diff --git a/.gitignore b/.gitignore
index f4c0b09..e213b27 100644
--- a/.gitignore
+++ b/.gitignore
@@ -34,6 +34,7 @@
 *.gcno
 modules.builtin
 Module.symvers
+*.dwo
 
 #
 # Top-level generic files
diff --git a/Documentation/device-mapper/switch.txt b/Documentation/device-mapper/switch.txt
index 2fa7493..8897d04 100644
--- a/Documentation/device-mapper/switch.txt
+++ b/Documentation/device-mapper/switch.txt
@@ -106,6 +106,11 @@
     The path number in the range 0 ... (<num_paths> - 1).
     Expressed in hexadecimal (WITHOUT any prefix like 0x).
 
+R<n>,<m>
+    This parameter allows repetitive patterns to be loaded quickly. <n> and <m>
+    are hexadecimal numbers. The last <n> mappings are repeated in the next <m>
+    slots.
+
 Status
 ======
 
@@ -124,3 +129,10 @@
 Set mappings for the first 7 entries to point to devices switch0, switch1,
 switch2, switch0, switch1, switch2, switch1:
     dmsetup message switch 0 set_region_mappings 0:0 :1 :2 :0 :1 :2 :1
+
+Set repetitive mapping. This command:
+    dmsetup message switch 0 set_region_mappings 1000:1 :2 R2,10
+is equivalent to:
+    dmsetup message switch 0 set_region_mappings 1000:1 :2 :1 :2 :1 :2 :1 :2 \
+	:1 :2 :1 :2 :1 :2 :1 :2 :1 :2
+
diff --git a/Documentation/devicetree/bindings/mmc/exynos-dw-mshc.txt b/Documentation/devicetree/bindings/mmc/exynos-dw-mshc.txt
index 532b1d4..6cd3525 100644
--- a/Documentation/devicetree/bindings/mmc/exynos-dw-mshc.txt
+++ b/Documentation/devicetree/bindings/mmc/exynos-dw-mshc.txt
@@ -46,13 +46,14 @@
       - if CIU clock divider value is 0 (that is divide by 1), both tx and rx
         phase shift clocks should be 0.
 
-Required properties for a slot:
+Required properties for a slot (Deprecated - Recommend to use one slot per host):
 
 * gpios: specifies a list of gpios used for command, clock and data bus. The
   first gpio is the command line and the second gpio is the clock line. The
   rest of the gpios (depending on the bus-width property) are the data lines in
   no particular order. The format of the gpio specifier depends on the gpio
   controller.
+(Deprecated - Refer to Documentation/devicetree/binding/pinctrl/samsung-pinctrl.txt)
 
 Example:
 
@@ -69,21 +70,13 @@
 
 	dwmmc0@12200000 {
 		num-slots = <1>;
-		supports-highspeed;
+		cap-mmc-highspeed;
+		cap-sd-highspeed;
 		broken-cd;
 		fifo-depth = <0x80>;
 		card-detect-delay = <200>;
 		samsung,dw-mshc-ciu-div = <3>;
 		samsung,dw-mshc-sdr-timing = <2 3>;
 		samsung,dw-mshc-ddr-timing = <1 2>;
-
-		slot@0 {
-			reg = <0>;
-			bus-width = <8>;
-			gpios = <&gpc0 0 2 0 3>, <&gpc0 1 2 0 3>,
-				<&gpc1 0 2 3 3>, <&gpc1 1 2 3 3>,
-				<&gpc1 2 2 3 3>, <&gpc1 3 2 3 3>,
-				<&gpc0 3 2 3 3>, <&gpc0 4 2 3 3>,
-				<&gpc0 5 2 3 3>, <&gpc0 6 2 3 3>;
-		};
+		bus-width = <8>;
 	};
diff --git a/Documentation/devicetree/bindings/mmc/k3-dw-mshc.txt b/Documentation/devicetree/bindings/mmc/k3-dw-mshc.txt
index e5bc49f..3b35449 100644
--- a/Documentation/devicetree/bindings/mmc/k3-dw-mshc.txt
+++ b/Documentation/devicetree/bindings/mmc/k3-dw-mshc.txt
@@ -34,13 +34,11 @@
 		num-slots = <1>;
 		vmmc-supply = <&ldo12>;
 		fifo-depth = <0x100>;
-		supports-highspeed;
 		pinctrl-names = "default";
 		pinctrl-0 = <&sd_pmx_pins &sd_cfg_func1 &sd_cfg_func2>;
-		slot@0 {
-			reg = <0>;
-			bus-width = <4>;
-			disable-wp;
-			cd-gpios = <&gpio10 3 0>;
-		};
+		bus-width = <4>;
+		disable-wp;
+		cd-gpios = <&gpio10 3 0>;
+		cap-mmc-highspeed;
+		cap-sd-highspeed;
 	};
diff --git a/Documentation/devicetree/bindings/mmc/mmc.txt b/Documentation/devicetree/bindings/mmc/mmc.txt
index 3c18001..431716e 100644
--- a/Documentation/devicetree/bindings/mmc/mmc.txt
+++ b/Documentation/devicetree/bindings/mmc/mmc.txt
@@ -34,8 +34,8 @@
 - cap-power-off-card: powering off the card is safe
 - cap-sdio-irq: enable SDIO IRQ signalling on this interface
 - full-pwr-cycle: full power cycle of the card is supported
-- mmc-highspeed-ddr-1_8v: eMMC high-speed DDR mode(1.8V I/O) is supported
-- mmc-highspeed-ddr-1_2v: eMMC high-speed DDR mode(1.2V I/O) is supported
+- mmc-ddr-1_8v: eMMC high-speed DDR mode(1.8V I/O) is supported
+- mmc-ddr-1_2v: eMMC high-speed DDR mode(1.2V I/O) is supported
 - mmc-hs200-1_8v: eMMC HS200 mode(1.8V I/O) is supported
 - mmc-hs200-1_2v: eMMC HS200 mode(1.2V I/O) is supported
 - mmc-hs400-1_8v: eMMC HS400 mode(1.8V I/O) is supported
diff --git a/Documentation/devicetree/bindings/mmc/renesas,mmcif.txt b/Documentation/devicetree/bindings/mmc/renesas,mmcif.txt
new file mode 100644
index 0000000..299081f
--- /dev/null
+++ b/Documentation/devicetree/bindings/mmc/renesas,mmcif.txt
@@ -0,0 +1,32 @@
+* Renesas Multi Media Card Interface (MMCIF) Controller
+
+This file documents differences between the core properties in mmc.txt
+and the properties used by the MMCIF device.
+
+
+Required properties:
+
+- compatible: must contain one of the following
+	- "renesas,mmcif-r8a7740" for the MMCIF found in r8a7740 SoCs
+	- "renesas,mmcif-r8a7790" for the MMCIF found in r8a7790 SoCs
+	- "renesas,mmcif-r8a7791" for the MMCIF found in r8a7791 SoCs
+	- "renesas,sh-mmcif" for the generic MMCIF
+
+- clocks: reference to the functional clock
+
+- dmas: reference to the DMA channels, one per channel name listed in the
+  dma-names property.
+- dma-names: must contain "tx" for the transmit DMA channel and "rx" for the
+  receive DMA channel.
+
+
+Example: R8A7790 (R-Car H2) MMCIF0
+
+	mmcif0: mmc@ee200000 {
+		compatible = "renesas,mmcif-r8a7790", "renesas,sh-mmcif";
+		reg = <0 0xee200000 0 0x80>;
+		interrupts = <0 169 IRQ_TYPE_LEVEL_HIGH>;
+		clocks = <&mstp3_clks R8A7790_CLK_MMCIF0>;
+		dmas = <&dmac0 0xd1>, <&dmac0 0xd2>;
+		dma-names = "tx", "rx";
+	};
diff --git a/Documentation/devicetree/bindings/mmc/sdhci-msm.txt b/Documentation/devicetree/bindings/mmc/sdhci-msm.txt
index 81b33b5..485483a 100644
--- a/Documentation/devicetree/bindings/mmc/sdhci-msm.txt
+++ b/Documentation/devicetree/bindings/mmc/sdhci-msm.txt
@@ -27,8 +27,8 @@
 		bus-width = <8>;
 		non-removable;
 
-		vmmc = <&pm8941_l20>;
-		vqmmc = <&pm8941_s3>;
+		vmmc-supply = <&pm8941_l20>;
+		vqmmc-supply = <&pm8941_s3>;
 
 		pinctrl-names = "default";
 		pinctrl-0 = <&sdc1_clk &sdc1_cmd &sdc1_data>;
@@ -44,8 +44,8 @@
 		bus-width = <4>;
 		cd-gpios = <&msmgpio 62 0x1>;
 
-		vmmc = <&pm8941_l21>;
-		vqmmc = <&pm8941_l13>;
+		vmmc-supply = <&pm8941_l21>;
+		vqmmc-supply = <&pm8941_l13>;
 
 		pinctrl-names = "default";
 		pinctrl-0 = <&sdc2_clk &sdc2_cmd &sdc2_data>;
diff --git a/Documentation/devicetree/bindings/mmc/sdhci-st.txt b/Documentation/devicetree/bindings/mmc/sdhci-st.txt
new file mode 100644
index 0000000..7527db4
--- /dev/null
+++ b/Documentation/devicetree/bindings/mmc/sdhci-st.txt
@@ -0,0 +1,33 @@
+* STMicroelectronics sdhci-st MMC/SD controller
+
+This file documents the differences between the core properties in
+Documentation/devicetree/bindings/mmc/mmc.txt and the properties
+used by the sdhci-st driver.
+
+Required properties:
+- compatible :  Must be "st,sdhci"
+- clock-names : Should be "mmc"
+                See: Documentation/devicetree/bindings/resource-names.txt
+- clocks :      Phandle of the clock used by the sdhci controler
+                See: Documentation/devicetree/bindings/clock/clock-bindings.txt
+
+Optional properties:
+- non-removable: non-removable slot
+                 See: Documentation/devicetree/bindings/mmc/mmc.txt
+- bus-width: Number of data lines
+                 See: Documentation/devicetree/bindings/mmc/mmc.txt
+
+Example:
+
+mmc0: sdhci@fe81e000 {
+	compatible	= "st,sdhci";
+	status		= "disabled";
+	reg		= <0xfe81e000 0x1000>;
+	interrupts	= <GIC_SPI 127 IRQ_TYPE_NONE>;
+	interrupt-names	= "mmcirq";
+	pinctrl-names	= "default";
+	pinctrl-0	= <&pinctrl_mmc0>;
+	clock-names	= "mmc";
+	clocks		= <&clk_s_a1_ls 1>;
+	bus-width       = <8>
+};
diff --git a/Documentation/devicetree/bindings/mmc/synopsys-dw-mshc.txt b/Documentation/devicetree/bindings/mmc/synopsys-dw-mshc.txt
index 2d4a725..346c609 100644
--- a/Documentation/devicetree/bindings/mmc/synopsys-dw-mshc.txt
+++ b/Documentation/devicetree/bindings/mmc/synopsys-dw-mshc.txt
@@ -67,7 +67,8 @@
 * card-detect-delay: Delay in milli-seconds before detecting card after card
   insert event. The default value is 0.
 
-* supports-highspeed: Enables support for high speed cards (up to 50MHz)
+* supports-highspeed (DEPRECATED): Enables support for high speed cards (up to 50MHz)
+			   (use "cap-mmc-highspeed" or "cap-sd-highspeed" instead)
 
 * broken-cd: as documented in mmc core bindings.
 
