb.c
@@ -378,8 +378,8 @@ static void intel_fb_plane_dims(const struct intel_framebuffer *fb, int color_pl
 	intel_fb_plane_get_subsampling(&main_hsub, &main_vsub, &fb->base, main_plane);
 	intel_fb_plane_get_subsampling(&hsub, &vsub, &fb->base, color_plane);
 
-	*w = main_width / main_hsub / hsub;
-	*h = main_height / main_vsub / vsub;
+	*w = DIV_ROUND_UP(main_width, main_hsub * hsub);
+	*h = DIV_ROUND_UP(main_height, main_vsub * vsub);
 }
 
 static u32 intel_adjust_tile_offset(int *x, int *y,
diff --git a/drivers/gpu/drm/i915/display/intel_hdmi.c b/drivers/gpu/drm/i915/display/intel_hdmi.c
index d2e61f6c..371736b 100644
--- a/drivers/gpu/drm/i915/display/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/display/intel_hdmi.c
@@ -1246,12 +1246,13 @@ static void hsw_set_infoframes(struct intel_encoder *encoder,
 void intel_dp_dual_mode_set_tmds_output(struct intel_hdmi *hdmi, bool enable)
 {
 	struct drm_i915_private *dev_priv = intel_hdmi_to_i915(hdmi);
-	struct i2c_adapter *adapter =
-		intel_gmbus_get_adapter(dev_priv, hdmi->ddc_bus);
+	struct i2c_adapter *adapter;
 
 	if (hdmi->dp_dual_mode.type < DRM_DP_DUAL_MODE_TYPE2_DVI)
 		return;
 
+	adapter = intel_gmbus_get_adapter(dev_priv, hdmi->ddc_bus);
+
 	drm_dbg_kms(&dev_priv->drm, "%s DP dual mode adaptor TMDS output\n",
 		    enable ? "Enabling" : "Disabling");
 
@@ -2258,6 +2259,17 @@ int intel_hdmi_compute_config(struct intel_encoder *encoder,
 	return 0;
 }
 
+void intel_hdmi_encoder_shutdown(struct intel_encoder *encoder)
+{
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+
+	/*
+	 * Give a hand to buggy BIOSen which forget to turn
+	 * the TMDS output buffers back on after a reboot.
+	 */
+	intel_dp_dual_mode_set_tmds_output(intel_hdmi, true);
+}
+
 static void
 intel_hdmi_unset_edid(struct drm_connector *connector)
 {
diff --git a/drivers/gpu/drm/i915/display/intel_hdmi.h b/drivers/gpu/drm/i915/display/intel_hdmi.h
index b43a180..2bf440e 100644
--- a/drivers/gpu/drm/i915/display/intel_hdmi.h
+++ b/drivers/gpu/drm/i915/display/intel_hdmi.h
@@ -28,6 +28,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *dig_port,
 int intel_hdmi_compute_config(struct intel_encoder *encoder,
 			      struct intel_crtc_state *pipe_config,
 			      struct drm_connector_state *conn_state);
+void intel_hdmi_encoder_shutdown(struct intel_encoder *encoder);
 bool intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder,
 				       struct drm_connector *connector,
 				       bool high_tmds_clock_ratio,
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c
index eae09d3..e8a58c9 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c
@@ -9,6 +9,8 @@
 #include <linux/dma-resv.h>
 #include <linux/module.h>
 
+#include <asm/smp.h>
+
 #include "i915_drv.h"
 #include "i915_gem_object.h"
 #include "i915_scatterlist.h"
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt.c b/drivers/gpu/drm/i915/gt/intel_ggtt.c
index f17383e..57c9755 100644
--- a/drivers/gpu/drm/i915/gt/intel_ggtt.c
+++ b/drivers/gpu/drm/i915/gt/intel_ggtt.c
@@ -1396,6 +1396,9 @@ remap_pages(struct drm_i915_gem_object *obj,
 {
 	unsigned int row;
 
+	if (!width || !height)
+		return sg;
+
 	if (alignment_pad) {
 		st->nents++;
 
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
index d7710de..38b47e7 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
@@ -2373,6 +2373,7 @@ static inline void guc_lrc_desc_unpin(struct intel_context *ce)
 	unsigned long flags;
 	bool disabled;
 
+	lockdep_assert_held(&guc->submission_state.lock);
 	GEM_BUG_ON(!intel_gt_pm_is_awake(gt));
 	GEM_BUG_ON(!lrc_desc_registered(guc, ce->guc_id.id));
 	GEM_BUG_ON(ce != __get_context(guc, ce->guc_id.id));
@@ -2388,7 +2389,7 @@ static inline void guc_lrc_desc_unpin(struct intel_context *ce)
 	}
 	spin_unlock_irqrestore(&ce->guc_state.lock, flags);
 	if (unlikely(disabled)) {
-		release_guc_id(guc, ce);
+		__release_guc_id(guc, ce);
 		__guc_context_destroy(ce);
 		return;
 	}
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 2c3cd6e6..820a1f3 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -1537,38 +1537,14 @@ i915_request_await_object(struct i915_request *to,
 			  struct drm_i915_gem_object *obj,
 			  bool write)
 {
-	struct dma_fence *excl;
+	struct dma_resv_iter cursor;
+	struct dma_fence *fence;
 	int ret = 0;
 
-	if (write) {
-		struct dma_fence **shared;
-		unsigned int count, i;
-
-		ret = dma_resv_get_fences(obj->base.resv, &excl, &count,
-					  &shared);
+	dma_resv_for_each_fence(&cursor, obj->base.resv, write, fence) {
+		ret = i915_request_await_dma_fence(to, fence);
 		if (ret)
-			return ret;
-
-		for (i = 0; i < count; i++) {
-			ret = i915_request_await_dma_fence(to, shared[i]);
-			if (ret)
-				break;
-
-			dma_fence_put(shared[i]);
-		}
-
-		for (; i < count; i++)
-			dma_fence_put(shared[i]);
-		kfree(shared);
-	} else {
-		excl = dma_resv_get_excl_unlocked(obj->base.resv);
-	}
-
-	if (excl) {
-		if (ret == 0)
-			ret = i915_request_await_dma_fence(to, excl);
-
-		dma_fence_put(excl);
+			break;
 	}
 
 	return ret;
diff --git a/drivers/gpu/drm/imx/imx-drm-core.c b/drivers/gpu/drm/imx/imx-drm-core.c
index 9558e9e..cb685fe 100644
--- a/drivers/gpu/drm/imx/imx-drm-core.c
+++ b/drivers/gpu/drm/imx/imx-drm-core.c
@@ -81,7 +81,6 @@ static void imx_drm_atomic_commit_tail(struct drm_atomic_state *state)
 	struct drm_plane_state *old_plane_state, *new_plane_state;
 	bool plane_disabling = false;
 	int i;
-	bool fence_cookie = dma_fence_begin_signalling();
 
 	drm_atomic_helper_commit_modeset_disables(dev, state);
 
@@ -112,7 +111,6 @@ static void imx_drm_atomic_commit_tail(struct drm_atomic_state *state)
 	}
 
 	drm_atomic_helper_commit_hw_done(state);
-	dma_fence_end_signalling(fence_cookie);
 }
 
 static const struct drm_mode_config_helper_funcs imx_drm_mode_config_helpers = {
diff --git a/drivers/gpu/drm/mxsfb/mxsfb_kms.c b/drivers/gpu/drm/mxsfb/mxsfb_kms.c
index 89dd618..0655582 100644
--- a/drivers/gpu/drm/mxsfb/mxsfb_kms.c
+++ b/drivers/gpu/drm/mxsfb/mxsfb_kms.c
@@ -88,7 +88,7 @@ static void mxsfb_set_formats(struct mxsfb_drm_private *mxsfb,
 		ctrl |= CTRL_BUS_WIDTH_24;
 		break;
 	default:
-		dev_err(drm->dev, "Unknown media bus format %d\n", bus_format);
+		dev_err(drm->dev, "Unknown media bus format 0x%x\n", bus_format);
 		break;
 	}
 
@@ -362,6 +362,12 @@ static void mxsfb_crtc_atomic_enable(struct drm_crtc *crtc,
 			drm_atomic_get_new_bridge_state(state,
 							mxsfb->bridge);
 		bus_format = bridge_state->input_bus_cfg.format;
+		if (bus_format == MEDIA_BUS_FMT_FIXED) {
+			dev_warn_once(drm->dev,
+				      "Bridge does not provide bus format, assuming MEDIA_BUS_FMT_RGB888_1X24.\n"
+				      "Please fix bridge driver by handling atomic_get_input_bus_fmts.\n");
+			bus_format = MEDIA_BUS_FMT_RGB888_1X24;
+		}
 	}
 
 	/* If there is no bridge, use bus format from connector */
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 12b107a..fa73fe5 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1249,7 +1249,6 @@ nouveau_ttm_tt_populate(struct ttm_device *bdev,
 {
 	struct ttm_tt *ttm_dma = (void *)ttm;
 	struct nouveau_drm *drm;
-	struct device *dev;
 	bool slave = !!(ttm->page_flags & TTM_TT_FLAG_EXTERNAL);
 
 	if (ttm_tt_is_populated(ttm))
@@ -1262,7 +1261,6 @@ nouveau_ttm_tt_populate(struct ttm_device *bdev,
 	}
 
 	drm = nouveau_bdev(bdev);
-	dev = drm->dev->dev;
 
 	return ttm_pool_alloc(&drm->ttm.bdev.pool, ttm, ctx);
 }
@@ -1272,7 +1270,6 @@ nouveau_ttm_tt_unpopulate(struct ttm_device *bdev,
 			  struct ttm_tt *ttm)
 {
 	struct nouveau_drm *drm;
-	struct device *dev;
 	bool slave = !!(ttm->page_flags & TTM_TT_FLAG_EXTERNAL);
 
 	if (slave)
@@ -1281,7 +1278,6 @@ nouveau_ttm_tt_unpopulate(struct ttm_device *bdev,
 	nouveau_ttm_tt_unbind(bdev, ttm);
 
 	drm = nouveau_bdev(bdev);
-	dev = drm->dev->dev;
 
 	return ttm_pool_free(&drm->ttm.bdev.pool, ttm);
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 6109cd9..e7efd9e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -562,6 +562,7 @@ nouveau_drm_device_init(struct drm_device *dev)
 		nvkm_dbgopt(nouveau_debug, "DRM");
 
 	INIT_LIST_HEAD(&drm->clients);
+	mutex_init(&drm->clients_lock);
 	spin_lock_init(&drm->tile.lock);
 
 	/* workaround an odd issue on nvc1 by disabling the device's
@@ -632,6 +633,7 @@ nouveau_drm_device_init(struct drm_device *dev)
 static void
 nouveau_drm_device_fini(struct drm_device *dev)
 {
+	struct nouveau_cli *cli, *temp_cli;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 
 	if (nouveau_pmops_runtime()) {
@@ -656,9 +658,28 @@ nouveau_drm_device_fini(struct drm_device *dev)
 	nouveau_ttm_fini(drm);
 	nouveau_vga_fini(drm);
 
+	/*
+	 * There may be existing clients from as-yet unclosed files. For now,
+	 * clean them up here rather than deferring until the file is closed,
+	 * but this likely not correct if we want to support hot-unplugging
+	 * properly.
+	 */
+	mutex_lock(&drm->clients_lock);
+	list_for_each_entry_safe(cli, temp_cli, &drm->clients, head) {
+		list_del(&cli->head);
+		mutex_lock(&cli->mutex);
+		if (cli->abi16)
+			nouveau_abi16_fini(cli->abi16);
+		mutex_unlock(&cli->mutex);
+		nouveau_cli_fini(cli);
+		kfree(cli);
+	}
+	mutex_unlock(&drm->clients_lock);
+
 	nouveau_cli_fini(&drm->client);
 	nouveau_cli_fini(&drm->master);
 	nvif_parent_dtor(&drm->parent);
+	mutex_destroy(&drm->clients_lock);
 	kfree(drm);
 }
 
@@ -796,7 +817,7 @@ nouveau_drm_device_remove(struct drm_device *dev)
 	struct nvkm_client *client;
 	struct nvkm_device *device;
 
-	drm_dev_unregister(dev);
+	drm_dev_unplug(dev);
 
 	client = nvxx_client(&drm->client.base);
 	device = nvkm_device_find(client->device);
@@ -1090,9 +1111,9 @@ nouveau_drm_open(struct drm_device *dev, struct drm_file *fpriv)
 
 	fpriv->driver_priv = cli;
 
-	mutex_lock(&drm->client.mutex);
+	mutex_lock(&drm->clients_lock);
 	list_add(&cli->head, &drm->clients);
-	mutex_unlock(&drm->client.mutex);
+	mutex_unlock(&drm->clients_lock);
 
 done:
 	if (ret && cli) {
@@ -1110,6 +1131,16 @@ nouveau_drm_postclose(struct drm_device *dev, struct drm_file *fpriv)
 {
 	struct nouveau_cli *cli = nouveau_cli(fpriv);
 	struct nouveau_drm *drm = nouveau_drm(dev);
+	int dev_index;
+
+	/*
+	 * The device is gone, and as it currently stands all clients are
+	 * cleaned up in the removal codepath. In the future this may change
+	 * so that we can support hot-unplugging, but for now we immediately
+	 * return to avoid a double-free situation.
+	 */
+	if (!drm_dev_enter(dev, &dev_index))
+		return;
 
 	pm_runtime_get_sync(dev->dev);
 
@@ -1118,14 +1149,15 @@ nouveau_drm_postclose(struct drm_device *dev, struct drm_file *fpriv)
 		nouveau_abi16_fini(cli->abi16);
 	mutex_unlock(&cli->mutex);
 
-	mutex_lock(&drm->client.mutex);
+	mutex_lock(&drm->clients_lock);
 	list_del(&cli->head);
-	mutex_unlock(&drm->client.mutex);
+	mutex_unlock(&drm->clients_lock);
 
 	nouveau_cli_fini(cli);
 	kfree(cli);
 	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
+	drm_dev_exit(dev_index);
 }
 
 static const struct drm_ioctl_desc
diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h
index ba65f13..b2a970a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drv.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drv.h
@@ -139,6 +139,11 @@ struct nouveau_drm {
 
 	struct list_head clients;
 
+	/**
+	 * @clients_lock: Protects access to the @clients list of &struct nouveau_cli.
+	 */
+	struct mutex clients_lock;
+
 	u8 old_pm_cap;
 
 	struct {
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 8c2ecc2..9416bee 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -56,7 +56,7 @@ static vm_fault_t nouveau_ttm_fault(struct vm_fault *vmf)
 
 	nouveau_bo_del_io_reserve_lru(bo);
 	prot = vm_get_page_prot(vma->vm_flags);
-	ret = ttm_bo_vm_fault_reserved(vmf, prot, TTM_BO_VM_NUM_PREFAULT, 1);
+	ret = ttm_bo_vm_fault_reserved(vmf, prot, TTM_BO_VM_NUM_PREFAULT);
 	nouveau_bo_add_io_reserve_lru(bo);
 	if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
 		return ret;
@@ -337,7 +337,7 @@ nouveau_gem_set_domain(struct drm_gem_object *gem, uint32_t read_domains,
 	struct ttm_buffer_object *bo = &nvbo->bo;
 	uint32_t domains = valid_domains & nvbo->valid_domains &
 		(write_domains ? write_domains : read_domains);
-	uint32_t pref_domains = 0;;
+	uint32_t pref_domains = 0;
 
 	if (!domains)
 		return -EINVAL;
diff --git a/drivers/gpu/drm/nouveau/nouveau_svm.c b/drivers/gpu/drm/nouveau/nouveau_svm.c
index 1a896a2..266809e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_svm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_svm.c
@@ -162,10 +162,14 @@ nouveau_svmm_bind(struct drm_device *dev, void *data,
 	 */
 
 	mm = get_task_mm(current);
+	if (!mm) {
+		return -EINVAL;
+	}
 	mmap_read_lock(mm);
 
 	if (!cli->svm.svmm) {
 		mmap_read_unlock(mm);
+		mmput(mm);
 		return -EINVAL;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gt215.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gt215.c
index 704df0f..09a112a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gt215.c
@@ -78,6 +78,6 @@ int
 gt215_ce_new(struct nvkm_device *device, enum nvkm_subdev_type type, int inst,
 	     struct nvkm_engine **pengine)
 {
-	return nvkm_falcon_new_(&gt215_ce, device, type, inst,
+	return nvkm_falcon_new_(&gt215_ce, device, type, -1,
 				(device->chipset != 0xaf), 0x104000, pengine);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c
index ca75c5f..b51d690 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c
@@ -3147,8 +3147,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 	WARN_ON(device->chip->ptr.inst & ~((1 << ARRAY_SIZE(device->ptr)) - 1));             \
 	for (j = 0; device->chip->ptr.inst && j < ARRAY_SIZE(device->ptr); j++) {            \
 		if ((device->chip->ptr.inst & BIT(j)) && (subdev_mask & BIT_ULL(type))) {    \
-			int inst = (device->chip->ptr.inst == 1) ? -1 : (j);                 \
-			ret = device->chip->ptr.ctor(device, (type), inst, &device->ptr[j]); \
+			ret = device->chip->ptr.ctor(device, (type), (j), &device->ptr[j]);  \
 			subdev = nvkm_device_subdev(device, (type), (j));                    \
 			if (ret) {                                                           \
 				nvkm_subdev_del(&subdev);                                    \
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/base.c
index c39e797..cf5dcfd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/nvenc/base.c
@@ -21,7 +21,6 @@
  */
 #include "priv.h"
 
-#include "priv.h"
 #include <core/firmware.h>
 
 static void *
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/uvmm.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/uvmm.c
index d6a1f8d..186b4e6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/uvmm.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/uvmm.c
@@ -299,7 +299,7 @@ nvkm_uvmm_mthd_page(struct nvkm_uvmm *uvmm, void *argv, u32 argc)
 	page = uvmm->vmm->func->page;
 	for (nr = 0; page[nr].shift; nr++);
 
-	if (!(ret = nvif_unpack(ret, &argv, &argc, args->v0, 0, 0, false))) {
+	if (!(nvif_unpack(ret, &argv, &argc, args->v0, 0, 0, false))) {
 		if ((index = args->v0.index) >= nr)
 			return -EINVAL;
 		type = page[index].type;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c
index b5e7337..17899fc 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c
@@ -488,7 +488,7 @@ gp100_vmm_fault_cancel(struct nvkm_vmm *vmm, void *argv, u32 argc)
 		struct gp100_vmm_fault_cancel_v0 v0;
 	} *args = argv;
 	int ret = -ENOSYS;
-	u32 inst, aper;
+	u32 aper;
 
 	if ((ret = nvif_unpack(ret, &argv, &argc, args->v0, 0, 0, false)))
 		return ret;
@@ -502,7 +502,7 @@ gp100_vmm_fault_cancel(struct nvkm_vmm *vmm, void *argv, u32 argc)
 	args->v0.inst |= 0x80000000;
 
 	if (!WARN_ON(nvkm_gr_ctxsw_pause(device))) {
-		if ((inst = nvkm_gr_ctxsw_inst(device)) == args->v0.inst) {
+		if (nvkm_gr_ctxsw_inst(device) == args->v0.inst) {
 			gf100_vmm_invalidate(vmm, 0x0000001b
 					     /* CANCEL_TARGETED. */ |
 					     (args->v0.hub    << 20) |
diff --git a/drivers/gpu/drm/panel/Kconfig b/drivers/gpu/drm/panel/Kconfig
index 2cb8eba..cfc8d64 100644
--- a/drivers/gpu/drm/panel/Kconfig
+++ b/drivers/gpu/drm/panel/Kconfig
@@ -520,6 +520,16 @@
 	  Say Y here if you want to enable support for Sharp LS043T1LE01 qHD
 	  (540x960) DSI panel as found on the Qualcomm APQ8074 Dragonboard
 
+config DRM_PANEL_SHARP_LS060T1SX01
+	tristate "Sharp LS060T1SX01 FullHD video mode panel"
+	depends on OF
+	depends on DRM_MIPI_DSI
+	depends on BACKLIGHT_CLASS_DEVICE
+	help
+	  Say Y here if you want to enable support for Sharp LS060T1SX01 6.0"
+	  FullHD (1080x1920) DSI panel as found in Dragonboard Display Adapter
+	  Bundle.
+
 config DRM_PANEL_SITRONIX_ST7701
 	tristate "Sitronix ST7701 panel driver"
 	depends on OF
diff --git a/drivers/gpu/drm/panel/Makefile b/drivers/gpu/drm/panel/Makefile
index 6e30640..bca4cc1 100644
--- a/drivers/gpu/drm/panel/Makefile
+++ b/drivers/gpu/drm/panel/Makefile
@@ -53,6 +53,7 @@
 obj-$(CONFIG_DRM_PANEL_SHARP_LQ101R1SX01) += panel-sharp-lq101r1sx01.o
 obj-$(CONFIG_DRM_PANEL_SHARP_LS037V7DW01) += panel-sharp-ls037v7dw01.o
 obj-$(CONFIG_DRM_PANEL_SHARP_LS043T1LE01) += panel-sharp-ls043t1le01.o
+obj-$(CONFIG_DRM_PANEL_SHARP_LS060T1SX01) += panel-sharp-ls060t1sx01.o
 obj-$(CONFIG_DRM_PANEL_SITRONIX_ST7701) += panel-sitronix-st7701.o
 obj-$(CONFIG_DRM_PANEL_SITRONIX_ST7703) += panel-sitronix-st7703.o
 obj-$(CONFIG_DRM_PANEL_SITRONIX_ST7789V) += panel-sitronix-st7789v.o
diff --git a/drivers/gpu/drm/panel/panel-mantix-mlaf057we51.c b/drivers/gpu/drm/panel/panel-mantix-mlaf057we51.c
index 30f28ad..31daae1d 100644
--- a/drivers/gpu/drm/panel/panel-mantix-mlaf057we51.c
+++ b/drivers/gpu/drm/panel/panel-mantix-mlaf057we51.c
@@ -8,6 +8,7 @@
 #include <linux/backlight.h>
 #include <linux/delay.h>
 #include <linux/gpio/consumer.h>
+#include <linux/media-bus-format.h>
 #include <linux/module.h>
 #include <linux/of_device.h>
 #include <linux/regulator/consumer.h>
@@ -220,6 +221,10 @@ static const struct drm_display_mode default_mode_ys = {
 	.height_mm   = 130,
 };
 
+static const u32 mantix_bus_formats[] = {
+	MEDIA_BUS_FMT_RGB888_1X24,
+};
+
 static int mantix_get_modes(struct drm_panel *panel,
 			    struct drm_connector *connector)
 {
@@ -241,6 +246,10 @@ static int mantix_get_modes(struct drm_panel *panel,
 	connector->display_info.height_mm = mode->height_mm;
 	drm_mode_probed_add(connector, mode);
 
+	drm_display_info_set_bus_formats(&connector->display_info,
+					 mantix_bus_formats,
+					 ARRAY_SIZE(mantix_bus_formats));
+
 	return 1;
 }
 
diff --git a/drivers/gpu/drm/panel/panel-samsung-s6e63m0-dsi.c b/drivers/gpu/drm/panel/panel-samsung-s6e63m0-dsi.c
index e0b1a7e..e0f7736 100644
--- a/drivers/gpu/drm/panel/panel-samsung-s6e63m0-dsi.c
+++ b/drivers/gpu/drm/panel/panel-samsung-s6e63m0-dsi.c
@@ -116,7 +116,8 @@ static int s6e63m0_dsi_probe(struct mipi_dsi_device *dsi)
 static int s6e63m0_dsi_remove(struct mipi_dsi_device *dsi)
 {
 	mipi_dsi_detach(dsi);
-	return s6e63m0_remove(&dsi->dev);
+	s6e63m0_remove(&dsi->dev);
+	return 0;
 }
 
 static const struct of_device_id s6e63m0_dsi_of_match[] = {
diff --git a/drivers/gpu/drm/panel/panel-samsung-s6e63m0-spi.c b/drivers/gpu/drm/panel/panel-samsung-s6e63m0-spi.c
index 3669cc3..c178d96 100644
--- a/drivers/gpu/drm/panel/panel-samsung-s6e63m0-spi.c
+++ b/drivers/gpu/drm/panel/panel-samsung-s6e63m0-spi.c
@@ -64,7 +64,8 @@ static int s6e63m0_spi_probe(struct spi_device *spi)
 
 static int s6e63m0_spi_remove(struct spi_device *spi)
 {
-	return s6e63m0_remove(&spi->dev);
+	s6e63m0_remove(&spi->dev);
+	return 0;
 }
 
 static const struct of_device_id s6e63m0_spi_of_match[] = {
diff --git a/drivers/gpu/drm/panel/panel-samsung-s6e63m0.c b/drivers/gpu/drm/panel/panel-samsung-s6e63m0.c
index 35d72ac..b34fa4d 100644
--- a/drivers/gpu/drm/panel/panel-samsung-s6e63m0.c
+++ b/drivers/gpu/drm/panel/panel-samsung-s6e63m0.c
@@ -749,13 +749,11 @@ int s6e63m0_probe(struct device *dev, void *trsp,
 }
 EXPORT_SYMBOL_GPL(s6e63m0_probe);
 
-int s6e63m0_remove(struct device *dev)
+void s6e63m0_remove(struct device *dev)
 {
 	struct s6e63m0 *ctx = dev_get_drvdata(dev);
 
 	drm_panel_remove(&ctx->panel);
-
-	return 0;
 }
 EXPORT_SYMBOL_GPL(s6e63m0_remove);
 
diff --git a/drivers/gpu/drm/panel/panel-samsung-s6e63m0.h b/drivers/gpu/drm/panel/panel-samsung-s6e63m0.h
index 306605e..c926eca 100644
--- a/drivers/gpu/drm/panel/panel-samsung-s6e63m0.h
+++ b/drivers/gpu/drm/panel/panel-samsung-s6e63m0.h
@@ -35,6 +35,6 @@ int s6e63m0_probe(struct device *dev, void *trsp,
 				   const u8 *data,
 				   size_t len),
 		  bool dsi_mode);
-int s6e63m0_remove(struct device *dev);
+void s6e63m0_remove(struct device *dev);
 
 #endif /* _PANEL_SAMSUNG_S6E63M0_H */
diff --git a/drivers/gpu/drm/panel/panel-sharp-ls060t1sx01.c b/drivers/gpu/drm/panel/panel-sharp-ls060t1sx01.c
new file mode 100644
index 0000000..e125705
--- /dev/null
+++ b/drivers/gpu/drm/panel/panel-sharp-ls060t1sx01.c
@@ -0,0 +1,333 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (c) 2021 Linaro Ltd.
+ * Generated with linux-mdss-dsi-panel-driver-generator from vendor device tree:
+ *   Copyright (c) 2013-2014, The Linux Foundation. All rights reserved.
+ */
+
+#include <linux/delay.h>
+#include <linux/gpio/consumer.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/regulator/consumer.h>
+
+#include <video/mipi_display.h>
+
+#include <drm/drm_mipi_dsi.h>
+#include <drm/drm_modes.h>
+#include <drm/drm_panel.h>
+
+struct sharp_ls060 {
+	struct drm_panel panel;
+	struct mipi_dsi_device *dsi;
+	struct regulator *vddi_supply;
+	struct regulator *vddh_supply;
+	struct regulator *avdd_supply;
+	struct regulator *avee_supply;
+	struct gpio_desc *reset_gpio;
+	bool prepared;
+};
+
+static inline struct sharp_ls060 *to_sharp_ls060(struct drm_panel *panel)
+{
+	return container_of(panel, struct sharp_ls060, panel);
+}
+
+#define dsi_dcs_write_seq(dsi, seq...) ({				\
+		static const u8 d[] = { seq };				\
+									\
+		mipi_dsi_dcs_write_buffer(dsi, d, ARRAY_SIZE(d));	\
+	})
+
+static void sharp_ls060_reset(struct sharp_ls060 *ctx)
+{
+	gpiod_set_value_cansleep(ctx->reset_gpio, 0);
+	usleep_range(10000, 11000);
+	gpiod_set_value_cansleep(ctx->reset_gpio, 1);
+	usleep_range(10000, 11000);
+	gpiod_set_value_cansleep(ctx->reset_gpio, 0);
+	usleep_range(10000, 11000);
+}
+
+static int sharp_ls060_on(struct sharp_ls060 *ctx)
+{
+	struct mipi_dsi_device *dsi = ctx->dsi;
+	struct device *dev = &dsi->dev;
+	int ret;
+
+	dsi->mode_flags |= MIPI_DSI_MODE_LPM;
+
+	ret = dsi_dcs_write_seq(dsi, 0xbb, 0x13);
+	if (ret < 0) {
+		dev_err(dev, "Failed to send command: %d\n", ret);
+		return ret;
+	}
+
+	ret = dsi_dcs_write_seq(dsi, MIPI_DCS_WRITE_MEMORY_START);
+	if (ret < 0) {
+		dev_err(dev, "Failed to send command: %d\n", ret);
+		return ret;
+	}
+
+	ret = mipi_dsi_dcs_exit_sleep_mode(dsi);
+	if (ret < 0) {
+		dev_err(dev, "Failed to exit sleep mode: %d\n", ret);
+		return ret;
+	}
+	msleep(120);
+
+	ret = mipi_dsi_dcs_set_display_on(dsi);
+	if (ret < 0) {
+		dev_err(dev, "Failed to set display on: %d\n", ret);
+		return ret;
+	}
+	msleep(50);
+
+	return 0;
+}
+
+static int sharp_ls060_off(struct sharp_ls060 *ctx)
+{
+	struct mipi_dsi_device *dsi = ctx->dsi;
+	struct device *dev = &dsi->dev;
+	int ret;
+
+	dsi->mode_flags &= ~MIPI_DSI_MODE_LPM;
+
+	ret = mipi_dsi_dcs_set_display_off(dsi);
+	if (ret < 0) {
+		dev_err(dev, "Failed to set display off: %d\n", ret);
+		return ret;
+	}
+	usleep_range(2000, 3000);
+
+	ret = mipi_dsi_dcs_enter_sleep_mode(dsi);
+	if (ret < 0) {
+		dev_err(dev, "Failed to enter sleep mode: %d\n", ret);
+		return ret;
+	}
+	msleep(121);
+
+	return 0;
+}
+
+static int sharp_ls060_prepare(struct drm_panel *panel)
+{
+	struct sharp_ls060 *ctx = to_sharp_ls060(panel);
+	struct device *dev = &ctx->dsi->dev;
+	int ret;
+
+	if (ctx->prepared)
+		return 0;
+
+	ret = regulator_enable(ctx->vddi_supply);
+	if (ret < 0)
+		return ret;
+
+	ret = regulator_enable(ctx->avdd_supply);
+	if (ret < 0)
+		goto err_avdd;
+
+	usleep_range(1000, 2000);
+
+	ret = regulator_enable(ctx->avee_supply);
+	if (ret < 0)
+		goto err_avee;
+
+	usleep_range(10000, 11000);
+
+	ret = regulator_enable(ctx->vddh_supply);
+	if (ret < 0)
+		goto err_vddh;
+
+	usleep_range(10000, 11000);
+
+	sharp_ls060_reset(ctx);
+
+	ret = sharp_ls060_on(ctx);
+	if (ret < 0) {
+		dev_err(dev, "Failed to initialize panel: %d\n", ret);
+		goto err_on;
+	}
+
+	ctx->prepared = true;
+
+	return 0;
+
+err_on:
+	regulator_disable(ctx->vddh_supply);
+
+	usleep_range(10000, 11000);
+
+err_vddh:
+	regulator_disable(ctx->avee_supply);
+
+err_avee:
+	regulator_disable(ctx->avdd_supply);
+
+	gpiod_set_value_cansleep(ctx->reset_gpio, 1);
+
+err_avdd:
+	regulator_disable(ctx->vddi_supply);
+
+	return ret;
+}
+
+static int sharp_ls060_unprepare(struct drm_panel *panel)
+{
+	struct sharp_ls060 *ctx = to_sharp_ls060(panel);
+	struct device *dev = &ctx->dsi->dev;
+	int ret;
+
+	if (!ctx->prepared)
+		return 0;
+
+	ret = sharp_ls060_off(ctx);
+	if (ret < 0)
+		dev_err(dev, "Failed to un-initialize panel: %d\n", ret);
+
+	regulator_disable(ctx->vddh_supply);
+
+	usleep_range(10000, 11000);
+
+	regulator_disable(ctx->avee_supply);
+	regulator_disable(ctx->avdd_supply);
+
+	gpiod_set_value_cansleep(ctx->reset_gpio, 1);
+
+	regulator_disable(ctx->vddi_supply);
+
+	ctx->prepared = false;
+	return 0;
+}
+
+static const struct drm_display_mode sharp_ls060_mode = {
+	.clock = (1080 + 96 + 16 + 64) * (1920 + 4 + 1 + 16) * 60 / 1000,
+	.hdisplay = 1080,
+	.hsync_start = 1080 + 96,
+	.hsync_end = 1080 + 96 + 16,
+	.htotal = 1080 + 96 + 16 + 64,
+	.vdisplay = 1920,
+	.vsync_start = 1920 + 4,
+	.vsync_end = 1920 + 4 + 1,
+	.vtotal = 1920 + 4 + 1 + 16,
+	.width_mm = 75,
+	.height_mm = 132,
+};
+
+static int sharp_ls060_get_modes(struct drm_panel *panel,
+				 struct drm_connector *connector)
+{
+	struct drm_display_mode *mode;
+
+	mode = drm_mode_duplicate(connector->dev, &sharp_ls060_mode);
+	if (!mode)
+		return -ENOMEM;
+
+	drm_mode_set_name(mode);
+
+	mode->type = DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED;
+	connector->display_info.width_mm = mode->width_mm;
+	connector->display_info.height_mm = mode->height_mm;
+	drm_mode_probed_add(connector, mode);
+
+	return 1;
+}
+
+static const struct drm_panel_funcs sharp_ls060_panel_funcs = {
+	.prepare = sharp_ls060_prepare,
+	.unprepare = sharp_ls060_unprepare,
+	.get_modes = sharp_ls060_get_modes,
+};
+
+static int sharp_ls060_probe(struct mipi_dsi_device *dsi)
+{
+	struct device *dev = &dsi->dev;
+	struct sharp_ls060 *ctx;
+	int ret;
+
+	ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	ctx->vddi_supply = devm_regulator_get(dev, "vddi");
+	if (IS_ERR(ctx->vddi_supply))
+		return PTR_ERR(ctx->vddi_supply);
+
+	ctx->vddh_supply = devm_regulator_get(dev, "vddh");
+	if (IS_ERR(ctx->vddh_supply))
+		return PTR_ERR(ctx->vddh_supply);
+
+	ctx->avdd_supply = devm_regulator_get(dev, "avdd");
+	if (IS_ERR(ctx->avdd_supply))
+		return PTR_ERR(ctx->avdd_supply);
+
+	ctx->avee_supply = devm_regulator_get(dev, "avee");
+	if (IS_ERR(ctx->avee_supply))
+		return PTR_ERR(ctx->avee_supply);
+
+	ctx->reset_gpio = devm_gpiod_get(dev, "reset", GPIOD_OUT_HIGH);
+	if (IS_ERR(ctx->reset_gpio))
+		return dev_err_probe(dev, PTR_ERR(ctx->reset_gpio),
+				     "Failed to get reset-gpios\n");
+
+	ctx->dsi = dsi;
+	mipi_dsi_set_drvdata(dsi, ctx);
+
+	dsi->lanes = 4;
+	dsi->format = MIPI_DSI_FMT_RGB888;
+	dsi->mode_flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_VIDEO_BURST |
+			  MIPI_DSI_MODE_NO_EOT_PACKET |
+			  MIPI_DSI_CLOCK_NON_CONTINUOUS;
+
+	drm_panel_init(&ctx->panel, dev, &sharp_ls060_panel_funcs,
+		       DRM_MODE_CONNECTOR_DSI);
+
+	ret = drm_panel_of_backlight(&ctx->panel);
+	if (ret)
+		return dev_err_probe(dev, ret, "Failed to get backlight\n");
+
+	drm_panel_add(&ctx->panel);
+
+	ret = mipi_dsi_attach(dsi);
+	if (ret < 0) {
+		dev_err(dev, "Failed to attach to DSI host: %d\n", ret);
+		drm_panel_remove(&ctx->panel);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int sharp_ls060_remove(struct mipi_dsi_device *dsi)
+{
+	struct sharp_ls060 *ctx = mipi_dsi_get_drvdata(dsi);
+	int ret;
+
+	ret = mipi_dsi_detach(dsi);
+	if (ret < 0)
+		dev_err(&dsi->dev, "Failed to detach from DSI host: %d\n", ret);
+
+	drm_panel_remove(&ctx->panel);
+
+	return 0;
+}
+
+static const struct of_device_id sharp_ls060t1sx01_of_match[] = {
+	{ .compatible = "sharp,ls060t1sx01" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, sharp_ls060t1sx01_of_match);
+
+static struct mipi_dsi_driver sharp_ls060_driver = {
+	.probe = sharp_ls060_probe,
+	.remove = sharp_ls060_remove,
+	.driver = {
+		.name = "panel-sharp-ls060t1sx01",
+		.of_match_table = sharp_ls060t1sx01_of_match,
+	},
+};
+module_mipi_dsi_driver(sharp_ls060_driver);
+
+MODULE_AUTHOR("Dmitry Baryshkov <dmitry.baryshkov@linaro.org>");
+MODULE_DESCRIPTION("DRM driver for Sharp LS060T1SX01 1080p video mode dsi panel");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c
index 7f3e1b8..eb475a3 100644
--- a/drivers/gpu/drm/panel/panel-simple.c
+++ b/drivers/gpu/drm/panel/panel-simple.c
@@ -2370,6 +2370,38 @@ static const struct panel_desc logictechno_lt170410_2whc = {
 	.connector_type = DRM_MODE_CONNECTOR_LVDS,
 };
 
+static const struct drm_display_mode logictechno_lttd800480070_l2rt_mode = {
+	.clock = 33000,
+	.hdisplay = 800,
+	.hsync_start = 800 + 112,
+	.hsync_end = 800 + 112 + 3,
+	.htotal = 800 + 112 + 3 + 85,
+	.vdisplay = 480,
+	.vsync_start = 480 + 38,
+	.vsync_end = 480 + 38 + 3,
+	.vtotal = 480 + 38 + 3 + 29,
+	.flags = DRM_MODE_FLAG_NVSYNC | DRM_MODE_FLAG_NHSYNC,
+};
+
+static const struct panel_desc logictechno_lttd800480070_l2rt = {
+	.modes = &logictechno_lttd800480070_l2rt_mode,
+	.num_modes = 1,
+	.bpc = 8,
+	.size = {
+		.width = 154,
+		.height = 86,
+	},
+	.delay = {
+		.prepare = 45,
+		.enable = 100,
+		.disable = 100,
+		.unprepare = 45
+	},
+	.bus_format = MEDIA_BUS_FMT_RGB888_1X24,
+	.bus_flags = DRM_BUS_FLAG_PIXDATA_SAMPLE_NEGEDGE,
+	.connector_type = DRM_MODE_CONNECTOR_DPI,
+};
+
 static const struct drm_display_mode logictechno_lttd800480070_l6wh_rt_mode = {
 	.clock = 33000,
 	.hdisplay = 800,
@@ -3751,6 +3783,9 @@ static const struct of_device_id platform_of_match[] = {
 		.compatible = "logictechno,lt170410-2whc",
 		.data = &logictechno_lt170410_2whc,
 	}, {
+		.compatible = "logictechno,lttd800480070-l2rt",
+		.data = &logictechno_lttd800480070_l2rt,
+	}, {
 		.compatible = "logictechno,lttd800480070-l6wh-rt",
 		.data = &logictechno_lttd800480070_l6wh_rt,
 	}, {
diff --git a/drivers/gpu/drm/panel/panel-sitronix-st7703.c b/drivers/gpu/drm/panel/panel-sitronix-st7703.c
index a2c303e..73f69c9 100644
--- a/drivers/gpu/drm/panel/panel-sitronix-st7703.c
+++ b/drivers/gpu/drm/panel/panel-sitronix-st7703.c
@@ -453,6 +453,10 @@ static int st7703_prepare(struct drm_panel *panel)
 	return ret;
 }
 
+static const u32 mantix_bus_formats[] = {
+	MEDIA_BUS_FMT_RGB888_1X24,
+};
+
 static int st7703_get_modes(struct drm_panel *panel,
 			    struct drm_connector *connector)
 {
@@ -474,6 +478,10 @@ static int st7703_get_modes(struct drm_panel *panel,
 	connector->display_info.height_mm = mode->height_mm;
 	drm_mode_probed_add(connector, mode);
 
+	drm_display_info_set_bus_formats(&connector->display_info,
+					 mantix_bus_formats,
+					 ARRAY_SIZE(mantix_bus_formats));
+
 	return 1;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index 458f92a..a36a4f2 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -61,7 +61,7 @@ static vm_fault_t radeon_gem_fault(struct vm_fault *vmf)
 		goto unlock_resv;
 
 	ret = ttm_bo_vm_fault_reserved(vmf, vmf->vma->vm_page_prot,
-				       TTM_BO_VM_NUM_PREFAULT, 1);
+				       TTM_BO_VM_NUM_PREFAULT);
 	if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
 		goto unlock_mclk;
 
diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c
index 042c16b..5bc5f77 100644
--- a/drivers/gpu/drm/scheduler/sched_main.c
+++ b/drivers/gpu/drm/scheduler/sched_main.c
@@ -699,30 +699,16 @@ int drm_sched_job_add_implicit_dependencies(struct drm_sched_job *job,
 					    struct drm_gem_object *obj,
 					    bool write)
 {
+	struct dma_resv_iter cursor;
+	struct dma_fence *fence;
 	int ret;
-	struct dma_fence **fences;
-	unsigned int i, fence_count;
 
-	if (!write) {
-		struct dma_fence *fence = dma_resv_get_excl_unlocked(obj->resv);
-
-		return drm_sched_job_add_dependency(job, fence);
-	}
-
-	ret = dma_resv_get_fences(obj->resv, NULL, &fence_count, &fences);
-	if (ret || !fence_count)
-		return ret;
-
-	for (i = 0; i < fence_count; i++) {
-		ret = drm_sched_job_add_dependency(job, fences[i]);
+	dma_resv_for_each_fence(&cursor, obj->resv, write, fence) {
+		ret = drm_sched_job_add_dependency(job, fence);
 		if (ret)
-			break;
+			return ret;
 	}
-
-	for (; i < fence_count; i++)
-		dma_fence_put(fences[i]);
-	kfree(fences);
-	return ret;
+	return 0;
 }
 EXPORT_SYMBOL(drm_sched_job_add_implicit_dependencies);
 
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index d62b201..739f11c 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -269,23 +269,15 @@ static int ttm_bo_individualize_resv(struct ttm_buffer_object *bo)
 static void ttm_bo_flush_all_fences(struct ttm_buffer_object *bo)
 {
 	struct dma_resv *resv = &bo->base._resv;
-	struct dma_resv_list *fobj;
+	struct dma_resv_iter cursor;
 	struct dma_fence *fence;
-	int i;
 
-	rcu_read_lock();
-	fobj = dma_resv_shared_list(resv);
-	fence = dma_resv_excl_fence(resv);
-	if (fence && !fence->ops->signaled)
-		dma_fence_enable_sw_signaling(fence);
-
-	for (i = 0; fobj && i < fobj->shared_count; ++i) {
-		fence = rcu_dereference(fobj->shared[i]);
-
+	dma_resv_iter_begin(&cursor, resv, true);
+	dma_resv_for_each_fence_unlocked(&cursor, fence) {
 		if (!fence->ops->signaled)
 			dma_fence_enable_sw_signaling(fence);
 	}
-	rcu_read_unlock();
+	dma_resv_iter_end(&cursor);
 }
 
 /**
@@ -627,7 +619,8 @@ static bool ttm_bo_evict_swapout_allowable(struct ttm_buffer_object *bo,
 			*busy = !ret;
 	}
 
-	if (ret && place && !bo->bdev->funcs->eviction_valuable(bo, place)) {
+	if (ret && place && (bo->resource->mem_type != place->mem_type ||
+		!bo->bdev->funcs->eviction_valuable(bo, place))) {
 		ret = false;
 		if (*locked) {
 			dma_resv_unlock(bo->base.resv);
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index 33680c9..08ba083 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -173,89 +173,6 @@ vm_fault_t ttm_bo_vm_reserve(struct ttm_buffer_object *bo,
 }
 EXPORT_SYMBOL(ttm_bo_vm_reserve);
 
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
-/**
- * ttm_bo_vm_insert_huge - Insert a pfn for PUD or PMD faults
- * @vmf: Fault data
- * @bo: The buffer object
- * @page_offset: Page offset from bo start
- * @fault_page_size: The size of the fault in pages.
- * @pgprot: The page protections.
- * Does additional checking whether it's possible to insert a PUD or PMD
- * pfn and performs the insertion.
- *
- * Return: VM_FAULT_NOPAGE on successful insertion, VM_FAULT_FALLBACK if
- * a huge fault was not possible, or on insertion error.
- */
-static vm_fault_t ttm_bo_vm_insert_huge(struct vm_fault *vmf,
-					struct ttm_buffer_object *bo,
-					pgoff_t page_offset,
-					pgoff_t fault_page_size,
-					pgprot_t pgprot)
-{
-	pgoff_t i;
-	vm_fault_t ret;
-	unsigned long pfn;
-	pfn_t pfnt;
-	struct ttm_tt *ttm = bo->ttm;
-	bool write = vmf->flags & FAULT_FLAG_WRITE;
-
-	/* Fault should not cross bo boundary. */
-	page_offset &= ~(fault_page_size - 1);
-	if (page_offset + fault_page_size > bo->resource->num_pages)
-		goto out_fallback;
-
-	if (bo->resource->bus.is_iomem)
-		pfn = ttm_bo_io_mem_pfn(bo, page_offset);
-	else
-		pfn = page_to_pfn(ttm->pages[page_offset]);
-
-	/* pfn must be fault_page_size aligned. */
-	if ((pfn & (fault_page_size - 1)) != 0)
-		goto out_fallback;
-
-	/* Check that memory is contiguous. */
-	if (!bo->resource->bus.is_iomem) {
-		for (i = 1; i < fault_page_size; ++i) {
-			if (page_to_pfn(ttm->pages[page_offset + i]) != pfn + i)
-				goto out_fallback;
-		}
-	} else if (bo->bdev->funcs->io_mem_pfn) {
-		for (i = 1; i < fault_page_size; ++i) {
-			if (ttm_bo_io_mem_pfn(bo, page_offset + i) != pfn + i)
-				goto out_fallback;
-		}
-	}
-
-	pfnt = __pfn_to_pfn_t(pfn, PFN_DEV);
-	if (fault_page_size == (HPAGE_PMD_SIZE >> PAGE_SHIFT))
-		ret = vmf_insert_pfn_pmd_prot(vmf, pfnt, pgprot, write);
-#ifdef CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
-	else if (fault_page_size == (HPAGE_PUD_SIZE >> PAGE_SHIFT))
-		ret = vmf_insert_pfn_pud_prot(vmf, pfnt, pgprot, write);
-#endif
-	else
-		WARN_ON_ONCE(ret = VM_FAULT_FALLBACK);
-
-	if (ret != VM_FAULT_NOPAGE)
-		goto out_fallback;
-
-	return VM_FAULT_NOPAGE;
-out_fallback:
-	count_vm_event(THP_FAULT_FALLBACK);
-	return VM_FAULT_FALLBACK;
-}
-#else
-static vm_fault_t ttm_bo_vm_insert_huge(struct vm_fault *vmf,
-					struct ttm_buffer_object *bo,
-					pgoff_t page_offset,
-					pgoff_t fault_page_size,
-					pgprot_t pgprot)
-{
-	return VM_FAULT_FALLBACK;
-}
-#endif
-
 /**
  * ttm_bo_vm_fault_reserved - TTM fault helper
  * @vmf: The struct vm_fault given as argument to the fault callback
@@ -263,7 +180,6 @@ static vm_fault_t ttm_bo_vm_insert_huge(struct vm_fault *vmf,
  * @num_prefault: Maximum number of prefault pages. The caller may want to
  * specify this based on madvice settings and the size of the GPU object
  * backed by the memory.
- * @fault_page_size: The size of the fault in pages.
  *
  * This function inserts one or more page table entries pointing to the
  * memory backing the buffer object, and then returns a return code
@@ -277,8 +193,7 @@ static vm_fault_t ttm_bo_vm_insert_huge(struct vm_fault *vmf,
  */
 vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 				    pgprot_t prot,
-				    pgoff_t num_prefault,
-				    pgoff_t fault_page_size)
+				    pgoff_t num_prefault)
 {
 	struct vm_area_struct *vma = vmf->vma;
 	struct ttm_buffer_object *bo = vma->vm_private_data;
@@ -329,11 +244,6 @@ vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 		prot = pgprot_decrypted(prot);
 	}
 
-	/* We don't prefault on huge faults. Yet. */
-	if (IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE) && fault_page_size != 1)
-		return ttm_bo_vm_insert_huge(vmf, bo, page_offset,
-					     fault_page_size, prot);
-
 	/*
 	 * Speculatively prefault a number of pages. Only error on
 	 * first page.
@@ -429,7 +339,7 @@ vm_fault_t ttm_bo_vm_fault(struct vm_fault *vmf)
 
 	prot = vma->vm_page_prot;
 	if (drm_dev_enter(ddev, &idx)) {
-		ret = ttm_bo_vm_fault_reserved(vmf, prot, TTM_BO_VM_NUM_PREFAULT, 1);
+		ret = ttm_bo_vm_fault_reserved(vmf, prot, TTM_BO_VM_NUM_PREFAULT);
 		drm_dev_exit(idx);
 	} else {
 		ret = ttm_bo_vm_dummy_page(vmf, prot);
diff --git a/drivers/gpu/drm/udl/udl_connector.c b/drivers/gpu/drm/udl/udl_connector.c
index 3750fd2..930574a 100644
--- a/drivers/gpu/drm/udl/udl_connector.c
+++ b/drivers/gpu/drm/udl/udl_connector.c
@@ -30,7 +30,7 @@ static int udl_get_edid_block(void *data, u8 *buf, unsigned int block,
 		int bval = (i + block * EDID_LENGTH) << 8;
 		ret = usb_control_msg(udev, usb_rcvctrlpipe(udev, 0),
 				      0x02, (0x80 | (0x02 << 5)), bval,
-				      0xA1, read_buff, 2, HZ);
+				      0xA1, read_buff, 2, 1000);
 		if (ret < 1) {
 			DRM_ERROR("Read EDID byte %d failed err %x\n", i, ret);
 			kfree(read_buff);
diff --git a/drivers/gpu/drm/v3d/v3d_gem.c b/drivers/gpu/drm/v3d/v3d_gem.c
index 6a000d7..e47ae40 100644
--- a/drivers/gpu/drm/v3d/v3d_gem.c
+++ b/drivers/gpu/drm/v3d/v3d_gem.c
@@ -487,8 +487,8 @@ v3d_job_init(struct v3d_dev *v3d, struct drm_file *file_priv,
 			for (i = 0; i < se->in_sync_count; i++) {
 				struct drm_v3d_sem in;
 
-				ret = copy_from_user(&in, handle++, sizeof(in));
-				if (ret) {
+				if (copy_from_user(&in, handle++, sizeof(in))) {
+					ret = -EFAULT;
 					DRM_DEBUG("Failed to copy wait dep handle.\n");
 					goto fail_deps;
 				}
@@ -609,8 +609,8 @@ v3d_get_multisync_post_deps(struct drm_file *file_priv,
 	for (i = 0; i < count; i++) {
 		struct drm_v3d_sem out;
 
-		ret = copy_from_user(&out, post_deps++, sizeof(out));
-		if (ret) {
+		if (copy_from_user(&out, post_deps++, sizeof(out))) {
+			ret = -EFAULT;
 			DRM_DEBUG("Failed to copy post dep handles\n");
 			goto fail;
 		}
@@ -646,9 +646,8 @@ v3d_get_multisync_submit_deps(struct drm_file *file_priv,
 	struct v3d_submit_ext *se = data;
 	int ret;
 
-	ret = copy_from_user(&multisync, ext, sizeof(multisync));
-	if (ret)
-		return ret;
+	if (copy_from_user(&multisync, ext, sizeof(multisync)))
+		return -EFAULT;
 
 	if (multisync.pad)
 		return -EINVAL;
diff --git a/drivers/gpu/drm/virtio/virtgpu_display.c b/drivers/gpu/drm/virtio/virtgpu_display.c
index a6caebd..5b00310a 100644
--- a/drivers/gpu/drm/virtio/virtgpu_display.c
+++ b/drivers/gpu/drm/virtio/virtgpu_display.c
@@ -308,8 +308,10 @@ virtio_gpu_user_framebuffer_create(struct drm_device *dev,
 		return ERR_PTR(-EINVAL);
 
 	virtio_gpu_fb = kzalloc(sizeof(*virtio_gpu_fb), GFP_KERNEL);
-	if (virtio_gpu_fb == NULL)
+	if (virtio_gpu_fb == NULL) {
+		drm_gem_object_put(obj);
 		return ERR_PTR(-ENOMEM);
+	}
 
 	ret = virtio_gpu_framebuffer_init(dev, virtio_gpu_fb, mode_cmd, obj);
 	if (ret) {
diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.c b/drivers/gpu/drm/virtio/virtgpu_drv.c
index 749db18..d86e1ad 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.c
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.c
@@ -163,10 +163,11 @@ static __poll_t virtio_gpu_poll(struct file *filp,
 	struct drm_file *drm_file = filp->private_data;
 	struct virtio_gpu_fpriv *vfpriv = drm_file->driver_priv;
 	struct drm_device *dev = drm_file->minor->dev;
+	struct virtio_gpu_device *vgdev = dev->dev_private;
 	struct drm_pending_event *e = NULL;
 	__poll_t mask = 0;
 
-	if (!vfpriv->ring_idx_mask)
+	if (!vgdev->has_virgl_3d || !vfpriv || !vfpriv->ring_idx_mask)
 		return drm_poll(filp, wait);
 
 	poll_wait(filp, &drm_file->event_wait, wait);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index a833751..858aff9 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -1550,10 +1550,6 @@ void vmw_bo_dirty_unmap(struct vmw_buffer_object *vbo,
 			pgoff_t start, pgoff_t end);
 vm_fault_t vmw_bo_vm_fault(struct vm_fault *vmf);
 vm_fault_t vmw_bo_vm_mkwrite(struct vm_fault *vmf);
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
-vm_fault_t vmw_bo_vm_huge_fault(struct vm_fault *vmf,
-				enum page_entry_size pe_size);
-#endif
 
 /* Transparent hugepage support - vmwgfx_thp.c */
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c b/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
index e5a9a5c..922317d 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
@@ -477,7 +477,7 @@ vm_fault_t vmw_bo_vm_fault(struct vm_fault *vmf)
 	else
 		prot = vm_get_page_prot(vma->vm_flags);
 
-	ret = ttm_bo_vm_fault_reserved(vmf, prot, num_prefault, 1);
+	ret = ttm_bo_vm_fault_reserved(vmf, prot, num_prefault);
 	if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
 		return ret;
 
@@ -486,73 +486,3 @@ vm_fault_t vmw_bo_vm_fault(struct vm_fault *vmf)
 
 	return ret;
 }
-
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
-vm_fault_t vmw_bo_vm_huge_fault(struct vm_fault *vmf,
-				enum page_entry_size pe_size)
-{
-	struct vm_area_struct *vma = vmf->vma;
-	struct ttm_buffer_object *bo = (struct ttm_buffer_object *)
-	    vma->vm_private_data;
-	struct vmw_buffer_object *vbo =
-		container_of(bo, struct vmw_buffer_object, base);
-	pgprot_t prot;
-	vm_fault_t ret;
-	pgoff_t fault_page_size;
-	bool write = vmf->flags & FAULT_FLAG_WRITE;
-
-	switch (pe_size) {
-	case PE_SIZE_PMD:
-		fault_page_size = HPAGE_PMD_SIZE >> PAGE_SHIFT;
-		break;
-#ifdef CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
-	case PE_SIZE_PUD:
-		fault_page_size = HPAGE_PUD_SIZE >> PAGE_SHIFT;
-		break;
-#endif
-	default:
-		WARN_ON_ONCE(1);
-		return VM_FAULT_FALLBACK;
-	}
-
-	/* Always do write dirty-tracking and COW on PTE level. */
-	if (write && (READ_ONCE(vbo->dirty) || is_cow_mapping(vma->vm_flags)))
-		return VM_FAULT_FALLBACK;
-
-	ret = ttm_bo_vm_reserve(bo, vmf);
-	if (ret)
-		return ret;
-
-	if (vbo->dirty) {
-		pgoff_t allowed_prefault;
-		unsigned long page_offset;
-
-		page_offset = vmf->pgoff -
-			drm_vma_node_start(&bo->base.vma_node);
-		if (page_offset >= bo->resource->num_pages ||
-		    vmw_resources_clean(vbo, page_offset,
-					page_offset + PAGE_SIZE,
-					&allowed_prefault)) {
-			ret = VM_FAULT_SIGBUS;
-			goto out_unlock;
-		}
-
-		/*
-		 * Write protect, so we get a new fault on write, and can
-		 * split.
-		 */
-		prot = vm_get_page_prot(vma->vm_flags & ~VM_SHARED);
-	} else {
-		prot = vm_get_page_prot(vma->vm_flags);
-	}
-
-	ret = ttm_bo_vm_fault_reserved(vmf, prot, 1, fault_page_size);
-	if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
-		return ret;
-
-out_unlock:
-	dma_resv_unlock(bo->base.resv);
-
-	return ret;
-}
-#endif
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
index e6b1f98..0a4c340 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
@@ -61,9 +61,6 @@ int vmw_mmap(struct file *filp, struct vm_area_struct *vma)
 		.fault = vmw_bo_vm_fault,
 		.open = ttm_bo_vm_open,
 		.close = ttm_bo_vm_close,
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
-		.huge_fault = vmw_bo_vm_huge_fault,
-#endif
 	};
 	struct drm_file *file_priv = filp->private_data;
 	struct vmw_private *dev_priv = vmw_priv(file_priv->minor->dev);
diff --git a/drivers/input/joystick/analog.c b/drivers/input/joystick/analog.c
index 882c3c8b..3088c5b 100644
--- a/drivers/input/joystick/analog.c
+++ b/drivers/input/joystick/analog.c
@@ -19,6 +19,7 @@
 #include <linux/input.h>
 #include <linux/gameport.h>
 #include <linux/jiffies.h>
+#include <linux/seq_buf.h>
 #include <linux/timex.h>
 #include <linux/timekeeping.h>
 
@@ -338,23 +339,24 @@ static void analog_calibrate_timer(struct analog_port *port)
 
 static void analog_name(struct analog *analog)
 {
-	snprintf(analog->name, sizeof(analog->name), "Analog %d-axis %d-button",
+	struct seq_buf s;
+
+	seq_buf_init(&s, analog->name, sizeof(analog->name));
+	seq_buf_printf(&s, "Analog %d-axis %d-button",
 		 hweight8(analog->mask & ANALOG_AXES_STD),
 		 hweight8(analog->mask & ANALOG_BTNS_STD) + !!(analog->mask & ANALOG_BTNS_CHF) * 2 +
 		 hweight16(analog->mask & ANALOG_BTNS_GAMEPAD) + !!(analog->mask & ANALOG_HBTN_CHF) * 4);
 
 	if (analog->mask & ANALOG_HATS_ALL)
-		snprintf(analog->name, sizeof(analog->name), "%s %d-hat",
-			 analog->name, hweight16(analog->mask & ANALOG_HATS_ALL));
+		seq_buf_printf(&s, " %d-hat",
+			       hweight16(analog->mask & ANALOG_HATS_ALL));
 
 	if (analog->mask & ANALOG_HAT_FCS)
-		strlcat(analog->name, " FCS", sizeof(analog->name));
+		seq_buf_printf(&s, " FCS");
 	if (analog->mask & ANALOG_ANY_CHF)
-		strlcat(analog->name, (analog->mask & ANALOG_SAITEK) ? " Saitek" : " CHF",
-			sizeof(analog->name));
+		seq_buf_printf(&s, (analog->mask & ANALOG_SAITEK) ? " Saitek" : " CHF");
 
-	strlcat(analog->name, (analog->mask & ANALOG_GAMEPAD) ? " gamepad": " joystick",
-		sizeof(analog->name));
+	seq_buf_printf(&s, (analog->mask & ANALOG_GAMEPAD) ? " gamepad" : " joystick");
 }
 
 /*
diff --git a/drivers/input/joystick/iforce/iforce-usb.c b/drivers/input/joystick/iforce/iforce-usb.c
index 6c554c1..ea58805 100644
--- a/drivers/input/joystick/iforce/iforce-usb.c
+++ b/drivers/input/joystick/iforce/iforce-usb.c
@@ -92,7 +92,7 @@ static int iforce_usb_get_id(struct iforce *iforce, u8 id,
 				 id,
 				 USB_TYPE_VENDOR | USB_DIR_IN |
 					USB_RECIP_INTERFACE,
-				 0, 0, buf, IFORCE_MAX_LENGTH, HZ);
+				 0, 0, buf, IFORCE_MAX_LENGTH, 1000);
 	if (status < 0) {
 		dev_err(&iforce_usb->intf->dev,
 			"usb_submit_urb failed: %d\n", status);
diff --git a/drivers/input/joystick/tmdc.c b/drivers/input/joystick/tmdc.c
index f89e9aa..7416de8 100644
--- a/drivers/input/joystick/tmdc.c
+++ b/drivers/input/joystick/tmdc.c
@@ -83,7 +83,7 @@ static const struct tmdc_model {
 	const signed char *axes;
 	const short *buttons;
 } tmdc_models[] = {
-	{   1, "ThrustMaster Millenium 3D Inceptor",	  6, 2, { 4, 2 }, { 4, 6 }, tmdc_abs, tmdc_btn_joy },
+	{   1, "ThrustMaster Millennium 3D Inceptor",	  6, 2, { 4, 2 }, { 4, 6 }, tmdc_abs, tmdc_btn_joy },
 	{   3, "ThrustMaster Rage 3D Gamepad",		  2, 0, { 8, 2 }, { 0, 0 }, tmdc_abs, tmdc_btn_pad },
 	{   4, "ThrustMaster Attack Throttle",		  5, 2, { 4, 6 }, { 4, 2 }, tmdc_abs_at, tmdc_btn_at },
 	{   8, "ThrustMaster FragMaster",		  4, 0, { 8, 2 }, { 0, 0 }, tmdc_abs_fm, tmdc_btn_fm },
diff --git a/drivers/input/keyboard/Kconfig b/drivers/input/keyboard/Kconfig
index e75650e..0c607da 100644
--- a/drivers/input/keyboard/Kconfig
+++ b/drivers/input/keyboard/Kconfig
@@ -788,4 +788,14 @@
 	  To compile this driver as a module, choose M here: the
 	  module will be called pmic-keys.
 
+config KEYBOARD_CYPRESS_SF
+	tristate "Cypress StreetFighter touchkey support"
+	depends on I2C
+	help
+	  Say Y here if you want to enable support for Cypress StreetFighter
+	  touchkeys.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called cypress-sf.
+
 endif
diff --git a/drivers/input/keyboard/Makefile b/drivers/input/keyboard/Makefile
index 1d689fd..e3c8648 100644
--- a/drivers/input/keyboard/Makefile
+++ b/drivers/input/keyboard/Makefile
@@ -17,6 +17,7 @@
 obj-$(CONFIG_KEYBOARD_CAP11XX)		+= cap11xx.o
 obj-$(CONFIG_KEYBOARD_CLPS711X)		+= clps711x-keypad.o
 obj-$(CONFIG_KEYBOARD_CROS_EC)		+= cros_ec_keyb.o
+obj-$(CONFIG_KEYBOARD_CYPRESS_SF)	+= cypress-sf.o
 obj-$(CONFIG_KEYBOARD_DAVINCI)		+= davinci_keyscan.o
 obj-$(CONFIG_KEYBOARD_DLINK_DIR685)	+= dlink-dir685-touchkeys.o
 obj-$(CONFIG_KEYBOARD_EP93XX)		+= ep93xx_keypad.o
diff --git a/drivers/input/keyboard/cap11xx.c b/drivers/input/keyboard/cap11xx.c
index 688e2be..7c85343 100644
--- a/drivers/input/keyboard/cap11xx.c
+++ b/drivers/input/keyboard/cap11xx.c
@@ -91,18 +91,21 @@ struct cap11xx_hw_model {
 	u8 product_id;
 	unsigned int num_channels;
 	unsigned int num_leds;
+	bool no_gain;
 };
 
 enum {
 	CAP1106,
 	CAP1126,
 	CAP1188,
+	CAP1206,
 };
 
 static const struct cap11xx_hw_model cap11xx_devices[] = {
-	[CAP1106] = { .product_id = 0x55, .num_channels = 6, .num_leds = 0 },
-	[CAP1126] = { .product_id = 0x53, .num_channels = 6, .num_leds = 2 },
-	[CAP1188] = { .product_id = 0x50, .num_channels = 8, .num_leds = 8 },
+	[CAP1106] = { .product_id = 0x55, .num_channels = 6, .num_leds = 0, .no_gain = false },
+	[CAP1126] = { .product_id = 0x53, .num_channels = 6, .num_leds = 2, .no_gain = false },
+	[CAP1188] = { .product_id = 0x50, .num_channels = 8, .num_leds = 8, .no_gain = false },
+	[CAP1206] = { .product_id = 0x67, .num_channels = 6, .num_leds = 0, .no_gain = true },
 };
 
 static const struct reg_default cap11xx_reg_defaults[] = {
@@ -378,17 +381,24 @@ static int cap11xx_i2c_probe(struct i2c_client *i2c_client,
 	node = dev->of_node;
 
 	if (!of_property_read_u32(node, "microchip,sensor-gain", &gain32)) {
-		if (is_power_of_2(gain32) && gain32 <= 8)
+		if (cap->no_gain)
+			dev_warn(dev,
+				 "This version doesn't support sensor gain\n");
+		else if (is_power_of_2(gain32) && gain32 <= 8)
 			gain = ilog2(gain32);
 		else
 			dev_err(dev, "Invalid sensor-gain value %d\n", gain32);
 	}
 
-	if (of_property_read_bool(node, "microchip,irq-active-high")) {
-		error = regmap_update_bits(priv->regmap, CAP11XX_REG_CONFIG2,
-					   CAP11XX_REG_CONFIG2_ALT_POL, 0);
-		if (error)
-			return error;
+	if (id->driver_data != CAP1206) {
+		if (of_property_read_bool(node, "microchip,irq-active-high")) {
+			error = regmap_update_bits(priv->regmap,
+						   CAP11XX_REG_CONFIG2,
+						   CAP11XX_REG_CONFIG2_ALT_POL,
+						   0);
+			if (error)
+				return error;
+		}
 	}
 
 	/* Provide some useful defaults */
@@ -398,11 +408,14 @@ static int cap11xx_i2c_probe(struct i2c_client *i2c_client,
 	of_property_read_u32_array(node, "linux,keycodes",
 				   priv->keycodes, cap->num_channels);
 
-	error = regmap_update_bits(priv->regmap, CAP11XX_REG_MAIN_CONTROL,
-				   CAP11XX_REG_MAIN_CONTROL_GAIN_MASK,
-				   gain << CAP11XX_REG_MAIN_CONTROL_GAIN_SHIFT);
-	if (error)
-		return error;
+	if (!cap->no_gain) {
+		error = regmap_update_bits(priv->regmap,
+				CAP11XX_REG_MAIN_CONTROL,
+				CAP11XX_REG_MAIN_CONTROL_GAIN_MASK,
+				gain << CAP11XX_REG_MAIN_CONTROL_GAIN_SHIFT);
+		if (error)
+			return error;
+	}
 
 	/* Disable autorepeat. The Linux input system has its own handling. */
 	error = regmap_write(priv->regmap, CAP11XX_REG_REPEAT_RATE, 0);
@@ -470,6 +483,7 @@ static const struct of_device_id cap11xx_dt_ids[] = {
 	{ .compatible = "microchip,cap1106", },
 	{ .compatible = "microchip,cap1126", },
 	{ .compatible = "microchip,cap1188", },
+	{ .compatible = "microchip,cap1206", },
 	{}
 };
 MODULE_DEVICE_TABLE(of, cap11xx_dt_ids);
@@ -478,6 +492,7 @@ static const struct i2c_device_id cap11xx_i2c_ids[] = {
 	{ "cap1106", CAP1106 },
 	{ "cap1126", CAP1126 },
 	{ "cap1188", CAP1188 },
+	{ "cap1206", CAP1206 },
 	{}
 };
 MODULE_DEVICE_TABLE(i2c, cap11xx_i2c_ids);
diff --git a/drivers/input/keyboard/cypress-sf.c b/drivers/input/keyboard/cypress-sf.c
new file mode 100644
index 0000000..c289960
--- /dev/null
+++ b/drivers/input/keyboard/cypress-sf.c
@@ -0,0 +1,224 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Cypress StreetFighter Touchkey Driver
+ *
+ * Copyright (c) 2021 Yassine Oudjana <y.oudjana@protonmail.com>
+ */
+
+#include <linux/bitmap.h>
+#include <linux/bitops.h>
+#include <linux/device.h>
+#include <linux/i2c.h>
+#include <linux/input.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/pm.h>
+#include <linux/regulator/consumer.h>
+
+#define CYPRESS_SF_DEV_NAME "cypress-sf"
+
+#define CYPRESS_SF_REG_BUTTON_STATUS	0x4a
+
+struct cypress_sf_data {
+	struct i2c_client *client;
+	struct input_dev *input_dev;
+	struct regulator_bulk_data regulators[2];
+	u32 *keycodes;
+	unsigned long keystates;
+	int num_keys;
+};
+
+static irqreturn_t cypress_sf_irq_handler(int irq, void *devid)
+{
+	struct cypress_sf_data *touchkey = devid;
+	unsigned long keystates, changed;
+	bool new_state;
+	int val, key;
+
+	val = i2c_smbus_read_byte_data(touchkey->client,
+				       CYPRESS_SF_REG_BUTTON_STATUS);
+	if (val < 0) {
+		dev_err(&touchkey->client->dev,
+			"Failed to read button status: %d", val);
+		return IRQ_NONE;
+	}
+	keystates = val;
+
+	bitmap_xor(&changed, &keystates, &touchkey->keystates,
+		   touchkey->num_keys);
+
+	for_each_set_bit(key, &changed, touchkey->num_keys) {
+		new_state = keystates & BIT(key);
+		dev_dbg(&touchkey->client->dev,
+			"Key %d changed to %d", key, new_state);
+		input_report_key(touchkey->input_dev,
+				 touchkey->keycodes[key], new_state);
+	}
+
+	input_sync(touchkey->input_dev);
+	touchkey->keystates = keystates;
+
+	return IRQ_HANDLED;
+}
+
+static int cypress_sf_probe(struct i2c_client *client)
+{
+	struct cypress_sf_data *touchkey;
+	int key, error;
+
+	touchkey = devm_kzalloc(&client->dev, sizeof(*touchkey), GFP_KERNEL);
+	if (!touchkey)
+		return -ENOMEM;
+
+	touchkey->client = client;
+	i2c_set_clientdata(client, touchkey);
+
+	touchkey->regulators[0].supply = "vdd";
+	touchkey->regulators[1].supply = "avdd";
+
+	error = devm_regulator_bulk_get(&client->dev,
+					ARRAY_SIZE(touchkey->regulators),
+					touchkey->regulators);
+	if (error) {
+		dev_err(&client->dev, "Failed to get regulators: %d\n", error);
+		return error;
+	}
+
+	touchkey->num_keys = device_property_read_u32_array(&client->dev,
+							    "linux,keycodes",
+							    NULL, 0);
+	if (touchkey->num_keys < 0) {
+		/* Default key count */
+		touchkey->num_keys = 2;
+	}
+
+	touchkey->keycodes = devm_kcalloc(&client->dev,
+					  touchkey->num_keys,
+					  sizeof(*touchkey->keycodes),
+					  GFP_KERNEL);
+	if (!touchkey->keycodes)
+		return -ENOMEM;
+
+	error = device_property_read_u32_array(&client->dev, "linux,keycodes",
+					       touchkey->keycodes,
+					       touchkey->num_keys);
+
+	if (error) {
+		dev_warn(&client->dev,
+			 "Failed to read keycodes: %d, using defaults\n",
+			 error);
+
+		/* Default keycodes */
+		touchkey->keycodes[0] = KEY_BACK;
+		touchkey->keycodes[1] = KEY_MENU;
+	}
+
+	error = regulator_bulk_enable(ARRAY_SIZE(touchkey->regulators),
+				      touchkey->regulators);
+	if (error) {
+		dev_err(&client->dev,
+			"Failed to enable regulators: %d\n", error);
+		return error;
+	}
+
+	touchkey->input_dev = devm_input_allocate_device(&client->dev);
+	if (!touchkey->input_dev) {
+		dev_err(&client->dev, "Failed to allocate input device\n");
+		return -ENOMEM;
+	}
+
+	touchkey->input_dev->name = CYPRESS_SF_DEV_NAME;
+	touchkey->input_dev->id.bustype = BUS_I2C;
+
+	for (key = 0; key < touchkey->num_keys; ++key)
+		input_set_capability(touchkey->input_dev,
+				     EV_KEY, touchkey->keycodes[key]);
+
+	error = input_register_device(touchkey->input_dev);
+	if (error) {
+		dev_err(&client->dev,
+			"Failed to register input device: %d\n", error);
+		return error;
+	}
+
+	error = devm_request_threaded_irq(&client->dev, client->irq,
+					  NULL, cypress_sf_irq_handler,
+					  IRQF_ONESHOT,
+					  CYPRESS_SF_DEV_NAME, touchkey);
+	if (error) {
+		dev_err(&client->dev,
+			"Failed to register threaded irq: %d", error);
+		return error;
+	}
+
+	return 0;
+};
+
+static int __maybe_unused cypress_sf_suspend(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct cypress_sf_data *touchkey = i2c_get_clientdata(client);
+	int error;
+
+	disable_irq(client->irq);
+
+	error = regulator_bulk_disable(ARRAY_SIZE(touchkey->regulators),
+				       touchkey->regulators);
+	if (error) {
+		dev_err(dev, "Failed to disable regulators: %d", error);
+		enable_irq(client->irq);
+		return error;
+	}
+
+	return 0;
+}
+
+static int __maybe_unused cypress_sf_resume(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct cypress_sf_data *touchkey = i2c_get_clientdata(client);
+	int error;
+
+	error = regulator_bulk_enable(ARRAY_SIZE(touchkey->regulators),
+				      touchkey->regulators);
+	if (error) {
+		dev_err(dev, "Failed to enable regulators: %d", error);
+		return error;
+	}
+
+	enable_irq(client->irq);
+
+	return 0;
+}
+
+static SIMPLE_DEV_PM_OPS(cypress_sf_pm_ops,
+			 cypress_sf_suspend, cypress_sf_resume);
+
+static struct i2c_device_id cypress_sf_id_table[] = {
+	{ CYPRESS_SF_DEV_NAME, 0 },
+	{ }
+};
+MODULE_DEVICE_TABLE(i2c, cypress_sf_id_table);
+
+#ifdef CONFIG_OF
+static const struct of_device_id cypress_sf_of_match[] = {
+	{ .compatible = "cypress,sf3155", },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, cypress_sf_of_match);
+#endif
+
+static struct i2c_driver cypress_sf_driver = {
+	.driver = {
+		.name = CYPRESS_SF_DEV_NAME,
+		.pm = &cypress_sf_pm_ops,
+		.of_match_table = of_match_ptr(cypress_sf_of_match),
+	},
+	.id_table = cypress_sf_id_table,
+	.probe_new = cypress_sf_probe,
+};
+module_i2c_driver(cypress_sf_driver);
+
+MODULE_AUTHOR("Yassine Oudjana <y.oudjana@protonmail.com>");
+MODULE_DESCRIPTION("Cypress StreetFighter Touchkey Driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/input/keyboard/ep93xx_keypad.c b/drivers/input/keyboard/ep93xx_keypad.c
index e0e931e..272a4f1 100644
--- a/drivers/input/keyboard/ep93xx_keypad.c
+++ b/drivers/input/keyboard/ep93xx_keypad.c
@@ -17,6 +17,7 @@
  * flag.
  */
 
+#include <linux/bits.h>
 #include <linux/module.h>
 #include <linux/platform_device.h>
 #include <linux/interrupt.h>
@@ -26,6 +27,7 @@
 #include <linux/slab.h>
 #include <linux/soc/cirrus/ep93xx.h>
 #include <linux/platform_data/keypad-ep93xx.h>
+#include <linux/pm_wakeirq.h>
 
 /*
  * Keypad Interface Register offsets
@@ -35,28 +37,28 @@
 #define KEY_REG			0x08	/* Key Value Capture register */
 
 /* Key Scan Initialization Register bit defines */
-#define KEY_INIT_DBNC_MASK	(0x00ff0000)
-#define KEY_INIT_DBNC_SHIFT	(16)
-#define KEY_INIT_DIS3KY		(1<<15)
-#define KEY_INIT_DIAG		(1<<14)
-#define KEY_INIT_BACK		(1<<13)
-#define KEY_INIT_T2		(1<<12)
-#define KEY_INIT_PRSCL_MASK	(0x000003ff)
-#define KEY_INIT_PRSCL_SHIFT	(0)
+#define KEY_INIT_DBNC_MASK	GENMASK(23, 16)
+#define KEY_INIT_DBNC_SHIFT	16
+#define KEY_INIT_DIS3KY		BIT(15)
+#define KEY_INIT_DIAG		BIT(14)
+#define KEY_INIT_BACK		BIT(13)
+#define KEY_INIT_T2		BIT(12)
+#define KEY_INIT_PRSCL_MASK	GENMASK(9, 0)
+#define KEY_INIT_PRSCL_SHIFT	0
 
 /* Key Scan Diagnostic Register bit defines */
-#define KEY_DIAG_MASK		(0x0000003f)
-#define KEY_DIAG_SHIFT		(0)
+#define KEY_DIAG_MASK		GENMASK(5, 0)
+#define KEY_DIAG_SHIFT		0
 
 /* Key Value Capture Register bit defines */
-#define KEY_REG_K		(1<<15)
-#define KEY_REG_INT		(1<<14)
-#define KEY_REG_2KEYS		(1<<13)
-#define KEY_REG_1KEY		(1<<12)
-#define KEY_REG_KEY2_MASK	(0x00000fc0)
-#define KEY_REG_KEY2_SHIFT	(6)
-#define KEY_REG_KEY1_MASK	(0x0000003f)
-#define KEY_REG_KEY1_SHIFT	(0)
+#define KEY_REG_K		BIT(15)
+#define KEY_REG_INT		BIT(14)
+#define KEY_REG_2KEYS		BIT(13)
+#define KEY_REG_1KEY		BIT(12)
+#define KEY_REG_KEY2_MASK	GENMASK(11, 6)
+#define KEY_REG_KEY2_SHIFT	6
+#define KEY_REG_KEY1_MASK	GENMASK(5, 0)
+#define KEY_REG_KEY1_SHIFT	0
 
 #define EP93XX_MATRIX_SIZE	(EP93XX_MATRIX_ROWS * EP93XX_MATRIX_COLS)
 
@@ -175,8 +177,7 @@ static void ep93xx_keypad_close(struct input_dev *pdev)
 }
 
 
-#ifdef CONFIG_PM_SLEEP
-static int ep93xx_keypad_suspend(struct device *dev)
+static int __maybe_unused ep93xx_keypad_suspend(struct device *dev)
 {
 	struct platform_device *pdev = to_platform_device(dev);
 	struct ep93xx_keypad *keypad = platform_get_drvdata(pdev);
@@ -191,21 +192,15 @@ static int ep93xx_keypad_suspend(struct device *dev)
 
 	mutex_unlock(&input_dev->mutex);
 
-	if (device_may_wakeup(&pdev->dev))
-		enable_irq_wake(keypad->irq);
-
 	return 0;
 }
 
-static int ep93xx_keypad_resume(struct device *dev)
+static int __maybe_unused ep93xx_keypad_resume(struct device *dev)
 {
 	struct platform_device *pdev = to_platform_device(dev);
 	struct ep93xx_keypad *keypad = platform_get_drvdata(pdev);
 	struct input_dev *input_dev = keypad->input_dev;
 
-	if (device_may_wakeup(&pdev->dev))
-		disable_irq_wake(keypad->irq);
-
 	mutex_lock(&input_dev->mutex);
 
 	if (input_device_enabled(input_dev)) {
@@ -220,11 +215,17 @@ static int ep93xx_keypad_resume(struct device *dev)
 
 	return 0;
 }
-#endif
 
 static SIMPLE_DEV_PM_OPS(ep93xx_keypad_pm_ops,
 			 ep93xx_keypad_suspend, ep93xx_keypad_resume);
 
+static void ep93xx_keypad_release_gpio_action(void *_pdev)
+{
+	struct platform_device *pdev = _pdev;
+
+	ep93xx_keypad_release_gpio(pdev);
+}
+
 static int ep93xx_keypad_probe(struct platform_device *pdev)
 {
 	struct ep93xx_keypad *keypad;
@@ -233,61 +234,46 @@ static int ep93xx_keypad_probe(struct platform_device *pdev)
 	struct resource *res;
 	int err;
 
-	keypad = kzalloc(sizeof(struct ep93xx_keypad), GFP_KERNEL);
+	keypad = devm_kzalloc(&pdev->dev, sizeof(*keypad), GFP_KERNEL);
 	if (!keypad)
 		return -ENOMEM;
 
 	keypad->pdata = dev_get_platdata(&pdev->dev);
-	if (!keypad->pdata) {
-		err = -EINVAL;
-		goto failed_free;
-	}
+	if (!keypad->pdata)
+		return -EINVAL;
 
 	keymap_data = keypad->pdata->keymap_data;
-	if (!keymap_data) {
-		err = -EINVAL;
-		goto failed_free;
-	}
+	if (!keymap_data)
+		return -EINVAL;
 
 	keypad->irq = platform_get_irq(pdev, 0);
-	if (keypad->irq < 0) {
-		err = keypad->irq;
-		goto failed_free;
-	}
+	if (keypad->irq < 0)
+		return keypad->irq;
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	if (!res) {
-		err = -ENXIO;
-		goto failed_free;
-	}
+	if (!res)
+		return -ENXIO;
 
-	res = request_mem_region(res->start, resource_size(res), pdev->name);
-	if (!res) {
-		err = -EBUSY;
-		goto failed_free;
-	}
-
-	keypad->mmio_base = ioremap(res->start, resource_size(res));
-	if (keypad->mmio_base == NULL) {
-		err = -ENXIO;
-		goto failed_free_mem;
-	}
+	keypad->mmio_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(keypad->mmio_base))
+		return PTR_ERR(keypad->mmio_base);
 
 	err = ep93xx_keypad_acquire_gpio(pdev);
 	if (err)
-		goto failed_free_io;
+		return err;
 
-	keypad->clk = clk_get(&pdev->dev, NULL);
-	if (IS_ERR(keypad->clk)) {
-		err = PTR_ERR(keypad->clk);
-		goto failed_free_gpio;
-	}
+	err = devm_add_action_or_reset(&pdev->dev,
+				       ep93xx_keypad_release_gpio_action, pdev);
+	if (err)
+		return err;
 
-	input_dev = input_allocate_device();
-	if (!input_dev) {
-		err = -ENOMEM;
-		goto failed_put_clk;
-	}
+	keypad->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(keypad->clk))
+		return PTR_ERR(keypad->clk);
+
+	input_dev = devm_input_allocate_device(&pdev->dev);
+	if (!input_dev)
+		return -ENOMEM;
 
 	keypad->input_dev = input_dev;
 
@@ -295,70 +281,40 @@ static int ep93xx_keypad_probe(struct platform_device *pdev)
 	input_dev->id.bustype = BUS_HOST;
 	input_dev->open = ep93xx_keypad_open;
 	input_dev->close = ep93xx_keypad_close;
-	input_dev->dev.parent = &pdev->dev;
 
 	err = matrix_keypad_build_keymap(keymap_data, NULL,
 					 EP93XX_MATRIX_ROWS, EP93XX_MATRIX_COLS,
 					 keypad->keycodes, input_dev);
 	if (err)
-		goto failed_free_dev;
+		return err;
 
 	if (keypad->pdata->flags & EP93XX_KEYPAD_AUTOREPEAT)
 		__set_bit(EV_REP, input_dev->evbit);
 	input_set_drvdata(input_dev, keypad);
 
-	err = request_irq(keypad->irq, ep93xx_keypad_irq_handler,
-			  0, pdev->name, keypad);
+	err = devm_request_irq(&pdev->dev, keypad->irq,
+			       ep93xx_keypad_irq_handler,
+			       0, pdev->name, keypad);
 	if (err)
-		goto failed_free_dev;
+		return err;
 
 	err = input_register_device(input_dev);
 	if (err)
-		goto failed_free_irq;
+		return err;
 
 	platform_set_drvdata(pdev, keypad);
+
 	device_init_wakeup(&pdev->dev, 1);
+	err = dev_pm_set_wake_irq(&pdev->dev, keypad->irq);
+	if (err)
+		dev_warn(&pdev->dev, "failed to set up wakeup irq: %d\n", err);
 
 	return 0;
-
-failed_free_irq:
-	free_irq(keypad->irq, keypad);
-failed_free_dev:
-	input_free_device(input_dev);
-failed_put_clk:
-	clk_put(keypad->clk);
-failed_free_gpio:
-	ep93xx_keypad_release_gpio(pdev);
-failed_free_io:
-	iounmap(keypad->mmio_base);
-failed_free_mem:
-	release_mem_region(res->start, resource_size(res));
-failed_free:
-	kfree(keypad);
-	return err;
 }
 
 static int ep93xx_keypad_remove(struct platform_device *pdev)
 {
-	struct ep93xx_keypad *keypad = platform_get_drvdata(pdev);
-	struct resource *res;
-
-	free_irq(keypad->irq, keypad);
-
-	if (keypad->enabled)
-		clk_disable(keypad->clk);
-	clk_put(keypad->clk);
-
-	input_unregister_device(keypad->input_dev);
-
-	ep93xx_keypad_release_gpio(pdev);
-
-	iounmap(keypad->mmio_base);
-
-	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	release_mem_region(res->start, resource_size(res));
-
-	kfree(keypad);
+	dev_pm_clear_wake_irq(&pdev->dev);
 
 	return 0;
 }
diff --git a/drivers/input/keyboard/mpr121_touchkey.c b/drivers/input/keyboard/mpr121_touchkey.c
index 40d6e50..230ab3d 100644
--- a/drivers/input/keyboard/mpr121_touchkey.c
+++ b/drivers/input/keyboard/mpr121_touchkey.c
@@ -107,9 +107,9 @@ static struct regulator *mpr121_vdd_supply_init(struct device *dev)
 		return ERR_PTR(err);
 	}
 
-	err = devm_add_action(dev, mpr121_vdd_supply_disable, vdd_supply);
+	err = devm_add_action_or_reset(dev, mpr121_vdd_supply_disable,
+				       vdd_supply);
 	if (err) {
-		regulator_disable(vdd_supply);
 		dev_err(dev, "failed to add disable regulator action: %d\n",
 			err);
 		return ERR_PTR(err);
diff --git a/drivers/input/keyboard/omap-keypad.c b/drivers/input/keyboard/omap-keypad.c
index dbe836c..eb3a687 100644
--- a/drivers/input/keyboard/omap-keypad.c
+++ b/drivers/input/keyboard/omap-keypad.c
@@ -190,8 +190,7 @@ static int omap_kp_probe(struct platform_device *pdev)
 	row_shift = get_count_order(pdata->cols);
 	keycodemax = pdata->rows << row_shift;
 
-	omap_kp = kzalloc(sizeof(struct omap_kp) +
-			keycodemax * sizeof(unsigned short), GFP_KERNEL);
+	omap_kp = kzalloc(struct_size(omap_kp, keymap, keycodemax), GFP_KERNEL);
 	input_dev = input_allocate_device();
 	if (!omap_kp || !input_dev) {
 		kfree(omap_kp);
diff --git a/drivers/input/keyboard/tm2-touchkey.c b/drivers/input/keyboard/tm2-touchkey.c
index 6218b1c..632cd6c 100644
--- a/drivers/input/keyboard/tm2-touchkey.c
+++ b/drivers/input/keyboard/tm2-touchkey.c
@@ -156,6 +156,8 @@ static irqreturn_t tm2_touchkey_irq_handler(int irq, void *devid)
 		goto out;
 	}
 
+	input_event(touchkey->input_dev, EV_MSC, MSC_SCAN, index);
+
 	if (data & TM2_TOUCHKEY_BIT_PRESS_EV) {
 		for (i = 0; i < touchkey->num_keycodes; i++)
 			input_report_key(touchkey->input_dev,
@@ -250,6 +252,11 @@ static int tm2_touchkey_probe(struct i2c_client *client,
 	touchkey->input_dev->name = TM2_TOUCHKEY_DEV_NAME;
 	touchkey->input_dev->id.bustype = BUS_I2C;
 
+	touchkey->input_dev->keycode = touchkey->keycodes;
+	touchkey->input_dev->keycodemax = touchkey->num_keycodes;
+	touchkey->input_dev->keycodesize = sizeof(touchkey->keycodes[0]);
+
+	input_set_capability(touchkey->input_dev, EV_MSC, MSC_SCAN);
 	for (i = 0; i < touchkey->num_keycodes; i++)
 		input_set_capability(touchkey->input_dev, EV_KEY,
 				     touchkey->keycodes[i]);
diff --git a/drivers/input/misc/adxl34x-i2c.c b/drivers/input/misc/adxl34x-i2c.c
index e64368a..a3b5f88 100644
--- a/drivers/input/misc/adxl34x-i2c.c
+++ b/drivers/input/misc/adxl34x-i2c.c
@@ -103,7 +103,9 @@ static int adxl34x_i2c_remove(struct i2c_client *client)
 {
 	struct adxl34x *ac = i2c_get_clientdata(client);
 
-	return adxl34x_remove(ac);
+	adxl34x_remove(ac);
+
+	return 0;
 }
 
 static int __maybe_unused adxl34x_i2c_suspend(struct device *dev)
diff --git a/drivers/input/misc/adxl34x-spi.c b/drivers/input/misc/adxl34x-spi.c
index df6afa4..6e51c9bc 100644
--- a/drivers/input/misc/adxl34x-spi.c
+++ b/drivers/input/misc/adxl34x-spi.c
@@ -91,7 +91,9 @@ static int adxl34x_spi_remove(struct spi_device *spi)
 {
 	struct adxl34x *ac = spi_get_drvdata(spi);
 
-	return adxl34x_remove(ac);
+	adxl34x_remove(ac);
+
+	return 0;
 }
 
 static int __maybe_unused adxl34x_spi_suspend(struct device *dev)
diff --git a/drivers/input/misc/adxl34x.c b/drivers/input/misc/adxl34x.c
index 4cc4e8f..a4af314 100644
--- a/drivers/input/misc/adxl34x.c
+++ b/drivers/input/misc/adxl34x.c
@@ -237,7 +237,7 @@ static const struct adxl34x_platform_data adxl34x_default_init = {
 
 static void adxl34x_get_triple(struct adxl34x *ac, struct axis_triple *axis)
 {
-	short buf[3];
+	__le16 buf[3];
 
 	ac->bops->read_block(ac->dev, DATAX0, DATAZ1 - DATAX0 + 1, buf);
 
@@ -896,15 +896,13 @@ struct adxl34x *adxl34x_probe(struct device *dev, int irq,
 }
 EXPORT_SYMBOL_GPL(adxl34x_probe);
 
-int adxl34x_remove(struct adxl34x *ac)
+void adxl34x_remove(struct adxl34x *ac)
 {
 	sysfs_remove_group(&ac->dev->kobj, &adxl34x_attr_group);
 	free_irq(ac->irq, ac);
 	input_unregister_device(ac->input);
 	dev_dbg(ac->dev, "unregistered accelerometer\n");
 	kfree(ac);
-
-	return 0;
 }
 EXPORT_SYMBOL_GPL(adxl34x_remove);
 
diff --git a/drivers/input/misc/adxl34x.h b/drivers/input/misc/adxl34x.h
index 83a0eec..febf852 100644
--- a/drivers/input/misc/adxl34x.h
+++ b/drivers/input/misc/adxl34x.h
@@ -25,6 +25,6 @@ void adxl34x_resume(struct adxl34x *ac);
 struct adxl34x *adxl34x_probe(struct device *dev, int irq,
 			      bool fifo_delay_default,
 			      const struct adxl34x_bus_ops *bops);
-int adxl34x_remove(struct adxl34x *ac);
+void adxl34x_remove(struct adxl34x *ac);
 
 #endif
diff --git a/drivers/input/misc/ariel-pwrbutton.c b/drivers/input/misc/ariel-pwrbutton.c
index 17bbaac..cdc8071 100644
--- a/drivers/input/misc/ariel-pwrbutton.c
+++ b/drivers/input/misc/ariel-pwrbutton.c
@@ -149,12 +149,19 @@ static const struct of_device_id ariel_pwrbutton_of_match[] = {
 };
 MODULE_DEVICE_TABLE(of, ariel_pwrbutton_of_match);
 
+static const struct spi_device_id ariel_pwrbutton_spi_ids[] = {
+	{ .name = "wyse-ariel-ec-input" },
+	{ }
+};
+MODULE_DEVICE_TABLE(spi, ariel_pwrbutton_spi_ids);
+
 static struct spi_driver ariel_pwrbutton_driver = {
 	.driver = {
 		.name = "dell-wyse-ariel-ec-input",
 		.of_match_table = ariel_pwrbutton_of_match,
 	},
 	.probe = ariel_pwrbutton_probe,
+	.id_table = ariel_pwrbutton_spi_ids,
 };
 module_spi_driver(ariel_pwrbutton_driver);
 
diff --git a/drivers/input/misc/cpcap-pwrbutton.c b/drivers/input/misc/cpcap-pwrbutton.c
index 0abef63..879790b 100644
--- a/drivers/input/misc/cpcap-pwrbutton.c
+++ b/drivers/input/misc/cpcap-pwrbutton.c
@@ -54,9 +54,13 @@ static irqreturn_t powerbutton_irq(int irq, void *_button)
 static int cpcap_power_button_probe(struct platform_device *pdev)
 {
 	struct cpcap_power_button *button;
-	int irq = platform_get_irq(pdev, 0);
+	int irq;
 	int err;
 
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
 	button = devm_kmalloc(&pdev->dev, sizeof(*button), GFP_KERNEL);
 	if (!button)
 		return -ENOMEM;
@@ -73,7 +77,6 @@ static int cpcap_power_button_probe(struct platform_device *pdev)
 
 	button->idev->name = "cpcap-pwrbutton";
 	button->idev->phys = "cpcap-pwrbutton/input0";
-	button->idev->dev.parent = button->dev;
 	input_set_capability(button->idev, EV_KEY, KEY_POWER);
 
 	err = devm_request_threaded_irq(&pdev->dev, irq, NULL,
diff --git a/drivers/input/misc/max77693-haptic.c b/drivers/input/misc/max77693-haptic.c
index 0d09ffe..4369d3c 100644
--- a/drivers/input/misc/max77693-haptic.c
+++ b/drivers/input/misc/max77693-haptic.c
@@ -424,5 +424,4 @@ module_platform_driver(max77693_haptic_driver);
 MODULE_AUTHOR("Jaewon Kim <jaewon02.kim@samsung.com>");
 MODULE_AUTHOR("Krzysztof Kozlowski <krzk@kernel.org>");
 MODULE_DESCRIPTION("MAXIM 77693/77843 Haptic driver");
-MODULE_ALIAS("platform:max77693-haptic");
 MODULE_LICENSE("GPL");
diff --git a/drivers/input/misc/max8925_onkey.c b/drivers/input/misc/max8925_onkey.c
index ffab4a4..4770cb5 100644
--- a/drivers/input/misc/max8925_onkey.c
+++ b/drivers/input/misc/max8925_onkey.c
@@ -1,4 +1,4 @@
-/**
+/*
  * MAX8925 ONKEY driver
  *
  * Copyright (C) 2009 Marvell International Ltd.
diff --git a/drivers/input/misc/palmas-pwrbutton.c b/drivers/input/misc/palmas-pwrbutton.c
index 1e1baed..f9b05cf 100644
--- a/drivers/input/misc/palmas-pwrbutton.c
+++ b/drivers/input/misc/palmas-pwrbutton.c
@@ -210,6 +210,11 @@ static int palmas_pwron_probe(struct platform_device *pdev)
 	INIT_DELAYED_WORK(&pwron->input_work, palmas_power_button_work);
 
 	pwron->irq = platform_get_irq(pdev, 0);
+	if (pwron->irq < 0) {
+		error = pwron->irq;
+		goto err_free_input;
+	}
+
 	error = request_threaded_irq(pwron->irq, NULL, pwron_irq,
 				     IRQF_TRIGGER_HIGH |
 					IRQF_TRIGGER_LOW |
diff --git a/drivers/input/misc/pm8941-pwrkey.c b/drivers/input/misc/pm8941-pwrkey.c
index 3360960..89af524 100644
--- a/drivers/input/misc/pm8941-pwrkey.c
+++ b/drivers/input/misc/pm8941-pwrkey.c
@@ -29,6 +29,7 @@
 #define PON_PS_HOLD_RST_CTL2		0x5b
 #define  PON_PS_HOLD_ENABLE		BIT(7)
 #define  PON_PS_HOLD_TYPE_MASK		0x0f
+#define  PON_PS_HOLD_TYPE_WARM_RESET	1
 #define  PON_PS_HOLD_TYPE_SHUTDOWN	4
 #define  PON_PS_HOLD_TYPE_HARD_RESET	7
 
@@ -99,7 +100,10 @@ static int pm8941_reboot_notify(struct notifier_block *nb,
 		break;
 	case SYS_RESTART:
 	default:
-		reset_type = PON_PS_HOLD_TYPE_HARD_RESET;
+		if (reboot_mode == REBOOT_WARM)
+			reset_type = PON_PS_HOLD_TYPE_WARM_RESET;
+		else
+			reset_type = PON_PS_HOLD_TYPE_HARD_RESET;
 		break;
 	}
 
diff --git a/drivers/input/mouse/elantech.c b/drivers/input/mouse/elantech.c
index 2d0bc02..956d9cd 100644
--- a/drivers/input/mouse/elantech.c
+++ b/drivers/input/mouse/elantech.c
@@ -517,6 +517,19 @@ static void elantech_report_trackpoint(struct psmouse *psmouse,
 	case 0x16008020U:
 	case 0x26800010U:
 	case 0x36808000U:
+
+		/*
+		 * This firmware misreport coordinates for trackpoint
+		 * occasionally. Discard packets outside of [-127, 127] range
+		 * to prevent cursor jumps.
+		 */
+		if (packet[4] == 0x80 || packet[5] == 0x80 ||
+		    packet[1] >> 7 == packet[4] >> 7 ||
+		    packet[2] >> 7 == packet[5] >> 7) {
+			elantech_debug("discarding packet [%6ph]\n", packet);
+			break;
+
+		}
 		x = packet[4] - (int)((packet[1]^0x80) << 1);
 		y = (int)((packet[2]^0x80) << 1) - packet[5];
 
diff --git a/drivers/input/rmi4/rmi_bus.c b/drivers/input/rmi4/rmi_bus.c
index 24f31a5..50a0134 100644
--- a/drivers/input/rmi4/rmi_bus.c
+++ b/drivers/input/rmi4/rmi_bus.c
@@ -90,6 +90,7 @@ int rmi_register_transport_device(struct rmi_transport_dev *xport)
 
 	rmi_dev->dev.bus = &rmi_bus_type;
 	rmi_dev->dev.type = &rmi_device_type;
+	rmi_dev->dev.parent = xport->dev;
 
 	xport->rmi_dev = rmi_dev;
 
diff --git a/drivers/input/serio/i8042-x86ia64io.h b/drivers/input/serio/i8042-x86ia64io.h
index a5a0035..aedd055 100644
--- a/drivers/input/serio/i8042-x86ia64io.h
+++ b/drivers/input/serio/i8042-x86ia64io.h
@@ -273,6 +273,13 @@ static const struct dmi_system_id __initconst i8042_dmi_nomux_table[] = {
 		},
 	},
 	{
+		/* Fujitsu Lifebook T725 laptop */
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "FUJITSU"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "LIFEBOOK T725"),
+		},
+	},
+	{
 		/* Fujitsu Lifebook U745 */
 		.matches = {
 			DMI_MATCH(DMI_SYS_VENDOR, "FUJITSU"),
@@ -841,6 +848,13 @@ static const struct dmi_system_id __initconst i8042_dmi_notimeout_table[] = {
 		},
 	},
 	{
+		/* Fujitsu Lifebook T725 laptop */
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "FUJITSU"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "LIFEBOOK T725"),
+		},
+	},
+	{
 		/* Fujitsu U574 laptop */
 		/* https://bugzilla.kernel.org/show_bug.cgi?id=69731 */
 		.matches = {
diff --git a/drivers/input/touchscreen/Kconfig b/drivers/input/touchscreen/Kconfig
index d4e7473..2f6adfb 100644
--- a/drivers/input/touchscreen/Kconfig
+++ b/drivers/input/touchscreen/Kconfig
@@ -425,6 +425,7 @@
 config TOUCHSCREEN_ILI210X
 	tristate "Ilitek ILI210X based touchscreen"
 	depends on I2C
+	select CRC_CCITT
 	help
 	  Say Y here if you have a ILI210X based touchscreen
 	  controller. This driver supports models ILI2102,
diff --git a/drivers/input/touchscreen/Makefile b/drivers/input/touchscreen/Makefile
index 7d34100..39a8127 100644
--- a/drivers/input/touchscreen/Makefile
+++ b/drivers/input/touchscreen/Makefile
@@ -6,6 +6,7 @@
 # Each configuration option enables a list of files.
 
 wm97xx-ts-y := wm97xx-core.o
+goodix_ts-y := goodix.o goodix_fwupload.o
 
 obj-$(CONFIG_TOUCHSCREEN_88PM860X)	+= 88pm860x-ts.o
 obj-$(CONFIG_TOUCHSCREEN_AD7877)	+= ad7877.o
@@ -44,7 +45,7 @@
 obj-$(CONFIG_TOUCHSCREEN_EGALAX_SERIAL)	+= egalax_ts_serial.o
 obj-$(CONFIG_TOUCHSCREEN_EXC3000)	+= exc3000.o
 obj-$(CONFIG_TOUCHSCREEN_FUJITSU)	+= fujitsu_ts.o
-obj-$(CONFIG_TOUCHSCREEN_GOODIX)	+= goodix.o
+obj-$(CONFIG_TOUCHSCREEN_GOODIX)	+= goodix_ts.o
 obj-$(CONFIG_TOUCHSCREEN_HIDEEP)	+= hideep.o
 obj-$(CONFIG_TOUCHSCREEN_ILI210X)	+= ili210x.o
 obj-$(CONFIG_TOUCHSCREEN_ILITEK)	+= ilitek_ts_i2c.o
diff --git a/drivers/input/touchscreen/ads7846.c b/drivers/input/touchscreen/ads7846.c
index f113a27..a25a77d 100644
--- a/drivers/input/touchscreen/ads7846.c
+++ b/drivers/input/touchscreen/ads7846.c
@@ -101,10 +101,6 @@ struct ads7846 {
 	struct spi_device	*spi;
 	struct regulator	*reg;
 
-#if IS_ENABLED(CONFIG_HWMON)
-	struct device		*hwmon;
-#endif
-
 	u16			model;
 	u16			vref_mv;
 	u16			vref_delay_usecs;
@@ -142,13 +138,18 @@ struct ads7846 {
 
 	int			(*filter)(void *data, int data_idx, int *val);
 	void			*filter_data;
-	void			(*filter_cleanup)(void *data);
 	int			(*get_pendown_state)(void);
 	int			gpio_pendown;
 
 	void			(*wait_for_sync)(void);
 };
 
+enum ads7846_filter {
+	ADS7846_FILTER_OK,
+	ADS7846_FILTER_REPEAT,
+	ADS7846_FILTER_IGNORE,
+};
+
 /* leave chip selected when we're done, for quicker re-select? */
 #if	0
 #define	CS_CHANGE(xfer)	((xfer).cs_change = 1)
@@ -549,6 +550,8 @@ __ATTRIBUTE_GROUPS(ads7846_attr);
 
 static int ads784x_hwmon_register(struct spi_device *spi, struct ads7846 *ts)
 {
+	struct device *hwmon;
+
 	/* hwmon sensors need a reference voltage */
 	switch (ts->model) {
 	case 7846:
@@ -569,17 +572,11 @@ static int ads784x_hwmon_register(struct spi_device *spi, struct ads7846 *ts)
 		break;
 	}
 
-	ts->hwmon = hwmon_device_register_with_groups(&spi->dev, spi->modalias,
-						      ts, ads7846_attr_groups);
+	hwmon = devm_hwmon_device_register_with_groups(&spi->dev,
+						       spi->modalias, ts,
+						       ads7846_attr_groups);
 
-	return PTR_ERR_OR_ZERO(ts->hwmon);
-}
-
-static void ads784x_hwmon_unregister(struct spi_device *spi,
-				     struct ads7846 *ts)
-{
-	if (ts->hwmon)
-		hwmon_device_unregister(ts->hwmon);
+	return PTR_ERR_OR_ZERO(hwmon);
 }
 
 #else
@@ -588,11 +585,6 @@ static inline int ads784x_hwmon_register(struct spi_device *spi,
 {
 	return 0;
 }
-
-static inline void ads784x_hwmon_unregister(struct spi_device *spi,
-					    struct ads7846 *ts)
-{
-}
 #endif
 
 static ssize_t ads7846_pen_down_show(struct device *dev,
@@ -1014,8 +1006,8 @@ static int ads7846_setup_pendown(struct spi_device *spi,
 		ts->get_pendown_state = pdata->get_pendown_state;
 	} else if (gpio_is_valid(pdata->gpio_pendown)) {
 
-		err = gpio_request_one(pdata->gpio_pendown, GPIOF_IN,
-				       "ads7846_pendown");
+		err = devm_gpio_request_one(&spi->dev, pdata->gpio_pendown,
+					    GPIOF_IN, "ads7846_pendown");
 		if (err) {
 			dev_err(&spi->dev,
 				"failed to request/setup pendown GPIO%d: %d\n",
@@ -1212,24 +1204,30 @@ static const struct ads7846_platform_data *ads7846_probe_dt(struct device *dev)
 }
 #endif
 
+static void ads7846_regulator_disable(void *regulator)
+{
+	regulator_disable(regulator);
+}
+
 static int ads7846_probe(struct spi_device *spi)
 {
 	const struct ads7846_platform_data *pdata;
 	struct ads7846 *ts;
+	struct device *dev = &spi->dev;
 	struct ads7846_packet *packet;
 	struct input_dev *input_dev;
 	unsigned long irq_flags;
 	int err;
 
 	if (!spi->irq) {
-		dev_dbg(&spi->dev, "no IRQ?\n");
+		dev_dbg(dev, "no IRQ?\n");
 		return -EINVAL;
 	}
 
 	/* don't exceed max specified sample rate */
 	if (spi->max_speed_hz > (125000 * SAMPLE_BITS)) {
-		dev_err(&spi->dev, "f(sample) %d KHz?\n",
-				(spi->max_speed_hz/SAMPLE_BITS)/1000);
+		dev_err(dev, "f(sample) %d KHz?\n",
+			(spi->max_speed_hz/SAMPLE_BITS)/1000);
 		return -EINVAL;
 	}
 
@@ -1245,13 +1243,17 @@ static int ads7846_probe(struct spi_device *spi)
 	if (err < 0)
 		return err;
 
-	ts = kzalloc(sizeof(struct ads7846), GFP_KERNEL);
-	packet = kzalloc(sizeof(struct ads7846_packet), GFP_KERNEL);
-	input_dev = input_allocate_device();
-	if (!ts || !packet || !input_dev) {
-		err = -ENOMEM;
-		goto err_free_mem;
-	}
+	ts = devm_kzalloc(dev, sizeof(struct ads7846), GFP_KERNEL);
+	if (!ts)
+		return -ENOMEM;
+
+	packet = devm_kzalloc(dev, sizeof(struct ads7846_packet), GFP_KERNEL);
+	if (!packet)
+		return -ENOMEM;
+
+	input_dev = devm_input_allocate_device(dev);
+	if (!input_dev)
+		return -ENOMEM;
 
 	spi_set_drvdata(spi, ts);
 
@@ -1262,13 +1264,11 @@ static int ads7846_probe(struct spi_device *spi)
 	mutex_init(&ts->lock);
 	init_waitqueue_head(&ts->wait);
 
-	pdata = dev_get_platdata(&spi->dev);
+	pdata = dev_get_platdata(dev);
 	if (!pdata) {
-		pdata = ads7846_probe_dt(&spi->dev);
-		if (IS_ERR(pdata)) {
-			err = PTR_ERR(pdata);
-			goto err_free_mem;
-		}
+		pdata = ads7846_probe_dt(dev);
+		if (IS_ERR(pdata))
+			return PTR_ERR(pdata);
 	}
 
 	ts->model = pdata->model ? : 7846;
@@ -1276,15 +1276,7 @@ static int ads7846_probe(struct spi_device *spi)
 	ts->x_plate_ohms = pdata->x_plate_ohms ? : 400;
 	ts->vref_mv = pdata->vref_mv;
 
-	if (pdata->filter != NULL) {
-		if (pdata->filter_init != NULL) {
-			err = pdata->filter_init(pdata, &ts->filter_data);
-			if (err < 0)
-				goto err_free_mem;
-		}
-		ts->filter = pdata->filter;
-		ts->filter_cleanup = pdata->filter_cleanup;
-	} else if (pdata->debounce_max) {
+	if (pdata->debounce_max) {
 		ts->debounce_max = pdata->debounce_max;
 		if (ts->debounce_max < 2)
 			ts->debounce_max = 2;
@@ -1298,7 +1290,7 @@ static int ads7846_probe(struct spi_device *spi)
 
 	err = ads7846_setup_pendown(spi, ts, pdata);
 	if (err)
-		goto err_cleanup_filter;
+		return err;
 
 	if (pdata->penirq_recheck_delay_usecs)
 		ts->penirq_recheck_delay_usecs =
@@ -1306,15 +1298,16 @@ static int ads7846_probe(struct spi_device *spi)
 
 	ts->wait_for_sync = pdata->wait_for_sync ? : null_wait_for_sync;
 
-	snprintf(ts->phys, sizeof(ts->phys), "%s/input0", dev_name(&spi->dev));
+	snprintf(ts->phys, sizeof(ts->phys), "%s/input0", dev_name(dev));
 	snprintf(ts->name, sizeof(ts->name), "ADS%d Touchscreen", ts->model);
 
 	input_dev->name = ts->name;
 	input_dev->phys = ts->phys;
-	input_dev->dev.parent = &spi->dev;
 
-	input_dev->evbit[0] = BIT_MASK(EV_KEY) | BIT_MASK(EV_ABS);
-	input_dev->keybit[BIT_WORD(BTN_TOUCH)] = BIT_MASK(BTN_TOUCH);
+	input_dev->id.bustype = BUS_SPI;
+	input_dev->id.product = pdata->model;
+
+	input_set_capability(input_dev, EV_KEY, BTN_TOUCH);
 	input_set_abs_params(input_dev, ABS_X,
 			pdata->x_min ? : 0,
 			pdata->x_max ? : MAX_12BIT,
@@ -1345,125 +1338,84 @@ static int ads7846_probe(struct spi_device *spi)
 
 	ads7846_setup_spi_msg(ts, pdata);
 
-	ts->reg = regulator_get(&spi->dev, "vcc");
+	ts->reg = devm_regulator_get(dev, "vcc");
 	if (IS_ERR(ts->reg)) {
 		err = PTR_ERR(ts->reg);
-		dev_err(&spi->dev, "unable to get regulator: %d\n", err);
-		goto err_free_gpio;
+		dev_err(dev, "unable to get regulator: %d\n", err);
+		return err;
 	}
 
 	err = regulator_enable(ts->reg);
 	if (err) {
-		dev_err(&spi->dev, "unable to enable regulator: %d\n", err);
-		goto err_put_regulator;
+		dev_err(dev, "unable to enable regulator: %d\n", err);
+		return err;
 	}
 
+	err = devm_add_action_or_reset(dev, ads7846_regulator_disable, ts->reg);
+	if (err)
+		return err;
+
 	irq_flags = pdata->irq_flags ? : IRQF_TRIGGER_FALLING;
 	irq_flags |= IRQF_ONESHOT;
 
-	err = request_threaded_irq(spi->irq, ads7846_hard_irq, ads7846_irq,
-				   irq_flags, spi->dev.driver->name, ts);
-	if (err && !pdata->irq_flags) {
-		dev_info(&spi->dev,
+	err = devm_request_threaded_irq(dev, spi->irq,
+					ads7846_hard_irq, ads7846_irq,
+					irq_flags, dev->driver->name, ts);
+	if (err && err != -EPROBE_DEFER && !pdata->irq_flags) {
+		dev_info(dev,
 			"trying pin change workaround on irq %d\n", spi->irq);
 		irq_flags |= IRQF_TRIGGER_RISING;
-		err = request_threaded_irq(spi->irq,
-				  ads7846_hard_irq, ads7846_irq,
-				  irq_flags, spi->dev.driver->name, ts);
+		err = devm_request_threaded_irq(dev, spi->irq,
+						ads7846_hard_irq, ads7846_irq,
+						irq_flags, dev->driver->name,
+						ts);
 	}
 
 	if (err) {
-		dev_dbg(&spi->dev, "irq %d busy?\n", spi->irq);
-		goto err_disable_regulator;
+		dev_dbg(dev, "irq %d busy?\n", spi->irq);
+		return err;
 	}
 
 	err = ads784x_hwmon_register(spi, ts);
 	if (err)
-		goto err_free_irq;
+		return err;
 
-	dev_info(&spi->dev, "touchscreen, irq %d\n", spi->irq);
+	dev_info(dev, "touchscreen, irq %d\n", spi->irq);
 
 	/*
 	 * Take a first sample, leaving nPENIRQ active and vREF off; avoid
 	 * the touchscreen, in case it's not connected.
 	 */
 	if (ts->model == 7845)
-		ads7845_read12_ser(&spi->dev, PWRDOWN);
+		ads7845_read12_ser(dev, PWRDOWN);
 	else
-		(void) ads7846_read12_ser(&spi->dev, READ_12BIT_SER(vaux));
+		(void) ads7846_read12_ser(dev, READ_12BIT_SER(vaux));
 
-	err = sysfs_create_group(&spi->dev.kobj, &ads784x_attr_group);
+	err = devm_device_add_group(dev, &ads784x_attr_group);
 	if (err)
-		goto err_remove_hwmon;
+		return err;
 
 	err = input_register_device(input_dev);
 	if (err)
-		goto err_remove_attr_group;
+		return err;
 
-	device_init_wakeup(&spi->dev, pdata->wakeup);
+	device_init_wakeup(dev, pdata->wakeup);
 
 	/*
 	 * If device does not carry platform data we must have allocated it
 	 * when parsing DT data.
 	 */
-	if (!dev_get_platdata(&spi->dev))
-		devm_kfree(&spi->dev, (void *)pdata);
+	if (!dev_get_platdata(dev))
+		devm_kfree(dev, (void *)pdata);
 
 	return 0;
-
- err_remove_attr_group:
-	sysfs_remove_group(&spi->dev.kobj, &ads784x_attr_group);
- err_remove_hwmon:
-	ads784x_hwmon_unregister(spi, ts);
- err_free_irq:
-	free_irq(spi->irq, ts);
- err_disable_regulator:
-	regulator_disable(ts->reg);
- err_put_regulator:
-	regulator_put(ts->reg);
- err_free_gpio:
-	if (!ts->get_pendown_state)
-		gpio_free(ts->gpio_pendown);
- err_cleanup_filter:
-	if (ts->filter_cleanup)
-		ts->filter_cleanup(ts->filter_data);
- err_free_mem:
-	input_free_device(input_dev);
-	kfree(packet);
-	kfree(ts);
-	return err;
 }
 
 static int ads7846_remove(struct spi_device *spi)
 {
 	struct ads7846 *ts = spi_get_drvdata(spi);
 
-	sysfs_remove_group(&spi->dev.kobj, &ads784x_attr_group);
-
-	ads7846_disable(ts);
-	free_irq(ts->spi->irq, ts);
-
-	input_unregister_device(ts->input);
-
-	ads784x_hwmon_unregister(spi, ts);
-
-	regulator_put(ts->reg);
-
-	if (!ts->get_pendown_state) {
-		/*
-		 * If we are not using specialized pendown method we must
-		 * have been relying on gpio we set up ourselves.
-		 */
-		gpio_free(ts->gpio_pendown);
-	}
-
-	if (ts->filter_cleanup)
-		ts->filter_cleanup(ts->filter_data);
-
-	kfree(ts->packet);
-	kfree(ts);
-
-	dev_dbg(&spi->dev, "unregistered touchscreen\n");
+	ads7846_stop(ts);
 
 	return 0;
 }
diff --git a/drivers/input/touchscreen/elants_i2c.c b/drivers/input/touchscreen/elants_i2c.c
index 68f542b..7e13a66 100644
--- a/drivers/input/touchscreen/elants_i2c.c
+++ b/drivers/input/touchscreen/elants_i2c.c
@@ -1439,11 +1439,11 @@ static int elants_i2c_probe(struct i2c_client *client)
 	if (error)
 		return error;
 
-	error = devm_add_action(&client->dev, elants_i2c_power_off, ts);
+	error = devm_add_action_or_reset(&client->dev,
+					 elants_i2c_power_off, ts);
 	if (error) {
 		dev_err(&client->dev,
 			"failed to install power off action: %d\n", error);
-		elants_i2c_power_off(ts);
 		return error;
 	}
 
diff --git a/drivers/input/touchscreen/goodix.c b/drivers/input/touchscreen/goodix.c
index 4f53d3c..b5cc917 100644
--- a/drivers/input/touchscreen/goodix.c
+++ b/drivers/input/touchscreen/goodix.c
@@ -14,20 +14,15 @@
 #include <linux/kernel.h>
 #include <linux/dmi.h>
 #include <linux/firmware.h>
-#include <linux/gpio/consumer.h>
-#include <linux/i2c.h>
-#include <linux/input.h>
-#include <linux/input/mt.h>
-#include <linux/input/touchscreen.h>
 #include <linux/module.h>
 #include <linux/delay.h>
 #include <linux/irq.h>
 #include <linux/interrupt.h>
-#include <linux/regulator/consumer.h>
 #include <linux/slab.h>
 #include <linux/acpi.h>
 #include <linux/of.h>
 #include <asm/unaligned.h>
+#include "goodix.h"
 
 #define GOODIX_GPIO_INT_NAME		"irq"
 #define GOODIX_GPIO_RST_NAME		"reset"
@@ -38,22 +33,11 @@
 #define GOODIX_CONTACT_SIZE		8
 #define GOODIX_MAX_CONTACT_SIZE		9
 #define GOODIX_MAX_CONTACTS		10
-#define GOODIX_MAX_KEYS			7
 
 #define GOODIX_CONFIG_MIN_LENGTH	186
 #define GOODIX_CONFIG_911_LENGTH	186
 #define GOODIX_CONFIG_967_LENGTH	228
 #define GOODIX_CONFIG_GT9X_LENGTH	240
-#define GOODIX_CONFIG_MAX_LENGTH	240
-
-/* Register defines */
-#define GOODIX_REG_COMMAND		0x8040
-#define GOODIX_CMD_SCREEN_OFF		0x05
-
-#define GOODIX_READ_COOR_ADDR		0x814E
-#define GOODIX_GT1X_REG_CONFIG_DATA	0x8050
-#define GOODIX_GT9X_REG_CONFIG_DATA	0x8047
-#define GOODIX_REG_ID			0x8140
 
 #define GOODIX_BUFFER_STATUS_READY	BIT(7)
 #define GOODIX_HAVE_KEY			BIT(4)
@@ -68,55 +52,11 @@
 #define ACPI_GPIO_SUPPORT
 #endif
 
-struct goodix_ts_data;
-
-enum goodix_irq_pin_access_method {
-	IRQ_PIN_ACCESS_NONE,
-	IRQ_PIN_ACCESS_GPIO,
-	IRQ_PIN_ACCESS_ACPI_GPIO,
-	IRQ_PIN_ACCESS_ACPI_METHOD,
-};
-
-struct goodix_chip_data {
-	u16 config_addr;
-	int config_len;
-	int (*check_config)(struct goodix_ts_data *ts, const u8 *cfg, int len);
-	void (*calc_config_checksum)(struct goodix_ts_data *ts);
-};
-
 struct goodix_chip_id {
 	const char *id;
 	const struct goodix_chip_data *data;
 };
 
-#define GOODIX_ID_MAX_LEN	4
-
-struct goodix_ts_data {
-	struct i2c_client *client;
-	struct input_dev *input_dev;
-	const struct goodix_chip_data *chip;
-	struct touchscreen_properties prop;
-	unsigned int max_touch_num;
-	unsigned int int_trigger_type;
-	struct regulator *avdd28;
-	struct regulator *vddio;
-	struct gpio_desc *gpiod_int;
-	struct gpio_desc *gpiod_rst;
-	int gpio_count;
-	int gpio_int_idx;
-	char id[GOODIX_ID_MAX_LEN + 1];
-	u16 version;
-	const char *cfg_name;
-	bool reset_controller_at_probe;
-	bool load_cfg_from_disk;
-	struct completion firmware_loading_complete;
-	unsigned long irq_flags;
-	enum goodix_irq_pin_access_method irq_pin_access_method;
-	unsigned int contact_size;
-	u8 config[GOODIX_CONFIG_MAX_LENGTH];
-	unsigned short keymap[GOODIX_MAX_KEYS];
-};
-
 static int goodix_check_cfg_8(struct goodix_ts_data *ts,
 			      const u8 *cfg, int len);
 static int goodix_check_cfg_16(struct goodix_ts_data *ts,
@@ -215,8 +155,7 @@ static const struct dmi_system_id inverted_x_screen[] = {
  * @buf: raw write data buffer.
  * @len: length of the buffer to write
  */
-static int goodix_i2c_read(struct i2c_client *client,
-			   u16 reg, u8 *buf, int len)
+int goodix_i2c_read(struct i2c_client *client, u16 reg, u8 *buf, int len)
 {
 	struct i2c_msg msgs[2];
 	__be16 wbuf = cpu_to_be16(reg);
@@ -233,7 +172,13 @@ static int goodix_i2c_read(struct i2c_client *client,
 	msgs[1].buf   = buf;
 
 	ret = i2c_transfer(client->adapter, msgs, 2);
-	return ret < 0 ? ret : (ret != ARRAY_SIZE(msgs) ? -EIO : 0);
+	if (ret >= 0)
+		ret = (ret == ARRAY_SIZE(msgs) ? 0 : -EIO);
+
+	if (ret)
+		dev_err(&client->dev, "Error reading %d bytes from 0x%04x: %d\n",
+			len, reg, ret);
+	return ret;
 }
 
 /**
@@ -244,8 +189,7 @@ static int goodix_i2c_read(struct i2c_client *client,
  * @buf: raw data buffer to write.
  * @len: length of the buffer to write
  */
-static int goodix_i2c_write(struct i2c_client *client, u16 reg, const u8 *buf,
-			    unsigned len)
+int goodix_i2c_write(struct i2c_client *client, u16 reg, const u8 *buf, int len)
 {
 	u8 *addr_buf;
 	struct i2c_msg msg;
@@ -265,11 +209,18 @@ static int goodix_i2c_write(struct i2c_client *client, u16 reg, const u8 *buf,
 	msg.len = len + 2;
 
 	ret = i2c_transfer(client->adapter, &msg, 1);
+	if (ret >= 0)
+		ret = (ret == 1 ? 0 : -EIO);
+
 	kfree(addr_buf);
-	return ret < 0 ? ret : (ret != 1 ? -EIO : 0);
+
+	if (ret)
+		dev_err(&client->dev, "Error writing %d bytes to 0x%04x: %d\n",
+			len, reg, ret);
+	return ret;
 }
 
-static int goodix_i2c_write_u8(struct i2c_client *client, u16 reg, u8 value)
+int goodix_i2c_write_u8(struct i2c_client *client, u16 reg, u8 value)
 {
 	return goodix_i2c_write(client, reg, &value, sizeof(value));
 }
@@ -308,11 +259,8 @@ static int goodix_ts_read_input_report(struct goodix_ts_data *ts, u8 *data)
 	do {
 		error = goodix_i2c_read(ts->client, addr, data,
 					header_contact_keycode_size);
-		if (error) {
-			dev_err(&ts->client->dev, "I2C transfer error: %d\n",
-					error);
+		if (error)
 			return error;
-		}
 
 		if (data[0] & GOODIX_BUFFER_STATUS_READY) {
 			touch_num = data[0] & 0x0f;
@@ -333,6 +281,11 @@ static int goodix_ts_read_input_report(struct goodix_ts_data *ts, u8 *data)
 			return touch_num;
 		}
 
+		if (data[0] == 0 && ts->firmware_name) {
+			if (goodix_handle_fw_request(ts))
+				return 0;
+		}
+
 		usleep_range(1000, 2000); /* Poll every 1 - 2 ms */
 	} while (time_before(jiffies, max_timeout));
 
@@ -435,9 +388,7 @@ static irqreturn_t goodix_ts_irq_handler(int irq, void *dev_id)
 	struct goodix_ts_data *ts = dev_id;
 
 	goodix_process_events(ts);
-
-	if (goodix_i2c_write_u8(ts->client, GOODIX_READ_COOR_ADDR, 0) < 0)
-		dev_err(&ts->client->dev, "I2C write end_cmd error\n");
+	goodix_i2c_write_u8(ts->client, GOODIX_READ_COOR_ADDR, 0);
 
 	return IRQ_HANDLED;
 }
@@ -553,7 +504,7 @@ static int goodix_check_cfg(struct goodix_ts_data *ts, const u8 *cfg, int len)
  * @cfg: config firmware to write to device
  * @len: config data length
  */
-static int goodix_send_cfg(struct goodix_ts_data *ts, const u8 *cfg, int len)
+int goodix_send_cfg(struct goodix_ts_data *ts, const u8 *cfg, int len)
 {
 	int error;
 
@@ -562,11 +513,9 @@ static int goodix_send_cfg(struct goodix_ts_data *ts, const u8 *cfg, int len)
 		return error;
 
 	error = goodix_i2c_write(ts->client, ts->chip->config_addr, cfg, len);
-	if (error) {
-		dev_err(&ts->client->dev, "Failed to write config data: %d",
-			error);
+	if (error)
 		return error;
-	}
+
 	dev_dbg(&ts->client->dev, "Config sent successfully.");
 
 	/* Let the firmware reconfigure itself, so sleep for 10ms */
@@ -651,21 +600,66 @@ static int goodix_irq_direction_input(struct goodix_ts_data *ts)
 	return -EINVAL; /* Never reached */
 }
 
-static int goodix_int_sync(struct goodix_ts_data *ts)
+int goodix_int_sync(struct goodix_ts_data *ts)
 {
 	int error;
 
 	error = goodix_irq_direction_output(ts, 0);
 	if (error)
-		return error;
+		goto error;
 
 	msleep(50);				/* T5: 50ms */
 
 	error = goodix_irq_direction_input(ts);
 	if (error)
-		return error;
+		goto error;
 
 	return 0;
+
+error:
+	dev_err(&ts->client->dev, "Controller irq sync failed.\n");
+	return error;
+}
+
+/**
+ * goodix_reset_no_int_sync - Reset device, leaving interrupt line in output mode
+ *
+ * @ts: goodix_ts_data pointer
+ */
+int goodix_reset_no_int_sync(struct goodix_ts_data *ts)
+{
+	int error;
+
+	/* begin select I2C slave addr */
+	error = gpiod_direction_output(ts->gpiod_rst, 0);
+	if (error)
+		goto error;
+
+	msleep(20);				/* T2: > 10ms */
+
+	/* HIGH: 0x28/0x29, LOW: 0xBA/0xBB */
+	error = goodix_irq_direction_output(ts, ts->client->addr == 0x14);
+	if (error)
+		goto error;
+
+	usleep_range(100, 2000);		/* T3: > 100us */
+
+	error = gpiod_direction_output(ts->gpiod_rst, 1);
+	if (error)
+		goto error;
+
+	usleep_range(6000, 10000);		/* T4: > 5ms */
+
+	/* end select I2C slave addr */
+	error = gpiod_direction_input(ts->gpiod_rst);
+	if (error)
+		goto error;
+
+	return 0;
+
+error:
+	dev_err(&ts->client->dev, "Controller reset failed.\n");
+	return error;
 }
 
 /**
@@ -677,36 +671,11 @@ static int goodix_reset(struct goodix_ts_data *ts)
 {
 	int error;
 
-	/* begin select I2C slave addr */
-	error = gpiod_direction_output(ts->gpiod_rst, 0);
+	error = goodix_reset_no_int_sync(ts);
 	if (error)
 		return error;
 
-	msleep(20);				/* T2: > 10ms */
-
-	/* HIGH: 0x28/0x29, LOW: 0xBA/0xBB */
-	error = goodix_irq_direction_output(ts, ts->client->addr == 0x14);
-	if (error)
-		return error;
-
-	usleep_range(100, 2000);		/* T3: > 100us */
-
-	error = gpiod_direction_output(ts->gpiod_rst, 1);
-	if (error)
-		return error;
-
-	usleep_range(6000, 10000);		/* T4: > 5ms */
-
-	/* end select I2C slave addr */
-	error = gpiod_direction_input(ts->gpiod_rst);
-	if (error)
-		return error;
-
-	error = goodix_int_sync(ts);
-	if (error)
-		return error;
-
-	return 0;
+	return goodix_int_sync(ts);
 }
 
 #ifdef ACPI_GPIO_SUPPORT
@@ -931,14 +900,19 @@ static void goodix_read_config(struct goodix_ts_data *ts)
 	int x_max, y_max;
 	int error;
 
-	error = goodix_i2c_read(ts->client, ts->chip->config_addr,
-				ts->config, ts->chip->config_len);
-	if (error) {
-		dev_warn(&ts->client->dev, "Error reading config: %d\n",
-			 error);
-		ts->int_trigger_type = GOODIX_INT_TRIGGER;
-		ts->max_touch_num = GOODIX_MAX_CONTACTS;
-		return;
+	/*
+	 * On controllers where we need to upload the firmware
+	 * (controllers without flash) ts->config already has the config
+	 * at this point and the controller itself does not have it yet!
+	 */
+	if (!ts->firmware_name) {
+		error = goodix_i2c_read(ts->client, ts->chip->config_addr,
+					ts->config, ts->chip->config_len);
+		if (error) {
+			ts->int_trigger_type = GOODIX_INT_TRIGGER;
+			ts->max_touch_num = GOODIX_MAX_CONTACTS;
+			return;
+		}
 	}
 
 	ts->int_trigger_type = ts->config[TRIGGER_LOC] & 0x03;
@@ -966,10 +940,8 @@ static int goodix_read_version(struct goodix_ts_data *ts)
 	char id_str[GOODIX_ID_MAX_LEN + 1];
 
 	error = goodix_i2c_read(ts->client, GOODIX_REG_ID, buf, sizeof(buf));
-	if (error) {
-		dev_err(&ts->client->dev, "read version failed: %d\n", error);
+	if (error)
 		return error;
-	}
 
 	memcpy(id_str, buf, GOODIX_ID_MAX_LEN);
 	id_str[GOODIX_ID_MAX_LEN] = 0;
@@ -995,13 +967,10 @@ static int goodix_i2c_test(struct i2c_client *client)
 	u8 test;
 
 	while (retry++ < 2) {
-		error = goodix_i2c_read(client, GOODIX_REG_ID,
-					&test, 1);
+		error = goodix_i2c_read(client, GOODIX_REG_ID, &test, 1);
 		if (!error)
 			return 0;
 
-		dev_err(&client->dev, "i2c test failed attempt %d: %d\n",
-			retry, error);
 		msleep(20);
 	}
 
@@ -1130,7 +1099,16 @@ static void goodix_config_cb(const struct firmware *cfg, void *ctx)
 	struct goodix_ts_data *ts = ctx;
 	int error;
 
-	if (cfg) {
+	if (ts->firmware_name) {
+		if (!cfg)
+			goto err_release_cfg;
+
+		error = goodix_check_cfg(ts, cfg->data, cfg->size);
+		if (error)
+			goto err_release_cfg;
+
+		memcpy(ts->config, cfg->data, cfg->size);
+	} else if (cfg) {
 		/* send device configuration to the firmware */
 		error = goodix_send_cfg(ts, cfg->data, cfg->size);
 		if (error)
@@ -1156,6 +1134,7 @@ static int goodix_ts_probe(struct i2c_client *client,
 			   const struct i2c_device_id *id)
 {
 	struct goodix_ts_data *ts;
+	const char *cfg_name;
 	int error;
 
 	dev_dbg(&client->dev, "I2C Address: 0x%02x\n", client->addr);
@@ -1205,10 +1184,8 @@ static int goodix_ts_probe(struct i2c_client *client,
 	if (ts->reset_controller_at_probe) {
 		/* reset the controller */
 		error = goodix_reset(ts);
-		if (error) {
-			dev_err(&client->dev, "Controller reset failed.\n");
+		if (error)
 			return error;
-		}
 	}
 
 	error = goodix_i2c_test(client);
@@ -1223,20 +1200,27 @@ static int goodix_ts_probe(struct i2c_client *client,
 		return error;
 	}
 
-	error = goodix_read_version(ts);
-	if (error) {
-		dev_err(&client->dev, "Read version failed.\n");
+	error = goodix_firmware_check(ts);
+	if (error)
 		return error;
-	}
+
+	error = goodix_read_version(ts);
+	if (error)
+		return error;
 
 	ts->chip = goodix_get_chip_data(ts->id);
 
 	if (ts->load_cfg_from_disk) {
 		/* update device config */
-		ts->cfg_name = devm_kasprintf(&client->dev, GFP_KERNEL,
-					      "goodix_%s_cfg.bin", ts->id);
-		if (!ts->cfg_name)
-			return -ENOMEM;
+		error = device_property_read_string(&client->dev,
+						    "goodix,config-name",
+						    &cfg_name);
+		if (!error)
+			snprintf(ts->cfg_name, sizeof(ts->cfg_name),
+				 "goodix/%s", cfg_name);
+		else
+			snprintf(ts->cfg_name, sizeof(ts->cfg_name),
+				 "goodix_%s_cfg.bin", ts->id);
 
 		error = request_firmware_nowait(THIS_MODULE, true, ts->cfg_name,
 						&client->dev, GFP_KERNEL, ts,
@@ -1286,6 +1270,9 @@ static int __maybe_unused goodix_suspend(struct device *dev)
 	/* Free IRQ as IRQ pin is used as output in the suspend sequence */
 	goodix_free_irq(ts);
 
+	/* Save reference (calibration) info if necessary */
+	goodix_save_bak_ref(ts);
+
 	/* Output LOW on the INT pin for 5 ms */
 	error = goodix_irq_direction_output(ts, 0);
 	if (error) {
@@ -1298,7 +1285,6 @@ static int __maybe_unused goodix_suspend(struct device *dev)
 	error = goodix_i2c_write_u8(ts->client, GOODIX_REG_COMMAND,
 				    GOODIX_CMD_SCREEN_OFF);
 	if (error) {
-		dev_err(&ts->client->dev, "Screen off command failed\n");
 		goodix_irq_direction_input(ts);
 		goodix_request_irq(ts);
 		return -EAGAIN;
@@ -1341,19 +1327,14 @@ static int __maybe_unused goodix_resume(struct device *dev)
 
 	error = goodix_i2c_read(ts->client, ts->chip->config_addr,
 				&config_ver, 1);
-	if (error)
-		dev_warn(dev, "Error reading config version: %d, resetting controller\n",
-			 error);
-	else if (config_ver != ts->config[0])
+	if (!error && config_ver != ts->config[0])
 		dev_info(dev, "Config version mismatch %d != %d, resetting controller\n",
 			 config_ver, ts->config[0]);
 
 	if (error != 0 || config_ver != ts->config[0]) {
 		error = goodix_reset(ts);
-		if (error) {
-			dev_err(dev, "Controller reset failed.\n");
+		if (error)
 			return error;
-		}
 
 		error = goodix_send_cfg(ts, ts->config, ts->chip->config_len);
 		if (error)
diff --git a/drivers/input/touchscreen/goodix.h b/drivers/input/touchscreen/goodix.h
new file mode 100644
index 0000000..62138f9
--- /dev/null
+++ b/drivers/input/touchscreen/goodix.h
@@ -0,0 +1,117 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+#ifndef __GOODIX_H__
+#define __GOODIX_H__
+
+#include <linux/gpio/consumer.h>
+#include <linux/i2c.h>
+#include <linux/input.h>
+#include <linux/input/mt.h>
+#include <linux/input/touchscreen.h>
+#include <linux/regulator/consumer.h>
+
+/* Register defines */
+#define GOODIX_REG_MISCTL_DSP_CTL		0x4010
+#define GOODIX_REG_MISCTL_SRAM_BANK		0x4048
+#define GOODIX_REG_MISCTL_MEM_CD_EN		0x4049
+#define GOODIX_REG_MISCTL_CACHE_EN		0x404B
+#define GOODIX_REG_MISCTL_TMR0_EN		0x40B0
+#define GOODIX_REG_MISCTL_SWRST			0x4180
+#define GOODIX_REG_MISCTL_CPU_SWRST_PULSE	0x4184
+#define GOODIX_REG_MISCTL_BOOTCTL		0x4190
+#define GOODIX_REG_MISCTL_BOOT_OPT		0x4218
+#define GOODIX_REG_MISCTL_BOOT_CTL		0x5094
+
+#define GOODIX_REG_FW_SIG			0x8000
+#define GOODIX_FW_SIG_LEN			10
+
+#define GOODIX_REG_MAIN_CLK			0x8020
+#define GOODIX_MAIN_CLK_LEN			6
+
+#define GOODIX_REG_COMMAND			0x8040
+#define GOODIX_CMD_SCREEN_OFF			0x05
+
+#define GOODIX_REG_SW_WDT			0x8041
+
+#define GOODIX_REG_REQUEST			0x8043
+#define GOODIX_RQST_RESPONDED			0x00
+#define GOODIX_RQST_CONFIG			0x01
+#define GOODIX_RQST_BAK_REF			0x02
+#define GOODIX_RQST_RESET			0x03
+#define GOODIX_RQST_MAIN_CLOCK			0x04
+/*
+ * Unknown request which gets send by the controller aprox.
+ * every 34 seconds once it is up and running.
+ */
+#define GOODIX_RQST_UNKNOWN			0x06
+#define GOODIX_RQST_IDLE			0xFF
+
+#define GOODIX_REG_STATUS			0x8044
+
+#define GOODIX_GT1X_REG_CONFIG_DATA		0x8050
+#define GOODIX_GT9X_REG_CONFIG_DATA		0x8047
+#define GOODIX_REG_ID				0x8140
+#define GOODIX_READ_COOR_ADDR			0x814E
+#define GOODIX_REG_BAK_REF			0x99D0
+
+#define GOODIX_ID_MAX_LEN			4
+#define GOODIX_CONFIG_MAX_LENGTH		240
+#define GOODIX_MAX_KEYS				7
+
+enum goodix_irq_pin_access_method {
+	IRQ_PIN_ACCESS_NONE,
+	IRQ_PIN_ACCESS_GPIO,
+	IRQ_PIN_ACCESS_ACPI_GPIO,
+	IRQ_PIN_ACCESS_ACPI_METHOD,
+};
+
+struct goodix_ts_data;
+
+struct goodix_chip_data {
+	u16 config_addr;
+	int config_len;
+	int (*check_config)(struct goodix_ts_data *ts, const u8 *cfg, int len);
+	void (*calc_config_checksum)(struct goodix_ts_data *ts);
+};
+
+struct goodix_ts_data {
+	struct i2c_client *client;
+	struct input_dev *input_dev;
+	const struct goodix_chip_data *chip;
+	const char *firmware_name;
+	struct touchscreen_properties prop;
+	unsigned int max_touch_num;
+	unsigned int int_trigger_type;
+	struct regulator *avdd28;
+	struct regulator *vddio;
+	struct gpio_desc *gpiod_int;
+	struct gpio_desc *gpiod_rst;
+	int gpio_count;
+	int gpio_int_idx;
+	char id[GOODIX_ID_MAX_LEN + 1];
+	char cfg_name[64];
+	u16 version;
+	bool reset_controller_at_probe;
+	bool load_cfg_from_disk;
+	struct completion firmware_loading_complete;
+	unsigned long irq_flags;
+	enum goodix_irq_pin_access_method irq_pin_access_method;
+	unsigned int contact_size;
+	u8 config[GOODIX_CONFIG_MAX_LENGTH];
+	unsigned short keymap[GOODIX_MAX_KEYS];
+	u8 main_clk[GOODIX_MAIN_CLK_LEN];
+	int bak_ref_len;
+	u8 *bak_ref;
+};
+
+int goodix_i2c_read(struct i2c_client *client, u16 reg, u8 *buf, int len);
+int goodix_i2c_write(struct i2c_client *client, u16 reg, const u8 *buf, int len);
+int goodix_i2c_write_u8(struct i2c_client *client, u16 reg, u8 value);
+int goodix_send_cfg(struct goodix_ts_data *ts, const u8 *cfg, int len);
+int goodix_int_sync(struct goodix_ts_data *ts);
+int goodix_reset_no_int_sync(struct goodix_ts_data *ts);
+
+int goodix_firmware_check(struct goodix_ts_data *ts);
+bool goodix_handle_fw_request(struct goodix_ts_data *ts);
+void goodix_save_bak_ref(struct goodix_ts_data *ts);
+
+#endif
diff --git a/drivers/input/touchscreen/goodix_fwupload.c b/drivers/input/touchscreen/goodix_fwupload.c
new file mode 100644
index 0000000..c1e7a24
--- /dev/null
+++ b/drivers/input/touchscreen/goodix_fwupload.c
@@ -0,0 +1,427 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Goodix Touchscreen firmware upload support
+ *
+ * Copyright (c) 2021 Hans de Goede <hdegoede@redhat.com>
+ *
+ * This is a rewrite of gt9xx_update.c from the Allwinner H3 BSP which is:
+ * Copyright (c) 2010 - 2012 Goodix Technology.
+ * Author: andrew@goodix.com
+ */
+
+#include <linux/device.h>
+#include <linux/firmware.h>
+#include <linux/i2c.h>
+#include "goodix.h"
+
+#define GOODIX_FW_HEADER_LENGTH		sizeof(struct goodix_fw_header)
+#define GOODIX_FW_SECTION_LENGTH	0x2000
+#define GOODIX_FW_DSP_LENGTH		0x1000
+#define GOODIX_FW_UPLOAD_ADDRESS	0xc000
+
+#define GOODIX_CFG_LOC_HAVE_KEY		 7
+#define GOODIX_CFG_LOC_DRVA_NUM		27
+#define GOODIX_CFG_LOC_DRVB_NUM		28
+#define GOODIX_CFG_LOC_SENS_NUM		29
+
+struct goodix_fw_header {
+	u8 hw_info[4];
+	u8 pid[8];
+	u8 vid[2];
+} __packed;
+
+static u16 goodix_firmware_checksum(const u8 *data, int size)
+{
+	u16 checksum = 0;
+	int i;
+
+	for (i = 0; i < size; i += 2)
+		checksum += (data[i] << 8) + data[i + 1];
+
+	return checksum;
+}
+
+static int goodix_firmware_verify(struct device *dev, const struct firmware *fw)
+{
+	const struct goodix_fw_header *fw_header;
+	size_t expected_size;
+	const u8 *data;
+	u16 checksum;
+	char buf[9];
+
+	expected_size = GOODIX_FW_HEADER_LENGTH + 4 * GOODIX_FW_SECTION_LENGTH +
+			GOODIX_FW_DSP_LENGTH;
+	if (fw->size != expected_size) {
+		dev_err(dev, "Firmware has wrong size, expected %zu got %zu\n",
+			expected_size, fw->size);
+		return -EINVAL;
+	}
+
+	data = fw->data + GOODIX_FW_HEADER_LENGTH;
+	checksum = goodix_firmware_checksum(data, 4 * GOODIX_FW_SECTION_LENGTH);
+	if (checksum) {
+		dev_err(dev, "Main firmware checksum error\n");
+		return -EINVAL;
+	}
+
+	data += 4 * GOODIX_FW_SECTION_LENGTH;
+	checksum = goodix_firmware_checksum(data, GOODIX_FW_DSP_LENGTH);
+	if (checksum) {
+		dev_err(dev, "DSP firmware checksum error\n");
+		return -EINVAL;
+	}
+
+	fw_header = (const struct goodix_fw_header *)fw->data;
+	dev_info(dev, "Firmware hardware info %02x%02x%02x%02x\n",
+		 fw_header->hw_info[0], fw_header->hw_info[1],
+		 fw_header->hw_info[2], fw_header->hw_info[3]);
+	/* pid is a 8 byte buffer containing a string, weird I know */
+	memcpy(buf, fw_header->pid, 8);
+	buf[8] = 0;
+	dev_info(dev, "Firmware PID: %s VID: %02x%02x\n", buf,
+		 fw_header->vid[0], fw_header->vid[1]);
+	return 0;
+}
+
+static int goodix_enter_upload_mode(struct i2c_client *client)
+{
+	int tries, error;
+	u8 val;
+
+	tries = 200;
+	do {
+		error = goodix_i2c_write_u8(client,
+					    GOODIX_REG_MISCTL_SWRST, 0x0c);
+		if (error)
+			return error;
+
+		error = goodix_i2c_read(client,
+					GOODIX_REG_MISCTL_SWRST, &val, 1);
+		if (error)
+			return error;
+
+		if (val == 0x0c)
+			break;
+	} while (--tries);
+
+	if (!tries) {
+		dev_err(&client->dev, "Error could not hold ss51 & dsp\n");
+		return -EIO;
+	}
+
+	/* DSP_CK and DSP_ALU_CK PowerOn */
+	error = goodix_i2c_write_u8(client, GOODIX_REG_MISCTL_DSP_CTL, 0x00);
+	if (error)
+		return error;
+
+	/* Disable watchdog */
+	error = goodix_i2c_write_u8(client, GOODIX_REG_MISCTL_TMR0_EN, 0x00);
+	if (error)
+		return error;
+
+	/* Clear cache enable */
+	error = goodix_i2c_write_u8(client, GOODIX_REG_MISCTL_CACHE_EN, 0x00);
+	if (error)
+		return error;
+
+	/* Set boot from SRAM */
+	error = goodix_i2c_write_u8(client, GOODIX_REG_MISCTL_BOOTCTL, 0x02);
+	if (error)
+		return error;
+
+	/* Software reboot */
+	error = goodix_i2c_write_u8(client,
+				    GOODIX_REG_MISCTL_CPU_SWRST_PULSE, 0x01);
+	if (error)
+		return error;
+
+	/* Clear control flag */
+	error = goodix_i2c_write_u8(client, GOODIX_REG_MISCTL_BOOTCTL, 0x00);
+	if (error)
+		return error;
+
+	/* Set scramble */
+	error = goodix_i2c_write_u8(client, GOODIX_REG_MISCTL_BOOT_OPT, 0x00);
+	if (error)
+		return error;
+
+	/* Enable accessing code */
+	error = goodix_i2c_write_u8(client, GOODIX_REG_MISCTL_MEM_CD_EN, 0x01);
+	if (error)
+		return error;
+
+	return 0;
+}
+
+static int goodix_start_firmware(struct i2c_client *client)
+{
+	int error;
+	u8 val;
+
+	/* Init software watchdog */
+	error = goodix_i2c_write_u8(client, GOODIX_REG_SW_WDT, 0xaa);
+	if (error)
+		return error;
+
+	/* Release SS51 & DSP */
+	error = goodix_i2c_write_u8(client, GOODIX_REG_MISCTL_SWRST, 0x00);
+	if (error)
+		return error;
+
+	error = goodix_i2c_read(client, GOODIX_REG_SW_WDT, &val, 1);
+	if (error)
+		return error;
+
+	/* The value we've written to SW_WDT should have been cleared now */
+	if (val == 0xaa) {
+		dev_err(&client->dev, "Error SW_WDT reg not cleared on fw startup\n");
+		return -EIO;
+	}
+
+	/* Re-init software watchdog */
+	error = goodix_i2c_write_u8(client, GOODIX_REG_SW_WDT, 0xaa);
+	if (error)
+		return error;
+
+	return 0;
+}
+
+static int goodix_firmware_upload(struct goodix_ts_data *ts)
+{
+	const struct firmware *fw;
+	char fw_name[64];
+	const u8 *data;
+	int error;
+
+	snprintf(fw_name, sizeof(fw_name), "goodix/%s", ts->firmware_name);
+
+	error = request_firmware(&fw, fw_name, &ts->client->dev);
+	if (error) {
+		dev_err(&ts->client->dev, "Firmware request error %d\n", error);
+		return error;
+	}
+
+	error = goodix_firmware_verify(&ts->client->dev, fw);
+	if (error)
+		goto release;
+
+	error = goodix_reset_no_int_sync(ts);
+	if (error)
+		return error;
+
+	error = goodix_enter_upload_mode(ts->client);
+	if (error)
+		goto release;
+
+	/* Select SRAM bank 0 and upload section 1 & 2 */
+	error = goodix_i2c_write_u8(ts->client,
+				    GOODIX_REG_MISCTL_SRAM_BANK, 0x00);
+	if (error)
+		goto release;
+
+	data = fw->data + GOODIX_FW_HEADER_LENGTH;
+	error = goodix_i2c_write(ts->client, GOODIX_FW_UPLOAD_ADDRESS,
+				 data, 2 * GOODIX_FW_SECTION_LENGTH);
+	if (error)
+		goto release;
+
+	/* Select SRAM bank 1 and upload section 3 & 4 */
+	error = goodix_i2c_write_u8(ts->client,
+				    GOODIX_REG_MISCTL_SRAM_BANK, 0x01);
+	if (error)
+		goto release;
+
+	data += 2 * GOODIX_FW_SECTION_LENGTH;
+	error = goodix_i2c_write(ts->client, GOODIX_FW_UPLOAD_ADDRESS,
+				 data, 2 * GOODIX_FW_SECTION_LENGTH);
+	if (error)
+		goto release;
+
+	/* Select SRAM bank 2 and upload the DSP firmware */
+	error = goodix_i2c_write_u8(ts->client,
+				    GOODIX_REG_MISCTL_SRAM_BANK, 0x02);
+	if (error)
+		goto release;
+
+	data += 2 * GOODIX_FW_SECTION_LENGTH;
+	error = goodix_i2c_write(ts->client, GOODIX_FW_UPLOAD_ADDRESS,
+				 data, GOODIX_FW_DSP_LENGTH);
+	if (error)
+		goto release;
+
+	error = goodix_start_firmware(ts->client);
+	if (error)
+		goto release;
+
+	error = goodix_int_sync(ts);
+release:
+	release_firmware(fw);
+	return error;
+}
+
+static int goodix_prepare_bak_ref(struct goodix_ts_data *ts)
+{
+	u8 have_key, driver_num, sensor_num;
+
+	if (ts->bak_ref)
+		return 0; /* Already done */
+
+	have_key = (ts->config[GOODIX_CFG_LOC_HAVE_KEY] & 0x01);
+
+	driver_num = (ts->config[GOODIX_CFG_LOC_DRVA_NUM] & 0x1f) +
+		     (ts->config[GOODIX_CFG_LOC_DRVB_NUM] & 0x1f);
+	if (have_key)
+		driver_num--;
+
+	sensor_num = (ts->config[GOODIX_CFG_LOC_SENS_NUM] & 0x0f) +
+		     ((ts->config[GOODIX_CFG_LOC_SENS_NUM] >> 4) & 0x0f);
+
+	dev_dbg(&ts->client->dev, "Drv %d Sen %d Key %d\n",
+		driver_num, sensor_num, have_key);
+
+	ts->bak_ref_len = (driver_num * (sensor_num - 2) + 2) * 2;
+
+	ts->bak_ref = devm_kzalloc(&ts->client->dev,
+				   ts->bak_ref_len, GFP_KERNEL);
+	if (!ts->bak_ref)
+		return -ENOMEM;
+
+	/*
+	 * The bak_ref array contains the backup of an array of (self/auto)
+	 * calibration related values which the Android version of the driver
+	 * stores on the filesystem so that it can be restored after reboot.
+	 * The mainline kernel never writes directly to the filesystem like
+	 * this, we always start will all the values which give a correction
+	 * factor in approx. the -20 - +20 range (in 2s complement) set to 0.
+	 *
+	 * Note the touchscreen works fine without restoring the reference
+	 * values after a reboot / power-cycle.
+	 *
+	 * The last 2 bytes are a 16 bits unsigned checksum which is expected
+	 * to make the addition al all 16 bit unsigned values in the array add
+	 * up to 1 (rather then the usual 0), so we must set the last byte to 1.
+	 */
+	ts->bak_ref[ts->bak_ref_len - 1] = 1;
+
+	return 0;
+}
+
+static int goodix_send_main_clock(struct goodix_ts_data *ts)
+{
+	u32 main_clk = 54; /* Default main clock */
+	u8 checksum = 0;
+	int i;
+
+	device_property_read_u32(&ts->client->dev,
+				 "goodix,main-clk", &main_clk);
+
+	for (i = 0; i < (GOODIX_MAIN_CLK_LEN - 1); i++) {
+		ts->main_clk[i] = main_clk;
+		checksum += main_clk;
+	}
+
+	/* The value of all bytes combines must be 0 */
+	ts->main_clk[GOODIX_MAIN_CLK_LEN - 1] = 256 - checksum;
+
+	return goodix_i2c_write(ts->client, GOODIX_REG_MAIN_CLK,
+				ts->main_clk, GOODIX_MAIN_CLK_LEN);
+}
+
+int goodix_firmware_check(struct goodix_ts_data *ts)
+{
+	device_property_read_string(&ts->client->dev,
+				    "firmware-name", &ts->firmware_name);
+	if (!ts->firmware_name)
+		return 0;
+
+	if (ts->irq_pin_access_method == IRQ_PIN_ACCESS_NONE) {
+		dev_err(&ts->client->dev, "Error no IRQ-pin access method, cannot upload fw.\n");
+		return -EINVAL;
+	}
+
+	dev_info(&ts->client->dev, "Touchscreen controller needs fw-upload\n");
+	ts->load_cfg_from_disk = true;
+
+	return goodix_firmware_upload(ts);
+}
+
+bool goodix_handle_fw_request(struct goodix_ts_data *ts)
+{
+	int error;
+	u8 val;
+
+	error = goodix_i2c_read(ts->client, GOODIX_REG_REQUEST, &val, 1);
+	if (error)
+		return false;
+
+	switch (val) {
+	case GOODIX_RQST_RESPONDED:
+		/*
+		 * If we read back our own last ack the IRQ was not for
+		 * a request.
+		 */
+		return false;
+	case GOODIX_RQST_CONFIG:
+		error = goodix_send_cfg(ts, ts->config, ts->chip->config_len);
+		if (error)
+			return false;
+
+		break;
+	case GOODIX_RQST_BAK_REF:
+		error = goodix_prepare_bak_ref(ts);
+		if (error)
+			return false;
+
+		error = goodix_i2c_write(ts->client, GOODIX_REG_BAK_REF,
+					 ts->bak_ref, ts->bak_ref_len);
+		if (error)
+			return false;
+
+		break;
+	case GOODIX_RQST_RESET:
+		error = goodix_firmware_upload(ts);
+		if (error)
+			return false;
+
+		break;
+	case GOODIX_RQST_MAIN_CLOCK:
+		error = goodix_send_main_clock(ts);
+		if (error)
+			return false;
+
+		break;
+	case GOODIX_RQST_UNKNOWN:
+	case GOODIX_RQST_IDLE:
+		break;
+	default:
+		dev_err_ratelimited(&ts->client->dev, "Unknown Request: 0x%02x\n", val);
+	}
+
+	/* Ack the request */
+	goodix_i2c_write_u8(ts->client,
+			    GOODIX_REG_REQUEST, GOODIX_RQST_RESPONDED);
+	return true;
+}
+
+void goodix_save_bak_ref(struct goodix_ts_data *ts)
+{
+	int error;
+	u8 val;
+
+	if (!ts->firmware_name)
+		return;
+
+	error = goodix_i2c_read(ts->client, GOODIX_REG_STATUS, &val, 1);
+	if (error)
+		return;
+
+	if (!(val & 0x80))
+		return;
+
+	error = goodix_i2c_read(ts->client, GOODIX_REG_BAK_REF,
+				ts->bak_ref, ts->bak_ref_len);
+	if (error) {
+		memset(ts->bak_ref, 0, ts->bak_ref_len);
+		ts->bak_ref[ts->bak_ref_len - 1] = 1;
+	}
+}
diff --git a/drivers/input/touchscreen/ili210x.c b/drivers/input/touchscreen/ili210x.c
index 30576a5..2bd407d 100644
--- a/drivers/input/touchscreen/ili210x.c
+++ b/drivers/input/touchscreen/ili210x.c
@@ -1,7 +1,9 @@
 // SPDX-License-Identifier: GPL-2.0-only
+#include <linux/crc-ccitt.h>
 #include <linux/delay.h>
 #include <linux/gpio/consumer.h>
 #include <linux/i2c.h>
+#include <linux/ihex.h>
 #include <linux/input.h>
 #include <linux/input/mt.h>
 #include <linux/input/touchscreen.h>
@@ -12,7 +14,7 @@
 #include <linux/slab.h>
 #include <asm/unaligned.h>
 
-#define ILI2XXX_POLL_PERIOD	20
+#define ILI2XXX_POLL_PERIOD	15
 
 #define ILI210X_DATA_SIZE	64
 #define ILI211X_DATA_SIZE	43
@@ -22,8 +24,23 @@
 /* Touchscreen commands */
 #define REG_TOUCHDATA		0x10
 #define REG_PANEL_INFO		0x20
+#define REG_FIRMWARE_VERSION	0x40
+#define REG_PROTOCOL_VERSION	0x42
+#define REG_KERNEL_VERSION	0x61
+#define REG_IC_BUSY		0x80
+#define REG_IC_BUSY_NOT_BUSY	0x50
+#define REG_GET_MODE		0xc0
+#define REG_GET_MODE_AP		0x5a
+#define REG_GET_MODE_BL		0x55
+#define REG_SET_MODE_AP		0xc1
+#define REG_SET_MODE_BL		0xc2
+#define REG_WRITE_DATA		0xc3
+#define REG_WRITE_ENABLE	0xc4
+#define REG_READ_DATA_CRC	0xc7
 #define REG_CALIBRATE		0xcc
 
+#define ILI251X_FW_FILENAME	"ilitek/ili251x.bin"
+
 struct ili2xxx_chip {
 	int (*read_reg)(struct i2c_client *client, u8 reg,
 			void *buf, size_t len);
@@ -35,6 +52,7 @@ struct ili2xxx_chip {
 	unsigned int max_touches;
 	unsigned int resolution;
 	bool has_calibrate_reg;
+	bool has_firmware_proto;
 	bool has_pressure_reg;
 };
 
@@ -44,6 +62,10 @@ struct ili210x {
 	struct gpio_desc *reset_gpio;
 	struct touchscreen_properties prop;
 	const struct ili2xxx_chip *chip;
+	u8 version_firmware[8];
+	u8 version_kernel[5];
+	u8 version_proto[2];
+	u8 ic_mode[2];
 	bool stop;
 };
 
@@ -202,15 +224,17 @@ static const struct ili2xxx_chip ili212x_chip = {
 	.has_calibrate_reg	= true,
 };
 
-static int ili251x_read_reg(struct i2c_client *client,
-			    u8 reg, void *buf, size_t len)
+static int ili251x_read_reg_common(struct i2c_client *client,
+				   u8 reg, void *buf, size_t len,
+				   unsigned int delay)
 {
 	int error;
 	int ret;
 
 	ret = i2c_master_send(client, &reg, 1);
 	if (ret == 1) {
-		usleep_range(5000, 5500);
+		if (delay)
+			usleep_range(delay, delay + 500);
 
 		ret = i2c_master_recv(client, buf, len);
 		if (ret == len)
@@ -222,12 +246,18 @@ static int ili251x_read_reg(struct i2c_client *client,
 	return ret;
 }
 
+static int ili251x_read_reg(struct i2c_client *client,
+			    u8 reg, void *buf, size_t len)
+{
+	return ili251x_read_reg_common(client, reg, buf, len, 5000);
+}
+
 static int ili251x_read_touch_data(struct i2c_client *client, u8 *data)
 {
 	int error;
 
-	error = ili251x_read_reg(client, REG_TOUCHDATA,
-				 data, ILI251X_DATA_SIZE1);
+	error = ili251x_read_reg_common(client, REG_TOUCHDATA,
+					data, ILI251X_DATA_SIZE1, 0);
 	if (!error && data[0] == 2) {
 		error = i2c_master_recv(client, data + ILI251X_DATA_SIZE1,
 					ILI251X_DATA_SIZE2);
@@ -268,6 +298,7 @@ static const struct ili2xxx_chip ili251x_chip = {
 	.continue_polling	= ili251x_check_continue_polling,
 	.max_touches		= 10,
 	.has_calibrate_reg	= true,
+	.has_firmware_proto	= true,
 	.has_pressure_reg	= true,
 };
 
@@ -303,10 +334,13 @@ static irqreturn_t ili210x_irq(int irq, void *irq_data)
 	const struct ili2xxx_chip *chip = priv->chip;
 	u8 touchdata[ILI210X_DATA_SIZE] = { 0 };
 	bool keep_polling;
+	ktime_t time_next;
+	s64 time_delta;
 	bool touch;
 	int error;
 
 	do {
+		time_next = ktime_add_ms(ktime_get(), ILI2XXX_POLL_PERIOD);
 		error = chip->get_touch_data(client, touchdata);
 		if (error) {
 			dev_err(&client->dev,
@@ -316,13 +350,201 @@ static irqreturn_t ili210x_irq(int irq, void *irq_data)
 
 		touch = ili210x_report_events(priv, touchdata);
 		keep_polling = chip->continue_polling(touchdata, touch);
-		if (keep_polling)
-			msleep(ILI2XXX_POLL_PERIOD);
+		if (keep_polling) {
+			time_delta = ktime_us_delta(time_next, ktime_get());
+			if (time_delta > 0)
+				usleep_range(time_delta, time_delta + 1000);
+		}
 	} while (!priv->stop && keep_polling);
 
 	return IRQ_HANDLED;
 }
 
+static int ili251x_firmware_update_resolution(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	u16 resx, resy;
+	u8 rs[10];
+	int error;
+
+	/* The firmware update blob might have changed the resolution. */
+	error = priv->chip->read_reg(client, REG_PANEL_INFO, &rs, sizeof(rs));
+	if (error)
+		return error;
+
+	resx = le16_to_cpup((__le16 *)rs);
+	resy = le16_to_cpup((__le16 *)(rs + 2));
+
+	/* The value reported by the firmware is invalid. */
+	if (!resx || resx == 0xffff || !resy || resy == 0xffff)
+		return -EINVAL;
+
+	input_abs_set_max(priv->input, ABS_X, resx - 1);
+	input_abs_set_max(priv->input, ABS_Y, resy - 1);
+	input_abs_set_max(priv->input, ABS_MT_POSITION_X, resx - 1);
+	input_abs_set_max(priv->input, ABS_MT_POSITION_Y, resy - 1);
+
+	return 0;
+}
+
+static ssize_t ili251x_firmware_update_firmware_version(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	int error;
+	u8 fw[8];
+
+	/* Get firmware version */
+	error = priv->chip->read_reg(client, REG_FIRMWARE_VERSION,
+				     &fw, sizeof(fw));
+	if (!error)
+		memcpy(priv->version_firmware, fw, sizeof(fw));
+
+	return error;
+}
+
+static ssize_t ili251x_firmware_update_kernel_version(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	int error;
+	u8 kv[5];
+
+	/* Get kernel version */
+	error = priv->chip->read_reg(client, REG_KERNEL_VERSION,
+				     &kv, sizeof(kv));
+	if (!error)
+		memcpy(priv->version_kernel, kv, sizeof(kv));
+
+	return error;
+}
+
+static ssize_t ili251x_firmware_update_protocol_version(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	int error;
+	u8 pv[2];
+
+	/* Get protocol version */
+	error = priv->chip->read_reg(client, REG_PROTOCOL_VERSION,
+				     &pv, sizeof(pv));
+	if (!error)
+		memcpy(priv->version_proto, pv, sizeof(pv));
+
+	return error;
+}
+
+static ssize_t ili251x_firmware_update_ic_mode(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	int error;
+	u8 md[2];
+
+	/* Get chip boot mode */
+	error = priv->chip->read_reg(client, REG_GET_MODE, &md, sizeof(md));
+	if (!error)
+		memcpy(priv->ic_mode, md, sizeof(md));
+
+	return error;
+}
+
+static int ili251x_firmware_update_cached_state(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	int error;
+
+	if (!priv->chip->has_firmware_proto)
+		return 0;
+
+	/* Wait for firmware to boot and stabilize itself. */
+	msleep(200);
+
+	/* Firmware does report valid information. */
+	error = ili251x_firmware_update_resolution(dev);
+	if (error)
+		return error;
+
+	error = ili251x_firmware_update_firmware_version(dev);
+	if (error)
+		return error;
+
+	error = ili251x_firmware_update_kernel_version(dev);
+	if (error)
+		return error;
+
+	error = ili251x_firmware_update_protocol_version(dev);
+	if (error)
+		return error;
+
+	error = ili251x_firmware_update_ic_mode(dev);
+	if (error)
+		return error;
+
+	return 0;
+}
+
+static ssize_t ili251x_firmware_version_show(struct device *dev,
+					     struct device_attribute *attr,
+					     char *buf)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	u8 *fw = priv->version_firmware;
+
+	return sysfs_emit(buf, "%02x%02x.%02x%02x.%02x%02x.%02x%02x\n",
+			  fw[0], fw[1], fw[2], fw[3],
+			  fw[4], fw[5], fw[6], fw[7]);
+}
+static DEVICE_ATTR(firmware_version, 0444, ili251x_firmware_version_show, NULL);
+
+static ssize_t ili251x_kernel_version_show(struct device *dev,
+					   struct device_attribute *attr,
+					   char *buf)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	u8 *kv = priv->version_kernel;
+
+	return sysfs_emit(buf, "%02x.%02x.%02x.%02x.%02x\n",
+			  kv[0], kv[1], kv[2], kv[3], kv[4]);
+}
+static DEVICE_ATTR(kernel_version, 0444, ili251x_kernel_version_show, NULL);
+
+static ssize_t ili251x_protocol_version_show(struct device *dev,
+					     struct device_attribute *attr,
+					     char *buf)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	u8 *pv = priv->version_proto;
+
+	return sysfs_emit(buf, "%02x.%02x\n", pv[0], pv[1]);
+}
+static DEVICE_ATTR(protocol_version, 0444, ili251x_protocol_version_show, NULL);
+
+static ssize_t ili251x_mode_show(struct device *dev,
+				 struct device_attribute *attr, char *buf)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	u8 *md = priv->ic_mode;
+	char *mode = "AP";
+
+	if (md[0] == REG_GET_MODE_AP)		/* Application Mode */
+		mode = "AP";
+	else if (md[0] == REG_GET_MODE_BL)	/* BootLoader Mode */
+		mode = "BL";
+	else					/* Unknown Mode */
+		mode = "??";
+
+	return sysfs_emit(buf, "%02x.%02x:%s\n", md[0], md[1], mode);
+}
+static DEVICE_ATTR(mode, 0444, ili251x_mode_show, NULL);
+
 static ssize_t ili210x_calibrate(struct device *dev,
 				 struct device_attribute *attr,
 				 const char *buf, size_t count)
@@ -349,24 +571,333 @@ static ssize_t ili210x_calibrate(struct device *dev,
 }
 static DEVICE_ATTR(calibrate, S_IWUSR, NULL, ili210x_calibrate);
 
+static int ili251x_firmware_to_buffer(const struct firmware *fw,
+				      u8 **buf, u16 *ac_end, u16 *df_end)
+{
+	const struct ihex_binrec *rec;
+	u32 fw_addr, fw_last_addr = 0;
+	u16 fw_len;
+	u8 *fw_buf;
+	int error;
+
+	/*
+	 * The firmware ihex blob can never be bigger than 64 kiB, so make this
+	 * simple -- allocate a 64 kiB buffer, iterate over the ihex blob records
+	 * once, copy them all into this buffer at the right locations, and then
+	 * do all operations on this linear buffer.
+	 */
+	fw_buf = kzalloc(SZ_64K, GFP_KERNEL);
+	if (!fw_buf)
+		return -ENOMEM;
+
+	rec = (const struct ihex_binrec *)fw->data;
+	while (rec) {
+		fw_addr = be32_to_cpu(rec->addr);
+		fw_len = be16_to_cpu(rec->len);
+
+		/* The last 32 Byte firmware block can be 0xffe0 */
+		if (fw_addr + fw_len > SZ_64K || fw_addr > SZ_64K - 32) {
+			error = -EFBIG;
+			goto err_big;
+		}
+
+		/* Find the last address before DF start address, that is AC end */
+		if (fw_addr == 0xf000)
+			*ac_end = fw_last_addr;
+		fw_last_addr = fw_addr + fw_len;
+
+		memcpy(fw_buf + fw_addr, rec->data, fw_len);
+		rec = ihex_next_binrec(rec);
+	}
+
+	/* DF end address is the last address in the firmware blob */
+	*df_end = fw_addr + fw_len;
+	*buf = fw_buf;
+	return 0;
+
+err_big:
+	kfree(fw_buf);
+	return error;
+}
+
+/* Switch mode between Application and BootLoader */
+static int ili251x_switch_ic_mode(struct i2c_client *client, u8 cmd_mode)
+{
+	struct ili210x *priv = i2c_get_clientdata(client);
+	u8 cmd_wren[3] = { REG_WRITE_ENABLE, 0x5a, 0xa5 };
+	u8 md[2];
+	int error;
+
+	error = priv->chip->read_reg(client, REG_GET_MODE, md, sizeof(md));
+	if (error)
+		return error;
+	/* Mode already set */
+	if ((cmd_mode == REG_SET_MODE_AP && md[0] == REG_GET_MODE_AP) ||
+	    (cmd_mode == REG_SET_MODE_BL && md[0] == REG_GET_MODE_BL))
+		return 0;
+
+	/* Unlock writes */
+	error = i2c_master_send(client, cmd_wren, sizeof(cmd_wren));
+	if (error != sizeof(cmd_wren))
+		return -EINVAL;
+
+	mdelay(20);
+
+	/* Select mode (BootLoader or Application) */
+	error = i2c_master_send(client, &cmd_mode, 1);
+	if (error != 1)
+		return -EINVAL;
+
+	mdelay(200);	/* Reboot into bootloader takes a lot of time ... */
+
+	/* Read back mode */
+	error = priv->chip->read_reg(client, REG_GET_MODE, md, sizeof(md));
+	if (error)
+		return error;
+	/* Check if mode is correct now. */
+	if ((cmd_mode == REG_SET_MODE_AP && md[0] == REG_GET_MODE_AP) ||
+	    (cmd_mode == REG_SET_MODE_BL && md[0] == REG_GET_MODE_BL))
+		return 0;
+
+	return -EINVAL;
+}
+
+static int ili251x_firmware_busy(struct i2c_client *client)
+{
+	struct ili210x *priv = i2c_get_clientdata(client);
+	int error, i = 0;
+	u8 data;
+
+	do {
+		/* The read_reg already contains suitable delay */
+		error = priv->chip->read_reg(client, REG_IC_BUSY, &data, 1);
+		if (error)
+			return error;
+		if (i++ == 100000)
+			return -ETIMEDOUT;
+	} while (data != REG_IC_BUSY_NOT_BUSY);
+
+	return 0;
+}
+
+static int ili251x_firmware_write_to_ic(struct device *dev, u8 *fwbuf,
+					u16 start, u16 end, u8 dataflash)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+	u8 cmd_crc = REG_READ_DATA_CRC;
+	u8 crcrb[4] = { 0 };
+	u8 fw_data[33];
+	u16 fw_addr;
+	int error;
+
+	/*
+	 * The DF (dataflash) needs 2 bytes offset for unknown reasons,
+	 * the AC (application) has 2 bytes CRC16-CCITT at the end.
+	 */
+	u16 crc = crc_ccitt(0, fwbuf + start + (dataflash ? 2 : 0),
+			    end - start - 2);
+
+	/* Unlock write to either AC (application) or DF (dataflash) area */
+	u8 cmd_wr[10] = {
+		REG_WRITE_ENABLE, 0x5a, 0xa5, dataflash,
+		(end >> 16) & 0xff, (end >> 8) & 0xff, end & 0xff,
+		(crc >> 16) & 0xff, (crc >> 8) & 0xff, crc & 0xff
+	};
+
+	error = i2c_master_send(client, cmd_wr, sizeof(cmd_wr));
+	if (error != sizeof(cmd_wr))
+		return -EINVAL;
+
+	error = ili251x_firmware_busy(client);
+	if (error)
+		return error;
+
+	for (fw_addr = start; fw_addr < end; fw_addr += 32) {
+		fw_data[0] = REG_WRITE_DATA;
+		memcpy(&(fw_data[1]), fwbuf + fw_addr, 32);
+		error = i2c_master_send(client, fw_data, 33);
+		if (error != sizeof(fw_data))
+			return error;
+		error = ili251x_firmware_busy(client);
+		if (error)
+			return error;
+	}
+
+	error = i2c_master_send(client, &cmd_crc, 1);
+	if (error != 1)
+		return -EINVAL;
+
+	error = ili251x_firmware_busy(client);
+	if (error)
+		return error;
+
+	error = priv->chip->read_reg(client, REG_READ_DATA_CRC,
+				   &crcrb, sizeof(crcrb));
+	if (error)
+		return error;
+
+	/* Check CRC readback */
+	if ((crcrb[0] != (crc & 0xff)) || crcrb[1] != ((crc >> 8) & 0xff))
+		return -EINVAL;
+
+	return 0;
+}
+
+static int ili251x_firmware_reset(struct i2c_client *client)
+{
+	u8 cmd_reset[2] = { 0xf2, 0x01 };
+	int error;
+
+	error = i2c_master_send(client, cmd_reset, sizeof(cmd_reset));
+	if (error != sizeof(cmd_reset))
+		return -EINVAL;
+
+	return ili251x_firmware_busy(client);
+}
+
+static void ili251x_hardware_reset(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct ili210x *priv = i2c_get_clientdata(client);
+
+	/* Reset the controller */
+	gpiod_set_value_cansleep(priv->reset_gpio, 1);
+	usleep_range(10000, 15000);
+	gpiod_set_value_cansleep(priv->reset_gpio, 0);
+	msleep(300);
+}
+
+static ssize_t ili210x_firmware_update_store(struct device *dev,
+					     struct device_attribute *attr,
+					     const char *buf, size_t count)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	const char *fwname = ILI251X_FW_FILENAME;
+	const struct firmware *fw;
+	u16 ac_end, df_end;
+	u8 *fwbuf;
+	int error;
+	int i;
+
+	error = request_ihex_firmware(&fw, fwname, dev);
+	if (error) {
+		dev_err(dev, "Failed to request firmware %s, error=%d\n",
+			fwname, error);
+		return error;
+	}
+
+	error = ili251x_firmware_to_buffer(fw, &fwbuf, &ac_end, &df_end);
+	release_firmware(fw);
+	if (error)
+		return error;
+
+	/*
+	 * Disable touchscreen IRQ, so that we would not get spurious touch
+	 * interrupt during firmware update, and so that the IRQ handler won't
+	 * trigger and interfere with the firmware update. There is no bit in
+	 * the touch controller to disable the IRQs during update, so we have
+	 * to do it this way here.
+	 */
+	disable_irq(client->irq);
+
+	dev_dbg(dev, "Firmware update started, firmware=%s\n", fwname);
+
+	ili251x_hardware_reset(dev);
+
+	error = ili251x_firmware_reset(client);
+	if (error)
+		goto exit;
+
+	/* This may not succeed on first try, so re-try a few times. */
+	for (i = 0; i < 5; i++) {
+		error = ili251x_switch_ic_mode(client, REG_SET_MODE_BL);
+		if (!error)
+			break;
+	}
+
+	if (error)
+		goto exit;
+
+	dev_dbg(dev, "IC is now in BootLoader mode\n");
+
+	msleep(200);	/* The bootloader seems to need some time too. */
+
+	error = ili251x_firmware_write_to_ic(dev, fwbuf, 0xf000, df_end, 1);
+	if (error) {
+		dev_err(dev, "DF firmware update failed, error=%d\n", error);
+		goto exit;
+	}
+
+	dev_dbg(dev, "DataFlash firmware written\n");
+
+	error = ili251x_firmware_write_to_ic(dev, fwbuf, 0x2000, ac_end, 0);
+	if (error) {
+		dev_err(dev, "AC firmware update failed, error=%d\n", error);
+		goto exit;
+	}
+
+	dev_dbg(dev, "Application firmware written\n");
+
+	/* This may not succeed on first try, so re-try a few times. */
+	for (i = 0; i < 5; i++) {
+		error = ili251x_switch_ic_mode(client, REG_SET_MODE_AP);
+		if (!error)
+			break;
+	}
+
+	if (error)
+		goto exit;
+
+	dev_dbg(dev, "IC is now in Application mode\n");
+
+	error = ili251x_firmware_update_cached_state(dev);
+	if (error)
+		goto exit;
+
+	error = count;
+
+exit:
+	ili251x_hardware_reset(dev);
+	dev_dbg(dev, "Firmware update ended, error=%i\n", error);
+	enable_irq(client->irq);
+	kfree(fwbuf);
+	return error;
+}
+
+static DEVICE_ATTR(firmware_update, 0200, NULL, ili210x_firmware_update_store);
+
 static struct attribute *ili210x_attributes[] = {
 	&dev_attr_calibrate.attr,
+	&dev_attr_firmware_update.attr,
+	&dev_attr_firmware_version.attr,
+	&dev_attr_kernel_version.attr,
+	&dev_attr_protocol_version.attr,
+	&dev_attr_mode.attr,
 	NULL,
 };
 
-static umode_t ili210x_calibrate_visible(struct kobject *kobj,
+static umode_t ili210x_attributes_visible(struct kobject *kobj,
 					  struct attribute *attr, int index)
 {
 	struct device *dev = kobj_to_dev(kobj);
 	struct i2c_client *client = to_i2c_client(dev);
 	struct ili210x *priv = i2c_get_clientdata(client);
 
-	return priv->chip->has_calibrate_reg ? attr->mode : 0;
+	/* Calibrate is present on all ILI2xxx which have calibrate register */
+	if (attr == &dev_attr_calibrate.attr)
+		return priv->chip->has_calibrate_reg ? attr->mode : 0;
+
+	/* Firmware/Kernel/Protocol/BootMode is implememted only for ILI251x */
+	if (!priv->chip->has_firmware_proto)
+		return 0;
+
+	return attr->mode;
 }
 
 static const struct attribute_group ili210x_attr_group = {
 	.attrs = ili210x_attributes,
-	.is_visible = ili210x_calibrate_visible,
+	.is_visible = ili210x_attributes_visible,
 };
 
 static void ili210x_power_down(void *data)
@@ -449,6 +980,12 @@ static int ili210x_i2c_probe(struct i2c_client *client,
 	input_set_abs_params(input, ABS_MT_POSITION_Y, 0, max_xy, 0, 0);
 	if (priv->chip->has_pressure_reg)
 		input_set_abs_params(input, ABS_MT_PRESSURE, 0, 0xa, 0, 0);
+	error = ili251x_firmware_update_cached_state(dev);
+	if (error) {
+		dev_err(dev, "Unable to cache firmware information, err: %d\n",
+			error);
+		return error;
+	}
 	touchscreen_parse_properties(input, true, &priv->prop);
 
 	error = input_mt_init_slots(input, priv->chip->max_touches,
diff --git a/drivers/input/touchscreen/raydium_i2c_ts.c b/drivers/input/touchscreen/raydium_i2c_ts.c
index 4d2d22a..3a49529 100644
--- a/drivers/input/touchscreen/raydium_i2c_ts.c
+++ b/drivers/input/touchscreen/raydium_i2c_ts.c
@@ -37,6 +37,7 @@
 #define RM_CMD_BOOT_READ	0x44		/* send wait bl data ready*/
 
 #define RM_BOOT_RDY		0xFF		/* bl data ready */
+#define RM_BOOT_CMD_READHWID	0x0E		/* read hwid */
 
 /* I2C main commands */
 #define RM_CMD_QUERY_BANK	0x2B
@@ -290,6 +291,44 @@ static int raydium_i2c_sw_reset(struct i2c_client *client)
 	return 0;
 }
 
+static int raydium_i2c_query_ts_bootloader_info(struct raydium_data *ts)
+{
+	struct i2c_client *client = ts->client;
+	static const u8 get_hwid[] = { RM_BOOT_CMD_READHWID,
+				       0x10, 0xc0, 0x01, 0x00, 0x04, 0x00 };
+	u8 rbuf[5] = { 0 };
+	u32 hw_ver;
+	int error;
+
+	error = raydium_i2c_send(client, RM_CMD_BOOT_WRT,
+				 get_hwid, sizeof(get_hwid));
+	if (error) {
+		dev_err(&client->dev, "WRT HWID command failed: %d\n", error);
+		return error;
+	}
+
+	error = raydium_i2c_send(client, RM_CMD_BOOT_ACK, rbuf, 1);
+	if (error) {
+		dev_err(&client->dev, "Ack HWID command failed: %d\n", error);
+		return error;
+	}
+
+	error = raydium_i2c_read(client, RM_CMD_BOOT_CHK, rbuf, sizeof(rbuf));
+	if (error) {
+		dev_err(&client->dev, "Read HWID command failed: %d (%4ph)\n",
+			error, rbuf + 1);
+		hw_ver = 0xffffffffUL;
+	} else {
+		hw_ver = get_unaligned_be32(rbuf + 1);
+	}
+
+	ts->info.hw_ver = cpu_to_le32(hw_ver);
+	ts->info.main_ver = 0xff;
+	ts->info.sub_ver = 0xff;
+
+	return error;
+}
+
 static int raydium_i2c_query_ts_info(struct raydium_data *ts)
 {
 	struct i2c_client *client = ts->client;
@@ -388,13 +427,10 @@ static int raydium_i2c_initialize(struct raydium_data *ts)
 	if (error)
 		ts->boot_mode = RAYDIUM_TS_BLDR;
 
-	if (ts->boot_mode == RAYDIUM_TS_BLDR) {
-		ts->info.hw_ver = cpu_to_le32(0xffffffffUL);
-		ts->info.main_ver = 0xff;
-		ts->info.sub_ver = 0xff;
-	} else {
+	if (ts->boot_mode == RAYDIUM_TS_BLDR)
+		raydium_i2c_query_ts_bootloader_info(ts);
+	else
 		raydium_i2c_query_ts_info(ts);
-	}
 
 	return error;
 }
@@ -1082,11 +1118,11 @@ static int raydium_i2c_probe(struct i2c_client *client,
 	if (error)
 		return error;
 
-	error = devm_add_action(&client->dev, raydium_i2c_power_off, ts);
+	error = devm_add_action_or_reset(&client->dev,
+					 raydium_i2c_power_off, ts);
 	if (error) {
 		dev_err(&client->dev,
 			"failed to install power off action: %d\n", error);
-		raydium_i2c_power_off(ts);
 		return error;
 	}
 
@@ -1218,7 +1254,7 @@ static SIMPLE_DEV_PM_OPS(raydium_i2c_pm_ops,
 			 raydium_i2c_suspend, raydium_i2c_resume);
 
 static const struct i2c_device_id raydium_i2c_id[] = {
-	{ "raydium_i2c" , 0 },
+	{ "raydium_i2c", 0 },
 	{ "rm32380", 0 },
 	{ /* sentinel */ }
 };
diff --git a/drivers/input/touchscreen/st1232.c b/drivers/input/touchscreen/st1232.c
index 6abae66..e38ba3e 100644
--- a/drivers/input/touchscreen/st1232.c
+++ b/drivers/input/touchscreen/st1232.c
@@ -92,7 +92,7 @@ static int st1232_ts_wait_ready(struct st1232_ts_data *ts)
 	unsigned int retries;
 	int error;
 
-	for (retries = 10; retries; retries--) {
+	for (retries = 100; retries; retries--) {
 		error = st1232_ts_read_data(ts, REG_STATUS, 1);
 		if (!error) {
 			switch (ts->read_buf[0]) {
@@ -389,6 +389,7 @@ static struct i2c_driver st1232_ts_driver = {
 	.driver = {
 		.name	= ST1232_TS_NAME,
 		.of_match_table = st1232_ts_dt_ids,
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
 		.pm	= &st1232_ts_pm_ops,
 	},
 };
diff --git a/drivers/input/touchscreen/tsc2004.c b/drivers/input/touchscreen/tsc2004.c
index 0272ced..9fdd870 100644
--- a/drivers/input/touchscreen/tsc2004.c
+++ b/drivers/input/touchscreen/tsc2004.c
@@ -45,7 +45,9 @@ static int tsc2004_probe(struct i2c_client *i2c,
 
 static int tsc2004_remove(struct i2c_client *i2c)
 {
-	return tsc200x_remove(&i2c->dev);
+	tsc200x_remove(&i2c->dev);
+
+	return 0;
 }
 
 static const struct i2c_device_id tsc2004_idtable[] = {
diff --git a/drivers/input/touchscreen/tsc2005.c b/drivers/input/touchscreen/tsc2005.c
index 923496bb..a2f5592 100644
--- a/drivers/input/touchscreen/tsc2005.c
+++ b/drivers/input/touchscreen/tsc2005.c
@@ -66,7 +66,9 @@ static int tsc2005_probe(struct spi_device *spi)
 
 static int tsc2005_remove(struct spi_device *spi)
 {
-	return tsc200x_remove(&spi->dev);
+	tsc200x_remove(&spi->dev);
+
+	return 0;
 }
 
 #ifdef CONFIG_OF
diff --git a/drivers/input/touchscreen/tsc200x-core.c b/drivers/input/touchscreen/tsc200x-core.c
index b8d720d..27810f6 100644
--- a/drivers/input/touchscreen/tsc200x-core.c
+++ b/drivers/input/touchscreen/tsc200x-core.c
@@ -577,15 +577,13 @@ int tsc200x_probe(struct device *dev, int irq, const struct input_id *tsc_id,
 }
 EXPORT_SYMBOL_GPL(tsc200x_probe);
 
-int tsc200x_remove(struct device *dev)
+void tsc200x_remove(struct device *dev)
 {
 	struct tsc200x *ts = dev_get_drvdata(dev);
 
 	sysfs_remove_group(&dev->kobj, &tsc200x_attr_group);
 
 	regulator_disable(ts->vio);
-
-	return 0;
 }
 EXPORT_SYMBOL_GPL(tsc200x_remove);
 
diff --git a/drivers/input/touchscreen/tsc200x-core.h b/drivers/input/touchscreen/tsc200x-core.h
index a43c08c..4ded344 100644
--- a/drivers/input/touchscreen/tsc200x-core.h
+++ b/drivers/input/touchscreen/tsc200x-core.h
@@ -74,6 +74,6 @@ extern const struct dev_pm_ops tsc200x_pm_ops;
 int tsc200x_probe(struct device *dev, int irq, const struct input_id *tsc_id,
 		  struct regmap *regmap,
 		  int (*tsc200x_cmd)(struct device *dev, u8 cmd));
-int tsc200x_remove(struct device *dev);
+void tsc200x_remove(struct device *dev);
 
 #endif
diff --git a/drivers/input/touchscreen/wacom_i2c.c b/drivers/input/touchscreen/wacom_i2c.c
index 22826c3..fe4ea62 100644
--- a/drivers/input/touchscreen/wacom_i2c.c
+++ b/drivers/input/touchscreen/wacom_i2c.c
@@ -6,6 +6,7 @@
  * <tobita.tatsunosuke@wacom.co.jp>
  */
 
+#include <linux/bits.h>
 #include <linux/module.h>
 #include <linux/input.h>
 #include <linux/i2c.h>
@@ -14,6 +15,15 @@
 #include <linux/interrupt.h>
 #include <asm/unaligned.h>
 
+/* Bitmasks (for data[3]) */
+#define WACOM_TIP_SWITCH	BIT(0)
+#define WACOM_BARREL_SWITCH	BIT(1)
+#define WACOM_ERASER		BIT(2)
+#define WACOM_INVERT		BIT(3)
+#define WACOM_BARREL_SWITCH_2	BIT(4)
+#define WACOM_IN_PROXIMITY	BIT(5)
+
+/* Registers */
 #define WACOM_CMD_QUERY0	0x04
 #define WACOM_CMD_QUERY1	0x00
 #define WACOM_CMD_QUERY2	0x33
@@ -99,19 +109,19 @@ static irqreturn_t wacom_i2c_irq(int irq, void *dev_id)
 	if (error < 0)
 		goto out;
 
-	tsw = data[3] & 0x01;
-	ers = data[3] & 0x04;
-	f1 = data[3] & 0x02;
-	f2 = data[3] & 0x10;
+	tsw = data[3] & WACOM_TIP_SWITCH;
+	ers = data[3] & WACOM_ERASER;
+	f1 = data[3] & WACOM_BARREL_SWITCH;
+	f2 = data[3] & WACOM_BARREL_SWITCH_2;
 	x = le16_to_cpup((__le16 *)&data[4]);
 	y = le16_to_cpup((__le16 *)&data[6]);
 	pressure = le16_to_cpup((__le16 *)&data[8]);
 
 	if (!wac_i2c->prox)
-		wac_i2c->tool = (data[3] & 0x0c) ?
+		wac_i2c->tool = (data[3] & (WACOM_ERASER | WACOM_INVERT)) ?
 			BTN_TOOL_RUBBER : BTN_TOOL_PEN;
 
-	wac_i2c->prox = data[3] & 0x20;
+	wac_i2c->prox = data[3] & WACOM_IN_PROXIMITY;
 
 	input_report_key(input, BTN_TOUCH, tsw || ers);
 	input_report_key(input, wac_i2c->tool, wac_i2c->prox);
diff --git a/drivers/irqchip/irq-csky-mpintc.c b/drivers/irqchip/irq-csky-mpintc.c
index cb403c9..4aebd67 100644
--- a/drivers/irqchip/irq-csky-mpintc.c
+++ b/drivers/irqchip/irq-csky-mpintc.c
@@ -78,7 +78,7 @@ static void csky_mpintc_handler(struct pt_regs *regs)
 		readl_relaxed(reg_base + INTCL_RDYIR));
 }
 
-static void csky_mpintc_enable(struct irq_data *d)
+static void csky_mpintc_unmask(struct irq_data *d)
 {
 	void __iomem *reg_base = this_cpu_read(intcl_reg);
 
@@ -87,7 +87,7 @@ static void csky_mpintc_enable(struct irq_data *d)
 	writel_relaxed(d->hwirq, reg_base + INTCL_SENR);
 }
 
-static void csky_mpintc_disable(struct irq_data *d)
+static void csky_mpintc_mask(struct irq_data *d)
 {
 	void __iomem *reg_base = this_cpu_read(intcl_reg);
 
@@ -164,8 +164,8 @@ static int csky_irq_set_affinity(struct irq_data *d,
 static struct irq_chip csky_irq_chip = {
 	.name           = "C-SKY SMP Intc",
 	.irq_eoi	= csky_mpintc_eoi,
-	.irq_enable	= csky_mpintc_enable,
-	.irq_disable	= csky_mpintc_disable,
+	.irq_unmask	= csky_mpintc_unmask,
+	.irq_mask	= csky_mpintc_mask,
 	.irq_set_type	= csky_mpintc_set_type,
 #ifdef CONFIG_SMP
 	.irq_set_affinity = csky_irq_set_affinity,
diff --git a/drivers/irqchip/irq-sifive-plic.c b/drivers/irqchip/irq-sifive-plic.c
index cf74cfa..259065d 100644
--- a/drivers/irqchip/irq-sifive-plic.c
+++ b/drivers/irqchip/irq-sifive-plic.c
@@ -163,7 +163,13 @@ static void plic_irq_eoi(struct irq_data *d)
 {
 	struct plic_handler *handler = this_cpu_ptr(&plic_handlers);
 
-	writel(d->hwirq, handler->hart_base + CONTEXT_CLAIM);
+	if (irqd_irq_masked(d)) {
+		plic_irq_unmask(d);
+		writel(d->hwirq, handler->hart_base + CONTEXT_CLAIM);
+		plic_irq_mask(d);
+	} else {
+		writel(d->hwirq, handler->hart_base + CONTEXT_CLAIM);
+	}
 }
 
 static struct irq_chip plic_chip = {
diff --git a/drivers/of/irq.c b/drivers/of/irq.c
index 32be5a0..b10f015b 100644
--- a/drivers/of/irq.c
+++ b/drivers/of/irq.c
@@ -161,9 +161,10 @@ int of_irq_parse_raw(const __be32 *addr, struct of_phandle_args *out_irq)
 		 * if it is then we are done, unless there is an
 		 * interrupt-map which takes precedence.
 		 */
+		bool intc = of_property_read_bool(ipar, "interrupt-controller");
+
 		imap = of_get_property(ipar, "interrupt-map", &imaplen);
-		if (imap == NULL &&
-		    of_property_read_bool(ipar, "interrupt-controller")) {
+		if (imap == NULL && intc) {
 			pr_debug(" -> got it !\n");
 			return 0;
 		}
@@ -244,8 +245,20 @@ int of_irq_parse_raw(const __be32 *addr, struct of_phandle_args *out_irq)
 
 			pr_debug(" -> imaplen=%d\n", imaplen);
 		}
-		if (!match)
+		if (!match) {
+			if (intc) {
+				/*
+				 * The PASEMI Nemo is a known offender, so
+				 * let's only warn for anyone else.
+				 */
+				WARN(!IS_ENABLED(CONFIG_PPC_PASEMI),
+				     "%pOF interrupt-map failed, using interrupt-controller\n",
+				     ipar);
+				return 0;
+			}
+
 			goto fail;
+		}
 
 		/*
 		 * Successfully parsed an interrrupt-map translation; copy new
diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
index e917bb3..93b1411 100644
--- a/drivers/pci/controller/Kconfig
+++ b/drivers/pci/controller/Kconfig
@@ -270,7 +270,8 @@
 
 config PCIE_BRCMSTB
 	tristate "Broadcom Brcmstb PCIe host controller"
-	depends on ARCH_BRCMSTB || ARCH_BCM2835 || ARCH_BCM4908 || COMPILE_TEST
+	depends on ARCH_BRCMSTB || ARCH_BCM2835 || ARCH_BCM4908 || \
+		   BMIPS_GENERIC || COMPILE_TEST
 	depends on OF
 	depends on PCI_MSI_IRQ_DOMAIN
 	default ARCH_BRCMSTB
diff --git a/drivers/pci/hotplug/s390_pci_hpc.c b/drivers/pci/hotplug/s390_pci_hpc.c
index dcefdb4..a89b7de 100644
--- a/drivers/pci/hotplug/s390_pci_hpc.c
+++ b/drivers/pci/hotplug/s390_pci_hpc.c
@@ -57,6 +57,29 @@ static int disable_slot(struct hotplug_slot *hotplug_slot)
 	return zpci_deconfigure_device(zdev);
 }
 
+static int reset_slot(struct hotplug_slot *hotplug_slot, bool probe)
+{
+	struct zpci_dev *zdev = container_of(hotplug_slot, struct zpci_dev,
+					     hotplug_slot);
+
+	if (zdev->state != ZPCI_FN_STATE_CONFIGURED)
+		return -EIO;
+	/*
+	 * We can't take the zdev->lock as reset_slot may be called during
+	 * probing and/or device removal which already happens under the
+	 * zdev->lock. Instead the user should use the higher level
+	 * pci_reset_function() or pci_bus_reset() which hold the PCI device
+	 * lock preventing concurrent removal. If not using these functions
+	 * holding the PCI device lock is required.
+	 */
+
+	/* As long as the function is configured we can reset */
+	if (probe)
+		return 0;
+
+	return zpci_hot_reset_device(zdev);
+}
+
 static int get_power_status(struct hotplug_slot *hotplug_slot, u8 *value)
 {
 	struct zpci_dev *zdev = container_of(hotplug_slot, struct zpci_dev,
@@ -76,6 +99,7 @@ static int get_adapter_status(struct hotplug_slot *hotplug_slot, u8 *value)
 static const struct hotplug_slot_ops s390_hotplug_slot_ops = {
 	.enable_slot =		enable_slot,
 	.disable_slot =		disable_slot,
+	.reset_slot =		reset_slot,
 	.get_power_status =	get_power_status,
 	.get_adapter_status =	get_adapter_status,
 };
diff --git a/drivers/pci/msi.c b/drivers/pci/msi.c
index 12e296d..48e3f4e 100644
--- a/drivers/pci/msi.c
+++ b/drivers/pci/msi.c
@@ -148,6 +148,9 @@ static noinline void pci_msi_update_mask(struct msi_desc *desc, u32 clear, u32 s
 	raw_spinlock_t *lock = &desc->dev->msi_lock;
 	unsigned long flags;
 
+	if (!desc->msi_attrib.can_mask)
+		return;
+
 	raw_spin_lock_irqsave(lock, flags);
 	desc->msi_mask &= ~clear;
 	desc->msi_mask |= set;
@@ -181,7 +184,8 @@ static void pci_msix_write_vector_ctrl(struct msi_desc *desc, u32 ctrl)
 {
 	void __iomem *desc_addr = pci_msix_desc_addr(desc);
 
-	writel(ctrl, desc_addr + PCI_MSIX_ENTRY_VECTOR_CTRL);
+	if (desc->msi_attrib.can_mask)
+		writel(ctrl, desc_addr + PCI_MSIX_ENTRY_VECTOR_CTRL);
 }
 
 static inline void pci_msix_mask(struct msi_desc *desc)
@@ -200,23 +204,17 @@ static inline void pci_msix_unmask(struct msi_desc *desc)
 
 static void __pci_msi_mask_desc(struct msi_desc *desc, u32 mask)
 {
-	if (pci_msi_ignore_mask || desc->msi_attrib.is_virtual)
-		return;
-
 	if (desc->msi_attrib.is_msix)
 		pci_msix_mask(desc);
-	else if (desc->msi_attrib.maskbit)
+	else
 		pci_msi_mask(desc, mask);
 }
 
 static void __pci_msi_unmask_desc(struct msi_desc *desc, u32 mask)
 {
-	if (pci_msi_ignore_mask || desc->msi_attrib.is_virtual)
-		return;
-
 	if (desc->msi_attrib.is_msix)
 		pci_msix_unmask(desc);
-	else if (desc->msi_attrib.maskbit)
+	else
 		pci_msi_unmask(desc, mask);
 }
 
@@ -370,6 +368,11 @@ static void free_msi_irqs(struct pci_dev *dev)
 			for (i = 0; i < entry->nvec_used; i++)
 				BUG_ON(irq_has_action(entry->irq + i));
 
+	if (dev->msi_irq_groups) {
+		msi_destroy_sysfs(&dev->dev, dev->msi_irq_groups);
+		dev->msi_irq_groups = NULL;
+	}
+
 	pci_msi_teardown_msi_irqs(dev);
 
 	list_for_each_entry_safe(entry, tmp, msi_list, list) {
@@ -381,11 +384,6 @@ static void free_msi_irqs(struct pci_dev *dev)
 		list_del(&entry->list);
 		free_msi_entry(entry);
 	}
-
-	if (dev->msi_irq_groups) {
-		msi_destroy_sysfs(&dev->dev, dev->msi_irq_groups);
-		dev->msi_irq_groups = NULL;
-	}
 }
 
 static void pci_intx_for_msi(struct pci_dev *dev, int enable)
@@ -479,12 +477,16 @@ msi_setup_entry(struct pci_dev *dev, int nvec, struct irq_affinity *affd)
 		goto out;
 
 	pci_read_config_word(dev, dev->msi_cap + PCI_MSI_FLAGS, &control);
+	/* Lies, damned lies, and MSIs */
+	if (dev->dev_flags & PCI_DEV_FLAGS_HAS_MSI_MASKING)
+		control |= PCI_MSI_FLAGS_MASKBIT;
 
 	entry->msi_attrib.is_msix	= 0;
 	entry->msi_attrib.is_64		= !!(control & PCI_MSI_FLAGS_64BIT);
 	entry->msi_attrib.is_virtual    = 0;
 	entry->msi_attrib.entry_nr	= 0;
-	entry->msi_attrib.maskbit	= !!(control & PCI_MSI_FLAGS_MASKBIT);
+	entry->msi_attrib.can_mask	= !pci_msi_ignore_mask &&
+					  !!(control & PCI_MSI_FLAGS_MASKBIT);
 	entry->msi_attrib.default_irq	= dev->irq;	/* Save IOAPIC IRQ */
 	entry->msi_attrib.multi_cap	= (control & PCI_MSI_FLAGS_QMASK) >> 1;
 	entry->msi_attrib.multiple	= ilog2(__roundup_pow_of_two(nvec));
@@ -495,7 +497,7 @@ msi_setup_entry(struct pci_dev *dev, int nvec, struct irq_affinity *affd)
 		entry->mask_pos = dev->msi_cap + PCI_MSI_MASK_32;
 
 	/* Save the initial mask status */
-	if (entry->msi_attrib.maskbit)
+	if (entry->msi_attrib.can_mask)
 		pci_read_config_dword(dev, entry->mask_pos, &entry->msi_mask);
 
 out:
@@ -639,10 +641,13 @@ static int msix_setup_entries(struct pci_dev *dev, void __iomem *base,
 		entry->msi_attrib.is_virtual =
 			entry->msi_attrib.entry_nr >= vec_count;
 
+		entry->msi_attrib.can_mask	= !pci_msi_ignore_mask &&
+						  !entry->msi_attrib.is_virtual;
+
 		entry->msi_attrib.default_irq	= dev->irq;
 		entry->mask_base		= base;
 
-		if (!entry->msi_attrib.is_virtual) {
+		if (entry->msi_attrib.can_mask) {
 			addr = pci_msix_desc_addr(entry);
 			entry->msix_ctrl = readl(addr + PCI_MSIX_ENTRY_VECTOR_CTRL);
 		}
diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
index 1579a37..3d2fb39 100644
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -5106,12 +5106,13 @@ static int pci_reset_bus_function(struct pci_dev *dev, bool probe)
 	return pci_parent_bus_reset(dev, probe);
 }
 
-static void pci_dev_lock(struct pci_dev *dev)
+void pci_dev_lock(struct pci_dev *dev)
 {
 	pci_cfg_access_lock(dev);
 	/* block PM suspend, driver probe, etc. */
 	device_lock(&dev->dev);
 }
+EXPORT_SYMBOL_GPL(pci_dev_lock);
 
 /* Return 1 on successful lock, 0 on contention */
 int pci_dev_trylock(struct pci_dev *dev)
diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
index aedb78c..003950c 100644
--- a/drivers/pci/quirks.c
+++ b/drivers/pci/quirks.c
@@ -5851,3 +5851,9 @@ DECLARE_PCI_FIXUP_ENABLE(PCI_VENDOR_ID_PERICOM, 0x2303,
 			 pci_fixup_pericom_acs_store_forward);
 DECLARE_PCI_FIXUP_RESUME(PCI_VENDOR_ID_PERICOM, 0x2303,
 			 pci_fixup_pericom_acs_store_forward);
+
+static void nvidia_ion_ahci_fixup(struct pci_dev *pdev)
+{
+	pdev->dev_flags |= PCI_DEV_FLAGS_HAS_MSI_MASKING;
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_NVIDIA, 0x0ab8, nvidia_ion_ahci_fixup);
diff --git a/drivers/pwm/Kconfig b/drivers/pwm/Kconfig
index aa29841..21e3b05 100644
--- a/drivers/pwm/Kconfig
+++ b/drivers/pwm/Kconfig
@@ -476,7 +476,9 @@
 	depends on PLAT_SAMSUNG || ARCH_S5PV210 || ARCH_EXYNOS || COMPILE_TEST
 	depends on HAS_IOMEM
 	help
-	  Generic PWM framework driver for Samsung.
+	  Generic PWM framework driver for Samsung S3C24xx, S3C64xx, S5Pv210
+	  and Exynos SoCs.
+	  Choose Y here only if you build for such Samsung SoC.
 
 	  To compile this driver as a module, choose M here: the module
 	  will be called pwm-samsung.
diff --git a/drivers/pwm/core.c b/drivers/pwm/core.c
index 4527f09..fb04a43 100644
--- a/drivers/pwm/core.c
+++ b/drivers/pwm/core.c
@@ -532,6 +532,15 @@ int pwm_apply_state(struct pwm_device *pwm, const struct pwm_state *state)
 	struct pwm_chip *chip;
 	int err;
 
+	/*
+	 * Some lowlevel driver's implementations of .apply() make use of
+	 * mutexes, also with some drivers only returning when the new
+	 * configuration is active calling pwm_apply_state() from atomic context
+	 * is a bad idea. So make it explicit that calling this function might
+	 * sleep.
+	 */
+	might_sleep();
+
 	if (!pwm || !state || !state->period ||
 	    state->duty_cycle > state->period)
 		return -EINVAL;
diff --git a/drivers/pwm/pwm-atmel.c b/drivers/pwm/pwm-atmel.c
index e748604..98b34ea 100644
--- a/drivers/pwm/pwm-atmel.c
+++ b/drivers/pwm/pwm-atmel.c
@@ -24,7 +24,6 @@
 #include <linux/err.h>
 #include <linux/io.h>
 #include <linux/module.h>
-#include <linux/mutex.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
 #include <linux/platform_device.h>
diff --git a/drivers/pwm/pwm-samsung.c b/drivers/pwm/pwm-samsung.c
index dd94c431..0a4ff55 100644
--- a/drivers/pwm/pwm-samsung.c
+++ b/drivers/pwm/pwm-samsung.c
@@ -117,6 +117,20 @@ static inline unsigned int to_tcon_channel(unsigned int channel)
 	return (channel == 0) ? 0 : (channel + 1);
 }
 
+static void __pwm_samsung_manual_update(struct samsung_pwm_chip *chip,
+				      struct pwm_device *pwm)
+{
+	unsigned int tcon_chan = to_tcon_channel(pwm->hwpwm);
+	u32 tcon;
+
+	tcon = readl(chip->base + REG_TCON);
+	tcon |= TCON_MANUALUPDATE(tcon_chan);
+	writel(tcon, chip->base + REG_TCON);
+
+	tcon &= ~TCON_MANUALUPDATE(tcon_chan);
+	writel(tcon, chip->base + REG_TCON);
+}
+
 static void pwm_samsung_set_divisor(struct samsung_pwm_chip *pwm,
 				    unsigned int channel, u8 divisor)
 {
@@ -276,6 +290,13 @@ static void pwm_samsung_disable(struct pwm_chip *chip, struct pwm_device *pwm)
 	tcon &= ~TCON_AUTORELOAD(tcon_chan);
 	writel(tcon, our_chip->base + REG_TCON);
 
+	/*
+	 * In case the PWM is at 100% duty cycle, force a manual
+	 * update to prevent the signal from staying high.
+	 */
+	if (readl(our_chip->base + REG_TCMPB(pwm->hwpwm)) == (u32)-1U)
+		__pwm_samsung_manual_update(our_chip, pwm);
+
 	our_chip->disabled_mask |= BIT(pwm->hwpwm);
 
 	spin_unlock_irqrestore(&samsung_pwm_lock, flags);
@@ -284,18 +305,11 @@ static void pwm_samsung_disable(struct pwm_chip *chip, struct pwm_device *pwm)
 static void pwm_samsung_manual_update(struct samsung_pwm_chip *chip,
 				      struct pwm_device *pwm)
 {
-	unsigned int tcon_chan = to_tcon_channel(pwm->hwpwm);
-	u32 tcon;
 	unsigned long flags;
 
 	spin_lock_irqsave(&samsung_pwm_lock, flags);
 
-	tcon = readl(chip->base + REG_TCON);
-	tcon |= TCON_MANUALUPDATE(tcon_chan);
-	writel(tcon, chip->base + REG_TCON);
-
-	tcon &= ~TCON_MANUALUPDATE(tcon_chan);
-	writel(tcon, chip->base + REG_TCON);
+	__pwm_samsung_manual_update(chip, pwm);
 
 	spin_unlock_irqrestore(&samsung_pwm_lock, flags);
 }
diff --git a/drivers/pwm/pwm-visconti.c b/drivers/pwm/pwm-visconti.c
index af4e37d..927c4cb 100644
--- a/drivers/pwm/pwm-visconti.c
+++ b/drivers/pwm/pwm-visconti.c
@@ -144,28 +144,17 @@ static int visconti_pwm_probe(struct platform_device *pdev)
 	if (IS_ERR(priv->base))
 		return PTR_ERR(priv->base);
 
-	platform_set_drvdata(pdev, priv);
-
 	priv->chip.dev = dev;
 	priv->chip.ops = &visconti_pwm_ops;
 	priv->chip.npwm = 4;
 
-	ret = pwmchip_add(&priv->chip);
+	ret = devm_pwmchip_add(&pdev->dev, &priv->chip);
 	if (ret < 0)
 		return dev_err_probe(&pdev->dev, ret, "Cannot register visconti PWM\n");
 
 	return 0;
 }
 
-static int visconti_pwm_remove(struct platform_device *pdev)
-{
-	struct visconti_pwm_chip *priv = platform_get_drvdata(pdev);
-
-	pwmchip_remove(&priv->chip);
-
-	return 0;
-}
-
 static const struct of_device_id visconti_pwm_of_match[] = {
 	{ .compatible = "toshiba,visconti-pwm", },
 	{ }
@@ -178,7 +167,6 @@ static struct platform_driver visconti_pwm_driver = {
 		.of_match_table = visconti_pwm_of_match,
 	},
 	.probe = visconti_pwm_probe,
-	.remove = visconti_pwm_remove,
 };
 module_platform_driver(visconti_pwm_driver);
 
diff --git a/drivers/pwm/pwm-vt8500.c b/drivers/pwm/pwm-vt8500.c
index ea2aa15..480bfc2 100644
--- a/drivers/pwm/pwm-vt8500.c
+++ b/drivers/pwm/pwm-vt8500.c
@@ -56,7 +56,7 @@ struct vt8500_chip {
 #define to_vt8500_chip(chip)	container_of(chip, struct vt8500_chip, chip)
 
 #define msecs_to_loops(t) (loops_per_jiffy / 1000 * HZ * t)
-static inline void pwm_busy_wait(struct vt8500_chip *vt8500, int nr, u8 bitmask)
+static inline void vt8500_pwm_busy_wait(struct vt8500_chip *vt8500, int nr, u8 bitmask)
 {
 	int loops = msecs_to_loops(10);
 	u32 mask = bitmask << (nr << 8);
@@ -106,18 +106,18 @@ static int vt8500_pwm_config(struct pwm_chip *chip, struct pwm_device *pwm,
 	dc = c;
 
 	writel(prescale, vt8500->base + REG_SCALAR(pwm->hwpwm));
-	pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_SCALAR_UPDATE);
+	vt8500_pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_SCALAR_UPDATE);
 
 	writel(pv, vt8500->base + REG_PERIOD(pwm->hwpwm));
-	pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_PERIOD_UPDATE);
+	vt8500_pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_PERIOD_UPDATE);
 
 	writel(dc, vt8500->base + REG_DUTY(pwm->hwpwm));
-	pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_DUTY_UPDATE);
+	vt8500_pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_DUTY_UPDATE);
 
 	val = readl(vt8500->base + REG_CTRL(pwm->hwpwm));
 	val |= CTRL_AUTOLOAD;
 	writel(val, vt8500->base + REG_CTRL(pwm->hwpwm));
-	pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_CTRL_UPDATE);
+	vt8500_pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_CTRL_UPDATE);
 
 	clk_disable(vt8500->clk);
 	return 0;
@@ -138,7 +138,7 @@ static int vt8500_pwm_enable(struct pwm_chip *chip, struct pwm_device *pwm)
 	val = readl(vt8500->base + REG_CTRL(pwm->hwpwm));
 	val |= CTRL_ENABLE;
 	writel(val, vt8500->base + REG_CTRL(pwm->hwpwm));
-	pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_CTRL_UPDATE);
+	vt8500_pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_CTRL_UPDATE);
 
 	return 0;
 }
@@ -151,7 +151,7 @@ static void vt8500_pwm_disable(struct pwm_chip *chip, struct pwm_device *pwm)
 	val = readl(vt8500->base + REG_CTRL(pwm->hwpwm));
 	val &= ~CTRL_ENABLE;
 	writel(val, vt8500->base + REG_CTRL(pwm->hwpwm));
-	pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_CTRL_UPDATE);
+	vt8500_pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_CTRL_UPDATE);
 
 	clk_disable(vt8500->clk);
 }
@@ -171,7 +171,7 @@ static int vt8500_pwm_set_polarity(struct pwm_chip *chip,
 		val &= ~CTRL_INVERT;
 
 	writel(val, vt8500->base + REG_CTRL(pwm->hwpwm));
-	pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_CTRL_UPDATE);
+	vt8500_pwm_busy_wait(vt8500, pwm->hwpwm, STATUS_CTRL_UPDATE);
 
 	return 0;
 }
diff --git a/drivers/rtc/Kconfig b/drivers/rtc/Kconfig
index 7208eeb8..058e56a1 100644
--- a/drivers/rtc/Kconfig
+++ b/drivers/rtc/Kconfig
@@ -441,6 +441,7 @@
 
 config RTC_DRV_PCF8523
 	tristate "NXP PCF8523"
+	select REGMAP_I2C
 	help
 	  If you say yes here you get support for the NXP PCF8523 RTC
 	  chips.
@@ -582,14 +583,6 @@
 	  This driver can also be built as a module. If so, the module
 	  will be called rtc-tps65910.
 
-config RTC_DRV_TPS80031
-	tristate "TI TPS80031/TPS80032 RTC driver"
-	depends on MFD_TPS80031
-	help
-	  TI Power Management IC TPS80031 supports RTC functionality
-	  along with alarm. This driver supports the RTC driver for
-	  the TPS80031 RTC module.
-
 config RTC_DRV_RC5T583
 	tristate "RICOH 5T583 RTC driver"
 	depends on MFD_RC5T583
@@ -1929,4 +1922,14 @@
 	  This can also be built as a module. If so, the module will
 	  be named "rtc_wilco_ec".
 
+config RTC_DRV_MSC313
+	tristate "MStar MSC313 RTC"
+        depends on ARCH_MSTARV7 || COMPILE_TEST
+	help
+	  If you say yes here you get support for the Mstar MSC313e On-Chip
+	  Real Time Clock.
+
+	  This driver can also be built as a module, if so, the module
+	  will be called "rtc-msc313".
+
 endif # RTC_CLASS
diff --git a/drivers/rtc/Makefile b/drivers/rtc/Makefile
index 5ceeafe..678a8ef 100644
--- a/drivers/rtc/Makefile
+++ b/drivers/rtc/Makefile
@@ -103,6 +103,7 @@
 obj-$(CONFIG_RTC_DRV_MESON)	+= rtc-meson.o
 obj-$(CONFIG_RTC_DRV_MOXART)	+= rtc-moxart.o
 obj-$(CONFIG_RTC_DRV_MPC5121)	+= rtc-mpc5121.o
+obj-$(CONFIG_RTC_DRV_MSC313)	+= rtc-msc313.o
 obj-$(CONFIG_RTC_DRV_MSM6242)	+= rtc-msm6242.o
 obj-$(CONFIG_RTC_DRV_MT2712)	+= rtc-mt2712.o
 obj-$(CONFIG_RTC_DRV_MT6397)	+= rtc-mt6397.o
@@ -169,7 +170,6 @@
 obj-$(CONFIG_RTC_DRV_TEST)	+= rtc-test.o
 obj-$(CONFIG_RTC_DRV_TPS6586X)	+= rtc-tps6586x.o
 obj-$(CONFIG_RTC_DRV_TPS65910)	+= rtc-tps65910.o
-obj-$(CONFIG_RTC_DRV_TPS80031)	+= rtc-tps80031.o
 obj-$(CONFIG_RTC_DRV_TWL4030)	+= rtc-twl.o
 obj-$(CONFIG_RTC_DRV_V3020)	+= rtc-v3020.o
 obj-$(CONFIG_RTC_DRV_VR41XX)	+= rtc-vr41xx.o
diff --git a/drivers/rtc/class.c b/drivers/rtc/class.c
index f77bc08..4b460c6 100644
--- a/drivers/rtc/class.c
+++ b/drivers/rtc/class.c
@@ -232,6 +232,7 @@ static struct rtc_device *rtc_allocate_device(void)
 	rtc->pie_enabled = 0;
 
 	set_bit(RTC_FEATURE_ALARM, rtc->features);
+	set_bit(RTC_FEATURE_UPDATE_INTERRUPT, rtc->features);
 
 	return rtc;
 }
@@ -334,7 +335,8 @@ static void devm_rtc_unregister_device(void *data)
 	 * letting any rtc_class_open() users access it again
 	 */
 	rtc_proc_del_device(rtc);
-	cdev_device_del(&rtc->char_dev, &rtc->dev);
+	if (!test_bit(RTC_NO_CDEV, &rtc->flags))
+		cdev_device_del(&rtc->char_dev, &rtc->dev);
 	rtc->ops = NULL;
 	mutex_unlock(&rtc->ops_lock);
 }
@@ -363,7 +365,9 @@ struct rtc_device *devm_rtc_allocate_device(struct device *dev)
 
 	rtc->id = id;
 	rtc->dev.parent = dev;
-	dev_set_name(&rtc->dev, "rtc%d", id);
+	err = dev_set_name(&rtc->dev, "rtc%d", id);
+	if (err)
+		return ERR_PTR(err);
 
 	err = devm_add_action_or_reset(dev, devm_rtc_release_device, rtc);
 	if (err)
@@ -386,6 +390,12 @@ int __devm_rtc_register_device(struct module *owner, struct rtc_device *rtc)
 	if (!rtc->ops->set_alarm)
 		clear_bit(RTC_FEATURE_ALARM, rtc->features);
 
+	if (rtc->uie_unsupported)
+		clear_bit(RTC_FEATURE_UPDATE_INTERRUPT, rtc->features);
+
+	if (rtc->ops->set_offset)
+		set_bit(RTC_FEATURE_CORRECTION, rtc->features);
+
 	rtc->owner = owner;
 	rtc_device_get_offset(rtc);
 
@@ -397,12 +407,14 @@ int __devm_rtc_register_device(struct module *owner, struct rtc_device *rtc)
 	rtc_dev_prepare(rtc);
 
 	err = cdev_device_add(&rtc->char_dev, &rtc->dev);
-	if (err)
+	if (err) {
+		set_bit(RTC_NO_CDEV, &rtc->flags);
 		dev_warn(rtc->dev.parent, "failed to add char device %d:%d\n",
 			 MAJOR(rtc->dev.devt), rtc->id);
-	else
+	} else {
 		dev_dbg(rtc->dev.parent, "char device (%d:%d)\n",
 			MAJOR(rtc->dev.devt), rtc->id);
+	}
 
 	rtc_proc_add_device(rtc);
 
diff --git a/drivers/rtc/dev.c b/drivers/rtc/dev.c
index 5b8ebe8..e104972 100644
--- a/drivers/rtc/dev.c
+++ b/drivers/rtc/dev.c
@@ -208,6 +208,7 @@ static long rtc_dev_ioctl(struct file *file,
 	const struct rtc_class_ops *ops = rtc->ops;
 	struct rtc_time tm;
 	struct rtc_wkalrm alarm;
+	struct rtc_param param;
 	void __user *uarg = (void __user *)arg;
 
 	err = mutex_lock_interruptible(&rtc->ops_lock);
@@ -221,6 +222,7 @@ static long rtc_dev_ioctl(struct file *file,
 	switch (cmd) {
 	case RTC_EPOCH_SET:
 	case RTC_SET_TIME:
+	case RTC_PARAM_SET:
 		if (!capable(CAP_SYS_TIME))
 			err = -EACCES;
 		break;
@@ -382,6 +384,69 @@ static long rtc_dev_ioctl(struct file *file,
 			err = -EFAULT;
 		return err;
 
+	case RTC_PARAM_GET:
+		if (copy_from_user(&param, uarg, sizeof(param))) {
+			mutex_unlock(&rtc->ops_lock);
+			return -EFAULT;
+		}
+
+		switch(param.param) {
+			long offset;
+		case RTC_PARAM_FEATURES:
+			if (param.index != 0)
+				err = -EINVAL;
+			param.uvalue = rtc->features[0];
+			break;
+
+		case RTC_PARAM_CORRECTION:
+			mutex_unlock(&rtc->ops_lock);
+			if (param.index != 0)
+				return -EINVAL;
+			err = rtc_read_offset(rtc, &offset);
+			mutex_lock(&rtc->ops_lock);
+			if (err == 0)
+				param.svalue = offset;
+			break;
+
+		default:
+			if (rtc->ops->param_get)
+				err = rtc->ops->param_get(rtc->dev.parent, &param);
+			else
+				err = -EINVAL;
+		}
+
+		if (!err)
+			if (copy_to_user(uarg, &param, sizeof(param)))
+				err = -EFAULT;
+
+		break;
+
+	case RTC_PARAM_SET:
+		if (copy_from_user(&param, uarg, sizeof(param))) {
+			mutex_unlock(&rtc->ops_lock);
+			return -EFAULT;
+		}
+
+		switch(param.param) {
+		case RTC_PARAM_FEATURES:
+			err = -EINVAL;
+			break;
+
+		case RTC_PARAM_CORRECTION:
+			mutex_unlock(&rtc->ops_lock);
+			if (param.index != 0)
+				return -EINVAL;
+			return rtc_set_offset(rtc, param.svalue);
+
+		default:
+			if (rtc->ops->param_set)
+				err = rtc->ops->param_set(rtc->dev.parent, &param);
+			else
+				err = -EINVAL;
+		}
+
+		break;
+
 	default:
 		/* Finally try the driver's ioctl interface */
 		if (ops->ioctl) {
diff --git a/drivers/rtc/interface.c b/drivers/rtc/interface.c
index 9a2bd49..d8e8357 100644
--- a/drivers/rtc/interface.c
+++ b/drivers/rtc/interface.c
@@ -423,6 +423,7 @@ static int __rtc_set_alarm(struct rtc_device *rtc, struct rtc_wkalrm *alarm)
 	if (err)
 		return err;
 	now = rtc_tm_to_time64(&tm);
+
 	if (scheduled <= now)
 		return -ETIME;
 	/*
@@ -447,6 +448,7 @@ static int __rtc_set_alarm(struct rtc_device *rtc, struct rtc_wkalrm *alarm)
 
 int rtc_set_alarm(struct rtc_device *rtc, struct rtc_wkalrm *alarm)
 {
+	ktime_t alarm_time;
 	int err;
 
 	if (!rtc->ops)
@@ -468,7 +470,15 @@ int rtc_set_alarm(struct rtc_device *rtc, struct rtc_wkalrm *alarm)
 	if (rtc->aie_timer.enabled)
 		rtc_timer_remove(rtc, &rtc->aie_timer);
 
-	rtc->aie_timer.node.expires = rtc_tm_to_ktime(alarm->time);
+	alarm_time = rtc_tm_to_ktime(alarm->time);
+	/*
+	 * Round down so we never miss a deadline, checking for past deadline is
+	 * done in __rtc_set_alarm
+	 */
+	if (test_bit(RTC_FEATURE_ALARM_RES_MINUTE, rtc->features))
+		alarm_time = ktime_sub_ns(alarm_time, (u64)alarm->time.tm_sec * NSEC_PER_SEC);
+
+	rtc->aie_timer.node.expires = alarm_time;
 	rtc->aie_timer.period = 0;
 	if (alarm->enabled)
 		err = rtc_timer_enqueue(rtc, &rtc->aie_timer);
@@ -561,7 +571,8 @@ int rtc_update_irq_enable(struct rtc_device *rtc, unsigned int enabled)
 	if (rtc->uie_rtctimer.enabled == enabled)
 		goto out;
 
-	if (rtc->uie_unsupported || !test_bit(RTC_FEATURE_ALARM, rtc->features)) {
+	if (!test_bit(RTC_FEATURE_UPDATE_INTERRUPT, rtc->features) ||
+	    !test_bit(RTC_FEATURE_ALARM, rtc->features)) {
 		mutex_unlock(&rtc->ops_lock);
 #ifdef CONFIG_RTC_INTF_DEV_UIE_EMUL
 		return rtc_dev_update_irq_enable_emul(rtc, enabled);
diff --git a/drivers/rtc/rtc-ab-eoz9.c b/drivers/rtc/rtc-ab-eoz9.c
index a9b3555..e188ab5 100644
--- a/drivers/rtc/rtc-ab-eoz9.c
+++ b/drivers/rtc/rtc-ab-eoz9.c
@@ -534,7 +534,6 @@ static int abeoz9_probe(struct i2c_client *client,
 	data->rtc->ops = &rtc_ops;
 	data->rtc->range_min = RTC_TIMESTAMP_BEGIN_2000;
 	data->rtc->range_max = RTC_TIMESTAMP_END_2099;
-	data->rtc->uie_unsupported = 1;
 	clear_bit(RTC_FEATURE_ALARM, data->rtc->features);
 
 	if (client->irq > 0) {
@@ -546,6 +545,8 @@ static int abeoz9_probe(struct i2c_client *client,
 			dev_err(dev, "failed to request alarm irq\n");
 			return ret;
 		}
+	} else {
+		clear_bit(RTC_FEATURE_UPDATE_INTERRUPT, data->rtc->features);
 	}
 
 	if (client->irq > 0 || device_property_read_bool(dev, "wakeup-source")) {
diff --git a/drivers/rtc/rtc-ab8500.c b/drivers/rtc/rtc-ab8500.c
index b400488..ea33e14 100644
--- a/drivers/rtc/rtc-ab8500.c
+++ b/drivers/rtc/rtc-ab8500.c
@@ -184,25 +184,9 @@ static int ab8500_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alarm)
 {
 	int retval, i;
 	unsigned char buf[ARRAY_SIZE(ab8500_rtc_alarm_regs)];
-	unsigned long mins, secs = 0, cursec = 0;
-	struct rtc_time curtm;
+	unsigned long mins;
 
-	/* Get the number of seconds since 1970 */
-	secs = rtc_tm_to_time64(&alarm->time);
-
-	/*
-	 * Check whether alarm is set less than 1min.
-	 * Since our RTC doesn't support alarm resolution less than 1min,
-	 * return -EINVAL, so UIE EMUL can take it up, incase of UIE_ON
-	 */
-	ab8500_rtc_read_time(dev, &curtm); /* Read current time */
-	cursec = rtc_tm_to_time64(&curtm);
-	if ((secs - cursec) < 59) {
-		dev_dbg(dev, "Alarm less than 1 minute not supported\r\n");
-		return -EINVAL;
-	}
-
-	mins = secs / 60;
+	mins = (unsigned long)rtc_tm_to_time64(&alarm->time) / 60;
 
 	buf[2] = mins & 0xFF;
 	buf[1] = (mins >> 8) & 0xFF;
@@ -394,7 +378,8 @@ static int ab8500_rtc_probe(struct platform_device *pdev)
 	dev_pm_set_wake_irq(&pdev->dev, irq);
 	platform_set_drvdata(pdev, rtc);
 
-	rtc->uie_unsupported = 1;
+	set_bit(RTC_FEATURE_ALARM_RES_MINUTE, rtc->features);
+	clear_bit(RTC_FEATURE_UPDATE_INTERRUPT, rtc->features);
 
 	rtc->range_max = (1ULL << 24) * 60 - 1; // 24-bit minutes + 59 secs
 	rtc->start_secs = RTC_TIMESTAMP_BEGIN_2000;
diff --git a/drivers/rtc/rtc-ds1302.c b/drivers/rtc/rtc-ds1302.c
index b3de6d2..2f83ade 100644
--- a/drivers/rtc/rtc-ds1302.c
+++ b/drivers/rtc/rtc-ds1302.c
@@ -199,11 +199,18 @@ static const struct of_device_id ds1302_dt_ids[] = {
 MODULE_DEVICE_TABLE(of, ds1302_dt_ids);
 #endif
 
+static const struct spi_device_id ds1302_spi_ids[] = {
+	{ .name = "ds1302", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(spi, ds1302_spi_ids);
+
 static struct spi_driver ds1302_driver = {
 	.driver.name	= "rtc-ds1302",
 	.driver.of_match_table = of_match_ptr(ds1302_dt_ids),
 	.probe		= ds1302_probe,
 	.remove		= ds1302_remove,
+	.id_table	= ds1302_spi_ids,
 };
 
 module_spi_driver(ds1302_driver);
diff --git a/drivers/rtc/rtc-ds1390.c b/drivers/rtc/rtc-ds1390.c
index 66fc861..93ce72b 100644
--- a/drivers/rtc/rtc-ds1390.c
+++ b/drivers/rtc/rtc-ds1390.c
@@ -219,12 +219,19 @@ static const struct of_device_id ds1390_of_match[] = {
 };
 MODULE_DEVICE_TABLE(of, ds1390_of_match);
 
+static const struct spi_device_id ds1390_spi_ids[] = {
+	{ .name = "ds1390" },
+	{}
+};
+MODULE_DEVICE_TABLE(spi, ds1390_spi_ids);
+
 static struct spi_driver ds1390_driver = {
 	.driver = {
 		.name	= "rtc-ds1390",
 		.of_match_table = of_match_ptr(ds1390_of_match),
 	},
 	.probe	= ds1390_probe,
+	.id_table = ds1390_spi_ids,
 };
 
 module_spi_driver(ds1390_driver);
diff --git a/drivers/rtc/rtc-m41t80.c b/drivers/rtc/rtc-m41t80.c
index f736f8c..6d383b6 100644
--- a/drivers/rtc/rtc-m41t80.c
+++ b/drivers/rtc/rtc-m41t80.c
@@ -557,7 +557,7 @@ static struct clk *m41t80_sqw_register_clk(struct m41t80_data *m41t80)
 		 * registered automatically when being referenced.
 		 */
 		of_node_put(fixed_clock);
-		return 0;
+		return NULL;
 	}
 
 	/* First disable the clock */
diff --git a/drivers/rtc/rtc-mcp795.c b/drivers/rtc/rtc-mcp795.c
index bad7792..0d515b3 100644
--- a/drivers/rtc/rtc-mcp795.c
+++ b/drivers/rtc/rtc-mcp795.c
@@ -430,12 +430,19 @@ static const struct of_device_id mcp795_of_match[] = {
 MODULE_DEVICE_TABLE(of, mcp795_of_match);
 #endif
 
+static const struct spi_device_id mcp795_spi_ids[] = {
+	{ .name = "mcp795" },
+	{ }
+};
+MODULE_DEVICE_TABLE(spi, mcp795_spi_ids);
+
 static struct spi_driver mcp795_driver = {
 		.driver = {
 				.name = "rtc-mcp795",
 				.of_match_table = of_match_ptr(mcp795_of_match),
 		},
 		.probe = mcp795_probe,
+		.id_table = mcp795_spi_ids,
 };
 
 module_spi_driver(mcp795_driver);
diff --git a/drivers/rtc/rtc-msc313.c b/drivers/rtc/rtc-msc313.c
new file mode 100644
index 0000000..f3fde01
--- /dev/null
+++ b/drivers/rtc/rtc-msc313.c
@@ -0,0 +1,259 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Real time clocks driver for MStar/SigmaStar ARMv7 SoCs.
+ * Based on "Real Time Clock driver for msb252x." that was contained
+ * in various MStar kernels.
+ *
+ * (C) 2019 Daniel Palmer
+ * (C) 2021 Romain Perier
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/mod_devicetable.h>
+#include <linux/platform_device.h>
+#include <linux/rtc.h>
+
+/* Registers */
+#define REG_RTC_CTRL		0x00
+#define REG_RTC_FREQ_CW_L	0x04
+#define REG_RTC_FREQ_CW_H	0x08
+#define REG_RTC_LOAD_VAL_L	0x0C
+#define REG_RTC_LOAD_VAL_H	0x10
+#define REG_RTC_MATCH_VAL_L	0x14
+#define REG_RTC_MATCH_VAL_H	0x18
+#define REG_RTC_STATUS_INT	0x1C
+#define REG_RTC_CNT_VAL_L	0x20
+#define REG_RTC_CNT_VAL_H	0x24
+
+/* Control bits for REG_RTC_CTRL */
+#define SOFT_RSTZ_BIT		BIT(0)
+#define CNT_EN_BIT		BIT(1)
+#define WRAP_EN_BIT		BIT(2)
+#define LOAD_EN_BIT		BIT(3)
+#define READ_EN_BIT		BIT(4)
+#define INT_MASK_BIT		BIT(5)
+#define INT_FORCE_BIT		BIT(6)
+#define INT_CLEAR_BIT		BIT(7)
+
+/* Control bits for REG_RTC_STATUS_INT */
+#define RAW_INT_BIT		BIT(0)
+#define ALM_INT_BIT		BIT(1)
+
+struct msc313_rtc {
+	struct rtc_device *rtc_dev;
+	void __iomem *rtc_base;
+};
+
+static int msc313_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alarm)
+{
+	struct msc313_rtc *priv = dev_get_drvdata(dev);
+	unsigned long seconds;
+
+	seconds = readw(priv->rtc_base + REG_RTC_MATCH_VAL_L)
+			| ((unsigned long)readw(priv->rtc_base + REG_RTC_MATCH_VAL_H) << 16);
+
+	rtc_time64_to_tm(seconds, &alarm->time);
+
+	if (!(readw(priv->rtc_base + REG_RTC_CTRL) & INT_MASK_BIT))
+		alarm->enabled = 1;
+
+	return 0;
+}
+
+static int msc313_rtc_alarm_irq_enable(struct device *dev, unsigned int enabled)
+{
+	struct msc313_rtc *priv = dev_get_drvdata(dev);
+	u16 reg;
+
+	reg = readw(priv->rtc_base + REG_RTC_CTRL);
+	if (enabled)
+		reg &= ~INT_MASK_BIT;
+	else
+		reg |= INT_MASK_BIT;
+	writew(reg, priv->rtc_base + REG_RTC_CTRL);
+	return 0;
+}
+
+static int msc313_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alarm)
+{
+	struct msc313_rtc *priv = dev_get_drvdata(dev);
+	unsigned long seconds;
+
+	seconds = rtc_tm_to_time64(&alarm->time);
+	writew((seconds & 0xFFFF), priv->rtc_base + REG_RTC_MATCH_VAL_L);
+	writew((seconds >> 16) & 0xFFFF, priv->rtc_base + REG_RTC_MATCH_VAL_H);
+
+	msc313_rtc_alarm_irq_enable(dev, alarm->enabled);
+
+	return 0;
+}
+
+static bool msc313_rtc_get_enabled(struct msc313_rtc *priv)
+{
+	return readw(priv->rtc_base + REG_RTC_CTRL) & CNT_EN_BIT;
+}
+
+static void msc313_rtc_set_enabled(struct msc313_rtc *priv)
+{
+	u16 reg;
+
+	reg = readw(priv->rtc_base + REG_RTC_CTRL);
+	reg |= CNT_EN_BIT;
+	writew(reg, priv->rtc_base + REG_RTC_CTRL);
+}
+
+static int msc313_rtc_read_time(struct device *dev, struct rtc_time *tm)
+{
+	struct msc313_rtc *priv = dev_get_drvdata(dev);
+	u32 seconds;
+	u16 reg;
+
+	if (!msc313_rtc_get_enabled(priv))
+		return -EINVAL;
+
+	reg = readw(priv->rtc_base + REG_RTC_CTRL);
+	writew(reg | READ_EN_BIT, priv->rtc_base + REG_RTC_CTRL);
+
+	/* Wait for HW latch done */
+	while (readw(priv->rtc_base + REG_RTC_CTRL) & READ_EN_BIT)
+		udelay(1);
+
+	seconds = readw(priv->rtc_base + REG_RTC_CNT_VAL_L)
+			| ((unsigned long)readw(priv->rtc_base + REG_RTC_CNT_VAL_H) << 16);
+
+	rtc_time64_to_tm(seconds, tm);
+
+	return 0;
+}
+
+static int msc313_rtc_set_time(struct device *dev, struct rtc_time *tm)
+{
+	struct msc313_rtc *priv = dev_get_drvdata(dev);
+	unsigned long seconds;
+	u16 reg;
+
+	seconds = rtc_tm_to_time64(tm);
+	writew(seconds & 0xFFFF, priv->rtc_base + REG_RTC_LOAD_VAL_L);
+	writew((seconds >> 16) & 0xFFFF, priv->rtc_base + REG_RTC_LOAD_VAL_H);
+
+	/* Enable load for loading value into internal RTC counter */
+	reg = readw(priv->rtc_base + REG_RTC_CTRL);
+	writew(reg | LOAD_EN_BIT, priv->rtc_base + REG_RTC_CTRL);
+
+	/* Wait for HW latch done */
+	while (readw(priv->rtc_base + REG_RTC_CTRL) & LOAD_EN_BIT)
+		udelay(1);
+	msc313_rtc_set_enabled(priv);
+	return 0;
+}
+
+static const struct rtc_class_ops msc313_rtc_ops = {
+	.read_time = msc313_rtc_read_time,
+	.set_time = msc313_rtc_set_time,
+	.read_alarm = msc313_rtc_read_alarm,
+	.set_alarm = msc313_rtc_set_alarm,
+	.alarm_irq_enable = msc313_rtc_alarm_irq_enable,
+};
+
+static irqreturn_t msc313_rtc_interrupt(s32 irq, void *dev_id)
+{
+	struct msc313_rtc *priv = dev_get_drvdata(dev_id);
+	u16 reg;
+
+	reg = readw(priv->rtc_base + REG_RTC_STATUS_INT);
+	if (!(reg & ALM_INT_BIT))
+		return IRQ_NONE;
+
+	reg = readw(priv->rtc_base + REG_RTC_CTRL);
+	reg |= INT_CLEAR_BIT;
+	reg &= ~INT_FORCE_BIT;
+	writew(reg, priv->rtc_base + REG_RTC_CTRL);
+
+	rtc_update_irq(priv->rtc_dev, 1, RTC_IRQF | RTC_AF);
+
+	return IRQ_HANDLED;
+}
+
+static int msc313_rtc_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct msc313_rtc *priv;
+	unsigned long rate;
+	struct clk *clk;
+	int ret;
+	int irq;
+
+	priv = devm_kzalloc(&pdev->dev, sizeof(struct msc313_rtc), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->rtc_base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(priv->rtc_base))
+		return PTR_ERR(priv->rtc_base);
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return -EINVAL;
+
+	priv->rtc_dev = devm_rtc_allocate_device(dev);
+	if (IS_ERR(priv->rtc_dev))
+		return PTR_ERR(priv->rtc_dev);
+
+	priv->rtc_dev->ops = &msc313_rtc_ops;
+	priv->rtc_dev->range_max = U32_MAX;
+
+	ret = devm_request_irq(dev, irq, msc313_rtc_interrupt, IRQF_SHARED,
+			       dev_name(&pdev->dev), &pdev->dev);
+	if (ret) {
+		dev_err(dev, "Could not request IRQ\n");
+		return ret;
+	}
+
+	clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(clk)) {
+		dev_err(dev, "No input reference clock\n");
+		return PTR_ERR(clk);
+	}
+
+	ret = clk_prepare_enable(clk);
+	if (ret) {
+		dev_err(dev, "Failed to enable the reference clock, %d\n", ret);
+		return ret;
+	}
+
+	ret = devm_add_action_or_reset(dev, (void (*) (void *))clk_disable_unprepare, clk);
+	if (ret)
+		return ret;
+
+	rate = clk_get_rate(clk);
+	writew(rate & 0xFFFF, priv->rtc_base + REG_RTC_FREQ_CW_L);
+	writew((rate >> 16) & 0xFFFF, priv->rtc_base + REG_RTC_FREQ_CW_H);
+
+	platform_set_drvdata(pdev, priv);
+
+	return devm_rtc_register_device(priv->rtc_dev);
+}
+
+static const struct of_device_id msc313_rtc_of_match_table[] = {
+	{ .compatible = "mstar,msc313-rtc" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, msc313_rtc_of_match_table);
+
+static struct platform_driver msc313_rtc_driver = {
+	.probe = msc313_rtc_probe,
+	.driver = {
+		.name = "msc313-rtc",
+		.of_match_table = msc313_rtc_of_match_table,
+	},
+};
+
+module_platform_driver(msc313_rtc_driver);
+
+MODULE_AUTHOR("Daniel Palmer <daniel@thingy.jp>");
+MODULE_AUTHOR("Romain Perier <romain.perier@gmail.com>");
+MODULE_DESCRIPTION("MStar RTC Driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/rtc/rtc-omap.c b/drivers/rtc/rtc-omap.c
index d46e0f0..4d4f3b1 100644
--- a/drivers/rtc/rtc-omap.c
+++ b/drivers/rtc/rtc-omap.c
@@ -1029,6 +1029,5 @@ static struct platform_driver omap_rtc_driver = {
 
 module_platform_driver(omap_rtc_driver);
 
-MODULE_ALIAS("platform:omap_rtc");
 MODULE_AUTHOR("George G. Davis (and others)");
 MODULE_LICENSE("GPL");
diff --git a/drivers/rtc/rtc-pcf2123.c b/drivers/rtc/rtc-pcf2123.c
index 0f58cac..7473e6c 100644
--- a/drivers/rtc/rtc-pcf2123.c
+++ b/drivers/rtc/rtc-pcf2123.c
@@ -451,12 +451,21 @@ static const struct of_device_id pcf2123_dt_ids[] = {
 MODULE_DEVICE_TABLE(of, pcf2123_dt_ids);
 #endif
 
+static const struct spi_device_id pcf2123_spi_ids[] = {
+	{ .name = "pcf2123", },
+	{ .name = "rv2123", },
+	{ .name = "rtc-pcf2123", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(spi, pcf2123_spi_ids);
+
 static struct spi_driver pcf2123_driver = {
 	.driver	= {
 			.name	= "rtc-pcf2123",
 			.of_match_table = of_match_ptr(pcf2123_dt_ids),
 	},
 	.probe	= pcf2123_probe,
+	.id_table = pcf2123_spi_ids,
 };
 
 module_spi_driver(pcf2123_driver);
diff --git a/drivers/rtc/rtc-pcf85063.c b/drivers/rtc/rtc-pcf85063.c
index 14da4ab..15e50bb 100644
--- a/drivers/rtc/rtc-pcf85063.c
+++ b/drivers/rtc/rtc-pcf85063.c
@@ -34,6 +34,7 @@
 #define PCF85063_REG_CTRL1		0x00 /* status */
 #define PCF85063_REG_CTRL1_CAP_SEL	BIT(0)
 #define PCF85063_REG_CTRL1_STOP		BIT(5)
+#define PCF85063_REG_CTRL1_EXT_TEST	BIT(7)
 
 #define PCF85063_REG_CTRL2		0x01
 #define PCF85063_CTRL2_AF		BIT(6)
@@ -117,6 +118,7 @@ static int pcf85063_rtc_set_time(struct device *dev, struct rtc_time *tm)
 	 * reset state until all time/date registers are written
 	 */
 	rc = regmap_update_bits(pcf85063->regmap, PCF85063_REG_CTRL1,
+				PCF85063_REG_CTRL1_EXT_TEST |
 				PCF85063_REG_CTRL1_STOP,
 				PCF85063_REG_CTRL1_STOP);
 	if (rc)
@@ -297,7 +299,7 @@ static int pcf85063_ioctl(struct device *dev, unsigned int cmd,
 		if (ret < 0)
 			return ret;
 
-		status = status & PCF85063_REG_SC_OS ? RTC_VL_DATA_INVALID : 0;
+		status = (status & PCF85063_REG_SC_OS) ? RTC_VL_DATA_INVALID : 0;
 
 		return put_user(status, (unsigned int __user *)arg);
 
@@ -479,6 +481,18 @@ static struct clk *pcf85063_clkout_register_clk(struct pcf85063 *pcf85063)
 	struct clk *clk;
 	struct clk_init_data init;
 	struct device_node *node = pcf85063->rtc->dev.parent->of_node;
+	struct device_node *fixed_clock;
+
+	fixed_clock = of_get_child_by_name(node, "clock");
+	if (fixed_clock) {
+		/*
+		 * skip registering square wave clock when a fixed
+		 * clock has been registered. The fixed clock is
+		 * registered automatically when being referenced.
+		 */
+		of_node_put(fixed_clock);
+		return NULL;
+	}
 
 	init.name = "pcf85063-clkout";
 	init.ops = &pcf85063_clkout_ops;
diff --git a/drivers/rtc/rtc-pcf8523.c b/drivers/rtc/rtc-pcf8523.c
index 8b6fb20..c93acad 100644
--- a/drivers/rtc/rtc-pcf8523.c
+++ b/drivers/rtc/rtc-pcf8523.c
@@ -4,8 +4,10 @@
  */
 
 #include <linux/bcd.h>
+#include <linux/bitfield.h>
 #include <linux/i2c.h>
 #include <linux/module.h>
+#include <linux/regmap.h>
 #include <linux/rtc.h>
 #include <linux/of.h>
 #include <linux/pm_wakeirq.h>
@@ -19,11 +21,10 @@
 #define PCF8523_CONTROL2_AF BIT(3)
 
 #define PCF8523_REG_CONTROL3 0x02
-#define PCF8523_CONTROL3_PM_BLD BIT(7) /* battery low detection disabled */
-#define PCF8523_CONTROL3_PM_VDD BIT(6) /* switch-over disabled */
-#define PCF8523_CONTROL3_PM_DSM BIT(5) /* direct switching mode */
-#define PCF8523_CONTROL3_PM_MASK 0xe0
+#define PCF8523_CONTROL3_PM  GENMASK(7,5)
+#define PCF8523_PM_STANDBY   0x7
 #define PCF8523_CONTROL3_BLF BIT(2) /* battery low bit, read-only */
+#define PCF8523_CONTROL3_BSF BIT(3)
 
 #define PCF8523_REG_SECONDS  0x03
 #define PCF8523_SECONDS_OS BIT(7)
@@ -48,127 +49,45 @@
 
 struct pcf8523 {
 	struct rtc_device *rtc;
-	struct i2c_client *client;
+	struct regmap *regmap;
 };
 
-static int pcf8523_read(struct i2c_client *client, u8 reg, u8 *valuep)
+static int pcf8523_load_capacitance(struct pcf8523 *pcf8523, struct device_node *node)
 {
-	struct i2c_msg msgs[2];
-	u8 value = 0;
-	int err;
-
-	msgs[0].addr = client->addr;
-	msgs[0].flags = 0;
-	msgs[0].len = sizeof(reg);
-	msgs[0].buf = &reg;
-
-	msgs[1].addr = client->addr;
-	msgs[1].flags = I2C_M_RD;
-	msgs[1].len = sizeof(value);
-	msgs[1].buf = &value;
-
-	err = i2c_transfer(client->adapter, msgs, ARRAY_SIZE(msgs));
-	if (err < 0)
-		return err;
-
-	*valuep = value;
-
-	return 0;
-}
-
-static int pcf8523_write(struct i2c_client *client, u8 reg, u8 value)
-{
-	u8 buffer[2] = { reg, value };
-	struct i2c_msg msg;
-	int err;
-
-	msg.addr = client->addr;
-	msg.flags = 0;
-	msg.len = sizeof(buffer);
-	msg.buf = buffer;
-
-	err = i2c_transfer(client->adapter, &msg, 1);
-	if (err < 0)
-		return err;
-
-	return 0;
-}
-
-static int pcf8523_voltage_low(struct i2c_client *client)
-{
-	u8 value;
-	int err;
-
-	err = pcf8523_read(client, PCF8523_REG_CONTROL3, &value);
-	if (err < 0)
-		return err;
-
-	return !!(value & PCF8523_CONTROL3_BLF);
-}
-
-static int pcf8523_load_capacitance(struct i2c_client *client)
-{
-	u32 load;
-	u8 value;
-	int err;
-
-	err = pcf8523_read(client, PCF8523_REG_CONTROL1, &value);
-	if (err < 0)
-		return err;
+	u32 load, value = 0;
 
 	load = 12500;
-	of_property_read_u32(client->dev.of_node, "quartz-load-femtofarads",
-			     &load);
+	of_property_read_u32(node, "quartz-load-femtofarads", &load);
 
 	switch (load) {
 	default:
-		dev_warn(&client->dev, "Unknown quartz-load-femtofarads value: %d. Assuming 12500",
+		dev_warn(&pcf8523->rtc->dev, "Unknown quartz-load-femtofarads value: %d. Assuming 12500",
 			 load);
 		fallthrough;
 	case 12500:
 		value |= PCF8523_CONTROL1_CAP_SEL;
 		break;
 	case 7000:
-		value &= ~PCF8523_CONTROL1_CAP_SEL;
 		break;
 	}
 
-	err = pcf8523_write(client, PCF8523_REG_CONTROL1, value);
-
-	return err;
-}
-
-static int pcf8523_set_pm(struct i2c_client *client, u8 pm)
-{
-	u8 value;
-	int err;
-
-	err = pcf8523_read(client, PCF8523_REG_CONTROL3, &value);
-	if (err < 0)
-		return err;
-
-	value = (value & ~PCF8523_CONTROL3_PM_MASK) | pm;
-
-	err = pcf8523_write(client, PCF8523_REG_CONTROL3, value);
-	if (err < 0)
-		return err;
-
-	return 0;
+	return regmap_update_bits(pcf8523->regmap, PCF8523_REG_CONTROL1,
+				  PCF8523_CONTROL1_CAP_SEL, value);
 }
 
 static irqreturn_t pcf8523_irq(int irq, void *dev_id)
 {
-	struct pcf8523 *pcf8523 = i2c_get_clientdata(dev_id);
-	u8 value;
+	struct pcf8523 *pcf8523 = dev_id;
+	u32 value;
 	int err;
 
-	err = pcf8523_read(pcf8523->client, PCF8523_REG_CONTROL2, &value);
+	err = regmap_read(pcf8523->regmap, PCF8523_REG_CONTROL2, &value);
 	if (err < 0)
 		return IRQ_HANDLED;
 
 	if (value & PCF8523_CONTROL2_AF) {
 		value &= ~PCF8523_CONTROL2_AF;
-		pcf8523_write(pcf8523->client, PCF8523_REG_CONTROL2, value);
+		regmap_write(pcf8523->regmap, PCF8523_REG_CONTROL2, value);
 		rtc_update_irq(pcf8523->rtc, 1, RTC_IRQF | RTC_AF);
 
 		return IRQ_HANDLED;
@@ -177,68 +96,14 @@ static irqreturn_t pcf8523_irq(int irq, void *dev_id)
 	return IRQ_NONE;
 }
 
-static int pcf8523_stop_rtc(struct i2c_client *client)
-{
-	u8 value;
-	int err;
-
-	err = pcf8523_read(client, PCF8523_REG_CONTROL1, &value);
-	if (err < 0)
-		return err;
-
-	value |= PCF8523_CONTROL1_STOP;
-
-	err = pcf8523_write(client, PCF8523_REG_CONTROL1, value);
-	if (err < 0)
-		return err;
-
-	return 0;
-}
-
-static int pcf8523_start_rtc(struct i2c_client *client)
-{
-	u8 value;
-	int err;
-
-	err = pcf8523_read(client, PCF8523_REG_CONTROL1, &value);
-	if (err < 0)
-		return err;
-
-	value &= ~PCF8523_CONTROL1_STOP;
-
-	err = pcf8523_write(client, PCF8523_REG_CONTROL1, value);
-	if (err < 0)
-		return err;
-
-	return 0;
-}
-
 static int pcf8523_rtc_read_time(struct device *dev, struct rtc_time *tm)
 {
-	struct i2c_client *client = to_i2c_client(dev);
-	u8 start = PCF8523_REG_SECONDS, regs[7];
-	struct i2c_msg msgs[2];
+	struct pcf8523 *pcf8523 = dev_get_drvdata(dev);
+	u8 regs[7];
 	int err;
 
-	err = pcf8523_voltage_low(client);
-	if (err < 0) {
-		return err;
-	} else if (err > 0) {
-		dev_err(dev, "low voltage detected, time is unreliable\n");
-		return -EINVAL;
-	}
-
-	msgs[0].addr = client->addr;
-	msgs[0].flags = 0;
-	msgs[0].len = 1;
-	msgs[0].buf = &start;
-
-	msgs[1].addr = client->addr;
-	msgs[1].flags = I2C_M_RD;
-	msgs[1].len = sizeof(regs);
-	msgs[1].buf = regs;
-
-	err = i2c_transfer(client->adapter, msgs, ARRAY_SIZE(msgs));
+	err = regmap_bulk_read(pcf8523->regmap, PCF8523_REG_SECONDS, regs,
+			       sizeof(regs));
 	if (err < 0)
 		return err;
 
@@ -258,63 +123,50 @@ static int pcf8523_rtc_read_time(struct device *dev, struct rtc_time *tm)
 
 static int pcf8523_rtc_set_time(struct device *dev, struct rtc_time *tm)
 {
-	struct i2c_client *client = to_i2c_client(dev);
-	struct i2c_msg msg;
-	u8 regs[8];
+	struct pcf8523 *pcf8523 = dev_get_drvdata(dev);
+	u8 regs[7];
 	int err;
 
-	err = pcf8523_stop_rtc(client);
+	err = regmap_update_bits(pcf8523->regmap, PCF8523_REG_CONTROL1,
+				 PCF8523_CONTROL1_STOP, PCF8523_CONTROL1_STOP);
 	if (err < 0)
 		return err;
 
-	regs[0] = PCF8523_REG_SECONDS;
 	/* This will purposely overwrite PCF8523_SECONDS_OS */
-	regs[1] = bin2bcd(tm->tm_sec);
-	regs[2] = bin2bcd(tm->tm_min);
-	regs[3] = bin2bcd(tm->tm_hour);
-	regs[4] = bin2bcd(tm->tm_mday);
-	regs[5] = tm->tm_wday;
-	regs[6] = bin2bcd(tm->tm_mon + 1);
-	regs[7] = bin2bcd(tm->tm_year - 100);
+	regs[0] = bin2bcd(tm->tm_sec);
+	regs[1] = bin2bcd(tm->tm_min);
+	regs[2] = bin2bcd(tm->tm_hour);
+	regs[3] = bin2bcd(tm->tm_mday);
+	regs[4] = tm->tm_wday;
+	regs[5] = bin2bcd(tm->tm_mon + 1);
+	regs[6] = bin2bcd(tm->tm_year - 100);
 
-	msg.addr = client->addr;
-	msg.flags = 0;
-	msg.len = sizeof(regs);
-	msg.buf = regs;
-
-	err = i2c_transfer(client->adapter, &msg, 1);
+	err = regmap_bulk_write(pcf8523->regmap, PCF8523_REG_SECONDS, regs,
+				sizeof(regs));
 	if (err < 0) {
 		/*
 		 * If the time cannot be set, restart the RTC anyway. Note
 		 * that errors are ignored if the RTC cannot be started so
 		 * that we have a chance to propagate the original error.
 		 */
-		pcf8523_start_rtc(client);
+		regmap_update_bits(pcf8523->regmap, PCF8523_REG_CONTROL1,
+				   PCF8523_CONTROL1_STOP, 0);
 		return err;
 	}
 
-	return pcf8523_start_rtc(client);
+	return regmap_update_bits(pcf8523->regmap, PCF8523_REG_CONTROL1,
+				 PCF8523_CONTROL1_STOP, 0);
 }
 
 static int pcf8523_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *tm)
 {
-	struct i2c_client *client = to_i2c_client(dev);
-	u8 start = PCF8523_REG_MINUTE_ALARM, regs[4];
-	struct i2c_msg msgs[2];
-	u8 value;
+	struct pcf8523 *pcf8523 = dev_get_drvdata(dev);
+	u8 regs[4];
+	u32 value;
 	int err;
 
-	msgs[0].addr = client->addr;
-	msgs[0].flags = 0;
-	msgs[0].len = 1;
-	msgs[0].buf = &start;
-
-	msgs[1].addr = client->addr;
-	msgs[1].flags = I2C_M_RD;
-	msgs[1].len = sizeof(regs);
-	msgs[1].buf = regs;
-
-	err = i2c_transfer(client->adapter, msgs, ARRAY_SIZE(msgs));
+	err = regmap_bulk_read(pcf8523->regmap, PCF8523_REG_MINUTE_ALARM, regs,
+			       sizeof(regs));
 	if (err < 0)
 		return err;
 
@@ -324,12 +176,12 @@ static int pcf8523_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *tm)
 	tm->time.tm_mday = bcd2bin(regs[2] & 0x3F);
 	tm->time.tm_wday = bcd2bin(regs[3] & 0x7);
 
-	err = pcf8523_read(client, PCF8523_REG_CONTROL1, &value);
+	err = regmap_read(pcf8523->regmap, PCF8523_REG_CONTROL1, &value);
 	if (err < 0)
 		return err;
 	tm->enabled = !!(value & PCF8523_CONTROL1_AIE);
 
-	err = pcf8523_read(client, PCF8523_REG_CONTROL2, &value);
+	err = regmap_read(pcf8523->regmap, PCF8523_REG_CONTROL2, &value);
 	if (err < 0)
 		return err;
 	tm->pending = !!(value & PCF8523_CONTROL2_AF);
@@ -339,30 +191,16 @@ static int pcf8523_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *tm)
 
 static int pcf8523_irq_enable(struct device *dev, unsigned int enabled)
 {
-	struct i2c_client *client = to_i2c_client(dev);
-	u8 value;
-	int err;
+	struct pcf8523 *pcf8523 = dev_get_drvdata(dev);
 
-	err = pcf8523_read(client, PCF8523_REG_CONTROL1, &value);
-	if (err < 0)
-		return err;
-
-	value &= PCF8523_CONTROL1_AIE;
-
-	if (enabled)
-		value |= PCF8523_CONTROL1_AIE;
-
-	err = pcf8523_write(client, PCF8523_REG_CONTROL1, value);
-	if (err < 0)
-		return err;
-
-	return 0;
+	return regmap_update_bits(pcf8523->regmap, PCF8523_REG_CONTROL1,
+				  PCF8523_CONTROL1_AIE, enabled ?
+				  PCF8523_CONTROL1_AIE : 0);
 }
 
 static int pcf8523_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *tm)
 {
-	struct i2c_client *client = to_i2c_client(dev);
-	struct i2c_msg msg;
+	struct pcf8523 *pcf8523 = dev_get_drvdata(dev);
 	u8 regs[5];
 	int err;
 
@@ -370,7 +208,7 @@ static int pcf8523_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *tm)
 	if (err)
 		return err;
 
-	err = pcf8523_write(client, PCF8523_REG_CONTROL2, 0);
+	err = regmap_write(pcf8523->regmap, PCF8523_REG_CONTROL2, 0);
 	if (err < 0)
 		return err;
 
@@ -382,16 +220,13 @@ static int pcf8523_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *tm)
 		rtc_time64_to_tm(alarm_time, &tm->time);
 	}
 
-	regs[0] = PCF8523_REG_MINUTE_ALARM;
-	regs[1] = bin2bcd(tm->time.tm_min);
-	regs[2] = bin2bcd(tm->time.tm_hour);
-	regs[3] = bin2bcd(tm->time.tm_mday);
-	regs[4] = ALARM_DIS;
-	msg.addr = client->addr;
-	msg.flags = 0;
-	msg.len = sizeof(regs);
-	msg.buf = regs;
-	err = i2c_transfer(client->adapter, &msg, 1);
+	regs[0] = bin2bcd(tm->time.tm_min);
+	regs[1] = bin2bcd(tm->time.tm_hour);
+	regs[2] = bin2bcd(tm->time.tm_mday);
+	regs[3] = ALARM_DIS;
+
+	err = regmap_bulk_write(pcf8523->regmap, PCF8523_REG_MINUTE_ALARM, regs,
+				sizeof(regs));
 	if (err < 0)
 		return err;
 
@@ -401,24 +236,101 @@ static int pcf8523_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *tm)
 	return 0;
 }
 
-#ifdef CONFIG_RTC_INTF_DEV
+static int pcf8523_param_get(struct device *dev, struct rtc_param *param)
+{
+	struct pcf8523 *pcf8523 = dev_get_drvdata(dev);
+	int ret;
+
+	switch(param->param) {
+		u32 value;
+
+	case RTC_PARAM_BACKUP_SWITCH_MODE:
+		ret = regmap_read(pcf8523->regmap, PCF8523_REG_CONTROL3, &value);
+		if (ret < 0)
+			return ret;
+
+		value = FIELD_GET(PCF8523_CONTROL3_PM, value);
+
+		switch(value) {
+		case 0x0:
+		case 0x4:
+			param->uvalue = RTC_BSM_LEVEL;
+			break;
+		case 0x1:
+		case 0x5:
+			param->uvalue = RTC_BSM_DIRECT;
+			break;
+		case PCF8523_PM_STANDBY:
+			param->uvalue = RTC_BSM_STANDBY;
+			break;
+		default:
+			param->uvalue = RTC_BSM_DISABLED;
+		}
+
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int pcf8523_param_set(struct device *dev, struct rtc_param *param)
+{
+	struct pcf8523 *pcf8523 = dev_get_drvdata(dev);
+
+	switch(param->param) {
+		u8 mode;
+	case RTC_PARAM_BACKUP_SWITCH_MODE:
+		switch (param->uvalue) {
+		case RTC_BSM_DISABLED:
+			mode = 0x2;
+			break;
+		case RTC_BSM_DIRECT:
+			mode = 0x1;
+			break;
+		case RTC_BSM_LEVEL:
+			mode = 0x0;
+			break;
+		case RTC_BSM_STANDBY:
+			mode = PCF8523_PM_STANDBY;
+			break;
+		default:
+			return -EINVAL;
+		}
+
+		return regmap_update_bits(pcf8523->regmap, PCF8523_REG_CONTROL3,
+					  PCF8523_CONTROL3_PM,
+					  FIELD_PREP(PCF8523_CONTROL3_PM, mode));
+
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static int pcf8523_rtc_ioctl(struct device *dev, unsigned int cmd,
 			     unsigned long arg)
 {
-	struct i2c_client *client = to_i2c_client(dev);
+	struct pcf8523 *pcf8523 = dev_get_drvdata(dev);
 	unsigned int flags = 0;
-	u8 value;
+	u32 value;
 	int ret;
 
 	switch (cmd) {
 	case RTC_VL_READ:
-		ret = pcf8523_voltage_low(client);
+		ret = regmap_read(pcf8523->regmap, PCF8523_REG_CONTROL3, &value);
 		if (ret < 0)
 			return ret;
-		if (ret)
+
+		if (value & PCF8523_CONTROL3_BLF)
 			flags |= RTC_VL_BACKUP_LOW;
 
-		ret = pcf8523_read(client, PCF8523_REG_SECONDS, &value);
+		ret = regmap_read(pcf8523->regmap, PCF8523_REG_SECONDS, &value);
 		if (ret < 0)
 			return ret;
 
@@ -431,18 +343,15 @@ static int pcf8523_rtc_ioctl(struct device *dev, unsigned int cmd,
 		return -ENOIOCTLCMD;
 	}
 }
-#else
-#define pcf8523_rtc_ioctl NULL
-#endif
 
 static int pcf8523_rtc_read_offset(struct device *dev, long *offset)
 {
-	struct i2c_client *client = to_i2c_client(dev);
+	struct pcf8523 *pcf8523 = dev_get_drvdata(dev);
 	int err;
-	u8 value;
+	u32 value;
 	s8 val;
 
-	err = pcf8523_read(client, PCF8523_REG_OFFSET, &value);
+	err = regmap_read(pcf8523->regmap, PCF8523_REG_OFFSET, &value);
 	if (err < 0)
 		return err;
 
@@ -455,9 +364,9 @@ static int pcf8523_rtc_read_offset(struct device *dev, long *offset)
 
 static int pcf8523_rtc_set_offset(struct device *dev, long offset)
 {
-	struct i2c_client *client = to_i2c_client(dev);
+	struct pcf8523 *pcf8523 = dev_get_drvdata(dev);
 	long reg_m0, reg_m1;
-	u8 value;
+	u32 value;
 
 	reg_m0 = clamp(DIV_ROUND_CLOSEST(offset, 4340), -64L, 63L);
 	reg_m1 = clamp(DIV_ROUND_CLOSEST(offset, 4069), -64L, 63L);
@@ -467,7 +376,7 @@ static int pcf8523_rtc_set_offset(struct device *dev, long offset)
 	else
 		value = (reg_m1 & 0x7f) | PCF8523_OFFSET_MODE;
 
-	return pcf8523_write(client, PCF8523_REG_OFFSET, value);
+	return regmap_write(pcf8523->regmap, PCF8523_REG_OFFSET, value);
 }
 
 static const struct rtc_class_ops pcf8523_rtc_ops = {
@@ -479,6 +388,14 @@ static const struct rtc_class_ops pcf8523_rtc_ops = {
 	.ioctl = pcf8523_rtc_ioctl,
 	.read_offset = pcf8523_rtc_read_offset,
 	.set_offset = pcf8523_rtc_set_offset,
+	.param_get = pcf8523_param_get,
+	.param_set = pcf8523_param_set,
+};
+
+static const struct regmap_config regmap_config = {
+        .reg_bits = 8,
+        .val_bits = 8,
+        .max_register = 0x13,
 };
 
 static int pcf8523_probe(struct i2c_client *client,
@@ -487,6 +404,7 @@ static int pcf8523_probe(struct i2c_client *client,
 	struct pcf8523 *pcf8523;
 	struct rtc_device *rtc;
 	bool wakeup_source = false;
+	u32 value;
 	int err;
 
 	if (!i2c_check_functionality(client->adapter, I2C_FUNC_I2C))
@@ -496,46 +414,60 @@ static int pcf8523_probe(struct i2c_client *client,
 	if (!pcf8523)
 		return -ENOMEM;
 
+	pcf8523->regmap = devm_regmap_init_i2c(client, &regmap_config);
+	if (IS_ERR(pcf8523->regmap))
+		return PTR_ERR(pcf8523->regmap);
+
 	i2c_set_clientdata(client, pcf8523);
-	pcf8523->client = client;
-
-	err = pcf8523_load_capacitance(client);
-	if (err < 0)
-		dev_warn(&client->dev, "failed to set xtal load capacitance: %d",
-			 err);
-
-	err = pcf8523_set_pm(client, 0);
-	if (err < 0)
-		return err;
 
 	rtc = devm_rtc_allocate_device(&client->dev);
 	if (IS_ERR(rtc))
 		return PTR_ERR(rtc);
-
 	pcf8523->rtc = rtc;
+
+	err = pcf8523_load_capacitance(pcf8523, client->dev.of_node);
+	if (err < 0)
+		dev_warn(&client->dev, "failed to set xtal load capacitance: %d",
+			 err);
+
+	err = regmap_read(pcf8523->regmap, PCF8523_REG_SECONDS, &value);
+	if (err < 0)
+		return err;
+
+	if (value & PCF8523_SECONDS_OS) {
+		err = regmap_read(pcf8523->regmap, PCF8523_REG_CONTROL3, &value);
+		if (err < 0)
+			return err;
+
+		if (FIELD_GET(PCF8523_CONTROL3_PM, value) == PCF8523_PM_STANDBY) {
+			err = regmap_write(pcf8523->regmap, PCF8523_REG_CONTROL3,
+					   value & ~PCF8523_CONTROL3_PM);
+			if (err < 0)
+				return err;
+		}
+	}
+
 	rtc->ops = &pcf8523_rtc_ops;
 	rtc->range_min = RTC_TIMESTAMP_BEGIN_2000;
 	rtc->range_max = RTC_TIMESTAMP_END_2099;
 	rtc->uie_unsupported = 1;
 
 	if (client->irq > 0) {
-		err = pcf8523_write(client, PCF8523_TMR_CLKOUT_CTRL, 0x38);
+		err = regmap_write(pcf8523->regmap, PCF8523_TMR_CLKOUT_CTRL, 0x38);
 		if (err < 0)
 			return err;
 
 		err = devm_request_threaded_irq(&client->dev, client->irq,
 						NULL, pcf8523_irq,
 						IRQF_SHARED | IRQF_ONESHOT | IRQF_TRIGGER_LOW,
-						dev_name(&rtc->dev), client);
+						dev_name(&rtc->dev), pcf8523);
 		if (err)
 			return err;
 
 		dev_pm_set_wake_irq(&client->dev, client->irq);
 	}
 
-#ifdef CONFIG_OF
 	wakeup_source = of_property_read_bool(client->dev.of_node, "wakeup-source");
-#endif
 	if (client->irq > 0 || wakeup_source)
 		device_init_wakeup(&client->dev, true);
 
@@ -548,19 +480,17 @@ static const struct i2c_device_id pcf8523_id[] = {
 };
 MODULE_DEVICE_TABLE(i2c, pcf8523_id);
 
-#ifdef CONFIG_OF
 static const struct of_device_id pcf8523_of_match[] = {
 	{ .compatible = "nxp,pcf8523" },
 	{ .compatible = "microcrystal,rv8523" },
 	{ }
 };
 MODULE_DEVICE_TABLE(of, pcf8523_of_match);
-#endif
 
 static struct i2c_driver pcf8523_driver = {
 	.driver = {
 		.name = "rtc-pcf8523",
-		.of_match_table = of_match_ptr(pcf8523_of_match),
+		.of_match_table = pcf8523_of_match,
 	},
 	.probe = pcf8523_probe,
 	.id_table = pcf8523_id,
diff --git a/drivers/rtc/rtc-rv3028.c b/drivers/rtc/rtc-rv3028.c
index 12c8073..cdc623b 100644
--- a/drivers/rtc/rtc-rv3028.c
+++ b/drivers/rtc/rtc-rv3028.c
@@ -10,6 +10,7 @@
 
 #include <linux/clk-provider.h>
 #include <linux/bcd.h>
+#include <linux/bitfield.h>
 #include <linux/bitops.h>
 #include <linux/i2c.h>
 #include <linux/interrupt.h>
@@ -80,6 +81,10 @@
 
 #define RV3028_BACKUP_TCE		BIT(5)
 #define RV3028_BACKUP_TCR_MASK		GENMASK(1,0)
+#define RV3028_BACKUP_BSM		GENMASK(3,2)
+
+#define RV3028_BACKUP_BSM_DSM		0x1
+#define RV3028_BACKUP_BSM_LSM		0x3
 
 #define OFFSET_STEP_PPT			953674
 
@@ -512,6 +517,71 @@ static int rv3028_set_offset(struct device *dev, long offset)
 
 }
 
+static int rv3028_param_get(struct device *dev, struct rtc_param *param)
+{
+	struct rv3028_data *rv3028 = dev_get_drvdata(dev);
+	int ret;
+
+	switch(param->param) {
+		u32 value;
+
+	case RTC_PARAM_BACKUP_SWITCH_MODE:
+		ret = regmap_read(rv3028->regmap, RV3028_BACKUP, &value);
+		if (ret < 0)
+			return ret;
+
+		value = FIELD_GET(RV3028_BACKUP_BSM, value);
+
+		switch(value) {
+		case RV3028_BACKUP_BSM_DSM:
+			param->uvalue = RTC_BSM_DIRECT;
+			break;
+		case RV3028_BACKUP_BSM_LSM:
+			param->uvalue = RTC_BSM_LEVEL;
+			break;
+		default:
+			param->uvalue = RTC_BSM_DISABLED;
+		}
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int rv3028_param_set(struct device *dev, struct rtc_param *param)
+{
+	struct rv3028_data *rv3028 = dev_get_drvdata(dev);
+
+	switch(param->param) {
+		u8 mode;
+	case RTC_PARAM_BACKUP_SWITCH_MODE:
+		switch (param->uvalue) {
+		case RTC_BSM_DISABLED:
+			mode = 0;
+			break;
+		case RTC_BSM_DIRECT:
+			mode = RV3028_BACKUP_BSM_DSM;
+			break;
+		case RTC_BSM_LEVEL:
+			mode = RV3028_BACKUP_BSM_LSM;
+			break;
+		default:
+			return -EINVAL;
+		}
+
+		return rv3028_update_cfg(rv3028, RV3028_BACKUP, RV3028_BACKUP_BSM,
+					 FIELD_PREP(RV3028_BACKUP_BSM, mode));
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static int rv3028_ioctl(struct device *dev, unsigned int cmd, unsigned long arg)
 {
 	struct rv3028_data *rv3028 = dev_get_drvdata(dev);
@@ -776,6 +846,8 @@ static const struct rtc_class_ops rv3028_rtc_ops = {
 	.read_offset = rv3028_read_offset,
 	.set_offset = rv3028_set_offset,
 	.ioctl = rv3028_ioctl,
+	.param_get = rv3028_param_get,
+	.param_set = rv3028_param_set,
 };
 
 static const struct regmap_config regmap_config = {
@@ -878,6 +950,8 @@ static int rv3028_probe(struct i2c_client *client)
 	if (ret)
 		return ret;
 
+	set_bit(RTC_FEATURE_BACKUP_SWITCH_MODE, rv3028->rtc->features);
+
 	rv3028->rtc->range_min = RTC_TIMESTAMP_BEGIN_2000;
 	rv3028->rtc->range_max = RTC_TIMESTAMP_END_2099;
 	rv3028->rtc->ops = &rv3028_rtc_ops;
diff --git a/drivers/rtc/rtc-rv3032.c b/drivers/rtc/rtc-rv3032.c
index d63102d..c3bee30 100644
--- a/drivers/rtc/rtc-rv3032.c
+++ b/drivers/rtc/rtc-rv3032.c
@@ -106,6 +106,7 @@
 struct rv3032_data {
 	struct regmap *regmap;
 	struct rtc_device *rtc;
+	bool trickle_charger_set;
 #ifdef CONFIG_COMMON_CLK
 	struct clk_hw clkout_hw;
 #endif
@@ -310,14 +311,6 @@ static int rv3032_set_alarm(struct device *dev, struct rtc_wkalrm *alrm)
 	u8 ctrl = 0;
 	int ret;
 
-	/* The alarm has no seconds, round up to nearest minute */
-	if (alrm->time.tm_sec) {
-		time64_t alarm_time = rtc_tm_to_time64(&alrm->time);
-
-		alarm_time += 60 - alrm->time.tm_sec;
-		rtc_time64_to_tm(alarm_time, &alrm->time);
-	}
-
 	ret = regmap_update_bits(rv3032->regmap, RV3032_CTRL2,
 				 RV3032_CTRL2_AIE | RV3032_CTRL2_UIE, 0);
 	if (ret)
@@ -402,6 +395,75 @@ static int rv3032_set_offset(struct device *dev, long offset)
 				 FIELD_PREP(RV3032_OFFSET_MSK, offset));
 }
 
+static int rv3032_param_get(struct device *dev, struct rtc_param *param)
+{
+	struct rv3032_data *rv3032 = dev_get_drvdata(dev);
+	int ret;
+
+	switch(param->param) {
+		u32 value;
+
+	case RTC_PARAM_BACKUP_SWITCH_MODE:
+		ret = regmap_read(rv3032->regmap, RV3032_PMU, &value);
+		if (ret < 0)
+			return ret;
+
+		value = FIELD_GET(RV3032_PMU_BSM, value);
+
+		switch(value) {
+		case RV3032_PMU_BSM_DSM:
+			param->uvalue = RTC_BSM_DIRECT;
+			break;
+		case RV3032_PMU_BSM_LSM:
+			param->uvalue = RTC_BSM_LEVEL;
+			break;
+		default:
+			param->uvalue = RTC_BSM_DISABLED;
+		}
+
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int rv3032_param_set(struct device *dev, struct rtc_param *param)
+{
+	struct rv3032_data *rv3032 = dev_get_drvdata(dev);
+
+	switch(param->param) {
+		u8 mode;
+	case RTC_PARAM_BACKUP_SWITCH_MODE:
+		if (rv3032->trickle_charger_set)
+			return -EINVAL;
+
+		switch (param->uvalue) {
+		case RTC_BSM_DISABLED:
+			mode = 0;
+			break;
+		case RTC_BSM_DIRECT:
+			mode = RV3032_PMU_BSM_DSM;
+			break;
+		case RTC_BSM_LEVEL:
+			mode = RV3032_PMU_BSM_LSM;
+			break;
+		default:
+			return -EINVAL;
+		}
+
+		return rv3032_update_cfg(rv3032, RV3032_PMU, RV3032_PMU_BSM,
+					 FIELD_PREP(RV3032_PMU_BSM, mode));
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static int rv3032_ioctl(struct device *dev, unsigned int cmd, unsigned long arg)
 {
 	struct rv3032_data *rv3032 = dev_get_drvdata(dev);
@@ -541,6 +603,8 @@ static int rv3032_trickle_charger_setup(struct device *dev, struct rv3032_data *
 		return 0;
 	}
 
+	rv3032->trickle_charger_set = true;
+
 	return rv3032_update_cfg(rv3032, RV3032_PMU,
 				 RV3032_PMU_TCR | RV3032_PMU_TCM | RV3032_PMU_BSM,
 				 val | FIELD_PREP(RV3032_PMU_TCR, i));
@@ -617,11 +681,11 @@ static int rv3032_clkout_set_rate(struct clk_hw *hw, unsigned long rate,
 
 	ret = rv3032_enter_eerd(rv3032, &eerd);
 	if (ret)
-		goto exit_eerd;
+		return ret;
 
 	ret = regmap_write(rv3032->regmap, RV3032_CLKOUT1, hfd & 0xff);
 	if (ret)
-		return ret;
+		goto exit_eerd;
 
 	ret = regmap_write(rv3032->regmap, RV3032_CLKOUT2, RV3032_CLKOUT2_OS |
 			    FIELD_PREP(RV3032_CLKOUT2_HFD_MSK, hfd >> 8));
@@ -813,6 +877,8 @@ static const struct rtc_class_ops rv3032_rtc_ops = {
 	.read_alarm = rv3032_get_alarm,
 	.set_alarm = rv3032_set_alarm,
 	.alarm_irq_enable = rv3032_alarm_irq_enable,
+	.param_get = rv3032_param_get,
+	.param_set = rv3032_param_set,
 };
 
 static const struct regmap_config regmap_config = {
@@ -883,6 +949,9 @@ static int rv3032_probe(struct i2c_client *client)
 
 	rv3032_trickle_charger_setup(&client->dev, rv3032);
 
+	set_bit(RTC_FEATURE_BACKUP_SWITCH_MODE, rv3032->rtc->features);
+	set_bit(RTC_FEATURE_ALARM_RES_MINUTE, rv3032->rtc->features);
+
 	rv3032->rtc->range_min = RTC_TIMESTAMP_BEGIN_2000;
 	rv3032->rtc->range_max = RTC_TIMESTAMP_END_2099;
 	rv3032->rtc->ops = &rv3032_rtc_ops;
diff --git a/drivers/rtc/rtc-rv8803.c b/drivers/rtc/rtc-rv8803.c
index 72adef5..0d5ed38 100644
--- a/drivers/rtc/rtc-rv8803.c
+++ b/drivers/rtc/rtc-rv8803.c
@@ -340,8 +340,8 @@ static int rv8803_set_alarm(struct device *dev, struct rtc_wkalrm *alrm)
 		}
 	}
 
-	ctrl[1] &= ~RV8803_FLAG_AF;
-	err = rv8803_write_reg(rv8803->client, RV8803_FLAG, ctrl[1]);
+	ctrl[0] &= ~RV8803_FLAG_AF;
+	err = rv8803_write_reg(rv8803->client, RV8803_FLAG, ctrl[0]);
 	mutex_unlock(&rv8803->flags_lock);
 	if (err)
 		return err;
diff --git a/drivers/rtc/rtc-rx6110.c b/drivers/rtc/rtc-rx6110.c
index f4d4250..758fd6e 100644
--- a/drivers/rtc/rtc-rx6110.c
+++ b/drivers/rtc/rtc-rx6110.c
@@ -422,7 +422,7 @@ static struct regmap_config regmap_i2c_config = {
 static int rx6110_i2c_probe(struct i2c_client *client,
 			    const struct i2c_device_id *id)
 {
-	struct i2c_adapter *adapter = to_i2c_adapter(client->dev.parent);
+	struct i2c_adapter *adapter = client->adapter;
 	struct rx6110_data *rx6110;
 
 	if (!i2c_check_functionality(adapter, I2C_FUNC_SMBUS_BYTE_DATA
diff --git a/drivers/rtc/rtc-rx8025.c b/drivers/rtc/rtc-rx8025.c
index d38aaf0..5bfdd34 100644
--- a/drivers/rtc/rtc-rx8025.c
+++ b/drivers/rtc/rtc-rx8025.c
@@ -248,9 +248,6 @@ static int rx8025_set_time(struct device *dev, struct rtc_time *dt)
 	u8 date[7];
 	int ret;
 
-	if ((dt->tm_year < 100) || (dt->tm_year > 199))
-		return -EINVAL;
-
 	/*
 	 * Here the read-only bits are written as "0".  I'm not sure if that
 	 * is sound.
@@ -318,9 +315,6 @@ static int rx8025_read_alarm(struct device *dev, struct rtc_wkalrm *t)
 	u8 ald[2];
 	int ctrl2, err;
 
-	if (client->irq <= 0)
-		return -EINVAL;
-
 	err = rx8025_read_regs(client, RX8025_REG_ALDMIN, 2, ald);
 	if (err)
 		return err;
@@ -355,20 +349,6 @@ static int rx8025_set_alarm(struct device *dev, struct rtc_wkalrm *t)
 	u8 ald[2];
 	int err;
 
-	if (client->irq <= 0)
-		return -EINVAL;
-
-	/*
-	 * Hardware alarm precision is 1 minute!
-	 * round up to nearest minute
-	 */
-	if (t->time.tm_sec) {
-		time64_t alarm_time = rtc_tm_to_time64(&t->time);
-
-		alarm_time += 60 - t->time.tm_sec;
-		rtc_time64_to_tm(alarm_time, &t->time);
-	}
-
 	ald[0] = bin2bcd(t->time.tm_min);
 	if (rx8025->ctrl1 & RX8025_BIT_CTRL1_1224)
 		ald[1] = bin2bcd(t->time.tm_hour);
@@ -423,17 +403,7 @@ static int rx8025_alarm_irq_enable(struct device *dev, unsigned int enabled)
 	return 0;
 }
 
-static const struct rtc_class_ops rx8025_rtc_ops = {
-	.read_time = rx8025_get_time,
-	.set_time = rx8025_set_time,
-	.read_alarm = rx8025_read_alarm,
-	.set_alarm = rx8025_set_alarm,
-	.alarm_irq_enable = rx8025_alarm_irq_enable,
-};
-
 /*
- * Clock precision adjustment support
- *
  * According to the RX8025 SA/NB application manual the frequency and
  * temperature characteristics can be approximated using the following
  * equation:
@@ -444,11 +414,8 @@ static const struct rtc_class_ops rx8025_rtc_ops = {
  *   a : Coefficient = (-35 +-5) * 10**-9
  *   ut: Ultimate temperature in degree = +25 +-5 degree
  *   t : Any temperature in degree
- *
- * Note that the clock adjustment in ppb must be entered (which is
- * the negative value of the deviation).
  */
-static int rx8025_get_clock_adjust(struct device *dev, int *adj)
+static int rx8025_read_offset(struct device *dev, long *offset)
 {
 	struct i2c_client *client = to_i2c_client(dev);
 	int digoff;
@@ -457,63 +424,75 @@ static int rx8025_get_clock_adjust(struct device *dev, int *adj)
 	if (digoff < 0)
 		return digoff;
 
-	*adj = digoff >= 64 ? digoff - 128 : digoff;
-	if (*adj > 0)
-		(*adj)--;
-	*adj *= -RX8025_ADJ_RESOLUTION;
+	*offset = digoff >= 64 ? digoff - 128 : digoff;
+	if (*offset > 0)
+		(*offset)--;
+	*offset *= RX8025_ADJ_RESOLUTION;
 
 	return 0;
 }
 
-static int rx8025_set_clock_adjust(struct device *dev, int adj)
+static int rx8025_set_offset(struct device *dev, long offset)
 {
 	struct i2c_client *client = to_i2c_client(dev);
 	u8 digoff;
 	int err;
 
-	adj /= -RX8025_ADJ_RESOLUTION;
-	if (adj > RX8025_ADJ_DATA_MAX)
-		adj = RX8025_ADJ_DATA_MAX;
-	else if (adj < RX8025_ADJ_DATA_MIN)
-		adj = RX8025_ADJ_DATA_MIN;
-	else if (adj > 0)
-		adj++;
-	else if (adj < 0)
-		adj += 128;
-	digoff = adj;
+	offset /= RX8025_ADJ_RESOLUTION;
+	if (offset > RX8025_ADJ_DATA_MAX)
+		offset = RX8025_ADJ_DATA_MAX;
+	else if (offset < RX8025_ADJ_DATA_MIN)
+		offset = RX8025_ADJ_DATA_MIN;
+	else if (offset > 0)
+		offset++;
+	else if (offset < 0)
+		offset += 128;
+	digoff = offset;
 
 	err = rx8025_write_reg(client, RX8025_REG_DIGOFF, digoff);
 	if (err)
 		return err;
 
-	dev_dbg(dev, "%s: write 0x%02x\n", __func__, digoff);
-
 	return 0;
 }
 
+static const struct rtc_class_ops rx8025_rtc_ops = {
+	.read_time = rx8025_get_time,
+	.set_time = rx8025_set_time,
+	.read_alarm = rx8025_read_alarm,
+	.set_alarm = rx8025_set_alarm,
+	.alarm_irq_enable = rx8025_alarm_irq_enable,
+	.read_offset = rx8025_read_offset,
+	.set_offset = rx8025_set_offset,
+};
+
 static ssize_t rx8025_sysfs_show_clock_adjust(struct device *dev,
 					      struct device_attribute *attr,
 					      char *buf)
 {
-	int err, adj;
+	long adj;
+	int err;
 
-	err = rx8025_get_clock_adjust(dev, &adj);
+	dev_warn_once(dev, "clock_adjust_ppb is deprecated, use offset\n");
+	err = rx8025_read_offset(dev, &adj);
 	if (err)
 		return err;
 
-	return sprintf(buf, "%d\n", adj);
+	return sprintf(buf, "%ld\n", -adj);
 }
 
 static ssize_t rx8025_sysfs_store_clock_adjust(struct device *dev,
 					       struct device_attribute *attr,
 					       const char *buf, size_t count)
 {
-	int adj, err;
+	long adj;
+	int err;
 
-	if (sscanf(buf, "%i", &adj) != 1)
+	dev_warn_once(dev, "clock_adjust_ppb is deprecated, use offset\n");
+	if (kstrtol(buf, 10, &adj) != 0)
 		return -EINVAL;
 
-	err = rx8025_set_clock_adjust(dev, adj);
+	err = rx8025_set_offset(dev, -adj);
 
 	return err ? err : count;
 }
@@ -522,15 +501,14 @@ static DEVICE_ATTR(clock_adjust_ppb, S_IRUGO | S_IWUSR,
 		   rx8025_sysfs_show_clock_adjust,
 		   rx8025_sysfs_store_clock_adjust);
 
-static int rx8025_sysfs_register(struct device *dev)
-{
-	return device_create_file(dev, &dev_attr_clock_adjust_ppb);
-}
+static struct attribute *rx8025_attrs[] = {
+	&dev_attr_clock_adjust_ppb.attr,
+	NULL
+};
 
-static void rx8025_sysfs_unregister(struct device *dev)
-{
-	device_remove_file(dev, &dev_attr_clock_adjust_ppb);
-}
+static const struct attribute_group rx8025_attr_group = {
+	.attrs	= rx8025_attrs,
+};
 
 static int rx8025_probe(struct i2c_client *client,
 			const struct i2c_device_id *id)
@@ -559,12 +537,13 @@ static int rx8025_probe(struct i2c_client *client,
 	if (err)
 		return err;
 
-	rx8025->rtc = devm_rtc_device_register(&client->dev, client->name,
-					  &rx8025_rtc_ops, THIS_MODULE);
-	if (IS_ERR(rx8025->rtc)) {
-		dev_err(&client->dev, "unable to register the class device\n");
+	rx8025->rtc = devm_rtc_allocate_device(&client->dev);
+	if (IS_ERR(rx8025->rtc))
 		return PTR_ERR(rx8025->rtc);
-	}
+
+	rx8025->rtc->ops = &rx8025_rtc_ops;
+	rx8025->rtc->range_min = RTC_TIMESTAMP_BEGIN_1900;
+	rx8025->rtc->range_max = RTC_TIMESTAMP_END_2099;
 
 	if (client->irq > 0) {
 		dev_info(&client->dev, "IRQ %d supplied\n", client->irq);
@@ -572,25 +551,20 @@ static int rx8025_probe(struct i2c_client *client,
 						rx8025_handle_irq,
 						IRQF_ONESHOT,
 						"rx8025", client);
-		if (err) {
-			dev_err(&client->dev, "unable to request IRQ, alarms disabled\n");
-			client->irq = 0;
-		}
+		if (err)
+			clear_bit(RTC_FEATURE_ALARM, rx8025->rtc->features);
 	}
 
 	rx8025->rtc->max_user_freq = 1;
 
-	/* the rx8025 alarm only supports a minute accuracy */
-	rx8025->rtc->uie_unsupported = 1;
+	set_bit(RTC_FEATURE_ALARM_RES_MINUTE, rx8025->rtc->features);
+	clear_bit(RTC_FEATURE_UPDATE_INTERRUPT, rx8025->rtc->features);
 
-	err = rx8025_sysfs_register(&client->dev);
-	return err;
-}
+	err = rtc_add_group(rx8025->rtc, &rx8025_attr_group);
+	if (err)
+		return err;
 
-static int rx8025_remove(struct i2c_client *client)
-{
-	rx8025_sysfs_unregister(&client->dev);
-	return 0;
+	return devm_rtc_register_device(rx8025->rtc);
 }
 
 static struct i2c_driver rx8025_driver = {
@@ -598,7 +572,6 @@ static struct i2c_driver rx8025_driver = {
 		.name = "rtc-rx8025",
 	},
 	.probe		= rx8025_probe,
-	.remove		= rx8025_remove,
 	.id_table	= rx8025_id,
 };
 
diff --git a/drivers/rtc/rtc-s35390a.c b/drivers/rtc/rtc-s35390a.c
index b5bdeda..26278c7 100644
--- a/drivers/rtc/rtc-s35390a.c
+++ b/drivers/rtc/rtc-s35390a.c
@@ -285,9 +285,6 @@ static int s35390a_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alm)
 		alm->time.tm_min, alm->time.tm_hour, alm->time.tm_mday,
 		alm->time.tm_mon, alm->time.tm_year, alm->time.tm_wday);
 
-	if (alm->time.tm_sec != 0)
-		dev_warn(&client->dev, "Alarms are only supported on a per minute basis!\n");
-
 	/* disable interrupt (which deasserts the irq line) */
 	err = s35390a_set_reg(s35390a, S35390A_CMD_STATUS2, &sts, sizeof(sts));
 	if (err < 0)
@@ -491,8 +488,8 @@ static int s35390a_probe(struct i2c_client *client,
 	s35390a->rtc->range_min = RTC_TIMESTAMP_BEGIN_2000;
 	s35390a->rtc->range_max = RTC_TIMESTAMP_END_2099;
 
-	/* supports per-minute alarms only, therefore set uie_unsupported */
-	s35390a->rtc->uie_unsupported = 1;
+	set_bit(RTC_FEATURE_ALARM_RES_MINUTE, s35390a->rtc->features);
+	clear_bit(RTC_FEATURE_UPDATE_INTERRUPT, s35390a->rtc->features );
 
 	if (status1 & S35390A_FLAG_INT2)
 		rtc_update_irq(s35390a->rtc, 1, RTC_AF);
diff --git a/drivers/rtc/rtc-s3c.c b/drivers/rtc/rtc-s3c.c
index e57d3ca..db529733 100644
--- a/drivers/rtc/rtc-s3c.c
+++ b/drivers/rtc/rtc-s3c.c
@@ -127,10 +127,9 @@ static int s3c_rtc_setaie(struct device *dev, unsigned int enabled)
 	return ret;
 }
 
-/* Time read/write */
-static int s3c_rtc_gettime(struct device *dev, struct rtc_time *rtc_tm)
+/* Read time from RTC and convert it from BCD */
+static int s3c_rtc_read_time(struct s3c_rtc *info, struct rtc_time *tm)
 {
-	struct s3c_rtc *info = dev_get_drvdata(dev);
 	unsigned int have_retried = 0;
 	int ret;
 
@@ -139,54 +138,40 @@ static int s3c_rtc_gettime(struct device *dev, struct rtc_time *rtc_tm)
 		return ret;
 
 retry_get_time:
-	rtc_tm->tm_min  = readb(info->base + S3C2410_RTCMIN);
-	rtc_tm->tm_hour = readb(info->base + S3C2410_RTCHOUR);
-	rtc_tm->tm_mday = readb(info->base + S3C2410_RTCDATE);
-	rtc_tm->tm_mon  = readb(info->base + S3C2410_RTCMON);
-	rtc_tm->tm_year = readb(info->base + S3C2410_RTCYEAR);
-	rtc_tm->tm_sec  = readb(info->base + S3C2410_RTCSEC);
+	tm->tm_min  = readb(info->base + S3C2410_RTCMIN);
+	tm->tm_hour = readb(info->base + S3C2410_RTCHOUR);
+	tm->tm_mday = readb(info->base + S3C2410_RTCDATE);
+	tm->tm_mon  = readb(info->base + S3C2410_RTCMON);
+	tm->tm_year = readb(info->base + S3C2410_RTCYEAR);
+	tm->tm_sec  = readb(info->base + S3C2410_RTCSEC);
 
-	/* the only way to work out whether the system was mid-update
+	/*
+	 * The only way to work out whether the system was mid-update
 	 * when we read it is to check the second counter, and if it
 	 * is zero, then we re-try the entire read
 	 */
-
-	if (rtc_tm->tm_sec == 0 && !have_retried) {
+	if (tm->tm_sec == 0 && !have_retried) {
 		have_retried = 1;
 		goto retry_get_time;
 	}
 
-	rtc_tm->tm_sec = bcd2bin(rtc_tm->tm_sec);
-	rtc_tm->tm_min = bcd2bin(rtc_tm->tm_min);
-	rtc_tm->tm_hour = bcd2bin(rtc_tm->tm_hour);
-	rtc_tm->tm_mday = bcd2bin(rtc_tm->tm_mday);
-	rtc_tm->tm_mon = bcd2bin(rtc_tm->tm_mon);
-	rtc_tm->tm_year = bcd2bin(rtc_tm->tm_year);
-
 	s3c_rtc_disable_clk(info);
 
-	rtc_tm->tm_year += 100;
-	rtc_tm->tm_mon -= 1;
+	tm->tm_sec  = bcd2bin(tm->tm_sec);
+	tm->tm_min  = bcd2bin(tm->tm_min);
+	tm->tm_hour = bcd2bin(tm->tm_hour);
+	tm->tm_mday = bcd2bin(tm->tm_mday);
+	tm->tm_mon  = bcd2bin(tm->tm_mon);
+	tm->tm_year = bcd2bin(tm->tm_year);
 
-	dev_dbg(dev, "read time %ptR\n", rtc_tm);
 	return 0;
 }
 
-static int s3c_rtc_settime(struct device *dev, struct rtc_time *tm)
+/* Convert time to BCD and write it to RTC */
+static int s3c_rtc_write_time(struct s3c_rtc *info, const struct rtc_time *tm)
 {
-	struct s3c_rtc *info = dev_get_drvdata(dev);
-	int year = tm->tm_year - 100;
 	int ret;
 
-	dev_dbg(dev, "set time %ptR\n", tm);
-
-	/* we get around y2k by simply not supporting it */
-
-	if (year < 0 || year >= 100) {
-		dev_err(dev, "rtc only supports 100 years\n");
-		return -EINVAL;
-	}
-
 	ret = s3c_rtc_enable_clk(info);
 	if (ret)
 		return ret;
@@ -195,14 +180,48 @@ static int s3c_rtc_settime(struct device *dev, struct rtc_time *tm)
 	writeb(bin2bcd(tm->tm_min),  info->base + S3C2410_RTCMIN);
 	writeb(bin2bcd(tm->tm_hour), info->base + S3C2410_RTCHOUR);
 	writeb(bin2bcd(tm->tm_mday), info->base + S3C2410_RTCDATE);
-	writeb(bin2bcd(tm->tm_mon + 1), info->base + S3C2410_RTCMON);
-	writeb(bin2bcd(year), info->base + S3C2410_RTCYEAR);
+	writeb(bin2bcd(tm->tm_mon),  info->base + S3C2410_RTCMON);
+	writeb(bin2bcd(tm->tm_year), info->base + S3C2410_RTCYEAR);
 
 	s3c_rtc_disable_clk(info);
 
 	return 0;
 }
 
+static int s3c_rtc_gettime(struct device *dev, struct rtc_time *tm)
+{
+	struct s3c_rtc *info = dev_get_drvdata(dev);
+	int ret;
+
+	ret = s3c_rtc_read_time(info, tm);
+	if (ret)
+		return ret;
+
+	/* Convert internal representation to actual date/time */
+	tm->tm_year += 100;
+	tm->tm_mon -= 1;
+
+	dev_dbg(dev, "read time %ptR\n", tm);
+	return 0;
+}
+
+static int s3c_rtc_settime(struct device *dev, struct rtc_time *tm)
+{
+	struct s3c_rtc *info = dev_get_drvdata(dev);
+	struct rtc_time rtc_tm = *tm;
+
+	dev_dbg(dev, "set time %ptR\n", tm);
+
+	/*
+	 * Convert actual date/time to internal representation.
+	 * We get around Y2K by simply not supporting it.
+	 */
+	rtc_tm.tm_year -= 100;
+	rtc_tm.tm_mon += 1;
+
+	return s3c_rtc_write_time(info, &rtc_tm);
+}
+
 static int s3c_rtc_getalarm(struct device *dev, struct rtc_wkalrm *alrm)
 {
 	struct s3c_rtc *info = dev_get_drvdata(dev);
@@ -447,15 +466,20 @@ static int s3c_rtc_probe(struct platform_device *pdev)
 
 	device_init_wakeup(&pdev->dev, 1);
 
-	/* register RTC and exit */
-	info->rtc = devm_rtc_device_register(&pdev->dev, "s3c", &s3c_rtcops,
-					     THIS_MODULE);
+	info->rtc = devm_rtc_allocate_device(&pdev->dev);
 	if (IS_ERR(info->rtc)) {
-		dev_err(&pdev->dev, "cannot attach rtc\n");
 		ret = PTR_ERR(info->rtc);
 		goto err_nortc;
 	}
 
+	info->rtc->ops = &s3c_rtcops;
+	info->rtc->range_min = RTC_TIMESTAMP_BEGIN_2000;
+	info->rtc->range_max = RTC_TIMESTAMP_END_2099;
+
+	ret = devm_rtc_register_device(info->rtc);
+	if (ret)
+		goto err_nortc;
+
 	ret = devm_request_irq(&pdev->dev, info->irq_alarm, s3c_rtc_alarmirq,
 			       0, "s3c2410-rtc alarm", info);
 	if (ret) {
diff --git a/drivers/rtc/rtc-s5m.c b/drivers/rtc/rtc-s5m.c
index fb9c6b7..4243fe6 100644
--- a/drivers/rtc/rtc-s5m.c
+++ b/drivers/rtc/rtc-s5m.c
@@ -861,4 +861,3 @@ module_platform_driver(s5m_rtc_driver);
 MODULE_AUTHOR("Sangbeom Kim <sbkim73@samsung.com>");
 MODULE_DESCRIPTION("Samsung S5M/S2MPS14 RTC driver");
 MODULE_LICENSE("GPL");
-MODULE_ALIAS("platform:s5m-rtc");
diff --git a/drivers/rtc/rtc-sun6i.c b/drivers/rtc/rtc-sun6i.c
index adec1b1..711832c 100644
--- a/drivers/rtc/rtc-sun6i.c
+++ b/drivers/rtc/rtc-sun6i.c
@@ -673,8 +673,17 @@ static int sun6i_rtc_probe(struct platform_device *pdev)
 	struct sun6i_rtc_dev *chip = sun6i_rtc;
 	int ret;
 
-	if (!chip)
-		return -ENODEV;
+	if (!chip) {
+		chip = devm_kzalloc(&pdev->dev, sizeof(*chip), GFP_KERNEL);
+		if (!chip)
+			return -ENOMEM;
+
+		spin_lock_init(&chip->lock);
+
+		chip->base = devm_platform_ioremap_resource(pdev, 0);
+		if (IS_ERR(chip->base))
+			return PTR_ERR(chip->base);
+	}
 
 	platform_set_drvdata(pdev, chip);
 
diff --git a/drivers/rtc/rtc-tps80031.c b/drivers/rtc/rtc-tps80031.c
deleted file mode 100644
index c77b8ea..0000000
--- a/drivers/rtc/rtc-tps80031.c
+++ /dev/null
@@ -1,324 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-only
-/*
- * rtc-tps80031.c -- TI TPS80031/TPS80032 RTC driver
- *
- * RTC driver for TI TPS80031/TPS80032 Fully Integrated
- * Power Management with Power Path and Battery Charger
- *
- * Copyright (c) 2012, NVIDIA Corporation.
- *
- * Author: Laxman Dewangan <ldewangan@nvidia.com>
- */
-
-#include <linux/bcd.h>
-#include <linux/device.h>
-#include <linux/err.h>
-#include <linux/init.h>
-#include <linux/kernel.h>
-#include <linux/module.h>
-#include <linux/mfd/tps80031.h>
-#include <linux/platform_device.h>
-#include <linux/pm.h>
-#include <linux/rtc.h>
-#include <linux/slab.h>
-
-#define ENABLE_ALARM_INT			0x08
-#define ALARM_INT_STATUS			0x40
-
-/**
- * Setting bit to 1 in STOP_RTC will run the RTC and
- * setting this bit to 0 will freeze RTC.
- */
-#define STOP_RTC				0x1
-
-/* Power on reset Values of RTC registers */
-#define TPS80031_RTC_POR_YEAR			0
-#define TPS80031_RTC_POR_MONTH			1
-#define TPS80031_RTC_POR_DAY			1
-
-/* Numbers of registers for time and alarms */
-#define TPS80031_RTC_TIME_NUM_REGS		7
-#define TPS80031_RTC_ALARM_NUM_REGS		6
-
-/**
- * PMU RTC have only 2 nibbles to store year information, so using an
- * offset of 100 to set the base year as 2000 for our driver.
- */
-#define RTC_YEAR_OFFSET 100
-
-struct tps80031_rtc {
-	struct rtc_device	*rtc;
-	int			irq;
-};
-
-static int tps80031_rtc_read_time(struct device *dev, struct rtc_time *tm)
-{
-	u8 buff[TPS80031_RTC_TIME_NUM_REGS];
-	int ret;
-
-	ret = tps80031_reads(dev->parent, TPS80031_SLAVE_ID1,
-			TPS80031_SECONDS_REG, TPS80031_RTC_TIME_NUM_REGS, buff);
-	if (ret < 0) {
-		dev_err(dev, "reading RTC_SECONDS_REG failed, err = %d\n", ret);
-		return ret;
-	}
-
-	tm->tm_sec = bcd2bin(buff[0]);
-	tm->tm_min = bcd2bin(buff[1]);
-	tm->tm_hour = bcd2bin(buff[2]);
-	tm->tm_mday = bcd2bin(buff[3]);
-	tm->tm_mon = bcd2bin(buff[4]) - 1;
-	tm->tm_year = bcd2bin(buff[5]) + RTC_YEAR_OFFSET;
-	tm->tm_wday = bcd2bin(buff[6]);
-	return 0;
-}
-
-static int tps80031_rtc_set_time(struct device *dev, struct rtc_time *tm)
-{
-	u8 buff[7];
-	int ret;
-
-	buff[0] = bin2bcd(tm->tm_sec);
-	buff[1] = bin2bcd(tm->tm_min);
-	buff[2] = bin2bcd(tm->tm_hour);
-	buff[3] = bin2bcd(tm->tm_mday);
-	buff[4] = bin2bcd(tm->tm_mon + 1);
-	buff[5] = bin2bcd(tm->tm_year % RTC_YEAR_OFFSET);
-	buff[6] = bin2bcd(tm->tm_wday);
-
-	/* Stop RTC while updating the RTC time registers */
-	ret = tps80031_clr_bits(dev->parent, TPS80031_SLAVE_ID1,
-				TPS80031_RTC_CTRL_REG, STOP_RTC);
-	if (ret < 0) {
-		dev_err(dev->parent, "Stop RTC failed, err = %d\n", ret);
-		return ret;
-	}
-
-	ret = tps80031_writes(dev->parent, TPS80031_SLAVE_ID1,
-			TPS80031_SECONDS_REG,
-			TPS80031_RTC_TIME_NUM_REGS, buff);
-	if (ret < 0) {
-		dev_err(dev, "writing RTC_SECONDS_REG failed, err %d\n", ret);
-		return ret;
-	}
-
-	ret = tps80031_set_bits(dev->parent, TPS80031_SLAVE_ID1,
-				TPS80031_RTC_CTRL_REG, STOP_RTC);
-	if (ret < 0)
-		dev_err(dev->parent, "Start RTC failed, err = %d\n", ret);
-	return ret;
-}
-
-static int tps80031_rtc_alarm_irq_enable(struct device *dev,
-					 unsigned int enable)
-{
-	int ret;
-
-	if (enable)
-		ret = tps80031_set_bits(dev->parent, TPS80031_SLAVE_ID1,
-				TPS80031_RTC_INTERRUPTS_REG, ENABLE_ALARM_INT);
-	else
-		ret = tps80031_clr_bits(dev->parent, TPS80031_SLAVE_ID1,
-				TPS80031_RTC_INTERRUPTS_REG, ENABLE_ALARM_INT);
-	if (ret < 0) {
-		dev_err(dev, "Update on RTC_INT failed, err = %d\n", ret);
-		return ret;
-	}
-	return 0;
-}
-
-static int tps80031_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alrm)
-{
-	u8 buff[TPS80031_RTC_ALARM_NUM_REGS];
-	int ret;
-
-	buff[0] = bin2bcd(alrm->time.tm_sec);
-	buff[1] = bin2bcd(alrm->time.tm_min);
-	buff[2] = bin2bcd(alrm->time.tm_hour);
-	buff[3] = bin2bcd(alrm->time.tm_mday);
-	buff[4] = bin2bcd(alrm->time.tm_mon + 1);
-	buff[5] = bin2bcd(alrm->time.tm_year % RTC_YEAR_OFFSET);
-	ret = tps80031_writes(dev->parent, TPS80031_SLAVE_ID1,
-			TPS80031_ALARM_SECONDS_REG,
-			TPS80031_RTC_ALARM_NUM_REGS, buff);
-	if (ret < 0) {
-		dev_err(dev, "Writing RTC_ALARM failed, err %d\n", ret);
-		return ret;
-	}
-	return tps80031_rtc_alarm_irq_enable(dev, alrm->enabled);
-}
-
-static int tps80031_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm)
-{
-	u8 buff[6];
-	int ret;
-
-	ret = tps80031_reads(dev->parent, TPS80031_SLAVE_ID1,
-			TPS80031_ALARM_SECONDS_REG,
-			TPS80031_RTC_ALARM_NUM_REGS, buff);
-	if (ret < 0) {
-		dev_err(dev->parent,
-			"reading RTC_ALARM failed, err = %d\n", ret);
-		return ret;
-	}
-
-	alrm->time.tm_sec = bcd2bin(buff[0]);
-	alrm->time.tm_min = bcd2bin(buff[1]);
-	alrm->time.tm_hour = bcd2bin(buff[2]);
-	alrm->time.tm_mday = bcd2bin(buff[3]);
-	alrm->time.tm_mon = bcd2bin(buff[4]) - 1;
-	alrm->time.tm_year = bcd2bin(buff[5]) + RTC_YEAR_OFFSET;
-	return 0;
-}
-
-static int clear_alarm_int_status(struct device *dev, struct tps80031_rtc *rtc)
-{
-	int ret;
-	u8 buf;
-
-	/**
-	 * As per datasheet, A dummy read of this  RTC_STATUS_REG register
-	 * is necessary before each I2C read in order to update the status
-	 * register value.
-	 */
-	ret = tps80031_read(dev->parent, TPS80031_SLAVE_ID1,
-				TPS80031_RTC_STATUS_REG, &buf);
-	if (ret < 0) {
-		dev_err(dev, "reading RTC_STATUS failed. err = %d\n", ret);
-		return ret;
-	}
-
-	/* clear Alarm status bits.*/
-	ret = tps80031_set_bits(dev->parent, TPS80031_SLAVE_ID1,
-			TPS80031_RTC_STATUS_REG, ALARM_INT_STATUS);
-	if (ret < 0) {
-		dev_err(dev, "clear Alarm INT failed, err = %d\n", ret);
-		return ret;
-	}
-	return 0;
-}
-
-static irqreturn_t tps80031_rtc_irq(int irq, void *data)
-{
-	struct device *dev = data;
-	struct tps80031_rtc *rtc = dev_get_drvdata(dev);
-	int ret;
-
-	ret = clear_alarm_int_status(dev, rtc);
-	if (ret < 0)
-		return ret;
-
-	rtc_update_irq(rtc->rtc, 1, RTC_IRQF | RTC_AF);
-	return IRQ_HANDLED;
-}
-
-static const struct rtc_class_ops tps80031_rtc_ops = {
-	.read_time = tps80031_rtc_read_time,
-	.set_time = tps80031_rtc_set_time,
-	.set_alarm = tps80031_rtc_set_alarm,
-	.read_alarm = tps80031_rtc_read_alarm,
-	.alarm_irq_enable = tps80031_rtc_alarm_irq_enable,
-};
-
-static int tps80031_rtc_probe(struct platform_device *pdev)
-{
-	struct tps80031_rtc *rtc;
-	struct rtc_time tm;
-	int ret;
-
-	rtc = devm_kzalloc(&pdev->dev, sizeof(*rtc), GFP_KERNEL);
-	if (!rtc)
-		return -ENOMEM;
-
-	rtc->irq = platform_get_irq(pdev, 0);
-	platform_set_drvdata(pdev, rtc);
-
-	/* Start RTC */
-	ret = tps80031_set_bits(pdev->dev.parent, TPS80031_SLAVE_ID1,
-			TPS80031_RTC_CTRL_REG, STOP_RTC);
-	if (ret < 0) {
-		dev_err(&pdev->dev, "failed to start RTC. err = %d\n", ret);
-		return ret;
-	}
-
-	/* If RTC have POR values, set time 01:01:2000 */
-	tps80031_rtc_read_time(&pdev->dev, &tm);
-	if ((tm.tm_year == RTC_YEAR_OFFSET + TPS80031_RTC_POR_YEAR) &&
-		(tm.tm_mon == (TPS80031_RTC_POR_MONTH - 1)) &&
-		(tm.tm_mday == TPS80031_RTC_POR_DAY)) {
-		tm.tm_year = 2000;
-		tm.tm_mday = 1;
-		tm.tm_mon = 1;
-		ret = tps80031_rtc_set_time(&pdev->dev, &tm);
-		if (ret < 0) {
-			dev_err(&pdev->dev,
-				"RTC set time failed, err = %d\n", ret);
-			return ret;
-		}
-	}
-
-	/* Clear alarm intretupt status if it is there */
-	ret = clear_alarm_int_status(&pdev->dev, rtc);
-	if (ret < 0) {
-		dev_err(&pdev->dev, "Clear alarm int failed, err = %d\n", ret);
-		return ret;
-	}
-
-	rtc->rtc = devm_rtc_device_register(&pdev->dev, pdev->name,
-			       &tps80031_rtc_ops, THIS_MODULE);
-	if (IS_ERR(rtc->rtc)) {
-		ret = PTR_ERR(rtc->rtc);
-		dev_err(&pdev->dev, "RTC registration failed, err %d\n", ret);
-		return ret;
-	}
-
-	ret = devm_request_threaded_irq(&pdev->dev, rtc->irq, NULL,
-			tps80031_rtc_irq,
-			IRQF_ONESHOT,
-			dev_name(&pdev->dev), rtc);
-	if (ret < 0) {
-		dev_err(&pdev->dev, "request IRQ:%d failed, err = %d\n",
-			 rtc->irq, ret);
-		return ret;
-	}
-	device_set_wakeup_capable(&pdev->dev, 1);
-	return 0;
-}
-
-#ifdef CONFIG_PM_SLEEP
-static int tps80031_rtc_suspend(struct device *dev)
-{
-	struct tps80031_rtc *rtc = dev_get_drvdata(dev);
-
-	if (device_may_wakeup(dev))
-		enable_irq_wake(rtc->irq);
-	return 0;
-}
-
-static int tps80031_rtc_resume(struct device *dev)
-{
-	struct tps80031_rtc *rtc = dev_get_drvdata(dev);
-
-	if (device_may_wakeup(dev))
-		disable_irq_wake(rtc->irq);
-	return 0;
-};
-#endif
-
-static SIMPLE_DEV_PM_OPS(tps80031_pm_ops, tps80031_rtc_suspend,
-			tps80031_rtc_resume);
-
-static struct platform_driver tps80031_rtc_driver = {
-	.driver	= {
-		.name	= "tps80031-rtc",
-		.pm	= &tps80031_pm_ops,
-	},
-	.probe	= tps80031_rtc_probe,
-};
-
-module_platform_driver(tps80031_rtc_driver);
-
-MODULE_ALIAS("platform:tps80031-rtc");
-MODULE_DESCRIPTION("TI TPS80031/TPS80032 RTC driver");
-MODULE_AUTHOR("Laxman Dewangan <ldewangan@nvidia.com>");
-MODULE_LICENSE("GPL v2");
diff --git a/drivers/s390/char/tape_std.c b/drivers/s390/char/tape_std.c
index 1f5fab6..f7e75d9 100644
--- a/drivers/s390/char/tape_std.c
+++ b/drivers/s390/char/tape_std.c
@@ -53,7 +53,6 @@ int
 tape_std_assign(struct tape_device *device)
 {
 	int                  rc;
-	struct timer_list    timeout;
 	struct tape_request *request;
 
 	request = tape_alloc_request(2, 11);
@@ -70,7 +69,7 @@ tape_std_assign(struct tape_device *device)
 	 * So we set up a timeout for this call.
 	 */
 	timer_setup(&request->timer, tape_std_assign_timeout, 0);
-	mod_timer(&timeout, jiffies + 2 * HZ);
+	mod_timer(&request->timer, jiffies + msecs_to_jiffies(2000));
 
 	rc = tape_do_io_interruptible(device, request);
 
diff --git a/drivers/s390/cio/css.c b/drivers/s390/cio/css.c
index 2bc55cc..ce9e751 100644
--- a/drivers/s390/cio/css.c
+++ b/drivers/s390/cio/css.c
@@ -437,8 +437,8 @@ static ssize_t dev_busid_show(struct device *dev,
 	struct subchannel *sch = to_subchannel(dev);
 	struct pmcw *pmcw = &sch->schib.pmcw;
 
-	if ((pmcw->st == SUBCHANNEL_TYPE_IO ||
-	     pmcw->st == SUBCHANNEL_TYPE_MSG) && pmcw->dnv)
+	if ((pmcw->st == SUBCHANNEL_TYPE_IO && pmcw->dnv) ||
+	    (pmcw->st == SUBCHANNEL_TYPE_MSG && pmcw->w))
 		return sysfs_emit(buf, "0.%x.%04x\n", sch->schid.ssid,
 				  pmcw->dev);
 	else
diff --git a/drivers/scsi/elx/libefc_sli/sli4.c b/drivers/scsi/elx/libefc_sli/sli4.c
index 6c6c04e..907d67a 100644
--- a/drivers/scsi/elx/libefc_sli/sli4.c
+++ b/drivers/scsi/elx/libefc_sli/sli4.c
@@ -4145,7 +4145,7 @@ static int
 sli_get_read_config(struct sli4 *sli4)
 {
 	struct sli4_rsp_read_config *conf = sli4->bmbx.virt;
-	u32 i, total, total_size;
+	u32 i, total;
 	u32 *base;
 
 	if (sli_cmd_read_config(sli4, sli4->bmbx.virt)) {
@@ -4203,8 +4203,7 @@ sli_get_read_config(struct sli4 *sli4)
 
 	for (i = 0; i < SLI4_RSRC_MAX; i++) {
 		total = sli4->ext[i].number * sli4->ext[i].size;
-		total_size = BITS_TO_LONGS(total) * sizeof(long);
-		sli4->ext[i].use_map = kzalloc(total_size, GFP_KERNEL);
+		sli4->ext[i].use_map = bitmap_zalloc(total, GFP_KERNEL);
 		if (!sli4->ext[i].use_map) {
 			efc_log_err(sli4, "bitmap memory allocation failed %d\n",
 				    i);
@@ -4743,7 +4742,7 @@ sli_reset(struct sli4 *sli4)
 	sli4->ext[0].base = NULL;
 
 	for (i = 0; i < SLI4_RSRC_MAX; i++) {
-		kfree(sli4->ext[i].use_map);
+		bitmap_free(sli4->ext[i].use_map);
 		sli4->ext[i].use_map = NULL;
 		sli4->ext[i].base = NULL;
 	}
@@ -4784,7 +4783,7 @@ sli_teardown(struct sli4 *sli4)
 	for (i = 0; i < SLI4_RSRC_MAX; i++) {
 		sli4->ext[i].base = NULL;
 
-		kfree(sli4->ext[i].use_map);
+		bitmap_free(sli4->ext[i].use_map);
 		sli4->ext[i].use_map = NULL;
 	}
 
diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c
index cda0135..8049b00 100644
--- a/drivers/scsi/hosts.c
+++ b/drivers/scsi/hosts.c
@@ -388,6 +388,7 @@ struct Scsi_Host *scsi_host_alloc(struct scsi_host_template *sht, int privsize)
 	shost->shost_state = SHOST_CREATED;
 	INIT_LIST_HEAD(&shost->__devices);
 	INIT_LIST_HEAD(&shost->__targets);
+	INIT_LIST_HEAD(&shost->eh_abort_list);
 	INIT_LIST_HEAD(&shost->eh_cmd_q);
 	INIT_LIST_HEAD(&shost->starved_list);
 	init_waitqueue_head(&shost->host_wait);
diff --git a/drivers/scsi/qla2xxx/qla_attr.c b/drivers/scsi/qla2xxx/qla_attr.c
index 30f9545..032efb2 100644
--- a/drivers/scsi/qla2xxx/qla_attr.c
+++ b/drivers/scsi/qla2xxx/qla_attr.c
@@ -2762,7 +2762,12 @@ qla2x00_terminate_rport_io(struct fc_rport *rport)
 			if (fcport->loop_id != FC_NO_LOOP_ID)
 				fcport->logout_on_delete = 1;
 
-			qlt_schedule_sess_for_deletion(fcport);
+			if (!EDIF_NEGOTIATION_PENDING(fcport)) {
+				ql_dbg(ql_dbg_disc, fcport->vha, 0x911e,
+				       "%s %d schedule session deletion\n", __func__,
+				       __LINE__);
+				qlt_schedule_sess_for_deletion(fcport);
+			}
 		} else {
 			qla2x00_port_logout(fcport->vha, fcport);
 		}
diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h
index 8924eeb9..9ebf4a2 100644
--- a/drivers/scsi/qla2xxx/qla_def.h
+++ b/drivers/scsi/qla2xxx/qla_def.h
@@ -639,9 +639,9 @@ struct qla_els_pt_arg {
 	u8 els_opcode;
 	u8 vp_idx;
 	__le16 nport_handle;
-	u16 control_flags;
+	u16 control_flags, ox_id;
 	__le32 rx_xchg_address;
-	port_id_t did;
+	port_id_t did, sid;
 	u32 tx_len, tx_byte_count, rx_len, rx_byte_count;
 	dma_addr_t tx_addr, rx_addr;
 
diff --git a/drivers/scsi/qla2xxx/qla_edif.c b/drivers/scsi/qla2xxx/qla_edif.c
index ad746c6..2e37b18 100644
--- a/drivers/scsi/qla2xxx/qla_edif.c
+++ b/drivers/scsi/qla2xxx/qla_edif.c
@@ -218,7 +218,7 @@ fc_port_t *fcport)
 		    "%s edif not enabled\n", __func__);
 		goto done;
 	}
-	if (vha->e_dbell.db_flags != EDB_ACTIVE) {
+	if (DBELL_INACTIVE(vha)) {
 		ql_dbg(ql_dbg_edif, vha, 0x09102,
 		    "%s doorbell not enabled\n", __func__);
 		goto done;
@@ -290,63 +290,6 @@ qla_edif_app_check(scsi_qla_host_t *vha, struct app_id appid)
 	return false;
 }
 
-static void qla_edif_reset_auth_wait(struct fc_port *fcport, int state,
-		int waitonly)
-{
-	int cnt, max_cnt = 200;
-	bool traced = false;
-
-	fcport->keep_nport_handle = 1;
-
-	if (!waitonly) {
-		qla2x00_set_fcport_disc_state(fcport, state);
-		qlt_schedule_sess_for_deletion(fcport);
-	} else {
-		qla2x00_set_fcport_disc_state(fcport, state);
-	}
-
-	ql_dbg(ql_dbg_edif, fcport->vha, 0xf086,
-		"%s: waiting for session, max_cnt=%u\n",
-		__func__, max_cnt);
-
-	cnt = 0;
-
-	if (waitonly) {
-		/* Marker wait min 10 msecs. */
-		msleep(50);
-		cnt += 50;
-	}
-	while (1) {
-		if (!traced) {
-			ql_dbg(ql_dbg_edif, fcport->vha, 0xf086,
-			    "%s: session sleep.\n",
-			    __func__);
-			traced = true;
-		}
-		msleep(20);
-		cnt++;
-		if (waitonly && (fcport->disc_state == state ||
-			fcport->disc_state == DSC_LOGIN_COMPLETE))
-			break;
-		if (fcport->disc_state == DSC_LOGIN_AUTH_PEND)
-			break;
-		if (cnt > max_cnt)
-			break;
-	}
-
-	if (!waitonly) {
-		ql_dbg(ql_dbg_edif, fcport->vha, 0xf086,
-		    "%s: waited for session - %8phC, loopid=%x portid=%06x fcport=%p state=%u, cnt=%u\n",
-		    __func__, fcport->port_name, fcport->loop_id,
-		    fcport->d_id.b24, fcport, fcport->disc_state, cnt);
-	} else {
-		ql_dbg(ql_dbg_edif, fcport->vha, 0xf086,
-		    "%s: waited ONLY for session - %8phC, loopid=%x portid=%06x fcport=%p state=%u, cnt=%u\n",
-		    __func__, fcport->port_name, fcport->loop_id,
-		    fcport->d_id.b24, fcport, fcport->disc_state, cnt);
-	}
-}
-
 static void
 qla_edif_free_sa_ctl(fc_port_t *fcport, struct edif_sa_ctl *sa_ctl,
 	int index)
@@ -529,7 +472,8 @@ qla_edif_app_start(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 	struct app_start_reply	appreply;
 	struct fc_port  *fcport, *tf;
 
-	ql_dbg(ql_dbg_edif, vha, 0x911d, "%s app start\n", __func__);
+	ql_log(ql_log_info, vha, 0x1313,
+	       "EDIF application registration with driver, FC device connections will be re-established.\n");
 
 	sg_copy_to_buffer(bsg_job->request_payload.sg_list,
 	    bsg_job->request_payload.sg_cnt, &appstart,
@@ -538,9 +482,9 @@ qla_edif_app_start(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 	ql_dbg(ql_dbg_edif, vha, 0x911d, "%s app_vid=%x app_start_flags %x\n",
 	     __func__, appstart.app_info.app_vid, appstart.app_start_flags);
 
-	if (vha->e_dbell.db_flags != EDB_ACTIVE) {
+	if (DBELL_INACTIVE(vha)) {
 		/* mark doorbell as active since an app is now present */
-		vha->e_dbell.db_flags = EDB_ACTIVE;
+		vha->e_dbell.db_flags |= EDB_ACTIVE;
 	} else {
 		ql_dbg(ql_dbg_edif, vha, 0x911e, "%s doorbell already active\n",
 		     __func__);
@@ -554,37 +498,36 @@ qla_edif_app_start(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 		qla2xxx_wake_dpc(vha);
 	} else {
 		list_for_each_entry_safe(fcport, tf, &vha->vp_fcports, list) {
+			ql_dbg(ql_dbg_edif, vha, 0x2058,
+			       "FCSP - nn %8phN pn %8phN portid=%06x.\n",
+			       fcport->node_name, fcport->port_name,
+			       fcport->d_id.b24);
 			ql_dbg(ql_dbg_edif, vha, 0xf084,
-			       "%s: sess %p %8phC lid %#04x s_id %06x logout %d\n",
-			       __func__, fcport, fcport->port_name,
-			       fcport->loop_id, fcport->d_id.b24,
-			       fcport->logout_on_delete);
-
-			ql_dbg(ql_dbg_edif, vha, 0xf084,
-			       "keep %d els_logo %d disc state %d auth state %d stop state %d\n",
-			       fcport->keep_nport_handle,
-			       fcport->send_els_logo, fcport->disc_state,
-			       fcport->edif.auth_state, fcport->edif.app_stop);
+			       "%s: se_sess %p / sess %p from port %8phC "
+			       "loop_id %#04x s_id %06x logout %d "
+			       "keep %d els_logo %d disc state %d auth state %d"
+			       "stop state %d\n",
+			       __func__, fcport->se_sess, fcport,
+			       fcport->port_name, fcport->loop_id,
+			       fcport->d_id.b24, fcport->logout_on_delete,
+			       fcport->keep_nport_handle, fcport->send_els_logo,
+			       fcport->disc_state, fcport->edif.auth_state,
+			       fcport->edif.app_stop);
 
 			if (atomic_read(&vha->loop_state) == LOOP_DOWN)
 				break;
-			if (!(fcport->flags & FCF_FCSP_DEVICE))
-				continue;
 
 			fcport->edif.app_started = 1;
-			if (fcport->edif.app_stop ||
-			    (fcport->disc_state != DSC_LOGIN_COMPLETE &&
-			     fcport->disc_state != DSC_LOGIN_PEND &&
-			     fcport->disc_state != DSC_DELETED)) {
-				/* no activity */
-				fcport->edif.app_stop = 0;
+			fcport->login_retry = vha->hw->login_retry_count;
 
-				ql_dbg(ql_dbg_edif, vha, 0x911e,
-				       "%s wwpn %8phC calling qla_edif_reset_auth_wait\n",
-				       __func__, fcport->port_name);
-				fcport->edif.app_sess_online = 1;
-				qla_edif_reset_auth_wait(fcport, DSC_LOGIN_PEND, 0);
-			}
+			/* no activity */
+			fcport->edif.app_stop = 0;
+
+			ql_dbg(ql_dbg_edif, vha, 0x911e,
+			       "%s wwpn %8phC calling qla_edif_reset_auth_wait\n",
+			       __func__, fcport->port_name);
+			fcport->edif.app_sess_online = 0;
+			qlt_schedule_sess_for_deletion(fcport);
 			qla_edif_sa_ctl_init(vha, fcport);
 		}
 	}
@@ -601,14 +544,14 @@ qla_edif_app_start(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 	appreply.edif_enode_active = vha->pur_cinfo.enode_flags;
 	appreply.edif_edb_active = vha->e_dbell.db_flags;
 
-	bsg_job->reply_len = sizeof(struct fc_bsg_reply) +
-	    sizeof(struct app_start_reply);
+	bsg_job->reply_len = sizeof(struct fc_bsg_reply);
 
 	SET_DID_STATUS(bsg_reply->result, DID_OK);
 
-	sg_copy_from_buffer(bsg_job->reply_payload.sg_list,
-	    bsg_job->reply_payload.sg_cnt, &appreply,
-	    sizeof(struct app_start_reply));
+	bsg_reply->reply_payload_rcv_len = sg_copy_from_buffer(bsg_job->reply_payload.sg_list,
+							       bsg_job->reply_payload.sg_cnt,
+							       &appreply,
+							       sizeof(struct app_start_reply));
 
 	ql_dbg(ql_dbg_edif, vha, 0x911d,
 	    "%s app start completed with 0x%x\n",
@@ -800,15 +743,15 @@ qla_edif_app_authok(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 		ql_dbg(ql_dbg_edif, vha, 0x911e,
 		    "%s AUTH complete - RESUME with prli for wwpn %8phC\n",
 		    __func__, fcport->port_name);
-		qla_edif_reset_auth_wait(fcport, DSC_LOGIN_PEND, 1);
 		qla24xx_post_prli_work(vha, fcport);
 	}
 
 errstate_exit:
 	bsg_job->reply_len = sizeof(struct fc_bsg_reply);
-	sg_copy_from_buffer(bsg_job->reply_payload.sg_list,
-	    bsg_job->reply_payload.sg_cnt, &appplogireply,
-	    sizeof(struct app_plogi_reply));
+	bsg_reply->reply_payload_rcv_len = sg_copy_from_buffer(bsg_job->reply_payload.sg_list,
+							       bsg_job->reply_payload.sg_cnt,
+							       &appplogireply,
+							       sizeof(struct app_plogi_reply));
 
 	return rval;
 }
@@ -873,7 +816,7 @@ qla_edif_app_authfail(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 
 		if (qla_ini_mode_enabled(fcport->vha)) {
 			fcport->send_els_logo = 1;
-			qla_edif_reset_auth_wait(fcport, DSC_LOGIN_PEND, 0);
+			qlt_schedule_sess_for_deletion(fcport);
 		}
 	}
 
@@ -891,7 +834,7 @@ static int
 qla_edif_app_getfcinfo(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 {
 	int32_t			rval = 0;
-	int32_t			num_cnt;
+	int32_t			pcnt = 0;
 	struct fc_bsg_reply	*bsg_reply = bsg_job->reply;
 	struct app_pinfo_req	app_req;
 	struct app_pinfo_reply	*app_reply;
@@ -903,16 +846,14 @@ qla_edif_app_getfcinfo(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 	    bsg_job->request_payload.sg_cnt, &app_req,
 	    sizeof(struct app_pinfo_req));
 
-	num_cnt = app_req.num_ports;	/* num of ports alloc'd by app */
-
 	app_reply = kzalloc((sizeof(struct app_pinfo_reply) +
-	    sizeof(struct app_pinfo) * num_cnt), GFP_KERNEL);
+	    sizeof(struct app_pinfo) * app_req.num_ports), GFP_KERNEL);
+
 	if (!app_reply) {
 		SET_DID_STATUS(bsg_reply->result, DID_ERROR);
 		rval = -1;
 	} else {
 		struct fc_port	*fcport = NULL, *tf;
-		uint32_t	pcnt = 0;
 
 		list_for_each_entry_safe(fcport, tf, &vha->vp_fcports, list) {
 			if (!(fcport->flags & FCF_FCSP_DEVICE))
@@ -981,9 +922,11 @@ qla_edif_app_getfcinfo(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 		SET_DID_STATUS(bsg_reply->result, DID_OK);
 	}
 
-	sg_copy_from_buffer(bsg_job->reply_payload.sg_list,
-	    bsg_job->reply_payload.sg_cnt, app_reply,
-	    sizeof(struct app_pinfo_reply) + sizeof(struct app_pinfo) * num_cnt);
+	bsg_job->reply_len = sizeof(struct fc_bsg_reply);
+	bsg_reply->reply_payload_rcv_len = sg_copy_from_buffer(bsg_job->reply_payload.sg_list,
+							       bsg_job->reply_payload.sg_cnt,
+							       app_reply,
+							       sizeof(struct app_pinfo_reply) + sizeof(struct app_pinfo) * pcnt);
 
 	kfree(app_reply);
 
@@ -1000,10 +943,11 @@ qla_edif_app_getstats(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 {
 	int32_t			rval = 0;
 	struct fc_bsg_reply	*bsg_reply = bsg_job->reply;
-	uint32_t ret_size, size;
+	uint32_t size;
 
 	struct app_sinfo_req	app_req;
 	struct app_stats_reply	*app_reply;
+	uint32_t pcnt = 0;
 
 	sg_copy_to_buffer(bsg_job->request_payload.sg_list,
 	    bsg_job->request_payload.sg_cnt, &app_req,
@@ -1019,18 +963,12 @@ qla_edif_app_getstats(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 	size = sizeof(struct app_stats_reply) +
 	    (sizeof(struct app_sinfo) * app_req.num_ports);
 