@@ -98,14 +99,11 @@
 		clock-frequency = <400000000>;
 		clock-freq-min-max = <400000 200000000>;
 		num-slots = <1>;
-		supports-highspeed;
 		broken-cd;
 		fifo-depth = <0x80>;
 		card-detect-delay = <200>;
 		vmmc-supply = <&buck8>;
-
-		slot@0 {
-			reg = <0>;
-			bus-width = <8>;
-		};
+		bus-width = <8>;
+		cap-mmc-highspeed;
+		cap-sd-highspeed;
 	};
diff --git a/Documentation/devicetree/bindings/mmc/ti-omap-hsmmc.txt b/Documentation/devicetree/bindings/mmc/ti-omap-hsmmc.txt
index ce80561..76bf087 100644
--- a/Documentation/devicetree/bindings/mmc/ti-omap-hsmmc.txt
+++ b/Documentation/devicetree/bindings/mmc/ti-omap-hsmmc.txt
@@ -12,6 +12,7 @@
  Should be "ti,omap3-hsmmc", for OMAP3 controllers
  Should be "ti,omap3-pre-es3-hsmmc" for OMAP3 controllers pre ES3.0
  Should be "ti,omap4-hsmmc", for OMAP4 controllers
+ Should be "ti,am33xx-hsmmc", for AM335x controllers
 - ti,hwmods: Must be "mmc<n>", n is controller instance starting 1
 
 Optional properties:
@@ -56,3 +57,56 @@
 			&edma 25>;
 		dma-names = "tx", "rx";
 	};
+
+[workaround for missing swakeup on am33xx]
+
+This SOC is missing the swakeup line, it will not detect SDIO irq
+while in suspend.
+
+                             ------
+                             | PRCM |
+                              ------
+                               ^ |
+                       swakeup | | fclk
+                               | v
+       ------                -------               -----
+      | card | -- CIRQ -->  | hsmmc | -- IRQ -->  | CPU |
+       ------                -------               -----
+
+In suspend the fclk is off and the module is disfunctional. Even register reads
+will fail. A small logic in the host will request fclk restore, when an
+external event is detected. Once the clock is restored, the host detects the
+event normally. Since am33xx doesn't have this line it never wakes from
+suspend.
+
+The workaround is to reconfigure the dat1 line as a GPIO upon suspend. To make
+this work, we need to set the named pinctrl states "default" and "idle".
+Prepare idle to remux dat1 as a gpio, and default to remux it back as sdio
+dat1. The MMC driver will then toggle between idle and default state during
+runtime.
+
+In summary:
+1. select matching 'compatible' section, see example below.
+2. specify pinctrl states "default" and "idle", "sleep" is optional.
+3. specify the gpio irq used for detecting sdio irq in suspend
+
+If configuration is incomplete, a warning message is emitted "falling back to
+polling". Also check the "sdio irq mode" in /sys/kernel/debug/mmc0/regs. Mind
+not every application needs SDIO irq, e.g. MMC cards.
+
+	mmc1: mmc@48060100 {
+		compatible = "ti,am33xx-hsmmc";
+		...
+		pinctrl-names = "default", "idle", "sleep"
+		pinctrl-0 = <&mmc1_pins>;
+		pinctrl-1 = <&mmc1_idle>;
+		pinctrl-2 = <&mmc1_sleep>;
+		...
+		interrupts-extended = <&intc 64 &gpio2 28 0>;
+	};
+
+	mmc1_idle : pinmux_cirq_pin {
+		pinctrl-single,pins = <
+		        0x0f8 0x3f      /* GPIO2_28 */
+		>;
+	};
diff --git a/Documentation/devicetree/bindings/mmc/tmio_mmc.txt b/Documentation/devicetree/bindings/mmc/tmio_mmc.txt
index 6a2a116..fa0f327 100644
--- a/Documentation/devicetree/bindings/mmc/tmio_mmc.txt
+++ b/Documentation/devicetree/bindings/mmc/tmio_mmc.txt
@@ -18,6 +18,7 @@
 		"renesas,sdhi-r8a7778" - SDHI IP on R8A7778 SoC
 		"renesas,sdhi-r8a7779" - SDHI IP on R8A7779 SoC
 		"renesas,sdhi-r8a7790" - SDHI IP on R8A7790 SoC
+		"renesas,sdhi-r8a7791" - SDHI IP on R8A7791 SoC
 
 Optional properties:
 - toshiba,mmc-wrprotect-disable: write-protect detection is unavailable
diff --git a/Documentation/devicetree/changesets.txt b/Documentation/devicetree/changesets.txt
new file mode 100644
index 0000000..935ba5a
--- /dev/null
+++ b/Documentation/devicetree/changesets.txt
@@ -0,0 +1,40 @@
+A DT changeset is a method which allows one to apply changes
+in the live tree in such a way that either the full set of changes
+will be applied, or none of them will be. If an error occurs partway
+through applying the changeset, then the tree will be rolled back to the
+previous state. A changeset can also be removed after it has been
+applied.
+
+When a changeset is applied, all of the changes get applied to the tree
+at once before emitting OF_RECONFIG notifiers. This is so that the
+receiver sees a complete and consistent state of the tree when it
+receives the notifier.
+
+The sequence of a changeset is as follows.
+
+1. of_changeset_init() - initializes a changeset
+
+2. A number of DT tree change calls, of_changeset_attach_node(),
+of_changeset_detach_node(), of_changeset_add_property(),
+of_changeset_remove_property, of_changeset_update_property() to prepare
+a set of changes. No changes to the active tree are made at this point.
+All the change operations are recorded in the of_changeset 'entries'
+list.
+
+3. mutex_lock(of_mutex) - starts a changeset; The global of_mutex
+ensures there can only be one editor at a time.
+
+4. of_changeset_apply() - Apply the changes to the tree. Either the
+entire changeset will get applied, or if there is an error the tree will
+be restored to the previous state
+
+5. mutex_unlock(of_mutex) - All operations complete, release the mutex
+
+If a successfully applied changeset needs to be removed, it can be done
+with the following sequence.
+
+1. mutex_lock(of_mutex)
+
+2. of_changeset_revert()
+
+3. mutex_unlock(of_mutex)
diff --git a/Documentation/devicetree/todo.txt b/Documentation/devicetree/todo.txt
new file mode 100644
index 0000000..c3cf065
--- /dev/null
+++ b/Documentation/devicetree/todo.txt
@@ -0,0 +1,11 @@
+Todo list for devicetree:
+
+=== General structure ===
+- Switch from custom lists to (h)list_head for nodes and properties structure
+- Remove of_allnodes list and iterate using list of child nodes alone
+
+=== CONFIG_OF_DYNAMIC ===
+- Switch to RCU for tree updates and get rid of global spinlock
+- Document node lifecycle for CONFIG_OF_DYNAMIC
+- Always set ->full_name at of_attach_node() time
+- pseries: Get rid of open-coded tree modification from arch/powerpc/platforms/pseries/dlpar.c
diff --git a/Documentation/infiniband/user_mad.txt b/Documentation/infiniband/user_mad.txt
index 8a36695..7aca13a 100644
--- a/Documentation/infiniband/user_mad.txt
+++ b/Documentation/infiniband/user_mad.txt
@@ -26,6 +26,11 @@
   ioctl.  Also, all agents registered through a file descriptor will
   be unregistered when the descriptor is closed.
 
+  2014 -- a new registration ioctl is now provided which allows additional
+       fields to be provided during registration.
+       Users of this registration call are implicitly setting the use of
+       pkey_index (see below).
+
 Receiving MADs
 
   MADs are received using read().  The receive side now supports
@@ -104,10 +109,10 @@
   The old ib_umad interface did not allow setting the P_Key index for
   MADs that are sent and did not provide a way for obtaining the P_Key
   index of received MADs.  A new layout for struct ib_user_mad_hdr
-  with a pkey_index member has been defined; however, to preserve
-  binary compatibility with older applications, this new layout will
-  not be used unless the IB_USER_MAD_ENABLE_PKEY ioctl is called
-  before a file descriptor is used for anything else.
+  with a pkey_index member has been defined; however, to preserve binary
+  compatibility with older applications, this new layout will not be used
+  unless one of IB_USER_MAD_ENABLE_PKEY or IB_USER_MAD_REGISTER_AGENT2 ioctl's
+  are called before a file descriptor is used for anything else.
 
   In September 2008, the IB_USER_MAD_ABI_VERSION will be incremented
   to 6, the new layout of struct ib_user_mad_hdr will be used by
diff --git a/Documentation/kbuild/00-INDEX b/Documentation/kbuild/00-INDEX
index e8d2b6d..8c5e6aa 100644
--- a/Documentation/kbuild/00-INDEX
+++ b/Documentation/kbuild/00-INDEX
@@ -1,5 +1,7 @@
 00-INDEX
 	- this file: info on the kernel build process
+headers_install.txt
+	- how to export Linux headers for use by userspace
 kbuild.txt
 	- developer information on kbuild
 kconfig.txt
diff --git a/Documentation/make/headers_install.txt b/Documentation/kbuild/headers_install.txt
similarity index 100%
rename from Documentation/make/headers_install.txt
rename to Documentation/kbuild/headers_install.txt
diff --git a/Documentation/kbuild/makefiles.txt b/Documentation/kbuild/makefiles.txt
index c600e2f..764f599 100644
--- a/Documentation/kbuild/makefiles.txt
+++ b/Documentation/kbuild/makefiles.txt
@@ -23,11 +23,10 @@
 	=== 4 Host Program support
 	   --- 4.1 Simple Host Program
 	   --- 4.2 Composite Host Programs
-	   --- 4.3 Defining shared libraries
-	   --- 4.4 Using C++ for host programs
-	   --- 4.5 Controlling compiler options for host programs
-	   --- 4.6 When host programs are actually built
-	   --- 4.7 Using hostprogs-$(CONFIG_FOO)
+	   --- 4.3 Using C++ for host programs
+	   --- 4.4 Controlling compiler options for host programs
+	   --- 4.5 When host programs are actually built
+	   --- 4.6 Using hostprogs-$(CONFIG_FOO)
 
 	=== 5 Kbuild clean infrastructure
 
@@ -643,29 +642,7 @@
 	Finally, the two .o files are linked to the executable, lxdialog.
 	Note: The syntax <executable>-y is not permitted for host-programs.
 
---- 4.3 Defining shared libraries
-
-	Objects with extension .so are considered shared libraries, and
-	will be compiled as position independent objects.
-	Kbuild provides support for shared libraries, but the usage
-	shall be restricted.
-	In the following example the libkconfig.so shared library is used
-	to link the executable conf.
-
-	Example:
-		#scripts/kconfig/Makefile
-		hostprogs-y     := conf
-		conf-objs       := conf.o libkconfig.so
-		libkconfig-objs := expr.o type.o
-
-	Shared libraries always require a corresponding -objs line, and
-	in the example above the shared library libkconfig is composed by
-	the two objects expr.o and type.o.
-	expr.o and type.o will be built as position independent code and
-	linked as a shared library libkconfig.so. C++ is not supported for
-	shared libraries.
-
---- 4.4 Using C++ for host programs
+--- 4.3 Using C++ for host programs
 
 	kbuild offers support for host programs written in C++. This was
 	introduced solely to support kconfig, and is not recommended
@@ -688,7 +665,7 @@
 		qconf-cxxobjs := qconf.o
 		qconf-objs    := check.o
 
---- 4.5 Controlling compiler options for host programs
+--- 4.4 Controlling compiler options for host programs
 
 	When compiling host programs, it is possible to set specific flags.
 	The programs will always be compiled utilising $(HOSTCC) passed
@@ -716,7 +693,7 @@
 	When linking qconf, it will be passed the extra option
 	"-L$(QTDIR)/lib".
 
---- 4.6 When host programs are actually built
+--- 4.5 When host programs are actually built
 
 	Kbuild will only build host-programs when they are referenced
 	as a prerequisite.
@@ -747,7 +724,7 @@
 	This will tell kbuild to build lxdialog even if not referenced in
 	any rule.
 
---- 4.7 Using hostprogs-$(CONFIG_FOO)
+--- 4.6 Using hostprogs-$(CONFIG_FOO)
 
 	A typical pattern in a Kbuild file looks like this:
 
diff --git a/Makefile b/Makefile
index a897c50..6aace67 100644
--- a/Makefile
+++ b/Makefile
@@ -372,6 +372,7 @@
 INSTALLKERNEL  := installkernel
 DEPMOD		= /sbin/depmod
 PERL		= perl
+PYTHON		= python
 CHECK		= sparse
 
 CHECKFLAGS     := -D__linux__ -Dlinux -D__STDC__ -Dunix -D__unix__ \
@@ -422,7 +423,7 @@
 export VERSION PATCHLEVEL SUBLEVEL KERNELRELEASE KERNELVERSION
 export ARCH SRCARCH CONFIG_SHELL HOSTCC HOSTCFLAGS CROSS_COMPILE AS LD CC
 export CPP AR NM STRIP OBJCOPY OBJDUMP
-export MAKE AWK GENKSYMS INSTALLKERNEL PERL UTS_MACHINE
+export MAKE AWK GENKSYMS INSTALLKERNEL PERL PYTHON UTS_MACHINE
 export HOSTCXX HOSTCXXFLAGS LDFLAGS_MODULE CHECK CHECKFLAGS
 
 export KBUILD_CPPFLAGS NOSTDINC_FLAGS LINUXINCLUDE OBJCOPYFLAGS LDFLAGS
@@ -687,6 +688,7 @@
 # source of a reference will be _MergedGlobals and not on of the whitelisted names.
 # See modpost pattern 2
 KBUILD_CFLAGS += $(call cc-option, -mno-global-merge,)
+KBUILD_CFLAGS += $(call cc-option, -fcatch-undefined-behavior)
 else
 
 # This warning generated too much noise in a regular build.
@@ -710,9 +712,16 @@
 KBUILD_CFLAGS   += $(call cc-option, -fno-var-tracking-assignments)
 
 ifdef CONFIG_DEBUG_INFO
+ifdef CONFIG_DEBUG_INFO_SPLIT
+KBUILD_CFLAGS   += $(call cc-option, -gsplit-dwarf, -g)
+else
 KBUILD_CFLAGS	+= -g
+endif
 KBUILD_AFLAGS	+= -Wa,-gdwarf-2
 endif
+ifdef CONFIG_DEBUG_INFO_DWARF4
+KBUILD_CFLAGS	+= $(call cc-option, -gdwarf-4,)
+endif
 
 ifdef CONFIG_DEBUG_INFO_REDUCED
 KBUILD_CFLAGS 	+= $(call cc-option, -femit-struct-debug-baseonly) \
@@ -1055,6 +1064,13 @@
 	$(Q)$(MAKE) $(hdr-inst)=arch/$(hdr-arch)/include/uapi/asm $(hdr-dst) HDRCHECK=1
 
 # ---------------------------------------------------------------------------
+# Kernel selftest
+
+PHONY += kselftest
+kselftest:
+	$(Q)$(MAKE) -C tools/testing/selftests run_tests
+
+# ---------------------------------------------------------------------------
 # Modules
 
 ifdef CONFIG_MODULES
@@ -1241,9 +1257,9 @@
 	@echo  '  tags/TAGS	  - Generate tags file for editors'
 	@echo  '  cscope	  - Generate cscope index'
 	@echo  '  gtags           - Generate GNU GLOBAL index'
-	@echo  '  kernelrelease	  - Output the release version string'
-	@echo  '  kernelversion	  - Output the version stored in Makefile'
-	@echo  '  image_name	  - Output the image name'
+	@echo  '  kernelrelease	  - Output the release version string (use with make -s)'
+	@echo  '  kernelversion	  - Output the version stored in Makefile (use with make -s)'
+	@echo  '  image_name	  - Output the image name (use with make -s)'
 	@echo  '  headers_install - Install sanitised kernel headers to INSTALL_HDR_PATH'; \
 	 echo  '                    (default: $(INSTALL_HDR_PATH))'; \
 	 echo  ''
@@ -1257,6 +1273,11 @@
 	@echo  '  headerdep       - Detect inclusion cycles in headers'
 	@$(MAKE) -f $(srctree)/scripts/Makefile.help checker-help
 	@echo  ''
+	@echo  'Kernel selftest'
+	@echo  '  kselftest       - Build and run kernel selftest (run as root)'
+	@echo  '                    Build, install, and boot kernel before'
+	@echo  '                    running kselftest on it'
+	@echo  ''
 	@echo  'Kernel packaging:'
 	@$(MAKE) $(build)=$(package-dir) help
 	@echo  ''
@@ -1398,6 +1419,7 @@
 	@find $(if $(KBUILD_EXTMOD), $(KBUILD_EXTMOD), .) $(RCS_FIND_IGNORE) \
 		\( -name '*.[oas]' -o -name '*.ko' -o -name '.*.cmd' \
 		-o -name '*.ko.*' \
+		-o -name '*.dwo'  \
 		-o -name '.*.d' -o -name '.*.tmp' -o -name '*.mod.c' \
 		-o -name '*.symtypes' -o -name 'modules.order' \
 		-o -name modules.builtin -o -name '.tmp_*.o.*' \
diff --git a/arch/arm/boot/dts/versatile-ab.dts b/arch/arm/boot/dts/versatile-ab.dts
index 36c771a..27d0d9c 100644
--- a/arch/arm/boot/dts/versatile-ab.dts
+++ b/arch/arm/boot/dts/versatile-ab.dts
@@ -15,6 +15,10 @@
 		i2c0 = &i2c0;
 	};
 
+	chosen {
+		stdout-path = &uart0;
+	};
+
 	memory {
 		reg = <0x0 0x08000000>;
 	};
diff --git a/arch/arm/boot/dts/versatile-pb.dts b/arch/arm/boot/dts/versatile-pb.dts
index d025048..e36c1e8 100644
--- a/arch/arm/boot/dts/versatile-pb.dts
+++ b/arch/arm/boot/dts/versatile-pb.dts
@@ -56,5 +56,3 @@
 		};
 	};
 };
-
-#include <testcases.dtsi>
diff --git a/arch/arm/xen/grant-table.c b/arch/arm/xen/grant-table.c
index 2c4041c..e437918 100644
--- a/arch/arm/xen/grant-table.c
+++ b/arch/arm/xen/grant-table.c
@@ -49,8 +49,3 @@
 {
 	return 0;
 }
-
-int arch_gnttab_init(unsigned long nr_shared, unsigned long nr_status)
-{
-	return 0;
-}
diff --git a/arch/ia64/Makefile b/arch/ia64/Makefile
index f37238f..5441b14 100644
--- a/arch/ia64/Makefile
+++ b/arch/ia64/Makefile
@@ -76,7 +76,7 @@
 	$(Q)$(MAKE) $(build)=$(boot) $@
 
 unwcheck: vmlinux
-	-$(Q)READELF=$(READELF) python $(srctree)/arch/ia64/scripts/unwcheck.py $<
+	-$(Q)READELF=$(READELF) $(PYTHON) $(srctree)/arch/ia64/scripts/unwcheck.py $<
 
 archclean:
 	$(Q)$(MAKE) $(clean)=$(boot)
diff --git a/arch/powerpc/boot/gunzip_util.c b/arch/powerpc/boot/gunzip_util.c
index ef2aed0..9dc5250 100644
--- a/arch/powerpc/boot/gunzip_util.c
+++ b/arch/powerpc/boot/gunzip_util.c
@@ -112,10 +112,10 @@
 		r = zlib_inflate(&state->s, Z_FULL_FLUSH);
 		if (r != Z_OK && r != Z_STREAM_END)
 			fatal("inflate returned %d msg: %s\n\r", r, state->s.msg);
-		len = state->s.next_out - (unsigned char *)dst;
+		len = state->s.next_out - (Byte *)dst;
 	} else {
 		/* uncompressed image */
-		len = min(state->s.avail_in, (unsigned)dstlen);
+		len = min(state->s.avail_in, (uLong)dstlen);
 		memcpy(dst, state->s.next_in, len);
 		state->s.next_in += len;
 		state->s.avail_in -= len;
diff --git a/arch/powerpc/include/asm/cputable.h b/arch/powerpc/include/asm/cputable.h
index 642e436..daa5af9 100644
--- a/arch/powerpc/include/asm/cputable.h
+++ b/arch/powerpc/include/asm/cputable.h
@@ -459,7 +459,8 @@
 #define CPU_FTRS_POSSIBLE	\
 	    (CPU_FTRS_POWER4 | CPU_FTRS_PPC970 | CPU_FTRS_POWER5 | \
 	     CPU_FTRS_POWER6 | CPU_FTRS_POWER7 | CPU_FTRS_POWER8E | \
-	     CPU_FTRS_POWER8 | CPU_FTRS_CELL | CPU_FTRS_PA6T | CPU_FTR_VSX)
+	     CPU_FTRS_POWER8 | CPU_FTRS_POWER8_DD1 | CPU_FTRS_CELL | \
+	     CPU_FTRS_PA6T | CPU_FTR_VSX)
 #endif
 #else
 enum {
@@ -509,7 +510,8 @@
 #define CPU_FTRS_ALWAYS		\
 	    (CPU_FTRS_POWER4 & CPU_FTRS_PPC970 & CPU_FTRS_POWER5 & \
 	     CPU_FTRS_POWER6 & CPU_FTRS_POWER7 & CPU_FTRS_CELL & \
-	     CPU_FTRS_PA6T & CPU_FTRS_POSSIBLE)
+	     CPU_FTRS_PA6T & CPU_FTRS_POWER8 & CPU_FTRS_POWER8E & \
+	     CPU_FTRS_POWER8_DD1 & CPU_FTRS_POSSIBLE)
 #endif
 #else
 enum {
diff --git a/arch/powerpc/include/asm/machdep.h b/arch/powerpc/include/asm/machdep.h
index 44e9051..b125cea 100644
--- a/arch/powerpc/include/asm/machdep.h
+++ b/arch/powerpc/include/asm/machdep.h
@@ -57,10 +57,10 @@
 	void            (*hpte_removebolted)(unsigned long ea,
 					     int psize, int ssize);
 	void		(*flush_hash_range)(unsigned long number, int local);
-	void		(*hugepage_invalidate)(struct mm_struct *mm,
+	void		(*hugepage_invalidate)(unsigned long vsid,
+					       unsigned long addr,
 					       unsigned char *hpte_slot_array,
-					       unsigned long addr, int psize);
-
+					       int psize, int ssize);
 	/* special for kexec, to be called in real mode, linear mapping is
 	 * destroyed as well */
 	void		(*hpte_clear_all)(void);
diff --git a/arch/powerpc/include/asm/opal.h b/arch/powerpc/include/asm/opal.h
index b2f8ce1..86055e5 100644
--- a/arch/powerpc/include/asm/opal.h
+++ b/arch/powerpc/include/asm/opal.h
@@ -149,6 +149,8 @@
 #define OPAL_DUMP_INFO2				94
 #define OPAL_PCI_EEH_FREEZE_SET			97
 #define OPAL_HANDLE_HMI				98
+#define OPAL_REGISTER_DUMP_REGION		101
+#define OPAL_UNREGISTER_DUMP_REGION		102
 
 #ifndef __ASSEMBLY__
 
@@ -920,6 +922,8 @@
 		uint64_t length);
 int64_t opal_sensor_read(uint32_t sensor_hndl, int token, __be32 *sensor_data);
 int64_t opal_handle_hmi(void);
+int64_t opal_register_dump_region(uint32_t id, uint64_t start, uint64_t end);
+int64_t opal_unregister_dump_region(uint32_t id);
 
 /* Internal functions */
 extern int early_init_dt_scan_opal(unsigned long node, const char *uname,
@@ -974,6 +978,13 @@
 					     unsigned long vmalloc_size);
 void opal_free_sg_list(struct opal_sg_list *sg);
 