-	if (size > bsg_job->reply_payload.payload_len)
-		ret_size = bsg_job->reply_payload.payload_len;
-	else
-		ret_size = size;
-
 	app_reply = kzalloc(size, GFP_KERNEL);
 	if (!app_reply) {
 		SET_DID_STATUS(bsg_reply->result, DID_ERROR);
 		rval = -1;
 	} else {
 		struct fc_port	*fcport = NULL, *tf;
-		uint32_t	pcnt = 0;
 
 		list_for_each_entry_safe(fcport, tf, &vha->vp_fcports, list) {
 			if (fcport->edif.enable) {
@@ -1054,9 +992,11 @@ qla_edif_app_getstats(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 		SET_DID_STATUS(bsg_reply->result, DID_OK);
 	}
 
+	bsg_job->reply_len = sizeof(struct fc_bsg_reply);
 	bsg_reply->reply_payload_rcv_len =
 	    sg_copy_from_buffer(bsg_job->reply_payload.sg_list,
-	       bsg_job->reply_payload.sg_cnt, app_reply, ret_size);
+	       bsg_job->reply_payload.sg_cnt, app_reply,
+	       sizeof(struct app_stats_reply) + (sizeof(struct app_sinfo) * pcnt));
 
 	kfree(app_reply);
 
@@ -1130,8 +1070,7 @@ qla_edif_app_mgmt(struct bsg_job *bsg_job)
 		    __func__,
 		    bsg_request->rqst_data.h_vendor.vendor_cmd[1]);
 		rval = EXT_STATUS_INVALID_PARAM;
-		bsg_job->reply_len = sizeof(struct fc_bsg_reply);
-		SET_DID_STATUS(bsg_reply->result, DID_ERROR);
+		done = false;
 		break;
 	}
 
@@ -1330,7 +1269,7 @@ qla24xx_sadb_update(struct bsg_job *bsg_job)
 		goto done;
 	}
 
-	if (vha->e_dbell.db_flags != EDB_ACTIVE) {
+	if (DBELL_INACTIVE(vha)) {
 		ql_log(ql_log_warn, vha, 0x70a1, "App not started\n");
 		rval = -EIO;
 		SET_DID_STATUS(bsg_reply->result, DID_ERROR);
@@ -1651,6 +1590,40 @@ qla_enode_stop(scsi_qla_host_t *vha)
 	spin_unlock_irqrestore(&vha->pur_cinfo.pur_lock, flags);
 }
 
+static void qla_enode_clear(scsi_qla_host_t *vha, port_id_t portid)
+{
+	unsigned    long flags;
+	struct enode    *e, *tmp;
+	struct purexevent   *purex;
+	LIST_HEAD(enode_list);
+
+	if (vha->pur_cinfo.enode_flags != ENODE_ACTIVE) {
+		ql_dbg(ql_dbg_edif, vha, 0x09102,
+		       "%s enode not active\n", __func__);
+		return;
+	}
+	spin_lock_irqsave(&vha->pur_cinfo.pur_lock, flags);
+	list_for_each_entry_safe(e, tmp, &vha->pur_cinfo.head, list) {
+		purex = &e->u.purexinfo;
+		if (purex->pur_info.pur_sid.b24 == portid.b24) {
+			ql_dbg(ql_dbg_edif, vha, 0x911d,
+			    "%s free ELS sid=%06x. xchg %x, nb=%xh\n",
+			    __func__, portid.b24,
+			    purex->pur_info.pur_rx_xchg_address,
+			    purex->pur_info.pur_bytes_rcvd);
+
+			list_del_init(&e->list);
+			list_add_tail(&e->list, &enode_list);
+		}
+	}
+	spin_unlock_irqrestore(&vha->pur_cinfo.pur_lock, flags);
+
+	list_for_each_entry_safe(e, tmp, &enode_list, list) {
+		list_del_init(&e->list);
+		qla_enode_free(vha, e);
+	}
+}
+
 /*
  *  allocate enode struct and populate buffer
  *  returns: enode pointer with buffers
@@ -1695,41 +1668,25 @@ static struct enode *
 qla_enode_find(scsi_qla_host_t *vha, uint32_t ntype, uint32_t p1, uint32_t p2)
 {
 	struct enode		*node_rtn = NULL;
-	struct enode		*list_node = NULL;
+	struct enode		*list_node, *q;
 	unsigned long		flags;
-	struct list_head	*pos, *q;
 	uint32_t		sid;
-	uint32_t		rw_flag;
 	struct purexevent	*purex;
 
 	/* secure the list from moving under us */
 	spin_lock_irqsave(&vha->pur_cinfo.pur_lock, flags);
 
-	list_for_each_safe(pos, q, &vha->pur_cinfo.head) {
-		list_node = list_entry(pos, struct enode, list);
+	list_for_each_entry_safe(list_node, q, &vha->pur_cinfo.head, list) {
 
 		/* node type determines what p1 and p2 are */
 		purex = &list_node->u.purexinfo;
 		sid = p1;
-		rw_flag = p2;
 
 		if (purex->pur_info.pur_sid.b24 == sid) {
-			if (purex->pur_info.pur_pend == 1 &&
-			    rw_flag == PUR_GET) {
-				/*
-				 * if the receive is in progress
-				 * and its a read/get then can't
-				 * transfer yet
-				 */
-				ql_dbg(ql_dbg_edif, vha, 0x9106,
-				    "%s purex xfer in progress for sid=%x\n",
-				    __func__, sid);
-			} else {
-				/* found it and its complete */
-				node_rtn = list_node;
-				list_del(pos);
-				break;
-			}
+			/* found it and its complete */
+			node_rtn = list_node;
+			list_del(&list_node->list);
+			break;
 		}
 	}
 
@@ -1802,7 +1759,8 @@ qla_els_reject_iocb(scsi_qla_host_t *vha, struct qla_qpair *qp,
 	qla_els_pt_iocb(vha, els_iocb, a);
 
 	ql_dbg(ql_dbg_edif, vha, 0x0183,
-	    "Sending ELS reject...\n");
+	    "Sending ELS reject ox_id %04x s:%06x -> d:%06x\n",
+	    a->ox_id, a->sid.b24, a->did.b24);
 	ql_dump_buffer(ql_dbg_edif + ql_dbg_verbose, vha, 0x0185,
 	    vha->hw->elsrej.c, sizeof(*vha->hw->elsrej.c));
 	/* flush iocb to mem before notifying hw doorbell */
@@ -1814,7 +1772,7 @@ qla_els_reject_iocb(scsi_qla_host_t *vha, struct qla_qpair *qp,
 void
 qla_edb_init(scsi_qla_host_t *vha)
 {
-	if (vha->e_dbell.db_flags == EDB_ACTIVE) {
+	if (DBELL_ACTIVE(vha)) {
 		/* list already init'd - error */
 		ql_dbg(ql_dbg_edif, vha, 0x09102,
 		    "edif db already initialized, cannot reinit\n");
@@ -1850,6 +1808,57 @@ qla_edb_node_free(scsi_qla_host_t *vha, struct edb_node *node)
 	node->ntype = N_UNDEF;
 }
 
+static void qla_edb_clear(scsi_qla_host_t *vha, port_id_t portid)
+{
+	unsigned long flags;
+	struct edb_node *e, *tmp;
+	port_id_t sid;
+	LIST_HEAD(edb_list);
+
+	if (DBELL_INACTIVE(vha)) {
+		/* doorbell list not enabled */
+		ql_dbg(ql_dbg_edif, vha, 0x09102,
+		       "%s doorbell not enabled\n", __func__);
+		return;
+	}
+
+	/* grab lock so list doesn't move */
+	spin_lock_irqsave(&vha->e_dbell.db_lock, flags);
+	list_for_each_entry_safe(e, tmp, &vha->e_dbell.head, list) {
+		switch (e->ntype) {
+		case VND_CMD_AUTH_STATE_NEEDED:
+		case VND_CMD_AUTH_STATE_SESSION_SHUTDOWN:
+			sid = e->u.plogi_did;
+			break;
+		case VND_CMD_AUTH_STATE_ELS_RCVD:
+			sid = e->u.els_sid;
+			break;
+		case VND_CMD_AUTH_STATE_SAUPDATE_COMPL:
+			/* app wants to see this  */
+			continue;
+		default:
+			ql_log(ql_log_warn, vha, 0x09102,
+			       "%s unknown node type: %x\n", __func__, e->ntype);
+			sid.b24 = 0;
+			break;
+		}
+		if (sid.b24 == portid.b24) {
+			ql_dbg(ql_dbg_edif, vha, 0x910f,
+			       "%s free doorbell event : node type = %x %p\n",
+			       __func__, e->ntype, e);
+			list_del_init(&e->list);
+			list_add_tail(&e->list, &edb_list);
+		}
+	}
+	spin_unlock_irqrestore(&vha->e_dbell.db_lock, flags);
+
+	list_for_each_entry_safe(e, tmp, &edb_list, list) {
+		qla_edb_node_free(vha, e);
+		list_del_init(&e->list);
+		kfree(e);
+	}
+}
+
 /* function called when app is stopping */
 
 void
@@ -1858,7 +1867,7 @@ qla_edb_stop(scsi_qla_host_t *vha)
 	unsigned long flags;
 	struct edb_node *node, *q;
 
-	if (vha->e_dbell.db_flags != EDB_ACTIVE) {
+	if (DBELL_INACTIVE(vha)) {
 		/* doorbell list not enabled */
 		ql_dbg(ql_dbg_edif, vha, 0x09102,
 		    "%s doorbell not enabled\n", __func__);
@@ -1909,7 +1918,7 @@ qla_edb_node_add(scsi_qla_host_t *vha, struct edb_node *ptr)
 {
 	unsigned long		flags;
 
-	if (vha->e_dbell.db_flags != EDB_ACTIVE) {
+	if (DBELL_INACTIVE(vha)) {
 		/* doorbell list not enabled */
 		ql_dbg(ql_dbg_edif, vha, 0x09102,
 		    "%s doorbell not enabled\n", __func__);
@@ -1940,7 +1949,7 @@ qla_edb_eventcreate(scsi_qla_host_t *vha, uint32_t dbtype,
 		return;
 	}
 
-	if (vha->e_dbell.db_flags != EDB_ACTIVE) {
+	if (DBELL_INACTIVE(vha)) {
 		if (fcport)
 			fcport->edif.auth_state = dbtype;
 		/* doorbell list not enabled */
@@ -2035,7 +2044,7 @@ qla_edif_timer(scsi_qla_host_t *vha)
 	struct qla_hw_data *ha = vha->hw;
 
 	if (!vha->vp_idx && N2N_TOPO(ha) && ha->flags.n2n_fw_acc_sec) {
-		if (vha->e_dbell.db_flags != EDB_ACTIVE &&
+		if (DBELL_INACTIVE(vha) &&
 		    ha->edif_post_stop_cnt_down) {
 			ha->edif_post_stop_cnt_down--;
 
@@ -2073,7 +2082,7 @@ edif_doorbell_show(struct device *dev, struct device_attribute *attr,
 	sz = 256;
 
 	/* stop new threads from waiting if we're not init'd */
-	if (vha->e_dbell.db_flags != EDB_ACTIVE) {
+	if (DBELL_INACTIVE(vha)) {
 		ql_dbg(ql_dbg_edif + ql_dbg_verbose, vha, 0x09122,
 		    "%s error - edif db not enabled\n", __func__);
 		return 0;
@@ -2346,6 +2355,7 @@ void qla24xx_auth_els(scsi_qla_host_t *vha, void **pkt, struct rsp_que **rsp)
 	a.tx_addr = vha->hw->elsrej.cdma;
 	a.vp_idx = vha->vp_idx;
 	a.control_flags = EPD_ELS_RJT;
+	a.ox_id = le16_to_cpu(p->ox_id);
 
 	sid = p->s_id[0] | (p->s_id[1] << 8) | (p->s_id[2] << 16);
 
@@ -2357,7 +2367,7 @@ void qla24xx_auth_els(scsi_qla_host_t *vha, void **pkt, struct rsp_que **rsp)
 		return;
 	}
 
-	if (totlen > MAX_PAYLOAD) {
+	if (totlen > ELS_MAX_PAYLOAD) {
 		ql_dbg(ql_dbg_edif, vha, 0x0910d,
 		    "%s WARNING: verbose ELS frame received (totlen=%x)\n",
 		    __func__, totlen);
@@ -2387,7 +2397,6 @@ void qla24xx_auth_els(scsi_qla_host_t *vha, void **pkt, struct rsp_que **rsp)
 
 	purex = &ptr->u.purexinfo;
 	purex->pur_info.pur_sid = a.did;
-	purex->pur_info.pur_pend = 0;
 	purex->pur_info.pur_bytes_rcvd = totlen;
 	purex->pur_info.pur_rx_xchg_address = le32_to_cpu(p->rx_xchg_addr);
 	purex->pur_info.pur_nphdl = le16_to_cpu(p->nport_handle);
@@ -2396,6 +2405,8 @@ void qla24xx_auth_els(scsi_qla_host_t *vha, void **pkt, struct rsp_que **rsp)
 	purex->pur_info.pur_did.b.al_pa =  p->d_id[0];
 	purex->pur_info.vp_idx = p->vp_idx;
 
+	a.sid = purex->pur_info.pur_did;
+
 	rc = __qla_copy_purex_to_buffer(vha, pkt, rsp, purex->msgp,
 		purex->msgp_len);
 	if (rc) {
@@ -2419,7 +2430,7 @@ void qla24xx_auth_els(scsi_qla_host_t *vha, void **pkt, struct rsp_que **rsp)
 
 	fcport = qla2x00_find_fcport_by_pid(host, &purex->pur_info.pur_sid);
 
-	if (host->e_dbell.db_flags != EDB_ACTIVE ||
+	if (DBELL_INACTIVE(vha) ||
 	    (fcport && EDIF_SESSION_DOWN(fcport))) {
 		ql_dbg(ql_dbg_edif, host, 0x0910c, "%s e_dbell.db_flags =%x %06x\n",
 		    __func__, host->e_dbell.db_flags,
@@ -2436,7 +2447,7 @@ void qla24xx_auth_els(scsi_qla_host_t *vha, void **pkt, struct rsp_que **rsp)
 	ql_dbg(ql_dbg_edif, host, 0x0910c,
 	    "%s COMPLETE purex->pur_info.pur_bytes_rcvd =%xh s:%06x -> d:%06x xchg=%xh\n",
 	    __func__, purex->pur_info.pur_bytes_rcvd, purex->pur_info.pur_sid.b24,
-	    purex->pur_info.pur_did.b24, p->rx_xchg_addr);
+	    purex->pur_info.pur_did.b24, purex->pur_info.pur_rx_xchg_address);
 
 	qla_edb_eventcreate(host, VND_CMD_AUTH_STATE_ELS_RCVD, sid, 0, NULL);
 }
@@ -3139,18 +3150,14 @@ static uint16_t qla_edif_sadb_get_sa_index(fc_port_t *fcport,
 /* release any sadb entries -- only done at teardown */
 void qla_edif_sadb_release(struct qla_hw_data *ha)
 {
-	struct list_head *pos;
-	struct list_head *tmp;
-	struct edif_sa_index_entry *entry;
+	struct edif_sa_index_entry *entry, *tmp;
 
-	list_for_each_safe(pos, tmp, &ha->sadb_rx_index_list) {
-		entry = list_entry(pos, struct edif_sa_index_entry, next);
+	list_for_each_entry_safe(entry, tmp, &ha->sadb_rx_index_list, next) {
 		list_del(&entry->next);
 		kfree(entry);
 	}
 
-	list_for_each_safe(pos, tmp, &ha->sadb_tx_index_list) {
-		entry = list_entry(pos, struct edif_sa_index_entry, next);
+	list_for_each_entry_safe(entry, tmp, &ha->sadb_tx_index_list, next) {
 		list_del(&entry->next);
 		kfree(entry);
 	}
@@ -3449,7 +3456,7 @@ int qla_edif_process_els(scsi_qla_host_t *vha, struct bsg_job *bsg_job)
 
 void qla_edif_sess_down(struct scsi_qla_host *vha, struct fc_port *sess)
 {
-	if (sess->edif.app_sess_online && vha->e_dbell.db_flags & EDB_ACTIVE) {
+	if (sess->edif.app_sess_online && DBELL_ACTIVE(vha)) {
 		ql_dbg(ql_dbg_disc, vha, 0xf09c,
 			"%s: sess %8phN send port_offline event\n",
 			__func__, sess->port_name);
@@ -3459,3 +3466,12 @@ void qla_edif_sess_down(struct scsi_qla_host *vha, struct fc_port *sess)
 		qla2x00_post_aen_work(vha, FCH_EVT_PORT_OFFLINE, sess->d_id.b24);
 	}
 }
+
+void qla_edif_clear_appdata(struct scsi_qla_host *vha, struct fc_port *fcport)
+{
+	if (!(fcport->flags & FCF_FCSP_DEVICE))
+		return;
+
+	qla_edb_clear(vha, fcport->d_id);
+	qla_enode_clear(vha, fcport->d_id);
+}
diff --git a/drivers/scsi/qla2xxx/qla_edif.h b/drivers/scsi/qla2xxx/qla_edif.h
index 9e8f28d..a965ca8 100644
--- a/drivers/scsi/qla2xxx/qla_edif.h
+++ b/drivers/scsi/qla2xxx/qla_edif.h
@@ -41,9 +41,12 @@ struct pur_core {
 };
 
 enum db_flags_t {
-	EDB_ACTIVE = 0x1,
+	EDB_ACTIVE = BIT_0,
 };
 
+#define DBELL_ACTIVE(_v) (_v->e_dbell.db_flags & EDB_ACTIVE)
+#define DBELL_INACTIVE(_v) (!(_v->e_dbell.db_flags & EDB_ACTIVE))
+
 struct edif_dbell {
 	enum db_flags_t		db_flags;
 	spinlock_t		db_lock;
@@ -93,7 +96,6 @@ struct sa_update_28xx {
 };
 
 #define        NUM_ENTRIES     256
-#define        MAX_PAYLOAD     1024
 #define        PUR_GET         1
 
 struct dinfo {
@@ -102,7 +104,6 @@ struct dinfo {
 };
 
 struct pur_ninfo {
-	unsigned int	pur_pend:1;
 	port_id_t       pur_sid;
 	port_id_t	pur_did;
 	uint8_t		vp_idx;
@@ -128,9 +129,15 @@ struct enode {
 	} u;
 };
 
+#define RX_ELS_SIZE (roundup(sizeof(struct enode) + ELS_MAX_PAYLOAD, SMP_CACHE_BYTES))
+
 #define EDIF_SESSION_DOWN(_s) \
 	(qla_ini_mode_enabled(_s->vha) && (_s->disc_state == DSC_DELETE_PEND || \
 	 _s->disc_state == DSC_DELETED || \
 	 !_s->edif.app_sess_online))
 
+#define EDIF_NEGOTIATION_PENDING(_fcport) \
+	(DBELL_ACTIVE(_fcport->vha) && \
+	 (_fcport->disc_state == DSC_LOGIN_AUTH_PEND))
+
 #endif	/* __QLA_EDIF_H */
diff --git a/drivers/scsi/qla2xxx/qla_edif_bsg.h b/drivers/scsi/qla2xxx/qla_edif_bsg.h
index 58b718d..53026d8 100644
--- a/drivers/scsi/qla2xxx/qla_edif_bsg.h
+++ b/drivers/scsi/qla2xxx/qla_edif_bsg.h
@@ -8,7 +8,7 @@
 #define __QLA_EDIF_BSG_H
 
 /* BSG Vendor specific commands */
-#define	ELS_MAX_PAYLOAD		1024
+#define	ELS_MAX_PAYLOAD		2112
 #ifndef	WWN_SIZE
 #define WWN_SIZE		8
 #endif
diff --git a/drivers/scsi/qla2xxx/qla_gbl.h b/drivers/scsi/qla2xxx/qla_gbl.h
index 3c4fa8b..8d8503a 100644
--- a/drivers/scsi/qla2xxx/qla_gbl.h
+++ b/drivers/scsi/qla2xxx/qla_gbl.h
@@ -142,6 +142,8 @@ void qlt_chk_edif_rx_sa_delete_pending(scsi_qla_host_t *vha, fc_port_t *fcport,
 void qla2x00_release_all_sadb(struct scsi_qla_host *vha, struct fc_port *fcport);
 int qla_edif_process_els(scsi_qla_host_t *vha, struct bsg_job *bsgjob);
 void qla_edif_sess_down(struct scsi_qla_host *vha, struct fc_port *sess);
+void qla_edif_clear_appdata(struct scsi_qla_host *vha,
+			    struct fc_port *fcport);
 const char *sc_to_str(uint16_t cmd);
 
 /*
@@ -171,7 +173,6 @@ extern int ql2xasynctmfenable;
 extern int ql2xgffidenable;
 extern int ql2xenabledif;
 extern int ql2xenablehba_err_chk;
-extern int ql2xtargetreset;
 extern int ql2xdontresethba;
 extern uint64_t ql2xmaxlun;
 extern int ql2xmdcapmask;
@@ -818,7 +819,6 @@ extern void qlafx00_abort_iocb(srb_t *, struct abort_iocb_entry_fx00 *);
 extern void qlafx00_fxdisc_iocb(srb_t *, struct fxdisc_entry_fx00 *);
 extern void qlafx00_timer_routine(scsi_qla_host_t *);
 extern int qlafx00_rescan_isp(scsi_qla_host_t *);
-extern int qlafx00_loop_reset(scsi_qla_host_t *vha);
 
 /* qla82xx related functions */
 
diff --git a/drivers/scsi/qla2xxx/qla_init.c b/drivers/scsi/qla2xxx/qla_init.c
index 908a72e..070b636 100644
--- a/drivers/scsi/qla2xxx/qla_init.c
+++ b/drivers/scsi/qla2xxx/qla_init.c
@@ -330,12 +330,9 @@ qla2x00_async_login(struct scsi_qla_host *vha, fc_port_t *fcport,
 		lio->u.logio.flags |= SRB_LOGIN_PRLI_ONLY;
 	} else {
 		if (vha->hw->flags.edif_enabled &&
-		    vha->e_dbell.db_flags & EDB_ACTIVE) {
+		    DBELL_ACTIVE(vha)) {
 			lio->u.logio.flags |=
 				(SRB_LOGIN_FCSP | SRB_LOGIN_SKIP_PRLI);
-			ql_dbg(ql_dbg_disc, vha, 0x2072,
-			    "Async-login: w/ FCSP %8phC hdl=%x, loopid=%x portid=%06x\n",
-			    fcport->port_name, sp->handle, fcport->loop_id, fcport->d_id.b24);
 		} else {
 			lio->u.logio.flags |= SRB_LOGIN_COND_PLOGI;
 		}
@@ -344,12 +341,14 @@ qla2x00_async_login(struct scsi_qla_host *vha, fc_port_t *fcport,
 	if (NVME_TARGET(vha->hw, fcport))
 		lio->u.logio.flags |= SRB_LOGIN_SKIP_PRLI;
 
-	ql_dbg(ql_dbg_disc, vha, 0x2072,
-	       "Async-login - %8phC hdl=%x, loopid=%x portid=%06x retries=%d.\n",
-	       fcport->port_name, sp->handle, fcport->loop_id,
-	       fcport->d_id.b24, fcport->login_retry);
-
 	rval = qla2x00_start_sp(sp);
+
+	ql_dbg(ql_dbg_disc, vha, 0x2072,
+	       "Async-login - %8phC hdl=%x, loopid=%x portid=%06x retries=%d %s.\n",
+	       fcport->port_name, sp->handle, fcport->loop_id,
+	       fcport->d_id.b24, fcport->login_retry,
+	       lio->u.logio.flags & SRB_LOGIN_FCSP ? "FCSP" : "");
+
 	if (rval != QLA_SUCCESS) {
 		fcport->flags |= FCF_LOGIN_NEEDED;
 		set_bit(RELOGIN_NEEDED, &vha->dpc_flags);
@@ -862,7 +861,7 @@ static void qla24xx_handle_gnl_done_event(scsi_qla_host_t *vha,
 				break;
 			case DSC_LS_PLOGI_COMP:
 				if (vha->hw->flags.edif_enabled &&
-				    vha->e_dbell.db_flags & EDB_ACTIVE) {
+				    DBELL_ACTIVE(vha)) {
 					/* check to see if App support secure or not */
 					qla24xx_post_gpdb_work(vha, fcport, 0);
 					break;
@@ -987,8 +986,6 @@ static void qla24xx_async_gnl_sp_done(srb_t *sp, int res)
 	    sp->name, res, sp->u.iocb_cmd.u.mbx.in_mb[1],
 	    sp->u.iocb_cmd.u.mbx.in_mb[2]);
 
-	if (res == QLA_FUNCTION_TIMEOUT)
-		return;
 
 	sp->fcport->flags &= ~(FCF_ASYNC_SENT|FCF_ASYNC_ACTIVE);
 	memset(&ea, 0, sizeof(ea));
@@ -1026,8 +1023,8 @@ static void qla24xx_async_gnl_sp_done(srb_t *sp, int res)
 	spin_unlock_irqrestore(&vha->hw->tgt.sess_lock, flags);
 
 	list_for_each_entry_safe(fcport, tf, &h, gnl_entry) {
-		list_del_init(&fcport->gnl_entry);
 		spin_lock_irqsave(&vha->hw->tgt.sess_lock, flags);
+		list_del_init(&fcport->gnl_entry);
 		fcport->flags &= ~(FCF_ASYNC_SENT | FCF_ASYNC_ACTIVE);
 		spin_unlock_irqrestore(&vha->hw->tgt.sess_lock, flags);
 		ea.fcport = fcport;
@@ -1454,7 +1451,7 @@ static int	qla_chk_secure_login(scsi_qla_host_t	*vha, fc_port_t *fcport,
 			qla2x00_post_aen_work(vha, FCH_EVT_PORT_ONLINE,
 			    fcport->d_id.b24);
 
-			if (vha->e_dbell.db_flags ==  EDB_ACTIVE) {
+			if (DBELL_ACTIVE(vha)) {
 				ql_dbg(ql_dbg_disc, vha, 0x20ef,
 				    "%s %d %8phC EDIF: post DB_AUTH: AUTH needed\n",
 				    __func__, __LINE__, fcport->port_name);
@@ -1786,16 +1783,72 @@ void qla2x00_handle_rscn(scsi_qla_host_t *vha, struct event_arg *ea)
 	fc_port_t *fcport;
 	unsigned long flags;
 
-	fcport = qla2x00_find_fcport_by_nportid(vha, &ea->id, 1);
-	if (fcport) {
-		if (fcport->flags & FCF_FCP2_DEVICE) {
-			ql_dbg(ql_dbg_disc, vha, 0x2115,
-			       "Delaying session delete for FCP2 portid=%06x %8phC ",
-			       fcport->d_id.b24, fcport->port_name);
-			return;
+	switch (ea->id.b.rsvd_1) {
+	case RSCN_PORT_ADDR:
+		fcport = qla2x00_find_fcport_by_nportid(vha, &ea->id, 1);
+		if (fcport) {
+			if (fcport->flags & FCF_FCP2_DEVICE) {
+				ql_dbg(ql_dbg_disc, vha, 0x2115,
+				       "Delaying session delete for FCP2 portid=%06x %8phC ",
+					fcport->d_id.b24, fcport->port_name);
+				return;
+			}
+
+			if (vha->hw->flags.edif_enabled && DBELL_ACTIVE(vha)) {
+				/*
+				 * On ipsec start by remote port, Target port
+				 * may use RSCN to trigger initiator to
+				 * relogin. If driver is already in the
+				 * process of a relogin, then ignore the RSCN
+				 * and allow the current relogin to continue.
+				 * This reduces thrashing of the connection.
+				 */
+				if (atomic_read(&fcport->state) == FCS_ONLINE) {
+					/*
+					 * If state = online, then set scan_needed=1 to do relogin.
+					 * Otherwise we're already in the middle of a relogin
+					 */
+					fcport->scan_needed = 1;
+					fcport->rscn_gen++;
+				}
+			} else {
+				fcport->scan_needed = 1;
+				fcport->rscn_gen++;
+			}
 		}
-		fcport->scan_needed = 1;
-		fcport->rscn_gen++;
+		break;
+	case RSCN_AREA_ADDR:
+		list_for_each_entry(fcport, &vha->vp_fcports, list) {
+			if (fcport->flags & FCF_FCP2_DEVICE)
+				continue;
+
+			if ((ea->id.b24 & 0xffff00) == (fcport->d_id.b24 & 0xffff00)) {
+				fcport->scan_needed = 1;
+				fcport->rscn_gen++;
+			}
+		}
+		break;
+	case RSCN_DOM_ADDR:
+		list_for_each_entry(fcport, &vha->vp_fcports, list) {
+			if (fcport->flags & FCF_FCP2_DEVICE)
+				continue;
+
+			if ((ea->id.b24 & 0xff0000) == (fcport->d_id.b24 & 0xff0000)) {
+				fcport->scan_needed = 1;
+				fcport->rscn_gen++;
+			}
+		}
+		break;
+	case RSCN_FAB_ADDR:
+	default:
+		list_for_each_entry(fcport, &vha->vp_fcports, list) {
+			if (fcport->flags & FCF_FCP2_DEVICE)
+				continue;
+
+			fcport->scan_needed = 1;
+			fcport->rscn_gen++;
+		}
+		break;
 	}
 
 	spin_lock_irqsave(&vha->work_lock, flags);
@@ -4187,7 +4240,7 @@ qla24xx_update_fw_options(scsi_qla_host_t *vha)
 		 * fw shal not send PRLI after PLOGI Acc
 		 */
 		if (ha->flags.edif_enabled &&
-		    vha->e_dbell.db_flags & EDB_ACTIVE) {
+		    DBELL_ACTIVE(vha)) {
 			ha->fw_options[3] |= BIT_15;
 			ha->flags.n2n_fw_acc_sec = 1;
 		} else {
@@ -4433,6 +4486,10 @@ qla2x00_init_rings(scsi_qla_host_t *vha)
 		    (ha->flags.fawwpn_enabled) ? "enabled" : "disabled");
 	}
 
+	/* ELS pass through payload is limit by frame size. */
+	if (ha->flags.edif_enabled)
+		mid_init_cb->init_cb.frame_payload_size = cpu_to_le16(ELS_MAX_PAYLOAD);
+
 	rval = qla2x00_init_firmware(vha, ha->init_cb_size);
 next_check:
 	if (rval) {
@@ -5339,8 +5396,7 @@ qla2x00_configure_loop(scsi_qla_host_t *vha)
 			 * use link up to wake up app to get ready for
 			 * authentication.
 			 */
-			if (ha->flags.edif_enabled &&
-			    !(vha->e_dbell.db_flags & EDB_ACTIVE))
+			if (ha->flags.edif_enabled && DBELL_INACTIVE(vha))
 				qla2x00_post_aen_work(vha, FCH_EVT_LINKUP,
 						      ha->link_data_rate);
 
@@ -5833,6 +5889,10 @@ void qla_register_fcport_fn(struct work_struct *work)
 
 	qla2x00_update_fcport(fcport->vha, fcport);
 
+	ql_dbg(ql_dbg_disc, fcport->vha, 0x911e,
+	       "%s rscn gen %d/%d next DS %d\n", __func__,
+	       rscn_gen, fcport->rscn_gen, fcport->next_disc_state);
+
 	if (rscn_gen != fcport->rscn_gen) {
 		/* RSCN(s) came in while registration */
 		switch (fcport->next_disc_state) {
diff --git a/drivers/scsi/qla2xxx/qla_iocb.c b/drivers/scsi/qla2xxx/qla_iocb.c
index 9d4ad1d..ed604f2 100644
--- a/drivers/scsi/qla2xxx/qla_iocb.c
+++ b/drivers/scsi/qla2xxx/qla_iocb.c
@@ -3034,8 +3034,7 @@ qla24xx_els_dcmd2_iocb(scsi_qla_host_t *vha, int els_opcode,
 	elsio->u.els_plogi.els_cmd = els_opcode;
 	elsio->u.els_plogi.els_plogi_pyld->opcode = els_opcode;
 
-	if (els_opcode == ELS_DCMD_PLOGI && vha->hw->flags.edif_enabled &&
-	    vha->e_dbell.db_flags & EDB_ACTIVE) {
+	if (els_opcode == ELS_DCMD_PLOGI && DBELL_ACTIVE(vha)) {
 		struct fc_els_flogi *p = ptr;
 
 		p->fl_csp.sp_features |= cpu_to_be16(FC_SP_FT_SEC);
diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c
index b26f269..aaf6504 100644
--- a/drivers/scsi/qla2xxx/qla_isr.c
+++ b/drivers/scsi/qla2xxx/qla_isr.c
@@ -2233,6 +2233,10 @@ qla24xx_els_ct_entry(scsi_qla_host_t *v, struct req_que *req,
 				}
 
 			} else if (comp_status == CS_PORT_LOGGED_OUT) {
+				ql_dbg(ql_dbg_disc, vha, 0x911e,
+				       "%s %d schedule session deletion\n",
+				       __func__, __LINE__);
+
 				els->u.els_plogi.len = 0;
 				res = DID_IMM_RETRY << 16;
 				qlt_schedule_sess_for_deletion(sp->fcport);
diff --git a/drivers/scsi/qla2xxx/qla_mr.c b/drivers/scsi/qla2xxx/qla_mr.c
index 6e920da..350b0c4 100644
--- a/drivers/scsi/qla2xxx/qla_mr.c
+++ b/drivers/scsi/qla2xxx/qla_mr.c
@@ -739,29 +739,6 @@ qlafx00_lun_reset(fc_port_t *fcport, uint64_t l, int tag)
 }
 
 int
-qlafx00_loop_reset(scsi_qla_host_t *vha)
-{
-	int ret;
-	struct fc_port *fcport;
-	struct qla_hw_data *ha = vha->hw;
-
-	if (ql2xtargetreset) {
-		list_for_each_entry(fcport, &vha->vp_fcports, list) {
-			if (fcport->port_type != FCT_TARGET)
-				continue;
-
-			ret = ha->isp_ops->target_reset(fcport, 0, 0);
-			if (ret != QLA_SUCCESS) {
-				ql_dbg(ql_dbg_taskm, vha, 0x803d,
-				    "Bus Reset failed: Reset=%d "
-				    "d_id=%x.\n", ret, fcport->d_id.b24);
-			}
-		}
-	}
-	return QLA_SUCCESS;
-}
-
-int
 qlafx00_iospace_config(struct qla_hw_data *ha)
 {
 	if (pci_request_selected_regions(ha->pdev, ha->bars,
diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
index c3ff50f..abcd309 100644
--- a/drivers/scsi/qla2xxx/qla_os.c
+++ b/drivers/scsi/qla2xxx/qla_os.c
@@ -202,12 +202,6 @@ MODULE_PARM_DESC(ql2xdbwr,
 		" 0 -- Regular doorbell.\n"
 		" 1 -- CAMRAM doorbell (faster).\n");
 
-int ql2xtargetreset = 1;
-module_param(ql2xtargetreset, int, S_IRUGO);
-MODULE_PARM_DESC(ql2xtargetreset,
-		 "Enable target reset."
-		 "Default is 1 - use hw defaults.");
-
 int ql2xgffidenable;
 module_param(ql2xgffidenable, int, S_IRUGO);
 MODULE_PARM_DESC(ql2xgffidenable,
@@ -1695,27 +1689,10 @@ int
 qla2x00_loop_reset(scsi_qla_host_t *vha)
 {
 	int ret;
-	struct fc_port *fcport;
 	struct qla_hw_data *ha = vha->hw;
 
-	if (IS_QLAFX00(ha)) {
-		return qlafx00_loop_reset(vha);
-	}
-
-	if (ql2xtargetreset == 1 && ha->flags.enable_target_reset) {
-		list_for_each_entry(fcport, &vha->vp_fcports, list) {
-			if (fcport->port_type != FCT_TARGET)
-				continue;
-
-			ret = ha->isp_ops->target_reset(fcport, 0, 0);
-			if (ret != QLA_SUCCESS) {
-				ql_dbg(ql_dbg_taskm, vha, 0x802c,
-				    "Bus Reset failed: Reset=%d "
-				    "d_id=%x.\n", ret, fcport->d_id.b24);
-			}
-		}
-	}
-
+	if (IS_QLAFX00(ha))
+		return QLA_SUCCESS;
 
 	if (ha->flags.enable_lip_full_login && !IS_CNA_CAPABLE(ha)) {
 		atomic_set(&vha->loop_state, LOOP_DOWN);
@@ -3908,13 +3885,13 @@ qla2x00_remove_one(struct pci_dev *pdev)
 static inline void
 qla24xx_free_purex_list(struct purex_list *list)
 {
-	struct list_head *item, *next;
+	struct purex_item *item, *next;
 	ulong flags;
 
 	spin_lock_irqsave(&list->lock, flags);
-	list_for_each_safe(item, next, &list->head) {
-		list_del(item);
-		kfree(list_entry(item, struct purex_item, list));
+	list_for_each_entry_safe(item, next, &list->head, list) {
+		list_del(&item->list);
+		kfree(item);
 	}
 	spin_unlock_irqrestore(&list->lock, flags);
 }
@@ -4375,7 +4352,7 @@ qla2x00_mem_alloc(struct qla_hw_data *ha, uint16_t req_len, uint16_t rsp_len,
 
 	/* allocate the purex dma pool */
 	ha->purex_dma_pool = dma_pool_create(name, &ha->pdev->dev,
-	    MAX_PAYLOAD, 8, 0);
+	    ELS_MAX_PAYLOAD, 8, 0);
 
 	if (!ha->purex_dma_pool) {
 		ql_dbg_pci(ql_dbg_init, ha->pdev, 0x011b,
diff --git a/drivers/scsi/qla2xxx/qla_target.c b/drivers/scsi/qla2xxx/qla_target.c
index 7d8242c..8993d43 100644
--- a/drivers/scsi/qla2xxx/qla_target.c
+++ b/drivers/scsi/qla2xxx/qla_target.c
@@ -1003,6 +1003,7 @@ void qlt_free_session_done(struct work_struct *work)
 					"%s bypassing release_all_sadb\n",
 					__func__);
 			}
+			qla_edif_clear_appdata(vha, sess);
 			qla_edif_sess_down(vha, sess);
 		}
 		qla2x00_mark_device_lost(vha, sess, 0);
@@ -4812,7 +4813,7 @@ static int qlt_handle_login(struct scsi_qla_host *vha,
 	}
 
 	if (vha->hw->flags.edif_enabled) {
-		if (!(vha->e_dbell.db_flags & EDB_ACTIVE)) {
+		if (DBELL_INACTIVE(vha)) {
 			ql_dbg(ql_dbg_disc, vha, 0xffff,
 			       "%s %d Term INOT due to app not started lid=%d, NportID %06X ",
 			       __func__, __LINE__, loop_id, port_id.b24);
diff --git a/drivers/scsi/qla2xxx/qla_version.h b/drivers/scsi/qla2xxx/qla_version.h
index 4b11716..27e440f 100644
--- a/drivers/scsi/qla2xxx/qla_version.h
+++ b/drivers/scsi/qla2xxx/qla_version.h
@@ -6,9 +6,9 @@
 /*
  * Driver version
  */
-#define QLA2XXX_VERSION      "10.02.07.100-k"
+#define QLA2XXX_VERSION      "10.02.07.200-k"
 
 #define QLA_DRIVER_MAJOR_VER	10
 #define QLA_DRIVER_MINOR_VER	2
 #define QLA_DRIVER_PATCH_VER	7
-#define QLA_DRIVER_BETA_VER	100
+#define QLA_DRIVER_BETA_VER	200
diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
index 06e7266..1d0278d 100644
--- a/drivers/scsi/scsi_debug.c
+++ b/drivers/scsi/scsi_debug.c
@@ -4259,6 +4259,8 @@ static int resp_verify(struct scsi_cmnd *scp, struct sdebug_dev_info *devip)
 		mk_sense_invalid_opcode(scp);
 		return check_condition_result;
 	}
+	if (vnum == 0)
+		return 0;	/* not an error */
 	a_num = is_bytchk3 ? 1 : vnum;
 	/* Treat following check like one for read (i.e. no write) access */
 	ret = check_device_access_params(scp, lba, a_num, false);
@@ -4322,6 +4324,8 @@ static int resp_report_zones(struct scsi_cmnd *scp,
 	}
 	zs_lba = get_unaligned_be64(cmd + 2);
 	alloc_len = get_unaligned_be32(cmd + 10);
+	if (alloc_len == 0)
+		return 0;	/* not an error */
 	rep_opts = cmd[14] & 0x3f;
 	partial = cmd[14] & 0x80;
 
diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c
index a531336..2371edb 100644
--- a/drivers/scsi/scsi_error.c
+++ b/drivers/scsi/scsi_error.c
@@ -133,6 +133,23 @@ static bool scsi_eh_should_retry_cmd(struct scsi_cmnd *cmd)
 	return true;
 }
 
+static void scsi_eh_complete_abort(struct scsi_cmnd *scmd, struct Scsi_Host *shost)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(shost->host_lock, flags);
+	list_del_init(&scmd->eh_entry);
+	/*
+	 * If the abort succeeds, and there is no further
+	 * EH action, clear the ->last_reset time.
+	 */
+	if (list_empty(&shost->eh_abort_list) &&
+	    list_empty(&shost->eh_cmd_q))
+		if (shost->eh_deadline != -1)
+			shost->last_reset = 0;
+	spin_unlock_irqrestore(shost->host_lock, flags);
+}
+
 /**
  * scmd_eh_abort_handler - Handle command aborts
  * @work:	command to be aborted.
@@ -150,6 +167,7 @@ scmd_eh_abort_handler(struct work_struct *work)
 		container_of(work, struct scsi_cmnd, abort_work.work);
 	struct scsi_device *sdev = scmd->device;
 	enum scsi_disposition rtn;
+	unsigned long flags;
 
 	if (scsi_host_eh_past_deadline(sdev->host)) {
 		SCSI_LOG_ERROR_RECOVERY(3,
@@ -173,12 +191,14 @@ scmd_eh_abort_handler(struct work_struct *work)
 				SCSI_LOG_ERROR_RECOVERY(3,
 					scmd_printk(KERN_WARNING, scmd,
 						    "retry aborted command\n"));
+				scsi_eh_complete_abort(scmd, sdev->host);
 				scsi_queue_insert(scmd, SCSI_MLQUEUE_EH_RETRY);
 				return;
 			} else {
 				SCSI_LOG_ERROR_RECOVERY(3,
 					scmd_printk(KERN_WARNING, scmd,
 						    "finish aborted command\n"));
+				scsi_eh_complete_abort(scmd, sdev->host);
 				scsi_finish_command(scmd);
 				return;
 			}
@@ -191,6 +211,9 @@ scmd_eh_abort_handler(struct work_struct *work)
 		}
 	}
 
+	spin_lock_irqsave(sdev->host->host_lock, flags);
+	list_del_init(&scmd->eh_entry);
+	spin_unlock_irqrestore(sdev->host->host_lock, flags);
 	scsi_eh_scmd_add(scmd);
 }
 
@@ -221,6 +244,8 @@ scsi_abort_command(struct scsi_cmnd *scmd)
 	spin_lock_irqsave(shost->host_lock, flags);
 	if (shost->eh_deadline != -1 && !shost->last_reset)
 		shost->last_reset = jiffies;
+	BUG_ON(!list_empty(&scmd->eh_entry));
+	list_add_tail(&scmd->eh_entry, &shost->eh_abort_list);
 	spin_unlock_irqrestore(shost->host_lock, flags);
 
 	scmd->eh_eflags |= SCSI_EH_ABORT_SCHEDULED;
diff --git a/drivers/scsi/scsi_ioctl.c b/drivers/scsi/scsi_ioctl.c
index 34412ea..400df33 100644
--- a/drivers/scsi/scsi_ioctl.c
+++ b/drivers/scsi/scsi_ioctl.c
@@ -347,6 +347,8 @@ static int scsi_fill_sghdr_rq(struct scsi_device *sdev, struct request *rq,
 {
 	struct scsi_request *req = scsi_req(rq);
 
+	if (hdr->cmd_len < 6)
+		return -EMSGSIZE;
 	if (copy_from_user(req->cmd, hdr->cmdp, hdr->cmd_len))
 		return -EFAULT;
 	if (!scsi_cmd_allowed(req->cmd, mode))
diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index b731c29..621d841 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -1153,6 +1153,7 @@ void scsi_init_command(struct scsi_device *dev, struct scsi_cmnd *cmd)
 	cmd->sense_buffer = buf;
 	cmd->prot_sdb = prot;
 	cmd->flags = flags;
+	INIT_LIST_HEAD(&cmd->eh_entry);
 	INIT_DELAYED_WORK(&cmd->abort_work, scmd_eh_abort_handler);
 	cmd->jiffies_at_alloc = jiffies_at_alloc;
 	cmd->retries = retries;
@@ -1184,8 +1185,6 @@ static blk_status_t scsi_setup_scsi_cmnd(struct scsi_device *sdev,
 	}
 
 	cmd->cmd_len = scsi_req(req)->cmd_len;
-	if (cmd->cmd_len == 0)
-		cmd->cmd_len = scsi_command_size(cmd->cmnd);
 	cmd->cmnd = scsi_req(req)->cmd;
 	cmd->transfersize = blk_rq_bytes(req);
 	cmd->allowed = scsi_req(req)->retries;
diff --git a/drivers/scsi/scsi_sysfs.c b/drivers/scsi/scsi_sysfs.c
index d3d3622..55addd7 100644
--- a/drivers/scsi/scsi_sysfs.c
+++ b/drivers/scsi/scsi_sysfs.c
@@ -1383,6 +1383,7 @@ int scsi_sysfs_add_sdev(struct scsi_device *sdev)
 			 * We're treating error on bsg register as non-fatal, so
 			 * pretend nothing went wrong.
 			 */
+			error = PTR_ERR(sdev->bsg_dev);
 			sdev_printk(KERN_INFO, sdev,
 				    "Failed to register bsg queue, errno=%d\n",
 				    error);
@@ -1580,7 +1581,6 @@ static struct device_type scsi_dev_type = {
 
 void scsi_sysfs_device_initialize(struct scsi_device *sdev)
 {
-	int i, j = 0;
 	unsigned long flags;
 	struct Scsi_Host *shost = sdev->host;
 	struct scsi_host_template *hostt = shost->hostt;
@@ -1592,15 +1592,7 @@ void scsi_sysfs_device_initialize(struct scsi_device *sdev)
 	scsi_enable_async_suspend(&sdev->sdev_gendev);
 	dev_set_name(&sdev->sdev_gendev, "%d:%d:%d:%llu",
 		     sdev->host->host_no, sdev->channel, sdev->id, sdev->lun);
-	sdev->gendev_attr_groups[j++] = &scsi_sdev_attr_group;
-	if (hostt->sdev_groups) {
-		for (i = 0; hostt->sdev_groups[i] &&
-			     j < ARRAY_SIZE(sdev->gendev_attr_groups);
-		     i++, j++) {
-			sdev->gendev_attr_groups[j] = hostt->sdev_groups[i];
-		}
-	}
-	WARN_ON_ONCE(j >= ARRAY_SIZE(sdev->gendev_attr_groups));
+	sdev->sdev_gendev.groups = hostt->sdev_groups;
 
 	device_initialize(&sdev->sdev_dev);
 	sdev->sdev_dev.parent = get_device(&sdev->sdev_gendev);
diff --git a/drivers/scsi/sr.c b/drivers/scsi/sr.c
index 50f8dc9..8e4af11 100644
--- a/drivers/scsi/sr.c
+++ b/drivers/scsi/sr.c
@@ -693,7 +693,6 @@ static int sr_probe(struct device *dev)
 	cd->device = sdev;
 	cd->disk = disk;
 	cd->driver = &sr_template;
-	cd->disk = disk;
 	cd->capacity = 0x1fffff;
 	cd->device->changed = 1;	/* force recheck CD type */
 	cd->media_present = 1;
diff --git a/drivers/scsi/ufs/ufs-debugfs.c b/drivers/scsi/ufs/ufs-debugfs.c
index 4e1ff20..4a0bbcf 100644
--- a/drivers/scsi/ufs/ufs-debugfs.c
+++ b/drivers/scsi/ufs/ufs-debugfs.c
@@ -8,6 +8,18 @@
 
 static struct dentry *ufs_debugfs_root;
 
+struct ufs_debugfs_attr {
+	const char			*name;
+	mode_t				mode;
+	const struct file_operations	*fops;
+};
+
+/* @file corresponds to a debugfs attribute in directory hba->debugfs_root. */
+static inline struct ufs_hba *hba_from_file(const struct file *file)
+{
+	return d_inode(file->f_path.dentry->d_parent)->i_private;
+}
+
 void __init ufs_debugfs_init(void)
 {
 	ufs_debugfs_root = debugfs_create_dir("ufshcd", NULL);
@@ -20,7 +32,7 @@ void ufs_debugfs_exit(void)
 
 static int ufs_debugfs_stats_show(struct seq_file *s, void *data)
 {
-	struct ufs_hba *hba = s->private;
+	struct ufs_hba *hba = hba_from_file(s->file);
 	struct ufs_event_hist *e = hba->ufs_stats.event;
 
 #define PRT(fmt, typ) \
@@ -126,13 +138,93 @@ static void ufs_debugfs_restart_ee(struct work_struct *work)
 	ufs_debugfs_put_user_access(hba);
 }
 
+static int ufs_saved_err_show(struct seq_file *s, void *data)
+{
+	struct ufs_debugfs_attr *attr = s->private;
+	struct ufs_hba *hba = hba_from_file(s->file);
+	const int *p;
+
+	if (strcmp(attr->name, "saved_err") == 0) {
+		p = &hba->saved_err;
+	} else if (strcmp(attr->name, "saved_uic_err") == 0) {
+		p = &hba->saved_uic_err;
+	} else {
+		return -ENOENT;
+	}
+
+	seq_printf(s, "%d\n", *p);
+	return 0;
+}
+
+static ssize_t ufs_saved_err_write(struct file *file, const char __user *buf,
+				   size_t count, loff_t *ppos)
+{
+	struct ufs_debugfs_attr *attr = file->f_inode->i_private;
+	struct ufs_hba *hba = hba_from_file(file);
+	char val_str[16] = { };
+	int val, ret;
+
+	if (count > sizeof(val_str))
+		return -EINVAL;
+	if (copy_from_user(val_str, buf, count))
+		return -EFAULT;
+	ret = kstrtoint(val_str, 0, &val);
+	if (ret < 0)
+		return ret;
+
+	spin_lock_irq(hba->host->host_lock);
+	if (strcmp(attr->name, "saved_err") == 0) {
+		hba->saved_err = val;
+	} else if (strcmp(attr->name, "saved_uic_err") == 0) {
+		hba->saved_uic_err = val;
+	} else {
+		ret = -ENOENT;
+	}
+	if (ret == 0)
+		ufshcd_schedule_eh_work(hba);
+	spin_unlock_irq(hba->host->host_lock);
+
+	return ret < 0 ? ret : count;
+}
+
+static int ufs_saved_err_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, ufs_saved_err_show, inode->i_private);
+}
+
+static const struct file_operations ufs_saved_err_fops = {
+	.owner		= THIS_MODULE,
+	.open		= ufs_saved_err_open,
+	.read		= seq_read,
+	.write		= ufs_saved_err_write,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+static const struct ufs_debugfs_attr ufs_attrs[] = {
+	{ "stats", 0400, &ufs_debugfs_stats_fops },
+	{ "saved_err", 0600, &ufs_saved_err_fops },
+	{ "saved_uic_err", 0600, &ufs_saved_err_fops },
+	{ }
+};
+
 void ufs_debugfs_hba_init(struct ufs_hba *hba)
 {
+	const struct ufs_debugfs_attr *attr;
+	struct dentry *root;
+
 	/* Set default exception event rate limit period to 20ms */
 	hba->debugfs_ee_rate_limit_ms = 20;
 	INIT_DELAYED_WORK(&hba->debugfs_ee_work, ufs_debugfs_restart_ee);
-	hba->debugfs_root = debugfs_create_dir(dev_name(hba->dev), ufs_debugfs_root);
-	debugfs_create_file("stats", 0400, hba->debugfs_root, hba, &ufs_debugfs_stats_fops);
+
+	root = debugfs_create_dir(dev_name(hba->dev), ufs_debugfs_root);
+	if (IS_ERR_OR_NULL(root))
+		return;
+	hba->debugfs_root = root;
+	d_inode(root)->i_private = hba;
+	for (attr = ufs_attrs; attr->name; attr++)
+		debugfs_create_file(attr->name, attr->mode, root, (void *)attr,
+				    attr->fops);
 	debugfs_create_file("exception_event_mask", 0600, hba->debugfs_root,
 			    hba, &ee_usr_mask_fops);
 	debugfs_create_u32("exception_event_rate_limit_ms", 0600, hba->debugfs_root,
diff --git a/drivers/scsi/ufs/ufs-exynos.c b/drivers/scsi/ufs/ufs-exynos.c
index 30d0c1a..cd26bc8 100644
--- a/drivers/scsi/ufs/ufs-exynos.c
+++ b/drivers/scsi/ufs/ufs-exynos.c
@@ -12,8 +12,10 @@
 #include <linux/module.h>
 #include <linux/of.h>
 #include <linux/of_address.h>
+#include <linux/mfd/syscon.h>
 #include <linux/phy/phy.h>
 #include <linux/platform_device.h>
+#include <linux/regmap.h>
 
 #include "ufshcd.h"
 #include "ufshcd-pltfrm.h"
@@ -48,10 +50,11 @@
 #define HCI_ERR_EN_T_LAYER	0x84
 #define HCI_ERR_EN_DME_LAYER	0x88
 #define HCI_CLKSTOP_CTRL	0xB0
+#define REFCLKOUT_STOP		BIT(4)
 #define REFCLK_STOP		BIT(2)
 #define UNIPRO_MCLK_STOP	BIT(1)
 #define UNIPRO_PCLK_STOP	BIT(0)
-#define CLK_STOP_MASK		(REFCLK_STOP |\
+#define CLK_STOP_MASK		(REFCLKOUT_STOP | REFCLK_STOP |\
 				 UNIPRO_MCLK_STOP |\
 				 UNIPRO_PCLK_STOP)
 #define HCI_MISC		0xB4
@@ -74,6 +77,52 @@
 				 UIC_TRANSPORT_NO_CONNECTION_RX |\
 				 UIC_TRANSPORT_BAD_TC)
 
+/* FSYS UFS Shareability */
+#define UFS_WR_SHARABLE		BIT(2)
+#define UFS_RD_SHARABLE		BIT(1)
+#define UFS_SHARABLE		(UFS_WR_SHARABLE | UFS_RD_SHARABLE)
+#define UFS_SHAREABILITY_OFFSET	0x710
+
+/* Multi-host registers */
+#define MHCTRL			0xC4
+#define MHCTRL_EN_VH_MASK	(0xE)
+#define MHCTRL_EN_VH(vh)	(vh << 1)
+#define PH2VH_MBOX		0xD8
+
+#define MH_MSG_MASK		(0xFF)
+
+#define MH_MSG(id, msg)		((id << 8) | (msg & 0xFF))
+#define MH_MSG_PH_READY		0x1
+#define MH_MSG_VH_READY		0x2
+
+#define ALLOW_INQUIRY		BIT(25)
+#define ALLOW_MODE_SELECT	BIT(24)
+#define ALLOW_MODE_SENSE	BIT(23)
+#define ALLOW_PRE_FETCH		GENMASK(22, 21)
+#define ALLOW_READ_CMD_ALL	GENMASK(20, 18)	/* read_6/10/16 */
+#define ALLOW_READ_BUFFER	BIT(17)
+#define ALLOW_READ_CAPACITY	GENMASK(16, 15)
+#define ALLOW_REPORT_LUNS	BIT(14)
+#define ALLOW_REQUEST_SENSE	BIT(13)
+#define ALLOW_SYNCHRONIZE_CACHE	GENMASK(8, 7)
+#define ALLOW_TEST_UNIT_READY	BIT(6)
+#define ALLOW_UNMAP		BIT(5)
+#define ALLOW_VERIFY		BIT(4)
+#define ALLOW_WRITE_CMD_ALL	GENMASK(3, 1)	/* write_6/10/16 */
+
+#define ALLOW_TRANS_VH_DEFAULT	(ALLOW_INQUIRY | ALLOW_MODE_SELECT | \
+				 ALLOW_MODE_SENSE | ALLOW_PRE_FETCH | \
+				 ALLOW_READ_CMD_ALL | ALLOW_READ_BUFFER | \
+				 ALLOW_READ_CAPACITY | ALLOW_REPORT_LUNS | \
+				 ALLOW_REQUEST_SENSE | ALLOW_SYNCHRONIZE_CACHE | \
+				 ALLOW_TEST_UNIT_READY | ALLOW_UNMAP | \
+				 ALLOW_VERIFY | ALLOW_WRITE_CMD_ALL)
+
+#define HCI_MH_ALLOWABLE_TRAN_OF_VH		0x30C
+#define HCI_MH_IID_IN_TASK_TAG			0X308
+
+#define PH_READY_TIMEOUT_MS			(5 * MSEC_PER_SEC)
+
 enum {
 	UNIPRO_L1_5 = 0,/* PHY Adapter */
 	UNIPRO_L2,	/* Data Link */
@@ -149,6 +198,117 @@ static int exynos7_ufs_drv_init(struct device *dev, struct exynos_ufs *ufs)
 	return 0;
 }
 
+static int exynosauto_ufs_drv_init(struct device *dev, struct exynos_ufs *ufs)
+{
+	struct exynos_ufs_uic_attr *attr = ufs->drv_data->uic_attr;
+
+	/* IO Coherency setting */
+	if (ufs->sysreg) {
+		return regmap_update_bits(ufs->sysreg,
+					  ufs->shareability_reg_offset,
+					  UFS_SHARABLE, UFS_SHARABLE);
+	}
+
+	attr->tx_dif_p_nsec = 3200000;
+
+	return 0;
+}
+
+static int exynosauto_ufs_post_hce_enable(struct exynos_ufs *ufs)
+{
+	struct ufs_hba *hba = ufs->hba;
+
+	/* Enable Virtual Host #1 */
+	ufshcd_rmwl(hba, MHCTRL_EN_VH_MASK, MHCTRL_EN_VH(1), MHCTRL);
+	/* Default VH Transfer permissions */
+	hci_writel(ufs, ALLOW_TRANS_VH_DEFAULT, HCI_MH_ALLOWABLE_TRAN_OF_VH);
+	/* IID information is replaced in TASKTAG[7:5] instead of IID in UCD */
+	hci_writel(ufs, 0x1, HCI_MH_IID_IN_TASK_TAG);
+
+	return 0;
+}
+
+static int exynosauto_ufs_pre_link(struct exynos_ufs *ufs)
+{
+	struct ufs_hba *hba = ufs->hba;
+	int i;
+	u32 tx_line_reset_period, rx_line_reset_period;
+
+	rx_line_reset_period = (RX_LINE_RESET_TIME * ufs->mclk_rate) / NSEC_PER_MSEC;
+	tx_line_reset_period = (TX_LINE_RESET_TIME * ufs->mclk_rate) / NSEC_PER_MSEC;
+
+	ufshcd_dme_set(hba, UIC_ARG_MIB(0x200), 0x40);
+	for_each_ufs_rx_lane(ufs, i) {
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(VND_RX_CLK_PRD, i),
+			       DIV_ROUND_UP(NSEC_PER_SEC, ufs->mclk_rate));
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(VND_RX_CLK_PRD_EN, i), 0x0);
+
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(VND_RX_LINERESET_VALUE2, i),
+			       (rx_line_reset_period >> 16) & 0xFF);
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(VND_RX_LINERESET_VALUE1, i),
+			       (rx_line_reset_period >> 8) & 0xFF);
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(VND_RX_LINERESET_VALUE0, i),
+			       (rx_line_reset_period) & 0xFF);
+
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(0x2f, i), 0x79);
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(0x84, i), 0x1);
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(0x25, i), 0xf6);
+	}
+
+	for_each_ufs_tx_lane(ufs, i) {
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(VND_TX_CLK_PRD, i),
+			       DIV_ROUND_UP(NSEC_PER_SEC, ufs->mclk_rate));
+		/* Not to affect VND_TX_LINERESET_PVALUE to VND_TX_CLK_PRD */
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(VND_TX_CLK_PRD_EN, i),
+			       0x02);
+
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(VND_TX_LINERESET_PVALUE2, i),
+			       (tx_line_reset_period >> 16) & 0xFF);
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(VND_TX_LINERESET_PVALUE1, i),
+			       (tx_line_reset_period >> 8) & 0xFF);
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(VND_TX_LINERESET_PVALUE0, i),
+			       (tx_line_reset_period) & 0xFF);
+
+		/* TX PWM Gear Capability / PWM_G1_ONLY */
+		ufshcd_dme_set(hba, UIC_ARG_MIB_SEL(0x04, i), 0x1);
+	}
+
+	ufshcd_dme_set(hba, UIC_ARG_MIB(0x200), 0x0);
+
+	ufshcd_dme_set(hba, UIC_ARG_MIB(PA_LOCAL_TX_LCC_ENABLE), 0x0);
+
+	ufshcd_dme_set(hba, UIC_ARG_MIB(0xa011), 0x8000);
+
+	return 0;
+}
+
+static int exynosauto_ufs_pre_pwr_change(struct exynos_ufs *ufs,
+					 struct ufs_pa_layer_attr *pwr)
+{
+	struct ufs_hba *hba = ufs->hba;
+
+	/* PACP_PWR_req and delivered to the remote DME */
+	ufshcd_dme_set(hba, UIC_ARG_MIB(PA_PWRMODEUSERDATA0), 12000);
+	ufshcd_dme_set(hba, UIC_ARG_MIB(PA_PWRMODEUSERDATA1), 32000);
+	ufshcd_dme_set(hba, UIC_ARG_MIB(PA_PWRMODEUSERDATA2), 16000);
+
+	return 0;
+}
+
+static int exynosauto_ufs_post_pwr_change(struct exynos_ufs *ufs,
+					  struct ufs_pa_layer_attr *pwr)
+{
+	struct ufs_hba *hba = ufs->hba;
+	u32 enabled_vh;
+
+	enabled_vh = ufshcd_readl(hba, MHCTRL) & MHCTRL_EN_VH_MASK;
+
+	/* Send physical host ready message to virtual hosts */
+	ufshcd_writel(hba, MH_MSG(enabled_vh, MH_MSG_PH_READY), PH2VH_MBOX);
+
+	return 0;
+}
+
 static int exynos7_ufs_pre_link(struct exynos_ufs *ufs)
 {
 	struct ufs_hba *hba = ufs->hba;
@@ -793,6 +953,27 @@ static void exynos_ufs_config_intr(struct exynos_ufs *ufs, u32 errs, u8 index)
 	}
 }
 
+static int exynos_ufs_setup_clocks(struct ufs_hba *hba, bool on,
+				   enum ufs_notify_change_status status)
+{
+	struct exynos_ufs *ufs = ufshcd_get_variant(hba);
+
+	if (!ufs)
+		return 0;
+
+	if (on && status == PRE_CHANGE) {
+		if (ufs->opts & EXYNOS_UFS_OPT_BROKEN_AUTO_CLK_CTRL)
+			exynos_ufs_disable_auto_ctrl_hcc(ufs);
+		exynos_ufs_ungate_clks(ufs);
+	} else if (!on && status == POST_CHANGE) {
+		exynos_ufs_gate_clks(ufs);
+		if (ufs->opts & EXYNOS_UFS_OPT_BROKEN_AUTO_CLK_CTRL)
+			exynos_ufs_enable_auto_ctrl_hcc(ufs);
+	}
+
+	return 0;
+}
+
 static int exynos_ufs_pre_link(struct ufs_hba *hba)
 {
 	struct exynos_ufs *ufs = ufshcd_get_variant(hba);
@@ -808,8 +989,12 @@ static int exynos_ufs_pre_link(struct ufs_hba *hba)
 
 	/* m-phy */
 	exynos_ufs_phy_init(ufs);
-	exynos_ufs_config_phy_time_attr(ufs);
-	exynos_ufs_config_phy_cap_attr(ufs);
+	if (!(ufs->opts & EXYNOS_UFS_OPT_SKIP_CONFIG_PHY_ATTR)) {
+		exynos_ufs_config_phy_time_attr(ufs);
+		exynos_ufs_config_phy_cap_attr(ufs);
+	}
+
+	exynos_ufs_setup_clocks(hba, true, PRE_CHANGE);
 
 	if (ufs->drv_data->pre_link)
 		ufs->drv_data->pre_link(ufs);
@@ -893,17 +1078,10 @@ static int exynos_ufs_post_link(struct ufs_hba *hba)
 static int exynos_ufs_parse_dt(struct device *dev, struct exynos_ufs *ufs)
 {
 	struct device_node *np = dev->of_node;
-	struct exynos_ufs_drv_data *drv_data = &exynos_ufs_drvs;
 	struct exynos_ufs_uic_attr *attr;
 	int ret = 0;
 
-	while (drv_data->compatible) {
-		if (of_device_is_compatible(np, drv_data->compatible)) {
-			ufs->drv_data = drv_data;
-			break;
-		}
-		drv_data++;
-	}
+	ufs->drv_data = device_get_match_data(dev);
 
 	if (ufs->drv_data && ufs->drv_data->uic_attr) {
 		attr = ufs->drv_data->uic_attr;
@@ -913,6 +1091,17 @@ static int exynos_ufs_parse_dt(struct device *dev, struct exynos_ufs *ufs)
 		goto out;
 	}
 
+	ufs->sysreg = syscon_regmap_lookup_by_phandle(np, "samsung,sysreg");
+	if (IS_ERR(ufs->sysreg))
+		ufs->sysreg = NULL;
+	else {
+		if (of_property_read_u32_index(np, "samsung,sysreg", 1,
+					       &ufs->shareability_reg_offset)) {
+			dev_warn(dev, "can't get an offset from sysreg. Set to default value\n");
+			ufs->shareability_reg_offset = UFS_SHAREABILITY_OFFSET;
+		}
+	}
+
 	ufs->pclk_avail_min = PCLK_AVAIL_MIN;
 	ufs->pclk_avail_max = PCLK_AVAIL_MAX;
 
@@ -927,6 +1116,18 @@ static int exynos_ufs_parse_dt(struct device *dev, struct exynos_ufs *ufs)
 	return ret;
 }
 
+static inline void exynos_ufs_priv_init(struct ufs_hba *hba,
+					struct exynos_ufs *ufs)
+{
+	ufs->hba = hba;
+	ufs->opts = ufs->drv_data->opts;
+	ufs->rx_sel_idx = PA_MAXDATALANES;
+	if (ufs->opts & EXYNOS_UFS_OPT_BROKEN_RX_SEL_IDX)
+		ufs->rx_sel_idx = 0;
+	hba->priv = (void *)ufs;
+	hba->quirks = ufs->drv_data->quirks;
+}
+
 static int exynos_ufs_init(struct ufs_hba *hba)
 {
 	struct device *dev = hba->dev;
@@ -976,13 +1177,8 @@ static int exynos_ufs_init(struct ufs_hba *hba)
 	if (ret)
 		goto phy_off;
 
-	ufs->hba = hba;
-	ufs->opts = ufs->drv_data->opts;
-	ufs->rx_sel_idx = PA_MAXDATALANES;
-	if (ufs->opts & EXYNOS_UFS_OPT_BROKEN_RX_SEL_IDX)
-		ufs->rx_sel_idx = 0;
-	hba->priv = (void *)ufs;
-	hba->quirks = ufs->drv_data->quirks;
+	exynos_ufs_priv_init(hba, ufs);
+
 	if (ufs->drv_data->drv_init) {
 		ret = ufs->drv_data->drv_init(dev, ufs);
 		if (ret) {
@@ -1110,6 +1306,12 @@ static int exynos_ufs_hce_enable_notify(struct ufs_hba *hba,
 
 	switch (status) {
 	case PRE_CHANGE:
+		if (ufs->drv_data->pre_hce_enable) {
+			ret = ufs->drv_data->pre_hce_enable(ufs);
+			if (ret)
+				return ret;
+		}
+
 		ret = exynos_ufs_host_reset(hba);
 		if (ret)
 			return ret;
@@ -1119,6 +1321,10 @@ static int exynos_ufs_hce_enable_notify(struct ufs_hba *hba,
 		exynos_ufs_calc_pwm_clk_div(ufs);
 		if (!(ufs->opts & EXYNOS_UFS_OPT_BROKEN_AUTO_CLK_CTRL))
 			exynos_ufs_enable_auto_ctrl_hcc(ufs);
+
+		if (ufs->drv_data->post_hce_enable)
+			ret = ufs->drv_data->post_hce_enable(ufs);
+
 		break;
 	}
 
@@ -1202,12 +1408,77 @@ static int exynos_ufs_resume(struct ufs_hba *hba, enum ufs_pm_op pm_op)
 	return 0;
 }
 
+static int exynosauto_ufs_vh_link_startup_notify(struct ufs_hba *hba,
+						 enum ufs_notify_change_status status)
+{
+	if (status == POST_CHANGE) {
+		ufshcd_set_link_active(hba);
+		ufshcd_set_ufs_dev_active(hba);
+	}
+
+	return 0;
+}
+
+static int exynosauto_ufs_vh_wait_ph_ready(struct ufs_hba *hba)
+{
+	u32 mbox;
+	ktime_t start, stop;
+
+	start = ktime_get();
+	stop = ktime_add(start, ms_to_ktime(PH_READY_TIMEOUT_MS));
+
+	do {
+		mbox = ufshcd_readl(hba, PH2VH_MBOX);
+		/* TODO: Mailbox message protocols between the PH and VHs are
+		 * not implemented yet. This will be supported later
+		 */
+		if ((mbox & MH_MSG_MASK) == MH_MSG_PH_READY)
+			return 0;
+
+		usleep_range(40, 50);
+	} while (ktime_before(ktime_get(), stop));
+
+	return -ETIME;
+}
+
+static int exynosauto_ufs_vh_init(struct ufs_hba *hba)
+{
+	struct device *dev = hba->dev;
+	struct platform_device *pdev = to_platform_device(dev);
+	struct exynos_ufs *ufs;
+	int ret;
+
+	ufs = devm_kzalloc(dev, sizeof(*ufs), GFP_KERNEL);
+	if (!ufs)
+		return -ENOMEM;
+
+	/* exynos-specific hci */
+	ufs->reg_hci = devm_platform_ioremap_resource_byname(pdev, "vs_hci");
+	if (IS_ERR(ufs->reg_hci)) {
+		dev_err(dev, "cannot ioremap for hci vendor register\n");
+		return PTR_ERR(ufs->reg_hci);
+	}
+
+	ret = exynosauto_ufs_vh_wait_ph_ready(hba);
+	if (ret)
+		return ret;
+
+	ufs->drv_data = device_get_match_data(dev);
+	if (!ufs->drv_data)
+		return -ENODEV;
+
+	exynos_ufs_priv_init(hba, ufs);
+
+	return 0;
+}
+
 static struct ufs_hba_variant_ops ufs_hba_exynos_ops = {
 	.name				= "exynos_ufs",
 	.init				= exynos_ufs_init,
 	.hce_enable_notify		= exynos_ufs_hce_enable_notify,
 	.link_startup_notify		= exynos_ufs_link_startup_notify,
 	.pwr_change_notify		= exynos_ufs_pwr_change_notify,
+	.setup_clocks			= exynos_ufs_setup_clocks,
 	.setup_xfer_req			= exynos_ufs_specify_nexus_t_xfer_req,
 	.setup_task_mgmt		= exynos_ufs_specify_nexus_t_tm_req,
 	.hibern8_notify			= exynos_ufs_hibern8_notify,
@@ -1215,12 +1486,24 @@ static struct ufs_hba_variant_ops ufs_hba_exynos_ops = {
 	.resume				= exynos_ufs_resume,
 };
 
+static struct ufs_hba_variant_ops ufs_hba_exynosauto_vh_ops = {
+	.name				= "exynosauto_ufs_vh",
+	.init				= exynosauto_ufs_vh_init,
+	.link_startup_notify		= exynosauto_ufs_vh_link_startup_notify,
+};
+
 static int exynos_ufs_probe(struct platform_device *pdev)
 {
 	int err;
 	struct device *dev = &pdev->dev;
+	const struct ufs_hba_variant_ops *vops = &ufs_hba_exynos_ops;
+	const struct exynos_ufs_drv_data *drv_data =
+		device_get_match_data(dev);
 
-	err = ufshcd_pltfrm_init(pdev, &ufs_hba_exynos_ops);
+	if (drv_data && drv_data->vops)
+		vops = drv_data->vops;
+
+	err = ufshcd_pltfrm_init(pdev, vops);
 	if (err)
 		dev_err(dev, "ufshcd_pltfrm_init() failed %d\n", err);
 
@@ -1261,8 +1544,35 @@ static struct exynos_ufs_uic_attr exynos7_uic_attr = {
 	.pa_dbg_option_suite		= 0x30103,
 };
 
+static struct exynos_ufs_drv_data exynosauto_ufs_drvs = {
+	.uic_attr		= &exynos7_uic_attr,
+	.quirks			= UFSHCD_QUIRK_PRDT_BYTE_GRAN |
+				  UFSHCI_QUIRK_SKIP_RESET_INTR_AGGR |
+				  UFSHCD_QUIRK_BROKEN_OCS_FATAL_ERROR |
+				  UFSHCD_QUIRK_SKIP_DEF_UNIPRO_TIMEOUT_SETTING,
+	.opts			= EXYNOS_UFS_OPT_BROKEN_AUTO_CLK_CTRL |
+				  EXYNOS_UFS_OPT_SKIP_CONFIG_PHY_ATTR |
+				  EXYNOS_UFS_OPT_BROKEN_RX_SEL_IDX,
+	.drv_init		= exynosauto_ufs_drv_init,
+	.post_hce_enable	= exynosauto_ufs_post_hce_enable,
+	.pre_link		= exynosauto_ufs_pre_link,
+	.pre_pwr_change		= exynosauto_ufs_pre_pwr_change,
+	.post_pwr_change	= exynosauto_ufs_post_pwr_change,
+};
+
+static struct exynos_ufs_drv_data exynosauto_ufs_vh_drvs = {
+	.vops			= &ufs_hba_exynosauto_vh_ops,
+	.quirks			= UFSHCD_QUIRK_PRDT_BYTE_GRAN |
+				  UFSHCI_QUIRK_SKIP_RESET_INTR_AGGR |
+				  UFSHCD_QUIRK_BROKEN_OCS_FATAL_ERROR |
+				  UFSHCI_QUIRK_BROKEN_HCE |
+				  UFSHCD_QUIRK_BROKEN_UIC_CMD |
+				  UFSHCD_QUIRK_SKIP_PH_CONFIGURATION |
+				  UFSHCD_QUIRK_SKIP_DEF_UNIPRO_TIMEOUT_SETTING,
+	.opts			= EXYNOS_UFS_OPT_BROKEN_RX_SEL_IDX,
+};
+
 static struct exynos_ufs_drv_data exynos_ufs_drvs = {
-	.compatible		= "samsung,exynos7-ufs",
 	.uic_attr		= &exynos7_uic_attr,
 	.quirks			= UFSHCD_QUIRK_PRDT_BYTE_GRAN |
 				  UFSHCI_QUIRK_BROKEN_REQ_LIST_CLR |
@@ -1287,6 +1597,10 @@ static struct exynos_ufs_drv_data exynos_ufs_drvs = {
 static const struct of_device_id exynos_ufs_of_match[] = {
 	{ .compatible = "samsung,exynos7-ufs",
 	  .data	      = &exynos_ufs_drvs },
+	{ .compatible = "samsung,exynosautov9-ufs",
+	  .data	      = &exynosauto_ufs_drvs },
+	{ .compatible = "samsung,exynosautov9-ufs-vh",
+	  .data	      = &exynosauto_ufs_vh_drvs },
 	{},
 };
 
diff --git a/drivers/scsi/ufs/ufs-exynos.h b/drivers/scsi/ufs/ufs-exynos.h
index dadf4fd..1c33e54 100644
--- a/drivers/scsi/ufs/ufs-exynos.h
+++ b/drivers/scsi/ufs/ufs-exynos.h
@@ -56,6 +56,22 @@
 #define TX_GRAN_NVAL_10_08	0x0296
 #define TX_GRAN_NVAL_H(v)	(((v) >> 8) & 0x3)
 
+#define VND_TX_CLK_PRD		0xAA
+#define VND_TX_CLK_PRD_EN	0xA9
+#define VND_TX_LINERESET_PVALUE0	0xAD
+#define VND_TX_LINERESET_PVALUE1	0xAC
+#define VND_TX_LINERESET_PVALUE2	0xAB
+
+#define TX_LINE_RESET_TIME	3200
+
+#define VND_RX_CLK_PRD		0x12
+#define VND_RX_CLK_PRD_EN	0x11
+#define VND_RX_LINERESET_VALUE0	0x1D
+#define VND_RX_LINERESET_VALUE1	0x1C
+#define VND_RX_LINERESET_VALUE2	0x1B
+
+#define RX_LINE_RESET_TIME	1000
+
 #define RX_FILLER_ENABLE	0x0316
 #define RX_FILLER_EN		(1 << 1)
 #define RX_LINERESET_VAL	0x0317
@@ -99,7 +115,7 @@ struct exynos_ufs;
 #define PA_HIBERN8TIME_VAL	0x20
 
 #define PCLK_AVAIL_MIN	70000000
-#define PCLK_AVAIL_MAX	133000000
+#define PCLK_AVAIL_MAX	167000000
 
 struct exynos_ufs_uic_attr {
 	/* TX Attributes */
@@ -142,7 +158,7 @@ struct exynos_ufs_uic_attr {
 };
 
 struct exynos_ufs_drv_data {
-	char *compatible;
+	const struct ufs_hba_variant_ops *vops;
 	struct exynos_ufs_uic_attr *uic_attr;
 	unsigned int quirks;
 	unsigned int opts;
@@ -154,6 +170,8 @@ struct exynos_ufs_drv_data {
 				struct ufs_pa_layer_attr *pwr);
 	int (*post_pwr_change)(struct exynos_ufs *ufs,
 				struct ufs_pa_layer_attr *pwr);
+	int (*pre_hce_enable)(struct exynos_ufs *ufs);
+	int (*post_hce_enable)(struct exynos_ufs *ufs);
 };
 
 struct ufs_phy_time_cfg {
@@ -191,7 +209,9 @@ struct exynos_ufs {
 	struct ufs_pa_layer_attr dev_req_params;
 	struct ufs_phy_time_cfg t_cfg;
 	ktime_t entry_hibern8_t;
-	struct exynos_ufs_drv_data *drv_data;
+	const struct exynos_ufs_drv_data *drv_data;
+	struct regmap *sysreg;
+	u32 shareability_reg_offset;
 
 	u32 opts;
 #define EXYNOS_UFS_OPT_HAS_APB_CLK_CTRL		BIT(0)
@@ -199,6 +219,7 @@ struct exynos_ufs {
 #define EXYNOS_UFS_OPT_BROKEN_AUTO_CLK_CTRL	BIT(2)
 #define EXYNOS_UFS_OPT_BROKEN_RX_SEL_IDX	BIT(3)
 #define EXYNOS_UFS_OPT_USE_SW_HIBERN8_TIMER	BIT(4)
+#define EXYNOS_UFS_OPT_SKIP_CONFIG_PHY_ATTR	BIT(5)
 };
 
 #define for_each_ufs_rx_lane(ufs, i) \
diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index 5c6a58a..afd3814 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -132,6 +132,14 @@ enum {
 	UFSHCD_CAN_QUEUE	= 32,
 };
 
+static const char *const ufshcd_state_name[] = {
+	[UFSHCD_STATE_RESET]			= "reset",
+	[UFSHCD_STATE_OPERATIONAL]		= "operational",
+	[UFSHCD_STATE_ERROR]			= "error",
+	[UFSHCD_STATE_EH_SCHEDULED_FATAL]	= "eh_fatal",
+	[UFSHCD_STATE_EH_SCHEDULED_NON_FATAL]	= "eh_non_fatal",
+};
+
 /* UFSHCD error handling flags */
 enum {
 	UFSHCD_EH_IN_PROGRESS = (1 << 0),
@@ -236,7 +244,6 @@ static int ufshcd_scale_clks(struct ufs_hba *hba, bool scale_up);
 static irqreturn_t ufshcd_intr(int irq, void *__hba);
 static int ufshcd_change_power_mode(struct ufs_hba *hba,
 			     struct ufs_pa_layer_attr *pwr_mode);
-static void ufshcd_schedule_eh_work(struct ufs_hba *hba);
 static int ufshcd_setup_hba_vreg(struct ufs_hba *hba, bool on);
 static int ufshcd_setup_vreg(struct ufs_hba *hba, bool on);
 static inline int ufshcd_config_vreg_hpm(struct ufs_hba *hba,
@@ -711,7 +718,7 @@ static inline bool ufshcd_is_device_present(struct ufs_hba *hba)
  * This function is used to get the OCS field from UTRD
  * Returns the OCS field in the UTRD
  */
-static inline int ufshcd_get_tr_ocs(struct ufshcd_lrb *lrbp)
+static enum utp_ocs ufshcd_get_tr_ocs(struct ufshcd_lrb *lrbp)
 {
 	return le32_to_cpu(lrbp->utr_descriptor_ptr->header.dword_2) & MASK_OCS;
 }
@@ -2323,6 +2330,9 @@ int ufshcd_send_uic_cmd(struct ufs_hba *hba, struct uic_command *uic_cmd)
 	int ret;
 	unsigned long flags;
 
+	if (hba->quirks & UFSHCD_QUIRK_BROKEN_UIC_CMD)
+		return 0;
+
 	ufshcd_hold(hba, false);
 	mutex_lock(&hba->uic_cmd_mutex);
 	ufshcd_add_delay_before_dme_cmd(hba);
@@ -2367,17 +2377,24 @@ static int ufshcd_map_sg(struct ufs_hba *hba, struct ufshcd_lrb *lrbp)
 					sizeof(struct ufshcd_sg_entry)));
 		else
 			lrbp->utr_descriptor_ptr->prd_table_length =
-				cpu_to_le16((u16) (sg_segments));
+				cpu_to_le16(sg_segments);
 
-		prd_table = (struct ufshcd_sg_entry *)lrbp->ucd_prdt_ptr;
+		prd_table = lrbp->ucd_prdt_ptr;
 
 		scsi_for_each_sg(cmd, sg, sg_segments, i) {
-			prd_table[i].size  =
-				cpu_to_le32(((u32) sg_dma_len(sg))-1);
-			prd_table[i].base_addr =
-				cpu_to_le32(lower_32_bits(sg->dma_address));
-			prd_table[i].upper_addr =
-				cpu_to_le32(upper_32_bits(sg->dma_address));
+			const unsigned int len = sg_dma_len(sg);
+
+			/*
+			 * From the UFSHCI spec: "Data Byte Count (DBC): A '0'
+			 * based value that indicates the length, in bytes, of
+			 * the data block. A maximum of length of 256KB may
+			 * exist for any entry. Bits 1:0 of this field shall be
+			 * 11b to indicate Dword granularity. A value of '3'
+			 * indicates 4 bytes, '7' indicates 8 bytes, etc."
+			 */
+			WARN_ONCE(len > 256 * 1024, "len = %#x\n", len);
+			prd_table[i].size = cpu_to_le32(len - 1);
+			prd_table[i].addr = cpu_to_le64(sg->dma_address);
 			prd_table[i].reserved = 0;
 		}
 	} else {
@@ -2661,7 +2678,7 @@ static void ufshcd_init_lrb(struct ufs_hba *hba, struct ufshcd_lrb *lrb, int i)
 	lrb->ucd_req_dma_addr = cmd_desc_element_addr;
 	lrb->ucd_rsp_ptr = (struct utp_upiu_rsp *)cmd_descp[i].response_upiu;
 	lrb->ucd_rsp_dma_addr = cmd_desc_element_addr + response_offset;
-	lrb->ucd_prdt_ptr = (struct ufshcd_sg_entry *)cmd_descp[i].prd_table;
+	lrb->ucd_prdt_ptr = cmd_descp[i].prd_table;
 	lrb->ucd_prdt_dma_addr = cmd_desc_element_addr + prdt_offset;
 }
 
@@ -5084,7 +5101,7 @@ ufshcd_transfer_rsp_status(struct ufs_hba *hba, struct ufshcd_lrb *lrbp)
 {
 	int result = 0;
 	int scsi_status;
-	int ocs;
+	enum utp_ocs ocs;
 
 	/* overall command status of utrd */
 	ocs = ufshcd_get_tr_ocs(lrbp);
@@ -5243,11 +5260,9 @@ static irqreturn_t ufshcd_uic_cmd_compl(struct ufs_hba *hba, u32 intr_status)
  * __ufshcd_transfer_req_compl - handle SCSI and query command completion
  * @hba: per adapter instance
  * @completed_reqs: bitmask that indicates which requests to complete
- * @retry_requests: whether to ask the SCSI core to retry completed requests
  */
 static void __ufshcd_transfer_req_compl(struct ufs_hba *hba,
-					unsigned long completed_reqs,
-					bool retry_requests)
+					unsigned long completed_reqs)
 {
 	struct ufshcd_lrb *lrbp;
 	struct scsi_cmnd *cmd;
@@ -5263,8 +5278,7 @@ static void __ufshcd_transfer_req_compl(struct ufs_hba *hba,
 			if (unlikely(ufshcd_should_inform_monitor(hba, lrbp)))
 				ufshcd_update_monitor(hba, lrbp);
 			ufshcd_add_command_trace(hba, index, UFS_CMD_COMP);
-			result = retry_requests ? DID_BUS_BUSY << 16 :
-				ufshcd_transfer_rsp_status(hba, lrbp);
+			result = ufshcd_transfer_rsp_status(hba, lrbp);
 			scsi_dma_unmap(cmd);
 			cmd->result = result;
 			/* Mark completed command as NULL in LRB */
@@ -5290,14 +5304,12 @@ static void __ufshcd_transfer_req_compl(struct ufs_hba *hba,
 /**
  * ufshcd_transfer_req_compl - handle SCSI and query command completion
  * @hba: per adapter instance
- * @retry_requests: whether or not to ask to retry requests
  *
  * Returns
  *  IRQ_HANDLED - If interrupt is valid
  *  IRQ_NONE    - If invalid interrupt
  */
-static irqreturn_t ufshcd_transfer_req_compl(struct ufs_hba *hba,
-					     bool retry_requests)
+static irqreturn_t ufshcd_transfer_req_compl(struct ufs_hba *hba)
 {
 	unsigned long completed_reqs, flags;
 	u32 tr_doorbell;
@@ -5326,8 +5338,7 @@ static irqreturn_t ufshcd_transfer_req_compl(struct ufs_hba *hba,
 	spin_unlock_irqrestore(&hba->outstanding_lock, flags);
 
 	if (completed_reqs) {
-		__ufshcd_transfer_req_compl(hba, completed_reqs,
-					    retry_requests);
+		__ufshcd_transfer_req_compl(hba, completed_reqs);
 		return IRQ_HANDLED;
 	} else {
 		return IRQ_NONE;
@@ -5826,13 +5837,7 @@ static void ufshcd_exception_event_handler(struct work_struct *work)
 /* Complete requests that have door-bell cleared */
 static void ufshcd_complete_requests(struct ufs_hba *hba)
 {
-	ufshcd_transfer_req_compl(hba, /*retry_requests=*/false);
-	ufshcd_tmc_handler(hba);
-}
-
-static void ufshcd_retry_aborted_requests(struct ufs_hba *hba)
-{
-	ufshcd_transfer_req_compl(hba, /*retry_requests=*/true);
+	ufshcd_transfer_req_compl(hba);
 	ufshcd_tmc_handler(hba);
 }
 
@@ -5914,9 +5919,10 @@ static inline bool ufshcd_is_saved_err_fatal(struct ufs_hba *hba)
 	       (hba->saved_err & (INT_FATAL_ERRORS | UFSHCD_UIC_HIBERN8_MASK));
 }
 
-/* host lock must be held before calling this func */
-static inline void ufshcd_schedule_eh_work(struct ufs_hba *hba)
+void ufshcd_schedule_eh_work(struct ufs_hba *hba)
 {
+	lockdep_assert_held(hba->host->host_lock);
+
 	/* handle fatal errors only when link is not in error state */
 	if (hba->ufshcd_state != UFSHCD_STATE_ERROR) {
 		if (hba->force_reset || ufshcd_is_link_broken(hba) ||
@@ -6076,6 +6082,13 @@ static void ufshcd_err_handler(struct work_struct *work)
 
 	hba = container_of(work, struct ufs_hba, eh_work);
 
+	dev_info(hba->dev,
+		 "%s started; HBA state %s; powered %d; shutting down %d; saved_err = %d; saved_uic_err = %d; force_reset = %d%s\n",
+		 __func__, ufshcd_state_name[hba->ufshcd_state],
+		 hba->is_powered, hba->shutting_down, hba->saved_err,
+		 hba->saved_uic_err, hba->force_reset,
+		 ufshcd_is_link_broken(hba) ? "; link is broken" : "");
+
 	down(&hba->host_sem);
 	spin_lock_irqsave(hba->host->host_lock, flags);
 	if (ufshcd_err_handling_should_stop(hba)) {
@@ -6170,6 +6183,8 @@ static void ufshcd_err_handler(struct work_struct *work)
 			err_xfer = true;
 			goto lock_skip_pending_xfer_clear;
 		}
+		dev_err(hba->dev, "Aborted tag %d / CDB %#02x\n", tag,
+			hba->lrb[tag].cmd ? hba->lrb[tag].cmd->cmnd[0] : -1);
 	}
 
 	/* Clear pending task management requests */
@@ -6181,7 +6196,8 @@ static void ufshcd_err_handler(struct work_struct *work)
 	}
 
 lock_skip_pending_xfer_clear:
-	ufshcd_retry_aborted_requests(hba);
+	/* Complete the requests that are cleared by s/w */
+	ufshcd_complete_requests(hba);
 
 	spin_lock_irqsave(hba->host->host_lock, flags);
 	hba->silence_err_logs = false;
@@ -6249,6 +6265,9 @@ static void ufshcd_err_handler(struct work_struct *work)
 	spin_unlock_irqrestore(hba->host->host_lock, flags);
 	ufshcd_err_handling_unprepare(hba);
 	up(&hba->host_sem);
+
+	dev_info(hba->dev, "%s finished; HBA state %s\n", __func__,
+		 ufshcd_state_name[hba->ufshcd_state]);
 }
 
 /**
@@ -6473,7 +6492,7 @@ static irqreturn_t ufshcd_sl_intr(struct ufs_hba *hba, u32 intr_status)
 		retval |= ufshcd_tmc_handler(hba);
 
 	if (intr_status & UTP_TRANSFER_REQ_COMPL)
-		retval |= ufshcd_transfer_req_compl(hba, /*retry_requests=*/false);
+		retval |= ufshcd_transfer_req_compl(hba);
 
 	return retval;
 }
@@ -6545,6 +6564,10 @@ static int ufshcd_clear_tm_cmd(struct ufs_hba *hba, int tag)
 	err = ufshcd_wait_for_register(hba,
 			REG_UTP_TASK_REQ_DOOR_BELL,
 			mask, 0, 1000, 1000);
+
+	dev_err(hba->dev, "Clearing task management function with tag %d %s\n",
+		tag, err ? "succeeded" : "failed");
+
 out:
 	return err;
 }
@@ -6637,7 +6660,8 @@ static int ufshcd_issue_tm_cmd(struct ufs_hba *hba, int lun_id, int task_id,
 		u8 tm_function, u8 *tm_response)
 {
 	struct utp_task_req_desc treq = { { 0 }, };
-	int ocs_value, err;
+	enum utp_ocs ocs_value;
+	int err;
 
 	/* Configure task request descriptor */
 	treq.header.dword_0 = cpu_to_le32(UTP_REQ_DESC_INT_CMD);
@@ -6815,7 +6839,7 @@ int ufshcd_exec_raw_upiu_cmd(struct ufs_hba *hba,
 	int err;
 	enum dev_cmd_type cmd_type = DEV_CMD_TYPE_QUERY;
 	struct utp_task_req_desc treq = { { 0 }, };
-	int ocs_value;
+	enum utp_ocs ocs_value;
 	u8 tm_f = be32_to_cpu(req_upiu->header.dword_1) >> 16 & MASK_TM_FUNC;
 
 	switch (msgcode) {
@@ -6893,7 +6917,7 @@ static int ufshcd_eh_device_reset_handler(struct scsi_cmnd *cmd)
 			err = ufshcd_clear_cmd(hba, pos);
 			if (err)
 				break;
-			__ufshcd_transfer_req_compl(hba, 1U << pos, false);
+			__ufshcd_transfer_req_compl(hba, 1U << pos);
 		}
 	}
 
@@ -7055,7 +7079,7 @@ static int ufshcd_abort(struct scsi_cmnd *cmd)
 		dev_err(hba->dev,
 		"%s: cmd was completed, but without a notifying intr, tag = %d",
 		__func__, tag);
-		__ufshcd_transfer_req_compl(hba, 1UL << tag, /*retry_requests=*/false);
+		__ufshcd_transfer_req_compl(hba, 1UL << tag);
 		goto release;
 	}
 
@@ -7121,7 +7145,7 @@ static int ufshcd_host_reset_and_restore(struct ufs_hba *hba)
 	ufshpb_reset_host(hba);
 	ufshcd_hba_stop(hba);
 	hba->silence_err_logs = true;
-	ufshcd_retry_aborted_requests(hba);
+	ufshcd_complete_requests(hba);
 	hba->silence_err_logs = false;
 
 	/* scale up clocks to max frequency before full reinitialization */
@@ -8002,6 +8026,9 @@ static int ufshcd_probe_hba(struct ufs_hba *hba, bool init_dev_params)
 	if (ret)
 		goto out;
 
+	if (hba->quirks & UFSHCD_QUIRK_SKIP_PH_CONFIGURATION)
+		goto out;
+
 	/* Debug counters initialization */
 	ufshcd_clear_dbg_ufs_stats(hba);
 
@@ -9772,6 +9799,11 @@ static int __init ufshcd_core_init(void)
 {
 	int ret;
 
+	/* Verify that there are no gaps in struct utp_transfer_cmd_desc. */
+	static_assert(sizeof(struct utp_transfer_cmd_desc) ==
+		      2 * ALIGNED_UPIU_SIZE +
+			      SG_ALL * sizeof(struct ufshcd_sg_entry));
+
 	ufs_debugfs_init();
 
 	ret = scsi_register_driver(&ufs_dev_wlun_template.gendrv);
diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
index dd76586..54750d7 100644
--- a/drivers/scsi/ufs/ufshcd.h
+++ b/drivers/scsi/ufs/ufshcd.h
@@ -589,6 +589,18 @@ enum ufshcd_quirks {
 	 * This quirk allows only sg entries aligned with page size.
 	 */
 	UFSHCD_QUIRK_ALIGN_SG_WITH_PAGE_SIZE		= 1 << 14,
+
+	/*
+	 * This quirk needs to be enabled if the host controller does not
+	 * support UIC command
+	 */
+	UFSHCD_QUIRK_BROKEN_UIC_CMD			= 1 << 15,
+
+	/*
+	 * This quirk needs to be enabled if the host controller cannot
+	 * support physical host configuration.
+	 */
+	UFSHCD_QUIRK_SKIP_PH_CONFIGURATION		= 1 << 16,
 };
 
 enum ufshcd_caps {
@@ -1023,6 +1035,7 @@ int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32 mask,
 void ufshcd_parse_dev_ref_clk_freq(struct ufs_hba *hba, struct clk *refclk);
 void ufshcd_update_evt_hist(struct ufs_hba *hba, u32 id, u32 val);
 void ufshcd_hba_stop(struct ufs_hba *hba);
+void ufshcd_schedule_eh_work(struct ufs_hba *hba);
 
 static inline void check_upiu_size(void)
 {
diff --git a/drivers/scsi/ufs/ufshci.h b/drivers/scsi/ufs/ufshci.h
index de95be5..6a295c8 100644
--- a/drivers/scsi/ufs/ufshci.h
+++ b/drivers/scsi/ufs/ufshci.h
@@ -389,7 +389,7 @@ enum {
 };
 
 /* Overall command status values */
-enum {
+enum utp_ocs {
 	OCS_SUCCESS			= 0x0,
 	OCS_INVALID_CMD_TABLE_ATTR	= 0x1,
 	OCS_INVALID_PRDT_ATTR		= 0x2,
@@ -402,6 +402,9 @@ enum {
 	OCS_INVALID_CRYPTO_CONFIG	= 0x9,
 	OCS_GENERAL_CRYPTO_ERROR	= 0xA,
 	OCS_INVALID_COMMAND_STATUS	= 0x0F,
+};
+
+enum {
 	MASK_OCS			= 0x0F,
 };
 
@@ -412,20 +415,18 @@ enum {
 
 /**
  * struct ufshcd_sg_entry - UFSHCI PRD Entry
- * @base_addr: Lower 32bit physical address DW-0
- * @upper_addr: Upper 32bit physical address DW-1
+ * @addr: Physical address; DW-0 and DW-1.
  * @reserved: Reserved for future use DW-2
  * @size: size of physical segment DW-3
  */
 struct ufshcd_sg_entry {
-	__le32    base_addr;
-	__le32    upper_addr;
+	__le64    addr;
 	__le32    reserved;
 	__le32    size;
 };
 
 /**
- * struct utp_transfer_cmd_desc - UFS Command Descriptor structure
+ * struct utp_transfer_cmd_desc - UTP Command Descriptor (UCD)
  * @command_upiu: Command UPIU Frame address
  * @response_upiu: Response UPIU Frame address
  * @prd_table: Physical Region Descriptor
@@ -451,7 +452,7 @@ struct request_desc_header {
 };
 
 /**
- * struct utp_transfer_req_desc - UTRD structure
+ * struct utp_transfer_req_desc - UTP Transfer Request Descriptor (UTRD)
  * @header: UTRD header DW-0 to DW-3
  * @command_desc_base_addr_lo: UCD base address low DW-4
  * @command_desc_base_addr_hi: UCD base address high DW-5
diff --git a/drivers/scsi/ufs/ufshpb.c b/drivers/scsi/ufs/ufshpb.c
index f4eca441..2e31e14 100644
--- a/drivers/scsi/ufs/ufshpb.c
+++ b/drivers/scsi/ufs/ufshpb.c
@@ -394,8 +394,6 @@ int ufshpb_prep(struct ufs_hba *hba, struct ufshcd_lrb *lrbp)
 	if (!ufshpb_is_supported_chunk(hpb, transfer_len))
 		return 0;
 
-	WARN_ON_ONCE(transfer_len > HPB_MULTI_CHUNK_HIGH);
-
 	if (hpb->is_hcm) {
 		/*
 		 * in host control mode, reads are the main source for
@@ -1572,7 +1570,7 @@ static void ufshpb_lu_parameter_init(struct ufs_hba *hba,
 	if (ufshpb_is_legacy(hba))
 		hpb->pre_req_max_tr_len = HPB_LEGACY_CHUNK_HIGH;
 	else
-		hpb->pre_req_max_tr_len = HPB_MULTI_CHUNK_HIGH;
+		hpb->pre_req_max_tr_len = hpb_dev_info->max_hpb_single_cmd;
 
 	hpb->lu_pinned_start = hpb_lu_info->pinned_start;
 	hpb->lu_pinned_end = hpb_lu_info->num_pinned ?
@@ -2582,7 +2580,7 @@ void ufshpb_get_dev_info(struct ufs_hba *hba, u8 *desc_buf)
 {
 	struct ufshpb_dev_info *hpb_dev_info = &hba->ufshpb_dev;
 	int version, ret;
-	u32 max_hpb_single_cmd = HPB_MULTI_CHUNK_LOW;
+	int max_single_cmd;
 
 	hpb_dev_info->control_mode = desc_buf[DEVICE_DESC_PARAM_HPB_CONTROL];
 
@@ -2598,18 +2596,22 @@ void ufshpb_get_dev_info(struct ufs_hba *hba, u8 *desc_buf)
 	if (version == HPB_SUPPORT_LEGACY_VERSION)
 		hpb_dev_info->is_legacy = true;
 
-	ret = ufshcd_query_attr_retry(hba, UPIU_QUERY_OPCODE_READ_ATTR,
-		QUERY_ATTR_IDN_MAX_HPB_SINGLE_CMD, 0, 0, &max_hpb_single_cmd);
-	if (ret)
-		dev_err(hba->dev, "%s: idn: read max size of single hpb cmd query request failed",
-			__func__);
-	hpb_dev_info->max_hpb_single_cmd = max_hpb_single_cmd;
-
 	/*
 	 * Get the number of user logical unit to check whether all
 	 * scsi_device finish initialization
 	 */
 	hpb_dev_info->num_lu = desc_buf[DEVICE_DESC_PARAM_NUM_LU];
+
+	if (hpb_dev_info->is_legacy)
+		return;
+
+	ret = ufshcd_query_attr_retry(hba, UPIU_QUERY_OPCODE_READ_ATTR,
+		QUERY_ATTR_IDN_MAX_HPB_SINGLE_CMD, 0, 0, &max_single_cmd);
+
+	if (ret)
+		hpb_dev_info->max_hpb_single_cmd = HPB_LEGACY_CHUNK_HIGH;
+	else
+		hpb_dev_info->max_hpb_single_cmd = min(max_single_cmd + 1, HPB_MULTI_CHUNK_HIGH);
 }
 
 void ufshpb_init(struct ufs_hba *hba)
diff --git a/drivers/scsi/ufs/ufshpb.h b/drivers/scsi/ufs/ufshpb.h
index f15d8fd..b475dbd7 100644
--- a/drivers/scsi/ufs/ufshpb.h
+++ b/drivers/scsi/ufs/ufshpb.h
@@ -31,7 +31,6 @@
 
 /* hpb support chunk size */
 #define HPB_LEGACY_CHUNK_HIGH			1
-#define HPB_MULTI_CHUNK_LOW			7
 #define HPB_MULTI_CHUNK_HIGH			255
 
 /* hpb vender defined opcode */
diff --git a/drivers/target/target_core_tmr.c b/drivers/target/target_core_tmr.c
index e7fcbc0..bac1114 100644
--- a/drivers/target/target_core_tmr.c
+++ b/drivers/target/target_core_tmr.c
@@ -50,15 +50,6 @@ EXPORT_SYMBOL(core_tmr_alloc_req);
 
 void core_tmr_release_req(struct se_tmr_req *tmr)
 {
-	struct se_device *dev = tmr->tmr_dev;
-	unsigned long flags;
-
-	if (dev) {
-		spin_lock_irqsave(&dev->se_tmr_lock, flags);
-		list_del_init(&tmr->tmr_list);
-		spin_unlock_irqrestore(&dev->se_tmr_lock, flags);
-	}
-
 	kfree(tmr);
 }
 
@@ -156,13 +147,6 @@ void core_tmr_abort_task(
 			se_cmd->state_active = false;
 			spin_unlock_irqrestore(&dev->queues[i].lock, flags);
 
-			/*
-			 * Ensure that this ABORT request is visible to the LU
-			 * RESET code.
-			 */
-			if (!tmr->tmr_dev)
-				WARN_ON_ONCE(transport_lookup_tmr_lun(tmr->task_cmd) < 0);
-
 			if (dev->transport->tmr_notify)
 				dev->transport->tmr_notify(dev, TMR_ABORT_TASK,
 							   &aborted_list);
@@ -234,6 +218,7 @@ static void core_tmr_drain_tmr_list(
 		}
 
 		list_move_tail(&tmr_p->tmr_list, &drain_tmr_list);
+		tmr_p->tmr_dev = NULL;
 	}
 	spin_unlock_irqrestore(&dev->se_tmr_lock, flags);
 
diff --git a/drivers/target/target_core_transport.c b/drivers/target/target_core_transport.c
index 4a2e749..7838dc2 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target/target_core_transport.c
@@ -676,6 +676,21 @@ static void target_remove_from_state_list(struct se_cmd *cmd)
 	spin_unlock_irqrestore(&dev->queues[cmd->cpuid].lock, flags);
 }
 
+static void target_remove_from_tmr_list(struct se_cmd *cmd)
+{
+	struct se_device *dev = NULL;
+	unsigned long flags;
+
+	if (cmd->se_cmd_flags & SCF_SCSI_TMR_CDB)
+		dev = cmd->se_tmr_req->tmr_dev;
+
+	if (dev) {
+		spin_lock_irqsave(&dev->se_tmr_lock, flags);
+		if (cmd->se_tmr_req->tmr_dev)
+			list_del_init(&cmd->se_tmr_req->tmr_list);
+		spin_unlock_irqrestore(&dev->se_tmr_lock, flags);
+	}
+}
 /*
  * This function is called by the target core after the target core has
  * finished processing a SCSI command or SCSI TMF. Both the regular command
@@ -687,13 +702,6 @@ static int transport_cmd_check_stop_to_fabric(struct se_cmd *cmd)
 {
 	unsigned long flags;
 
-	target_remove_from_state_list(cmd);
-
-	/*
-	 * Clear struct se_cmd->se_lun before the handoff to FE.
-	 */
-	cmd->se_lun = NULL;
-
 	spin_lock_irqsave(&cmd->t_state_lock, flags);
 	/*
 	 * Determine if frontend context caller is requesting the stopping of
@@ -728,8 +736,16 @@ static void transport_lun_remove_cmd(struct se_cmd *cmd)
 	if (!lun)
 		return;
 
+	target_remove_from_state_list(cmd);
+	target_remove_from_tmr_list(cmd);
+
 	if (cmpxchg(&cmd->lun_ref_active, true, false))
 		percpu_ref_put(&lun->lun_ref);
+
+	/*
+	 * Clear struct se_cmd->se_lun before the handoff to FE.
+	 */
+	cmd->se_lun = NULL;
 }
 
 static void target_complete_failure_work(struct work_struct *work)
diff --git a/drivers/thermal/intel/int340x_thermal/processor_thermal_mbox.c b/drivers/thermal/intel/int340x_thermal/processor_thermal_mbox.c
index a865219..01008ae 100644
--- a/drivers/thermal/intel/int340x_thermal/processor_thermal_mbox.c
+++ b/drivers/thermal/intel/int340x_thermal/processor_thermal_mbox.c
@@ -7,6 +7,7 @@
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/pci.h>
+#include <linux/io-64-nonatomic-lo-hi.h>
 #include "processor_thermal_device.h"
 
 #define MBOX_CMD_WORKLOAD_TYPE_READ	0x0E
diff --git a/drivers/video/fbdev/core/bitblit.c b/drivers/video/fbdev/core/bitblit.c
index f98e8f2..01fae2c 100644
--- a/drivers/video/fbdev/core/bitblit.c
+++ b/drivers/video/fbdev/core/bitblit.c
@@ -43,21 +43,6 @@ static void update_attr(u8 *dst, u8 *src, int attribute,
 	}
 }
 
-static void bit_bmove(struct vc_data *vc, struct fb_info *info, int sy,
-		      int sx, int dy, int dx, int height, int width)
-{
-	struct fb_copyarea area;
-
-	area.sx = sx * vc->vc_font.width;
-	area.sy = sy * vc->vc_font.height;
-	area.dx = dx * vc->vc_font.width;
-	area.dy = dy * vc->vc_font.height;
-	area.height = height * vc->vc_font.height;
-	area.width = width * vc->vc_font.width;
-
-	info->fbops->fb_copyarea(info, &area);
-}
-
 static void bit_clear(struct vc_data *vc, struct fb_info *info, int sy,
 		      int sx, int height, int width)
 {
@@ -393,7 +378,6 @@ static int bit_update_start(struct fb_info *info)
 
 void fbcon_set_bitops(struct fbcon_ops *ops)
 {
-	ops->bmove = bit_bmove;
 	ops->clear = bit_clear;
 	ops->putcs = bit_putcs;
 	ops->clear_margins = bit_clear_margins;
diff --git a/drivers/video/fbdev/core/fbcon.c b/drivers/video/fbdev/core/fbcon.c
index 22bb389..99ecd9a 100644
--- a/drivers/video/fbdev/core/fbcon.c
+++ b/drivers/video/fbdev/core/fbcon.c
@@ -173,8 +173,6 @@ static void fbcon_putcs(struct vc_data *vc, const unsigned short *s,
 			int count, int ypos, int xpos);
 static void fbcon_clear_margins(struct vc_data *vc, int bottom_only);
 static void fbcon_cursor(struct vc_data *vc, int mode);
-static void fbcon_bmove(struct vc_data *vc, int sy, int sx, int dy, int dx,
-			int height, int width);
 static int fbcon_switch(struct vc_data *vc);
 static int fbcon_blank(struct vc_data *vc, int blank, int mode_switch);
 static void fbcon_set_palette(struct vc_data *vc, const unsigned char *table);
@@ -182,16 +180,8 @@ static void fbcon_set_palette(struct vc_data *vc, const unsigned char *table);
 /*
  *  Internal routines
  */
-static __inline__ void ywrap_up(struct vc_data *vc, int count);
-static __inline__ void ywrap_down(struct vc_data *vc, int count);
-static __inline__ void ypan_up(struct vc_data *vc, int count);
-static __inline__ void ypan_down(struct vc_data *vc, int count);
-static void fbcon_bmove_rec(struct vc_data *vc, struct fbcon_display *p, int sy, int sx,
-			    int dy, int dx, int height, int width, u_int y_break);
 static void fbcon_set_disp(struct fb_info *info, struct fb_var_screeninfo *var,
 			   int unit);
-static void fbcon_redraw_move(struct vc_data *vc, struct fbcon_display *p,
-			      int line, int count, int dy);
 static void fbcon_modechanged(struct fb_info *info);
 static void fbcon_set_all_vcs(struct fb_info *info);
 static void fbcon_start(void);
@@ -1136,14 +1126,6 @@ static void fbcon_init(struct vc_data *vc, int init)
 	ops->graphics = 0;
 
 	/*
-	 * No more hw acceleration for fbcon.
-	 *
-	 * FIXME: Garbage collect all the now dead code after sufficient time
-	 * has passed.
-	 */
-	p->scrollmode = SCROLL_REDRAW;
-
-	/*
 	 *  ++guenther: console.c:vc_allocate() relies on initializing
 	 *  vc_{cols,rows}, but we must not set those if we are only
 	 *  resizing the console.
@@ -1229,14 +1211,13 @@ static void fbcon_deinit(struct vc_data *vc)
  *  This system is now divided into two levels because of complications
  *  caused by hardware scrolling. Top level functions:
  *
- *	fbcon_bmove(), fbcon_clear(), fbcon_putc(), fbcon_clear_margins()
+ *	fbcon_clear(), fbcon_putc(), fbcon_clear_margins()
  *
  *  handles y values in range [0, scr_height-1] that correspond to real
  *  screen positions. y_wrap shift means that first line of bitmap may be
  *  anywhere on this display. These functions convert lineoffsets to
  *  bitmap offsets and deal with the wrap-around case by splitting blits.
  *
- *	fbcon_bmove_physical_8()    -- These functions fast implementations
  *	fbcon_clear_physical_8()    -- of original fbcon_XXX fns.
  *	fbcon_putc_physical_8()	    -- (font width != 8) may be added later
  *
@@ -1409,224 +1390,6 @@ static void fbcon_set_disp(struct fb_info *info, struct fb_var_screeninfo *var,
 	}
 }
 
-static __inline__ void ywrap_up(struct vc_data *vc, int count)
-{
-	struct fb_info *info = registered_fb[con2fb_map[vc->vc_num]];
-	struct fbcon_ops *ops = info->fbcon_par;
-	struct fbcon_display *p = &fb_display[vc->vc_num];
-	
-	p->yscroll += count;
-	if (p->yscroll >= p->vrows)	/* Deal with wrap */
-		p->yscroll -= p->vrows;
-	ops->var.xoffset = 0;
-	ops->var.yoffset = p->yscroll * vc->vc_font.height;
-	ops->var.vmode |= FB_VMODE_YWRAP;
-	ops->update_start(info);
-	scrollback_max += count;
-	if (scrollback_max > scrollback_phys_max)
-		scrollback_max = scrollback_phys_max;
-	scrollback_current = 0;
-}
-
-static __inline__ void ywrap_down(struct vc_data *vc, int count)
-{
-	struct fb_info *info = registered_fb[con2fb_map[vc->vc_num]];
-	struct fbcon_ops *ops = info->fbcon_par;
-	struct fbcon_display *p = &fb_display[vc->vc_num];
-	
-	p->yscroll -= count;
-	if (p->yscroll < 0)	/* Deal with wrap */
-		p->yscroll += p->vrows;
-	ops->var.xoffset = 0;
-	ops->var.yoffset = p->yscroll * vc->vc_font.height;
-	ops->var.vmode |= FB_VMODE_YWRAP;
-	ops->update_start(info);
-	scrollback_max -= count;
-	if (scrollback_max < 0)
-		scrollback_max = 0;
-	scrollback_current = 0;
-}
-
-static __inline__ void ypan_up(struct vc_data *vc, int count)
-{
-	struct fb_info *info = registered_fb[con2fb_map[vc->vc_num]];
-	struct fbcon_display *p = &fb_display[vc->vc_num];
-	struct fbcon_ops *ops = info->fbcon_par;
-
-	p->yscroll += count;
-	if (p->yscroll > p->vrows - vc->vc_rows) {
-		ops->bmove(vc, info, p->vrows - vc->vc_rows,
-			    0, 0, 0, vc->vc_rows, vc->vc_cols);
-		p->yscroll -= p->vrows - vc->vc_rows;
-	}
-
-	ops->var.xoffset = 0;
-	ops->var.yoffset = p->yscroll * vc->vc_font.height;
-	ops->var.vmode &= ~FB_VMODE_YWRAP;
-	ops->update_start(info);
-	fbcon_clear_margins(vc, 1);
-	scrollback_max += count;
-	if (scrollback_max > scrollback_phys_max)
-		scrollback_max = scrollback_phys_max;
-	scrollback_current = 0;
-}
-
-static __inline__ void ypan_up_redraw(struct vc_data *vc, int t, int count)
-{
-	struct fb_info *info = registered_fb[con2fb_map[vc->vc_num]];
-	struct fbcon_ops *ops = info->fbcon_par;
-	struct fbcon_display *p = &fb_display[vc->vc_num];
-
-	p->yscroll += count;
-
-	if (p->yscroll > p->vrows - vc->vc_rows) {
-		p->yscroll -= p->vrows - vc->vc_rows;
-		fbcon_redraw_move(vc, p, t + count, vc->vc_rows - count, t);
-	}
-
-	ops->var.xoffset = 0;
-	ops->var.yoffset = p->yscroll * vc->vc_font.height;
-	ops->var.vmode &= ~FB_VMODE_YWRAP;
-	ops->update_start(info);
-	fbcon_clear_margins(vc, 1);
-	scrollback_max += count;
-	if (scrollback_max > scrollback_phys_max)
-		scrollback_max = scrollback_phys_max;
-	scrollback_current = 0;
-}
-
-static __inline__ void ypan_down(struct vc_data *vc, int count)
-{
-	struct fb_info *info = registered_fb[con2fb_map[vc->vc_num]];
-	struct fbcon_display *p = &fb_display[vc->vc_num];
-	struct fbcon_ops *ops = info->fbcon_par;
-	
-	p->yscroll -= count;
-	if (p->yscroll < 0) {
-		ops->bmove(vc, info, 0, 0, p->vrows - vc->vc_rows,
-			    0, vc->vc_rows, vc->vc_cols);
-		p->yscroll += p->vrows - vc->vc_rows;
-	}
-
-	ops->var.xoffset = 0;
-	ops->var.yoffset = p->yscroll * vc->vc_font.height;
-	ops->var.vmode &= ~FB_VMODE_YWRAP;
-	ops->update_start(info);
-	fbcon_clear_margins(vc, 1);
-	scrollback_max -= count;
-	if (scrollback_max < 0)
-		scrollback_max = 0;
-	scrollback_current = 0;
-}
-
-static __inline__ void ypan_down_redraw(struct vc_data *vc, int t, int count)
-{
-	struct fb_info *info = registered_fb[con2fb_map[vc->vc_num]];
-	struct fbcon_ops *ops = info->fbcon_par;
-	struct fbcon_display *p = &fb_display[vc->vc_num];
-
-	p->yscroll -= count;
-
-	if (p->yscroll < 0) {
-		p->yscroll += p->vrows - vc->vc_rows;
-		fbcon_redraw_move(vc, p, t, vc->vc_rows - count, t + count);
-	}
-
-	ops->var.xoffset = 0;
-	ops->var.yoffset = p->yscroll * vc->vc_font.height;
-	ops->var.vmode &= ~FB_VMODE_YWRAP;
-	ops->update_start(info);
-	fbcon_clear_margins(vc, 1);
-	scrollback_max -= count;
-	if (scrollback_max < 0)
-		scrollback_max = 0;
-	scrollback_current = 0;
-}
-
-static void fbcon_redraw_move(struct vc_data *vc, struct fbcon_display *p,
-			      int line, int count, int dy)
-{
-	unsigned short *s = (unsigned short *)
-		(vc->vc_origin + vc->vc_size_row * line);
-
-	while (count--) {
-		unsigned short *start = s;
-		unsigned short *le = advance_row(s, 1);
-		unsigned short c;
-		int x = 0;
-		unsigned short attr = 1;
-
-		do {
-			c = scr_readw(s);
-			if (attr != (c & 0xff00)) {
-				attr = c & 0xff00;
-				if (s > start) {
-					fbcon_putcs(vc, start, s - start,
-						    dy, x);
-					x += s - start;
-					start = s;
-				}
-			}
-			console_conditional_schedule();
-			s++;
-		} while (s < le);
-		if (s > start)
-			fbcon_putcs(vc, start, s - start, dy, x);
-		console_conditional_schedule();
-		dy++;
-	}
-}
-
-static void fbcon_redraw_blit(struct vc_data *vc, struct fb_info *info,
-			struct fbcon_display *p, int line, int count, int ycount)
-{
-	int offset = ycount * vc->vc_cols;
-	unsigned short *d = (unsigned short *)
-	    (vc->vc_origin + vc->vc_size_row * line);
-	unsigned short *s = d + offset;
-	struct fbcon_ops *ops = info->fbcon_par;
-
-	while (count--) {
-		unsigned short *start = s;
-		unsigned short *le = advance_row(s, 1);
-		unsigned short c;
-		int x = 0;
-
-		do {
-			c = scr_readw(s);
-
-			if (c == scr_readw(d)) {
-				if (s > start) {
-					ops->bmove(vc, info, line + ycount, x,
-						   line, x, 1, s-start);
-					x += s - start + 1;
-					start = s + 1;
-				} else {
-					x++;
-					start++;
-				}
-			}
-
-			scr_writew(c, d);
-			console_conditional_schedule();
-			s++;
-			d++;
-		} while (s < le);
-		if (s > start)
-			ops->bmove(vc, info, line + ycount, x, line, x, 1,
-				   s-start);
-		console_conditional_schedule();
-		if (ycount > 0)
-			line++;
-		else {
-			line--;
-			/* NOTE: We subtract two lines from these pointers */
-			s -= vc->vc_size_row;
-			d -= vc->vc_size_row;
-		}
-	}
-}
-
 static void fbcon_redraw(struct vc_data *vc, struct fbcon_display *p,
 			 int line, int count, int offset)
 {
@@ -1687,7 +1450,6 @@ static bool fbcon_scroll(struct vc_data *vc, unsigned int t, unsigned int b,
 {
 	struct fb_info *info = registered_fb[con2fb_map[vc->vc_num]];
 	struct fbcon_display *p = &fb_display[vc->vc_num];
-	int scroll_partial = info->flags & FBINFO_PARTIAL_PAN_OK;
 
 	if (fbcon_is_inactive(vc, info))
 		return true;
@@ -1704,249 +1466,32 @@ static bool fbcon_scroll(struct vc_data *vc, unsigned int t, unsigned int b,
 	case SM_UP:
 		if (count > vc->vc_rows)	/* Maximum realistic size */
 			count = vc->vc_rows;
-		if (logo_shown >= 0)
-			goto redraw_up;
-		switch (p->scrollmode) {
-		case SCROLL_MOVE:
-			fbcon_redraw_blit(vc, info, p, t, b - t - count,
-				     count);
-			fbcon_clear(vc, b - count, 0, count, vc->vc_cols);
-			scr_memsetw((unsigned short *) (vc->vc_origin +
-							vc->vc_size_row *
-							(b - count)),
-				    vc->vc_video_erase_char,
-				    vc->vc_size_row * count);
-			return true;
-
-		case SCROLL_WRAP_MOVE:
-			if (b - t - count > 3 * vc->vc_rows >> 2) {
-				if (t > 0)
-					fbcon_bmove(vc, 0, 0, count, 0, t,
-						    vc->vc_cols);
-				ywrap_up(vc, count);
-				if (vc->vc_rows - b > 0)
-					fbcon_bmove(vc, b - count, 0, b, 0,
-						    vc->vc_rows - b,
-						    vc->vc_cols);
-			} else if (info->flags & FBINFO_READS_FAST)
-				fbcon_bmove(vc, t + count, 0, t, 0,
-					    b - t - count, vc->vc_cols);
-			else
-				goto redraw_up;
-			fbcon_clear(vc, b - count, 0, count, vc->vc_cols);
-			break;
-
-		case SCROLL_PAN_REDRAW:
-			if ((p->yscroll + count <=
-			     2 * (p->vrows - vc->vc_rows))
-			    && ((!scroll_partial && (b - t == vc->vc_rows))
-				|| (scroll_partial
-				    && (b - t - count >
-					3 * vc->vc_rows >> 2)))) {
-				if (t > 0)
-					fbcon_redraw_move(vc, p, 0, t, count);
-				ypan_up_redraw(vc, t, count);
-				if (vc->vc_rows - b > 0)
-					fbcon_redraw_move(vc, p, b,
-							  vc->vc_rows - b, b);
-			} else
-				fbcon_redraw_move(vc, p, t + count, b - t - count, t);
-			fbcon_clear(vc, b - count, 0, count, vc->vc_cols);
-			break;
-
-		case SCROLL_PAN_MOVE:
-			if ((p->yscroll + count <=
-			     2 * (p->vrows - vc->vc_rows))
-			    && ((!scroll_partial && (b - t == vc->vc_rows))
-				|| (scroll_partial
-				    && (b - t - count >
-					3 * vc->vc_rows >> 2)))) {
-				if (t > 0)
-					fbcon_bmove(vc, 0, 0, count, 0, t,
-						    vc->vc_cols);
-				ypan_up(vc, count);
-				if (vc->vc_rows - b > 0)
-					fbcon_bmove(vc, b - count, 0, b, 0,
-						    vc->vc_rows - b,
-						    vc->vc_cols);
-			} else if (info->flags & FBINFO_READS_FAST)
-				fbcon_bmove(vc, t + count, 0, t, 0,
-					    b - t - count, vc->vc_cols);
-			else
-				goto redraw_up;
-			fbcon_clear(vc, b - count, 0, count, vc->vc_cols);
-			break;
-
-		case SCROLL_REDRAW:
-		      redraw_up:
-			fbcon_redraw(vc, p, t, b - t - count,
-				     count * vc->vc_cols);
-			fbcon_clear(vc, b - count, 0, count, vc->vc_cols);
-			scr_memsetw((unsigned short *) (vc->vc_origin +
-							vc->vc_size_row *
-							(b - count)),
-				    vc->vc_video_erase_char,
-				    vc->vc_size_row * count);
-			return true;
-		}
-		break;
+		fbcon_redraw(vc, p, t, b - t - count,
+			     count * vc->vc_cols);
+		fbcon_clear(vc, b - count, 0, count, vc->vc_cols);
+		scr_memsetw((unsigned short *) (vc->vc_origin +
+						vc->vc_size_row *
+						(b - count)),
+			    vc->vc_video_erase_char,
+			    vc->vc_size_row * count);
+		return true;
 
 	case SM_DOWN:
 		if (count > vc->vc_rows)	/* Maximum realistic size */
 			count = vc->vc_rows;
-		if (logo_shown >= 0)
-			goto redraw_down;
-		switch (p->scrollmode) {
-		case SCROLL_MOVE:
-			fbcon_redraw_blit(vc, info, p, b - 1, b - t - count,
-				     -count);
-			fbcon_clear(vc, t, 0, count, vc->vc_cols);
-			scr_memsetw((unsigned short *) (vc->vc_origin +
-							vc->vc_size_row *
-							t),
-				    vc->vc_video_erase_char,
-				    vc->vc_size_row * count);
-			return true;
-
-		case SCROLL_WRAP_MOVE:
-			if (b - t - count > 3 * vc->vc_rows >> 2) {
-				if (vc->vc_rows - b > 0)
-					fbcon_bmove(vc, b, 0, b - count, 0,
-						    vc->vc_rows - b,
-						    vc->vc_cols);
-				ywrap_down(vc, count);
-				if (t > 0)
-					fbcon_bmove(vc, count, 0, 0, 0, t,
-						    vc->vc_cols);
-			} else if (info->flags & FBINFO_READS_FAST)
-				fbcon_bmove(vc, t, 0, t + count, 0,
-					    b - t - count, vc->vc_cols);
-			else
-				goto redraw_down;
-			fbcon_clear(vc, t, 0, count, vc->vc_cols);
-			break;
-
-		case SCROLL_PAN_MOVE:
-			if ((count - p->yscroll <= p->vrows - vc->vc_rows)
-			    && ((!scroll_partial && (b - t == vc->vc_rows))
-				|| (scroll_partial
-				    && (b - t - count >
-					3 * vc->vc_rows >> 2)))) {
-				if (vc->vc_rows - b > 0)
-					fbcon_bmove(vc, b, 0, b - count, 0,
-						    vc->vc_rows - b,
-						    vc->vc_cols);
-				ypan_down(vc, count);
-				if (t > 0)
-					fbcon_bmove(vc, count, 0, 0, 0, t,
-						    vc->vc_cols);
-			} else if (info->flags & FBINFO_READS_FAST)
-				fbcon_bmove(vc, t, 0, t + count, 0,
-					    b - t - count, vc->vc_cols);
-			else
-				goto redraw_down;
-			fbcon_clear(vc, t, 0, count, vc->vc_cols);
-			break;
-
-		case SCROLL_PAN_REDRAW:
-			if ((count - p->yscroll <= p->vrows - vc->vc_rows)
-			    && ((!scroll_partial && (b - t == vc->vc_rows))
-				|| (scroll_partial
-				    && (b - t - count >
-					3 * vc->vc_rows >> 2)))) {
-				if (vc->vc_rows - b > 0)
-					fbcon_redraw_move(vc, p, b, vc->vc_rows - b,
-							  b - count);
-				ypan_down_redraw(vc, t, count);
-				if (t > 0)
-					fbcon_redraw_move(vc, p, count, t, 0);
-			} else
-				fbcon_redraw_move(vc, p, t, b - t - count, t + count);
-			fbcon_clear(vc, t, 0, count, vc->vc_cols);
-			break;
-
-		case SCROLL_REDRAW:
-		      redraw_down:
-			fbcon_redraw(vc, p, b - 1, b - t - count,
-				     -count * vc->vc_cols);
-			fbcon_clear(vc, t, 0, count, vc->vc_cols);
-			scr_memsetw((unsigned short *) (vc->vc_origin +
-							vc->vc_size_row *
-							t),
-				    vc->vc_video_erase_char,
-				    vc->vc_size_row * count);
-			return true;
-		}
+		fbcon_redraw(vc, p, b - 1, b - t - count,
+			     -count * vc->vc_cols);
+		fbcon_clear(vc, t, 0, count, vc->vc_cols);
+		scr_memsetw((unsigned short *) (vc->vc_origin +
+						vc->vc_size_row *
+						t),
+			    vc->vc_video_erase_char,
+			    vc->vc_size_row * count);
+		return true;
 	}
 	return false;
 }
 
-
-static void fbcon_bmove(struct vc_data *vc, int sy, int sx, int dy, int dx,
-			int height, int width)
-{
-	struct fb_info *info = registered_fb[con2fb_map[vc->vc_num]];
-	struct fbcon_display *p = &fb_display[vc->vc_num];
-	
-	if (fbcon_is_inactive(vc, info))
-		return;
-
-	if (!width || !height)
-		return;
-
-	/*  Split blits that cross physical y_wrap case.
-	 *  Pathological case involves 4 blits, better to use recursive
-	 *  code rather than unrolled case
-	 *
-	 *  Recursive invocations don't need to erase the cursor over and
-	 *  over again, so we use fbcon_bmove_rec()
-	 */
-	fbcon_bmove_rec(vc, p, sy, sx, dy, dx, height, width,
-			p->vrows - p->yscroll);
-}
-
-static void fbcon_bmove_rec(struct vc_data *vc, struct fbcon_display *p, int sy, int sx,
-			    int dy, int dx, int height, int width, u_int y_break)
-{
-	struct fb_info *info = registered_fb[con2fb_map[vc->vc_num]];
-	struct fbcon_ops *ops = info->fbcon_par;
-	u_int b;
-
-	if (sy < y_break && sy + height > y_break) {
-		b = y_break - sy;
-		if (dy < sy) {	/* Avoid trashing self */
-			fbcon_bmove_rec(vc, p, sy, sx, dy, dx, b, width,
-					y_break);
-			fbcon_bmove_rec(vc, p, sy + b, sx, dy + b, dx,
-					height - b, width, y_break);
-		} else {
-			fbcon_bmove_rec(vc, p, sy + b, sx, dy + b, dx,
-					height - b, width, y_break);
-			fbcon_bmove_rec(vc, p, sy, sx, dy, dx, b, width,
-					y_break);
-		}
-		return;
-	}
-
-	if (dy < y_break && dy + height > y_break) {
-		b = y_break - dy;
-		if (dy < sy) {	/* Avoid trashing self */
-			fbcon_bmove_rec(vc, p, sy, sx, dy, dx, b, width,
-					y_break);
-			fbcon_bmove_rec(vc, p, sy + b, sx, dy + b, dx,
-					height - b, width, y_break);
-		} else {
-			fbcon_bmove_rec(vc, p, sy + b, sx, dy + b, dx,
-					height - b, width, y_break);
-			fbcon_bmove_rec(vc, p, sy, sx, dy, dx, b, width,
-					y_break);
-		}
-		return;
-	}
-	ops->bmove(vc, info, real_y(p, sy), sx, real_y(p, dy), dx,
-		   height, width);
-}
-
 static void updatescrollmode(struct fbcon_display *p,
 					struct fb_info *info,
 					struct vc_data *vc)
@@ -2119,21 +1664,7 @@ static int fbcon_switch(struct vc_data *vc)
 
 	updatescrollmode(p, info, vc);
 
-	switch (p->scrollmode) {
-	case SCROLL_WRAP_MOVE:
-		scrollback_phys_max = p->vrows - vc->vc_rows;
-		break;
-	case SCROLL_PAN_MOVE:
-	case SCROLL_PAN_REDRAW:
-		scrollback_phys_max = p->vrows - 2 * vc->vc_rows;
-		if (scrollback_phys_max < 0)
-			scrollback_phys_max = 0;
-		break;
-	default:
-		scrollback_phys_max = 0;
-		break;
-	}
-
+	scrollback_phys_max = 0;
 	scrollback_max = 0;
 	scrollback_current = 0;
 
diff --git a/drivers/video/fbdev/core/fbcon.h b/drivers/video/fbdev/core/fbcon.h
index 9315b36..a00603b 100644
--- a/drivers/video/fbdev/core/fbcon.h
+++ b/drivers/video/fbdev/core/fbcon.h
@@ -29,7 +29,6 @@ struct fbcon_display {
     /* Filled in by the low-level console driver */
     const u_char *fontdata;
     int userfont;                   /* != 0 if fontdata kmalloc()ed */
-    u_short scrollmode;             /* Scroll Method */
     u_short inverse;                /* != 0 text black on white as default */
     short yscroll;                  /* Hardware scrolling */
     int vrows;                      /* number of virtual rows */
@@ -52,8 +51,6 @@ struct fbcon_display {
 };
 
 struct fbcon_ops {
-	void (*bmove)(struct vc_data *vc, struct fb_info *info, int sy,
-		      int sx, int dy, int dx, int height, int width);
 	void (*clear)(struct vc_data *vc, struct fb_info *info, int sy,
 		      int sx, int height, int width);
 	void (*putcs)(struct vc_data *vc, struct fb_info *info,
@@ -152,62 +149,6 @@ static inline int attr_col_ec(int shift, struct vc_data *vc,
 #define attr_bgcol_ec(bgshift, vc, info) attr_col_ec(bgshift, vc, info, 0)
 #define attr_fgcol_ec(fgshift, vc, info) attr_col_ec(fgshift, vc, info, 1)
 
-    /*
-     *  Scroll Method
-     */
-     
-/* There are several methods fbcon can use to move text around the screen:
- *
- *                     Operation   Pan    Wrap
- *---------------------------------------------
- * SCROLL_MOVE         copyarea    No     No
- * SCROLL_PAN_MOVE     copyarea    Yes    No
- * SCROLL_WRAP_MOVE    copyarea    No     Yes
- * SCROLL_REDRAW       imageblit   No     No
- * SCROLL_PAN_REDRAW   imageblit   Yes    No
- * SCROLL_WRAP_REDRAW  imageblit   No     Yes
- *
- * (SCROLL_WRAP_REDRAW is not implemented yet)
- *
- * In general, fbcon will choose the best scrolling
- * method based on the rule below:
- *
- * Pan/Wrap > accel imageblit > accel copyarea >
- * soft imageblit > (soft copyarea)
- *
- * Exception to the rule: Pan + accel copyarea is
- * preferred over Pan + accel imageblit.
- *
- * The above is typical for PCI/AGP cards. Unless
- * overridden, fbcon will never use soft copyarea.
- *
- * If you need to override the above rule, set the
- * appropriate flags in fb_info->flags.  For example,
- * to prefer copyarea over imageblit, set
- * FBINFO_READS_FAST.
- *
- * Other notes:
- * + use the hardware engine to move the text
- *    (hw-accelerated copyarea() and fillrect())
- * + use hardware-supported panning on a large virtual screen
- * + amifb can not only pan, but also wrap the display by N lines
- *    (i.e. visible line i = physical line (i+N) % yres).
- * + read what's already rendered on the screen and
- *     write it in a different place (this is cfb_copyarea())
- * + re-render the text to the screen
- *
- * Whether to use wrapping or panning can only be figured out at
- * runtime (when we know whether our font height is a multiple
- * of the pan/wrap step)
- *
- */
-
-#define SCROLL_MOVE	   0x001
-#define SCROLL_PAN_MOVE	   0x002
-#define SCROLL_WRAP_MOVE   0x003
-#define SCROLL_REDRAW	   0x004
-#define SCROLL_PAN_REDRAW  0x005
-
 #ifdef CONFIG_FB_TILEBLITTING
 extern void fbcon_set_tileops(struct vc_data *vc, struct fb_info *info);
 #endif
diff --git a/drivers/video/fbdev/core/fbcon_ccw.c b/drivers/video/fbdev/core/fbcon_ccw.c
index 9cd2c4b..ffa7893 100644
--- a/drivers/video/fbdev/core/fbcon_ccw.c
+++ b/drivers/video/fbdev/core/fbcon_ccw.c
@@ -59,31 +59,12 @@ static void ccw_update_attr(u8 *dst, u8 *src, int attribute,
 	}
 }
 
-
-static void ccw_bmove(struct vc_data *vc, struct fb_info *info, int sy,
-		     int sx, int dy, int dx, int height, int width)
-{
-	struct fbcon_ops *ops = info->fbcon_par;
-	struct fb_copyarea area;
-	u32 vyres = GETVYRES(ops->p->scrollmode, info);
-
-	area.sx = sy * vc->vc_font.height;
-	area.sy = vyres - ((sx + width) * vc->vc_font.width);
-	area.dx = dy * vc->vc_font.height;
-	area.dy = vyres - ((dx + width) * vc->vc_font.width);
-	area.width = height * vc->vc_font.height;
-	area.height  = width * vc->vc_font.width;
-
-	info->fbops->fb_copyarea(info, &area);
-}
-
 static void ccw_clear(struct vc_data *vc, struct fb_info *info, int sy,
 		     int sx, int height, int width)
 {
-	struct fbcon_ops *ops = info->fbcon_par;
 	struct fb_fillrect region;
 	int bgshift = (vc->vc_hi_font_mask) ? 13 : 12;
-	u32 vyres = GETVYRES(ops->p->scrollmode, info);
+	u32 vyres = info->var.yres;
 
 	region.color = attr_bgcol_ec(bgshift,vc,info);
 	region.dx = sy * vc->vc_font.height;
@@ -140,7 +121,7 @@ static void ccw_putcs(struct vc_data *vc, struct fb_info *info,
 	u32 cnt, pitch, size;
 	u32 attribute = get_attribute(info, scr_readw(s));
 	u8 *dst, *buf = NULL;
-	u32 vyres = GETVYRES(ops->p->scrollmode, info);
+	u32 vyres = info->var.yres;
 
 	if (!ops->fontbuffer)
 		return;
@@ -229,7 +210,7 @@ static void ccw_cursor(struct vc_data *vc, struct fb_info *info, int mode,
 	int attribute, use_sw = vc->vc_cursor_type & CUR_SW;
 	int err = 1, dx, dy;
 	char *src;
-	u32 vyres = GETVYRES(ops->p->scrollmode, info);
+	u32 vyres = info->var.yres;
 
 	if (!ops->fontbuffer)
 		return;
@@ -387,7 +368,7 @@ static int ccw_update_start(struct fb_info *info)
 {
 	struct fbcon_ops *ops = info->fbcon_par;
 	u32 yoffset;
-	u32 vyres = GETVYRES(ops->p->scrollmode, info);
+	u32 vyres = info->var.yres;
 	int err;
 
 	yoffset = (vyres - info->var.yres) - ops->var.xoffset;
@@ -402,7 +383,6 @@ static int ccw_update_start(struct fb_info *info)
 
 void fbcon_rotate_ccw(struct fbcon_ops *ops)
 {
-	ops->bmove = ccw_bmove;
 	ops->clear = ccw_clear;
 	ops->putcs = ccw_putcs;
 	ops->clear_margins = ccw_clear_margins;
diff --git a/drivers/video/fbdev/core/fbcon_cw.c b/drivers/video/fbdev/core/fbcon_cw.c
index 88d89fa..92e5b7f 100644
--- a/drivers/video/fbdev/core/fbcon_cw.c
+++ b/drivers/video/fbdev/core/fbcon_cw.c
@@ -44,31 +44,12 @@ static void cw_update_attr(u8 *dst, u8 *src, int attribute,
 	}
 }
 
-
-static void cw_bmove(struct vc_data *vc, struct fb_info *info, int sy,
-		     int sx, int dy, int dx, int height, int width)
-{
-	struct fbcon_ops *ops = info->fbcon_par;
-	struct fb_copyarea area;
-	u32 vxres = GETVXRES(ops->p->scrollmode, info);
-
-	area.sx = vxres - ((sy + height) * vc->vc_font.height);
-	area.sy = sx * vc->vc_font.width;
-	area.dx = vxres - ((dy + height) * vc->vc_font.height);
-	area.dy = dx * vc->vc_font.width;
-	area.width = height * vc->vc_font.height;
-	area.height  = width * vc->vc_font.width;
-
-	info->fbops->fb_copyarea(info, &area);
-}
-
 static void cw_clear(struct vc_data *vc, struct fb_info *info, int sy,
 		     int sx, int height, int width)
 {
-	struct fbcon_ops *ops = info->fbcon_par;
 	struct fb_fillrect region;
 	int bgshift = (vc->vc_hi_font_mask) ? 13 : 12;
-	u32 vxres = GETVXRES(ops->p->scrollmode, info);
+	u32 vxres = info->var.xres;
 
 	region.color = attr_bgcol_ec(bgshift,vc,info);
 	region.dx = vxres - ((sy + height) * vc->vc_font.height);
@@ -125,7 +106,7 @@ static void cw_putcs(struct vc_data *vc, struct fb_info *info,
 	u32 cnt, pitch, size;
 	u32 attribute = get_attribute(info, scr_readw(s));
 	u8 *dst, *buf = NULL;
-	u32 vxres = GETVXRES(ops->p->scrollmode, info);
+	u32 vxres = info->var.xres;
 
 	if (!ops->fontbuffer)
 		return;
@@ -212,7 +193,7 @@ static void cw_cursor(struct vc_data *vc, struct fb_info *info, int mode,
 	int attribute, use_sw = vc->vc_cursor_type & CUR_SW;
 	int err = 1, dx, dy;
 	char *src;
-	u32 vxres = GETVXRES(ops->p->scrollmode, info);
+	u32 vxres = info->var.xres;
 
 	if (!ops->fontbuffer)
 		return;
@@ -369,7 +350,7 @@ static void cw_cursor(struct vc_data *vc, struct fb_info *info, int mode,
 static int cw_update_start(struct fb_info *info)
 {
 	struct fbcon_ops *ops = info->fbcon_par;
-	u32 vxres = GETVXRES(ops->p->scrollmode, info);
+	u32 vxres = info->var.xres;
 	u32 xoffset;
 	int err;
 
@@ -385,7 +366,6 @@ static int cw_update_start(struct fb_info *info)
 
 void fbcon_rotate_cw(struct fbcon_ops *ops)
 {
-	ops->bmove = cw_bmove;
 	ops->clear = cw_clear;
 	ops->putcs = cw_putcs;
 	ops->clear_margins = cw_clear_margins;
diff --git a/drivers/video/fbdev/core/fbcon_rotate.h b/drivers/video/fbdev/core/fbcon_rotate.h
index e233444..b528b2e 100644
--- a/drivers/video/fbdev/core/fbcon_rotate.h
+++ b/drivers/video/fbdev/core/fbcon_rotate.h
@@ -11,15 +11,6 @@
 #ifndef _FBCON_ROTATE_H
 #define _FBCON_ROTATE_H
 
-#define GETVYRES(s,i) ({                           \
-        (s == SCROLL_REDRAW || s == SCROLL_MOVE) ? \
-        (i)->var.yres : (i)->var.yres_virtual; })
-
-#define GETVXRES(s,i) ({                           \
-        (s == SCROLL_REDRAW || s == SCROLL_MOVE || !(i)->fix.xpanstep) ? \
-        (i)->var.xres : (i)->var.xres_virtual; })
-
-
 static inline int pattern_test_bit(u32 x, u32 y, u32 pitch, const char *pat)
 {
 	u32 tmp = (y * pitch) + x, index = tmp / 8,  bit = tmp % 8;
diff --git a/drivers/video/fbdev/core/fbcon_ud.c b/drivers/video/fbdev/core/fbcon_ud.c
index 8d5e66b..09619bd 100644
--- a/drivers/video/fbdev/core/fbcon_ud.c
+++ b/drivers/video/fbdev/core/fbcon_ud.c
@@ -44,33 +44,13 @@ static void ud_update_attr(u8 *dst, u8 *src, int attribute,
 	}
 }
 
-
-static void ud_bmove(struct vc_data *vc, struct fb_info *info, int sy,
-		     int sx, int dy, int dx, int height, int width)
-{
-	struct fbcon_ops *ops = info->fbcon_par;
-	struct fb_copyarea area;
-	u32 vyres = GETVYRES(ops->p->scrollmode, info);
-	u32 vxres = GETVXRES(ops->p->scrollmode, info);
-
-	area.sy = vyres - ((sy + height) * vc->vc_font.height);
-	area.sx = vxres - ((sx + width) * vc->vc_font.width);
-	area.dy = vyres - ((dy + height) * vc->vc_font.height);
-	area.dx = vxres - ((dx + width) * vc->vc_font.width);
-	area.height = height * vc->vc_font.height;
-	area.width  = width * vc->vc_font.width;
-
-	info->fbops->fb_copyarea(info, &area);
-}
-
 static void ud_clear(struct vc_data *vc, struct fb_info *info, int sy,
 		     int sx, int height, int width)
 {
-	struct fbcon_ops *ops = info->fbcon_par;
 	struct fb_fillrect region;
 	int bgshift = (vc->vc_hi_font_mask) ? 13 : 12;
-	u32 vyres = GETVYRES(ops->p->scrollmode, info);
-	u32 vxres = GETVXRES(ops->p->scrollmode, info);
+	u32 vyres = info->var.yres;
+	u32 vxres = info->var.xres;
 
 	region.color = attr_bgcol_ec(bgshift,vc,info);
 	region.dy = vyres - ((sy + height) * vc->vc_font.height);
@@ -162,8 +142,8 @@ static void ud_putcs(struct vc_data *vc, struct fb_info *info,
 	u32 mod = vc->vc_font.width % 8, cnt, pitch, size;
 	u32 attribute = get_attribute(info, scr_readw(s));
 	u8 *dst, *buf = NULL;
-	u32 vyres = GETVYRES(ops->p->scrollmode, info);
-	u32 vxres = GETVXRES(ops->p->scrollmode, info);
+	u32 vyres = info->var.yres;
+	u32 vxres = info->var.xres;
 
 	if (!ops->fontbuffer)
 		return;
@@ -259,8 +239,8 @@ static void ud_cursor(struct vc_data *vc, struct fb_info *info, int mode,
 	int attribute, use_sw = vc->vc_cursor_type & CUR_SW;
 	int err = 1, dx, dy;
 	char *src;
-	u32 vyres = GETVYRES(ops->p->scrollmode, info);
-	u32 vxres = GETVXRES(ops->p->scrollmode, info);
+	u32 vyres = info->var.yres;
+	u32 vxres = info->var.xres;
 
 	if (!ops->fontbuffer)
 		return;
@@ -410,8 +390,8 @@ static int ud_update_start(struct fb_info *info)
 {
 	struct fbcon_ops *ops = info->fbcon_par;
 	int xoffset, yoffset;
-	u32 vyres = GETVYRES(ops->p->scrollmode, info);
-	u32 vxres = GETVXRES(ops->p->scrollmode, info);
+	u32 vyres = info->var.yres;
+	u32 vxres = info->var.xres;
 	int err;
 
 	xoffset = vxres - info->var.xres - ops->var.xoffset;
@@ -429,7 +409,6 @@ static int ud_update_start(struct fb_info *info)
 
 void fbcon_rotate_ud(struct fbcon_ops *ops)
 {
-	ops->bmove = ud_bmove;
 	ops->clear = ud_clear;
 	ops->putcs = ud_putcs;
 	ops->clear_margins = ud_clear_margins;
diff --git a/drivers/video/fbdev/core/fbmem.c b/drivers/video/fbdev/core/fbmem.c
index 7420d2c..826175ad 100644
--- a/drivers/video/fbdev/core/fbmem.c
+++ b/drivers/video/fbdev/core/fbmem.c
@@ -1702,8 +1702,11 @@ static void do_unregister_framebuffer(struct fb_info *fb_info)
 {
 	unlink_framebuffer(fb_info);
 	if (fb_info->pixmap.addr &&
-	    (fb_info->pixmap.flags & FB_PIXMAP_DEFAULT))
+	    (fb_info->pixmap.flags & FB_PIXMAP_DEFAULT)) {
 		kfree(fb_info->pixmap.addr);
+		fb_info->pixmap.addr = NULL;
+	}
+
 	fb_destroy_modelist(&fb_info->modelist);
 	registered_fb[fb_info->node] = NULL;
 	num_registered_fb--;
diff --git a/drivers/video/fbdev/core/tileblit.c b/drivers/video/fbdev/core/tileblit.c
index 2768eff..72af950 100644
--- a/drivers/video/fbdev/core/tileblit.c
+++ b/drivers/video/fbdev/core/tileblit.c
@@ -16,21 +16,6 @@
 #include <asm/types.h>
 #include "fbcon.h"
 
-static void tile_bmove(struct vc_data *vc, struct fb_info *info, int sy,
-		       int sx, int dy, int dx, int height, int width)
-{
-	struct fb_tilearea area;
-
-	area.sx = sx;
-	area.sy = sy;
-	area.dx = dx;
-	area.dy = dy;
-	area.height = height;
-	area.width = width;
-
-	info->tileops->fb_tilecopy(info, &area);
-}
-
 static void tile_clear(struct vc_data *vc, struct fb_info *info, int sy,
 		       int sx, int height, int width)
 {
@@ -133,7 +118,6 @@ void fbcon_set_tileops(struct vc_data *vc, struct fb_info *info)
 	struct fb_tilemap map;
 	struct fbcon_ops *ops = info->fbcon_par;
 
-	ops->bmove = tile_bmove;
 	ops->clear = tile_clear;
 	ops->putcs = tile_putcs;
 	ops->clear_margins = tile_clear_margins;
diff --git a/drivers/video/fbdev/skeletonfb.c b/drivers/video/fbdev/skeletonfb.c
index bcacfb6..0fe922f 100644
--- a/drivers/video/fbdev/skeletonfb.c
+++ b/drivers/video/fbdev/skeletonfb.c
@@ -505,15 +505,15 @@ void xxxfb_fillrect(struct fb_info *p, const struct fb_fillrect *region)
 }
 
 /**
- *      xxxfb_copyarea - REQUIRED function. Can use generic routines if
- *                       non acclerated hardware and packed pixel based.
+ *      xxxfb_copyarea - OBSOLETE function.
  *                       Copies one area of the screen to another area.
+ *                       Will be deleted in a future version
  *
  *      @info: frame buffer structure that represents a single frame buffer
  *      @area: Structure providing the data to copy the framebuffer contents
  *	       from one region to another.
  *
- *      This drawing operation copies a rectangular area from one area of the
+ *      This drawing operation copied a rectangular area from one area of the
  *	screen to another area.
  */
 void xxxfb_copyarea(struct fb_info *p, const struct fb_copyarea *area) 
@@ -645,9 +645,9 @@ static const struct fb_ops xxxfb_ops = {
 	.fb_setcolreg	= xxxfb_setcolreg,
 	.fb_blank	= xxxfb_blank,
 	.fb_pan_display	= xxxfb_pan_display,
-	.fb_fillrect	= xxxfb_fillrect, 	/* Needed !!! */
-	.fb_copyarea	= xxxfb_copyarea,	/* Needed !!! */
-	.fb_imageblit	= xxxfb_imageblit,	/* Needed !!! */
+	.fb_fillrect	= xxxfb_fillrect,	/* Needed !!!   */
+	.fb_copyarea	= xxxfb_copyarea,	/* Obsolete     */
+	.fb_imageblit	= xxxfb_imageblit,	/* Needed !!!   */
 	.fb_cursor	= xxxfb_cursor,		/* Optional !!! */
 	.fb_sync	= xxxfb_sync,
 	.fb_ioctl	= xxxfb_ioctl,
diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index 0da0af2..96e5a87 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -2889,6 +2889,7 @@ static unsigned int virtio_mem_features[] = {
 #if defined(CONFIG_NUMA) && defined(CONFIG_ACPI_NUMA)
 	VIRTIO_MEM_F_ACPI_PXM,
 #endif
+	VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE,
 };
 
 static const struct virtio_device_id virtio_mem_id_table[] = {
diff --git a/fs/9p/vfs_addr.c b/fs/9p/vfs_addr.c
index adafdf8..fac918c 100644
--- a/fs/9p/vfs_addr.c
+++ b/fs/9p/vfs_addr.c
@@ -108,7 +108,9 @@ static const struct netfs_read_request_ops v9fs_req_ops = {
  */
 static int v9fs_vfs_readpage(struct file *file, struct page *page)
 {
-	return netfs_readpage(file, page, &v9fs_req_ops, NULL);
+	struct folio *folio = page_folio(page);
+
+	return netfs_readpage(file, folio, &v9fs_req_ops, NULL);
 }
 
 /**
@@ -130,13 +132,15 @@ static void v9fs_vfs_readahead(struct readahead_control *ractl)
 
 static int v9fs_release_page(struct page *page, gfp_t gfp)
 {
-	if (PagePrivate(page))
+	struct folio *folio = page_folio(page);
+
+	if (folio_test_private(folio))
 		return 0;
 #ifdef CONFIG_9P_FSCACHE
-	if (PageFsCache(page)) {
+	if (folio_test_fscache(folio)) {
 		if (!(gfp & __GFP_DIRECT_RECLAIM) || !(gfp & __GFP_FS))
 			return 0;
-		wait_on_page_fscache(page);
+		folio_wait_fscache(folio);
 	}
 #endif
 	return 1;
@@ -152,55 +156,58 @@ static int v9fs_release_page(struct page *page, gfp_t gfp)
 static void v9fs_invalidate_page(struct page *page, unsigned int offset,
 				 unsigned int length)
 {
-	wait_on_page_fscache(page);
+	struct folio *folio = page_folio(page);
+
+	folio_wait_fscache(folio);
 }
 
-static int v9fs_vfs_writepage_locked(struct page *page)
+static int v9fs_vfs_write_folio_locked(struct folio *folio)
 {
-	struct inode *inode = page->mapping->host;
+	struct inode *inode = folio_inode(folio);
 	struct v9fs_inode *v9inode = V9FS_I(inode);
-	loff_t start = page_offset(page);
-	loff_t size = i_size_read(inode);
+	loff_t start = folio_pos(folio);
+	loff_t i_size = i_size_read(inode);
 	struct iov_iter from;
-	int err, len;
+	size_t len = folio_size(folio);
+	int err;
 
-	if (page->index == size >> PAGE_SHIFT)
-		len = size & ~PAGE_MASK;
-	else
-		len = PAGE_SIZE;
+	if (start >= i_size)
+		return 0; /* Simultaneous truncation occurred */
 
-	iov_iter_xarray(&from, WRITE, &page->mapping->i_pages, start, len);
+	len = min_t(loff_t, i_size - start, len);
+
+	iov_iter_xarray(&from, WRITE, &folio_mapping(folio)->i_pages, start, len);
 
 	/* We should have writeback_fid always set */
 	BUG_ON(!v9inode->writeback_fid);
 
-	set_page_writeback(page);
+	folio_start_writeback(folio);
 
 	p9_client_write(v9inode->writeback_fid, start, &from, &err);
 
-	end_page_writeback(page);
+	folio_end_writeback(folio);
 	return err;
 }
 
 static int v9fs_vfs_writepage(struct page *page, struct writeback_control *wbc)
 {
+	struct folio *folio = page_folio(page);
 	int retval;
 
-	p9_debug(P9_DEBUG_VFS, "page %p\n", page);
+	p9_debug(P9_DEBUG_VFS, "folio %p\n", folio);
 
-	retval = v9fs_vfs_writepage_locked(page);
+	retval = v9fs_vfs_write_folio_locked(folio);
 	if (retval < 0) {
 		if (retval == -EAGAIN) {
-			redirty_page_for_writepage(wbc, page);
+			folio_redirty_for_writepage(wbc, folio);
 			retval = 0;
 		} else {
-			SetPageError(page);
-			mapping_set_error(page->mapping, retval);
+			mapping_set_error(folio_mapping(folio), retval);
 		}
 	} else
 		retval = 0;
 
-	unlock_page(page);
+	folio_unlock(folio);
 	return retval;
 }
 
@@ -213,14 +220,15 @@ static int v9fs_vfs_writepage(struct page *page, struct writeback_control *wbc)
 
 static int v9fs_launder_page(struct page *page)
 {
+	struct folio *folio = page_folio(page);
 	int retval;
 
-	if (clear_page_dirty_for_io(page)) {
-		retval = v9fs_vfs_writepage_locked(page);
+	if (folio_clear_dirty_for_io(folio)) {
+		retval = v9fs_vfs_write_folio_locked(folio);
 		if (retval)
 			return retval;
 	}
-	wait_on_page_fscache(page);
+	folio_wait_fscache(folio);
 	return 0;
 }
 
@@ -265,10 +273,10 @@ v9fs_direct_IO(struct kiocb *iocb, struct iov_iter *iter)
 
 static int v9fs_write_begin(struct file *filp, struct address_space *mapping,
 			    loff_t pos, unsigned int len, unsigned int flags,
-			    struct page **pagep, void **fsdata)
+			    struct page **subpagep, void **fsdata)
 {
 	int retval;
-	struct page *page;
+	struct folio *folio;
 	struct v9fs_inode *v9inode = V9FS_I(mapping->host);
 
 	p9_debug(P9_DEBUG_VFS, "filp %p, mapping %p\n", filp, mapping);
@@ -279,31 +287,32 @@ static int v9fs_write_begin(struct file *filp, struct address_space *mapping,
 	 * file.  We need to do this before we get a lock on the page in case
 	 * there's more than one writer competing for the same cache block.
 	 */
-	retval = netfs_write_begin(filp, mapping, pos, len, flags, &page, fsdata,
+	retval = netfs_write_begin(filp, mapping, pos, len, flags, &folio, fsdata,
 				   &v9fs_req_ops, NULL);
 	if (retval < 0)
 		return retval;
 
-	*pagep = find_subpage(page, pos / PAGE_SIZE);
+	*subpagep = &folio->page;
 	return retval;
 }
 
 static int v9fs_write_end(struct file *filp, struct address_space *mapping,
 			  loff_t pos, unsigned int len, unsigned int copied,
-			  struct page *page, void *fsdata)
+			  struct page *subpage, void *fsdata)
 {
 	loff_t last_pos = pos + copied;
-	struct inode *inode = page->mapping->host;
+	struct folio *folio = page_folio(subpage);
+	struct inode *inode = mapping->host;
 
 	p9_debug(P9_DEBUG_VFS, "filp %p, mapping %p\n", filp, mapping);
 
-	if (!PageUptodate(page)) {
+	if (!folio_test_uptodate(folio)) {
 		if (unlikely(copied < len)) {
 			copied = 0;
 			goto out;
 		}
 
-		SetPageUptodate(page);
+		folio_mark_uptodate(folio);
 	}
 
 	/*
@@ -314,10 +323,10 @@ static int v9fs_write_end(struct file *filp, struct address_space *mapping,
 		inode_add_bytes(inode, last_pos - inode->i_size);
 		i_size_write(inode, last_pos);
 	}
-	set_page_dirty(page);
+	folio_mark_dirty(folio);
 out:
-	unlock_page(page);
-	put_page(page);
+	folio_unlock(folio);
+	folio_put(folio);
 
 	return copied;
 }
diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c
index 4244d48..612e297 100644
--- a/fs/9p/vfs_file.c
+++ b/fs/9p/vfs_file.c
@@ -528,13 +528,13 @@ static vm_fault_t
 v9fs_vm_page_mkwrite(struct vm_fault *vmf)
 {
 	struct v9fs_inode *v9inode;
-	struct page *page = vmf->page;
+	struct folio *folio = page_folio(vmf->page);
 	struct file *filp = vmf->vma->vm_file;
 	struct inode *inode = file_inode(filp);
 
 
-	p9_debug(P9_DEBUG_VFS, "page %p fid %lx\n",
-		 page, (unsigned long)filp->private_data);
+	p9_debug(P9_DEBUG_VFS, "folio %p fid %lx\n",
+		 folio, (unsigned long)filp->private_data);
 
 	v9inode = V9FS_I(inode);
 
@@ -542,24 +542,24 @@ v9fs_vm_page_mkwrite(struct vm_fault *vmf)
 	 * be modified.  We then assume the entire page will need writing back.
 	 */
 #ifdef CONFIG_9P_FSCACHE
-	if (PageFsCache(page) &&
-	    wait_on_page_fscache_killable(page) < 0)
-		return VM_FAULT_RETRY;
+	if (folio_test_fscache(folio) &&
+	    folio_wait_fscache_killable(folio) < 0)
+		return VM_FAULT_NOPAGE;
 #endif
 
 	/* Update file times before taking page lock */
 	file_update_time(filp);
 
 	BUG_ON(!v9inode->writeback_fid);
-	if (lock_page_killable(page) < 0)
+	if (folio_lock_killable(folio) < 0)
 		return VM_FAULT_RETRY;
-	if (page->mapping != inode->i_mapping)
+	if (folio_mapping(folio) != inode->i_mapping)
 		goto out_unlock;
-	wait_for_stable_page(page);
+	folio_wait_stable(folio);
 
 	return VM_FAULT_LOCKED;
 out_unlock:
-	unlock_page(page);
+	folio_unlock(folio);
 	return VM_FAULT_NOPAGE;
 }
 
diff --git a/fs/afs/dir.c b/fs/afs/dir.c
index 4579bbd..da9b4f8 100644
--- a/fs/afs/dir.c
+++ b/fs/afs/dir.c
@@ -103,13 +103,13 @@ struct afs_lookup_cookie {
 };
 
 /*
- * Drop the refs that we're holding on the pages we were reading into.  We've
+ * Drop the refs that we're holding on the folios we were reading into.  We've
  * got refs on the first nr_pages pages.
  */
 static void afs_dir_read_cleanup(struct afs_read *req)
 {
 	struct address_space *mapping = req->vnode->vfs_inode.i_mapping;
-	struct page *page;
+	struct folio *folio;
 	pgoff_t last = req->nr_pages - 1;
 
 	XA_STATE(xas, &mapping->i_pages, 0);
@@ -118,65 +118,56 @@ static void afs_dir_read_cleanup(struct afs_read *req)
 		return;
 
 	rcu_read_lock();
-	xas_for_each(&xas, page, last) {
-		if (xas_retry(&xas, page))
+	xas_for_each(&xas, folio, last) {
+		if (xas_retry(&xas, folio))
 			continue;
-		BUG_ON(xa_is_value(page));
-		BUG_ON(PageCompound(page));
-		ASSERTCMP(page->mapping, ==, mapping);
+		BUG_ON(xa_is_value(folio));
+		ASSERTCMP(folio_file_mapping(folio), ==, mapping);
 
-		put_page(page);
+		folio_put(folio);
 	}
 
 	rcu_read_unlock();
 }
 
 /*
- * check that a directory page is valid
+ * check that a directory folio is valid
  */
-static bool afs_dir_check_page(struct afs_vnode *dvnode, struct page *page,
-			       loff_t i_size)
+static bool afs_dir_check_folio(struct afs_vnode *dvnode, struct folio *folio,
+				loff_t i_size)
 {
-	struct afs_xdr_dir_page *dbuf;
-	loff_t latter, off;
-	int tmp, qty;
+	union afs_xdr_dir_block *block;
+	size_t offset, size;
+	loff_t pos;
 
-	/* Determine how many magic numbers there should be in this page, but
+	/* Determine how many magic numbers there should be in this folio, but
 	 * we must take care because the directory may change size under us.
 	 */
-	off = page_offset(page);
-	if (i_size <= off)
+	pos = folio_pos(folio);
+	if (i_size <= pos)
 		goto checked;
 
-	latter = i_size - off;
-	if (latter >= PAGE_SIZE)
-		qty = PAGE_SIZE;
-	else
-		qty = latter;
-	qty /= sizeof(union afs_xdr_dir_block);
-
-	/* check them */
-	dbuf = kmap_atomic(page);
-	for (tmp = 0; tmp < qty; tmp++) {
-		if (dbuf->blocks[tmp].hdr.magic != AFS_DIR_MAGIC) {
-			printk("kAFS: %s(%lx): bad magic %d/%d is %04hx\n",
-			       __func__, dvnode->vfs_inode.i_ino, tmp, qty,
-			       ntohs(dbuf->blocks[tmp].hdr.magic));
-			trace_afs_dir_check_failed(dvnode, off, i_size);
-			kunmap(page);
+	size = min_t(loff_t, folio_size(folio), i_size - pos);
+	for (offset = 0; offset < size; offset += sizeof(*block)) {
+		block = kmap_local_folio(folio, offset);
+		if (block->hdr.magic != AFS_DIR_MAGIC) {
+			printk("kAFS: %s(%lx): [%llx] bad magic %zx/%zx is %04hx\n",
+			       __func__, dvnode->vfs_inode.i_ino,
+			       pos, offset, size, ntohs(block->hdr.magic));
+			trace_afs_dir_check_failed(dvnode, pos + offset, i_size);
+			kunmap_local(block);
 			trace_afs_file_error(dvnode, -EIO, afs_file_error_dir_bad_magic);
 			goto error;
 		}
 
 		/* Make sure each block is NUL terminated so we can reasonably
-		 * use string functions on it.  The filenames in the page
+		 * use string functions on it.  The filenames in the folio
 		 * *should* be NUL-terminated anyway.
 		 */
-		((u8 *)&dbuf->blocks[tmp])[AFS_DIR_BLOCK_SIZE - 1] = 0;
+		((u8 *)block)[AFS_DIR_BLOCK_SIZE - 1] = 0;
+
+		kunmap_local(block);
 	}
-
-	kunmap_atomic(dbuf);
-
 checked:
 	afs_stat_v(dvnode, n_read_dir);
 	return true;
@@ -190,11 +181,11 @@ static bool afs_dir_check_page(struct afs_vnode *dvnode, struct page *page,
  */
 static void afs_dir_dump(struct afs_vnode *dvnode, struct afs_read *req)
 {
-	struct afs_xdr_dir_page *dbuf;
+	union afs_xdr_dir_block *block;
 	struct address_space *mapping = dvnode->vfs_inode.i_mapping;
-	struct page *page;
-	unsigned int i, qty = PAGE_SIZE / sizeof(union afs_xdr_dir_block);
+	struct folio *folio;
 	pgoff_t last = req->nr_pages - 1;
+	size_t offset, size;
 
 	XA_STATE(xas, &mapping->i_pages, 0);
 
@@ -205,30 +196,28 @@ static void afs_dir_dump(struct afs_vnode *dvnode, struct afs_read *req)
 		req->pos, req->nr_pages,
 		req->iter->iov_offset,  iov_iter_count(req->iter));
 
-	xas_for_each(&xas, page, last) {
-		if (xas_retry(&xas, page))
+	xas_for_each(&xas, folio, last) {
+		if (xas_retry(&xas, folio))
 			continue;
 
-		BUG_ON(PageCompound(page));
-		BUG_ON(page->mapping != mapping);
+		BUG_ON(folio_file_mapping(folio) != mapping);
 
-		dbuf = kmap_atomic(page);
-		for (i = 0; i < qty; i++) {
-			union afs_xdr_dir_block *block = &dbuf->blocks[i];
-
-			pr_warn("[%02lx] %32phN\n", page->index * qty + i, block);
+		size = min_t(loff_t, folio_size(folio), req->actual_len - folio_pos(folio));
+		for (offset = 0; offset < size; offset += sizeof(*block)) {
+			block = kmap_local_folio(folio, offset);
+			pr_warn("[%02lx] %32phN\n", folio_index(folio) + offset, block);
+			kunmap_local(block);
 		}
-		kunmap_atomic(dbuf);
 	}
 }
 
 /*
- * Check all the pages in a directory.  All the pages are held pinned.
+ * Check all the blocks in a directory.  All the folios are held pinned.
  */
 static int afs_dir_check(struct afs_vnode *dvnode, struct afs_read *req)
 {
 	struct address_space *mapping = dvnode->vfs_inode.i_mapping;
-	struct page *page;
+	struct folio *folio;
 	pgoff_t last = req->nr_pages - 1;
 	int ret = 0;
 
@@ -238,14 +227,13 @@ static int afs_dir_check(struct afs_vnode *dvnode, struct afs_read *req)
 		return 0;
 
 	rcu_read_lock();
-	xas_for_each(&xas, page, last) {
-		if (xas_retry(&xas, page))
+	xas_for_each(&xas, folio, last) {
+		if (xas_retry(&xas, folio))
 			continue;
 
-		BUG_ON(PageCompound(page));
-		BUG_ON(page->mapping != mapping);
+		BUG_ON(folio_file_mapping(folio) != mapping);
 
-		if (!afs_dir_check_page(dvnode, page, req->file_size)) {
+		if (!afs_dir_check_folio(dvnode, folio, req->actual_len)) {
 			afs_dir_dump(dvnode, req);
 			ret = -EIO;
 			break;
@@ -274,15 +262,16 @@ static int afs_dir_open(struct inode *inode, struct file *file)
 
 /*
  * Read the directory into the pagecache in one go, scrubbing the previous
- * contents.  The list of pages is returned, pinning them so that they don't
+ * contents.  The list of folios is returned, pinning them so that they don't
  * get reclaimed during the iteration.
  */
 static struct afs_read *afs_read_dir(struct afs_vnode *dvnode, struct key *key)
 	__acquires(&dvnode->validate_lock)
 {
+	struct address_space *mapping = dvnode->vfs_inode.i_mapping;
 	struct afs_read *req;
 	loff_t i_size;
-	int nr_pages, i, n;
+	int nr_pages, i;
 	int ret;
 
 	_enter("");
@@ -320,43 +309,30 @@ static struct afs_read *afs_read_dir(struct afs_vnode *dvnode, struct key *key)
 	req->iter = &req->def_iter;
 
 	/* Fill in any gaps that we might find where the memory reclaimer has
-	 * been at work and pin all the pages.  If there are any gaps, we will
+	 * been at work and pin all the folios.  If there are any gaps, we will
 	 * need to reread the entire directory contents.
 	 */
 	i = req->nr_pages;
 	while (i < nr_pages) {
-		struct page *pages[8], *page;
+		struct folio *folio;
 
-		n = find_get_pages_contig(dvnode->vfs_inode.i_mapping, i,
-					  min_t(unsigned int, nr_pages - i,
-						ARRAY_SIZE(pages)),
-					  pages);
-		_debug("find %u at %u/%u", n, i, nr_pages);
-
-		if (n == 0) {
-			gfp_t gfp = dvnode->vfs_inode.i_mapping->gfp_mask;
-
+		folio = filemap_get_folio(mapping, i);
+		if (!folio) {
 			if (test_and_clear_bit(AFS_VNODE_DIR_VALID, &dvnode->flags))
 				afs_stat_v(dvnode, n_inval);
 
 			ret = -ENOMEM;
-			page = __page_cache_alloc(gfp);
-			if (!page)
+			folio = __filemap_get_folio(mapping,
+						    i, FGP_LOCK | FGP_CREAT,
+						    mapping->gfp_mask);
+			if (!folio)
 				goto error;
-			ret = add_to_page_cache_lru(page,
-						    dvnode->vfs_inode.i_mapping,
-						    i, gfp);
-			if (ret < 0)
-				goto error;
-
-			attach_page_private(page, (void *)1);
-			unlock_page(page);
-			req->nr_pages++;
-			i++;
-		} else {
-			req->nr_pages += n;
-			i += n;
+			folio_attach_private(folio, (void *)1);
+			folio_unlock(folio);
 		}
+
+		req->nr_pages += folio_nr_pages(folio);
+		i += folio_nr_pages(folio);
 	}
 
 	/* If we're going to reload, we need to lock all the pages to prevent
@@ -424,7 +400,7 @@ static int afs_dir_iterate_block(struct afs_vnode *dvnode,
 	size_t nlen;
 	int tmp;
 
-	_enter("%u,%x,%p,,",(unsigned)ctx->pos,blkoff,block);
+	_enter("%llx,%x", ctx->pos, blkoff);
 
 	curr = (ctx->pos - blkoff) / sizeof(union afs_xdr_dirent);
 
@@ -513,12 +489,10 @@ static int afs_dir_iterate(struct inode *dir, struct dir_context *ctx,
 			   struct key *key, afs_dataversion_t *_dir_version)
 {
 	struct afs_vnode *dvnode = AFS_FS_I(dir);
-	struct afs_xdr_dir_page *dbuf;
 	union afs_xdr_dir_block *dblock;
 	struct afs_read *req;
-	struct page *page;
-	unsigned blkoff, limit;
-	void __rcu **slot;
+	struct folio *folio;
+	unsigned offset, size;
 	int ret;
 
 	_enter("{%lu},%u,,", dir->i_ino, (unsigned)ctx->pos);
@@ -540,43 +514,30 @@ static int afs_dir_iterate(struct inode *dir, struct dir_context *ctx,
 	/* walk through the blocks in sequence */
 	ret = 0;
 	while (ctx->pos < req->actual_len) {
-		blkoff = ctx->pos & ~(sizeof(union afs_xdr_dir_block) - 1);
-
-		/* Fetch the appropriate page from the directory and re-add it
+		/* Fetch the appropriate folio from the directory and re-add it
 		 * to the LRU.  We have all the pages pinned with an extra ref.
 		 */
-		rcu_read_lock();
-		page = NULL;
-		slot = radix_tree_lookup_slot(&dvnode->vfs_inode.i_mapping->i_pages,
-					      blkoff / PAGE_SIZE);
-		if (slot)
-			page = radix_tree_deref_slot(slot);
-		rcu_read_unlock();
-		if (!page) {
+		folio = __filemap_get_folio(dir->i_mapping, ctx->pos / PAGE_SIZE,
+					    FGP_ACCESSED, 0);
+		if (!folio) {
 			ret = afs_bad(dvnode, afs_file_error_dir_missing_page);
 			break;
 		}
-		mark_page_accessed(page);
 
-		limit = blkoff & ~(PAGE_SIZE - 1);
+		offset = round_down(ctx->pos, sizeof(*dblock)) - folio_file_pos(folio);
+		size = min_t(loff_t, folio_size(folio),
+			     req->actual_len - folio_file_pos(folio));
 
-		dbuf = kmap(page);
-
-		/* deal with the individual blocks stashed on this page */
 		do {
-			dblock = &dbuf->blocks[(blkoff % PAGE_SIZE) /
-					       sizeof(union afs_xdr_dir_block)];
-			ret = afs_dir_iterate_block(dvnode, ctx, dblock, blkoff);
-			if (ret != 1) {
-				kunmap(page);
+			dblock = kmap_local_folio(folio, offset);
+			ret = afs_dir_iterate_block(dvnode, ctx, dblock,
+						    folio_file_pos(folio) + offset);
+			kunmap_local(dblock);
+			if (ret != 1)
 				goto out;
-			}
 
-			blkoff += sizeof(union afs_xdr_dir_block);
+		} while (offset += sizeof(*dblock), offset < size);
 
-		} while (ctx->pos < dir->i_size && blkoff < limit);
-
-		kunmap(page);
 		ret = 0;
 	}
 
@@ -2037,42 +1998,42 @@ static int afs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 }
 
 /*
- * Release a directory page and clean up its private state if it's not busy
- * - return true if the page can now be released, false if not
+ * Release a directory folio and clean up its private state if it's not busy
+ * - return true if the folio can now be released, false if not
  */
-static int afs_dir_releasepage(struct page *page, gfp_t gfp_flags)
+static int afs_dir_releasepage(struct page *subpage, gfp_t gfp_flags)
 {
-	struct afs_vnode *dvnode = AFS_FS_I(page->mapping->host);
+	struct folio *folio = page_folio(subpage);
+	struct afs_vnode *dvnode = AFS_FS_I(folio_inode(folio));
 
-	_enter("{{%llx:%llu}[%lu]}", dvnode->fid.vid, dvnode->fid.vnode, page->index);
+	_enter("{{%llx:%llu}[%lu]}", dvnode->fid.vid, dvnode->fid.vnode, folio_index(folio));
 
-	detach_page_private(page);
+	folio_detach_private(folio);
 
 	/* The directory will need reloading. */
 	if (test_and_clear_bit(AFS_VNODE_DIR_VALID, &dvnode->flags))
 		afs_stat_v(dvnode, n_relpg);
-	return 1;
+	return true;
 }
 
 /*
- * invalidate part or all of a page
- * - release a page and clean up its private data if offset is 0 (indicating
- *   the entire page)
+ * Invalidate part or all of a folio.
  */
-static void afs_dir_invalidatepage(struct page *page, unsigned int offset,
+static void afs_dir_invalidatepage(struct page *subpage, unsigned int offset,
 				   unsigned int length)
 {
-	struct afs_vnode *dvnode = AFS_FS_I(page->mapping->host);
+	struct folio *folio = page_folio(subpage);
+	struct afs_vnode *dvnode = AFS_FS_I(folio_inode(folio));
 
-	_enter("{%lu},%u,%u", page->index, offset, length);
+	_enter("{%lu},%u,%u", folio_index(folio), offset, length);
 
-	BUG_ON(!PageLocked(page));
+	BUG_ON(!folio_test_locked(folio));
 
 	/* The directory will need reloading. */
 	if (test_and_clear_bit(AFS_VNODE_DIR_VALID, &dvnode->flags))
 		afs_stat_v(dvnode, n_inval);
 
-	/* we clean up only if the entire page is being invalidated */
-	if (offset == 0 && length == thp_size(page))
-		detach_page_private(page);
+	/* we clean up only if the entire folio is being invalidated */
+	if (offset == 0 && length == folio_size(folio))
+		folio_detach_private(folio);
 }
diff --git a/fs/afs/dir_edit.c b/fs/afs/dir_edit.c
index 540b9fc..d98e109 100644
--- a/fs/afs/dir_edit.c
+++ b/fs/afs/dir_edit.c
@@ -105,6 +105,25 @@ static void afs_clear_contig_bits(union afs_xdr_dir_block *block,
 }
 
 /*
+ * Get a new directory folio.
+ */
+static struct folio *afs_dir_get_folio(struct afs_vnode *vnode, pgoff_t index)
+{
+	struct address_space *mapping = vnode->vfs_inode.i_mapping;
+	struct folio *folio;
+
+	folio = __filemap_get_folio(mapping, index,
+				    FGP_LOCK | FGP_ACCESSED | FGP_CREAT,
+				    mapping->gfp_mask);
+	if (!folio)
+		clear_bit(AFS_VNODE_DIR_VALID, &vnode->flags);
+	else if (folio && !folio_test_private(folio))
+		folio_attach_private(folio, (void *)1);
+
+	return folio;
+}
+
+/*
  * Scan a directory block looking for a dirent of the right name.
  */
 static int afs_dir_scan_block(union afs_xdr_dir_block *block, struct qstr *name,
@@ -188,13 +207,11 @@ void afs_edit_dir_add(struct afs_vnode *vnode,
 		      enum afs_edit_dir_reason why)
 {
 	union afs_xdr_dir_block *meta, *block;
-	struct afs_xdr_dir_page *meta_page, *dir_page;
 	union afs_xdr_dirent *de;
-	struct page *page0, *page;
+	struct folio *folio0, *folio;
 	unsigned int need_slots, nr_blocks, b;
 	pgoff_t index;
 	loff_t i_size;
-	gfp_t gfp;
 	int slot;
 
 	_enter(",,{%d,%s},", name->len, name->name);
@@ -206,10 +223,8 @@ void afs_edit_dir_add(struct afs_vnode *vnode,
 		return;
 	}
 
-	gfp = vnode->vfs_inode.i_mapping->gfp_mask;
-	page0 = find_or_create_page(vnode->vfs_inode.i_mapping, 0, gfp);
-	if (!page0) {
-		clear_bit(AFS_VNODE_DIR_VALID, &vnode->flags);
+	folio0 = afs_dir_get_folio(vnode, 0);
+	if (!folio0) {
 		_leave(" [fgp]");
 		return;
 	}
@@ -217,42 +232,35 @@ void afs_edit_dir_add(struct afs_vnode *vnode,
 	/* Work out how many slots we're going to need. */
 	need_slots = afs_dir_calc_slots(name->len);
 
-	meta_page = kmap(page0);
-	meta = &meta_page->blocks[0];
+	meta = kmap_local_folio(folio0, 0);
 	if (i_size == 0)
 		goto new_directory;
 	nr_blocks = i_size / AFS_DIR_BLOCK_SIZE;
 
-	/* Find a block that has sufficient slots available.  Each VM page
+	/* Find a block that has sufficient slots available.  Each folio
 	 * contains two or more directory blocks.
 	 */
 	for (b = 0; b < nr_blocks + 1; b++) {
-		/* If the directory extended into a new page, then we need to
-		 * tack a new page on the end.
+		/* If the directory extended into a new folio, then we need to
+		 * tack a new folio on the end.
 		 */
 		index = b / AFS_DIR_BLOCKS_PER_PAGE;
-		if (index == 0) {
-			page = page0;
-			dir_page = meta_page;
+		if (nr_blocks >= AFS_DIR_MAX_BLOCKS)
+			goto error;
+		if (index >= folio_nr_pages(folio0)) {
+			folio = afs_dir_get_folio(vnode, index);
+			if (!folio)
+				goto error;
 		} else {
-			if (nr_blocks >= AFS_DIR_MAX_BLOCKS)
-				goto error;
-			gfp = vnode->vfs_inode.i_mapping->gfp_mask;
-			page = find_or_create_page(vnode->vfs_inode.i_mapping,
-						   index, gfp);
-			if (!page)
-				goto error;
-			if (!PagePrivate(page))
-				attach_page_private(page, (void *)1);
-			dir_page = kmap(page);
+			folio = folio0;
 		}
 
+		block = kmap_local_folio(folio, b * AFS_DIR_BLOCK_SIZE - folio_file_pos(folio));
+
 		/* Abandon the edit if we got a callback break. */
 		if (!test_bit(AFS_VNODE_DIR_VALID, &vnode->flags))
 			goto invalidated;
 
-		block = &dir_page->blocks[b % AFS_DIR_BLOCKS_PER_PAGE];
-
 		_debug("block %u: %2u %3u %u",
 		       b,
 		       (b < AFS_DIR_BLOCKS_WITH_CTR) ? meta->meta.alloc_ctrs[b] : 99,
@@ -266,7 +274,7 @@ void afs_edit_dir_add(struct afs_vnode *vnode,
 			afs_set_i_size(vnode, (b + 1) * AFS_DIR_BLOCK_SIZE);
 		}
 
-		/* Only lower dir pages have a counter in the header. */
+		/* Only lower dir blocks have a counter in the header. */
 		if (b >= AFS_DIR_BLOCKS_WITH_CTR ||
 		    meta->meta.alloc_ctrs[b] >= need_slots) {
 			/* We need to try and find one or more consecutive
@@ -279,10 +287,10 @@ void afs_edit_dir_add(struct afs_vnode *vnode,
 			}
 		}
 
-		if (page != page0) {
-			unlock_page(page);
-			kunmap(page);
-			put_page(page);
+		kunmap_local(block);
+		if (folio != folio0) {
+			folio_unlock(folio);
+			folio_put(folio);
 		}
 	}
 
@@ -298,8 +306,8 @@ void afs_edit_dir_add(struct afs_vnode *vnode,
 	i_size = AFS_DIR_BLOCK_SIZE;
 	afs_set_i_size(vnode, i_size);
 	slot = AFS_DIR_RESV_BLOCKS0;
-	page = page0;
-	block = meta;
+	folio = folio0;
+	block = kmap_local_folio(folio, 0);
 	nr_blocks = 1;
 	b = 0;
 
@@ -318,10 +326,10 @@ void afs_edit_dir_add(struct afs_vnode *vnode,
 
 	/* Adjust the bitmap. */
 	afs_set_contig_bits(block, slot, need_slots);
-	if (page != page0) {
-		unlock_page(page);
-		kunmap(page);
-		put_page(page);
+	kunmap_local(block);
+	if (folio != folio0) {
+		folio_unlock(folio);
+		folio_put(folio);
 	}
 
 	/* Adjust the allocation counter. */
@@ -333,18 +341,19 @@ void afs_edit_dir_add(struct afs_vnode *vnode,
 	_debug("Insert %s in %u[%u]", name->name, b, slot);
 
 out_unmap:
-	unlock_page(page0);
-	kunmap(page0);
-	put_page(page0);
+	kunmap_local(meta);
+	folio_unlock(folio0);
+	folio_put(folio0);
 	_leave("");
 	return;
 
 invalidated:
 	trace_afs_edit_dir(vnode, why, afs_edit_dir_create_inval, 0, 0, 0, 0, name->name);
 	clear_bit(AFS_VNODE_DIR_VALID, &vnode->flags);
-	if (page != page0) {
-		kunmap(page);
-		put_page(page);
+	kunmap_local(block);
+	if (folio != folio0) {
+		folio_unlock(folio);
+		folio_put(folio);
 	}
 	goto out_unmap;
 
@@ -364,10 +373,9 @@ void afs_edit_dir_add(struct afs_vnode *vnode,
 void afs_edit_dir_remove(struct afs_vnode *vnode,
 			 struct qstr *name, enum afs_edit_dir_reason why)
 {
-	struct afs_xdr_dir_page *meta_page, *dir_page;
 	union afs_xdr_dir_block *meta, *block;
 	union afs_xdr_dirent *de;
-	struct page *page0, *page;
+	struct folio *folio0, *folio;
 	unsigned int need_slots, nr_blocks, b;
 	pgoff_t index;
 	loff_t i_size;
@@ -384,9 +392,8 @@ void afs_edit_dir_remove(struct afs_vnode *vnode,
 	}
 	nr_blocks = i_size / AFS_DIR_BLOCK_SIZE;
 
-	page0 = find_lock_page(vnode->vfs_inode.i_mapping, 0);
-	if (!page0) {
-		clear_bit(AFS_VNODE_DIR_VALID, &vnode->flags);
+	folio0 = afs_dir_get_folio(vnode, 0);
+	if (!folio0) {
 		_leave(" [fgp]");
 		return;
 	}
@@ -394,30 +401,27 @@ void afs_edit_dir_remove(struct afs_vnode *vnode,
 	/* Work out how many slots we're going to discard. */
 	need_slots = afs_dir_calc_slots(name->len);
 
-	meta_page = kmap(page0);
-	meta = &meta_page->blocks[0];
+	meta = kmap_local_folio(folio0, 0);
 
-	/* Find a page that has sufficient slots available.  Each VM page
+	/* Find a block that has sufficient slots available.  Each folio
 	 * contains two or more directory blocks.
 	 */
 	for (b = 0; b < nr_blocks; b++) {
 		index = b / AFS_DIR_BLOCKS_PER_PAGE;
-		if (index != 0) {
-			page = find_lock_page(vnode->vfs_inode.i_mapping, index);
-			if (!page)
+		if (index >= folio_nr_pages(folio0)) {
+			folio = afs_dir_get_folio(vnode, index);
+			if (!folio)
 				goto error;
-			dir_page = kmap(page);
 		} else {
-			page = page0;
-			dir_page = meta_page;
+			folio = folio0;
 		}
 
+		block = kmap_local_folio(folio, b * AFS_DIR_BLOCK_SIZE - folio_file_pos(folio));
+
 		/* Abandon the edit if we got a callback break. */
 		if (!test_bit(AFS_VNODE_DIR_VALID, &vnode->flags))
 			goto invalidated;
 
-		block = &dir_page->blocks[b % AFS_DIR_BLOCKS_PER_PAGE];
-
 		if (b > AFS_DIR_BLOCKS_WITH_CTR ||
 		    meta->meta.alloc_ctrs[b] <= AFS_DIR_SLOTS_PER_BLOCK - 1 - need_slots) {
 			slot = afs_dir_scan_block(block, name, b);
@@ -425,10 +429,10 @@ void afs_edit_dir_remove(struct afs_vnode *vnode,
 				goto found_dirent;
 		}
 
-		if (page != page0) {
-			unlock_page(page);
-			kunmap(page);
-			put_page(page);
+		kunmap_local(block);
+		if (folio != folio0) {
+			folio_unlock(folio);
+			folio_put(folio);
 		}
 	}
 
@@ -449,10 +453,10 @@ void afs_edit_dir_remove(struct afs_vnode *vnode,
 
 	/* Adjust the bitmap. */
 	afs_clear_contig_bits(block, slot, need_slots);
-	if (page != page0) {
-		unlock_page(page);
-		kunmap(page);
-		put_page(page);
+	kunmap_local(block);
+	if (folio != folio0) {
+		folio_unlock(folio);
+		folio_put(folio);
 	}
 
 	/* Adjust the allocation counter. */
@@ -464,9 +468,9 @@ void afs_edit_dir_remove(struct afs_vnode *vnode,
 	_debug("Remove %s from %u[%u]", name->name, b, slot);
 
 out_unmap:
-	unlock_page(page0);
-	kunmap(page0);
-	put_page(page0);
+	kunmap_local(meta);
+	folio_unlock(folio0);
+	folio_put(folio0);
 	_leave("");
 	return;
 
@@ -474,10 +478,10 @@ void afs_edit_dir_remove(struct afs_vnode *vnode,
 	trace_afs_edit_dir(vnode, why, afs_edit_dir_delete_inval,
 			   0, 0, 0, 0, name->name);
 	clear_bit(AFS_VNODE_DIR_VALID, &vnode->flags);
-	if (page != page0) {
-		unlock_page(page);
-		kunmap(page);
-		put_page(page);
+	kunmap_local(block);
+	if (folio != folio0) {
+		folio_unlock(folio);
+		folio_put(folio);
 	}
 	goto out_unmap;
 
diff --git a/fs/afs/file.c b/fs/afs/file.c
index eb11d04..cb6ad61 100644
--- a/fs/afs/file.c
+++ b/fs/afs/file.c
@@ -324,21 +324,24 @@ static int afs_symlink_readpage(struct file *file, struct page *page)
 {
 	struct afs_vnode *vnode = AFS_FS_I(page->mapping->host);
 	struct afs_read *fsreq;
+	struct folio *folio = page_folio(page);
 	int ret;
 
 	fsreq = afs_alloc_read(GFP_NOFS);
 	if (!fsreq)
 		return -ENOMEM;
 
-	fsreq->pos	= page->index * PAGE_SIZE;
-	fsreq->len	= PAGE_SIZE;
+	fsreq->pos	= folio_pos(folio);
+	fsreq->len	= folio_size(folio);
 	fsreq->vnode	= vnode;
 	fsreq->iter	= &fsreq->def_iter;
 	iov_iter_xarray(&fsreq->def_iter, READ, &page->mapping->i_pages,
 			fsreq->pos, fsreq->len);
 
 	ret = afs_fetch_data(fsreq->vnode, fsreq);
-	page_endio(page, false, ret);
+	if (ret == 0)
+		SetPageUptodate(page);
+	unlock_page(page);
 	return ret;
 }
 
@@ -362,7 +365,7 @@ static int afs_begin_cache_operation(struct netfs_read_request *rreq)
 }
 
 static int afs_check_write_begin(struct file *file, loff_t pos, unsigned len,
-				 struct page *page, void **_fsdata)
+				 struct folio *folio, void **_fsdata)
 {
 	struct afs_vnode *vnode = AFS_FS_I(file_inode(file));
 
@@ -385,7 +388,9 @@ const struct netfs_read_request_ops afs_req_ops = {
 
 static int afs_readpage(struct file *file, struct page *page)
 {
-	return netfs_readpage(file, page, &afs_req_ops, NULL);
+	struct folio *folio = page_folio(page);
+
+	return netfs_readpage(file, folio, &afs_req_ops, NULL);
 }
 
 static void afs_readahead(struct readahead_control *ractl)
@@ -397,29 +402,29 @@ static void afs_readahead(struct readahead_control *ractl)
  * Adjust the dirty region of the page on truncation or full invalidation,
  * getting rid of the markers altogether if the region is entirely invalidated.
  */
-static void afs_invalidate_dirty(struct page *page, unsigned int offset,
+static void afs_invalidate_dirty(struct folio *folio, unsigned int offset,
 				 unsigned int length)
 {
-	struct afs_vnode *vnode = AFS_FS_I(page->mapping->host);
+	struct afs_vnode *vnode = AFS_FS_I(folio_inode(folio));
 	unsigned long priv;
 	unsigned int f, t, end = offset + length;
 
-	priv = page_private(page);
+	priv = (unsigned long)folio_get_private(folio);
 
 	/* we clean up only if the entire page is being invalidated */
-	if (offset == 0 && length == thp_size(page))
+	if (offset == 0 && length == folio_size(folio))
 		goto full_invalidate;
 
 	 /* If the page was dirtied by page_mkwrite(), the PTE stays writable
 	  * and we don't get another notification to tell us to expand it
 	  * again.
 	  */
-	if (afs_is_page_dirty_mmapped(priv))
+	if (afs_is_folio_dirty_mmapped(priv))
 		return;
 
 	/* We may need to shorten the dirty region */
-	f = afs_page_dirty_from(page, priv);
-	t = afs_page_dirty_to(page, priv);
+	f = afs_folio_dirty_from(folio, priv);
+	t = afs_folio_dirty_to(folio, priv);
 
 	if (t <= offset || f >= end)
 		return; /* Doesn't overlap */
@@ -437,17 +442,17 @@ static void afs_invalidate_dirty(struct page *page, unsigned int offset,
 	if (f == t)
 		goto undirty;
 
-	priv = afs_page_dirty(page, f, t);
-	set_page_private(page, priv);
-	trace_afs_page_dirty(vnode, tracepoint_string("trunc"), page);
+	priv = afs_folio_dirty(folio, f, t);
+	folio_change_private(folio, (void *)priv);
+	trace_afs_folio_dirty(vnode, tracepoint_string("trunc"), folio);
 	return;
 
 undirty:
-	trace_afs_page_dirty(vnode, tracepoint_string("undirty"), page);
-	clear_page_dirty_for_io(page);
+	trace_afs_folio_dirty(vnode, tracepoint_string("undirty"), folio);
+	folio_clear_dirty_for_io(folio);
 full_invalidate:
-	trace_afs_page_dirty(vnode, tracepoint_string("inval"), page);
-	detach_page_private(page);
+	trace_afs_folio_dirty(vnode, tracepoint_string("inval"), folio);
+	folio_detach_private(folio);
 }
 
 /*
@@ -458,14 +463,16 @@ static void afs_invalidate_dirty(struct page *page, unsigned int offset,
 static void afs_invalidatepage(struct page *page, unsigned int offset,
 			       unsigned int length)
 {
-	_enter("{%lu},%u,%u", page->index, offset, length);
+	struct folio *folio = page_folio(page);
+
+	_enter("{%lu},%u,%u", folio_index(folio), offset, length);
 
 	BUG_ON(!PageLocked(page));
 
 	if (PagePrivate(page))
-		afs_invalidate_dirty(page, offset, length);
+		afs_invalidate_dirty(folio, offset, length);
 
-	wait_on_page_fscache(page);
+	folio_wait_fscache(folio);
 	_leave("");
 }
 
@@ -475,30 +482,31 @@ static void afs_invalidatepage(struct page *page, unsigned int offset,
  */
 static int afs_releasepage(struct page *page, gfp_t gfp_flags)
 {
-	struct afs_vnode *vnode = AFS_FS_I(page->mapping->host);
+	struct folio *folio = page_folio(page);
+	struct afs_vnode *vnode = AFS_FS_I(folio_inode(folio));
 
 	_enter("{{%llx:%llu}[%lu],%lx},%x",
-	       vnode->fid.vid, vnode->fid.vnode, page->index, page->flags,
+	       vnode->fid.vid, vnode->fid.vnode, folio_index(folio), folio->flags,
 	       gfp_flags);
 
 	/* deny if page is being written to the cache and the caller hasn't
 	 * elected to wait */
 #ifdef CONFIG_AFS_FSCACHE
-	if (PageFsCache(page)) {
+	if (folio_test_fscache(folio)) {
 		if (!(gfp_flags & __GFP_DIRECT_RECLAIM) || !(gfp_flags & __GFP_FS))
 			return false;
-		wait_on_page_fscache(page);
+		folio_wait_fscache(folio);
 	}
 #endif
 
-	if (PagePrivate(page)) {
-		trace_afs_page_dirty(vnode, tracepoint_string("rel"), page);
-		detach_page_private(page);
+	if (folio_test_private(folio)) {
+		trace_afs_folio_dirty(vnode, tracepoint_string("rel"), folio);
+		folio_detach_private(folio);
 	}
 
-	/* indicate that the page can be released */
+	/* Indicate that the folio can be released */
 	_leave(" = T");
-	return 1;
+	return true;
 }
 
 static void afs_add_open_mmap(struct afs_vnode *vnode)
diff --git a/fs/afs/internal.h b/fs/afs/internal.h
index 9357c53..aa4c0d6 100644
--- a/fs/afs/internal.h
+++ b/fs/afs/internal.h
@@ -876,59 +876,59 @@ struct afs_vnode_cache_aux {
 } __packed;
 
 /*
- * We use page->private to hold the amount of the page that we've written to,
+ * We use folio->private to hold the amount of the folio that we've written to,
  * splitting the field into two parts.  However, we need to represent a range
- * 0...PAGE_SIZE, so we reduce the resolution if the size of the page
+ * 0...FOLIO_SIZE, so we reduce the resolution if the size of the folio
  * exceeds what we can encode.
  */
 #ifdef CONFIG_64BIT
-#define __AFS_PAGE_PRIV_MASK	0x7fffffffUL
-#define __AFS_PAGE_PRIV_SHIFT	32
-#define __AFS_PAGE_PRIV_MMAPPED	0x80000000UL
+#define __AFS_FOLIO_PRIV_MASK		0x7fffffffUL
+#define __AFS_FOLIO_PRIV_SHIFT		32
+#define __AFS_FOLIO_PRIV_MMAPPED	0x80000000UL
 #else
-#define __AFS_PAGE_PRIV_MASK	0x7fffUL
-#define __AFS_PAGE_PRIV_SHIFT	16
-#define __AFS_PAGE_PRIV_MMAPPED	0x8000UL
+#define __AFS_FOLIO_PRIV_MASK		0x7fffUL
+#define __AFS_FOLIO_PRIV_SHIFT		16
+#define __AFS_FOLIO_PRIV_MMAPPED	0x8000UL
 #endif
 
-static inline unsigned int afs_page_dirty_resolution(struct page *page)
+static inline unsigned int afs_folio_dirty_resolution(struct folio *folio)
 {
-	int shift = thp_order(page) + PAGE_SHIFT - (__AFS_PAGE_PRIV_SHIFT - 1);
+	int shift = folio_shift(folio) - (__AFS_FOLIO_PRIV_SHIFT - 1);
 	return (shift > 0) ? shift : 0;
 }
 
-static inline size_t afs_page_dirty_from(struct page *page, unsigned long priv)
+static inline size_t afs_folio_dirty_from(struct folio *folio, unsigned long priv)
 {
-	unsigned long x = priv & __AFS_PAGE_PRIV_MASK;
+	unsigned long x = priv & __AFS_FOLIO_PRIV_MASK;
 
 	/* The lower bound is inclusive */
-	return x << afs_page_dirty_resolution(page);
+	return x << afs_folio_dirty_resolution(folio);
 }
 
-static inline size_t afs_page_dirty_to(struct page *page, unsigned long priv)
+static inline size_t afs_folio_dirty_to(struct folio *folio, unsigned long priv)
 {
-	unsigned long x = (priv >> __AFS_PAGE_PRIV_SHIFT) & __AFS_PAGE_PRIV_MASK;
+	unsigned long x = (priv >> __AFS_FOLIO_PRIV_SHIFT) & __AFS_FOLIO_PRIV_MASK;
 
 	/* The upper bound is immediately beyond the region */
-	return (x + 1) << afs_page_dirty_resolution(page);
+	return (x + 1) << afs_folio_dirty_resolution(folio);
 }
 
-static inline unsigned long afs_page_dirty(struct page *page, size_t from, size_t to)
+static inline unsigned long afs_folio_dirty(struct folio *folio, size_t from, size_t to)
 {
-	unsigned int res = afs_page_dirty_resolution(page);
+	unsigned int res = afs_folio_dirty_resolution(folio);
 	from >>= res;
 	to = (to - 1) >> res;
-	return (to << __AFS_PAGE_PRIV_SHIFT) | from;
+	return (to << __AFS_FOLIO_PRIV_SHIFT) | from;
 }
 
-static inline unsigned long afs_page_dirty_mmapped(unsigned long priv)
+static inline unsigned long afs_folio_dirty_mmapped(unsigned long priv)
 {
-	return priv | __AFS_PAGE_PRIV_MMAPPED;
+	return priv | __AFS_FOLIO_PRIV_MMAPPED;
 }
 
-static inline bool afs_is_page_dirty_mmapped(unsigned long priv)
+static inline bool afs_is_folio_dirty_mmapped(unsigned long priv)
 {
-	return priv & __AFS_PAGE_PRIV_MMAPPED;
+	return priv & __AFS_FOLIO_PRIV_MMAPPED;
 }
 
 #include <trace/events/afs.h>
diff --git a/fs/afs/write.c b/fs/afs/write.c
index 8b1d9c2..ca4909b 100644
--- a/fs/afs/write.c
+++ b/fs/afs/write.c
@@ -32,7 +32,7 @@ int afs_write_begin(struct file *file, struct address_space *mapping,
 		    struct page **_page, void **fsdata)
 {
 	struct afs_vnode *vnode = AFS_FS_I(file_inode(file));
-	struct page *page;
+	struct folio *folio;
 	unsigned long priv;
 	unsigned f, from;
 	unsigned t, to;
@@ -46,12 +46,12 @@ int afs_write_begin(struct file *file, struct address_space *mapping,
 	 * file.  We need to do this before we get a lock on the page in case
 	 * there's more than one writer competing for the same cache block.
 	 */
-	ret = netfs_write_begin(file, mapping, pos, len, flags, &page, fsdata,
+	ret = netfs_write_begin(file, mapping, pos, len, flags, &folio, fsdata,
 				&afs_req_ops, NULL);
 	if (ret < 0)
 		return ret;
 
-	index = page->index;
+	index = folio_index(folio);
 	from = pos - index * PAGE_SIZE;
 	to = from + len;
 
@@ -59,14 +59,14 @@ int afs_write_begin(struct file *file, struct address_space *mapping,
 	/* See if this page is already partially written in a way that we can
 	 * merge the new write with.
 	 */
-	if (PagePrivate(page)) {
-		priv = page_private(page);
-		f = afs_page_dirty_from(page, priv);
-		t = afs_page_dirty_to(page, priv);
+	if (folio_test_private(folio)) {
+		priv = (unsigned long)folio_get_private(folio);
+		f = afs_folio_dirty_from(folio, priv);
+		t = afs_folio_dirty_to(folio, priv);
 		ASSERTCMP(f, <=, t);
 
-		if (PageWriteback(page)) {
-			trace_afs_page_dirty(vnode, tracepoint_string("alrdy"), page);
+		if (folio_test_writeback(folio)) {
+			trace_afs_folio_dirty(vnode, tracepoint_string("alrdy"), folio);
 			goto flush_conflicting_write;
 		}
 		/* If the file is being filled locally, allow inter-write
@@ -78,7 +78,7 @@ int afs_write_begin(struct file *file, struct address_space *mapping,
 			goto flush_conflicting_write;
 	}
 
-	*_page = page;
+	*_page = &folio->page;
 	_leave(" = 0");
 	return 0;
 
@@ -87,17 +87,17 @@ int afs_write_begin(struct file *file, struct address_space *mapping,
 	 */
 flush_conflicting_write:
 	_debug("flush conflict");
-	ret = write_one_page(page);
+	ret = folio_write_one(folio);
 	if (ret < 0)
 		goto error;
 
-	ret = lock_page_killable(page);
+	ret = folio_lock_killable(folio);
 	if (ret < 0)
 		goto error;
 	goto try_again;
 
 error:
-	put_page(page);
+	folio_put(folio);
 	_leave(" = %d", ret);
 	return ret;
 }
@@ -107,24 +107,25 @@ int afs_write_begin(struct file *file, struct address_space *mapping,
  */
 int afs_write_end(struct file *file, struct address_space *mapping,
 		  loff_t pos, unsigned len, unsigned copied,
-		  struct page *page, void *fsdata)
+		  struct page *subpage, void *fsdata)
 {
+	struct folio *folio = page_folio(subpage);
 	struct afs_vnode *vnode = AFS_FS_I(file_inode(file));
 	unsigned long priv;
-	unsigned int f, from = pos & (thp_size(page) - 1);
+	unsigned int f, from = offset_in_folio(folio, pos);
 	unsigned int t, to = from + copied;
 	loff_t i_size, maybe_i_size;
 
 	_enter("{%llx:%llu},{%lx}",
-	       vnode->fid.vid, vnode->fid.vnode, page->index);
+	       vnode->fid.vid, vnode->fid.vnode, folio_index(folio));
 
-	if (!PageUptodate(page)) {
+	if (!folio_test_uptodate(folio)) {
 		if (copied < len) {
 			copied = 0;
 			goto out;
 		}
 
-		SetPageUptodate(page);
+		folio_mark_uptodate(folio);
 	}
 
 	if (copied == 0)
@@ -141,29 +142,29 @@ int afs_write_end(struct file *file, struct address_space *mapping,
 		write_sequnlock(&vnode->cb_lock);
 	}
 
-	if (PagePrivate(page)) {
-		priv = page_private(page);
-		f = afs_page_dirty_from(page, priv);
-		t = afs_page_dirty_to(page, priv);
+	if (folio_test_private(folio)) {
+		priv = (unsigned long)folio_get_private(folio);
+		f = afs_folio_dirty_from(folio, priv);
+		t = afs_folio_dirty_to(folio, priv);
 		if (from < f)
 			f = from;
 		if (to > t)
 			t = to;
-		priv = afs_page_dirty(page, f, t);
-		set_page_private(page, priv);
-		trace_afs_page_dirty(vnode, tracepoint_string("dirty+"), page);
+		priv = afs_folio_dirty(folio, f, t);
+		folio_change_private(folio, (void *)priv);
+		trace_afs_folio_dirty(vnode, tracepoint_string("dirty+"), folio);
 	} else {
-		priv = afs_page_dirty(page, from, to);
-		attach_page_private(page, (void *)priv);
-		trace_afs_page_dirty(vnode, tracepoint_string("dirty"), page);
+		priv = afs_folio_dirty(folio, from, to);
+		folio_attach_private(folio, (void *)priv);
+		trace_afs_folio_dirty(vnode, tracepoint_string("dirty"), folio);
 	}
 
-	if (set_page_dirty(page))
-		_debug("dirtied %lx", page->index);
+	if (folio_mark_dirty(folio))
+		_debug("dirtied %lx", folio_index(folio));
 
 out:
-	unlock_page(page);
-	put_page(page);
+	folio_unlock(folio);
+	folio_put(folio);
 	return copied;
 }
 
@@ -174,40 +175,32 @@ static void afs_kill_pages(struct address_space *mapping,
 			   loff_t start, loff_t len)
 {
 	struct afs_vnode *vnode = AFS_FS_I(mapping->host);
-	struct pagevec pv;
-	unsigned int loop, psize;
+	struct folio *folio;
+	pgoff_t index = start / PAGE_SIZE;
+	pgoff_t last = (start + len - 1) / PAGE_SIZE, next;
 
 	_enter("{%llx:%llu},%llx @%llx",
 	       vnode->fid.vid, vnode->fid.vnode, len, start);
 
-	pagevec_init(&pv);
-
 	do {
-		_debug("kill %llx @%llx", len, start);
+		_debug("kill %lx (to %lx)", index, last);
 
-		pv.nr = find_get_pages_contig(mapping, start / PAGE_SIZE,
-					      PAGEVEC_SIZE, pv.pages);
-		if (pv.nr == 0)
-			break;
-
-		for (loop = 0; loop < pv.nr; loop++) {
-			struct page *page = pv.pages[loop];
-
-			if (page->index * PAGE_SIZE >= start + len)
-				break;
-
-			psize = thp_size(page);
-			start += psize;
-			len -= psize;
-			ClearPageUptodate(page);
-			end_page_writeback(page);
-			lock_page(page);
-			generic_error_remove_page(mapping, page);
-			unlock_page(page);
+		folio = filemap_get_folio(mapping, index);
+		if (!folio) {
+			next = index + 1;
+			continue;
 		}
 
-		__pagevec_release(&pv);
-	} while (len > 0);
+		next = folio_next_index(folio);
+
+		folio_clear_uptodate(folio);
+		folio_end_writeback(folio);
+		folio_lock(folio);
+		generic_error_remove_page(mapping, &folio->page);
+		folio_unlock(folio);
+		folio_put(folio);
+
+	} while (index = next, index <= last);
 
 	_leave("");
 }
@@ -220,37 +213,27 @@ static void afs_redirty_pages(struct writeback_control *wbc,
 			      loff_t start, loff_t len)
 {
 	struct afs_vnode *vnode = AFS_FS_I(mapping->host);
-	struct pagevec pv;
-	unsigned int loop, psize;
+	struct folio *folio;
+	pgoff_t index = start / PAGE_SIZE;
+	pgoff_t last = (start + len - 1) / PAGE_SIZE, next;
 
 	_enter("{%llx:%llu},%llx @%llx",
 	       vnode->fid.vid, vnode->fid.vnode, len, start);
 
-	pagevec_init(&pv);
-
 	do {
 		_debug("redirty %llx @%llx", len, start);
 
-		pv.nr = find_get_pages_contig(mapping, start / PAGE_SIZE,
-					      PAGEVEC_SIZE, pv.pages);
-		if (pv.nr == 0)
-			break;
-
-		for (loop = 0; loop < pv.nr; loop++) {
-			struct page *page = pv.pages[loop];
-
-			if (page->index * PAGE_SIZE >= start + len)
-				break;
-
-			psize = thp_size(page);
-			start += psize;
-			len -= psize;
-			redirty_page_for_writepage(wbc, page);
-			end_page_writeback(page);
+		folio = filemap_get_folio(mapping, index);
+		if (!folio) {
+			next = index + 1;
+			continue;
 		}
 
-		__pagevec_release(&pv);
-	} while (len > 0);
+		next = index + folio_nr_pages(folio);
+		folio_redirty_for_writepage(wbc, folio);
+		folio_end_writeback(folio);
+		folio_put(folio);
+	} while (index = next, index <= last);
 
 	_leave("");
 }
@@ -261,7 +244,7 @@ static void afs_redirty_pages(struct writeback_control *wbc,
 static void afs_pages_written_back(struct afs_vnode *vnode, loff_t start, unsigned int len)
 {
 	struct address_space *mapping = vnode->vfs_inode.i_mapping;
-	struct page *page;
+	struct folio *folio;
 	pgoff_t end;
 
 	XA_STATE(xas, &mapping->i_pages, start / PAGE_SIZE);
@@ -272,15 +255,16 @@ static void afs_pages_written_back(struct afs_vnode *vnode, loff_t start, unsign
 	rcu_read_lock();
 
 	end = (start + len - 1) / PAGE_SIZE;
-	xas_for_each(&xas, page, end) {
-		if (!PageWriteback(page)) {
-			kdebug("bad %x @%llx page %lx %lx", len, start, page->index, end);
-			ASSERT(PageWriteback(page));
+	xas_for_each(&xas, folio, end) {
+		if (!folio_test_writeback(folio)) {
+			kdebug("bad %x @%llx page %lx %lx",
+			       len, start, folio_index(folio), end);
+			ASSERT(folio_test_writeback(folio));
 		}
 
-		trace_afs_page_dirty(vnode, tracepoint_string("clear"), page);
-		detach_page_private(page);
-		page_endio(page, true, 0);
+		trace_afs_folio_dirty(vnode, tracepoint_string("clear"), folio);
+		folio_detach_private(folio);
+		folio_end_writeback(folio);
 	}
 
 	rcu_read_unlock();
@@ -437,7 +421,7 @@ static void afs_extend_writeback(struct address_space *mapping,
 				 unsigned int *_len)
 {
 	struct pagevec pvec;
-	struct page *page;
+	struct folio *folio;
 	unsigned long priv;
 	unsigned int psize, filler = 0;
 	unsigned int f, t;
@@ -456,43 +440,43 @@ static void afs_extend_writeback(struct address_space *mapping,
 		 */
 		rcu_read_lock();
 
-		xas_for_each(&xas, page, ULONG_MAX) {
+		xas_for_each(&xas, folio, ULONG_MAX) {
 			stop = true;
-			if (xas_retry(&xas, page))
+			if (xas_retry(&xas, folio))
 				continue;
-			if (xa_is_value(page))
+			if (xa_is_value(folio))
 				break;
-			if (page->index != index)
+			if (folio_index(folio) != index)
 				break;
 
-			if (!page_cache_get_speculative(page)) {
+			if (!folio_try_get_rcu(folio)) {
 				xas_reset(&xas);
 				continue;
 			}
 
 			/* Has the page moved or been split? */
-			if (unlikely(page != xas_reload(&xas))) {
-				put_page(page);
+			if (unlikely(folio != xas_reload(&xas))) {
+				folio_put(folio);
 				break;
 			}
 
-			if (!trylock_page(page)) {
-				put_page(page);
+			if (!folio_trylock(folio)) {
+				folio_put(folio);
 				break;
 			}
-			if (!PageDirty(page) || PageWriteback(page)) {
-				unlock_page(page);
-				put_page(page);
+			if (!folio_test_dirty(folio) || folio_test_writeback(folio)) {
+				folio_unlock(folio);
+				folio_put(folio);
 				break;
 			}
 
-			psize = thp_size(page);
-			priv = page_private(page);
-			f = afs_page_dirty_from(page, priv);
-			t = afs_page_dirty_to(page, priv);
+			psize = folio_size(folio);
+			priv = (unsigned long)folio_get_private(folio);
+			f = afs_folio_dirty_from(folio, priv);
+			t = afs_folio_dirty_to(folio, priv);
 			if (f != 0 && !new_content) {
-				unlock_page(page);
-				put_page(page);
+				folio_unlock(folio);
+				folio_put(folio);
 				break;
 			}
 
@@ -503,8 +487,8 @@ static void afs_extend_writeback(struct address_space *mapping,
 			else if (t == psize || new_content)
 				stop = false;
 
-			index += thp_nr_pages(page);
-			if (!pagevec_add(&pvec, page))
+			index += folio_nr_pages(folio);
+			if (!pagevec_add(&pvec, &folio->page))
 				break;
 			if (stop)
 				break;
@@ -521,16 +505,16 @@ static void afs_extend_writeback(struct address_space *mapping,
 			break;
 
 		for (i = 0; i < pagevec_count(&pvec); i++) {
-			page = pvec.pages[i];
-			trace_afs_page_dirty(vnode, tracepoint_string("store+"), page);
+			folio = page_folio(pvec.pages[i]);
+			trace_afs_folio_dirty(vnode, tracepoint_string("store+"), folio);
 
-			if (!clear_page_dirty_for_io(page))
+			if (!folio_clear_dirty_for_io(folio))
 				BUG();
-			if (test_set_page_writeback(page))
+			if (folio_start_writeback(folio))
 				BUG();
 
-			*_count -= thp_nr_pages(page);
-			unlock_page(page);
+			*_count -= folio_nr_pages(folio);
+			folio_unlock(folio);
 		}
 
 		pagevec_release(&pvec);
@@ -544,10 +528,10 @@ static void afs_extend_writeback(struct address_space *mapping,
  * Synchronously write back the locked page and any subsequent non-locked dirty
  * pages.
  */
-static ssize_t afs_write_back_from_locked_page(struct address_space *mapping,
-					       struct writeback_control *wbc,
-					       struct page *page,
-					       loff_t start, loff_t end)
+static ssize_t afs_write_back_from_locked_folio(struct address_space *mapping,
+						struct writeback_control *wbc,
+						struct folio *folio,
+						loff_t start, loff_t end)
 {
 	struct afs_vnode *vnode = AFS_FS_I(mapping->host);
 	struct iov_iter iter;
@@ -558,22 +542,22 @@ static ssize_t afs_write_back_from_locked_page(struct address_space *mapping,
 	long count = wbc->nr_to_write;
 	int ret;
 
-	_enter(",%lx,%llx-%llx", page->index, start, end);
+	_enter(",%lx,%llx-%llx", folio_index(folio), start, end);
 
-	if (test_set_page_writeback(page))
+	if (folio_start_writeback(folio))
 		BUG();
 
-	count -= thp_nr_pages(page);
+	count -= folio_nr_pages(folio);
 
 	/* Find all consecutive lockable dirty pages that have contiguous
 	 * written regions, stopping when we find a page that is not
 	 * immediately lockable, is not dirty or is missing, or we reach the
 	 * end of the range.
 	 */
-	priv = page_private(page);
-	offset = afs_page_dirty_from(page, priv);
-	to = afs_page_dirty_to(page, priv);
-	trace_afs_page_dirty(vnode, tracepoint_string("store"), page);
+	priv = (unsigned long)folio_get_private(folio);
+	offset = afs_folio_dirty_from(folio, priv);
+	to = afs_folio_dirty_to(folio, priv);
+	trace_afs_folio_dirty(vnode, tracepoint_string("store"), folio);
 
 	len = to - offset;
 	start += offset;
@@ -586,7 +570,7 @@ static ssize_t afs_write_back_from_locked_page(struct address_space *mapping,
 		max_len = min_t(unsigned long long, max_len, i_size - start);
 
 		if (len < max_len &&
-		    (to == thp_size(page) || new_content))
+		    (to == folio_size(folio) || new_content))
 			afs_extend_writeback(mapping, vnode, &count,
 					     start, max_len, new_content, &len);
 		len = min_t(loff_t, len, max_len);
@@ -596,7 +580,7 @@ static ssize_t afs_write_back_from_locked_page(struct address_space *mapping,
 	 * set; the first page is still locked at this point, but all the rest
 	 * have been unlocked.
 	 */
-	unlock_page(page);
+	folio_unlock(folio);
 
 	if (start < i_size) {
 		_debug("write back %x @%llx [%llx]", len, start, i_size);
@@ -657,16 +641,17 @@ static ssize_t afs_write_back_from_locked_page(struct address_space *mapping,
  * write a page back to the server
  * - the caller locked the page for us
  */
-int afs_writepage(struct page *page, struct writeback_control *wbc)
+int afs_writepage(struct page *subpage, struct writeback_control *wbc)
 {
+	struct folio *folio = page_folio(subpage);
 	ssize_t ret;
 	loff_t start;
 
-	_enter("{%lx},", page->index);
+	_enter("{%lx},", folio_index(folio));
 
-	start = page->index * PAGE_SIZE;
-	ret = afs_write_back_from_locked_page(page->mapping, wbc, page,
-					      start, LLONG_MAX - start);
+	start = folio_index(folio) * PAGE_SIZE;
+	ret = afs_write_back_from_locked_folio(folio_mapping(folio), wbc,
+					       folio, start, LLONG_MAX - start);
 	if (ret < 0) {
 		_leave(" = %zd", ret);
 		return ret;
@@ -683,7 +668,8 @@ static int afs_writepages_region(struct address_space *mapping,
 				 struct writeback_control *wbc,
 				 loff_t start, loff_t end, loff_t *_next)
 {
-	struct page *page;
+	struct folio *folio;
+	struct page *head_page;
 	ssize_t ret;
 	int n;
 
@@ -693,13 +679,14 @@ static int afs_writepages_region(struct address_space *mapping,
 		pgoff_t index = start / PAGE_SIZE;
 
 		n = find_get_pages_range_tag(mapping, &index, end / PAGE_SIZE,
-					     PAGECACHE_TAG_DIRTY, 1, &page);
+					     PAGECACHE_TAG_DIRTY, 1, &head_page);
 		if (!n)
 			break;
 
-		start = (loff_t)page->index * PAGE_SIZE; /* May regress with THPs */
+		folio = page_folio(head_page);
+		start = folio_pos(folio); /* May regress with THPs */
 
-		_debug("wback %lx", page->index);
+		_debug("wback %lx", folio_index(folio));
 
 		/* At this point we hold neither the i_pages lock nor the
 		 * page lock: the page may be truncated or invalidated
@@ -707,37 +694,38 @@ static int afs_writepages_region(struct address_space *mapping,
 		 * back from swapper_space to tmpfs file mapping
 		 */
 		if (wbc->sync_mode != WB_SYNC_NONE) {
-			ret = lock_page_killable(page);
+			ret = folio_lock_killable(folio);
 			if (ret < 0) {
-				put_page(page);
+				folio_put(folio);
 				return ret;
 			}
 		} else {
-			if (!trylock_page(page)) {
-				put_page(page);
+			if (!folio_trylock(folio)) {
+				folio_put(folio);
 				return 0;
 			}
 		}
 
-		if (page->mapping != mapping || !PageDirty(page)) {
-			start += thp_size(page);
-			unlock_page(page);
-			put_page(page);
+		if (folio_mapping(folio) != mapping ||
+		    !folio_test_dirty(folio)) {
+			start += folio_size(folio);
+			folio_unlock(folio);
+			folio_put(folio);
 			continue;
 		}
 
-		if (PageWriteback(page)) {
-			unlock_page(page);
+		if (folio_test_writeback(folio)) {
+			folio_unlock(folio);
 			if (wbc->sync_mode != WB_SYNC_NONE)
-				wait_on_page_writeback(page);
-			put_page(page);
+				folio_wait_writeback(folio);
+			folio_put(folio);
 			continue;
 		}
 
-		if (!clear_page_dirty_for_io(page))
+		if (!folio_clear_dirty_for_io(folio))
 			BUG();
-		ret = afs_write_back_from_locked_page(mapping, wbc, page, start, end);
-		put_page(page);
+		ret = afs_write_back_from_locked_folio(mapping, wbc, folio, start, end);
+		folio_put(folio);
 		if (ret < 0) {
 			_leave(" = %zd", ret);
 			return ret;
@@ -862,7 +850,6 @@ int afs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 vm_fault_t afs_page_mkwrite(struct vm_fault *vmf)
 {
 	struct folio *folio = page_folio(vmf->page);
-	struct page *page = &folio->page;
 	struct file *file = vmf->vma->vm_file;
 	struct inode *inode = file_inode(file);
 	struct afs_vnode *vnode = AFS_FS_I(inode);
@@ -870,7 +857,7 @@ vm_fault_t afs_page_mkwrite(struct vm_fault *vmf)
 	unsigned long priv;
 	vm_fault_t ret = VM_FAULT_RETRY;
 
-	_enter("{{%llx:%llu}},{%lx}", vnode->fid.vid, vnode->fid.vnode, page->index);
+	_enter("{{%llx:%llu}},{%lx}", vnode->fid.vid, vnode->fid.vnode, folio_index(folio));
 
 	afs_validate(vnode, af->key);
 
@@ -880,18 +867,18 @@ vm_fault_t afs_page_mkwrite(struct vm_fault *vmf)
 	 * be modified.  We then assume the entire page will need writing back.
 	 */
 #ifdef CONFIG_AFS_FSCACHE
-	if (PageFsCache(page) &&
-	    wait_on_page_fscache_killable(page) < 0)
+	if (folio_test_fscache(folio) &&
+	    folio_wait_fscache_killable(folio) < 0)
 		goto out;
 #endif
 
 	if (folio_wait_writeback_killable(folio))
 		goto out;
 
-	if (lock_page_killable(page) < 0)
+	if (folio_lock_killable(folio) < 0)
 		goto out;
 
-	/* We mustn't change page->private until writeback is complete as that
+	/* We mustn't change folio->private until writeback is complete as that
 	 * details the portion of the page we need to write back and we might
 	 * need to redirty the page if there's a problem.
 	 */
@@ -900,14 +887,14 @@ vm_fault_t afs_page_mkwrite(struct vm_fault *vmf)
 		goto out;
 	}
 
-	priv = afs_page_dirty(page, 0, thp_size(page));
-	priv = afs_page_dirty_mmapped(priv);
-	if (PagePrivate(page)) {
-		set_page_private(page, priv);
-		trace_afs_page_dirty(vnode, tracepoint_string("mkwrite+"), page);
+	priv = afs_folio_dirty(folio, 0, folio_size(folio));
+	priv = afs_folio_dirty_mmapped(priv);
+	if (folio_test_private(folio)) {
+		folio_change_private(folio, (void *)priv);
+		trace_afs_folio_dirty(vnode, tracepoint_string("mkwrite+"), folio);
 	} else {
-		attach_page_private(page, (void *)priv);
-		trace_afs_page_dirty(vnode, tracepoint_string("mkwrite"), page);
+		folio_attach_private(folio, (void *)priv);
+		trace_afs_folio_dirty(vnode, tracepoint_string("mkwrite"), folio);
 	}
 	file_update_time(file);
 
@@ -948,38 +935,38 @@ void afs_prune_wb_keys(struct afs_vnode *vnode)
 /*
  * Clean up a page during invalidation.
  */
-int afs_launder_page(struct page *page)
+int afs_launder_page(struct page *subpage)
 {
-	struct address_space *mapping = page->mapping;
-	struct afs_vnode *vnode = AFS_FS_I(mapping->host);
+	struct folio *folio = page_folio(subpage);
+	struct afs_vnode *vnode = AFS_FS_I(folio_inode(folio));
 	struct iov_iter iter;
 	struct bio_vec bv[1];
 	unsigned long priv;
 	unsigned int f, t;
 	int ret = 0;
 
-	_enter("{%lx}", page->index);
+	_enter("{%lx}", folio_index(folio));
 
-	priv = page_private(page);
-	if (clear_page_dirty_for_io(page)) {
+	priv = (unsigned long)folio_get_private(folio);
+	if (folio_clear_dirty_for_io(folio)) {
 		f = 0;
-		t = thp_size(page);
-		if (PagePrivate(page)) {
-			f = afs_page_dirty_from(page, priv);
-			t = afs_page_dirty_to(page, priv);
+		t = folio_size(folio);
+		if (folio_test_private(folio)) {
+			f = afs_folio_dirty_from(folio, priv);
+			t = afs_folio_dirty_to(folio, priv);
 		}
 
-		bv[0].bv_page = page;
+		bv[0].bv_page = &folio->page;
 		bv[0].bv_offset = f;
 		bv[0].bv_len = t - f;
 		iov_iter_bvec(&iter, WRITE, bv, 1, bv[0].bv_len);
 
-		trace_afs_page_dirty(vnode, tracepoint_string("launder"), page);
-		ret = afs_store_data(vnode, &iter, page_offset(page) + f, true);
+		trace_afs_folio_dirty(vnode, tracepoint_string("launder"), folio);
+		ret = afs_store_data(vnode, &iter, folio_pos(folio) + f, true);
 	}
 
-	trace_afs_page_dirty(vnode, tracepoint_string("laundered"), page);
-	detach_page_private(page);
-	wait_on_page_fscache(page);
+	trace_afs_folio_dirty(vnode, tracepoint_string("laundered"), folio);
+	folio_detach_private(folio);
+	folio_wait_fscache(folio);
 	return ret;
 }
diff --git a/fs/btrfs/zstd.c b/fs/btrfs/zstd.c
index f06b680..fc42dd0 100644
--- a/fs/btrfs/zstd.c
+++ b/fs/btrfs/zstd.c
@@ -28,10 +28,10 @@
 /* 307s to avoid pathologically clashing with transaction commit */
 #define ZSTD_BTRFS_RECLAIM_JIFFIES (307 * HZ)
 
-static ZSTD_parameters zstd_get_btrfs_parameters(unsigned int level,
+static zstd_parameters zstd_get_btrfs_parameters(unsigned int level,
 						 size_t src_len)
 {
-	ZSTD_parameters params = ZSTD_getParams(level, src_len, 0);
+	zstd_parameters params = zstd_get_params(level, src_len);
 
 	if (params.cParams.windowLog > ZSTD_BTRFS_MAX_WINDOWLOG)
 		params.cParams.windowLog = ZSTD_BTRFS_MAX_WINDOWLOG;
@@ -48,8 +48,8 @@ struct workspace {
 	unsigned long last_used; /* jiffies */
 	struct list_head list;
 	struct list_head lru_list;
-	ZSTD_inBuffer in_buf;
-	ZSTD_outBuffer out_buf;
+	zstd_in_buffer in_buf;
+	zstd_out_buffer out_buf;
 };
 
 /*
@@ -155,12 +155,12 @@ static void zstd_calc_ws_mem_sizes(void)
 	unsigned int level;
 
 	for (level = 1; level <= ZSTD_BTRFS_MAX_LEVEL; level++) {
-		ZSTD_parameters params =
+		zstd_parameters params =
 			zstd_get_btrfs_parameters(level, ZSTD_BTRFS_MAX_INPUT);
 		size_t level_size =
 			max_t(size_t,
-			      ZSTD_CStreamWorkspaceBound(params.cParams),
-			      ZSTD_DStreamWorkspaceBound(ZSTD_BTRFS_MAX_INPUT));
+			      zstd_cstream_workspace_bound(&params.cParams),
+			      zstd_dstream_workspace_bound(ZSTD_BTRFS_MAX_INPUT));
 
 		max_size = max_t(size_t, max_size, level_size);
 		zstd_ws_mem_sizes[level - 1] = max_size;
@@ -371,7 +371,7 @@ int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
 		unsigned long *total_in, unsigned long *total_out)
 {
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
-	ZSTD_CStream *stream;
+	zstd_cstream *stream;
 	int ret = 0;
 	int nr_pages = 0;
 	struct page *in_page = NULL;  /* The current page to read */
@@ -381,7 +381,7 @@ int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
 	unsigned long len = *total_out;
 	const unsigned long nr_dest_pages = *out_pages;
 	unsigned long max_out = nr_dest_pages * PAGE_SIZE;
-	ZSTD_parameters params = zstd_get_btrfs_parameters(workspace->req_level,
+	zstd_parameters params = zstd_get_btrfs_parameters(workspace->req_level,
 							   len);
 
 	*out_pages = 0;
@@ -389,10 +389,10 @@ int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
 	*total_in = 0;
 
 	/* Initialize the stream */
-	stream = ZSTD_initCStream(params, len, workspace->mem,
+	stream = zstd_init_cstream(&params, len, workspace->mem,
 			workspace->size);
 	if (!stream) {
-		pr_warn("BTRFS: ZSTD_initCStream failed\n");
+		pr_warn("BTRFS: zstd_init_cstream failed\n");
 		ret = -EIO;
 		goto out;
 	}
@@ -418,11 +418,11 @@ int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
 	while (1) {
 		size_t ret2;
 
-		ret2 = ZSTD_compressStream(stream, &workspace->out_buf,
+		ret2 = zstd_compress_stream(stream, &workspace->out_buf,
 				&workspace->in_buf);
-		if (ZSTD_isError(ret2)) {
-			pr_debug("BTRFS: ZSTD_compressStream returned %d\n",
-					ZSTD_getErrorCode(ret2));
+		if (zstd_is_error(ret2)) {
+			pr_debug("BTRFS: zstd_compress_stream returned %d\n",
+					zstd_get_error_code(ret2));
 			ret = -EIO;
 			goto out;
 		}
@@ -487,10 +487,10 @@ int zstd_compress_pages(struct list_head *ws, struct address_space *mapping,
 	while (1) {
 		size_t ret2;
 
-		ret2 = ZSTD_endStream(stream, &workspace->out_buf);
-		if (ZSTD_isError(ret2)) {
-			pr_debug("BTRFS: ZSTD_endStream returned %d\n",
-					ZSTD_getErrorCode(ret2));
+		ret2 = zstd_end_stream(stream, &workspace->out_buf);
+		if (zstd_is_error(ret2)) {
+			pr_debug("BTRFS: zstd_end_stream returned %d\n",
+					zstd_get_error_code(ret2));
 			ret = -EIO;
 			goto out;
 		}
@@ -548,17 +548,17 @@ int zstd_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
 	struct page **pages_in = cb->compressed_pages;
 	size_t srclen = cb->compressed_len;
-	ZSTD_DStream *stream;
+	zstd_dstream *stream;
 	int ret = 0;
 	unsigned long page_in_index = 0;
 	unsigned long total_pages_in = DIV_ROUND_UP(srclen, PAGE_SIZE);
 	unsigned long buf_start;
 	unsigned long total_out = 0;
 
-	stream = ZSTD_initDStream(
+	stream = zstd_init_dstream(
 			ZSTD_BTRFS_MAX_INPUT, workspace->mem, workspace->size);
 	if (!stream) {
-		pr_debug("BTRFS: ZSTD_initDStream failed\n");
+		pr_debug("BTRFS: zstd_init_dstream failed\n");
 		ret = -EIO;
 		goto done;
 	}
@@ -574,11 +574,11 @@ int zstd_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 	while (1) {
 		size_t ret2;
 
-		ret2 = ZSTD_decompressStream(stream, &workspace->out_buf,
+		ret2 = zstd_decompress_stream(stream, &workspace->out_buf,
 				&workspace->in_buf);
-		if (ZSTD_isError(ret2)) {
-			pr_debug("BTRFS: ZSTD_decompressStream returned %d\n",
-					ZSTD_getErrorCode(ret2));
+		if (zstd_is_error(ret2)) {
+			pr_debug("BTRFS: zstd_decompress_stream returned %d\n",
+					zstd_get_error_code(ret2));
 			ret = -EIO;
 			goto done;
 		}
@@ -624,16 +624,16 @@ int zstd_decompress(struct list_head *ws, unsigned char *data_in,
 		size_t destlen)
 {
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
-	ZSTD_DStream *stream;
+	zstd_dstream *stream;
 	int ret = 0;
 	size_t ret2;
 	unsigned long total_out = 0;
 	unsigned long pg_offset = 0;
 
-	stream = ZSTD_initDStream(
+	stream = zstd_init_dstream(
 			ZSTD_BTRFS_MAX_INPUT, workspace->mem, workspace->size);
 	if (!stream) {
-		pr_warn("BTRFS: ZSTD_initDStream failed\n");
+		pr_warn("BTRFS: zstd_init_dstream failed\n");
 		ret = -EIO;
 		goto finish;
 	}
@@ -657,15 +657,15 @@ int zstd_decompress(struct list_head *ws, unsigned char *data_in,
 
 		/* Check if the frame is over and we still need more input */
 		if (ret2 == 0) {
-			pr_debug("BTRFS: ZSTD_decompressStream ended early\n");
+			pr_debug("BTRFS: zstd_decompress_stream ended early\n");
 			ret = -EIO;
 			goto finish;
 		}
-		ret2 = ZSTD_decompressStream(stream, &workspace->out_buf,
+		ret2 = zstd_decompress_stream(stream, &workspace->out_buf,
 				&workspace->in_buf);
-		if (ZSTD_isError(ret2)) {
-			pr_debug("BTRFS: ZSTD_decompressStream returned %d\n",
-					ZSTD_getErrorCode(ret2));
+		if (zstd_is_error(ret2)) {
+			pr_debug("BTRFS: zstd_decompress_stream returned %d\n",
+					zstd_get_error_code(ret2));
 			ret = -EIO;
 			goto finish;
 		}
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index 99b80b5..e53c854 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -63,7 +63,7 @@
 	 (CONGESTION_ON_THRESH(congestion_kb) >> 2))
 
 static int ceph_netfs_check_write_begin(struct file *file, loff_t pos, unsigned int len,
-					struct page *page, void **_fsdata);
+					struct folio *folio, void **_fsdata);
 
 static inline struct ceph_snap_context *page_snap_context(struct page *page)
 {
@@ -317,13 +317,14 @@ static const struct netfs_read_request_ops ceph_netfs_read_ops = {
 };
 
 /* read a single page, without unlocking it. */
-static int ceph_readpage(struct file *file, struct page *page)
+static int ceph_readpage(struct file *file, struct page *subpage)
 {
+	struct folio *folio = page_folio(subpage);
 	struct inode *inode = file_inode(file);
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	struct ceph_vino vino = ceph_vino(inode);
-	u64 off = page_offset(page);
-	u64 len = thp_size(page);
+	size_t len = folio_size(folio);
+	u64 off = folio_file_pos(folio);
 
 	if (ci->i_inline_version != CEPH_INLINE_NONE) {
 		/*
@@ -331,19 +332,19 @@ static int ceph_readpage(struct file *file, struct page *page)
 		 * into page cache while getting Fcr caps.
 		 */
 		if (off == 0) {
-			unlock_page(page);
+			folio_unlock(folio);
 			return -EINVAL;
 		}
-		zero_user_segment(page, 0, thp_size(page));
-		SetPageUptodate(page);
-		unlock_page(page);
+		zero_user_segment(&folio->page, 0, folio_size(folio));
+		folio_mark_uptodate(folio);
+		folio_unlock(folio);
 		return 0;
 	}
 
-	dout("readpage ino %llx.%llx file %p off %llu len %llu page %p index %lu\n",
-	     vino.ino, vino.snap, file, off, len, page, page->index);
+	dout("readpage ino %llx.%llx file %p off %llu len %zu folio %p index %lu\n",
+	     vino.ino, vino.snap, file, off, len, folio, folio_index(folio));
 
-	return netfs_readpage(file, page, &ceph_netfs_read_ops, NULL);
+	return netfs_readpage(file, folio, &ceph_netfs_read_ops, NULL);
 }
 
 static void ceph_readahead(struct readahead_control *ractl)
@@ -724,7 +725,7 @@ static int ceph_writepages_start(struct address_space *mapping,
 	     wbc->sync_mode == WB_SYNC_NONE ? "NONE" :
 	     (wbc->sync_mode == WB_SYNC_ALL ? "ALL" : "HOLD"));
 
-	if (READ_ONCE(fsc->mount_state) >= CEPH_MOUNT_SHUTDOWN) {
+	if (ceph_inode_is_shutdown(inode)) {
 		if (ci->i_wrbuffer_ref > 0) {
 			pr_warn_ratelimited(
 				"writepage_start %p %lld forced umount\n",
@@ -1145,12 +1146,12 @@ static struct ceph_snap_context *
 ceph_find_incompatible(struct page *page)
 {
 	struct inode *inode = page->mapping->host;
-	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
 	struct ceph_inode_info *ci = ceph_inode(inode);
 
-	if (READ_ONCE(fsc->mount_state) >= CEPH_MOUNT_SHUTDOWN) {
-		dout(" page %p forced umount\n", page);
-		return ERR_PTR(-EIO);
+	if (ceph_inode_is_shutdown(inode)) {
+		dout(" page %p %llx:%llx is shutdown\n", page,
+		     ceph_vinop(inode));
+		return ERR_PTR(-ESTALE);
 	}
 
 	for (;;) {
@@ -1187,18 +1188,18 @@ ceph_find_incompatible(struct page *page)
 }
 
 static int ceph_netfs_check_write_begin(struct file *file, loff_t pos, unsigned int len,
-					struct page *page, void **_fsdata)
+					struct folio *folio, void **_fsdata)
 {
 	struct inode *inode = file_inode(file);
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	struct ceph_snap_context *snapc;
 
-	snapc = ceph_find_incompatible(page);
+	snapc = ceph_find_incompatible(folio_page(folio, 0));
 	if (snapc) {
 		int r;
 
-		unlock_page(page);
-		put_page(page);
+		folio_unlock(folio);
+		folio_put(folio);
 		if (IS_ERR(snapc))
 			return PTR_ERR(snapc);
 
@@ -1216,12 +1217,12 @@ static int ceph_netfs_check_write_begin(struct file *file, loff_t pos, unsigned
  * clean, or already dirty within the same snap context.
  */
 static int ceph_write_begin(struct file *file, struct address_space *mapping,
-			    loff_t pos, unsigned len, unsigned flags,
+			    loff_t pos, unsigned len, unsigned aop_flags,
 			    struct page **pagep, void **fsdata)
 {
 	struct inode *inode = file_inode(file);
 	struct ceph_inode_info *ci = ceph_inode(inode);
-	struct page *page = NULL;
+	struct folio *folio = NULL;
 	pgoff_t index = pos >> PAGE_SHIFT;
 	int r;
 
@@ -1230,39 +1231,43 @@ static int ceph_write_begin(struct file *file, struct address_space *mapping,
 	 * for inline_version sent to the MDS.
 	 */
 	if (ci->i_inline_version != CEPH_INLINE_NONE) {
-		page = grab_cache_page_write_begin(mapping, index, flags);
-		if (!page)
+		unsigned int fgp_flags = FGP_LOCK | FGP_WRITE | FGP_CREAT | FGP_STABLE;
+		if (aop_flags & AOP_FLAG_NOFS)
+			fgp_flags |= FGP_NOFS;
+		folio = __filemap_get_folio(mapping, index, fgp_flags,
+					    mapping_gfp_mask(mapping));
+		if (!folio)
 			return -ENOMEM;
 
 		/*
 		 * The inline_version on a new inode is set to 1. If that's the
-		 * case, then the page is brand new and isn't yet Uptodate.
+		 * case, then the folio is brand new and isn't yet Uptodate.
 		 */
 		r = 0;
 		if (index == 0 && ci->i_inline_version != 1) {
-			if (!PageUptodate(page)) {
+			if (!folio_test_uptodate(folio)) {
 				WARN_ONCE(1, "ceph: write_begin called on still-inlined inode (inline_version %llu)!\n",
 					  ci->i_inline_version);
 				r = -EINVAL;
 			}
 			goto out;
 		}
-		zero_user_segment(page, 0, thp_size(page));
-		SetPageUptodate(page);
+		zero_user_segment(&folio->page, 0, folio_size(folio));
+		folio_mark_uptodate(folio);
 		goto out;
 	}
 
-	r = netfs_write_begin(file, inode->i_mapping, pos, len, 0, &page, NULL,
+	r = netfs_write_begin(file, inode->i_mapping, pos, len, 0, &folio, NULL,
 			      &ceph_netfs_read_ops, NULL);
 out:
 	if (r == 0)
-		wait_on_page_fscache(page);
+		folio_wait_fscache(folio);
 	if (r < 0) {
-		if (page)
-			put_page(page);
+		if (folio)
+			folio_put(folio);
 	} else {
-		WARN_ON_ONCE(!PageLocked(page));
-		*pagep = page;
+		WARN_ON_ONCE(!folio_test_locked(folio));
+		*pagep = &folio->page;
 	}
 	return r;
 }
@@ -1273,32 +1278,33 @@ static int ceph_write_begin(struct file *file, struct address_space *mapping,
  */
 static int ceph_write_end(struct file *file, struct address_space *mapping,
 			  loff_t pos, unsigned len, unsigned copied,
-			  struct page *page, void *fsdata)
+			  struct page *subpage, void *fsdata)
 {
+	struct folio *folio = page_folio(subpage);
 	struct inode *inode = file_inode(file);
 	bool check_cap = false;
 
-	dout("write_end file %p inode %p page %p %d~%d (%d)\n", file,
-	     inode, page, (int)pos, (int)copied, (int)len);
+	dout("write_end file %p inode %p folio %p %d~%d (%d)\n", file,
+	     inode, folio, (int)pos, (int)copied, (int)len);
 
-	if (!PageUptodate(page)) {
+	if (!folio_test_uptodate(folio)) {
 		/* just return that nothing was copied on a short copy */
 		if (copied < len) {
 			copied = 0;
 			goto out;
 		}
-		SetPageUptodate(page);
+		folio_mark_uptodate(folio);
 	}
 
 	/* did file size increase? */
 	if (pos+copied > i_size_read(inode))
 		check_cap = ceph_inode_set_size(inode, pos+copied);
 
-	set_page_dirty(page);
+	folio_mark_dirty(folio);
 
 out:
-	unlock_page(page);
-	put_page(page);
+	folio_unlock(folio);
+	folio_put(folio);
 
 	if (check_cap)
 		ceph_check_caps(ceph_inode(inode), CHECK_CAPS_AUTHONLY, NULL);
@@ -1306,17 +1312,6 @@ static int ceph_write_end(struct file *file, struct address_space *mapping,
 	return copied;
 }
 
-/*
- * we set .direct_IO to indicate direct io is supported, but since we
- * intercept O_DIRECT reads and writes early, this function should
- * never get called.
- */
-static ssize_t ceph_direct_io(struct kiocb *iocb, struct iov_iter *iter)
-{
-	WARN_ON(1);
-	return -EINVAL;
-}
-
 const struct address_space_operations ceph_aops = {
 	.readpage = ceph_readpage,
 	.readahead = ceph_readahead,
@@ -1327,7 +1322,7 @@ const struct address_space_operations ceph_aops = {
 	.set_page_dirty = ceph_set_page_dirty,
 	.invalidatepage = ceph_invalidatepage,
 	.releasepage = ceph_releasepage,
-	.direct_IO = ceph_direct_io,
+	.direct_IO = noop_direct_IO,
 };
 
 static void ceph_block_sigs(sigset_t *oldset)
@@ -1356,6 +1351,9 @@ static vm_fault_t ceph_filemap_fault(struct vm_fault *vmf)
 	sigset_t oldset;
 	vm_fault_t ret = VM_FAULT_SIGBUS;
 
+	if (ceph_inode_is_shutdown(inode))
+		return ret;
+
 	ceph_block_sigs(&oldset);
 
 	dout("filemap_fault %p %llx.%llx %llu trying to get caps\n",
@@ -1447,6 +1445,9 @@ static vm_fault_t ceph_page_mkwrite(struct vm_fault *vmf)
 	sigset_t oldset;
 	vm_fault_t ret = VM_FAULT_SIGBUS;
 
+	if (ceph_inode_is_shutdown(inode))
+		return ret;
+
 	prealloc_cf = ceph_alloc_cap_flush();
 	if (!prealloc_cf)
 		return VM_FAULT_OOM;
diff --git a/fs/ceph/cache.c b/fs/ceph/cache.c
index 9cfadbb..457afda 100644
--- a/fs/ceph/cache.c
+++ b/fs/ceph/cache.c
@@ -12,12 +12,6 @@
 #include "super.h"
 #include "cache.h"
 
-struct ceph_aux_inode {
-	u64 	version;
-	u64	mtime_sec;
-	u64	mtime_nsec;
-};
-
 struct fscache_netfs ceph_cache_netfs = {
 	.name		= "ceph",
 	.version	= 0,
@@ -109,20 +103,14 @@ static enum fscache_checkaux ceph_fscache_inode_check_aux(
 	void *cookie_netfs_data, const void *data, uint16_t dlen,
 	loff_t object_size)
 {
-	struct ceph_aux_inode aux;
 	struct ceph_inode_info* ci = cookie_netfs_data;
 	struct inode* inode = &ci->vfs_inode;
 
-	if (dlen != sizeof(aux) ||
+	if (dlen != sizeof(ci->i_version) ||
 	    i_size_read(inode) != object_size)
 		return FSCACHE_CHECKAUX_OBSOLETE;
 
-	memset(&aux, 0, sizeof(aux));
-	aux.version = ci->i_version;
-	aux.mtime_sec = inode->i_mtime.tv_sec;
-	aux.mtime_nsec = inode->i_mtime.tv_nsec;
-
-	if (memcmp(data, &aux, sizeof(aux)) != 0)
+	if (*(u64 *)data != ci->i_version)
 		return FSCACHE_CHECKAUX_OBSOLETE;
 
 	dout("ceph inode 0x%p cached okay\n", ci);
@@ -139,7 +127,6 @@ void ceph_fscache_register_inode_cookie(struct inode *inode)
 {
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
-	struct ceph_aux_inode aux;
 
 	/* No caching for filesystem */
 	if (!fsc->fscache)
@@ -151,14 +138,10 @@ void ceph_fscache_register_inode_cookie(struct inode *inode)
 
 	inode_lock_nested(inode, I_MUTEX_CHILD);
 	if (!ci->fscache) {
-		memset(&aux, 0, sizeof(aux));
-		aux.version = ci->i_version;
-		aux.mtime_sec = inode->i_mtime.tv_sec;
-		aux.mtime_nsec = inode->i_mtime.tv_nsec;
 		ci->fscache = fscache_acquire_cookie(fsc->fscache,
 						     &ceph_fscache_inode_object_def,
 						     &ci->i_vino, sizeof(ci->i_vino),
-						     &aux, sizeof(aux),
+						     &ci->i_version, sizeof(ci->i_version),
 						     ci, i_size_read(inode), false);
 	}
 	inode_unlock(inode);
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 8f537f1..b9460b6 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -1188,11 +1188,11 @@ void ceph_remove_cap(struct ceph_cap *cap, bool queue_release)
 
 	lockdep_assert_held(&ci->i_ceph_lock);
 
-	fsc = ceph_sb_to_client(ci->vfs_inode.i_sb);
+	fsc = ceph_inode_to_client(&ci->vfs_inode);
 	WARN_ON_ONCE(ci->i_auth_cap == cap &&
 		     !list_empty(&ci->i_dirty_item) &&
 		     !fsc->blocklisted &&
-		     READ_ONCE(fsc->mount_state) != CEPH_MOUNT_SHUTDOWN);
+		     !ceph_inode_is_shutdown(&ci->vfs_inode));
 
 	__ceph_remove_cap(cap, queue_release);
 }
@@ -1968,8 +1968,8 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
 		}
 	}
 
-	dout("check_caps %p file_want %s used %s dirty %s flushing %s"
-	     " issued %s revoking %s retain %s %s%s\n", inode,
+	dout("check_caps %llx.%llx file_want %s used %s dirty %s flushing %s"
+	     " issued %s revoking %s retain %s %s%s\n", ceph_vinop(inode),
 	     ceph_cap_string(file_wanted),
 	     ceph_cap_string(used), ceph_cap_string(ci->i_dirty_caps),
 	     ceph_cap_string(ci->i_flushing_caps),
@@ -1990,7 +1990,8 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
 	    (revoking & (CEPH_CAP_FILE_CACHE|
 			 CEPH_CAP_FILE_LAZYIO)) && /*  or revoking cache */
 	    !tried_invalidate) {
-		dout("check_caps trying to invalidate on %p\n", inode);
+		dout("check_caps trying to invalidate on %llx.%llx\n",
+		     ceph_vinop(inode));
 		if (try_nonblocking_invalidate(inode) < 0) {
 			dout("check_caps queuing invalidate\n");
 			queue_invalidate = true;
@@ -2629,9 +2630,9 @@ void ceph_take_cap_refs(struct ceph_inode_info *ci, int got,
  *
  * Returns 0 if caps were not able to be acquired (yet), 1 if succeed,
  * or a negative error code. There are 3 speical error codes:
- *  -EAGAIN: need to sleep but non-blocking is specified
- *  -EFBIG:  ask caller to call check_max_size() and try again.
- *  -ESTALE: ask caller to call ceph_renew_caps() and try again.
+ *  -EAGAIN:  need to sleep but non-blocking is specified
+ *  -EFBIG:   ask caller to call check_max_size() and try again.
+ *  -EUCLEAN: ask caller to call ceph_renew_caps() and try again.
  */
 enum {
 	/* first 8 bits are reserved for CEPH_FILE_MODE_FOO */
@@ -2679,7 +2680,7 @@ static int try_get_cap_refs(struct inode *inode, int need, int want,
 			dout("get_cap_refs %p endoff %llu > maxsize %llu\n",
 			     inode, endoff, ci->i_max_size);
 			if (endoff > ci->i_requested_max_size)
-				ret = ci->i_auth_cap ? -EFBIG : -ESTALE;
+				ret = ci->i_auth_cap ? -EFBIG : -EUCLEAN;
 			goto out_unlock;
 		}
 		/*
@@ -2749,9 +2750,9 @@ static int try_get_cap_refs(struct inode *inode, int need, int want,
 			goto out_unlock;
 		}
 
-		if (READ_ONCE(mdsc->fsc->mount_state) >= CEPH_MOUNT_SHUTDOWN) {
-			dout("get_cap_refs %p forced umount\n", inode);
-			ret = -EIO;
+		if (ceph_inode_is_shutdown(inode)) {
+			dout("get_cap_refs %p inode is shutdown\n", inode);
+			ret = -ESTALE;
 			goto out_unlock;
 		}
 		mds_wanted = __ceph_caps_mds_wanted(ci, false);
@@ -2759,7 +2760,7 @@ static int try_get_cap_refs(struct inode *inode, int need, int want,
 			dout("get_cap_refs %p need %s > mds_wanted %s\n",
 			     inode, ceph_cap_string(need),
 			     ceph_cap_string(mds_wanted));
-			ret = -ESTALE;
+			ret = -EUCLEAN;
 			goto out_unlock;
 		}
 
@@ -2843,7 +2844,7 @@ int ceph_try_get_caps(struct inode *inode, int need, int want,
 
 	ret = try_get_cap_refs(inode, need, want, 0, flags, got);
 	/* three special error codes */
-	if (ret == -EAGAIN || ret == -EFBIG || ret == -ESTALE)
+	if (ret == -EAGAIN || ret == -EFBIG || ret == -EUCLEAN)
 		ret = 0;
 	return ret;
 }
@@ -2926,7 +2927,7 @@ int ceph_get_caps(struct file *filp, int need, int want, loff_t endoff, int *got
 		}
 
 		if (ret < 0) {
-			if (ret == -EFBIG || ret == -ESTALE) {
+			if (ret == -EFBIG || ret == -EUCLEAN) {
 				int ret2 = ceph_wait_on_async_create(inode);
 				if (ret2 < 0)
 					return ret2;
@@ -2935,7 +2936,7 @@ int ceph_get_caps(struct file *filp, int need, int want, loff_t endoff, int *got
 				check_max_size(inode, endoff);
 				continue;
 			}
-			if (ret == -ESTALE) {
+			if (ret == -EUCLEAN) {
 				/* session was killed, try renew caps */
 				ret = ceph_renew_caps(inode, flags);
 				if (ret == 0)
@@ -4315,7 +4316,7 @@ static void flush_dirty_session_caps(struct ceph_mds_session *s)
 				      i_dirty_item);
 		inode = &ci->vfs_inode;
 		ihold(inode);
-		dout("flush_dirty_caps %p\n", inode);
+		dout("flush_dirty_caps %llx.%llx\n", ceph_vinop(inode));
 		spin_unlock(&mdsc->cap_dirty_lock);
 		ceph_check_caps(ci, CHECK_CAPS_FLUSH, NULL);
 		iput(inode);
@@ -4560,3 +4561,119 @@ int ceph_encode_dentry_release(void **p, struct dentry *dentry,
 	spin_unlock(&dentry->d_lock);
 	return ret;
 }
+
+static int remove_capsnaps(struct ceph_mds_client *mdsc, struct inode *inode)
+{
+	struct ceph_inode_info *ci = ceph_inode(inode);
+	struct ceph_cap_snap *capsnap;
+	int capsnap_release = 0;
+
+	lockdep_assert_held(&ci->i_ceph_lock);
+
+	dout("removing capsnaps, ci is %p, inode is %p\n", ci, inode);
+
+	while (!list_empty(&ci->i_cap_snaps)) {
+		capsnap = list_first_entry(&ci->i_cap_snaps,
+					   struct ceph_cap_snap, ci_item);
+		__ceph_remove_capsnap(inode, capsnap, NULL, NULL);
+		ceph_put_snap_context(capsnap->context);
+		ceph_put_cap_snap(capsnap);
+		capsnap_release++;
+	}
+	wake_up_all(&ci->i_cap_wq);
+	wake_up_all(&mdsc->cap_flushing_wq);
+	return capsnap_release;
+}
+
+int ceph_purge_inode_cap(struct inode *inode, struct ceph_cap *cap, bool *invalidate)
+{
+	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
+	struct ceph_mds_client *mdsc = fsc->mdsc;
+	struct ceph_inode_info *ci = ceph_inode(inode);
+	bool is_auth;
+	bool dirty_dropped = false;
+	int iputs = 0;
+
+	lockdep_assert_held(&ci->i_ceph_lock);
+
+	dout("removing cap %p, ci is %p, inode is %p\n",
+	     cap, ci, &ci->vfs_inode);
+
+	is_auth = (cap == ci->i_auth_cap);
+	__ceph_remove_cap(cap, false);
+	if (is_auth) {
+		struct ceph_cap_flush *cf;
+
+		if (ceph_inode_is_shutdown(inode)) {
+			if (inode->i_data.nrpages > 0)
+				*invalidate = true;
+			if (ci->i_wrbuffer_ref > 0)
+				mapping_set_error(&inode->i_data, -EIO);
+		}
+
+		spin_lock(&mdsc->cap_dirty_lock);
+
+		/* trash all of the cap flushes for this inode */
+		while (!list_empty(&ci->i_cap_flush_list)) {
+			cf = list_first_entry(&ci->i_cap_flush_list,
+					      struct ceph_cap_flush, i_list);
+			list_del_init(&cf->g_list);
+			list_del_init(&cf->i_list);
+			if (!cf->is_capsnap)
+				ceph_free_cap_flush(cf);
+		}
+
+		if (!list_empty(&ci->i_dirty_item)) {
+			pr_warn_ratelimited(
+				" dropping dirty %s state for %p %lld\n",
+				ceph_cap_string(ci->i_dirty_caps),
+				inode, ceph_ino(inode));
+			ci->i_dirty_caps = 0;
+			list_del_init(&ci->i_dirty_item);
+			dirty_dropped = true;
+		}
+		if (!list_empty(&ci->i_flushing_item)) {
+			pr_warn_ratelimited(
+				" dropping dirty+flushing %s state for %p %lld\n",
+				ceph_cap_string(ci->i_flushing_caps),
+				inode, ceph_ino(inode));
+			ci->i_flushing_caps = 0;
+			list_del_init(&ci->i_flushing_item);
+			mdsc->num_cap_flushing--;
+			dirty_dropped = true;
+		}
+		spin_unlock(&mdsc->cap_dirty_lock);
+
+		if (dirty_dropped) {
+			mapping_set_error(inode->i_mapping, -EIO);
+
+			if (ci->i_wrbuffer_ref_head == 0 &&
+			    ci->i_wr_ref == 0 &&
+			    ci->i_dirty_caps == 0 &&
+			    ci->i_flushing_caps == 0) {
+				ceph_put_snap_context(ci->i_head_snapc);
+				ci->i_head_snapc = NULL;
+			}
+		}
+
+		if (atomic_read(&ci->i_filelock_ref) > 0) {
+			/* make further file lock syscall return -EIO */
+			ci->i_ceph_flags |= CEPH_I_ERROR_FILELOCK;
+			pr_warn_ratelimited(" dropping file locks for %p %lld\n",
+					    inode, ceph_ino(inode));
+		}
+
+		if (!ci->i_dirty_caps && ci->i_prealloc_cap_flush) {
+			cf = ci->i_prealloc_cap_flush;
+			ci->i_prealloc_cap_flush = NULL;
+			if (!cf->is_capsnap)
+				ceph_free_cap_flush(cf);
+		}
+
+		if (!list_empty(&ci->i_cap_snaps))
+			iputs = remove_capsnaps(mdsc, inode);
+	}
+	if (dirty_dropped)
+		++iputs;
+	return iputs;
+}
diff --git a/fs/ceph/debugfs.c b/fs/ceph/debugfs.c
index 38b78b45..3cf7c9c 100644
--- a/fs/ceph/debugfs.c
+++ b/fs/ceph/debugfs.c
@@ -146,82 +146,93 @@ static int mdsc_show(struct seq_file *s, void *p)
 		   name, total, avg, _min, max, sum);			\
 }
 
-static int metric_show(struct seq_file *s, void *p)
+static int metrics_file_show(struct seq_file *s, void *p)
 {
 	struct ceph_fs_client *fsc = s->private;
-	struct ceph_mds_client *mdsc = fsc->mdsc;
-	struct ceph_client_metric *m = &mdsc->metric;
-	int nr_caps = 0;
-	s64 total, sum, avg, min, max, sq;
-	u64 sum_sz, avg_sz, min_sz, max_sz;
+	struct ceph_client_metric *m = &fsc->mdsc->metric;
 
-	sum = percpu_counter_sum(&m->total_inodes);
 	seq_printf(s, "item                               total\n");
 	seq_printf(s, "------------------------------------------\n");
-	seq_printf(s, "%-35s%lld / %lld\n", "opened files  / total inodes",
-		   atomic64_read(&m->opened_files), sum);
-	seq_printf(s, "%-35s%lld / %lld\n", "pinned i_caps / total inodes",
-		   atomic64_read(&m->total_caps), sum);
-	seq_printf(s, "%-35s%lld / %lld\n", "opened inodes / total inodes",
-		   percpu_counter_sum(&m->opened_inodes), sum);
+	seq_printf(s, "%-35s%lld\n", "total inodes",
+		   percpu_counter_sum(&m->total_inodes));
+	seq_printf(s, "%-35s%lld\n", "opened files",
+		   atomic64_read(&m->opened_files));
+	seq_printf(s, "%-35s%lld\n", "pinned i_caps",
+		   atomic64_read(&m->total_caps));
+	seq_printf(s, "%-35s%lld\n", "opened inodes",
+		   percpu_counter_sum(&m->opened_inodes));
+	return 0;
+}
 
-	seq_printf(s, "\n");
+static const char * const metric_str[] = {
+	"read",
+	"write",
+	"metadata",
+	"copyfrom"
+};
+static int metrics_latency_show(struct seq_file *s, void *p)
+{
+	struct ceph_fs_client *fsc = s->private;
+	struct ceph_client_metric *cm = &fsc->mdsc->metric;
+	struct ceph_metric *m;
+	s64 total, sum, avg, min, max, sq;
+	int i;
+
 	seq_printf(s, "item          total       avg_lat(us)     min_lat(us)     max_lat(us)     stdev(us)\n");
 	seq_printf(s, "-----------------------------------------------------------------------------------\n");
 
-	spin_lock(&m->read_metric_lock);
-	total = m->total_reads;
-	sum = m->read_latency_sum;
-	avg = total > 0 ? DIV64_U64_ROUND_CLOSEST(sum, total) : 0;
-	min = m->read_latency_min;
-	max = m->read_latency_max;
-	sq = m->read_latency_sq_sum;
-	spin_unlock(&m->read_metric_lock);
-	CEPH_LAT_METRIC_SHOW("read", total, avg, min, max, sq);
+	for (i = 0; i < METRIC_MAX; i++) {
+		m = &cm->metric[i];
+		spin_lock(&m->lock);
+		total = m->total;
+		sum = m->latency_sum;
+		avg = total > 0 ? DIV64_U64_ROUND_CLOSEST(sum, total) : 0;
+		min = m->latency_min;
+		max = m->latency_max;
+		sq = m->latency_sq_sum;
+		spin_unlock(&m->lock);
+		CEPH_LAT_METRIC_SHOW(metric_str[i], total, avg, min, max, sq);
+	}
 
-	spin_lock(&m->write_metric_lock);
-	total = m->total_writes;
-	sum = m->write_latency_sum;
-	avg = total > 0 ? DIV64_U64_ROUND_CLOSEST(sum, total) : 0;
-	min = m->write_latency_min;
-	max = m->write_latency_max;
-	sq = m->write_latency_sq_sum;
-	spin_unlock(&m->write_metric_lock);
-	CEPH_LAT_METRIC_SHOW("write", total, avg, min, max, sq);
+	return 0;
+}
 
-	spin_lock(&m->metadata_metric_lock);
-	total = m->total_metadatas;
-	sum = m->metadata_latency_sum;
-	avg = total > 0 ? DIV64_U64_ROUND_CLOSEST(sum, total) : 0;
-	min = m->metadata_latency_min;
-	max = m->metadata_latency_max;
-	sq = m->metadata_latency_sq_sum;
-	spin_unlock(&m->metadata_metric_lock);
-	CEPH_LAT_METRIC_SHOW("metadata", total, avg, min, max, sq);
+static int metrics_size_show(struct seq_file *s, void *p)
+{
+	struct ceph_fs_client *fsc = s->private;
+	struct ceph_client_metric *cm = &fsc->mdsc->metric;
+	struct ceph_metric *m;
+	s64 total;
+	u64 sum, avg, min, max;
+	int i;
 
-	seq_printf(s, "\n");
 	seq_printf(s, "item          total       avg_sz(bytes)   min_sz(bytes)   max_sz(bytes)  total_sz(bytes)\n");
 	seq_printf(s, "----------------------------------------------------------------------------------------\n");
 
-	spin_lock(&m->read_metric_lock);
-	total = m->total_reads;
-	sum_sz = m->read_size_sum;
-	avg_sz = total > 0 ? DIV64_U64_ROUND_CLOSEST(sum_sz, total) : 0;
-	min_sz = m->read_size_min;
-	max_sz = m->read_size_max;
-	spin_unlock(&m->read_metric_lock);
-	CEPH_SZ_METRIC_SHOW("read", total, avg_sz, min_sz, max_sz, sum_sz);
+	for (i = 0; i < METRIC_MAX; i++) {
+		/* skip 'metadata' as it doesn't use the size metric */
+		if (i == METRIC_METADATA)
+			continue;
+		m = &cm->metric[i];
+		spin_lock(&m->lock);
+		total = m->total;
+		sum = m->size_sum;
+		avg = total > 0 ? DIV64_U64_ROUND_CLOSEST(sum, total) : 0;
+		min = m->size_min;
+		max = m->size_max;
+		spin_unlock(&m->lock);
+		CEPH_SZ_METRIC_SHOW(metric_str[i], total, avg, min, max, sum);
+	}
 
-	spin_lock(&m->write_metric_lock);
-	total = m->total_writes;
-	sum_sz = m->write_size_sum;
-	avg_sz = total > 0 ? DIV64_U64_ROUND_CLOSEST(sum_sz, total) : 0;
-	min_sz = m->write_size_min;
-	max_sz = m->write_size_max;
-	spin_unlock(&m->write_metric_lock);
-	CEPH_SZ_METRIC_SHOW("write", total, avg_sz, min_sz, max_sz, sum_sz);
+	return 0;
+}
 
-	seq_printf(s, "\n");
+static int metrics_caps_show(struct seq_file *s, void *p)
+{
+	struct ceph_fs_client *fsc = s->private;
+	struct ceph_client_metric *m = &fsc->mdsc->metric;
+	int nr_caps = 0;
+
 	seq_printf(s, "item          total           miss            hit\n");
 	seq_printf(s, "-------------------------------------------------\n");
 
@@ -350,8 +361,11 @@ DEFINE_SHOW_ATTRIBUTE(mdsmap);
 DEFINE_SHOW_ATTRIBUTE(mdsc);
 DEFINE_SHOW_ATTRIBUTE(caps);
 DEFINE_SHOW_ATTRIBUTE(mds_sessions);
-DEFINE_SHOW_ATTRIBUTE(metric);
 DEFINE_SHOW_ATTRIBUTE(status);
+DEFINE_SHOW_ATTRIBUTE(metrics_file);
+DEFINE_SHOW_ATTRIBUTE(metrics_latency);
+DEFINE_SHOW_ATTRIBUTE(metrics_size);
+DEFINE_SHOW_ATTRIBUTE(metrics_caps);
 
 
 /*
@@ -385,8 +399,9 @@ void ceph_fs_debugfs_cleanup(struct ceph_fs_client *fsc)
 	debugfs_remove(fsc->debugfs_mdsmap);
 	debugfs_remove(fsc->debugfs_mds_sessions);
 	debugfs_remove(fsc->debugfs_caps);
-	debugfs_remove(fsc->debugfs_metric);
+	debugfs_remove(fsc->debugfs_status);
 	debugfs_remove(fsc->debugfs_mdsc);
+	debugfs_remove_recursive(fsc->debugfs_metrics_dir);
 }
 
 void ceph_fs_debugfs_init(struct ceph_fs_client *fsc)
@@ -426,12 +441,6 @@ void ceph_fs_debugfs_init(struct ceph_fs_client *fsc)
 						fsc,
 						&mdsc_fops);
 
-	fsc->debugfs_metric = debugfs_create_file("metrics",
-						  0400,
-						  fsc->client->debugfs_dir,
-						  fsc,
-						  &metric_fops);
-
 	fsc->debugfs_caps = debugfs_create_file("caps",
 						0400,
 						fsc->client->debugfs_dir,
@@ -443,6 +452,18 @@ void ceph_fs_debugfs_init(struct ceph_fs_client *fsc)
 						  fsc->client->debugfs_dir,
 						  fsc,
 						  &status_fops);
+
+	fsc->debugfs_metrics_dir = debugfs_create_dir("metrics",
+						      fsc->client->debugfs_dir);
+
+	debugfs_create_file("file", 0400, fsc->debugfs_metrics_dir, fsc,
+			    &metrics_file_fops);
+	debugfs_create_file("latency", 0400, fsc->debugfs_metrics_dir, fsc,
+			    &metrics_latency_fops);
+	debugfs_create_file("size", 0400, fsc->debugfs_metrics_dir, fsc,
+			    &metrics_size_fops);
+	debugfs_create_file("caps", 0400, fsc->debugfs_metrics_dir, fsc,
+			    &metrics_caps_fops);
 }
 
 
diff --git a/fs/ceph/export.c b/fs/ceph/export.c
index 1d65934..e0fa66ac 100644
--- a/fs/ceph/export.c
+++ b/fs/ceph/export.c
@@ -157,6 +157,11 @@ static struct inode *__lookup_inode(struct super_block *sb, u64 ino)
 		ceph_mdsc_put_request(req);
 		if (!inode)
 			return err < 0 ? ERR_PTR(err) : ERR_PTR(-ESTALE);
+	} else {
+		if (ceph_inode_is_shutdown(inode)) {
+			iput(inode);
+			return ERR_PTR(-ESTALE);
+		}
 	}
 	return inode;
 }
@@ -223,8 +228,13 @@ static struct dentry *__snapfh_to_dentry(struct super_block *sb,
 		return ERR_PTR(-ESTALE);
 
 	inode = ceph_find_inode(sb, vino);
-	if (inode)
+	if (inode) {
+		if (ceph_inode_is_shutdown(inode)) {
+			iput(inode);
+			return ERR_PTR(-ESTALE);
+		}
 		return d_obtain_alias(inode);
+	}
 
 	req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_LOOKUPINO,
 				       USE_ANY_MDS);
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index b129ea5..02a0a0f 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -525,6 +525,7 @@ static void ceph_async_create_cb(struct ceph_mds_client *mdsc,
 
 	if (result) {
 		struct dentry *dentry = req->r_dentry;
+		struct inode *inode = d_inode(dentry);
 		int pathlen = 0;
 		u64 base = 0;
 		char *path = ceph_mdsc_build_path(req->r_dentry, &pathlen,
@@ -534,7 +535,8 @@ static void ceph_async_create_cb(struct ceph_mds_client *mdsc,
 		if (!d_unhashed(dentry))
 			d_drop(dentry);
 
-		/* FIXME: start returning I/O errors on all accesses? */
+		ceph_inode_shutdown(inode);
+
 		pr_warn("ceph: async create failure path=(%llx)%s result=%d!\n",
 			base, IS_ERR(path) ? "<<bad>>" : path, result);
 		ceph_mdsc_free_path(path, pathlen);
@@ -556,7 +558,7 @@ static void ceph_async_create_cb(struct ceph_mds_client *mdsc,
 		}
 		ceph_kick_flushing_inode_caps(req->r_session, ci);
 		spin_unlock(&ci->i_ceph_lock);
-	} else {
+	} else if (!result) {
 		pr_warn("%s: no req->r_target_inode for 0x%llx\n", __func__,
 			req->r_deleg_ino);
 	}
@@ -845,6 +847,7 @@ static ssize_t ceph_sync_read(struct kiocb *iocb, struct iov_iter *to,
 	ssize_t ret;
 	u64 off = iocb->ki_pos;
 	u64 len = iov_iter_count(to);
+	u64 i_size;
 
 	dout("sync_read on file %p %llu~%u %s\n", file, off, (unsigned)len,
 	     (file->f_flags & O_DIRECT) ? "O_DIRECT" : "");
@@ -868,7 +871,6 @@ static ssize_t ceph_sync_read(struct kiocb *iocb, struct iov_iter *to,
 		struct page **pages;
 		int num_pages;
 		size_t page_off;
-		u64 i_size;
 		bool more;
 		int idx;
 		size_t left;
@@ -951,11 +953,14 @@ static ssize_t ceph_sync_read(struct kiocb *iocb, struct iov_iter *to,
 	}
 
 	if (off > iocb->ki_pos) {
-		if (ret >= 0 &&
-		    iov_iter_count(to) > 0 && off >= i_size_read(inode))
+		if (off >= i_size) {
 			*retry_op = CHECK_EOF;
-		ret = off - iocb->ki_pos;
-		iocb->ki_pos = off;
+			ret = i_size - iocb->ki_pos;
+			iocb->ki_pos = i_size;
+		} else {
+			ret = off - iocb->ki_pos;
+			iocb->ki_pos = off;
+		}
 	}
 
 	dout("sync_read result %zd retry_op %d\n", ret, *retry_op);
@@ -1526,6 +1531,9 @@ static ssize_t ceph_read_iter(struct kiocb *iocb, struct iov_iter *to)
 	dout("aio_read %p %llx.%llx %llu~%u trying to get caps on %p\n",
 	     inode, ceph_vinop(inode), iocb->ki_pos, (unsigned)len, inode);
 
+	if (ceph_inode_is_shutdown(inode))
+		return -ESTALE;
+
 	if (direct_lock)
 		ceph_start_io_direct(inode);
 	else
@@ -1678,6 +1686,9 @@ static ssize_t ceph_write_iter(struct kiocb *iocb, struct iov_iter *from)
 	loff_t pos;
 	loff_t limit = max(i_size_read(inode), fsc->max_file_size);
 
+	if (ceph_inode_is_shutdown(inode))
+		return -ESTALE;
+
 	if (ceph_snap(inode) != CEPH_NOSNAP)
 		return -EROFS;
 
@@ -2200,6 +2211,54 @@ static int is_file_size_ok(struct inode *src_inode, struct inode *dst_inode,
 	return 0;
 }
 
+static struct ceph_osd_request *
+ceph_alloc_copyfrom_request(struct ceph_osd_client *osdc,
+			    u64 src_snapid,
+			    struct ceph_object_id *src_oid,
+			    struct ceph_object_locator *src_oloc,
+			    struct ceph_object_id *dst_oid,
+			    struct ceph_object_locator *dst_oloc,
+			    u32 truncate_seq, u64 truncate_size)
+{
+	struct ceph_osd_request *req;
+	int ret;
+	u32 src_fadvise_flags =
+		CEPH_OSD_OP_FLAG_FADVISE_SEQUENTIAL |
+		CEPH_OSD_OP_FLAG_FADVISE_NOCACHE;
+	u32 dst_fadvise_flags =
+		CEPH_OSD_OP_FLAG_FADVISE_SEQUENTIAL |
+		CEPH_OSD_OP_FLAG_FADVISE_DONTNEED;
+
+	req = ceph_osdc_alloc_request(osdc, NULL, 1, false, GFP_KERNEL);
+	if (!req)
+		return ERR_PTR(-ENOMEM);
+
+	req->r_flags = CEPH_OSD_FLAG_WRITE;
+
+	ceph_oloc_copy(&req->r_t.base_oloc, dst_oloc);
+	ceph_oid_copy(&req->r_t.base_oid, dst_oid);
+
+	ret = osd_req_op_copy_from_init(req, src_snapid, 0,
+					src_oid, src_oloc,
+					src_fadvise_flags,
+					dst_fadvise_flags,
+					truncate_seq,
+					truncate_size,
+					CEPH_OSD_COPY_FROM_FLAG_TRUNCATE_SEQ);
+	if (ret)
+		goto out;
+
+	ret = ceph_osdc_alloc_messages(req, GFP_KERNEL);
+	if (ret)
+		goto out;
+
+	return req;
+
+out:
+	ceph_osdc_put_request(req);
+	return ERR_PTR(ret);
+}
+
 static ssize_t ceph_do_objects_copy(struct ceph_inode_info *src_ci, u64 *src_off,
 				    struct ceph_inode_info *dst_ci, u64 *dst_off,
 				    struct ceph_fs_client *fsc,
@@ -2207,6 +2266,8 @@ static ssize_t ceph_do_objects_copy(struct ceph_inode_info *src_ci, u64 *src_off
 {
 	struct ceph_object_locator src_oloc, dst_oloc;
 	struct ceph_object_id src_oid, dst_oid;
+	struct ceph_osd_client *osdc;
+	struct ceph_osd_request *req;
 	size_t bytes = 0;
 	u64 src_objnum, src_objoff, dst_objnum, dst_objoff;
 	u32 src_objlen, dst_objlen;
@@ -2217,6 +2278,7 @@ static ssize_t ceph_do_objects_copy(struct ceph_inode_info *src_ci, u64 *src_off
 	src_oloc.pool_ns = ceph_try_get_string(src_ci->i_layout.pool_ns);
 	dst_oloc.pool = dst_ci->i_layout.pool_id;
 	dst_oloc.pool_ns = ceph_try_get_string(dst_ci->i_layout.pool_ns);
+	osdc = &fsc->client->osdc;
 
 	while (len >= object_size) {
 		ceph_calc_file_object_mapping(&src_ci->i_layout, *src_off,
@@ -2232,17 +2294,22 @@ static ssize_t ceph_do_objects_copy(struct ceph_inode_info *src_ci, u64 *src_off
 		ceph_oid_printf(&dst_oid, "%llx.%08llx",
 				dst_ci->i_vino.ino, dst_objnum);
 		/* Do an object remote copy */
-		ret = ceph_osdc_copy_from(&fsc->client->osdc,
-					  src_ci->i_vino.snap, 0,
-					  &src_oid, &src_oloc,
-					  CEPH_OSD_OP_FLAG_FADVISE_SEQUENTIAL |
-					  CEPH_OSD_OP_FLAG_FADVISE_NOCACHE,
-					  &dst_oid, &dst_oloc,
-					  CEPH_OSD_OP_FLAG_FADVISE_SEQUENTIAL |
-					  CEPH_OSD_OP_FLAG_FADVISE_DONTNEED,
-					  dst_ci->i_truncate_seq,
-					  dst_ci->i_truncate_size,
-					  CEPH_OSD_COPY_FROM_FLAG_TRUNCATE_SEQ);
+		req = ceph_alloc_copyfrom_request(osdc, src_ci->i_vino.snap,
+						  &src_oid, &src_oloc,
+						  &dst_oid, &dst_oloc,
+						  dst_ci->i_truncate_seq,
+						  dst_ci->i_truncate_size);
+		if (IS_ERR(req))
+			ret = PTR_ERR(req);
+		else {
+			ceph_osdc_start_request(osdc, req, false);
+			ret = ceph_osdc_wait_request(osdc, req);
+			ceph_update_copyfrom_metrics(&fsc->mdsc->metric,
+						     req->r_start_latency,
+						     req->r_end_latency,
+						     object_size, ret);
+			ceph_osdc_put_request(req);
+		}
 		if (ret) {
 			if (ret == -EOPNOTSUPP) {
 				fsc->have_copy_from2 = false;
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 1c75741..e3322fc 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -1841,15 +1841,14 @@ void ceph_queue_inode_work(struct inode *inode, int work_bit)
 static void ceph_do_invalidate_pages(struct inode *inode)
 {
 	struct ceph_inode_info *ci = ceph_inode(inode);
-	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
 	u32 orig_gen;
 	int check = 0;
 
 	mutex_lock(&ci->i_truncate_mutex);
 
-	if (READ_ONCE(fsc->mount_state) >= CEPH_MOUNT_SHUTDOWN) {
-		pr_warn_ratelimited("invalidate_pages %p %lld forced umount\n",
-				    inode, ceph_ino(inode));
+	if (ceph_inode_is_shutdown(inode)) {
+		pr_warn_ratelimited("%s: inode %llx.%llx is shut down\n",
+				    __func__, ceph_vinop(inode));
 		mapping_set_error(inode->i_mapping, -EIO);
 		truncate_pagecache(inode, 0);
 		mutex_unlock(&ci->i_truncate_mutex);
@@ -1871,7 +1870,8 @@ static void ceph_do_invalidate_pages(struct inode *inode)
 
 	ceph_fscache_invalidate(inode);
 	if (invalidate_inode_pages2(inode->i_mapping) < 0) {
-		pr_err("invalidate_pages %p fails\n", inode);
+		pr_err("invalidate_inode_pages2 %llx.%llx failed\n",
+		       ceph_vinop(inode));
 	}
 
 	spin_lock(&ci->i_ceph_lock);
@@ -2103,12 +2103,14 @@ int __ceph_setattr(struct inode *inode, struct iattr *attr)
 		loff_t isize = i_size_read(inode);
 
 		dout("setattr %p size %lld -> %lld\n", inode, isize, attr->ia_size);
-		if ((issued & CEPH_CAP_FILE_EXCL) && attr->ia_size > isize) {
-			i_size_write(inode, attr->ia_size);
-			inode->i_blocks = calc_inode_blocks(attr->ia_size);
-			ci->i_reported_size = attr->ia_size;
-			dirtied |= CEPH_CAP_FILE_EXCL;
-			ia_valid |= ATTR_MTIME;
+		if ((issued & CEPH_CAP_FILE_EXCL) && attr->ia_size >= isize) {
+			if (attr->ia_size > isize) {
+				i_size_write(inode, attr->ia_size);
+				inode->i_blocks = calc_inode_blocks(attr->ia_size);
+				ci->i_reported_size = attr->ia_size;
+				dirtied |= CEPH_CAP_FILE_EXCL;
+				ia_valid |= ATTR_MTIME;
+			}
 		} else if ((issued & CEPH_CAP_FILE_SHARED) == 0 ||
 			   attr->ia_size != isize) {
 			req->r_args.setattr.size = cpu_to_le64(attr->ia_size);
@@ -2217,6 +2219,9 @@ int ceph_setattr(struct user_namespace *mnt_userns, struct dentry *dentry,
 	if (ceph_snap(inode) != CEPH_NOSNAP)
 		return -EROFS;
 
+	if (ceph_inode_is_shutdown(inode))
+		return -ESTALE;
+
 	err = setattr_prepare(&init_user_ns, dentry, attr);
 	if (err != 0)
 		return err;
@@ -2347,6 +2352,9 @@ int ceph_getattr(struct user_namespace *mnt_userns, const struct path *path,
 	u32 valid_mask = STATX_BASIC_STATS;
 	int err = 0;
 
+	if (ceph_inode_is_shutdown(inode))
+		return -ESTALE;
+
 	/* Skip the getattr altogether if we're asked not to sync */
 	if (!(flags & AT_STATX_DONT_SYNC)) {
 		err = ceph_do_getattr(inode,
@@ -2394,3 +2402,27 @@ int ceph_getattr(struct user_namespace *mnt_userns, const struct path *path,
 	stat->result_mask = request_mask & valid_mask;
 	return err;
 }
+
+void ceph_inode_shutdown(struct inode *inode)
+{
+	struct ceph_inode_info *ci = ceph_inode(inode);
+	struct rb_node *p;
+	int iputs = 0;
+	bool invalidate = false;
+
+	spin_lock(&ci->i_ceph_lock);
+	ci->i_ceph_flags |= CEPH_I_SHUTDOWN;
+	p = rb_first(&ci->i_caps);
+	while (p) {
+		struct ceph_cap *cap = rb_entry(p, struct ceph_cap, ci_node);
+
+		p = rb_next(p);
+		iputs += ceph_purge_inode_cap(inode, cap, &invalidate);
+	}
+	spin_unlock(&ci->i_ceph_lock);
+
+	if (invalidate)
+		ceph_queue_invalidate(inode);
+	while (iputs--)
+		iput(inode);
+}
diff --git a/fs/ceph/locks.c b/fs/ceph/locks.c
index d8c3106..d1f154a 100644
--- a/fs/ceph/locks.c
+++ b/fs/ceph/locks.c
@@ -241,6 +241,9 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl)
 	if (!(fl->fl_flags & FL_POSIX))
 		return -ENOLCK;
 
+	if (ceph_inode_is_shutdown(inode))
+		return -ESTALE;
+
 	dout("ceph_lock, fl_owner: %p\n", fl->fl_owner);
 
 	/* set wait bit as appropriate, then make command as Ceph expects it*/
@@ -303,6 +306,9 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl)
 	if (!(fl->fl_flags & FL_FLOCK))
 		return -ENOLCK;
 
+	if (ceph_inode_is_shutdown(inode))
+		return -ESTALE;
+
 	dout("ceph_flock, fl_file: %p\n", fl->fl_file);
 
 	spin_lock(&ci->i_ceph_lock);
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index d64413a..250aad3 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -1590,129 +1590,23 @@ int ceph_iterate_session_caps(struct ceph_mds_session *session,
 	return ret;
 }
 
-static int remove_capsnaps(struct ceph_mds_client *mdsc, struct inode *inode)
-{
-	struct ceph_inode_info *ci = ceph_inode(inode);
-	struct ceph_cap_snap *capsnap;
-	int capsnap_release = 0;
-
-	lockdep_assert_held(&ci->i_ceph_lock);
-
-	dout("removing capsnaps, ci is %p, inode is %p\n", ci, inode);
-
-	while (!list_empty(&ci->i_cap_snaps)) {
-		capsnap = list_first_entry(&ci->i_cap_snaps,
-					   struct ceph_cap_snap, ci_item);
-		__ceph_remove_capsnap(inode, capsnap, NULL, NULL);
-		ceph_put_snap_context(capsnap->context);
-		ceph_put_cap_snap(capsnap);
-		capsnap_release++;
-	}
-	wake_up_all(&ci->i_cap_wq);
-	wake_up_all(&mdsc->cap_flushing_wq);
-	return capsnap_release;
-}
-
 static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 				  void *arg)
 {
-	struct ceph_fs_client *fsc = (struct ceph_fs_client *)arg;
-	struct ceph_mds_client *mdsc = fsc->mdsc;
 	struct ceph_inode_info *ci = ceph_inode(inode);
-	LIST_HEAD(to_remove);
-	bool dirty_dropped = false;
 	bool invalidate = false;
-	int capsnap_release = 0;
+	int iputs;
 
 	dout("removing cap %p, ci is %p, inode is %p\n",
 	     cap, ci, &ci->vfs_inode);
 	spin_lock(&ci->i_ceph_lock);
-	__ceph_remove_cap(cap, false);
-	if (!ci->i_auth_cap) {
-		struct ceph_cap_flush *cf;
-
-		if (READ_ONCE(fsc->mount_state) >= CEPH_MOUNT_SHUTDOWN) {
-			if (inode->i_data.nrpages > 0)
-				invalidate = true;
-			if (ci->i_wrbuffer_ref > 0)
-				mapping_set_error(&inode->i_data, -EIO);
-		}
-
-		while (!list_empty(&ci->i_cap_flush_list)) {
-			cf = list_first_entry(&ci->i_cap_flush_list,
-					      struct ceph_cap_flush, i_list);
-			list_move(&cf->i_list, &to_remove);
-		}
-
-		spin_lock(&mdsc->cap_dirty_lock);
-
-		list_for_each_entry(cf, &to_remove, i_list)
-			list_del_init(&cf->g_list);
-
-		if (!list_empty(&ci->i_dirty_item)) {
-			pr_warn_ratelimited(
-				" dropping dirty %s state for %p %lld\n",
-				ceph_cap_string(ci->i_dirty_caps),
-				inode, ceph_ino(inode));
-			ci->i_dirty_caps = 0;
-			list_del_init(&ci->i_dirty_item);
-			dirty_dropped = true;
-		}
-		if (!list_empty(&ci->i_flushing_item)) {
-			pr_warn_ratelimited(
-				" dropping dirty+flushing %s state for %p %lld\n",
-				ceph_cap_string(ci->i_flushing_caps),
-				inode, ceph_ino(inode));
-			ci->i_flushing_caps = 0;
-			list_del_init(&ci->i_flushing_item);
-			mdsc->num_cap_flushing--;
-			dirty_dropped = true;
-		}
-		spin_unlock(&mdsc->cap_dirty_lock);
-
-		if (dirty_dropped) {
-			mapping_set_error(inode->i_mapping, -EIO);
-
-			if (ci->i_wrbuffer_ref_head == 0 &&
-			    ci->i_wr_ref == 0 &&
-			    ci->i_dirty_caps == 0 &&
-			    ci->i_flushing_caps == 0) {
-				ceph_put_snap_context(ci->i_head_snapc);
-				ci->i_head_snapc = NULL;
-			}
-		}
-
-		if (atomic_read(&ci->i_filelock_ref) > 0) {
-			/* make further file lock syscall return -EIO */
-			ci->i_ceph_flags |= CEPH_I_ERROR_FILELOCK;
-			pr_warn_ratelimited(" dropping file locks for %p %lld\n",
-					    inode, ceph_ino(inode));
-		}
-
-		if (!ci->i_dirty_caps && ci->i_prealloc_cap_flush) {
-			list_add(&ci->i_prealloc_cap_flush->i_list, &to_remove);
-			ci->i_prealloc_cap_flush = NULL;
-		}
-
-		if (!list_empty(&ci->i_cap_snaps))
-			capsnap_release = remove_capsnaps(mdsc, inode);
-	}
+	iputs = ceph_purge_inode_cap(inode, cap, &invalidate);
 	spin_unlock(&ci->i_ceph_lock);
-	while (!list_empty(&to_remove)) {
-		struct ceph_cap_flush *cf;
-		cf = list_first_entry(&to_remove,
-				      struct ceph_cap_flush, i_list);
-		list_del_init(&cf->i_list);
-		if (!cf->is_capsnap)
-			ceph_free_cap_flush(cf);
-	}
 
 	wake_up_all(&ci->i_cap_wq);
 	if (invalidate)
 		ceph_queue_invalidate(inode);
-	if (dirty_dropped)
-		iput(inode);
-	while (capsnap_release--)
+	while (iputs--)
 		iput(inode);
 	return 0;
 }
@@ -3467,9 +3361,14 @@ static void handle_session(struct ceph_mds_session *session,
 
 	if (msg_version >= 3) {
 		u32 len;
-		/* version >= 2, metadata */
-		if (__decode_session_metadata(&p, end, &blocklisted) < 0)
+		/* version >= 2 and < 5, decode metadata, skip otherwise
+		 * as it's handled via flags.
+		 */
+		if (msg_version >= 5)
+			ceph_decode_skip_map(&p, end, string, string, bad);
+		else if (__decode_session_metadata(&p, end, &blocklisted) < 0)
 			goto bad;
+
 		/* version >= 3, feature bits */
 		ceph_decode_32_safe(&p, end, len, bad);
 		if (len) {
@@ -3478,6 +3377,18 @@ static void handle_session(struct ceph_mds_session *session,
 		}
 	}
 
+	if (msg_version >= 5) {
+		u32 flags;
+		/* version >= 4, struct_v, struct_cv, len, metric_spec */
+	        ceph_decode_skip_n(&p, end, 2 + sizeof(u32) * 2, bad);
+		/* version >= 5, flags   */
+                ceph_decode_32_safe(&p, end, flags, bad);
+		if (flags & CEPH_SESSION_BLOCKLISTED) {
+		        pr_warn("mds%d session blocklisted\n", session->s_mds);
+			blocklisted = true;
+		}
+	}
+
 	mutex_lock(&mdsc->mutex);
 	if (op == CEPH_SESSION_CLOSE) {
 		ceph_get_mds_session(session);
@@ -5072,7 +4983,8 @@ void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg)
 	return;
 
 bad:
-	pr_err("error decoding fsmap\n");
+	pr_err("error decoding fsmap %d. Shutting down mount.\n", err);
+	ceph_umount_begin(mdsc->fsc->sb);
 err_out:
 	mutex_lock(&mdsc->mutex);
 	mdsc->mdsmap_err = err;
@@ -5139,7 +5051,8 @@ void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg)
 bad_unlock:
 	mutex_unlock(&mdsc->mutex);
 bad:
-	pr_err("error decoding mdsmap %d\n", err);
+	pr_err("error decoding mdsmap %d. Shutting down mount.\n", err);
+	ceph_umount_begin(mdsc->fsc->sb);
 	return;
 }
 
diff --git a/fs/ceph/mdsmap.c b/fs/ceph/mdsmap.c
index 61d67cb..3038773 100644
--- a/fs/ceph/mdsmap.c
+++ b/fs/ceph/mdsmap.c
@@ -263,10 +263,6 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end, bool msgr2)
 				goto nomem;
 			for (j = 0; j < num_export_targets; j++) {
 				target = ceph_decode_32(&pexport_targets);
-				if (target >= m->possible_max_rank) {
-					err = -EIO;
-					goto corrupt;
-				}
 				info->export_targets[j] = target;
 			}
 		} else {
diff --git a/fs/ceph/metric.c b/fs/ceph/metric.c
index 04d5df2..c57699d 100644
--- a/fs/ceph/metric.c
+++ b/fs/ceph/metric.c
@@ -62,7 +62,7 @@ static bool ceph_mdsc_send_metrics(struct ceph_mds_client *mdsc,
 	read->header.ver = 1;
 	read->header.compat = 1;
 	read->header.data_len = cpu_to_le32(sizeof(*read) - header_len);
-	sum = m->read_latency_sum;
+	sum = m->metric[METRIC_READ].latency_sum;
 	jiffies_to_timespec64(sum, &ts);
 	read->sec = cpu_to_le32(ts.tv_sec);
 	read->nsec = cpu_to_le32(ts.tv_nsec);
@@ -74,7 +74,7 @@ static bool ceph_mdsc_send_metrics(struct ceph_mds_client *mdsc,
 	write->header.ver = 1;
 	write->header.compat = 1;
 	write->header.data_len = cpu_to_le32(sizeof(*write) - header_len);
-	sum = m->write_latency_sum;
+	sum = m->metric[METRIC_WRITE].latency_sum;
 	jiffies_to_timespec64(sum, &ts);
 	write->sec = cpu_to_le32(ts.tv_sec);
 	write->nsec = cpu_to_le32(ts.tv_nsec);
@@ -86,7 +86,7 @@ static bool ceph_mdsc_send_metrics(struct ceph_mds_client *mdsc,
 	meta->header.ver = 1;
 	meta->header.compat = 1;
 	meta->header.data_len = cpu_to_le32(sizeof(*meta) - header_len);
-	sum = m->metadata_latency_sum;
+	sum = m->metric[METRIC_METADATA].latency_sum;
 	jiffies_to_timespec64(sum, &ts);
 	meta->sec = cpu_to_le32(ts.tv_sec);
 	meta->nsec = cpu_to_le32(ts.tv_nsec);
@@ -141,8 +141,8 @@ static bool ceph_mdsc_send_metrics(struct ceph_mds_client *mdsc,
 	rsize->header.ver = 1;
 	rsize->header.compat = 1;
 	rsize->header.data_len = cpu_to_le32(sizeof(*rsize) - header_len);
-	rsize->total_ops = cpu_to_le64(m->total_reads);
-	rsize->total_size = cpu_to_le64(m->read_size_sum);
+	rsize->total_ops = cpu_to_le64(m->metric[METRIC_READ].total);
+	rsize->total_size = cpu_to_le64(m->metric[METRIC_READ].size_sum);
 	items++;
 
 	/* encode the write io size metric */
@@ -151,8 +151,8 @@ static bool ceph_mdsc_send_metrics(struct ceph_mds_client *mdsc,
 	wsize->header.ver = 1;
 	wsize->header.compat = 1;
 	wsize->header.data_len = cpu_to_le32(sizeof(*wsize) - header_len);
-	wsize->total_ops = cpu_to_le64(m->total_writes);
-	wsize->total_size = cpu_to_le64(m->write_size_sum);
+	wsize->total_ops = cpu_to_le64(m->metric[METRIC_WRITE].total);
+	wsize->total_size = cpu_to_le64(m->metric[METRIC_WRITE].size_sum);
 	items++;
 
 	put_unaligned_le32(items, &head->num);
@@ -220,7 +220,8 @@ static void metric_delayed_work(struct work_struct *work)
 
 int ceph_metric_init(struct ceph_client_metric *m)
 {
-	int ret;
+	struct ceph_metric *metric;
+	int ret, i;
 
 	if (!m)
 		return -EINVAL;
@@ -243,32 +244,18 @@ int ceph_metric_init(struct ceph_client_metric *m)
 	if (ret)
 		goto err_i_caps_mis;
 
-	spin_lock_init(&m->read_metric_lock);
-	m->read_latency_sq_sum = 0;
-	m->read_latency_min = KTIME_MAX;
-	m->read_latency_max = 0;
-	m->total_reads = 0;
-	m->read_latency_sum = 0;
-	m->read_size_min = U64_MAX;
-	m->read_size_max = 0;
-	m->read_size_sum = 0;
-
-	spin_lock_init(&m->write_metric_lock);
-	m->write_latency_sq_sum = 0;
-	m->write_latency_min = KTIME_MAX;
-	m->write_latency_max = 0;
-	m->total_writes = 0;
-	m->write_latency_sum = 0;
-	m->write_size_min = U64_MAX;
-	m->write_size_max = 0;
-	m->write_size_sum = 0;
-
-	spin_lock_init(&m->metadata_metric_lock);
-	m->metadata_latency_sq_sum = 0;
-	m->metadata_latency_min = KTIME_MAX;
-	m->metadata_latency_max = 0;
-	m->total_metadatas = 0;
-	m->metadata_latency_sum = 0;
+	for (i = 0; i < METRIC_MAX; i++) {
+		metric = &m->metric[i];
+		spin_lock_init(&metric->lock);
+		metric->size_sum = 0;
+		metric->size_min = U64_MAX;
+		metric->size_max = 0;
+		metric->total = 0;
+		metric->latency_sum = 0;
+		metric->latency_sq_sum = 0;
+		metric->latency_min = KTIME_MAX;
+		metric->latency_max = 0;
+	}
 
 	atomic64_set(&m->opened_files, 0);
 	ret = percpu_counter_init(&m->opened_inodes, 0, GFP_KERNEL);
@@ -338,9 +325,9 @@ static inline void __update_stdev(ktime_t total, ktime_t lsum,
 	*sq_sump += sq;
 }
 
-void ceph_update_read_metrics(struct ceph_client_metric *m,
-			      ktime_t r_start, ktime_t r_end,
-			      unsigned int size, int rc)
+void ceph_update_metrics(struct ceph_metric *m,
+			 ktime_t r_start, ktime_t r_end,
+			 unsigned int size, int rc)
 {
 	ktime_t lat = ktime_sub(r_end, r_start);
 	ktime_t total;
@@ -348,63 +335,12 @@ void ceph_update_read_metrics(struct ceph_client_metric *m,
 	if (unlikely(rc < 0 && rc != -ENOENT && rc != -ETIMEDOUT))
 		return;
 
-	spin_lock(&m->read_metric_lock);
-	total = ++m->total_reads;
-	m->read_size_sum += size;
-	m->read_latency_sum += lat;
-	METRIC_UPDATE_MIN_MAX(m->read_size_min,
-			      m->read_size_max,
-			      size);
-	METRIC_UPDATE_MIN_MAX(m->read_latency_min,
-			      m->read_latency_max,
-			      lat);
-	__update_stdev(total, m->read_latency_sum,
-		       &m->read_latency_sq_sum, lat);
-	spin_unlock(&m->read_metric_lock);
-}
-
-void ceph_update_write_metrics(struct ceph_client_metric *m,
-			       ktime_t r_start, ktime_t r_end,
-			       unsigned int size, int rc)
-{
-	ktime_t lat = ktime_sub(r_end, r_start);
-	ktime_t total;
-
-	if (unlikely(rc && rc != -ETIMEDOUT))
-		return;
-
-	spin_lock(&m->write_metric_lock);
-	total = ++m->total_writes;
-	m->write_size_sum += size;
-	m->write_latency_sum += lat;
-	METRIC_UPDATE_MIN_MAX(m->write_size_min,
-			      m->write_size_max,
-			      size);
-	METRIC_UPDATE_MIN_MAX(m->write_latency_min,
-			      m->write_latency_max,
-			      lat);
-	__update_stdev(total, m->write_latency_sum,
-		       &m->write_latency_sq_sum, lat);
-	spin_unlock(&m->write_metric_lock);
-}
-
-void ceph_update_metadata_metrics(struct ceph_client_metric *m,
-				  ktime_t r_start, ktime_t r_end,
-				  int rc)
-{
-	ktime_t lat = ktime_sub(r_end, r_start);
-	ktime_t total;
-
-	if (unlikely(rc && rc != -ENOENT))
-		return;
-
-	spin_lock(&m->metadata_metric_lock);
-	total = ++m->total_metadatas;
-	m->metadata_latency_sum += lat;
-	METRIC_UPDATE_MIN_MAX(m->metadata_latency_min,
-			      m->metadata_latency_max,
-			      lat);
-	__update_stdev(total, m->metadata_latency_sum,
-		       &m->metadata_latency_sq_sum, lat);
-	spin_unlock(&m->metadata_metric_lock);
+	spin_lock(&m->lock);
+	total = ++m->total;
+	m->size_sum += size;
+	METRIC_UPDATE_MIN_MAX(m->size_min, m->size_max, size);
+	m->latency_sum += lat;
+	METRIC_UPDATE_MIN_MAX(m->latency_min, m->latency_max, lat);
+	__update_stdev(total, m->latency_sum, &m->latency_sq_sum, lat);
+	spin_unlock(&m->lock);
 }
diff --git a/fs/ceph/metric.h b/fs/ceph/metric.h
index 0133955..bb45608 100644
--- a/fs/ceph/metric.h
+++ b/fs/ceph/metric.h
@@ -125,6 +125,26 @@ struct ceph_metric_head {
 	__le32 num;	/* the number of metrics that will be sent */
 } __packed;
 
+enum metric_type {
+	METRIC_READ,
+	METRIC_WRITE,
+	METRIC_METADATA,
+	METRIC_COPYFROM,
+	METRIC_MAX
+};
+
+struct ceph_metric {
+	spinlock_t lock;
+	u64 total;
+	u64 size_sum;
+	u64 size_min;
+	u64 size_max;
+	ktime_t latency_sum;
+	ktime_t latency_sq_sum;
+	ktime_t latency_min;
+	ktime_t latency_max;
+};
+
 /* This is the global metrics */
 struct ceph_client_metric {
 	atomic64_t            total_dentries;
@@ -135,32 +155,7 @@ struct ceph_client_metric {
 	struct percpu_counter i_caps_hit;
 	struct percpu_counter i_caps_mis;
 
-	spinlock_t read_metric_lock;
-	u64 total_reads;
-	u64 read_size_sum;
-	u64 read_size_min;
-	u64 read_size_max;
-	ktime_t read_latency_sum;
-	ktime_t read_latency_sq_sum;
-	ktime_t read_latency_min;
-	ktime_t read_latency_max;
-
-	spinlock_t write_metric_lock;
-	u64 total_writes;
-	u64 write_size_sum;
-	u64 write_size_min;
-	u64 write_size_max;
-	ktime_t write_latency_sum;
-	ktime_t write_latency_sq_sum;
-	ktime_t write_latency_min;
-	ktime_t write_latency_max;
-
-	spinlock_t metadata_metric_lock;
-	u64 total_metadatas;
-	ktime_t metadata_latency_sum;
-	ktime_t metadata_latency_sq_sum;
-	ktime_t metadata_latency_min;
-	ktime_t metadata_latency_max;
+	struct ceph_metric metric[METRIC_MAX];
 
 	/* The total number of directories and files that are opened */
 	atomic64_t opened_files;
@@ -195,13 +190,36 @@ static inline void ceph_update_cap_mis(struct ceph_client_metric *m)
 	percpu_counter_inc(&m->i_caps_mis);
 }
 
-extern void ceph_update_read_metrics(struct ceph_client_metric *m,
-				     ktime_t r_start, ktime_t r_end,
-				     unsigned int size, int rc);
-extern void ceph_update_write_metrics(struct ceph_client_metric *m,
-				      ktime_t r_start, ktime_t r_end,
-				      unsigned int size, int rc);
-extern void ceph_update_metadata_metrics(struct ceph_client_metric *m,
-				         ktime_t r_start, ktime_t r_end,
-					 int rc);
+extern void ceph_update_metrics(struct ceph_metric *m,
+				ktime_t r_start, ktime_t r_end,
+				unsigned int size, int rc);
+
+static inline void ceph_update_read_metrics(struct ceph_client_metric *m,
+					    ktime_t r_start, ktime_t r_end,
+					    unsigned int size, int rc)
+{
+	ceph_update_metrics(&m->metric[METRIC_READ],
+			    r_start, r_end, size, rc);
+}
+static inline void ceph_update_write_metrics(struct ceph_client_metric *m,
+					     ktime_t r_start, ktime_t r_end,
+					     unsigned int size, int rc)
+{
+	ceph_update_metrics(&m->metric[METRIC_WRITE],
+			    r_start, r_end, size, rc);
+}
+static inline void ceph_update_metadata_metrics(struct ceph_client_metric *m,
+						ktime_t r_start, ktime_t r_end,
+						int rc)
+{
+	ceph_update_metrics(&m->metric[METRIC_METADATA],
+			    r_start, r_end, 0, rc);
+}
+static inline void ceph_update_copyfrom_metrics(struct ceph_client_metric *m,
+						ktime_t r_start, ktime_t r_end,
+						unsigned int size, int rc)
+{
+	ceph_update_metrics(&m->metric[METRIC_COPYFROM],
+			    r_start, r_end, size, rc);
+}
 #endif /* _FS_CEPH_MDS_METRIC_H */
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index fd8742b..bab6123 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -52,8 +52,7 @@ static int ceph_statfs(struct dentry *dentry, struct kstatfs *buf)
 	struct ceph_fs_client *fsc = ceph_inode_to_client(d_inode(dentry));
 	struct ceph_mon_client *monc = &fsc->client->monc;
 	struct ceph_statfs st;
-	u64 fsid;
-	int err;
+	int i, err;
 	u64 data_pool;
 
 	if (fsc->mdsc->mdsmap->m_num_data_pg_pools == 1) {
@@ -99,12 +98,14 @@ static int ceph_statfs(struct dentry *dentry, struct kstatfs *buf)
 	buf->f_namelen = NAME_MAX;
 
 	/* Must convert the fsid, for consistent values across arches */
+	buf->f_fsid.val[0] = 0;
 	mutex_lock(&monc->mutex);
-	fsid = le64_to_cpu(*(__le64 *)(&monc->monmap->fsid)) ^
-	       le64_to_cpu(*((__le64 *)&monc->monmap->fsid + 1));
+	for (i = 0 ; i < sizeof(monc->monmap->fsid) / sizeof(__le32) ; ++i)
+		buf->f_fsid.val[0] ^= le32_to_cpu(((__le32 *)&monc->monmap->fsid)[i]);
 	mutex_unlock(&monc->mutex);
 
-	buf->f_fsid = u64_to_fsid(fsid);
+	/* fold the fs_cluster_id into the upper bits */
+	buf->f_fsid.val[1] = monc->fs_cluster_id;
 
 	return 0;
 }
@@ -577,8 +578,8 @@ static int ceph_show_options(struct seq_file *m, struct dentry *root)
 	if (fsopt->flags & CEPH_MOUNT_OPT_CLEANRECOVER)
 		seq_show_option(m, "recover_session", "clean");
 
-	if (fsopt->flags & CEPH_MOUNT_OPT_ASYNC_DIROPS)
-		seq_puts(m, ",nowsync");
+	if (!(fsopt->flags & CEPH_MOUNT_OPT_ASYNC_DIROPS))
+		seq_puts(m, ",wsync");
 
 	if (fsopt->wsize != CEPH_MAX_WRITE_SIZE)
 		seq_printf(m, ",wsize=%u", fsopt->wsize);
@@ -842,7 +843,7 @@ static void __ceph_umount_begin(struct ceph_fs_client *fsc)
  * ceph_umount_begin - initiate forced umount.  Tear down the
  * mount, skipping steps that may hang while waiting for server(s).
  */
-static void ceph_umount_begin(struct super_block *sb)
+void ceph_umount_begin(struct super_block *sb)
 {
 	struct ceph_fs_client *fsc = ceph_sb_to_client(sb);
 
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index 14f951c..ac331aa 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -48,7 +48,8 @@
 
 #define CEPH_MOUNT_OPT_DEFAULT			\
 	(CEPH_MOUNT_OPT_DCACHE |		\
-	 CEPH_MOUNT_OPT_NOCOPYFROM)
+	 CEPH_MOUNT_OPT_NOCOPYFROM |		\
+	 CEPH_MOUNT_OPT_ASYNC_DIROPS)
 
 #define ceph_set_mount_opt(fsc, opt) \
 	(fsc)->mount_options->flags |= CEPH_MOUNT_OPT_##opt
@@ -128,9 +129,9 @@ struct ceph_fs_client {
 	struct dentry *debugfs_congestion_kb;
 	struct dentry *debugfs_bdi;
 	struct dentry *debugfs_mdsc, *debugfs_mdsmap;
-	struct dentry *debugfs_metric;
 	struct dentry *debugfs_status;
 	struct dentry *debugfs_mds_sessions;
+	struct dentry *debugfs_metrics_dir;
 #endif
 
 #ifdef CONFIG_CEPH_FSCACHE
@@ -580,6 +581,7 @@ static inline struct inode *ceph_find_inode(struct super_block *sb,
 #define CEPH_I_ODIRECT		(1 << 11) /* inode in direct I/O mode */
 #define CEPH_ASYNC_CREATE_BIT	(12)	  /* async create in flight for this */
 #define CEPH_I_ASYNC_CREATE	(1 << CEPH_ASYNC_CREATE_BIT)
+#define CEPH_I_SHUTDOWN		(1 << 13) /* inode is no longer usable */
 
 /*
  * Masks of ceph inode work.
@@ -939,6 +941,7 @@ extern void ceph_put_snapid_map(struct ceph_mds_client* mdsc,
 				struct ceph_snapid_map *sm);
 extern void ceph_trim_snapid_map(struct ceph_mds_client *mdsc);
 extern void ceph_cleanup_snapid_map(struct ceph_mds_client *mdsc);
+void ceph_umount_begin(struct super_block *sb);
 
 
 /*
@@ -1027,6 +1030,16 @@ extern int ceph_setattr(struct user_namespace *mnt_userns,
 extern int ceph_getattr(struct user_namespace *mnt_userns,
 			const struct path *path, struct kstat *stat,
 			u32 request_mask, unsigned int flags);
+void ceph_inode_shutdown(struct inode *inode);
+
+static inline bool ceph_inode_is_shutdown(struct inode *inode)
+{
+	unsigned long flags = READ_ONCE(ceph_inode(inode)->i_ceph_flags);
+	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
+	int state = READ_ONCE(fsc->mount_state);
+
+	return (flags & CEPH_I_SHUTDOWN) || state >= CEPH_MOUNT_SHUTDOWN;
+}
 
 /* xattr.c */
 int __ceph_setxattr(struct inode *, const char *, const void *, size_t, int);
@@ -1198,6 +1211,7 @@ extern int ceph_mmap(struct file *file, struct vm_area_struct *vma);
 extern int ceph_uninline_data(struct file *filp, struct page *locked_page);
 extern int ceph_pool_perm_check(struct inode *inode, int need);
 extern void ceph_pool_perm_destroy(struct ceph_mds_client* mdsc);
+int ceph_purge_inode_cap(struct inode *inode, struct ceph_cap *cap, bool *invalidate);
 
 /* file.c */
 extern const struct file_operations ceph_file_fops;
diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c
index de2c12b..d282caf 100644
--- a/fs/cifs/cifs_debug.c
+++ b/fs/cifs/cifs_debug.c
@@ -271,7 +271,8 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v)
 	c = 0;
 	spin_lock(&cifs_tcp_ses_lock);
 	list_for_each_entry(server, &cifs_tcp_ses_list, tcp_ses_list) {
-		if (server->is_channel)
+		/* channel info will be printed as a part of sessions below */
+		if (CIFS_SERVER_IS_CHAN(server))
 			continue;
 
 		c++;
@@ -358,6 +359,8 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v)
 			seq_printf(m, " signed");
 		if (server->posix_ext_supported)
 			seq_printf(m, " posix");
+		if (server->nosharesock)
+			seq_printf(m, " nosharesock");
 
 		if (server->rdma)
 			seq_printf(m, "\nRDMA ");
@@ -412,12 +415,14 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v)
 				   from_kuid(&init_user_ns, ses->linux_uid),
 				   from_kuid(&init_user_ns, ses->cred_uid));
 
+			spin_lock(&ses->chan_lock);
 			if (ses->chan_count > 1) {
 				seq_printf(m, "\n\n\tExtra Channels: %zu ",
 					   ses->chan_count-1);
 				for (j = 1; j < ses->chan_count; j++)
 					cifs_dump_channel(m, j, &ses->chans[j]);
 			}
+			spin_unlock(&ses->chan_lock);
 
 			seq_puts(m, "\n\n\tShares: ");
 			j = 0;
diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c
index 007427b..b0864da 100644
--- a/fs/cifs/cifs_dfs_ref.c
+++ b/fs/cifs/cifs_dfs_ref.c
@@ -307,12 +307,8 @@ static struct vfsmount *cifs_dfs_do_mount(struct dentry *mntpt,
 static struct vfsmount *cifs_dfs_do_automount(struct dentry *mntpt)
 {
 	struct cifs_sb_info *cifs_sb;
-	struct cifs_ses *ses;
-	struct cifs_tcon *tcon;
 	void *page;
-	char *full_path, *root_path;
-	unsigned int xid;
-	int rc;
+	char *full_path;
 	struct vfsmount *mnt;
 
 	cifs_dbg(FYI, "in %s\n", __func__);
@@ -324,8 +320,6 @@ static struct vfsmount *cifs_dfs_do_automount(struct dentry *mntpt)
 	 * the double backslashes usually used in the UNC. This function
 	 * gives us the latter, so we must adjust the result.
 	 */
-	mnt = ERR_PTR(-ENOMEM);
-
 	cifs_sb = CIFS_SB(mntpt->d_sb);
 	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_DFS) {
 		mnt = ERR_PTR(-EREMOTE);
@@ -341,60 +335,11 @@ static struct vfsmount *cifs_dfs_do_automount(struct dentry *mntpt)
 	}
 
 	convert_delimiter(full_path, '\\');
-
 	cifs_dbg(FYI, "%s: full_path: %s\n", __func__, full_path);
 
-	if (!cifs_sb_master_tlink(cifs_sb)) {
-		cifs_dbg(FYI, "%s: master tlink is NULL\n", __func__);
-		goto free_full_path;
-	}
-
-	tcon = cifs_sb_master_tcon(cifs_sb);
-	if (!tcon) {
-		cifs_dbg(FYI, "%s: master tcon is NULL\n", __func__);
-		goto free_full_path;
-	}
-
-	root_path = kstrdup(tcon->treeName, GFP_KERNEL);
-	if (!root_path) {
-		mnt = ERR_PTR(-ENOMEM);
-		goto free_full_path;
-	}
-	cifs_dbg(FYI, "%s: root path: %s\n", __func__, root_path);
-
-	ses = tcon->ses;
-	xid = get_xid();
-
-	/*
-	 * If DFS root has been expired, then unconditionally fetch it again to
-	 * refresh DFS referral cache.
-	 */
-	rc = dfs_cache_find(xid, ses, cifs_sb->local_nls, cifs_remap(cifs_sb),
-			    root_path + 1, NULL, NULL);
-	if (!rc) {
-		rc = dfs_cache_find(xid, ses, cifs_sb->local_nls,
-				    cifs_remap(cifs_sb), full_path + 1,
-				    NULL, NULL);
-	}
-
-	free_xid(xid);
-
-	if (rc) {
-		mnt = ERR_PTR(rc);
-		goto free_root_path;
-	}
-	/*
-	 * OK - we were able to get and cache a referral for @full_path.
-	 *
-	 * Now, pass it down to cifs_mount() and it will retry every available
-	 * node server in case of failures - no need to do it here.
-	 */
 	mnt = cifs_dfs_do_mount(mntpt, cifs_sb, full_path);
-	cifs_dbg(FYI, "%s: cifs_dfs_do_mount:%s , mnt:%p\n", __func__,
-		 full_path + 1, mnt);
+	cifs_dbg(FYI, "%s: cifs_dfs_do_mount:%s , mnt:%p\n", __func__, full_path + 1, mnt);
 
-free_root_path:
-	kfree(root_path);
 free_full_path:
 	free_dentry_path(page);
 cdda_exit:
diff --git a/fs/cifs/cifs_fs_sb.h b/fs/cifs/cifs_fs_sb.h
index f9740752..013a4bd 100644
--- a/fs/cifs/cifs_fs_sb.h
+++ b/fs/cifs/cifs_fs_sb.h
@@ -61,11 +61,6 @@ struct cifs_sb_info {
 	/* only used when CIFS_MOUNT_USE_PREFIX_PATH is set */
 	char *prepath;
 
-	/*
-	 * Canonical DFS path initially provided by the mount call. We might connect to something
-	 * different via DFS but we want to keep it to do failover properly.
-	 */
-	char *origin_fullpath; /* \\HOST\SHARE\[OPTIONAL PATH] */
 	/* randomly generated 128-bit number for indexing dfs mount groups in referral cache */
 	uuid_t dfs_mount_id;
 	/*
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index abff31d..be74606 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -15,6 +15,7 @@
 #include <linux/slab.h>
 #include <linux/mempool.h>
 #include <linux/workqueue.h>
+#include <linux/utsname.h>
 #include "cifs_fs_sb.h"
 #include "cifsacl.h"
 #include <crypto/internal/hash.h>
@@ -75,7 +76,8 @@
 #define SMB_ECHO_INTERVAL_MAX 600
 #define SMB_ECHO_INTERVAL_DEFAULT 60
 
-/* dns resolution interval in seconds */
+/* dns resolution intervals in seconds */
+#define SMB_DNS_RESOLVE_INTERVAL_MIN     120
 #define SMB_DNS_RESOLVE_INTERVAL_DEFAULT 600
 
 /* maximum number of PDUs in one compound */
@@ -99,6 +101,8 @@
 #define XATTR_DOS_ATTRIB "user.DOSATTRIB"
 #endif
 
+#define CIFS_MAX_WORKSTATION_LEN  (__NEW_UTS_LEN + 1)  /* reasonable max for client */
+
 /*
  * CIFS vfs client Status information (based on what we know.)
  */
@@ -592,6 +596,7 @@ struct TCP_Server_Info {
 	struct list_head pending_mid_q;
 	bool noblocksnd;		/* use blocking sendmsg */
 	bool noautotune;		/* do not autotune send buf sizes */
+	bool nosharesock;
 	bool tcp_nodelay;
 	unsigned int credits;  /* send no more requests at once */
 	unsigned int max_credits; /* can override large 32000 default at mnt */
@@ -685,13 +690,34 @@ struct TCP_Server_Info {
 	 */
 	int nr_targets;
 	bool noblockcnt; /* use non-blocking connect() */
-	bool is_channel; /* if a session channel */
+
+	/*
+	 * If this is a session channel,
+	 * primary_server holds the ref-counted
+	 * pointer to primary channel connection for the session.
+	 */
+#define CIFS_SERVER_IS_CHAN(server)	(!!(server)->primary_server)
+	struct TCP_Server_Info *primary_server;
+
 #ifdef CONFIG_CIFS_SWN_UPCALL
 	bool use_swn_dstaddr;
 	struct sockaddr_storage swn_dstaddr;
 #endif
 #ifdef CONFIG_CIFS_DFS_UPCALL
 	bool is_dfs_conn; /* if a dfs connection */
+	struct mutex refpath_lock; /* protects leaf_fullpath */
+	/*
+	 * Canonical DFS full paths that were used to chase referrals in mount and reconnect.
+	 *
+	 * origin_fullpath: first or original referral path
+	 * leaf_fullpath: last referral path (might be changed due to nested links in reconnect)
+	 *
+	 * current_fullpath: pointer to either origin_fullpath or leaf_fullpath
+	 * NOTE: cannot be accessed outside cifs_reconnect() and smb2_reconnect()
+	 *
+	 * format: \\HOST\SHARE\[OPTIONAL PATH]
+	 */
+	char *origin_fullpath, *leaf_fullpath, *current_fullpath;
 #endif
 };
 
@@ -908,6 +934,7 @@ struct cifs_ses {
 				   and after mount option parsing we fill it */
 	char *domainName;
 	char *password;
+	char *workstation_name;
 	struct session_key auth_key;
 	struct ntlmssp_auth *ntlmssp; /* ciphertext, flags, server challenge */
 	enum securityEnum sectype; /* what security flavor was specified? */
@@ -933,16 +960,21 @@ struct cifs_ses {
 	 * iface_lock should be taken when accessing any of these fields
 	 */
 	spinlock_t iface_lock;
+	/* ========= begin: protected by iface_lock ======== */
 	struct cifs_server_iface *iface_list;
 	size_t iface_count;
 	unsigned long iface_last_update; /* jiffies */
+	/* ========= end: protected by iface_lock ======== */
 
+	spinlock_t chan_lock;
+	/* ========= begin: protected by chan_lock ======== */
 #define CIFS_MAX_CHANNELS 16
 	struct cifs_chan chans[CIFS_MAX_CHANNELS];
 	struct cifs_chan *binding_chan;
 	size_t chan_count;
 	size_t chan_max;
 	atomic_t chan_seq; /* round robin state */
+	/* ========= end: protected by chan_lock ======== */
 };
 
 /*
@@ -1091,7 +1123,6 @@ struct cifs_tcon {
 	struct cached_fid crfid; /* Cached root fid */
 	/* BB add field for back pointer to sb struct(s)? */
 #ifdef CONFIG_CIFS_DFS_UPCALL
-	char *dfs_path; /* canonical DFS path */
 	struct list_head ulist; /* cache update list */
 #endif
 };
@@ -1942,4 +1973,14 @@ static inline bool is_tcon_dfs(struct cifs_tcon *tcon)
 		tcon->share_flags & (SHI1005_FLAGS_DFS | SHI1005_FLAGS_DFS_ROOT);
 }
 
+static inline bool cifs_is_referral_server(struct cifs_tcon *tcon,
+					   const struct dfs_info3_param *ref)
+{
+	/*
+	 * Check if all targets are capable of handling DFS referrals as per
+	 * MS-DFSC 2.2.4 RESP_GET_DFS_REFERRAL.
+	 */
+	return is_tcon_dfs(tcon) || (ref && (ref->flags & DFSREF_REFERRAL_SERVER));
+}
+
 #endif	/* _CIFS_GLOB_H */
diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
index d0f85b66..f3073a6 100644
--- a/fs/cifs/cifsproto.h
+++ b/fs/cifs/cifsproto.h
@@ -269,8 +269,9 @@ extern void cifs_close_all_deferred_files(struct cifs_tcon *cifs_tcon);
 
 extern void cifs_close_deferred_file_under_dentry(struct cifs_tcon *cifs_tcon,
 				const char *path);
-
-extern struct TCP_Server_Info *cifs_get_tcp_session(struct smb3_fs_context *ctx);
+extern struct TCP_Server_Info *
+cifs_get_tcp_session(struct smb3_fs_context *ctx,
+		     struct TCP_Server_Info *primary_server);
 extern void cifs_put_tcp_session(struct TCP_Server_Info *server,
 				 int from_reconnect);
 extern void cifs_put_tcon(struct cifs_tcon *tcon);
@@ -607,7 +608,7 @@ int smb2_parse_query_directory(struct cifs_tcon *tcon, struct kvec *rsp_iov,
 
 struct super_block *cifs_get_tcp_super(struct TCP_Server_Info *server);
 void cifs_put_tcp_super(struct super_block *sb);
-int update_super_prepath(struct cifs_tcon *tcon, char *prefix);
+int cifs_update_super_prepath(struct cifs_sb_info *cifs_sb, char *prefix);
 char *extract_hostname(const char *unc);
 char *extract_sharename(const char *unc);
 
@@ -634,4 +635,7 @@ static inline int cifs_create_options(struct cifs_sb_info *cifs_sb, int options)
 		return options;
 }
 
+struct super_block *cifs_get_tcon_super(struct cifs_tcon *tcon);
+void cifs_put_tcon_super(struct super_block *sb);
+
 #endif			/* _CIFSPROTO_H */
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 0abbff4..82577a7 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -61,6 +61,20 @@ extern bool disable_legacy_dialects;
 /* Drop the connection to not overload the server */
 #define NUM_STATUS_IO_TIMEOUT   5
 
+struct mount_ctx {
+	struct cifs_sb_info *cifs_sb;
+	struct smb3_fs_context *fs_ctx;
+	unsigned int xid;
+	struct TCP_Server_Info *server;
+	struct cifs_ses *ses;
+	struct cifs_tcon *tcon;
+#ifdef CONFIG_CIFS_DFS_UPCALL
+	struct cifs_ses *root_ses;
+	uuid_t mount_id;
+	char *origin_fullpath, *leaf_fullpath;
+#endif
+};
+
 static int ip_connect(struct TCP_Server_Info *server);
 static int generic_ip_connect(struct TCP_Server_Info *server);
 static void tlink_rb_insert(struct rb_root *root, struct tcon_link *new_tlink);
@@ -115,7 +129,7 @@ static int reconn_set_ipaddr_from_hostname(struct TCP_Server_Info *server)
 			 * To make sure we don't use the cached entry, retry 1s
 			 * after expiry.
 			 */
-			ttl = (expiry - now + 1);
+			ttl = max_t(unsigned long, expiry - now, SMB_DNS_RESOLVE_INTERVAL_MIN) + 1;
 	}
 	rc = !rc ? -1 : 0;
 
@@ -148,139 +162,38 @@ static void cifs_resolve_server(struct work_struct *work)
 	mutex_unlock(&server->srv_mutex);
 }
 
-#ifdef CONFIG_CIFS_DFS_UPCALL
-/* These functions must be called with server->srv_mutex held */
-static void reconn_set_next_dfs_target(struct TCP_Server_Info *server,
-				       struct cifs_sb_info *cifs_sb,
-				       struct dfs_cache_tgt_list *tgt_list,
-				       struct dfs_cache_tgt_iterator **tgt_it)
-{
-	const char *name;
-	int rc;
-
-	if (!cifs_sb || !cifs_sb->origin_fullpath)
-		return;
-
-	if (!*tgt_it) {
-		*tgt_it = dfs_cache_get_tgt_iterator(tgt_list);
-	} else {
-		*tgt_it = dfs_cache_get_next_tgt(tgt_list, *tgt_it);
-		if (!*tgt_it)
-			*tgt_it = dfs_cache_get_tgt_iterator(tgt_list);
-	}
-
-	cifs_dbg(FYI, "%s: UNC: %s\n", __func__, cifs_sb->origin_fullpath);
-
-	name = dfs_cache_get_tgt_name(*tgt_it);
-
-	kfree(server->hostname);
-
-	server->hostname = extract_hostname(name);
-	if (IS_ERR(server->hostname)) {
-		cifs_dbg(FYI,
-			 "%s: failed to extract hostname from target: %ld\n",
-			 __func__, PTR_ERR(server->hostname));
-		return;
-	}
-
-	rc = reconn_set_ipaddr_from_hostname(server);
-	if (rc) {
-		cifs_dbg(FYI, "%s: failed to resolve hostname: %d\n",
-			 __func__, rc);
-	}
-}
-
-static inline int reconn_setup_dfs_targets(struct cifs_sb_info *cifs_sb,
-					   struct dfs_cache_tgt_list *tl)
-{
-	if (!cifs_sb->origin_fullpath)
-		return -EOPNOTSUPP;
-	return dfs_cache_noreq_find(cifs_sb->origin_fullpath + 1, NULL, tl);
-}
-#endif
-
-/*
- * cifs tcp session reconnection
+/**
+ * Mark all sessions and tcons for reconnect.
  *
- * mark tcp session as reconnecting so temporarily locked
- * mark all smb sessions as reconnecting for tcp session
- * reconnect tcp session
- * wake up waiters on reconnection? - (not needed currently)
+ * @server needs to be previously set to CifsNeedReconnect.
  */
-int
-cifs_reconnect(struct TCP_Server_Info *server)
+static void cifs_mark_tcp_ses_conns_for_reconnect(struct TCP_Server_Info *server)
 {
-	int rc = 0;
-	struct list_head *tmp, *tmp2;
 	struct cifs_ses *ses;
 	struct cifs_tcon *tcon;
-	struct mid_q_entry *mid_entry;
+	struct mid_q_entry *mid, *nmid;
 	struct list_head retry_list;
-#ifdef CONFIG_CIFS_DFS_UPCALL
-	struct super_block *sb = NULL;
-	struct cifs_sb_info *cifs_sb = NULL;
-	struct dfs_cache_tgt_list tgt_list = DFS_CACHE_TGT_LIST_INIT(tgt_list);
-	struct dfs_cache_tgt_iterator *tgt_it = NULL;
-#endif
+	struct TCP_Server_Info *pserver;
 
-	spin_lock(&GlobalMid_Lock);
-	server->nr_targets = 1;
-#ifdef CONFIG_CIFS_DFS_UPCALL
-	spin_unlock(&GlobalMid_Lock);
-	sb = cifs_get_tcp_super(server);
-	if (IS_ERR(sb)) {
-		rc = PTR_ERR(sb);
-		cifs_dbg(FYI, "%s: will not do DFS failover: rc = %d\n",
-			 __func__, rc);
-		sb = NULL;
-	} else {
-		cifs_sb = CIFS_SB(sb);
-		rc = reconn_setup_dfs_targets(cifs_sb, &tgt_list);
-		if (rc) {
-			cifs_sb = NULL;
-			if (rc != -EOPNOTSUPP) {
-				cifs_server_dbg(VFS, "%s: no target servers for DFS failover\n",
-						__func__);
-			}
-		} else {
-			server->nr_targets = dfs_cache_get_nr_tgts(&tgt_list);
-		}
-	}
-	cifs_dbg(FYI, "%s: will retry %d target(s)\n", __func__,
-		 server->nr_targets);
-	spin_lock(&GlobalMid_Lock);
-#endif
-	if (server->tcpStatus == CifsExiting) {
-		/* the demux thread will exit normally
-		next time through the loop */
-		spin_unlock(&GlobalMid_Lock);
-#ifdef CONFIG_CIFS_DFS_UPCALL
-		dfs_cache_free_tgts(&tgt_list);
-		cifs_put_tcp_super(sb);
-#endif
-		wake_up(&server->response_q);
-		return rc;
-	} else
-		server->tcpStatus = CifsNeedReconnect;
-	spin_unlock(&GlobalMid_Lock);
 	server->maxBuf = 0;
 	server->max_read = 0;
 
 	cifs_dbg(FYI, "Mark tcp session as need reconnect\n");
 	trace_smb3_reconnect(server->CurrentMid, server->conn_id, server->hostname);
+	/*
+	 * before reconnecting the tcp session, mark the smb session (uid) and the tid bad so they
+	 * are not used until reconnected.
+	 */
+	cifs_dbg(FYI, "%s: marking sessions and tcons for reconnect\n", __func__);
 
-	/* before reconnecting the tcp session, mark the smb session (uid)
-		and the tid bad so they are not used until reconnected */
-	cifs_dbg(FYI, "%s: marking sessions and tcons for reconnect\n",
-		 __func__);
+	/* If server is a channel, select the primary channel */
+	pserver = CIFS_SERVER_IS_CHAN(server) ? server->primary_server : server;
+
 	spin_lock(&cifs_tcp_ses_lock);
-	list_for_each(tmp, &server->smb_ses_list) {
-		ses = list_entry(tmp, struct cifs_ses, smb_ses_list);
+	list_for_each_entry(ses, &pserver->smb_ses_list, smb_ses_list) {
 		ses->need_reconnect = true;
-		list_for_each(tmp2, &ses->tcon_list) {
-			tcon = list_entry(tmp2, struct cifs_tcon, tcon_list);
+		list_for_each_entry(tcon, &ses->tcon_list, tcon_list)
 			tcon->need_reconnect = true;
-		}
 		if (ses->tcon_ipc)
 			ses->tcon_ipc->need_reconnect = true;
 	}
@@ -290,11 +203,11 @@ cifs_reconnect(struct TCP_Server_Info *server)
 	cifs_dbg(FYI, "%s: tearing down socket\n", __func__);
 	mutex_lock(&server->srv_mutex);
 	if (server->ssocket) {
-		cifs_dbg(FYI, "State: 0x%x Flags: 0x%lx\n",
-			 server->ssocket->state, server->ssocket->flags);
+		cifs_dbg(FYI, "State: 0x%x Flags: 0x%lx\n", server->ssocket->state,
+			 server->ssocket->flags);
 		kernel_sock_shutdown(server->ssocket, SHUT_WR);
-		cifs_dbg(FYI, "Post shutdown state: 0x%x Flags: 0x%lx\n",
-			 server->ssocket->state, server->ssocket->flags);
+		cifs_dbg(FYI, "Post shutdown state: 0x%x Flags: 0x%lx\n", server->ssocket->state,
+			 server->ssocket->flags);
 		sock_release(server->ssocket);
 		server->ssocket = NULL;
 	}
@@ -309,23 +222,21 @@ cifs_reconnect(struct TCP_Server_Info *server)
 	INIT_LIST_HEAD(&retry_list);
 	cifs_dbg(FYI, "%s: moving mids to private list\n", __func__);
 	spin_lock(&GlobalMid_Lock);
-	list_for_each_safe(tmp, tmp2, &server->pending_mid_q) {
-		mid_entry = list_entry(tmp, struct mid_q_entry, qhead);
-		kref_get(&mid_entry->refcount);
-		if (mid_entry->mid_state == MID_REQUEST_SUBMITTED)
-			mid_entry->mid_state = MID_RETRY_NEEDED;
-		list_move(&mid_entry->qhead, &retry_list);
-		mid_entry->mid_flags |= MID_DELETED;
+	list_for_each_entry_safe(mid, nmid, &server->pending_mid_q, qhead) {
+		kref_get(&mid->refcount);
+		if (mid->mid_state == MID_REQUEST_SUBMITTED)
+			mid->mid_state = MID_RETRY_NEEDED;
+		list_move(&mid->qhead, &retry_list);
+		mid->mid_flags |= MID_DELETED;
 	}
 	spin_unlock(&GlobalMid_Lock);
 	mutex_unlock(&server->srv_mutex);
 
 	cifs_dbg(FYI, "%s: issuing mid callbacks\n", __func__);
-	list_for_each_safe(tmp, tmp2, &retry_list) {
-		mid_entry = list_entry(tmp, struct mid_q_entry, qhead);
-		list_del_init(&mid_entry->qhead);
-		mid_entry->callback(mid_entry);
-		cifs_mid_q_entry_release(mid_entry);
+	list_for_each_entry_safe(mid, nmid, &retry_list, qhead) {
+		list_del_init(&mid->qhead);
+		mid->callback(mid);
+		cifs_mid_q_entry_release(mid);
 	}
 
 	if (cifs_rdma_enabled(server)) {
@@ -333,38 +244,48 @@ cifs_reconnect(struct TCP_Server_Info *server)
 		smbd_destroy(server);
 		mutex_unlock(&server->srv_mutex);
 	}
+}
+
+static bool cifs_tcp_ses_needs_reconnect(struct TCP_Server_Info *server, int num_targets)
+{
+	spin_lock(&GlobalMid_Lock);
+	server->nr_targets = num_targets;
+	if (server->tcpStatus == CifsExiting) {
+		/* the demux thread will exit normally next time through the loop */
+		spin_unlock(&GlobalMid_Lock);
+		wake_up(&server->response_q);
+		return false;
+	}
+	server->tcpStatus = CifsNeedReconnect;
+	spin_unlock(&GlobalMid_Lock);
+	return true;
+}
+
+/*
+ * cifs tcp session reconnection
+ *
+ * mark tcp session as reconnecting so temporarily locked
+ * mark all smb sessions as reconnecting for tcp session
+ * reconnect tcp session
+ * wake up waiters on reconnection? - (not needed currently)
+ */
+static int __cifs_reconnect(struct TCP_Server_Info *server)
+{
+	int rc = 0;
+
+	if (!cifs_tcp_ses_needs_reconnect(server, 1))
+		return 0;
+
+	cifs_mark_tcp_ses_conns_for_reconnect(server);
 
 	do {
 		try_to_freeze();
-
 		mutex_lock(&server->srv_mutex);
 