+/*
+ * Dump region ID range usable by the OS
+ */
+#define OPAL_DUMP_REGION_HOST_START		0x80
+#define OPAL_DUMP_REGION_LOG_BUF		0x80
+#define OPAL_DUMP_REGION_HOST_END		0xFF
+
 #endif /* __ASSEMBLY__ */
 
 #endif /* __OPAL_H */
diff --git a/arch/powerpc/include/asm/pgtable-ppc64.h b/arch/powerpc/include/asm/pgtable-ppc64.h
index eb92610..7b3d54f 100644
--- a/arch/powerpc/include/asm/pgtable-ppc64.h
+++ b/arch/powerpc/include/asm/pgtable-ppc64.h
@@ -413,7 +413,7 @@
 }
 
 extern void hpte_do_hugepage_flush(struct mm_struct *mm, unsigned long addr,
-				   pmd_t *pmdp);
+				   pmd_t *pmdp, unsigned long old_pmd);
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 extern pmd_t pfn_pmd(unsigned long pfn, pgprot_t pgprot);
 extern pmd_t mk_pmd(struct page *page, pgprot_t pgprot);
diff --git a/arch/powerpc/include/asm/pte-hash64-64k.h b/arch/powerpc/include/asm/pte-hash64-64k.h
index b6d2d42..4f4ec2a 100644
--- a/arch/powerpc/include/asm/pte-hash64-64k.h
+++ b/arch/powerpc/include/asm/pte-hash64-64k.h
@@ -46,11 +46,31 @@
  * in order to deal with 64K made of 4K HW pages. Thus we override the
  * generic accessors and iterators here
  */
-#define __real_pte(e,p) 	((real_pte_t) { \
-			(e), (pte_val(e) & _PAGE_COMBO) ? \
-				(pte_val(*((p) + PTRS_PER_PTE))) : 0 })
-#define __rpte_to_hidx(r,index)	((pte_val((r).pte) & _PAGE_COMBO) ? \
-        (((r).hidx >> ((index)<<2)) & 0xf) : ((pte_val((r).pte) >> 12) & 0xf))
+#define __real_pte __real_pte
+static inline real_pte_t __real_pte(pte_t pte, pte_t *ptep)
+{
+	real_pte_t rpte;
+
+	rpte.pte = pte;
+	rpte.hidx = 0;
+	if (pte_val(pte) & _PAGE_COMBO) {
+		/*
+		 * Make sure we order the hidx load against the _PAGE_COMBO
+		 * check. The store side ordering is done in __hash_page_4K
+		 */
+		smp_rmb();
+		rpte.hidx = pte_val(*((ptep) + PTRS_PER_PTE));
+	}
+	return rpte;
+}
+
+static inline unsigned long __rpte_to_hidx(real_pte_t rpte, unsigned long index)
+{
+	if ((pte_val(rpte.pte) & _PAGE_COMBO))
+		return (rpte.hidx >> (index<<2)) & 0xf;
+	return (pte_val(rpte.pte) >> 12) & 0xf;
+}
+
 #define __rpte_to_pte(r)	((r).pte)
 #define __rpte_sub_valid(rpte, index) \
 	(pte_val(rpte.pte) & (_PAGE_HPTE_SUB0 >> (index)))
diff --git a/arch/powerpc/include/asm/reg.h b/arch/powerpc/include/asm/reg.h
index 1c987bf..0c05059 100644
--- a/arch/powerpc/include/asm/reg.h
+++ b/arch/powerpc/include/asm/reg.h
@@ -213,9 +213,8 @@
 #define SPRN_ACOP	0x1F	/* Available Coprocessor Register */
 #define SPRN_TFIAR	0x81	/* Transaction Failure Inst Addr   */
 #define SPRN_TEXASR	0x82	/* Transaction EXception & Summary */
-#define   TEXASR_FS	__MASK(63-36)	/* Transaction Failure Summary */
 #define SPRN_TEXASRU	0x83	/* ''	   ''	   ''	 Upper 32  */
-#define   TEXASR_FS     __MASK(63-36) /* TEXASR Failure Summary */
+#define   TEXASR_FS	__MASK(63-36) /* TEXASR Failure Summary */
 #define SPRN_TFHAR	0x80	/* Transaction Failure Handler Addr */
 #define SPRN_CTRLF	0x088
 #define SPRN_CTRLT	0x098
diff --git a/arch/powerpc/include/asm/spinlock.h b/arch/powerpc/include/asm/spinlock.h
index 35aa339..4dbe072 100644
--- a/arch/powerpc/include/asm/spinlock.h
+++ b/arch/powerpc/include/asm/spinlock.h
@@ -61,6 +61,7 @@
 
 static inline int arch_spin_is_locked(arch_spinlock_t *lock)
 {
+	smp_mb();
 	return !arch_spin_value_unlocked(*lock);
 }
 
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index 6144d5a..050f79a 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -592,61 +592,6 @@
 	MASKABLE_EXCEPTION_HV_OOL(0xe62, hmi_exception)
 	KVM_HANDLER(PACA_EXGEN, EXC_HV, 0xe62)
 
-	.globl hmi_exception_early
-hmi_exception_early:
-	EXCEPTION_PROLOG_1(PACA_EXGEN, NOTEST, 0xe60)
-	mr	r10,r1			/* Save r1			*/
-	ld	r1,PACAEMERGSP(r13)	/* Use emergency stack		*/
-	subi	r1,r1,INT_FRAME_SIZE	/* alloc stack frame		*/
-	std	r9,_CCR(r1)		/* save CR in stackframe	*/
-	mfspr	r11,SPRN_HSRR0		/* Save HSRR0 */
-	std	r11,_NIP(r1)		/* save HSRR0 in stackframe	*/
-	mfspr	r12,SPRN_HSRR1		/* Save SRR1 */
-	std	r12,_MSR(r1)		/* save SRR1 in stackframe	*/
-	std	r10,0(r1)		/* make stack chain pointer	*/
-	std	r0,GPR0(r1)		/* save r0 in stackframe	*/
-	std	r10,GPR1(r1)		/* save r1 in stackframe	*/
-	EXCEPTION_PROLOG_COMMON_2(PACA_EXGEN)
-	EXCEPTION_PROLOG_COMMON_3(0xe60)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	hmi_exception_realmode
-	/* Windup the stack. */
-	/* Clear MSR_RI before setting SRR0 and SRR1. */
-	li	r0,MSR_RI
-	mfmsr	r9			/* get MSR value */
-	andc	r9,r9,r0
-	mtmsrd	r9,1			/* Clear MSR_RI */
-	/* Move original HSRR0 and HSRR1 into the respective regs */
-	ld	r9,_MSR(r1)
-	mtspr	SPRN_HSRR1,r9
-	ld	r3,_NIP(r1)
-	mtspr	SPRN_HSRR0,r3
-	ld	r9,_CTR(r1)
-	mtctr	r9
-	ld	r9,_XER(r1)
-	mtxer	r9
-	ld	r9,_LINK(r1)
-	mtlr	r9
-	REST_GPR(0, r1)
-	REST_8GPRS(2, r1)
-	REST_GPR(10, r1)
-	ld	r11,_CCR(r1)
-	mtcr	r11
-	REST_GPR(11, r1)
-	REST_2GPRS(12, r1)
-	/* restore original r1. */
-	ld	r1,GPR1(r1)
-
-	/*
-	 * Go to virtual mode and pull the HMI event information from
-	 * firmware.
-	 */
-	.globl hmi_exception_after_realmode
-hmi_exception_after_realmode:
-	SET_SCRATCH0(r13)
-	EXCEPTION_PROLOG_0(PACA_EXGEN)
-	b	hmi_exception_hv
-
 	MASKABLE_EXCEPTION_HV_OOL(0xe82, h_doorbell)
 	KVM_HANDLER(PACA_EXGEN, EXC_HV, 0xe82)
 
@@ -1306,6 +1251,61 @@
 	. = 0x8000
 #endif /* defined(CONFIG_PPC_PSERIES) || defined(CONFIG_PPC_POWERNV) */
 
+	.globl hmi_exception_early
+hmi_exception_early:
+	EXCEPTION_PROLOG_1(PACA_EXGEN, NOTEST, 0xe60)
+	mr	r10,r1			/* Save r1			*/
+	ld	r1,PACAEMERGSP(r13)	/* Use emergency stack		*/
+	subi	r1,r1,INT_FRAME_SIZE	/* alloc stack frame		*/
+	std	r9,_CCR(r1)		/* save CR in stackframe	*/
+	mfspr	r11,SPRN_HSRR0		/* Save HSRR0 */
+	std	r11,_NIP(r1)		/* save HSRR0 in stackframe	*/
+	mfspr	r12,SPRN_HSRR1		/* Save SRR1 */
+	std	r12,_MSR(r1)		/* save SRR1 in stackframe	*/
+	std	r10,0(r1)		/* make stack chain pointer	*/
+	std	r0,GPR0(r1)		/* save r0 in stackframe	*/
+	std	r10,GPR1(r1)		/* save r1 in stackframe	*/
+	EXCEPTION_PROLOG_COMMON_2(PACA_EXGEN)
+	EXCEPTION_PROLOG_COMMON_3(0xe60)
+	addi	r3,r1,STACK_FRAME_OVERHEAD
+	bl	hmi_exception_realmode
+	/* Windup the stack. */
+	/* Clear MSR_RI before setting SRR0 and SRR1. */
+	li	r0,MSR_RI
+	mfmsr	r9			/* get MSR value */
+	andc	r9,r9,r0
+	mtmsrd	r9,1			/* Clear MSR_RI */
+	/* Move original HSRR0 and HSRR1 into the respective regs */
+	ld	r9,_MSR(r1)
+	mtspr	SPRN_HSRR1,r9
+	ld	r3,_NIP(r1)
+	mtspr	SPRN_HSRR0,r3
+	ld	r9,_CTR(r1)
+	mtctr	r9
+	ld	r9,_XER(r1)
+	mtxer	r9
+	ld	r9,_LINK(r1)
+	mtlr	r9
+	REST_GPR(0, r1)
+	REST_8GPRS(2, r1)
+	REST_GPR(10, r1)
+	ld	r11,_CCR(r1)
+	mtcr	r11
+	REST_GPR(11, r1)
+	REST_2GPRS(12, r1)
+	/* restore original r1. */
+	ld	r1,GPR1(r1)
+
+	/*
+	 * Go to virtual mode and pull the HMI event information from
+	 * firmware.
+	 */
+	.globl hmi_exception_after_realmode
+hmi_exception_after_realmode:
+	SET_SCRATCH0(r13)
+	EXCEPTION_PROLOG_0(PACA_EXGEN)
+	b	hmi_exception_hv
+
 #ifdef CONFIG_PPC_POWERNV
 _GLOBAL(opal_mc_secondary_handler)
 	HMT_MEDIUM_PPR_DISCARD
diff --git a/arch/powerpc/kernel/head_44x.S b/arch/powerpc/kernel/head_44x.S
index c334f53..b5061ab 100644
--- a/arch/powerpc/kernel/head_44x.S
+++ b/arch/powerpc/kernel/head_44x.S
@@ -1210,10 +1210,12 @@
 
 	/* We configure icbi to invalidate 128 bytes at a time since the
 	 * current 32-bit kernel code isn't too happy with icache != dcache
-	 * block size
+	 * block size. We also disable the BTAC as this can cause errors
+	 * in some circumstances (see IBM Erratum 47).
 	 */
 	mfspr	r3,SPRN_CCR0
 	oris	r3,r3,0x0020
+	ori	r3,r3,0x0040
 	mtspr	SPRN_CCR0,r3
 	isync
 
diff --git a/arch/powerpc/kernel/iommu.c b/arch/powerpc/kernel/iommu.c
index f84f799..a10642a 100644
--- a/arch/powerpc/kernel/iommu.c
+++ b/arch/powerpc/kernel/iommu.c
@@ -1120,37 +1120,41 @@
 int iommu_add_device(struct device *dev)
 {
 	struct iommu_table *tbl;
-	int ret = 0;
 
-	if (WARN_ON(dev->iommu_group)) {
-		pr_warn("iommu_tce: device %s is already in iommu group %d, skipping\n",
-				dev_name(dev),
-				iommu_group_id(dev->iommu_group));
+	/*
+	 * The sysfs entries should be populated before
+	 * binding IOMMU group. If sysfs entries isn't
+	 * ready, we simply bail.
+	 */
+	if (!device_is_registered(dev))
+		return -ENOENT;
+
+	if (dev->iommu_group) {
+		pr_debug("%s: Skipping device %s with iommu group %d\n",
+			 __func__, dev_name(dev),
+			 iommu_group_id(dev->iommu_group));
 		return -EBUSY;
 	}
 
 	tbl = get_iommu_table_base(dev);
 	if (!tbl || !tbl->it_group) {
-		pr_debug("iommu_tce: skipping device %s with no tbl\n",
-				dev_name(dev));
+		pr_debug("%s: Skipping device %s with no tbl\n",
+			 __func__, dev_name(dev));
 		return 0;
 	}
 
-	pr_debug("iommu_tce: adding %s to iommu group %d\n",
-			dev_name(dev), iommu_group_id(tbl->it_group));
+	pr_debug("%s: Adding %s to iommu group %d\n",
+		 __func__, dev_name(dev),
+		 iommu_group_id(tbl->it_group));
 
 	if (PAGE_SIZE < IOMMU_PAGE_SIZE(tbl)) {
-		pr_err("iommu_tce: unsupported iommu page size.");
-		pr_err("%s has not been added\n", dev_name(dev));
+		pr_err("%s: Invalid IOMMU page size %lx (%lx) on %s\n",
+		       __func__, IOMMU_PAGE_SIZE(tbl),
+		       PAGE_SIZE, dev_name(dev));
 		return -EINVAL;
 	}
 
-	ret = iommu_group_add_device(tbl->it_group, dev);
-	if (ret < 0)
-		pr_err("iommu_tce: %s has not been added, ret=%d\n",
-				dev_name(dev), ret);
-
-	return ret;
+	return iommu_group_add_device(tbl->it_group, dev);
 }
 EXPORT_SYMBOL_GPL(iommu_add_device);
 
diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c
index 1a3b105..4e139f8 100644
--- a/arch/powerpc/kernel/prom.c
+++ b/arch/powerpc/kernel/prom.c
@@ -818,76 +818,6 @@
 }
 EXPORT_SYMBOL(cpu_to_chip_id);
 
-#ifdef CONFIG_PPC_PSERIES
-/*
- * Fix up the uninitialized fields in a new device node:
- * name, type and pci-specific fields
- */
-
-static int of_finish_dynamic_node(struct device_node *node)
-{
-	struct device_node *parent = of_get_parent(node);
-	int err = 0;
-	const phandle *ibm_phandle;
-
-	node->name = of_get_property(node, "name", NULL);
-	node->type = of_get_property(node, "device_type", NULL);
-
-	if (!node->name)
-		node->name = "<NULL>";
-	if (!node->type)
-		node->type = "<NULL>";
-
-	if (!parent) {
-		err = -ENODEV;
-		goto out;
-	}
-
-	/* We don't support that function on PowerMac, at least
-	 * not yet
-	 */
-	if (machine_is(powermac))
-		return -ENODEV;
-
-	/* fix up new node's phandle field */
-	if ((ibm_phandle = of_get_property(node, "ibm,phandle", NULL)))
-		node->phandle = *ibm_phandle;
-
-out:
-	of_node_put(parent);
-	return err;
-}
-
-static int prom_reconfig_notifier(struct notifier_block *nb,
-				  unsigned long action, void *node)
-{
-	int err;
-
-	switch (action) {
-	case OF_RECONFIG_ATTACH_NODE:
-		err = of_finish_dynamic_node(node);
-		if (err < 0)
-			printk(KERN_ERR "finish_node returned %d\n", err);
-		break;
-	default:
-		err = 0;
-		break;
-	}
-	return notifier_from_errno(err);
-}
-
-static struct notifier_block prom_reconfig_nb = {
-	.notifier_call = prom_reconfig_notifier,
-	.priority = 10, /* This one needs to run first */
-};
-
-static int __init prom_reconfig_setup(void)
-{
-	return of_reconfig_notifier_register(&prom_reconfig_nb);
-}
-__initcall(prom_reconfig_setup);
-#endif
-
 bool arch_match_cpu_phys_id(int cpu, u64 phys_id)
 {
 	return (int)phys_id == get_hard_smp_processor_id(cpu);
diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
index 1007fb8..a0738af 100644
--- a/arch/powerpc/kernel/smp.c
+++ b/arch/powerpc/kernel/smp.c
@@ -376,6 +376,11 @@
 					GFP_KERNEL, cpu_to_node(cpu));
 		zalloc_cpumask_var_node(&per_cpu(cpu_core_map, cpu),
 					GFP_KERNEL, cpu_to_node(cpu));
+		/*
+		 * numa_node_id() works after this.
+		 */
+		set_cpu_numa_node(cpu, numa_cpu_lookup_table[cpu]);
+		set_cpu_numa_mem(cpu, local_memory_node(numa_cpu_lookup_table[cpu]));
 	}
 
 	cpumask_set_cpu(boot_cpuid, cpu_sibling_mask(boot_cpuid));
@@ -723,12 +728,6 @@
 	}
 	traverse_core_siblings(cpu, true);
 
-	/*
-	 * numa_node_id() works after this.
-	 */
-	set_numa_node(numa_cpu_lookup_table[cpu]);
-	set_numa_mem(local_memory_node(numa_cpu_lookup_table[cpu]));
-
 	smp_wmb();
 	notify_cpu_starting(cpu);
 	set_cpu_online(cpu, true);
diff --git a/arch/powerpc/lib/locks.c b/arch/powerpc/lib/locks.c
index 0c9c8d7..170a034 100644
--- a/arch/powerpc/lib/locks.c
+++ b/arch/powerpc/lib/locks.c
@@ -70,12 +70,16 @@
 
 void arch_spin_unlock_wait(arch_spinlock_t *lock)
 {
+	smp_mb();
+
 	while (lock->slock) {
 		HMT_low();
 		if (SHARED_PROCESSOR)
 			__spin_yield(lock);
 	}
 	HMT_medium();
+
+	smp_mb();
 }
 
 EXPORT_SYMBOL(arch_spin_unlock_wait);
diff --git a/arch/powerpc/mm/hash_native_64.c b/arch/powerpc/mm/hash_native_64.c
index cf1d325..afc0a82 100644
--- a/arch/powerpc/mm/hash_native_64.c
+++ b/arch/powerpc/mm/hash_native_64.c
@@ -412,18 +412,18 @@
 	local_irq_restore(flags);
 }
 
-static void native_hugepage_invalidate(struct mm_struct *mm,
+static void native_hugepage_invalidate(unsigned long vsid,
+				       unsigned long addr,
 				       unsigned char *hpte_slot_array,
-				       unsigned long addr, int psize)
+				       int psize, int ssize)
 {
-	int ssize = 0, i;
-	int lock_tlbie;
+	int i;
 	struct hash_pte *hptep;
 	int actual_psize = MMU_PAGE_16M;
 	unsigned int max_hpte_count, valid;
 	unsigned long flags, s_addr = addr;
 	unsigned long hpte_v, want_v, shift;
-	unsigned long hidx, vpn = 0, vsid, hash, slot;
+	unsigned long hidx, vpn = 0, hash, slot;
 
 	shift = mmu_psize_defs[psize].shift;
 	max_hpte_count = 1U << (PMD_SHIFT - shift);
@@ -437,15 +437,6 @@
 
 		/* get the vpn */
 		addr = s_addr + (i * (1ul << shift));
-		if (!is_kernel_addr(addr)) {
-			ssize = user_segment_size(addr);
-			vsid = get_vsid(mm->context.id, addr, ssize);
-			WARN_ON(vsid == 0);
-		} else {
-			vsid = get_kernel_vsid(addr, mmu_kernel_ssize);
-			ssize = mmu_kernel_ssize;
-		}
-
 		vpn = hpt_vpn(addr, vsid, ssize);
 		hash = hpt_hash(vpn, shift, ssize);
 		if (hidx & _PTEIDX_SECONDARY)
@@ -465,22 +456,13 @@
 		else
 			/* Invalidate the hpte. NOTE: this also unlocks it */
 			hptep->v = 0;
+		/*
+		 * We need to do tlb invalidate for all the address, tlbie
+		 * instruction compares entry_VA in tlb with the VA specified
+		 * here
+		 */
+		tlbie(vpn, psize, actual_psize, ssize, 0);
 	}
-	/*
-	 * Since this is a hugepage, we just need a single tlbie.
-	 * use the last vpn.
-	 */
-	lock_tlbie = !mmu_has_feature(MMU_FTR_LOCKLESS_TLBIE);
-	if (lock_tlbie)
-		raw_spin_lock(&native_tlbie_lock);
-
-	asm volatile("ptesync":::"memory");
-	__tlbie(vpn, psize, actual_psize, ssize);
-	asm volatile("eieio; tlbsync; ptesync":::"memory");
-
-	if (lock_tlbie)
-		raw_spin_unlock(&native_tlbie_lock);
-
 	local_irq_restore(flags);
 }
 
diff --git a/arch/powerpc/mm/hugepage-hash64.c b/arch/powerpc/mm/hugepage-hash64.c
index 826893f..5f5e632 100644
--- a/arch/powerpc/mm/hugepage-hash64.c
+++ b/arch/powerpc/mm/hugepage-hash64.c
@@ -18,6 +18,57 @@
 #include <linux/mm.h>
 #include <asm/machdep.h>
 
+static void invalidate_old_hpte(unsigned long vsid, unsigned long addr,
+				pmd_t *pmdp, unsigned int psize, int ssize)
+{
+	int i, max_hpte_count, valid;
+	unsigned long s_addr;
+	unsigned char *hpte_slot_array;
+	unsigned long hidx, shift, vpn, hash, slot;
+
+	s_addr = addr & HPAGE_PMD_MASK;
+	hpte_slot_array = get_hpte_slot_array(pmdp);
+	/*
+	 * IF we try to do a HUGE PTE update after a withdraw is done.
+	 * we will find the below NULL. This happens when we do
+	 * split_huge_page_pmd
+	 */
+	if (!hpte_slot_array)
+		return;
+
+	if (ppc_md.hugepage_invalidate)
+		return ppc_md.hugepage_invalidate(vsid, s_addr, hpte_slot_array,
+						  psize, ssize);
+	/*
+	 * No bluk hpte removal support, invalidate each entry
+	 */
+	shift = mmu_psize_defs[psize].shift;
+	max_hpte_count = HPAGE_PMD_SIZE >> shift;
+	for (i = 0; i < max_hpte_count; i++) {
+		/*
+		 * 8 bits per each hpte entries
+		 * 000| [ secondary group (one bit) | hidx (3 bits) | valid bit]
+		 */
+		valid = hpte_valid(hpte_slot_array, i);
+		if (!valid)
+			continue;
+		hidx =  hpte_hash_index(hpte_slot_array, i);
+
+		/* get the vpn */
+		addr = s_addr + (i * (1ul << shift));
+		vpn = hpt_vpn(addr, vsid, ssize);
+		hash = hpt_hash(vpn, shift, ssize);
+		if (hidx & _PTEIDX_SECONDARY)
+			hash = ~hash;
+
+		slot = (hash & htab_hash_mask) * HPTES_PER_GROUP;
+		slot += hidx & _PTEIDX_GROUP_IX;
+		ppc_md.hpte_invalidate(slot, vpn, psize,
+				       MMU_PAGE_16M, ssize, 0);
+	}
+}
+
+
 int __hash_page_thp(unsigned long ea, unsigned long access, unsigned long vsid,
 		    pmd_t *pmdp, unsigned long trap, int local, int ssize,
 		    unsigned int psize)
@@ -33,7 +84,9 @@
 	 * atomically mark the linux large page PMD busy and dirty
 	 */
 	do {
-		old_pmd = pmd_val(*pmdp);
+		pmd_t pmd = ACCESS_ONCE(*pmdp);
+
+		old_pmd = pmd_val(pmd);
 		/* If PMD busy, retry the access */
 		if (unlikely(old_pmd & _PAGE_BUSY))
 			return 0;
@@ -85,6 +138,15 @@
 	vpn = hpt_vpn(ea, vsid, ssize);
 	hash = hpt_hash(vpn, shift, ssize);
 	hpte_slot_array = get_hpte_slot_array(pmdp);
+	if (psize == MMU_PAGE_4K) {
+		/*
+		 * invalidate the old hpte entry if we have that mapped via 64K
+		 * base page size. This is because demote_segment won't flush
+		 * hash page table entries.
+		 */
+		if ((old_pmd & _PAGE_HASHPTE) && !(old_pmd & _PAGE_COMBO))
+			invalidate_old_hpte(vsid, ea, pmdp, MMU_PAGE_64K, ssize);
+	}
 