-
 		if (!cifs_swn_set_server_dstaddr(server)) {
-#ifdef CONFIG_CIFS_DFS_UPCALL
-		if (cifs_sb && cifs_sb->origin_fullpath)
-			/*
-			 * Set up next DFS target server (if any) for reconnect. If DFS
-			 * feature is disabled, then we will retry last server we
-			 * connected to before.
-			 */
-			reconn_set_next_dfs_target(server, cifs_sb, &tgt_list, &tgt_it);
-		else {
-#endif
-			/*
-			 * Resolve the hostname again to make sure that IP address is up-to-date.
-			 */
+			/* resolve the hostname again to make sure that IP address is up-to-date */
 			rc = reconn_set_ipaddr_from_hostname(server);
-			if (rc) {
-				cifs_dbg(FYI, "%s: failed to resolve hostname: %d\n",
-						__func__, rc);
-			}
-
-#ifdef CONFIG_CIFS_DFS_UPCALL
-		}
-#endif
-
-
+			cifs_dbg(FYI, "%s: reconn_set_ipaddr_from_hostname: rc=%d\n", __func__, rc);
 		}
 
 		if (cifs_rdma_enabled(server))
@@ -372,8 +293,8 @@ cifs_reconnect(struct TCP_Server_Info *server)
 		else
 			rc = generic_ip_connect(server);
 		if (rc) {
-			cifs_dbg(FYI, "reconnect error %d\n", rc);
 			mutex_unlock(&server->srv_mutex);
+			cifs_dbg(FYI, "%s: reconnect error %d\n", __func__, rc);
 			msleep(3000);
 		} else {
 			atomic_inc(&tcpSesReconnectCount);
@@ -387,19 +308,6 @@ cifs_reconnect(struct TCP_Server_Info *server)
 		}
 	} while (server->tcpStatus == CifsNeedReconnect);
 
-#ifdef CONFIG_CIFS_DFS_UPCALL
-	if (tgt_it) {
-		rc = dfs_cache_noreq_update_tgthint(cifs_sb->origin_fullpath + 1,
-						    tgt_it);
-		if (rc) {
-			cifs_server_dbg(VFS, "%s: failed to update DFS target hint: rc = %d\n",
-				 __func__, rc);
-		}
-		dfs_cache_free_tgts(&tgt_list);
-	}
-
-	cifs_put_tcp_super(sb);
-#endif
 	if (server->tcpStatus == CifsNeedNegotiate)
 		mod_delayed_work(cifsiod_wq, &server->echo, 0);
 
@@ -407,6 +315,147 @@ cifs_reconnect(struct TCP_Server_Info *server)
 	return rc;
 }
 
+#ifdef CONFIG_CIFS_DFS_UPCALL
+static int __reconnect_target_unlocked(struct TCP_Server_Info *server, const char *target)
+{
+	int rc;
+	char *hostname;
+
+	if (!cifs_swn_set_server_dstaddr(server)) {
+		if (server->hostname != target) {
+			hostname = extract_hostname(target);
+			if (!IS_ERR(hostname)) {
+				kfree(server->hostname);
+				server->hostname = hostname;
+			} else {
+				cifs_dbg(FYI, "%s: couldn't extract hostname or address from dfs target: %ld\n",
+					 __func__, PTR_ERR(hostname));
+				cifs_dbg(FYI, "%s: default to last target server: %s\n", __func__,
+					 server->hostname);
+			}
+		}
+		/* resolve the hostname again to make sure that IP address is up-to-date. */
+		rc = reconn_set_ipaddr_from_hostname(server);
+		cifs_dbg(FYI, "%s: reconn_set_ipaddr_from_hostname: rc=%d\n", __func__, rc);
+	}
+	/* Reconnect the socket */
+	if (cifs_rdma_enabled(server))
+		rc = smbd_reconnect(server);
+	else
+		rc = generic_ip_connect(server);
+
+	return rc;
+}
+
+static int reconnect_target_unlocked(struct TCP_Server_Info *server, struct dfs_cache_tgt_list *tl,
+				     struct dfs_cache_tgt_iterator **target_hint)
+{
+	int rc;
+	struct dfs_cache_tgt_iterator *tit;
+
+	*target_hint = NULL;
+
+	/* If dfs target list is empty, then reconnect to last server */
+	tit = dfs_cache_get_tgt_iterator(tl);
+	if (!tit)
+		return __reconnect_target_unlocked(server, server->hostname);
+
+	/* Otherwise, try every dfs target in @tl */
+	for (; tit; tit = dfs_cache_get_next_tgt(tl, tit)) {
+		rc = __reconnect_target_unlocked(server, dfs_cache_get_tgt_name(tit));
+		if (!rc) {
+			*target_hint = tit;
+			break;
+		}
+	}
+	return rc;
+}
+
+static int reconnect_dfs_server(struct TCP_Server_Info *server)
+{
+	int rc = 0;
+	const char *refpath = server->current_fullpath + 1;
+	struct dfs_cache_tgt_list tl = DFS_CACHE_TGT_LIST_INIT(tl);
+	struct dfs_cache_tgt_iterator *target_hint = NULL;
+	int num_targets = 0;
+
+	/*
+	 * Determine the number of dfs targets the referral path in @cifs_sb resolves to.
+	 *
+	 * smb2_reconnect() needs to know how long it should wait based upon the number of dfs
+	 * targets (server->nr_targets).  It's also possible that the cached referral was cleared
+	 * through /proc/fs/cifs/dfscache or the target list is empty due to server settings after
+	 * refreshing the referral, so, in this case, default it to 1.
+	 */
+	if (!dfs_cache_noreq_find(refpath, NULL, &tl))
+		num_targets = dfs_cache_get_nr_tgts(&tl);
+	if (!num_targets)
+		num_targets = 1;
+
+	if (!cifs_tcp_ses_needs_reconnect(server, num_targets))
+		return 0;
+
+	cifs_mark_tcp_ses_conns_for_reconnect(server);
+
+	do {
+		try_to_freeze();
+		mutex_lock(&server->srv_mutex);
+
+		rc = reconnect_target_unlocked(server, &tl, &target_hint);
+		if (rc) {
+			/* Failed to reconnect socket */
+			mutex_unlock(&server->srv_mutex);
+			cifs_dbg(FYI, "%s: reconnect error %d\n", __func__, rc);
+			msleep(3000);
+			continue;
+		}
+		/*
+		 * Socket was created.  Update tcp session status to CifsNeedNegotiate so that a
+		 * process waiting for reconnect will know it needs to re-establish session and tcon
+		 * through the reconnected target server.
+		 */
+		atomic_inc(&tcpSesReconnectCount);
+		set_credits(server, 1);
+		spin_lock(&GlobalMid_Lock);
+		if (server->tcpStatus != CifsExiting)
+			server->tcpStatus = CifsNeedNegotiate;
+		spin_unlock(&GlobalMid_Lock);
+		cifs_swn_reset_server_dstaddr(server);
+		mutex_unlock(&server->srv_mutex);
+	} while (server->tcpStatus == CifsNeedReconnect);
+
+	if (target_hint)
+		dfs_cache_noreq_update_tgthint(refpath, target_hint);
+
+	dfs_cache_free_tgts(&tl);
+
+	/* Need to set up echo worker again once connection has been established */
+	if (server->tcpStatus == CifsNeedNegotiate)
+		mod_delayed_work(cifsiod_wq, &server->echo, 0);
+
+	wake_up(&server->response_q);
+	return rc;
+}
+
+int cifs_reconnect(struct TCP_Server_Info *server)
+{
+	/* If tcp session is not an dfs connection, then reconnect to last target server */
+	spin_lock(&cifs_tcp_ses_lock);
+	if (!server->is_dfs_conn || !server->origin_fullpath || !server->leaf_fullpath) {
+		spin_unlock(&cifs_tcp_ses_lock);
+		return __cifs_reconnect(server);
+	}
+	spin_unlock(&cifs_tcp_ses_lock);
+
+	return reconnect_dfs_server(server);
+}
+#else
+int cifs_reconnect(struct TCP_Server_Info *server)
+{
+	return __cifs_reconnect(server);
+}
+#endif
+
 static void
 cifs_echo_request(struct work_struct *work)
 {
@@ -665,13 +714,14 @@ dequeue_mid(struct mid_q_entry *mid, bool malformed)
 	 * Trying to handle/dequeue a mid after the send_recv()
 	 * function has finished processing it is a bug.
 	 */
-	if (mid->mid_flags & MID_DELETED)
+	if (mid->mid_flags & MID_DELETED) {
+		spin_unlock(&GlobalMid_Lock);
 		pr_warn_once("trying to dequeue a deleted mid\n");
-	else {
+	} else {
 		list_del_init(&mid->qhead);
 		mid->mid_flags |= MID_DELETED;
+		spin_unlock(&GlobalMid_Lock);
 	}
-	spin_unlock(&GlobalMid_Lock);
 }
 
 static unsigned int
@@ -794,6 +844,10 @@ static void clean_demultiplex_info(struct TCP_Server_Info *server)
 		 */
 	}
 
+#ifdef CONFIG_CIFS_DFS_UPCALL
+	kfree(server->origin_fullpath);
+	kfree(server->leaf_fullpath);
+#endif
 	kfree(server);
 
 	length = atomic_dec_return(&tcpSesAllocCount);
@@ -1217,7 +1271,13 @@ static int match_server(struct TCP_Server_Info *server, struct smb3_fs_context *
 {
 	struct sockaddr *addr = (struct sockaddr *)&ctx->dstaddr;
 
-	if (ctx->nosharesock)
+	if (ctx->nosharesock) {
+		server->nosharesock = true;
+		return 0;
+	}
+
+	/* this server does not share socket */
+	if (server->nosharesock)
 		return 0;
 
 	/* If multidialect negotiation see if existing sessions match one */
@@ -1283,7 +1343,7 @@ cifs_find_tcp_session(struct smb3_fs_context *ctx)
 		 * Skip ses channels since they're only handled in lower layers
 		 * (e.g. cifs_send_recv).
 		 */
-		if (server->is_channel || !match_server(server, ctx))
+		if (CIFS_SERVER_IS_CHAN(server) || !match_server(server, ctx))
 			continue;
 
 		++server->srv_count;
@@ -1314,6 +1374,10 @@ cifs_put_tcp_session(struct TCP_Server_Info *server, int from_reconnect)
 	list_del_init(&server->tcp_ses_list);
 	spin_unlock(&cifs_tcp_ses_lock);
 
+	/* For secondary channels, we pick up ref-count on the primary server */
+	if (CIFS_SERVER_IS_CHAN(server))
+		cifs_put_tcp_session(server->primary_server, from_reconnect);
+
 	cancel_delayed_work_sync(&server->echo);
 	cancel_delayed_work_sync(&server->resolve);
 
@@ -1333,7 +1397,10 @@ cifs_put_tcp_session(struct TCP_Server_Info *server, int from_reconnect)
 	spin_unlock(&GlobalMid_Lock);
 
 	cifs_crypto_secmech_release(server);
-	cifs_fscache_release_client_cookie(server);
+
+	/* fscache server cookies are based on primary channel only */
+	if (!CIFS_SERVER_IS_CHAN(server))
+		cifs_fscache_release_client_cookie(server);
 
 	kfree(server->session_key.response);
 	server->session_key.response = NULL;
@@ -1346,7 +1413,8 @@ cifs_put_tcp_session(struct TCP_Server_Info *server, int from_reconnect)
 }
 
 struct TCP_Server_Info *
-cifs_get_tcp_session(struct smb3_fs_context *ctx)
+cifs_get_tcp_session(struct smb3_fs_context *ctx,
+		     struct TCP_Server_Info *primary_server)
 {
 	struct TCP_Server_Info *tcp_ses = NULL;
 	int rc;
@@ -1383,6 +1451,10 @@ cifs_get_tcp_session(struct smb3_fs_context *ctx)
 	tcp_ses->in_flight = 0;
 	tcp_ses->max_in_flight = 0;
 	tcp_ses->credits = 1;
+	if (primary_server) {
+		++primary_server->srv_count;
+		tcp_ses->primary_server = primary_server;
+	}
 	init_waitqueue_head(&tcp_ses->response_q);
 	init_waitqueue_head(&tcp_ses->request_q);
 	INIT_LIST_HEAD(&tcp_ses->pending_mid_q);
@@ -1403,6 +1475,9 @@ cifs_get_tcp_session(struct smb3_fs_context *ctx)
 	INIT_DELAYED_WORK(&tcp_ses->resolve, cifs_resolve_server);
 	INIT_DELAYED_WORK(&tcp_ses->reconnect, smb2_reconnect_server);
 	mutex_init(&tcp_ses->reconnect_mutex);
+#ifdef CONFIG_CIFS_DFS_UPCALL
+	mutex_init(&tcp_ses->refpath_lock);
+#endif
 	memcpy(&tcp_ses->srcaddr, &ctx->srcaddr,
 	       sizeof(tcp_ses->srcaddr));
 	memcpy(&tcp_ses->dstaddr, &ctx->dstaddr,
@@ -1481,7 +1556,9 @@ cifs_get_tcp_session(struct smb3_fs_context *ctx)
 	list_add(&tcp_ses->tcp_ses_list, &cifs_tcp_ses_list);
 	spin_unlock(&cifs_tcp_ses_lock);
 
-	cifs_fscache_get_client_cookie(tcp_ses);
+	/* fscache server cookies are based on primary channel only */
+	if (!CIFS_SERVER_IS_CHAN(tcp_ses))
+		cifs_fscache_get_client_cookie(tcp_ses);
 
 	/* queue echo request delayed work */
 	queue_delayed_work(cifsiod_wq, &tcp_ses->echo, tcp_ses->echo_interval);
@@ -1501,6 +1578,8 @@ cifs_get_tcp_session(struct smb3_fs_context *ctx)
 
 out_err:
 	if (tcp_ses) {
+		if (CIFS_SERVER_IS_CHAN(tcp_ses))
+			cifs_put_tcp_session(tcp_ses->primary_server, false);
 		kfree(tcp_ses->hostname);
 		if (tcp_ses->ssocket)
 			sock_release(tcp_ses->ssocket);
@@ -1519,8 +1598,12 @@ static int match_session(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 	 * If an existing session is limited to less channels than
 	 * requested, it should not be reused
 	 */
-	if (ses->chan_max < ctx->max_channels)
+	spin_lock(&ses->chan_lock);
+	if (ses->chan_max < ctx->max_channels) {
+		spin_unlock(&ses->chan_lock);
 		return 0;
+	}
+	spin_unlock(&ses->chan_lock);
 
 	switch (ses->sectype) {
 	case Kerberos:
@@ -1655,6 +1738,7 @@ cifs_find_smb_ses(struct TCP_Server_Info *server, struct smb3_fs_context *ctx)
 void cifs_put_smb_ses(struct cifs_ses *ses)
 {
 	unsigned int rc, xid;
+	unsigned int chan_count;
 	struct TCP_Server_Info *server = ses->server;
 	cifs_dbg(FYI, "%s: ses_count=%d\n", __func__, ses->ses_count);
 
@@ -1696,12 +1780,24 @@ void cifs_put_smb_ses(struct cifs_ses *ses)
 	list_del_init(&ses->smb_ses_list);
 	spin_unlock(&cifs_tcp_ses_lock);
 
+	spin_lock(&ses->chan_lock);
+	chan_count = ses->chan_count;
+	spin_unlock(&ses->chan_lock);
+
 	/* close any extra channels */
-	if (ses->chan_count > 1) {
+	if (chan_count > 1) {
 		int i;
 
-		for (i = 1; i < ses->chan_count; i++)
+		for (i = 1; i < chan_count; i++) {
+			/*
+			 * note: for now, we're okay accessing ses->chans
+			 * without chan_lock. But when chans can go away, we'll
+			 * need to introduce ref counting to make sure that chan
+			 * is not freed from under us.
+			 */
 			cifs_put_tcp_session(ses->chans[i].server, 0);
+			ses->chans[i].server = NULL;
+		}
 	}
 
 	sesInfoFree(ses);
@@ -1885,16 +1981,18 @@ cifs_get_smb_ses(struct TCP_Server_Info *server, struct smb3_fs_context *ctx)
 			 ses->status);
 
 		mutex_lock(&ses->session_mutex);
-		rc = cifs_negotiate_protocol(xid, ses);
-		if (rc) {
-			mutex_unlock(&ses->session_mutex);
-			/* problem -- put our ses reference */
-			cifs_put_smb_ses(ses);
-			free_xid(xid);
-			return ERR_PTR(rc);
-		}
 		if (ses->need_reconnect) {
 			cifs_dbg(FYI, "Session needs reconnect\n");
+
+			rc = cifs_negotiate_protocol(xid, ses);
+			if (rc) {
+				mutex_unlock(&ses->session_mutex);
+				/* problem -- put our ses reference */
+				cifs_put_smb_ses(ses);
+				free_xid(xid);
+				return ERR_PTR(rc);
+			}
+
 			rc = cifs_setup_session(xid, ses,
 						ctx->local_nls);
 			if (rc) {
@@ -1942,6 +2040,12 @@ cifs_get_smb_ses(struct TCP_Server_Info *server, struct smb3_fs_context *ctx)
 		if (!ses->domainName)
 			goto get_ses_fail;
 	}
+	if (ctx->workstation_name) {
+		ses->workstation_name = kstrdup(ctx->workstation_name,
+						GFP_KERNEL);
+		if (!ses->workstation_name)
+			goto get_ses_fail;
+	}
 	if (ctx->domainauto)
 		ses->domainAuto = ctx->domainauto;
 	ses->cred_uid = ctx->cred_uid;
@@ -1952,9 +2056,11 @@ cifs_get_smb_ses(struct TCP_Server_Info *server, struct smb3_fs_context *ctx)
 	mutex_lock(&ses->session_mutex);
 
 	/* add server as first channel */
+	spin_lock(&ses->chan_lock);
 	ses->chans[0].server = server;
 	ses->chan_count = 1;
 	ses->chan_max = ctx->multichannel ? ctx->max_channels:1;
+	spin_unlock(&ses->chan_lock);
 
 	rc = cifs_negotiate_protocol(xid, ses);
 	if (!rc)
@@ -2286,8 +2392,6 @@ cifs_get_tcon(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 	list_add(&tcon->tcon_list, &ses->tcon_list);
 	spin_unlock(&cifs_tcp_ses_lock);
 
-	cifs_fscache_get_super_cookie(tcon);
-
 	return tcon;
 
 out_fail:
@@ -2849,73 +2953,64 @@ int cifs_setup_cifs_sb(struct cifs_sb_info *cifs_sb)
 }
 
 /* Release all succeed connections */
-static inline void mount_put_conns(struct cifs_sb_info *cifs_sb,
-				   unsigned int xid,
-				   struct TCP_Server_Info *server,
-				   struct cifs_ses *ses, struct cifs_tcon *tcon)
+static inline void mount_put_conns(struct mount_ctx *mnt_ctx)
 {
 	int rc = 0;
 
-	if (tcon)
-		cifs_put_tcon(tcon);
-	else if (ses)
-		cifs_put_smb_ses(ses);
-	else if (server)
-		cifs_put_tcp_session(server, 0);
-	cifs_sb->mnt_cifs_flags &= ~CIFS_MOUNT_POSIX_PATHS;
-	free_xid(xid);
+	if (mnt_ctx->tcon)
+		cifs_put_tcon(mnt_ctx->tcon);
+	else if (mnt_ctx->ses)
+		cifs_put_smb_ses(mnt_ctx->ses);
+	else if (mnt_ctx->server)
+		cifs_put_tcp_session(mnt_ctx->server, 0);
+	mnt_ctx->cifs_sb->mnt_cifs_flags &= ~CIFS_MOUNT_POSIX_PATHS;
+	free_xid(mnt_ctx->xid);
 }
 
 /* Get connections for tcp, ses and tcon */
-static int mount_get_conns(struct smb3_fs_context *ctx, struct cifs_sb_info *cifs_sb,
-			   unsigned int *xid,
-			   struct TCP_Server_Info **nserver,
-			   struct cifs_ses **nses, struct cifs_tcon **ntcon)
+static int mount_get_conns(struct mount_ctx *mnt_ctx)
 {
 	int rc = 0;
-	struct TCP_Server_Info *server;
-	struct cifs_ses *ses;
-	struct cifs_tcon *tcon;
+	struct TCP_Server_Info *server = NULL;
+	struct cifs_ses *ses = NULL;
+	struct cifs_tcon *tcon = NULL;
+	struct smb3_fs_context *ctx = mnt_ctx->fs_ctx;
+	struct cifs_sb_info *cifs_sb = mnt_ctx->cifs_sb;
+	unsigned int xid;
 
-	*nserver = NULL;
-	*nses = NULL;
-	*ntcon = NULL;
-
-	*xid = get_xid();
+	xid = get_xid();
 
 	/* get a reference to a tcp session */
-	server = cifs_get_tcp_session(ctx);
+	server = cifs_get_tcp_session(ctx, NULL);
 	if (IS_ERR(server)) {
 		rc = PTR_ERR(server);
-		return rc;
+		server = NULL;
+		goto out;
 	}
 
-	*nserver = server;
-
 	/* get a reference to a SMB session */
 	ses = cifs_get_smb_ses(server, ctx);
 	if (IS_ERR(ses)) {
 		rc = PTR_ERR(ses);
-		return rc;
+		ses = NULL;
+		goto out;
 	}
 
-	*nses = ses;
-
 	if ((ctx->persistent == true) && (!(ses->server->capabilities &
 					    SMB2_GLOBAL_CAP_PERSISTENT_HANDLES))) {
 		cifs_server_dbg(VFS, "persistent handles not supported by server\n");
-		return -EOPNOTSUPP;
+		rc = -EOPNOTSUPP;
+		goto out;
 	}
 
 	/* search for existing tcon to this server share */
 	tcon = cifs_get_tcon(ses, ctx);
 	if (IS_ERR(tcon)) {
 		rc = PTR_ERR(tcon);
-		return rc;
+		tcon = NULL;
+		goto out;
 	}
 
-	*ntcon = tcon;
-
 	/* if new SMB3.11 POSIX extensions are supported do not remap / and \ */
 	if (tcon->posix_extensions)
 		cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_POSIX_PATHS;
@@ -2926,17 +3021,19 @@ static int mount_get_conns(struct smb3_fs_context *ctx, struct cifs_sb_info *cif
 		 * reset of caps checks mount to see if unix extensions disabled
 		 * for just this mount.
 		 */
-		reset_cifs_unix_caps(*xid, tcon, cifs_sb, ctx);
+		reset_cifs_unix_caps(xid, tcon, cifs_sb, ctx);
 		if ((tcon->ses->server->tcpStatus == CifsNeedReconnect) &&
 		    (le64_to_cpu(tcon->fsUnixInfo.Capability) &
-		     CIFS_UNIX_TRANSPORT_ENCRYPTION_MANDATORY_CAP))
-			return -EACCES;
+		     CIFS_UNIX_TRANSPORT_ENCRYPTION_MANDATORY_CAP)) {
+			rc = -EACCES;
+			goto out;
+		}
 	} else
 		tcon->unix_ext = 0; /* server does not support them */
 
 	/* do not care if a following call succeed - informational */
 	if (!tcon->pipe && server->ops->qfs_tcon) {
-		server->ops->qfs_tcon(*xid, tcon, cifs_sb);
+		server->ops->qfs_tcon(xid, tcon, cifs_sb);
 		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RO_CACHE) {
 			if (tcon->fsDevInfo.DeviceCharacteristics &
 			    cpu_to_le32(FILE_READ_ONLY_DEVICE))
@@ -2946,6 +3043,12 @@ static int mount_get_conns(struct smb3_fs_context *ctx, struct cifs_sb_info *cif
 				cifs_dbg(VFS, "read only mount of RW share\n");
 			/* no need to log a RW mount of a typical RW share */
 		}
+		/*
+		 * The cookie is initialized from volume info returned above.
+		 * Inside cifs_fscache_get_super_cookie it checks
+		 * that we do not get super cookie twice.
+		 */
+		cifs_fscache_get_super_cookie(tcon);
 	}
 
 	/*
@@ -2960,7 +3063,13 @@ static int mount_get_conns(struct smb3_fs_context *ctx, struct cifs_sb_info *cif
 	    (cifs_sb->ctx->rsize > server->ops->negotiate_rsize(tcon, ctx)))
 		cifs_sb->ctx->rsize = server->ops->negotiate_rsize(tcon, ctx);
 
-	return 0;
+out:
+	mnt_ctx->server = server;
+	mnt_ctx->ses = ses;
+	mnt_ctx->tcon = tcon;
+	mnt_ctx->xid = xid;
+
+	return rc;
 }
 
 static int mount_setup_tlink(struct cifs_sb_info *cifs_sb, struct cifs_ses *ses,
@@ -2990,18 +3099,17 @@ static int mount_setup_tlink(struct cifs_sb_info *cifs_sb, struct cifs_ses *ses,
 }
 
 #ifdef CONFIG_CIFS_DFS_UPCALL
-static int mount_get_dfs_conns(struct smb3_fs_context *ctx, struct cifs_sb_info *cifs_sb,
-			       unsigned int *xid, struct TCP_Server_Info **nserver,
-			       struct cifs_ses **nses, struct cifs_tcon **ntcon)
+/* Get unique dfs connections */
+static int mount_get_dfs_conns(struct mount_ctx *mnt_ctx)
 {
 	int rc;
 
-	ctx->nosharesock = true;
-	rc = mount_get_conns(ctx, cifs_sb, xid, nserver, nses, ntcon);
-	if (*nserver) {
+	mnt_ctx->fs_ctx->nosharesock = true;
+	rc = mount_get_conns(mnt_ctx);
+	if (mnt_ctx->server) {
 		cifs_dbg(FYI, "%s: marking tcp session as a dfs connection\n", __func__);
 		spin_lock(&cifs_tcp_ses_lock);
-		(*nserver)->is_dfs_conn = true;
+		mnt_ctx->server->is_dfs_conn = true;
 		spin_unlock(&cifs_tcp_ses_lock);
 	}
 	return rc;
@@ -3043,190 +3151,38 @@ build_unc_path_to_root(const struct smb3_fs_context *ctx,
 }
 
 /*
- * expand_dfs_referral - Perform a dfs referral query and update the cifs_sb
+ * expand_dfs_referral - Update cifs_sb from dfs referral path
  *
- * If a referral is found, cifs_sb->ctx->mount_options will be (re-)allocated
- * to a string containing updated options for the submount.  Otherwise it
- * will be left untouched.
- *
- * Returns the rc from get_dfs_path to the caller, which can be used to
- * determine whether there were referrals.
+ * cifs_sb->ctx->mount_options will be (re-)allocated to a string containing updated options for the
+ * submount.  Otherwise it will be left untouched.
  */
-static int
-expand_dfs_referral(const unsigned int xid, struct cifs_ses *ses,
-		    struct smb3_fs_context *ctx, struct cifs_sb_info *cifs_sb,
-		    char *ref_path)
+static int expand_dfs_referral(struct mount_ctx *mnt_ctx, const char *full_path,
+			       struct dfs_info3_param *referral)
 {
 	int rc;
-	struct dfs_info3_param referral = {0};
-	char *full_path = NULL, *mdata = NULL;
+	struct cifs_sb_info *cifs_sb = mnt_ctx->cifs_sb;
+	struct smb3_fs_context *ctx = mnt_ctx->fs_ctx;
+	char *fake_devname = NULL, *mdata = NULL;
 
-	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_DFS)
-		return -EREMOTE;
-
-	full_path = build_unc_path_to_root(ctx, cifs_sb, true);
-	if (IS_ERR(full_path))
-		return PTR_ERR(full_path);
-
-	rc = dfs_cache_find(xid, ses, cifs_sb->local_nls, cifs_remap(cifs_sb),
-			    ref_path, &referral, NULL);
-	if (!rc) {
-		char *fake_devname = NULL;
-
-		mdata = cifs_compose_mount_options(cifs_sb->ctx->mount_options,
-						   full_path + 1, &referral,
-						   &fake_devname);
-		free_dfs_info_param(&referral);
-
-		if (IS_ERR(mdata)) {
-			rc = PTR_ERR(mdata);
-			mdata = NULL;
-		} else {
-			/*
-			 * We can not clear out the whole structure since we
-			 * no longer have an explicit function to parse
-			 * a mount-string. Instead we need to clear out the
-			 * individual fields that are no longer valid.
-			 */
-			kfree(ctx->prepath);
-			ctx->prepath = NULL;
-			rc = cifs_setup_volume_info(ctx, mdata, fake_devname);
-		}
-		kfree(fake_devname);
-		kfree(cifs_sb->ctx->mount_options);
-		cifs_sb->ctx->mount_options = mdata;
-	}
-	kfree(full_path);
-	return rc;
-}
-
-static int get_next_dfs_tgt(struct dfs_cache_tgt_list *tgt_list,
-			    struct dfs_cache_tgt_iterator **tgt_it)
-{
-	if (!*tgt_it)
-		*tgt_it = dfs_cache_get_tgt_iterator(tgt_list);
-	else
-		*tgt_it = dfs_cache_get_next_tgt(tgt_list, *tgt_it);
-	return !*tgt_it ? -EHOSTDOWN : 0;
-}
-
-static int update_vol_info(const struct dfs_cache_tgt_iterator *tgt_it,
-			   struct smb3_fs_context *fake_ctx, struct smb3_fs_context *ctx)
-{
-	const char *tgt = dfs_cache_get_tgt_name(tgt_it);
-	int len = strlen(tgt) + 2;
-	char *new_unc;
-
-	new_unc = kmalloc(len, GFP_KERNEL);
-	if (!new_unc)
-		return -ENOMEM;
-	scnprintf(new_unc, len, "\\%s", tgt);
-
-	kfree(ctx->UNC);
-	ctx->UNC = new_unc;
-
-	if (fake_ctx->prepath) {
-		kfree(ctx->prepath);
-		ctx->prepath = fake_ctx->prepath;
-		fake_ctx->prepath = NULL;
-	}
-	memcpy(&ctx->dstaddr, &fake_ctx->dstaddr, sizeof(ctx->dstaddr));
-
-	return 0;
-}
-
-static int do_dfs_failover(const char *path, const char *full_path, struct cifs_sb_info *cifs_sb,
-			   struct smb3_fs_context *ctx, struct cifs_ses *root_ses,
-			   unsigned int *xid, struct TCP_Server_Info **server,
-			   struct cifs_ses **ses, struct cifs_tcon **tcon)
-{
-	int rc;
-	char *npath = NULL;
-	struct dfs_cache_tgt_list tgt_list = DFS_CACHE_TGT_LIST_INIT(tgt_list);
-	struct dfs_cache_tgt_iterator *tgt_it = NULL;
-	struct smb3_fs_context tmp_ctx = {NULL};
-
-	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_DFS)
-		return -EOPNOTSUPP;
-
-	npath = dfs_cache_canonical_path(path, cifs_sb->local_nls, cifs_remap(cifs_sb));
-	if (IS_ERR(npath))
-		return PTR_ERR(npath);
-
-	cifs_dbg(FYI, "%s: path=%s full_path=%s\n", __func__, npath, full_path);
-
-	rc = dfs_cache_noreq_find(npath, NULL, &tgt_list);
-	if (rc)
-		goto out;
-	/*
-	 * We use a 'tmp_ctx' here because we need pass it down to the mount_{get,put} functions to
-	 * test connection against new DFS targets.
-	 */
-	rc = smb3_fs_context_dup(&tmp_ctx, ctx);
-	if (rc)
-		goto out;
-
-	for (;;) {
-		struct dfs_info3_param ref = {0};
-		char *fake_devname = NULL, *mdata = NULL;
-
-		/* Get next DFS target server - if any */
-		rc = get_next_dfs_tgt(&tgt_list, &tgt_it);
-		if (rc)
-			break;
-
-		rc = dfs_cache_get_tgt_referral(npath, tgt_it, &ref);
-		if (rc)
-			break;
-
-		cifs_dbg(FYI, "%s: old ctx: UNC=%s prepath=%s\n", __func__, tmp_ctx.UNC,
-			 tmp_ctx.prepath);
-
-		mdata = cifs_compose_mount_options(cifs_sb->ctx->mount_options, full_path + 1, &ref,
-						   &fake_devname);
-		free_dfs_info_param(&ref);
-
-		if (IS_ERR(mdata)) {
-			rc = PTR_ERR(mdata);
-			mdata = NULL;
-		} else
-			rc = cifs_setup_volume_info(&tmp_ctx, mdata, fake_devname);
-
-		kfree(mdata);
-		kfree(fake_devname);
-
-		if (rc)
-			break;
-
-		cifs_dbg(FYI, "%s: new ctx: UNC=%s prepath=%s\n", __func__, tmp_ctx.UNC,
-			 tmp_ctx.prepath);
-
-		mount_put_conns(cifs_sb, *xid, *server, *ses, *tcon);
-		rc = mount_get_dfs_conns(&tmp_ctx, cifs_sb, xid, server, ses, tcon);
-		if (!rc || (*server && *ses)) {
-			/*
-			 * We were able to connect to new target server. Update current context with
-			 * new target server.
-			 */
-			rc = update_vol_info(tgt_it, &tmp_ctx, ctx);
-			break;
-		}
-	}
-	if (!rc) {
-		cifs_dbg(FYI, "%s: final ctx: UNC=%s prepath=%s\n", __func__, tmp_ctx.UNC,
-			 tmp_ctx.prepath);
+	mdata = cifs_compose_mount_options(cifs_sb->ctx->mount_options, full_path + 1, referral,
+					   &fake_devname);
+	if (IS_ERR(mdata)) {
+		rc = PTR_ERR(mdata);
+		mdata = NULL;
+	} else {
 		/*
-		 * Update DFS target hint in DFS referral cache with the target server we
-		 * successfully reconnected to.
+		 * We can not clear out the whole structure since we no longer have an explicit
+		 * function to parse a mount-string. Instead we need to clear out the individual
+		 * fields that are no longer valid.
 		 */
-		rc = dfs_cache_update_tgthint(*xid, root_ses ? root_ses : *ses, cifs_sb->local_nls,
-					      cifs_remap(cifs_sb), path, tgt_it);
+		kfree(ctx->prepath);
+		ctx->prepath = NULL;
+		rc = cifs_setup_volume_info(ctx, mdata, fake_devname);
 	}
+	kfree(fake_devname);
+	kfree(cifs_sb->ctx->mount_options);
+	cifs_sb->ctx->mount_options = mdata;
 
-out:
-	kfree(npath);
-	smb3_cleanup_fs_context_contents(&tmp_ctx);
-	dfs_cache_free_tgts(&tgt_list);
 	return rc;
 }
 #endif
@@ -3333,12 +3289,14 @@ cifs_are_all_path_components_accessible(struct TCP_Server_Info *server,
  * Check if path is remote (e.g. a DFS share). Return -EREMOTE if it is,
  * otherwise 0.
  */
-static int is_path_remote(struct cifs_sb_info *cifs_sb, struct smb3_fs_context *ctx,
-			  const unsigned int xid,
-			  struct TCP_Server_Info *server,
-			  struct cifs_tcon *tcon)
+static int is_path_remote(struct mount_ctx *mnt_ctx)
 {
 	int rc;
+	struct cifs_sb_info *cifs_sb = mnt_ctx->cifs_sb;
+	struct TCP_Server_Info *server = mnt_ctx->server;
+	unsigned int xid = mnt_ctx->xid;
+	struct cifs_tcon *tcon = mnt_ctx->tcon;
+	struct smb3_fs_context *ctx = mnt_ctx->fs_ctx;
 	char *full_path;
 
 	if (!server->ops->is_path_accessible)
@@ -3376,111 +3334,26 @@ static int is_path_remote(struct cifs_sb_info *cifs_sb, struct smb3_fs_context *
 }
 
 #ifdef CONFIG_CIFS_DFS_UPCALL
-static void set_root_ses(struct cifs_sb_info *cifs_sb, const uuid_t *mount_id, struct cifs_ses *ses,
-			 struct cifs_ses **root_ses)
+static void set_root_ses(struct mount_ctx *mnt_ctx)
 {
-	if (ses) {
+	if (mnt_ctx->ses) {
 		spin_lock(&cifs_tcp_ses_lock);
-		ses->ses_count++;
+		mnt_ctx->ses->ses_count++;
 		spin_unlock(&cifs_tcp_ses_lock);
-		dfs_cache_add_refsrv_session(mount_id, ses);
+		dfs_cache_add_refsrv_session(&mnt_ctx->mount_id, mnt_ctx->ses);
 	}
-	*root_ses = ses;
+	mnt_ctx->root_ses = mnt_ctx->ses;
 }
 
-/* Set up next dfs prefix path in @dfs_path */
-static int next_dfs_prepath(struct cifs_sb_info *cifs_sb, struct smb3_fs_context *ctx,
-			    const unsigned int xid, struct TCP_Server_Info *server,
-			    struct cifs_tcon *tcon, char **dfs_path)
-{
-	char *path, *npath;
-	int added_treename = is_tcon_dfs(tcon);
-	int rc;
-
-	path = cifs_build_path_to_root(ctx, cifs_sb, tcon, added_treename);
-	if (!path)
-		return -ENOMEM;
-
-	rc = is_path_remote(cifs_sb, ctx, xid, server, tcon);
-	if (rc == -EREMOTE) {
-		struct smb3_fs_context v = {NULL};
-		/* if @path contains a tree name, skip it in the prefix path */
-		if (added_treename) {
-			rc = smb3_parse_devname(path, &v);
-			if (rc)
-				goto out;
-			npath = build_unc_path_to_root(&v, cifs_sb, true);
-			smb3_cleanup_fs_context_contents(&v);
-		} else {
-			v.UNC = ctx->UNC;
-			v.prepath = path + 1;
-			npath = build_unc_path_to_root(&v, cifs_sb, true);
-		}
-
-		if (IS_ERR(npath)) {
-			rc = PTR_ERR(npath);
-			goto out;
-		}
-
-		kfree(*dfs_path);
-		*dfs_path = npath;
-		rc = -EREMOTE;
-	}
-
-out:
-	kfree(path);
-	return rc;
-}
-
-/* Check if resolved targets can handle any DFS referrals */
-static int is_referral_server(const char *ref_path, struct cifs_sb_info *cifs_sb,
-			      struct cifs_tcon *tcon, bool *ref_server)
+static int is_dfs_mount(struct mount_ctx *mnt_ctx, bool *isdfs, struct dfs_cache_tgt_list *root_tl)
 {
 	int rc;
-	struct dfs_info3_param ref = {0};
+	struct cifs_sb_info *cifs_sb = mnt_ctx->cifs_sb;
+	struct smb3_fs_context *ctx = mnt_ctx->fs_ctx;
 
-	cifs_dbg(FYI, "%s: ref_path=%s\n", __func__, ref_path);
+	*isdfs = true;
 
-	if (is_tcon_dfs(tcon)) {
-		*ref_server = true;
-	} else {
-		char *npath;
-
-		npath = dfs_cache_canonical_path(ref_path, cifs_sb->local_nls, cifs_remap(cifs_sb));
-		if (IS_ERR(npath))
-			return PTR_ERR(npath);
-
-		rc = dfs_cache_noreq_find(npath, &ref, NULL);
-		kfree(npath);
-		if (rc) {
-			cifs_dbg(VFS, "%s: dfs_cache_noreq_find: failed (rc=%d)\n", __func__, rc);
-			return rc;
-		}
-		cifs_dbg(FYI, "%s: ref.flags=0x%x\n", __func__, ref.flags);
-		/*
-		 * Check if all targets are capable of handling DFS referrals as per
-		 * MS-DFSC 2.2.4 RESP_GET_DFS_REFERRAL.
-		 */
-		*ref_server = !!(ref.flags & DFSREF_REFERRAL_SERVER);
-		free_dfs_info_param(&ref);
-	}
-	return 0;
-}
-
-int cifs_mount(struct cifs_sb_info *cifs_sb, struct smb3_fs_context *ctx)
-{
-	int rc = 0;
-	unsigned int xid;
-	struct TCP_Server_Info *server = NULL;
-	struct cifs_ses *ses = NULL, *root_ses = NULL;
-	struct cifs_tcon *tcon = NULL;
-	int count = 0;
-	uuid_t mount_id = {0};
-	char *ref_path = NULL, *full_path = NULL;
-	char *oldmnt = NULL;
-	bool ref_server = false;
-
-	rc = mount_get_conns(ctx, cifs_sb, &xid, &server, &ses, &tcon);
+	rc = mount_get_conns(mnt_ctx);
 	/*
 	 * If called with 'nodfs' mount option, then skip DFS resolving.  Otherwise unconditionally
 	 * try to get an DFS referral (even cached) to determine whether it is an DFS mount.
@@ -3489,167 +3362,261 @@ int cifs_mount(struct cifs_sb_info *cifs_sb, struct smb3_fs_context *ctx)
 	 * to respond with PATH_NOT_COVERED to requests that include the prefix.
 	 */
 	if ((cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_DFS) ||
-	    dfs_cache_find(xid, ses, cifs_sb->local_nls, cifs_remap(cifs_sb), ctx->UNC + 1, NULL,
-			   NULL)) {
+	    dfs_cache_find(mnt_ctx->xid, mnt_ctx->ses, cifs_sb->local_nls, cifs_remap(cifs_sb),
+			   ctx->UNC + 1, NULL, root_tl)) {
 		if (rc)
-			goto error;
+			return rc;
 		/* Check if it is fully accessible and then mount it */
-		rc = is_path_remote(cifs_sb, ctx, xid, server, tcon);
+		rc = is_path_remote(mnt_ctx);
 		if (!rc)
-			goto out;
-		if (rc != -EREMOTE)
-			goto error;
+			*isdfs = false;
+		else if (rc != -EREMOTE)
+			return rc;
+	}
+	return 0;
+}
+
+static int connect_dfs_target(struct mount_ctx *mnt_ctx, const char *full_path,
+			      const char *ref_path, struct dfs_cache_tgt_iterator *tit)
+{
+	int rc;
+	struct dfs_info3_param ref = {};
+	struct cifs_sb_info *cifs_sb = mnt_ctx->cifs_sb;
+	char *oldmnt = cifs_sb->ctx->mount_options;
+
+	rc = dfs_cache_get_tgt_referral(ref_path, tit, &ref);
+	if (rc)
+		goto out;
+
+	rc = expand_dfs_referral(mnt_ctx, full_path, &ref);
+	if (rc)
+		goto out;
+
+	/* Connect to new target only if we were redirected (e.g. mount options changed) */
+	if (oldmnt != cifs_sb->ctx->mount_options) {
+		mount_put_conns(mnt_ctx);
+		rc = mount_get_dfs_conns(mnt_ctx);
+	}
+	if (!rc) {
+		if (cifs_is_referral_server(mnt_ctx->tcon, &ref))
+			set_root_ses(mnt_ctx);
+		rc = dfs_cache_update_tgthint(mnt_ctx->xid, mnt_ctx->root_ses, cifs_sb->local_nls,
+					      cifs_remap(cifs_sb), ref_path, tit);
 	}
 
-	mount_put_conns(cifs_sb, xid, server, ses, tcon);
-	/*
-	 * Ignore error check here because we may failover to other targets from cached a
-	 * referral.
-	 */
-	(void)mount_get_dfs_conns(ctx, cifs_sb, &xid, &server, &ses, &tcon);
+out:
+	free_dfs_info_param(&ref);
+	return rc;
+}
 
-	/* Get path of DFS root */
-	ref_path = build_unc_path_to_root(ctx, cifs_sb, false);
-	if (IS_ERR(ref_path)) {
-		rc = PTR_ERR(ref_path);
-		ref_path = NULL;
-		goto error;
+static int connect_dfs_root(struct mount_ctx *mnt_ctx, struct dfs_cache_tgt_list *root_tl)
+{
+	int rc;
+	char *full_path;
+	struct cifs_sb_info *cifs_sb = mnt_ctx->cifs_sb;
+	struct smb3_fs_context *ctx = mnt_ctx->fs_ctx;
+	struct dfs_cache_tgt_iterator *tit;
+
+	/* Put initial connections as they might be shared with other mounts.  We need unique dfs
+	 * connections per mount to properly failover, so mount_get_dfs_conns() must be used from
+	 * now on.
+	 */
+	mount_put_conns(mnt_ctx);
+	mount_get_dfs_conns(mnt_ctx);
+
+	full_path = build_unc_path_to_root(ctx, cifs_sb, true);
+	if (IS_ERR(full_path))
+		return PTR_ERR(full_path);
+
+	mnt_ctx->origin_fullpath = dfs_cache_canonical_path(ctx->UNC, cifs_sb->local_nls,
+							    cifs_remap(cifs_sb));
+	if (IS_ERR(mnt_ctx->origin_fullpath)) {
+		rc = PTR_ERR(mnt_ctx->origin_fullpath);
+		mnt_ctx->origin_fullpath = NULL;
+		goto out;
 	}
 
-	uuid_gen(&mount_id);
-	set_root_ses(cifs_sb, &mount_id, ses, &root_ses);
-	do {
-		/* Save full path of last DFS path we used to resolve final target server */
-		kfree(full_path);
-		full_path = build_unc_path_to_root(ctx, cifs_sb, !!count);
-		if (IS_ERR(full_path)) {
-			rc = PTR_ERR(full_path);
-			full_path = NULL;
+	/* Try all dfs root targets */
+	for (rc = -ENOENT, tit = dfs_cache_get_tgt_iterator(root_tl);
+	     tit; tit = dfs_cache_get_next_tgt(root_tl, tit)) {
+		rc = connect_dfs_target(mnt_ctx, full_path, mnt_ctx->origin_fullpath + 1, tit);
+		if (!rc) {
+			mnt_ctx->leaf_fullpath = kstrdup(mnt_ctx->origin_fullpath, GFP_KERNEL);
+			if (!mnt_ctx->leaf_fullpath)
+				rc = -ENOMEM;
 			break;
 		}
-		/* Chase referral */
-		oldmnt = cifs_sb->ctx->mount_options;
-		rc = expand_dfs_referral(xid, root_ses, ctx, cifs_sb, ref_path + 1);
-		if (rc)
-			break;
-		/* Connect to new DFS target only if we were redirected */
-		if (oldmnt != cifs_sb->ctx->mount_options) {
-			mount_put_conns(cifs_sb, xid, server, ses, tcon);
-			rc = mount_get_dfs_conns(ctx, cifs_sb, &xid, &server, &ses, &tcon);
-		}
-		if (rc && !server && !ses) {
-			/* Failed to connect. Try to connect to other targets in the referral. */
-			rc = do_dfs_failover(ref_path + 1, full_path, cifs_sb, ctx, root_ses, &xid,
-					     &server, &ses, &tcon);
-		}
-		if (rc == -EACCES || rc == -EOPNOTSUPP || !server || !ses)
-			break;
-		if (!tcon)
-			continue;
+	}
 
-		/* Make sure that requests go through new root servers */
-		rc = is_referral_server(ref_path + 1, cifs_sb, tcon, &ref_server);
-		if (rc)
-			break;
-		if (ref_server)
-			set_root_ses(cifs_sb, &mount_id, ses, &root_ses);
-
-		/* Get next dfs path and then continue chasing them if -EREMOTE */
-		rc = next_dfs_prepath(cifs_sb, ctx, xid, server, tcon, &ref_path);
-		/* Prevent recursion on broken link referrals */
-		if (rc == -EREMOTE && ++count > MAX_NESTED_LINKS)
-			rc = -ELOOP;
-	} while (rc == -EREMOTE);
-
-	if (rc || !tcon || !ses)
-		goto error;
-
-	kfree(ref_path);
-	/*
-	 * Store DFS full path in both superblock and tree connect structures.
-	 *
-	 * For DFS root mounts, the prefix path (cifs_sb->prepath) is preserved during reconnect so
-	 * only the root path is set in cifs_sb->origin_fullpath and tcon->dfs_path. And for DFS
-	 * links, the prefix path is included in both and may be changed during reconnect.  See
-	 * cifs_tree_connect().
-	 */
-	ref_path = dfs_cache_canonical_path(full_path, cifs_sb->local_nls, cifs_remap(cifs_sb));
+out:
 	kfree(full_path);
-	full_path = NULL;
+	return rc;
+}
 
-	if (IS_ERR(ref_path)) {
-		rc = PTR_ERR(ref_path);
-		ref_path = NULL;
-		goto error;
+static int __follow_dfs_link(struct mount_ctx *mnt_ctx)
+{
+	int rc;
+	struct cifs_sb_info *cifs_sb = mnt_ctx->cifs_sb;
+	struct smb3_fs_context *ctx = mnt_ctx->fs_ctx;
+	char *full_path;
+	struct dfs_cache_tgt_list tl = DFS_CACHE_TGT_LIST_INIT(tl);
+	struct dfs_cache_tgt_iterator *tit;
+
+	full_path = build_unc_path_to_root(ctx, cifs_sb, true);
+	if (IS_ERR(full_path))
+		return PTR_ERR(full_path);
+
+	kfree(mnt_ctx->leaf_fullpath);
+	mnt_ctx->leaf_fullpath = dfs_cache_canonical_path(full_path, cifs_sb->local_nls,
+							  cifs_remap(cifs_sb));
+	if (IS_ERR(mnt_ctx->leaf_fullpath)) {
+		rc = PTR_ERR(mnt_ctx->leaf_fullpath);
+		mnt_ctx->leaf_fullpath = NULL;
+		goto out;
 	}
-	cifs_sb->origin_fullpath = ref_path;
 
-	ref_path = kstrdup(cifs_sb->origin_fullpath, GFP_KERNEL);
-	if (!ref_path) {
-		rc = -ENOMEM;
-		goto error;
+	/* Get referral from dfs link */
+	rc = dfs_cache_find(mnt_ctx->xid, mnt_ctx->root_ses, cifs_sb->local_nls,
+			    cifs_remap(cifs_sb), mnt_ctx->leaf_fullpath + 1, NULL, &tl);
+	if (rc)
+		goto out;
+
+	/* Try all dfs link targets */
+	for (rc = -ENOENT, tit = dfs_cache_get_tgt_iterator(&tl);
+	     tit; tit = dfs_cache_get_next_tgt(&tl, tit)) {
+		rc = connect_dfs_target(mnt_ctx, full_path, mnt_ctx->leaf_fullpath + 1, tit);
+		if (!rc) {
+			rc = is_path_remote(mnt_ctx);
+			break;
+		}
 	}
-	spin_lock(&cifs_tcp_ses_lock);
-	tcon->dfs_path = ref_path;
-	ref_path = NULL;
-	spin_unlock(&cifs_tcp_ses_lock);
 
+out:
+	kfree(full_path);
+	dfs_cache_free_tgts(&tl);
+	return rc;
+}
+
+static int follow_dfs_link(struct mount_ctx *mnt_ctx)
+{
+	int rc;
+	struct cifs_sb_info *cifs_sb = mnt_ctx->cifs_sb;
+	struct smb3_fs_context *ctx = mnt_ctx->fs_ctx;
+	char *full_path;
+	int num_links = 0;
+
+	full_path = build_unc_path_to_root(ctx, cifs_sb, true);
+	if (IS_ERR(full_path))
+		return PTR_ERR(full_path);
+
+	kfree(mnt_ctx->origin_fullpath);
+	mnt_ctx->origin_fullpath = dfs_cache_canonical_path(full_path, cifs_sb->local_nls,
+							    cifs_remap(cifs_sb));
+	kfree(full_path);
+
+	if (IS_ERR(mnt_ctx->origin_fullpath)) {
+		rc = PTR_ERR(mnt_ctx->origin_fullpath);
+		mnt_ctx->origin_fullpath = NULL;
+		return rc;
+	}
+
+	do {
+		rc = __follow_dfs_link(mnt_ctx);
+		if (!rc || rc != -EREMOTE)
+			break;
+	} while (rc = -ELOOP, ++num_links < MAX_NESTED_LINKS);
+
+	return rc;
+}
+
+/* Set up DFS referral paths for failover */
+static void setup_server_referral_paths(struct mount_ctx *mnt_ctx)
+{
+	struct TCP_Server_Info *server = mnt_ctx->server;
+
+	server->origin_fullpath = mnt_ctx->origin_fullpath;
+	server->leaf_fullpath = mnt_ctx->leaf_fullpath;
+	server->current_fullpath = mnt_ctx->leaf_fullpath;
+	mnt_ctx->origin_fullpath = mnt_ctx->leaf_fullpath = NULL;
+}
+
+int cifs_mount(struct cifs_sb_info *cifs_sb, struct smb3_fs_context *ctx)
+{
+	int rc;
+	struct mount_ctx mnt_ctx = { .cifs_sb = cifs_sb, .fs_ctx = ctx, };
+	struct dfs_cache_tgt_list tl = DFS_CACHE_TGT_LIST_INIT(tl);
+	bool isdfs;
+
+	rc = is_dfs_mount(&mnt_ctx, &isdfs, &tl);
+	if (rc)
+		goto error;
+	if (!isdfs)
+		goto out;
+
+	uuid_gen(&mnt_ctx.mount_id);
+	rc = connect_dfs_root(&mnt_ctx, &tl);
+	dfs_cache_free_tgts(&tl);
+
+	if (rc)
+		goto error;
+
+	rc = is_path_remote(&mnt_ctx);
+	if (rc == -EREMOTE)
+		rc = follow_dfs_link(&mnt_ctx);
+	if (rc)
+		goto error;
+
+	setup_server_referral_paths(&mnt_ctx);
 	/*
-	 * After reconnecting to a different server, unique ids won't
-	 * match anymore, so we disable serverino. This prevents
-	 * dentry revalidation to think the dentry are stale (ESTALE).
+	 * After reconnecting to a different server, unique ids won't match anymore, so we disable
+	 * serverino. This prevents dentry revalidation to think the dentry are stale (ESTALE).
 	 */
 	cifs_autodisable_serverino(cifs_sb);
 	/*
-	 * Force the use of prefix path to support failover on DFS paths that
-	 * resolve to targets that have different prefix paths.
+	 * Force the use of prefix path to support failover on DFS paths that resolve to targets
+	 * that have different prefix paths.
 	 */
 	cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_USE_PREFIX_PATH;
 	kfree(cifs_sb->prepath);
 	cifs_sb->prepath = ctx->prepath;
 	ctx->prepath = NULL;
-	uuid_copy(&cifs_sb->dfs_mount_id, &mount_id);
+	uuid_copy(&cifs_sb->dfs_mount_id, &mnt_ctx.mount_id);
 
 out:
-	free_xid(xid);
-	cifs_try_adding_channels(cifs_sb, ses);
-	return mount_setup_tlink(cifs_sb, ses, tcon);
+	free_xid(mnt_ctx.xid);
+	cifs_try_adding_channels(cifs_sb, mnt_ctx.ses);
+	return mount_setup_tlink(cifs_sb, mnt_ctx.ses, mnt_ctx.tcon);
 
 error:
-	kfree(ref_path);
-	kfree(full_path);
-	kfree(cifs_sb->origin_fullpath);
-	dfs_cache_put_refsrv_sessions(&mount_id);
-	mount_put_conns(cifs_sb, xid, server, ses, tcon);
+	dfs_cache_put_refsrv_sessions(&mnt_ctx.mount_id);
+	kfree(mnt_ctx.origin_fullpath);
+	kfree(mnt_ctx.leaf_fullpath);
+	mount_put_conns(&mnt_ctx);
 	return rc;
 }
 #else
 int cifs_mount(struct cifs_sb_info *cifs_sb, struct smb3_fs_context *ctx)
 {
 	int rc = 0;
-	unsigned int xid;
-	struct cifs_ses *ses;
-	struct cifs_tcon *tcon;
-	struct TCP_Server_Info *server;
+	struct mount_ctx mnt_ctx = { .cifs_sb = cifs_sb, .fs_ctx = ctx, };
 
-	rc = mount_get_conns(ctx, cifs_sb, &xid, &server, &ses, &tcon);
+	rc = mount_get_conns(&mnt_ctx);
 	if (rc)
 		goto error;
 
-	if (tcon) {
-		rc = is_path_remote(cifs_sb, ctx, xid, server, tcon);
+	if (mnt_ctx.tcon) {
+		rc = is_path_remote(&mnt_ctx);
 		if (rc == -EREMOTE)
 			rc = -EOPNOTSUPP;
 		if (rc)
 			goto error;
 	}
 
-	free_xid(xid);
-
-	return mount_setup_tlink(cifs_sb, ses, tcon);
+	free_xid(mnt_ctx.xid);
+	return mount_setup_tlink(cifs_sb, mnt_ctx.ses, mnt_ctx.tcon);
 
 error:
-	mount_put_conns(cifs_sb, xid, server, ses, tcon);
+	mount_put_conns(&mnt_ctx);
 	return rc;
 }
 #endif
@@ -3818,7 +3785,6 @@ cifs_umount(struct cifs_sb_info *cifs_sb)
 	kfree(cifs_sb->prepath);
 #ifdef CONFIG_CIFS_DFS_UPCALL
 	dfs_cache_put_refsrv_sessions(&cifs_sb->dfs_mount_id);
-	kfree(cifs_sb->origin_fullpath);
 #endif
 	call_rcu(&cifs_sb->rcu, delayed_free);
 }
@@ -4145,104 +4111,246 @@ cifs_prune_tlinks(struct work_struct *work)
 }
 
 #ifdef CONFIG_CIFS_DFS_UPCALL
+static void mark_tcon_tcp_ses_for_reconnect(struct cifs_tcon *tcon)
+{
+	int i;
+
+	for (i = 0; i < tcon->ses->chan_count; i++) {
+		spin_lock(&GlobalMid_Lock);
+		if (tcon->ses->chans[i].server->tcpStatus != CifsExiting)
+			tcon->ses->chans[i].server->tcpStatus = CifsNeedReconnect;
+		spin_unlock(&GlobalMid_Lock);
+	}
+}
+
+/* Update dfs referral path of superblock */
+static int update_server_fullpath(struct TCP_Server_Info *server, struct cifs_sb_info *cifs_sb,
+				  const char *target)
+{
+	int rc = 0;
+	size_t len = strlen(target);
+	char *refpath, *npath;
+
+	if (unlikely(len < 2 || *target != '\\'))
+		return -EINVAL;
+
+	if (target[1] == '\\') {
+		len += 1;
+		refpath = kmalloc(len, GFP_KERNEL);
+		if (!refpath)
+			return -ENOMEM;
+
+		scnprintf(refpath, len, "%s", target);
+	} else {
+		len += sizeof("\\");
+		refpath = kmalloc(len, GFP_KERNEL);
+		if (!refpath)
+			return -ENOMEM;
+
+		scnprintf(refpath, len, "\\%s", target);
+	}
+
+	npath = dfs_cache_canonical_path(refpath, cifs_sb->local_nls, cifs_remap(cifs_sb));
+	kfree(refpath);
+
+	if (IS_ERR(npath)) {
+		rc = PTR_ERR(npath);
+	} else {
+		mutex_lock(&server->refpath_lock);
+		kfree(server->leaf_fullpath);
+		server->leaf_fullpath = npath;
+		mutex_unlock(&server->refpath_lock);
+		server->current_fullpath = server->leaf_fullpath;
+	}
+	return rc;
+}
+
+static int target_share_matches_server(struct TCP_Server_Info *server, const char *tcp_host,
+				       size_t tcp_host_len, char *share, bool *target_match)
+{
+	int rc = 0;
+	const char *dfs_host;
+	size_t dfs_host_len;
+
+	*target_match = true;
+	extract_unc_hostname(share, &dfs_host, &dfs_host_len);
+
+	/* Check if hostnames or addresses match */
+	if (dfs_host_len != tcp_host_len || strncasecmp(dfs_host, tcp_host, dfs_host_len) != 0) {
+		cifs_dbg(FYI, "%s: %.*s doesn't match %.*s\n", __func__, (int)dfs_host_len,
+			 dfs_host, (int)tcp_host_len, tcp_host);
+		rc = match_target_ip(server, dfs_host, dfs_host_len, target_match);
+		if (rc)
+			cifs_dbg(VFS, "%s: failed to match target ip: %d\n", __func__, rc);
+	}
+	return rc;
+}
+
+static int __tree_connect_dfs_target(const unsigned int xid, struct cifs_tcon *tcon,
+				     struct cifs_sb_info *cifs_sb, char *tree, bool islink,
+				     struct dfs_cache_tgt_list *tl)
+{
+	int rc;
+	struct TCP_Server_Info *server = tcon->ses->server;
+	const struct smb_version_operations *ops = server->ops;
+	struct cifs_tcon *ipc = tcon->ses->tcon_ipc;
+	char *share = NULL, *prefix = NULL;
+	const char *tcp_host;
+	size_t tcp_host_len;
+	struct dfs_cache_tgt_iterator *tit;
+	bool target_match;
+
+	extract_unc_hostname(server->hostname, &tcp_host, &tcp_host_len);
+
+	tit = dfs_cache_get_tgt_iterator(tl);
+	if (!tit) {
+		rc = -ENOENT;
+		goto out;
+	}
+
+	/* Try to tree connect to all dfs targets */
+	for (; tit; tit = dfs_cache_get_next_tgt(tl, tit)) {
+		const char *target = dfs_cache_get_tgt_name(tit);
+		struct dfs_cache_tgt_list ntl = DFS_CACHE_TGT_LIST_INIT(ntl);
+
+		kfree(share);
+		kfree(prefix);
+		share = prefix = NULL;
+
+		/* Check if share matches with tcp ses */
+		rc = dfs_cache_get_tgt_share(server->current_fullpath + 1, tit, &share, &prefix);
+		if (rc) {
+			cifs_dbg(VFS, "%s: failed to parse target share: %d\n", __func__, rc);
+			break;
+		}
+
+		rc = target_share_matches_server(server, tcp_host, tcp_host_len, share,
+						 &target_match);
+		if (rc)
+			break;
+		if (!target_match) {
+			rc = -EHOSTUNREACH;
+			continue;
+		}
+
+		if (ipc->need_reconnect) {
+			scnprintf(tree, MAX_TREE_SIZE, "\\\\%s\\IPC$", server->hostname);
+			rc = ops->tree_connect(xid, ipc->ses, tree, ipc, cifs_sb->local_nls);
+			if (rc)
+				break;
+		}
+
+		scnprintf(tree, MAX_TREE_SIZE, "\\%s", share);
+		if (!islink) {
+			rc = ops->tree_connect(xid, tcon->ses, tree, tcon, cifs_sb->local_nls);
+			break;
+		}
+		/*
+		 * If no dfs referrals were returned from link target, then just do a TREE_CONNECT
+		 * to it.  Otherwise, cache the dfs referral and then mark current tcp ses for
+		 * reconnect so either the demultiplex thread or the echo worker will reconnect to
+		 * newly resolved target.
+		 */
+		if (dfs_cache_find(xid, tcon->ses, cifs_sb->local_nls, cifs_remap(cifs_sb), target,
+				   NULL, &ntl)) {
+			rc = ops->tree_connect(xid, tcon->ses, tree, tcon, cifs_sb->local_nls);
+			if (rc)
+				continue;
+			rc = dfs_cache_noreq_update_tgthint(server->current_fullpath + 1, tit);
+			if (!rc)
+				rc = cifs_update_super_prepath(cifs_sb, prefix);
+		} else {
+			/* Target is another dfs share */
+			rc = update_server_fullpath(server, cifs_sb, target);
+			dfs_cache_free_tgts(tl);
+
+			if (!rc) {
+				rc = -EREMOTE;
+				list_replace_init(&ntl.tl_list, &tl->tl_list);
+			} else
+				dfs_cache_free_tgts(&ntl);
+		}
+		break;
+	}
+
+out:
+	kfree(share);
+	kfree(prefix);
+
+	return rc;
+}
+
+static int tree_connect_dfs_target(const unsigned int xid, struct cifs_tcon *tcon,
+				   struct cifs_sb_info *cifs_sb, char *tree, bool islink,
+				   struct dfs_cache_tgt_list *tl)
+{
+	int rc;
+	int num_links = 0;
+	struct TCP_Server_Info *server = tcon->ses->server;
+
+	do {
+		rc = __tree_connect_dfs_target(xid, tcon, cifs_sb, tree, islink, tl);
+		if (!rc || rc != -EREMOTE)
+			break;
+	} while (rc = -ELOOP, ++num_links < MAX_NESTED_LINKS);
+	/*
+	 * If we couldn't tree connect to any targets from last referral path, then retry from
+	 * original referral path.
+	 */
+	if (rc && server->current_fullpath != server->origin_fullpath) {
+		server->current_fullpath = server->origin_fullpath;
+		mark_tcon_tcp_ses_for_reconnect(tcon);
+	}
+
+	dfs_cache_free_tgts(tl);
+	return rc;
+}
+
 int cifs_tree_connect(const unsigned int xid, struct cifs_tcon *tcon, const struct nls_table *nlsc)
 {
 	int rc;
 	struct TCP_Server_Info *server = tcon->ses->server;
 	const struct smb_version_operations *ops = server->ops;
-	struct dfs_cache_tgt_list tl;
-	struct dfs_cache_tgt_iterator *it = NULL;
+	struct super_block *sb = NULL;
+	struct cifs_sb_info *cifs_sb;
+	struct dfs_cache_tgt_list tl = DFS_CACHE_TGT_LIST_INIT(tl);
 	char *tree;
-	const char *tcp_host;
-	size_t tcp_host_len;
-	const char *dfs_host;
-	size_t dfs_host_len;
-	char *share = NULL, *prefix = NULL;
 	struct dfs_info3_param ref = {0};
-	bool isroot;
 
 	tree = kzalloc(MAX_TREE_SIZE, GFP_KERNEL);
 	if (!tree)
 		return -ENOMEM;
 
-	/* If it is not dfs or there was no cached dfs referral, then reconnect to same share */
-	if (!tcon->dfs_path || dfs_cache_noreq_find(tcon->dfs_path + 1, &ref, &tl)) {
-		if (tcon->ipc) {
-			scnprintf(tree, MAX_TREE_SIZE, "\\\\%s\\IPC$", server->hostname);
-			rc = ops->tree_connect(xid, tcon->ses, tree, tcon, nlsc);
-		} else {
-			rc = ops->tree_connect(xid, tcon->ses, tcon->treeName, tcon, nlsc);
-		}
+	if (tcon->ipc) {
+		scnprintf(tree, MAX_TREE_SIZE, "\\\\%s\\IPC$", server->hostname);
+		rc = ops->tree_connect(xid, tcon->ses, tree, tcon, nlsc);
 		goto out;
 	}
 
-	isroot = ref.server_type == DFS_TYPE_ROOT;
+	sb = cifs_get_tcp_super(server);
+	if (IS_ERR(sb)) {
+		rc = PTR_ERR(sb);
+		cifs_dbg(VFS, "%s: could not find superblock: %d\n", __func__, rc);
+		goto out;
+	}
+
+	cifs_sb = CIFS_SB(sb);
+
+	/* If it is not dfs or there was no cached dfs referral, then reconnect to same share */
+	if (!server->current_fullpath ||
+	    dfs_cache_noreq_find(server->current_fullpath + 1, &ref, &tl)) {
+		rc = ops->tree_connect(xid, tcon->ses, tcon->treeName, tcon, cifs_sb->local_nls);
+		goto out;
+	}
+
+	rc = tree_connect_dfs_target(xid, tcon, cifs_sb, tree, ref.server_type == DFS_TYPE_LINK,
+				     &tl);
 	free_dfs_info_param(&ref);
 
-	extract_unc_hostname(server->hostname, &tcp_host, &tcp_host_len);
-
-	for (it = dfs_cache_get_tgt_iterator(&tl); it; it = dfs_cache_get_next_tgt(&tl, it)) {
-		bool target_match;
-
-		kfree(share);
-		kfree(prefix);
-		share = NULL;
-		prefix = NULL;
-
-		rc = dfs_cache_get_tgt_share(tcon->dfs_path + 1, it, &share, &prefix);
-		if (rc) {
-			cifs_dbg(VFS, "%s: failed to parse target share %d\n",
-				 __func__, rc);
-			continue;
-		}
-
-		extract_unc_hostname(share, &dfs_host, &dfs_host_len);
-
-		if (dfs_host_len != tcp_host_len
-		    || strncasecmp(dfs_host, tcp_host, dfs_host_len) != 0) {
-			cifs_dbg(FYI, "%s: %.*s doesn't match %.*s\n", __func__, (int)dfs_host_len,
-				 dfs_host, (int)tcp_host_len, tcp_host);
-
-			rc = match_target_ip(server, dfs_host, dfs_host_len, &target_match);
-			if (rc) {
-				cifs_dbg(VFS, "%s: failed to match target ip: %d\n", __func__, rc);
-				break;
-			}
-
-			if (!target_match) {
-				cifs_dbg(FYI, "%s: skipping target\n", __func__);
-				continue;
-			}
-		}
-
-		if (tcon->ipc) {
-			scnprintf(tree, MAX_TREE_SIZE, "\\\\%s\\IPC$", share);
-			rc = ops->tree_connect(xid, tcon->ses, tree, tcon, nlsc);
-		} else {
-			scnprintf(tree, MAX_TREE_SIZE, "\\%s", share);
-			rc = ops->tree_connect(xid, tcon->ses, tree, tcon, nlsc);
-			/* Only handle prefix paths of DFS link targets */
-			if (!rc && !isroot) {
-				rc = update_super_prepath(tcon, prefix);
-				break;
-			}
-		}
-		if (rc == -EREMOTE)
-			break;
-	}
-
-	kfree(share);
-	kfree(prefix);
-
-	if (!rc) {
-		if (it)
-			rc = dfs_cache_noreq_update_tgthint(tcon->dfs_path + 1, it);
-		else
-			rc = -ENOENT;
-	}
-	dfs_cache_free_tgts(&tl);
 out:
 	kfree(tree);
+	cifs_put_tcp_super(sb);
+
 	return rc;
 }
 #else
diff --git a/fs/cifs/dfs_cache.c b/fs/cifs/dfs_cache.c
index 2837455..5c1259d 100644
--- a/fs/cifs/dfs_cache.c
+++ b/fs/cifs/dfs_cache.c
@@ -283,7 +283,7 @@ static int dfscache_proc_show(struct seq_file *m, void *v)
 			seq_printf(m,
 				   "cache entry: path=%s,type=%s,ttl=%d,etime=%ld,hdr_flags=0x%x,ref_flags=0x%x,interlink=%s,path_consumed=%d,expired=%s\n",
 				   ce->path, ce->srvtype == DFS_TYPE_ROOT ? "root" : "link",
-				   ce->ttl, ce->etime.tv_nsec, ce->ref_flags, ce->hdr_flags,
+				   ce->ttl, ce->etime.tv_nsec, ce->hdr_flags, ce->ref_flags,
 				   IS_DFS_INTERLINK(ce->hdr_flags) ? "yes" : "no",
 				   ce->path_consumed, cache_entry_expired(ce) ? "yes" : "no");
 
@@ -1364,9 +1364,9 @@ static void mark_for_reconnect_if_needed(struct cifs_tcon *tcon, struct dfs_cach
 }
 
 /* Refresh dfs referral of tcon and mark it for reconnect if needed */
-static int refresh_tcon(struct cifs_ses **sessions, struct cifs_tcon *tcon, bool force_refresh)
+static int __refresh_tcon(const char *path, struct cifs_ses **sessions, struct cifs_tcon *tcon,
+			  bool force_refresh)
 {
-	const char *path = tcon->dfs_path + 1;
 	struct cifs_ses *ses;
 	struct cache_entry *ce;
 	struct dfs_info3_param *refs = NULL;
@@ -1422,6 +1422,20 @@ static int refresh_tcon(struct cifs_ses **sessions, struct cifs_tcon *tcon, bool
 	return rc;
 }
 
+static int refresh_tcon(struct cifs_ses **sessions, struct cifs_tcon *tcon, bool force_refresh)
+{
+	struct TCP_Server_Info *server = tcon->ses->server;
+
+	mutex_lock(&server->refpath_lock);
+	if (strcasecmp(server->leaf_fullpath, server->origin_fullpath))
+		__refresh_tcon(server->leaf_fullpath + 1, sessions, tcon, force_refresh);
+	mutex_unlock(&server->refpath_lock);
+
+	__refresh_tcon(server->origin_fullpath + 1, sessions, tcon, force_refresh);
+
+	return 0;
+}
+
 /**
  * dfs_cache_remount_fs - remount a DFS share
  *
@@ -1435,6 +1449,7 @@ static int refresh_tcon(struct cifs_ses **sessions, struct cifs_tcon *tcon, bool
 int dfs_cache_remount_fs(struct cifs_sb_info *cifs_sb)
 {
 	struct cifs_tcon *tcon;
+	struct TCP_Server_Info *server;
 	struct mount_group *mg;
 	struct cifs_ses *sessions[CACHE_MAX_ENTRIES + 1] = {NULL};
 	int rc;
@@ -1443,13 +1458,15 @@ int dfs_cache_remount_fs(struct cifs_sb_info *cifs_sb)
 		return -EINVAL;
 
 	tcon = cifs_sb_master_tcon(cifs_sb);
-	if (!tcon->dfs_path) {
-		cifs_dbg(FYI, "%s: not a dfs tcon\n", __func__);
+	server = tcon->ses->server;
+
+	if (!server->origin_fullpath) {
+		cifs_dbg(FYI, "%s: not a dfs mount\n", __func__);
 		return 0;
 	}
 
 	if (uuid_is_null(&cifs_sb->dfs_mount_id)) {
-		cifs_dbg(FYI, "%s: tcon has no dfs mount group id\n", __func__);
+		cifs_dbg(FYI, "%s: no dfs mount group id\n", __func__);
 		return -EINVAL;
 	}
 
@@ -1457,7 +1474,7 @@ int dfs_cache_remount_fs(struct cifs_sb_info *cifs_sb)
 	mg = find_mount_group_locked(&cifs_sb->dfs_mount_id);
 	if (IS_ERR(mg)) {
 		mutex_unlock(&mount_group_list_lock);
-		cifs_dbg(FYI, "%s: tcon has ipc session to refresh referral\n", __func__);
+		cifs_dbg(FYI, "%s: no ipc session for refreshing referral\n", __func__);
 		return PTR_ERR(mg);
 	}
 	kref_get(&mg->refcount);
@@ -1498,9 +1515,12 @@ static void refresh_mounts(struct cifs_ses **sessions)
 
 	spin_lock(&cifs_tcp_ses_lock);
 	list_for_each_entry(server, &cifs_tcp_ses_list, tcp_ses_list) {
+		if (!server->is_dfs_conn)
+			continue;
+
 		list_for_each_entry(ses, &server->smb_ses_list, smb_ses_list) {
 			list_for_each_entry(tcon, &ses->tcon_list, tcon_list) {
-				if (tcon->dfs_path) {
+				if (!tcon->ipc && !tcon->need_reconnect) {
 					tcon->tc_count++;
 					list_add_tail(&tcon->ulist, &tcons);
 				}
@@ -1510,8 +1530,16 @@ static void refresh_mounts(struct cifs_ses **sessions)
 	spin_unlock(&cifs_tcp_ses_lock);
 
 	list_for_each_entry_safe(tcon, ntcon, &tcons, ulist) {
+		struct TCP_Server_Info *server = tcon->ses->server;
+
 		list_del_init(&tcon->ulist);
-		refresh_tcon(sessions, tcon, false);
+
+		mutex_lock(&server->refpath_lock);
+		if (strcasecmp(server->leaf_fullpath, server->origin_fullpath))
+			__refresh_tcon(server->leaf_fullpath + 1, sessions, tcon, false);
+		mutex_unlock(&server->refpath_lock);
+
+		__refresh_tcon(server->origin_fullpath + 1, sessions, tcon, false);
 		cifs_put_tcon(tcon);
 	}
 }
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 1b855fc..9fee3af 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -2692,12 +2692,23 @@ int cifs_strict_fsync(struct file *file, loff_t start, loff_t end,
 	tcon = tlink_tcon(smbfile->tlink);
 	if (!(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOSSYNC)) {
 		server = tcon->ses->server;
-		if (server->ops->flush)
-			rc = server->ops->flush(xid, tcon, &smbfile->fid);
-		else
+		if (server->ops->flush == NULL) {
 			rc = -ENOSYS;
+			goto strict_fsync_exit;
+		}
+
+		if ((OPEN_FMODE(smbfile->f_flags) & FMODE_WRITE) == 0) {
+			smbfile = find_writable_file(CIFS_I(inode), FIND_WR_ANY);
+			if (smbfile) {
+				rc = server->ops->flush(xid, tcon, &smbfile->fid);
+				cifsFileInfo_put(smbfile);
+			} else
+				cifs_dbg(FYI, "ignore fsync for file not open for write\n");
+		} else
+			rc = server->ops->flush(xid, tcon, &smbfile->fid);
 	}
 
+strict_fsync_exit:
 	free_xid(xid);
 	return rc;
 }
@@ -2709,6 +2720,7 @@ int cifs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 	struct cifs_tcon *tcon;
 	struct TCP_Server_Info *server;
 	struct cifsFileInfo *smbfile = file->private_data;
+	struct inode *inode = file_inode(file);
 	struct cifs_sb_info *cifs_sb = CIFS_FILE_SB(file);
 
 	rc = file_write_and_wait_range(file, start, end);
@@ -2725,12 +2737,23 @@ int cifs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 	tcon = tlink_tcon(smbfile->tlink);
 	if (!(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOSSYNC)) {
 		server = tcon->ses->server;
-		if (server->ops->flush)
-			rc = server->ops->flush(xid, tcon, &smbfile->fid);
-		else
+		if (server->ops->flush == NULL) {
 			rc = -ENOSYS;
+			goto fsync_exit;
+		}
+
+		if ((OPEN_FMODE(smbfile->f_flags) & FMODE_WRITE) == 0) {
+			smbfile = find_writable_file(CIFS_I(inode), FIND_WR_ANY);
+			if (smbfile) {
+				rc = server->ops->flush(xid, tcon, &smbfile->fid);
+				cifsFileInfo_put(smbfile);
+			} else
+				cifs_dbg(FYI, "ignore fsync for file not open for write\n");
+		} else
+			rc = server->ops->flush(xid, tcon, &smbfile->fid);
 	}
 
+fsync_exit:
 	free_xid(xid);
 	return rc;
 }
diff --git a/fs/cifs/fs_context.c b/fs/cifs/fs_context.c
index 38d96a4..6a179ae 100644
--- a/fs/cifs/fs_context.c
+++ b/fs/cifs/fs_context.c
@@ -308,7 +308,9 @@ smb3_fs_context_dup(struct smb3_fs_context *new_ctx, struct smb3_fs_context *ctx
 	new_ctx->nodename = NULL;
 	new_ctx->username = NULL;
 	new_ctx->password = NULL;
+	new_ctx->server_hostname = NULL;
 	new_ctx->domainname = NULL;
+	new_ctx->workstation_name = NULL;
 	new_ctx->UNC = NULL;
 	new_ctx->source = NULL;
 	new_ctx->iocharset = NULL;
@@ -323,6 +325,7 @@ smb3_fs_context_dup(struct smb3_fs_context *new_ctx, struct smb3_fs_context *ctx
 	DUP_CTX_STR(UNC);
 	DUP_CTX_STR(source);
 	DUP_CTX_STR(domainname);
+	DUP_CTX_STR(workstation_name);
 	DUP_CTX_STR(nodename);
 	DUP_CTX_STR(iocharset);
 
@@ -459,6 +462,7 @@ smb3_parse_devname(const char *devname, struct smb3_fs_context *ctx)
 		return -EINVAL;
 
 	/* record the server hostname */
+	kfree(ctx->server_hostname);
 	ctx->server_hostname = kstrndup(devname + 2, pos - devname - 2, GFP_KERNEL);
 	if (!ctx->server_hostname)
 		return -ENOMEM;
@@ -720,6 +724,11 @@ static int smb3_verify_reconfigure_ctx(struct fs_context *fc,
 		cifs_errorf(fc, "can not change domainname during remount\n");
 		return -EINVAL;
 	}
+	if (new_ctx->workstation_name &&
+	    (!old_ctx->workstation_name || strcmp(new_ctx->workstation_name, old_ctx->workstation_name))) {
+		cifs_errorf(fc, "can not change workstation_name during remount\n");
+		return -EINVAL;
+	}
 	if (new_ctx->nodename &&
 	    (!old_ctx->nodename || strcmp(new_ctx->nodename, old_ctx->nodename))) {
 		cifs_errorf(fc, "can not change nodename during remount\n");
@@ -753,7 +762,8 @@ static int smb3_reconfigure(struct fs_context *fc)
 		return rc;
 
 	/*
-	 * We can not change UNC/username/password/domainname/nodename/iocharset
+	 * We can not change UNC/username/password/domainname/
+	 * workstation_name/nodename/iocharset
 	 * during reconnect so ignore what we have in the new context and
 	 * just use what we already have in cifs_sb->ctx.
 	 */
@@ -762,6 +772,7 @@ static int smb3_reconfigure(struct fs_context *fc)
 	STEAL_STRING(cifs_sb, ctx, username);
 	STEAL_STRING(cifs_sb, ctx, password);
 	STEAL_STRING(cifs_sb, ctx, domainname);
+	STEAL_STRING(cifs_sb, ctx, workstation_name);
 	STEAL_STRING(cifs_sb, ctx, nodename);
 	STEAL_STRING(cifs_sb, ctx, iocharset);
 
@@ -1414,13 +1425,22 @@ static int smb3_fs_context_parse_param(struct fs_context *fc,
 
 int smb3_init_fs_context(struct fs_context *fc)
 {
+	int rc;
 	struct smb3_fs_context *ctx;
 	char *nodename = utsname()->nodename;
 	int i;
 
 	ctx = kzalloc(sizeof(struct smb3_fs_context), GFP_KERNEL);
-	if (unlikely(!ctx))
-		return -ENOMEM;
+	if (unlikely(!ctx)) {
+		rc = -ENOMEM;
+		goto err_exit;
+	}
+
+	ctx->workstation_name = kstrdup(nodename, GFP_KERNEL);
+	if (unlikely(!ctx->workstation_name)) {
+		rc = -ENOMEM;
+		goto err_exit;
+	}
 
 	/*
 	 * does not have to be perfect mapping since field is
@@ -1493,6 +1513,14 @@ int smb3_init_fs_context(struct fs_context *fc)
 	fc->fs_private = ctx;
 	fc->ops = &smb3_fs_context_ops;
 	return 0;
+
+err_exit:
+	if (ctx) {
+		kfree(ctx->workstation_name);
+		kfree(ctx);
+	}
+
+	return rc;
 }
 
 void
@@ -1518,6 +1546,8 @@ smb3_cleanup_fs_context_contents(struct smb3_fs_context *ctx)
 	ctx->source = NULL;
 	kfree(ctx->domainname);
 	ctx->domainname = NULL;
+	kfree(ctx->workstation_name);
+	ctx->workstation_name = NULL;
 	kfree(ctx->nodename);
 	ctx->nodename = NULL;
 	kfree(ctx->iocharset);
diff --git a/fs/cifs/fs_context.h b/fs/cifs/fs_context.h
index b2d22cf..e54090d 100644
--- a/fs/cifs/fs_context.h
+++ b/fs/cifs/fs_context.h
@@ -170,6 +170,7 @@ struct smb3_fs_context {
 	char *server_hostname;
 	char *UNC;
 	char *nodename;
+	char *workstation_name;
 	char *iocharset;  /* local code page for mapping to and from Unicode */
 	char source_rfc1001_name[RFC1001_NAME_LEN_WITH_NULL]; /* clnt nb name */
 	char target_rfc1001_name[RFC1001_NAME_LEN_WITH_NULL]; /* srvr nb name */
diff --git a/fs/cifs/fscache.c b/fs/cifs/fscache.c
index 8eedd20..7e409a3 100644
--- a/fs/cifs/fscache.c
+++ b/fs/cifs/fscache.c
@@ -87,6 +87,14 @@ void cifs_fscache_get_super_cookie(struct cifs_tcon *tcon)
 	char *sharename;
 	struct cifs_fscache_super_auxdata auxdata;
 
+	/*
+	 * Check if cookie was already initialized so don't reinitialize it.
+	 * In the future, as we integrate with newer fscache features,
+	 * we may want to instead add a check if cookie has changed
+	 */
+	if (tcon->fscache == NULL)
+		return;
+
 	sharename = extract_sharename(tcon->treeName);
 	if (IS_ERR(sharename)) {
 		cifs_dbg(FYI, "%s: couldn't extract sharename\n", __func__);
diff --git a/fs/cifs/misc.c b/fs/cifs/misc.c
index ba2c3e8..5148d48 100644
--- a/fs/cifs/misc.c
+++ b/fs/cifs/misc.c
@@ -75,6 +75,7 @@ sesInfoAlloc(void)
 		INIT_LIST_HEAD(&ret_buf->tcon_list);
 		mutex_init(&ret_buf->session_mutex);
 		spin_lock_init(&ret_buf->iface_lock);
+		spin_lock_init(&ret_buf->chan_lock);
 	}
 	return ret_buf;
 }
@@ -94,6 +95,7 @@ sesInfoFree(struct cifs_ses *buf_to_free)
 	kfree_sensitive(buf_to_free->password);
 	kfree(buf_to_free->user_name);
 	kfree(buf_to_free->domainName);
+	kfree(buf_to_free->workstation_name);
 	kfree_sensitive(buf_to_free->auth_key.response);
 	kfree(buf_to_free->iface_list);
 	kfree_sensitive(buf_to_free);
@@ -138,9 +140,6 @@ tconInfoFree(struct cifs_tcon *buf_to_free)
 	kfree(buf_to_free->nativeFileSystem);
 	kfree_sensitive(buf_to_free->password);
 	kfree(buf_to_free->crfid.fid);
-#ifdef CONFIG_CIFS_DFS_UPCALL
-	kfree(buf_to_free->dfs_path);
-#endif
 	kfree(buf_to_free);
 }
 
@@ -1287,69 +1286,20 @@ int match_target_ip(struct TCP_Server_Info *server,
 	return rc;
 }
 
-static void tcon_super_cb(struct super_block *sb, void *arg)
+int cifs_update_super_prepath(struct cifs_sb_info *cifs_sb, char *prefix)
 {
-	struct super_cb_data *sd = arg;
-	struct cifs_tcon *tcon = sd->data;
-	struct cifs_sb_info *cifs_sb;
-
-	if (sd->sb)
-		return;
-
-	cifs_sb = CIFS_SB(sb);
-	if (tcon->dfs_path && cifs_sb->origin_fullpath &&
-	    !strcasecmp(tcon->dfs_path, cifs_sb->origin_fullpath))
-		sd->sb = sb;
-}
-
-static inline struct super_block *cifs_get_tcon_super(struct cifs_tcon *tcon)
-{
-	return __cifs_get_super(tcon_super_cb, tcon);
-}
-
-static inline void cifs_put_tcon_super(struct super_block *sb)
-{
-	__cifs_put_super(sb);
-}
-#else
-static inline struct super_block *cifs_get_tcon_super(struct cifs_tcon *tcon)
-{
-	return ERR_PTR(-EOPNOTSUPP);
-}
-
-static inline void cifs_put_tcon_super(struct super_block *sb)
-{
-}
-#endif
-
-int update_super_prepath(struct cifs_tcon *tcon, char *prefix)
-{
-	struct super_block *sb;
-	struct cifs_sb_info *cifs_sb;
-	int rc = 0;
-
-	sb = cifs_get_tcon_super(tcon);
-	if (IS_ERR(sb))
-		return PTR_ERR(sb);
-
-	cifs_sb = CIFS_SB(sb);
-
 	kfree(cifs_sb->prepath);
 
 	if (prefix && *prefix) {
 		cifs_sb->prepath = kstrdup(prefix, GFP_ATOMIC);
-		if (!cifs_sb->prepath) {
-			rc = -ENOMEM;
-			goto out;
-		}
+		if (!cifs_sb->prepath)
+			return -ENOMEM;
 
 		convert_delimiter(cifs_sb->prepath, CIFS_DIR_SEP(cifs_sb));
 	} else
 		cifs_sb->prepath = NULL;
 
 	cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_USE_PREFIX_PATH;
-
-out:
-	cifs_put_tcon_super(sb);
-	return rc;
+	return 0;
 }
+#endif
diff --git a/fs/cifs/ntlmssp.h b/fs/cifs/ntlmssp.h
index 25a2b8e..fe707f4 100644
--- a/fs/cifs/ntlmssp.h
+++ b/fs/cifs/ntlmssp.h
@@ -119,7 +119,9 @@ typedef struct _AUTHENTICATE_MESSAGE {
  */
 
 int decode_ntlmssp_challenge(char *bcc_ptr, int blob_len, struct cifs_ses *ses);
-void build_ntlmssp_negotiate_blob(unsigned char *pbuffer, struct cifs_ses *ses);
+int build_ntlmssp_negotiate_blob(unsigned char **pbuffer, u16 *buflen,
+				 struct cifs_ses *ses,
+				 const struct nls_table *nls_cp);
 int build_ntlmssp_auth_blob(unsigned char **pbuffer, u16 *buflen,
 			struct cifs_ses *ses,
 			const struct nls_table *nls_cp);
diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c
index 23e02db..2c10b18 100644
--- a/fs/cifs/sess.c
+++ b/fs/cifs/sess.c
@@ -54,41 +54,53 @@ bool is_ses_using_iface(struct cifs_ses *ses, struct cifs_server_iface *iface)
 {
 	int i;
 
+	spin_lock(&ses->chan_lock);
 	for (i = 0; i < ses->chan_count; i++) {
-		if (is_server_using_iface(ses->chans[i].server, iface))
+		if (is_server_using_iface(ses->chans[i].server, iface)) {
+			spin_unlock(&ses->chan_lock);
 			return true;
+		}
 	}
+	spin_unlock(&ses->chan_lock);
 	return false;
 }
 
 /* returns number of channels added */
 int cifs_try_adding_channels(struct cifs_sb_info *cifs_sb, struct cifs_ses *ses)
 {
-	int old_chan_count = ses->chan_count;
-	int left = ses->chan_max - ses->chan_count;
+	int old_chan_count, new_chan_count;
+	int left;
 	int i = 0;
 	int rc = 0;
 	int tries = 0;
 	struct cifs_server_iface *ifaces = NULL;
 	size_t iface_count;
 
+	if (ses->server->dialect < SMB30_PROT_ID) {
+		cifs_dbg(VFS, "multichannel is not supported on this protocol version, use 3.0 or above\n");
+		return 0;
+	}
+
+	spin_lock(&ses->chan_lock);
+
+	new_chan_count = old_chan_count = ses->chan_count;
+	left = ses->chan_max - ses->chan_count;
+
 	if (left <= 0) {
 		cifs_dbg(FYI,
 			 "ses already at max_channels (%zu), nothing to open\n",
 			 ses->chan_max);
-		return 0;
-	}
-
-	if (ses->server->dialect < SMB30_PROT_ID) {
-		cifs_dbg(VFS, "multichannel is not supported on this protocol version, use 3.0 or above\n");
+		spin_unlock(&ses->chan_lock);
 		return 0;
 	}
 
 	if (!(ses->server->capabilities & SMB2_GLOBAL_CAP_MULTI_CHANNEL)) {
 		cifs_dbg(VFS, "server %s does not support multichannel\n", ses->server->hostname);
 		ses->chan_max = 1;
+		spin_unlock(&ses->chan_lock);
 		return 0;
 	}
+	spin_unlock(&ses->chan_lock);
 
 	/*
 	 * Make a copy of the iface list at the time and use that
@@ -142,10 +154,11 @@ int cifs_try_adding_channels(struct cifs_sb_info *cifs_sb, struct cifs_ses *ses)
 		cifs_dbg(FYI, "successfully opened new channel on iface#%d\n",
 			 i);
 		left--;
+		new_chan_count++;
 	}
 
 	kfree(ifaces);
-	return ses->chan_count - old_chan_count;
+	return new_chan_count - old_chan_count;
 }
 
 /*
@@ -157,10 +170,14 @@ cifs_ses_find_chan(struct cifs_ses *ses, struct TCP_Server_Info *server)
 {
 	int i;
 
+	spin_lock(&ses->chan_lock);
 	for (i = 0; i < ses->chan_count; i++) {
-		if (ses->chans[i].server == server)
+		if (ses->chans[i].server == server) {
+			spin_unlock(&ses->chan_lock);
 			return &ses->chans[i];
+		}
 	}
+	spin_unlock(&ses->chan_lock);
 	return NULL;
 }
 
@@ -168,6 +185,7 @@ static int
 cifs_ses_add_channel(struct cifs_sb_info *cifs_sb, struct cifs_ses *ses,
 		     struct cifs_server_iface *iface)
 {
+	struct TCP_Server_Info *chan_server;
 	struct cifs_chan *chan;
 	struct smb3_fs_context ctx = {NULL};
 	static const char unc_fmt[] = "\\%s\\foo";
@@ -240,18 +258,19 @@ cifs_ses_add_channel(struct cifs_sb_info *cifs_sb, struct cifs_ses *ses,
 	       SMB2_CLIENT_GUID_SIZE);
 	ctx.use_client_guid = true;
 
-	mutex_lock(&ses->session_mutex);
+	chan_server = cifs_get_tcp_session(&ctx, ses->server);
 
+	mutex_lock(&ses->session_mutex);
+	spin_lock(&ses->chan_lock);
 	chan = ses->binding_chan = &ses->chans[ses->chan_count];
-	chan->server = cifs_get_tcp_session(&ctx);
+	chan->server = chan_server;
 	if (IS_ERR(chan->server)) {
 		rc = PTR_ERR(chan->server);
 		chan->server = NULL;
+		spin_unlock(&ses->chan_lock);
 		goto out;
 	}
-	spin_lock(&cifs_tcp_ses_lock);
-	chan->server->is_channel = true;
-	spin_unlock(&cifs_tcp_ses_lock);
+	spin_unlock(&ses->chan_lock);
 
 	/*
 	 * We need to allocate the server crypto now as we will need
@@ -283,8 +302,11 @@ cifs_ses_add_channel(struct cifs_sb_info *cifs_sb, struct cifs_ses *ses,
 	 * ses to the new server.
 	 */
 
+	spin_lock(&ses->chan_lock);
 	ses->chan_count++;
 	atomic_set(&ses->chan_seq, 0);
+	spin_unlock(&ses->chan_lock);
+
 out:
 	ses->binding = false;
 	ses->binding_chan = NULL;
@@ -599,18 +621,85 @@ int decode_ntlmssp_challenge(char *bcc_ptr, int blob_len,
 	return 0;
 }
 
+static int size_of_ntlmssp_blob(struct cifs_ses *ses, int base_size)
+{
+	int sz = base_size + ses->auth_key.len
+		- CIFS_SESS_KEY_SIZE + CIFS_CPHTXT_SIZE + 2;
+
+	if (ses->domainName)
+		sz += sizeof(__le16) * strnlen(ses->domainName, CIFS_MAX_DOMAINNAME_LEN);
+	else
+		sz += sizeof(__le16);
+
+	if (ses->user_name)
+		sz += sizeof(__le16) * strnlen(ses->user_name, CIFS_MAX_USERNAME_LEN);
+	else
+		sz += sizeof(__le16);
+
+	sz += sizeof(__le16) * strnlen(ses->workstation_name, CIFS_MAX_WORKSTATION_LEN);
+
+	return sz;
+}
+
+static inline void cifs_security_buffer_from_str(SECURITY_BUFFER *pbuf,
+						 char *str_value,
+						 int str_length,
+						 unsigned char *pstart,
+						 unsigned char **pcur,
+						 const struct nls_table *nls_cp)
+{
+	unsigned char *tmp = pstart;
+	int len;
+
+	if (!pbuf)
+		return;
+
+	if (!pcur)
+		pcur = &tmp;
+
+	if (!str_value) {
+		pbuf->BufferOffset = cpu_to_le32(*pcur - pstart);
+		pbuf->Length = 0;
+		pbuf->MaximumLength = 0;
+		*pcur += sizeof(__le16);
+	} else {
+		len = cifs_strtoUTF16((__le16 *)*pcur,
+				      str_value,
+				      str_length,
+				      nls_cp);
+		len *= sizeof(__le16);
+		pbuf->BufferOffset = cpu_to_le32(*pcur - pstart);
+		pbuf->Length = cpu_to_le16(len);
+		pbuf->MaximumLength = cpu_to_le16(len);
+		*pcur += len;
+	}
+}
+
 /* BB Move to ntlmssp.c eventually */
 
-/* We do not malloc the blob, it is passed in pbuffer, because
-   it is fixed size, and small, making this approach cleaner */
-void build_ntlmssp_negotiate_blob(unsigned char *pbuffer,
-					 struct cifs_ses *ses)
+int build_ntlmssp_negotiate_blob(unsigned char **pbuffer,
+				 u16 *buflen,
+				 struct cifs_ses *ses,
+				 const struct nls_table *nls_cp)
 {
+	int rc = 0;
 	struct TCP_Server_Info *server = cifs_ses_server(ses);
-	NEGOTIATE_MESSAGE *sec_blob = (NEGOTIATE_MESSAGE *)pbuffer;
+	NEGOTIATE_MESSAGE *sec_blob;
 	__u32 flags;
+	unsigned char *tmp;
+	int len;
 
-	memset(pbuffer, 0, sizeof(NEGOTIATE_MESSAGE));
+	len = size_of_ntlmssp_blob(ses, sizeof(NEGOTIATE_MESSAGE));
+	*pbuffer = kmalloc(len, GFP_KERNEL);
+	if (!*pbuffer) {
+		rc = -ENOMEM;
+		cifs_dbg(VFS, "Error %d during NTLMSSP allocation\n", rc);
+		*buflen = 0;
+		goto setup_ntlm_neg_ret;
+	}
+	sec_blob = (NEGOTIATE_MESSAGE *)*pbuffer;
+
+	memset(*pbuffer, 0, sizeof(NEGOTIATE_MESSAGE));
 	memcpy(sec_blob->Signature, NTLMSSP_SIGNATURE, 8);
 	sec_blob->MessageType = NtLmNegotiate;
 
@@ -624,34 +713,25 @@ void build_ntlmssp_negotiate_blob(unsigned char *pbuffer,
 	if (!server->session_estab || ses->ntlmssp->sesskey_per_smbsess)
 		flags |= NTLMSSP_NEGOTIATE_KEY_XCH;
 
+	tmp = *pbuffer + sizeof(NEGOTIATE_MESSAGE);
 	sec_blob->NegotiateFlags = cpu_to_le32(flags);
 
-	sec_blob->WorkstationName.BufferOffset = 0;
-	sec_blob->WorkstationName.Length = 0;
-	sec_blob->WorkstationName.MaximumLength = 0;
+	/* these fields should be null in negotiate phase MS-NLMP 3.1.5.1.1 */
+	cifs_security_buffer_from_str(&sec_blob->DomainName,
+				      NULL,
+				      CIFS_MAX_DOMAINNAME_LEN,
+				      *pbuffer, &tmp,
+				      nls_cp);
 
-	/* Domain name is sent on the Challenge not Negotiate NTLMSSP request */
-	sec_blob->DomainName.BufferOffset = 0;
-	sec_blob->DomainName.Length = 0;
-	sec_blob->DomainName.MaximumLength = 0;
-}
+	cifs_security_buffer_from_str(&sec_blob->WorkstationName,
+				      NULL,
+				      CIFS_MAX_WORKSTATION_LEN,
+				      *pbuffer, &tmp,
+				      nls_cp);
 
-static int size_of_ntlmssp_blob(struct cifs_ses *ses)
-{
-	int sz = sizeof(AUTHENTICATE_MESSAGE) + ses->auth_key.len
-		- CIFS_SESS_KEY_SIZE + CIFS_CPHTXT_SIZE + 2;
-
-	if (ses->domainName)
-		sz += 2 * strnlen(ses->domainName, CIFS_MAX_DOMAINNAME_LEN);
-	else
-		sz += 2;
-
-	if (ses->user_name)
-		sz += 2 * strnlen(ses->user_name, CIFS_MAX_USERNAME_LEN);
-	else
-		sz += 2;
-
-	return sz;
+	*buflen = tmp - *pbuffer;
+setup_ntlm_neg_ret:
+	return rc;
 }
 
 int build_ntlmssp_auth_blob(unsigned char **pbuffer,
@@ -663,6 +743,7 @@ int build_ntlmssp_auth_blob(unsigned char **pbuffer,
 	AUTHENTICATE_MESSAGE *sec_blob;
 	__u32 flags;
 	unsigned char *tmp;
+	int len;
 
 	rc = setup_ntlmv2_rsp(ses, nls_cp);
 	if (rc) {
@@ -670,7 +751,9 @@ int build_ntlmssp_auth_blob(unsigned char **pbuffer,
 		*buflen = 0;
 		goto setup_ntlmv2_ret;
 	}
-	*pbuffer = kmalloc(size_of_ntlmssp_blob(ses), GFP_KERNEL);
+
+	len = size_of_ntlmssp_blob(ses, sizeof(AUTHENTICATE_MESSAGE));
+	*pbuffer = kmalloc(len, GFP_KERNEL);
 	if (!*pbuffer) {
 		rc = -ENOMEM;
 		cifs_dbg(VFS, "Error %d during NTLMSSP allocation\n", rc);
@@ -686,7 +769,7 @@ int build_ntlmssp_auth_blob(unsigned char **pbuffer,
 		NTLMSSP_REQUEST_TARGET | NTLMSSP_NEGOTIATE_TARGET_INFO |
 		NTLMSSP_NEGOTIATE_128 | NTLMSSP_NEGOTIATE_UNICODE |
 		NTLMSSP_NEGOTIATE_NTLM | NTLMSSP_NEGOTIATE_EXTENDED_SEC |
-		NTLMSSP_NEGOTIATE_SEAL;
+		NTLMSSP_NEGOTIATE_SEAL | NTLMSSP_NEGOTIATE_WORKSTATION_SUPPLIED;
 	if (ses->server->sign)
 		flags |= NTLMSSP_NEGOTIATE_SIGN;
 	if (!ses->server->session_estab || ses->ntlmssp->sesskey_per_smbsess)
@@ -719,42 +802,23 @@ int build_ntlmssp_auth_blob(unsigned char **pbuffer,
 		sec_blob->NtChallengeResponse.MaximumLength = 0;
 	}
 
-	if (ses->domainName == NULL) {
-		sec_blob->DomainName.BufferOffset = cpu_to_le32(tmp - *pbuffer);
-		sec_blob->DomainName.Length = 0;
-		sec_blob->DomainName.MaximumLength = 0;
-		tmp += 2;
-	} else {
-		int len;
-		len = cifs_strtoUTF16((__le16 *)tmp, ses->domainName,
-				      CIFS_MAX_DOMAINNAME_LEN, nls_cp);
-		len *= 2; /* unicode is 2 bytes each */
-		sec_blob->DomainName.BufferOffset = cpu_to_le32(tmp - *pbuffer);
-		sec_blob->DomainName.Length = cpu_to_le16(len);
-		sec_blob->DomainName.MaximumLength = cpu_to_le16(len);
-		tmp += len;
-	}
+	cifs_security_buffer_from_str(&sec_blob->DomainName,
+				      ses->domainName,
+				      CIFS_MAX_DOMAINNAME_LEN,
+				      *pbuffer, &tmp,
+				      nls_cp);
 
-	if (ses->user_name == NULL) {
-		sec_blob->UserName.BufferOffset = cpu_to_le32(tmp - *pbuffer);
-		sec_blob->UserName.Length = 0;
-		sec_blob->UserName.MaximumLength = 0;
-		tmp += 2;
-	} else {
-		int len;
-		len = cifs_strtoUTF16((__le16 *)tmp, ses->user_name,
-				      CIFS_MAX_USERNAME_LEN, nls_cp);
-		len *= 2; /* unicode is 2 bytes each */
-		sec_blob->UserName.BufferOffset = cpu_to_le32(tmp - *pbuffer);
-		sec_blob->UserName.Length = cpu_to_le16(len);
-		sec_blob->UserName.MaximumLength = cpu_to_le16(len);
-		tmp += len;
-	}
+	cifs_security_buffer_from_str(&sec_blob->UserName,
+				      ses->user_name,
+				      CIFS_MAX_USERNAME_LEN,
+				      *pbuffer, &tmp,
+				      nls_cp);
 
-	sec_blob->WorkstationName.BufferOffset = cpu_to_le32(tmp - *pbuffer);
-	sec_blob->WorkstationName.Length = 0;
-	sec_blob->WorkstationName.MaximumLength = 0;
-	tmp += 2;
+	cifs_security_buffer_from_str(&sec_blob->WorkstationName,
+				      ses->workstation_name,
+				      CIFS_MAX_WORKSTATION_LEN,
+				      *pbuffer, &tmp,
+				      nls_cp);
 
 	if (((ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_KEY_XCH) ||
 		(ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_EXTENDED_SEC))
@@ -1230,6 +1294,7 @@ sess_auth_rawntlmssp_negotiate(struct sess_data *sess_data)
 	struct cifs_ses *ses = sess_data->ses;
 	__u16 bytes_remaining;
 	char *bcc_ptr;
+	unsigned char *ntlmsspblob = NULL;
 	u16 blob_len;
 
 	cifs_dbg(FYI, "rawntlmssp session setup negotiate phase\n");
@@ -1253,10 +1318,15 @@ sess_auth_rawntlmssp_negotiate(struct sess_data *sess_data)
 	pSMB = (SESSION_SETUP_ANDX *)sess_data->iov[0].iov_base;
 
 	/* Build security blob before we assemble the request */
-	build_ntlmssp_negotiate_blob(pSMB->req.SecurityBlob, ses);
-	sess_data->iov[1].iov_len = sizeof(NEGOTIATE_MESSAGE);
-	sess_data->iov[1].iov_base = pSMB->req.SecurityBlob;
-	pSMB->req.SecurityBlobLength = cpu_to_le16(sizeof(NEGOTIATE_MESSAGE));
+	rc = build_ntlmssp_negotiate_blob(&ntlmsspblob,
+				     &blob_len, ses,
+				     sess_data->nls_cp);
+	if (rc)
+		goto out;
+
+	sess_data->iov[1].iov_len = blob_len;
+	sess_data->iov[1].iov_base = ntlmsspblob;
+	pSMB->req.SecurityBlobLength = cpu_to_le16(blob_len);
 
 	rc = _sess_auth_rawntlmssp_assemble_req(sess_data);
 	if (rc)
diff --git a/fs/cifs/smb2inode.c b/fs/cifs/smb2inode.c
index 8297703..fe5bfa2 100644
--- a/fs/cifs/smb2inode.c
+++ b/fs/cifs/smb2inode.c
@@ -46,6 +46,10 @@ struct cop_vars {
 	struct smb2_file_link_info link_info;
 };
 
+/*
+ * note: If cfile is passed, the reference to it is dropped here.
+ * So make sure that you do not reuse cfile after return from this func.
+ */
 static int
 smb2_compound_op(const unsigned int xid, struct cifs_tcon *tcon,
 		 struct cifs_sb_info *cifs_sb, const char *full_path,
@@ -536,10 +540,11 @@ smb2_query_path_info(const unsigned int xid, struct cifs_tcon *tcon,
 		create_options |= OPEN_REPARSE_POINT;
 
 		/* Failed on a symbolic link - query a reparse point info */
+		cifs_get_readable_path(tcon, full_path, &cfile);
 		rc = smb2_compound_op(xid, tcon, cifs_sb, full_path,
 				      FILE_READ_ATTRIBUTES, FILE_OPEN,
 				      create_options, ACL_NO_MODE,
-				      smb2_data, SMB2_OP_QUERY_INFO, NULL);
+				      smb2_data, SMB2_OP_QUERY_INFO, cfile);
 	}
 	if (rc)
 		goto out;
@@ -587,10 +592,11 @@ smb311_posix_query_path_info(const unsigned int xid, struct cifs_tcon *tcon,
 		create_options |= OPEN_REPARSE_POINT;
 
 		/* Failed on a symbolic link - query a reparse point info */
+		cifs_get_readable_path(tcon, full_path, &cfile);
 		rc = smb2_compound_op(xid, tcon, cifs_sb, full_path,
 				      FILE_READ_ATTRIBUTES, FILE_OPEN,
 				      create_options, ACL_NO_MODE,
-				      smb2_data, SMB2_OP_POSIX_QUERY_INFO, NULL);
+				      smb2_data, SMB2_OP_POSIX_QUERY_INFO, cfile);
 	}
 	if (rc)
 		goto out;
@@ -707,10 +713,12 @@ smb2_set_path_size(const unsigned int xid, struct cifs_tcon *tcon,
 		   struct cifs_sb_info *cifs_sb, bool set_alloc)
 {
 	__le64 eof = cpu_to_le64(size);
+	struct cifsFileInfo *cfile;
 
+	cifs_get_writable_path(tcon, full_path, FIND_WR_ANY, &cfile);
 	return smb2_compound_op(xid, tcon, cifs_sb, full_path,
 				FILE_WRITE_DATA, FILE_OPEN, 0, ACL_NO_MODE,
-				&eof, SMB2_OP_SET_EOF, NULL);
+				&eof, SMB2_OP_SET_EOF, cfile);
 }
 
 int
@@ -719,6 +727,8 @@ smb2_set_file_info(struct inode *inode, const char *full_path,
 {
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
 	struct tcon_link *tlink;
+	struct cifs_tcon *tcon;
+	struct cifsFileInfo *cfile;
 	int rc;
 
 	if ((buf->CreationTime == 0) && (buf->LastAccessTime == 0) &&
@@ -729,10 +739,12 @@ smb2_set_file_info(struct inode *inode, const char *full_path,
 	tlink = cifs_sb_tlink(cifs_sb);
 	if (IS_ERR(tlink))
 		return PTR_ERR(tlink);
+	tcon = tlink_tcon(tlink);
 
-	rc = smb2_compound_op(xid, tlink_tcon(tlink), cifs_sb, full_path,
+	cifs_get_writable_path(tcon, full_path, FIND_WR_ANY, &cfile);
+	rc = smb2_compound_op(xid, tcon, cifs_sb, full_path,
 			      FILE_WRITE_ATTRIBUTES, FILE_OPEN,
-			      0, ACL_NO_MODE, buf, SMB2_OP_SET_INFO, NULL);
+			      0, ACL_NO_MODE, buf, SMB2_OP_SET_INFO, cfile);
 	cifs_put_tlink(tlink);
 	return rc;
 }
diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c
index 7acf71de..c5b1dea 100644
--- a/fs/cifs/smb2ops.c
+++ b/fs/cifs/smb2ops.c
@@ -2844,6 +2844,7 @@ smb2_get_dfs_refer(const unsigned int xid, struct cifs_ses *ses,
 	struct fsctl_get_dfs_referral_req *dfs_req = NULL;
 	struct get_dfs_referral_rsp *dfs_rsp = NULL;
 	u32 dfs_req_size = 0, dfs_rsp_size = 0;
+	int retry_count = 0;
 
 	cifs_dbg(FYI, "%s: path: %s\n", __func__, search_name);
 
@@ -2895,11 +2896,14 @@ smb2_get_dfs_refer(const unsigned int xid, struct cifs_ses *ses,
 				true /* is_fsctl */,
 				(char *)dfs_req, dfs_req_size, CIFSMaxBufSize,
 				(char **)&dfs_rsp, &dfs_rsp_size);
-	} while (rc == -EAGAIN);
+		if (!is_retryable_error(rc))
+			break;
+		usleep_range(512, 2048);
+	} while (++retry_count < 5);
 
 	if (rc) {
-		if ((rc != -ENOENT) && (rc != -EOPNOTSUPP))
-			cifs_tcon_dbg(VFS, "ioctl error in %s rc=%d\n", __func__, rc);
+		if (!is_retryable_error(rc) && rc != -ENOENT && rc != -EOPNOTSUPP)
+			cifs_tcon_dbg(VFS, "%s: ioctl error: rc=%d\n", __func__, rc);
 		goto out;
 	}
 
diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c
index d2ecb2e..2f5f2c4 100644
--- a/fs/cifs/smb2pdu.c
+++ b/fs/cifs/smb2pdu.c
@@ -155,7 +155,11 @@ smb2_reconnect(__le16 smb2_command, struct cifs_tcon *tcon,
 	if (tcon == NULL)
 		return 0;
 
-	if (smb2_command == SMB2_TREE_CONNECT)
+	/*
+	 * Need to also skip SMB2_IOCTL because it is used for checking nested dfs links in
+	 * cifs_tree_connect().
+	 */
+	if (smb2_command == SMB2_TREE_CONNECT || smb2_command == SMB2_IOCTL)
 		return 0;
 
 	if (tcon->tidStatus == CifsExiting) {
@@ -253,7 +257,7 @@ smb2_reconnect(__le16 smb2_command, struct cifs_tcon *tcon,
 	/*
 	 * If we are reconnecting an extra channel, bind
 	 */
-	if (server->is_channel) {
+	if (CIFS_SERVER_IS_CHAN(server)) {
 		ses->binding = true;
 		ses->binding_chan = cifs_ses_find_chan(ses, server);
 	}
@@ -1456,7 +1460,7 @@ SMB2_sess_auth_rawntlmssp_negotiate(struct SMB2_sess_data *sess_data)
 	int rc;
 	struct cifs_ses *ses = sess_data->ses;
 	struct smb2_sess_setup_rsp *rsp = NULL;
-	char *ntlmssp_blob = NULL;
+	unsigned char *ntlmssp_blob = NULL;
 	bool use_spnego = false; /* else use raw ntlmssp */
 	u16 blob_length = 0;
 
@@ -1475,22 +1479,17 @@ SMB2_sess_auth_rawntlmssp_negotiate(struct SMB2_sess_data *sess_data)
 	if (rc)
 		goto out_err;
 
-	ntlmssp_blob = kmalloc(sizeof(struct _NEGOTIATE_MESSAGE),
-			       GFP_KERNEL);
-	if (ntlmssp_blob == NULL) {
-		rc = -ENOMEM;
-		goto out;
-	}
+	rc = build_ntlmssp_negotiate_blob(&ntlmssp_blob,
+					  &blob_length, ses,
+					  sess_data->nls_cp);
+	if (rc)
+		goto out_err;
 
-	build_ntlmssp_negotiate_blob(ntlmssp_blob, ses);
 	if (use_spnego) {
 		/* BB eventually need to add this */
 		cifs_dbg(VFS, "spnego not supported for SMB2 yet\n");
 		rc = -EOPNOTSUPP;
 		goto out;
-	} else {
-		blob_length = sizeof(struct _NEGOTIATE_MESSAGE);
-		/* with raw NTLMSSP we don't encapsulate in SPNEGO */
 	}
 	sess_data->iov[1].iov_base = ntlmssp_blob;
 	sess_data->iov[1].iov_len = blob_length;
@@ -1841,7 +1840,7 @@ SMB2_tcon(const unsigned int xid, struct cifs_ses *ses, const char *tree,
 	cifs_small_buf_release(req);
 	rsp = (struct smb2_tree_connect_rsp *)rsp_iov.iov_base;
 	trace_smb3_tcon(xid, tcon->tid, ses->Suid, tree, rc);
-	if (rc != 0) {
+	if ((rc != 0) || (rsp == NULL)) {
 		cifs_stats_fail_inc(tcon, SMB2_TREE_CONNECT_HE);
 		tcon->need_reconnect = true;
 		goto tcon_error_exit;
@@ -2669,7 +2668,18 @@ int smb311_posix_mkdir(const unsigned int xid, struct inode *inode,
 		goto err_free_rsp_buf;
 	}
 
+	/*
+	 * Although unlikely to be possible for rsp to be null and rc not set,
+	 * adding check below is slightly safer long term (and quiets Coverity
+	 * warning)
+	 */
 	rsp = (struct smb2_create_rsp *)rsp_iov.iov_base;
+	if (rsp == NULL) {
+		rc = -EIO;
+		kfree(pc_buf);
+		goto err_free_req;
+	}
+
 	trace_smb3_posix_mkdir_done(xid, le64_to_cpu(rsp->PersistentFileId),
 				    tcon->tid,
 				    ses->Suid, CREATE_NOT_FILE,
@@ -2942,7 +2952,9 @@ SMB2_open(const unsigned int xid, struct cifs_open_parms *oparms, __le16 *path,
 			tcon->need_reconnect = true;
 		}
 		goto creat_exit;
-	} else
+	} else if (rsp == NULL) /* unlikely to happen, but safer to check */
+		goto creat_exit;
+	else
 		trace_smb3_open_done(xid, le64_to_cpu(rsp->PersistentFileId),
 				     tcon->tid,
 				     ses->Suid, oparms->create_options,
@@ -3163,6 +3175,16 @@ SMB2_ioctl(const unsigned int xid, struct cifs_tcon *tcon, u64 persistent_fid,
 	if ((plen == NULL) || (out_data == NULL))
 		goto ioctl_exit;
 
+	/*
+	 * Although unlikely to be possible for rsp to be null and rc not set,
+	 * adding check below is slightly safer long term (and quiets Coverity
+	 * warning)
+	 */
+	if (rsp == NULL) {
+		rc = -EIO;
+		goto ioctl_exit;
+	}
+
 	*plen = le32_to_cpu(rsp->OutputCount);
 
 	/* We check for obvious errors in the output buffer length and offset */
diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c
index b737932..61ea3d3 100644
--- a/fs/cifs/transport.c
+++ b/fs/cifs/transport.c
@@ -1044,14 +1044,17 @@ struct TCP_Server_Info *cifs_pick_channel(struct cifs_ses *ses)
 	if (!ses)
 		return NULL;
 
+	spin_lock(&ses->chan_lock);
 	if (!ses->binding) {
 		/* round robin */
 		if (ses->chan_count > 1) {
 			index = (uint)atomic_inc_return(&ses->chan_seq);
 			index %= ses->chan_count;
 		}
+		spin_unlock(&ses->chan_lock);
 		return ses->chans[index].server;
 	} else {
+		spin_unlock(&ses->chan_lock);
 		return cifs_ses_server(ses);
 	}
 }
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index bcb1b91..9a249bf 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -96,16 +96,9 @@ static void z_erofs_free_pcluster(struct z_erofs_pcluster *pcl)
 	DBG_BUGON(1);
 }
 
-/*
- * a compressed_pages[] placeholder in order to avoid
- * being filled with file pages for in-place decompression.
- */
-#define PAGE_UNALLOCATED     ((void *)0x5F0E4B1D)
-
 /* how to allocate cached pages for a pcluster */
 enum z_erofs_cache_alloctype {
 	DONTALLOC,	/* don't allocate any cached pages */
-	DELAYEDALLOC,	/* delayed allocation (at the time of submitting io) */
 	/*
 	 * try to use cached I/O if page allocation succeeds or fallback
 	 * to in-place I/O instead to avoid any direct reclaim.
@@ -267,10 +260,6 @@ static void preload_compressed_pages(struct z_erofs_collector *clt,
 			/* I/O is needed, no possible to decompress directly */
 			standalone = false;
 			switch (type) {
-			case DELAYEDALLOC:
-				t = tagptr_init(compressed_page_t,
-						PAGE_UNALLOCATED);
-				break;
 			case TRYALLOC:
 				newpage = erofs_allocpage(pagepool, gfp);
 				if (!newpage)
@@ -371,8 +360,8 @@ static bool z_erofs_try_inplace_io(struct z_erofs_collector *clt,
 
 /* callers must be with collection lock held */
 static int z_erofs_attach_page(struct z_erofs_collector *clt,
-			       struct page *page,
-			       enum z_erofs_page_type type)
+			       struct page *page, enum z_erofs_page_type type,
+			       bool pvec_safereuse)
 {
 	int ret;
 
@@ -382,9 +371,9 @@ static int z_erofs_attach_page(struct z_erofs_collector *clt,
 	    z_erofs_try_inplace_io(clt, page))
 		return 0;
 
-	ret = z_erofs_pagevec_enqueue(&clt->vector, page, type);
+	ret = z_erofs_pagevec_enqueue(&clt->vector, page, type,
+				      pvec_safereuse);
 	clt->cl->vcnt += (unsigned int)ret;
-
 	return ret ? 0 : -EAGAIN;
 }
 
@@ -727,7 +716,8 @@ static int z_erofs_do_read_page(struct z_erofs_decompress_frontend *fe,
 		tight &= (clt->mode >= COLLECT_PRIMARY_FOLLOWED);
 
 retry:
-	err = z_erofs_attach_page(clt, page, page_type);
+	err = z_erofs_attach_page(clt, page, page_type,
+				  clt->mode >= COLLECT_PRIMARY_FOLLOWED);
 	/* should allocate an additional short-lived page for pagevec */
 	if (err == -EAGAIN) {
 		struct page *const newpage =
@@ -735,7 +725,7 @@ static int z_erofs_do_read_page(struct z_erofs_decompress_frontend *fe,
 
 		set_page_private(newpage, Z_EROFS_SHORTLIVED_PAGE);
 		err = z_erofs_attach_page(clt, newpage,
-					  Z_EROFS_PAGE_TYPE_EXCLUSIVE);
+					  Z_EROFS_PAGE_TYPE_EXCLUSIVE, true);
 		if (!err)
 			goto retry;
 	}
@@ -1089,15 +1079,6 @@ static struct page *pickup_page_for_submission(struct z_erofs_pcluster *pcl,
 	if (!page)
 		goto out_allocpage;
 
-	/*
-	 * the cached page has not been allocated and
-	 * an placeholder is out there, prepare it now.
-	 */
-	if (page == PAGE_UNALLOCATED) {
-		tocache = true;
-		goto out_allocpage;
-	}
-
 	/* process the target tagged pointer */
 	t = tagptr_init(compressed_page_t, page);
 	justfound = tagptr_unfold_tags(t);
diff --git a/fs/erofs/zdata.h b/fs/erofs/zdata.h
index 879df53..4a69515 100644
--- a/fs/erofs/zdata.h
+++ b/fs/erofs/zdata.h
@@ -179,4 +179,3 @@ static inline void z_erofs_onlinepage_endio(struct page *page)
 #define Z_EROFS_VMAP_GLOBAL_PAGES	2048
 
 #endif
-
diff --git a/fs/erofs/zpvec.h b/fs/erofs/zpvec.h
index dfd7fe0..b05464f 100644
--- a/fs/erofs/zpvec.h
+++ b/fs/erofs/zpvec.h
@@ -106,11 +106,18 @@ static inline void z_erofs_pagevec_ctor_init(struct z_erofs_pagevec_ctor *ctor,
 
 static inline bool z_erofs_pagevec_enqueue(struct z_erofs_pagevec_ctor *ctor,
 					   struct page *page,
-					   enum z_erofs_page_type type)
+					   enum z_erofs_page_type type,
+					   bool pvec_safereuse)
 {
-	if (!ctor->next && type)
-		if (ctor->index + 1 == ctor->nr)
+	if (!ctor->next) {
+		/* some pages cannot be reused as pvec safely without I/O */
+		if (type == Z_EROFS_PAGE_TYPE_EXCLUSIVE && !pvec_safereuse)
+			type = Z_EROFS_VLE_PAGE_TYPE_TAIL_SHARED;
+
+		if (type != Z_EROFS_PAGE_TYPE_EXCLUSIVE &&
+		    ctor->index + 1 == ctor->nr)
 			return false;
+	}
 
 	if (ctor->index >= ctor->nr)
 		z_erofs_pagevec_ctor_pagedown(ctor, false);
diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c
index 83e9bc0..f1693d4 100644
--- a/fs/f2fs/checkpoint.c
+++ b/fs/f2fs/checkpoint.c
@@ -653,7 +653,7 @@ static int recover_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
 		return PTR_ERR(inode);
 	}
 
-	err = dquot_initialize(inode);
+	err = f2fs_dquot_initialize(inode);
 	if (err) {
 		iput(inode);
 		goto err_out;
@@ -705,9 +705,6 @@ int f2fs_recover_orphan_inodes(struct f2fs_sb_info *sbi)
 	}
 
 #ifdef CONFIG_QUOTA
-	/* Needed for iput() to work correctly and not trash data */
-	sbi->sb->s_flags |= SB_ACTIVE;
-
 	/*
 	 * Turn on quotas which were not enabled for read-only mounts if
 	 * filesystem has quota feature, so that they are updated correctly.
@@ -1162,7 +1159,8 @@ static bool __need_flush_quota(struct f2fs_sb_info *sbi)
 	if (!is_journalled_quota(sbi))
 		return false;
 
-	down_write(&sbi->quota_sem);
+	if (!down_write_trylock(&sbi->quota_sem))
+		return true;
 	if (is_sbi_flag_set(sbi, SBI_QUOTA_SKIP_FLUSH)) {
 		ret = false;
 	} else if (is_sbi_flag_set(sbi, SBI_QUOTA_NEED_REPAIR)) {
diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c
index 20a083d..49121a2 100644
--- a/fs/f2fs/compress.c
+++ b/fs/f2fs/compress.c
@@ -336,8 +336,8 @@ static const struct f2fs_compress_ops f2fs_lz4_ops = {
 
 static int zstd_init_compress_ctx(struct compress_ctx *cc)
 {
-	ZSTD_parameters params;
-	ZSTD_CStream *stream;
+	zstd_parameters params;
+	zstd_cstream *stream;
 	void *workspace;
 	unsigned int workspace_size;
 	unsigned char level = F2FS_I(cc->inode)->i_compress_flag >>
@@ -346,17 +346,17 @@ static int zstd_init_compress_ctx(struct compress_ctx *cc)
 	if (!level)
 		level = F2FS_ZSTD_DEFAULT_CLEVEL;
 
-	params = ZSTD_getParams(level, cc->rlen, 0);
-	workspace_size = ZSTD_CStreamWorkspaceBound(params.cParams);
+	params = zstd_get_params(F2FS_ZSTD_DEFAULT_CLEVEL, cc->rlen);
+	workspace_size = zstd_cstream_workspace_bound(&params.cParams);
 
 	workspace = f2fs_kvmalloc(F2FS_I_SB(cc->inode),
 					workspace_size, GFP_NOFS);
 	if (!workspace)
 		return -ENOMEM;
 
-	stream = ZSTD_initCStream(params, 0, workspace, workspace_size);
+	stream = zstd_init_cstream(&params, 0, workspace, workspace_size);
 	if (!stream) {
-		printk_ratelimited("%sF2FS-fs (%s): %s ZSTD_initCStream failed\n",
+		printk_ratelimited("%sF2FS-fs (%s): %s zstd_init_cstream failed\n",
 				KERN_ERR, F2FS_I_SB(cc->inode)->sb->s_id,
 				__func__);
 		kvfree(workspace);
@@ -379,9 +379,9 @@ static void zstd_destroy_compress_ctx(struct compress_ctx *cc)
 
 static int zstd_compress_pages(struct compress_ctx *cc)
 {
-	ZSTD_CStream *stream = cc->private2;
-	ZSTD_inBuffer inbuf;
-	ZSTD_outBuffer outbuf;
+	zstd_cstream *stream = cc->private2;
+	zstd_in_buffer inbuf;
+	zstd_out_buffer outbuf;
 	int src_size = cc->rlen;
 	int dst_size = src_size - PAGE_SIZE - COMPRESS_HEADER_SIZE;
 	int ret;
@@ -394,19 +394,19 @@ static int zstd_compress_pages(struct compress_ctx *cc)
 	outbuf.dst = cc->cbuf->cdata;
 	outbuf.size = dst_size;
 
-	ret = ZSTD_compressStream(stream, &outbuf, &inbuf);
-	if (ZSTD_isError(ret)) {
-		printk_ratelimited("%sF2FS-fs (%s): %s ZSTD_compressStream failed, ret: %d\n",
+	ret = zstd_compress_stream(stream, &outbuf, &inbuf);
+	if (zstd_is_error(ret)) {
+		printk_ratelimited("%sF2FS-fs (%s): %s zstd_compress_stream failed, ret: %d\n",
 				KERN_ERR, F2FS_I_SB(cc->inode)->sb->s_id,
-				__func__, ZSTD_getErrorCode(ret));
+				__func__, zstd_get_error_code(ret));
 		return -EIO;
 	}
 
-	ret = ZSTD_endStream(stream, &outbuf);
-	if (ZSTD_isError(ret)) {
-		printk_ratelimited("%sF2FS-fs (%s): %s ZSTD_endStream returned %d\n",
+	ret = zstd_end_stream(stream, &outbuf);
+	if (zstd_is_error(ret)) {
+		printk_ratelimited("%sF2FS-fs (%s): %s zstd_end_stream returned %d\n",
 				KERN_ERR, F2FS_I_SB(cc->inode)->sb->s_id,
-				__func__, ZSTD_getErrorCode(ret));
+				__func__, zstd_get_error_code(ret));
 		return -EIO;
 	}
 
@@ -423,22 +423,22 @@ static int zstd_compress_pages(struct compress_ctx *cc)
 
 static int zstd_init_decompress_ctx(struct decompress_io_ctx *dic)
 {
-	ZSTD_DStream *stream;
+	zstd_dstream *stream;
 	void *workspace;
 	unsigned int workspace_size;
 	unsigned int max_window_size =
 			MAX_COMPRESS_WINDOW_SIZE(dic->log_cluster_size);
 
-	workspace_size = ZSTD_DStreamWorkspaceBound(max_window_size);
+	workspace_size = zstd_dstream_workspace_bound(max_window_size);
 
 	workspace = f2fs_kvmalloc(F2FS_I_SB(dic->inode),
 					workspace_size, GFP_NOFS);
 	if (!workspace)
 		return -ENOMEM;
 
-	stream = ZSTD_initDStream(max_window_size, workspace, workspace_size);
+	stream = zstd_init_dstream(max_window_size, workspace, workspace_size);
 	if (!stream) {
-		printk_ratelimited("%sF2FS-fs (%s): %s ZSTD_initDStream failed\n",
+		printk_ratelimited("%sF2FS-fs (%s): %s zstd_init_dstream failed\n",
 				KERN_ERR, F2FS_I_SB(dic->inode)->sb->s_id,
 				__func__);
 		kvfree(workspace);
@@ -460,9 +460,9 @@ static void zstd_destroy_decompress_ctx(struct decompress_io_ctx *dic)
 
 static int zstd_decompress_pages(struct decompress_io_ctx *dic)
 {
-	ZSTD_DStream *stream = dic->private2;
-	ZSTD_inBuffer inbuf;
-	ZSTD_outBuffer outbuf;
+	zstd_dstream *stream = dic->private2;
+	zstd_in_buffer inbuf;
+	zstd_out_buffer outbuf;
 	int ret;
 
 	inbuf.pos = 0;
@@ -473,11 +473,11 @@ static int zstd_decompress_pages(struct decompress_io_ctx *dic)
 	outbuf.dst = dic->rbuf;
 	outbuf.size = dic->rlen;
 
-	ret = ZSTD_decompressStream(stream, &outbuf, &inbuf);
-	if (ZSTD_isError(ret)) {
-		printk_ratelimited("%sF2FS-fs (%s): %s ZSTD_compressStream failed, ret: %d\n",
+	ret = zstd_decompress_stream(stream, &outbuf, &inbuf);
+	if (zstd_is_error(ret)) {
+		printk_ratelimited("%sF2FS-fs (%s): %s zstd_decompress_stream failed, ret: %d\n",
 				KERN_ERR, F2FS_I_SB(dic->inode)->sb->s_id,
-				__func__, ZSTD_getErrorCode(ret));
+				__func__, zstd_get_error_code(ret));
 		return -EIO;
 	}
 
@@ -882,6 +882,25 @@ bool f2fs_cluster_can_merge_page(struct compress_ctx *cc, pgoff_t index)
 	return is_page_in_cluster(cc, index);
 }
 
+bool f2fs_all_cluster_page_loaded(struct compress_ctx *cc, struct pagevec *pvec,
+				int index, int nr_pages)
+{
+	unsigned long pgidx;
+	int i;
+
+	if (nr_pages - index < cc->cluster_size)
+		return false;
+
+	pgidx = pvec->pages[index]->index;
+
+	for (i = 1; i < cc->cluster_size; i++) {
+		if (pvec->pages[index + i]->index != pgidx + i)
+			return false;
+	}
+
+	return true;
+}
+
 static bool cluster_has_invalid_data(struct compress_ctx *cc)
 {
 	loff_t i_size = i_size_read(cc->inode);
@@ -1531,6 +1550,7 @@ int f2fs_write_multi_pages(struct compress_ctx *cc,
 	if (cluster_may_compress(cc)) {
 		err = f2fs_compress_pages(cc);
 		if (err == -EAGAIN) {
+			add_compr_block_stat(cc->inode, cc->cluster_size);
 			goto write;
 		} else if (err) {
 			f2fs_put_rpages_wbc(cc, wbc, true, 1);
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index f4fd6c2..9f754aa 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -1465,10 +1465,15 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 	struct extent_info ei = {0, };
 	block_t blkaddr;
 	unsigned int start_pgofs;
+	int bidx = 0;
 
 	if (!maxblocks)
 		return 0;
 
+	map->m_bdev = inode->i_sb->s_bdev;
+	map->m_multidev_dio =
+		f2fs_allow_multi_device_dio(F2FS_I_SB(inode), flag);
+
 	map->m_len = 0;
 	map->m_flags = 0;
 
@@ -1491,6 +1496,21 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 		if (flag == F2FS_GET_BLOCK_DIO)
 			f2fs_wait_on_block_writeback_range(inode,
 						map->m_pblk, map->m_len);
+
+		if (map->m_multidev_dio) {
+			block_t blk_addr = map->m_pblk;
+
+			bidx = f2fs_target_device_index(sbi, map->m_pblk);
+
+			map->m_bdev = FDEV(bidx).bdev;
+			map->m_pblk -= FDEV(bidx).start_blk;
+			map->m_len = min(map->m_len,
+				FDEV(bidx).end_blk + 1 - map->m_pblk);
+
+			if (map->m_may_create)
+				f2fs_update_device_state(sbi, inode->i_ino,
+							blk_addr, map->m_len);
+		}
 		goto out;
 	}
 
@@ -1609,6 +1629,9 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 	if (flag == F2FS_GET_BLOCK_PRE_AIO)
 		goto skip;
 
+	if (map->m_multidev_dio)
+		bidx = f2fs_target_device_index(sbi, blkaddr);
+
 	if (map->m_len == 0) {
 		/* preallocated unwritten block should be mapped for fiemap. */
 		if (blkaddr == NEW_ADDR)
@@ -1617,10 +1640,15 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 
 		map->m_pblk = blkaddr;
 		map->m_len = 1;
+
+		if (map->m_multidev_dio)
+			map->m_bdev = FDEV(bidx).bdev;
 	} else if ((map->m_pblk != NEW_ADDR &&
 			blkaddr == (map->m_pblk + ofs)) ||
 			(map->m_pblk == NEW_ADDR && blkaddr == NEW_ADDR) ||
 			flag == F2FS_GET_BLOCK_PRE_DIO) {
+		if (map->m_multidev_dio && map->m_bdev != FDEV(bidx).bdev)
+			goto sync_out;
 		ofs++;
 		map->m_len++;
 	} else {
@@ -1673,10 +1701,32 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 
 sync_out:
 
-	/* for hardware encryption, but to avoid potential issue in future */
-	if (flag == F2FS_GET_BLOCK_DIO && map->m_flags & F2FS_MAP_MAPPED)
+	if (flag == F2FS_GET_BLOCK_DIO && map->m_flags & F2FS_MAP_MAPPED) {
+		/*
+		 * for hardware encryption, but to avoid potential issue
+		 * in future
+		 */
 		f2fs_wait_on_block_writeback_range(inode,
 						map->m_pblk, map->m_len);
+		invalidate_mapping_pages(META_MAPPING(sbi),
+						map->m_pblk, map->m_pblk);
+
+		if (map->m_multidev_dio) {
+			block_t blk_addr = map->m_pblk;
+
+			bidx = f2fs_target_device_index(sbi, map->m_pblk);
+
+			map->m_bdev = FDEV(bidx).bdev;
+			map->m_pblk -= FDEV(bidx).start_blk;
+
+			if (map->m_may_create)
+				f2fs_update_device_state(sbi, inode->i_ino,
+							blk_addr, map->m_len);
+
+			f2fs_bug_on(sbi, blk_addr + map->m_len >
+						FDEV(bidx).end_blk + 1);
+		}
+	}
 
 	if (flag == F2FS_GET_BLOCK_PRECACHE) {
 		if (map->m_flags & F2FS_MAP_MAPPED) {
@@ -1696,7 +1746,7 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 		f2fs_balance_fs(sbi, dn.node_changed);
 	}
 out:
-	trace_f2fs_map_blocks(inode, map, err);
+	trace_f2fs_map_blocks(inode, map, create, flag, err);
 	return err;
 }
 
@@ -1755,6 +1805,9 @@ static int __get_data_block(struct inode *inode, sector_t iblock,
 		map_bh(bh, inode->i_sb, map.m_pblk);
 		bh->b_state = (bh->b_state & ~F2FS_MAP_FLAGS) | map.m_flags;
 		bh->b_size = blks_to_bytes(inode, map.m_len);
+
+		if (map.m_multidev_dio)
+			bh->b_bdev = map.m_bdev;
 	}
 	return err;
 }
@@ -2989,6 +3042,10 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
 			need_readd = false;
 #ifdef CONFIG_F2FS_FS_COMPRESSION
 			if (f2fs_compressed_file(inode)) {
+				void *fsdata = NULL;
+				struct page *pagep;
+				int ret2;
+
 				ret = f2fs_init_compress_ctx(&cc);
 				if (ret) {
 					done = 1;
@@ -3007,27 +3064,23 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
 				if (unlikely(f2fs_cp_error(sbi)))
 					goto lock_page;
 
-				if (f2fs_cluster_is_empty(&cc)) {
-					void *fsdata = NULL;
-					struct page *pagep;
-					int ret2;
+				if (!f2fs_cluster_is_empty(&cc))
+					goto lock_page;
 
-					ret2 = f2fs_prepare_compress_overwrite(
+				ret2 = f2fs_prepare_compress_overwrite(
 							inode, &pagep,
 							page->index, &fsdata);
-					if (ret2 < 0) {
-						ret = ret2;
-						done = 1;
-						break;
-					} else if (ret2 &&
-						!f2fs_compress_write_end(inode,
-								fsdata, page->index,
-								1)) {
-						retry = 1;
-						break;
-					}
-				} else {
-					goto lock_page;
+				if (ret2 < 0) {
+					ret = ret2;
+					done = 1;
+					break;
+				} else if (ret2 &&
+					(!f2fs_compress_write_end(inode,
+						fsdata, page->index, 1) ||
+					 !f2fs_all_cluster_page_loaded(&cc,
+						&pvec, i, nr_pages))) {
+					retry = 1;
+					break;
 				}
 			}
 #endif
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index b339ae8..ce9fc9f 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -55,6 +55,7 @@ enum {
 	FAULT_DISCARD,
 	FAULT_WRITE_IO,
 	FAULT_SLAB_ALLOC,
+	FAULT_DQUOT_INIT,
 	FAULT_MAX,
 };
 
@@ -561,6 +562,9 @@ enum {
 
 #define MAX_DIR_RA_PAGES	4	/* maximum ra pages of dir */
 
+/* dirty segments threshold for triggering CP */
+#define DEFAULT_DIRTY_THRESHOLD		4
+
 /* for in-memory extent cache entry */
 #define F2FS_MIN_EXTENT_LEN	64	/* minimum extent length */
 
@@ -617,6 +621,7 @@ struct extent_tree {
 				F2FS_MAP_UNWRITTEN)
 
 struct f2fs_map_blocks {
+	struct block_device *m_bdev;	/* for multi-device dio */
 	block_t m_pblk;
 	block_t m_lblk;
 	unsigned int m_len;
@@ -625,6 +630,7 @@ struct f2fs_map_blocks {
 	pgoff_t *m_next_extent;		/* point to next possible extent */
 	int m_seg_type;
 	bool m_may_create;		/* indicate it is from write path */
+	bool m_multidev_dio;		/* indicate it allows multi-device dio */
 };
 
 /* for flag in get_data_block */
@@ -1284,8 +1290,10 @@ enum {
 };
 
 enum {
-	FS_MODE_ADAPTIVE,	/* use both lfs/ssr allocation */
-	FS_MODE_LFS,		/* use lfs allocation only */
+	FS_MODE_ADAPTIVE,		/* use both lfs/ssr allocation */
+	FS_MODE_LFS,			/* use lfs allocation only */
+	FS_MODE_FRAGMENT_SEG,		/* segment fragmentation mode */
+	FS_MODE_FRAGMENT_BLK,		/* block fragmentation mode */
 };
 
 enum {
@@ -1728,12 +1736,15 @@ struct f2fs_sb_info {
 
 	/* For shrinker support */
 	struct list_head s_list;
+	struct mutex umount_mutex;
+	unsigned int shrinker_run_no;
+
+	/* For multi devices */
 	int s_ndevs;				/* number of devices */
 	struct f2fs_dev_info *devs;		/* for device list */
 	unsigned int dirty_device;		/* for checkpoint data flush */
 	spinlock_t dev_lock;			/* protect dirty_device */
-	struct mutex umount_mutex;
-	unsigned int shrinker_run_no;
+	bool aligned_blksize;			/* all devices has the same logical blksize */
 
 	/* For write statistics */
 	u64 sectors_written_start;
@@ -1756,6 +1767,9 @@ struct f2fs_sb_info {
 
 	unsigned long seq_file_ra_mul;		/* multiplier for ra_pages of seq. files in fadvise */
 
+	int max_fragment_chunk;			/* max chunk size for block fragmentation mode */
+	int max_fragment_hole;			/* max hole size for block fragmentation mode */
+
 #ifdef CONFIG_F2FS_FS_COMPRESSION
 	struct kmem_cache *page_array_slab;	/* page array entry */
 	unsigned int page_array_slab_size;	/* default page array slab size */
@@ -3363,6 +3377,7 @@ static inline int f2fs_add_link(struct dentry *dentry, struct inode *inode)
  */
 int f2fs_inode_dirtied(struct inode *inode, bool sync);
 void f2fs_inode_synced(struct inode *inode);
+int f2fs_dquot_initialize(struct inode *inode);
 int f2fs_enable_quota_files(struct f2fs_sb_info *sbi, bool rdonly);
 int f2fs_quota_sync(struct super_block *sb, int type);
 loff_t max_file_blocks(struct inode *inode);
@@ -3492,6 +3507,8 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
 			block_t old_blkaddr, block_t *new_blkaddr,
 			struct f2fs_summary *sum, int type,
 			struct f2fs_io_info *fio);
+void f2fs_update_device_state(struct f2fs_sb_info *sbi, nid_t ino,
+					block_t blkaddr, unsigned int blkcnt);
 void f2fs_wait_on_page_writeback(struct page *page,
 			enum page_type type, bool ordered, bool locked);
 void f2fs_wait_on_block_writeback(struct inode *inode, block_t blkaddr);
@@ -3516,6 +3533,16 @@ unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi,
 unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi,
 			unsigned int segno);
 
+#define DEF_FRAGMENT_SIZE	4
+#define MIN_FRAGMENT_SIZE	1
+#define MAX_FRAGMENT_SIZE	512
+
+static inline bool f2fs_need_rand_seg(struct f2fs_sb_info *sbi)
+{
+	return F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_SEG ||
+		F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK;
+}
+
 /*
  * checkpoint.c
  */
@@ -4027,6 +4054,8 @@ void f2fs_end_read_compressed_page(struct page *page, bool failed,
 							block_t blkaddr);
 bool f2fs_cluster_is_empty(struct compress_ctx *cc);
 bool f2fs_cluster_can_merge_page(struct compress_ctx *cc, pgoff_t index);
+bool f2fs_all_cluster_page_loaded(struct compress_ctx *cc, struct pagevec *pvec,
+				int index, int nr_pages);
 bool f2fs_sanity_check_cluster(struct dnode_of_data *dn);
 void f2fs_compress_ctx_add_page(struct compress_ctx *cc, struct page *page);
 int f2fs_write_multi_pages(struct compress_ctx *cc,
@@ -4152,8 +4181,7 @@ static inline bool f2fs_disable_compressed_file(struct inode *inode)
 
 	if (!f2fs_compressed_file(inode))
 		return true;
-	if (S_ISREG(inode->i_mode) &&
-		(get_dirty_pages(inode) || atomic_read(&fi->i_compr_blocks)))
+	if (S_ISREG(inode->i_mode) && F2FS_HAS_BLOCKS(inode))
 		return false;
 
 	fi->i_flags &= ~F2FS_COMPR_FL;
@@ -4302,6 +4330,16 @@ static inline int block_unaligned_IO(struct inode *inode,
 	return align & blocksize_mask;
 }
 
+static inline bool f2fs_allow_multi_device_dio(struct f2fs_sb_info *sbi,
+								int flag)
+{
+	if (!f2fs_is_multi_device(sbi))
+		return false;
+	if (flag != F2FS_GET_BLOCK_DIO)
+		return false;
+	return sbi->aligned_blksize;
+}
+
 static inline bool f2fs_force_buffered_io(struct inode *inode,
 				struct kiocb *iocb, struct iov_iter *iter)
 {
@@ -4310,7 +4348,9 @@ static inline bool f2fs_force_buffered_io(struct inode *inode,
 
 	if (f2fs_post_read_required(inode))
 		return true;
-	if (f2fs_is_multi_device(sbi))
+
+	/* disallow direct IO if any of devices has unaligned blksize */
+	if (f2fs_is_multi_device(sbi) && !sbi->aligned_blksize)
 		return true;
 	/*
 	 * for blkzoned device, fallback direct IO to buffered IO, so
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index eb971e1..92ec269 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -786,7 +786,7 @@ int f2fs_truncate(struct inode *inode)
 		return -EIO;
 	}
 
-	err = dquot_initialize(inode);
+	err = f2fs_dquot_initialize(inode);
 	if (err)
 		return err;
 
@@ -916,7 +916,7 @@ int f2fs_setattr(struct user_namespace *mnt_userns, struct dentry *dentry,
 		return err;
 
 	if (is_quota_modification(inode, attr)) {
-		err = dquot_initialize(inode);
+		err = f2fs_dquot_initialize(inode);
 		if (err)
 			return err;
 	}
@@ -3020,7 +3020,7 @@ static int f2fs_ioc_setproject(struct inode *inode, __u32 projid)
 	}
 	f2fs_put_page(ipage, 1);
 
-	err = dquot_initialize(inode);
+	err = f2fs_dquot_initialize(inode);
 	if (err)
 		return err;
 
diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
index 77391e3..a946ce0 100644
--- a/fs/f2fs/gc.c
+++ b/fs/f2fs/gc.c
@@ -14,6 +14,7 @@
 #include <linux/delay.h>
 #include <linux/freezer.h>
 #include <linux/sched/signal.h>
+#include <linux/random.h>
 
 #include "f2fs.h"
 #include "node.h"
@@ -257,7 +258,9 @@ static void select_policy(struct f2fs_sb_info *sbi, int gc_type,
 		p->max_search = sbi->max_victim_search;
 
 	/* let's select beginning hot/small space first in no_heap mode*/
-	if (test_opt(sbi, NOHEAP) &&
+	if (f2fs_need_rand_seg(sbi))
+		p->offset = prandom_u32() % (MAIN_SECS(sbi) * sbi->segs_per_sec);
+	else if (test_opt(sbi, NOHEAP) &&
 		(type == CURSEG_HOT_DATA || IS_NODESEG(type)))
 		p->offset = 0;
 	else
diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c
index 56a20d5..ea08f0d 100644
--- a/fs/f2fs/inline.c
+++ b/fs/f2fs/inline.c
@@ -192,7 +192,7 @@ int f2fs_convert_inline_inode(struct inode *inode)
 			f2fs_hw_is_readonly(sbi) || f2fs_readonly(sbi->sb))
 		return 0;
 
-	err = dquot_initialize(inode);
+	err = f2fs_dquot_initialize(inode);
 	if (err)
 		return err;
 
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index 9141147..0f8b2df 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -527,7 +527,7 @@ struct inode *f2fs_iget(struct super_block *sb, unsigned long ino)
 		inode->i_op = &f2fs_dir_inode_operations;
 		inode->i_fop = &f2fs_dir_operations;
 		inode->i_mapping->a_ops = &f2fs_dblock_aops;
-		inode_nohighmem(inode);
+		mapping_set_gfp_mask(inode->i_mapping, GFP_NOFS);
 	} else if (S_ISLNK(inode->i_mode)) {
 		if (file_is_encrypt(inode))
 			inode->i_op = &f2fs_encrypted_symlink_inode_operations;
@@ -754,7 +754,7 @@ void f2fs_evict_inode(struct inode *inode)
 	if (inode->i_nlink || is_bad_inode(inode))
 		goto no_delete;
 
-	err = dquot_initialize(inode);
+	err = f2fs_dquot_initialize(inode);
 	if (err) {
 		err = 0;
 		set_sbi_flag(sbi, SBI_QUOTA_NEED_REPAIR);
diff --git a/fs/f2fs/namei.c b/fs/f2fs/namei.c
index 9c528e5..a728a0a 100644
--- a/fs/f2fs/namei.c
+++ b/fs/f2fs/namei.c
@@ -74,7 +74,7 @@ static struct inode *f2fs_new_inode(struct inode *dir, umode_t mode)
 	if (err)
 		goto fail_drop;
 
-	err = dquot_initialize(inode);
+	err = f2fs_dquot_initialize(inode);
 	if (err)
 		goto fail_drop;
 
@@ -345,7 +345,7 @@ static int f2fs_create(struct user_namespace *mnt_userns, struct inode *dir,
 	if (!f2fs_is_checkpoint_ready(sbi))
 		return -ENOSPC;
 
-	err = dquot_initialize(dir);
+	err = f2fs_dquot_initialize(dir);
 	if (err)
 		return err;
 
@@ -404,7 +404,7 @@ static int f2fs_link(struct dentry *old_dentry, struct inode *dir,
 			F2FS_I(old_dentry->d_inode)->i_projid)))
 		return -EXDEV;
 
-	err = dquot_initialize(dir);
+	err = f2fs_dquot_initialize(dir);
 	if (err)
 		return err;
 
@@ -460,7 +460,7 @@ static int __recover_dot_dentries(struct inode *dir, nid_t pino)
 		return 0;
 	}
 
-	err = dquot_initialize(dir);
+	err = f2fs_dquot_initialize(dir);
 	if (err)
 		return err;
 
@@ -598,10 +598,10 @@ static int f2fs_unlink(struct inode *dir, struct dentry *dentry)
 		goto fail;
 	}
 
-	err = dquot_initialize(dir);
+	err = f2fs_dquot_initialize(dir);
 	if (err)
 		goto fail;
-	err = dquot_initialize(inode);
+	err = f2fs_dquot_initialize(inode);
 	if (err)
 		goto fail;
 
@@ -675,7 +675,7 @@ static int f2fs_symlink(struct user_namespace *mnt_userns, struct inode *dir,
 	if (err)
 		return err;
 
-	err = dquot_initialize(dir);
+	err = f2fs_dquot_initialize(dir);
 	if (err)
 		return err;
 
@@ -746,7 +746,7 @@ static int f2fs_mkdir(struct user_namespace *mnt_userns, struct inode *dir,
 	if (unlikely(f2fs_cp_error(sbi)))
 		return -EIO;
 
-	err = dquot_initialize(dir);
+	err = f2fs_dquot_initialize(dir);
 	if (err)
 		return err;
 
@@ -757,7 +757,7 @@ static int f2fs_mkdir(struct user_namespace *mnt_userns, struct inode *dir,
 	inode->i_op = &f2fs_dir_inode_operations;
 	inode->i_fop = &f2fs_dir_operations;
 	inode->i_mapping->a_ops = &f2fs_dblock_aops;
-	inode_nohighmem(inode);
+	mapping_set_gfp_mask(inode->i_mapping, GFP_NOFS);
 
 	set_inode_flag(inode, FI_INC_LINK);
 	f2fs_lock_op(sbi);
@@ -803,7 +803,7 @@ static int f2fs_mknod(struct user_namespace *mnt_userns, struct inode *dir,
 	if (!f2fs_is_checkpoint_ready(sbi))
 		return -ENOSPC;
 
-	err = dquot_initialize(dir);
+	err = f2fs_dquot_initialize(dir);
 	if (err)
 		return err;
 
@@ -841,7 +841,7 @@ static int __f2fs_tmpfile(struct inode *dir, struct dentry *dentry,
 	struct inode *inode;
 	int err;
 
-	err = dquot_initialize(dir);
+	err = f2fs_dquot_initialize(dir);
 	if (err)
 		return err;
 
@@ -965,16 +965,16 @@ static int f2fs_rename(struct inode *old_dir, struct dentry *old_dentry,
 			return err;
 	}
 
-	err = dquot_initialize(old_dir);
+	err = f2fs_dquot_initialize(old_dir);
 	if (err)
 		goto out;
 
-	err = dquot_initialize(new_dir);
+	err = f2fs_dquot_initialize(new_dir);
 	if (err)
 		goto out;
 
 	if (new_inode) {
-		err = dquot_initialize(new_inode);
+		err = f2fs_dquot_initialize(new_inode);
 		if (err)
 			goto out;
 	}
@@ -1138,11 +1138,11 @@ static int f2fs_cross_rename(struct inode *old_dir, struct dentry *old_dentry,
 			F2FS_I(new_dentry->d_inode)->i_projid)))
 		return -EXDEV;
 
-	err = dquot_initialize(old_dir);
+	err = f2fs_dquot_initialize(old_dir);
 	if (err)
 		goto out;
 
-	err = dquot_initialize(new_dir);
+	err = f2fs_dquot_initialize(new_dir);
 	if (err)
 		goto out;
 
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index e863136..556fcd8 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -1443,6 +1443,7 @@ static struct page *__get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid,
 			  nid, nid_of_node(page), ino_of_node(page),
 			  ofs_of_node(page), cpver_of_node(page),
 			  next_blkaddr_of_node(page));
+		set_sbi_flag(sbi, SBI_NEED_FSCK);
 		err = -EINVAL;
 out_err:
 		ClearPageUptodate(page);
diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
index ff14a6e..18b98cf 100644
--- a/fs/f2fs/node.h
+++ b/fs/f2fs/node.h
@@ -138,11 +138,6 @@ static inline bool excess_cached_nats(struct f2fs_sb_info *sbi)
 	return NM_I(sbi)->nat_cnt[TOTAL_NAT] >= DEF_NAT_CACHE_THRESHOLD;
 }
 
-static inline bool excess_dirty_nodes(struct f2fs_sb_info *sbi)
-{
-	return get_pages(sbi, F2FS_DIRTY_NODES) >= sbi->blocks_per_seg * 8;
-}
-
 enum mem_type {
 	FREE_NIDS,	/* indicates the free nid list */
 	NAT_ENTRIES,	/* indicates the cached nat entry */
diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c
index 0465551..6a1b466 100644
--- a/fs/f2fs/recovery.c
+++ b/fs/f2fs/recovery.c
@@ -81,7 +81,7 @@ static struct fsync_inode_entry *add_fsync_inode(struct f2fs_sb_info *sbi,
 	if (IS_ERR(inode))
 		return ERR_CAST(inode);
 
-	err = dquot_initialize(inode);
+	err = f2fs_dquot_initialize(inode);
 	if (err)
 		goto err_out;
 
@@ -203,7 +203,7 @@ static int recover_dentry(struct inode *inode, struct page *ipage,
 			goto out_put;
 		}
 
-		err = dquot_initialize(einode);
+		err = f2fs_dquot_initialize(einode);
 		if (err) {
 			iput(einode);
 			goto out_put;
@@ -508,7 +508,7 @@ static int check_index_in_prev_nodes(struct f2fs_sb_info *sbi,
 		if (IS_ERR(inode))
 			return PTR_ERR(inode);
 
-		ret = dquot_initialize(inode);
+		ret = f2fs_dquot_initialize(inode);
 		if (ret) {
 			iput(inode);
 			return ret;
@@ -787,8 +787,6 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only)
 	}
 
 #ifdef CONFIG_QUOTA
-	/* Needed for iput() to work correctly and not trash data */
-	sbi->sb->s_flags |= SB_ACTIVE;
 	/* Turn on quotas so that they are updated correctly */
 	quota_enabled = f2fs_enable_quota_files(sbi, s_flags & SB_RDONLY);
 #endif
@@ -816,10 +814,8 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only)
 	err = recover_data(sbi, &inode_list, &tmp_inode_list, &dir_list);
 	if (!err)
 		f2fs_bug_on(sbi, !list_empty(&inode_list));
-	else {
-		/* restore s_flags to let iput() trash data */
-		sbi->sb->s_flags = s_flags;
-	}
+	else
+		f2fs_bug_on(sbi, sbi->sb->s_flags & SB_ACTIVE);
 skip:
 	fix_curseg_write_pointer = !check_only || list_empty(&inode_list);
 
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index a135d22..df9ed75 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -15,6 +15,7 @@
 #include <linux/timer.h>
 #include <linux/freezer.h>
 #include <linux/sched/signal.h>
+#include <linux/random.h>
 
 #include "f2fs.h"
 #include "segment.h"
@@ -529,6 +530,25 @@ void f2fs_balance_fs(struct f2fs_sb_info *sbi, bool need)
 	}
 }
 
+static inline bool excess_dirty_threshold(struct f2fs_sb_info *sbi)
+{
+	int factor = rwsem_is_locked(&sbi->cp_rwsem) ? 3 : 2;
+	unsigned int dents = get_pages(sbi, F2FS_DIRTY_DENTS);
+	unsigned int qdata = get_pages(sbi, F2FS_DIRTY_QDATA);
+	unsigned int nodes = get_pages(sbi, F2FS_DIRTY_NODES);
+	unsigned int meta = get_pages(sbi, F2FS_DIRTY_META);
+	unsigned int imeta = get_pages(sbi, F2FS_DIRTY_IMETA);
+	unsigned int threshold = sbi->blocks_per_seg * factor *
+					DEFAULT_DIRTY_THRESHOLD;
+	unsigned int global_threshold = threshold * 3 / 2;
+
+	if (dents >= threshold || qdata >= threshold ||
+		nodes >= threshold || meta >= threshold ||
+		imeta >= threshold)
+		return true;
+	return dents + qdata + nodes + meta + imeta >  global_threshold;
+}
+
 void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
 {
 	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
@@ -547,8 +567,8 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
 	else
 		f2fs_build_free_nids(sbi, false, false);
 
-	if (excess_dirty_nats(sbi) || excess_dirty_nodes(sbi) ||
-		excess_prefree_segs(sbi))
+	if (excess_dirty_nats(sbi) || excess_dirty_threshold(sbi) ||
+		excess_prefree_segs(sbi) || !f2fs_space_for_roll_forward(sbi))
 		goto do_sync;
 
 	/* there is background inflight IO or foreground operation recently */
@@ -561,7 +581,7 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
 		goto do_sync;
 
 	/* checkpoint is the only way to shrink partial cached entries */
-	if (f2fs_available_free_memory(sbi, NAT_ENTRIES) ||
+	if (f2fs_available_free_memory(sbi, NAT_ENTRIES) &&
 		f2fs_available_free_memory(sbi, INO_ENTRIES))
 		return;
 
@@ -2630,6 +2650,8 @@ static unsigned int __get_next_segno(struct f2fs_sb_info *sbi, int type)
 	unsigned short seg_type = curseg->seg_type;
 
 	sanity_check_seg_type(sbi, seg_type);
+	if (f2fs_need_rand_seg(sbi))
+		return prandom_u32() % (MAIN_SECS(sbi) * sbi->segs_per_sec);
 
 	/* if segs_per_sec is large than 1, we need to keep original policy. */
 	if (__is_large_section(sbi))
@@ -2681,6 +2703,9 @@ static void new_curseg(struct f2fs_sb_info *sbi, int type, bool new_sec)
 	curseg->next_segno = segno;
 	reset_curseg(sbi, type, 1);
 	curseg->alloc_type = LFS;
+	if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK)
+		curseg->fragment_remained_chunk =
+				prandom_u32() % sbi->max_fragment_chunk + 1;
 }
 
 static int __next_free_blkoff(struct f2fs_sb_info *sbi,
@@ -2707,12 +2732,22 @@ static int __next_free_blkoff(struct f2fs_sb_info *sbi,
 static void __refresh_next_blkoff(struct f2fs_sb_info *sbi,
 				struct curseg_info *seg)
 {
-	if (seg->alloc_type == SSR)
+	if (seg->alloc_type == SSR) {
 		seg->next_blkoff =
 			__next_free_blkoff(sbi, seg->segno,
 						seg->next_blkoff + 1);
-	else
+	} else {
 		seg->next_blkoff++;
+		if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK) {
+			/* To allocate block chunks in different sizes, use random number */
+			if (--seg->fragment_remained_chunk <= 0) {
+				seg->fragment_remained_chunk =
+				   prandom_u32() % sbi->max_fragment_chunk + 1;
+				seg->next_blkoff +=
+				   prandom_u32() % sbi->max_fragment_hole + 1;
+			}
+		}
+	}
 }
 
 bool f2fs_segment_has_free_slot(struct f2fs_sb_info *sbi, int segno)
@@ -3485,24 +3520,30 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
 	up_read(&SM_I(sbi)->curseg_lock);
 }
 
-static void update_device_state(struct f2fs_io_info *fio)
+void f2fs_update_device_state(struct f2fs_sb_info *sbi, nid_t ino,
+					block_t blkaddr, unsigned int blkcnt)
 {
-	struct f2fs_sb_info *sbi = fio->sbi;
-	unsigned int devidx;
-
 	if (!f2fs_is_multi_device(sbi))
 		return;
 
-	devidx = f2fs_target_device_index(sbi, fio->new_blkaddr);
+	while (1) {
+		unsigned int devidx = f2fs_target_device_index(sbi, blkaddr);
+		unsigned int blks = FDEV(devidx).end_blk - blkaddr + 1;
 
-	/* update device state for fsync */
-	f2fs_set_dirty_device(sbi, fio->ino, devidx, FLUSH_INO);
+		/* update device state for fsync */
+		f2fs_set_dirty_device(sbi, ino, devidx, FLUSH_INO);
 
-	/* update device state for checkpoint */
-	if (!f2fs_test_bit(devidx, (char *)&sbi->dirty_device)) {
-		spin_lock(&sbi->dev_lock);
-		f2fs_set_bit(devidx, (char *)&sbi->dirty_device);
-		spin_unlock(&sbi->dev_lock);
+		/* update device state for checkpoint */
+		if (!f2fs_test_bit(devidx, (char *)&sbi->dirty_device)) {
+			spin_lock(&sbi->dev_lock);
+			f2fs_set_bit(devidx, (char *)&sbi->dirty_device);
+			spin_unlock(&sbi->dev_lock);
+		}
+
+		if (blkcnt <= blks)
+			break;
+		blkcnt -= blks;
+		blkaddr += blks;
 	}
 }
 
@@ -3529,7 +3570,7 @@ static void do_write_page(struct f2fs_summary *sum, struct f2fs_io_info *fio)
 		goto reallocate;
 	}
 
-	update_device_state(fio);
+	f2fs_update_device_state(fio->sbi, fio->ino, fio->new_blkaddr, 1);
 
 	if (keep_order)
 		up_read(&fio->sbi->io_order_lock);
@@ -3611,6 +3652,9 @@ int f2fs_inplace_write_data(struct f2fs_io_info *fio)
 		goto drop_bio;
 	}
 
+	invalidate_mapping_pages(META_MAPPING(sbi),
+				fio->new_blkaddr, fio->new_blkaddr);
+
 	stat_inc_inplace_blocks(fio->sbi);
 
 	if (fio->bio && !(SM_I(sbi)->ipu_policy & (1 << F2FS_IPU_NOCACHE)))
@@ -3618,7 +3662,8 @@ int f2fs_inplace_write_data(struct f2fs_io_info *fio)
 	else
 		err = f2fs_submit_page_bio(fio);
 	if (!err) {
-		update_device_state(fio);
+		f2fs_update_device_state(fio->sbi, fio->ino,
+						fio->new_blkaddr, 1);
 		f2fs_update_iostat(fio->sbi, fio->io_type, F2FS_BLKSIZE);
 	}
 
diff --git a/fs/f2fs/segment.h b/fs/f2fs/segment.h
index 89fff25..46fde9f 100644
--- a/fs/f2fs/segment.h
+++ b/fs/f2fs/segment.h
@@ -314,6 +314,7 @@ struct curseg_info {
 	unsigned short next_blkoff;		/* next block offset to write */
 	unsigned int zone;			/* current zone number */
 	unsigned int next_segno;		/* preallocated segment */
+	int fragment_remained_chunk;		/* remained block size in a chunk for block fragmentation mode */
 	bool inited;				/* indicate inmem log is inited */
 };
 
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index cf049a0..040b6d0 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -58,6 +58,7 @@ const char *f2fs_fault_name[FAULT_MAX] = {
 	[FAULT_DISCARD]		= "discard error",
 	[FAULT_WRITE_IO]	= "write IO error",
 	[FAULT_SLAB_ALLOC]	= "slab alloc",
+	[FAULT_DQUOT_INIT]	= "dquot initialize",
 };
 
 void f2fs_build_fault_attr(struct f2fs_sb_info *sbi, unsigned int rate,
@@ -592,7 +593,7 @@ static int f2fs_set_zstd_level(struct f2fs_sb_info *sbi, const char *str)
 	if (kstrtouint(str + 1, 10, &level))
 		return -EINVAL;
 
-	if (!level || level > ZSTD_maxCLevel()) {
+	if (!level || level > zstd_max_clevel()) {
 		f2fs_info(sbi, "invalid zstd compress level: %d", level);
 		return -EINVAL;
 	}
@@ -817,6 +818,10 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
 				F2FS_OPTION(sbi).fs_mode = FS_MODE_ADAPTIVE;
 			} else if (!strcmp(name, "lfs")) {
 				F2FS_OPTION(sbi).fs_mode = FS_MODE_LFS;
+			} else if (!strcmp(name, "fragment:segment")) {
+				F2FS_OPTION(sbi).fs_mode = FS_MODE_FRAGMENT_SEG;
+			} else if (!strcmp(name, "fragment:block")) {
+				F2FS_OPTION(sbi).fs_mode = FS_MODE_FRAGMENT_BLK;
 			} else {
 				kfree(name);
 				return -EINVAL;
@@ -1292,7 +1297,7 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
 	/* Not pass down write hints if the number of active logs is lesser
 	 * than NR_CURSEG_PERSIST_TYPE.
 	 */
-	if (F2FS_OPTION(sbi).active_logs != NR_CURSEG_TYPE)
+	if (F2FS_OPTION(sbi).active_logs != NR_CURSEG_PERSIST_TYPE)
 		F2FS_OPTION(sbi).whint_mode = WHINT_MODE_OFF;
 
 	if (f2fs_sb_has_readonly(sbi) && !f2fs_readonly(sbi->sb)) {
@@ -1896,6 +1901,10 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
 		seq_puts(seq, "adaptive");
 	else if (F2FS_OPTION(sbi).fs_mode == FS_MODE_LFS)
 		seq_puts(seq, "lfs");
+	else if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_SEG)
+		seq_puts(seq, "fragment:segment");
+	else if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK)
+		seq_puts(seq, "fragment:block");
 	seq_printf(seq, ",active_logs=%u", F2FS_OPTION(sbi).active_logs);
 	if (test_opt(sbi, RESERVE_ROOT))
 		seq_printf(seq, ",reserve_root=%u,resuid=%u,resgid=%u",
@@ -2491,6 +2500,16 @@ static ssize_t f2fs_quota_write(struct super_block *sb, int type,
 	return len - towrite;
 }
 
+int f2fs_dquot_initialize(struct inode *inode)
+{
+	if (time_to_inject(F2FS_I_SB(inode), FAULT_DQUOT_INIT)) {
+		f2fs_show_injection_info(F2FS_I_SB(inode), FAULT_DQUOT_INIT);
+		return -ESRCH;
+	}
+
+	return dquot_initialize(inode);
+}
+
 static struct dquot **f2fs_get_dquots(struct inode *inode)
 {
 	return F2FS_I(inode)->i_dquot;
@@ -2875,6 +2894,11 @@ static const struct quotactl_ops f2fs_quotactl_ops = {
 	.get_nextdqblk	= dquot_get_next_dqblk,
 };
 #else
+int f2fs_dquot_initialize(struct inode *inode)
+{
+	return 0;
+}
+
 int f2fs_quota_sync(struct super_block *sb, int type)
 {
 	return 0;
@@ -3486,7 +3510,7 @@ int f2fs_sanity_check_ckpt(struct f2fs_sb_info *sbi)
 		NR_CURSEG_PERSIST_TYPE + nat_bits_blocks >= blocks_per_seg)) {
 		f2fs_warn(sbi, "Insane cp_payload: %u, nat_bits_blocks: %u)",
 			  cp_payload, nat_bits_blocks);
-		return -EFSCORRUPTED;
+		return 1;
 	}
 
 	if (unlikely(f2fs_cp_error(sbi))) {
@@ -3522,6 +3546,8 @@ static void init_sb_info(struct f2fs_sb_info *sbi)
 	sbi->max_victim_search = DEF_MAX_VICTIM_SEARCH;
 	sbi->migration_granularity = sbi->segs_per_sec;
 	sbi->seq_file_ra_mul = MIN_RA_MUL;
+	sbi->max_fragment_chunk = DEF_FRAGMENT_SIZE;
+	sbi->max_fragment_hole = DEF_FRAGMENT_SIZE;
 
 	sbi->dir_level = DEF_DIR_LEVEL;
 	sbi->interval_time[CP_TIME] = DEF_CP_INTERVAL;
@@ -3746,6 +3772,7 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi)
 {
 	struct f2fs_super_block *raw_super = F2FS_RAW_SUPER(sbi);
 	unsigned int max_devices = MAX_DEVICES;
+	unsigned int logical_blksize;
 	int i;
 
 	/* Initialize single device information */
@@ -3766,6 +3793,9 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi)
 	if (!sbi->devs)
 		return -ENOMEM;
 
+	logical_blksize = bdev_logical_block_size(sbi->sb->s_bdev);
+	sbi->aligned_blksize = true;
+
 	for (i = 0; i < max_devices; i++) {
 
 		if (i > 0 && !RDEV(i).path[0])
@@ -3802,6 +3832,9 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi)
 		/* to release errored devices */
 		sbi->s_ndevs = i + 1;
 
+		if (logical_blksize != bdev_logical_block_size(FDEV(i).bdev))
+			sbi->aligned_blksize = false;
+
 #ifdef CONFIG_BLK_DEV_ZONED
 		if (bdev_zoned_model(FDEV(i).bdev) == BLK_ZONED_HM &&
 				!f2fs_sb_has_blkzoned(sbi)) {
@@ -4351,6 +4384,8 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 free_stats:
 	f2fs_destroy_stats(sbi);
 free_nm:
+	/* stop discard thread before destroying node manager */
+	f2fs_stop_discard_thread(sbi);
 	f2fs_destroy_node_manager(sbi);
 free_sm:
 	f2fs_destroy_segment_manager(sbi);
diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
index a32fe31..7d28924 100644
--- a/fs/f2fs/sysfs.c
+++ b/fs/f2fs/sysfs.c
@@ -196,7 +196,7 @@ static ssize_t encoding_show(struct f2fs_attr *a,
 	struct super_block *sb = sbi->sb;
 
 	if (f2fs_sb_has_casefold(sbi))
-		return snprintf(buf, PAGE_SIZE, "%s (%d.%d.%d)\n",
+		return sysfs_emit(buf, "%s (%d.%d.%d)\n",
 			sb->s_encoding->charset,
 			(sb->s_encoding->version >> 16) & 0xff,
 			(sb->s_encoding->version >> 8) & 0xff,
@@ -245,7 +245,7 @@ static ssize_t avg_vblocks_show(struct f2fs_attr *a,
 static ssize_t main_blkaddr_show(struct f2fs_attr *a,
 				struct f2fs_sb_info *sbi, char *buf)
 {
-	return snprintf(buf, PAGE_SIZE, "%llu\n",
+	return sysfs_emit(buf, "%llu\n",
 			(unsigned long long)MAIN_BLKADDR(sbi));
 }
 
@@ -551,6 +551,22 @@ static ssize_t __sbi_store(struct f2fs_attr *a,
 		return count;
 	}
 
+	if (!strcmp(a->attr.name, "max_fragment_chunk")) {
+		if (t >= MIN_FRAGMENT_SIZE && t <= MAX_FRAGMENT_SIZE)
+			sbi->max_fragment_chunk = t;
+		else
+			return -EINVAL;
+		return count;
+	}
+
+	if (!strcmp(a->attr.name, "max_fragment_hole")) {
+		if (t >= MIN_FRAGMENT_SIZE && t <= MAX_FRAGMENT_SIZE)
+			sbi->max_fragment_hole = t;
+		else
+			return -EINVAL;
+		return count;
+	}
+
 	*ui = (unsigned int)t;
 
 	return count;
@@ -781,6 +797,8 @@ F2FS_RW_ATTR(ATGC_INFO, atgc_management, atgc_age_threshold, age_threshold);
 F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, seq_file_ra_mul, seq_file_ra_mul);
 F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, gc_segment_mode, gc_segment_mode);
 F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, gc_reclaimed_segments, gc_reclaimed_segs);
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, max_fragment_chunk, max_fragment_chunk);
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, max_fragment_hole, max_fragment_hole);
 
 #define ATTR_LIST(name) (&f2fs_attr_##name.attr)
 static struct attribute *f2fs_attrs[] = {
@@ -859,6 +877,8 @@ static struct attribute *f2fs_attrs[] = {
 	ATTR_LIST(seq_file_ra_mul),
 	ATTR_LIST(gc_segment_mode),
 	ATTR_LIST(gc_reclaimed_segments),
+	ATTR_LIST(max_fragment_chunk),
+	ATTR_LIST(max_fragment_hole),
 	NULL,
 };
 ATTRIBUTE_GROUPS(f2fs);
diff --git a/fs/f2fs/verity.c b/fs/f2fs/verity.c
index 03549b5..fe5acdc 100644
--- a/fs/f2fs/verity.c
+++ b/fs/f2fs/verity.c
@@ -136,7 +136,7 @@ static int f2fs_begin_enable_verity(struct file *filp)
 	 * here and not rely on ->open() doing it.  This must be done before
 	 * evicting the inline data.
 	 */
-	err = dquot_initialize(inode);
+	err = f2fs_dquot_initialize(inode);
 	if (err)
 		return err;
 
diff --git a/fs/f2fs/xattr.c b/fs/f2fs/xattr.c
index 1d2d29d..e348f33 100644
--- a/fs/f2fs/xattr.c
+++ b/fs/f2fs/xattr.c
@@ -773,7 +773,7 @@ int f2fs_setxattr(struct inode *inode, int index, const char *name,
 	if (!f2fs_is_checkpoint_ready(sbi))
 		return -ENOSPC;
 
-	err = dquot_initialize(inode);
+	err = f2fs_dquot_initialize(inode);
 	if (err)
 		return err;
 
diff --git a/fs/io-wq.c b/fs/io-wq.c
index afd955d..88202de 100644
--- a/fs/io-wq.c
+++ b/fs/io-wq.c
@@ -423,9 +423,10 @@ static inline unsigned int io_get_work_hash(struct io_wq_work *work)
 	return work->flags >> IO_WQ_HASH_SHIFT;
 }
 
-static void io_wait_on_hash(struct io_wqe *wqe, unsigned int hash)
+static bool io_wait_on_hash(struct io_wqe *wqe, unsigned int hash)
 {
 	struct io_wq *wq = wqe->wq;
+	bool ret = false;
 
 	spin_lock_irq(&wq->hash->wait.lock);
 	if (list_empty(&wqe->wait.entry)) {
@@ -433,9 +434,11 @@ static void io_wait_on_hash(struct io_wqe *wqe, unsigned int hash)
 		if (!test_bit(hash, &wq->hash->map)) {
 			__set_current_state(TASK_RUNNING);
 			list_del_init(&wqe->wait.entry);
+			ret = true;
 		}
 	}
 	spin_unlock_irq(&wq->hash->wait.lock);
+	return ret;
 }
 
 static struct io_wq_work *io_get_next_work(struct io_wqe_acct *acct,
@@ -475,14 +478,21 @@ static struct io_wq_work *io_get_next_work(struct io_wqe_acct *acct,
 	}
 
 	if (stall_hash != -1U) {
+		bool unstalled;
+
 		/*
 		 * Set this before dropping the lock to avoid racing with new
 		 * work being added and clearing the stalled bit.
 		 */
 		set_bit(IO_ACCT_STALLED_BIT, &acct->flags);
 		raw_spin_unlock(&wqe->lock);
-		io_wait_on_hash(wqe, stall_hash);
+		unstalled = io_wait_on_hash(wqe, stall_hash);
 		raw_spin_lock(&wqe->lock);
+		if (unstalled) {
+			clear_bit(IO_ACCT_STALLED_BIT, &acct->flags);
+			if (wq_has_sleeper(&wqe->wq->hash->wait))
+				wake_up(&wqe->wq->hash->wait);
+		}
 	}
 
 	return NULL;
@@ -564,8 +574,11 @@ static void io_worker_handle_work(struct io_worker *worker)
 				io_wqe_enqueue(wqe, linked);
 
 			if (hash != -1U && !next_hashed) {
+				/* serialize hash clear with wake_up() */
+				spin_lock_irq(&wq->hash->wait.lock);
 				clear_bit(hash, &wq->hash->map);
 				clear_bit(IO_ACCT_STALLED_BIT, &acct->flags);
+				spin_unlock_irq(&wq->hash->wait.lock);
 				if (wq_has_sleeper(&wq->hash->wait))
 					wake_up(&wq->hash->wait);
 				raw_spin_lock(&wqe->lock);
diff --git a/fs/ksmbd/Kconfig b/fs/ksmbd/Kconfig
index b83cbd7..e1fe177 100644
--- a/fs/ksmbd/Kconfig
+++ b/fs/ksmbd/Kconfig
@@ -6,7 +6,6 @@
 	select NLS
 	select NLS_UTF8
 	select CRYPTO
-	select CRYPTO_MD4
 	select CRYPTO_MD5
 	select CRYPTO_HMAC
 	select CRYPTO_ECB
@@ -19,6 +18,7 @@
 	select CRYPTO_GCM
 	select ASN1
 	select OID_REGISTRY
+	select CRC32
 	default n
 	help
 	  Choose Y here if you want to allow SMB3 compliant clients
diff --git a/fs/ksmbd/auth.c b/fs/ksmbd/auth.c
index 30a92dd..3503b1c 100644
--- a/fs/ksmbd/auth.c
+++ b/fs/ksmbd/auth.c
@@ -873,9 +873,9 @@ int ksmbd_gen_preauth_integrity_hash(struct ksmbd_conn *conn, char *buf,
 				     __u8 *pi_hash)
 {
 	int rc;
-	struct smb2_hdr *rcv_hdr = (struct smb2_hdr *)buf;
+	struct smb2_hdr *rcv_hdr = smb2_get_msg(buf);
 	char *all_bytes_msg = (char *)&rcv_hdr->ProtocolId;
-	int msg_size = be32_to_cpu(rcv_hdr->smb2_buf_length);
+	int msg_size = get_rfc1002_len(buf);
 	struct ksmbd_crypto_ctx *ctx = NULL;
 
 	if (conn->preauth_info->Preauth_HashId !=
@@ -983,7 +983,7 @@ static struct scatterlist *ksmbd_init_sg(struct kvec *iov, unsigned int nvec,
 					 u8 *sign)
 {
 	struct scatterlist *sg;
-	unsigned int assoc_data_len = sizeof(struct smb2_transform_hdr) - 24;
+	unsigned int assoc_data_len = sizeof(struct smb2_transform_hdr) - 20;
 	int i, nr_entries[3] = {0}, total_entries = 0, sg_idx = 0;
 
 	if (!nvec)
@@ -1047,9 +1047,8 @@ static struct scatterlist *ksmbd_init_sg(struct kvec *iov, unsigned int nvec,
 int ksmbd_crypt_message(struct ksmbd_conn *conn, struct kvec *iov,
 			unsigned int nvec, int enc)
 {
-	struct smb2_transform_hdr *tr_hdr =
-		(struct smb2_transform_hdr *)iov[0].iov_base;
-	unsigned int assoc_data_len = sizeof(struct smb2_transform_hdr) - 24;
+	struct smb2_transform_hdr *tr_hdr = smb2_get_msg(iov[0].iov_base);
+	unsigned int assoc_data_len = sizeof(struct smb2_transform_hdr) - 20;
 	int rc;
 	struct scatterlist *sg;
 	u8 sign[SMB2_SIGNATURE_SIZE] = {};
diff --git a/fs/ksmbd/connection.c b/fs/ksmbd/connection.c
index b57a0d8..83a94d0 100644
--- a/fs/ksmbd/connection.c
+++ b/fs/ksmbd/connection.c
@@ -158,26 +158,25 @@ void ksmbd_conn_wait_idle(struct ksmbd_conn *conn)
 int ksmbd_conn_write(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
-	struct smb_hdr *rsp_hdr = work->response_buf;
 	size_t len = 0;
 	int sent;
 	struct kvec iov[3];
 	int iov_idx = 0;
 
 	ksmbd_conn_try_dequeue_request(work);
-	if (!rsp_hdr) {
+	if (!work->response_buf) {
 		pr_err("NULL response header\n");
 		return -EINVAL;
 	}
 
 	if (work->tr_buf) {
 		iov[iov_idx] = (struct kvec) { work->tr_buf,
-				sizeof(struct smb2_transform_hdr) };
+				sizeof(struct smb2_transform_hdr) + 4 };
 		len += iov[iov_idx++].iov_len;
 	}
 
 	if (work->aux_payload_sz) {
-		iov[iov_idx] = (struct kvec) { rsp_hdr, work->resp_hdr_sz };
+		iov[iov_idx] = (struct kvec) { work->response_buf, work->resp_hdr_sz };
 		len += iov[iov_idx++].iov_len;
 		iov[iov_idx] = (struct kvec) { work->aux_payload_buf, work->aux_payload_sz };
 		len += iov[iov_idx++].iov_len;
@@ -185,8 +184,8 @@ int ksmbd_conn_write(struct ksmbd_work *work)
 		if (work->tr_buf)
 			iov[iov_idx].iov_len = work->resp_hdr_sz;
 		else
-			iov[iov_idx].iov_len = get_rfc1002_len(rsp_hdr) + 4;
-		iov[iov_idx].iov_base = rsp_hdr;
+			iov[iov_idx].iov_len = get_rfc1002_len(work->response_buf) + 4;
+		iov[iov_idx].iov_base = work->response_buf;
 		len += iov[iov_idx++].iov_len;
 	}
 
diff --git a/fs/ksmbd/ksmbd_work.c b/fs/ksmbd/ksmbd_work.c
index fd58eb4..14b9cae 100644
--- a/fs/ksmbd/ksmbd_work.c
+++ b/fs/ksmbd/ksmbd_work.c
@@ -69,7 +69,6 @@ int ksmbd_workqueue_init(void)
 
 void ksmbd_workqueue_destroy(void)
 {
-	flush_workqueue(ksmbd_wq);
 	destroy_workqueue(ksmbd_wq);
 	ksmbd_wq = NULL;
 }
diff --git a/fs/ksmbd/ksmbd_work.h b/fs/ksmbd/ksmbd_work.h
index f7156bc..5ece58e 100644
--- a/fs/ksmbd/ksmbd_work.h
+++ b/fs/ksmbd/ksmbd_work.h
@@ -92,7 +92,7 @@ struct ksmbd_work {
  */
 static inline void *ksmbd_resp_buf_next(struct ksmbd_work *work)
 {
-	return work->response_buf + work->next_smb2_rsp_hdr_off;
+	return work->response_buf + work->next_smb2_rsp_hdr_off + 4;
 }
 
 /**
@@ -101,7 +101,7 @@ static inline void *ksmbd_resp_buf_next(struct ksmbd_work *work)
  */
 static inline void *ksmbd_req_buf_next(struct ksmbd_work *work)
 {
-	return work->request_buf + work->next_smb2_rcv_hdr_off;
+	return work->request_buf + work->next_smb2_rcv_hdr_off + 4;
 }
 
 struct ksmbd_work *ksmbd_alloc_work_struct(void);
diff --git a/fs/ksmbd/oplock.c b/fs/ksmbd/oplock.c
index f9dae6e..077b8761 100644
--- a/fs/ksmbd/oplock.c
+++ b/fs/ksmbd/oplock.c
@@ -629,10 +629,10 @@ static void __smb2_oplock_break_noti(struct work_struct *wk)
 		return;
 	}
 
-	rsp_hdr = work->response_buf;
+	rsp_hdr = smb2_get_msg(work->response_buf);
 	memset(rsp_hdr, 0, sizeof(struct smb2_hdr) + 2);
-	rsp_hdr->smb2_buf_length =
-		cpu_to_be32(smb2_hdr_size_no_buflen(conn->vals));
+	*(__be32 *)work->response_buf =
+		cpu_to_be32(conn->vals->header_size);
 	rsp_hdr->ProtocolId = SMB2_PROTO_NUMBER;
 	rsp_hdr->StructureSize = SMB2_HEADER_STRUCTURE_SIZE;
 	rsp_hdr->CreditRequest = cpu_to_le16(0);
@@ -645,7 +645,7 @@ static void __smb2_oplock_break_noti(struct work_struct *wk)
 	rsp_hdr->SessionId = 0;
 	memset(rsp_hdr->Signature, 0, 16);
 
-	rsp = work->response_buf;
+	rsp = smb2_get_msg(work->response_buf);
 
 	rsp->StructureSize = cpu_to_le16(24);
 	if (!br_info->open_trunc &&
@@ -659,7 +659,7 @@ static void __smb2_oplock_break_noti(struct work_struct *wk)
 	rsp->PersistentFid = cpu_to_le64(fp->persistent_id);
 	rsp->VolatileFid = cpu_to_le64(fp->volatile_id);
 
-	inc_rfc1001_len(rsp, 24);
+	inc_rfc1001_len(work->response_buf, 24);
 
 	ksmbd_debug(OPLOCK,
 		    "sending oplock break v_id %llu p_id = %llu lock level = %d\n",
@@ -736,10 +736,10 @@ static void __smb2_lease_break_noti(struct work_struct *wk)
 		return;
 	}
 
-	rsp_hdr = work->response_buf;
+	rsp_hdr = smb2_get_msg(work->response_buf);
 	memset(rsp_hdr, 0, sizeof(struct smb2_hdr) + 2);
-	rsp_hdr->smb2_buf_length =
-		cpu_to_be32(smb2_hdr_size_no_buflen(conn->vals));
+	*(__be32 *)work->response_buf =
+		cpu_to_be32(conn->vals->header_size);
 	rsp_hdr->ProtocolId = SMB2_PROTO_NUMBER;
 	rsp_hdr->StructureSize = SMB2_HEADER_STRUCTURE_SIZE;
 	rsp_hdr->CreditRequest = cpu_to_le16(0);
@@ -752,7 +752,7 @@ static void __smb2_lease_break_noti(struct work_struct *wk)
 	rsp_hdr->SessionId = 0;
 	memset(rsp_hdr->Signature, 0, 16);
 
-	rsp = work->response_buf;
+	rsp = smb2_get_msg(work->response_buf);
 	rsp->StructureSize = cpu_to_le16(44);
 	rsp->Epoch = br_info->epoch;
 	rsp->Flags = 0;
@@ -768,7 +768,7 @@ static void __smb2_lease_break_noti(struct work_struct *wk)
 	rsp->AccessMaskHint = 0;
 	rsp->ShareMaskHint = 0;
 
-	inc_rfc1001_len(rsp, 44);
+	inc_rfc1001_len(work->response_buf, 44);
 
 	ksmbd_conn_write(work);
 	ksmbd_free_work_struct(work);
@@ -1335,19 +1335,16 @@ __u8 smb2_map_lease_to_oplock(__le32 lease_state)
  */
 void create_lease_buf(u8 *rbuf, struct lease *lease)
 {
-	char *LeaseKey = (char *)&lease->lease_key;
-
 	if (lease->version == 2) {
 		struct create_lease_v2 *buf = (struct create_lease_v2 *)rbuf;
-		char *ParentLeaseKey = (char *)&lease->parent_lease_key;
 
 		memset(buf, 0, sizeof(struct create_lease_v2));
-		buf->lcontext.LeaseKeyLow = *((__le64 *)LeaseKey);
-		buf->lcontext.LeaseKeyHigh = *((__le64 *)(LeaseKey + 8));
+		memcpy(buf->lcontext.LeaseKey, lease->lease_key,
+		       SMB2_LEASE_KEY_SIZE);
 		buf->lcontext.LeaseFlags = lease->flags;
 		buf->lcontext.LeaseState = lease->state;
-		buf->lcontext.ParentLeaseKeyLow = *((__le64 *)ParentLeaseKey);
-		buf->lcontext.ParentLeaseKeyHigh = *((__le64 *)(ParentLeaseKey + 8));
+		memcpy(buf->lcontext.ParentLeaseKey, lease->parent_lease_key,
+		       SMB2_LEASE_KEY_SIZE);
 		buf->ccontext.DataOffset = cpu_to_le16(offsetof
 				(struct create_lease_v2, lcontext));
 		buf->ccontext.DataLength = cpu_to_le32(sizeof(struct lease_context_v2));
@@ -1362,8 +1359,7 @@ void create_lease_buf(u8 *rbuf, struct lease *lease)
 		struct create_lease *buf = (struct create_lease *)rbuf;
 
 		memset(buf, 0, sizeof(struct create_lease));
-		buf->lcontext.LeaseKeyLow = *((__le64 *)LeaseKey);
-		buf->lcontext.LeaseKeyHigh = *((__le64 *)(LeaseKey + 8));
+		memcpy(buf->lcontext.LeaseKey, lease->lease_key, SMB2_LEASE_KEY_SIZE);
 		buf->lcontext.LeaseFlags = lease->flags;
 		buf->lcontext.LeaseState = lease->state;
 		buf->ccontext.DataOffset = cpu_to_le16(offsetof
@@ -1398,7 +1394,7 @@ struct lease_ctx_info *parse_lease_state(void *open_req)
 	if (!lreq)
 		return NULL;
 
-	data_offset = (char *)req + 4 + le32_to_cpu(req->CreateContextsOffset);
+	data_offset = (char *)req + le32_to_cpu(req->CreateContextsOffset);
 	cc = (struct create_context *)data_offset;
 	do {
 		cc = (struct create_context *)((char *)cc + next);
@@ -1416,19 +1412,17 @@ struct lease_ctx_info *parse_lease_state(void *open_req)
 		if (sizeof(struct lease_context_v2) == le32_to_cpu(cc->DataLength)) {
 			struct create_lease_v2 *lc = (struct create_lease_v2 *)cc;
 
-			*((__le64 *)lreq->lease_key) = lc->lcontext.LeaseKeyLow;
-			*((__le64 *)(lreq->lease_key + 8)) = lc->lcontext.LeaseKeyHigh;
+			memcpy(lreq->lease_key, lc->lcontext.LeaseKey, SMB2_LEASE_KEY_SIZE);
 			lreq->req_state = lc->lcontext.LeaseState;
 			lreq->flags = lc->lcontext.LeaseFlags;
 			lreq->duration = lc->lcontext.LeaseDuration;
-			*((__le64 *)lreq->parent_lease_key) = lc->lcontext.ParentLeaseKeyLow;
-			*((__le64 *)(lreq->parent_lease_key + 8)) = lc->lcontext.ParentLeaseKeyHigh;
+			memcpy(lreq->parent_lease_key, lc->lcontext.ParentLeaseKey,
+			       SMB2_LEASE_KEY_SIZE);
 			lreq->version = 2;
 		} else {
 			struct create_lease *lc = (struct create_lease *)cc;
 
-			*((__le64 *)lreq->lease_key) = lc->lcontext.LeaseKeyLow;
-			*((__le64 *)(lreq->lease_key + 8)) = lc->lcontext.LeaseKeyHigh;
+			memcpy(lreq->lease_key, lc->lcontext.LeaseKey, SMB2_LEASE_KEY_SIZE);
 			lreq->req_state = lc->lcontext.LeaseState;
 			lreq->flags = lc->lcontext.LeaseFlags;
 			lreq->duration = lc->lcontext.LeaseDuration;
@@ -1462,7 +1456,7 @@ struct create_context *smb2_find_context_vals(void *open_req, const char *tag)
 	 * CreateContextsOffset and CreateContextsLength are guaranteed to
 	 * be valid because of ksmbd_smb2_check_message().
 	 */
-	cc = (struct create_context *)((char *)req + 4 +
+	cc = (struct create_context *)((char *)req +
 				       le32_to_cpu(req->CreateContextsOffset));
 	remain_len = le32_to_cpu(req->CreateContextsLength);
 	do {
diff --git a/fs/ksmbd/oplock.h b/fs/ksmbd/oplock.h
index 119b804..0cf7a2b 100644
--- a/fs/ksmbd/oplock.h
+++ b/fs/ksmbd/oplock.h
@@ -28,8 +28,6 @@
 #define OPLOCK_WRITE_TO_NONE		0x04
 #define OPLOCK_READ_TO_NONE		0x08
 
-#define SMB2_LEASE_KEY_SIZE		16
-
 struct lease_ctx_info {
 	__u8			lease_key[SMB2_LEASE_KEY_SIZE];
 	__le32			req_state;
diff --git a/fs/ksmbd/server.c b/fs/ksmbd/server.c
index 2a2b213..2e12f6d 100644
--- a/fs/ksmbd/server.c
+++ b/fs/ksmbd/server.c
@@ -622,7 +622,6 @@ MODULE_DESCRIPTION("Linux kernel CIFS/SMB SERVER");
 MODULE_LICENSE("GPL");
 MODULE_SOFTDEP("pre: ecb");
 MODULE_SOFTDEP("pre: hmac");
-MODULE_SOFTDEP("pre: md4");
 MODULE_SOFTDEP("pre: md5");
 MODULE_SOFTDEP("pre: nls");
 MODULE_SOFTDEP("pre: aes");
@@ -632,5 +631,6 @@ MODULE_SOFTDEP("pre: sha512");
 MODULE_SOFTDEP("pre: aead2");
 MODULE_SOFTDEP("pre: ccm");
 MODULE_SOFTDEP("pre: gcm");
+MODULE_SOFTDEP("pre: crc32");
 module_init(ksmbd_server_init)
 module_exit(ksmbd_server_exit)
diff --git a/fs/ksmbd/smb2misc.c b/fs/ksmbd/smb2misc.c
index 030ca57..50d0b10 100644
--- a/fs/ksmbd/smb2misc.c
+++ b/fs/ksmbd/smb2misc.c
@@ -6,7 +6,6 @@
 
 #include "glob.h"
 #include "nterr.h"
-#include "smb2pdu.h"
 #include "smb_common.h"
 #include "smbstatus.h"
 #include "mgmt/user_session.h"
@@ -347,23 +346,16 @@ static int smb2_validate_credit_charge(struct ksmbd_conn *conn,
 
 int ksmbd_smb2_check_message(struct ksmbd_work *work)
 {
-	struct smb2_pdu *pdu = work->request_buf;
+	struct smb2_pdu *pdu = ksmbd_req_buf_next(work);
 	struct smb2_hdr *hdr = &pdu->hdr;
 	int command;
 	__u32 clc_len;  /* calculated length */
-	__u32 len = get_rfc1002_len(pdu);
+	__u32 len = get_rfc1002_len(work->request_buf);
 
-	if (work->next_smb2_rcv_hdr_off) {
-		pdu = ksmbd_req_buf_next(work);
-		hdr = &pdu->hdr;
-	}
-
-	if (le32_to_cpu(hdr->NextCommand) > 0) {
+	if (le32_to_cpu(hdr->NextCommand) > 0)
 		len = le32_to_cpu(hdr->NextCommand);
-	} else if (work->next_smb2_rcv_hdr_off) {
+	else if (work->next_smb2_rcv_hdr_off)
 		len -= work->next_smb2_rcv_hdr_off;
-		len = round_up(len, 8);
-	}
 
 	if (check_smb2_hdr(hdr))
 		return 1;
diff --git a/fs/ksmbd/smb2ops.c b/fs/ksmbd/smb2ops.c
index fb6a65d..0a5d845 100644
--- a/fs/ksmbd/smb2ops.c
+++ b/fs/ksmbd/smb2ops.c
@@ -6,7 +6,6 @@
 
 #include <linux/slab.h>
 #include "glob.h"
-#include "smb2pdu.h"
 
 #include "auth.h"
 #include "connection.h"
@@ -199,7 +198,7 @@ void init_smb2_1_server(struct ksmbd_conn *conn)
 	conn->cmds = smb2_0_server_cmds;
 	conn->max_cmds = ARRAY_SIZE(smb2_0_server_cmds);
 	conn->max_credits = SMB2_MAX_CREDITS;
-	conn->signing_algorithm = SIGNING_ALG_HMAC_SHA256;
+	conn->signing_algorithm = SIGNING_ALG_HMAC_SHA256_LE;
 
 	if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_LEASES)
 		conn->vals->capabilities |= SMB2_GLOBAL_CAP_LEASING;
@@ -217,7 +216,7 @@ void init_smb3_0_server(struct ksmbd_conn *conn)
 	conn->cmds = smb2_0_server_cmds;
 	conn->max_cmds = ARRAY_SIZE(smb2_0_server_cmds);
 	conn->max_credits = SMB2_MAX_CREDITS;
-	conn->signing_algorithm = SIGNING_ALG_AES_CMAC;
+	conn->signing_algorithm = SIGNING_ALG_AES_CMAC_LE;
 
 	if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_LEASES)
 		conn->vals->capabilities |= SMB2_GLOBAL_CAP_LEASING;
@@ -242,7 +241,7 @@ void init_smb3_02_server(struct ksmbd_conn *conn)
 	conn->cmds = smb2_0_server_cmds;
 	conn->max_cmds = ARRAY_SIZE(smb2_0_server_cmds);
 	conn->max_credits = SMB2_MAX_CREDITS;
-	conn->signing_algorithm = SIGNING_ALG_AES_CMAC;
+	conn->signing_algorithm = SIGNING_ALG_AES_CMAC_LE;
 
 	if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_LEASES)
 		conn->vals->capabilities |= SMB2_GLOBAL_CAP_LEASING;
@@ -267,7 +266,7 @@ int init_smb3_11_server(struct ksmbd_conn *conn)
 	conn->cmds = smb2_0_server_cmds;
 	conn->max_cmds = ARRAY_SIZE(smb2_0_server_cmds);
 	conn->max_credits = SMB2_MAX_CREDITS;
-	conn->signing_algorithm = SIGNING_ALG_AES_CMAC;
+	conn->signing_algorithm = SIGNING_ALG_AES_CMAC_LE;
 
 	if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB2_LEASES)
 		conn->vals->capabilities |= SMB2_GLOBAL_CAP_LEASING;
diff --git a/fs/ksmbd/smb2pdu.c b/fs/ksmbd/smb2pdu.c
index 7e448df..121f8e8 100644
--- a/fs/ksmbd/smb2pdu.c
+++ b/fs/ksmbd/smb2pdu.c
@@ -13,7 +13,6 @@
 #include <linux/falloc.h>
 
 #include "glob.h"
-#include "smb2pdu.h"
 #include "smbfsctl.h"
 #include "oplock.h"
 #include "smbacl.h"
@@ -44,8 +43,8 @@ static void __wbuf(struct ksmbd_work *work, void **req, void **rsp)
 		*req = ksmbd_req_buf_next(work);
 		*rsp = ksmbd_resp_buf_next(work);
 	} else {
-		*req = work->request_buf;
-		*rsp = work->response_buf;
+		*req = smb2_get_msg(work->request_buf);
+		*rsp = smb2_get_msg(work->response_buf);
 	}
 }
 
@@ -93,13 +92,14 @@ struct channel *lookup_chann_list(struct ksmbd_session *sess, struct ksmbd_conn
  */
 int smb2_get_ksmbd_tcon(struct ksmbd_work *work)
 {
-	struct smb2_hdr *req_hdr = work->request_buf;
+	struct smb2_hdr *req_hdr = smb2_get_msg(work->request_buf);
+	unsigned int cmd = le16_to_cpu(req_hdr->Command);
 	int tree_id;
 
 	work->tcon = NULL;
-	if (work->conn->ops->get_cmd_val(work) == SMB2_TREE_CONNECT_HE ||
-	    work->conn->ops->get_cmd_val(work) ==  SMB2_CANCEL_HE ||
-	    work->conn->ops->get_cmd_val(work) ==  SMB2_LOGOFF_HE) {
+	if (cmd == SMB2_TREE_CONNECT_HE ||
+	    cmd ==  SMB2_CANCEL_HE ||
+	    cmd ==  SMB2_LOGOFF_HE) {
 		ksmbd_debug(SMB, "skip to check tree connect request\n");
 		return 0;
 	}
@@ -130,7 +130,7 @@ void smb2_set_err_rsp(struct ksmbd_work *work)
 	if (work->next_smb2_rcv_hdr_off)
 		err_rsp = ksmbd_resp_buf_next(work);
 	else
-		err_rsp = work->response_buf;
+		err_rsp = smb2_get_msg(work->response_buf);
 
 	if (err_rsp->hdr.Status != STATUS_STOPPED_ON_SYMLINK) {
 		err_rsp->StructureSize = SMB2_ERROR_STRUCTURE_SIZE2_LE;
@@ -150,7 +150,7 @@ void smb2_set_err_rsp(struct ksmbd_work *work)
  */
 bool is_smb2_neg_cmd(struct ksmbd_work *work)
 {
-	struct smb2_hdr *hdr = work->request_buf;
+	struct smb2_hdr *hdr = smb2_get_msg(work->request_buf);
 
 	/* is it SMB2 header ? */
 	if (hdr->ProtocolId != SMB2_PROTO_NUMBER)
@@ -174,7 +174,7 @@ bool is_smb2_neg_cmd(struct ksmbd_work *work)
  */
 bool is_smb2_rsp(struct ksmbd_work *work)
 {
-	struct smb2_hdr *hdr = work->response_buf;
+	struct smb2_hdr *hdr = smb2_get_msg(work->response_buf);
 
 	/* is it SMB2 header ? */
 	if (hdr->ProtocolId != SMB2_PROTO_NUMBER)
@@ -200,7 +200,7 @@ u16 get_smb2_cmd_val(struct ksmbd_work *work)
 	if (work->next_smb2_rcv_hdr_off)
 		rcv_hdr = ksmbd_req_buf_next(work);
 	else
-		rcv_hdr = work->request_buf;
+		rcv_hdr = smb2_get_msg(work->request_buf);
 	return le16_to_cpu(rcv_hdr->Command);
 }
 
@@ -216,7 +216,7 @@ void set_smb2_rsp_status(struct ksmbd_work *work, __le32 err)
 	if (work->next_smb2_rcv_hdr_off)
 		rsp_hdr = ksmbd_resp_buf_next(work);
 	else
-		rsp_hdr = work->response_buf;
+		rsp_hdr = smb2_get_msg(work->response_buf);
 	rsp_hdr->Status = err;
 	smb2_set_err_rsp(work);
 }
@@ -237,13 +237,11 @@ int init_smb2_neg_rsp(struct ksmbd_work *work)
 	if (conn->need_neg == false)
 		return -EINVAL;
 
-	rsp_hdr = work->response_buf;
+	*(__be32 *)work->response_buf =
+		cpu_to_be32(conn->vals->header_size);
 
+	rsp_hdr = smb2_get_msg(work->response_buf);
 	memset(rsp_hdr, 0, sizeof(struct smb2_hdr) + 2);
-
-	rsp_hdr->smb2_buf_length =
-		cpu_to_be32(smb2_hdr_size_no_buflen(conn->vals));
-
 	rsp_hdr->ProtocolId = SMB2_PROTO_NUMBER;
 	rsp_hdr->StructureSize = SMB2_HEADER_STRUCTURE_SIZE;
 	rsp_hdr->CreditRequest = cpu_to_le16(2);
@@ -256,7 +254,7 @@ int init_smb2_neg_rsp(struct ksmbd_work *work)
 	rsp_hdr->SessionId = 0;
 	memset(rsp_hdr->Signature, 0, 16);
 
-	rsp = work->response_buf;
+	rsp = smb2_get_msg(work->response_buf);
 
 	WARN_ON(ksmbd_conn_good(work));
 
@@ -277,12 +275,12 @@ int init_smb2_neg_rsp(struct ksmbd_work *work)
 
 	rsp->SecurityBufferOffset = cpu_to_le16(128);
 	rsp->SecurityBufferLength = cpu_to_le16(AUTH_GSS_LENGTH);
-	ksmbd_copy_gss_neg_header(((char *)(&rsp->hdr) +
-		sizeof(rsp->hdr.smb2_buf_length)) +
+	ksmbd_copy_gss_neg_header((char *)(&rsp->hdr) +
 		le16_to_cpu(rsp->SecurityBufferOffset));
-	inc_rfc1001_len(rsp, sizeof(struct smb2_negotiate_rsp) -
-		sizeof(struct smb2_hdr) - sizeof(rsp->Buffer) +
-		AUTH_GSS_LENGTH);
+	inc_rfc1001_len(work->response_buf,
+			sizeof(struct smb2_negotiate_rsp) -
+			sizeof(struct smb2_hdr) - sizeof(rsp->Buffer) +
+			AUTH_GSS_LENGTH);
 	rsp->SecurityMode = SMB2_NEGOTIATE_SIGNING_ENABLED_LE;
 	if (server_conf.signing == KSMBD_CONFIG_OPT_MANDATORY)
 		rsp->SecurityMode |= SMB2_NEGOTIATE_SIGNING_REQUIRED_LE;
@@ -387,8 +385,8 @@ static void init_chained_smb2_rsp(struct ksmbd_work *work)
 	next_hdr_offset = le32_to_cpu(req->NextCommand);
 
 	new_len = ALIGN(len, 8);
-	inc_rfc1001_len(work->response_buf, ((sizeof(struct smb2_hdr) - 4)
-			+ new_len - len));
+	inc_rfc1001_len(work->response_buf,
+			sizeof(struct smb2_hdr) + new_len - len);
 	rsp->NextCommand = cpu_to_le32(new_len);
 
 	work->next_smb2_rcv_hdr_off += next_hdr_offset;
@@ -406,7 +404,7 @@ static void init_chained_smb2_rsp(struct ksmbd_work *work)
 		work->compound_fid = KSMBD_NO_FID;
 		work->compound_pfid = KSMBD_NO_FID;
 	}
-	memset((char *)rsp_hdr + 4, 0, sizeof(struct smb2_hdr) + 2);
+	memset((char *)rsp_hdr, 0, sizeof(struct smb2_hdr) + 2);
 	rsp_hdr->ProtocolId = SMB2_PROTO_NUMBER;
 	rsp_hdr->StructureSize = SMB2_HEADER_STRUCTURE_SIZE;
 	rsp_hdr->Command = rcv_hdr->Command;
@@ -432,7 +430,7 @@ static void init_chained_smb2_rsp(struct ksmbd_work *work)
  */
 bool is_chained_smb2_message(struct ksmbd_work *work)
 {
-	struct smb2_hdr *hdr = work->request_buf;
+	struct smb2_hdr *hdr = smb2_get_msg(work->request_buf);
 	unsigned int len, next_cmd;
 
 	if (hdr->ProtocolId != SMB2_PROTO_NUMBER)
@@ -483,13 +481,13 @@ bool is_chained_smb2_message(struct ksmbd_work *work)
  */
 int init_smb2_rsp_hdr(struct ksmbd_work *work)
 {
-	struct smb2_hdr *rsp_hdr = work->response_buf;
-	struct smb2_hdr *rcv_hdr = work->request_buf;
+	struct smb2_hdr *rsp_hdr = smb2_get_msg(work->response_buf);
+	struct smb2_hdr *rcv_hdr = smb2_get_msg(work->request_buf);
 	struct ksmbd_conn *conn = work->conn;
 
 	memset(rsp_hdr, 0, sizeof(struct smb2_hdr) + 2);
-	rsp_hdr->smb2_buf_length =
-		cpu_to_be32(smb2_hdr_size_no_buflen(conn->vals));
+	*(__be32 *)work->response_buf =
+		cpu_to_be32(conn->vals->header_size);
 	rsp_hdr->ProtocolId = rcv_hdr->ProtocolId;
 	rsp_hdr->StructureSize = SMB2_HEADER_STRUCTURE_SIZE;
 	rsp_hdr->Command = rcv_hdr->Command;
@@ -522,7 +520,7 @@ int init_smb2_rsp_hdr(struct ksmbd_work *work)
  */
 int smb2_allocate_rsp_buf(struct ksmbd_work *work)
 {
-	struct smb2_hdr *hdr = work->request_buf;
+	struct smb2_hdr *hdr = smb2_get_msg(work->request_buf);
 	size_t small_sz = MAX_CIFS_SMALL_BUFFER_SIZE;
 	size_t large_sz = small_sz + work->conn->vals->max_trans_size;
 	size_t sz = small_sz;
@@ -534,7 +532,7 @@ int smb2_allocate_rsp_buf(struct ksmbd_work *work)
 	if (cmd == SMB2_QUERY_INFO_HE) {
 		struct smb2_query_info_req *req;
 
-		req = work->request_buf;
+		req = smb2_get_msg(work->request_buf);
 		if (req->InfoType == SMB2_O_INFO_FILE &&
 		    (req->FileInfoClass == FILE_FULL_EA_INFORMATION ||
 		     req->FileInfoClass == FILE_ALL_INFORMATION))
@@ -561,7 +559,7 @@ int smb2_allocate_rsp_buf(struct ksmbd_work *work)
  */
 int smb2_check_user_session(struct ksmbd_work *work)
 {
-	struct smb2_hdr *req_hdr = work->request_buf;
+	struct smb2_hdr *req_hdr = smb2_get_msg(work->request_buf);
 	struct ksmbd_conn *conn = work->conn;
 	unsigned int cmd = conn->ops->get_cmd_val(work);
 	unsigned long long sess_id;
@@ -642,7 +640,7 @@ int setup_async_work(struct ksmbd_work *work, void (*fn)(void **), void **arg)
 	struct ksmbd_conn *conn = work->conn;
 	int id;
 
-	rsp_hdr = work->response_buf;
+	rsp_hdr = smb2_get_msg(work->response_buf);
 	rsp_hdr->Flags |= SMB2_FLAGS_ASYNC_COMMAND;
 
 	id = ksmbd_acquire_async_msg_id(&conn->async_ida);
@@ -674,7 +672,7 @@ void smb2_send_interim_resp(struct ksmbd_work *work, __le32 status)
 {
 	struct smb2_hdr *rsp_hdr;
 
-	rsp_hdr = work->response_buf;
+	rsp_hdr = smb2_get_msg(work->response_buf);
 	smb2_set_err_rsp(work);
 	rsp_hdr->Status = status;
 
@@ -715,17 +713,17 @@ static int smb2_get_dos_mode(struct kstat *stat, int attribute)
 	int attr = 0;
 
 	if (S_ISDIR(stat->mode)) {
-		attr = ATTR_DIRECTORY |
-			(attribute & (ATTR_HIDDEN | ATTR_SYSTEM));
+		attr = FILE_ATTRIBUTE_DIRECTORY |
+			(attribute & (FILE_ATTRIBUTE_HIDDEN | FILE_ATTRIBUTE_SYSTEM));
 	} else {
-		attr = (attribute & 0x00005137) | ATTR_ARCHIVE;
-		attr &= ~(ATTR_DIRECTORY);
+		attr = (attribute & 0x00005137) | FILE_ATTRIBUTE_ARCHIVE;
+		attr &= ~(FILE_ATTRIBUTE_DIRECTORY);
 		if (S_ISREG(stat->mode) && (server_conf.share_fake_fscaps &
 				FILE_SUPPORTS_SPARSE_FILES))
-			attr |= ATTR_SPARSE;
+			attr |= FILE_ATTRIBUTE_SPARSE_FILE;
 
 		if (smb2_get_reparse_tag_special_file(stat->mode))
-			attr |= ATTR_REPARSE;
+			attr |= FILE_ATTRIBUTE_REPARSE_POINT;
 	}
 
 	return attr;
@@ -753,16 +751,16 @@ static void build_encrypt_ctxt(struct smb2_encryption_neg_context *pneg_ctxt,
 	pneg_ctxt->Ciphers[0] = cipher_type;
 }
 
-static void build_compression_ctxt(struct smb2_compression_ctx *pneg_ctxt,
+static void build_compression_ctxt(struct smb2_compression_capabilities_context *pneg_ctxt,
 				   __le16 comp_algo)
 {
 	pneg_ctxt->ContextType = SMB2_COMPRESSION_CAPABILITIES;
 	pneg_ctxt->DataLength =
-		cpu_to_le16(sizeof(struct smb2_compression_ctx)
+		cpu_to_le16(sizeof(struct smb2_compression_capabilities_context)
 			- sizeof(struct smb2_neg_context));
 	pneg_ctxt->Reserved = cpu_to_le32(0);
 	pneg_ctxt->CompressionAlgorithmCount = cpu_to_le16(1);
-	pneg_ctxt->Reserved1 = cpu_to_le32(0);
+	pneg_ctxt->Flags = cpu_to_le32(0);
 	pneg_ctxt->CompressionAlgorithms[0] = comp_algo;
 }
 
@@ -802,11 +800,11 @@ static void build_posix_ctxt(struct smb2_posix_neg_context *pneg_ctxt)
 }
 
 static void assemble_neg_contexts(struct ksmbd_conn *conn,
-				  struct smb2_negotiate_rsp *rsp)
+				  struct smb2_negotiate_rsp *rsp,
+				  void *smb2_buf_len)
 {
-	/* +4 is to account for the RFC1001 len field */
 	char *pneg_ctxt = (char *)rsp +
-			le32_to_cpu(rsp->NegotiateContextOffset) + 4;
+			le32_to_cpu(rsp->NegotiateContextOffset);
 	int neg_ctxt_cnt = 1;
 	int ctxt_size;
 
@@ -815,7 +813,7 @@ static void assemble_neg_contexts(struct ksmbd_conn *conn,
 	build_preauth_ctxt((struct smb2_preauth_neg_context *)pneg_ctxt,
 			   conn->preauth_info->Preauth_HashId);
 	rsp->NegotiateContextCount = cpu_to_le16(neg_ctxt_cnt);
-	inc_rfc1001_len(rsp, AUTH_GSS_PADDING);
+	inc_rfc1001_len(smb2_buf_len, AUTH_GSS_PADDING);
 	ctxt_size = sizeof(struct smb2_preauth_neg_context);
 	/* Round to 8 byte boundary */
 	pneg_ctxt += round_up(sizeof(struct smb2_preauth_neg_context), 8);
@@ -839,12 +837,12 @@ static void assemble_neg_contexts(struct ksmbd_conn *conn,
 		ksmbd_debug(SMB,
 			    "assemble SMB2_COMPRESSION_CAPABILITIES context\n");
 		/* Temporarily set to SMB3_COMPRESS_NONE */
-		build_compression_ctxt((struct smb2_compression_ctx *)pneg_ctxt,
+		build_compression_ctxt((struct smb2_compression_capabilities_context *)pneg_ctxt,
 				       conn->compress_algorithm);
 		rsp->NegotiateContextCount = cpu_to_le16(++neg_ctxt_cnt);
-		ctxt_size += sizeof(struct smb2_compression_ctx) + 2;
+		ctxt_size += sizeof(struct smb2_compression_capabilities_context) + 2;
 		/* Round to 8 byte boundary */
-		pneg_ctxt += round_up(sizeof(struct smb2_compression_ctx) + 2,
+		pneg_ctxt += round_up(sizeof(struct smb2_compression_capabilities_context) + 2,
 				      8);
 	}
 
@@ -869,7 +867,7 @@ static void assemble_neg_contexts(struct ksmbd_conn *conn,
 		ctxt_size += sizeof(struct smb2_signing_capabilities) + 2;
 	}
 
-	inc_rfc1001_len(rsp, ctxt_size);
+	inc_rfc1001_len(smb2_buf_len, ctxt_size);
 }
 
 static __le32 decode_preauth_ctxt(struct ksmbd_conn *conn,
@@ -918,7 +916,7 @@ static void decode_encrypt_ctxt(struct ksmbd_conn *conn,
 }
 
 static void decode_compress_ctxt(struct ksmbd_conn *conn,
-				 struct smb2_compression_ctx *pneg_ctxt)
+				 struct smb2_compression_capabilities_context *pneg_ctxt)
 {
 	conn->compress_algorithm = SMB3_COMPRESS_NONE;
 }
@@ -939,8 +937,8 @@ static void decode_sign_cap_ctxt(struct ksmbd_conn *conn,
 	}
 
 	for (i = 0; i < sign_algo_cnt; i++) {
-		if (pneg_ctxt->SigningAlgorithms[i] == SIGNING_ALG_HMAC_SHA256 ||
-		    pneg_ctxt->SigningAlgorithms[i] == SIGNING_ALG_AES_CMAC) {
+		if (pneg_ctxt->SigningAlgorithms[i] == SIGNING_ALG_HMAC_SHA256_LE ||
+		    pneg_ctxt->SigningAlgorithms[i] == SIGNING_ALG_AES_CMAC_LE) {
 			ksmbd_debug(SMB, "Signing Algorithm ID = 0x%x\n",
 				    pneg_ctxt->SigningAlgorithms[i]);
 			conn->signing_negotiated = true;
@@ -952,14 +950,14 @@ static void decode_sign_cap_ctxt(struct ksmbd_conn *conn,
 }
 
 static __le32 deassemble_neg_contexts(struct ksmbd_conn *conn,
-				      struct smb2_negotiate_req *req)
+				      struct smb2_negotiate_req *req,
+				      int len_of_smb)
 {
 	/* +4 is to account for the RFC1001 len field */
-	struct smb2_neg_context *pctx = (struct smb2_neg_context *)((char *)req + 4);
+	struct smb2_neg_context *pctx = (struct smb2_neg_context *)req;
 	int i = 0, len_of_ctxts;
 	int offset = le32_to_cpu(req->NegotiateContextOffset);
 	int neg_ctxt_cnt = le16_to_cpu(req->NegotiateContextCount);
-	int len_of_smb = be32_to_cpu(req->hdr.smb2_buf_length);
 	__le32 status = STATUS_INVALID_PARAMETER;
 
 	ksmbd_debug(SMB, "decoding %d negotiate contexts\n", neg_ctxt_cnt);
@@ -1011,7 +1009,7 @@ static __le32 deassemble_neg_contexts(struct ksmbd_conn *conn,
 				break;
 
 			decode_compress_ctxt(conn,
-					     (struct smb2_compression_ctx *)pctx);
+					     (struct smb2_compression_capabilities_context *)pctx);
 		} else if (pctx->ContextType == SMB2_NETNAME_NEGOTIATE_CONTEXT_ID) {
 			ksmbd_debug(SMB,
 				    "deassemble SMB2_NETNAME_NEGOTIATE_CONTEXT_ID context\n");
@@ -1044,8 +1042,8 @@ static __le32 deassemble_neg_contexts(struct ksmbd_conn *conn,
 int smb2_handle_negotiate(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
-	struct smb2_negotiate_req *req = work->request_buf;
-	struct smb2_negotiate_rsp *rsp = work->response_buf;
+	struct smb2_negotiate_req *req = smb2_get_msg(work->request_buf);
+	struct smb2_negotiate_rsp *rsp = smb2_get_msg(work->response_buf);
 	int rc = 0;
 	unsigned int smb2_buf_len, smb2_neg_size;
 	__le32 status;
@@ -1066,7 +1064,7 @@ int smb2_handle_negotiate(struct ksmbd_work *work)
 	}
 
 	smb2_buf_len = get_rfc1002_len(work->request_buf);
-	smb2_neg_size = offsetof(struct smb2_negotiate_req, Dialects) - 4;
+	smb2_neg_size = offsetof(struct smb2_negotiate_req, Dialects);
 	if (smb2_neg_size > smb2_buf_len) {
 		rsp->hdr.Status = STATUS_INVALID_PARAMETER;
 		rc = -EINVAL;
@@ -1115,7 +1113,8 @@ int smb2_handle_negotiate(struct ksmbd_work *work)
 			goto err_out;
 		}
 
-		status = deassemble_neg_contexts(conn, req);
+		status = deassemble_neg_contexts(conn, req,
+						 get_rfc1002_len(work->request_buf));
 		if (status != STATUS_SUCCESS) {
 			pr_err("deassemble_neg_contexts error(0x%x)\n",
 			       status);
@@ -1135,7 +1134,7 @@ int smb2_handle_negotiate(struct ksmbd_work *work)
 						 conn->preauth_info->Preauth_HashValue);
 		rsp->NegotiateContextOffset =
 				cpu_to_le32(OFFSET_OF_NEG_CONTEXT);
-		assemble_neg_contexts(conn, rsp);
+		assemble_neg_contexts(conn, rsp, work->response_buf);
 		break;
 	case SMB302_PROT_ID:
 		init_smb3_02_server(conn);
@@ -1183,10 +1182,9 @@ int smb2_handle_negotiate(struct ksmbd_work *work)
 
 	rsp->SecurityBufferOffset = cpu_to_le16(128);
 	rsp->SecurityBufferLength = cpu_to_le16(AUTH_GSS_LENGTH);
-	ksmbd_copy_gss_neg_header(((char *)(&rsp->hdr) +
-				  sizeof(rsp->hdr.smb2_buf_length)) +
-				   le16_to_cpu(rsp->SecurityBufferOffset));
-	inc_rfc1001_len(rsp, sizeof(struct smb2_negotiate_rsp) -
+	ksmbd_copy_gss_neg_header((char *)(&rsp->hdr) +
+				  le16_to_cpu(rsp->SecurityBufferOffset));
+	inc_rfc1001_len(work->response_buf, sizeof(struct smb2_negotiate_rsp) -
 			sizeof(struct smb2_hdr) - sizeof(rsp->Buffer) +
 			 AUTH_GSS_LENGTH);
 	rsp->SecurityMode = SMB2_NEGOTIATE_SIGNING_ENABLED_LE;
@@ -1278,7 +1276,7 @@ static int ntlm_negotiate(struct ksmbd_work *work,
 			  struct negotiate_message *negblob,
 			  size_t negblob_len)
 {
-	struct smb2_sess_setup_rsp *rsp = work->response_buf;
+	struct smb2_sess_setup_rsp *rsp = smb2_get_msg(work->response_buf);
 	struct challenge_message *chgblob;
 	unsigned char *spnego_blob = NULL;
 	u16 spnego_blob_len;
@@ -1386,8 +1384,8 @@ static struct ksmbd_user *session_user(struct ksmbd_conn *conn,
 
 static int ntlm_authenticate(struct ksmbd_work *work)
 {
-	struct smb2_sess_setup_req *req = work->request_buf;
-	struct smb2_sess_setup_rsp *rsp = work->response_buf;
+	struct smb2_sess_setup_req *req = smb2_get_msg(work->request_buf);
+	struct smb2_sess_setup_rsp *rsp = smb2_get_msg(work->response_buf);
 	struct ksmbd_conn *conn = work->conn;
 	struct ksmbd_session *sess = work->sess;
 	struct channel *chann = NULL;
@@ -1410,7 +1408,7 @@ static int ntlm_authenticate(struct ksmbd_work *work)
 		memcpy((char *)&rsp->hdr.ProtocolId + sz, spnego_blob, spnego_blob_len);
 		rsp->SecurityBufferLength = cpu_to_le16(spnego_blob_len);
 		kfree(spnego_blob);
-		inc_rfc1001_len(rsp, spnego_blob_len - 1);
+		inc_rfc1001_len(work->response_buf, spnego_blob_len - 1);
 	}
 
 	user = session_user(conn, req);
@@ -1522,8 +1520,8 @@ static int ntlm_authenticate(struct ksmbd_work *work)
 #ifdef CONFIG_SMB_SERVER_KERBEROS5
 static int krb5_authenticate(struct ksmbd_work *work)
 {
-	struct smb2_sess_setup_req *req = work->request_buf;
-	struct smb2_sess_setup_rsp *rsp = work->response_buf;
+	struct smb2_sess_setup_req *req = smb2_get_msg(work->request_buf);
+	struct smb2_sess_setup_rsp *rsp = smb2_get_msg(work->response_buf);
 	struct ksmbd_conn *conn = work->conn;
 	struct ksmbd_session *sess = work->sess;
 	char *in_blob, *out_blob;
@@ -1538,8 +1536,7 @@ static int krb5_authenticate(struct ksmbd_work *work)
 	out_blob = (char *)&rsp->hdr.ProtocolId +
 		le16_to_cpu(rsp->SecurityBufferOffset);
 	out_len = work->response_sz -
-		offsetof(struct smb2_hdr, smb2_buf_length) -
-		le16_to_cpu(rsp->SecurityBufferOffset);
+		(le16_to_cpu(rsp->SecurityBufferOffset) + 4);
 
 	/* Check previous session */
 	prev_sess_id = le64_to_cpu(req->PreviousSessionId);
@@ -1556,7 +1553,7 @@ static int krb5_authenticate(struct ksmbd_work *work)
 		return -EINVAL;
 	}
 	rsp->SecurityBufferLength = cpu_to_le16(out_len);
-	inc_rfc1001_len(rsp, out_len - 1);
+	inc_rfc1001_len(work->response_buf, out_len - 1);
 
 	if ((conn->sign || server_conf.enforced_signing) ||
 	    (req->SecurityMode & SMB2_NEGOTIATE_SIGNING_REQUIRED))
@@ -1612,8 +1609,8 @@ static int krb5_authenticate(struct ksmbd_work *work)
 int smb2_sess_setup(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
-	struct smb2_sess_setup_req *req = work->request_buf;
-	struct smb2_sess_setup_rsp *rsp = work->response_buf;
+	struct smb2_sess_setup_req *req = smb2_get_msg(work->request_buf);
+	struct smb2_sess_setup_rsp *rsp = smb2_get_msg(work->response_buf);
 	struct ksmbd_session *sess;
 	struct negotiate_message *negblob;
 	unsigned int negblob_len, negblob_off;
@@ -1625,7 +1622,7 @@ int smb2_sess_setup(struct ksmbd_work *work)
 	rsp->SessionFlags = 0;
 	rsp->SecurityBufferOffset = cpu_to_le16(72);
 	rsp->SecurityBufferLength = 0;
-	inc_rfc1001_len(rsp, 9);
+	inc_rfc1001_len(work->response_buf, 9);
 
 	if (!req->hdr.SessionId) {
 		sess = ksmbd_smb2_session_create();
@@ -1699,7 +1696,7 @@ int smb2_sess_setup(struct ksmbd_work *work)
 
 	negblob_off = le16_to_cpu(req->SecurityBufferOffset);
 	negblob_len = le16_to_cpu(req->SecurityBufferLength);
-	if (negblob_off < (offsetof(struct smb2_sess_setup_req, Buffer) - 4) ||
+	if (negblob_off < offsetof(struct smb2_sess_setup_req, Buffer) ||
 	    negblob_len < offsetof(struct negotiate_message, NegotiateFlags))
 		return -EINVAL;
 
@@ -1739,7 +1736,8 @@ int smb2_sess_setup(struct ksmbd_work *work)
 				 * Note: here total size -1 is done as an
 				 * adjustment for 0 size blob
 				 */
-				inc_rfc1001_len(rsp, le16_to_cpu(rsp->SecurityBufferLength) - 1);
+				inc_rfc1001_len(work->response_buf,
+						le16_to_cpu(rsp->SecurityBufferLength) - 1);
 
 			} else if (negblob->MessageType == NtLmAuthenticate) {
 				rc = ntlm_authenticate(work);
@@ -1828,8 +1826,8 @@ int smb2_sess_setup(struct ksmbd_work *work)
 int smb2_tree_connect(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
-	struct smb2_tree_connect_req *req = work->request_buf;
-	struct smb2_tree_connect_rsp *rsp = work->response_buf;
+	struct smb2_tree_connect_req *req = smb2_get_msg(work->request_buf);
+	struct smb2_tree_connect_rsp *rsp = smb2_get_msg(work->response_buf);
 	struct ksmbd_session *sess = work->sess;
 	char *treename = NULL, *name = NULL;
 	struct ksmbd_tree_conn_status status;
@@ -1894,7 +1892,7 @@ int smb2_tree_connect(struct ksmbd_work *work)
 	rsp->Reserved = 0;
 	/* default manual caching */
 	rsp->ShareFlags = SMB2_SHAREFLAG_MANUAL_CACHING;
-	inc_rfc1001_len(rsp, 16);
+	inc_rfc1001_len(work->response_buf, 16);
 
 	if (!IS_ERR(treename))
 		kfree(treename);
@@ -1999,17 +1997,18 @@ static int smb2_create_open_flags(bool file_present, __le32 access,
  */
 int smb2_tree_disconnect(struct ksmbd_work *work)
 {
-	struct smb2_tree_disconnect_rsp *rsp = work->response_buf;
+	struct smb2_tree_disconnect_rsp *rsp = smb2_get_msg(work->response_buf);
 	struct ksmbd_session *sess = work->sess;
 	struct ksmbd_tree_connect *tcon = work->tcon;
 
 	rsp->StructureSize = cpu_to_le16(4);
-	inc_rfc1001_len(rsp, 4);
+	inc_rfc1001_len(work->response_buf, 4);
 
 	ksmbd_debug(SMB, "request\n");
 
 	if (!tcon) {
-		struct smb2_tree_disconnect_req *req = work->request_buf;
+		struct smb2_tree_disconnect_req *req =
+			smb2_get_msg(work->request_buf);
 
 		ksmbd_debug(SMB, "Invalid tid %d\n", req->hdr.Id.SyncId.TreeId);
 		rsp->hdr.Status = STATUS_NETWORK_NAME_DELETED;
@@ -2031,11 +2030,11 @@ int smb2_tree_disconnect(struct ksmbd_work *work)
 int smb2_session_logoff(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
-	struct smb2_logoff_rsp *rsp = work->response_buf;
+	struct smb2_logoff_rsp *rsp = smb2_get_msg(work->response_buf);
 	struct ksmbd_session *sess = work->sess;
 
 	rsp->StructureSize = cpu_to_le16(4);
-	inc_rfc1001_len(rsp, 4);
+	inc_rfc1001_len(work->response_buf, 4);
 
 	ksmbd_debug(SMB, "request\n");
 
@@ -2048,7 +2047,7 @@ int smb2_session_logoff(struct ksmbd_work *work)
 	ksmbd_conn_wait_idle(conn);
 
 	if (ksmbd_tree_conn_session_logoff(sess)) {
-		struct smb2_logoff_req *req = work->request_buf;
+		struct smb2_logoff_req *req = smb2_get_msg(work->request_buf);
 
 		ksmbd_debug(SMB, "Invalid tid %d\n", req->hdr.Id.SyncId.TreeId);
 		rsp->hdr.Status = STATUS_NETWORK_NAME_DELETED;
@@ -2075,8 +2074,8 @@ int smb2_session_logoff(struct ksmbd_work *work)
  */
 static noinline int create_smb2_pipe(struct ksmbd_work *work)
 {
-	struct smb2_create_rsp *rsp = work->response_buf;
-	struct smb2_create_req *req = work->request_buf;
+	struct smb2_create_rsp *rsp = smb2_get_msg(work->response_buf);
+	struct smb2_create_req *req = smb2_get_msg(work->request_buf);
 	int id;
 	int err;
 	char *name;
@@ -2099,7 +2098,7 @@ static noinline int create_smb2_pipe(struct ksmbd_work *work)
 	rsp->hdr.Status = STATUS_SUCCESS;
 	rsp->StructureSize = cpu_to_le16(89);
 	rsp->OplockLevel = SMB2_OPLOCK_LEVEL_NONE;
-	rsp->Reserved = 0;
+	rsp->Flags = 0;
 	rsp->CreateAction = cpu_to_le32(FILE_OPENED);
 
 	rsp->CreationTime = cpu_to_le64(0);
@@ -2107,14 +2106,14 @@ static noinline int create_smb2_pipe(struct ksmbd_work *work)
 	rsp->ChangeTime = cpu_to_le64(0);
 	rsp->AllocationSize = cpu_to_le64(0);
 	rsp->EndofFile = cpu_to_le64(0);
-	rsp->FileAttributes = ATTR_NORMAL_LE;
+	rsp->FileAttributes = FILE_ATTRIBUTE_NORMAL_LE;
 	rsp->Reserved2 = 0;
 	rsp->VolatileFileId = cpu_to_le64(id);
 	rsp->PersistentFileId = 0;
 	rsp->CreateContextsOffset = 0;
 	rsp->CreateContextsLength = 0;
 
-	inc_rfc1001_len(rsp, 88); /* StructureSize - 1*/
+	inc_rfc1001_len(work->response_buf, 88); /* StructureSize - 1*/
 	kfree(name);
 	return 0;
 
@@ -2353,7 +2352,7 @@ static void smb2_update_xattrs(struct ksmbd_tree_connect *tcon,
 	struct xattr_dos_attrib da;
 	int rc;
 
-	fp->f_ci->m_fattr &= ~(ATTR_HIDDEN_LE | ATTR_SYSTEM_LE);
+	fp->f_ci->m_fattr &= ~(FILE_ATTRIBUTE_HIDDEN_LE | FILE_ATTRIBUTE_SYSTEM_LE);
 
 	/* get FileAttributes from XATTR_NAME_DOS_ATTRIBUTE */
 	if (!test_share_config_flag(tcon->share_conf,
@@ -2463,7 +2462,7 @@ int smb2_open(struct ksmbd_work *work)
 	struct ksmbd_session *sess = work->sess;
 	struct ksmbd_tree_connect *tcon = work->tcon;
 	struct smb2_create_req *req;
-	struct smb2_create_rsp *rsp, *rsp_org;
+	struct smb2_create_rsp *rsp;
 	struct path path;
 	struct ksmbd_share_config *share = tcon->share_conf;
 	struct ksmbd_file *fp = NULL;
@@ -2489,7 +2488,6 @@ int smb2_open(struct ksmbd_work *work)
 	umode_t posix_mode = 0;
 	__le32 daccess, maximal_access = 0;
 
-	rsp_org = work->response_buf;
 	WORK_BUFFERS(work, req, rsp);
 
 	if (req->hdr.NextCommand && !work->next_smb2_rcv_hdr_off &&
@@ -2559,7 +2557,7 @@ int smb2_open(struct ksmbd_work *work)
 	if (req_op_level == SMB2_OPLOCK_LEVEL_LEASE)
 		lc = parse_lease_state(req);
 
-	if (le32_to_cpu(req->ImpersonationLevel) > le32_to_cpu(IL_DELEGATE_LE)) {
+	if (le32_to_cpu(req->ImpersonationLevel) > le32_to_cpu(IL_DELEGATE)) {
 		pr_err("Invalid impersonationlevel : 0x%x\n",
 		       le32_to_cpu(req->ImpersonationLevel));
 		rc = -EIO;
@@ -2567,7 +2565,7 @@ int smb2_open(struct ksmbd_work *work)
 		goto err_out1;
 	}
 
-	if (req->CreateOptions && !(req->CreateOptions & CREATE_OPTIONS_MASK)) {
+	if (req->CreateOptions && !(req->CreateOptions & CREATE_OPTIONS_MASK_LE)) {
 		pr_err("Invalid create options : 0x%x\n",
 		       le32_to_cpu(req->CreateOptions));
 		rc = -EINVAL;
@@ -2609,7 +2607,7 @@ int smb2_open(struct ksmbd_work *work)
 		goto err_out1;
 	}
 
-	if (req->FileAttributes && !(req->FileAttributes & ATTR_MASK_LE)) {
+	if (req->FileAttributes && !(req->FileAttributes & FILE_ATTRIBUTE_MASK_LE)) {
 		pr_err("Invalid file attribute : 0x%x\n",
 		       le32_to_cpu(req->FileAttributes));
 		rc = -EINVAL;
@@ -2740,7 +2738,7 @@ int smb2_open(struct ksmbd_work *work)
 		}
 
 		if (req->CreateOptions & FILE_DIRECTORY_FILE_LE &&
-		    req->FileAttributes & ATTR_NORMAL_LE) {
+		    req->FileAttributes & FILE_ATTRIBUTE_NORMAL_LE) {
 			rsp->hdr.Status = STATUS_NOT_A_DIRECTORY;
 			rc = -EIO;
 		}
@@ -3119,7 +3117,7 @@ int smb2_open(struct ksmbd_work *work)
 	opinfo = rcu_dereference(fp->f_opinfo);
 	rsp->OplockLevel = opinfo != NULL ? opinfo->level : 0;
 	rcu_read_unlock();
-	rsp->Reserved = 0;
+	rsp->Flags = 0;
 	rsp->CreateAction = cpu_to_le32(file_info);
 	rsp->CreationTime = cpu_to_le64(fp->create_time);
 	time = ksmbd_UnixTimeToNT(stat.atime);
@@ -3140,7 +3138,7 @@ int smb2_open(struct ksmbd_work *work)
 
 	rsp->CreateContextsOffset = 0;
 	rsp->CreateContextsLength = 0;
-	inc_rfc1001_len(rsp_org, 88); /* StructureSize - 1*/
+	inc_rfc1001_len(work->response_buf, 88); /* StructureSize - 1*/
 
 	/* If lease is request send lease context response */
 	if (opinfo && opinfo->is_lease) {
@@ -3155,7 +3153,8 @@ int smb2_open(struct ksmbd_work *work)
 		create_lease_buf(rsp->Buffer, opinfo->o_lease);
 		le32_add_cpu(&rsp->CreateContextsLength,
 			     conn->vals->create_lease_size);
-		inc_rfc1001_len(rsp_org, conn->vals->create_lease_size);
+		inc_rfc1001_len(work->response_buf,
+				conn->vals->create_lease_size);
 		next_ptr = &lease_ccontext->Next;
 		next_off = conn->vals->create_lease_size;
 	}
@@ -3175,7 +3174,8 @@ int smb2_open(struct ksmbd_work *work)
 				le32_to_cpu(maximal_access));
 		le32_add_cpu(&rsp->CreateContextsLength,
 			     conn->vals->create_mxac_size);
-		inc_rfc1001_len(rsp_org, conn->vals->create_mxac_size);
+		inc_rfc1001_len(work->response_buf,
+				conn->vals->create_mxac_size);
 		if (next_ptr)
 			*next_ptr = cpu_to_le32(next_off);
 		next_ptr = &mxac_ccontext->Next;
@@ -3193,7 +3193,8 @@ int smb2_open(struct ksmbd_work *work)
 				stat.ino, tcon->id);
 		le32_add_cpu(&rsp->CreateContextsLength,
 			     conn->vals->create_disk_id_size);
-		inc_rfc1001_len(rsp_org, conn->vals->create_disk_id_size);
+		inc_rfc1001_len(work->response_buf,
+				conn->vals->create_disk_id_size);
 		if (next_ptr)
 			*next_ptr = cpu_to_le32(next_off);
 		next_ptr = &disk_id_ccontext->Next;
@@ -3207,15 +3208,15 @@ int smb2_open(struct ksmbd_work *work)
 				fp);
 		le32_add_cpu(&rsp->CreateContextsLength,
 			     conn->vals->create_posix_size);
-		inc_rfc1001_len(rsp_org, conn->vals->create_posix_size);
+		inc_rfc1001_len(work->response_buf,
+				conn->vals->create_posix_size);
 		if (next_ptr)
 			*next_ptr = cpu_to_le32(next_off);
 	}
 
 	if (contxt_cnt > 0) {
 		rsp->CreateContextsOffset =
-			cpu_to_le32(offsetof(struct smb2_create_rsp, Buffer)
-			- 4);
+			cpu_to_le32(offsetof(struct smb2_create_rsp, Buffer));
 	}
 
 err_out:
@@ -3422,9 +3423,9 @@ static int smb2_populate_readdir_entry(struct ksmbd_conn *conn, int info_level,
 		ffdinfo->EaSize =
 			smb2_get_reparse_tag_special_file(ksmbd_kstat->kstat->mode);
 		if (ffdinfo->EaSize)
-			ffdinfo->ExtFileAttributes = ATTR_REPARSE_POINT_LE;
+			ffdinfo->ExtFileAttributes = FILE_ATTRIBUTE_REPARSE_POINT_LE;
 		if (d_info->hide_dot_file && d_info->name[0] == '.')
-			ffdinfo->ExtFileAttributes |= ATTR_HIDDEN_LE;
+			ffdinfo->ExtFileAttributes |= FILE_ATTRIBUTE_HIDDEN_LE;
 		memcpy(ffdinfo->FileName, conv_name, conv_len);
 		ffdinfo->NextEntryOffset = cpu_to_le32(next_entry_offset);
 		break;
@@ -3438,11 +3439,11 @@ static int smb2_populate_readdir_entry(struct ksmbd_conn *conn, int info_level,
 		fbdinfo->EaSize =
 			smb2_get_reparse_tag_special_file(ksmbd_kstat->kstat->mode);
 		if (fbdinfo->EaSize)
-			fbdinfo->ExtFileAttributes = ATTR_REPARSE_POINT_LE;
+			fbdinfo->ExtFileAttributes = FILE_ATTRIBUTE_REPARSE_POINT_LE;
 		fbdinfo->ShortNameLength = 0;
 		fbdinfo->Reserved = 0;
 		if (d_info->hide_dot_file && d_info->name[0] == '.')
-			fbdinfo->ExtFileAttributes |= ATTR_HIDDEN_LE;
+			fbdinfo->ExtFileAttributes |= FILE_ATTRIBUTE_HIDDEN_LE;
 		memcpy(fbdinfo->FileName, conv_name, conv_len);
 		fbdinfo->NextEntryOffset = cpu_to_le32(next_entry_offset);
 		break;
@@ -3454,7 +3455,7 @@ static int smb2_populate_readdir_entry(struct ksmbd_conn *conn, int info_level,
 		fdinfo = (struct file_directory_info *)kstat;
 		fdinfo->FileNameLength = cpu_to_le32(conv_len);
 		if (d_info->hide_dot_file && d_info->name[0] == '.')
-			fdinfo->ExtFileAttributes |= ATTR_HIDDEN_LE;
+			fdinfo->ExtFileAttributes |= FILE_ATTRIBUTE_HIDDEN_LE;
 		memcpy(fdinfo->FileName, conv_name, conv_len);
 		fdinfo->NextEntryOffset = cpu_to_le32(next_entry_offset);
 		break;
@@ -3478,11 +3479,11 @@ static int smb2_populate_readdir_entry(struct ksmbd_conn *conn, int info_level,
 		dinfo->EaSize =
 			smb2_get_reparse_tag_special_file(ksmbd_kstat->kstat->mode);
 		if (dinfo->EaSize)
-			dinfo->ExtFileAttributes = ATTR_REPARSE_POINT_LE;
+			dinfo->ExtFileAttributes = FILE_ATTRIBUTE_REPARSE_POINT_LE;
 		dinfo->Reserved = 0;
 		dinfo->UniqueId = cpu_to_le64(ksmbd_kstat->kstat->ino);
 		if (d_info->hide_dot_file && d_info->name[0] == '.')
-			dinfo->ExtFileAttributes |= ATTR_HIDDEN_LE;
+			dinfo->ExtFileAttributes |= FILE_ATTRIBUTE_HIDDEN_LE;
 		memcpy(dinfo->FileName, conv_name, conv_len);
 		dinfo->NextEntryOffset = cpu_to_le32(next_entry_offset);
 		break;
@@ -3496,13 +3497,13 @@ static int smb2_populate_readdir_entry(struct ksmbd_conn *conn, int info_level,
 		fibdinfo->EaSize =
 			smb2_get_reparse_tag_special_file(ksmbd_kstat->kstat->mode);
 		if (fibdinfo->EaSize)
-			fibdinfo->ExtFileAttributes = ATTR_REPARSE_POINT_LE;
+			fibdinfo->ExtFileAttributes = FILE_ATTRIBUTE_REPARSE_POINT_LE;
 		fibdinfo->UniqueId = cpu_to_le64(ksmbd_kstat->kstat->ino);
 		fibdinfo->ShortNameLength = 0;
 		fibdinfo->Reserved = 0;
 		fibdinfo->Reserved2 = cpu_to_le16(0);
 		if (d_info->hide_dot_file && d_info->name[0] == '.')
-			fibdinfo->ExtFileAttributes |= ATTR_HIDDEN_LE;
+			fibdinfo->ExtFileAttributes |= FILE_ATTRIBUTE_HIDDEN_LE;
 		memcpy(fibdinfo->FileName, conv_name, conv_len);
 		fibdinfo->NextEntryOffset = cpu_to_le32(next_entry_offset);
 		break;
@@ -3528,9 +3529,10 @@ static int smb2_populate_readdir_entry(struct ksmbd_conn *conn, int info_level,
 		posix_info->Mode = cpu_to_le32(ksmbd_kstat->kstat->mode);
 		posix_info->Inode = cpu_to_le64(ksmbd_kstat->kstat->ino);
 		posix_info->DosAttributes =
-			S_ISDIR(ksmbd_kstat->kstat->mode) ? ATTR_DIRECTORY_LE : ATTR_ARCHIVE_LE;
+			S_ISDIR(ksmbd_kstat->kstat->mode) ?
+				FILE_ATTRIBUTE_DIRECTORY_LE : FILE_ATTRIBUTE_ARCHIVE_LE;
 		if (d_info->hide_dot_file && d_info->name[0] == '.')
-			posix_info->DosAttributes |= ATTR_HIDDEN_LE;
+			posix_info->DosAttributes |= FILE_ATTRIBUTE_HIDDEN_LE;
 		id_to_sid(from_kuid_munged(&init_user_ns, ksmbd_kstat->kstat->uid),
 			  SIDNFS_USER, (struct smb_sid *)&posix_info->SidBuffer[0]);
 		id_to_sid(from_kgid_munged(&init_user_ns, ksmbd_kstat->kstat->gid),
@@ -3816,7 +3818,7 @@ int smb2_query_dir(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
 	struct smb2_query_directory_req *req;
-	struct smb2_query_directory_rsp *rsp, *rsp_org;
+	struct smb2_query_directory_rsp *rsp;
 	struct ksmbd_share_config *share = work->tcon->share_conf;
 	struct ksmbd_file *dir_fp = NULL;
 	struct ksmbd_dir_info d_info;
@@ -3826,7 +3828,6 @@ int smb2_query_dir(struct ksmbd_work *work)
 	int buffer_sz;
 	struct smb2_query_dir_private query_dir_private = {NULL, };
 
-	rsp_org = work->response_buf;
 	WORK_BUFFERS(work, req, rsp);
 
 	if (ksmbd_override_fsids(work)) {
@@ -3947,7 +3948,7 @@ int smb2_query_dir(struct ksmbd_work *work)
 		rsp->OutputBufferOffset = cpu_to_le16(0);
 		rsp->OutputBufferLength = cpu_to_le32(0);
 		rsp->Buffer[0] = 0;
-		inc_rfc1001_len(rsp_org, 9);
+		inc_rfc1001_len(work->response_buf, 9);
 	} else {
 		((struct file_directory_info *)
 		((char *)rsp->Buffer + d_info.last_entry_offset))
@@ -3956,7 +3957,7 @@ int smb2_query_dir(struct ksmbd_work *work)
 		rsp->StructureSize = cpu_to_le16(9);
 		rsp->OutputBufferOffset = cpu_to_le16(72);
 		rsp->OutputBufferLength = cpu_to_le32(d_info.data_count);
-		inc_rfc1001_len(rsp_org, 8 + d_info.data_count);
+		inc_rfc1001_len(work->response_buf, 8 + d_info.data_count);
 	}
 
 	kfree(srch_ptr);
@@ -3999,26 +4000,28 @@ int smb2_query_dir(struct ksmbd_work *work)
  * Return:	0 on success, otherwise error
  */
 static int buffer_check_err(int reqOutputBufferLength,
-			    struct smb2_query_info_rsp *rsp, int infoclass_size)
+			    struct smb2_query_info_rsp *rsp,
+			    void *rsp_org, int infoclass_size)
 {
 	if (reqOutputBufferLength < le32_to_cpu(rsp->OutputBufferLength)) {
 		if (reqOutputBufferLength < infoclass_size) {
 			pr_err("Invalid Buffer Size Requested\n");
 			rsp->hdr.Status = STATUS_INFO_LENGTH_MISMATCH;
-			rsp->hdr.smb2_buf_length = cpu_to_be32(sizeof(struct smb2_hdr) - 4);
+			*(__be32 *)rsp_org = cpu_to_be32(sizeof(struct smb2_hdr));
 			return -EINVAL;
 		}
 
 		ksmbd_debug(SMB, "Buffer Overflow\n");
 		rsp->hdr.Status = STATUS_BUFFER_OVERFLOW;
-		rsp->hdr.smb2_buf_length = cpu_to_be32(sizeof(struct smb2_hdr) - 4 +
+		*(__be32 *)rsp_org = cpu_to_be32(sizeof(struct smb2_hdr) +
 				reqOutputBufferLength);
 		rsp->OutputBufferLength = cpu_to_le32(reqOutputBufferLength);
 	}
 	return 0;
 }
 
-static void get_standard_info_pipe(struct smb2_query_info_rsp *rsp)
+static void get_standard_info_pipe(struct smb2_query_info_rsp *rsp,
+				   void *rsp_org)
 {
 	struct smb2_file_standard_info *sinfo;
 
@@ -4031,10 +4034,11 @@ static void get_standard_info_pipe(struct smb2_query_info_rsp *rsp)
 	sinfo->Directory = 0;
 	rsp->OutputBufferLength =
 		cpu_to_le32(sizeof(struct smb2_file_standard_info));
-	inc_rfc1001_len(rsp, sizeof(struct smb2_file_standard_info));
+	inc_rfc1001_len(rsp_org, sizeof(struct smb2_file_standard_info));
 }
 
-static void get_internal_info_pipe(struct smb2_query_info_rsp *rsp, u64 num)
+static void get_internal_info_pipe(struct smb2_query_info_rsp *rsp, u64 num,
+				   void *rsp_org)
 {
 	struct smb2_file_internal_info *file_info;
 
@@ -4044,12 +4048,13 @@ static void get_internal_info_pipe(struct smb2_query_info_rsp *rsp, u64 num)
 	file_info->IndexNumber = cpu_to_le64(num | (1ULL << 63));
 	rsp->OutputBufferLength =
 		cpu_to_le32(sizeof(struct smb2_file_internal_info));
-	inc_rfc1001_len(rsp, sizeof(struct smb2_file_internal_info));
+	inc_rfc1001_len(rsp_org, sizeof(struct smb2_file_internal_info));
 }
 
 static int smb2_get_info_file_pipe(struct ksmbd_session *sess,
 				   struct smb2_query_info_req *req,
-				   struct smb2_query_info_rsp *rsp)
+				   struct smb2_query_info_rsp *rsp,
+				   void *rsp_org)
 {
 	u64 id;
 	int rc;
@@ -4067,14 +4072,16 @@ static int smb2_get_info_file_pipe(struct ksmbd_session *sess,
 
 	switch (req->FileInfoClass) {
 	case FILE_STANDARD_INFORMATION:
-		get_standard_info_pipe(rsp);
+		get_standard_info_pipe(rsp, rsp_org);
 		rc = buffer_check_err(le32_to_cpu(req->OutputBufferLength),
-				      rsp, FILE_STANDARD_INFORMATION_SIZE);
+				      rsp, rsp_org,
+				      FILE_STANDARD_INFORMATION_SIZE);
 		break;
 	case FILE_INTERNAL_INFORMATION:
-		get_internal_info_pipe(rsp, id);
+		get_internal_info_pipe(rsp, id, rsp_org);
 		rc = buffer_check_err(le32_to_cpu(req->OutputBufferLength),
-				      rsp, FILE_INTERNAL_INFORMATION_SIZE);
+				      rsp, rsp_org,
+				      FILE_INTERNAL_INFORMATION_SIZE);
 		break;
 	default:
 		ksmbd_debug(SMB, "smb2_info_file_pipe for %u not supported\n",
@@ -4688,7 +4695,7 @@ static int find_file_posix_info(struct smb2_query_info_rsp *rsp,
 
 static int smb2_get_info_file(struct ksmbd_work *work,
 			      struct smb2_query_info_req *req,
-			      struct smb2_query_info_rsp *rsp, void *rsp_org)
+			      struct smb2_query_info_rsp *rsp)
 {
 	struct ksmbd_file *fp;
 	int fileinfoclass = 0;
@@ -4699,7 +4706,8 @@ static int smb2_get_info_file(struct ksmbd_work *work,
 	if (test_share_config_flag(work->tcon->share_conf,
 				   KSMBD_SHARE_FLAG_PIPE)) {
 		/* smb2 info file called for pipe */
-		return smb2_get_info_file_pipe(work->sess, req, rsp);
+		return smb2_get_info_file_pipe(work->sess, req, rsp,
+					       work->response_buf);
 	}
 
 	if (work->next_smb2_rcv_hdr_off) {
@@ -4724,77 +4732,77 @@ static int smb2_get_info_file(struct ksmbd_work *work,
 
 	switch (fileinfoclass) {
 	case FILE_ACCESS_INFORMATION:
-		get_file_access_info(rsp, fp, rsp_org);
+		get_file_access_info(rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_ACCESS_INFORMATION_SIZE;
 		break;
 
 	case FILE_BASIC_INFORMATION:
-		rc = get_file_basic_info(rsp, fp, rsp_org);
+		rc = get_file_basic_info(rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_BASIC_INFORMATION_SIZE;
 		break;
 
 	case FILE_STANDARD_INFORMATION:
-		get_file_standard_info(rsp, fp, rsp_org);
+		get_file_standard_info(rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_STANDARD_INFORMATION_SIZE;
 		break;
 
 	case FILE_ALIGNMENT_INFORMATION:
-		get_file_alignment_info(rsp, rsp_org);
+		get_file_alignment_info(rsp, work->response_buf);
 		file_infoclass_size = FILE_ALIGNMENT_INFORMATION_SIZE;
 		break;
 
 	case FILE_ALL_INFORMATION:
-		rc = get_file_all_info(work, rsp, fp, rsp_org);
+		rc = get_file_all_info(work, rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_ALL_INFORMATION_SIZE;
 		break;
 
 	case FILE_ALTERNATE_NAME_INFORMATION:
-		get_file_alternate_info(work, rsp, fp, rsp_org);
+		get_file_alternate_info(work, rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_ALTERNATE_NAME_INFORMATION_SIZE;
 		break;
 
 	case FILE_STREAM_INFORMATION:
-		get_file_stream_info(work, rsp, fp, rsp_org);
+		get_file_stream_info(work, rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_STREAM_INFORMATION_SIZE;
 		break;
 
 	case FILE_INTERNAL_INFORMATION:
-		get_file_internal_info(rsp, fp, rsp_org);
+		get_file_internal_info(rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_INTERNAL_INFORMATION_SIZE;
 		break;
 
 	case FILE_NETWORK_OPEN_INFORMATION:
-		rc = get_file_network_open_info(rsp, fp, rsp_org);
+		rc = get_file_network_open_info(rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_NETWORK_OPEN_INFORMATION_SIZE;
 		break;
 
 	case FILE_EA_INFORMATION:
-		get_file_ea_info(rsp, rsp_org);
+		get_file_ea_info(rsp, work->response_buf);
 		file_infoclass_size = FILE_EA_INFORMATION_SIZE;
 		break;
 
 	case FILE_FULL_EA_INFORMATION:
-		rc = smb2_get_ea(work, fp, req, rsp, rsp_org);
+		rc = smb2_get_ea(work, fp, req, rsp, work->response_buf);
 		file_infoclass_size = FILE_FULL_EA_INFORMATION_SIZE;
 		break;
 
 	case FILE_POSITION_INFORMATION:
-		get_file_position_info(rsp, fp, rsp_org);
+		get_file_position_info(rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_POSITION_INFORMATION_SIZE;
 		break;
 
 	case FILE_MODE_INFORMATION:
-		get_file_mode_info(rsp, fp, rsp_org);
+		get_file_mode_info(rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_MODE_INFORMATION_SIZE;
 		break;
 
 	case FILE_COMPRESSION_INFORMATION:
-		get_file_compression_info(rsp, fp, rsp_org);
+		get_file_compression_info(rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_COMPRESSION_INFORMATION_SIZE;
 		break;
 
 	case FILE_ATTRIBUTE_TAG_INFORMATION:
-		rc = get_file_attribute_tag_info(rsp, fp, rsp_org);
+		rc = get_file_attribute_tag_info(rsp, fp, work->response_buf);
 		file_infoclass_size = FILE_ATTRIBUTE_TAG_INFORMATION_SIZE;
 		break;
 	case SMB_FIND_FILE_POSIX_INFO:
@@ -4802,7 +4810,7 @@ static int smb2_get_info_file(struct ksmbd_work *work,
 			pr_err("client doesn't negotiate with SMB3.1.1 POSIX Extensions\n");
 			rc = -EOPNOTSUPP;
 		} else {
-			rc = find_file_posix_info(rsp, fp, rsp_org);
+			rc = find_file_posix_info(rsp, fp, work->response_buf);
 			file_infoclass_size = sizeof(struct smb311_posix_qinfo);
 		}
 		break;
@@ -4813,7 +4821,7 @@ static int smb2_get_info_file(struct ksmbd_work *work,
 	}
 	if (!rc)
 		rc = buffer_check_err(le32_to_cpu(req->OutputBufferLength),
-				      rsp,
+				      rsp, work->response_buf,
 				      file_infoclass_size);
 	ksmbd_fd_put(work, fp);
 	return rc;
@@ -4821,7 +4829,7 @@ static int smb2_get_info_file(struct ksmbd_work *work,
 
 static int smb2_get_info_filesystem(struct ksmbd_work *work,
 				    struct smb2_query_info_req *req,
-				    struct smb2_query_info_rsp *rsp, void *rsp_org)
+				    struct smb2_query_info_rsp *rsp)
 {
 	struct ksmbd_session *sess = work->sess;
 	struct ksmbd_conn *conn = sess->conn;
@@ -4857,7 +4865,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 		info->DeviceType = cpu_to_le32(stfs.f_type);
 		info->DeviceCharacteristics = cpu_to_le32(0x00000020);
 		rsp->OutputBufferLength = cpu_to_le32(8);
-		inc_rfc1001_len(rsp_org, 8);
+		inc_rfc1001_len(work->response_buf, 8);
 		fs_infoclass_size = FS_DEVICE_INFORMATION_SIZE;
 		break;
 	}
@@ -4883,7 +4891,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 		info->FileSystemNameLen = cpu_to_le32(len);
 		sz = sizeof(struct filesystem_attribute_info) - 2 + len;
 		rsp->OutputBufferLength = cpu_to_le32(sz);
-		inc_rfc1001_len(rsp_org, sz);
+		inc_rfc1001_len(work->response_buf, sz);
 		fs_infoclass_size = FS_ATTRIBUTE_INFORMATION_SIZE;
 		break;
 	}
@@ -4891,11 +4899,18 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 	{
 		struct filesystem_vol_info *info;
 		size_t sz;
+		unsigned int serial_crc = 0;
 
 		info = (struct filesystem_vol_info *)(rsp->Buffer);
 		info->VolumeCreationTime = 0;
+		serial_crc = crc32_le(serial_crc, share->name,
+				      strlen(share->name));
+		serial_crc = crc32_le(serial_crc, share->path,
+				      strlen(share->path));
+		serial_crc = crc32_le(serial_crc, ksmbd_netbios_name(),
+				      strlen(ksmbd_netbios_name()));
 		/* Taking dummy value of serial number*/
-		info->SerialNumber = cpu_to_le32(0xbc3ac512);
+		info->SerialNumber = cpu_to_le32(serial_crc);
 		len = smbConvertToUTF16((__le16 *)info->VolumeLabel,
 					share->name, PATH_MAX,
 					conn->local_nls, 0);
@@ -4904,7 +4919,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 		info->Reserved = 0;
 		sz = sizeof(struct filesystem_vol_info) - 2 + len;
 		rsp->OutputBufferLength = cpu_to_le32(sz);
-		inc_rfc1001_len(rsp_org, sz);
+		inc_rfc1001_len(work->response_buf, sz);
 		fs_infoclass_size = FS_VOLUME_INFORMATION_SIZE;
 		break;
 	}
@@ -4918,7 +4933,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 		info->SectorsPerAllocationUnit = cpu_to_le32(1);
 		info->BytesPerSector = cpu_to_le32(stfs.f_bsize);
 		rsp->OutputBufferLength = cpu_to_le32(24);
-		inc_rfc1001_len(rsp_org, 24);
+		inc_rfc1001_len(work->response_buf, 24);
 		fs_infoclass_size = FS_SIZE_INFORMATION_SIZE;
 		break;
 	}
@@ -4935,7 +4950,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 		info->SectorsPerAllocationUnit = cpu_to_le32(1);
 		info->BytesPerSector = cpu_to_le32(stfs.f_bsize);
 		rsp->OutputBufferLength = cpu_to_le32(32);
-		inc_rfc1001_len(rsp_org, 32);
+		inc_rfc1001_len(work->response_buf, 32);
 		fs_infoclass_size = FS_FULL_SIZE_INFORMATION_SIZE;
 		break;
 	}
@@ -4956,7 +4971,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 		info->extended_info.rel_date = 0;
 		memcpy(info->extended_info.version_string, "1.1.0", strlen("1.1.0"));
 		rsp->OutputBufferLength = cpu_to_le32(64);
-		inc_rfc1001_len(rsp_org, 64);
+		inc_rfc1001_len(work->response_buf, 64);
 		fs_infoclass_size = FS_OBJECT_ID_INFORMATION_SIZE;
 		break;
 	}
@@ -4977,7 +4992,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 		info->ByteOffsetForSectorAlignment = 0;
 		info->ByteOffsetForPartitionAlignment = 0;
 		rsp->OutputBufferLength = cpu_to_le32(28);
-		inc_rfc1001_len(rsp_org, 28);
+		inc_rfc1001_len(work->response_buf, 28);
 		fs_infoclass_size = FS_SECTOR_SIZE_INFORMATION_SIZE;
 		break;
 	}
@@ -4999,7 +5014,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 		info->DefaultQuotaLimit = cpu_to_le64(SMB2_NO_FID);
 		info->Padding = 0;
 		rsp->OutputBufferLength = cpu_to_le32(48);
-		inc_rfc1001_len(rsp_org, 48);
+		inc_rfc1001_len(work->response_buf, 48);
 		fs_infoclass_size = FS_CONTROL_INFORMATION_SIZE;
 		break;
 	}
@@ -5020,7 +5035,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 			info->TotalFileNodes = cpu_to_le64(stfs.f_files);
 			info->FreeFileNodes = cpu_to_le64(stfs.f_ffree);
 			rsp->OutputBufferLength = cpu_to_le32(56);
-			inc_rfc1001_len(rsp_org, 56);
+			inc_rfc1001_len(work->response_buf, 56);
 			fs_infoclass_size = FS_POSIX_INFORMATION_SIZE;
 		}
 		break;
@@ -5030,7 +5045,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 		return -EOPNOTSUPP;
 	}
 	rc = buffer_check_err(le32_to_cpu(req->OutputBufferLength),
-			      rsp,
+			      rsp, work->response_buf,
 			      fs_infoclass_size);
 	path_put(&path);
 	return rc;
@@ -5038,7 +5053,7 @@ static int smb2_get_info_filesystem(struct ksmbd_work *work,
 
 static int smb2_get_info_sec(struct ksmbd_work *work,
 			     struct smb2_query_info_req *req,
-			     struct smb2_query_info_rsp *rsp, void *rsp_org)
+			     struct smb2_query_info_rsp *rsp)
 {
 	struct ksmbd_file *fp;
 	struct user_namespace *user_ns;
@@ -5065,7 +5080,7 @@ static int smb2_get_info_sec(struct ksmbd_work *work,
 
 		secdesclen = sizeof(struct smb_ntsd);
 		rsp->OutputBufferLength = cpu_to_le32(secdesclen);
-		inc_rfc1001_len(rsp_org, secdesclen);
+		inc_rfc1001_len(work->response_buf, secdesclen);
 
 		return 0;
 	}
@@ -5107,7 +5122,7 @@ static int smb2_get_info_sec(struct ksmbd_work *work,
 		return rc;
 
 	rsp->OutputBufferLength = cpu_to_le32(secdesclen);
-	inc_rfc1001_len(rsp_org, secdesclen);
+	inc_rfc1001_len(work->response_buf, secdesclen);
 	return 0;
 }
 
@@ -5120,10 +5135,9 @@ static int smb2_get_info_sec(struct ksmbd_work *work,
 int smb2_query_info(struct ksmbd_work *work)
 {
 	struct smb2_query_info_req *req;
-	struct smb2_query_info_rsp *rsp, *rsp_org;
+	struct smb2_query_info_rsp *rsp;
 	int rc = 0;
 
-	rsp_org = work->response_buf;
 	WORK_BUFFERS(work, req, rsp);
 
 	ksmbd_debug(SMB, "GOT query info request\n");
@@ -5131,15 +5145,15 @@ int smb2_query_info(struct ksmbd_work *work)
 	switch (req->InfoType) {
 	case SMB2_O_INFO_FILE:
 		ksmbd_debug(SMB, "GOT SMB2_O_INFO_FILE\n");
-		rc = smb2_get_info_file(work, req, rsp, (void *)rsp_org);
+		rc = smb2_get_info_file(work, req, rsp);
 		break;
 	case SMB2_O_INFO_FILESYSTEM:
 		ksmbd_debug(SMB, "GOT SMB2_O_INFO_FILESYSTEM\n");
-		rc = smb2_get_info_filesystem(work, req, rsp, (void *)rsp_org);
+		rc = smb2_get_info_filesystem(work, req, rsp);
 		break;
 	case SMB2_O_INFO_SECURITY:
 		ksmbd_debug(SMB, "GOT SMB2_O_INFO_SECURITY\n");
-		rc = smb2_get_info_sec(work, req, rsp, (void *)rsp_org);
+		rc = smb2_get_info_sec(work, req, rsp);
 		break;
 	default:
 		ksmbd_debug(SMB, "InfoType %d not supported yet\n",
@@ -5164,7 +5178,7 @@ int smb2_query_info(struct ksmbd_work *work)
 	}
 	rsp->StructureSize = cpu_to_le16(9);
 	rsp->OutputBufferOffset = cpu_to_le16(72);
-	inc_rfc1001_len(rsp_org, 8);
+	inc_rfc1001_len(work->response_buf, 8);
 	return 0;
 }
 
@@ -5177,8 +5191,8 @@ int smb2_query_info(struct ksmbd_work *work)
 static noinline int smb2_close_pipe(struct ksmbd_work *work)
 {
 	u64 id;
-	struct smb2_close_req *req = work->request_buf;
-	struct smb2_close_rsp *rsp = work->response_buf;
+	struct smb2_close_req *req = smb2_get_msg(work->request_buf);
+	struct smb2_close_rsp *rsp = smb2_get_msg(work->response_buf);
 
 	id = le64_to_cpu(req->VolatileFileId);
 	ksmbd_session_rpc_close(work->sess, id);
@@ -5193,7 +5207,7 @@ static noinline int smb2_close_pipe(struct ksmbd_work *work)
 	rsp->AllocationSize = 0;
 	rsp->EndOfFile = 0;
 	rsp->Attributes = 0;
-	inc_rfc1001_len(rsp, 60);
+	inc_rfc1001_len(work->response_buf, 60);
 	return 0;
 }
 
@@ -5209,14 +5223,12 @@ int smb2_close(struct ksmbd_work *work)
 	u64 sess_id;
 	struct smb2_close_req *req;
 	struct smb2_close_rsp *rsp;
-	struct smb2_close_rsp *rsp_org;
 	struct ksmbd_conn *conn = work->conn;
 	struct ksmbd_file *fp;
 	struct inode *inode;
 	u64 time;
 	int err = 0;
 
-	rsp_org = work->response_buf;
 	WORK_BUFFERS(work, req, rsp);
 
 	if (test_share_config_flag(work->tcon->share_conf,
@@ -5306,7 +5318,7 @@ int smb2_close(struct ksmbd_work *work)
 			rsp->hdr.Status = STATUS_FILE_CLOSED;
 		smb2_set_err_rsp(work);
 	} else {
-		inc_rfc1001_len(rsp_org, 60);
+		inc_rfc1001_len(work->response_buf, 60);
 	}
 
 	return 0;
@@ -5320,11 +5332,11 @@ int smb2_close(struct ksmbd_work *work)
  */
 int smb2_echo(struct ksmbd_work *work)
 {
-	struct smb2_echo_rsp *rsp = work->response_buf;
+	struct smb2_echo_rsp *rsp = smb2_get_msg(work->response_buf);
 
 	rsp->StructureSize = cpu_to_le16(4);
 	rsp->Reserved = 0;
-	inc_rfc1001_len(rsp, 4);
+	inc_rfc1001_len(work->response_buf, 4);
 	return 0;
 }
 
@@ -5566,14 +5578,14 @@ static int set_file_basic_info(struct ksmbd_file *fp,
 
 	if (file_info->Attributes) {
 		if (!S_ISDIR(inode->i_mode) &&
-		    file_info->Attributes & ATTR_DIRECTORY_LE) {
+		    file_info->Attributes & FILE_ATTRIBUTE_DIRECTORY_LE) {
 			pr_err("can't change a file to a directory\n");
 			return -EINVAL;
 		}
 
-		if (!(S_ISDIR(inode->i_mode) && file_info->Attributes == ATTR_NORMAL_LE))
+		if (!(S_ISDIR(inode->i_mode) && file_info->Attributes == FILE_ATTRIBUTE_NORMAL_LE))
 			fp->f_ci->m_fattr = file_info->Attributes |
-				(fp->f_ci->m_fattr & ATTR_DIRECTORY_LE);
+				(fp->f_ci->m_fattr & FILE_ATTRIBUTE_DIRECTORY_LE);
 	}
 
 	if (test_share_config_flag(share, KSMBD_SHARE_FLAG_STORE_DOS_ATTRS) &&
@@ -5794,9 +5806,7 @@ static int set_file_mode_info(struct ksmbd_file *fp,
 
 	mode = file_info->Mode;
 
-	if ((mode & ~FILE_MODE_INFO_MASK) ||
-	    (mode & FILE_SYNCHRONOUS_IO_ALERT_LE &&
-	     mode & FILE_SYNCHRONOUS_IO_NONALERT_LE)) {
+	if ((mode & ~FILE_MODE_INFO_MASK)) {
 		pr_err("Mode is not valid : 0x%x\n", le32_to_cpu(mode));
 		return -EINVAL;
 	}
@@ -5943,14 +5953,13 @@ static int smb2_set_info_sec(struct ksmbd_file *fp, int addition_info,
 int smb2_set_info(struct ksmbd_work *work)
 {
 	struct smb2_set_info_req *req;
-	struct smb2_set_info_rsp *rsp, *rsp_org;
+	struct smb2_set_info_rsp *rsp;
 	struct ksmbd_file *fp;
 	int rc = 0;
 	unsigned int id = KSMBD_NO_FID, pid = KSMBD_NO_FID;
 
 	ksmbd_debug(SMB, "Received set info request\n");
 
-	rsp_org = work->response_buf;
 	if (work->next_smb2_rcv_hdr_off) {
 		req = ksmbd_req_buf_next(work);
 		rsp = ksmbd_resp_buf_next(work);
@@ -5961,8 +5970,8 @@ int smb2_set_info(struct ksmbd_work *work)
 			pid = work->compound_pfid;
 		}
 	} else {
-		req = work->request_buf;
-		rsp = work->response_buf;
+		req = smb2_get_msg(work->request_buf);
+		rsp = smb2_get_msg(work->response_buf);
 	}
 
 	if (!has_file_id(id)) {
@@ -6002,7 +6011,7 @@ int smb2_set_info(struct ksmbd_work *work)
 		goto err_out;
 
 	rsp->StructureSize = cpu_to_le16(2);
-	inc_rfc1001_len(rsp_org, 2);
+	inc_rfc1001_len(work->response_buf, 2);
 	ksmbd_fd_put(work, fp);
 	return 0;
 
@@ -6042,12 +6051,12 @@ static noinline int smb2_read_pipe(struct ksmbd_work *work)
 	int nbytes = 0, err;
 	u64 id;
 	struct ksmbd_rpc_command *rpc_resp;
-	struct smb2_read_req *req = work->request_buf;
-	struct smb2_read_rsp *rsp = work->response_buf;
+	struct smb2_read_req *req = smb2_get_msg(work->request_buf);
+	struct smb2_read_rsp *rsp = smb2_get_msg(work->response_buf);
 
 	id = le64_to_cpu(req->VolatileFileId);
 
-	inc_rfc1001_len(rsp, 16);
+	inc_rfc1001_len(work->response_buf, 16);
 	rpc_resp = ksmbd_rpc_read(work->sess, id);
 	if (rpc_resp) {
 		if (rpc_resp->flags != KSMBD_RPC_OK) {
@@ -6066,7 +6075,7 @@ static noinline int smb2_read_pipe(struct ksmbd_work *work)
 		       rpc_resp->payload_sz);
 
 		nbytes = rpc_resp->payload_sz;
-		work->resp_hdr_sz = get_rfc1002_len(rsp) + 4;
+		work->resp_hdr_sz = get_rfc1002_len(work->response_buf) + 4;
 		work->aux_payload_sz = nbytes;
 		kvfree(rpc_resp);
 	}
@@ -6076,8 +6085,8 @@ static noinline int smb2_read_pipe(struct ksmbd_work *work)
 	rsp->Reserved = 0;
 	rsp->DataLength = cpu_to_le32(nbytes);
 	rsp->DataRemaining = 0;
-	rsp->Reserved2 = 0;
-	inc_rfc1001_len(rsp, nbytes);
+	rsp->Flags = 0;
+	inc_rfc1001_len(work->response_buf, nbytes);
 	return 0;
 
 out:
@@ -6127,14 +6136,13 @@ int smb2_read(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
 	struct smb2_read_req *req;
-	struct smb2_read_rsp *rsp, *rsp_org;
+	struct smb2_read_rsp *rsp;
 	struct ksmbd_file *fp;
 	loff_t offset;
 	size_t length, mincount;
 	ssize_t nbytes = 0, remain_bytes = 0;
 	int err = 0;
 
-	rsp_org = work->response_buf;
 	WORK_BUFFERS(work, req, rsp);
 
 	if (test_share_config_flag(work->tcon->share_conf,
@@ -6215,11 +6223,11 @@ int smb2_read(struct ksmbd_work *work)
 	rsp->Reserved = 0;
 	rsp->DataLength = cpu_to_le32(nbytes);
 	rsp->DataRemaining = cpu_to_le32(remain_bytes);
-	rsp->Reserved2 = 0;
-	inc_rfc1001_len(rsp_org, 16);
-	work->resp_hdr_sz = get_rfc1002_len(rsp_org) + 4;
+	rsp->Flags = 0;
+	inc_rfc1001_len(work->response_buf, 16);
+	work->resp_hdr_sz = get_rfc1002_len(work->response_buf) + 4;
 	work->aux_payload_sz = nbytes;
-	inc_rfc1001_len(rsp_org, nbytes);
+	inc_rfc1001_len(work->response_buf, nbytes);
 	ksmbd_fd_put(work, fp);
 	return 0;
 
@@ -6254,8 +6262,8 @@ int smb2_read(struct ksmbd_work *work)
  */
 static noinline int smb2_write_pipe(struct ksmbd_work *work)
 {
-	struct smb2_write_req *req = work->request_buf;
-	struct smb2_write_rsp *rsp = work->response_buf;
+	struct smb2_write_req *req = smb2_get_msg(work->request_buf);
+	struct smb2_write_rsp *rsp = smb2_get_msg(work->response_buf);
 	struct ksmbd_rpc_command *rpc_resp;
 	u64 id = 0;
 	int err = 0, ret = 0;
@@ -6266,13 +6274,14 @@ static noinline int smb2_write_pipe(struct ksmbd_work *work)
 	id = le64_to_cpu(req->VolatileFileId);
 
 	if (le16_to_cpu(req->DataOffset) ==
-	    (offsetof(struct smb2_write_req, Buffer) - 4)) {
+	    offsetof(struct smb2_write_req, Buffer)) {
 		data_buf = (char *)&req->Buffer[0];
 	} else {
-		if ((u64)le16_to_cpu(req->DataOffset) + length > get_rfc1002_len(req)) {
+		if ((u64)le16_to_cpu(req->DataOffset) + length >
+		    get_rfc1002_len(work->request_buf)) {
 			pr_err("invalid write data offset %u, smb_len %u\n",
 			       le16_to_cpu(req->DataOffset),
-			       get_rfc1002_len(req));
+			       get_rfc1002_len(work->request_buf));
 			err = -EINVAL;
 			goto out;
 		}
@@ -6304,7 +6313,7 @@ static noinline int smb2_write_pipe(struct ksmbd_work *work)
 	rsp->DataLength = cpu_to_le32(length);
 	rsp->DataRemaining = 0;
 	rsp->Reserved2 = 0;
-	inc_rfc1001_len(rsp, 16);
+	inc_rfc1001_len(work->response_buf, 16);
 	return 0;
 out:
 	if (err) {
@@ -6372,7 +6381,7 @@ static ssize_t smb2_write_rdma_channel(struct ksmbd_work *work,
 int smb2_write(struct ksmbd_work *work)
 {
 	struct smb2_write_req *req;
-	struct smb2_write_rsp *rsp, *rsp_org;
+	struct smb2_write_rsp *rsp;
 	struct ksmbd_file *fp = NULL;
 	loff_t offset;
 	size_t length;
@@ -6381,7 +6390,6 @@ int smb2_write(struct ksmbd_work *work)
 	bool writethrough = false;
 	int err = 0;
 
-	rsp_org = work->response_buf;
 	WORK_BUFFERS(work, req, rsp);
 
 	if (test_share_config_flag(work->tcon->share_conf, KSMBD_SHARE_FLAG_PIPE)) {
@@ -6424,13 +6432,14 @@ int smb2_write(struct ksmbd_work *work)
 	if (req->Channel != SMB2_CHANNEL_RDMA_V1 &&
 	    req->Channel != SMB2_CHANNEL_RDMA_V1_INVALIDATE) {
 		if (le16_to_cpu(req->DataOffset) ==
-		    (offsetof(struct smb2_write_req, Buffer) - 4)) {
+		    offsetof(struct smb2_write_req, Buffer)) {
 			data_buf = (char *)&req->Buffer[0];
 		} else {
-			if ((u64)le16_to_cpu(req->DataOffset) + length > get_rfc1002_len(req)) {
+			if ((u64)le16_to_cpu(req->DataOffset) + length >
+			    get_rfc1002_len(work->request_buf)) {
 				pr_err("invalid write data offset %u, smb_len %u\n",
 				       le16_to_cpu(req->DataOffset),
-				       get_rfc1002_len(req));
+				       get_rfc1002_len(work->request_buf));
 				err = -EINVAL;
 				goto out;
 			}
@@ -6468,7 +6477,7 @@ int smb2_write(struct ksmbd_work *work)
 	rsp->DataLength = cpu_to_le32(nbytes);
 	rsp->DataRemaining = 0;
 	rsp->Reserved2 = 0;
-	inc_rfc1001_len(rsp_org, 16);
+	inc_rfc1001_len(work->response_buf, 16);
 	ksmbd_fd_put(work, fp);
 	return 0;
 
@@ -6502,10 +6511,9 @@ int smb2_write(struct ksmbd_work *work)
 int smb2_flush(struct ksmbd_work *work)
 {
 	struct smb2_flush_req *req;
-	struct smb2_flush_rsp *rsp, *rsp_org;
+	struct smb2_flush_rsp *rsp;
 	int err;
 
-	rsp_org = work->response_buf;
 	WORK_BUFFERS(work, req, rsp);
 
 	ksmbd_debug(SMB, "SMB2_FLUSH called for fid %llu\n",
@@ -6519,7 +6527,7 @@ int smb2_flush(struct ksmbd_work *work)
 
 	rsp->StructureSize = cpu_to_le16(4);
 	rsp->Reserved = 0;
-	inc_rfc1001_len(rsp_org, 4);
+	inc_rfc1001_len(work->response_buf, 4);
 	return 0;
 
 out:
@@ -6540,7 +6548,7 @@ int smb2_flush(struct ksmbd_work *work)
 int smb2_cancel(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
-	struct smb2_hdr *hdr = work->request_buf;
+	struct smb2_hdr *hdr = smb2_get_msg(work->request_buf);
 	struct smb2_hdr *chdr;
 	struct ksmbd_work *cancel_work = NULL;
 	int canceled = 0;
@@ -6555,7 +6563,7 @@ int smb2_cancel(struct ksmbd_work *work)
 		spin_lock(&conn->request_lock);
 		list_for_each_entry(cancel_work, command_list,
 				    async_request_entry) {
-			chdr = cancel_work->request_buf;
+			chdr = smb2_get_msg(cancel_work->request_buf);
 
 			if (cancel_work->async_id !=
 			    le64_to_cpu(hdr->Id.AsyncId))
@@ -6574,7 +6582,7 @@ int smb2_cancel(struct ksmbd_work *work)
 
 		spin_lock(&conn->request_lock);
 		list_for_each_entry(cancel_work, command_list, request_entry) {
-			chdr = cancel_work->request_buf;
+			chdr = smb2_get_msg(cancel_work->request_buf);
 
 			if (chdr->MessageId != hdr->MessageId ||
 			    cancel_work == work)
@@ -6709,8 +6717,8 @@ static inline bool lock_defer_pending(struct file_lock *fl)
  */
 int smb2_lock(struct ksmbd_work *work)
 {
-	struct smb2_lock_req *req = work->request_buf;
-	struct smb2_lock_rsp *rsp = work->response_buf;
+	struct smb2_lock_req *req = smb2_get_msg(work->request_buf);
+	struct smb2_lock_rsp *rsp = smb2_get_msg(work->response_buf);
 	struct smb2_lock_element *lock_ele;
 	struct ksmbd_file *fp = NULL;
 	struct file_lock *flock = NULL;
@@ -7017,7 +7025,7 @@ int smb2_lock(struct ksmbd_work *work)
 	ksmbd_debug(SMB, "successful in taking lock\n");
 	rsp->hdr.Status = STATUS_SUCCESS;
 	rsp->Reserved = 0;
-	inc_rfc1001_len(rsp, 4);
+	inc_rfc1001_len(work->response_buf, 4);
 	ksmbd_fd_put(work, fp);
 	return 0;
 
@@ -7312,7 +7320,7 @@ static int fsctl_validate_negotiate_info(struct ksmbd_conn *conn,
 	int ret = 0;
 	int dialect;
 
-	if (in_buf_len < sizeof(struct validate_negotiate_info_req) +
+	if (in_buf_len < offsetof(struct validate_negotiate_info_req, Dialects) +
 			le16_to_cpu(neg_req->DialectCount) * sizeof(__le16))
 		return -EINVAL;
 
@@ -7435,9 +7443,9 @@ static inline int fsctl_set_sparse(struct ksmbd_work *work, u64 id,
 
 	old_fattr = fp->f_ci->m_fattr;
 	if (sparse->SetSparse)
-		fp->f_ci->m_fattr |= ATTR_SPARSE_FILE_LE;
+		fp->f_ci->m_fattr |= FILE_ATTRIBUTE_SPARSE_FILE_LE;
 	else
-		fp->f_ci->m_fattr &= ~ATTR_SPARSE_FILE_LE;
+		fp->f_ci->m_fattr &= ~FILE_ATTRIBUTE_SPARSE_FILE_LE;
 
 	if (fp->f_ci->m_fattr != old_fattr &&
 	    test_share_config_flag(work->tcon->share_conf,
@@ -7490,13 +7498,12 @@ static int fsctl_request_resume_key(struct ksmbd_work *work,
 int smb2_ioctl(struct ksmbd_work *work)
 {
 	struct smb2_ioctl_req *req;
-	struct smb2_ioctl_rsp *rsp, *rsp_org;
+	struct smb2_ioctl_rsp *rsp;
 	unsigned int cnt_code, nbytes = 0, out_buf_len, in_buf_len;
 	u64 id = KSMBD_NO_FID;
 	struct ksmbd_conn *conn = work->conn;
 	int ret = 0;
 
-	rsp_org = work->response_buf;
 	if (work->next_smb2_rcv_hdr_off) {
 		req = ksmbd_req_buf_next(work);
 		rsp = ksmbd_resp_buf_next(work);
@@ -7506,8 +7513,8 @@ int smb2_ioctl(struct ksmbd_work *work)
 			id = work->compound_fid;
 		}
 	} else {
-		req = work->request_buf;
-		rsp = work->response_buf;
+		req = smb2_get_msg(work->request_buf);
+		rsp = smb2_get_msg(work->response_buf);
 	}
 
 	if (!has_file_id(id))
@@ -7787,7 +7794,7 @@ int smb2_ioctl(struct ksmbd_work *work)
 	rsp->Reserved = cpu_to_le16(0);
 	rsp->Flags = cpu_to_le32(0);
 	rsp->Reserved2 = cpu_to_le32(0);
-	inc_rfc1001_len(rsp_org, 48 + nbytes);
+	inc_rfc1001_len(work->response_buf, 48 + nbytes);
 
 	return 0;
 
@@ -7814,8 +7821,8 @@ int smb2_ioctl(struct ksmbd_work *work)
  */
 static void smb20_oplock_break_ack(struct ksmbd_work *work)
 {
-	struct smb2_oplock_break *req = work->request_buf;
-	struct smb2_oplock_break *rsp = work->response_buf;
+	struct smb2_oplock_break *req = smb2_get_msg(work->request_buf);
+	struct smb2_oplock_break *rsp = smb2_get_msg(work->response_buf);
 	struct ksmbd_file *fp;
 	struct oplock_info *opinfo = NULL;
 	__le32 err = 0;
@@ -7922,7 +7929,7 @@ static void smb20_oplock_break_ack(struct ksmbd_work *work)
 	rsp->Reserved2 = 0;
 	rsp->VolatileFid = cpu_to_le64(volatile_id);
 	rsp->PersistentFid = cpu_to_le64(persistent_id);
-	inc_rfc1001_len(rsp, 24);
+	inc_rfc1001_len(work->response_buf, 24);
 	return;
 
 err_out:
@@ -7958,8 +7965,8 @@ static int check_lease_state(struct lease *lease, __le32 req_state)
 static void smb21_lease_break_ack(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
-	struct smb2_lease_ack *req = work->request_buf;
-	struct smb2_lease_ack *rsp = work->response_buf;
+	struct smb2_lease_ack *req = smb2_get_msg(work->request_buf);
+	struct smb2_lease_ack *rsp = smb2_get_msg(work->response_buf);
 	struct oplock_info *opinfo;
 	__le32 err = 0;
 	int ret = 0;
@@ -8071,7 +8078,7 @@ static void smb21_lease_break_ack(struct ksmbd_work *work)
 	memcpy(rsp->LeaseKey, req->LeaseKey, 16);
 	rsp->LeaseState = lease_state;
 	rsp->LeaseDuration = 0;
-	inc_rfc1001_len(rsp, 36);
+	inc_rfc1001_len(work->response_buf, 36);
 	return;
 
 err_out:
@@ -8092,8 +8099,8 @@ static void smb21_lease_break_ack(struct ksmbd_work *work)
  */
 int smb2_oplock_break(struct ksmbd_work *work)
 {
-	struct smb2_oplock_break *req = work->request_buf;
-	struct smb2_oplock_break *rsp = work->response_buf;
+	struct smb2_oplock_break *req = smb2_get_msg(work->request_buf);
+	struct smb2_oplock_break *rsp = smb2_get_msg(work->response_buf);
 
 	switch (le16_to_cpu(req->StructureSize)) {
 	case OP_BREAK_STRUCT_SIZE_20:
@@ -8120,8 +8127,8 @@ int smb2_oplock_break(struct ksmbd_work *work)
  */
 int smb2_notify(struct ksmbd_work *work)
 {
-	struct smb2_notify_req *req;
-	struct smb2_notify_rsp *rsp;
+	struct smb2_change_notify_req *req;
+	struct smb2_change_notify_rsp *rsp;
 
 	WORK_BUFFERS(work, req, rsp);
 
@@ -8145,7 +8152,7 @@ int smb2_notify(struct ksmbd_work *work)
  */
 bool smb2_is_sign_req(struct ksmbd_work *work, unsigned int command)
 {
-	struct smb2_hdr *rcv_hdr2 = work->request_buf;
+	struct smb2_hdr *rcv_hdr2 = smb2_get_msg(work->request_buf);
 
 	if ((rcv_hdr2->Flags & SMB2_FLAGS_SIGNED) &&
 	    command != SMB2_NEGOTIATE_HE &&
@@ -8164,22 +8171,22 @@ bool smb2_is_sign_req(struct ksmbd_work *work, unsigned int command)
  */
 int smb2_check_sign_req(struct ksmbd_work *work)
 {
-	struct smb2_hdr *hdr, *hdr_org;
+	struct smb2_hdr *hdr;
 	char signature_req[SMB2_SIGNATURE_SIZE];
 	char signature[SMB2_HMACSHA256_SIZE];
 	struct kvec iov[1];
 	size_t len;
 
-	hdr_org = hdr = work->request_buf;
+	hdr = smb2_get_msg(work->request_buf);
 	if (work->next_smb2_rcv_hdr_off)
 		hdr = ksmbd_req_buf_next(work);
 
 	if (!hdr->NextCommand && !work->next_smb2_rcv_hdr_off)
-		len = be32_to_cpu(hdr_org->smb2_buf_length);
+		len = get_rfc1002_len(work->request_buf);
 	else if (hdr->NextCommand)
 		len = le32_to_cpu(hdr->NextCommand);
 	else
-		len = be32_to_cpu(hdr_org->smb2_buf_length) -
+		len = get_rfc1002_len(work->request_buf) -
 			work->next_smb2_rcv_hdr_off;
 
 	memcpy(signature_req, hdr->Signature, SMB2_SIGNATURE_SIZE);
@@ -8207,25 +8214,26 @@ int smb2_check_sign_req(struct ksmbd_work *work)
  */
 void smb2_set_sign_rsp(struct ksmbd_work *work)
 {
-	struct smb2_hdr *hdr, *hdr_org;
+	struct smb2_hdr *hdr;
 	struct smb2_hdr *req_hdr;
 	char signature[SMB2_HMACSHA256_SIZE];
 	struct kvec iov[2];
 	size_t len;
 	int n_vec = 1;
 
-	hdr_org = hdr = work->response_buf;
+	hdr = smb2_get_msg(work->response_buf);
 	if (work->next_smb2_rsp_hdr_off)
 		hdr = ksmbd_resp_buf_next(work);
 
 	req_hdr = ksmbd_req_buf_next(work);
 
 	if (!work->next_smb2_rsp_hdr_off) {
-		len = get_rfc1002_len(hdr_org);
+		len = get_rfc1002_len(work->response_buf);
 		if (req_hdr->NextCommand)
 			len = ALIGN(len, 8);
 	} else {
-		len = get_rfc1002_len(hdr_org) - work->next_smb2_rsp_hdr_off;
+		len = get_rfc1002_len(work->response_buf) -
+			work->next_smb2_rsp_hdr_off;
 		len = ALIGN(len, 8);
 	}
 
@@ -8261,23 +8269,23 @@ int smb3_check_sign_req(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
 	char *signing_key;
-	struct smb2_hdr *hdr, *hdr_org;
+	struct smb2_hdr *hdr;
 	struct channel *chann;
 	char signature_req[SMB2_SIGNATURE_SIZE];
 	char signature[SMB2_CMACAES_SIZE];
 	struct kvec iov[1];
 	size_t len;
 
-	hdr_org = hdr = work->request_buf;
+	hdr = smb2_get_msg(work->request_buf);
 	if (work->next_smb2_rcv_hdr_off)
 		hdr = ksmbd_req_buf_next(work);
 
 	if (!hdr->NextCommand && !work->next_smb2_rcv_hdr_off)
-		len = be32_to_cpu(hdr_org->smb2_buf_length);
+		len = get_rfc1002_len(work->request_buf);
 	else if (hdr->NextCommand)
 		len = le32_to_cpu(hdr->NextCommand);
 	else
-		len = be32_to_cpu(hdr_org->smb2_buf_length) -
+		len = get_rfc1002_len(work->request_buf) -
 			work->next_smb2_rcv_hdr_off;
 
 	if (le16_to_cpu(hdr->Command) == SMB2_SESSION_SETUP_HE) {
@@ -8318,8 +8326,7 @@ int smb3_check_sign_req(struct ksmbd_work *work)
 void smb3_set_sign_rsp(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
-	struct smb2_hdr *req_hdr;
-	struct smb2_hdr *hdr, *hdr_org;
+	struct smb2_hdr *req_hdr, *hdr;
 	struct channel *chann;
 	char signature[SMB2_CMACAES_SIZE];
 	struct kvec iov[2];
@@ -8327,18 +8334,19 @@ void smb3_set_sign_rsp(struct ksmbd_work *work)
 	size_t len;
 	char *signing_key;
 
-	hdr_org = hdr = work->response_buf;
+	hdr = smb2_get_msg(work->response_buf);
 	if (work->next_smb2_rsp_hdr_off)
 		hdr = ksmbd_resp_buf_next(work);
 
 	req_hdr = ksmbd_req_buf_next(work);
 
 	if (!work->next_smb2_rsp_hdr_off) {
-		len = get_rfc1002_len(hdr_org);
+		len = get_rfc1002_len(work->response_buf);
 		if (req_hdr->NextCommand)
 			len = ALIGN(len, 8);
 	} else {
-		len = get_rfc1002_len(hdr_org) - work->next_smb2_rsp_hdr_off;
+		len = get_rfc1002_len(work->response_buf) -
+			work->next_smb2_rsp_hdr_off;
 		len = ALIGN(len, 8);
 	}
 
@@ -8391,7 +8399,7 @@ void smb3_preauth_hash_rsp(struct ksmbd_work *work)
 
 	if (le16_to_cpu(req->Command) == SMB2_NEGOTIATE_HE &&
 	    conn->preauth_info)
-		ksmbd_gen_preauth_integrity_hash(conn, (char *)rsp,
+		ksmbd_gen_preauth_integrity_hash(conn, work->response_buf,
 						 conn->preauth_info->Preauth_HashValue);
 
 	if (le16_to_cpu(rsp->Command) == SMB2_SESSION_SETUP_HE && sess) {
@@ -8409,35 +8417,34 @@ void smb3_preauth_hash_rsp(struct ksmbd_work *work)
 			if (!hash_value)
 				return;
 		}
-		ksmbd_gen_preauth_integrity_hash(conn, (char *)rsp,
+		ksmbd_gen_preauth_integrity_hash(conn, work->response_buf,
 						 hash_value);
 	}
 }
 
-static void fill_transform_hdr(struct smb2_transform_hdr *tr_hdr, char *old_buf,
-			       __le16 cipher_type)
+static void fill_transform_hdr(void *tr_buf, char *old_buf, __le16 cipher_type)
 {
-	struct smb2_hdr *hdr = (struct smb2_hdr *)old_buf;
+	struct smb2_transform_hdr *tr_hdr = tr_buf + 4;
+	struct smb2_hdr *hdr = smb2_get_msg(old_buf);
 	unsigned int orig_len = get_rfc1002_len(old_buf);
 
-	memset(tr_hdr, 0, sizeof(struct smb2_transform_hdr));
+	memset(tr_buf, 0, sizeof(struct smb2_transform_hdr) + 4);
 	tr_hdr->ProtocolId = SMB2_TRANSFORM_PROTO_NUM;
 	tr_hdr->OriginalMessageSize = cpu_to_le32(orig_len);
-	tr_hdr->Flags = cpu_to_le16(0x01);
+	tr_hdr->Flags = cpu_to_le16(TRANSFORM_FLAG_ENCRYPTED);
 	if (cipher_type == SMB2_ENCRYPTION_AES128_GCM ||
 	    cipher_type == SMB2_ENCRYPTION_AES256_GCM)
 		get_random_bytes(&tr_hdr->Nonce, SMB3_AES_GCM_NONCE);
 	else
 		get_random_bytes(&tr_hdr->Nonce, SMB3_AES_CCM_NONCE);
 	memcpy(&tr_hdr->SessionId, &hdr->SessionId, 8);
-	inc_rfc1001_len(tr_hdr, sizeof(struct smb2_transform_hdr) - 4);
-	inc_rfc1001_len(tr_hdr, orig_len);
+	inc_rfc1001_len(tr_buf, sizeof(struct smb2_transform_hdr));
+	inc_rfc1001_len(tr_buf, orig_len);
 }
 
 int smb3_encrypt_resp(struct ksmbd_work *work)
 {
 	char *buf = work->response_buf;
-	struct smb2_transform_hdr *tr_hdr;
 	struct kvec iov[3];
 	int rc = -ENOMEM;
 	int buf_size = 0, rq_nvec = 2 + (work->aux_payload_sz ? 1 : 0);
@@ -8445,15 +8452,15 @@ int smb3_encrypt_resp(struct ksmbd_work *work)
 	if (ARRAY_SIZE(iov) < rq_nvec)
 		return -ENOMEM;
 
-	tr_hdr = kzalloc(sizeof(struct smb2_transform_hdr), GFP_KERNEL);
-	if (!tr_hdr)
+	work->tr_buf = kzalloc(sizeof(struct smb2_transform_hdr) + 4, GFP_KERNEL);
+	if (!work->tr_buf)
 		return rc;
 
 	/* fill transform header */
-	fill_transform_hdr(tr_hdr, buf, work->conn->cipher_type);
+	fill_transform_hdr(work->tr_buf, buf, work->conn->cipher_type);
 
-	iov[0].iov_base = tr_hdr;
-	iov[0].iov_len = sizeof(struct smb2_transform_hdr);
+	iov[0].iov_base = work->tr_buf;
+	iov[0].iov_len = sizeof(struct smb2_transform_hdr) + 4;
 	buf_size += iov[0].iov_len - 4;
 
 	iov[1].iov_base = buf + 4;
@@ -8473,15 +8480,14 @@ int smb3_encrypt_resp(struct ksmbd_work *work)
 		return rc;
 
 	memmove(buf, iov[1].iov_base, iov[1].iov_len);
-	tr_hdr->smb2_buf_length = cpu_to_be32(buf_size);
-	work->tr_buf = tr_hdr;
+	*(__be32 *)work->tr_buf = cpu_to_be32(buf_size);
 
 	return rc;
 }
 
 bool smb3_is_transform_hdr(void *buf)
 {
-	struct smb2_transform_hdr *trhdr = buf;
+	struct smb2_transform_hdr *trhdr = smb2_get_msg(buf);
 
 	return trhdr->ProtocolId == SMB2_TRANSFORM_PROTO_NUM;
 }
@@ -8491,12 +8497,10 @@ int smb3_decrypt_req(struct ksmbd_work *work)
 	struct ksmbd_conn *conn = work->conn;
 	struct ksmbd_session *sess;
 	char *buf = work->request_buf;
-	struct smb2_hdr *hdr;
 	unsigned int pdu_length = get_rfc1002_len(buf);
 	struct kvec iov[2];
-	int buf_data_size = pdu_length + 4 -
-		sizeof(struct smb2_transform_hdr);
-	struct smb2_transform_hdr *tr_hdr = (struct smb2_transform_hdr *)buf;
+	int buf_data_size = pdu_length - sizeof(struct smb2_transform_hdr);
+	struct smb2_transform_hdr *tr_hdr = smb2_get_msg(buf);
 	int rc = 0;
 
 	if (buf_data_size < sizeof(struct smb2_hdr)) {
@@ -8518,16 +8522,15 @@ int smb3_decrypt_req(struct ksmbd_work *work)
 	}
 
 	iov[0].iov_base = buf;
-	iov[0].iov_len = sizeof(struct smb2_transform_hdr);
-	iov[1].iov_base = buf + sizeof(struct smb2_transform_hdr);
+	iov[0].iov_len = sizeof(struct smb2_transform_hdr) + 4;
+	iov[1].iov_base = buf + sizeof(struct smb2_transform_hdr) + 4;
 	iov[1].iov_len = buf_data_size;
 	rc = ksmbd_crypt_message(conn, iov, 2, 0);
 	if (rc)
 		return rc;
 
 	memmove(buf + 4, iov[1].iov_base, buf_data_size);
-	hdr = (struct smb2_hdr *)buf;
-	hdr->smb2_buf_length = cpu_to_be32(buf_data_size);
+	*(__be32 *)buf = cpu_to_be32(buf_data_size);
 
 	return rc;
 }
@@ -8535,7 +8538,7 @@ int smb3_decrypt_req(struct ksmbd_work *work)
 bool smb3_11_final_sess_setup_resp(struct ksmbd_work *work)
 {
 	struct ksmbd_conn *conn = work->conn;
-	struct smb2_hdr *rsp = work->response_buf;
+	struct smb2_hdr *rsp = smb2_get_msg(work->response_buf);
 
 	if (conn->dialect < SMB30_PROT_ID)
 		return false;
diff --git a/fs/ksmbd/smb2pdu.h b/fs/ksmbd/smb2pdu.h
index ff5a2f0..4a3e433 100644
--- a/fs/ksmbd/smb2pdu.h
+++ b/fs/ksmbd/smb2pdu.h
@@ -10,60 +10,6 @@
 #include "ntlmssp.h"
 #include "smbacl.h"
 
-/*
- * Note that, due to trying to use names similar to the protocol specifications,
- * there are many mixed case field names in the structures below.  Although
- * this does not match typical Linux kernel style, it is necessary to be
- * able to match against the protocol specfication.
- *
- * SMB2 commands
- * Some commands have minimal (wct=0,bcc=0), or uninteresting, responses
- * (ie no useful data other than the SMB error code itself) and are marked such.
- * Knowing this helps avoid response buffer allocations and copy in some cases.
- */
-
-/* List of commands in host endian */
-#define SMB2_NEGOTIATE_HE	0x0000
-#define SMB2_SESSION_SETUP_HE	0x0001
-#define SMB2_LOGOFF_HE		0x0002 /* trivial request/resp */
-#define SMB2_TREE_CONNECT_HE	0x0003
-#define SMB2_TREE_DISCONNECT_HE	0x0004 /* trivial req/resp */
-#define SMB2_CREATE_HE		0x0005
-#define SMB2_CLOSE_HE		0x0006
-#define SMB2_FLUSH_HE		0x0007 /* trivial resp */
-#define SMB2_READ_HE		0x0008
-#define SMB2_WRITE_HE		0x0009
-#define SMB2_LOCK_HE		0x000A
-#define SMB2_IOCTL_HE		0x000B
-#define SMB2_CANCEL_HE		0x000C
-#define SMB2_ECHO_HE		0x000D
-#define SMB2_QUERY_DIRECTORY_HE	0x000E
-#define SMB2_CHANGE_NOTIFY_HE	0x000F
-#define SMB2_QUERY_INFO_HE	0x0010
-#define SMB2_SET_INFO_HE	0x0011
-#define SMB2_OPLOCK_BREAK_HE	0x0012
-
-/* The same list in little endian */
-#define SMB2_NEGOTIATE		cpu_to_le16(SMB2_NEGOTIATE_HE)
-#define SMB2_SESSION_SETUP	cpu_to_le16(SMB2_SESSION_SETUP_HE)
-#define SMB2_LOGOFF		cpu_to_le16(SMB2_LOGOFF_HE)
-#define SMB2_TREE_CONNECT	cpu_to_le16(SMB2_TREE_CONNECT_HE)
-#define SMB2_TREE_DISCONNECT	cpu_to_le16(SMB2_TREE_DISCONNECT_HE)
-#define SMB2_CREATE		cpu_to_le16(SMB2_CREATE_HE)
-#define SMB2_CLOSE		cpu_to_le16(SMB2_CLOSE_HE)
-#define SMB2_FLUSH		cpu_to_le16(SMB2_FLUSH_HE)
-#define SMB2_READ		cpu_to_le16(SMB2_READ_HE)
-#define SMB2_WRITE		cpu_to_le16(SMB2_WRITE_HE)
-#define SMB2_LOCK		cpu_to_le16(SMB2_LOCK_HE)
-#define SMB2_IOCTL		cpu_to_le16(SMB2_IOCTL_HE)
-#define SMB2_CANCEL		cpu_to_le16(SMB2_CANCEL_HE)
-#define SMB2_ECHO		cpu_to_le16(SMB2_ECHO_HE)
-#define SMB2_QUERY_DIRECTORY	cpu_to_le16(SMB2_QUERY_DIRECTORY_HE)
-#define SMB2_CHANGE_NOTIFY	cpu_to_le16(SMB2_CHANGE_NOTIFY_HE)
-#define SMB2_QUERY_INFO		cpu_to_le16(SMB2_QUERY_INFO_HE)
-#define SMB2_SET_INFO		cpu_to_le16(SMB2_SET_INFO_HE)
-#define SMB2_OPLOCK_BREAK	cpu_to_le16(SMB2_OPLOCK_BREAK_HE)
-
 /*Create Action Flags*/
 #define FILE_SUPERSEDED                0x00000000
 #define FILE_OPENED            0x00000001
@@ -96,9 +42,6 @@
 /* SMB2 Max Credits */
 #define SMB2_MAX_CREDITS		8192
 
-#define SMB2_CLIENT_GUID_SIZE		16
-#define SMB2_CREATE_GUID_SIZE		16
-
 /* Maximum buffer size value we can send with 1 credit */
 #define SMB2_MAX_BUFFER_SIZE 65536
 
@@ -107,9 +50,6 @@
 /* BB FIXME - analyze following length BB */
 #define MAX_SMB2_HDR_SIZE 0x78 /* 4 len + 64 hdr + (2*24 wct) + 2 bct + 2 pad */
 
-#define SMB2_PROTO_NUMBER cpu_to_le32(0x424d53fe) /* 'B''M''S' */
-#define SMB2_TRANSFORM_PROTO_NUM cpu_to_le32(0x424d53fd)
-
 #define SMB21_DEFAULT_IOSIZE	(1024 * 1024)
 #define SMB3_DEFAULT_IOSIZE	(4 * 1024 * 1024)
 #define SMB3_DEFAULT_TRANS_SIZE	(1024 * 1024)
@@ -117,78 +57,6 @@
 #define SMB3_MAX_IOSIZE	(8 * 1024 * 1024)
 
 /*
- * SMB2 Header Definition
- *
- * "MBZ" :  Must be Zero
- * "BB"  :  BugBug, Something to check/review/analyze later
- * "PDU" :  "Protocol Data Unit" (ie a network "frame")
- *
- */
-
-#define __SMB2_HEADER_STRUCTURE_SIZE	64
-#define SMB2_HEADER_STRUCTURE_SIZE				\
-	cpu_to_le16(__SMB2_HEADER_STRUCTURE_SIZE)
-
-struct smb2_hdr {
-	__be32 smb2_buf_length;	/* big endian on wire */
-				/*
-				 * length is only two or three bytes - with
-				 * one or two byte type preceding it that MBZ
-				 */
-	__le32 ProtocolId;	/* 0xFE 'S' 'M' 'B' */
-	__le16 StructureSize;	/* 64 */
-	__le16 CreditCharge;	/* MBZ */
-	__le32 Status;		/* Error from server */
-	__le16 Command;
-	__le16 CreditRequest;	/* CreditResponse */
-	__le32 Flags;
-	__le32 NextCommand;
-	__le64 MessageId;
-	union {
-		struct {
-			__le32 ProcessId;
-			__le32  TreeId;
-		} __packed SyncId;
-		__le64  AsyncId;
-	} __packed Id;
-	__le64  SessionId;
-	__u8   Signature[16];
-} __packed;
-
-struct smb2_pdu {
-	struct smb2_hdr hdr;
-	__le16 StructureSize2; /* size of wct area (varies, request specific) */
-} __packed;
-
-#define SMB3_AES_CCM_NONCE 11
-#define SMB3_AES_GCM_NONCE 12
-
-struct smb2_transform_hdr {
-	__be32 smb2_buf_length; /* big endian on wire */
-	/*
-	 * length is only two or three bytes - with
-	 * one or two byte type preceding it that MBZ
-	 */
-	__le32 ProtocolId;      /* 0xFD 'S' 'M' 'B' */
-	__u8   Signature[16];
-	__u8   Nonce[16];
-	__le32 OriginalMessageSize;
-	__u16  Reserved1;
-	__le16 Flags; /* EncryptionAlgorithm */
-	__le64  SessionId;
-} __packed;
-
-/*
- *	SMB2 flag definitions
- */
-#define SMB2_FLAGS_SERVER_TO_REDIR	cpu_to_le32(0x00000001)
-#define SMB2_FLAGS_ASYNC_COMMAND	cpu_to_le32(0x00000002)
-#define SMB2_FLAGS_RELATED_OPERATIONS	cpu_to_le32(0x00000004)
-#define SMB2_FLAGS_SIGNED		cpu_to_le32(0x00000008)
-#define SMB2_FLAGS_DFS_OPERATIONS	cpu_to_le32(0x10000000)
-#define SMB2_FLAGS_REPLAY_OPERATIONS	cpu_to_le32(0x20000000)
-
-/*
  *	Definitions for SMB2 Protocol Data Units (network frames)
  *
  *  See MS-SMB2.PDF specification for protocol details.
@@ -209,425 +77,30 @@ struct smb2_err_rsp {
 	__u8   ErrorData[1];  /* variable length */
 } __packed;
 
-struct smb2_negotiate_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize; /* Must be 36 */
-	__le16 DialectCount;
-	__le16 SecurityMode;
-	__le16 Reserved;	/* MBZ */
-	__le32 Capabilities;
-	__u8   ClientGUID[SMB2_CLIENT_GUID_SIZE];
-	/* In SMB3.02 and earlier next three were MBZ le64 ClientStartTime */
-	__le32 NegotiateContextOffset; /* SMB3.1.1 only. MBZ earlier */
-	__le16 NegotiateContextCount;  /* SMB3.1.1 only. MBZ earlier */
-	__le16 Reserved2;
-	__le16 Dialects[1]; /* One dialect (vers=) at a time for now */
-} __packed;
-
-/* SecurityMode flags */
-#define SMB2_NEGOTIATE_SIGNING_ENABLED_LE	cpu_to_le16(0x0001)
-#define SMB2_NEGOTIATE_SIGNING_REQUIRED		0x0002
-#define SMB2_NEGOTIATE_SIGNING_REQUIRED_LE	cpu_to_le16(0x0002)
-/* Capabilities flags */
-#define SMB2_GLOBAL_CAP_DFS		0x00000001
-#define SMB2_GLOBAL_CAP_LEASING		0x00000002 /* Resp only New to SMB2.1 */
-#define SMB2_GLOBAL_CAP_LARGE_MTU	0X00000004 /* Resp only New to SMB2.1 */
-#define SMB2_GLOBAL_CAP_MULTI_CHANNEL	0x00000008 /* New to SMB3 */
-#define SMB2_GLOBAL_CAP_PERSISTENT_HANDLES 0x00000010 /* New to SMB3 */
-#define SMB2_GLOBAL_CAP_DIRECTORY_LEASING  0x00000020 /* New to SMB3 */
-#define SMB2_GLOBAL_CAP_ENCRYPTION	0x00000040 /* New to SMB3 */
-/* Internal types */
-#define SMB2_NT_FIND			0x00100000
-#define SMB2_LARGE_FILES		0x00200000
-
-#define SMB311_SALT_SIZE			32
-/* Hash Algorithm Types */
-#define SMB2_PREAUTH_INTEGRITY_SHA512	cpu_to_le16(0x0001)
-
-#define PREAUTH_HASHVALUE_SIZE		64
-
 struct preauth_integrity_info {
 	/* PreAuth integrity Hash ID */
 	__le16			Preauth_HashId;
 	/* PreAuth integrity Hash Value */
-	__u8			Preauth_HashValue[PREAUTH_HASHVALUE_SIZE];
+	__u8			Preauth_HashValue[SMB2_PREAUTH_HASH_SIZE];
 };
 
-/* offset is sizeof smb2_negotiate_rsp - 4 but rounded up to 8 bytes. */
+/* offset is sizeof smb2_negotiate_rsp but rounded up to 8 bytes. */
 #ifdef CONFIG_SMB_SERVER_KERBEROS5
-/* sizeof(struct smb2_negotiate_rsp) - 4 =
+/* sizeof(struct smb2_negotiate_rsp) =
  * header(64) + response(64) + GSS_LENGTH(96) + GSS_PADDING(0)
  */
 #define OFFSET_OF_NEG_CONTEXT	0xe0
 #else
-/* sizeof(struct smb2_negotiate_rsp) - 4 =
+/* sizeof(struct smb2_negotiate_rsp) =
  * header(64) + response(64) + GSS_LENGTH(74) + GSS_PADDING(6)
  */
 #define OFFSET_OF_NEG_CONTEXT	0xd0
 #endif
 
-#define SMB2_PREAUTH_INTEGRITY_CAPABILITIES	cpu_to_le16(1)
-#define SMB2_ENCRYPTION_CAPABILITIES		cpu_to_le16(2)
-#define SMB2_COMPRESSION_CAPABILITIES		cpu_to_le16(3)
-#define SMB2_NETNAME_NEGOTIATE_CONTEXT_ID	cpu_to_le16(5)
-#define SMB2_SIGNING_CAPABILITIES		cpu_to_le16(8)
-#define SMB2_POSIX_EXTENSIONS_AVAILABLE		cpu_to_le16(0x100)
-
-struct smb2_neg_context {
-	__le16  ContextType;
-	__le16  DataLength;
-	__le32  Reserved;
-	/* Followed by array of data */
-} __packed;
-
-struct smb2_preauth_neg_context {
-	__le16	ContextType; /* 1 */
-	__le16	DataLength;
-	__le32	Reserved;
-	__le16	HashAlgorithmCount; /* 1 */
-	__le16	SaltLength;
-	__le16	HashAlgorithms; /* HashAlgorithms[0] since only one defined */
-	__u8	Salt[SMB311_SALT_SIZE];
-} __packed;
-
-/* Encryption Algorithms Ciphers */
-#define SMB2_ENCRYPTION_AES128_CCM	cpu_to_le16(0x0001)
-#define SMB2_ENCRYPTION_AES128_GCM	cpu_to_le16(0x0002)
-#define SMB2_ENCRYPTION_AES256_CCM	cpu_to_le16(0x0003)
-#define SMB2_ENCRYPTION_AES256_GCM	cpu_to_le16(0x0004)
-
-struct smb2_encryption_neg_context {
-	__le16	ContextType; /* 2 */
-	__le16	DataLength;
-	__le32	Reserved;
-	/* CipherCount usally 2, but can be 3 when AES256-GCM enabled */
-	__le16	CipherCount; /* AES-128-GCM and AES-128-CCM by default */
-	__le16	Ciphers[];
-} __packed;
-
-#define SMB3_COMPRESS_NONE	cpu_to_le16(0x0000)
-#define SMB3_COMPRESS_LZNT1	cpu_to_le16(0x0001)
-#define SMB3_COMPRESS_LZ77	cpu_to_le16(0x0002)
-#define SMB3_COMPRESS_LZ77_HUFF	cpu_to_le16(0x0003)
-
-struct smb2_compression_ctx {
-	__le16	ContextType; /* 3 */
-	__le16  DataLength;
-	__le32	Reserved;
-	__le16	CompressionAlgorithmCount;
-	__u16	Padding;
-	__le32	Reserved1;
-	__le16	CompressionAlgorithms[];
-} __packed;
-
-#define POSIX_CTXT_DATA_LEN     16
-struct smb2_posix_neg_context {
-	__le16	ContextType; /* 0x100 */
-	__le16	DataLength;
-	__le32	Reserved;
-	__u8	Name[16]; /* POSIX ctxt GUID 93AD25509CB411E7B42383DE968BCD7C */
-} __packed;
-
-struct smb2_netname_neg_context {
-	__le16	ContextType; /* 0x100 */
-	__le16	DataLength;
-	__le32	Reserved;
-	__le16	NetName[]; /* hostname of target converted to UCS-2 */
-} __packed;
-
-/* Signing algorithms */
-#define SIGNING_ALG_HMAC_SHA256		cpu_to_le16(0)
-#define SIGNING_ALG_AES_CMAC		cpu_to_le16(1)
-#define SIGNING_ALG_AES_GMAC		cpu_to_le16(2)
-
-struct smb2_signing_capabilities {
-	__le16	ContextType; /* 8 */
-	__le16	DataLength;
-	__le32	Reserved;
-	__le16	SigningAlgorithmCount;
-	__le16	SigningAlgorithms[];
-} __packed;
-
-struct smb2_negotiate_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 65 */
-	__le16 SecurityMode;
-	__le16 DialectRevision;
-	__le16 NegotiateContextCount; /* Prior to SMB3.1.1 was Reserved & MBZ */
-	__u8   ServerGUID[16];
-	__le32 Capabilities;
-	__le32 MaxTransactSize;
-	__le32 MaxReadSize;
-	__le32 MaxWriteSize;
-	__le64 SystemTime;	/* MBZ */
-	__le64 ServerStartTime;
-	__le16 SecurityBufferOffset;
-	__le16 SecurityBufferLength;
-	__le32 NegotiateContextOffset;	/* Pre:SMB3.1.1 was reserved/ignored */
-	__u8   Buffer[1];	/* variable length GSS security buffer */
-} __packed;
-
-/* Flags */
-#define SMB2_SESSION_REQ_FLAG_BINDING		0x01
-#define SMB2_SESSION_REQ_FLAG_ENCRYPT_DATA	0x04
-
 #define SMB2_SESSION_EXPIRED		(0)
 #define SMB2_SESSION_IN_PROGRESS	BIT(0)
 #define SMB2_SESSION_VALID		BIT(1)
 
-/* Flags */
-#define SMB2_SESSION_REQ_FLAG_BINDING		0x01
-#define SMB2_SESSION_REQ_FLAG_ENCRYPT_DATA	0x04
-
-struct smb2_sess_setup_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize; /* Must be 25 */
-	__u8   Flags;
-	__u8   SecurityMode;
-	__le32 Capabilities;
-	__le32 Channel;
-	__le16 SecurityBufferOffset;
-	__le16 SecurityBufferLength;
-	__le64 PreviousSessionId;
-	__u8   Buffer[1];	/* variable length GSS security buffer */
-} __packed;
-
-/* Flags/Reserved for SMB3.1.1 */
-#define SMB2_SHAREFLAG_CLUSTER_RECONNECT	0x0001
-
-/* Currently defined SessionFlags */
-#define SMB2_SESSION_FLAG_IS_GUEST_LE		cpu_to_le16(0x0001)
-#define SMB2_SESSION_FLAG_IS_NULL_LE		cpu_to_le16(0x0002)
-#define SMB2_SESSION_FLAG_ENCRYPT_DATA_LE	cpu_to_le16(0x0004)
-struct smb2_sess_setup_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize; /* Must be 9 */
-	__le16 SessionFlags;
-	__le16 SecurityBufferOffset;
-	__le16 SecurityBufferLength;
-	__u8   Buffer[1];	/* variable length GSS security buffer */
-} __packed;
-
-struct smb2_logoff_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 4 */
-	__le16 Reserved;
-} __packed;
-
-struct smb2_logoff_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 4 */
-	__le16 Reserved;
-} __packed;
-
-struct smb2_tree_connect_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 9 */
-	__le16 Reserved;	/* Flags in SMB3.1.1 */
-	__le16 PathOffset;
-	__le16 PathLength;
-	__u8   Buffer[1];	/* variable length */
-} __packed;
-
-struct smb2_tree_connect_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 16 */
-	__u8   ShareType;  /* see below */
-	__u8   Reserved;
-	__le32 ShareFlags; /* see below */
-	__le32 Capabilities; /* see below */
-	__le32 MaximalAccess;
-} __packed;
-
-/* Possible ShareType values */
-#define SMB2_SHARE_TYPE_DISK	0x01
-#define SMB2_SHARE_TYPE_PIPE	0x02
-#define	SMB2_SHARE_TYPE_PRINT	0x03
-
-/*
- * Possible ShareFlags - exactly one and only one of the first 4 caching flags
- * must be set (any of the remaining, SHI1005, flags may be set individually
- * or in combination.
- */
-#define SMB2_SHAREFLAG_MANUAL_CACHING			0x00000000
-#define SMB2_SHAREFLAG_AUTO_CACHING			0x00000010
-#define SMB2_SHAREFLAG_VDO_CACHING			0x00000020
-#define SMB2_SHAREFLAG_NO_CACHING			0x00000030
-#define SHI1005_FLAGS_DFS				0x00000001
-#define SHI1005_FLAGS_DFS_ROOT				0x00000002
-#define SHI1005_FLAGS_RESTRICT_EXCLUSIVE_OPENS		0x00000100
-#define SHI1005_FLAGS_FORCE_SHARED_DELETE		0x00000200
-#define SHI1005_FLAGS_ALLOW_NAMESPACE_CACHING		0x00000400
-#define SHI1005_FLAGS_ACCESS_BASED_DIRECTORY_ENUM	0x00000800
-#define SHI1005_FLAGS_FORCE_LEVELII_OPLOCK		0x00001000
-#define SHI1005_FLAGS_ENABLE_HASH			0x00002000
-
-/* Possible share capabilities */
-#define SMB2_SHARE_CAP_DFS	cpu_to_le32(0x00000008)
-
-struct smb2_tree_disconnect_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 4 */
-	__le16 Reserved;
-} __packed;
-
-struct smb2_tree_disconnect_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 4 */
-	__le16 Reserved;
-} __packed;
-
-#define ATTR_READONLY_LE	cpu_to_le32(ATTR_READONLY)
-#define ATTR_HIDDEN_LE		cpu_to_le32(ATTR_HIDDEN)
-#define ATTR_SYSTEM_LE		cpu_to_le32(ATTR_SYSTEM)
-#define ATTR_DIRECTORY_LE	cpu_to_le32(ATTR_DIRECTORY)
-#define ATTR_ARCHIVE_LE		cpu_to_le32(ATTR_ARCHIVE)
-#define ATTR_NORMAL_LE		cpu_to_le32(ATTR_NORMAL)
-#define ATTR_TEMPORARY_LE	cpu_to_le32(ATTR_TEMPORARY)
-#define ATTR_SPARSE_FILE_LE	cpu_to_le32(ATTR_SPARSE)
-#define ATTR_REPARSE_POINT_LE	cpu_to_le32(ATTR_REPARSE)
-#define ATTR_COMPRESSED_LE	cpu_to_le32(ATTR_COMPRESSED)
-#define ATTR_OFFLINE_LE		cpu_to_le32(ATTR_OFFLINE)
-#define ATTR_NOT_CONTENT_INDEXED_LE	cpu_to_le32(ATTR_NOT_CONTENT_INDEXED)
-#define ATTR_ENCRYPTED_LE	cpu_to_le32(ATTR_ENCRYPTED)
-#define ATTR_INTEGRITY_STREAML_LE	cpu_to_le32(0x00008000)
-#define ATTR_NO_SCRUB_DATA_LE	cpu_to_le32(0x00020000)
-#define ATTR_MASK_LE		cpu_to_le32(0x00007FB7)
-
-/* Oplock levels */
-#define SMB2_OPLOCK_LEVEL_NONE		0x00
-#define SMB2_OPLOCK_LEVEL_II		0x01
-#define SMB2_OPLOCK_LEVEL_EXCLUSIVE	0x08
-#define SMB2_OPLOCK_LEVEL_BATCH		0x09
-#define SMB2_OPLOCK_LEVEL_LEASE		0xFF
-/* Non-spec internal type */
-#define SMB2_OPLOCK_LEVEL_NOCHANGE	0x99
-
-/* Desired Access Flags */
-#define FILE_READ_DATA_LE		cpu_to_le32(0x00000001)
-#define FILE_LIST_DIRECTORY_LE		cpu_to_le32(0x00000001)
-#define FILE_WRITE_DATA_LE		cpu_to_le32(0x00000002)
-#define FILE_ADD_FILE_LE		cpu_to_le32(0x00000002)
-#define FILE_APPEND_DATA_LE		cpu_to_le32(0x00000004)
-#define FILE_ADD_SUBDIRECTORY_LE	cpu_to_le32(0x00000004)
-#define FILE_READ_EA_LE			cpu_to_le32(0x00000008)
-#define FILE_WRITE_EA_LE		cpu_to_le32(0x00000010)
-#define FILE_EXECUTE_LE			cpu_to_le32(0x00000020)
-#define FILE_TRAVERSE_LE		cpu_to_le32(0x00000020)
-#define FILE_DELETE_CHILD_LE		cpu_to_le32(0x00000040)
-#define FILE_READ_ATTRIBUTES_LE		cpu_to_le32(0x00000080)
-#define FILE_WRITE_ATTRIBUTES_LE	cpu_to_le32(0x00000100)
-#define FILE_DELETE_LE			cpu_to_le32(0x00010000)
-#define FILE_READ_CONTROL_LE		cpu_to_le32(0x00020000)
-#define FILE_WRITE_DAC_LE		cpu_to_le32(0x00040000)
-#define FILE_WRITE_OWNER_LE		cpu_to_le32(0x00080000)
-#define FILE_SYNCHRONIZE_LE		cpu_to_le32(0x00100000)
-#define FILE_ACCESS_SYSTEM_SECURITY_LE	cpu_to_le32(0x01000000)
-#define FILE_MAXIMAL_ACCESS_LE		cpu_to_le32(0x02000000)
-#define FILE_GENERIC_ALL_LE		cpu_to_le32(0x10000000)
-#define FILE_GENERIC_EXECUTE_LE		cpu_to_le32(0x20000000)
-#define FILE_GENERIC_WRITE_LE		cpu_to_le32(0x40000000)
-#define FILE_GENERIC_READ_LE		cpu_to_le32(0x80000000)
-#define DESIRED_ACCESS_MASK		cpu_to_le32(0xF21F01FF)
-
-/* ShareAccess Flags */
-#define FILE_SHARE_READ_LE		cpu_to_le32(0x00000001)
-#define FILE_SHARE_WRITE_LE		cpu_to_le32(0x00000002)
-#define FILE_SHARE_DELETE_LE		cpu_to_le32(0x00000004)
-#define FILE_SHARE_ALL_LE		cpu_to_le32(0x00000007)
-
-/* CreateDisposition Flags */
-#define FILE_SUPERSEDE_LE		cpu_to_le32(0x00000000)
-#define FILE_OPEN_LE			cpu_to_le32(0x00000001)
-#define FILE_CREATE_LE			cpu_to_le32(0x00000002)
-#define	FILE_OPEN_IF_LE			cpu_to_le32(0x00000003)
-#define FILE_OVERWRITE_LE		cpu_to_le32(0x00000004)
-#define FILE_OVERWRITE_IF_LE		cpu_to_le32(0x00000005)
-#define FILE_CREATE_MASK_LE		cpu_to_le32(0x00000007)
-
-#define FILE_READ_DESIRED_ACCESS_LE	(FILE_READ_DATA_LE |		\
-					FILE_READ_EA_LE |		\
-					FILE_GENERIC_READ_LE)
-#define FILE_WRITE_DESIRE_ACCESS_LE	(FILE_WRITE_DATA_LE |		\
-					FILE_APPEND_DATA_LE |		\
-					FILE_WRITE_EA_LE |		\
-					FILE_WRITE_ATTRIBUTES_LE |	\
-					FILE_GENERIC_WRITE_LE)
-
-/* Impersonation Levels */
-#define IL_ANONYMOUS_LE		cpu_to_le32(0x00000000)
-#define IL_IDENTIFICATION_LE	cpu_to_le32(0x00000001)
-#define IL_IMPERSONATION_LE	cpu_to_le32(0x00000002)
-#define IL_DELEGATE_LE		cpu_to_le32(0x00000003)
-
-/* Create Context Values */
-#define SMB2_CREATE_EA_BUFFER			"ExtA" /* extended attributes */
-#define SMB2_CREATE_SD_BUFFER			"SecD" /* security descriptor */
-#define SMB2_CREATE_DURABLE_HANDLE_REQUEST	"DHnQ"
-#define SMB2_CREATE_DURABLE_HANDLE_RECONNECT	"DHnC"
-#define SMB2_CREATE_ALLOCATION_SIZE		"AlSi"
-#define SMB2_CREATE_QUERY_MAXIMAL_ACCESS_REQUEST "MxAc"
-#define SMB2_CREATE_TIMEWARP_REQUEST		"TWrp"
-#define SMB2_CREATE_QUERY_ON_DISK_ID		"QFid"
-#define SMB2_CREATE_REQUEST_LEASE		"RqLs"
-#define SMB2_CREATE_DURABLE_HANDLE_REQUEST_V2   "DH2Q"
-#define SMB2_CREATE_DURABLE_HANDLE_RECONNECT_V2 "DH2C"
-#define SMB2_CREATE_APP_INSTANCE_ID     "\x45\xBC\xA6\x6A\xEF\xA7\xF7\x4A\x90\x08\xFA\x46\x2E\x14\x4D\x74"
- #define SMB2_CREATE_APP_INSTANCE_VERSION	"\xB9\x82\xD0\xB7\x3B\x56\x07\x4F\xA0\x7B\x52\x4A\x81\x16\xA0\x10"
-#define SVHDX_OPEN_DEVICE_CONTEXT       0x83CE6F1AD851E0986E34401CC9BCFCE9
-#define SMB2_CREATE_TAG_POSIX		"\x93\xAD\x25\x50\x9C\xB4\x11\xE7\xB4\x23\x83\xDE\x96\x8B\xCD\x7C"
-
-struct smb2_create_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 57 */
-	__u8   SecurityFlags;
-	__u8   RequestedOplockLevel;
-	__le32 ImpersonationLevel;
-	__le64 SmbCreateFlags;
-	__le64 Reserved;
-	__le32 DesiredAccess;
-	__le32 FileAttributes;
-	__le32 ShareAccess;
-	__le32 CreateDisposition;
-	__le32 CreateOptions;
-	__le16 NameOffset;
-	__le16 NameLength;
-	__le32 CreateContextsOffset;
-	__le32 CreateContextsLength;
-	__u8   Buffer[0];
-} __packed;
-
-struct smb2_create_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 89 */
-	__u8   OplockLevel;
-	__u8   Reserved;
-	__le32 CreateAction;
-	__le64 CreationTime;
-	__le64 LastAccessTime;
-	__le64 LastWriteTime;
-	__le64 ChangeTime;
-	__le64 AllocationSize;
-	__le64 EndofFile;
-	__le32 FileAttributes;
-	__le32 Reserved2;
-	__le64  PersistentFileId;
-	__le64  VolatileFileId;
-	__le32 CreateContextsOffset;
-	__le32 CreateContextsLength;
-	__u8   Buffer[1];
-} __packed;
-
-struct create_context {
-	__le32 Next;
-	__le16 NameOffset;
-	__le16 NameLength;
-	__le16 Reserved;
-	__le16 DataOffset;
-	__le32 DataLength;
-	__u8 Buffer[0];
-} __packed;
-
 struct create_durable_req_v2 {
 	struct create_context ccontext;
 	__u8   Name[8];
@@ -743,22 +216,21 @@ struct create_posix_rsp {
 
 #define SMB2_LEASE_FLAG_BREAK_IN_PROGRESS_LE	cpu_to_le32(0x02)
 
+#define SMB2_LEASE_KEY_SIZE			16
+
 struct lease_context {
-	__le64 LeaseKeyLow;
-	__le64 LeaseKeyHigh;
+	__u8 LeaseKey[SMB2_LEASE_KEY_SIZE];
 	__le32 LeaseState;
 	__le32 LeaseFlags;
 	__le64 LeaseDuration;
 } __packed;
 
 struct lease_context_v2 {
-	__le64 LeaseKeyLow;
-	__le64 LeaseKeyHigh;
+	__u8 LeaseKey[SMB2_LEASE_KEY_SIZE];
 	__le32 LeaseState;
 	__le32 LeaseFlags;
 	__le64 LeaseDuration;
-	__le64 ParentLeaseKeyLow;
-	__le64 ParentLeaseKeyHigh;
+	__u8 ParentLeaseKey[SMB2_LEASE_KEY_SIZE];
 	__le16 Epoch;
 	__le16 Reserved;
 } __packed;
@@ -776,114 +248,12 @@ struct create_lease_v2 {
 	__u8   Pad[4];
 } __packed;
 
-/* Currently defined values for close flags */
-#define SMB2_CLOSE_FLAG_POSTQUERY_ATTRIB	cpu_to_le16(0x0001)
-struct smb2_close_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 24 */
-	__le16 Flags;
-	__le32 Reserved;
-	__le64  PersistentFileId;
-	__le64  VolatileFileId;
-} __packed;
-
-struct smb2_close_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize; /* 60 */
-	__le16 Flags;
-	__le32 Reserved;
-	__le64 CreationTime;
-	__le64 LastAccessTime;
-	__le64 LastWriteTime;
-	__le64 ChangeTime;
-	__le64 AllocationSize;	/* Beginning of FILE_STANDARD_INFO equivalent */
-	__le64 EndOfFile;
-	__le32 Attributes;
-} __packed;
-
-struct smb2_flush_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;	/* Must be 24 */
-	__le16 Reserved1;
-	__le32 Reserved2;
-	__le64  PersistentFileId;
-	__le64  VolatileFileId;
-} __packed;
-
-struct smb2_flush_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize;
-	__le16 Reserved;
-} __packed;
-
 struct smb2_buffer_desc_v1 {
 	__le64 offset;
 	__le32 token;
 	__le32 length;
 } __packed;
 