 	valid = hpte_valid(hpte_slot_array, index);
 	if (valid) {
@@ -107,11 +169,8 @@
 			 * safely update this here.
 			 */
 			valid = 0;
-			new_pmd &= ~_PAGE_HPTEFLAGS;
 			hpte_slot_array[index] = 0;
-		} else
-			/* clear the busy bits and set the hash pte bits */
-			new_pmd = (new_pmd & ~_PAGE_HPTEFLAGS) | _PAGE_HASHPTE;
+		}
 	}
 
 	if (!valid) {
@@ -119,11 +178,7 @@
 
 		/* insert new entry */
 		pa = pmd_pfn(__pmd(old_pmd)) << PAGE_SHIFT;
-repeat:
-		hpte_group = ((hash & htab_hash_mask) * HPTES_PER_GROUP) & ~0x7UL;
-
-		/* clear the busy bits and set the hash pte bits */
-		new_pmd = (new_pmd & ~_PAGE_HPTEFLAGS) | _PAGE_HASHPTE;
+		new_pmd |= _PAGE_HASHPTE;
 
 		/* Add in WIMG bits */
 		rflags |= (new_pmd & (_PAGE_WRITETHRU | _PAGE_NO_CACHE |
@@ -132,6 +187,8 @@
 		 * enable the memory coherence always
 		 */
 		rflags |= HPTE_R_M;
+repeat:
+		hpte_group = ((hash & htab_hash_mask) * HPTES_PER_GROUP) & ~0x7UL;
 
 		/* Insert into the hash table, primary slot */
 		slot = ppc_md.hpte_insert(hpte_group, vpn, pa, rflags, 0,
@@ -172,8 +229,17 @@
 		mark_hpte_slot_valid(hpte_slot_array, index, slot);
 	}
 	/*
-	 * No need to use ldarx/stdcx here
+	 * Mark the pte with _PAGE_COMBO, if we are trying to hash it with
+	 * base page size 4k.
 	 */
+	if (psize == MMU_PAGE_4K)
+		new_pmd |= _PAGE_COMBO;
+	/*
+	 * The hpte valid is stored in the pgtable whose address is in the
+	 * second half of the PMD. Order this against clearing of the busy bit in
+	 * huge pmd.
+	 */
+	smp_wmb();
 	*pmdp = __pmd(new_pmd & ~_PAGE_BUSY);
 	return 0;
 }
diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
index d3e9a78..d7737a5 100644
--- a/arch/powerpc/mm/numa.c
+++ b/arch/powerpc/mm/numa.c
@@ -1049,7 +1049,7 @@
 
 void __init do_init_bootmem(void)
 {
-	int nid;
+	int nid, cpu;
 
 	min_low_pfn = 0;
 	max_low_pfn = memblock_end_of_DRAM() >> PAGE_SHIFT;
@@ -1122,8 +1122,15 @@
 
 	reset_numa_cpu_lookup_table();
 	register_cpu_notifier(&ppc64_numa_nb);
-	cpu_numa_callback(&ppc64_numa_nb, CPU_UP_PREPARE,
-			  (void *)(unsigned long)boot_cpuid);
+	/*
+	 * We need the numa_cpu_lookup_table to be accurate for all CPUs,
+	 * even before we online them, so that we can use cpu_to_{node,mem}
+	 * early in boot, cf. smp_prepare_cpus().
+	 */
+	for_each_possible_cpu(cpu) {
+		cpu_numa_callback(&ppc64_numa_nb, CPU_UP_PREPARE,
+				  (void *)(unsigned long)cpu);
+	}
 }
 
 void __init paging_init(void)
diff --git a/arch/powerpc/mm/pgtable_64.c b/arch/powerpc/mm/pgtable_64.c
index 3b3c4d3..c8d709a 100644
--- a/arch/powerpc/mm/pgtable_64.c
+++ b/arch/powerpc/mm/pgtable_64.c
@@ -54,6 +54,9 @@
 
 #include "mmu_decl.h"
 
+#define CREATE_TRACE_POINTS
+#include <trace/events/thp.h>
+
 /* Some sanity checking */
 #if TASK_SIZE_USER64 > PGTABLE_RANGE
 #error TASK_SIZE_USER64 exceeds pagetable range
@@ -537,8 +540,9 @@
 	old = pmd_val(*pmdp);
 	*pmdp = __pmd((old & ~clr) | set);
 #endif
+	trace_hugepage_update(addr, old, clr, set);
 	if (old & _PAGE_HASHPTE)
-		hpte_do_hugepage_flush(mm, addr, pmdp);
+		hpte_do_hugepage_flush(mm, addr, pmdp, old);
 	return old;
 }
 
@@ -642,10 +646,11 @@
 	 * If we didn't had the splitting flag set, go and flush the
 	 * HPTE entries.
 	 */
+	trace_hugepage_splitting(address, old);
 	if (!(old & _PAGE_SPLITTING)) {
 		/* We need to flush the hpte */
 		if (old & _PAGE_HASHPTE)
-			hpte_do_hugepage_flush(vma->vm_mm, address, pmdp);
+			hpte_do_hugepage_flush(vma->vm_mm, address, pmdp, old);
 	}
 	/*
 	 * This ensures that generic code that rely on IRQ disabling
@@ -709,6 +714,7 @@
 	assert_spin_locked(&mm->page_table_lock);
 	WARN_ON(!pmd_trans_huge(pmd));
 #endif
+	trace_hugepage_set_pmd(addr, pmd);
 	return set_pte_at(mm, addr, pmdp_ptep(pmdp), pmd_pte(pmd));
 }
 
@@ -723,7 +729,7 @@
  * neesd to be flushed.
  */
 void hpte_do_hugepage_flush(struct mm_struct *mm, unsigned long addr,
-			    pmd_t *pmdp)
+			    pmd_t *pmdp, unsigned long old_pmd)
 {
 	int ssize, i;
 	unsigned long s_addr;
@@ -745,12 +751,29 @@
 	if (!hpte_slot_array)
 		return;
 
-	/* get the base page size */
+	/* get the base page size,vsid and segment size */
+#ifdef CONFIG_DEBUG_VM
 	psize = get_slice_psize(mm, s_addr);
+	BUG_ON(psize == MMU_PAGE_16M);
+#endif
+	if (old_pmd & _PAGE_COMBO)
+		psize = MMU_PAGE_4K;
+	else
+		psize = MMU_PAGE_64K;
+
+	if (!is_kernel_addr(s_addr)) {
+		ssize = user_segment_size(s_addr);
+		vsid = get_vsid(mm->context.id, s_addr, ssize);
+		WARN_ON(vsid == 0);
+	} else {
+		vsid = get_kernel_vsid(s_addr, mmu_kernel_ssize);
+		ssize = mmu_kernel_ssize;
+	}
 
 	if (ppc_md.hugepage_invalidate)
-		return ppc_md.hugepage_invalidate(mm, hpte_slot_array,
-						  s_addr, psize);
+		return ppc_md.hugepage_invalidate(vsid, s_addr,
+						  hpte_slot_array,
+						  psize, ssize);
 	/*
 	 * No bluk hpte removal support, invalidate each entry
 	 */
@@ -768,15 +791,6 @@
 
 		/* get the vpn */
 		addr = s_addr + (i * (1ul << shift));
-		if (!is_kernel_addr(addr)) {
-			ssize = user_segment_size(addr);
-			vsid = get_vsid(mm->context.id, addr, ssize);
-			WARN_ON(vsid == 0);
-		} else {
-			vsid = get_kernel_vsid(addr, mmu_kernel_ssize);
-			ssize = mmu_kernel_ssize;
-		}
-
 		vpn = hpt_vpn(addr, vsid, ssize);
 		hash = hpt_hash(vpn, shift, ssize);
 		if (hidx & _PTEIDX_SECONDARY)
diff --git a/arch/powerpc/mm/tlb_hash64.c b/arch/powerpc/mm/tlb_hash64.c
index c99f651..d2a94b8 100644
--- a/arch/powerpc/mm/tlb_hash64.c
+++ b/arch/powerpc/mm/tlb_hash64.c
@@ -30,6 +30,8 @@
 #include <asm/tlb.h>
 #include <asm/bug.h>
 
+#include <trace/events/thp.h>
+
 DEFINE_PER_CPU(struct ppc64_tlb_batch, ppc64_tlb_batch);
 
 /*
@@ -213,10 +215,12 @@
 		if (ptep == NULL)
 			continue;
 		pte = pte_val(*ptep);
+		if (hugepage_shift)
+			trace_hugepage_invalidate(start, pte_val(pte));
 		if (!(pte & _PAGE_HASHPTE))
 			continue;
 		if (unlikely(hugepage_shift && pmd_trans_huge(*(pmd_t *)pte)))
-			hpte_do_hugepage_flush(mm, start, (pmd_t *)pte);
+			hpte_do_hugepage_flush(mm, start, (pmd_t *)ptep, pte);
 		else
 			hpte_need_flush(mm, start, ptep, pte, 0);
 	}
diff --git a/arch/powerpc/mm/tlb_nohash.c b/arch/powerpc/mm/tlb_nohash.c
index 92cb18d..f38ea4d 100644
--- a/arch/powerpc/mm/tlb_nohash.c
+++ b/arch/powerpc/mm/tlb_nohash.c
@@ -581,42 +581,10 @@
 /*
  * Early initialization of the MMU TLB code
  */
-static void __early_init_mmu(int boot_cpu)
+static void early_init_this_mmu(void)
 {
 	unsigned int mas4;
 
-	/* XXX This will have to be decided at runtime, but right
-	 * now our boot and TLB miss code hard wires it. Ideally
-	 * we should find out a suitable page size and patch the
-	 * TLB miss code (either that or use the PACA to store
-	 * the value we want)
-	 */
-	mmu_linear_psize = MMU_PAGE_1G;
-
-	/* XXX This should be decided at runtime based on supported
-	 * page sizes in the TLB, but for now let's assume 16M is
-	 * always there and a good fit (which it probably is)
-	 *
-	 * Freescale booke only supports 4K pages in TLB0, so use that.
-	 */
-	if (mmu_has_feature(MMU_FTR_TYPE_FSL_E))
-		mmu_vmemmap_psize = MMU_PAGE_4K;
-	else
-		mmu_vmemmap_psize = MMU_PAGE_16M;
-
-	/* XXX This code only checks for TLB 0 capabilities and doesn't
-	 *     check what page size combos are supported by the HW. It
-	 *     also doesn't handle the case where a separate array holds
-	 *     the IND entries from the array loaded by the PT.
-	 */
-	if (boot_cpu) {
-		/* Look for supported page sizes */
-		setup_page_sizes();
-
-		/* Look for HW tablewalk support */
-		setup_mmu_htw();
-	}
-
 	/* Set MAS4 based on page table setting */
 
 	mas4 = 0x4 << MAS4_WIMGED_SHIFT;
@@ -650,11 +618,6 @@
 	}
 	mtspr(SPRN_MAS4, mas4);
 