-#define SMB2_CHANNEL_NONE		cpu_to_le32(0x00000000)
-#define SMB2_CHANNEL_RDMA_V1		cpu_to_le32(0x00000001)
-#define SMB2_CHANNEL_RDMA_V1_INVALIDATE cpu_to_le32(0x00000002)
-
-struct smb2_read_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize; /* Must be 49 */
-	__u8   Padding; /* offset from start of SMB2 header to place read */
-	__u8   Reserved;
-	__le32 Length;
-	__le64 Offset;
-	__le64  PersistentFileId;
-	__le64  VolatileFileId;
-	__le32 MinimumCount;
-	__le32 Channel; /* Reserved MBZ */
-	__le32 RemainingBytes;
-	__le16 ReadChannelInfoOffset; /* Reserved MBZ */
-	__le16 ReadChannelInfoLength; /* Reserved MBZ */
-	__u8   Buffer[1];
-} __packed;
-
-struct smb2_read_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize; /* Must be 17 */
-	__u8   DataOffset;
-	__u8   Reserved;
-	__le32 DataLength;
-	__le32 DataRemaining;
-	__u32  Reserved2;
-	__u8   Buffer[1];
-} __packed;
-
-/* For write request Flags field below the following flag is defined: */
-#define SMB2_WRITEFLAG_WRITE_THROUGH 0x00000001
-
-struct smb2_write_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize; /* Must be 49 */
-	__le16 DataOffset; /* offset from start of SMB2 header to write data */
-	__le32 Length;
-	__le64 Offset;
-	__le64  PersistentFileId;
-	__le64  VolatileFileId;
-	__le32 Channel; /* Reserved MBZ */
-	__le32 RemainingBytes;
-	__le16 WriteChannelInfoOffset; /* Reserved MBZ */
-	__le16 WriteChannelInfoLength; /* Reserved MBZ */
-	__le32 Flags;
-	__u8   Buffer[1];
-} __packed;
-
-struct smb2_write_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize; /* Must be 17 */
-	__u8   DataOffset;
-	__u8   Reserved;
-	__le32 DataLength;
-	__le32 DataRemaining;
-	__u32  Reserved2;
-	__u8   Buffer[1];
-} __packed;
-
 #define SMB2_0_IOCTL_IS_FSCTL 0x00000001
 
 struct duplicate_extents_to_file {
@@ -1033,43 +403,6 @@ struct reparse_data_buffer {
 	__u8	DataBuffer[]; /* Variable Length */
 } __packed;
 
-/* Completion Filter flags for Notify */
-#define FILE_NOTIFY_CHANGE_FILE_NAME	0x00000001
-#define FILE_NOTIFY_CHANGE_DIR_NAME	0x00000002
-#define FILE_NOTIFY_CHANGE_NAME		0x00000003
-#define FILE_NOTIFY_CHANGE_ATTRIBUTES	0x00000004
-#define FILE_NOTIFY_CHANGE_SIZE		0x00000008
-#define FILE_NOTIFY_CHANGE_LAST_WRITE	0x00000010
-#define FILE_NOTIFY_CHANGE_LAST_ACCESS	0x00000020
-#define FILE_NOTIFY_CHANGE_CREATION	0x00000040
-#define FILE_NOTIFY_CHANGE_EA		0x00000080
-#define FILE_NOTIFY_CHANGE_SECURITY	0x00000100
-#define FILE_NOTIFY_CHANGE_STREAM_NAME	0x00000200
-#define FILE_NOTIFY_CHANGE_STREAM_SIZE	0x00000400
-#define FILE_NOTIFY_CHANGE_STREAM_WRITE	0x00000800
-
-/* Flags */
-#define SMB2_WATCH_TREE	0x0001
-
-struct smb2_notify_req {
-	struct smb2_hdr hdr;
-	__le16 StructureSize; /* Must be 32 */
-	__le16 Flags;
-	__le32 OutputBufferLength;
-	__le64 PersistentFileId;
-	__le64 VolatileFileId;
-	__u32 CompletionFileter;
-	__u32 Reserved;
-} __packed;
-
-struct smb2_notify_rsp {
-	struct smb2_hdr hdr;
-	__le16 StructureSize; /* Must be 9 */
-	__le16 OutputBufferOffset;
-	__le32 OutputBufferLength;
-	__u8 Buffer[1];
-} __packed;
-
 /* SMB2 Notify Action Flags */
 #define FILE_ACTION_ADDED		0x00000001
 #define FILE_ACTION_REMOVED		0x00000002
@@ -1528,7 +861,7 @@ struct smb2_file_pos_info {
 	__le64 CurrentByteOffset;
 } __packed;
 
-#define FILE_MODE_INFO_MASK cpu_to_le32(0x0000103e)
+#define FILE_MODE_INFO_MASK cpu_to_le32(0x0000100e)
 
 struct smb2_file_mode_info {
 	__le32 Mode;
@@ -1705,4 +1038,13 @@ int smb2_ioctl(struct ksmbd_work *work);
 int smb2_oplock_break(struct ksmbd_work *work);
 int smb2_notify(struct ksmbd_work *ksmbd_work);
 
+/*
+ * Get the body of the smb2 message excluding the 4 byte rfc1002 headers
+ * from request/response buffer.
+ */
+static inline void *smb2_get_msg(void *buf)
+{
+	return buf + 4;
+}
+
 #endif	/* _SMB2PDU_H */
diff --git a/fs/ksmbd/smb_common.c b/fs/ksmbd/smb_common.c
index 707490a..ef7f42b 100644
--- a/fs/ksmbd/smb_common.c
+++ b/fs/ksmbd/smb_common.c
@@ -132,7 +132,7 @@ int ksmbd_lookup_protocol_idx(char *str)
  */
 int ksmbd_verify_smb_message(struct ksmbd_work *work)
 {
-	struct smb2_hdr *smb2_hdr = work->request_buf + work->next_smb2_rcv_hdr_off;
+	struct smb2_hdr *smb2_hdr = ksmbd_req_buf_next(work);
 	struct smb_hdr *hdr;
 
 	if (smb2_hdr->ProtocolId == SMB2_PROTO_NUMBER)
@@ -239,14 +239,14 @@ int ksmbd_lookup_dialect_by_id(__le16 *cli_dialects, __le16 dialects_count)
 static int ksmbd_negotiate_smb_dialect(void *buf)
 {
 	int smb_buf_length = get_rfc1002_len(buf);
-	__le32 proto = ((struct smb2_hdr *)buf)->ProtocolId;
+	__le32 proto = ((struct smb2_hdr *)smb2_get_msg(buf))->ProtocolId;
 
 	if (proto == SMB2_PROTO_NUMBER) {
 		struct smb2_negotiate_req *req;
 		int smb2_neg_size =
-			offsetof(struct smb2_negotiate_req, Dialects) - 4;
+			offsetof(struct smb2_negotiate_req, Dialects);
 
-		req = (struct smb2_negotiate_req *)buf;
+		req = (struct smb2_negotiate_req *)smb2_get_msg(buf);
 		if (smb2_neg_size > smb_buf_length)
 			goto err_out;
 
@@ -445,11 +445,12 @@ int ksmbd_smb_negotiate_common(struct ksmbd_work *work, unsigned int command)
 	struct ksmbd_conn *conn = work->conn;
 	int ret;
 
-	conn->dialect = ksmbd_negotiate_smb_dialect(work->request_buf);
+	conn->dialect =
+		ksmbd_negotiate_smb_dialect(work->request_buf);
 	ksmbd_debug(SMB, "conn->dialect 0x%x\n", conn->dialect);
 
 	if (command == SMB2_NEGOTIATE_HE) {
-		struct smb2_hdr *smb2_hdr = work->request_buf;
+		struct smb2_hdr *smb2_hdr = smb2_get_msg(work->request_buf);
 
 		if (smb2_hdr->ProtocolId != SMB2_PROTO_NUMBER) {
 			ksmbd_debug(SMB, "Downgrade to SMB1 negotiation\n");
diff --git a/fs/ksmbd/smb_common.h b/fs/ksmbd/smb_common.h
index 6e79e75..5059084 100644
--- a/fs/ksmbd/smb_common.h
+++ b/fs/ksmbd/smb_common.h
@@ -10,6 +10,7 @@
 
 #include "glob.h"
 #include "nterr.h"
+#include "../smbfs_common/smb2pdu.h"
 #include "smb2pdu.h"
 
 /* ksmbd's Specific ERRNO */
@@ -32,17 +33,6 @@
 #define SMB302_VERSION_STRING	"3.02"
 #define SMB311_VERSION_STRING	"3.1.1"
 
-/* Dialects */
-#define SMB10_PROT_ID		0x00
-#define SMB20_PROT_ID		0x0202
-#define SMB21_PROT_ID		0x0210
-/* multi-protocol negotiate request */
-#define SMB2X_PROT_ID		0x02FF
-#define SMB30_PROT_ID		0x0300
-#define SMB302_PROT_ID		0x0302
-#define SMB311_PROT_ID		0x0311
-#define BAD_PROT_ID		0xFFFF
-
 #define SMB_ECHO_INTERVAL	(60 * HZ)
 
 #define CIFS_DEFAULT_IOSIZE	(64 * 1024)
@@ -59,21 +49,6 @@
 /*
  * File Attribute flags
  */
-#define ATTR_READONLY			0x0001
-#define ATTR_HIDDEN			0x0002
-#define ATTR_SYSTEM			0x0004
-#define ATTR_VOLUME			0x0008
-#define ATTR_DIRECTORY			0x0010
-#define ATTR_ARCHIVE			0x0020
-#define ATTR_DEVICE			0x0040
-#define ATTR_NORMAL			0x0080
-#define ATTR_TEMPORARY			0x0100
-#define ATTR_SPARSE			0x0200
-#define ATTR_REPARSE			0x0400
-#define ATTR_COMPRESSED			0x0800
-#define ATTR_OFFLINE			0x1000
-#define ATTR_NOT_CONTENT_INDEXED	0x2000
-#define ATTR_ENCRYPTED			0x4000
 #define ATTR_POSIX_SEMANTICS		0x01000000
 #define ATTR_BACKUP_SEMANTICS		0x02000000
 #define ATTR_DELETE_ON_CLOSE		0x04000000
@@ -82,23 +57,6 @@
 #define ATTR_NO_BUFFERING		0x20000000
 #define ATTR_WRITE_THROUGH		0x80000000
 
-#define ATTR_READONLY_LE		cpu_to_le32(ATTR_READONLY)
-#define ATTR_HIDDEN_LE			cpu_to_le32(ATTR_HIDDEN)
-#define ATTR_SYSTEM_LE			cpu_to_le32(ATTR_SYSTEM)
-#define ATTR_DIRECTORY_LE		cpu_to_le32(ATTR_DIRECTORY)
-#define ATTR_ARCHIVE_LE			cpu_to_le32(ATTR_ARCHIVE)
-#define ATTR_NORMAL_LE			cpu_to_le32(ATTR_NORMAL)
-#define ATTR_TEMPORARY_LE		cpu_to_le32(ATTR_TEMPORARY)
-#define ATTR_SPARSE_FILE_LE		cpu_to_le32(ATTR_SPARSE)
-#define ATTR_REPARSE_POINT_LE		cpu_to_le32(ATTR_REPARSE)
-#define ATTR_COMPRESSED_LE		cpu_to_le32(ATTR_COMPRESSED)
-#define ATTR_OFFLINE_LE			cpu_to_le32(ATTR_OFFLINE)
-#define ATTR_NOT_CONTENT_INDEXED_LE	cpu_to_le32(ATTR_NOT_CONTENT_INDEXED)
-#define ATTR_ENCRYPTED_LE		cpu_to_le32(ATTR_ENCRYPTED)
-#define ATTR_INTEGRITY_STREAML_LE	cpu_to_le32(0x00008000)
-#define ATTR_NO_SCRUB_DATA_LE		cpu_to_le32(0x00020000)
-#define ATTR_MASK_LE			cpu_to_le32(0x00007FB7)
-
 /* List of FileSystemAttributes - see 2.5.1 of MS-FSCC */
 #define FILE_SUPPORTS_SPARSE_VDL	0x10000000 /* faster nonsparse extend */
 #define FILE_SUPPORTS_BLOCK_REFCOUNTING	0x08000000 /* allow ioctl dup extents */
@@ -160,11 +118,6 @@
 /* file_execute, file_read_attributes*/
 /* write_dac, and delete.           */
 
-#define FILE_READ_RIGHTS (FILE_READ_DATA | FILE_READ_EA | FILE_READ_ATTRIBUTES)
-#define FILE_WRITE_RIGHTS (FILE_WRITE_DATA | FILE_APPEND_DATA \
-		| FILE_WRITE_EA | FILE_WRITE_ATTRIBUTES)
-#define FILE_EXEC_RIGHTS (FILE_EXECUTE)
-
 #define SET_FILE_READ_RIGHTS (FILE_READ_DATA | FILE_READ_EA \
 		| FILE_READ_ATTRIBUTES \
 		| DELETE | READ_CONTROL | WRITE_DAC \
@@ -477,12 +430,6 @@ struct smb_version_cmds {
 	int (*proc)(struct ksmbd_work *swork);
 };
 
-static inline size_t
-smb2_hdr_size_no_buflen(struct smb_version_values *vals)
-{
-	return vals->header_size - 4;
-}
-
 int ksmbd_min_protocol(void);
 int ksmbd_max_protocol(void);
 
diff --git a/fs/ksmbd/transport_rdma.c b/fs/ksmbd/transport_rdma.c
index a2fd5a4..7e57cbb 100644
--- a/fs/ksmbd/transport_rdma.c
+++ b/fs/ksmbd/transport_rdma.c
@@ -484,7 +484,7 @@ static int smb_direct_check_recvmsg(struct smb_direct_recvmsg *recvmsg)
 		struct smb_direct_data_transfer *req =
 			(struct smb_direct_data_transfer *)recvmsg->packet;
 		struct smb2_hdr *hdr = (struct smb2_hdr *)(recvmsg->packet
-				+ le32_to_cpu(req->data_offset) - 4);
+				+ le32_to_cpu(req->data_offset));
 		ksmbd_debug(RDMA,
 			    "CreditGranted: %u, CreditRequested: %u, DataLength: %u, RemainingDataLength: %u, SMB: %x, Command: %u\n",
 			    le16_to_cpu(req->credits_granted),
@@ -2043,7 +2043,6 @@ int ksmbd_rdma_destroy(void)
 	smb_direct_listener.cm_id = NULL;
 
 	if (smb_direct_wq) {
-		flush_workqueue(smb_direct_wq);
 		destroy_workqueue(smb_direct_wq);
 		smb_direct_wq = NULL;
 	}
diff --git a/fs/ksmbd/vfs.c b/fs/ksmbd/vfs.c
index 835b384..19d3639 100644
--- a/fs/ksmbd/vfs.c
+++ b/fs/ksmbd/vfs.c
@@ -1013,7 +1013,7 @@ int ksmbd_vfs_zero_data(struct ksmbd_work *work, struct ksmbd_file *fp,
 			loff_t off, loff_t len)
 {
 	smb_break_all_levII_oplock(work, fp, 1);
-	if (fp->f_ci->m_fattr & ATTR_SPARSE_FILE_LE)
+	if (fp->f_ci->m_fattr & FILE_ATTRIBUTE_SPARSE_FILE_LE)
 		return vfs_fallocate(fp->filp,
 				     FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE,
 				     off, len);
@@ -1624,7 +1624,7 @@ void *ksmbd_vfs_init_kstat(char **p, struct ksmbd_kstat *ksmbd_kstat)
 	time = ksmbd_UnixTimeToNT(kstat->ctime);
 	info->ChangeTime = cpu_to_le64(time);
 
-	if (ksmbd_kstat->file_attributes & ATTR_DIRECTORY_LE) {
+	if (ksmbd_kstat->file_attributes & FILE_ATTRIBUTE_DIRECTORY_LE) {
 		info->EndOfFile = 0;
 		info->AllocationSize = 0;
 	} else {
@@ -1654,9 +1654,9 @@ int ksmbd_vfs_fill_dentry_attrs(struct ksmbd_work *work,
 	 * or that acl is disable in server's filesystem and the config is yes.
 	 */
 	if (S_ISDIR(ksmbd_kstat->kstat->mode))
-		ksmbd_kstat->file_attributes = ATTR_DIRECTORY_LE;
+		ksmbd_kstat->file_attributes = FILE_ATTRIBUTE_DIRECTORY_LE;
 	else
-		ksmbd_kstat->file_attributes = ATTR_ARCHIVE_LE;
+		ksmbd_kstat->file_attributes = FILE_ATTRIBUTE_ARCHIVE_LE;
 
 	if (test_share_config_flag(work->tcon->share_conf,
 				   KSMBD_SHARE_FLAG_STORE_DOS_ATTRS)) {
diff --git a/fs/ksmbd/vfs.h b/fs/ksmbd/vfs.h
index b0d5b8f..adf94a4 100644
--- a/fs/ksmbd/vfs.h
+++ b/fs/ksmbd/vfs.h
@@ -25,48 +25,9 @@ enum {
 };
 
 /* CreateOptions */
-/* Flag is set, it must not be a file , valid for directory only */
-#define FILE_DIRECTORY_FILE_LE			cpu_to_le32(0x00000001)
-#define FILE_WRITE_THROUGH_LE			cpu_to_le32(0x00000002)
-#define FILE_SEQUENTIAL_ONLY_LE			cpu_to_le32(0x00000004)
-
-/* Should not buffer on server*/
-#define FILE_NO_INTERMEDIATE_BUFFERING_LE	cpu_to_le32(0x00000008)
-/* MBZ */
-#define FILE_SYNCHRONOUS_IO_ALERT_LE		cpu_to_le32(0x00000010)
-/* MBZ */
-#define FILE_SYNCHRONOUS_IO_NONALERT_LE		cpu_to_le32(0x00000020)
-
-/* Flaf must not be set for directory */
-#define FILE_NON_DIRECTORY_FILE_LE		cpu_to_le32(0x00000040)
-
-/* Should be zero */
 #define CREATE_TREE_CONNECTION			cpu_to_le32(0x00000080)
-#define FILE_COMPLETE_IF_OPLOCKED_LE		cpu_to_le32(0x00000100)
-#define FILE_NO_EA_KNOWLEDGE_LE			cpu_to_le32(0x00000200)
-#define FILE_OPEN_REMOTE_INSTANCE		cpu_to_le32(0x00000400)
-
-/**
- * Doc says this is obsolete "open for recovery" flag should be zero
- * in any case.
- */
-#define CREATE_OPEN_FOR_RECOVERY		cpu_to_le32(0x00000400)
-#define FILE_RANDOM_ACCESS_LE			cpu_to_le32(0x00000800)
-#define FILE_DELETE_ON_CLOSE_LE			cpu_to_le32(0x00001000)
-#define FILE_OPEN_BY_FILE_ID_LE			cpu_to_le32(0x00002000)
-#define FILE_OPEN_FOR_BACKUP_INTENT_LE		cpu_to_le32(0x00004000)
-#define FILE_NO_COMPRESSION_LE			cpu_to_le32(0x00008000)
-
-/* Should be zero*/
-#define FILE_OPEN_REQUIRING_OPLOCK		cpu_to_le32(0x00010000)
-#define FILE_DISALLOW_EXCLUSIVE			cpu_to_le32(0x00020000)
 #define FILE_RESERVE_OPFILTER_LE		cpu_to_le32(0x00100000)
-#define FILE_OPEN_REPARSE_POINT_LE		cpu_to_le32(0x00200000)
-#define FILE_OPEN_NO_RECALL_LE			cpu_to_le32(0x00400000)
 
-/* Should be zero */
-#define FILE_OPEN_FOR_FREE_SPACE_QUERY_LE	cpu_to_le32(0x00800000)
-#define CREATE_OPTIONS_MASK			cpu_to_le32(0x00FFFFFF)
 #define CREATE_OPTION_READONLY			0x10000000
 /* system. NB not sent over wire */
 #define CREATE_OPTION_SPECIAL			0x20000000
diff --git a/fs/netfs/read_helper.c b/fs/netfs/read_helper.c
index 994ec22..9320a42 100644
--- a/fs/netfs/read_helper.c
+++ b/fs/netfs/read_helper.c
@@ -230,7 +230,7 @@ static void netfs_rreq_completed(struct netfs_read_request *rreq, bool was_async
 
 /*
  * Deal with the completion of writing the data to the cache.  We have to clear
- * the PG_fscache bits on the pages involved and release the caller's ref.
+ * the PG_fscache bits on the folios involved and release the caller's ref.
  *
  * May be called in softirq mode and we inherit a ref from the caller.
  */
@@ -238,7 +238,7 @@ static void netfs_rreq_unmark_after_write(struct netfs_read_request *rreq,
 					  bool was_async)
 {
 	struct netfs_read_subrequest *subreq;
-	struct page *page;
+	struct folio *folio;
 	pgoff_t unlocked = 0;
 	bool have_unlocked = false;
 
@@ -247,14 +247,14 @@ static void netfs_rreq_unmark_after_write(struct netfs_read_request *rreq,
 	list_for_each_entry(subreq, &rreq->subrequests, rreq_link) {
 		XA_STATE(xas, &rreq->mapping->i_pages, subreq->start / PAGE_SIZE);
 
-		xas_for_each(&xas, page, (subreq->start + subreq->len - 1) / PAGE_SIZE) {
+		xas_for_each(&xas, folio, (subreq->start + subreq->len - 1) / PAGE_SIZE) {
 			/* We might have multiple writes from the same huge
-			 * page, but we mustn't unlock a page more than once.
+			 * folio, but we mustn't unlock a folio more than once.
 			 */
-			if (have_unlocked && page->index <= unlocked)
+			if (have_unlocked && folio_index(folio) <= unlocked)
 				continue;
-			unlocked = page->index;
-			end_page_fscache(page);
+			unlocked = folio_index(folio);
+			folio_end_fscache(folio);
 			have_unlocked = true;
 		}
 	}
@@ -367,18 +367,17 @@ static void netfs_rreq_write_to_cache(struct netfs_read_request *rreq,
 }
 
 /*
- * Unlock the pages in a read operation.  We need to set PG_fscache on any
- * pages we're going to write back before we unlock them.
+ * Unlock the folios in a read operation.  We need to set PG_fscache on any
+ * folios we're going to write back before we unlock them.
  */
 static void netfs_rreq_unlock(struct netfs_read_request *rreq)
 {
 	struct netfs_read_subrequest *subreq;
-	struct page *page;
+	struct folio *folio;
 	unsigned int iopos, account = 0;
 	pgoff_t start_page = rreq->start / PAGE_SIZE;
 	pgoff_t last_page = ((rreq->start + rreq->len) / PAGE_SIZE) - 1;
 	bool subreq_failed = false;
-	int i;
 
 	XA_STATE(xas, &rreq->mapping->i_pages, start_page);
 
@@ -403,9 +402,9 @@ static void netfs_rreq_unlock(struct netfs_read_request *rreq)
 	trace_netfs_rreq(rreq, netfs_rreq_trace_unlock);
 
 	rcu_read_lock();
-	xas_for_each(&xas, page, last_page) {
-		unsigned int pgpos = (page->index - start_page) * PAGE_SIZE;
-		unsigned int pgend = pgpos + thp_size(page);
+	xas_for_each(&xas, folio, last_page) {
+		unsigned int pgpos = (folio_index(folio) - start_page) * PAGE_SIZE;
+		unsigned int pgend = pgpos + folio_size(folio);
 		bool pg_failed = false;
 
 		for (;;) {
@@ -414,7 +413,7 @@ static void netfs_rreq_unlock(struct netfs_read_request *rreq)
 				break;
 			}
 			if (test_bit(NETFS_SREQ_WRITE_TO_CACHE, &subreq->flags))
-				set_page_fscache(page);
+				folio_start_fscache(folio);
 			pg_failed |= subreq_failed;
 			if (pgend < iopos + subreq->len)
 				break;
@@ -433,17 +432,16 @@ static void netfs_rreq_unlock(struct netfs_read_request *rreq)
 		}
 
 		if (!pg_failed) {
-			for (i = 0; i < thp_nr_pages(page); i++)
-				flush_dcache_page(page);
-			SetPageUptodate(page);
+			flush_dcache_folio(folio);
+			folio_mark_uptodate(folio);
 		}
 
-		if (!test_bit(NETFS_RREQ_DONT_UNLOCK_PAGES, &rreq->flags)) {
-			if (page->index == rreq->no_unlock_page &&
-			    test_bit(NETFS_RREQ_NO_UNLOCK_PAGE, &rreq->flags))
+		if (!test_bit(NETFS_RREQ_DONT_UNLOCK_FOLIOS, &rreq->flags)) {
+			if (folio_index(folio) == rreq->no_unlock_folio &&
+			    test_bit(NETFS_RREQ_NO_UNLOCK_FOLIO, &rreq->flags))
 				_debug("no unlock");
 			else
-				unlock_page(page);
+				folio_unlock(folio);
 		}
 	}
 	rcu_read_unlock();
@@ -876,7 +874,6 @@ void netfs_readahead(struct readahead_control *ractl,
 		     void *netfs_priv)
 {
 	struct netfs_read_request *rreq;
-	struct page *page;
 	unsigned int debug_index = 0;
 	int ret;
 
@@ -911,11 +908,11 @@ void netfs_readahead(struct readahead_control *ractl,
 
 	} while (rreq->submitted < rreq->len);
 
-	/* Drop the refs on the pages here rather than in the cache or
+	/* Drop the refs on the folios here rather than in the cache or
 	 * filesystem.  The locks will be dropped in netfs_rreq_unlock().
 	 */
-	while ((page = readahead_page(ractl)))
-		put_page(page);
+	while (readahead_folio(ractl))
+		;
 
 	/* If we decrement nr_rd_ops to 0, the ref belongs to us. */
 	if (atomic_dec_and_test(&rreq->nr_rd_ops))
@@ -935,7 +932,7 @@ EXPORT_SYMBOL(netfs_readahead);
 /**
  * netfs_readpage - Helper to manage a readpage request
  * @file: The file to read from
- * @page: The page to read
+ * @folio: The folio to read
  * @ops: The network filesystem's operations for the helper to use
  * @netfs_priv: Private netfs data to be retained in the request
  *
@@ -950,7 +947,7 @@ EXPORT_SYMBOL(netfs_readahead);
  * This is usable whether or not caching is enabled.
  */
 int netfs_readpage(struct file *file,
-		   struct page *page,
+		   struct folio *folio,
 		   const struct netfs_read_request_ops *ops,
 		   void *netfs_priv)
 {
@@ -958,23 +955,23 @@ int netfs_readpage(struct file *file,
 	unsigned int debug_index = 0;
 	int ret;
 
-	_enter("%lx", page_index(page));
+	_enter("%lx", folio_index(folio));
 
 	rreq = netfs_alloc_read_request(ops, netfs_priv, file);
 	if (!rreq) {
 		if (netfs_priv)
-			ops->cleanup(netfs_priv, page_file_mapping(page));
-		unlock_page(page);
+			ops->cleanup(netfs_priv, folio_file_mapping(folio));
+		folio_unlock(folio);
 		return -ENOMEM;
 	}
-	rreq->mapping	= page_file_mapping(page);
-	rreq->start	= page_file_offset(page);
-	rreq->len	= thp_size(page);
+	rreq->mapping	= folio_file_mapping(folio);
+	rreq->start	= folio_file_pos(folio);
+	rreq->len	= folio_size(folio);
 
 	if (ops->begin_cache_operation) {
 		ret = ops->begin_cache_operation(rreq);
 		if (ret == -ENOMEM || ret == -EINTR || ret == -ERESTARTSYS) {
-			unlock_page(page);
+			folio_unlock(folio);
 			goto out;
 		}
 	}
@@ -1012,40 +1009,40 @@ int netfs_readpage(struct file *file,
 EXPORT_SYMBOL(netfs_readpage);
 
 /**
- * netfs_skip_page_read - prep a page for writing without reading first
- * @page: page being prepared
+ * netfs_skip_folio_read - prep a folio for writing without reading first
+ * @folio: The folio being prepared
  * @pos: starting position for the write
  * @len: length of write
  *
  * In some cases, write_begin doesn't need to read at all:
- * - full page write
- * - write that lies in a page that is completely beyond EOF
- * - write that covers the the page from start to EOF or beyond it
+ * - full folio write
+ * - write that lies in a folio that is completely beyond EOF
+ * - write that covers the folio from start to EOF or beyond it
  *
  * If any of these criteria are met, then zero out the unwritten parts
- * of the page and return true. Otherwise, return false.
+ * of the folio and return true. Otherwise, return false.
  */
-static bool netfs_skip_page_read(struct page *page, loff_t pos, size_t len)
+static bool netfs_skip_folio_read(struct folio *folio, loff_t pos, size_t len)
 {
-	struct inode *inode = page->mapping->host;
+	struct inode *inode = folio_inode(folio);
 	loff_t i_size = i_size_read(inode);
-	size_t offset = offset_in_thp(page, pos);
+	size_t offset = offset_in_folio(folio, pos);
 
-	/* Full page write */
-	if (offset == 0 && len >= thp_size(page))
+	/* Full folio write */
+	if (offset == 0 && len >= folio_size(folio))
 		return true;
 
-	/* pos beyond last page in the file */
+	/* pos beyond last folio in the file */
 	if (pos - offset >= i_size)
 		goto zero_out;
 
-	/* Write that covers from the start of the page to EOF or beyond */
+	/* Write that covers from the start of the folio to EOF or beyond */
 	if (offset == 0 && (pos + len) >= i_size)
 		goto zero_out;
 
 	return false;
 zero_out:
-	zero_user_segments(page, 0, offset, offset + len, thp_size(page));
+	zero_user_segments(&folio->page, 0, offset, offset + len, folio_size(folio));
 	return true;
 }
 
@@ -1054,9 +1051,9 @@ static bool netfs_skip_page_read(struct page *page, loff_t pos, size_t len)
  * @file: The file to read from
  * @mapping: The mapping to read from
  * @pos: File position at which the write will begin
- * @len: The length of the write (may extend beyond the end of the page chosen)
- * @flags: AOP_* flags
- * @_page: Where to put the resultant page
+ * @len: The length of the write (may extend beyond the end of the folio chosen)
+ * @aop_flags: AOP_* flags
+ * @_folio: Where to put the resultant folio
  * @_fsdata: Place for the netfs to store a cookie
  * @ops: The network filesystem's operations for the helper to use
  * @netfs_priv: Private netfs data to be retained in the request
@@ -1072,37 +1069,41 @@ static bool netfs_skip_page_read(struct page *page, loff_t pos, size_t len)
  * issue_op, is mandatory.
  *
  * The check_write_begin() operation can be provided to check for and flush
- * conflicting writes once the page is grabbed and locked.  It is passed a
+ * conflicting writes once the folio is grabbed and locked.  It is passed a
  * pointer to the fsdata cookie that gets returned to the VM to be passed to
  * write_end.  It is permitted to sleep.  It should return 0 if the request
- * should go ahead; unlock the page and return -EAGAIN to cause the page to be
- * regot; or return an error.
+ * should go ahead; unlock the folio and return -EAGAIN to cause the folio to
+ * be regot; or return an error.
  *
  * This is usable whether or not caching is enabled.
  */
 int netfs_write_begin(struct file *file, struct address_space *mapping,
-		      loff_t pos, unsigned int len, unsigned int flags,
-		      struct page **_page, void **_fsdata,
+		      loff_t pos, unsigned int len, unsigned int aop_flags,
+		      struct folio **_folio, void **_fsdata,
 		      const struct netfs_read_request_ops *ops,
 		      void *netfs_priv)
 {
 	struct netfs_read_request *rreq;
-	struct page *page, *xpage;
+	struct folio *folio;
 	struct inode *inode = file_inode(file);
-	unsigned int debug_index = 0;
+	unsigned int debug_index = 0, fgp_flags;
 	pgoff_t index = pos >> PAGE_SHIFT;
 	int ret;
 
 	DEFINE_READAHEAD(ractl, file, NULL, mapping, index);
 
 retry:
-	page = grab_cache_page_write_begin(mapping, index, flags);
-	if (!page)
+	fgp_flags = FGP_LOCK | FGP_WRITE | FGP_CREAT | FGP_STABLE;
+	if (aop_flags & AOP_FLAG_NOFS)
+		fgp_flags |= FGP_NOFS;
+	folio = __filemap_get_folio(mapping, index, fgp_flags,
+				    mapping_gfp_mask(mapping));
+	if (!folio)
 		return -ENOMEM;
 
 	if (ops->check_write_begin) {
 		/* Allow the netfs (eg. ceph) to flush conflicts. */
-		ret = ops->check_write_begin(file, pos, len, page, _fsdata);
+		ret = ops->check_write_begin(file, pos, len, folio, _fsdata);
 		if (ret < 0) {
 			trace_netfs_failure(NULL, NULL, ret, netfs_fail_check_write_begin);
 			if (ret == -EAGAIN)
@@ -1111,28 +1112,28 @@ int netfs_write_begin(struct file *file, struct address_space *mapping,
 		}
 	}
 
-	if (PageUptodate(page))
-		goto have_page;
+	if (folio_test_uptodate(folio))
+		goto have_folio;
 
 	/* If the page is beyond the EOF, we want to clear it - unless it's
 	 * within the cache granule containing the EOF, in which case we need
 	 * to preload the granule.
 	 */
 	if (!ops->is_cache_enabled(inode) &&
-	    netfs_skip_page_read(page, pos, len)) {
+	    netfs_skip_folio_read(folio, pos, len)) {
 		netfs_stat(&netfs_n_rh_write_zskip);
-		goto have_page_no_wait;
+		goto have_folio_no_wait;
 	}
 
 	ret = -ENOMEM;
 	rreq = netfs_alloc_read_request(ops, netfs_priv, file);
 	if (!rreq)
 		goto error;
-	rreq->mapping		= page->mapping;
-	rreq->start		= page_offset(page);
-	rreq->len		= thp_size(page);
-	rreq->no_unlock_page	= page->index;
-	__set_bit(NETFS_RREQ_NO_UNLOCK_PAGE, &rreq->flags);
+	rreq->mapping		= folio_file_mapping(folio);
+	rreq->start		= folio_file_pos(folio);
+	rreq->len		= folio_size(folio);
+	rreq->no_unlock_folio	= folio_index(folio);
+	__set_bit(NETFS_RREQ_NO_UNLOCK_FOLIO, &rreq->flags);
 	netfs_priv = NULL;
 
 	if (ops->begin_cache_operation) {
@@ -1147,14 +1148,14 @@ int netfs_write_begin(struct file *file, struct address_space *mapping,
 	/* Expand the request to meet caching requirements and download
 	 * preferences.
 	 */
-	ractl._nr_pages = thp_nr_pages(page);
+	ractl._nr_pages = folio_nr_pages(folio);
 	netfs_rreq_expand(rreq, &ractl);
 	netfs_get_read_request(rreq);
 
-	/* We hold the page locks, so we can drop the references */
-	while ((xpage = readahead_page(&ractl)))
-		if (xpage != page)
-			put_page(xpage);
+	/* We hold the folio locks, so we can drop the references */
+	folio_get(folio);
+	while (readahead_folio(&ractl))
+		;
 
 	atomic_set(&rreq->nr_rd_ops, 1);
 	do {
@@ -1184,22 +1185,22 @@ int netfs_write_begin(struct file *file, struct address_space *mapping,
 	if (ret < 0)
 		goto error;
 
-have_page:
-	ret = wait_on_page_fscache_killable(page);
+have_folio:
+	ret = folio_wait_fscache_killable(folio);
 	if (ret < 0)
 		goto error;
-have_page_no_wait:
+have_folio_no_wait:
 	if (netfs_priv)
 		ops->cleanup(netfs_priv, mapping);
-	*_page = page;
+	*_folio = folio;
 	_leave(" = 0");
 	return 0;
 
 error_put:
 	netfs_put_read_request(rreq, false);
 error:
-	unlock_page(page);
-	put_page(page);
+	folio_unlock(folio);
+	folio_put(folio);
 	if (netfs_priv)
 		ops->cleanup(netfs_priv, mapping);
 	_leave(" = %d", ret);
diff --git a/fs/pstore/platform.c b/fs/pstore/platform.c
index b9614db..f243cb5 100644
--- a/fs/pstore/platform.c
+++ b/fs/pstore/platform.c
@@ -218,7 +218,7 @@ static int zbufsize_842(size_t size)
 #if IS_ENABLED(CONFIG_PSTORE_ZSTD_COMPRESS)
 static int zbufsize_zstd(size_t size)
 {
-	return ZSTD_compressBound(size);
+	return zstd_compress_bound(size);
 }
 #endif
 
diff --git a/fs/squashfs/zstd_wrapper.c b/fs/squashfs/zstd_wrapper.c
index 0015cf8..c40445d 100644
--- a/fs/squashfs/zstd_wrapper.c
+++ b/fs/squashfs/zstd_wrapper.c
@@ -34,7 +34,7 @@ static void *zstd_init(struct squashfs_sb_info *msblk, void *buff)
 		goto failed;
 	wksp->window_size = max_t(size_t,
 			msblk->block_size, SQUASHFS_METADATA_SIZE);
-	wksp->mem_size = ZSTD_DStreamWorkspaceBound(wksp->window_size);
+	wksp->mem_size = zstd_dstream_workspace_bound(wksp->window_size);
 	wksp->mem = vmalloc(wksp->mem_size);
 	if (wksp->mem == NULL)
 		goto failed;
@@ -63,15 +63,15 @@ static int zstd_uncompress(struct squashfs_sb_info *msblk, void *strm,
 	struct squashfs_page_actor *output)
 {
 	struct workspace *wksp = strm;
-	ZSTD_DStream *stream;
+	zstd_dstream *stream;
 	size_t total_out = 0;
 	int error = 0;
-	ZSTD_inBuffer in_buf = { NULL, 0, 0 };
-	ZSTD_outBuffer out_buf = { NULL, 0, 0 };
+	zstd_in_buffer in_buf = { NULL, 0, 0 };
+	zstd_out_buffer out_buf = { NULL, 0, 0 };
 	struct bvec_iter_all iter_all = {};
 	struct bio_vec *bvec = bvec_init_iter_all(&iter_all);
 
-	stream = ZSTD_initDStream(wksp->window_size, wksp->mem, wksp->mem_size);
+	stream = zstd_init_dstream(wksp->window_size, wksp->mem, wksp->mem_size);
 
 	if (!stream) {
 		ERROR("Failed to initialize zstd decompressor\n");
@@ -116,14 +116,14 @@ static int zstd_uncompress(struct squashfs_sb_info *msblk, void *strm,
 		}
 
 		total_out -= out_buf.pos;
-		zstd_err = ZSTD_decompressStream(stream, &out_buf, &in_buf);
+		zstd_err = zstd_decompress_stream(stream, &out_buf, &in_buf);
 		total_out += out_buf.pos; /* add the additional data produced */
 		if (zstd_err == 0)
 			break;
 
-		if (ZSTD_isError(zstd_err)) {
+		if (zstd_is_error(zstd_err)) {
 			ERROR("zstd decompression error: %d\n",
-					(int)ZSTD_getErrorCode(zstd_err));
+					(int)zstd_get_error_code(zstd_err));
 			error = -EIO;
 			break;
 		}
diff --git a/include/drm/drm_modeset_lock.h b/include/drm/drm_modeset_lock.h
index aafd073..b84693f 100644
--- a/include/drm/drm_modeset_lock.h
+++ b/include/drm/drm_modeset_lock.h
@@ -24,6 +24,8 @@
 #ifndef DRM_MODESET_LOCK_H_
 #define DRM_MODESET_LOCK_H_
 
+#include <linux/types.h> /* stackdepot.h is not self-contained */
+#include <linux/stackdepot.h>
 #include <linux/ww_mutex.h>
 
 struct drm_modeset_lock;
@@ -52,6 +54,12 @@ struct drm_modeset_acquire_ctx {
 	struct drm_modeset_lock *contended;
 
 	/*
+	 * Stack depot for debugging when a contended lock was not backed off
+	 * from.
+	 */
+	depot_stack_handle_t stack_depot;
+
+	/*
 	 * list of held locks (drm_modeset_lock)
 	 */
 	struct list_head locked;
diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h
index 76d7c33..cd785cf 100644
--- a/include/drm/ttm/ttm_bo_api.h
+++ b/include/drm/ttm/ttm_bo_api.h
@@ -351,9 +351,10 @@ bool ttm_bo_eviction_valuable(struct ttm_buffer_object *bo,
  * @bo: Pointer to a ttm_buffer_object to be initialized.
  * @size: Requested size of buffer object.
  * @type: Requested type of buffer object.
- * @flags: Initial placement flags.
+ * @placement: Initial placement for buffer object.
  * @page_alignment: Data alignment in pages.
  * @ctx: TTM operation context for memory allocation.
+ * @sg: Scatter-gather table.
  * @resv: Pointer to a dma_resv, or NULL to let ttm allocate one.
  * @destroy: Destroy function. Use NULL for kfree().
  *
@@ -394,7 +395,7 @@ int ttm_bo_init_reserved(struct ttm_device *bdev,
  * @bo: Pointer to a ttm_buffer_object to be initialized.
  * @size: Requested size of buffer object.
  * @type: Requested type of buffer object.
- * @flags: Initial placement flags.
+ * @placement: Initial placement for buffer object.
  * @page_alignment: Data alignment in pages.
  * @interruptible: If needing to sleep to wait for GPU resources,
  * sleep interruptible.
@@ -402,6 +403,7 @@ int ttm_bo_init_reserved(struct ttm_device *bdev,
  * holds a pointer to a persistent shmem object. Typically, this would
  * point to the shmem object backing a GEM object if TTM is used to back a
  * GEM user interface.
+ * @sg: Scatter-gather table.
  * @resv: Pointer to a dma_resv, or NULL to let ttm allocate one.
  * @destroy: Destroy function. Use NULL for kfree().
  *
@@ -582,8 +584,7 @@ vm_fault_t ttm_bo_vm_reserve(struct ttm_buffer_object *bo,
 
 vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 				    pgprot_t prot,
-				    pgoff_t num_prefault,
-				    pgoff_t fault_page_size);
+				    pgoff_t num_prefault);
 
 vm_fault_t ttm_bo_vm_fault(struct vm_fault *vmf);
 
diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h
index bc2699f..7ad6c3d 100644
--- a/include/linux/ceph/ceph_fs.h
+++ b/include/linux/ceph/ceph_fs.h
@@ -302,6 +302,8 @@ enum {
 	CEPH_SESSION_REQUEST_FLUSH_MDLOG,
 };
 
+#define CEPH_SESSION_BLOCKLISTED	(1 << 0)  /* session blocklisted */
+
 extern const char *ceph_session_op_name(int op);
 
 struct ceph_mds_session_head {
diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h
index 83fa08a..3431011 100644
--- a/include/linux/ceph/osd_client.h
+++ b/include/linux/ceph/osd_client.h
@@ -475,6 +475,14 @@ extern void osd_req_op_alloc_hint_init(struct ceph_osd_request *osd_req,
 				       u64 expected_object_size,
 				       u64 expected_write_size,
 				       u32 flags);
+extern int osd_req_op_copy_from_init(struct ceph_osd_request *req,
+				     u64 src_snapid, u64 src_version,
+				     struct ceph_object_id *src_oid,
+				     struct ceph_object_locator *src_oloc,
+				     u32 src_fadvise_flags,
+				     u32 dst_fadvise_flags,
+				     u32 truncate_seq, u64 truncate_size,
+				     u8 copy_from_flags);
 
 extern struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
 					       struct ceph_snap_context *snapc,
@@ -515,17 +523,6 @@ int ceph_osdc_call(struct ceph_osd_client *osdc,
 		   struct page *req_page, size_t req_len,
 		   struct page **resp_pages, size_t *resp_len);
 
-int ceph_osdc_copy_from(struct ceph_osd_client *osdc,
-			u64 src_snapid, u64 src_version,
-			struct ceph_object_id *src_oid,
-			struct ceph_object_locator *src_oloc,
-			u32 src_fadvise_flags,
-			struct ceph_object_id *dst_oid,
-			struct ceph_object_locator *dst_oloc,
-			u32 dst_fadvise_flags,
-			u32 truncate_seq, u64 truncate_size,
-			u8 copy_from_flags);
-
 /* watch/notify */
 struct ceph_osd_linger_request *
 ceph_osdc_watch(struct ceph_osd_client *osdc,
diff --git a/include/linux/dma-resv.h b/include/linux/dma-resv.h
index 8b6c2063..dbd235a 100644
--- a/include/linux/dma-resv.h
+++ b/include/linux/dma-resv.h
@@ -170,15 +170,20 @@ struct dma_resv_iter {
 	/** @index: index into the shared fences */
 	unsigned int index;
 
-	/** @fences: the shared fences */
+	/** @fences: the shared fences; private, *MUST* not dereference  */
 	struct dma_resv_list *fences;
 
+	/** @shared_count: number of shared fences */
+	unsigned int shared_count;
+
 	/** @is_restarted: true if this is the first returned fence */
 	bool is_restarted;
 };
 
 struct dma_fence *dma_resv_iter_first_unlocked(struct dma_resv_iter *cursor);
 struct dma_fence *dma_resv_iter_next_unlocked(struct dma_resv_iter *cursor);
+struct dma_fence *dma_resv_iter_first(struct dma_resv_iter *cursor);
+struct dma_fence *dma_resv_iter_next(struct dma_resv_iter *cursor);
 
 /**
  * dma_resv_iter_begin - initialize a dma_resv_iter object
@@ -244,6 +249,24 @@ static inline bool dma_resv_iter_is_restarted(struct dma_resv_iter *cursor)
 	for (fence = dma_resv_iter_first_unlocked(cursor);		\
 	     fence; fence = dma_resv_iter_next_unlocked(cursor))
 
+/**
+ * dma_resv_for_each_fence - fence iterator
+ * @cursor: a struct dma_resv_iter pointer
+ * @obj: a dma_resv object pointer
+ * @all_fences: true if all fences should be returned
+ * @fence: the current fence
+ *
+ * Iterate over the fences in a struct dma_resv object while holding the
+ * &dma_resv.lock. @all_fences controls if the shared fences are returned as
+ * well. The cursor initialisation is part of the iterator and the fence stays
+ * valid as long as the lock is held and so no extra reference to the fence is
+ * taken.
+ */
+#define dma_resv_for_each_fence(cursor, obj, all_fences, fence)	\
+	for (dma_resv_iter_begin(cursor, obj, all_fences),	\
+	     fence = dma_resv_iter_first(cursor); fence;	\
+	     fence = dma_resv_iter_next(cursor))
+
 #define dma_resv_held(obj) lockdep_is_held(&(obj)->lock.base)
 #define dma_resv_assert_held(obj) lockdep_assert_held(&(obj)->lock.base)
 
diff --git a/include/linux/efi.h b/include/linux/efi.h
index 6b5d36b..dbd39b2 100644
--- a/include/linux/efi.h
+++ b/include/linux/efi.h
@@ -362,6 +362,7 @@ void efi_native_runtime_setup(void);
 
 /* OEM GUIDs */
 #define DELLEMC_EFI_RCI2_TABLE_GUID		EFI_GUID(0x2d9f28a2, 0xa886, 0x456a,  0x97, 0xa8, 0xf1, 0x1e, 0xf2, 0x4f, 0xf4, 0x55)
+#define AMD_SEV_MEM_ENCRYPT_GUID		EFI_GUID(0x0cf29b71, 0x9e51, 0x433a,  0xa3, 0xb7, 0x81, 0xf3, 0xab, 0x16, 0xb8, 0x75)
 
 typedef struct {
 	efi_guid_t guid;
diff --git a/include/linux/fb.h b/include/linux/fb.h
index 5950f8f..6f3db99 100644
--- a/include/linux/fb.h
+++ b/include/linux/fb.h
@@ -262,7 +262,7 @@ struct fb_ops {
 
 	/* Draws a rectangle */
 	void (*fb_fillrect) (struct fb_info *info, const struct fb_fillrect *rect);
-	/* Copy data from area to another */
+	/* Copy data from area to another. Obsolete. */
 	void (*fb_copyarea) (struct fb_info *info, const struct fb_copyarea *region);
 	/* Draws a image to the display */
 	void (*fb_imageblit) (struct fb_info *info, const struct fb_image *image);
diff --git a/include/linux/genhd.h b/include/linux/genhd.h
index 462634b..74c4102 100644
--- a/include/linux/genhd.h
+++ b/include/linux/genhd.h
@@ -205,9 +205,9 @@ static inline dev_t disk_devt(struct gendisk *disk)
 void disk_uevent(struct gendisk *disk, enum kobject_action action);
 
 /* block/genhd.c */
-int device_add_disk(struct device *parent, struct gendisk *disk,
-		const struct attribute_group **groups);
-static inline int add_disk(struct gendisk *disk)
+int __must_check device_add_disk(struct device *parent, struct gendisk *disk,
+				 const struct attribute_group **groups);
+static inline int __must_check add_disk(struct gendisk *disk)
 {
 	return device_add_disk(NULL, disk, NULL);
 }
diff --git a/include/linux/input/cy8ctmg110_pdata.h b/include/linux/input/cy8ctmg110_pdata.h
deleted file mode 100644
index ee1d445..0000000
--- a/include/linux/input/cy8ctmg110_pdata.h
+++ /dev/null
@@ -1,10 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0 */
-#ifndef _LINUX_CY8CTMG110_PDATA_H
-#define _LINUX_CY8CTMG110_PDATA_H
-
-struct cy8ctmg110_pdata
-{
-	int reset_pin;		/* Reset pin is wired to this GPIO (optional) */
-};
-
-#endif
diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index 968b4c4..77755ac 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -85,7 +85,7 @@
 struct completion;
 struct user;
 
-#ifdef CONFIG_PREEMPT_VOLUNTARY
+#ifdef CONFIG_PREEMPT_VOLUNTARY_BUILD
 
 extern int __cond_resched(void);
 # define might_resched() __cond_resched()
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 60a35d9..9e0667e 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -150,7 +150,7 @@ static inline bool is_error_page(struct page *page)
 #define KVM_REQ_MMU_RELOAD        (1 | KVM_REQUEST_WAIT | KVM_REQUEST_NO_WAKEUP)
 #define KVM_REQ_UNBLOCK           2
 #define KVM_REQ_UNHALT            3
-#define KVM_REQ_VM_BUGGED         (4 | KVM_REQUEST_WAIT | KVM_REQUEST_NO_WAKEUP)
+#define KVM_REQ_VM_DEAD           (4 | KVM_REQUEST_WAIT | KVM_REQUEST_NO_WAKEUP)
 #define KVM_REQUEST_ARCH_BASE     8
 
 #define KVM_ARCH_REQ_FLAGS(nr, flags) ({ \
@@ -617,6 +617,7 @@ struct kvm {
 	unsigned int max_halt_poll_ns;
 	u32 dirty_ring_size;
 	bool vm_bugged;
+	bool vm_dead;
 
 #ifdef CONFIG_HAVE_KVM_PM_NOTIFIER
 	struct notifier_block pm_notifier;
@@ -650,12 +651,19 @@ struct kvm {
 #define vcpu_err(vcpu, fmt, ...)					\
 	kvm_err("vcpu%i " fmt, (vcpu)->vcpu_id, ## __VA_ARGS__)
 
+static inline void kvm_vm_dead(struct kvm *kvm)
+{
+	kvm->vm_dead = true;
+	kvm_make_all_cpus_request(kvm, KVM_REQ_VM_DEAD);
+}
+
 static inline void kvm_vm_bugged(struct kvm *kvm)
 {
 	kvm->vm_bugged = true;
-	kvm_make_all_cpus_request(kvm, KVM_REQ_VM_BUGGED);
+	kvm_vm_dead(kvm);
 }
 
+
 #define KVM_BUG(cond, kvm, fmt...)				\
 ({								\
 	int __ret = (cond);					\
diff --git a/include/linux/libata.h b/include/linux/libata.h
index 2884383..2a8404b 100644
--- a/include/linux/libata.h
+++ b/include/linux/libata.h
@@ -394,7 +394,7 @@ enum {
 	/* This should match the actual table size of
 	 * ata_eh_cmd_timeout_table in libata-eh.c.
 	 */
-	ATA_EH_CMD_TIMEOUT_TABLE_SIZE = 6,
+	ATA_EH_CMD_TIMEOUT_TABLE_SIZE = 7,
 
 	/* Horkage types. May be set by libata or controller on drives
 	   (some horkage may be drive/controller pair dependent */
@@ -427,6 +427,7 @@ enum {
 	ATA_HORKAGE_MAX_SEC_1024 = (1 << 25),	/* Limit max sects to 1024 */
 	ATA_HORKAGE_MAX_TRIM_128M = (1 << 26),	/* Limit max trim size to 128M */
 	ATA_HORKAGE_NO_NCQ_ON_ATI = (1 << 27),	/* Disable NCQ on ATI chipset */
+	ATA_HORKAGE_NO_ID_DEV_LOG = (1 << 28),	/* Identify device log missing */
 
 	 /* DMA mask for user DMA control: User visible values; DO NOT
 	    renumber */
diff --git a/include/linux/lsm_hook_defs.h b/include/linux/lsm_hook_defs.h
index 442a611..df8de62 100644
--- a/include/linux/lsm_hook_defs.h
+++ b/include/linux/lsm_hook_defs.h
@@ -335,8 +335,6 @@ LSM_HOOK(int, 0, sctp_bind_connect, struct sock *sk, int optname,
 	 struct sockaddr *address, int addrlen)
 LSM_HOOK(void, LSM_RET_VOID, sctp_sk_clone, struct sctp_association *asoc,
 	 struct sock *sk, struct sock *newsk)
-LSM_HOOK(void, LSM_RET_VOID, sctp_assoc_established, struct sctp_association *asoc,
-	 struct sk_buff *skb)
 #endif /* CONFIG_SECURITY_NETWORK */
 
 #ifdef CONFIG_SECURITY_INFINIBAND
diff --git a/include/linux/lsm_hooks.h b/include/linux/lsm_hooks.h
index d682321..d45b6f6 100644
--- a/include/linux/lsm_hooks.h
+++ b/include/linux/lsm_hooks.h
@@ -1050,11 +1050,6 @@
  *	@asoc pointer to current sctp association structure.
  *	@sk pointer to current sock structure.
  *	@newsk pointer to new sock structure.
- * @sctp_assoc_established:
- *	Passes the @asoc and @chunk->skb of the association COOKIE_ACK packet
- *	to the security module.
- *	@asoc pointer to sctp association structure.
- *	@skb pointer to skbuff of association packet.
  *
  * Security hooks for Infiniband
  *
diff --git a/include/linux/msi.h b/include/linux/msi.h
index 49cf6eb..e616f94 100644
--- a/include/linux/msi.h
+++ b/include/linux/msi.h
@@ -148,7 +148,7 @@ struct msi_desc {
 				u8	is_msix		: 1;
 				u8	multiple	: 3;
 				u8	multi_cap	: 3;
-				u8	maskbit		: 1;
+				u8	can_mask	: 1;
 				u8	is_64		: 1;
 				u8	is_virtual	: 1;
 				u16	entry_nr;
diff --git a/include/linux/netfs.h b/include/linux/netfs.h
index 12c4177..ca0683b 100644
--- a/include/linux/netfs.h
+++ b/include/linux/netfs.h
@@ -166,13 +166,13 @@ struct netfs_read_request {
 	short			error;		/* 0 or error that occurred */
 	loff_t			i_size;		/* Size of the file */
 	loff_t			start;		/* Start position */
-	pgoff_t			no_unlock_page;	/* Don't unlock this page after read */
+	pgoff_t			no_unlock_folio; /* Don't unlock this folio after read */
 	refcount_t		usage;
 	unsigned long		flags;
 #define NETFS_RREQ_INCOMPLETE_IO	0	/* Some ioreqs terminated short or with error */
 #define NETFS_RREQ_WRITE_TO_CACHE	1	/* Need to write to the cache */
-#define NETFS_RREQ_NO_UNLOCK_PAGE	2	/* Don't unlock no_unlock_page on completion */
-#define NETFS_RREQ_DONT_UNLOCK_PAGES	3	/* Don't unlock the pages on completion */
+#define NETFS_RREQ_NO_UNLOCK_FOLIO	2	/* Don't unlock no_unlock_folio on completion */
+#define NETFS_RREQ_DONT_UNLOCK_FOLIOS	3	/* Don't unlock the folios on completion */
 #define NETFS_RREQ_FAILED		4	/* The request failed */
 #define NETFS_RREQ_IN_PROGRESS		5	/* Unlocked when the request completes */
 	const struct netfs_read_request_ops *netfs_ops;
@@ -190,7 +190,7 @@ struct netfs_read_request_ops {
 	void (*issue_op)(struct netfs_read_subrequest *subreq);
 	bool (*is_still_valid)(struct netfs_read_request *rreq);
 	int (*check_write_begin)(struct file *file, loff_t pos, unsigned len,
-				 struct page *page, void **_fsdata);
+				 struct folio *folio, void **_fsdata);
 	void (*done)(struct netfs_read_request *rreq);
 	void (*cleanup)(struct address_space *mapping, void *netfs_priv);
 };
@@ -240,11 +240,11 @@ extern void netfs_readahead(struct readahead_control *,
 			    const struct netfs_read_request_ops *,
 			    void *);
 extern int netfs_readpage(struct file *,
-			  struct page *,
+			  struct folio *,
 			  const struct netfs_read_request_ops *,
 			  void *);
 extern int netfs_write_begin(struct file *, struct address_space *,
-			     loff_t, unsigned int, unsigned int, struct page **,
+			     loff_t, unsigned int, unsigned int, struct folio **,
 			     void **,
 			     const struct netfs_read_request_ops *,
 			     void *);
diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
index 6a30916..1a0c646 100644
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -253,6 +253,20 @@ static inline struct address_space *page_mapping_file(struct page *page)
 	return folio_mapping(folio);
 }
 
+/**
+ * folio_inode - Get the host inode for this folio.
+ * @folio: The folio.
+ *
+ * For folios which are in the page cache, return the inode that this folio
+ * belongs to.
+ *
+ * Do not call this for folios which aren't in the page cache.
+ */
+static inline struct inode *folio_inode(struct folio *folio)
+{
+	return folio->mapping->host;
+}
+
 static inline bool page_cache_add_speculative(struct page *page, int count)
 {
 	VM_BUG_ON_PAGE(PageTail(page), page);
@@ -280,6 +294,25 @@ static inline void folio_attach_private(struct folio *folio, void *data)
 }
 
 /**
+ * folio_change_private - Change private data on a folio.
+ * @folio: Folio to change the data on.
+ * @data: Data to set on the folio.
+ *
+ * Change the private data attached to a folio and return the old
+ * data.  The page must previously have had data attached and the data
+ * must be detached before the folio will be freed.
+ *
+ * Return: Data that was previously attached to the folio.
+ */
+static inline void *folio_change_private(struct folio *folio, void *data)
+{
+	void *old = folio_get_private(folio);
+
+	folio->private = data;
+	return old;
+}
+
+/**
  * folio_detach_private - Detach private data from a folio.
  * @folio: Folio to detach data from.
  *
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 138d764..18a75c8e 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -233,6 +233,8 @@ enum pci_dev_flags {
 	PCI_DEV_FLAGS_NO_FLR_RESET = (__force pci_dev_flags_t) (1 << 10),
 	/* Don't use Relaxed Ordering for TLPs directed at this device */
 	PCI_DEV_FLAGS_NO_RELAXED_ORDERING = (__force pci_dev_flags_t) (1 << 11),
+	/* Device does honor MSI masking despite saying otherwise */
+	PCI_DEV_FLAGS_HAS_MSI_MASKING = (__force pci_dev_flags_t) (1 << 12),
 };
 
 enum pci_irq_reroute_variant {
@@ -1666,6 +1668,7 @@ void pci_cfg_access_lock(struct pci_dev *dev);
 bool pci_cfg_access_trylock(struct pci_dev *dev);
 void pci_cfg_access_unlock(struct pci_dev *dev);
 
+void pci_dev_lock(struct pci_dev *dev);
 int pci_dev_trylock(struct pci_dev *dev);
 void pci_dev_unlock(struct pci_dev *dev);
 
diff --git a/include/linux/posix-timers.h b/include/linux/posix-timers.h
index 00fef00..5bbcd28 100644
--- a/include/linux/posix-timers.h
+++ b/include/linux/posix-timers.h
@@ -184,8 +184,10 @@ static inline void posix_cputimers_group_init(struct posix_cputimers *pct,
 #endif
 
 #ifdef CONFIG_POSIX_CPU_TIMERS_TASK_WORK
+void clear_posix_cputimers_work(struct task_struct *p);
 void posix_cputimers_init_work(void);
 #else
+static inline void clear_posix_cputimers_work(struct task_struct *p) { }
 static inline void posix_cputimers_init_work(void) { }
 #endif
 
diff --git a/include/linux/pwm.h b/include/linux/pwm.h
index 725c9b7..e6dac95 100644
--- a/include/linux/pwm.h
+++ b/include/linux/pwm.h
@@ -429,16 +429,19 @@ struct pwm_device *devm_fwnode_pwm_get(struct device *dev,
 #else
 static inline struct pwm_device *pwm_request(int pwm_id, const char *label)
 {
+	might_sleep();
 	return ERR_PTR(-ENODEV);
 }
 
 static inline void pwm_free(struct pwm_device *pwm)
 {
+	might_sleep();
 }
 
 static inline int pwm_apply_state(struct pwm_device *pwm,
 				  const struct pwm_state *state)
 {
+	might_sleep();
 	return -ENOTSUPP;
 }
 
@@ -450,6 +453,7 @@ static inline int pwm_adjust_config(struct pwm_device *pwm)
 static inline int pwm_config(struct pwm_device *pwm, int duty_ns,
 			     int period_ns)
 {
+	might_sleep();
 	return -EINVAL;
 }
 
@@ -462,11 +466,13 @@ static inline int pwm_capture(struct pwm_device *pwm,
 
 static inline int pwm_enable(struct pwm_device *pwm)
 {
+	might_sleep();
 	return -EINVAL;
 }
 
 static inline void pwm_disable(struct pwm_device *pwm)
 {
+	might_sleep();
 }
 
 static inline int pwm_set_chip_data(struct pwm_device *pwm, void *data)
@@ -493,12 +499,14 @@ static inline struct pwm_device *pwm_request_from_chip(struct pwm_chip *chip,
 						       unsigned int index,
 						       const char *label)
 {
+	might_sleep();
 	return ERR_PTR(-ENODEV);
 }
 
 static inline struct pwm_device *pwm_get(struct device *dev,
 					 const char *consumer)
 {
+	might_sleep();
 	return ERR_PTR(-ENODEV);
 }
 
@@ -506,16 +514,19 @@ static inline struct pwm_device *of_pwm_get(struct device *dev,
 					    struct device_node *np,
 					    const char *con_id)
 {
+	might_sleep();
 	return ERR_PTR(-ENODEV);
 }
 
 static inline void pwm_put(struct pwm_device *pwm)
 {
+	might_sleep();
 }
 
 static inline struct pwm_device *devm_pwm_get(struct device *dev,
 					      const char *consumer)
 {
+	might_sleep();
 	return ERR_PTR(-ENODEV);
 }
 
@@ -523,6 +534,7 @@ static inline struct pwm_device *devm_of_pwm_get(struct device *dev,
 						 struct device_node *np,
 						 const char *con_id)
 {
+	might_sleep();
 	return ERR_PTR(-ENODEV);
 }
 
@@ -530,6 +542,7 @@ static inline struct pwm_device *
 devm_fwnode_pwm_get(struct device *dev, struct fwnode_handle *fwnode,
 		    const char *con_id)
 {
+	might_sleep();
 	return ERR_PTR(-ENODEV);
 }
 #endif
diff --git a/include/linux/rtc.h b/include/linux/rtc.h
index bd611e2..47fd1c2 100644
--- a/include/linux/rtc.h
+++ b/include/linux/rtc.h
@@ -66,6 +66,8 @@ struct rtc_class_ops {
 	int (*alarm_irq_enable)(struct device *, unsigned int enabled);
 	int (*read_offset)(struct device *, long *offset);
 	int (*set_offset)(struct device *, long offset);
+	int (*param_get)(struct device *, struct rtc_param *param);
+	int (*param_set)(struct device *, struct rtc_param *param);
 };
 
 struct rtc_device;
@@ -80,6 +82,7 @@ struct rtc_timer {
 
 /* flags */
 #define RTC_DEV_BUSY 0
+#define RTC_NO_CDEV  1
 
 struct rtc_device {
 	struct device dev;
diff --git a/include/linux/security.h b/include/linux/security.h
index 06eac4e..bbf44a4 100644
--- a/include/linux/security.h
+++ b/include/linux/security.h
@@ -1430,8 +1430,6 @@ int security_sctp_bind_connect(struct sock *sk, int optname,
 			       struct sockaddr *address, int addrlen);
 void security_sctp_sk_clone(struct sctp_association *asoc, struct sock *sk,
 			    struct sock *newsk);
-void security_sctp_assoc_established(struct sctp_association *asoc,
-				     struct sk_buff *skb);
 
 #else	/* CONFIG_SECURITY_NETWORK */
 static inline int security_unix_stream_connect(struct sock *sock,
@@ -1651,11 +1649,6 @@ static inline void security_sctp_sk_clone(struct sctp_association *asoc,
 					  struct sock *newsk)
 {
 }
-
-static inline void security_sctp_assoc_established(struct sctp_association *asoc,
-						   struct sk_buff *skb)
-{
-}
 #endif	/* CONFIG_SECURITY_NETWORK */
 
 #ifdef CONFIG_SECURITY_INFINIBAND
diff --git a/include/linux/spi/ads7846.h b/include/linux/spi/ads7846.h
index 1a5eaef..d424c1a 100644
--- a/include/linux/spi/ads7846.h
+++ b/include/linux/spi/ads7846.h
@@ -1,17 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0 */
 /* linux/spi/ads7846.h */
 
-/* Touchscreen characteristics vary between boards and models.  The
- * platform_data for the device's "struct device" holds this information.
- *
- * It's OK if the min/max values are zero.
- */
-enum ads7846_filter {
-	ADS7846_FILTER_OK,
-	ADS7846_FILTER_REPEAT,
-	ADS7846_FILTER_IGNORE,
-};
-
 struct ads7846_platform_data {
 	u16	model;			/* 7843, 7845, 7846, 7873. */
 	u16	vref_delay_usecs;	/* 0 for external vref; etc */
@@ -51,10 +40,6 @@ struct ads7846_platform_data {
 	int	gpio_pendown_debounce;	/* platform specific debounce time for
 					 * the gpio_pendown */
 	int	(*get_pendown_state)(void);
-	int	(*filter_init)	(const struct ads7846_platform_data *pdata,
-				 void **filter_data);
-	int	(*filter)	(void *filter_data, int data_idx, int *val);
-	void	(*filter_cleanup)(void *filter_data);
 	void	(*wait_for_sync)(void);
 	bool	wakeup;
 	unsigned long irq_flags;
diff --git a/include/linux/vermagic.h b/include/linux/vermagic.h
index 1eaaa93..329d63b 100644
--- a/include/linux/vermagic.h
+++ b/include/linux/vermagic.h
@@ -15,7 +15,7 @@
 #else
 #define MODULE_VERMAGIC_SMP ""
 #endif
-#ifdef CONFIG_PREEMPT
+#ifdef CONFIG_PREEMPT_BUILD
 #define MODULE_VERMAGIC_PREEMPT "preempt "
 #elif defined(CONFIG_PREEMPT_RT)
 #define MODULE_VERMAGIC_PREEMPT "preempt_rt "
diff --git a/include/linux/zstd.h b/include/linux/zstd.h
index e87f78c..113408e 100644
--- a/include/linux/zstd.h
+++ b/include/linux/zstd.h
@@ -1,138 +1,97 @@
+/* SPDX-License-Identifier: GPL-2.0+ OR BSD-3-Clause */
 /*
- * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * Copyright (c) Yann Collet, Facebook, Inc.
  * All rights reserved.
  *
- * This source code is licensed under the BSD-style license found in the
- * LICENSE file in the root directory of https://github.com/facebook/zstd.
- * An additional grant of patent rights can be found in the PATENTS file in the
- * same directory.
- *
- * This program is free software; you can redistribute it and/or modify it under
- * the terms of the GNU General Public License version 2 as published by the
- * Free Software Foundation. This program is dual-licensed; you may select
- * either version 2 of the GNU General Public License ("GPL") or BSD license
- * ("BSD").
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of https://github.com/facebook/zstd) and
+ * the GPLv2 (found in the COPYING file in the root directory of
+ * https://github.com/facebook/zstd). You may select, at your option, one of the
+ * above-listed licenses.
  */
 
-#ifndef ZSTD_H
-#define ZSTD_H
-
-/* ======   Dependency   ======*/
-#include <linux/types.h>   /* size_t */
-
-
-/*-*****************************************************************************
- * Introduction
- *
- * zstd, short for Zstandard, is a fast lossless compression algorithm,
- * targeting real-time compression scenarios at zlib-level and better
- * compression ratios. The zstd compression library provides in-memory
- * compression and decompression functions. The library supports compression
- * levels from 1 up to ZSTD_maxCLevel() which is 22. Levels >= 20, labeled
- * ultra, should be used with caution, as they require more memory.
- * Compression can be done in:
- *  - a single step, reusing a context (described as Explicit memory management)
- *  - unbounded multiple steps (described as Streaming compression)
- * The compression ratio achievable on small data can be highly improved using
- * compression with a dictionary in:
- *  - a single step (described as Simple dictionary API)
- *  - a single step, reusing a dictionary (described as Fast dictionary API)
- ******************************************************************************/
-
-/*======  Helper functions  ======*/
+#ifndef LINUX_ZSTD_H
+#define LINUX_ZSTD_H
 
 /**
- * enum ZSTD_ErrorCode - zstd error codes
- *
- * Functions that return size_t can be checked for errors using ZSTD_isError()
- * and the ZSTD_ErrorCode can be extracted using ZSTD_getErrorCode().
+ * This is a kernel-style API that wraps the upstream zstd API, which cannot be
+ * used directly because the symbols aren't exported. It exposes the minimal
+ * functionality which is currently required by users of zstd in the kernel.
+ * Expose extra functions from lib/zstd/zstd.h as needed.
  */
-typedef enum {
-	ZSTD_error_no_error,
-	ZSTD_error_GENERIC,
-	ZSTD_error_prefix_unknown,
-	ZSTD_error_version_unsupported,
-	ZSTD_error_parameter_unknown,
-	ZSTD_error_frameParameter_unsupported,
-	ZSTD_error_frameParameter_unsupportedBy32bits,
-	ZSTD_error_frameParameter_windowTooLarge,
-	ZSTD_error_compressionParameter_unsupported,
-	ZSTD_error_init_missing,
-	ZSTD_error_memory_allocation,
-	ZSTD_error_stage_wrong,
-	ZSTD_error_dstSize_tooSmall,
-	ZSTD_error_srcSize_wrong,
-	ZSTD_error_corruption_detected,
-	ZSTD_error_checksum_wrong,
-	ZSTD_error_tableLog_tooLarge,
-	ZSTD_error_maxSymbolValue_tooLarge,
-	ZSTD_error_maxSymbolValue_tooSmall,
-	ZSTD_error_dictionary_corrupted,
-	ZSTD_error_dictionary_wrong,
-	ZSTD_error_dictionaryCreation_failed,
-	ZSTD_error_maxCode
-} ZSTD_ErrorCode;
+
+/* ======   Dependency   ====== */
+#include <linux/types.h>
+#include <linux/zstd_errors.h>
+#include <linux/zstd_lib.h>
+
+/* ======   Helper Functions   ====== */
+/**
+ * zstd_compress_bound() - maximum compressed size in worst case scenario
+ * @src_size: The size of the data to compress.
+ *
+ * Return:    The maximum compressed size in the worst case scenario.
+ */
+size_t zstd_compress_bound(size_t src_size);
 
 /**
- * ZSTD_maxCLevel() - maximum compression level available
- *
- * Return: Maximum compression level available.
- */
-int ZSTD_maxCLevel(void);
-/**
- * ZSTD_compressBound() - maximum compressed size in worst case scenario
- * @srcSize: The size of the data to compress.
- *
- * Return:   The maximum compressed size in the worst case scenario.
- */
-size_t ZSTD_compressBound(size_t srcSize);
-/**
- * ZSTD_isError() - tells if a size_t function result is an error code
+ * zstd_is_error() - tells if a size_t function result is an error code
  * @code:  The function result to check for error.
  *
  * Return: Non-zero iff the code is an error.
  */
-static __attribute__((unused)) unsigned int ZSTD_isError(size_t code)
-{
-	return code > (size_t)-ZSTD_error_maxCode;
-}
-/**
- * ZSTD_getErrorCode() - translates an error function result to a ZSTD_ErrorCode
- * @functionResult: The result of a function for which ZSTD_isError() is true.
- *
- * Return:          The ZSTD_ErrorCode corresponding to the functionResult or 0
- *                  if the functionResult isn't an error.
- */
-static __attribute__((unused)) ZSTD_ErrorCode ZSTD_getErrorCode(
-	size_t functionResult)
-{
-	if (!ZSTD_isError(functionResult))
-		return (ZSTD_ErrorCode)0;
-	return (ZSTD_ErrorCode)(0 - functionResult);
-}
+unsigned int zstd_is_error(size_t code);
 
 /**
- * enum ZSTD_strategy - zstd compression search strategy
- *
- * From faster to stronger.
+ * enum zstd_error_code - zstd error codes
  */
-typedef enum {
-	ZSTD_fast,
-	ZSTD_dfast,
-	ZSTD_greedy,
-	ZSTD_lazy,
-	ZSTD_lazy2,
-	ZSTD_btlazy2,
-	ZSTD_btopt,
-	ZSTD_btopt2
-} ZSTD_strategy;
+typedef ZSTD_ErrorCode zstd_error_code;
 
 /**
- * struct ZSTD_compressionParameters - zstd compression parameters
+ * zstd_get_error_code() - translates an error function result to an error code
+ * @code:  The function result for which zstd_is_error(code) is true.
+ *
+ * Return: A unique error code for this error.
+ */
+zstd_error_code zstd_get_error_code(size_t code);
+
+/**
+ * zstd_get_error_name() - translates an error function result to a string
+ * @code:  The function result for which zstd_is_error(code) is true.
+ *
+ * Return: An error string corresponding to the error code.
+ */
+const char *zstd_get_error_name(size_t code);
+
+/**
+ * zstd_min_clevel() - minimum allowed compression level
+ *
+ * Return: The minimum allowed compression level.
+ */
+int zstd_min_clevel(void);
+
+/**
+ * zstd_max_clevel() - maximum allowed compression level
+ *
+ * Return: The maximum allowed compression level.
+ */
+int zstd_max_clevel(void);
+
+/* ======   Parameter Selection   ====== */
+
+/**
+ * enum zstd_strategy - zstd compression search strategy
+ *
+ * From faster to stronger. See zstd_lib.h.
+ */
+typedef ZSTD_strategy zstd_strategy;
+
+/**
+ * struct zstd_compression_parameters - zstd compression parameters
  * @windowLog:    Log of the largest match distance. Larger means more
  *                compression, and more memory needed during decompression.
- * @chainLog:     Fully searched segment. Larger means more compression, slower,
- *                and more memory (useless for fast).
+ * @chainLog:     Fully searched segment. Larger means more compression,
+ *                slower, and more memory (useless for fast).
  * @hashLog:      Dispatch table. Larger means more compression,
  *                slower, and more memory.
  * @searchLog:    Number of searches. Larger means more compression and slower.
@@ -141,1017 +100,348 @@ typedef enum {
  * @targetLength: Acceptable match size for optimal parser (only). Larger means
  *                more compression, and slower.
  * @strategy:     The zstd compression strategy.
- */
-typedef struct {
-	unsigned int windowLog;
-	unsigned int chainLog;
-	unsigned int hashLog;
-	unsigned int searchLog;
-	unsigned int searchLength;
-	unsigned int targetLength;
-	ZSTD_strategy strategy;
-} ZSTD_compressionParameters;
-
-/**
- * struct ZSTD_frameParameters - zstd frame parameters
- * @contentSizeFlag: Controls whether content size will be present in the frame
- *                   header (when known).
- * @checksumFlag:    Controls whether a 32-bit checksum is generated at the end
- *                   of the frame for error detection.
- * @noDictIDFlag:    Controls whether dictID will be saved into the frame header
- *                   when using dictionary compression.
  *
- * The default value is all fields set to 0.
+ * See zstd_lib.h.
  */
-typedef struct {
-	unsigned int contentSizeFlag;
-	unsigned int checksumFlag;
-	unsigned int noDictIDFlag;
-} ZSTD_frameParameters;
+typedef ZSTD_compressionParameters zstd_compression_parameters;
 
 /**
- * struct ZSTD_parameters - zstd parameters
+ * struct zstd_frame_parameters - zstd frame parameters
+ * @contentSizeFlag: Controls whether content size will be present in the
+ *                   frame header (when known).
+ * @checksumFlag:    Controls whether a 32-bit checksum is generated at the
+ *                   end of the frame for error detection.
+ * @noDictIDFlag:    Controls whether dictID will be saved into the frame
+ *                   header when using dictionary compression.
+ *
+ * The default value is all fields set to 0. See zstd_lib.h.
+ */
+typedef ZSTD_frameParameters zstd_frame_parameters;
+
+/**
+ * struct zstd_parameters - zstd parameters
  * @cParams: The compression parameters.
  * @fParams: The frame parameters.
  */
-typedef struct {
-	ZSTD_compressionParameters cParams;
-	ZSTD_frameParameters fParams;
-} ZSTD_parameters;
+typedef ZSTD_parameters zstd_parameters;
 
 /**
- * ZSTD_getCParams() - returns ZSTD_compressionParameters for selected level
- * @compressionLevel: The compression level from 1 to ZSTD_maxCLevel().
- * @estimatedSrcSize: The estimated source size to compress or 0 if unknown.
- * @dictSize:         The dictionary size or 0 if a dictionary isn't being used.
+ * zstd_get_params() - returns zstd_parameters for selected level
+ * @level:              The compression level
+ * @estimated_src_size: The estimated source size to compress or 0
+ *                      if unknown.
  *
- * Return:            The selected ZSTD_compressionParameters.
+ * Return:              The selected zstd_parameters.
  */
-ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel,
-	unsigned long long estimatedSrcSize, size_t dictSize);
+zstd_parameters zstd_get_params(int level,
+	unsigned long long estimated_src_size);
+
+/* ======   Single-pass Compression   ====== */
+
+typedef ZSTD_CCtx zstd_cctx;
 
 /**
- * ZSTD_getParams() - returns ZSTD_parameters for selected level
- * @compressionLevel: The compression level from 1 to ZSTD_maxCLevel().
- * @estimatedSrcSize: The estimated source size to compress or 0 if unknown.
- * @dictSize:         The dictionary size or 0 if a dictionary isn't being used.
- *
- * The same as ZSTD_getCParams() except also selects the default frame
- * parameters (all zero).
- *
- * Return:            The selected ZSTD_parameters.
- */
-ZSTD_parameters ZSTD_getParams(int compressionLevel,
-	unsigned long long estimatedSrcSize, size_t dictSize);
-
-/*-*************************************
- * Explicit memory management
- **************************************/
-
-/**
- * ZSTD_CCtxWorkspaceBound() - amount of memory needed to initialize a ZSTD_CCtx
- * @cParams: The compression parameters to be used for compression.
+ * zstd_cctx_workspace_bound() - max memory needed to initialize a zstd_cctx
+ * @parameters: The compression parameters to be used.
  *
  * If multiple compression parameters might be used, the caller must call
- * ZSTD_CCtxWorkspaceBound() for each set of parameters and use the maximum
+ * zstd_cctx_workspace_bound() for each set of parameters and use the maximum
  * size.
  *
- * Return:   A lower bound on the size of the workspace that is passed to
- *           ZSTD_initCCtx().
+ * Return:      A lower bound on the size of the workspace that is passed to
+ *              zstd_init_cctx().
  */
-size_t ZSTD_CCtxWorkspaceBound(ZSTD_compressionParameters cParams);
+size_t zstd_cctx_workspace_bound(const zstd_compression_parameters *parameters);
 
 /**
- * struct ZSTD_CCtx - the zstd compression context
+ * zstd_init_cctx() - initialize a zstd compression context
+ * @workspace:      The workspace to emplace the context into. It must outlive
+ *                  the returned context.
+ * @workspace_size: The size of workspace. Use zstd_cctx_workspace_bound() to
+ *                  determine how large the workspace must be.
  *
- * When compressing many times it is recommended to allocate a context just once
- * and reuse it for each successive compression operation.
+ * Return:          A zstd compression context or NULL on error.
  */
-typedef struct ZSTD_CCtx_s ZSTD_CCtx;
-/**
- * ZSTD_initCCtx() - initialize a zstd compression context
- * @workspace:     The workspace to emplace the context into. It must outlive
- *                 the returned context.
- * @workspaceSize: The size of workspace. Use ZSTD_CCtxWorkspaceBound() to
- *                 determine how large the workspace must be.
- *
- * Return:         A compression context emplaced into workspace.
- */
-ZSTD_CCtx *ZSTD_initCCtx(void *workspace, size_t workspaceSize);
+zstd_cctx *zstd_init_cctx(void *workspace, size_t workspace_size);
 
 /**
- * ZSTD_compressCCtx() - compress src into dst
- * @ctx:         The context. Must have been initialized with a workspace at
- *               least as large as ZSTD_CCtxWorkspaceBound(params.cParams).
- * @dst:         The buffer to compress src into.
- * @dstCapacity: The size of the destination buffer. May be any size, but
- *               ZSTD_compressBound(srcSize) is guaranteed to be large enough.
- * @src:         The data to compress.
- * @srcSize:     The size of the data to compress.
- * @params:      The parameters to use for compression. See ZSTD_getParams().
+ * zstd_compress_cctx() - compress src into dst with the initialized parameters
+ * @cctx:         The context. Must have been initialized with zstd_init_cctx().
+ * @dst:          The buffer to compress src into.
+ * @dst_capacity: The size of the destination buffer. May be any size, but
+ *                ZSTD_compressBound(srcSize) is guaranteed to be large enough.
+ * @src:          The data to compress.
+ * @src_size:     The size of the data to compress.
+ * @parameters:   The compression parameters to be used.
  *
- * Return:       The compressed size or an error, which can be checked using
- *               ZSTD_isError().
+ * Return:        The compressed size or an error, which can be checked using
+ *                zstd_is_error().
  */
-size_t ZSTD_compressCCtx(ZSTD_CCtx *ctx, void *dst, size_t dstCapacity,
-	const void *src, size_t srcSize, ZSTD_parameters params);
+size_t zstd_compress_cctx(zstd_cctx *cctx, void *dst, size_t dst_capacity,
+	const void *src, size_t src_size, const zstd_parameters *parameters);
+
+/* ======   Single-pass Decompression   ====== */
+
+typedef ZSTD_DCtx zstd_dctx;
 
 /**
- * ZSTD_DCtxWorkspaceBound() - amount of memory needed to initialize a ZSTD_DCtx
+ * zstd_dctx_workspace_bound() - max memory needed to initialize a zstd_dctx
  *
  * Return: A lower bound on the size of the workspace that is passed to
- *         ZSTD_initDCtx().
+ *         zstd_init_dctx().
  */
-size_t ZSTD_DCtxWorkspaceBound(void);
+size_t zstd_dctx_workspace_bound(void);
 
 /**
- * struct ZSTD_DCtx - the zstd decompression context
+ * zstd_init_dctx() - initialize a zstd decompression context
+ * @workspace:      The workspace to emplace the context into. It must outlive
+ *                  the returned context.
+ * @workspace_size: The size of workspace. Use zstd_dctx_workspace_bound() to
+ *                  determine how large the workspace must be.
  *
- * When decompressing many times it is recommended to allocate a context just
- * once and reuse it for each successive decompression operation.
+ * Return:          A zstd decompression context or NULL on error.
  */
-typedef struct ZSTD_DCtx_s ZSTD_DCtx;
-/**
- * ZSTD_initDCtx() - initialize a zstd decompression context
- * @workspace:     The workspace to emplace the context into. It must outlive
- *                 the returned context.
- * @workspaceSize: The size of workspace. Use ZSTD_DCtxWorkspaceBound() to
- *                 determine how large the workspace must be.
- *
- * Return:         A decompression context emplaced into workspace.
- */
-ZSTD_DCtx *ZSTD_initDCtx(void *workspace, size_t workspaceSize);
+zstd_dctx *zstd_init_dctx(void *workspace, size_t workspace_size);
 
 /**
- * ZSTD_decompressDCtx() - decompress zstd compressed src into dst
- * @ctx:         The decompression context.
- * @dst:         The buffer to decompress src into.
- * @dstCapacity: The size of the destination buffer. Must be at least as large
- *               as the decompressed size. If the caller cannot upper bound the
- *               decompressed size, then it's better to use the streaming API.
- * @src:         The zstd compressed data to decompress. Multiple concatenated
- *               frames and skippable frames are allowed.
- * @srcSize:     The exact size of the data to decompress.
+ * zstd_decompress_dctx() - decompress zstd compressed src into dst
+ * @dctx:         The decompression context.
+ * @dst:          The buffer to decompress src into.
+ * @dst_capacity: The size of the destination buffer. Must be at least as large
+ *                as the decompressed size. If the caller cannot upper bound the
+ *                decompressed size, then it's better to use the streaming API.
+ * @src:          The zstd compressed data to decompress. Multiple concatenated
+ *                frames and skippable frames are allowed.
+ * @src_size:     The exact size of the data to decompress.
  *
- * Return:       The decompressed size or an error, which can be checked using
- *               ZSTD_isError().
+ * Return:        The decompressed size or an error, which can be checked using
+ *                zstd_is_error().
  */
-size_t ZSTD_decompressDCtx(ZSTD_DCtx *ctx, void *dst, size_t dstCapacity,
-	const void *src, size_t srcSize);
+size_t zstd_decompress_dctx(zstd_dctx *dctx, void *dst, size_t dst_capacity,
+	const void *src, size_t src_size);
 
-/*-************************
- * Simple dictionary API
- **************************/
+/* ======   Streaming Buffers   ====== */
 
 /**
- * ZSTD_compress_usingDict() - compress src into dst using a dictionary
- * @ctx:         The context. Must have been initialized with a workspace at
- *               least as large as ZSTD_CCtxWorkspaceBound(params.cParams).
- * @dst:         The buffer to compress src into.
- * @dstCapacity: The size of the destination buffer. May be any size, but
- *               ZSTD_compressBound(srcSize) is guaranteed to be large enough.
- * @src:         The data to compress.
- * @srcSize:     The size of the data to compress.
- * @dict:        The dictionary to use for compression.
- * @dictSize:    The size of the dictionary.
- * @params:      The parameters to use for compression. See ZSTD_getParams().
- *
- * Compression using a predefined dictionary. The same dictionary must be used
- * during decompression.
- *
- * Return:       The compressed size or an error, which can be checked using
- *               ZSTD_isError().
- */
-size_t ZSTD_compress_usingDict(ZSTD_CCtx *ctx, void *dst, size_t dstCapacity,
-	const void *src, size_t srcSize, const void *dict, size_t dictSize,
-	ZSTD_parameters params);
-
-/**
- * ZSTD_decompress_usingDict() - decompress src into dst using a dictionary
- * @ctx:         The decompression context.
- * @dst:         The buffer to decompress src into.
- * @dstCapacity: The size of the destination buffer. Must be at least as large
- *               as the decompressed size. If the caller cannot upper bound the
- *               decompressed size, then it's better to use the streaming API.
- * @src:         The zstd compressed data to decompress. Multiple concatenated
- *               frames and skippable frames are allowed.
- * @srcSize:     The exact size of the data to decompress.
- * @dict:        The dictionary to use for decompression. The same dictionary
- *               must've been used to compress the data.
- * @dictSize:    The size of the dictionary.
- *
- * Return:       The decompressed size or an error, which can be checked using
- *               ZSTD_isError().
- */
-size_t ZSTD_decompress_usingDict(ZSTD_DCtx *ctx, void *dst, size_t dstCapacity,
-	const void *src, size_t srcSize, const void *dict, size_t dictSize);
-
-/*-**************************
- * Fast dictionary API
- ***************************/
-
-/**
- * ZSTD_CDictWorkspaceBound() - memory needed to initialize a ZSTD_CDict
- * @cParams: The compression parameters to be used for compression.
- *
- * Return:   A lower bound on the size of the workspace that is passed to
- *           ZSTD_initCDict().
- */
-size_t ZSTD_CDictWorkspaceBound(ZSTD_compressionParameters cParams);
-
-/**
- * struct ZSTD_CDict - a digested dictionary to be used for compression
- */
-typedef struct ZSTD_CDict_s ZSTD_CDict;
-
-/**
- * ZSTD_initCDict() - initialize a digested dictionary for compression
- * @dictBuffer:    The dictionary to digest. The buffer is referenced by the
- *                 ZSTD_CDict so it must outlive the returned ZSTD_CDict.
- * @dictSize:      The size of the dictionary.
- * @params:        The parameters to use for compression. See ZSTD_getParams().
- * @workspace:     The workspace. It must outlive the returned ZSTD_CDict.
- * @workspaceSize: The workspace size. Must be at least
- *                 ZSTD_CDictWorkspaceBound(params.cParams).
- *
- * When compressing multiple messages / blocks with the same dictionary it is
- * recommended to load it just once. The ZSTD_CDict merely references the
- * dictBuffer, so it must outlive the returned ZSTD_CDict.
- *
- * Return:         The digested dictionary emplaced into workspace.
- */
-ZSTD_CDict *ZSTD_initCDict(const void *dictBuffer, size_t dictSize,
-	ZSTD_parameters params, void *workspace, size_t workspaceSize);
-
-/**
- * ZSTD_compress_usingCDict() - compress src into dst using a ZSTD_CDict
- * @ctx:         The context. Must have been initialized with a workspace at
- *               least as large as ZSTD_CCtxWorkspaceBound(cParams) where
- *               cParams are the compression parameters used to initialize the
- *               cdict.
- * @dst:         The buffer to compress src into.
- * @dstCapacity: The size of the destination buffer. May be any size, but
- *               ZSTD_compressBound(srcSize) is guaranteed to be large enough.
- * @src:         The data to compress.
- * @srcSize:     The size of the data to compress.
- * @cdict:       The digested dictionary to use for compression.
- * @params:      The parameters to use for compression. See ZSTD_getParams().
- *
- * Compression using a digested dictionary. The same dictionary must be used
- * during decompression.
- *
- * Return:       The compressed size or an error, which can be checked using
- *               ZSTD_isError().
- */
-size_t ZSTD_compress_usingCDict(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity,
-	const void *src, size_t srcSize, const ZSTD_CDict *cdict);
-
-
-/**
- * ZSTD_DDictWorkspaceBound() - memory needed to initialize a ZSTD_DDict
- *
- * Return:  A lower bound on the size of the workspace that is passed to
- *          ZSTD_initDDict().
- */
-size_t ZSTD_DDictWorkspaceBound(void);
-
-/**
- * struct ZSTD_DDict - a digested dictionary to be used for decompression
- */
-typedef struct ZSTD_DDict_s ZSTD_DDict;
-
-/**
- * ZSTD_initDDict() - initialize a digested dictionary for decompression
- * @dictBuffer:    The dictionary to digest. The buffer is referenced by the
- *                 ZSTD_DDict so it must outlive the returned ZSTD_DDict.
- * @dictSize:      The size of the dictionary.
- * @workspace:     The workspace. It must outlive the returned ZSTD_DDict.
- * @workspaceSize: The workspace size. Must be at least
- *                 ZSTD_DDictWorkspaceBound().
- *
- * When decompressing multiple messages / blocks with the same dictionary it is
- * recommended to load it just once. The ZSTD_DDict merely references the
- * dictBuffer, so it must outlive the returned ZSTD_DDict.
- *
- * Return:         The digested dictionary emplaced into workspace.
- */
-ZSTD_DDict *ZSTD_initDDict(const void *dictBuffer, size_t dictSize,
-	void *workspace, size_t workspaceSize);
-
-/**
- * ZSTD_decompress_usingDDict() - decompress src into dst using a ZSTD_DDict
- * @ctx:         The decompression context.
- * @dst:         The buffer to decompress src into.
- * @dstCapacity: The size of the destination buffer. Must be at least as large
- *               as the decompressed size. If the caller cannot upper bound the
- *               decompressed size, then it's better to use the streaming API.
- * @src:         The zstd compressed data to decompress. Multiple concatenated
- *               frames and skippable frames are allowed.
- * @srcSize:     The exact size of the data to decompress.
- * @ddict:       The digested dictionary to use for decompression. The same
- *               dictionary must've been used to compress the data.
- *
- * Return:       The decompressed size or an error, which can be checked using
- *               ZSTD_isError().
- */
-size_t ZSTD_decompress_usingDDict(ZSTD_DCtx *dctx, void *dst,
-	size_t dstCapacity, const void *src, size_t srcSize,
-	const ZSTD_DDict *ddict);
-
-
-/*-**************************
- * Streaming
- ***************************/
-
-/**
- * struct ZSTD_inBuffer - input buffer for streaming
+ * struct zstd_in_buffer - input buffer for streaming
  * @src:  Start of the input buffer.
  * @size: Size of the input buffer.
  * @pos:  Position where reading stopped. Will be updated.
  *        Necessarily 0 <= pos <= size.
+ *
+ * See zstd_lib.h.
  */
-typedef struct ZSTD_inBuffer_s {
-	const void *src;
-	size_t size;
-	size_t pos;
-} ZSTD_inBuffer;
+typedef ZSTD_inBuffer zstd_in_buffer;
 
 /**
- * struct ZSTD_outBuffer - output buffer for streaming
+ * struct zstd_out_buffer - output buffer for streaming
  * @dst:  Start of the output buffer.
  * @size: Size of the output buffer.
  * @pos:  Position where writing stopped. Will be updated.
  *        Necessarily 0 <= pos <= size.
+ *
+ * See zstd_lib.h.
  */
-typedef struct ZSTD_outBuffer_s {
-	void *dst;
-	size_t size;
-	size_t pos;
-} ZSTD_outBuffer;
+typedef ZSTD_outBuffer zstd_out_buffer;
 
+/* ======   Streaming Compression   ====== */
 
-
-/*-*****************************************************************************
- * Streaming compression - HowTo
- *
- * A ZSTD_CStream object is required to track streaming operation.
- * Use ZSTD_initCStream() to initialize a ZSTD_CStream object.
- * ZSTD_CStream objects can be reused multiple times on consecutive compression
- * operations. It is recommended to re-use ZSTD_CStream in situations where many
- * streaming operations will be achieved consecutively. Use one separate
- * ZSTD_CStream per thread for parallel execution.
- *
- * Use ZSTD_compressStream() repetitively to consume input stream.
- * The function will automatically update both `pos` fields.
- * Note that it may not consume the entire input, in which case `pos < size`,
- * and it's up to the caller to present again remaining data.
- * It returns a hint for the preferred number of bytes to use as an input for
- * the next function call.
- *
- * At any moment, it's possible to flush whatever data remains within internal
- * buffer, using ZSTD_flushStream(). `output->pos` will be updated. There might
- * still be some content left within the internal buffer if `output->size` is
- * too small. It returns the number of bytes left in the internal buffer and
- * must be called until it returns 0.
- *
- * ZSTD_endStream() instructs to finish a frame. It will perform a flush and
- * write frame epilogue. The epilogue is required for decoders to consider a
- * frame completed. Similar to ZSTD_flushStream(), it may not be able to flush
- * the full content if `output->size` is too small. In which case, call again
- * ZSTD_endStream() to complete the flush. It returns the number of bytes left
- * in the internal buffer and must be called until it returns 0.
- ******************************************************************************/
+typedef ZSTD_CStream zstd_cstream;
 
 /**
- * ZSTD_CStreamWorkspaceBound() - memory needed to initialize a ZSTD_CStream
- * @cParams: The compression parameters to be used for compression.
+ * zstd_cstream_workspace_bound() - memory needed to initialize a zstd_cstream
+ * @cparams: The compression parameters to be used for compression.
  *
  * Return:   A lower bound on the size of the workspace that is passed to
- *           ZSTD_initCStream() and ZSTD_initCStream_usingCDict().
+ *           zstd_init_cstream().
  */
-size_t ZSTD_CStreamWorkspaceBound(ZSTD_compressionParameters cParams);
+size_t zstd_cstream_workspace_bound(const zstd_compression_parameters *cparams);
 
 /**
- * struct ZSTD_CStream - the zstd streaming compression context
- */
-typedef struct ZSTD_CStream_s ZSTD_CStream;
-
-/*===== ZSTD_CStream management functions =====*/
-/**
- * ZSTD_initCStream() - initialize a zstd streaming compression context
- * @params:         The zstd compression parameters.
- * @pledgedSrcSize: If params.fParams.contentSizeFlag == 1 then the caller must
- *                  pass the source size (zero means empty source). Otherwise,
- *                  the caller may optionally pass the source size, or zero if
- *                  unknown.
- * @workspace:      The workspace to emplace the context into. It must outlive
- *                  the returned context.
- * @workspaceSize:  The size of workspace.
- *                  Use ZSTD_CStreamWorkspaceBound(params.cParams) to determine
- *                  how large the workspace must be.
+ * zstd_init_cstream() - initialize a zstd streaming compression context
+ * @parameters        The zstd parameters to use for compression.
+ * @pledged_src_size: If params.fParams.contentSizeFlag == 1 then the caller
+ *                    must pass the source size (zero means empty source).
+ *                    Otherwise, the caller may optionally pass the source
+ *                    size, or zero if unknown.
+ * @workspace:        The workspace to emplace the context into. It must outlive
+ *                    the returned context.
+ * @workspace_size:   The size of workspace.
+ *                    Use zstd_cstream_workspace_bound(params->cparams) to
+ *                    determine how large the workspace must be.
  *
- * Return:          The zstd streaming compression context.
+ * Return:            The zstd streaming compression context or NULL on error.
  */
-ZSTD_CStream *ZSTD_initCStream(ZSTD_parameters params,
-	unsigned long long pledgedSrcSize, void *workspace,
-	size_t workspaceSize);
+zstd_cstream *zstd_init_cstream(const zstd_parameters *parameters,
+	unsigned long long pledged_src_size, void *workspace, size_t workspace_size);
 
 /**
- * ZSTD_initCStream_usingCDict() - initialize a streaming compression context
- * @cdict:          The digested dictionary to use for compression.
- * @pledgedSrcSize: Optionally the source size, or zero if unknown.
- * @workspace:      The workspace to emplace the context into. It must outlive
- *                  the returned context.
- * @workspaceSize:  The size of workspace. Call ZSTD_CStreamWorkspaceBound()
- *                  with the cParams used to initialize the cdict to determine
- *                  how large the workspace must be.
- *
- * Return:          The zstd streaming compression context.
- */
-ZSTD_CStream *ZSTD_initCStream_usingCDict(const ZSTD_CDict *cdict,
-	unsigned long long pledgedSrcSize, void *workspace,
-	size_t workspaceSize);
-
-/*===== Streaming compression functions =====*/
-/**
- * ZSTD_resetCStream() - reset the context using parameters from creation
- * @zcs:            The zstd streaming compression context to reset.
- * @pledgedSrcSize: Optionally the source size, or zero if unknown.
+ * zstd_reset_cstream() - reset the context using parameters from creation
+ * @cstream:          The zstd streaming compression context to reset.
+ * @pledged_src_size: Optionally the source size, or zero if unknown.
  *
  * Resets the context using the parameters from creation. Skips dictionary
- * loading, since it can be reused. If `pledgedSrcSize` is non-zero the frame
+ * loading, since it can be reused. If `pledged_src_size` is non-zero the frame
  * content size is always written into the frame header.
  *
- * Return:          Zero or an error, which can be checked using ZSTD_isError().
+ * Return:            Zero or an error, which can be checked using
+ *                    zstd_is_error().
  */
-size_t ZSTD_resetCStream(ZSTD_CStream *zcs, unsigned long long pledgedSrcSize);
+size_t zstd_reset_cstream(zstd_cstream *cstream,
+	unsigned long long pledged_src_size);
+
 /**
- * ZSTD_compressStream() - streaming compress some of input into output
- * @zcs:    The zstd streaming compression context.
- * @output: Destination buffer. `output->pos` is updated to indicate how much
- *          compressed data was written.
- * @input:  Source buffer. `input->pos` is updated to indicate how much data was
- *          read. Note that it may not consume the entire input, in which case
- *          `input->pos < input->size`, and it's up to the caller to present
- *          remaining data again.
+ * zstd_compress_stream() - streaming compress some of input into output
+ * @cstream: The zstd streaming compression context.
+ * @output:  Destination buffer. `output->pos` is updated to indicate how much
+ *           compressed data was written.
+ * @input:   Source buffer. `input->pos` is updated to indicate how much data
+ *           was read. Note that it may not consume the entire input, in which
+ *           case `input->pos < input->size`, and it's up to the caller to
+ *           present remaining data again.
  *
  * The `input` and `output` buffers may be any size. Guaranteed to make some
  * forward progress if `input` and `output` are not empty.
  *
- * Return:  A hint for the number of bytes to use as the input for the next
- *          function call or an error, which can be checked using
- *          ZSTD_isError().
+ * Return:   A hint for the number of bytes to use as the input for the next
+ *           function call or an error, which can be checked using
+ *           zstd_is_error().
  */
-size_t ZSTD_compressStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output,
-	ZSTD_inBuffer *input);
+size_t zstd_compress_stream(zstd_cstream *cstream, zstd_out_buffer *output,
+	zstd_in_buffer *input);
+
 /**
- * ZSTD_flushStream() - flush internal buffers into output
- * @zcs:    The zstd streaming compression context.
- * @output: Destination buffer. `output->pos` is updated to indicate how much
- *          compressed data was written.
+ * zstd_flush_stream() - flush internal buffers into output
+ * @cstream: The zstd streaming compression context.
+ * @output:  Destination buffer. `output->pos` is updated to indicate how much
+ *           compressed data was written.
  *
- * ZSTD_flushStream() must be called until it returns 0, meaning all the data
- * has been flushed. Since ZSTD_flushStream() causes a block to be ended,
+ * zstd_flush_stream() must be called until it returns 0, meaning all the data
+ * has been flushed. Since zstd_flush_stream() causes a block to be ended,
  * calling it too often will degrade the compression ratio.
  *
- * Return:  The number of bytes still present within internal buffers or an
- *          error, which can be checked using ZSTD_isError().
+ * Return:   The number of bytes still present within internal buffers or an
+ *           error, which can be checked using zstd_is_error().
  */
-size_t ZSTD_flushStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output);
+size_t zstd_flush_stream(zstd_cstream *cstream, zstd_out_buffer *output);
+
 /**
- * ZSTD_endStream() - flush internal buffers into output and end the frame
- * @zcs:    The zstd streaming compression context.
- * @output: Destination buffer. `output->pos` is updated to indicate how much
- *          compressed data was written.
+ * zstd_end_stream() - flush internal buffers into output and end the frame
+ * @cstream: The zstd streaming compression context.
+ * @output:  Destination buffer. `output->pos` is updated to indicate how much
+ *           compressed data was written.
  *
- * ZSTD_endStream() must be called until it returns 0, meaning all the data has
+ * zstd_end_stream() must be called until it returns 0, meaning all the data has
  * been flushed and the frame epilogue has been written.
  *
- * Return:  The number of bytes still present within internal buffers or an
- *          error, which can be checked using ZSTD_isError().
+ * Return:   The number of bytes still present within internal buffers or an
+ *           error, which can be checked using zstd_is_error().
  */
-size_t ZSTD_endStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output);
+size_t zstd_end_stream(zstd_cstream *cstream, zstd_out_buffer *output);
+
+/* ======   Streaming Decompression   ====== */
+
+typedef ZSTD_DStream zstd_dstream;
 
 /**
- * ZSTD_CStreamInSize() - recommended size for the input buffer
+ * zstd_dstream_workspace_bound() - memory needed to initialize a zstd_dstream
+ * @max_window_size: The maximum window size allowed for compressed frames.
  *
- * Return: The recommended size for the input buffer.
+ * Return:           A lower bound on the size of the workspace that is passed
+ *                   to zstd_init_dstream().
  */
-size_t ZSTD_CStreamInSize(void);
-/**
- * ZSTD_CStreamOutSize() - recommended size for the output buffer
- *
- * When the output buffer is at least this large, it is guaranteed to be large
- * enough to flush at least one complete compressed block.
- *
- * Return: The recommended size for the output buffer.
- */
-size_t ZSTD_CStreamOutSize(void);
-
-
-
-/*-*****************************************************************************
- * Streaming decompression - HowTo
- *
- * A ZSTD_DStream object is required to track streaming operations.
- * Use ZSTD_initDStream() to initialize a ZSTD_DStream object.
- * ZSTD_DStream objects can be re-used multiple times.
- *
- * Use ZSTD_decompressStream() repetitively to consume your input.
- * The function will update both `pos` fields.
- * If `input->pos < input->size`, some input has not been consumed.
- * It's up to the caller to present again remaining data.
- * If `output->pos < output->size`, decoder has flushed everything it could.
- * Returns 0 iff a frame is completely decoded and fully flushed.
- * Otherwise it returns a suggested next input size that will never load more
- * than the current frame.
- ******************************************************************************/
+size_t zstd_dstream_workspace_bound(size_t max_window_size);
 
 /**
- * ZSTD_DStreamWorkspaceBound() - memory needed to initialize a ZSTD_DStream
- * @maxWindowSize: The maximum window size allowed for compressed frames.
+ * zstd_init_dstream() - initialize a zstd streaming decompression context
+ * @max_window_size: The maximum window size allowed for compressed frames.
+ * @workspace:       The workspace to emplace the context into. It must outlive
+ *                   the returned context.
+ * @workspaceSize:   The size of workspace.
+ *                   Use zstd_dstream_workspace_bound(max_window_size) to
+ *                   determine how large the workspace must be.
  *
- * Return:         A lower bound on the size of the workspace that is passed to
- *                 ZSTD_initDStream() and ZSTD_initDStream_usingDDict().
+ * Return:           The zstd streaming decompression context.
  */
-size_t ZSTD_DStreamWorkspaceBound(size_t maxWindowSize);
+zstd_dstream *zstd_init_dstream(size_t max_window_size, void *workspace,
+	size_t workspace_size);
 
 /**
- * struct ZSTD_DStream - the zstd streaming decompression context
- */
-typedef struct ZSTD_DStream_s ZSTD_DStream;
-/*===== ZSTD_DStream management functions =====*/
-/**
- * ZSTD_initDStream() - initialize a zstd streaming decompression context
- * @maxWindowSize: The maximum window size allowed for compressed frames.
- * @workspace:     The workspace to emplace the context into. It must outlive
- *                 the returned context.
- * @workspaceSize: The size of workspace.
- *                 Use ZSTD_DStreamWorkspaceBound(maxWindowSize) to determine
- *                 how large the workspace must be.
- *
- * Return:         The zstd streaming decompression context.
- */
-ZSTD_DStream *ZSTD_initDStream(size_t maxWindowSize, void *workspace,
-	size_t workspaceSize);
-/**
- * ZSTD_initDStream_usingDDict() - initialize streaming decompression context
- * @maxWindowSize: The maximum window size allowed for compressed frames.
- * @ddict:         The digested dictionary to use for decompression.
- * @workspace:     The workspace to emplace the context into. It must outlive
- *                 the returned context.
- * @workspaceSize: The size of workspace.
- *                 Use ZSTD_DStreamWorkspaceBound(maxWindowSize) to determine
- *                 how large the workspace must be.
- *
- * Return:         The zstd streaming decompression context.
- */
-ZSTD_DStream *ZSTD_initDStream_usingDDict(size_t maxWindowSize,
-	const ZSTD_DDict *ddict, void *workspace, size_t workspaceSize);
-
-/*===== Streaming decompression functions =====*/
-/**
- * ZSTD_resetDStream() - reset the context using parameters from creation
- * @zds:   The zstd streaming decompression context to reset.
+ * zstd_reset_dstream() - reset the context using parameters from creation
+ * @dstream: The zstd streaming decompression context to reset.
  *
  * Resets the context using the parameters from creation. Skips dictionary
  * loading, since it can be reused.
  *
- * Return: Zero or an error, which can be checked using ZSTD_isError().
+ * Return:   Zero or an error, which can be checked using zstd_is_error().
  */
-size_t ZSTD_resetDStream(ZSTD_DStream *zds);
+size_t zstd_reset_dstream(zstd_dstream *dstream);
+
 /**
- * ZSTD_decompressStream() - streaming decompress some of input into output
- * @zds:    The zstd streaming decompression context.
- * @output: Destination buffer. `output.pos` is updated to indicate how much
- *          decompressed data was written.
- * @input:  Source buffer. `input.pos` is updated to indicate how much data was
- *          read. Note that it may not consume the entire input, in which case
- *          `input.pos < input.size`, and it's up to the caller to present
- *          remaining data again.
+ * zstd_decompress_stream() - streaming decompress some of input into output
+ * @dstream: The zstd streaming decompression context.
+ * @output:  Destination buffer. `output.pos` is updated to indicate how much
+ *           decompressed data was written.
+ * @input:   Source buffer. `input.pos` is updated to indicate how much data was
+ *           read. Note that it may not consume the entire input, in which case
+ *           `input.pos < input.size`, and it's up to the caller to present
+ *           remaining data again.
  *
  * The `input` and `output` buffers may be any size. Guaranteed to make some
  * forward progress if `input` and `output` are not empty.
- * ZSTD_decompressStream() will not consume the last byte of the frame until
+ * zstd_decompress_stream() will not consume the last byte of the frame until
  * the entire frame is flushed.
  *
- * Return:  Returns 0 iff a frame is completely decoded and fully flushed.
- *          Otherwise returns a hint for the number of bytes to use as the input
- *          for the next function call or an error, which can be checked using
- *          ZSTD_isError(). The size hint will never load more than the frame.
+ * Return:   Returns 0 iff a frame is completely decoded and fully flushed.
+ *           Otherwise returns a hint for the number of bytes to use as the
+ *           input for the next function call or an error, which can be checked
+ *           using zstd_is_error(). The size hint will never load more than the
+ *           frame.
  */
-size_t ZSTD_decompressStream(ZSTD_DStream *zds, ZSTD_outBuffer *output,
-	ZSTD_inBuffer *input);
+size_t zstd_decompress_stream(zstd_dstream *dstream, zstd_out_buffer *output,
+	zstd_in_buffer *input);
+
+/* ======   Frame Inspection Functions ====== */
 
 /**
- * ZSTD_DStreamInSize() - recommended size for the input buffer
+ * zstd_find_frame_compressed_size() - returns the size of a compressed frame
+ * @src:      Source buffer. It should point to the start of a zstd encoded
+ *            frame or a skippable frame.
+ * @src_size: The size of the source buffer. It must be at least as large as the
+ *            size of the frame.
  *
- * Return: The recommended size for the input buffer.
+ * Return:    The compressed size of the frame pointed to by `src` or an error,
+ *            which can be check with zstd_is_error().
+ *            Suitable to pass to ZSTD_decompress() or similar functions.
  */
-size_t ZSTD_DStreamInSize(void);
-/**
- * ZSTD_DStreamOutSize() - recommended size for the output buffer
- *
- * When the output buffer is at least this large, it is guaranteed to be large
- * enough to flush at least one complete decompressed block.
- *
- * Return: The recommended size for the output buffer.
- */
-size_t ZSTD_DStreamOutSize(void);
-
-
-/* --- Constants ---*/
-#define ZSTD_MAGICNUMBER            0xFD2FB528   /* >= v0.8.0 */
-#define ZSTD_MAGIC_SKIPPABLE_START  0x184D2A50U
-
-#define ZSTD_CONTENTSIZE_UNKNOWN (0ULL - 1)
-#define ZSTD_CONTENTSIZE_ERROR   (0ULL - 2)
-
-#define ZSTD_WINDOWLOG_MAX_32  27
-#define ZSTD_WINDOWLOG_MAX_64  27
-#define ZSTD_WINDOWLOG_MAX \
-	((unsigned int)(sizeof(size_t) == 4 \
-		? ZSTD_WINDOWLOG_MAX_32 \
-		: ZSTD_WINDOWLOG_MAX_64))
-#define ZSTD_WINDOWLOG_MIN 10
-#define ZSTD_HASHLOG_MAX ZSTD_WINDOWLOG_MAX
-#define ZSTD_HASHLOG_MIN        6
-#define ZSTD_CHAINLOG_MAX     (ZSTD_WINDOWLOG_MAX+1)
-#define ZSTD_CHAINLOG_MIN      ZSTD_HASHLOG_MIN
-#define ZSTD_HASHLOG3_MAX      17
-#define ZSTD_SEARCHLOG_MAX    (ZSTD_WINDOWLOG_MAX-1)
-#define ZSTD_SEARCHLOG_MIN      1
-/* only for ZSTD_fast, other strategies are limited to 6 */
-#define ZSTD_SEARCHLENGTH_MAX   7
-/* only for ZSTD_btopt, other strategies are limited to 4 */
-#define ZSTD_SEARCHLENGTH_MIN   3
-#define ZSTD_TARGETLENGTH_MIN   4
-#define ZSTD_TARGETLENGTH_MAX 999
-
-/* for static allocation */
-#define ZSTD_FRAMEHEADERSIZE_MAX 18
-#define ZSTD_FRAMEHEADERSIZE_MIN  6
-#define ZSTD_frameHeaderSize_prefix 5
-#define ZSTD_frameHeaderSize_min ZSTD_FRAMEHEADERSIZE_MIN
-#define ZSTD_frameHeaderSize_max ZSTD_FRAMEHEADERSIZE_MAX
-/* magic number + skippable frame length */
-#define ZSTD_skippableHeaderSize 8
-
-
-/*-*************************************
- * Compressed size functions
- **************************************/
+size_t zstd_find_frame_compressed_size(const void *src, size_t src_size);
 
 /**
- * ZSTD_findFrameCompressedSize() - returns the size of a compressed frame
- * @src:     Source buffer. It should point to the start of a zstd encoded frame
- *           or a skippable frame.
- * @srcSize: The size of the source buffer. It must be at least as large as the
- *           size of the frame.
- *
- * Return:   The compressed size of the frame pointed to by `src` or an error,
- *           which can be check with ZSTD_isError().
- *           Suitable to pass to ZSTD_decompress() or similar functions.
- */
-size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize);
-
-/*-*************************************
- * Decompressed size functions
- **************************************/
-/**
- * ZSTD_getFrameContentSize() - returns the content size in a zstd frame header
- * @src:     It should point to the start of a zstd encoded frame.
- * @srcSize: The size of the source buffer. It must be at least as large as the
- *           frame header. `ZSTD_frameHeaderSize_max` is always large enough.
- *
- * Return:   The frame content size stored in the frame header if known.
- *           `ZSTD_CONTENTSIZE_UNKNOWN` if the content size isn't stored in the
- *           frame header. `ZSTD_CONTENTSIZE_ERROR` on invalid input.
- */
-unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize);
-
-/**
- * ZSTD_findDecompressedSize() - returns decompressed size of a series of frames
- * @src:     It should point to the start of a series of zstd encoded and/or
- *           skippable frames.
- * @srcSize: The exact size of the series of frames.
- *
- * If any zstd encoded frame in the series doesn't have the frame content size
- * set, `ZSTD_CONTENTSIZE_UNKNOWN` is returned. But frame content size is always
- * set when using ZSTD_compress(). The decompressed size can be very large.
- * If the source is untrusted, the decompressed size could be wrong or
- * intentionally modified. Always ensure the result fits within the
- * application's authorized limits. ZSTD_findDecompressedSize() handles multiple
- * frames, and so it must traverse the input to read each frame header. This is
- * efficient as most of the data is skipped, however it does mean that all frame
- * data must be present and valid.
- *
- * Return:   Decompressed size of all the data contained in the frames if known.
- *           `ZSTD_CONTENTSIZE_UNKNOWN` if the decompressed size is unknown.
- *           `ZSTD_CONTENTSIZE_ERROR` if an error occurred.
- */
-unsigned long long ZSTD_findDecompressedSize(const void *src, size_t srcSize);
-
-/*-*************************************
- * Advanced compression functions
- **************************************/
-/**
- * ZSTD_checkCParams() - ensure parameter values remain within authorized range
- * @cParams: The zstd compression parameters.
- *
- * Return:   Zero or an error, which can be checked using ZSTD_isError().
- */
-size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams);
-
-/**
- * ZSTD_adjustCParams() - optimize parameters for a given srcSize and dictSize
- * @srcSize:  Optionally the estimated source size, or zero if unknown.
- * @dictSize: Optionally the estimated dictionary size, or zero if unknown.
- *
- * Return:    The optimized parameters.
- */
-ZSTD_compressionParameters ZSTD_adjustCParams(
-	ZSTD_compressionParameters cParams, unsigned long long srcSize,
-	size_t dictSize);
-
-/*--- Advanced decompression functions ---*/
-
-/**
- * ZSTD_isFrame() - returns true iff the buffer starts with a valid frame
- * @buffer: The source buffer to check.
- * @size:   The size of the source buffer, must be at least 4 bytes.
- *
- * Return: True iff the buffer starts with a zstd or skippable frame identifier.
- */
-unsigned int ZSTD_isFrame(const void *buffer, size_t size);
-
-/**
- * ZSTD_getDictID_fromDict() - returns the dictionary id stored in a dictionary
- * @dict:     The dictionary buffer.
- * @dictSize: The size of the dictionary buffer.
- *
- * Return:    The dictionary id stored within the dictionary or 0 if the
- *            dictionary is not a zstd dictionary. If it returns 0 the
- *            dictionary can still be loaded as a content-only dictionary.
- */
-unsigned int ZSTD_getDictID_fromDict(const void *dict, size_t dictSize);
-
-/**
- * ZSTD_getDictID_fromDDict() - returns the dictionary id stored in a ZSTD_DDict
- * @ddict: The ddict to find the id of.
- *
- * Return: The dictionary id stored within `ddict` or 0 if the dictionary is not
- *         a zstd dictionary. If it returns 0 `ddict` will be loaded as a
- *         content-only dictionary.
- */
-unsigned int ZSTD_getDictID_fromDDict(const ZSTD_DDict *ddict);
-
-/**
- * ZSTD_getDictID_fromFrame() - returns the dictionary id stored in a zstd frame
- * @src:     Source buffer. It must be a zstd encoded frame.
- * @srcSize: The size of the source buffer. It must be at least as large as the
- *           frame header. `ZSTD_frameHeaderSize_max` is always large enough.
- *
- * Return:   The dictionary id required to decompress the frame stored within
- *           `src` or 0 if the dictionary id could not be decoded. It can return
- *           0 if the frame does not require a dictionary, the dictionary id
- *           wasn't stored in the frame, `src` is not a zstd frame, or `srcSize`
- *           is too small.
- */
-unsigned int ZSTD_getDictID_fromFrame(const void *src, size_t srcSize);
-
-/**
- * struct ZSTD_frameParams - zstd frame parameters stored in the frame header
- * @frameContentSize: The frame content size, or 0 if not present.
+ * struct zstd_frame_params - zstd frame parameters stored in the frame header
+ * @frameContentSize: The frame content size, or ZSTD_CONTENTSIZE_UNKNOWN if not
+ *                    present.
  * @windowSize:       The window size, or 0 if the frame is a skippable frame.
+ * @blockSizeMax:     The maximum block size.
+ * @frameType:        The frame type (zstd or skippable)
+ * @headerSize:       The size of the frame header.
  * @dictID:           The dictionary id, or 0 if not present.
  * @checksumFlag:     Whether a checksum was used.
+ *
+ * See zstd_lib.h.
  */
-typedef struct {
-	unsigned long long frameContentSize;
-	unsigned int windowSize;
-	unsigned int dictID;
-	unsigned int checksumFlag;
-} ZSTD_frameParams;
+typedef ZSTD_frameHeader zstd_frame_header;
 
 /**
- * ZSTD_getFrameParams() - extracts parameters from a zstd or skippable frame
- * @fparamsPtr: On success the frame parameters are written here.
- * @src:        The source buffer. It must point to a zstd or skippable frame.
- * @srcSize:    The size of the source buffer. `ZSTD_frameHeaderSize_max` is
- *              always large enough to succeed.
+ * zstd_get_frame_header() - extracts parameters from a zstd or skippable frame
+ * @params:   On success the frame parameters are written here.
+ * @src:      The source buffer. It must point to a zstd or skippable frame.
+ * @src_size: The size of the source buffer.
  *
- * Return:      0 on success. If more data is required it returns how many bytes
- *              must be provided to make forward progress. Otherwise it returns
- *              an error, which can be checked using ZSTD_isError().
+ * Return:    0 on success. If more data is required it returns how many bytes
+ *            must be provided to make forward progress. Otherwise it returns
+ *            an error, which can be checked using zstd_is_error().
  */
-size_t ZSTD_getFrameParams(ZSTD_frameParams *fparamsPtr, const void *src,
-	size_t srcSize);
+size_t zstd_get_frame_header(zstd_frame_header *params, const void *src,
+	size_t src_size);
 
-/*-*****************************************************************************
- * Buffer-less and synchronous inner streaming functions
- *
- * This is an advanced API, giving full control over buffer management, for
- * users which need direct control over memory.
- * But it's also a complex one, with many restrictions (documented below).
- * Prefer using normal streaming API for an easier experience
- ******************************************************************************/
-
-/*-*****************************************************************************
- * Buffer-less streaming compression (synchronous mode)
- *
- * A ZSTD_CCtx object is required to track streaming operations.
- * Use ZSTD_initCCtx() to initialize a context.
- * ZSTD_CCtx object can be re-used multiple times within successive compression
- * operations.
- *
- * Start by initializing a context.
- * Use ZSTD_compressBegin(), or ZSTD_compressBegin_usingDict() for dictionary
- * compression,
- * or ZSTD_compressBegin_advanced(), for finer parameter control.
- * It's also possible to duplicate a reference context which has already been
- * initialized, using ZSTD_copyCCtx()
- *
- * Then, consume your input using ZSTD_compressContinue().
- * There are some important considerations to keep in mind when using this
- * advanced function :
- * - ZSTD_compressContinue() has no internal buffer. It uses externally provided
- *   buffer only.
- * - Interface is synchronous : input is consumed entirely and produce 1+
- *   (or more) compressed blocks.
- * - Caller must ensure there is enough space in `dst` to store compressed data
- *   under worst case scenario. Worst case evaluation is provided by
- *   ZSTD_compressBound().
- *   ZSTD_compressContinue() doesn't guarantee recover after a failed
- *   compression.
- * - ZSTD_compressContinue() presumes prior input ***is still accessible and
- *   unmodified*** (up to maximum distance size, see WindowLog).
- *   It remembers all previous contiguous blocks, plus one separated memory
- *   segment (which can itself consists of multiple contiguous blocks)
- * - ZSTD_compressContinue() detects that prior input has been overwritten when
- *   `src` buffer overlaps. In which case, it will "discard" the relevant memory
- *   section from its history.
- *
- * Finish a frame with ZSTD_compressEnd(), which will write the last block(s)
- * and optional checksum. It's possible to use srcSize==0, in which case, it
- * will write a final empty block to end the frame. Without last block mark,
- * frames will be considered unfinished (corrupted) by decoders.
- *
- * `ZSTD_CCtx` object can be re-used (ZSTD_compressBegin()) to compress some new
- * frame.
- ******************************************************************************/
-
-/*=====   Buffer-less streaming compression functions  =====*/
-size_t ZSTD_compressBegin(ZSTD_CCtx *cctx, int compressionLevel);
-size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx *cctx, const void *dict,
-	size_t dictSize, int compressionLevel);
-size_t ZSTD_compressBegin_advanced(ZSTD_CCtx *cctx, const void *dict,
-	size_t dictSize, ZSTD_parameters params,
-	unsigned long long pledgedSrcSize);
-size_t ZSTD_copyCCtx(ZSTD_CCtx *cctx, const ZSTD_CCtx *preparedCCtx,
-	unsigned long long pledgedSrcSize);
-size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx *cctx, const ZSTD_CDict *cdict,
-	unsigned long long pledgedSrcSize);
-size_t ZSTD_compressContinue(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity,
-	const void *src, size_t srcSize);
-size_t ZSTD_compressEnd(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity,
-	const void *src, size_t srcSize);
-
-
-
-/*-*****************************************************************************
- * Buffer-less streaming decompression (synchronous mode)
- *
- * A ZSTD_DCtx object is required to track streaming operations.
- * Use ZSTD_initDCtx() to initialize a context.
- * A ZSTD_DCtx object can be re-used multiple times.
- *
- * First typical operation is to retrieve frame parameters, using
- * ZSTD_getFrameParams(). It fills a ZSTD_frameParams structure which provide
- * important information to correctly decode the frame, such as the minimum
- * rolling buffer size to allocate to decompress data (`windowSize`), and the
- * dictionary ID used.
- * Note: content size is optional, it may not be present. 0 means unknown.
- * Note that these values could be wrong, either because of data malformation,
- * or because an attacker is spoofing deliberate false information. As a
- * consequence, check that values remain within valid application range,
- * especially `windowSize`, before allocation. Each application can set its own
- * limit, depending on local restrictions. For extended interoperability, it is
- * recommended to support at least 8 MB.
- * Frame parameters are extracted from the beginning of the compressed frame.
- * Data fragment must be large enough to ensure successful decoding, typically
- * `ZSTD_frameHeaderSize_max` bytes.
- * Result: 0: successful decoding, the `ZSTD_frameParams` structure is filled.
- *        >0: `srcSize` is too small, provide at least this many bytes.
- *        errorCode, which can be tested using ZSTD_isError().
- *
- * Start decompression, with ZSTD_decompressBegin() or
- * ZSTD_decompressBegin_usingDict(). Alternatively, you can copy a prepared
- * context, using ZSTD_copyDCtx().
- *
- * Then use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue()
- * alternatively.
- * ZSTD_nextSrcSizeToDecompress() tells how many bytes to provide as 'srcSize'
- * to ZSTD_decompressContinue().
- * ZSTD_decompressContinue() requires this _exact_ amount of bytes, or it will
- * fail.
- *
- * The result of ZSTD_decompressContinue() is the number of bytes regenerated
- * within 'dst' (necessarily <= dstCapacity). It can be zero, which is not an
- * error; it just means ZSTD_decompressContinue() has decoded some metadata
- * item. It can also be an error code, which can be tested with ZSTD_isError().
- *
- * ZSTD_decompressContinue() needs previous data blocks during decompression, up
- * to `windowSize`. They should preferably be located contiguously, prior to
- * current block. Alternatively, a round buffer of sufficient size is also
- * possible. Sufficient size is determined by frame parameters.
- * ZSTD_decompressContinue() is very sensitive to contiguity, if 2 blocks don't
- * follow each other, make sure that either the compressor breaks contiguity at
- * the same place, or that previous contiguous segment is large enough to
- * properly handle maximum back-reference.
- *
- * A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero.
- * Context can then be reset to start a new decompression.
- *
- * Note: it's possible to know if next input to present is a header or a block,
- * using ZSTD_nextInputType(). This information is not required to properly
- * decode a frame.
- *
- * == Special case: skippable frames ==
- *
- * Skippable frames allow integration of user-defined data into a flow of
- * concatenated frames. Skippable frames will be ignored (skipped) by a
- * decompressor. The format of skippable frames is as follows:
- * a) Skippable frame ID - 4 Bytes, Little endian format, any value from
- *    0x184D2A50 to 0x184D2A5F
- * b) Frame Size - 4 Bytes, Little endian format, unsigned 32-bits
- * c) Frame Content - any content (User Data) of length equal to Frame Size
- * For skippable frames ZSTD_decompressContinue() always returns 0.
- * For skippable frames ZSTD_getFrameParams() returns fparamsPtr->windowLog==0
- * what means that a frame is skippable.
- * Note: If fparamsPtr->frameContentSize==0, it is ambiguous: the frame might
- *       actually be a zstd encoded frame with no content. For purposes of
- *       decompression, it is valid in both cases to skip the frame using
- *       ZSTD_findFrameCompressedSize() to find its size in bytes.
- * It also returns frame size as fparamsPtr->frameContentSize.
- ******************************************************************************/
-
-/*=====   Buffer-less streaming decompression functions  =====*/
-size_t ZSTD_decompressBegin(ZSTD_DCtx *dctx);
-size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx *dctx, const void *dict,
-	size_t dictSize);
-void   ZSTD_copyDCtx(ZSTD_DCtx *dctx, const ZSTD_DCtx *preparedDCtx);
-size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx *dctx);
-size_t ZSTD_decompressContinue(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity,
-	const void *src, size_t srcSize);
-typedef enum {
-	ZSTDnit_frameHeader,
-	ZSTDnit_blockHeader,
-	ZSTDnit_block,
-	ZSTDnit_lastBlock,
-	ZSTDnit_checksum,
-	ZSTDnit_skippableFrame
-} ZSTD_nextInputType_e;
-ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx *dctx);
-
-/*-*****************************************************************************
- * Block functions
- *
- * Block functions produce and decode raw zstd blocks, without frame metadata.
- * Frame metadata cost is typically ~18 bytes, which can be non-negligible for
- * very small blocks (< 100 bytes). User will have to take in charge required
- * information to regenerate data, such as compressed and content sizes.
- *
- * A few rules to respect:
- * - Compressing and decompressing require a context structure
- *   + Use ZSTD_initCCtx() and ZSTD_initDCtx()
- * - It is necessary to init context before starting
- *   + compression : ZSTD_compressBegin()
- *   + decompression : ZSTD_decompressBegin()
- *   + variants _usingDict() are also allowed
- *   + copyCCtx() and copyDCtx() work too
- * - Block size is limited, it must be <= ZSTD_getBlockSizeMax()
- *   + If you need to compress more, cut data into multiple blocks
- *   + Consider using the regular ZSTD_compress() instead, as frame metadata
- *     costs become negligible when source size is large.
- * - When a block is considered not compressible enough, ZSTD_compressBlock()
- *   result will be zero. In which case, nothing is produced into `dst`.
- *   + User must test for such outcome and deal directly with uncompressed data
- *   + ZSTD_decompressBlock() doesn't accept uncompressed data as input!!!
- *   + In case of multiple successive blocks, decoder must be informed of
- *     uncompressed block existence to follow proper history. Use
- *     ZSTD_insertBlock() in such a case.
- ******************************************************************************/
-
-/* Define for static allocation */
-#define ZSTD_BLOCKSIZE_ABSOLUTEMAX (128 * 1024)
-/*=====   Raw zstd block functions  =====*/
-size_t ZSTD_getBlockSizeMax(ZSTD_CCtx *cctx);
-size_t ZSTD_compressBlock(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity,
-	const void *src, size_t srcSize);
-size_t ZSTD_decompressBlock(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity,
-	const void *src, size_t srcSize);
-size_t ZSTD_insertBlock(ZSTD_DCtx *dctx, const void *blockStart,
-	size_t blockSize);
-
-#endif  /* ZSTD_H */
+#endif  /* LINUX_ZSTD_H */
diff --git a/include/linux/zstd_errors.h b/include/linux/zstd_errors.h
new file mode 100644
index 0000000..58b6dd4
--- /dev/null
+++ b/include/linux/zstd_errors.h
@@ -0,0 +1,77 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_ERRORS_H_398273423
+#define ZSTD_ERRORS_H_398273423
+
+
+/*===== dependency =====*/
+#include <linux/types.h>   /* size_t */
+
+
+/* =====   ZSTDERRORLIB_API : control library symbols visibility   ===== */
+#define ZSTDERRORLIB_VISIBILITY 
+#define ZSTDERRORLIB_API ZSTDERRORLIB_VISIBILITY
+
+/*-*********************************************
+ *  Error codes list
+ *-*********************************************
+ *  Error codes _values_ are pinned down since v1.3.1 only.
+ *  Therefore, don't rely on values if you may link to any version < v1.3.1.
+ *
+ *  Only values < 100 are considered stable.
+ *
+ *  note 1 : this API shall be used with static linking only.
+ *           dynamic linking is not yet officially supported.
+ *  note 2 : Prefer relying on the enum than on its value whenever possible
+ *           This is the only supported way to use the error list < v1.3.1
+ *  note 3 : ZSTD_isError() is always correct, whatever the library version.
+ **********************************************/
+typedef enum {
+  ZSTD_error_no_error = 0,
+  ZSTD_error_GENERIC  = 1,
+  ZSTD_error_prefix_unknown                = 10,
+  ZSTD_error_version_unsupported           = 12,
+  ZSTD_error_frameParameter_unsupported    = 14,
+  ZSTD_error_frameParameter_windowTooLarge = 16,
+  ZSTD_error_corruption_detected = 20,
+  ZSTD_error_checksum_wrong      = 22,
+  ZSTD_error_dictionary_corrupted      = 30,
+  ZSTD_error_dictionary_wrong          = 32,
+  ZSTD_error_dictionaryCreation_failed = 34,
+  ZSTD_error_parameter_unsupported   = 40,
+  ZSTD_error_parameter_outOfBound    = 42,
+  ZSTD_error_tableLog_tooLarge       = 44,
+  ZSTD_error_maxSymbolValue_tooLarge = 46,
+  ZSTD_error_maxSymbolValue_tooSmall = 48,
+  ZSTD_error_stage_wrong       = 60,
+  ZSTD_error_init_missing      = 62,
+  ZSTD_error_memory_allocation = 64,
+  ZSTD_error_workSpace_tooSmall= 66,
+  ZSTD_error_dstSize_tooSmall = 70,
+  ZSTD_error_srcSize_wrong    = 72,
+  ZSTD_error_dstBuffer_null   = 74,
+  /* following error codes are __NOT STABLE__, they can be removed or changed in future versions */
+  ZSTD_error_frameIndex_tooLarge = 100,
+  ZSTD_error_seekableIO          = 102,
+  ZSTD_error_dstBuffer_wrong     = 104,
+  ZSTD_error_srcBuffer_wrong     = 105,
+  ZSTD_error_maxCode = 120  /* never EVER use this value directly, it can change in future versions! Use ZSTD_isError() instead */
+} ZSTD_ErrorCode;
+
+/*! ZSTD_getErrorCode() :
+    convert a `size_t` function result into a `ZSTD_ErrorCode` enum type,
+    which can be used to compare with enum list published above */
+ZSTDERRORLIB_API ZSTD_ErrorCode ZSTD_getErrorCode(size_t functionResult);
+ZSTDERRORLIB_API const char* ZSTD_getErrorString(ZSTD_ErrorCode code);   /*< Same as ZSTD_getErrorName, but using a `ZSTD_ErrorCode` enum argument */
+
+
+
+#endif /* ZSTD_ERRORS_H_398273423 */
diff --git a/include/linux/zstd_lib.h b/include/linux/zstd_lib.h
new file mode 100644
index 0000000..b8c7dbf9
--- /dev/null
+++ b/include/linux/zstd_lib.h
@@ -0,0 +1,2432 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_H_235446
+#define ZSTD_H_235446
+
+/* ======   Dependency   ======*/
+#include <linux/limits.h>   /* INT_MAX */
+#include <linux/types.h>   /* size_t */
+
+
+/* =====   ZSTDLIB_API : control library symbols visibility   ===== */
+#define ZSTDLIB_VISIBILITY 
+#define ZSTDLIB_API ZSTDLIB_VISIBILITY
+
+
+/* *****************************************************************************
+  Introduction
+
+  zstd, short for Zstandard, is a fast lossless compression algorithm, targeting
+  real-time compression scenarios at zlib-level and better compression ratios.
+  The zstd compression library provides in-memory compression and decompression
+  functions.
+
+  The library supports regular compression levels from 1 up to ZSTD_maxCLevel(),
+  which is currently 22. Levels >= 20, labeled `--ultra`, should be used with
+  caution, as they require more memory. The library also offers negative
+  compression levels, which extend the range of speed vs. ratio preferences.
+  The lower the level, the faster the speed (at the cost of compression).
+
+  Compression can be done in:
+    - a single step (described as Simple API)
+    - a single step, reusing a context (described as Explicit context)
+    - unbounded multiple steps (described as Streaming compression)
+
+  The compression ratio achievable on small data can be highly improved using
+  a dictionary. Dictionary compression can be performed in:
+    - a single step (described as Simple dictionary API)
+    - a single step, reusing a dictionary (described as Bulk-processing
+      dictionary API)
+
+  Advanced experimental functions can be accessed using
+  `#define ZSTD_STATIC_LINKING_ONLY` before including zstd.h.
+
+  Advanced experimental APIs should never be used with a dynamically-linked
+  library. They are not "stable"; their definitions or signatures may change in
+  the future. Only static linking is allowed.
+*******************************************************************************/
+
+/*------   Version   ------*/
+#define ZSTD_VERSION_MAJOR    1
+#define ZSTD_VERSION_MINOR    4
+#define ZSTD_VERSION_RELEASE  10
+#define ZSTD_VERSION_NUMBER  (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE)
+
+/*! ZSTD_versionNumber() :
+ *  Return runtime library version, the value is (MAJOR*100*100 + MINOR*100 + RELEASE). */
+ZSTDLIB_API unsigned ZSTD_versionNumber(void);
+
+#define ZSTD_LIB_VERSION ZSTD_VERSION_MAJOR.ZSTD_VERSION_MINOR.ZSTD_VERSION_RELEASE
+#define ZSTD_QUOTE(str) #str
+#define ZSTD_EXPAND_AND_QUOTE(str) ZSTD_QUOTE(str)
+#define ZSTD_VERSION_STRING ZSTD_EXPAND_AND_QUOTE(ZSTD_LIB_VERSION)
+
+/*! ZSTD_versionString() :
+ *  Return runtime library version, like "1.4.5". Requires v1.3.0+. */
+ZSTDLIB_API const char* ZSTD_versionString(void);
+
+/* *************************************
+ *  Default constant
+ ***************************************/
+#ifndef ZSTD_CLEVEL_DEFAULT
+#  define ZSTD_CLEVEL_DEFAULT 3
+#endif
+
+/* *************************************
+ *  Constants
+ ***************************************/
+
+/* All magic numbers are supposed read/written to/from files/memory using little-endian convention */
+#define ZSTD_MAGICNUMBER            0xFD2FB528    /* valid since v0.8.0 */
+#define ZSTD_MAGIC_DICTIONARY       0xEC30A437    /* valid since v0.7.0 */
+#define ZSTD_MAGIC_SKIPPABLE_START  0x184D2A50    /* all 16 values, from 0x184D2A50 to 0x184D2A5F, signal the beginning of a skippable frame */
+#define ZSTD_MAGIC_SKIPPABLE_MASK   0xFFFFFFF0
+
+#define ZSTD_BLOCKSIZELOG_MAX  17
+#define ZSTD_BLOCKSIZE_MAX     (1<<ZSTD_BLOCKSIZELOG_MAX)
+
+
+
+/* *************************************
+*  Simple API
+***************************************/
+/*! ZSTD_compress() :
+ *  Compresses `src` content as a single zstd compressed frame into already allocated `dst`.
+ *  Hint : compression runs faster if `dstCapacity` >=  `ZSTD_compressBound(srcSize)`.
+ *  @return : compressed size written into `dst` (<= `dstCapacity),
+ *            or an error code if it fails (which can be tested using ZSTD_isError()). */
+ZSTDLIB_API size_t ZSTD_compress( void* dst, size_t dstCapacity,
+                            const void* src, size_t srcSize,
+                                  int compressionLevel);
+
+/*! ZSTD_decompress() :
+ *  `compressedSize` : must be the _exact_ size of some number of compressed and/or skippable frames.
+ *  `dstCapacity` is an upper bound of originalSize to regenerate.
+ *  If user cannot imply a maximum upper bound, it's better to use streaming mode to decompress data.
+ *  @return : the number of bytes decompressed into `dst` (<= `dstCapacity`),
+ *            or an errorCode if it fails (which can be tested using ZSTD_isError()). */
+ZSTDLIB_API size_t ZSTD_decompress( void* dst, size_t dstCapacity,
+                              const void* src, size_t compressedSize);
+
+/*! ZSTD_getFrameContentSize() : requires v1.3.0+
+ *  `src` should point to the start of a ZSTD encoded frame.
+ *  `srcSize` must be at least as large as the frame header.
+ *            hint : any size >= `ZSTD_frameHeaderSize_max` is large enough.
+ *  @return : - decompressed size of `src` frame content, if known
+ *            - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined
+ *            - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small)
+ *   note 1 : a 0 return value means the frame is valid but "empty".
+ *   note 2 : decompressed size is an optional field, it may not be present, typically in streaming mode.
+ *            When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size.
+ *            In which case, it's necessary to use streaming mode to decompress data.
+ *            Optionally, application can rely on some implicit limit,
+ *            as ZSTD_decompress() only needs an upper bound of decompressed size.
+ *            (For example, data could be necessarily cut into blocks <= 16 KB).
+ *   note 3 : decompressed size is always present when compression is completed using single-pass functions,
+ *            such as ZSTD_compress(), ZSTD_compressCCtx() ZSTD_compress_usingDict() or ZSTD_compress_usingCDict().
+ *   note 4 : decompressed size can be very large (64-bits value),
+ *            potentially larger than what local system can handle as a single memory segment.
+ *            In which case, it's necessary to use streaming mode to decompress data.
+ *   note 5 : If source is untrusted, decompressed size could be wrong or intentionally modified.
+ *            Always ensure return value fits within application's authorized limits.
+ *            Each application can set its own limits.
+ *   note 6 : This function replaces ZSTD_getDecompressedSize() */
+#define ZSTD_CONTENTSIZE_UNKNOWN (0ULL - 1)
+#define ZSTD_CONTENTSIZE_ERROR   (0ULL - 2)
+ZSTDLIB_API unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize);
+
+/*! ZSTD_getDecompressedSize() :
+ *  NOTE: This function is now obsolete, in favor of ZSTD_getFrameContentSize().
+ *  Both functions work the same way, but ZSTD_getDecompressedSize() blends
+ *  "empty", "unknown" and "error" results to the same return value (0),
+ *  while ZSTD_getFrameContentSize() gives them separate return values.
+ * @return : decompressed size of `src` frame content _if known and not empty_, 0 otherwise. */
+ZSTDLIB_API unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize);
+
+/*! ZSTD_findFrameCompressedSize() :
+ * `src` should point to the start of a ZSTD frame or skippable frame.
+ * `srcSize` must be >= first frame size
+ * @return : the compressed size of the first frame starting at `src`,
+ *           suitable to pass as `srcSize` to `ZSTD_decompress` or similar,
+ *        or an error code if input is invalid */
+ZSTDLIB_API size_t ZSTD_findFrameCompressedSize(const void* src, size_t srcSize);
+
+
+/*======  Helper functions  ======*/
+#define ZSTD_COMPRESSBOUND(srcSize)   ((srcSize) + ((srcSize)>>8) + (((srcSize) < (128<<10)) ? (((128<<10) - (srcSize)) >> 11) /* margin, from 64 to 0 */ : 0))  /* this formula ensures that bound(A) + bound(B) <= bound(A+B) as long as A and B >= 128 KB */
+ZSTDLIB_API size_t      ZSTD_compressBound(size_t srcSize); /*!< maximum compressed size in worst case single-pass scenario */
+ZSTDLIB_API unsigned    ZSTD_isError(size_t code);          /*!< tells if a `size_t` function result is an error code */
+ZSTDLIB_API const char* ZSTD_getErrorName(size_t code);     /*!< provides readable string from an error code */
+ZSTDLIB_API int         ZSTD_minCLevel(void);               /*!< minimum negative compression level allowed */
+ZSTDLIB_API int         ZSTD_maxCLevel(void);               /*!< maximum compression level available */
+
+
+/* *************************************
+*  Explicit context
+***************************************/
+/*= Compression context
+ *  When compressing many times,
+ *  it is recommended to allocate a context just once,
+ *  and re-use it for each successive compression operation.
+ *  This will make workload friendlier for system's memory.
+ *  Note : re-using context is just a speed / resource optimization.
+ *         It doesn't change the compression ratio, which remains identical.
+ *  Note 2 : In multi-threaded environments,
+ *         use one different context per thread for parallel execution.
+ */
+typedef struct ZSTD_CCtx_s ZSTD_CCtx;
+ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx(void);
+ZSTDLIB_API size_t     ZSTD_freeCCtx(ZSTD_CCtx* cctx);  /* accept NULL pointer */
+
+/*! ZSTD_compressCCtx() :
+ *  Same as ZSTD_compress(), using an explicit ZSTD_CCtx.
+ *  Important : in order to behave similarly to `ZSTD_compress()`,
+ *  this function compresses at requested compression level,
+ *  __ignoring any other parameter__ .
+ *  If any advanced parameter was set using the advanced API,
+ *  they will all be reset. Only `compressionLevel` remains.
+ */
+ZSTDLIB_API size_t ZSTD_compressCCtx(ZSTD_CCtx* cctx,
+                                     void* dst, size_t dstCapacity,
+                               const void* src, size_t srcSize,
+                                     int compressionLevel);
+
+/*= Decompression context
+ *  When decompressing many times,
+ *  it is recommended to allocate a context only once,
+ *  and re-use it for each successive compression operation.
+ *  This will make workload friendlier for system's memory.
+ *  Use one context per thread for parallel execution. */
+typedef struct ZSTD_DCtx_s ZSTD_DCtx;
+ZSTDLIB_API ZSTD_DCtx* ZSTD_createDCtx(void);
+ZSTDLIB_API size_t     ZSTD_freeDCtx(ZSTD_DCtx* dctx);  /* accept NULL pointer */
+
+/*! ZSTD_decompressDCtx() :
+ *  Same as ZSTD_decompress(),
+ *  requires an allocated ZSTD_DCtx.
+ *  Compatible with sticky parameters.
+ */
+ZSTDLIB_API size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx,
+                                       void* dst, size_t dstCapacity,
+                                 const void* src, size_t srcSize);
+
+
+/* *************************************
+*  Advanced compression API
+***************************************/
+
+/* API design :
+ *   Parameters are pushed one by one into an existing context,
+ *   using ZSTD_CCtx_set*() functions.
+ *   Pushed parameters are sticky : they are valid for next compressed frame, and any subsequent frame.
+ *   "sticky" parameters are applicable to `ZSTD_compress2()` and `ZSTD_compressStream*()` !
+ *   __They do not apply to "simple" one-shot variants such as ZSTD_compressCCtx()__ .
+ *
+ *   It's possible to reset all parameters to "default" using ZSTD_CCtx_reset().
+ *
+ *   This API supercedes all other "advanced" API entry points in the experimental section.
+ *   In the future, we expect to remove from experimental API entry points which are redundant with this API.
+ */
+
+
+/* Compression strategies, listed from fastest to strongest */
+typedef enum { ZSTD_fast=1,
+               ZSTD_dfast=2,
+               ZSTD_greedy=3,
+               ZSTD_lazy=4,
+               ZSTD_lazy2=5,
+               ZSTD_btlazy2=6,
+               ZSTD_btopt=7,
+               ZSTD_btultra=8,
+               ZSTD_btultra2=9
+               /* note : new strategies _might_ be added in the future.
+                         Only the order (from fast to strong) is guaranteed */
+} ZSTD_strategy;
+
+
+typedef enum {
+
+    /* compression parameters
+     * Note: When compressing with a ZSTD_CDict these parameters are superseded
+     * by the parameters used to construct the ZSTD_CDict.
+     * See ZSTD_CCtx_refCDict() for more info (superseded-by-cdict). */
+    ZSTD_c_compressionLevel=100, /* Set compression parameters according to pre-defined cLevel table.
+                              * Note that exact compression parameters are dynamically determined,
+                              * depending on both compression level and srcSize (when known).
+                              * Default level is ZSTD_CLEVEL_DEFAULT==3.
+                              * Special: value 0 means default, which is controlled by ZSTD_CLEVEL_DEFAULT.
+                              * Note 1 : it's possible to pass a negative compression level.
+                              * Note 2 : setting a level does not automatically set all other compression parameters
+                              *   to default. Setting this will however eventually dynamically impact the compression
+                              *   parameters which have not been manually set. The manually set
+                              *   ones will 'stick'. */
+    /* Advanced compression parameters :
+     * It's possible to pin down compression parameters to some specific values.
+     * In which case, these values are no longer dynamically selected by the compressor */
+    ZSTD_c_windowLog=101,    /* Maximum allowed back-reference distance, expressed as power of 2.
+                              * This will set a memory budget for streaming decompression,
+                              * with larger values requiring more memory
+                              * and typically compressing more.
+                              * Must be clamped between ZSTD_WINDOWLOG_MIN and ZSTD_WINDOWLOG_MAX.
+                              * Special: value 0 means "use default windowLog".
+                              * Note: Using a windowLog greater than ZSTD_WINDOWLOG_LIMIT_DEFAULT
+                              *       requires explicitly allowing such size at streaming decompression stage. */
+    ZSTD_c_hashLog=102,      /* Size of the initial probe table, as a power of 2.
+                              * Resulting memory usage is (1 << (hashLog+2)).
+                              * Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX.
+                              * Larger tables improve compression ratio of strategies <= dFast,
+                              * and improve speed of strategies > dFast.
+                              * Special: value 0 means "use default hashLog". */
+    ZSTD_c_chainLog=103,     /* Size of the multi-probe search table, as a power of 2.
+                              * Resulting memory usage is (1 << (chainLog+2)).
+                              * Must be clamped between ZSTD_CHAINLOG_MIN and ZSTD_CHAINLOG_MAX.
+                              * Larger tables result in better and slower compression.
+                              * This parameter is useless for "fast" strategy.
+                              * It's still useful when using "dfast" strategy,
+                              * in which case it defines a secondary probe table.
+                              * Special: value 0 means "use default chainLog". */
+    ZSTD_c_searchLog=104,    /* Number of search attempts, as a power of 2.
+                              * More attempts result in better and slower compression.
+                              * This parameter is useless for "fast" and "dFast" strategies.
+                              * Special: value 0 means "use default searchLog". */
+    ZSTD_c_minMatch=105,     /* Minimum size of searched matches.
+                              * Note that Zstandard can still find matches of smaller size,
+                              * it just tweaks its search algorithm to look for this size and larger.
+                              * Larger values increase compression and decompression speed, but decrease ratio.
+                              * Must be clamped between ZSTD_MINMATCH_MIN and ZSTD_MINMATCH_MAX.
+                              * Note that currently, for all strategies < btopt, effective minimum is 4.
+                              *                    , for all strategies > fast, effective maximum is 6.
+                              * Special: value 0 means "use default minMatchLength". */
+    ZSTD_c_targetLength=106, /* Impact of this field depends on strategy.
+                              * For strategies btopt, btultra & btultra2:
+                              *     Length of Match considered "good enough" to stop search.
+                              *     Larger values make compression stronger, and slower.
+                              * For strategy fast:
+                              *     Distance between match sampling.
+                              *     Larger values make compression faster, and weaker.
+                              * Special: value 0 means "use default targetLength". */
+    ZSTD_c_strategy=107,     /* See ZSTD_strategy enum definition.
+                              * The higher the value of selected strategy, the more complex it is,
+                              * resulting in stronger and slower compression.
+                              * Special: value 0 means "use default strategy". */
+
+    /* LDM mode parameters */
+    ZSTD_c_enableLongDistanceMatching=160, /* Enable long distance matching.
+                                     * This parameter is designed to improve compression ratio
+                                     * for large inputs, by finding large matches at long distance.
+                                     * It increases memory usage and window size.
+                                     * Note: enabling this parameter increases default ZSTD_c_windowLog to 128 MB
+                                     * except when expressly set to a different value.
+                                     * Note: will be enabled by default if ZSTD_c_windowLog >= 128 MB and
+                                     * compression strategy >= ZSTD_btopt (== compression level 16+) */
+    ZSTD_c_ldmHashLog=161,   /* Size of the table for long distance matching, as a power of 2.
+                              * Larger values increase memory usage and compression ratio,
+                              * but decrease compression speed.
+                              * Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX
+                              * default: windowlog - 7.
+                              * Special: value 0 means "automatically determine hashlog". */
+    ZSTD_c_ldmMinMatch=162,  /* Minimum match size for long distance matcher.
+                              * Larger/too small values usually decrease compression ratio.
+                              * Must be clamped between ZSTD_LDM_MINMATCH_MIN and ZSTD_LDM_MINMATCH_MAX.
+                              * Special: value 0 means "use default value" (default: 64). */
+    ZSTD_c_ldmBucketSizeLog=163, /* Log size of each bucket in the LDM hash table for collision resolution.
+                              * Larger values improve collision resolution but decrease compression speed.
+                              * The maximum value is ZSTD_LDM_BUCKETSIZELOG_MAX.
+                              * Special: value 0 means "use default value" (default: 3). */
+    ZSTD_c_ldmHashRateLog=164, /* Frequency of inserting/looking up entries into the LDM hash table.
+                              * Must be clamped between 0 and (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN).
+                              * Default is MAX(0, (windowLog - ldmHashLog)), optimizing hash table usage.
+                              * Larger values improve compression speed.
+                              * Deviating far from default value will likely result in a compression ratio decrease.
+                              * Special: value 0 means "automatically determine hashRateLog". */
+
+    /* frame parameters */
+    ZSTD_c_contentSizeFlag=200, /* Content size will be written into frame header _whenever known_ (default:1)
+                              * Content size must be known at the beginning of compression.
+                              * This is automatically the case when using ZSTD_compress2(),
+                              * For streaming scenarios, content size must be provided with ZSTD_CCtx_setPledgedSrcSize() */
+    ZSTD_c_checksumFlag=201, /* A 32-bits checksum of content is written at end of frame (default:0) */
+    ZSTD_c_dictIDFlag=202,   /* When applicable, dictionary's ID is written into frame header (default:1) */
+
+    /* multi-threading parameters */
+    /* These parameters are only active if multi-threading is enabled (compiled with build macro ZSTD_MULTITHREAD).
+     * Otherwise, trying to set any other value than default (0) will be a no-op and return an error.
+     * In a situation where it's unknown if the linked library supports multi-threading or not,
+     * setting ZSTD_c_nbWorkers to any value >= 1 and consulting the return value provides a quick way to check this property.
+     */
+    ZSTD_c_nbWorkers=400,    /* Select how many threads will be spawned to compress in parallel.
+                              * When nbWorkers >= 1, triggers asynchronous mode when invoking ZSTD_compressStream*() :
+                              * ZSTD_compressStream*() consumes input and flush output if possible, but immediately gives back control to caller,
+                              * while compression is performed in parallel, within worker thread(s).
+                              * (note : a strong exception to this rule is when first invocation of ZSTD_compressStream2() sets ZSTD_e_end :
+                              *  in which case, ZSTD_compressStream2() delegates to ZSTD_compress2(), which is always a blocking call).
+                              * More workers improve speed, but also increase memory usage.
+                              * Default value is `0`, aka "single-threaded mode" : no worker is spawned,
+                              * compression is performed inside Caller's thread, and all invocations are blocking */
+    ZSTD_c_jobSize=401,      /* Size of a compression job. This value is enforced only when nbWorkers >= 1.
+                              * Each compression job is completed in parallel, so this value can indirectly impact the nb of active threads.
+                              * 0 means default, which is dynamically determined based on compression parameters.
+                              * Job size must be a minimum of overlap size, or 1 MB, whichever is largest.
+                              * The minimum size is automatically and transparently enforced. */
+    ZSTD_c_overlapLog=402,   /* Control the overlap size, as a fraction of window size.
+                              * The overlap size is an amount of data reloaded from previous job at the beginning of a new job.
+                              * It helps preserve compression ratio, while each job is compressed in parallel.
+                              * This value is enforced only when nbWorkers >= 1.
+                              * Larger values increase compression ratio, but decrease speed.
+                              * Possible values range from 0 to 9 :
+                              * - 0 means "default" : value will be determined by the library, depending on strategy
+                              * - 1 means "no overlap"
+                              * - 9 means "full overlap", using a full window size.
+                              * Each intermediate rank increases/decreases load size by a factor 2 :
+                              * 9: full window;  8: w/2;  7: w/4;  6: w/8;  5:w/16;  4: w/32;  3:w/64;  2:w/128;  1:no overlap;  0:default
+                              * default value varies between 6 and 9, depending on strategy */
+
+    /* note : additional experimental parameters are also available
+     * within the experimental section of the API.
+     * At the time of this writing, they include :
+     * ZSTD_c_rsyncable
+     * ZSTD_c_format
+     * ZSTD_c_forceMaxWindow
+     * ZSTD_c_forceAttachDict
+     * ZSTD_c_literalCompressionMode
+     * ZSTD_c_targetCBlockSize
+     * ZSTD_c_srcSizeHint
+     * ZSTD_c_enableDedicatedDictSearch
+     * ZSTD_c_stableInBuffer
+     * ZSTD_c_stableOutBuffer
+     * ZSTD_c_blockDelimiters
+     * ZSTD_c_validateSequences
+     * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.
+     * note : never ever use experimentalParam? names directly;
+     *        also, the enums values themselves are unstable and can still change.
+     */
+     ZSTD_c_experimentalParam1=500,
+     ZSTD_c_experimentalParam2=10,
+     ZSTD_c_experimentalParam3=1000,
+     ZSTD_c_experimentalParam4=1001,
+     ZSTD_c_experimentalParam5=1002,
+     ZSTD_c_experimentalParam6=1003,
+     ZSTD_c_experimentalParam7=1004,
+     ZSTD_c_experimentalParam8=1005,
+     ZSTD_c_experimentalParam9=1006,
+     ZSTD_c_experimentalParam10=1007,
+     ZSTD_c_experimentalParam11=1008,
+     ZSTD_c_experimentalParam12=1009
+} ZSTD_cParameter;
+
+typedef struct {
+    size_t error;
+    int lowerBound;
+    int upperBound;
+} ZSTD_bounds;
+
+/*! ZSTD_cParam_getBounds() :
+ *  All parameters must belong to an interval with lower and upper bounds,
+ *  otherwise they will either trigger an error or be automatically clamped.
+ * @return : a structure, ZSTD_bounds, which contains
+ *         - an error status field, which must be tested using ZSTD_isError()
+ *         - lower and upper bounds, both inclusive
+ */
+ZSTDLIB_API ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_cParameter cParam);
+
+/*! ZSTD_CCtx_setParameter() :
+ *  Set one compression parameter, selected by enum ZSTD_cParameter.
+ *  All parameters have valid bounds. Bounds can be queried using ZSTD_cParam_getBounds().
+ *  Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter).
+ *  Setting a parameter is generally only possible during frame initialization (before starting compression).
+ *  Exception : when using multi-threading mode (nbWorkers >= 1),
+ *              the following parameters can be updated _during_ compression (within same frame):
+ *              => compressionLevel, hashLog, chainLog, searchLog, minMatch, targetLength and strategy.
+ *              new parameters will be active for next job only (after a flush()).
+ * @return : an error code (which can be tested using ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value);
+
+/*! ZSTD_CCtx_setPledgedSrcSize() :
+ *  Total input data size to be compressed as a single frame.
+ *  Value will be written in frame header, unless if explicitly forbidden using ZSTD_c_contentSizeFlag.
+ *  This value will also be controlled at end of frame, and trigger an error if not respected.
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ *  Note 1 : pledgedSrcSize==0 actually means zero, aka an empty frame.
+ *           In order to mean "unknown content size", pass constant ZSTD_CONTENTSIZE_UNKNOWN.
+ *           ZSTD_CONTENTSIZE_UNKNOWN is default value for any new frame.
+ *  Note 2 : pledgedSrcSize is only valid once, for the next frame.
+ *           It's discarded at the end of the frame, and replaced by ZSTD_CONTENTSIZE_UNKNOWN.
+ *  Note 3 : Whenever all input data is provided and consumed in a single round,
+ *           for example with ZSTD_compress2(),
+ *           or invoking immediately ZSTD_compressStream2(,,,ZSTD_e_end),
+ *           this value is automatically overridden by srcSize instead.
+ */
+ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize);
+
+typedef enum {
+    ZSTD_reset_session_only = 1,
+    ZSTD_reset_parameters = 2,
+    ZSTD_reset_session_and_parameters = 3
+} ZSTD_ResetDirective;
+
+/*! ZSTD_CCtx_reset() :
+ *  There are 2 different things that can be reset, independently or jointly :
+ *  - The session : will stop compressing current frame, and make CCtx ready to start a new one.
+ *                  Useful after an error, or to interrupt any ongoing compression.
+ *                  Any internal data not yet flushed is cancelled.
+ *                  Compression parameters and dictionary remain unchanged.
+ *                  They will be used to compress next frame.
+ *                  Resetting session never fails.
+ *  - The parameters : changes all parameters back to "default".
+ *                  This removes any reference to any dictionary too.
+ *                  Parameters can only be changed between 2 sessions (i.e. no compression is currently ongoing)
+ *                  otherwise the reset fails, and function returns an error value (which can be tested using ZSTD_isError())
+ *  - Both : similar to resetting the session, followed by resetting parameters.
+ */
+ZSTDLIB_API size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset);
+
+/*! ZSTD_compress2() :
+ *  Behave the same as ZSTD_compressCCtx(), but compression parameters are set using the advanced API.
+ *  ZSTD_compress2() always starts a new frame.
+ *  Should cctx hold data from a previously unfinished frame, everything about it is forgotten.
+ *  - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
+ *  - The function is always blocking, returns when compression is completed.
+ *  Hint : compression runs faster if `dstCapacity` >=  `ZSTD_compressBound(srcSize)`.
+ * @return : compressed size written into `dst` (<= `dstCapacity),
+ *           or an error code if it fails (which can be tested using ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_compress2( ZSTD_CCtx* cctx,
+                                   void* dst, size_t dstCapacity,
+                             const void* src, size_t srcSize);
+
+
+/* *************************************
+*  Advanced decompression API
+***************************************/
+
+/* The advanced API pushes parameters one by one into an existing DCtx context.
+ * Parameters are sticky, and remain valid for all following frames
+ * using the same DCtx context.
+ * It's possible to reset parameters to default values using ZSTD_DCtx_reset().
+ * Note : This API is compatible with existing ZSTD_decompressDCtx() and ZSTD_decompressStream().
+ *        Therefore, no new decompression function is necessary.
+ */
+
+typedef enum {
+
+    ZSTD_d_windowLogMax=100, /* Select a size limit (in power of 2) beyond which
+                              * the streaming API will refuse to allocate memory buffer
+                              * in order to protect the host from unreasonable memory requirements.
+                              * This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode.
+                              * By default, a decompression context accepts window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT).
+                              * Special: value 0 means "use default maximum windowLog". */
+
+    /* note : additional experimental parameters are also available
+     * within the experimental section of the API.
+     * At the time of this writing, they include :
+     * ZSTD_d_format
+     * ZSTD_d_stableOutBuffer
+     * ZSTD_d_forceIgnoreChecksum
+     * ZSTD_d_refMultipleDDicts
+     * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.
+     * note : never ever use experimentalParam? names directly
+     */
+     ZSTD_d_experimentalParam1=1000,
+     ZSTD_d_experimentalParam2=1001,
+     ZSTD_d_experimentalParam3=1002,
+     ZSTD_d_experimentalParam4=1003
+
+} ZSTD_dParameter;
+
+/*! ZSTD_dParam_getBounds() :
+ *  All parameters must belong to an interval with lower and upper bounds,
+ *  otherwise they will either trigger an error or be automatically clamped.
+ * @return : a structure, ZSTD_bounds, which contains
+ *         - an error status field, which must be tested using ZSTD_isError()
+ *         - both lower and upper bounds, inclusive
+ */
+ZSTDLIB_API ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam);
+
+/*! ZSTD_DCtx_setParameter() :
+ *  Set one compression parameter, selected by enum ZSTD_dParameter.
+ *  All parameters have valid bounds. Bounds can be queried using ZSTD_dParam_getBounds().
+ *  Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter).
+ *  Setting a parameter is only possible during frame initialization (before starting decompression).
+ * @return : 0, or an error code (which can be tested using ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int value);
+
+/*! ZSTD_DCtx_reset() :
+ *  Return a DCtx to clean state.
+ *  Session and parameters can be reset jointly or separately.
+ *  Parameters can only be reset when no active frame is being decompressed.
+ * @return : 0, or an error code, which can be tested with ZSTD_isError()
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset);
+
+
+/* **************************
+*  Streaming
+****************************/
+
+typedef struct ZSTD_inBuffer_s {
+  const void* src;    /*< start of input buffer */
+  size_t size;        /*< size of input buffer */
+  size_t pos;         /*< position where reading stopped. Will be updated. Necessarily 0 <= pos <= size */
+} ZSTD_inBuffer;
+
+typedef struct ZSTD_outBuffer_s {
+  void*  dst;         /*< start of output buffer */
+  size_t size;        /*< size of output buffer */
+  size_t pos;         /*< position where writing stopped. Will be updated. Necessarily 0 <= pos <= size */
+} ZSTD_outBuffer;
+
+
+
+/*-***********************************************************************
+*  Streaming compression - HowTo
+*
+*  A ZSTD_CStream object is required to track streaming operation.
+*  Use ZSTD_createCStream() and ZSTD_freeCStream() to create/release resources.
+*  ZSTD_CStream objects can be reused multiple times on consecutive compression operations.
+*  It is recommended to re-use ZSTD_CStream since it will play nicer with system's memory, by re-using already allocated memory.
+*
+*  For parallel execution, use one separate ZSTD_CStream per thread.
+*
+*  note : since v1.3.0, ZSTD_CStream and ZSTD_CCtx are the same thing.
+*
+*  Parameters are sticky : when starting a new compression on the same context,
+*  it will re-use the same sticky parameters as previous compression session.
+*  When in doubt, it's recommended to fully initialize the context before usage.
+*  Use ZSTD_CCtx_reset() to reset the context and ZSTD_CCtx_setParameter(),
+*  ZSTD_CCtx_setPledgedSrcSize(), or ZSTD_CCtx_loadDictionary() and friends to
+*  set more specific parameters, the pledged source size, or load a dictionary.
+*
+*  Use ZSTD_compressStream2() with ZSTD_e_continue as many times as necessary to
+*  consume input stream. The function will automatically update both `pos`
+*  fields within `input` and `output`.
+*  Note that the function may not consume the entire input, for example, because
+*  the output buffer is already full, in which case `input.pos < input.size`.
+*  The caller must check if input has been entirely consumed.
+*  If not, the caller must make some room to receive more compressed data,
+*  and then present again remaining input data.
+*  note: ZSTD_e_continue is guaranteed to make some forward progress when called,
+*        but doesn't guarantee maximal forward progress. This is especially relevant
+*        when compressing with multiple threads. The call won't block if it can
+*        consume some input, but if it can't it will wait for some, but not all,
+*        output to be flushed.
+* @return : provides a minimum amount of data remaining to be flushed from internal buffers
+*           or an error code, which can be tested using ZSTD_isError().
+*
+*  At any moment, it's possible to flush whatever data might remain stuck within internal buffer,
+*  using ZSTD_compressStream2() with ZSTD_e_flush. `output->pos` will be updated.
+*  Note that, if `output->size` is too small, a single invocation with ZSTD_e_flush might not be enough (return code > 0).
+*  In which case, make some room to receive more compressed data, and call again ZSTD_compressStream2() with ZSTD_e_flush.
+*  You must continue calling ZSTD_compressStream2() with ZSTD_e_flush until it returns 0, at which point you can change the
+*  operation.
+*  note: ZSTD_e_flush will flush as much output as possible, meaning when compressing with multiple threads, it will
+*        block until the flush is complete or the output buffer is full.
+*  @return : 0 if internal buffers are entirely flushed,
+*            >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),
+*            or an error code, which can be tested using ZSTD_isError().
+*
+*  Calling ZSTD_compressStream2() with ZSTD_e_end instructs to finish a frame.
+*  It will perform a flush and write frame epilogue.
+*  The epilogue is required for decoders to consider a frame completed.
+*  flush operation is the same, and follows same rules as calling ZSTD_compressStream2() with ZSTD_e_flush.
+*  You must continue calling ZSTD_compressStream2() with ZSTD_e_end until it returns 0, at which point you are free to
+*  start a new frame.
+*  note: ZSTD_e_end will flush as much output as possible, meaning when compressing with multiple threads, it will
+*        block until the flush is complete or the output buffer is full.
+*  @return : 0 if frame fully completed and fully flushed,
+*            >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),
+*            or an error code, which can be tested using ZSTD_isError().
+*
+* *******************************************************************/
+
+typedef ZSTD_CCtx ZSTD_CStream;  /*< CCtx and CStream are now effectively same object (>= v1.3.0) */
+                                 /* Continue to distinguish them for compatibility with older versions <= v1.2.0 */
+/*===== ZSTD_CStream management functions =====*/
+ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream(void);
+ZSTDLIB_API size_t ZSTD_freeCStream(ZSTD_CStream* zcs);  /* accept NULL pointer */
+
+/*===== Streaming compression functions =====*/
+typedef enum {
+    ZSTD_e_continue=0, /* collect more data, encoder decides when to output compressed result, for optimal compression ratio */
+    ZSTD_e_flush=1,    /* flush any data provided so far,
+                        * it creates (at least) one new block, that can be decoded immediately on reception;
+                        * frame will continue: any future data can still reference previously compressed data, improving compression.
+                        * note : multithreaded compression will block to flush as much output as possible. */
+    ZSTD_e_end=2       /* flush any remaining data _and_ close current frame.
+                        * note that frame is only closed after compressed data is fully flushed (return value == 0).
+                        * After that point, any additional data starts a new frame.
+                        * note : each frame is independent (does not reference any content from previous frame).
+                        : note : multithreaded compression will block to flush as much output as possible. */
+} ZSTD_EndDirective;
+
+/*! ZSTD_compressStream2() :
+ *  Behaves about the same as ZSTD_compressStream, with additional control on end directive.
+ *  - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
+ *  - Compression parameters cannot be changed once compression is started (save a list of exceptions in multi-threading mode)
+ *  - output->pos must be <= dstCapacity, input->pos must be <= srcSize
+ *  - output->pos and input->pos will be updated. They are guaranteed to remain below their respective limit.
+ *  - endOp must be a valid directive
+ *  - When nbWorkers==0 (default), function is blocking : it completes its job before returning to caller.
+ *  - When nbWorkers>=1, function is non-blocking : it copies a portion of input, distributes jobs to internal worker threads, flush to output whatever is available,
+ *                                                  and then immediately returns, just indicating that there is some data remaining to be flushed.
+ *                                                  The function nonetheless guarantees forward progress : it will return only after it reads or write at least 1+ byte.
+ *  - Exception : if the first call requests a ZSTD_e_end directive and provides enough dstCapacity, the function delegates to ZSTD_compress2() which is always blocking.
+ *  - @return provides a minimum amount of data remaining to be flushed from internal buffers
+ *            or an error code, which can be tested using ZSTD_isError().
+ *            if @return != 0, flush is not fully completed, there is still some data left within internal buffers.
+ *            This is useful for ZSTD_e_flush, since in this case more flushes are necessary to empty all buffers.
+ *            For ZSTD_e_end, @return == 0 when internal buffers are fully flushed and frame is completed.
+ *  - after a ZSTD_e_end directive, if internal buffer is not fully flushed (@return != 0),
+ *            only ZSTD_e_end or ZSTD_e_flush operations are allowed.
+ *            Before starting a new compression job, or changing compression parameters,
+ *            it is required to fully flush internal buffers.
+ */
+ZSTDLIB_API size_t ZSTD_compressStream2( ZSTD_CCtx* cctx,
+                                         ZSTD_outBuffer* output,
+                                         ZSTD_inBuffer* input,
+                                         ZSTD_EndDirective endOp);
+
+
+/* These buffer sizes are softly recommended.
+ * They are not required : ZSTD_compressStream*() happily accepts any buffer size, for both input and output.
+ * Respecting the recommended size just makes it a bit easier for ZSTD_compressStream*(),
+ * reducing the amount of memory shuffling and buffering, resulting in minor performance savings.
+ *
+ * However, note that these recommendations are from the perspective of a C caller program.
+ * If the streaming interface is invoked from some other language,
+ * especially managed ones such as Java or Go, through a foreign function interface such as jni or cgo,
+ * a major performance rule is to reduce crossing such interface to an absolute minimum.
+ * It's not rare that performance ends being spent more into the interface, rather than compression itself.
+ * In which cases, prefer using large buffers, as large as practical,
+ * for both input and output, to reduce the nb of roundtrips.
+ */
+ZSTDLIB_API size_t ZSTD_CStreamInSize(void);    /*< recommended size for input buffer */
+ZSTDLIB_API size_t ZSTD_CStreamOutSize(void);   /*< recommended size for output buffer. Guarantee to successfully flush at least one complete compressed block. */
+
+
+/* *****************************************************************************
+ * This following is a legacy streaming API.
+ * It can be replaced by ZSTD_CCtx_reset() and ZSTD_compressStream2().
+ * It is redundant, but remains fully supported.
+ * Advanced parameters and dictionary compression can only be used through the
+ * new API.
+ ******************************************************************************/
+
+/*!
+ * Equivalent to:
+ *
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any)
+ *     ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
+ */
+ZSTDLIB_API size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel);
+/*!
+ * Alternative for ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue).
+ * NOTE: The return value is different. ZSTD_compressStream() returns a hint for
+ * the next read size (if non-zero and not an error). ZSTD_compressStream2()
+ * returns the minimum nb of bytes left to flush (if non-zero and not an error).
+ */
+ZSTDLIB_API size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
+/*! Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_flush). */
+ZSTDLIB_API size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
+/*! Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_end). */
+ZSTDLIB_API size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
+
+
+/*-***************************************************************************
+*  Streaming decompression - HowTo
+*
+*  A ZSTD_DStream object is required to track streaming operations.
+*  Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources.
+*  ZSTD_DStream objects can be re-used multiple times.
+*
+*  Use ZSTD_initDStream() to start a new decompression operation.
+* @return : recommended first input size
+*  Alternatively, use advanced API to set specific properties.
+*
+*  Use ZSTD_decompressStream() repetitively to consume your input.
+*  The function will update both `pos` fields.
+*  If `input.pos < input.size`, some input has not been consumed.
+*  It's up to the caller to present again remaining data.
+*  The function tries to flush all data decoded immediately, respecting output buffer size.
+*  If `output.pos < output.size`, decoder has flushed everything it could.
+*  But if `output.pos == output.size`, there might be some data left within internal buffers.,
+*  In which case, call ZSTD_decompressStream() again to flush whatever remains in the buffer.
+*  Note : with no additional input provided, amount of data flushed is necessarily <= ZSTD_BLOCKSIZE_MAX.
+* @return : 0 when a frame is completely decoded and fully flushed,
+*        or an error code, which can be tested using ZSTD_isError(),
+*        or any other value > 0, which means there is still some decoding or flushing to do to complete current frame :
+*                                the return value is a suggested next input size (just a hint for better latency)
+*                                that will never request more than the remaining frame size.
+* *******************************************************************************/
+
+typedef ZSTD_DCtx ZSTD_DStream;  /*< DCtx and DStream are now effectively same object (>= v1.3.0) */
+                                 /* For compatibility with versions <= v1.2.0, prefer differentiating them. */
+/*===== ZSTD_DStream management functions =====*/
+ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream(void);
+ZSTDLIB_API size_t ZSTD_freeDStream(ZSTD_DStream* zds);  /* accept NULL pointer */
+
+/*===== Streaming decompression functions =====*/
+
+/* This function is redundant with the advanced API and equivalent to:
+ *
+ *     ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
+ *     ZSTD_DCtx_refDDict(zds, NULL);
+ */
+ZSTDLIB_API size_t ZSTD_initDStream(ZSTD_DStream* zds);
+
+ZSTDLIB_API size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
+
+ZSTDLIB_API size_t ZSTD_DStreamInSize(void);    /*!< recommended size for input buffer */
+ZSTDLIB_API size_t ZSTD_DStreamOutSize(void);   /*!< recommended size for output buffer. Guarantee to successfully flush at least one complete block in all circumstances. */
+
+
+/* ************************
+*  Simple dictionary API
+***************************/
+/*! ZSTD_compress_usingDict() :
+ *  Compression at an explicit compression level using a Dictionary.
+ *  A dictionary can be any arbitrary data segment (also called a prefix),
+ *  or a buffer with specified information (see dictBuilder/zdict.h).
+ *  Note : This function loads the dictionary, resulting in significant startup delay.
+ *         It's intended for a dictionary used only once.
+ *  Note 2 : When `dict == NULL || dictSize < 8` no dictionary is used. */
+ZSTDLIB_API size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx,
+                                           void* dst, size_t dstCapacity,
+                                     const void* src, size_t srcSize,
+                                     const void* dict,size_t dictSize,
+                                           int compressionLevel);
+
+/*! ZSTD_decompress_usingDict() :
+ *  Decompression using a known Dictionary.
+ *  Dictionary must be identical to the one used during compression.
+ *  Note : This function loads the dictionary, resulting in significant startup delay.
+ *         It's intended for a dictionary used only once.
+ *  Note : When `dict == NULL || dictSize < 8` no dictionary is used. */
+ZSTDLIB_API size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,
+                                             void* dst, size_t dstCapacity,
+                                       const void* src, size_t srcSize,
+                                       const void* dict,size_t dictSize);
+
+
+/* *********************************
+ *  Bulk processing dictionary API
+ **********************************/
+typedef struct ZSTD_CDict_s ZSTD_CDict;
+
+/*! ZSTD_createCDict() :
+ *  When compressing multiple messages or blocks using the same dictionary,
+ *  it's recommended to digest the dictionary only once, since it's a costly operation.
+ *  ZSTD_createCDict() will create a state from digesting a dictionary.
+ *  The resulting state can be used for future compression operations with very limited startup cost.
+ *  ZSTD_CDict can be created once and shared by multiple threads concurrently, since its usage is read-only.
+ * @dictBuffer can be released after ZSTD_CDict creation, because its content is copied within CDict.
+ *  Note 1 : Consider experimental function `ZSTD_createCDict_byReference()` if you prefer to not duplicate @dictBuffer content.
+ *  Note 2 : A ZSTD_CDict can be created from an empty @dictBuffer,
+ *      in which case the only thing that it transports is the @compressionLevel.
+ *      This can be useful in a pipeline featuring ZSTD_compress_usingCDict() exclusively,
+ *      expecting a ZSTD_CDict parameter with any data, including those without a known dictionary. */
+ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict(const void* dictBuffer, size_t dictSize,
+                                         int compressionLevel);
+
+/*! ZSTD_freeCDict() :
+ *  Function frees memory allocated by ZSTD_createCDict().
+ *  If a NULL pointer is passed, no operation is performed. */
+ZSTDLIB_API size_t      ZSTD_freeCDict(ZSTD_CDict* CDict);
+
+/*! ZSTD_compress_usingCDict() :
+ *  Compression using a digested Dictionary.
+ *  Recommended when same dictionary is used multiple times.
+ *  Note : compression level is _decided at dictionary creation time_,
+ *     and frame parameters are hardcoded (dictID=yes, contentSize=yes, checksum=no) */
+ZSTDLIB_API size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,
+                                            void* dst, size_t dstCapacity,
+                                      const void* src, size_t srcSize,
+                                      const ZSTD_CDict* cdict);
+
+
+typedef struct ZSTD_DDict_s ZSTD_DDict;
+
+/*! ZSTD_createDDict() :
+ *  Create a digested dictionary, ready to start decompression operation without startup delay.
+ *  dictBuffer can be released after DDict creation, as its content is copied inside DDict. */
+ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict(const void* dictBuffer, size_t dictSize);
+
+/*! ZSTD_freeDDict() :
+ *  Function frees memory allocated with ZSTD_createDDict()
+ *  If a NULL pointer is passed, no operation is performed. */
+ZSTDLIB_API size_t      ZSTD_freeDDict(ZSTD_DDict* ddict);
+
+/*! ZSTD_decompress_usingDDict() :
+ *  Decompression using a digested Dictionary.
+ *  Recommended when same dictionary is used multiple times. */
+ZSTDLIB_API size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,
+                                              void* dst, size_t dstCapacity,
+                                        const void* src, size_t srcSize,
+                                        const ZSTD_DDict* ddict);
+
+
+/* ******************************
+ *  Dictionary helper functions
+ *******************************/
+
+/*! ZSTD_getDictID_fromDict() :
+ *  Provides the dictID stored within dictionary.
+ *  if @return == 0, the dictionary is not conformant with Zstandard specification.
+ *  It can still be loaded, but as a content-only dictionary. */
+ZSTDLIB_API unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize);
+
+/*! ZSTD_getDictID_fromDDict() :
+ *  Provides the dictID of the dictionary loaded into `ddict`.
+ *  If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
+ *  Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */
+ZSTDLIB_API unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict);
+
+/*! ZSTD_getDictID_fromFrame() :
+ *  Provides the dictID required to decompressed the frame stored within `src`.
+ *  If @return == 0, the dictID could not be decoded.
+ *  This could for one of the following reasons :
+ *  - The frame does not require a dictionary to be decoded (most common case).
+ *  - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information.
+ *    Note : this use case also happens when using a non-conformant dictionary.
+ *  - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`).
+ *  - This is not a Zstandard frame.
+ *  When identifying the exact failure cause, it's possible to use ZSTD_getFrameHeader(), which will provide a more precise error code. */
+ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);
+
+
+/* *****************************************************************************
+ * Advanced dictionary and prefix API
+ *
+ * This API allows dictionaries to be used with ZSTD_compress2(),
+ * ZSTD_compressStream2(), and ZSTD_decompress(). Dictionaries are sticky, and
+ * only reset with the context is reset with ZSTD_reset_parameters or
+ * ZSTD_reset_session_and_parameters. Prefixes are single-use.
+ ******************************************************************************/
+
+
+/*! ZSTD_CCtx_loadDictionary() :
+ *  Create an internal CDict from `dict` buffer.
+ *  Decompression will have to use same dictionary.
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ *  Special: Loading a NULL (or 0-size) dictionary invalidates previous dictionary,
+ *           meaning "return to no-dictionary mode".
+ *  Note 1 : Dictionary is sticky, it will be used for all future compressed frames.
+ *           To return to "no-dictionary" situation, load a NULL dictionary (or reset parameters).
+ *  Note 2 : Loading a dictionary involves building tables.
+ *           It's also a CPU consuming operation, with non-negligible impact on latency.
+ *           Tables are dependent on compression parameters, and for this reason,
+ *           compression parameters can no longer be changed after loading a dictionary.
+ *  Note 3 :`dict` content will be copied internally.
+ *           Use experimental ZSTD_CCtx_loadDictionary_byReference() to reference content instead.
+ *           In such a case, dictionary buffer must outlive its users.
+ *  Note 4 : Use ZSTD_CCtx_loadDictionary_advanced()
+ *           to precisely select how dictionary content must be interpreted. */
+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
+
+/*! ZSTD_CCtx_refCDict() :
+ *  Reference a prepared dictionary, to be used for all next compressed frames.
+ *  Note that compression parameters are enforced from within CDict,
+ *  and supersede any compression parameter previously set within CCtx.
+ *  The parameters ignored are labelled as "superseded-by-cdict" in the ZSTD_cParameter enum docs.
+ *  The ignored parameters will be used again if the CCtx is returned to no-dictionary mode.
+ *  The dictionary will remain valid for future compressed frames using same CCtx.
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ *  Special : Referencing a NULL CDict means "return to no-dictionary mode".
+ *  Note 1 : Currently, only one dictionary can be managed.
+ *           Referencing a new dictionary effectively "discards" any previous one.
+ *  Note 2 : CDict is just referenced, its lifetime must outlive its usage within CCtx. */
+ZSTDLIB_API size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict);
+
+/*! ZSTD_CCtx_refPrefix() :
+ *  Reference a prefix (single-usage dictionary) for next compressed frame.
+ *  A prefix is **only used once**. Tables are discarded at end of frame (ZSTD_e_end).
+ *  Decompression will need same prefix to properly regenerate data.
+ *  Compressing with a prefix is similar in outcome as performing a diff and compressing it,
+ *  but performs much faster, especially during decompression (compression speed is tunable with compression level).
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ *  Special: Adding any prefix (including NULL) invalidates any previous prefix or dictionary
+ *  Note 1 : Prefix buffer is referenced. It **must** outlive compression.
+ *           Its content must remain unmodified during compression.
+ *  Note 2 : If the intention is to diff some large src data blob with some prior version of itself,
+ *           ensure that the window size is large enough to contain the entire source.
+ *           See ZSTD_c_windowLog.
+ *  Note 3 : Referencing a prefix involves building tables, which are dependent on compression parameters.
+ *           It's a CPU consuming operation, with non-negligible impact on latency.
+ *           If there is a need to use the same prefix multiple times, consider loadDictionary instead.
+ *  Note 4 : By default, the prefix is interpreted as raw content (ZSTD_dct_rawContent).
+ *           Use experimental ZSTD_CCtx_refPrefix_advanced() to alter dictionary interpretation. */
+ZSTDLIB_API size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx,
+                                 const void* prefix, size_t prefixSize);
+
+/*! ZSTD_DCtx_loadDictionary() :
+ *  Create an internal DDict from dict buffer,
+ *  to be used to decompress next frames.
+ *  The dictionary remains valid for all future frames, until explicitly invalidated.
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ *  Special : Adding a NULL (or 0-size) dictionary invalidates any previous dictionary,
+ *            meaning "return to no-dictionary mode".
+ *  Note 1 : Loading a dictionary involves building tables,
+ *           which has a non-negligible impact on CPU usage and latency.
+ *           It's recommended to "load once, use many times", to amortize the cost
+ *  Note 2 :`dict` content will be copied internally, so `dict` can be released after loading.
+ *           Use ZSTD_DCtx_loadDictionary_byReference() to reference dictionary content instead.
+ *  Note 3 : Use ZSTD_DCtx_loadDictionary_advanced() to take control of
+ *           how dictionary content is loaded and interpreted.
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);
+
+/*! ZSTD_DCtx_refDDict() :
+ *  Reference a prepared dictionary, to be used to decompress next frames.
+ *  The dictionary remains active for decompression of future frames using same DCtx.
+ *
+ *  If called with ZSTD_d_refMultipleDDicts enabled, repeated calls of this function
+ *  will store the DDict references in a table, and the DDict used for decompression
+ *  will be determined at decompression time, as per the dict ID in the frame.
+ *  The memory for the table is allocated on the first call to refDDict, and can be
+ *  freed with ZSTD_freeDCtx().
+ *
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ *  Note 1 : Currently, only one dictionary can be managed.
+ *           Referencing a new dictionary effectively "discards" any previous one.
+ *  Special: referencing a NULL DDict means "return to no-dictionary mode".
+ *  Note 2 : DDict is just referenced, its lifetime must outlive its usage from DCtx.
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);
+
+/*! ZSTD_DCtx_refPrefix() :
+ *  Reference a prefix (single-usage dictionary) to decompress next frame.
+ *  This is the reverse operation of ZSTD_CCtx_refPrefix(),
+ *  and must use the same prefix as the one used during compression.
+ *  Prefix is **only used once**. Reference is discarded at end of frame.
+ *  End of frame is reached when ZSTD_decompressStream() returns 0.
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ *  Note 1 : Adding any prefix (including NULL) invalidates any previously set prefix or dictionary
+ *  Note 2 : Prefix buffer is referenced. It **must** outlive decompression.
+ *           Prefix buffer must remain unmodified up to the end of frame,
+ *           reached when ZSTD_decompressStream() returns 0.
+ *  Note 3 : By default, the prefix is treated as raw content (ZSTD_dct_rawContent).
+ *           Use ZSTD_CCtx_refPrefix_advanced() to alter dictMode (Experimental section)
+ *  Note 4 : Referencing a raw content prefix has almost no cpu nor memory cost.
+ *           A full dictionary is more costly, as it requires building tables.
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx,
+                                 const void* prefix, size_t prefixSize);
+
+/* ===   Memory management   === */
+
+/*! ZSTD_sizeof_*() :
+ *  These functions give the _current_ memory usage of selected object.
+ *  Note that object memory usage can evolve (increase or decrease) over time. */
+ZSTDLIB_API size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx);
+ZSTDLIB_API size_t ZSTD_sizeof_DCtx(const ZSTD_DCtx* dctx);
+ZSTDLIB_API size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs);
+ZSTDLIB_API size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds);
+ZSTDLIB_API size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict);
+ZSTDLIB_API size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);
+
+#endif  /* ZSTD_H_235446 */
+
+
+/* **************************************************************************************
+ *   ADVANCED AND EXPERIMENTAL FUNCTIONS
+ ****************************************************************************************
+ * The definitions in the following section are considered experimental.
+ * They are provided for advanced scenarios.
+ * They should never be used with a dynamic library, as prototypes may change in the future.
+ * Use them only in association with static linking.
+ * ***************************************************************************************/
+
+#if !defined(ZSTD_H_ZSTD_STATIC_LINKING_ONLY)
+#define ZSTD_H_ZSTD_STATIC_LINKING_ONLY
+
+/* **************************************************************************************
+ *   experimental API (static linking only)
+ ****************************************************************************************
+ * The following symbols and constants
+ * are not planned to join "stable API" status in the near future.
+ * They can still change in future versions.
+ * Some of them are planned to remain in the static_only section indefinitely.
+ * Some of them might be removed in the future (especially when redundant with existing stable functions)
+ * ***************************************************************************************/
+
+#define ZSTD_FRAMEHEADERSIZE_PREFIX(format) ((format) == ZSTD_f_zstd1 ? 5 : 1)   /* minimum input size required to query frame header size */
+#define ZSTD_FRAMEHEADERSIZE_MIN(format)    ((format) == ZSTD_f_zstd1 ? 6 : 2)
+#define ZSTD_FRAMEHEADERSIZE_MAX   18   /* can be useful for static allocation */
+#define ZSTD_SKIPPABLEHEADERSIZE    8
+
+/* compression parameter bounds */
+#define ZSTD_WINDOWLOG_MAX_32    30
+#define ZSTD_WINDOWLOG_MAX_64    31
+#define ZSTD_WINDOWLOG_MAX     ((int)(sizeof(size_t) == 4 ? ZSTD_WINDOWLOG_MAX_32 : ZSTD_WINDOWLOG_MAX_64))
+#define ZSTD_WINDOWLOG_MIN       10
+#define ZSTD_HASHLOG_MAX       ((ZSTD_WINDOWLOG_MAX < 30) ? ZSTD_WINDOWLOG_MAX : 30)
+#define ZSTD_HASHLOG_MIN          6
+#define ZSTD_CHAINLOG_MAX_32     29
+#define ZSTD_CHAINLOG_MAX_64     30
+#define ZSTD_CHAINLOG_MAX      ((int)(sizeof(size_t) == 4 ? ZSTD_CHAINLOG_MAX_32 : ZSTD_CHAINLOG_MAX_64))
+#define ZSTD_CHAINLOG_MIN        ZSTD_HASHLOG_MIN
+#define ZSTD_SEARCHLOG_MAX      (ZSTD_WINDOWLOG_MAX-1)
+#define ZSTD_SEARCHLOG_MIN        1
+#define ZSTD_MINMATCH_MAX         7   /* only for ZSTD_fast, other strategies are limited to 6 */
+#define ZSTD_MINMATCH_MIN         3   /* only for ZSTD_btopt+, faster strategies are limited to 4 */
+#define ZSTD_TARGETLENGTH_MAX    ZSTD_BLOCKSIZE_MAX
+#define ZSTD_TARGETLENGTH_MIN     0   /* note : comparing this constant to an unsigned results in a tautological test */
+#define ZSTD_STRATEGY_MIN        ZSTD_fast
+#define ZSTD_STRATEGY_MAX        ZSTD_btultra2
+
+
+#define ZSTD_OVERLAPLOG_MIN       0
+#define ZSTD_OVERLAPLOG_MAX       9
+
+#define ZSTD_WINDOWLOG_LIMIT_DEFAULT 27   /* by default, the streaming decoder will refuse any frame
+                                           * requiring larger than (1<<ZSTD_WINDOWLOG_LIMIT_DEFAULT) window size,
+                                           * to preserve host's memory from unreasonable requirements.
+                                           * This limit can be overridden using ZSTD_DCtx_setParameter(,ZSTD_d_windowLogMax,).
+                                           * The limit does not apply for one-pass decoders (such as ZSTD_decompress()), since no additional memory is allocated */
+
+
+/* LDM parameter bounds */
+#define ZSTD_LDM_HASHLOG_MIN      ZSTD_HASHLOG_MIN
+#define ZSTD_LDM_HASHLOG_MAX      ZSTD_HASHLOG_MAX
+#define ZSTD_LDM_MINMATCH_MIN        4
+#define ZSTD_LDM_MINMATCH_MAX     4096
+#define ZSTD_LDM_BUCKETSIZELOG_MIN   1
+#define ZSTD_LDM_BUCKETSIZELOG_MAX   8
+#define ZSTD_LDM_HASHRATELOG_MIN     0
+#define ZSTD_LDM_HASHRATELOG_MAX (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN)
+
+/* Advanced parameter bounds */
+#define ZSTD_TARGETCBLOCKSIZE_MIN   64
+#define ZSTD_TARGETCBLOCKSIZE_MAX   ZSTD_BLOCKSIZE_MAX
+#define ZSTD_SRCSIZEHINT_MIN        0
+#define ZSTD_SRCSIZEHINT_MAX        INT_MAX
+
+/* internal */
+#define ZSTD_HASHLOG3_MAX           17
+
+
+/* ---  Advanced types  --- */
+
+typedef struct ZSTD_CCtx_params_s ZSTD_CCtx_params;
+
+typedef struct {
+    unsigned int offset;      /* The offset of the match. (NOT the same as the offset code)
+                               * If offset == 0 and matchLength == 0, this sequence represents the last
+                               * literals in the block of litLength size.
+                               */
+
+    unsigned int litLength;   /* Literal length of the sequence. */
+    unsigned int matchLength; /* Match length of the sequence. */
+
+                              /* Note: Users of this API may provide a sequence with matchLength == litLength == offset == 0.
+                               * In this case, we will treat the sequence as a marker for a block boundary.
+                               */
+
+    unsigned int rep;         /* Represents which repeat offset is represented by the field 'offset'.
+                               * Ranges from [0, 3].
+                               *
+                               * Repeat offsets are essentially previous offsets from previous sequences sorted in
+                               * recency order. For more detail, see doc/zstd_compression_format.md
+                               *
+                               * If rep == 0, then 'offset' does not contain a repeat offset.
+                               * If rep > 0:
+                               *  If litLength != 0:
+                               *      rep == 1 --> offset == repeat_offset_1
+                               *      rep == 2 --> offset == repeat_offset_2
+                               *      rep == 3 --> offset == repeat_offset_3
+                               *  If litLength == 0:
+                               *      rep == 1 --> offset == repeat_offset_2
+                               *      rep == 2 --> offset == repeat_offset_3
+                               *      rep == 3 --> offset == repeat_offset_1 - 1
+                               *
+                               * Note: This field is optional. ZSTD_generateSequences() will calculate the value of
+                               * 'rep', but repeat offsets do not necessarily need to be calculated from an external
+                               * sequence provider's perspective. For example, ZSTD_compressSequences() does not
+                               * use this 'rep' field at all (as of now).
+                               */
+} ZSTD_Sequence;
+
+typedef struct {
+    unsigned windowLog;       /*< largest match distance : larger == more compression, more memory needed during decompression */
+    unsigned chainLog;        /*< fully searched segment : larger == more compression, slower, more memory (useless for fast) */
+    unsigned hashLog;         /*< dispatch table : larger == faster, more memory */
+    unsigned searchLog;       /*< nb of searches : larger == more compression, slower */
+    unsigned minMatch;        /*< match length searched : larger == faster decompression, sometimes less compression */
+    unsigned targetLength;    /*< acceptable match size for optimal parser (only) : larger == more compression, slower */
+    ZSTD_strategy strategy;   /*< see ZSTD_strategy definition above */
+} ZSTD_compressionParameters;
+
+typedef struct {
+    int contentSizeFlag; /*< 1: content size will be in frame header (when known) */
+    int checksumFlag;    /*< 1: generate a 32-bits checksum using XXH64 algorithm at end of frame, for error detection */
+    int noDictIDFlag;    /*< 1: no dictID will be saved into frame header (dictID is only useful for dictionary compression) */
+} ZSTD_frameParameters;
+
+typedef struct {
+    ZSTD_compressionParameters cParams;
+    ZSTD_frameParameters fParams;
+} ZSTD_parameters;
+
+typedef enum {
+    ZSTD_dct_auto = 0,       /* dictionary is "full" when starting with ZSTD_MAGIC_DICTIONARY, otherwise it is "rawContent" */
+    ZSTD_dct_rawContent = 1, /* ensures dictionary is always loaded as rawContent, even if it starts with ZSTD_MAGIC_DICTIONARY */
+    ZSTD_dct_fullDict = 2    /* refuses to load a dictionary if it does not respect Zstandard's specification, starting with ZSTD_MAGIC_DICTIONARY */
+} ZSTD_dictContentType_e;
+
+typedef enum {
+    ZSTD_dlm_byCopy = 0,  /*< Copy dictionary content internally */
+    ZSTD_dlm_byRef = 1    /*< Reference dictionary content -- the dictionary buffer must outlive its users. */
+} ZSTD_dictLoadMethod_e;
+
+typedef enum {
+    ZSTD_f_zstd1 = 0,           /* zstd frame format, specified in zstd_compression_format.md (default) */
+    ZSTD_f_zstd1_magicless = 1  /* Variant of zstd frame format, without initial 4-bytes magic number.
+                                 * Useful to save 4 bytes per generated frame.
+                                 * Decoder cannot recognise automatically this format, requiring this instruction. */
+} ZSTD_format_e;
+
+typedef enum {
+    /* Note: this enum controls ZSTD_d_forceIgnoreChecksum */
+    ZSTD_d_validateChecksum = 0,
+    ZSTD_d_ignoreChecksum = 1
+} ZSTD_forceIgnoreChecksum_e;
+
+typedef enum {
+    /* Note: this enum controls ZSTD_d_refMultipleDDicts */
+    ZSTD_rmd_refSingleDDict = 0,
+    ZSTD_rmd_refMultipleDDicts = 1
+} ZSTD_refMultipleDDicts_e;
+
+typedef enum {
+    /* Note: this enum and the behavior it controls are effectively internal
+     * implementation details of the compressor. They are expected to continue
+     * to evolve and should be considered only in the context of extremely
+     * advanced performance tuning.
+     *
+     * Zstd currently supports the use of a CDict in three ways:
+     *
+     * - The contents of the CDict can be copied into the working context. This
+     *   means that the compression can search both the dictionary and input
+     *   while operating on a single set of internal tables. This makes
+     *   the compression faster per-byte of input. However, the initial copy of
+     *   the CDict's tables incurs a fixed cost at the beginning of the
+     *   compression. For small compressions (< 8 KB), that copy can dominate
+     *   the cost of the compression.
+     *
+     * - The CDict's tables can be used in-place. In this model, compression is
+     *   slower per input byte, because the compressor has to search two sets of
+     *   tables. However, this model incurs no start-up cost (as long as the
+     *   working context's tables can be reused). For small inputs, this can be
+     *   faster than copying the CDict's tables.
+     *
+     * - The CDict's tables are not used at all, and instead we use the working
+     *   context alone to reload the dictionary and use params based on the source
+     *   size. See ZSTD_compress_insertDictionary() and ZSTD_compress_usingDict().
+     *   This method is effective when the dictionary sizes are very small relative
+     *   to the input size, and the input size is fairly large to begin with.
+     *
+     * Zstd has a simple internal heuristic that selects which strategy to use
+     * at the beginning of a compression. However, if experimentation shows that
+     * Zstd is making poor choices, it is possible to override that choice with
+     * this enum.
+     */
+    ZSTD_dictDefaultAttach = 0, /* Use the default heuristic. */
+    ZSTD_dictForceAttach   = 1, /* Never copy the dictionary. */
+    ZSTD_dictForceCopy     = 2, /* Always copy the dictionary. */
+    ZSTD_dictForceLoad     = 3  /* Always reload the dictionary */
+} ZSTD_dictAttachPref_e;
+
+typedef enum {
+  ZSTD_lcm_auto = 0,          /*< Automatically determine the compression mode based on the compression level.
+                               *   Negative compression levels will be uncompressed, and positive compression
+                               *   levels will be compressed. */
+  ZSTD_lcm_huffman = 1,       /*< Always attempt Huffman compression. Uncompressed literals will still be
+                               *   emitted if Huffman compression is not profitable. */
+  ZSTD_lcm_uncompressed = 2   /*< Always emit uncompressed literals. */
+} ZSTD_literalCompressionMode_e;
+
+
+/* *************************************
+*  Frame size functions
+***************************************/
+
+/*! ZSTD_findDecompressedSize() :
+ *  `src` should point to the start of a series of ZSTD encoded and/or skippable frames
+ *  `srcSize` must be the _exact_ size of this series
+ *       (i.e. there should be a frame boundary at `src + srcSize`)
+ *  @return : - decompressed size of all data in all successive frames
+ *            - if the decompressed size cannot be determined: ZSTD_CONTENTSIZE_UNKNOWN
+ *            - if an error occurred: ZSTD_CONTENTSIZE_ERROR
+ *
+ *   note 1 : decompressed size is an optional field, that may not be present, especially in streaming mode.
+ *            When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size.
+ *            In which case, it's necessary to use streaming mode to decompress data.
+ *   note 2 : decompressed size is always present when compression is done with ZSTD_compress()
+ *   note 3 : decompressed size can be very large (64-bits value),
+ *            potentially larger than what local system can handle as a single memory segment.
+ *            In which case, it's necessary to use streaming mode to decompress data.
+ *   note 4 : If source is untrusted, decompressed size could be wrong or intentionally modified.
+ *            Always ensure result fits within application's authorized limits.
+ *            Each application can set its own limits.
+ *   note 5 : ZSTD_findDecompressedSize handles multiple frames, and so it must traverse the input to
+ *            read each contained frame header.  This is fast as most of the data is skipped,
+ *            however it does mean that all frame data must be present and valid. */
+ZSTDLIB_API unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize);
+
+/*! ZSTD_decompressBound() :
+ *  `src` should point to the start of a series of ZSTD encoded and/or skippable frames
+ *  `srcSize` must be the _exact_ size of this series
+ *       (i.e. there should be a frame boundary at `src + srcSize`)
+ *  @return : - upper-bound for the decompressed size of all data in all successive frames
+ *            - if an error occurred: ZSTD_CONTENTSIZE_ERROR
+ *
+ *  note 1  : an error can occur if `src` contains an invalid or incorrectly formatted frame.
+ *  note 2  : the upper-bound is exact when the decompressed size field is available in every ZSTD encoded frame of `src`.
+ *            in this case, `ZSTD_findDecompressedSize` and `ZSTD_decompressBound` return the same value.
+ *  note 3  : when the decompressed size field isn't available, the upper-bound for that frame is calculated by:
+ *              upper-bound = # blocks * min(128 KB, Window_Size)
+ */
+ZSTDLIB_API unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize);
+
+/*! ZSTD_frameHeaderSize() :
+ *  srcSize must be >= ZSTD_FRAMEHEADERSIZE_PREFIX.
+ * @return : size of the Frame Header,
+ *           or an error code (if srcSize is too small) */
+ZSTDLIB_API size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize);
+
+typedef enum {
+  ZSTD_sf_noBlockDelimiters = 0,         /* Representation of ZSTD_Sequence has no block delimiters, sequences only */
+  ZSTD_sf_explicitBlockDelimiters = 1    /* Representation of ZSTD_Sequence contains explicit block delimiters */
+} ZSTD_sequenceFormat_e;
+
+/*! ZSTD_generateSequences() :
+ * Generate sequences using ZSTD_compress2, given a source buffer.
+ *
+ * Each block will end with a dummy sequence
+ * with offset == 0, matchLength == 0, and litLength == length of last literals.
+ * litLength may be == 0, and if so, then the sequence of (of: 0 ml: 0 ll: 0)
+ * simply acts as a block delimiter.
+ *
+ * zc can be used to insert custom compression params.
+ * This function invokes ZSTD_compress2
+ *
+ * The output of this function can be fed into ZSTD_compressSequences() with CCtx
+ * setting of ZSTD_c_blockDelimiters as ZSTD_sf_explicitBlockDelimiters
+ * @return : number of sequences generated
+ */
+
+ZSTDLIB_API size_t ZSTD_generateSequences(ZSTD_CCtx* zc, ZSTD_Sequence* outSeqs,
+                                          size_t outSeqsSize, const void* src, size_t srcSize);
+
+/*! ZSTD_mergeBlockDelimiters() :
+ * Given an array of ZSTD_Sequence, remove all sequences that represent block delimiters/last literals
+ * by merging them into into the literals of the next sequence.
+ *
+ * As such, the final generated result has no explicit representation of block boundaries,
+ * and the final last literals segment is not represented in the sequences.
+ *
+ * The output of this function can be fed into ZSTD_compressSequences() with CCtx
+ * setting of ZSTD_c_blockDelimiters as ZSTD_sf_noBlockDelimiters
+ * @return : number of sequences left after merging
+ */
+ZSTDLIB_API size_t ZSTD_mergeBlockDelimiters(ZSTD_Sequence* sequences, size_t seqsSize);
+
+/*! ZSTD_compressSequences() :
+ * Compress an array of ZSTD_Sequence, generated from the original source buffer, into dst.
+ * If a dictionary is included, then the cctx should reference the dict. (see: ZSTD_CCtx_refCDict(), ZSTD_CCtx_loadDictionary(), etc.)
+ * The entire source is compressed into a single frame.
+ *
+ * The compression behavior changes based on cctx params. In particular:
+ *    If ZSTD_c_blockDelimiters == ZSTD_sf_noBlockDelimiters, the array of ZSTD_Sequence is expected to contain
+ *    no block delimiters (defined in ZSTD_Sequence). Block boundaries are roughly determined based on
+ *    the block size derived from the cctx, and sequences may be split. This is the default setting.
+ *
+ *    If ZSTD_c_blockDelimiters == ZSTD_sf_explicitBlockDelimiters, the array of ZSTD_Sequence is expected to contain
+ *    block delimiters (defined in ZSTD_Sequence). Behavior is undefined if no block delimiters are provided.
+ *
+ *    If ZSTD_c_validateSequences == 0, this function will blindly accept the sequences provided. Invalid sequences cause undefined
+ *    behavior. If ZSTD_c_validateSequences == 1, then if sequence is invalid (see doc/zstd_compression_format.md for
+ *    specifics regarding offset/matchlength requirements) then the function will bail out and return an error.
+ *
+ *    In addition to the two adjustable experimental params, there are other important cctx params.
+ *    - ZSTD_c_minMatch MUST be set as less than or equal to the smallest match generated by the match finder. It has a minimum value of ZSTD_MINMATCH_MIN.
+ *    - ZSTD_c_compressionLevel accordingly adjusts the strength of the entropy coder, as it would in typical compression.
+ *    - ZSTD_c_windowLog affects offset validation: this function will return an error at higher debug levels if a provided offset
+ *      is larger than what the spec allows for a given window log and dictionary (if present). See: doc/zstd_compression_format.md
+ *
+ * Note: Repcodes are, as of now, always re-calculated within this function, so ZSTD_Sequence::rep is unused.
+ * Note 2: Once we integrate ability to ingest repcodes, the explicit block delims mode must respect those repcodes exactly,
+ *         and cannot emit an RLE block that disagrees with the repcode history
+ * @return : final compressed size or a ZSTD error.
+ */
+ZSTDLIB_API size_t ZSTD_compressSequences(ZSTD_CCtx* const cctx, void* dst, size_t dstSize,
+                                  const ZSTD_Sequence* inSeqs, size_t inSeqsSize,
+                                  const void* src, size_t srcSize);
+
+
+/*! ZSTD_writeSkippableFrame() :
+ * Generates a zstd skippable frame containing data given by src, and writes it to dst buffer.
+ *
+ * Skippable frames begin with a a 4-byte magic number. There are 16 possible choices of magic number,
+ * ranging from ZSTD_MAGIC_SKIPPABLE_START to ZSTD_MAGIC_SKIPPABLE_START+15.
+ * As such, the parameter magicVariant controls the exact skippable frame magic number variant used, so
+ * the magic number used will be ZSTD_MAGIC_SKIPPABLE_START + magicVariant.
+ *
+ * Returns an error if destination buffer is not large enough, if the source size is not representable
+ * with a 4-byte unsigned int, or if the parameter magicVariant is greater than 15 (and therefore invalid).
+ *
+ * @return : number of bytes written or a ZSTD error.
+ */
+ZSTDLIB_API size_t ZSTD_writeSkippableFrame(void* dst, size_t dstCapacity,
+                                            const void* src, size_t srcSize, unsigned magicVariant);
+
+
+/* *************************************
+*  Memory management
+***************************************/
+
+/*! ZSTD_estimate*() :
+ *  These functions make it possible to estimate memory usage
+ *  of a future {D,C}Ctx, before its creation.
+ *
+ *  ZSTD_estimateCCtxSize() will provide a memory budget large enough
+ *  for any compression level up to selected one.
+ *  Note : Unlike ZSTD_estimateCStreamSize*(), this estimate
+ *         does not include space for a window buffer.
+ *         Therefore, the estimation is only guaranteed for single-shot compressions, not streaming.
+ *  The estimate will assume the input may be arbitrarily large,
+ *  which is the worst case.
+ *
+ *  When srcSize can be bound by a known and rather "small" value,
+ *  this fact can be used to provide a tighter estimation
+ *  because the CCtx compression context will need less memory.
+ *  This tighter estimation can be provided by more advanced functions
+ *  ZSTD_estimateCCtxSize_usingCParams(), which can be used in tandem with ZSTD_getCParams(),
+ *  and ZSTD_estimateCCtxSize_usingCCtxParams(), which can be used in tandem with ZSTD_CCtxParams_setParameter().
+ *  Both can be used to estimate memory using custom compression parameters and arbitrary srcSize limits.
+ *
+ *  Note 2 : only single-threaded compression is supported.
+ *  ZSTD_estimateCCtxSize_usingCCtxParams() will return an error code if ZSTD_c_nbWorkers is >= 1.
+ */
+ZSTDLIB_API size_t ZSTD_estimateCCtxSize(int compressionLevel);
+ZSTDLIB_API size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams);
+ZSTDLIB_API size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params);
+ZSTDLIB_API size_t ZSTD_estimateDCtxSize(void);
+
+/*! ZSTD_estimateCStreamSize() :
+ *  ZSTD_estimateCStreamSize() will provide a budget large enough for any compression level up to selected one.
+ *  It will also consider src size to be arbitrarily "large", which is worst case.
+ *  If srcSize is known to always be small, ZSTD_estimateCStreamSize_usingCParams() can provide a tighter estimation.
+ *  ZSTD_estimateCStreamSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
+ *  ZSTD_estimateCStreamSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParams_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1.
+ *  Note : CStream size estimation is only correct for single-threaded compression.
+ *  ZSTD_DStream memory budget depends on window Size.
+ *  This information can be passed manually, using ZSTD_estimateDStreamSize,
+ *  or deducted from a valid frame Header, using ZSTD_estimateDStreamSize_fromFrame();
+ *  Note : if streaming is init with function ZSTD_init?Stream_usingDict(),
+ *         an internal ?Dict will be created, which additional size is not estimated here.
+ *         In this case, get total size by adding ZSTD_estimate?DictSize */
+ZSTDLIB_API size_t ZSTD_estimateCStreamSize(int compressionLevel);
+ZSTDLIB_API size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams);
+ZSTDLIB_API size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params);
+ZSTDLIB_API size_t ZSTD_estimateDStreamSize(size_t windowSize);
+ZSTDLIB_API size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize);
+
+/*! ZSTD_estimate?DictSize() :
+ *  ZSTD_estimateCDictSize() will bet that src size is relatively "small", and content is copied, like ZSTD_createCDict().
+ *  ZSTD_estimateCDictSize_advanced() makes it possible to control compression parameters precisely, like ZSTD_createCDict_advanced().
+ *  Note : dictionaries created by reference (`ZSTD_dlm_byRef`) are logically smaller.
+ */
+ZSTDLIB_API size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel);
+ZSTDLIB_API size_t ZSTD_estimateCDictSize_advanced(size_t dictSize, ZSTD_compressionParameters cParams, ZSTD_dictLoadMethod_e dictLoadMethod);
+ZSTDLIB_API size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod);
+
+/*! ZSTD_initStatic*() :
+ *  Initialize an object using a pre-allocated fixed-size buffer.
+ *  workspace: The memory area to emplace the object into.
+ *             Provided pointer *must be 8-bytes aligned*.
+ *             Buffer must outlive object.
+ *  workspaceSize: Use ZSTD_estimate*Size() to determine
+ *                 how large workspace must be to support target scenario.
+ * @return : pointer to object (same address as workspace, just different type),
+ *           or NULL if error (size too small, incorrect alignment, etc.)
+ *  Note : zstd will never resize nor malloc() when using a static buffer.
+ *         If the object requires more memory than available,
+ *         zstd will just error out (typically ZSTD_error_memory_allocation).
+ *  Note 2 : there is no corresponding "free" function.
+ *           Since workspace is allocated externally, it must be freed externally too.
+ *  Note 3 : cParams : use ZSTD_getCParams() to convert a compression level
+ *           into its associated cParams.
+ *  Limitation 1 : currently not compatible with internal dictionary creation, triggered by
+ *                 ZSTD_CCtx_loadDictionary(), ZSTD_initCStream_usingDict() or ZSTD_initDStream_usingDict().
+ *  Limitation 2 : static cctx currently not compatible with multi-threading.
+ *  Limitation 3 : static dctx is incompatible with legacy support.
+ */
+ZSTDLIB_API ZSTD_CCtx*    ZSTD_initStaticCCtx(void* workspace, size_t workspaceSize);
+ZSTDLIB_API ZSTD_CStream* ZSTD_initStaticCStream(void* workspace, size_t workspaceSize);    /*< same as ZSTD_initStaticCCtx() */
+
+ZSTDLIB_API ZSTD_DCtx*    ZSTD_initStaticDCtx(void* workspace, size_t workspaceSize);
+ZSTDLIB_API ZSTD_DStream* ZSTD_initStaticDStream(void* workspace, size_t workspaceSize);    /*< same as ZSTD_initStaticDCtx() */
+
+ZSTDLIB_API const ZSTD_CDict* ZSTD_initStaticCDict(
+                                        void* workspace, size_t workspaceSize,
+                                        const void* dict, size_t dictSize,
+                                        ZSTD_dictLoadMethod_e dictLoadMethod,
+                                        ZSTD_dictContentType_e dictContentType,
+                                        ZSTD_compressionParameters cParams);
+
+ZSTDLIB_API const ZSTD_DDict* ZSTD_initStaticDDict(
+                                        void* workspace, size_t workspaceSize,
+                                        const void* dict, size_t dictSize,
+                                        ZSTD_dictLoadMethod_e dictLoadMethod,
+                                        ZSTD_dictContentType_e dictContentType);
+
+
+/*! Custom memory allocation :
+ *  These prototypes make it possible to pass your own allocation/free functions.
+ *  ZSTD_customMem is provided at creation time, using ZSTD_create*_advanced() variants listed below.
+ *  All allocation/free operations will be completed using these custom variants instead of regular <stdlib.h> ones.
+ */
+typedef void* (*ZSTD_allocFunction) (void* opaque, size_t size);
+typedef void  (*ZSTD_freeFunction) (void* opaque, void* address);
+typedef struct { ZSTD_allocFunction customAlloc; ZSTD_freeFunction customFree; void* opaque; } ZSTD_customMem;
+static
+__attribute__((__unused__))
+ZSTD_customMem const ZSTD_defaultCMem = { NULL, NULL, NULL };  /*< this constant defers to stdlib's functions */
+
+ZSTDLIB_API ZSTD_CCtx*    ZSTD_createCCtx_advanced(ZSTD_customMem customMem);
+ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream_advanced(ZSTD_customMem customMem);
+ZSTDLIB_API ZSTD_DCtx*    ZSTD_createDCtx_advanced(ZSTD_customMem customMem);
+ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream_advanced(ZSTD_customMem customMem);
+
+ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_advanced(const void* dict, size_t dictSize,
+                                                  ZSTD_dictLoadMethod_e dictLoadMethod,
+                                                  ZSTD_dictContentType_e dictContentType,
+                                                  ZSTD_compressionParameters cParams,
+                                                  ZSTD_customMem customMem);
+
+/* ! Thread pool :
+ * These prototypes make it possible to share a thread pool among multiple compression contexts.
+ * This can limit resources for applications with multiple threads where each one uses
+ * a threaded compression mode (via ZSTD_c_nbWorkers parameter).
+ * ZSTD_createThreadPool creates a new thread pool with a given number of threads.
+ * Note that the lifetime of such pool must exist while being used.
+ * ZSTD_CCtx_refThreadPool assigns a thread pool to a context (use NULL argument value
+ * to use an internal thread pool).
+ * ZSTD_freeThreadPool frees a thread pool, accepts NULL pointer.
+ */
+typedef struct POOL_ctx_s ZSTD_threadPool;
+ZSTDLIB_API ZSTD_threadPool* ZSTD_createThreadPool(size_t numThreads);
+ZSTDLIB_API void ZSTD_freeThreadPool (ZSTD_threadPool* pool);  /* accept NULL pointer */
+ZSTDLIB_API size_t ZSTD_CCtx_refThreadPool(ZSTD_CCtx* cctx, ZSTD_threadPool* pool);
+
+
+/*
+ * This API is temporary and is expected to change or disappear in the future!
+ */
+ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_advanced2(
+    const void* dict, size_t dictSize,
+    ZSTD_dictLoadMethod_e dictLoadMethod,
+    ZSTD_dictContentType_e dictContentType,
+    const ZSTD_CCtx_params* cctxParams,
+    ZSTD_customMem customMem);
+
+ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_advanced(
+    const void* dict, size_t dictSize,
+    ZSTD_dictLoadMethod_e dictLoadMethod,
+    ZSTD_dictContentType_e dictContentType,
+    ZSTD_customMem customMem);
+
+
+/* *************************************
+*  Advanced compression functions
+***************************************/
+
+/*! ZSTD_createCDict_byReference() :
+ *  Create a digested dictionary for compression
+ *  Dictionary content is just referenced, not duplicated.
+ *  As a consequence, `dictBuffer` **must** outlive CDict,
+ *  and its content must remain unmodified throughout the lifetime of CDict.
+ *  note: equivalent to ZSTD_createCDict_advanced(), with dictLoadMethod==ZSTD_dlm_byRef */
+ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_byReference(const void* dictBuffer, size_t dictSize, int compressionLevel);
+
+/*! ZSTD_getDictID_fromCDict() :
+ *  Provides the dictID of the dictionary loaded into `cdict`.
+ *  If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
+ *  Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */
+ZSTDLIB_API unsigned ZSTD_getDictID_fromCDict(const ZSTD_CDict* cdict);
+
+/*! ZSTD_getCParams() :
+ * @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize.
+ * `estimatedSrcSize` value is optional, select 0 if not known */
+ZSTDLIB_API ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
+
+/*! ZSTD_getParams() :
+ *  same as ZSTD_getCParams(), but @return a full `ZSTD_parameters` object instead of sub-component `ZSTD_compressionParameters`.
+ *  All fields of `ZSTD_frameParameters` are set to default : contentSize=1, checksum=0, noDictID=0 */
+ZSTDLIB_API ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
+
+/*! ZSTD_checkCParams() :
+ *  Ensure param values remain within authorized range.
+ * @return 0 on success, or an error code (can be checked with ZSTD_isError()) */
+ZSTDLIB_API size_t ZSTD_checkCParams(ZSTD_compressionParameters params);
+
+/*! ZSTD_adjustCParams() :
+ *  optimize params for a given `srcSize` and `dictSize`.
+ * `srcSize` can be unknown, in which case use ZSTD_CONTENTSIZE_UNKNOWN.
+ * `dictSize` must be `0` when there is no dictionary.
+ *  cPar can be invalid : all parameters will be clamped within valid range in the @return struct.
+ *  This function never fails (wide contract) */
+ZSTDLIB_API ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize);
+
+/*! ZSTD_compress_advanced() :
+ *  Note : this function is now DEPRECATED.
+ *         It can be replaced by ZSTD_compress2(), in combination with ZSTD_CCtx_setParameter() and other parameter setters.
+ *  This prototype will be marked as deprecated and generate compilation warning on reaching v1.5.x */
+ZSTDLIB_API size_t ZSTD_compress_advanced(ZSTD_CCtx* cctx,
+                                          void* dst, size_t dstCapacity,
+                                    const void* src, size_t srcSize,
+                                    const void* dict,size_t dictSize,
+                                          ZSTD_parameters params);
+
+/*! ZSTD_compress_usingCDict_advanced() :
+ *  Note : this function is now REDUNDANT.
+ *         It can be replaced by ZSTD_compress2(), in combination with ZSTD_CCtx_loadDictionary() and other parameter setters.
+ *  This prototype will be marked as deprecated and generate compilation warning in some future version */
+ZSTDLIB_API size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx,
+                                              void* dst, size_t dstCapacity,
+                                        const void* src, size_t srcSize,
+                                        const ZSTD_CDict* cdict,
+                                              ZSTD_frameParameters fParams);
+
+
+/*! ZSTD_CCtx_loadDictionary_byReference() :
+ *  Same as ZSTD_CCtx_loadDictionary(), but dictionary content is referenced, instead of being copied into CCtx.
+ *  It saves some memory, but also requires that `dict` outlives its usage within `cctx` */
+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_byReference(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
+
+/*! ZSTD_CCtx_loadDictionary_advanced() :
+ *  Same as ZSTD_CCtx_loadDictionary(), but gives finer control over
+ *  how to load the dictionary (by copy ? by reference ?)
+ *  and how to interpret it (automatic ? force raw mode ? full mode only ?) */
+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType);
+
+/*! ZSTD_CCtx_refPrefix_advanced() :
+ *  Same as ZSTD_CCtx_refPrefix(), but gives finer control over
+ *  how to interpret prefix content (automatic ? force raw mode (default) ? full mode only ?) */
+ZSTDLIB_API size_t ZSTD_CCtx_refPrefix_advanced(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType);
+
+/* ===   experimental parameters   === */
+/* these parameters can be used with ZSTD_setParameter()
+ * they are not guaranteed to remain supported in the future */
+
+ /* Enables rsyncable mode,
+  * which makes compressed files more rsync friendly
+  * by adding periodic synchronization points to the compressed data.
+  * The target average block size is ZSTD_c_jobSize / 2.
+  * It's possible to modify the job size to increase or decrease
+  * the granularity of the synchronization point.
+  * Once the jobSize is smaller than the window size,
+  * it will result in compression ratio degradation.
+  * NOTE 1: rsyncable mode only works when multithreading is enabled.
+  * NOTE 2: rsyncable performs poorly in combination with long range mode,
+  * since it will decrease the effectiveness of synchronization points,
+  * though mileage may vary.
+  * NOTE 3: Rsyncable mode limits maximum compression speed to ~400 MB/s.
+  * If the selected compression level is already running significantly slower,
+  * the overall speed won't be significantly impacted.
+  */
+ #define ZSTD_c_rsyncable ZSTD_c_experimentalParam1
+
+/* Select a compression format.
+ * The value must be of type ZSTD_format_e.
+ * See ZSTD_format_e enum definition for details */
+#define ZSTD_c_format ZSTD_c_experimentalParam2
+
+/* Force back-reference distances to remain < windowSize,
+ * even when referencing into Dictionary content (default:0) */
+#define ZSTD_c_forceMaxWindow ZSTD_c_experimentalParam3
+
+/* Controls whether the contents of a CDict
+ * are used in place, or copied into the working context.
+ * Accepts values from the ZSTD_dictAttachPref_e enum.
+ * See the comments on that enum for an explanation of the feature. */
+#define ZSTD_c_forceAttachDict ZSTD_c_experimentalParam4
+
+/* Controls how the literals are compressed (default is auto).
+ * The value must be of type ZSTD_literalCompressionMode_e.
+ * See ZSTD_literalCompressionMode_t enum definition for details.
+ */
+#define ZSTD_c_literalCompressionMode ZSTD_c_experimentalParam5
+
+/* Tries to fit compressed block size to be around targetCBlockSize.
+ * No target when targetCBlockSize == 0.
+ * There is no guarantee on compressed block size (default:0) */
+#define ZSTD_c_targetCBlockSize ZSTD_c_experimentalParam6
+
+/* User's best guess of source size.
+ * Hint is not valid when srcSizeHint == 0.
+ * There is no guarantee that hint is close to actual source size,
+ * but compression ratio may regress significantly if guess considerably underestimates */
+#define ZSTD_c_srcSizeHint ZSTD_c_experimentalParam7
+
+/* Controls whether the new and experimental "dedicated dictionary search
+ * structure" can be used. This feature is still rough around the edges, be
+ * prepared for surprising behavior!
+ *
+ * How to use it:
+ *
+ * When using a CDict, whether to use this feature or not is controlled at
+ * CDict creation, and it must be set in a CCtxParams set passed into that
+ * construction (via ZSTD_createCDict_advanced2()). A compression will then
+ * use the feature or not based on how the CDict was constructed; the value of
+ * this param, set in the CCtx, will have no effect.
+ *
+ * However, when a dictionary buffer is passed into a CCtx, such as via
+ * ZSTD_CCtx_loadDictionary(), this param can be set on the CCtx to control
+ * whether the CDict that is created internally can use the feature or not.
+ *
+ * What it does:
+ *
+ * Normally, the internal data structures of the CDict are analogous to what
+ * would be stored in a CCtx after compressing the contents of a dictionary.
+ * To an approximation, a compression using a dictionary can then use those
+ * data structures to simply continue what is effectively a streaming
+ * compression where the simulated compression of the dictionary left off.
+ * Which is to say, the search structures in the CDict are normally the same
+ * format as in the CCtx.
+ *
+ * It is possible to do better, since the CDict is not like a CCtx: the search
+ * structures are written once during CDict creation, and then are only read
+ * after that, while the search structures in the CCtx are both read and
+ * written as the compression goes along. This means we can choose a search
+ * structure for the dictionary that is read-optimized.
+ *
+ * This feature enables the use of that different structure.
+ *
+ * Note that some of the members of the ZSTD_compressionParameters struct have
+ * different semantics and constraints in the dedicated search structure. It is
+ * highly recommended that you simply set a compression level in the CCtxParams
+ * you pass into the CDict creation call, and avoid messing with the cParams
+ * directly.
+ *
+ * Effects:
+ *
+ * This will only have any effect when the selected ZSTD_strategy
+ * implementation supports this feature. Currently, that's limited to
+ * ZSTD_greedy, ZSTD_lazy, and ZSTD_lazy2.
+ *
+ * Note that this means that the CDict tables can no longer be copied into the
+ * CCtx, so the dict attachment mode ZSTD_dictForceCopy will no longer be
+ * useable. The dictionary can only be attached or reloaded.
+ *
+ * In general, you should expect compression to be faster--sometimes very much
+ * so--and CDict creation to be slightly slower. Eventually, we will probably
+ * make this mode the default.
+ */
+#define ZSTD_c_enableDedicatedDictSearch ZSTD_c_experimentalParam8
+
+/* ZSTD_c_stableInBuffer
+ * Experimental parameter.
+ * Default is 0 == disabled. Set to 1 to enable.
+ *
+ * Tells the compressor that the ZSTD_inBuffer will ALWAYS be the same
+ * between calls, except for the modifications that zstd makes to pos (the
+ * caller must not modify pos). This is checked by the compressor, and
+ * compression will fail if it ever changes. This means the only flush
+ * mode that makes sense is ZSTD_e_end, so zstd will error if ZSTD_e_end
+ * is not used. The data in the ZSTD_inBuffer in the range [src, src + pos)
+ * MUST not be modified during compression or you will get data corruption.
+ *
+ * When this flag is enabled zstd won't allocate an input window buffer,
+ * because the user guarantees it can reference the ZSTD_inBuffer until
+ * the frame is complete. But, it will still allocate an output buffer
+ * large enough to fit a block (see ZSTD_c_stableOutBuffer). This will also
+ * avoid the memcpy() from the input buffer to the input window buffer.
+ *
+ * NOTE: ZSTD_compressStream2() will error if ZSTD_e_end is not used.
+ * That means this flag cannot be used with ZSTD_compressStream().
+ *
+ * NOTE: So long as the ZSTD_inBuffer always points to valid memory, using
+ * this flag is ALWAYS memory safe, and will never access out-of-bounds
+ * memory. However, compression WILL fail if you violate the preconditions.
+ *
+ * WARNING: The data in the ZSTD_inBuffer in the range [dst, dst + pos) MUST
+ * not be modified during compression or you will get data corruption. This
+ * is because zstd needs to reference data in the ZSTD_inBuffer to find
+ * matches. Normally zstd maintains its own window buffer for this purpose,
+ * but passing this flag tells zstd to use the user provided buffer.
+ */
+#define ZSTD_c_stableInBuffer ZSTD_c_experimentalParam9
+
+/* ZSTD_c_stableOutBuffer
+ * Experimental parameter.
+ * Default is 0 == disabled. Set to 1 to enable.
+ *
+ * Tells he compressor that the ZSTD_outBuffer will not be resized between
+ * calls. Specifically: (out.size - out.pos) will never grow. This gives the
+ * compressor the freedom to say: If the compressed data doesn't fit in the
+ * output buffer then return ZSTD_error_dstSizeTooSmall. This allows us to
+ * always decompress directly into the output buffer, instead of decompressing
+ * into an internal buffer and copying to the output buffer.
+ *
+ * When this flag is enabled zstd won't allocate an output buffer, because
+ * it can write directly to the ZSTD_outBuffer. It will still allocate the
+ * input window buffer (see ZSTD_c_stableInBuffer).
+ *
+ * Zstd will check that (out.size - out.pos) never grows and return an error
+ * if it does. While not strictly necessary, this should prevent surprises.
+ */
+#define ZSTD_c_stableOutBuffer ZSTD_c_experimentalParam10
+
+/* ZSTD_c_blockDelimiters
+ * Default is 0 == ZSTD_sf_noBlockDelimiters.
+ *
+ * For use with sequence compression API: ZSTD_compressSequences().
+ *
+ * Designates whether or not the given array of ZSTD_Sequence contains block delimiters
+ * and last literals, which are defined as sequences with offset == 0 and matchLength == 0.
+ * See the definition of ZSTD_Sequence for more specifics.
+ */
+#define ZSTD_c_blockDelimiters ZSTD_c_experimentalParam11
+
+/* ZSTD_c_validateSequences
+ * Default is 0 == disabled. Set to 1 to enable sequence validation.
+ *
+ * For use with sequence compression API: ZSTD_compressSequences().
+ * Designates whether or not we validate sequences provided to ZSTD_compressSequences()
+ * during function execution.
+ *
+ * Without validation, providing a sequence that does not conform to the zstd spec will cause
+ * undefined behavior, and may produce a corrupted block.
+ *
+ * With validation enabled, a if sequence is invalid (see doc/zstd_compression_format.md for
+ * specifics regarding offset/matchlength requirements) then the function will bail out and
+ * return an error.
+ *
+ */
+#define ZSTD_c_validateSequences ZSTD_c_experimentalParam12
+
+/*! ZSTD_CCtx_getParameter() :
+ *  Get the requested compression parameter value, selected by enum ZSTD_cParameter,
+ *  and store it into int* value.
+ * @return : 0, or an error code (which can be tested with ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_CCtx_getParameter(const ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value);
+
+
+/*! ZSTD_CCtx_params :
+ *  Quick howto :
+ *  - ZSTD_createCCtxParams() : Create a ZSTD_CCtx_params structure
+ *  - ZSTD_CCtxParams_setParameter() : Push parameters one by one into
+ *                                     an existing ZSTD_CCtx_params structure.
+ *                                     This is similar to
+ *                                     ZSTD_CCtx_setParameter().
+ *  - ZSTD_CCtx_setParametersUsingCCtxParams() : Apply parameters to
+ *                                    an existing CCtx.
+ *                                    These parameters will be applied to
+ *                                    all subsequent frames.
+ *  - ZSTD_compressStream2() : Do compression using the CCtx.
+ *  - ZSTD_freeCCtxParams() : Free the memory, accept NULL pointer.
+ *
+ *  This can be used with ZSTD_estimateCCtxSize_advanced_usingCCtxParams()
+ *  for static allocation of CCtx for single-threaded compression.
+ */
+ZSTDLIB_API ZSTD_CCtx_params* ZSTD_createCCtxParams(void);
+ZSTDLIB_API size_t ZSTD_freeCCtxParams(ZSTD_CCtx_params* params);  /* accept NULL pointer */
+
+/*! ZSTD_CCtxParams_reset() :
+ *  Reset params to default values.
+ */
+ZSTDLIB_API size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_params* params);
+
+/*! ZSTD_CCtxParams_init() :
+ *  Initializes the compression parameters of cctxParams according to
+ *  compression level. All other parameters are reset to their default values.
+ */
+ZSTDLIB_API size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel);
+
+/*! ZSTD_CCtxParams_init_advanced() :
+ *  Initializes the compression and frame parameters of cctxParams according to
+ *  params. All other parameters are reset to their default values.
+ */
+ZSTDLIB_API size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params);
+
+/*! ZSTD_CCtxParams_setParameter() :
+ *  Similar to ZSTD_CCtx_setParameter.
+ *  Set one compression parameter, selected by enum ZSTD_cParameter.
+ *  Parameters must be applied to a ZSTD_CCtx using
+ *  ZSTD_CCtx_setParametersUsingCCtxParams().
+ * @result : a code representing success or failure (which can be tested with
+ *           ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value);
+
+/*! ZSTD_CCtxParams_getParameter() :
+ * Similar to ZSTD_CCtx_getParameter.
+ * Get the requested value of one compression parameter, selected by enum ZSTD_cParameter.
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_CCtxParams_getParameter(const ZSTD_CCtx_params* params, ZSTD_cParameter param, int* value);
+
+/*! ZSTD_CCtx_setParametersUsingCCtxParams() :
+ *  Apply a set of ZSTD_CCtx_params to the compression context.
+ *  This can be done even after compression is started,
+ *    if nbWorkers==0, this will have no impact until a new compression is started.
+ *    if nbWorkers>=1, new parameters will be picked up at next job,
+ *       with a few restrictions (windowLog, pledgedSrcSize, nbWorkers, jobSize, and overlapLog are not updated).
+ */
+ZSTDLIB_API size_t ZSTD_CCtx_setParametersUsingCCtxParams(
+        ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params);
+
+/*! ZSTD_compressStream2_simpleArgs() :
+ *  Same as ZSTD_compressStream2(),
+ *  but using only integral types as arguments.
+ *  This variant might be helpful for binders from dynamic languages
+ *  which have troubles handling structures containing memory pointers.
+ */
+ZSTDLIB_API size_t ZSTD_compressStream2_simpleArgs (
+                            ZSTD_CCtx* cctx,
+                            void* dst, size_t dstCapacity, size_t* dstPos,
+                      const void* src, size_t srcSize, size_t* srcPos,
+                            ZSTD_EndDirective endOp);
+
+
+/* *************************************
+*  Advanced decompression functions
+***************************************/
+
+/*! ZSTD_isFrame() :
+ *  Tells if the content of `buffer` starts with a valid Frame Identifier.
+ *  Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0.
+ *  Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled.
+ *  Note 3 : Skippable Frame Identifiers are considered valid. */
+ZSTDLIB_API unsigned ZSTD_isFrame(const void* buffer, size_t size);
+
+/*! ZSTD_createDDict_byReference() :
+ *  Create a digested dictionary, ready to start decompression operation without startup delay.
+ *  Dictionary content is referenced, and therefore stays in dictBuffer.
+ *  It is important that dictBuffer outlives DDict,
+ *  it must remain read accessible throughout the lifetime of DDict */
+ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize);
+
+/*! ZSTD_DCtx_loadDictionary_byReference() :
+ *  Same as ZSTD_DCtx_loadDictionary(),
+ *  but references `dict` content instead of copying it into `dctx`.
+ *  This saves memory if `dict` remains around.,
+ *  However, it's imperative that `dict` remains accessible (and unmodified) while being used, so it must outlive decompression. */
+ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);
+
+/*! ZSTD_DCtx_loadDictionary_advanced() :
+ *  Same as ZSTD_DCtx_loadDictionary(),
+ *  but gives direct control over
+ *  how to load the dictionary (by copy ? by reference ?)
+ *  and how to interpret it (automatic ? force raw mode ? full mode only ?). */
+ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType);
+
+/*! ZSTD_DCtx_refPrefix_advanced() :
+ *  Same as ZSTD_DCtx_refPrefix(), but gives finer control over
+ *  how to interpret prefix content (automatic ? force raw mode (default) ? full mode only ?) */
+ZSTDLIB_API size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType);
+
+/*! ZSTD_DCtx_setMaxWindowSize() :
+ *  Refuses allocating internal buffers for frames requiring a window size larger than provided limit.
+ *  This protects a decoder context from reserving too much memory for itself (potential attack scenario).
+ *  This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode.
+ *  By default, a decompression context accepts all window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT)
+ * @return : 0, or an error code (which can be tested using ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize);
+
+/*! ZSTD_DCtx_getParameter() :
+ *  Get the requested decompression parameter value, selected by enum ZSTD_dParameter,
+ *  and store it into int* value.
+ * @return : 0, or an error code (which can be tested with ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_getParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int* value);
+
+/* ZSTD_d_format
+ * experimental parameter,
+ * allowing selection between ZSTD_format_e input compression formats
+ */
+#define ZSTD_d_format ZSTD_d_experimentalParam1
+/* ZSTD_d_stableOutBuffer
+ * Experimental parameter.
+ * Default is 0 == disabled. Set to 1 to enable.
+ *
+ * Tells the decompressor that the ZSTD_outBuffer will ALWAYS be the same
+ * between calls, except for the modifications that zstd makes to pos (the
+ * caller must not modify pos). This is checked by the decompressor, and
+ * decompression will fail if it ever changes. Therefore the ZSTD_outBuffer
+ * MUST be large enough to fit the entire decompressed frame. This will be
+ * checked when the frame content size is known. The data in the ZSTD_outBuffer
+ * in the range [dst, dst + pos) MUST not be modified during decompression
+ * or you will get data corruption.
+ *
+ * When this flags is enabled zstd won't allocate an output buffer, because
+ * it can write directly to the ZSTD_outBuffer, but it will still allocate
+ * an input buffer large enough to fit any compressed block. This will also
+ * avoid the memcpy() from the internal output buffer to the ZSTD_outBuffer.
+ * If you need to avoid the input buffer allocation use the buffer-less
+ * streaming API.
+ *
+ * NOTE: So long as the ZSTD_outBuffer always points to valid memory, using
+ * this flag is ALWAYS memory safe, and will never access out-of-bounds
+ * memory. However, decompression WILL fail if you violate the preconditions.
+ *
+ * WARNING: The data in the ZSTD_outBuffer in the range [dst, dst + pos) MUST
+ * not be modified during decompression or you will get data corruption. This
+ * is because zstd needs to reference data in the ZSTD_outBuffer to regenerate
+ * matches. Normally zstd maintains its own buffer for this purpose, but passing
+ * this flag tells zstd to use the user provided buffer.
+ */
+#define ZSTD_d_stableOutBuffer ZSTD_d_experimentalParam2
+
+/* ZSTD_d_forceIgnoreChecksum
+ * Experimental parameter.
+ * Default is 0 == disabled. Set to 1 to enable
+ *
+ * Tells the decompressor to skip checksum validation during decompression, regardless
+ * of whether checksumming was specified during compression. This offers some
+ * slight performance benefits, and may be useful for debugging.
+ * Param has values of type ZSTD_forceIgnoreChecksum_e
+ */
+#define ZSTD_d_forceIgnoreChecksum ZSTD_d_experimentalParam3
+
+/* ZSTD_d_refMultipleDDicts
+ * Experimental parameter.
+ * Default is 0 == disabled. Set to 1 to enable
+ *
+ * If enabled and dctx is allocated on the heap, then additional memory will be allocated
+ * to store references to multiple ZSTD_DDict. That is, multiple calls of ZSTD_refDDict()
+ * using a given ZSTD_DCtx, rather than overwriting the previous DDict reference, will instead
+ * store all references. At decompression time, the appropriate dictID is selected
+ * from the set of DDicts based on the dictID in the frame.
+ *
+ * Usage is simply calling ZSTD_refDDict() on multiple dict buffers.
+ *
+ * Param has values of byte ZSTD_refMultipleDDicts_e
+ *
+ * WARNING: Enabling this parameter and calling ZSTD_DCtx_refDDict(), will trigger memory
+ * allocation for the hash table. ZSTD_freeDCtx() also frees this memory.
+ * Memory is allocated as per ZSTD_DCtx::customMem.
+ *
+ * Although this function allocates memory for the table, the user is still responsible for
+ * memory management of the underlying ZSTD_DDict* themselves.
+ */
+#define ZSTD_d_refMultipleDDicts ZSTD_d_experimentalParam4
+
+
+/*! ZSTD_DCtx_setFormat() :
+ *  Instruct the decoder context about what kind of data to decode next.
+ *  This instruction is mandatory to decode data without a fully-formed header,
+ *  such ZSTD_f_zstd1_magicless for example.
+ * @return : 0, or an error code (which can be tested using ZSTD_isError()). */
+ZSTDLIB_API size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format);
+
+/*! ZSTD_decompressStream_simpleArgs() :
+ *  Same as ZSTD_decompressStream(),
+ *  but using only integral types as arguments.
+ *  This can be helpful for binders from dynamic languages
+ *  which have troubles handling structures containing memory pointers.
+ */
+ZSTDLIB_API size_t ZSTD_decompressStream_simpleArgs (
+                            ZSTD_DCtx* dctx,
+                            void* dst, size_t dstCapacity, size_t* dstPos,
+                      const void* src, size_t srcSize, size_t* srcPos);
+
+
+/* ******************************************************************
+*  Advanced streaming functions
+*  Warning : most of these functions are now redundant with the Advanced API.
+*  Once Advanced API reaches "stable" status,
+*  redundant functions will be deprecated, and then at some point removed.
+********************************************************************/
+
+/*=====   Advanced Streaming compression functions  =====*/
+
+/*! ZSTD_initCStream_srcSize() :
+ * This function is deprecated, and equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any)
+ *     ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
+ *     ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
+ *
+ * pledgedSrcSize must be correct. If it is not known at init time, use
+ * ZSTD_CONTENTSIZE_UNKNOWN. Note that, for compatibility with older programs,
+ * "0" also disables frame content size field. It may be enabled in the future.
+ * Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
+ */
+ZSTDLIB_API size_t
+ZSTD_initCStream_srcSize(ZSTD_CStream* zcs,
+                         int compressionLevel,
+                         unsigned long long pledgedSrcSize);
+
+/*! ZSTD_initCStream_usingDict() :
+ * This function is deprecated, and is equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
+ *     ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);
+ *
+ * Creates of an internal CDict (incompatible with static CCtx), except if
+ * dict == NULL or dictSize < 8, in which case no dict is used.
+ * Note: dict is loaded with ZSTD_dct_auto (treated as a full zstd dictionary if
+ * it begins with ZSTD_MAGIC_DICTIONARY, else as raw content) and ZSTD_dlm_byCopy.
+ * Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
+ */
+ZSTDLIB_API size_t
+ZSTD_initCStream_usingDict(ZSTD_CStream* zcs,
+                     const void* dict, size_t dictSize,
+                           int compressionLevel);
+
+/*! ZSTD_initCStream_advanced() :
+ * This function is deprecated, and is approximately equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     // Pseudocode: Set each zstd parameter and leave the rest as-is.
+ *     for ((param, value) : params) {
+ *         ZSTD_CCtx_setParameter(zcs, param, value);
+ *     }
+ *     ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
+ *     ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);
+ *
+ * dict is loaded with ZSTD_dct_auto and ZSTD_dlm_byCopy.
+ * pledgedSrcSize must be correct.
+ * If srcSize is not known at init time, use value ZSTD_CONTENTSIZE_UNKNOWN.
+ * Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
+ */
+ZSTDLIB_API size_t
+ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
+                    const void* dict, size_t dictSize,
+                          ZSTD_parameters params,
+                          unsigned long long pledgedSrcSize);
+
+/*! ZSTD_initCStream_usingCDict() :
+ * This function is deprecated, and equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_refCDict(zcs, cdict);
+ *
+ * note : cdict will just be referenced, and must outlive compression session
+ * Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
+ */
+ZSTDLIB_API size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict);
+
+/*! ZSTD_initCStream_usingCDict_advanced() :
+ *   This function is DEPRECATED, and is approximately equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     // Pseudocode: Set each zstd frame parameter and leave the rest as-is.
+ *     for ((fParam, value) : fParams) {
+ *         ZSTD_CCtx_setParameter(zcs, fParam, value);
+ *     }
+ *     ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
+ *     ZSTD_CCtx_refCDict(zcs, cdict);
+ *
+ * same as ZSTD_initCStream_usingCDict(), with control over frame parameters.
+ * pledgedSrcSize must be correct. If srcSize is not known at init time, use
+ * value ZSTD_CONTENTSIZE_UNKNOWN.
+ * Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
+ */
+ZSTDLIB_API size_t
+ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs,
+                               const ZSTD_CDict* cdict,
+                                     ZSTD_frameParameters fParams,
+                                     unsigned long long pledgedSrcSize);
+
+/*! ZSTD_resetCStream() :
+ * This function is deprecated, and is equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
+ *
+ *  start a new frame, using same parameters from previous frame.
+ *  This is typically useful to skip dictionary loading stage, since it will re-use it in-place.
+ *  Note that zcs must be init at least once before using ZSTD_resetCStream().
+ *  If pledgedSrcSize is not known at reset time, use macro ZSTD_CONTENTSIZE_UNKNOWN.
+ *  If pledgedSrcSize > 0, its value must be correct, as it will be written in header, and controlled at the end.
+ *  For the time being, pledgedSrcSize==0 is interpreted as "srcSize unknown" for compatibility with older programs,
+ *  but it will change to mean "empty" in future version, so use macro ZSTD_CONTENTSIZE_UNKNOWN instead.
+ * @return : 0, or an error code (which can be tested using ZSTD_isError())
+ *  Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
+ */
+ZSTDLIB_API size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize);
+
+
+typedef struct {
+    unsigned long long ingested;   /* nb input bytes read and buffered */
+    unsigned long long consumed;   /* nb input bytes actually compressed */
+    unsigned long long produced;   /* nb of compressed bytes generated and buffered */
+    unsigned long long flushed;    /* nb of compressed bytes flushed : not provided; can be tracked from caller side */
+    unsigned currentJobID;         /* MT only : latest started job nb */
+    unsigned nbActiveWorkers;      /* MT only : nb of workers actively compressing at probe time */
+} ZSTD_frameProgression;
+
+/* ZSTD_getFrameProgression() :
+ * tells how much data has been ingested (read from input)
+ * consumed (input actually compressed) and produced (output) for current frame.
+ * Note : (ingested - consumed) is amount of input data buffered internally, not yet compressed.
+ * Aggregates progression inside active worker threads.
+ */
+ZSTDLIB_API ZSTD_frameProgression ZSTD_getFrameProgression(const ZSTD_CCtx* cctx);
+
+/*! ZSTD_toFlushNow() :
+ *  Tell how many bytes are ready to be flushed immediately.
+ *  Useful for multithreading scenarios (nbWorkers >= 1).
+ *  Probe the oldest active job, defined as oldest job not yet entirely flushed,
+ *  and check its output buffer.
+ * @return : amount of data stored in oldest job and ready to be flushed immediately.
+ *  if @return == 0, it means either :
+ *  + there is no active job (could be checked with ZSTD_frameProgression()), or
+ *  + oldest job is still actively compressing data,
+ *    but everything it has produced has also been flushed so far,
+ *    therefore flush speed is limited by production speed of oldest job
+ *    irrespective of the speed of concurrent (and newer) jobs.
+ */
+ZSTDLIB_API size_t ZSTD_toFlushNow(ZSTD_CCtx* cctx);
+
+
+/*=====   Advanced Streaming decompression functions  =====*/
+
+/*!
+ * This function is deprecated, and is equivalent to:
+ *
+ *     ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
+ *     ZSTD_DCtx_loadDictionary(zds, dict, dictSize);
+ *
+ * note: no dictionary will be used if dict == NULL or dictSize < 8
+ * Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
+ */
+ZSTDLIB_API size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize);
+
+/*!
+ * This function is deprecated, and is equivalent to:
+ *
+ *     ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
+ *     ZSTD_DCtx_refDDict(zds, ddict);
+ *
+ * note : ddict is referenced, it must outlive decompression session
+ * Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
+ */
+ZSTDLIB_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict);
+
+/*!
+ * This function is deprecated, and is equivalent to:
+ *
+ *     ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
+ *
+ * re-use decompression parameters from previous init; saves dictionary loading
+ * Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
+ */
+ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);
+
+
+/* *******************************************************************
+*  Buffer-less and synchronous inner streaming functions
+*
+*  This is an advanced API, giving full control over buffer management, for users which need direct control over memory.
+*  But it's also a complex one, with several restrictions, documented below.
+*  Prefer normal streaming API for an easier experience.
+********************************************************************* */
+
+/*
+  Buffer-less streaming compression (synchronous mode)
+
+  A ZSTD_CCtx object is required to track streaming operations.
+  Use ZSTD_createCCtx() / ZSTD_freeCCtx() to manage resource.
+  ZSTD_CCtx object can be re-used multiple times within successive compression operations.
+
+  Start by initializing a context.
+  Use ZSTD_compressBegin(), or ZSTD_compressBegin_usingDict() for dictionary compression,
+  or ZSTD_compressBegin_advanced(), for finer parameter control.
+  It's also possible to duplicate a reference context which has already been initialized, using ZSTD_copyCCtx()
+
+  Then, consume your input using ZSTD_compressContinue().
+  There are some important considerations to keep in mind when using this advanced function :
+  - ZSTD_compressContinue() has no internal buffer. It uses externally provided buffers only.
+  - Interface is synchronous : input is consumed entirely and produces 1+ compressed blocks.
+  - Caller must ensure there is enough space in `dst` to store compressed data under worst case scenario.
+    Worst case evaluation is provided by ZSTD_compressBound().
+    ZSTD_compressContinue() doesn't guarantee recover after a failed compression.
+  - ZSTD_compressContinue() presumes prior input ***is still accessible and unmodified*** (up to maximum distance size, see WindowLog).
+    It remembers all previous contiguous blocks, plus one separated memory segment (which can itself consists of multiple contiguous blocks)
+  - ZSTD_compressContinue() detects that prior input has been overwritten when `src` buffer overlaps.
+    In which case, it will "discard" the relevant memory section from its history.
+
+  Finish a frame with ZSTD_compressEnd(), which will write the last block(s) and optional checksum.
+  It's possible to use srcSize==0, in which case, it will write a final empty block to end the frame.
+  Without last block mark, frames are considered unfinished (hence corrupted) by compliant decoders.
+
+  `ZSTD_CCtx` object can be re-used (ZSTD_compressBegin()) to compress again.
+*/
+
+/*=====   Buffer-less streaming compression functions  =====*/
+ZSTDLIB_API size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel);
+ZSTDLIB_API size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel);
+ZSTDLIB_API size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_parameters params, unsigned long long pledgedSrcSize); /*< pledgedSrcSize : If srcSize is not known at init time, use ZSTD_CONTENTSIZE_UNKNOWN */
+ZSTDLIB_API size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict); /*< note: fails if cdict==NULL */
+ZSTDLIB_API size_t ZSTD_compressBegin_usingCDict_advanced(ZSTD_CCtx* const cctx, const ZSTD_CDict* const cdict, ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize);   /* compression parameters are already set within cdict. pledgedSrcSize must be correct. If srcSize is not known, use macro ZSTD_CONTENTSIZE_UNKNOWN */
+ZSTDLIB_API size_t ZSTD_copyCCtx(ZSTD_CCtx* cctx, const ZSTD_CCtx* preparedCCtx, unsigned long long pledgedSrcSize); /*<  note: if pledgedSrcSize is not known, use ZSTD_CONTENTSIZE_UNKNOWN */
+
+ZSTDLIB_API size_t ZSTD_compressContinue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+ZSTDLIB_API size_t ZSTD_compressEnd(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+
+
+/*
+  Buffer-less streaming decompression (synchronous mode)
+
+  A ZSTD_DCtx object is required to track streaming operations.
+  Use ZSTD_createDCtx() / ZSTD_freeDCtx() to manage it.
+  A ZSTD_DCtx object can be re-used multiple times.
+
+  First typical operation is to retrieve frame parameters, using ZSTD_getFrameHeader().
+  Frame header is extracted from the beginning of compressed frame, so providing only the frame's beginning is enough.
+  Data fragment must be large enough to ensure successful decoding.
+ `ZSTD_frameHeaderSize_max` bytes is guaranteed to always be large enough.
+  @result : 0 : successful decoding, the `ZSTD_frameHeader` structure is correctly filled.
+           >0 : `srcSize` is too small, please provide at least @result bytes on next attempt.
+           errorCode, which can be tested using ZSTD_isError().
+
+  It fills a ZSTD_frameHeader structure with important information to correctly decode the frame,
+  such as the dictionary ID, content size, or maximum back-reference distance (`windowSize`).
+  Note that these values could be wrong, either because of data corruption, or because a 3rd party deliberately spoofs false information.
+  As a consequence, check that values remain within valid application range.
+  For example, do not allocate memory blindly, check that `windowSize` is within expectation.
+  Each application can set its own limits, depending on local restrictions.
+  For extended interoperability, it is recommended to support `windowSize` of at least 8 MB.
+
+  ZSTD_decompressContinue() needs previous data blocks during decompression, up to `windowSize` bytes.
+  ZSTD_decompressContinue() is very sensitive to contiguity,
+  if 2 blocks don't follow each other, make sure that either the compressor breaks contiguity at the same place,
+  or that previous contiguous segment is large enough to properly handle maximum back-reference distance.
+  There are multiple ways to guarantee this condition.
+
+  The most memory efficient way is to use a round buffer of sufficient size.
+  Sufficient size is determined by invoking ZSTD_decodingBufferSize_min(),
+  which can @return an error code if required value is too large for current system (in 32-bits mode).
+  In a round buffer methodology, ZSTD_decompressContinue() decompresses each block next to previous one,
+  up to the moment there is not enough room left in the buffer to guarantee decoding another full block,
+  which maximum size is provided in `ZSTD_frameHeader` structure, field `blockSizeMax`.
+  At which point, decoding can resume from the beginning of the buffer.
+  Note that already decoded data stored in the buffer should be flushed before being overwritten.
+
+  There are alternatives possible, for example using two or more buffers of size `windowSize` each, though they consume more memory.
+
+  Finally, if you control the compression process, you can also ignore all buffer size rules,
+  as long as the encoder and decoder progress in "lock-step",
+  aka use exactly the same buffer sizes, break contiguity at the same place, etc.
+
+  Once buffers are setup, start decompression, with ZSTD_decompressBegin().
+  If decompression requires a dictionary, use ZSTD_decompressBegin_usingDict() or ZSTD_decompressBegin_usingDDict().
+
+  Then use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue() alternatively.
+  ZSTD_nextSrcSizeToDecompress() tells how many bytes to provide as 'srcSize' to ZSTD_decompressContinue().
+  ZSTD_decompressContinue() requires this _exact_ amount of bytes, or it will fail.
+
+ @result of ZSTD_decompressContinue() is the number of bytes regenerated within 'dst' (necessarily <= dstCapacity).
+  It can be zero : it just means ZSTD_decompressContinue() has decoded some metadata item.
+  It can also be an error code, which can be tested with ZSTD_isError().
+
+  A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero.
+  Context can then be reset to start a new decompression.
+
+  Note : it's possible to know if next input to present is a header or a block, using ZSTD_nextInputType().
+  This information is not required to properly decode a frame.
+
+  == Special case : skippable frames ==
+
+  Skippable frames allow integration of user-defined data into a flow of concatenated frames.
+  Skippable frames will be ignored (skipped) by decompressor.
+  The format of skippable frames is as follows :
+  a) Skippable frame ID - 4 Bytes, Little endian format, any value from 0x184D2A50 to 0x184D2A5F
+  b) Frame Size - 4 Bytes, Little endian format, unsigned 32-bits
+  c) Frame Content - any content (User Data) of length equal to Frame Size
+  For skippable frames ZSTD_getFrameHeader() returns zfhPtr->frameType==ZSTD_skippableFrame.
+  For skippable frames ZSTD_decompressContinue() always returns 0 : it only skips the content.
+*/
+
+/*=====   Buffer-less streaming decompression functions  =====*/
+typedef enum { ZSTD_frame, ZSTD_skippableFrame } ZSTD_frameType_e;
+typedef struct {
+    unsigned long long frameContentSize; /* if == ZSTD_CONTENTSIZE_UNKNOWN, it means this field is not available. 0 means "empty" */
+    unsigned long long windowSize;       /* can be very large, up to <= frameContentSize */
+    unsigned blockSizeMax;
+    ZSTD_frameType_e frameType;          /* if == ZSTD_skippableFrame, frameContentSize is the size of skippable content */
+    unsigned headerSize;
+    unsigned dictID;
+    unsigned checksumFlag;
+} ZSTD_frameHeader;
+
+/*! ZSTD_getFrameHeader() :
+ *  decode Frame Header, or requires larger `srcSize`.
+ * @return : 0, `zfhPtr` is correctly filled,
+ *          >0, `srcSize` is too small, value is wanted `srcSize` amount,
+ *           or an error code, which can be tested using ZSTD_isError() */
+ZSTDLIB_API size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize);   /*< doesn't consume input */
+/*! ZSTD_getFrameHeader_advanced() :
+ *  same as ZSTD_getFrameHeader(),
+ *  with added capability to select a format (like ZSTD_f_zstd1_magicless) */
+ZSTDLIB_API size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format);
+ZSTDLIB_API size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize);  /*< when frame content size is not known, pass in frameContentSize == ZSTD_CONTENTSIZE_UNKNOWN */
+
+ZSTDLIB_API size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx);
+ZSTDLIB_API size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);
+ZSTDLIB_API size_t ZSTD_decompressBegin_usingDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);
+
+ZSTDLIB_API size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx);
+ZSTDLIB_API size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+
+/* misc */
+ZSTDLIB_API void   ZSTD_copyDCtx(ZSTD_DCtx* dctx, const ZSTD_DCtx* preparedDCtx);
+typedef enum { ZSTDnit_frameHeader, ZSTDnit_blockHeader, ZSTDnit_block, ZSTDnit_lastBlock, ZSTDnit_checksum, ZSTDnit_skippableFrame } ZSTD_nextInputType_e;
+ZSTDLIB_API ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx);
+
+
+
+
+/* ============================ */
+/*       Block level API       */
+/* ============================ */
+
+/*!
+    Block functions produce and decode raw zstd blocks, without frame metadata.
+    Frame metadata cost is typically ~12 bytes, which can be non-negligible for very small blocks (< 100 bytes).
+    But users will have to take in charge needed metadata to regenerate data, such as compressed and content sizes.
+
+    A few rules to respect :
+    - Compressing and decompressing require a context structure
+      + Use ZSTD_createCCtx() and ZSTD_createDCtx()
+    - It is necessary to init context before starting
+      + compression : any ZSTD_compressBegin*() variant, including with dictionary
+      + decompression : any ZSTD_decompressBegin*() variant, including with dictionary
+      + copyCCtx() and copyDCtx() can be used too
+    - Block size is limited, it must be <= ZSTD_getBlockSize() <= ZSTD_BLOCKSIZE_MAX == 128 KB
+      + If input is larger than a block size, it's necessary to split input data into multiple blocks
+      + For inputs larger than a single block, consider using regular ZSTD_compress() instead.
+        Frame metadata is not that costly, and quickly becomes negligible as source size grows larger than a block.
+    - When a block is considered not compressible enough, ZSTD_compressBlock() result will be 0 (zero) !
+      ===> In which case, nothing is produced into `dst` !
+      + User __must__ test for such outcome and deal directly with uncompressed data
+      + A block cannot be declared incompressible if ZSTD_compressBlock() return value was != 0.
+        Doing so would mess up with statistics history, leading to potential data corruption.
+      + ZSTD_decompressBlock() _doesn't accept uncompressed data as input_ !!
+      + In case of multiple successive blocks, should some of them be uncompressed,
+        decoder must be informed of their existence in order to follow proper history.
+        Use ZSTD_insertBlock() for such a case.
+*/
+
+/*=====   Raw zstd block functions  =====*/
+ZSTDLIB_API size_t ZSTD_getBlockSize   (const ZSTD_CCtx* cctx);
+ZSTDLIB_API size_t ZSTD_compressBlock  (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+ZSTDLIB_API size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+ZSTDLIB_API size_t ZSTD_insertBlock    (ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize);  /*< insert uncompressed block into `dctx` history. Useful for multi-blocks decompression. */
+
+
+#endif   /* ZSTD_H_ZSTD_STATIC_LINKING_ONLY */
+
diff --git a/include/scsi/scsi_cmnd.h b/include/scsi/scsi_cmnd.h
index cd9eaa83..477a800 100644
--- a/include/scsi/scsi_cmnd.h
+++ b/include/scsi/scsi_cmnd.h
@@ -73,7 +73,7 @@ enum scsi_cmnd_submitter {
 struct scsi_cmnd {
 	struct scsi_request req;
 	struct scsi_device *device;
-	struct list_head eh_entry; /* entry for the host eh_cmd_q */
+	struct list_head eh_entry; /* entry for the host eh_abort_list/eh_cmd_q */
 	struct delayed_work abort_work;
 
 	struct rcu_head rcu;
diff --git a/include/scsi/scsi_device.h b/include/scsi/scsi_device.h
index 83a7890..d1c6fc8 100644
--- a/include/scsi/scsi_device.h
+++ b/include/scsi/scsi_device.h
@@ -226,12 +226,6 @@ struct scsi_device {
 
 	struct device		sdev_gendev,
 				sdev_dev;
-	/*
-	 * The array size 6 provides space for one attribute group for the
-	 * SCSI core, four attribute groups defined by SCSI LLDs and one
-	 * terminating NULL pointer.
-	 */
-	const struct attribute_group *gendev_attr_groups[6];
 
 	struct execute_work	ew; /* used to get process context on put */
 	struct work_struct	requeue_work;
diff --git a/include/scsi/scsi_host.h b/include/scsi/scsi_host.h
index ae71595..ebe059b 100644
--- a/include/scsi/scsi_host.h
+++ b/include/scsi/scsi_host.h
@@ -551,6 +551,7 @@ struct Scsi_Host {
 
 	struct mutex		scan_mutex;/* serialize scanning activity */
 
+	struct list_head	eh_abort_list;
 	struct list_head	eh_cmd_q;
 	struct task_struct    * ehandler;  /* Error recovery thread. */
 	struct completion     * eh_action; /* Wait for specific actions on the
diff --git a/include/sound/memalloc.h b/include/sound/memalloc.h
index 653dfff..1051b84 100644
--- a/include/sound/memalloc.h
+++ b/include/sound/memalloc.h
@@ -36,6 +36,13 @@ struct snd_dma_device {
 #define SNDRV_DMA_TYPE_CONTINUOUS	1	/* continuous no-DMA memory */
 #define SNDRV_DMA_TYPE_DEV		2	/* generic device continuous */
 #define SNDRV_DMA_TYPE_DEV_WC		5	/* continuous write-combined */
+#ifdef CONFIG_SND_DMA_SGBUF
+#define SNDRV_DMA_TYPE_DEV_SG		3	/* generic device SG-buffer */
+#define SNDRV_DMA_TYPE_DEV_WC_SG	6	/* SG write-combined */
+#else
+#define SNDRV_DMA_TYPE_DEV_SG	SNDRV_DMA_TYPE_DEV /* no SG-buf support */
+#define SNDRV_DMA_TYPE_DEV_WC_SG	SNDRV_DMA_TYPE_DEV_WC
+#endif
 #ifdef CONFIG_GENERIC_ALLOCATOR
 #define SNDRV_DMA_TYPE_DEV_IRAM		4	/* generic device iram-buffer */
 #else
@@ -44,13 +51,6 @@ struct snd_dma_device {
 #define SNDRV_DMA_TYPE_VMALLOC		7	/* vmalloc'ed buffer */
 #define SNDRV_DMA_TYPE_NONCONTIG	8	/* non-coherent SG buffer */
 #define SNDRV_DMA_TYPE_NONCOHERENT	9	/* non-coherent buffer */
-#ifdef CONFIG_SND_DMA_SGBUF
-#define SNDRV_DMA_TYPE_DEV_SG		SNDRV_DMA_TYPE_NONCONTIG
-#define SNDRV_DMA_TYPE_DEV_WC_SG	6	/* SG write-combined */
-#else
-#define SNDRV_DMA_TYPE_DEV_SG	SNDRV_DMA_TYPE_DEV /* no SG-buf support */
-#define SNDRV_DMA_TYPE_DEV_WC_SG	SNDRV_DMA_TYPE_DEV_WC
-#endif
 
 /*
  * info for buffer allocation
diff --git a/include/trace/events/afs.h b/include/trace/events/afs.h
index bca73e8..499f5fa 100644
--- a/include/trace/events/afs.h
+++ b/include/trace/events/afs.h
@@ -1016,31 +1016,32 @@ TRACE_EVENT(afs_dir_check_failed,
 		      __entry->vnode, __entry->off, __entry->i_size)
 	    );
 
-TRACE_EVENT(afs_page_dirty,
-	    TP_PROTO(struct afs_vnode *vnode, const char *where, struct page *page),
+TRACE_EVENT(afs_folio_dirty,
+	    TP_PROTO(struct afs_vnode *vnode, const char *where, struct folio *folio),
 
-	    TP_ARGS(vnode, where, page),
+	    TP_ARGS(vnode, where, folio),
 
 	    TP_STRUCT__entry(
 		    __field(struct afs_vnode *,		vnode		)
 		    __field(const char *,		where		)
-		    __field(pgoff_t,			page		)
+		    __field(pgoff_t,			index		)
 		    __field(unsigned long,		from		)
 		    __field(unsigned long,		to		)
 			     ),
 
 	    TP_fast_assign(
+		    unsigned long priv = (unsigned long)folio_get_private(folio);
 		    __entry->vnode = vnode;
 		    __entry->where = where;
-		    __entry->page = page->index;
-		    __entry->from = afs_page_dirty_from(page, page->private);
-		    __entry->to = afs_page_dirty_to(page, page->private);
-		    __entry->to |= (afs_is_page_dirty_mmapped(page->private) ?
-				    (1UL << (BITS_PER_LONG - 1)) : 0);
+		    __entry->index = folio_index(folio);
+		    __entry->from  = afs_folio_dirty_from(folio, priv);
+		    __entry->to    = afs_folio_dirty_to(folio, priv);
+		    __entry->to   |= (afs_is_folio_dirty_mmapped(priv) ?
+				      (1UL << (BITS_PER_LONG - 1)) : 0);
 			   ),
 
 	    TP_printk("vn=%p %lx %s %lx-%lx%s",
-		      __entry->vnode, __entry->page, __entry->where,
+		      __entry->vnode, __entry->index, __entry->where,
 		      __entry->from,
 		      __entry->to & ~(1UL << (BITS_PER_LONG - 1)),
 		      __entry->to & (1UL << (BITS_PER_LONG - 1)) ? " M" : "")
diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h
index 4e881d9..f8cb916 100644
--- a/include/trace/events/f2fs.h
+++ b/include/trace/events/f2fs.h
@@ -570,9 +570,10 @@ TRACE_EVENT(f2fs_file_write_iter,
 );
 
 TRACE_EVENT(f2fs_map_blocks,
-	TP_PROTO(struct inode *inode, struct f2fs_map_blocks *map, int ret),
+	TP_PROTO(struct inode *inode, struct f2fs_map_blocks *map,
+				int create, int flag, int ret),
 
-	TP_ARGS(inode, map, ret),
+	TP_ARGS(inode, map, create, flag, ret),
 
 	TP_STRUCT__entry(
 		__field(dev_t,	dev)
@@ -583,11 +584,14 @@ TRACE_EVENT(f2fs_map_blocks,
 		__field(unsigned int,	m_flags)
 		__field(int,	m_seg_type)
 		__field(bool,	m_may_create)
+		__field(bool,	m_multidev_dio)
+		__field(int,	create)
+		__field(int,	flag)
 		__field(int,	ret)
 	),
 
 	TP_fast_assign(
-		__entry->dev		= inode->i_sb->s_dev;
+		__entry->dev		= map->m_bdev->bd_dev;
 		__entry->ino		= inode->i_ino;
 		__entry->m_lblk		= map->m_lblk;
 		__entry->m_pblk		= map->m_pblk;
@@ -595,12 +599,16 @@ TRACE_EVENT(f2fs_map_blocks,
 		__entry->m_flags	= map->m_flags;
 		__entry->m_seg_type	= map->m_seg_type;
 		__entry->m_may_create	= map->m_may_create;
+		__entry->m_multidev_dio	= map->m_multidev_dio;
+		__entry->create		= create;
+		__entry->flag		= flag;
 		__entry->ret		= ret;
 	),
 
 	TP_printk("dev = (%d,%d), ino = %lu, file offset = %llu, "
-		"start blkaddr = 0x%llx, len = 0x%llx, flags = %u,"
-		"seg_type = %d, may_create = %d, err = %d",
+		"start blkaddr = 0x%llx, len = 0x%llx, flags = %u, "
+		"seg_type = %d, may_create = %d, multidevice = %d, "
+		"create = %d, flag = %d, err = %d",
 		show_dev_ino(__entry),
 		(unsigned long long)__entry->m_lblk,
 		(unsigned long long)__entry->m_pblk,
@@ -608,6 +616,9 @@ TRACE_EVENT(f2fs_map_blocks,
 		__entry->m_flags,
 		__entry->m_seg_type,
 		__entry->m_may_create,
+		__entry->m_multidev_dio,
+		__entry->create,
+		__entry->flag,
 		__entry->ret)
 );
 
@@ -807,20 +818,20 @@ TRACE_EVENT(f2fs_lookup_start,
 	TP_STRUCT__entry(
 		__field(dev_t,	dev)
 		__field(ino_t,	ino)
-		__field(const char *,	name)
+		__string(name,	dentry->d_name.name)
 		__field(unsigned int, flags)
 	),
 
 	TP_fast_assign(
 		__entry->dev	= dir->i_sb->s_dev;
 		__entry->ino	= dir->i_ino;
-		__entry->name	= dentry->d_name.name;
+		__assign_str(name, dentry->d_name.name);
 		__entry->flags	= flags;
 	),
 
 	TP_printk("dev = (%d,%d), pino = %lu, name:%s, flags:%u",
 		show_dev_ino(__entry),
-		__entry->name,
+		__get_str(name),
 		__entry->flags)
 );
 
@@ -834,7 +845,7 @@ TRACE_EVENT(f2fs_lookup_end,
 	TP_STRUCT__entry(
 		__field(dev_t,	dev)
 		__field(ino_t,	ino)
-		__field(const char *,	name)
+		__string(name,	dentry->d_name.name)
 		__field(nid_t,	cino)
 		__field(int,	err)
 	),
@@ -842,14 +853,14 @@ TRACE_EVENT(f2fs_lookup_end,
 	TP_fast_assign(
 		__entry->dev	= dir->i_sb->s_dev;
 		__entry->ino	= dir->i_ino;
-		__entry->name	= dentry->d_name.name;
+		__assign_str(name, dentry->d_name.name);
 		__entry->cino	= ino;
 		__entry->err	= err;
 	),
 
 	TP_printk("dev = (%d,%d), pino = %lu, name:%s, ino:%u, err:%d",
 		show_dev_ino(__entry),
-		__entry->name,
+		__get_str(name),
 		__entry->cino,
 		__entry->err)
 );
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index 78f0719..1daa452 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -1130,6 +1130,7 @@ struct kvm_ppc_resize_hpt {
 #define KVM_CAP_BINARY_STATS_FD 203
 #define KVM_CAP_EXIT_ON_EMULATION_FAILURE 204
 #define KVM_CAP_ARM_MTE 205
+#define KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM 206
 
 #ifdef KVM_CAP_IRQ_ROUTING
 
diff --git a/include/uapi/linux/rtc.h b/include/uapi/linux/rtc.h
index f950bff..03e5b77 100644
--- a/include/uapi/linux/rtc.h
+++ b/include/uapi/linux/rtc.h
@@ -14,6 +14,7 @@
 
 #include <linux/const.h>
 #include <linux/ioctl.h>
+#include <linux/types.h>
 
 /*
  * The struct used to pass data via the following ioctl. Similar to the
@@ -66,6 +67,17 @@ struct rtc_pll_info {
 	long pll_clock;     /* base PLL frequency */
 };
 
+struct rtc_param {
+	__u64 param;
+	union {
+		__u64 uvalue;
+		__s64 svalue;
+		__u64 ptr;
+	};
+	__u32 index;
+	__u32 __pad;
+};
+
 /*
  * ioctl calls that are permitted to the /dev/rtc interface, if
  * any of the RTC drivers are enabled.
@@ -95,6 +107,9 @@ struct rtc_pll_info {
 #define RTC_PLL_GET	_IOR('p', 0x11, struct rtc_pll_info)  /* Get PLL correction */
 #define RTC_PLL_SET	_IOW('p', 0x12, struct rtc_pll_info)  /* Set PLL correction */
 
+#define RTC_PARAM_GET	_IOW('p', 0x13, struct rtc_param)  /* Get parameter */
+#define RTC_PARAM_SET	_IOW('p', 0x14, struct rtc_param)  /* Set parameter */
+
 #define RTC_VL_DATA_INVALID	_BITUL(0) /* Voltage too low, RTC data is invalid */
 #define RTC_VL_BACKUP_LOW	_BITUL(1) /* Backup voltage is low */
 #define RTC_VL_BACKUP_EMPTY	_BITUL(2) /* Backup empty or not present */
@@ -114,7 +129,21 @@ struct rtc_pll_info {
 #define RTC_FEATURE_ALARM		0
 #define RTC_FEATURE_ALARM_RES_MINUTE	1
 #define RTC_FEATURE_NEED_WEEK_DAY	2
-#define RTC_FEATURE_CNT			3
+#define RTC_FEATURE_ALARM_RES_2S	3
+#define RTC_FEATURE_UPDATE_INTERRUPT	4
+#define RTC_FEATURE_CORRECTION		5
+#define RTC_FEATURE_BACKUP_SWITCH_MODE	6
+#define RTC_FEATURE_CNT			7
+
+/* parameter list */
+#define RTC_PARAM_FEATURES		0
+#define RTC_PARAM_CORRECTION		1
+#define RTC_PARAM_BACKUP_SWITCH_MODE	2
+
+#define RTC_BSM_DISABLED	0
+#define RTC_BSM_DIRECT		1
+#define RTC_BSM_LEVEL		2
+#define RTC_BSM_STANDBY		3
 
 #define RTC_MAX_FREQ	8192
 
diff --git a/include/uapi/linux/virtio_mem.h b/include/uapi/linux/virtio_mem.h
index 70e01c6..e9122f1 100644
--- a/include/uapi/linux/virtio_mem.h
+++ b/include/uapi/linux/virtio_mem.h
@@ -68,9 +68,10 @@
  * explicitly triggered (VIRTIO_MEM_REQ_UNPLUG).
  *
  * There are no guarantees what will happen if unplugged memory is
- * read/written. Such memory should, in general, not be touched. E.g.,
- * even writing might succeed, but the values will simply be discarded at
- * random points in time.
+ * read/written. In general, unplugged memory should not be touched, because
+ * the resulting action is undefined. There is one exception: without
+ * VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE, unplugged memory inside the usable
+ * region can be read, to simplify creation of memory dumps.
  *
  * It can happen that the device cannot process a request, because it is
  * busy. The device driver has to retry later.
@@ -87,6 +88,8 @@
 
 /* node_id is an ACPI PXM and is valid */
 #define VIRTIO_MEM_F_ACPI_PXM		0
+/* unplugged memory must not be accessed */
+#define VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE	1
 
 
 /* --- virtio-mem: guest -> host requests --- */
diff --git a/init/Makefile b/init/Makefile
index 2846113..04eeee12 100644
--- a/init/Makefile
+++ b/init/Makefile
@@ -30,7 +30,7 @@
 quiet_cmd_compile.h = CHK     $@
       cmd_compile.h = \
 	$(CONFIG_SHELL) $(srctree)/scripts/mkcompile_h $@	\
-	"$(UTS_MACHINE)" "$(CONFIG_SMP)" "$(CONFIG_PREEMPT)"	\
+	"$(UTS_MACHINE)" "$(CONFIG_SMP)" "$(CONFIG_PREEMPT_BUILD)"	\
 	"$(CONFIG_PREEMPT_RT)" $(CONFIG_CC_VERSION_TEXT) "$(LD)"
 
 include/generated/compile.h: FORCE
diff --git a/kernel/Kconfig.preempt b/kernel/Kconfig.preempt
index 60f1bfc..ce77f02 100644
--- a/kernel/Kconfig.preempt
+++ b/kernel/Kconfig.preempt
@@ -1,12 +1,23 @@
 # SPDX-License-Identifier: GPL-2.0-only
 
+config PREEMPT_NONE_BUILD
+	bool
+
+config PREEMPT_VOLUNTARY_BUILD
+	bool
+
+config PREEMPT_BUILD
+	bool
+	select PREEMPTION
+	select UNINLINE_SPIN_UNLOCK if !ARCH_INLINE_SPIN_UNLOCK
+
 choice
 	prompt "Preemption Model"
-	default PREEMPT_NONE_BEHAVIOUR
+	default PREEMPT_NONE
 
-config PREEMPT_NONE_BEHAVIOUR
+config PREEMPT_NONE
 	bool "No Forced Preemption (Server)"
-	select PREEMPT_NONE if !PREEMPT_DYNAMIC
+	select PREEMPT_NONE_BUILD if !PREEMPT_DYNAMIC
 	help
 	  This is the traditional Linux preemption model, geared towards
 	  throughput. It will still provide good latencies most of the
@@ -18,10 +29,10 @@
 	  raw processing power of the kernel, irrespective of scheduling
 	  latencies.
 
-config PREEMPT_VOLUNTARY_BEHAVIOUR
+config PREEMPT_VOLUNTARY
 	bool "Voluntary Kernel Preemption (Desktop)"
 	depends on !ARCH_NO_PREEMPT
-	select PREEMPT_VOLUNTARY if !PREEMPT_DYNAMIC
+	select PREEMPT_VOLUNTARY_BUILD if !PREEMPT_DYNAMIC
 	help
 	  This option reduces the latency of the kernel by adding more
 	  "explicit preemption points" to the kernel code. These new
@@ -37,10 +48,10 @@
 
 	  Select this if you are building a kernel for a desktop system.
 
-config PREEMPT_BEHAVIOUR
+config PREEMPT
 	bool "Preemptible Kernel (Low-Latency Desktop)"
 	depends on !ARCH_NO_PREEMPT
-	select PREEMPT
+	select PREEMPT_BUILD
 	help
 	  This option reduces the latency of the kernel by making
 	  all kernel code (that is not executing in a critical section)
@@ -58,7 +69,7 @@
 
 config PREEMPT_RT
 	bool "Fully Preemptible Kernel (Real-Time)"
-	depends on EXPERT && ARCH_SUPPORTS_RT && !PREEMPT_DYNAMIC
+	depends on EXPERT && ARCH_SUPPORTS_RT
 	select PREEMPTION
 	help
 	  This option turns the kernel into a real-time kernel by replacing
@@ -75,17 +86,6 @@
 
 endchoice
 
-config PREEMPT_NONE
-	bool
-
-config PREEMPT_VOLUNTARY
-	bool
-
-config PREEMPT
-	bool
-	select PREEMPTION
-	select UNINLINE_SPIN_UNLOCK if !ARCH_INLINE_SPIN_UNLOCK
-
 config PREEMPT_COUNT
        bool
 
@@ -95,8 +95,8 @@
 
 config PREEMPT_DYNAMIC
 	bool "Preemption behaviour defined on boot"
-	depends on HAVE_PREEMPT_DYNAMIC
-	select PREEMPT
+	depends on HAVE_PREEMPT_DYNAMIC && !PREEMPT_RT
+	select PREEMPT_BUILD
 	default y
 	help
 	  This option allows to define the preemption model on the kernel
diff --git a/kernel/events/core.c b/kernel/events/core.c
index f2253ea..523106a 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -7154,7 +7154,6 @@ void perf_output_sample(struct perf_output_handle *handle,
 static u64 perf_virt_to_phys(u64 virt)
 {
 	u64 phys_addr = 0;
-	struct page *p = NULL;
 
 	if (!virt)
 		return 0;
@@ -7173,14 +7172,15 @@ static u64 perf_virt_to_phys(u64 virt)
 		 * If failed, leave phys_addr as 0.
 		 */
 		if (current->mm != NULL) {
+			struct page *p;
+
 			pagefault_disable();
-			if (get_user_page_fast_only(virt, 0, &p))
+			if (get_user_page_fast_only(virt, 0, &p)) {
 				phys_addr = page_to_phys(p) + virt % PAGE_SIZE;
+				put_page(p);
+			}
 			pagefault_enable();
 		}
-
-		if (p)
-			put_page(p);
 	}
 
 	return phys_addr;
diff --git a/kernel/fork.c b/kernel/fork.c
index 5de23f3..3244cc5 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -2277,6 +2277,7 @@ static __latent_entropy struct task_struct *copy_process(
 	p->pdeath_signal = 0;
 	INIT_LIST_HEAD(&p->thread_group);
 	p->task_works = NULL;
+	clear_posix_cputimers_work(p);
 
 #ifdef CONFIG_KRETPROBES
 	p->kretprobe_instances.first = NULL;
diff --git a/kernel/irq/msi.c b/kernel/irq/msi.c
index 6a5ecee..7f350ae 100644
--- a/kernel/irq/msi.c
+++ b/kernel/irq/msi.c
@@ -529,10 +529,10 @@ static bool msi_check_reservation_mode(struct irq_domain *domain,
 
 	/*
 	 * Checking the first MSI descriptor is sufficient. MSIX supports
-	 * masking and MSI does so when the maskbit is set.
+	 * masking and MSI does so when the can_mask attribute is set.
 	 */
 	desc = first_msi_entry(dev);
-	return desc->msi_attrib.is_msix || desc->msi_attrib.maskbit;
+	return desc->msi_attrib.is_msix || desc->msi_attrib.can_mask;
 }
 
 int __msi_domain_alloc_irqs(struct irq_domain *domain, struct device *dev,
diff --git a/kernel/reboot.c b/kernel/reboot.c
index d6e0f9f..6bcc5d6 100644
--- a/kernel/reboot.c
+++ b/kernel/reboot.c
@@ -33,6 +33,7 @@ EXPORT_SYMBOL(cad_pid);
 #define DEFAULT_REBOOT_MODE
 #endif
 enum reboot_mode reboot_mode DEFAULT_REBOOT_MODE;
+EXPORT_SYMBOL_GPL(reboot_mode);
 enum reboot_mode panic_reboot_mode = REBOOT_UNDEFINED;
 
 /*
diff --git a/kernel/sched/autogroup.c b/kernel/sched/autogroup.c
index 2067080..8629b37 100644
--- a/kernel/sched/autogroup.c
+++ b/kernel/sched/autogroup.c
@@ -31,7 +31,7 @@ static inline void autogroup_destroy(struct kref *kref)
 	ag->tg->rt_se = NULL;
 	ag->tg->rt_rq = NULL;
 #endif
-	sched_offline_group(ag->tg);
+	sched_release_group(ag->tg);
 	sched_destroy_group(ag->tg);
 }
 
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 523fd60..3c9b0fd 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -3726,6 +3726,9 @@ void wake_up_if_idle(int cpu)
 
 bool cpus_share_cache(int this_cpu, int that_cpu)
 {
+	if (this_cpu == that_cpu)
+		return true;
+
 	return per_cpu(sd_llc_id, this_cpu) == per_cpu(sd_llc_id, that_cpu);
 }
 
@@ -6625,13 +6628,13 @@ __setup("preempt=", setup_preempt_mode);
 static void __init preempt_dynamic_init(void)
 {
 	if (preempt_dynamic_mode == preempt_dynamic_undefined) {
-		if (IS_ENABLED(CONFIG_PREEMPT_NONE_BEHAVIOUR)) {
+		if (IS_ENABLED(CONFIG_PREEMPT_NONE)) {
 			sched_dynamic_update(preempt_dynamic_none);
-		} else if (IS_ENABLED(CONFIG_PREEMPT_VOLUNTARY_BEHAVIOUR)) {
+		} else if (IS_ENABLED(CONFIG_PREEMPT_VOLUNTARY)) {
 			sched_dynamic_update(preempt_dynamic_voluntary);
 		} else {
 			/* Default static call setting, nothing to do */
-			WARN_ON_ONCE(!IS_ENABLED(CONFIG_PREEMPT_BEHAVIOUR));
+			WARN_ON_ONCE(!IS_ENABLED(CONFIG_PREEMPT));
 			preempt_dynamic_mode = preempt_dynamic_full;
 			pr_info("Dynamic Preempt: full\n");
 		}
@@ -9716,6 +9719,22 @@ static void sched_free_group(struct task_group *tg)
 	kmem_cache_free(task_group_cache, tg);
 }
 
+static void sched_free_group_rcu(struct rcu_head *rcu)
+{
+	sched_free_group(container_of(rcu, struct task_group, rcu));
+}
+
+static void sched_unregister_group(struct task_group *tg)
+{
+	unregister_fair_sched_group(tg);
+	unregister_rt_sched_group(tg);
+	/*
+	 * We have to wait for yet another RCU grace period to expire, as
+	 * print_cfs_stats() might run concurrently.
+	 */
+	call_rcu(&tg->rcu, sched_free_group_rcu);
+}
+
 /* allocate runqueue etc for a new task group */
 struct task_group *sched_create_group(struct task_group *parent)
 {
@@ -9759,25 +9778,35 @@ void sched_online_group(struct task_group *tg, struct task_group *parent)
 }
 
 /* rcu callback to free various structures associated with a task group */
-static void sched_free_group_rcu(struct rcu_head *rhp)
+static void sched_unregister_group_rcu(struct rcu_head *rhp)
 {
 	/* Now it should be safe to free those cfs_rqs: */
-	sched_free_group(container_of(rhp, struct task_group, rcu));
+	sched_unregister_group(container_of(rhp, struct task_group, rcu));
 }
 
 void sched_destroy_group(struct task_group *tg)
 {
 	/* Wait for possible concurrent references to cfs_rqs complete: */
-	call_rcu(&tg->rcu, sched_free_group_rcu);
+	call_rcu(&tg->rcu, sched_unregister_group_rcu);
 }
 
-void sched_offline_group(struct task_group *tg)
+void sched_release_group(struct task_group *tg)
 {
 	unsigned long flags;
 
-	/* End participation in shares distribution: */
-	unregister_fair_sched_group(tg);
-
+	/*
+	 * Unlink first, to avoid walk_tg_tree_from() from finding us (via
+	 * sched_cfs_period_timer()).
+	 *
+	 * For this to be effective, we have to wait for all pending users of
+	 * this task group to leave their RCU critical section to ensure no new
+	 * user will see our dying task group any more. Specifically ensure
+	 * that tg_unthrottle_up() won't add decayed cfs_rq's to it.
+	 *
+	 * We therefore defer calling unregister_fair_sched_group() to
+	 * sched_unregister_group() which is guarantied to get called only after the
+	 * current RCU grace period has expired.
+	 */
 	spin_lock_irqsave(&task_group_lock, flags);
 	list_del_rcu(&tg->list);
 	list_del_rcu(&tg->siblings);
@@ -9896,7 +9925,7 @@ static void cpu_cgroup_css_released(struct cgroup_subsys_state *css)
 {
 	struct task_group *tg = css_tg(css);
 
-	sched_offline_group(tg);
+	sched_release_group(tg);
 }
 
 static void cpu_cgroup_css_free(struct cgroup_subsys_state *css)
@@ -9906,7 +9935,7 @@ static void cpu_cgroup_css_free(struct cgroup_subsys_state *css)
 	/*
 	 * Relies on the RCU grace period between css_released() and this.
 	 */
-	sched_free_group(tg);
+	sched_unregister_group(tg);
 }
 
 /*
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 13950be..6e476f6 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -11456,8 +11456,6 @@ void free_fair_sched_group(struct task_group *tg)
 {
 	int i;
 
-	destroy_cfs_bandwidth(tg_cfs_bandwidth(tg));
-
 	for_each_possible_cpu(i) {
 		if (tg->cfs_rq)
 			kfree(tg->cfs_rq[i]);
@@ -11534,6 +11532,8 @@ void unregister_fair_sched_group(struct task_group *tg)
 	struct rq *rq;
 	int cpu;
 
+	destroy_cfs_bandwidth(tg_cfs_bandwidth(tg));
+
 	for_each_possible_cpu(cpu) {
 		if (tg->se[cpu])
 			remove_entity_load_avg(tg->se[cpu]);
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index bb945f8..b48baab 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -137,13 +137,17 @@ static inline struct rq *rq_of_rt_se(struct sched_rt_entity *rt_se)
 	return rt_rq->rq;
 }
 
+void unregister_rt_sched_group(struct task_group *tg)
+{
+	if (tg->rt_se)
+		destroy_rt_bandwidth(&tg->rt_bandwidth);
+
+}
+
 void free_rt_sched_group(struct task_group *tg)
 {
 	int i;
 
-	if (tg->rt_se)
-		destroy_rt_bandwidth(&tg->rt_bandwidth);
-
 	for_each_possible_cpu(i) {
 		if (tg->rt_rq)
 			kfree(tg->rt_rq[i]);
@@ -250,6 +254,8 @@ static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
 	return &rq->rt;
 }
 
+void unregister_rt_sched_group(struct task_group *tg) { }
+
 void free_rt_sched_group(struct task_group *tg) { }
 
 int alloc_rt_sched_group(struct task_group *tg, struct task_group *parent)
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 7f1612d..0e66749 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -488,6 +488,7 @@ extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
 extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
 extern void unthrottle_cfs_rq(struct cfs_rq *cfs_rq);
 
+extern void unregister_rt_sched_group(struct task_group *tg);
 extern void free_rt_sched_group(struct task_group *tg);
 extern int alloc_rt_sched_group(struct task_group *tg, struct task_group *parent);
 extern void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq,
@@ -503,7 +504,7 @@ extern struct task_group *sched_create_group(struct task_group *parent);
 extern void sched_online_group(struct task_group *tg,
 			       struct task_group *parent);
 extern void sched_destroy_group(struct task_group *tg);
-extern void sched_offline_group(struct task_group *tg);
+extern void sched_release_group(struct task_group *tg);
 
 extern void sched_move_task(struct task_struct *tsk);
 
diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 643d412..96b4e78 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -1159,13 +1159,28 @@ static void posix_cpu_timers_work(struct callback_head *work)
 }
 
 /*
+ * Clear existing posix CPU timers task work.
+ */
+void clear_posix_cputimers_work(struct task_struct *p)
+{
+	/*
+	 * A copied work entry from the old task is not meaningful, clear it.
+	 * N.B. init_task_work will not do this.
+	 */
+	memset(&p->posix_cputimers_work.work, 0,
+	       sizeof(p->posix_cputimers_work.work));
+	init_task_work(&p->posix_cputimers_work.work,
+		       posix_cpu_timers_work);
+	p->posix_cputimers_work.scheduled = false;
+}
+
+/*
  * Initialize posix CPU timers task work in init task. Out of line to
  * keep the callback static and to avoid header recursion hell.
  */
 void __init posix_cputimers_init_work(void)
 {
-	init_task_work(&current->posix_cputimers_work.work,
-		       posix_cpu_timers_work);
+	clear_posix_cputimers_work(current);
 }
 
 /*
diff --git a/kernel/trace/trace_events_hist.c b/kernel/trace/trace_events_hist.c
index 0abc9a4..8a10046 100644
--- a/kernel/trace/trace_events_hist.c
+++ b/kernel/trace/trace_events_hist.c
@@ -1953,9 +1953,10 @@ static struct hist_field *create_hist_field(struct hist_trigger_data *hist_data,
 		if (!hist_field->type)
 			goto free;
 
-		if (field->filter_type == FILTER_STATIC_STRING)
+		if (field->filter_type == FILTER_STATIC_STRING) {
 			hist_field->fn = hist_field_string;
-		else if (field->filter_type == FILTER_DYN_STRING)
+			hist_field->size = field->size;
+		} else if (field->filter_type == FILTER_DYN_STRING)
 			hist_field->fn = hist_field_dynstring;
 		else
 			hist_field->fn = hist_field_pstring;
@@ -2580,7 +2581,8 @@ static struct hist_field *parse_expr(struct hist_trigger_data *hist_data,
 	operand1_str = str;
 	str = sep+1;
 
-	if (!operand1_str || !str)
+	/* Binary operator requires both operands */
+	if (*operand1_str == '\0' || *str == '\0')
 		goto free;
 
 	operand_flags = 0;
@@ -3025,7 +3027,7 @@ static inline void __update_field_vars(struct tracing_map_elt *elt,
 			char *str = elt_data->field_var_str[j++];
 			char *val_str = (char *)(uintptr_t)var_val;
 
-			strscpy(str, val_str, STR_VAR_LEN_MAX);
+			strscpy(str, val_str, val->size);
 			var_val = (u64)(uintptr_t)str;
 		}
 		tracing_map_set_var(elt, var_idx, var_val);
@@ -4920,7 +4922,7 @@ static void hist_trigger_elt_update(struct hist_trigger_data *hist_data,
 
 				str = elt_data->field_var_str[idx];
 				val_str = (char *)(uintptr_t)hist_val;
-				strscpy(str, val_str, STR_VAR_LEN_MAX);
+				strscpy(str, val_str, hist_field->size);
 
 				hist_val = (u64)(uintptr_t)str;
 			}
diff --git a/kernel/trace/trace_osnoise.c b/kernel/trace/trace_osnoise.c
index 3e4a165..7520d43 100644
--- a/kernel/trace/trace_osnoise.c
+++ b/kernel/trace/trace_osnoise.c
@@ -55,7 +55,8 @@ struct osnoise_instance {
 	struct list_head	list;
 	struct trace_array	*tr;
 };
-struct list_head osnoise_instances;
+
+static struct list_head osnoise_instances;
 
 static bool osnoise_has_registered_instances(void)
 {
diff --git a/lib/decompress_unzstd.c b/lib/decompress_unzstd.c
index 6b629ab..a512b99 100644
--- a/lib/decompress_unzstd.c
+++ b/lib/decompress_unzstd.c
@@ -68,11 +68,7 @@
 #ifdef STATIC
 # define UNZSTD_PREBOOT
 # include "xxhash.c"
-# include "zstd/entropy_common.c"
-# include "zstd/fse_decompress.c"
-# include "zstd/huf_decompress.c"
-# include "zstd/zstd_common.c"
-# include "zstd/decompress.c"
+# include "zstd/decompress_sources.h"
 #endif
 
 #include <linux/decompress/mm.h>
@@ -91,11 +87,15 @@
 
 static int INIT handle_zstd_error(size_t ret, void (*error)(char *x))
 {
-	const int err = ZSTD_getErrorCode(ret);
+	const zstd_error_code err = zstd_get_error_code(ret);
 
-	if (!ZSTD_isError(ret))
+	if (!zstd_is_error(ret))
 		return 0;
 
+	/*
+	 * zstd_get_error_name() cannot be used because error takes a char *
+	 * not a const char *
+	 */
 	switch (err) {
 	case ZSTD_error_memory_allocation:
 		error("ZSTD decompressor ran out of memory");
@@ -124,28 +124,28 @@ static int INIT decompress_single(const u8 *in_buf, long in_len, u8 *out_buf,
 				  long out_len, long *in_pos,
 				  void (*error)(char *x))
 {
-	const size_t wksp_size = ZSTD_DCtxWorkspaceBound();
+	const size_t wksp_size = zstd_dctx_workspace_bound();
 	void *wksp = large_malloc(wksp_size);
-	ZSTD_DCtx *dctx = ZSTD_initDCtx(wksp, wksp_size);
+	zstd_dctx *dctx = zstd_init_dctx(wksp, wksp_size);
 	int err;
 	size_t ret;
 
 	if (dctx == NULL) {
-		error("Out of memory while allocating ZSTD_DCtx");
+		error("Out of memory while allocating zstd_dctx");
 		err = -1;
 		goto out;
 	}
 	/*
 	 * Find out how large the frame actually is, there may be junk at
-	 * the end of the frame that ZSTD_decompressDCtx() can't handle.
+	 * the end of the frame that zstd_decompress_dctx() can't handle.
 	 */
-	ret = ZSTD_findFrameCompressedSize(in_buf, in_len);
+	ret = zstd_find_frame_compressed_size(in_buf, in_len);
 	err = handle_zstd_error(ret, error);
 	if (err)
 		goto out;
 	in_len = (long)ret;
 
-	ret = ZSTD_decompressDCtx(dctx, out_buf, out_len, in_buf, in_len);
+	ret = zstd_decompress_dctx(dctx, out_buf, out_len, in_buf, in_len);
 	err = handle_zstd_error(ret, error);
 	if (err)
 		goto out;
@@ -167,14 +167,14 @@ static int INIT __unzstd(unsigned char *in_buf, long in_len,
 			 long *in_pos,
 			 void (*error)(char *x))
 {
-	ZSTD_inBuffer in;
-	ZSTD_outBuffer out;
-	ZSTD_frameParams params;
+	zstd_in_buffer in;
+	zstd_out_buffer out;
+	zstd_frame_header header;
 	void *in_allocated = NULL;
 	void *out_allocated = NULL;
 	void *wksp = NULL;
 	size_t wksp_size;
-	ZSTD_DStream *dstream;
+	zstd_dstream *dstream;
 	int err;
 	size_t ret;
 
@@ -238,13 +238,13 @@ static int INIT __unzstd(unsigned char *in_buf, long in_len,
 	out.size = out_len;
 
 	/*
-	 * We need to know the window size to allocate the ZSTD_DStream.
+	 * We need to know the window size to allocate the zstd_dstream.
 	 * Since we are streaming, we need to allocate a buffer for the sliding
 	 * window. The window size varies from 1 KB to ZSTD_WINDOWSIZE_MAX
 	 * (8 MB), so it is important to use the actual value so as not to
 	 * waste memory when it is smaller.
 	 */
-	ret = ZSTD_getFrameParams(&params, in.src, in.size);
+	ret = zstd_get_frame_header(&header, in.src, in.size);
 	err = handle_zstd_error(ret, error);
 	if (err)
 		goto out;
@@ -253,19 +253,19 @@ static int INIT __unzstd(unsigned char *in_buf, long in_len,
 		err = -1;
 		goto out;
 	}
-	if (params.windowSize > ZSTD_WINDOWSIZE_MAX) {
+	if (header.windowSize > ZSTD_WINDOWSIZE_MAX) {
 		error("ZSTD-compressed data has too large a window size");
 		err = -1;
 		goto out;
 	}
 
 	/*
-	 * Allocate the ZSTD_DStream now that we know how much memory is
+	 * Allocate the zstd_dstream now that we know how much memory is
 	 * required.
 	 */
-	wksp_size = ZSTD_DStreamWorkspaceBound(params.windowSize);
+	wksp_size = zstd_dstream_workspace_bound(header.windowSize);
 	wksp = large_malloc(wksp_size);
-	dstream = ZSTD_initDStream(params.windowSize, wksp, wksp_size);
+	dstream = zstd_init_dstream(header.windowSize, wksp, wksp_size);
 	if (dstream == NULL) {
 		error("Out of memory while allocating ZSTD_DStream");
 		err = -1;
@@ -298,7 +298,7 @@ static int INIT __unzstd(unsigned char *in_buf, long in_len,
 			in.size = in_len;
 		}
 		/* Returns zero when the frame is complete. */
-		ret = ZSTD_decompressStream(dstream, &out, &in);
+		ret = zstd_decompress_stream(dstream, &out, &in);
 		err = handle_zstd_error(ret, error);
 		if (err)
 			goto out;
diff --git a/lib/zstd/Makefile b/lib/zstd/Makefile
index f5d778e..65218ec 100644
--- a/lib/zstd/Makefile
+++ b/lib/zstd/Makefile
@@ -1,10 +1,46 @@
-# SPDX-License-Identifier: GPL-2.0-only
+# SPDX-License-Identifier: GPL-2.0+ OR BSD-3-Clause
+# ################################################################
+# Copyright (c) Facebook, Inc.
+# All rights reserved.
+#
+# This source code is licensed under both the BSD-style license (found in the
+# LICENSE file in the root directory of this source tree) and the GPLv2 (found
+# in the COPYING file in the root directory of this source tree).
+# You may select, at your option, one of the above-listed licenses.
+# ################################################################
 obj-$(CONFIG_ZSTD_COMPRESS) += zstd_compress.o
 obj-$(CONFIG_ZSTD_DECOMPRESS) += zstd_decompress.o
 