-	/* Set the global containing the top of the linear mapping
-	 * for use by the TLB miss code
-	 */
-	linear_map_top = memblock_end_of_DRAM();
-
 #ifdef CONFIG_PPC_FSL_BOOK3E
 	if (mmu_has_feature(MMU_FTR_TYPE_FSL_E)) {
 		unsigned int num_cams;
@@ -662,10 +625,49 @@
 		/* use a quarter of the TLBCAM for bolted linear map */
 		num_cams = (mfspr(SPRN_TLB1CFG) & TLBnCFG_N_ENTRY) / 4;
 		linear_map_top = map_mem_in_cams(linear_map_top, num_cams);
+	}
+#endif
 
-		/* limit memory so we dont have linear faults */
-		memblock_enforce_memory_limit(linear_map_top);
+	/* A sync won't hurt us after mucking around with
+	 * the MMU configuration
+	 */
+	mb();
+}
 
+static void __init early_init_mmu_global(void)
+{
+	/* XXX This will have to be decided at runtime, but right
+	 * now our boot and TLB miss code hard wires it. Ideally
+	 * we should find out a suitable page size and patch the
+	 * TLB miss code (either that or use the PACA to store
+	 * the value we want)
+	 */
+	mmu_linear_psize = MMU_PAGE_1G;
+
+	/* XXX This should be decided at runtime based on supported
+	 * page sizes in the TLB, but for now let's assume 16M is
+	 * always there and a good fit (which it probably is)
+	 *
+	 * Freescale booke only supports 4K pages in TLB0, so use that.
+	 */
+	if (mmu_has_feature(MMU_FTR_TYPE_FSL_E))
+		mmu_vmemmap_psize = MMU_PAGE_4K;
+	else
+		mmu_vmemmap_psize = MMU_PAGE_16M;
+
+	/* XXX This code only checks for TLB 0 capabilities and doesn't
+	 *     check what page size combos are supported by the HW. It
+	 *     also doesn't handle the case where a separate array holds
+	 *     the IND entries from the array loaded by the PT.
+	 */
+	/* Look for supported page sizes */
+	setup_page_sizes();
+
+	/* Look for HW tablewalk support */
+	setup_mmu_htw();
+
+#ifdef CONFIG_PPC_FSL_BOOK3E
+	if (mmu_has_feature(MMU_FTR_TYPE_FSL_E)) {
 		if (book3e_htw_mode == PPC_HTW_NONE) {
 			extlb_level_exc = EX_TLB_SIZE;
 			patch_exception(0x1c0, exc_data_tlb_miss_bolted_book3e);
@@ -675,22 +677,41 @@
 	}
 #endif
 
-	/* A sync won't hurt us after mucking around with
-	 * the MMU configuration
+	/* Set the global containing the top of the linear mapping
+	 * for use by the TLB miss code
 	 */
-	mb();
+	linear_map_top = memblock_end_of_DRAM();
+}
+
+static void __init early_mmu_set_memory_limit(void)
+{
+#ifdef CONFIG_PPC_FSL_BOOK3E
+	if (mmu_has_feature(MMU_FTR_TYPE_FSL_E)) {
+		/*
+		 * Limit memory so we dont have linear faults.
+		 * Unlike memblock_set_current_limit, which limits
+		 * memory available during early boot, this permanently
+		 * reduces the memory available to Linux.  We need to
+		 * do this because highmem is not supported on 64-bit.
+		 */
+		memblock_enforce_memory_limit(linear_map_top);
+	}
+#endif
 
 	memblock_set_current_limit(linear_map_top);
 }
 
+/* boot cpu only */
 void __init early_init_mmu(void)
 {
-	__early_init_mmu(1);
+	early_init_mmu_global();
+	early_init_this_mmu();
+	early_mmu_set_memory_limit();
 }
 
 void early_init_mmu_secondary(void)
 {
-	__early_init_mmu(0);
+	early_init_this_mmu();
 }
 
 void setup_initial_memory_limit(phys_addr_t first_memblock_base,
diff --git a/arch/powerpc/perf/hv-24x7.c b/arch/powerpc/perf/hv-24x7.c
index 66d0f17..70d4f74 100644
--- a/arch/powerpc/perf/hv-24x7.c
+++ b/arch/powerpc/perf/hv-24x7.c
@@ -223,7 +223,7 @@
 		pr_err("h_get_24x7_catalog_page(ver=%lld, page=%lld) failed:"
 		       " rc=%ld\n",
 		       catalog_version_num, page_offset, hret);
-	kfree(page);
+	kmem_cache_free(hv_page_cache, page);
 
 	pr_devel("catalog_read: offset=%lld(%lld) count=%zu(%zu) catalog_len=%zu(%zu) => %zd\n",
 			offset, page_offset, count, page_count, catalog_len,
diff --git a/arch/powerpc/platforms/powermac/feature.c b/arch/powerpc/platforms/powermac/feature.c
index 1413e72..4882bfd 100644
--- a/arch/powerpc/platforms/powermac/feature.c
+++ b/arch/powerpc/platforms/powermac/feature.c
@@ -2805,25 +2805,20 @@
 		/* Enable GMAC for now for PCI probing. It will be disabled
 		 * later on after PCI probe
 		 */
-		np = of_find_node_by_name(NULL, "ethernet");
-		while(np) {
+		for_each_node_by_name(np, "ethernet")
 			if (of_device_is_compatible(np, "K2-GMAC"))
 				g5_gmac_enable(np, 0, 1);
-			np = of_find_node_by_name(np, "ethernet");
-		}
 
 		/* Enable FW before PCI probe. Will be disabled later on
 		 * Note: We should have a batter way to check that we are
 		 * dealing with uninorth internal cell and not a PCI cell
 		 * on the external PCI. The code below works though.
 		 */
-		np = of_find_node_by_name(NULL, "firewire");
-		while(np) {
+		for_each_node_by_name(np, "firewire") {
 			if (of_device_is_compatible(np, "pci106b,5811")) {
 				macio_chips[0].flags |= MACIO_FLAG_FW_SUPPORTED;
 				g5_fw_enable(np, 0, 1);
 			}
-			np = of_find_node_by_name(np, "firewire");
 		}
 	}
 #else /* CONFIG_PPC64 */
@@ -2834,13 +2829,11 @@
 		/* Enable GMAC for now for PCI probing. It will be disabled
 		 * later on after PCI probe
 		 */
-		np = of_find_node_by_name(NULL, "ethernet");
-		while(np) {
+		for_each_node_by_name(np, "ethernet") {
 			if (np->parent
 			    && of_device_is_compatible(np->parent, "uni-north")
 			    && of_device_is_compatible(np, "gmac"))
 				core99_gmac_enable(np, 0, 1);
-			np = of_find_node_by_name(np, "ethernet");
 		}
 
 		/* Enable FW before PCI probe. Will be disabled later on
@@ -2848,8 +2841,7 @@
 		 * dealing with uninorth internal cell and not a PCI cell
 		 * on the external PCI. The code below works though.
 		 */
-		np = of_find_node_by_name(NULL, "firewire");
-		while(np) {
+		for_each_node_by_name(np, "firewire") {
 			if (np->parent
 			    && of_device_is_compatible(np->parent, "uni-north")
 			    && (of_device_is_compatible(np, "pci106b,18") ||
@@ -2858,18 +2850,16 @@
 				macio_chips[0].flags |= MACIO_FLAG_FW_SUPPORTED;
 				core99_firewire_enable(np, 0, 1);
 			}
-			np = of_find_node_by_name(np, "firewire");
 		}
 
 		/* Enable ATA-100 before PCI probe. */
 		np = of_find_node_by_name(NULL, "ata-6");
-		while(np) {
+		for_each_node_by_name(np, "ata-6") {
 			if (np->parent
 			    && of_device_is_compatible(np->parent, "uni-north")
 			    && of_device_is_compatible(np, "kauai-ata")) {
 				core99_ata100_enable(np, 1);
 			}
-			np = of_find_node_by_name(np, "ata-6");
 		}
 
 		/* Switch airport off */
diff --git a/arch/powerpc/platforms/powermac/pci.c b/arch/powerpc/platforms/powermac/pci.c
index cf7009b..7e868cc 100644
--- a/arch/powerpc/platforms/powermac/pci.c
+++ b/arch/powerpc/platforms/powermac/pci.c
@@ -698,7 +698,7 @@
 {
 	struct device_node *nec;
 
-	for (nec = NULL; (nec = of_find_node_by_name(nec, "usb")) != NULL;) {
+	for_each_node_by_name(nec, "usb") {
 		struct pci_controller *hose;
 		u32 data;
 		const u32 *prop;
diff --git a/arch/powerpc/platforms/powermac/smp.c b/arch/powerpc/platforms/powermac/smp.c
index 5cbd4d6..af094ae 100644
--- a/arch/powerpc/platforms/powermac/smp.c
+++ b/arch/powerpc/platforms/powermac/smp.c
@@ -577,7 +577,7 @@
 	int ok;
 
 	/* Look for the clock chip */
-	while ((cc = of_find_node_by_name(cc, "i2c-hwclock")) != NULL) {
+	for_each_node_by_name(cc, "i2c-hwclock") {
 		p = of_get_parent(cc);
 		ok = p && of_device_is_compatible(p, "uni-n-i2c");
 		of_node_put(p);
diff --git a/arch/powerpc/platforms/powermac/udbg_adb.c b/arch/powerpc/platforms/powermac/udbg_adb.c
index 44e0b55..366bd22 100644
--- a/arch/powerpc/platforms/powermac/udbg_adb.c
+++ b/arch/powerpc/platforms/powermac/udbg_adb.c
@@ -191,7 +191,7 @@
 	 * of type "adb". If not, we return a failure, but we keep the
 	 * bext output set for now
 	 */
-	for (np = NULL; (np = of_find_node_by_name(np, "keyboard")) != NULL;) {
+	for_each_node_by_name(np, "keyboard") {
 		struct device_node *parent = of_get_parent(np);
 		int found = (parent && strcmp(parent->type, "adb") == 0);
 		of_node_put(parent);
diff --git a/arch/powerpc/platforms/powernv/opal-wrappers.S b/arch/powerpc/platforms/powernv/opal-wrappers.S
index a328be4..2e6ce1b 100644
--- a/arch/powerpc/platforms/powernv/opal-wrappers.S
+++ b/arch/powerpc/platforms/powernv/opal-wrappers.S
@@ -245,3 +245,5 @@
 OPAL_CALL(opal_get_param,			OPAL_GET_PARAM);
 OPAL_CALL(opal_set_param,			OPAL_SET_PARAM);
 OPAL_CALL(opal_handle_hmi,			OPAL_HANDLE_HMI);
+OPAL_CALL(opal_register_dump_region,		OPAL_REGISTER_DUMP_REGION);
+OPAL_CALL(opal_unregister_dump_region,		OPAL_UNREGISTER_DUMP_REGION);
diff --git a/arch/powerpc/platforms/powernv/opal.c b/arch/powerpc/platforms/powernv/opal.c
index f0a01a4..b44eec3 100644
--- a/arch/powerpc/platforms/powernv/opal.c
+++ b/arch/powerpc/platforms/powernv/opal.c
@@ -605,6 +605,24 @@
 	return 0;
 }
 