 ccflags-y += -O3
 
-zstd_compress-y := fse_compress.o huf_compress.o compress.o \
-		   entropy_common.o fse_decompress.o zstd_common.o
-zstd_decompress-y := huf_decompress.o decompress.o \
-		     entropy_common.o fse_decompress.o zstd_common.o
+zstd_compress-y := \
+		zstd_compress_module.o \
+		common/debug.o \
+		common/entropy_common.o \
+		common/error_private.o \
+		common/fse_decompress.o \
+		common/zstd_common.o \
+		compress/fse_compress.o \
+		compress/hist.o \
+		compress/huf_compress.o \
+		compress/zstd_compress.o \
+		compress/zstd_compress_literals.o \
+		compress/zstd_compress_sequences.o \
+		compress/zstd_compress_superblock.o \
+		compress/zstd_double_fast.o \
+		compress/zstd_fast.o \
+		compress/zstd_lazy.o \
+		compress/zstd_ldm.o \
+		compress/zstd_opt.o \
+
+zstd_decompress-y := \
+		zstd_decompress_module.o \
+		common/debug.o \
+		common/entropy_common.o \
+		common/error_private.o \
+		common/fse_decompress.o \
+		common/zstd_common.o \
+		decompress/huf_decompress.o \
+		decompress/zstd_ddict.o \
+		decompress/zstd_decompress.o \
+		decompress/zstd_decompress_block.o \
diff --git a/lib/zstd/bitstream.h b/lib/zstd/bitstream.h
deleted file mode 100644
index 5d6343c..0000000
--- a/lib/zstd/bitstream.h
+++ /dev/null
@@ -1,380 +0,0 @@
-/*
- * bitstream
- * Part of FSE library
- * header file (to include)
- * Copyright (C) 2013-2016, Yann Collet.
- *
- * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are
- * met:
- *
- *   * Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- *   * Redistributions in binary form must reproduce the above
- * copyright notice, this list of conditions and the following disclaimer
- * in the documentation and/or other materials provided with the
- * distribution.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- *
- * This program is free software; you can redistribute it and/or modify it under
- * the terms of the GNU General Public License version 2 as published by the
- * Free Software Foundation. This program is dual-licensed; you may select
- * either version 2 of the GNU General Public License ("GPL") or BSD license
- * ("BSD").
- *
- * You can contact the author at :
- * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
- */
-#ifndef BITSTREAM_H_MODULE
-#define BITSTREAM_H_MODULE
-
-/*
-*  This API consists of small unitary functions, which must be inlined for best performance.
-*  Since link-time-optimization is not available for all compilers,
-*  these functions are defined into a .h to be included.
-*/
-
-/*-****************************************
-*  Dependencies
-******************************************/
-#include "error_private.h" /* error codes and messages */
-#include "mem.h"	   /* unaligned access routines */
-
-/*=========================================
-*  Target specific
-=========================================*/
-#define STREAM_ACCUMULATOR_MIN_32 25
-#define STREAM_ACCUMULATOR_MIN_64 57
-#define STREAM_ACCUMULATOR_MIN ((U32)(ZSTD_32bits() ? STREAM_ACCUMULATOR_MIN_32 : STREAM_ACCUMULATOR_MIN_64))
-
-/*-******************************************
-*  bitStream encoding API (write forward)
-********************************************/
-/* bitStream can mix input from multiple sources.
-*  A critical property of these streams is that they encode and decode in **reverse** direction.
-*  So the first bit sequence you add will be the last to be read, like a LIFO stack.
-*/
-typedef struct {
-	size_t bitContainer;
-	int bitPos;
-	char *startPtr;
-	char *ptr;
-	char *endPtr;
-} BIT_CStream_t;
-
-ZSTD_STATIC size_t BIT_initCStream(BIT_CStream_t *bitC, void *dstBuffer, size_t dstCapacity);
-ZSTD_STATIC void BIT_addBits(BIT_CStream_t *bitC, size_t value, unsigned nbBits);
-ZSTD_STATIC void BIT_flushBits(BIT_CStream_t *bitC);
-ZSTD_STATIC size_t BIT_closeCStream(BIT_CStream_t *bitC);
-
-/* Start with initCStream, providing the size of buffer to write into.
-*  bitStream will never write outside of this buffer.
-*  `dstCapacity` must be >= sizeof(bitD->bitContainer), otherwise @return will be an error code.
-*
-*  bits are first added to a local register.
-*  Local register is size_t, hence 64-bits on 64-bits systems, or 32-bits on 32-bits systems.
-*  Writing data into memory is an explicit operation, performed by the flushBits function.
-*  Hence keep track how many bits are potentially stored into local register to avoid register overflow.
-*  After a flushBits, a maximum of 7 bits might still be stored into local register.
-*
-*  Avoid storing elements of more than 24 bits if you want compatibility with 32-bits bitstream readers.
-*
-*  Last operation is to close the bitStream.
-*  The function returns the final size of CStream in bytes.
-*  If data couldn't fit into `dstBuffer`, it will return a 0 ( == not storable)
-*/
-
-/*-********************************************
-*  bitStream decoding API (read backward)
-**********************************************/
-typedef struct {
-	size_t bitContainer;
-	unsigned bitsConsumed;
-	const char *ptr;
-	const char *start;
-} BIT_DStream_t;
-
-typedef enum {
-	BIT_DStream_unfinished = 0,
-	BIT_DStream_endOfBuffer = 1,
-	BIT_DStream_completed = 2,
-	BIT_DStream_overflow = 3
-} BIT_DStream_status; /* result of BIT_reloadDStream() */
-/* 1,2,4,8 would be better for bitmap combinations, but slows down performance a bit ... :( */
-
-ZSTD_STATIC size_t BIT_initDStream(BIT_DStream_t *bitD, const void *srcBuffer, size_t srcSize);
-ZSTD_STATIC size_t BIT_readBits(BIT_DStream_t *bitD, unsigned nbBits);
-ZSTD_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t *bitD);
-ZSTD_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t *bitD);
-
-/* Start by invoking BIT_initDStream().
-*  A chunk of the bitStream is then stored into a local register.
-*  Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (size_t).
-*  You can then retrieve bitFields stored into the local register, **in reverse order**.
-*  Local register is explicitly reloaded from memory by the BIT_reloadDStream() method.
-*  A reload guarantee a minimum of ((8*sizeof(bitD->bitContainer))-7) bits when its result is BIT_DStream_unfinished.
-*  Otherwise, it can be less than that, so proceed accordingly.
-*  Checking if DStream has reached its end can be performed with BIT_endOfDStream().
-*/
-
-/*-****************************************
-*  unsafe API
-******************************************/
-ZSTD_STATIC void BIT_addBitsFast(BIT_CStream_t *bitC, size_t value, unsigned nbBits);
-/* faster, but works only if value is "clean", meaning all high bits above nbBits are 0 */
-
-ZSTD_STATIC void BIT_flushBitsFast(BIT_CStream_t *bitC);
-/* unsafe version; does not check buffer overflow */
-
-ZSTD_STATIC size_t BIT_readBitsFast(BIT_DStream_t *bitD, unsigned nbBits);
-/* faster, but works only if nbBits >= 1 */
-
-/*-**************************************************************
-*  Internal functions
-****************************************************************/
-ZSTD_STATIC unsigned BIT_highbit32(register U32 val) { return 31 - __builtin_clz(val); }
-
-/*=====    Local Constants   =====*/
-static const unsigned BIT_mask[] = {0,       1,       3,       7,	0xF,      0x1F,     0x3F,     0x7F,      0xFF,
-				    0x1FF,   0x3FF,   0x7FF,   0xFFF,    0x1FFF,   0x3FFF,   0x7FFF,   0xFFFF,    0x1FFFF,
-				    0x3FFFF, 0x7FFFF, 0xFFFFF, 0x1FFFFF, 0x3FFFFF, 0x7FFFFF, 0xFFFFFF, 0x1FFFFFF, 0x3FFFFFF}; /* up to 26 bits */
-
-/*-**************************************************************
-*  bitStream encoding
-****************************************************************/
-/*! BIT_initCStream() :
- *  `dstCapacity` must be > sizeof(void*)
- *  @return : 0 if success,
-			  otherwise an error code (can be tested using ERR_isError() ) */
-ZSTD_STATIC size_t BIT_initCStream(BIT_CStream_t *bitC, void *startPtr, size_t dstCapacity)
-{
-	bitC->bitContainer = 0;
-	bitC->bitPos = 0;
-	bitC->startPtr = (char *)startPtr;
-	bitC->ptr = bitC->startPtr;
-	bitC->endPtr = bitC->startPtr + dstCapacity - sizeof(bitC->ptr);
-	if (dstCapacity <= sizeof(bitC->ptr))
-		return ERROR(dstSize_tooSmall);
-	return 0;
-}
-
-/*! BIT_addBits() :
-	can add up to 26 bits into `bitC`.
-	Does not check for register overflow ! */
-ZSTD_STATIC void BIT_addBits(BIT_CStream_t *bitC, size_t value, unsigned nbBits)
-{
-	bitC->bitContainer |= (value & BIT_mask[nbBits]) << bitC->bitPos;
-	bitC->bitPos += nbBits;
-}
-
-/*! BIT_addBitsFast() :
- *  works only if `value` is _clean_, meaning all high bits above nbBits are 0 */
-ZSTD_STATIC void BIT_addBitsFast(BIT_CStream_t *bitC, size_t value, unsigned nbBits)
-{
-	bitC->bitContainer |= value << bitC->bitPos;
-	bitC->bitPos += nbBits;
-}
-
-/*! BIT_flushBitsFast() :
- *  unsafe version; does not check buffer overflow */
-ZSTD_STATIC void BIT_flushBitsFast(BIT_CStream_t *bitC)
-{
-	size_t const nbBytes = bitC->bitPos >> 3;
-	ZSTD_writeLEST(bitC->ptr, bitC->bitContainer);
-	bitC->ptr += nbBytes;
-	bitC->bitPos &= 7;
-	bitC->bitContainer >>= nbBytes * 8; /* if bitPos >= sizeof(bitContainer)*8 --> undefined behavior */
-}
-
-/*! BIT_flushBits() :
- *  safe version; check for buffer overflow, and prevents it.
- *  note : does not signal buffer overflow. This will be revealed later on using BIT_closeCStream() */
-ZSTD_STATIC void BIT_flushBits(BIT_CStream_t *bitC)
-{
-	size_t const nbBytes = bitC->bitPos >> 3;
-	ZSTD_writeLEST(bitC->ptr, bitC->bitContainer);
-	bitC->ptr += nbBytes;
-	if (bitC->ptr > bitC->endPtr)
-		bitC->ptr = bitC->endPtr;
-	bitC->bitPos &= 7;
-	bitC->bitContainer >>= nbBytes * 8; /* if bitPos >= sizeof(bitContainer)*8 --> undefined behavior */
-}
-
-/*! BIT_closeCStream() :
- *  @return : size of CStream, in bytes,
-			  or 0 if it could not fit into dstBuffer */
-ZSTD_STATIC size_t BIT_closeCStream(BIT_CStream_t *bitC)
-{
-	BIT_addBitsFast(bitC, 1, 1); /* endMark */
-	BIT_flushBits(bitC);
-
-	if (bitC->ptr >= bitC->endPtr)
-		return 0; /* doesn't fit within authorized budget : cancel */
-
-	return (bitC->ptr - bitC->startPtr) + (bitC->bitPos > 0);
-}
-
-/*-********************************************************
-* bitStream decoding
-**********************************************************/
-/*! BIT_initDStream() :
-*   Initialize a BIT_DStream_t.
-*   `bitD` : a pointer to an already allocated BIT_DStream_t structure.
-*   `srcSize` must be the *exact* size of the bitStream, in bytes.
-*   @return : size of stream (== srcSize) or an errorCode if a problem is detected
-*/
-ZSTD_STATIC size_t BIT_initDStream(BIT_DStream_t *bitD, const void *srcBuffer, size_t srcSize)
-{
-	if (srcSize < 1) {
-		memset(bitD, 0, sizeof(*bitD));
-		return ERROR(srcSize_wrong);
-	}
-
-	if (srcSize >= sizeof(bitD->bitContainer)) { /* normal case */
-		bitD->start = (const char *)srcBuffer;
-		bitD->ptr = (const char *)srcBuffer + srcSize - sizeof(bitD->bitContainer);
-		bitD->bitContainer = ZSTD_readLEST(bitD->ptr);
-		{
-			BYTE const lastByte = ((const BYTE *)srcBuffer)[srcSize - 1];
-			bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0; /* ensures bitsConsumed is always set */
-			if (lastByte == 0)
-				return ERROR(GENERIC); /* endMark not present */
-		}
-	} else {
-		bitD->start = (const char *)srcBuffer;
-		bitD->ptr = bitD->start;
-		bitD->bitContainer = *(const BYTE *)(bitD->start);
-		switch (srcSize) {
-		case 7: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[6]) << (sizeof(bitD->bitContainer) * 8 - 16);
-			fallthrough;
-		case 6: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[5]) << (sizeof(bitD->bitContainer) * 8 - 24);
-			fallthrough;
-		case 5: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[4]) << (sizeof(bitD->bitContainer) * 8 - 32);
-			fallthrough;
-		case 4: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[3]) << 24;
-			fallthrough;
-		case 3: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[2]) << 16;
-			fallthrough;
-		case 2: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[1]) << 8;
-			fallthrough;
-		default:;
-		}
-		{
-			BYTE const lastByte = ((const BYTE *)srcBuffer)[srcSize - 1];
-			bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0;
-			if (lastByte == 0)
-				return ERROR(GENERIC); /* endMark not present */
-		}
-		bitD->bitsConsumed += (U32)(sizeof(bitD->bitContainer) - srcSize) * 8;
-	}
-
-	return srcSize;
-}
-
-ZSTD_STATIC size_t BIT_getUpperBits(size_t bitContainer, U32 const start) { return bitContainer >> start; }
-
-ZSTD_STATIC size_t BIT_getMiddleBits(size_t bitContainer, U32 const start, U32 const nbBits) { return (bitContainer >> start) & BIT_mask[nbBits]; }
-
-ZSTD_STATIC size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits) { return bitContainer & BIT_mask[nbBits]; }
-
-/*! BIT_lookBits() :
- *  Provides next n bits from local register.
- *  local register is not modified.
- *  On 32-bits, maxNbBits==24.
- *  On 64-bits, maxNbBits==56.
- *  @return : value extracted
- */
-ZSTD_STATIC size_t BIT_lookBits(const BIT_DStream_t *bitD, U32 nbBits)
-{
-	U32 const bitMask = sizeof(bitD->bitContainer) * 8 - 1;
-	return ((bitD->bitContainer << (bitD->bitsConsumed & bitMask)) >> 1) >> ((bitMask - nbBits) & bitMask);
-}
-
-/*! BIT_lookBitsFast() :
-*   unsafe version; only works only if nbBits >= 1 */
-ZSTD_STATIC size_t BIT_lookBitsFast(const BIT_DStream_t *bitD, U32 nbBits)
-{
-	U32 const bitMask = sizeof(bitD->bitContainer) * 8 - 1;
-	return (bitD->bitContainer << (bitD->bitsConsumed & bitMask)) >> (((bitMask + 1) - nbBits) & bitMask);
-}
-
-ZSTD_STATIC void BIT_skipBits(BIT_DStream_t *bitD, U32 nbBits) { bitD->bitsConsumed += nbBits; }
-
-/*! BIT_readBits() :
- *  Read (consume) next n bits from local register and update.
- *  Pay attention to not read more than nbBits contained into local register.
- *  @return : extracted value.
- */
-ZSTD_STATIC size_t BIT_readBits(BIT_DStream_t *bitD, U32 nbBits)
-{
-	size_t const value = BIT_lookBits(bitD, nbBits);
-	BIT_skipBits(bitD, nbBits);
-	return value;
-}
-
-/*! BIT_readBitsFast() :
-*   unsafe version; only works only if nbBits >= 1 */
-ZSTD_STATIC size_t BIT_readBitsFast(BIT_DStream_t *bitD, U32 nbBits)
-{
-	size_t const value = BIT_lookBitsFast(bitD, nbBits);
-	BIT_skipBits(bitD, nbBits);
-	return value;
-}
-
-/*! BIT_reloadDStream() :
-*   Refill `bitD` from buffer previously set in BIT_initDStream() .
-*   This function is safe, it guarantees it will not read beyond src buffer.
-*   @return : status of `BIT_DStream_t` internal register.
-			  if status == BIT_DStream_unfinished, internal register is filled with >= (sizeof(bitD->bitContainer)*8 - 7) bits */
-ZSTD_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t *bitD)
-{
-	if (bitD->bitsConsumed > (sizeof(bitD->bitContainer) * 8)) /* should not happen => corruption detected */
-		return BIT_DStream_overflow;
-
-	if (bitD->ptr >= bitD->start + sizeof(bitD->bitContainer)) {
-		bitD->ptr -= bitD->bitsConsumed >> 3;
-		bitD->bitsConsumed &= 7;
-		bitD->bitContainer = ZSTD_readLEST(bitD->ptr);
-		return BIT_DStream_unfinished;
-	}
-	if (bitD->ptr == bitD->start) {
-		if (bitD->bitsConsumed < sizeof(bitD->bitContainer) * 8)
-			return BIT_DStream_endOfBuffer;
-		return BIT_DStream_completed;
-	}
-	{
-		U32 nbBytes = bitD->bitsConsumed >> 3;
-		BIT_DStream_status result = BIT_DStream_unfinished;
-		if (bitD->ptr - nbBytes < bitD->start) {
-			nbBytes = (U32)(bitD->ptr - bitD->start); /* ptr > start */
-			result = BIT_DStream_endOfBuffer;
-		}
-		bitD->ptr -= nbBytes;
-		bitD->bitsConsumed -= nbBytes * 8;
-		bitD->bitContainer = ZSTD_readLEST(bitD->ptr); /* reminder : srcSize > sizeof(bitD) */
-		return result;
-	}
-}
-
-/*! BIT_endOfDStream() :
-*   @return Tells if DStream has exactly reached its end (all bits consumed).
-*/
-ZSTD_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t *DStream)
-{
-	return ((DStream->ptr == DStream->start) && (DStream->bitsConsumed == sizeof(DStream->bitContainer) * 8));
-}
-
-#endif /* BITSTREAM_H_MODULE */
diff --git a/lib/zstd/common/bitstream.h b/lib/zstd/common/bitstream.h
new file mode 100644
index 0000000..28248ab
--- /dev/null
+++ b/lib/zstd/common/bitstream.h
@@ -0,0 +1,437 @@
+/* ******************************************************************
+ * bitstream
+ * Part of FSE library
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ * You can contact the author at :
+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+#ifndef BITSTREAM_H_MODULE
+#define BITSTREAM_H_MODULE
+
+/*
+*  This API consists of small unitary functions, which must be inlined for best performance.
+*  Since link-time-optimization is not available for all compilers,
+*  these functions are defined into a .h to be included.
+*/
+
+/*-****************************************
+*  Dependencies
+******************************************/
+#include "mem.h"            /* unaligned access routines */
+#include "compiler.h"       /* UNLIKELY() */
+#include "debug.h"          /* assert(), DEBUGLOG(), RAWLOG() */
+#include "error_private.h"  /* error codes and messages */
+
+
+/*=========================================
+*  Target specific
+=========================================*/
+
+#define STREAM_ACCUMULATOR_MIN_32  25
+#define STREAM_ACCUMULATOR_MIN_64  57
+#define STREAM_ACCUMULATOR_MIN    ((U32)(MEM_32bits() ? STREAM_ACCUMULATOR_MIN_32 : STREAM_ACCUMULATOR_MIN_64))
+
+
+/*-******************************************
+*  bitStream encoding API (write forward)
+********************************************/
+/* bitStream can mix input from multiple sources.
+ * A critical property of these streams is that they encode and decode in **reverse** direction.
+ * So the first bit sequence you add will be the last to be read, like a LIFO stack.
+ */
+typedef struct {
+    size_t bitContainer;
+    unsigned bitPos;
+    char*  startPtr;
+    char*  ptr;
+    char*  endPtr;
+} BIT_CStream_t;
+
+MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC, void* dstBuffer, size_t dstCapacity);
+MEM_STATIC void   BIT_addBits(BIT_CStream_t* bitC, size_t value, unsigned nbBits);
+MEM_STATIC void   BIT_flushBits(BIT_CStream_t* bitC);
+MEM_STATIC size_t BIT_closeCStream(BIT_CStream_t* bitC);
+
+/* Start with initCStream, providing the size of buffer to write into.
+*  bitStream will never write outside of this buffer.
+*  `dstCapacity` must be >= sizeof(bitD->bitContainer), otherwise @return will be an error code.
+*
+*  bits are first added to a local register.
+*  Local register is size_t, hence 64-bits on 64-bits systems, or 32-bits on 32-bits systems.
+*  Writing data into memory is an explicit operation, performed by the flushBits function.
+*  Hence keep track how many bits are potentially stored into local register to avoid register overflow.
+*  After a flushBits, a maximum of 7 bits might still be stored into local register.
+*
+*  Avoid storing elements of more than 24 bits if you want compatibility with 32-bits bitstream readers.
+*
+*  Last operation is to close the bitStream.
+*  The function returns the final size of CStream in bytes.
+*  If data couldn't fit into `dstBuffer`, it will return a 0 ( == not storable)
+*/
+
+
+/*-********************************************
+*  bitStream decoding API (read backward)
+**********************************************/
+typedef struct {
+    size_t   bitContainer;
+    unsigned bitsConsumed;
+    const char* ptr;
+    const char* start;
+    const char* limitPtr;
+} BIT_DStream_t;
+
+typedef enum { BIT_DStream_unfinished = 0,
+               BIT_DStream_endOfBuffer = 1,
+               BIT_DStream_completed = 2,
+               BIT_DStream_overflow = 3 } BIT_DStream_status;  /* result of BIT_reloadDStream() */
+               /* 1,2,4,8 would be better for bitmap combinations, but slows down performance a bit ... :( */
+
+MEM_STATIC size_t   BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize);
+MEM_STATIC size_t   BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits);
+MEM_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD);
+MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* bitD);
+
+
+/* Start by invoking BIT_initDStream().
+*  A chunk of the bitStream is then stored into a local register.
+*  Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (size_t).
+*  You can then retrieve bitFields stored into the local register, **in reverse order**.
+*  Local register is explicitly reloaded from memory by the BIT_reloadDStream() method.
+*  A reload guarantee a minimum of ((8*sizeof(bitD->bitContainer))-7) bits when its result is BIT_DStream_unfinished.
+*  Otherwise, it can be less than that, so proceed accordingly.
+*  Checking if DStream has reached its end can be performed with BIT_endOfDStream().
+*/
+
+
+/*-****************************************
+*  unsafe API
+******************************************/
+MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC, size_t value, unsigned nbBits);
+/* faster, but works only if value is "clean", meaning all high bits above nbBits are 0 */
+
+MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC);
+/* unsafe version; does not check buffer overflow */
+
+MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits);
+/* faster, but works only if nbBits >= 1 */
+
+
+
+/*-**************************************************************
+*  Internal functions
+****************************************************************/
+MEM_STATIC unsigned BIT_highbit32 (U32 val)
+{
+    assert(val != 0);
+    {
+#   if (__GNUC__ >= 3)   /* Use GCC Intrinsic */
+        return __builtin_clz (val) ^ 31;
+#   else   /* Software version */
+        static const unsigned DeBruijnClz[32] = { 0,  9,  1, 10, 13, 21,  2, 29,
+                                                 11, 14, 16, 18, 22, 25,  3, 30,
+                                                  8, 12, 20, 28, 15, 17, 24,  7,
+                                                 19, 27, 23,  6, 26,  5,  4, 31 };
+        U32 v = val;
+        v |= v >> 1;
+        v |= v >> 2;
+        v |= v >> 4;
+        v |= v >> 8;
+        v |= v >> 16;
+        return DeBruijnClz[ (U32) (v * 0x07C4ACDDU) >> 27];
+#   endif
+    }
+}
+
+/*=====    Local Constants   =====*/
+static const unsigned BIT_mask[] = {
+    0,          1,         3,         7,         0xF,       0x1F,
+    0x3F,       0x7F,      0xFF,      0x1FF,     0x3FF,     0x7FF,
+    0xFFF,      0x1FFF,    0x3FFF,    0x7FFF,    0xFFFF,    0x1FFFF,
+    0x3FFFF,    0x7FFFF,   0xFFFFF,   0x1FFFFF,  0x3FFFFF,  0x7FFFFF,
+    0xFFFFFF,   0x1FFFFFF, 0x3FFFFFF, 0x7FFFFFF, 0xFFFFFFF, 0x1FFFFFFF,
+    0x3FFFFFFF, 0x7FFFFFFF}; /* up to 31 bits */
+#define BIT_MASK_SIZE (sizeof(BIT_mask) / sizeof(BIT_mask[0]))
+
+/*-**************************************************************
+*  bitStream encoding
+****************************************************************/
+/*! BIT_initCStream() :
+ *  `dstCapacity` must be > sizeof(size_t)
+ *  @return : 0 if success,
+ *            otherwise an error code (can be tested using ERR_isError()) */
+MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC,
+                                  void* startPtr, size_t dstCapacity)
+{
+    bitC->bitContainer = 0;
+    bitC->bitPos = 0;
+    bitC->startPtr = (char*)startPtr;
+    bitC->ptr = bitC->startPtr;
+    bitC->endPtr = bitC->startPtr + dstCapacity - sizeof(bitC->bitContainer);
+    if (dstCapacity <= sizeof(bitC->bitContainer)) return ERROR(dstSize_tooSmall);
+    return 0;
+}
+
+/*! BIT_addBits() :
+ *  can add up to 31 bits into `bitC`.
+ *  Note : does not check for register overflow ! */
+MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC,
+                            size_t value, unsigned nbBits)
+{
+    DEBUG_STATIC_ASSERT(BIT_MASK_SIZE == 32);
+    assert(nbBits < BIT_MASK_SIZE);
+    assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8);
+    bitC->bitContainer |= (value & BIT_mask[nbBits]) << bitC->bitPos;
+    bitC->bitPos += nbBits;
+}
+
+/*! BIT_addBitsFast() :
+ *  works only if `value` is _clean_,
+ *  meaning all high bits above nbBits are 0 */
+MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC,
+                                size_t value, unsigned nbBits)
+{
+    assert((value>>nbBits) == 0);
+    assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8);
+    bitC->bitContainer |= value << bitC->bitPos;
+    bitC->bitPos += nbBits;
+}
+
+/*! BIT_flushBitsFast() :
+ *  assumption : bitContainer has not overflowed
+ *  unsafe version; does not check buffer overflow */
+MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC)
+{
+    size_t const nbBytes = bitC->bitPos >> 3;
+    assert(bitC->bitPos < sizeof(bitC->bitContainer) * 8);
+    assert(bitC->ptr <= bitC->endPtr);
+    MEM_writeLEST(bitC->ptr, bitC->bitContainer);
+    bitC->ptr += nbBytes;
+    bitC->bitPos &= 7;
+    bitC->bitContainer >>= nbBytes*8;
+}
+
+/*! BIT_flushBits() :
+ *  assumption : bitContainer has not overflowed
+ *  safe version; check for buffer overflow, and prevents it.
+ *  note : does not signal buffer overflow.
+ *  overflow will be revealed later on using BIT_closeCStream() */
+MEM_STATIC void BIT_flushBits(BIT_CStream_t* bitC)
+{
+    size_t const nbBytes = bitC->bitPos >> 3;
+    assert(bitC->bitPos < sizeof(bitC->bitContainer) * 8);
+    assert(bitC->ptr <= bitC->endPtr);
+    MEM_writeLEST(bitC->ptr, bitC->bitContainer);
+    bitC->ptr += nbBytes;
+    if (bitC->ptr > bitC->endPtr) bitC->ptr = bitC->endPtr;
+    bitC->bitPos &= 7;
+    bitC->bitContainer >>= nbBytes*8;
+}
+
+/*! BIT_closeCStream() :
+ *  @return : size of CStream, in bytes,
+ *            or 0 if it could not fit into dstBuffer */
+MEM_STATIC size_t BIT_closeCStream(BIT_CStream_t* bitC)
+{
+    BIT_addBitsFast(bitC, 1, 1);   /* endMark */
+    BIT_flushBits(bitC);
+    if (bitC->ptr >= bitC->endPtr) return 0; /* overflow detected */
+    return (bitC->ptr - bitC->startPtr) + (bitC->bitPos > 0);
+}
+
+
+/*-********************************************************
+*  bitStream decoding
+**********************************************************/
+/*! BIT_initDStream() :
+ *  Initialize a BIT_DStream_t.
+ * `bitD` : a pointer to an already allocated BIT_DStream_t structure.
+ * `srcSize` must be the *exact* size of the bitStream, in bytes.
+ * @return : size of stream (== srcSize), or an errorCode if a problem is detected
+ */
+MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize)
+{
+    if (srcSize < 1) { ZSTD_memset(bitD, 0, sizeof(*bitD)); return ERROR(srcSize_wrong); }
+
+    bitD->start = (const char*)srcBuffer;
+    bitD->limitPtr = bitD->start + sizeof(bitD->bitContainer);
+
+    if (srcSize >=  sizeof(bitD->bitContainer)) {  /* normal case */
+        bitD->ptr   = (const char*)srcBuffer + srcSize - sizeof(bitD->bitContainer);
+        bitD->bitContainer = MEM_readLEST(bitD->ptr);
+        { BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];
+          bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0;  /* ensures bitsConsumed is always set */
+          if (lastByte == 0) return ERROR(GENERIC); /* endMark not present */ }
+    } else {
+        bitD->ptr   = bitD->start;
+        bitD->bitContainer = *(const BYTE*)(bitD->start);
+        switch(srcSize)
+        {
+        case 7: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[6]) << (sizeof(bitD->bitContainer)*8 - 16);
+                ZSTD_FALLTHROUGH;
+
+        case 6: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[5]) << (sizeof(bitD->bitContainer)*8 - 24);
+                ZSTD_FALLTHROUGH;
+
+        case 5: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[4]) << (sizeof(bitD->bitContainer)*8 - 32);
+                ZSTD_FALLTHROUGH;
+
+        case 4: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[3]) << 24;
+                ZSTD_FALLTHROUGH;
+
+        case 3: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[2]) << 16;
+                ZSTD_FALLTHROUGH;
+
+        case 2: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[1]) <<  8;
+                ZSTD_FALLTHROUGH;
+
+        default: break;
+        }
+        {   BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];
+            bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0;
+            if (lastByte == 0) return ERROR(corruption_detected);  /* endMark not present */
+        }
+        bitD->bitsConsumed += (U32)(sizeof(bitD->bitContainer) - srcSize)*8;
+    }
+
+    return srcSize;
+}
+
+MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getUpperBits(size_t bitContainer, U32 const start)
+{
+    return bitContainer >> start;
+}
+
+MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getMiddleBits(size_t bitContainer, U32 const start, U32 const nbBits)
+{
+    U32 const regMask = sizeof(bitContainer)*8 - 1;
+    /* if start > regMask, bitstream is corrupted, and result is undefined */
+    assert(nbBits < BIT_MASK_SIZE);
+    return (bitContainer >> (start & regMask)) & BIT_mask[nbBits];
+}
+
+MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits)
+{
+    assert(nbBits < BIT_MASK_SIZE);
+    return bitContainer & BIT_mask[nbBits];
+}
+
+/*! BIT_lookBits() :
+ *  Provides next n bits from local register.
+ *  local register is not modified.
+ *  On 32-bits, maxNbBits==24.
+ *  On 64-bits, maxNbBits==56.
+ * @return : value extracted */
+MEM_STATIC  FORCE_INLINE_ATTR size_t BIT_lookBits(const BIT_DStream_t*  bitD, U32 nbBits)
+{
+    /* arbitrate between double-shift and shift+mask */
+#if 1
+    /* if bitD->bitsConsumed + nbBits > sizeof(bitD->bitContainer)*8,
+     * bitstream is likely corrupted, and result is undefined */
+    return BIT_getMiddleBits(bitD->bitContainer, (sizeof(bitD->bitContainer)*8) - bitD->bitsConsumed - nbBits, nbBits);
+#else
+    /* this code path is slower on my os-x laptop */
+    U32 const regMask = sizeof(bitD->bitContainer)*8 - 1;
+    return ((bitD->bitContainer << (bitD->bitsConsumed & regMask)) >> 1) >> ((regMask-nbBits) & regMask);
+#endif
+}
+
+/*! BIT_lookBitsFast() :
+ *  unsafe version; only works if nbBits >= 1 */
+MEM_STATIC size_t BIT_lookBitsFast(const BIT_DStream_t* bitD, U32 nbBits)
+{
+    U32 const regMask = sizeof(bitD->bitContainer)*8 - 1;
+    assert(nbBits >= 1);
+    return (bitD->bitContainer << (bitD->bitsConsumed & regMask)) >> (((regMask+1)-nbBits) & regMask);
+}
+
+MEM_STATIC FORCE_INLINE_ATTR void BIT_skipBits(BIT_DStream_t* bitD, U32 nbBits)
+{
+    bitD->bitsConsumed += nbBits;
+}
+
+/*! BIT_readBits() :
+ *  Read (consume) next n bits from local register and update.
+ *  Pay attention to not read more than nbBits contained into local register.
+ * @return : extracted value. */
+MEM_STATIC FORCE_INLINE_ATTR size_t BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits)
+{
+    size_t const value = BIT_lookBits(bitD, nbBits);
+    BIT_skipBits(bitD, nbBits);
+    return value;
+}
+
+/*! BIT_readBitsFast() :
+ *  unsafe version; only works only if nbBits >= 1 */
+MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits)
+{
+    size_t const value = BIT_lookBitsFast(bitD, nbBits);
+    assert(nbBits >= 1);
+    BIT_skipBits(bitD, nbBits);
+    return value;
+}
+
+/*! BIT_reloadDStreamFast() :
+ *  Similar to BIT_reloadDStream(), but with two differences:
+ *  1. bitsConsumed <= sizeof(bitD->bitContainer)*8 must hold!
+ *  2. Returns BIT_DStream_overflow when bitD->ptr < bitD->limitPtr, at this
+ *     point you must use BIT_reloadDStream() to reload.
+ */
+MEM_STATIC BIT_DStream_status BIT_reloadDStreamFast(BIT_DStream_t* bitD)
+{
+    if (UNLIKELY(bitD->ptr < bitD->limitPtr))
+        return BIT_DStream_overflow;
+    assert(bitD->bitsConsumed <= sizeof(bitD->bitContainer)*8);
+    bitD->ptr -= bitD->bitsConsumed >> 3;
+    bitD->bitsConsumed &= 7;
+    bitD->bitContainer = MEM_readLEST(bitD->ptr);
+    return BIT_DStream_unfinished;
+}
+
+/*! BIT_reloadDStream() :
+ *  Refill `bitD` from buffer previously set in BIT_initDStream() .
+ *  This function is safe, it guarantees it will not read beyond src buffer.
+ * @return : status of `BIT_DStream_t` internal register.
+ *           when status == BIT_DStream_unfinished, internal register is filled with at least 25 or 57 bits */
+MEM_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD)
+{
+    if (bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8))  /* overflow detected, like end of stream */
+        return BIT_DStream_overflow;
+
+    if (bitD->ptr >= bitD->limitPtr) {
+        return BIT_reloadDStreamFast(bitD);
+    }
+    if (bitD->ptr == bitD->start) {
+        if (bitD->bitsConsumed < sizeof(bitD->bitContainer)*8) return BIT_DStream_endOfBuffer;
+        return BIT_DStream_completed;
+    }
+    /* start < ptr < limitPtr */
+    {   U32 nbBytes = bitD->bitsConsumed >> 3;
+        BIT_DStream_status result = BIT_DStream_unfinished;
+        if (bitD->ptr - nbBytes < bitD->start) {
+            nbBytes = (U32)(bitD->ptr - bitD->start);  /* ptr > start */
+            result = BIT_DStream_endOfBuffer;
+        }
+        bitD->ptr -= nbBytes;
+        bitD->bitsConsumed -= nbBytes*8;
+        bitD->bitContainer = MEM_readLEST(bitD->ptr);   /* reminder : srcSize > sizeof(bitD->bitContainer), otherwise bitD->ptr == bitD->start */
+        return result;
+    }
+}
+
+/*! BIT_endOfDStream() :
+ * @return : 1 if DStream has _exactly_ reached its end (all bits consumed).
+ */
+MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* DStream)
+{
+    return ((DStream->ptr == DStream->start) && (DStream->bitsConsumed == sizeof(DStream->bitContainer)*8));
+}
+
+
+#endif /* BITSTREAM_H_MODULE */
diff --git a/lib/zstd/common/compiler.h b/lib/zstd/common/compiler.h
new file mode 100644
index 0000000..a1a051e
--- /dev/null
+++ b/lib/zstd/common/compiler.h
@@ -0,0 +1,170 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_COMPILER_H
+#define ZSTD_COMPILER_H
+
+/*-*******************************************************
+*  Compiler specifics
+*********************************************************/
+/* force inlining */
+
+#if (defined(__GNUC__) && !defined(__STRICT_ANSI__)) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L   /* C99 */
+#  define INLINE_KEYWORD inline
+#else
+#  define INLINE_KEYWORD
+#endif
+
+#define FORCE_INLINE_ATTR __attribute__((always_inline))
+
+
+/*
+  On MSVC qsort requires that functions passed into it use the __cdecl calling conversion(CC).
+  This explictly marks such functions as __cdecl so that the code will still compile
+  if a CC other than __cdecl has been made the default.
+*/
+#define WIN_CDECL
+
+/*
+ * FORCE_INLINE_TEMPLATE is used to define C "templates", which take constant
+ * parameters. They must be inlined for the compiler to eliminate the constant
+ * branches.
+ */
+#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR
+/*
+ * HINT_INLINE is used to help the compiler generate better code. It is *not*
+ * used for "templates", so it can be tweaked based on the compilers
+ * performance.
+ *
+ * gcc-4.8 and gcc-4.9 have been shown to benefit from leaving off the
+ * always_inline attribute.
+ *
+ * clang up to 5.0.0 (trunk) benefit tremendously from the always_inline
+ * attribute.
+ */
+#if !defined(__clang__) && defined(__GNUC__) && __GNUC__ >= 4 && __GNUC_MINOR__ >= 8 && __GNUC__ < 5
+#  define HINT_INLINE static INLINE_KEYWORD
+#else
+#  define HINT_INLINE static INLINE_KEYWORD FORCE_INLINE_ATTR
+#endif
+
+/* UNUSED_ATTR tells the compiler it is okay if the function is unused. */
+#define UNUSED_ATTR __attribute__((unused))
+
+/* force no inlining */
+#define FORCE_NOINLINE static __attribute__((__noinline__))
+
+
+/* target attribute */
+#ifndef __has_attribute
+  #define __has_attribute(x) 0  /* Compatibility with non-clang compilers. */
+#endif
+#define TARGET_ATTRIBUTE(target) __attribute__((__target__(target)))
+
+/* Enable runtime BMI2 dispatch based on the CPU.
+ * Enabled for clang & gcc >=4.8 on x86 when BMI2 isn't enabled by default.
+ */
+#ifndef DYNAMIC_BMI2
+  #if ((defined(__clang__) && __has_attribute(__target__)) \
+      || (defined(__GNUC__) \
+          && (__GNUC__ >= 5 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 8)))) \
+      && (defined(__x86_64__) || defined(_M_X86)) \
+      && !defined(__BMI2__)
+  #  define DYNAMIC_BMI2 1
+  #else
+  #  define DYNAMIC_BMI2 0
+  #endif
+#endif
+
+/* prefetch
+ * can be disabled, by declaring NO_PREFETCH build macro */
+#if ( (__GNUC__ >= 4) || ( (__GNUC__ == 3) && (__GNUC_MINOR__ >= 1) ) )
+#  define PREFETCH_L1(ptr)  __builtin_prefetch((ptr), 0 /* rw==read */, 3 /* locality */)
+#  define PREFETCH_L2(ptr)  __builtin_prefetch((ptr), 0 /* rw==read */, 2 /* locality */)
+#elif defined(__aarch64__)
+#  define PREFETCH_L1(ptr)  __asm__ __volatile__("prfm pldl1keep, %0" ::"Q"(*(ptr)))
+#  define PREFETCH_L2(ptr)  __asm__ __volatile__("prfm pldl2keep, %0" ::"Q"(*(ptr)))
+#else
+#  define PREFETCH_L1(ptr) (void)(ptr)  /* disabled */
+#  define PREFETCH_L2(ptr) (void)(ptr)  /* disabled */
+#endif  /* NO_PREFETCH */
+
+#define CACHELINE_SIZE 64
+
+#define PREFETCH_AREA(p, s)  {            \
+    const char* const _ptr = (const char*)(p);  \
+    size_t const _size = (size_t)(s);     \
+    size_t _pos;                          \
+    for (_pos=0; _pos<_size; _pos+=CACHELINE_SIZE) {  \
+        PREFETCH_L2(_ptr + _pos);         \
+    }                                     \
+}
+
+/* vectorization
+ * older GCC (pre gcc-4.3 picked as the cutoff) uses a different syntax */
+#if !defined(__INTEL_COMPILER) && !defined(__clang__) && defined(__GNUC__)
+#  if (__GNUC__ == 4 && __GNUC_MINOR__ > 3) || (__GNUC__ >= 5)
+#    define DONT_VECTORIZE __attribute__((optimize("no-tree-vectorize")))
+#  else
+#    define DONT_VECTORIZE _Pragma("GCC optimize(\"no-tree-vectorize\")")
+#  endif
+#else
+#  define DONT_VECTORIZE
+#endif
+
+/* Tell the compiler that a branch is likely or unlikely.
+ * Only use these macros if it causes the compiler to generate better code.
+ * If you can remove a LIKELY/UNLIKELY annotation without speed changes in gcc
+ * and clang, please do.
+ */
+#define LIKELY(x) (__builtin_expect((x), 1))
+#define UNLIKELY(x) (__builtin_expect((x), 0))
+
+/* disable warnings */
+
+/*Like DYNAMIC_BMI2 but for compile time determination of BMI2 support*/
+
+
+/* compat. with non-clang compilers */
+#ifndef __has_builtin
+#  define __has_builtin(x) 0
+#endif
+
+/* compat. with non-clang compilers */
+#ifndef __has_feature
+#  define __has_feature(x) 0
+#endif
+
+/* C-language Attributes are added in C23. */
+#if defined(__STDC_VERSION__) && (__STDC_VERSION__ > 201710L) && defined(__has_c_attribute)
+# define ZSTD_HAS_C_ATTRIBUTE(x) __has_c_attribute(x)
+#else
+# define ZSTD_HAS_C_ATTRIBUTE(x) 0
+#endif
+
+/* Only use C++ attributes in C++. Some compilers report support for C++
+ * attributes when compiling with C.
+ */
+#define ZSTD_HAS_CPP_ATTRIBUTE(x) 0
+
+/* Define ZSTD_FALLTHROUGH macro for annotating switch case with the 'fallthrough' attribute.
+ * - C23: https://en.cppreference.com/w/c/language/attributes/fallthrough
+ * - CPP17: https://en.cppreference.com/w/cpp/language/attributes/fallthrough
+ * - Else: __attribute__((__fallthrough__))
+ */
+#define ZSTD_FALLTHROUGH fallthrough
+
+/* detects whether we are being compiled under msan */
+
+
+/* detects whether we are being compiled under asan */
+
+
+#endif /* ZSTD_COMPILER_H */
diff --git a/lib/zstd/common/cpu.h b/lib/zstd/common/cpu.h
new file mode 100644
index 0000000..0db7b42
--- /dev/null
+++ b/lib/zstd/common/cpu.h
@@ -0,0 +1,194 @@
+/*
+ * Copyright (c) Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_COMMON_CPU_H
+#define ZSTD_COMMON_CPU_H
+
+/*
+ * Implementation taken from folly/CpuId.h
+ * https://github.com/facebook/folly/blob/master/folly/CpuId.h
+ */
+
+#include "mem.h"
+
+
+typedef struct {
+    U32 f1c;
+    U32 f1d;
+    U32 f7b;
+    U32 f7c;
+} ZSTD_cpuid_t;
+
+MEM_STATIC ZSTD_cpuid_t ZSTD_cpuid(void) {
+    U32 f1c = 0;
+    U32 f1d = 0;
+    U32 f7b = 0;
+    U32 f7c = 0;
+#if defined(__i386__) && defined(__PIC__) && !defined(__clang__) && defined(__GNUC__)
+    /* The following block like the normal cpuid branch below, but gcc
+     * reserves ebx for use of its pic register so we must specially
+     * handle the save and restore to avoid clobbering the register
+     */
+    U32 n;
+    __asm__(
+        "pushl %%ebx\n\t"
+        "cpuid\n\t"
+        "popl %%ebx\n\t"
+        : "=a"(n)
+        : "a"(0)
+        : "ecx", "edx");
+    if (n >= 1) {
+      U32 f1a;
+      __asm__(
+          "pushl %%ebx\n\t"
+          "cpuid\n\t"
+          "popl %%ebx\n\t"
+          : "=a"(f1a), "=c"(f1c), "=d"(f1d)
+          : "a"(1));
+    }
+    if (n >= 7) {
+      __asm__(
+          "pushl %%ebx\n\t"
+          "cpuid\n\t"
+          "movl %%ebx, %%eax\n\t"
+          "popl %%ebx"
+          : "=a"(f7b), "=c"(f7c)
+          : "a"(7), "c"(0)
+          : "edx");
+    }
+#elif defined(__x86_64__) || defined(_M_X64) || defined(__i386__)
+    U32 n;
+    __asm__("cpuid" : "=a"(n) : "a"(0) : "ebx", "ecx", "edx");
+    if (n >= 1) {
+      U32 f1a;
+      __asm__("cpuid" : "=a"(f1a), "=c"(f1c), "=d"(f1d) : "a"(1) : "ebx");
+    }
+    if (n >= 7) {
+      U32 f7a;
+      __asm__("cpuid"
+              : "=a"(f7a), "=b"(f7b), "=c"(f7c)
+              : "a"(7), "c"(0)
+              : "edx");
+    }
+#endif
+    {
+        ZSTD_cpuid_t cpuid;
+        cpuid.f1c = f1c;
+        cpuid.f1d = f1d;
+        cpuid.f7b = f7b;
+        cpuid.f7c = f7c;
+        return cpuid;
+    }
+}
+
+#define X(name, r, bit)                                                        \
+  MEM_STATIC int ZSTD_cpuid_##name(ZSTD_cpuid_t const cpuid) {                 \
+    return ((cpuid.r) & (1U << bit)) != 0;                                     \
+  }
+
+/* cpuid(1): Processor Info and Feature Bits. */
+#define C(name, bit) X(name, f1c, bit)
+  C(sse3, 0)
+  C(pclmuldq, 1)
+  C(dtes64, 2)
+  C(monitor, 3)
+  C(dscpl, 4)
+  C(vmx, 5)
+  C(smx, 6)
+  C(eist, 7)
+  C(tm2, 8)
+  C(ssse3, 9)
+  C(cnxtid, 10)
+  C(fma, 12)
+  C(cx16, 13)
+  C(xtpr, 14)
+  C(pdcm, 15)
+  C(pcid, 17)
+  C(dca, 18)
+  C(sse41, 19)
+  C(sse42, 20)
+  C(x2apic, 21)
+  C(movbe, 22)
+  C(popcnt, 23)
+  C(tscdeadline, 24)
+  C(aes, 25)
+  C(xsave, 26)
+  C(osxsave, 27)
+  C(avx, 28)
+  C(f16c, 29)
+  C(rdrand, 30)
+#undef C
+#define D(name, bit) X(name, f1d, bit)
+  D(fpu, 0)
+  D(vme, 1)
+  D(de, 2)
+  D(pse, 3)
+  D(tsc, 4)
+  D(msr, 5)
+  D(pae, 6)
+  D(mce, 7)
+  D(cx8, 8)
+  D(apic, 9)
+  D(sep, 11)
+  D(mtrr, 12)
+  D(pge, 13)
+  D(mca, 14)
+  D(cmov, 15)
+  D(pat, 16)
+  D(pse36, 17)
+  D(psn, 18)
+  D(clfsh, 19)
+  D(ds, 21)
+  D(acpi, 22)
+  D(mmx, 23)
+  D(fxsr, 24)
+  D(sse, 25)
+  D(sse2, 26)
+  D(ss, 27)
+  D(htt, 28)
+  D(tm, 29)
+  D(pbe, 31)
+#undef D
+
+/* cpuid(7): Extended Features. */
+#define B(name, bit) X(name, f7b, bit)
+  B(bmi1, 3)
+  B(hle, 4)
+  B(avx2, 5)
+  B(smep, 7)
+  B(bmi2, 8)
+  B(erms, 9)
+  B(invpcid, 10)
+  B(rtm, 11)
+  B(mpx, 14)
+  B(avx512f, 16)
+  B(avx512dq, 17)
+  B(rdseed, 18)
+  B(adx, 19)
+  B(smap, 20)
+  B(avx512ifma, 21)
+  B(pcommit, 22)
+  B(clflushopt, 23)
+  B(clwb, 24)
+  B(avx512pf, 26)
+  B(avx512er, 27)
+  B(avx512cd, 28)
+  B(sha, 29)
+  B(avx512bw, 30)
+  B(avx512vl, 31)
+#undef B
+#define C(name, bit) X(name, f7c, bit)
+  C(prefetchwt1, 0)
+  C(avx512vbmi, 1)
+#undef C
+
+#undef X
+
+#endif /* ZSTD_COMMON_CPU_H */
diff --git a/lib/zstd/common/debug.c b/lib/zstd/common/debug.c
new file mode 100644
index 0000000..bb863c9
--- /dev/null
+++ b/lib/zstd/common/debug.c
@@ -0,0 +1,24 @@
+/* ******************************************************************
+ * debug
+ * Part of FSE library
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ * You can contact the author at :
+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+
+/*
+ * This module only hosts one global variable
+ * which can be used to dynamically influence the verbosity of traces,
+ * such as DEBUGLOG and RAWLOG
+ */
+
+#include "debug.h"
+
+int g_debuglevel = DEBUGLEVEL;
diff --git a/lib/zstd/common/debug.h b/lib/zstd/common/debug.h
new file mode 100644
index 0000000..6dd88d1
--- /dev/null
+++ b/lib/zstd/common/debug.h
@@ -0,0 +1,101 @@
+/* ******************************************************************
+ * debug
+ * Part of FSE library
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ * You can contact the author at :
+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+
+/*
+ * The purpose of this header is to enable debug functions.
+ * They regroup assert(), DEBUGLOG() and RAWLOG() for run-time,
+ * and DEBUG_STATIC_ASSERT() for compile-time.
+ *
+ * By default, DEBUGLEVEL==0, which means run-time debug is disabled.
+ *
+ * Level 1 enables assert() only.
+ * Starting level 2, traces can be generated and pushed to stderr.
+ * The higher the level, the more verbose the traces.
+ *
+ * It's possible to dynamically adjust level using variable g_debug_level,
+ * which is only declared if DEBUGLEVEL>=2,
+ * and is a global variable, not multi-thread protected (use with care)
+ */
+
+#ifndef DEBUG_H_12987983217
+#define DEBUG_H_12987983217
+
+
+
+/* static assert is triggered at compile time, leaving no runtime artefact.
+ * static assert only works with compile-time constants.
+ * Also, this variant can only be used inside a function. */
+#define DEBUG_STATIC_ASSERT(c) (void)sizeof(char[(c) ? 1 : -1])
+
+
+/* DEBUGLEVEL is expected to be defined externally,
+ * typically through compiler command line.
+ * Value must be a number. */
+#ifndef DEBUGLEVEL
+#  define DEBUGLEVEL 0
+#endif
+
+
+/* recommended values for DEBUGLEVEL :
+ * 0 : release mode, no debug, all run-time checks disabled
+ * 1 : enables assert() only, no display
+ * 2 : reserved, for currently active debug path
+ * 3 : events once per object lifetime (CCtx, CDict, etc.)
+ * 4 : events once per frame
+ * 5 : events once per block
+ * 6 : events once per sequence (verbose)
+ * 7+: events at every position (*very* verbose)
+ *
+ * It's generally inconvenient to output traces > 5.
+ * In which case, it's possible to selectively trigger high verbosity levels
+ * by modifying g_debug_level.
+ */
+
+#if (DEBUGLEVEL>=1)
+#  define ZSTD_DEPS_NEED_ASSERT
+#  include "zstd_deps.h"
+#else
+#  ifndef assert   /* assert may be already defined, due to prior #include <assert.h> */
+#    define assert(condition) ((void)0)   /* disable assert (default) */
+#  endif
+#endif
+
+#if (DEBUGLEVEL>=2)
+#  define ZSTD_DEPS_NEED_IO
+#  include "zstd_deps.h"
+extern int g_debuglevel; /* the variable is only declared,
+                            it actually lives in debug.c,
+                            and is shared by the whole process.
+                            It's not thread-safe.
+                            It's useful when enabling very verbose levels
+                            on selective conditions (such as position in src) */
+
+#  define RAWLOG(l, ...) {                                       \
+                if (l<=g_debuglevel) {                           \
+                    ZSTD_DEBUG_PRINT(__VA_ARGS__);               \
+            }   }
+#  define DEBUGLOG(l, ...) {                                     \
+                if (l<=g_debuglevel) {                           \
+                    ZSTD_DEBUG_PRINT(__FILE__ ": " __VA_ARGS__); \
+                    ZSTD_DEBUG_PRINT(" \n");                     \
+            }   }
+#else
+#  define RAWLOG(l, ...)      {}    /* disabled */
+#  define DEBUGLOG(l, ...)    {}    /* disabled */
+#endif
+
+
+
+#endif /* DEBUG_H_12987983217 */
diff --git a/lib/zstd/common/entropy_common.c b/lib/zstd/common/entropy_common.c
new file mode 100644
index 0000000..53b47a2
--- /dev/null
+++ b/lib/zstd/common/entropy_common.c
@@ -0,0 +1,357 @@
+/* ******************************************************************
+ * Common functions of New Generation Entropy library
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ *  You can contact the author at :
+ *  - FSE+HUF source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *  - Public forum : https://groups.google.com/forum/#!forum/lz4c
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+/* *************************************
+*  Dependencies
+***************************************/
+#include "mem.h"
+#include "error_private.h"       /* ERR_*, ERROR */
+#define FSE_STATIC_LINKING_ONLY  /* FSE_MIN_TABLELOG */
+#include "fse.h"
+#define HUF_STATIC_LINKING_ONLY  /* HUF_TABLELOG_ABSOLUTEMAX */
+#include "huf.h"
+
+
+/*===   Version   ===*/
+unsigned FSE_versionNumber(void) { return FSE_VERSION_NUMBER; }
+
+
+/*===   Error Management   ===*/
+unsigned FSE_isError(size_t code) { return ERR_isError(code); }
+const char* FSE_getErrorName(size_t code) { return ERR_getErrorName(code); }
+
+unsigned HUF_isError(size_t code) { return ERR_isError(code); }
+const char* HUF_getErrorName(size_t code) { return ERR_getErrorName(code); }
+
+
+/*-**************************************************************
+*  FSE NCount encoding-decoding
+****************************************************************/
+static U32 FSE_ctz(U32 val)
+{
+    assert(val != 0);
+    {
+#   if (__GNUC__ >= 3)   /* GCC Intrinsic */
+        return __builtin_ctz(val);
+#   else   /* Software version */
+        U32 count = 0;
+        while ((val & 1) == 0) {
+            val >>= 1;
+            ++count;
+        }
+        return count;
+#   endif
+    }
+}
+
+FORCE_INLINE_TEMPLATE
+size_t FSE_readNCount_body(short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
+                           const void* headerBuffer, size_t hbSize)
+{
+    const BYTE* const istart = (const BYTE*) headerBuffer;
+    const BYTE* const iend = istart + hbSize;
+    const BYTE* ip = istart;
+    int nbBits;
+    int remaining;
+    int threshold;
+    U32 bitStream;
+    int bitCount;
+    unsigned charnum = 0;
+    unsigned const maxSV1 = *maxSVPtr + 1;
+    int previous0 = 0;
+
+    if (hbSize < 8) {
+        /* This function only works when hbSize >= 8 */
+        char buffer[8] = {0};
+        ZSTD_memcpy(buffer, headerBuffer, hbSize);
+        {   size_t const countSize = FSE_readNCount(normalizedCounter, maxSVPtr, tableLogPtr,
+                                                    buffer, sizeof(buffer));
+            if (FSE_isError(countSize)) return countSize;
+            if (countSize > hbSize) return ERROR(corruption_detected);
+            return countSize;
+    }   }
+    assert(hbSize >= 8);
+
+    /* init */
+    ZSTD_memset(normalizedCounter, 0, (*maxSVPtr+1) * sizeof(normalizedCounter[0]));   /* all symbols not present in NCount have a frequency of 0 */
+    bitStream = MEM_readLE32(ip);
+    nbBits = (bitStream & 0xF) + FSE_MIN_TABLELOG;   /* extract tableLog */
+    if (nbBits > FSE_TABLELOG_ABSOLUTE_MAX) return ERROR(tableLog_tooLarge);
+    bitStream >>= 4;
+    bitCount = 4;
+    *tableLogPtr = nbBits;
+    remaining = (1<<nbBits)+1;
+    threshold = 1<<nbBits;
+    nbBits++;
+
+    for (;;) {
+        if (previous0) {
+            /* Count the number of repeats. Each time the
+             * 2-bit repeat code is 0b11 there is another
+             * repeat.
+             * Avoid UB by setting the high bit to 1.
+             */
+            int repeats = FSE_ctz(~bitStream | 0x80000000) >> 1;
+            while (repeats >= 12) {
+                charnum += 3 * 12;
+                if (LIKELY(ip <= iend-7)) {
+                    ip += 3;
+                } else {
+                    bitCount -= (int)(8 * (iend - 7 - ip));
+                    bitCount &= 31;
+                    ip = iend - 4;
+                }
+                bitStream = MEM_readLE32(ip) >> bitCount;
+                repeats = FSE_ctz(~bitStream | 0x80000000) >> 1;
+            }
+            charnum += 3 * repeats;
+            bitStream >>= 2 * repeats;
+            bitCount += 2 * repeats;
+
+            /* Add the final repeat which isn't 0b11. */
+            assert((bitStream & 3) < 3);
+            charnum += bitStream & 3;
+            bitCount += 2;
+
+            /* This is an error, but break and return an error
+             * at the end, because returning out of a loop makes
+             * it harder for the compiler to optimize.
+             */
+            if (charnum >= maxSV1) break;
+
+            /* We don't need to set the normalized count to 0
+             * because we already memset the whole buffer to 0.
+             */
+
+            if (LIKELY(ip <= iend-7) || (ip + (bitCount>>3) <= iend-4)) {
+                assert((bitCount >> 3) <= 3); /* For first condition to work */
+                ip += bitCount>>3;
+                bitCount &= 7;
+            } else {
+                bitCount -= (int)(8 * (iend - 4 - ip));
+                bitCount &= 31;
+                ip = iend - 4;
+            }
+            bitStream = MEM_readLE32(ip) >> bitCount;
+        }
+        {
+            int const max = (2*threshold-1) - remaining;
+            int count;
+
+            if ((bitStream & (threshold-1)) < (U32)max) {
+                count = bitStream & (threshold-1);
+                bitCount += nbBits-1;
+            } else {
+                count = bitStream & (2*threshold-1);
+                if (count >= threshold) count -= max;
+                bitCount += nbBits;
+            }
+
+            count--;   /* extra accuracy */
+            /* When it matters (small blocks), this is a
+             * predictable branch, because we don't use -1.
+             */
+            if (count >= 0) {
+                remaining -= count;
+            } else {
+                assert(count == -1);
+                remaining += count;
+            }
+            normalizedCounter[charnum++] = (short)count;
+            previous0 = !count;
+
+            assert(threshold > 1);
+            if (remaining < threshold) {
+                /* This branch can be folded into the
+                 * threshold update condition because we
+                 * know that threshold > 1.
+                 */
+                if (remaining <= 1) break;
+                nbBits = BIT_highbit32(remaining) + 1;
+                threshold = 1 << (nbBits - 1);
+            }
+            if (charnum >= maxSV1) break;
+
+            if (LIKELY(ip <= iend-7) || (ip + (bitCount>>3) <= iend-4)) {
+                ip += bitCount>>3;
+                bitCount &= 7;
+            } else {
+                bitCount -= (int)(8 * (iend - 4 - ip));
+                bitCount &= 31;
+                ip = iend - 4;
+            }
+            bitStream = MEM_readLE32(ip) >> bitCount;
+    }   }
+    if (remaining != 1) return ERROR(corruption_detected);
+    /* Only possible when there are too many zeros. */
+    if (charnum > maxSV1) return ERROR(maxSymbolValue_tooSmall);
+    if (bitCount > 32) return ERROR(corruption_detected);
+    *maxSVPtr = charnum-1;
+
+    ip += (bitCount+7)>>3;
+    return ip-istart;
+}
+
+/* Avoids the FORCE_INLINE of the _body() function. */
+static size_t FSE_readNCount_body_default(
+        short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
+        const void* headerBuffer, size_t hbSize)
+{
+    return FSE_readNCount_body(normalizedCounter, maxSVPtr, tableLogPtr, headerBuffer, hbSize);
+}
+
+#if DYNAMIC_BMI2
+TARGET_ATTRIBUTE("bmi2") static size_t FSE_readNCount_body_bmi2(
+        short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
+        const void* headerBuffer, size_t hbSize)
+{
+    return FSE_readNCount_body(normalizedCounter, maxSVPtr, tableLogPtr, headerBuffer, hbSize);
+}
+#endif
+
+size_t FSE_readNCount_bmi2(
+        short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
+        const void* headerBuffer, size_t hbSize, int bmi2)
+{
+#if DYNAMIC_BMI2
+    if (bmi2) {
+        return FSE_readNCount_body_bmi2(normalizedCounter, maxSVPtr, tableLogPtr, headerBuffer, hbSize);
+    }
+#endif
+    (void)bmi2;
+    return FSE_readNCount_body_default(normalizedCounter, maxSVPtr, tableLogPtr, headerBuffer, hbSize);
+}
+
+size_t FSE_readNCount(
+        short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
+        const void* headerBuffer, size_t hbSize)
+{
+    return FSE_readNCount_bmi2(normalizedCounter, maxSVPtr, tableLogPtr, headerBuffer, hbSize, /* bmi2 */ 0);
+}
+
+
+/*! HUF_readStats() :
+    Read compact Huffman tree, saved by HUF_writeCTable().
+    `huffWeight` is destination buffer.
+    `rankStats` is assumed to be a table of at least HUF_TABLELOG_MAX U32.
+    @return : size read from `src` , or an error Code .
+    Note : Needed by HUF_readCTable() and HUF_readDTableX?() .
+*/
+size_t HUF_readStats(BYTE* huffWeight, size_t hwSize, U32* rankStats,
+                     U32* nbSymbolsPtr, U32* tableLogPtr,
+                     const void* src, size_t srcSize)
+{
+    U32 wksp[HUF_READ_STATS_WORKSPACE_SIZE_U32];
+    return HUF_readStats_wksp(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, wksp, sizeof(wksp), /* bmi2 */ 0);
+}
+
+FORCE_INLINE_TEMPLATE size_t
+HUF_readStats_body(BYTE* huffWeight, size_t hwSize, U32* rankStats,
+                   U32* nbSymbolsPtr, U32* tableLogPtr,
+                   const void* src, size_t srcSize,
+                   void* workSpace, size_t wkspSize,
+                   int bmi2)
+{
+    U32 weightTotal;
+    const BYTE* ip = (const BYTE*) src;
+    size_t iSize;
+    size_t oSize;
+
+    if (!srcSize) return ERROR(srcSize_wrong);
+    iSize = ip[0];
+    /* ZSTD_memset(huffWeight, 0, hwSize);   *//* is not necessary, even though some analyzer complain ... */
+
+    if (iSize >= 128) {  /* special header */
+        oSize = iSize - 127;
+        iSize = ((oSize+1)/2);
+        if (iSize+1 > srcSize) return ERROR(srcSize_wrong);
+        if (oSize >= hwSize) return ERROR(corruption_detected);
+        ip += 1;
+        {   U32 n;
+            for (n=0; n<oSize; n+=2) {
+                huffWeight[n]   = ip[n/2] >> 4;
+                huffWeight[n+1] = ip[n/2] & 15;
+    }   }   }
+    else  {   /* header compressed with FSE (normal case) */
+        if (iSize+1 > srcSize) return ERROR(srcSize_wrong);
+        /* max (hwSize-1) values decoded, as last one is implied */
+        oSize = FSE_decompress_wksp_bmi2(huffWeight, hwSize-1, ip+1, iSize, 6, workSpace, wkspSize, bmi2);
+        if (FSE_isError(oSize)) return oSize;
+    }
+
+    /* collect weight stats */
+    ZSTD_memset(rankStats, 0, (HUF_TABLELOG_MAX + 1) * sizeof(U32));
+    weightTotal = 0;
+    {   U32 n; for (n=0; n<oSize; n++) {
+            if (huffWeight[n] >= HUF_TABLELOG_MAX) return ERROR(corruption_detected);
+            rankStats[huffWeight[n]]++;
+            weightTotal += (1 << huffWeight[n]) >> 1;
+    }   }
+    if (weightTotal == 0) return ERROR(corruption_detected);
+
+    /* get last non-null symbol weight (implied, total must be 2^n) */
+    {   U32 const tableLog = BIT_highbit32(weightTotal) + 1;
+        if (tableLog > HUF_TABLELOG_MAX) return ERROR(corruption_detected);
+        *tableLogPtr = tableLog;
+        /* determine last weight */
+        {   U32 const total = 1 << tableLog;
+            U32 const rest = total - weightTotal;
+            U32 const verif = 1 << BIT_highbit32(rest);
+            U32 const lastWeight = BIT_highbit32(rest) + 1;
+            if (verif != rest) return ERROR(corruption_detected);    /* last value must be a clean power of 2 */
+            huffWeight[oSize] = (BYTE)lastWeight;
+            rankStats[lastWeight]++;
+    }   }
+
+    /* check tree construction validity */
+    if ((rankStats[1] < 2) || (rankStats[1] & 1)) return ERROR(corruption_detected);   /* by construction : at least 2 elts of rank 1, must be even */
+
+    /* results */
+    *nbSymbolsPtr = (U32)(oSize+1);
+    return iSize+1;
+}
+
+/* Avoids the FORCE_INLINE of the _body() function. */
+static size_t HUF_readStats_body_default(BYTE* huffWeight, size_t hwSize, U32* rankStats,
+                     U32* nbSymbolsPtr, U32* tableLogPtr,
+                     const void* src, size_t srcSize,
+                     void* workSpace, size_t wkspSize)
+{
+    return HUF_readStats_body(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, workSpace, wkspSize, 0);
+}
+
+#if DYNAMIC_BMI2
+static TARGET_ATTRIBUTE("bmi2") size_t HUF_readStats_body_bmi2(BYTE* huffWeight, size_t hwSize, U32* rankStats,
+                     U32* nbSymbolsPtr, U32* tableLogPtr,
+                     const void* src, size_t srcSize,
+                     void* workSpace, size_t wkspSize)
+{
+    return HUF_readStats_body(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, workSpace, wkspSize, 1);
+}
+#endif
+
+size_t HUF_readStats_wksp(BYTE* huffWeight, size_t hwSize, U32* rankStats,
+                     U32* nbSymbolsPtr, U32* tableLogPtr,
+                     const void* src, size_t srcSize,
+                     void* workSpace, size_t wkspSize,
+                     int bmi2)
+{
+#if DYNAMIC_BMI2
+    if (bmi2) {
+        return HUF_readStats_body_bmi2(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, workSpace, wkspSize);
+    }
+#endif
+    (void)bmi2;
+    return HUF_readStats_body_default(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, workSpace, wkspSize);
+}
diff --git a/lib/zstd/common/error_private.c b/lib/zstd/common/error_private.c
new file mode 100644
index 0000000..6d1135f
--- /dev/null
+++ b/lib/zstd/common/error_private.c
@@ -0,0 +1,56 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/* The purpose of this file is to have a single list of error strings embedded in binary */
+
+#include "error_private.h"
+
+const char* ERR_getErrorString(ERR_enum code)
+{
+#ifdef ZSTD_STRIP_ERROR_STRINGS
+    (void)code;
+    return "Error strings stripped";
+#else
+    static const char* const notErrorCode = "Unspecified error code";
+    switch( code )
+    {
+    case PREFIX(no_error): return "No error detected";
+    case PREFIX(GENERIC):  return "Error (generic)";
+    case PREFIX(prefix_unknown): return "Unknown frame descriptor";
+    case PREFIX(version_unsupported): return "Version not supported";
+    case PREFIX(frameParameter_unsupported): return "Unsupported frame parameter";
+    case PREFIX(frameParameter_windowTooLarge): return "Frame requires too much memory for decoding";
+    case PREFIX(corruption_detected): return "Corrupted block detected";
+    case PREFIX(checksum_wrong): return "Restored data doesn't match checksum";
+    case PREFIX(parameter_unsupported): return "Unsupported parameter";
+    case PREFIX(parameter_outOfBound): return "Parameter is out of bound";
+    case PREFIX(init_missing): return "Context should be init first";
+    case PREFIX(memory_allocation): return "Allocation error : not enough memory";
+    case PREFIX(workSpace_tooSmall): return "workSpace buffer is not large enough";
+    case PREFIX(stage_wrong): return "Operation not authorized at current processing stage";
+    case PREFIX(tableLog_tooLarge): return "tableLog requires too much memory : unsupported";
+    case PREFIX(maxSymbolValue_tooLarge): return "Unsupported max Symbol Value : too large";
+    case PREFIX(maxSymbolValue_tooSmall): return "Specified maxSymbolValue is too small";
+    case PREFIX(dictionary_corrupted): return "Dictionary is corrupted";
+    case PREFIX(dictionary_wrong): return "Dictionary mismatch";
+    case PREFIX(dictionaryCreation_failed): return "Cannot create Dictionary from provided samples";
+    case PREFIX(dstSize_tooSmall): return "Destination buffer is too small";
+    case PREFIX(srcSize_wrong): return "Src size is incorrect";
+    case PREFIX(dstBuffer_null): return "Operation on NULL destination buffer";
+        /* following error codes are not stable and may be removed or changed in a future version */
+    case PREFIX(frameIndex_tooLarge): return "Frame index is too large";
+    case PREFIX(seekableIO): return "An I/O error occurred when reading/seeking";
+    case PREFIX(dstBuffer_wrong): return "Destination buffer is wrong";
+    case PREFIX(srcBuffer_wrong): return "Source buffer is wrong";
+    case PREFIX(maxCode):
+    default: return notErrorCode;
+    }
+#endif
+}
diff --git a/lib/zstd/common/error_private.h b/lib/zstd/common/error_private.h
new file mode 100644
index 0000000..d14e686
--- /dev/null
+++ b/lib/zstd/common/error_private.h
@@ -0,0 +1,66 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/* Note : this module is expected to remain private, do not expose it */
+
+#ifndef ERROR_H_MODULE
+#define ERROR_H_MODULE
+
+
+
+/* ****************************************
+*  Dependencies
+******************************************/
+#include "zstd_deps.h"    /* size_t */
+#include <linux/zstd_errors.h>  /* enum list */
+
+
+/* ****************************************
+*  Compiler-specific
+******************************************/
+#define ERR_STATIC static __attribute__((unused))
+
+
+/*-****************************************
+*  Customization (error_public.h)
+******************************************/
+typedef ZSTD_ErrorCode ERR_enum;
+#define PREFIX(name) ZSTD_error_##name
+
+
+/*-****************************************
+*  Error codes handling
+******************************************/
+#undef ERROR   /* already defined on Visual Studio */
+#define ERROR(name) ZSTD_ERROR(name)
+#define ZSTD_ERROR(name) ((size_t)-PREFIX(name))
+
+ERR_STATIC unsigned ERR_isError(size_t code) { return (code > ERROR(maxCode)); }
+
+ERR_STATIC ERR_enum ERR_getErrorCode(size_t code) { if (!ERR_isError(code)) return (ERR_enum)0; return (ERR_enum) (0-code); }
+
+/* check and forward error code */
+#define CHECK_V_F(e, f) size_t const e = f; if (ERR_isError(e)) return e
+#define CHECK_F(f)   { CHECK_V_F(_var_err__, f); }
+
+
+/*-****************************************
+*  Error Strings
+******************************************/
+
+const char* ERR_getErrorString(ERR_enum code);   /* error_private.c */
+
+ERR_STATIC const char* ERR_getErrorName(size_t code)
+{
+    return ERR_getErrorString(ERR_getErrorCode(code));
+}
+
+
+#endif /* ERROR_H_MODULE */
diff --git a/lib/zstd/common/fse.h b/lib/zstd/common/fse.h
new file mode 100644
index 0000000..0bb174c
--- /dev/null
+++ b/lib/zstd/common/fse.h
@@ -0,0 +1,710 @@
+/* ******************************************************************
+ * FSE : Finite State Entropy codec
+ * Public Prototypes declaration
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ * You can contact the author at :
+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+
+#ifndef FSE_H
+#define FSE_H
+
+
+/*-*****************************************
+*  Dependencies
+******************************************/
+#include "zstd_deps.h"    /* size_t, ptrdiff_t */
+
+
+/*-*****************************************
+*  FSE_PUBLIC_API : control library symbols visibility
+******************************************/
+#if defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) && defined(__GNUC__) && (__GNUC__ >= 4)
+#  define FSE_PUBLIC_API __attribute__ ((visibility ("default")))
+#elif defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1)   /* Visual expected */
+#  define FSE_PUBLIC_API __declspec(dllexport)
+#elif defined(FSE_DLL_IMPORT) && (FSE_DLL_IMPORT==1)
+#  define FSE_PUBLIC_API __declspec(dllimport) /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/
+#else
+#  define FSE_PUBLIC_API
+#endif
+
+/*------   Version   ------*/
+#define FSE_VERSION_MAJOR    0
+#define FSE_VERSION_MINOR    9
+#define FSE_VERSION_RELEASE  0
+
+#define FSE_LIB_VERSION FSE_VERSION_MAJOR.FSE_VERSION_MINOR.FSE_VERSION_RELEASE
+#define FSE_QUOTE(str) #str
+#define FSE_EXPAND_AND_QUOTE(str) FSE_QUOTE(str)
+#define FSE_VERSION_STRING FSE_EXPAND_AND_QUOTE(FSE_LIB_VERSION)
+
+#define FSE_VERSION_NUMBER  (FSE_VERSION_MAJOR *100*100 + FSE_VERSION_MINOR *100 + FSE_VERSION_RELEASE)
+FSE_PUBLIC_API unsigned FSE_versionNumber(void);   /*< library version number; to be used when checking dll version */
+
+
+/*-****************************************
+*  FSE simple functions
+******************************************/
+/*! FSE_compress() :
+    Compress content of buffer 'src', of size 'srcSize', into destination buffer 'dst'.
+    'dst' buffer must be already allocated. Compression runs faster is dstCapacity >= FSE_compressBound(srcSize).
+    @return : size of compressed data (<= dstCapacity).
+    Special values : if return == 0, srcData is not compressible => Nothing is stored within dst !!!
+                     if return == 1, srcData is a single byte symbol * srcSize times. Use RLE compression instead.
+                     if FSE_isError(return), compression failed (more details using FSE_getErrorName())
+*/
+FSE_PUBLIC_API size_t FSE_compress(void* dst, size_t dstCapacity,
+                             const void* src, size_t srcSize);
+
+/*! FSE_decompress():
+    Decompress FSE data from buffer 'cSrc', of size 'cSrcSize',
+    into already allocated destination buffer 'dst', of size 'dstCapacity'.
+    @return : size of regenerated data (<= maxDstSize),
+              or an error code, which can be tested using FSE_isError() .
+
+    ** Important ** : FSE_decompress() does not decompress non-compressible nor RLE data !!!
+    Why ? : making this distinction requires a header.
+    Header management is intentionally delegated to the user layer, which can better manage special cases.
+*/
+FSE_PUBLIC_API size_t FSE_decompress(void* dst,  size_t dstCapacity,
+                               const void* cSrc, size_t cSrcSize);
+
+
+/*-*****************************************
+*  Tool functions
+******************************************/
+FSE_PUBLIC_API size_t FSE_compressBound(size_t size);       /* maximum compressed size */
+
+/* Error Management */
+FSE_PUBLIC_API unsigned    FSE_isError(size_t code);        /* tells if a return value is an error code */
+FSE_PUBLIC_API const char* FSE_getErrorName(size_t code);   /* provides error code string (useful for debugging) */
+
+
+/*-*****************************************
+*  FSE advanced functions
+******************************************/
+/*! FSE_compress2() :
+    Same as FSE_compress(), but allows the selection of 'maxSymbolValue' and 'tableLog'
+    Both parameters can be defined as '0' to mean : use default value
+    @return : size of compressed data
+    Special values : if return == 0, srcData is not compressible => Nothing is stored within cSrc !!!
+                     if return == 1, srcData is a single byte symbol * srcSize times. Use RLE compression.
+                     if FSE_isError(return), it's an error code.
+*/
+FSE_PUBLIC_API size_t FSE_compress2 (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog);
+
+
+/*-*****************************************
+*  FSE detailed API
+******************************************/
+/*!
+FSE_compress() does the following:
+1. count symbol occurrence from source[] into table count[] (see hist.h)
+2. normalize counters so that sum(count[]) == Power_of_2 (2^tableLog)
+3. save normalized counters to memory buffer using writeNCount()
+4. build encoding table 'CTable' from normalized counters
+5. encode the data stream using encoding table 'CTable'
+
+FSE_decompress() does the following:
+1. read normalized counters with readNCount()
+2. build decoding table 'DTable' from normalized counters
+3. decode the data stream using decoding table 'DTable'
+
+The following API allows targeting specific sub-functions for advanced tasks.
+For example, it's possible to compress several blocks using the same 'CTable',
+or to save and provide normalized distribution using external method.
+*/
+
+/* *** COMPRESSION *** */
+
+/*! FSE_optimalTableLog():
+    dynamically downsize 'tableLog' when conditions are met.
+    It saves CPU time, by using smaller tables, while preserving or even improving compression ratio.
+    @return : recommended tableLog (necessarily <= 'maxTableLog') */
+FSE_PUBLIC_API unsigned FSE_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue);
+
+/*! FSE_normalizeCount():
+    normalize counts so that sum(count[]) == Power_of_2 (2^tableLog)
+    'normalizedCounter' is a table of short, of minimum size (maxSymbolValue+1).
+    useLowProbCount is a boolean parameter which trades off compressed size for
+    faster header decoding. When it is set to 1, the compressed data will be slightly
+    smaller. And when it is set to 0, FSE_readNCount() and FSE_buildDTable() will be
+    faster. If you are compressing a small amount of data (< 2 KB) then useLowProbCount=0
+    is a good default, since header deserialization makes a big speed difference.
+    Otherwise, useLowProbCount=1 is a good default, since the speed difference is small.
+    @return : tableLog,
+              or an errorCode, which can be tested using FSE_isError() */
+FSE_PUBLIC_API size_t FSE_normalizeCount(short* normalizedCounter, unsigned tableLog,
+                    const unsigned* count, size_t srcSize, unsigned maxSymbolValue, unsigned useLowProbCount);
+
+/*! FSE_NCountWriteBound():
+    Provides the maximum possible size of an FSE normalized table, given 'maxSymbolValue' and 'tableLog'.
+    Typically useful for allocation purpose. */
+FSE_PUBLIC_API size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog);
+
+/*! FSE_writeNCount():
+    Compactly save 'normalizedCounter' into 'buffer'.
+    @return : size of the compressed table,
+              or an errorCode, which can be tested using FSE_isError(). */
+FSE_PUBLIC_API size_t FSE_writeNCount (void* buffer, size_t bufferSize,
+                                 const short* normalizedCounter,
+                                 unsigned maxSymbolValue, unsigned tableLog);
+
+/*! Constructor and Destructor of FSE_CTable.
+    Note that FSE_CTable size depends on 'tableLog' and 'maxSymbolValue' */
+typedef unsigned FSE_CTable;   /* don't allocate that. It's only meant to be more restrictive than void* */
+FSE_PUBLIC_API FSE_CTable* FSE_createCTable (unsigned maxSymbolValue, unsigned tableLog);
+FSE_PUBLIC_API void        FSE_freeCTable (FSE_CTable* ct);
+
+/*! FSE_buildCTable():
+    Builds `ct`, which must be already allocated, using FSE_createCTable().
+    @return : 0, or an errorCode, which can be tested using FSE_isError() */
+FSE_PUBLIC_API size_t FSE_buildCTable(FSE_CTable* ct, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog);
+
+/*! FSE_compress_usingCTable():
+    Compress `src` using `ct` into `dst` which must be already allocated.
+    @return : size of compressed data (<= `dstCapacity`),
+              or 0 if compressed data could not fit into `dst`,
+              or an errorCode, which can be tested using FSE_isError() */
+FSE_PUBLIC_API size_t FSE_compress_usingCTable (void* dst, size_t dstCapacity, const void* src, size_t srcSize, const FSE_CTable* ct);
+
+/*!
+Tutorial :
+----------
+The first step is to count all symbols. FSE_count() does this job very fast.
+Result will be saved into 'count', a table of unsigned int, which must be already allocated, and have 'maxSymbolValuePtr[0]+1' cells.
+'src' is a table of bytes of size 'srcSize'. All values within 'src' MUST be <= maxSymbolValuePtr[0]
+maxSymbolValuePtr[0] will be updated, with its real value (necessarily <= original value)
+FSE_count() will return the number of occurrence of the most frequent symbol.
+This can be used to know if there is a single symbol within 'src', and to quickly evaluate its compressibility.
+If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError()).
+
+The next step is to normalize the frequencies.
+FSE_normalizeCount() will ensure that sum of frequencies is == 2 ^'tableLog'.
+It also guarantees a minimum of 1 to any Symbol with frequency >= 1.
+You can use 'tableLog'==0 to mean "use default tableLog value".
+If you are unsure of which tableLog value to use, you can ask FSE_optimalTableLog(),
+which will provide the optimal valid tableLog given sourceSize, maxSymbolValue, and a user-defined maximum (0 means "default").
+
+The result of FSE_normalizeCount() will be saved into a table,
+called 'normalizedCounter', which is a table of signed short.
+'normalizedCounter' must be already allocated, and have at least 'maxSymbolValue+1' cells.
+The return value is tableLog if everything proceeded as expected.
+It is 0 if there is a single symbol within distribution.
+If there is an error (ex: invalid tableLog value), the function will return an ErrorCode (which can be tested using FSE_isError()).
+
+'normalizedCounter' can be saved in a compact manner to a memory area using FSE_writeNCount().
+'buffer' must be already allocated.
+For guaranteed success, buffer size must be at least FSE_headerBound().
+The result of the function is the number of bytes written into 'buffer'.
+If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError(); ex : buffer size too small).
+
+'normalizedCounter' can then be used to create the compression table 'CTable'.
+The space required by 'CTable' must be already allocated, using FSE_createCTable().
+You can then use FSE_buildCTable() to fill 'CTable'.
+If there is an error, both functions will return an ErrorCode (which can be tested using FSE_isError()).
+
+'CTable' can then be used to compress 'src', with FSE_compress_usingCTable().
+Similar to FSE_count(), the convention is that 'src' is assumed to be a table of char of size 'srcSize'
+The function returns the size of compressed data (without header), necessarily <= `dstCapacity`.
+If it returns '0', compressed data could not fit into 'dst'.
+If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError()).
+*/
+
+
+/* *** DECOMPRESSION *** */
+
+/*! FSE_readNCount():
+    Read compactly saved 'normalizedCounter' from 'rBuffer'.
+    @return : size read from 'rBuffer',
+              or an errorCode, which can be tested using FSE_isError().
+              maxSymbolValuePtr[0] and tableLogPtr[0] will also be updated with their respective values */
+FSE_PUBLIC_API size_t FSE_readNCount (short* normalizedCounter,
+                           unsigned* maxSymbolValuePtr, unsigned* tableLogPtr,
+                           const void* rBuffer, size_t rBuffSize);
+
+/*! FSE_readNCount_bmi2():
+ * Same as FSE_readNCount() but pass bmi2=1 when your CPU supports BMI2 and 0 otherwise.
+ */
+FSE_PUBLIC_API size_t FSE_readNCount_bmi2(short* normalizedCounter,
+                           unsigned* maxSymbolValuePtr, unsigned* tableLogPtr,
+                           const void* rBuffer, size_t rBuffSize, int bmi2);
+
+/*! Constructor and Destructor of FSE_DTable.
+    Note that its size depends on 'tableLog' */
+typedef unsigned FSE_DTable;   /* don't allocate that. It's just a way to be more restrictive than void* */
+FSE_PUBLIC_API FSE_DTable* FSE_createDTable(unsigned tableLog);
+FSE_PUBLIC_API void        FSE_freeDTable(FSE_DTable* dt);
+
+/*! FSE_buildDTable():
+    Builds 'dt', which must be already allocated, using FSE_createDTable().
+    return : 0, or an errorCode, which can be tested using FSE_isError() */
+FSE_PUBLIC_API size_t FSE_buildDTable (FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog);
+
+/*! FSE_decompress_usingDTable():
+    Decompress compressed source `cSrc` of size `cSrcSize` using `dt`
+    into `dst` which must be already allocated.
+    @return : size of regenerated data (necessarily <= `dstCapacity`),
+              or an errorCode, which can be tested using FSE_isError() */
+FSE_PUBLIC_API size_t FSE_decompress_usingDTable(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, const FSE_DTable* dt);
+
+/*!
+Tutorial :
+----------
+(Note : these functions only decompress FSE-compressed blocks.
+ If block is uncompressed, use memcpy() instead
+ If block is a single repeated byte, use memset() instead )
+
+The first step is to obtain the normalized frequencies of symbols.
+This can be performed by FSE_readNCount() if it was saved using FSE_writeNCount().
+'normalizedCounter' must be already allocated, and have at least 'maxSymbolValuePtr[0]+1' cells of signed short.
+In practice, that means it's necessary to know 'maxSymbolValue' beforehand,
+or size the table to handle worst case situations (typically 256).
+FSE_readNCount() will provide 'tableLog' and 'maxSymbolValue'.
+The result of FSE_readNCount() is the number of bytes read from 'rBuffer'.
+Note that 'rBufferSize' must be at least 4 bytes, even if useful information is less than that.
+If there is an error, the function will return an error code, which can be tested using FSE_isError().
+
+The next step is to build the decompression tables 'FSE_DTable' from 'normalizedCounter'.
+This is performed by the function FSE_buildDTable().
+The space required by 'FSE_DTable' must be already allocated using FSE_createDTable().
+If there is an error, the function will return an error code, which can be tested using FSE_isError().
+
+`FSE_DTable` can then be used to decompress `cSrc`, with FSE_decompress_usingDTable().
+`cSrcSize` must be strictly correct, otherwise decompression will fail.
+FSE_decompress_usingDTable() result will tell how many bytes were regenerated (<=`dstCapacity`).
+If there is an error, the function will return an error code, which can be tested using FSE_isError(). (ex: dst buffer too small)
+*/
+
+#endif  /* FSE_H */
+
+#if !defined(FSE_H_FSE_STATIC_LINKING_ONLY)
+#define FSE_H_FSE_STATIC_LINKING_ONLY
+
+/* *** Dependency *** */
+#include "bitstream.h"
+
+
+/* *****************************************
+*  Static allocation
+*******************************************/
+/* FSE buffer bounds */
+#define FSE_NCOUNTBOUND 512
+#define FSE_BLOCKBOUND(size) ((size) + ((size)>>7) + 4 /* fse states */ + sizeof(size_t) /* bitContainer */)
+#define FSE_COMPRESSBOUND(size) (FSE_NCOUNTBOUND + FSE_BLOCKBOUND(size))   /* Macro version, useful for static allocation */
+
+/* It is possible to statically allocate FSE CTable/DTable as a table of FSE_CTable/FSE_DTable using below macros */
+#define FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue)   (1 + (1<<((maxTableLog)-1)) + (((maxSymbolValue)+1)*2))
+#define FSE_DTABLE_SIZE_U32(maxTableLog)                   (1 + (1<<(maxTableLog)))
+
+/* or use the size to malloc() space directly. Pay attention to alignment restrictions though */
+#define FSE_CTABLE_SIZE(maxTableLog, maxSymbolValue)   (FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) * sizeof(FSE_CTable))
+#define FSE_DTABLE_SIZE(maxTableLog)                   (FSE_DTABLE_SIZE_U32(maxTableLog) * sizeof(FSE_DTable))
+
+
+/* *****************************************
+ *  FSE advanced API
+ ***************************************** */
+
+unsigned FSE_optimalTableLog_internal(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, unsigned minus);
+/*< same as FSE_optimalTableLog(), which used `minus==2` */
+
+/* FSE_compress_wksp() :
+ * Same as FSE_compress2(), but using an externally allocated scratch buffer (`workSpace`).
+ * FSE_COMPRESS_WKSP_SIZE_U32() provides the minimum size required for `workSpace` as a table of FSE_CTable.
+ */
+#define FSE_COMPRESS_WKSP_SIZE_U32(maxTableLog, maxSymbolValue)   ( FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) + ((maxTableLog > 12) ? (1 << (maxTableLog - 2)) : 1024) )
+size_t FSE_compress_wksp (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize);
+
+size_t FSE_buildCTable_raw (FSE_CTable* ct, unsigned nbBits);
+/*< build a fake FSE_CTable, designed for a flat distribution, where each symbol uses nbBits */
+
+size_t FSE_buildCTable_rle (FSE_CTable* ct, unsigned char symbolValue);
+/*< build a fake FSE_CTable, designed to compress always the same symbolValue */
+
+/* FSE_buildCTable_wksp() :
+ * Same as FSE_buildCTable(), but using an externally allocated scratch buffer (`workSpace`).
+ * `wkspSize` must be >= `FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(maxSymbolValue, tableLog)` of `unsigned`.
+ */
+#define FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(maxSymbolValue, tableLog) (maxSymbolValue + 2 + (1ull << (tableLog - 2)))
+#define FSE_BUILD_CTABLE_WORKSPACE_SIZE(maxSymbolValue, tableLog) (sizeof(unsigned) * FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(maxSymbolValue, tableLog))
+size_t FSE_buildCTable_wksp(FSE_CTable* ct, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize);
+
+#define FSE_BUILD_DTABLE_WKSP_SIZE(maxTableLog, maxSymbolValue) (sizeof(short) * (maxSymbolValue + 1) + (1ULL << maxTableLog) + 8)
+#define FSE_BUILD_DTABLE_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) ((FSE_BUILD_DTABLE_WKSP_SIZE(maxTableLog, maxSymbolValue) + sizeof(unsigned) - 1) / sizeof(unsigned))
+FSE_PUBLIC_API size_t FSE_buildDTable_wksp(FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize);
+/*< Same as FSE_buildDTable(), using an externally allocated `workspace` produced with `FSE_BUILD_DTABLE_WKSP_SIZE_U32(maxSymbolValue)` */
+
+size_t FSE_buildDTable_raw (FSE_DTable* dt, unsigned nbBits);
+/*< build a fake FSE_DTable, designed to read a flat distribution where each symbol uses nbBits */
+
+size_t FSE_buildDTable_rle (FSE_DTable* dt, unsigned char symbolValue);
+/*< build a fake FSE_DTable, designed to always generate the same symbolValue */
+
+#define FSE_DECOMPRESS_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) (FSE_DTABLE_SIZE_U32(maxTableLog) + FSE_BUILD_DTABLE_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) + (FSE_MAX_SYMBOL_VALUE + 1) / 2 + 1)
+#define FSE_DECOMPRESS_WKSP_SIZE(maxTableLog, maxSymbolValue) (FSE_DECOMPRESS_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) * sizeof(unsigned))
+size_t FSE_decompress_wksp(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize);
+/*< same as FSE_decompress(), using an externally allocated `workSpace` produced with `FSE_DECOMPRESS_WKSP_SIZE_U32(maxLog, maxSymbolValue)` */
+
+size_t FSE_decompress_wksp_bmi2(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize, int bmi2);
+/*< Same as FSE_decompress_wksp() but with dynamic BMI2 support. Pass 1 if your CPU supports BMI2 or 0 if it doesn't. */
+
+typedef enum {
+   FSE_repeat_none,  /*< Cannot use the previous table */
+   FSE_repeat_check, /*< Can use the previous table but it must be checked */
+   FSE_repeat_valid  /*< Can use the previous table and it is assumed to be valid */
+ } FSE_repeat;
+
+/* *****************************************
+*  FSE symbol compression API
+*******************************************/
+/*!
+   This API consists of small unitary functions, which highly benefit from being inlined.
+   Hence their body are included in next section.
+*/
+typedef struct {
+    ptrdiff_t   value;
+    const void* stateTable;
+    const void* symbolTT;
+    unsigned    stateLog;
+} FSE_CState_t;
+
+static void FSE_initCState(FSE_CState_t* CStatePtr, const FSE_CTable* ct);
+
+static void FSE_encodeSymbol(BIT_CStream_t* bitC, FSE_CState_t* CStatePtr, unsigned symbol);
+
+static void FSE_flushCState(BIT_CStream_t* bitC, const FSE_CState_t* CStatePtr);
+
+/*<
+These functions are inner components of FSE_compress_usingCTable().
+They allow the creation of custom streams, mixing multiple tables and bit sources.
+
+A key property to keep in mind is that encoding and decoding are done **in reverse direction**.
+So the first symbol you will encode is the last you will decode, like a LIFO stack.
+
+You will need a few variables to track your CStream. They are :
+
+FSE_CTable    ct;         // Provided by FSE_buildCTable()
+BIT_CStream_t bitStream;  // bitStream tracking structure
+FSE_CState_t  state;      // State tracking structure (can have several)
+
+
+The first thing to do is to init bitStream and state.
+    size_t errorCode = BIT_initCStream(&bitStream, dstBuffer, maxDstSize);
+    FSE_initCState(&state, ct);
+
+Note that BIT_initCStream() can produce an error code, so its result should be tested, using FSE_isError();
+You can then encode your input data, byte after byte.
+FSE_encodeSymbol() outputs a maximum of 'tableLog' bits at a time.
+Remember decoding will be done in reverse direction.
+    FSE_encodeByte(&bitStream, &state, symbol);
+
+At any time, you can also add any bit sequence.
+Note : maximum allowed nbBits is 25, for compatibility with 32-bits decoders
+    BIT_addBits(&bitStream, bitField, nbBits);
+
+The above methods don't commit data to memory, they just store it into local register, for speed.
+Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (size_t).
+Writing data to memory is a manual operation, performed by the flushBits function.
+    BIT_flushBits(&bitStream);
+
+Your last FSE encoding operation shall be to flush your last state value(s).
+    FSE_flushState(&bitStream, &state);
+
+Finally, you must close the bitStream.
+The function returns the size of CStream in bytes.
+If data couldn't fit into dstBuffer, it will return a 0 ( == not compressible)
+If there is an error, it returns an errorCode (which can be tested using FSE_isError()).
+    size_t size = BIT_closeCStream(&bitStream);
+*/
+
+
+/* *****************************************
+*  FSE symbol decompression API
+*******************************************/
+typedef struct {
+    size_t      state;
+    const void* table;   /* precise table may vary, depending on U16 */
+} FSE_DState_t;
+
+
+static void     FSE_initDState(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD, const FSE_DTable* dt);
+
+static unsigned char FSE_decodeSymbol(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD);
+
+static unsigned FSE_endOfDState(const FSE_DState_t* DStatePtr);
+
+/*<
+Let's now decompose FSE_decompress_usingDTable() into its unitary components.
+You will decode FSE-encoded symbols from the bitStream,
+and also any other bitFields you put in, **in reverse order**.
+
+You will need a few variables to track your bitStream. They are :
+
+BIT_DStream_t DStream;    // Stream context
+FSE_DState_t  DState;     // State context. Multiple ones are possible
+FSE_DTable*   DTablePtr;  // Decoding table, provided by FSE_buildDTable()
+
+The first thing to do is to init the bitStream.
+    errorCode = BIT_initDStream(&DStream, srcBuffer, srcSize);
+
+You should then retrieve your initial state(s)
+(in reverse flushing order if you have several ones) :
+    errorCode = FSE_initDState(&DState, &DStream, DTablePtr);
+
+You can then decode your data, symbol after symbol.
+For information the maximum number of bits read by FSE_decodeSymbol() is 'tableLog'.
+Keep in mind that symbols are decoded in reverse order, like a LIFO stack (last in, first out).
+    unsigned char symbol = FSE_decodeSymbol(&DState, &DStream);
+
+You can retrieve any bitfield you eventually stored into the bitStream (in reverse order)
+Note : maximum allowed nbBits is 25, for 32-bits compatibility
+    size_t bitField = BIT_readBits(&DStream, nbBits);
+
+All above operations only read from local register (which size depends on size_t).
+Refueling the register from memory is manually performed by the reload method.
+    endSignal = FSE_reloadDStream(&DStream);
+
+BIT_reloadDStream() result tells if there is still some more data to read from DStream.
+BIT_DStream_unfinished : there is still some data left into the DStream.
+BIT_DStream_endOfBuffer : Dstream reached end of buffer. Its container may no longer be completely filled.
+BIT_DStream_completed : Dstream reached its exact end, corresponding in general to decompression completed.
+BIT_DStream_tooFar : Dstream went too far. Decompression result is corrupted.
+
+When reaching end of buffer (BIT_DStream_endOfBuffer), progress slowly, notably if you decode multiple symbols per loop,
+to properly detect the exact end of stream.
+After each decoded symbol, check if DStream is fully consumed using this simple test :
+    BIT_reloadDStream(&DStream) >= BIT_DStream_completed
+
+When it's done, verify decompression is fully completed, by checking both DStream and the relevant states.
+Checking if DStream has reached its end is performed by :
+    BIT_endOfDStream(&DStream);
+Check also the states. There might be some symbols left there, if some high probability ones (>50%) are possible.
+    FSE_endOfDState(&DState);
+*/
+
+
+/* *****************************************
+*  FSE unsafe API
+*******************************************/
+static unsigned char FSE_decodeSymbolFast(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD);
+/* faster, but works only if nbBits is always >= 1 (otherwise, result will be corrupted) */
+
+
+/* *****************************************
+*  Implementation of inlined functions
+*******************************************/
+typedef struct {
+    int deltaFindState;
+    U32 deltaNbBits;
+} FSE_symbolCompressionTransform; /* total 8 bytes */
+
+MEM_STATIC void FSE_initCState(FSE_CState_t* statePtr, const FSE_CTable* ct)
+{
+    const void* ptr = ct;
+    const U16* u16ptr = (const U16*) ptr;
+    const U32 tableLog = MEM_read16(ptr);
+    statePtr->value = (ptrdiff_t)1<<tableLog;
+    statePtr->stateTable = u16ptr+2;
+    statePtr->symbolTT = ct + 1 + (tableLog ? (1<<(tableLog-1)) : 1);
+    statePtr->stateLog = tableLog;
+}
+
+
+/*! FSE_initCState2() :
+*   Same as FSE_initCState(), but the first symbol to include (which will be the last to be read)
+*   uses the smallest state value possible, saving the cost of this symbol */
+MEM_STATIC void FSE_initCState2(FSE_CState_t* statePtr, const FSE_CTable* ct, U32 symbol)
+{
+    FSE_initCState(statePtr, ct);
+    {   const FSE_symbolCompressionTransform symbolTT = ((const FSE_symbolCompressionTransform*)(statePtr->symbolTT))[symbol];
+        const U16* stateTable = (const U16*)(statePtr->stateTable);
+        U32 nbBitsOut  = (U32)((symbolTT.deltaNbBits + (1<<15)) >> 16);
+        statePtr->value = (nbBitsOut << 16) - symbolTT.deltaNbBits;
+        statePtr->value = stateTable[(statePtr->value >> nbBitsOut) + symbolTT.deltaFindState];
+    }
+}
+
+MEM_STATIC void FSE_encodeSymbol(BIT_CStream_t* bitC, FSE_CState_t* statePtr, unsigned symbol)
+{
+    FSE_symbolCompressionTransform const symbolTT = ((const FSE_symbolCompressionTransform*)(statePtr->symbolTT))[symbol];
+    const U16* const stateTable = (const U16*)(statePtr->stateTable);
+    U32 const nbBitsOut  = (U32)((statePtr->value + symbolTT.deltaNbBits) >> 16);
+    BIT_addBits(bitC, statePtr->value, nbBitsOut);
+    statePtr->value = stateTable[ (statePtr->value >> nbBitsOut) + symbolTT.deltaFindState];
+}
+
+MEM_STATIC void FSE_flushCState(BIT_CStream_t* bitC, const FSE_CState_t* statePtr)
+{
+    BIT_addBits(bitC, statePtr->value, statePtr->stateLog);
+    BIT_flushBits(bitC);
+}
+
+
+/* FSE_getMaxNbBits() :
+ * Approximate maximum cost of a symbol, in bits.
+ * Fractional get rounded up (i.e : a symbol with a normalized frequency of 3 gives the same result as a frequency of 2)
+ * note 1 : assume symbolValue is valid (<= maxSymbolValue)
+ * note 2 : if freq[symbolValue]==0, @return a fake cost of tableLog+1 bits */
+MEM_STATIC U32 FSE_getMaxNbBits(const void* symbolTTPtr, U32 symbolValue)
+{
+    const FSE_symbolCompressionTransform* symbolTT = (const FSE_symbolCompressionTransform*) symbolTTPtr;
+    return (symbolTT[symbolValue].deltaNbBits + ((1<<16)-1)) >> 16;
+}
+
+/* FSE_bitCost() :
+ * Approximate symbol cost, as fractional value, using fixed-point format (accuracyLog fractional bits)
+ * note 1 : assume symbolValue is valid (<= maxSymbolValue)
+ * note 2 : if freq[symbolValue]==0, @return a fake cost of tableLog+1 bits */
+MEM_STATIC U32 FSE_bitCost(const void* symbolTTPtr, U32 tableLog, U32 symbolValue, U32 accuracyLog)
+{
+    const FSE_symbolCompressionTransform* symbolTT = (const FSE_symbolCompressionTransform*) symbolTTPtr;
+    U32 const minNbBits = symbolTT[symbolValue].deltaNbBits >> 16;
+    U32 const threshold = (minNbBits+1) << 16;
+    assert(tableLog < 16);
+    assert(accuracyLog < 31-tableLog);  /* ensure enough room for renormalization double shift */
+    {   U32 const tableSize = 1 << tableLog;
+        U32 const deltaFromThreshold = threshold - (symbolTT[symbolValue].deltaNbBits + tableSize);
+        U32 const normalizedDeltaFromThreshold = (deltaFromThreshold << accuracyLog) >> tableLog;   /* linear interpolation (very approximate) */
+        U32 const bitMultiplier = 1 << accuracyLog;
+        assert(symbolTT[symbolValue].deltaNbBits + tableSize <= threshold);
+        assert(normalizedDeltaFromThreshold <= bitMultiplier);
+        return (minNbBits+1)*bitMultiplier - normalizedDeltaFromThreshold;
+    }
+}
+
+
+/* ======    Decompression    ====== */
+
+typedef struct {
+    U16 tableLog;
+    U16 fastMode;
+} FSE_DTableHeader;   /* sizeof U32 */
+
+typedef struct
+{
+    unsigned short newState;
+    unsigned char  symbol;
+    unsigned char  nbBits;
+} FSE_decode_t;   /* size == U32 */
+
+MEM_STATIC void FSE_initDState(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD, const FSE_DTable* dt)
+{
+    const void* ptr = dt;
+    const FSE_DTableHeader* const DTableH = (const FSE_DTableHeader*)ptr;
+    DStatePtr->state = BIT_readBits(bitD, DTableH->tableLog);
+    BIT_reloadDStream(bitD);
+    DStatePtr->table = dt + 1;
+}
+
+MEM_STATIC BYTE FSE_peekSymbol(const FSE_DState_t* DStatePtr)
+{
+    FSE_decode_t const DInfo = ((const FSE_decode_t*)(DStatePtr->table))[DStatePtr->state];
+    return DInfo.symbol;
+}
+
+MEM_STATIC void FSE_updateState(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD)
+{
+    FSE_decode_t const DInfo = ((const FSE_decode_t*)(DStatePtr->table))[DStatePtr->state];
+    U32 const nbBits = DInfo.nbBits;
+    size_t const lowBits = BIT_readBits(bitD, nbBits);
+    DStatePtr->state = DInfo.newState + lowBits;
+}
+
+MEM_STATIC BYTE FSE_decodeSymbol(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD)
+{
+    FSE_decode_t const DInfo = ((const FSE_decode_t*)(DStatePtr->table))[DStatePtr->state];
+    U32 const nbBits = DInfo.nbBits;
+    BYTE const symbol = DInfo.symbol;
+    size_t const lowBits = BIT_readBits(bitD, nbBits);
+
+    DStatePtr->state = DInfo.newState + lowBits;
+    return symbol;
+}
+
+/*! FSE_decodeSymbolFast() :
+    unsafe, only works if no symbol has a probability > 50% */
+MEM_STATIC BYTE FSE_decodeSymbolFast(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD)
+{
+    FSE_decode_t const DInfo = ((const FSE_decode_t*)(DStatePtr->table))[DStatePtr->state];
+    U32 const nbBits = DInfo.nbBits;
+    BYTE const symbol = DInfo.symbol;
+    size_t const lowBits = BIT_readBitsFast(bitD, nbBits);
+
+    DStatePtr->state = DInfo.newState + lowBits;
+    return symbol;
+}
+
+MEM_STATIC unsigned FSE_endOfDState(const FSE_DState_t* DStatePtr)
+{
+    return DStatePtr->state == 0;
+}
+
+
+
+#ifndef FSE_COMMONDEFS_ONLY
+
+/* **************************************************************
+*  Tuning parameters
+****************************************************************/
+/*!MEMORY_USAGE :
+*  Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.)
+*  Increasing memory usage improves compression ratio
+*  Reduced memory usage can improve speed, due to cache effect
+*  Recommended max value is 14, for 16KB, which nicely fits into Intel x86 L1 cache */
+#ifndef FSE_MAX_MEMORY_USAGE
+#  define FSE_MAX_MEMORY_USAGE 14
+#endif
+#ifndef FSE_DEFAULT_MEMORY_USAGE
+#  define FSE_DEFAULT_MEMORY_USAGE 13
+#endif
+#if (FSE_DEFAULT_MEMORY_USAGE > FSE_MAX_MEMORY_USAGE)
+#  error "FSE_DEFAULT_MEMORY_USAGE must be <= FSE_MAX_MEMORY_USAGE"
+#endif
+
+/*!FSE_MAX_SYMBOL_VALUE :
+*  Maximum symbol value authorized.
+*  Required for proper stack allocation */
+#ifndef FSE_MAX_SYMBOL_VALUE
+#  define FSE_MAX_SYMBOL_VALUE 255
+#endif
+
+/* **************************************************************
+*  template functions type & suffix
+****************************************************************/
+#define FSE_FUNCTION_TYPE BYTE
+#define FSE_FUNCTION_EXTENSION
+#define FSE_DECODE_TYPE FSE_decode_t
+
+
+#endif   /* !FSE_COMMONDEFS_ONLY */
+
+
+/* ***************************************************************
+*  Constants
+*****************************************************************/
+#define FSE_MAX_TABLELOG  (FSE_MAX_MEMORY_USAGE-2)
+#define FSE_MAX_TABLESIZE (1U<<FSE_MAX_TABLELOG)
+#define FSE_MAXTABLESIZE_MASK (FSE_MAX_TABLESIZE-1)
+#define FSE_DEFAULT_TABLELOG (FSE_DEFAULT_MEMORY_USAGE-2)
+#define FSE_MIN_TABLELOG 5
+
+#define FSE_TABLELOG_ABSOLUTE_MAX 15
+#if FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX
+#  error "FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX is not supported"
+#endif
+
+#define FSE_TABLESTEP(tableSize) (((tableSize)>>1) + ((tableSize)>>3) + 3)
+
+
+#endif /* FSE_STATIC_LINKING_ONLY */
+
+
diff --git a/lib/zstd/common/fse_decompress.c b/lib/zstd/common/fse_decompress.c
new file mode 100644
index 0000000..2c8bbe3
--- /dev/null
+++ b/lib/zstd/common/fse_decompress.c
@@ -0,0 +1,390 @@
+/* ******************************************************************
+ * FSE : Finite State Entropy decoder
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ *  You can contact the author at :
+ *  - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *  - Public forum : https://groups.google.com/forum/#!forum/lz4c
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+
+/* **************************************************************
+*  Includes
+****************************************************************/
+#include "debug.h"      /* assert */
+#include "bitstream.h"
+#include "compiler.h"
+#define FSE_STATIC_LINKING_ONLY
+#include "fse.h"
+#include "error_private.h"
+#define ZSTD_DEPS_NEED_MALLOC
+#include "zstd_deps.h"
+
+
+/* **************************************************************
+*  Error Management
+****************************************************************/
+#define FSE_isError ERR_isError
+#define FSE_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c)   /* use only *after* variable declarations */
+
+
+/* **************************************************************
+*  Templates
+****************************************************************/
+/*
+  designed to be included
+  for type-specific functions (template emulation in C)
+  Objective is to write these functions only once, for improved maintenance
+*/
+
+/* safety checks */
+#ifndef FSE_FUNCTION_EXTENSION
+#  error "FSE_FUNCTION_EXTENSION must be defined"
+#endif
+#ifndef FSE_FUNCTION_TYPE
+#  error "FSE_FUNCTION_TYPE must be defined"
+#endif
+
+/* Function names */
+#define FSE_CAT(X,Y) X##Y
+#define FSE_FUNCTION_NAME(X,Y) FSE_CAT(X,Y)
+#define FSE_TYPE_NAME(X,Y) FSE_CAT(X,Y)
+
+
+/* Function templates */
+FSE_DTable* FSE_createDTable (unsigned tableLog)
+{
+    if (tableLog > FSE_TABLELOG_ABSOLUTE_MAX) tableLog = FSE_TABLELOG_ABSOLUTE_MAX;
+    return (FSE_DTable*)ZSTD_malloc( FSE_DTABLE_SIZE_U32(tableLog) * sizeof (U32) );
+}
+
+void FSE_freeDTable (FSE_DTable* dt)
+{
+    ZSTD_free(dt);
+}
+
+static size_t FSE_buildDTable_internal(FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize)
+{
+    void* const tdPtr = dt+1;   /* because *dt is unsigned, 32-bits aligned on 32-bits */
+    FSE_DECODE_TYPE* const tableDecode = (FSE_DECODE_TYPE*) (tdPtr);
+    U16* symbolNext = (U16*)workSpace;
+    BYTE* spread = (BYTE*)(symbolNext + maxSymbolValue + 1);
+
+    U32 const maxSV1 = maxSymbolValue + 1;
+    U32 const tableSize = 1 << tableLog;
+    U32 highThreshold = tableSize-1;
+
+    /* Sanity Checks */
+    if (FSE_BUILD_DTABLE_WKSP_SIZE(tableLog, maxSymbolValue) > wkspSize) return ERROR(maxSymbolValue_tooLarge);
+    if (maxSymbolValue > FSE_MAX_SYMBOL_VALUE) return ERROR(maxSymbolValue_tooLarge);
+    if (tableLog > FSE_MAX_TABLELOG) return ERROR(tableLog_tooLarge);
+
+    /* Init, lay down lowprob symbols */
+    {   FSE_DTableHeader DTableH;
+        DTableH.tableLog = (U16)tableLog;
+        DTableH.fastMode = 1;
+        {   S16 const largeLimit= (S16)(1 << (tableLog-1));
+            U32 s;
+            for (s=0; s<maxSV1; s++) {
+                if (normalizedCounter[s]==-1) {
+                    tableDecode[highThreshold--].symbol = (FSE_FUNCTION_TYPE)s;
+                    symbolNext[s] = 1;
+                } else {
+                    if (normalizedCounter[s] >= largeLimit) DTableH.fastMode=0;
+                    symbolNext[s] = normalizedCounter[s];
+        }   }   }
+        ZSTD_memcpy(dt, &DTableH, sizeof(DTableH));
+    }
+
+    /* Spread symbols */
+    if (highThreshold == tableSize - 1) {
+        size_t const tableMask = tableSize-1;
+        size_t const step = FSE_TABLESTEP(tableSize);
+        /* First lay down the symbols in order.
+         * We use a uint64_t to lay down 8 bytes at a time. This reduces branch
+         * misses since small blocks generally have small table logs, so nearly
+         * all symbols have counts <= 8. We ensure we have 8 bytes at the end of
+         * our buffer to handle the over-write.
+         */
+        {
+            U64 const add = 0x0101010101010101ull;
+            size_t pos = 0;
+            U64 sv = 0;
+            U32 s;
+            for (s=0; s<maxSV1; ++s, sv += add) {
+                int i;
+                int const n = normalizedCounter[s];
+                MEM_write64(spread + pos, sv);
+                for (i = 8; i < n; i += 8) {
+                    MEM_write64(spread + pos + i, sv);
+                }
+                pos += n;
+            }
+        }
+        /* Now we spread those positions across the table.
+         * The benefit of doing it in two stages is that we avoid the the
+         * variable size inner loop, which caused lots of branch misses.
+         * Now we can run through all the positions without any branch misses.
+         * We unroll the loop twice, since that is what emperically worked best.
+         */
+        {
+            size_t position = 0;
+            size_t s;
+            size_t const unroll = 2;
+            assert(tableSize % unroll == 0); /* FSE_MIN_TABLELOG is 5 */
+            for (s = 0; s < (size_t)tableSize; s += unroll) {
+                size_t u;
+                for (u = 0; u < unroll; ++u) {
+                    size_t const uPosition = (position + (u * step)) & tableMask;
+                    tableDecode[uPosition].symbol = spread[s + u];
+                }
+                position = (position + (unroll * step)) & tableMask;
+            }
+            assert(position == 0);
+        }
+    } else {
+        U32 const tableMask = tableSize-1;
+        U32 const step = FSE_TABLESTEP(tableSize);
+        U32 s, position = 0;
+        for (s=0; s<maxSV1; s++) {
+            int i;
+            for (i=0; i<normalizedCounter[s]; i++) {
+                tableDecode[position].symbol = (FSE_FUNCTION_TYPE)s;
+                position = (position + step) & tableMask;
+                while (position > highThreshold) position = (position + step) & tableMask;   /* lowprob area */
+        }   }
+        if (position!=0) return ERROR(GENERIC);   /* position must reach all cells once, otherwise normalizedCounter is incorrect */
+    }
+
+    /* Build Decoding table */
+    {   U32 u;
+        for (u=0; u<tableSize; u++) {
+            FSE_FUNCTION_TYPE const symbol = (FSE_FUNCTION_TYPE)(tableDecode[u].symbol);
+            U32 const nextState = symbolNext[symbol]++;
+            tableDecode[u].nbBits = (BYTE) (tableLog - BIT_highbit32(nextState) );
+            tableDecode[u].newState = (U16) ( (nextState << tableDecode[u].nbBits) - tableSize);
+    }   }
+
+    return 0;
+}
+
+size_t FSE_buildDTable_wksp(FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize)
+{
+    return FSE_buildDTable_internal(dt, normalizedCounter, maxSymbolValue, tableLog, workSpace, wkspSize);
+}
+
+
+#ifndef FSE_COMMONDEFS_ONLY
+
+/*-*******************************************************
+*  Decompression (Byte symbols)
+*********************************************************/
+size_t FSE_buildDTable_rle (FSE_DTable* dt, BYTE symbolValue)
+{
+    void* ptr = dt;
+    FSE_DTableHeader* const DTableH = (FSE_DTableHeader*)ptr;
+    void* dPtr = dt + 1;
+    FSE_decode_t* const cell = (FSE_decode_t*)dPtr;
+
+    DTableH->tableLog = 0;
+    DTableH->fastMode = 0;
+
+    cell->newState = 0;
+    cell->symbol = symbolValue;
+    cell->nbBits = 0;
+
+    return 0;
+}
+
+
+size_t FSE_buildDTable_raw (FSE_DTable* dt, unsigned nbBits)
+{
+    void* ptr = dt;
+    FSE_DTableHeader* const DTableH = (FSE_DTableHeader*)ptr;
+    void* dPtr = dt + 1;
+    FSE_decode_t* const dinfo = (FSE_decode_t*)dPtr;
+    const unsigned tableSize = 1 << nbBits;
+    const unsigned tableMask = tableSize - 1;
+    const unsigned maxSV1 = tableMask+1;
+    unsigned s;
+
+    /* Sanity checks */
+    if (nbBits < 1) return ERROR(GENERIC);         /* min size */
+
+    /* Build Decoding Table */
+    DTableH->tableLog = (U16)nbBits;
+    DTableH->fastMode = 1;
+    for (s=0; s<maxSV1; s++) {
+        dinfo[s].newState = 0;
+        dinfo[s].symbol = (BYTE)s;
+        dinfo[s].nbBits = (BYTE)nbBits;
+    }
+
+    return 0;
+}
+
+FORCE_INLINE_TEMPLATE size_t FSE_decompress_usingDTable_generic(
+          void* dst, size_t maxDstSize,
+    const void* cSrc, size_t cSrcSize,
+    const FSE_DTable* dt, const unsigned fast)
+{
+    BYTE* const ostart = (BYTE*) dst;
+    BYTE* op = ostart;
+    BYTE* const omax = op + maxDstSize;
+    BYTE* const olimit = omax-3;
+
+    BIT_DStream_t bitD;
+    FSE_DState_t state1;
+    FSE_DState_t state2;
+
+    /* Init */
+    CHECK_F(BIT_initDStream(&bitD, cSrc, cSrcSize));
+
+    FSE_initDState(&state1, &bitD, dt);
+    FSE_initDState(&state2, &bitD, dt);
+
+#define FSE_GETSYMBOL(statePtr) fast ? FSE_decodeSymbolFast(statePtr, &bitD) : FSE_decodeSymbol(statePtr, &bitD)
+
+    /* 4 symbols per loop */
+    for ( ; (BIT_reloadDStream(&bitD)==BIT_DStream_unfinished) & (op<olimit) ; op+=4) {
+        op[0] = FSE_GETSYMBOL(&state1);
+
+        if (FSE_MAX_TABLELOG*2+7 > sizeof(bitD.bitContainer)*8)    /* This test must be static */
+            BIT_reloadDStream(&bitD);
+
+        op[1] = FSE_GETSYMBOL(&state2);
+
+        if (FSE_MAX_TABLELOG*4+7 > sizeof(bitD.bitContainer)*8)    /* This test must be static */
+            { if (BIT_reloadDStream(&bitD) > BIT_DStream_unfinished) { op+=2; break; } }
+
+        op[2] = FSE_GETSYMBOL(&state1);
+
+        if (FSE_MAX_TABLELOG*2+7 > sizeof(bitD.bitContainer)*8)    /* This test must be static */
+            BIT_reloadDStream(&bitD);
+
+        op[3] = FSE_GETSYMBOL(&state2);
+    }
+
+    /* tail */
+    /* note : BIT_reloadDStream(&bitD) >= FSE_DStream_partiallyFilled; Ends at exactly BIT_DStream_completed */
+    while (1) {
+        if (op>(omax-2)) return ERROR(dstSize_tooSmall);
+        *op++ = FSE_GETSYMBOL(&state1);
+        if (BIT_reloadDStream(&bitD)==BIT_DStream_overflow) {
+            *op++ = FSE_GETSYMBOL(&state2);
+            break;
+        }
+
+        if (op>(omax-2)) return ERROR(dstSize_tooSmall);
+        *op++ = FSE_GETSYMBOL(&state2);
+        if (BIT_reloadDStream(&bitD)==BIT_DStream_overflow) {
+            *op++ = FSE_GETSYMBOL(&state1);
+            break;
+    }   }
+
+    return op-ostart;
+}
+
+
+size_t FSE_decompress_usingDTable(void* dst, size_t originalSize,
+                            const void* cSrc, size_t cSrcSize,
+                            const FSE_DTable* dt)
+{
+    const void* ptr = dt;
+    const FSE_DTableHeader* DTableH = (const FSE_DTableHeader*)ptr;
+    const U32 fastMode = DTableH->fastMode;
+
+    /* select fast mode (static) */
+    if (fastMode) return FSE_decompress_usingDTable_generic(dst, originalSize, cSrc, cSrcSize, dt, 1);
+    return FSE_decompress_usingDTable_generic(dst, originalSize, cSrc, cSrcSize, dt, 0);
+}
+
+
+size_t FSE_decompress_wksp(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize)
+{
+    return FSE_decompress_wksp_bmi2(dst, dstCapacity, cSrc, cSrcSize, maxLog, workSpace, wkspSize, /* bmi2 */ 0);
+}
+
+typedef struct {
+    short ncount[FSE_MAX_SYMBOL_VALUE + 1];
+    FSE_DTable dtable[1]; /* Dynamically sized */
+} FSE_DecompressWksp;
+
+
+FORCE_INLINE_TEMPLATE size_t FSE_decompress_wksp_body(
+        void* dst, size_t dstCapacity,
+        const void* cSrc, size_t cSrcSize,
+        unsigned maxLog, void* workSpace, size_t wkspSize,
+        int bmi2)
+{
+    const BYTE* const istart = (const BYTE*)cSrc;
+    const BYTE* ip = istart;
+    unsigned tableLog;
+    unsigned maxSymbolValue = FSE_MAX_SYMBOL_VALUE;
+    FSE_DecompressWksp* const wksp = (FSE_DecompressWksp*)workSpace;
+
+    DEBUG_STATIC_ASSERT((FSE_MAX_SYMBOL_VALUE + 1) % 2 == 0);
+    if (wkspSize < sizeof(*wksp)) return ERROR(GENERIC);
+
+    /* normal FSE decoding mode */
+    {
+        size_t const NCountLength = FSE_readNCount_bmi2(wksp->ncount, &maxSymbolValue, &tableLog, istart, cSrcSize, bmi2);
+        if (FSE_isError(NCountLength)) return NCountLength;
+        if (tableLog > maxLog) return ERROR(tableLog_tooLarge);
+        assert(NCountLength <= cSrcSize);
+        ip += NCountLength;
+        cSrcSize -= NCountLength;
+    }
+
+    if (FSE_DECOMPRESS_WKSP_SIZE(tableLog, maxSymbolValue) > wkspSize) return ERROR(tableLog_tooLarge);
+    workSpace = wksp->dtable + FSE_DTABLE_SIZE_U32(tableLog);
+    wkspSize -= sizeof(*wksp) + FSE_DTABLE_SIZE(tableLog);
+
+    CHECK_F( FSE_buildDTable_internal(wksp->dtable, wksp->ncount, maxSymbolValue, tableLog, workSpace, wkspSize) );
+
+    {
+        const void* ptr = wksp->dtable;
+        const FSE_DTableHeader* DTableH = (const FSE_DTableHeader*)ptr;
+        const U32 fastMode = DTableH->fastMode;
+
+        /* select fast mode (static) */
+        if (fastMode) return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, wksp->dtable, 1);
+        return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, wksp->dtable, 0);
+    }
+}
+
+/* Avoids the FORCE_INLINE of the _body() function. */
+static size_t FSE_decompress_wksp_body_default(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize)
+{
+    return FSE_decompress_wksp_body(dst, dstCapacity, cSrc, cSrcSize, maxLog, workSpace, wkspSize, 0);
+}
+
+#if DYNAMIC_BMI2
+TARGET_ATTRIBUTE("bmi2") static size_t FSE_decompress_wksp_body_bmi2(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize)
+{
+    return FSE_decompress_wksp_body(dst, dstCapacity, cSrc, cSrcSize, maxLog, workSpace, wkspSize, 1);
+}
+#endif
+
+size_t FSE_decompress_wksp_bmi2(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize, int bmi2)
+{
+#if DYNAMIC_BMI2
+    if (bmi2) {
+        return FSE_decompress_wksp_body_bmi2(dst, dstCapacity, cSrc, cSrcSize, maxLog, workSpace, wkspSize);
+    }
+#endif
+    (void)bmi2;
+    return FSE_decompress_wksp_body_default(dst, dstCapacity, cSrc, cSrcSize, maxLog, workSpace, wkspSize);
+}
+
+
+typedef FSE_DTable DTable_max_t[FSE_DTABLE_SIZE_U32(FSE_MAX_TABLELOG)];
+
+
+
+#endif   /* FSE_COMMONDEFS_ONLY */
diff --git a/lib/zstd/common/huf.h b/lib/zstd/common/huf.h
new file mode 100644
index 0000000..88c5586
--- /dev/null
+++ b/lib/zstd/common/huf.h
@@ -0,0 +1,356 @@
+/* ******************************************************************
+ * huff0 huffman codec,
+ * part of Finite State Entropy library
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ * You can contact the author at :
+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+
+#ifndef HUF_H_298734234
+#define HUF_H_298734234
+
+/* *** Dependencies *** */
+#include "zstd_deps.h"    /* size_t */
+
+
+/* *** library symbols visibility *** */
+/* Note : when linking with -fvisibility=hidden on gcc, or by default on Visual,
+ *        HUF symbols remain "private" (internal symbols for library only).
+ *        Set macro FSE_DLL_EXPORT to 1 if you want HUF symbols visible on DLL interface */
+#if defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) && defined(__GNUC__) && (__GNUC__ >= 4)
+#  define HUF_PUBLIC_API __attribute__ ((visibility ("default")))
+#elif defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1)   /* Visual expected */
+#  define HUF_PUBLIC_API __declspec(dllexport)
+#elif defined(FSE_DLL_IMPORT) && (FSE_DLL_IMPORT==1)
+#  define HUF_PUBLIC_API __declspec(dllimport)  /* not required, just to generate faster code (saves a function pointer load from IAT and an indirect jump) */
+#else
+#  define HUF_PUBLIC_API
+#endif
+
+
+/* ========================== */
+/* ***  simple functions  *** */
+/* ========================== */
+
+/* HUF_compress() :
+ *  Compress content from buffer 'src', of size 'srcSize', into buffer 'dst'.
+ * 'dst' buffer must be already allocated.
+ *  Compression runs faster if `dstCapacity` >= HUF_compressBound(srcSize).
+ * `srcSize` must be <= `HUF_BLOCKSIZE_MAX` == 128 KB.
+ * @return : size of compressed data (<= `dstCapacity`).
+ *  Special values : if return == 0, srcData is not compressible => Nothing is stored within dst !!!
+ *                   if HUF_isError(return), compression failed (more details using HUF_getErrorName())
+ */
+HUF_PUBLIC_API size_t HUF_compress(void* dst, size_t dstCapacity,
+                             const void* src, size_t srcSize);
+
+/* HUF_decompress() :
+ *  Decompress HUF data from buffer 'cSrc', of size 'cSrcSize',
+ *  into already allocated buffer 'dst', of minimum size 'dstSize'.
+ * `originalSize` : **must** be the ***exact*** size of original (uncompressed) data.
+ *  Note : in contrast with FSE, HUF_decompress can regenerate
+ *         RLE (cSrcSize==1) and uncompressed (cSrcSize==dstSize) data,
+ *         because it knows size to regenerate (originalSize).
+ * @return : size of regenerated data (== originalSize),
+ *           or an error code, which can be tested using HUF_isError()
+ */
+HUF_PUBLIC_API size_t HUF_decompress(void* dst,  size_t originalSize,
+                               const void* cSrc, size_t cSrcSize);
+
+
+/* ***   Tool functions *** */
+#define HUF_BLOCKSIZE_MAX (128 * 1024)                  /*< maximum input size for a single block compressed with HUF_compress */
+HUF_PUBLIC_API size_t HUF_compressBound(size_t size);   /*< maximum compressed size (worst case) */
+
+/* Error Management */
+HUF_PUBLIC_API unsigned    HUF_isError(size_t code);       /*< tells if a return value is an error code */
+HUF_PUBLIC_API const char* HUF_getErrorName(size_t code);  /*< provides error code string (useful for debugging) */
+
+
+/* ***   Advanced function   *** */
+
+/* HUF_compress2() :
+ *  Same as HUF_compress(), but offers control over `maxSymbolValue` and `tableLog`.
+ * `maxSymbolValue` must be <= HUF_SYMBOLVALUE_MAX .
+ * `tableLog` must be `<= HUF_TABLELOG_MAX` . */
+HUF_PUBLIC_API size_t HUF_compress2 (void* dst, size_t dstCapacity,
+                               const void* src, size_t srcSize,
+                               unsigned maxSymbolValue, unsigned tableLog);
+
+/* HUF_compress4X_wksp() :
+ *  Same as HUF_compress2(), but uses externally allocated `workSpace`.
+ * `workspace` must have minimum alignment of 4, and be at least as large as HUF_WORKSPACE_SIZE */
+#define HUF_WORKSPACE_SIZE ((6 << 10) + 256)
+#define HUF_WORKSPACE_SIZE_U32 (HUF_WORKSPACE_SIZE / sizeof(U32))
+HUF_PUBLIC_API size_t HUF_compress4X_wksp (void* dst, size_t dstCapacity,
+                                     const void* src, size_t srcSize,
+                                     unsigned maxSymbolValue, unsigned tableLog,
+                                     void* workSpace, size_t wkspSize);
+
+#endif   /* HUF_H_298734234 */
+
+/* ******************************************************************
+ *  WARNING !!
+ *  The following section contains advanced and experimental definitions
+ *  which shall never be used in the context of a dynamic library,
+ *  because they are not guaranteed to remain stable in the future.
+ *  Only consider them in association with static linking.
+ * *****************************************************************/
+#if !defined(HUF_H_HUF_STATIC_LINKING_ONLY)
+#define HUF_H_HUF_STATIC_LINKING_ONLY
+
+/* *** Dependencies *** */
+#include "mem.h"   /* U32 */
+#define FSE_STATIC_LINKING_ONLY
+#include "fse.h"
+
+
+/* *** Constants *** */
+#define HUF_TABLELOG_MAX      12      /* max runtime value of tableLog (due to static allocation); can be modified up to HUF_ABSOLUTEMAX_TABLELOG */
+#define HUF_TABLELOG_DEFAULT  11      /* default tableLog value when none specified */
+#define HUF_SYMBOLVALUE_MAX  255
+
+#define HUF_TABLELOG_ABSOLUTEMAX  15  /* absolute limit of HUF_MAX_TABLELOG. Beyond that value, code does not work */
+#if (HUF_TABLELOG_MAX > HUF_TABLELOG_ABSOLUTEMAX)
+#  error "HUF_TABLELOG_MAX is too large !"
+#endif
+
+
+/* ****************************************
+*  Static allocation
+******************************************/
+/* HUF buffer bounds */
+#define HUF_CTABLEBOUND 129
+#define HUF_BLOCKBOUND(size) (size + (size>>8) + 8)   /* only true when incompressible is pre-filtered with fast heuristic */
+#define HUF_COMPRESSBOUND(size) (HUF_CTABLEBOUND + HUF_BLOCKBOUND(size))   /* Macro version, useful for static allocation */
+
+/* static allocation of HUF's Compression Table */
+/* this is a private definition, just exposed for allocation and strict aliasing purpose. never EVER access its members directly */
+struct HUF_CElt_s {
+  U16  val;
+  BYTE nbBits;
+};   /* typedef'd to HUF_CElt */
+typedef struct HUF_CElt_s HUF_CElt;   /* consider it an incomplete type */
+#define HUF_CTABLE_SIZE_U32(maxSymbolValue)   ((maxSymbolValue)+1)   /* Use tables of U32, for proper alignment */
+#define HUF_CTABLE_SIZE(maxSymbolValue)       (HUF_CTABLE_SIZE_U32(maxSymbolValue) * sizeof(U32))
+#define HUF_CREATE_STATIC_CTABLE(name, maxSymbolValue) \
+    HUF_CElt name[HUF_CTABLE_SIZE_U32(maxSymbolValue)] /* no final ; */
+
+/* static allocation of HUF's DTable */
+typedef U32 HUF_DTable;
+#define HUF_DTABLE_SIZE(maxTableLog)   (1 + (1<<(maxTableLog)))
+#define HUF_CREATE_STATIC_DTABLEX1(DTable, maxTableLog) \
+        HUF_DTable DTable[HUF_DTABLE_SIZE((maxTableLog)-1)] = { ((U32)((maxTableLog)-1) * 0x01000001) }
+#define HUF_CREATE_STATIC_DTABLEX2(DTable, maxTableLog) \
+        HUF_DTable DTable[HUF_DTABLE_SIZE(maxTableLog)] = { ((U32)(maxTableLog) * 0x01000001) }
+
+
+/* ****************************************
+*  Advanced decompression functions
+******************************************/
+size_t HUF_decompress4X1 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /*< single-symbol decoder */
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress4X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /*< double-symbols decoder */
+#endif
+
+size_t HUF_decompress4X_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /*< decodes RLE and uncompressed */
+size_t HUF_decompress4X_hufOnly(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /*< considers RLE and uncompressed as errors */
+size_t HUF_decompress4X_hufOnly_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /*< considers RLE and uncompressed as errors */
+size_t HUF_decompress4X1_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /*< single-symbol decoder */
+size_t HUF_decompress4X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);   /*< single-symbol decoder */
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress4X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /*< double-symbols decoder */
+size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);   /*< double-symbols decoder */
+#endif
+
+
+/* ****************************************
+ *  HUF detailed API
+ * ****************************************/
+
+/*! HUF_compress() does the following:
+ *  1. count symbol occurrence from source[] into table count[] using FSE_count() (exposed within "fse.h")
+ *  2. (optional) refine tableLog using HUF_optimalTableLog()
+ *  3. build Huffman table from count using HUF_buildCTable()
+ *  4. save Huffman table to memory buffer using HUF_writeCTable()
+ *  5. encode the data stream using HUF_compress4X_usingCTable()
+ *
+ *  The following API allows targeting specific sub-functions for advanced tasks.
+ *  For example, it's possible to compress several blocks using the same 'CTable',
+ *  or to save and regenerate 'CTable' using external methods.
+ */
+unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue);
+size_t HUF_buildCTable (HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue, unsigned maxNbBits);   /* @return : maxNbBits; CTable and count can overlap. In which case, CTable will overwrite count content */
+size_t HUF_writeCTable (void* dst, size_t maxDstSize, const HUF_CElt* CTable, unsigned maxSymbolValue, unsigned huffLog);
+size_t HUF_writeCTable_wksp(void* dst, size_t maxDstSize, const HUF_CElt* CTable, unsigned maxSymbolValue, unsigned huffLog, void* workspace, size_t workspaceSize);
+size_t HUF_compress4X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable);
+size_t HUF_estimateCompressedSize(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue);
+int HUF_validateCTable(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue);
+
+typedef enum {
+   HUF_repeat_none,  /*< Cannot use the previous table */
+   HUF_repeat_check, /*< Can use the previous table but it must be checked. Note : The previous table must have been constructed by HUF_compress{1, 4}X_repeat */
+   HUF_repeat_valid  /*< Can use the previous table and it is assumed to be valid */
+ } HUF_repeat;
+/* HUF_compress4X_repeat() :
+ *  Same as HUF_compress4X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none.
+ *  If it uses hufTable it does not modify hufTable or repeat.
+ *  If it doesn't, it sets *repeat = HUF_repeat_none, and it sets hufTable to the table used.
+ *  If preferRepeat then the old table will always be used if valid. */
+size_t HUF_compress4X_repeat(void* dst, size_t dstSize,
+                       const void* src, size_t srcSize,
+                       unsigned maxSymbolValue, unsigned tableLog,
+                       void* workSpace, size_t wkspSize,    /*< `workSpace` must be aligned on 4-bytes boundaries, `wkspSize` must be >= HUF_WORKSPACE_SIZE */
+                       HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2);
+
+/* HUF_buildCTable_wksp() :
+ *  Same as HUF_buildCTable(), but using externally allocated scratch buffer.
+ * `workSpace` must be aligned on 4-bytes boundaries, and its size must be >= HUF_CTABLE_WORKSPACE_SIZE.
+ */
+#define HUF_CTABLE_WORKSPACE_SIZE_U32 (2*HUF_SYMBOLVALUE_MAX +1 +1)
+#define HUF_CTABLE_WORKSPACE_SIZE (HUF_CTABLE_WORKSPACE_SIZE_U32 * sizeof(unsigned))
+size_t HUF_buildCTable_wksp (HUF_CElt* tree,
+                       const unsigned* count, U32 maxSymbolValue, U32 maxNbBits,
+                             void* workSpace, size_t wkspSize);
+
+/*! HUF_readStats() :
+ *  Read compact Huffman tree, saved by HUF_writeCTable().
+ * `huffWeight` is destination buffer.
+ * @return : size read from `src` , or an error Code .
+ *  Note : Needed by HUF_readCTable() and HUF_readDTableXn() . */
+size_t HUF_readStats(BYTE* huffWeight, size_t hwSize,
+                     U32* rankStats, U32* nbSymbolsPtr, U32* tableLogPtr,
+                     const void* src, size_t srcSize);
+
+/*! HUF_readStats_wksp() :
+ * Same as HUF_readStats() but takes an external workspace which must be
+ * 4-byte aligned and its size must be >= HUF_READ_STATS_WORKSPACE_SIZE.
+ * If the CPU has BMI2 support, pass bmi2=1, otherwise pass bmi2=0.
+ */
+#define HUF_READ_STATS_WORKSPACE_SIZE_U32 FSE_DECOMPRESS_WKSP_SIZE_U32(6, HUF_TABLELOG_MAX-1)
+#define HUF_READ_STATS_WORKSPACE_SIZE (HUF_READ_STATS_WORKSPACE_SIZE_U32 * sizeof(unsigned))
+size_t HUF_readStats_wksp(BYTE* huffWeight, size_t hwSize,
+                          U32* rankStats, U32* nbSymbolsPtr, U32* tableLogPtr,
+                          const void* src, size_t srcSize,
+                          void* workspace, size_t wkspSize,
+                          int bmi2);
+
+/* HUF_readCTable() :
+ *  Loading a CTable saved with HUF_writeCTable() */
+size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize, unsigned *hasZeroWeights);
+
+/* HUF_getNbBits() :
+ *  Read nbBits from CTable symbolTable, for symbol `symbolValue` presumed <= HUF_SYMBOLVALUE_MAX
+ *  Note 1 : is not inlined, as HUF_CElt definition is private
+ *  Note 2 : const void* used, so that it can provide a statically allocated table as argument (which uses type U32) */
+U32 HUF_getNbBits(const void* symbolTable, U32 symbolValue);
+
+/*
+ * HUF_decompress() does the following:
+ * 1. select the decompression algorithm (X1, X2) based on pre-computed heuristics
+ * 2. build Huffman table from save, using HUF_readDTableX?()
+ * 3. decode 1 or 4 segments in parallel using HUF_decompress?X?_usingDTable()
+ */
+
+/* HUF_selectDecoder() :
+ *  Tells which decoder is likely to decode faster,
+ *  based on a set of pre-computed metrics.
+ * @return : 0==HUF_decompress4X1, 1==HUF_decompress4X2 .
+ *  Assumption : 0 < dstSize <= 128 KB */
+U32 HUF_selectDecoder (size_t dstSize, size_t cSrcSize);
+
+/*
+ *  The minimum workspace size for the `workSpace` used in
+ *  HUF_readDTableX1_wksp() and HUF_readDTableX2_wksp().
+ *
+ *  The space used depends on HUF_TABLELOG_MAX, ranging from ~1500 bytes when
+ *  HUF_TABLE_LOG_MAX=12 to ~1850 bytes when HUF_TABLE_LOG_MAX=15.
+ *  Buffer overflow errors may potentially occur if code modifications result in
+ *  a required workspace size greater than that specified in the following
+ *  macro.
+ */
+#define HUF_DECOMPRESS_WORKSPACE_SIZE ((2 << 10) + (1 << 9))
+#define HUF_DECOMPRESS_WORKSPACE_SIZE_U32 (HUF_DECOMPRESS_WORKSPACE_SIZE / sizeof(U32))
+
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_readDTableX1 (HUF_DTable* DTable, const void* src, size_t srcSize);
+size_t HUF_readDTableX1_wksp (HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize);
+#endif
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_readDTableX2 (HUF_DTable* DTable, const void* src, size_t srcSize);
+size_t HUF_readDTableX2_wksp (HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize);
+#endif
+
+size_t HUF_decompress4X_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_decompress4X1_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+#endif
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress4X2_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+#endif
+
+
+/* ====================== */
+/* single stream variants */
+/* ====================== */
+
+size_t HUF_compress1X (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog);
+size_t HUF_compress1X_wksp (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize);  /*< `workSpace` must be a table of at least HUF_WORKSPACE_SIZE_U32 unsigned */
+size_t HUF_compress1X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable);
+/* HUF_compress1X_repeat() :
+ *  Same as HUF_compress1X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none.
+ *  If it uses hufTable it does not modify hufTable or repeat.
+ *  If it doesn't, it sets *repeat = HUF_repeat_none, and it sets hufTable to the table used.
+ *  If preferRepeat then the old table will always be used if valid. */
+size_t HUF_compress1X_repeat(void* dst, size_t dstSize,
+                       const void* src, size_t srcSize,
+                       unsigned maxSymbolValue, unsigned tableLog,
+                       void* workSpace, size_t wkspSize,   /*< `workSpace` must be aligned on 4-bytes boundaries, `wkspSize` must be >= HUF_WORKSPACE_SIZE */
+                       HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2);
+
+size_t HUF_decompress1X1 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /* single-symbol decoder */
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress1X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /* double-symbol decoder */
+#endif
+
+size_t HUF_decompress1X_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);
+size_t HUF_decompress1X_DCtx_wksp (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_decompress1X1_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /*< single-symbol decoder */
+size_t HUF_decompress1X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);   /*< single-symbol decoder */
+#endif
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress1X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /*< double-symbols decoder */
+size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);   /*< double-symbols decoder */
+#endif
+
+size_t HUF_decompress1X_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);   /*< automatic selection of sing or double symbol decoder, based on DTable */
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_decompress1X1_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+#endif
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress1X2_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+#endif
+
+/* BMI2 variants.
+ * If the CPU has BMI2 support, pass bmi2=1, otherwise pass bmi2=0.
+ */
+size_t HUF_decompress1X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2);
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_decompress1X1_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2);
+#endif
+size_t HUF_decompress4X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2);
+size_t HUF_decompress4X_hufOnly_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2);
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_readDTableX1_wksp_bmi2(HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize, int bmi2);
+#endif
+
+#endif /* HUF_STATIC_LINKING_ONLY */
+
diff --git a/lib/zstd/common/mem.h b/lib/zstd/common/mem.h
new file mode 100644
index 0000000..dcdd586
--- /dev/null
+++ b/lib/zstd/common/mem.h
@@ -0,0 +1,259 @@
+/* SPDX-License-Identifier: GPL-2.0+ OR BSD-3-Clause */
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef MEM_H_MODULE
+#define MEM_H_MODULE
+
+/*-****************************************
+*  Dependencies
+******************************************/
+#include <asm/unaligned.h>  /* get_unaligned, put_unaligned* */
+#include <linux/compiler.h>  /* inline */
+#include <linux/swab.h>  /* swab32, swab64 */
+#include <linux/types.h>  /* size_t, ptrdiff_t */
+#include "debug.h"  /* DEBUG_STATIC_ASSERT */
+
+/*-****************************************
+*  Compiler specifics
+******************************************/
+#define MEM_STATIC static inline
+
+/*-**************************************************************
+*  Basic Types
+*****************************************************************/
+typedef uint8_t  BYTE;
+typedef uint16_t U16;
+typedef int16_t  S16;
+typedef uint32_t U32;
+typedef int32_t  S32;
+typedef uint64_t U64;
+typedef int64_t  S64;
+
+/*-**************************************************************
+*  Memory I/O API
+*****************************************************************/
+/*=== Static platform detection ===*/
+MEM_STATIC unsigned MEM_32bits(void);
+MEM_STATIC unsigned MEM_64bits(void);
+MEM_STATIC unsigned MEM_isLittleEndian(void);
+
+/*=== Native unaligned read/write ===*/
+MEM_STATIC U16 MEM_read16(const void* memPtr);
+MEM_STATIC U32 MEM_read32(const void* memPtr);
+MEM_STATIC U64 MEM_read64(const void* memPtr);
+MEM_STATIC size_t MEM_readST(const void* memPtr);
+
+MEM_STATIC void MEM_write16(void* memPtr, U16 value);
+MEM_STATIC void MEM_write32(void* memPtr, U32 value);
+MEM_STATIC void MEM_write64(void* memPtr, U64 value);
+
+/*=== Little endian unaligned read/write ===*/
+MEM_STATIC U16 MEM_readLE16(const void* memPtr);
+MEM_STATIC U32 MEM_readLE24(const void* memPtr);
+MEM_STATIC U32 MEM_readLE32(const void* memPtr);
+MEM_STATIC U64 MEM_readLE64(const void* memPtr);
+MEM_STATIC size_t MEM_readLEST(const void* memPtr);
+
+MEM_STATIC void MEM_writeLE16(void* memPtr, U16 val);
+MEM_STATIC void MEM_writeLE24(void* memPtr, U32 val);
+MEM_STATIC void MEM_writeLE32(void* memPtr, U32 val32);
+MEM_STATIC void MEM_writeLE64(void* memPtr, U64 val64);
+MEM_STATIC void MEM_writeLEST(void* memPtr, size_t val);
+
+/*=== Big endian unaligned read/write ===*/
+MEM_STATIC U32 MEM_readBE32(const void* memPtr);
+MEM_STATIC U64 MEM_readBE64(const void* memPtr);
+MEM_STATIC size_t MEM_readBEST(const void* memPtr);
+
+MEM_STATIC void MEM_writeBE32(void* memPtr, U32 val32);
+MEM_STATIC void MEM_writeBE64(void* memPtr, U64 val64);
+MEM_STATIC void MEM_writeBEST(void* memPtr, size_t val);
+
+/*=== Byteswap ===*/
+MEM_STATIC U32 MEM_swap32(U32 in);
+MEM_STATIC U64 MEM_swap64(U64 in);
+MEM_STATIC size_t MEM_swapST(size_t in);
+
+/*-**************************************************************
+*  Memory I/O Implementation
+*****************************************************************/
+MEM_STATIC unsigned MEM_32bits(void)
+{
+    return sizeof(size_t) == 4;
+}
+
+MEM_STATIC unsigned MEM_64bits(void)
+{
+    return sizeof(size_t) == 8;
+}
+
+#if defined(__LITTLE_ENDIAN)
+#define MEM_LITTLE_ENDIAN 1
+#else
+#define MEM_LITTLE_ENDIAN 0
+#endif
+
+MEM_STATIC unsigned MEM_isLittleEndian(void)
+{
+    return MEM_LITTLE_ENDIAN;
+}
+
+MEM_STATIC U16 MEM_read16(const void *memPtr)
+{
+    return get_unaligned((const U16 *)memPtr);
+}
+
+MEM_STATIC U32 MEM_read32(const void *memPtr)
+{
+    return get_unaligned((const U32 *)memPtr);
+}
+
+MEM_STATIC U64 MEM_read64(const void *memPtr)
+{
+    return get_unaligned((const U64 *)memPtr);
+}
+
+MEM_STATIC size_t MEM_readST(const void *memPtr)
+{
+    return get_unaligned((const size_t *)memPtr);
+}
+
+MEM_STATIC void MEM_write16(void *memPtr, U16 value)
+{
+    put_unaligned(value, (U16 *)memPtr);
+}
+
+MEM_STATIC void MEM_write32(void *memPtr, U32 value)
+{
+    put_unaligned(value, (U32 *)memPtr);
+}
+
+MEM_STATIC void MEM_write64(void *memPtr, U64 value)
+{
+    put_unaligned(value, (U64 *)memPtr);
+}
+
+/*=== Little endian r/w ===*/
+
+MEM_STATIC U16 MEM_readLE16(const void *memPtr)
+{
+    return get_unaligned_le16(memPtr);
+}
+
+MEM_STATIC void MEM_writeLE16(void *memPtr, U16 val)
+{
+    put_unaligned_le16(val, memPtr);
+}
+
+MEM_STATIC U32 MEM_readLE24(const void *memPtr)
+{
+    return MEM_readLE16(memPtr) + (((const BYTE *)memPtr)[2] << 16);
+}
+
+MEM_STATIC void MEM_writeLE24(void *memPtr, U32 val)
+{
+	MEM_writeLE16(memPtr, (U16)val);
+	((BYTE *)memPtr)[2] = (BYTE)(val >> 16);
+}
+
+MEM_STATIC U32 MEM_readLE32(const void *memPtr)
+{
+    return get_unaligned_le32(memPtr);
+}
+
+MEM_STATIC void MEM_writeLE32(void *memPtr, U32 val32)
+{
+    put_unaligned_le32(val32, memPtr);
+}
+
+MEM_STATIC U64 MEM_readLE64(const void *memPtr)
+{
+    return get_unaligned_le64(memPtr);
+}
+
+MEM_STATIC void MEM_writeLE64(void *memPtr, U64 val64)
+{
+    put_unaligned_le64(val64, memPtr);
+}
+
+MEM_STATIC size_t MEM_readLEST(const void *memPtr)
+{
+	if (MEM_32bits())
+		return (size_t)MEM_readLE32(memPtr);
+	else
+		return (size_t)MEM_readLE64(memPtr);
+}
+
+MEM_STATIC void MEM_writeLEST(void *memPtr, size_t val)
+{
+	if (MEM_32bits())
+		MEM_writeLE32(memPtr, (U32)val);
+	else
+		MEM_writeLE64(memPtr, (U64)val);
+}
+
+/*=== Big endian r/w ===*/
+
+MEM_STATIC U32 MEM_readBE32(const void *memPtr)
+{
+    return get_unaligned_be32(memPtr);
+}
+
+MEM_STATIC void MEM_writeBE32(void *memPtr, U32 val32)
+{
+    put_unaligned_be32(val32, memPtr);
+}
+
+MEM_STATIC U64 MEM_readBE64(const void *memPtr)
+{
+    return get_unaligned_be64(memPtr);
+}
+
+MEM_STATIC void MEM_writeBE64(void *memPtr, U64 val64)
+{
+    put_unaligned_be64(val64, memPtr);
+}
+
+MEM_STATIC size_t MEM_readBEST(const void *memPtr)
+{
+	if (MEM_32bits())
+		return (size_t)MEM_readBE32(memPtr);
+	else
+		return (size_t)MEM_readBE64(memPtr);
+}
+
+MEM_STATIC void MEM_writeBEST(void *memPtr, size_t val)
+{
+	if (MEM_32bits())
+		MEM_writeBE32(memPtr, (U32)val);
+	else
+		MEM_writeBE64(memPtr, (U64)val);
+}
+
+MEM_STATIC U32 MEM_swap32(U32 in)
+{
+    return swab32(in);
+}
+
+MEM_STATIC U64 MEM_swap64(U64 in)
+{
+    return swab64(in);
+}
+
+MEM_STATIC size_t MEM_swapST(size_t in)
+{
+    if (MEM_32bits())
+        return (size_t)MEM_swap32((U32)in);
+    else
+        return (size_t)MEM_swap64((U64)in);
+}
+
+#endif /* MEM_H_MODULE */
diff --git a/lib/zstd/common/zstd_common.c b/lib/zstd/common/zstd_common.c
new file mode 100644
index 0000000..3d7e35b
--- /dev/null
+++ b/lib/zstd/common/zstd_common.c
@@ -0,0 +1,83 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+
+/*-*************************************
+*  Dependencies
+***************************************/
+#define ZSTD_DEPS_NEED_MALLOC
+#include "zstd_deps.h"   /* ZSTD_malloc, ZSTD_calloc, ZSTD_free, ZSTD_memset */
+#include "error_private.h"
+#include "zstd_internal.h"
+
+
+/*-****************************************
+*  Version
+******************************************/
+unsigned ZSTD_versionNumber(void) { return ZSTD_VERSION_NUMBER; }
+
+const char* ZSTD_versionString(void) { return ZSTD_VERSION_STRING; }
+
+
+/*-****************************************
+*  ZSTD Error Management
+******************************************/
+#undef ZSTD_isError   /* defined within zstd_internal.h */
+/*! ZSTD_isError() :
+ *  tells if a return value is an error code
+ *  symbol is required for external callers */
+unsigned ZSTD_isError(size_t code) { return ERR_isError(code); }
+
+/*! ZSTD_getErrorName() :
+ *  provides error code string from function result (useful for debugging) */
+const char* ZSTD_getErrorName(size_t code) { return ERR_getErrorName(code); }
+
+/*! ZSTD_getError() :
+ *  convert a `size_t` function result into a proper ZSTD_errorCode enum */
+ZSTD_ErrorCode ZSTD_getErrorCode(size_t code) { return ERR_getErrorCode(code); }
+
+/*! ZSTD_getErrorString() :
+ *  provides error code string from enum */
+const char* ZSTD_getErrorString(ZSTD_ErrorCode code) { return ERR_getErrorString(code); }
+
+
+
+/*=**************************************************************
+*  Custom allocator
+****************************************************************/
+void* ZSTD_customMalloc(size_t size, ZSTD_customMem customMem)
+{
+    if (customMem.customAlloc)
+        return customMem.customAlloc(customMem.opaque, size);
+    return ZSTD_malloc(size);
+}
+
+void* ZSTD_customCalloc(size_t size, ZSTD_customMem customMem)
+{
+    if (customMem.customAlloc) {
+        /* calloc implemented as malloc+memset;
+         * not as efficient as calloc, but next best guess for custom malloc */
+        void* const ptr = customMem.customAlloc(customMem.opaque, size);
+        ZSTD_memset(ptr, 0, size);
+        return ptr;
+    }
+    return ZSTD_calloc(1, size);
+}
+
+void ZSTD_customFree(void* ptr, ZSTD_customMem customMem)
+{
+    if (ptr!=NULL) {
+        if (customMem.customFree)
+            customMem.customFree(customMem.opaque, ptr);
+        else
+            ZSTD_free(ptr);
+    }
+}
diff --git a/lib/zstd/common/zstd_deps.h b/lib/zstd/common/zstd_deps.h
new file mode 100644
index 0000000..7a5bf44
--- /dev/null
+++ b/lib/zstd/common/zstd_deps.h
@@ -0,0 +1,125 @@
+/* SPDX-License-Identifier: GPL-2.0+ OR BSD-3-Clause */
+/*
+ * Copyright (c) Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/*
+ * This file provides common libc dependencies that zstd requires.
+ * The purpose is to allow replacing this file with a custom implementation
+ * to compile zstd without libc support.
+ */
+
+/* Need:
+ * NULL
+ * INT_MAX
+ * UINT_MAX
+ * ZSTD_memcpy()
+ * ZSTD_memset()
+ * ZSTD_memmove()
+ */
+#ifndef ZSTD_DEPS_COMMON
+#define ZSTD_DEPS_COMMON
+
+#include <linux/limits.h>
+#include <linux/stddef.h>
+
+#define ZSTD_memcpy(d,s,n) __builtin_memcpy((d),(s),(n))
+#define ZSTD_memmove(d,s,n) __builtin_memmove((d),(s),(n))
+#define ZSTD_memset(d,s,n) __builtin_memset((d),(s),(n))
+
+#endif /* ZSTD_DEPS_COMMON */
+
+/*
+ * Define malloc as always failing. That means the user must
+ * either use ZSTD_customMem or statically allocate memory.
+ * Need:
+ * ZSTD_malloc()
+ * ZSTD_free()
+ * ZSTD_calloc()
+ */
+#ifdef ZSTD_DEPS_NEED_MALLOC
+#ifndef ZSTD_DEPS_MALLOC
+#define ZSTD_DEPS_MALLOC
+
+#define ZSTD_malloc(s) ({ (void)(s); NULL; })
+#define ZSTD_free(p) ((void)(p))
+#define ZSTD_calloc(n,s) ({ (void)(n); (void)(s); NULL; })
+
+#endif /* ZSTD_DEPS_MALLOC */
+#endif /* ZSTD_DEPS_NEED_MALLOC */
+
+/*
+ * Provides 64-bit math support.
+ * Need:
+ * U64 ZSTD_div64(U64 dividend, U32 divisor)
+ */
+#ifdef ZSTD_DEPS_NEED_MATH64
+#ifndef ZSTD_DEPS_MATH64
+#define ZSTD_DEPS_MATH64
+
+#include <linux/math64.h>
+
+static uint64_t ZSTD_div64(uint64_t dividend, uint32_t divisor) {
+  return div_u64(dividend, divisor);
+}
+
+#endif /* ZSTD_DEPS_MATH64 */
+#endif /* ZSTD_DEPS_NEED_MATH64 */
+
+/*
+ * This is only requested when DEBUGLEVEL >= 1, meaning
+ * it is disabled in production.
+ * Need:
+ * assert()
+ */
+#ifdef ZSTD_DEPS_NEED_ASSERT
+#ifndef ZSTD_DEPS_ASSERT
+#define ZSTD_DEPS_ASSERT
+
+#include <linux/kernel.h>
+
+#define assert(x) WARN_ON((x))
+
+#endif /* ZSTD_DEPS_ASSERT */
+#endif /* ZSTD_DEPS_NEED_ASSERT */
+
+/*
+ * This is only requested when DEBUGLEVEL >= 2, meaning
+ * it is disabled in production.
+ * Need:
+ * ZSTD_DEBUG_PRINT()
+ */
+#ifdef ZSTD_DEPS_NEED_IO
+#ifndef ZSTD_DEPS_IO
+#define ZSTD_DEPS_IO
+
+#include <linux/printk.h>
+
+#define ZSTD_DEBUG_PRINT(...) pr_debug(__VA_ARGS__)
+
+#endif /* ZSTD_DEPS_IO */
+#endif /* ZSTD_DEPS_NEED_IO */
+
+/*
+ * Only requested when MSAN is enabled.
+ * Need:
+ * intptr_t
+ */
+#ifdef ZSTD_DEPS_NEED_STDINT
+#ifndef ZSTD_DEPS_STDINT
+#define ZSTD_DEPS_STDINT
+
+/*
+ * The Linux Kernel doesn't provide intptr_t, only uintptr_t, which
+ * is an unsigned long.
+ */
+typedef long intptr_t;
+
+#endif /* ZSTD_DEPS_STDINT */
+#endif /* ZSTD_DEPS_NEED_STDINT */
diff --git a/lib/zstd/common/zstd_internal.h b/lib/zstd/common/zstd_internal.h
new file mode 100644
index 0000000..fc6f3a9
--- /dev/null
+++ b/lib/zstd/common/zstd_internal.h
@@ -0,0 +1,450 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_CCOMMON_H_MODULE
+#define ZSTD_CCOMMON_H_MODULE
+
+/* this module contains definitions which must be identical
+ * across compression, decompression and dictBuilder.
+ * It also contains a few functions useful to at least 2 of them
+ * and which benefit from being inlined */
+
+/*-*************************************
+*  Dependencies
+***************************************/
+#include "compiler.h"
+#include "mem.h"
+#include "debug.h"                 /* assert, DEBUGLOG, RAWLOG, g_debuglevel */
+#include "error_private.h"
+#define ZSTD_STATIC_LINKING_ONLY
+#include <linux/zstd.h>
+#define FSE_STATIC_LINKING_ONLY
+#include "fse.h"
+#define HUF_STATIC_LINKING_ONLY
+#include "huf.h"
+#include <linux/xxhash.h>                /* XXH_reset, update, digest */
+#define ZSTD_TRACE 0
+
+
+/* ---- static assert (debug) --- */
+#define ZSTD_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c)
+#define ZSTD_isError ERR_isError   /* for inlining */
+#define FSE_isError  ERR_isError
+#define HUF_isError  ERR_isError
+
+
+/*-*************************************
+*  shared macros
+***************************************/
+#undef MIN
+#undef MAX
+#define MIN(a,b) ((a)<(b) ? (a) : (b))
+#define MAX(a,b) ((a)>(b) ? (a) : (b))
+
+/*
+ * Ignore: this is an internal helper.
+ *
+ * This is a helper function to help force C99-correctness during compilation.
+ * Under strict compilation modes, variadic macro arguments can't be empty.
+ * However, variadic function arguments can be. Using a function therefore lets
+ * us statically check that at least one (string) argument was passed,
+ * independent of the compilation flags.
+ */
+static INLINE_KEYWORD UNUSED_ATTR
+void _force_has_format_string(const char *format, ...) {
+  (void)format;
+}
+
+/*
+ * Ignore: this is an internal helper.
+ *
+ * We want to force this function invocation to be syntactically correct, but
+ * we don't want to force runtime evaluation of its arguments.
+ */
+#define _FORCE_HAS_FORMAT_STRING(...) \
+  if (0) { \
+    _force_has_format_string(__VA_ARGS__); \
+  }
+
+/*
+ * Return the specified error if the condition evaluates to true.
+ *
+ * In debug modes, prints additional information.
+ * In order to do that (particularly, printing the conditional that failed),
+ * this can't just wrap RETURN_ERROR().
+ */
+#define RETURN_ERROR_IF(cond, err, ...) \
+  if (cond) { \
+    RAWLOG(3, "%s:%d: ERROR!: check %s failed, returning %s", \
+           __FILE__, __LINE__, ZSTD_QUOTE(cond), ZSTD_QUOTE(ERROR(err))); \
+    _FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
+    RAWLOG(3, ": " __VA_ARGS__); \
+    RAWLOG(3, "\n"); \
+    return ERROR(err); \
+  }
+
+/*
+ * Unconditionally return the specified error.
+ *
+ * In debug modes, prints additional information.
+ */
+#define RETURN_ERROR(err, ...) \
+  do { \
+    RAWLOG(3, "%s:%d: ERROR!: unconditional check failed, returning %s", \
+           __FILE__, __LINE__, ZSTD_QUOTE(ERROR(err))); \
+    _FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
+    RAWLOG(3, ": " __VA_ARGS__); \
+    RAWLOG(3, "\n"); \
+    return ERROR(err); \
+  } while(0);
+
+/*
+ * If the provided expression evaluates to an error code, returns that error code.
+ *
+ * In debug modes, prints additional information.
+ */
+#define FORWARD_IF_ERROR(err, ...) \
+  do { \
+    size_t const err_code = (err); \
+    if (ERR_isError(err_code)) { \
+      RAWLOG(3, "%s:%d: ERROR!: forwarding error in %s: %s", \
+             __FILE__, __LINE__, ZSTD_QUOTE(err), ERR_getErrorName(err_code)); \
+      _FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
+      RAWLOG(3, ": " __VA_ARGS__); \
+      RAWLOG(3, "\n"); \
+      return err_code; \
+    } \
+  } while(0);
+
+
+/*-*************************************
+*  Common constants
+***************************************/
+#define ZSTD_OPT_NUM    (1<<12)
+
+#define ZSTD_REP_NUM      3                 /* number of repcodes */
+#define ZSTD_REP_MOVE     (ZSTD_REP_NUM-1)
+static UNUSED_ATTR const U32 repStartValue[ZSTD_REP_NUM] = { 1, 4, 8 };
+
+#define KB *(1 <<10)
+#define MB *(1 <<20)
+#define GB *(1U<<30)
+
+#define BIT7 128
+#define BIT6  64
+#define BIT5  32
+#define BIT4  16
+#define BIT1   2
+#define BIT0   1
+
+#define ZSTD_WINDOWLOG_ABSOLUTEMIN 10
+static UNUSED_ATTR const size_t ZSTD_fcs_fieldSize[4] = { 0, 2, 4, 8 };
+static UNUSED_ATTR const size_t ZSTD_did_fieldSize[4] = { 0, 1, 2, 4 };
+
+#define ZSTD_FRAMEIDSIZE 4   /* magic number size */
+
+#define ZSTD_BLOCKHEADERSIZE 3   /* C standard doesn't allow `static const` variable to be init using another `static const` variable */
+static UNUSED_ATTR const size_t ZSTD_blockHeaderSize = ZSTD_BLOCKHEADERSIZE;
+typedef enum { bt_raw, bt_rle, bt_compressed, bt_reserved } blockType_e;
+
+#define ZSTD_FRAMECHECKSUMSIZE 4
+
+#define MIN_SEQUENCES_SIZE 1 /* nbSeq==0 */
+#define MIN_CBLOCK_SIZE (1 /*litCSize*/ + 1 /* RLE or RAW */ + MIN_SEQUENCES_SIZE /* nbSeq==0 */)   /* for a non-null block */
+
+#define HufLog 12
+typedef enum { set_basic, set_rle, set_compressed, set_repeat } symbolEncodingType_e;
+
+#define LONGNBSEQ 0x7F00
+
+#define MINMATCH 3
+
+#define Litbits  8
+#define MaxLit ((1<<Litbits) - 1)
+#define MaxML   52
+#define MaxLL   35
+#define DefaultMaxOff 28
+#define MaxOff  31
+#define MaxSeq MAX(MaxLL, MaxML)   /* Assumption : MaxOff < MaxLL,MaxML */
+#define MLFSELog    9
+#define LLFSELog    9
+#define OffFSELog   8
+#define MaxFSELog  MAX(MAX(MLFSELog, LLFSELog), OffFSELog)
+
+#define ZSTD_MAX_HUF_HEADER_SIZE 128 /* header + <= 127 byte tree description */
+/* Each table cannot take more than #symbols * FSELog bits */
+#define ZSTD_MAX_FSE_HEADERS_SIZE (((MaxML + 1) * MLFSELog + (MaxLL + 1) * LLFSELog + (MaxOff + 1) * OffFSELog + 7) / 8)
+
+static UNUSED_ATTR const U32 LL_bits[MaxLL+1] = {
+     0, 0, 0, 0, 0, 0, 0, 0,
+     0, 0, 0, 0, 0, 0, 0, 0,
+     1, 1, 1, 1, 2, 2, 3, 3,
+     4, 6, 7, 8, 9,10,11,12,
+    13,14,15,16
+};
+static UNUSED_ATTR const S16 LL_defaultNorm[MaxLL+1] = {
+     4, 3, 2, 2, 2, 2, 2, 2,
+     2, 2, 2, 2, 2, 1, 1, 1,
+     2, 2, 2, 2, 2, 2, 2, 2,
+     2, 3, 2, 1, 1, 1, 1, 1,
+    -1,-1,-1,-1
+};
+#define LL_DEFAULTNORMLOG 6  /* for static allocation */
+static UNUSED_ATTR const U32 LL_defaultNormLog = LL_DEFAULTNORMLOG;
+
+static UNUSED_ATTR const U32 ML_bits[MaxML+1] = {
+     0, 0, 0, 0, 0, 0, 0, 0,
+     0, 0, 0, 0, 0, 0, 0, 0,
+     0, 0, 0, 0, 0, 0, 0, 0,
+     0, 0, 0, 0, 0, 0, 0, 0,
+     1, 1, 1, 1, 2, 2, 3, 3,
+     4, 4, 5, 7, 8, 9,10,11,
+    12,13,14,15,16
+};
+static UNUSED_ATTR const S16 ML_defaultNorm[MaxML+1] = {
+     1, 4, 3, 2, 2, 2, 2, 2,
+     2, 1, 1, 1, 1, 1, 1, 1,
+     1, 1, 1, 1, 1, 1, 1, 1,
+     1, 1, 1, 1, 1, 1, 1, 1,
+     1, 1, 1, 1, 1, 1, 1, 1,
+     1, 1, 1, 1, 1, 1,-1,-1,
+    -1,-1,-1,-1,-1
+};
+#define ML_DEFAULTNORMLOG 6  /* for static allocation */
+static UNUSED_ATTR const U32 ML_defaultNormLog = ML_DEFAULTNORMLOG;
+
+static UNUSED_ATTR const S16 OF_defaultNorm[DefaultMaxOff+1] = {
+     1, 1, 1, 1, 1, 1, 2, 2,
+     2, 1, 1, 1, 1, 1, 1, 1,
+     1, 1, 1, 1, 1, 1, 1, 1,
+    -1,-1,-1,-1,-1
+};
+#define OF_DEFAULTNORMLOG 5  /* for static allocation */
+static UNUSED_ATTR const U32 OF_defaultNormLog = OF_DEFAULTNORMLOG;
+
+
+/*-*******************************************
+*  Shared functions to include for inlining
+*********************************************/
+static void ZSTD_copy8(void* dst, const void* src) {
+    ZSTD_memcpy(dst, src, 8);
+}
+
+#define COPY8(d,s) { ZSTD_copy8(d,s); d+=8; s+=8; }
+static void ZSTD_copy16(void* dst, const void* src) {
+    ZSTD_memcpy(dst, src, 16);
+}
+#define COPY16(d,s) { ZSTD_copy16(d,s); d+=16; s+=16; }
+
+#define WILDCOPY_OVERLENGTH 32
+#define WILDCOPY_VECLEN 16
+
+typedef enum {
+    ZSTD_no_overlap,
+    ZSTD_overlap_src_before_dst
+    /*  ZSTD_overlap_dst_before_src, */
+} ZSTD_overlap_e;
+
+/*! ZSTD_wildcopy() :
+ *  Custom version of ZSTD_memcpy(), can over read/write up to WILDCOPY_OVERLENGTH bytes (if length==0)
+ *  @param ovtype controls the overlap detection
+ *         - ZSTD_no_overlap: The source and destination are guaranteed to be at least WILDCOPY_VECLEN bytes apart.
+ *         - ZSTD_overlap_src_before_dst: The src and dst may overlap, but they MUST be at least 8 bytes apart.
+ *           The src buffer must be before the dst buffer.
+ */
+MEM_STATIC FORCE_INLINE_ATTR
+void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e const ovtype)
+{
+    ptrdiff_t diff = (BYTE*)dst - (const BYTE*)src;
+    const BYTE* ip = (const BYTE*)src;
+    BYTE* op = (BYTE*)dst;
+    BYTE* const oend = op + length;
+
+    assert(diff >= 8 || (ovtype == ZSTD_no_overlap && diff <= -WILDCOPY_VECLEN));
+
+    if (ovtype == ZSTD_overlap_src_before_dst && diff < WILDCOPY_VECLEN) {
+        /* Handle short offset copies. */
+        do {
+            COPY8(op, ip)
+        } while (op < oend);
+    } else {
+        assert(diff >= WILDCOPY_VECLEN || diff <= -WILDCOPY_VECLEN);
+        /* Separate out the first COPY16() call because the copy length is
+         * almost certain to be short, so the branches have different
+         * probabilities. Since it is almost certain to be short, only do
+         * one COPY16() in the first call. Then, do two calls per loop since
+         * at that point it is more likely to have a high trip count.
+         */
+#ifdef __aarch64__
+        do {
+            COPY16(op, ip);
+        }
+        while (op < oend);
+#else
+        ZSTD_copy16(op, ip);
+        if (16 >= length) return;
+        op += 16;
+        ip += 16;
+        do {
+            COPY16(op, ip);
+            COPY16(op, ip);
+        }
+        while (op < oend);
+#endif
+    }
+}
+
+MEM_STATIC size_t ZSTD_limitCopy(void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+    size_t const length = MIN(dstCapacity, srcSize);
+    if (length > 0) {
+        ZSTD_memcpy(dst, src, length);
+    }
+    return length;
+}
+
+/* define "workspace is too large" as this number of times larger than needed */
+#define ZSTD_WORKSPACETOOLARGE_FACTOR 3
+
+/* when workspace is continuously too large
+ * during at least this number of times,
+ * context's memory usage is considered wasteful,
+ * because it's sized to handle a worst case scenario which rarely happens.
+ * In which case, resize it down to free some memory */
+#define ZSTD_WORKSPACETOOLARGE_MAXDURATION 128
+
+/* Controls whether the input/output buffer is buffered or stable. */
+typedef enum {
+    ZSTD_bm_buffered = 0,  /* Buffer the input/output */
+    ZSTD_bm_stable = 1     /* ZSTD_inBuffer/ZSTD_outBuffer is stable */
+} ZSTD_bufferMode_e;
+
+
+/*-*******************************************
+*  Private declarations
+*********************************************/
+typedef struct seqDef_s {
+    U32 offset;         /* Offset code of the sequence */
+    U16 litLength;
+    U16 matchLength;
+} seqDef;
+
+typedef struct {
+    seqDef* sequencesStart;
+    seqDef* sequences;      /* ptr to end of sequences */
+    BYTE* litStart;
+    BYTE* lit;              /* ptr to end of literals */
+    BYTE* llCode;
+    BYTE* mlCode;
+    BYTE* ofCode;
+    size_t maxNbSeq;
+    size_t maxNbLit;
+
+    /* longLengthPos and longLengthID to allow us to represent either a single litLength or matchLength
+     * in the seqStore that has a value larger than U16 (if it exists). To do so, we increment
+     * the existing value of the litLength or matchLength by 0x10000.
+     */
+    U32   longLengthID;   /* 0 == no longLength; 1 == Represent the long literal; 2 == Represent the long match; */
+    U32   longLengthPos;  /* Index of the sequence to apply long length modification to */
+} seqStore_t;
+
+typedef struct {
+    U32 litLength;
+    U32 matchLength;
+} ZSTD_sequenceLength;
+
+/*
+ * Returns the ZSTD_sequenceLength for the given sequences. It handles the decoding of long sequences
+ * indicated by longLengthPos and longLengthID, and adds MINMATCH back to matchLength.
+ */
+MEM_STATIC ZSTD_sequenceLength ZSTD_getSequenceLength(seqStore_t const* seqStore, seqDef const* seq)
+{
+    ZSTD_sequenceLength seqLen;
+    seqLen.litLength = seq->litLength;
+    seqLen.matchLength = seq->matchLength + MINMATCH;
+    if (seqStore->longLengthPos == (U32)(seq - seqStore->sequencesStart)) {
+        if (seqStore->longLengthID == 1) {
+            seqLen.litLength += 0xFFFF;
+        }
+        if (seqStore->longLengthID == 2) {
+            seqLen.matchLength += 0xFFFF;
+        }
+    }
+    return seqLen;
+}
+
+/*
+ * Contains the compressed frame size and an upper-bound for the decompressed frame size.
+ * Note: before using `compressedSize`, check for errors using ZSTD_isError().
+ *       similarly, before using `decompressedBound`, check for errors using:
+ *          `decompressedBound != ZSTD_CONTENTSIZE_ERROR`
+ */
+typedef struct {
+    size_t compressedSize;
+    unsigned long long decompressedBound;
+} ZSTD_frameSizeInfo;   /* decompress & legacy */
+
+const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx);   /* compress & dictBuilder */
+void ZSTD_seqToCodes(const seqStore_t* seqStorePtr);   /* compress, dictBuilder, decodeCorpus (shouldn't get its definition from here) */
+
+/* custom memory allocation functions */
+void* ZSTD_customMalloc(size_t size, ZSTD_customMem customMem);
+void* ZSTD_customCalloc(size_t size, ZSTD_customMem customMem);
+void ZSTD_customFree(void* ptr, ZSTD_customMem customMem);
+
+
+MEM_STATIC U32 ZSTD_highbit32(U32 val)   /* compress, dictBuilder, decodeCorpus */
+{
+    assert(val != 0);
+    {
+#   if (__GNUC__ >= 3)   /* GCC Intrinsic */
+        return __builtin_clz (val) ^ 31;
+#   else   /* Software version */
+        static const U32 DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30, 8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31 };
+        U32 v = val;
+        v |= v >> 1;
+        v |= v >> 2;
+        v |= v >> 4;
+        v |= v >> 8;
+        v |= v >> 16;
+        return DeBruijnClz[(v * 0x07C4ACDDU) >> 27];
+#   endif
+    }
+}
+
+
+/* ZSTD_invalidateRepCodes() :
+ * ensures next compression will not use repcodes from previous block.
+ * Note : only works with regular variant;
+ *        do not use with extDict variant ! */
+void ZSTD_invalidateRepCodes(ZSTD_CCtx* cctx);   /* zstdmt, adaptive_compression (shouldn't get this definition from here) */
+
+
+typedef struct {
+    blockType_e blockType;
+    U32 lastBlock;
+    U32 origSize;
+} blockProperties_t;   /* declared here for decompress and fullbench */
+
+/*! ZSTD_getcBlockSize() :
+ *  Provides the size of compressed block from block header `src` */
+/* Used by: decompress, fullbench (does not get its definition from here) */
+size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
+                          blockProperties_t* bpPtr);
+
+/*! ZSTD_decodeSeqHeaders() :
+ *  decode sequence header from src */
+/* Used by: decompress, fullbench (does not get its definition from here) */
+size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,
+                       const void* src, size_t srcSize);
+
+
+
+#endif   /* ZSTD_CCOMMON_H_MODULE */
diff --git a/lib/zstd/compress.c b/lib/zstd/compress.c
deleted file mode 100644
index b080264..0000000
--- a/lib/zstd/compress.c
+++ /dev/null
@@ -1,3485 +0,0 @@
-/**
- * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
- * All rights reserved.
- *
- * This source code is licensed under the BSD-style license found in the
- * LICENSE file in the root directory of https://github.com/facebook/zstd.
- * An additional grant of patent rights can be found in the PATENTS file in the
- * same directory.
- *
- * This program is free software; you can redistribute it and/or modify it under
- * the terms of the GNU General Public License version 2 as published by the
- * Free Software Foundation. This program is dual-licensed; you may select
- * either version 2 of the GNU General Public License ("GPL") or BSD license
- * ("BSD").
- */
-
-/*-*************************************
-*  Dependencies
-***************************************/
-#include "fse.h"
-#include "huf.h"
-#include "mem.h"
-#include "zstd_internal.h" /* includes zstd.h */
-#include <linux/kernel.h>
-#include <linux/module.h>
-#include <linux/string.h> /* memset */
-
-/*-*************************************
-*  Constants
-***************************************/
-static const U32 g_searchStrength = 8; /* control skip over incompressible data */
-#define HASH_READ_SIZE 8
-typedef enum { ZSTDcs_created = 0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZSTD_compressionStage_e;
-
-/*-*************************************
-*  Helper functions
-***************************************/
-size_t ZSTD_compressBound(size_t srcSize) { return FSE_compressBound(srcSize) + 12; }
-
-/*-*************************************
-*  Sequence storage
-***************************************/
-static void ZSTD_resetSeqStore(seqStore_t *ssPtr)
-{
-	ssPtr->lit = ssPtr->litStart;
-	ssPtr->sequences = ssPtr->sequencesStart;
-	ssPtr->longLengthID = 0;
-}
-
-/*-*************************************
-*  Context memory management
-***************************************/
-struct ZSTD_CCtx_s {
-	const BYTE *nextSrc;  /* next block here to continue on curr prefix */
-	const BYTE *base;     /* All regular indexes relative to this position */
-	const BYTE *dictBase; /* extDict indexes relative to this position */
-	U32 dictLimit;	/* below that point, need extDict */
-	U32 lowLimit;	 /* below that point, no more data */
-	U32 nextToUpdate;     /* index from which to continue dictionary update */
-	U32 nextToUpdate3;    /* index from which to continue dictionary update */
-	U32 hashLog3;	 /* dispatch table : larger == faster, more memory */
-	U32 loadedDictEnd;    /* index of end of dictionary */
-	U32 forceWindow;      /* force back-references to respect limit of 1<<wLog, even for dictionary */
-	U32 forceRawDict;     /* Force loading dictionary in "content-only" mode (no header analysis) */
-	ZSTD_compressionStage_e stage;
-	U32 rep[ZSTD_REP_NUM];
-	U32 repToConfirm[ZSTD_REP_NUM];
-	U32 dictID;
-	ZSTD_parameters params;
-	void *workSpace;
-	size_t workSpaceSize;
-	size_t blockSize;
-	U64 frameContentSize;
-	struct xxh64_state xxhState;
-	ZSTD_customMem customMem;
-
-	seqStore_t seqStore; /* sequences storage ptrs */
-	U32 *hashTable;
-	U32 *hashTable3;
-	U32 *chainTable;
-	HUF_CElt *hufTable;
-	U32 flagStaticTables;
-	HUF_repeat flagStaticHufTable;
-	FSE_CTable offcodeCTable[FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)];
-	FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)];
-	FSE_CTable litlengthCTable[FSE_CTABLE_SIZE_U32(LLFSELog, MaxLL)];
-	unsigned tmpCounters[HUF_COMPRESS_WORKSPACE_SIZE_U32];
-};
-
-size_t ZSTD_CCtxWorkspaceBound(ZSTD_compressionParameters cParams)
-{
-	size_t const blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, (size_t)1 << cParams.windowLog);
-	U32 const divider = (cParams.searchLength == 3) ? 3 : 4;
-	size_t const maxNbSeq = blockSize / divider;
-	size_t const tokenSpace = blockSize + 11 * maxNbSeq;
-	size_t const chainSize = (cParams.strategy == ZSTD_fast) ? 0 : (1 << cParams.chainLog);
-	size_t const hSize = ((size_t)1) << cParams.hashLog;
-	U32 const hashLog3 = (cParams.searchLength > 3) ? 0 : MIN(ZSTD_HASHLOG3_MAX, cParams.windowLog);
-	size_t const h3Size = ((size_t)1) << hashLog3;
-	size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
-	size_t const optSpace =
-	    ((MaxML + 1) + (MaxLL + 1) + (MaxOff + 1) + (1 << Litbits)) * sizeof(U32) + (ZSTD_OPT_NUM + 1) * (sizeof(ZSTD_match_t) + sizeof(ZSTD_optimal_t));
-	size_t const workspaceSize = tableSpace + (256 * sizeof(U32)) /* huffTable */ + tokenSpace +
-				     (((cParams.strategy == ZSTD_btopt) || (cParams.strategy == ZSTD_btopt2)) ? optSpace : 0);
-
-	return ZSTD_ALIGN(sizeof(ZSTD_stack)) + ZSTD_ALIGN(sizeof(ZSTD_CCtx)) + ZSTD_ALIGN(workspaceSize);
-}
-
-static ZSTD_CCtx *ZSTD_createCCtx_advanced(ZSTD_customMem customMem)
-{
-	ZSTD_CCtx *cctx;
-	if (!customMem.customAlloc || !customMem.customFree)
-		return NULL;
-	cctx = (ZSTD_CCtx *)ZSTD_malloc(sizeof(ZSTD_CCtx), customMem);
-	if (!cctx)
-		return NULL;
-	memset(cctx, 0, sizeof(ZSTD_CCtx));
-	cctx->customMem = customMem;
-	return cctx;
-}
-
-ZSTD_CCtx *ZSTD_initCCtx(void *workspace, size_t workspaceSize)
-{
-	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-	ZSTD_CCtx *cctx = ZSTD_createCCtx_advanced(stackMem);
-	if (cctx) {
-		cctx->workSpace = ZSTD_stackAllocAll(cctx->customMem.opaque, &cctx->workSpaceSize);
-	}
-	return cctx;
-}
-
-size_t ZSTD_freeCCtx(ZSTD_CCtx *cctx)
-{
-	if (cctx == NULL)
-		return 0; /* support free on NULL */
-	ZSTD_free(cctx->workSpace, cctx->customMem);
-	ZSTD_free(cctx, cctx->customMem);
-	return 0; /* reserved as a potential error code in the future */
-}
-
-const seqStore_t *ZSTD_getSeqStore(const ZSTD_CCtx *ctx) /* hidden interface */ { return &(ctx->seqStore); }
-
-static ZSTD_parameters ZSTD_getParamsFromCCtx(const ZSTD_CCtx *cctx) { return cctx->params; }
-
-/** ZSTD_checkParams() :
-	ensure param values remain within authorized range.
-	@return : 0, or an error code if one value is beyond authorized range */
-size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams)
-{
-#define CLAMPCHECK(val, min, max)                                       \
-	{                                                               \
-		if ((val < min) | (val > max))                          \
-			return ERROR(compressionParameter_unsupported); \
-	}
-	CLAMPCHECK(cParams.windowLog, ZSTD_WINDOWLOG_MIN, ZSTD_WINDOWLOG_MAX);
-	CLAMPCHECK(cParams.chainLog, ZSTD_CHAINLOG_MIN, ZSTD_CHAINLOG_MAX);
-	CLAMPCHECK(cParams.hashLog, ZSTD_HASHLOG_MIN, ZSTD_HASHLOG_MAX);
-	CLAMPCHECK(cParams.searchLog, ZSTD_SEARCHLOG_MIN, ZSTD_SEARCHLOG_MAX);
-	CLAMPCHECK(cParams.searchLength, ZSTD_SEARCHLENGTH_MIN, ZSTD_SEARCHLENGTH_MAX);
-	CLAMPCHECK(cParams.targetLength, ZSTD_TARGETLENGTH_MIN, ZSTD_TARGETLENGTH_MAX);
-	if ((U32)(cParams.strategy) > (U32)ZSTD_btopt2)
-		return ERROR(compressionParameter_unsupported);
-	return 0;
-}
-
-/** ZSTD_cycleLog() :
- *  condition for correct operation : hashLog > 1 */
-static U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat)
-{
-	U32 const btScale = ((U32)strat >= (U32)ZSTD_btlazy2);
-	return hashLog - btScale;
-}
-
-/** ZSTD_adjustCParams() :
-	optimize `cPar` for a given input (`srcSize` and `dictSize`).
-	mostly downsizing to reduce memory consumption and initialization.
-	Both `srcSize` and `dictSize` are optional (use 0 if unknown),
-	but if both are 0, no optimization can be done.
-	Note : cPar is considered validated at this stage. Use ZSTD_checkParams() to ensure that. */
-ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize)
-{
-	if (srcSize + dictSize == 0)
-		return cPar; /* no size information available : no adjustment */
-
-	/* resize params, to use less memory when necessary */
-	{
-		U32 const minSrcSize = (srcSize == 0) ? 500 : 0;
-		U64 const rSize = srcSize + dictSize + minSrcSize;
-		if (rSize < ((U64)1 << ZSTD_WINDOWLOG_MAX)) {
-			U32 const srcLog = MAX(ZSTD_HASHLOG_MIN, ZSTD_highbit32((U32)(rSize)-1) + 1);
-			if (cPar.windowLog > srcLog)
-				cPar.windowLog = srcLog;
-		}
-	}
-	if (cPar.hashLog > cPar.windowLog)
-		cPar.hashLog = cPar.windowLog;
-	{
-		U32 const cycleLog = ZSTD_cycleLog(cPar.chainLog, cPar.strategy);
-		if (cycleLog > cPar.windowLog)
-			cPar.chainLog -= (cycleLog - cPar.windowLog);
-	}
-
-	if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN)
-		cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN; /* required for frame header */
-
-	return cPar;
-}
-
-static U32 ZSTD_equivalentParams(ZSTD_parameters param1, ZSTD_parameters param2)
-{
-	return (param1.cParams.hashLog == param2.cParams.hashLog) & (param1.cParams.chainLog == param2.cParams.chainLog) &
-	       (param1.cParams.strategy == param2.cParams.strategy) & ((param1.cParams.searchLength == 3) == (param2.cParams.searchLength == 3));
-}
-
-/*! ZSTD_continueCCtx() :
-	reuse CCtx without reset (note : requires no dictionary) */
-static size_t ZSTD_continueCCtx(ZSTD_CCtx *cctx, ZSTD_parameters params, U64 frameContentSize)
-{
-	U32 const end = (U32)(cctx->nextSrc - cctx->base);
-	cctx->params = params;
-	cctx->frameContentSize = frameContentSize;
-	cctx->lowLimit = end;
-	cctx->dictLimit = end;
-	cctx->nextToUpdate = end + 1;
-	cctx->stage = ZSTDcs_init;
-	cctx->dictID = 0;
-	cctx->loadedDictEnd = 0;
-	{
-		int i;
-		for (i = 0; i < ZSTD_REP_NUM; i++)
-			cctx->rep[i] = repStartValue[i];
-	}
-	cctx->seqStore.litLengthSum = 0; /* force reset of btopt stats */
-	xxh64_reset(&cctx->xxhState, 0);
-	return 0;
-}
-
-typedef enum { ZSTDcrp_continue, ZSTDcrp_noMemset, ZSTDcrp_fullReset } ZSTD_compResetPolicy_e;
-
-/*! ZSTD_resetCCtx_advanced() :
-	note : `params` must be validated */
-static size_t ZSTD_resetCCtx_advanced(ZSTD_CCtx *zc, ZSTD_parameters params, U64 frameContentSize, ZSTD_compResetPolicy_e const crp)
-{
-	if (crp == ZSTDcrp_continue)
-		if (ZSTD_equivalentParams(params, zc->params)) {
-			zc->flagStaticTables = 0;
-			zc->flagStaticHufTable = HUF_repeat_none;
-			return ZSTD_continueCCtx(zc, params, frameContentSize);
-		}
-
-	{
-		size_t const blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, (size_t)1 << params.cParams.windowLog);
-		U32 const divider = (params.cParams.searchLength == 3) ? 3 : 4;
-		size_t const maxNbSeq = blockSize / divider;
-		size_t const tokenSpace = blockSize + 11 * maxNbSeq;
-		size_t const chainSize = (params.cParams.strategy == ZSTD_fast) ? 0 : (1 << params.cParams.chainLog);
-		size_t const hSize = ((size_t)1) << params.cParams.hashLog;
-		U32 const hashLog3 = (params.cParams.searchLength > 3) ? 0 : MIN(ZSTD_HASHLOG3_MAX, params.cParams.windowLog);
-		size_t const h3Size = ((size_t)1) << hashLog3;
-		size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
-		void *ptr;
-
-		/* Check if workSpace is large enough, alloc a new one if needed */
-		{
-			size_t const optSpace = ((MaxML + 1) + (MaxLL + 1) + (MaxOff + 1) + (1 << Litbits)) * sizeof(U32) +
-						(ZSTD_OPT_NUM + 1) * (sizeof(ZSTD_match_t) + sizeof(ZSTD_optimal_t));
-			size_t const neededSpace = tableSpace + (256 * sizeof(U32)) /* huffTable */ + tokenSpace +
-						   (((params.cParams.strategy == ZSTD_btopt) || (params.cParams.strategy == ZSTD_btopt2)) ? optSpace : 0);
-			if (zc->workSpaceSize < neededSpace) {
-				ZSTD_free(zc->workSpace, zc->customMem);
-				zc->workSpace = ZSTD_malloc(neededSpace, zc->customMem);
-				if (zc->workSpace == NULL)
-					return ERROR(memory_allocation);
-				zc->workSpaceSize = neededSpace;
-			}
-		}
-
-		if (crp != ZSTDcrp_noMemset)
-			memset(zc->workSpace, 0, tableSpace); /* reset tables only */
-		xxh64_reset(&zc->xxhState, 0);
-		zc->hashLog3 = hashLog3;
-		zc->hashTable = (U32 *)(zc->workSpace);
-		zc->chainTable = zc->hashTable + hSize;
-		zc->hashTable3 = zc->chainTable + chainSize;
-		ptr = zc->hashTable3 + h3Size;
-		zc->hufTable = (HUF_CElt *)ptr;
-		zc->flagStaticTables = 0;
-		zc->flagStaticHufTable = HUF_repeat_none;
-		ptr = ((U32 *)ptr) + 256; /* note : HUF_CElt* is incomplete type, size is simulated using U32 */
-
-		zc->nextToUpdate = 1;
-		zc->nextSrc = NULL;
-		zc->base = NULL;
-		zc->dictBase = NULL;
-		zc->dictLimit = 0;
-		zc->lowLimit = 0;
-		zc->params = params;
-		zc->blockSize = blockSize;
-		zc->frameContentSize = frameContentSize;
-		{
-			int i;
-			for (i = 0; i < ZSTD_REP_NUM; i++)
-				zc->rep[i] = repStartValue[i];
-		}
-
-		if ((params.cParams.strategy == ZSTD_btopt) || (params.cParams.strategy == ZSTD_btopt2)) {
-			zc->seqStore.litFreq = (U32 *)ptr;
-			zc->seqStore.litLengthFreq = zc->seqStore.litFreq + (1 << Litbits);
-			zc->seqStore.matchLengthFreq = zc->seqStore.litLengthFreq + (MaxLL + 1);
-			zc->seqStore.offCodeFreq = zc->seqStore.matchLengthFreq + (MaxML + 1);
-			ptr = zc->seqStore.offCodeFreq + (MaxOff + 1);
-			zc->seqStore.matchTable = (ZSTD_match_t *)ptr;
-			ptr = zc->seqStore.matchTable + ZSTD_OPT_NUM + 1;
-			zc->seqStore.priceTable = (ZSTD_optimal_t *)ptr;
-			ptr = zc->seqStore.priceTable + ZSTD_OPT_NUM + 1;
-			zc->seqStore.litLengthSum = 0;
-		}
-		zc->seqStore.sequencesStart = (seqDef *)ptr;
-		ptr = zc->seqStore.sequencesStart + maxNbSeq;
-		zc->seqStore.llCode = (BYTE *)ptr;
-		zc->seqStore.mlCode = zc->seqStore.llCode + maxNbSeq;
-		zc->seqStore.ofCode = zc->seqStore.mlCode + maxNbSeq;
-		zc->seqStore.litStart = zc->seqStore.ofCode + maxNbSeq;
-
-		zc->stage = ZSTDcs_init;
-		zc->dictID = 0;
-		zc->loadedDictEnd = 0;
-
-		return 0;
-	}
-}
-
-/* ZSTD_invalidateRepCodes() :
- * ensures next compression will not use repcodes from previous block.
- * Note : only works with regular variant;
- *        do not use with extDict variant ! */
-void ZSTD_invalidateRepCodes(ZSTD_CCtx *cctx)
-{
-	int i;
-	for (i = 0; i < ZSTD_REP_NUM; i++)
-		cctx->rep[i] = 0;
-}
-
-/*! ZSTD_copyCCtx() :
-*   Duplicate an existing context `srcCCtx` into another one `dstCCtx`.
-*   Only works during stage ZSTDcs_init (i.e. after creation, but before first call to ZSTD_compressContinue()).
-*   @return : 0, or an error code */
-size_t ZSTD_copyCCtx(ZSTD_CCtx *dstCCtx, const ZSTD_CCtx *srcCCtx, unsigned long long pledgedSrcSize)
-{
-	if (srcCCtx->stage != ZSTDcs_init)
-		return ERROR(stage_wrong);
-
-	memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem));
-	{
-		ZSTD_parameters params = srcCCtx->params;
-		params.fParams.contentSizeFlag = (pledgedSrcSize > 0);
-		ZSTD_resetCCtx_advanced(dstCCtx, params, pledgedSrcSize, ZSTDcrp_noMemset);
-	}
-
-	/* copy tables */
-	{
-		size_t const chainSize = (srcCCtx->params.cParams.strategy == ZSTD_fast) ? 0 : (1 << srcCCtx->params.cParams.chainLog);
-		size_t const hSize = ((size_t)1) << srcCCtx->params.cParams.hashLog;
-		size_t const h3Size = (size_t)1 << srcCCtx->hashLog3;
-		size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
-		memcpy(dstCCtx->workSpace, srcCCtx->workSpace, tableSpace);
-	}
-
-	/* copy dictionary offsets */
-	dstCCtx->nextToUpdate = srcCCtx->nextToUpdate;
-	dstCCtx->nextToUpdate3 = srcCCtx->nextToUpdate3;
-	dstCCtx->nextSrc = srcCCtx->nextSrc;
-	dstCCtx->base = srcCCtx->base;
-	dstCCtx->dictBase = srcCCtx->dictBase;
-	dstCCtx->dictLimit = srcCCtx->dictLimit;
-	dstCCtx->lowLimit = srcCCtx->lowLimit;
-	dstCCtx->loadedDictEnd = srcCCtx->loadedDictEnd;
-	dstCCtx->dictID = srcCCtx->dictID;
-
-	/* copy entropy tables */
-	dstCCtx->flagStaticTables = srcCCtx->flagStaticTables;
-	dstCCtx->flagStaticHufTable = srcCCtx->flagStaticHufTable;
-	if (srcCCtx->flagStaticTables) {
-		memcpy(dstCCtx->litlengthCTable, srcCCtx->litlengthCTable, sizeof(dstCCtx->litlengthCTable));
-		memcpy(dstCCtx->matchlengthCTable, srcCCtx->matchlengthCTable, sizeof(dstCCtx->matchlengthCTable));
-		memcpy(dstCCtx->offcodeCTable, srcCCtx->offcodeCTable, sizeof(dstCCtx->offcodeCTable));
-	}
-	if (srcCCtx->flagStaticHufTable) {
-		memcpy(dstCCtx->hufTable, srcCCtx->hufTable, 256 * 4);
-	}
-
-	return 0;
-}
-
-/*! ZSTD_reduceTable() :
-*   reduce table indexes by `reducerValue` */
-static void ZSTD_reduceTable(U32 *const table, U32 const size, U32 const reducerValue)
-{
-	U32 u;
-	for (u = 0; u < size; u++) {
-		if (table[u] < reducerValue)
-			table[u] = 0;
-		else
-			table[u] -= reducerValue;
-	}
-}
-
-/*! ZSTD_reduceIndex() :
-*   rescale all indexes to avoid future overflow (indexes are U32) */
-static void ZSTD_reduceIndex(ZSTD_CCtx *zc, const U32 reducerValue)
-{
-	{
-		U32 const hSize = 1 << zc->params.cParams.hashLog;
-		ZSTD_reduceTable(zc->hashTable, hSize, reducerValue);
-	}
-
-	{
-		U32 const chainSize = (zc->params.cParams.strategy == ZSTD_fast) ? 0 : (1 << zc->params.cParams.chainLog);
-		ZSTD_reduceTable(zc->chainTable, chainSize, reducerValue);
-	}
-
-	{
-		U32 const h3Size = (zc->hashLog3) ? 1 << zc->hashLog3 : 0;
-		ZSTD_reduceTable(zc->hashTable3, h3Size, reducerValue);
-	}
-}
-
-/*-*******************************************************
-*  Block entropic compression
-*********************************************************/
-
-/* See doc/zstd_compression_format.md for detailed format description */
-
-size_t ZSTD_noCompressBlock(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	if (srcSize + ZSTD_blockHeaderSize > dstCapacity)
-		return ERROR(dstSize_tooSmall);
-	memcpy((BYTE *)dst + ZSTD_blockHeaderSize, src, srcSize);
-	ZSTD_writeLE24(dst, (U32)(srcSize << 2) + (U32)bt_raw);
-	return ZSTD_blockHeaderSize + srcSize;
-}
-
-static size_t ZSTD_noCompressLiterals(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	BYTE *const ostart = (BYTE * const)dst;
-	U32 const flSize = 1 + (srcSize > 31) + (srcSize > 4095);
-
-	if (srcSize + flSize > dstCapacity)
-		return ERROR(dstSize_tooSmall);
-
-	switch (flSize) {
-	case 1: /* 2 - 1 - 5 */ ostart[0] = (BYTE)((U32)set_basic + (srcSize << 3)); break;
-	case 2: /* 2 - 2 - 12 */ ZSTD_writeLE16(ostart, (U16)((U32)set_basic + (1 << 2) + (srcSize << 4))); break;
-	default: /*note : should not be necessary : flSize is within {1,2,3} */
-	case 3: /* 2 - 2 - 20 */ ZSTD_writeLE32(ostart, (U32)((U32)set_basic + (3 << 2) + (srcSize << 4))); break;
-	}
-
-	memcpy(ostart + flSize, src, srcSize);
-	return srcSize + flSize;
-}
-
-static size_t ZSTD_compressRleLiteralsBlock(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	BYTE *const ostart = (BYTE * const)dst;
-	U32 const flSize = 1 + (srcSize > 31) + (srcSize > 4095);
-
-	(void)dstCapacity; /* dstCapacity already guaranteed to be >=4, hence large enough */
-
-	switch (flSize) {
-	case 1: /* 2 - 1 - 5 */ ostart[0] = (BYTE)((U32)set_rle + (srcSize << 3)); break;
-	case 2: /* 2 - 2 - 12 */ ZSTD_writeLE16(ostart, (U16)((U32)set_rle + (1 << 2) + (srcSize << 4))); break;
-	default: /*note : should not be necessary : flSize is necessarily within {1,2,3} */
-	case 3: /* 2 - 2 - 20 */ ZSTD_writeLE32(ostart, (U32)((U32)set_rle + (3 << 2) + (srcSize << 4))); break;
-	}
-
-	ostart[flSize] = *(const BYTE *)src;
-	return flSize + 1;
-}
-
-static size_t ZSTD_minGain(size_t srcSize) { return (srcSize >> 6) + 2; }
-
-static size_t ZSTD_compressLiterals(ZSTD_CCtx *zc, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	size_t const minGain = ZSTD_minGain(srcSize);
-	size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB);
-	BYTE *const ostart = (BYTE *)dst;
-	U32 singleStream = srcSize < 256;
-	symbolEncodingType_e hType = set_compressed;
-	size_t cLitSize;
-
-/* small ? don't even attempt compression (speed opt) */
-#define LITERAL_NOENTROPY 63
-	{
-		size_t const minLitSize = zc->flagStaticHufTable == HUF_repeat_valid ? 6 : LITERAL_NOENTROPY;
-		if (srcSize <= minLitSize)
-			return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
-	}
-
-	if (dstCapacity < lhSize + 1)
-		return ERROR(dstSize_tooSmall); /* not enough space for compression */
-	{
-		HUF_repeat repeat = zc->flagStaticHufTable;
-		int const preferRepeat = zc->params.cParams.strategy < ZSTD_lazy ? srcSize <= 1024 : 0;
-		if (repeat == HUF_repeat_valid && lhSize == 3)
-			singleStream = 1;
-		cLitSize = singleStream ? HUF_compress1X_repeat(ostart + lhSize, dstCapacity - lhSize, src, srcSize, 255, 11, zc->tmpCounters,
-								sizeof(zc->tmpCounters), zc->hufTable, &repeat, preferRepeat)
-					: HUF_compress4X_repeat(ostart + lhSize, dstCapacity - lhSize, src, srcSize, 255, 11, zc->tmpCounters,
-								sizeof(zc->tmpCounters), zc->hufTable, &repeat, preferRepeat);
-		if (repeat != HUF_repeat_none) {
-			hType = set_repeat;
-		} /* reused the existing table */
-		else {
-			zc->flagStaticHufTable = HUF_repeat_check;
-		} /* now have a table to reuse */
-	}
-
-	if ((cLitSize == 0) | (cLitSize >= srcSize - minGain)) {
-		zc->flagStaticHufTable = HUF_repeat_none;
-		return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
-	}
-	if (cLitSize == 1) {
-		zc->flagStaticHufTable = HUF_repeat_none;
-		return ZSTD_compressRleLiteralsBlock(dst, dstCapacity, src, srcSize);
-	}
-
-	/* Build header */
-	switch (lhSize) {
-	case 3: /* 2 - 2 - 10 - 10 */
-	{
-		U32 const lhc = hType + ((!singleStream) << 2) + ((U32)srcSize << 4) + ((U32)cLitSize << 14);
-		ZSTD_writeLE24(ostart, lhc);
-		break;
-	}
-	case 4: /* 2 - 2 - 14 - 14 */
-	{
-		U32 const lhc = hType + (2 << 2) + ((U32)srcSize << 4) + ((U32)cLitSize << 18);
-		ZSTD_writeLE32(ostart, lhc);
-		break;
-	}
-	default: /* should not be necessary, lhSize is only {3,4,5} */
-	case 5:  /* 2 - 2 - 18 - 18 */
-	{
-		U32 const lhc = hType + (3 << 2) + ((U32)srcSize << 4) + ((U32)cLitSize << 22);
-		ZSTD_writeLE32(ostart, lhc);
-		ostart[4] = (BYTE)(cLitSize >> 10);
-		break;
-	}
-	}
-	return lhSize + cLitSize;
-}
-
-static const BYTE LL_Code[64] = {0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  10, 11, 12, 13, 14, 15, 16, 16, 17, 17, 18, 18,
-				 19, 19, 20, 20, 20, 20, 21, 21, 21, 21, 22, 22, 22, 22, 22, 22, 22, 22, 23, 23, 23, 23,
-				 23, 23, 23, 23, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24};
-
-static const BYTE ML_Code[128] = {0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,
-				  26, 27, 28, 29, 30, 31, 32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 36, 36, 37, 37, 37, 37, 38, 38, 38, 38,
-				  38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40,
-				  40, 40, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 42, 42, 42, 42, 42, 42, 42, 42,
-				  42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42};
-
-void ZSTD_seqToCodes(const seqStore_t *seqStorePtr)
-{
-	BYTE const LL_deltaCode = 19;
-	BYTE const ML_deltaCode = 36;
-	const seqDef *const sequences = seqStorePtr->sequencesStart;
-	BYTE *const llCodeTable = seqStorePtr->llCode;
-	BYTE *const ofCodeTable = seqStorePtr->ofCode;
-	BYTE *const mlCodeTable = seqStorePtr->mlCode;
-	U32 const nbSeq = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
-	U32 u;
-	for (u = 0; u < nbSeq; u++) {
-		U32 const llv = sequences[u].litLength;
-		U32 const mlv = sequences[u].matchLength;
-		llCodeTable[u] = (llv > 63) ? (BYTE)ZSTD_highbit32(llv) + LL_deltaCode : LL_Code[llv];
-		ofCodeTable[u] = (BYTE)ZSTD_highbit32(sequences[u].offset);
-		mlCodeTable[u] = (mlv > 127) ? (BYTE)ZSTD_highbit32(mlv) + ML_deltaCode : ML_Code[mlv];
-	}
-	if (seqStorePtr->longLengthID == 1)
-		llCodeTable[seqStorePtr->longLengthPos] = MaxLL;
-	if (seqStorePtr->longLengthID == 2)
-		mlCodeTable[seqStorePtr->longLengthPos] = MaxML;
-}
-
-ZSTD_STATIC size_t ZSTD_compressSequences_internal(ZSTD_CCtx *zc, void *dst, size_t dstCapacity)
-{
-	const int longOffsets = zc->params.cParams.windowLog > STREAM_ACCUMULATOR_MIN;
-	const seqStore_t *seqStorePtr = &(zc->seqStore);
-	FSE_CTable *CTable_LitLength = zc->litlengthCTable;
-	FSE_CTable *CTable_OffsetBits = zc->offcodeCTable;
-	FSE_CTable *CTable_MatchLength = zc->matchlengthCTable;
-	U32 LLtype, Offtype, MLtype; /* compressed, raw or rle */
-	const seqDef *const sequences = seqStorePtr->sequencesStart;
-	const BYTE *const ofCodeTable = seqStorePtr->ofCode;
-	const BYTE *const llCodeTable = seqStorePtr->llCode;
-	const BYTE *const mlCodeTable = seqStorePtr->mlCode;
-	BYTE *const ostart = (BYTE *)dst;
-	BYTE *const oend = ostart + dstCapacity;
-	BYTE *op = ostart;
-	size_t const nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart;
-	BYTE *seqHead;
-
-	U32 *count;
-	S16 *norm;
-	U32 *workspace;
-	size_t workspaceSize = sizeof(zc->tmpCounters);
-	{
-		size_t spaceUsed32 = 0;
-		count = (U32 *)zc->tmpCounters + spaceUsed32;
-		spaceUsed32 += MaxSeq + 1;
-		norm = (S16 *)((U32 *)zc->tmpCounters + spaceUsed32);
-		spaceUsed32 += ALIGN(sizeof(S16) * (MaxSeq + 1), sizeof(U32)) >> 2;
-
-		workspace = (U32 *)zc->tmpCounters + spaceUsed32;
-		workspaceSize -= (spaceUsed32 << 2);
-	}
-
-	/* Compress literals */
-	{
-		const BYTE *const literals = seqStorePtr->litStart;
-		size_t const litSize = seqStorePtr->lit - literals;
-		size_t const cSize = ZSTD_compressLiterals(zc, op, dstCapacity, literals, litSize);
-		if (ZSTD_isError(cSize))
-			return cSize;
-		op += cSize;
-	}
-
-	/* Sequences Header */
-	if ((oend - op) < 3 /*max nbSeq Size*/ + 1 /*seqHead */)
-		return ERROR(dstSize_tooSmall);
-	if (nbSeq < 0x7F)
-		*op++ = (BYTE)nbSeq;
-	else if (nbSeq < LONGNBSEQ)
-		op[0] = (BYTE)((nbSeq >> 8) + 0x80), op[1] = (BYTE)nbSeq, op += 2;
-	else
-		op[0] = 0xFF, ZSTD_writeLE16(op + 1, (U16)(nbSeq - LONGNBSEQ)), op += 3;
-	if (nbSeq == 0)
-		return op - ostart;
-
-	/* seqHead : flags for FSE encoding type */
-	seqHead = op++;
-
-#define MIN_SEQ_FOR_DYNAMIC_FSE 64
-#define MAX_SEQ_FOR_STATIC_FSE 1000
-
-	/* convert length/distances into codes */
-	ZSTD_seqToCodes(seqStorePtr);
-
-	/* CTable for Literal Lengths */
-	{
-		U32 max = MaxLL;
-		size_t const mostFrequent = FSE_countFast_wksp(count, &max, llCodeTable, nbSeq, workspace);
-		if ((mostFrequent == nbSeq) && (nbSeq > 2)) {
-			*op++ = llCodeTable[0];
-			FSE_buildCTable_rle(CTable_LitLength, (BYTE)max);
-			LLtype = set_rle;
-		} else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
-			LLtype = set_repeat;
-		} else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (LL_defaultNormLog - 1)))) {
-			FSE_buildCTable_wksp(CTable_LitLength, LL_defaultNorm, MaxLL, LL_defaultNormLog, workspace, workspaceSize);
-			LLtype = set_basic;
-		} else {
-			size_t nbSeq_1 = nbSeq;
-			const U32 tableLog = FSE_optimalTableLog(LLFSELog, nbSeq, max);
-			if (count[llCodeTable[nbSeq - 1]] > 1) {
-				count[llCodeTable[nbSeq - 1]]--;
-				nbSeq_1--;
-			}
-			FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max);
-			{
-				size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog); /* overflow protected */
-				if (FSE_isError(NCountSize))
-					return NCountSize;
-				op += NCountSize;
-			}
-			FSE_buildCTable_wksp(CTable_LitLength, norm, max, tableLog, workspace, workspaceSize);
-			LLtype = set_compressed;
-		}
-	}
-
-	/* CTable for Offsets */
-	{
-		U32 max = MaxOff;
-		size_t const mostFrequent = FSE_countFast_wksp(count, &max, ofCodeTable, nbSeq, workspace);
-		if ((mostFrequent == nbSeq) && (nbSeq > 2)) {
-			*op++ = ofCodeTable[0];
-			FSE_buildCTable_rle(CTable_OffsetBits, (BYTE)max);
-			Offtype = set_rle;
-		} else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
-			Offtype = set_repeat;
-		} else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (OF_defaultNormLog - 1)))) {
-			FSE_buildCTable_wksp(CTable_OffsetBits, OF_defaultNorm, MaxOff, OF_defaultNormLog, workspace, workspaceSize);
-			Offtype = set_basic;
-		} else {
-			size_t nbSeq_1 = nbSeq;
-			const U32 tableLog = FSE_optimalTableLog(OffFSELog, nbSeq, max);
-			if (count[ofCodeTable[nbSeq - 1]] > 1) {
-				count[ofCodeTable[nbSeq - 1]]--;
-				nbSeq_1--;
-			}
-			FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max);
-			{
-				size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog); /* overflow protected */
-				if (FSE_isError(NCountSize))
-					return NCountSize;
-				op += NCountSize;
-			}
-			FSE_buildCTable_wksp(CTable_OffsetBits, norm, max, tableLog, workspace, workspaceSize);
-			Offtype = set_compressed;
-		}
-	}
-
-	/* CTable for MatchLengths */
-	{
-		U32 max = MaxML;
-		size_t const mostFrequent = FSE_countFast_wksp(count, &max, mlCodeTable, nbSeq, workspace);
-		if ((mostFrequent == nbSeq) && (nbSeq > 2)) {
-			*op++ = *mlCodeTable;
-			FSE_buildCTable_rle(CTable_MatchLength, (BYTE)max);
-			MLtype = set_rle;
-		} else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
-			MLtype = set_repeat;
-		} else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (ML_defaultNormLog - 1)))) {
-			FSE_buildCTable_wksp(CTable_MatchLength, ML_defaultNorm, MaxML, ML_defaultNormLog, workspace, workspaceSize);
-			MLtype = set_basic;
-		} else {
-			size_t nbSeq_1 = nbSeq;
-			const U32 tableLog = FSE_optimalTableLog(MLFSELog, nbSeq, max);
-			if (count[mlCodeTable[nbSeq - 1]] > 1) {
-				count[mlCodeTable[nbSeq - 1]]--;
-				nbSeq_1--;
-			}
-			FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max);
-			{
-				size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog); /* overflow protected */
-				if (FSE_isError(NCountSize))
-					return NCountSize;
-				op += NCountSize;
-			}
-			FSE_buildCTable_wksp(CTable_MatchLength, norm, max, tableLog, workspace, workspaceSize);
-			MLtype = set_compressed;
-		}
-	}
-
-	*seqHead = (BYTE)((LLtype << 6) + (Offtype << 4) + (MLtype << 2));
-	zc->flagStaticTables = 0;
-
-	/* Encoding Sequences */
-	{
-		BIT_CStream_t blockStream;
-		FSE_CState_t stateMatchLength;
-		FSE_CState_t stateOffsetBits;
-		FSE_CState_t stateLitLength;
-
-		CHECK_E(BIT_initCStream(&blockStream, op, oend - op), dstSize_tooSmall); /* not enough space remaining */
-
-		/* first symbols */
-		FSE_initCState2(&stateMatchLength, CTable_MatchLength, mlCodeTable[nbSeq - 1]);
-		FSE_initCState2(&stateOffsetBits, CTable_OffsetBits, ofCodeTable[nbSeq - 1]);
-		FSE_initCState2(&stateLitLength, CTable_LitLength, llCodeTable[nbSeq - 1]);
-		BIT_addBits(&blockStream, sequences[nbSeq - 1].litLength, LL_bits[llCodeTable[nbSeq - 1]]);
-		if (ZSTD_32bits())
-			BIT_flushBits(&blockStream);
-		BIT_addBits(&blockStream, sequences[nbSeq - 1].matchLength, ML_bits[mlCodeTable[nbSeq - 1]]);
-		if (ZSTD_32bits())
-			BIT_flushBits(&blockStream);
-		if (longOffsets) {
-			U32 const ofBits = ofCodeTable[nbSeq - 1];
-			int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN - 1);
-			if (extraBits) {
-				BIT_addBits(&blockStream, sequences[nbSeq - 1].offset, extraBits);
-				BIT_flushBits(&blockStream);
-			}
-			BIT_addBits(&blockStream, sequences[nbSeq - 1].offset >> extraBits, ofBits - extraBits);
-		} else {
-			BIT_addBits(&blockStream, sequences[nbSeq - 1].offset, ofCodeTable[nbSeq - 1]);
-		}
-		BIT_flushBits(&blockStream);
-
-		{
-			size_t n;
-			for (n = nbSeq - 2; n < nbSeq; n--) { /* intentional underflow */
-				BYTE const llCode = llCodeTable[n];
-				BYTE const ofCode = ofCodeTable[n];
-				BYTE const mlCode = mlCodeTable[n];
-				U32 const llBits = LL_bits[llCode];
-				U32 const ofBits = ofCode; /* 32b*/ /* 64b*/
-				U32 const mlBits = ML_bits[mlCode];
-				/* (7)*/							    /* (7)*/
-				FSE_encodeSymbol(&blockStream, &stateOffsetBits, ofCode); /* 15 */  /* 15 */
-				FSE_encodeSymbol(&blockStream, &stateMatchLength, mlCode); /* 24 */ /* 24 */
-				if (ZSTD_32bits())
-					BIT_flushBits(&blockStream);				  /* (7)*/
-				FSE_encodeSymbol(&blockStream, &stateLitLength, llCode); /* 16 */ /* 33 */
-				if (ZSTD_32bits() || (ofBits + mlBits + llBits >= 64 - 7 - (LLFSELog + MLFSELog + OffFSELog)))
-					BIT_flushBits(&blockStream); /* (7)*/
-				BIT_addBits(&blockStream, sequences[n].litLength, llBits);
-				if (ZSTD_32bits() && ((llBits + mlBits) > 24))
-					BIT_flushBits(&blockStream);
-				BIT_addBits(&blockStream, sequences[n].matchLength, mlBits);
-				if (ZSTD_32bits())
-					BIT_flushBits(&blockStream); /* (7)*/
-				if (longOffsets) {
-					int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN - 1);
-					if (extraBits) {
-						BIT_addBits(&blockStream, sequences[n].offset, extraBits);
-						BIT_flushBits(&blockStream); /* (7)*/
-					}
-					BIT_addBits(&blockStream, sequences[n].offset >> extraBits, ofBits - extraBits); /* 31 */
-				} else {
-					BIT_addBits(&blockStream, sequences[n].offset, ofBits); /* 31 */
-				}
-				BIT_flushBits(&blockStream); /* (7)*/
-			}
-		}
-
-		FSE_flushCState(&blockStream, &stateMatchLength);
-		FSE_flushCState(&blockStream, &stateOffsetBits);
-		FSE_flushCState(&blockStream, &stateLitLength);
-
-		{
-			size_t const streamSize = BIT_closeCStream(&blockStream);
-			if (streamSize == 0)
-				return ERROR(dstSize_tooSmall); /* not enough space */
-			op += streamSize;
-		}
-	}
-	return op - ostart;
-}
-
-ZSTD_STATIC size_t ZSTD_compressSequences(ZSTD_CCtx *zc, void *dst, size_t dstCapacity, size_t srcSize)
-{
-	size_t const cSize = ZSTD_compressSequences_internal(zc, dst, dstCapacity);
-	size_t const minGain = ZSTD_minGain(srcSize);
-	size_t const maxCSize = srcSize - minGain;
-	/* If the srcSize <= dstCapacity, then there is enough space to write a
-	 * raw uncompressed block. Since we ran out of space, the block must not
-	 * be compressible, so fall back to a raw uncompressed block.
-	 */
-	int const uncompressibleError = cSize == ERROR(dstSize_tooSmall) && srcSize <= dstCapacity;
-	int i;
-
-	if (ZSTD_isError(cSize) && !uncompressibleError)
-		return cSize;
-	if (cSize >= maxCSize || uncompressibleError) {
-		zc->flagStaticHufTable = HUF_repeat_none;
-		return 0;
-	}
-	/* confirm repcodes */
-	for (i = 0; i < ZSTD_REP_NUM; i++)
-		zc->rep[i] = zc->repToConfirm[i];
-	return cSize;
-}
-
-/*! ZSTD_storeSeq() :
-	Store a sequence (literal length, literals, offset code and match length code) into seqStore_t.
-	`offsetCode` : distance to match, or 0 == repCode.
-	`matchCode` : matchLength - MINMATCH
-*/
-ZSTD_STATIC void ZSTD_storeSeq(seqStore_t *seqStorePtr, size_t litLength, const void *literals, U32 offsetCode, size_t matchCode)
-{
-	/* copy Literals */
-	ZSTD_wildcopy(seqStorePtr->lit, literals, litLength);
-	seqStorePtr->lit += litLength;
-
-	/* literal Length */
-	if (litLength > 0xFFFF) {
-		seqStorePtr->longLengthID = 1;
-		seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
-	}
-	seqStorePtr->sequences[0].litLength = (U16)litLength;
-
-	/* match offset */
-	seqStorePtr->sequences[0].offset = offsetCode + 1;
-
-	/* match Length */
-	if (matchCode > 0xFFFF) {
-		seqStorePtr->longLengthID = 2;
-		seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
-	}
-	seqStorePtr->sequences[0].matchLength = (U16)matchCode;
-
-	seqStorePtr->sequences++;
-}
-
-/*-*************************************
-*  Match length counter
-***************************************/
-static unsigned ZSTD_NbCommonBytes(register size_t val)
-{
-	if (ZSTD_isLittleEndian()) {
-		if (ZSTD_64bits()) {
-			return (__builtin_ctzll((U64)val) >> 3);
-		} else { /* 32 bits */
-			return (__builtin_ctz((U32)val) >> 3);
-		}
-	} else { /* Big Endian CPU */
-		if (ZSTD_64bits()) {
-			return (__builtin_clzll(val) >> 3);
-		} else { /* 32 bits */
-			return (__builtin_clz((U32)val) >> 3);
-		}
-	}
-}
-
-static size_t ZSTD_count(const BYTE *pIn, const BYTE *pMatch, const BYTE *const pInLimit)
-{
-	const BYTE *const pStart = pIn;
-	const BYTE *const pInLoopLimit = pInLimit - (sizeof(size_t) - 1);
-
-	while (pIn < pInLoopLimit) {
-		size_t const diff = ZSTD_readST(pMatch) ^ ZSTD_readST(pIn);
-		if (!diff) {
-			pIn += sizeof(size_t);
-			pMatch += sizeof(size_t);
-			continue;
-		}
-		pIn += ZSTD_NbCommonBytes(diff);
-		return (size_t)(pIn - pStart);
-	}
-	if (ZSTD_64bits())
-		if ((pIn < (pInLimit - 3)) && (ZSTD_read32(pMatch) == ZSTD_read32(pIn))) {
-			pIn += 4;
-			pMatch += 4;
-		}
-	if ((pIn < (pInLimit - 1)) && (ZSTD_read16(pMatch) == ZSTD_read16(pIn))) {
-		pIn += 2;
-		pMatch += 2;
-	}
-	if ((pIn < pInLimit) && (*pMatch == *pIn))
-		pIn++;
-	return (size_t)(pIn - pStart);
-}
-
-/** ZSTD_count_2segments() :
-*   can count match length with `ip` & `match` in 2 different segments.
-*   convention : on reaching mEnd, match count continue starting from iStart
-*/
-static size_t ZSTD_count_2segments(const BYTE *ip, const BYTE *match, const BYTE *iEnd, const BYTE *mEnd, const BYTE *iStart)
-{
-	const BYTE *const vEnd = MIN(ip + (mEnd - match), iEnd);
-	size_t const matchLength = ZSTD_count(ip, match, vEnd);
-	if (match + matchLength != mEnd)
-		return matchLength;
-	return matchLength + ZSTD_count(ip + matchLength, iStart, iEnd);
-}
-
-/*-*************************************
-*  Hashes
-***************************************/
-static const U32 prime3bytes = 506832829U;
-static U32 ZSTD_hash3(U32 u, U32 h) { return ((u << (32 - 24)) * prime3bytes) >> (32 - h); }
-ZSTD_STATIC size_t ZSTD_hash3Ptr(const void *ptr, U32 h) { return ZSTD_hash3(ZSTD_readLE32(ptr), h); } /* only in zstd_opt.h */
-
-static const U32 prime4bytes = 2654435761U;
-static U32 ZSTD_hash4(U32 u, U32 h) { return (u * prime4bytes) >> (32 - h); }
-static size_t ZSTD_hash4Ptr(const void *ptr, U32 h) { return ZSTD_hash4(ZSTD_read32(ptr), h); }
-
-static const U64 prime5bytes = 889523592379ULL;
-static size_t ZSTD_hash5(U64 u, U32 h) { return (size_t)(((u << (64 - 40)) * prime5bytes) >> (64 - h)); }
-static size_t ZSTD_hash5Ptr(const void *p, U32 h) { return ZSTD_hash5(ZSTD_readLE64(p), h); }
-
-static const U64 prime6bytes = 227718039650203ULL;
-static size_t ZSTD_hash6(U64 u, U32 h) { return (size_t)(((u << (64 - 48)) * prime6bytes) >> (64 - h)); }
-static size_t ZSTD_hash6Ptr(const void *p, U32 h) { return ZSTD_hash6(ZSTD_readLE64(p), h); }
-
-static const U64 prime7bytes = 58295818150454627ULL;
-static size_t ZSTD_hash7(U64 u, U32 h) { return (size_t)(((u << (64 - 56)) * prime7bytes) >> (64 - h)); }
-static size_t ZSTD_hash7Ptr(const void *p, U32 h) { return ZSTD_hash7(ZSTD_readLE64(p), h); }
-
-static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL;
-static size_t ZSTD_hash8(U64 u, U32 h) { return (size_t)(((u)*prime8bytes) >> (64 - h)); }
-static size_t ZSTD_hash8Ptr(const void *p, U32 h) { return ZSTD_hash8(ZSTD_readLE64(p), h); }
-
-static size_t ZSTD_hashPtr(const void *p, U32 hBits, U32 mls)
-{
-	switch (mls) {
-	// case 3: return ZSTD_hash3Ptr(p, hBits);
-	default:
-	case 4: return ZSTD_hash4Ptr(p, hBits);
-	case 5: return ZSTD_hash5Ptr(p, hBits);
-	case 6: return ZSTD_hash6Ptr(p, hBits);
-	case 7: return ZSTD_hash7Ptr(p, hBits);
-	case 8: return ZSTD_hash8Ptr(p, hBits);
-	}
-}
-
-/*-*************************************
-*  Fast Scan
-***************************************/
-static void ZSTD_fillHashTable(ZSTD_CCtx *zc, const void *end, const U32 mls)
-{
-	U32 *const hashTable = zc->hashTable;
-	U32 const hBits = zc->params.cParams.hashLog;
-	const BYTE *const base = zc->base;
-	const BYTE *ip = base + zc->nextToUpdate;
-	const BYTE *const iend = ((const BYTE *)end) - HASH_READ_SIZE;
-	const size_t fastHashFillStep = 3;
-
-	while (ip <= iend) {
-		hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip - base);
-		ip += fastHashFillStep;
-	}
-}
-
-FORCE_INLINE
-void ZSTD_compressBlock_fast_generic(ZSTD_CCtx *cctx, const void *src, size_t srcSize, const U32 mls)
-{
-	U32 *const hashTable = cctx->hashTable;
-	U32 const hBits = cctx->params.cParams.hashLog;
-	seqStore_t *seqStorePtr = &(cctx->seqStore);
-	const BYTE *const base = cctx->base;
-	const BYTE *const istart = (const BYTE *)src;
-	const BYTE *ip = istart;
-	const BYTE *anchor = istart;
-	const U32 lowestIndex = cctx->dictLimit;
-	const BYTE *const lowest = base + lowestIndex;
-	const BYTE *const iend = istart + srcSize;
-	const BYTE *const ilimit = iend - HASH_READ_SIZE;
-	U32 offset_1 = cctx->rep[0], offset_2 = cctx->rep[1];
-	U32 offsetSaved = 0;
-
-	/* init */
-	ip += (ip == lowest);
-	{
-		U32 const maxRep = (U32)(ip - lowest);
-		if (offset_2 > maxRep)
-			offsetSaved = offset_2, offset_2 = 0;
-		if (offset_1 > maxRep)
-			offsetSaved = offset_1, offset_1 = 0;
-	}
-
-	/* Main Search Loop */
-	while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */
-		size_t mLength;
-		size_t const h = ZSTD_hashPtr(ip, hBits, mls);
-		U32 const curr = (U32)(ip - base);
-		U32 const matchIndex = hashTable[h];
-		const BYTE *match = base + matchIndex;
-		hashTable[h] = curr; /* update hash table */
-
-		if ((offset_1 > 0) & (ZSTD_read32(ip + 1 - offset_1) == ZSTD_read32(ip + 1))) {
-			mLength = ZSTD_count(ip + 1 + 4, ip + 1 + 4 - offset_1, iend) + 4;
-			ip++;
-			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, 0, mLength - MINMATCH);
-		} else {
-			U32 offset;
-			if ((matchIndex <= lowestIndex) || (ZSTD_read32(match) != ZSTD_read32(ip))) {
-				ip += ((ip - anchor) >> g_searchStrength) + 1;
-				continue;
-			}
-			mLength = ZSTD_count(ip + 4, match + 4, iend) + 4;
-			offset = (U32)(ip - match);
-			while (((ip > anchor) & (match > lowest)) && (ip[-1] == match[-1])) {
-				ip--;
-				match--;
-				mLength++;
-			} /* catch up */
-			offset_2 = offset_1;
-			offset_1 = offset;
-
-			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, offset + ZSTD_REP_MOVE, mLength - MINMATCH);
-		}
-
-		/* match found */
-		ip += mLength;
-		anchor = ip;
-
-		if (ip <= ilimit) {
-			/* Fill Table */
-			hashTable[ZSTD_hashPtr(base + curr + 2, hBits, mls)] = curr + 2; /* here because curr+2 could be > iend-8 */
-			hashTable[ZSTD_hashPtr(ip - 2, hBits, mls)] = (U32)(ip - 2 - base);
-			/* check immediate repcode */
-			while ((ip <= ilimit) && ((offset_2 > 0) & (ZSTD_read32(ip) == ZSTD_read32(ip - offset_2)))) {
-				/* store sequence */
-				size_t const rLength = ZSTD_count(ip + 4, ip + 4 - offset_2, iend) + 4;
-				{
-					U32 const tmpOff = offset_2;
-					offset_2 = offset_1;
-					offset_1 = tmpOff;
-				} /* swap offset_2 <=> offset_1 */
-				hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip - base);
-				ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength - MINMATCH);
-				ip += rLength;
-				anchor = ip;
-				continue; /* faster when present ... (?) */
-			}
-		}
-	}
-
-	/* save reps for next block */
-	cctx->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
-	cctx->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
-
-	/* Last Literals */
-	{
-		size_t const lastLLSize = iend - anchor;
-		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-		seqStorePtr->lit += lastLLSize;
-	}
-}
-
-static void ZSTD_compressBlock_fast(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-	const U32 mls = ctx->params.cParams.searchLength;
-	switch (mls) {
-	default: /* includes case 3 */
-	case 4: ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 4); return;
-	case 5: ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 5); return;
-	case 6: ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 6); return;
-	case 7: ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 7); return;
-	}
-}
-
-static void ZSTD_compressBlock_fast_extDict_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const U32 mls)
-{
-	U32 *hashTable = ctx->hashTable;
-	const U32 hBits = ctx->params.cParams.hashLog;
-	seqStore_t *seqStorePtr = &(ctx->seqStore);
-	const BYTE *const base = ctx->base;
-	const BYTE *const dictBase = ctx->dictBase;
-	const BYTE *const istart = (const BYTE *)src;
-	const BYTE *ip = istart;
-	const BYTE *anchor = istart;
-	const U32 lowestIndex = ctx->lowLimit;
-	const BYTE *const dictStart = dictBase + lowestIndex;
-	const U32 dictLimit = ctx->dictLimit;
-	const BYTE *const lowPrefixPtr = base + dictLimit;
-	const BYTE *const dictEnd = dictBase + dictLimit;
-	const BYTE *const iend = istart + srcSize;
-	const BYTE *const ilimit = iend - 8;
-	U32 offset_1 = ctx->rep[0], offset_2 = ctx->rep[1];
-
-	/* Search Loop */
-	while (ip < ilimit) { /* < instead of <=, because (ip+1) */
-		const size_t h = ZSTD_hashPtr(ip, hBits, mls);
-		const U32 matchIndex = hashTable[h];
-		const BYTE *matchBase = matchIndex < dictLimit ? dictBase : base;
-		const BYTE *match = matchBase + matchIndex;
-		const U32 curr = (U32)(ip - base);
-		const U32 repIndex = curr + 1 - offset_1; /* offset_1 expected <= curr +1 */
-		const BYTE *repBase = repIndex < dictLimit ? dictBase : base;
-		const BYTE *repMatch = repBase + repIndex;
-		size_t mLength;
-		hashTable[h] = curr; /* update hash table */
-
-		if ((((U32)((dictLimit - 1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex)) &&
-		    (ZSTD_read32(repMatch) == ZSTD_read32(ip + 1))) {
-			const BYTE *repMatchEnd = repIndex < dictLimit ? dictEnd : iend;
-			mLength = ZSTD_count_2segments(ip + 1 + EQUAL_READ32, repMatch + EQUAL_READ32, iend, repMatchEnd, lowPrefixPtr) + EQUAL_READ32;
-			ip++;
-			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, 0, mLength - MINMATCH);
-		} else {
-			if ((matchIndex < lowestIndex) || (ZSTD_read32(match) != ZSTD_read32(ip))) {
-				ip += ((ip - anchor) >> g_searchStrength) + 1;
-				continue;
-			}
-			{
-				const BYTE *matchEnd = matchIndex < dictLimit ? dictEnd : iend;
-				const BYTE *lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;
-				U32 offset;
-				mLength = ZSTD_count_2segments(ip + EQUAL_READ32, match + EQUAL_READ32, iend, matchEnd, lowPrefixPtr) + EQUAL_READ32;
-				while (((ip > anchor) & (match > lowMatchPtr)) && (ip[-1] == match[-1])) {
-					ip--;
-					match--;
-					mLength++;
-				} /* catch up */
-				offset = curr - matchIndex;
-				offset_2 = offset_1;
-				offset_1 = offset;
-				ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, offset + ZSTD_REP_MOVE, mLength - MINMATCH);
-			}
-		}
-
-		/* found a match : store it */
-		ip += mLength;
-		anchor = ip;
-
-		if (ip <= ilimit) {
-			/* Fill Table */
-			hashTable[ZSTD_hashPtr(base + curr + 2, hBits, mls)] = curr + 2;
-			hashTable[ZSTD_hashPtr(ip - 2, hBits, mls)] = (U32)(ip - 2 - base);
-			/* check immediate repcode */
-			while (ip <= ilimit) {
-				U32 const curr2 = (U32)(ip - base);
-				U32 const repIndex2 = curr2 - offset_2;
-				const BYTE *repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;
-				if ((((U32)((dictLimit - 1) - repIndex2) >= 3) & (repIndex2 > lowestIndex)) /* intentional overflow */
-				    && (ZSTD_read32(repMatch2) == ZSTD_read32(ip))) {
-					const BYTE *const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;
-					size_t repLength2 =
-					    ZSTD_count_2segments(ip + EQUAL_READ32, repMatch2 + EQUAL_READ32, iend, repEnd2, lowPrefixPtr) + EQUAL_READ32;
-					U32 tmpOffset = offset_2;
-					offset_2 = offset_1;
-					offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
-					ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2 - MINMATCH);
-					hashTable[ZSTD_hashPtr(ip, hBits, mls)] = curr2;
-					ip += repLength2;
-					anchor = ip;
-					continue;
-				}
-				break;
-			}
-		}
-	}
-
-	/* save reps for next block */
-	ctx->repToConfirm[0] = offset_1;
-	ctx->repToConfirm[1] = offset_2;
-
-	/* Last Literals */
-	{
-		size_t const lastLLSize = iend - anchor;
-		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-		seqStorePtr->lit += lastLLSize;
-	}
-}
-
-static void ZSTD_compressBlock_fast_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-	U32 const mls = ctx->params.cParams.searchLength;
-	switch (mls) {
-	default: /* includes case 3 */
-	case 4: ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 4); return;
-	case 5: ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 5); return;
-	case 6: ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 6); return;
-	case 7: ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 7); return;
-	}
-}
-
-/*-*************************************
-*  Double Fast
-***************************************/
-static void ZSTD_fillDoubleHashTable(ZSTD_CCtx *cctx, const void *end, const U32 mls)
-{
-	U32 *const hashLarge = cctx->hashTable;
-	U32 const hBitsL = cctx->params.cParams.hashLog;
-	U32 *const hashSmall = cctx->chainTable;
-	U32 const hBitsS = cctx->params.cParams.chainLog;
-	const BYTE *const base = cctx->base;
-	const BYTE *ip = base + cctx->nextToUpdate;
-	const BYTE *const iend = ((const BYTE *)end) - HASH_READ_SIZE;
-	const size_t fastHashFillStep = 3;
-
-	while (ip <= iend) {
-		hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip - base);
-		hashLarge[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip - base);
-		ip += fastHashFillStep;
-	}
-}
-
-FORCE_INLINE
-void ZSTD_compressBlock_doubleFast_generic(ZSTD_CCtx *cctx, const void *src, size_t srcSize, const U32 mls)
-{
-	U32 *const hashLong = cctx->hashTable;
-	const U32 hBitsL = cctx->params.cParams.hashLog;
-	U32 *const hashSmall = cctx->chainTable;
-	const U32 hBitsS = cctx->params.cParams.chainLog;
-	seqStore_t *seqStorePtr = &(cctx->seqStore);
-	const BYTE *const base = cctx->base;
-	const BYTE *const istart = (const BYTE *)src;
-	const BYTE *ip = istart;
-	const BYTE *anchor = istart;
-	const U32 lowestIndex = cctx->dictLimit;
-	const BYTE *const lowest = base + lowestIndex;
-	const BYTE *const iend = istart + srcSize;
-	const BYTE *const ilimit = iend - HASH_READ_SIZE;
-	U32 offset_1 = cctx->rep[0], offset_2 = cctx->rep[1];
-	U32 offsetSaved = 0;
-
-	/* init */
-	ip += (ip == lowest);
-	{
-		U32 const maxRep = (U32)(ip - lowest);
-		if (offset_2 > maxRep)
-			offsetSaved = offset_2, offset_2 = 0;
-		if (offset_1 > maxRep)
-			offsetSaved = offset_1, offset_1 = 0;
-	}
-
-	/* Main Search Loop */
-	while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */
-		size_t mLength;
-		size_t const h2 = ZSTD_hashPtr(ip, hBitsL, 8);
-		size_t const h = ZSTD_hashPtr(ip, hBitsS, mls);
-		U32 const curr = (U32)(ip - base);
-		U32 const matchIndexL = hashLong[h2];
-		U32 const matchIndexS = hashSmall[h];
-		const BYTE *matchLong = base + matchIndexL;
-		const BYTE *match = base + matchIndexS;
-		hashLong[h2] = hashSmall[h] = curr; /* update hash tables */
-
-		if ((offset_1 > 0) & (ZSTD_read32(ip + 1 - offset_1) == ZSTD_read32(ip + 1))) { /* note : by construction, offset_1 <= curr */
-			mLength = ZSTD_count(ip + 1 + 4, ip + 1 + 4 - offset_1, iend) + 4;
-			ip++;
-			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, 0, mLength - MINMATCH);
-		} else {
-			U32 offset;
-			if ((matchIndexL > lowestIndex) && (ZSTD_read64(matchLong) == ZSTD_read64(ip))) {
-				mLength = ZSTD_count(ip + 8, matchLong + 8, iend) + 8;
-				offset = (U32)(ip - matchLong);
-				while (((ip > anchor) & (matchLong > lowest)) && (ip[-1] == matchLong[-1])) {
-					ip--;
-					matchLong--;
-					mLength++;
-				} /* catch up */
-			} else if ((matchIndexS > lowestIndex) && (ZSTD_read32(match) == ZSTD_read32(ip))) {
-				size_t const h3 = ZSTD_hashPtr(ip + 1, hBitsL, 8);
-				U32 const matchIndex3 = hashLong[h3];
-				const BYTE *match3 = base + matchIndex3;
-				hashLong[h3] = curr + 1;
-				if ((matchIndex3 > lowestIndex) && (ZSTD_read64(match3) == ZSTD_read64(ip + 1))) {
-					mLength = ZSTD_count(ip + 9, match3 + 8, iend) + 8;
-					ip++;
-					offset = (U32)(ip - match3);
-					while (((ip > anchor) & (match3 > lowest)) && (ip[-1] == match3[-1])) {
-						ip--;
-						match3--;
-						mLength++;
-					} /* catch up */
-				} else {
-					mLength = ZSTD_count(ip + 4, match + 4, iend) + 4;
-					offset = (U32)(ip - match);
-					while (((ip > anchor) & (match > lowest)) && (ip[-1] == match[-1])) {
-						ip--;
-						match--;
-						mLength++;
-					} /* catch up */
-				}
-			} else {
-				ip += ((ip - anchor) >> g_searchStrength) + 1;
-				continue;
-			}
-
-			offset_2 = offset_1;
-			offset_1 = offset;
-
-			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, offset + ZSTD_REP_MOVE, mLength - MINMATCH);
-		}
-
-		/* match found */
-		ip += mLength;
-		anchor = ip;
-
-		if (ip <= ilimit) {
-			/* Fill Table */
-			hashLong[ZSTD_hashPtr(base + curr + 2, hBitsL, 8)] = hashSmall[ZSTD_hashPtr(base + curr + 2, hBitsS, mls)] =
-			    curr + 2; /* here because curr+2 could be > iend-8 */
-			hashLong[ZSTD_hashPtr(ip - 2, hBitsL, 8)] = hashSmall[ZSTD_hashPtr(ip - 2, hBitsS, mls)] = (U32)(ip - 2 - base);
-
-			/* check immediate repcode */
-			while ((ip <= ilimit) && ((offset_2 > 0) & (ZSTD_read32(ip) == ZSTD_read32(ip - offset_2)))) {
-				/* store sequence */
-				size_t const rLength = ZSTD_count(ip + 4, ip + 4 - offset_2, iend) + 4;
-				{
-					U32 const tmpOff = offset_2;
-					offset_2 = offset_1;
-					offset_1 = tmpOff;
-				} /* swap offset_2 <=> offset_1 */
-				hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip - base);
-				hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip - base);
-				ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength - MINMATCH);
-				ip += rLength;
-				anchor = ip;
-				continue; /* faster when present ... (?) */
-			}
-		}
-	}
-
-	/* save reps for next block */
-	cctx->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
-	cctx->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
-
-	/* Last Literals */
-	{
-		size_t const lastLLSize = iend - anchor;
-		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-		seqStorePtr->lit += lastLLSize;
-	}
-}
-
-static void ZSTD_compressBlock_doubleFast(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-	const U32 mls = ctx->params.cParams.searchLength;
-	switch (mls) {
-	default: /* includes case 3 */
-	case 4: ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 4); return;
-	case 5: ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 5); return;
-	case 6: ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 6); return;
-	case 7: ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 7); return;
-	}
-}
-
-static void ZSTD_compressBlock_doubleFast_extDict_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const U32 mls)
-{
-	U32 *const hashLong = ctx->hashTable;
-	U32 const hBitsL = ctx->params.cParams.hashLog;
-	U32 *const hashSmall = ctx->chainTable;
-	U32 const hBitsS = ctx->params.cParams.chainLog;
-	seqStore_t *seqStorePtr = &(ctx->seqStore);
-	const BYTE *const base = ctx->base;
-	const BYTE *const dictBase = ctx->dictBase;
-	const BYTE *const istart = (const BYTE *)src;
-	const BYTE *ip = istart;
-	const BYTE *anchor = istart;
-	const U32 lowestIndex = ctx->lowLimit;
-	const BYTE *const dictStart = dictBase + lowestIndex;
-	const U32 dictLimit = ctx->dictLimit;
-	const BYTE *const lowPrefixPtr = base + dictLimit;
-	const BYTE *const dictEnd = dictBase + dictLimit;
-	const BYTE *const iend = istart + srcSize;
-	const BYTE *const ilimit = iend - 8;
-	U32 offset_1 = ctx->rep[0], offset_2 = ctx->rep[1];
-
-	/* Search Loop */
-	while (ip < ilimit) { /* < instead of <=, because (ip+1) */
-		const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);
-		const U32 matchIndex = hashSmall[hSmall];
-		const BYTE *matchBase = matchIndex < dictLimit ? dictBase : base;
-		const BYTE *match = matchBase + matchIndex;
-
-		const size_t hLong = ZSTD_hashPtr(ip, hBitsL, 8);
-		const U32 matchLongIndex = hashLong[hLong];
-		const BYTE *matchLongBase = matchLongIndex < dictLimit ? dictBase : base;
-		const BYTE *matchLong = matchLongBase + matchLongIndex;
-
-		const U32 curr = (U32)(ip - base);
-		const U32 repIndex = curr + 1 - offset_1; /* offset_1 expected <= curr +1 */
-		const BYTE *repBase = repIndex < dictLimit ? dictBase : base;
-		const BYTE *repMatch = repBase + repIndex;
-		size_t mLength;
-		hashSmall[hSmall] = hashLong[hLong] = curr; /* update hash table */
-
-		if ((((U32)((dictLimit - 1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex)) &&
-		    (ZSTD_read32(repMatch) == ZSTD_read32(ip + 1))) {
-			const BYTE *repMatchEnd = repIndex < dictLimit ? dictEnd : iend;
-			mLength = ZSTD_count_2segments(ip + 1 + 4, repMatch + 4, iend, repMatchEnd, lowPrefixPtr) + 4;
-			ip++;
-			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, 0, mLength - MINMATCH);
-		} else {
-			if ((matchLongIndex > lowestIndex) && (ZSTD_read64(matchLong) == ZSTD_read64(ip))) {
-				const BYTE *matchEnd = matchLongIndex < dictLimit ? dictEnd : iend;
-				const BYTE *lowMatchPtr = matchLongIndex < dictLimit ? dictStart : lowPrefixPtr;
-				U32 offset;
-				mLength = ZSTD_count_2segments(ip + 8, matchLong + 8, iend, matchEnd, lowPrefixPtr) + 8;
-				offset = curr - matchLongIndex;
-				while (((ip > anchor) & (matchLong > lowMatchPtr)) && (ip[-1] == matchLong[-1])) {
-					ip--;
-					matchLong--;
-					mLength++;
-				} /* catch up */
-				offset_2 = offset_1;
-				offset_1 = offset;
-				ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, offset + ZSTD_REP_MOVE, mLength - MINMATCH);
-
-			} else if ((matchIndex > lowestIndex) && (ZSTD_read32(match) == ZSTD_read32(ip))) {
-				size_t const h3 = ZSTD_hashPtr(ip + 1, hBitsL, 8);
-				U32 const matchIndex3 = hashLong[h3];
-				const BYTE *const match3Base = matchIndex3 < dictLimit ? dictBase : base;
-				const BYTE *match3 = match3Base + matchIndex3;
-				U32 offset;
-				hashLong[h3] = curr + 1;
-				if ((matchIndex3 > lowestIndex) && (ZSTD_read64(match3) == ZSTD_read64(ip + 1))) {
-					const BYTE *matchEnd = matchIndex3 < dictLimit ? dictEnd : iend;
-					const BYTE *lowMatchPtr = matchIndex3 < dictLimit ? dictStart : lowPrefixPtr;
-					mLength = ZSTD_count_2segments(ip + 9, match3 + 8, iend, matchEnd, lowPrefixPtr) + 8;
-					ip++;
-					offset = curr + 1 - matchIndex3;
-					while (((ip > anchor) & (match3 > lowMatchPtr)) && (ip[-1] == match3[-1])) {
-						ip--;
-						match3--;
-						mLength++;
-					} /* catch up */
-				} else {
-					const BYTE *matchEnd = matchIndex < dictLimit ? dictEnd : iend;
-					const BYTE *lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;
-					mLength = ZSTD_count_2segments(ip + 4, match + 4, iend, matchEnd, lowPrefixPtr) + 4;
-					offset = curr - matchIndex;
-					while (((ip > anchor) & (match > lowMatchPtr)) && (ip[-1] == match[-1])) {
-						ip--;
-						match--;
-						mLength++;
-					} /* catch up */
-				}
-				offset_2 = offset_1;
-				offset_1 = offset;
-				ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, offset + ZSTD_REP_MOVE, mLength - MINMATCH);
-
-			} else {
-				ip += ((ip - anchor) >> g_searchStrength) + 1;
-				continue;
-			}
-		}
-
-		/* found a match : store it */
-		ip += mLength;
-		anchor = ip;
-
-		if (ip <= ilimit) {
-			/* Fill Table */
-			hashSmall[ZSTD_hashPtr(base + curr + 2, hBitsS, mls)] = curr + 2;
-			hashLong[ZSTD_hashPtr(base + curr + 2, hBitsL, 8)] = curr + 2;
-			hashSmall[ZSTD_hashPtr(ip - 2, hBitsS, mls)] = (U32)(ip - 2 - base);
-			hashLong[ZSTD_hashPtr(ip - 2, hBitsL, 8)] = (U32)(ip - 2 - base);
-			/* check immediate repcode */
-			while (ip <= ilimit) {
-				U32 const curr2 = (U32)(ip - base);
-				U32 const repIndex2 = curr2 - offset_2;
-				const BYTE *repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;
-				if ((((U32)((dictLimit - 1) - repIndex2) >= 3) & (repIndex2 > lowestIndex)) /* intentional overflow */
-				    && (ZSTD_read32(repMatch2) == ZSTD_read32(ip))) {
-					const BYTE *const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;
-					size_t const repLength2 =
-					    ZSTD_count_2segments(ip + EQUAL_READ32, repMatch2 + EQUAL_READ32, iend, repEnd2, lowPrefixPtr) + EQUAL_READ32;
-					U32 tmpOffset = offset_2;
-					offset_2 = offset_1;
-					offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
-					ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2 - MINMATCH);
-					hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = curr2;
-					hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = curr2;
-					ip += repLength2;
-					anchor = ip;
-					continue;
-				}
-				break;
-			}
-		}
-	}
-
-	/* save reps for next block */
-	ctx->repToConfirm[0] = offset_1;
-	ctx->repToConfirm[1] = offset_2;
-
-	/* Last Literals */
-	{
-		size_t const lastLLSize = iend - anchor;
-		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-		seqStorePtr->lit += lastLLSize;
-	}
-}
-
-static void ZSTD_compressBlock_doubleFast_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-	U32 const mls = ctx->params.cParams.searchLength;
-	switch (mls) {
-	default: /* includes case 3 */
-	case 4: ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 4); return;
-	case 5: ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 5); return;
-	case 6: ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 6); return;
-	case 7: ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 7); return;
-	}
-}
-
-/*-*************************************
-*  Binary Tree search
-***************************************/
-/** ZSTD_insertBt1() : add one or multiple positions to tree.
-*   ip : assumed <= iend-8 .
-*   @return : nb of positions added */
-static U32 ZSTD_insertBt1(ZSTD_CCtx *zc, const BYTE *const ip, const U32 mls, const BYTE *const iend, U32 nbCompares, U32 extDict)
-{
-	U32 *const hashTable = zc->hashTable;
-	U32 const hashLog = zc->params.cParams.hashLog;
-	size_t const h = ZSTD_hashPtr(ip, hashLog, mls);
-	U32 *const bt = zc->chainTable;
-	U32 const btLog = zc->params.cParams.chainLog - 1;
-	U32 const btMask = (1 << btLog) - 1;
-	U32 matchIndex = hashTable[h];
-	size_t commonLengthSmaller = 0, commonLengthLarger = 0;
-	const BYTE *const base = zc->base;
-	const BYTE *const dictBase = zc->dictBase;
-	const U32 dictLimit = zc->dictLimit;
-	const BYTE *const dictEnd = dictBase + dictLimit;
-	const BYTE *const prefixStart = base + dictLimit;
-	const BYTE *match;
-	const U32 curr = (U32)(ip - base);
-	const U32 btLow = btMask >= curr ? 0 : curr - btMask;
-	U32 *smallerPtr = bt + 2 * (curr & btMask);
-	U32 *largerPtr = smallerPtr + 1;
-	U32 dummy32; /* to be nullified at the end */
-	U32 const windowLow = zc->lowLimit;
-	U32 matchEndIdx = curr + 8;
-	size_t bestLength = 8;
-
-	hashTable[h] = curr; /* Update Hash Table */
-
-	while (nbCompares-- && (matchIndex > windowLow)) {
-		U32 *const nextPtr = bt + 2 * (matchIndex & btMask);
-		size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
-
-		if ((!extDict) || (matchIndex + matchLength >= dictLimit)) {
-			match = base + matchIndex;
-			if (match[matchLength] == ip[matchLength])
-				matchLength += ZSTD_count(ip + matchLength + 1, match + matchLength + 1, iend) + 1;
-		} else {
-			match = dictBase + matchIndex;
-			matchLength += ZSTD_count_2segments(ip + matchLength, match + matchLength, iend, dictEnd, prefixStart);
-			if (matchIndex + matchLength >= dictLimit)
-				match = base + matchIndex; /* to prepare for next usage of match[matchLength] */
-		}
-
-		if (matchLength > bestLength) {
-			bestLength = matchLength;
-			if (matchLength > matchEndIdx - matchIndex)
-				matchEndIdx = matchIndex + (U32)matchLength;
-		}
-
-		if (ip + matchLength == iend) /* equal : no way to know if inf or sup */
-			break;		      /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt the tree */
-
-		if (match[matchLength] < ip[matchLength]) { /* necessarily within correct buffer */
-			/* match is smaller than curr */
-			*smallerPtr = matchIndex;	  /* update smaller idx */
-			commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */
-			if (matchIndex <= btLow) {
-				smallerPtr = &dummy32;
-				break;
-			}			  /* beyond tree size, stop the search */
-			smallerPtr = nextPtr + 1; /* new "smaller" => larger of match */
-			matchIndex = nextPtr[1];  /* new matchIndex larger than previous (closer to curr) */
-		} else {
-			/* match is larger than curr */
-			*largerPtr = matchIndex;
-			commonLengthLarger = matchLength;
-			if (matchIndex <= btLow) {
-				largerPtr = &dummy32;
-				break;
-			} /* beyond tree size, stop the search */
-			largerPtr = nextPtr;
-			matchIndex = nextPtr[0];
-		}
-	}
-
-	*smallerPtr = *largerPtr = 0;
-	if (bestLength > 384)
-		return MIN(192, (U32)(bestLength - 384)); /* speed optimization */
-	if (matchEndIdx > curr + 8)
-		return matchEndIdx - curr - 8;
-	return 1;
-}
-
-static size_t ZSTD_insertBtAndFindBestMatch(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iend, size_t *offsetPtr, U32 nbCompares, const U32 mls,
-					    U32 extDict)
-{
-	U32 *const hashTable = zc->hashTable;
-	U32 const hashLog = zc->params.cParams.hashLog;
-	size_t const h = ZSTD_hashPtr(ip, hashLog, mls);
-	U32 *const bt = zc->chainTable;
-	U32 const btLog = zc->params.cParams.chainLog - 1;
-	U32 const btMask = (1 << btLog) - 1;
-	U32 matchIndex = hashTable[h];
-	size_t commonLengthSmaller = 0, commonLengthLarger = 0;
-	const BYTE *const base = zc->base;
-	const BYTE *const dictBase = zc->dictBase;
-	const U32 dictLimit = zc->dictLimit;
-	const BYTE *const dictEnd = dictBase + dictLimit;
-	const BYTE *const prefixStart = base + dictLimit;
-	const U32 curr = (U32)(ip - base);
-	const U32 btLow = btMask >= curr ? 0 : curr - btMask;
-	const U32 windowLow = zc->lowLimit;
-	U32 *smallerPtr = bt + 2 * (curr & btMask);
-	U32 *largerPtr = bt + 2 * (curr & btMask) + 1;
-	U32 matchEndIdx = curr + 8;
-	U32 dummy32; /* to be nullified at the end */
-	size_t bestLength = 0;
-
-	hashTable[h] = curr; /* Update Hash Table */
-
-	while (nbCompares-- && (matchIndex > windowLow)) {
-		U32 *const nextPtr = bt + 2 * (matchIndex & btMask);
-		size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
-		const BYTE *match;
-
-		if ((!extDict) || (matchIndex + matchLength >= dictLimit)) {
-			match = base + matchIndex;
-			if (match[matchLength] == ip[matchLength])
-				matchLength += ZSTD_count(ip + matchLength + 1, match + matchLength + 1, iend) + 1;
-		} else {
-			match = dictBase + matchIndex;
-			matchLength += ZSTD_count_2segments(ip + matchLength, match + matchLength, iend, dictEnd, prefixStart);
-			if (matchIndex + matchLength >= dictLimit)
-				match = base + matchIndex; /* to prepare for next usage of match[matchLength] */
-		}
-
-		if (matchLength > bestLength) {
-			if (matchLength > matchEndIdx - matchIndex)
-				matchEndIdx = matchIndex + (U32)matchLength;
-			if ((4 * (int)(matchLength - bestLength)) > (int)(ZSTD_highbit32(curr - matchIndex + 1) - ZSTD_highbit32((U32)offsetPtr[0] + 1)))
-				bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + curr - matchIndex;
-			if (ip + matchLength == iend) /* equal : no way to know if inf or sup */
-				break;		      /* drop, to guarantee consistency (miss a little bit of compression) */
-		}
-
-		if (match[matchLength] < ip[matchLength]) {
-			/* match is smaller than curr */
-			*smallerPtr = matchIndex;	  /* update smaller idx */
-			commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */
-			if (matchIndex <= btLow) {
-				smallerPtr = &dummy32;
-				break;
-			}			  /* beyond tree size, stop the search */
-			smallerPtr = nextPtr + 1; /* new "smaller" => larger of match */
-			matchIndex = nextPtr[1];  /* new matchIndex larger than previous (closer to curr) */
-		} else {
-			/* match is larger than curr */
-			*largerPtr = matchIndex;
-			commonLengthLarger = matchLength;
-			if (matchIndex <= btLow) {
-				largerPtr = &dummy32;
-				break;
-			} /* beyond tree size, stop the search */
-			largerPtr = nextPtr;
-			matchIndex = nextPtr[0];
-		}
-	}
-
-	*smallerPtr = *largerPtr = 0;
-
-	zc->nextToUpdate = (matchEndIdx > curr + 8) ? matchEndIdx - 8 : curr + 1;
-	return bestLength;
-}
-
-static void ZSTD_updateTree(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iend, const U32 nbCompares, const U32 mls)
-{
-	const BYTE *const base = zc->base;
-	const U32 target = (U32)(ip - base);
-	U32 idx = zc->nextToUpdate;
-
-	while (idx < target)
-		idx += ZSTD_insertBt1(zc, base + idx, mls, iend, nbCompares, 0);
-}
-
-/** ZSTD_BtFindBestMatch() : Tree updater, providing best match */
-static size_t ZSTD_BtFindBestMatch(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts, const U32 mls)
-{
-	if (ip < zc->base + zc->nextToUpdate)
-		return 0; /* skipped area */
-	ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls);
-	return ZSTD_insertBtAndFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, mls, 0);
-}
-
-static size_t ZSTD_BtFindBestMatch_selectMLS(ZSTD_CCtx *zc, /* Index table will be updated */
-					     const BYTE *ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts, const U32 matchLengthSearch)
-{
-	switch (matchLengthSearch) {
-	default: /* includes case 3 */
-	case 4: return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4);
-	case 5: return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5);
-	case 7:
-	case 6: return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6);
-	}
-}
-
-static void ZSTD_updateTree_extDict(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iend, const U32 nbCompares, const U32 mls)
-{
-	const BYTE *const base = zc->base;
-	const U32 target = (U32)(ip - base);
-	U32 idx = zc->nextToUpdate;
-
-	while (idx < target)
-		idx += ZSTD_insertBt1(zc, base + idx, mls, iend, nbCompares, 1);
-}
-
-/** Tree updater, providing best match */
-static size_t ZSTD_BtFindBestMatch_extDict(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts,
-					   const U32 mls)
-{
-	if (ip < zc->base + zc->nextToUpdate)
-		return 0; /* skipped area */
-	ZSTD_updateTree_extDict(zc, ip, iLimit, maxNbAttempts, mls);
-	return ZSTD_insertBtAndFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, mls, 1);
-}
-
-static size_t ZSTD_BtFindBestMatch_selectMLS_extDict(ZSTD_CCtx *zc, /* Index table will be updated */
-						     const BYTE *ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts,
-						     const U32 matchLengthSearch)
-{
-	switch (matchLengthSearch) {
-	default: /* includes case 3 */
-	case 4: return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4);
-	case 5: return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5);
-	case 7:
-	case 6: return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6);
-	}
-}
-
-/* *********************************
-*  Hash Chain
-***********************************/
-#define NEXT_IN_CHAIN(d, mask) chainTable[(d)&mask]
-
-/* Update chains up to ip (excluded)
-   Assumption : always within prefix (i.e. not within extDict) */
-FORCE_INLINE
-U32 ZSTD_insertAndFindFirstIndex(ZSTD_CCtx *zc, const BYTE *ip, U32 mls)
-{
-	U32 *const hashTable = zc->hashTable;
-	const U32 hashLog = zc->params.cParams.hashLog;
-	U32 *const chainTable = zc->chainTable;
-	const U32 chainMask = (1 << zc->params.cParams.chainLog) - 1;
-	const BYTE *const base = zc->base;
-	const U32 target = (U32)(ip - base);
-	U32 idx = zc->nextToUpdate;
-
-	while (idx < target) { /* catch up */
-		size_t const h = ZSTD_hashPtr(base + idx, hashLog, mls);
-		NEXT_IN_CHAIN(idx, chainMask) = hashTable[h];
-		hashTable[h] = idx;
-		idx++;
-	}
-
-	zc->nextToUpdate = target;
-	return hashTable[ZSTD_hashPtr(ip, hashLog, mls)];
-}
-
-/* inlining is important to hardwire a hot branch (template emulation) */
-FORCE_INLINE
-size_t ZSTD_HcFindBestMatch_generic(ZSTD_CCtx *zc, /* Index table will be updated */
-				    const BYTE *const ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts, const U32 mls,
-				    const U32 extDict)
-{
-	U32 *const chainTable = zc->chainTable;
-	const U32 chainSize = (1 << zc->params.cParams.chainLog);
-	const U32 chainMask = chainSize - 1;
-	const BYTE *const base = zc->base;
-	const BYTE *const dictBase = zc->dictBase;
-	const U32 dictLimit = zc->dictLimit;
-	const BYTE *const prefixStart = base + dictLimit;
-	const BYTE *const dictEnd = dictBase + dictLimit;
-	const U32 lowLimit = zc->lowLimit;
-	const U32 curr = (U32)(ip - base);
-	const U32 minChain = curr > chainSize ? curr - chainSize : 0;
-	int nbAttempts = maxNbAttempts;
-	size_t ml = EQUAL_READ32 - 1;
-
-	/* HC4 match finder */
-	U32 matchIndex = ZSTD_insertAndFindFirstIndex(zc, ip, mls);
-
-	for (; (matchIndex > lowLimit) & (nbAttempts > 0); nbAttempts--) {
-		const BYTE *match;
-		size_t currMl = 0;
-		if ((!extDict) || matchIndex >= dictLimit) {
-			match = base + matchIndex;
-			if (match[ml] == ip[ml]) /* potentially better */
-				currMl = ZSTD_count(ip, match, iLimit);
-		} else {
-			match = dictBase + matchIndex;
-			if (ZSTD_read32(match) == ZSTD_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */
-				currMl = ZSTD_count_2segments(ip + EQUAL_READ32, match + EQUAL_READ32, iLimit, dictEnd, prefixStart) + EQUAL_READ32;
-		}
-
-		/* save best solution */
-		if (currMl > ml) {
-			ml = currMl;
-			*offsetPtr = curr - matchIndex + ZSTD_REP_MOVE;
-			if (ip + currMl == iLimit)
-				break; /* best possible, and avoid read overflow*/
-		}
-
-		if (matchIndex <= minChain)
-			break;
-		matchIndex = NEXT_IN_CHAIN(matchIndex, chainMask);
-	}
-
-	return ml;
-}
-
-FORCE_INLINE size_t ZSTD_HcFindBestMatch_selectMLS(ZSTD_CCtx *zc, const BYTE *ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts,
-						   const U32 matchLengthSearch)
-{
-	switch (matchLengthSearch) {
-	default: /* includes case 3 */
-	case 4: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4, 0);
-	case 5: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5, 0);
-	case 7:
-	case 6: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6, 0);
-	}
-}
-
-FORCE_INLINE size_t ZSTD_HcFindBestMatch_extDict_selectMLS(ZSTD_CCtx *zc, const BYTE *ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts,
-							   const U32 matchLengthSearch)
-{
-	switch (matchLengthSearch) {
-	default: /* includes case 3 */
-	case 4: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4, 1);
-	case 5: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5, 1);
-	case 7:
-	case 6: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6, 1);
-	}
-}
-
-/* *******************************
-*  Common parser - lazy strategy
-*********************************/
-FORCE_INLINE
-void ZSTD_compressBlock_lazy_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const U32 searchMethod, const U32 depth)
-{
-	seqStore_t *seqStorePtr = &(ctx->seqStore);
-	const BYTE *const istart = (const BYTE *)src;
-	const BYTE *ip = istart;
-	const BYTE *anchor = istart;
-	const BYTE *const iend = istart + srcSize;
-	const BYTE *const ilimit = iend - 8;
-	const BYTE *const base = ctx->base + ctx->dictLimit;
-
-	U32 const maxSearches = 1 << ctx->params.cParams.searchLog;
-	U32 const mls = ctx->params.cParams.searchLength;
-
-	typedef size_t (*searchMax_f)(ZSTD_CCtx * zc, const BYTE *ip, const BYTE *iLimit, size_t *offsetPtr, U32 maxNbAttempts, U32 matchLengthSearch);
-	searchMax_f const searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS : ZSTD_HcFindBestMatch_selectMLS;
-	U32 offset_1 = ctx->rep[0], offset_2 = ctx->rep[1], savedOffset = 0;
-
-	/* init */
-	ip += (ip == base);
-	ctx->nextToUpdate3 = ctx->nextToUpdate;
-	{
-		U32 const maxRep = (U32)(ip - base);
-		if (offset_2 > maxRep)
-			savedOffset = offset_2, offset_2 = 0;
-		if (offset_1 > maxRep)
-			savedOffset = offset_1, offset_1 = 0;
-	}
-
-	/* Match Loop */
-	while (ip < ilimit) {
-		size_t matchLength = 0;
-		size_t offset = 0;
-		const BYTE *start = ip + 1;
-
-		/* check repCode */
-		if ((offset_1 > 0) & (ZSTD_read32(ip + 1) == ZSTD_read32(ip + 1 - offset_1))) {
-			/* repcode : we take it */
-			matchLength = ZSTD_count(ip + 1 + EQUAL_READ32, ip + 1 + EQUAL_READ32 - offset_1, iend) + EQUAL_READ32;
-			if (depth == 0)
-				goto _storeSequence;
-		}
-
-		/* first search (depth 0) */
-		{
-			size_t offsetFound = 99999999;
-			size_t const ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);
-			if (ml2 > matchLength)
-				matchLength = ml2, start = ip, offset = offsetFound;
-		}
-
-		if (matchLength < EQUAL_READ32) {
-			ip += ((ip - anchor) >> g_searchStrength) + 1; /* jump faster over incompressible sections */
-			continue;
-		}
-
-		/* let's try to find a better solution */
-		if (depth >= 1)
-			while (ip < ilimit) {
-				ip++;
-				if ((offset) && ((offset_1 > 0) & (ZSTD_read32(ip) == ZSTD_read32(ip - offset_1)))) {
-					size_t const mlRep = ZSTD_count(ip + EQUAL_READ32, ip + EQUAL_READ32 - offset_1, iend) + EQUAL_READ32;
-					int const gain2 = (int)(mlRep * 3);
-					int const gain1 = (int)(matchLength * 3 - ZSTD_highbit32((U32)offset + 1) + 1);
-					if ((mlRep >= EQUAL_READ32) && (gain2 > gain1))
-						matchLength = mlRep, offset = 0, start = ip;
-				}
-				{
-					size_t offset2 = 99999999;
-					size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-					int const gain2 = (int)(ml2 * 4 - ZSTD_highbit32((U32)offset2 + 1)); /* raw approx */
-					int const gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 4);
-					if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {
-						matchLength = ml2, offset = offset2, start = ip;
-						continue; /* search a better one */
-					}
-				}
-
-				/* let's find an even better one */
-				if ((depth == 2) && (ip < ilimit)) {
-					ip++;
-					if ((offset) && ((offset_1 > 0) & (ZSTD_read32(ip) == ZSTD_read32(ip - offset_1)))) {
-						size_t const ml2 = ZSTD_count(ip + EQUAL_READ32, ip + EQUAL_READ32 - offset_1, iend) + EQUAL_READ32;
-						int const gain2 = (int)(ml2 * 4);
-						int const gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 1);
-						if ((ml2 >= EQUAL_READ32) && (gain2 > gain1))
-							matchLength = ml2, offset = 0, start = ip;
-					}
-					{
-						size_t offset2 = 99999999;
-						size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-						int const gain2 = (int)(ml2 * 4 - ZSTD_highbit32((U32)offset2 + 1)); /* raw approx */
-						int const gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 7);
-						if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {
-							matchLength = ml2, offset = offset2, start = ip;
-							continue;
-						}
-					}
-				}
-				break; /* nothing found : store previous solution */
-			}
-
-		/* NOTE:
-		 * start[-offset+ZSTD_REP_MOVE-1] is undefined behavior.
-		 * (-offset+ZSTD_REP_MOVE-1) is unsigned, and is added to start, which
-		 * overflows the pointer, which is undefined behavior.
-		 */
-		/* catch up */
-		if (offset) {
-			while ((start > anchor) && (start > base + offset - ZSTD_REP_MOVE) &&
-			       (start[-1] == (start-offset+ZSTD_REP_MOVE)[-1])) /* only search for offset within prefix */
-			{
-				start--;
-				matchLength++;
-			}
-			offset_2 = offset_1;
-			offset_1 = (U32)(offset - ZSTD_REP_MOVE);
-		}
-
-	/* store sequence */
-_storeSequence:
-		{
-			size_t const litLength = start - anchor;
-			ZSTD_storeSeq(seqStorePtr, litLength, anchor, (U32)offset, matchLength - MINMATCH);
-			anchor = ip = start + matchLength;
-		}
-
-		/* check immediate repcode */
-		while ((ip <= ilimit) && ((offset_2 > 0) & (ZSTD_read32(ip) == ZSTD_read32(ip - offset_2)))) {
-			/* store sequence */
-			matchLength = ZSTD_count(ip + EQUAL_READ32, ip + EQUAL_READ32 - offset_2, iend) + EQUAL_READ32;
-			offset = offset_2;
-			offset_2 = offset_1;
-			offset_1 = (U32)offset; /* swap repcodes */
-			ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength - MINMATCH);
-			ip += matchLength;
-			anchor = ip;
-			continue; /* faster when present ... (?) */
-		}
-	}
-
-	/* Save reps for next block */
-	ctx->repToConfirm[0] = offset_1 ? offset_1 : savedOffset;
-	ctx->repToConfirm[1] = offset_2 ? offset_2 : savedOffset;
-
-	/* Last Literals */
-	{
-		size_t const lastLLSize = iend - anchor;
-		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-		seqStorePtr->lit += lastLLSize;
-	}
-}
-
-static void ZSTD_compressBlock_btlazy2(ZSTD_CCtx *ctx, const void *src, size_t srcSize) { ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 1, 2); }
-
-static void ZSTD_compressBlock_lazy2(ZSTD_CCtx *ctx, const void *src, size_t srcSize) { ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 2); }
-
-static void ZSTD_compressBlock_lazy(ZSTD_CCtx *ctx, const void *src, size_t srcSize) { ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 1); }
-
-static void ZSTD_compressBlock_greedy(ZSTD_CCtx *ctx, const void *src, size_t srcSize) { ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 0); }
-
-FORCE_INLINE
-void ZSTD_compressBlock_lazy_extDict_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const U32 searchMethod, const U32 depth)
-{
-	seqStore_t *seqStorePtr = &(ctx->seqStore);
-	const BYTE *const istart = (const BYTE *)src;
-	const BYTE *ip = istart;
-	const BYTE *anchor = istart;
-	const BYTE *const iend = istart + srcSize;
-	const BYTE *const ilimit = iend - 8;
-	const BYTE *const base = ctx->base;
-	const U32 dictLimit = ctx->dictLimit;
-	const U32 lowestIndex = ctx->lowLimit;
-	const BYTE *const prefixStart = base + dictLimit;
-	const BYTE *const dictBase = ctx->dictBase;
-	const BYTE *const dictEnd = dictBase + dictLimit;
-	const BYTE *const dictStart = dictBase + ctx->lowLimit;
-
-	const U32 maxSearches = 1 << ctx->params.cParams.searchLog;
-	const U32 mls = ctx->params.cParams.searchLength;
-
-	typedef size_t (*searchMax_f)(ZSTD_CCtx * zc, const BYTE *ip, const BYTE *iLimit, size_t *offsetPtr, U32 maxNbAttempts, U32 matchLengthSearch);
-	searchMax_f searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS_extDict : ZSTD_HcFindBestMatch_extDict_selectMLS;
-
-	U32 offset_1 = ctx->rep[0], offset_2 = ctx->rep[1];
-
-	/* init */
-	ctx->nextToUpdate3 = ctx->nextToUpdate;
-	ip += (ip == prefixStart);
-
-	/* Match Loop */
-	while (ip < ilimit) {
-		size_t matchLength = 0;
-		size_t offset = 0;
-		const BYTE *start = ip + 1;
-		U32 curr = (U32)(ip - base);
-
-		/* check repCode */
-		{
-			const U32 repIndex = (U32)(curr + 1 - offset_1);
-			const BYTE *const repBase = repIndex < dictLimit ? dictBase : base;
-			const BYTE *const repMatch = repBase + repIndex;
-			if (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */
-				if (ZSTD_read32(ip + 1) == ZSTD_read32(repMatch)) {
-					/* repcode detected we should take it */
-					const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend;
-					matchLength =
-					    ZSTD_count_2segments(ip + 1 + EQUAL_READ32, repMatch + EQUAL_READ32, iend, repEnd, prefixStart) + EQUAL_READ32;
-					if (depth == 0)
-						goto _storeSequence;
-				}
-		}
-
-		/* first search (depth 0) */
-		{
-			size_t offsetFound = 99999999;
-			size_t const ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);
-			if (ml2 > matchLength)
-				matchLength = ml2, start = ip, offset = offsetFound;
-		}
-
-		if (matchLength < EQUAL_READ32) {
-			ip += ((ip - anchor) >> g_searchStrength) + 1; /* jump faster over incompressible sections */
-			continue;
-		}
-
-		/* let's try to find a better solution */
-		if (depth >= 1)
-			while (ip < ilimit) {
-				ip++;
-				curr++;
-				/* check repCode */
-				if (offset) {
-					const U32 repIndex = (U32)(curr - offset_1);
-					const BYTE *const repBase = repIndex < dictLimit ? dictBase : base;
-					const BYTE *const repMatch = repBase + repIndex;
-					if (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */
-						if (ZSTD_read32(ip) == ZSTD_read32(repMatch)) {
-							/* repcode detected */
-							const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend;
-							size_t const repLength =
-							    ZSTD_count_2segments(ip + EQUAL_READ32, repMatch + EQUAL_READ32, iend, repEnd, prefixStart) +
-							    EQUAL_READ32;
-							int const gain2 = (int)(repLength * 3);
-							int const gain1 = (int)(matchLength * 3 - ZSTD_highbit32((U32)offset + 1) + 1);
-							if ((repLength >= EQUAL_READ32) && (gain2 > gain1))
-								matchLength = repLength, offset = 0, start = ip;
-						}
-				}
-
-				/* search match, depth 1 */
-				{
-					size_t offset2 = 99999999;
-					size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-					int const gain2 = (int)(ml2 * 4 - ZSTD_highbit32((U32)offset2 + 1)); /* raw approx */
-					int const gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 4);
-					if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {
-						matchLength = ml2, offset = offset2, start = ip;
-						continue; /* search a better one */
-					}
-				}
-
-				/* let's find an even better one */
-				if ((depth == 2) && (ip < ilimit)) {
-					ip++;
-					curr++;
-					/* check repCode */
-					if (offset) {
-						const U32 repIndex = (U32)(curr - offset_1);
-						const BYTE *const repBase = repIndex < dictLimit ? dictBase : base;
-						const BYTE *const repMatch = repBase + repIndex;
-						if (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */
-							if (ZSTD_read32(ip) == ZSTD_read32(repMatch)) {
-								/* repcode detected */
-								const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend;
-								size_t repLength = ZSTD_count_2segments(ip + EQUAL_READ32, repMatch + EQUAL_READ32, iend,
-													repEnd, prefixStart) +
-										   EQUAL_READ32;
-								int gain2 = (int)(repLength * 4);
-								int gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 1);
-								if ((repLength >= EQUAL_READ32) && (gain2 > gain1))
-									matchLength = repLength, offset = 0, start = ip;
-							}
-					}
-
-					/* search match, depth 2 */
-					{
-						size_t offset2 = 99999999;
-						size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-						int const gain2 = (int)(ml2 * 4 - ZSTD_highbit32((U32)offset2 + 1)); /* raw approx */
-						int const gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 7);
-						if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {
-							matchLength = ml2, offset = offset2, start = ip;
-							continue;
-						}
-					}
-				}
-				break; /* nothing found : store previous solution */
-			}
-
-		/* catch up */
-		if (offset) {
-			U32 const matchIndex = (U32)((start - base) - (offset - ZSTD_REP_MOVE));
-			const BYTE *match = (matchIndex < dictLimit) ? dictBase + matchIndex : base + matchIndex;
-			const BYTE *const mStart = (matchIndex < dictLimit) ? dictStart : prefixStart;
-			while ((start > anchor) && (match > mStart) && (start[-1] == match[-1])) {
-				start--;
-				match--;
-				matchLength++;
-			} /* catch up */
-			offset_2 = offset_1;
-			offset_1 = (U32)(offset - ZSTD_REP_MOVE);
-		}
-
-	/* store sequence */
-	_storeSequence : {
-		size_t const litLength = start - anchor;
-		ZSTD_storeSeq(seqStorePtr, litLength, anchor, (U32)offset, matchLength - MINMATCH);
-		anchor = ip = start + matchLength;
-	}
-
-		/* check immediate repcode */
-		while (ip <= ilimit) {
-			const U32 repIndex = (U32)((ip - base) - offset_2);
-			const BYTE *const repBase = repIndex < dictLimit ? dictBase : base;
-			const BYTE *const repMatch = repBase + repIndex;
-			if (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */
-				if (ZSTD_read32(ip) == ZSTD_read32(repMatch)) {
-					/* repcode detected we should take it */
-					const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend;
-					matchLength =
-					    ZSTD_count_2segments(ip + EQUAL_READ32, repMatch + EQUAL_READ32, iend, repEnd, prefixStart) + EQUAL_READ32;
-					offset = offset_2;
-					offset_2 = offset_1;
-					offset_1 = (U32)offset; /* swap offset history */
-					ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength - MINMATCH);
-					ip += matchLength;
-					anchor = ip;
-					continue; /* faster when present ... (?) */
-				}
-			break;
-		}
-	}
-
-	/* Save reps for next block */
-	ctx->repToConfirm[0] = offset_1;
-	ctx->repToConfirm[1] = offset_2;
-
-	/* Last Literals */
-	{
-		size_t const lastLLSize = iend - anchor;
-		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-		seqStorePtr->lit += lastLLSize;
-	}
-}
-
-void ZSTD_compressBlock_greedy_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize) { ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 0); }
-
-static void ZSTD_compressBlock_lazy_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-	ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 1);
-}
-
-static void ZSTD_compressBlock_lazy2_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-	ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 2);
-}
-
-static void ZSTD_compressBlock_btlazy2_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-	ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 1, 2);
-}
-
-/* The optimal parser */
-#include "zstd_opt.h"
-
-static void ZSTD_compressBlock_btopt(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-#ifdef ZSTD_OPT_H_91842398743
-	ZSTD_compressBlock_opt_generic(ctx, src, srcSize, 0);
-#else
-	(void)ctx;
-	(void)src;
-	(void)srcSize;
-	return;
-#endif
-}
-
-static void ZSTD_compressBlock_btopt2(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-#ifdef ZSTD_OPT_H_91842398743
-	ZSTD_compressBlock_opt_generic(ctx, src, srcSize, 1);
-#else
-	(void)ctx;
-	(void)src;
-	(void)srcSize;
-	return;
-#endif
-}
-
-static void ZSTD_compressBlock_btopt_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-#ifdef ZSTD_OPT_H_91842398743
-	ZSTD_compressBlock_opt_extDict_generic(ctx, src, srcSize, 0);
-#else
-	(void)ctx;
-	(void)src;
-	(void)srcSize;
-	return;
-#endif
-}
-
-static void ZSTD_compressBlock_btopt2_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-{
-#ifdef ZSTD_OPT_H_91842398743
-	ZSTD_compressBlock_opt_extDict_generic(ctx, src, srcSize, 1);
-#else
-	(void)ctx;
-	(void)src;
-	(void)srcSize;
-	return;
-#endif
-}
-
-typedef void (*ZSTD_blockCompressor)(ZSTD_CCtx *ctx, const void *src, size_t srcSize);
-
-static ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, int extDict)
-{
-	static const ZSTD_blockCompressor blockCompressor[2][8] = {
-	    {ZSTD_compressBlock_fast, ZSTD_compressBlock_doubleFast, ZSTD_compressBlock_greedy, ZSTD_compressBlock_lazy, ZSTD_compressBlock_lazy2,
-	     ZSTD_compressBlock_btlazy2, ZSTD_compressBlock_btopt, ZSTD_compressBlock_btopt2},
-	    {ZSTD_compressBlock_fast_extDict, ZSTD_compressBlock_doubleFast_extDict, ZSTD_compressBlock_greedy_extDict, ZSTD_compressBlock_lazy_extDict,
-	     ZSTD_compressBlock_lazy2_extDict, ZSTD_compressBlock_btlazy2_extDict, ZSTD_compressBlock_btopt_extDict, ZSTD_compressBlock_btopt2_extDict}};
-
-	return blockCompressor[extDict][(U32)strat];
-}
-
-static size_t ZSTD_compressBlock_internal(ZSTD_CCtx *zc, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->params.cParams.strategy, zc->lowLimit < zc->dictLimit);
-	const BYTE *const base = zc->base;
-	const BYTE *const istart = (const BYTE *)src;
-	const U32 curr = (U32)(istart - base);
-	if (srcSize < MIN_CBLOCK_SIZE + ZSTD_blockHeaderSize + 1)
-		return 0; /* don't even attempt compression below a certain srcSize */
-	ZSTD_resetSeqStore(&(zc->seqStore));
-	if (curr > zc->nextToUpdate + 384)
-		zc->nextToUpdate = curr - MIN(192, (U32)(curr - zc->nextToUpdate - 384)); /* update tree not updated after finding very long rep matches */
-	blockCompressor(zc, src, srcSize);
-	return ZSTD_compressSequences(zc, dst, dstCapacity, srcSize);
-}
-
-/*! ZSTD_compress_generic() :
-*   Compress a chunk of data into one or multiple blocks.
-*   All blocks will be terminated, all input will be consumed.
-*   Function will issue an error if there is not enough `dstCapacity` to hold the compressed content.
-*   Frame is supposed already started (header already produced)
-*   @return : compressed size, or an error code
-*/
-static size_t ZSTD_compress_generic(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, U32 lastFrameChunk)
-{
-	size_t blockSize = cctx->blockSize;
-	size_t remaining = srcSize;
-	const BYTE *ip = (const BYTE *)src;
-	BYTE *const ostart = (BYTE *)dst;
-	BYTE *op = ostart;
-	U32 const maxDist = 1 << cctx->params.cParams.windowLog;
-
-	if (cctx->params.fParams.checksumFlag && srcSize)
-		xxh64_update(&cctx->xxhState, src, srcSize);
-
-	while (remaining) {
-		U32 const lastBlock = lastFrameChunk & (blockSize >= remaining);
-		size_t cSize;
-
-		if (dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE)
-			return ERROR(dstSize_tooSmall); /* not enough space to store compressed block */
-		if (remaining < blockSize)
-			blockSize = remaining;
-
-		/* preemptive overflow correction */
-		if (cctx->lowLimit > (3U << 29)) {
-			U32 const cycleMask = (1 << ZSTD_cycleLog(cctx->params.cParams.hashLog, cctx->params.cParams.strategy)) - 1;
-			U32 const curr = (U32)(ip - cctx->base);
-			U32 const newCurr = (curr & cycleMask) + (1 << cctx->params.cParams.windowLog);
-			U32 const correction = curr - newCurr;
-			ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_64 <= 30);
-			ZSTD_reduceIndex(cctx, correction);
-			cctx->base += correction;
-			cctx->dictBase += correction;
-			cctx->lowLimit -= correction;
-			cctx->dictLimit -= correction;
-			if (cctx->nextToUpdate < correction)
-				cctx->nextToUpdate = 0;
-			else
-				cctx->nextToUpdate -= correction;
-		}
-
-		if ((U32)(ip + blockSize - cctx->base) > cctx->loadedDictEnd + maxDist) {
-			/* enforce maxDist */
-			U32 const newLowLimit = (U32)(ip + blockSize - cctx->base) - maxDist;
-			if (cctx->lowLimit < newLowLimit)
-				cctx->lowLimit = newLowLimit;
-			if (cctx->dictLimit < cctx->lowLimit)
-				cctx->dictLimit = cctx->lowLimit;
-		}
-
-		cSize = ZSTD_compressBlock_internal(cctx, op + ZSTD_blockHeaderSize, dstCapacity - ZSTD_blockHeaderSize, ip, blockSize);
-		if (ZSTD_isError(cSize))
-			return cSize;
-
-		if (cSize == 0) { /* block is not compressible */
-			U32 const cBlockHeader24 = lastBlock + (((U32)bt_raw) << 1) + (U32)(blockSize << 3);
-			if (blockSize + ZSTD_blockHeaderSize > dstCapacity)
-				return ERROR(dstSize_tooSmall);
-			ZSTD_writeLE32(op, cBlockHeader24); /* no pb, 4th byte will be overwritten */
-			memcpy(op + ZSTD_blockHeaderSize, ip, blockSize);
-			cSize = ZSTD_blockHeaderSize + blockSize;
-		} else {
-			U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed) << 1) + (U32)(cSize << 3);
-			ZSTD_writeLE24(op, cBlockHeader24);
-			cSize += ZSTD_blockHeaderSize;
-		}
-
-		remaining -= blockSize;
-		dstCapacity -= cSize;
-		ip += blockSize;
-		op += cSize;
-	}
-
-	if (lastFrameChunk && (op > ostart))
-		cctx->stage = ZSTDcs_ending;
-	return op - ostart;
-}
-
-static size_t ZSTD_writeFrameHeader(void *dst, size_t dstCapacity, ZSTD_parameters params, U64 pledgedSrcSize, U32 dictID)
-{
-	BYTE *const op = (BYTE *)dst;
-	U32 const dictIDSizeCode = (dictID > 0) + (dictID >= 256) + (dictID >= 65536); /* 0-3 */
-	U32 const checksumFlag = params.fParams.checksumFlag > 0;
-	U32 const windowSize = 1U << params.cParams.windowLog;
-	U32 const singleSegment = params.fParams.contentSizeFlag && (windowSize >= pledgedSrcSize);
-	BYTE const windowLogByte = (BYTE)((params.cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3);
-	U32 const fcsCode =
-	    params.fParams.contentSizeFlag ? (pledgedSrcSize >= 256) + (pledgedSrcSize >= 65536 + 256) + (pledgedSrcSize >= 0xFFFFFFFFU) : 0; /* 0-3 */
-	BYTE const frameHeaderDecriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag << 2) + (singleSegment << 5) + (fcsCode << 6));
-	size_t pos;
-
-	if (dstCapacity < ZSTD_frameHeaderSize_max)
-		return ERROR(dstSize_tooSmall);
-
-	ZSTD_writeLE32(dst, ZSTD_MAGICNUMBER);
-	op[4] = frameHeaderDecriptionByte;
-	pos = 5;
-	if (!singleSegment)
-		op[pos++] = windowLogByte;
-	switch (dictIDSizeCode) {
-	default: /* impossible */
-	case 0: break;
-	case 1:
-		op[pos] = (BYTE)(dictID);
-		pos++;
-		break;
-	case 2:
-		ZSTD_writeLE16(op + pos, (U16)dictID);
-		pos += 2;
-		break;
-	case 3:
-		ZSTD_writeLE32(op + pos, dictID);
-		pos += 4;
-		break;
-	}
-	switch (fcsCode) {
-	default: /* impossible */
-	case 0:
-		if (singleSegment)
-			op[pos++] = (BYTE)(pledgedSrcSize);
-		break;
-	case 1:
-		ZSTD_writeLE16(op + pos, (U16)(pledgedSrcSize - 256));
-		pos += 2;
-		break;
-	case 2:
-		ZSTD_writeLE32(op + pos, (U32)(pledgedSrcSize));
-		pos += 4;
-		break;
-	case 3:
-		ZSTD_writeLE64(op + pos, (U64)(pledgedSrcSize));
-		pos += 8;
-		break;
-	}
-	return pos;
-}
-
-static size_t ZSTD_compressContinue_internal(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, U32 frame, U32 lastFrameChunk)
-{
-	const BYTE *const ip = (const BYTE *)src;
-	size_t fhSize = 0;
-
-	if (cctx->stage == ZSTDcs_created)
-		return ERROR(stage_wrong); /* missing init (ZSTD_compressBegin) */
-
-	if (frame && (cctx->stage == ZSTDcs_init)) {
-		fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->params, cctx->frameContentSize, cctx->dictID);
-		if (ZSTD_isError(fhSize))
-			return fhSize;
-		dstCapacity -= fhSize;
-		dst = (char *)dst + fhSize;
-		cctx->stage = ZSTDcs_ongoing;
-	}
-
-	/* Check if blocks follow each other */
-	if (src != cctx->nextSrc) {
-		/* not contiguous */
-		ptrdiff_t const delta = cctx->nextSrc - ip;
-		cctx->lowLimit = cctx->dictLimit;
-		cctx->dictLimit = (U32)(cctx->nextSrc - cctx->base);
-		cctx->dictBase = cctx->base;
-		cctx->base -= delta;
-		cctx->nextToUpdate = cctx->dictLimit;
-		if (cctx->dictLimit - cctx->lowLimit < HASH_READ_SIZE)
-			cctx->lowLimit = cctx->dictLimit; /* too small extDict */
-	}
-
-	/* if input and dictionary overlap : reduce dictionary (area presumed modified by input) */
-	if ((ip + srcSize > cctx->dictBase + cctx->lowLimit) & (ip < cctx->dictBase + cctx->dictLimit)) {
-		ptrdiff_t const highInputIdx = (ip + srcSize) - cctx->dictBase;
-		U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)cctx->dictLimit) ? cctx->dictLimit : (U32)highInputIdx;
-		cctx->lowLimit = lowLimitMax;
-	}
-
-	cctx->nextSrc = ip + srcSize;
-
-	if (srcSize) {
-		size_t const cSize = frame ? ZSTD_compress_generic(cctx, dst, dstCapacity, src, srcSize, lastFrameChunk)
-					   : ZSTD_compressBlock_internal(cctx, dst, dstCapacity, src, srcSize);
-		if (ZSTD_isError(cSize))
-			return cSize;
-		return cSize + fhSize;
-	} else
-		return fhSize;
-}
-
-size_t ZSTD_compressContinue(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1, 0);
-}
-
-size_t ZSTD_getBlockSizeMax(ZSTD_CCtx *cctx) { return MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, 1 << cctx->params.cParams.windowLog); }
-
-size_t ZSTD_compressBlock(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	size_t const blockSizeMax = ZSTD_getBlockSizeMax(cctx);
-	if (srcSize > blockSizeMax)
-		return ERROR(srcSize_wrong);
-	return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 0, 0);
-}
-
-/*! ZSTD_loadDictionaryContent() :
- *  @return : 0, or an error code
- */
-static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx *zc, const void *src, size_t srcSize)
-{
-	const BYTE *const ip = (const BYTE *)src;
-	const BYTE *const iend = ip + srcSize;
-
-	/* input becomes curr prefix */
-	zc->lowLimit = zc->dictLimit;
-	zc->dictLimit = (U32)(zc->nextSrc - zc->base);
-	zc->dictBase = zc->base;
-	zc->base += ip - zc->nextSrc;
-	zc->nextToUpdate = zc->dictLimit;
-	zc->loadedDictEnd = zc->forceWindow ? 0 : (U32)(iend - zc->base);
-
-	zc->nextSrc = iend;
-	if (srcSize <= HASH_READ_SIZE)
-		return 0;
-
-	switch (zc->params.cParams.strategy) {
-	case ZSTD_fast: ZSTD_fillHashTable(zc, iend, zc->params.cParams.searchLength); break;
-
-	case ZSTD_dfast: ZSTD_fillDoubleHashTable(zc, iend, zc->params.cParams.searchLength); break;
-
-	case ZSTD_greedy:
-	case ZSTD_lazy:
-	case ZSTD_lazy2:
-		if (srcSize >= HASH_READ_SIZE)
-			ZSTD_insertAndFindFirstIndex(zc, iend - HASH_READ_SIZE, zc->params.cParams.searchLength);
-		break;
-
-	case ZSTD_btlazy2:
-	case ZSTD_btopt:
-	case ZSTD_btopt2:
-		if (srcSize >= HASH_READ_SIZE)
-			ZSTD_updateTree(zc, iend - HASH_READ_SIZE, iend, 1 << zc->params.cParams.searchLog, zc->params.cParams.searchLength);
-		break;
-
-	default:
-		return ERROR(GENERIC); /* strategy doesn't exist; impossible */
-	}
-
-	zc->nextToUpdate = (U32)(iend - zc->base);
-	return 0;
-}
-
-/* Dictionaries that assign zero probability to symbols that show up causes problems
-   when FSE encoding.  Refuse dictionaries that assign zero probability to symbols
-   that we may encounter during compression.
-   NOTE: This behavior is not standard and could be improved in the future. */
-static size_t ZSTD_checkDictNCount(short *normalizedCounter, unsigned dictMaxSymbolValue, unsigned maxSymbolValue)
-{
-	U32 s;
-	if (dictMaxSymbolValue < maxSymbolValue)
-		return ERROR(dictionary_corrupted);
-	for (s = 0; s <= maxSymbolValue; ++s) {
-		if (normalizedCounter[s] == 0)
-			return ERROR(dictionary_corrupted);
-	}
-	return 0;
-}
-
-/* Dictionary format :
- * See :
- * https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#dictionary-format
- */
-/*! ZSTD_loadZstdDictionary() :
- * @return : 0, or an error code
- *  assumptions : magic number supposed already checked
- *                dictSize supposed > 8
- */
-static size_t ZSTD_loadZstdDictionary(ZSTD_CCtx *cctx, const void *dict, size_t dictSize)
-{
-	const BYTE *dictPtr = (const BYTE *)dict;
-	const BYTE *const dictEnd = dictPtr + dictSize;
-	short offcodeNCount[MaxOff + 1];
-	unsigned offcodeMaxValue = MaxOff;
-
-	dictPtr += 4; /* skip magic number */
-	cctx->dictID = cctx->params.fParams.noDictIDFlag ? 0 : ZSTD_readLE32(dictPtr);
-	dictPtr += 4;
-
-	{
-		size_t const hufHeaderSize = HUF_readCTable_wksp(cctx->hufTable, 255, dictPtr, dictEnd - dictPtr, cctx->tmpCounters, sizeof(cctx->tmpCounters));
-		if (HUF_isError(hufHeaderSize))
-			return ERROR(dictionary_corrupted);
-		dictPtr += hufHeaderSize;
-	}
-
-	{
-		unsigned offcodeLog;
-		size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd - dictPtr);
-		if (FSE_isError(offcodeHeaderSize))
-			return ERROR(dictionary_corrupted);
-		if (offcodeLog > OffFSELog)
-			return ERROR(dictionary_corrupted);
-		/* Defer checking offcodeMaxValue because we need to know the size of the dictionary content */
-		CHECK_E(FSE_buildCTable_wksp(cctx->offcodeCTable, offcodeNCount, offcodeMaxValue, offcodeLog, cctx->tmpCounters, sizeof(cctx->tmpCounters)),
-			dictionary_corrupted);
-		dictPtr += offcodeHeaderSize;
-	}
-
-	{
-		short matchlengthNCount[MaxML + 1];
-		unsigned matchlengthMaxValue = MaxML, matchlengthLog;
-		size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd - dictPtr);
-		if (FSE_isError(matchlengthHeaderSize))
-			return ERROR(dictionary_corrupted);
-		if (matchlengthLog > MLFSELog)
-			return ERROR(dictionary_corrupted);
-		/* Every match length code must have non-zero probability */
-		CHECK_F(ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML));
-		CHECK_E(
-		    FSE_buildCTable_wksp(cctx->matchlengthCTable, matchlengthNCount, matchlengthMaxValue, matchlengthLog, cctx->tmpCounters, sizeof(cctx->tmpCounters)),
-		    dictionary_corrupted);
-		dictPtr += matchlengthHeaderSize;
-	}
-
-	{
-		short litlengthNCount[MaxLL + 1];
-		unsigned litlengthMaxValue = MaxLL, litlengthLog;
-		size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd - dictPtr);
-		if (FSE_isError(litlengthHeaderSize))
-			return ERROR(dictionary_corrupted);
-		if (litlengthLog > LLFSELog)
-			return ERROR(dictionary_corrupted);
-		/* Every literal length code must have non-zero probability */
-		CHECK_F(ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL));
-		CHECK_E(FSE_buildCTable_wksp(cctx->litlengthCTable, litlengthNCount, litlengthMaxValue, litlengthLog, cctx->tmpCounters, sizeof(cctx->tmpCounters)),
-			dictionary_corrupted);
-		dictPtr += litlengthHeaderSize;
-	}
-
-	if (dictPtr + 12 > dictEnd)
-		return ERROR(dictionary_corrupted);
-	cctx->rep[0] = ZSTD_readLE32(dictPtr + 0);
-	cctx->rep[1] = ZSTD_readLE32(dictPtr + 4);
-	cctx->rep[2] = ZSTD_readLE32(dictPtr + 8);
-	dictPtr += 12;
-
-	{
-		size_t const dictContentSize = (size_t)(dictEnd - dictPtr);
-		U32 offcodeMax = MaxOff;
-		if (dictContentSize <= ((U32)-1) - 128 KB) {
-			U32 const maxOffset = (U32)dictContentSize + 128 KB; /* The maximum offset that must be supported */
-			offcodeMax = ZSTD_highbit32(maxOffset);		     /* Calculate minimum offset code required to represent maxOffset */
-		}
-		/* All offset values <= dictContentSize + 128 KB must be representable */
-		CHECK_F(ZSTD_checkDictNCount(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff)));
-		/* All repCodes must be <= dictContentSize and != 0*/
-		{
-			U32 u;
-			for (u = 0; u < 3; u++) {
-				if (cctx->rep[u] == 0)
-					return ERROR(dictionary_corrupted);
-				if (cctx->rep[u] > dictContentSize)
-					return ERROR(dictionary_corrupted);
-			}
-		}
-
-		cctx->flagStaticTables = 1;
-		cctx->flagStaticHufTable = HUF_repeat_valid;
-		return ZSTD_loadDictionaryContent(cctx, dictPtr, dictContentSize);
-	}
-}
-
-/** ZSTD_compress_insertDictionary() :
-*   @return : 0, or an error code */
-static size_t ZSTD_compress_insertDictionary(ZSTD_CCtx *cctx, const void *dict, size_t dictSize)
-{
-	if ((dict == NULL) || (dictSize <= 8))
-		return 0;
-
-	/* dict as pure content */
-	if ((ZSTD_readLE32(dict) != ZSTD_DICT_MAGIC) || (cctx->forceRawDict))
-		return ZSTD_loadDictionaryContent(cctx, dict, dictSize);
-
-	/* dict as zstd dictionary */
-	return ZSTD_loadZstdDictionary(cctx, dict, dictSize);
-}
-
-/*! ZSTD_compressBegin_internal() :
-*   @return : 0, or an error code */
-static size_t ZSTD_compressBegin_internal(ZSTD_CCtx *cctx, const void *dict, size_t dictSize, ZSTD_parameters params, U64 pledgedSrcSize)
-{
-	ZSTD_compResetPolicy_e const crp = dictSize ? ZSTDcrp_fullReset : ZSTDcrp_continue;
-	CHECK_F(ZSTD_resetCCtx_advanced(cctx, params, pledgedSrcSize, crp));
-	return ZSTD_compress_insertDictionary(cctx, dict, dictSize);
-}
-
-/*! ZSTD_compressBegin_advanced() :
-*   @return : 0, or an error code */
-size_t ZSTD_compressBegin_advanced(ZSTD_CCtx *cctx, const void *dict, size_t dictSize, ZSTD_parameters params, unsigned long long pledgedSrcSize)
-{
-	/* compression parameters verification and optimization */
-	CHECK_F(ZSTD_checkCParams(params.cParams));
-	return ZSTD_compressBegin_internal(cctx, dict, dictSize, params, pledgedSrcSize);
-}
-
-size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx *cctx, const void *dict, size_t dictSize, int compressionLevel)
-{
-	ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, dictSize);
-	return ZSTD_compressBegin_internal(cctx, dict, dictSize, params, 0);
-}
-
-size_t ZSTD_compressBegin(ZSTD_CCtx *cctx, int compressionLevel) { return ZSTD_compressBegin_usingDict(cctx, NULL, 0, compressionLevel); }
-
-/*! ZSTD_writeEpilogue() :
-*   Ends a frame.
-*   @return : nb of bytes written into dst (or an error code) */
-static size_t ZSTD_writeEpilogue(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity)
-{
-	BYTE *const ostart = (BYTE *)dst;
-	BYTE *op = ostart;
-	size_t fhSize = 0;
-
-	if (cctx->stage == ZSTDcs_created)
-		return ERROR(stage_wrong); /* init missing */
-
-	/* special case : empty frame */
-	if (cctx->stage == ZSTDcs_init) {
-		fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->params, 0, 0);
-		if (ZSTD_isError(fhSize))
-			return fhSize;
-		dstCapacity -= fhSize;
-		op += fhSize;
-		cctx->stage = ZSTDcs_ongoing;
-	}
-
-	if (cctx->stage != ZSTDcs_ending) {
-		/* write one last empty block, make it the "last" block */
-		U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw) << 1) + 0;
-		if (dstCapacity < 4)
-			return ERROR(dstSize_tooSmall);
-		ZSTD_writeLE32(op, cBlockHeader24);
-		op += ZSTD_blockHeaderSize;
-		dstCapacity -= ZSTD_blockHeaderSize;
-	}
-
-	if (cctx->params.fParams.checksumFlag) {
-		U32 const checksum = (U32)xxh64_digest(&cctx->xxhState);
-		if (dstCapacity < 4)
-			return ERROR(dstSize_tooSmall);
-		ZSTD_writeLE32(op, checksum);
-		op += 4;
-	}
-
-	cctx->stage = ZSTDcs_created; /* return to "created but no init" status */
-	return op - ostart;
-}
-
-size_t ZSTD_compressEnd(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	size_t endResult;
-	size_t const cSize = ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1, 1);
-	if (ZSTD_isError(cSize))
-		return cSize;
-	endResult = ZSTD_writeEpilogue(cctx, (char *)dst + cSize, dstCapacity - cSize);
-	if (ZSTD_isError(endResult))
-		return endResult;
-	return cSize + endResult;
-}
-
-static size_t ZSTD_compress_internal(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const void *dict, size_t dictSize,
-				     ZSTD_parameters params)
-{
-	CHECK_F(ZSTD_compressBegin_internal(cctx, dict, dictSize, params, srcSize));
-	return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
-}
-
-size_t ZSTD_compress_usingDict(ZSTD_CCtx *ctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const void *dict, size_t dictSize,
-			       ZSTD_parameters params)
-{
-	return ZSTD_compress_internal(ctx, dst, dstCapacity, src, srcSize, dict, dictSize, params);
-}
-
-size_t ZSTD_compressCCtx(ZSTD_CCtx *ctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, ZSTD_parameters params)
-{
-	return ZSTD_compress_internal(ctx, dst, dstCapacity, src, srcSize, NULL, 0, params);
-}
-
-/* =====  Dictionary API  ===== */
-
-struct ZSTD_CDict_s {
-	void *dictBuffer;
-	const void *dictContent;
-	size_t dictContentSize;
-	ZSTD_CCtx *refContext;
-}; /* typedef'd tp ZSTD_CDict within "zstd.h" */
-
-size_t ZSTD_CDictWorkspaceBound(ZSTD_compressionParameters cParams) { return ZSTD_CCtxWorkspaceBound(cParams) + ZSTD_ALIGN(sizeof(ZSTD_CDict)); }
-
-static ZSTD_CDict *ZSTD_createCDict_advanced(const void *dictBuffer, size_t dictSize, unsigned byReference, ZSTD_parameters params, ZSTD_customMem customMem)
-{
-	if (!customMem.customAlloc || !customMem.customFree)
-		return NULL;
-
-	{
-		ZSTD_CDict *const cdict = (ZSTD_CDict *)ZSTD_malloc(sizeof(ZSTD_CDict), customMem);
-		ZSTD_CCtx *const cctx = ZSTD_createCCtx_advanced(customMem);
-
-		if (!cdict || !cctx) {
-			ZSTD_free(cdict, customMem);
-			ZSTD_freeCCtx(cctx);
-			return NULL;
-		}
-
-		if ((byReference) || (!dictBuffer) || (!dictSize)) {
-			cdict->dictBuffer = NULL;
-			cdict->dictContent = dictBuffer;
-		} else {
-			void *const internalBuffer = ZSTD_malloc(dictSize, customMem);
-			if (!internalBuffer) {
-				ZSTD_free(cctx, customMem);
-				ZSTD_free(cdict, customMem);
-				return NULL;
-			}
-			memcpy(internalBuffer, dictBuffer, dictSize);
-			cdict->dictBuffer = internalBuffer;
-			cdict->dictContent = internalBuffer;
-		}
-
-		{
-			size_t const errorCode = ZSTD_compressBegin_advanced(cctx, cdict->dictContent, dictSize, params, 0);
-			if (ZSTD_isError(errorCode)) {
-				ZSTD_free(cdict->dictBuffer, customMem);
-				ZSTD_free(cdict, customMem);
-				ZSTD_freeCCtx(cctx);
-				return NULL;
-			}
-		}
-
-		cdict->refContext = cctx;
-		cdict->dictContentSize = dictSize;
-		return cdict;
-	}
-}
-
-ZSTD_CDict *ZSTD_initCDict(const void *dict, size_t dictSize, ZSTD_parameters params, void *workspace, size_t workspaceSize)
-{
-	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-	return ZSTD_createCDict_advanced(dict, dictSize, 1, params, stackMem);
-}
-
-size_t ZSTD_freeCDict(ZSTD_CDict *cdict)
-{
-	if (cdict == NULL)
-		return 0; /* support free on NULL */
-	{
-		ZSTD_customMem const cMem = cdict->refContext->customMem;
-		ZSTD_freeCCtx(cdict->refContext);
-		ZSTD_free(cdict->dictBuffer, cMem);
-		ZSTD_free(cdict, cMem);
-		return 0;
-	}
-}
-
-static ZSTD_parameters ZSTD_getParamsFromCDict(const ZSTD_CDict *cdict) { return ZSTD_getParamsFromCCtx(cdict->refContext); }
-
-size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx *cctx, const ZSTD_CDict *cdict, unsigned long long pledgedSrcSize)
-{
-	if (cdict->dictContentSize)
-		CHECK_F(ZSTD_copyCCtx(cctx, cdict->refContext, pledgedSrcSize))
-	else {
-		ZSTD_parameters params = cdict->refContext->params;
-		params.fParams.contentSizeFlag = (pledgedSrcSize > 0);
-		CHECK_F(ZSTD_compressBegin_advanced(cctx, NULL, 0, params, pledgedSrcSize));
-	}
-	return 0;
-}
-
-/*! ZSTD_compress_usingCDict() :
-*   Compression using a digested Dictionary.
-*   Faster startup than ZSTD_compress_usingDict(), recommended when same dictionary is used multiple times.
-*   Note that compression level is decided during dictionary creation */
-size_t ZSTD_compress_usingCDict(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const ZSTD_CDict *cdict)
-{
-	CHECK_F(ZSTD_compressBegin_usingCDict(cctx, cdict, srcSize));
-
-	if (cdict->refContext->params.fParams.contentSizeFlag == 1) {
-		cctx->params.fParams.contentSizeFlag = 1;
-		cctx->frameContentSize = srcSize;
-	} else {
-		cctx->params.fParams.contentSizeFlag = 0;
-	}
-
-	return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
-}
-
-/* ******************************************************************
-*  Streaming
-********************************************************************/
-
-typedef enum { zcss_init, zcss_load, zcss_flush, zcss_final } ZSTD_cStreamStage;
-
-struct ZSTD_CStream_s {
-	ZSTD_CCtx *cctx;
-	ZSTD_CDict *cdictLocal;
-	const ZSTD_CDict *cdict;
-	char *inBuff;
-	size_t inBuffSize;
-	size_t inToCompress;
-	size_t inBuffPos;
-	size_t inBuffTarget;
-	size_t blockSize;
-	char *outBuff;
-	size_t outBuffSize;
-	size_t outBuffContentSize;
-	size_t outBuffFlushedSize;
-	ZSTD_cStreamStage stage;
-	U32 checksum;
-	U32 frameEnded;
-	U64 pledgedSrcSize;
-	U64 inputProcessed;
-	ZSTD_parameters params;
-	ZSTD_customMem customMem;
-}; /* typedef'd to ZSTD_CStream within "zstd.h" */
-
-size_t ZSTD_CStreamWorkspaceBound(ZSTD_compressionParameters cParams)
-{
-	size_t const inBuffSize = (size_t)1 << cParams.windowLog;
-	size_t const blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, inBuffSize);
-	size_t const outBuffSize = ZSTD_compressBound(blockSize) + 1;
-
-	return ZSTD_CCtxWorkspaceBound(cParams) + ZSTD_ALIGN(sizeof(ZSTD_CStream)) + ZSTD_ALIGN(inBuffSize) + ZSTD_ALIGN(outBuffSize);
-}
-
-ZSTD_CStream *ZSTD_createCStream_advanced(ZSTD_customMem customMem)
-{
-	ZSTD_CStream *zcs;
-
-	if (!customMem.customAlloc || !customMem.customFree)
-		return NULL;
-
-	zcs = (ZSTD_CStream *)ZSTD_malloc(sizeof(ZSTD_CStream), customMem);
-	if (zcs == NULL)
-		return NULL;
-	memset(zcs, 0, sizeof(ZSTD_CStream));
-	memcpy(&zcs->customMem, &customMem, sizeof(ZSTD_customMem));
-	zcs->cctx = ZSTD_createCCtx_advanced(customMem);
-	if (zcs->cctx == NULL) {
-		ZSTD_freeCStream(zcs);
-		return NULL;
-	}
-	return zcs;
-}
-
-size_t ZSTD_freeCStream(ZSTD_CStream *zcs)
-{
-	if (zcs == NULL)
-		return 0; /* support free on NULL */
-	{
-		ZSTD_customMem const cMem = zcs->customMem;
-		ZSTD_freeCCtx(zcs->cctx);
-		zcs->cctx = NULL;
-		ZSTD_freeCDict(zcs->cdictLocal);
-		zcs->cdictLocal = NULL;
-		ZSTD_free(zcs->inBuff, cMem);
-		zcs->inBuff = NULL;
-		ZSTD_free(zcs->outBuff, cMem);
-		zcs->outBuff = NULL;
-		ZSTD_free(zcs, cMem);
-		return 0;
-	}
-}
-
-/*======   Initialization   ======*/
-
-size_t ZSTD_CStreamInSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX; }
-size_t ZSTD_CStreamOutSize(void) { return ZSTD_compressBound(ZSTD_BLOCKSIZE_ABSOLUTEMAX) + ZSTD_blockHeaderSize + 4 /* 32-bits hash */; }
-
-static size_t ZSTD_resetCStream_internal(ZSTD_CStream *zcs, unsigned long long pledgedSrcSize)
-{
-	if (zcs->inBuffSize == 0)
-		return ERROR(stage_wrong); /* zcs has not been init at least once => can't reset */
-
-	if (zcs->cdict)
-		CHECK_F(ZSTD_compressBegin_usingCDict(zcs->cctx, zcs->cdict, pledgedSrcSize))
-	else
-		CHECK_F(ZSTD_compressBegin_advanced(zcs->cctx, NULL, 0, zcs->params, pledgedSrcSize));
-
-	zcs->inToCompress = 0;
-	zcs->inBuffPos = 0;
-	zcs->inBuffTarget = zcs->blockSize;
-	zcs->outBuffContentSize = zcs->outBuffFlushedSize = 0;
-	zcs->stage = zcss_load;
-	zcs->frameEnded = 0;
-	zcs->pledgedSrcSize = pledgedSrcSize;
-	zcs->inputProcessed = 0;
-	return 0; /* ready to go */
-}
-
-size_t ZSTD_resetCStream(ZSTD_CStream *zcs, unsigned long long pledgedSrcSize)
-{
-
-	zcs->params.fParams.contentSizeFlag = (pledgedSrcSize > 0);
-
-	return ZSTD_resetCStream_internal(zcs, pledgedSrcSize);
-}
-
-static size_t ZSTD_initCStream_advanced(ZSTD_CStream *zcs, const void *dict, size_t dictSize, ZSTD_parameters params, unsigned long long pledgedSrcSize)
-{
-	/* allocate buffers */
-	{
-		size_t const neededInBuffSize = (size_t)1 << params.cParams.windowLog;
-		if (zcs->inBuffSize < neededInBuffSize) {
-			zcs->inBuffSize = neededInBuffSize;
-			ZSTD_free(zcs->inBuff, zcs->customMem);
-			zcs->inBuff = (char *)ZSTD_malloc(neededInBuffSize, zcs->customMem);
-			if (zcs->inBuff == NULL)
-				return ERROR(memory_allocation);
-		}
-		zcs->blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, neededInBuffSize);
-	}
-	if (zcs->outBuffSize < ZSTD_compressBound(zcs->blockSize) + 1) {
-		zcs->outBuffSize = ZSTD_compressBound(zcs->blockSize) + 1;
-		ZSTD_free(zcs->outBuff, zcs->customMem);
-		zcs->outBuff = (char *)ZSTD_malloc(zcs->outBuffSize, zcs->customMem);
-		if (zcs->outBuff == NULL)
-			return ERROR(memory_allocation);
-	}
-
-	if (dict && dictSize >= 8) {
-		ZSTD_freeCDict(zcs->cdictLocal);
-		zcs->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize, 0, params, zcs->customMem);
-		if (zcs->cdictLocal == NULL)
-			return ERROR(memory_allocation);
-		zcs->cdict = zcs->cdictLocal;
-	} else
-		zcs->cdict = NULL;
-
-	zcs->checksum = params.fParams.checksumFlag > 0;
-	zcs->params = params;
-
-	return ZSTD_resetCStream_internal(zcs, pledgedSrcSize);
-}
-
-ZSTD_CStream *ZSTD_initCStream(ZSTD_parameters params, unsigned long long pledgedSrcSize, void *workspace, size_t workspaceSize)
-{
-	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-	ZSTD_CStream *const zcs = ZSTD_createCStream_advanced(stackMem);
-	if (zcs) {
-		size_t const code = ZSTD_initCStream_advanced(zcs, NULL, 0, params, pledgedSrcSize);
-		if (ZSTD_isError(code)) {
-			return NULL;
-		}
-	}
-	return zcs;
-}
-
-ZSTD_CStream *ZSTD_initCStream_usingCDict(const ZSTD_CDict *cdict, unsigned long long pledgedSrcSize, void *workspace, size_t workspaceSize)
-{
-	ZSTD_parameters const params = ZSTD_getParamsFromCDict(cdict);
-	ZSTD_CStream *const zcs = ZSTD_initCStream(params, pledgedSrcSize, workspace, workspaceSize);
-	if (zcs) {
-		zcs->cdict = cdict;
-		if (ZSTD_isError(ZSTD_resetCStream_internal(zcs, pledgedSrcSize))) {
-			return NULL;
-		}
-	}
-	return zcs;
-}
-
-/*======   Compression   ======*/
-
-typedef enum { zsf_gather, zsf_flush, zsf_end } ZSTD_flush_e;
-
-ZSTD_STATIC size_t ZSTD_limitCopy(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	size_t const length = MIN(dstCapacity, srcSize);
-	memcpy(dst, src, length);
-	return length;
-}
-
-static size_t ZSTD_compressStream_generic(ZSTD_CStream *zcs, void *dst, size_t *dstCapacityPtr, const void *src, size_t *srcSizePtr, ZSTD_flush_e const flush)
-{
-	U32 someMoreWork = 1;
-	const char *const istart = (const char *)src;
-	const char *const iend = istart + *srcSizePtr;
-	const char *ip = istart;
-	char *const ostart = (char *)dst;
-	char *const oend = ostart + *dstCapacityPtr;
-	char *op = ostart;
-
-	while (someMoreWork) {
-		switch (zcs->stage) {
-		case zcss_init:
-			return ERROR(init_missing); /* call ZBUFF_compressInit() first ! */
-
-		case zcss_load:
-			/* complete inBuffer */
-			{
-				size_t const toLoad = zcs->inBuffTarget - zcs->inBuffPos;
-				size_t const loaded = ZSTD_limitCopy(zcs->inBuff + zcs->inBuffPos, toLoad, ip, iend - ip);
-				zcs->inBuffPos += loaded;
-				ip += loaded;
-				if ((zcs->inBuffPos == zcs->inToCompress) || (!flush && (toLoad != loaded))) {
-					someMoreWork = 0;
-					break; /* not enough input to get a full block : stop there, wait for more */
-				}
-			}
-			/* compress curr block (note : this stage cannot be stopped in the middle) */
-			{
-				void *cDst;
-				size_t cSize;
-				size_t const iSize = zcs->inBuffPos - zcs->inToCompress;
-				size_t oSize = oend - op;
-				if (oSize >= ZSTD_compressBound(iSize))
-					cDst = op; /* compress directly into output buffer (avoid flush stage) */
-				else
-					cDst = zcs->outBuff, oSize = zcs->outBuffSize;
-				cSize = (flush == zsf_end) ? ZSTD_compressEnd(zcs->cctx, cDst, oSize, zcs->inBuff + zcs->inToCompress, iSize)
-							   : ZSTD_compressContinue(zcs->cctx, cDst, oSize, zcs->inBuff + zcs->inToCompress, iSize);
-				if (ZSTD_isError(cSize))
-					return cSize;
-				if (flush == zsf_end)
-					zcs->frameEnded = 1;
-				/* prepare next block */
-				zcs->inBuffTarget = zcs->inBuffPos + zcs->blockSize;
-				if (zcs->inBuffTarget > zcs->inBuffSize)
-					zcs->inBuffPos = 0, zcs->inBuffTarget = zcs->blockSize; /* note : inBuffSize >= blockSize */
-				zcs->inToCompress = zcs->inBuffPos;
-				if (cDst == op) {
-					op += cSize;
-					break;
-				} /* no need to flush */
-				zcs->outBuffContentSize = cSize;
-				zcs->outBuffFlushedSize = 0;
-				zcs->stage = zcss_flush; /* pass-through to flush stage */
-			}
-			fallthrough;
-
-		case zcss_flush: {
-			size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;
-			size_t const flushed = ZSTD_limitCopy(op, oend - op, zcs->outBuff + zcs->outBuffFlushedSize, toFlush);
-			op += flushed;
-			zcs->outBuffFlushedSize += flushed;
-			if (toFlush != flushed) {
-				someMoreWork = 0;
-				break;
-			} /* dst too small to store flushed data : stop there */
-			zcs->outBuffContentSize = zcs->outBuffFlushedSize = 0;
-			zcs->stage = zcss_load;
-			break;
-		}
-
-		case zcss_final:
-			someMoreWork = 0; /* do nothing */
-			break;
-
-		default:
-			return ERROR(GENERIC); /* impossible */
-		}
-	}
-
-	*srcSizePtr = ip - istart;
-	*dstCapacityPtr = op - ostart;
-	zcs->inputProcessed += *srcSizePtr;
-	if (zcs->frameEnded)
-		return 0;
-	{
-		size_t hintInSize = zcs->inBuffTarget - zcs->inBuffPos;
-		if (hintInSize == 0)
-			hintInSize = zcs->blockSize;
-		return hintInSize;
-	}
-}
-
-size_t ZSTD_compressStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output, ZSTD_inBuffer *input)
-{
-	size_t sizeRead = input->size - input->pos;
-	size_t sizeWritten = output->size - output->pos;
-	size_t const result =
-	    ZSTD_compressStream_generic(zcs, (char *)(output->dst) + output->pos, &sizeWritten, (const char *)(input->src) + input->pos, &sizeRead, zsf_gather);
-	input->pos += sizeRead;
-	output->pos += sizeWritten;
-	return result;
-}
-
-/*======   Finalize   ======*/
-
-/*! ZSTD_flushStream() :
-*   @return : amount of data remaining to flush */
-size_t ZSTD_flushStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output)
-{
-	size_t srcSize = 0;
-	size_t sizeWritten = output->size - output->pos;
-	size_t const result = ZSTD_compressStream_generic(zcs, (char *)(output->dst) + output->pos, &sizeWritten, &srcSize,
-							  &srcSize, /* use a valid src address instead of NULL */
-							  zsf_flush);
-	output->pos += sizeWritten;
-	if (ZSTD_isError(result))
-		return result;
-	return zcs->outBuffContentSize - zcs->outBuffFlushedSize; /* remaining to flush */
-}
-
-size_t ZSTD_endStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output)
-{
-	BYTE *const ostart = (BYTE *)(output->dst) + output->pos;
-	BYTE *const oend = (BYTE *)(output->dst) + output->size;
-	BYTE *op = ostart;
-
-	if ((zcs->pledgedSrcSize) && (zcs->inputProcessed != zcs->pledgedSrcSize))
-		return ERROR(srcSize_wrong); /* pledgedSrcSize not respected */
-
-	if (zcs->stage != zcss_final) {
-		/* flush whatever remains */
-		size_t srcSize = 0;
-		size_t sizeWritten = output->size - output->pos;
-		size_t const notEnded =
-		    ZSTD_compressStream_generic(zcs, ostart, &sizeWritten, &srcSize, &srcSize, zsf_end); /* use a valid src address instead of NULL */
-		size_t const remainingToFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;
-		op += sizeWritten;
-		if (remainingToFlush) {
-			output->pos += sizeWritten;
-			return remainingToFlush + ZSTD_BLOCKHEADERSIZE /* final empty block */ + (zcs->checksum * 4);
-		}
-		/* create epilogue */
-		zcs->stage = zcss_final;
-		zcs->outBuffContentSize = !notEnded ? 0 : ZSTD_compressEnd(zcs->cctx, zcs->outBuff, zcs->outBuffSize, NULL,
-									   0); /* write epilogue, including final empty block, into outBuff */
-	}
-
-	/* flush epilogue */
-	{
-		size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;
-		size_t const flushed = ZSTD_limitCopy(op, oend - op, zcs->outBuff + zcs->outBuffFlushedSize, toFlush);
-		op += flushed;
-		zcs->outBuffFlushedSize += flushed;
-		output->pos += op - ostart;
-		if (toFlush == flushed)
-			zcs->stage = zcss_init; /* end reached */
-		return toFlush - flushed;
-	}
-}
-
-/*-=====  Pre-defined compression levels  =====-*/
-
-#define ZSTD_DEFAULT_CLEVEL 1
-#define ZSTD_MAX_CLEVEL 22
-int ZSTD_maxCLevel(void) { return ZSTD_MAX_CLEVEL; }
-
-static const ZSTD_compressionParameters ZSTD_defaultCParameters[4][ZSTD_MAX_CLEVEL + 1] = {
-    {
-	/* "default" */
-	/* W,  C,  H,  S,  L, TL, strat */
-	{18, 12, 12, 1, 7, 16, ZSTD_fast},    /* level  0 - never used */
-	{19, 13, 14, 1, 7, 16, ZSTD_fast},    /* level  1 */
-	{19, 15, 16, 1, 6, 16, ZSTD_fast},    /* level  2 */
-	{20, 16, 17, 1, 5, 16, ZSTD_dfast},   /* level  3.*/
-	{20, 18, 18, 1, 5, 16, ZSTD_dfast},   /* level  4.*/
-	{20, 15, 18, 3, 5, 16, ZSTD_greedy},  /* level  5 */
-	{21, 16, 19, 2, 5, 16, ZSTD_lazy},    /* level  6 */
-	{21, 17, 20, 3, 5, 16, ZSTD_lazy},    /* level  7 */
-	{21, 18, 20, 3, 5, 16, ZSTD_lazy2},   /* level  8 */
-	{21, 20, 20, 3, 5, 16, ZSTD_lazy2},   /* level  9 */
-	{21, 19, 21, 4, 5, 16, ZSTD_lazy2},   /* level 10 */
-	{22, 20, 22, 4, 5, 16, ZSTD_lazy2},   /* level 11 */
-	{22, 20, 22, 5, 5, 16, ZSTD_lazy2},   /* level 12 */
-	{22, 21, 22, 5, 5, 16, ZSTD_lazy2},   /* level 13 */
-	{22, 21, 22, 6, 5, 16, ZSTD_lazy2},   /* level 14 */
-	{22, 21, 21, 5, 5, 16, ZSTD_btlazy2}, /* level 15 */
-	{23, 22, 22, 5, 5, 16, ZSTD_btlazy2}, /* level 16 */
-	{23, 21, 22, 4, 5, 24, ZSTD_btopt},   /* level 17 */
-	{23, 23, 22, 6, 5, 32, ZSTD_btopt},   /* level 18 */
-	{23, 23, 22, 6, 3, 48, ZSTD_btopt},   /* level 19 */
-	{25, 25, 23, 7, 3, 64, ZSTD_btopt2},  /* level 20 */
-	{26, 26, 23, 7, 3, 256, ZSTD_btopt2}, /* level 21 */
-	{27, 27, 25, 9, 3, 512, ZSTD_btopt2}, /* level 22 */
-    },
-    {
-	/* for srcSize <= 256 KB */
-	/* W,  C,  H,  S,  L,  T, strat */
-	{0, 0, 0, 0, 0, 0, ZSTD_fast},	 /* level  0 - not used */
-	{18, 13, 14, 1, 6, 8, ZSTD_fast},      /* level  1 */
-	{18, 14, 13, 1, 5, 8, ZSTD_dfast},     /* level  2 */
-	{18, 16, 15, 1, 5, 8, ZSTD_dfast},     /* level  3 */
-	{18, 15, 17, 1, 5, 8, ZSTD_greedy},    /* level  4.*/
-	{18, 16, 17, 4, 5, 8, ZSTD_greedy},    /* level  5.*/
-	{18, 16, 17, 3, 5, 8, ZSTD_lazy},      /* level  6.*/
-	{18, 17, 17, 4, 4, 8, ZSTD_lazy},      /* level  7 */
-	{18, 17, 17, 4, 4, 8, ZSTD_lazy2},     /* level  8 */
-	{18, 17, 17, 5, 4, 8, ZSTD_lazy2},     /* level  9 */
-	{18, 17, 17, 6, 4, 8, ZSTD_lazy2},     /* level 10 */
-	{18, 18, 17, 6, 4, 8, ZSTD_lazy2},     /* level 11.*/
-	{18, 18, 17, 7, 4, 8, ZSTD_lazy2},     /* level 12.*/
-	{18, 19, 17, 6, 4, 8, ZSTD_btlazy2},   /* level 13 */
-	{18, 18, 18, 4, 4, 16, ZSTD_btopt},    /* level 14.*/
-	{18, 18, 18, 4, 3, 16, ZSTD_btopt},    /* level 15.*/
-	{18, 19, 18, 6, 3, 32, ZSTD_btopt},    /* level 16.*/
-	{18, 19, 18, 8, 3, 64, ZSTD_btopt},    /* level 17.*/
-	{18, 19, 18, 9, 3, 128, ZSTD_btopt},   /* level 18.*/
-	{18, 19, 18, 10, 3, 256, ZSTD_btopt},  /* level 19.*/
-	{18, 19, 18, 11, 3, 512, ZSTD_btopt2}, /* level 20.*/
-	{18, 19, 18, 12, 3, 512, ZSTD_btopt2}, /* level 21.*/
-	{18, 19, 18, 13, 3, 512, ZSTD_btopt2}, /* level 22.*/
-    },
-    {
-	/* for srcSize <= 128 KB */
-	/* W,  C,  H,  S,  L,  T, strat */
-	{17, 12, 12, 1, 7, 8, ZSTD_fast},      /* level  0 - not used */
-	{17, 12, 13, 1, 6, 8, ZSTD_fast},      /* level  1 */
-	{17, 13, 16, 1, 5, 8, ZSTD_fast},      /* level  2 */
-	{17, 16, 16, 2, 5, 8, ZSTD_dfast},     /* level  3 */
-	{17, 13, 15, 3, 4, 8, ZSTD_greedy},    /* level  4 */
-	{17, 15, 17, 4, 4, 8, ZSTD_greedy},    /* level  5 */
-	{17, 16, 17, 3, 4, 8, ZSTD_lazy},      /* level  6 */
-	{17, 15, 17, 4, 4, 8, ZSTD_lazy2},     /* level  7 */
-	{17, 17, 17, 4, 4, 8, ZSTD_lazy2},     /* level  8 */
-	{17, 17, 17, 5, 4, 8, ZSTD_lazy2},     /* level  9 */
-	{17, 17, 17, 6, 4, 8, ZSTD_lazy2},     /* level 10 */
-	{17, 17, 17, 7, 4, 8, ZSTD_lazy2},     /* level 11 */
-	{17, 17, 17, 8, 4, 8, ZSTD_lazy2},     /* level 12 */
-	{17, 18, 17, 6, 4, 8, ZSTD_btlazy2},   /* level 13.*/
-	{17, 17, 17, 7, 3, 8, ZSTD_btopt},     /* level 14.*/
-	{17, 17, 17, 7, 3, 16, ZSTD_btopt},    /* level 15.*/
-	{17, 18, 17, 7, 3, 32, ZSTD_btopt},    /* level 16.*/
-	{17, 18, 17, 7, 3, 64, ZSTD_btopt},    /* level 17.*/
-	{17, 18, 17, 7, 3, 256, ZSTD_btopt},   /* level 18.*/
-	{17, 18, 17, 8, 3, 256, ZSTD_btopt},   /* level 19.*/
-	{17, 18, 17, 9, 3, 256, ZSTD_btopt2},  /* level 20.*/
-	{17, 18, 17, 10, 3, 256, ZSTD_btopt2}, /* level 21.*/
-	{17, 18, 17, 11, 3, 512, ZSTD_btopt2}, /* level 22.*/
-    },
-    {
-	/* for srcSize <= 16 KB */
-	/* W,  C,  H,  S,  L,  T, strat */
-	{14, 12, 12, 1, 7, 6, ZSTD_fast},      /* level  0 - not used */
-	{14, 14, 14, 1, 6, 6, ZSTD_fast},      /* level  1 */
-	{14, 14, 14, 1, 4, 6, ZSTD_fast},      /* level  2 */
-	{14, 14, 14, 1, 4, 6, ZSTD_dfast},     /* level  3.*/
-	{14, 14, 14, 4, 4, 6, ZSTD_greedy},    /* level  4.*/
-	{14, 14, 14, 3, 4, 6, ZSTD_lazy},      /* level  5.*/
-	{14, 14, 14, 4, 4, 6, ZSTD_lazy2},     /* level  6 */
-	{14, 14, 14, 5, 4, 6, ZSTD_lazy2},     /* level  7 */
-	{14, 14, 14, 6, 4, 6, ZSTD_lazy2},     /* level  8.*/
-	{14, 15, 14, 6, 4, 6, ZSTD_btlazy2},   /* level  9.*/
-	{14, 15, 14, 3, 3, 6, ZSTD_btopt},     /* level 10.*/
-	{14, 15, 14, 6, 3, 8, ZSTD_btopt},     /* level 11.*/
-	{14, 15, 14, 6, 3, 16, ZSTD_btopt},    /* level 12.*/
-	{14, 15, 14, 6, 3, 24, ZSTD_btopt},    /* level 13.*/
-	{14, 15, 15, 6, 3, 48, ZSTD_btopt},    /* level 14.*/
-	{14, 15, 15, 6, 3, 64, ZSTD_btopt},    /* level 15.*/
-	{14, 15, 15, 6, 3, 96, ZSTD_btopt},    /* level 16.*/
-	{14, 15, 15, 6, 3, 128, ZSTD_btopt},   /* level 17.*/
-	{14, 15, 15, 6, 3, 256, ZSTD_btopt},   /* level 18.*/
-	{14, 15, 15, 7, 3, 256, ZSTD_btopt},   /* level 19.*/
-	{14, 15, 15, 8, 3, 256, ZSTD_btopt2},  /* level 20.*/
-	{14, 15, 15, 9, 3, 256, ZSTD_btopt2},  /* level 21.*/
-	{14, 15, 15, 10, 3, 256, ZSTD_btopt2}, /* level 22.*/
-    },
-};
-
-/*! ZSTD_getCParams() :
-*   @return ZSTD_compressionParameters structure for a selected compression level, `srcSize` and `dictSize`.
-*   Size values are optional, provide 0 if not known or unused */
-ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long srcSize, size_t dictSize)
-{
-	ZSTD_compressionParameters cp;
-	size_t const addedSize = srcSize ? 0 : 500;
-	U64 const rSize = srcSize + dictSize ? srcSize + dictSize + addedSize : (U64)-1;
-	U32 const tableID = (rSize <= 256 KB) + (rSize <= 128 KB) + (rSize <= 16 KB); /* intentional underflow for srcSizeHint == 0 */
-	if (compressionLevel <= 0)
-		compressionLevel = ZSTD_DEFAULT_CLEVEL; /* 0 == default; no negative compressionLevel yet */
-	if (compressionLevel > ZSTD_MAX_CLEVEL)
-		compressionLevel = ZSTD_MAX_CLEVEL;
-	cp = ZSTD_defaultCParameters[tableID][compressionLevel];
-	if (ZSTD_32bits()) { /* auto-correction, for 32-bits mode */
-		if (cp.windowLog > ZSTD_WINDOWLOG_MAX)
-			cp.windowLog = ZSTD_WINDOWLOG_MAX;
-		if (cp.chainLog > ZSTD_CHAINLOG_MAX)
-			cp.chainLog = ZSTD_CHAINLOG_MAX;
-		if (cp.hashLog > ZSTD_HASHLOG_MAX)
-			cp.hashLog = ZSTD_HASHLOG_MAX;
-	}
-	cp = ZSTD_adjustCParams(cp, srcSize, dictSize);
-	return cp;
-}
-
-/*! ZSTD_getParams() :
-*   same as ZSTD_getCParams(), but @return a `ZSTD_parameters` object (instead of `ZSTD_compressionParameters`).
-*   All fields of `ZSTD_frameParameters` are set to default (0) */
-ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSize, size_t dictSize)
-{
-	ZSTD_parameters params;
-	ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, srcSize, dictSize);
-	memset(&params, 0, sizeof(params));
-	params.cParams = cParams;
-	return params;
-}
-
-EXPORT_SYMBOL(ZSTD_maxCLevel);
-EXPORT_SYMBOL(ZSTD_compressBound);
-
-EXPORT_SYMBOL(ZSTD_CCtxWorkspaceBound);
-EXPORT_SYMBOL(ZSTD_initCCtx);
-EXPORT_SYMBOL(ZSTD_compressCCtx);
-EXPORT_SYMBOL(ZSTD_compress_usingDict);
-
-EXPORT_SYMBOL(ZSTD_CDictWorkspaceBound);
-EXPORT_SYMBOL(ZSTD_initCDict);
-EXPORT_SYMBOL(ZSTD_compress_usingCDict);
-
-EXPORT_SYMBOL(ZSTD_CStreamWorkspaceBound);
-EXPORT_SYMBOL(ZSTD_initCStream);
-EXPORT_SYMBOL(ZSTD_initCStream_usingCDict);
-EXPORT_SYMBOL(ZSTD_resetCStream);
-EXPORT_SYMBOL(ZSTD_compressStream);
-EXPORT_SYMBOL(ZSTD_flushStream);
-EXPORT_SYMBOL(ZSTD_endStream);
-EXPORT_SYMBOL(ZSTD_CStreamInSize);
-EXPORT_SYMBOL(ZSTD_CStreamOutSize);
-
-EXPORT_SYMBOL(ZSTD_getCParams);
-EXPORT_SYMBOL(ZSTD_getParams);
-EXPORT_SYMBOL(ZSTD_checkCParams);
-EXPORT_SYMBOL(ZSTD_adjustCParams);
-
-EXPORT_SYMBOL(ZSTD_compressBegin);
-EXPORT_SYMBOL(ZSTD_compressBegin_usingDict);
-EXPORT_SYMBOL(ZSTD_compressBegin_advanced);
-EXPORT_SYMBOL(ZSTD_copyCCtx);
-EXPORT_SYMBOL(ZSTD_compressBegin_usingCDict);
-EXPORT_SYMBOL(ZSTD_compressContinue);
-EXPORT_SYMBOL(ZSTD_compressEnd);
-
-EXPORT_SYMBOL(ZSTD_getBlockSizeMax);
-EXPORT_SYMBOL(ZSTD_compressBlock);
-
-MODULE_LICENSE("Dual BSD/GPL");
-MODULE_DESCRIPTION("Zstd Compressor");
diff --git a/lib/zstd/compress/fse_compress.c b/lib/zstd/compress/fse_compress.c
new file mode 100644
index 0000000..436985b
--- /dev/null
+++ b/lib/zstd/compress/fse_compress.c
@@ -0,0 +1,625 @@
+/* ******************************************************************
+ * FSE : Finite State Entropy encoder
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ *  You can contact the author at :
+ *  - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *  - Public forum : https://groups.google.com/forum/#!forum/lz4c
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+/* **************************************************************
+*  Includes
+****************************************************************/
+#include "../common/compiler.h"
+#include "../common/mem.h"        /* U32, U16, etc. */
+#include "../common/debug.h"      /* assert, DEBUGLOG */
+#include "hist.h"       /* HIST_count_wksp */
+#include "../common/bitstream.h"
+#define FSE_STATIC_LINKING_ONLY
+#include "../common/fse.h"
+#include "../common/error_private.h"
+#define ZSTD_DEPS_NEED_MALLOC
+#define ZSTD_DEPS_NEED_MATH64
+#include "../common/zstd_deps.h"  /* ZSTD_malloc, ZSTD_free, ZSTD_memcpy, ZSTD_memset */
+
+
+/* **************************************************************
+*  Error Management
+****************************************************************/
+#define FSE_isError ERR_isError
+
+
+/* **************************************************************
+*  Templates
+****************************************************************/
+/*
+  designed to be included
+  for type-specific functions (template emulation in C)
+  Objective is to write these functions only once, for improved maintenance
+*/
+
+/* safety checks */
+#ifndef FSE_FUNCTION_EXTENSION
+#  error "FSE_FUNCTION_EXTENSION must be defined"
+#endif
+#ifndef FSE_FUNCTION_TYPE
+#  error "FSE_FUNCTION_TYPE must be defined"
+#endif
+
+/* Function names */
+#define FSE_CAT(X,Y) X##Y
+#define FSE_FUNCTION_NAME(X,Y) FSE_CAT(X,Y)
+#define FSE_TYPE_NAME(X,Y) FSE_CAT(X,Y)
+
+
+/* Function templates */
+
+/* FSE_buildCTable_wksp() :
+ * Same as FSE_buildCTable(), but using an externally allocated scratch buffer (`workSpace`).
+ * wkspSize should be sized to handle worst case situation, which is `1<<max_tableLog * sizeof(FSE_FUNCTION_TYPE)`
+ * workSpace must also be properly aligned with FSE_FUNCTION_TYPE requirements
+ */
+size_t FSE_buildCTable_wksp(FSE_CTable* ct,
+                      const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog,
+                            void* workSpace, size_t wkspSize)
+{
+    U32 const tableSize = 1 << tableLog;
+    U32 const tableMask = tableSize - 1;
+    void* const ptr = ct;
+    U16* const tableU16 = ( (U16*) ptr) + 2;
+    void* const FSCT = ((U32*)ptr) + 1 /* header */ + (tableLog ? tableSize>>1 : 1) ;
+    FSE_symbolCompressionTransform* const symbolTT = (FSE_symbolCompressionTransform*) (FSCT);
+    U32 const step = FSE_TABLESTEP(tableSize);
+
+    U32* cumul = (U32*)workSpace;
+    FSE_FUNCTION_TYPE* tableSymbol = (FSE_FUNCTION_TYPE*)(cumul + (maxSymbolValue + 2));
+
+    U32 highThreshold = tableSize-1;
+
+    if ((size_t)workSpace & 3) return ERROR(GENERIC); /* Must be 4 byte aligned */
+    if (FSE_BUILD_CTABLE_WORKSPACE_SIZE(maxSymbolValue, tableLog) > wkspSize) return ERROR(tableLog_tooLarge);
+    /* CTable header */
+    tableU16[-2] = (U16) tableLog;
+    tableU16[-1] = (U16) maxSymbolValue;
+    assert(tableLog < 16);   /* required for threshold strategy to work */
+
+    /* For explanations on how to distribute symbol values over the table :
+     * http://fastcompression.blogspot.fr/2014/02/fse-distributing-symbol-values.html */
+
+     #ifdef __clang_analyzer__
+     ZSTD_memset(tableSymbol, 0, sizeof(*tableSymbol) * tableSize);   /* useless initialization, just to keep scan-build happy */
+     #endif
+
+    /* symbol start positions */
+    {   U32 u;
+        cumul[0] = 0;
+        for (u=1; u <= maxSymbolValue+1; u++) {
+            if (normalizedCounter[u-1]==-1) {  /* Low proba symbol */
+                cumul[u] = cumul[u-1] + 1;
+                tableSymbol[highThreshold--] = (FSE_FUNCTION_TYPE)(u-1);
+            } else {
+                cumul[u] = cumul[u-1] + normalizedCounter[u-1];
+        }   }
+        cumul[maxSymbolValue+1] = tableSize+1;
+    }
+
+    /* Spread symbols */
+    {   U32 position = 0;
+        U32 symbol;
+        for (symbol=0; symbol<=maxSymbolValue; symbol++) {
+            int nbOccurrences;
+            int const freq = normalizedCounter[symbol];
+            for (nbOccurrences=0; nbOccurrences<freq; nbOccurrences++) {
+                tableSymbol[position] = (FSE_FUNCTION_TYPE)symbol;
+                position = (position + step) & tableMask;
+                while (position > highThreshold)
+                    position = (position + step) & tableMask;   /* Low proba area */
+        }   }
+
+        assert(position==0);  /* Must have initialized all positions */
+    }
+
+    /* Build table */
+    {   U32 u; for (u=0; u<tableSize; u++) {
+        FSE_FUNCTION_TYPE s = tableSymbol[u];   /* note : static analyzer may not understand tableSymbol is properly initialized */
+        tableU16[cumul[s]++] = (U16) (tableSize+u);   /* TableU16 : sorted by symbol order; gives next state value */
+    }   }
+
+    /* Build Symbol Transformation Table */
+    {   unsigned total = 0;
+        unsigned s;
+        for (s=0; s<=maxSymbolValue; s++) {
+            switch (normalizedCounter[s])
+            {
+            case  0:
+                /* filling nonetheless, for compatibility with FSE_getMaxNbBits() */
+                symbolTT[s].deltaNbBits = ((tableLog+1) << 16) - (1<<tableLog);
+                break;
+
+            case -1:
+            case  1:
+                symbolTT[s].deltaNbBits = (tableLog << 16) - (1<<tableLog);
+                symbolTT[s].deltaFindState = total - 1;
+                total ++;
+                break;
+            default :
+                {
+                    U32 const maxBitsOut = tableLog - BIT_highbit32 (normalizedCounter[s]-1);
+                    U32 const minStatePlus = normalizedCounter[s] << maxBitsOut;
+                    symbolTT[s].deltaNbBits = (maxBitsOut << 16) - minStatePlus;
+                    symbolTT[s].deltaFindState = total - normalizedCounter[s];
+                    total +=  normalizedCounter[s];
+    }   }   }   }
+
+#if 0  /* debug : symbol costs */
+    DEBUGLOG(5, "\n --- table statistics : ");
+    {   U32 symbol;
+        for (symbol=0; symbol<=maxSymbolValue; symbol++) {
+            DEBUGLOG(5, "%3u: w=%3i,   maxBits=%u, fracBits=%.2f",
+                symbol, normalizedCounter[symbol],
+                FSE_getMaxNbBits(symbolTT, symbol),
+                (double)FSE_bitCost(symbolTT, tableLog, symbol, 8) / 256);
+        }
+    }
+#endif
+
+    return 0;
+}
+
+
+
+
+#ifndef FSE_COMMONDEFS_ONLY
+
+
+/*-**************************************************************
+*  FSE NCount encoding
+****************************************************************/
+size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog)
+{
+    size_t const maxHeaderSize = (((maxSymbolValue+1) * tableLog) >> 3) + 3;
+    return maxSymbolValue ? maxHeaderSize : FSE_NCOUNTBOUND;  /* maxSymbolValue==0 ? use default */
+}
+
+static size_t
+FSE_writeNCount_generic (void* header, size_t headerBufferSize,
+                   const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog,
+                         unsigned writeIsSafe)
+{
+    BYTE* const ostart = (BYTE*) header;
+    BYTE* out = ostart;
+    BYTE* const oend = ostart + headerBufferSize;
+    int nbBits;
+    const int tableSize = 1 << tableLog;
+    int remaining;
+    int threshold;
+    U32 bitStream = 0;
+    int bitCount = 0;
+    unsigned symbol = 0;
+    unsigned const alphabetSize = maxSymbolValue + 1;
+    int previousIs0 = 0;
+
+    /* Table Size */
+    bitStream += (tableLog-FSE_MIN_TABLELOG) << bitCount;
+    bitCount  += 4;
+
+    /* Init */
+    remaining = tableSize+1;   /* +1 for extra accuracy */
+    threshold = tableSize;
+    nbBits = tableLog+1;
+
+    while ((symbol < alphabetSize) && (remaining>1)) {  /* stops at 1 */
+        if (previousIs0) {
+            unsigned start = symbol;
+            while ((symbol < alphabetSize) && !normalizedCounter[symbol]) symbol++;
+            if (symbol == alphabetSize) break;   /* incorrect distribution */
+            while (symbol >= start+24) {
+                start+=24;
+                bitStream += 0xFFFFU << bitCount;
+                if ((!writeIsSafe) && (out > oend-2))
+                    return ERROR(dstSize_tooSmall);   /* Buffer overflow */
+                out[0] = (BYTE) bitStream;
+                out[1] = (BYTE)(bitStream>>8);
+                out+=2;
+                bitStream>>=16;
+            }
+            while (symbol >= start+3) {
+                start+=3;
+                bitStream += 3 << bitCount;
+                bitCount += 2;
+            }
+            bitStream += (symbol-start) << bitCount;
+            bitCount += 2;
+            if (bitCount>16) {
+                if ((!writeIsSafe) && (out > oend - 2))
+                    return ERROR(dstSize_tooSmall);   /* Buffer overflow */
+                out[0] = (BYTE)bitStream;
+                out[1] = (BYTE)(bitStream>>8);
+                out += 2;
+                bitStream >>= 16;
+                bitCount -= 16;
+        }   }
+        {   int count = normalizedCounter[symbol++];
+            int const max = (2*threshold-1) - remaining;
+            remaining -= count < 0 ? -count : count;
+            count++;   /* +1 for extra accuracy */
+            if (count>=threshold)
+                count += max;   /* [0..max[ [max..threshold[ (...) [threshold+max 2*threshold[ */
+            bitStream += count << bitCount;
+            bitCount  += nbBits;
+            bitCount  -= (count<max);
+            previousIs0  = (count==1);
+            if (remaining<1) return ERROR(GENERIC);
+            while (remaining<threshold) { nbBits--; threshold>>=1; }
+        }
+        if (bitCount>16) {
+            if ((!writeIsSafe) && (out > oend - 2))
+                return ERROR(dstSize_tooSmall);   /* Buffer overflow */
+            out[0] = (BYTE)bitStream;
+            out[1] = (BYTE)(bitStream>>8);
+            out += 2;
+            bitStream >>= 16;
+            bitCount -= 16;
+    }   }
+
+    if (remaining != 1)
+        return ERROR(GENERIC);  /* incorrect normalized distribution */
+    assert(symbol <= alphabetSize);
+
+    /* flush remaining bitStream */
+    if ((!writeIsSafe) && (out > oend - 2))
+        return ERROR(dstSize_tooSmall);   /* Buffer overflow */
+    out[0] = (BYTE)bitStream;
+    out[1] = (BYTE)(bitStream>>8);
+    out+= (bitCount+7) /8;
+
+    return (out-ostart);
+}
+
+
+size_t FSE_writeNCount (void* buffer, size_t bufferSize,
+                  const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog)
+{
+    if (tableLog > FSE_MAX_TABLELOG) return ERROR(tableLog_tooLarge);   /* Unsupported */
+    if (tableLog < FSE_MIN_TABLELOG) return ERROR(GENERIC);   /* Unsupported */
+
+    if (bufferSize < FSE_NCountWriteBound(maxSymbolValue, tableLog))
+        return FSE_writeNCount_generic(buffer, bufferSize, normalizedCounter, maxSymbolValue, tableLog, 0);
+
+    return FSE_writeNCount_generic(buffer, bufferSize, normalizedCounter, maxSymbolValue, tableLog, 1 /* write in buffer is safe */);
+}
+
+
+/*-**************************************************************
+*  FSE Compression Code
+****************************************************************/
+
+FSE_CTable* FSE_createCTable (unsigned maxSymbolValue, unsigned tableLog)
+{
+    size_t size;
+    if (tableLog > FSE_TABLELOG_ABSOLUTE_MAX) tableLog = FSE_TABLELOG_ABSOLUTE_MAX;
+    size = FSE_CTABLE_SIZE_U32 (tableLog, maxSymbolValue) * sizeof(U32);
+    return (FSE_CTable*)ZSTD_malloc(size);
+}
+
+void FSE_freeCTable (FSE_CTable* ct) { ZSTD_free(ct); }
+
+/* provides the minimum logSize to safely represent a distribution */
+static unsigned FSE_minTableLog(size_t srcSize, unsigned maxSymbolValue)
+{
+    U32 minBitsSrc = BIT_highbit32((U32)(srcSize)) + 1;
+    U32 minBitsSymbols = BIT_highbit32(maxSymbolValue) + 2;
+    U32 minBits = minBitsSrc < minBitsSymbols ? minBitsSrc : minBitsSymbols;
+    assert(srcSize > 1); /* Not supported, RLE should be used instead */
+    return minBits;
+}
+
+unsigned FSE_optimalTableLog_internal(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, unsigned minus)
+{
+    U32 maxBitsSrc = BIT_highbit32((U32)(srcSize - 1)) - minus;
+    U32 tableLog = maxTableLog;
+    U32 minBits = FSE_minTableLog(srcSize, maxSymbolValue);
+    assert(srcSize > 1); /* Not supported, RLE should be used instead */
+    if (tableLog==0) tableLog = FSE_DEFAULT_TABLELOG;
+    if (maxBitsSrc < tableLog) tableLog = maxBitsSrc;   /* Accuracy can be reduced */
+    if (minBits > tableLog) tableLog = minBits;   /* Need a minimum to safely represent all symbol values */
+    if (tableLog < FSE_MIN_TABLELOG) tableLog = FSE_MIN_TABLELOG;
+    if (tableLog > FSE_MAX_TABLELOG) tableLog = FSE_MAX_TABLELOG;
+    return tableLog;
+}
+
+unsigned FSE_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue)
+{
+    return FSE_optimalTableLog_internal(maxTableLog, srcSize, maxSymbolValue, 2);
+}
+
+/* Secondary normalization method.
+   To be used when primary method fails. */
+
+static size_t FSE_normalizeM2(short* norm, U32 tableLog, const unsigned* count, size_t total, U32 maxSymbolValue, short lowProbCount)
+{
+    short const NOT_YET_ASSIGNED = -2;
+    U32 s;
+    U32 distributed = 0;
+    U32 ToDistribute;
+
+    /* Init */
+    U32 const lowThreshold = (U32)(total >> tableLog);
+    U32 lowOne = (U32)((total * 3) >> (tableLog + 1));
+
+    for (s=0; s<=maxSymbolValue; s++) {
+        if (count[s] == 0) {
+            norm[s]=0;
+            continue;
+        }
+        if (count[s] <= lowThreshold) {
+            norm[s] = lowProbCount;
+            distributed++;
+            total -= count[s];
+            continue;
+        }
+        if (count[s] <= lowOne) {
+            norm[s] = 1;
+            distributed++;
+            total -= count[s];
+            continue;
+        }
+
+        norm[s]=NOT_YET_ASSIGNED;
+    }
+    ToDistribute = (1 << tableLog) - distributed;
+
+    if (ToDistribute == 0)
+        return 0;
+
+    if ((total / ToDistribute) > lowOne) {
+        /* risk of rounding to zero */
+        lowOne = (U32)((total * 3) / (ToDistribute * 2));
+        for (s=0; s<=maxSymbolValue; s++) {
+            if ((norm[s] == NOT_YET_ASSIGNED) && (count[s] <= lowOne)) {
+                norm[s] = 1;
+                distributed++;
+                total -= count[s];
+                continue;
+        }   }
+        ToDistribute = (1 << tableLog) - distributed;
+    }
+
+    if (distributed == maxSymbolValue+1) {
+        /* all values are pretty poor;
+           probably incompressible data (should have already been detected);
+           find max, then give all remaining points to max */
+        U32 maxV = 0, maxC = 0;
+        for (s=0; s<=maxSymbolValue; s++)
+            if (count[s] > maxC) { maxV=s; maxC=count[s]; }
+        norm[maxV] += (short)ToDistribute;
+        return 0;
+    }
+
+    if (total == 0) {
+        /* all of the symbols were low enough for the lowOne or lowThreshold */
+        for (s=0; ToDistribute > 0; s = (s+1)%(maxSymbolValue+1))
+            if (norm[s] > 0) { ToDistribute--; norm[s]++; }
+        return 0;
+    }
+
+    {   U64 const vStepLog = 62 - tableLog;
+        U64 const mid = (1ULL << (vStepLog-1)) - 1;
+        U64 const rStep = ZSTD_div64((((U64)1<<vStepLog) * ToDistribute) + mid, (U32)total);   /* scale on remaining */
+        U64 tmpTotal = mid;
+        for (s=0; s<=maxSymbolValue; s++) {
+            if (norm[s]==NOT_YET_ASSIGNED) {
+                U64 const end = tmpTotal + (count[s] * rStep);
+                U32 const sStart = (U32)(tmpTotal >> vStepLog);
+                U32 const sEnd = (U32)(end >> vStepLog);
+                U32 const weight = sEnd - sStart;
+                if (weight < 1)
+                    return ERROR(GENERIC);
+                norm[s] = (short)weight;
+                tmpTotal = end;
+    }   }   }
+
+    return 0;
+}
+
+size_t FSE_normalizeCount (short* normalizedCounter, unsigned tableLog,
+                           const unsigned* count, size_t total,
+                           unsigned maxSymbolValue, unsigned useLowProbCount)
+{
+    /* Sanity checks */
+    if (tableLog==0) tableLog = FSE_DEFAULT_TABLELOG;
+    if (tableLog < FSE_MIN_TABLELOG) return ERROR(GENERIC);   /* Unsupported size */
+    if (tableLog > FSE_MAX_TABLELOG) return ERROR(tableLog_tooLarge);   /* Unsupported size */
+    if (tableLog < FSE_minTableLog(total, maxSymbolValue)) return ERROR(GENERIC);   /* Too small tableLog, compression potentially impossible */
+
+    {   static U32 const rtbTable[] = {     0, 473195, 504333, 520860, 550000, 700000, 750000, 830000 };
+        short const lowProbCount = useLowProbCount ? -1 : 1;
+        U64 const scale = 62 - tableLog;
+        U64 const step = ZSTD_div64((U64)1<<62, (U32)total);   /* <== here, one division ! */
+        U64 const vStep = 1ULL<<(scale-20);
+        int stillToDistribute = 1<<tableLog;
+        unsigned s;
+        unsigned largest=0;
+        short largestP=0;
+        U32 lowThreshold = (U32)(total >> tableLog);
+
+        for (s=0; s<=maxSymbolValue; s++) {
+            if (count[s] == total) return 0;   /* rle special case */
+            if (count[s] == 0) { normalizedCounter[s]=0; continue; }
+            if (count[s] <= lowThreshold) {
+                normalizedCounter[s] = lowProbCount;
+                stillToDistribute--;
+            } else {
+                short proba = (short)((count[s]*step) >> scale);
+                if (proba<8) {
+                    U64 restToBeat = vStep * rtbTable[proba];
+                    proba += (count[s]*step) - ((U64)proba<<scale) > restToBeat;
+                }
+                if (proba > largestP) { largestP=proba; largest=s; }
+                normalizedCounter[s] = proba;
+                stillToDistribute -= proba;
+        }   }
+        if (-stillToDistribute >= (normalizedCounter[largest] >> 1)) {
+            /* corner case, need another normalization method */
+            size_t const errorCode = FSE_normalizeM2(normalizedCounter, tableLog, count, total, maxSymbolValue, lowProbCount);
+            if (FSE_isError(errorCode)) return errorCode;
+        }
+        else normalizedCounter[largest] += (short)stillToDistribute;
+    }
+
+#if 0
+    {   /* Print Table (debug) */
+        U32 s;
+        U32 nTotal = 0;
+        for (s=0; s<=maxSymbolValue; s++)
+            RAWLOG(2, "%3i: %4i \n", s, normalizedCounter[s]);
+        for (s=0; s<=maxSymbolValue; s++)
+            nTotal += abs(normalizedCounter[s]);
+        if (nTotal != (1U<<tableLog))
+            RAWLOG(2, "Warning !!! Total == %u != %u !!!", nTotal, 1U<<tableLog);
+        getchar();
+    }
+#endif
+
+    return tableLog;
+}
+
+
+/* fake FSE_CTable, for raw (uncompressed) input */
+size_t FSE_buildCTable_raw (FSE_CTable* ct, unsigned nbBits)
+{
+    const unsigned tableSize = 1 << nbBits;
+    const unsigned tableMask = tableSize - 1;
+    const unsigned maxSymbolValue = tableMask;
+    void* const ptr = ct;
+    U16* const tableU16 = ( (U16*) ptr) + 2;
+    void* const FSCT = ((U32*)ptr) + 1 /* header */ + (tableSize>>1);   /* assumption : tableLog >= 1 */
+    FSE_symbolCompressionTransform* const symbolTT = (FSE_symbolCompressionTransform*) (FSCT);
+    unsigned s;
+
+    /* Sanity checks */
+    if (nbBits < 1) return ERROR(GENERIC);             /* min size */
+
+    /* header */
+    tableU16[-2] = (U16) nbBits;
+    tableU16[-1] = (U16) maxSymbolValue;
+
+    /* Build table */
+    for (s=0; s<tableSize; s++)
+        tableU16[s] = (U16)(tableSize + s);
+
+    /* Build Symbol Transformation Table */
+    {   const U32 deltaNbBits = (nbBits << 16) - (1 << nbBits);
+        for (s=0; s<=maxSymbolValue; s++) {
+            symbolTT[s].deltaNbBits = deltaNbBits;
+            symbolTT[s].deltaFindState = s-1;
+    }   }
+
+    return 0;
+}
+
+/* fake FSE_CTable, for rle input (always same symbol) */
+size_t FSE_buildCTable_rle (FSE_CTable* ct, BYTE symbolValue)
+{
+    void* ptr = ct;
+    U16* tableU16 = ( (U16*) ptr) + 2;
+    void* FSCTptr = (U32*)ptr + 2;
+    FSE_symbolCompressionTransform* symbolTT = (FSE_symbolCompressionTransform*) FSCTptr;
+
+    /* header */
+    tableU16[-2] = (U16) 0;
+    tableU16[-1] = (U16) symbolValue;
+
+    /* Build table */
+    tableU16[0] = 0;
+    tableU16[1] = 0;   /* just in case */
+
+    /* Build Symbol Transformation Table */
+    symbolTT[symbolValue].deltaNbBits = 0;
+    symbolTT[symbolValue].deltaFindState = 0;
+
+    return 0;
+}
+
+
+static size_t FSE_compress_usingCTable_generic (void* dst, size_t dstSize,
+                           const void* src, size_t srcSize,
+                           const FSE_CTable* ct, const unsigned fast)
+{
+    const BYTE* const istart = (const BYTE*) src;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* ip=iend;
+
+    BIT_CStream_t bitC;
+    FSE_CState_t CState1, CState2;
+
+    /* init */
+    if (srcSize <= 2) return 0;
+    { size_t const initError = BIT_initCStream(&bitC, dst, dstSize);
+      if (FSE_isError(initError)) return 0; /* not enough space available to write a bitstream */ }
+
+#define FSE_FLUSHBITS(s)  (fast ? BIT_flushBitsFast(s) : BIT_flushBits(s))
+
+    if (srcSize & 1) {
+        FSE_initCState2(&CState1, ct, *--ip);
+        FSE_initCState2(&CState2, ct, *--ip);
+        FSE_encodeSymbol(&bitC, &CState1, *--ip);
+        FSE_FLUSHBITS(&bitC);
+    } else {
+        FSE_initCState2(&CState2, ct, *--ip);
+        FSE_initCState2(&CState1, ct, *--ip);
+    }
+
+    /* join to mod 4 */
+    srcSize -= 2;
+    if ((sizeof(bitC.bitContainer)*8 > FSE_MAX_TABLELOG*4+7 ) && (srcSize & 2)) {  /* test bit 2 */
+        FSE_encodeSymbol(&bitC, &CState2, *--ip);
+        FSE_encodeSymbol(&bitC, &CState1, *--ip);
+        FSE_FLUSHBITS(&bitC);
+    }
+
+    /* 2 or 4 encoding per loop */
+    while ( ip>istart ) {
+
+        FSE_encodeSymbol(&bitC, &CState2, *--ip);
+
+        if (sizeof(bitC.bitContainer)*8 < FSE_MAX_TABLELOG*2+7 )   /* this test must be static */
+            FSE_FLUSHBITS(&bitC);
+
+        FSE_encodeSymbol(&bitC, &CState1, *--ip);
+
+        if (sizeof(bitC.bitContainer)*8 > FSE_MAX_TABLELOG*4+7 ) {  /* this test must be static */
+            FSE_encodeSymbol(&bitC, &CState2, *--ip);
+            FSE_encodeSymbol(&bitC, &CState1, *--ip);
+        }
+
+        FSE_FLUSHBITS(&bitC);
+    }
+
+    FSE_flushCState(&bitC, &CState2);
+    FSE_flushCState(&bitC, &CState1);
+    return BIT_closeCStream(&bitC);
+}
+
+size_t FSE_compress_usingCTable (void* dst, size_t dstSize,
+                           const void* src, size_t srcSize,
+                           const FSE_CTable* ct)
+{
+    unsigned const fast = (dstSize >= FSE_BLOCKBOUND(srcSize));
+
+    if (fast)
+        return FSE_compress_usingCTable_generic(dst, dstSize, src, srcSize, ct, 1);
+    else
+        return FSE_compress_usingCTable_generic(dst, dstSize, src, srcSize, ct, 0);
+}
+
+
+size_t FSE_compressBound(size_t size) { return FSE_COMPRESSBOUND(size); }
+
+
+#endif   /* FSE_COMMONDEFS_ONLY */
diff --git a/lib/zstd/compress/hist.c b/lib/zstd/compress/hist.c
new file mode 100644
index 0000000..3ddc6df
--- /dev/null
+++ b/lib/zstd/compress/hist.c
@@ -0,0 +1,165 @@
+/* ******************************************************************
+ * hist : Histogram functions
+ * part of Finite State Entropy project
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ *  You can contact the author at :
+ *  - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *  - Public forum : https://groups.google.com/forum/#!forum/lz4c
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+/* --- dependencies --- */
+#include "../common/mem.h"             /* U32, BYTE, etc. */
+#include "../common/debug.h"           /* assert, DEBUGLOG */
+#include "../common/error_private.h"   /* ERROR */
+#include "hist.h"
+
+
+/* --- Error management --- */
+unsigned HIST_isError(size_t code) { return ERR_isError(code); }
+
+/*-**************************************************************
+ *  Histogram functions
+ ****************************************************************/
+unsigned HIST_count_simple(unsigned* count, unsigned* maxSymbolValuePtr,
+                           const void* src, size_t srcSize)
+{
+    const BYTE* ip = (const BYTE*)src;
+    const BYTE* const end = ip + srcSize;
+    unsigned maxSymbolValue = *maxSymbolValuePtr;
+    unsigned largestCount=0;
+
+    ZSTD_memset(count, 0, (maxSymbolValue+1) * sizeof(*count));
+    if (srcSize==0) { *maxSymbolValuePtr = 0; return 0; }
+
+    while (ip<end) {
+        assert(*ip <= maxSymbolValue);
+        count[*ip++]++;
+    }
+
+    while (!count[maxSymbolValue]) maxSymbolValue--;
+    *maxSymbolValuePtr = maxSymbolValue;
+
+    {   U32 s;
+        for (s=0; s<=maxSymbolValue; s++)
+            if (count[s] > largestCount) largestCount = count[s];
+    }
+
+    return largestCount;
+}
+
+typedef enum { trustInput, checkMaxSymbolValue } HIST_checkInput_e;
+
+/* HIST_count_parallel_wksp() :
+ * store histogram into 4 intermediate tables, recombined at the end.
+ * this design makes better use of OoO cpus,
+ * and is noticeably faster when some values are heavily repeated.
+ * But it needs some additional workspace for intermediate tables.
+ * `workSpace` must be a U32 table of size >= HIST_WKSP_SIZE_U32.
+ * @return : largest histogram frequency,
+ *           or an error code (notably when histogram's alphabet is larger than *maxSymbolValuePtr) */
+static size_t HIST_count_parallel_wksp(
+                                unsigned* count, unsigned* maxSymbolValuePtr,
+                                const void* source, size_t sourceSize,
+                                HIST_checkInput_e check,
+                                U32* const workSpace)
+{
+    const BYTE* ip = (const BYTE*)source;
+    const BYTE* const iend = ip+sourceSize;
+    size_t const countSize = (*maxSymbolValuePtr + 1) * sizeof(*count);
+    unsigned max=0;
+    U32* const Counting1 = workSpace;
+    U32* const Counting2 = Counting1 + 256;
+    U32* const Counting3 = Counting2 + 256;
+    U32* const Counting4 = Counting3 + 256;
+
+    /* safety checks */
+    assert(*maxSymbolValuePtr <= 255);
+    if (!sourceSize) {
+        ZSTD_memset(count, 0, countSize);
+        *maxSymbolValuePtr = 0;
+        return 0;
+    }
+    ZSTD_memset(workSpace, 0, 4*256*sizeof(unsigned));
+
+    /* by stripes of 16 bytes */
+    {   U32 cached = MEM_read32(ip); ip += 4;
+        while (ip < iend-15) {
+            U32 c = cached; cached = MEM_read32(ip); ip += 4;
+            Counting1[(BYTE) c     ]++;
+            Counting2[(BYTE)(c>>8) ]++;
+            Counting3[(BYTE)(c>>16)]++;
+            Counting4[       c>>24 ]++;
+            c = cached; cached = MEM_read32(ip); ip += 4;
+            Counting1[(BYTE) c     ]++;
+            Counting2[(BYTE)(c>>8) ]++;
+            Counting3[(BYTE)(c>>16)]++;
+            Counting4[       c>>24 ]++;
+            c = cached; cached = MEM_read32(ip); ip += 4;
+            Counting1[(BYTE) c     ]++;
+            Counting2[(BYTE)(c>>8) ]++;
+            Counting3[(BYTE)(c>>16)]++;
+            Counting4[       c>>24 ]++;
+            c = cached; cached = MEM_read32(ip); ip += 4;
+            Counting1[(BYTE) c     ]++;
+            Counting2[(BYTE)(c>>8) ]++;
+            Counting3[(BYTE)(c>>16)]++;
+            Counting4[       c>>24 ]++;
+        }
+        ip-=4;
+    }
+
+    /* finish last symbols */
+    while (ip<iend) Counting1[*ip++]++;
+
+    {   U32 s;
+        for (s=0; s<256; s++) {
+            Counting1[s] += Counting2[s] + Counting3[s] + Counting4[s];
+            if (Counting1[s] > max) max = Counting1[s];
+    }   }
+
+    {   unsigned maxSymbolValue = 255;
+        while (!Counting1[maxSymbolValue]) maxSymbolValue--;
+        if (check && maxSymbolValue > *maxSymbolValuePtr) return ERROR(maxSymbolValue_tooSmall);
+        *maxSymbolValuePtr = maxSymbolValue;
+        ZSTD_memmove(count, Counting1, countSize);   /* in case count & Counting1 are overlapping */
+    }
+    return (size_t)max;
+}
+
+/* HIST_countFast_wksp() :
+ * Same as HIST_countFast(), but using an externally provided scratch buffer.
+ * `workSpace` is a writable buffer which must be 4-bytes aligned,
+ * `workSpaceSize` must be >= HIST_WKSP_SIZE
+ */
+size_t HIST_countFast_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
+                          const void* source, size_t sourceSize,
+                          void* workSpace, size_t workSpaceSize)
+{
+    if (sourceSize < 1500) /* heuristic threshold */
+        return HIST_count_simple(count, maxSymbolValuePtr, source, sourceSize);
+    if ((size_t)workSpace & 3) return ERROR(GENERIC);  /* must be aligned on 4-bytes boundaries */
+    if (workSpaceSize < HIST_WKSP_SIZE) return ERROR(workSpace_tooSmall);
+    return HIST_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, trustInput, (U32*)workSpace);
+}
+
+/* HIST_count_wksp() :
+ * Same as HIST_count(), but using an externally provided scratch buffer.
+ * `workSpace` size must be table of >= HIST_WKSP_SIZE_U32 unsigned */
+size_t HIST_count_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
+                       const void* source, size_t sourceSize,
+                       void* workSpace, size_t workSpaceSize)
+{
+    if ((size_t)workSpace & 3) return ERROR(GENERIC);  /* must be aligned on 4-bytes boundaries */
+    if (workSpaceSize < HIST_WKSP_SIZE) return ERROR(workSpace_tooSmall);
+    if (*maxSymbolValuePtr < 255)
+        return HIST_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, checkMaxSymbolValue, (U32*)workSpace);
+    *maxSymbolValuePtr = 255;
+    return HIST_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, workSpace, workSpaceSize);
+}
+
diff --git a/lib/zstd/compress/hist.h b/lib/zstd/compress/hist.h
new file mode 100644
index 0000000..fc1830a
--- /dev/null
+++ b/lib/zstd/compress/hist.h
@@ -0,0 +1,75 @@
+/* ******************************************************************
+ * hist : Histogram functions
+ * part of Finite State Entropy project
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ *  You can contact the author at :
+ *  - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *  - Public forum : https://groups.google.com/forum/#!forum/lz4c
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+/* --- dependencies --- */
+#include "../common/zstd_deps.h"   /* size_t */
+
+
+/* --- simple histogram functions --- */
+
+/*! HIST_count():
+ *  Provides the precise count of each byte within a table 'count'.
+ * 'count' is a table of unsigned int, of minimum size (*maxSymbolValuePtr+1).
+ *  Updates *maxSymbolValuePtr with actual largest symbol value detected.
+ * @return : count of the most frequent symbol (which isn't identified).
+ *           or an error code, which can be tested using HIST_isError().
+ *           note : if return == srcSize, there is only one symbol.
+ */
+size_t HIST_count(unsigned* count, unsigned* maxSymbolValuePtr,
+                  const void* src, size_t srcSize);
+
+unsigned HIST_isError(size_t code);  /*< tells if a return value is an error code */
+
+
+/* --- advanced histogram functions --- */
+
+#define HIST_WKSP_SIZE_U32 1024
+#define HIST_WKSP_SIZE    (HIST_WKSP_SIZE_U32 * sizeof(unsigned))
+/* HIST_count_wksp() :
+ *  Same as HIST_count(), but using an externally provided scratch buffer.
+ *  Benefit is this function will use very little stack space.
+ * `workSpace` is a writable buffer which must be 4-bytes aligned,
+ * `workSpaceSize` must be >= HIST_WKSP_SIZE
+ */
+size_t HIST_count_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
+                       const void* src, size_t srcSize,
+                       void* workSpace, size_t workSpaceSize);
+
+/* HIST_countFast() :
+ *  same as HIST_count(), but blindly trusts that all byte values within src are <= *maxSymbolValuePtr.
+ *  This function is unsafe, and will segfault if any value within `src` is `> *maxSymbolValuePtr`
+ */
+size_t HIST_countFast(unsigned* count, unsigned* maxSymbolValuePtr,
+                      const void* src, size_t srcSize);
+
+/* HIST_countFast_wksp() :
+ *  Same as HIST_countFast(), but using an externally provided scratch buffer.
+ * `workSpace` is a writable buffer which must be 4-bytes aligned,
+ * `workSpaceSize` must be >= HIST_WKSP_SIZE
+ */
+size_t HIST_countFast_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
+                           const void* src, size_t srcSize,
+                           void* workSpace, size_t workSpaceSize);
+
+/*! HIST_count_simple() :
+ *  Same as HIST_countFast(), this function is unsafe,
+ *  and will segfault if any value within `src` is `> *maxSymbolValuePtr`.
+ *  It is also a bit slower for large inputs.
+ *  However, it does not need any additional memory (not even on stack).
+ * @return : count of the most frequent symbol.
+ *  Note this function doesn't produce any error (i.e. it must succeed).
+ */
+unsigned HIST_count_simple(unsigned* count, unsigned* maxSymbolValuePtr,
+                           const void* src, size_t srcSize);
diff --git a/lib/zstd/compress/huf_compress.c b/lib/zstd/compress/huf_compress.c
new file mode 100644
index 0000000..f76a526
--- /dev/null
+++ b/lib/zstd/compress/huf_compress.c
@@ -0,0 +1,905 @@
+/* ******************************************************************
+ * Huffman encoder, part of New Generation Entropy library
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ *  You can contact the author at :
+ *  - FSE+HUF source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *  - Public forum : https://groups.google.com/forum/#!forum/lz4c
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+/* **************************************************************
+*  Compiler specifics
+****************************************************************/
+
+
+/* **************************************************************
+*  Includes
+****************************************************************/
+#include "../common/zstd_deps.h"     /* ZSTD_memcpy, ZSTD_memset */
+#include "../common/compiler.h"
+#include "../common/bitstream.h"
+#include "hist.h"
+#define FSE_STATIC_LINKING_ONLY   /* FSE_optimalTableLog_internal */
+#include "../common/fse.h"        /* header compression */
+#define HUF_STATIC_LINKING_ONLY
+#include "../common/huf.h"
+#include "../common/error_private.h"
+
+
+/* **************************************************************
+*  Error Management
+****************************************************************/
+#define HUF_isError ERR_isError
+#define HUF_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c)   /* use only *after* variable declarations */
+
+
+/* **************************************************************
+*  Utils
+****************************************************************/
+unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue)
+{
+    return FSE_optimalTableLog_internal(maxTableLog, srcSize, maxSymbolValue, 1);
+}
+
+
+/* *******************************************************
+*  HUF : Huffman block compression
+*********************************************************/
+/* HUF_compressWeights() :
+ * Same as FSE_compress(), but dedicated to huff0's weights compression.
+ * The use case needs much less stack memory.
+ * Note : all elements within weightTable are supposed to be <= HUF_TABLELOG_MAX.
+ */
+#define MAX_FSE_TABLELOG_FOR_HUFF_HEADER 6
+
+typedef struct {
+    FSE_CTable CTable[FSE_CTABLE_SIZE_U32(MAX_FSE_TABLELOG_FOR_HUFF_HEADER, HUF_TABLELOG_MAX)];
+    U32 scratchBuffer[FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(HUF_TABLELOG_MAX, MAX_FSE_TABLELOG_FOR_HUFF_HEADER)];
+    unsigned count[HUF_TABLELOG_MAX+1];
+    S16 norm[HUF_TABLELOG_MAX+1];
+} HUF_CompressWeightsWksp;
+
+static size_t HUF_compressWeights(void* dst, size_t dstSize, const void* weightTable, size_t wtSize, void* workspace, size_t workspaceSize)
+{
+    BYTE* const ostart = (BYTE*) dst;
+    BYTE* op = ostart;
+    BYTE* const oend = ostart + dstSize;
+
+    unsigned maxSymbolValue = HUF_TABLELOG_MAX;
+    U32 tableLog = MAX_FSE_TABLELOG_FOR_HUFF_HEADER;
+    HUF_CompressWeightsWksp* wksp = (HUF_CompressWeightsWksp*)workspace;
+
+    if (workspaceSize < sizeof(HUF_CompressWeightsWksp)) return ERROR(GENERIC);
+
+    /* init conditions */
+    if (wtSize <= 1) return 0;  /* Not compressible */
+
+    /* Scan input and build symbol stats */
+    {   unsigned const maxCount = HIST_count_simple(wksp->count, &maxSymbolValue, weightTable, wtSize);   /* never fails */
+        if (maxCount == wtSize) return 1;   /* only a single symbol in src : rle */
+        if (maxCount == 1) return 0;        /* each symbol present maximum once => not compressible */
+    }
+
+    tableLog = FSE_optimalTableLog(tableLog, wtSize, maxSymbolValue);
+    CHECK_F( FSE_normalizeCount(wksp->norm, tableLog, wksp->count, wtSize, maxSymbolValue, /* useLowProbCount */ 0) );
+
+    /* Write table description header */
+    {   CHECK_V_F(hSize, FSE_writeNCount(op, (size_t)(oend-op), wksp->norm, maxSymbolValue, tableLog) );
+        op += hSize;
+    }
+
+    /* Compress */
+    CHECK_F( FSE_buildCTable_wksp(wksp->CTable, wksp->norm, maxSymbolValue, tableLog, wksp->scratchBuffer, sizeof(wksp->scratchBuffer)) );
+    {   CHECK_V_F(cSize, FSE_compress_usingCTable(op, (size_t)(oend - op), weightTable, wtSize, wksp->CTable) );
+        if (cSize == 0) return 0;   /* not enough space for compressed data */
+        op += cSize;
+    }
+
+    return (size_t)(op-ostart);
+}
+
+
+typedef struct {
+    HUF_CompressWeightsWksp wksp;
+    BYTE bitsToWeight[HUF_TABLELOG_MAX + 1];   /* precomputed conversion table */
+    BYTE huffWeight[HUF_SYMBOLVALUE_MAX];
+} HUF_WriteCTableWksp;
+
+size_t HUF_writeCTable_wksp(void* dst, size_t maxDstSize,
+                            const HUF_CElt* CTable, unsigned maxSymbolValue, unsigned huffLog,
+                            void* workspace, size_t workspaceSize)
+{
+    BYTE* op = (BYTE*)dst;
+    U32 n;
+    HUF_WriteCTableWksp* wksp = (HUF_WriteCTableWksp*)workspace;
+
+    /* check conditions */
+    if (workspaceSize < sizeof(HUF_WriteCTableWksp)) return ERROR(GENERIC);
+    if (maxSymbolValue > HUF_SYMBOLVALUE_MAX) return ERROR(maxSymbolValue_tooLarge);
+
+    /* convert to weight */
+    wksp->bitsToWeight[0] = 0;
+    for (n=1; n<huffLog+1; n++)
+        wksp->bitsToWeight[n] = (BYTE)(huffLog + 1 - n);
+    for (n=0; n<maxSymbolValue; n++)
+        wksp->huffWeight[n] = wksp->bitsToWeight[CTable[n].nbBits];
+
+    /* attempt weights compression by FSE */
+    {   CHECK_V_F(hSize, HUF_compressWeights(op+1, maxDstSize-1, wksp->huffWeight, maxSymbolValue, &wksp->wksp, sizeof(wksp->wksp)) );
+        if ((hSize>1) & (hSize < maxSymbolValue/2)) {   /* FSE compressed */
+            op[0] = (BYTE)hSize;
+            return hSize+1;
+    }   }
+
+    /* write raw values as 4-bits (max : 15) */
+    if (maxSymbolValue > (256-128)) return ERROR(GENERIC);   /* should not happen : likely means source cannot be compressed */
+    if (((maxSymbolValue+1)/2) + 1 > maxDstSize) return ERROR(dstSize_tooSmall);   /* not enough space within dst buffer */
+    op[0] = (BYTE)(128 /*special case*/ + (maxSymbolValue-1));
+    wksp->huffWeight[maxSymbolValue] = 0;   /* to be sure it doesn't cause msan issue in final combination */
+    for (n=0; n<maxSymbolValue; n+=2)
+        op[(n/2)+1] = (BYTE)((wksp->huffWeight[n] << 4) + wksp->huffWeight[n+1]);
+    return ((maxSymbolValue+1)/2) + 1;
+}
+
+/*! HUF_writeCTable() :
+    `CTable` : Huffman tree to save, using huf representation.
+    @return : size of saved CTable */
+size_t HUF_writeCTable (void* dst, size_t maxDstSize,
+                        const HUF_CElt* CTable, unsigned maxSymbolValue, unsigned huffLog)
+{
+    HUF_WriteCTableWksp wksp;
+    return HUF_writeCTable_wksp(dst, maxDstSize, CTable, maxSymbolValue, huffLog, &wksp, sizeof(wksp));
+}
+
+
+size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize, unsigned* hasZeroWeights)
+{
+    BYTE huffWeight[HUF_SYMBOLVALUE_MAX + 1];   /* init not required, even though some static analyzer may complain */
+    U32 rankVal[HUF_TABLELOG_ABSOLUTEMAX + 1];   /* large enough for values from 0 to 16 */
+    U32 tableLog = 0;
+    U32 nbSymbols = 0;
+
+    /* get symbol weights */
+    CHECK_V_F(readSize, HUF_readStats(huffWeight, HUF_SYMBOLVALUE_MAX+1, rankVal, &nbSymbols, &tableLog, src, srcSize));
+    *hasZeroWeights = (rankVal[0] > 0);
+
+    /* check result */
+    if (tableLog > HUF_TABLELOG_MAX) return ERROR(tableLog_tooLarge);
+    if (nbSymbols > *maxSymbolValuePtr+1) return ERROR(maxSymbolValue_tooSmall);
+
+    /* Prepare base value per rank */
+    {   U32 n, nextRankStart = 0;
+        for (n=1; n<=tableLog; n++) {
+            U32 curr = nextRankStart;
+            nextRankStart += (rankVal[n] << (n-1));
+            rankVal[n] = curr;
+    }   }
+
+    /* fill nbBits */
+    {   U32 n; for (n=0; n<nbSymbols; n++) {
+            const U32 w = huffWeight[n];
+            CTable[n].nbBits = (BYTE)(tableLog + 1 - w) & -(w != 0);
+    }   }
+
+    /* fill val */
+    {   U16 nbPerRank[HUF_TABLELOG_MAX+2]  = {0};  /* support w=0=>n=tableLog+1 */
+        U16 valPerRank[HUF_TABLELOG_MAX+2] = {0};
+        { U32 n; for (n=0; n<nbSymbols; n++) nbPerRank[CTable[n].nbBits]++; }
+        /* determine stating value per rank */
+        valPerRank[tableLog+1] = 0;   /* for w==0 */
+        {   U16 min = 0;
+            U32 n; for (n=tableLog; n>0; n--) {  /* start at n=tablelog <-> w=1 */
+                valPerRank[n] = min;     /* get starting value within each rank */
+                min += nbPerRank[n];
+                min >>= 1;
+        }   }
+        /* assign value within rank, symbol order */
+        { U32 n; for (n=0; n<nbSymbols; n++) CTable[n].val = valPerRank[CTable[n].nbBits]++; }
+    }
+
+    *maxSymbolValuePtr = nbSymbols - 1;
+    return readSize;
+}
+
+U32 HUF_getNbBits(const void* symbolTable, U32 symbolValue)
+{
+    const HUF_CElt* table = (const HUF_CElt*)symbolTable;
+    assert(symbolValue <= HUF_SYMBOLVALUE_MAX);
+    return table[symbolValue].nbBits;
+}
+
+
+typedef struct nodeElt_s {
+    U32 count;
+    U16 parent;
+    BYTE byte;
+    BYTE nbBits;
+} nodeElt;
+
+/*
+ * HUF_setMaxHeight():
+ * Enforces maxNbBits on the Huffman tree described in huffNode.
+ *
+ * It sets all nodes with nbBits > maxNbBits to be maxNbBits. Then it adjusts
+ * the tree to so that it is a valid canonical Huffman tree.
+ *
+ * @pre               The sum of the ranks of each symbol == 2^largestBits,
+ *                    where largestBits == huffNode[lastNonNull].nbBits.
+ * @post              The sum of the ranks of each symbol == 2^largestBits,
+ *                    where largestBits is the return value <= maxNbBits.
+ *
+ * @param huffNode    The Huffman tree modified in place to enforce maxNbBits.
+ * @param lastNonNull The symbol with the lowest count in the Huffman tree.
+ * @param maxNbBits   The maximum allowed number of bits, which the Huffman tree
+ *                    may not respect. After this function the Huffman tree will
+ *                    respect maxNbBits.
+ * @return            The maximum number of bits of the Huffman tree after adjustment,
+ *                    necessarily no more than maxNbBits.
+ */
+static U32 HUF_setMaxHeight(nodeElt* huffNode, U32 lastNonNull, U32 maxNbBits)
+{
+    const U32 largestBits = huffNode[lastNonNull].nbBits;
+    /* early exit : no elt > maxNbBits, so the tree is already valid. */
+    if (largestBits <= maxNbBits) return largestBits;
+
+    /* there are several too large elements (at least >= 2) */
+    {   int totalCost = 0;
+        const U32 baseCost = 1 << (largestBits - maxNbBits);
+        int n = (int)lastNonNull;
+
+        /* Adjust any ranks > maxNbBits to maxNbBits.
+         * Compute totalCost, which is how far the sum of the ranks is
+         * we are over 2^largestBits after adjust the offending ranks.
+         */
+        while (huffNode[n].nbBits > maxNbBits) {
+            totalCost += baseCost - (1 << (largestBits - huffNode[n].nbBits));
+            huffNode[n].nbBits = (BYTE)maxNbBits;
+            n--;
+        }
+        /* n stops at huffNode[n].nbBits <= maxNbBits */
+        assert(huffNode[n].nbBits <= maxNbBits);
+        /* n end at index of smallest symbol using < maxNbBits */
+        while (huffNode[n].nbBits == maxNbBits) --n;
+
+        /* renorm totalCost from 2^largestBits to 2^maxNbBits
+         * note : totalCost is necessarily a multiple of baseCost */
+        assert((totalCost & (baseCost - 1)) == 0);
+        totalCost >>= (largestBits - maxNbBits);
+        assert(totalCost > 0);
+
+        /* repay normalized cost */
+        {   U32 const noSymbol = 0xF0F0F0F0;
+            U32 rankLast[HUF_TABLELOG_MAX+2];
+
+            /* Get pos of last (smallest = lowest cum. count) symbol per rank */
+            ZSTD_memset(rankLast, 0xF0, sizeof(rankLast));
+            {   U32 currentNbBits = maxNbBits;
+                int pos;
+                for (pos=n ; pos >= 0; pos--) {
+                    if (huffNode[pos].nbBits >= currentNbBits) continue;
+                    currentNbBits = huffNode[pos].nbBits;   /* < maxNbBits */
+                    rankLast[maxNbBits-currentNbBits] = (U32)pos;
+            }   }
+
+            while (totalCost > 0) {
+                /* Try to reduce the next power of 2 above totalCost because we
+                 * gain back half the rank.
+                 */
+                U32 nBitsToDecrease = BIT_highbit32((U32)totalCost) + 1;
+                for ( ; nBitsToDecrease > 1; nBitsToDecrease--) {
+                    U32 const highPos = rankLast[nBitsToDecrease];
+                    U32 const lowPos = rankLast[nBitsToDecrease-1];
+                    if (highPos == noSymbol) continue;
+                    /* Decrease highPos if no symbols of lowPos or if it is
+                     * not cheaper to remove 2 lowPos than highPos.
+                     */
+                    if (lowPos == noSymbol) break;
+                    {   U32 const highTotal = huffNode[highPos].count;
+                        U32 const lowTotal = 2 * huffNode[lowPos].count;
+                        if (highTotal <= lowTotal) break;
+                }   }
+                /* only triggered when no more rank 1 symbol left => find closest one (note : there is necessarily at least one !) */
+                assert(rankLast[nBitsToDecrease] != noSymbol || nBitsToDecrease == 1);
+                /* HUF_MAX_TABLELOG test just to please gcc 5+; but it should not be necessary */
+                while ((nBitsToDecrease<=HUF_TABLELOG_MAX) && (rankLast[nBitsToDecrease] == noSymbol))
+                    nBitsToDecrease++;
+                assert(rankLast[nBitsToDecrease] != noSymbol);
+                /* Increase the number of bits to gain back half the rank cost. */
+                totalCost -= 1 << (nBitsToDecrease-1);
+                huffNode[rankLast[nBitsToDecrease]].nbBits++;
+
+                /* Fix up the new rank.
+                 * If the new rank was empty, this symbol is now its smallest.
+                 * Otherwise, this symbol will be the largest in the new rank so no adjustment.
+                 */
+                if (rankLast[nBitsToDecrease-1] == noSymbol)
+                    rankLast[nBitsToDecrease-1] = rankLast[nBitsToDecrease];
+                /* Fix up the old rank.
+                 * If the symbol was at position 0, meaning it was the highest weight symbol in the tree,
+                 * it must be the only symbol in its rank, so the old rank now has no symbols.
+                 * Otherwise, since the Huffman nodes are sorted by count, the previous position is now
+                 * the smallest node in the rank. If the previous position belongs to a different rank,
+                 * then the rank is now empty.
+                 */
+                if (rankLast[nBitsToDecrease] == 0)    /* special case, reached largest symbol */
+                    rankLast[nBitsToDecrease] = noSymbol;
+                else {
+                    rankLast[nBitsToDecrease]--;
+                    if (huffNode[rankLast[nBitsToDecrease]].nbBits != maxNbBits-nBitsToDecrease)
+                        rankLast[nBitsToDecrease] = noSymbol;   /* this rank is now empty */
+                }
+            }   /* while (totalCost > 0) */
+
+            /* If we've removed too much weight, then we have to add it back.
+             * To avoid overshooting again, we only adjust the smallest rank.
+             * We take the largest nodes from the lowest rank 0 and move them
+             * to rank 1. There's guaranteed to be enough rank 0 symbols because
+             * TODO.
+             */
+            while (totalCost < 0) {  /* Sometimes, cost correction overshoot */
+                /* special case : no rank 1 symbol (using maxNbBits-1);
+                 * let's create one from largest rank 0 (using maxNbBits).
+                 */
+                if (rankLast[1] == noSymbol) {
+                    while (huffNode[n].nbBits == maxNbBits) n--;
+                    huffNode[n+1].nbBits--;
+                    assert(n >= 0);
+                    rankLast[1] = (U32)(n+1);
+                    totalCost++;
+                    continue;
+                }
+                huffNode[ rankLast[1] + 1 ].nbBits--;
+                rankLast[1]++;
+                totalCost ++;
+            }
+        }   /* repay normalized cost */
+    }   /* there are several too large elements (at least >= 2) */
+
+    return maxNbBits;
+}
+
+typedef struct {
+    U32 base;
+    U32 curr;
+} rankPos;
+
+typedef nodeElt huffNodeTable[HUF_CTABLE_WORKSPACE_SIZE_U32];
+
+#define RANK_POSITION_TABLE_SIZE 32
+
+typedef struct {
+  huffNodeTable huffNodeTbl;
+  rankPos rankPosition[RANK_POSITION_TABLE_SIZE];
+} HUF_buildCTable_wksp_tables;
+
+/*
+ * HUF_sort():
+ * Sorts the symbols [0, maxSymbolValue] by count[symbol] in decreasing order.
+ *
+ * @param[out] huffNode       Sorted symbols by decreasing count. Only members `.count` and `.byte` are filled.
+ *                            Must have (maxSymbolValue + 1) entries.
+ * @param[in]  count          Histogram of the symbols.
+ * @param[in]  maxSymbolValue Maximum symbol value.
+ * @param      rankPosition   This is a scratch workspace. Must have RANK_POSITION_TABLE_SIZE entries.
+ */
+static void HUF_sort(nodeElt* huffNode, const unsigned* count, U32 maxSymbolValue, rankPos* rankPosition)
+{
+    int n;
+    int const maxSymbolValue1 = (int)maxSymbolValue + 1;
+
+    /* Compute base and set curr to base.
+     * For symbol s let lowerRank = BIT_highbit32(count[n]+1) and rank = lowerRank + 1.
+     * Then 2^lowerRank <= count[n]+1 <= 2^rank.
+     * We attribute each symbol to lowerRank's base value, because we want to know where
+     * each rank begins in the output, so for rank R we want to count ranks R+1 and above.
+     */
+    ZSTD_memset(rankPosition, 0, sizeof(*rankPosition) * RANK_POSITION_TABLE_SIZE);
+    for (n = 0; n < maxSymbolValue1; ++n) {
+        U32 lowerRank = BIT_highbit32(count[n] + 1);
+        rankPosition[lowerRank].base++;
+    }
+    assert(rankPosition[RANK_POSITION_TABLE_SIZE - 1].base == 0);
+    for (n = RANK_POSITION_TABLE_SIZE - 1; n > 0; --n) {
+        rankPosition[n-1].base += rankPosition[n].base;
+        rankPosition[n-1].curr = rankPosition[n-1].base;
+    }
+    /* Sort */
+    for (n = 0; n < maxSymbolValue1; ++n) {
+        U32 const c = count[n];
+        U32 const r = BIT_highbit32(c+1) + 1;
+        U32 pos = rankPosition[r].curr++;
+        /* Insert into the correct position in the rank.
+         * We have at most 256 symbols, so this insertion should be fine.
+         */
+        while ((pos > rankPosition[r].base) && (c > huffNode[pos-1].count)) {
+            huffNode[pos] = huffNode[pos-1];
+            pos--;
+        }
+        huffNode[pos].count = c;
+        huffNode[pos].byte  = (BYTE)n;
+    }
+}
+
+
+/* HUF_buildCTable_wksp() :
+ *  Same as HUF_buildCTable(), but using externally allocated scratch buffer.
+ *  `workSpace` must be aligned on 4-bytes boundaries, and be at least as large as sizeof(HUF_buildCTable_wksp_tables).
+ */
+#define STARTNODE (HUF_SYMBOLVALUE_MAX+1)
+
+/* HUF_buildTree():
+ * Takes the huffNode array sorted by HUF_sort() and builds an unlimited-depth Huffman tree.
+ *
+ * @param huffNode        The array sorted by HUF_sort(). Builds the Huffman tree in this array.
+ * @param maxSymbolValue  The maximum symbol value.
+ * @return                The smallest node in the Huffman tree (by count).
+ */
+static int HUF_buildTree(nodeElt* huffNode, U32 maxSymbolValue)
+{
+    nodeElt* const huffNode0 = huffNode - 1;
+    int nonNullRank;
+    int lowS, lowN;
+    int nodeNb = STARTNODE;
+    int n, nodeRoot;
+    /* init for parents */
+    nonNullRank = (int)maxSymbolValue;
+    while(huffNode[nonNullRank].count == 0) nonNullRank--;
+    lowS = nonNullRank; nodeRoot = nodeNb + lowS - 1; lowN = nodeNb;
+    huffNode[nodeNb].count = huffNode[lowS].count + huffNode[lowS-1].count;
+    huffNode[lowS].parent = huffNode[lowS-1].parent = (U16)nodeNb;
+    nodeNb++; lowS-=2;
+    for (n=nodeNb; n<=nodeRoot; n++) huffNode[n].count = (U32)(1U<<30);
+    huffNode0[0].count = (U32)(1U<<31);  /* fake entry, strong barrier */
+
+    /* create parents */
+    while (nodeNb <= nodeRoot) {
+        int const n1 = (huffNode[lowS].count < huffNode[lowN].count) ? lowS-- : lowN++;
+        int const n2 = (huffNode[lowS].count < huffNode[lowN].count) ? lowS-- : lowN++;
+        huffNode[nodeNb].count = huffNode[n1].count + huffNode[n2].count;
+        huffNode[n1].parent = huffNode[n2].parent = (U16)nodeNb;
+        nodeNb++;
+    }
+
+    /* distribute weights (unlimited tree height) */
+    huffNode[nodeRoot].nbBits = 0;
+    for (n=nodeRoot-1; n>=STARTNODE; n--)
+        huffNode[n].nbBits = huffNode[ huffNode[n].parent ].nbBits + 1;
+    for (n=0; n<=nonNullRank; n++)
+        huffNode[n].nbBits = huffNode[ huffNode[n].parent ].nbBits + 1;
+
+    return nonNullRank;
+}
+
+/*
+ * HUF_buildCTableFromTree():
+ * Build the CTable given the Huffman tree in huffNode.
+ *
+ * @param[out] CTable         The output Huffman CTable.
+ * @param      huffNode       The Huffman tree.
+ * @param      nonNullRank    The last and smallest node in the Huffman tree.
+ * @param      maxSymbolValue The maximum symbol value.
+ * @param      maxNbBits      The exact maximum number of bits used in the Huffman tree.
+ */
+static void HUF_buildCTableFromTree(HUF_CElt* CTable, nodeElt const* huffNode, int nonNullRank, U32 maxSymbolValue, U32 maxNbBits)
+{
+    /* fill result into ctable (val, nbBits) */
+    int n;
+    U16 nbPerRank[HUF_TABLELOG_MAX+1] = {0};
+    U16 valPerRank[HUF_TABLELOG_MAX+1] = {0};
+    int const alphabetSize = (int)(maxSymbolValue + 1);
+    for (n=0; n<=nonNullRank; n++)
+        nbPerRank[huffNode[n].nbBits]++;
+    /* determine starting value per rank */
+    {   U16 min = 0;
+        for (n=(int)maxNbBits; n>0; n--) {
+            valPerRank[n] = min;      /* get starting value within each rank */
+            min += nbPerRank[n];
+            min >>= 1;
+    }   }
+    for (n=0; n<alphabetSize; n++)
+        CTable[huffNode[n].byte].nbBits = huffNode[n].nbBits;   /* push nbBits per symbol, symbol order */
+    for (n=0; n<alphabetSize; n++)
+        CTable[n].val = valPerRank[CTable[n].nbBits]++;   /* assign value within rank, symbol order */
+}
+
+size_t HUF_buildCTable_wksp (HUF_CElt* tree, const unsigned* count, U32 maxSymbolValue, U32 maxNbBits, void* workSpace, size_t wkspSize)
+{
+    HUF_buildCTable_wksp_tables* const wksp_tables = (HUF_buildCTable_wksp_tables*)workSpace;
+    nodeElt* const huffNode0 = wksp_tables->huffNodeTbl;
+    nodeElt* const huffNode = huffNode0+1;
+    int nonNullRank;
+
+    /* safety checks */
+    if (((size_t)workSpace & 3) != 0) return ERROR(GENERIC);  /* must be aligned on 4-bytes boundaries */
+    if (wkspSize < sizeof(HUF_buildCTable_wksp_tables))
+      return ERROR(workSpace_tooSmall);
+    if (maxNbBits == 0) maxNbBits = HUF_TABLELOG_DEFAULT;
+    if (maxSymbolValue > HUF_SYMBOLVALUE_MAX)
+      return ERROR(maxSymbolValue_tooLarge);
+    ZSTD_memset(huffNode0, 0, sizeof(huffNodeTable));
+
+    /* sort, decreasing order */
+    HUF_sort(huffNode, count, maxSymbolValue, wksp_tables->rankPosition);
+
+    /* build tree */
+    nonNullRank = HUF_buildTree(huffNode, maxSymbolValue);
+
+    /* enforce maxTableLog */
+    maxNbBits = HUF_setMaxHeight(huffNode, (U32)nonNullRank, maxNbBits);
+    if (maxNbBits > HUF_TABLELOG_MAX) return ERROR(GENERIC);   /* check fit into table */
+
+    HUF_buildCTableFromTree(tree, huffNode, nonNullRank, maxSymbolValue, maxNbBits);
+
+    return maxNbBits;
+}
+
+size_t HUF_estimateCompressedSize(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue)
+{
+    size_t nbBits = 0;
+    int s;
+    for (s = 0; s <= (int)maxSymbolValue; ++s) {
+        nbBits += CTable[s].nbBits * count[s];
+    }
+    return nbBits >> 3;
+}
+
+int HUF_validateCTable(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue) {
+  int bad = 0;
+  int s;
+  for (s = 0; s <= (int)maxSymbolValue; ++s) {
+    bad |= (count[s] != 0) & (CTable[s].nbBits == 0);
+  }
+  return !bad;
+}
+
+size_t HUF_compressBound(size_t size) { return HUF_COMPRESSBOUND(size); }
+
+FORCE_INLINE_TEMPLATE void
+HUF_encodeSymbol(BIT_CStream_t* bitCPtr, U32 symbol, const HUF_CElt* CTable)
+{
+    BIT_addBitsFast(bitCPtr, CTable[symbol].val, CTable[symbol].nbBits);
+}
+
+#define HUF_FLUSHBITS(s)  BIT_flushBits(s)
+
+#define HUF_FLUSHBITS_1(stream) \
+    if (sizeof((stream)->bitContainer)*8 < HUF_TABLELOG_MAX*2+7) HUF_FLUSHBITS(stream)
+
+#define HUF_FLUSHBITS_2(stream) \
+    if (sizeof((stream)->bitContainer)*8 < HUF_TABLELOG_MAX*4+7) HUF_FLUSHBITS(stream)
+
+FORCE_INLINE_TEMPLATE size_t
+HUF_compress1X_usingCTable_internal_body(void* dst, size_t dstSize,
+                                   const void* src, size_t srcSize,
+                                   const HUF_CElt* CTable)
+{
+    const BYTE* ip = (const BYTE*) src;
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* const oend = ostart + dstSize;
+    BYTE* op = ostart;
+    size_t n;
+    BIT_CStream_t bitC;
+
+    /* init */
+    if (dstSize < 8) return 0;   /* not enough space to compress */
+    { size_t const initErr = BIT_initCStream(&bitC, op, (size_t)(oend-op));
+      if (HUF_isError(initErr)) return 0; }
+
+    n = srcSize & ~3;  /* join to mod 4 */
+    switch (srcSize & 3)
+    {
+        case 3:
+            HUF_encodeSymbol(&bitC, ip[n+ 2], CTable);
+            HUF_FLUSHBITS_2(&bitC);
+            ZSTD_FALLTHROUGH;
+        case 2:
+            HUF_encodeSymbol(&bitC, ip[n+ 1], CTable);
+            HUF_FLUSHBITS_1(&bitC);
+            ZSTD_FALLTHROUGH;
+        case 1:
+            HUF_encodeSymbol(&bitC, ip[n+ 0], CTable);
+            HUF_FLUSHBITS(&bitC);
+            ZSTD_FALLTHROUGH;
+        case 0: ZSTD_FALLTHROUGH;
+        default: break;
+    }
+
+    for (; n>0; n-=4) {  /* note : n&3==0 at this stage */
+        HUF_encodeSymbol(&bitC, ip[n- 1], CTable);
+        HUF_FLUSHBITS_1(&bitC);
+        HUF_encodeSymbol(&bitC, ip[n- 2], CTable);
+        HUF_FLUSHBITS_2(&bitC);
+        HUF_encodeSymbol(&bitC, ip[n- 3], CTable);
+        HUF_FLUSHBITS_1(&bitC);
+        HUF_encodeSymbol(&bitC, ip[n- 4], CTable);
+        HUF_FLUSHBITS(&bitC);
+    }
+
+    return BIT_closeCStream(&bitC);
+}
+
+#if DYNAMIC_BMI2
+
+static TARGET_ATTRIBUTE("bmi2") size_t
+HUF_compress1X_usingCTable_internal_bmi2(void* dst, size_t dstSize,
+                                   const void* src, size_t srcSize,
+                                   const HUF_CElt* CTable)
+{
+    return HUF_compress1X_usingCTable_internal_body(dst, dstSize, src, srcSize, CTable);
+}
+
+static size_t
+HUF_compress1X_usingCTable_internal_default(void* dst, size_t dstSize,
+                                      const void* src, size_t srcSize,
+                                      const HUF_CElt* CTable)
+{
+    return HUF_compress1X_usingCTable_internal_body(dst, dstSize, src, srcSize, CTable);
+}
+
+static size_t
+HUF_compress1X_usingCTable_internal(void* dst, size_t dstSize,
+                              const void* src, size_t srcSize,
+                              const HUF_CElt* CTable, const int bmi2)
+{
+    if (bmi2) {
+        return HUF_compress1X_usingCTable_internal_bmi2(dst, dstSize, src, srcSize, CTable);
+    }
+    return HUF_compress1X_usingCTable_internal_default(dst, dstSize, src, srcSize, CTable);
+}
+
+#else
+
+static size_t
+HUF_compress1X_usingCTable_internal(void* dst, size_t dstSize,
+                              const void* src, size_t srcSize,
+                              const HUF_CElt* CTable, const int bmi2)
+{
+    (void)bmi2;
+    return HUF_compress1X_usingCTable_internal_body(dst, dstSize, src, srcSize, CTable);
+}
+
+#endif
+
+size_t HUF_compress1X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable)
+{
+    return HUF_compress1X_usingCTable_internal(dst, dstSize, src, srcSize, CTable, /* bmi2 */ 0);
+}
+
+
+static size_t
+HUF_compress4X_usingCTable_internal(void* dst, size_t dstSize,
+                              const void* src, size_t srcSize,
+                              const HUF_CElt* CTable, int bmi2)
+{
+    size_t const segmentSize = (srcSize+3)/4;   /* first 3 segments */
+    const BYTE* ip = (const BYTE*) src;
+    const BYTE* const iend = ip + srcSize;
+    BYTE* const ostart = (BYTE*) dst;
+    BYTE* const oend = ostart + dstSize;
+    BYTE* op = ostart;
+
+    if (dstSize < 6 + 1 + 1 + 1 + 8) return 0;   /* minimum space to compress successfully */
+    if (srcSize < 12) return 0;   /* no saving possible : too small input */
+    op += 6;   /* jumpTable */
+
+    assert(op <= oend);
+    {   CHECK_V_F(cSize, HUF_compress1X_usingCTable_internal(op, (size_t)(oend-op), ip, segmentSize, CTable, bmi2) );
+        if (cSize==0) return 0;
+        assert(cSize <= 65535);
+        MEM_writeLE16(ostart, (U16)cSize);
+        op += cSize;
+    }
+
+    ip += segmentSize;
+    assert(op <= oend);
+    {   CHECK_V_F(cSize, HUF_compress1X_usingCTable_internal(op, (size_t)(oend-op), ip, segmentSize, CTable, bmi2) );
+        if (cSize==0) return 0;
+        assert(cSize <= 65535);
+        MEM_writeLE16(ostart+2, (U16)cSize);
+        op += cSize;
+    }
+
+    ip += segmentSize;
+    assert(op <= oend);
+    {   CHECK_V_F(cSize, HUF_compress1X_usingCTable_internal(op, (size_t)(oend-op), ip, segmentSize, CTable, bmi2) );
+        if (cSize==0) return 0;
+        assert(cSize <= 65535);
+        MEM_writeLE16(ostart+4, (U16)cSize);
+        op += cSize;
+    }
+
+    ip += segmentSize;
+    assert(op <= oend);
+    assert(ip <= iend);
+    {   CHECK_V_F(cSize, HUF_compress1X_usingCTable_internal(op, (size_t)(oend-op), ip, (size_t)(iend-ip), CTable, bmi2) );
+        if (cSize==0) return 0;
+        op += cSize;
+    }
+
+    return (size_t)(op-ostart);
+}
+
+size_t HUF_compress4X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable)
+{
+    return HUF_compress4X_usingCTable_internal(dst, dstSize, src, srcSize, CTable, /* bmi2 */ 0);
+}
+
+typedef enum { HUF_singleStream, HUF_fourStreams } HUF_nbStreams_e;
+
+static size_t HUF_compressCTable_internal(
+                BYTE* const ostart, BYTE* op, BYTE* const oend,
+                const void* src, size_t srcSize,
+                HUF_nbStreams_e nbStreams, const HUF_CElt* CTable, const int bmi2)
+{
+    size_t const cSize = (nbStreams==HUF_singleStream) ?
+                         HUF_compress1X_usingCTable_internal(op, (size_t)(oend - op), src, srcSize, CTable, bmi2) :
+                         HUF_compress4X_usingCTable_internal(op, (size_t)(oend - op), src, srcSize, CTable, bmi2);
+    if (HUF_isError(cSize)) { return cSize; }
+    if (cSize==0) { return 0; }   /* uncompressible */
+    op += cSize;
+    /* check compressibility */
+    assert(op >= ostart);
+    if ((size_t)(op-ostart) >= srcSize-1) { return 0; }
+    return (size_t)(op-ostart);
+}
+
+typedef struct {
+    unsigned count[HUF_SYMBOLVALUE_MAX + 1];
+    HUF_CElt CTable[HUF_SYMBOLVALUE_MAX + 1];
+    union {
+        HUF_buildCTable_wksp_tables buildCTable_wksp;
+        HUF_WriteCTableWksp writeCTable_wksp;
+    } wksps;
+} HUF_compress_tables_t;
+
+/* HUF_compress_internal() :
+ * `workSpace_align4` must be aligned on 4-bytes boundaries,
+ * and occupies the same space as a table of HUF_WORKSPACE_SIZE_U32 unsigned */
+static size_t
+HUF_compress_internal (void* dst, size_t dstSize,
+                 const void* src, size_t srcSize,
+                       unsigned maxSymbolValue, unsigned huffLog,
+                       HUF_nbStreams_e nbStreams,
+                       void* workSpace_align4, size_t wkspSize,
+                       HUF_CElt* oldHufTable, HUF_repeat* repeat, int preferRepeat,
+                 const int bmi2)
+{
+    HUF_compress_tables_t* const table = (HUF_compress_tables_t*)workSpace_align4;
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* const oend = ostart + dstSize;
+    BYTE* op = ostart;
+
+    HUF_STATIC_ASSERT(sizeof(*table) <= HUF_WORKSPACE_SIZE);
+    assert(((size_t)workSpace_align4 & 3) == 0);   /* must be aligned on 4-bytes boundaries */
+
+    /* checks & inits */
+    if (wkspSize < HUF_WORKSPACE_SIZE) return ERROR(workSpace_tooSmall);
+    if (!srcSize) return 0;  /* Uncompressed */
+    if (!dstSize) return 0;  /* cannot fit anything within dst budget */
+    if (srcSize > HUF_BLOCKSIZE_MAX) return ERROR(srcSize_wrong);   /* current block size limit */
+    if (huffLog > HUF_TABLELOG_MAX) return ERROR(tableLog_tooLarge);
+    if (maxSymbolValue > HUF_SYMBOLVALUE_MAX) return ERROR(maxSymbolValue_tooLarge);
+    if (!maxSymbolValue) maxSymbolValue = HUF_SYMBOLVALUE_MAX;
+    if (!huffLog) huffLog = HUF_TABLELOG_DEFAULT;
+
+    /* Heuristic : If old table is valid, use it for small inputs */
+    if (preferRepeat && repeat && *repeat == HUF_repeat_valid) {
+        return HUF_compressCTable_internal(ostart, op, oend,
+                                           src, srcSize,
+                                           nbStreams, oldHufTable, bmi2);
+    }
+
+    /* Scan input and build symbol stats */
+    {   CHECK_V_F(largest, HIST_count_wksp (table->count, &maxSymbolValue, (const BYTE*)src, srcSize, workSpace_align4, wkspSize) );
+        if (largest == srcSize) { *ostart = ((const BYTE*)src)[0]; return 1; }   /* single symbol, rle */
+        if (largest <= (srcSize >> 7)+4) return 0;   /* heuristic : probably not compressible enough */
+    }
+
+    /* Check validity of previous table */
+    if ( repeat
+      && *repeat == HUF_repeat_check
+      && !HUF_validateCTable(oldHufTable, table->count, maxSymbolValue)) {
+        *repeat = HUF_repeat_none;
+    }
+    /* Heuristic : use existing table for small inputs */
+    if (preferRepeat && repeat && *repeat != HUF_repeat_none) {
+        return HUF_compressCTable_internal(ostart, op, oend,
+                                           src, srcSize,
+                                           nbStreams, oldHufTable, bmi2);
+    }
+
+    /* Build Huffman Tree */
+    huffLog = HUF_optimalTableLog(huffLog, srcSize, maxSymbolValue);
+    {   size_t const maxBits = HUF_buildCTable_wksp(table->CTable, table->count,
+                                            maxSymbolValue, huffLog,
+                                            &table->wksps.buildCTable_wksp, sizeof(table->wksps.buildCTable_wksp));
+        CHECK_F(maxBits);
+        huffLog = (U32)maxBits;
+        /* Zero unused symbols in CTable, so we can check it for validity */
+        ZSTD_memset(table->CTable + (maxSymbolValue + 1), 0,
+               sizeof(table->CTable) - ((maxSymbolValue + 1) * sizeof(HUF_CElt)));
+    }
+
+    /* Write table description header */
+    {   CHECK_V_F(hSize, HUF_writeCTable_wksp(op, dstSize, table->CTable, maxSymbolValue, huffLog,
+                                              &table->wksps.writeCTable_wksp, sizeof(table->wksps.writeCTable_wksp)) );
+        /* Check if using previous huffman table is beneficial */
+        if (repeat && *repeat != HUF_repeat_none) {
+            size_t const oldSize = HUF_estimateCompressedSize(oldHufTable, table->count, maxSymbolValue);
+            size_t const newSize = HUF_estimateCompressedSize(table->CTable, table->count, maxSymbolValue);
+            if (oldSize <= hSize + newSize || hSize + 12 >= srcSize) {
+                return HUF_compressCTable_internal(ostart, op, oend,
+                                                   src, srcSize,
+                                                   nbStreams, oldHufTable, bmi2);
+        }   }
+
+        /* Use the new huffman table */
+        if (hSize + 12ul >= srcSize) { return 0; }
+        op += hSize;
+        if (repeat) { *repeat = HUF_repeat_none; }
+        if (oldHufTable)
+            ZSTD_memcpy(oldHufTable, table->CTable, sizeof(table->CTable));  /* Save new table */
+    }
+    return HUF_compressCTable_internal(ostart, op, oend,
+                                       src, srcSize,
+                                       nbStreams, table->CTable, bmi2);
+}
+
+
+size_t HUF_compress1X_wksp (void* dst, size_t dstSize,
+                      const void* src, size_t srcSize,
+                      unsigned maxSymbolValue, unsigned huffLog,
+                      void* workSpace, size_t wkspSize)
+{
+    return HUF_compress_internal(dst, dstSize, src, srcSize,
+                                 maxSymbolValue, huffLog, HUF_singleStream,
+                                 workSpace, wkspSize,
+                                 NULL, NULL, 0, 0 /*bmi2*/);
+}
+
+size_t HUF_compress1X_repeat (void* dst, size_t dstSize,
+                      const void* src, size_t srcSize,
+                      unsigned maxSymbolValue, unsigned huffLog,
+                      void* workSpace, size_t wkspSize,
+                      HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2)
+{
+    return HUF_compress_internal(dst, dstSize, src, srcSize,
+                                 maxSymbolValue, huffLog, HUF_singleStream,
+                                 workSpace, wkspSize, hufTable,
+                                 repeat, preferRepeat, bmi2);
+}
+
+/* HUF_compress4X_repeat():
+ * compress input using 4 streams.
+ * provide workspace to generate compression tables */
+size_t HUF_compress4X_wksp (void* dst, size_t dstSize,
+                      const void* src, size_t srcSize,
+                      unsigned maxSymbolValue, unsigned huffLog,
+                      void* workSpace, size_t wkspSize)
+{
+    return HUF_compress_internal(dst, dstSize, src, srcSize,
+                                 maxSymbolValue, huffLog, HUF_fourStreams,
+                                 workSpace, wkspSize,
+                                 NULL, NULL, 0, 0 /*bmi2*/);
+}
+
+/* HUF_compress4X_repeat():
+ * compress input using 4 streams.
+ * re-use an existing huffman compression table */
+size_t HUF_compress4X_repeat (void* dst, size_t dstSize,
+                      const void* src, size_t srcSize,
+                      unsigned maxSymbolValue, unsigned huffLog,
+                      void* workSpace, size_t wkspSize,
+                      HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2)
+{
+    return HUF_compress_internal(dst, dstSize, src, srcSize,
+                                 maxSymbolValue, huffLog, HUF_fourStreams,
+                                 workSpace, wkspSize,
+                                 hufTable, repeat, preferRepeat, bmi2);
+}
+
diff --git a/lib/zstd/compress/zstd_compress.c b/lib/zstd/compress/zstd_compress.c
new file mode 100644
index 0000000..a4e9160
--- /dev/null
+++ b/lib/zstd/compress/zstd_compress.c
@@ -0,0 +1,5109 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/*-*************************************
+*  Dependencies
+***************************************/
+#include "../common/zstd_deps.h"  /* INT_MAX, ZSTD_memset, ZSTD_memcpy */
+#include "../common/cpu.h"
+#include "../common/mem.h"
+#include "hist.h"           /* HIST_countFast_wksp */
+#define FSE_STATIC_LINKING_ONLY   /* FSE_encodeSymbol */
+#include "../common/fse.h"
+#define HUF_STATIC_LINKING_ONLY
+#include "../common/huf.h"
+#include "zstd_compress_internal.h"
+#include "zstd_compress_sequences.h"
+#include "zstd_compress_literals.h"
+#include "zstd_fast.h"
+#include "zstd_double_fast.h"
+#include "zstd_lazy.h"
+#include "zstd_opt.h"
+#include "zstd_ldm.h"
+#include "zstd_compress_superblock.h"
+
+/* ***************************************************************
+*  Tuning parameters
+*****************************************************************/
+/*!
+ * COMPRESS_HEAPMODE :
+ * Select how default decompression function ZSTD_compress() allocates its context,
+ * on stack (0, default), or into heap (1).
+ * Note that functions with explicit context such as ZSTD_compressCCtx() are unaffected.
+ */
+
+
+/*-*************************************
+*  Helper functions
+***************************************/
+/* ZSTD_compressBound()
+ * Note that the result from this function is only compatible with the "normal"
+ * full-block strategy.
+ * When there are a lot of small blocks due to frequent flush in streaming mode
+ * the overhead of headers can make the compressed data to be larger than the
+ * return value of ZSTD_compressBound().
+ */
+size_t ZSTD_compressBound(size_t srcSize) {
+    return ZSTD_COMPRESSBOUND(srcSize);
+}
+
+
+/*-*************************************
+*  Context memory management
+***************************************/
+struct ZSTD_CDict_s {
+    const void* dictContent;
+    size_t dictContentSize;
+    ZSTD_dictContentType_e dictContentType; /* The dictContentType the CDict was created with */
+    U32* entropyWorkspace; /* entropy workspace of HUF_WORKSPACE_SIZE bytes */
+    ZSTD_cwksp workspace;
+    ZSTD_matchState_t matchState;
+    ZSTD_compressedBlockState_t cBlockState;
+    ZSTD_customMem customMem;
+    U32 dictID;
+    int compressionLevel; /* 0 indicates that advanced API was used to select CDict params */
+};  /* typedef'd to ZSTD_CDict within "zstd.h" */
+
+ZSTD_CCtx* ZSTD_createCCtx(void)
+{
+    return ZSTD_createCCtx_advanced(ZSTD_defaultCMem);
+}
+
+static void ZSTD_initCCtx(ZSTD_CCtx* cctx, ZSTD_customMem memManager)
+{
+    assert(cctx != NULL);
+    ZSTD_memset(cctx, 0, sizeof(*cctx));
+    cctx->customMem = memManager;
+    cctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid());
+    {   size_t const err = ZSTD_CCtx_reset(cctx, ZSTD_reset_parameters);
+        assert(!ZSTD_isError(err));
+        (void)err;
+    }
+}
+
+ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem)
+{
+    ZSTD_STATIC_ASSERT(zcss_init==0);
+    ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN==(0ULL - 1));
+    if ((!customMem.customAlloc) ^ (!customMem.customFree)) return NULL;
+    {   ZSTD_CCtx* const cctx = (ZSTD_CCtx*)ZSTD_customMalloc(sizeof(ZSTD_CCtx), customMem);
+        if (!cctx) return NULL;
+        ZSTD_initCCtx(cctx, customMem);
+        return cctx;
+    }
+}
+
+ZSTD_CCtx* ZSTD_initStaticCCtx(void* workspace, size_t workspaceSize)
+{
+    ZSTD_cwksp ws;
+    ZSTD_CCtx* cctx;
+    if (workspaceSize <= sizeof(ZSTD_CCtx)) return NULL;  /* minimum size */
+    if ((size_t)workspace & 7) return NULL;  /* must be 8-aligned */
+    ZSTD_cwksp_init(&ws, workspace, workspaceSize, ZSTD_cwksp_static_alloc);
+
+    cctx = (ZSTD_CCtx*)ZSTD_cwksp_reserve_object(&ws, sizeof(ZSTD_CCtx));
+    if (cctx == NULL) return NULL;
+
+    ZSTD_memset(cctx, 0, sizeof(ZSTD_CCtx));
+    ZSTD_cwksp_move(&cctx->workspace, &ws);
+    cctx->staticSize = workspaceSize;
+
+    /* statically sized space. entropyWorkspace never moves (but prev/next block swap places) */
+    if (!ZSTD_cwksp_check_available(&cctx->workspace, ENTROPY_WORKSPACE_SIZE + 2 * sizeof(ZSTD_compressedBlockState_t))) return NULL;
+    cctx->blockState.prevCBlock = (ZSTD_compressedBlockState_t*)ZSTD_cwksp_reserve_object(&cctx->workspace, sizeof(ZSTD_compressedBlockState_t));
+    cctx->blockState.nextCBlock = (ZSTD_compressedBlockState_t*)ZSTD_cwksp_reserve_object(&cctx->workspace, sizeof(ZSTD_compressedBlockState_t));
+    cctx->entropyWorkspace = (U32*)ZSTD_cwksp_reserve_object(&cctx->workspace, ENTROPY_WORKSPACE_SIZE);
+    cctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid());
+    return cctx;
+}
+
+/*
+ * Clears and frees all of the dictionaries in the CCtx.
+ */
+static void ZSTD_clearAllDicts(ZSTD_CCtx* cctx)
+{
+    ZSTD_customFree(cctx->localDict.dictBuffer, cctx->customMem);
+    ZSTD_freeCDict(cctx->localDict.cdict);
+    ZSTD_memset(&cctx->localDict, 0, sizeof(cctx->localDict));
+    ZSTD_memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict));
+    cctx->cdict = NULL;
+}
+
+static size_t ZSTD_sizeof_localDict(ZSTD_localDict dict)
+{
+    size_t const bufferSize = dict.dictBuffer != NULL ? dict.dictSize : 0;
+    size_t const cdictSize = ZSTD_sizeof_CDict(dict.cdict);
+    return bufferSize + cdictSize;
+}
+
+static void ZSTD_freeCCtxContent(ZSTD_CCtx* cctx)
+{
+    assert(cctx != NULL);
+    assert(cctx->staticSize == 0);
+    ZSTD_clearAllDicts(cctx);
+    ZSTD_cwksp_free(&cctx->workspace, cctx->customMem);
+}
+
+size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx)
+{
+    if (cctx==NULL) return 0;   /* support free on NULL */
+    RETURN_ERROR_IF(cctx->staticSize, memory_allocation,
+                    "not compatible with static CCtx");
+    {
+        int cctxInWorkspace = ZSTD_cwksp_owns_buffer(&cctx->workspace, cctx);
+        ZSTD_freeCCtxContent(cctx);
+        if (!cctxInWorkspace) {
+            ZSTD_customFree(cctx, cctx->customMem);
+        }
+    }
+    return 0;
+}
+
+
+static size_t ZSTD_sizeof_mtctx(const ZSTD_CCtx* cctx)
+{
+    (void)cctx;
+    return 0;
+}
+
+
+size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx)
+{
+    if (cctx==NULL) return 0;   /* support sizeof on NULL */
+    /* cctx may be in the workspace */
+    return (cctx->workspace.workspace == cctx ? 0 : sizeof(*cctx))
+           + ZSTD_cwksp_sizeof(&cctx->workspace)
+           + ZSTD_sizeof_localDict(cctx->localDict)
+           + ZSTD_sizeof_mtctx(cctx);
+}
+
+size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs)
+{
+    return ZSTD_sizeof_CCtx(zcs);  /* same object */
+}
+
+/* private API call, for dictBuilder only */
+const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx) { return &(ctx->seqStore); }
+
+/* Returns 1 if compression parameters are such that we should
+ * enable long distance matching (wlog >= 27, strategy >= btopt).
+ * Returns 0 otherwise.
+ */
+static U32 ZSTD_CParams_shouldEnableLdm(const ZSTD_compressionParameters* const cParams) {
+    return cParams->strategy >= ZSTD_btopt && cParams->windowLog >= 27;
+}
+
+static ZSTD_CCtx_params ZSTD_makeCCtxParamsFromCParams(
+        ZSTD_compressionParameters cParams)
+{
+    ZSTD_CCtx_params cctxParams;
+    /* should not matter, as all cParams are presumed properly defined */
+    ZSTD_CCtxParams_init(&cctxParams, ZSTD_CLEVEL_DEFAULT);
+    cctxParams.cParams = cParams;
+
+    if (ZSTD_CParams_shouldEnableLdm(&cParams)) {
+        DEBUGLOG(4, "ZSTD_makeCCtxParamsFromCParams(): Including LDM into cctx params");
+        cctxParams.ldmParams.enableLdm = 1;
+        /* LDM is enabled by default for optimal parser and window size >= 128MB */
+        ZSTD_ldm_adjustParameters(&cctxParams.ldmParams, &cParams);
+        assert(cctxParams.ldmParams.hashLog >= cctxParams.ldmParams.bucketSizeLog);
+        assert(cctxParams.ldmParams.hashRateLog < 32);
+    }
+
+    assert(!ZSTD_checkCParams(cParams));
+    return cctxParams;
+}
+
+static ZSTD_CCtx_params* ZSTD_createCCtxParams_advanced(
+        ZSTD_customMem customMem)
+{
+    ZSTD_CCtx_params* params;
+    if ((!customMem.customAlloc) ^ (!customMem.customFree)) return NULL;
+    params = (ZSTD_CCtx_params*)ZSTD_customCalloc(
+            sizeof(ZSTD_CCtx_params), customMem);
+    if (!params) { return NULL; }
+    ZSTD_CCtxParams_init(params, ZSTD_CLEVEL_DEFAULT);
+    params->customMem = customMem;
+    return params;
+}
+
+ZSTD_CCtx_params* ZSTD_createCCtxParams(void)
+{
+    return ZSTD_createCCtxParams_advanced(ZSTD_defaultCMem);
+}
+
+size_t ZSTD_freeCCtxParams(ZSTD_CCtx_params* params)
+{
+    if (params == NULL) { return 0; }
+    ZSTD_customFree(params, params->customMem);
+    return 0;
+}
+
+size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_params* params)
+{
+    return ZSTD_CCtxParams_init(params, ZSTD_CLEVEL_DEFAULT);
+}
+
+size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel) {
+    RETURN_ERROR_IF(!cctxParams, GENERIC, "NULL pointer!");
+    ZSTD_memset(cctxParams, 0, sizeof(*cctxParams));
+    cctxParams->compressionLevel = compressionLevel;
+    cctxParams->fParams.contentSizeFlag = 1;
+    return 0;
+}
+
+#define ZSTD_NO_CLEVEL 0
+
+/*
+ * Initializes the cctxParams from params and compressionLevel.
+ * @param compressionLevel If params are derived from a compression level then that compression level, otherwise ZSTD_NO_CLEVEL.
+ */
+static void ZSTD_CCtxParams_init_internal(ZSTD_CCtx_params* cctxParams, ZSTD_parameters const* params, int compressionLevel)
+{
+    assert(!ZSTD_checkCParams(params->cParams));
+    ZSTD_memset(cctxParams, 0, sizeof(*cctxParams));
+    cctxParams->cParams = params->cParams;
+    cctxParams->fParams = params->fParams;
+    /* Should not matter, as all cParams are presumed properly defined.
+     * But, set it for tracing anyway.
+     */
+    cctxParams->compressionLevel = compressionLevel;
+}
+
+size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params)
+{
+    RETURN_ERROR_IF(!cctxParams, GENERIC, "NULL pointer!");
+    FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) , "");
+    ZSTD_CCtxParams_init_internal(cctxParams, &params, ZSTD_NO_CLEVEL);
+    return 0;
+}
+
+/*
+ * Sets cctxParams' cParams and fParams from params, but otherwise leaves them alone.
+ * @param param Validated zstd parameters.
+ */
+static void ZSTD_CCtxParams_setZstdParams(
+        ZSTD_CCtx_params* cctxParams, const ZSTD_parameters* params)
+{
+    assert(!ZSTD_checkCParams(params->cParams));
+    cctxParams->cParams = params->cParams;
+    cctxParams->fParams = params->fParams;
+    /* Should not matter, as all cParams are presumed properly defined.
+     * But, set it for tracing anyway.
+     */
+    cctxParams->compressionLevel = ZSTD_NO_CLEVEL;
+}
+
+ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_cParameter param)
+{
+    ZSTD_bounds bounds = { 0, 0, 0 };
+
+    switch(param)
+    {
+    case ZSTD_c_compressionLevel:
+        bounds.lowerBound = ZSTD_minCLevel();
+        bounds.upperBound = ZSTD_maxCLevel();
+        return bounds;
+
+    case ZSTD_c_windowLog:
+        bounds.lowerBound = ZSTD_WINDOWLOG_MIN;
+        bounds.upperBound = ZSTD_WINDOWLOG_MAX;
+        return bounds;
+
+    case ZSTD_c_hashLog:
+        bounds.lowerBound = ZSTD_HASHLOG_MIN;
+        bounds.upperBound = ZSTD_HASHLOG_MAX;
+        return bounds;
+
+    case ZSTD_c_chainLog:
+        bounds.lowerBound = ZSTD_CHAINLOG_MIN;
+        bounds.upperBound = ZSTD_CHAINLOG_MAX;
+        return bounds;
+
+    case ZSTD_c_searchLog:
+        bounds.lowerBound = ZSTD_SEARCHLOG_MIN;
+        bounds.upperBound = ZSTD_SEARCHLOG_MAX;
+        return bounds;
+
+    case ZSTD_c_minMatch:
+        bounds.lowerBound = ZSTD_MINMATCH_MIN;
+        bounds.upperBound = ZSTD_MINMATCH_MAX;
+        return bounds;
+
+    case ZSTD_c_targetLength:
+        bounds.lowerBound = ZSTD_TARGETLENGTH_MIN;
+        bounds.upperBound = ZSTD_TARGETLENGTH_MAX;
+        return bounds;
+
+    case ZSTD_c_strategy:
+        bounds.lowerBound = ZSTD_STRATEGY_MIN;
+        bounds.upperBound = ZSTD_STRATEGY_MAX;
+        return bounds;
+
+    case ZSTD_c_contentSizeFlag:
+        bounds.lowerBound = 0;
+        bounds.upperBound = 1;
+        return bounds;
+
+    case ZSTD_c_checksumFlag:
+        bounds.lowerBound = 0;
+        bounds.upperBound = 1;
+        return bounds;
+
+    case ZSTD_c_dictIDFlag:
+        bounds.lowerBound = 0;
+        bounds.upperBound = 1;
+        return bounds;
+
+    case ZSTD_c_nbWorkers:
+        bounds.lowerBound = 0;
+        bounds.upperBound = 0;
+        return bounds;
+
+    case ZSTD_c_jobSize:
+        bounds.lowerBound = 0;
+        bounds.upperBound = 0;
+        return bounds;
+
+    case ZSTD_c_overlapLog:
+        bounds.lowerBound = 0;
+        bounds.upperBound = 0;
+        return bounds;
+
+    case ZSTD_c_enableDedicatedDictSearch:
+        bounds.lowerBound = 0;
+        bounds.upperBound = 1;
+        return bounds;
+
+    case ZSTD_c_enableLongDistanceMatching:
+        bounds.lowerBound = 0;
+        bounds.upperBound = 1;
+        return bounds;
+
+    case ZSTD_c_ldmHashLog:
+        bounds.lowerBound = ZSTD_LDM_HASHLOG_MIN;
+        bounds.upperBound = ZSTD_LDM_HASHLOG_MAX;
+        return bounds;
+
+    case ZSTD_c_ldmMinMatch:
+        bounds.lowerBound = ZSTD_LDM_MINMATCH_MIN;
+        bounds.upperBound = ZSTD_LDM_MINMATCH_MAX;
+        return bounds;
+
+    case ZSTD_c_ldmBucketSizeLog:
+        bounds.lowerBound = ZSTD_LDM_BUCKETSIZELOG_MIN;
+        bounds.upperBound = ZSTD_LDM_BUCKETSIZELOG_MAX;
+        return bounds;
+
+    case ZSTD_c_ldmHashRateLog:
+        bounds.lowerBound = ZSTD_LDM_HASHRATELOG_MIN;
+        bounds.upperBound = ZSTD_LDM_HASHRATELOG_MAX;
+        return bounds;
+
+    /* experimental parameters */
+    case ZSTD_c_rsyncable:
+        bounds.lowerBound = 0;
+        bounds.upperBound = 1;
+        return bounds;
+
+    case ZSTD_c_forceMaxWindow :
+        bounds.lowerBound = 0;
+        bounds.upperBound = 1;
+        return bounds;
+
+    case ZSTD_c_format:
+        ZSTD_STATIC_ASSERT(ZSTD_f_zstd1 < ZSTD_f_zstd1_magicless);
+        bounds.lowerBound = ZSTD_f_zstd1;
+        bounds.upperBound = ZSTD_f_zstd1_magicless;   /* note : how to ensure at compile time that this is the highest value enum ? */
+        return bounds;
+
+    case ZSTD_c_forceAttachDict:
+        ZSTD_STATIC_ASSERT(ZSTD_dictDefaultAttach < ZSTD_dictForceLoad);
+        bounds.lowerBound = ZSTD_dictDefaultAttach;
+        bounds.upperBound = ZSTD_dictForceLoad;       /* note : how to ensure at compile time that this is the highest value enum ? */
+        return bounds;
+
+    case ZSTD_c_literalCompressionMode:
+        ZSTD_STATIC_ASSERT(ZSTD_lcm_auto < ZSTD_lcm_huffman && ZSTD_lcm_huffman < ZSTD_lcm_uncompressed);
+        bounds.lowerBound = ZSTD_lcm_auto;
+        bounds.upperBound = ZSTD_lcm_uncompressed;
+        return bounds;
+
+    case ZSTD_c_targetCBlockSize:
+        bounds.lowerBound = ZSTD_TARGETCBLOCKSIZE_MIN;
+        bounds.upperBound = ZSTD_TARGETCBLOCKSIZE_MAX;
+        return bounds;
+
+    case ZSTD_c_srcSizeHint:
+        bounds.lowerBound = ZSTD_SRCSIZEHINT_MIN;
+        bounds.upperBound = ZSTD_SRCSIZEHINT_MAX;
+        return bounds;
+
+    case ZSTD_c_stableInBuffer:
+    case ZSTD_c_stableOutBuffer:
+        bounds.lowerBound = (int)ZSTD_bm_buffered;
+        bounds.upperBound = (int)ZSTD_bm_stable;
+        return bounds;
+
+    case ZSTD_c_blockDelimiters:
+        bounds.lowerBound = (int)ZSTD_sf_noBlockDelimiters;
+        bounds.upperBound = (int)ZSTD_sf_explicitBlockDelimiters;
+        return bounds;
+
+    case ZSTD_c_validateSequences:
+        bounds.lowerBound = 0;
+        bounds.upperBound = 1;
+        return bounds;
+
+    default:
+        bounds.error = ERROR(parameter_unsupported);
+        return bounds;
+    }
+}
+
+/* ZSTD_cParam_clampBounds:
+ * Clamps the value into the bounded range.
+ */
+static size_t ZSTD_cParam_clampBounds(ZSTD_cParameter cParam, int* value)
+{
+    ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam);
+    if (ZSTD_isError(bounds.error)) return bounds.error;
+    if (*value < bounds.lowerBound) *value = bounds.lowerBound;
+    if (*value > bounds.upperBound) *value = bounds.upperBound;
+    return 0;
+}
+
+#define BOUNDCHECK(cParam, val) { \
+    RETURN_ERROR_IF(!ZSTD_cParam_withinBounds(cParam,val), \
+                    parameter_outOfBound, "Param out of bounds"); \
+}
+
+
+static int ZSTD_isUpdateAuthorized(ZSTD_cParameter param)
+{
+    switch(param)
+    {
+    case ZSTD_c_compressionLevel:
+    case ZSTD_c_hashLog:
+    case ZSTD_c_chainLog:
+    case ZSTD_c_searchLog:
+    case ZSTD_c_minMatch:
+    case ZSTD_c_targetLength:
+    case ZSTD_c_strategy:
+        return 1;
+
+    case ZSTD_c_format:
+    case ZSTD_c_windowLog:
+    case ZSTD_c_contentSizeFlag:
+    case ZSTD_c_checksumFlag:
+    case ZSTD_c_dictIDFlag:
+    case ZSTD_c_forceMaxWindow :
+    case ZSTD_c_nbWorkers:
+    case ZSTD_c_jobSize:
+    case ZSTD_c_overlapLog:
+    case ZSTD_c_rsyncable:
+    case ZSTD_c_enableDedicatedDictSearch:
+    case ZSTD_c_enableLongDistanceMatching:
+    case ZSTD_c_ldmHashLog:
+    case ZSTD_c_ldmMinMatch:
+    case ZSTD_c_ldmBucketSizeLog:
+    case ZSTD_c_ldmHashRateLog:
+    case ZSTD_c_forceAttachDict:
+    case ZSTD_c_literalCompressionMode:
+    case ZSTD_c_targetCBlockSize:
+    case ZSTD_c_srcSizeHint:
+    case ZSTD_c_stableInBuffer:
+    case ZSTD_c_stableOutBuffer:
+    case ZSTD_c_blockDelimiters:
+    case ZSTD_c_validateSequences:
+    default:
+        return 0;
+    }
+}
+
+size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value)
+{
+    DEBUGLOG(4, "ZSTD_CCtx_setParameter (%i, %i)", (int)param, value);
+    if (cctx->streamStage != zcss_init) {
+        if (ZSTD_isUpdateAuthorized(param)) {
+            cctx->cParamsChanged = 1;
+        } else {
+            RETURN_ERROR(stage_wrong, "can only set params in ctx init stage");
+    }   }
+
+    switch(param)
+    {
+    case ZSTD_c_nbWorkers:
+        RETURN_ERROR_IF((value!=0) && cctx->staticSize, parameter_unsupported,
+                        "MT not compatible with static alloc");
+        break;
+
+    case ZSTD_c_compressionLevel:
+    case ZSTD_c_windowLog:
+    case ZSTD_c_hashLog:
+    case ZSTD_c_chainLog:
+    case ZSTD_c_searchLog:
+    case ZSTD_c_minMatch:
+    case ZSTD_c_targetLength:
+    case ZSTD_c_strategy:
+    case ZSTD_c_ldmHashRateLog:
+    case ZSTD_c_format:
+    case ZSTD_c_contentSizeFlag:
+    case ZSTD_c_checksumFlag:
+    case ZSTD_c_dictIDFlag:
+    case ZSTD_c_forceMaxWindow:
+    case ZSTD_c_forceAttachDict:
+    case ZSTD_c_literalCompressionMode:
+    case ZSTD_c_jobSize:
+    case ZSTD_c_overlapLog:
+    case ZSTD_c_rsyncable:
+    case ZSTD_c_enableDedicatedDictSearch:
+    case ZSTD_c_enableLongDistanceMatching:
+    case ZSTD_c_ldmHashLog:
+    case ZSTD_c_ldmMinMatch:
+    case ZSTD_c_ldmBucketSizeLog:
+    case ZSTD_c_targetCBlockSize:
+    case ZSTD_c_srcSizeHint:
+    case ZSTD_c_stableInBuffer:
+    case ZSTD_c_stableOutBuffer:
+    case ZSTD_c_blockDelimiters:
+    case ZSTD_c_validateSequences:
+        break;
+
+    default: RETURN_ERROR(parameter_unsupported, "unknown parameter");
+    }
+    return ZSTD_CCtxParams_setParameter(&cctx->requestedParams, param, value);
+}
+
+size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
+                                    ZSTD_cParameter param, int value)
+{
+    DEBUGLOG(4, "ZSTD_CCtxParams_setParameter (%i, %i)", (int)param, value);
+    switch(param)
+    {
+    case ZSTD_c_format :
+        BOUNDCHECK(ZSTD_c_format, value);
+        CCtxParams->format = (ZSTD_format_e)value;
+        return (size_t)CCtxParams->format;
+
+    case ZSTD_c_compressionLevel : {
+        FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value), "");
+        if (value == 0)
+            CCtxParams->compressionLevel = ZSTD_CLEVEL_DEFAULT; /* 0 == default */
+        else
+            CCtxParams->compressionLevel = value;
+        if (CCtxParams->compressionLevel >= 0) return (size_t)CCtxParams->compressionLevel;
+        return 0;  /* return type (size_t) cannot represent negative values */
+    }
+
+    case ZSTD_c_windowLog :
+        if (value!=0)   /* 0 => use default */
+            BOUNDCHECK(ZSTD_c_windowLog, value);
+        CCtxParams->cParams.windowLog = (U32)value;
+        return CCtxParams->cParams.windowLog;
+
+    case ZSTD_c_hashLog :
+        if (value!=0)   /* 0 => use default */
+            BOUNDCHECK(ZSTD_c_hashLog, value);
+        CCtxParams->cParams.hashLog = (U32)value;
+        return CCtxParams->cParams.hashLog;
+
+    case ZSTD_c_chainLog :
+        if (value!=0)   /* 0 => use default */
+            BOUNDCHECK(ZSTD_c_chainLog, value);
+        CCtxParams->cParams.chainLog = (U32)value;
+        return CCtxParams->cParams.chainLog;
+
+    case ZSTD_c_searchLog :
+        if (value!=0)   /* 0 => use default */
+            BOUNDCHECK(ZSTD_c_searchLog, value);
+        CCtxParams->cParams.searchLog = (U32)value;
+        return (size_t)value;
+
+    case ZSTD_c_minMatch :
+        if (value!=0)   /* 0 => use default */
+            BOUNDCHECK(ZSTD_c_minMatch, value);
+        CCtxParams->cParams.minMatch = value;
+        return CCtxParams->cParams.minMatch;
+
+    case ZSTD_c_targetLength :
+        BOUNDCHECK(ZSTD_c_targetLength, value);
+        CCtxParams->cParams.targetLength = value;
+        return CCtxParams->cParams.targetLength;
+
+    case ZSTD_c_strategy :
+        if (value!=0)   /* 0 => use default */
+            BOUNDCHECK(ZSTD_c_strategy, value);
+        CCtxParams->cParams.strategy = (ZSTD_strategy)value;
+        return (size_t)CCtxParams->cParams.strategy;
+
+    case ZSTD_c_contentSizeFlag :
+        /* Content size written in frame header _when known_ (default:1) */
+        DEBUGLOG(4, "set content size flag = %u", (value!=0));
+        CCtxParams->fParams.contentSizeFlag = value != 0;
+        return CCtxParams->fParams.contentSizeFlag;
+
+    case ZSTD_c_checksumFlag :
+        /* A 32-bits content checksum will be calculated and written at end of frame (default:0) */
+        CCtxParams->fParams.checksumFlag = value != 0;
+        return CCtxParams->fParams.checksumFlag;
+
+    case ZSTD_c_dictIDFlag : /* When applicable, dictionary's dictID is provided in frame header (default:1) */
+        DEBUGLOG(4, "set dictIDFlag = %u", (value!=0));
+        CCtxParams->fParams.noDictIDFlag = !value;
+        return !CCtxParams->fParams.noDictIDFlag;
+
+    case ZSTD_c_forceMaxWindow :
+        CCtxParams->forceWindow = (value != 0);
+        return CCtxParams->forceWindow;
+
+    case ZSTD_c_forceAttachDict : {
+        const ZSTD_dictAttachPref_e pref = (ZSTD_dictAttachPref_e)value;
+        BOUNDCHECK(ZSTD_c_forceAttachDict, pref);
+        CCtxParams->attachDictPref = pref;
+        return CCtxParams->attachDictPref;
+    }
+
+    case ZSTD_c_literalCompressionMode : {
+        const ZSTD_literalCompressionMode_e lcm = (ZSTD_literalCompressionMode_e)value;
+        BOUNDCHECK(ZSTD_c_literalCompressionMode, lcm);
+        CCtxParams->literalCompressionMode = lcm;
+        return CCtxParams->literalCompressionMode;
+    }
+
+    case ZSTD_c_nbWorkers :
+        RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
+        return 0;
+
+    case ZSTD_c_jobSize :
+        RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
+        return 0;
+
+    case ZSTD_c_overlapLog :
+        RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
+        return 0;
+
+    case ZSTD_c_rsyncable :
+        RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
+        return 0;
+
+    case ZSTD_c_enableDedicatedDictSearch :
+        CCtxParams->enableDedicatedDictSearch = (value!=0);
+        return CCtxParams->enableDedicatedDictSearch;
+
+    case ZSTD_c_enableLongDistanceMatching :
+        CCtxParams->ldmParams.enableLdm = (value!=0);
+        return CCtxParams->ldmParams.enableLdm;
+
+    case ZSTD_c_ldmHashLog :
+        if (value!=0)   /* 0 ==> auto */
+            BOUNDCHECK(ZSTD_c_ldmHashLog, value);
+        CCtxParams->ldmParams.hashLog = value;
+        return CCtxParams->ldmParams.hashLog;
+
+    case ZSTD_c_ldmMinMatch :
+        if (value!=0)   /* 0 ==> default */
+            BOUNDCHECK(ZSTD_c_ldmMinMatch, value);
+        CCtxParams->ldmParams.minMatchLength = value;
+        return CCtxParams->ldmParams.minMatchLength;
+
+    case ZSTD_c_ldmBucketSizeLog :
+        if (value!=0)   /* 0 ==> default */
+            BOUNDCHECK(ZSTD_c_ldmBucketSizeLog, value);
+        CCtxParams->ldmParams.bucketSizeLog = value;
+        return CCtxParams->ldmParams.bucketSizeLog;
+
+    case ZSTD_c_ldmHashRateLog :
+        if (value!=0)   /* 0 ==> default */
+            BOUNDCHECK(ZSTD_c_ldmHashRateLog, value);
+        CCtxParams->ldmParams.hashRateLog = value;
+        return CCtxParams->ldmParams.hashRateLog;
+
+    case ZSTD_c_targetCBlockSize :
+        if (value!=0)   /* 0 ==> default */
+            BOUNDCHECK(ZSTD_c_targetCBlockSize, value);
+        CCtxParams->targetCBlockSize = value;
+        return CCtxParams->targetCBlockSize;
+
+    case ZSTD_c_srcSizeHint :
+        if (value!=0)    /* 0 ==> default */
+            BOUNDCHECK(ZSTD_c_srcSizeHint, value);
+        CCtxParams->srcSizeHint = value;
+        return CCtxParams->srcSizeHint;
+
+    case ZSTD_c_stableInBuffer:
+        BOUNDCHECK(ZSTD_c_stableInBuffer, value);
+        CCtxParams->inBufferMode = (ZSTD_bufferMode_e)value;
+        return CCtxParams->inBufferMode;
+
+    case ZSTD_c_stableOutBuffer:
+        BOUNDCHECK(ZSTD_c_stableOutBuffer, value);
+        CCtxParams->outBufferMode = (ZSTD_bufferMode_e)value;
+        return CCtxParams->outBufferMode;
+
+    case ZSTD_c_blockDelimiters:
+        BOUNDCHECK(ZSTD_c_blockDelimiters, value);
+        CCtxParams->blockDelimiters = (ZSTD_sequenceFormat_e)value;
+        return CCtxParams->blockDelimiters;
+
+    case ZSTD_c_validateSequences:
+        BOUNDCHECK(ZSTD_c_validateSequences, value);
+        CCtxParams->validateSequences = value;
+        return CCtxParams->validateSequences;
+
+    default: RETURN_ERROR(parameter_unsupported, "unknown parameter");
+    }
+}
+
+size_t ZSTD_CCtx_getParameter(ZSTD_CCtx const* cctx, ZSTD_cParameter param, int* value)
+{
+    return ZSTD_CCtxParams_getParameter(&cctx->requestedParams, param, value);
+}
+
+size_t ZSTD_CCtxParams_getParameter(
+        ZSTD_CCtx_params const* CCtxParams, ZSTD_cParameter param, int* value)
+{
+    switch(param)
+    {
+    case ZSTD_c_format :
+        *value = CCtxParams->format;
+        break;
+    case ZSTD_c_compressionLevel :
+        *value = CCtxParams->compressionLevel;
+        break;
+    case ZSTD_c_windowLog :
+        *value = (int)CCtxParams->cParams.windowLog;
+        break;
+    case ZSTD_c_hashLog :
+        *value = (int)CCtxParams->cParams.hashLog;
+        break;
+    case ZSTD_c_chainLog :
+        *value = (int)CCtxParams->cParams.chainLog;
+        break;
+    case ZSTD_c_searchLog :
+        *value = CCtxParams->cParams.searchLog;
+        break;
+    case ZSTD_c_minMatch :
+        *value = CCtxParams->cParams.minMatch;
+        break;
+    case ZSTD_c_targetLength :
+        *value = CCtxParams->cParams.targetLength;
+        break;
+    case ZSTD_c_strategy :
+        *value = (unsigned)CCtxParams->cParams.strategy;
+        break;
+    case ZSTD_c_contentSizeFlag :
+        *value = CCtxParams->fParams.contentSizeFlag;
+        break;
+    case ZSTD_c_checksumFlag :
+        *value = CCtxParams->fParams.checksumFlag;
+        break;
+    case ZSTD_c_dictIDFlag :
+        *value = !CCtxParams->fParams.noDictIDFlag;
+        break;
+    case ZSTD_c_forceMaxWindow :
+        *value = CCtxParams->forceWindow;
+        break;
+    case ZSTD_c_forceAttachDict :
+        *value = CCtxParams->attachDictPref;
+        break;
+    case ZSTD_c_literalCompressionMode :
+        *value = CCtxParams->literalCompressionMode;
+        break;
+    case ZSTD_c_nbWorkers :
+        assert(CCtxParams->nbWorkers == 0);
+        *value = CCtxParams->nbWorkers;
+        break;
+    case ZSTD_c_jobSize :
+        RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");
+    case ZSTD_c_overlapLog :
+        RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");
+    case ZSTD_c_rsyncable :
+        RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");
+    case ZSTD_c_enableDedicatedDictSearch :
+        *value = CCtxParams->enableDedicatedDictSearch;
+        break;
+    case ZSTD_c_enableLongDistanceMatching :
+        *value = CCtxParams->ldmParams.enableLdm;
+        break;
+    case ZSTD_c_ldmHashLog :
+        *value = CCtxParams->ldmParams.hashLog;
+        break;
+    case ZSTD_c_ldmMinMatch :
+        *value = CCtxParams->ldmParams.minMatchLength;
+        break;
+    case ZSTD_c_ldmBucketSizeLog :
+        *value = CCtxParams->ldmParams.bucketSizeLog;
+        break;
+    case ZSTD_c_ldmHashRateLog :
+        *value = CCtxParams->ldmParams.hashRateLog;
+        break;
+    case ZSTD_c_targetCBlockSize :
+        *value = (int)CCtxParams->targetCBlockSize;
+        break;
+    case ZSTD_c_srcSizeHint :
+        *value = (int)CCtxParams->srcSizeHint;
+        break;
+    case ZSTD_c_stableInBuffer :
+        *value = (int)CCtxParams->inBufferMode;
+        break;
+    case ZSTD_c_stableOutBuffer :
+        *value = (int)CCtxParams->outBufferMode;
+        break;
+    case ZSTD_c_blockDelimiters :
+        *value = (int)CCtxParams->blockDelimiters;
+        break;
+    case ZSTD_c_validateSequences :
+        *value = (int)CCtxParams->validateSequences;
+        break;
+    default: RETURN_ERROR(parameter_unsupported, "unknown parameter");
+    }
+    return 0;
+}
+
+/* ZSTD_CCtx_setParametersUsingCCtxParams() :
+ *  just applies `params` into `cctx`
+ *  no action is performed, parameters are merely stored.
+ *  If ZSTDMT is enabled, parameters are pushed to cctx->mtctx.
+ *    This is possible even if a compression is ongoing.
+ *    In which case, new parameters will be applied on the fly, starting with next compression job.
+ */
+size_t ZSTD_CCtx_setParametersUsingCCtxParams(
+        ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params)
+{
+    DEBUGLOG(4, "ZSTD_CCtx_setParametersUsingCCtxParams");
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
+                    "The context is in the wrong stage!");
+    RETURN_ERROR_IF(cctx->cdict, stage_wrong,
+                    "Can't override parameters with cdict attached (some must "
+                    "be inherited from the cdict).");
+
+    cctx->requestedParams = *params;
+    return 0;
+}
+
+ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize)
+{
+    DEBUGLOG(4, "ZSTD_CCtx_setPledgedSrcSize to %u bytes", (U32)pledgedSrcSize);
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
+                    "Can't set pledgedSrcSize when not in init stage.");
+    cctx->pledgedSrcSizePlusOne = pledgedSrcSize+1;
+    return 0;
+}
+
+static ZSTD_compressionParameters ZSTD_dedicatedDictSearch_getCParams(
+        int const compressionLevel,
+        size_t const dictSize);
+static int ZSTD_dedicatedDictSearch_isSupported(
+        const ZSTD_compressionParameters* cParams);
+static void ZSTD_dedicatedDictSearch_revertCParams(
+        ZSTD_compressionParameters* cParams);
+
+/*
+ * Initializes the local dict using the requested parameters.
+ * NOTE: This does not use the pledged src size, because it may be used for more
+ * than one compression.
+ */
+static size_t ZSTD_initLocalDict(ZSTD_CCtx* cctx)
+{
+    ZSTD_localDict* const dl = &cctx->localDict;
+    if (dl->dict == NULL) {
+        /* No local dictionary. */
+        assert(dl->dictBuffer == NULL);
+        assert(dl->cdict == NULL);
+        assert(dl->dictSize == 0);
+        return 0;
+    }
+    if (dl->cdict != NULL) {
+        assert(cctx->cdict == dl->cdict);
+        /* Local dictionary already initialized. */
+        return 0;
+    }
+    assert(dl->dictSize > 0);
+    assert(cctx->cdict == NULL);
+    assert(cctx->prefixDict.dict == NULL);
+
+    dl->cdict = ZSTD_createCDict_advanced2(
+            dl->dict,
+            dl->dictSize,
+            ZSTD_dlm_byRef,
+            dl->dictContentType,
+            &cctx->requestedParams,
+            cctx->customMem);
+    RETURN_ERROR_IF(!dl->cdict, memory_allocation, "ZSTD_createCDict_advanced failed");
+    cctx->cdict = dl->cdict;
+    return 0;
+}
+
+size_t ZSTD_CCtx_loadDictionary_advanced(
+        ZSTD_CCtx* cctx, const void* dict, size_t dictSize,
+        ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType)
+{
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
+                    "Can't load a dictionary when ctx is not in init stage.");
+    DEBUGLOG(4, "ZSTD_CCtx_loadDictionary_advanced (size: %u)", (U32)dictSize);
+    ZSTD_clearAllDicts(cctx);  /* in case one already exists */
+    if (dict == NULL || dictSize == 0)  /* no dictionary mode */
+        return 0;
+    if (dictLoadMethod == ZSTD_dlm_byRef) {
+        cctx->localDict.dict = dict;
+    } else {
+        void* dictBuffer;
+        RETURN_ERROR_IF(cctx->staticSize, memory_allocation,
+                        "no malloc for static CCtx");
+        dictBuffer = ZSTD_customMalloc(dictSize, cctx->customMem);
+        RETURN_ERROR_IF(!dictBuffer, memory_allocation, "NULL pointer!");
+        ZSTD_memcpy(dictBuffer, dict, dictSize);
+        cctx->localDict.dictBuffer = dictBuffer;
+        cctx->localDict.dict = dictBuffer;
+    }
+    cctx->localDict.dictSize = dictSize;
+    cctx->localDict.dictContentType = dictContentType;
+    return 0;
+}
+
+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_byReference(
+      ZSTD_CCtx* cctx, const void* dict, size_t dictSize)
+{
+    return ZSTD_CCtx_loadDictionary_advanced(
+            cctx, dict, dictSize, ZSTD_dlm_byRef, ZSTD_dct_auto);
+}
+
+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize)
+{
+    return ZSTD_CCtx_loadDictionary_advanced(
+            cctx, dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dct_auto);
+}
+
+
+size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict)
+{
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
+                    "Can't ref a dict when ctx not in init stage.");
+    /* Free the existing local cdict (if any) to save memory. */
+    ZSTD_clearAllDicts(cctx);
+    cctx->cdict = cdict;
+    return 0;
+}
+
+size_t ZSTD_CCtx_refThreadPool(ZSTD_CCtx* cctx, ZSTD_threadPool* pool)
+{
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
+                    "Can't ref a pool when ctx not in init stage.");
+    cctx->pool = pool;
+    return 0;
+}
+
+size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize)
+{
+    return ZSTD_CCtx_refPrefix_advanced(cctx, prefix, prefixSize, ZSTD_dct_rawContent);
+}
+
+size_t ZSTD_CCtx_refPrefix_advanced(
+        ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType)
+{
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
+                    "Can't ref a prefix when ctx not in init stage.");
+    ZSTD_clearAllDicts(cctx);
+    if (prefix != NULL && prefixSize > 0) {
+        cctx->prefixDict.dict = prefix;
+        cctx->prefixDict.dictSize = prefixSize;
+        cctx->prefixDict.dictContentType = dictContentType;
+    }
+    return 0;
+}
+
+/*! ZSTD_CCtx_reset() :
+ *  Also dumps dictionary */
+size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset)
+{
+    if ( (reset == ZSTD_reset_session_only)
+      || (reset == ZSTD_reset_session_and_parameters) ) {
+        cctx->streamStage = zcss_init;
+        cctx->pledgedSrcSizePlusOne = 0;
+    }
+    if ( (reset == ZSTD_reset_parameters)
+      || (reset == ZSTD_reset_session_and_parameters) ) {
+        RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
+                        "Can't reset parameters only when not in init stage.");
+        ZSTD_clearAllDicts(cctx);
+        return ZSTD_CCtxParams_reset(&cctx->requestedParams);
+    }
+    return 0;
+}
+
+
+/* ZSTD_checkCParams() :
+    control CParam values remain within authorized range.
+    @return : 0, or an error code if one value is beyond authorized range */
+size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams)
+{
+    BOUNDCHECK(ZSTD_c_windowLog, (int)cParams.windowLog);
+    BOUNDCHECK(ZSTD_c_chainLog,  (int)cParams.chainLog);
+    BOUNDCHECK(ZSTD_c_hashLog,   (int)cParams.hashLog);
+    BOUNDCHECK(ZSTD_c_searchLog, (int)cParams.searchLog);
+    BOUNDCHECK(ZSTD_c_minMatch,  (int)cParams.minMatch);
+    BOUNDCHECK(ZSTD_c_targetLength,(int)cParams.targetLength);
+    BOUNDCHECK(ZSTD_c_strategy,  cParams.strategy);
+    return 0;
+}
+
+/* ZSTD_clampCParams() :
+ *  make CParam values within valid range.
+ *  @return : valid CParams */
+static ZSTD_compressionParameters
+ZSTD_clampCParams(ZSTD_compressionParameters cParams)
+{
+#   define CLAMP_TYPE(cParam, val, type) {                                \
+        ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam);         \
+        if ((int)val<bounds.lowerBound) val=(type)bounds.lowerBound;      \
+        else if ((int)val>bounds.upperBound) val=(type)bounds.upperBound; \
+    }
+#   define CLAMP(cParam, val) CLAMP_TYPE(cParam, val, unsigned)
+    CLAMP(ZSTD_c_windowLog, cParams.windowLog);
+    CLAMP(ZSTD_c_chainLog,  cParams.chainLog);
+    CLAMP(ZSTD_c_hashLog,   cParams.hashLog);
+    CLAMP(ZSTD_c_searchLog, cParams.searchLog);
+    CLAMP(ZSTD_c_minMatch,  cParams.minMatch);
+    CLAMP(ZSTD_c_targetLength,cParams.targetLength);
+    CLAMP_TYPE(ZSTD_c_strategy,cParams.strategy, ZSTD_strategy);
+    return cParams;
+}
+
+/* ZSTD_cycleLog() :
+ *  condition for correct operation : hashLog > 1 */
+U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat)
+{
+    U32 const btScale = ((U32)strat >= (U32)ZSTD_btlazy2);
+    return hashLog - btScale;
+}
+
+/* ZSTD_dictAndWindowLog() :
+ * Returns an adjusted window log that is large enough to fit the source and the dictionary.
+ * The zstd format says that the entire dictionary is valid if one byte of the dictionary
+ * is within the window. So the hashLog and chainLog should be large enough to reference both
+ * the dictionary and the window. So we must use this adjusted dictAndWindowLog when downsizing
+ * the hashLog and windowLog.
+ * NOTE: srcSize must not be ZSTD_CONTENTSIZE_UNKNOWN.
+ */
+static U32 ZSTD_dictAndWindowLog(U32 windowLog, U64 srcSize, U64 dictSize)
+{
+    const U64 maxWindowSize = 1ULL << ZSTD_WINDOWLOG_MAX;
+    /* No dictionary ==> No change */
+    if (dictSize == 0) {
+        return windowLog;
+    }
+    assert(windowLog <= ZSTD_WINDOWLOG_MAX);
+    assert(srcSize != ZSTD_CONTENTSIZE_UNKNOWN); /* Handled in ZSTD_adjustCParams_internal() */
+    {
+        U64 const windowSize = 1ULL << windowLog;
+        U64 const dictAndWindowSize = dictSize + windowSize;
+        /* If the window size is already large enough to fit both the source and the dictionary
+         * then just use the window size. Otherwise adjust so that it fits the dictionary and
+         * the window.
+         */
+        if (windowSize >= dictSize + srcSize) {
+            return windowLog; /* Window size large enough already */
+        } else if (dictAndWindowSize >= maxWindowSize) {
+            return ZSTD_WINDOWLOG_MAX; /* Larger than max window log */
+        } else  {
+            return ZSTD_highbit32((U32)dictAndWindowSize - 1) + 1;
+        }
+    }
+}
+
+/* ZSTD_adjustCParams_internal() :
+ *  optimize `cPar` for a specified input (`srcSize` and `dictSize`).
+ *  mostly downsize to reduce memory consumption and initialization latency.
+ * `srcSize` can be ZSTD_CONTENTSIZE_UNKNOWN when not known.
+ * `mode` is the mode for parameter adjustment. See docs for `ZSTD_cParamMode_e`.
+ *  note : `srcSize==0` means 0!
+ *  condition : cPar is presumed validated (can be checked using ZSTD_checkCParams()). */
+static ZSTD_compressionParameters
+ZSTD_adjustCParams_internal(ZSTD_compressionParameters cPar,
+                            unsigned long long srcSize,
+                            size_t dictSize,
+                            ZSTD_cParamMode_e mode)
+{
+    const U64 minSrcSize = 513; /* (1<<9) + 1 */
+    const U64 maxWindowResize = 1ULL << (ZSTD_WINDOWLOG_MAX-1);
+    assert(ZSTD_checkCParams(cPar)==0);
+
+    switch (mode) {
+    case ZSTD_cpm_unknown:
+    case ZSTD_cpm_noAttachDict:
+        /* If we don't know the source size, don't make any
+         * assumptions about it. We will already have selected
+         * smaller parameters if a dictionary is in use.
+         */
+        break;
+    case ZSTD_cpm_createCDict:
+        /* Assume a small source size when creating a dictionary
+         * with an unkown source size.
+         */
+        if (dictSize && srcSize == ZSTD_CONTENTSIZE_UNKNOWN)
+            srcSize = minSrcSize;
+        break;
+    case ZSTD_cpm_attachDict:
+        /* Dictionary has its own dedicated parameters which have
+         * already been selected. We are selecting parameters
+         * for only the source.
+         */
+        dictSize = 0;
+        break;
+    default:
+        assert(0);
+        break;
+    }
+
+    /* resize windowLog if input is small enough, to use less memory */
+    if ( (srcSize < maxWindowResize)
+      && (dictSize < maxWindowResize) )  {
+        U32 const tSize = (U32)(srcSize + dictSize);
+        static U32 const hashSizeMin = 1 << ZSTD_HASHLOG_MIN;
+        U32 const srcLog = (tSize < hashSizeMin) ? ZSTD_HASHLOG_MIN :
+                            ZSTD_highbit32(tSize-1) + 1;
+        if (cPar.windowLog > srcLog) cPar.windowLog = srcLog;
+    }
+    if (srcSize != ZSTD_CONTENTSIZE_UNKNOWN) {
+        U32 const dictAndWindowLog = ZSTD_dictAndWindowLog(cPar.windowLog, (U64)srcSize, (U64)dictSize);
+        U32 const cycleLog = ZSTD_cycleLog(cPar.chainLog, cPar.strategy);
+        if (cPar.hashLog > dictAndWindowLog+1) cPar.hashLog = dictAndWindowLog+1;
+        if (cycleLog > dictAndWindowLog)
+            cPar.chainLog -= (cycleLog - dictAndWindowLog);
+    }
+
+    if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN)
+        cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN;  /* minimum wlog required for valid frame header */
+
+    return cPar;
+}
+
+ZSTD_compressionParameters
+ZSTD_adjustCParams(ZSTD_compressionParameters cPar,
+                   unsigned long long srcSize,
+                   size_t dictSize)
+{
+    cPar = ZSTD_clampCParams(cPar);   /* resulting cPar is necessarily valid (all parameters within range) */
+    if (srcSize == 0) srcSize = ZSTD_CONTENTSIZE_UNKNOWN;
+    return ZSTD_adjustCParams_internal(cPar, srcSize, dictSize, ZSTD_cpm_unknown);
+}
+
+static ZSTD_compressionParameters ZSTD_getCParams_internal(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize, ZSTD_cParamMode_e mode);
+static ZSTD_parameters ZSTD_getParams_internal(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize, ZSTD_cParamMode_e mode);
+
+static void ZSTD_overrideCParams(
+              ZSTD_compressionParameters* cParams,
+        const ZSTD_compressionParameters* overrides)
+{
+    if (overrides->windowLog)    cParams->windowLog    = overrides->windowLog;
+    if (overrides->hashLog)      cParams->hashLog      = overrides->hashLog;
+    if (overrides->chainLog)     cParams->chainLog     = overrides->chainLog;
+    if (overrides->searchLog)    cParams->searchLog    = overrides->searchLog;
+    if (overrides->minMatch)     cParams->minMatch     = overrides->minMatch;
+    if (overrides->targetLength) cParams->targetLength = overrides->targetLength;
+    if (overrides->strategy)     cParams->strategy     = overrides->strategy;
+}
+
+ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams(
+        const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize, ZSTD_cParamMode_e mode)
+{
+    ZSTD_compressionParameters cParams;
+    if (srcSizeHint == ZSTD_CONTENTSIZE_UNKNOWN && CCtxParams->srcSizeHint > 0) {
+      srcSizeHint = CCtxParams->srcSizeHint;
+    }
+    cParams = ZSTD_getCParams_internal(CCtxParams->compressionLevel, srcSizeHint, dictSize, mode);
+    if (CCtxParams->ldmParams.enableLdm) cParams.windowLog = ZSTD_LDM_DEFAULT_WINDOW_LOG;
+    ZSTD_overrideCParams(&cParams, &CCtxParams->cParams);
+    assert(!ZSTD_checkCParams(cParams));
+    /* srcSizeHint == 0 means 0 */
+    return ZSTD_adjustCParams_internal(cParams, srcSizeHint, dictSize, mode);
+}
+
+static size_t
+ZSTD_sizeof_matchState(const ZSTD_compressionParameters* const cParams,
+                       const U32 forCCtx)
+{
+    size_t const chainSize = (cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cParams->chainLog);
+    size_t const hSize = ((size_t)1) << cParams->hashLog;
+    U32    const hashLog3 = (forCCtx && cParams->minMatch==3) ? MIN(ZSTD_HASHLOG3_MAX, cParams->windowLog) : 0;
+    size_t const h3Size = hashLog3 ? ((size_t)1) << hashLog3 : 0;
+    /* We don't use ZSTD_cwksp_alloc_size() here because the tables aren't
+     * surrounded by redzones in ASAN. */
+    size_t const tableSpace = chainSize * sizeof(U32)
+                            + hSize * sizeof(U32)
+                            + h3Size * sizeof(U32);
+    size_t const optPotentialSpace =
+        ZSTD_cwksp_alloc_size((MaxML+1) * sizeof(U32))
+      + ZSTD_cwksp_alloc_size((MaxLL+1) * sizeof(U32))
+      + ZSTD_cwksp_alloc_size((MaxOff+1) * sizeof(U32))
+      + ZSTD_cwksp_alloc_size((1<<Litbits) * sizeof(U32))
+      + ZSTD_cwksp_alloc_size((ZSTD_OPT_NUM+1) * sizeof(ZSTD_match_t))
+      + ZSTD_cwksp_alloc_size((ZSTD_OPT_NUM+1) * sizeof(ZSTD_optimal_t));
+    size_t const optSpace = (forCCtx && (cParams->strategy >= ZSTD_btopt))
+                                ? optPotentialSpace
+                                : 0;
+    DEBUGLOG(4, "chainSize: %u - hSize: %u - h3Size: %u",
+                (U32)chainSize, (U32)hSize, (U32)h3Size);
+    return tableSpace + optSpace;
+}
+
+static size_t ZSTD_estimateCCtxSize_usingCCtxParams_internal(
+        const ZSTD_compressionParameters* cParams,
+        const ldmParams_t* ldmParams,
+        const int isStatic,
+        const size_t buffInSize,
+        const size_t buffOutSize,
+        const U64 pledgedSrcSize)
+{
+    size_t const windowSize = MAX(1, (size_t)MIN(((U64)1 << cParams->windowLog), pledgedSrcSize));
+    size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, windowSize);
+    U32    const divider = (cParams->minMatch==3) ? 3 : 4;
+    size_t const maxNbSeq = blockSize / divider;
+    size_t const tokenSpace = ZSTD_cwksp_alloc_size(WILDCOPY_OVERLENGTH + blockSize)
+                            + ZSTD_cwksp_alloc_size(maxNbSeq * sizeof(seqDef))
+                            + 3 * ZSTD_cwksp_alloc_size(maxNbSeq * sizeof(BYTE));
+    size_t const entropySpace = ZSTD_cwksp_alloc_size(ENTROPY_WORKSPACE_SIZE);
+    size_t const blockStateSpace = 2 * ZSTD_cwksp_alloc_size(sizeof(ZSTD_compressedBlockState_t));
+    size_t const matchStateSize = ZSTD_sizeof_matchState(cParams, /* forCCtx */ 1);
+
+    size_t const ldmSpace = ZSTD_ldm_getTableSize(*ldmParams);
+    size_t const maxNbLdmSeq = ZSTD_ldm_getMaxNbSeq(*ldmParams, blockSize);
+    size_t const ldmSeqSpace = ldmParams->enableLdm ?
+        ZSTD_cwksp_alloc_size(maxNbLdmSeq * sizeof(rawSeq)) : 0;
+
+
+    size_t const bufferSpace = ZSTD_cwksp_alloc_size(buffInSize)
+                             + ZSTD_cwksp_alloc_size(buffOutSize);
+
+    size_t const cctxSpace = isStatic ? ZSTD_cwksp_alloc_size(sizeof(ZSTD_CCtx)) : 0;
+
+    size_t const neededSpace =
+        cctxSpace +
+        entropySpace +
+        blockStateSpace +
+        ldmSpace +
+        ldmSeqSpace +
+        matchStateSize +
+        tokenSpace +
+        bufferSpace;
+
+    DEBUGLOG(5, "estimate workspace : %u", (U32)neededSpace);
+    return neededSpace;
+}
+
+size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params)
+{
+    ZSTD_compressionParameters const cParams =
+                ZSTD_getCParamsFromCCtxParams(params, ZSTD_CONTENTSIZE_UNKNOWN, 0, ZSTD_cpm_noAttachDict);
+
+    RETURN_ERROR_IF(params->nbWorkers > 0, GENERIC, "Estimate CCtx size is supported for single-threaded compression only.");
+    /* estimateCCtxSize is for one-shot compression. So no buffers should
+     * be needed. However, we still allocate two 0-sized buffers, which can
+     * take space under ASAN. */
+    return ZSTD_estimateCCtxSize_usingCCtxParams_internal(
+        &cParams, &params->ldmParams, 1, 0, 0, ZSTD_CONTENTSIZE_UNKNOWN);
+}
+
+size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams)
+{
+    ZSTD_CCtx_params const params = ZSTD_makeCCtxParamsFromCParams(cParams);
+    return ZSTD_estimateCCtxSize_usingCCtxParams(&params);
+}
+
+static size_t ZSTD_estimateCCtxSize_internal(int compressionLevel)
+{
+    int tier = 0;
+    size_t largestSize = 0;
+    static const unsigned long long srcSizeTiers[4] = {16 KB, 128 KB, 256 KB, ZSTD_CONTENTSIZE_UNKNOWN};
+    for (; tier < 4; ++tier) {
+        /* Choose the set of cParams for a given level across all srcSizes that give the largest cctxSize */
+        ZSTD_compressionParameters const cParams = ZSTD_getCParams_internal(compressionLevel, srcSizeTiers[tier], 0, ZSTD_cpm_noAttachDict);
+        largestSize = MAX(ZSTD_estimateCCtxSize_usingCParams(cParams), largestSize);
+    }
+    return largestSize;
+}
+
+size_t ZSTD_estimateCCtxSize(int compressionLevel)
+{
+    int level;
+    size_t memBudget = 0;
+    for (level=MIN(compressionLevel, 1); level<=compressionLevel; level++) {
+        /* Ensure monotonically increasing memory usage as compression level increases */
+        size_t const newMB = ZSTD_estimateCCtxSize_internal(level);
+        if (newMB > memBudget) memBudget = newMB;
+    }
+    return memBudget;
+}
+
+size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params)
+{
+    RETURN_ERROR_IF(params->nbWorkers > 0, GENERIC, "Estimate CCtx size is supported for single-threaded compression only.");
+    {   ZSTD_compressionParameters const cParams =
+                ZSTD_getCParamsFromCCtxParams(params, ZSTD_CONTENTSIZE_UNKNOWN, 0, ZSTD_cpm_noAttachDict);
+        size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog);
+        size_t const inBuffSize = (params->inBufferMode == ZSTD_bm_buffered)
+                ? ((size_t)1 << cParams.windowLog) + blockSize
+                : 0;
+        size_t const outBuffSize = (params->outBufferMode == ZSTD_bm_buffered)
+                ? ZSTD_compressBound(blockSize) + 1
+                : 0;
+
+        return ZSTD_estimateCCtxSize_usingCCtxParams_internal(
+            &cParams, &params->ldmParams, 1, inBuffSize, outBuffSize,
+            ZSTD_CONTENTSIZE_UNKNOWN);
+    }
+}
+
+size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams)
+{
+    ZSTD_CCtx_params const params = ZSTD_makeCCtxParamsFromCParams(cParams);
+    return ZSTD_estimateCStreamSize_usingCCtxParams(&params);
+}
+
+static size_t ZSTD_estimateCStreamSize_internal(int compressionLevel)
+{
+    ZSTD_compressionParameters const cParams = ZSTD_getCParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, 0, ZSTD_cpm_noAttachDict);
+    return ZSTD_estimateCStreamSize_usingCParams(cParams);
+}
+
+size_t ZSTD_estimateCStreamSize(int compressionLevel)
+{
+    int level;
+    size_t memBudget = 0;
+    for (level=MIN(compressionLevel, 1); level<=compressionLevel; level++) {
+        size_t const newMB = ZSTD_estimateCStreamSize_internal(level);
+        if (newMB > memBudget) memBudget = newMB;
+    }
+    return memBudget;
+}
+
+/* ZSTD_getFrameProgression():
+ * tells how much data has been consumed (input) and produced (output) for current frame.
+ * able to count progression inside worker threads (non-blocking mode).
+ */
+ZSTD_frameProgression ZSTD_getFrameProgression(const ZSTD_CCtx* cctx)
+{
+    {   ZSTD_frameProgression fp;
+        size_t const buffered = (cctx->inBuff == NULL) ? 0 :
+                                cctx->inBuffPos - cctx->inToCompress;
+        if (buffered) assert(cctx->inBuffPos >= cctx->inToCompress);
+        assert(buffered <= ZSTD_BLOCKSIZE_MAX);
+        fp.ingested = cctx->consumedSrcSize + buffered;
+        fp.consumed = cctx->consumedSrcSize;
+        fp.produced = cctx->producedCSize;
+        fp.flushed  = cctx->producedCSize;   /* simplified; some data might still be left within streaming output buffer */
+        fp.currentJobID = 0;
+        fp.nbActiveWorkers = 0;
+        return fp;
+}   }
+
+/*! ZSTD_toFlushNow()
+ *  Only useful for multithreading scenarios currently (nbWorkers >= 1).
+ */
+size_t ZSTD_toFlushNow(ZSTD_CCtx* cctx)
+{
+    (void)cctx;
+    return 0;   /* over-simplification; could also check if context is currently running in streaming mode, and in which case, report how many bytes are left to be flushed within output buffer */
+}
+
+static void ZSTD_assertEqualCParams(ZSTD_compressionParameters cParams1,
+                                    ZSTD_compressionParameters cParams2)
+{
+    (void)cParams1;
+    (void)cParams2;
+    assert(cParams1.windowLog    == cParams2.windowLog);
+    assert(cParams1.chainLog     == cParams2.chainLog);
+    assert(cParams1.hashLog      == cParams2.hashLog);
+    assert(cParams1.searchLog    == cParams2.searchLog);
+    assert(cParams1.minMatch     == cParams2.minMatch);
+    assert(cParams1.targetLength == cParams2.targetLength);
+    assert(cParams1.strategy     == cParams2.strategy);
+}
+
+void ZSTD_reset_compressedBlockState(ZSTD_compressedBlockState_t* bs)
+{
+    int i;
+    for (i = 0; i < ZSTD_REP_NUM; ++i)
+        bs->rep[i] = repStartValue[i];
+    bs->entropy.huf.repeatMode = HUF_repeat_none;
+    bs->entropy.fse.offcode_repeatMode = FSE_repeat_none;
+    bs->entropy.fse.matchlength_repeatMode = FSE_repeat_none;
+    bs->entropy.fse.litlength_repeatMode = FSE_repeat_none;
+}
+
+/*! ZSTD_invalidateMatchState()
+ *  Invalidate all the matches in the match finder tables.
+ *  Requires nextSrc and base to be set (can be NULL).
+ */
+static void ZSTD_invalidateMatchState(ZSTD_matchState_t* ms)
+{
+    ZSTD_window_clear(&ms->window);
+
+    ms->nextToUpdate = ms->window.dictLimit;
+    ms->loadedDictEnd = 0;
+    ms->opt.litLengthSum = 0;  /* force reset of btopt stats */
+    ms->dictMatchState = NULL;
+}
+
+/*
+ * Controls, for this matchState reset, whether the tables need to be cleared /
+ * prepared for the coming compression (ZSTDcrp_makeClean), or whether the
+ * tables can be left unclean (ZSTDcrp_leaveDirty), because we know that a
+ * subsequent operation will overwrite the table space anyways (e.g., copying
+ * the matchState contents in from a CDict).
+ */
+typedef enum {
+    ZSTDcrp_makeClean,
+    ZSTDcrp_leaveDirty
+} ZSTD_compResetPolicy_e;
+
+/*
+ * Controls, for this matchState reset, whether indexing can continue where it
+ * left off (ZSTDirp_continue), or whether it needs to be restarted from zero
+ * (ZSTDirp_reset).
+ */
+typedef enum {
+    ZSTDirp_continue,
+    ZSTDirp_reset
+} ZSTD_indexResetPolicy_e;
+
+typedef enum {
+    ZSTD_resetTarget_CDict,
+    ZSTD_resetTarget_CCtx
+} ZSTD_resetTarget_e;
+
+static size_t
+ZSTD_reset_matchState(ZSTD_matchState_t* ms,
+                      ZSTD_cwksp* ws,
+                const ZSTD_compressionParameters* cParams,
+                const ZSTD_compResetPolicy_e crp,
+                const ZSTD_indexResetPolicy_e forceResetIndex,
+                const ZSTD_resetTarget_e forWho)
+{
+    size_t const chainSize = (cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cParams->chainLog);
+    size_t const hSize = ((size_t)1) << cParams->hashLog;
+    U32    const hashLog3 = ((forWho == ZSTD_resetTarget_CCtx) && cParams->minMatch==3) ? MIN(ZSTD_HASHLOG3_MAX, cParams->windowLog) : 0;
+    size_t const h3Size = hashLog3 ? ((size_t)1) << hashLog3 : 0;
+
+    DEBUGLOG(4, "reset indices : %u", forceResetIndex == ZSTDirp_reset);
+    if (forceResetIndex == ZSTDirp_reset) {
+        ZSTD_window_init(&ms->window);
+        ZSTD_cwksp_mark_tables_dirty(ws);
+    }
+
+    ms->hashLog3 = hashLog3;
+
+    ZSTD_invalidateMatchState(ms);
+
+    assert(!ZSTD_cwksp_reserve_failed(ws)); /* check that allocation hasn't already failed */
+
+    ZSTD_cwksp_clear_tables(ws);
+
+    DEBUGLOG(5, "reserving table space");
+    /* table Space */
+    ms->hashTable = (U32*)ZSTD_cwksp_reserve_table(ws, hSize * sizeof(U32));
+    ms->chainTable = (U32*)ZSTD_cwksp_reserve_table(ws, chainSize * sizeof(U32));
+    ms->hashTable3 = (U32*)ZSTD_cwksp_reserve_table(ws, h3Size * sizeof(U32));
+    RETURN_ERROR_IF(ZSTD_cwksp_reserve_failed(ws), memory_allocation,
+                    "failed a workspace allocation in ZSTD_reset_matchState");
+
+    DEBUGLOG(4, "reset table : %u", crp!=ZSTDcrp_leaveDirty);
+    if (crp!=ZSTDcrp_leaveDirty) {
+        /* reset tables only */
+        ZSTD_cwksp_clean_tables(ws);
+    }
+
+    /* opt parser space */
+    if ((forWho == ZSTD_resetTarget_CCtx) && (cParams->strategy >= ZSTD_btopt)) {
+        DEBUGLOG(4, "reserving optimal parser space");
+        ms->opt.litFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (1<<Litbits) * sizeof(unsigned));
+        ms->opt.litLengthFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxLL+1) * sizeof(unsigned));
+        ms->opt.matchLengthFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxML+1) * sizeof(unsigned));
+        ms->opt.offCodeFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxOff+1) * sizeof(unsigned));
+        ms->opt.matchTable = (ZSTD_match_t*)ZSTD_cwksp_reserve_aligned(ws, (ZSTD_OPT_NUM+1) * sizeof(ZSTD_match_t));
+        ms->opt.priceTable = (ZSTD_optimal_t*)ZSTD_cwksp_reserve_aligned(ws, (ZSTD_OPT_NUM+1) * sizeof(ZSTD_optimal_t));
+    }
+
+    ms->cParams = *cParams;
+
+    RETURN_ERROR_IF(ZSTD_cwksp_reserve_failed(ws), memory_allocation,
+                    "failed a workspace allocation in ZSTD_reset_matchState");
+
+    return 0;
+}
+
+/* ZSTD_indexTooCloseToMax() :
+ * minor optimization : prefer memset() rather than reduceIndex()
+ * which is measurably slow in some circumstances (reported for Visual Studio).
+ * Works when re-using a context for a lot of smallish inputs :
+ * if all inputs are smaller than ZSTD_INDEXOVERFLOW_MARGIN,
+ * memset() will be triggered before reduceIndex().
+ */
+#define ZSTD_INDEXOVERFLOW_MARGIN (16 MB)
+static int ZSTD_indexTooCloseToMax(ZSTD_window_t w)
+{
+    return (size_t)(w.nextSrc - w.base) > (ZSTD_CURRENT_MAX - ZSTD_INDEXOVERFLOW_MARGIN);
+}
+
+/*! ZSTD_resetCCtx_internal() :
+    note : `params` are assumed fully validated at this stage */
+static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
+                                      ZSTD_CCtx_params params,
+                                      U64 const pledgedSrcSize,
+                                      ZSTD_compResetPolicy_e const crp,
+                                      ZSTD_buffered_policy_e const zbuff)
+{
+    ZSTD_cwksp* const ws = &zc->workspace;
+    DEBUGLOG(4, "ZSTD_resetCCtx_internal: pledgedSrcSize=%u, wlog=%u",
+                (U32)pledgedSrcSize, params.cParams.windowLog);
+    assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
+
+    zc->isFirstBlock = 1;
+
+    if (params.ldmParams.enableLdm) {
+        /* Adjust long distance matching parameters */
+        ZSTD_ldm_adjustParameters(&params.ldmParams, &params.cParams);
+        assert(params.ldmParams.hashLog >= params.ldmParams.bucketSizeLog);
+        assert(params.ldmParams.hashRateLog < 32);
+    }
+
+    {   size_t const windowSize = MAX(1, (size_t)MIN(((U64)1 << params.cParams.windowLog), pledgedSrcSize));
+        size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, windowSize);
+        U32    const divider = (params.cParams.minMatch==3) ? 3 : 4;
+        size_t const maxNbSeq = blockSize / divider;
+        size_t const buffOutSize = (zbuff == ZSTDb_buffered && params.outBufferMode == ZSTD_bm_buffered)
+                ? ZSTD_compressBound(blockSize) + 1
+                : 0;
+        size_t const buffInSize = (zbuff == ZSTDb_buffered && params.inBufferMode == ZSTD_bm_buffered)
+                ? windowSize + blockSize
+                : 0;
+        size_t const maxNbLdmSeq = ZSTD_ldm_getMaxNbSeq(params.ldmParams, blockSize);
+
+        int const indexTooClose = ZSTD_indexTooCloseToMax(zc->blockState.matchState.window);
+        ZSTD_indexResetPolicy_e needsIndexReset =
+            (!indexTooClose && zc->initialized) ? ZSTDirp_continue : ZSTDirp_reset;
+
+        size_t const neededSpace =
+            ZSTD_estimateCCtxSize_usingCCtxParams_internal(
+                &params.cParams, &params.ldmParams, zc->staticSize != 0,
+                buffInSize, buffOutSize, pledgedSrcSize);
+        FORWARD_IF_ERROR(neededSpace, "cctx size estimate failed!");
+
+        if (!zc->staticSize) ZSTD_cwksp_bump_oversized_duration(ws, 0);
+
+        /* Check if workspace is large enough, alloc a new one if needed */
+        {
+            int const workspaceTooSmall = ZSTD_cwksp_sizeof(ws) < neededSpace;
+            int const workspaceWasteful = ZSTD_cwksp_check_wasteful(ws, neededSpace);
+
+            DEBUGLOG(4, "Need %zu B workspace", neededSpace);
+            DEBUGLOG(4, "windowSize: %zu - blockSize: %zu", windowSize, blockSize);
+
+            if (workspaceTooSmall || workspaceWasteful) {
+                DEBUGLOG(4, "Resize workspaceSize from %zuKB to %zuKB",
+                            ZSTD_cwksp_sizeof(ws) >> 10,
+                            neededSpace >> 10);
+
+                RETURN_ERROR_IF(zc->staticSize, memory_allocation, "static cctx : no resize");
+
+                needsIndexReset = ZSTDirp_reset;
+
+                ZSTD_cwksp_free(ws, zc->customMem);
+                FORWARD_IF_ERROR(ZSTD_cwksp_create(ws, neededSpace, zc->customMem), "");
+
+                DEBUGLOG(5, "reserving object space");
+                /* Statically sized space.
+                 * entropyWorkspace never moves,
+                 * though prev/next block swap places */
+                assert(ZSTD_cwksp_check_available(ws, 2 * sizeof(ZSTD_compressedBlockState_t)));
+                zc->blockState.prevCBlock = (ZSTD_compressedBlockState_t*) ZSTD_cwksp_reserve_object(ws, sizeof(ZSTD_compressedBlockState_t));
+                RETURN_ERROR_IF(zc->blockState.prevCBlock == NULL, memory_allocation, "couldn't allocate prevCBlock");
+                zc->blockState.nextCBlock = (ZSTD_compressedBlockState_t*) ZSTD_cwksp_reserve_object(ws, sizeof(ZSTD_compressedBlockState_t));
+                RETURN_ERROR_IF(zc->blockState.nextCBlock == NULL, memory_allocation, "couldn't allocate nextCBlock");
+                zc->entropyWorkspace = (U32*) ZSTD_cwksp_reserve_object(ws, ENTROPY_WORKSPACE_SIZE);
+                RETURN_ERROR_IF(zc->blockState.nextCBlock == NULL, memory_allocation, "couldn't allocate entropyWorkspace");
+        }   }
+
+        ZSTD_cwksp_clear(ws);
+
+        /* init params */
+        zc->appliedParams = params;
+        zc->blockState.matchState.cParams = params.cParams;
+        zc->pledgedSrcSizePlusOne = pledgedSrcSize+1;
+        zc->consumedSrcSize = 0;
+        zc->producedCSize = 0;
+        if (pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN)
+            zc->appliedParams.fParams.contentSizeFlag = 0;
+        DEBUGLOG(4, "pledged content size : %u ; flag : %u",
+            (unsigned)pledgedSrcSize, zc->appliedParams.fParams.contentSizeFlag);
+        zc->blockSize = blockSize;
+
+        xxh64_reset(&zc->xxhState, 0);
+        zc->stage = ZSTDcs_init;
+        zc->dictID = 0;
+        zc->dictContentSize = 0;
+
+        ZSTD_reset_compressedBlockState(zc->blockState.prevCBlock);
+
+        /* ZSTD_wildcopy() is used to copy into the literals buffer,
+         * so we have to oversize the buffer by WILDCOPY_OVERLENGTH bytes.
+         */
+        zc->seqStore.litStart = ZSTD_cwksp_reserve_buffer(ws, blockSize + WILDCOPY_OVERLENGTH);
+        zc->seqStore.maxNbLit = blockSize;
+
+        /* buffers */
+        zc->bufferedPolicy = zbuff;
+        zc->inBuffSize = buffInSize;
+        zc->inBuff = (char*)ZSTD_cwksp_reserve_buffer(ws, buffInSize);
+        zc->outBuffSize = buffOutSize;
+        zc->outBuff = (char*)ZSTD_cwksp_reserve_buffer(ws, buffOutSize);
+
+        /* ldm bucketOffsets table */
+        if (params.ldmParams.enableLdm) {
+            /* TODO: avoid memset? */
+            size_t const numBuckets =
+                  ((size_t)1) << (params.ldmParams.hashLog -
+                                  params.ldmParams.bucketSizeLog);
+            zc->ldmState.bucketOffsets = ZSTD_cwksp_reserve_buffer(ws, numBuckets);
+            ZSTD_memset(zc->ldmState.bucketOffsets, 0, numBuckets);
+        }
+
+        /* sequences storage */
+        ZSTD_referenceExternalSequences(zc, NULL, 0);
+        zc->seqStore.maxNbSeq = maxNbSeq;
+        zc->seqStore.llCode = ZSTD_cwksp_reserve_buffer(ws, maxNbSeq * sizeof(BYTE));
+        zc->seqStore.mlCode = ZSTD_cwksp_reserve_buffer(ws, maxNbSeq * sizeof(BYTE));
+        zc->seqStore.ofCode = ZSTD_cwksp_reserve_buffer(ws, maxNbSeq * sizeof(BYTE));
+        zc->seqStore.sequencesStart = (seqDef*)ZSTD_cwksp_reserve_aligned(ws, maxNbSeq * sizeof(seqDef));
+
+        FORWARD_IF_ERROR(ZSTD_reset_matchState(
+            &zc->blockState.matchState,
+            ws,
+            &params.cParams,
+            crp,
+            needsIndexReset,
+            ZSTD_resetTarget_CCtx), "");
+
+        /* ldm hash table */
+        if (params.ldmParams.enableLdm) {
+            /* TODO: avoid memset? */
+            size_t const ldmHSize = ((size_t)1) << params.ldmParams.hashLog;
+            zc->ldmState.hashTable = (ldmEntry_t*)ZSTD_cwksp_reserve_aligned(ws, ldmHSize * sizeof(ldmEntry_t));
+            ZSTD_memset(zc->ldmState.hashTable, 0, ldmHSize * sizeof(ldmEntry_t));
+            zc->ldmSequences = (rawSeq*)ZSTD_cwksp_reserve_aligned(ws, maxNbLdmSeq * sizeof(rawSeq));
+            zc->maxNbLdmSequences = maxNbLdmSeq;
+
+            ZSTD_window_init(&zc->ldmState.window);
+            ZSTD_window_clear(&zc->ldmState.window);
+            zc->ldmState.loadedDictEnd = 0;
+        }
+
+        /* Due to alignment, when reusing a workspace, we can actually consume
+         * up to 3 extra bytes for alignment. See the comments in zstd_cwksp.h
+         */
+        assert(ZSTD_cwksp_used(ws) >= neededSpace &&
+               ZSTD_cwksp_used(ws) <= neededSpace + 3);
+
+        DEBUGLOG(3, "wksp: finished allocating, %zd bytes remain available", ZSTD_cwksp_available_space(ws));
+        zc->initialized = 1;
+
+        return 0;
+    }
+}
+
+/* ZSTD_invalidateRepCodes() :
+ * ensures next compression will not use repcodes from previous block.
+ * Note : only works with regular variant;
+ *        do not use with extDict variant ! */
+void ZSTD_invalidateRepCodes(ZSTD_CCtx* cctx) {
+    int i;
+    for (i=0; i<ZSTD_REP_NUM; i++) cctx->blockState.prevCBlock->rep[i] = 0;
+    assert(!ZSTD_window_hasExtDict(cctx->blockState.matchState.window));
+}
+
+/* These are the approximate sizes for each strategy past which copying the
+ * dictionary tables into the working context is faster than using them
+ * in-place.
+ */
+static const size_t attachDictSizeCutoffs[ZSTD_STRATEGY_MAX+1] = {
+    8 KB,  /* unused */
+    8 KB,  /* ZSTD_fast */
+    16 KB, /* ZSTD_dfast */
+    32 KB, /* ZSTD_greedy */
+    32 KB, /* ZSTD_lazy */
+    32 KB, /* ZSTD_lazy2 */
+    32 KB, /* ZSTD_btlazy2 */
+    32 KB, /* ZSTD_btopt */
+    8 KB,  /* ZSTD_btultra */
+    8 KB   /* ZSTD_btultra2 */
+};
+
+static int ZSTD_shouldAttachDict(const ZSTD_CDict* cdict,
+                                 const ZSTD_CCtx_params* params,
+                                 U64 pledgedSrcSize)
+{
+    size_t cutoff = attachDictSizeCutoffs[cdict->matchState.cParams.strategy];
+    int const dedicatedDictSearch = cdict->matchState.dedicatedDictSearch;
+    return dedicatedDictSearch
+        || ( ( pledgedSrcSize <= cutoff
+            || pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN
+            || params->attachDictPref == ZSTD_dictForceAttach )
+          && params->attachDictPref != ZSTD_dictForceCopy
+          && !params->forceWindow ); /* dictMatchState isn't correctly
+                                      * handled in _enforceMaxDist */
+}
+
+static size_t
+ZSTD_resetCCtx_byAttachingCDict(ZSTD_CCtx* cctx,
+                        const ZSTD_CDict* cdict,
+                        ZSTD_CCtx_params params,
+                        U64 pledgedSrcSize,
+                        ZSTD_buffered_policy_e zbuff)
+{
+    {
+        ZSTD_compressionParameters adjusted_cdict_cParams = cdict->matchState.cParams;
+        unsigned const windowLog = params.cParams.windowLog;
+        assert(windowLog != 0);
+        /* Resize working context table params for input only, since the dict
+         * has its own tables. */
+        /* pledgedSrcSize == 0 means 0! */
+
+        if (cdict->matchState.dedicatedDictSearch) {
+            ZSTD_dedicatedDictSearch_revertCParams(&adjusted_cdict_cParams);
+        }
+
+        params.cParams = ZSTD_adjustCParams_internal(adjusted_cdict_cParams, pledgedSrcSize,
+                                                     cdict->dictContentSize, ZSTD_cpm_attachDict);
+        params.cParams.windowLog = windowLog;
+        FORWARD_IF_ERROR(ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize,
+                                                 ZSTDcrp_makeClean, zbuff), "");
+        assert(cctx->appliedParams.cParams.strategy == adjusted_cdict_cParams.strategy);
+    }
+
+    {   const U32 cdictEnd = (U32)( cdict->matchState.window.nextSrc
+                                  - cdict->matchState.window.base);
+        const U32 cdictLen = cdictEnd - cdict->matchState.window.dictLimit;
+        if (cdictLen == 0) {
+            /* don't even attach dictionaries with no contents */
+            DEBUGLOG(4, "skipping attaching empty dictionary");
+        } else {
+            DEBUGLOG(4, "attaching dictionary into context");
+            cctx->blockState.matchState.dictMatchState = &cdict->matchState;
+
+            /* prep working match state so dict matches never have negative indices
+             * when they are translated to the working context's index space. */
+            if (cctx->blockState.matchState.window.dictLimit < cdictEnd) {
+                cctx->blockState.matchState.window.nextSrc =
+                    cctx->blockState.matchState.window.base + cdictEnd;
+                ZSTD_window_clear(&cctx->blockState.matchState.window);
+            }
+            /* loadedDictEnd is expressed within the referential of the active context */
+            cctx->blockState.matchState.loadedDictEnd = cctx->blockState.matchState.window.dictLimit;
+    }   }
+
+    cctx->dictID = cdict->dictID;
+    cctx->dictContentSize = cdict->dictContentSize;
+
+    /* copy block state */
+    ZSTD_memcpy(cctx->blockState.prevCBlock, &cdict->cBlockState, sizeof(cdict->cBlockState));
+
+    return 0;
+}
+
+static size_t ZSTD_resetCCtx_byCopyingCDict(ZSTD_CCtx* cctx,
+                            const ZSTD_CDict* cdict,
+                            ZSTD_CCtx_params params,
+                            U64 pledgedSrcSize,
+                            ZSTD_buffered_policy_e zbuff)
+{
+    const ZSTD_compressionParameters *cdict_cParams = &cdict->matchState.cParams;
+
+    assert(!cdict->matchState.dedicatedDictSearch);
+
+    DEBUGLOG(4, "copying dictionary into context");
+
+    {   unsigned const windowLog = params.cParams.windowLog;
+        assert(windowLog != 0);
+        /* Copy only compression parameters related to tables. */
+        params.cParams = *cdict_cParams;
+        params.cParams.windowLog = windowLog;
+        FORWARD_IF_ERROR(ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize,
+                                                 ZSTDcrp_leaveDirty, zbuff), "");
+        assert(cctx->appliedParams.cParams.strategy == cdict_cParams->strategy);
+        assert(cctx->appliedParams.cParams.hashLog == cdict_cParams->hashLog);
+        assert(cctx->appliedParams.cParams.chainLog == cdict_cParams->chainLog);
+    }
+
+    ZSTD_cwksp_mark_tables_dirty(&cctx->workspace);
+
+    /* copy tables */
+    {   size_t const chainSize = (cdict_cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cdict_cParams->chainLog);
+        size_t const hSize =  (size_t)1 << cdict_cParams->hashLog;
+
+        ZSTD_memcpy(cctx->blockState.matchState.hashTable,
+               cdict->matchState.hashTable,
+               hSize * sizeof(U32));
+        ZSTD_memcpy(cctx->blockState.matchState.chainTable,
+               cdict->matchState.chainTable,
+               chainSize * sizeof(U32));
+    }
+
+    /* Zero the hashTable3, since the cdict never fills it */
+    {   int const h3log = cctx->blockState.matchState.hashLog3;
+        size_t const h3Size = h3log ? ((size_t)1 << h3log) : 0;
+        assert(cdict->matchState.hashLog3 == 0);
+        ZSTD_memset(cctx->blockState.matchState.hashTable3, 0, h3Size * sizeof(U32));
+    }
+
+    ZSTD_cwksp_mark_tables_clean(&cctx->workspace);
+
+    /* copy dictionary offsets */
+    {   ZSTD_matchState_t const* srcMatchState = &cdict->matchState;
+        ZSTD_matchState_t* dstMatchState = &cctx->blockState.matchState;
+        dstMatchState->window       = srcMatchState->window;
+        dstMatchState->nextToUpdate = srcMatchState->nextToUpdate;
+        dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd;
+    }
+
+    cctx->dictID = cdict->dictID;
+    cctx->dictContentSize = cdict->dictContentSize;
+
+    /* copy block state */
+    ZSTD_memcpy(cctx->blockState.prevCBlock, &cdict->cBlockState, sizeof(cdict->cBlockState));
+
+    return 0;
+}
+
+/* We have a choice between copying the dictionary context into the working
+ * context, or referencing the dictionary context from the working context
+ * in-place. We decide here which strategy to use. */
+static size_t ZSTD_resetCCtx_usingCDict(ZSTD_CCtx* cctx,
+                            const ZSTD_CDict* cdict,
+                            const ZSTD_CCtx_params* params,
+                            U64 pledgedSrcSize,
+                            ZSTD_buffered_policy_e zbuff)
+{
+
+    DEBUGLOG(4, "ZSTD_resetCCtx_usingCDict (pledgedSrcSize=%u)",
+                (unsigned)pledgedSrcSize);
+
+    if (ZSTD_shouldAttachDict(cdict, params, pledgedSrcSize)) {
+        return ZSTD_resetCCtx_byAttachingCDict(
+            cctx, cdict, *params, pledgedSrcSize, zbuff);
+    } else {
+        return ZSTD_resetCCtx_byCopyingCDict(
+            cctx, cdict, *params, pledgedSrcSize, zbuff);
+    }
+}
+
+/*! ZSTD_copyCCtx_internal() :
+ *  Duplicate an existing context `srcCCtx` into another one `dstCCtx`.
+ *  Only works during stage ZSTDcs_init (i.e. after creation, but before first call to ZSTD_compressContinue()).
+ *  The "context", in this case, refers to the hash and chain tables,
+ *  entropy tables, and dictionary references.
+ * `windowLog` value is enforced if != 0, otherwise value is copied from srcCCtx.
+ * @return : 0, or an error code */
+static size_t ZSTD_copyCCtx_internal(ZSTD_CCtx* dstCCtx,
+                            const ZSTD_CCtx* srcCCtx,
+                            ZSTD_frameParameters fParams,
+                            U64 pledgedSrcSize,
+                            ZSTD_buffered_policy_e zbuff)
+{
+    DEBUGLOG(5, "ZSTD_copyCCtx_internal");
+    RETURN_ERROR_IF(srcCCtx->stage!=ZSTDcs_init, stage_wrong,
+                    "Can't copy a ctx that's not in init stage.");
+
+    ZSTD_memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem));
+    {   ZSTD_CCtx_params params = dstCCtx->requestedParams;
+        /* Copy only compression parameters related to tables. */
+        params.cParams = srcCCtx->appliedParams.cParams;
+        params.fParams = fParams;
+        ZSTD_resetCCtx_internal(dstCCtx, params, pledgedSrcSize,
+                                ZSTDcrp_leaveDirty, zbuff);
+        assert(dstCCtx->appliedParams.cParams.windowLog == srcCCtx->appliedParams.cParams.windowLog);
+        assert(dstCCtx->appliedParams.cParams.strategy == srcCCtx->appliedParams.cParams.strategy);
+        assert(dstCCtx->appliedParams.cParams.hashLog == srcCCtx->appliedParams.cParams.hashLog);
+        assert(dstCCtx->appliedParams.cParams.chainLog == srcCCtx->appliedParams.cParams.chainLog);
+        assert(dstCCtx->blockState.matchState.hashLog3 == srcCCtx->blockState.matchState.hashLog3);
+    }
+
+    ZSTD_cwksp_mark_tables_dirty(&dstCCtx->workspace);
+
+    /* copy tables */
+    {   size_t const chainSize = (srcCCtx->appliedParams.cParams.strategy == ZSTD_fast) ? 0 : ((size_t)1 << srcCCtx->appliedParams.cParams.chainLog);
+        size_t const hSize =  (size_t)1 << srcCCtx->appliedParams.cParams.hashLog;
+        int const h3log = srcCCtx->blockState.matchState.hashLog3;
+        size_t const h3Size = h3log ? ((size_t)1 << h3log) : 0;
+
+        ZSTD_memcpy(dstCCtx->blockState.matchState.hashTable,
+               srcCCtx->blockState.matchState.hashTable,
+               hSize * sizeof(U32));
+        ZSTD_memcpy(dstCCtx->blockState.matchState.chainTable,
+               srcCCtx->blockState.matchState.chainTable,
+               chainSize * sizeof(U32));
+        ZSTD_memcpy(dstCCtx->blockState.matchState.hashTable3,
+               srcCCtx->blockState.matchState.hashTable3,
+               h3Size * sizeof(U32));
+    }
+
+    ZSTD_cwksp_mark_tables_clean(&dstCCtx->workspace);
+
+    /* copy dictionary offsets */
+    {
+        const ZSTD_matchState_t* srcMatchState = &srcCCtx->blockState.matchState;
+        ZSTD_matchState_t* dstMatchState = &dstCCtx->blockState.matchState;
+        dstMatchState->window       = srcMatchState->window;
+        dstMatchState->nextToUpdate = srcMatchState->nextToUpdate;
+        dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd;
+    }
+    dstCCtx->dictID = srcCCtx->dictID;
+    dstCCtx->dictContentSize = srcCCtx->dictContentSize;
+
+    /* copy block state */
+    ZSTD_memcpy(dstCCtx->blockState.prevCBlock, srcCCtx->blockState.prevCBlock, sizeof(*srcCCtx->blockState.prevCBlock));
+
+    return 0;
+}
+
+/*! ZSTD_copyCCtx() :
+ *  Duplicate an existing context `srcCCtx` into another one `dstCCtx`.
+ *  Only works during stage ZSTDcs_init (i.e. after creation, but before first call to ZSTD_compressContinue()).
+ *  pledgedSrcSize==0 means "unknown".
+*   @return : 0, or an error code */
+size_t ZSTD_copyCCtx(ZSTD_CCtx* dstCCtx, const ZSTD_CCtx* srcCCtx, unsigned long long pledgedSrcSize)
+{
+    ZSTD_frameParameters fParams = { 1 /*content*/, 0 /*checksum*/, 0 /*noDictID*/ };
+    ZSTD_buffered_policy_e const zbuff = srcCCtx->bufferedPolicy;
+    ZSTD_STATIC_ASSERT((U32)ZSTDb_buffered==1);
+    if (pledgedSrcSize==0) pledgedSrcSize = ZSTD_CONTENTSIZE_UNKNOWN;
+    fParams.contentSizeFlag = (pledgedSrcSize != ZSTD_CONTENTSIZE_UNKNOWN);
+
+    return ZSTD_copyCCtx_internal(dstCCtx, srcCCtx,
+                                fParams, pledgedSrcSize,
+                                zbuff);
+}
+
+
+#define ZSTD_ROWSIZE 16
+/*! ZSTD_reduceTable() :
+ *  reduce table indexes by `reducerValue`, or squash to zero.
+ *  PreserveMark preserves "unsorted mark" for btlazy2 strategy.
+ *  It must be set to a clear 0/1 value, to remove branch during inlining.
+ *  Presume table size is a multiple of ZSTD_ROWSIZE
+ *  to help auto-vectorization */
+FORCE_INLINE_TEMPLATE void
+ZSTD_reduceTable_internal (U32* const table, U32 const size, U32 const reducerValue, int const preserveMark)
+{
+    int const nbRows = (int)size / ZSTD_ROWSIZE;
+    int cellNb = 0;
+    int rowNb;
+    assert((size & (ZSTD_ROWSIZE-1)) == 0);  /* multiple of ZSTD_ROWSIZE */
+    assert(size < (1U<<31));   /* can be casted to int */
+
+
+    for (rowNb=0 ; rowNb < nbRows ; rowNb++) {
+        int column;
+        for (column=0; column<ZSTD_ROWSIZE; column++) {
+            if (preserveMark) {
+                U32 const adder = (table[cellNb] == ZSTD_DUBT_UNSORTED_MARK) ? reducerValue : 0;
+                table[cellNb] += adder;
+            }
+            if (table[cellNb] < reducerValue) table[cellNb] = 0;
+            else table[cellNb] -= reducerValue;
+            cellNb++;
+    }   }
+}
+
+static void ZSTD_reduceTable(U32* const table, U32 const size, U32 const reducerValue)
+{
+    ZSTD_reduceTable_internal(table, size, reducerValue, 0);
+}
+
+static void ZSTD_reduceTable_btlazy2(U32* const table, U32 const size, U32 const reducerValue)
+{
+    ZSTD_reduceTable_internal(table, size, reducerValue, 1);
+}
+
+/*! ZSTD_reduceIndex() :
+*   rescale all indexes to avoid future overflow (indexes are U32) */
+static void ZSTD_reduceIndex (ZSTD_matchState_t* ms, ZSTD_CCtx_params const* params, const U32 reducerValue)
+{
+    {   U32 const hSize = (U32)1 << params->cParams.hashLog;
+        ZSTD_reduceTable(ms->hashTable, hSize, reducerValue);
+    }
+
+    if (params->cParams.strategy != ZSTD_fast) {
+        U32 const chainSize = (U32)1 << params->cParams.chainLog;
+        if (params->cParams.strategy == ZSTD_btlazy2)
+            ZSTD_reduceTable_btlazy2(ms->chainTable, chainSize, reducerValue);
+        else
+            ZSTD_reduceTable(ms->chainTable, chainSize, reducerValue);
+    }
+
+    if (ms->hashLog3) {
+        U32 const h3Size = (U32)1 << ms->hashLog3;
+        ZSTD_reduceTable(ms->hashTable3, h3Size, reducerValue);
+    }
+}
+
+
+/*-*******************************************************
+*  Block entropic compression
+*********************************************************/
+
+/* See doc/zstd_compression_format.md for detailed format description */
+
+void ZSTD_seqToCodes(const seqStore_t* seqStorePtr)
+{
+    const seqDef* const sequences = seqStorePtr->sequencesStart;
+    BYTE* const llCodeTable = seqStorePtr->llCode;
+    BYTE* const ofCodeTable = seqStorePtr->ofCode;
+    BYTE* const mlCodeTable = seqStorePtr->mlCode;
+    U32 const nbSeq = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
+    U32 u;
+    assert(nbSeq <= seqStorePtr->maxNbSeq);
+    for (u=0; u<nbSeq; u++) {
+        U32 const llv = sequences[u].litLength;
+        U32 const mlv = sequences[u].matchLength;
+        llCodeTable[u] = (BYTE)ZSTD_LLcode(llv);
+        ofCodeTable[u] = (BYTE)ZSTD_highbit32(sequences[u].offset);
+        mlCodeTable[u] = (BYTE)ZSTD_MLcode(mlv);
+    }
+    if (seqStorePtr->longLengthID==1)
+        llCodeTable[seqStorePtr->longLengthPos] = MaxLL;
+    if (seqStorePtr->longLengthID==2)
+        mlCodeTable[seqStorePtr->longLengthPos] = MaxML;
+}
+
+/* ZSTD_useTargetCBlockSize():
+ * Returns if target compressed block size param is being used.
+ * If used, compression will do best effort to make a compressed block size to be around targetCBlockSize.
+ * Returns 1 if true, 0 otherwise. */
+static int ZSTD_useTargetCBlockSize(const ZSTD_CCtx_params* cctxParams)
+{
+    DEBUGLOG(5, "ZSTD_useTargetCBlockSize (targetCBlockSize=%zu)", cctxParams->targetCBlockSize);
+    return (cctxParams->targetCBlockSize != 0);
+}
+
+/* ZSTD_entropyCompressSequences_internal():
+ * actually compresses both literals and sequences */
+MEM_STATIC size_t
+ZSTD_entropyCompressSequences_internal(seqStore_t* seqStorePtr,
+                          const ZSTD_entropyCTables_t* prevEntropy,
+                                ZSTD_entropyCTables_t* nextEntropy,
+                          const ZSTD_CCtx_params* cctxParams,
+                                void* dst, size_t dstCapacity,
+                                void* entropyWorkspace, size_t entropyWkspSize,
+                          const int bmi2)
+{
+    const int longOffsets = cctxParams->cParams.windowLog > STREAM_ACCUMULATOR_MIN;
+    ZSTD_strategy const strategy = cctxParams->cParams.strategy;
+    unsigned* count = (unsigned*)entropyWorkspace;
+    FSE_CTable* CTable_LitLength = nextEntropy->fse.litlengthCTable;
+    FSE_CTable* CTable_OffsetBits = nextEntropy->fse.offcodeCTable;
+    FSE_CTable* CTable_MatchLength = nextEntropy->fse.matchlengthCTable;
+    U32 LLtype, Offtype, MLtype;   /* compressed, raw or rle */
+    const seqDef* const sequences = seqStorePtr->sequencesStart;
+    const BYTE* const ofCodeTable = seqStorePtr->ofCode;
+    const BYTE* const llCodeTable = seqStorePtr->llCode;
+    const BYTE* const mlCodeTable = seqStorePtr->mlCode;
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* const oend = ostart + dstCapacity;
+    BYTE* op = ostart;
+    size_t const nbSeq = (size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
+    BYTE* seqHead;
+    BYTE* lastNCount = NULL;
+
+    entropyWorkspace = count + (MaxSeq + 1);
+    entropyWkspSize -= (MaxSeq + 1) * sizeof(*count);
+
+    DEBUGLOG(4, "ZSTD_entropyCompressSequences_internal (nbSeq=%zu)", nbSeq);
+    ZSTD_STATIC_ASSERT(HUF_WORKSPACE_SIZE >= (1<<MAX(MLFSELog,LLFSELog)));
+    assert(entropyWkspSize >= HUF_WORKSPACE_SIZE);
+
+    /* Compress literals */
+    {   const BYTE* const literals = seqStorePtr->litStart;
+        size_t const litSize = (size_t)(seqStorePtr->lit - literals);
+        size_t const cSize = ZSTD_compressLiterals(
+                                    &prevEntropy->huf, &nextEntropy->huf,
+                                    cctxParams->cParams.strategy,
+                                    ZSTD_disableLiteralsCompression(cctxParams),
+                                    op, dstCapacity,
+                                    literals, litSize,
+                                    entropyWorkspace, entropyWkspSize,
+                                    bmi2);
+        FORWARD_IF_ERROR(cSize, "ZSTD_compressLiterals failed");
+        assert(cSize <= dstCapacity);
+        op += cSize;
+    }
+
+    /* Sequences Header */
+    RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/,
+                    dstSize_tooSmall, "Can't fit seq hdr in output buf!");
+    if (nbSeq < 128) {
+        *op++ = (BYTE)nbSeq;
+    } else if (nbSeq < LONGNBSEQ) {
+        op[0] = (BYTE)((nbSeq>>8) + 0x80);
+        op[1] = (BYTE)nbSeq;
+        op+=2;
+    } else {
+        op[0]=0xFF;
+        MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ));
+        op+=3;
+    }
+    assert(op <= oend);
+    if (nbSeq==0) {
+        /* Copy the old tables over as if we repeated them */
+        ZSTD_memcpy(&nextEntropy->fse, &prevEntropy->fse, sizeof(prevEntropy->fse));
+        return (size_t)(op - ostart);
+    }
+
+    /* seqHead : flags for FSE encoding type */
+    seqHead = op++;
+    assert(op <= oend);
+
+    /* convert length/distances into codes */
+    ZSTD_seqToCodes(seqStorePtr);
+    /* build CTable for Literal Lengths */
+    {   unsigned max = MaxLL;
+        size_t const mostFrequent = HIST_countFast_wksp(count, &max, llCodeTable, nbSeq, entropyWorkspace, entropyWkspSize);   /* can't fail */
+        DEBUGLOG(5, "Building LL table");
+        nextEntropy->fse.litlength_repeatMode = prevEntropy->fse.litlength_repeatMode;
+        LLtype = ZSTD_selectEncodingType(&nextEntropy->fse.litlength_repeatMode,
+                                        count, max, mostFrequent, nbSeq,
+                                        LLFSELog, prevEntropy->fse.litlengthCTable,
+                                        LL_defaultNorm, LL_defaultNormLog,
+                                        ZSTD_defaultAllowed, strategy);
+        assert(set_basic < set_compressed && set_rle < set_compressed);
+        assert(!(LLtype < set_compressed && nextEntropy->fse.litlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */
+        {   size_t const countSize = ZSTD_buildCTable(
+                op, (size_t)(oend - op),
+                CTable_LitLength, LLFSELog, (symbolEncodingType_e)LLtype,
+                count, max, llCodeTable, nbSeq,
+                LL_defaultNorm, LL_defaultNormLog, MaxLL,
+                prevEntropy->fse.litlengthCTable,
+                sizeof(prevEntropy->fse.litlengthCTable),
+                entropyWorkspace, entropyWkspSize);
+            FORWARD_IF_ERROR(countSize, "ZSTD_buildCTable for LitLens failed");
+            if (LLtype == set_compressed)
+                lastNCount = op;
+            op += countSize;
+            assert(op <= oend);
+    }   }
+    /* build CTable for Offsets */
+    {   unsigned max = MaxOff;
+        size_t const mostFrequent = HIST_countFast_wksp(
+            count, &max, ofCodeTable, nbSeq, entropyWorkspace, entropyWkspSize);  /* can't fail */
+        /* We can only use the basic table if max <= DefaultMaxOff, otherwise the offsets are too large */
+        ZSTD_defaultPolicy_e const defaultPolicy = (max <= DefaultMaxOff) ? ZSTD_defaultAllowed : ZSTD_defaultDisallowed;
+        DEBUGLOG(5, "Building OF table");
+        nextEntropy->fse.offcode_repeatMode = prevEntropy->fse.offcode_repeatMode;
+        Offtype = ZSTD_selectEncodingType(&nextEntropy->fse.offcode_repeatMode,
+                                        count, max, mostFrequent, nbSeq,
+                                        OffFSELog, prevEntropy->fse.offcodeCTable,
+                                        OF_defaultNorm, OF_defaultNormLog,
+                                        defaultPolicy, strategy);
+        assert(!(Offtype < set_compressed && nextEntropy->fse.offcode_repeatMode != FSE_repeat_none)); /* We don't copy tables */
+        {   size_t const countSize = ZSTD_buildCTable(
+                op, (size_t)(oend - op),
+                CTable_OffsetBits, OffFSELog, (symbolEncodingType_e)Offtype,
+                count, max, ofCodeTable, nbSeq,
+                OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff,
+                prevEntropy->fse.offcodeCTable,
+                sizeof(prevEntropy->fse.offcodeCTable),
+                entropyWorkspace, entropyWkspSize);
+            FORWARD_IF_ERROR(countSize, "ZSTD_buildCTable for Offsets failed");
+            if (Offtype == set_compressed)
+                lastNCount = op;
+            op += countSize;
+            assert(op <= oend);
+    }   }
+    /* build CTable for MatchLengths */
+    {   unsigned max = MaxML;
+        size_t const mostFrequent = HIST_countFast_wksp(
+            count, &max, mlCodeTable, nbSeq, entropyWorkspace, entropyWkspSize);   /* can't fail */
+        DEBUGLOG(5, "Building ML table (remaining space : %i)", (int)(oend-op));
+        nextEntropy->fse.matchlength_repeatMode = prevEntropy->fse.matchlength_repeatMode;
+        MLtype = ZSTD_selectEncodingType(&nextEntropy->fse.matchlength_repeatMode,
+                                        count, max, mostFrequent, nbSeq,
+                                        MLFSELog, prevEntropy->fse.matchlengthCTable,
+                                        ML_defaultNorm, ML_defaultNormLog,
+                                        ZSTD_defaultAllowed, strategy);
+        assert(!(MLtype < set_compressed && nextEntropy->fse.matchlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */
+        {   size_t const countSize = ZSTD_buildCTable(
+                op, (size_t)(oend - op),
+                CTable_MatchLength, MLFSELog, (symbolEncodingType_e)MLtype,
+                count, max, mlCodeTable, nbSeq,
+                ML_defaultNorm, ML_defaultNormLog, MaxML,
+                prevEntropy->fse.matchlengthCTable,
+                sizeof(prevEntropy->fse.matchlengthCTable),
+                entropyWorkspace, entropyWkspSize);
+            FORWARD_IF_ERROR(countSize, "ZSTD_buildCTable for MatchLengths failed");
+            if (MLtype == set_compressed)
+                lastNCount = op;
+            op += countSize;
+            assert(op <= oend);
+    }   }
+
+    *seqHead = (BYTE)((LLtype<<6) + (Offtype<<4) + (MLtype<<2));
+
+    {   size_t const bitstreamSize = ZSTD_encodeSequences(
+                                        op, (size_t)(oend - op),
+                                        CTable_MatchLength, mlCodeTable,
+                                        CTable_OffsetBits, ofCodeTable,
+                                        CTable_LitLength, llCodeTable,
+                                        sequences, nbSeq,
+                                        longOffsets, bmi2);
+        FORWARD_IF_ERROR(bitstreamSize, "ZSTD_encodeSequences failed");
+        op += bitstreamSize;
+        assert(op <= oend);
+        /* zstd versions <= 1.3.4 mistakenly report corruption when
+         * FSE_readNCount() receives a buffer < 4 bytes.
+         * Fixed by https://github.com/facebook/zstd/pull/1146.
+         * This can happen when the last set_compressed table present is 2
+         * bytes and the bitstream is only one byte.
+         * In this exceedingly rare case, we will simply emit an uncompressed
+         * block, since it isn't worth optimizing.
+         */
+        if (lastNCount && (op - lastNCount) < 4) {
+            /* NCountSize >= 2 && bitstreamSize > 0 ==> lastCountSize == 3 */
+            assert(op - lastNCount == 3);
+            DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.3.4 by "
+                        "emitting an uncompressed block.");
+            return 0;
+        }
+    }
+
+    DEBUGLOG(5, "compressed block size : %u", (unsigned)(op - ostart));
+    return (size_t)(op - ostart);
+}
+
+MEM_STATIC size_t
+ZSTD_entropyCompressSequences(seqStore_t* seqStorePtr,
+                       const ZSTD_entropyCTables_t* prevEntropy,
+                             ZSTD_entropyCTables_t* nextEntropy,
+                       const ZSTD_CCtx_params* cctxParams,
+                             void* dst, size_t dstCapacity,
+                             size_t srcSize,
+                             void* entropyWorkspace, size_t entropyWkspSize,
+                             int bmi2)
+{
+    size_t const cSize = ZSTD_entropyCompressSequences_internal(
+                            seqStorePtr, prevEntropy, nextEntropy, cctxParams,
+                            dst, dstCapacity,
+                            entropyWorkspace, entropyWkspSize, bmi2);
+    if (cSize == 0) return 0;
+    /* When srcSize <= dstCapacity, there is enough space to write a raw uncompressed block.
+     * Since we ran out of space, block must be not compressible, so fall back to raw uncompressed block.
+     */
+    if ((cSize == ERROR(dstSize_tooSmall)) & (srcSize <= dstCapacity))
+        return 0;  /* block not compressed */
+    FORWARD_IF_ERROR(cSize, "ZSTD_entropyCompressSequences_internal failed");
+
+    /* Check compressibility */
+    {   size_t const maxCSize = srcSize - ZSTD_minGain(srcSize, cctxParams->cParams.strategy);
+        if (cSize >= maxCSize) return 0;  /* block not compressed */
+    }
+    DEBUGLOG(4, "ZSTD_entropyCompressSequences() cSize: %zu\n", cSize);
+    return cSize;
+}
+
+/* ZSTD_selectBlockCompressor() :
+ * Not static, but internal use only (used by long distance matcher)
+ * assumption : strat is a valid strategy */
+ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_dictMode_e dictMode)
+{
+    static const ZSTD_blockCompressor blockCompressor[4][ZSTD_STRATEGY_MAX+1] = {
+        { ZSTD_compressBlock_fast  /* default for 0 */,
+          ZSTD_compressBlock_fast,
+          ZSTD_compressBlock_doubleFast,
+          ZSTD_compressBlock_greedy,
+          ZSTD_compressBlock_lazy,
+          ZSTD_compressBlock_lazy2,
+          ZSTD_compressBlock_btlazy2,
+          ZSTD_compressBlock_btopt,
+          ZSTD_compressBlock_btultra,
+          ZSTD_compressBlock_btultra2 },
+        { ZSTD_compressBlock_fast_extDict  /* default for 0 */,
+          ZSTD_compressBlock_fast_extDict,
+          ZSTD_compressBlock_doubleFast_extDict,
+          ZSTD_compressBlock_greedy_extDict,
+          ZSTD_compressBlock_lazy_extDict,
+          ZSTD_compressBlock_lazy2_extDict,
+          ZSTD_compressBlock_btlazy2_extDict,
+          ZSTD_compressBlock_btopt_extDict,
+          ZSTD_compressBlock_btultra_extDict,
+          ZSTD_compressBlock_btultra_extDict },
+        { ZSTD_compressBlock_fast_dictMatchState  /* default for 0 */,
+          ZSTD_compressBlock_fast_dictMatchState,
+          ZSTD_compressBlock_doubleFast_dictMatchState,
+          ZSTD_compressBlock_greedy_dictMatchState,
+          ZSTD_compressBlock_lazy_dictMatchState,
+          ZSTD_compressBlock_lazy2_dictMatchState,
+          ZSTD_compressBlock_btlazy2_dictMatchState,
+          ZSTD_compressBlock_btopt_dictMatchState,
+          ZSTD_compressBlock_btultra_dictMatchState,
+          ZSTD_compressBlock_btultra_dictMatchState },
+        { NULL  /* default for 0 */,
+          NULL,
+          NULL,
+          ZSTD_compressBlock_greedy_dedicatedDictSearch,
+          ZSTD_compressBlock_lazy_dedicatedDictSearch,
+          ZSTD_compressBlock_lazy2_dedicatedDictSearch,
+          NULL,
+          NULL,
+          NULL,
+          NULL }
+    };
+    ZSTD_blockCompressor selectedCompressor;
+    ZSTD_STATIC_ASSERT((unsigned)ZSTD_fast == 1);
+
+    assert(ZSTD_cParam_withinBounds(ZSTD_c_strategy, strat));
+    selectedCompressor = blockCompressor[(int)dictMode][(int)strat];
+    assert(selectedCompressor != NULL);
+    return selectedCompressor;
+}
+
+static void ZSTD_storeLastLiterals(seqStore_t* seqStorePtr,
+                                   const BYTE* anchor, size_t lastLLSize)
+{
+    ZSTD_memcpy(seqStorePtr->lit, anchor, lastLLSize);
+    seqStorePtr->lit += lastLLSize;
+}
+
+void ZSTD_resetSeqStore(seqStore_t* ssPtr)
+{
+    ssPtr->lit = ssPtr->litStart;
+    ssPtr->sequences = ssPtr->sequencesStart;
+    ssPtr->longLengthID = 0;
+}
+
+typedef enum { ZSTDbss_compress, ZSTDbss_noCompress } ZSTD_buildSeqStore_e;
+
+static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
+{
+    ZSTD_matchState_t* const ms = &zc->blockState.matchState;
+    DEBUGLOG(5, "ZSTD_buildSeqStore (srcSize=%zu)", srcSize);
+    assert(srcSize <= ZSTD_BLOCKSIZE_MAX);
+    /* Assert that we have correctly flushed the ctx params into the ms's copy */
+    ZSTD_assertEqualCParams(zc->appliedParams.cParams, ms->cParams);
+    if (srcSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) {
+        if (zc->appliedParams.cParams.strategy >= ZSTD_btopt) {
+            ZSTD_ldm_skipRawSeqStoreBytes(&zc->externSeqStore, srcSize);
+        } else {
+            ZSTD_ldm_skipSequences(&zc->externSeqStore, srcSize, zc->appliedParams.cParams.minMatch);
+        }
+        return ZSTDbss_noCompress; /* don't even attempt compression below a certain srcSize */
+    }
+    ZSTD_resetSeqStore(&(zc->seqStore));
+    /* required for optimal parser to read stats from dictionary */
+    ms->opt.symbolCosts = &zc->blockState.prevCBlock->entropy;
+    /* tell the optimal parser how we expect to compress literals */
+    ms->opt.literalCompressionMode = zc->appliedParams.literalCompressionMode;
+    /* a gap between an attached dict and the current window is not safe,
+     * they must remain adjacent,
+     * and when that stops being the case, the dict must be unset */
+    assert(ms->dictMatchState == NULL || ms->loadedDictEnd == ms->window.dictLimit);
+
+    /* limited update after a very long match */
+    {   const BYTE* const base = ms->window.base;
+        const BYTE* const istart = (const BYTE*)src;
+        const U32 curr = (U32)(istart-base);
+        if (sizeof(ptrdiff_t)==8) assert(istart - base < (ptrdiff_t)(U32)(-1));   /* ensure no overflow */
+        if (curr > ms->nextToUpdate + 384)
+            ms->nextToUpdate = curr - MIN(192, (U32)(curr - ms->nextToUpdate - 384));
+    }
+
+    /* select and store sequences */
+    {   ZSTD_dictMode_e const dictMode = ZSTD_matchState_dictMode(ms);
+        size_t lastLLSize;
+        {   int i;
+            for (i = 0; i < ZSTD_REP_NUM; ++i)
+                zc->blockState.nextCBlock->rep[i] = zc->blockState.prevCBlock->rep[i];
+        }
+        if (zc->externSeqStore.pos < zc->externSeqStore.size) {
+            assert(!zc->appliedParams.ldmParams.enableLdm);
+            /* Updates ldmSeqStore.pos */
+            lastLLSize =
+                ZSTD_ldm_blockCompress(&zc->externSeqStore,
+                                       ms, &zc->seqStore,
+                                       zc->blockState.nextCBlock->rep,
+                                       src, srcSize);
+            assert(zc->externSeqStore.pos <= zc->externSeqStore.size);
+        } else if (zc->appliedParams.ldmParams.enableLdm) {
+            rawSeqStore_t ldmSeqStore = kNullRawSeqStore;
+
+            ldmSeqStore.seq = zc->ldmSequences;
+            ldmSeqStore.capacity = zc->maxNbLdmSequences;
+            /* Updates ldmSeqStore.size */
+            FORWARD_IF_ERROR(ZSTD_ldm_generateSequences(&zc->ldmState, &ldmSeqStore,
+                                               &zc->appliedParams.ldmParams,
+                                               src, srcSize), "");
+            /* Updates ldmSeqStore.pos */
+            lastLLSize =
+                ZSTD_ldm_blockCompress(&ldmSeqStore,
+                                       ms, &zc->seqStore,
+                                       zc->blockState.nextCBlock->rep,
+                                       src, srcSize);
+            assert(ldmSeqStore.pos == ldmSeqStore.size);
+        } else {   /* not long range mode */
+            ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy, dictMode);
+            ms->ldmSeqStore = NULL;
+            lastLLSize = blockCompressor(ms, &zc->seqStore, zc->blockState.nextCBlock->rep, src, srcSize);
+        }
+        {   const BYTE* const lastLiterals = (const BYTE*)src + srcSize - lastLLSize;
+            ZSTD_storeLastLiterals(&zc->seqStore, lastLiterals, lastLLSize);
+    }   }
+    return ZSTDbss_compress;
+}
+
+static void ZSTD_copyBlockSequences(ZSTD_CCtx* zc)
+{
+    const seqStore_t* seqStore = ZSTD_getSeqStore(zc);
+    const seqDef* seqStoreSeqs = seqStore->sequencesStart;
+    size_t seqStoreSeqSize = seqStore->sequences - seqStoreSeqs;
+    size_t seqStoreLiteralsSize = (size_t)(seqStore->lit - seqStore->litStart);
+    size_t literalsRead = 0;
+    size_t lastLLSize;
+
+    ZSTD_Sequence* outSeqs = &zc->seqCollector.seqStart[zc->seqCollector.seqIndex];
+    size_t i;
+    repcodes_t updatedRepcodes;
+
+    assert(zc->seqCollector.seqIndex + 1 < zc->seqCollector.maxSequences);
+    /* Ensure we have enough space for last literals "sequence" */
+    assert(zc->seqCollector.maxSequences >= seqStoreSeqSize + 1);
+    ZSTD_memcpy(updatedRepcodes.rep, zc->blockState.prevCBlock->rep, sizeof(repcodes_t));
+    for (i = 0; i < seqStoreSeqSize; ++i) {
+        U32 rawOffset = seqStoreSeqs[i].offset - ZSTD_REP_NUM;
+        outSeqs[i].litLength = seqStoreSeqs[i].litLength;
+        outSeqs[i].matchLength = seqStoreSeqs[i].matchLength + MINMATCH;
+        outSeqs[i].rep = 0;
+
+        if (i == seqStore->longLengthPos) {
+            if (seqStore->longLengthID == 1) {
+                outSeqs[i].litLength += 0x10000;
+            } else if (seqStore->longLengthID == 2) {
+                outSeqs[i].matchLength += 0x10000;
+            }
+        }
+
+        if (seqStoreSeqs[i].offset <= ZSTD_REP_NUM) {
+            /* Derive the correct offset corresponding to a repcode */
+            outSeqs[i].rep = seqStoreSeqs[i].offset;
+            if (outSeqs[i].litLength != 0) {
+                rawOffset = updatedRepcodes.rep[outSeqs[i].rep - 1];
+            } else {
+                if (outSeqs[i].rep == 3) {
+                    rawOffset = updatedRepcodes.rep[0] - 1;
+                } else {
+                    rawOffset = updatedRepcodes.rep[outSeqs[i].rep];
+                }
+            }
+        }
+        outSeqs[i].offset = rawOffset;
+        /* seqStoreSeqs[i].offset == offCode+1, and ZSTD_updateRep() expects offCode
+           so we provide seqStoreSeqs[i].offset - 1 */
+        updatedRepcodes = ZSTD_updateRep(updatedRepcodes.rep,
+                                         seqStoreSeqs[i].offset - 1,
+                                         seqStoreSeqs[i].litLength == 0);
+        literalsRead += outSeqs[i].litLength;
+    }
+    /* Insert last literals (if any exist) in the block as a sequence with ml == off == 0.
+     * If there are no last literals, then we'll emit (of: 0, ml: 0, ll: 0), which is a marker
+     * for the block boundary, according to the API.
+     */
+    assert(seqStoreLiteralsSize >= literalsRead);
+    lastLLSize = seqStoreLiteralsSize - literalsRead;
+    outSeqs[i].litLength = (U32)lastLLSize;
+    outSeqs[i].matchLength = outSeqs[i].offset = outSeqs[i].rep = 0;
+    seqStoreSeqSize++;
+    zc->seqCollector.seqIndex += seqStoreSeqSize;
+}
+
+size_t ZSTD_generateSequences(ZSTD_CCtx* zc, ZSTD_Sequence* outSeqs,
+                              size_t outSeqsSize, const void* src, size_t srcSize)
+{
+    const size_t dstCapacity = ZSTD_compressBound(srcSize);
+    void* dst = ZSTD_customMalloc(dstCapacity, ZSTD_defaultCMem);
+    SeqCollector seqCollector;
+
+    RETURN_ERROR_IF(dst == NULL, memory_allocation, "NULL pointer!");
+
+    seqCollector.collectSequences = 1;
+    seqCollector.seqStart = outSeqs;
+    seqCollector.seqIndex = 0;
+    seqCollector.maxSequences = outSeqsSize;
+    zc->seqCollector = seqCollector;
+
+    ZSTD_compress2(zc, dst, dstCapacity, src, srcSize);
+    ZSTD_customFree(dst, ZSTD_defaultCMem);
+    return zc->seqCollector.seqIndex;
+}
+
+size_t ZSTD_mergeBlockDelimiters(ZSTD_Sequence* sequences, size_t seqsSize) {
+    size_t in = 0;
+    size_t out = 0;
+    for (; in < seqsSize; ++in) {
+        if (sequences[in].offset == 0 && sequences[in].matchLength == 0) {
+            if (in != seqsSize - 1) {
+                sequences[in+1].litLength += sequences[in].litLength;
+            }
+        } else {
+            sequences[out] = sequences[in];
+            ++out;
+        }
+    }
+    return out;
+}
+
+/* Unrolled loop to read four size_ts of input at a time. Returns 1 if is RLE, 0 if not. */
+static int ZSTD_isRLE(const BYTE* src, size_t length) {
+    const BYTE* ip = src;
+    const BYTE value = ip[0];
+    const size_t valueST = (size_t)((U64)value * 0x0101010101010101ULL);
+    const size_t unrollSize = sizeof(size_t) * 4;
+    const size_t unrollMask = unrollSize - 1;
+    const size_t prefixLength = length & unrollMask;
+    size_t i;
+    size_t u;
+    if (length == 1) return 1;
+    /* Check if prefix is RLE first before using unrolled loop */
+    if (prefixLength && ZSTD_count(ip+1, ip, ip+prefixLength) != prefixLength-1) {
+        return 0;
+    }
+    for (i = prefixLength; i != length; i += unrollSize) {
+        for (u = 0; u < unrollSize; u += sizeof(size_t)) {
+            if (MEM_readST(ip + i + u) != valueST) {
+                return 0;
+            }
+        }
+    }
+    return 1;
+}
+
+/* Returns true if the given block may be RLE.
+ * This is just a heuristic based on the compressibility.
+ * It may return both false positives and false negatives.
+ */
+static int ZSTD_maybeRLE(seqStore_t const* seqStore)
+{
+    size_t const nbSeqs = (size_t)(seqStore->sequences - seqStore->sequencesStart);
+    size_t const nbLits = (size_t)(seqStore->lit - seqStore->litStart);
+
+    return nbSeqs < 4 && nbLits < 10;
+}
+
+static void ZSTD_confirmRepcodesAndEntropyTables(ZSTD_CCtx* zc)
+{
+    ZSTD_compressedBlockState_t* const tmp = zc->blockState.prevCBlock;
+    zc->blockState.prevCBlock = zc->blockState.nextCBlock;
+    zc->blockState.nextCBlock = tmp;
+}
+
+static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc,
+                                        void* dst, size_t dstCapacity,
+                                        const void* src, size_t srcSize, U32 frame)
+{
+    /* This the upper bound for the length of an rle block.
+     * This isn't the actual upper bound. Finding the real threshold
+     * needs further investigation.
+     */
+    const U32 rleMaxLength = 25;
+    size_t cSize;
+    const BYTE* ip = (const BYTE*)src;
+    BYTE* op = (BYTE*)dst;
+    DEBUGLOG(5, "ZSTD_compressBlock_internal (dstCapacity=%u, dictLimit=%u, nextToUpdate=%u)",
+                (unsigned)dstCapacity, (unsigned)zc->blockState.matchState.window.dictLimit,
+                (unsigned)zc->blockState.matchState.nextToUpdate);
+
+    {   const size_t bss = ZSTD_buildSeqStore(zc, src, srcSize);
+        FORWARD_IF_ERROR(bss, "ZSTD_buildSeqStore failed");
+        if (bss == ZSTDbss_noCompress) { cSize = 0; goto out; }
+    }
+
+    if (zc->seqCollector.collectSequences) {
+        ZSTD_copyBlockSequences(zc);
+        ZSTD_confirmRepcodesAndEntropyTables(zc);
+        return 0;
+    }
+
+    /* encode sequences and literals */
+    cSize = ZSTD_entropyCompressSequences(&zc->seqStore,
+            &zc->blockState.prevCBlock->entropy, &zc->blockState.nextCBlock->entropy,
+            &zc->appliedParams,
+            dst, dstCapacity,
+            srcSize,
+            zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE /* statically allocated in resetCCtx */,
+            zc->bmi2);
+
+    if (zc->seqCollector.collectSequences) {
+        ZSTD_copyBlockSequences(zc);
+        return 0;
+    }
+
+
+    if (frame &&
+        /* We don't want to emit our first block as a RLE even if it qualifies because
+         * doing so will cause the decoder (cli only) to throw a "should consume all input error."
+         * This is only an issue for zstd <= v1.4.3
+         */
+        !zc->isFirstBlock &&
+        cSize < rleMaxLength &&
+        ZSTD_isRLE(ip, srcSize))
+    {
+        cSize = 1;
+        op[0] = ip[0];
+    }
+
+out:
+    if (!ZSTD_isError(cSize) && cSize > 1) {
+        ZSTD_confirmRepcodesAndEntropyTables(zc);
+    }
+    /* We check that dictionaries have offset codes available for the first
+     * block. After the first block, the offcode table might not have large
+     * enough codes to represent the offsets in the data.
+     */
+    if (zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode == FSE_repeat_valid)
+        zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode = FSE_repeat_check;
+
+    return cSize;
+}
+
+static size_t ZSTD_compressBlock_targetCBlockSize_body(ZSTD_CCtx* zc,
+                               void* dst, size_t dstCapacity,
+                               const void* src, size_t srcSize,
+                               const size_t bss, U32 lastBlock)
+{
+    DEBUGLOG(6, "Attempting ZSTD_compressSuperBlock()");
+    if (bss == ZSTDbss_compress) {
+        if (/* We don't want to emit our first block as a RLE even if it qualifies because
+            * doing so will cause the decoder (cli only) to throw a "should consume all input error."
+            * This is only an issue for zstd <= v1.4.3
+            */
+            !zc->isFirstBlock &&
+            ZSTD_maybeRLE(&zc->seqStore) &&
+            ZSTD_isRLE((BYTE const*)src, srcSize))
+        {
+            return ZSTD_rleCompressBlock(dst, dstCapacity, *(BYTE const*)src, srcSize, lastBlock);
+        }
+        /* Attempt superblock compression.
+         *
+         * Note that compressed size of ZSTD_compressSuperBlock() is not bound by the
+         * standard ZSTD_compressBound(). This is a problem, because even if we have
+         * space now, taking an extra byte now could cause us to run out of space later
+         * and violate ZSTD_compressBound().
+         *
+         * Define blockBound(blockSize) = blockSize + ZSTD_blockHeaderSize.
+         *
+         * In order to respect ZSTD_compressBound() we must attempt to emit a raw
+         * uncompressed block in these cases:
+         *   * cSize == 0: Return code for an uncompressed block.
+         *   * cSize == dstSize_tooSmall: We may have expanded beyond blockBound(srcSize).
+         *     ZSTD_noCompressBlock() will return dstSize_tooSmall if we are really out of
+         *     output space.
+         *   * cSize >= blockBound(srcSize): We have expanded the block too much so
+         *     emit an uncompressed block.
+         */
+        {
+            size_t const cSize = ZSTD_compressSuperBlock(zc, dst, dstCapacity, src, srcSize, lastBlock);
+            if (cSize != ERROR(dstSize_tooSmall)) {
+                size_t const maxCSize = srcSize - ZSTD_minGain(srcSize, zc->appliedParams.cParams.strategy);
+                FORWARD_IF_ERROR(cSize, "ZSTD_compressSuperBlock failed");
+                if (cSize != 0 && cSize < maxCSize + ZSTD_blockHeaderSize) {
+                    ZSTD_confirmRepcodesAndEntropyTables(zc);
+                    return cSize;
+                }
+            }
+        }
+    }
+
+    DEBUGLOG(6, "Resorting to ZSTD_noCompressBlock()");
+    /* Superblock compression failed, attempt to emit a single no compress block.
+     * The decoder will be able to stream this block since it is uncompressed.
+     */
+    return ZSTD_noCompressBlock(dst, dstCapacity, src, srcSize, lastBlock);
+}
+
+static size_t ZSTD_compressBlock_targetCBlockSize(ZSTD_CCtx* zc,
+                               void* dst, size_t dstCapacity,
+                               const void* src, size_t srcSize,
+                               U32 lastBlock)
+{
+    size_t cSize = 0;
+    const size_t bss = ZSTD_buildSeqStore(zc, src, srcSize);
+    DEBUGLOG(5, "ZSTD_compressBlock_targetCBlockSize (dstCapacity=%u, dictLimit=%u, nextToUpdate=%u, srcSize=%zu)",
+                (unsigned)dstCapacity, (unsigned)zc->blockState.matchState.window.dictLimit, (unsigned)zc->blockState.matchState.nextToUpdate, srcSize);
+    FORWARD_IF_ERROR(bss, "ZSTD_buildSeqStore failed");
+
+    cSize = ZSTD_compressBlock_targetCBlockSize_body(zc, dst, dstCapacity, src, srcSize, bss, lastBlock);
+    FORWARD_IF_ERROR(cSize, "ZSTD_compressBlock_targetCBlockSize_body failed");
+
+    if (zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode == FSE_repeat_valid)
+        zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode = FSE_repeat_check;
+
+    return cSize;
+}
+
+static void ZSTD_overflowCorrectIfNeeded(ZSTD_matchState_t* ms,
+                                         ZSTD_cwksp* ws,
+                                         ZSTD_CCtx_params const* params,
+                                         void const* ip,
+                                         void const* iend)
+{
+    if (ZSTD_window_needOverflowCorrection(ms->window, iend)) {
+        U32 const maxDist = (U32)1 << params->cParams.windowLog;
+        U32 const cycleLog = ZSTD_cycleLog(params->cParams.chainLog, params->cParams.strategy);
+        U32 const correction = ZSTD_window_correctOverflow(&ms->window, cycleLog, maxDist, ip);
+        ZSTD_STATIC_ASSERT(ZSTD_CHAINLOG_MAX <= 30);
+        ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_32 <= 30);
+        ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX <= 31);
+        ZSTD_cwksp_mark_tables_dirty(ws);
+        ZSTD_reduceIndex(ms, params, correction);
+        ZSTD_cwksp_mark_tables_clean(ws);
+        if (ms->nextToUpdate < correction) ms->nextToUpdate = 0;
+        else ms->nextToUpdate -= correction;
+        /* invalidate dictionaries on overflow correction */
+        ms->loadedDictEnd = 0;
+        ms->dictMatchState = NULL;
+    }
+}
+
+/*! ZSTD_compress_frameChunk() :
+*   Compress a chunk of data into one or multiple blocks.
+*   All blocks will be terminated, all input will be consumed.
+*   Function will issue an error if there is not enough `dstCapacity` to hold the compressed content.
+*   Frame is supposed already started (header already produced)
+*   @return : compressed size, or an error code
+*/
+static size_t ZSTD_compress_frameChunk (ZSTD_CCtx* cctx,
+                                     void* dst, size_t dstCapacity,
+                               const void* src, size_t srcSize,
+                                     U32 lastFrameChunk)
+{
+    size_t blockSize = cctx->blockSize;
+    size_t remaining = srcSize;
+    const BYTE* ip = (const BYTE*)src;
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* op = ostart;
+    U32 const maxDist = (U32)1 << cctx->appliedParams.cParams.windowLog;
+
+    assert(cctx->appliedParams.cParams.windowLog <= ZSTD_WINDOWLOG_MAX);
+
+    DEBUGLOG(4, "ZSTD_compress_frameChunk (blockSize=%u)", (unsigned)blockSize);
+    if (cctx->appliedParams.fParams.checksumFlag && srcSize)
+        xxh64_update(&cctx->xxhState, src, srcSize);
+
+    while (remaining) {
+        ZSTD_matchState_t* const ms = &cctx->blockState.matchState;
+        U32 const lastBlock = lastFrameChunk & (blockSize >= remaining);
+
+        RETURN_ERROR_IF(dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE,
+                        dstSize_tooSmall,
+                        "not enough space to store compressed block");
+        if (remaining < blockSize) blockSize = remaining;
+
+        ZSTD_overflowCorrectIfNeeded(
+            ms, &cctx->workspace, &cctx->appliedParams, ip, ip + blockSize);
+        ZSTD_checkDictValidity(&ms->window, ip + blockSize, maxDist, &ms->loadedDictEnd, &ms->dictMatchState);
+
+        /* Ensure hash/chain table insertion resumes no sooner than lowlimit */
+        if (ms->nextToUpdate < ms->window.lowLimit) ms->nextToUpdate = ms->window.lowLimit;
+
+        {   size_t cSize;
+            if (ZSTD_useTargetCBlockSize(&cctx->appliedParams)) {
+                cSize = ZSTD_compressBlock_targetCBlockSize(cctx, op, dstCapacity, ip, blockSize, lastBlock);
+                FORWARD_IF_ERROR(cSize, "ZSTD_compressBlock_targetCBlockSize failed");
+                assert(cSize > 0);
+                assert(cSize <= blockSize + ZSTD_blockHeaderSize);
+            } else {
+                cSize = ZSTD_compressBlock_internal(cctx,
+                                        op+ZSTD_blockHeaderSize, dstCapacity-ZSTD_blockHeaderSize,
+                                        ip, blockSize, 1 /* frame */);
+                FORWARD_IF_ERROR(cSize, "ZSTD_compressBlock_internal failed");
+
+                if (cSize == 0) {  /* block is not compressible */
+                    cSize = ZSTD_noCompressBlock(op, dstCapacity, ip, blockSize, lastBlock);
+                    FORWARD_IF_ERROR(cSize, "ZSTD_noCompressBlock failed");
+                } else {
+                    U32 const cBlockHeader = cSize == 1 ?
+                        lastBlock + (((U32)bt_rle)<<1) + (U32)(blockSize << 3) :
+                        lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3);
+                    MEM_writeLE24(op, cBlockHeader);
+                    cSize += ZSTD_blockHeaderSize;
+                }
+            }
+
+
+            ip += blockSize;
+            assert(remaining >= blockSize);
+            remaining -= blockSize;
+            op += cSize;
+            assert(dstCapacity >= cSize);
+            dstCapacity -= cSize;
+            cctx->isFirstBlock = 0;
+            DEBUGLOG(5, "ZSTD_compress_frameChunk: adding a block of size %u",
+                        (unsigned)cSize);
+    }   }
+
+    if (lastFrameChunk && (op>ostart)) cctx->stage = ZSTDcs_ending;
+    return (size_t)(op-ostart);
+}
+
+
+static size_t ZSTD_writeFrameHeader(void* dst, size_t dstCapacity,
+                                    const ZSTD_CCtx_params* params, U64 pledgedSrcSize, U32 dictID)
+{   BYTE* const op = (BYTE*)dst;
+    U32   const dictIDSizeCodeLength = (dictID>0) + (dictID>=256) + (dictID>=65536);   /* 0-3 */
+    U32   const dictIDSizeCode = params->fParams.noDictIDFlag ? 0 : dictIDSizeCodeLength;   /* 0-3 */
+    U32   const checksumFlag = params->fParams.checksumFlag>0;
+    U32   const windowSize = (U32)1 << params->cParams.windowLog;
+    U32   const singleSegment = params->fParams.contentSizeFlag && (windowSize >= pledgedSrcSize);
+    BYTE  const windowLogByte = (BYTE)((params->cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3);
+    U32   const fcsCode = params->fParams.contentSizeFlag ?
+                     (pledgedSrcSize>=256) + (pledgedSrcSize>=65536+256) + (pledgedSrcSize>=0xFFFFFFFFU) : 0;  /* 0-3 */
+    BYTE  const frameHeaderDescriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag<<2) + (singleSegment<<5) + (fcsCode<<6) );
+    size_t pos=0;
+
+    assert(!(params->fParams.contentSizeFlag && pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN));
+    RETURN_ERROR_IF(dstCapacity < ZSTD_FRAMEHEADERSIZE_MAX, dstSize_tooSmall,
+                    "dst buf is too small to fit worst-case frame header size.");
+    DEBUGLOG(4, "ZSTD_writeFrameHeader : dictIDFlag : %u ; dictID : %u ; dictIDSizeCode : %u",
+                !params->fParams.noDictIDFlag, (unsigned)dictID, (unsigned)dictIDSizeCode);
+    if (params->format == ZSTD_f_zstd1) {
+        MEM_writeLE32(dst, ZSTD_MAGICNUMBER);
+        pos = 4;
+    }
+    op[pos++] = frameHeaderDescriptionByte;
+    if (!singleSegment) op[pos++] = windowLogByte;
+    switch(dictIDSizeCode)
+    {
+        default:
+            assert(0); /* impossible */
+            ZSTD_FALLTHROUGH;
+        case 0 : break;
+        case 1 : op[pos] = (BYTE)(dictID); pos++; break;
+        case 2 : MEM_writeLE16(op+pos, (U16)dictID); pos+=2; break;
+        case 3 : MEM_writeLE32(op+pos, dictID); pos+=4; break;
+    }
+    switch(fcsCode)
+    {
+        default:
+            assert(0); /* impossible */
+            ZSTD_FALLTHROUGH;
+        case 0 : if (singleSegment) op[pos++] = (BYTE)(pledgedSrcSize); break;
+        case 1 : MEM_writeLE16(op+pos, (U16)(pledgedSrcSize-256)); pos+=2; break;
+        case 2 : MEM_writeLE32(op+pos, (U32)(pledgedSrcSize)); pos+=4; break;
+        case 3 : MEM_writeLE64(op+pos, (U64)(pledgedSrcSize)); pos+=8; break;
+    }
+    return pos;
+}
+
+/* ZSTD_writeSkippableFrame_advanced() :
+ * Writes out a skippable frame with the specified magic number variant (16 are supported),
+ * from ZSTD_MAGIC_SKIPPABLE_START to ZSTD_MAGIC_SKIPPABLE_START+15, and the desired source data.
+ *
+ * Returns the total number of bytes written, or a ZSTD error code.
+ */
+size_t ZSTD_writeSkippableFrame(void* dst, size_t dstCapacity,
+                                const void* src, size_t srcSize, unsigned magicVariant) {
+    BYTE* op = (BYTE*)dst;
+    RETURN_ERROR_IF(dstCapacity < srcSize + ZSTD_SKIPPABLEHEADERSIZE /* Skippable frame overhead */,
+                    dstSize_tooSmall, "Not enough room for skippable frame");
+    RETURN_ERROR_IF(srcSize > (unsigned)0xFFFFFFFF, srcSize_wrong, "Src size too large for skippable frame");
+    RETURN_ERROR_IF(magicVariant > 15, parameter_outOfBound, "Skippable frame magic number variant not supported");
+
+    MEM_writeLE32(op, (U32)(ZSTD_MAGIC_SKIPPABLE_START + magicVariant));
+    MEM_writeLE32(op+4, (U32)srcSize);
+    ZSTD_memcpy(op+8, src, srcSize);
+    return srcSize + ZSTD_SKIPPABLEHEADERSIZE;
+}
+
+/* ZSTD_writeLastEmptyBlock() :
+ * output an empty Block with end-of-frame mark to complete a frame
+ * @return : size of data written into `dst` (== ZSTD_blockHeaderSize (defined in zstd_internal.h))
+ *           or an error code if `dstCapacity` is too small (<ZSTD_blockHeaderSize)
+ */
+size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity)
+{
+    RETURN_ERROR_IF(dstCapacity < ZSTD_blockHeaderSize, dstSize_tooSmall,
+                    "dst buf is too small to write frame trailer empty block.");
+    {   U32 const cBlockHeader24 = 1 /*lastBlock*/ + (((U32)bt_raw)<<1);  /* 0 size */
+        MEM_writeLE24(dst, cBlockHeader24);
+        return ZSTD_blockHeaderSize;
+    }
+}
+
+size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq)
+{
+    RETURN_ERROR_IF(cctx->stage != ZSTDcs_init, stage_wrong,
+                    "wrong cctx stage");
+    RETURN_ERROR_IF(cctx->appliedParams.ldmParams.enableLdm,
+                    parameter_unsupported,
+                    "incompatible with ldm");
+    cctx->externSeqStore.seq = seq;
+    cctx->externSeqStore.size = nbSeq;
+    cctx->externSeqStore.capacity = nbSeq;
+    cctx->externSeqStore.pos = 0;
+    cctx->externSeqStore.posInSequence = 0;
+    return 0;
+}
+
+
+static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* cctx,
+                              void* dst, size_t dstCapacity,
+                        const void* src, size_t srcSize,
+                               U32 frame, U32 lastFrameChunk)
+{
+    ZSTD_matchState_t* const ms = &cctx->blockState.matchState;
+    size_t fhSize = 0;
+
+    DEBUGLOG(5, "ZSTD_compressContinue_internal, stage: %u, srcSize: %u",
+                cctx->stage, (unsigned)srcSize);
+    RETURN_ERROR_IF(cctx->stage==ZSTDcs_created, stage_wrong,
+                    "missing init (ZSTD_compressBegin)");
+
+    if (frame && (cctx->stage==ZSTDcs_init)) {
+        fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, &cctx->appliedParams,
+                                       cctx->pledgedSrcSizePlusOne-1, cctx->dictID);
+        FORWARD_IF_ERROR(fhSize, "ZSTD_writeFrameHeader failed");
+        assert(fhSize <= dstCapacity);
+        dstCapacity -= fhSize;
+        dst = (char*)dst + fhSize;
+        cctx->stage = ZSTDcs_ongoing;
+    }
+
+    if (!srcSize) return fhSize;  /* do not generate an empty block if no input */
+
+    if (!ZSTD_window_update(&ms->window, src, srcSize)) {
+        ms->nextToUpdate = ms->window.dictLimit;
+    }
+    if (cctx->appliedParams.ldmParams.enableLdm) {
+        ZSTD_window_update(&cctx->ldmState.window, src, srcSize);
+    }
+
+    if (!frame) {
+        /* overflow check and correction for block mode */
+        ZSTD_overflowCorrectIfNeeded(
+            ms, &cctx->workspace, &cctx->appliedParams,
+            src, (BYTE const*)src + srcSize);
+    }
+
+    DEBUGLOG(5, "ZSTD_compressContinue_internal (blockSize=%u)", (unsigned)cctx->blockSize);
+    {   size_t const cSize = frame ?
+                             ZSTD_compress_frameChunk (cctx, dst, dstCapacity, src, srcSize, lastFrameChunk) :
+                             ZSTD_compressBlock_internal (cctx, dst, dstCapacity, src, srcSize, 0 /* frame */);
+        FORWARD_IF_ERROR(cSize, "%s", frame ? "ZSTD_compress_frameChunk failed" : "ZSTD_compressBlock_internal failed");
+        cctx->consumedSrcSize += srcSize;
+        cctx->producedCSize += (cSize + fhSize);
+        assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0));
+        if (cctx->pledgedSrcSizePlusOne != 0) {  /* control src size */
+            ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1);
+            RETURN_ERROR_IF(
+                cctx->consumedSrcSize+1 > cctx->pledgedSrcSizePlusOne,
+                srcSize_wrong,
+                "error : pledgedSrcSize = %u, while realSrcSize >= %u",
+                (unsigned)cctx->pledgedSrcSizePlusOne-1,
+                (unsigned)cctx->consumedSrcSize);
+        }
+        return cSize + fhSize;
+    }
+}
+
+size_t ZSTD_compressContinue (ZSTD_CCtx* cctx,
+                              void* dst, size_t dstCapacity,
+                        const void* src, size_t srcSize)
+{
+    DEBUGLOG(5, "ZSTD_compressContinue (srcSize=%u)", (unsigned)srcSize);
+    return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1 /* frame mode */, 0 /* last chunk */);
+}
+
+
+size_t ZSTD_getBlockSize(const ZSTD_CCtx* cctx)
+{
+    ZSTD_compressionParameters const cParams = cctx->appliedParams.cParams;
+    assert(!ZSTD_checkCParams(cParams));
+    return MIN (ZSTD_BLOCKSIZE_MAX, (U32)1 << cParams.windowLog);
+}
+
+size_t ZSTD_compressBlock(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+    DEBUGLOG(5, "ZSTD_compressBlock: srcSize = %u", (unsigned)srcSize);
+    { size_t const blockSizeMax = ZSTD_getBlockSize(cctx);
+      RETURN_ERROR_IF(srcSize > blockSizeMax, srcSize_wrong, "input is larger than a block"); }
+
+    return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 0 /* frame mode */, 0 /* last chunk */);
+}
+
+/*! ZSTD_loadDictionaryContent() :
+ *  @return : 0, or an error code
+ */
+static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms,
+                                         ldmState_t* ls,
+                                         ZSTD_cwksp* ws,
+                                         ZSTD_CCtx_params const* params,
+                                         const void* src, size_t srcSize,
+                                         ZSTD_dictTableLoadMethod_e dtlm)
+{
+    const BYTE* ip = (const BYTE*) src;
+    const BYTE* const iend = ip + srcSize;
+
+    ZSTD_window_update(&ms->window, src, srcSize);
+    ms->loadedDictEnd = params->forceWindow ? 0 : (U32)(iend - ms->window.base);
+
+    if (params->ldmParams.enableLdm && ls != NULL) {
+        ZSTD_window_update(&ls->window, src, srcSize);
+        ls->loadedDictEnd = params->forceWindow ? 0 : (U32)(iend - ls->window.base);
+    }
+
+    /* Assert that we the ms params match the params we're being given */
+    ZSTD_assertEqualCParams(params->cParams, ms->cParams);
+
+    if (srcSize <= HASH_READ_SIZE) return 0;
+
+    while (iend - ip > HASH_READ_SIZE) {
+        size_t const remaining = (size_t)(iend - ip);
+        size_t const chunk = MIN(remaining, ZSTD_CHUNKSIZE_MAX);
+        const BYTE* const ichunk = ip + chunk;
+
+        ZSTD_overflowCorrectIfNeeded(ms, ws, params, ip, ichunk);
+
+        if (params->ldmParams.enableLdm && ls != NULL)
+            ZSTD_ldm_fillHashTable(ls, (const BYTE*)src, (const BYTE*)src + srcSize, &params->ldmParams);
+
+        switch(params->cParams.strategy)
+        {
+        case ZSTD_fast:
+            ZSTD_fillHashTable(ms, ichunk, dtlm);
+            break;
+        case ZSTD_dfast:
+            ZSTD_fillDoubleHashTable(ms, ichunk, dtlm);
+            break;
+
+        case ZSTD_greedy:
+        case ZSTD_lazy:
+        case ZSTD_lazy2:
+            if (chunk >= HASH_READ_SIZE && ms->dedicatedDictSearch) {
+                assert(chunk == remaining); /* must load everything in one go */
+                ZSTD_dedicatedDictSearch_lazy_loadDictionary(ms, ichunk-HASH_READ_SIZE);
+            } else if (chunk >= HASH_READ_SIZE) {
+                ZSTD_insertAndFindFirstIndex(ms, ichunk-HASH_READ_SIZE);
+            }
+            break;
+
+        case ZSTD_btlazy2:   /* we want the dictionary table fully sorted */
+        case ZSTD_btopt:
+        case ZSTD_btultra:
+        case ZSTD_btultra2:
+            if (chunk >= HASH_READ_SIZE)
+                ZSTD_updateTree(ms, ichunk-HASH_READ_SIZE, ichunk);
+            break;
+
+        default:
+            assert(0);  /* not possible : not a valid strategy id */
+        }
+
+        ip = ichunk;
+    }
+
+    ms->nextToUpdate = (U32)(iend - ms->window.base);
+    return 0;
+}
+
+
+/* Dictionaries that assign zero probability to symbols that show up causes problems
+ * when FSE encoding. Mark dictionaries with zero probability symbols as FSE_repeat_check
+ * and only dictionaries with 100% valid symbols can be assumed valid.
+ */
+static FSE_repeat ZSTD_dictNCountRepeat(short* normalizedCounter, unsigned dictMaxSymbolValue, unsigned maxSymbolValue)
+{
+    U32 s;
+    if (dictMaxSymbolValue < maxSymbolValue) {
+        return FSE_repeat_check;
+    }
+    for (s = 0; s <= maxSymbolValue; ++s) {
+        if (normalizedCounter[s] == 0) {
+            return FSE_repeat_check;
+        }
+    }
+    return FSE_repeat_valid;
+}
+
+size_t ZSTD_loadCEntropy(ZSTD_compressedBlockState_t* bs, void* workspace,
+                         const void* const dict, size_t dictSize)
+{
+    short offcodeNCount[MaxOff+1];
+    unsigned offcodeMaxValue = MaxOff;
+    const BYTE* dictPtr = (const BYTE*)dict;    /* skip magic num and dict ID */
+    const BYTE* const dictEnd = dictPtr + dictSize;
+    dictPtr += 8;
+    bs->entropy.huf.repeatMode = HUF_repeat_check;
+
+    {   unsigned maxSymbolValue = 255;
+        unsigned hasZeroWeights = 1;
+        size_t const hufHeaderSize = HUF_readCTable((HUF_CElt*)bs->entropy.huf.CTable, &maxSymbolValue, dictPtr,
+            dictEnd-dictPtr, &hasZeroWeights);
+
+        /* We only set the loaded table as valid if it contains all non-zero
+         * weights. Otherwise, we set it to check */
+        if (!hasZeroWeights)
+            bs->entropy.huf.repeatMode = HUF_repeat_valid;
+
+        RETURN_ERROR_IF(HUF_isError(hufHeaderSize), dictionary_corrupted, "");
+        RETURN_ERROR_IF(maxSymbolValue < 255, dictionary_corrupted, "");
+        dictPtr += hufHeaderSize;
+    }
+
+    {   unsigned offcodeLog;
+        size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr);
+        RETURN_ERROR_IF(FSE_isError(offcodeHeaderSize), dictionary_corrupted, "");
+        RETURN_ERROR_IF(offcodeLog > OffFSELog, dictionary_corrupted, "");
+        /* fill all offset symbols to avoid garbage at end of table */
+        RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(
+                bs->entropy.fse.offcodeCTable,
+                offcodeNCount, MaxOff, offcodeLog,
+                workspace, HUF_WORKSPACE_SIZE)),
+            dictionary_corrupted, "");
+        /* Defer checking offcodeMaxValue because we need to know the size of the dictionary content */
+        dictPtr += offcodeHeaderSize;
+    }
+
+    {   short matchlengthNCount[MaxML+1];
+        unsigned matchlengthMaxValue = MaxML, matchlengthLog;
+        size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr);
+        RETURN_ERROR_IF(FSE_isError(matchlengthHeaderSize), dictionary_corrupted, "");
+        RETURN_ERROR_IF(matchlengthLog > MLFSELog, dictionary_corrupted, "");
+        RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(
+                bs->entropy.fse.matchlengthCTable,
+                matchlengthNCount, matchlengthMaxValue, matchlengthLog,
+                workspace, HUF_WORKSPACE_SIZE)),
+            dictionary_corrupted, "");
+        bs->entropy.fse.matchlength_repeatMode = ZSTD_dictNCountRepeat(matchlengthNCount, matchlengthMaxValue, MaxML);
+        dictPtr += matchlengthHeaderSize;
+    }
+
+    {   short litlengthNCount[MaxLL+1];
+        unsigned litlengthMaxValue = MaxLL, litlengthLog;
+        size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr);
+        RETURN_ERROR_IF(FSE_isError(litlengthHeaderSize), dictionary_corrupted, "");
+        RETURN_ERROR_IF(litlengthLog > LLFSELog, dictionary_corrupted, "");
+        RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(
+                bs->entropy.fse.litlengthCTable,
+                litlengthNCount, litlengthMaxValue, litlengthLog,
+                workspace, HUF_WORKSPACE_SIZE)),
+            dictionary_corrupted, "");
+        bs->entropy.fse.litlength_repeatMode = ZSTD_dictNCountRepeat(litlengthNCount, litlengthMaxValue, MaxLL);
+        dictPtr += litlengthHeaderSize;
+    }
+
+    RETURN_ERROR_IF(dictPtr+12 > dictEnd, dictionary_corrupted, "");
+    bs->rep[0] = MEM_readLE32(dictPtr+0);
+    bs->rep[1] = MEM_readLE32(dictPtr+4);
+    bs->rep[2] = MEM_readLE32(dictPtr+8);
+    dictPtr += 12;
+
+    {   size_t const dictContentSize = (size_t)(dictEnd - dictPtr);
+        U32 offcodeMax = MaxOff;
+        if (dictContentSize <= ((U32)-1) - 128 KB) {
+            U32 const maxOffset = (U32)dictContentSize + 128 KB; /* The maximum offset that must be supported */
+            offcodeMax = ZSTD_highbit32(maxOffset); /* Calculate minimum offset code required to represent maxOffset */
+        }
+        /* All offset values <= dictContentSize + 128 KB must be representable for a valid table */
+        bs->entropy.fse.offcode_repeatMode = ZSTD_dictNCountRepeat(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff));
+
+        /* All repCodes must be <= dictContentSize and != 0 */
+        {   U32 u;
+            for (u=0; u<3; u++) {
+                RETURN_ERROR_IF(bs->rep[u] == 0, dictionary_corrupted, "");
+                RETURN_ERROR_IF(bs->rep[u] > dictContentSize, dictionary_corrupted, "");
+    }   }   }
+
+    return dictPtr - (const BYTE*)dict;
+}
+
+/* Dictionary format :
+ * See :
+ * https://github.com/facebook/zstd/blob/release/doc/zstd_compression_format.md#dictionary-format
+ */
+/*! ZSTD_loadZstdDictionary() :
+ * @return : dictID, or an error code
+ *  assumptions : magic number supposed already checked
+ *                dictSize supposed >= 8
+ */
+static size_t ZSTD_loadZstdDictionary(ZSTD_compressedBlockState_t* bs,
+                                      ZSTD_matchState_t* ms,
+                                      ZSTD_cwksp* ws,
+                                      ZSTD_CCtx_params const* params,
+                                      const void* dict, size_t dictSize,
+                                      ZSTD_dictTableLoadMethod_e dtlm,
+                                      void* workspace)
+{
+    const BYTE* dictPtr = (const BYTE*)dict;
+    const BYTE* const dictEnd = dictPtr + dictSize;
+    size_t dictID;
+    size_t eSize;
+
+    ZSTD_STATIC_ASSERT(HUF_WORKSPACE_SIZE >= (1<<MAX(MLFSELog,LLFSELog)));
+    assert(dictSize >= 8);
+    assert(MEM_readLE32(dictPtr) == ZSTD_MAGIC_DICTIONARY);
+
+    dictID = params->fParams.noDictIDFlag ? 0 :  MEM_readLE32(dictPtr + 4 /* skip magic number */ );
+    eSize = ZSTD_loadCEntropy(bs, workspace, dict, dictSize);
+    FORWARD_IF_ERROR(eSize, "ZSTD_loadCEntropy failed");
+    dictPtr += eSize;
+
+    {
+        size_t const dictContentSize = (size_t)(dictEnd - dictPtr);
+        FORWARD_IF_ERROR(ZSTD_loadDictionaryContent(
+            ms, NULL, ws, params, dictPtr, dictContentSize, dtlm), "");
+    }
+    return dictID;
+}
+
+/* ZSTD_compress_insertDictionary() :
+*   @return : dictID, or an error code */
+static size_t
+ZSTD_compress_insertDictionary(ZSTD_compressedBlockState_t* bs,
+                               ZSTD_matchState_t* ms,
+                               ldmState_t* ls,
+                               ZSTD_cwksp* ws,
+                         const ZSTD_CCtx_params* params,
+                         const void* dict, size_t dictSize,
+                               ZSTD_dictContentType_e dictContentType,
+                               ZSTD_dictTableLoadMethod_e dtlm,
+                               void* workspace)
+{
+    DEBUGLOG(4, "ZSTD_compress_insertDictionary (dictSize=%u)", (U32)dictSize);
+    if ((dict==NULL) || (dictSize<8)) {
+        RETURN_ERROR_IF(dictContentType == ZSTD_dct_fullDict, dictionary_wrong, "");
+        return 0;
+    }
+
+    ZSTD_reset_compressedBlockState(bs);
+
+    /* dict restricted modes */
+    if (dictContentType == ZSTD_dct_rawContent)
+        return ZSTD_loadDictionaryContent(ms, ls, ws, params, dict, dictSize, dtlm);
+
+    if (MEM_readLE32(dict) != ZSTD_MAGIC_DICTIONARY) {
+        if (dictContentType == ZSTD_dct_auto) {
+            DEBUGLOG(4, "raw content dictionary detected");
+            return ZSTD_loadDictionaryContent(
+                ms, ls, ws, params, dict, dictSize, dtlm);
+        }
+        RETURN_ERROR_IF(dictContentType == ZSTD_dct_fullDict, dictionary_wrong, "");
+        assert(0);   /* impossible */
+    }
+
+    /* dict as full zstd dictionary */
+    return ZSTD_loadZstdDictionary(
+        bs, ms, ws, params, dict, dictSize, dtlm, workspace);
+}
+
+#define ZSTD_USE_CDICT_PARAMS_SRCSIZE_CUTOFF (128 KB)
+#define ZSTD_USE_CDICT_PARAMS_DICTSIZE_MULTIPLIER (6ULL)
+
+/*! ZSTD_compressBegin_internal() :
+ * @return : 0, or an error code */
+static size_t ZSTD_compressBegin_internal(ZSTD_CCtx* cctx,
+                                    const void* dict, size_t dictSize,
+                                    ZSTD_dictContentType_e dictContentType,
+                                    ZSTD_dictTableLoadMethod_e dtlm,
+                                    const ZSTD_CDict* cdict,
+                                    const ZSTD_CCtx_params* params, U64 pledgedSrcSize,
+                                    ZSTD_buffered_policy_e zbuff)
+{
+    DEBUGLOG(4, "ZSTD_compressBegin_internal: wlog=%u", params->cParams.windowLog);
+    /* params are supposed to be fully validated at this point */
+    assert(!ZSTD_isError(ZSTD_checkCParams(params->cParams)));
+    assert(!((dict) && (cdict)));  /* either dict or cdict, not both */
+    if ( (cdict)
+      && (cdict->dictContentSize > 0)
+      && ( pledgedSrcSize < ZSTD_USE_CDICT_PARAMS_SRCSIZE_CUTOFF
+        || pledgedSrcSize < cdict->dictContentSize * ZSTD_USE_CDICT_PARAMS_DICTSIZE_MULTIPLIER
+        || pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN
+        || cdict->compressionLevel == 0)
+      && (params->attachDictPref != ZSTD_dictForceLoad) ) {
+        return ZSTD_resetCCtx_usingCDict(cctx, cdict, params, pledgedSrcSize, zbuff);
+    }
+
+    FORWARD_IF_ERROR( ZSTD_resetCCtx_internal(cctx, *params, pledgedSrcSize,
+                                     ZSTDcrp_makeClean, zbuff) , "");
+    {   size_t const dictID = cdict ?
+                ZSTD_compress_insertDictionary(
+                        cctx->blockState.prevCBlock, &cctx->blockState.matchState,
+                        &cctx->ldmState, &cctx->workspace, &cctx->appliedParams, cdict->dictContent,
+                        cdict->dictContentSize, cdict->dictContentType, dtlm,
+                        cctx->entropyWorkspace)
+              : ZSTD_compress_insertDictionary(
+                        cctx->blockState.prevCBlock, &cctx->blockState.matchState,
+                        &cctx->ldmState, &cctx->workspace, &cctx->appliedParams, dict, dictSize,
+                        dictContentType, dtlm, cctx->entropyWorkspace);
+        FORWARD_IF_ERROR(dictID, "ZSTD_compress_insertDictionary failed");
+        assert(dictID <= UINT_MAX);
+        cctx->dictID = (U32)dictID;
+        cctx->dictContentSize = cdict ? cdict->dictContentSize : dictSize;
+    }
+    return 0;
+}
+
+size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx,
+                                    const void* dict, size_t dictSize,
+                                    ZSTD_dictContentType_e dictContentType,
+                                    ZSTD_dictTableLoadMethod_e dtlm,
+                                    const ZSTD_CDict* cdict,
+                                    const ZSTD_CCtx_params* params,
+                                    unsigned long long pledgedSrcSize)
+{
+    DEBUGLOG(4, "ZSTD_compressBegin_advanced_internal: wlog=%u", params->cParams.windowLog);
+    /* compression parameters verification and optimization */
+    FORWARD_IF_ERROR( ZSTD_checkCParams(params->cParams) , "");
+    return ZSTD_compressBegin_internal(cctx,
+                                       dict, dictSize, dictContentType, dtlm,
+                                       cdict,
+                                       params, pledgedSrcSize,
+                                       ZSTDb_not_buffered);
+}
+
+/*! ZSTD_compressBegin_advanced() :
+*   @return : 0, or an error code */
+size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx,
+                             const void* dict, size_t dictSize,
+                                   ZSTD_parameters params, unsigned long long pledgedSrcSize)
+{
+    ZSTD_CCtx_params cctxParams;
+    ZSTD_CCtxParams_init_internal(&cctxParams, &params, ZSTD_NO_CLEVEL);
+    return ZSTD_compressBegin_advanced_internal(cctx,
+                                            dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast,
+                                            NULL /*cdict*/,
+                                            &cctxParams, pledgedSrcSize);
+}
+
+size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel)
+{
+    ZSTD_CCtx_params cctxParams;
+    {
+        ZSTD_parameters const params = ZSTD_getParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, dictSize, ZSTD_cpm_noAttachDict);
+        ZSTD_CCtxParams_init_internal(&cctxParams, &params, (compressionLevel == 0) ? ZSTD_CLEVEL_DEFAULT : compressionLevel);
+    }
+    DEBUGLOG(4, "ZSTD_compressBegin_usingDict (dictSize=%u)", (unsigned)dictSize);
+    return ZSTD_compressBegin_internal(cctx, dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, NULL,
+                                       &cctxParams, ZSTD_CONTENTSIZE_UNKNOWN, ZSTDb_not_buffered);
+}
+
+size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel)
+{
+    return ZSTD_compressBegin_usingDict(cctx, NULL, 0, compressionLevel);
+}
+
+
+/*! ZSTD_writeEpilogue() :
+*   Ends a frame.
+*   @return : nb of bytes written into dst (or an error code) */
+static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity)
+{
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* op = ostart;
+    size_t fhSize = 0;
+
+    DEBUGLOG(4, "ZSTD_writeEpilogue");
+    RETURN_ERROR_IF(cctx->stage == ZSTDcs_created, stage_wrong, "init missing");
+
+    /* special case : empty frame */
+    if (cctx->stage == ZSTDcs_init) {
+        fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, &cctx->appliedParams, 0, 0);
+        FORWARD_IF_ERROR(fhSize, "ZSTD_writeFrameHeader failed");
+        dstCapacity -= fhSize;
+        op += fhSize;
+        cctx->stage = ZSTDcs_ongoing;
+    }
+
+    if (cctx->stage != ZSTDcs_ending) {
+        /* write one last empty block, make it the "last" block */
+        U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw)<<1) + 0;
+        RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall, "no room for epilogue");
+        MEM_writeLE32(op, cBlockHeader24);
+        op += ZSTD_blockHeaderSize;
+        dstCapacity -= ZSTD_blockHeaderSize;
+    }
+
+    if (cctx->appliedParams.fParams.checksumFlag) {
+        U32 const checksum = (U32) xxh64_digest(&cctx->xxhState);
+        RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall, "no room for checksum");
+        DEBUGLOG(4, "ZSTD_writeEpilogue: write checksum : %08X", (unsigned)checksum);
+        MEM_writeLE32(op, checksum);
+        op += 4;
+    }
+
+    cctx->stage = ZSTDcs_created;  /* return to "created but no init" status */
+    return op-ostart;
+}
+
+void ZSTD_CCtx_trace(ZSTD_CCtx* cctx, size_t extraCSize)
+{
+    (void)cctx;
+    (void)extraCSize;
+}
+
+size_t ZSTD_compressEnd (ZSTD_CCtx* cctx,
+                         void* dst, size_t dstCapacity,
+                   const void* src, size_t srcSize)
+{
+    size_t endResult;
+    size_t const cSize = ZSTD_compressContinue_internal(cctx,
+                                dst, dstCapacity, src, srcSize,
+                                1 /* frame mode */, 1 /* last chunk */);
+    FORWARD_IF_ERROR(cSize, "ZSTD_compressContinue_internal failed");
+    endResult = ZSTD_writeEpilogue(cctx, (char*)dst + cSize, dstCapacity-cSize);
+    FORWARD_IF_ERROR(endResult, "ZSTD_writeEpilogue failed");
+    assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0));
+    if (cctx->pledgedSrcSizePlusOne != 0) {  /* control src size */
+        ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1);
+        DEBUGLOG(4, "end of frame : controlling src size");
+        RETURN_ERROR_IF(
+            cctx->pledgedSrcSizePlusOne != cctx->consumedSrcSize+1,
+            srcSize_wrong,
+             "error : pledgedSrcSize = %u, while realSrcSize = %u",
+            (unsigned)cctx->pledgedSrcSizePlusOne-1,
+            (unsigned)cctx->consumedSrcSize);
+    }
+    ZSTD_CCtx_trace(cctx, endResult);
+    return cSize + endResult;
+}
+
+size_t ZSTD_compress_advanced (ZSTD_CCtx* cctx,
+                               void* dst, size_t dstCapacity,
+                         const void* src, size_t srcSize,
+                         const void* dict,size_t dictSize,
+                               ZSTD_parameters params)
+{
+    ZSTD_CCtx_params cctxParams;
+    DEBUGLOG(4, "ZSTD_compress_advanced");
+    FORWARD_IF_ERROR(ZSTD_checkCParams(params.cParams), "");
+    ZSTD_CCtxParams_init_internal(&cctxParams, &params, ZSTD_NO_CLEVEL);
+    return ZSTD_compress_advanced_internal(cctx,
+                                           dst, dstCapacity,
+                                           src, srcSize,
+                                           dict, dictSize,
+                                           &cctxParams);
+}
+
+/* Internal */
+size_t ZSTD_compress_advanced_internal(
+        ZSTD_CCtx* cctx,
+        void* dst, size_t dstCapacity,
+        const void* src, size_t srcSize,
+        const void* dict,size_t dictSize,
+        const ZSTD_CCtx_params* params)
+{
+    DEBUGLOG(4, "ZSTD_compress_advanced_internal (srcSize:%u)", (unsigned)srcSize);
+    FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx,
+                         dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, NULL,
+                         params, srcSize, ZSTDb_not_buffered) , "");
+    return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
+}
+
+size_t ZSTD_compress_usingDict(ZSTD_CCtx* cctx,
+                               void* dst, size_t dstCapacity,
+                         const void* src, size_t srcSize,
+                         const void* dict, size_t dictSize,
+                               int compressionLevel)
+{
+    ZSTD_CCtx_params cctxParams;
+    {
+        ZSTD_parameters const params = ZSTD_getParams_internal(compressionLevel, srcSize, dict ? dictSize : 0, ZSTD_cpm_noAttachDict);
+        assert(params.fParams.contentSizeFlag == 1);
+        ZSTD_CCtxParams_init_internal(&cctxParams, &params, (compressionLevel == 0) ? ZSTD_CLEVEL_DEFAULT: compressionLevel);
+    }
+    DEBUGLOG(4, "ZSTD_compress_usingDict (srcSize=%u)", (unsigned)srcSize);
+    return ZSTD_compress_advanced_internal(cctx, dst, dstCapacity, src, srcSize, dict, dictSize, &cctxParams);
+}
+
+size_t ZSTD_compressCCtx(ZSTD_CCtx* cctx,
+                         void* dst, size_t dstCapacity,
+                   const void* src, size_t srcSize,
+                         int compressionLevel)
+{
+    DEBUGLOG(4, "ZSTD_compressCCtx (srcSize=%u)", (unsigned)srcSize);
+    assert(cctx != NULL);
+    return ZSTD_compress_usingDict(cctx, dst, dstCapacity, src, srcSize, NULL, 0, compressionLevel);
+}
+
+size_t ZSTD_compress(void* dst, size_t dstCapacity,
+               const void* src, size_t srcSize,
+                     int compressionLevel)
+{
+    size_t result;
+    ZSTD_CCtx* cctx = ZSTD_createCCtx();
+    RETURN_ERROR_IF(!cctx, memory_allocation, "ZSTD_createCCtx failed");
+    result = ZSTD_compressCCtx(cctx, dst, dstCapacity, src, srcSize, compressionLevel);
+    ZSTD_freeCCtx(cctx);
+    return result;
+}
+
+
+/* =====  Dictionary API  ===== */
+
+/*! ZSTD_estimateCDictSize_advanced() :
+ *  Estimate amount of memory that will be needed to create a dictionary with following arguments */
+size_t ZSTD_estimateCDictSize_advanced(
+        size_t dictSize, ZSTD_compressionParameters cParams,
+        ZSTD_dictLoadMethod_e dictLoadMethod)
+{
+    DEBUGLOG(5, "sizeof(ZSTD_CDict) : %u", (unsigned)sizeof(ZSTD_CDict));
+    return ZSTD_cwksp_alloc_size(sizeof(ZSTD_CDict))
+         + ZSTD_cwksp_alloc_size(HUF_WORKSPACE_SIZE)
+         + ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 0)
+         + (dictLoadMethod == ZSTD_dlm_byRef ? 0
+            : ZSTD_cwksp_alloc_size(ZSTD_cwksp_align(dictSize, sizeof(void *))));
+}
+
+size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel)
+{
+    ZSTD_compressionParameters const cParams = ZSTD_getCParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, dictSize, ZSTD_cpm_createCDict);
+    return ZSTD_estimateCDictSize_advanced(dictSize, cParams, ZSTD_dlm_byCopy);
+}
+
+size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict)
+{
+    if (cdict==NULL) return 0;   /* support sizeof on NULL */
+    DEBUGLOG(5, "sizeof(*cdict) : %u", (unsigned)sizeof(*cdict));
+    /* cdict may be in the workspace */
+    return (cdict->workspace.workspace == cdict ? 0 : sizeof(*cdict))
+        + ZSTD_cwksp_sizeof(&cdict->workspace);
+}
+
+static size_t ZSTD_initCDict_internal(
+                    ZSTD_CDict* cdict,
+              const void* dictBuffer, size_t dictSize,
+                    ZSTD_dictLoadMethod_e dictLoadMethod,
+                    ZSTD_dictContentType_e dictContentType,
+                    ZSTD_CCtx_params params)
+{
+    DEBUGLOG(3, "ZSTD_initCDict_internal (dictContentType:%u)", (unsigned)dictContentType);
+    assert(!ZSTD_checkCParams(params.cParams));
+    cdict->matchState.cParams = params.cParams;
+    cdict->matchState.dedicatedDictSearch = params.enableDedicatedDictSearch;
+    if (cdict->matchState.dedicatedDictSearch && dictSize > ZSTD_CHUNKSIZE_MAX) {
+        cdict->matchState.dedicatedDictSearch = 0;
+    }
+    if ((dictLoadMethod == ZSTD_dlm_byRef) || (!dictBuffer) || (!dictSize)) {
+        cdict->dictContent = dictBuffer;
+    } else {
+         void *internalBuffer = ZSTD_cwksp_reserve_object(&cdict->workspace, ZSTD_cwksp_align(dictSize, sizeof(void*)));
+        RETURN_ERROR_IF(!internalBuffer, memory_allocation, "NULL pointer!");
+        cdict->dictContent = internalBuffer;
+        ZSTD_memcpy(internalBuffer, dictBuffer, dictSize);
+    }
+    cdict->dictContentSize = dictSize;
+    cdict->dictContentType = dictContentType;
+
+    cdict->entropyWorkspace = (U32*)ZSTD_cwksp_reserve_object(&cdict->workspace, HUF_WORKSPACE_SIZE);
+
+
+    /* Reset the state to no dictionary */
+    ZSTD_reset_compressedBlockState(&cdict->cBlockState);
+    FORWARD_IF_ERROR(ZSTD_reset_matchState(
+        &cdict->matchState,
+        &cdict->workspace,
+        &params.cParams,
+        ZSTDcrp_makeClean,
+        ZSTDirp_reset,
+        ZSTD_resetTarget_CDict), "");
+    /* (Maybe) load the dictionary
+     * Skips loading the dictionary if it is < 8 bytes.
+     */
+    {   params.compressionLevel = ZSTD_CLEVEL_DEFAULT;
+        params.fParams.contentSizeFlag = 1;
+        {   size_t const dictID = ZSTD_compress_insertDictionary(
+                    &cdict->cBlockState, &cdict->matchState, NULL, &cdict->workspace,
+                    &params, cdict->dictContent, cdict->dictContentSize,
+                    dictContentType, ZSTD_dtlm_full, cdict->entropyWorkspace);
+            FORWARD_IF_ERROR(dictID, "ZSTD_compress_insertDictionary failed");
+            assert(dictID <= (size_t)(U32)-1);
+            cdict->dictID = (U32)dictID;
+        }
+    }
+
+    return 0;
+}
+
+static ZSTD_CDict* ZSTD_createCDict_advanced_internal(size_t dictSize,
+                                      ZSTD_dictLoadMethod_e dictLoadMethod,
+                                      ZSTD_compressionParameters cParams, ZSTD_customMem customMem)
+{
+    if ((!customMem.customAlloc) ^ (!customMem.customFree)) return NULL;
+
+    {   size_t const workspaceSize =
+            ZSTD_cwksp_alloc_size(sizeof(ZSTD_CDict)) +
+            ZSTD_cwksp_alloc_size(HUF_WORKSPACE_SIZE) +
+            ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 0) +
+            (dictLoadMethod == ZSTD_dlm_byRef ? 0
+             : ZSTD_cwksp_alloc_size(ZSTD_cwksp_align(dictSize, sizeof(void*))));
+        void* const workspace = ZSTD_customMalloc(workspaceSize, customMem);
+        ZSTD_cwksp ws;
+        ZSTD_CDict* cdict;
+
+        if (!workspace) {
+            ZSTD_customFree(workspace, customMem);
+            return NULL;
+        }
+
+        ZSTD_cwksp_init(&ws, workspace, workspaceSize, ZSTD_cwksp_dynamic_alloc);
+
+        cdict = (ZSTD_CDict*)ZSTD_cwksp_reserve_object(&ws, sizeof(ZSTD_CDict));
+        assert(cdict != NULL);
+        ZSTD_cwksp_move(&cdict->workspace, &ws);
+        cdict->customMem = customMem;
+        cdict->compressionLevel = ZSTD_NO_CLEVEL; /* signals advanced API usage */
+
+        return cdict;
+    }
+}
+
+ZSTD_CDict* ZSTD_createCDict_advanced(const void* dictBuffer, size_t dictSize,
+                                      ZSTD_dictLoadMethod_e dictLoadMethod,
+                                      ZSTD_dictContentType_e dictContentType,
+                                      ZSTD_compressionParameters cParams,
+                                      ZSTD_customMem customMem)
+{
+    ZSTD_CCtx_params cctxParams;
+    ZSTD_memset(&cctxParams, 0, sizeof(cctxParams));
+    ZSTD_CCtxParams_init(&cctxParams, 0);
+    cctxParams.cParams = cParams;
+    cctxParams.customMem = customMem;
+    return ZSTD_createCDict_advanced2(
+        dictBuffer, dictSize,
+        dictLoadMethod, dictContentType,
+        &cctxParams, customMem);
+}
+
+ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_advanced2(
+        const void* dict, size_t dictSize,
+        ZSTD_dictLoadMethod_e dictLoadMethod,
+        ZSTD_dictContentType_e dictContentType,
+        const ZSTD_CCtx_params* originalCctxParams,
+        ZSTD_customMem customMem)
+{
+    ZSTD_CCtx_params cctxParams = *originalCctxParams;
+    ZSTD_compressionParameters cParams;
+    ZSTD_CDict* cdict;
+
+    DEBUGLOG(3, "ZSTD_createCDict_advanced2, mode %u", (unsigned)dictContentType);
+    if (!customMem.customAlloc ^ !customMem.customFree) return NULL;
+
+    if (cctxParams.enableDedicatedDictSearch) {
+        cParams = ZSTD_dedicatedDictSearch_getCParams(
+            cctxParams.compressionLevel, dictSize);
+        ZSTD_overrideCParams(&cParams, &cctxParams.cParams);
+    } else {
+        cParams = ZSTD_getCParamsFromCCtxParams(
+            &cctxParams, ZSTD_CONTENTSIZE_UNKNOWN, dictSize, ZSTD_cpm_createCDict);
+    }
+
+    if (!ZSTD_dedicatedDictSearch_isSupported(&cParams)) {
+        /* Fall back to non-DDSS params */
+        cctxParams.enableDedicatedDictSearch = 0;
+        cParams = ZSTD_getCParamsFromCCtxParams(
+            &cctxParams, ZSTD_CONTENTSIZE_UNKNOWN, dictSize, ZSTD_cpm_createCDict);
+    }
+
+    cctxParams.cParams = cParams;
+
+    cdict = ZSTD_createCDict_advanced_internal(dictSize,
+                        dictLoadMethod, cctxParams.cParams,
+                        customMem);
+
+    if (ZSTD_isError( ZSTD_initCDict_internal(cdict,
+                                    dict, dictSize,
+                                    dictLoadMethod, dictContentType,
+                                    cctxParams) )) {
+        ZSTD_freeCDict(cdict);
+        return NULL;
+    }
+
+    return cdict;
+}
+
+ZSTD_CDict* ZSTD_createCDict(const void* dict, size_t dictSize, int compressionLevel)
+{
+    ZSTD_compressionParameters cParams = ZSTD_getCParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, dictSize, ZSTD_cpm_createCDict);
+    ZSTD_CDict* const cdict = ZSTD_createCDict_advanced(dict, dictSize,
+                                                  ZSTD_dlm_byCopy, ZSTD_dct_auto,
+                                                  cParams, ZSTD_defaultCMem);
+    if (cdict)
+        cdict->compressionLevel = (compressionLevel == 0) ? ZSTD_CLEVEL_DEFAULT : compressionLevel;
+    return cdict;
+}
+
+ZSTD_CDict* ZSTD_createCDict_byReference(const void* dict, size_t dictSize, int compressionLevel)
+{
+    ZSTD_compressionParameters cParams = ZSTD_getCParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, dictSize, ZSTD_cpm_createCDict);
+    ZSTD_CDict* const cdict = ZSTD_createCDict_advanced(dict, dictSize,
+                                     ZSTD_dlm_byRef, ZSTD_dct_auto,
+                                     cParams, ZSTD_defaultCMem);
+    if (cdict)
+        cdict->compressionLevel = (compressionLevel == 0) ? ZSTD_CLEVEL_DEFAULT : compressionLevel;
+    return cdict;
+}
+
+size_t ZSTD_freeCDict(ZSTD_CDict* cdict)
+{
+    if (cdict==NULL) return 0;   /* support free on NULL */
+    {   ZSTD_customMem const cMem = cdict->customMem;
+        int cdictInWorkspace = ZSTD_cwksp_owns_buffer(&cdict->workspace, cdict);
+        ZSTD_cwksp_free(&cdict->workspace, cMem);
+        if (!cdictInWorkspace) {
+            ZSTD_customFree(cdict, cMem);
+        }
+        return 0;
+    }
+}
+
+/*! ZSTD_initStaticCDict_advanced() :
+ *  Generate a digested dictionary in provided memory area.
+ *  workspace: The memory area to emplace the dictionary into.
+ *             Provided pointer must 8-bytes aligned.
+ *             It must outlive dictionary usage.
+ *  workspaceSize: Use ZSTD_estimateCDictSize()
+ *                 to determine how large workspace must be.
+ *  cParams : use ZSTD_getCParams() to transform a compression level
+ *            into its relevants cParams.
+ * @return : pointer to ZSTD_CDict*, or NULL if error (size too small)
+ *  Note : there is no corresponding "free" function.
+ *         Since workspace was allocated externally, it must be freed externally.
+ */
+const ZSTD_CDict* ZSTD_initStaticCDict(
+                                 void* workspace, size_t workspaceSize,
+                           const void* dict, size_t dictSize,
+                                 ZSTD_dictLoadMethod_e dictLoadMethod,
+                                 ZSTD_dictContentType_e dictContentType,
+                                 ZSTD_compressionParameters cParams)
+{
+    size_t const matchStateSize = ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 0);
+    size_t const neededSize = ZSTD_cwksp_alloc_size(sizeof(ZSTD_CDict))
+                            + (dictLoadMethod == ZSTD_dlm_byRef ? 0
+                               : ZSTD_cwksp_alloc_size(ZSTD_cwksp_align(dictSize, sizeof(void*))))
+                            + ZSTD_cwksp_alloc_size(HUF_WORKSPACE_SIZE)
+                            + matchStateSize;
+    ZSTD_CDict* cdict;
+    ZSTD_CCtx_params params;
+
+    if ((size_t)workspace & 7) return NULL;  /* 8-aligned */
+
+    {
+        ZSTD_cwksp ws;
+        ZSTD_cwksp_init(&ws, workspace, workspaceSize, ZSTD_cwksp_static_alloc);
+        cdict = (ZSTD_CDict*)ZSTD_cwksp_reserve_object(&ws, sizeof(ZSTD_CDict));
+        if (cdict == NULL) return NULL;
+        ZSTD_cwksp_move(&cdict->workspace, &ws);
+    }
+
+    DEBUGLOG(4, "(workspaceSize < neededSize) : (%u < %u) => %u",
+        (unsigned)workspaceSize, (unsigned)neededSize, (unsigned)(workspaceSize < neededSize));
+    if (workspaceSize < neededSize) return NULL;
+
+    ZSTD_CCtxParams_init(&params, 0);
+    params.cParams = cParams;
+
+    if (ZSTD_isError( ZSTD_initCDict_internal(cdict,
+                                              dict, dictSize,
+                                              dictLoadMethod, dictContentType,
+                                              params) ))
+        return NULL;
+
+    return cdict;
+}
+
+ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict)
+{
+    assert(cdict != NULL);
+    return cdict->matchState.cParams;
+}
+
+/*! ZSTD_getDictID_fromCDict() :
+ *  Provides the dictID of the dictionary loaded into `cdict`.
+ *  If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
+ *  Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */
+unsigned ZSTD_getDictID_fromCDict(const ZSTD_CDict* cdict)
+{
+    if (cdict==NULL) return 0;
+    return cdict->dictID;
+}
+
+
+/* ZSTD_compressBegin_usingCDict_advanced() :
+ * cdict must be != NULL */
+size_t ZSTD_compressBegin_usingCDict_advanced(
+    ZSTD_CCtx* const cctx, const ZSTD_CDict* const cdict,
+    ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize)
+{
+    ZSTD_CCtx_params cctxParams;
+    DEBUGLOG(4, "ZSTD_compressBegin_usingCDict_advanced");
+    RETURN_ERROR_IF(cdict==NULL, dictionary_wrong, "NULL pointer!");
+    /* Initialize the cctxParams from the cdict */
+    {
+        ZSTD_parameters params;
+        params.fParams = fParams;
+        params.cParams = ( pledgedSrcSize < ZSTD_USE_CDICT_PARAMS_SRCSIZE_CUTOFF
+                        || pledgedSrcSize < cdict->dictContentSize * ZSTD_USE_CDICT_PARAMS_DICTSIZE_MULTIPLIER
+                        || pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN
+                        || cdict->compressionLevel == 0 ) ?
+                ZSTD_getCParamsFromCDict(cdict)
+              : ZSTD_getCParams(cdict->compressionLevel,
+                                pledgedSrcSize,
+                                cdict->dictContentSize);
+        ZSTD_CCtxParams_init_internal(&cctxParams, &params, cdict->compressionLevel);
+    }
+    /* Increase window log to fit the entire dictionary and source if the
+     * source size is known. Limit the increase to 19, which is the
+     * window log for compression level 1 with the largest source size.
+     */
+    if (pledgedSrcSize != ZSTD_CONTENTSIZE_UNKNOWN) {
+        U32 const limitedSrcSize = (U32)MIN(pledgedSrcSize, 1U << 19);
+        U32 const limitedSrcLog = limitedSrcSize > 1 ? ZSTD_highbit32(limitedSrcSize - 1) + 1 : 1;
+        cctxParams.cParams.windowLog = MAX(cctxParams.cParams.windowLog, limitedSrcLog);
+    }
+    return ZSTD_compressBegin_internal(cctx,
+                                        NULL, 0, ZSTD_dct_auto, ZSTD_dtlm_fast,
+                                        cdict,
+                                        &cctxParams, pledgedSrcSize,
+                                        ZSTDb_not_buffered);
+}
+
+/* ZSTD_compressBegin_usingCDict() :
+ * pledgedSrcSize=0 means "unknown"
+ * if pledgedSrcSize>0, it will enable contentSizeFlag */
+size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict)
+{
+    ZSTD_frameParameters const fParams = { 0 /*content*/, 0 /*checksum*/, 0 /*noDictID*/ };
+    DEBUGLOG(4, "ZSTD_compressBegin_usingCDict : dictIDFlag == %u", !fParams.noDictIDFlag);
+    return ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, ZSTD_CONTENTSIZE_UNKNOWN);
+}
+
+size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx,
+                                void* dst, size_t dstCapacity,
+                                const void* src, size_t srcSize,
+                                const ZSTD_CDict* cdict, ZSTD_frameParameters fParams)
+{
+    FORWARD_IF_ERROR(ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, srcSize), "");   /* will check if cdict != NULL */
+    return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
+}
+
+/*! ZSTD_compress_usingCDict() :
+ *  Compression using a digested Dictionary.
+ *  Faster startup than ZSTD_compress_usingDict(), recommended when same dictionary is used multiple times.
+ *  Note that compression parameters are decided at CDict creation time
+ *  while frame parameters are hardcoded */
+size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,
+                                void* dst, size_t dstCapacity,
+                                const void* src, size_t srcSize,
+                                const ZSTD_CDict* cdict)
+{
+    ZSTD_frameParameters const fParams = { 1 /*content*/, 0 /*checksum*/, 0 /*noDictID*/ };
+    return ZSTD_compress_usingCDict_advanced(cctx, dst, dstCapacity, src, srcSize, cdict, fParams);
+}
+
+
+
+/* ******************************************************************
+*  Streaming
+********************************************************************/
+
+ZSTD_CStream* ZSTD_createCStream(void)
+{
+    DEBUGLOG(3, "ZSTD_createCStream");
+    return ZSTD_createCStream_advanced(ZSTD_defaultCMem);
+}
+
+ZSTD_CStream* ZSTD_initStaticCStream(void *workspace, size_t workspaceSize)
+{
+    return ZSTD_initStaticCCtx(workspace, workspaceSize);
+}
+
+ZSTD_CStream* ZSTD_createCStream_advanced(ZSTD_customMem customMem)
+{   /* CStream and CCtx are now same object */
+    return ZSTD_createCCtx_advanced(customMem);
+}
+
+size_t ZSTD_freeCStream(ZSTD_CStream* zcs)
+{
+    return ZSTD_freeCCtx(zcs);   /* same object */
+}
+
+
+
+/*======   Initialization   ======*/
+
+size_t ZSTD_CStreamInSize(void)  { return ZSTD_BLOCKSIZE_MAX; }
+
+size_t ZSTD_CStreamOutSize(void)
+{
+    return ZSTD_compressBound(ZSTD_BLOCKSIZE_MAX) + ZSTD_blockHeaderSize + 4 /* 32-bits hash */ ;
+}
+
+static ZSTD_cParamMode_e ZSTD_getCParamMode(ZSTD_CDict const* cdict, ZSTD_CCtx_params const* params, U64 pledgedSrcSize)
+{
+    if (cdict != NULL && ZSTD_shouldAttachDict(cdict, params, pledgedSrcSize))
+        return ZSTD_cpm_attachDict;
+    else
+        return ZSTD_cpm_noAttachDict;
+}
+
+/* ZSTD_resetCStream():
+ * pledgedSrcSize == 0 means "unknown" */
+size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pss)
+{
+    /* temporary : 0 interpreted as "unknown" during transition period.
+     * Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN.
+     * 0 will be interpreted as "empty" in the future.
+     */
+    U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss;
+    DEBUGLOG(4, "ZSTD_resetCStream: pledgedSrcSize = %u", (unsigned)pledgedSrcSize);
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) , "");
+    return 0;
+}
+
+/*! ZSTD_initCStream_internal() :
+ *  Note : for lib/compress only. Used by zstdmt_compress.c.
+ *  Assumption 1 : params are valid
+ *  Assumption 2 : either dict, or cdict, is defined, not both */
+size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,
+                    const void* dict, size_t dictSize, const ZSTD_CDict* cdict,
+                    const ZSTD_CCtx_params* params,
+                    unsigned long long pledgedSrcSize)
+{
+    DEBUGLOG(4, "ZSTD_initCStream_internal");
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) , "");
+    assert(!ZSTD_isError(ZSTD_checkCParams(params->cParams)));
+    zcs->requestedParams = *params;
+    assert(!((dict) && (cdict)));  /* either dict or cdict, not both */
+    if (dict) {
+        FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) , "");
+    } else {
+        /* Dictionary is cleared if !cdict */
+        FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) , "");
+    }
+    return 0;
+}
+
+/* ZSTD_initCStream_usingCDict_advanced() :
+ * same as ZSTD_initCStream_usingCDict(), with control over frame parameters */
+size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs,
+                                            const ZSTD_CDict* cdict,
+                                            ZSTD_frameParameters fParams,
+                                            unsigned long long pledgedSrcSize)
+{
+    DEBUGLOG(4, "ZSTD_initCStream_usingCDict_advanced");
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) , "");
+    zcs->requestedParams.fParams = fParams;
+    FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) , "");
+    return 0;
+}
+
+/* note : cdict must outlive compression session */
+size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict)
+{
+    DEBUGLOG(4, "ZSTD_initCStream_usingCDict");
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) , "");
+    return 0;
+}
+
+
+/* ZSTD_initCStream_advanced() :
+ * pledgedSrcSize must be exact.
+ * if srcSize is not known at init time, use value ZSTD_CONTENTSIZE_UNKNOWN.
+ * dict is loaded with default parameters ZSTD_dct_auto and ZSTD_dlm_byCopy. */
+size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
+                                 const void* dict, size_t dictSize,
+                                 ZSTD_parameters params, unsigned long long pss)
+{
+    /* for compatibility with older programs relying on this behavior.
+     * Users should now specify ZSTD_CONTENTSIZE_UNKNOWN.
+     * This line will be removed in the future.
+     */
+    U64 const pledgedSrcSize = (pss==0 && params.fParams.contentSizeFlag==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss;
+    DEBUGLOG(4, "ZSTD_initCStream_advanced");
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) , "");
+    FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) , "");
+    ZSTD_CCtxParams_setZstdParams(&zcs->requestedParams, &params);
+    FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) , "");
+    return 0;
+}
+
+size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel)
+{
+    DEBUGLOG(4, "ZSTD_initCStream_usingDict");
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) , "");
+    return 0;
+}
+
+size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pss)
+{
+    /* temporary : 0 interpreted as "unknown" during transition period.
+     * Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN.
+     * 0 will be interpreted as "empty" in the future.
+     */
+    U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss;
+    DEBUGLOG(4, "ZSTD_initCStream_srcSize");
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, NULL) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) , "");
+    return 0;
+}
+
+size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel)
+{
+    DEBUGLOG(4, "ZSTD_initCStream");
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, NULL) , "");
+    FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) , "");
+    return 0;
+}
+
+/*======   Compression   ======*/
+
+static size_t ZSTD_nextInputSizeHint(const ZSTD_CCtx* cctx)
+{
+    size_t hintInSize = cctx->inBuffTarget - cctx->inBuffPos;
+    if (hintInSize==0) hintInSize = cctx->blockSize;
+    return hintInSize;
+}
+
+/* ZSTD_compressStream_generic():
+ *  internal function for all *compressStream*() variants
+ *  non-static, because can be called from zstdmt_compress.c
+ * @return : hint size for next input */
+static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
+                                          ZSTD_outBuffer* output,
+                                          ZSTD_inBuffer* input,
+                                          ZSTD_EndDirective const flushMode)
+{
+    const char* const istart = (const char*)input->src;
+    const char* const iend = input->size != 0 ? istart + input->size : istart;
+    const char* ip = input->pos != 0 ? istart + input->pos : istart;
+    char* const ostart = (char*)output->dst;
+    char* const oend = output->size != 0 ? ostart + output->size : ostart;
+    char* op = output->pos != 0 ? ostart + output->pos : ostart;
+    U32 someMoreWork = 1;
+
+    /* check expectations */
+    DEBUGLOG(5, "ZSTD_compressStream_generic, flush=%u", (unsigned)flushMode);
+    if (zcs->appliedParams.inBufferMode == ZSTD_bm_buffered) {
+        assert(zcs->inBuff != NULL);
+        assert(zcs->inBuffSize > 0);
+    }
+    if (zcs->appliedParams.outBufferMode == ZSTD_bm_buffered) {
+        assert(zcs->outBuff !=  NULL);
+        assert(zcs->outBuffSize > 0);
+    }
+    assert(output->pos <= output->size);
+    assert(input->pos <= input->size);
+    assert((U32)flushMode <= (U32)ZSTD_e_end);
+
+    while (someMoreWork) {
+        switch(zcs->streamStage)
+        {
+        case zcss_init:
+            RETURN_ERROR(init_missing, "call ZSTD_initCStream() first!");
+
+        case zcss_load:
+            if ( (flushMode == ZSTD_e_end)
+              && ( (size_t)(oend-op) >= ZSTD_compressBound(iend-ip)     /* Enough output space */
+                || zcs->appliedParams.outBufferMode == ZSTD_bm_stable)  /* OR we are allowed to return dstSizeTooSmall */
+              && (zcs->inBuffPos == 0) ) {
+                /* shortcut to compression pass directly into output buffer */
+                size_t const cSize = ZSTD_compressEnd(zcs,
+                                                op, oend-op, ip, iend-ip);
+                DEBUGLOG(4, "ZSTD_compressEnd : cSize=%u", (unsigned)cSize);
+                FORWARD_IF_ERROR(cSize, "ZSTD_compressEnd failed");
+                ip = iend;
+                op += cSize;
+                zcs->frameEnded = 1;
+                ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+                someMoreWork = 0; break;
+            }
+            /* complete loading into inBuffer in buffered mode */
+            if (zcs->appliedParams.inBufferMode == ZSTD_bm_buffered) {
+                size_t const toLoad = zcs->inBuffTarget - zcs->inBuffPos;
+                size_t const loaded = ZSTD_limitCopy(
+                                        zcs->inBuff + zcs->inBuffPos, toLoad,
+                                        ip, iend-ip);
+                zcs->inBuffPos += loaded;
+                if (loaded != 0)
+                    ip += loaded;
+                if ( (flushMode == ZSTD_e_continue)
+                  && (zcs->inBuffPos < zcs->inBuffTarget) ) {
+                    /* not enough input to fill full block : stop here */
+                    someMoreWork = 0; break;
+                }
+                if ( (flushMode == ZSTD_e_flush)
+                  && (zcs->inBuffPos == zcs->inToCompress) ) {
+                    /* empty */
+                    someMoreWork = 0; break;
+                }
+            }
+            /* compress current block (note : this stage cannot be stopped in the middle) */
+            DEBUGLOG(5, "stream compression stage (flushMode==%u)", flushMode);
+            {   int const inputBuffered = (zcs->appliedParams.inBufferMode == ZSTD_bm_buffered);
+                void* cDst;
+                size_t cSize;
+                size_t oSize = oend-op;
+                size_t const iSize = inputBuffered
+                    ? zcs->inBuffPos - zcs->inToCompress
+                    : MIN((size_t)(iend - ip), zcs->blockSize);
+                if (oSize >= ZSTD_compressBound(iSize) || zcs->appliedParams.outBufferMode == ZSTD_bm_stable)
+                    cDst = op;   /* compress into output buffer, to skip flush stage */
+                else
+                    cDst = zcs->outBuff, oSize = zcs->outBuffSize;
+                if (inputBuffered) {
+                    unsigned const lastBlock = (flushMode == ZSTD_e_end) && (ip==iend);
+                    cSize = lastBlock ?
+                            ZSTD_compressEnd(zcs, cDst, oSize,
+                                        zcs->inBuff + zcs->inToCompress, iSize) :
+                            ZSTD_compressContinue(zcs, cDst, oSize,
+                                        zcs->inBuff + zcs->inToCompress, iSize);
+                    FORWARD_IF_ERROR(cSize, "%s", lastBlock ? "ZSTD_compressEnd failed" : "ZSTD_compressContinue failed");
+                    zcs->frameEnded = lastBlock;
+                    /* prepare next block */
+                    zcs->inBuffTarget = zcs->inBuffPos + zcs->blockSize;
+                    if (zcs->inBuffTarget > zcs->inBuffSize)
+                        zcs->inBuffPos = 0, zcs->inBuffTarget = zcs->blockSize;
+                    DEBUGLOG(5, "inBuffTarget:%u / inBuffSize:%u",
+                            (unsigned)zcs->inBuffTarget, (unsigned)zcs->inBuffSize);
+                    if (!lastBlock)
+                        assert(zcs->inBuffTarget <= zcs->inBuffSize);
+                    zcs->inToCompress = zcs->inBuffPos;
+                } else {
+                    unsigned const lastBlock = (ip + iSize == iend);
+                    assert(flushMode == ZSTD_e_end /* Already validated */);
+                    cSize = lastBlock ?
+                            ZSTD_compressEnd(zcs, cDst, oSize, ip, iSize) :
+                            ZSTD_compressContinue(zcs, cDst, oSize, ip, iSize);
+                    /* Consume the input prior to error checking to mirror buffered mode. */
+                    if (iSize > 0)
+                        ip += iSize;
+                    FORWARD_IF_ERROR(cSize, "%s", lastBlock ? "ZSTD_compressEnd failed" : "ZSTD_compressContinue failed");
+                    zcs->frameEnded = lastBlock;
+                    if (lastBlock)
+                        assert(ip == iend);
+                }
+                if (cDst == op) {  /* no need to flush */
+                    op += cSize;
+                    if (zcs->frameEnded) {
+                        DEBUGLOG(5, "Frame completed directly in outBuffer");
+                        someMoreWork = 0;
+                        ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+                    }
+                    break;
+                }
+                zcs->outBuffContentSize = cSize;
+                zcs->outBuffFlushedSize = 0;
+                zcs->streamStage = zcss_flush; /* pass-through to flush stage */
+            }
+	    ZSTD_FALLTHROUGH;
+        case zcss_flush:
+            DEBUGLOG(5, "flush stage");
+            assert(zcs->appliedParams.outBufferMode == ZSTD_bm_buffered);
+            {   size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;
+                size_t const flushed = ZSTD_limitCopy(op, (size_t)(oend-op),
+                            zcs->outBuff + zcs->outBuffFlushedSize, toFlush);
+                DEBUGLOG(5, "toFlush: %u into %u ==> flushed: %u",
+                            (unsigned)toFlush, (unsigned)(oend-op), (unsigned)flushed);
+                if (flushed)
+                    op += flushed;
+                zcs->outBuffFlushedSize += flushed;
+                if (toFlush!=flushed) {
+                    /* flush not fully completed, presumably because dst is too small */
+                    assert(op==oend);
+                    someMoreWork = 0;
+                    break;
+                }
+                zcs->outBuffContentSize = zcs->outBuffFlushedSize = 0;
+                if (zcs->frameEnded) {
+                    DEBUGLOG(5, "Frame completed on flush");
+                    someMoreWork = 0;
+                    ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+                    break;
+                }
+                zcs->streamStage = zcss_load;
+                break;
+            }
+
+        default: /* impossible */
+            assert(0);
+        }
+    }
+
+    input->pos = ip - istart;
+    output->pos = op - ostart;
+    if (zcs->frameEnded) return 0;
+    return ZSTD_nextInputSizeHint(zcs);
+}
+
+static size_t ZSTD_nextInputSizeHint_MTorST(const ZSTD_CCtx* cctx)
+{
+    return ZSTD_nextInputSizeHint(cctx);
+
+}
+
+size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input)
+{
+    FORWARD_IF_ERROR( ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue) , "");
+    return ZSTD_nextInputSizeHint_MTorST(zcs);
+}
+
+/* After a compression call set the expected input/output buffer.
+ * This is validated at the start of the next compression call.
+ */
+static void ZSTD_setBufferExpectations(ZSTD_CCtx* cctx, ZSTD_outBuffer const* output, ZSTD_inBuffer const* input)
+{
+    if (cctx->appliedParams.inBufferMode == ZSTD_bm_stable) {
+        cctx->expectedInBuffer = *input;
+    }
+    if (cctx->appliedParams.outBufferMode == ZSTD_bm_stable) {
+        cctx->expectedOutBufferSize = output->size - output->pos;
+    }
+}
+
+/* Validate that the input/output buffers match the expectations set by
+ * ZSTD_setBufferExpectations.
+ */
+static size_t ZSTD_checkBufferStability(ZSTD_CCtx const* cctx,
+                                        ZSTD_outBuffer const* output,
+                                        ZSTD_inBuffer const* input,
+                                        ZSTD_EndDirective endOp)
+{
+    if (cctx->appliedParams.inBufferMode == ZSTD_bm_stable) {
+        ZSTD_inBuffer const expect = cctx->expectedInBuffer;
+        if (expect.src != input->src || expect.pos != input->pos || expect.size != input->size)
+            RETURN_ERROR(srcBuffer_wrong, "ZSTD_c_stableInBuffer enabled but input differs!");
+        if (endOp != ZSTD_e_end)
+            RETURN_ERROR(srcBuffer_wrong, "ZSTD_c_stableInBuffer can only be used with ZSTD_e_end!");
+    }
+    if (cctx->appliedParams.outBufferMode == ZSTD_bm_stable) {
+        size_t const outBufferSize = output->size - output->pos;
+        if (cctx->expectedOutBufferSize != outBufferSize)
+            RETURN_ERROR(dstBuffer_wrong, "ZSTD_c_stableOutBuffer enabled but output size differs!");
+    }
+    return 0;
+}
+
+static size_t ZSTD_CCtx_init_compressStream2(ZSTD_CCtx* cctx,
+                                             ZSTD_EndDirective endOp,
+                                             size_t inSize) {
+    ZSTD_CCtx_params params = cctx->requestedParams;
+    ZSTD_prefixDict const prefixDict = cctx->prefixDict;
+    FORWARD_IF_ERROR( ZSTD_initLocalDict(cctx) , ""); /* Init the local dict if present. */
+    ZSTD_memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict));   /* single usage */
+    assert(prefixDict.dict==NULL || cctx->cdict==NULL);    /* only one can be set */
+    if (cctx->cdict)
+        params.compressionLevel = cctx->cdict->compressionLevel; /* let cdict take priority in terms of compression level */
+    DEBUGLOG(4, "ZSTD_compressStream2 : transparent init stage");
+    if (endOp == ZSTD_e_end) cctx->pledgedSrcSizePlusOne = inSize + 1;  /* auto-fix pledgedSrcSize */
+    {
+        size_t const dictSize = prefixDict.dict
+                ? prefixDict.dictSize
+                : (cctx->cdict ? cctx->cdict->dictContentSize : 0);
+        ZSTD_cParamMode_e const mode = ZSTD_getCParamMode(cctx->cdict, &params, cctx->pledgedSrcSizePlusOne - 1);
+        params.cParams = ZSTD_getCParamsFromCCtxParams(
+                &params, cctx->pledgedSrcSizePlusOne-1,
+                dictSize, mode);
+    }
+
+    if (ZSTD_CParams_shouldEnableLdm(&params.cParams)) {
+        /* Enable LDM by default for optimal parser and window size >= 128MB */
+        DEBUGLOG(4, "LDM enabled by default (window size >= 128MB, strategy >= btopt)");
+        params.ldmParams.enableLdm = 1;
+    }
+
+    {   U64 const pledgedSrcSize = cctx->pledgedSrcSizePlusOne - 1;
+        assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
+        FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx,
+                prefixDict.dict, prefixDict.dictSize, prefixDict.dictContentType, ZSTD_dtlm_fast,
+                cctx->cdict,
+                &params, pledgedSrcSize,
+                ZSTDb_buffered) , "");
+        assert(cctx->appliedParams.nbWorkers == 0);
+        cctx->inToCompress = 0;
+        cctx->inBuffPos = 0;
+        if (cctx->appliedParams.inBufferMode == ZSTD_bm_buffered) {
+            /* for small input: avoid automatic flush on reaching end of block, since
+            * it would require to add a 3-bytes null block to end frame
+            */
+            cctx->inBuffTarget = cctx->blockSize + (cctx->blockSize == pledgedSrcSize);
+        } else {
+            cctx->inBuffTarget = 0;
+        }
+        cctx->outBuffContentSize = cctx->outBuffFlushedSize = 0;
+        cctx->streamStage = zcss_load;
+        cctx->frameEnded = 0;
+    }
+    return 0;
+}
+
+size_t ZSTD_compressStream2( ZSTD_CCtx* cctx,
+                             ZSTD_outBuffer* output,
+                             ZSTD_inBuffer* input,
+                             ZSTD_EndDirective endOp)
+{
+    DEBUGLOG(5, "ZSTD_compressStream2, endOp=%u ", (unsigned)endOp);
+    /* check conditions */
+    RETURN_ERROR_IF(output->pos > output->size, dstSize_tooSmall, "invalid output buffer");
+    RETURN_ERROR_IF(input->pos  > input->size, srcSize_wrong, "invalid input buffer");
+    RETURN_ERROR_IF((U32)endOp > (U32)ZSTD_e_end, parameter_outOfBound, "invalid endDirective");
+    assert(cctx != NULL);
+
+    /* transparent initialization stage */
+    if (cctx->streamStage == zcss_init) {
+        FORWARD_IF_ERROR(ZSTD_CCtx_init_compressStream2(cctx, endOp, input->size), "CompressStream2 initialization failed");
+        ZSTD_setBufferExpectations(cctx, output, input);    /* Set initial buffer expectations now that we've initialized */
+    }
+    /* end of transparent initialization stage */
+
+    FORWARD_IF_ERROR(ZSTD_checkBufferStability(cctx, output, input, endOp), "invalid buffers");
+    /* compression stage */
+    FORWARD_IF_ERROR( ZSTD_compressStream_generic(cctx, output, input, endOp) , "");
+    DEBUGLOG(5, "completed ZSTD_compressStream2");
+    ZSTD_setBufferExpectations(cctx, output, input);
+    return cctx->outBuffContentSize - cctx->outBuffFlushedSize; /* remaining to flush */
+}
+
+size_t ZSTD_compressStream2_simpleArgs (
+                            ZSTD_CCtx* cctx,
+                            void* dst, size_t dstCapacity, size_t* dstPos,
+                      const void* src, size_t srcSize, size_t* srcPos,
+                            ZSTD_EndDirective endOp)
+{
+    ZSTD_outBuffer output = { dst, dstCapacity, *dstPos };
+    ZSTD_inBuffer  input  = { src, srcSize, *srcPos };
+    /* ZSTD_compressStream2() will check validity of dstPos and srcPos */
+    size_t const cErr = ZSTD_compressStream2(cctx, &output, &input, endOp);
+    *dstPos = output.pos;
+    *srcPos = input.pos;
+    return cErr;
+}
+
+size_t ZSTD_compress2(ZSTD_CCtx* cctx,
+                      void* dst, size_t dstCapacity,
+                      const void* src, size_t srcSize)
+{
+    ZSTD_bufferMode_e const originalInBufferMode = cctx->requestedParams.inBufferMode;
+    ZSTD_bufferMode_e const originalOutBufferMode = cctx->requestedParams.outBufferMode;
+    DEBUGLOG(4, "ZSTD_compress2 (srcSize=%u)", (unsigned)srcSize);
+    ZSTD_CCtx_reset(cctx, ZSTD_reset_session_only);
+    /* Enable stable input/output buffers. */
+    cctx->requestedParams.inBufferMode = ZSTD_bm_stable;
+    cctx->requestedParams.outBufferMode = ZSTD_bm_stable;
+    {   size_t oPos = 0;
+        size_t iPos = 0;
+        size_t const result = ZSTD_compressStream2_simpleArgs(cctx,
+                                        dst, dstCapacity, &oPos,
+                                        src, srcSize, &iPos,
+                                        ZSTD_e_end);
+        /* Reset to the original values. */
+        cctx->requestedParams.inBufferMode = originalInBufferMode;
+        cctx->requestedParams.outBufferMode = originalOutBufferMode;
+        FORWARD_IF_ERROR(result, "ZSTD_compressStream2_simpleArgs failed");
+        if (result != 0) {  /* compression not completed, due to lack of output space */
+            assert(oPos == dstCapacity);
+            RETURN_ERROR(dstSize_tooSmall, "");
+        }
+        assert(iPos == srcSize);   /* all input is expected consumed */
+        return oPos;
+    }
+}
+
+typedef struct {
+    U32 idx;             /* Index in array of ZSTD_Sequence */
+    U32 posInSequence;   /* Position within sequence at idx */
+    size_t posInSrc;        /* Number of bytes given by sequences provided so far */
+} ZSTD_sequencePosition;
+
+/* Returns a ZSTD error code if sequence is not valid */
+static size_t ZSTD_validateSequence(U32 offCode, U32 matchLength,
+                                    size_t posInSrc, U32 windowLog, size_t dictSize, U32 minMatch) {
+    size_t offsetBound;
+    U32 windowSize = 1 << windowLog;
+    /* posInSrc represents the amount of data the the decoder would decode up to this point.
+     * As long as the amount of data decoded is less than or equal to window size, offsets may be
+     * larger than the total length of output decoded in order to reference the dict, even larger than
+     * window size. After output surpasses windowSize, we're limited to windowSize offsets again.
+     */
+    offsetBound = posInSrc > windowSize ? (size_t)windowSize : posInSrc + (size_t)dictSize;
+    RETURN_ERROR_IF(offCode > offsetBound + ZSTD_REP_MOVE, corruption_detected, "Offset too large!");
+    RETURN_ERROR_IF(matchLength < minMatch, corruption_detected, "Matchlength too small");
+    return 0;
+}
+
+/* Returns an offset code, given a sequence's raw offset, the ongoing repcode array, and whether litLength == 0 */
+static U32 ZSTD_finalizeOffCode(U32 rawOffset, const U32 rep[ZSTD_REP_NUM], U32 ll0) {
+    U32 offCode = rawOffset + ZSTD_REP_MOVE;
+    U32 repCode = 0;
+
+    if (!ll0 && rawOffset == rep[0]) {
+        repCode = 1;
+    } else if (rawOffset == rep[1]) {
+        repCode = 2 - ll0;
+    } else if (rawOffset == rep[2]) {
+        repCode = 3 - ll0;
+    } else if (ll0 && rawOffset == rep[0] - 1) {
+        repCode = 3;
+    }
+    if (repCode) {
+        /* ZSTD_storeSeq expects a number in the range [0, 2] to represent a repcode */
+        offCode = repCode - 1;
+    }
+    return offCode;
+}
+
+/* Returns 0 on success, and a ZSTD_error otherwise. This function scans through an array of
+ * ZSTD_Sequence, storing the sequences it finds, until it reaches a block delimiter.
+ */
+static size_t ZSTD_copySequencesToSeqStoreExplicitBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos,
+                                                             const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
+                                                             const void* src, size_t blockSize) {
+    U32 idx = seqPos->idx;
+    BYTE const* ip = (BYTE const*)(src);
+    const BYTE* const iend = ip + blockSize;
+    repcodes_t updatedRepcodes;
+    U32 dictSize;
+    U32 litLength;
+    U32 matchLength;
+    U32 ll0;
+    U32 offCode;
+
+    if (cctx->cdict) {
+        dictSize = (U32)cctx->cdict->dictContentSize;
+    } else if (cctx->prefixDict.dict) {
+        dictSize = (U32)cctx->prefixDict.dictSize;
+    } else {
+        dictSize = 0;
+    }
+    ZSTD_memcpy(updatedRepcodes.rep, cctx->blockState.prevCBlock->rep, sizeof(repcodes_t));
+    for (; (inSeqs[idx].matchLength != 0 || inSeqs[idx].offset != 0) && idx < inSeqsSize; ++idx) {
+        litLength = inSeqs[idx].litLength;
+        matchLength = inSeqs[idx].matchLength;
+        ll0 = litLength == 0;
+        offCode = ZSTD_finalizeOffCode(inSeqs[idx].offset, updatedRepcodes.rep, ll0);
+        updatedRepcodes = ZSTD_updateRep(updatedRepcodes.rep, offCode, ll0);
+
+        DEBUGLOG(6, "Storing sequence: (of: %u, ml: %u, ll: %u)", offCode, matchLength, litLength);
+        if (cctx->appliedParams.validateSequences) {
+            seqPos->posInSrc += litLength + matchLength;
+            FORWARD_IF_ERROR(ZSTD_validateSequence(offCode, matchLength, seqPos->posInSrc,
+                                                cctx->appliedParams.cParams.windowLog, dictSize,
+                                                cctx->appliedParams.cParams.minMatch),
+                                                "Sequence validation failed");
+        }
+        RETURN_ERROR_IF(idx - seqPos->idx > cctx->seqStore.maxNbSeq, memory_allocation,
+                        "Not enough memory allocated. Try adjusting ZSTD_c_minMatch.");
+        ZSTD_storeSeq(&cctx->seqStore, litLength, ip, iend, offCode, matchLength - MINMATCH);
+        ip += matchLength + litLength;
+    }
+    ZSTD_memcpy(cctx->blockState.nextCBlock->rep, updatedRepcodes.rep, sizeof(repcodes_t));
+
+    if (inSeqs[idx].litLength) {
+        DEBUGLOG(6, "Storing last literals of size: %u", inSeqs[idx].litLength);
+        ZSTD_storeLastLiterals(&cctx->seqStore, ip, inSeqs[idx].litLength);
+        ip += inSeqs[idx].litLength;
+        seqPos->posInSrc += inSeqs[idx].litLength;
+    }
+    RETURN_ERROR_IF(ip != iend, corruption_detected, "Blocksize doesn't agree with block delimiter!");
+    seqPos->idx = idx+1;
+    return 0;
+}
+
+/* Returns the number of bytes to move the current read position back by. Only non-zero
+ * if we ended up splitting a sequence. Otherwise, it may return a ZSTD error if something
+ * went wrong.
+ *
+ * This function will attempt to scan through blockSize bytes represented by the sequences
+ * in inSeqs, storing any (partial) sequences.
+ *
+ * Occasionally, we may want to change the actual number of bytes we consumed from inSeqs to
+ * avoid splitting a match, or to avoid splitting a match such that it would produce a match
+ * smaller than MINMATCH. In this case, we return the number of bytes that we didn't read from this block.
+ */
+static size_t ZSTD_copySequencesToSeqStoreNoBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos,
+                                                       const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
+                                                       const void* src, size_t blockSize) {
+    U32 idx = seqPos->idx;
+    U32 startPosInSequence = seqPos->posInSequence;
+    U32 endPosInSequence = seqPos->posInSequence + (U32)blockSize;
+    size_t dictSize;
+    BYTE const* ip = (BYTE const*)(src);
+    BYTE const* iend = ip + blockSize;  /* May be adjusted if we decide to process fewer than blockSize bytes */
+    repcodes_t updatedRepcodes;
+    U32 bytesAdjustment = 0;
+    U32 finalMatchSplit = 0;
+    U32 litLength;
+    U32 matchLength;
+    U32 rawOffset;
+    U32 offCode;
+
+    if (cctx->cdict) {
+        dictSize = cctx->cdict->dictContentSize;
+    } else if (cctx->prefixDict.dict) {
+        dictSize = cctx->prefixDict.dictSize;
+    } else {
+        dictSize = 0;
+    }
+    DEBUGLOG(5, "ZSTD_copySequencesToSeqStore: idx: %u PIS: %u blockSize: %zu", idx, startPosInSequence, blockSize);
+    DEBUGLOG(5, "Start seq: idx: %u (of: %u ml: %u ll: %u)", idx, inSeqs[idx].offset, inSeqs[idx].matchLength, inSeqs[idx].litLength);
+    ZSTD_memcpy(updatedRepcodes.rep, cctx->blockState.prevCBlock->rep, sizeof(repcodes_t));
+    while (endPosInSequence && idx < inSeqsSize && !finalMatchSplit) {
+        const ZSTD_Sequence currSeq = inSeqs[idx];
+        litLength = currSeq.litLength;
+        matchLength = currSeq.matchLength;
+        rawOffset = currSeq.offset;
+
+        /* Modify the sequence depending on where endPosInSequence lies */
+        if (endPosInSequence >= currSeq.litLength + currSeq.matchLength) {
+            if (startPosInSequence >= litLength) {
+                startPosInSequence -= litLength;
+                litLength = 0;
+                matchLength -= startPosInSequence;
+            } else {
+                litLength -= startPosInSequence;
+            }
+            /* Move to the next sequence */
+            endPosInSequence -= currSeq.litLength + currSeq.matchLength;
+            startPosInSequence = 0;
+            idx++;
+        } else {
+            /* This is the final (partial) sequence we're adding from inSeqs, and endPosInSequence
+               does not reach the end of the match. So, we have to split the sequence */
+            DEBUGLOG(6, "Require a split: diff: %u, idx: %u PIS: %u",
+                     currSeq.litLength + currSeq.matchLength - endPosInSequence, idx, endPosInSequence);
+            if (endPosInSequence > litLength) {
+                U32 firstHalfMatchLength;
+                litLength = startPosInSequence >= litLength ? 0 : litLength - startPosInSequence;
+                firstHalfMatchLength = endPosInSequence - startPosInSequence - litLength;
+                if (matchLength > blockSize && firstHalfMatchLength >= cctx->appliedParams.cParams.minMatch) {
+                    /* Only ever split the match if it is larger than the block size */
+                    U32 secondHalfMatchLength = currSeq.matchLength + currSeq.litLength - endPosInSequence;
+                    if (secondHalfMatchLength < cctx->appliedParams.cParams.minMatch) {
+                        /* Move the endPosInSequence backward so that it creates match of minMatch length */
+                        endPosInSequence -= cctx->appliedParams.cParams.minMatch - secondHalfMatchLength;
+                        bytesAdjustment = cctx->appliedParams.cParams.minMatch - secondHalfMatchLength;
+                        firstHalfMatchLength -= bytesAdjustment;
+                    }
+                    matchLength = firstHalfMatchLength;
+                    /* Flag that we split the last match - after storing the sequence, exit the loop,
+                       but keep the value of endPosInSequence */
+                    finalMatchSplit = 1;
+                } else {
+                    /* Move the position in sequence backwards so that we don't split match, and break to store
+                     * the last literals. We use the original currSeq.litLength as a marker for where endPosInSequence
+                     * should go. We prefer to do this whenever it is not necessary to split the match, or if doing so
+                     * would cause the first half of the match to be too small
+                     */
+                    bytesAdjustment = endPosInSequence - currSeq.litLength;
+                    endPosInSequence = currSeq.litLength;
+                    break;
+                }
+            } else {
+                /* This sequence ends inside the literals, break to store the last literals */
+                break;
+            }
+        }
+        /* Check if this offset can be represented with a repcode */
+        {   U32 ll0 = (litLength == 0);
+            offCode = ZSTD_finalizeOffCode(rawOffset, updatedRepcodes.rep, ll0);
+            updatedRepcodes = ZSTD_updateRep(updatedRepcodes.rep, offCode, ll0);
+        }
+
+        if (cctx->appliedParams.validateSequences) {
+            seqPos->posInSrc += litLength + matchLength;
+            FORWARD_IF_ERROR(ZSTD_validateSequence(offCode, matchLength, seqPos->posInSrc,
+                                                   cctx->appliedParams.cParams.windowLog, dictSize,
+                                                   cctx->appliedParams.cParams.minMatch),
+                                                   "Sequence validation failed");
+        }
+        DEBUGLOG(6, "Storing sequence: (of: %u, ml: %u, ll: %u)", offCode, matchLength, litLength);
+        RETURN_ERROR_IF(idx - seqPos->idx > cctx->seqStore.maxNbSeq, memory_allocation,
+                        "Not enough memory allocated. Try adjusting ZSTD_c_minMatch.");
+        ZSTD_storeSeq(&cctx->seqStore, litLength, ip, iend, offCode, matchLength - MINMATCH);
+        ip += matchLength + litLength;
+    }
+    DEBUGLOG(5, "Ending seq: idx: %u (of: %u ml: %u ll: %u)", idx, inSeqs[idx].offset, inSeqs[idx].matchLength, inSeqs[idx].litLength);
+    assert(idx == inSeqsSize || endPosInSequence <= inSeqs[idx].litLength + inSeqs[idx].matchLength);
+    seqPos->idx = idx;
+    seqPos->posInSequence = endPosInSequence;
+    ZSTD_memcpy(cctx->blockState.nextCBlock->rep, updatedRepcodes.rep, sizeof(repcodes_t));
+
+    iend -= bytesAdjustment;
+    if (ip != iend) {
+        /* Store any last literals */
+        U32 lastLLSize = (U32)(iend - ip);
+        assert(ip <= iend);
+        DEBUGLOG(6, "Storing last literals of size: %u", lastLLSize);
+        ZSTD_storeLastLiterals(&cctx->seqStore, ip, lastLLSize);
+        seqPos->posInSrc += lastLLSize;
+    }
+
+    return bytesAdjustment;
+}
+
+typedef size_t (*ZSTD_sequenceCopier) (ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos,
+                                       const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
+                                       const void* src, size_t blockSize);
+static ZSTD_sequenceCopier ZSTD_selectSequenceCopier(ZSTD_sequenceFormat_e mode) {
+    ZSTD_sequenceCopier sequenceCopier = NULL;
+    assert(ZSTD_cParam_withinBounds(ZSTD_c_blockDelimiters, mode));
+    if (mode == ZSTD_sf_explicitBlockDelimiters) {
+        return ZSTD_copySequencesToSeqStoreExplicitBlockDelim;
+    } else if (mode == ZSTD_sf_noBlockDelimiters) {
+        return ZSTD_copySequencesToSeqStoreNoBlockDelim;
+    }
+    assert(sequenceCopier != NULL);
+    return sequenceCopier;
+}
+
+/* Compress, block-by-block, all of the sequences given.
+ *
+ * Returns the cumulative size of all compressed blocks (including their headers), otherwise a ZSTD error.
+ */
+static size_t ZSTD_compressSequences_internal(ZSTD_CCtx* cctx,
+                                              void* dst, size_t dstCapacity,
+                                              const ZSTD_Sequence* inSeqs, size_t inSeqsSize,
+                                              const void* src, size_t srcSize) {
+    size_t cSize = 0;
+    U32 lastBlock;
+    size_t blockSize;
+    size_t compressedSeqsSize;
+    size_t remaining = srcSize;
+    ZSTD_sequencePosition seqPos = {0, 0, 0};
+
+    BYTE const* ip = (BYTE const*)src;
+    BYTE* op = (BYTE*)dst;
+    ZSTD_sequenceCopier sequenceCopier = ZSTD_selectSequenceCopier(cctx->appliedParams.blockDelimiters);
+
+    DEBUGLOG(4, "ZSTD_compressSequences_internal srcSize: %zu, inSeqsSize: %zu", srcSize, inSeqsSize);
+    /* Special case: empty frame */
+    if (remaining == 0) {
+        U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw)<<1);
+        RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall, "No room for empty frame block header");
+        MEM_writeLE32(op, cBlockHeader24);
+        op += ZSTD_blockHeaderSize;
+        dstCapacity -= ZSTD_blockHeaderSize;
+        cSize += ZSTD_blockHeaderSize;
+    }
+
+    while (remaining) {
+        size_t cBlockSize;
+        size_t additionalByteAdjustment;
+        lastBlock = remaining <= cctx->blockSize;
+        blockSize = lastBlock ? (U32)remaining : (U32)cctx->blockSize;
+        ZSTD_resetSeqStore(&cctx->seqStore);
+        DEBUGLOG(4, "Working on new block. Blocksize: %zu", blockSize);
+
+        additionalByteAdjustment = sequenceCopier(cctx, &seqPos, inSeqs, inSeqsSize, ip, blockSize);
+        FORWARD_IF_ERROR(additionalByteAdjustment, "Bad sequence copy");
+        blockSize -= additionalByteAdjustment;
+
+        /* If blocks are too small, emit as a nocompress block */
+        if (blockSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) {
+            cBlockSize = ZSTD_noCompressBlock(op, dstCapacity, ip, blockSize, lastBlock);
+            FORWARD_IF_ERROR(cBlockSize, "Nocompress block failed");
+            DEBUGLOG(4, "Block too small, writing out nocompress block: cSize: %zu", cBlockSize);
+            cSize += cBlockSize;
+            ip += blockSize;
+            op += cBlockSize;
+            remaining -= blockSize;
+            dstCapacity -= cBlockSize;
+            continue;
+        }
+
+        compressedSeqsSize = ZSTD_entropyCompressSequences(&cctx->seqStore,
+                                &cctx->blockState.prevCBlock->entropy, &cctx->blockState.nextCBlock->entropy,
+                                &cctx->appliedParams,
+                                op + ZSTD_blockHeaderSize /* Leave space for block header */, dstCapacity - ZSTD_blockHeaderSize,
+                                blockSize,
+                                cctx->entropyWorkspace, ENTROPY_WORKSPACE_SIZE /* statically allocated in resetCCtx */,
+                                cctx->bmi2);
+        FORWARD_IF_ERROR(compressedSeqsSize, "Compressing sequences of block failed");
+        DEBUGLOG(4, "Compressed sequences size: %zu", compressedSeqsSize);
+
+        if (!cctx->isFirstBlock &&
+            ZSTD_maybeRLE(&cctx->seqStore) &&
+            ZSTD_isRLE((BYTE const*)src, srcSize)) {
+            /* We don't want to emit our first block as a RLE even if it qualifies because
+            * doing so will cause the decoder (cli only) to throw a "should consume all input error."
+            * This is only an issue for zstd <= v1.4.3
+            */
+            compressedSeqsSize = 1;
+        }
+
+        if (compressedSeqsSize == 0) {
+            /* ZSTD_noCompressBlock writes the block header as well */
+            cBlockSize = ZSTD_noCompressBlock(op, dstCapacity, ip, blockSize, lastBlock);
+            FORWARD_IF_ERROR(cBlockSize, "Nocompress block failed");
+            DEBUGLOG(4, "Writing out nocompress block, size: %zu", cBlockSize);
+        } else if (compressedSeqsSize == 1) {
+            cBlockSize = ZSTD_rleCompressBlock(op, dstCapacity, *ip, blockSize, lastBlock);
+            FORWARD_IF_ERROR(cBlockSize, "RLE compress block failed");
+            DEBUGLOG(4, "Writing out RLE block, size: %zu", cBlockSize);
+        } else {
+            U32 cBlockHeader;
+            /* Error checking and repcodes update */
+            ZSTD_confirmRepcodesAndEntropyTables(cctx);
+            if (cctx->blockState.prevCBlock->entropy.fse.offcode_repeatMode == FSE_repeat_valid)
+                cctx->blockState.prevCBlock->entropy.fse.offcode_repeatMode = FSE_repeat_check;
+
+            /* Write block header into beginning of block*/
+            cBlockHeader = lastBlock + (((U32)bt_compressed)<<1) + (U32)(compressedSeqsSize << 3);
+            MEM_writeLE24(op, cBlockHeader);
+            cBlockSize = ZSTD_blockHeaderSize + compressedSeqsSize;
+            DEBUGLOG(4, "Writing out compressed block, size: %zu", cBlockSize);
+        }
+
+        cSize += cBlockSize;
+        DEBUGLOG(4, "cSize running total: %zu", cSize);
+
+        if (lastBlock) {
+            break;
+        } else {
+            ip += blockSize;
+            op += cBlockSize;
+            remaining -= blockSize;
+            dstCapacity -= cBlockSize;
+            cctx->isFirstBlock = 0;
+        }
+    }
+
+    return cSize;
+}
+
+size_t ZSTD_compressSequences(ZSTD_CCtx* const cctx, void* dst, size_t dstCapacity,
+                              const ZSTD_Sequence* inSeqs, size_t inSeqsSize,
+                              const void* src, size_t srcSize) {
+    BYTE* op = (BYTE*)dst;
+    size_t cSize = 0;
+    size_t compressedBlocksSize = 0;
+    size_t frameHeaderSize = 0;
+
+    /* Transparent initialization stage, same as compressStream2() */
+    DEBUGLOG(3, "ZSTD_compressSequences()");
+    assert(cctx != NULL);
+    FORWARD_IF_ERROR(ZSTD_CCtx_init_compressStream2(cctx, ZSTD_e_end, srcSize), "CCtx initialization failed");
+    /* Begin writing output, starting with frame header */
+    frameHeaderSize = ZSTD_writeFrameHeader(op, dstCapacity, &cctx->appliedParams, srcSize, cctx->dictID);
+    op += frameHeaderSize;
+    dstCapacity -= frameHeaderSize;
+    cSize += frameHeaderSize;
+    if (cctx->appliedParams.fParams.checksumFlag && srcSize) {
+        xxh64_update(&cctx->xxhState, src, srcSize);
+    }
+    /* cSize includes block header size and compressed sequences size */
+    compressedBlocksSize = ZSTD_compressSequences_internal(cctx,
+                                                           op, dstCapacity,
+                                                           inSeqs, inSeqsSize,
+                                                           src, srcSize);
+    FORWARD_IF_ERROR(compressedBlocksSize, "Compressing blocks failed!");
+    cSize += compressedBlocksSize;
+    dstCapacity -= compressedBlocksSize;
+
+    if (cctx->appliedParams.fParams.checksumFlag) {
+        U32 const checksum = (U32) xxh64_digest(&cctx->xxhState);
+        RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall, "no room for checksum");
+        DEBUGLOG(4, "Write checksum : %08X", (unsigned)checksum);
+        MEM_writeLE32((char*)dst + cSize, checksum);
+        cSize += 4;
+    }
+
+    DEBUGLOG(3, "Final compressed size: %zu", cSize);
+    return cSize;
+}
+
+/*======   Finalize   ======*/
+
+/*! ZSTD_flushStream() :
+ * @return : amount of data remaining to flush */
+size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output)
+{
+    ZSTD_inBuffer input = { NULL, 0, 0 };
+    return ZSTD_compressStream2(zcs, output, &input, ZSTD_e_flush);
+}
+
+
+size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output)
+{
+    ZSTD_inBuffer input = { NULL, 0, 0 };
+    size_t const remainingToFlush = ZSTD_compressStream2(zcs, output, &input, ZSTD_e_end);
+    FORWARD_IF_ERROR( remainingToFlush , "ZSTD_compressStream2 failed");
+    if (zcs->appliedParams.nbWorkers > 0) return remainingToFlush;   /* minimal estimation */
+    /* single thread mode : attempt to calculate remaining to flush more precisely */
+    {   size_t const lastBlockSize = zcs->frameEnded ? 0 : ZSTD_BLOCKHEADERSIZE;
+        size_t const checksumSize = (size_t)(zcs->frameEnded ? 0 : zcs->appliedParams.fParams.checksumFlag * 4);
+        size_t const toFlush = remainingToFlush + lastBlockSize + checksumSize;
+        DEBUGLOG(4, "ZSTD_endStream : remaining to flush : %u", (unsigned)toFlush);
+        return toFlush;
+    }
+}
+
+
+/*-=====  Pre-defined compression levels  =====-*/
+
+#define ZSTD_MAX_CLEVEL     22
+int ZSTD_maxCLevel(void) { return ZSTD_MAX_CLEVEL; }
+int ZSTD_minCLevel(void) { return (int)-ZSTD_TARGETLENGTH_MAX; }
+
+static const ZSTD_compressionParameters ZSTD_defaultCParameters[4][ZSTD_MAX_CLEVEL+1] = {
+{   /* "default" - for any srcSize > 256 KB */
+    /* W,  C,  H,  S,  L, TL, strat */
+    { 19, 12, 13,  1,  6,  1, ZSTD_fast    },  /* base for negative levels */
+    { 19, 13, 14,  1,  7,  0, ZSTD_fast    },  /* level  1 */
+    { 20, 15, 16,  1,  6,  0, ZSTD_fast    },  /* level  2 */
+    { 21, 16, 17,  1,  5,  0, ZSTD_dfast   },  /* level  3 */
+    { 21, 18, 18,  1,  5,  0, ZSTD_dfast   },  /* level  4 */
+    { 21, 18, 19,  2,  5,  2, ZSTD_greedy  },  /* level  5 */
+    { 21, 19, 19,  3,  5,  4, ZSTD_greedy  },  /* level  6 */
+    { 21, 19, 19,  3,  5,  8, ZSTD_lazy    },  /* level  7 */
+    { 21, 19, 19,  3,  5, 16, ZSTD_lazy2   },  /* level  8 */
+    { 21, 19, 20,  4,  5, 16, ZSTD_lazy2   },  /* level  9 */
+    { 22, 20, 21,  4,  5, 16, ZSTD_lazy2   },  /* level 10 */
+    { 22, 21, 22,  4,  5, 16, ZSTD_lazy2   },  /* level 11 */
+    { 22, 21, 22,  5,  5, 16, ZSTD_lazy2   },  /* level 12 */
+    { 22, 21, 22,  5,  5, 32, ZSTD_btlazy2 },  /* level 13 */
+    { 22, 22, 23,  5,  5, 32, ZSTD_btlazy2 },  /* level 14 */
+    { 22, 23, 23,  6,  5, 32, ZSTD_btlazy2 },  /* level 15 */
+    { 22, 22, 22,  5,  5, 48, ZSTD_btopt   },  /* level 16 */
+    { 23, 23, 22,  5,  4, 64, ZSTD_btopt   },  /* level 17 */
+    { 23, 23, 22,  6,  3, 64, ZSTD_btultra },  /* level 18 */
+    { 23, 24, 22,  7,  3,256, ZSTD_btultra2},  /* level 19 */
+    { 25, 25, 23,  7,  3,256, ZSTD_btultra2},  /* level 20 */
+    { 26, 26, 24,  7,  3,512, ZSTD_btultra2},  /* level 21 */
+    { 27, 27, 25,  9,  3,999, ZSTD_btultra2},  /* level 22 */
+},
+{   /* for srcSize <= 256 KB */
+    /* W,  C,  H,  S,  L,  T, strat */
+    { 18, 12, 13,  1,  5,  1, ZSTD_fast    },  /* base for negative levels */
+    { 18, 13, 14,  1,  6,  0, ZSTD_fast    },  /* level  1 */
+    { 18, 14, 14,  1,  5,  0, ZSTD_dfast   },  /* level  2 */
+    { 18, 16, 16,  1,  4,  0, ZSTD_dfast   },  /* level  3 */
+    { 18, 16, 17,  2,  5,  2, ZSTD_greedy  },  /* level  4.*/
+    { 18, 18, 18,  3,  5,  2, ZSTD_greedy  },  /* level  5.*/
+    { 18, 18, 19,  3,  5,  4, ZSTD_lazy    },  /* level  6.*/
+    { 18, 18, 19,  4,  4,  4, ZSTD_lazy    },  /* level  7 */
+    { 18, 18, 19,  4,  4,  8, ZSTD_lazy2   },  /* level  8 */
+    { 18, 18, 19,  5,  4,  8, ZSTD_lazy2   },  /* level  9 */
+    { 18, 18, 19,  6,  4,  8, ZSTD_lazy2   },  /* level 10 */
+    { 18, 18, 19,  5,  4, 12, ZSTD_btlazy2 },  /* level 11.*/
+    { 18, 19, 19,  7,  4, 12, ZSTD_btlazy2 },  /* level 12.*/
+    { 18, 18, 19,  4,  4, 16, ZSTD_btopt   },  /* level 13 */
+    { 18, 18, 19,  4,  3, 32, ZSTD_btopt   },  /* level 14.*/
+    { 18, 18, 19,  6,  3,128, ZSTD_btopt   },  /* level 15.*/
+    { 18, 19, 19,  6,  3,128, ZSTD_btultra },  /* level 16.*/
+    { 18, 19, 19,  8,  3,256, ZSTD_btultra },  /* level 17.*/
+    { 18, 19, 19,  6,  3,128, ZSTD_btultra2},  /* level 18.*/
+    { 18, 19, 19,  8,  3,256, ZSTD_btultra2},  /* level 19.*/
+    { 18, 19, 19, 10,  3,512, ZSTD_btultra2},  /* level 20.*/
+    { 18, 19, 19, 12,  3,512, ZSTD_btultra2},  /* level 21.*/
+    { 18, 19, 19, 13,  3,999, ZSTD_btultra2},  /* level 22.*/
+},
+{   /* for srcSize <= 128 KB */
+    /* W,  C,  H,  S,  L,  T, strat */
+    { 17, 12, 12,  1,  5,  1, ZSTD_fast    },  /* base for negative levels */
+    { 17, 12, 13,  1,  6,  0, ZSTD_fast    },  /* level  1 */
+    { 17, 13, 15,  1,  5,  0, ZSTD_fast    },  /* level  2 */
+    { 17, 15, 16,  2,  5,  0, ZSTD_dfast   },  /* level  3 */
+    { 17, 17, 17,  2,  4,  0, ZSTD_dfast   },  /* level  4 */
+    { 17, 16, 17,  3,  4,  2, ZSTD_greedy  },  /* level  5 */
+    { 17, 17, 17,  3,  4,  4, ZSTD_lazy    },  /* level  6 */
+    { 17, 17, 17,  3,  4,  8, ZSTD_lazy2   },  /* level  7 */
+    { 17, 17, 17,  4,  4,  8, ZSTD_lazy2   },  /* level  8 */
+    { 17, 17, 17,  5,  4,  8, ZSTD_lazy2   },  /* level  9 */
+    { 17, 17, 17,  6,  4,  8, ZSTD_lazy2   },  /* level 10 */
+    { 17, 17, 17,  5,  4,  8, ZSTD_btlazy2 },  /* level 11 */
+    { 17, 18, 17,  7,  4, 12, ZSTD_btlazy2 },  /* level 12 */
+    { 17, 18, 17,  3,  4, 12, ZSTD_btopt   },  /* level 13.*/
+    { 17, 18, 17,  4,  3, 32, ZSTD_btopt   },  /* level 14.*/
+    { 17, 18, 17,  6,  3,256, ZSTD_btopt   },  /* level 15.*/
+    { 17, 18, 17,  6,  3,128, ZSTD_btultra },  /* level 16.*/
+    { 17, 18, 17,  8,  3,256, ZSTD_btultra },  /* level 17.*/
+    { 17, 18, 17, 10,  3,512, ZSTD_btultra },  /* level 18.*/
+    { 17, 18, 17,  5,  3,256, ZSTD_btultra2},  /* level 19.*/
+    { 17, 18, 17,  7,  3,512, ZSTD_btultra2},  /* level 20.*/
+    { 17, 18, 17,  9,  3,512, ZSTD_btultra2},  /* level 21.*/
+    { 17, 18, 17, 11,  3,999, ZSTD_btultra2},  /* level 22.*/
+},
+{   /* for srcSize <= 16 KB */
+    /* W,  C,  H,  S,  L,  T, strat */
+    { 14, 12, 13,  1,  5,  1, ZSTD_fast    },  /* base for negative levels */
+    { 14, 14, 15,  1,  5,  0, ZSTD_fast    },  /* level  1 */
+    { 14, 14, 15,  1,  4,  0, ZSTD_fast    },  /* level  2 */
+    { 14, 14, 15,  2,  4,  0, ZSTD_dfast   },  /* level  3 */
+    { 14, 14, 14,  4,  4,  2, ZSTD_greedy  },  /* level  4 */
+    { 14, 14, 14,  3,  4,  4, ZSTD_lazy    },  /* level  5.*/
+    { 14, 14, 14,  4,  4,  8, ZSTD_lazy2   },  /* level  6 */
+    { 14, 14, 14,  6,  4,  8, ZSTD_lazy2   },  /* level  7 */
+    { 14, 14, 14,  8,  4,  8, ZSTD_lazy2   },  /* level  8.*/
+    { 14, 15, 14,  5,  4,  8, ZSTD_btlazy2 },  /* level  9.*/
+    { 14, 15, 14,  9,  4,  8, ZSTD_btlazy2 },  /* level 10.*/
+    { 14, 15, 14,  3,  4, 12, ZSTD_btopt   },  /* level 11.*/
+    { 14, 15, 14,  4,  3, 24, ZSTD_btopt   },  /* level 12.*/
+    { 14, 15, 14,  5,  3, 32, ZSTD_btultra },  /* level 13.*/
+    { 14, 15, 15,  6,  3, 64, ZSTD_btultra },  /* level 14.*/
+    { 14, 15, 15,  7,  3,256, ZSTD_btultra },  /* level 15.*/
+    { 14, 15, 15,  5,  3, 48, ZSTD_btultra2},  /* level 16.*/
+    { 14, 15, 15,  6,  3,128, ZSTD_btultra2},  /* level 17.*/
+    { 14, 15, 15,  7,  3,256, ZSTD_btultra2},  /* level 18.*/
+    { 14, 15, 15,  8,  3,256, ZSTD_btultra2},  /* level 19.*/
+    { 14, 15, 15,  8,  3,512, ZSTD_btultra2},  /* level 20.*/
+    { 14, 15, 15,  9,  3,512, ZSTD_btultra2},  /* level 21.*/
+    { 14, 15, 15, 10,  3,999, ZSTD_btultra2},  /* level 22.*/
+},
+};
+
+static ZSTD_compressionParameters ZSTD_dedicatedDictSearch_getCParams(int const compressionLevel, size_t const dictSize)
+{
+    ZSTD_compressionParameters cParams = ZSTD_getCParams_internal(compressionLevel, 0, dictSize, ZSTD_cpm_createCDict);
+    switch (cParams.strategy) {
+        case ZSTD_fast:
+        case ZSTD_dfast:
+            break;
+        case ZSTD_greedy:
+        case ZSTD_lazy:
+        case ZSTD_lazy2:
+            cParams.hashLog += ZSTD_LAZY_DDSS_BUCKET_LOG;
+            break;
+        case ZSTD_btlazy2:
+        case ZSTD_btopt:
+        case ZSTD_btultra:
+        case ZSTD_btultra2:
+            break;
+    }
+    return cParams;
+}
+
+static int ZSTD_dedicatedDictSearch_isSupported(
+        ZSTD_compressionParameters const* cParams)
+{
+    return (cParams->strategy >= ZSTD_greedy)
+        && (cParams->strategy <= ZSTD_lazy2)
+        && (cParams->hashLog >= cParams->chainLog)
+        && (cParams->chainLog <= 24);
+}
+
+/*
+ * Reverses the adjustment applied to cparams when enabling dedicated dict
+ * search. This is used to recover the params set to be used in the working
+ * context. (Otherwise, those tables would also grow.)
+ */
+static void ZSTD_dedicatedDictSearch_revertCParams(
+        ZSTD_compressionParameters* cParams) {
+    switch (cParams->strategy) {
+        case ZSTD_fast:
+        case ZSTD_dfast:
+            break;
+        case ZSTD_greedy:
+        case ZSTD_lazy:
+        case ZSTD_lazy2:
+            cParams->hashLog -= ZSTD_LAZY_DDSS_BUCKET_LOG;
+            break;
+        case ZSTD_btlazy2:
+        case ZSTD_btopt:
+        case ZSTD_btultra:
+        case ZSTD_btultra2:
+            break;
+    }
+}
+
+static U64 ZSTD_getCParamRowSize(U64 srcSizeHint, size_t dictSize, ZSTD_cParamMode_e mode)
+{
+    switch (mode) {
+    case ZSTD_cpm_unknown:
+    case ZSTD_cpm_noAttachDict:
+    case ZSTD_cpm_createCDict:
+        break;
+    case ZSTD_cpm_attachDict:
+        dictSize = 0;
+        break;
+    default:
+        assert(0);
+        break;
+    }
+    {   int const unknown = srcSizeHint == ZSTD_CONTENTSIZE_UNKNOWN;
+        size_t const addedSize = unknown && dictSize > 0 ? 500 : 0;
+        return unknown && dictSize == 0 ? ZSTD_CONTENTSIZE_UNKNOWN : srcSizeHint+dictSize+addedSize;
+    }
+}
+
+/*! ZSTD_getCParams_internal() :
+ * @return ZSTD_compressionParameters structure for a selected compression level, srcSize and dictSize.
+ *  Note: srcSizeHint 0 means 0, use ZSTD_CONTENTSIZE_UNKNOWN for unknown.
+ *        Use dictSize == 0 for unknown or unused.
+ *  Note: `mode` controls how we treat the `dictSize`. See docs for `ZSTD_cParamMode_e`. */
+static ZSTD_compressionParameters ZSTD_getCParams_internal(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize, ZSTD_cParamMode_e mode)
+{
+    U64 const rSize = ZSTD_getCParamRowSize(srcSizeHint, dictSize, mode);
+    U32 const tableID = (rSize <= 256 KB) + (rSize <= 128 KB) + (rSize <= 16 KB);
+    int row;
+    DEBUGLOG(5, "ZSTD_getCParams_internal (cLevel=%i)", compressionLevel);
+
+    /* row */
+    if (compressionLevel == 0) row = ZSTD_CLEVEL_DEFAULT;   /* 0 == default */
+    else if (compressionLevel < 0) row = 0;   /* entry 0 is baseline for fast mode */
+    else if (compressionLevel > ZSTD_MAX_CLEVEL) row = ZSTD_MAX_CLEVEL;
+    else row = compressionLevel;
+
+    {   ZSTD_compressionParameters cp = ZSTD_defaultCParameters[tableID][row];
+        /* acceleration factor */
+        if (compressionLevel < 0) {
+            int const clampedCompressionLevel = MAX(ZSTD_minCLevel(), compressionLevel);
+            cp.targetLength = (unsigned)(-clampedCompressionLevel);
+        }
+        /* refine parameters based on srcSize & dictSize */
+        return ZSTD_adjustCParams_internal(cp, srcSizeHint, dictSize, mode);
+    }
+}
+
+/*! ZSTD_getCParams() :
+ * @return ZSTD_compressionParameters structure for a selected compression level, srcSize and dictSize.
+ *  Size values are optional, provide 0 if not known or unused */
+ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize)
+{
+    if (srcSizeHint == 0) srcSizeHint = ZSTD_CONTENTSIZE_UNKNOWN;
+    return ZSTD_getCParams_internal(compressionLevel, srcSizeHint, dictSize, ZSTD_cpm_unknown);
+}
+
+/*! ZSTD_getParams() :
+ *  same idea as ZSTD_getCParams()
+ * @return a `ZSTD_parameters` structure (instead of `ZSTD_compressionParameters`).
+ *  Fields of `ZSTD_frameParameters` are set to default values */
+static ZSTD_parameters ZSTD_getParams_internal(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize, ZSTD_cParamMode_e mode) {
+    ZSTD_parameters params;
+    ZSTD_compressionParameters const cParams = ZSTD_getCParams_internal(compressionLevel, srcSizeHint, dictSize, mode);
+    DEBUGLOG(5, "ZSTD_getParams (cLevel=%i)", compressionLevel);
+    ZSTD_memset(&params, 0, sizeof(params));
+    params.cParams = cParams;
+    params.fParams.contentSizeFlag = 1;
+    return params;
+}
+
+/*! ZSTD_getParams() :
+ *  same idea as ZSTD_getCParams()
+ * @return a `ZSTD_parameters` structure (instead of `ZSTD_compressionParameters`).
+ *  Fields of `ZSTD_frameParameters` are set to default values */
+ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize) {
+    if (srcSizeHint == 0) srcSizeHint = ZSTD_CONTENTSIZE_UNKNOWN;
+    return ZSTD_getParams_internal(compressionLevel, srcSizeHint, dictSize, ZSTD_cpm_unknown);
+}
diff --git a/lib/zstd/compress/zstd_compress_internal.h b/lib/zstd/compress/zstd_compress_internal.h
new file mode 100644
index 0000000..685d2f9
--- /dev/null
+++ b/lib/zstd/compress/zstd_compress_internal.h
@@ -0,0 +1,1188 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/* This header contains definitions
+ * that shall **only** be used by modules within lib/compress.
+ */
+
+#ifndef ZSTD_COMPRESS_H
+#define ZSTD_COMPRESS_H
+
+/*-*************************************
+*  Dependencies
+***************************************/
+#include "../common/zstd_internal.h"
+#include "zstd_cwksp.h"
+
+
+/*-*************************************
+*  Constants
+***************************************/
+#define kSearchStrength      8
+#define HASH_READ_SIZE       8
+#define ZSTD_DUBT_UNSORTED_MARK 1   /* For btlazy2 strategy, index ZSTD_DUBT_UNSORTED_MARK==1 means "unsorted".
+                                       It could be confused for a real successor at index "1", if sorted as larger than its predecessor.
+                                       It's not a big deal though : candidate will just be sorted again.
+                                       Additionally, candidate position 1 will be lost.
+                                       But candidate 1 cannot hide a large tree of candidates, so it's a minimal loss.
+                                       The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be mishandled after table re-use with a different strategy.
+                                       This constant is required by ZSTD_compressBlock_btlazy2() and ZSTD_reduceTable_internal() */
+
+
+/*-*************************************
+*  Context memory management
+***************************************/
+typedef enum { ZSTDcs_created=0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZSTD_compressionStage_e;
+typedef enum { zcss_init=0, zcss_load, zcss_flush } ZSTD_cStreamStage;
+
+typedef struct ZSTD_prefixDict_s {
+    const void* dict;
+    size_t dictSize;
+    ZSTD_dictContentType_e dictContentType;
+} ZSTD_prefixDict;
+
+typedef struct {
+    void* dictBuffer;
+    void const* dict;
+    size_t dictSize;
+    ZSTD_dictContentType_e dictContentType;
+    ZSTD_CDict* cdict;
+} ZSTD_localDict;
+
+typedef struct {
+    HUF_CElt CTable[HUF_CTABLE_SIZE_U32(255)];
+    HUF_repeat repeatMode;
+} ZSTD_hufCTables_t;
+
+typedef struct {
+    FSE_CTable offcodeCTable[FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)];
+    FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)];
+    FSE_CTable litlengthCTable[FSE_CTABLE_SIZE_U32(LLFSELog, MaxLL)];
+    FSE_repeat offcode_repeatMode;
+    FSE_repeat matchlength_repeatMode;
+    FSE_repeat litlength_repeatMode;
+} ZSTD_fseCTables_t;
+
+typedef struct {
+    ZSTD_hufCTables_t huf;
+    ZSTD_fseCTables_t fse;
+} ZSTD_entropyCTables_t;
+
+typedef struct {
+    U32 off;            /* Offset code (offset + ZSTD_REP_MOVE) for the match */
+    U32 len;            /* Raw length of match */
+} ZSTD_match_t;
+
+typedef struct {
+    U32 offset;         /* Offset of sequence */
+    U32 litLength;      /* Length of literals prior to match */
+    U32 matchLength;    /* Raw length of match */
+} rawSeq;
+
+typedef struct {
+  rawSeq* seq;          /* The start of the sequences */
+  size_t pos;           /* The index in seq where reading stopped. pos <= size. */
+  size_t posInSequence; /* The position within the sequence at seq[pos] where reading
+                           stopped. posInSequence <= seq[pos].litLength + seq[pos].matchLength */
+  size_t size;          /* The number of sequences. <= capacity. */
+  size_t capacity;      /* The capacity starting from `seq` pointer */
+} rawSeqStore_t;
+
+UNUSED_ATTR static const rawSeqStore_t kNullRawSeqStore = {NULL, 0, 0, 0, 0};
+
+typedef struct {
+    int price;
+    U32 off;
+    U32 mlen;
+    U32 litlen;
+    U32 rep[ZSTD_REP_NUM];
+} ZSTD_optimal_t;
+
+typedef enum { zop_dynamic=0, zop_predef } ZSTD_OptPrice_e;
+
+typedef struct {
+    /* All tables are allocated inside cctx->workspace by ZSTD_resetCCtx_internal() */
+    unsigned* litFreq;           /* table of literals statistics, of size 256 */
+    unsigned* litLengthFreq;     /* table of litLength statistics, of size (MaxLL+1) */
+    unsigned* matchLengthFreq;   /* table of matchLength statistics, of size (MaxML+1) */
+    unsigned* offCodeFreq;       /* table of offCode statistics, of size (MaxOff+1) */
+    ZSTD_match_t* matchTable;    /* list of found matches, of size ZSTD_OPT_NUM+1 */
+    ZSTD_optimal_t* priceTable;  /* All positions tracked by optimal parser, of size ZSTD_OPT_NUM+1 */
+
+    U32  litSum;                 /* nb of literals */
+    U32  litLengthSum;           /* nb of litLength codes */
+    U32  matchLengthSum;         /* nb of matchLength codes */
+    U32  offCodeSum;             /* nb of offset codes */
+    U32  litSumBasePrice;        /* to compare to log2(litfreq) */
+    U32  litLengthSumBasePrice;  /* to compare to log2(llfreq)  */
+    U32  matchLengthSumBasePrice;/* to compare to log2(mlfreq)  */
+    U32  offCodeSumBasePrice;    /* to compare to log2(offreq)  */
+    ZSTD_OptPrice_e priceType;   /* prices can be determined dynamically, or follow a pre-defined cost structure */
+    const ZSTD_entropyCTables_t* symbolCosts;  /* pre-calculated dictionary statistics */
+    ZSTD_literalCompressionMode_e literalCompressionMode;
+} optState_t;
+
+typedef struct {
+  ZSTD_entropyCTables_t entropy;
+  U32 rep[ZSTD_REP_NUM];
+} ZSTD_compressedBlockState_t;
+
+typedef struct {
+    BYTE const* nextSrc;    /* next block here to continue on current prefix */
+    BYTE const* base;       /* All regular indexes relative to this position */
+    BYTE const* dictBase;   /* extDict indexes relative to this position */
+    U32 dictLimit;          /* below that point, need extDict */
+    U32 lowLimit;           /* below that point, no more valid data */
+} ZSTD_window_t;
+
+typedef struct ZSTD_matchState_t ZSTD_matchState_t;
+struct ZSTD_matchState_t {
+    ZSTD_window_t window;   /* State for window round buffer management */
+    U32 loadedDictEnd;      /* index of end of dictionary, within context's referential.
+                             * When loadedDictEnd != 0, a dictionary is in use, and still valid.
+                             * This relies on a mechanism to set loadedDictEnd=0 when dictionary is no longer within distance.
+                             * Such mechanism is provided within ZSTD_window_enforceMaxDist() and ZSTD_checkDictValidity().
+                             * When dict referential is copied into active context (i.e. not attached),
+                             * loadedDictEnd == dictSize, since referential starts from zero.
+                             */
+    U32 nextToUpdate;       /* index from which to continue table update */
+    U32 hashLog3;           /* dispatch table for matches of len==3 : larger == faster, more memory */
+    U32* hashTable;
+    U32* hashTable3;
+    U32* chainTable;
+    int dedicatedDictSearch;  /* Indicates whether this matchState is using the
+                               * dedicated dictionary search structure.
+                               */
+    optState_t opt;         /* optimal parser state */
+    const ZSTD_matchState_t* dictMatchState;
+    ZSTD_compressionParameters cParams;
+    const rawSeqStore_t* ldmSeqStore;
+};
+
+typedef struct {
+    ZSTD_compressedBlockState_t* prevCBlock;
+    ZSTD_compressedBlockState_t* nextCBlock;
+    ZSTD_matchState_t matchState;
+} ZSTD_blockState_t;
+
+typedef struct {
+    U32 offset;
+    U32 checksum;
+} ldmEntry_t;
+
+typedef struct {
+    BYTE const* split;
+    U32 hash;
+    U32 checksum;
+    ldmEntry_t* bucket;
+} ldmMatchCandidate_t;
+
+#define LDM_BATCH_SIZE 64
+
+typedef struct {
+    ZSTD_window_t window;   /* State for the window round buffer management */
+    ldmEntry_t* hashTable;
+    U32 loadedDictEnd;
+    BYTE* bucketOffsets;    /* Next position in bucket to insert entry */
+    size_t splitIndices[LDM_BATCH_SIZE];
+    ldmMatchCandidate_t matchCandidates[LDM_BATCH_SIZE];
+} ldmState_t;
+
+typedef struct {
+    U32 enableLdm;          /* 1 if enable long distance matching */
+    U32 hashLog;            /* Log size of hashTable */
+    U32 bucketSizeLog;      /* Log bucket size for collision resolution, at most 8 */
+    U32 minMatchLength;     /* Minimum match length */
+    U32 hashRateLog;       /* Log number of entries to skip */
+    U32 windowLog;          /* Window log for the LDM */
+} ldmParams_t;
+
+typedef struct {
+    int collectSequences;
+    ZSTD_Sequence* seqStart;
+    size_t seqIndex;
+    size_t maxSequences;
+} SeqCollector;
+
+struct ZSTD_CCtx_params_s {
+    ZSTD_format_e format;
+    ZSTD_compressionParameters cParams;
+    ZSTD_frameParameters fParams;
+
+    int compressionLevel;
+    int forceWindow;           /* force back-references to respect limit of
+                                * 1<<wLog, even for dictionary */
+    size_t targetCBlockSize;   /* Tries to fit compressed block size to be around targetCBlockSize.
+                                * No target when targetCBlockSize == 0.
+                                * There is no guarantee on compressed block size */
+    int srcSizeHint;           /* User's best guess of source size.
+                                * Hint is not valid when srcSizeHint == 0.
+                                * There is no guarantee that hint is close to actual source size */
+
+    ZSTD_dictAttachPref_e attachDictPref;
+    ZSTD_literalCompressionMode_e literalCompressionMode;
+
+    /* Multithreading: used to pass parameters to mtctx */
+    int nbWorkers;
+    size_t jobSize;
+    int overlapLog;
+    int rsyncable;
+
+    /* Long distance matching parameters */
+    ldmParams_t ldmParams;
+
+    /* Dedicated dict search algorithm trigger */
+    int enableDedicatedDictSearch;
+
+    /* Input/output buffer modes */
+    ZSTD_bufferMode_e inBufferMode;
+    ZSTD_bufferMode_e outBufferMode;
+
+    /* Sequence compression API */
+    ZSTD_sequenceFormat_e blockDelimiters;
+    int validateSequences;
+
+    /* Internal use, for createCCtxParams() and freeCCtxParams() only */
+    ZSTD_customMem customMem;
+};  /* typedef'd to ZSTD_CCtx_params within "zstd.h" */
+
+#define COMPRESS_SEQUENCES_WORKSPACE_SIZE (sizeof(unsigned) * (MaxSeq + 2))
+#define ENTROPY_WORKSPACE_SIZE (HUF_WORKSPACE_SIZE + COMPRESS_SEQUENCES_WORKSPACE_SIZE)
+
+/*
+ * Indicates whether this compression proceeds directly from user-provided
+ * source buffer to user-provided destination buffer (ZSTDb_not_buffered), or
+ * whether the context needs to buffer the input/output (ZSTDb_buffered).
+ */
+typedef enum {
+    ZSTDb_not_buffered,
+    ZSTDb_buffered
+} ZSTD_buffered_policy_e;
+
+struct ZSTD_CCtx_s {
+    ZSTD_compressionStage_e stage;
+    int cParamsChanged;                  /* == 1 if cParams(except wlog) or compression level are changed in requestedParams. Triggers transmission of new params to ZSTDMT (if available) then reset to 0. */
+    int bmi2;                            /* == 1 if the CPU supports BMI2 and 0 otherwise. CPU support is determined dynamically once per context lifetime. */
+    ZSTD_CCtx_params requestedParams;
+    ZSTD_CCtx_params appliedParams;
+    U32   dictID;
+    size_t dictContentSize;
+
+    ZSTD_cwksp workspace; /* manages buffer for dynamic allocations */
+    size_t blockSize;
+    unsigned long long pledgedSrcSizePlusOne;  /* this way, 0 (default) == unknown */
+    unsigned long long consumedSrcSize;
+    unsigned long long producedCSize;
+    struct xxh64_state xxhState;
+    ZSTD_customMem customMem;
+    ZSTD_threadPool* pool;
+    size_t staticSize;
+    SeqCollector seqCollector;
+    int isFirstBlock;
+    int initialized;
+
+    seqStore_t seqStore;      /* sequences storage ptrs */
+    ldmState_t ldmState;      /* long distance matching state */
+    rawSeq* ldmSequences;     /* Storage for the ldm output sequences */
+    size_t maxNbLdmSequences;
+    rawSeqStore_t externSeqStore; /* Mutable reference to external sequences */
+    ZSTD_blockState_t blockState;
+    U32* entropyWorkspace;  /* entropy workspace of ENTROPY_WORKSPACE_SIZE bytes */
+
+    /* Wether we are streaming or not */
+    ZSTD_buffered_policy_e bufferedPolicy;
+
+    /* streaming */
+    char*  inBuff;
+    size_t inBuffSize;
+    size_t inToCompress;
+    size_t inBuffPos;
+    size_t inBuffTarget;
+    char*  outBuff;
+    size_t outBuffSize;
+    size_t outBuffContentSize;
+    size_t outBuffFlushedSize;
+    ZSTD_cStreamStage streamStage;
+    U32    frameEnded;
+
+    /* Stable in/out buffer verification */
+    ZSTD_inBuffer expectedInBuffer;
+    size_t expectedOutBufferSize;
+
+    /* Dictionary */
+    ZSTD_localDict localDict;
+    const ZSTD_CDict* cdict;
+    ZSTD_prefixDict prefixDict;   /* single-usage dictionary */
+
+    /* Multi-threading */
+
+    /* Tracing */
+};
+
+typedef enum { ZSTD_dtlm_fast, ZSTD_dtlm_full } ZSTD_dictTableLoadMethod_e;
+
+typedef enum {
+    ZSTD_noDict = 0,
+    ZSTD_extDict = 1,
+    ZSTD_dictMatchState = 2,
+    ZSTD_dedicatedDictSearch = 3
+} ZSTD_dictMode_e;
+
+typedef enum {
+    ZSTD_cpm_noAttachDict = 0,  /* Compression with ZSTD_noDict or ZSTD_extDict.
+                                 * In this mode we use both the srcSize and the dictSize
+                                 * when selecting and adjusting parameters.
+                                 */
+    ZSTD_cpm_attachDict = 1,    /* Compression with ZSTD_dictMatchState or ZSTD_dedicatedDictSearch.
+                                 * In this mode we only take the srcSize into account when selecting
+                                 * and adjusting parameters.
+                                 */
+    ZSTD_cpm_createCDict = 2,   /* Creating a CDict.
+                                 * In this mode we take both the source size and the dictionary size
+                                 * into account when selecting and adjusting the parameters.
+                                 */
+    ZSTD_cpm_unknown = 3,       /* ZSTD_getCParams, ZSTD_getParams, ZSTD_adjustParams.
+                                 * We don't know what these parameters are for. We default to the legacy
+                                 * behavior of taking both the source size and the dict size into account
+                                 * when selecting and adjusting parameters.
+                                 */
+} ZSTD_cParamMode_e;
+
+typedef size_t (*ZSTD_blockCompressor) (
+        ZSTD_matchState_t* bs, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_dictMode_e dictMode);
+
+
+MEM_STATIC U32 ZSTD_LLcode(U32 litLength)
+{
+    static const BYTE LL_Code[64] = {  0,  1,  2,  3,  4,  5,  6,  7,
+                                       8,  9, 10, 11, 12, 13, 14, 15,
+                                      16, 16, 17, 17, 18, 18, 19, 19,
+                                      20, 20, 20, 20, 21, 21, 21, 21,
+                                      22, 22, 22, 22, 22, 22, 22, 22,
+                                      23, 23, 23, 23, 23, 23, 23, 23,
+                                      24, 24, 24, 24, 24, 24, 24, 24,
+                                      24, 24, 24, 24, 24, 24, 24, 24 };
+    static const U32 LL_deltaCode = 19;
+    return (litLength > 63) ? ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength];
+}
+
+/* ZSTD_MLcode() :
+ * note : mlBase = matchLength - MINMATCH;
+ *        because it's the format it's stored in seqStore->sequences */
+MEM_STATIC U32 ZSTD_MLcode(U32 mlBase)
+{
+    static const BYTE ML_Code[128] = { 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15,
+                                      16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31,
+                                      32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 36, 36, 37, 37, 37, 37,
+                                      38, 38, 38, 38, 38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39,
+                                      40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40,
+                                      41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41,
+                                      42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42,
+                                      42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42 };
+    static const U32 ML_deltaCode = 36;
+    return (mlBase > 127) ? ZSTD_highbit32(mlBase) + ML_deltaCode : ML_Code[mlBase];
+}
+
+typedef struct repcodes_s {
+    U32 rep[3];
+} repcodes_t;
+
+MEM_STATIC repcodes_t ZSTD_updateRep(U32 const rep[3], U32 const offset, U32 const ll0)
+{
+    repcodes_t newReps;
+    if (offset >= ZSTD_REP_NUM) {  /* full offset */
+        newReps.rep[2] = rep[1];
+        newReps.rep[1] = rep[0];
+        newReps.rep[0] = offset - ZSTD_REP_MOVE;
+    } else {   /* repcode */
+        U32 const repCode = offset + ll0;
+        if (repCode > 0) {  /* note : if repCode==0, no change */
+            U32 const currentOffset = (repCode==ZSTD_REP_NUM) ? (rep[0] - 1) : rep[repCode];
+            newReps.rep[2] = (repCode >= 2) ? rep[1] : rep[2];
+            newReps.rep[1] = rep[0];
+            newReps.rep[0] = currentOffset;
+        } else {   /* repCode == 0 */
+            ZSTD_memcpy(&newReps, rep, sizeof(newReps));
+        }
+    }
+    return newReps;
+}
+
+/* ZSTD_cParam_withinBounds:
+ * @return 1 if value is within cParam bounds,
+ * 0 otherwise */
+MEM_STATIC int ZSTD_cParam_withinBounds(ZSTD_cParameter cParam, int value)
+{
+    ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam);
+    if (ZSTD_isError(bounds.error)) return 0;
+    if (value < bounds.lowerBound) return 0;
+    if (value > bounds.upperBound) return 0;
+    return 1;
+}
+
+/* ZSTD_noCompressBlock() :
+ * Writes uncompressed block to dst buffer from given src.
+ * Returns the size of the block */
+MEM_STATIC size_t ZSTD_noCompressBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize, U32 lastBlock)
+{
+    U32 const cBlockHeader24 = lastBlock + (((U32)bt_raw)<<1) + (U32)(srcSize << 3);
+    RETURN_ERROR_IF(srcSize + ZSTD_blockHeaderSize > dstCapacity,
+                    dstSize_tooSmall, "dst buf too small for uncompressed block");
+    MEM_writeLE24(dst, cBlockHeader24);
+    ZSTD_memcpy((BYTE*)dst + ZSTD_blockHeaderSize, src, srcSize);
+    return ZSTD_blockHeaderSize + srcSize;
+}
+
+MEM_STATIC size_t ZSTD_rleCompressBlock (void* dst, size_t dstCapacity, BYTE src, size_t srcSize, U32 lastBlock)
+{
+    BYTE* const op = (BYTE*)dst;
+    U32 const cBlockHeader = lastBlock + (((U32)bt_rle)<<1) + (U32)(srcSize << 3);
+    RETURN_ERROR_IF(dstCapacity < 4, dstSize_tooSmall, "");
+    MEM_writeLE24(op, cBlockHeader);
+    op[3] = src;
+    return 4;
+}
+
+
+/* ZSTD_minGain() :
+ * minimum compression required
+ * to generate a compress block or a compressed literals section.
+ * note : use same formula for both situations */
+MEM_STATIC size_t ZSTD_minGain(size_t srcSize, ZSTD_strategy strat)
+{
+    U32 const minlog = (strat>=ZSTD_btultra) ? (U32)(strat) - 1 : 6;
+    ZSTD_STATIC_ASSERT(ZSTD_btultra == 8);
+    assert(ZSTD_cParam_withinBounds(ZSTD_c_strategy, strat));
+    return (srcSize >> minlog) + 2;
+}
+
+MEM_STATIC int ZSTD_disableLiteralsCompression(const ZSTD_CCtx_params* cctxParams)
+{
+    switch (cctxParams->literalCompressionMode) {
+    case ZSTD_lcm_huffman:
+        return 0;
+    case ZSTD_lcm_uncompressed:
+        return 1;
+    default:
+        assert(0 /* impossible: pre-validated */);
+        ZSTD_FALLTHROUGH;
+    case ZSTD_lcm_auto:
+        return (cctxParams->cParams.strategy == ZSTD_fast) && (cctxParams->cParams.targetLength > 0);
+    }
+}
+
+/*! ZSTD_safecopyLiterals() :
+ *  memcpy() function that won't read beyond more than WILDCOPY_OVERLENGTH bytes past ilimit_w.
+ *  Only called when the sequence ends past ilimit_w, so it only needs to be optimized for single
+ *  large copies.
+ */
+static void ZSTD_safecopyLiterals(BYTE* op, BYTE const* ip, BYTE const* const iend, BYTE const* ilimit_w) {
+    assert(iend > ilimit_w);
+    if (ip <= ilimit_w) {
+        ZSTD_wildcopy(op, ip, ilimit_w - ip, ZSTD_no_overlap);
+        op += ilimit_w - ip;
+        ip = ilimit_w;
+    }
+    while (ip < iend) *op++ = *ip++;
+}
+
+/*! ZSTD_storeSeq() :
+ *  Store a sequence (litlen, litPtr, offCode and mlBase) into seqStore_t.
+ *  `offCode` : distance to match + ZSTD_REP_MOVE (values <= ZSTD_REP_MOVE are repCodes).
+ *  `mlBase` : matchLength - MINMATCH
+ *  Allowed to overread literals up to litLimit.
+*/
+HINT_INLINE UNUSED_ATTR
+void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const BYTE* literals, const BYTE* litLimit, U32 offCode, size_t mlBase)
+{
+    BYTE const* const litLimit_w = litLimit - WILDCOPY_OVERLENGTH;
+    BYTE const* const litEnd = literals + litLength;
+#if defined(DEBUGLEVEL) && (DEBUGLEVEL >= 6)
+    static const BYTE* g_start = NULL;
+    if (g_start==NULL) g_start = (const BYTE*)literals;  /* note : index only works for compression within a single segment */
+    {   U32 const pos = (U32)((const BYTE*)literals - g_start);
+        DEBUGLOG(6, "Cpos%7u :%3u literals, match%4u bytes at offCode%7u",
+               pos, (U32)litLength, (U32)mlBase+MINMATCH, (U32)offCode);
+    }
+#endif
+    assert((size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart) < seqStorePtr->maxNbSeq);
+    /* copy Literals */
+    assert(seqStorePtr->maxNbLit <= 128 KB);
+    assert(seqStorePtr->lit + litLength <= seqStorePtr->litStart + seqStorePtr->maxNbLit);
+    assert(literals + litLength <= litLimit);
+    if (litEnd <= litLimit_w) {
+        /* Common case we can use wildcopy.
+	 * First copy 16 bytes, because literals are likely short.
+	 */
+        assert(WILDCOPY_OVERLENGTH >= 16);
+        ZSTD_copy16(seqStorePtr->lit, literals);
+        if (litLength > 16) {
+            ZSTD_wildcopy(seqStorePtr->lit+16, literals+16, (ptrdiff_t)litLength-16, ZSTD_no_overlap);
+        }
+    } else {
+        ZSTD_safecopyLiterals(seqStorePtr->lit, literals, litEnd, litLimit_w);
+    }
+    seqStorePtr->lit += litLength;
+
+    /* literal Length */
+    if (litLength>0xFFFF) {
+        assert(seqStorePtr->longLengthID == 0); /* there can only be a single long length */
+        seqStorePtr->longLengthID = 1;
+        seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
+    }
+    seqStorePtr->sequences[0].litLength = (U16)litLength;
+
+    /* match offset */
+    seqStorePtr->sequences[0].offset = offCode + 1;
+
+    /* match Length */
+    if (mlBase>0xFFFF) {
+        assert(seqStorePtr->longLengthID == 0); /* there can only be a single long length */
+        seqStorePtr->longLengthID = 2;
+        seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
+    }
+    seqStorePtr->sequences[0].matchLength = (U16)mlBase;
+
+    seqStorePtr->sequences++;
+}
+
+
+/*-*************************************
+*  Match length counter
+***************************************/
+static unsigned ZSTD_NbCommonBytes (size_t val)
+{
+    if (MEM_isLittleEndian()) {
+        if (MEM_64bits()) {
+#       if (__GNUC__ >= 4)
+            return (__builtin_ctzll((U64)val) >> 3);
+#       else
+            static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2,
+                                                     0, 3, 1, 3, 1, 4, 2, 7,
+                                                     0, 2, 3, 6, 1, 5, 3, 5,
+                                                     1, 3, 4, 4, 2, 5, 6, 7,
+                                                     7, 0, 1, 2, 3, 3, 4, 6,
+                                                     2, 6, 5, 5, 3, 4, 5, 6,
+                                                     7, 1, 2, 4, 6, 4, 4, 5,
+                                                     7, 2, 6, 5, 7, 6, 7, 7 };
+            return DeBruijnBytePos[((U64)((val & -(long long)val) * 0x0218A392CDABBD3FULL)) >> 58];
+#       endif
+        } else { /* 32 bits */
+#       if (__GNUC__ >= 3)
+            return (__builtin_ctz((U32)val) >> 3);
+#       else
+            static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0,
+                                                     3, 2, 2, 1, 3, 2, 0, 1,
+                                                     3, 3, 1, 2, 2, 2, 2, 0,
+                                                     3, 1, 2, 0, 1, 0, 1, 1 };
+            return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
+#       endif
+        }
+    } else {  /* Big Endian CPU */
+        if (MEM_64bits()) {
+#       if (__GNUC__ >= 4)
+            return (__builtin_clzll(val) >> 3);
+#       else
+            unsigned r;
+            const unsigned n32 = sizeof(size_t)*4;   /* calculate this way due to compiler complaining in 32-bits mode */
+            if (!(val>>n32)) { r=4; } else { r=0; val>>=n32; }
+            if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
+            r += (!val);
+            return r;
+#       endif
+        } else { /* 32 bits */
+#       if (__GNUC__ >= 3)
+            return (__builtin_clz((U32)val) >> 3);
+#       else
+            unsigned r;
+            if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
+            r += (!val);
+            return r;
+#       endif
+    }   }
+}
+
+
+MEM_STATIC size_t ZSTD_count(const BYTE* pIn, const BYTE* pMatch, const BYTE* const pInLimit)
+{
+    const BYTE* const pStart = pIn;
+    const BYTE* const pInLoopLimit = pInLimit - (sizeof(size_t)-1);
+
+    if (pIn < pInLoopLimit) {
+        { size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn);
+          if (diff) return ZSTD_NbCommonBytes(diff); }
+        pIn+=sizeof(size_t); pMatch+=sizeof(size_t);
+        while (pIn < pInLoopLimit) {
+            size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn);
+            if (!diff) { pIn+=sizeof(size_t); pMatch+=sizeof(size_t); continue; }
+            pIn += ZSTD_NbCommonBytes(diff);
+            return (size_t)(pIn - pStart);
+    }   }
+    if (MEM_64bits() && (pIn<(pInLimit-3)) && (MEM_read32(pMatch) == MEM_read32(pIn))) { pIn+=4; pMatch+=4; }
+    if ((pIn<(pInLimit-1)) && (MEM_read16(pMatch) == MEM_read16(pIn))) { pIn+=2; pMatch+=2; }
+    if ((pIn<pInLimit) && (*pMatch == *pIn)) pIn++;
+    return (size_t)(pIn - pStart);
+}
+
+/* ZSTD_count_2segments() :
+ *  can count match length with `ip` & `match` in 2 different segments.
+ *  convention : on reaching mEnd, match count continue starting from iStart
+ */
+MEM_STATIC size_t
+ZSTD_count_2segments(const BYTE* ip, const BYTE* match,
+                     const BYTE* iEnd, const BYTE* mEnd, const BYTE* iStart)
+{
+    const BYTE* const vEnd = MIN( ip + (mEnd - match), iEnd);
+    size_t const matchLength = ZSTD_count(ip, match, vEnd);
+    if (match + matchLength != mEnd) return matchLength;
+    DEBUGLOG(7, "ZSTD_count_2segments: found a 2-parts match (current length==%zu)", matchLength);
+    DEBUGLOG(7, "distance from match beginning to end dictionary = %zi", mEnd - match);
+    DEBUGLOG(7, "distance from current pos to end buffer = %zi", iEnd - ip);
+    DEBUGLOG(7, "next byte : ip==%02X, istart==%02X", ip[matchLength], *iStart);
+    DEBUGLOG(7, "final match length = %zu", matchLength + ZSTD_count(ip+matchLength, iStart, iEnd));
+    return matchLength + ZSTD_count(ip+matchLength, iStart, iEnd);
+}
+
+
+/*-*************************************
+ *  Hashes
+ ***************************************/
+static const U32 prime3bytes = 506832829U;
+static U32    ZSTD_hash3(U32 u, U32 h) { return ((u << (32-24)) * prime3bytes)  >> (32-h) ; }
+MEM_STATIC size_t ZSTD_hash3Ptr(const void* ptr, U32 h) { return ZSTD_hash3(MEM_readLE32(ptr), h); } /* only in zstd_opt.h */
+
+static const U32 prime4bytes = 2654435761U;
+static U32    ZSTD_hash4(U32 u, U32 h) { return (u * prime4bytes) >> (32-h) ; }
+static size_t ZSTD_hash4Ptr(const void* ptr, U32 h) { return ZSTD_hash4(MEM_read32(ptr), h); }
+
+static const U64 prime5bytes = 889523592379ULL;
+static size_t ZSTD_hash5(U64 u, U32 h) { return (size_t)(((u  << (64-40)) * prime5bytes) >> (64-h)) ; }
+static size_t ZSTD_hash5Ptr(const void* p, U32 h) { return ZSTD_hash5(MEM_readLE64(p), h); }
+
+static const U64 prime6bytes = 227718039650203ULL;
+static size_t ZSTD_hash6(U64 u, U32 h) { return (size_t)(((u  << (64-48)) * prime6bytes) >> (64-h)) ; }
+static size_t ZSTD_hash6Ptr(const void* p, U32 h) { return ZSTD_hash6(MEM_readLE64(p), h); }
+
+static const U64 prime7bytes = 58295818150454627ULL;
+static size_t ZSTD_hash7(U64 u, U32 h) { return (size_t)(((u  << (64-56)) * prime7bytes) >> (64-h)) ; }
+static size_t ZSTD_hash7Ptr(const void* p, U32 h) { return ZSTD_hash7(MEM_readLE64(p), h); }
+
+static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL;
+static size_t ZSTD_hash8(U64 u, U32 h) { return (size_t)(((u) * prime8bytes) >> (64-h)) ; }
+static size_t ZSTD_hash8Ptr(const void* p, U32 h) { return ZSTD_hash8(MEM_readLE64(p), h); }
+
+MEM_STATIC FORCE_INLINE_ATTR
+size_t ZSTD_hashPtr(const void* p, U32 hBits, U32 mls)
+{
+    switch(mls)
+    {
+    default:
+    case 4: return ZSTD_hash4Ptr(p, hBits);
+    case 5: return ZSTD_hash5Ptr(p, hBits);
+    case 6: return ZSTD_hash6Ptr(p, hBits);
+    case 7: return ZSTD_hash7Ptr(p, hBits);
+    case 8: return ZSTD_hash8Ptr(p, hBits);
+    }
+}
+
+/* ZSTD_ipow() :
+ * Return base^exponent.
+ */
+static U64 ZSTD_ipow(U64 base, U64 exponent)
+{
+    U64 power = 1;
+    while (exponent) {
+      if (exponent & 1) power *= base;
+      exponent >>= 1;
+      base *= base;
+    }
+    return power;
+}
+
+#define ZSTD_ROLL_HASH_CHAR_OFFSET 10
+
+/* ZSTD_rollingHash_append() :
+ * Add the buffer to the hash value.
+ */
+static U64 ZSTD_rollingHash_append(U64 hash, void const* buf, size_t size)
+{
+    BYTE const* istart = (BYTE const*)buf;
+    size_t pos;
+    for (pos = 0; pos < size; ++pos) {
+        hash *= prime8bytes;
+        hash += istart[pos] + ZSTD_ROLL_HASH_CHAR_OFFSET;
+    }
+    return hash;
+}
+
+/* ZSTD_rollingHash_compute() :
+ * Compute the rolling hash value of the buffer.
+ */
+MEM_STATIC U64 ZSTD_rollingHash_compute(void const* buf, size_t size)
+{
+    return ZSTD_rollingHash_append(0, buf, size);
+}
+
+/* ZSTD_rollingHash_primePower() :
+ * Compute the primePower to be passed to ZSTD_rollingHash_rotate() for a hash
+ * over a window of length bytes.
+ */
+MEM_STATIC U64 ZSTD_rollingHash_primePower(U32 length)
+{
+    return ZSTD_ipow(prime8bytes, length - 1);
+}
+
+/* ZSTD_rollingHash_rotate() :
+ * Rotate the rolling hash by one byte.
+ */
+MEM_STATIC U64 ZSTD_rollingHash_rotate(U64 hash, BYTE toRemove, BYTE toAdd, U64 primePower)
+{
+    hash -= (toRemove + ZSTD_ROLL_HASH_CHAR_OFFSET) * primePower;
+    hash *= prime8bytes;
+    hash += toAdd + ZSTD_ROLL_HASH_CHAR_OFFSET;
+    return hash;
+}
+
+/*-*************************************
+*  Round buffer management
+***************************************/
+#if (ZSTD_WINDOWLOG_MAX_64 > 31)
+# error "ZSTD_WINDOWLOG_MAX is too large : would overflow ZSTD_CURRENT_MAX"
+#endif
+/* Max current allowed */
+#define ZSTD_CURRENT_MAX ((3U << 29) + (1U << ZSTD_WINDOWLOG_MAX))
+/* Maximum chunk size before overflow correction needs to be called again */
+#define ZSTD_CHUNKSIZE_MAX                                                     \
+    ( ((U32)-1)                  /* Maximum ending current index */            \
+    - ZSTD_CURRENT_MAX)          /* Maximum beginning lowLimit */
+
+/*
+ * ZSTD_window_clear():
+ * Clears the window containing the history by simply setting it to empty.
+ */
+MEM_STATIC void ZSTD_window_clear(ZSTD_window_t* window)
+{
+    size_t const endT = (size_t)(window->nextSrc - window->base);
+    U32 const end = (U32)endT;
+
+    window->lowLimit = end;
+    window->dictLimit = end;
+}
+
+/*
+ * ZSTD_window_hasExtDict():
+ * Returns non-zero if the window has a non-empty extDict.
+ */
+MEM_STATIC U32 ZSTD_window_hasExtDict(ZSTD_window_t const window)
+{
+    return window.lowLimit < window.dictLimit;
+}
+
+/*
+ * ZSTD_matchState_dictMode():
+ * Inspects the provided matchState and figures out what dictMode should be
+ * passed to the compressor.
+ */
+MEM_STATIC ZSTD_dictMode_e ZSTD_matchState_dictMode(const ZSTD_matchState_t *ms)
+{
+    return ZSTD_window_hasExtDict(ms->window) ?
+        ZSTD_extDict :
+        ms->dictMatchState != NULL ?
+            (ms->dictMatchState->dedicatedDictSearch ? ZSTD_dedicatedDictSearch : ZSTD_dictMatchState) :
+            ZSTD_noDict;
+}
+
+/*
+ * ZSTD_window_needOverflowCorrection():
+ * Returns non-zero if the indices are getting too large and need overflow
+ * protection.
+ */
+MEM_STATIC U32 ZSTD_window_needOverflowCorrection(ZSTD_window_t const window,
+                                                  void const* srcEnd)
+{
+    U32 const curr = (U32)((BYTE const*)srcEnd - window.base);
+    return curr > ZSTD_CURRENT_MAX;
+}
+
+/*
+ * ZSTD_window_correctOverflow():
+ * Reduces the indices to protect from index overflow.
+ * Returns the correction made to the indices, which must be applied to every
+ * stored index.
+ *
+ * The least significant cycleLog bits of the indices must remain the same,
+ * which may be 0. Every index up to maxDist in the past must be valid.
+ * NOTE: (maxDist & cycleMask) must be zero.
+ */
+MEM_STATIC U32 ZSTD_window_correctOverflow(ZSTD_window_t* window, U32 cycleLog,
+                                           U32 maxDist, void const* src)
+{
+    /* preemptive overflow correction:
+     * 1. correction is large enough:
+     *    lowLimit > (3<<29) ==> current > 3<<29 + 1<<windowLog
+     *    1<<windowLog <= newCurrent < 1<<chainLog + 1<<windowLog
+     *
+     *    current - newCurrent
+     *    > (3<<29 + 1<<windowLog) - (1<<windowLog + 1<<chainLog)
+     *    > (3<<29) - (1<<chainLog)
+     *    > (3<<29) - (1<<30)             (NOTE: chainLog <= 30)
+     *    > 1<<29
+     *
+     * 2. (ip+ZSTD_CHUNKSIZE_MAX - cctx->base) doesn't overflow:
+     *    After correction, current is less than (1<<chainLog + 1<<windowLog).
+     *    In 64-bit mode we are safe, because we have 64-bit ptrdiff_t.
+     *    In 32-bit mode we are safe, because (chainLog <= 29), so
+     *    ip+ZSTD_CHUNKSIZE_MAX - cctx->base < 1<<32.
+     * 3. (cctx->lowLimit + 1<<windowLog) < 1<<32:
+     *    windowLog <= 31 ==> 3<<29 + 1<<windowLog < 7<<29 < 1<<32.
+     */
+    U32 const cycleMask = (1U << cycleLog) - 1;
+    U32 const curr = (U32)((BYTE const*)src - window->base);
+    U32 const currentCycle0 = curr & cycleMask;
+    /* Exclude zero so that newCurrent - maxDist >= 1. */
+    U32 const currentCycle1 = currentCycle0 == 0 ? (1U << cycleLog) : currentCycle0;
+    U32 const newCurrent = currentCycle1 + maxDist;
+    U32 const correction = curr - newCurrent;
+    assert((maxDist & cycleMask) == 0);
+    assert(curr > newCurrent);
+    /* Loose bound, should be around 1<<29 (see above) */
+    assert(correction > 1<<28);
+
+    window->base += correction;
+    window->dictBase += correction;
+    if (window->lowLimit <= correction) window->lowLimit = 1;
+    else window->lowLimit -= correction;
+    if (window->dictLimit <= correction) window->dictLimit = 1;
+    else window->dictLimit -= correction;
+
+    /* Ensure we can still reference the full window. */
+    assert(newCurrent >= maxDist);
+    assert(newCurrent - maxDist >= 1);
+    /* Ensure that lowLimit and dictLimit didn't underflow. */
+    assert(window->lowLimit <= newCurrent);
+    assert(window->dictLimit <= newCurrent);
+
+    DEBUGLOG(4, "Correction of 0x%x bytes to lowLimit=0x%x", correction,
+             window->lowLimit);
+    return correction;
+}
+
+/*
+ * ZSTD_window_enforceMaxDist():
+ * Updates lowLimit so that:
+ *    (srcEnd - base) - lowLimit == maxDist + loadedDictEnd
+ *
+ * It ensures index is valid as long as index >= lowLimit.
+ * This must be called before a block compression call.
+ *
+ * loadedDictEnd is only defined if a dictionary is in use for current compression.
+ * As the name implies, loadedDictEnd represents the index at end of dictionary.
+ * The value lies within context's referential, it can be directly compared to blockEndIdx.
+ *
+ * If loadedDictEndPtr is NULL, no dictionary is in use, and we use loadedDictEnd == 0.
+ * If loadedDictEndPtr is not NULL, we set it to zero after updating lowLimit.
+ * This is because dictionaries are allowed to be referenced fully
+ * as long as the last byte of the dictionary is in the window.
+ * Once input has progressed beyond window size, dictionary cannot be referenced anymore.
+ *
+ * In normal dict mode, the dictionary lies between lowLimit and dictLimit.
+ * In dictMatchState mode, lowLimit and dictLimit are the same,
+ * and the dictionary is below them.
+ * forceWindow and dictMatchState are therefore incompatible.
+ */
+MEM_STATIC void
+ZSTD_window_enforceMaxDist(ZSTD_window_t* window,
+                     const void* blockEnd,
+                           U32   maxDist,
+                           U32*  loadedDictEndPtr,
+                     const ZSTD_matchState_t** dictMatchStatePtr)
+{
+    U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base);
+    U32 const loadedDictEnd = (loadedDictEndPtr != NULL) ? *loadedDictEndPtr : 0;
+    DEBUGLOG(5, "ZSTD_window_enforceMaxDist: blockEndIdx=%u, maxDist=%u, loadedDictEnd=%u",
+                (unsigned)blockEndIdx, (unsigned)maxDist, (unsigned)loadedDictEnd);
+
+    /* - When there is no dictionary : loadedDictEnd == 0.
+         In which case, the test (blockEndIdx > maxDist) is merely to avoid
+         overflowing next operation `newLowLimit = blockEndIdx - maxDist`.
+       - When there is a standard dictionary :
+         Index referential is copied from the dictionary,
+         which means it starts from 0.
+         In which case, loadedDictEnd == dictSize,
+         and it makes sense to compare `blockEndIdx > maxDist + dictSize`
+         since `blockEndIdx` also starts from zero.
+       - When there is an attached dictionary :
+         loadedDictEnd is expressed within the referential of the context,
+         so it can be directly compared against blockEndIdx.
+    */
+    if (blockEndIdx > maxDist + loadedDictEnd) {
+        U32 const newLowLimit = blockEndIdx - maxDist;
+        if (window->lowLimit < newLowLimit) window->lowLimit = newLowLimit;
+        if (window->dictLimit < window->lowLimit) {
+            DEBUGLOG(5, "Update dictLimit to match lowLimit, from %u to %u",
+                        (unsigned)window->dictLimit, (unsigned)window->lowLimit);
+            window->dictLimit = window->lowLimit;
+        }
+        /* On reaching window size, dictionaries are invalidated */
+        if (loadedDictEndPtr) *loadedDictEndPtr = 0;
+        if (dictMatchStatePtr) *dictMatchStatePtr = NULL;
+    }
+}
+
+/* Similar to ZSTD_window_enforceMaxDist(),
+ * but only invalidates dictionary
+ * when input progresses beyond window size.
+ * assumption : loadedDictEndPtr and dictMatchStatePtr are valid (non NULL)
+ *              loadedDictEnd uses same referential as window->base
+ *              maxDist is the window size */
+MEM_STATIC void
+ZSTD_checkDictValidity(const ZSTD_window_t* window,
+                       const void* blockEnd,
+                             U32   maxDist,
+                             U32*  loadedDictEndPtr,
+                       const ZSTD_matchState_t** dictMatchStatePtr)
+{
+    assert(loadedDictEndPtr != NULL);
+    assert(dictMatchStatePtr != NULL);
+    {   U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base);
+        U32 const loadedDictEnd = *loadedDictEndPtr;
+        DEBUGLOG(5, "ZSTD_checkDictValidity: blockEndIdx=%u, maxDist=%u, loadedDictEnd=%u",
+                    (unsigned)blockEndIdx, (unsigned)maxDist, (unsigned)loadedDictEnd);
+        assert(blockEndIdx >= loadedDictEnd);
+
+        if (blockEndIdx > loadedDictEnd + maxDist) {
+            /* On reaching window size, dictionaries are invalidated.
+             * For simplification, if window size is reached anywhere within next block,
+             * the dictionary is invalidated for the full block.
+             */
+            DEBUGLOG(6, "invalidating dictionary for current block (distance > windowSize)");
+            *loadedDictEndPtr = 0;
+            *dictMatchStatePtr = NULL;
+        } else {
+            if (*loadedDictEndPtr != 0) {
+                DEBUGLOG(6, "dictionary considered valid for current block");
+    }   }   }
+}
+
+MEM_STATIC void ZSTD_window_init(ZSTD_window_t* window) {
+    ZSTD_memset(window, 0, sizeof(*window));
+    window->base = (BYTE const*)"";
+    window->dictBase = (BYTE const*)"";
+    window->dictLimit = 1;    /* start from 1, so that 1st position is valid */
+    window->lowLimit = 1;     /* it ensures first and later CCtx usages compress the same */
+    window->nextSrc = window->base + 1;   /* see issue #1241 */
+}
+
+/*
+ * ZSTD_window_update():
+ * Updates the window by appending [src, src + srcSize) to the window.
+ * If it is not contiguous, the current prefix becomes the extDict, and we
+ * forget about the extDict. Handles overlap of the prefix and extDict.
+ * Returns non-zero if the segment is contiguous.
+ */
+MEM_STATIC U32 ZSTD_window_update(ZSTD_window_t* window,
+                                  void const* src, size_t srcSize)
+{
+    BYTE const* const ip = (BYTE const*)src;
+    U32 contiguous = 1;
+    DEBUGLOG(5, "ZSTD_window_update");
+    if (srcSize == 0)
+        return contiguous;
+    assert(window->base != NULL);
+    assert(window->dictBase != NULL);
+    /* Check if blocks follow each other */
+    if (src != window->nextSrc) {
+        /* not contiguous */
+        size_t const distanceFromBase = (size_t)(window->nextSrc - window->base);
+        DEBUGLOG(5, "Non contiguous blocks, new segment starts at %u", window->dictLimit);
+        window->lowLimit = window->dictLimit;
+        assert(distanceFromBase == (size_t)(U32)distanceFromBase);  /* should never overflow */
+        window->dictLimit = (U32)distanceFromBase;
+        window->dictBase = window->base;
+        window->base = ip - distanceFromBase;
+        /* ms->nextToUpdate = window->dictLimit; */
+        if (window->dictLimit - window->lowLimit < HASH_READ_SIZE) window->lowLimit = window->dictLimit;   /* too small extDict */
+        contiguous = 0;
+    }
+    window->nextSrc = ip + srcSize;
+    /* if input and dictionary overlap : reduce dictionary (area presumed modified by input) */
+    if ( (ip+srcSize > window->dictBase + window->lowLimit)
+       & (ip < window->dictBase + window->dictLimit)) {
+        ptrdiff_t const highInputIdx = (ip + srcSize) - window->dictBase;
+        U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)window->dictLimit) ? window->dictLimit : (U32)highInputIdx;
+        window->lowLimit = lowLimitMax;
+        DEBUGLOG(5, "Overlapping extDict and input : new lowLimit = %u", window->lowLimit);
+    }
+    return contiguous;
+}
+
+/*
+ * Returns the lowest allowed match index. It may either be in the ext-dict or the prefix.
+ */
+MEM_STATIC U32 ZSTD_getLowestMatchIndex(const ZSTD_matchState_t* ms, U32 curr, unsigned windowLog)
+{
+    U32    const maxDistance = 1U << windowLog;
+    U32    const lowestValid = ms->window.lowLimit;
+    U32    const withinWindow = (curr - lowestValid > maxDistance) ? curr - maxDistance : lowestValid;
+    U32    const isDictionary = (ms->loadedDictEnd != 0);
+    /* When using a dictionary the entire dictionary is valid if a single byte of the dictionary
+     * is within the window. We invalidate the dictionary (and set loadedDictEnd to 0) when it isn't
+     * valid for the entire block. So this check is sufficient to find the lowest valid match index.
+     */
+    U32    const matchLowest = isDictionary ? lowestValid : withinWindow;
+    return matchLowest;
+}
+
+/*
+ * Returns the lowest allowed match index in the prefix.
+ */
+MEM_STATIC U32 ZSTD_getLowestPrefixIndex(const ZSTD_matchState_t* ms, U32 curr, unsigned windowLog)
+{
+    U32    const maxDistance = 1U << windowLog;
+    U32    const lowestValid = ms->window.dictLimit;
+    U32    const withinWindow = (curr - lowestValid > maxDistance) ? curr - maxDistance : lowestValid;
+    U32    const isDictionary = (ms->loadedDictEnd != 0);
+    /* When computing the lowest prefix index we need to take the dictionary into account to handle
+     * the edge case where the dictionary and the source are contiguous in memory.
+     */
+    U32    const matchLowest = isDictionary ? lowestValid : withinWindow;
+    return matchLowest;
+}
+
+
+
+/* debug functions */
+#if (DEBUGLEVEL>=2)
+
+MEM_STATIC double ZSTD_fWeight(U32 rawStat)
+{
+    U32 const fp_accuracy = 8;
+    U32 const fp_multiplier = (1 << fp_accuracy);
+    U32 const newStat = rawStat + 1;
+    U32 const hb = ZSTD_highbit32(newStat);
+    U32 const BWeight = hb * fp_multiplier;
+    U32 const FWeight = (newStat << fp_accuracy) >> hb;
+    U32 const weight = BWeight + FWeight;
+    assert(hb + fp_accuracy < 31);
+    return (double)weight / fp_multiplier;
+}
+
+/* display a table content,
+ * listing each element, its frequency, and its predicted bit cost */
+MEM_STATIC void ZSTD_debugTable(const U32* table, U32 max)
+{
+    unsigned u, sum;
+    for (u=0, sum=0; u<=max; u++) sum += table[u];
+    DEBUGLOG(2, "total nb elts: %u", sum);
+    for (u=0; u<=max; u++) {
+        DEBUGLOG(2, "%2u: %5u  (%.2f)",
+                u, table[u], ZSTD_fWeight(sum) - ZSTD_fWeight(table[u]) );
+    }
+}
+
+#endif
+
+
+
+/* ===============================================================
+ * Shared internal declarations
+ * These prototypes may be called from sources not in lib/compress
+ * =============================================================== */
+
+/* ZSTD_loadCEntropy() :
+ * dict : must point at beginning of a valid zstd dictionary.
+ * return : size of dictionary header (size of magic number + dict ID + entropy tables)
+ * assumptions : magic number supposed already checked
+ *               and dictSize >= 8 */
+size_t ZSTD_loadCEntropy(ZSTD_compressedBlockState_t* bs, void* workspace,
+                         const void* const dict, size_t dictSize);
+
+void ZSTD_reset_compressedBlockState(ZSTD_compressedBlockState_t* bs);
+
+/* ==============================================================
+ * Private declarations
+ * These prototypes shall only be called from within lib/compress
+ * ============================================================== */
+
+/* ZSTD_getCParamsFromCCtxParams() :
+ * cParams are built depending on compressionLevel, src size hints,
+ * LDM and manually set compression parameters.
+ * Note: srcSizeHint == 0 means 0!
+ */
+ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams(
+        const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize, ZSTD_cParamMode_e mode);
+
+/*! ZSTD_initCStream_internal() :
+ *  Private use only. Init streaming operation.
+ *  expects params to be valid.
+ *  must receive dict, or cdict, or none, but not both.
+ *  @return : 0, or an error code */
+size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,
+                     const void* dict, size_t dictSize,
+                     const ZSTD_CDict* cdict,
+                     const ZSTD_CCtx_params* params, unsigned long long pledgedSrcSize);
+
+void ZSTD_resetSeqStore(seqStore_t* ssPtr);
+
+/*! ZSTD_getCParamsFromCDict() :
+ *  as the name implies */
+ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict);
+
+/* ZSTD_compressBegin_advanced_internal() :
+ * Private use only. To be called from zstdmt_compress.c. */
+size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx,
+                                    const void* dict, size_t dictSize,
+                                    ZSTD_dictContentType_e dictContentType,
+                                    ZSTD_dictTableLoadMethod_e dtlm,
+                                    const ZSTD_CDict* cdict,
+                                    const ZSTD_CCtx_params* params,
+                                    unsigned long long pledgedSrcSize);
+
+/* ZSTD_compress_advanced_internal() :
+ * Private use only. To be called from zstdmt_compress.c. */
+size_t ZSTD_compress_advanced_internal(ZSTD_CCtx* cctx,
+                                       void* dst, size_t dstCapacity,
+                                 const void* src, size_t srcSize,
+                                 const void* dict,size_t dictSize,
+                                 const ZSTD_CCtx_params* params);
+
+
+/* ZSTD_writeLastEmptyBlock() :
+ * output an empty Block with end-of-frame mark to complete a frame
+ * @return : size of data written into `dst` (== ZSTD_blockHeaderSize (defined in zstd_internal.h))
+ *           or an error code if `dstCapacity` is too small (<ZSTD_blockHeaderSize)
+ */
+size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity);
+
+
+/* ZSTD_referenceExternalSequences() :
+ * Must be called before starting a compression operation.
+ * seqs must parse a prefix of the source.
+ * This cannot be used when long range matching is enabled.
+ * Zstd will use these sequences, and pass the literals to a secondary block
+ * compressor.
+ * @return : An error code on failure.
+ * NOTE: seqs are not verified! Invalid sequences can cause out-of-bounds memory
+ * access and data corruption.
+ */
+size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq);
+
+/* ZSTD_cycleLog() :
+ *  condition for correct operation : hashLog > 1 */
+U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat);
+
+/* ZSTD_CCtx_trace() :
+ *  Trace the end of a compression call.
+ */
+void ZSTD_CCtx_trace(ZSTD_CCtx* cctx, size_t extraCSize);
+
+#endif /* ZSTD_COMPRESS_H */
diff --git a/lib/zstd/compress/zstd_compress_literals.c b/lib/zstd/compress/zstd_compress_literals.c
new file mode 100644
index 0000000..655bcda
--- /dev/null
+++ b/lib/zstd/compress/zstd_compress_literals.c
@@ -0,0 +1,158 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+ /*-*************************************
+ *  Dependencies
+ ***************************************/
+#include "zstd_compress_literals.h"
+
+size_t ZSTD_noCompressLiterals (void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+    BYTE* const ostart = (BYTE*)dst;
+    U32   const flSize = 1 + (srcSize>31) + (srcSize>4095);
+
+    RETURN_ERROR_IF(srcSize + flSize > dstCapacity, dstSize_tooSmall, "");
+
+    switch(flSize)
+    {
+        case 1: /* 2 - 1 - 5 */
+            ostart[0] = (BYTE)((U32)set_basic + (srcSize<<3));
+            break;
+        case 2: /* 2 - 2 - 12 */
+            MEM_writeLE16(ostart, (U16)((U32)set_basic + (1<<2) + (srcSize<<4)));
+            break;
+        case 3: /* 2 - 2 - 20 */
+            MEM_writeLE32(ostart, (U32)((U32)set_basic + (3<<2) + (srcSize<<4)));
+            break;
+        default:   /* not necessary : flSize is {1,2,3} */
+            assert(0);
+    }
+
+    ZSTD_memcpy(ostart + flSize, src, srcSize);
+    DEBUGLOG(5, "Raw literals: %u -> %u", (U32)srcSize, (U32)(srcSize + flSize));
+    return srcSize + flSize;
+}
+
+size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+    BYTE* const ostart = (BYTE*)dst;
+    U32   const flSize = 1 + (srcSize>31) + (srcSize>4095);
+
+    (void)dstCapacity;  /* dstCapacity already guaranteed to be >=4, hence large enough */
+
+    switch(flSize)
+    {
+        case 1: /* 2 - 1 - 5 */
+            ostart[0] = (BYTE)((U32)set_rle + (srcSize<<3));
+            break;
+        case 2: /* 2 - 2 - 12 */
+            MEM_writeLE16(ostart, (U16)((U32)set_rle + (1<<2) + (srcSize<<4)));
+            break;
+        case 3: /* 2 - 2 - 20 */
+            MEM_writeLE32(ostart, (U32)((U32)set_rle + (3<<2) + (srcSize<<4)));
+            break;
+        default:   /* not necessary : flSize is {1,2,3} */
+            assert(0);
+    }
+
+    ostart[flSize] = *(const BYTE*)src;
+    DEBUGLOG(5, "RLE literals: %u -> %u", (U32)srcSize, (U32)flSize + 1);
+    return flSize+1;
+}
+
+size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,
+                              ZSTD_hufCTables_t* nextHuf,
+                              ZSTD_strategy strategy, int disableLiteralCompression,
+                              void* dst, size_t dstCapacity,
+                        const void* src, size_t srcSize,
+                              void* entropyWorkspace, size_t entropyWorkspaceSize,
+                        const int bmi2)
+{
+    size_t const minGain = ZSTD_minGain(srcSize, strategy);
+    size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB);
+    BYTE*  const ostart = (BYTE*)dst;
+    U32 singleStream = srcSize < 256;
+    symbolEncodingType_e hType = set_compressed;
+    size_t cLitSize;
+
+    DEBUGLOG(5,"ZSTD_compressLiterals (disableLiteralCompression=%i srcSize=%u)",
+                disableLiteralCompression, (U32)srcSize);
+
+    /* Prepare nextEntropy assuming reusing the existing table */
+    ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
+
+    if (disableLiteralCompression)
+        return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
+
+    /* small ? don't even attempt compression (speed opt) */
+#   define COMPRESS_LITERALS_SIZE_MIN 63
+    {   size_t const minLitSize = (prevHuf->repeatMode == HUF_repeat_valid) ? 6 : COMPRESS_LITERALS_SIZE_MIN;
+        if (srcSize <= minLitSize) return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
+    }
+
+    RETURN_ERROR_IF(dstCapacity < lhSize+1, dstSize_tooSmall, "not enough space for compression");
+    {   HUF_repeat repeat = prevHuf->repeatMode;
+        int const preferRepeat = strategy < ZSTD_lazy ? srcSize <= 1024 : 0;
+        if (repeat == HUF_repeat_valid && lhSize == 3) singleStream = 1;
+        cLitSize = singleStream ?
+            HUF_compress1X_repeat(
+                ostart+lhSize, dstCapacity-lhSize, src, srcSize,
+                HUF_SYMBOLVALUE_MAX, HUF_TABLELOG_DEFAULT, entropyWorkspace, entropyWorkspaceSize,
+                (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2) :
+            HUF_compress4X_repeat(
+                ostart+lhSize, dstCapacity-lhSize, src, srcSize,
+                HUF_SYMBOLVALUE_MAX, HUF_TABLELOG_DEFAULT, entropyWorkspace, entropyWorkspaceSize,
+                (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2);
+        if (repeat != HUF_repeat_none) {
+            /* reused the existing table */
+            DEBUGLOG(5, "Reusing previous huffman table");
+            hType = set_repeat;
+        }
+    }
+
+    if ((cLitSize==0) | (cLitSize >= srcSize - minGain) | ERR_isError(cLitSize)) {
+        ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
+        return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
+    }
+    if (cLitSize==1) {
+        ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
+        return ZSTD_compressRleLiteralsBlock(dst, dstCapacity, src, srcSize);
+    }
+
+    if (hType == set_compressed) {
+        /* using a newly constructed table */
+        nextHuf->repeatMode = HUF_repeat_check;
+    }
+
+    /* Build header */
+    switch(lhSize)
+    {
+    case 3: /* 2 - 2 - 10 - 10 */
+        {   U32 const lhc = hType + ((!singleStream) << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<14);
+            MEM_writeLE24(ostart, lhc);
+            break;
+        }
+    case 4: /* 2 - 2 - 14 - 14 */
+        {   U32 const lhc = hType + (2 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<18);
+            MEM_writeLE32(ostart, lhc);
+            break;
+        }
+    case 5: /* 2 - 2 - 18 - 18 */
+        {   U32 const lhc = hType + (3 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<22);
+            MEM_writeLE32(ostart, lhc);
+            ostart[4] = (BYTE)(cLitSize >> 10);
+            break;
+        }
+    default:  /* not possible : lhSize is {3,4,5} */
+        assert(0);
+    }
+    DEBUGLOG(5, "Compressed literals: %u -> %u", (U32)srcSize, (U32)(lhSize+cLitSize));
+    return lhSize+cLitSize;
+}
diff --git a/lib/zstd/compress/zstd_compress_literals.h b/lib/zstd/compress/zstd_compress_literals.h
new file mode 100644
index 0000000..9904c0c
--- /dev/null
+++ b/lib/zstd/compress/zstd_compress_literals.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_COMPRESS_LITERALS_H
+#define ZSTD_COMPRESS_LITERALS_H
+
+#include "zstd_compress_internal.h" /* ZSTD_hufCTables_t, ZSTD_minGain() */
+
+
+size_t ZSTD_noCompressLiterals (void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+
+size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+
+size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,
+                              ZSTD_hufCTables_t* nextHuf,
+                              ZSTD_strategy strategy, int disableLiteralCompression,
+                              void* dst, size_t dstCapacity,
+                        const void* src, size_t srcSize,
+                              void* entropyWorkspace, size_t entropyWorkspaceSize,
+                        const int bmi2);
+
+#endif /* ZSTD_COMPRESS_LITERALS_H */
diff --git a/lib/zstd/compress/zstd_compress_sequences.c b/lib/zstd/compress/zstd_compress_sequences.c
new file mode 100644
index 0000000..dcfcdc9
--- /dev/null
+++ b/lib/zstd/compress/zstd_compress_sequences.c
@@ -0,0 +1,439 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+ /*-*************************************
+ *  Dependencies
+ ***************************************/
+#include "zstd_compress_sequences.h"
+
+/*
+ * -log2(x / 256) lookup table for x in [0, 256).
+ * If x == 0: Return 0
+ * Else: Return floor(-log2(x / 256) * 256)
+ */
+static unsigned const kInverseProbabilityLog256[256] = {
+    0,    2048, 1792, 1642, 1536, 1453, 1386, 1329, 1280, 1236, 1197, 1162,
+    1130, 1100, 1073, 1047, 1024, 1001, 980,  960,  941,  923,  906,  889,
+    874,  859,  844,  830,  817,  804,  791,  779,  768,  756,  745,  734,
+    724,  714,  704,  694,  685,  676,  667,  658,  650,  642,  633,  626,
+    618,  610,  603,  595,  588,  581,  574,  567,  561,  554,  548,  542,
+    535,  529,  523,  517,  512,  506,  500,  495,  489,  484,  478,  473,
+    468,  463,  458,  453,  448,  443,  438,  434,  429,  424,  420,  415,
+    411,  407,  402,  398,  394,  390,  386,  382,  377,  373,  370,  366,
+    362,  358,  354,  350,  347,  343,  339,  336,  332,  329,  325,  322,
+    318,  315,  311,  308,  305,  302,  298,  295,  292,  289,  286,  282,
+    279,  276,  273,  270,  267,  264,  261,  258,  256,  253,  250,  247,
+    244,  241,  239,  236,  233,  230,  228,  225,  222,  220,  217,  215,
+    212,  209,  207,  204,  202,  199,  197,  194,  192,  190,  187,  185,
+    182,  180,  178,  175,  173,  171,  168,  166,  164,  162,  159,  157,
+    155,  153,  151,  149,  146,  144,  142,  140,  138,  136,  134,  132,
+    130,  128,  126,  123,  121,  119,  117,  115,  114,  112,  110,  108,
+    106,  104,  102,  100,  98,   96,   94,   93,   91,   89,   87,   85,
+    83,   82,   80,   78,   76,   74,   73,   71,   69,   67,   66,   64,
+    62,   61,   59,   57,   55,   54,   52,   50,   49,   47,   46,   44,
+    42,   41,   39,   37,   36,   34,   33,   31,   30,   28,   26,   25,
+    23,   22,   20,   19,   17,   16,   14,   13,   11,   10,   8,    7,
+    5,    4,    2,    1,
+};
+
+static unsigned ZSTD_getFSEMaxSymbolValue(FSE_CTable const* ctable) {
+  void const* ptr = ctable;
+  U16 const* u16ptr = (U16 const*)ptr;
+  U32 const maxSymbolValue = MEM_read16(u16ptr + 1);
+  return maxSymbolValue;
+}
+
+/*
+ * Returns true if we should use ncount=-1 else we should
+ * use ncount=1 for low probability symbols instead.
+ */
+static unsigned ZSTD_useLowProbCount(size_t const nbSeq)
+{
+    /* Heuristic: This should cover most blocks <= 16K and
+     * start to fade out after 16K to about 32K depending on
+     * comprssibility.
+     */
+    return nbSeq >= 2048;
+}
+
+/*
+ * Returns the cost in bytes of encoding the normalized count header.
+ * Returns an error if any of the helper functions return an error.
+ */
+static size_t ZSTD_NCountCost(unsigned const* count, unsigned const max,
+                              size_t const nbSeq, unsigned const FSELog)
+{
+    BYTE wksp[FSE_NCOUNTBOUND];
+    S16 norm[MaxSeq + 1];
+    const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max);
+    FORWARD_IF_ERROR(FSE_normalizeCount(norm, tableLog, count, nbSeq, max, ZSTD_useLowProbCount(nbSeq)), "");
+    return FSE_writeNCount(wksp, sizeof(wksp), norm, max, tableLog);
+}
+
+/*
+ * Returns the cost in bits of encoding the distribution described by count
+ * using the entropy bound.
+ */
+static size_t ZSTD_entropyCost(unsigned const* count, unsigned const max, size_t const total)
+{
+    unsigned cost = 0;
+    unsigned s;
+    for (s = 0; s <= max; ++s) {
+        unsigned norm = (unsigned)((256 * count[s]) / total);
+        if (count[s] != 0 && norm == 0)
+            norm = 1;
+        assert(count[s] < total);
+        cost += count[s] * kInverseProbabilityLog256[norm];
+    }
+    return cost >> 8;
+}
+
+/*
+ * Returns the cost in bits of encoding the distribution in count using ctable.
+ * Returns an error if ctable cannot represent all the symbols in count.
+ */
+size_t ZSTD_fseBitCost(
+    FSE_CTable const* ctable,
+    unsigned const* count,
+    unsigned const max)
+{
+    unsigned const kAccuracyLog = 8;
+    size_t cost = 0;
+    unsigned s;
+    FSE_CState_t cstate;
+    FSE_initCState(&cstate, ctable);
+    if (ZSTD_getFSEMaxSymbolValue(ctable) < max) {
+        DEBUGLOG(5, "Repeat FSE_CTable has maxSymbolValue %u < %u",
+                    ZSTD_getFSEMaxSymbolValue(ctable), max);
+        return ERROR(GENERIC);
+    }
+    for (s = 0; s <= max; ++s) {
+        unsigned const tableLog = cstate.stateLog;
+        unsigned const badCost = (tableLog + 1) << kAccuracyLog;
+        unsigned const bitCost = FSE_bitCost(cstate.symbolTT, tableLog, s, kAccuracyLog);
+        if (count[s] == 0)
+            continue;
+        if (bitCost >= badCost) {
+            DEBUGLOG(5, "Repeat FSE_CTable has Prob[%u] == 0", s);
+            return ERROR(GENERIC);
+        }
+        cost += (size_t)count[s] * bitCost;
+    }
+    return cost >> kAccuracyLog;
+}
+
+/*
+ * Returns the cost in bits of encoding the distribution in count using the
+ * table described by norm. The max symbol support by norm is assumed >= max.
+ * norm must be valid for every symbol with non-zero probability in count.
+ */
+size_t ZSTD_crossEntropyCost(short const* norm, unsigned accuracyLog,
+                             unsigned const* count, unsigned const max)
+{
+    unsigned const shift = 8 - accuracyLog;
+    size_t cost = 0;
+    unsigned s;
+    assert(accuracyLog <= 8);
+    for (s = 0; s <= max; ++s) {
+        unsigned const normAcc = (norm[s] != -1) ? (unsigned)norm[s] : 1;
+        unsigned const norm256 = normAcc << shift;
+        assert(norm256 > 0);
+        assert(norm256 < 256);
+        cost += count[s] * kInverseProbabilityLog256[norm256];
+    }
+    return cost >> 8;
+}
+
+symbolEncodingType_e
+ZSTD_selectEncodingType(
+        FSE_repeat* repeatMode, unsigned const* count, unsigned const max,
+        size_t const mostFrequent, size_t nbSeq, unsigned const FSELog,
+        FSE_CTable const* prevCTable,
+        short const* defaultNorm, U32 defaultNormLog,
+        ZSTD_defaultPolicy_e const isDefaultAllowed,
+        ZSTD_strategy const strategy)
+{
+    ZSTD_STATIC_ASSERT(ZSTD_defaultDisallowed == 0 && ZSTD_defaultAllowed != 0);
+    if (mostFrequent == nbSeq) {
+        *repeatMode = FSE_repeat_none;
+        if (isDefaultAllowed && nbSeq <= 2) {
+            /* Prefer set_basic over set_rle when there are 2 or less symbols,
+             * since RLE uses 1 byte, but set_basic uses 5-6 bits per symbol.
+             * If basic encoding isn't possible, always choose RLE.
+             */
+            DEBUGLOG(5, "Selected set_basic");
+            return set_basic;
+        }
+        DEBUGLOG(5, "Selected set_rle");
+        return set_rle;
+    }
+    if (strategy < ZSTD_lazy) {
+        if (isDefaultAllowed) {
+            size_t const staticFse_nbSeq_max = 1000;
+            size_t const mult = 10 - strategy;
+            size_t const baseLog = 3;
+            size_t const dynamicFse_nbSeq_min = (((size_t)1 << defaultNormLog) * mult) >> baseLog;  /* 28-36 for offset, 56-72 for lengths */
+            assert(defaultNormLog >= 5 && defaultNormLog <= 6);  /* xx_DEFAULTNORMLOG */
+            assert(mult <= 9 && mult >= 7);
+            if ( (*repeatMode == FSE_repeat_valid)
+              && (nbSeq < staticFse_nbSeq_max) ) {
+                DEBUGLOG(5, "Selected set_repeat");
+                return set_repeat;
+            }
+            if ( (nbSeq < dynamicFse_nbSeq_min)
+              || (mostFrequent < (nbSeq >> (defaultNormLog-1))) ) {
+                DEBUGLOG(5, "Selected set_basic");
+                /* The format allows default tables to be repeated, but it isn't useful.
+                 * When using simple heuristics to select encoding type, we don't want
+                 * to confuse these tables with dictionaries. When running more careful
+                 * analysis, we don't need to waste time checking both repeating tables
+                 * and default tables.
+                 */
+                *repeatMode = FSE_repeat_none;
+                return set_basic;
+            }
+        }
+    } else {
+        size_t const basicCost = isDefaultAllowed ? ZSTD_crossEntropyCost(defaultNorm, defaultNormLog, count, max) : ERROR(GENERIC);
+        size_t const repeatCost = *repeatMode != FSE_repeat_none ? ZSTD_fseBitCost(prevCTable, count, max) : ERROR(GENERIC);
+        size_t const NCountCost = ZSTD_NCountCost(count, max, nbSeq, FSELog);
+        size_t const compressedCost = (NCountCost << 3) + ZSTD_entropyCost(count, max, nbSeq);
+
+        if (isDefaultAllowed) {
+            assert(!ZSTD_isError(basicCost));
+            assert(!(*repeatMode == FSE_repeat_valid && ZSTD_isError(repeatCost)));
+        }
+        assert(!ZSTD_isError(NCountCost));
+        assert(compressedCost < ERROR(maxCode));
+        DEBUGLOG(5, "Estimated bit costs: basic=%u\trepeat=%u\tcompressed=%u",
+                    (unsigned)basicCost, (unsigned)repeatCost, (unsigned)compressedCost);
+        if (basicCost <= repeatCost && basicCost <= compressedCost) {
+            DEBUGLOG(5, "Selected set_basic");
+            assert(isDefaultAllowed);
+            *repeatMode = FSE_repeat_none;
+            return set_basic;
+        }
+        if (repeatCost <= compressedCost) {
+            DEBUGLOG(5, "Selected set_repeat");
+            assert(!ZSTD_isError(repeatCost));
+            return set_repeat;
+        }
+        assert(compressedCost < basicCost && compressedCost < repeatCost);
+    }
+    DEBUGLOG(5, "Selected set_compressed");
+    *repeatMode = FSE_repeat_check;
+    return set_compressed;
+}
+
+typedef struct {
+    S16 norm[MaxSeq + 1];
+    U32 wksp[FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(MaxSeq, MaxFSELog)];
+} ZSTD_BuildCTableWksp;
+
+size_t
+ZSTD_buildCTable(void* dst, size_t dstCapacity,
+                FSE_CTable* nextCTable, U32 FSELog, symbolEncodingType_e type,
+                unsigned* count, U32 max,
+                const BYTE* codeTable, size_t nbSeq,
+                const S16* defaultNorm, U32 defaultNormLog, U32 defaultMax,
+                const FSE_CTable* prevCTable, size_t prevCTableSize,
+                void* entropyWorkspace, size_t entropyWorkspaceSize)
+{
+    BYTE* op = (BYTE*)dst;
+    const BYTE* const oend = op + dstCapacity;
+    DEBUGLOG(6, "ZSTD_buildCTable (dstCapacity=%u)", (unsigned)dstCapacity);
+
+    switch (type) {
+    case set_rle:
+        FORWARD_IF_ERROR(FSE_buildCTable_rle(nextCTable, (BYTE)max), "");
+        RETURN_ERROR_IF(dstCapacity==0, dstSize_tooSmall, "not enough space");
+        *op = codeTable[0];
+        return 1;
+    case set_repeat:
+        ZSTD_memcpy(nextCTable, prevCTable, prevCTableSize);
+        return 0;
+    case set_basic:
+        FORWARD_IF_ERROR(FSE_buildCTable_wksp(nextCTable, defaultNorm, defaultMax, defaultNormLog, entropyWorkspace, entropyWorkspaceSize), "");  /* note : could be pre-calculated */
+        return 0;
+    case set_compressed: {
+        ZSTD_BuildCTableWksp* wksp = (ZSTD_BuildCTableWksp*)entropyWorkspace;
+        size_t nbSeq_1 = nbSeq;
+        const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max);
+        if (count[codeTable[nbSeq-1]] > 1) {
+            count[codeTable[nbSeq-1]]--;
+            nbSeq_1--;
+        }
+        assert(nbSeq_1 > 1);
+        assert(entropyWorkspaceSize >= sizeof(ZSTD_BuildCTableWksp));
+        (void)entropyWorkspaceSize;
+        FORWARD_IF_ERROR(FSE_normalizeCount(wksp->norm, tableLog, count, nbSeq_1, max, ZSTD_useLowProbCount(nbSeq_1)), "");
+        {   size_t const NCountSize = FSE_writeNCount(op, oend - op, wksp->norm, max, tableLog);   /* overflow protected */
+            FORWARD_IF_ERROR(NCountSize, "FSE_writeNCount failed");
+            FORWARD_IF_ERROR(FSE_buildCTable_wksp(nextCTable, wksp->norm, max, tableLog, wksp->wksp, sizeof(wksp->wksp)), "");
+            return NCountSize;
+        }
+    }
+    default: assert(0); RETURN_ERROR(GENERIC, "impossible to reach");
+    }
+}
+
+FORCE_INLINE_TEMPLATE size_t
+ZSTD_encodeSequences_body(
+            void* dst, size_t dstCapacity,
+            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
+            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
+            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
+            seqDef const* sequences, size_t nbSeq, int longOffsets)
+{
+    BIT_CStream_t blockStream;
+    FSE_CState_t  stateMatchLength;
+    FSE_CState_t  stateOffsetBits;
+    FSE_CState_t  stateLitLength;
+
+    RETURN_ERROR_IF(
+        ERR_isError(BIT_initCStream(&blockStream, dst, dstCapacity)),
+        dstSize_tooSmall, "not enough space remaining");
+    DEBUGLOG(6, "available space for bitstream : %i  (dstCapacity=%u)",
+                (int)(blockStream.endPtr - blockStream.startPtr),
+                (unsigned)dstCapacity);
+
+    /* first symbols */
+    FSE_initCState2(&stateMatchLength, CTable_MatchLength, mlCodeTable[nbSeq-1]);
+    FSE_initCState2(&stateOffsetBits,  CTable_OffsetBits,  ofCodeTable[nbSeq-1]);
+    FSE_initCState2(&stateLitLength,   CTable_LitLength,   llCodeTable[nbSeq-1]);
+    BIT_addBits(&blockStream, sequences[nbSeq-1].litLength, LL_bits[llCodeTable[nbSeq-1]]);
+    if (MEM_32bits()) BIT_flushBits(&blockStream);
+    BIT_addBits(&blockStream, sequences[nbSeq-1].matchLength, ML_bits[mlCodeTable[nbSeq-1]]);
+    if (MEM_32bits()) BIT_flushBits(&blockStream);
+    if (longOffsets) {
+        U32 const ofBits = ofCodeTable[nbSeq-1];
+        unsigned const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1);
+        if (extraBits) {
+            BIT_addBits(&blockStream, sequences[nbSeq-1].offset, extraBits);
+            BIT_flushBits(&blockStream);
+        }
+        BIT_addBits(&blockStream, sequences[nbSeq-1].offset >> extraBits,
+                    ofBits - extraBits);
+    } else {
+        BIT_addBits(&blockStream, sequences[nbSeq-1].offset, ofCodeTable[nbSeq-1]);
+    }
+    BIT_flushBits(&blockStream);
+
+    {   size_t n;
+        for (n=nbSeq-2 ; n<nbSeq ; n--) {      /* intentional underflow */
+            BYTE const llCode = llCodeTable[n];
+            BYTE const ofCode = ofCodeTable[n];
+            BYTE const mlCode = mlCodeTable[n];
+            U32  const llBits = LL_bits[llCode];
+            U32  const ofBits = ofCode;
+            U32  const mlBits = ML_bits[mlCode];
+            DEBUGLOG(6, "encoding: litlen:%2u - matchlen:%2u - offCode:%7u",
+                        (unsigned)sequences[n].litLength,
+                        (unsigned)sequences[n].matchLength + MINMATCH,
+                        (unsigned)sequences[n].offset);
+                                                                            /* 32b*/  /* 64b*/
+                                                                            /* (7)*/  /* (7)*/
+            FSE_encodeSymbol(&blockStream, &stateOffsetBits, ofCode);       /* 15 */  /* 15 */
+            FSE_encodeSymbol(&blockStream, &stateMatchLength, mlCode);      /* 24 */  /* 24 */
+            if (MEM_32bits()) BIT_flushBits(&blockStream);                  /* (7)*/
+            FSE_encodeSymbol(&blockStream, &stateLitLength, llCode);        /* 16 */  /* 33 */
+            if (MEM_32bits() || (ofBits+mlBits+llBits >= 64-7-(LLFSELog+MLFSELog+OffFSELog)))
+                BIT_flushBits(&blockStream);                                /* (7)*/
+            BIT_addBits(&blockStream, sequences[n].litLength, llBits);
+            if (MEM_32bits() && ((llBits+mlBits)>24)) BIT_flushBits(&blockStream);
+            BIT_addBits(&blockStream, sequences[n].matchLength, mlBits);
+            if (MEM_32bits() || (ofBits+mlBits+llBits > 56)) BIT_flushBits(&blockStream);
+            if (longOffsets) {
+                unsigned const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1);
+                if (extraBits) {
+                    BIT_addBits(&blockStream, sequences[n].offset, extraBits);
+                    BIT_flushBits(&blockStream);                            /* (7)*/
+                }
+                BIT_addBits(&blockStream, sequences[n].offset >> extraBits,
+                            ofBits - extraBits);                            /* 31 */
+            } else {
+                BIT_addBits(&blockStream, sequences[n].offset, ofBits);     /* 31 */
+            }
+            BIT_flushBits(&blockStream);                                    /* (7)*/
+            DEBUGLOG(7, "remaining space : %i", (int)(blockStream.endPtr - blockStream.ptr));
+    }   }
+
+    DEBUGLOG(6, "ZSTD_encodeSequences: flushing ML state with %u bits", stateMatchLength.stateLog);
+    FSE_flushCState(&blockStream, &stateMatchLength);
+    DEBUGLOG(6, "ZSTD_encodeSequences: flushing Off state with %u bits", stateOffsetBits.stateLog);
+    FSE_flushCState(&blockStream, &stateOffsetBits);
+    DEBUGLOG(6, "ZSTD_encodeSequences: flushing LL state with %u bits", stateLitLength.stateLog);
+    FSE_flushCState(&blockStream, &stateLitLength);
+
+    {   size_t const streamSize = BIT_closeCStream(&blockStream);
+        RETURN_ERROR_IF(streamSize==0, dstSize_tooSmall, "not enough space");
+        return streamSize;
+    }
+}
+
+static size_t
+ZSTD_encodeSequences_default(
+            void* dst, size_t dstCapacity,
+            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
+            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
+            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
+            seqDef const* sequences, size_t nbSeq, int longOffsets)
+{
+    return ZSTD_encodeSequences_body(dst, dstCapacity,
+                                    CTable_MatchLength, mlCodeTable,
+                                    CTable_OffsetBits, ofCodeTable,
+                                    CTable_LitLength, llCodeTable,
+                                    sequences, nbSeq, longOffsets);
+}
+
+
+#if DYNAMIC_BMI2
+
+static TARGET_ATTRIBUTE("bmi2") size_t
+ZSTD_encodeSequences_bmi2(
+            void* dst, size_t dstCapacity,
+            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
+            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
+            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
+            seqDef const* sequences, size_t nbSeq, int longOffsets)
+{
+    return ZSTD_encodeSequences_body(dst, dstCapacity,
+                                    CTable_MatchLength, mlCodeTable,
+                                    CTable_OffsetBits, ofCodeTable,
+                                    CTable_LitLength, llCodeTable,
+                                    sequences, nbSeq, longOffsets);
+}
+
+#endif
+
+size_t ZSTD_encodeSequences(
+            void* dst, size_t dstCapacity,
+            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
+            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
+            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
+            seqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2)
+{
+    DEBUGLOG(5, "ZSTD_encodeSequences: dstCapacity = %u", (unsigned)dstCapacity);
+#if DYNAMIC_BMI2
+    if (bmi2) {
+        return ZSTD_encodeSequences_bmi2(dst, dstCapacity,
+                                         CTable_MatchLength, mlCodeTable,
+                                         CTable_OffsetBits, ofCodeTable,
+                                         CTable_LitLength, llCodeTable,
+                                         sequences, nbSeq, longOffsets);
+    }
+#endif
+    (void)bmi2;
+    return ZSTD_encodeSequences_default(dst, dstCapacity,
+                                        CTable_MatchLength, mlCodeTable,
+                                        CTable_OffsetBits, ofCodeTable,
+                                        CTable_LitLength, llCodeTable,
+                                        sequences, nbSeq, longOffsets);
+}
diff --git a/lib/zstd/compress/zstd_compress_sequences.h b/lib/zstd/compress/zstd_compress_sequences.h
new file mode 100644
index 0000000..7991364
--- /dev/null
+++ b/lib/zstd/compress/zstd_compress_sequences.h
@@ -0,0 +1,54 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_COMPRESS_SEQUENCES_H
+#define ZSTD_COMPRESS_SEQUENCES_H
+
+#include "../common/fse.h" /* FSE_repeat, FSE_CTable */
+#include "../common/zstd_internal.h" /* symbolEncodingType_e, ZSTD_strategy */
+
+typedef enum {
+    ZSTD_defaultDisallowed = 0,
+    ZSTD_defaultAllowed = 1
+} ZSTD_defaultPolicy_e;
+
+symbolEncodingType_e
+ZSTD_selectEncodingType(
+        FSE_repeat* repeatMode, unsigned const* count, unsigned const max,
+        size_t const mostFrequent, size_t nbSeq, unsigned const FSELog,
+        FSE_CTable const* prevCTable,
+        short const* defaultNorm, U32 defaultNormLog,
+        ZSTD_defaultPolicy_e const isDefaultAllowed,
+        ZSTD_strategy const strategy);
+
+size_t
+ZSTD_buildCTable(void* dst, size_t dstCapacity,
+                FSE_CTable* nextCTable, U32 FSELog, symbolEncodingType_e type,
+                unsigned* count, U32 max,
+                const BYTE* codeTable, size_t nbSeq,
+                const S16* defaultNorm, U32 defaultNormLog, U32 defaultMax,
+                const FSE_CTable* prevCTable, size_t prevCTableSize,
+                void* entropyWorkspace, size_t entropyWorkspaceSize);
+
+size_t ZSTD_encodeSequences(
+            void* dst, size_t dstCapacity,
+            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
+            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
+            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
+            seqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2);
+
+size_t ZSTD_fseBitCost(
+    FSE_CTable const* ctable,
+    unsigned const* count,
+    unsigned const max);
+
+size_t ZSTD_crossEntropyCost(short const* norm, unsigned accuracyLog,
+                             unsigned const* count, unsigned const max);
+#endif /* ZSTD_COMPRESS_SEQUENCES_H */
diff --git a/lib/zstd/compress/zstd_compress_superblock.c b/lib/zstd/compress/zstd_compress_superblock.c
new file mode 100644
index 0000000..ee03e0a
--- /dev/null
+++ b/lib/zstd/compress/zstd_compress_superblock.c
@@ -0,0 +1,850 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+ /*-*************************************
+ *  Dependencies
+ ***************************************/
+#include "zstd_compress_superblock.h"
+
+#include "../common/zstd_internal.h"  /* ZSTD_getSequenceLength */
+#include "hist.h"                     /* HIST_countFast_wksp */
+#include "zstd_compress_internal.h"
+#include "zstd_compress_sequences.h"
+#include "zstd_compress_literals.h"
+
+/*-*************************************
+*  Superblock entropy buffer structs
+***************************************/
+/* ZSTD_hufCTablesMetadata_t :
+ *  Stores Literals Block Type for a super-block in hType, and
+ *  huffman tree description in hufDesBuffer.
+ *  hufDesSize refers to the size of huffman tree description in bytes.
+ *  This metadata is populated in ZSTD_buildSuperBlockEntropy_literal() */
+typedef struct {
+    symbolEncodingType_e hType;
+    BYTE hufDesBuffer[ZSTD_MAX_HUF_HEADER_SIZE];
+    size_t hufDesSize;
+} ZSTD_hufCTablesMetadata_t;
+
+/* ZSTD_fseCTablesMetadata_t :
+ *  Stores symbol compression modes for a super-block in {ll, ol, ml}Type, and
+ *  fse tables in fseTablesBuffer.
+ *  fseTablesSize refers to the size of fse tables in bytes.
+ *  This metadata is populated in ZSTD_buildSuperBlockEntropy_sequences() */
+typedef struct {
+    symbolEncodingType_e llType;
+    symbolEncodingType_e ofType;
+    symbolEncodingType_e mlType;
+    BYTE fseTablesBuffer[ZSTD_MAX_FSE_HEADERS_SIZE];
+    size_t fseTablesSize;
+    size_t lastCountSize; /* This is to account for bug in 1.3.4. More detail in ZSTD_compressSubBlock_sequences() */
+} ZSTD_fseCTablesMetadata_t;
+
+typedef struct {
+    ZSTD_hufCTablesMetadata_t hufMetadata;
+    ZSTD_fseCTablesMetadata_t fseMetadata;
+} ZSTD_entropyCTablesMetadata_t;
+
+
+/* ZSTD_buildSuperBlockEntropy_literal() :
+ *  Builds entropy for the super-block literals.
+ *  Stores literals block type (raw, rle, compressed, repeat) and
+ *  huffman description table to hufMetadata.
+ *  @return : size of huffman description table or error code */
+static size_t ZSTD_buildSuperBlockEntropy_literal(void* const src, size_t srcSize,
+                                            const ZSTD_hufCTables_t* prevHuf,
+                                                  ZSTD_hufCTables_t* nextHuf,
+                                                  ZSTD_hufCTablesMetadata_t* hufMetadata,
+                                                  const int disableLiteralsCompression,
+                                                  void* workspace, size_t wkspSize)
+{
+    BYTE* const wkspStart = (BYTE*)workspace;
+    BYTE* const wkspEnd = wkspStart + wkspSize;
+    BYTE* const countWkspStart = wkspStart;
+    unsigned* const countWksp = (unsigned*)workspace;
+    const size_t countWkspSize = (HUF_SYMBOLVALUE_MAX + 1) * sizeof(unsigned);
+    BYTE* const nodeWksp = countWkspStart + countWkspSize;
+    const size_t nodeWkspSize = wkspEnd-nodeWksp;
+    unsigned maxSymbolValue = 255;
+    unsigned huffLog = HUF_TABLELOG_DEFAULT;
+    HUF_repeat repeat = prevHuf->repeatMode;
+
+    DEBUGLOG(5, "ZSTD_buildSuperBlockEntropy_literal (srcSize=%zu)", srcSize);
+
+    /* Prepare nextEntropy assuming reusing the existing table */
+    ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
+
+    if (disableLiteralsCompression) {
+        DEBUGLOG(5, "set_basic - disabled");
+        hufMetadata->hType = set_basic;
+        return 0;
+    }
+
+    /* small ? don't even attempt compression (speed opt) */
+#   define COMPRESS_LITERALS_SIZE_MIN 63
+    {   size_t const minLitSize = (prevHuf->repeatMode == HUF_repeat_valid) ? 6 : COMPRESS_LITERALS_SIZE_MIN;
+        if (srcSize <= minLitSize) {
+            DEBUGLOG(5, "set_basic - too small");
+            hufMetadata->hType = set_basic;
+            return 0;
+        }
+    }
+
+    /* Scan input and build symbol stats */
+    {   size_t const largest = HIST_count_wksp (countWksp, &maxSymbolValue, (const BYTE*)src, srcSize, workspace, wkspSize);
+        FORWARD_IF_ERROR(largest, "HIST_count_wksp failed");
+        if (largest == srcSize) {
+            DEBUGLOG(5, "set_rle");
+            hufMetadata->hType = set_rle;
+            return 0;
+        }
+        if (largest <= (srcSize >> 7)+4) {
+            DEBUGLOG(5, "set_basic - no gain");
+            hufMetadata->hType = set_basic;
+            return 0;
+        }
+    }
+
+    /* Validate the previous Huffman table */
+    if (repeat == HUF_repeat_check && !HUF_validateCTable((HUF_CElt const*)prevHuf->CTable, countWksp, maxSymbolValue)) {
+        repeat = HUF_repeat_none;
+    }
+
+    /* Build Huffman Tree */
+    ZSTD_memset(nextHuf->CTable, 0, sizeof(nextHuf->CTable));
+    huffLog = HUF_optimalTableLog(huffLog, srcSize, maxSymbolValue);
+    {   size_t const maxBits = HUF_buildCTable_wksp((HUF_CElt*)nextHuf->CTable, countWksp,
+                                                    maxSymbolValue, huffLog,
+                                                    nodeWksp, nodeWkspSize);
+        FORWARD_IF_ERROR(maxBits, "HUF_buildCTable_wksp");
+        huffLog = (U32)maxBits;
+        {   /* Build and write the CTable */
+            size_t const newCSize = HUF_estimateCompressedSize(
+                    (HUF_CElt*)nextHuf->CTable, countWksp, maxSymbolValue);
+            size_t const hSize = HUF_writeCTable_wksp(
+                    hufMetadata->hufDesBuffer, sizeof(hufMetadata->hufDesBuffer),
+                    (HUF_CElt*)nextHuf->CTable, maxSymbolValue, huffLog,
+                    nodeWksp, nodeWkspSize);
+            /* Check against repeating the previous CTable */
+            if (repeat != HUF_repeat_none) {
+                size_t const oldCSize = HUF_estimateCompressedSize(
+                        (HUF_CElt const*)prevHuf->CTable, countWksp, maxSymbolValue);
+                if (oldCSize < srcSize && (oldCSize <= hSize + newCSize || hSize + 12 >= srcSize)) {
+                    DEBUGLOG(5, "set_repeat - smaller");
+                    ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
+                    hufMetadata->hType = set_repeat;
+                    return 0;
+                }
+            }
+            if (newCSize + hSize >= srcSize) {
+                DEBUGLOG(5, "set_basic - no gains");
+                ZSTD_memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
+                hufMetadata->hType = set_basic;
+                return 0;
+            }
+            DEBUGLOG(5, "set_compressed (hSize=%u)", (U32)hSize);
+            hufMetadata->hType = set_compressed;
+            nextHuf->repeatMode = HUF_repeat_check;
+            return hSize;
+        }
+    }
+}
+
+/* ZSTD_buildSuperBlockEntropy_sequences() :
+ *  Builds entropy for the super-block sequences.
+ *  Stores symbol compression modes and fse table to fseMetadata.
+ *  @return : size of fse tables or error code */
+static size_t ZSTD_buildSuperBlockEntropy_sequences(seqStore_t* seqStorePtr,
+                                              const ZSTD_fseCTables_t* prevEntropy,
+                                                    ZSTD_fseCTables_t* nextEntropy,
+                                              const ZSTD_CCtx_params* cctxParams,
+                                                    ZSTD_fseCTablesMetadata_t* fseMetadata,
+                                                    void* workspace, size_t wkspSize)
+{
+    BYTE* const wkspStart = (BYTE*)workspace;
+    BYTE* const wkspEnd = wkspStart + wkspSize;
+    BYTE* const countWkspStart = wkspStart;
+    unsigned* const countWksp = (unsigned*)workspace;
+    const size_t countWkspSize = (MaxSeq + 1) * sizeof(unsigned);
+    BYTE* const cTableWksp = countWkspStart + countWkspSize;
+    const size_t cTableWkspSize = wkspEnd-cTableWksp;
+    ZSTD_strategy const strategy = cctxParams->cParams.strategy;
+    FSE_CTable* CTable_LitLength = nextEntropy->litlengthCTable;
+    FSE_CTable* CTable_OffsetBits = nextEntropy->offcodeCTable;
+    FSE_CTable* CTable_MatchLength = nextEntropy->matchlengthCTable;
+    const BYTE* const ofCodeTable = seqStorePtr->ofCode;
+    const BYTE* const llCodeTable = seqStorePtr->llCode;
+    const BYTE* const mlCodeTable = seqStorePtr->mlCode;
+    size_t const nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart;
+    BYTE* const ostart = fseMetadata->fseTablesBuffer;
+    BYTE* const oend = ostart + sizeof(fseMetadata->fseTablesBuffer);
+    BYTE* op = ostart;
+
+    assert(cTableWkspSize >= (1 << MaxFSELog) * sizeof(FSE_FUNCTION_TYPE));
+    DEBUGLOG(5, "ZSTD_buildSuperBlockEntropy_sequences (nbSeq=%zu)", nbSeq);
+    ZSTD_memset(workspace, 0, wkspSize);
+
+    fseMetadata->lastCountSize = 0;
+    /* convert length/distances into codes */
+    ZSTD_seqToCodes(seqStorePtr);
+    /* build CTable for Literal Lengths */
+    {   U32 LLtype;
+        unsigned max = MaxLL;
+        size_t const mostFrequent = HIST_countFast_wksp(countWksp, &max, llCodeTable, nbSeq, workspace, wkspSize);  /* can't fail */
+        DEBUGLOG(5, "Building LL table");
+        nextEntropy->litlength_repeatMode = prevEntropy->litlength_repeatMode;
+        LLtype = ZSTD_selectEncodingType(&nextEntropy->litlength_repeatMode,
+                                        countWksp, max, mostFrequent, nbSeq,
+                                        LLFSELog, prevEntropy->litlengthCTable,
+                                        LL_defaultNorm, LL_defaultNormLog,
+                                        ZSTD_defaultAllowed, strategy);
+        assert(set_basic < set_compressed && set_rle < set_compressed);
+        assert(!(LLtype < set_compressed && nextEntropy->litlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */
+        {   size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_LitLength, LLFSELog, (symbolEncodingType_e)LLtype,
+                                                    countWksp, max, llCodeTable, nbSeq, LL_defaultNorm, LL_defaultNormLog, MaxLL,
+                                                    prevEntropy->litlengthCTable, sizeof(prevEntropy->litlengthCTable),
+                                                    cTableWksp, cTableWkspSize);
+            FORWARD_IF_ERROR(countSize, "ZSTD_buildCTable for LitLens failed");
+            if (LLtype == set_compressed)
+                fseMetadata->lastCountSize = countSize;
+            op += countSize;
+            fseMetadata->llType = (symbolEncodingType_e) LLtype;
+    }   }
+    /* build CTable for Offsets */
+    {   U32 Offtype;
+        unsigned max = MaxOff;
+        size_t const mostFrequent = HIST_countFast_wksp(countWksp, &max, ofCodeTable, nbSeq, workspace, wkspSize);  /* can't fail */
+        /* We can only use the basic table if max <= DefaultMaxOff, otherwise the offsets are too large */
+        ZSTD_defaultPolicy_e const defaultPolicy = (max <= DefaultMaxOff) ? ZSTD_defaultAllowed : ZSTD_defaultDisallowed;
+        DEBUGLOG(5, "Building OF table");
+        nextEntropy->offcode_repeatMode = prevEntropy->offcode_repeatMode;
+        Offtype = ZSTD_selectEncodingType(&nextEntropy->offcode_repeatMode,
+                                        countWksp, max, mostFrequent, nbSeq,
+                                        OffFSELog, prevEntropy->offcodeCTable,
+                                        OF_defaultNorm, OF_defaultNormLog,
+                                        defaultPolicy, strategy);
+        assert(!(Offtype < set_compressed && nextEntropy->offcode_repeatMode != FSE_repeat_none)); /* We don't copy tables */
+        {   size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_OffsetBits, OffFSELog, (symbolEncodingType_e)Offtype,
+                                                    countWksp, max, ofCodeTable, nbSeq, OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff,
+                                                    prevEntropy->offcodeCTable, sizeof(prevEntropy->offcodeCTable),
+                                                    cTableWksp, cTableWkspSize);
+            FORWARD_IF_ERROR(countSize, "ZSTD_buildCTable for Offsets failed");
+            if (Offtype == set_compressed)
+                fseMetadata->lastCountSize = countSize;
+            op += countSize;
+            fseMetadata->ofType = (symbolEncodingType_e) Offtype;
+    }   }
+    /* build CTable for MatchLengths */
+    {   U32 MLtype;
+        unsigned max = MaxML;
+        size_t const mostFrequent = HIST_countFast_wksp(countWksp, &max, mlCodeTable, nbSeq, workspace, wkspSize);   /* can't fail */
+        DEBUGLOG(5, "Building ML table (remaining space : %i)", (int)(oend-op));
+        nextEntropy->matchlength_repeatMode = prevEntropy->matchlength_repeatMode;
+        MLtype = ZSTD_selectEncodingType(&nextEntropy->matchlength_repeatMode,
+                                        countWksp, max, mostFrequent, nbSeq,
+                                        MLFSELog, prevEntropy->matchlengthCTable,
+                                        ML_defaultNorm, ML_defaultNormLog,
+                                        ZSTD_defaultAllowed, strategy);
+        assert(!(MLtype < set_compressed && nextEntropy->matchlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */
+        {   size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_MatchLength, MLFSELog, (symbolEncodingType_e)MLtype,
+                                                    countWksp, max, mlCodeTable, nbSeq, ML_defaultNorm, ML_defaultNormLog, MaxML,
+                                                    prevEntropy->matchlengthCTable, sizeof(prevEntropy->matchlengthCTable),
+                                                    cTableWksp, cTableWkspSize);
+            FORWARD_IF_ERROR(countSize, "ZSTD_buildCTable for MatchLengths failed");
+            if (MLtype == set_compressed)
+                fseMetadata->lastCountSize = countSize;
+            op += countSize;
+            fseMetadata->mlType = (symbolEncodingType_e) MLtype;
+    }   }
+    assert((size_t) (op-ostart) <= sizeof(fseMetadata->fseTablesBuffer));
+    return op-ostart;
+}
+
+
+/* ZSTD_buildSuperBlockEntropy() :
+ *  Builds entropy for the super-block.
+ *  @return : 0 on success or error code */
+static size_t
+ZSTD_buildSuperBlockEntropy(seqStore_t* seqStorePtr,
+                      const ZSTD_entropyCTables_t* prevEntropy,
+                            ZSTD_entropyCTables_t* nextEntropy,
+                      const ZSTD_CCtx_params* cctxParams,
+                            ZSTD_entropyCTablesMetadata_t* entropyMetadata,
+                            void* workspace, size_t wkspSize)
+{
+    size_t const litSize = seqStorePtr->lit - seqStorePtr->litStart;
+    DEBUGLOG(5, "ZSTD_buildSuperBlockEntropy");
+    entropyMetadata->hufMetadata.hufDesSize =
+        ZSTD_buildSuperBlockEntropy_literal(seqStorePtr->litStart, litSize,
+                                            &prevEntropy->huf, &nextEntropy->huf,
+                                            &entropyMetadata->hufMetadata,
+                                            ZSTD_disableLiteralsCompression(cctxParams),
+                                            workspace, wkspSize);
+    FORWARD_IF_ERROR(entropyMetadata->hufMetadata.hufDesSize, "ZSTD_buildSuperBlockEntropy_literal failed");
+    entropyMetadata->fseMetadata.fseTablesSize =
+        ZSTD_buildSuperBlockEntropy_sequences(seqStorePtr,
+                                              &prevEntropy->fse, &nextEntropy->fse,
+                                              cctxParams,
+                                              &entropyMetadata->fseMetadata,
+                                              workspace, wkspSize);
+    FORWARD_IF_ERROR(entropyMetadata->fseMetadata.fseTablesSize, "ZSTD_buildSuperBlockEntropy_sequences failed");
+    return 0;
+}
+
+/* ZSTD_compressSubBlock_literal() :
+ *  Compresses literals section for a sub-block.
+ *  When we have to write the Huffman table we will sometimes choose a header
+ *  size larger than necessary. This is because we have to pick the header size
+ *  before we know the table size + compressed size, so we have a bound on the
+ *  table size. If we guessed incorrectly, we fall back to uncompressed literals.
+ *
+ *  We write the header when writeEntropy=1 and set entropyWritten=1 when we succeeded
+ *  in writing the header, otherwise it is set to 0.
+ *
+ *  hufMetadata->hType has literals block type info.
+ *      If it is set_basic, all sub-blocks literals section will be Raw_Literals_Block.
+ *      If it is set_rle, all sub-blocks literals section will be RLE_Literals_Block.
+ *      If it is set_compressed, first sub-block's literals section will be Compressed_Literals_Block
+ *      If it is set_compressed, first sub-block's literals section will be Treeless_Literals_Block
+ *      and the following sub-blocks' literals sections will be Treeless_Literals_Block.
+ *  @return : compressed size of literals section of a sub-block
+ *            Or 0 if it unable to compress.
+ *            Or error code */
+static size_t ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable,
+                                    const ZSTD_hufCTablesMetadata_t* hufMetadata,
+                                    const BYTE* literals, size_t litSize,
+                                    void* dst, size_t dstSize,
+                                    const int bmi2, int writeEntropy, int* entropyWritten)
+{
+    size_t const header = writeEntropy ? 200 : 0;
+    size_t const lhSize = 3 + (litSize >= (1 KB - header)) + (litSize >= (16 KB - header));
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* const oend = ostart + dstSize;
+    BYTE* op = ostart + lhSize;
+    U32 const singleStream = lhSize == 3;
+    symbolEncodingType_e hType = writeEntropy ? hufMetadata->hType : set_repeat;
+    size_t cLitSize = 0;
+
+    (void)bmi2; /* TODO bmi2... */
+
+    DEBUGLOG(5, "ZSTD_compressSubBlock_literal (litSize=%zu, lhSize=%zu, writeEntropy=%d)", litSize, lhSize, writeEntropy);
+
+    *entropyWritten = 0;
+    if (litSize == 0 || hufMetadata->hType == set_basic) {
+      DEBUGLOG(5, "ZSTD_compressSubBlock_literal using raw literal");
+      return ZSTD_noCompressLiterals(dst, dstSize, literals, litSize);
+    } else if (hufMetadata->hType == set_rle) {
+      DEBUGLOG(5, "ZSTD_compressSubBlock_literal using rle literal");
+      return ZSTD_compressRleLiteralsBlock(dst, dstSize, literals, litSize);
+    }
+
+    assert(litSize > 0);
+    assert(hufMetadata->hType == set_compressed || hufMetadata->hType == set_repeat);
+
+    if (writeEntropy && hufMetadata->hType == set_compressed) {
+        ZSTD_memcpy(op, hufMetadata->hufDesBuffer, hufMetadata->hufDesSize);
+        op += hufMetadata->hufDesSize;
+        cLitSize += hufMetadata->hufDesSize;
+        DEBUGLOG(5, "ZSTD_compressSubBlock_literal (hSize=%zu)", hufMetadata->hufDesSize);
+    }
+
+    /* TODO bmi2 */
+    {   const size_t cSize = singleStream ? HUF_compress1X_usingCTable(op, oend-op, literals, litSize, hufTable)
+                                          : HUF_compress4X_usingCTable(op, oend-op, literals, litSize, hufTable);
+        op += cSize;
+        cLitSize += cSize;
+        if (cSize == 0 || ERR_isError(cSize)) {
+            DEBUGLOG(5, "Failed to write entropy tables %s", ZSTD_getErrorName(cSize));
+            return 0;
+        }
+        /* If we expand and we aren't writing a header then emit uncompressed */
+        if (!writeEntropy && cLitSize >= litSize) {
+            DEBUGLOG(5, "ZSTD_compressSubBlock_literal using raw literal because uncompressible");
+            return ZSTD_noCompressLiterals(dst, dstSize, literals, litSize);
+        }
+        /* If we are writing headers then allow expansion that doesn't change our header size. */
+        if (lhSize < (size_t)(3 + (cLitSize >= 1 KB) + (cLitSize >= 16 KB))) {
+            assert(cLitSize > litSize);
+            DEBUGLOG(5, "Literals expanded beyond allowed header size");
+            return ZSTD_noCompressLiterals(dst, dstSize, literals, litSize);
+        }
+        DEBUGLOG(5, "ZSTD_compressSubBlock_literal (cSize=%zu)", cSize);
+    }
+
+    /* Build header */
+    switch(lhSize)
+    {
+    case 3: /* 2 - 2 - 10 - 10 */
+        {   U32 const lhc = hType + ((!singleStream) << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<14);
+            MEM_writeLE24(ostart, lhc);
+            break;
+        }
+    case 4: /* 2 - 2 - 14 - 14 */
+        {   U32 const lhc = hType + (2 << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<18);
+            MEM_writeLE32(ostart, lhc);
+            break;
+        }
+    case 5: /* 2 - 2 - 18 - 18 */
+        {   U32 const lhc = hType + (3 << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<22);
+            MEM_writeLE32(ostart, lhc);
+            ostart[4] = (BYTE)(cLitSize >> 10);
+            break;
+        }
+    default:  /* not possible : lhSize is {3,4,5} */
+        assert(0);
+    }
+    *entropyWritten = 1;
+    DEBUGLOG(5, "Compressed literals: %u -> %u", (U32)litSize, (U32)(op-ostart));
+    return op-ostart;
+}
+
+static size_t ZSTD_seqDecompressedSize(seqStore_t const* seqStore, const seqDef* sequences, size_t nbSeq, size_t litSize, int lastSequence) {
+    const seqDef* const sstart = sequences;
+    const seqDef* const send = sequences + nbSeq;
+    const seqDef* sp = sstart;
+    size_t matchLengthSum = 0;
+    size_t litLengthSum = 0;
+    while (send-sp > 0) {
+        ZSTD_sequenceLength const seqLen = ZSTD_getSequenceLength(seqStore, sp);
+        litLengthSum += seqLen.litLength;
+        matchLengthSum += seqLen.matchLength;
+        sp++;
+    }
+    assert(litLengthSum <= litSize);
+    if (!lastSequence) {
+        assert(litLengthSum == litSize);
+    }
+    return matchLengthSum + litSize;
+}
+
+/* ZSTD_compressSubBlock_sequences() :
+ *  Compresses sequences section for a sub-block.
+ *  fseMetadata->llType, fseMetadata->ofType, and fseMetadata->mlType have
+ *  symbol compression modes for the super-block.
+ *  The first successfully compressed block will have these in its header.
+ *  We set entropyWritten=1 when we succeed in compressing the sequences.
+ *  The following sub-blocks will always have repeat mode.
+ *  @return : compressed size of sequences section of a sub-block
+ *            Or 0 if it is unable to compress
+ *            Or error code. */
+static size_t ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables,
+                                              const ZSTD_fseCTablesMetadata_t* fseMetadata,
+                                              const seqDef* sequences, size_t nbSeq,
+                                              const BYTE* llCode, const BYTE* mlCode, const BYTE* ofCode,
+                                              const ZSTD_CCtx_params* cctxParams,
+                                              void* dst, size_t dstCapacity,
+                                              const int bmi2, int writeEntropy, int* entropyWritten)
+{
+    const int longOffsets = cctxParams->cParams.windowLog > STREAM_ACCUMULATOR_MIN;
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* const oend = ostart + dstCapacity;
+    BYTE* op = ostart;
+    BYTE* seqHead;
+
+    DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (nbSeq=%zu, writeEntropy=%d, longOffsets=%d)", nbSeq, writeEntropy, longOffsets);
+
+    *entropyWritten = 0;
+    /* Sequences Header */
+    RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/,
+                    dstSize_tooSmall, "");
+    if (nbSeq < 0x7F)
+        *op++ = (BYTE)nbSeq;
+    else if (nbSeq < LONGNBSEQ)
+        op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2;
+    else
+        op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3;
+    if (nbSeq==0) {
+        return op - ostart;
+    }
+
+    /* seqHead : flags for FSE encoding type */
+    seqHead = op++;
+
+    DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (seqHeadSize=%u)", (unsigned)(op-ostart));
+
+    if (writeEntropy) {
+        const U32 LLtype = fseMetadata->llType;
+        const U32 Offtype = fseMetadata->ofType;
+        const U32 MLtype = fseMetadata->mlType;
+        DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (fseTablesSize=%zu)", fseMetadata->fseTablesSize);
+        *seqHead = (BYTE)((LLtype<<6) + (Offtype<<4) + (MLtype<<2));
+        ZSTD_memcpy(op, fseMetadata->fseTablesBuffer, fseMetadata->fseTablesSize);
+        op += fseMetadata->fseTablesSize;
+    } else {
+        const U32 repeat = set_repeat;
+        *seqHead = (BYTE)((repeat<<6) + (repeat<<4) + (repeat<<2));
+    }
+
+    {   size_t const bitstreamSize = ZSTD_encodeSequences(
+                                        op, oend - op,
+                                        fseTables->matchlengthCTable, mlCode,
+                                        fseTables->offcodeCTable, ofCode,
+                                        fseTables->litlengthCTable, llCode,
+                                        sequences, nbSeq,
+                                        longOffsets, bmi2);
+        FORWARD_IF_ERROR(bitstreamSize, "ZSTD_encodeSequences failed");
+        op += bitstreamSize;
+        /* zstd versions <= 1.3.4 mistakenly report corruption when
+         * FSE_readNCount() receives a buffer < 4 bytes.
+         * Fixed by https://github.com/facebook/zstd/pull/1146.
+         * This can happen when the last set_compressed table present is 2
+         * bytes and the bitstream is only one byte.
+         * In this exceedingly rare case, we will simply emit an uncompressed
+         * block, since it isn't worth optimizing.
+         */
+#ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+        if (writeEntropy && fseMetadata->lastCountSize && fseMetadata->lastCountSize + bitstreamSize < 4) {
+            /* NCountSize >= 2 && bitstreamSize > 0 ==> lastCountSize == 3 */
+            assert(fseMetadata->lastCountSize + bitstreamSize == 3);
+            DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.3.4 by "
+                        "emitting an uncompressed block.");
+            return 0;
+        }
+#endif
+        DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (bitstreamSize=%zu)", bitstreamSize);
+    }
+
+    /* zstd versions <= 1.4.0 mistakenly report error when
+     * sequences section body size is less than 3 bytes.
+     * Fixed by https://github.com/facebook/zstd/pull/1664.
+     * This can happen when the previous sequences section block is compressed
+     * with rle mode and the current block's sequences section is compressed
+     * with repeat mode where sequences section body size can be 1 byte.
+     */
+#ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+    if (op-seqHead < 4) {
+        DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.4.0 by emitting "
+                    "an uncompressed block when sequences are < 4 bytes");
+        return 0;
+    }
+#endif
+
+    *entropyWritten = 1;
+    return op - ostart;
+}
+
+/* ZSTD_compressSubBlock() :
+ *  Compresses a single sub-block.
+ *  @return : compressed size of the sub-block
+ *            Or 0 if it failed to compress. */
+static size_t ZSTD_compressSubBlock(const ZSTD_entropyCTables_t* entropy,
+                                    const ZSTD_entropyCTablesMetadata_t* entropyMetadata,
+                                    const seqDef* sequences, size_t nbSeq,
+                                    const BYTE* literals, size_t litSize,
+                                    const BYTE* llCode, const BYTE* mlCode, const BYTE* ofCode,
+                                    const ZSTD_CCtx_params* cctxParams,
+                                    void* dst, size_t dstCapacity,
+                                    const int bmi2,
+                                    int writeLitEntropy, int writeSeqEntropy,
+                                    int* litEntropyWritten, int* seqEntropyWritten,
+                                    U32 lastBlock)
+{
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* const oend = ostart + dstCapacity;
+    BYTE* op = ostart + ZSTD_blockHeaderSize;
+    DEBUGLOG(5, "ZSTD_compressSubBlock (litSize=%zu, nbSeq=%zu, writeLitEntropy=%d, writeSeqEntropy=%d, lastBlock=%d)",
+                litSize, nbSeq, writeLitEntropy, writeSeqEntropy, lastBlock);
+    {   size_t cLitSize = ZSTD_compressSubBlock_literal((const HUF_CElt*)entropy->huf.CTable,
+                                                        &entropyMetadata->hufMetadata, literals, litSize,
+                                                        op, oend-op, bmi2, writeLitEntropy, litEntropyWritten);
+        FORWARD_IF_ERROR(cLitSize, "ZSTD_compressSubBlock_literal failed");
+        if (cLitSize == 0) return 0;
+        op += cLitSize;
+    }
+    {   size_t cSeqSize = ZSTD_compressSubBlock_sequences(&entropy->fse,
+                                                  &entropyMetadata->fseMetadata,
+                                                  sequences, nbSeq,
+                                                  llCode, mlCode, ofCode,
+                                                  cctxParams,
+                                                  op, oend-op,
+                                                  bmi2, writeSeqEntropy, seqEntropyWritten);
+        FORWARD_IF_ERROR(cSeqSize, "ZSTD_compressSubBlock_sequences failed");
+        if (cSeqSize == 0) return 0;
+        op += cSeqSize;
+    }
+    /* Write block header */
+    {   size_t cSize = (op-ostart)-ZSTD_blockHeaderSize;
+        U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3);
+        MEM_writeLE24(ostart, cBlockHeader24);
+    }
+    return op-ostart;
+}
+
+static size_t ZSTD_estimateSubBlockSize_literal(const BYTE* literals, size_t litSize,
+                                                const ZSTD_hufCTables_t* huf,
+                                                const ZSTD_hufCTablesMetadata_t* hufMetadata,
+                                                void* workspace, size_t wkspSize,
+                                                int writeEntropy)
+{
+    unsigned* const countWksp = (unsigned*)workspace;
+    unsigned maxSymbolValue = 255;
+    size_t literalSectionHeaderSize = 3; /* Use hard coded size of 3 bytes */
+
+    if (hufMetadata->hType == set_basic) return litSize;
+    else if (hufMetadata->hType == set_rle) return 1;
+    else if (hufMetadata->hType == set_compressed || hufMetadata->hType == set_repeat) {
+        size_t const largest = HIST_count_wksp (countWksp, &maxSymbolValue, (const BYTE*)literals, litSize, workspace, wkspSize);
+        if (ZSTD_isError(largest)) return litSize;
+        {   size_t cLitSizeEstimate = HUF_estimateCompressedSize((const HUF_CElt*)huf->CTable, countWksp, maxSymbolValue);
+            if (writeEntropy) cLitSizeEstimate += hufMetadata->hufDesSize;
+            return cLitSizeEstimate + literalSectionHeaderSize;
+    }   }
+    assert(0); /* impossible */
+    return 0;
+}
+
+static size_t ZSTD_estimateSubBlockSize_symbolType(symbolEncodingType_e type,
+                        const BYTE* codeTable, unsigned maxCode,
+                        size_t nbSeq, const FSE_CTable* fseCTable,
+                        const U32* additionalBits,
+                        short const* defaultNorm, U32 defaultNormLog, U32 defaultMax,
+                        void* workspace, size_t wkspSize)
+{
+    unsigned* const countWksp = (unsigned*)workspace;
+    const BYTE* ctp = codeTable;
+    const BYTE* const ctStart = ctp;
+    const BYTE* const ctEnd = ctStart + nbSeq;
+    size_t cSymbolTypeSizeEstimateInBits = 0;
+    unsigned max = maxCode;
+
+    HIST_countFast_wksp(countWksp, &max, codeTable, nbSeq, workspace, wkspSize);  /* can't fail */
+    if (type == set_basic) {
+        /* We selected this encoding type, so it must be valid. */
+        assert(max <= defaultMax);
+        cSymbolTypeSizeEstimateInBits = max <= defaultMax
+                ? ZSTD_crossEntropyCost(defaultNorm, defaultNormLog, countWksp, max)
+                : ERROR(GENERIC);
+    } else if (type == set_rle) {
+        cSymbolTypeSizeEstimateInBits = 0;
+    } else if (type == set_compressed || type == set_repeat) {
+        cSymbolTypeSizeEstimateInBits = ZSTD_fseBitCost(fseCTable, countWksp, max);
+    }
+    if (ZSTD_isError(cSymbolTypeSizeEstimateInBits)) return nbSeq * 10;
+    while (ctp < ctEnd) {
+        if (additionalBits) cSymbolTypeSizeEstimateInBits += additionalBits[*ctp];
+        else cSymbolTypeSizeEstimateInBits += *ctp; /* for offset, offset code is also the number of additional bits */
+        ctp++;
+    }
+    return cSymbolTypeSizeEstimateInBits / 8;
+}
+
+static size_t ZSTD_estimateSubBlockSize_sequences(const BYTE* ofCodeTable,
+                                                  const BYTE* llCodeTable,
+                                                  const BYTE* mlCodeTable,
+                                                  size_t nbSeq,
+                                                  const ZSTD_fseCTables_t* fseTables,
+                                                  const ZSTD_fseCTablesMetadata_t* fseMetadata,
+                                                  void* workspace, size_t wkspSize,
+                                                  int writeEntropy)
+{
+    size_t sequencesSectionHeaderSize = 3; /* Use hard coded size of 3 bytes */
+    size_t cSeqSizeEstimate = 0;
+    cSeqSizeEstimate += ZSTD_estimateSubBlockSize_symbolType(fseMetadata->ofType, ofCodeTable, MaxOff,
+                                         nbSeq, fseTables->offcodeCTable, NULL,
+                                         OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff,
+                                         workspace, wkspSize);
+    cSeqSizeEstimate += ZSTD_estimateSubBlockSize_symbolType(fseMetadata->llType, llCodeTable, MaxLL,
+                                         nbSeq, fseTables->litlengthCTable, LL_bits,
+                                         LL_defaultNorm, LL_defaultNormLog, MaxLL,
+                                         workspace, wkspSize);
+    cSeqSizeEstimate += ZSTD_estimateSubBlockSize_symbolType(fseMetadata->mlType, mlCodeTable, MaxML,
+                                         nbSeq, fseTables->matchlengthCTable, ML_bits,
+                                         ML_defaultNorm, ML_defaultNormLog, MaxML,
+                                         workspace, wkspSize);
+    if (writeEntropy) cSeqSizeEstimate += fseMetadata->fseTablesSize;
+    return cSeqSizeEstimate + sequencesSectionHeaderSize;
+}
+
+static size_t ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize,
+                                        const BYTE* ofCodeTable,
+                                        const BYTE* llCodeTable,
+                                        const BYTE* mlCodeTable,
+                                        size_t nbSeq,
+                                        const ZSTD_entropyCTables_t* entropy,
+                                        const ZSTD_entropyCTablesMetadata_t* entropyMetadata,
+                                        void* workspace, size_t wkspSize,
+                                        int writeLitEntropy, int writeSeqEntropy) {
+    size_t cSizeEstimate = 0;
+    cSizeEstimate += ZSTD_estimateSubBlockSize_literal(literals, litSize,
+                                                         &entropy->huf, &entropyMetadata->hufMetadata,
+                                                         workspace, wkspSize, writeLitEntropy);
+    cSizeEstimate += ZSTD_estimateSubBlockSize_sequences(ofCodeTable, llCodeTable, mlCodeTable,
+                                                         nbSeq, &entropy->fse, &entropyMetadata->fseMetadata,
+                                                         workspace, wkspSize, writeSeqEntropy);
+    return cSizeEstimate + ZSTD_blockHeaderSize;
+}
+
+static int ZSTD_needSequenceEntropyTables(ZSTD_fseCTablesMetadata_t const* fseMetadata)
+{
+    if (fseMetadata->llType == set_compressed || fseMetadata->llType == set_rle)
+        return 1;
+    if (fseMetadata->mlType == set_compressed || fseMetadata->mlType == set_rle)
+        return 1;
+    if (fseMetadata->ofType == set_compressed || fseMetadata->ofType == set_rle)
+        return 1;
+    return 0;
+}
+
+/* ZSTD_compressSubBlock_multi() :
+ *  Breaks super-block into multiple sub-blocks and compresses them.
+ *  Entropy will be written to the first block.
+ *  The following blocks will use repeat mode to compress.
+ *  All sub-blocks are compressed blocks (no raw or rle blocks).
+ *  @return : compressed size of the super block (which is multiple ZSTD blocks)
+ *            Or 0 if it failed to compress. */
+static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr,
+                            const ZSTD_compressedBlockState_t* prevCBlock,
+                            ZSTD_compressedBlockState_t* nextCBlock,
+                            const ZSTD_entropyCTablesMetadata_t* entropyMetadata,
+                            const ZSTD_CCtx_params* cctxParams,
+                                  void* dst, size_t dstCapacity,
+                            const void* src, size_t srcSize,
+                            const int bmi2, U32 lastBlock,
+                            void* workspace, size_t wkspSize)
+{
+    const seqDef* const sstart = seqStorePtr->sequencesStart;
+    const seqDef* const send = seqStorePtr->sequences;
+    const seqDef* sp = sstart;
+    const BYTE* const lstart = seqStorePtr->litStart;
+    const BYTE* const lend = seqStorePtr->lit;
+    const BYTE* lp = lstart;
+    BYTE const* ip = (BYTE const*)src;
+    BYTE const* const iend = ip + srcSize;
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* const oend = ostart + dstCapacity;
+    BYTE* op = ostart;
+    const BYTE* llCodePtr = seqStorePtr->llCode;
+    const BYTE* mlCodePtr = seqStorePtr->mlCode;
+    const BYTE* ofCodePtr = seqStorePtr->ofCode;
+    size_t targetCBlockSize = cctxParams->targetCBlockSize;
+    size_t litSize, seqCount;
+    int writeLitEntropy = entropyMetadata->hufMetadata.hType == set_compressed;
+    int writeSeqEntropy = 1;
+    int lastSequence = 0;
+
+    DEBUGLOG(5, "ZSTD_compressSubBlock_multi (litSize=%u, nbSeq=%u)",
+                (unsigned)(lend-lp), (unsigned)(send-sstart));
+
+    litSize = 0;
+    seqCount = 0;
+    do {
+        size_t cBlockSizeEstimate = 0;
+        if (sstart == send) {
+            lastSequence = 1;
+        } else {
+            const seqDef* const sequence = sp + seqCount;
+            lastSequence = sequence == send - 1;
+            litSize += ZSTD_getSequenceLength(seqStorePtr, sequence).litLength;
+            seqCount++;
+        }
+        if (lastSequence) {
+            assert(lp <= lend);
+            assert(litSize <= (size_t)(lend - lp));
+            litSize = (size_t)(lend - lp);
+        }
+        /* I think there is an optimization opportunity here.
+         * Calling ZSTD_estimateSubBlockSize for every sequence can be wasteful
+         * since it recalculates estimate from scratch.
+         * For example, it would recount literal distribution and symbol codes everytime.
+         */
+        cBlockSizeEstimate = ZSTD_estimateSubBlockSize(lp, litSize, ofCodePtr, llCodePtr, mlCodePtr, seqCount,
+                                                       &nextCBlock->entropy, entropyMetadata,
+                                                       workspace, wkspSize, writeLitEntropy, writeSeqEntropy);
+        if (cBlockSizeEstimate > targetCBlockSize || lastSequence) {
+            int litEntropyWritten = 0;
+            int seqEntropyWritten = 0;
+            const size_t decompressedSize = ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, lastSequence);
+            const size_t cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata,
+                                                       sp, seqCount,
+                                                       lp, litSize,
+                                                       llCodePtr, mlCodePtr, ofCodePtr,
+                                                       cctxParams,
+                                                       op, oend-op,
+                                                       bmi2, writeLitEntropy, writeSeqEntropy,
+                                                       &litEntropyWritten, &seqEntropyWritten,
+                                                       lastBlock && lastSequence);
+            FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed");
+            if (cSize > 0 && cSize < decompressedSize) {
+                DEBUGLOG(5, "Committed the sub-block");
+                assert(ip + decompressedSize <= iend);
+                ip += decompressedSize;
+                sp += seqCount;
+                lp += litSize;
+                op += cSize;
+                llCodePtr += seqCount;
+                mlCodePtr += seqCount;
+                ofCodePtr += seqCount;
+                litSize = 0;
+                seqCount = 0;
+                /* Entropy only needs to be written once */
+                if (litEntropyWritten) {
+                    writeLitEntropy = 0;
+                }
+                if (seqEntropyWritten) {
+                    writeSeqEntropy = 0;
+                }
+            }
+        }
+    } while (!lastSequence);
+    if (writeLitEntropy) {
+        DEBUGLOG(5, "ZSTD_compressSubBlock_multi has literal entropy tables unwritten");
+        ZSTD_memcpy(&nextCBlock->entropy.huf, &prevCBlock->entropy.huf, sizeof(prevCBlock->entropy.huf));
+    }
+    if (writeSeqEntropy && ZSTD_needSequenceEntropyTables(&entropyMetadata->fseMetadata)) {
+        /* If we haven't written our entropy tables, then we've violated our contract and
+         * must emit an uncompressed block.
+         */
+        DEBUGLOG(5, "ZSTD_compressSubBlock_multi has sequence entropy tables unwritten");
+        return 0;
+    }
+    if (ip < iend) {
+        size_t const cSize = ZSTD_noCompressBlock(op, oend - op, ip, iend - ip, lastBlock);
+        DEBUGLOG(5, "ZSTD_compressSubBlock_multi last sub-block uncompressed, %zu bytes", (size_t)(iend - ip));
+        FORWARD_IF_ERROR(cSize, "ZSTD_noCompressBlock failed");
+        assert(cSize != 0);
+        op += cSize;
+        /* We have to regenerate the repcodes because we've skipped some sequences */
+        if (sp < send) {
+            seqDef const* seq;
+            repcodes_t rep;
+            ZSTD_memcpy(&rep, prevCBlock->rep, sizeof(rep));
+            for (seq = sstart; seq < sp; ++seq) {
+                rep = ZSTD_updateRep(rep.rep, seq->offset - 1, ZSTD_getSequenceLength(seqStorePtr, seq).litLength == 0);
+            }
+            ZSTD_memcpy(nextCBlock->rep, &rep, sizeof(rep));
+        }
+    }
+    DEBUGLOG(5, "ZSTD_compressSubBlock_multi compressed");
+    return op-ostart;
+}
+
+size_t ZSTD_compressSuperBlock(ZSTD_CCtx* zc,
+                               void* dst, size_t dstCapacity,
+                               void const* src, size_t srcSize,
+                               unsigned lastBlock) {
+    ZSTD_entropyCTablesMetadata_t entropyMetadata;
+
+    FORWARD_IF_ERROR(ZSTD_buildSuperBlockEntropy(&zc->seqStore,
+          &zc->blockState.prevCBlock->entropy,
+          &zc->blockState.nextCBlock->entropy,
+          &zc->appliedParams,
+          &entropyMetadata,
+          zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE /* statically allocated in resetCCtx */), "");
+
+    return ZSTD_compressSubBlock_multi(&zc->seqStore,
+            zc->blockState.prevCBlock,
+            zc->blockState.nextCBlock,
+            &entropyMetadata,
+            &zc->appliedParams,
+            dst, dstCapacity,
+            src, srcSize,
+            zc->bmi2, lastBlock,
+            zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE /* statically allocated in resetCCtx */);
+}
diff --git a/lib/zstd/compress/zstd_compress_superblock.h b/lib/zstd/compress/zstd_compress_superblock.h
new file mode 100644
index 0000000..224ece7
--- /dev/null
+++ b/lib/zstd/compress/zstd_compress_superblock.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_COMPRESS_ADVANCED_H
+#define ZSTD_COMPRESS_ADVANCED_H
+
+/*-*************************************
+*  Dependencies
+***************************************/
+
+#include <linux/zstd.h> /* ZSTD_CCtx */
+
+/*-*************************************
+*  Target Compressed Block Size
+***************************************/
+
+/* ZSTD_compressSuperBlock() :
+ * Used to compress a super block when targetCBlockSize is being used.
+ * The given block will be compressed into multiple sub blocks that are around targetCBlockSize. */
+size_t ZSTD_compressSuperBlock(ZSTD_CCtx* zc,
+                               void* dst, size_t dstCapacity,
+                               void const* src, size_t srcSize,
+                               unsigned lastBlock);
+
+#endif /* ZSTD_COMPRESS_ADVANCED_H */
diff --git a/lib/zstd/compress/zstd_cwksp.h b/lib/zstd/compress/zstd_cwksp.h
new file mode 100644
index 0000000..98e359a
--- /dev/null
+++ b/lib/zstd/compress/zstd_cwksp.h
@@ -0,0 +1,482 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_CWKSP_H
+#define ZSTD_CWKSP_H
+
+/*-*************************************
+*  Dependencies
+***************************************/
+#include "../common/zstd_internal.h"
+
+
+/*-*************************************
+*  Constants
+***************************************/
+
+/* Since the workspace is effectively its own little malloc implementation /
+ * arena, when we run under ASAN, we should similarly insert redzones between
+ * each internal element of the workspace, so ASAN will catch overruns that
+ * reach outside an object but that stay inside the workspace.
+ *
+ * This defines the size of that redzone.
+ */
+#ifndef ZSTD_CWKSP_ASAN_REDZONE_SIZE
+#define ZSTD_CWKSP_ASAN_REDZONE_SIZE 128
+#endif
+
+/*-*************************************
+*  Structures
+***************************************/
+typedef enum {
+    ZSTD_cwksp_alloc_objects,
+    ZSTD_cwksp_alloc_buffers,
+    ZSTD_cwksp_alloc_aligned
+} ZSTD_cwksp_alloc_phase_e;
+
+/*
+ * Used to describe whether the workspace is statically allocated (and will not
+ * necessarily ever be freed), or if it's dynamically allocated and we can
+ * expect a well-formed caller to free this.
+ */
+typedef enum {
+    ZSTD_cwksp_dynamic_alloc,
+    ZSTD_cwksp_static_alloc
+} ZSTD_cwksp_static_alloc_e;
+
+/*
+ * Zstd fits all its internal datastructures into a single continuous buffer,
+ * so that it only needs to perform a single OS allocation (or so that a buffer
+ * can be provided to it and it can perform no allocations at all). This buffer
+ * is called the workspace.
+ *
+ * Several optimizations complicate that process of allocating memory ranges
+ * from this workspace for each internal datastructure:
+ *
+ * - These different internal datastructures have different setup requirements:
+ *
+ *   - The static objects need to be cleared once and can then be trivially
+ *     reused for each compression.
+ *
+ *   - Various buffers don't need to be initialized at all--they are always
+ *     written into before they're read.
+ *
+ *   - The matchstate tables have a unique requirement that they don't need
+ *     their memory to be totally cleared, but they do need the memory to have
+ *     some bound, i.e., a guarantee that all values in the memory they've been
+ *     allocated is less than some maximum value (which is the starting value
+ *     for the indices that they will then use for compression). When this
+ *     guarantee is provided to them, they can use the memory without any setup
+ *     work. When it can't, they have to clear the area.
+ *
+ * - These buffers also have different alignment requirements.
+ *
+ * - We would like to reuse the objects in the workspace for multiple
+ *   compressions without having to perform any expensive reallocation or
+ *   reinitialization work.
+ *
+ * - We would like to be able to efficiently reuse the workspace across
+ *   multiple compressions **even when the compression parameters change** and
+ *   we need to resize some of the objects (where possible).
+ *
+ * To attempt to manage this buffer, given these constraints, the ZSTD_cwksp
+ * abstraction was created. It works as follows:
+ *
+ * Workspace Layout:
+ *
+ * [                        ... workspace ...                         ]
+ * [objects][tables ... ->] free space [<- ... aligned][<- ... buffers]
+ *
+ * The various objects that live in the workspace are divided into the
+ * following categories, and are allocated separately:
+ *
+ * - Static objects: this is optionally the enclosing ZSTD_CCtx or ZSTD_CDict,
+ *   so that literally everything fits in a single buffer. Note: if present,
+ *   this must be the first object in the workspace, since ZSTD_customFree{CCtx,
+ *   CDict}() rely on a pointer comparison to see whether one or two frees are
+ *   required.
+ *
+ * - Fixed size objects: these are fixed-size, fixed-count objects that are
+ *   nonetheless "dynamically" allocated in the workspace so that we can
+ *   control how they're initialized separately from the broader ZSTD_CCtx.
+ *   Examples:
+ *   - Entropy Workspace
+ *   - 2 x ZSTD_compressedBlockState_t
+ *   - CDict dictionary contents
+ *
+ * - Tables: these are any of several different datastructures (hash tables,
+ *   chain tables, binary trees) that all respect a common format: they are
+ *   uint32_t arrays, all of whose values are between 0 and (nextSrc - base).
+ *   Their sizes depend on the cparams.
+ *
+ * - Aligned: these buffers are used for various purposes that require 4 byte
+ *   alignment, but don't require any initialization before they're used.
+ *
+ * - Buffers: these buffers are used for various purposes that don't require
+ *   any alignment or initialization before they're used. This means they can
+ *   be moved around at no cost for a new compression.
+ *
+ * Allocating Memory:
+ *
+ * The various types of objects must be allocated in order, so they can be
+ * correctly packed into the workspace buffer. That order is:
+ *
+ * 1. Objects
+ * 2. Buffers
+ * 3. Aligned
+ * 4. Tables
+ *
+ * Attempts to reserve objects of different types out of order will fail.
+ */
+typedef struct {
+    void* workspace;
+    void* workspaceEnd;
+
+    void* objectEnd;
+    void* tableEnd;
+    void* tableValidEnd;
+    void* allocStart;
+
+    BYTE allocFailed;
+    int workspaceOversizedDuration;
+    ZSTD_cwksp_alloc_phase_e phase;
+    ZSTD_cwksp_static_alloc_e isStatic;
+} ZSTD_cwksp;
+
+/*-*************************************
+*  Functions
+***************************************/
+
+MEM_STATIC size_t ZSTD_cwksp_available_space(ZSTD_cwksp* ws);
+
+MEM_STATIC void ZSTD_cwksp_assert_internal_consistency(ZSTD_cwksp* ws) {
+    (void)ws;
+    assert(ws->workspace <= ws->objectEnd);
+    assert(ws->objectEnd <= ws->tableEnd);
+    assert(ws->objectEnd <= ws->tableValidEnd);
+    assert(ws->tableEnd <= ws->allocStart);
+    assert(ws->tableValidEnd <= ws->allocStart);
+    assert(ws->allocStart <= ws->workspaceEnd);
+}
+
+/*
+ * Align must be a power of 2.
+ */
+MEM_STATIC size_t ZSTD_cwksp_align(size_t size, size_t const align) {
+    size_t const mask = align - 1;
+    assert((align & mask) == 0);
+    return (size + mask) & ~mask;
+}
+
+/*
+ * Use this to determine how much space in the workspace we will consume to
+ * allocate this object. (Normally it should be exactly the size of the object,
+ * but under special conditions, like ASAN, where we pad each object, it might
+ * be larger.)
+ *
+ * Since tables aren't currently redzoned, you don't need to call through this
+ * to figure out how much space you need for the matchState tables. Everything
+ * else is though.
+ */
+MEM_STATIC size_t ZSTD_cwksp_alloc_size(size_t size) {
+    if (size == 0)
+        return 0;
+    return size;
+}
+
+MEM_STATIC void ZSTD_cwksp_internal_advance_phase(
+        ZSTD_cwksp* ws, ZSTD_cwksp_alloc_phase_e phase) {
+    assert(phase >= ws->phase);
+    if (phase > ws->phase) {
+        if (ws->phase < ZSTD_cwksp_alloc_buffers &&
+                phase >= ZSTD_cwksp_alloc_buffers) {
+            ws->tableValidEnd = ws->objectEnd;
+        }
+        if (ws->phase < ZSTD_cwksp_alloc_aligned &&
+                phase >= ZSTD_cwksp_alloc_aligned) {
+            /* If unaligned allocations down from a too-large top have left us
+             * unaligned, we need to realign our alloc ptr. Technically, this
+             * can consume space that is unaccounted for in the neededSpace
+             * calculation. However, I believe this can only happen when the
+             * workspace is too large, and specifically when it is too large
+             * by a larger margin than the space that will be consumed. */
+            /* TODO: cleaner, compiler warning friendly way to do this??? */
+            ws->allocStart = (BYTE*)ws->allocStart - ((size_t)ws->allocStart & (sizeof(U32)-1));
+            if (ws->allocStart < ws->tableValidEnd) {
+                ws->tableValidEnd = ws->allocStart;
+            }
+        }
+        ws->phase = phase;
+    }
+}
+
+/*
+ * Returns whether this object/buffer/etc was allocated in this workspace.
+ */
+MEM_STATIC int ZSTD_cwksp_owns_buffer(const ZSTD_cwksp* ws, const void* ptr) {
+    return (ptr != NULL) && (ws->workspace <= ptr) && (ptr <= ws->workspaceEnd);
+}
+
+/*
+ * Internal function. Do not use directly.
+ */
+MEM_STATIC void* ZSTD_cwksp_reserve_internal(
+        ZSTD_cwksp* ws, size_t bytes, ZSTD_cwksp_alloc_phase_e phase) {
+    void* alloc;
+    void* bottom = ws->tableEnd;
+    ZSTD_cwksp_internal_advance_phase(ws, phase);
+    alloc = (BYTE *)ws->allocStart - bytes;
+
+    if (bytes == 0)
+        return NULL;
+
+
+    DEBUGLOG(5, "cwksp: reserving %p %zd bytes, %zd bytes remaining",
+        alloc, bytes, ZSTD_cwksp_available_space(ws) - bytes);
+    ZSTD_cwksp_assert_internal_consistency(ws);
+    assert(alloc >= bottom);
+    if (alloc < bottom) {
+        DEBUGLOG(4, "cwksp: alloc failed!");
+        ws->allocFailed = 1;
+        return NULL;
+    }
+    if (alloc < ws->tableValidEnd) {
+        ws->tableValidEnd = alloc;
+    }
+    ws->allocStart = alloc;
+
+
+    return alloc;
+}
+
+/*
+ * Reserves and returns unaligned memory.
+ */
+MEM_STATIC BYTE* ZSTD_cwksp_reserve_buffer(ZSTD_cwksp* ws, size_t bytes) {
+    return (BYTE*)ZSTD_cwksp_reserve_internal(ws, bytes, ZSTD_cwksp_alloc_buffers);
+}
+
+/*
+ * Reserves and returns memory sized on and aligned on sizeof(unsigned).
+ */
+MEM_STATIC void* ZSTD_cwksp_reserve_aligned(ZSTD_cwksp* ws, size_t bytes) {
+    assert((bytes & (sizeof(U32)-1)) == 0);
+    return ZSTD_cwksp_reserve_internal(ws, ZSTD_cwksp_align(bytes, sizeof(U32)), ZSTD_cwksp_alloc_aligned);
+}
+
+/*
+ * Aligned on sizeof(unsigned). These buffers have the special property that
+ * their values remain constrained, allowing us to re-use them without
+ * memset()-ing them.
+ */
+MEM_STATIC void* ZSTD_cwksp_reserve_table(ZSTD_cwksp* ws, size_t bytes) {
+    const ZSTD_cwksp_alloc_phase_e phase = ZSTD_cwksp_alloc_aligned;
+    void* alloc = ws->tableEnd;
+    void* end = (BYTE *)alloc + bytes;
+    void* top = ws->allocStart;
+
+    DEBUGLOG(5, "cwksp: reserving %p table %zd bytes, %zd bytes remaining",
+        alloc, bytes, ZSTD_cwksp_available_space(ws) - bytes);
+    assert((bytes & (sizeof(U32)-1)) == 0);
+    ZSTD_cwksp_internal_advance_phase(ws, phase);
+    ZSTD_cwksp_assert_internal_consistency(ws);
+    assert(end <= top);
+    if (end > top) {
+        DEBUGLOG(4, "cwksp: table alloc failed!");
+        ws->allocFailed = 1;
+        return NULL;
+    }
+    ws->tableEnd = end;
+
+
+    return alloc;
+}
+
+/*
+ * Aligned on sizeof(void*).
+ */
+MEM_STATIC void* ZSTD_cwksp_reserve_object(ZSTD_cwksp* ws, size_t bytes) {
+    size_t roundedBytes = ZSTD_cwksp_align(bytes, sizeof(void*));
+    void* alloc = ws->objectEnd;
+    void* end = (BYTE*)alloc + roundedBytes;
+
+
+    DEBUGLOG(5,
+        "cwksp: reserving %p object %zd bytes (rounded to %zd), %zd bytes remaining",
+        alloc, bytes, roundedBytes, ZSTD_cwksp_available_space(ws) - roundedBytes);
+    assert(((size_t)alloc & (sizeof(void*)-1)) == 0);
+    assert((bytes & (sizeof(void*)-1)) == 0);
+    ZSTD_cwksp_assert_internal_consistency(ws);
+    /* we must be in the first phase, no advance is possible */
+    if (ws->phase != ZSTD_cwksp_alloc_objects || end > ws->workspaceEnd) {
+        DEBUGLOG(4, "cwksp: object alloc failed!");
+        ws->allocFailed = 1;
+        return NULL;
+    }
+    ws->objectEnd = end;
+    ws->tableEnd = end;
+    ws->tableValidEnd = end;
+
+
+    return alloc;
+}
+
+MEM_STATIC void ZSTD_cwksp_mark_tables_dirty(ZSTD_cwksp* ws) {
+    DEBUGLOG(4, "cwksp: ZSTD_cwksp_mark_tables_dirty");
+
+
+    assert(ws->tableValidEnd >= ws->objectEnd);
+    assert(ws->tableValidEnd <= ws->allocStart);
+    ws->tableValidEnd = ws->objectEnd;
+    ZSTD_cwksp_assert_internal_consistency(ws);
+}
+
+MEM_STATIC void ZSTD_cwksp_mark_tables_clean(ZSTD_cwksp* ws) {
+    DEBUGLOG(4, "cwksp: ZSTD_cwksp_mark_tables_clean");
+    assert(ws->tableValidEnd >= ws->objectEnd);
+    assert(ws->tableValidEnd <= ws->allocStart);
+    if (ws->tableValidEnd < ws->tableEnd) {
+        ws->tableValidEnd = ws->tableEnd;
+    }
+    ZSTD_cwksp_assert_internal_consistency(ws);
+}
+
+/*
+ * Zero the part of the allocated tables not already marked clean.
+ */
+MEM_STATIC void ZSTD_cwksp_clean_tables(ZSTD_cwksp* ws) {
+    DEBUGLOG(4, "cwksp: ZSTD_cwksp_clean_tables");
+    assert(ws->tableValidEnd >= ws->objectEnd);
+    assert(ws->tableValidEnd <= ws->allocStart);
+    if (ws->tableValidEnd < ws->tableEnd) {
+        ZSTD_memset(ws->tableValidEnd, 0, (BYTE*)ws->tableEnd - (BYTE*)ws->tableValidEnd);
+    }
+    ZSTD_cwksp_mark_tables_clean(ws);
+}
+
+/*
+ * Invalidates table allocations.
+ * All other allocations remain valid.
+ */
+MEM_STATIC void ZSTD_cwksp_clear_tables(ZSTD_cwksp* ws) {
+    DEBUGLOG(4, "cwksp: clearing tables!");
+
+
+    ws->tableEnd = ws->objectEnd;
+    ZSTD_cwksp_assert_internal_consistency(ws);
+}
+
+/*
+ * Invalidates all buffer, aligned, and table allocations.
+ * Object allocations remain valid.
+ */
+MEM_STATIC void ZSTD_cwksp_clear(ZSTD_cwksp* ws) {
+    DEBUGLOG(4, "cwksp: clearing!");
+
+
+
+    ws->tableEnd = ws->objectEnd;
+    ws->allocStart = ws->workspaceEnd;
+    ws->allocFailed = 0;
+    if (ws->phase > ZSTD_cwksp_alloc_buffers) {
+        ws->phase = ZSTD_cwksp_alloc_buffers;
+    }
+    ZSTD_cwksp_assert_internal_consistency(ws);
+}
+
+/*
+ * The provided workspace takes ownership of the buffer [start, start+size).
+ * Any existing values in the workspace are ignored (the previously managed
+ * buffer, if present, must be separately freed).
+ */
+MEM_STATIC void ZSTD_cwksp_init(ZSTD_cwksp* ws, void* start, size_t size, ZSTD_cwksp_static_alloc_e isStatic) {
+    DEBUGLOG(4, "cwksp: init'ing workspace with %zd bytes", size);
+    assert(((size_t)start & (sizeof(void*)-1)) == 0); /* ensure correct alignment */
+    ws->workspace = start;
+    ws->workspaceEnd = (BYTE*)start + size;
+    ws->objectEnd = ws->workspace;
+    ws->tableValidEnd = ws->objectEnd;
+    ws->phase = ZSTD_cwksp_alloc_objects;
+    ws->isStatic = isStatic;
+    ZSTD_cwksp_clear(ws);
+    ws->workspaceOversizedDuration = 0;
+    ZSTD_cwksp_assert_internal_consistency(ws);
+}
+
+MEM_STATIC size_t ZSTD_cwksp_create(ZSTD_cwksp* ws, size_t size, ZSTD_customMem customMem) {
+    void* workspace = ZSTD_customMalloc(size, customMem);
+    DEBUGLOG(4, "cwksp: creating new workspace with %zd bytes", size);
+    RETURN_ERROR_IF(workspace == NULL, memory_allocation, "NULL pointer!");
+    ZSTD_cwksp_init(ws, workspace, size, ZSTD_cwksp_dynamic_alloc);
+    return 0;
+}
+
+MEM_STATIC void ZSTD_cwksp_free(ZSTD_cwksp* ws, ZSTD_customMem customMem) {
+    void *ptr = ws->workspace;
+    DEBUGLOG(4, "cwksp: freeing workspace");
+    ZSTD_memset(ws, 0, sizeof(ZSTD_cwksp));
+    ZSTD_customFree(ptr, customMem);
+}
+
+/*
+ * Moves the management of a workspace from one cwksp to another. The src cwksp
+ * is left in an invalid state (src must be re-init()'ed before it's used again).
+ */
+MEM_STATIC void ZSTD_cwksp_move(ZSTD_cwksp* dst, ZSTD_cwksp* src) {
+    *dst = *src;
+    ZSTD_memset(src, 0, sizeof(ZSTD_cwksp));
+}
+
+MEM_STATIC size_t ZSTD_cwksp_sizeof(const ZSTD_cwksp* ws) {
+    return (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->workspace);
+}
+
+MEM_STATIC size_t ZSTD_cwksp_used(const ZSTD_cwksp* ws) {
+    return (size_t)((BYTE*)ws->tableEnd - (BYTE*)ws->workspace)
+         + (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->allocStart);
+}
+
+MEM_STATIC int ZSTD_cwksp_reserve_failed(const ZSTD_cwksp* ws) {
+    return ws->allocFailed;
+}
+
+/*-*************************************
+*  Functions Checking Free Space
+***************************************/
+
+MEM_STATIC size_t ZSTD_cwksp_available_space(ZSTD_cwksp* ws) {
+    return (size_t)((BYTE*)ws->allocStart - (BYTE*)ws->tableEnd);
+}
+
+MEM_STATIC int ZSTD_cwksp_check_available(ZSTD_cwksp* ws, size_t additionalNeededSpace) {
+    return ZSTD_cwksp_available_space(ws) >= additionalNeededSpace;
+}
+
+MEM_STATIC int ZSTD_cwksp_check_too_large(ZSTD_cwksp* ws, size_t additionalNeededSpace) {
+    return ZSTD_cwksp_check_available(
+        ws, additionalNeededSpace * ZSTD_WORKSPACETOOLARGE_FACTOR);
+}
+
+MEM_STATIC int ZSTD_cwksp_check_wasteful(ZSTD_cwksp* ws, size_t additionalNeededSpace) {
+    return ZSTD_cwksp_check_too_large(ws, additionalNeededSpace)
+        && ws->workspaceOversizedDuration > ZSTD_WORKSPACETOOLARGE_MAXDURATION;
+}
+
+MEM_STATIC void ZSTD_cwksp_bump_oversized_duration(
+        ZSTD_cwksp* ws, size_t additionalNeededSpace) {
+    if (ZSTD_cwksp_check_too_large(ws, additionalNeededSpace)) {
+        ws->workspaceOversizedDuration++;
+    } else {
+        ws->workspaceOversizedDuration = 0;
+    }
+}
+
+
+#endif /* ZSTD_CWKSP_H */
diff --git a/lib/zstd/compress/zstd_double_fast.c b/lib/zstd/compress/zstd_double_fast.c
new file mode 100644
index 0000000..b0424d2
--- /dev/null
+++ b/lib/zstd/compress/zstd_double_fast.c
@@ -0,0 +1,519 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#include "zstd_compress_internal.h"
+#include "zstd_double_fast.h"
+
+
+void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms,
+                              void const* end, ZSTD_dictTableLoadMethod_e dtlm)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32* const hashLarge = ms->hashTable;
+    U32  const hBitsL = cParams->hashLog;
+    U32  const mls = cParams->minMatch;
+    U32* const hashSmall = ms->chainTable;
+    U32  const hBitsS = cParams->chainLog;
+    const BYTE* const base = ms->window.base;
+    const BYTE* ip = base + ms->nextToUpdate;
+    const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;
+    const U32 fastHashFillStep = 3;
+
+    /* Always insert every fastHashFillStep position into the hash tables.
+     * Insert the other positions into the large hash table if their entry
+     * is empty.
+     */
+    for (; ip + fastHashFillStep - 1 <= iend; ip += fastHashFillStep) {
+        U32 const curr = (U32)(ip - base);
+        U32 i;
+        for (i = 0; i < fastHashFillStep; ++i) {
+            size_t const smHash = ZSTD_hashPtr(ip + i, hBitsS, mls);
+            size_t const lgHash = ZSTD_hashPtr(ip + i, hBitsL, 8);
+            if (i == 0)
+                hashSmall[smHash] = curr + i;
+            if (i == 0 || hashLarge[lgHash] == 0)
+                hashLarge[lgHash] = curr + i;
+            /* Only load extra positions for ZSTD_dtlm_full */
+            if (dtlm == ZSTD_dtlm_fast)
+                break;
+    }   }
+}
+
+
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_doubleFast_generic(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize,
+        U32 const mls /* template */, ZSTD_dictMode_e const dictMode)
+{
+    ZSTD_compressionParameters const* cParams = &ms->cParams;
+    U32* const hashLong = ms->hashTable;
+    const U32 hBitsL = cParams->hashLog;
+    U32* const hashSmall = ms->chainTable;
+    const U32 hBitsS = cParams->chainLog;
+    const BYTE* const base = ms->window.base;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
+    /* presumes that, if there is a dictionary, it must be using Attach mode */
+    const U32 prefixLowestIndex = ZSTD_getLowestPrefixIndex(ms, endIndex, cParams->windowLog);
+    const BYTE* const prefixLowest = base + prefixLowestIndex;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - HASH_READ_SIZE;
+    U32 offset_1=rep[0], offset_2=rep[1];
+    U32 offsetSaved = 0;
+
+    const ZSTD_matchState_t* const dms = ms->dictMatchState;
+    const ZSTD_compressionParameters* const dictCParams =
+                                     dictMode == ZSTD_dictMatchState ?
+                                     &dms->cParams : NULL;
+    const U32* const dictHashLong  = dictMode == ZSTD_dictMatchState ?
+                                     dms->hashTable : NULL;
+    const U32* const dictHashSmall = dictMode == ZSTD_dictMatchState ?
+                                     dms->chainTable : NULL;
+    const U32 dictStartIndex       = dictMode == ZSTD_dictMatchState ?
+                                     dms->window.dictLimit : 0;
+    const BYTE* const dictBase     = dictMode == ZSTD_dictMatchState ?
+                                     dms->window.base : NULL;
+    const BYTE* const dictStart    = dictMode == ZSTD_dictMatchState ?
+                                     dictBase + dictStartIndex : NULL;
+    const BYTE* const dictEnd      = dictMode == ZSTD_dictMatchState ?
+                                     dms->window.nextSrc : NULL;
+    const U32 dictIndexDelta       = dictMode == ZSTD_dictMatchState ?
+                                     prefixLowestIndex - (U32)(dictEnd - dictBase) :
+                                     0;
+    const U32 dictHBitsL           = dictMode == ZSTD_dictMatchState ?
+                                     dictCParams->hashLog : hBitsL;
+    const U32 dictHBitsS           = dictMode == ZSTD_dictMatchState ?
+                                     dictCParams->chainLog : hBitsS;
+    const U32 dictAndPrefixLength  = (U32)((ip - prefixLowest) + (dictEnd - dictStart));
+
+    DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_generic");
+
+    assert(dictMode == ZSTD_noDict || dictMode == ZSTD_dictMatchState);
+
+    /* if a dictionary is attached, it must be within window range */
+    if (dictMode == ZSTD_dictMatchState) {
+        assert(ms->window.dictLimit + (1U << cParams->windowLog) >= endIndex);
+    }
+
+    /* init */
+    ip += (dictAndPrefixLength == 0);
+    if (dictMode == ZSTD_noDict) {
+        U32 const curr = (U32)(ip - base);
+        U32 const windowLow = ZSTD_getLowestPrefixIndex(ms, curr, cParams->windowLog);
+        U32 const maxRep = curr - windowLow;
+        if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
+        if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
+    }
+    if (dictMode == ZSTD_dictMatchState) {
+        /* dictMatchState repCode checks don't currently handle repCode == 0
+         * disabling. */
+        assert(offset_1 <= dictAndPrefixLength);
+        assert(offset_2 <= dictAndPrefixLength);
+    }
+
+    /* Main Search Loop */
+    while (ip < ilimit) {   /* < instead of <=, because repcode check at (ip+1) */
+        size_t mLength;
+        U32 offset;
+        size_t const h2 = ZSTD_hashPtr(ip, hBitsL, 8);
+        size_t const h = ZSTD_hashPtr(ip, hBitsS, mls);
+        size_t const dictHL = ZSTD_hashPtr(ip, dictHBitsL, 8);
+        size_t const dictHS = ZSTD_hashPtr(ip, dictHBitsS, mls);
+        U32 const curr = (U32)(ip-base);
+        U32 const matchIndexL = hashLong[h2];
+        U32 matchIndexS = hashSmall[h];
+        const BYTE* matchLong = base + matchIndexL;
+        const BYTE* match = base + matchIndexS;
+        const U32 repIndex = curr + 1 - offset_1;
+        const BYTE* repMatch = (dictMode == ZSTD_dictMatchState
+                            && repIndex < prefixLowestIndex) ?
+                               dictBase + (repIndex - dictIndexDelta) :
+                               base + repIndex;
+        hashLong[h2] = hashSmall[h] = curr;   /* update hash tables */
+
+        /* check dictMatchState repcode */
+        if (dictMode == ZSTD_dictMatchState
+            && ((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)
+            && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
+            const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
+            mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
+            ip++;
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength-MINMATCH);
+            goto _match_stored;
+        }
+
+        /* check noDict repcode */
+        if ( dictMode == ZSTD_noDict
+          && ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1)))) {
+            mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
+            ip++;
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength-MINMATCH);
+            goto _match_stored;
+        }
+
+        if (matchIndexL > prefixLowestIndex) {
+            /* check prefix long match */
+            if (MEM_read64(matchLong) == MEM_read64(ip)) {
+                mLength = ZSTD_count(ip+8, matchLong+8, iend) + 8;
+                offset = (U32)(ip-matchLong);
+                while (((ip>anchor) & (matchLong>prefixLowest)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
+                goto _match_found;
+            }
+        } else if (dictMode == ZSTD_dictMatchState) {
+            /* check dictMatchState long match */
+            U32 const dictMatchIndexL = dictHashLong[dictHL];
+            const BYTE* dictMatchL = dictBase + dictMatchIndexL;
+            assert(dictMatchL < dictEnd);
+
+            if (dictMatchL > dictStart && MEM_read64(dictMatchL) == MEM_read64(ip)) {
+                mLength = ZSTD_count_2segments(ip+8, dictMatchL+8, iend, dictEnd, prefixLowest) + 8;
+                offset = (U32)(curr - dictMatchIndexL - dictIndexDelta);
+                while (((ip>anchor) & (dictMatchL>dictStart)) && (ip[-1] == dictMatchL[-1])) { ip--; dictMatchL--; mLength++; } /* catch up */
+                goto _match_found;
+        }   }
+
+        if (matchIndexS > prefixLowestIndex) {
+            /* check prefix short match */
+            if (MEM_read32(match) == MEM_read32(ip)) {
+                goto _search_next_long;
+            }
+        } else if (dictMode == ZSTD_dictMatchState) {
+            /* check dictMatchState short match */
+            U32 const dictMatchIndexS = dictHashSmall[dictHS];
+            match = dictBase + dictMatchIndexS;
+            matchIndexS = dictMatchIndexS + dictIndexDelta;
+
+            if (match > dictStart && MEM_read32(match) == MEM_read32(ip)) {
+                goto _search_next_long;
+        }   }
+
+        ip += ((ip-anchor) >> kSearchStrength) + 1;
+#if defined(__aarch64__)
+        PREFETCH_L1(ip+256);
+#endif
+        continue;
+
+_search_next_long:
+
+        {   size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
+            size_t const dictHLNext = ZSTD_hashPtr(ip+1, dictHBitsL, 8);
+            U32 const matchIndexL3 = hashLong[hl3];
+            const BYTE* matchL3 = base + matchIndexL3;
+            hashLong[hl3] = curr + 1;
+
+            /* check prefix long +1 match */
+            if (matchIndexL3 > prefixLowestIndex) {
+                if (MEM_read64(matchL3) == MEM_read64(ip+1)) {
+                    mLength = ZSTD_count(ip+9, matchL3+8, iend) + 8;
+                    ip++;
+                    offset = (U32)(ip-matchL3);
+                    while (((ip>anchor) & (matchL3>prefixLowest)) && (ip[-1] == matchL3[-1])) { ip--; matchL3--; mLength++; } /* catch up */
+                    goto _match_found;
+                }
+            } else if (dictMode == ZSTD_dictMatchState) {
+                /* check dict long +1 match */
+                U32 const dictMatchIndexL3 = dictHashLong[dictHLNext];
+                const BYTE* dictMatchL3 = dictBase + dictMatchIndexL3;
+                assert(dictMatchL3 < dictEnd);
+                if (dictMatchL3 > dictStart && MEM_read64(dictMatchL3) == MEM_read64(ip+1)) {
+                    mLength = ZSTD_count_2segments(ip+1+8, dictMatchL3+8, iend, dictEnd, prefixLowest) + 8;
+                    ip++;
+                    offset = (U32)(curr + 1 - dictMatchIndexL3 - dictIndexDelta);
+                    while (((ip>anchor) & (dictMatchL3>dictStart)) && (ip[-1] == dictMatchL3[-1])) { ip--; dictMatchL3--; mLength++; } /* catch up */
+                    goto _match_found;
+        }   }   }
+
+        /* if no long +1 match, explore the short match we found */
+        if (dictMode == ZSTD_dictMatchState && matchIndexS < prefixLowestIndex) {
+            mLength = ZSTD_count_2segments(ip+4, match+4, iend, dictEnd, prefixLowest) + 4;
+            offset = (U32)(curr - matchIndexS);
+            while (((ip>anchor) & (match>dictStart)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
+        } else {
+            mLength = ZSTD_count(ip+4, match+4, iend) + 4;
+            offset = (U32)(ip - match);
+            while (((ip>anchor) & (match>prefixLowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
+        }
+
+_match_found:
+        offset_2 = offset_1;
+        offset_1 = offset;
+
+        ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+
+_match_stored:
+        /* match found */
+        ip += mLength;
+        anchor = ip;
+
+        if (ip <= ilimit) {
+            /* Complementary insertion */
+            /* done after iLimit test, as candidates could be > iend-8 */
+            {   U32 const indexToInsert = curr+2;
+                hashLong[ZSTD_hashPtr(base+indexToInsert, hBitsL, 8)] = indexToInsert;
+                hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
+                hashSmall[ZSTD_hashPtr(base+indexToInsert, hBitsS, mls)] = indexToInsert;
+                hashSmall[ZSTD_hashPtr(ip-1, hBitsS, mls)] = (U32)(ip-1-base);
+            }
+
+            /* check immediate repcode */
+            if (dictMode == ZSTD_dictMatchState) {
+                while (ip <= ilimit) {
+                    U32 const current2 = (U32)(ip-base);
+                    U32 const repIndex2 = current2 - offset_2;
+                    const BYTE* repMatch2 = dictMode == ZSTD_dictMatchState
+                        && repIndex2 < prefixLowestIndex ?
+                            dictBase + repIndex2 - dictIndexDelta :
+                            base + repIndex2;
+                    if ( ((U32)((prefixLowestIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
+                       && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
+                        const BYTE* const repEnd2 = repIndex2 < prefixLowestIndex ? dictEnd : iend;
+                        size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixLowest) + 4;
+                        U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
+                        ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, repLength2-MINMATCH);
+                        hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;
+                        hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;
+                        ip += repLength2;
+                        anchor = ip;
+                        continue;
+                    }
+                    break;
+            }   }
+
+            if (dictMode == ZSTD_noDict) {
+                while ( (ip <= ilimit)
+                     && ( (offset_2>0)
+                        & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
+                    /* store sequence */
+                    size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
+                    U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff;  /* swap offset_2 <=> offset_1 */
+                    hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip-base);
+                    hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip-base);
+                    ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, rLength-MINMATCH);
+                    ip += rLength;
+                    anchor = ip;
+                    continue;   /* faster when present ... (?) */
+        }   }   }
+    }   /* while (ip < ilimit) */
+
+    /* save reps for next block */
+    rep[0] = offset_1 ? offset_1 : offsetSaved;
+    rep[1] = offset_2 ? offset_2 : offsetSaved;
+
+    /* Return the last literals size */
+    return (size_t)(iend - anchor);
+}
+
+
+size_t ZSTD_compressBlock_doubleFast(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    const U32 mls = ms->cParams.minMatch;
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 4, ZSTD_noDict);
+    case 5 :
+        return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 5, ZSTD_noDict);
+    case 6 :
+        return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 6, ZSTD_noDict);
+    case 7 :
+        return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 7, ZSTD_noDict);
+    }
+}
+
+
+size_t ZSTD_compressBlock_doubleFast_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    const U32 mls = ms->cParams.minMatch;
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 4, ZSTD_dictMatchState);
+    case 5 :
+        return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 5, ZSTD_dictMatchState);
+    case 6 :
+        return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 6, ZSTD_dictMatchState);
+    case 7 :
+        return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 7, ZSTD_dictMatchState);
+    }
+}
+
+
+static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize,
+        U32 const mls /* template */)
+{
+    ZSTD_compressionParameters const* cParams = &ms->cParams;
+    U32* const hashLong = ms->hashTable;
+    U32  const hBitsL = cParams->hashLog;
+    U32* const hashSmall = ms->chainTable;
+    U32  const hBitsS = cParams->chainLog;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    const BYTE* const base = ms->window.base;
+    const U32   endIndex = (U32)((size_t)(istart - base) + srcSize);
+    const U32   lowLimit = ZSTD_getLowestMatchIndex(ms, endIndex, cParams->windowLog);
+    const U32   dictStartIndex = lowLimit;
+    const U32   dictLimit = ms->window.dictLimit;
+    const U32   prefixStartIndex = (dictLimit > lowLimit) ? dictLimit : lowLimit;
+    const BYTE* const prefixStart = base + prefixStartIndex;
+    const BYTE* const dictBase = ms->window.dictBase;
+    const BYTE* const dictStart = dictBase + dictStartIndex;
+    const BYTE* const dictEnd = dictBase + prefixStartIndex;
+    U32 offset_1=rep[0], offset_2=rep[1];
+
+    DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_extDict_generic (srcSize=%zu)", srcSize);
+
+    /* if extDict is invalidated due to maxDistance, switch to "regular" variant */
+    if (prefixStartIndex == dictStartIndex)
+        return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, mls, ZSTD_noDict);
+
+    /* Search Loop */
+    while (ip < ilimit) {  /* < instead of <=, because (ip+1) */
+        const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);
+        const U32 matchIndex = hashSmall[hSmall];
+        const BYTE* const matchBase = matchIndex < prefixStartIndex ? dictBase : base;
+        const BYTE* match = matchBase + matchIndex;
+
+        const size_t hLong = ZSTD_hashPtr(ip, hBitsL, 8);
+        const U32 matchLongIndex = hashLong[hLong];
+        const BYTE* const matchLongBase = matchLongIndex < prefixStartIndex ? dictBase : base;
+        const BYTE* matchLong = matchLongBase + matchLongIndex;
+
+        const U32 curr = (U32)(ip-base);
+        const U32 repIndex = curr + 1 - offset_1;   /* offset_1 expected <= curr +1 */
+        const BYTE* const repBase = repIndex < prefixStartIndex ? dictBase : base;
+        const BYTE* const repMatch = repBase + repIndex;
+        size_t mLength;
+        hashSmall[hSmall] = hashLong[hLong] = curr;   /* update hash table */
+
+        if ((((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex doesn't overlap dict + prefix */
+            & (repIndex > dictStartIndex))
+          && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
+            const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
+            mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
+            ip++;
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength-MINMATCH);
+        } else {
+            if ((matchLongIndex > dictStartIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {
+                const BYTE* const matchEnd = matchLongIndex < prefixStartIndex ? dictEnd : iend;
+                const BYTE* const lowMatchPtr = matchLongIndex < prefixStartIndex ? dictStart : prefixStart;
+                U32 offset;
+                mLength = ZSTD_count_2segments(ip+8, matchLong+8, iend, matchEnd, prefixStart) + 8;
+                offset = curr - matchLongIndex;
+                while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; }   /* catch up */
+                offset_2 = offset_1;
+                offset_1 = offset;
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+
+            } else if ((matchIndex > dictStartIndex) && (MEM_read32(match) == MEM_read32(ip))) {
+                size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
+                U32 const matchIndex3 = hashLong[h3];
+                const BYTE* const match3Base = matchIndex3 < prefixStartIndex ? dictBase : base;
+                const BYTE* match3 = match3Base + matchIndex3;
+                U32 offset;
+                hashLong[h3] = curr + 1;
+                if ( (matchIndex3 > dictStartIndex) && (MEM_read64(match3) == MEM_read64(ip+1)) ) {
+                    const BYTE* const matchEnd = matchIndex3 < prefixStartIndex ? dictEnd : iend;
+                    const BYTE* const lowMatchPtr = matchIndex3 < prefixStartIndex ? dictStart : prefixStart;
+                    mLength = ZSTD_count_2segments(ip+9, match3+8, iend, matchEnd, prefixStart) + 8;
+                    ip++;
+                    offset = curr+1 - matchIndex3;
+                    while (((ip>anchor) & (match3>lowMatchPtr)) && (ip[-1] == match3[-1])) { ip--; match3--; mLength++; } /* catch up */
+                } else {
+                    const BYTE* const matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend;
+                    const BYTE* const lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart;
+                    mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4;
+                    offset = curr - matchIndex;
+                    while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; }   /* catch up */
+                }
+                offset_2 = offset_1;
+                offset_1 = offset;
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+
+            } else {
+                ip += ((ip-anchor) >> kSearchStrength) + 1;
+                continue;
+        }   }
+
+        /* move to next sequence start */
+        ip += mLength;
+        anchor = ip;
+
+        if (ip <= ilimit) {
+            /* Complementary insertion */
+            /* done after iLimit test, as candidates could be > iend-8 */
+            {   U32 const indexToInsert = curr+2;
+                hashLong[ZSTD_hashPtr(base+indexToInsert, hBitsL, 8)] = indexToInsert;
+                hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
+                hashSmall[ZSTD_hashPtr(base+indexToInsert, hBitsS, mls)] = indexToInsert;
+                hashSmall[ZSTD_hashPtr(ip-1, hBitsS, mls)] = (U32)(ip-1-base);
+            }
+
+            /* check immediate repcode */
+            while (ip <= ilimit) {
+                U32 const current2 = (U32)(ip-base);
+                U32 const repIndex2 = current2 - offset_2;
+                const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;
+                if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3)   /* intentional overflow : ensure repIndex2 doesn't overlap dict + prefix */
+                    & (repIndex2 > dictStartIndex))
+                  && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
+                    const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
+                    size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
+                    U32 const tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
+                    ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, repLength2-MINMATCH);
+                    hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;
+                    hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;
+                    ip += repLength2;
+                    anchor = ip;
+                    continue;
+                }
+                break;
+    }   }   }
+
+    /* save reps for next block */
+    rep[0] = offset_1;
+    rep[1] = offset_2;
+
+    /* Return the last literals size */
+    return (size_t)(iend - anchor);
+}
+
+
+size_t ZSTD_compressBlock_doubleFast_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    U32 const mls = ms->cParams.minMatch;
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_doubleFast_extDict_generic(ms, seqStore, rep, src, srcSize, 4);
+    case 5 :
+        return ZSTD_compressBlock_doubleFast_extDict_generic(ms, seqStore, rep, src, srcSize, 5);
+    case 6 :
+        return ZSTD_compressBlock_doubleFast_extDict_generic(ms, seqStore, rep, src, srcSize, 6);
+    case 7 :
+        return ZSTD_compressBlock_doubleFast_extDict_generic(ms, seqStore, rep, src, srcSize, 7);
+    }
+}
diff --git a/lib/zstd/compress/zstd_double_fast.h b/lib/zstd/compress/zstd_double_fast.h
new file mode 100644
index 0000000..6822bde
--- /dev/null
+++ b/lib/zstd/compress/zstd_double_fast.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_DOUBLE_FAST_H
+#define ZSTD_DOUBLE_FAST_H
+
+
+#include "../common/mem.h"      /* U32 */
+#include "zstd_compress_internal.h"     /* ZSTD_CCtx, size_t */
+
+void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms,
+                              void const* end, ZSTD_dictTableLoadMethod_e dtlm);
+size_t ZSTD_compressBlock_doubleFast(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_doubleFast_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_doubleFast_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+
+
+
+#endif /* ZSTD_DOUBLE_FAST_H */
diff --git a/lib/zstd/compress/zstd_fast.c b/lib/zstd/compress/zstd_fast.c
new file mode 100644
index 0000000..96b7d48
--- /dev/null
+++ b/lib/zstd/compress/zstd_fast.c
@@ -0,0 +1,496 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#include "zstd_compress_internal.h"  /* ZSTD_hashPtr, ZSTD_count, ZSTD_storeSeq */
+#include "zstd_fast.h"
+
+
+void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
+                        const void* const end,
+                        ZSTD_dictTableLoadMethod_e dtlm)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32* const hashTable = ms->hashTable;
+    U32  const hBits = cParams->hashLog;
+    U32  const mls = cParams->minMatch;
+    const BYTE* const base = ms->window.base;
+    const BYTE* ip = base + ms->nextToUpdate;
+    const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;
+    const U32 fastHashFillStep = 3;
+
+    /* Always insert every fastHashFillStep position into the hash table.
+     * Insert the other positions if their hash entry is empty.
+     */
+    for ( ; ip + fastHashFillStep < iend + 2; ip += fastHashFillStep) {
+        U32 const curr = (U32)(ip - base);
+        size_t const hash0 = ZSTD_hashPtr(ip, hBits, mls);
+        hashTable[hash0] = curr;
+        if (dtlm == ZSTD_dtlm_fast) continue;
+        /* Only load extra positions for ZSTD_dtlm_full */
+        {   U32 p;
+            for (p = 1; p < fastHashFillStep; ++p) {
+                size_t const hash = ZSTD_hashPtr(ip + p, hBits, mls);
+                if (hashTable[hash] == 0) {  /* not yet filled */
+                    hashTable[hash] = curr + p;
+    }   }   }   }
+}
+
+
+FORCE_INLINE_TEMPLATE size_t
+ZSTD_compressBlock_fast_generic(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize,
+        U32 const mls)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32* const hashTable = ms->hashTable;
+    U32 const hlog = cParams->hashLog;
+    /* support stepSize of 0 */
+    size_t const stepSize = cParams->targetLength + !(cParams->targetLength) + 1;
+    const BYTE* const base = ms->window.base;
+    const BYTE* const istart = (const BYTE*)src;
+    /* We check ip0 (ip + 0) and ip1 (ip + 1) each loop */
+    const BYTE* ip0 = istart;
+    const BYTE* ip1;
+    const BYTE* anchor = istart;
+    const U32   endIndex = (U32)((size_t)(istart - base) + srcSize);
+    const U32   prefixStartIndex = ZSTD_getLowestPrefixIndex(ms, endIndex, cParams->windowLog);
+    const BYTE* const prefixStart = base + prefixStartIndex;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - HASH_READ_SIZE;
+    U32 offset_1=rep[0], offset_2=rep[1];
+    U32 offsetSaved = 0;
+
+    /* init */
+    DEBUGLOG(5, "ZSTD_compressBlock_fast_generic");
+    ip0 += (ip0 == prefixStart);
+    ip1 = ip0 + 1;
+    {   U32 const curr = (U32)(ip0 - base);
+        U32 const windowLow = ZSTD_getLowestPrefixIndex(ms, curr, cParams->windowLog);
+        U32 const maxRep = curr - windowLow;
+        if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
+        if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
+    }
+
+    /* Main Search Loop */
+#ifdef __INTEL_COMPILER
+    /* From intel 'The vector pragma indicates that the loop should be
+     * vectorized if it is legal to do so'. Can be used together with
+     * #pragma ivdep (but have opted to exclude that because intel
+     * warns against using it).*/
+    #pragma vector always
+#endif
+    while (ip1 < ilimit) {   /* < instead of <=, because check at ip0+2 */
+        size_t mLength;
+        BYTE const* ip2 = ip0 + 2;
+        size_t const h0 = ZSTD_hashPtr(ip0, hlog, mls);
+        U32 const val0 = MEM_read32(ip0);
+        size_t const h1 = ZSTD_hashPtr(ip1, hlog, mls);
+        U32 const val1 = MEM_read32(ip1);
+        U32 const current0 = (U32)(ip0-base);
+        U32 const current1 = (U32)(ip1-base);
+        U32 const matchIndex0 = hashTable[h0];
+        U32 const matchIndex1 = hashTable[h1];
+        BYTE const* repMatch = ip2 - offset_1;
+        const BYTE* match0 = base + matchIndex0;
+        const BYTE* match1 = base + matchIndex1;
+        U32 offcode;
+
+#if defined(__aarch64__)
+        PREFETCH_L1(ip0+256);
+#endif
+
+        hashTable[h0] = current0;   /* update hash table */
+        hashTable[h1] = current1;   /* update hash table */
+
+        assert(ip0 + 1 == ip1);
+
+        if ((offset_1 > 0) & (MEM_read32(repMatch) == MEM_read32(ip2))) {
+            mLength = (ip2[-1] == repMatch[-1]) ? 1 : 0;
+            ip0 = ip2 - mLength;
+            match0 = repMatch - mLength;
+            mLength += 4;
+            offcode = 0;
+            goto _match;
+        }
+        if ((matchIndex0 > prefixStartIndex) && MEM_read32(match0) == val0) {
+            /* found a regular match */
+            goto _offset;
+        }
+        if ((matchIndex1 > prefixStartIndex) && MEM_read32(match1) == val1) {
+            /* found a regular match after one literal */
+            ip0 = ip1;
+            match0 = match1;
+            goto _offset;
+        }
+        {   size_t const step = ((size_t)(ip0-anchor) >> (kSearchStrength - 1)) + stepSize;
+            assert(step >= 2);
+            ip0 += step;
+            ip1 += step;
+            continue;
+        }
+_offset: /* Requires: ip0, match0 */
+        /* Compute the offset code */
+        offset_2 = offset_1;
+        offset_1 = (U32)(ip0-match0);
+        offcode = offset_1 + ZSTD_REP_MOVE;
+        mLength = 4;
+        /* Count the backwards match length */
+        while (((ip0>anchor) & (match0>prefixStart))
+             && (ip0[-1] == match0[-1])) { ip0--; match0--; mLength++; } /* catch up */
+
+_match: /* Requires: ip0, match0, offcode */
+        /* Count the forward length */
+        mLength += ZSTD_count(ip0+mLength, match0+mLength, iend);
+        ZSTD_storeSeq(seqStore, (size_t)(ip0-anchor), anchor, iend, offcode, mLength-MINMATCH);
+        /* match found */
+        ip0 += mLength;
+        anchor = ip0;
+
+        if (ip0 <= ilimit) {
+            /* Fill Table */
+            assert(base+current0+2 > istart);  /* check base overflow */
+            hashTable[ZSTD_hashPtr(base+current0+2, hlog, mls)] = current0+2;  /* here because current+2 could be > iend-8 */
+            hashTable[ZSTD_hashPtr(ip0-2, hlog, mls)] = (U32)(ip0-2-base);
+
+            if (offset_2 > 0) { /* offset_2==0 means offset_2 is invalidated */
+                while ( (ip0 <= ilimit) && (MEM_read32(ip0) == MEM_read32(ip0 - offset_2)) ) {
+                    /* store sequence */
+                    size_t const rLength = ZSTD_count(ip0+4, ip0+4-offset_2, iend) + 4;
+                    { U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; } /* swap offset_2 <=> offset_1 */
+                    hashTable[ZSTD_hashPtr(ip0, hlog, mls)] = (U32)(ip0-base);
+                    ip0 += rLength;
+                    ZSTD_storeSeq(seqStore, 0 /*litLen*/, anchor, iend, 0 /*offCode*/, rLength-MINMATCH);
+                    anchor = ip0;
+                    continue;   /* faster when present (confirmed on gcc-8) ... (?) */
+        }   }   }
+        ip1 = ip0 + 1;
+    }
+
+    /* save reps for next block */
+    rep[0] = offset_1 ? offset_1 : offsetSaved;
+    rep[1] = offset_2 ? offset_2 : offsetSaved;
+
+    /* Return the last literals size */
+    return (size_t)(iend - anchor);
+}
+
+
+size_t ZSTD_compressBlock_fast(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    U32 const mls = ms->cParams.minMatch;
+    assert(ms->dictMatchState == NULL);
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 4);
+    case 5 :
+        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5);
+    case 6 :
+        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6);
+    case 7 :
+        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7);
+    }
+}
+
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_fast_dictMatchState_generic(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize, U32 const mls)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32* const hashTable = ms->hashTable;
+    U32 const hlog = cParams->hashLog;
+    /* support stepSize of 0 */
+    U32 const stepSize = cParams->targetLength + !(cParams->targetLength);
+    const BYTE* const base = ms->window.base;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const U32   prefixStartIndex = ms->window.dictLimit;
+    const BYTE* const prefixStart = base + prefixStartIndex;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - HASH_READ_SIZE;
+    U32 offset_1=rep[0], offset_2=rep[1];
+    U32 offsetSaved = 0;
+
+    const ZSTD_matchState_t* const dms = ms->dictMatchState;
+    const ZSTD_compressionParameters* const dictCParams = &dms->cParams ;
+    const U32* const dictHashTable = dms->hashTable;
+    const U32 dictStartIndex       = dms->window.dictLimit;
+    const BYTE* const dictBase     = dms->window.base;
+    const BYTE* const dictStart    = dictBase + dictStartIndex;
+    const BYTE* const dictEnd      = dms->window.nextSrc;
+    const U32 dictIndexDelta       = prefixStartIndex - (U32)(dictEnd - dictBase);
+    const U32 dictAndPrefixLength  = (U32)(ip - prefixStart + dictEnd - dictStart);
+    const U32 dictHLog             = dictCParams->hashLog;
+
+    /* if a dictionary is still attached, it necessarily means that
+     * it is within window size. So we just check it. */
+    const U32 maxDistance = 1U << cParams->windowLog;
+    const U32 endIndex = (U32)((size_t)(ip - base) + srcSize);
+    assert(endIndex - prefixStartIndex <= maxDistance);
+    (void)maxDistance; (void)endIndex;   /* these variables are not used when assert() is disabled */
+
+    /* ensure there will be no underflow
+     * when translating a dict index into a local index */
+    assert(prefixStartIndex >= (U32)(dictEnd - dictBase));
+
+    /* init */
+    DEBUGLOG(5, "ZSTD_compressBlock_fast_dictMatchState_generic");
+    ip += (dictAndPrefixLength == 0);
+    /* dictMatchState repCode checks don't currently handle repCode == 0
+     * disabling. */
+    assert(offset_1 <= dictAndPrefixLength);
+    assert(offset_2 <= dictAndPrefixLength);
+
+    /* Main Search Loop */
+    while (ip < ilimit) {   /* < instead of <=, because repcode check at (ip+1) */
+        size_t mLength;
+        size_t const h = ZSTD_hashPtr(ip, hlog, mls);
+        U32 const curr = (U32)(ip-base);
+        U32 const matchIndex = hashTable[h];
+        const BYTE* match = base + matchIndex;
+        const U32 repIndex = curr + 1 - offset_1;
+        const BYTE* repMatch = (repIndex < prefixStartIndex) ?
+                               dictBase + (repIndex - dictIndexDelta) :
+                               base + repIndex;
+        hashTable[h] = curr;   /* update hash table */
+
+        if ( ((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex isn't overlapping dict + prefix */
+          && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
+            const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
+            mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
+            ip++;
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength-MINMATCH);
+        } else if ( (matchIndex <= prefixStartIndex) ) {
+            size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);
+            U32 const dictMatchIndex = dictHashTable[dictHash];
+            const BYTE* dictMatch = dictBase + dictMatchIndex;
+            if (dictMatchIndex <= dictStartIndex ||
+                MEM_read32(dictMatch) != MEM_read32(ip)) {
+                assert(stepSize >= 1);
+                ip += ((ip-anchor) >> kSearchStrength) + stepSize;
+                continue;
+            } else {
+                /* found a dict match */
+                U32 const offset = (U32)(curr-dictMatchIndex-dictIndexDelta);
+                mLength = ZSTD_count_2segments(ip+4, dictMatch+4, iend, dictEnd, prefixStart) + 4;
+                while (((ip>anchor) & (dictMatch>dictStart))
+                     && (ip[-1] == dictMatch[-1])) {
+                    ip--; dictMatch--; mLength++;
+                } /* catch up */
+                offset_2 = offset_1;
+                offset_1 = offset;
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+            }
+        } else if (MEM_read32(match) != MEM_read32(ip)) {
+            /* it's not a match, and we're not going to check the dictionary */
+            assert(stepSize >= 1);
+            ip += ((ip-anchor) >> kSearchStrength) + stepSize;
+            continue;
+        } else {
+            /* found a regular match */
+            U32 const offset = (U32)(ip-match);
+            mLength = ZSTD_count(ip+4, match+4, iend) + 4;
+            while (((ip>anchor) & (match>prefixStart))
+                 && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
+            offset_2 = offset_1;
+            offset_1 = offset;
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+        }
+
+        /* match found */
+        ip += mLength;
+        anchor = ip;
+
+        if (ip <= ilimit) {
+            /* Fill Table */
+            assert(base+curr+2 > istart);  /* check base overflow */
+            hashTable[ZSTD_hashPtr(base+curr+2, hlog, mls)] = curr+2;  /* here because curr+2 could be > iend-8 */
+            hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);
+
+            /* check immediate repcode */
+            while (ip <= ilimit) {
+                U32 const current2 = (U32)(ip-base);
+                U32 const repIndex2 = current2 - offset_2;
+                const BYTE* repMatch2 = repIndex2 < prefixStartIndex ?
+                        dictBase - dictIndexDelta + repIndex2 :
+                        base + repIndex2;
+                if ( ((U32)((prefixStartIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
+                   && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
+                    const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
+                    size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
+                    U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
+                    ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, repLength2-MINMATCH);
+                    hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
+                    ip += repLength2;
+                    anchor = ip;
+                    continue;
+                }
+                break;
+            }
+        }
+    }
+
+    /* save reps for next block */
+    rep[0] = offset_1 ? offset_1 : offsetSaved;
+    rep[1] = offset_2 ? offset_2 : offsetSaved;
+
+    /* Return the last literals size */
+    return (size_t)(iend - anchor);
+}
+
+size_t ZSTD_compressBlock_fast_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    U32 const mls = ms->cParams.minMatch;
+    assert(ms->dictMatchState != NULL);
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 4);
+    case 5 :
+        return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 5);
+    case 6 :
+        return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 6);
+    case 7 :
+        return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 7);
+    }
+}
+
+
+static size_t ZSTD_compressBlock_fast_extDict_generic(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize, U32 const mls)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32* const hashTable = ms->hashTable;
+    U32 const hlog = cParams->hashLog;
+    /* support stepSize of 0 */
+    U32 const stepSize = cParams->targetLength + !(cParams->targetLength);
+    const BYTE* const base = ms->window.base;
+    const BYTE* const dictBase = ms->window.dictBase;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const U32   endIndex = (U32)((size_t)(istart - base) + srcSize);
+    const U32   lowLimit = ZSTD_getLowestMatchIndex(ms, endIndex, cParams->windowLog);
+    const U32   dictStartIndex = lowLimit;
+    const BYTE* const dictStart = dictBase + dictStartIndex;
+    const U32   dictLimit = ms->window.dictLimit;
+    const U32   prefixStartIndex = dictLimit < lowLimit ? lowLimit : dictLimit;
+    const BYTE* const prefixStart = base + prefixStartIndex;
+    const BYTE* const dictEnd = dictBase + prefixStartIndex;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    U32 offset_1=rep[0], offset_2=rep[1];
+
+    DEBUGLOG(5, "ZSTD_compressBlock_fast_extDict_generic (offset_1=%u)", offset_1);
+
+    /* switch to "regular" variant if extDict is invalidated due to maxDistance */
+    if (prefixStartIndex == dictStartIndex)
+        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, mls);
+
+    /* Search Loop */
+    while (ip < ilimit) {  /* < instead of <=, because (ip+1) */
+        const size_t h = ZSTD_hashPtr(ip, hlog, mls);
+        const U32    matchIndex = hashTable[h];
+        const BYTE* const matchBase = matchIndex < prefixStartIndex ? dictBase : base;
+        const BYTE*  match = matchBase + matchIndex;
+        const U32    curr = (U32)(ip-base);
+        const U32    repIndex = curr + 1 - offset_1;
+        const BYTE* const repBase = repIndex < prefixStartIndex ? dictBase : base;
+        const BYTE* const repMatch = repBase + repIndex;
+        hashTable[h] = curr;   /* update hash table */
+        DEBUGLOG(7, "offset_1 = %u , curr = %u", offset_1, curr);
+        assert(offset_1 <= curr +1);   /* check repIndex */
+
+        if ( (((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > dictStartIndex))
+           && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
+            const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
+            size_t const rLength = ZSTD_count_2segments(ip+1 +4, repMatch +4, iend, repMatchEnd, prefixStart) + 4;
+            ip++;
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, rLength-MINMATCH);
+            ip += rLength;
+            anchor = ip;
+        } else {
+            if ( (matchIndex < dictStartIndex) ||
+                 (MEM_read32(match) != MEM_read32(ip)) ) {
+                assert(stepSize >= 1);
+                ip += ((ip-anchor) >> kSearchStrength) + stepSize;
+                continue;
+            }
+            {   const BYTE* const matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend;
+                const BYTE* const lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart;
+                U32 const offset = curr - matchIndex;
+                size_t mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4;
+                while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; }   /* catch up */
+                offset_2 = offset_1; offset_1 = offset;  /* update offset history */
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+                ip += mLength;
+                anchor = ip;
+        }   }
+
+        if (ip <= ilimit) {
+            /* Fill Table */
+            hashTable[ZSTD_hashPtr(base+curr+2, hlog, mls)] = curr+2;
+            hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);
+            /* check immediate repcode */
+            while (ip <= ilimit) {
+                U32 const current2 = (U32)(ip-base);
+                U32 const repIndex2 = current2 - offset_2;
+                const BYTE* const repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;
+                if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) & (repIndex2 > dictStartIndex))  /* intentional overflow */
+                   && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
+                    const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
+                    size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
+                    { U32 const tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; }  /* swap offset_2 <=> offset_1 */
+                    ZSTD_storeSeq(seqStore, 0 /*litlen*/, anchor, iend, 0 /*offcode*/, repLength2-MINMATCH);
+                    hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
+                    ip += repLength2;
+                    anchor = ip;
+                    continue;
+                }
+                break;
+    }   }   }
+
+    /* save reps for next block */
+    rep[0] = offset_1;
+    rep[1] = offset_2;
+
+    /* Return the last literals size */
+    return (size_t)(iend - anchor);
+}
+
+
+size_t ZSTD_compressBlock_fast_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    U32 const mls = ms->cParams.minMatch;
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 4);
+    case 5 :
+        return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 5);
+    case 6 :
+        return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 6);
+    case 7 :
+        return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 7);
+    }
+}
diff --git a/lib/zstd/compress/zstd_fast.h b/lib/zstd/compress/zstd_fast.h
new file mode 100644
index 0000000..fddc2f5
--- /dev/null
+++ b/lib/zstd/compress/zstd_fast.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_FAST_H
+#define ZSTD_FAST_H
+
+
+#include "../common/mem.h"      /* U32 */
+#include "zstd_compress_internal.h"
+
+void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
+                        void const* end, ZSTD_dictTableLoadMethod_e dtlm);
+size_t ZSTD_compressBlock_fast(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_fast_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_fast_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+
+
+#endif /* ZSTD_FAST_H */
diff --git a/lib/zstd/compress/zstd_lazy.c b/lib/zstd/compress/zstd_lazy.c
new file mode 100644
index 0000000..fb54d4e
--- /dev/null
+++ b/lib/zstd/compress/zstd_lazy.c
@@ -0,0 +1,1414 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#include "zstd_compress_internal.h"
+#include "zstd_lazy.h"
+
+
+/*-*************************************
+*  Binary Tree search
+***************************************/
+
+static void
+ZSTD_updateDUBT(ZSTD_matchState_t* ms,
+                const BYTE* ip, const BYTE* iend,
+                U32 mls)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32* const hashTable = ms->hashTable;
+    U32  const hashLog = cParams->hashLog;
+
+    U32* const bt = ms->chainTable;
+    U32  const btLog  = cParams->chainLog - 1;
+    U32  const btMask = (1 << btLog) - 1;
+
+    const BYTE* const base = ms->window.base;
+    U32 const target = (U32)(ip - base);
+    U32 idx = ms->nextToUpdate;
+
+    if (idx != target)
+        DEBUGLOG(7, "ZSTD_updateDUBT, from %u to %u (dictLimit:%u)",
+                    idx, target, ms->window.dictLimit);
+    assert(ip + 8 <= iend);   /* condition for ZSTD_hashPtr */
+    (void)iend;
+
+    assert(idx >= ms->window.dictLimit);   /* condition for valid base+idx */
+    for ( ; idx < target ; idx++) {
+        size_t const h  = ZSTD_hashPtr(base + idx, hashLog, mls);   /* assumption : ip + 8 <= iend */
+        U32    const matchIndex = hashTable[h];
+
+        U32*   const nextCandidatePtr = bt + 2*(idx&btMask);
+        U32*   const sortMarkPtr  = nextCandidatePtr + 1;
+
+        DEBUGLOG(8, "ZSTD_updateDUBT: insert %u", idx);
+        hashTable[h] = idx;   /* Update Hash Table */
+        *nextCandidatePtr = matchIndex;   /* update BT like a chain */
+        *sortMarkPtr = ZSTD_DUBT_UNSORTED_MARK;
+    }
+    ms->nextToUpdate = target;
+}
+
+
+/* ZSTD_insertDUBT1() :
+ *  sort one already inserted but unsorted position
+ *  assumption : curr >= btlow == (curr - btmask)
+ *  doesn't fail */
+static void
+ZSTD_insertDUBT1(ZSTD_matchState_t* ms,
+                 U32 curr, const BYTE* inputEnd,
+                 U32 nbCompares, U32 btLow,
+                 const ZSTD_dictMode_e dictMode)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32* const bt = ms->chainTable;
+    U32  const btLog  = cParams->chainLog - 1;
+    U32  const btMask = (1 << btLog) - 1;
+    size_t commonLengthSmaller=0, commonLengthLarger=0;
+    const BYTE* const base = ms->window.base;
+    const BYTE* const dictBase = ms->window.dictBase;
+    const U32 dictLimit = ms->window.dictLimit;
+    const BYTE* const ip = (curr>=dictLimit) ? base + curr : dictBase + curr;
+    const BYTE* const iend = (curr>=dictLimit) ? inputEnd : dictBase + dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    const BYTE* match;
+    U32* smallerPtr = bt + 2*(curr&btMask);
+    U32* largerPtr  = smallerPtr + 1;
+    U32 matchIndex = *smallerPtr;   /* this candidate is unsorted : next sorted candidate is reached through *smallerPtr, while *largerPtr contains previous unsorted candidate (which is already saved and can be overwritten) */
+    U32 dummy32;   /* to be nullified at the end */
+    U32 const windowValid = ms->window.lowLimit;
+    U32 const maxDistance = 1U << cParams->windowLog;
+    U32 const windowLow = (curr - windowValid > maxDistance) ? curr - maxDistance : windowValid;
+
+
+    DEBUGLOG(8, "ZSTD_insertDUBT1(%u) (dictLimit=%u, lowLimit=%u)",
+                curr, dictLimit, windowLow);
+    assert(curr >= btLow);
+    assert(ip < iend);   /* condition for ZSTD_count */
+
+    for (; nbCompares && (matchIndex > windowLow); --nbCompares) {
+        U32* const nextPtr = bt + 2*(matchIndex & btMask);
+        size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
+        assert(matchIndex < curr);
+        /* note : all candidates are now supposed sorted,
+         * but it's still possible to have nextPtr[1] == ZSTD_DUBT_UNSORTED_MARK
+         * when a real index has the same value as ZSTD_DUBT_UNSORTED_MARK */
+
+        if ( (dictMode != ZSTD_extDict)
+          || (matchIndex+matchLength >= dictLimit)  /* both in current segment*/
+          || (curr < dictLimit) /* both in extDict */) {
+            const BYTE* const mBase = ( (dictMode != ZSTD_extDict)
+                                     || (matchIndex+matchLength >= dictLimit)) ?
+                                        base : dictBase;
+            assert( (matchIndex+matchLength >= dictLimit)   /* might be wrong if extDict is incorrectly set to 0 */
+                 || (curr < dictLimit) );
+            match = mBase + matchIndex;
+            matchLength += ZSTD_count(ip+matchLength, match+matchLength, iend);
+        } else {
+            match = dictBase + matchIndex;
+            matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
+            if (matchIndex+matchLength >= dictLimit)
+                match = base + matchIndex;   /* preparation for next read of match[matchLength] */
+        }
+
+        DEBUGLOG(8, "ZSTD_insertDUBT1: comparing %u with %u : found %u common bytes ",
+                    curr, matchIndex, (U32)matchLength);
+
+        if (ip+matchLength == iend) {   /* equal : no way to know if inf or sup */
+            break;   /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt tree */
+        }
+
+        if (match[matchLength] < ip[matchLength]) {  /* necessarily within buffer */
+            /* match is smaller than current */
+            *smallerPtr = matchIndex;             /* update smaller idx */
+            commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
+            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop searching */
+            DEBUGLOG(8, "ZSTD_insertDUBT1: %u (>btLow=%u) is smaller : next => %u",
+                        matchIndex, btLow, nextPtr[1]);
+            smallerPtr = nextPtr+1;               /* new "candidate" => larger than match, which was smaller than target */
+            matchIndex = nextPtr[1];              /* new matchIndex, larger than previous and closer to current */
+        } else {
+            /* match is larger than current */
+            *largerPtr = matchIndex;
+            commonLengthLarger = matchLength;
+            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop searching */
+            DEBUGLOG(8, "ZSTD_insertDUBT1: %u (>btLow=%u) is larger => %u",
+                        matchIndex, btLow, nextPtr[0]);
+            largerPtr = nextPtr;
+            matchIndex = nextPtr[0];
+    }   }
+
+    *smallerPtr = *largerPtr = 0;
+}
+
+
+static size_t
+ZSTD_DUBT_findBetterDictMatch (
+        ZSTD_matchState_t* ms,
+        const BYTE* const ip, const BYTE* const iend,
+        size_t* offsetPtr,
+        size_t bestLength,
+        U32 nbCompares,
+        U32 const mls,
+        const ZSTD_dictMode_e dictMode)
+{
+    const ZSTD_matchState_t * const dms = ms->dictMatchState;
+    const ZSTD_compressionParameters* const dmsCParams = &dms->cParams;
+    const U32 * const dictHashTable = dms->hashTable;
+    U32         const hashLog = dmsCParams->hashLog;
+    size_t      const h  = ZSTD_hashPtr(ip, hashLog, mls);
+    U32               dictMatchIndex = dictHashTable[h];
+
+    const BYTE* const base = ms->window.base;
+    const BYTE* const prefixStart = base + ms->window.dictLimit;
+    U32         const curr = (U32)(ip-base);
+    const BYTE* const dictBase = dms->window.base;
+    const BYTE* const dictEnd = dms->window.nextSrc;
+    U32         const dictHighLimit = (U32)(dms->window.nextSrc - dms->window.base);
+    U32         const dictLowLimit = dms->window.lowLimit;
+    U32         const dictIndexDelta = ms->window.lowLimit - dictHighLimit;
+
+    U32*        const dictBt = dms->chainTable;
+    U32         const btLog  = dmsCParams->chainLog - 1;
+    U32         const btMask = (1 << btLog) - 1;
+    U32         const btLow = (btMask >= dictHighLimit - dictLowLimit) ? dictLowLimit : dictHighLimit - btMask;
+
+    size_t commonLengthSmaller=0, commonLengthLarger=0;
+
+    (void)dictMode;
+    assert(dictMode == ZSTD_dictMatchState);
+
+    for (; nbCompares && (dictMatchIndex > dictLowLimit); --nbCompares) {
+        U32* const nextPtr = dictBt + 2*(dictMatchIndex & btMask);
+        size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
+        const BYTE* match = dictBase + dictMatchIndex;
+        matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
+        if (dictMatchIndex+matchLength >= dictHighLimit)
+            match = base + dictMatchIndex + dictIndexDelta;   /* to prepare for next usage of match[matchLength] */
+
+        if (matchLength > bestLength) {
+            U32 matchIndex = dictMatchIndex + dictIndexDelta;
+            if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(curr-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) ) {
+                DEBUGLOG(9, "ZSTD_DUBT_findBetterDictMatch(%u) : found better match length %u -> %u and offsetCode %u -> %u (dictMatchIndex %u, matchIndex %u)",
+                    curr, (U32)bestLength, (U32)matchLength, (U32)*offsetPtr, ZSTD_REP_MOVE + curr - matchIndex, dictMatchIndex, matchIndex);
+                bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + curr - matchIndex;
+            }
+            if (ip+matchLength == iend) {   /* reached end of input : ip[matchLength] is not valid, no way to know if it's larger or smaller than match */
+                break;   /* drop, to guarantee consistency (miss a little bit of compression) */
+            }
+        }
+
+        if (match[matchLength] < ip[matchLength]) {
+            if (dictMatchIndex <= btLow) { break; }   /* beyond tree size, stop the search */
+            commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
+            dictMatchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
+        } else {
+            /* match is larger than current */
+            if (dictMatchIndex <= btLow) { break; }   /* beyond tree size, stop the search */
+            commonLengthLarger = matchLength;
+            dictMatchIndex = nextPtr[0];
+        }
+    }
+
+    if (bestLength >= MINMATCH) {
+        U32 const mIndex = curr - ((U32)*offsetPtr - ZSTD_REP_MOVE); (void)mIndex;
+        DEBUGLOG(8, "ZSTD_DUBT_findBetterDictMatch(%u) : found match of length %u and offsetCode %u (pos %u)",
+                    curr, (U32)bestLength, (U32)*offsetPtr, mIndex);
+    }
+    return bestLength;
+
+}
+
+
+static size_t
+ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms,
+                        const BYTE* const ip, const BYTE* const iend,
+                        size_t* offsetPtr,
+                        U32 const mls,
+                        const ZSTD_dictMode_e dictMode)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32*   const hashTable = ms->hashTable;
+    U32    const hashLog = cParams->hashLog;
+    size_t const h  = ZSTD_hashPtr(ip, hashLog, mls);
+    U32          matchIndex  = hashTable[h];
+
+    const BYTE* const base = ms->window.base;
+    U32    const curr = (U32)(ip-base);
+    U32    const windowLow = ZSTD_getLowestMatchIndex(ms, curr, cParams->windowLog);
+
+    U32*   const bt = ms->chainTable;
+    U32    const btLog  = cParams->chainLog - 1;
+    U32    const btMask = (1 << btLog) - 1;
+    U32    const btLow = (btMask >= curr) ? 0 : curr - btMask;
+    U32    const unsortLimit = MAX(btLow, windowLow);
+
+    U32*         nextCandidate = bt + 2*(matchIndex&btMask);
+    U32*         unsortedMark = bt + 2*(matchIndex&btMask) + 1;
+    U32          nbCompares = 1U << cParams->searchLog;
+    U32          nbCandidates = nbCompares;
+    U32          previousCandidate = 0;
+
+    DEBUGLOG(7, "ZSTD_DUBT_findBestMatch (%u) ", curr);
+    assert(ip <= iend-8);   /* required for h calculation */
+    assert(dictMode != ZSTD_dedicatedDictSearch);
+
+    /* reach end of unsorted candidates list */
+    while ( (matchIndex > unsortLimit)
+         && (*unsortedMark == ZSTD_DUBT_UNSORTED_MARK)
+         && (nbCandidates > 1) ) {
+        DEBUGLOG(8, "ZSTD_DUBT_findBestMatch: candidate %u is unsorted",
+                    matchIndex);
+        *unsortedMark = previousCandidate;  /* the unsortedMark becomes a reversed chain, to move up back to original position */
+        previousCandidate = matchIndex;
+        matchIndex = *nextCandidate;
+        nextCandidate = bt + 2*(matchIndex&btMask);
+        unsortedMark = bt + 2*(matchIndex&btMask) + 1;
+        nbCandidates --;
+    }
+
+    /* nullify last candidate if it's still unsorted
+     * simplification, detrimental to compression ratio, beneficial for speed */
+    if ( (matchIndex > unsortLimit)
+      && (*unsortedMark==ZSTD_DUBT_UNSORTED_MARK) ) {
+        DEBUGLOG(7, "ZSTD_DUBT_findBestMatch: nullify last unsorted candidate %u",
+                    matchIndex);
+        *nextCandidate = *unsortedMark = 0;
+    }
+
+    /* batch sort stacked candidates */
+    matchIndex = previousCandidate;
+    while (matchIndex) {  /* will end on matchIndex == 0 */
+        U32* const nextCandidateIdxPtr = bt + 2*(matchIndex&btMask) + 1;
+        U32 const nextCandidateIdx = *nextCandidateIdxPtr;
+        ZSTD_insertDUBT1(ms, matchIndex, iend,
+                         nbCandidates, unsortLimit, dictMode);
+        matchIndex = nextCandidateIdx;
+        nbCandidates++;
+    }
+
+    /* find longest match */
+    {   size_t commonLengthSmaller = 0, commonLengthLarger = 0;
+        const BYTE* const dictBase = ms->window.dictBase;
+        const U32 dictLimit = ms->window.dictLimit;
+        const BYTE* const dictEnd = dictBase + dictLimit;
+        const BYTE* const prefixStart = base + dictLimit;
+        U32* smallerPtr = bt + 2*(curr&btMask);
+        U32* largerPtr  = bt + 2*(curr&btMask) + 1;
+        U32 matchEndIdx = curr + 8 + 1;
+        U32 dummy32;   /* to be nullified at the end */
+        size_t bestLength = 0;
+
+        matchIndex  = hashTable[h];
+        hashTable[h] = curr;   /* Update Hash Table */
+
+        for (; nbCompares && (matchIndex > windowLow); --nbCompares) {
+            U32* const nextPtr = bt + 2*(matchIndex & btMask);
+            size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
+            const BYTE* match;
+
+            if ((dictMode != ZSTD_extDict) || (matchIndex+matchLength >= dictLimit)) {
+                match = base + matchIndex;
+                matchLength += ZSTD_count(ip+matchLength, match+matchLength, iend);
+            } else {
+                match = dictBase + matchIndex;
+                matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
+                if (matchIndex+matchLength >= dictLimit)
+                    match = base + matchIndex;   /* to prepare for next usage of match[matchLength] */
+            }
+
+            if (matchLength > bestLength) {
+                if (matchLength > matchEndIdx - matchIndex)
+                    matchEndIdx = matchIndex + (U32)matchLength;
+                if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(curr-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) )
+                    bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + curr - matchIndex;
+                if (ip+matchLength == iend) {   /* equal : no way to know if inf or sup */
+                    if (dictMode == ZSTD_dictMatchState) {
+                        nbCompares = 0; /* in addition to avoiding checking any
+                                         * further in this loop, make sure we
+                                         * skip checking in the dictionary. */
+                    }
+                    break;   /* drop, to guarantee consistency (miss a little bit of compression) */
+                }
+            }
+
+            if (match[matchLength] < ip[matchLength]) {
+                /* match is smaller than current */
+                *smallerPtr = matchIndex;             /* update smaller idx */
+                commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
+                if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+                smallerPtr = nextPtr+1;               /* new "smaller" => larger of match */
+                matchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
+            } else {
+                /* match is larger than current */
+                *largerPtr = matchIndex;
+                commonLengthLarger = matchLength;
+                if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+                largerPtr = nextPtr;
+                matchIndex = nextPtr[0];
+        }   }
+
+        *smallerPtr = *largerPtr = 0;
+
+        assert(nbCompares <= (1U << ZSTD_SEARCHLOG_MAX)); /* Check we haven't underflowed. */
+        if (dictMode == ZSTD_dictMatchState && nbCompares) {
+            bestLength = ZSTD_DUBT_findBetterDictMatch(
+                    ms, ip, iend,
+                    offsetPtr, bestLength, nbCompares,
+                    mls, dictMode);
+        }
+
+        assert(matchEndIdx > curr+8); /* ensure nextToUpdate is increased */
+        ms->nextToUpdate = matchEndIdx - 8;   /* skip repetitive patterns */
+        if (bestLength >= MINMATCH) {
+            U32 const mIndex = curr - ((U32)*offsetPtr - ZSTD_REP_MOVE); (void)mIndex;
+            DEBUGLOG(8, "ZSTD_DUBT_findBestMatch(%u) : found match of length %u and offsetCode %u (pos %u)",
+                        curr, (U32)bestLength, (U32)*offsetPtr, mIndex);
+        }
+        return bestLength;
+    }
+}
+
+
+/* ZSTD_BtFindBestMatch() : Tree updater, providing best match */
+FORCE_INLINE_TEMPLATE size_t
+ZSTD_BtFindBestMatch( ZSTD_matchState_t* ms,
+                const BYTE* const ip, const BYTE* const iLimit,
+                      size_t* offsetPtr,
+                const U32 mls /* template */,
+                const ZSTD_dictMode_e dictMode)
+{
+    DEBUGLOG(7, "ZSTD_BtFindBestMatch");
+    if (ip < ms->window.base + ms->nextToUpdate) return 0;   /* skipped area */
+    ZSTD_updateDUBT(ms, ip, iLimit, mls);
+    return ZSTD_DUBT_findBestMatch(ms, ip, iLimit, offsetPtr, mls, dictMode);
+}
+
+
+static size_t
+ZSTD_BtFindBestMatch_selectMLS (  ZSTD_matchState_t* ms,
+                            const BYTE* ip, const BYTE* const iLimit,
+                                  size_t* offsetPtr)
+{
+    switch(ms->cParams.minMatch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 4, ZSTD_noDict);
+    case 5 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 5, ZSTD_noDict);
+    case 7 :
+    case 6 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 6, ZSTD_noDict);
+    }
+}
+
+
+static size_t ZSTD_BtFindBestMatch_dictMatchState_selectMLS (
+                        ZSTD_matchState_t* ms,
+                        const BYTE* ip, const BYTE* const iLimit,
+                        size_t* offsetPtr)
+{
+    switch(ms->cParams.minMatch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 4, ZSTD_dictMatchState);
+    case 5 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 5, ZSTD_dictMatchState);
+    case 7 :
+    case 6 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 6, ZSTD_dictMatchState);
+    }
+}
+
+
+static size_t ZSTD_BtFindBestMatch_extDict_selectMLS (
+                        ZSTD_matchState_t* ms,
+                        const BYTE* ip, const BYTE* const iLimit,
+                        size_t* offsetPtr)
+{
+    switch(ms->cParams.minMatch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 4, ZSTD_extDict);
+    case 5 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 5, ZSTD_extDict);
+    case 7 :
+    case 6 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 6, ZSTD_extDict);
+    }
+}
+
+
+
+/* *********************************
+*  Hash Chain
+***********************************/
+#define NEXT_IN_CHAIN(d, mask)   chainTable[(d) & (mask)]
+
+/* Update chains up to ip (excluded)
+   Assumption : always within prefix (i.e. not within extDict) */
+FORCE_INLINE_TEMPLATE U32 ZSTD_insertAndFindFirstIndex_internal(
+                        ZSTD_matchState_t* ms,
+                        const ZSTD_compressionParameters* const cParams,
+                        const BYTE* ip, U32 const mls)
+{
+    U32* const hashTable  = ms->hashTable;
+    const U32 hashLog = cParams->hashLog;
+    U32* const chainTable = ms->chainTable;
+    const U32 chainMask = (1 << cParams->chainLog) - 1;
+    const BYTE* const base = ms->window.base;
+    const U32 target = (U32)(ip - base);
+    U32 idx = ms->nextToUpdate;
+
+    while(idx < target) { /* catch up */
+        size_t const h = ZSTD_hashPtr(base+idx, hashLog, mls);
+        NEXT_IN_CHAIN(idx, chainMask) = hashTable[h];
+        hashTable[h] = idx;
+        idx++;
+    }
+
+    ms->nextToUpdate = target;
+    return hashTable[ZSTD_hashPtr(ip, hashLog, mls)];
+}
+
+U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip) {
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    return ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, ms->cParams.minMatch);
+}
+
+void ZSTD_dedicatedDictSearch_lazy_loadDictionary(ZSTD_matchState_t* ms, const BYTE* const ip)
+{
+    const BYTE* const base = ms->window.base;
+    U32 const target = (U32)(ip - base);
+    U32* const hashTable = ms->hashTable;
+    U32* const chainTable = ms->chainTable;
+    U32 const chainSize = 1 << ms->cParams.chainLog;
+    U32 idx = ms->nextToUpdate;
+    U32 const minChain = chainSize < target ? target - chainSize : idx;
+    U32 const bucketSize = 1 << ZSTD_LAZY_DDSS_BUCKET_LOG;
+    U32 const cacheSize = bucketSize - 1;
+    U32 const chainAttempts = (1 << ms->cParams.searchLog) - cacheSize;
+    U32 const chainLimit = chainAttempts > 255 ? 255 : chainAttempts;
+
+    /* We know the hashtable is oversized by a factor of `bucketSize`.
+     * We are going to temporarily pretend `bucketSize == 1`, keeping only a
+     * single entry. We will use the rest of the space to construct a temporary
+     * chaintable.
+     */
+    U32 const hashLog = ms->cParams.hashLog - ZSTD_LAZY_DDSS_BUCKET_LOG;
+    U32* const tmpHashTable = hashTable;
+    U32* const tmpChainTable = hashTable + ((size_t)1 << hashLog);
+    U32 const tmpChainSize = ((1 << ZSTD_LAZY_DDSS_BUCKET_LOG) - 1) << hashLog;
+    U32 const tmpMinChain = tmpChainSize < target ? target - tmpChainSize : idx;
+
+    U32 hashIdx;
+
+    assert(ms->cParams.chainLog <= 24);
+    assert(ms->cParams.hashLog >= ms->cParams.chainLog);
+    assert(idx != 0);
+    assert(tmpMinChain <= minChain);
+
+    /* fill conventional hash table and conventional chain table */
+    for ( ; idx < target; idx++) {
+        U32 const h = (U32)ZSTD_hashPtr(base + idx, hashLog, ms->cParams.minMatch);
+        if (idx >= tmpMinChain) {
+            tmpChainTable[idx - tmpMinChain] = hashTable[h];
+        }
+        tmpHashTable[h] = idx;
+    }
+
+    /* sort chains into ddss chain table */
+    {
+        U32 chainPos = 0;
+        for (hashIdx = 0; hashIdx < (1U << hashLog); hashIdx++) {
+            U32 count;
+            U32 countBeyondMinChain = 0;
+            U32 i = tmpHashTable[hashIdx];
+            for (count = 0; i >= tmpMinChain && count < cacheSize; count++) {
+                /* skip through the chain to the first position that won't be
+                 * in the hash cache bucket */
+                if (i < minChain) {
+                    countBeyondMinChain++;
+                }
+                i = tmpChainTable[i - tmpMinChain];
+            }
+            if (count == cacheSize) {
+                for (count = 0; count < chainLimit;) {
+                    if (i < minChain) {
+                        if (!i || countBeyondMinChain++ > cacheSize) {
+                            /* only allow pulling `cacheSize` number of entries
+                             * into the cache or chainTable beyond `minChain`,
+                             * to replace the entries pulled out of the
+                             * chainTable into the cache. This lets us reach
+                             * back further without increasing the total number
+                             * of entries in the chainTable, guaranteeing the
+                             * DDSS chain table will fit into the space
+                             * allocated for the regular one. */
+                            break;
+                        }
+                    }
+                    chainTable[chainPos++] = i;
+                    count++;
+                    if (i < tmpMinChain) {
+                        break;
+                    }
+                    i = tmpChainTable[i - tmpMinChain];
+                }
+            } else {
+                count = 0;
+            }
+            if (count) {
+                tmpHashTable[hashIdx] = ((chainPos - count) << 8) + count;
+            } else {
+                tmpHashTable[hashIdx] = 0;
+            }
+        }
+        assert(chainPos <= chainSize); /* I believe this is guaranteed... */
+    }
+
+    /* move chain pointers into the last entry of each hash bucket */
+    for (hashIdx = (1 << hashLog); hashIdx; ) {
+        U32 const bucketIdx = --hashIdx << ZSTD_LAZY_DDSS_BUCKET_LOG;
+        U32 const chainPackedPointer = tmpHashTable[hashIdx];
+        U32 i;
+        for (i = 0; i < cacheSize; i++) {
+            hashTable[bucketIdx + i] = 0;
+        }
+        hashTable[bucketIdx + bucketSize - 1] = chainPackedPointer;
+    }
+
+    /* fill the buckets of the hash table */
+    for (idx = ms->nextToUpdate; idx < target; idx++) {
+        U32 const h = (U32)ZSTD_hashPtr(base + idx, hashLog, ms->cParams.minMatch)
+                   << ZSTD_LAZY_DDSS_BUCKET_LOG;
+        U32 i;
+        /* Shift hash cache down 1. */
+        for (i = cacheSize - 1; i; i--)
+            hashTable[h + i] = hashTable[h + i - 1];
+        hashTable[h] = idx;
+    }
+
+    ms->nextToUpdate = target;
+}
+
+
+/* inlining is important to hardwire a hot branch (template emulation) */
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_HcFindBestMatch_generic (
+                        ZSTD_matchState_t* ms,
+                        const BYTE* const ip, const BYTE* const iLimit,
+                        size_t* offsetPtr,
+                        const U32 mls, const ZSTD_dictMode_e dictMode)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32* const chainTable = ms->chainTable;
+    const U32 chainSize = (1 << cParams->chainLog);
+    const U32 chainMask = chainSize-1;
+    const BYTE* const base = ms->window.base;
+    const BYTE* const dictBase = ms->window.dictBase;
+    const U32 dictLimit = ms->window.dictLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const U32 curr = (U32)(ip-base);
+    const U32 maxDistance = 1U << cParams->windowLog;
+    const U32 lowestValid = ms->window.lowLimit;
+    const U32 withinMaxDistance = (curr - lowestValid > maxDistance) ? curr - maxDistance : lowestValid;
+    const U32 isDictionary = (ms->loadedDictEnd != 0);
+    const U32 lowLimit = isDictionary ? lowestValid : withinMaxDistance;
+    const U32 minChain = curr > chainSize ? curr - chainSize : 0;
+    U32 nbAttempts = 1U << cParams->searchLog;
+    size_t ml=4-1;
+
+    const ZSTD_matchState_t* const dms = ms->dictMatchState;
+    const U32 ddsHashLog = dictMode == ZSTD_dedicatedDictSearch
+                         ? dms->cParams.hashLog - ZSTD_LAZY_DDSS_BUCKET_LOG : 0;
+    const size_t ddsIdx = dictMode == ZSTD_dedicatedDictSearch
+                        ? ZSTD_hashPtr(ip, ddsHashLog, mls) << ZSTD_LAZY_DDSS_BUCKET_LOG : 0;
+
+    U32 matchIndex;
+
+    if (dictMode == ZSTD_dedicatedDictSearch) {
+        const U32* entry = &dms->hashTable[ddsIdx];
+        PREFETCH_L1(entry);
+    }
+
+    /* HC4 match finder */
+    matchIndex = ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, mls);
+
+    for ( ; (matchIndex>=lowLimit) & (nbAttempts>0) ; nbAttempts--) {
+        size_t currentMl=0;
+        if ((dictMode != ZSTD_extDict) || matchIndex >= dictLimit) {
+            const BYTE* const match = base + matchIndex;
+            assert(matchIndex >= dictLimit);   /* ensures this is true if dictMode != ZSTD_extDict */
+            if (match[ml] == ip[ml])   /* potentially better */
+                currentMl = ZSTD_count(ip, match, iLimit);
+        } else {
+            const BYTE* const match = dictBase + matchIndex;
+            assert(match+4 <= dictEnd);
+            if (MEM_read32(match) == MEM_read32(ip))   /* assumption : matchIndex <= dictLimit-4 (by table construction) */
+                currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dictEnd, prefixStart) + 4;
+        }
+
+        /* save best solution */
+        if (currentMl > ml) {
+            ml = currentMl;
+            *offsetPtr = curr - matchIndex + ZSTD_REP_MOVE;
+            if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
+        }
+
+        if (matchIndex <= minChain) break;
+        matchIndex = NEXT_IN_CHAIN(matchIndex, chainMask);
+    }
+
+    assert(nbAttempts <= (1U << ZSTD_SEARCHLOG_MAX)); /* Check we haven't underflowed. */
+    if (dictMode == ZSTD_dedicatedDictSearch) {
+        const U32 ddsLowestIndex  = dms->window.dictLimit;
+        const BYTE* const ddsBase = dms->window.base;
+        const BYTE* const ddsEnd  = dms->window.nextSrc;
+        const U32 ddsSize         = (U32)(ddsEnd - ddsBase);
+        const U32 ddsIndexDelta   = dictLimit - ddsSize;
+        const U32 bucketSize      = (1 << ZSTD_LAZY_DDSS_BUCKET_LOG);
+        const U32 bucketLimit     = nbAttempts < bucketSize - 1 ? nbAttempts : bucketSize - 1;
+        U32 ddsAttempt;
+
+        for (ddsAttempt = 0; ddsAttempt < bucketSize - 1; ddsAttempt++) {
+            PREFETCH_L1(ddsBase + dms->hashTable[ddsIdx + ddsAttempt]);
+        }
+
+        {
+            U32 const chainPackedPointer = dms->hashTable[ddsIdx + bucketSize - 1];
+            U32 const chainIndex = chainPackedPointer >> 8;
+
+            PREFETCH_L1(&dms->chainTable[chainIndex]);
+        }
+
+        for (ddsAttempt = 0; ddsAttempt < bucketLimit; ddsAttempt++) {
+            size_t currentMl=0;
+            const BYTE* match;
+            matchIndex = dms->hashTable[ddsIdx + ddsAttempt];
+            match = ddsBase + matchIndex;
+
+            if (!matchIndex) {
+                return ml;
+            }
+
+            /* guaranteed by table construction */
+            (void)ddsLowestIndex;
+            assert(matchIndex >= ddsLowestIndex);
+            assert(match+4 <= ddsEnd);
+            if (MEM_read32(match) == MEM_read32(ip)) {
+                /* assumption : matchIndex <= dictLimit-4 (by table construction) */
+                currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, ddsEnd, prefixStart) + 4;
+            }
+
+            /* save best solution */
+            if (currentMl > ml) {
+                ml = currentMl;
+                *offsetPtr = curr - (matchIndex + ddsIndexDelta) + ZSTD_REP_MOVE;
+                if (ip+currentMl == iLimit) {
+                    /* best possible, avoids read overflow on next attempt */
+                    return ml;
+                }
+            }
+        }
+
+        {
+            U32 const chainPackedPointer = dms->hashTable[ddsIdx + bucketSize - 1];
+            U32 chainIndex = chainPackedPointer >> 8;
+            U32 const chainLength = chainPackedPointer & 0xFF;
+            U32 const chainAttempts = nbAttempts - ddsAttempt;
+            U32 const chainLimit = chainAttempts > chainLength ? chainLength : chainAttempts;
+            U32 chainAttempt;
+
+            for (chainAttempt = 0 ; chainAttempt < chainLimit; chainAttempt++) {
+                PREFETCH_L1(ddsBase + dms->chainTable[chainIndex + chainAttempt]);
+            }
+
+            for (chainAttempt = 0 ; chainAttempt < chainLimit; chainAttempt++, chainIndex++) {
+                size_t currentMl=0;
+                const BYTE* match;
+                matchIndex = dms->chainTable[chainIndex];
+                match = ddsBase + matchIndex;
+
+                /* guaranteed by table construction */
+                assert(matchIndex >= ddsLowestIndex);
+                assert(match+4 <= ddsEnd);
+                if (MEM_read32(match) == MEM_read32(ip)) {
+                    /* assumption : matchIndex <= dictLimit-4 (by table construction) */
+                    currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, ddsEnd, prefixStart) + 4;
+                }
+
+                /* save best solution */
+                if (currentMl > ml) {
+                    ml = currentMl;
+                    *offsetPtr = curr - (matchIndex + ddsIndexDelta) + ZSTD_REP_MOVE;
+                    if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
+                }
+            }
+        }
+    } else if (dictMode == ZSTD_dictMatchState) {
+        const U32* const dmsChainTable = dms->chainTable;
+        const U32 dmsChainSize         = (1 << dms->cParams.chainLog);
+        const U32 dmsChainMask         = dmsChainSize - 1;
+        const U32 dmsLowestIndex       = dms->window.dictLimit;
+        const BYTE* const dmsBase      = dms->window.base;
+        const BYTE* const dmsEnd       = dms->window.nextSrc;
+        const U32 dmsSize              = (U32)(dmsEnd - dmsBase);
+        const U32 dmsIndexDelta        = dictLimit - dmsSize;
+        const U32 dmsMinChain = dmsSize > dmsChainSize ? dmsSize - dmsChainSize : 0;
+
+        matchIndex = dms->hashTable[ZSTD_hashPtr(ip, dms->cParams.hashLog, mls)];
+
+        for ( ; (matchIndex>=dmsLowestIndex) & (nbAttempts>0) ; nbAttempts--) {
+            size_t currentMl=0;
+            const BYTE* const match = dmsBase + matchIndex;
+            assert(match+4 <= dmsEnd);
+            if (MEM_read32(match) == MEM_read32(ip))   /* assumption : matchIndex <= dictLimit-4 (by table construction) */
+                currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dmsEnd, prefixStart) + 4;
+
+            /* save best solution */
+            if (currentMl > ml) {
+                ml = currentMl;
+                *offsetPtr = curr - (matchIndex + dmsIndexDelta) + ZSTD_REP_MOVE;
+                if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
+            }
+
+            if (matchIndex <= dmsMinChain) break;
+
+            matchIndex = dmsChainTable[matchIndex & dmsChainMask];
+        }
+    }
+
+    return ml;
+}
+
+
+FORCE_INLINE_TEMPLATE size_t ZSTD_HcFindBestMatch_selectMLS (
+                        ZSTD_matchState_t* ms,
+                        const BYTE* ip, const BYTE* const iLimit,
+                        size_t* offsetPtr)
+{
+    switch(ms->cParams.minMatch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 4, ZSTD_noDict);
+    case 5 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 5, ZSTD_noDict);
+    case 7 :
+    case 6 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 6, ZSTD_noDict);
+    }
+}
+
+
+static size_t ZSTD_HcFindBestMatch_dictMatchState_selectMLS (
+                        ZSTD_matchState_t* ms,
+                        const BYTE* ip, const BYTE* const iLimit,
+                        size_t* offsetPtr)
+{
+    switch(ms->cParams.minMatch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 4, ZSTD_dictMatchState);
+    case 5 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 5, ZSTD_dictMatchState);
+    case 7 :
+    case 6 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 6, ZSTD_dictMatchState);
+    }
+}
+
+
+static size_t ZSTD_HcFindBestMatch_dedicatedDictSearch_selectMLS (
+                        ZSTD_matchState_t* ms,
+                        const BYTE* ip, const BYTE* const iLimit,
+                        size_t* offsetPtr)
+{
+    switch(ms->cParams.minMatch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 4, ZSTD_dedicatedDictSearch);
+    case 5 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 5, ZSTD_dedicatedDictSearch);
+    case 7 :
+    case 6 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 6, ZSTD_dedicatedDictSearch);
+    }
+}
+
+
+FORCE_INLINE_TEMPLATE size_t ZSTD_HcFindBestMatch_extDict_selectMLS (
+                        ZSTD_matchState_t* ms,
+                        const BYTE* ip, const BYTE* const iLimit,
+                        size_t* offsetPtr)
+{
+    switch(ms->cParams.minMatch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 4, ZSTD_extDict);
+    case 5 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 5, ZSTD_extDict);
+    case 7 :
+    case 6 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 6, ZSTD_extDict);
+    }
+}
+
+
+/* *******************************
+*  Common parser - lazy strategy
+*********************************/
+typedef enum { search_hashChain, search_binaryTree } searchMethod_e;
+
+FORCE_INLINE_TEMPLATE size_t
+ZSTD_compressBlock_lazy_generic(
+                        ZSTD_matchState_t* ms, seqStore_t* seqStore,
+                        U32 rep[ZSTD_REP_NUM],
+                        const void* src, size_t srcSize,
+                        const searchMethod_e searchMethod, const U32 depth,
+                        ZSTD_dictMode_e const dictMode)
+{
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    const BYTE* const base = ms->window.base;
+    const U32 prefixLowestIndex = ms->window.dictLimit;
+    const BYTE* const prefixLowest = base + prefixLowestIndex;
+
+    typedef size_t (*searchMax_f)(
+                        ZSTD_matchState_t* ms,
+                        const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr);
+
+    /*
+     * This table is indexed first by the four ZSTD_dictMode_e values, and then
+     * by the two searchMethod_e values. NULLs are placed for configurations
+     * that should never occur (extDict modes go to the other implementation
+     * below and there is no DDSS for binary tree search yet).
+     */
+    const searchMax_f searchFuncs[4][2] = {
+        {
+            ZSTD_HcFindBestMatch_selectMLS,
+            ZSTD_BtFindBestMatch_selectMLS
+        },
+        {
+            NULL,
+            NULL
+        },
+        {
+            ZSTD_HcFindBestMatch_dictMatchState_selectMLS,
+            ZSTD_BtFindBestMatch_dictMatchState_selectMLS
+        },
+        {
+            ZSTD_HcFindBestMatch_dedicatedDictSearch_selectMLS,
+            NULL
+        }
+    };
+
+    searchMax_f const searchMax = searchFuncs[dictMode][searchMethod == search_binaryTree];
+    U32 offset_1 = rep[0], offset_2 = rep[1], savedOffset=0;
+
+    const int isDMS = dictMode == ZSTD_dictMatchState;
+    const int isDDS = dictMode == ZSTD_dedicatedDictSearch;
+    const int isDxS = isDMS || isDDS;
+    const ZSTD_matchState_t* const dms = ms->dictMatchState;
+    const U32 dictLowestIndex      = isDxS ? dms->window.dictLimit : 0;
+    const BYTE* const dictBase     = isDxS ? dms->window.base : NULL;
+    const BYTE* const dictLowest   = isDxS ? dictBase + dictLowestIndex : NULL;
+    const BYTE* const dictEnd      = isDxS ? dms->window.nextSrc : NULL;
+    const U32 dictIndexDelta       = isDxS ?
+                                     prefixLowestIndex - (U32)(dictEnd - dictBase) :
+                                     0;
+    const U32 dictAndPrefixLength = (U32)((ip - prefixLowest) + (dictEnd - dictLowest));
+
+    assert(searchMax != NULL);
+
+    DEBUGLOG(5, "ZSTD_compressBlock_lazy_generic (dictMode=%u)", (U32)dictMode);
+
+    /* init */
+    ip += (dictAndPrefixLength == 0);
+    if (dictMode == ZSTD_noDict) {
+        U32 const curr = (U32)(ip - base);
+        U32 const windowLow = ZSTD_getLowestPrefixIndex(ms, curr, ms->cParams.windowLog);
+        U32 const maxRep = curr - windowLow;
+        if (offset_2 > maxRep) savedOffset = offset_2, offset_2 = 0;
+        if (offset_1 > maxRep) savedOffset = offset_1, offset_1 = 0;
+    }
+    if (isDxS) {
+        /* dictMatchState repCode checks don't currently handle repCode == 0
+         * disabling. */
+        assert(offset_1 <= dictAndPrefixLength);
+        assert(offset_2 <= dictAndPrefixLength);
+    }
+
+    /* Match Loop */
+#if defined(__x86_64__)
+    /* I've measured random a 5% speed loss on levels 5 & 6 (greedy) when the
+     * code alignment is perturbed. To fix the instability align the loop on 32-bytes.
+     */
+    __asm__(".p2align 5");
+#endif
+    while (ip < ilimit) {
+        size_t matchLength=0;
+        size_t offset=0;
+        const BYTE* start=ip+1;
+
+        /* check repCode */
+        if (isDxS) {
+            const U32 repIndex = (U32)(ip - base) + 1 - offset_1;
+            const BYTE* repMatch = ((dictMode == ZSTD_dictMatchState || dictMode == ZSTD_dedicatedDictSearch)
+                                && repIndex < prefixLowestIndex) ?
+                                   dictBase + (repIndex - dictIndexDelta) :
+                                   base + repIndex;
+            if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)
+                && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
+                const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
+                matchLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
+                if (depth==0) goto _storeSequence;
+            }
+        }
+        if ( dictMode == ZSTD_noDict
+          && ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1)))) {
+            matchLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
+            if (depth==0) goto _storeSequence;
+        }
+
+        /* first search (depth 0) */
+        {   size_t offsetFound = 999999999;
+            size_t const ml2 = searchMax(ms, ip, iend, &offsetFound);
+            if (ml2 > matchLength)
+                matchLength = ml2, start = ip, offset=offsetFound;
+        }
+
+        if (matchLength < 4) {
+            ip += ((ip-anchor) >> kSearchStrength) + 1;   /* jump faster over incompressible sections */
+            continue;
+        }
+
+        /* let's try to find a better solution */
+        if (depth>=1)
+        while (ip<ilimit) {
+            ip ++;
+            if ( (dictMode == ZSTD_noDict)
+              && (offset) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) {
+                size_t const mlRep = ZSTD_count(ip+4, ip+4-offset_1, iend) + 4;
+                int const gain2 = (int)(mlRep * 3);
+                int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);
+                if ((mlRep >= 4) && (gain2 > gain1))
+                    matchLength = mlRep, offset = 0, start = ip;
+            }
+            if (isDxS) {
+                const U32 repIndex = (U32)(ip - base) - offset_1;
+                const BYTE* repMatch = repIndex < prefixLowestIndex ?
+                               dictBase + (repIndex - dictIndexDelta) :
+                               base + repIndex;
+                if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)
+                    && (MEM_read32(repMatch) == MEM_read32(ip)) ) {
+                    const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
+                    size_t const mlRep = ZSTD_count_2segments(ip+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
+                    int const gain2 = (int)(mlRep * 3);
+                    int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);
+                    if ((mlRep >= 4) && (gain2 > gain1))
+                        matchLength = mlRep, offset = 0, start = ip;
+                }
+            }
+            {   size_t offset2=999999999;
+                size_t const ml2 = searchMax(ms, ip, iend, &offset2);
+                int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
+                int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 4);
+                if ((ml2 >= 4) && (gain2 > gain1)) {
+                    matchLength = ml2, offset = offset2, start = ip;
+                    continue;   /* search a better one */
+            }   }
+
+            /* let's find an even better one */
+            if ((depth==2) && (ip<ilimit)) {
+                ip ++;
+                if ( (dictMode == ZSTD_noDict)
+                  && (offset) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) {
+                    size_t const mlRep = ZSTD_count(ip+4, ip+4-offset_1, iend) + 4;
+                    int const gain2 = (int)(mlRep * 4);
+                    int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);
+                    if ((mlRep >= 4) && (gain2 > gain1))
+                        matchLength = mlRep, offset = 0, start = ip;
+                }
+                if (isDxS) {
+                    const U32 repIndex = (U32)(ip - base) - offset_1;
+                    const BYTE* repMatch = repIndex < prefixLowestIndex ?
+                                   dictBase + (repIndex - dictIndexDelta) :
+                                   base + repIndex;
+                    if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)
+                        && (MEM_read32(repMatch) == MEM_read32(ip)) ) {
+                        const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
+                        size_t const mlRep = ZSTD_count_2segments(ip+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
+                        int const gain2 = (int)(mlRep * 4);
+                        int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);
+                        if ((mlRep >= 4) && (gain2 > gain1))
+                            matchLength = mlRep, offset = 0, start = ip;
+                    }
+                }
+                {   size_t offset2=999999999;
+                    size_t const ml2 = searchMax(ms, ip, iend, &offset2);
+                    int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
+                    int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 7);
+                    if ((ml2 >= 4) && (gain2 > gain1)) {
+                        matchLength = ml2, offset = offset2, start = ip;
+                        continue;
+            }   }   }
+            break;  /* nothing found : store previous solution */
+        }
+
+        /* NOTE:
+         * start[-offset+ZSTD_REP_MOVE-1] is undefined behavior.
+         * (-offset+ZSTD_REP_MOVE-1) is unsigned, and is added to start, which
+         * overflows the pointer, which is undefined behavior.
+         */
+        /* catch up */
+        if (offset) {
+            if (dictMode == ZSTD_noDict) {
+                while ( ((start > anchor) & (start - (offset-ZSTD_REP_MOVE) > prefixLowest))
+                     && (start[-1] == (start-(offset-ZSTD_REP_MOVE))[-1]) )  /* only search for offset within prefix */
+                    { start--; matchLength++; }
+            }
+            if (isDxS) {
+                U32 const matchIndex = (U32)((start-base) - (offset - ZSTD_REP_MOVE));
+                const BYTE* match = (matchIndex < prefixLowestIndex) ? dictBase + matchIndex - dictIndexDelta : base + matchIndex;
+                const BYTE* const mStart = (matchIndex < prefixLowestIndex) ? dictLowest : prefixLowest;
+                while ((start>anchor) && (match>mStart) && (start[-1] == match[-1])) { start--; match--; matchLength++; }  /* catch up */
+            }
+            offset_2 = offset_1; offset_1 = (U32)(offset - ZSTD_REP_MOVE);
+        }
+        /* store sequence */
+_storeSequence:
+        {   size_t const litLength = start - anchor;
+            ZSTD_storeSeq(seqStore, litLength, anchor, iend, (U32)offset, matchLength-MINMATCH);
+            anchor = ip = start + matchLength;
+        }
+
+        /* check immediate repcode */
+        if (isDxS) {
+            while (ip <= ilimit) {
+                U32 const current2 = (U32)(ip-base);
+                U32 const repIndex = current2 - offset_2;
+                const BYTE* repMatch = repIndex < prefixLowestIndex ?
+                        dictBase - dictIndexDelta + repIndex :
+                        base + repIndex;
+                if ( ((U32)((prefixLowestIndex-1) - (U32)repIndex) >= 3 /* intentional overflow */)
+                   && (MEM_read32(repMatch) == MEM_read32(ip)) ) {
+                    const BYTE* const repEnd2 = repIndex < prefixLowestIndex ? dictEnd : iend;
+                    matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd2, prefixLowest) + 4;
+                    offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset;   /* swap offset_2 <=> offset_1 */
+                    ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, matchLength-MINMATCH);
+                    ip += matchLength;
+                    anchor = ip;
+                    continue;
+                }
+                break;
+            }
+        }
+
+        if (dictMode == ZSTD_noDict) {
+            while ( ((ip <= ilimit) & (offset_2>0))
+                 && (MEM_read32(ip) == MEM_read32(ip - offset_2)) ) {
+                /* store sequence */
+                matchLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
+                offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap repcodes */
+                ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, matchLength-MINMATCH);
+                ip += matchLength;
+                anchor = ip;
+                continue;   /* faster when present ... (?) */
+    }   }   }
+
+    /* Save reps for next block */
+    rep[0] = offset_1 ? offset_1 : savedOffset;
+    rep[1] = offset_2 ? offset_2 : savedOffset;
+
+    /* Return the last literals size */
+    return (size_t)(iend - anchor);
+}
+
+
+size_t ZSTD_compressBlock_btlazy2(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_noDict);
+}
+
+size_t ZSTD_compressBlock_lazy2(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_noDict);
+}
+
+size_t ZSTD_compressBlock_lazy(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_noDict);
+}
+
+size_t ZSTD_compressBlock_greedy(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_noDict);
+}
+
+size_t ZSTD_compressBlock_btlazy2_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_dictMatchState);
+}
+
+size_t ZSTD_compressBlock_lazy2_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dictMatchState);
+}
+
+size_t ZSTD_compressBlock_lazy_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dictMatchState);
+}
+
+size_t ZSTD_compressBlock_greedy_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dictMatchState);
+}
+
+
+size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dedicatedDictSearch);
+}
+
+size_t ZSTD_compressBlock_lazy_dedicatedDictSearch(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dedicatedDictSearch);
+}
+
+size_t ZSTD_compressBlock_greedy_dedicatedDictSearch(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dedicatedDictSearch);
+}
+
+
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_lazy_extDict_generic(
+                        ZSTD_matchState_t* ms, seqStore_t* seqStore,
+                        U32 rep[ZSTD_REP_NUM],
+                        const void* src, size_t srcSize,
+                        const searchMethod_e searchMethod, const U32 depth)
+{
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    const BYTE* const base = ms->window.base;
+    const U32 dictLimit = ms->window.dictLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    const BYTE* const dictBase = ms->window.dictBase;
+    const BYTE* const dictEnd  = dictBase + dictLimit;
+    const BYTE* const dictStart  = dictBase + ms->window.lowLimit;
+    const U32 windowLog = ms->cParams.windowLog;
+
+    typedef size_t (*searchMax_f)(
+                        ZSTD_matchState_t* ms,
+                        const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr);
+    searchMax_f searchMax = searchMethod==search_binaryTree ? ZSTD_BtFindBestMatch_extDict_selectMLS : ZSTD_HcFindBestMatch_extDict_selectMLS;
+
+    U32 offset_1 = rep[0], offset_2 = rep[1];
+
+    DEBUGLOG(5, "ZSTD_compressBlock_lazy_extDict_generic");
+
+    /* init */
+    ip += (ip == prefixStart);
+
+    /* Match Loop */
+#if defined(__x86_64__)
+    /* I've measured random a 5% speed loss on levels 5 & 6 (greedy) when the
+     * code alignment is perturbed. To fix the instability align the loop on 32-bytes.
+     */
+    __asm__(".p2align 5");
+#endif
+    while (ip < ilimit) {
+        size_t matchLength=0;
+        size_t offset=0;
+        const BYTE* start=ip+1;
+        U32 curr = (U32)(ip-base);
+
+        /* check repCode */
+        {   const U32 windowLow = ZSTD_getLowestMatchIndex(ms, curr+1, windowLog);
+            const U32 repIndex = (U32)(curr+1 - offset_1);
+            const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
+            const BYTE* const repMatch = repBase + repIndex;
+            if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > windowLow))   /* intentional overflow */
+            if (MEM_read32(ip+1) == MEM_read32(repMatch)) {
+                /* repcode detected we should take it */
+                const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
+                matchLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repEnd, prefixStart) + 4;
+                if (depth==0) goto _storeSequence;
+        }   }
+
+        /* first search (depth 0) */
+        {   size_t offsetFound = 999999999;
+            size_t const ml2 = searchMax(ms, ip, iend, &offsetFound);
+            if (ml2 > matchLength)
+                matchLength = ml2, start = ip, offset=offsetFound;
+        }
+
+        if (matchLength < 4) {
+            ip += ((ip-anchor) >> kSearchStrength) + 1;   /* jump faster over incompressible sections */
+            continue;
+        }
+
+        /* let's try to find a better solution */
+        if (depth>=1)
+        while (ip<ilimit) {
+            ip ++;
+            curr++;
+            /* check repCode */
+            if (offset) {
+                const U32 windowLow = ZSTD_getLowestMatchIndex(ms, curr, windowLog);
+                const U32 repIndex = (U32)(curr - offset_1);
+                const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
+                const BYTE* const repMatch = repBase + repIndex;
+                if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > windowLow))  /* intentional overflow */
+                if (MEM_read32(ip) == MEM_read32(repMatch)) {
+                    /* repcode detected */
+                    const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
+                    size_t const repLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;
+                    int const gain2 = (int)(repLength * 3);
+                    int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);
+                    if ((repLength >= 4) && (gain2 > gain1))
+                        matchLength = repLength, offset = 0, start = ip;
+            }   }
+
+            /* search match, depth 1 */
+            {   size_t offset2=999999999;
+                size_t const ml2 = searchMax(ms, ip, iend, &offset2);
+                int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
+                int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 4);
+                if ((ml2 >= 4) && (gain2 > gain1)) {
+                    matchLength = ml2, offset = offset2, start = ip;
+                    continue;   /* search a better one */
+            }   }
+
+            /* let's find an even better one */
+            if ((depth==2) && (ip<ilimit)) {
+                ip ++;
+                curr++;
+                /* check repCode */
+                if (offset) {
+                    const U32 windowLow = ZSTD_getLowestMatchIndex(ms, curr, windowLog);
+                    const U32 repIndex = (U32)(curr - offset_1);
+                    const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
+                    const BYTE* const repMatch = repBase + repIndex;
+                    if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > windowLow))  /* intentional overflow */
+                    if (MEM_read32(ip) == MEM_read32(repMatch)) {
+                        /* repcode detected */
+                        const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
+                        size_t const repLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;
+                        int const gain2 = (int)(repLength * 4);
+                        int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);
+                        if ((repLength >= 4) && (gain2 > gain1))
+                            matchLength = repLength, offset = 0, start = ip;
+                }   }
+
+                /* search match, depth 2 */
+                {   size_t offset2=999999999;
+                    size_t const ml2 = searchMax(ms, ip, iend, &offset2);
+                    int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
+                    int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 7);
+                    if ((ml2 >= 4) && (gain2 > gain1)) {
+                        matchLength = ml2, offset = offset2, start = ip;
+                        continue;
+            }   }   }
+            break;  /* nothing found : store previous solution */
+        }
+
+        /* catch up */
+        if (offset) {
+            U32 const matchIndex = (U32)((start-base) - (offset - ZSTD_REP_MOVE));
+            const BYTE* match = (matchIndex < dictLimit) ? dictBase + matchIndex : base + matchIndex;
+            const BYTE* const mStart = (matchIndex < dictLimit) ? dictStart : prefixStart;
+            while ((start>anchor) && (match>mStart) && (start[-1] == match[-1])) { start--; match--; matchLength++; }  /* catch up */
+            offset_2 = offset_1; offset_1 = (U32)(offset - ZSTD_REP_MOVE);
+        }
+
+        /* store sequence */
+_storeSequence:
+        {   size_t const litLength = start - anchor;
+            ZSTD_storeSeq(seqStore, litLength, anchor, iend, (U32)offset, matchLength-MINMATCH);
+            anchor = ip = start + matchLength;
+        }
+
+        /* check immediate repcode */
+        while (ip <= ilimit) {
+            const U32 repCurrent = (U32)(ip-base);
+            const U32 windowLow = ZSTD_getLowestMatchIndex(ms, repCurrent, windowLog);
+            const U32 repIndex = repCurrent - offset_2;
+            const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
+            const BYTE* const repMatch = repBase + repIndex;
+            if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > windowLow))  /* intentional overflow */
+            if (MEM_read32(ip) == MEM_read32(repMatch)) {
+                /* repcode detected we should take it */
+                const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
+                matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;
+                offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset;   /* swap offset history */
+                ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, matchLength-MINMATCH);
+                ip += matchLength;
+                anchor = ip;
+                continue;   /* faster when present ... (?) */
+            }
+            break;
+    }   }
+
+    /* Save reps for next block */
+    rep[0] = offset_1;
+    rep[1] = offset_2;
+
+    /* Return the last literals size */
+    return (size_t)(iend - anchor);
+}
+
+
+size_t ZSTD_compressBlock_greedy_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0);
+}
+
+size_t ZSTD_compressBlock_lazy_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1);
+}
+
+size_t ZSTD_compressBlock_lazy2_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2);
+}
+
+size_t ZSTD_compressBlock_btlazy2_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2);
+}
diff --git a/lib/zstd/compress/zstd_lazy.h b/lib/zstd/compress/zstd_lazy.h
new file mode 100644
index 0000000..2fc5a61
--- /dev/null
+++ b/lib/zstd/compress/zstd_lazy.h
@@ -0,0 +1,81 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_LAZY_H
+#define ZSTD_LAZY_H
+
+
+#include "zstd_compress_internal.h"
+
+/*
+ * Dedicated Dictionary Search Structure bucket log. In the
+ * ZSTD_dedicatedDictSearch mode, the hashTable has
+ * 2 ** ZSTD_LAZY_DDSS_BUCKET_LOG entries in each bucket, rather than just
+ * one.
+ */
+#define ZSTD_LAZY_DDSS_BUCKET_LOG 2
+
+U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip);
+
+void ZSTD_dedicatedDictSearch_lazy_loadDictionary(ZSTD_matchState_t* ms, const BYTE* const ip);
+
+void ZSTD_preserveUnsortedMark (U32* const table, U32 const size, U32 const reducerValue);  /*! used in ZSTD_reduceIndex(). preemptively increase value of ZSTD_DUBT_UNSORTED_MARK */
+
+size_t ZSTD_compressBlock_btlazy2(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_greedy(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+
+size_t ZSTD_compressBlock_btlazy2_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_greedy_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+
+size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy_dedicatedDictSearch(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_greedy_dedicatedDictSearch(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+
+size_t ZSTD_compressBlock_greedy_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_btlazy2_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+
+
+#endif /* ZSTD_LAZY_H */
diff --git a/lib/zstd/compress/zstd_ldm.c b/lib/zstd/compress/zstd_ldm.c
new file mode 100644
index 0000000..8ef7e88
--- /dev/null
+++ b/lib/zstd/compress/zstd_ldm.c
@@ -0,0 +1,686 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#include "zstd_ldm.h"
+
+#include "../common/debug.h"
+#include <linux/xxhash.h>
+#include "zstd_fast.h"          /* ZSTD_fillHashTable() */
+#include "zstd_double_fast.h"   /* ZSTD_fillDoubleHashTable() */
+#include "zstd_ldm_geartab.h"
+
+#define LDM_BUCKET_SIZE_LOG 3
+#define LDM_MIN_MATCH_LENGTH 64
+#define LDM_HASH_RLOG 7
+
+typedef struct {
+    U64 rolling;
+    U64 stopMask;
+} ldmRollingHashState_t;
+
+/* ZSTD_ldm_gear_init():
+ *
+ * Initializes the rolling hash state such that it will honor the
+ * settings in params. */
+static void ZSTD_ldm_gear_init(ldmRollingHashState_t* state, ldmParams_t const* params)
+{
+    unsigned maxBitsInMask = MIN(params->minMatchLength, 64);
+    unsigned hashRateLog = params->hashRateLog;
+
+    state->rolling = ~(U32)0;
+
+    /* The choice of the splitting criterion is subject to two conditions:
+     *   1. it has to trigger on average every 2^(hashRateLog) bytes;
+     *   2. ideally, it has to depend on a window of minMatchLength bytes.
+     *
+     * In the gear hash algorithm, bit n depends on the last n bytes;
+     * so in order to obtain a good quality splitting criterion it is
+     * preferable to use bits with high weight.
+     *
+     * To match condition 1 we use a mask with hashRateLog bits set
+     * and, because of the previous remark, we make sure these bits
+     * have the highest possible weight while still respecting
+     * condition 2.
+     */
+    if (hashRateLog > 0 && hashRateLog <= maxBitsInMask) {
+        state->stopMask = (((U64)1 << hashRateLog) - 1) << (maxBitsInMask - hashRateLog);
+    } else {
+        /* In this degenerate case we simply honor the hash rate. */
+        state->stopMask = ((U64)1 << hashRateLog) - 1;
+    }
+}
+
+/* ZSTD_ldm_gear_feed():
+ *
+ * Registers in the splits array all the split points found in the first
+ * size bytes following the data pointer. This function terminates when
+ * either all the data has been processed or LDM_BATCH_SIZE splits are
+ * present in the splits array.
+ *
+ * Precondition: The splits array must not be full.
+ * Returns: The number of bytes processed. */
+static size_t ZSTD_ldm_gear_feed(ldmRollingHashState_t* state,
+                                 BYTE const* data, size_t size,
+                                 size_t* splits, unsigned* numSplits)
+{
+    size_t n;
+    U64 hash, mask;
+
+    hash = state->rolling;
+    mask = state->stopMask;
+    n = 0;
+
+#define GEAR_ITER_ONCE() do { \
+        hash = (hash << 1) + ZSTD_ldm_gearTab[data[n] & 0xff]; \
+        n += 1; \
+        if (UNLIKELY((hash & mask) == 0)) { \
+            splits[*numSplits] = n; \
+            *numSplits += 1; \
+            if (*numSplits == LDM_BATCH_SIZE) \
+                goto done; \
+        } \
+    } while (0)
+
+    while (n + 3 < size) {
+        GEAR_ITER_ONCE();
+        GEAR_ITER_ONCE();
+        GEAR_ITER_ONCE();
+        GEAR_ITER_ONCE();
+    }
+    while (n < size) {
+        GEAR_ITER_ONCE();
+    }
+
+#undef GEAR_ITER_ONCE
+
+done:
+    state->rolling = hash;
+    return n;
+}
+
+void ZSTD_ldm_adjustParameters(ldmParams_t* params,
+                               ZSTD_compressionParameters const* cParams)
+{
+    params->windowLog = cParams->windowLog;
+    ZSTD_STATIC_ASSERT(LDM_BUCKET_SIZE_LOG <= ZSTD_LDM_BUCKETSIZELOG_MAX);
+    DEBUGLOG(4, "ZSTD_ldm_adjustParameters");
+    if (!params->bucketSizeLog) params->bucketSizeLog = LDM_BUCKET_SIZE_LOG;
+    if (!params->minMatchLength) params->minMatchLength = LDM_MIN_MATCH_LENGTH;
+    if (params->hashLog == 0) {
+        params->hashLog = MAX(ZSTD_HASHLOG_MIN, params->windowLog - LDM_HASH_RLOG);
+        assert(params->hashLog <= ZSTD_HASHLOG_MAX);
+    }
+    if (params->hashRateLog == 0) {
+        params->hashRateLog = params->windowLog < params->hashLog
+                                   ? 0
+                                   : params->windowLog - params->hashLog;
+    }
+    params->bucketSizeLog = MIN(params->bucketSizeLog, params->hashLog);
+}
+
+size_t ZSTD_ldm_getTableSize(ldmParams_t params)
+{
+    size_t const ldmHSize = ((size_t)1) << params.hashLog;
+    size_t const ldmBucketSizeLog = MIN(params.bucketSizeLog, params.hashLog);
+    size_t const ldmBucketSize = ((size_t)1) << (params.hashLog - ldmBucketSizeLog);
+    size_t const totalSize = ZSTD_cwksp_alloc_size(ldmBucketSize)
+                           + ZSTD_cwksp_alloc_size(ldmHSize * sizeof(ldmEntry_t));
+    return params.enableLdm ? totalSize : 0;
+}
+
+size_t ZSTD_ldm_getMaxNbSeq(ldmParams_t params, size_t maxChunkSize)
+{
+    return params.enableLdm ? (maxChunkSize / params.minMatchLength) : 0;
+}
+
+/* ZSTD_ldm_getBucket() :
+ *  Returns a pointer to the start of the bucket associated with hash. */
+static ldmEntry_t* ZSTD_ldm_getBucket(
+        ldmState_t* ldmState, size_t hash, ldmParams_t const ldmParams)
+{
+    return ldmState->hashTable + (hash << ldmParams.bucketSizeLog);
+}
+
+/* ZSTD_ldm_insertEntry() :
+ *  Insert the entry with corresponding hash into the hash table */
+static void ZSTD_ldm_insertEntry(ldmState_t* ldmState,
+                                 size_t const hash, const ldmEntry_t entry,
+                                 ldmParams_t const ldmParams)
+{
+    BYTE* const pOffset = ldmState->bucketOffsets + hash;
+    unsigned const offset = *pOffset;
+
+    *(ZSTD_ldm_getBucket(ldmState, hash, ldmParams) + offset) = entry;
+    *pOffset = (BYTE)((offset + 1) & ((1u << ldmParams.bucketSizeLog) - 1));
+
+}
+
+/* ZSTD_ldm_countBackwardsMatch() :
+ *  Returns the number of bytes that match backwards before pIn and pMatch.
+ *
+ *  We count only bytes where pMatch >= pBase and pIn >= pAnchor. */
+static size_t ZSTD_ldm_countBackwardsMatch(
+            const BYTE* pIn, const BYTE* pAnchor,
+            const BYTE* pMatch, const BYTE* pMatchBase)
+{
+    size_t matchLength = 0;
+    while (pIn > pAnchor && pMatch > pMatchBase && pIn[-1] == pMatch[-1]) {
+        pIn--;
+        pMatch--;
+        matchLength++;
+    }
+    return matchLength;
+}
+
+/* ZSTD_ldm_countBackwardsMatch_2segments() :
+ *  Returns the number of bytes that match backwards from pMatch,
+ *  even with the backwards match spanning 2 different segments.
+ *
+ *  On reaching `pMatchBase`, start counting from mEnd */
+static size_t ZSTD_ldm_countBackwardsMatch_2segments(
+                    const BYTE* pIn, const BYTE* pAnchor,
+                    const BYTE* pMatch, const BYTE* pMatchBase,
+                    const BYTE* pExtDictStart, const BYTE* pExtDictEnd)
+{
+    size_t matchLength = ZSTD_ldm_countBackwardsMatch(pIn, pAnchor, pMatch, pMatchBase);
+    if (pMatch - matchLength != pMatchBase || pMatchBase == pExtDictStart) {
+        /* If backwards match is entirely in the extDict or prefix, immediately return */
+        return matchLength;
+    }
+    DEBUGLOG(7, "ZSTD_ldm_countBackwardsMatch_2segments: found 2-parts backwards match (length in prefix==%zu)", matchLength);
+    matchLength += ZSTD_ldm_countBackwardsMatch(pIn - matchLength, pAnchor, pExtDictEnd, pExtDictStart);
+    DEBUGLOG(7, "final backwards match length = %zu", matchLength);
+    return matchLength;
+}
+
+/* ZSTD_ldm_fillFastTables() :
+ *
+ *  Fills the relevant tables for the ZSTD_fast and ZSTD_dfast strategies.
+ *  This is similar to ZSTD_loadDictionaryContent.
+ *
+ *  The tables for the other strategies are filled within their
+ *  block compressors. */
+static size_t ZSTD_ldm_fillFastTables(ZSTD_matchState_t* ms,
+                                      void const* end)
+{
+    const BYTE* const iend = (const BYTE*)end;
+
+    switch(ms->cParams.strategy)
+    {
+    case ZSTD_fast:
+        ZSTD_fillHashTable(ms, iend, ZSTD_dtlm_fast);
+        break;
+
+    case ZSTD_dfast:
+        ZSTD_fillDoubleHashTable(ms, iend, ZSTD_dtlm_fast);
+        break;
+
+    case ZSTD_greedy:
+    case ZSTD_lazy:
+    case ZSTD_lazy2:
+    case ZSTD_btlazy2:
+    case ZSTD_btopt:
+    case ZSTD_btultra:
+    case ZSTD_btultra2:
+        break;
+    default:
+        assert(0);  /* not possible : not a valid strategy id */
+    }
+
+    return 0;
+}
+
+void ZSTD_ldm_fillHashTable(
+            ldmState_t* ldmState, const BYTE* ip,
+            const BYTE* iend, ldmParams_t const* params)
+{
+    U32 const minMatchLength = params->minMatchLength;
+    U32 const hBits = params->hashLog - params->bucketSizeLog;
+    BYTE const* const base = ldmState->window.base;
+    BYTE const* const istart = ip;
+    ldmRollingHashState_t hashState;
+    size_t* const splits = ldmState->splitIndices;
+    unsigned numSplits;
+
+    DEBUGLOG(5, "ZSTD_ldm_fillHashTable");
+
+    ZSTD_ldm_gear_init(&hashState, params);
+    while (ip < iend) {
+        size_t hashed;
+        unsigned n;
+        
+        numSplits = 0;
+        hashed = ZSTD_ldm_gear_feed(&hashState, ip, iend - ip, splits, &numSplits);
+
+        for (n = 0; n < numSplits; n++) {
+            if (ip + splits[n] >= istart + minMatchLength) {
+                BYTE const* const split = ip + splits[n] - minMatchLength;
+                U64 const xxhash = xxh64(split, minMatchLength, 0);
+                U32 const hash = (U32)(xxhash & (((U32)1 << hBits) - 1));
+                ldmEntry_t entry;
+
+                entry.offset = (U32)(split - base);
+                entry.checksum = (U32)(xxhash >> 32);
+                ZSTD_ldm_insertEntry(ldmState, hash, entry, *params);
+            }
+        }
+
+        ip += hashed;
+    }
+}
+
+
+/* ZSTD_ldm_limitTableUpdate() :
+ *
+ *  Sets cctx->nextToUpdate to a position corresponding closer to anchor
+ *  if it is far way
+ *  (after a long match, only update tables a limited amount). */
+static void ZSTD_ldm_limitTableUpdate(ZSTD_matchState_t* ms, const BYTE* anchor)
+{
+    U32 const curr = (U32)(anchor - ms->window.base);
+    if (curr > ms->nextToUpdate + 1024) {
+        ms->nextToUpdate =
+            curr - MIN(512, curr - ms->nextToUpdate - 1024);
+    }
+}
+
+static size_t ZSTD_ldm_generateSequences_internal(
+        ldmState_t* ldmState, rawSeqStore_t* rawSeqStore,
+        ldmParams_t const* params, void const* src, size_t srcSize)
+{
+    /* LDM parameters */
+    int const extDict = ZSTD_window_hasExtDict(ldmState->window);
+    U32 const minMatchLength = params->minMatchLength;
+    U32 const entsPerBucket = 1U << params->bucketSizeLog;
+    U32 const hBits = params->hashLog - params->bucketSizeLog;
+    /* Prefix and extDict parameters */
+    U32 const dictLimit = ldmState->window.dictLimit;
+    U32 const lowestIndex = extDict ? ldmState->window.lowLimit : dictLimit;
+    BYTE const* const base = ldmState->window.base;
+    BYTE const* const dictBase = extDict ? ldmState->window.dictBase : NULL;
+    BYTE const* const dictStart = extDict ? dictBase + lowestIndex : NULL;
+    BYTE const* const dictEnd = extDict ? dictBase + dictLimit : NULL;
+    BYTE const* const lowPrefixPtr = base + dictLimit;
+    /* Input bounds */
+    BYTE const* const istart = (BYTE const*)src;
+    BYTE const* const iend = istart + srcSize;
+    BYTE const* const ilimit = iend - HASH_READ_SIZE;
+    /* Input positions */
+    BYTE const* anchor = istart;
+    BYTE const* ip = istart;
+    /* Rolling hash state */
+    ldmRollingHashState_t hashState;
+    /* Arrays for staged-processing */
+    size_t* const splits = ldmState->splitIndices;
+    ldmMatchCandidate_t* const candidates = ldmState->matchCandidates;
+    unsigned numSplits;
+
+    if (srcSize < minMatchLength)
+        return iend - anchor;
+
+    /* Initialize the rolling hash state with the first minMatchLength bytes */
+    ZSTD_ldm_gear_init(&hashState, params);
+    {
+        size_t n = 0;
+
+        while (n < minMatchLength) {
+            numSplits = 0;
+            n += ZSTD_ldm_gear_feed(&hashState, ip + n, minMatchLength - n,
+                                    splits, &numSplits);
+        }
+        ip += minMatchLength;
+    }
+
+    while (ip < ilimit) {
+        size_t hashed;
+        unsigned n;
+
+        numSplits = 0;
+        hashed = ZSTD_ldm_gear_feed(&hashState, ip, ilimit - ip,
+                                    splits, &numSplits);
+
+        for (n = 0; n < numSplits; n++) {
+            BYTE const* const split = ip + splits[n] - minMatchLength;
+            U64 const xxhash = xxh64(split, minMatchLength, 0);
+            U32 const hash = (U32)(xxhash & (((U32)1 << hBits) - 1));
+
+            candidates[n].split = split;
+            candidates[n].hash = hash;
+            candidates[n].checksum = (U32)(xxhash >> 32);
+            candidates[n].bucket = ZSTD_ldm_getBucket(ldmState, hash, *params);
+            PREFETCH_L1(candidates[n].bucket);
+        }
+
+        for (n = 0; n < numSplits; n++) {
+            size_t forwardMatchLength = 0, backwardMatchLength = 0,
+                   bestMatchLength = 0, mLength;
+            BYTE const* const split = candidates[n].split;
+            U32 const checksum = candidates[n].checksum;
+            U32 const hash = candidates[n].hash;
+            ldmEntry_t* const bucket = candidates[n].bucket;
+            ldmEntry_t const* cur;
+            ldmEntry_t const* bestEntry = NULL;
+            ldmEntry_t newEntry;
+
+            newEntry.offset = (U32)(split - base);
+            newEntry.checksum = checksum;
+
+            /* If a split point would generate a sequence overlapping with
+             * the previous one, we merely register it in the hash table and
+             * move on */
+            if (split < anchor) {
+                ZSTD_ldm_insertEntry(ldmState, hash, newEntry, *params);
+                continue;
+            }
+
+            for (cur = bucket; cur < bucket + entsPerBucket; cur++) {
+                size_t curForwardMatchLength, curBackwardMatchLength,
+                       curTotalMatchLength;
+                if (cur->checksum != checksum || cur->offset <= lowestIndex) {
+                    continue;
+                }
+                if (extDict) {
+                    BYTE const* const curMatchBase =
+                        cur->offset < dictLimit ? dictBase : base;
+                    BYTE const* const pMatch = curMatchBase + cur->offset;
+                    BYTE const* const matchEnd =
+                        cur->offset < dictLimit ? dictEnd : iend;
+                    BYTE const* const lowMatchPtr =
+                        cur->offset < dictLimit ? dictStart : lowPrefixPtr;
+                    curForwardMatchLength =
+                        ZSTD_count_2segments(split, pMatch, iend, matchEnd, lowPrefixPtr);
+                    if (curForwardMatchLength < minMatchLength) {
+                        continue;
+                    }
+                    curBackwardMatchLength = ZSTD_ldm_countBackwardsMatch_2segments(
+                            split, anchor, pMatch, lowMatchPtr, dictStart, dictEnd);
+                } else { /* !extDict */
+                    BYTE const* const pMatch = base + cur->offset;
+                    curForwardMatchLength = ZSTD_count(split, pMatch, iend);
+                    if (curForwardMatchLength < minMatchLength) {
+                        continue;
+                    }
+                    curBackwardMatchLength =
+                        ZSTD_ldm_countBackwardsMatch(split, anchor, pMatch, lowPrefixPtr);
+                }
+                curTotalMatchLength = curForwardMatchLength + curBackwardMatchLength;
+
+                if (curTotalMatchLength > bestMatchLength) {
+                    bestMatchLength = curTotalMatchLength;
+                    forwardMatchLength = curForwardMatchLength;
+                    backwardMatchLength = curBackwardMatchLength;
+                    bestEntry = cur;
+                }
+            }
+
+            /* No match found -- insert an entry into the hash table
+             * and process the next candidate match */
+            if (bestEntry == NULL) {
+                ZSTD_ldm_insertEntry(ldmState, hash, newEntry, *params);
+                continue;
+            }
+
+            /* Match found */
+            mLength = forwardMatchLength + backwardMatchLength;
+            {
+                U32 const offset = (U32)(split - base) - bestEntry->offset;
+                rawSeq* const seq = rawSeqStore->seq + rawSeqStore->size;
+
+                /* Out of sequence storage */
+                if (rawSeqStore->size == rawSeqStore->capacity)
+                    return ERROR(dstSize_tooSmall);
+                seq->litLength = (U32)(split - backwardMatchLength - anchor);
+                seq->matchLength = (U32)mLength;
+                seq->offset = offset;
+                rawSeqStore->size++;
+            }
+
+            /* Insert the current entry into the hash table --- it must be
+             * done after the previous block to avoid clobbering bestEntry */
+            ZSTD_ldm_insertEntry(ldmState, hash, newEntry, *params);
+
+            anchor = split + forwardMatchLength;
+        }
+
+        ip += hashed;
+    }
+
+    return iend - anchor;
+}
+
+/*! ZSTD_ldm_reduceTable() :
+ *  reduce table indexes by `reducerValue` */
+static void ZSTD_ldm_reduceTable(ldmEntry_t* const table, U32 const size,
+                                 U32 const reducerValue)
+{
+    U32 u;
+    for (u = 0; u < size; u++) {
+        if (table[u].offset < reducerValue) table[u].offset = 0;
+        else table[u].offset -= reducerValue;
+    }
+}
+
+size_t ZSTD_ldm_generateSequences(
+        ldmState_t* ldmState, rawSeqStore_t* sequences,
+        ldmParams_t const* params, void const* src, size_t srcSize)
+{
+    U32 const maxDist = 1U << params->windowLog;
+    BYTE const* const istart = (BYTE const*)src;
+    BYTE const* const iend = istart + srcSize;
+    size_t const kMaxChunkSize = 1 << 20;
+    size_t const nbChunks = (srcSize / kMaxChunkSize) + ((srcSize % kMaxChunkSize) != 0);
+    size_t chunk;
+    size_t leftoverSize = 0;
+
+    assert(ZSTD_CHUNKSIZE_MAX >= kMaxChunkSize);
+    /* Check that ZSTD_window_update() has been called for this chunk prior
+     * to passing it to this function.
+     */
+    assert(ldmState->window.nextSrc >= (BYTE const*)src + srcSize);
+    /* The input could be very large (in zstdmt), so it must be broken up into
+     * chunks to enforce the maximum distance and handle overflow correction.
+     */
+    assert(sequences->pos <= sequences->size);
+    assert(sequences->size <= sequences->capacity);
+    for (chunk = 0; chunk < nbChunks && sequences->size < sequences->capacity; ++chunk) {
+        BYTE const* const chunkStart = istart + chunk * kMaxChunkSize;
+        size_t const remaining = (size_t)(iend - chunkStart);
+        BYTE const *const chunkEnd =
+            (remaining < kMaxChunkSize) ? iend : chunkStart + kMaxChunkSize;
+        size_t const chunkSize = chunkEnd - chunkStart;
+        size_t newLeftoverSize;
+        size_t const prevSize = sequences->size;
+
+        assert(chunkStart < iend);
+        /* 1. Perform overflow correction if necessary. */
+        if (ZSTD_window_needOverflowCorrection(ldmState->window, chunkEnd)) {
+            U32 const ldmHSize = 1U << params->hashLog;
+            U32 const correction = ZSTD_window_correctOverflow(
+                &ldmState->window, /* cycleLog */ 0, maxDist, chunkStart);
+            ZSTD_ldm_reduceTable(ldmState->hashTable, ldmHSize, correction);
+            /* invalidate dictionaries on overflow correction */
+            ldmState->loadedDictEnd = 0;
+        }
+        /* 2. We enforce the maximum offset allowed.
+         *
+         * kMaxChunkSize should be small enough that we don't lose too much of
+         * the window through early invalidation.
+         * TODO: * Test the chunk size.
+         *       * Try invalidation after the sequence generation and test the
+         *         the offset against maxDist directly.
+         *
+         * NOTE: Because of dictionaries + sequence splitting we MUST make sure
+         * that any offset used is valid at the END of the sequence, since it may
+         * be split into two sequences. This condition holds when using
+         * ZSTD_window_enforceMaxDist(), but if we move to checking offsets
+         * against maxDist directly, we'll have to carefully handle that case.
+         */
+        ZSTD_window_enforceMaxDist(&ldmState->window, chunkEnd, maxDist, &ldmState->loadedDictEnd, NULL);
+        /* 3. Generate the sequences for the chunk, and get newLeftoverSize. */
+        newLeftoverSize = ZSTD_ldm_generateSequences_internal(
+            ldmState, sequences, params, chunkStart, chunkSize);
+        if (ZSTD_isError(newLeftoverSize))
+            return newLeftoverSize;
+        /* 4. We add the leftover literals from previous iterations to the first
+         *    newly generated sequence, or add the `newLeftoverSize` if none are
+         *    generated.
+         */
+        /* Prepend the leftover literals from the last call */
+        if (prevSize < sequences->size) {
+            sequences->seq[prevSize].litLength += (U32)leftoverSize;
+            leftoverSize = newLeftoverSize;
+        } else {
+            assert(newLeftoverSize == chunkSize);
+            leftoverSize += chunkSize;
+        }
+    }
+    return 0;
+}
+
+void ZSTD_ldm_skipSequences(rawSeqStore_t* rawSeqStore, size_t srcSize, U32 const minMatch) {
+    while (srcSize > 0 && rawSeqStore->pos < rawSeqStore->size) {
+        rawSeq* seq = rawSeqStore->seq + rawSeqStore->pos;
+        if (srcSize <= seq->litLength) {
+            /* Skip past srcSize literals */
+            seq->litLength -= (U32)srcSize;
+            return;
+        }
+        srcSize -= seq->litLength;
+        seq->litLength = 0;
+        if (srcSize < seq->matchLength) {
+            /* Skip past the first srcSize of the match */
+            seq->matchLength -= (U32)srcSize;
+            if (seq->matchLength < minMatch) {
+                /* The match is too short, omit it */
+                if (rawSeqStore->pos + 1 < rawSeqStore->size) {
+                    seq[1].litLength += seq[0].matchLength;
+                }
+                rawSeqStore->pos++;
+            }
+            return;
+        }
+        srcSize -= seq->matchLength;
+        seq->matchLength = 0;
+        rawSeqStore->pos++;
+    }
+}
+
+/*
+ * If the sequence length is longer than remaining then the sequence is split
+ * between this block and the next.
+ *
+ * Returns the current sequence to handle, or if the rest of the block should
+ * be literals, it returns a sequence with offset == 0.
+ */
+static rawSeq maybeSplitSequence(rawSeqStore_t* rawSeqStore,
+                                 U32 const remaining, U32 const minMatch)
+{
+    rawSeq sequence = rawSeqStore->seq[rawSeqStore->pos];
+    assert(sequence.offset > 0);
+    /* Likely: No partial sequence */
+    if (remaining >= sequence.litLength + sequence.matchLength) {
+        rawSeqStore->pos++;
+        return sequence;
+    }
+    /* Cut the sequence short (offset == 0 ==> rest is literals). */
+    if (remaining <= sequence.litLength) {
+        sequence.offset = 0;
+    } else if (remaining < sequence.litLength + sequence.matchLength) {
+        sequence.matchLength = remaining - sequence.litLength;
+        if (sequence.matchLength < minMatch) {
+            sequence.offset = 0;
+        }
+    }
+    /* Skip past `remaining` bytes for the future sequences. */
+    ZSTD_ldm_skipSequences(rawSeqStore, remaining, minMatch);
+    return sequence;
+}
+
+void ZSTD_ldm_skipRawSeqStoreBytes(rawSeqStore_t* rawSeqStore, size_t nbBytes) {
+    U32 currPos = (U32)(rawSeqStore->posInSequence + nbBytes);
+    while (currPos && rawSeqStore->pos < rawSeqStore->size) {
+        rawSeq currSeq = rawSeqStore->seq[rawSeqStore->pos];
+        if (currPos >= currSeq.litLength + currSeq.matchLength) {
+            currPos -= currSeq.litLength + currSeq.matchLength;
+            rawSeqStore->pos++;
+        } else {
+            rawSeqStore->posInSequence = currPos;
+            break;
+        }
+    }
+    if (currPos == 0 || rawSeqStore->pos == rawSeqStore->size) {
+        rawSeqStore->posInSequence = 0;
+    }
+}
+
+size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
+    ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+    void const* src, size_t srcSize)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    unsigned const minMatch = cParams->minMatch;
+    ZSTD_blockCompressor const blockCompressor =
+        ZSTD_selectBlockCompressor(cParams->strategy, ZSTD_matchState_dictMode(ms));
+    /* Input bounds */
+    BYTE const* const istart = (BYTE const*)src;
+    BYTE const* const iend = istart + srcSize;
+    /* Input positions */
+    BYTE const* ip = istart;
+
+    DEBUGLOG(5, "ZSTD_ldm_blockCompress: srcSize=%zu", srcSize);
+    /* If using opt parser, use LDMs only as candidates rather than always accepting them */
+    if (cParams->strategy >= ZSTD_btopt) {
+        size_t lastLLSize;
+        ms->ldmSeqStore = rawSeqStore;
+        lastLLSize = blockCompressor(ms, seqStore, rep, src, srcSize);
+        ZSTD_ldm_skipRawSeqStoreBytes(rawSeqStore, srcSize);
+        return lastLLSize;
+    }
+
+    assert(rawSeqStore->pos <= rawSeqStore->size);
+    assert(rawSeqStore->size <= rawSeqStore->capacity);
+    /* Loop through each sequence and apply the block compressor to the literals */
+    while (rawSeqStore->pos < rawSeqStore->size && ip < iend) {
+        /* maybeSplitSequence updates rawSeqStore->pos */
+        rawSeq const sequence = maybeSplitSequence(rawSeqStore,
+                                                   (U32)(iend - ip), minMatch);
+        int i;
+        /* End signal */
+        if (sequence.offset == 0)
+            break;
+
+        assert(ip + sequence.litLength + sequence.matchLength <= iend);
+
+        /* Fill tables for block compressor */
+        ZSTD_ldm_limitTableUpdate(ms, ip);
+        ZSTD_ldm_fillFastTables(ms, ip);
+        /* Run the block compressor */
+        DEBUGLOG(5, "pos %u : calling block compressor on segment of size %u", (unsigned)(ip-istart), sequence.litLength);
+        {
+            size_t const newLitLength =
+                blockCompressor(ms, seqStore, rep, ip, sequence.litLength);
+            ip += sequence.litLength;
+            /* Update the repcodes */
+            for (i = ZSTD_REP_NUM - 1; i > 0; i--)
+                rep[i] = rep[i-1];
+            rep[0] = sequence.offset;
+            /* Store the sequence */
+            ZSTD_storeSeq(seqStore, newLitLength, ip - newLitLength, iend,
+                          sequence.offset + ZSTD_REP_MOVE,
+                          sequence.matchLength - MINMATCH);
+            ip += sequence.matchLength;
+        }
+    }
+    /* Fill the tables for the block compressor */
+    ZSTD_ldm_limitTableUpdate(ms, ip);
+    ZSTD_ldm_fillFastTables(ms, ip);
+    /* Compress the last literals */
+    return blockCompressor(ms, seqStore, rep, ip, iend - ip);
+}
diff --git a/lib/zstd/compress/zstd_ldm.h b/lib/zstd/compress/zstd_ldm.h
new file mode 100644
index 0000000..25b2527
--- /dev/null
+++ b/lib/zstd/compress/zstd_ldm.h
@@ -0,0 +1,110 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_LDM_H
+#define ZSTD_LDM_H
+
+
+#include "zstd_compress_internal.h"   /* ldmParams_t, U32 */
+#include <linux/zstd.h>   /* ZSTD_CCtx, size_t */
+
+/*-*************************************
+*  Long distance matching
+***************************************/
+
+#define ZSTD_LDM_DEFAULT_WINDOW_LOG ZSTD_WINDOWLOG_LIMIT_DEFAULT
+
+void ZSTD_ldm_fillHashTable(
+            ldmState_t* state, const BYTE* ip,
+            const BYTE* iend, ldmParams_t const* params);
+
+/*
+ * ZSTD_ldm_generateSequences():
+ *
+ * Generates the sequences using the long distance match finder.
+ * Generates long range matching sequences in `sequences`, which parse a prefix
+ * of the source. `sequences` must be large enough to store every sequence,
+ * which can be checked with `ZSTD_ldm_getMaxNbSeq()`.
+ * @returns 0 or an error code.
+ *
+ * NOTE: The user must have called ZSTD_window_update() for all of the input
+ * they have, even if they pass it to ZSTD_ldm_generateSequences() in chunks.
+ * NOTE: This function returns an error if it runs out of space to store
+ *       sequences.
+ */
+size_t ZSTD_ldm_generateSequences(
+            ldmState_t* ldms, rawSeqStore_t* sequences,
+            ldmParams_t const* params, void const* src, size_t srcSize);
+
+/*
+ * ZSTD_ldm_blockCompress():
+ *
+ * Compresses a block using the predefined sequences, along with a secondary
+ * block compressor. The literals section of every sequence is passed to the
+ * secondary block compressor, and those sequences are interspersed with the
+ * predefined sequences. Returns the length of the last literals.
+ * Updates `rawSeqStore.pos` to indicate how many sequences have been consumed.
+ * `rawSeqStore.seq` may also be updated to split the last sequence between two
+ * blocks.
+ * @return The length of the last literals.
+ *
+ * NOTE: The source must be at most the maximum block size, but the predefined
+ * sequences can be any size, and may be longer than the block. In the case that
+ * they are longer than the block, the last sequences may need to be split into
+ * two. We handle that case correctly, and update `rawSeqStore` appropriately.
+ * NOTE: This function does not return any errors.
+ */
+size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
+            ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+            void const* src, size_t srcSize);
+
+/*
+ * ZSTD_ldm_skipSequences():
+ *
+ * Skip past `srcSize` bytes worth of sequences in `rawSeqStore`.
+ * Avoids emitting matches less than `minMatch` bytes.
+ * Must be called for data that is not passed to ZSTD_ldm_blockCompress().
+ */
+void ZSTD_ldm_skipSequences(rawSeqStore_t* rawSeqStore, size_t srcSize,
+    U32 const minMatch);
+
+/* ZSTD_ldm_skipRawSeqStoreBytes():
+ * Moves forward in rawSeqStore by nbBytes, updating fields 'pos' and 'posInSequence'.
+ * Not to be used in conjunction with ZSTD_ldm_skipSequences().
+ * Must be called for data with is not passed to ZSTD_ldm_blockCompress().
+ */
+void ZSTD_ldm_skipRawSeqStoreBytes(rawSeqStore_t* rawSeqStore, size_t nbBytes);
+
+/* ZSTD_ldm_getTableSize() :
+ *  Estimate the space needed for long distance matching tables or 0 if LDM is
+ *  disabled.
+ */
+size_t ZSTD_ldm_getTableSize(ldmParams_t params);
+
+/* ZSTD_ldm_getSeqSpace() :
+ *  Return an upper bound on the number of sequences that can be produced by
+ *  the long distance matcher, or 0 if LDM is disabled.
+ */
+size_t ZSTD_ldm_getMaxNbSeq(ldmParams_t params, size_t maxChunkSize);
+
+/* ZSTD_ldm_adjustParameters() :
+ *  If the params->hashRateLog is not set, set it to its default value based on
+ *  windowLog and params->hashLog.
+ *
+ *  Ensures that params->bucketSizeLog is <= params->hashLog (setting it to
+ *  params->hashLog if it is not).
+ *
+ *  Ensures that the minMatchLength >= targetLength during optimal parsing.
+ */
+void ZSTD_ldm_adjustParameters(ldmParams_t* params,
+                               ZSTD_compressionParameters const* cParams);
+
+
+#endif /* ZSTD_FAST_H */
diff --git a/lib/zstd/compress/zstd_ldm_geartab.h b/lib/zstd/compress/zstd_ldm_geartab.h
new file mode 100644
index 0000000..e5c24d8
--- /dev/null
+++ b/lib/zstd/compress/zstd_ldm_geartab.h
@@ -0,0 +1,103 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_LDM_GEARTAB_H
+#define ZSTD_LDM_GEARTAB_H
+
+static U64 ZSTD_ldm_gearTab[256] = {
+    0xf5b8f72c5f77775c, 0x84935f266b7ac412, 0xb647ada9ca730ccc,
+    0xb065bb4b114fb1de, 0x34584e7e8c3a9fd0, 0x4e97e17c6ae26b05,
+    0x3a03d743bc99a604, 0xcecd042422c4044f, 0x76de76c58524259e,
+    0x9c8528f65badeaca, 0x86563706e2097529, 0x2902475fa375d889,
+    0xafb32a9739a5ebe6, 0xce2714da3883e639, 0x21eaf821722e69e,
+    0x37b628620b628,    0x49a8d455d88caf5,  0x8556d711e6958140,
+    0x4f7ae74fc605c1f,  0x829f0c3468bd3a20, 0x4ffdc885c625179e,
+    0x8473de048a3daf1b, 0x51008822b05646b2, 0x69d75d12b2d1cc5f,
+    0x8c9d4a19159154bc, 0xc3cc10f4abbd4003, 0xd06ddc1cecb97391,
+    0xbe48e6e7ed80302e, 0x3481db31cee03547, 0xacc3f67cdaa1d210,
+    0x65cb771d8c7f96cc, 0x8eb27177055723dd, 0xc789950d44cd94be,
+    0x934feadc3700b12b, 0x5e485f11edbdf182, 0x1e2e2a46fd64767a,
+    0x2969ca71d82efa7c, 0x9d46e9935ebbba2e, 0xe056b67e05e6822b,
+    0x94d73f55739d03a0, 0xcd7010bdb69b5a03, 0x455ef9fcd79b82f4,
+    0x869cb54a8749c161, 0x38d1a4fa6185d225, 0xb475166f94bbe9bb,
+    0xa4143548720959f1, 0x7aed4780ba6b26ba, 0xd0ce264439e02312,
+    0x84366d746078d508, 0xa8ce973c72ed17be, 0x21c323a29a430b01,
+    0x9962d617e3af80ee, 0xab0ce91d9c8cf75b, 0x530e8ee6d19a4dbc,
+    0x2ef68c0cf53f5d72, 0xc03a681640a85506, 0x496e4e9f9c310967,
+    0x78580472b59b14a0, 0x273824c23b388577, 0x66bf923ad45cb553,
+    0x47ae1a5a2492ba86, 0x35e304569e229659, 0x4765182a46870b6f,
+    0x6cbab625e9099412, 0xddac9a2e598522c1, 0x7172086e666624f2,
+    0xdf5003ca503b7837, 0x88c0c1db78563d09, 0x58d51865acfc289d,
+    0x177671aec65224f1, 0xfb79d8a241e967d7, 0x2be1e101cad9a49a,
+    0x6625682f6e29186b, 0x399553457ac06e50, 0x35dffb4c23abb74,
+    0x429db2591f54aade, 0xc52802a8037d1009, 0x6acb27381f0b25f3,
+    0xf45e2551ee4f823b, 0x8b0ea2d99580c2f7, 0x3bed519cbcb4e1e1,
+    0xff452823dbb010a,  0x9d42ed614f3dd267, 0x5b9313c06257c57b,
+    0xa114b8008b5e1442, 0xc1fe311c11c13d4b, 0x66e8763ea34c5568,
+    0x8b982af1c262f05d, 0xee8876faaa75fbb7, 0x8a62a4d0d172bb2a,
+    0xc13d94a3b7449a97, 0x6dbbba9dc15d037c, 0xc786101f1d92e0f1,
+    0xd78681a907a0b79b, 0xf61aaf2962c9abb9, 0x2cfd16fcd3cb7ad9,
+    0x868c5b6744624d21, 0x25e650899c74ddd7, 0xba042af4a7c37463,
+    0x4eb1a539465a3eca, 0xbe09dbf03b05d5ca, 0x774e5a362b5472ba,
+    0x47a1221229d183cd, 0x504b0ca18ef5a2df, 0xdffbdfbde2456eb9,
+    0x46cd2b2fbee34634, 0xf2aef8fe819d98c3, 0x357f5276d4599d61,
+    0x24a5483879c453e3, 0x88026889192b4b9,  0x28da96671782dbec,
+    0x4ef37c40588e9aaa, 0x8837b90651bc9fb3, 0xc164f741d3f0e5d6,
+    0xbc135a0a704b70ba, 0x69cd868f7622ada,  0xbc37ba89e0b9c0ab,
+    0x47c14a01323552f6, 0x4f00794bacee98bb, 0x7107de7d637a69d5,
+    0x88af793bb6f2255e, 0xf3c6466b8799b598, 0xc288c616aa7f3b59,
+    0x81ca63cf42fca3fd, 0x88d85ace36a2674b, 0xd056bd3792389e7,
+    0xe55c396c4e9dd32d, 0xbefb504571e6c0a6, 0x96ab32115e91e8cc,
+    0xbf8acb18de8f38d1, 0x66dae58801672606, 0x833b6017872317fb,
+    0xb87c16f2d1c92864, 0xdb766a74e58b669c, 0x89659f85c61417be,
+    0xc8daad856011ea0c, 0x76a4b565b6fe7eae, 0xa469d085f6237312,
+    0xaaf0365683a3e96c, 0x4dbb746f8424f7b8, 0x638755af4e4acc1,
+    0x3d7807f5bde64486, 0x17be6d8f5bbb7639, 0x903f0cd44dc35dc,
+    0x67b672eafdf1196c, 0xa676ff93ed4c82f1, 0x521d1004c5053d9d,
+    0x37ba9ad09ccc9202, 0x84e54d297aacfb51, 0xa0b4b776a143445,
+    0x820d471e20b348e,  0x1874383cb83d46dc, 0x97edeec7a1efe11c,
+    0xb330e50b1bdc42aa, 0x1dd91955ce70e032, 0xa514cdb88f2939d5,
+    0x2791233fd90db9d3, 0x7b670a4cc50f7a9b, 0x77c07d2a05c6dfa5,
+    0xe3778b6646d0a6fa, 0xb39c8eda47b56749, 0x933ed448addbef28,
+    0xaf846af6ab7d0bf4, 0xe5af208eb666e49,  0x5e6622f73534cd6a,
+    0x297daeca42ef5b6e, 0x862daef3d35539a6, 0xe68722498f8e1ea9,
+    0x981c53093dc0d572, 0xfa09b0bfbf86fbf5, 0x30b1e96166219f15,
+    0x70e7d466bdc4fb83, 0x5a66736e35f2a8e9, 0xcddb59d2b7c1baef,
+    0xd6c7d247d26d8996, 0xea4e39eac8de1ba3, 0x539c8bb19fa3aff2,
+    0x9f90e4c5fd508d8,  0xa34e5956fbaf3385, 0x2e2f8e151d3ef375,
+    0x173691e9b83faec1, 0xb85a8d56bf016379, 0x8382381267408ae3,
+    0xb90f901bbdc0096d, 0x7c6ad32933bcec65, 0x76bb5e2f2c8ad595,
+    0x390f851a6cf46d28, 0xc3e6064da1c2da72, 0xc52a0c101cfa5389,
+    0xd78eaf84a3fbc530, 0x3781b9e2288b997e, 0x73c2f6dea83d05c4,
+    0x4228e364c5b5ed7,  0x9d7a3edf0da43911, 0x8edcfeda24686756,
+    0x5e7667a7b7a9b3a1, 0x4c4f389fa143791d, 0xb08bc1023da7cddc,
+    0x7ab4be3ae529b1cc, 0x754e6132dbe74ff9, 0x71635442a839df45,
+    0x2f6fb1643fbe52de, 0x961e0a42cf7a8177, 0xf3b45d83d89ef2ea,
+    0xee3de4cf4a6e3e9b, 0xcd6848542c3295e7, 0xe4cee1664c78662f,
+    0x9947548b474c68c4, 0x25d73777a5ed8b0b, 0xc915b1d636b7fc,
+    0x21c2ba75d9b0d2da, 0x5f6b5dcf608a64a1, 0xdcf333255ff9570c,
+    0x633b922418ced4ee, 0xc136dde0b004b34a, 0x58cc83b05d4b2f5a,
+    0x5eb424dda28e42d2, 0x62df47369739cd98, 0xb4e0b42485e4ce17,
+    0x16e1f0c1f9a8d1e7, 0x8ec3916707560ebf, 0x62ba6e2df2cc9db3,
+    0xcbf9f4ff77d83a16, 0x78d9d7d07d2bbcc4, 0xef554ce1e02c41f4,
+    0x8d7581127eccf94d, 0xa9b53336cb3c8a05, 0x38c42c0bf45c4f91,
+    0x640893cdf4488863, 0x80ec34bc575ea568, 0x39f324f5b48eaa40,
+    0xe9d9ed1f8eff527f, 0x9224fc058cc5a214, 0xbaba00b04cfe7741,
+    0x309a9f120fcf52af, 0xa558f3ec65626212, 0x424bec8b7adabe2f,
+    0x41622513a6aea433, 0xb88da2d5324ca798, 0xd287733b245528a4,
+    0x9a44697e6d68aec3, 0x7b1093be2f49bb28, 0x50bbec632e3d8aad,
+    0x6cd90723e1ea8283, 0x897b9e7431b02bf3, 0x219efdcb338a7047,
+    0x3b0311f0a27c0656, 0xdb17bf91c0db96e7, 0x8cd4fd6b4e85a5b2,
+    0xfab071054ba6409d, 0x40d6fe831fa9dfd9, 0xaf358debad7d791e,
+    0xeb8d0e25a65e3e58, 0xbbcbd3df14e08580, 0xcf751f27ecdab2b,
+    0x2b4da14f2613d8f4
+};
+
+#endif /* ZSTD_LDM_GEARTAB_H */
diff --git a/lib/zstd/compress/zstd_opt.c b/lib/zstd/compress/zstd_opt.c
new file mode 100644
index 0000000..0433705
--- /dev/null
+++ b/lib/zstd/compress/zstd_opt.c
@@ -0,0 +1,1346 @@
+/*
+ * Copyright (c) Przemyslaw Skibinski, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#include "zstd_compress_internal.h"
+#include "hist.h"
+#include "zstd_opt.h"
+
+
+#define ZSTD_LITFREQ_ADD    2   /* scaling factor for litFreq, so that frequencies adapt faster to new stats */
+#define ZSTD_FREQ_DIV       4   /* log factor when using previous stats to init next stats */
+#define ZSTD_MAX_PRICE     (1<<30)
+
+#define ZSTD_PREDEF_THRESHOLD 1024   /* if srcSize < ZSTD_PREDEF_THRESHOLD, symbols' cost is assumed static, directly determined by pre-defined distributions */
+
+
+/*-*************************************
+*  Price functions for optimal parser
+***************************************/
+
+#if 0    /* approximation at bit level */
+#  define BITCOST_ACCURACY 0
+#  define BITCOST_MULTIPLIER (1 << BITCOST_ACCURACY)
+#  define WEIGHT(stat)  ((void)opt, ZSTD_bitWeight(stat))
+#elif 0  /* fractional bit accuracy */
+#  define BITCOST_ACCURACY 8
+#  define BITCOST_MULTIPLIER (1 << BITCOST_ACCURACY)
+#  define WEIGHT(stat,opt) ((void)opt, ZSTD_fracWeight(stat))
+#else    /* opt==approx, ultra==accurate */
+#  define BITCOST_ACCURACY 8
+#  define BITCOST_MULTIPLIER (1 << BITCOST_ACCURACY)
+#  define WEIGHT(stat,opt) (opt ? ZSTD_fracWeight(stat) : ZSTD_bitWeight(stat))
+#endif
+
+MEM_STATIC U32 ZSTD_bitWeight(U32 stat)
+{
+    return (ZSTD_highbit32(stat+1) * BITCOST_MULTIPLIER);
+}
+
+MEM_STATIC U32 ZSTD_fracWeight(U32 rawStat)
+{
+    U32 const stat = rawStat + 1;
+    U32 const hb = ZSTD_highbit32(stat);
+    U32 const BWeight = hb * BITCOST_MULTIPLIER;
+    U32 const FWeight = (stat << BITCOST_ACCURACY) >> hb;
+    U32 const weight = BWeight + FWeight;
+    assert(hb + BITCOST_ACCURACY < 31);
+    return weight;
+}
+
+#if (DEBUGLEVEL>=2)
+/* debugging function,
+ * @return price in bytes as fractional value
+ * for debug messages only */
+MEM_STATIC double ZSTD_fCost(U32 price)
+{
+    return (double)price / (BITCOST_MULTIPLIER*8);
+}
+#endif
+
+static int ZSTD_compressedLiterals(optState_t const* const optPtr)
+{
+    return optPtr->literalCompressionMode != ZSTD_lcm_uncompressed;
+}
+
+static void ZSTD_setBasePrices(optState_t* optPtr, int optLevel)
+{
+    if (ZSTD_compressedLiterals(optPtr))
+        optPtr->litSumBasePrice = WEIGHT(optPtr->litSum, optLevel);
+    optPtr->litLengthSumBasePrice = WEIGHT(optPtr->litLengthSum, optLevel);
+    optPtr->matchLengthSumBasePrice = WEIGHT(optPtr->matchLengthSum, optLevel);
+    optPtr->offCodeSumBasePrice = WEIGHT(optPtr->offCodeSum, optLevel);
+}
+
+
+/* ZSTD_downscaleStat() :
+ * reduce all elements in table by a factor 2^(ZSTD_FREQ_DIV+malus)
+ * return the resulting sum of elements */
+static U32 ZSTD_downscaleStat(unsigned* table, U32 lastEltIndex, int malus)
+{
+    U32 s, sum=0;
+    DEBUGLOG(5, "ZSTD_downscaleStat (nbElts=%u)", (unsigned)lastEltIndex+1);
+    assert(ZSTD_FREQ_DIV+malus > 0 && ZSTD_FREQ_DIV+malus < 31);
+    for (s=0; s<lastEltIndex+1; s++) {
+        table[s] = 1 + (table[s] >> (ZSTD_FREQ_DIV+malus));
+        sum += table[s];
+    }
+    return sum;
+}
+
+/* ZSTD_rescaleFreqs() :
+ * if first block (detected by optPtr->litLengthSum == 0) : init statistics
+ *    take hints from dictionary if there is one
+ *    or init from zero, using src for literals stats, or flat 1 for match symbols
+ * otherwise downscale existing stats, to be used as seed for next block.
+ */
+static void
+ZSTD_rescaleFreqs(optState_t* const optPtr,
+            const BYTE* const src, size_t const srcSize,
+                  int const optLevel)
+{
+    int const compressedLiterals = ZSTD_compressedLiterals(optPtr);
+    DEBUGLOG(5, "ZSTD_rescaleFreqs (srcSize=%u)", (unsigned)srcSize);
+    optPtr->priceType = zop_dynamic;
+
+    if (optPtr->litLengthSum == 0) {  /* first block : init */
+        if (srcSize <= ZSTD_PREDEF_THRESHOLD) {  /* heuristic */
+            DEBUGLOG(5, "(srcSize <= ZSTD_PREDEF_THRESHOLD) => zop_predef");
+            optPtr->priceType = zop_predef;
+        }
+
+        assert(optPtr->symbolCosts != NULL);
+        if (optPtr->symbolCosts->huf.repeatMode == HUF_repeat_valid) {
+            /* huffman table presumed generated by dictionary */
+            optPtr->priceType = zop_dynamic;
+
+            if (compressedLiterals) {
+                unsigned lit;
+                assert(optPtr->litFreq != NULL);
+                optPtr->litSum = 0;
+                for (lit=0; lit<=MaxLit; lit++) {
+                    U32 const scaleLog = 11;   /* scale to 2K */
+                    U32 const bitCost = HUF_getNbBits(optPtr->symbolCosts->huf.CTable, lit);
+                    assert(bitCost <= scaleLog);
+                    optPtr->litFreq[lit] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/;
+                    optPtr->litSum += optPtr->litFreq[lit];
+            }   }
+
+            {   unsigned ll;
+                FSE_CState_t llstate;
+                FSE_initCState(&llstate, optPtr->symbolCosts->fse.litlengthCTable);
+                optPtr->litLengthSum = 0;
+                for (ll=0; ll<=MaxLL; ll++) {
+                    U32 const scaleLog = 10;   /* scale to 1K */
+                    U32 const bitCost = FSE_getMaxNbBits(llstate.symbolTT, ll);
+                    assert(bitCost < scaleLog);
+                    optPtr->litLengthFreq[ll] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/;
+                    optPtr->litLengthSum += optPtr->litLengthFreq[ll];
+            }   }
+
+            {   unsigned ml;
+                FSE_CState_t mlstate;
+                FSE_initCState(&mlstate, optPtr->symbolCosts->fse.matchlengthCTable);
+                optPtr->matchLengthSum = 0;
+                for (ml=0; ml<=MaxML; ml++) {
+                    U32 const scaleLog = 10;
+                    U32 const bitCost = FSE_getMaxNbBits(mlstate.symbolTT, ml);
+                    assert(bitCost < scaleLog);
+                    optPtr->matchLengthFreq[ml] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/;
+                    optPtr->matchLengthSum += optPtr->matchLengthFreq[ml];
+            }   }
+
+            {   unsigned of;
+                FSE_CState_t ofstate;
+                FSE_initCState(&ofstate, optPtr->symbolCosts->fse.offcodeCTable);
+                optPtr->offCodeSum = 0;
+                for (of=0; of<=MaxOff; of++) {
+                    U32 const scaleLog = 10;
+                    U32 const bitCost = FSE_getMaxNbBits(ofstate.symbolTT, of);
+                    assert(bitCost < scaleLog);
+                    optPtr->offCodeFreq[of] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/;
+                    optPtr->offCodeSum += optPtr->offCodeFreq[of];
+            }   }
+
+        } else {  /* not a dictionary */
+
+            assert(optPtr->litFreq != NULL);
+            if (compressedLiterals) {
+                unsigned lit = MaxLit;
+                HIST_count_simple(optPtr->litFreq, &lit, src, srcSize);   /* use raw first block to init statistics */
+                optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1);
+            }
+
+            {   unsigned ll;
+                for (ll=0; ll<=MaxLL; ll++)
+                    optPtr->litLengthFreq[ll] = 1;
+            }
+            optPtr->litLengthSum = MaxLL+1;
+
+            {   unsigned ml;
+                for (ml=0; ml<=MaxML; ml++)
+                    optPtr->matchLengthFreq[ml] = 1;
+            }
+            optPtr->matchLengthSum = MaxML+1;
+
+            {   unsigned of;
+                for (of=0; of<=MaxOff; of++)
+                    optPtr->offCodeFreq[of] = 1;
+            }
+            optPtr->offCodeSum = MaxOff+1;
+
+        }
+
+    } else {   /* new block : re-use previous statistics, scaled down */
+
+        if (compressedLiterals)
+            optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1);
+        optPtr->litLengthSum = ZSTD_downscaleStat(optPtr->litLengthFreq, MaxLL, 0);
+        optPtr->matchLengthSum = ZSTD_downscaleStat(optPtr->matchLengthFreq, MaxML, 0);
+        optPtr->offCodeSum = ZSTD_downscaleStat(optPtr->offCodeFreq, MaxOff, 0);
+    }
+
+    ZSTD_setBasePrices(optPtr, optLevel);
+}
+
+/* ZSTD_rawLiteralsCost() :
+ * price of literals (only) in specified segment (which length can be 0).
+ * does not include price of literalLength symbol */
+static U32 ZSTD_rawLiteralsCost(const BYTE* const literals, U32 const litLength,
+                                const optState_t* const optPtr,
+                                int optLevel)
+{
+    if (litLength == 0) return 0;
+
+    if (!ZSTD_compressedLiterals(optPtr))
+        return (litLength << 3) * BITCOST_MULTIPLIER;  /* Uncompressed - 8 bytes per literal. */
+
+    if (optPtr->priceType == zop_predef)
+        return (litLength*6) * BITCOST_MULTIPLIER;  /* 6 bit per literal - no statistic used */
+
+    /* dynamic statistics */
+    {   U32 price = litLength * optPtr->litSumBasePrice;
+        U32 u;
+        for (u=0; u < litLength; u++) {
+            assert(WEIGHT(optPtr->litFreq[literals[u]], optLevel) <= optPtr->litSumBasePrice);   /* literal cost should never be negative */
+            price -= WEIGHT(optPtr->litFreq[literals[u]], optLevel);
+        }
+        return price;
+    }
+}
+
+/* ZSTD_litLengthPrice() :
+ * cost of literalLength symbol */
+static U32 ZSTD_litLengthPrice(U32 const litLength, const optState_t* const optPtr, int optLevel)
+{
+    if (optPtr->priceType == zop_predef) return WEIGHT(litLength, optLevel);
+
+    /* dynamic statistics */
+    {   U32 const llCode = ZSTD_LLcode(litLength);
+        return (LL_bits[llCode] * BITCOST_MULTIPLIER)
+             + optPtr->litLengthSumBasePrice
+             - WEIGHT(optPtr->litLengthFreq[llCode], optLevel);
+    }
+}
+
+/* ZSTD_getMatchPrice() :
+ * Provides the cost of the match part (offset + matchLength) of a sequence
+ * Must be combined with ZSTD_fullLiteralsCost() to get the full cost of a sequence.
+ * optLevel: when <2, favors small offset for decompression speed (improved cache efficiency) */
+FORCE_INLINE_TEMPLATE U32
+ZSTD_getMatchPrice(U32 const offset,
+                   U32 const matchLength,
+             const optState_t* const optPtr,
+                   int const optLevel)
+{
+    U32 price;
+    U32 const offCode = ZSTD_highbit32(offset+1);
+    U32 const mlBase = matchLength - MINMATCH;
+    assert(matchLength >= MINMATCH);
+
+    if (optPtr->priceType == zop_predef)  /* fixed scheme, do not use statistics */
+        return WEIGHT(mlBase, optLevel) + ((16 + offCode) * BITCOST_MULTIPLIER);
+
+    /* dynamic statistics */
+    price = (offCode * BITCOST_MULTIPLIER) + (optPtr->offCodeSumBasePrice - WEIGHT(optPtr->offCodeFreq[offCode], optLevel));
+    if ((optLevel<2) /*static*/ && offCode >= 20)
+        price += (offCode-19)*2 * BITCOST_MULTIPLIER; /* handicap for long distance offsets, favor decompression speed */
+
+    /* match Length */
+    {   U32 const mlCode = ZSTD_MLcode(mlBase);
+        price += (ML_bits[mlCode] * BITCOST_MULTIPLIER) + (optPtr->matchLengthSumBasePrice - WEIGHT(optPtr->matchLengthFreq[mlCode], optLevel));
+    }
+
+    price += BITCOST_MULTIPLIER / 5;   /* heuristic : make matches a bit more costly to favor less sequences -> faster decompression speed */
+
+    DEBUGLOG(8, "ZSTD_getMatchPrice(ml:%u) = %u", matchLength, price);
+    return price;
+}
+
+/* ZSTD_updateStats() :
+ * assumption : literals + litLengtn <= iend */
+static void ZSTD_updateStats(optState_t* const optPtr,
+                             U32 litLength, const BYTE* literals,
+                             U32 offsetCode, U32 matchLength)
+{
+    /* literals */
+    if (ZSTD_compressedLiterals(optPtr)) {
+        U32 u;
+        for (u=0; u < litLength; u++)
+            optPtr->litFreq[literals[u]] += ZSTD_LITFREQ_ADD;
+        optPtr->litSum += litLength*ZSTD_LITFREQ_ADD;
+    }
+
+    /* literal Length */
+    {   U32 const llCode = ZSTD_LLcode(litLength);
+        optPtr->litLengthFreq[llCode]++;
+        optPtr->litLengthSum++;
+    }
+
+    /* match offset code (0-2=>repCode; 3+=>offset+2) */
+    {   U32 const offCode = ZSTD_highbit32(offsetCode+1);
+        assert(offCode <= MaxOff);
+        optPtr->offCodeFreq[offCode]++;
+        optPtr->offCodeSum++;
+    }
+
+    /* match Length */
+    {   U32 const mlBase = matchLength - MINMATCH;
+        U32 const mlCode = ZSTD_MLcode(mlBase);
+        optPtr->matchLengthFreq[mlCode]++;
+        optPtr->matchLengthSum++;
+    }
+}
+
+
+/* ZSTD_readMINMATCH() :
+ * function safe only for comparisons
+ * assumption : memPtr must be at least 4 bytes before end of buffer */
+MEM_STATIC U32 ZSTD_readMINMATCH(const void* memPtr, U32 length)
+{
+    switch (length)
+    {
+    default :
+    case 4 : return MEM_read32(memPtr);
+    case 3 : if (MEM_isLittleEndian())
+                return MEM_read32(memPtr)<<8;
+             else
+                return MEM_read32(memPtr)>>8;
+    }
+}
+
+
+/* Update hashTable3 up to ip (excluded)
+   Assumption : always within prefix (i.e. not within extDict) */
+static U32 ZSTD_insertAndFindFirstIndexHash3 (ZSTD_matchState_t* ms,
+                                              U32* nextToUpdate3,
+                                              const BYTE* const ip)
+{
+    U32* const hashTable3 = ms->hashTable3;
+    U32 const hashLog3 = ms->hashLog3;
+    const BYTE* const base = ms->window.base;
+    U32 idx = *nextToUpdate3;
+    U32 const target = (U32)(ip - base);
+    size_t const hash3 = ZSTD_hash3Ptr(ip, hashLog3);
+    assert(hashLog3 > 0);
+
+    while(idx < target) {
+        hashTable3[ZSTD_hash3Ptr(base+idx, hashLog3)] = idx;
+        idx++;
+    }
+
+    *nextToUpdate3 = target;
+    return hashTable3[hash3];
+}
+
+
+/*-*************************************
+*  Binary Tree search
+***************************************/
+/* ZSTD_insertBt1() : add one or multiple positions to tree.
+ *  ip : assumed <= iend-8 .
+ * @return : nb of positions added */
+static U32 ZSTD_insertBt1(
+                ZSTD_matchState_t* ms,
+                const BYTE* const ip, const BYTE* const iend,
+                U32 const mls, const int extDict)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32*   const hashTable = ms->hashTable;
+    U32    const hashLog = cParams->hashLog;
+    size_t const h  = ZSTD_hashPtr(ip, hashLog, mls);
+    U32*   const bt = ms->chainTable;
+    U32    const btLog  = cParams->chainLog - 1;
+    U32    const btMask = (1 << btLog) - 1;
+    U32 matchIndex = hashTable[h];
+    size_t commonLengthSmaller=0, commonLengthLarger=0;
+    const BYTE* const base = ms->window.base;
+    const BYTE* const dictBase = ms->window.dictBase;
+    const U32 dictLimit = ms->window.dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    const BYTE* match;
+    const U32 curr = (U32)(ip-base);
+    const U32 btLow = btMask >= curr ? 0 : curr - btMask;
+    U32* smallerPtr = bt + 2*(curr&btMask);
+    U32* largerPtr  = smallerPtr + 1;
+    U32 dummy32;   /* to be nullified at the end */
+    U32 const windowLow = ms->window.lowLimit;
+    U32 matchEndIdx = curr+8+1;
+    size_t bestLength = 8;
+    U32 nbCompares = 1U << cParams->searchLog;
+#ifdef ZSTD_C_PREDICT
+    U32 predictedSmall = *(bt + 2*((curr-1)&btMask) + 0);
+    U32 predictedLarge = *(bt + 2*((curr-1)&btMask) + 1);
+    predictedSmall += (predictedSmall>0);
+    predictedLarge += (predictedLarge>0);
+#endif /* ZSTD_C_PREDICT */
+
+    DEBUGLOG(8, "ZSTD_insertBt1 (%u)", curr);
+
+    assert(ip <= iend-8);   /* required for h calculation */
+    hashTable[h] = curr;   /* Update Hash Table */
+
+    assert(windowLow > 0);
+    for (; nbCompares && (matchIndex >= windowLow); --nbCompares) {
+        U32* const nextPtr = bt + 2*(matchIndex & btMask);
+        size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
+        assert(matchIndex < curr);
+
+#ifdef ZSTD_C_PREDICT   /* note : can create issues when hlog small <= 11 */
+        const U32* predictPtr = bt + 2*((matchIndex-1) & btMask);   /* written this way, as bt is a roll buffer */
+        if (matchIndex == predictedSmall) {
+            /* no need to check length, result known */
+            *smallerPtr = matchIndex;
+            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+            smallerPtr = nextPtr+1;               /* new "smaller" => larger of match */
+            matchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
+            predictedSmall = predictPtr[1] + (predictPtr[1]>0);
+            continue;
+        }
+        if (matchIndex == predictedLarge) {
+            *largerPtr = matchIndex;
+            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+            largerPtr = nextPtr;
+            matchIndex = nextPtr[0];
+            predictedLarge = predictPtr[0] + (predictPtr[0]>0);
+            continue;
+        }
+#endif
+
+        if (!extDict || (matchIndex+matchLength >= dictLimit)) {
+            assert(matchIndex+matchLength >= dictLimit);   /* might be wrong if actually extDict */
+            match = base + matchIndex;
+            matchLength += ZSTD_count(ip+matchLength, match+matchLength, iend);
+        } else {
+            match = dictBase + matchIndex;
+            matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
+            if (matchIndex+matchLength >= dictLimit)
+                match = base + matchIndex;   /* to prepare for next usage of match[matchLength] */
+        }
+
+        if (matchLength > bestLength) {
+            bestLength = matchLength;
+            if (matchLength > matchEndIdx - matchIndex)
+                matchEndIdx = matchIndex + (U32)matchLength;
+        }
+
+        if (ip+matchLength == iend) {   /* equal : no way to know if inf or sup */
+            break;   /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt tree */
+        }
+
+        if (match[matchLength] < ip[matchLength]) {  /* necessarily within buffer */
+            /* match is smaller than current */
+            *smallerPtr = matchIndex;             /* update smaller idx */
+            commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
+            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop searching */
+            smallerPtr = nextPtr+1;               /* new "candidate" => larger than match, which was smaller than target */
+            matchIndex = nextPtr[1];              /* new matchIndex, larger than previous and closer to current */
+        } else {
+            /* match is larger than current */
+            *largerPtr = matchIndex;
+            commonLengthLarger = matchLength;
+            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop searching */
+            largerPtr = nextPtr;
+            matchIndex = nextPtr[0];
+    }   }
+
+    *smallerPtr = *largerPtr = 0;
+    {   U32 positions = 0;
+        if (bestLength > 384) positions = MIN(192, (U32)(bestLength - 384));   /* speed optimization */
+        assert(matchEndIdx > curr + 8);
+        return MAX(positions, matchEndIdx - (curr + 8));
+    }
+}
+
+FORCE_INLINE_TEMPLATE
+void ZSTD_updateTree_internal(
+                ZSTD_matchState_t* ms,
+                const BYTE* const ip, const BYTE* const iend,
+                const U32 mls, const ZSTD_dictMode_e dictMode)
+{
+    const BYTE* const base = ms->window.base;
+    U32 const target = (U32)(ip - base);
+    U32 idx = ms->nextToUpdate;
+    DEBUGLOG(6, "ZSTD_updateTree_internal, from %u to %u  (dictMode:%u)",
+                idx, target, dictMode);
+
+    while(idx < target) {
+        U32 const forward = ZSTD_insertBt1(ms, base+idx, iend, mls, dictMode == ZSTD_extDict);
+        assert(idx < (U32)(idx + forward));
+        idx += forward;
+    }
+    assert((size_t)(ip - base) <= (size_t)(U32)(-1));
+    assert((size_t)(iend - base) <= (size_t)(U32)(-1));
+    ms->nextToUpdate = target;
+}
+
+void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend) {
+    ZSTD_updateTree_internal(ms, ip, iend, ms->cParams.minMatch, ZSTD_noDict);
+}
+
+FORCE_INLINE_TEMPLATE
+U32 ZSTD_insertBtAndGetAllMatches (
+                    ZSTD_match_t* matches,   /* store result (found matches) in this table (presumed large enough) */
+                    ZSTD_matchState_t* ms,
+                    U32* nextToUpdate3,
+                    const BYTE* const ip, const BYTE* const iLimit, const ZSTD_dictMode_e dictMode,
+                    const U32 rep[ZSTD_REP_NUM],
+                    U32 const ll0,   /* tells if associated literal length is 0 or not. This value must be 0 or 1 */
+                    const U32 lengthToBeat,
+                    U32 const mls /* template */)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32 const sufficient_len = MIN(cParams->targetLength, ZSTD_OPT_NUM -1);
+    const BYTE* const base = ms->window.base;
+    U32 const curr = (U32)(ip-base);
+    U32 const hashLog = cParams->hashLog;
+    U32 const minMatch = (mls==3) ? 3 : 4;
+    U32* const hashTable = ms->hashTable;
+    size_t const h  = ZSTD_hashPtr(ip, hashLog, mls);
+    U32 matchIndex  = hashTable[h];
+    U32* const bt   = ms->chainTable;
+    U32 const btLog = cParams->chainLog - 1;
+    U32 const btMask= (1U << btLog) - 1;
+    size_t commonLengthSmaller=0, commonLengthLarger=0;
+    const BYTE* const dictBase = ms->window.dictBase;
+    U32 const dictLimit = ms->window.dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    U32 const btLow = (btMask >= curr) ? 0 : curr - btMask;
+    U32 const windowLow = ZSTD_getLowestMatchIndex(ms, curr, cParams->windowLog);
+    U32 const matchLow = windowLow ? windowLow : 1;
+    U32* smallerPtr = bt + 2*(curr&btMask);
+    U32* largerPtr  = bt + 2*(curr&btMask) + 1;
+    U32 matchEndIdx = curr+8+1;   /* farthest referenced position of any match => detects repetitive patterns */
+    U32 dummy32;   /* to be nullified at the end */
+    U32 mnum = 0;
+    U32 nbCompares = 1U << cParams->searchLog;
+
+    const ZSTD_matchState_t* dms    = dictMode == ZSTD_dictMatchState ? ms->dictMatchState : NULL;
+    const ZSTD_compressionParameters* const dmsCParams =
+                                      dictMode == ZSTD_dictMatchState ? &dms->cParams : NULL;
+    const BYTE* const dmsBase       = dictMode == ZSTD_dictMatchState ? dms->window.base : NULL;
+    const BYTE* const dmsEnd        = dictMode == ZSTD_dictMatchState ? dms->window.nextSrc : NULL;
+    U32         const dmsHighLimit  = dictMode == ZSTD_dictMatchState ? (U32)(dmsEnd - dmsBase) : 0;
+    U32         const dmsLowLimit   = dictMode == ZSTD_dictMatchState ? dms->window.lowLimit : 0;
+    U32         const dmsIndexDelta = dictMode == ZSTD_dictMatchState ? windowLow - dmsHighLimit : 0;
+    U32         const dmsHashLog    = dictMode == ZSTD_dictMatchState ? dmsCParams->hashLog : hashLog;
+    U32         const dmsBtLog      = dictMode == ZSTD_dictMatchState ? dmsCParams->chainLog - 1 : btLog;
+    U32         const dmsBtMask     = dictMode == ZSTD_dictMatchState ? (1U << dmsBtLog) - 1 : 0;
+    U32         const dmsBtLow      = dictMode == ZSTD_dictMatchState && dmsBtMask < dmsHighLimit - dmsLowLimit ? dmsHighLimit - dmsBtMask : dmsLowLimit;
+
+    size_t bestLength = lengthToBeat-1;
+    DEBUGLOG(8, "ZSTD_insertBtAndGetAllMatches: current=%u", curr);
+
+    /* check repCode */
+    assert(ll0 <= 1);   /* necessarily 1 or 0 */
+    {   U32 const lastR = ZSTD_REP_NUM + ll0;
+        U32 repCode;
+        for (repCode = ll0; repCode < lastR; repCode++) {
+            U32 const repOffset = (repCode==ZSTD_REP_NUM) ? (rep[0] - 1) : rep[repCode];
+            U32 const repIndex = curr - repOffset;
+            U32 repLen = 0;
+            assert(curr >= dictLimit);
+            if (repOffset-1 /* intentional overflow, discards 0 and -1 */ < curr-dictLimit) {  /* equivalent to `curr > repIndex >= dictLimit` */
+                /* We must validate the repcode offset because when we're using a dictionary the
+                 * valid offset range shrinks when the dictionary goes out of bounds.
+                 */
+                if ((repIndex >= windowLow) & (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(ip - repOffset, minMatch))) {
+                    repLen = (U32)ZSTD_count(ip+minMatch, ip+minMatch-repOffset, iLimit) + minMatch;
+                }
+            } else {  /* repIndex < dictLimit || repIndex >= curr */
+                const BYTE* const repMatch = dictMode == ZSTD_dictMatchState ?
+                                             dmsBase + repIndex - dmsIndexDelta :
+                                             dictBase + repIndex;
+                assert(curr >= windowLow);
+                if ( dictMode == ZSTD_extDict
+                  && ( ((repOffset-1) /*intentional overflow*/ < curr - windowLow)  /* equivalent to `curr > repIndex >= windowLow` */
+                     & (((U32)((dictLimit-1) - repIndex) >= 3) ) /* intentional overflow : do not test positions overlapping 2 memory segments */)
+                  && (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch)) ) {
+                    repLen = (U32)ZSTD_count_2segments(ip+minMatch, repMatch+minMatch, iLimit, dictEnd, prefixStart) + minMatch;
+                }
+                if (dictMode == ZSTD_dictMatchState
+                  && ( ((repOffset-1) /*intentional overflow*/ < curr - (dmsLowLimit + dmsIndexDelta))  /* equivalent to `curr > repIndex >= dmsLowLimit` */
+                     & ((U32)((dictLimit-1) - repIndex) >= 3) ) /* intentional overflow : do not test positions overlapping 2 memory segments */
+                  && (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch)) ) {
+                    repLen = (U32)ZSTD_count_2segments(ip+minMatch, repMatch+minMatch, iLimit, dmsEnd, prefixStart) + minMatch;
+            }   }
+            /* save longer solution */
+            if (repLen > bestLength) {
+                DEBUGLOG(8, "found repCode %u (ll0:%u, offset:%u) of length %u",
+                            repCode, ll0, repOffset, repLen);
+                bestLength = repLen;
+                matches[mnum].off = repCode - ll0;
+                matches[mnum].len = (U32)repLen;
+                mnum++;
+                if ( (repLen > sufficient_len)
+                   | (ip+repLen == iLimit) ) {  /* best possible */
+                    return mnum;
+    }   }   }   }
+
+    /* HC3 match finder */
+    if ((mls == 3) /*static*/ && (bestLength < mls)) {
+        U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3(ms, nextToUpdate3, ip);
+        if ((matchIndex3 >= matchLow)
+          & (curr - matchIndex3 < (1<<18)) /*heuristic : longer distance likely too expensive*/ ) {
+            size_t mlen;
+            if ((dictMode == ZSTD_noDict) /*static*/ || (dictMode == ZSTD_dictMatchState) /*static*/ || (matchIndex3 >= dictLimit)) {
+                const BYTE* const match = base + matchIndex3;
+                mlen = ZSTD_count(ip, match, iLimit);
+            } else {
+                const BYTE* const match = dictBase + matchIndex3;
+                mlen = ZSTD_count_2segments(ip, match, iLimit, dictEnd, prefixStart);
+            }
+
+            /* save best solution */
+            if (mlen >= mls /* == 3 > bestLength */) {
+                DEBUGLOG(8, "found small match with hlog3, of length %u",
+                            (U32)mlen);
+                bestLength = mlen;
+                assert(curr > matchIndex3);
+                assert(mnum==0);  /* no prior solution */
+                matches[0].off = (curr - matchIndex3) + ZSTD_REP_MOVE;
+                matches[0].len = (U32)mlen;
+                mnum = 1;
+                if ( (mlen > sufficient_len) |
+                     (ip+mlen == iLimit) ) {  /* best possible length */
+                    ms->nextToUpdate = curr+1;  /* skip insertion */
+                    return 1;
+        }   }   }
+        /* no dictMatchState lookup: dicts don't have a populated HC3 table */
+    }
+
+    hashTable[h] = curr;   /* Update Hash Table */
+
+    for (; nbCompares && (matchIndex >= matchLow); --nbCompares) {
+        U32* const nextPtr = bt + 2*(matchIndex & btMask);
+        const BYTE* match;
+        size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
+        assert(curr > matchIndex);
+
+        if ((dictMode == ZSTD_noDict) || (dictMode == ZSTD_dictMatchState) || (matchIndex+matchLength >= dictLimit)) {
+            assert(matchIndex+matchLength >= dictLimit);  /* ensure the condition is correct when !extDict */
+            match = base + matchIndex;
+            if (matchIndex >= dictLimit) assert(memcmp(match, ip, matchLength) == 0);  /* ensure early section of match is equal as expected */
+            matchLength += ZSTD_count(ip+matchLength, match+matchLength, iLimit);
+        } else {
+            match = dictBase + matchIndex;
+            assert(memcmp(match, ip, matchLength) == 0);  /* ensure early section of match is equal as expected */
+            matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iLimit, dictEnd, prefixStart);
+            if (matchIndex+matchLength >= dictLimit)
+                match = base + matchIndex;   /* prepare for match[matchLength] read */
+        }
+
+        if (matchLength > bestLength) {
+            DEBUGLOG(8, "found match of length %u at distance %u (offCode=%u)",
+                    (U32)matchLength, curr - matchIndex, curr - matchIndex + ZSTD_REP_MOVE);
+            assert(matchEndIdx > matchIndex);
+            if (matchLength > matchEndIdx - matchIndex)
+                matchEndIdx = matchIndex + (U32)matchLength;
+            bestLength = matchLength;
+            matches[mnum].off = (curr - matchIndex) + ZSTD_REP_MOVE;
+            matches[mnum].len = (U32)matchLength;
+            mnum++;
+            if ( (matchLength > ZSTD_OPT_NUM)
+               | (ip+matchLength == iLimit) /* equal : no way to know if inf or sup */) {
+                if (dictMode == ZSTD_dictMatchState) nbCompares = 0; /* break should also skip searching dms */
+                break; /* drop, to preserve bt consistency (miss a little bit of compression) */
+            }
+        }
+
+        if (match[matchLength] < ip[matchLength]) {
+            /* match smaller than current */
+            *smallerPtr = matchIndex;             /* update smaller idx */
+            commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
+            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+            smallerPtr = nextPtr+1;               /* new candidate => larger than match, which was smaller than current */
+            matchIndex = nextPtr[1];              /* new matchIndex, larger than previous, closer to current */
+        } else {
+            *largerPtr = matchIndex;
+            commonLengthLarger = matchLength;
+            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+            largerPtr = nextPtr;
+            matchIndex = nextPtr[0];
+    }   }
+
+    *smallerPtr = *largerPtr = 0;
+
+    assert(nbCompares <= (1U << ZSTD_SEARCHLOG_MAX)); /* Check we haven't underflowed. */
+    if (dictMode == ZSTD_dictMatchState && nbCompares) {
+        size_t const dmsH = ZSTD_hashPtr(ip, dmsHashLog, mls);
+        U32 dictMatchIndex = dms->hashTable[dmsH];
+        const U32* const dmsBt = dms->chainTable;
+        commonLengthSmaller = commonLengthLarger = 0;
+        for (; nbCompares && (dictMatchIndex > dmsLowLimit); --nbCompares) {
+            const U32* const nextPtr = dmsBt + 2*(dictMatchIndex & dmsBtMask);
+            size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
+            const BYTE* match = dmsBase + dictMatchIndex;
+            matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iLimit, dmsEnd, prefixStart);
+            if (dictMatchIndex+matchLength >= dmsHighLimit)
+                match = base + dictMatchIndex + dmsIndexDelta;   /* to prepare for next usage of match[matchLength] */
+
+            if (matchLength > bestLength) {
+                matchIndex = dictMatchIndex + dmsIndexDelta;
+                DEBUGLOG(8, "found dms match of length %u at distance %u (offCode=%u)",
+                        (U32)matchLength, curr - matchIndex, curr - matchIndex + ZSTD_REP_MOVE);
+                if (matchLength > matchEndIdx - matchIndex)
+                    matchEndIdx = matchIndex + (U32)matchLength;
+                bestLength = matchLength;
+                matches[mnum].off = (curr - matchIndex) + ZSTD_REP_MOVE;
+                matches[mnum].len = (U32)matchLength;
+                mnum++;
+                if ( (matchLength > ZSTD_OPT_NUM)
+                   | (ip+matchLength == iLimit) /* equal : no way to know if inf or sup */) {
+                    break;   /* drop, to guarantee consistency (miss a little bit of compression) */
+                }
+            }
+
+            if (dictMatchIndex <= dmsBtLow) { break; }   /* beyond tree size, stop the search */
+            if (match[matchLength] < ip[matchLength]) {
+                commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
+                dictMatchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
+            } else {
+                /* match is larger than current */
+                commonLengthLarger = matchLength;
+                dictMatchIndex = nextPtr[0];
+            }
+        }
+    }
+
+    assert(matchEndIdx > curr+8);
+    ms->nextToUpdate = matchEndIdx - 8;  /* skip repetitive patterns */
+    return mnum;
+}
+
+
+FORCE_INLINE_TEMPLATE U32 ZSTD_BtGetAllMatches (
+                        ZSTD_match_t* matches,   /* store result (match found, increasing size) in this table */
+                        ZSTD_matchState_t* ms,
+                        U32* nextToUpdate3,
+                        const BYTE* ip, const BYTE* const iHighLimit, const ZSTD_dictMode_e dictMode,
+                        const U32 rep[ZSTD_REP_NUM],
+                        U32 const ll0,
+                        U32 const lengthToBeat)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32 const matchLengthSearch = cParams->minMatch;
+    DEBUGLOG(8, "ZSTD_BtGetAllMatches");
+    if (ip < ms->window.base + ms->nextToUpdate) return 0;   /* skipped area */
+    ZSTD_updateTree_internal(ms, ip, iHighLimit, matchLengthSearch, dictMode);
+    switch(matchLengthSearch)
+    {
+    case 3 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 3);
+    default :
+    case 4 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 4);
+    case 5 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 5);
+    case 7 :
+    case 6 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 6);
+    }
+}
+
+/* ***********************
+*  LDM helper functions  *
+*************************/
+
+/* Struct containing info needed to make decision about ldm inclusion */
+typedef struct {
+    rawSeqStore_t seqStore;         /* External match candidates store for this block */
+    U32 startPosInBlock;            /* Start position of the current match candidate */
+    U32 endPosInBlock;              /* End position of the current match candidate */
+    U32 offset;                     /* Offset of the match candidate */
+} ZSTD_optLdm_t;
+
+/* ZSTD_optLdm_skipRawSeqStoreBytes():
+ * Moves forward in rawSeqStore by nbBytes, which will update the fields 'pos' and 'posInSequence'.
+ */
+static void ZSTD_optLdm_skipRawSeqStoreBytes(rawSeqStore_t* rawSeqStore, size_t nbBytes) {
+    U32 currPos = (U32)(rawSeqStore->posInSequence + nbBytes);
+    while (currPos && rawSeqStore->pos < rawSeqStore->size) {
+        rawSeq currSeq = rawSeqStore->seq[rawSeqStore->pos];
+        if (currPos >= currSeq.litLength + currSeq.matchLength) {
+            currPos -= currSeq.litLength + currSeq.matchLength;
+            rawSeqStore->pos++;
+        } else {
+            rawSeqStore->posInSequence = currPos;
+            break;
+        }
+    }
+    if (currPos == 0 || rawSeqStore->pos == rawSeqStore->size) {
+        rawSeqStore->posInSequence = 0;
+    }
+}
+
+/* ZSTD_opt_getNextMatchAndUpdateSeqStore():
+ * Calculates the beginning and end of the next match in the current block.
+ * Updates 'pos' and 'posInSequence' of the ldmSeqStore.
+ */
+static void ZSTD_opt_getNextMatchAndUpdateSeqStore(ZSTD_optLdm_t* optLdm, U32 currPosInBlock,
+                                                   U32 blockBytesRemaining) {
+    rawSeq currSeq;
+    U32 currBlockEndPos;
+    U32 literalsBytesRemaining;
+    U32 matchBytesRemaining;
+
+    /* Setting match end position to MAX to ensure we never use an LDM during this block */
+    if (optLdm->seqStore.size == 0 || optLdm->seqStore.pos >= optLdm->seqStore.size) {
+        optLdm->startPosInBlock = UINT_MAX;
+        optLdm->endPosInBlock = UINT_MAX;
+        return;
+    }
+    /* Calculate appropriate bytes left in matchLength and litLength after adjusting
+       based on ldmSeqStore->posInSequence */
+    currSeq = optLdm->seqStore.seq[optLdm->seqStore.pos];
+    assert(optLdm->seqStore.posInSequence <= currSeq.litLength + currSeq.matchLength);
+    currBlockEndPos = currPosInBlock + blockBytesRemaining;
+    literalsBytesRemaining = (optLdm->seqStore.posInSequence < currSeq.litLength) ?
+            currSeq.litLength - (U32)optLdm->seqStore.posInSequence :
+            0;
+    matchBytesRemaining = (literalsBytesRemaining == 0) ?
+            currSeq.matchLength - ((U32)optLdm->seqStore.posInSequence - currSeq.litLength) :
+            currSeq.matchLength;
+
+    /* If there are more literal bytes than bytes remaining in block, no ldm is possible */
+    if (literalsBytesRemaining >= blockBytesRemaining) {
+        optLdm->startPosInBlock = UINT_MAX;
+        optLdm->endPosInBlock = UINT_MAX;
+        ZSTD_optLdm_skipRawSeqStoreBytes(&optLdm->seqStore, blockBytesRemaining);
+        return;
+    }
+
+    /* Matches may be < MINMATCH by this process. In that case, we will reject them
+       when we are deciding whether or not to add the ldm */
+    optLdm->startPosInBlock = currPosInBlock + literalsBytesRemaining;
+    optLdm->endPosInBlock = optLdm->startPosInBlock + matchBytesRemaining;
+    optLdm->offset = currSeq.offset;
+
+    if (optLdm->endPosInBlock > currBlockEndPos) {
+        /* Match ends after the block ends, we can't use the whole match */
+        optLdm->endPosInBlock = currBlockEndPos;
+        ZSTD_optLdm_skipRawSeqStoreBytes(&optLdm->seqStore, currBlockEndPos - currPosInBlock);
+    } else {
+        /* Consume nb of bytes equal to size of sequence left */
+        ZSTD_optLdm_skipRawSeqStoreBytes(&optLdm->seqStore, literalsBytesRemaining + matchBytesRemaining);
+    }
+}
+
+/* ZSTD_optLdm_maybeAddMatch():
+ * Adds a match if it's long enough, based on it's 'matchStartPosInBlock'
+ * and 'matchEndPosInBlock', into 'matches'. Maintains the correct ordering of 'matches'
+ */
+static void ZSTD_optLdm_maybeAddMatch(ZSTD_match_t* matches, U32* nbMatches,
+                                      ZSTD_optLdm_t* optLdm, U32 currPosInBlock) {
+    U32 posDiff = currPosInBlock - optLdm->startPosInBlock;
+    /* Note: ZSTD_match_t actually contains offCode and matchLength (before subtracting MINMATCH) */
+    U32 candidateMatchLength = optLdm->endPosInBlock - optLdm->startPosInBlock - posDiff;
+    U32 candidateOffCode = optLdm->offset + ZSTD_REP_MOVE;
+
+    /* Ensure that current block position is not outside of the match */
+    if (currPosInBlock < optLdm->startPosInBlock
+      || currPosInBlock >= optLdm->endPosInBlock
+      || candidateMatchLength < MINMATCH) {
+        return;
+    }
+
+    if (*nbMatches == 0 || ((candidateMatchLength > matches[*nbMatches-1].len) && *nbMatches < ZSTD_OPT_NUM)) {
+        DEBUGLOG(6, "ZSTD_optLdm_maybeAddMatch(): Adding ldm candidate match (offCode: %u matchLength %u) at block position=%u",
+                 candidateOffCode, candidateMatchLength, currPosInBlock);
+        matches[*nbMatches].len = candidateMatchLength;
+        matches[*nbMatches].off = candidateOffCode;
+        (*nbMatches)++;
+    }
+}
+
+/* ZSTD_optLdm_processMatchCandidate():
+ * Wrapper function to update ldm seq store and call ldm functions as necessary.
+ */
+static void ZSTD_optLdm_processMatchCandidate(ZSTD_optLdm_t* optLdm, ZSTD_match_t* matches, U32* nbMatches,
+                                              U32 currPosInBlock, U32 remainingBytes) {
+    if (optLdm->seqStore.size == 0 || optLdm->seqStore.pos >= optLdm->seqStore.size) {
+        return;
+    }
+
+    if (currPosInBlock >= optLdm->endPosInBlock) {
+        if (currPosInBlock > optLdm->endPosInBlock) {
+            /* The position at which ZSTD_optLdm_processMatchCandidate() is called is not necessarily
+             * at the end of a match from the ldm seq store, and will often be some bytes
+             * over beyond matchEndPosInBlock. As such, we need to correct for these "overshoots"
+             */
+            U32 posOvershoot = currPosInBlock - optLdm->endPosInBlock;
+            ZSTD_optLdm_skipRawSeqStoreBytes(&optLdm->seqStore, posOvershoot);
+        } 
+        ZSTD_opt_getNextMatchAndUpdateSeqStore(optLdm, currPosInBlock, remainingBytes);
+    }
+    ZSTD_optLdm_maybeAddMatch(matches, nbMatches, optLdm, currPosInBlock);
+}
+
+/*-*******************************
+*  Optimal parser
+*********************************/
+
+
+static U32 ZSTD_totalLen(ZSTD_optimal_t sol)
+{
+    return sol.litlen + sol.mlen;
+}
+
+#if 0 /* debug */
+
+static void
+listStats(const U32* table, int lastEltID)
+{
+    int const nbElts = lastEltID + 1;
+    int enb;
+    for (enb=0; enb < nbElts; enb++) {
+        (void)table;
+        /* RAWLOG(2, "%3i:%3i,  ", enb, table[enb]); */
+        RAWLOG(2, "%4i,", table[enb]);
+    }
+    RAWLOG(2, " \n");
+}
+
+#endif
+
+FORCE_INLINE_TEMPLATE size_t
+ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
+                               seqStore_t* seqStore,
+                               U32 rep[ZSTD_REP_NUM],
+                         const void* src, size_t srcSize,
+                         const int optLevel,
+                         const ZSTD_dictMode_e dictMode)
+{
+    optState_t* const optStatePtr = &ms->opt;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    const BYTE* const base = ms->window.base;
+    const BYTE* const prefixStart = base + ms->window.dictLimit;
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+
+    U32 const sufficient_len = MIN(cParams->targetLength, ZSTD_OPT_NUM -1);
+    U32 const minMatch = (cParams->minMatch == 3) ? 3 : 4;
+    U32 nextToUpdate3 = ms->nextToUpdate;
+
+    ZSTD_optimal_t* const opt = optStatePtr->priceTable;
+    ZSTD_match_t* const matches = optStatePtr->matchTable;
+    ZSTD_optimal_t lastSequence;
+    ZSTD_optLdm_t optLdm;
+
+    optLdm.seqStore = ms->ldmSeqStore ? *ms->ldmSeqStore : kNullRawSeqStore;
+    optLdm.endPosInBlock = optLdm.startPosInBlock = optLdm.offset = 0;
+    ZSTD_opt_getNextMatchAndUpdateSeqStore(&optLdm, (U32)(ip-istart), (U32)(iend-ip));
+
+    /* init */
+    DEBUGLOG(5, "ZSTD_compressBlock_opt_generic: current=%u, prefix=%u, nextToUpdate=%u",
+                (U32)(ip - base), ms->window.dictLimit, ms->nextToUpdate);
+    assert(optLevel <= 2);
+    ZSTD_rescaleFreqs(optStatePtr, (const BYTE*)src, srcSize, optLevel);
+    ip += (ip==prefixStart);
+
+    /* Match Loop */
+    while (ip < ilimit) {
+        U32 cur, last_pos = 0;
+
+        /* find first match */
+        {   U32 const litlen = (U32)(ip - anchor);
+            U32 const ll0 = !litlen;
+            U32 nbMatches = ZSTD_BtGetAllMatches(matches, ms, &nextToUpdate3, ip, iend, dictMode, rep, ll0, minMatch);
+            ZSTD_optLdm_processMatchCandidate(&optLdm, matches, &nbMatches,
+                                              (U32)(ip-istart), (U32)(iend - ip));
+            if (!nbMatches) { ip++; continue; }
+
+            /* initialize opt[0] */
+            { U32 i ; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; }
+            opt[0].mlen = 0;  /* means is_a_literal */
+            opt[0].litlen = litlen;
+            /* We don't need to include the actual price of the literals because
+             * it is static for the duration of the forward pass, and is included
+             * in every price. We include the literal length to avoid negative
+             * prices when we subtract the previous literal length.
+             */
+            opt[0].price = ZSTD_litLengthPrice(litlen, optStatePtr, optLevel);
+
+            /* large match -> immediate encoding */
+            {   U32 const maxML = matches[nbMatches-1].len;
+                U32 const maxOffset = matches[nbMatches-1].off;
+                DEBUGLOG(6, "found %u matches of maxLength=%u and maxOffCode=%u at cPos=%u => start new series",
+                            nbMatches, maxML, maxOffset, (U32)(ip-prefixStart));
+
+                if (maxML > sufficient_len) {
+                    lastSequence.litlen = litlen;
+                    lastSequence.mlen = maxML;
+                    lastSequence.off = maxOffset;
+                    DEBUGLOG(6, "large match (%u>%u), immediate encoding",
+                                maxML, sufficient_len);
+                    cur = 0;
+                    last_pos = ZSTD_totalLen(lastSequence);
+                    goto _shortestPath;
+            }   }
+
+            /* set prices for first matches starting position == 0 */
+            {   U32 const literalsPrice = opt[0].price + ZSTD_litLengthPrice(0, optStatePtr, optLevel);
+                U32 pos;
+                U32 matchNb;
+                for (pos = 1; pos < minMatch; pos++) {
+                    opt[pos].price = ZSTD_MAX_PRICE;   /* mlen, litlen and price will be fixed during forward scanning */
+                }
+                for (matchNb = 0; matchNb < nbMatches; matchNb++) {
+                    U32 const offset = matches[matchNb].off;
+                    U32 const end = matches[matchNb].len;
+                    for ( ; pos <= end ; pos++ ) {
+                        U32 const matchPrice = ZSTD_getMatchPrice(offset, pos, optStatePtr, optLevel);
+                        U32 const sequencePrice = literalsPrice + matchPrice;
+                        DEBUGLOG(7, "rPos:%u => set initial price : %.2f",
+                                    pos, ZSTD_fCost(sequencePrice));
+                        opt[pos].mlen = pos;
+                        opt[pos].off = offset;
+                        opt[pos].litlen = litlen;
+                        opt[pos].price = sequencePrice;
+                }   }
+                last_pos = pos-1;
+            }
+        }
+
+        /* check further positions */
+        for (cur = 1; cur <= last_pos; cur++) {
+            const BYTE* const inr = ip + cur;
+            assert(cur < ZSTD_OPT_NUM);
+            DEBUGLOG(7, "cPos:%zi==rPos:%u", inr-istart, cur)
+
+            /* Fix current position with one literal if cheaper */
+            {   U32 const litlen = (opt[cur-1].mlen == 0) ? opt[cur-1].litlen + 1 : 1;
+                int const price = opt[cur-1].price
+                                + ZSTD_rawLiteralsCost(ip+cur-1, 1, optStatePtr, optLevel)
+                                + ZSTD_litLengthPrice(litlen, optStatePtr, optLevel)
+                                - ZSTD_litLengthPrice(litlen-1, optStatePtr, optLevel);
+                assert(price < 1000000000); /* overflow check */
+                if (price <= opt[cur].price) {
+                    DEBUGLOG(7, "cPos:%zi==rPos:%u : better price (%.2f<=%.2f) using literal (ll==%u) (hist:%u,%u,%u)",
+                                inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price), litlen,
+                                opt[cur-1].rep[0], opt[cur-1].rep[1], opt[cur-1].rep[2]);
+                    opt[cur].mlen = 0;
+                    opt[cur].off = 0;
+                    opt[cur].litlen = litlen;
+                    opt[cur].price = price;
+                } else {
+                    DEBUGLOG(7, "cPos:%zi==rPos:%u : literal would cost more (%.2f>%.2f) (hist:%u,%u,%u)",
+                                inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price),
+                                opt[cur].rep[0], opt[cur].rep[1], opt[cur].rep[2]);
+                }
+            }
+
+            /* Set the repcodes of the current position. We must do it here
+             * because we rely on the repcodes of the 2nd to last sequence being
+             * correct to set the next chunks repcodes during the backward
+             * traversal.
+             */
+            ZSTD_STATIC_ASSERT(sizeof(opt[cur].rep) == sizeof(repcodes_t));
+            assert(cur >= opt[cur].mlen);
+            if (opt[cur].mlen != 0) {
+                U32 const prev = cur - opt[cur].mlen;
+                repcodes_t newReps = ZSTD_updateRep(opt[prev].rep, opt[cur].off, opt[cur].litlen==0);
+                ZSTD_memcpy(opt[cur].rep, &newReps, sizeof(repcodes_t));
+            } else {
+                ZSTD_memcpy(opt[cur].rep, opt[cur - 1].rep, sizeof(repcodes_t));
+            }
+
+            /* last match must start at a minimum distance of 8 from oend */
+            if (inr > ilimit) continue;
+
+            if (cur == last_pos) break;
+
+            if ( (optLevel==0) /*static_test*/
+              && (opt[cur+1].price <= opt[cur].price + (BITCOST_MULTIPLIER/2)) ) {
+                DEBUGLOG(7, "move to next rPos:%u : price is <=", cur+1);
+                continue;  /* skip unpromising positions; about ~+6% speed, -0.01 ratio */
+            }
+
+            {   U32 const ll0 = (opt[cur].mlen != 0);
+                U32 const litlen = (opt[cur].mlen == 0) ? opt[cur].litlen : 0;
+                U32 const previousPrice = opt[cur].price;
+                U32 const basePrice = previousPrice + ZSTD_litLengthPrice(0, optStatePtr, optLevel);
+                U32 nbMatches = ZSTD_BtGetAllMatches(matches, ms, &nextToUpdate3, inr, iend, dictMode, opt[cur].rep, ll0, minMatch);
+                U32 matchNb;
+
+                ZSTD_optLdm_processMatchCandidate(&optLdm, matches, &nbMatches,
+                                                  (U32)(inr-istart), (U32)(iend-inr));
+
+                if (!nbMatches) {
+                    DEBUGLOG(7, "rPos:%u : no match found", cur);
+                    continue;
+                }
+
+                {   U32 const maxML = matches[nbMatches-1].len;
+                    DEBUGLOG(7, "cPos:%zi==rPos:%u, found %u matches, of maxLength=%u",
+                                inr-istart, cur, nbMatches, maxML);
+
+                    if ( (maxML > sufficient_len)
+                      || (cur + maxML >= ZSTD_OPT_NUM) ) {
+                        lastSequence.mlen = maxML;
+                        lastSequence.off = matches[nbMatches-1].off;
+                        lastSequence.litlen = litlen;
+                        cur -= (opt[cur].mlen==0) ? opt[cur].litlen : 0;  /* last sequence is actually only literals, fix cur to last match - note : may underflow, in which case, it's first sequence, and it's okay */
+                        last_pos = cur + ZSTD_totalLen(lastSequence);
+                        if (cur > ZSTD_OPT_NUM) cur = 0;   /* underflow => first match */
+                        goto _shortestPath;
+                }   }
+
+                /* set prices using matches found at position == cur */
+                for (matchNb = 0; matchNb < nbMatches; matchNb++) {
+                    U32 const offset = matches[matchNb].off;
+                    U32 const lastML = matches[matchNb].len;
+                    U32 const startML = (matchNb>0) ? matches[matchNb-1].len+1 : minMatch;
+                    U32 mlen;
+
+                    DEBUGLOG(7, "testing match %u => offCode=%4u, mlen=%2u, llen=%2u",
+                                matchNb, matches[matchNb].off, lastML, litlen);
+
+                    for (mlen = lastML; mlen >= startML; mlen--) {  /* scan downward */
+                        U32 const pos = cur + mlen;
+                        int const price = basePrice + ZSTD_getMatchPrice(offset, mlen, optStatePtr, optLevel);
+
+                        if ((pos > last_pos) || (price < opt[pos].price)) {
+                            DEBUGLOG(7, "rPos:%u (ml=%2u) => new better price (%.2f<%.2f)",
+                                        pos, mlen, ZSTD_fCost(price), ZSTD_fCost(opt[pos].price));
+                            while (last_pos < pos) { opt[last_pos+1].price = ZSTD_MAX_PRICE; last_pos++; }   /* fill empty positions */
+                            opt[pos].mlen = mlen;
+                            opt[pos].off = offset;
+                            opt[pos].litlen = litlen;
+                            opt[pos].price = price;
+                        } else {
+                            DEBUGLOG(7, "rPos:%u (ml=%2u) => new price is worse (%.2f>=%.2f)",
+                                        pos, mlen, ZSTD_fCost(price), ZSTD_fCost(opt[pos].price));
+                            if (optLevel==0) break;  /* early update abort; gets ~+10% speed for about -0.01 ratio loss */
+                        }
+            }   }   }
+        }  /* for (cur = 1; cur <= last_pos; cur++) */
+
+        lastSequence = opt[last_pos];
+        cur = last_pos > ZSTD_totalLen(lastSequence) ? last_pos - ZSTD_totalLen(lastSequence) : 0;  /* single sequence, and it starts before `ip` */
+        assert(cur < ZSTD_OPT_NUM);  /* control overflow*/
+
+_shortestPath:   /* cur, last_pos, best_mlen, best_off have to be set */
+        assert(opt[0].mlen == 0);
+
+        /* Set the next chunk's repcodes based on the repcodes of the beginning
+         * of the last match, and the last sequence. This avoids us having to
+         * update them while traversing the sequences.
+         */
+        if (lastSequence.mlen != 0) {
+            repcodes_t reps = ZSTD_updateRep(opt[cur].rep, lastSequence.off, lastSequence.litlen==0);
+            ZSTD_memcpy(rep, &reps, sizeof(reps));
+        } else {
+            ZSTD_memcpy(rep, opt[cur].rep, sizeof(repcodes_t));
+        }
+
+        {   U32 const storeEnd = cur + 1;
+            U32 storeStart = storeEnd;
+            U32 seqPos = cur;
+
+            DEBUGLOG(6, "start reverse traversal (last_pos:%u, cur:%u)",
+                        last_pos, cur); (void)last_pos;
+            assert(storeEnd < ZSTD_OPT_NUM);
+            DEBUGLOG(6, "last sequence copied into pos=%u (llen=%u,mlen=%u,ofc=%u)",
+                        storeEnd, lastSequence.litlen, lastSequence.mlen, lastSequence.off);
+            opt[storeEnd] = lastSequence;
+            while (seqPos > 0) {
+                U32 const backDist = ZSTD_totalLen(opt[seqPos]);
+                storeStart--;
+                DEBUGLOG(6, "sequence from rPos=%u copied into pos=%u (llen=%u,mlen=%u,ofc=%u)",
+                            seqPos, storeStart, opt[seqPos].litlen, opt[seqPos].mlen, opt[seqPos].off);
+                opt[storeStart] = opt[seqPos];
+                seqPos = (seqPos > backDist) ? seqPos - backDist : 0;
+            }
+
+            /* save sequences */
+            DEBUGLOG(6, "sending selected sequences into seqStore")
+            {   U32 storePos;
+                for (storePos=storeStart; storePos <= storeEnd; storePos++) {
+                    U32 const llen = opt[storePos].litlen;
+                    U32 const mlen = opt[storePos].mlen;
+                    U32 const offCode = opt[storePos].off;
+                    U32 const advance = llen + mlen;
+                    DEBUGLOG(6, "considering seq starting at %zi, llen=%u, mlen=%u",
+                                anchor - istart, (unsigned)llen, (unsigned)mlen);
+
+                    if (mlen==0) {  /* only literals => must be last "sequence", actually starting a new stream of sequences */
+                        assert(storePos == storeEnd);   /* must be last sequence */
+                        ip = anchor + llen;     /* last "sequence" is a bunch of literals => don't progress anchor */
+                        continue;   /* will finish */
+                    }
+
+                    assert(anchor + llen <= iend);
+                    ZSTD_updateStats(optStatePtr, llen, anchor, offCode, mlen);
+                    ZSTD_storeSeq(seqStore, llen, anchor, iend, offCode, mlen-MINMATCH);
+                    anchor += advance;
+                    ip = anchor;
+            }   }
+            ZSTD_setBasePrices(optStatePtr, optLevel);
+        }
+    }   /* while (ip < ilimit) */
+
+    /* Return the last literals size */
+    return (size_t)(iend - anchor);
+}
+
+
+size_t ZSTD_compressBlock_btopt(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        const void* src, size_t srcSize)
+{
+    DEBUGLOG(5, "ZSTD_compressBlock_btopt");
+    return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 0 /*optLevel*/, ZSTD_noDict);
+}
+
+
+/* used in 2-pass strategy */
+static U32 ZSTD_upscaleStat(unsigned* table, U32 lastEltIndex, int bonus)
+{
+    U32 s, sum=0;
+    assert(ZSTD_FREQ_DIV+bonus >= 0);
+    for (s=0; s<lastEltIndex+1; s++) {
+        table[s] <<= ZSTD_FREQ_DIV+bonus;
+        table[s]--;
+        sum += table[s];
+    }
+    return sum;
+}
+
+/* used in 2-pass strategy */
+MEM_STATIC void ZSTD_upscaleStats(optState_t* optPtr)
+{
+    if (ZSTD_compressedLiterals(optPtr))
+        optPtr->litSum = ZSTD_upscaleStat(optPtr->litFreq, MaxLit, 0);
+    optPtr->litLengthSum = ZSTD_upscaleStat(optPtr->litLengthFreq, MaxLL, 0);
+    optPtr->matchLengthSum = ZSTD_upscaleStat(optPtr->matchLengthFreq, MaxML, 0);
+    optPtr->offCodeSum = ZSTD_upscaleStat(optPtr->offCodeFreq, MaxOff, 0);
+}
+
+/* ZSTD_initStats_ultra():
+ * make a first compression pass, just to seed stats with more accurate starting values.
+ * only works on first block, with no dictionary and no ldm.
+ * this function cannot error, hence its contract must be respected.
+ */
+static void
+ZSTD_initStats_ultra(ZSTD_matchState_t* ms,
+                     seqStore_t* seqStore,
+                     U32 rep[ZSTD_REP_NUM],
+               const void* src, size_t srcSize)
+{
+    U32 tmpRep[ZSTD_REP_NUM];  /* updated rep codes will sink here */
+    ZSTD_memcpy(tmpRep, rep, sizeof(tmpRep));
+
+    DEBUGLOG(4, "ZSTD_initStats_ultra (srcSize=%zu)", srcSize);
+    assert(ms->opt.litLengthSum == 0);    /* first block */
+    assert(seqStore->sequences == seqStore->sequencesStart);   /* no ldm */
+    assert(ms->window.dictLimit == ms->window.lowLimit);   /* no dictionary */
+    assert(ms->window.dictLimit - ms->nextToUpdate <= 1);  /* no prefix (note: intentional overflow, defined as 2-complement) */
+
+    ZSTD_compressBlock_opt_generic(ms, seqStore, tmpRep, src, srcSize, 2 /*optLevel*/, ZSTD_noDict);   /* generate stats into ms->opt*/
+
+    /* invalidate first scan from history */
+    ZSTD_resetSeqStore(seqStore);
+    ms->window.base -= srcSize;
+    ms->window.dictLimit += (U32)srcSize;
+    ms->window.lowLimit = ms->window.dictLimit;
+    ms->nextToUpdate = ms->window.dictLimit;
+
+    /* re-inforce weight of collected statistics */
+    ZSTD_upscaleStats(&ms->opt);
+}
+
+size_t ZSTD_compressBlock_btultra(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        const void* src, size_t srcSize)
+{
+    DEBUGLOG(5, "ZSTD_compressBlock_btultra (srcSize=%zu)", srcSize);
+    return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /*optLevel*/, ZSTD_noDict);
+}
+
+size_t ZSTD_compressBlock_btultra2(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        const void* src, size_t srcSize)
+{
+    U32 const curr = (U32)((const BYTE*)src - ms->window.base);
+    DEBUGLOG(5, "ZSTD_compressBlock_btultra2 (srcSize=%zu)", srcSize);
+
+    /* 2-pass strategy:
+     * this strategy makes a first pass over first block to collect statistics
+     * and seed next round's statistics with it.
+     * After 1st pass, function forgets everything, and starts a new block.
+     * Consequently, this can only work if no data has been previously loaded in tables,
+     * aka, no dictionary, no prefix, no ldm preprocessing.
+     * The compression ratio gain is generally small (~0.5% on first block),
+     * the cost is 2x cpu time on first block. */
+    assert(srcSize <= ZSTD_BLOCKSIZE_MAX);
+    if ( (ms->opt.litLengthSum==0)   /* first block */
+      && (seqStore->sequences == seqStore->sequencesStart)  /* no ldm */
+      && (ms->window.dictLimit == ms->window.lowLimit)   /* no dictionary */
+      && (curr == ms->window.dictLimit)   /* start of frame, nothing already loaded nor skipped */
+      && (srcSize > ZSTD_PREDEF_THRESHOLD)
+      ) {
+        ZSTD_initStats_ultra(ms, seqStore, rep, src, srcSize);
+    }
+
+    return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /*optLevel*/, ZSTD_noDict);
+}
+
+size_t ZSTD_compressBlock_btopt_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 0 /*optLevel*/, ZSTD_dictMatchState);
+}
+
+size_t ZSTD_compressBlock_btultra_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /*optLevel*/, ZSTD_dictMatchState);
+}
+
+size_t ZSTD_compressBlock_btopt_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 0 /*optLevel*/, ZSTD_extDict);
+}
+
+size_t ZSTD_compressBlock_btultra_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /*optLevel*/, ZSTD_extDict);
+}
+
+/* note : no btultra2 variant for extDict nor dictMatchState,
+ * because btultra2 is not meant to work with dictionaries
+ * and is only specific for the first block (no prefix) */
diff --git a/lib/zstd/compress/zstd_opt.h b/lib/zstd/compress/zstd_opt.h
new file mode 100644
index 0000000..22b8628
--- /dev/null
+++ b/lib/zstd/compress/zstd_opt.h
@@ -0,0 +1,50 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_OPT_H
+#define ZSTD_OPT_H
+
+
+#include "zstd_compress_internal.h"
+
+/* used in ZSTD_loadDictionaryContent() */
+void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend);
+
+size_t ZSTD_compressBlock_btopt(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_btultra(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_btultra2(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+
+
+size_t ZSTD_compressBlock_btopt_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_btultra_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+
+size_t ZSTD_compressBlock_btopt_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_btultra_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+
+        /* note : no btultra2 variant for extDict nor dictMatchState,
+         * because btultra2 is not meant to work with dictionaries
+         * and is only specific for the first block (no prefix) */
+
+
+#endif /* ZSTD_OPT_H */
diff --git a/lib/zstd/decompress.c b/lib/zstd/decompress.c
deleted file mode 100644
index 66cd487a..0000000
--- a/lib/zstd/decompress.c
+++ /dev/null
@@ -1,2531 +0,0 @@
-/**
- * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
- * All rights reserved.
- *
- * This source code is licensed under the BSD-style license found in the
- * LICENSE file in the root directory of https://github.com/facebook/zstd.
- * An additional grant of patent rights can be found in the PATENTS file in the
- * same directory.
- *
- * This program is free software; you can redistribute it and/or modify it under
- * the terms of the GNU General Public License version 2 as published by the
- * Free Software Foundation. This program is dual-licensed; you may select
- * either version 2 of the GNU General Public License ("GPL") or BSD license
- * ("BSD").
- */
-
-/* ***************************************************************
-*  Tuning parameters
-*****************************************************************/
-/*!
-*  MAXWINDOWSIZE_DEFAULT :
-*  maximum window size accepted by DStream, by default.
-*  Frames requiring more memory will be rejected.
-*/
-#ifndef ZSTD_MAXWINDOWSIZE_DEFAULT
-#define ZSTD_MAXWINDOWSIZE_DEFAULT ((1 << ZSTD_WINDOWLOG_MAX) + 1) /* defined within zstd.h */
-#endif
-
-/*-*******************************************************
-*  Dependencies
-*********************************************************/
-#include "fse.h"
-#include "huf.h"
-#include "mem.h" /* low level memory routines */
-#include "zstd_internal.h"
-#include <linux/kernel.h>
-#include <linux/module.h>
-#include <linux/string.h> /* memcpy, memmove, memset */
-
-#define ZSTD_PREFETCH(ptr) __builtin_prefetch(ptr, 0, 0)
-
-/*-*************************************
-*  Macros
-***************************************/
-#define ZSTD_isError ERR_isError /* for inlining */
-#define FSE_isError ERR_isError
-#define HUF_isError ERR_isError
-
-/*_*******************************************************
-*  Memory operations
-**********************************************************/
-static void ZSTD_copy4(void *dst, const void *src) { memcpy(dst, src, 4); }
-
-/*-*************************************************************
-*   Context management
-***************************************************************/
-typedef enum {
-	ZSTDds_getFrameHeaderSize,
-	ZSTDds_decodeFrameHeader,
-	ZSTDds_decodeBlockHeader,
-	ZSTDds_decompressBlock,
-	ZSTDds_decompressLastBlock,
-	ZSTDds_checkChecksum,
-	ZSTDds_decodeSkippableHeader,
-	ZSTDds_skipFrame
-} ZSTD_dStage;
-
-typedef struct {
-	FSE_DTable LLTable[FSE_DTABLE_SIZE_U32(LLFSELog)];
-	FSE_DTable OFTable[FSE_DTABLE_SIZE_U32(OffFSELog)];
-	FSE_DTable MLTable[FSE_DTABLE_SIZE_U32(MLFSELog)];
-	HUF_DTable hufTable[HUF_DTABLE_SIZE(HufLog)]; /* can accommodate HUF_decompress4X */
-	U64 workspace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32 / 2];
-	U32 rep[ZSTD_REP_NUM];
-} ZSTD_entropyTables_t;
-
-struct ZSTD_DCtx_s {
-	const FSE_DTable *LLTptr;
-	const FSE_DTable *MLTptr;
-	const FSE_DTable *OFTptr;
-	const HUF_DTable *HUFptr;
-	ZSTD_entropyTables_t entropy;
-	const void *previousDstEnd; /* detect continuity */
-	const void *base;	   /* start of curr segment */
-	const void *vBase;	  /* virtual start of previous segment if it was just before curr one */
-	const void *dictEnd;	/* end of previous segment */
-	size_t expected;
-	ZSTD_frameParams fParams;
-	blockType_e bType; /* used in ZSTD_decompressContinue(), to transfer blockType between header decoding and block decoding stages */
-	ZSTD_dStage stage;
-	U32 litEntropy;
-	U32 fseEntropy;
-	struct xxh64_state xxhState;
-	size_t headerSize;
-	U32 dictID;
-	const BYTE *litPtr;
-	ZSTD_customMem customMem;
-	size_t litSize;
-	size_t rleSize;
-	BYTE litBuffer[ZSTD_BLOCKSIZE_ABSOLUTEMAX + WILDCOPY_OVERLENGTH];
-	BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX];
-}; /* typedef'd to ZSTD_DCtx within "zstd.h" */
-
-size_t ZSTD_DCtxWorkspaceBound(void) { return ZSTD_ALIGN(sizeof(ZSTD_stack)) + ZSTD_ALIGN(sizeof(ZSTD_DCtx)); }
-
-size_t ZSTD_decompressBegin(ZSTD_DCtx *dctx)
-{
-	dctx->expected = ZSTD_frameHeaderSize_prefix;
-	dctx->stage = ZSTDds_getFrameHeaderSize;
-	dctx->previousDstEnd = NULL;
-	dctx->base = NULL;
-	dctx->vBase = NULL;
-	dctx->dictEnd = NULL;
-	dctx->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */
-	dctx->litEntropy = dctx->fseEntropy = 0;
-	dctx->dictID = 0;
-	ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.rep) == sizeof(repStartValue));
-	memcpy(dctx->entropy.rep, repStartValue, sizeof(repStartValue)); /* initial repcodes */
-	dctx->LLTptr = dctx->entropy.LLTable;
-	dctx->MLTptr = dctx->entropy.MLTable;
-	dctx->OFTptr = dctx->entropy.OFTable;
-	dctx->HUFptr = dctx->entropy.hufTable;
-	return 0;
-}
-
-ZSTD_DCtx *ZSTD_createDCtx_advanced(ZSTD_customMem customMem)
-{
-	ZSTD_DCtx *dctx;
-
-	if (!customMem.customAlloc || !customMem.customFree)
-		return NULL;
-
-	dctx = (ZSTD_DCtx *)ZSTD_malloc(sizeof(ZSTD_DCtx), customMem);
-	if (!dctx)
-		return NULL;
-	memcpy(&dctx->customMem, &customMem, sizeof(customMem));
-	ZSTD_decompressBegin(dctx);
-	return dctx;
-}
-
-ZSTD_DCtx *ZSTD_initDCtx(void *workspace, size_t workspaceSize)
-{
-	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-	return ZSTD_createDCtx_advanced(stackMem);
-}
-
-size_t ZSTD_freeDCtx(ZSTD_DCtx *dctx)
-{
-	if (dctx == NULL)
-		return 0; /* support free on NULL */
-	ZSTD_free(dctx, dctx->customMem);
-	return 0; /* reserved as a potential error code in the future */
-}
-
-void ZSTD_copyDCtx(ZSTD_DCtx *dstDCtx, const ZSTD_DCtx *srcDCtx)
-{
-	size_t const workSpaceSize = (ZSTD_BLOCKSIZE_ABSOLUTEMAX + WILDCOPY_OVERLENGTH) + ZSTD_frameHeaderSize_max;
-	memcpy(dstDCtx, srcDCtx, sizeof(ZSTD_DCtx) - workSpaceSize); /* no need to copy workspace */
-}
-
-static void ZSTD_refDDict(ZSTD_DCtx *dstDCtx, const ZSTD_DDict *ddict);
-
-/*-*************************************************************
-*   Decompression section
-***************************************************************/
-
-/*! ZSTD_isFrame() :
- *  Tells if the content of `buffer` starts with a valid Frame Identifier.
- *  Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0.
- *  Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled.
- *  Note 3 : Skippable Frame Identifiers are considered valid. */
-unsigned ZSTD_isFrame(const void *buffer, size_t size)
-{
-	if (size < 4)
-		return 0;
-	{
-		U32 const magic = ZSTD_readLE32(buffer);
-		if (magic == ZSTD_MAGICNUMBER)
-			return 1;
-		if ((magic & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START)
-			return 1;
-	}
-	return 0;
-}
-
-/** ZSTD_frameHeaderSize() :
-*   srcSize must be >= ZSTD_frameHeaderSize_prefix.
-*   @return : size of the Frame Header */
-static size_t ZSTD_frameHeaderSize(const void *src, size_t srcSize)
-{
-	if (srcSize < ZSTD_frameHeaderSize_prefix)
-		return ERROR(srcSize_wrong);
-	{
-		BYTE const fhd = ((const BYTE *)src)[4];
-		U32 const dictID = fhd & 3;
-		U32 const singleSegment = (fhd >> 5) & 1;
-		U32 const fcsId = fhd >> 6;
-		return ZSTD_frameHeaderSize_prefix + !singleSegment + ZSTD_did_fieldSize[dictID] + ZSTD_fcs_fieldSize[fcsId] + (singleSegment && !fcsId);
-	}
-}
-
-/** ZSTD_getFrameParams() :
-*   decode Frame Header, or require larger `srcSize`.
-*   @return : 0, `fparamsPtr` is correctly filled,
-*            >0, `srcSize` is too small, result is expected `srcSize`,
-*             or an error code, which can be tested using ZSTD_isError() */
-size_t ZSTD_getFrameParams(ZSTD_frameParams *fparamsPtr, const void *src, size_t srcSize)
-{
-	const BYTE *ip = (const BYTE *)src;
-
-	if (srcSize < ZSTD_frameHeaderSize_prefix)
-		return ZSTD_frameHeaderSize_prefix;
-	if (ZSTD_readLE32(src) != ZSTD_MAGICNUMBER) {
-		if ((ZSTD_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
-			if (srcSize < ZSTD_skippableHeaderSize)
-				return ZSTD_skippableHeaderSize; /* magic number + skippable frame length */
-			memset(fparamsPtr, 0, sizeof(*fparamsPtr));
-			fparamsPtr->frameContentSize = ZSTD_readLE32((const char *)src + 4);
-			fparamsPtr->windowSize = 0; /* windowSize==0 means a frame is skippable */
-			return 0;
-		}
-		return ERROR(prefix_unknown);
-	}
-
-	/* ensure there is enough `srcSize` to fully read/decode frame header */
-	{
-		size_t const fhsize = ZSTD_frameHeaderSize(src, srcSize);
-		if (srcSize < fhsize)
-			return fhsize;
-	}
-
-	{
-		BYTE const fhdByte = ip[4];
-		size_t pos = 5;
-		U32 const dictIDSizeCode = fhdByte & 3;
-		U32 const checksumFlag = (fhdByte >> 2) & 1;
-		U32 const singleSegment = (fhdByte >> 5) & 1;
-		U32 const fcsID = fhdByte >> 6;
-		U32 const windowSizeMax = 1U << ZSTD_WINDOWLOG_MAX;
-		U32 windowSize = 0;
-		U32 dictID = 0;
-		U64 frameContentSize = 0;
-		if ((fhdByte & 0x08) != 0)
-			return ERROR(frameParameter_unsupported); /* reserved bits, which must be zero */
-		if (!singleSegment) {
-			BYTE const wlByte = ip[pos++];
-			U32 const windowLog = (wlByte >> 3) + ZSTD_WINDOWLOG_ABSOLUTEMIN;
-			if (windowLog > ZSTD_WINDOWLOG_MAX)
-				return ERROR(frameParameter_windowTooLarge); /* avoids issue with 1 << windowLog */
-			windowSize = (1U << windowLog);
-			windowSize += (windowSize >> 3) * (wlByte & 7);
-		}
-
-		switch (dictIDSizeCode) {
-		default: /* impossible */
-		case 0: break;
-		case 1:
-			dictID = ip[pos];
-			pos++;
-			break;
-		case 2:
-			dictID = ZSTD_readLE16(ip + pos);
-			pos += 2;
-			break;
-		case 3:
-			dictID = ZSTD_readLE32(ip + pos);
-			pos += 4;
-			break;
-		}
-		switch (fcsID) {
-		default: /* impossible */
-		case 0:
-			if (singleSegment)
-				frameContentSize = ip[pos];
-			break;
-		case 1: frameContentSize = ZSTD_readLE16(ip + pos) + 256; break;
-		case 2: frameContentSize = ZSTD_readLE32(ip + pos); break;
-		case 3: frameContentSize = ZSTD_readLE64(ip + pos); break;
-		}
-		if (!windowSize)
-			windowSize = (U32)frameContentSize;
-		if (windowSize > windowSizeMax)
-			return ERROR(frameParameter_windowTooLarge);
-		fparamsPtr->frameContentSize = frameContentSize;
-		fparamsPtr->windowSize = windowSize;
-		fparamsPtr->dictID = dictID;
-		fparamsPtr->checksumFlag = checksumFlag;
-	}
-	return 0;
-}
-
-/** ZSTD_getFrameContentSize() :
-*   compatible with legacy mode
-*   @return : decompressed size of the single frame pointed to be `src` if known, otherwise
-*             - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined
-*             - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small) */
-unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize)
-{
-	{
-		ZSTD_frameParams fParams;
-		if (ZSTD_getFrameParams(&fParams, src, srcSize) != 0)
-			return ZSTD_CONTENTSIZE_ERROR;
-		if (fParams.windowSize == 0) {
-			/* Either skippable or empty frame, size == 0 either way */
-			return 0;
-		} else if (fParams.frameContentSize != 0) {
-			return fParams.frameContentSize;
-		} else {
-			return ZSTD_CONTENTSIZE_UNKNOWN;
-		}
-	}
-}
-
-/** ZSTD_findDecompressedSize() :
- *  compatible with legacy mode
- *  `srcSize` must be the exact length of some number of ZSTD compressed and/or
- *      skippable frames
- *  @return : decompressed size of the frames contained */
-unsigned long long ZSTD_findDecompressedSize(const void *src, size_t srcSize)
-{
-	{
-		unsigned long long totalDstSize = 0;
-		while (srcSize >= ZSTD_frameHeaderSize_prefix) {
-			const U32 magicNumber = ZSTD_readLE32(src);
-
-			if ((magicNumber & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
-				size_t skippableSize;
-				if (srcSize < ZSTD_skippableHeaderSize)
-					return ERROR(srcSize_wrong);
-				skippableSize = ZSTD_readLE32((const BYTE *)src + 4) + ZSTD_skippableHeaderSize;
-				if (srcSize < skippableSize) {
-					return ZSTD_CONTENTSIZE_ERROR;
-				}
-
-				src = (const BYTE *)src + skippableSize;
-				srcSize -= skippableSize;
-				continue;
-			}
-
-			{
-				unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize);
-				if (ret >= ZSTD_CONTENTSIZE_ERROR)
-					return ret;
-
-				/* check for overflow */
-				if (totalDstSize + ret < totalDstSize)
-					return ZSTD_CONTENTSIZE_ERROR;
-				totalDstSize += ret;
-			}
-			{
-				size_t const frameSrcSize = ZSTD_findFrameCompressedSize(src, srcSize);
-				if (ZSTD_isError(frameSrcSize)) {
-					return ZSTD_CONTENTSIZE_ERROR;
-				}
-
-				src = (const BYTE *)src + frameSrcSize;
-				srcSize -= frameSrcSize;
-			}
-		}
-
-		if (srcSize) {
-			return ZSTD_CONTENTSIZE_ERROR;
-		}
-
-		return totalDstSize;
-	}
-}
-
-/** ZSTD_decodeFrameHeader() :
-*   `headerSize` must be the size provided by ZSTD_frameHeaderSize().
-*   @return : 0 if success, or an error code, which can be tested using ZSTD_isError() */
-static size_t ZSTD_decodeFrameHeader(ZSTD_DCtx *dctx, const void *src, size_t headerSize)
-{
-	size_t const result = ZSTD_getFrameParams(&(dctx->fParams), src, headerSize);
-	if (ZSTD_isError(result))
-		return result; /* invalid header */
-	if (result > 0)
-		return ERROR(srcSize_wrong); /* headerSize too small */
-	if (dctx->fParams.dictID && (dctx->dictID != dctx->fParams.dictID))
-		return ERROR(dictionary_wrong);
-	if (dctx->fParams.checksumFlag)
-		xxh64_reset(&dctx->xxhState, 0);
-	return 0;
-}
-
-typedef struct {
-	blockType_e blockType;
-	U32 lastBlock;
-	U32 origSize;
-} blockProperties_t;
-
-/*! ZSTD_getcBlockSize() :
-*   Provides the size of compressed block from block header `src` */
-size_t ZSTD_getcBlockSize(const void *src, size_t srcSize, blockProperties_t *bpPtr)
-{
-	if (srcSize < ZSTD_blockHeaderSize)
-		return ERROR(srcSize_wrong);
-	{
-		U32 const cBlockHeader = ZSTD_readLE24(src);
-		U32 const cSize = cBlockHeader >> 3;
-		bpPtr->lastBlock = cBlockHeader & 1;
-		bpPtr->blockType = (blockType_e)((cBlockHeader >> 1) & 3);
-		bpPtr->origSize = cSize; /* only useful for RLE */
-		if (bpPtr->blockType == bt_rle)
-			return 1;
-		if (bpPtr->blockType == bt_reserved)
-			return ERROR(corruption_detected);
-		return cSize;
-	}
-}
-
-static size_t ZSTD_copyRawBlock(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	if (srcSize > dstCapacity)
-		return ERROR(dstSize_tooSmall);
-	memcpy(dst, src, srcSize);
-	return srcSize;
-}
-
-static size_t ZSTD_setRleBlock(void *dst, size_t dstCapacity, const void *src, size_t srcSize, size_t regenSize)
-{
-	if (srcSize != 1)
-		return ERROR(srcSize_wrong);
-	if (regenSize > dstCapacity)
-		return ERROR(dstSize_tooSmall);
-	memset(dst, *(const BYTE *)src, regenSize);
-	return regenSize;
-}
-
-/*! ZSTD_decodeLiteralsBlock() :
-	@return : nb of bytes read from src (< srcSize ) */
-size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx *dctx, const void *src, size_t srcSize) /* note : srcSize < BLOCKSIZE */
-{
-	if (srcSize < MIN_CBLOCK_SIZE)
-		return ERROR(corruption_detected);
-
-	{
-		const BYTE *const istart = (const BYTE *)src;
-		symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3);
-
-		switch (litEncType) {
-		case set_repeat:
-			if (dctx->litEntropy == 0)
-				return ERROR(dictionary_corrupted);
-			fallthrough;
-		case set_compressed:
-			if (srcSize < 5)
-				return ERROR(corruption_detected); /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need up to 5 for case 3 */
-			{
-				size_t lhSize, litSize, litCSize;
-				U32 singleStream = 0;
-				U32 const lhlCode = (istart[0] >> 2) & 3;
-				U32 const lhc = ZSTD_readLE32(istart);
-				switch (lhlCode) {
-				case 0:
-				case 1:
-				default: /* note : default is impossible, since lhlCode into [0..3] */
-					/* 2 - 2 - 10 - 10 */
-					singleStream = !lhlCode;
-					lhSize = 3;
-					litSize = (lhc >> 4) & 0x3FF;
-					litCSize = (lhc >> 14) & 0x3FF;
-					break;
-				case 2:
-					/* 2 - 2 - 14 - 14 */
-					lhSize = 4;
-					litSize = (lhc >> 4) & 0x3FFF;
-					litCSize = lhc >> 18;
-					break;
-				case 3:
-					/* 2 - 2 - 18 - 18 */
-					lhSize = 5;
-					litSize = (lhc >> 4) & 0x3FFFF;
-					litCSize = (lhc >> 22) + (istart[4] << 10);
-					break;
-				}
-				if (litSize > ZSTD_BLOCKSIZE_ABSOLUTEMAX)
-					return ERROR(corruption_detected);
-				if (litCSize + lhSize > srcSize)
-					return ERROR(corruption_detected);
-
-				if (HUF_isError(
-					(litEncType == set_repeat)
-					    ? (singleStream ? HUF_decompress1X_usingDTable(dctx->litBuffer, litSize, istart + lhSize, litCSize, dctx->HUFptr)
-							    : HUF_decompress4X_usingDTable(dctx->litBuffer, litSize, istart + lhSize, litCSize, dctx->HUFptr))
-					    : (singleStream
-						   ? HUF_decompress1X2_DCtx_wksp(dctx->entropy.hufTable, dctx->litBuffer, litSize, istart + lhSize, litCSize,
-										 dctx->entropy.workspace, sizeof(dctx->entropy.workspace))
-						   : HUF_decompress4X_hufOnly_wksp(dctx->entropy.hufTable, dctx->litBuffer, litSize, istart + lhSize, litCSize,
-										   dctx->entropy.workspace, sizeof(dctx->entropy.workspace)))))
-					return ERROR(corruption_detected);
-
-				dctx->litPtr = dctx->litBuffer;
-				dctx->litSize = litSize;
-				dctx->litEntropy = 1;
-				if (litEncType == set_compressed)
-					dctx->HUFptr = dctx->entropy.hufTable;
-				memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);
-				return litCSize + lhSize;
-			}
-
-		case set_basic: {
-			size_t litSize, lhSize;
-			U32 const lhlCode = ((istart[0]) >> 2) & 3;
-			switch (lhlCode) {
-			case 0:
-			case 2:
-			default: /* note : default is impossible, since lhlCode into [0..3] */
-				lhSize = 1;
-				litSize = istart[0] >> 3;
-				break;
-			case 1:
-				lhSize = 2;
-				litSize = ZSTD_readLE16(istart) >> 4;
-				break;
-			case 3:
-				lhSize = 3;
-				litSize = ZSTD_readLE24(istart) >> 4;
-				break;
-			}
-
-			if (lhSize + litSize + WILDCOPY_OVERLENGTH > srcSize) { /* risk reading beyond src buffer with wildcopy */
-				if (litSize + lhSize > srcSize)
-					return ERROR(corruption_detected);
-				memcpy(dctx->litBuffer, istart + lhSize, litSize);
-				dctx->litPtr = dctx->litBuffer;
-				dctx->litSize = litSize;
-				memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);
-				return lhSize + litSize;
-			}
-			/* direct reference into compressed stream */
-			dctx->litPtr = istart + lhSize;
-			dctx->litSize = litSize;
-			return lhSize + litSize;
-		}
-
-		case set_rle: {
-			U32 const lhlCode = ((istart[0]) >> 2) & 3;
-			size_t litSize, lhSize;
-			switch (lhlCode) {
-			case 0:
-			case 2:
-			default: /* note : default is impossible, since lhlCode into [0..3] */
-				lhSize = 1;
-				litSize = istart[0] >> 3;
-				break;
-			case 1:
-				lhSize = 2;
-				litSize = ZSTD_readLE16(istart) >> 4;
-				break;
-			case 3:
-				lhSize = 3;
-				litSize = ZSTD_readLE24(istart) >> 4;
-				if (srcSize < 4)
-					return ERROR(corruption_detected); /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need lhSize+1 = 4 */
-				break;
-			}
-			if (litSize > ZSTD_BLOCKSIZE_ABSOLUTEMAX)
-				return ERROR(corruption_detected);
-			memset(dctx->litBuffer, istart[lhSize], litSize + WILDCOPY_OVERLENGTH);
-			dctx->litPtr = dctx->litBuffer;
-			dctx->litSize = litSize;
-			return lhSize + 1;
-		}
-		default:
-			return ERROR(corruption_detected); /* impossible */
-		}
-	}
-}
-
-typedef union {
-	FSE_decode_t realData;
-	U32 alignedBy4;
-} FSE_decode_t4;
-
-static const FSE_decode_t4 LL_defaultDTable[(1 << LL_DEFAULTNORMLOG) + 1] = {
-    {{LL_DEFAULTNORMLOG, 1, 1}}, /* header : tableLog, fastMode, fastMode */
-    {{0, 0, 4}},		 /* 0 : base, symbol, bits */
-    {{16, 0, 4}},
-    {{32, 1, 5}},
-    {{0, 3, 5}},
-    {{0, 4, 5}},
-    {{0, 6, 5}},
-    {{0, 7, 5}},
-    {{0, 9, 5}},
-    {{0, 10, 5}},
-    {{0, 12, 5}},
-    {{0, 14, 6}},
-    {{0, 16, 5}},
-    {{0, 18, 5}},
-    {{0, 19, 5}},
-    {{0, 21, 5}},
-    {{0, 22, 5}},
-    {{0, 24, 5}},
-    {{32, 25, 5}},
-    {{0, 26, 5}},
-    {{0, 27, 6}},
-    {{0, 29, 6}},
-    {{0, 31, 6}},
-    {{32, 0, 4}},
-    {{0, 1, 4}},
-    {{0, 2, 5}},
-    {{32, 4, 5}},
-    {{0, 5, 5}},
-    {{32, 7, 5}},
-    {{0, 8, 5}},
-    {{32, 10, 5}},
-    {{0, 11, 5}},
-    {{0, 13, 6}},
-    {{32, 16, 5}},
-    {{0, 17, 5}},
-    {{32, 19, 5}},
-    {{0, 20, 5}},
-    {{32, 22, 5}},
-    {{0, 23, 5}},
-    {{0, 25, 4}},
-    {{16, 25, 4}},
-    {{32, 26, 5}},
-    {{0, 28, 6}},
-    {{0, 30, 6}},
-    {{48, 0, 4}},
-    {{16, 1, 4}},
-    {{32, 2, 5}},
-    {{32, 3, 5}},
-    {{32, 5, 5}},
-    {{32, 6, 5}},
-    {{32, 8, 5}},
-    {{32, 9, 5}},
-    {{32, 11, 5}},
-    {{32, 12, 5}},
-    {{0, 15, 6}},
-    {{32, 17, 5}},
-    {{32, 18, 5}},
-    {{32, 20, 5}},
-    {{32, 21, 5}},
-    {{32, 23, 5}},
-    {{32, 24, 5}},
-    {{0, 35, 6}},
-    {{0, 34, 6}},
-    {{0, 33, 6}},
-    {{0, 32, 6}},
-}; /* LL_defaultDTable */
-
-static const FSE_decode_t4 ML_defaultDTable[(1 << ML_DEFAULTNORMLOG) + 1] = {
-    {{ML_DEFAULTNORMLOG, 1, 1}}, /* header : tableLog, fastMode, fastMode */
-    {{0, 0, 6}},		 /* 0 : base, symbol, bits */
-    {{0, 1, 4}},
-    {{32, 2, 5}},
-    {{0, 3, 5}},
-    {{0, 5, 5}},
-    {{0, 6, 5}},
-    {{0, 8, 5}},
-    {{0, 10, 6}},
-    {{0, 13, 6}},
-    {{0, 16, 6}},
-    {{0, 19, 6}},
-    {{0, 22, 6}},
-    {{0, 25, 6}},
-    {{0, 28, 6}},
-    {{0, 31, 6}},
-    {{0, 33, 6}},
-    {{0, 35, 6}},
-    {{0, 37, 6}},
-    {{0, 39, 6}},
-    {{0, 41, 6}},
-    {{0, 43, 6}},
-    {{0, 45, 6}},
-    {{16, 1, 4}},
-    {{0, 2, 4}},
-    {{32, 3, 5}},
-    {{0, 4, 5}},
-    {{32, 6, 5}},
-    {{0, 7, 5}},
-    {{0, 9, 6}},
-    {{0, 12, 6}},
-    {{0, 15, 6}},
-    {{0, 18, 6}},
-    {{0, 21, 6}},
-    {{0, 24, 6}},
-    {{0, 27, 6}},
-    {{0, 30, 6}},
-    {{0, 32, 6}},
-    {{0, 34, 6}},
-    {{0, 36, 6}},
-    {{0, 38, 6}},
-    {{0, 40, 6}},
-    {{0, 42, 6}},
-    {{0, 44, 6}},
-    {{32, 1, 4}},
-    {{48, 1, 4}},
-    {{16, 2, 4}},
-    {{32, 4, 5}},
-    {{32, 5, 5}},
-    {{32, 7, 5}},
-    {{32, 8, 5}},
-    {{0, 11, 6}},
-    {{0, 14, 6}},
-    {{0, 17, 6}},
-    {{0, 20, 6}},
-    {{0, 23, 6}},
-    {{0, 26, 6}},
-    {{0, 29, 6}},
-    {{0, 52, 6}},
-    {{0, 51, 6}},
-    {{0, 50, 6}},
-    {{0, 49, 6}},
-    {{0, 48, 6}},
-    {{0, 47, 6}},
-    {{0, 46, 6}},
-}; /* ML_defaultDTable */
-
-static const FSE_decode_t4 OF_defaultDTable[(1 << OF_DEFAULTNORMLOG) + 1] = {
-    {{OF_DEFAULTNORMLOG, 1, 1}}, /* header : tableLog, fastMode, fastMode */
-    {{0, 0, 5}},		 /* 0 : base, symbol, bits */
-    {{0, 6, 4}},
-    {{0, 9, 5}},
-    {{0, 15, 5}},
-    {{0, 21, 5}},
-    {{0, 3, 5}},
-    {{0, 7, 4}},
-    {{0, 12, 5}},
-    {{0, 18, 5}},
-    {{0, 23, 5}},
-    {{0, 5, 5}},
-    {{0, 8, 4}},
-    {{0, 14, 5}},
-    {{0, 20, 5}},
-    {{0, 2, 5}},
-    {{16, 7, 4}},
-    {{0, 11, 5}},
-    {{0, 17, 5}},
-    {{0, 22, 5}},
-    {{0, 4, 5}},
-    {{16, 8, 4}},
-    {{0, 13, 5}},
-    {{0, 19, 5}},
-    {{0, 1, 5}},
-    {{16, 6, 4}},
-    {{0, 10, 5}},
-    {{0, 16, 5}},
-    {{0, 28, 5}},
-    {{0, 27, 5}},
-    {{0, 26, 5}},
-    {{0, 25, 5}},
-    {{0, 24, 5}},
-}; /* OF_defaultDTable */
-
-/*! ZSTD_buildSeqTable() :
-	@return : nb bytes read from src,
-			  or an error code if it fails, testable with ZSTD_isError()
-*/
-static size_t ZSTD_buildSeqTable(FSE_DTable *DTableSpace, const FSE_DTable **DTablePtr, symbolEncodingType_e type, U32 max, U32 maxLog, const void *src,
-				 size_t srcSize, const FSE_decode_t4 *defaultTable, U32 flagRepeatTable, void *workspace, size_t workspaceSize)
-{
-	const void *const tmpPtr = defaultTable; /* bypass strict aliasing */
-	switch (type) {
-	case set_rle:
-		if (!srcSize)
-			return ERROR(srcSize_wrong);
-		if ((*(const BYTE *)src) > max)
-			return ERROR(corruption_detected);
-		FSE_buildDTable_rle(DTableSpace, *(const BYTE *)src);
-		*DTablePtr = DTableSpace;
-		return 1;
-	case set_basic: *DTablePtr = (const FSE_DTable *)tmpPtr; return 0;
-	case set_repeat:
-		if (!flagRepeatTable)
-			return ERROR(corruption_detected);
-		return 0;
-	default: /* impossible */
-	case set_compressed: {
-		U32 tableLog;
-		S16 *norm = (S16 *)workspace;
-		size_t const spaceUsed32 = ALIGN(sizeof(S16) * (MaxSeq + 1), sizeof(U32)) >> 2;
-
-		if ((spaceUsed32 << 2) > workspaceSize)
-			return ERROR(GENERIC);
-		workspace = (U32 *)workspace + spaceUsed32;
-		workspaceSize -= (spaceUsed32 << 2);
-		{
-			size_t const headerSize = FSE_readNCount(norm, &max, &tableLog, src, srcSize);
-			if (FSE_isError(headerSize))
-				return ERROR(corruption_detected);
-			if (tableLog > maxLog)
-				return ERROR(corruption_detected);
-			FSE_buildDTable_wksp(DTableSpace, norm, max, tableLog, workspace, workspaceSize);
-			*DTablePtr = DTableSpace;
-			return headerSize;
-		}
-	}
-	}
-}
-
-size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx *dctx, int *nbSeqPtr, const void *src, size_t srcSize)
-{
-	const BYTE *const istart = (const BYTE *const)src;
-	const BYTE *const iend = istart + srcSize;
-	const BYTE *ip = istart;
-
-	/* check */
-	if (srcSize < MIN_SEQUENCES_SIZE)
-		return ERROR(srcSize_wrong);
-
-	/* SeqHead */
-	{
-		int nbSeq = *ip++;
-		if (!nbSeq) {
-			*nbSeqPtr = 0;
-			return 1;
-		}
-		if (nbSeq > 0x7F) {
-			if (nbSeq == 0xFF) {
-				if (ip + 2 > iend)
-					return ERROR(srcSize_wrong);
-				nbSeq = ZSTD_readLE16(ip) + LONGNBSEQ, ip += 2;
-			} else {
-				if (ip >= iend)
-					return ERROR(srcSize_wrong);
-				nbSeq = ((nbSeq - 0x80) << 8) + *ip++;
-			}
-		}
-		*nbSeqPtr = nbSeq;
-	}
-
-	/* FSE table descriptors */
-	if (ip + 4 > iend)
-		return ERROR(srcSize_wrong); /* minimum possible size */
-	{
-		symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6);
-		symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3);
-		symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3);
-		ip++;
-
-		/* Build DTables */
-		{
-			size_t const llhSize = ZSTD_buildSeqTable(dctx->entropy.LLTable, &dctx->LLTptr, LLtype, MaxLL, LLFSELog, ip, iend - ip,
-								  LL_defaultDTable, dctx->fseEntropy, dctx->entropy.workspace, sizeof(dctx->entropy.workspace));
-			if (ZSTD_isError(llhSize))
-				return ERROR(corruption_detected);
-			ip += llhSize;
-		}
-		{
-			size_t const ofhSize = ZSTD_buildSeqTable(dctx->entropy.OFTable, &dctx->OFTptr, OFtype, MaxOff, OffFSELog, ip, iend - ip,
-								  OF_defaultDTable, dctx->fseEntropy, dctx->entropy.workspace, sizeof(dctx->entropy.workspace));
-			if (ZSTD_isError(ofhSize))
-				return ERROR(corruption_detected);
-			ip += ofhSize;
-		}
-		{
-			size_t const mlhSize = ZSTD_buildSeqTable(dctx->entropy.MLTable, &dctx->MLTptr, MLtype, MaxML, MLFSELog, ip, iend - ip,
-								  ML_defaultDTable, dctx->fseEntropy, dctx->entropy.workspace, sizeof(dctx->entropy.workspace));
-			if (ZSTD_isError(mlhSize))
-				return ERROR(corruption_detected);
-			ip += mlhSize;
-		}
-	}
-
-	return ip - istart;
-}
-
-typedef struct {
-	size_t litLength;
-	size_t matchLength;
-	size_t offset;
-	const BYTE *match;
-} seq_t;
-
-typedef struct {
-	BIT_DStream_t DStream;
-	FSE_DState_t stateLL;
-	FSE_DState_t stateOffb;
-	FSE_DState_t stateML;
-	size_t prevOffset[ZSTD_REP_NUM];
-	const BYTE *base;
-	size_t pos;
-	uPtrDiff gotoDict;
-} seqState_t;
-
-FORCE_NOINLINE
-size_t ZSTD_execSequenceLast7(BYTE *op, BYTE *const oend, seq_t sequence, const BYTE **litPtr, const BYTE *const litLimit, const BYTE *const base,
-			      const BYTE *const vBase, const BYTE *const dictEnd)
-{
-	BYTE *const oLitEnd = op + sequence.litLength;
-	size_t const sequenceLength = sequence.litLength + sequence.matchLength;
-	BYTE *const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */
-	BYTE *const oend_w = oend - WILDCOPY_OVERLENGTH;
-	const BYTE *const iLitEnd = *litPtr + sequence.litLength;
-	const BYTE *match = oLitEnd - sequence.offset;
-
-	/* check */
-	if (oMatchEnd > oend)
-		return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */
-	if (iLitEnd > litLimit)
-		return ERROR(corruption_detected); /* over-read beyond lit buffer */
-	if (oLitEnd <= oend_w)
-		return ERROR(GENERIC); /* Precondition */
-
-	/* copy literals */
-	if (op < oend_w) {
-		ZSTD_wildcopy(op, *litPtr, oend_w - op);
-		*litPtr += oend_w - op;
-		op = oend_w;
-	}
-	while (op < oLitEnd)
-		*op++ = *(*litPtr)++;
-
-	/* copy Match */
-	if (sequence.offset > (size_t)(oLitEnd - base)) {
-		/* offset beyond prefix */
-		if (sequence.offset > (size_t)(oLitEnd - vBase))
-			return ERROR(corruption_detected);
-		match = dictEnd - (base - match);
-		if (match + sequence.matchLength <= dictEnd) {
-			memmove(oLitEnd, match, sequence.matchLength);
-			return sequenceLength;
-		}
-		/* span extDict & currPrefixSegment */
-		{
-			size_t const length1 = dictEnd - match;
-			memmove(oLitEnd, match, length1);
-			op = oLitEnd + length1;
-			sequence.matchLength -= length1;
-			match = base;
-		}
-	}
-	while (op < oMatchEnd)
-		*op++ = *match++;
-	return sequenceLength;
-}
-
-static seq_t ZSTD_decodeSequence(seqState_t *seqState)
-{
-	seq_t seq;
-
-	U32 const llCode = FSE_peekSymbol(&seqState->stateLL);
-	U32 const mlCode = FSE_peekSymbol(&seqState->stateML);
-	U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb); /* <= maxOff, by table construction */
-
-	U32 const llBits = LL_bits[llCode];
-	U32 const mlBits = ML_bits[mlCode];
-	U32 const ofBits = ofCode;
-	U32 const totalBits = llBits + mlBits + ofBits;
-
-	static const U32 LL_base[MaxLL + 1] = {0,  1,  2,  3,  4,  5,  6,  7,  8,    9,     10,    11,    12,    13,     14,     15,     16,     18,
-					       20, 22, 24, 28, 32, 40, 48, 64, 0x80, 0x100, 0x200, 0x400, 0x800, 0x1000, 0x2000, 0x4000, 0x8000, 0x10000};
-
-	static const U32 ML_base[MaxML + 1] = {3,  4,  5,  6,  7,  8,  9,  10,   11,    12,    13,    14,    15,     16,     17,     18,     19,     20,
-					       21, 22, 23, 24, 25, 26, 27, 28,   29,    30,    31,    32,    33,     34,     35,     37,     39,     41,
-					       43, 47, 51, 59, 67, 83, 99, 0x83, 0x103, 0x203, 0x403, 0x803, 0x1003, 0x2003, 0x4003, 0x8003, 0x10003};
-
-	static const U32 OF_base[MaxOff + 1] = {0,       1,	1,	5,	0xD,      0x1D,      0x3D,      0x7D,      0xFD,     0x1FD,
-						0x3FD,   0x7FD,    0xFFD,    0x1FFD,   0x3FFD,   0x7FFD,    0xFFFD,    0x1FFFD,   0x3FFFD,  0x7FFFD,
-						0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD, 0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD};
-
-	/* sequence */
-	{
-		size_t offset;
-		if (!ofCode)
-			offset = 0;
-		else {
-			offset = OF_base[ofCode] + BIT_readBitsFast(&seqState->DStream, ofBits); /* <=  (ZSTD_WINDOWLOG_MAX-1) bits */
-			if (ZSTD_32bits())
-				BIT_reloadDStream(&seqState->DStream);
-		}
-
-		if (ofCode <= 1) {
-			offset += (llCode == 0);
-			if (offset) {
-				size_t temp = (offset == 3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];
-				temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */
-				if (offset != 1)
-					seqState->prevOffset[2] = seqState->prevOffset[1];
-				seqState->prevOffset[1] = seqState->prevOffset[0];
-				seqState->prevOffset[0] = offset = temp;
-			} else {
-				offset = seqState->prevOffset[0];
-			}
-		} else {
-			seqState->prevOffset[2] = seqState->prevOffset[1];
-			seqState->prevOffset[1] = seqState->prevOffset[0];
-			seqState->prevOffset[0] = offset;
-		}
-		seq.offset = offset;
-	}
-
-	seq.matchLength = ML_base[mlCode] + ((mlCode > 31) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <=  16 bits */
-	if (ZSTD_32bits() && (mlBits + llBits > 24))
-		BIT_reloadDStream(&seqState->DStream);
-
-	seq.litLength = LL_base[llCode] + ((llCode > 15) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <=  16 bits */
-	if (ZSTD_32bits() || (totalBits > 64 - 7 - (LLFSELog + MLFSELog + OffFSELog)))
-		BIT_reloadDStream(&seqState->DStream);
-
-	/* ANS state update */
-	FSE_updateState(&seqState->stateLL, &seqState->DStream); /* <=  9 bits */
-	FSE_updateState(&seqState->stateML, &seqState->DStream); /* <=  9 bits */
-	if (ZSTD_32bits())
-		BIT_reloadDStream(&seqState->DStream);		   /* <= 18 bits */
-	FSE_updateState(&seqState->stateOffb, &seqState->DStream); /* <=  8 bits */
-
-	seq.match = NULL;
-
-	return seq;
-}
-
-FORCE_INLINE
-size_t ZSTD_execSequence(BYTE *op, BYTE *const oend, seq_t sequence, const BYTE **litPtr, const BYTE *const litLimit, const BYTE *const base,
-			 const BYTE *const vBase, const BYTE *const dictEnd)
-{
-	BYTE *const oLitEnd = op + sequence.litLength;
-	size_t const sequenceLength = sequence.litLength + sequence.matchLength;
-	BYTE *const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */
-	BYTE *const oend_w = oend - WILDCOPY_OVERLENGTH;
-	const BYTE *const iLitEnd = *litPtr + sequence.litLength;
-	const BYTE *match = oLitEnd - sequence.offset;
-
-	/* check */
-	if (oMatchEnd > oend)
-		return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */
-	if (iLitEnd > litLimit)
-		return ERROR(corruption_detected); /* over-read beyond lit buffer */
-	if (oLitEnd > oend_w)
-		return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, base, vBase, dictEnd);
-
-	/* copy Literals */
-	ZSTD_copy8(op, *litPtr);
-	if (sequence.litLength > 8)
-		ZSTD_wildcopy(op + 8, (*litPtr) + 8,
-			      sequence.litLength - 8); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */
-	op = oLitEnd;
-	*litPtr = iLitEnd; /* update for next sequence */
-
-	/* copy Match */
-	if (sequence.offset > (size_t)(oLitEnd - base)) {
-		/* offset beyond prefix */
-		if (sequence.offset > (size_t)(oLitEnd - vBase))
-			return ERROR(corruption_detected);
-		match = dictEnd + (match - base);
-		if (match + sequence.matchLength <= dictEnd) {
-			memmove(oLitEnd, match, sequence.matchLength);
-			return sequenceLength;
-		}
-		/* span extDict & currPrefixSegment */
-		{
-			size_t const length1 = dictEnd - match;
-			memmove(oLitEnd, match, length1);
-			op = oLitEnd + length1;
-			sequence.matchLength -= length1;
-			match = base;
-			if (op > oend_w || sequence.matchLength < MINMATCH) {
-				U32 i;
-				for (i = 0; i < sequence.matchLength; ++i)
-					op[i] = match[i];
-				return sequenceLength;
-			}
-		}
-	}
-	/* Requirement: op <= oend_w && sequence.matchLength >= MINMATCH */
-
-	/* match within prefix */
-	if (sequence.offset < 8) {
-		/* close range match, overlap */
-		static const U32 dec32table[] = {0, 1, 2, 1, 4, 4, 4, 4};   /* added */
-		static const int dec64table[] = {8, 8, 8, 7, 8, 9, 10, 11}; /* subtracted */
-		int const sub2 = dec64table[sequence.offset];
-		op[0] = match[0];
-		op[1] = match[1];
-		op[2] = match[2];
-		op[3] = match[3];
-		match += dec32table[sequence.offset];
-		ZSTD_copy4(op + 4, match);
-		match -= sub2;
-	} else {
-		ZSTD_copy8(op, match);
-	}
-	op += 8;
-	match += 8;
-
-	if (oMatchEnd > oend - (16 - MINMATCH)) {
-		if (op < oend_w) {
-			ZSTD_wildcopy(op, match, oend_w - op);
-			match += oend_w - op;
-			op = oend_w;
-		}
-		while (op < oMatchEnd)
-			*op++ = *match++;
-	} else {
-		ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength - 8); /* works even if matchLength < 8 */
-	}
-	return sequenceLength;
-}
-
-static size_t ZSTD_decompressSequences(ZSTD_DCtx *dctx, void *dst, size_t maxDstSize, const void *seqStart, size_t seqSize)
-{
-	const BYTE *ip = (const BYTE *)seqStart;
-	const BYTE *const iend = ip + seqSize;
-	BYTE *const ostart = (BYTE * const)dst;
-	BYTE *const oend = ostart + maxDstSize;
-	BYTE *op = ostart;
-	const BYTE *litPtr = dctx->litPtr;
-	const BYTE *const litEnd = litPtr + dctx->litSize;
-	const BYTE *const base = (const BYTE *)(dctx->base);
-	const BYTE *const vBase = (const BYTE *)(dctx->vBase);
-	const BYTE *const dictEnd = (const BYTE *)(dctx->dictEnd);
-	int nbSeq;
-
-	/* Build Decoding Tables */
-	{
-		size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize);
-		if (ZSTD_isError(seqHSize))
-			return seqHSize;
-		ip += seqHSize;
-	}
-
-	/* Regen sequences */
-	if (nbSeq) {
-		seqState_t seqState;
-		dctx->fseEntropy = 1;
-		{
-			U32 i;
-			for (i = 0; i < ZSTD_REP_NUM; i++)
-				seqState.prevOffset[i] = dctx->entropy.rep[i];
-		}
-		CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend - ip), corruption_detected);
-		FSE_initDState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
-		FSE_initDState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
-		FSE_initDState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
-
-		for (; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq;) {
-			nbSeq--;
-			{
-				seq_t const sequence = ZSTD_decodeSequence(&seqState);
-				size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, base, vBase, dictEnd);
-				if (ZSTD_isError(oneSeqSize))
-					return oneSeqSize;
-				op += oneSeqSize;
-			}
-		}
-
-		/* check if reached exact end */
-		if (nbSeq)
-			return ERROR(corruption_detected);
-		/* save reps for next block */
-		{
-			U32 i;
-			for (i = 0; i < ZSTD_REP_NUM; i++)
-				dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]);
-		}
-	}
-
-	/* last literal segment */
-	{
-		size_t const lastLLSize = litEnd - litPtr;
-		if (lastLLSize > (size_t)(oend - op))
-			return ERROR(dstSize_tooSmall);
-		memcpy(op, litPtr, lastLLSize);
-		op += lastLLSize;
-	}
-
-	return op - ostart;
-}
-
-FORCE_INLINE seq_t ZSTD_decodeSequenceLong_generic(seqState_t *seqState, int const longOffsets)
-{
-	seq_t seq;
-
-	U32 const llCode = FSE_peekSymbol(&seqState->stateLL);
-	U32 const mlCode = FSE_peekSymbol(&seqState->stateML);
-	U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb); /* <= maxOff, by table construction */
-
-	U32 const llBits = LL_bits[llCode];
-	U32 const mlBits = ML_bits[mlCode];
-	U32 const ofBits = ofCode;
-	U32 const totalBits = llBits + mlBits + ofBits;
-
-	static const U32 LL_base[MaxLL + 1] = {0,  1,  2,  3,  4,  5,  6,  7,  8,    9,     10,    11,    12,    13,     14,     15,     16,     18,
-					       20, 22, 24, 28, 32, 40, 48, 64, 0x80, 0x100, 0x200, 0x400, 0x800, 0x1000, 0x2000, 0x4000, 0x8000, 0x10000};
-
-	static const U32 ML_base[MaxML + 1] = {3,  4,  5,  6,  7,  8,  9,  10,   11,    12,    13,    14,    15,     16,     17,     18,     19,     20,
-					       21, 22, 23, 24, 25, 26, 27, 28,   29,    30,    31,    32,    33,     34,     35,     37,     39,     41,
-					       43, 47, 51, 59, 67, 83, 99, 0x83, 0x103, 0x203, 0x403, 0x803, 0x1003, 0x2003, 0x4003, 0x8003, 0x10003};
-
-	static const U32 OF_base[MaxOff + 1] = {0,       1,	1,	5,	0xD,      0x1D,      0x3D,      0x7D,      0xFD,     0x1FD,
-						0x3FD,   0x7FD,    0xFFD,    0x1FFD,   0x3FFD,   0x7FFD,    0xFFFD,    0x1FFFD,   0x3FFFD,  0x7FFFD,
-						0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD, 0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD};
-
-	/* sequence */
-	{
-		size_t offset;
-		if (!ofCode)
-			offset = 0;
-		else {
-			if (longOffsets) {
-				int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN);
-				offset = OF_base[ofCode] + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);
-				if (ZSTD_32bits() || extraBits)
-					BIT_reloadDStream(&seqState->DStream);
-				if (extraBits)
-					offset += BIT_readBitsFast(&seqState->DStream, extraBits);
-			} else {
-				offset = OF_base[ofCode] + BIT_readBitsFast(&seqState->DStream, ofBits); /* <=  (ZSTD_WINDOWLOG_MAX-1) bits */
-				if (ZSTD_32bits())
-					BIT_reloadDStream(&seqState->DStream);
-			}
-		}
-
-		if (ofCode <= 1) {
-			offset += (llCode == 0);
-			if (offset) {
-				size_t temp = (offset == 3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];
-				temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */
-				if (offset != 1)
-					seqState->prevOffset[2] = seqState->prevOffset[1];
-				seqState->prevOffset[1] = seqState->prevOffset[0];
-				seqState->prevOffset[0] = offset = temp;
-			} else {
-				offset = seqState->prevOffset[0];
-			}
-		} else {
-			seqState->prevOffset[2] = seqState->prevOffset[1];
-			seqState->prevOffset[1] = seqState->prevOffset[0];
-			seqState->prevOffset[0] = offset;
-		}
-		seq.offset = offset;
-	}
-
-	seq.matchLength = ML_base[mlCode] + ((mlCode > 31) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <=  16 bits */
-	if (ZSTD_32bits() && (mlBits + llBits > 24))
-		BIT_reloadDStream(&seqState->DStream);
-
-	seq.litLength = LL_base[llCode] + ((llCode > 15) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <=  16 bits */
-	if (ZSTD_32bits() || (totalBits > 64 - 7 - (LLFSELog + MLFSELog + OffFSELog)))
-		BIT_reloadDStream(&seqState->DStream);
-
-	{
-		size_t const pos = seqState->pos + seq.litLength;
-		seq.match = seqState->base + pos - seq.offset; /* single memory segment */
-		if (seq.offset > pos)
-			seq.match += seqState->gotoDict; /* separate memory segment */
-		seqState->pos = pos + seq.matchLength;
-	}
-
-	/* ANS state update */
-	FSE_updateState(&seqState->stateLL, &seqState->DStream); /* <=  9 bits */
-	FSE_updateState(&seqState->stateML, &seqState->DStream); /* <=  9 bits */
-	if (ZSTD_32bits())
-		BIT_reloadDStream(&seqState->DStream);		   /* <= 18 bits */
-	FSE_updateState(&seqState->stateOffb, &seqState->DStream); /* <=  8 bits */
-
-	return seq;
-}
-
-static seq_t ZSTD_decodeSequenceLong(seqState_t *seqState, unsigned const windowSize)
-{
-	if (ZSTD_highbit32(windowSize) > STREAM_ACCUMULATOR_MIN) {
-		return ZSTD_decodeSequenceLong_generic(seqState, 1);
-	} else {
-		return ZSTD_decodeSequenceLong_generic(seqState, 0);
-	}
-}
-
-FORCE_INLINE
-size_t ZSTD_execSequenceLong(BYTE *op, BYTE *const oend, seq_t sequence, const BYTE **litPtr, const BYTE *const litLimit, const BYTE *const base,
-			     const BYTE *const vBase, const BYTE *const dictEnd)
-{
-	BYTE *const oLitEnd = op + sequence.litLength;
-	size_t const sequenceLength = sequence.litLength + sequence.matchLength;
-	BYTE *const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */
-	BYTE *const oend_w = oend - WILDCOPY_OVERLENGTH;
-	const BYTE *const iLitEnd = *litPtr + sequence.litLength;
-	const BYTE *match = sequence.match;
-
-	/* check */
-	if (oMatchEnd > oend)
-		return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */
-	if (iLitEnd > litLimit)
-		return ERROR(corruption_detected); /* over-read beyond lit buffer */
-	if (oLitEnd > oend_w)
-		return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, base, vBase, dictEnd);
-
-	/* copy Literals */
-	ZSTD_copy8(op, *litPtr);
-	if (sequence.litLength > 8)
-		ZSTD_wildcopy(op + 8, (*litPtr) + 8,
-			      sequence.litLength - 8); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */
-	op = oLitEnd;
-	*litPtr = iLitEnd; /* update for next sequence */
-
-	/* copy Match */
-	if (sequence.offset > (size_t)(oLitEnd - base)) {
-		/* offset beyond prefix */
-		if (sequence.offset > (size_t)(oLitEnd - vBase))
-			return ERROR(corruption_detected);
-		if (match + sequence.matchLength <= dictEnd) {
-			memmove(oLitEnd, match, sequence.matchLength);
-			return sequenceLength;
-		}
-		/* span extDict & currPrefixSegment */
-		{
-			size_t const length1 = dictEnd - match;
-			memmove(oLitEnd, match, length1);
-			op = oLitEnd + length1;
-			sequence.matchLength -= length1;
-			match = base;
-			if (op > oend_w || sequence.matchLength < MINMATCH) {
-				U32 i;
-				for (i = 0; i < sequence.matchLength; ++i)
-					op[i] = match[i];
-				return sequenceLength;
-			}
-		}
-	}
-	/* Requirement: op <= oend_w && sequence.matchLength >= MINMATCH */
-
-	/* match within prefix */
-	if (sequence.offset < 8) {
-		/* close range match, overlap */
-		static const U32 dec32table[] = {0, 1, 2, 1, 4, 4, 4, 4};   /* added */
-		static const int dec64table[] = {8, 8, 8, 7, 8, 9, 10, 11}; /* subtracted */
-		int const sub2 = dec64table[sequence.offset];
-		op[0] = match[0];
-		op[1] = match[1];
-		op[2] = match[2];
-		op[3] = match[3];
-		match += dec32table[sequence.offset];
-		ZSTD_copy4(op + 4, match);
-		match -= sub2;
-	} else {
-		ZSTD_copy8(op, match);
-	}
-	op += 8;
-	match += 8;
-
-	if (oMatchEnd > oend - (16 - MINMATCH)) {
-		if (op < oend_w) {
-			ZSTD_wildcopy(op, match, oend_w - op);
-			match += oend_w - op;
-			op = oend_w;
-		}
-		while (op < oMatchEnd)
-			*op++ = *match++;
-	} else {
-		ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength - 8); /* works even if matchLength < 8 */
-	}
-	return sequenceLength;
-}
-
-static size_t ZSTD_decompressSequencesLong(ZSTD_DCtx *dctx, void *dst, size_t maxDstSize, const void *seqStart, size_t seqSize)
-{
-	const BYTE *ip = (const BYTE *)seqStart;
-	const BYTE *const iend = ip + seqSize;
-	BYTE *const ostart = (BYTE * const)dst;
-	BYTE *const oend = ostart + maxDstSize;
-	BYTE *op = ostart;
-	const BYTE *litPtr = dctx->litPtr;
-	const BYTE *const litEnd = litPtr + dctx->litSize;
-	const BYTE *const base = (const BYTE *)(dctx->base);
-	const BYTE *const vBase = (const BYTE *)(dctx->vBase);
-	const BYTE *const dictEnd = (const BYTE *)(dctx->dictEnd);
-	unsigned const windowSize = dctx->fParams.windowSize;
-	int nbSeq;
-
-	/* Build Decoding Tables */
-	{
-		size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize);
-		if (ZSTD_isError(seqHSize))
-			return seqHSize;
-		ip += seqHSize;
-	}
-
-	/* Regen sequences */
-	if (nbSeq) {
-#define STORED_SEQS 4
-#define STOSEQ_MASK (STORED_SEQS - 1)
-#define ADVANCED_SEQS 4
-		seq_t *sequences = (seq_t *)dctx->entropy.workspace;
-		int const seqAdvance = MIN(nbSeq, ADVANCED_SEQS);
-		seqState_t seqState;
-		int seqNb;
-		ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.workspace) >= sizeof(seq_t) * STORED_SEQS);
-		dctx->fseEntropy = 1;
-		{
-			U32 i;
-			for (i = 0; i < ZSTD_REP_NUM; i++)
-				seqState.prevOffset[i] = dctx->entropy.rep[i];
-		}
-		seqState.base = base;
-		seqState.pos = (size_t)(op - base);
-		seqState.gotoDict = (uPtrDiff)dictEnd - (uPtrDiff)base; /* cast to avoid undefined behaviour */
-		CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend - ip), corruption_detected);
-		FSE_initDState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
-		FSE_initDState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
-		FSE_initDState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
-
-		/* prepare in advance */
-		for (seqNb = 0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && seqNb < seqAdvance; seqNb++) {
-			sequences[seqNb] = ZSTD_decodeSequenceLong(&seqState, windowSize);
-		}
-		if (seqNb < seqAdvance)
-			return ERROR(corruption_detected);
-
-		/* decode and decompress */
-		for (; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && seqNb < nbSeq; seqNb++) {
-			seq_t const sequence = ZSTD_decodeSequenceLong(&seqState, windowSize);
-			size_t const oneSeqSize =
-			    ZSTD_execSequenceLong(op, oend, sequences[(seqNb - ADVANCED_SEQS) & STOSEQ_MASK], &litPtr, litEnd, base, vBase, dictEnd);
-			if (ZSTD_isError(oneSeqSize))
-				return oneSeqSize;
-			ZSTD_PREFETCH(sequence.match);
-			sequences[seqNb & STOSEQ_MASK] = sequence;
-			op += oneSeqSize;
-		}
-		if (seqNb < nbSeq)
-			return ERROR(corruption_detected);
-
-		/* finish queue */
-		seqNb -= seqAdvance;
-		for (; seqNb < nbSeq; seqNb++) {
-			size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[seqNb & STOSEQ_MASK], &litPtr, litEnd, base, vBase, dictEnd);
-			if (ZSTD_isError(oneSeqSize))
-				return oneSeqSize;
-			op += oneSeqSize;
-		}
-
-		/* save reps for next block */
-		{
-			U32 i;
-			for (i = 0; i < ZSTD_REP_NUM; i++)
-				dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]);
-		}
-	}
-
-	/* last literal segment */
-	{
-		size_t const lastLLSize = litEnd - litPtr;
-		if (lastLLSize > (size_t)(oend - op))
-			return ERROR(dstSize_tooSmall);
-		memcpy(op, litPtr, lastLLSize);
-		op += lastLLSize;
-	}
-
-	return op - ostart;
-}
-
-static size_t ZSTD_decompressBlock_internal(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{ /* blockType == blockCompressed */
-	const BYTE *ip = (const BYTE *)src;
-
-	if (srcSize >= ZSTD_BLOCKSIZE_ABSOLUTEMAX)
-		return ERROR(srcSize_wrong);
-
-	/* Decode literals section */
-	{
-		size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize);
-		if (ZSTD_isError(litCSize))
-			return litCSize;
-		ip += litCSize;
-		srcSize -= litCSize;
-	}
-	if (sizeof(size_t) > 4) /* do not enable prefetching on 32-bits x86, as it's performance detrimental */
-				/* likely because of register pressure */
-				/* if that's the correct cause, then 32-bits ARM should be affected differently */
-				/* it would be good to test this on ARM real hardware, to see if prefetch version improves speed */
-		if (dctx->fParams.windowSize > (1 << 23))
-			return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize);
-	return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize);
-}
-
-static void ZSTD_checkContinuity(ZSTD_DCtx *dctx, const void *dst)
-{
-	if (dst != dctx->previousDstEnd) { /* not contiguous */
-		dctx->dictEnd = dctx->previousDstEnd;
-		dctx->vBase = (const char *)dst - ((const char *)(dctx->previousDstEnd) - (const char *)(dctx->base));
-		dctx->base = dst;
-		dctx->previousDstEnd = dst;
-	}
-}
-
-size_t ZSTD_decompressBlock(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	size_t dSize;
-	ZSTD_checkContinuity(dctx, dst);
-	dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize);
-	dctx->previousDstEnd = (char *)dst + dSize;
-	return dSize;
-}
-
-/** ZSTD_insertBlock() :
-	insert `src` block into `dctx` history. Useful to track uncompressed blocks. */
-size_t ZSTD_insertBlock(ZSTD_DCtx *dctx, const void *blockStart, size_t blockSize)
-{
-	ZSTD_checkContinuity(dctx, blockStart);
-	dctx->previousDstEnd = (const char *)blockStart + blockSize;
-	return blockSize;
-}
-
-size_t ZSTD_generateNxBytes(void *dst, size_t dstCapacity, BYTE byte, size_t length)
-{
-	if (length > dstCapacity)
-		return ERROR(dstSize_tooSmall);
-	memset(dst, byte, length);
-	return length;
-}
-
-/** ZSTD_findFrameCompressedSize() :
- *  compatible with legacy mode
- *  `src` must point to the start of a ZSTD frame, ZSTD legacy frame, or skippable frame
- *  `srcSize` must be at least as large as the frame contained
- *  @return : the compressed size of the frame starting at `src` */
-size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize)
-{
-	if (srcSize >= ZSTD_skippableHeaderSize && (ZSTD_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
-		return ZSTD_skippableHeaderSize + ZSTD_readLE32((const BYTE *)src + 4);
-	} else {
-		const BYTE *ip = (const BYTE *)src;
-		const BYTE *const ipstart = ip;
-		size_t remainingSize = srcSize;
-		ZSTD_frameParams fParams;
-
-		size_t const headerSize = ZSTD_frameHeaderSize(ip, remainingSize);
-		if (ZSTD_isError(headerSize))
-			return headerSize;
-
-		/* Frame Header */
-		{
-			size_t const ret = ZSTD_getFrameParams(&fParams, ip, remainingSize);
-			if (ZSTD_isError(ret))
-				return ret;
-			if (ret > 0)
-				return ERROR(srcSize_wrong);
-		}
-
-		ip += headerSize;
-		remainingSize -= headerSize;
-
-		/* Loop on each block */
-		while (1) {
-			blockProperties_t blockProperties;
-			size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties);
-			if (ZSTD_isError(cBlockSize))
-				return cBlockSize;
-
-			if (ZSTD_blockHeaderSize + cBlockSize > remainingSize)
-				return ERROR(srcSize_wrong);
-
-			ip += ZSTD_blockHeaderSize + cBlockSize;
-			remainingSize -= ZSTD_blockHeaderSize + cBlockSize;
-
-			if (blockProperties.lastBlock)
-				break;
-		}
-
-		if (fParams.checksumFlag) { /* Frame content checksum */
-			if (remainingSize < 4)
-				return ERROR(srcSize_wrong);
-			ip += 4;
-			remainingSize -= 4;
-		}
-
-		return ip - ipstart;
-	}
-}
-
-/*! ZSTD_decompressFrame() :
-*   @dctx must be properly initialized */
-static size_t ZSTD_decompressFrame(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void **srcPtr, size_t *srcSizePtr)
-{
-	const BYTE *ip = (const BYTE *)(*srcPtr);
-	BYTE *const ostart = (BYTE * const)dst;
-	BYTE *const oend = ostart + dstCapacity;
-	BYTE *op = ostart;
-	size_t remainingSize = *srcSizePtr;
-
-	/* check */
-	if (remainingSize < ZSTD_frameHeaderSize_min + ZSTD_blockHeaderSize)
-		return ERROR(srcSize_wrong);
-
-	/* Frame Header */
-	{
-		size_t const frameHeaderSize = ZSTD_frameHeaderSize(ip, ZSTD_frameHeaderSize_prefix);
-		if (ZSTD_isError(frameHeaderSize))
-			return frameHeaderSize;
-		if (remainingSize < frameHeaderSize + ZSTD_blockHeaderSize)
-			return ERROR(srcSize_wrong);
-		CHECK_F(ZSTD_decodeFrameHeader(dctx, ip, frameHeaderSize));
-		ip += frameHeaderSize;
-		remainingSize -= frameHeaderSize;
-	}
-
-	/* Loop on each block */
-	while (1) {
-		size_t decodedSize;
-		blockProperties_t blockProperties;
-		size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties);
-		if (ZSTD_isError(cBlockSize))
-			return cBlockSize;
-
-		ip += ZSTD_blockHeaderSize;
-		remainingSize -= ZSTD_blockHeaderSize;
-		if (cBlockSize > remainingSize)
-			return ERROR(srcSize_wrong);
-
-		switch (blockProperties.blockType) {
-		case bt_compressed: decodedSize = ZSTD_decompressBlock_internal(dctx, op, oend - op, ip, cBlockSize); break;
-		case bt_raw: decodedSize = ZSTD_copyRawBlock(op, oend - op, ip, cBlockSize); break;
-		case bt_rle: decodedSize = ZSTD_generateNxBytes(op, oend - op, *ip, blockProperties.origSize); break;
-		case bt_reserved:
-		default: return ERROR(corruption_detected);
-		}
-
-		if (ZSTD_isError(decodedSize))
-			return decodedSize;
-		if (dctx->fParams.checksumFlag)
-			xxh64_update(&dctx->xxhState, op, decodedSize);
-		op += decodedSize;
-		ip += cBlockSize;
-		remainingSize -= cBlockSize;
-		if (blockProperties.lastBlock)
-			break;
-	}
-
-	if (dctx->fParams.checksumFlag) { /* Frame content checksum verification */
-		U32 const checkCalc = (U32)xxh64_digest(&dctx->xxhState);
-		U32 checkRead;
-		if (remainingSize < 4)
-			return ERROR(checksum_wrong);
-		checkRead = ZSTD_readLE32(ip);
-		if (checkRead != checkCalc)
-			return ERROR(checksum_wrong);
-		ip += 4;
-		remainingSize -= 4;
-	}
-
-	/* Allow caller to get size read */
-	*srcPtr = ip;
-	*srcSizePtr = remainingSize;
-	return op - ostart;
-}
-
-static const void *ZSTD_DDictDictContent(const ZSTD_DDict *ddict);
-static size_t ZSTD_DDictDictSize(const ZSTD_DDict *ddict);
-
-static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const void *dict, size_t dictSize,
-					const ZSTD_DDict *ddict)
-{
-	void *const dststart = dst;
-
-	if (ddict) {
-		if (dict) {
-			/* programmer error, these two cases should be mutually exclusive */
-			return ERROR(GENERIC);
-		}
-
-		dict = ZSTD_DDictDictContent(ddict);
-		dictSize = ZSTD_DDictDictSize(ddict);
-	}
-
-	while (srcSize >= ZSTD_frameHeaderSize_prefix) {
-		U32 magicNumber;
-
-		magicNumber = ZSTD_readLE32(src);
-		if (magicNumber != ZSTD_MAGICNUMBER) {
-			if ((magicNumber & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
-				size_t skippableSize;
-				if (srcSize < ZSTD_skippableHeaderSize)
-					return ERROR(srcSize_wrong);
-				skippableSize = ZSTD_readLE32((const BYTE *)src + 4) + ZSTD_skippableHeaderSize;
-				if (srcSize < skippableSize) {
-					return ERROR(srcSize_wrong);
-				}
-
-				src = (const BYTE *)src + skippableSize;
-				srcSize -= skippableSize;
-				continue;
-			} else {
-				return ERROR(prefix_unknown);
-			}
-		}
-
-		if (ddict) {
-			/* we were called from ZSTD_decompress_usingDDict */
-			ZSTD_refDDict(dctx, ddict);
-		} else {
-			/* this will initialize correctly with no dict if dict == NULL, so
-			 * use this in all cases but ddict */
-			CHECK_F(ZSTD_decompressBegin_usingDict(dctx, dict, dictSize));
-		}
-		ZSTD_checkContinuity(dctx, dst);
-
-		{
-			const size_t res = ZSTD_decompressFrame(dctx, dst, dstCapacity, &src, &srcSize);
-			if (ZSTD_isError(res))
-				return res;
-			/* don't need to bounds check this, ZSTD_decompressFrame will have
-			 * already */
-			dst = (BYTE *)dst + res;
-			dstCapacity -= res;
-		}
-	}
-
-	if (srcSize)
-		return ERROR(srcSize_wrong); /* input not entirely consumed */
-
-	return (BYTE *)dst - (BYTE *)dststart;
-}
-
-size_t ZSTD_decompress_usingDict(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const void *dict, size_t dictSize)
-{
-	return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize, dict, dictSize, NULL);
-}
-
-size_t ZSTD_decompressDCtx(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	return ZSTD_decompress_usingDict(dctx, dst, dstCapacity, src, srcSize, NULL, 0);
-}
-
-/*-**************************************
-*   Advanced Streaming Decompression API
-*   Bufferless and synchronous
-****************************************/
-size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx *dctx) { return dctx->expected; }
-
-ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx *dctx)
-{
-	switch (dctx->stage) {
-	default: /* should not happen */
-	case ZSTDds_getFrameHeaderSize:
-	case ZSTDds_decodeFrameHeader: return ZSTDnit_frameHeader;
-	case ZSTDds_decodeBlockHeader: return ZSTDnit_blockHeader;
-	case ZSTDds_decompressBlock: return ZSTDnit_block;
-	case ZSTDds_decompressLastBlock: return ZSTDnit_lastBlock;
-	case ZSTDds_checkChecksum: return ZSTDnit_checksum;
-	case ZSTDds_decodeSkippableHeader:
-	case ZSTDds_skipFrame: return ZSTDnit_skippableFrame;
-	}
-}
-
-int ZSTD_isSkipFrame(ZSTD_DCtx *dctx) { return dctx->stage == ZSTDds_skipFrame; } /* for zbuff */
-
-/** ZSTD_decompressContinue() :
-*   @return : nb of bytes generated into `dst` (necessarily <= `dstCapacity)
-*             or an error code, which can be tested using ZSTD_isError() */
-size_t ZSTD_decompressContinue(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	/* Sanity check */
-	if (srcSize != dctx->expected)
-		return ERROR(srcSize_wrong);
-	if (dstCapacity)
-		ZSTD_checkContinuity(dctx, dst);
-
-	switch (dctx->stage) {
-	case ZSTDds_getFrameHeaderSize:
-		if (srcSize != ZSTD_frameHeaderSize_prefix)
-			return ERROR(srcSize_wrong);					/* impossible */
-		if ((ZSTD_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */
-			memcpy(dctx->headerBuffer, src, ZSTD_frameHeaderSize_prefix);
-			dctx->expected = ZSTD_skippableHeaderSize - ZSTD_frameHeaderSize_prefix; /* magic number + skippable frame length */
-			dctx->stage = ZSTDds_decodeSkippableHeader;
-			return 0;
-		}
-		dctx->headerSize = ZSTD_frameHeaderSize(src, ZSTD_frameHeaderSize_prefix);
-		if (ZSTD_isError(dctx->headerSize))
-			return dctx->headerSize;
-		memcpy(dctx->headerBuffer, src, ZSTD_frameHeaderSize_prefix);
-		if (dctx->headerSize > ZSTD_frameHeaderSize_prefix) {
-			dctx->expected = dctx->headerSize - ZSTD_frameHeaderSize_prefix;
-			dctx->stage = ZSTDds_decodeFrameHeader;
-			return 0;
-		}
-		dctx->expected = 0; /* not necessary to copy more */
-		fallthrough;
-
-	case ZSTDds_decodeFrameHeader:
-		memcpy(dctx->headerBuffer + ZSTD_frameHeaderSize_prefix, src, dctx->expected);
-		CHECK_F(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize));
-		dctx->expected = ZSTD_blockHeaderSize;
-		dctx->stage = ZSTDds_decodeBlockHeader;
-		return 0;
-
-	case ZSTDds_decodeBlockHeader: {
-		blockProperties_t bp;
-		size_t const cBlockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp);
-		if (ZSTD_isError(cBlockSize))
-			return cBlockSize;
-		dctx->expected = cBlockSize;
-		dctx->bType = bp.blockType;
-		dctx->rleSize = bp.origSize;
-		if (cBlockSize) {
-			dctx->stage = bp.lastBlock ? ZSTDds_decompressLastBlock : ZSTDds_decompressBlock;
-			return 0;
-		}
-		/* empty block */
-		if (bp.lastBlock) {
-			if (dctx->fParams.checksumFlag) {
-				dctx->expected = 4;
-				dctx->stage = ZSTDds_checkChecksum;
-			} else {
-				dctx->expected = 0; /* end of frame */
-				dctx->stage = ZSTDds_getFrameHeaderSize;
-			}
-		} else {
-			dctx->expected = 3; /* go directly to next header */
-			dctx->stage = ZSTDds_decodeBlockHeader;
-		}
-		return 0;
-	}
-	case ZSTDds_decompressLastBlock:
-	case ZSTDds_decompressBlock: {
-		size_t rSize;
-		switch (dctx->bType) {
-		case bt_compressed: rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize); break;
-		case bt_raw: rSize = ZSTD_copyRawBlock(dst, dstCapacity, src, srcSize); break;
-		case bt_rle: rSize = ZSTD_setRleBlock(dst, dstCapacity, src, srcSize, dctx->rleSize); break;
-		case bt_reserved: /* should never happen */
-		default: return ERROR(corruption_detected);
-		}
-		if (ZSTD_isError(rSize))
-			return rSize;
-		if (dctx->fParams.checksumFlag)
-			xxh64_update(&dctx->xxhState, dst, rSize);
-
-		if (dctx->stage == ZSTDds_decompressLastBlock) { /* end of frame */
-			if (dctx->fParams.checksumFlag) {	/* another round for frame checksum */
-				dctx->expected = 4;
-				dctx->stage = ZSTDds_checkChecksum;
-			} else {
-				dctx->expected = 0; /* ends here */
-				dctx->stage = ZSTDds_getFrameHeaderSize;
-			}
-		} else {
-			dctx->stage = ZSTDds_decodeBlockHeader;
-			dctx->expected = ZSTD_blockHeaderSize;
-			dctx->previousDstEnd = (char *)dst + rSize;
-		}
-		return rSize;
-	}
-	case ZSTDds_checkChecksum: {
-		U32 const h32 = (U32)xxh64_digest(&dctx->xxhState);
-		U32 const check32 = ZSTD_readLE32(src); /* srcSize == 4, guaranteed by dctx->expected */
-		if (check32 != h32)
-			return ERROR(checksum_wrong);
-		dctx->expected = 0;
-		dctx->stage = ZSTDds_getFrameHeaderSize;
-		return 0;
-	}
-	case ZSTDds_decodeSkippableHeader: {
-		memcpy(dctx->headerBuffer + ZSTD_frameHeaderSize_prefix, src, dctx->expected);
-		dctx->expected = ZSTD_readLE32(dctx->headerBuffer + 4);
-		dctx->stage = ZSTDds_skipFrame;
-		return 0;
-	}
-	case ZSTDds_skipFrame: {
-		dctx->expected = 0;
-		dctx->stage = ZSTDds_getFrameHeaderSize;
-		return 0;
-	}
-	default:
-		return ERROR(GENERIC); /* impossible */
-	}
-}
-
-static size_t ZSTD_refDictContent(ZSTD_DCtx *dctx, const void *dict, size_t dictSize)
-{
-	dctx->dictEnd = dctx->previousDstEnd;
-	dctx->vBase = (const char *)dict - ((const char *)(dctx->previousDstEnd) - (const char *)(dctx->base));
-	dctx->base = dict;
-	dctx->previousDstEnd = (const char *)dict + dictSize;
-	return 0;
-}
-
-/* ZSTD_loadEntropy() :
- * dict : must point at beginning of a valid zstd dictionary
- * @return : size of entropy tables read */
-static size_t ZSTD_loadEntropy(ZSTD_entropyTables_t *entropy, const void *const dict, size_t const dictSize)
-{
-	const BYTE *dictPtr = (const BYTE *)dict;
-	const BYTE *const dictEnd = dictPtr + dictSize;
-
-	if (dictSize <= 8)
-		return ERROR(dictionary_corrupted);
-	dictPtr += 8; /* skip header = magic + dictID */
-
-	{
-		size_t const hSize = HUF_readDTableX4_wksp(entropy->hufTable, dictPtr, dictEnd - dictPtr, entropy->workspace, sizeof(entropy->workspace));
-		if (HUF_isError(hSize))
-			return ERROR(dictionary_corrupted);
-		dictPtr += hSize;
-	}
-
-	{
-		short offcodeNCount[MaxOff + 1];
-		U32 offcodeMaxValue = MaxOff, offcodeLog;
-		size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd - dictPtr);
-		if (FSE_isError(offcodeHeaderSize))
-			return ERROR(dictionary_corrupted);
-		if (offcodeLog > OffFSELog)
-			return ERROR(dictionary_corrupted);
-		CHECK_E(FSE_buildDTable_wksp(entropy->OFTable, offcodeNCount, offcodeMaxValue, offcodeLog, entropy->workspace, sizeof(entropy->workspace)), dictionary_corrupted);
-		dictPtr += offcodeHeaderSize;
-	}
-
-	{
-		short matchlengthNCount[MaxML + 1];
-		unsigned matchlengthMaxValue = MaxML, matchlengthLog;
-		size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd - dictPtr);
-		if (FSE_isError(matchlengthHeaderSize))
-			return ERROR(dictionary_corrupted);
-		if (matchlengthLog > MLFSELog)
-			return ERROR(dictionary_corrupted);
-		CHECK_E(FSE_buildDTable_wksp(entropy->MLTable, matchlengthNCount, matchlengthMaxValue, matchlengthLog, entropy->workspace, sizeof(entropy->workspace)), dictionary_corrupted);
-		dictPtr += matchlengthHeaderSize;
-	}
-
-	{
-		short litlengthNCount[MaxLL + 1];
-		unsigned litlengthMaxValue = MaxLL, litlengthLog;
-		size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd - dictPtr);
-		if (FSE_isError(litlengthHeaderSize))
-			return ERROR(dictionary_corrupted);
-		if (litlengthLog > LLFSELog)
-			return ERROR(dictionary_corrupted);
-		CHECK_E(FSE_buildDTable_wksp(entropy->LLTable, litlengthNCount, litlengthMaxValue, litlengthLog, entropy->workspace, sizeof(entropy->workspace)), dictionary_corrupted);
-		dictPtr += litlengthHeaderSize;
-	}
-
-	if (dictPtr + 12 > dictEnd)
-		return ERROR(dictionary_corrupted);
-	{
-		int i;
-		size_t const dictContentSize = (size_t)(dictEnd - (dictPtr + 12));
-		for (i = 0; i < 3; i++) {
-			U32 const rep = ZSTD_readLE32(dictPtr);
-			dictPtr += 4;
-			if (rep == 0 || rep >= dictContentSize)
-				return ERROR(dictionary_corrupted);
-			entropy->rep[i] = rep;
-		}
-	}
-
-	return dictPtr - (const BYTE *)dict;
-}
-
-static size_t ZSTD_decompress_insertDictionary(ZSTD_DCtx *dctx, const void *dict, size_t dictSize)
-{
-	if (dictSize < 8)
-		return ZSTD_refDictContent(dctx, dict, dictSize);
-	{
-		U32 const magic = ZSTD_readLE32(dict);
-		if (magic != ZSTD_DICT_MAGIC) {
-			return ZSTD_refDictContent(dctx, dict, dictSize); /* pure content mode */
-		}
-	}
-	dctx->dictID = ZSTD_readLE32((const char *)dict + 4);
-
-	/* load entropy tables */
-	{
-		size_t const eSize = ZSTD_loadEntropy(&dctx->entropy, dict, dictSize);
-		if (ZSTD_isError(eSize))
-			return ERROR(dictionary_corrupted);
-		dict = (const char *)dict + eSize;
-		dictSize -= eSize;
-	}
-	dctx->litEntropy = dctx->fseEntropy = 1;
-
-	/* reference dictionary content */
-	return ZSTD_refDictContent(dctx, dict, dictSize);
-}
-
-size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx *dctx, const void *dict, size_t dictSize)
-{
-	CHECK_F(ZSTD_decompressBegin(dctx));
-	if (dict && dictSize)
-		CHECK_E(ZSTD_decompress_insertDictionary(dctx, dict, dictSize), dictionary_corrupted);
-	return 0;
-}
-
-/* ======   ZSTD_DDict   ====== */
-
-struct ZSTD_DDict_s {
-	void *dictBuffer;
-	const void *dictContent;
-	size_t dictSize;
-	ZSTD_entropyTables_t entropy;
-	U32 dictID;
-	U32 entropyPresent;
-	ZSTD_customMem cMem;
-}; /* typedef'd to ZSTD_DDict within "zstd.h" */
-
-size_t ZSTD_DDictWorkspaceBound(void) { return ZSTD_ALIGN(sizeof(ZSTD_stack)) + ZSTD_ALIGN(sizeof(ZSTD_DDict)); }
-
-static const void *ZSTD_DDictDictContent(const ZSTD_DDict *ddict) { return ddict->dictContent; }
-
-static size_t ZSTD_DDictDictSize(const ZSTD_DDict *ddict) { return ddict->dictSize; }
-
-static void ZSTD_refDDict(ZSTD_DCtx *dstDCtx, const ZSTD_DDict *ddict)
-{
-	ZSTD_decompressBegin(dstDCtx); /* init */
-	if (ddict) {		       /* support refDDict on NULL */
-		dstDCtx->dictID = ddict->dictID;
-		dstDCtx->base = ddict->dictContent;
-		dstDCtx->vBase = ddict->dictContent;
-		dstDCtx->dictEnd = (const BYTE *)ddict->dictContent + ddict->dictSize;
-		dstDCtx->previousDstEnd = dstDCtx->dictEnd;
-		if (ddict->entropyPresent) {
-			dstDCtx->litEntropy = 1;
-			dstDCtx->fseEntropy = 1;
-			dstDCtx->LLTptr = ddict->entropy.LLTable;
-			dstDCtx->MLTptr = ddict->entropy.MLTable;
-			dstDCtx->OFTptr = ddict->entropy.OFTable;
-			dstDCtx->HUFptr = ddict->entropy.hufTable;
-			dstDCtx->entropy.rep[0] = ddict->entropy.rep[0];
-			dstDCtx->entropy.rep[1] = ddict->entropy.rep[1];
-			dstDCtx->entropy.rep[2] = ddict->entropy.rep[2];
-		} else {
-			dstDCtx->litEntropy = 0;
-			dstDCtx->fseEntropy = 0;
-		}
-	}
-}
-
-static size_t ZSTD_loadEntropy_inDDict(ZSTD_DDict *ddict)
-{
-	ddict->dictID = 0;
-	ddict->entropyPresent = 0;
-	if (ddict->dictSize < 8)
-		return 0;
-	{
-		U32 const magic = ZSTD_readLE32(ddict->dictContent);
-		if (magic != ZSTD_DICT_MAGIC)
-			return 0; /* pure content mode */
-	}
-	ddict->dictID = ZSTD_readLE32((const char *)ddict->dictContent + 4);
-
-	/* load entropy tables */
-	CHECK_E(ZSTD_loadEntropy(&ddict->entropy, ddict->dictContent, ddict->dictSize), dictionary_corrupted);
-	ddict->entropyPresent = 1;
-	return 0;
-}
-
-static ZSTD_DDict *ZSTD_createDDict_advanced(const void *dict, size_t dictSize, unsigned byReference, ZSTD_customMem customMem)
-{
-	if (!customMem.customAlloc || !customMem.customFree)
-		return NULL;
-
-	{
-		ZSTD_DDict *const ddict = (ZSTD_DDict *)ZSTD_malloc(sizeof(ZSTD_DDict), customMem);
-		if (!ddict)
-			return NULL;
-		ddict->cMem = customMem;
-
-		if ((byReference) || (!dict) || (!dictSize)) {
-			ddict->dictBuffer = NULL;
-			ddict->dictContent = dict;
-		} else {
-			void *const internalBuffer = ZSTD_malloc(dictSize, customMem);
-			if (!internalBuffer) {
-				ZSTD_freeDDict(ddict);
-				return NULL;
-			}
-			memcpy(internalBuffer, dict, dictSize);
-			ddict->dictBuffer = internalBuffer;
-			ddict->dictContent = internalBuffer;
-		}
-		ddict->dictSize = dictSize;
-		ddict->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */
-		/* parse dictionary content */
-		{
-			size_t const errorCode = ZSTD_loadEntropy_inDDict(ddict);
-			if (ZSTD_isError(errorCode)) {
-				ZSTD_freeDDict(ddict);
-				return NULL;
-			}
-		}
-
-		return ddict;
-	}
-}
-
-/*! ZSTD_initDDict() :
-*   Create a digested dictionary, to start decompression without startup delay.
-*   `dict` content is copied inside DDict.
-*   Consequently, `dict` can be released after `ZSTD_DDict` creation */
-ZSTD_DDict *ZSTD_initDDict(const void *dict, size_t dictSize, void *workspace, size_t workspaceSize)
-{
-	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-	return ZSTD_createDDict_advanced(dict, dictSize, 1, stackMem);
-}
-
-size_t ZSTD_freeDDict(ZSTD_DDict *ddict)
-{
-	if (ddict == NULL)
-		return 0; /* support free on NULL */
-	{
-		ZSTD_customMem const cMem = ddict->cMem;
-		ZSTD_free(ddict->dictBuffer, cMem);
-		ZSTD_free(ddict, cMem);
-		return 0;
-	}
-}
-
-/*! ZSTD_getDictID_fromDict() :
- *  Provides the dictID stored within dictionary.
- *  if @return == 0, the dictionary is not conformant with Zstandard specification.
- *  It can still be loaded, but as a content-only dictionary. */
-unsigned ZSTD_getDictID_fromDict(const void *dict, size_t dictSize)
-{
-	if (dictSize < 8)
-		return 0;
-	if (ZSTD_readLE32(dict) != ZSTD_DICT_MAGIC)
-		return 0;
-	return ZSTD_readLE32((const char *)dict + 4);
-}
-
-/*! ZSTD_getDictID_fromDDict() :
- *  Provides the dictID of the dictionary loaded into `ddict`.
- *  If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
- *  Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */
-unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict *ddict)
-{
-	if (ddict == NULL)
-		return 0;
-	return ZSTD_getDictID_fromDict(ddict->dictContent, ddict->dictSize);
-}
-
-/*! ZSTD_getDictID_fromFrame() :
- *  Provides the dictID required to decompressed the frame stored within `src`.
- *  If @return == 0, the dictID could not be decoded.
- *  This could for one of the following reasons :
- *  - The frame does not require a dictionary to be decoded (most common case).
- *  - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information.
- *    Note : this use case also happens when using a non-conformant dictionary.
- *  - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`).
- *  - This is not a Zstandard frame.
- *  When identifying the exact failure cause, it's possible to used ZSTD_getFrameParams(), which will provide a more precise error code. */
-unsigned ZSTD_getDictID_fromFrame(const void *src, size_t srcSize)
-{
-	ZSTD_frameParams zfp = {0, 0, 0, 0};
-	size_t const hError = ZSTD_getFrameParams(&zfp, src, srcSize);
-	if (ZSTD_isError(hError))
-		return 0;
-	return zfp.dictID;
-}
-
-/*! ZSTD_decompress_usingDDict() :
-*   Decompression using a pre-digested Dictionary
-*   Use dictionary without significant overhead. */
-size_t ZSTD_decompress_usingDDict(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const ZSTD_DDict *ddict)
-{
-	/* pass content and size in case legacy frames are encountered */
-	return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize, NULL, 0, ddict);
-}
-
-/*=====================================
-*   Streaming decompression
-*====================================*/
-
-typedef enum { zdss_init, zdss_loadHeader, zdss_read, zdss_load, zdss_flush } ZSTD_dStreamStage;
-
-/* *** Resource management *** */
-struct ZSTD_DStream_s {
-	ZSTD_DCtx *dctx;
-	ZSTD_DDict *ddictLocal;
-	const ZSTD_DDict *ddict;
-	ZSTD_frameParams fParams;
-	ZSTD_dStreamStage stage;
-	char *inBuff;
-	size_t inBuffSize;
-	size_t inPos;
-	size_t maxWindowSize;
-	char *outBuff;
-	size_t outBuffSize;
-	size_t outStart;
-	size_t outEnd;
-	size_t blockSize;
-	BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX]; /* tmp buffer to store frame header */
-	size_t lhSize;
-	ZSTD_customMem customMem;
-	void *legacyContext;
-	U32 previousLegacyVersion;
-	U32 legacyVersion;
-	U32 hostageByte;
-}; /* typedef'd to ZSTD_DStream within "zstd.h" */
-
-size_t ZSTD_DStreamWorkspaceBound(size_t maxWindowSize)
-{
-	size_t const blockSize = MIN(maxWindowSize, ZSTD_BLOCKSIZE_ABSOLUTEMAX);
-	size_t const inBuffSize = blockSize;
-	size_t const outBuffSize = maxWindowSize + blockSize + WILDCOPY_OVERLENGTH * 2;
-	return ZSTD_DCtxWorkspaceBound() + ZSTD_ALIGN(sizeof(ZSTD_DStream)) + ZSTD_ALIGN(inBuffSize) + ZSTD_ALIGN(outBuffSize);
-}
-
-static ZSTD_DStream *ZSTD_createDStream_advanced(ZSTD_customMem customMem)
-{
-	ZSTD_DStream *zds;
-
-	if (!customMem.customAlloc || !customMem.customFree)
-		return NULL;
-
-	zds = (ZSTD_DStream *)ZSTD_malloc(sizeof(ZSTD_DStream), customMem);
-	if (zds == NULL)
-		return NULL;
-	memset(zds, 0, sizeof(ZSTD_DStream));
-	memcpy(&zds->customMem, &customMem, sizeof(ZSTD_customMem));
-	zds->dctx = ZSTD_createDCtx_advanced(customMem);
-	if (zds->dctx == NULL) {
-		ZSTD_freeDStream(zds);
-		return NULL;
-	}
-	zds->stage = zdss_init;
-	zds->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;
-	return zds;
-}
-
-ZSTD_DStream *ZSTD_initDStream(size_t maxWindowSize, void *workspace, size_t workspaceSize)
-{
-	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-	ZSTD_DStream *zds = ZSTD_createDStream_advanced(stackMem);
-	if (!zds) {
-		return NULL;
-	}
-
-	zds->maxWindowSize = maxWindowSize;
-	zds->stage = zdss_loadHeader;
-	zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0;
-	ZSTD_freeDDict(zds->ddictLocal);
-	zds->ddictLocal = NULL;
-	zds->ddict = zds->ddictLocal;
-	zds->legacyVersion = 0;
-	zds->hostageByte = 0;
-
-	{
-		size_t const blockSize = MIN(zds->maxWindowSize, ZSTD_BLOCKSIZE_ABSOLUTEMAX);
-		size_t const neededOutSize = zds->maxWindowSize + blockSize + WILDCOPY_OVERLENGTH * 2;
-
-		zds->inBuff = (char *)ZSTD_malloc(blockSize, zds->customMem);
-		zds->inBuffSize = blockSize;
-		zds->outBuff = (char *)ZSTD_malloc(neededOutSize, zds->customMem);
-		zds->outBuffSize = neededOutSize;
-		if (zds->inBuff == NULL || zds->outBuff == NULL) {
-			ZSTD_freeDStream(zds);
-			return NULL;
-		}
-	}
-	return zds;
-}
-
-ZSTD_DStream *ZSTD_initDStream_usingDDict(size_t maxWindowSize, const ZSTD_DDict *ddict, void *workspace, size_t workspaceSize)
-{
-	ZSTD_DStream *zds = ZSTD_initDStream(maxWindowSize, workspace, workspaceSize);
-	if (zds) {
-		zds->ddict = ddict;
-	}
-	return zds;
-}
-
-size_t ZSTD_freeDStream(ZSTD_DStream *zds)
-{
-	if (zds == NULL)
-		return 0; /* support free on null */
-	{
-		ZSTD_customMem const cMem = zds->customMem;
-		ZSTD_freeDCtx(zds->dctx);
-		zds->dctx = NULL;
-		ZSTD_freeDDict(zds->ddictLocal);
-		zds->ddictLocal = NULL;
-		ZSTD_free(zds->inBuff, cMem);
-		zds->inBuff = NULL;
-		ZSTD_free(zds->outBuff, cMem);
-		zds->outBuff = NULL;
-		ZSTD_free(zds, cMem);
-		return 0;
-	}
-}
-
-/* *** Initialization *** */
-
-size_t ZSTD_DStreamInSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX + ZSTD_blockHeaderSize; }
-size_t ZSTD_DStreamOutSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX; }
-
-size_t ZSTD_resetDStream(ZSTD_DStream *zds)
-{
-	zds->stage = zdss_loadHeader;
-	zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0;
-	zds->legacyVersion = 0;
-	zds->hostageByte = 0;
-	return ZSTD_frameHeaderSize_prefix;
-}
-
-/* *****   Decompression   ***** */
-
-ZSTD_STATIC size_t ZSTD_limitCopy(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-{
-	size_t const length = MIN(dstCapacity, srcSize);
-	memcpy(dst, src, length);
-	return length;
-}
-
-size_t ZSTD_decompressStream(ZSTD_DStream *zds, ZSTD_outBuffer *output, ZSTD_inBuffer *input)
-{
-	const char *const istart = (const char *)(input->src) + input->pos;
-	const char *const iend = (const char *)(input->src) + input->size;
-	const char *ip = istart;
-	char *const ostart = (char *)(output->dst) + output->pos;
-	char *const oend = (char *)(output->dst) + output->size;
-	char *op = ostart;
-	U32 someMoreWork = 1;
-
-	while (someMoreWork) {
-		switch (zds->stage) {
-		case zdss_init:
-			ZSTD_resetDStream(zds); /* transparent reset on starting decoding a new frame */
-			fallthrough;
-
-		case zdss_loadHeader: {
-			size_t const hSize = ZSTD_getFrameParams(&zds->fParams, zds->headerBuffer, zds->lhSize);
-			if (ZSTD_isError(hSize))
-				return hSize;
-			if (hSize != 0) {				   /* need more input */
-				size_t const toLoad = hSize - zds->lhSize; /* if hSize!=0, hSize > zds->lhSize */
-				if (toLoad > (size_t)(iend - ip)) {	/* not enough input to load full header */
-					memcpy(zds->headerBuffer + zds->lhSize, ip, iend - ip);
-					zds->lhSize += iend - ip;
-					input->pos = input->size;
-					return (MAX(ZSTD_frameHeaderSize_min, hSize) - zds->lhSize) +
-					       ZSTD_blockHeaderSize; /* remaining header bytes + next block header */
-				}
-				memcpy(zds->headerBuffer + zds->lhSize, ip, toLoad);
-				zds->lhSize = hSize;
-				ip += toLoad;
-				break;
-			}
-
-			/* check for single-pass mode opportunity */
-			if (zds->fParams.frameContentSize && zds->fParams.windowSize /* skippable frame if == 0 */
-			    && (U64)(size_t)(oend - op) >= zds->fParams.frameContentSize) {
-				size_t const cSize = ZSTD_findFrameCompressedSize(istart, iend - istart);
-				if (cSize <= (size_t)(iend - istart)) {
-					size_t const decompressedSize = ZSTD_decompress_usingDDict(zds->dctx, op, oend - op, istart, cSize, zds->ddict);
-					if (ZSTD_isError(decompressedSize))
-						return decompressedSize;
-					ip = istart + cSize;
-					op += decompressedSize;
-					zds->dctx->expected = 0;
-					zds->stage = zdss_init;
-					someMoreWork = 0;
-					break;
-				}
-			}
-
-			/* Consume header */
-			ZSTD_refDDict(zds->dctx, zds->ddict);
-			{
-				size_t const h1Size = ZSTD_nextSrcSizeToDecompress(zds->dctx); /* == ZSTD_frameHeaderSize_prefix */
-				CHECK_F(ZSTD_decompressContinue(zds->dctx, NULL, 0, zds->headerBuffer, h1Size));
-				{
-					size_t const h2Size = ZSTD_nextSrcSizeToDecompress(zds->dctx);
-					CHECK_F(ZSTD_decompressContinue(zds->dctx, NULL, 0, zds->headerBuffer + h1Size, h2Size));
-				}
-			}
-
-			zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN);
-			if (zds->fParams.windowSize > zds->maxWindowSize)
-				return ERROR(frameParameter_windowTooLarge);
-
-			/* Buffers are preallocated, but double check */
-			{
-				size_t const blockSize = MIN(zds->maxWindowSize, ZSTD_BLOCKSIZE_ABSOLUTEMAX);
-				size_t const neededOutSize = zds->maxWindowSize + blockSize + WILDCOPY_OVERLENGTH * 2;
-				if (zds->inBuffSize < blockSize) {
-					return ERROR(GENERIC);
-				}
-				if (zds->outBuffSize < neededOutSize) {
-					return ERROR(GENERIC);
-				}
-				zds->blockSize = blockSize;
-			}
-			zds->stage = zdss_read;
-		}
-			fallthrough;
-
-		case zdss_read: {
-			size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds->dctx);
-			if (neededInSize == 0) { /* end of frame */
-				zds->stage = zdss_init;
-				someMoreWork = 0;
-				break;
-			}
-			if ((size_t)(iend - ip) >= neededInSize) { /* decode directly from src */
-				const int isSkipFrame = ZSTD_isSkipFrame(zds->dctx);
-				size_t const decodedSize = ZSTD_decompressContinue(zds->dctx, zds->outBuff + zds->outStart,
-										   (isSkipFrame ? 0 : zds->outBuffSize - zds->outStart), ip, neededInSize);
-				if (ZSTD_isError(decodedSize))
-					return decodedSize;
-				ip += neededInSize;
-				if (!decodedSize && !isSkipFrame)
-					break; /* this was just a header */
-				zds->outEnd = zds->outStart + decodedSize;
-				zds->stage = zdss_flush;
-				break;
-			}
-			if (ip == iend) {
-				someMoreWork = 0;
-				break;
-			} /* no more input */
-			zds->stage = zdss_load;
-			/* pass-through */
-		}
-			fallthrough;
-
-		case zdss_load: {
-			size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds->dctx);
-			size_t const toLoad = neededInSize - zds->inPos; /* should always be <= remaining space within inBuff */
-			size_t loadedSize;
-			if (toLoad > zds->inBuffSize - zds->inPos)
-				return ERROR(corruption_detected); /* should never happen */
-			loadedSize = ZSTD_limitCopy(zds->inBuff + zds->inPos, toLoad, ip, iend - ip);
-			ip += loadedSize;
-			zds->inPos += loadedSize;
-			if (loadedSize < toLoad) {
-				someMoreWork = 0;
-				break;
-			} /* not enough input, wait for more */
-
-			/* decode loaded input */
-			{
-				const int isSkipFrame = ZSTD_isSkipFrame(zds->dctx);
-				size_t const decodedSize = ZSTD_decompressContinue(zds->dctx, zds->outBuff + zds->outStart, zds->outBuffSize - zds->outStart,
-										   zds->inBuff, neededInSize);
-				if (ZSTD_isError(decodedSize))
-					return decodedSize;
-				zds->inPos = 0; /* input is consumed */
-				if (!decodedSize && !isSkipFrame) {
-					zds->stage = zdss_read;
-					break;
-				} /* this was just a header */
-				zds->outEnd = zds->outStart + decodedSize;
-				zds->stage = zdss_flush;
-				/* pass-through */
-			}
-		}
-			fallthrough;
-
-		case zdss_flush: {
-			size_t const toFlushSize = zds->outEnd - zds->outStart;
-			size_t const flushedSize = ZSTD_limitCopy(op, oend - op, zds->outBuff + zds->outStart, toFlushSize);
-			op += flushedSize;
-			zds->outStart += flushedSize;
-			if (flushedSize == toFlushSize) { /* flush completed */
-				zds->stage = zdss_read;
-				if (zds->outStart + zds->blockSize > zds->outBuffSize)
-					zds->outStart = zds->outEnd = 0;
-				break;
-			}
-			/* cannot complete flush */
-			someMoreWork = 0;
-			break;
-		}
-		default:
-			return ERROR(GENERIC); /* impossible */
-		}
-	}
-
-	/* result */
-	input->pos += (size_t)(ip - istart);
-	output->pos += (size_t)(op - ostart);
-	{
-		size_t nextSrcSizeHint = ZSTD_nextSrcSizeToDecompress(zds->dctx);
-		if (!nextSrcSizeHint) {			    /* frame fully decoded */
-			if (zds->outEnd == zds->outStart) { /* output fully flushed */
-				if (zds->hostageByte) {
-					if (input->pos >= input->size) {
-						zds->stage = zdss_read;
-						return 1;
-					}	     /* can't release hostage (not present) */
-					input->pos++; /* release hostage */
-				}
-				return 0;
-			}
-			if (!zds->hostageByte) { /* output not fully flushed; keep last byte as hostage; will be released when all output is flushed */
-				input->pos--;    /* note : pos > 0, otherwise, impossible to finish reading last block */
-				zds->hostageByte = 1;
-			}
-			return 1;
-		}
-		nextSrcSizeHint += ZSTD_blockHeaderSize * (ZSTD_nextInputType(zds->dctx) == ZSTDnit_block); /* preload header of next block */
-		if (zds->inPos > nextSrcSizeHint)
-			return ERROR(GENERIC); /* should never happen */
-		nextSrcSizeHint -= zds->inPos; /* already loaded*/
-		return nextSrcSizeHint;
-	}
-}
-
-EXPORT_SYMBOL(ZSTD_DCtxWorkspaceBound);
-EXPORT_SYMBOL(ZSTD_initDCtx);
-EXPORT_SYMBOL(ZSTD_decompressDCtx);
-EXPORT_SYMBOL(ZSTD_decompress_usingDict);
-
-EXPORT_SYMBOL(ZSTD_DDictWorkspaceBound);
-EXPORT_SYMBOL(ZSTD_initDDict);
-EXPORT_SYMBOL(ZSTD_decompress_usingDDict);
-
-EXPORT_SYMBOL(ZSTD_DStreamWorkspaceBound);
-EXPORT_SYMBOL(ZSTD_initDStream);
-EXPORT_SYMBOL(ZSTD_initDStream_usingDDict);
-EXPORT_SYMBOL(ZSTD_resetDStream);
-EXPORT_SYMBOL(ZSTD_decompressStream);
-EXPORT_SYMBOL(ZSTD_DStreamInSize);
-EXPORT_SYMBOL(ZSTD_DStreamOutSize);
-
-EXPORT_SYMBOL(ZSTD_findFrameCompressedSize);
-EXPORT_SYMBOL(ZSTD_getFrameContentSize);
-EXPORT_SYMBOL(ZSTD_findDecompressedSize);
-
-EXPORT_SYMBOL(ZSTD_isFrame);
-EXPORT_SYMBOL(ZSTD_getDictID_fromDict);
-EXPORT_SYMBOL(ZSTD_getDictID_fromDDict);
-EXPORT_SYMBOL(ZSTD_getDictID_fromFrame);
-
-EXPORT_SYMBOL(ZSTD_getFrameParams);
-EXPORT_SYMBOL(ZSTD_decompressBegin);
-EXPORT_SYMBOL(ZSTD_decompressBegin_usingDict);
-EXPORT_SYMBOL(ZSTD_copyDCtx);
-EXPORT_SYMBOL(ZSTD_nextSrcSizeToDecompress);
-EXPORT_SYMBOL(ZSTD_decompressContinue);
-EXPORT_SYMBOL(ZSTD_nextInputType);
-
-EXPORT_SYMBOL(ZSTD_decompressBlock);
-EXPORT_SYMBOL(ZSTD_insertBlock);
-
-MODULE_LICENSE("Dual BSD/GPL");
-MODULE_DESCRIPTION("Zstd Decompressor");
diff --git a/lib/zstd/decompress/huf_decompress.c b/lib/zstd/decompress/huf_decompress.c
new file mode 100644
index 0000000..5105e59
--- /dev/null
+++ b/lib/zstd/decompress/huf_decompress.c
@@ -0,0 +1,1206 @@
+/* ******************************************************************
+ * huff0 huffman decoder,
+ * part of Finite State Entropy library
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ *
+ *  You can contact the author at :
+ *  - FSE+HUF source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+/* **************************************************************
+*  Dependencies
+****************************************************************/
+#include "../common/zstd_deps.h"  /* ZSTD_memcpy, ZSTD_memset */
+#include "../common/compiler.h"
+#include "../common/bitstream.h"  /* BIT_* */
+#include "../common/fse.h"        /* to compress headers */
+#define HUF_STATIC_LINKING_ONLY
+#include "../common/huf.h"
+#include "../common/error_private.h"
+
+/* **************************************************************
+*  Macros
+****************************************************************/
+
+/* These two optional macros force the use one way or another of the two
+ * Huffman decompression implementations. You can't force in both directions
+ * at the same time.
+ */
+#if defined(HUF_FORCE_DECOMPRESS_X1) && \
+    defined(HUF_FORCE_DECOMPRESS_X2)
+#error "Cannot force the use of the X1 and X2 decoders at the same time!"
+#endif
+
+
+/* **************************************************************
+*  Error Management
+****************************************************************/
+#define HUF_isError ERR_isError
+
+
+/* **************************************************************
+*  Byte alignment for workSpace management
+****************************************************************/
+#define HUF_ALIGN(x, a)         HUF_ALIGN_MASK((x), (a) - 1)
+#define HUF_ALIGN_MASK(x, mask) (((x) + (mask)) & ~(mask))
+
+
+/* **************************************************************
+*  BMI2 Variant Wrappers
+****************************************************************/
+#if DYNAMIC_BMI2
+
+#define HUF_DGEN(fn)                                                        \
+                                                                            \
+    static size_t fn##_default(                                             \
+                  void* dst,  size_t dstSize,                               \
+            const void* cSrc, size_t cSrcSize,                              \
+            const HUF_DTable* DTable)                                       \
+    {                                                                       \
+        return fn##_body(dst, dstSize, cSrc, cSrcSize, DTable);             \
+    }                                                                       \
+                                                                            \
+    static TARGET_ATTRIBUTE("bmi2") size_t fn##_bmi2(                       \
+                  void* dst,  size_t dstSize,                               \
+            const void* cSrc, size_t cSrcSize,                              \
+            const HUF_DTable* DTable)                                       \
+    {                                                                       \
+        return fn##_body(dst, dstSize, cSrc, cSrcSize, DTable);             \
+    }                                                                       \
+                                                                            \
+    static size_t fn(void* dst, size_t dstSize, void const* cSrc,           \
+                     size_t cSrcSize, HUF_DTable const* DTable, int bmi2)   \
+    {                                                                       \
+        if (bmi2) {                                                         \
+            return fn##_bmi2(dst, dstSize, cSrc, cSrcSize, DTable);         \
+        }                                                                   \
+        return fn##_default(dst, dstSize, cSrc, cSrcSize, DTable);          \
+    }
+
+#else
+
+#define HUF_DGEN(fn)                                                        \
+    static size_t fn(void* dst, size_t dstSize, void const* cSrc,           \
+                     size_t cSrcSize, HUF_DTable const* DTable, int bmi2)   \
+    {                                                                       \
+        (void)bmi2;                                                         \
+        return fn##_body(dst, dstSize, cSrc, cSrcSize, DTable);             \
+    }
+
+#endif
+
+
+/*-***************************/
+/*  generic DTableDesc       */
+/*-***************************/
+typedef struct { BYTE maxTableLog; BYTE tableType; BYTE tableLog; BYTE reserved; } DTableDesc;
+
+static DTableDesc HUF_getDTableDesc(const HUF_DTable* table)
+{
+    DTableDesc dtd;
+    ZSTD_memcpy(&dtd, table, sizeof(dtd));
+    return dtd;
+}
+
+
+#ifndef HUF_FORCE_DECOMPRESS_X2
+
+/*-***************************/
+/*  single-symbol decoding   */
+/*-***************************/
+typedef struct { BYTE byte; BYTE nbBits; } HUF_DEltX1;   /* single-symbol decoding */
+
+/*
+ * Packs 4 HUF_DEltX1 structs into a U64. This is used to lay down 4 entries at
+ * a time.
+ */
+static U64 HUF_DEltX1_set4(BYTE symbol, BYTE nbBits) {
+    U64 D4;
+    if (MEM_isLittleEndian()) {
+        D4 = symbol + (nbBits << 8);
+    } else {
+        D4 = (symbol << 8) + nbBits;
+    }
+    D4 *= 0x0001000100010001ULL;
+    return D4;
+}
+
+typedef struct {
+        U32 rankVal[HUF_TABLELOG_ABSOLUTEMAX + 1];
+        U32 rankStart[HUF_TABLELOG_ABSOLUTEMAX + 1];
+        U32 statsWksp[HUF_READ_STATS_WORKSPACE_SIZE_U32];
+        BYTE symbols[HUF_SYMBOLVALUE_MAX + 1];
+        BYTE huffWeight[HUF_SYMBOLVALUE_MAX + 1];
+} HUF_ReadDTableX1_Workspace;
+
+
+size_t HUF_readDTableX1_wksp(HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize)
+{
+    return HUF_readDTableX1_wksp_bmi2(DTable, src, srcSize, workSpace, wkspSize, /* bmi2 */ 0);
+}
+
+size_t HUF_readDTableX1_wksp_bmi2(HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize, int bmi2)
+{
+    U32 tableLog = 0;
+    U32 nbSymbols = 0;
+    size_t iSize;
+    void* const dtPtr = DTable + 1;
+    HUF_DEltX1* const dt = (HUF_DEltX1*)dtPtr;
+    HUF_ReadDTableX1_Workspace* wksp = (HUF_ReadDTableX1_Workspace*)workSpace;
+
+    DEBUG_STATIC_ASSERT(HUF_DECOMPRESS_WORKSPACE_SIZE >= sizeof(*wksp));
+    if (sizeof(*wksp) > wkspSize) return ERROR(tableLog_tooLarge);
+
+    DEBUG_STATIC_ASSERT(sizeof(DTableDesc) == sizeof(HUF_DTable));
+    /* ZSTD_memset(huffWeight, 0, sizeof(huffWeight)); */   /* is not necessary, even though some analyzer complain ... */
+
+    iSize = HUF_readStats_wksp(wksp->huffWeight, HUF_SYMBOLVALUE_MAX + 1, wksp->rankVal, &nbSymbols, &tableLog, src, srcSize, wksp->statsWksp, sizeof(wksp->statsWksp), bmi2);
+    if (HUF_isError(iSize)) return iSize;
+
+    /* Table header */
+    {   DTableDesc dtd = HUF_getDTableDesc(DTable);
+        if (tableLog > (U32)(dtd.maxTableLog+1)) return ERROR(tableLog_tooLarge);   /* DTable too small, Huffman tree cannot fit in */
+        dtd.tableType = 0;
+        dtd.tableLog = (BYTE)tableLog;
+        ZSTD_memcpy(DTable, &dtd, sizeof(dtd));
+    }
+
+    /* Compute symbols and rankStart given rankVal:
+     *
+     * rankVal already contains the number of values of each weight.
+     *
+     * symbols contains the symbols ordered by weight. First are the rankVal[0]
+     * weight 0 symbols, followed by the rankVal[1] weight 1 symbols, and so on.
+     * symbols[0] is filled (but unused) to avoid a branch.
+     *
+     * rankStart contains the offset where each rank belongs in the DTable.
+     * rankStart[0] is not filled because there are no entries in the table for
+     * weight 0.
+     */
+    {
+        int n;
+        int nextRankStart = 0;
+        int const unroll = 4;
+        int const nLimit = (int)nbSymbols - unroll + 1;
+        for (n=0; n<(int)tableLog+1; n++) {
+            U32 const curr = nextRankStart;
+            nextRankStart += wksp->rankVal[n];
+            wksp->rankStart[n] = curr;
+        }
+        for (n=0; n < nLimit; n += unroll) {
+            int u;
+            for (u=0; u < unroll; ++u) {
+                size_t const w = wksp->huffWeight[n+u];
+                wksp->symbols[wksp->rankStart[w]++] = (BYTE)(n+u);
+            }
+        }
+        for (; n < (int)nbSymbols; ++n) {
+            size_t const w = wksp->huffWeight[n];
+            wksp->symbols[wksp->rankStart[w]++] = (BYTE)n;
+        }
+    }
+
+    /* fill DTable
+     * We fill all entries of each weight in order.
+     * That way length is a constant for each iteration of the outter loop.
+     * We can switch based on the length to a different inner loop which is
+     * optimized for that particular case.
+     */
+    {
+        U32 w;
+        int symbol=wksp->rankVal[0];
+        int rankStart=0;
+        for (w=1; w<tableLog+1; ++w) {
+            int const symbolCount = wksp->rankVal[w];
+            int const length = (1 << w) >> 1;
+            int uStart = rankStart;
+            BYTE const nbBits = (BYTE)(tableLog + 1 - w);
+            int s;
+            int u;
+            switch (length) {
+            case 1:
+                for (s=0; s<symbolCount; ++s) {
+                    HUF_DEltX1 D;
+                    D.byte = wksp->symbols[symbol + s];
+                    D.nbBits = nbBits;
+                    dt[uStart] = D;
+                    uStart += 1;
+                }
+                break;
+            case 2:
+                for (s=0; s<symbolCount; ++s) {
+                    HUF_DEltX1 D;
+                    D.byte = wksp->symbols[symbol + s];
+                    D.nbBits = nbBits;
+                    dt[uStart+0] = D;
+                    dt[uStart+1] = D;
+                    uStart += 2;
+                }
+                break;
+            case 4:
+                for (s=0; s<symbolCount; ++s) {
+                    U64 const D4 = HUF_DEltX1_set4(wksp->symbols[symbol + s], nbBits);
+                    MEM_write64(dt + uStart, D4);
+                    uStart += 4;
+                }
+                break;
+            case 8:
+                for (s=0; s<symbolCount; ++s) {
+                    U64 const D4 = HUF_DEltX1_set4(wksp->symbols[symbol + s], nbBits);
+                    MEM_write64(dt + uStart, D4);
+                    MEM_write64(dt + uStart + 4, D4);
+                    uStart += 8;
+                }
+                break;
+            default:
+                for (s=0; s<symbolCount; ++s) {
+                    U64 const D4 = HUF_DEltX1_set4(wksp->symbols[symbol + s], nbBits);
+                    for (u=0; u < length; u += 16) {
+                        MEM_write64(dt + uStart + u + 0, D4);
+                        MEM_write64(dt + uStart + u + 4, D4);
+                        MEM_write64(dt + uStart + u + 8, D4);
+                        MEM_write64(dt + uStart + u + 12, D4);
+                    }
+                    assert(u == length);
+                    uStart += length;
+                }
+                break;
+            }
+            symbol += symbolCount;
+            rankStart += symbolCount * length;
+        }
+    }
+    return iSize;
+}
+
+FORCE_INLINE_TEMPLATE BYTE
+HUF_decodeSymbolX1(BIT_DStream_t* Dstream, const HUF_DEltX1* dt, const U32 dtLog)
+{
+    size_t const val = BIT_lookBitsFast(Dstream, dtLog); /* note : dtLog >= 1 */
+    BYTE const c = dt[val].byte;
+    BIT_skipBits(Dstream, dt[val].nbBits);
+    return c;
+}
+
+#define HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr) \
+    *ptr++ = HUF_decodeSymbolX1(DStreamPtr, dt, dtLog)
+
+#define HUF_DECODE_SYMBOLX1_1(ptr, DStreamPtr)  \
+    if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
+        HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr)
+
+#define HUF_DECODE_SYMBOLX1_2(ptr, DStreamPtr) \
+    if (MEM_64bits()) \
+        HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr)
+
+HINT_INLINE size_t
+HUF_decodeStreamX1(BYTE* p, BIT_DStream_t* const bitDPtr, BYTE* const pEnd, const HUF_DEltX1* const dt, const U32 dtLog)
+{
+    BYTE* const pStart = p;
+
+    /* up to 4 symbols at a time */
+    while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p < pEnd-3)) {
+        HUF_DECODE_SYMBOLX1_2(p, bitDPtr);
+        HUF_DECODE_SYMBOLX1_1(p, bitDPtr);
+        HUF_DECODE_SYMBOLX1_2(p, bitDPtr);
+        HUF_DECODE_SYMBOLX1_0(p, bitDPtr);
+    }
+
+    /* [0-3] symbols remaining */
+    if (MEM_32bits())
+        while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p < pEnd))
+            HUF_DECODE_SYMBOLX1_0(p, bitDPtr);
+
+    /* no more data to retrieve from bitstream, no need to reload */
+    while (p < pEnd)
+        HUF_DECODE_SYMBOLX1_0(p, bitDPtr);
+
+    return pEnd-pStart;
+}
+
+FORCE_INLINE_TEMPLATE size_t
+HUF_decompress1X1_usingDTable_internal_body(
+          void* dst,  size_t dstSize,
+    const void* cSrc, size_t cSrcSize,
+    const HUF_DTable* DTable)
+{
+    BYTE* op = (BYTE*)dst;
+    BYTE* const oend = op + dstSize;
+    const void* dtPtr = DTable + 1;
+    const HUF_DEltX1* const dt = (const HUF_DEltX1*)dtPtr;
+    BIT_DStream_t bitD;
+    DTableDesc const dtd = HUF_getDTableDesc(DTable);
+    U32 const dtLog = dtd.tableLog;
+
+    CHECK_F( BIT_initDStream(&bitD, cSrc, cSrcSize) );
+
+    HUF_decodeStreamX1(op, &bitD, oend, dt, dtLog);
+
+    if (!BIT_endOfDStream(&bitD)) return ERROR(corruption_detected);
+
+    return dstSize;
+}
+
+FORCE_INLINE_TEMPLATE size_t
+HUF_decompress4X1_usingDTable_internal_body(
+          void* dst,  size_t dstSize,
+    const void* cSrc, size_t cSrcSize,
+    const HUF_DTable* DTable)
+{
+    /* Check */
+    if (cSrcSize < 10) return ERROR(corruption_detected);  /* strict minimum : jump table + 1 byte per stream */
+
+    {   const BYTE* const istart = (const BYTE*) cSrc;
+        BYTE* const ostart = (BYTE*) dst;
+        BYTE* const oend = ostart + dstSize;
+        BYTE* const olimit = oend - 3;
+        const void* const dtPtr = DTable + 1;
+        const HUF_DEltX1* const dt = (const HUF_DEltX1*)dtPtr;
+
+        /* Init */
+        BIT_DStream_t bitD1;
+        BIT_DStream_t bitD2;
+        BIT_DStream_t bitD3;
+        BIT_DStream_t bitD4;
+        size_t const length1 = MEM_readLE16(istart);
+        size_t const length2 = MEM_readLE16(istart+2);
+        size_t const length3 = MEM_readLE16(istart+4);
+        size_t const length4 = cSrcSize - (length1 + length2 + length3 + 6);
+        const BYTE* const istart1 = istart + 6;  /* jumpTable */
+        const BYTE* const istart2 = istart1 + length1;
+        const BYTE* const istart3 = istart2 + length2;
+        const BYTE* const istart4 = istart3 + length3;
+        const size_t segmentSize = (dstSize+3) / 4;
+        BYTE* const opStart2 = ostart + segmentSize;
+        BYTE* const opStart3 = opStart2 + segmentSize;
+        BYTE* const opStart4 = opStart3 + segmentSize;
+        BYTE* op1 = ostart;
+        BYTE* op2 = opStart2;
+        BYTE* op3 = opStart3;
+        BYTE* op4 = opStart4;
+        DTableDesc const dtd = HUF_getDTableDesc(DTable);
+        U32 const dtLog = dtd.tableLog;
+        U32 endSignal = 1;
+
+        if (length4 > cSrcSize) return ERROR(corruption_detected);   /* overflow */
+        CHECK_F( BIT_initDStream(&bitD1, istart1, length1) );
+        CHECK_F( BIT_initDStream(&bitD2, istart2, length2) );
+        CHECK_F( BIT_initDStream(&bitD3, istart3, length3) );
+        CHECK_F( BIT_initDStream(&bitD4, istart4, length4) );
+
+        /* up to 16 symbols per loop (4 symbols per stream) in 64-bit mode */
+        for ( ; (endSignal) & (op4 < olimit) ; ) {
+            HUF_DECODE_SYMBOLX1_2(op1, &bitD1);
+            HUF_DECODE_SYMBOLX1_2(op2, &bitD2);
+            HUF_DECODE_SYMBOLX1_2(op3, &bitD3);
+            HUF_DECODE_SYMBOLX1_2(op4, &bitD4);
+            HUF_DECODE_SYMBOLX1_1(op1, &bitD1);
+            HUF_DECODE_SYMBOLX1_1(op2, &bitD2);
+            HUF_DECODE_SYMBOLX1_1(op3, &bitD3);
+            HUF_DECODE_SYMBOLX1_1(op4, &bitD4);
+            HUF_DECODE_SYMBOLX1_2(op1, &bitD1);
+            HUF_DECODE_SYMBOLX1_2(op2, &bitD2);
+            HUF_DECODE_SYMBOLX1_2(op3, &bitD3);
+            HUF_DECODE_SYMBOLX1_2(op4, &bitD4);
+            HUF_DECODE_SYMBOLX1_0(op1, &bitD1);
+            HUF_DECODE_SYMBOLX1_0(op2, &bitD2);
+            HUF_DECODE_SYMBOLX1_0(op3, &bitD3);
+            HUF_DECODE_SYMBOLX1_0(op4, &bitD4);
+            endSignal &= BIT_reloadDStreamFast(&bitD1) == BIT_DStream_unfinished;
+            endSignal &= BIT_reloadDStreamFast(&bitD2) == BIT_DStream_unfinished;
+            endSignal &= BIT_reloadDStreamFast(&bitD3) == BIT_DStream_unfinished;
+            endSignal &= BIT_reloadDStreamFast(&bitD4) == BIT_DStream_unfinished;
+        }
+
+        /* check corruption */
+        /* note : should not be necessary : op# advance in lock step, and we control op4.
+         *        but curiously, binary generated by gcc 7.2 & 7.3 with -mbmi2 runs faster when >=1 test is present */
+        if (op1 > opStart2) return ERROR(corruption_detected);
+        if (op2 > opStart3) return ERROR(corruption_detected);
+        if (op3 > opStart4) return ERROR(corruption_detected);
+        /* note : op4 supposed already verified within main loop */
+
+        /* finish bitStreams one by one */
+        HUF_decodeStreamX1(op1, &bitD1, opStart2, dt, dtLog);
+        HUF_decodeStreamX1(op2, &bitD2, opStart3, dt, dtLog);
+        HUF_decodeStreamX1(op3, &bitD3, opStart4, dt, dtLog);
+        HUF_decodeStreamX1(op4, &bitD4, oend,     dt, dtLog);
+
+        /* check */
+        { U32 const endCheck = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);
+          if (!endCheck) return ERROR(corruption_detected); }
+
+        /* decoded size */
+        return dstSize;
+    }
+}
+
+
+typedef size_t (*HUF_decompress_usingDTable_t)(void *dst, size_t dstSize,
+                                               const void *cSrc,
+                                               size_t cSrcSize,
+                                               const HUF_DTable *DTable);
+
+HUF_DGEN(HUF_decompress1X1_usingDTable_internal)
+HUF_DGEN(HUF_decompress4X1_usingDTable_internal)
+
+
+
+size_t HUF_decompress1X1_usingDTable(
+          void* dst,  size_t dstSize,
+    const void* cSrc, size_t cSrcSize,
+    const HUF_DTable* DTable)
+{
+    DTableDesc dtd = HUF_getDTableDesc(DTable);
+    if (dtd.tableType != 0) return ERROR(GENERIC);
+    return HUF_decompress1X1_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
+}
+
+size_t HUF_decompress1X1_DCtx_wksp(HUF_DTable* DCtx, void* dst, size_t dstSize,
+                                   const void* cSrc, size_t cSrcSize,
+                                   void* workSpace, size_t wkspSize)
+{
+    const BYTE* ip = (const BYTE*) cSrc;
+
+    size_t const hSize = HUF_readDTableX1_wksp(DCtx, cSrc, cSrcSize, workSpace, wkspSize);
+    if (HUF_isError(hSize)) return hSize;
+    if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
+    ip += hSize; cSrcSize -= hSize;
+
+    return HUF_decompress1X1_usingDTable_internal(dst, dstSize, ip, cSrcSize, DCtx, /* bmi2 */ 0);
+}
+
+
+size_t HUF_decompress4X1_usingDTable(
+          void* dst,  size_t dstSize,
+    const void* cSrc, size_t cSrcSize,
+    const HUF_DTable* DTable)
+{
+    DTableDesc dtd = HUF_getDTableDesc(DTable);
+    if (dtd.tableType != 0) return ERROR(GENERIC);
+    return HUF_decompress4X1_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
+}
+
+static size_t HUF_decompress4X1_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize,
+                                   const void* cSrc, size_t cSrcSize,
+                                   void* workSpace, size_t wkspSize, int bmi2)
+{
+    const BYTE* ip = (const BYTE*) cSrc;
+
+    size_t const hSize = HUF_readDTableX1_wksp_bmi2(dctx, cSrc, cSrcSize, workSpace, wkspSize, bmi2);
+    if (HUF_isError(hSize)) return hSize;
+    if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
+    ip += hSize; cSrcSize -= hSize;
+
+    return HUF_decompress4X1_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx, bmi2);
+}
+
+size_t HUF_decompress4X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,
+                                   const void* cSrc, size_t cSrcSize,
+                                   void* workSpace, size_t wkspSize)
+{
+    return HUF_decompress4X1_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, 0);
+}
+
+
+#endif /* HUF_FORCE_DECOMPRESS_X2 */
+
+
+#ifndef HUF_FORCE_DECOMPRESS_X1
+
+/* *************************/
+/* double-symbols decoding */
+/* *************************/
+
+typedef struct { U16 sequence; BYTE nbBits; BYTE length; } HUF_DEltX2;  /* double-symbols decoding */
+typedef struct { BYTE symbol; BYTE weight; } sortedSymbol_t;
+typedef U32 rankValCol_t[HUF_TABLELOG_MAX + 1];
+typedef rankValCol_t rankVal_t[HUF_TABLELOG_MAX];
+
+
+/* HUF_fillDTableX2Level2() :
+ * `rankValOrigin` must be a table of at least (HUF_TABLELOG_MAX + 1) U32 */
+static void HUF_fillDTableX2Level2(HUF_DEltX2* DTable, U32 sizeLog, const U32 consumed,
+                           const U32* rankValOrigin, const int minWeight,
+                           const sortedSymbol_t* sortedSymbols, const U32 sortedListSize,
+                           U32 nbBitsBaseline, U16 baseSeq, U32* wksp, size_t wkspSize)
+{
+    HUF_DEltX2 DElt;
+    U32* rankVal = wksp;
+
+    assert(wkspSize >= HUF_TABLELOG_MAX + 1);
+    (void)wkspSize;
+    /* get pre-calculated rankVal */
+    ZSTD_memcpy(rankVal, rankValOrigin, sizeof(U32) * (HUF_TABLELOG_MAX + 1));
+
+    /* fill skipped values */
+    if (minWeight>1) {
+        U32 i, skipSize = rankVal[minWeight];
+        MEM_writeLE16(&(DElt.sequence), baseSeq);
+        DElt.nbBits   = (BYTE)(consumed);
+        DElt.length   = 1;
+        for (i = 0; i < skipSize; i++)
+            DTable[i] = DElt;
+    }
+
+    /* fill DTable */
+    {   U32 s; for (s=0; s<sortedListSize; s++) {   /* note : sortedSymbols already skipped */
+            const U32 symbol = sortedSymbols[s].symbol;
+            const U32 weight = sortedSymbols[s].weight;
+            const U32 nbBits = nbBitsBaseline - weight;
+            const U32 length = 1 << (sizeLog-nbBits);
+            const U32 start = rankVal[weight];
+            U32 i = start;
+            const U32 end = start + length;
+
+            MEM_writeLE16(&(DElt.sequence), (U16)(baseSeq + (symbol << 8)));
+            DElt.nbBits = (BYTE)(nbBits + consumed);
+            DElt.length = 2;
+            do { DTable[i++] = DElt; } while (i<end);   /* since length >= 1 */
+
+            rankVal[weight] += length;
+    }   }
+}
+
+
+static void HUF_fillDTableX2(HUF_DEltX2* DTable, const U32 targetLog,
+                           const sortedSymbol_t* sortedList, const U32 sortedListSize,
+                           const U32* rankStart, rankVal_t rankValOrigin, const U32 maxWeight,
+                           const U32 nbBitsBaseline, U32* wksp, size_t wkspSize)
+{
+    U32* rankVal = wksp;
+    const int scaleLog = nbBitsBaseline - targetLog;   /* note : targetLog >= srcLog, hence scaleLog <= 1 */
+    const U32 minBits  = nbBitsBaseline - maxWeight;
+    U32 s;
+
+    assert(wkspSize >= HUF_TABLELOG_MAX + 1);
+    wksp += HUF_TABLELOG_MAX + 1;
+    wkspSize -= HUF_TABLELOG_MAX + 1;
+
+    ZSTD_memcpy(rankVal, rankValOrigin, sizeof(U32) * (HUF_TABLELOG_MAX + 1));
+
+    /* fill DTable */
+    for (s=0; s<sortedListSize; s++) {
+        const U16 symbol = sortedList[s].symbol;
+        const U32 weight = sortedList[s].weight;
+        const U32 nbBits = nbBitsBaseline - weight;
+        const U32 start = rankVal[weight];
+        const U32 length = 1 << (targetLog-nbBits);
+
+        if (targetLog-nbBits >= minBits) {   /* enough room for a second symbol */
+            U32 sortedRank;
+            int minWeight = nbBits + scaleLog;
+            if (minWeight < 1) minWeight = 1;
+            sortedRank = rankStart[minWeight];
+            HUF_fillDTableX2Level2(DTable+start, targetLog-nbBits, nbBits,
+                           rankValOrigin[nbBits], minWeight,
+                           sortedList+sortedRank, sortedListSize-sortedRank,
+                           nbBitsBaseline, symbol, wksp, wkspSize);
+        } else {
+            HUF_DEltX2 DElt;
+            MEM_writeLE16(&(DElt.sequence), symbol);
+            DElt.nbBits = (BYTE)(nbBits);
+            DElt.length = 1;
+            {   U32 const end = start + length;
+                U32 u;
+                for (u = start; u < end; u++) DTable[u] = DElt;
+        }   }
+        rankVal[weight] += length;
+    }
+}
+
+typedef struct {
+    rankValCol_t rankVal[HUF_TABLELOG_MAX];
+    U32 rankStats[HUF_TABLELOG_MAX + 1];
+    U32 rankStart0[HUF_TABLELOG_MAX + 2];
+    sortedSymbol_t sortedSymbol[HUF_SYMBOLVALUE_MAX + 1];
+    BYTE weightList[HUF_SYMBOLVALUE_MAX + 1];
+    U32 calleeWksp[HUF_READ_STATS_WORKSPACE_SIZE_U32];
+} HUF_ReadDTableX2_Workspace;
+
+size_t HUF_readDTableX2_wksp(HUF_DTable* DTable,
+                       const void* src, size_t srcSize,
+                             void* workSpace, size_t wkspSize)
+{
+    U32 tableLog, maxW, sizeOfSort, nbSymbols;
+    DTableDesc dtd = HUF_getDTableDesc(DTable);
+    U32 const maxTableLog = dtd.maxTableLog;
+    size_t iSize;
+    void* dtPtr = DTable+1;   /* force compiler to avoid strict-aliasing */
+    HUF_DEltX2* const dt = (HUF_DEltX2*)dtPtr;
+    U32 *rankStart;
+
+    HUF_ReadDTableX2_Workspace* const wksp = (HUF_ReadDTableX2_Workspace*)workSpace;
+
+    if (sizeof(*wksp) > wkspSize) return ERROR(GENERIC);
+
+    rankStart = wksp->rankStart0 + 1;
+    ZSTD_memset(wksp->rankStats, 0, sizeof(wksp->rankStats));
+    ZSTD_memset(wksp->rankStart0, 0, sizeof(wksp->rankStart0));
+
+    DEBUG_STATIC_ASSERT(sizeof(HUF_DEltX2) == sizeof(HUF_DTable));   /* if compiler fails here, assertion is wrong */
+    if (maxTableLog > HUF_TABLELOG_MAX) return ERROR(tableLog_tooLarge);
+    /* ZSTD_memset(weightList, 0, sizeof(weightList)); */  /* is not necessary, even though some analyzer complain ... */
+
+    iSize = HUF_readStats_wksp(wksp->weightList, HUF_SYMBOLVALUE_MAX + 1, wksp->rankStats, &nbSymbols, &tableLog, src, srcSize, wksp->calleeWksp, sizeof(wksp->calleeWksp), /* bmi2 */ 0);
+    if (HUF_isError(iSize)) return iSize;
+
+    /* check result */
+    if (tableLog > maxTableLog) return ERROR(tableLog_tooLarge);   /* DTable can't fit code depth */
+
+    /* find maxWeight */
+    for (maxW = tableLog; wksp->rankStats[maxW]==0; maxW--) {}  /* necessarily finds a solution before 0 */
+
+    /* Get start index of each weight */
+    {   U32 w, nextRankStart = 0;
+        for (w=1; w<maxW+1; w++) {
+            U32 curr = nextRankStart;
+            nextRankStart += wksp->rankStats[w];
+            rankStart[w] = curr;
+        }
+        rankStart[0] = nextRankStart;   /* put all 0w symbols at the end of sorted list*/
+        sizeOfSort = nextRankStart;
+    }
+
+    /* sort symbols by weight */
+    {   U32 s;
+        for (s=0; s<nbSymbols; s++) {
+            U32 const w = wksp->weightList[s];
+            U32 const r = rankStart[w]++;
+            wksp->sortedSymbol[r].symbol = (BYTE)s;
+            wksp->sortedSymbol[r].weight = (BYTE)w;
+        }
+        rankStart[0] = 0;   /* forget 0w symbols; this is beginning of weight(1) */
+    }
+
+    /* Build rankVal */
+    {   U32* const rankVal0 = wksp->rankVal[0];
+        {   int const rescale = (maxTableLog-tableLog) - 1;   /* tableLog <= maxTableLog */
+            U32 nextRankVal = 0;
+            U32 w;
+            for (w=1; w<maxW+1; w++) {
+                U32 curr = nextRankVal;
+                nextRankVal += wksp->rankStats[w] << (w+rescale);
+                rankVal0[w] = curr;
+        }   }
+        {   U32 const minBits = tableLog+1 - maxW;
+            U32 consumed;
+            for (consumed = minBits; consumed < maxTableLog - minBits + 1; consumed++) {
+                U32* const rankValPtr = wksp->rankVal[consumed];
+                U32 w;
+                for (w = 1; w < maxW+1; w++) {
+                    rankValPtr[w] = rankVal0[w] >> consumed;
+    }   }   }   }
+
+    HUF_fillDTableX2(dt, maxTableLog,
+                   wksp->sortedSymbol, sizeOfSort,
+                   wksp->rankStart0, wksp->rankVal, maxW,
+                   tableLog+1,
+                   wksp->calleeWksp, sizeof(wksp->calleeWksp) / sizeof(U32));
+
+    dtd.tableLog = (BYTE)maxTableLog;
+    dtd.tableType = 1;
+    ZSTD_memcpy(DTable, &dtd, sizeof(dtd));
+    return iSize;
+}
+
+
+FORCE_INLINE_TEMPLATE U32
+HUF_decodeSymbolX2(void* op, BIT_DStream_t* DStream, const HUF_DEltX2* dt, const U32 dtLog)
+{
+    size_t const val = BIT_lookBitsFast(DStream, dtLog);   /* note : dtLog >= 1 */
+    ZSTD_memcpy(op, dt+val, 2);
+    BIT_skipBits(DStream, dt[val].nbBits);
+    return dt[val].length;
+}
+
+FORCE_INLINE_TEMPLATE U32
+HUF_decodeLastSymbolX2(void* op, BIT_DStream_t* DStream, const HUF_DEltX2* dt, const U32 dtLog)
+{
+    size_t const val = BIT_lookBitsFast(DStream, dtLog);   /* note : dtLog >= 1 */
+    ZSTD_memcpy(op, dt+val, 1);
+    if (dt[val].length==1) BIT_skipBits(DStream, dt[val].nbBits);
+    else {
+        if (DStream->bitsConsumed < (sizeof(DStream->bitContainer)*8)) {
+            BIT_skipBits(DStream, dt[val].nbBits);
+            if (DStream->bitsConsumed > (sizeof(DStream->bitContainer)*8))
+                /* ugly hack; works only because it's the last symbol. Note : can't easily extract nbBits from just this symbol */
+                DStream->bitsConsumed = (sizeof(DStream->bitContainer)*8);
+    }   }
+    return 1;
+}
+
+#define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) \
+    ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog)
+
+#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr) \
+    if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
+        ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog)
+
+#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \
+    if (MEM_64bits()) \
+        ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog)
+
+HINT_INLINE size_t
+HUF_decodeStreamX2(BYTE* p, BIT_DStream_t* bitDPtr, BYTE* const pEnd,
+                const HUF_DEltX2* const dt, const U32 dtLog)
+{
+    BYTE* const pStart = p;
+
+    /* up to 8 symbols at a time */
+    while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p < pEnd-(sizeof(bitDPtr->bitContainer)-1))) {
+        HUF_DECODE_SYMBOLX2_2(p, bitDPtr);
+        HUF_DECODE_SYMBOLX2_1(p, bitDPtr);
+        HUF_DECODE_SYMBOLX2_2(p, bitDPtr);
+        HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
+    }
+
+    /* closer to end : up to 2 symbols at a time */
+    while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p <= pEnd-2))
+        HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
+
+    while (p <= pEnd-2)
+        HUF_DECODE_SYMBOLX2_0(p, bitDPtr);   /* no need to reload : reached the end of DStream */
+
+    if (p < pEnd)
+        p += HUF_decodeLastSymbolX2(p, bitDPtr, dt, dtLog);
+
+    return p-pStart;
+}
+
+FORCE_INLINE_TEMPLATE size_t
+HUF_decompress1X2_usingDTable_internal_body(
+          void* dst,  size_t dstSize,
+    const void* cSrc, size_t cSrcSize,
+    const HUF_DTable* DTable)
+{
+    BIT_DStream_t bitD;
+
+    /* Init */
+    CHECK_F( BIT_initDStream(&bitD, cSrc, cSrcSize) );
+
+    /* decode */
+    {   BYTE* const ostart = (BYTE*) dst;
+        BYTE* const oend = ostart + dstSize;
+        const void* const dtPtr = DTable+1;   /* force compiler to not use strict-aliasing */
+        const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr;
+        DTableDesc const dtd = HUF_getDTableDesc(DTable);
+        HUF_decodeStreamX2(ostart, &bitD, oend, dt, dtd.tableLog);
+    }
+
+    /* check */
+    if (!BIT_endOfDStream(&bitD)) return ERROR(corruption_detected);
+
+    /* decoded size */
+    return dstSize;
+}
+
+FORCE_INLINE_TEMPLATE size_t
+HUF_decompress4X2_usingDTable_internal_body(
+          void* dst,  size_t dstSize,
+    const void* cSrc, size_t cSrcSize,
+    const HUF_DTable* DTable)
+{
+    if (cSrcSize < 10) return ERROR(corruption_detected);   /* strict minimum : jump table + 1 byte per stream */
+
+    {   const BYTE* const istart = (const BYTE*) cSrc;
+        BYTE* const ostart = (BYTE*) dst;
+        BYTE* const oend = ostart + dstSize;
+        BYTE* const olimit = oend - (sizeof(size_t)-1);
+        const void* const dtPtr = DTable+1;
+        const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr;
+
+        /* Init */
+        BIT_DStream_t bitD1;
+        BIT_DStream_t bitD2;
+        BIT_DStream_t bitD3;
+        BIT_DStream_t bitD4;
+        size_t const length1 = MEM_readLE16(istart);
+        size_t const length2 = MEM_readLE16(istart+2);
+        size_t const length3 = MEM_readLE16(istart+4);
+        size_t const length4 = cSrcSize - (length1 + length2 + length3 + 6);
+        const BYTE* const istart1 = istart + 6;  /* jumpTable */
+        const BYTE* const istart2 = istart1 + length1;
+        const BYTE* const istart3 = istart2 + length2;
+        const BYTE* const istart4 = istart3 + length3;
+        size_t const segmentSize = (dstSize+3) / 4;
+        BYTE* const opStart2 = ostart + segmentSize;
+        BYTE* const opStart3 = opStart2 + segmentSize;
+        BYTE* const opStart4 = opStart3 + segmentSize;
+        BYTE* op1 = ostart;
+        BYTE* op2 = opStart2;
+        BYTE* op3 = opStart3;
+        BYTE* op4 = opStart4;
+        U32 endSignal = 1;
+        DTableDesc const dtd = HUF_getDTableDesc(DTable);
+        U32 const dtLog = dtd.tableLog;
+
+        if (length4 > cSrcSize) return ERROR(corruption_detected);   /* overflow */
+        CHECK_F( BIT_initDStream(&bitD1, istart1, length1) );
+        CHECK_F( BIT_initDStream(&bitD2, istart2, length2) );
+        CHECK_F( BIT_initDStream(&bitD3, istart3, length3) );
+        CHECK_F( BIT_initDStream(&bitD4, istart4, length4) );
+
+        /* 16-32 symbols per loop (4-8 symbols per stream) */
+        for ( ; (endSignal) & (op4 < olimit); ) {
+#if defined(__clang__) && (defined(__x86_64__) || defined(__i386__))
+            HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
+            HUF_DECODE_SYMBOLX2_1(op1, &bitD1);
+            HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
+            HUF_DECODE_SYMBOLX2_0(op1, &bitD1);
+            HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
+            HUF_DECODE_SYMBOLX2_1(op2, &bitD2);
+            HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
+            HUF_DECODE_SYMBOLX2_0(op2, &bitD2);
+            endSignal &= BIT_reloadDStreamFast(&bitD1) == BIT_DStream_unfinished;
+            endSignal &= BIT_reloadDStreamFast(&bitD2) == BIT_DStream_unfinished;
+            HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
+            HUF_DECODE_SYMBOLX2_1(op3, &bitD3);
+            HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
+            HUF_DECODE_SYMBOLX2_0(op3, &bitD3);
+            HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
+            HUF_DECODE_SYMBOLX2_1(op4, &bitD4);
+            HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
+            HUF_DECODE_SYMBOLX2_0(op4, &bitD4);
+            endSignal &= BIT_reloadDStreamFast(&bitD3) == BIT_DStream_unfinished;
+            endSignal &= BIT_reloadDStreamFast(&bitD4) == BIT_DStream_unfinished;
+#else
+            HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
+            HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
+            HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
+            HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
+            HUF_DECODE_SYMBOLX2_1(op1, &bitD1);
+            HUF_DECODE_SYMBOLX2_1(op2, &bitD2);
+            HUF_DECODE_SYMBOLX2_1(op3, &bitD3);
+            HUF_DECODE_SYMBOLX2_1(op4, &bitD4);
+            HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
+            HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
+            HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
+            HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
+            HUF_DECODE_SYMBOLX2_0(op1, &bitD1);
+            HUF_DECODE_SYMBOLX2_0(op2, &bitD2);
+            HUF_DECODE_SYMBOLX2_0(op3, &bitD3);
+            HUF_DECODE_SYMBOLX2_0(op4, &bitD4);
+            endSignal = (U32)LIKELY((U32)
+                        (BIT_reloadDStreamFast(&bitD1) == BIT_DStream_unfinished)
+                      & (BIT_reloadDStreamFast(&bitD2) == BIT_DStream_unfinished)
+                      & (BIT_reloadDStreamFast(&bitD3) == BIT_DStream_unfinished)
+                      & (BIT_reloadDStreamFast(&bitD4) == BIT_DStream_unfinished));
+#endif
+        }
+
+        /* check corruption */
+        if (op1 > opStart2) return ERROR(corruption_detected);
+        if (op2 > opStart3) return ERROR(corruption_detected);
+        if (op3 > opStart4) return ERROR(corruption_detected);
+        /* note : op4 already verified within main loop */
+
+        /* finish bitStreams one by one */
+        HUF_decodeStreamX2(op1, &bitD1, opStart2, dt, dtLog);
+        HUF_decodeStreamX2(op2, &bitD2, opStart3, dt, dtLog);
+        HUF_decodeStreamX2(op3, &bitD3, opStart4, dt, dtLog);
+        HUF_decodeStreamX2(op4, &bitD4, oend,     dt, dtLog);
+
+        /* check */
+        { U32 const endCheck = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);
+          if (!endCheck) return ERROR(corruption_detected); }
+
+        /* decoded size */
+        return dstSize;
+    }
+}
+
+HUF_DGEN(HUF_decompress1X2_usingDTable_internal)
+HUF_DGEN(HUF_decompress4X2_usingDTable_internal)
+
+size_t HUF_decompress1X2_usingDTable(
+          void* dst,  size_t dstSize,
+    const void* cSrc, size_t cSrcSize,
+    const HUF_DTable* DTable)
+{
+    DTableDesc dtd = HUF_getDTableDesc(DTable);
+    if (dtd.tableType != 1) return ERROR(GENERIC);
+    return HUF_decompress1X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
+}
+
+size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable* DCtx, void* dst, size_t dstSize,
+                                   const void* cSrc, size_t cSrcSize,
+                                   void* workSpace, size_t wkspSize)
+{
+    const BYTE* ip = (const BYTE*) cSrc;
+
+    size_t const hSize = HUF_readDTableX2_wksp(DCtx, cSrc, cSrcSize,
+                                               workSpace, wkspSize);
+    if (HUF_isError(hSize)) return hSize;
+    if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
+    ip += hSize; cSrcSize -= hSize;
+
+    return HUF_decompress1X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, DCtx, /* bmi2 */ 0);
+}
+
+
+size_t HUF_decompress4X2_usingDTable(
+          void* dst,  size_t dstSize,
+    const void* cSrc, size_t cSrcSize,
+    const HUF_DTable* DTable)
+{
+    DTableDesc dtd = HUF_getDTableDesc(DTable);
+    if (dtd.tableType != 1) return ERROR(GENERIC);
+    return HUF_decompress4X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
+}
+
+static size_t HUF_decompress4X2_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize,
+                                   const void* cSrc, size_t cSrcSize,
+                                   void* workSpace, size_t wkspSize, int bmi2)
+{
+    const BYTE* ip = (const BYTE*) cSrc;
+
+    size_t hSize = HUF_readDTableX2_wksp(dctx, cSrc, cSrcSize,
+                                         workSpace, wkspSize);
+    if (HUF_isError(hSize)) return hSize;
+    if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
+    ip += hSize; cSrcSize -= hSize;
+
+    return HUF_decompress4X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx, bmi2);
+}
+
+size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,
+                                   const void* cSrc, size_t cSrcSize,
+                                   void* workSpace, size_t wkspSize)
+{
+    return HUF_decompress4X2_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, /* bmi2 */ 0);
+}
+
+
+#endif /* HUF_FORCE_DECOMPRESS_X1 */
+
+
+/* ***********************************/
+/* Universal decompression selectors */
+/* ***********************************/
+
+size_t HUF_decompress1X_usingDTable(void* dst, size_t maxDstSize,
+                                    const void* cSrc, size_t cSrcSize,
+                                    const HUF_DTable* DTable)
+{
+    DTableDesc const dtd = HUF_getDTableDesc(DTable);
+#if defined(HUF_FORCE_DECOMPRESS_X1)
+    (void)dtd;
+    assert(dtd.tableType == 0);
+    return HUF_decompress1X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
+#elif defined(HUF_FORCE_DECOMPRESS_X2)
+    (void)dtd;
+    assert(dtd.tableType == 1);
+    return HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
+#else
+    return dtd.tableType ? HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0) :
+                           HUF_decompress1X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
+#endif
+}
+
+size_t HUF_decompress4X_usingDTable(void* dst, size_t maxDstSize,
+                                    const void* cSrc, size_t cSrcSize,
+                                    const HUF_DTable* DTable)
+{
+    DTableDesc const dtd = HUF_getDTableDesc(DTable);
+#if defined(HUF_FORCE_DECOMPRESS_X1)
+    (void)dtd;
+    assert(dtd.tableType == 0);
+    return HUF_decompress4X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
+#elif defined(HUF_FORCE_DECOMPRESS_X2)
+    (void)dtd;
+    assert(dtd.tableType == 1);
+    return HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
+#else
+    return dtd.tableType ? HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0) :
+                           HUF_decompress4X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
+#endif
+}
+
+
+#if !defined(HUF_FORCE_DECOMPRESS_X1) && !defined(HUF_FORCE_DECOMPRESS_X2)
+typedef struct { U32 tableTime; U32 decode256Time; } algo_time_t;
+static const algo_time_t algoTime[16 /* Quantization */][3 /* single, double, quad */] =
+{
+    /* single, double, quad */
+    {{0,0}, {1,1}, {2,2}},  /* Q==0 : impossible */
+    {{0,0}, {1,1}, {2,2}},  /* Q==1 : impossible */
+    {{  38,130}, {1313, 74}, {2151, 38}},   /* Q == 2 : 12-18% */
+    {{ 448,128}, {1353, 74}, {2238, 41}},   /* Q == 3 : 18-25% */
+    {{ 556,128}, {1353, 74}, {2238, 47}},   /* Q == 4 : 25-32% */
+    {{ 714,128}, {1418, 74}, {2436, 53}},   /* Q == 5 : 32-38% */
+    {{ 883,128}, {1437, 74}, {2464, 61}},   /* Q == 6 : 38-44% */
+    {{ 897,128}, {1515, 75}, {2622, 68}},   /* Q == 7 : 44-50% */
+    {{ 926,128}, {1613, 75}, {2730, 75}},   /* Q == 8 : 50-56% */
+    {{ 947,128}, {1729, 77}, {3359, 77}},   /* Q == 9 : 56-62% */
+    {{1107,128}, {2083, 81}, {4006, 84}},   /* Q ==10 : 62-69% */
+    {{1177,128}, {2379, 87}, {4785, 88}},   /* Q ==11 : 69-75% */
+    {{1242,128}, {2415, 93}, {5155, 84}},   /* Q ==12 : 75-81% */
+    {{1349,128}, {2644,106}, {5260,106}},   /* Q ==13 : 81-87% */
+    {{1455,128}, {2422,124}, {4174,124}},   /* Q ==14 : 87-93% */
+    {{ 722,128}, {1891,145}, {1936,146}},   /* Q ==15 : 93-99% */
+};
+#endif
+
+/* HUF_selectDecoder() :
+ *  Tells which decoder is likely to decode faster,
+ *  based on a set of pre-computed metrics.
+ * @return : 0==HUF_decompress4X1, 1==HUF_decompress4X2 .
+ *  Assumption : 0 < dstSize <= 128 KB */
+U32 HUF_selectDecoder (size_t dstSize, size_t cSrcSize)
+{
+    assert(dstSize > 0);
+    assert(dstSize <= 128*1024);
+#if defined(HUF_FORCE_DECOMPRESS_X1)
+    (void)dstSize;
+    (void)cSrcSize;
+    return 0;
+#elif defined(HUF_FORCE_DECOMPRESS_X2)
+    (void)dstSize;
+    (void)cSrcSize;
+    return 1;
+#else
+    /* decoder timing evaluation */
+    {   U32 const Q = (cSrcSize >= dstSize) ? 15 : (U32)(cSrcSize * 16 / dstSize);   /* Q < 16 */
+        U32 const D256 = (U32)(dstSize >> 8);
+        U32 const DTime0 = algoTime[Q][0].tableTime + (algoTime[Q][0].decode256Time * D256);
+        U32 DTime1 = algoTime[Q][1].tableTime + (algoTime[Q][1].decode256Time * D256);
+        DTime1 += DTime1 >> 3;  /* advantage to algorithm using less memory, to reduce cache eviction */
+        return DTime1 < DTime0;
+    }
+#endif
+}
+
+
+size_t HUF_decompress4X_hufOnly_wksp(HUF_DTable* dctx, void* dst,
+                                     size_t dstSize, const void* cSrc,
+                                     size_t cSrcSize, void* workSpace,
+                                     size_t wkspSize)
+{
+    /* validation checks */
+    if (dstSize == 0) return ERROR(dstSize_tooSmall);
+    if (cSrcSize == 0) return ERROR(corruption_detected);
+
+    {   U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);
+#if defined(HUF_FORCE_DECOMPRESS_X1)
+        (void)algoNb;
+        assert(algoNb == 0);
+        return HUF_decompress4X1_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize);
+#elif defined(HUF_FORCE_DECOMPRESS_X2)
+        (void)algoNb;
+        assert(algoNb == 1);
+        return HUF_decompress4X2_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize);
+#else
+        return algoNb ? HUF_decompress4X2_DCtx_wksp(dctx, dst, dstSize, cSrc,
+                            cSrcSize, workSpace, wkspSize):
+                        HUF_decompress4X1_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize);
+#endif
+    }
+}
+
+size_t HUF_decompress1X_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,
+                                  const void* cSrc, size_t cSrcSize,
+                                  void* workSpace, size_t wkspSize)
+{
+    /* validation checks */
+    if (dstSize == 0) return ERROR(dstSize_tooSmall);
+    if (cSrcSize > dstSize) return ERROR(corruption_detected);   /* invalid */
+    if (cSrcSize == dstSize) { ZSTD_memcpy(dst, cSrc, dstSize); return dstSize; }   /* not compressed */
+    if (cSrcSize == 1) { ZSTD_memset(dst, *(const BYTE*)cSrc, dstSize); return dstSize; }   /* RLE */
+
+    {   U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);
+#if defined(HUF_FORCE_DECOMPRESS_X1)
+        (void)algoNb;
+        assert(algoNb == 0);
+        return HUF_decompress1X1_DCtx_wksp(dctx, dst, dstSize, cSrc,
+                                cSrcSize, workSpace, wkspSize);
+#elif defined(HUF_FORCE_DECOMPRESS_X2)
+        (void)algoNb;
+        assert(algoNb == 1);
+        return HUF_decompress1X2_DCtx_wksp(dctx, dst, dstSize, cSrc,
+                                cSrcSize, workSpace, wkspSize);
+#else
+        return algoNb ? HUF_decompress1X2_DCtx_wksp(dctx, dst, dstSize, cSrc,
+                                cSrcSize, workSpace, wkspSize):
+                        HUF_decompress1X1_DCtx_wksp(dctx, dst, dstSize, cSrc,
+                                cSrcSize, workSpace, wkspSize);
+#endif
+    }
+}
+
+
+size_t HUF_decompress1X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2)
+{
+    DTableDesc const dtd = HUF_getDTableDesc(DTable);
+#if defined(HUF_FORCE_DECOMPRESS_X1)
+    (void)dtd;
+    assert(dtd.tableType == 0);
+    return HUF_decompress1X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);
+#elif defined(HUF_FORCE_DECOMPRESS_X2)
+    (void)dtd;
+    assert(dtd.tableType == 1);
+    return HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);
+#else
+    return dtd.tableType ? HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2) :
+                           HUF_decompress1X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);
+#endif
+}
+
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_decompress1X1_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2)
+{
+    const BYTE* ip = (const BYTE*) cSrc;
+
+    size_t const hSize = HUF_readDTableX1_wksp_bmi2(dctx, cSrc, cSrcSize, workSpace, wkspSize, bmi2);
+    if (HUF_isError(hSize)) return hSize;
+    if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
+    ip += hSize; cSrcSize -= hSize;
+
+    return HUF_decompress1X1_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx, bmi2);
+}
+#endif
+
+size_t HUF_decompress4X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2)
+{
+    DTableDesc const dtd = HUF_getDTableDesc(DTable);
+#if defined(HUF_FORCE_DECOMPRESS_X1)
+    (void)dtd;
+    assert(dtd.tableType == 0);
+    return HUF_decompress4X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);
+#elif defined(HUF_FORCE_DECOMPRESS_X2)
+    (void)dtd;
+    assert(dtd.tableType == 1);
+    return HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);
+#else
+    return dtd.tableType ? HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2) :
+                           HUF_decompress4X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);
+#endif
+}
+
+size_t HUF_decompress4X_hufOnly_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2)
+{
+    /* validation checks */
+    if (dstSize == 0) return ERROR(dstSize_tooSmall);
+    if (cSrcSize == 0) return ERROR(corruption_detected);
+
+    {   U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);
+#if defined(HUF_FORCE_DECOMPRESS_X1)
+        (void)algoNb;
+        assert(algoNb == 0);
+        return HUF_decompress4X1_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, bmi2);
+#elif defined(HUF_FORCE_DECOMPRESS_X2)
+        (void)algoNb;
+        assert(algoNb == 1);
+        return HUF_decompress4X2_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, bmi2);
+#else
+        return algoNb ? HUF_decompress4X2_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, bmi2) :
+                        HUF_decompress4X1_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, bmi2);
+#endif
+    }
+}
+
diff --git a/lib/zstd/decompress/zstd_ddict.c b/lib/zstd/decompress/zstd_ddict.c
new file mode 100644
index 0000000..dbbc791
--- /dev/null
+++ b/lib/zstd/decompress/zstd_ddict.c
@@ -0,0 +1,241 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/* zstd_ddict.c :
+ * concentrates all logic that needs to know the internals of ZSTD_DDict object */
+
+/*-*******************************************************
+*  Dependencies
+*********************************************************/
+#include "../common/zstd_deps.h"   /* ZSTD_memcpy, ZSTD_memmove, ZSTD_memset */
+#include "../common/cpu.h"         /* bmi2 */
+#include "../common/mem.h"         /* low level memory routines */
+#define FSE_STATIC_LINKING_ONLY
+#include "../common/fse.h"
+#define HUF_STATIC_LINKING_ONLY
+#include "../common/huf.h"
+#include "zstd_decompress_internal.h"
+#include "zstd_ddict.h"
+
+
+
+
+/*-*******************************************************
+*  Types
+*********************************************************/
+struct ZSTD_DDict_s {
+    void* dictBuffer;
+    const void* dictContent;
+    size_t dictSize;
+    ZSTD_entropyDTables_t entropy;
+    U32 dictID;
+    U32 entropyPresent;
+    ZSTD_customMem cMem;
+};  /* typedef'd to ZSTD_DDict within "zstd.h" */
+
+const void* ZSTD_DDict_dictContent(const ZSTD_DDict* ddict)
+{
+    assert(ddict != NULL);
+    return ddict->dictContent;
+}
+
+size_t ZSTD_DDict_dictSize(const ZSTD_DDict* ddict)
+{
+    assert(ddict != NULL);
+    return ddict->dictSize;
+}
+
+void ZSTD_copyDDictParameters(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict)
+{
+    DEBUGLOG(4, "ZSTD_copyDDictParameters");
+    assert(dctx != NULL);
+    assert(ddict != NULL);
+    dctx->dictID = ddict->dictID;
+    dctx->prefixStart = ddict->dictContent;
+    dctx->virtualStart = ddict->dictContent;
+    dctx->dictEnd = (const BYTE*)ddict->dictContent + ddict->dictSize;
+    dctx->previousDstEnd = dctx->dictEnd;
+#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+    dctx->dictContentBeginForFuzzing = dctx->prefixStart;
+    dctx->dictContentEndForFuzzing = dctx->previousDstEnd;
+#endif
+    if (ddict->entropyPresent) {
+        dctx->litEntropy = 1;
+        dctx->fseEntropy = 1;
+        dctx->LLTptr = ddict->entropy.LLTable;
+        dctx->MLTptr = ddict->entropy.MLTable;
+        dctx->OFTptr = ddict->entropy.OFTable;
+        dctx->HUFptr = ddict->entropy.hufTable;
+        dctx->entropy.rep[0] = ddict->entropy.rep[0];
+        dctx->entropy.rep[1] = ddict->entropy.rep[1];
+        dctx->entropy.rep[2] = ddict->entropy.rep[2];
+    } else {
+        dctx->litEntropy = 0;
+        dctx->fseEntropy = 0;
+    }
+}
+
+
+static size_t
+ZSTD_loadEntropy_intoDDict(ZSTD_DDict* ddict,
+                           ZSTD_dictContentType_e dictContentType)
+{
+    ddict->dictID = 0;
+    ddict->entropyPresent = 0;
+    if (dictContentType == ZSTD_dct_rawContent) return 0;
+
+    if (ddict->dictSize < 8) {
+        if (dictContentType == ZSTD_dct_fullDict)
+            return ERROR(dictionary_corrupted);   /* only accept specified dictionaries */
+        return 0;   /* pure content mode */
+    }
+    {   U32 const magic = MEM_readLE32(ddict->dictContent);
+        if (magic != ZSTD_MAGIC_DICTIONARY) {
+            if (dictContentType == ZSTD_dct_fullDict)
+                return ERROR(dictionary_corrupted);   /* only accept specified dictionaries */
+            return 0;   /* pure content mode */
+        }
+    }
+    ddict->dictID = MEM_readLE32((const char*)ddict->dictContent + ZSTD_FRAMEIDSIZE);
+
+    /* load entropy tables */
+    RETURN_ERROR_IF(ZSTD_isError(ZSTD_loadDEntropy(
+            &ddict->entropy, ddict->dictContent, ddict->dictSize)),
+        dictionary_corrupted, "");
+    ddict->entropyPresent = 1;
+    return 0;
+}
+
+
+static size_t ZSTD_initDDict_internal(ZSTD_DDict* ddict,
+                                      const void* dict, size_t dictSize,
+                                      ZSTD_dictLoadMethod_e dictLoadMethod,
+                                      ZSTD_dictContentType_e dictContentType)
+{
+    if ((dictLoadMethod == ZSTD_dlm_byRef) || (!dict) || (!dictSize)) {
+        ddict->dictBuffer = NULL;
+        ddict->dictContent = dict;
+        if (!dict) dictSize = 0;
+    } else {
+        void* const internalBuffer = ZSTD_customMalloc(dictSize, ddict->cMem);
+        ddict->dictBuffer = internalBuffer;
+        ddict->dictContent = internalBuffer;
+        if (!internalBuffer) return ERROR(memory_allocation);
+        ZSTD_memcpy(internalBuffer, dict, dictSize);
+    }
+    ddict->dictSize = dictSize;
+    ddict->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001);  /* cover both little and big endian */
+
+    /* parse dictionary content */
+    FORWARD_IF_ERROR( ZSTD_loadEntropy_intoDDict(ddict, dictContentType) , "");
+
+    return 0;
+}
+
+ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize,
+                                      ZSTD_dictLoadMethod_e dictLoadMethod,
+                                      ZSTD_dictContentType_e dictContentType,
+                                      ZSTD_customMem customMem)
+{
+    if ((!customMem.customAlloc) ^ (!customMem.customFree)) return NULL;
+
+    {   ZSTD_DDict* const ddict = (ZSTD_DDict*) ZSTD_customMalloc(sizeof(ZSTD_DDict), customMem);
+        if (ddict == NULL) return NULL;
+        ddict->cMem = customMem;
+        {   size_t const initResult = ZSTD_initDDict_internal(ddict,
+                                            dict, dictSize,
+                                            dictLoadMethod, dictContentType);
+            if (ZSTD_isError(initResult)) {
+                ZSTD_freeDDict(ddict);
+                return NULL;
+        }   }
+        return ddict;
+    }
+}
+
+/*! ZSTD_createDDict() :
+*   Create a digested dictionary, to start decompression without startup delay.
+*   `dict` content is copied inside DDict.
+*   Consequently, `dict` can be released after `ZSTD_DDict` creation */
+ZSTD_DDict* ZSTD_createDDict(const void* dict, size_t dictSize)
+{
+    ZSTD_customMem const allocator = { NULL, NULL, NULL };
+    return ZSTD_createDDict_advanced(dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dct_auto, allocator);
+}
+
+/*! ZSTD_createDDict_byReference() :
+ *  Create a digested dictionary, to start decompression without startup delay.
+ *  Dictionary content is simply referenced, it will be accessed during decompression.
+ *  Warning : dictBuffer must outlive DDict (DDict must be freed before dictBuffer) */
+ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize)
+{
+    ZSTD_customMem const allocator = { NULL, NULL, NULL };
+    return ZSTD_createDDict_advanced(dictBuffer, dictSize, ZSTD_dlm_byRef, ZSTD_dct_auto, allocator);
+}
+
+
+const ZSTD_DDict* ZSTD_initStaticDDict(
+                                void* sBuffer, size_t sBufferSize,
+                                const void* dict, size_t dictSize,
+                                ZSTD_dictLoadMethod_e dictLoadMethod,
+                                ZSTD_dictContentType_e dictContentType)
+{
+    size_t const neededSpace = sizeof(ZSTD_DDict)
+                             + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize);
+    ZSTD_DDict* const ddict = (ZSTD_DDict*)sBuffer;
+    assert(sBuffer != NULL);
+    assert(dict != NULL);
+    if ((size_t)sBuffer & 7) return NULL;   /* 8-aligned */
+    if (sBufferSize < neededSpace) return NULL;
+    if (dictLoadMethod == ZSTD_dlm_byCopy) {
+        ZSTD_memcpy(ddict+1, dict, dictSize);  /* local copy */
+        dict = ddict+1;
+    }
+    if (ZSTD_isError( ZSTD_initDDict_internal(ddict,
+                                              dict, dictSize,
+                                              ZSTD_dlm_byRef, dictContentType) ))
+        return NULL;
+    return ddict;
+}
+
+
+size_t ZSTD_freeDDict(ZSTD_DDict* ddict)
+{
+    if (ddict==NULL) return 0;   /* support free on NULL */
+    {   ZSTD_customMem const cMem = ddict->cMem;
+        ZSTD_customFree(ddict->dictBuffer, cMem);
+        ZSTD_customFree(ddict, cMem);
+        return 0;
+    }
+}
+
+/*! ZSTD_estimateDDictSize() :
+ *  Estimate amount of memory that will be needed to create a dictionary for decompression.
+ *  Note : dictionary created by reference using ZSTD_dlm_byRef are smaller */
+size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod)
+{
+    return sizeof(ZSTD_DDict) + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize);
+}
+
+size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict)
+{
+    if (ddict==NULL) return 0;   /* support sizeof on NULL */
+    return sizeof(*ddict) + (ddict->dictBuffer ? ddict->dictSize : 0) ;
+}
+
+/*! ZSTD_getDictID_fromDDict() :
+ *  Provides the dictID of the dictionary loaded into `ddict`.
+ *  If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
+ *  Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */
+unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict)
+{
+    if (ddict==NULL) return 0;
+    return ZSTD_getDictID_fromDict(ddict->dictContent, ddict->dictSize);
+}
diff --git a/lib/zstd/decompress/zstd_ddict.h b/lib/zstd/decompress/zstd_ddict.h
new file mode 100644
index 0000000..8c1a79d
--- /dev/null
+++ b/lib/zstd/decompress/zstd_ddict.h
@@ -0,0 +1,44 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+#ifndef ZSTD_DDICT_H
+#define ZSTD_DDICT_H
+
+/*-*******************************************************
+ *  Dependencies
+ *********************************************************/
+#include "../common/zstd_deps.h"   /* size_t */
+#include <linux/zstd.h>     /* ZSTD_DDict, and several public functions */
+
+
+/*-*******************************************************
+ *  Interface
+ *********************************************************/
+
+/* note: several prototypes are already published in `zstd.h` :
+ * ZSTD_createDDict()
+ * ZSTD_createDDict_byReference()
+ * ZSTD_createDDict_advanced()
+ * ZSTD_freeDDict()
+ * ZSTD_initStaticDDict()
+ * ZSTD_sizeof_DDict()
+ * ZSTD_estimateDDictSize()
+ * ZSTD_getDictID_fromDict()
+ */
+
+const void* ZSTD_DDict_dictContent(const ZSTD_DDict* ddict);
+size_t ZSTD_DDict_dictSize(const ZSTD_DDict* ddict);
+
+void ZSTD_copyDDictParameters(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);
+
+
+
+#endif /* ZSTD_DDICT_H */
diff --git a/lib/zstd/decompress/zstd_decompress.c b/lib/zstd/decompress/zstd_decompress.c
new file mode 100644
index 0000000..b4d81d8
--- /dev/null
+++ b/lib/zstd/decompress/zstd_decompress.c
@@ -0,0 +1,2085 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+/* ***************************************************************
+*  Tuning parameters
+*****************************************************************/
+/*!
+ * HEAPMODE :
+ * Select how default decompression function ZSTD_decompress() allocates its context,
+ * on stack (0), or into heap (1, default; requires malloc()).
+ * Note that functions with explicit context such as ZSTD_decompressDCtx() are unaffected.
+ */
+#ifndef ZSTD_HEAPMODE
+#  define ZSTD_HEAPMODE 1
+#endif
+
+/*!
+*  LEGACY_SUPPORT :
+*  if set to 1+, ZSTD_decompress() can decode older formats (v0.1+)
+*/
+
+/*!
+ *  MAXWINDOWSIZE_DEFAULT :
+ *  maximum window size accepted by DStream __by default__.
+ *  Frames requiring more memory will be rejected.
+ *  It's possible to set a different limit using ZSTD_DCtx_setMaxWindowSize().
+ */
+#ifndef ZSTD_MAXWINDOWSIZE_DEFAULT
+#  define ZSTD_MAXWINDOWSIZE_DEFAULT (((U32)1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT) + 1)
+#endif
+
+/*!
+ *  NO_FORWARD_PROGRESS_MAX :
+ *  maximum allowed nb of calls to ZSTD_decompressStream()
+ *  without any forward progress
+ *  (defined as: no byte read from input, and no byte flushed to output)
+ *  before triggering an error.
+ */
+#ifndef ZSTD_NO_FORWARD_PROGRESS_MAX
+#  define ZSTD_NO_FORWARD_PROGRESS_MAX 16
+#endif
+
+
+/*-*******************************************************
+*  Dependencies
+*********************************************************/
+#include "../common/zstd_deps.h"   /* ZSTD_memcpy, ZSTD_memmove, ZSTD_memset */
+#include "../common/cpu.h"         /* bmi2 */
+#include "../common/mem.h"         /* low level memory routines */
+#define FSE_STATIC_LINKING_ONLY
+#include "../common/fse.h"
+#define HUF_STATIC_LINKING_ONLY
+#include "../common/huf.h"
+#include <linux/xxhash.h> /* xxh64_reset, xxh64_update, xxh64_digest, XXH64 */
+#include "../common/zstd_internal.h"  /* blockProperties_t */
+#include "zstd_decompress_internal.h"   /* ZSTD_DCtx */
+#include "zstd_ddict.h"  /* ZSTD_DDictDictContent */
+#include "zstd_decompress_block.h"   /* ZSTD_decompressBlock_internal */
+
+
+
+
+/* ***********************************
+ * Multiple DDicts Hashset internals *
+ *************************************/
+
+#define DDICT_HASHSET_MAX_LOAD_FACTOR_COUNT_MULT 4
+#define DDICT_HASHSET_MAX_LOAD_FACTOR_SIZE_MULT 3   /* These two constants represent SIZE_MULT/COUNT_MULT load factor without using a float.
+                                                     * Currently, that means a 0.75 load factor.
+                                                     * So, if count * COUNT_MULT / size * SIZE_MULT != 0, then we've exceeded
+                                                     * the load factor of the ddict hash set.
+                                                     */
+
+#define DDICT_HASHSET_TABLE_BASE_SIZE 64
+#define DDICT_HASHSET_RESIZE_FACTOR 2
+
+/* Hash function to determine starting position of dict insertion within the table
+ * Returns an index between [0, hashSet->ddictPtrTableSize]
+ */
+static size_t ZSTD_DDictHashSet_getIndex(const ZSTD_DDictHashSet* hashSet, U32 dictID) {
+    const U64 hash = xxh64(&dictID, sizeof(U32), 0);
+    /* DDict ptr table size is a multiple of 2, use size - 1 as mask to get index within [0, hashSet->ddictPtrTableSize) */
+    return hash & (hashSet->ddictPtrTableSize - 1);
+}
+
+/* Adds DDict to a hashset without resizing it.
+ * If inserting a DDict with a dictID that already exists in the set, replaces the one in the set.
+ * Returns 0 if successful, or a zstd error code if something went wrong.
+ */
+static size_t ZSTD_DDictHashSet_emplaceDDict(ZSTD_DDictHashSet* hashSet, const ZSTD_DDict* ddict) {
+    const U32 dictID = ZSTD_getDictID_fromDDict(ddict);
+    size_t idx = ZSTD_DDictHashSet_getIndex(hashSet, dictID);
+    const size_t idxRangeMask = hashSet->ddictPtrTableSize - 1;
+    RETURN_ERROR_IF(hashSet->ddictPtrCount == hashSet->ddictPtrTableSize, GENERIC, "Hash set is full!");
+    DEBUGLOG(4, "Hashed index: for dictID: %u is %zu", dictID, idx);
+    while (hashSet->ddictPtrTable[idx] != NULL) {
+        /* Replace existing ddict if inserting ddict with same dictID */
+        if (ZSTD_getDictID_fromDDict(hashSet->ddictPtrTable[idx]) == dictID) {
+            DEBUGLOG(4, "DictID already exists, replacing rather than adding");
+            hashSet->ddictPtrTable[idx] = ddict;
+            return 0;
+        }
+        idx &= idxRangeMask;
+        idx++;
+    }
+    DEBUGLOG(4, "Final idx after probing for dictID %u is: %zu", dictID, idx);
+    hashSet->ddictPtrTable[idx] = ddict;
+    hashSet->ddictPtrCount++;
+    return 0;
+}
+
+/* Expands hash table by factor of DDICT_HASHSET_RESIZE_FACTOR and
+ * rehashes all values, allocates new table, frees old table.
+ * Returns 0 on success, otherwise a zstd error code.
+ */
+static size_t ZSTD_DDictHashSet_expand(ZSTD_DDictHashSet* hashSet, ZSTD_customMem customMem) {
+    size_t newTableSize = hashSet->ddictPtrTableSize * DDICT_HASHSET_RESIZE_FACTOR;
+    const ZSTD_DDict** newTable = (const ZSTD_DDict**)ZSTD_customCalloc(sizeof(ZSTD_DDict*) * newTableSize, customMem);
+    const ZSTD_DDict** oldTable = hashSet->ddictPtrTable;
+    size_t oldTableSize = hashSet->ddictPtrTableSize;
+    size_t i;
+
+    DEBUGLOG(4, "Expanding DDict hash table! Old size: %zu new size: %zu", oldTableSize, newTableSize);
+    RETURN_ERROR_IF(!newTable, memory_allocation, "Expanded hashset allocation failed!");
+    hashSet->ddictPtrTable = newTable;
+    hashSet->ddictPtrTableSize = newTableSize;
+    hashSet->ddictPtrCount = 0;
+    for (i = 0; i < oldTableSize; ++i) {
+        if (oldTable[i] != NULL) {
+            FORWARD_IF_ERROR(ZSTD_DDictHashSet_emplaceDDict(hashSet, oldTable[i]), "");
+        }
+    }
+    ZSTD_customFree((void*)oldTable, customMem);
+    DEBUGLOG(4, "Finished re-hash");
+    return 0;
+}
+
+/* Fetches a DDict with the given dictID
+ * Returns the ZSTD_DDict* with the requested dictID. If it doesn't exist, then returns NULL.
+ */
+static const ZSTD_DDict* ZSTD_DDictHashSet_getDDict(ZSTD_DDictHashSet* hashSet, U32 dictID) {
+    size_t idx = ZSTD_DDictHashSet_getIndex(hashSet, dictID);
+    const size_t idxRangeMask = hashSet->ddictPtrTableSize - 1;
+    DEBUGLOG(4, "Hashed index: for dictID: %u is %zu", dictID, idx);
+    for (;;) {
+        size_t currDictID = ZSTD_getDictID_fromDDict(hashSet->ddictPtrTable[idx]);
+        if (currDictID == dictID || currDictID == 0) {
+            /* currDictID == 0 implies a NULL ddict entry */
+            break;
+        } else {
+            idx &= idxRangeMask;    /* Goes to start of table when we reach the end */
+            idx++;
+        }
+    }
+    DEBUGLOG(4, "Final idx after probing for dictID %u is: %zu", dictID, idx);
+    return hashSet->ddictPtrTable[idx];
+}
+
+/* Allocates space for and returns a ddict hash set
+ * The hash set's ZSTD_DDict* table has all values automatically set to NULL to begin with.
+ * Returns NULL if allocation failed.
+ */
+static ZSTD_DDictHashSet* ZSTD_createDDictHashSet(ZSTD_customMem customMem) {
+    ZSTD_DDictHashSet* ret = (ZSTD_DDictHashSet*)ZSTD_customMalloc(sizeof(ZSTD_DDictHashSet), customMem);
+    DEBUGLOG(4, "Allocating new hash set");
+    if (!ret)
+        return NULL;
+    ret->ddictPtrTable = (const ZSTD_DDict**)ZSTD_customCalloc(DDICT_HASHSET_TABLE_BASE_SIZE * sizeof(ZSTD_DDict*), customMem);
+    if (!ret->ddictPtrTable) {
+        ZSTD_customFree(ret, customMem);
+        return NULL;
+    }
+    ret->ddictPtrTableSize = DDICT_HASHSET_TABLE_BASE_SIZE;
+    ret->ddictPtrCount = 0;
+    return ret;
+}
+
+/* Frees the table of ZSTD_DDict* within a hashset, then frees the hashset itself.
+ * Note: The ZSTD_DDict* within the table are NOT freed.
+ */
+static void ZSTD_freeDDictHashSet(ZSTD_DDictHashSet* hashSet, ZSTD_customMem customMem) {
+    DEBUGLOG(4, "Freeing ddict hash set");
+    if (hashSet && hashSet->ddictPtrTable) {
+        ZSTD_customFree((void*)hashSet->ddictPtrTable, customMem);
+    }
+    if (hashSet) {
+        ZSTD_customFree(hashSet, customMem);
+    }
+}
+
+/* Public function: Adds a DDict into the ZSTD_DDictHashSet, possibly triggering a resize of the hash set.
+ * Returns 0 on success, or a ZSTD error.
+ */
+static size_t ZSTD_DDictHashSet_addDDict(ZSTD_DDictHashSet* hashSet, const ZSTD_DDict* ddict, ZSTD_customMem customMem) {
+    DEBUGLOG(4, "Adding dict ID: %u to hashset with - Count: %zu Tablesize: %zu", ZSTD_getDictID_fromDDict(ddict), hashSet->ddictPtrCount, hashSet->ddictPtrTableSize);
+    if (hashSet->ddictPtrCount * DDICT_HASHSET_MAX_LOAD_FACTOR_COUNT_MULT / hashSet->ddictPtrTableSize * DDICT_HASHSET_MAX_LOAD_FACTOR_SIZE_MULT != 0) {
+        FORWARD_IF_ERROR(ZSTD_DDictHashSet_expand(hashSet, customMem), "");
+    }
+    FORWARD_IF_ERROR(ZSTD_DDictHashSet_emplaceDDict(hashSet, ddict), "");
+    return 0;
+}
+
+/*-*************************************************************
+*   Context management
+***************************************************************/
+size_t ZSTD_sizeof_DCtx (const ZSTD_DCtx* dctx)
+{
+    if (dctx==NULL) return 0;   /* support sizeof NULL */
+    return sizeof(*dctx)
+           + ZSTD_sizeof_DDict(dctx->ddictLocal)
+           + dctx->inBuffSize + dctx->outBuffSize;
+}
+
+size_t ZSTD_estimateDCtxSize(void) { return sizeof(ZSTD_DCtx); }
+
+
+static size_t ZSTD_startingInputLength(ZSTD_format_e format)
+{
+    size_t const startingInputLength = ZSTD_FRAMEHEADERSIZE_PREFIX(format);
+    /* only supports formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless */
+    assert( (format == ZSTD_f_zstd1) || (format == ZSTD_f_zstd1_magicless) );
+    return startingInputLength;
+}
+
+static void ZSTD_DCtx_resetParameters(ZSTD_DCtx* dctx)
+{
+    assert(dctx->streamStage == zdss_init);
+    dctx->format = ZSTD_f_zstd1;
+    dctx->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;
+    dctx->outBufferMode = ZSTD_bm_buffered;
+    dctx->forceIgnoreChecksum = ZSTD_d_validateChecksum;
+    dctx->refMultipleDDicts = ZSTD_rmd_refSingleDDict;
+}
+
+static void ZSTD_initDCtx_internal(ZSTD_DCtx* dctx)
+{
+    dctx->staticSize  = 0;
+    dctx->ddict       = NULL;
+    dctx->ddictLocal  = NULL;
+    dctx->dictEnd     = NULL;
+    dctx->ddictIsCold = 0;
+    dctx->dictUses = ZSTD_dont_use;
+    dctx->inBuff      = NULL;
+    dctx->inBuffSize  = 0;
+    dctx->outBuffSize = 0;
+    dctx->streamStage = zdss_init;
+    dctx->legacyContext = NULL;
+    dctx->previousLegacyVersion = 0;
+    dctx->noForwardProgress = 0;
+    dctx->oversizedDuration = 0;
+    dctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid());
+    dctx->ddictSet = NULL;
+    ZSTD_DCtx_resetParameters(dctx);
+#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+    dctx->dictContentEndForFuzzing = NULL;
+#endif
+}
+
+ZSTD_DCtx* ZSTD_initStaticDCtx(void *workspace, size_t workspaceSize)
+{
+    ZSTD_DCtx* const dctx = (ZSTD_DCtx*) workspace;
+
+    if ((size_t)workspace & 7) return NULL;  /* 8-aligned */
+    if (workspaceSize < sizeof(ZSTD_DCtx)) return NULL;  /* minimum size */
+
+    ZSTD_initDCtx_internal(dctx);
+    dctx->staticSize = workspaceSize;
+    dctx->inBuff = (char*)(dctx+1);
+    return dctx;
+}
+
+ZSTD_DCtx* ZSTD_createDCtx_advanced(ZSTD_customMem customMem)
+{
+    if ((!customMem.customAlloc) ^ (!customMem.customFree)) return NULL;
+
+    {   ZSTD_DCtx* const dctx = (ZSTD_DCtx*)ZSTD_customMalloc(sizeof(*dctx), customMem);
+        if (!dctx) return NULL;
+        dctx->customMem = customMem;
+        ZSTD_initDCtx_internal(dctx);
+        return dctx;
+    }
+}
+
+ZSTD_DCtx* ZSTD_createDCtx(void)
+{
+    DEBUGLOG(3, "ZSTD_createDCtx");
+    return ZSTD_createDCtx_advanced(ZSTD_defaultCMem);
+}
+
+static void ZSTD_clearDict(ZSTD_DCtx* dctx)
+{
+    ZSTD_freeDDict(dctx->ddictLocal);
+    dctx->ddictLocal = NULL;
+    dctx->ddict = NULL;
+    dctx->dictUses = ZSTD_dont_use;
+}
+
+size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx)
+{
+    if (dctx==NULL) return 0;   /* support free on NULL */
+    RETURN_ERROR_IF(dctx->staticSize, memory_allocation, "not compatible with static DCtx");
+    {   ZSTD_customMem const cMem = dctx->customMem;
+        ZSTD_clearDict(dctx);
+        ZSTD_customFree(dctx->inBuff, cMem);
+        dctx->inBuff = NULL;
+        if (dctx->ddictSet) {
+            ZSTD_freeDDictHashSet(dctx->ddictSet, cMem);
+            dctx->ddictSet = NULL;
+        }
+        ZSTD_customFree(dctx, cMem);
+        return 0;
+    }
+}
+
+/* no longer useful */
+void ZSTD_copyDCtx(ZSTD_DCtx* dstDCtx, const ZSTD_DCtx* srcDCtx)
+{
+    size_t const toCopy = (size_t)((char*)(&dstDCtx->inBuff) - (char*)dstDCtx);
+    ZSTD_memcpy(dstDCtx, srcDCtx, toCopy);  /* no need to copy workspace */
+}
+
+/* Given a dctx with a digested frame params, re-selects the correct ZSTD_DDict based on
+ * the requested dict ID from the frame. If there exists a reference to the correct ZSTD_DDict, then
+ * accordingly sets the ddict to be used to decompress the frame.
+ *
+ * If no DDict is found, then no action is taken, and the ZSTD_DCtx::ddict remains as-is.
+ *
+ * ZSTD_d_refMultipleDDicts must be enabled for this function to be called.
+ */
+static void ZSTD_DCtx_selectFrameDDict(ZSTD_DCtx* dctx) {
+    assert(dctx->refMultipleDDicts && dctx->ddictSet);
+    DEBUGLOG(4, "Adjusting DDict based on requested dict ID from frame");
+    if (dctx->ddict) {
+        const ZSTD_DDict* frameDDict = ZSTD_DDictHashSet_getDDict(dctx->ddictSet, dctx->fParams.dictID);
+        if (frameDDict) {
+            DEBUGLOG(4, "DDict found!");
+            ZSTD_clearDict(dctx);
+            dctx->dictID = dctx->fParams.dictID;
+            dctx->ddict = frameDDict;
+            dctx->dictUses = ZSTD_use_indefinitely;
+        }
+    }
+}
+
+
+/*-*************************************************************
+ *   Frame header decoding
+ ***************************************************************/
+
+/*! ZSTD_isFrame() :
+ *  Tells if the content of `buffer` starts with a valid Frame Identifier.
+ *  Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0.
+ *  Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled.
+ *  Note 3 : Skippable Frame Identifiers are considered valid. */
+unsigned ZSTD_isFrame(const void* buffer, size_t size)
+{
+    if (size < ZSTD_FRAMEIDSIZE) return 0;
+    {   U32 const magic = MEM_readLE32(buffer);
+        if (magic == ZSTD_MAGICNUMBER) return 1;
+        if ((magic & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) return 1;
+    }
+    return 0;
+}
+
+/* ZSTD_frameHeaderSize_internal() :
+ *  srcSize must be large enough to reach header size fields.
+ *  note : only works for formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless.
+ * @return : size of the Frame Header
+ *           or an error code, which can be tested with ZSTD_isError() */
+static size_t ZSTD_frameHeaderSize_internal(const void* src, size_t srcSize, ZSTD_format_e format)
+{
+    size_t const minInputSize = ZSTD_startingInputLength(format);
+    RETURN_ERROR_IF(srcSize < minInputSize, srcSize_wrong, "");
+
+    {   BYTE const fhd = ((const BYTE*)src)[minInputSize-1];
+        U32 const dictID= fhd & 3;
+        U32 const singleSegment = (fhd >> 5) & 1;
+        U32 const fcsId = fhd >> 6;
+        return minInputSize + !singleSegment
+             + ZSTD_did_fieldSize[dictID] + ZSTD_fcs_fieldSize[fcsId]
+             + (singleSegment && !fcsId);
+    }
+}
+
+/* ZSTD_frameHeaderSize() :
+ *  srcSize must be >= ZSTD_frameHeaderSize_prefix.
+ * @return : size of the Frame Header,
+ *           or an error code (if srcSize is too small) */
+size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize)
+{
+    return ZSTD_frameHeaderSize_internal(src, srcSize, ZSTD_f_zstd1);
+}
+
+
+/* ZSTD_getFrameHeader_advanced() :
+ *  decode Frame Header, or require larger `srcSize`.
+ *  note : only works for formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless
+ * @return : 0, `zfhPtr` is correctly filled,
+ *          >0, `srcSize` is too small, value is wanted `srcSize` amount,
+ *           or an error code, which can be tested using ZSTD_isError() */
+size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format)
+{
+    const BYTE* ip = (const BYTE*)src;
+    size_t const minInputSize = ZSTD_startingInputLength(format);
+
+    ZSTD_memset(zfhPtr, 0, sizeof(*zfhPtr));   /* not strictly necessary, but static analyzer do not understand that zfhPtr is only going to be read only if return value is zero, since they are 2 different signals */
+    if (srcSize < minInputSize) return minInputSize;
+    RETURN_ERROR_IF(src==NULL, GENERIC, "invalid parameter");
+
+    if ( (format != ZSTD_f_zstd1_magicless)
+      && (MEM_readLE32(src) != ZSTD_MAGICNUMBER) ) {
+        if ((MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
+            /* skippable frame */
+            if (srcSize < ZSTD_SKIPPABLEHEADERSIZE)
+                return ZSTD_SKIPPABLEHEADERSIZE; /* magic number + frame length */
+            ZSTD_memset(zfhPtr, 0, sizeof(*zfhPtr));
+            zfhPtr->frameContentSize = MEM_readLE32((const char *)src + ZSTD_FRAMEIDSIZE);
+            zfhPtr->frameType = ZSTD_skippableFrame;
+            return 0;
+        }
+        RETURN_ERROR(prefix_unknown, "");
+    }
+
+    /* ensure there is enough `srcSize` to fully read/decode frame header */
+    {   size_t const fhsize = ZSTD_frameHeaderSize_internal(src, srcSize, format);
+        if (srcSize < fhsize) return fhsize;
+        zfhPtr->headerSize = (U32)fhsize;
+    }
+
+    {   BYTE const fhdByte = ip[minInputSize-1];
+        size_t pos = minInputSize;
+        U32 const dictIDSizeCode = fhdByte&3;
+        U32 const checksumFlag = (fhdByte>>2)&1;
+        U32 const singleSegment = (fhdByte>>5)&1;
+        U32 const fcsID = fhdByte>>6;
+        U64 windowSize = 0;
+        U32 dictID = 0;
+        U64 frameContentSize = ZSTD_CONTENTSIZE_UNKNOWN;
+        RETURN_ERROR_IF((fhdByte & 0x08) != 0, frameParameter_unsupported,
+                        "reserved bits, must be zero");
+
+        if (!singleSegment) {
+            BYTE const wlByte = ip[pos++];
+            U32 const windowLog = (wlByte >> 3) + ZSTD_WINDOWLOG_ABSOLUTEMIN;
+            RETURN_ERROR_IF(windowLog > ZSTD_WINDOWLOG_MAX, frameParameter_windowTooLarge, "");
+            windowSize = (1ULL << windowLog);
+            windowSize += (windowSize >> 3) * (wlByte&7);
+        }
+        switch(dictIDSizeCode)
+        {
+            default:
+                assert(0);  /* impossible */
+                ZSTD_FALLTHROUGH;
+            case 0 : break;
+            case 1 : dictID = ip[pos]; pos++; break;
+            case 2 : dictID = MEM_readLE16(ip+pos); pos+=2; break;
+            case 3 : dictID = MEM_readLE32(ip+pos); pos+=4; break;
+        }
+        switch(fcsID)
+        {
+            default:
+                assert(0);  /* impossible */
+                ZSTD_FALLTHROUGH;
+            case 0 : if (singleSegment) frameContentSize = ip[pos]; break;
+            case 1 : frameContentSize = MEM_readLE16(ip+pos)+256; break;
+            case 2 : frameContentSize = MEM_readLE32(ip+pos); break;
+            case 3 : frameContentSize = MEM_readLE64(ip+pos); break;
+        }
+        if (singleSegment) windowSize = frameContentSize;
+
+        zfhPtr->frameType = ZSTD_frame;
+        zfhPtr->frameContentSize = frameContentSize;
+        zfhPtr->windowSize = windowSize;
+        zfhPtr->blockSizeMax = (unsigned) MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
+        zfhPtr->dictID = dictID;
+        zfhPtr->checksumFlag = checksumFlag;
+    }
+    return 0;
+}
+
+/* ZSTD_getFrameHeader() :
+ *  decode Frame Header, or require larger `srcSize`.
+ *  note : this function does not consume input, it only reads it.
+ * @return : 0, `zfhPtr` is correctly filled,
+ *          >0, `srcSize` is too small, value is wanted `srcSize` amount,
+ *           or an error code, which can be tested using ZSTD_isError() */
+size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize)
+{
+    return ZSTD_getFrameHeader_advanced(zfhPtr, src, srcSize, ZSTD_f_zstd1);
+}
+
+
+/* ZSTD_getFrameContentSize() :
+ *  compatible with legacy mode
+ * @return : decompressed size of the single frame pointed to be `src` if known, otherwise
+ *         - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined
+ *         - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small) */
+unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize)
+{
+    {   ZSTD_frameHeader zfh;
+        if (ZSTD_getFrameHeader(&zfh, src, srcSize) != 0)
+            return ZSTD_CONTENTSIZE_ERROR;
+        if (zfh.frameType == ZSTD_skippableFrame) {
+            return 0;
+        } else {
+            return zfh.frameContentSize;
+    }   }
+}
+
+static size_t readSkippableFrameSize(void const* src, size_t srcSize)
+{
+    size_t const skippableHeaderSize = ZSTD_SKIPPABLEHEADERSIZE;
+    U32 sizeU32;
+
+    RETURN_ERROR_IF(srcSize < ZSTD_SKIPPABLEHEADERSIZE, srcSize_wrong, "");
+
+    sizeU32 = MEM_readLE32((BYTE const*)src + ZSTD_FRAMEIDSIZE);
+    RETURN_ERROR_IF((U32)(sizeU32 + ZSTD_SKIPPABLEHEADERSIZE) < sizeU32,
+                    frameParameter_unsupported, "");
+    {
+        size_t const skippableSize = skippableHeaderSize + sizeU32;
+        RETURN_ERROR_IF(skippableSize > srcSize, srcSize_wrong, "");
+        return skippableSize;
+    }
+}
+
+/* ZSTD_findDecompressedSize() :
+ *  compatible with legacy mode
+ *  `srcSize` must be the exact length of some number of ZSTD compressed and/or
+ *      skippable frames
+ *  @return : decompressed size of the frames contained */
+unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize)
+{
+    unsigned long long totalDstSize = 0;
+
+    while (srcSize >= ZSTD_startingInputLength(ZSTD_f_zstd1)) {
+        U32 const magicNumber = MEM_readLE32(src);
+
+        if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
+            size_t const skippableSize = readSkippableFrameSize(src, srcSize);
+            if (ZSTD_isError(skippableSize)) {
+                return ZSTD_CONTENTSIZE_ERROR;
+            }
+            assert(skippableSize <= srcSize);
+
+            src = (const BYTE *)src + skippableSize;
+            srcSize -= skippableSize;
+            continue;
+        }
+
+        {   unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize);
+            if (ret >= ZSTD_CONTENTSIZE_ERROR) return ret;
+
+            /* check for overflow */
+            if (totalDstSize + ret < totalDstSize) return ZSTD_CONTENTSIZE_ERROR;
+            totalDstSize += ret;
+        }
+        {   size_t const frameSrcSize = ZSTD_findFrameCompressedSize(src, srcSize);
+            if (ZSTD_isError(frameSrcSize)) {
+                return ZSTD_CONTENTSIZE_ERROR;
+            }
+
+            src = (const BYTE *)src + frameSrcSize;
+            srcSize -= frameSrcSize;
+        }
+    }  /* while (srcSize >= ZSTD_frameHeaderSize_prefix) */
+
+    if (srcSize) return ZSTD_CONTENTSIZE_ERROR;
+
+    return totalDstSize;
+}
+
+/* ZSTD_getDecompressedSize() :
+ *  compatible with legacy mode
+ * @return : decompressed size if known, 0 otherwise
+             note : 0 can mean any of the following :
+                   - frame content is empty
+                   - decompressed size field is not present in frame header
+                   - frame header unknown / not supported
+                   - frame header not complete (`srcSize` too small) */
+unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize)
+{
+    unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize);
+    ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_ERROR < ZSTD_CONTENTSIZE_UNKNOWN);
+    return (ret >= ZSTD_CONTENTSIZE_ERROR) ? 0 : ret;
+}
+
+
+/* ZSTD_decodeFrameHeader() :
+ * `headerSize` must be the size provided by ZSTD_frameHeaderSize().
+ * If multiple DDict references are enabled, also will choose the correct DDict to use.
+ * @return : 0 if success, or an error code, which can be tested using ZSTD_isError() */
+static size_t ZSTD_decodeFrameHeader(ZSTD_DCtx* dctx, const void* src, size_t headerSize)
+{
+    size_t const result = ZSTD_getFrameHeader_advanced(&(dctx->fParams), src, headerSize, dctx->format);
+    if (ZSTD_isError(result)) return result;    /* invalid header */
+    RETURN_ERROR_IF(result>0, srcSize_wrong, "headerSize too small");
+
+    /* Reference DDict requested by frame if dctx references multiple ddicts */
+    if (dctx->refMultipleDDicts == ZSTD_rmd_refMultipleDDicts && dctx->ddictSet) {
+        ZSTD_DCtx_selectFrameDDict(dctx);
+    }
+
+#ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+    /* Skip the dictID check in fuzzing mode, because it makes the search
+     * harder.
+     */
+    RETURN_ERROR_IF(dctx->fParams.dictID && (dctx->dictID != dctx->fParams.dictID),
+                    dictionary_wrong, "");
+#endif
+    dctx->validateChecksum = (dctx->fParams.checksumFlag && !dctx->forceIgnoreChecksum) ? 1 : 0;
+    if (dctx->validateChecksum) xxh64_reset(&dctx->xxhState, 0);
+    dctx->processedCSize += headerSize;
+    return 0;
+}
+
+static ZSTD_frameSizeInfo ZSTD_errorFrameSizeInfo(size_t ret)
+{
+    ZSTD_frameSizeInfo frameSizeInfo;
+    frameSizeInfo.compressedSize = ret;
+    frameSizeInfo.decompressedBound = ZSTD_CONTENTSIZE_ERROR;
+    return frameSizeInfo;
+}
+
+static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize)
+{
+    ZSTD_frameSizeInfo frameSizeInfo;
+    ZSTD_memset(&frameSizeInfo, 0, sizeof(ZSTD_frameSizeInfo));
+
+
+    if ((srcSize >= ZSTD_SKIPPABLEHEADERSIZE)
+        && (MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
+        frameSizeInfo.compressedSize = readSkippableFrameSize(src, srcSize);
+        assert(ZSTD_isError(frameSizeInfo.compressedSize) ||
+               frameSizeInfo.compressedSize <= srcSize);
+        return frameSizeInfo;
+    } else {
+        const BYTE* ip = (const BYTE*)src;
+        const BYTE* const ipstart = ip;
+        size_t remainingSize = srcSize;
+        size_t nbBlocks = 0;
+        ZSTD_frameHeader zfh;
+
+        /* Extract Frame Header */
+        {   size_t const ret = ZSTD_getFrameHeader(&zfh, src, srcSize);
+            if (ZSTD_isError(ret))
+                return ZSTD_errorFrameSizeInfo(ret);
+            if (ret > 0)
+                return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));
+        }
+
+        ip += zfh.headerSize;
+        remainingSize -= zfh.headerSize;
+
+        /* Iterate over each block */
+        while (1) {
+            blockProperties_t blockProperties;
+            size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties);
+            if (ZSTD_isError(cBlockSize))
+                return ZSTD_errorFrameSizeInfo(cBlockSize);
+
+            if (ZSTD_blockHeaderSize + cBlockSize > remainingSize)
+                return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));
+
+            ip += ZSTD_blockHeaderSize + cBlockSize;
+            remainingSize -= ZSTD_blockHeaderSize + cBlockSize;
+            nbBlocks++;
+
+            if (blockProperties.lastBlock) break;
+        }
+
+        /* Final frame content checksum */
+        if (zfh.checksumFlag) {
+            if (remainingSize < 4)
+                return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));
+            ip += 4;
+        }
+
+        frameSizeInfo.compressedSize = (size_t)(ip - ipstart);
+        frameSizeInfo.decompressedBound = (zfh.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN)
+                                        ? zfh.frameContentSize
+                                        : nbBlocks * zfh.blockSizeMax;
+        return frameSizeInfo;
+    }
+}
+
+/* ZSTD_findFrameCompressedSize() :
+ *  compatible with legacy mode
+ *  `src` must point to the start of a ZSTD frame, ZSTD legacy frame, or skippable frame
+ *  `srcSize` must be at least as large as the frame contained
+ *  @return : the compressed size of the frame starting at `src` */
+size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize)
+{
+    ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);
+    return frameSizeInfo.compressedSize;
+}
+
+/* ZSTD_decompressBound() :
+ *  compatible with legacy mode
+ *  `src` must point to the start of a ZSTD frame or a skippeable frame
+ *  `srcSize` must be at least as large as the frame contained
+ *  @return : the maximum decompressed size of the compressed source
+ */
+unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize)
+{
+    unsigned long long bound = 0;
+    /* Iterate over each frame */
+    while (srcSize > 0) {
+        ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);
+        size_t const compressedSize = frameSizeInfo.compressedSize;
+        unsigned long long const decompressedBound = frameSizeInfo.decompressedBound;
+        if (ZSTD_isError(compressedSize) || decompressedBound == ZSTD_CONTENTSIZE_ERROR)
+            return ZSTD_CONTENTSIZE_ERROR;
+        assert(srcSize >= compressedSize);
+        src = (const BYTE*)src + compressedSize;
+        srcSize -= compressedSize;
+        bound += decompressedBound;
+    }
+    return bound;
+}
+
+
+/*-*************************************************************
+ *   Frame decoding
+ ***************************************************************/
+
+/* ZSTD_insertBlock() :
+ *  insert `src` block into `dctx` history. Useful to track uncompressed blocks. */
+size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize)
+{
+    DEBUGLOG(5, "ZSTD_insertBlock: %u bytes", (unsigned)blockSize);
+    ZSTD_checkContinuity(dctx, blockStart, blockSize);
+    dctx->previousDstEnd = (const char*)blockStart + blockSize;
+    return blockSize;
+}
+
+
+static size_t ZSTD_copyRawBlock(void* dst, size_t dstCapacity,
+                          const void* src, size_t srcSize)
+{
+    DEBUGLOG(5, "ZSTD_copyRawBlock");
+    RETURN_ERROR_IF(srcSize > dstCapacity, dstSize_tooSmall, "");
+    if (dst == NULL) {
+        if (srcSize == 0) return 0;
+        RETURN_ERROR(dstBuffer_null, "");
+    }
+    ZSTD_memcpy(dst, src, srcSize);
+    return srcSize;
+}
+
+static size_t ZSTD_setRleBlock(void* dst, size_t dstCapacity,
+                               BYTE b,
+                               size_t regenSize)
+{
+    RETURN_ERROR_IF(regenSize > dstCapacity, dstSize_tooSmall, "");
+    if (dst == NULL) {
+        if (regenSize == 0) return 0;
+        RETURN_ERROR(dstBuffer_null, "");
+    }
+    ZSTD_memset(dst, b, regenSize);
+    return regenSize;
+}
+
+static void ZSTD_DCtx_trace_end(ZSTD_DCtx const* dctx, U64 uncompressedSize, U64 compressedSize, unsigned streaming)
+{
+    (void)dctx;
+    (void)uncompressedSize;
+    (void)compressedSize;
+    (void)streaming;
+}
+
+
+/*! ZSTD_decompressFrame() :
+ * @dctx must be properly initialized
+ *  will update *srcPtr and *srcSizePtr,
+ *  to make *srcPtr progress by one frame. */
+static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
+                                   void* dst, size_t dstCapacity,
+                             const void** srcPtr, size_t *srcSizePtr)
+{
+    const BYTE* const istart = (const BYTE*)(*srcPtr);
+    const BYTE* ip = istart;
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* const oend = dstCapacity != 0 ? ostart + dstCapacity : ostart;
+    BYTE* op = ostart;
+    size_t remainingSrcSize = *srcSizePtr;
+
+    DEBUGLOG(4, "ZSTD_decompressFrame (srcSize:%i)", (int)*srcSizePtr);
+
+    /* check */
+    RETURN_ERROR_IF(
+        remainingSrcSize < ZSTD_FRAMEHEADERSIZE_MIN(dctx->format)+ZSTD_blockHeaderSize,
+        srcSize_wrong, "");
+
+    /* Frame Header */
+    {   size_t const frameHeaderSize = ZSTD_frameHeaderSize_internal(
+                ip, ZSTD_FRAMEHEADERSIZE_PREFIX(dctx->format), dctx->format);
+        if (ZSTD_isError(frameHeaderSize)) return frameHeaderSize;
+        RETURN_ERROR_IF(remainingSrcSize < frameHeaderSize+ZSTD_blockHeaderSize,
+                        srcSize_wrong, "");
+        FORWARD_IF_ERROR( ZSTD_decodeFrameHeader(dctx, ip, frameHeaderSize) , "");
+        ip += frameHeaderSize; remainingSrcSize -= frameHeaderSize;
+    }
+
+    /* Loop on each block */
+    while (1) {
+        size_t decodedSize;
+        blockProperties_t blockProperties;
+        size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSrcSize, &blockProperties);
+        if (ZSTD_isError(cBlockSize)) return cBlockSize;
+
+        ip += ZSTD_blockHeaderSize;
+        remainingSrcSize -= ZSTD_blockHeaderSize;
+        RETURN_ERROR_IF(cBlockSize > remainingSrcSize, srcSize_wrong, "");
+
+        switch(blockProperties.blockType)
+        {
+        case bt_compressed:
+            decodedSize = ZSTD_decompressBlock_internal(dctx, op, (size_t)(oend-op), ip, cBlockSize, /* frame */ 1);
+            break;
+        case bt_raw :
+            decodedSize = ZSTD_copyRawBlock(op, (size_t)(oend-op), ip, cBlockSize);
+            break;
+        case bt_rle :
+            decodedSize = ZSTD_setRleBlock(op, (size_t)(oend-op), *ip, blockProperties.origSize);
+            break;
+        case bt_reserved :
+        default:
+            RETURN_ERROR(corruption_detected, "invalid block type");
+        }
+
+        if (ZSTD_isError(decodedSize)) return decodedSize;
+        if (dctx->validateChecksum)
+            xxh64_update(&dctx->xxhState, op, decodedSize);
+        if (decodedSize != 0)
+            op += decodedSize;
+        assert(ip != NULL);
+        ip += cBlockSize;
+        remainingSrcSize -= cBlockSize;
+        if (blockProperties.lastBlock) break;
+    }
+
+    if (dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN) {
+        RETURN_ERROR_IF((U64)(op-ostart) != dctx->fParams.frameContentSize,
+                        corruption_detected, "");
+    }
+    if (dctx->fParams.checksumFlag) { /* Frame content checksum verification */
+        RETURN_ERROR_IF(remainingSrcSize<4, checksum_wrong, "");
+        if (!dctx->forceIgnoreChecksum) {
+            U32 const checkCalc = (U32)xxh64_digest(&dctx->xxhState);
+            U32 checkRead;
+            checkRead = MEM_readLE32(ip);
+            RETURN_ERROR_IF(checkRead != checkCalc, checksum_wrong, "");
+        }
+        ip += 4;
+        remainingSrcSize -= 4;
+    }
+    ZSTD_DCtx_trace_end(dctx, (U64)(op-ostart), (U64)(ip-istart), /* streaming */ 0);
+    /* Allow caller to get size read */
+    *srcPtr = ip;
+    *srcSizePtr = remainingSrcSize;
+    return (size_t)(op-ostart);
+}
+
+static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
+                                        void* dst, size_t dstCapacity,
+                                  const void* src, size_t srcSize,
+                                  const void* dict, size_t dictSize,
+                                  const ZSTD_DDict* ddict)
+{
+    void* const dststart = dst;
+    int moreThan1Frame = 0;
+
+    DEBUGLOG(5, "ZSTD_decompressMultiFrame");
+    assert(dict==NULL || ddict==NULL);  /* either dict or ddict set, not both */
+
+    if (ddict) {
+        dict = ZSTD_DDict_dictContent(ddict);
+        dictSize = ZSTD_DDict_dictSize(ddict);
+    }
+
+    while (srcSize >= ZSTD_startingInputLength(dctx->format)) {
+
+
+        {   U32 const magicNumber = MEM_readLE32(src);
+            DEBUGLOG(4, "reading magic number %08X (expecting %08X)",
+                        (unsigned)magicNumber, ZSTD_MAGICNUMBER);
+            if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
+                size_t const skippableSize = readSkippableFrameSize(src, srcSize);
+                FORWARD_IF_ERROR(skippableSize, "readSkippableFrameSize failed");
+                assert(skippableSize <= srcSize);
+
+                src = (const BYTE *)src + skippableSize;
+                srcSize -= skippableSize;
+                continue;
+        }   }
+
+        if (ddict) {
+            /* we were called from ZSTD_decompress_usingDDict */
+            FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDDict(dctx, ddict), "");
+        } else {
+            /* this will initialize correctly with no dict if dict == NULL, so
+             * use this in all cases but ddict */
+            FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDict(dctx, dict, dictSize), "");
+        }
+        ZSTD_checkContinuity(dctx, dst, dstCapacity);
+
+        {   const size_t res = ZSTD_decompressFrame(dctx, dst, dstCapacity,
+                                                    &src, &srcSize);
+            RETURN_ERROR_IF(
+                (ZSTD_getErrorCode(res) == ZSTD_error_prefix_unknown)
+             && (moreThan1Frame==1),
+                srcSize_wrong,
+                "At least one frame successfully completed, "
+                "but following bytes are garbage: "
+                "it's more likely to be a srcSize error, "
+                "specifying more input bytes than size of frame(s). "
+                "Note: one could be unlucky, it might be a corruption error instead, "
+                "happening right at the place where we expect zstd magic bytes. "
+                "But this is _much_ less likely than a srcSize field error.");
+            if (ZSTD_isError(res)) return res;
+            assert(res <= dstCapacity);
+            if (res != 0)
+                dst = (BYTE*)dst + res;
+            dstCapacity -= res;
+        }
+        moreThan1Frame = 1;
+    }  /* while (srcSize >= ZSTD_frameHeaderSize_prefix) */
+
+    RETURN_ERROR_IF(srcSize, srcSize_wrong, "input not entirely consumed");
+
+    return (size_t)((BYTE*)dst - (BYTE*)dststart);
+}
+
+size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,
+                                 void* dst, size_t dstCapacity,
+                           const void* src, size_t srcSize,
+                           const void* dict, size_t dictSize)
+{
+    return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize, dict, dictSize, NULL);
+}
+
+
+static ZSTD_DDict const* ZSTD_getDDict(ZSTD_DCtx* dctx)
+{
+    switch (dctx->dictUses) {
+    default:
+        assert(0 /* Impossible */);
+        ZSTD_FALLTHROUGH;
+    case ZSTD_dont_use:
+        ZSTD_clearDict(dctx);
+        return NULL;
+    case ZSTD_use_indefinitely:
+        return dctx->ddict;
+    case ZSTD_use_once:
+        dctx->dictUses = ZSTD_dont_use;
+        return dctx->ddict;
+    }
+}
+
+size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+    return ZSTD_decompress_usingDDict(dctx, dst, dstCapacity, src, srcSize, ZSTD_getDDict(dctx));
+}
+
+
+size_t ZSTD_decompress(void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+#if defined(ZSTD_HEAPMODE) && (ZSTD_HEAPMODE>=1)
+    size_t regenSize;
+    ZSTD_DCtx* const dctx = ZSTD_createDCtx();
+    RETURN_ERROR_IF(dctx==NULL, memory_allocation, "NULL pointer!");
+    regenSize = ZSTD_decompressDCtx(dctx, dst, dstCapacity, src, srcSize);
+    ZSTD_freeDCtx(dctx);
+    return regenSize;
+#else   /* stack mode */
+    ZSTD_DCtx dctx;
+    ZSTD_initDCtx_internal(&dctx);
+    return ZSTD_decompressDCtx(&dctx, dst, dstCapacity, src, srcSize);
+#endif
+}
+
+
+/*-**************************************
+*   Advanced Streaming Decompression API
+*   Bufferless and synchronous
+****************************************/
+size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx) { return dctx->expected; }
+
+/*
+ * Similar to ZSTD_nextSrcSizeToDecompress(), but when when a block input can be streamed,
+ * we allow taking a partial block as the input. Currently only raw uncompressed blocks can
+ * be streamed.
+ *
+ * For blocks that can be streamed, this allows us to reduce the latency until we produce
+ * output, and avoid copying the input.
+ *
+ * @param inputSize - The total amount of input that the caller currently has.
+ */
+static size_t ZSTD_nextSrcSizeToDecompressWithInputSize(ZSTD_DCtx* dctx, size_t inputSize) {
+    if (!(dctx->stage == ZSTDds_decompressBlock || dctx->stage == ZSTDds_decompressLastBlock))
+        return dctx->expected;
+    if (dctx->bType != bt_raw)
+        return dctx->expected;
+    return MIN(MAX(inputSize, 1), dctx->expected);
+}
+
+ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx) {
+    switch(dctx->stage)
+    {
+    default:   /* should not happen */
+        assert(0);
+        ZSTD_FALLTHROUGH;
+    case ZSTDds_getFrameHeaderSize:
+        ZSTD_FALLTHROUGH;
+    case ZSTDds_decodeFrameHeader:
+        return ZSTDnit_frameHeader;
+    case ZSTDds_decodeBlockHeader:
+        return ZSTDnit_blockHeader;
+    case ZSTDds_decompressBlock:
+        return ZSTDnit_block;
+    case ZSTDds_decompressLastBlock:
+        return ZSTDnit_lastBlock;
+    case ZSTDds_checkChecksum:
+        return ZSTDnit_checksum;
+    case ZSTDds_decodeSkippableHeader:
+        ZSTD_FALLTHROUGH;
+    case ZSTDds_skipFrame:
+        return ZSTDnit_skippableFrame;
+    }
+}
+
+static int ZSTD_isSkipFrame(ZSTD_DCtx* dctx) { return dctx->stage == ZSTDds_skipFrame; }
+
+/* ZSTD_decompressContinue() :
+ *  srcSize : must be the exact nb of bytes expected (see ZSTD_nextSrcSizeToDecompress())
+ *  @return : nb of bytes generated into `dst` (necessarily <= `dstCapacity)
+ *            or an error code, which can be tested using ZSTD_isError() */
+size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+    DEBUGLOG(5, "ZSTD_decompressContinue (srcSize:%u)", (unsigned)srcSize);
+    /* Sanity check */
+    RETURN_ERROR_IF(srcSize != ZSTD_nextSrcSizeToDecompressWithInputSize(dctx, srcSize), srcSize_wrong, "not allowed");
+    ZSTD_checkContinuity(dctx, dst, dstCapacity);
+
+    dctx->processedCSize += srcSize;
+
+    switch (dctx->stage)
+    {
+    case ZSTDds_getFrameHeaderSize :
+        assert(src != NULL);
+        if (dctx->format == ZSTD_f_zstd1) {  /* allows header */
+            assert(srcSize >= ZSTD_FRAMEIDSIZE);  /* to read skippable magic number */
+            if ((MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {        /* skippable frame */
+                ZSTD_memcpy(dctx->headerBuffer, src, srcSize);
+                dctx->expected = ZSTD_SKIPPABLEHEADERSIZE - srcSize;  /* remaining to load to get full skippable frame header */
+                dctx->stage = ZSTDds_decodeSkippableHeader;
+                return 0;
+        }   }
+        dctx->headerSize = ZSTD_frameHeaderSize_internal(src, srcSize, dctx->format);
+        if (ZSTD_isError(dctx->headerSize)) return dctx->headerSize;
+        ZSTD_memcpy(dctx->headerBuffer, src, srcSize);
+        dctx->expected = dctx->headerSize - srcSize;
+        dctx->stage = ZSTDds_decodeFrameHeader;
+        return 0;
+
+    case ZSTDds_decodeFrameHeader:
+        assert(src != NULL);
+        ZSTD_memcpy(dctx->headerBuffer + (dctx->headerSize - srcSize), src, srcSize);
+        FORWARD_IF_ERROR(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize), "");
+        dctx->expected = ZSTD_blockHeaderSize;
+        dctx->stage = ZSTDds_decodeBlockHeader;
+        return 0;
+
+    case ZSTDds_decodeBlockHeader:
+        {   blockProperties_t bp;
+            size_t const cBlockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp);
+            if (ZSTD_isError(cBlockSize)) return cBlockSize;
+            RETURN_ERROR_IF(cBlockSize > dctx->fParams.blockSizeMax, corruption_detected, "Block Size Exceeds Maximum");
+            dctx->expected = cBlockSize;
+            dctx->bType = bp.blockType;
+            dctx->rleSize = bp.origSize;
+            if (cBlockSize) {
+                dctx->stage = bp.lastBlock ? ZSTDds_decompressLastBlock : ZSTDds_decompressBlock;
+                return 0;
+            }
+            /* empty block */
+            if (bp.lastBlock) {
+                if (dctx->fParams.checksumFlag) {
+                    dctx->expected = 4;
+                    dctx->stage = ZSTDds_checkChecksum;
+                } else {
+                    dctx->expected = 0; /* end of frame */
+                    dctx->stage = ZSTDds_getFrameHeaderSize;
+                }
+            } else {
+                dctx->expected = ZSTD_blockHeaderSize;  /* jump to next header */
+                dctx->stage = ZSTDds_decodeBlockHeader;
+            }
+            return 0;
+        }
+
+    case ZSTDds_decompressLastBlock:
+    case ZSTDds_decompressBlock:
+        DEBUGLOG(5, "ZSTD_decompressContinue: case ZSTDds_decompressBlock");
+        {   size_t rSize;
+            switch(dctx->bType)
+            {
+            case bt_compressed:
+                DEBUGLOG(5, "ZSTD_decompressContinue: case bt_compressed");
+                rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, /* frame */ 1);
+                dctx->expected = 0;  /* Streaming not supported */
+                break;
+            case bt_raw :
+                assert(srcSize <= dctx->expected);
+                rSize = ZSTD_copyRawBlock(dst, dstCapacity, src, srcSize);
+                FORWARD_IF_ERROR(rSize, "ZSTD_copyRawBlock failed");
+                assert(rSize == srcSize);
+                dctx->expected -= rSize;
+                break;
+            case bt_rle :
+                rSize = ZSTD_setRleBlock(dst, dstCapacity, *(const BYTE*)src, dctx->rleSize);
+                dctx->expected = 0;  /* Streaming not supported */
+                break;
+            case bt_reserved :   /* should never happen */
+            default:
+                RETURN_ERROR(corruption_detected, "invalid block type");
+            }
+            FORWARD_IF_ERROR(rSize, "");
+            RETURN_ERROR_IF(rSize > dctx->fParams.blockSizeMax, corruption_detected, "Decompressed Block Size Exceeds Maximum");
+            DEBUGLOG(5, "ZSTD_decompressContinue: decoded size from block : %u", (unsigned)rSize);
+            dctx->decodedSize += rSize;
+            if (dctx->validateChecksum) xxh64_update(&dctx->xxhState, dst, rSize);
+            dctx->previousDstEnd = (char*)dst + rSize;
+
+            /* Stay on the same stage until we are finished streaming the block. */
+            if (dctx->expected > 0) {
+                return rSize;
+            }
+
+            if (dctx->stage == ZSTDds_decompressLastBlock) {   /* end of frame */
+                DEBUGLOG(4, "ZSTD_decompressContinue: decoded size from frame : %u", (unsigned)dctx->decodedSize);
+                RETURN_ERROR_IF(
+                    dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN
+                 && dctx->decodedSize != dctx->fParams.frameContentSize,
+                    corruption_detected, "");
+                if (dctx->fParams.checksumFlag) {  /* another round for frame checksum */
+                    dctx->expected = 4;
+                    dctx->stage = ZSTDds_checkChecksum;
+                } else {
+                    ZSTD_DCtx_trace_end(dctx, dctx->decodedSize, dctx->processedCSize, /* streaming */ 1);
+                    dctx->expected = 0;   /* ends here */
+                    dctx->stage = ZSTDds_getFrameHeaderSize;
+                }
+            } else {
+                dctx->stage = ZSTDds_decodeBlockHeader;
+                dctx->expected = ZSTD_blockHeaderSize;
+            }
+            return rSize;
+        }
+
+    case ZSTDds_checkChecksum:
+        assert(srcSize == 4);  /* guaranteed by dctx->expected */
+        {
+            if (dctx->validateChecksum) {
+                U32 const h32 = (U32)xxh64_digest(&dctx->xxhState);
+                U32 const check32 = MEM_readLE32(src);
+                DEBUGLOG(4, "ZSTD_decompressContinue: checksum : calculated %08X :: %08X read", (unsigned)h32, (unsigned)check32);
+                RETURN_ERROR_IF(check32 != h32, checksum_wrong, "");
+            }
+            ZSTD_DCtx_trace_end(dctx, dctx->decodedSize, dctx->processedCSize, /* streaming */ 1);
+            dctx->expected = 0;
+            dctx->stage = ZSTDds_getFrameHeaderSize;
+            return 0;
+        }
+
+    case ZSTDds_decodeSkippableHeader:
+        assert(src != NULL);
+        assert(srcSize <= ZSTD_SKIPPABLEHEADERSIZE);
+        ZSTD_memcpy(dctx->headerBuffer + (ZSTD_SKIPPABLEHEADERSIZE - srcSize), src, srcSize);   /* complete skippable header */
+        dctx->expected = MEM_readLE32(dctx->headerBuffer + ZSTD_FRAMEIDSIZE);   /* note : dctx->expected can grow seriously large, beyond local buffer size */
+        dctx->stage = ZSTDds_skipFrame;
+        return 0;
+
+    case ZSTDds_skipFrame:
+        dctx->expected = 0;
+        dctx->stage = ZSTDds_getFrameHeaderSize;
+        return 0;
+
+    default:
+        assert(0);   /* impossible */
+        RETURN_ERROR(GENERIC, "impossible to reach");   /* some compiler require default to do something */
+    }
+}
+
+
+static size_t ZSTD_refDictContent(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)
+{
+    dctx->dictEnd = dctx->previousDstEnd;
+    dctx->virtualStart = (const char*)dict - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->prefixStart));
+    dctx->prefixStart = dict;
+    dctx->previousDstEnd = (const char*)dict + dictSize;
+#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+    dctx->dictContentBeginForFuzzing = dctx->prefixStart;
+    dctx->dictContentEndForFuzzing = dctx->previousDstEnd;
+#endif
+    return 0;
+}
+
+/*! ZSTD_loadDEntropy() :
+ *  dict : must point at beginning of a valid zstd dictionary.
+ * @return : size of entropy tables read */
+size_t
+ZSTD_loadDEntropy(ZSTD_entropyDTables_t* entropy,
+                  const void* const dict, size_t const dictSize)
+{
+    const BYTE* dictPtr = (const BYTE*)dict;
+    const BYTE* const dictEnd = dictPtr + dictSize;
+
+    RETURN_ERROR_IF(dictSize <= 8, dictionary_corrupted, "dict is too small");
+    assert(MEM_readLE32(dict) == ZSTD_MAGIC_DICTIONARY);   /* dict must be valid */
+    dictPtr += 8;   /* skip header = magic + dictID */
+
+    ZSTD_STATIC_ASSERT(offsetof(ZSTD_entropyDTables_t, OFTable) == offsetof(ZSTD_entropyDTables_t, LLTable) + sizeof(entropy->LLTable));
+    ZSTD_STATIC_ASSERT(offsetof(ZSTD_entropyDTables_t, MLTable) == offsetof(ZSTD_entropyDTables_t, OFTable) + sizeof(entropy->OFTable));
+    ZSTD_STATIC_ASSERT(sizeof(entropy->LLTable) + sizeof(entropy->OFTable) + sizeof(entropy->MLTable) >= HUF_DECOMPRESS_WORKSPACE_SIZE);
+    {   void* const workspace = &entropy->LLTable;   /* use fse tables as temporary workspace; implies fse tables are grouped together */
+        size_t const workspaceSize = sizeof(entropy->LLTable) + sizeof(entropy->OFTable) + sizeof(entropy->MLTable);
+#ifdef HUF_FORCE_DECOMPRESS_X1
+        /* in minimal huffman, we always use X1 variants */
+        size_t const hSize = HUF_readDTableX1_wksp(entropy->hufTable,
+                                                dictPtr, dictEnd - dictPtr,
+                                                workspace, workspaceSize);
+#else
+        size_t const hSize = HUF_readDTableX2_wksp(entropy->hufTable,
+                                                dictPtr, (size_t)(dictEnd - dictPtr),
+                                                workspace, workspaceSize);
+#endif
+        RETURN_ERROR_IF(HUF_isError(hSize), dictionary_corrupted, "");
+        dictPtr += hSize;
+    }
+
+    {   short offcodeNCount[MaxOff+1];
+        unsigned offcodeMaxValue = MaxOff, offcodeLog;
+        size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, (size_t)(dictEnd-dictPtr));
+        RETURN_ERROR_IF(FSE_isError(offcodeHeaderSize), dictionary_corrupted, "");
+        RETURN_ERROR_IF(offcodeMaxValue > MaxOff, dictionary_corrupted, "");
+        RETURN_ERROR_IF(offcodeLog > OffFSELog, dictionary_corrupted, "");
+        ZSTD_buildFSETable( entropy->OFTable,
+                            offcodeNCount, offcodeMaxValue,
+                            OF_base, OF_bits,
+                            offcodeLog,
+                            entropy->workspace, sizeof(entropy->workspace),
+                            /* bmi2 */0);
+        dictPtr += offcodeHeaderSize;
+    }
+
+    {   short matchlengthNCount[MaxML+1];
+        unsigned matchlengthMaxValue = MaxML, matchlengthLog;
+        size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, (size_t)(dictEnd-dictPtr));
+        RETURN_ERROR_IF(FSE_isError(matchlengthHeaderSize), dictionary_corrupted, "");
+        RETURN_ERROR_IF(matchlengthMaxValue > MaxML, dictionary_corrupted, "");
+        RETURN_ERROR_IF(matchlengthLog > MLFSELog, dictionary_corrupted, "");
+        ZSTD_buildFSETable( entropy->MLTable,
+                            matchlengthNCount, matchlengthMaxValue,
+                            ML_base, ML_bits,
+                            matchlengthLog,
+                            entropy->workspace, sizeof(entropy->workspace),
+                            /* bmi2 */ 0);
+        dictPtr += matchlengthHeaderSize;
+    }
+
+    {   short litlengthNCount[MaxLL+1];
+        unsigned litlengthMaxValue = MaxLL, litlengthLog;
+        size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, (size_t)(dictEnd-dictPtr));
+        RETURN_ERROR_IF(FSE_isError(litlengthHeaderSize), dictionary_corrupted, "");
+        RETURN_ERROR_IF(litlengthMaxValue > MaxLL, dictionary_corrupted, "");
+        RETURN_ERROR_IF(litlengthLog > LLFSELog, dictionary_corrupted, "");
+        ZSTD_buildFSETable( entropy->LLTable,
+                            litlengthNCount, litlengthMaxValue,
+                            LL_base, LL_bits,
+                            litlengthLog,
+                            entropy->workspace, sizeof(entropy->workspace),
+                            /* bmi2 */ 0);
+        dictPtr += litlengthHeaderSize;
+    }
+
+    RETURN_ERROR_IF(dictPtr+12 > dictEnd, dictionary_corrupted, "");
+    {   int i;
+        size_t const dictContentSize = (size_t)(dictEnd - (dictPtr+12));
+        for (i=0; i<3; i++) {
+            U32 const rep = MEM_readLE32(dictPtr); dictPtr += 4;
+            RETURN_ERROR_IF(rep==0 || rep > dictContentSize,
+                            dictionary_corrupted, "");
+            entropy->rep[i] = rep;
+    }   }
+
+    return (size_t)(dictPtr - (const BYTE*)dict);
+}
+
+static size_t ZSTD_decompress_insertDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)
+{
+    if (dictSize < 8) return ZSTD_refDictContent(dctx, dict, dictSize);
+    {   U32 const magic = MEM_readLE32(dict);
+        if (magic != ZSTD_MAGIC_DICTIONARY) {
+            return ZSTD_refDictContent(dctx, dict, dictSize);   /* pure content mode */
+    }   }
+    dctx->dictID = MEM_readLE32((const char*)dict + ZSTD_FRAMEIDSIZE);
+
+    /* load entropy tables */
+    {   size_t const eSize = ZSTD_loadDEntropy(&dctx->entropy, dict, dictSize);
+        RETURN_ERROR_IF(ZSTD_isError(eSize), dictionary_corrupted, "");
+        dict = (const char*)dict + eSize;
+        dictSize -= eSize;
+    }
+    dctx->litEntropy = dctx->fseEntropy = 1;
+
+    /* reference dictionary content */
+    return ZSTD_refDictContent(dctx, dict, dictSize);
+}
+
+size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx)
+{
+    assert(dctx != NULL);
+    dctx->expected = ZSTD_startingInputLength(dctx->format);  /* dctx->format must be properly set */
+    dctx->stage = ZSTDds_getFrameHeaderSize;
+    dctx->processedCSize = 0;
+    dctx->decodedSize = 0;
+    dctx->previousDstEnd = NULL;
+    dctx->prefixStart = NULL;
+    dctx->virtualStart = NULL;
+    dctx->dictEnd = NULL;
+    dctx->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001);  /* cover both little and big endian */
+    dctx->litEntropy = dctx->fseEntropy = 0;
+    dctx->dictID = 0;
+    dctx->bType = bt_reserved;
+    ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.rep) == sizeof(repStartValue));
+    ZSTD_memcpy(dctx->entropy.rep, repStartValue, sizeof(repStartValue));  /* initial repcodes */
+    dctx->LLTptr = dctx->entropy.LLTable;
+    dctx->MLTptr = dctx->entropy.MLTable;
+    dctx->OFTptr = dctx->entropy.OFTable;
+    dctx->HUFptr = dctx->entropy.hufTable;
+    return 0;
+}
+
+size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)
+{
+    FORWARD_IF_ERROR( ZSTD_decompressBegin(dctx) , "");
+    if (dict && dictSize)
+        RETURN_ERROR_IF(
+            ZSTD_isError(ZSTD_decompress_insertDictionary(dctx, dict, dictSize)),
+            dictionary_corrupted, "");
+    return 0;
+}
+
+
+/* ======   ZSTD_DDict   ====== */
+
+size_t ZSTD_decompressBegin_usingDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict)
+{
+    DEBUGLOG(4, "ZSTD_decompressBegin_usingDDict");
+    assert(dctx != NULL);
+    if (ddict) {
+        const char* const dictStart = (const char*)ZSTD_DDict_dictContent(ddict);
+        size_t const dictSize = ZSTD_DDict_dictSize(ddict);
+        const void* const dictEnd = dictStart + dictSize;
+        dctx->ddictIsCold = (dctx->dictEnd != dictEnd);
+        DEBUGLOG(4, "DDict is %s",
+                    dctx->ddictIsCold ? "~cold~" : "hot!");
+    }
+    FORWARD_IF_ERROR( ZSTD_decompressBegin(dctx) , "");
+    if (ddict) {   /* NULL ddict is equivalent to no dictionary */
+        ZSTD_copyDDictParameters(dctx, ddict);
+    }
+    return 0;
+}
+
+/*! ZSTD_getDictID_fromDict() :
+ *  Provides the dictID stored within dictionary.
+ *  if @return == 0, the dictionary is not conformant with Zstandard specification.
+ *  It can still be loaded, but as a content-only dictionary. */
+unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize)
+{
+    if (dictSize < 8) return 0;
+    if (MEM_readLE32(dict) != ZSTD_MAGIC_DICTIONARY) return 0;
+    return MEM_readLE32((const char*)dict + ZSTD_FRAMEIDSIZE);
+}
+
+/*! ZSTD_getDictID_fromFrame() :
+ *  Provides the dictID required to decompress frame stored within `src`.
+ *  If @return == 0, the dictID could not be decoded.
+ *  This could for one of the following reasons :
+ *  - The frame does not require a dictionary (most common case).
+ *  - The frame was built with dictID intentionally removed.
+ *    Needed dictionary is a hidden information.
+ *    Note : this use case also happens when using a non-conformant dictionary.
+ *  - `srcSize` is too small, and as a result, frame header could not be decoded.
+ *    Note : possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`.
+ *  - This is not a Zstandard frame.
+ *  When identifying the exact failure cause, it's possible to use
+ *  ZSTD_getFrameHeader(), which will provide a more precise error code. */
+unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize)
+{
+    ZSTD_frameHeader zfp = { 0, 0, 0, ZSTD_frame, 0, 0, 0 };
+    size_t const hError = ZSTD_getFrameHeader(&zfp, src, srcSize);
+    if (ZSTD_isError(hError)) return 0;
+    return zfp.dictID;
+}
+
+
+/*! ZSTD_decompress_usingDDict() :
+*   Decompression using a pre-digested Dictionary
+*   Use dictionary without significant overhead. */
+size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,
+                                  void* dst, size_t dstCapacity,
+                            const void* src, size_t srcSize,
+                            const ZSTD_DDict* ddict)
+{
+    /* pass content and size in case legacy frames are encountered */
+    return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize,
+                                     NULL, 0,
+                                     ddict);
+}
+
+
+/*=====================================
+*   Streaming decompression
+*====================================*/
+
+ZSTD_DStream* ZSTD_createDStream(void)
+{
+    DEBUGLOG(3, "ZSTD_createDStream");
+    return ZSTD_createDStream_advanced(ZSTD_defaultCMem);
+}
+
+ZSTD_DStream* ZSTD_initStaticDStream(void *workspace, size_t workspaceSize)
+{
+    return ZSTD_initStaticDCtx(workspace, workspaceSize);
+}
+
+ZSTD_DStream* ZSTD_createDStream_advanced(ZSTD_customMem customMem)
+{
+    return ZSTD_createDCtx_advanced(customMem);
+}
+
+size_t ZSTD_freeDStream(ZSTD_DStream* zds)
+{
+    return ZSTD_freeDCtx(zds);
+}
+
+
+/* ***  Initialization  *** */
+
+size_t ZSTD_DStreamInSize(void)  { return ZSTD_BLOCKSIZE_MAX + ZSTD_blockHeaderSize; }
+size_t ZSTD_DStreamOutSize(void) { return ZSTD_BLOCKSIZE_MAX; }
+
+size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx,
+                                   const void* dict, size_t dictSize,
+                                         ZSTD_dictLoadMethod_e dictLoadMethod,
+                                         ZSTD_dictContentType_e dictContentType)
+{
+    RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong, "");
+    ZSTD_clearDict(dctx);
+    if (dict && dictSize != 0) {
+        dctx->ddictLocal = ZSTD_createDDict_advanced(dict, dictSize, dictLoadMethod, dictContentType, dctx->customMem);
+        RETURN_ERROR_IF(dctx->ddictLocal == NULL, memory_allocation, "NULL pointer!");
+        dctx->ddict = dctx->ddictLocal;
+        dctx->dictUses = ZSTD_use_indefinitely;
+    }
+    return 0;
+}
+
+size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)
+{
+    return ZSTD_DCtx_loadDictionary_advanced(dctx, dict, dictSize, ZSTD_dlm_byRef, ZSTD_dct_auto);
+}
+
+size_t ZSTD_DCtx_loadDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)
+{
+    return ZSTD_DCtx_loadDictionary_advanced(dctx, dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dct_auto);
+}
+
+size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType)
+{
+    FORWARD_IF_ERROR(ZSTD_DCtx_loadDictionary_advanced(dctx, prefix, prefixSize, ZSTD_dlm_byRef, dictContentType), "");
+    dctx->dictUses = ZSTD_use_once;
+    return 0;
+}
+
+size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize)
+{
+    return ZSTD_DCtx_refPrefix_advanced(dctx, prefix, prefixSize, ZSTD_dct_rawContent);
+}
+
+
+/* ZSTD_initDStream_usingDict() :
+ * return : expected size, aka ZSTD_startingInputLength().
+ * this function cannot fail */
+size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize)
+{
+    DEBUGLOG(4, "ZSTD_initDStream_usingDict");
+    FORWARD_IF_ERROR( ZSTD_DCtx_reset(zds, ZSTD_reset_session_only) , "");
+    FORWARD_IF_ERROR( ZSTD_DCtx_loadDictionary(zds, dict, dictSize) , "");
+    return ZSTD_startingInputLength(zds->format);
+}
+
+/* note : this variant can't fail */
+size_t ZSTD_initDStream(ZSTD_DStream* zds)
+{
+    DEBUGLOG(4, "ZSTD_initDStream");
+    return ZSTD_initDStream_usingDDict(zds, NULL);
+}
+
+/* ZSTD_initDStream_usingDDict() :
+ * ddict will just be referenced, and must outlive decompression session
+ * this function cannot fail */
+size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* dctx, const ZSTD_DDict* ddict)
+{
+    FORWARD_IF_ERROR( ZSTD_DCtx_reset(dctx, ZSTD_reset_session_only) , "");
+    FORWARD_IF_ERROR( ZSTD_DCtx_refDDict(dctx, ddict) , "");
+    return ZSTD_startingInputLength(dctx->format);
+}
+
+/* ZSTD_resetDStream() :
+ * return : expected size, aka ZSTD_startingInputLength().
+ * this function cannot fail */
+size_t ZSTD_resetDStream(ZSTD_DStream* dctx)
+{
+    FORWARD_IF_ERROR(ZSTD_DCtx_reset(dctx, ZSTD_reset_session_only), "");
+    return ZSTD_startingInputLength(dctx->format);
+}
+
+
+size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict)
+{
+    RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong, "");
+    ZSTD_clearDict(dctx);
+    if (ddict) {
+        dctx->ddict = ddict;
+        dctx->dictUses = ZSTD_use_indefinitely;
+        if (dctx->refMultipleDDicts == ZSTD_rmd_refMultipleDDicts) {
+            if (dctx->ddictSet == NULL) {
+                dctx->ddictSet = ZSTD_createDDictHashSet(dctx->customMem);
+                if (!dctx->ddictSet) {
+                    RETURN_ERROR(memory_allocation, "Failed to allocate memory for hash set!");
+                }
+            }
+            assert(!dctx->staticSize);  /* Impossible: ddictSet cannot have been allocated if static dctx */
+            FORWARD_IF_ERROR(ZSTD_DDictHashSet_addDDict(dctx->ddictSet, ddict, dctx->customMem), "");
+        }
+    }
+    return 0;
+}
+
+/* ZSTD_DCtx_setMaxWindowSize() :
+ * note : no direct equivalence in ZSTD_DCtx_setParameter,
+ * since this version sets windowSize, and the other sets windowLog */
+size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize)
+{
+    ZSTD_bounds const bounds = ZSTD_dParam_getBounds(ZSTD_d_windowLogMax);
+    size_t const min = (size_t)1 << bounds.lowerBound;
+    size_t const max = (size_t)1 << bounds.upperBound;
+    RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong, "");
+    RETURN_ERROR_IF(maxWindowSize < min, parameter_outOfBound, "");
+    RETURN_ERROR_IF(maxWindowSize > max, parameter_outOfBound, "");
+    dctx->maxWindowSize = maxWindowSize;
+    return 0;
+}
+
+size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format)
+{
+    return ZSTD_DCtx_setParameter(dctx, ZSTD_d_format, (int)format);
+}
+
+ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam)
+{
+    ZSTD_bounds bounds = { 0, 0, 0 };
+    switch(dParam) {
+        case ZSTD_d_windowLogMax:
+            bounds.lowerBound = ZSTD_WINDOWLOG_ABSOLUTEMIN;
+            bounds.upperBound = ZSTD_WINDOWLOG_MAX;
+            return bounds;
+        case ZSTD_d_format:
+            bounds.lowerBound = (int)ZSTD_f_zstd1;
+            bounds.upperBound = (int)ZSTD_f_zstd1_magicless;
+            ZSTD_STATIC_ASSERT(ZSTD_f_zstd1 < ZSTD_f_zstd1_magicless);
+            return bounds;
+        case ZSTD_d_stableOutBuffer:
+            bounds.lowerBound = (int)ZSTD_bm_buffered;
+            bounds.upperBound = (int)ZSTD_bm_stable;
+            return bounds;
+        case ZSTD_d_forceIgnoreChecksum:
+            bounds.lowerBound = (int)ZSTD_d_validateChecksum;
+            bounds.upperBound = (int)ZSTD_d_ignoreChecksum;
+            return bounds;
+        case ZSTD_d_refMultipleDDicts:
+            bounds.lowerBound = (int)ZSTD_rmd_refSingleDDict;
+            bounds.upperBound = (int)ZSTD_rmd_refMultipleDDicts;
+            return bounds;
+        default:;
+    }
+    bounds.error = ERROR(parameter_unsupported);
+    return bounds;
+}
+
+/* ZSTD_dParam_withinBounds:
+ * @return 1 if value is within dParam bounds,
+ * 0 otherwise */
+static int ZSTD_dParam_withinBounds(ZSTD_dParameter dParam, int value)
+{
+    ZSTD_bounds const bounds = ZSTD_dParam_getBounds(dParam);
+    if (ZSTD_isError(bounds.error)) return 0;
+    if (value < bounds.lowerBound) return 0;
+    if (value > bounds.upperBound) return 0;
+    return 1;
+}
+
+#define CHECK_DBOUNDS(p,v) {                \
+    RETURN_ERROR_IF(!ZSTD_dParam_withinBounds(p, v), parameter_outOfBound, ""); \
+}
+
+size_t ZSTD_DCtx_getParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int* value)
+{
+    switch (param) {
+        case ZSTD_d_windowLogMax:
+            *value = (int)ZSTD_highbit32((U32)dctx->maxWindowSize);
+            return 0;
+        case ZSTD_d_format:
+            *value = (int)dctx->format;
+            return 0;
+        case ZSTD_d_stableOutBuffer:
+            *value = (int)dctx->outBufferMode;
+            return 0;
+        case ZSTD_d_forceIgnoreChecksum:
+            *value = (int)dctx->forceIgnoreChecksum;
+            return 0;
+        case ZSTD_d_refMultipleDDicts:
+            *value = (int)dctx->refMultipleDDicts;
+            return 0;
+        default:;
+    }
+    RETURN_ERROR(parameter_unsupported, "");
+}
+
+size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter dParam, int value)
+{
+    RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong, "");
+    switch(dParam) {
+        case ZSTD_d_windowLogMax:
+            if (value == 0) value = ZSTD_WINDOWLOG_LIMIT_DEFAULT;
+            CHECK_DBOUNDS(ZSTD_d_windowLogMax, value);
+            dctx->maxWindowSize = ((size_t)1) << value;
+            return 0;
+        case ZSTD_d_format:
+            CHECK_DBOUNDS(ZSTD_d_format, value);
+            dctx->format = (ZSTD_format_e)value;
+            return 0;
+        case ZSTD_d_stableOutBuffer:
+            CHECK_DBOUNDS(ZSTD_d_stableOutBuffer, value);
+            dctx->outBufferMode = (ZSTD_bufferMode_e)value;
+            return 0;
+        case ZSTD_d_forceIgnoreChecksum:
+            CHECK_DBOUNDS(ZSTD_d_forceIgnoreChecksum, value);
+            dctx->forceIgnoreChecksum = (ZSTD_forceIgnoreChecksum_e)value;
+            return 0;
+        case ZSTD_d_refMultipleDDicts:
+            CHECK_DBOUNDS(ZSTD_d_refMultipleDDicts, value);
+            if (dctx->staticSize != 0) {
+                RETURN_ERROR(parameter_unsupported, "Static dctx does not support multiple DDicts!");
+            }
+            dctx->refMultipleDDicts = (ZSTD_refMultipleDDicts_e)value;
+            return 0;
+        default:;
+    }
+    RETURN_ERROR(parameter_unsupported, "");
+}
+
+size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset)
+{
+    if ( (reset == ZSTD_reset_session_only)
+      || (reset == ZSTD_reset_session_and_parameters) ) {
+        dctx->streamStage = zdss_init;
+        dctx->noForwardProgress = 0;
+    }
+    if ( (reset == ZSTD_reset_parameters)
+      || (reset == ZSTD_reset_session_and_parameters) ) {
+        RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong, "");
+        ZSTD_clearDict(dctx);
+        ZSTD_DCtx_resetParameters(dctx);
+    }
+    return 0;
+}
+
+
+size_t ZSTD_sizeof_DStream(const ZSTD_DStream* dctx)
+{
+    return ZSTD_sizeof_DCtx(dctx);
+}
+
+size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize)
+{
+    size_t const blockSize = (size_t) MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
+    unsigned long long const neededRBSize = windowSize + blockSize + (WILDCOPY_OVERLENGTH * 2);
+    unsigned long long const neededSize = MIN(frameContentSize, neededRBSize);
+    size_t const minRBSize = (size_t) neededSize;
+    RETURN_ERROR_IF((unsigned long long)minRBSize != neededSize,
+                    frameParameter_windowTooLarge, "");
+    return minRBSize;
+}
+
+size_t ZSTD_estimateDStreamSize(size_t windowSize)
+{
+    size_t const blockSize = MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
+    size_t const inBuffSize = blockSize;  /* no block can be larger */
+    size_t const outBuffSize = ZSTD_decodingBufferSize_min(windowSize, ZSTD_CONTENTSIZE_UNKNOWN);
+    return ZSTD_estimateDCtxSize() + inBuffSize + outBuffSize;
+}
+
+size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize)
+{
+    U32 const windowSizeMax = 1U << ZSTD_WINDOWLOG_MAX;   /* note : should be user-selectable, but requires an additional parameter (or a dctx) */
+    ZSTD_frameHeader zfh;
+    size_t const err = ZSTD_getFrameHeader(&zfh, src, srcSize);
+    if (ZSTD_isError(err)) return err;
+    RETURN_ERROR_IF(err>0, srcSize_wrong, "");
+    RETURN_ERROR_IF(zfh.windowSize > windowSizeMax,
+                    frameParameter_windowTooLarge, "");
+    return ZSTD_estimateDStreamSize((size_t)zfh.windowSize);
+}
+
+
+/* *****   Decompression   ***** */
+
+static int ZSTD_DCtx_isOverflow(ZSTD_DStream* zds, size_t const neededInBuffSize, size_t const neededOutBuffSize)
+{
+    return (zds->inBuffSize + zds->outBuffSize) >= (neededInBuffSize + neededOutBuffSize) * ZSTD_WORKSPACETOOLARGE_FACTOR;
+}
+
+static void ZSTD_DCtx_updateOversizedDuration(ZSTD_DStream* zds, size_t const neededInBuffSize, size_t const neededOutBuffSize)
+{
+    if (ZSTD_DCtx_isOverflow(zds, neededInBuffSize, neededOutBuffSize))
+        zds->oversizedDuration++;
+    else
+        zds->oversizedDuration = 0;
+}
+
+static int ZSTD_DCtx_isOversizedTooLong(ZSTD_DStream* zds)
+{
+    return zds->oversizedDuration >= ZSTD_WORKSPACETOOLARGE_MAXDURATION;
+}
+
+/* Checks that the output buffer hasn't changed if ZSTD_obm_stable is used. */
+static size_t ZSTD_checkOutBuffer(ZSTD_DStream const* zds, ZSTD_outBuffer const* output)
+{
+    ZSTD_outBuffer const expect = zds->expectedOutBuffer;
+    /* No requirement when ZSTD_obm_stable is not enabled. */
+    if (zds->outBufferMode != ZSTD_bm_stable)
+        return 0;
+    /* Any buffer is allowed in zdss_init, this must be the same for every other call until
+     * the context is reset.
+     */
+    if (zds->streamStage == zdss_init)
+        return 0;
+    /* The buffer must match our expectation exactly. */
+    if (expect.dst == output->dst && expect.pos == output->pos && expect.size == output->size)
+        return 0;
+    RETURN_ERROR(dstBuffer_wrong, "ZSTD_d_stableOutBuffer enabled but output differs!");
+}
+
+/* Calls ZSTD_decompressContinue() with the right parameters for ZSTD_decompressStream()
+ * and updates the stage and the output buffer state. This call is extracted so it can be
+ * used both when reading directly from the ZSTD_inBuffer, and in buffered input mode.
+ * NOTE: You must break after calling this function since the streamStage is modified.
+ */
+static size_t ZSTD_decompressContinueStream(
+            ZSTD_DStream* zds, char** op, char* oend,
+            void const* src, size_t srcSize) {
+    int const isSkipFrame = ZSTD_isSkipFrame(zds);
+    if (zds->outBufferMode == ZSTD_bm_buffered) {
+        size_t const dstSize = isSkipFrame ? 0 : zds->outBuffSize - zds->outStart;
+        size_t const decodedSize = ZSTD_decompressContinue(zds,
+                zds->outBuff + zds->outStart, dstSize, src, srcSize);
+        FORWARD_IF_ERROR(decodedSize, "");
+        if (!decodedSize && !isSkipFrame) {
+            zds->streamStage = zdss_read;
+        } else {
+            zds->outEnd = zds->outStart + decodedSize;
+            zds->streamStage = zdss_flush;
+        }
+    } else {
+        /* Write directly into the output buffer */
+        size_t const dstSize = isSkipFrame ? 0 : (size_t)(oend - *op);
+        size_t const decodedSize = ZSTD_decompressContinue(zds, *op, dstSize, src, srcSize);
+        FORWARD_IF_ERROR(decodedSize, "");
+        *op += decodedSize;
+        /* Flushing is not needed. */
+        zds->streamStage = zdss_read;
+        assert(*op <= oend);
+        assert(zds->outBufferMode == ZSTD_bm_stable);
+    }
+    return 0;
+}
+
+size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input)
+{
+    const char* const src = (const char*)input->src;
+    const char* const istart = input->pos != 0 ? src + input->pos : src;
+    const char* const iend = input->size != 0 ? src + input->size : src;
+    const char* ip = istart;
+    char* const dst = (char*)output->dst;
+    char* const ostart = output->pos != 0 ? dst + output->pos : dst;
+    char* const oend = output->size != 0 ? dst + output->size : dst;
+    char* op = ostart;
+    U32 someMoreWork = 1;
+
+    DEBUGLOG(5, "ZSTD_decompressStream");
+    RETURN_ERROR_IF(
+        input->pos > input->size,
+        srcSize_wrong,
+        "forbidden. in: pos: %u   vs size: %u",
+        (U32)input->pos, (U32)input->size);
+    RETURN_ERROR_IF(
+        output->pos > output->size,
+        dstSize_tooSmall,
+        "forbidden. out: pos: %u   vs size: %u",
+        (U32)output->pos, (U32)output->size);
+    DEBUGLOG(5, "input size : %u", (U32)(input->size - input->pos));
+    FORWARD_IF_ERROR(ZSTD_checkOutBuffer(zds, output), "");
+
+    while (someMoreWork) {
+        switch(zds->streamStage)
+        {
+        case zdss_init :
+            DEBUGLOG(5, "stage zdss_init => transparent reset ");
+            zds->streamStage = zdss_loadHeader;
+            zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0;
+            zds->legacyVersion = 0;
+            zds->hostageByte = 0;
+            zds->expectedOutBuffer = *output;
+            ZSTD_FALLTHROUGH;
+
+        case zdss_loadHeader :
+            DEBUGLOG(5, "stage zdss_loadHeader (srcSize : %u)", (U32)(iend - ip));
+            {   size_t const hSize = ZSTD_getFrameHeader_advanced(&zds->fParams, zds->headerBuffer, zds->lhSize, zds->format);
+                if (zds->refMultipleDDicts && zds->ddictSet) {
+                    ZSTD_DCtx_selectFrameDDict(zds);
+                }
+                DEBUGLOG(5, "header size : %u", (U32)hSize);
+                if (ZSTD_isError(hSize)) {
+                    return hSize;   /* error */
+                }
+                if (hSize != 0) {   /* need more input */
+                    size_t const toLoad = hSize - zds->lhSize;   /* if hSize!=0, hSize > zds->lhSize */
+                    size_t const remainingInput = (size_t)(iend-ip);
+                    assert(iend >= ip);
+                    if (toLoad > remainingInput) {   /* not enough input to load full header */
+                        if (remainingInput > 0) {
+                            ZSTD_memcpy(zds->headerBuffer + zds->lhSize, ip, remainingInput);
+                            zds->lhSize += remainingInput;
+                        }
+                        input->pos = input->size;
+                        return (MAX((size_t)ZSTD_FRAMEHEADERSIZE_MIN(zds->format), hSize) - zds->lhSize) + ZSTD_blockHeaderSize;   /* remaining header bytes + next block header */
+                    }
+                    assert(ip != NULL);
+                    ZSTD_memcpy(zds->headerBuffer + zds->lhSize, ip, toLoad); zds->lhSize = hSize; ip += toLoad;
+                    break;
+            }   }
+
+            /* check for single-pass mode opportunity */
+            if (zds->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN
+                && zds->fParams.frameType != ZSTD_skippableFrame
+                && (U64)(size_t)(oend-op) >= zds->fParams.frameContentSize) {
+                size_t const cSize = ZSTD_findFrameCompressedSize(istart, (size_t)(iend-istart));
+                if (cSize <= (size_t)(iend-istart)) {
+                    /* shortcut : using single-pass mode */
+                    size_t const decompressedSize = ZSTD_decompress_usingDDict(zds, op, (size_t)(oend-op), istart, cSize, ZSTD_getDDict(zds));
+                    if (ZSTD_isError(decompressedSize)) return decompressedSize;
+                    DEBUGLOG(4, "shortcut to single-pass ZSTD_decompress_usingDDict()")
+                    ip = istart + cSize;
+                    op += decompressedSize;
+                    zds->expected = 0;
+                    zds->streamStage = zdss_init;
+                    someMoreWork = 0;
+                    break;
+            }   }
+
+            /* Check output buffer is large enough for ZSTD_odm_stable. */
+            if (zds->outBufferMode == ZSTD_bm_stable
+                && zds->fParams.frameType != ZSTD_skippableFrame
+                && zds->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN
+                && (U64)(size_t)(oend-op) < zds->fParams.frameContentSize) {
+                RETURN_ERROR(dstSize_tooSmall, "ZSTD_obm_stable passed but ZSTD_outBuffer is too small");
+            }
+
+            /* Consume header (see ZSTDds_decodeFrameHeader) */
+            DEBUGLOG(4, "Consume header");
+            FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDDict(zds, ZSTD_getDDict(zds)), "");
+
+            if ((MEM_readLE32(zds->headerBuffer) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {  /* skippable frame */
+                zds->expected = MEM_readLE32(zds->headerBuffer + ZSTD_FRAMEIDSIZE);
+                zds->stage = ZSTDds_skipFrame;
+            } else {
+                FORWARD_IF_ERROR(ZSTD_decodeFrameHeader(zds, zds->headerBuffer, zds->lhSize), "");
+                zds->expected = ZSTD_blockHeaderSize;
+                zds->stage = ZSTDds_decodeBlockHeader;
+            }
+
+            /* control buffer memory usage */
+            DEBUGLOG(4, "Control max memory usage (%u KB <= max %u KB)",
+                        (U32)(zds->fParams.windowSize >>10),
+                        (U32)(zds->maxWindowSize >> 10) );
+            zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN);
+            RETURN_ERROR_IF(zds->fParams.windowSize > zds->maxWindowSize,
+                            frameParameter_windowTooLarge, "");
+
+            /* Adapt buffer sizes to frame header instructions */
+            {   size_t const neededInBuffSize = MAX(zds->fParams.blockSizeMax, 4 /* frame checksum */);
+                size_t const neededOutBuffSize = zds->outBufferMode == ZSTD_bm_buffered
+                        ? ZSTD_decodingBufferSize_min(zds->fParams.windowSize, zds->fParams.frameContentSize)
+                        : 0;
+
+                ZSTD_DCtx_updateOversizedDuration(zds, neededInBuffSize, neededOutBuffSize);
+
+                {   int const tooSmall = (zds->inBuffSize < neededInBuffSize) || (zds->outBuffSize < neededOutBuffSize);
+                    int const tooLarge = ZSTD_DCtx_isOversizedTooLong(zds);
+
+                    if (tooSmall || tooLarge) {
+                        size_t const bufferSize = neededInBuffSize + neededOutBuffSize;
+                        DEBUGLOG(4, "inBuff  : from %u to %u",
+                                    (U32)zds->inBuffSize, (U32)neededInBuffSize);
+                        DEBUGLOG(4, "outBuff : from %u to %u",
+                                    (U32)zds->outBuffSize, (U32)neededOutBuffSize);
+                        if (zds->staticSize) {  /* static DCtx */
+                            DEBUGLOG(4, "staticSize : %u", (U32)zds->staticSize);
+                            assert(zds->staticSize >= sizeof(ZSTD_DCtx));  /* controlled at init */
+                            RETURN_ERROR_IF(
+                                bufferSize > zds->staticSize - sizeof(ZSTD_DCtx),
+                                memory_allocation, "");
+                        } else {
+                            ZSTD_customFree(zds->inBuff, zds->customMem);
+                            zds->inBuffSize = 0;
+                            zds->outBuffSize = 0;
+                            zds->inBuff = (char*)ZSTD_customMalloc(bufferSize, zds->customMem);
+                            RETURN_ERROR_IF(zds->inBuff == NULL, memory_allocation, "");
+                        }
+                        zds->inBuffSize = neededInBuffSize;
+                        zds->outBuff = zds->inBuff + zds->inBuffSize;
+                        zds->outBuffSize = neededOutBuffSize;
+            }   }   }
+            zds->streamStage = zdss_read;
+            ZSTD_FALLTHROUGH;
+
+        case zdss_read:
+            DEBUGLOG(5, "stage zdss_read");
+            {   size_t const neededInSize = ZSTD_nextSrcSizeToDecompressWithInputSize(zds, (size_t)(iend - ip));
+                DEBUGLOG(5, "neededInSize = %u", (U32)neededInSize);
+                if (neededInSize==0) {  /* end of frame */
+                    zds->streamStage = zdss_init;
+                    someMoreWork = 0;
+                    break;
+                }
+                if ((size_t)(iend-ip) >= neededInSize) {  /* decode directly from src */
+                    FORWARD_IF_ERROR(ZSTD_decompressContinueStream(zds, &op, oend, ip, neededInSize), "");
+                    ip += neededInSize;
+                    /* Function modifies the stage so we must break */
+                    break;
+            }   }
+            if (ip==iend) { someMoreWork = 0; break; }   /* no more input */
+            zds->streamStage = zdss_load;
+            ZSTD_FALLTHROUGH;
+
+        case zdss_load:
+            {   size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds);
+                size_t const toLoad = neededInSize - zds->inPos;
+                int const isSkipFrame = ZSTD_isSkipFrame(zds);
+                size_t loadedSize;
+                /* At this point we shouldn't be decompressing a block that we can stream. */
+                assert(neededInSize == ZSTD_nextSrcSizeToDecompressWithInputSize(zds, iend - ip));
+                if (isSkipFrame) {
+                    loadedSize = MIN(toLoad, (size_t)(iend-ip));
+                } else {
+                    RETURN_ERROR_IF(toLoad > zds->inBuffSize - zds->inPos,
+                                    corruption_detected,
+                                    "should never happen");
+                    loadedSize = ZSTD_limitCopy(zds->inBuff + zds->inPos, toLoad, ip, (size_t)(iend-ip));
+                }
+                ip += loadedSize;
+                zds->inPos += loadedSize;
+                if (loadedSize < toLoad) { someMoreWork = 0; break; }   /* not enough input, wait for more */
+
+                /* decode loaded input */
+                zds->inPos = 0;   /* input is consumed */
+                FORWARD_IF_ERROR(ZSTD_decompressContinueStream(zds, &op, oend, zds->inBuff, neededInSize), "");
+                /* Function modifies the stage so we must break */
+                break;
+            }
+        case zdss_flush:
+            {   size_t const toFlushSize = zds->outEnd - zds->outStart;
+                size_t const flushedSize = ZSTD_limitCopy(op, (size_t)(oend-op), zds->outBuff + zds->outStart, toFlushSize);
+                op += flushedSize;
+                zds->outStart += flushedSize;
+                if (flushedSize == toFlushSize) {  /* flush completed */
+                    zds->streamStage = zdss_read;
+                    if ( (zds->outBuffSize < zds->fParams.frameContentSize)
+                      && (zds->outStart + zds->fParams.blockSizeMax > zds->outBuffSize) ) {
+                        DEBUGLOG(5, "restart filling outBuff from beginning (left:%i, needed:%u)",
+                                (int)(zds->outBuffSize - zds->outStart),
+                                (U32)zds->fParams.blockSizeMax);
+                        zds->outStart = zds->outEnd = 0;
+                    }
+                    break;
+            }   }
+            /* cannot complete flush */
+            someMoreWork = 0;
+            break;
+
+        default:
+            assert(0);    /* impossible */
+            RETURN_ERROR(GENERIC, "impossible to reach");   /* some compiler require default to do something */
+    }   }
+
+    /* result */
+    input->pos = (size_t)(ip - (const char*)(input->src));
+    output->pos = (size_t)(op - (char*)(output->dst));
+
+    /* Update the expected output buffer for ZSTD_obm_stable. */
+    zds->expectedOutBuffer = *output;
+
+    if ((ip==istart) && (op==ostart)) {  /* no forward progress */
+        zds->noForwardProgress ++;
+        if (zds->noForwardProgress >= ZSTD_NO_FORWARD_PROGRESS_MAX) {
+            RETURN_ERROR_IF(op==oend, dstSize_tooSmall, "");
+            RETURN_ERROR_IF(ip==iend, srcSize_wrong, "");
+            assert(0);
+        }
+    } else {
+        zds->noForwardProgress = 0;
+    }
+    {   size_t nextSrcSizeHint = ZSTD_nextSrcSizeToDecompress(zds);
+        if (!nextSrcSizeHint) {   /* frame fully decoded */
+            if (zds->outEnd == zds->outStart) {  /* output fully flushed */
+                if (zds->hostageByte) {
+                    if (input->pos >= input->size) {
+                        /* can't release hostage (not present) */
+                        zds->streamStage = zdss_read;
+                        return 1;
+                    }
+                    input->pos++;  /* release hostage */
+                }   /* zds->hostageByte */
+                return 0;
+            }  /* zds->outEnd == zds->outStart */
+            if (!zds->hostageByte) { /* output not fully flushed; keep last byte as hostage; will be released when all output is flushed */
+                input->pos--;   /* note : pos > 0, otherwise, impossible to finish reading last block */
+                zds->hostageByte=1;
+            }
+            return 1;
+        }  /* nextSrcSizeHint==0 */
+        nextSrcSizeHint += ZSTD_blockHeaderSize * (ZSTD_nextInputType(zds) == ZSTDnit_block);   /* preload header of next block */
+        assert(zds->inPos <= nextSrcSizeHint);
+        nextSrcSizeHint -= zds->inPos;   /* part already loaded*/
+        return nextSrcSizeHint;
+    }
+}
+
+size_t ZSTD_decompressStream_simpleArgs (
+                            ZSTD_DCtx* dctx,
+                            void* dst, size_t dstCapacity, size_t* dstPos,
+                      const void* src, size_t srcSize, size_t* srcPos)
+{
+    ZSTD_outBuffer output = { dst, dstCapacity, *dstPos };
+    ZSTD_inBuffer  input  = { src, srcSize, *srcPos };
+    /* ZSTD_compress_generic() will check validity of dstPos and srcPos */
+    size_t const cErr = ZSTD_decompressStream(dctx, &output, &input);
+    *dstPos = output.pos;
+    *srcPos = input.pos;
+    return cErr;
+}
diff --git a/lib/zstd/decompress/zstd_decompress_block.c b/lib/zstd/decompress/zstd_decompress_block.c
new file mode 100644
index 0000000..2d101d9
--- /dev/null
+++ b/lib/zstd/decompress/zstd_decompress_block.c
@@ -0,0 +1,1540 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/* zstd_decompress_block :
+ * this module takes care of decompressing _compressed_ block */
+
+/*-*******************************************************
+*  Dependencies
+*********************************************************/
+#include "../common/zstd_deps.h"   /* ZSTD_memcpy, ZSTD_memmove, ZSTD_memset */
+#include "../common/compiler.h"    /* prefetch */
+#include "../common/cpu.h"         /* bmi2 */
+#include "../common/mem.h"         /* low level memory routines */
+#define FSE_STATIC_LINKING_ONLY
+#include "../common/fse.h"
+#define HUF_STATIC_LINKING_ONLY
+#include "../common/huf.h"
+#include "../common/zstd_internal.h"
+#include "zstd_decompress_internal.h"   /* ZSTD_DCtx */
+#include "zstd_ddict.h"  /* ZSTD_DDictDictContent */
+#include "zstd_decompress_block.h"
+
+/*_*******************************************************
+*  Macros
+**********************************************************/
+
+/* These two optional macros force the use one way or another of the two
+ * ZSTD_decompressSequences implementations. You can't force in both directions
+ * at the same time.
+ */
+#if defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT) && \
+    defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG)
+#error "Cannot force the use of the short and the long ZSTD_decompressSequences variants!"
+#endif
+
+
+/*_*******************************************************
+*  Memory operations
+**********************************************************/
+static void ZSTD_copy4(void* dst, const void* src) { ZSTD_memcpy(dst, src, 4); }
+
+
+/*-*************************************************************
+ *   Block decoding
+ ***************************************************************/
+
+/*! ZSTD_getcBlockSize() :
+ *  Provides the size of compressed block from block header `src` */
+size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
+                          blockProperties_t* bpPtr)
+{
+    RETURN_ERROR_IF(srcSize < ZSTD_blockHeaderSize, srcSize_wrong, "");
+
+    {   U32 const cBlockHeader = MEM_readLE24(src);
+        U32 const cSize = cBlockHeader >> 3;
+        bpPtr->lastBlock = cBlockHeader & 1;
+        bpPtr->blockType = (blockType_e)((cBlockHeader >> 1) & 3);
+        bpPtr->origSize = cSize;   /* only useful for RLE */
+        if (bpPtr->blockType == bt_rle) return 1;
+        RETURN_ERROR_IF(bpPtr->blockType == bt_reserved, corruption_detected, "");
+        return cSize;
+    }
+}
+
+
+/* Hidden declaration for fullbench */
+size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
+                          const void* src, size_t srcSize);
+/*! ZSTD_decodeLiteralsBlock() :
+ * @return : nb of bytes read from src (< srcSize )
+ *  note : symbol not declared but exposed for fullbench */
+size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
+                          const void* src, size_t srcSize)   /* note : srcSize < BLOCKSIZE */
+{
+    DEBUGLOG(5, "ZSTD_decodeLiteralsBlock");
+    RETURN_ERROR_IF(srcSize < MIN_CBLOCK_SIZE, corruption_detected, "");
+
+    {   const BYTE* const istart = (const BYTE*) src;
+        symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3);
+
+        switch(litEncType)
+        {
+        case set_repeat:
+            DEBUGLOG(5, "set_repeat flag : re-using stats from previous compressed literals block");
+            RETURN_ERROR_IF(dctx->litEntropy==0, dictionary_corrupted, "");
+            ZSTD_FALLTHROUGH;
+
+        case set_compressed:
+            RETURN_ERROR_IF(srcSize < 5, corruption_detected, "srcSize >= MIN_CBLOCK_SIZE == 3; here we need up to 5 for case 3");
+            {   size_t lhSize, litSize, litCSize;
+                U32 singleStream=0;
+                U32 const lhlCode = (istart[0] >> 2) & 3;
+                U32 const lhc = MEM_readLE32(istart);
+                size_t hufSuccess;
+                switch(lhlCode)
+                {
+                case 0: case 1: default:   /* note : default is impossible, since lhlCode into [0..3] */
+                    /* 2 - 2 - 10 - 10 */
+                    singleStream = !lhlCode;
+                    lhSize = 3;
+                    litSize  = (lhc >> 4) & 0x3FF;
+                    litCSize = (lhc >> 14) & 0x3FF;
+                    break;
+                case 2:
+                    /* 2 - 2 - 14 - 14 */
+                    lhSize = 4;
+                    litSize  = (lhc >> 4) & 0x3FFF;
+                    litCSize = lhc >> 18;
+                    break;
+                case 3:
+                    /* 2 - 2 - 18 - 18 */
+                    lhSize = 5;
+                    litSize  = (lhc >> 4) & 0x3FFFF;
+                    litCSize = (lhc >> 22) + ((size_t)istart[4] << 10);
+                    break;
+                }
+                RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected, "");
+                RETURN_ERROR_IF(litCSize + lhSize > srcSize, corruption_detected, "");
+
+                /* prefetch huffman table if cold */
+                if (dctx->ddictIsCold && (litSize > 768 /* heuristic */)) {
+                    PREFETCH_AREA(dctx->HUFptr, sizeof(dctx->entropy.hufTable));
+                }
+
+                if (litEncType==set_repeat) {
+                    if (singleStream) {
+                        hufSuccess = HUF_decompress1X_usingDTable_bmi2(
+                            dctx->litBuffer, litSize, istart+lhSize, litCSize,
+                            dctx->HUFptr, dctx->bmi2);
+                    } else {
+                        hufSuccess = HUF_decompress4X_usingDTable_bmi2(
+                            dctx->litBuffer, litSize, istart+lhSize, litCSize,
+                            dctx->HUFptr, dctx->bmi2);
+                    }
+                } else {
+                    if (singleStream) {
+#if defined(HUF_FORCE_DECOMPRESS_X2)
+                        hufSuccess = HUF_decompress1X_DCtx_wksp(
+                            dctx->entropy.hufTable, dctx->litBuffer, litSize,
+                            istart+lhSize, litCSize, dctx->workspace,
+                            sizeof(dctx->workspace));
+#else
+                        hufSuccess = HUF_decompress1X1_DCtx_wksp_bmi2(
+                            dctx->entropy.hufTable, dctx->litBuffer, litSize,
+                            istart+lhSize, litCSize, dctx->workspace,
+                            sizeof(dctx->workspace), dctx->bmi2);
+#endif
+                    } else {
+                        hufSuccess = HUF_decompress4X_hufOnly_wksp_bmi2(
+                            dctx->entropy.hufTable, dctx->litBuffer, litSize,
+                            istart+lhSize, litCSize, dctx->workspace,
+                            sizeof(dctx->workspace), dctx->bmi2);
+                    }
+                }
+
+                RETURN_ERROR_IF(HUF_isError(hufSuccess), corruption_detected, "");
+
+                dctx->litPtr = dctx->litBuffer;
+                dctx->litSize = litSize;
+                dctx->litEntropy = 1;
+                if (litEncType==set_compressed) dctx->HUFptr = dctx->entropy.hufTable;
+                ZSTD_memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);
+                return litCSize + lhSize;
+            }
+
+        case set_basic:
+            {   size_t litSize, lhSize;
+                U32 const lhlCode = ((istart[0]) >> 2) & 3;
+                switch(lhlCode)
+                {
+                case 0: case 2: default:   /* note : default is impossible, since lhlCode into [0..3] */
+                    lhSize = 1;
+                    litSize = istart[0] >> 3;
+                    break;
+                case 1:
+                    lhSize = 2;
+                    litSize = MEM_readLE16(istart) >> 4;
+                    break;
+                case 3:
+                    lhSize = 3;
+                    litSize = MEM_readLE24(istart) >> 4;
+                    break;
+                }
+
+                if (lhSize+litSize+WILDCOPY_OVERLENGTH > srcSize) {  /* risk reading beyond src buffer with wildcopy */
+                    RETURN_ERROR_IF(litSize+lhSize > srcSize, corruption_detected, "");
+                    ZSTD_memcpy(dctx->litBuffer, istart+lhSize, litSize);
+                    dctx->litPtr = dctx->litBuffer;
+                    dctx->litSize = litSize;
+                    ZSTD_memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);
+                    return lhSize+litSize;
+                }
+                /* direct reference into compressed stream */
+                dctx->litPtr = istart+lhSize;
+                dctx->litSize = litSize;
+                return lhSize+litSize;
+            }
+
+        case set_rle:
+            {   U32 const lhlCode = ((istart[0]) >> 2) & 3;
+                size_t litSize, lhSize;
+                switch(lhlCode)
+                {
+                case 0: case 2: default:   /* note : default is impossible, since lhlCode into [0..3] */
+                    lhSize = 1;
+                    litSize = istart[0] >> 3;
+                    break;
+                case 1:
+                    lhSize = 2;
+                    litSize = MEM_readLE16(istart) >> 4;
+                    break;
+                case 3:
+                    lhSize = 3;
+                    litSize = MEM_readLE24(istart) >> 4;
+                    RETURN_ERROR_IF(srcSize<4, corruption_detected, "srcSize >= MIN_CBLOCK_SIZE == 3; here we need lhSize+1 = 4");
+                    break;
+                }
+                RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected, "");
+                ZSTD_memset(dctx->litBuffer, istart[lhSize], litSize + WILDCOPY_OVERLENGTH);
+                dctx->litPtr = dctx->litBuffer;
+                dctx->litSize = litSize;
+                return lhSize+1;
+            }
+        default:
+            RETURN_ERROR(corruption_detected, "impossible");
+        }
+    }
+}
+
+/* Default FSE distribution tables.
+ * These are pre-calculated FSE decoding tables using default distributions as defined in specification :
+ * https://github.com/facebook/zstd/blob/release/doc/zstd_compression_format.md#default-distributions
+ * They were generated programmatically with following method :
+ * - start from default distributions, present in /lib/common/zstd_internal.h
+ * - generate tables normally, using ZSTD_buildFSETable()
+ * - printout the content of tables
+ * - pretify output, report below, test with fuzzer to ensure it's correct */
+
+/* Default FSE distribution table for Literal Lengths */
+static const ZSTD_seqSymbol LL_defaultDTable[(1<<LL_DEFAULTNORMLOG)+1] = {
+     {  1,  1,  1, LL_DEFAULTNORMLOG},  /* header : fastMode, tableLog */
+     /* nextState, nbAddBits, nbBits, baseVal */
+     {  0,  0,  4,    0},  { 16,  0,  4,    0},
+     { 32,  0,  5,    1},  {  0,  0,  5,    3},
+     {  0,  0,  5,    4},  {  0,  0,  5,    6},
+     {  0,  0,  5,    7},  {  0,  0,  5,    9},
+     {  0,  0,  5,   10},  {  0,  0,  5,   12},
+     {  0,  0,  6,   14},  {  0,  1,  5,   16},
+     {  0,  1,  5,   20},  {  0,  1,  5,   22},
+     {  0,  2,  5,   28},  {  0,  3,  5,   32},
+     {  0,  4,  5,   48},  { 32,  6,  5,   64},
+     {  0,  7,  5,  128},  {  0,  8,  6,  256},
+     {  0, 10,  6, 1024},  {  0, 12,  6, 4096},
+     { 32,  0,  4,    0},  {  0,  0,  4,    1},
+     {  0,  0,  5,    2},  { 32,  0,  5,    4},
+     {  0,  0,  5,    5},  { 32,  0,  5,    7},
+     {  0,  0,  5,    8},  { 32,  0,  5,   10},
+     {  0,  0,  5,   11},  {  0,  0,  6,   13},
+     { 32,  1,  5,   16},  {  0,  1,  5,   18},
+     { 32,  1,  5,   22},  {  0,  2,  5,   24},
+     { 32,  3,  5,   32},  {  0,  3,  5,   40},
+     {  0,  6,  4,   64},  { 16,  6,  4,   64},
+     { 32,  7,  5,  128},  {  0,  9,  6,  512},
+     {  0, 11,  6, 2048},  { 48,  0,  4,    0},
+     { 16,  0,  4,    1},  { 32,  0,  5,    2},
+     { 32,  0,  5,    3},  { 32,  0,  5,    5},
+     { 32,  0,  5,    6},  { 32,  0,  5,    8},
+     { 32,  0,  5,    9},  { 32,  0,  5,   11},
+     { 32,  0,  5,   12},  {  0,  0,  6,   15},
+     { 32,  1,  5,   18},  { 32,  1,  5,   20},
+     { 32,  2,  5,   24},  { 32,  2,  5,   28},
+     { 32,  3,  5,   40},  { 32,  4,  5,   48},
+     {  0, 16,  6,65536},  {  0, 15,  6,32768},
+     {  0, 14,  6,16384},  {  0, 13,  6, 8192},
+};   /* LL_defaultDTable */
+
+/* Default FSE distribution table for Offset Codes */
+static const ZSTD_seqSymbol OF_defaultDTable[(1<<OF_DEFAULTNORMLOG)+1] = {
+    {  1,  1,  1, OF_DEFAULTNORMLOG},  /* header : fastMode, tableLog */
+    /* nextState, nbAddBits, nbBits, baseVal */
+    {  0,  0,  5,    0},     {  0,  6,  4,   61},
+    {  0,  9,  5,  509},     {  0, 15,  5,32765},
+    {  0, 21,  5,2097149},   {  0,  3,  5,    5},
+    {  0,  7,  4,  125},     {  0, 12,  5, 4093},
+    {  0, 18,  5,262141},    {  0, 23,  5,8388605},
+    {  0,  5,  5,   29},     {  0,  8,  4,  253},
+    {  0, 14,  5,16381},     {  0, 20,  5,1048573},
+    {  0,  2,  5,    1},     { 16,  7,  4,  125},
+    {  0, 11,  5, 2045},     {  0, 17,  5,131069},
+    {  0, 22,  5,4194301},   {  0,  4,  5,   13},
+    { 16,  8,  4,  253},     {  0, 13,  5, 8189},
+    {  0, 19,  5,524285},    {  0,  1,  5,    1},
+    { 16,  6,  4,   61},     {  0, 10,  5, 1021},
+    {  0, 16,  5,65533},     {  0, 28,  5,268435453},
+    {  0, 27,  5,134217725}, {  0, 26,  5,67108861},
+    {  0, 25,  5,33554429},  {  0, 24,  5,16777213},
+};   /* OF_defaultDTable */
+
+
+/* Default FSE distribution table for Match Lengths */
+static const ZSTD_seqSymbol ML_defaultDTable[(1<<ML_DEFAULTNORMLOG)+1] = {
+    {  1,  1,  1, ML_DEFAULTNORMLOG},  /* header : fastMode, tableLog */
+    /* nextState, nbAddBits, nbBits, baseVal */
+    {  0,  0,  6,    3},  {  0,  0,  4,    4},
+    { 32,  0,  5,    5},  {  0,  0,  5,    6},
+    {  0,  0,  5,    8},  {  0,  0,  5,    9},
+    {  0,  0,  5,   11},  {  0,  0,  6,   13},
+    {  0,  0,  6,   16},  {  0,  0,  6,   19},
+    {  0,  0,  6,   22},  {  0,  0,  6,   25},
+    {  0,  0,  6,   28},  {  0,  0,  6,   31},
+    {  0,  0,  6,   34},  {  0,  1,  6,   37},
+    {  0,  1,  6,   41},  {  0,  2,  6,   47},
+    {  0,  3,  6,   59},  {  0,  4,  6,   83},
+    {  0,  7,  6,  131},  {  0,  9,  6,  515},
+    { 16,  0,  4,    4},  {  0,  0,  4,    5},
+    { 32,  0,  5,    6},  {  0,  0,  5,    7},
+    { 32,  0,  5,    9},  {  0,  0,  5,   10},
+    {  0,  0,  6,   12},  {  0,  0,  6,   15},
+    {  0,  0,  6,   18},  {  0,  0,  6,   21},
+    {  0,  0,  6,   24},  {  0,  0,  6,   27},
+    {  0,  0,  6,   30},  {  0,  0,  6,   33},
+    {  0,  1,  6,   35},  {  0,  1,  6,   39},
+    {  0,  2,  6,   43},  {  0,  3,  6,   51},
+    {  0,  4,  6,   67},  {  0,  5,  6,   99},
+    {  0,  8,  6,  259},  { 32,  0,  4,    4},
+    { 48,  0,  4,    4},  { 16,  0,  4,    5},
+    { 32,  0,  5,    7},  { 32,  0,  5,    8},
+    { 32,  0,  5,   10},  { 32,  0,  5,   11},
+    {  0,  0,  6,   14},  {  0,  0,  6,   17},
+    {  0,  0,  6,   20},  {  0,  0,  6,   23},
+    {  0,  0,  6,   26},  {  0,  0,  6,   29},
+    {  0,  0,  6,   32},  {  0, 16,  6,65539},
+    {  0, 15,  6,32771},  {  0, 14,  6,16387},
+    {  0, 13,  6, 8195},  {  0, 12,  6, 4099},
+    {  0, 11,  6, 2051},  {  0, 10,  6, 1027},
+};   /* ML_defaultDTable */
+
+
+static void ZSTD_buildSeqTable_rle(ZSTD_seqSymbol* dt, U32 baseValue, U32 nbAddBits)
+{
+    void* ptr = dt;
+    ZSTD_seqSymbol_header* const DTableH = (ZSTD_seqSymbol_header*)ptr;
+    ZSTD_seqSymbol* const cell = dt + 1;
+
+    DTableH->tableLog = 0;
+    DTableH->fastMode = 0;
+
+    cell->nbBits = 0;
+    cell->nextState = 0;
+    assert(nbAddBits < 255);
+    cell->nbAdditionalBits = (BYTE)nbAddBits;
+    cell->baseValue = baseValue;
+}
+
+
+/* ZSTD_buildFSETable() :
+ * generate FSE decoding table for one symbol (ll, ml or off)
+ * cannot fail if input is valid =>
+ * all inputs are presumed validated at this stage */
+FORCE_INLINE_TEMPLATE
+void ZSTD_buildFSETable_body(ZSTD_seqSymbol* dt,
+            const short* normalizedCounter, unsigned maxSymbolValue,
+            const U32* baseValue, const U32* nbAdditionalBits,
+            unsigned tableLog, void* wksp, size_t wkspSize)
+{
+    ZSTD_seqSymbol* const tableDecode = dt+1;
+    U32 const maxSV1 = maxSymbolValue + 1;
+    U32 const tableSize = 1 << tableLog;
+
+    U16* symbolNext = (U16*)wksp;
+    BYTE* spread = (BYTE*)(symbolNext + MaxSeq + 1);
+    U32 highThreshold = tableSize - 1;
+
+
+    /* Sanity Checks */
+    assert(maxSymbolValue <= MaxSeq);
+    assert(tableLog <= MaxFSELog);
+    assert(wkspSize >= ZSTD_BUILD_FSE_TABLE_WKSP_SIZE);
+    (void)wkspSize;
+    /* Init, lay down lowprob symbols */
+    {   ZSTD_seqSymbol_header DTableH;
+        DTableH.tableLog = tableLog;
+        DTableH.fastMode = 1;
+        {   S16 const largeLimit= (S16)(1 << (tableLog-1));
+            U32 s;
+            for (s=0; s<maxSV1; s++) {
+                if (normalizedCounter[s]==-1) {
+                    tableDecode[highThreshold--].baseValue = s;
+                    symbolNext[s] = 1;
+                } else {
+                    if (normalizedCounter[s] >= largeLimit) DTableH.fastMode=0;
+                    assert(normalizedCounter[s]>=0);
+                    symbolNext[s] = (U16)normalizedCounter[s];
+        }   }   }
+        ZSTD_memcpy(dt, &DTableH, sizeof(DTableH));
+    }
+
+    /* Spread symbols */
+    assert(tableSize <= 512);
+    /* Specialized symbol spreading for the case when there are
+     * no low probability (-1 count) symbols. When compressing
+     * small blocks we avoid low probability symbols to hit this
+     * case, since header decoding speed matters more.
+     */
+    if (highThreshold == tableSize - 1) {
+        size_t const tableMask = tableSize-1;
+        size_t const step = FSE_TABLESTEP(tableSize);
+        /* First lay down the symbols in order.
+         * We use a uint64_t to lay down 8 bytes at a time. This reduces branch
+         * misses since small blocks generally have small table logs, so nearly
+         * all symbols have counts <= 8. We ensure we have 8 bytes at the end of
+         * our buffer to handle the over-write.
+         */
+        {
+            U64 const add = 0x0101010101010101ull;
+            size_t pos = 0;
+            U64 sv = 0;
+            U32 s;
+            for (s=0; s<maxSV1; ++s, sv += add) {
+                int i;
+                int const n = normalizedCounter[s];
+                MEM_write64(spread + pos, sv);
+                for (i = 8; i < n; i += 8) {
+                    MEM_write64(spread + pos + i, sv);
+                }
+                pos += n;
+            }
+        }
+        /* Now we spread those positions across the table.
+         * The benefit of doing it in two stages is that we avoid the the
+         * variable size inner loop, which caused lots of branch misses.
+         * Now we can run through all the positions without any branch misses.
+         * We unroll the loop twice, since that is what emperically worked best.
+         */
+        {
+            size_t position = 0;
+            size_t s;
+            size_t const unroll = 2;
+            assert(tableSize % unroll == 0); /* FSE_MIN_TABLELOG is 5 */
+            for (s = 0; s < (size_t)tableSize; s += unroll) {
+                size_t u;
+                for (u = 0; u < unroll; ++u) {
+                    size_t const uPosition = (position + (u * step)) & tableMask;
+                    tableDecode[uPosition].baseValue = spread[s + u];
+                }
+                position = (position + (unroll * step)) & tableMask;
+            }
+            assert(position == 0);
+        }
+    } else {
+        U32 const tableMask = tableSize-1;
+        U32 const step = FSE_TABLESTEP(tableSize);
+        U32 s, position = 0;
+        for (s=0; s<maxSV1; s++) {
+            int i;
+            int const n = normalizedCounter[s];
+            for (i=0; i<n; i++) {
+                tableDecode[position].baseValue = s;
+                position = (position + step) & tableMask;
+                while (position > highThreshold) position = (position + step) & tableMask;   /* lowprob area */
+        }   }
+        assert(position == 0); /* position must reach all cells once, otherwise normalizedCounter is incorrect */
+    }
+
+    /* Build Decoding table */
+    {
+        U32 u;
+        for (u=0; u<tableSize; u++) {
+            U32 const symbol = tableDecode[u].baseValue;
+            U32 const nextState = symbolNext[symbol]++;
+            tableDecode[u].nbBits = (BYTE) (tableLog - BIT_highbit32(nextState) );
+            tableDecode[u].nextState = (U16) ( (nextState << tableDecode[u].nbBits) - tableSize);
+            assert(nbAdditionalBits[symbol] < 255);
+            tableDecode[u].nbAdditionalBits = (BYTE)nbAdditionalBits[symbol];
+            tableDecode[u].baseValue = baseValue[symbol];
+        }
+    }
+}
+
+/* Avoids the FORCE_INLINE of the _body() function. */
+static void ZSTD_buildFSETable_body_default(ZSTD_seqSymbol* dt,
+            const short* normalizedCounter, unsigned maxSymbolValue,
+            const U32* baseValue, const U32* nbAdditionalBits,
+            unsigned tableLog, void* wksp, size_t wkspSize)
+{
+    ZSTD_buildFSETable_body(dt, normalizedCounter, maxSymbolValue,
+            baseValue, nbAdditionalBits, tableLog, wksp, wkspSize);
+}
+
+#if DYNAMIC_BMI2
+TARGET_ATTRIBUTE("bmi2") static void ZSTD_buildFSETable_body_bmi2(ZSTD_seqSymbol* dt,
+            const short* normalizedCounter, unsigned maxSymbolValue,
+            const U32* baseValue, const U32* nbAdditionalBits,
+            unsigned tableLog, void* wksp, size_t wkspSize)
+{
+    ZSTD_buildFSETable_body(dt, normalizedCounter, maxSymbolValue,
+            baseValue, nbAdditionalBits, tableLog, wksp, wkspSize);
+}
+#endif
+
+void ZSTD_buildFSETable(ZSTD_seqSymbol* dt,
+            const short* normalizedCounter, unsigned maxSymbolValue,
+            const U32* baseValue, const U32* nbAdditionalBits,
+            unsigned tableLog, void* wksp, size_t wkspSize, int bmi2)
+{
+#if DYNAMIC_BMI2
+    if (bmi2) {
+        ZSTD_buildFSETable_body_bmi2(dt, normalizedCounter, maxSymbolValue,
+                baseValue, nbAdditionalBits, tableLog, wksp, wkspSize);
+        return;
+    }
+#endif
+    (void)bmi2;
+    ZSTD_buildFSETable_body_default(dt, normalizedCounter, maxSymbolValue,
+            baseValue, nbAdditionalBits, tableLog, wksp, wkspSize);
+}
+
+
+/*! ZSTD_buildSeqTable() :
+ * @return : nb bytes read from src,
+ *           or an error code if it fails */
+static size_t ZSTD_buildSeqTable(ZSTD_seqSymbol* DTableSpace, const ZSTD_seqSymbol** DTablePtr,
+                                 symbolEncodingType_e type, unsigned max, U32 maxLog,
+                                 const void* src, size_t srcSize,
+                                 const U32* baseValue, const U32* nbAdditionalBits,
+                                 const ZSTD_seqSymbol* defaultTable, U32 flagRepeatTable,
+                                 int ddictIsCold, int nbSeq, U32* wksp, size_t wkspSize,
+                                 int bmi2)
+{
+    switch(type)
+    {
+    case set_rle :
+        RETURN_ERROR_IF(!srcSize, srcSize_wrong, "");
+        RETURN_ERROR_IF((*(const BYTE*)src) > max, corruption_detected, "");
+        {   U32 const symbol = *(const BYTE*)src;
+            U32 const baseline = baseValue[symbol];
+            U32 const nbBits = nbAdditionalBits[symbol];
+            ZSTD_buildSeqTable_rle(DTableSpace, baseline, nbBits);
+        }
+        *DTablePtr = DTableSpace;
+        return 1;
+    case set_basic :
+        *DTablePtr = defaultTable;
+        return 0;
+    case set_repeat:
+        RETURN_ERROR_IF(!flagRepeatTable, corruption_detected, "");
+        /* prefetch FSE table if used */
+        if (ddictIsCold && (nbSeq > 24 /* heuristic */)) {
+            const void* const pStart = *DTablePtr;
+            size_t const pSize = sizeof(ZSTD_seqSymbol) * (SEQSYMBOL_TABLE_SIZE(maxLog));
+            PREFETCH_AREA(pStart, pSize);
+        }
+        return 0;
+    case set_compressed :
+        {   unsigned tableLog;
+            S16 norm[MaxSeq+1];
+            size_t const headerSize = FSE_readNCount(norm, &max, &tableLog, src, srcSize);
+            RETURN_ERROR_IF(FSE_isError(headerSize), corruption_detected, "");
+            RETURN_ERROR_IF(tableLog > maxLog, corruption_detected, "");
+            ZSTD_buildFSETable(DTableSpace, norm, max, baseValue, nbAdditionalBits, tableLog, wksp, wkspSize, bmi2);
+            *DTablePtr = DTableSpace;
+            return headerSize;
+        }
+    default :
+        assert(0);
+        RETURN_ERROR(GENERIC, "impossible");
+    }
+}
+
+size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,
+                             const void* src, size_t srcSize)
+{
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* ip = istart;
+    int nbSeq;
+    DEBUGLOG(5, "ZSTD_decodeSeqHeaders");
+
+    /* check */
+    RETURN_ERROR_IF(srcSize < MIN_SEQUENCES_SIZE, srcSize_wrong, "");
+
+    /* SeqHead */
+    nbSeq = *ip++;
+    if (!nbSeq) {
+        *nbSeqPtr=0;
+        RETURN_ERROR_IF(srcSize != 1, srcSize_wrong, "");
+        return 1;
+    }
+    if (nbSeq > 0x7F) {
+        if (nbSeq == 0xFF) {
+            RETURN_ERROR_IF(ip+2 > iend, srcSize_wrong, "");
+            nbSeq = MEM_readLE16(ip) + LONGNBSEQ;
+            ip+=2;
+        } else {
+            RETURN_ERROR_IF(ip >= iend, srcSize_wrong, "");
+            nbSeq = ((nbSeq-0x80)<<8) + *ip++;
+        }
+    }
+    *nbSeqPtr = nbSeq;
+
+    /* FSE table descriptors */
+    RETURN_ERROR_IF(ip+1 > iend, srcSize_wrong, ""); /* minimum possible size: 1 byte for symbol encoding types */
+    {   symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6);
+        symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3);
+        symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3);
+        ip++;
+
+        /* Build DTables */
+        {   size_t const llhSize = ZSTD_buildSeqTable(dctx->entropy.LLTable, &dctx->LLTptr,
+                                                      LLtype, MaxLL, LLFSELog,
+                                                      ip, iend-ip,
+                                                      LL_base, LL_bits,
+                                                      LL_defaultDTable, dctx->fseEntropy,
+                                                      dctx->ddictIsCold, nbSeq,
+                                                      dctx->workspace, sizeof(dctx->workspace),
+                                                      dctx->bmi2);
+            RETURN_ERROR_IF(ZSTD_isError(llhSize), corruption_detected, "ZSTD_buildSeqTable failed");
+            ip += llhSize;
+        }
+
+        {   size_t const ofhSize = ZSTD_buildSeqTable(dctx->entropy.OFTable, &dctx->OFTptr,
+                                                      OFtype, MaxOff, OffFSELog,
+                                                      ip, iend-ip,
+                                                      OF_base, OF_bits,
+                                                      OF_defaultDTable, dctx->fseEntropy,
+                                                      dctx->ddictIsCold, nbSeq,
+                                                      dctx->workspace, sizeof(dctx->workspace),
+                                                      dctx->bmi2);
+            RETURN_ERROR_IF(ZSTD_isError(ofhSize), corruption_detected, "ZSTD_buildSeqTable failed");
+            ip += ofhSize;
+        }
+
+        {   size_t const mlhSize = ZSTD_buildSeqTable(dctx->entropy.MLTable, &dctx->MLTptr,
+                                                      MLtype, MaxML, MLFSELog,
+                                                      ip, iend-ip,
+                                                      ML_base, ML_bits,
+                                                      ML_defaultDTable, dctx->fseEntropy,
+                                                      dctx->ddictIsCold, nbSeq,
+                                                      dctx->workspace, sizeof(dctx->workspace),
+                                                      dctx->bmi2);
+            RETURN_ERROR_IF(ZSTD_isError(mlhSize), corruption_detected, "ZSTD_buildSeqTable failed");
+            ip += mlhSize;
+        }
+    }
+
+    return ip-istart;
+}
+
+
+typedef struct {
+    size_t litLength;
+    size_t matchLength;
+    size_t offset;
+    const BYTE* match;
+} seq_t;
+
+typedef struct {
+    size_t state;
+    const ZSTD_seqSymbol* table;
+} ZSTD_fseState;
+
+typedef struct {
+    BIT_DStream_t DStream;
+    ZSTD_fseState stateLL;
+    ZSTD_fseState stateOffb;
+    ZSTD_fseState stateML;
+    size_t prevOffset[ZSTD_REP_NUM];
+    const BYTE* prefixStart;
+    const BYTE* dictEnd;
+    size_t pos;
+} seqState_t;
+
+/*! ZSTD_overlapCopy8() :
+ *  Copies 8 bytes from ip to op and updates op and ip where ip <= op.
+ *  If the offset is < 8 then the offset is spread to at least 8 bytes.
+ *
+ *  Precondition: *ip <= *op
+ *  Postcondition: *op - *op >= 8
+ */
+HINT_INLINE void ZSTD_overlapCopy8(BYTE** op, BYTE const** ip, size_t offset) {
+    assert(*ip <= *op);
+    if (offset < 8) {
+        /* close range match, overlap */
+        static const U32 dec32table[] = { 0, 1, 2, 1, 4, 4, 4, 4 };   /* added */
+        static const int dec64table[] = { 8, 8, 8, 7, 8, 9,10,11 };   /* subtracted */
+        int const sub2 = dec64table[offset];
+        (*op)[0] = (*ip)[0];
+        (*op)[1] = (*ip)[1];
+        (*op)[2] = (*ip)[2];
+        (*op)[3] = (*ip)[3];
+        *ip += dec32table[offset];
+        ZSTD_copy4(*op+4, *ip);
+        *ip -= sub2;
+    } else {
+        ZSTD_copy8(*op, *ip);
+    }
+    *ip += 8;
+    *op += 8;
+    assert(*op - *ip >= 8);
+}
+
+/*! ZSTD_safecopy() :
+ *  Specialized version of memcpy() that is allowed to READ up to WILDCOPY_OVERLENGTH past the input buffer
+ *  and write up to 16 bytes past oend_w (op >= oend_w is allowed).
+ *  This function is only called in the uncommon case where the sequence is near the end of the block. It
+ *  should be fast for a single long sequence, but can be slow for several short sequences.
+ *
+ *  @param ovtype controls the overlap detection
+ *         - ZSTD_no_overlap: The source and destination are guaranteed to be at least WILDCOPY_VECLEN bytes apart.
+ *         - ZSTD_overlap_src_before_dst: The src and dst may overlap and may be any distance apart.
+ *           The src buffer must be before the dst buffer.
+ */
+static void ZSTD_safecopy(BYTE* op, BYTE* const oend_w, BYTE const* ip, ptrdiff_t length, ZSTD_overlap_e ovtype) {
+    ptrdiff_t const diff = op - ip;
+    BYTE* const oend = op + length;
+
+    assert((ovtype == ZSTD_no_overlap && (diff <= -8 || diff >= 8 || op >= oend_w)) ||
+           (ovtype == ZSTD_overlap_src_before_dst && diff >= 0));
+
+    if (length < 8) {
+        /* Handle short lengths. */
+        while (op < oend) *op++ = *ip++;
+        return;
+    }
+    if (ovtype == ZSTD_overlap_src_before_dst) {
+        /* Copy 8 bytes and ensure the offset >= 8 when there can be overlap. */
+        assert(length >= 8);
+        ZSTD_overlapCopy8(&op, &ip, diff);
+        assert(op - ip >= 8);
+        assert(op <= oend);
+    }
+
+    if (oend <= oend_w) {
+        /* No risk of overwrite. */
+        ZSTD_wildcopy(op, ip, length, ovtype);
+        return;
+    }
+    if (op <= oend_w) {
+        /* Wildcopy until we get close to the end. */
+        assert(oend > oend_w);
+        ZSTD_wildcopy(op, ip, oend_w - op, ovtype);
+        ip += oend_w - op;
+        op = oend_w;
+    }
+    /* Handle the leftovers. */
+    while (op < oend) *op++ = *ip++;
+}
+
+/* ZSTD_execSequenceEnd():
+ * This version handles cases that are near the end of the output buffer. It requires
+ * more careful checks to make sure there is no overflow. By separating out these hard
+ * and unlikely cases, we can speed up the common cases.
+ *
+ * NOTE: This function needs to be fast for a single long sequence, but doesn't need
+ * to be optimized for many small sequences, since those fall into ZSTD_execSequence().
+ */
+FORCE_NOINLINE
+size_t ZSTD_execSequenceEnd(BYTE* op,
+                            BYTE* const oend, seq_t sequence,
+                            const BYTE** litPtr, const BYTE* const litLimit,
+                            const BYTE* const prefixStart, const BYTE* const virtualStart, const BYTE* const dictEnd)
+{
+    BYTE* const oLitEnd = op + sequence.litLength;
+    size_t const sequenceLength = sequence.litLength + sequence.matchLength;
+    const BYTE* const iLitEnd = *litPtr + sequence.litLength;
+    const BYTE* match = oLitEnd - sequence.offset;
+    BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH;
+
+    /* bounds checks : careful of address space overflow in 32-bit mode */
+    RETURN_ERROR_IF(sequenceLength > (size_t)(oend - op), dstSize_tooSmall, "last match must fit within dstBuffer");
+    RETURN_ERROR_IF(sequence.litLength > (size_t)(litLimit - *litPtr), corruption_detected, "try to read beyond literal buffer");
+    assert(op < op + sequenceLength);
+    assert(oLitEnd < op + sequenceLength);
+
+    /* copy literals */
+    ZSTD_safecopy(op, oend_w, *litPtr, sequence.litLength, ZSTD_no_overlap);
+    op = oLitEnd;
+    *litPtr = iLitEnd;
+
+    /* copy Match */
+    if (sequence.offset > (size_t)(oLitEnd - prefixStart)) {
+        /* offset beyond prefix */
+        RETURN_ERROR_IF(sequence.offset > (size_t)(oLitEnd - virtualStart), corruption_detected, "");
+        match = dictEnd - (prefixStart-match);
+        if (match + sequence.matchLength <= dictEnd) {
+            ZSTD_memmove(oLitEnd, match, sequence.matchLength);
+            return sequenceLength;
+        }
+        /* span extDict & currentPrefixSegment */
+        {   size_t const length1 = dictEnd - match;
+            ZSTD_memmove(oLitEnd, match, length1);
+            op = oLitEnd + length1;
+            sequence.matchLength -= length1;
+            match = prefixStart;
+    }   }
+    ZSTD_safecopy(op, oend_w, match, sequence.matchLength, ZSTD_overlap_src_before_dst);
+    return sequenceLength;
+}
+
+HINT_INLINE
+size_t ZSTD_execSequence(BYTE* op,
+                         BYTE* const oend, seq_t sequence,
+                         const BYTE** litPtr, const BYTE* const litLimit,
+                         const BYTE* const prefixStart, const BYTE* const virtualStart, const BYTE* const dictEnd)
+{
+    BYTE* const oLitEnd = op + sequence.litLength;
+    size_t const sequenceLength = sequence.litLength + sequence.matchLength;
+    BYTE* const oMatchEnd = op + sequenceLength;   /* risk : address space overflow (32-bits) */
+    BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH;   /* risk : address space underflow on oend=NULL */
+    const BYTE* const iLitEnd = *litPtr + sequence.litLength;
+    const BYTE* match = oLitEnd - sequence.offset;
+
+    assert(op != NULL /* Precondition */);
+    assert(oend_w < oend /* No underflow */);
+    /* Handle edge cases in a slow path:
+     *   - Read beyond end of literals
+     *   - Match end is within WILDCOPY_OVERLIMIT of oend
+     *   - 32-bit mode and the match length overflows
+     */
+    if (UNLIKELY(
+            iLitEnd > litLimit ||
+            oMatchEnd > oend_w ||
+            (MEM_32bits() && (size_t)(oend - op) < sequenceLength + WILDCOPY_OVERLENGTH)))
+        return ZSTD_execSequenceEnd(op, oend, sequence, litPtr, litLimit, prefixStart, virtualStart, dictEnd);
+
+    /* Assumptions (everything else goes into ZSTD_execSequenceEnd()) */
+    assert(op <= oLitEnd /* No overflow */);
+    assert(oLitEnd < oMatchEnd /* Non-zero match & no overflow */);
+    assert(oMatchEnd <= oend /* No underflow */);
+    assert(iLitEnd <= litLimit /* Literal length is in bounds */);
+    assert(oLitEnd <= oend_w /* Can wildcopy literals */);
+    assert(oMatchEnd <= oend_w /* Can wildcopy matches */);
+
+    /* Copy Literals:
+     * Split out litLength <= 16 since it is nearly always true. +1.6% on gcc-9.
+     * We likely don't need the full 32-byte wildcopy.
+     */
+    assert(WILDCOPY_OVERLENGTH >= 16);
+    ZSTD_copy16(op, (*litPtr));
+    if (UNLIKELY(sequence.litLength > 16)) {
+        ZSTD_wildcopy(op+16, (*litPtr)+16, sequence.litLength-16, ZSTD_no_overlap);
+    }
+    op = oLitEnd;
+    *litPtr = iLitEnd;   /* update for next sequence */
+
+    /* Copy Match */
+    if (sequence.offset > (size_t)(oLitEnd - prefixStart)) {
+        /* offset beyond prefix -> go into extDict */
+        RETURN_ERROR_IF(UNLIKELY(sequence.offset > (size_t)(oLitEnd - virtualStart)), corruption_detected, "");
+        match = dictEnd + (match - prefixStart);
+        if (match + sequence.matchLength <= dictEnd) {
+            ZSTD_memmove(oLitEnd, match, sequence.matchLength);
+            return sequenceLength;
+        }
+        /* span extDict & currentPrefixSegment */
+        {   size_t const length1 = dictEnd - match;
+            ZSTD_memmove(oLitEnd, match, length1);
+            op = oLitEnd + length1;
+            sequence.matchLength -= length1;
+            match = prefixStart;
+    }   }
+    /* Match within prefix of 1 or more bytes */
+    assert(op <= oMatchEnd);
+    assert(oMatchEnd <= oend_w);
+    assert(match >= prefixStart);
+    assert(sequence.matchLength >= 1);
+
+    /* Nearly all offsets are >= WILDCOPY_VECLEN bytes, which means we can use wildcopy
+     * without overlap checking.
+     */
+    if (LIKELY(sequence.offset >= WILDCOPY_VECLEN)) {
+        /* We bet on a full wildcopy for matches, since we expect matches to be
+         * longer than literals (in general). In silesia, ~10% of matches are longer
+         * than 16 bytes.
+         */
+        ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength, ZSTD_no_overlap);
+        return sequenceLength;
+    }
+    assert(sequence.offset < WILDCOPY_VECLEN);
+
+    /* Copy 8 bytes and spread the offset to be >= 8. */
+    ZSTD_overlapCopy8(&op, &match, sequence.offset);
+
+    /* If the match length is > 8 bytes, then continue with the wildcopy. */
+    if (sequence.matchLength > 8) {
+        assert(op < oMatchEnd);
+        ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8, ZSTD_overlap_src_before_dst);
+    }
+    return sequenceLength;
+}
+
+static void
+ZSTD_initFseState(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD, const ZSTD_seqSymbol* dt)
+{
+    const void* ptr = dt;
+    const ZSTD_seqSymbol_header* const DTableH = (const ZSTD_seqSymbol_header*)ptr;
+    DStatePtr->state = BIT_readBits(bitD, DTableH->tableLog);
+    DEBUGLOG(6, "ZSTD_initFseState : val=%u using %u bits",
+                (U32)DStatePtr->state, DTableH->tableLog);
+    BIT_reloadDStream(bitD);
+    DStatePtr->table = dt + 1;
+}
+
+FORCE_INLINE_TEMPLATE void
+ZSTD_updateFseState(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD)
+{
+    ZSTD_seqSymbol const DInfo = DStatePtr->table[DStatePtr->state];
+    U32 const nbBits = DInfo.nbBits;
+    size_t const lowBits = BIT_readBits(bitD, nbBits);
+    DStatePtr->state = DInfo.nextState + lowBits;
+}
+
+FORCE_INLINE_TEMPLATE void
+ZSTD_updateFseStateWithDInfo(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD, ZSTD_seqSymbol const DInfo)
+{
+    U32 const nbBits = DInfo.nbBits;
+    size_t const lowBits = BIT_readBits(bitD, nbBits);
+    DStatePtr->state = DInfo.nextState + lowBits;
+}
+
+/* We need to add at most (ZSTD_WINDOWLOG_MAX_32 - 1) bits to read the maximum
+ * offset bits. But we can only read at most (STREAM_ACCUMULATOR_MIN_32 - 1)
+ * bits before reloading. This value is the maximum number of bytes we read
+ * after reloading when we are decoding long offsets.
+ */
+#define LONG_OFFSETS_MAX_EXTRA_BITS_32                       \
+    (ZSTD_WINDOWLOG_MAX_32 > STREAM_ACCUMULATOR_MIN_32       \
+        ? ZSTD_WINDOWLOG_MAX_32 - STREAM_ACCUMULATOR_MIN_32  \
+        : 0)
+
+typedef enum { ZSTD_lo_isRegularOffset, ZSTD_lo_isLongOffset=1 } ZSTD_longOffset_e;
+typedef enum { ZSTD_p_noPrefetch=0, ZSTD_p_prefetch=1 } ZSTD_prefetch_e;
+
+FORCE_INLINE_TEMPLATE seq_t
+ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets, const ZSTD_prefetch_e prefetch)
+{
+    seq_t seq;
+    ZSTD_seqSymbol const llDInfo = seqState->stateLL.table[seqState->stateLL.state];
+    ZSTD_seqSymbol const mlDInfo = seqState->stateML.table[seqState->stateML.state];
+    ZSTD_seqSymbol const ofDInfo = seqState->stateOffb.table[seqState->stateOffb.state];
+    U32 const llBase = llDInfo.baseValue;
+    U32 const mlBase = mlDInfo.baseValue;
+    U32 const ofBase = ofDInfo.baseValue;
+    BYTE const llBits = llDInfo.nbAdditionalBits;
+    BYTE const mlBits = mlDInfo.nbAdditionalBits;
+    BYTE const ofBits = ofDInfo.nbAdditionalBits;
+    BYTE const totalBits = llBits+mlBits+ofBits;
+
+    /* sequence */
+    {   size_t offset;
+        if (ofBits > 1) {
+            ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);
+            ZSTD_STATIC_ASSERT(LONG_OFFSETS_MAX_EXTRA_BITS_32 == 5);
+            assert(ofBits <= MaxOff);
+            if (MEM_32bits() && longOffsets && (ofBits >= STREAM_ACCUMULATOR_MIN_32)) {
+                U32 const extraBits = ofBits - MIN(ofBits, 32 - seqState->DStream.bitsConsumed);
+                offset = ofBase + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);
+                BIT_reloadDStream(&seqState->DStream);
+                if (extraBits) offset += BIT_readBitsFast(&seqState->DStream, extraBits);
+                assert(extraBits <= LONG_OFFSETS_MAX_EXTRA_BITS_32);   /* to avoid another reload */
+            } else {
+                offset = ofBase + BIT_readBitsFast(&seqState->DStream, ofBits/*>0*/);   /* <=  (ZSTD_WINDOWLOG_MAX-1) bits */
+                if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);
+            }
+            seqState->prevOffset[2] = seqState->prevOffset[1];
+            seqState->prevOffset[1] = seqState->prevOffset[0];
+            seqState->prevOffset[0] = offset;
+        } else {
+            U32 const ll0 = (llBase == 0);
+            if (LIKELY((ofBits == 0))) {
+                if (LIKELY(!ll0))
+                    offset = seqState->prevOffset[0];
+                else {
+                    offset = seqState->prevOffset[1];
+                    seqState->prevOffset[1] = seqState->prevOffset[0];
+                    seqState->prevOffset[0] = offset;
+                }
+            } else {
+                offset = ofBase + ll0 + BIT_readBitsFast(&seqState->DStream, 1);
+                {   size_t temp = (offset==3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];
+                    temp += !temp;   /* 0 is not valid; input is corrupted; force offset to 1 */
+                    if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];
+                    seqState->prevOffset[1] = seqState->prevOffset[0];
+                    seqState->prevOffset[0] = offset = temp;
+        }   }   }
+        seq.offset = offset;
+    }
+
+    seq.matchLength = mlBase;
+    if (mlBits > 0)
+        seq.matchLength += BIT_readBitsFast(&seqState->DStream, mlBits/*>0*/);
+
+    if (MEM_32bits() && (mlBits+llBits >= STREAM_ACCUMULATOR_MIN_32-LONG_OFFSETS_MAX_EXTRA_BITS_32))
+        BIT_reloadDStream(&seqState->DStream);
+    if (MEM_64bits() && UNLIKELY(totalBits >= STREAM_ACCUMULATOR_MIN_64-(LLFSELog+MLFSELog+OffFSELog)))
+        BIT_reloadDStream(&seqState->DStream);
+    /* Ensure there are enough bits to read the rest of data in 64-bit mode. */
+    ZSTD_STATIC_ASSERT(16+LLFSELog+MLFSELog+OffFSELog < STREAM_ACCUMULATOR_MIN_64);
+
+    seq.litLength = llBase;
+    if (llBits > 0)
+        seq.litLength += BIT_readBitsFast(&seqState->DStream, llBits/*>0*/);
+
+    if (MEM_32bits())
+        BIT_reloadDStream(&seqState->DStream);
+
+    DEBUGLOG(6, "seq: litL=%u, matchL=%u, offset=%u",
+                (U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);
+
+    if (prefetch == ZSTD_p_prefetch) {
+        size_t const pos = seqState->pos + seq.litLength;
+        const BYTE* const matchBase = (seq.offset > pos) ? seqState->dictEnd : seqState->prefixStart;
+        seq.match = matchBase + pos - seq.offset;  /* note : this operation can overflow when seq.offset is really too large, which can only happen when input is corrupted.
+                                                    * No consequence though : no memory access will occur, offset is only used for prefetching */
+        seqState->pos = pos + seq.matchLength;
+    }
+
+    /* ANS state update
+     * gcc-9.0.0 does 2.5% worse with ZSTD_updateFseStateWithDInfo().
+     * clang-9.2.0 does 7% worse with ZSTD_updateFseState().
+     * Naturally it seems like ZSTD_updateFseStateWithDInfo() should be the
+     * better option, so it is the default for other compilers. But, if you
+     * measure that it is worse, please put up a pull request.
+     */
+    {
+#if !defined(__clang__)
+        const int kUseUpdateFseState = 1;
+#else
+        const int kUseUpdateFseState = 0;
+#endif
+        if (kUseUpdateFseState) {
+            ZSTD_updateFseState(&seqState->stateLL, &seqState->DStream);    /* <=  9 bits */
+            ZSTD_updateFseState(&seqState->stateML, &seqState->DStream);    /* <=  9 bits */
+            if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);    /* <= 18 bits */
+            ZSTD_updateFseState(&seqState->stateOffb, &seqState->DStream);  /* <=  8 bits */
+        } else {
+            ZSTD_updateFseStateWithDInfo(&seqState->stateLL, &seqState->DStream, llDInfo);    /* <=  9 bits */
+            ZSTD_updateFseStateWithDInfo(&seqState->stateML, &seqState->DStream, mlDInfo);    /* <=  9 bits */
+            if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);    /* <= 18 bits */
+            ZSTD_updateFseStateWithDInfo(&seqState->stateOffb, &seqState->DStream, ofDInfo);  /* <=  8 bits */
+        }
+    }
+
+    return seq;
+}
+
+#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+MEM_STATIC int ZSTD_dictionaryIsActive(ZSTD_DCtx const* dctx, BYTE const* prefixStart, BYTE const* oLitEnd)
+{
+    size_t const windowSize = dctx->fParams.windowSize;
+    /* No dictionary used. */
+    if (dctx->dictContentEndForFuzzing == NULL) return 0;
+    /* Dictionary is our prefix. */
+    if (prefixStart == dctx->dictContentBeginForFuzzing) return 1;
+    /* Dictionary is not our ext-dict. */
+    if (dctx->dictEnd != dctx->dictContentEndForFuzzing) return 0;
+    /* Dictionary is not within our window size. */
+    if ((size_t)(oLitEnd - prefixStart) >= windowSize) return 0;
+    /* Dictionary is active. */
+    return 1;
+}
+
+MEM_STATIC void ZSTD_assertValidSequence(
+        ZSTD_DCtx const* dctx,
+        BYTE const* op, BYTE const* oend,
+        seq_t const seq,
+        BYTE const* prefixStart, BYTE const* virtualStart)
+{
+#if DEBUGLEVEL >= 1
+    size_t const windowSize = dctx->fParams.windowSize;
+    size_t const sequenceSize = seq.litLength + seq.matchLength;
+    BYTE const* const oLitEnd = op + seq.litLength;
+    DEBUGLOG(6, "Checking sequence: litL=%u matchL=%u offset=%u",
+            (U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);
+    assert(op <= oend);
+    assert((size_t)(oend - op) >= sequenceSize);
+    assert(sequenceSize <= ZSTD_BLOCKSIZE_MAX);
+    if (ZSTD_dictionaryIsActive(dctx, prefixStart, oLitEnd)) {
+        size_t const dictSize = (size_t)((char const*)dctx->dictContentEndForFuzzing - (char const*)dctx->dictContentBeginForFuzzing);
+        /* Offset must be within the dictionary. */
+        assert(seq.offset <= (size_t)(oLitEnd - virtualStart));
+        assert(seq.offset <= windowSize + dictSize);
+    } else {
+        /* Offset must be within our window. */
+        assert(seq.offset <= windowSize);
+    }
+#else
+    (void)dctx, (void)op, (void)oend, (void)seq, (void)prefixStart, (void)virtualStart;
+#endif
+}
+#endif
+
+#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
+FORCE_INLINE_TEMPLATE size_t
+DONT_VECTORIZE
+ZSTD_decompressSequences_body( ZSTD_DCtx* dctx,
+                               void* dst, size_t maxDstSize,
+                         const void* seqStart, size_t seqSize, int nbSeq,
+                         const ZSTD_longOffset_e isLongOffset,
+                         const int frame)
+{
+    const BYTE* ip = (const BYTE*)seqStart;
+    const BYTE* const iend = ip + seqSize;
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* const oend = ostart + maxDstSize;
+    BYTE* op = ostart;
+    const BYTE* litPtr = dctx->litPtr;
+    const BYTE* const litEnd = litPtr + dctx->litSize;
+    const BYTE* const prefixStart = (const BYTE*) (dctx->prefixStart);
+    const BYTE* const vBase = (const BYTE*) (dctx->virtualStart);
+    const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
+    DEBUGLOG(5, "ZSTD_decompressSequences_body");
+    (void)frame;
+
+    /* Regen sequences */
+    if (nbSeq) {
+        seqState_t seqState;
+        size_t error = 0;
+        dctx->fseEntropy = 1;
+        { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }
+        RETURN_ERROR_IF(
+            ERR_isError(BIT_initDStream(&seqState.DStream, ip, iend-ip)),
+            corruption_detected, "");
+        ZSTD_initFseState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
+        ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
+        ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
+        assert(dst != NULL);
+
+        ZSTD_STATIC_ASSERT(
+                BIT_DStream_unfinished < BIT_DStream_completed &&
+                BIT_DStream_endOfBuffer < BIT_DStream_completed &&
+                BIT_DStream_completed < BIT_DStream_overflow);
+
+#if defined(__x86_64__)
+        /* Align the decompression loop to 32 + 16 bytes.
+         *
+         * zstd compiled with gcc-9 on an Intel i9-9900k shows 10% decompression
+         * speed swings based on the alignment of the decompression loop. This
+         * performance swing is caused by parts of the decompression loop falling
+         * out of the DSB. The entire decompression loop should fit in the DSB,
+         * when it can't we get much worse performance. You can measure if you've
+         * hit the good case or the bad case with this perf command for some
+         * compressed file test.zst:
+         *
+         *   perf stat -e cycles -e instructions -e idq.all_dsb_cycles_any_uops \
+         *             -e idq.all_mite_cycles_any_uops -- ./zstd -tq test.zst
+         *
+         * If you see most cycles served out of the MITE you've hit the bad case.
+         * If you see most cycles served out of the DSB you've hit the good case.
+         * If it is pretty even then you may be in an okay case.
+         *
+         * I've been able to reproduce this issue on the following CPUs:
+         *   - Kabylake: Macbook Pro (15-inch, 2019) 2.4 GHz Intel Core i9
+         *               Use Instruments->Counters to get DSB/MITE cycles.
+         *               I never got performance swings, but I was able to
+         *               go from the good case of mostly DSB to half of the
+         *               cycles served from MITE.
+         *   - Coffeelake: Intel i9-9900k
+         *
+         * I haven't been able to reproduce the instability or DSB misses on any
+         * of the following CPUS:
+         *   - Haswell
+         *   - Broadwell: Intel(R) Xeon(R) CPU E5-2680 v4 @ 2.40GH
+         *   - Skylake
+         *
+         * If you are seeing performance stability this script can help test.
+         * It tests on 4 commits in zstd where I saw performance change.
+         *
+         *   https://gist.github.com/terrelln/9889fc06a423fd5ca6e99351564473f4
+         */
+        __asm__(".p2align 5");
+        __asm__("nop");
+        __asm__(".p2align 4");
+#endif
+        for ( ; ; ) {
+            seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset, ZSTD_p_noPrefetch);
+            size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, prefixStart, vBase, dictEnd);
+#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
+            assert(!ZSTD_isError(oneSeqSize));
+            if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
+#endif
+            DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
+            BIT_reloadDStream(&(seqState.DStream));
+            op += oneSeqSize;
+            /* gcc and clang both don't like early returns in this loop.
+             * Instead break and check for an error at the end of the loop.
+             */
+            if (UNLIKELY(ZSTD_isError(oneSeqSize))) {
+                error = oneSeqSize;
+                break;
+            }
+            if (UNLIKELY(!--nbSeq)) break;
+        }
+
+        /* check if reached exact end */
+        DEBUGLOG(5, "ZSTD_decompressSequences_body: after decode loop, remaining nbSeq : %i", nbSeq);
+        if (ZSTD_isError(error)) return error;
+        RETURN_ERROR_IF(nbSeq, corruption_detected, "");
+        RETURN_ERROR_IF(BIT_reloadDStream(&seqState.DStream) < BIT_DStream_completed, corruption_detected, "");
+        /* save reps for next block */
+        { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
+    }
+
+    /* last literal segment */
+    {   size_t const lastLLSize = litEnd - litPtr;
+        RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall, "");
+        if (op != NULL) {
+            ZSTD_memcpy(op, litPtr, lastLLSize);
+            op += lastLLSize;
+        }
+    }
+
+    return op-ostart;
+}
+
+static size_t
+ZSTD_decompressSequences_default(ZSTD_DCtx* dctx,
+                                 void* dst, size_t maxDstSize,
+                           const void* seqStart, size_t seqSize, int nbSeq,
+                           const ZSTD_longOffset_e isLongOffset,
+                           const int frame)
+{
+    return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+}
+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
+
+#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT
+FORCE_INLINE_TEMPLATE size_t
+ZSTD_decompressSequencesLong_body(
+                               ZSTD_DCtx* dctx,
+                               void* dst, size_t maxDstSize,
+                         const void* seqStart, size_t seqSize, int nbSeq,
+                         const ZSTD_longOffset_e isLongOffset,
+                         const int frame)
+{
+    const BYTE* ip = (const BYTE*)seqStart;
+    const BYTE* const iend = ip + seqSize;
+    BYTE* const ostart = (BYTE*)dst;
+    BYTE* const oend = ostart + maxDstSize;
+    BYTE* op = ostart;
+    const BYTE* litPtr = dctx->litPtr;
+    const BYTE* const litEnd = litPtr + dctx->litSize;
+    const BYTE* const prefixStart = (const BYTE*) (dctx->prefixStart);
+    const BYTE* const dictStart = (const BYTE*) (dctx->virtualStart);
+    const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
+    (void)frame;
+
+    /* Regen sequences */
+    if (nbSeq) {
+#define STORED_SEQS 4
+#define STORED_SEQS_MASK (STORED_SEQS-1)
+#define ADVANCED_SEQS 4
+        seq_t sequences[STORED_SEQS];
+        int const seqAdvance = MIN(nbSeq, ADVANCED_SEQS);
+        seqState_t seqState;
+        int seqNb;
+        dctx->fseEntropy = 1;
+        { int i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }
+        seqState.prefixStart = prefixStart;
+        seqState.pos = (size_t)(op-prefixStart);
+        seqState.dictEnd = dictEnd;
+        assert(dst != NULL);
+        assert(iend >= ip);
+        RETURN_ERROR_IF(
+            ERR_isError(BIT_initDStream(&seqState.DStream, ip, iend-ip)),
+            corruption_detected, "");
+        ZSTD_initFseState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
+        ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
+        ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
+
+        /* prepare in advance */
+        for (seqNb=0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && (seqNb<seqAdvance); seqNb++) {
+            sequences[seqNb] = ZSTD_decodeSequence(&seqState, isLongOffset, ZSTD_p_prefetch);
+            PREFETCH_L1(sequences[seqNb].match); PREFETCH_L1(sequences[seqNb].match + sequences[seqNb].matchLength - 1); /* note : it's safe to invoke PREFETCH() on any memory address, including invalid ones */
+        }
+        RETURN_ERROR_IF(seqNb<seqAdvance, corruption_detected, "");
+
+        /* decode and decompress */
+        for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && (seqNb<nbSeq) ; seqNb++) {
+            seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset, ZSTD_p_prefetch);
+            size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequences[(seqNb-ADVANCED_SEQS) & STORED_SEQS_MASK], &litPtr, litEnd, prefixStart, dictStart, dictEnd);
+#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
+            assert(!ZSTD_isError(oneSeqSize));
+            if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequences[(seqNb-ADVANCED_SEQS) & STORED_SEQS_MASK], prefixStart, dictStart);
+#endif
+            if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
+            PREFETCH_L1(sequence.match); PREFETCH_L1(sequence.match + sequence.matchLength - 1); /* note : it's safe to invoke PREFETCH() on any memory address, including invalid ones */
+            sequences[seqNb & STORED_SEQS_MASK] = sequence;
+            op += oneSeqSize;
+        }
+        RETURN_ERROR_IF(seqNb<nbSeq, corruption_detected, "");
+
+        /* finish queue */
+        seqNb -= seqAdvance;
+        for ( ; seqNb<nbSeq ; seqNb++) {
+            size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequences[seqNb&STORED_SEQS_MASK], &litPtr, litEnd, prefixStart, dictStart, dictEnd);
+#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
+            assert(!ZSTD_isError(oneSeqSize));
+            if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequences[seqNb&STORED_SEQS_MASK], prefixStart, dictStart);
+#endif
+            if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
+            op += oneSeqSize;
+        }
+
+        /* save reps for next block */
+        { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
+    }
+
+    /* last literal segment */
+    {   size_t const lastLLSize = litEnd - litPtr;
+        RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall, "");
+        if (op != NULL) {
+            ZSTD_memcpy(op, litPtr, lastLLSize);
+            op += lastLLSize;
+        }
+    }
+
+    return op-ostart;
+}
+
+static size_t
+ZSTD_decompressSequencesLong_default(ZSTD_DCtx* dctx,
+                                 void* dst, size_t maxDstSize,
+                           const void* seqStart, size_t seqSize, int nbSeq,
+                           const ZSTD_longOffset_e isLongOffset,
+                           const int frame)
+{
+    return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+}
+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */
+
+
+
+#if DYNAMIC_BMI2
+
+#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
+static TARGET_ATTRIBUTE("bmi2") size_t
+DONT_VECTORIZE
+ZSTD_decompressSequences_bmi2(ZSTD_DCtx* dctx,
+                                 void* dst, size_t maxDstSize,
+                           const void* seqStart, size_t seqSize, int nbSeq,
+                           const ZSTD_longOffset_e isLongOffset,
+                           const int frame)
+{
+    return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+}
+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
+
+#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT
+static TARGET_ATTRIBUTE("bmi2") size_t
+ZSTD_decompressSequencesLong_bmi2(ZSTD_DCtx* dctx,
+                                 void* dst, size_t maxDstSize,
+                           const void* seqStart, size_t seqSize, int nbSeq,
+                           const ZSTD_longOffset_e isLongOffset,
+                           const int frame)
+{
+    return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+}
+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */
+
+#endif /* DYNAMIC_BMI2 */
+
+typedef size_t (*ZSTD_decompressSequences_t)(
+                            ZSTD_DCtx* dctx,
+                            void* dst, size_t maxDstSize,
+                            const void* seqStart, size_t seqSize, int nbSeq,
+                            const ZSTD_longOffset_e isLongOffset,
+                            const int frame);
+
+#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
+static size_t
+ZSTD_decompressSequences(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize,
+                   const void* seqStart, size_t seqSize, int nbSeq,
+                   const ZSTD_longOffset_e isLongOffset,
+                   const int frame)
+{
+    DEBUGLOG(5, "ZSTD_decompressSequences");
+#if DYNAMIC_BMI2
+    if (dctx->bmi2) {
+        return ZSTD_decompressSequences_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+    }
+#endif
+  return ZSTD_decompressSequences_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+}
+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
+
+
+#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT
+/* ZSTD_decompressSequencesLong() :
+ * decompression function triggered when a minimum share of offsets is considered "long",
+ * aka out of cache.
+ * note : "long" definition seems overloaded here, sometimes meaning "wider than bitstream register", and sometimes meaning "farther than memory cache distance".
+ * This function will try to mitigate main memory latency through the use of prefetching */
+static size_t
+ZSTD_decompressSequencesLong(ZSTD_DCtx* dctx,
+                             void* dst, size_t maxDstSize,
+                             const void* seqStart, size_t seqSize, int nbSeq,
+                             const ZSTD_longOffset_e isLongOffset,
+                             const int frame)
+{
+    DEBUGLOG(5, "ZSTD_decompressSequencesLong");
+#if DYNAMIC_BMI2
+    if (dctx->bmi2) {
+        return ZSTD_decompressSequencesLong_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+    }
+#endif
+  return ZSTD_decompressSequencesLong_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+}
+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */
+
+
+
+#if !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT) && \
+    !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG)
+/* ZSTD_getLongOffsetsShare() :
+ * condition : offTable must be valid
+ * @return : "share" of long offsets (arbitrarily defined as > (1<<23))
+ *           compared to maximum possible of (1<<OffFSELog) */
+static unsigned
+ZSTD_getLongOffsetsShare(const ZSTD_seqSymbol* offTable)
+{
+    const void* ptr = offTable;
+    U32 const tableLog = ((const ZSTD_seqSymbol_header*)ptr)[0].tableLog;
+    const ZSTD_seqSymbol* table = offTable + 1;
+    U32 const max = 1 << tableLog;
+    U32 u, total = 0;
+    DEBUGLOG(5, "ZSTD_getLongOffsetsShare: (tableLog=%u)", tableLog);
+
+    assert(max <= (1 << OffFSELog));  /* max not too large */
+    for (u=0; u<max; u++) {
+        if (table[u].nbAdditionalBits > 22) total += 1;
+    }
+
+    assert(tableLog <= OffFSELog);
+    total <<= (OffFSELog - tableLog);  /* scale to OffFSELog */
+
+    return total;
+}
+#endif
+
+size_t
+ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
+                              void* dst, size_t dstCapacity,
+                        const void* src, size_t srcSize, const int frame)
+{   /* blockType == blockCompressed */
+    const BYTE* ip = (const BYTE*)src;
+    /* isLongOffset must be true if there are long offsets.
+     * Offsets are long if they are larger than 2^STREAM_ACCUMULATOR_MIN.
+     * We don't expect that to be the case in 64-bit mode.
+     * In block mode, window size is not known, so we have to be conservative.
+     * (note: but it could be evaluated from current-lowLimit)
+     */
+    ZSTD_longOffset_e const isLongOffset = (ZSTD_longOffset_e)(MEM_32bits() && (!frame || (dctx->fParams.windowSize > (1ULL << STREAM_ACCUMULATOR_MIN))));
+    DEBUGLOG(5, "ZSTD_decompressBlock_internal (size : %u)", (U32)srcSize);
+
+    RETURN_ERROR_IF(srcSize >= ZSTD_BLOCKSIZE_MAX, srcSize_wrong, "");
+
+    /* Decode literals section */
+    {   size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize);
+        DEBUGLOG(5, "ZSTD_decodeLiteralsBlock : %u", (U32)litCSize);
+        if (ZSTD_isError(litCSize)) return litCSize;
+        ip += litCSize;
+        srcSize -= litCSize;
+    }
+
+    /* Build Decoding Tables */
+    {
+        /* These macros control at build-time which decompressor implementation
+         * we use. If neither is defined, we do some inspection and dispatch at
+         * runtime.
+         */
+#if !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT) && \
+    !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG)
+        int usePrefetchDecoder = dctx->ddictIsCold;
+#endif
+        int nbSeq;
+        size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, srcSize);
+        if (ZSTD_isError(seqHSize)) return seqHSize;
+        ip += seqHSize;
+        srcSize -= seqHSize;
+
+        RETURN_ERROR_IF(dst == NULL && nbSeq > 0, dstSize_tooSmall, "NULL not handled");
+
+#if !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT) && \
+    !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG)
+        if ( !usePrefetchDecoder
+          && (!frame || (dctx->fParams.windowSize > (1<<24)))
+          && (nbSeq>ADVANCED_SEQS) ) {  /* could probably use a larger nbSeq limit */
+            U32 const shareLongOffsets = ZSTD_getLongOffsetsShare(dctx->OFTptr);
+            U32 const minShare = MEM_64bits() ? 7 : 20; /* heuristic values, correspond to 2.73% and 7.81% */
+            usePrefetchDecoder = (shareLongOffsets >= minShare);
+        }
+#endif
+
+        dctx->ddictIsCold = 0;
+
+#if !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT) && \
+    !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG)
+        if (usePrefetchDecoder)
+#endif
+#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT
+            return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset, frame);
+#endif
+
+#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
+        /* else */
+        return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset, frame);
+#endif
+    }
+}
+
+
+void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst, size_t dstSize)
+{
+    if (dst != dctx->previousDstEnd && dstSize > 0) {   /* not contiguous */
+        dctx->dictEnd = dctx->previousDstEnd;
+        dctx->virtualStart = (const char*)dst - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->prefixStart));
+        dctx->prefixStart = dst;
+        dctx->previousDstEnd = dst;
+    }
+}
+
+
+size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx,
+                            void* dst, size_t dstCapacity,
+                      const void* src, size_t srcSize)
+{
+    size_t dSize;
+    ZSTD_checkContinuity(dctx, dst, dstCapacity);
+    dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, /* frame */ 0);
+    dctx->previousDstEnd = (char*)dst + dSize;
+    return dSize;
+}
diff --git a/lib/zstd/decompress/zstd_decompress_block.h b/lib/zstd/decompress/zstd_decompress_block.h
new file mode 100644
index 0000000..e7f5f66
--- /dev/null
+++ b/lib/zstd/decompress/zstd_decompress_block.h
@@ -0,0 +1,62 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+#ifndef ZSTD_DEC_BLOCK_H
+#define ZSTD_DEC_BLOCK_H
+
+/*-*******************************************************
+ *  Dependencies
+ *********************************************************/
+#include "../common/zstd_deps.h"   /* size_t */
+#include <linux/zstd.h>    /* DCtx, and some public functions */
+#include "../common/zstd_internal.h"  /* blockProperties_t, and some public functions */
+#include "zstd_decompress_internal.h"  /* ZSTD_seqSymbol */
+
+
+/* ===   Prototypes   === */
+
+/* note: prototypes already published within `zstd.h` :
+ * ZSTD_decompressBlock()
+ */
+
+/* note: prototypes already published within `zstd_internal.h` :
+ * ZSTD_getcBlockSize()
+ * ZSTD_decodeSeqHeaders()
+ */
+
+
+/* ZSTD_decompressBlock_internal() :
+ * decompress block, starting at `src`,
+ * into destination buffer `dst`.
+ * @return : decompressed block size,
+ *           or an error code (which can be tested using ZSTD_isError())
+ */
+size_t ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
+                               void* dst, size_t dstCapacity,
+                         const void* src, size_t srcSize, const int frame);
+
+/* ZSTD_buildFSETable() :
+ * generate FSE decoding table for one symbol (ll, ml or off)
+ * this function must be called with valid parameters only
+ * (dt is large enough, normalizedCounter distribution total is a power of 2, max is within range, etc.)
+ * in which case it cannot fail.
+ * The workspace must be 4-byte aligned and at least ZSTD_BUILD_FSE_TABLE_WKSP_SIZE bytes, which is
+ * defined in zstd_decompress_internal.h.
+ * Internal use only.
+ */
+void ZSTD_buildFSETable(ZSTD_seqSymbol* dt,
+             const short* normalizedCounter, unsigned maxSymbolValue,
+             const U32* baseValue, const U32* nbAdditionalBits,
+                   unsigned tableLog, void* wksp, size_t wkspSize,
+                   int bmi2);
+
+
+#endif /* ZSTD_DEC_BLOCK_H */
diff --git a/lib/zstd/decompress/zstd_decompress_internal.h b/lib/zstd/decompress/zstd_decompress_internal.h
new file mode 100644
index 0000000..4b9052f
--- /dev/null
+++ b/lib/zstd/decompress/zstd_decompress_internal.h
@@ -0,0 +1,202 @@
+/*
+ * Copyright (c) Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+/* zstd_decompress_internal:
+ * objects and definitions shared within lib/decompress modules */
+
+ #ifndef ZSTD_DECOMPRESS_INTERNAL_H
+ #define ZSTD_DECOMPRESS_INTERNAL_H
+
+
+/*-*******************************************************
+ *  Dependencies
+ *********************************************************/
+#include "../common/mem.h"             /* BYTE, U16, U32 */
+#include "../common/zstd_internal.h"   /* ZSTD_seqSymbol */
+
+
+
+/*-*******************************************************
+ *  Constants
+ *********************************************************/
+static UNUSED_ATTR const U32 LL_base[MaxLL+1] = {
+                 0,    1,    2,     3,     4,     5,     6,      7,
+                 8,    9,   10,    11,    12,    13,    14,     15,
+                16,   18,   20,    22,    24,    28,    32,     40,
+                48,   64, 0x80, 0x100, 0x200, 0x400, 0x800, 0x1000,
+                0x2000, 0x4000, 0x8000, 0x10000 };
+
+static UNUSED_ATTR const U32 OF_base[MaxOff+1] = {
+                 0,        1,       1,       5,     0xD,     0x1D,     0x3D,     0x7D,
+                 0xFD,   0x1FD,   0x3FD,   0x7FD,   0xFFD,   0x1FFD,   0x3FFD,   0x7FFD,
+                 0xFFFD, 0x1FFFD, 0x3FFFD, 0x7FFFD, 0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD,
+                 0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD, 0x1FFFFFFD, 0x3FFFFFFD, 0x7FFFFFFD };
+
+static UNUSED_ATTR const U32 OF_bits[MaxOff+1] = {
+                     0,  1,  2,  3,  4,  5,  6,  7,
+                     8,  9, 10, 11, 12, 13, 14, 15,
+                    16, 17, 18, 19, 20, 21, 22, 23,
+                    24, 25, 26, 27, 28, 29, 30, 31 };
+
+static UNUSED_ATTR const U32 ML_base[MaxML+1] = {
+                     3,  4,  5,    6,     7,     8,     9,    10,
+                    11, 12, 13,   14,    15,    16,    17,    18,
+                    19, 20, 21,   22,    23,    24,    25,    26,
+                    27, 28, 29,   30,    31,    32,    33,    34,
+                    35, 37, 39,   41,    43,    47,    51,    59,
+                    67, 83, 99, 0x83, 0x103, 0x203, 0x403, 0x803,
+                    0x1003, 0x2003, 0x4003, 0x8003, 0x10003 };
+
+
+/*-*******************************************************
+ *  Decompression types
+ *********************************************************/
+ typedef struct {
+     U32 fastMode;
+     U32 tableLog;
+ } ZSTD_seqSymbol_header;
+
+ typedef struct {
+     U16  nextState;
+     BYTE nbAdditionalBits;
+     BYTE nbBits;
+     U32  baseValue;
+ } ZSTD_seqSymbol;
+
+ #define SEQSYMBOL_TABLE_SIZE(log)   (1 + (1 << (log)))
+
+#define ZSTD_BUILD_FSE_TABLE_WKSP_SIZE (sizeof(S16) * (MaxSeq + 1) + (1u << MaxFSELog) + sizeof(U64))
+#define ZSTD_BUILD_FSE_TABLE_WKSP_SIZE_U32 ((ZSTD_BUILD_FSE_TABLE_WKSP_SIZE + sizeof(U32) - 1) / sizeof(U32))
+
+typedef struct {
+    ZSTD_seqSymbol LLTable[SEQSYMBOL_TABLE_SIZE(LLFSELog)];    /* Note : Space reserved for FSE Tables */
+    ZSTD_seqSymbol OFTable[SEQSYMBOL_TABLE_SIZE(OffFSELog)];   /* is also used as temporary workspace while building hufTable during DDict creation */
+    ZSTD_seqSymbol MLTable[SEQSYMBOL_TABLE_SIZE(MLFSELog)];    /* and therefore must be at least HUF_DECOMPRESS_WORKSPACE_SIZE large */
+    HUF_DTable hufTable[HUF_DTABLE_SIZE(HufLog)];  /* can accommodate HUF_decompress4X */
+    U32 rep[ZSTD_REP_NUM];
+    U32 workspace[ZSTD_BUILD_FSE_TABLE_WKSP_SIZE_U32];
+} ZSTD_entropyDTables_t;
+
+typedef enum { ZSTDds_getFrameHeaderSize, ZSTDds_decodeFrameHeader,
+               ZSTDds_decodeBlockHeader, ZSTDds_decompressBlock,
+               ZSTDds_decompressLastBlock, ZSTDds_checkChecksum,
+               ZSTDds_decodeSkippableHeader, ZSTDds_skipFrame } ZSTD_dStage;
+
+typedef enum { zdss_init=0, zdss_loadHeader,
+               zdss_read, zdss_load, zdss_flush } ZSTD_dStreamStage;
+
+typedef enum {
+    ZSTD_use_indefinitely = -1,  /* Use the dictionary indefinitely */
+    ZSTD_dont_use = 0,           /* Do not use the dictionary (if one exists free it) */
+    ZSTD_use_once = 1            /* Use the dictionary once and set to ZSTD_dont_use */
+} ZSTD_dictUses_e;
+
+/* Hashset for storing references to multiple ZSTD_DDict within ZSTD_DCtx */
+typedef struct {
+    const ZSTD_DDict** ddictPtrTable;
+    size_t ddictPtrTableSize;
+    size_t ddictPtrCount;
+} ZSTD_DDictHashSet;
+
+struct ZSTD_DCtx_s
+{
+    const ZSTD_seqSymbol* LLTptr;
+    const ZSTD_seqSymbol* MLTptr;
+    const ZSTD_seqSymbol* OFTptr;
+    const HUF_DTable* HUFptr;
+    ZSTD_entropyDTables_t entropy;
+    U32 workspace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32];   /* space needed when building huffman tables */
+    const void* previousDstEnd;   /* detect continuity */
+    const void* prefixStart;      /* start of current segment */
+    const void* virtualStart;     /* virtual start of previous segment if it was just before current one */
+    const void* dictEnd;          /* end of previous segment */
+    size_t expected;
+    ZSTD_frameHeader fParams;
+    U64 processedCSize;
+    U64 decodedSize;
+    blockType_e bType;            /* used in ZSTD_decompressContinue(), store blockType between block header decoding and block decompression stages */
+    ZSTD_dStage stage;
+    U32 litEntropy;
+    U32 fseEntropy;
+    struct xxh64_state xxhState;
+    size_t headerSize;
+    ZSTD_format_e format;
+    ZSTD_forceIgnoreChecksum_e forceIgnoreChecksum;   /* User specified: if == 1, will ignore checksums in compressed frame. Default == 0 */
+    U32 validateChecksum;         /* if == 1, will validate checksum. Is == 1 if (fParams.checksumFlag == 1) and (forceIgnoreChecksum == 0). */
+    const BYTE* litPtr;
+    ZSTD_customMem customMem;
+    size_t litSize;
+    size_t rleSize;
+    size_t staticSize;
+    int bmi2;                     /* == 1 if the CPU supports BMI2 and 0 otherwise. CPU support is determined dynamically once per context lifetime. */
+
+    /* dictionary */
+    ZSTD_DDict* ddictLocal;
+    const ZSTD_DDict* ddict;     /* set by ZSTD_initDStream_usingDDict(), or ZSTD_DCtx_refDDict() */
+    U32 dictID;
+    int ddictIsCold;             /* if == 1 : dictionary is "new" for working context, and presumed "cold" (not in cpu cache) */
+    ZSTD_dictUses_e dictUses;
+    ZSTD_DDictHashSet* ddictSet;                    /* Hash set for multiple ddicts */
+    ZSTD_refMultipleDDicts_e refMultipleDDicts;     /* User specified: if == 1, will allow references to multiple DDicts. Default == 0 (disabled) */
+
+    /* streaming */
+    ZSTD_dStreamStage streamStage;
+    char*  inBuff;
+    size_t inBuffSize;
+    size_t inPos;
+    size_t maxWindowSize;
+    char*  outBuff;
+    size_t outBuffSize;
+    size_t outStart;
+    size_t outEnd;
+    size_t lhSize;
+    void* legacyContext;
+    U32 previousLegacyVersion;
+    U32 legacyVersion;
+    U32 hostageByte;
+    int noForwardProgress;
+    ZSTD_bufferMode_e outBufferMode;
+    ZSTD_outBuffer expectedOutBuffer;
+
+    /* workspace */
+    BYTE litBuffer[ZSTD_BLOCKSIZE_MAX + WILDCOPY_OVERLENGTH];
+    BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX];
+
+    size_t oversizedDuration;
+
+#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+    void const* dictContentBeginForFuzzing;
+    void const* dictContentEndForFuzzing;
+#endif
+
+    /* Tracing */
+};  /* typedef'd to ZSTD_DCtx within "zstd.h" */
+
+
+/*-*******************************************************
+ *  Shared internal functions
+ *********************************************************/
+
+/*! ZSTD_loadDEntropy() :
+ *  dict : must point at beginning of a valid zstd dictionary.
+ * @return : size of dictionary header (size of magic number + dict ID + entropy tables) */
+size_t ZSTD_loadDEntropy(ZSTD_entropyDTables_t* entropy,
+                   const void* const dict, size_t const dictSize);
+
+/*! ZSTD_checkContinuity() :
+ *  check if next `dst` follows previous position, where decompression ended.
+ *  If yes, do nothing (continue on current segment).
+ *  If not, classify previous segment as "external dictionary", and start a new segment.
+ *  This function cannot fail. */
+void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst, size_t dstSize);
+
+
+#endif /* ZSTD_DECOMPRESS_INTERNAL_H */
diff --git a/lib/zstd/decompress_sources.h b/lib/zstd/decompress_sources.h
new file mode 100644
index 0000000..0fbec50
--- /dev/null
+++ b/lib/zstd/decompress_sources.h
@@ -0,0 +1,28 @@
+/* SPDX-License-Identifier: GPL-2.0+ OR BSD-3-Clause */
+/*
+ * Copyright (c) Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/*
+ * This file includes every .c file needed for decompression.
+ * It is used by lib/decompress_unzstd.c to include the decompression
+ * source into the translation-unit, so it can be used for kernel
+ * decompression.
+ */
+
+#include "common/debug.c"
+#include "common/entropy_common.c"
+#include "common/error_private.c"
+#include "common/fse_decompress.c"
+#include "common/zstd_common.c"
+#include "decompress/huf_decompress.c"
+#include "decompress/zstd_ddict.c"
+#include "decompress/zstd_decompress.c"
+#include "decompress/zstd_decompress_block.c"
+#include "zstd_decompress_module.c"
diff --git a/lib/zstd/entropy_common.c b/lib/zstd/entropy_common.c
deleted file mode 100644
index 2b0a643..0000000
--- a/lib/zstd/entropy_common.c
+++ /dev/null
@@ -1,243 +0,0 @@
-/*
- * Common functions of New Generation Entropy library
- * Copyright (C) 2016, Yann Collet.
- *
- * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are
- * met:
- *
- *   * Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- *   * Redistributions in binary form must reproduce the above
- * copyright notice, this list of conditions and the following disclaimer
- * in the documentation and/or other materials provided with the
- * distribution.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- *
- * This program is free software; you can redistribute it and/or modify it under
- * the terms of the GNU General Public License version 2 as published by the
- * Free Software Foundation. This program is dual-licensed; you may select
- * either version 2 of the GNU General Public License ("GPL") or BSD license
- * ("BSD").
- *
- * You can contact the author at :
- * - Source repository : https://github.com/Cyan4973/