+static void __init opal_dump_region_init(void)
+{
+	void *addr;
+	uint64_t size;
+	int rc;
+
+	/* Register kernel log buffer */
+	addr = log_buf_addr_get();
+	size = log_buf_len_get();
+	rc = opal_register_dump_region(OPAL_DUMP_REGION_LOG_BUF,
+				       __pa(addr), size);
+	/* Don't warn if this is just an older OPAL that doesn't
+	 * know about that call
+	 */
+	if (rc && rc != OPAL_UNSUPPORTED)
+		pr_warn("DUMP: Failed to register kernel log buffer. "
+			"rc = %d\n", rc);
+}
 static int __init opal_init(void)
 {
 	struct device_node *np, *consoles;
@@ -654,6 +672,8 @@
 	/* Create "opal" kobject under /sys/firmware */
 	rc = opal_sysfs_init();
 	if (rc == 0) {
+		/* Setup dump region interface */
+		opal_dump_region_init();
 		/* Setup error log interface */
 		rc = opal_elog_init();
 		/* Setup code update interface */
@@ -694,6 +714,9 @@
 		else
 			mdelay(10);
 	}
+
+	/* Unregister memory dump region */
+	opal_unregister_dump_region(OPAL_DUMP_REGION_LOG_BUF);
 }
 
 /* Export this so that test modules can use it */
diff --git a/arch/powerpc/platforms/powernv/pci-ioda.c b/arch/powerpc/platforms/powernv/pci-ioda.c
index b136108..df241b1 100644
--- a/arch/powerpc/platforms/powernv/pci-ioda.c
+++ b/arch/powerpc/platforms/powernv/pci-ioda.c
@@ -857,7 +857,7 @@
 
 	pe = &phb->ioda.pe_array[pdn->pe_number];
 	WARN_ON(get_dma_ops(&pdev->dev) != &dma_iommu_ops);
-	set_iommu_table_base(&pdev->dev, &pe->tce32_table);
+	set_iommu_table_base_and_group(&pdev->dev, &pe->tce32_table);
 }
 
 static int pnv_pci_ioda_dma_set_mask(struct pnv_phb *phb,
diff --git a/arch/powerpc/platforms/pseries/hotplug-memory.c b/arch/powerpc/platforms/pseries/hotplug-memory.c
index 7995135..c904583 100644
--- a/arch/powerpc/platforms/pseries/hotplug-memory.c
+++ b/arch/powerpc/platforms/pseries/hotplug-memory.c
@@ -146,7 +146,7 @@
 }
 static inline int pseries_remove_mem_node(struct device_node *np)
 {
-	return -EOPNOTSUPP;
+	return 0;
 }
 #endif /* CONFIG_MEMORY_HOTREMOVE */
 
@@ -194,7 +194,7 @@
 	if (!memblock_size)
 		return -EINVAL;
 
-	p = (u32 *)of_get_property(pr->dn, "ibm,dynamic-memory", NULL);
+	p = (u32 *) pr->old_prop->value;
 	if (!p)
 		return -EINVAL;
 
diff --git a/arch/powerpc/platforms/pseries/hvcserver.c b/arch/powerpc/platforms/pseries/hvcserver.c
index 4557e91..eedb645 100644
--- a/arch/powerpc/platforms/pseries/hvcserver.c
+++ b/arch/powerpc/platforms/pseries/hvcserver.c
@@ -163,8 +163,8 @@
 			return retval;
 		}
 
-		last_p_partition_ID = pi_buff[0];
-		last_p_unit_address = pi_buff[1];
+		last_p_partition_ID = be64_to_cpu(pi_buff[0]);
+		last_p_unit_address = be64_to_cpu(pi_buff[1]);
 
 		/* This indicates that there are no further partners */
 		if (last_p_partition_ID == ~0UL
diff --git a/arch/powerpc/platforms/pseries/iommu.c b/arch/powerpc/platforms/pseries/iommu.c
index 33b552f..4642d6a 100644
--- a/arch/powerpc/platforms/pseries/iommu.c
+++ b/arch/powerpc/platforms/pseries/iommu.c
@@ -721,13 +721,13 @@
 
 early_param("disable_ddw", disable_ddw_setup);
 
-static void remove_ddw(struct device_node *np)
+static void remove_ddw(struct device_node *np, bool remove_prop)
 {
 	struct dynamic_dma_window_prop *dwp;
 	struct property *win64;
 	const u32 *ddw_avail;
 	u64 liobn;
-	int len, ret;
+	int len, ret = 0;
 
 	ddw_avail = of_get_property(np, "ibm,ddw-applicable", &len);
 	win64 = of_find_property(np, DIRECT64_PROPNAME, NULL);
@@ -761,7 +761,8 @@
 			np->full_name, ret, ddw_avail[2], liobn);
 
 delprop:
-	ret = of_remove_property(np, win64);
+	if (remove_prop)
+		ret = of_remove_property(np, win64);
 	if (ret)
 		pr_warning("%s: failed to remove direct window property: %d\n",
 			np->full_name, ret);
@@ -805,7 +806,7 @@
 		window = kzalloc(sizeof(*window), GFP_KERNEL);
 		if (!window || len < sizeof(struct dynamic_dma_window_prop)) {
 			kfree(window);
-			remove_ddw(pdn);
+			remove_ddw(pdn, true);
 			continue;
 		}
 
@@ -1045,7 +1046,7 @@
 	kfree(window);
 
 out_clear_window:
-	remove_ddw(pdn);
+	remove_ddw(pdn, true);
 
 out_free_prop:
 	kfree(win64->name);
@@ -1255,7 +1256,14 @@
 
 	switch (action) {
 	case OF_RECONFIG_DETACH_NODE:
-		remove_ddw(np);
+		/*
+		 * Removing the property will invoke the reconfig
+		 * notifier again, which causes dead-lock on the
+		 * read-write semaphore of the notifier chain. So
+		 * we have to remove the property when releasing
+		 * the device node.
+		 */
+		remove_ddw(np, false);
 		if (pci && pci->iommu_table)
 			iommu_free_table(pci->iommu_table, np->full_name);
 
diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c
index fbfcef5..34e6423 100644
--- a/arch/powerpc/platforms/pseries/lpar.c
+++ b/arch/powerpc/platforms/pseries/lpar.c
@@ -431,16 +431,17 @@
 		spin_unlock_irqrestore(&pSeries_lpar_tlbie_lock, flags);
 }
 
-static void pSeries_lpar_hugepage_invalidate(struct mm_struct *mm,
-				       unsigned char *hpte_slot_array,
-				       unsigned long addr, int psize)
+static void pSeries_lpar_hugepage_invalidate(unsigned long vsid,
+					     unsigned long addr,
+					     unsigned char *hpte_slot_array,
+					     int psize, int ssize)
 {
-	int ssize = 0, i, index = 0;
+	int i, index = 0;
 	unsigned long s_addr = addr;
 	unsigned int max_hpte_count, valid;
 	unsigned long vpn_array[PPC64_HUGE_HPTE_BATCH];
 	unsigned long slot_array[PPC64_HUGE_HPTE_BATCH];
-	unsigned long shift, hidx, vpn = 0, vsid, hash, slot;
+	unsigned long shift, hidx, vpn = 0, hash, slot;
 
 	shift = mmu_psize_defs[psize].shift;
 	max_hpte_count = 1U << (PMD_SHIFT - shift);
@@ -453,15 +454,6 @@
 
 		/* get the vpn */
 		addr = s_addr + (i * (1ul << shift));
-		if (!is_kernel_addr(addr)) {
-			ssize = user_segment_size(addr);
-			vsid = get_vsid(mm->context.id, addr, ssize);
-			WARN_ON(vsid == 0);
-		} else {
-			vsid = get_kernel_vsid(addr, mmu_kernel_ssize);
-			ssize = mmu_kernel_ssize;
-		}
-
 		vpn = hpt_vpn(addr, vsid, ssize);
 		hash = hpt_hash(vpn, shift, ssize);
 		if (hidx & _PTEIDX_SECONDARY)
diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c
index cfe8a63..e724d31 100644
--- a/arch/powerpc/platforms/pseries/setup.c
+++ b/arch/powerpc/platforms/pseries/setup.c
@@ -232,8 +232,7 @@
 	struct device_node *np;
 	const char *typep;
 
-	for (np = NULL; (np = of_find_node_by_name(np,
-						   "interrupt-controller"));) {
+	for_each_node_by_name(np, "interrupt-controller") {
 		typep = of_get_property(np, "compatible", NULL);
 		if (strstr(typep, "open-pic")) {
 			pSeries_mpic_node = of_node_get(np);
diff --git a/arch/powerpc/xmon/xmon.c b/arch/powerpc/xmon/xmon.c
index 8d198b5..b988b5a 100644
--- a/arch/powerpc/xmon/xmon.c
+++ b/arch/powerpc/xmon/xmon.c
@@ -24,6 +24,7 @@
 #include <linux/interrupt.h>
 #include <linux/irq.h>
 #include <linux/bug.h>
+#include <linux/nmi.h>
 
 #include <asm/ptrace.h>
 #include <asm/string.h>
@@ -374,6 +375,7 @@
 #endif
 
 	local_irq_save(flags);
+	hard_irq_disable();
 
 	bp = in_breakpoint_table(regs->nip, &offset);
 	if (bp != NULL) {
@@ -558,6 +560,7 @@
 #endif
 	insert_cpu_bpts();
 
+	touch_nmi_watchdog();
 	local_irq_restore(flags);
 
 	return cmd != 'X' && cmd != EOF;
diff --git a/arch/x86/xen/grant-table.c b/arch/x86/xen/grant-table.c
index c041304..1580e7a 100644
--- a/arch/x86/xen/grant-table.c
+++ b/arch/x86/xen/grant-table.c
@@ -118,6 +118,7 @@
 {
 	struct page **pages;
 	xen_pfn_t *pfns;
+	void *vaddr;
 	int rc;
 	unsigned int i;
 	unsigned long nr_grant_frames = gnttab_max_grant_frames();
@@ -143,21 +144,20 @@
 	for (i = 0; i < nr_grant_frames; i++)
 		pfns[i] = page_to_pfn(pages[i]);
 
-	rc = arch_gnttab_map_shared(pfns, nr_grant_frames, nr_grant_frames,
-				    &xen_auto_xlat_grant_frames.vaddr);
-
-	if (rc) {
+	vaddr = vmap(pages, nr_grant_frames, 0, PAGE_KERNEL);
+	if (!vaddr) {
 		pr_warn("%s Couldn't map %ld pfns rc:%d\n", __func__,
 			nr_grant_frames, rc);
 		free_xenballooned_pages(nr_grant_frames, pages);
 		kfree(pages);
 		kfree(pfns);
-		return rc;
+		return -ENOMEM;
 	}
 	kfree(pages);
 
 	xen_auto_xlat_grant_frames.pfn = pfns;
 	xen_auto_xlat_grant_frames.count = nr_grant_frames;
+	xen_auto_xlat_grant_frames.vaddr = vaddr;
 
 	return 0;
 }
diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c
index 7b78f88..5718b0b 100644
--- a/arch/x86/xen/time.c
+++ b/arch/x86/xen/time.c
@@ -444,7 +444,7 @@
 
 	irq = bind_virq_to_irqhandler(VIRQ_TIMER, cpu, xen_timer_interrupt,
 				      IRQF_PERCPU|IRQF_NOBALANCING|IRQF_TIMER|
-				      IRQF_FORCE_RESUME,
+				      IRQF_FORCE_RESUME|IRQF_EARLY_RESUME,
 				      name, NULL);
 	(void)xen_set_irq_priority(irq, XEN_IRQ_PRIORITY_MAX);
 
diff --git a/block/bio-integrity.c b/block/bio-integrity.c
index 9e24106..bc423f7b 100644
--- a/block/bio-integrity.c
+++ b/block/bio-integrity.c
@@ -70,8 +70,10 @@
 					  bs->bvec_integrity_pool);
 		if (!bip->bip_vec)
 			goto err;
+		bip->bip_max_vcnt = bvec_nr_vecs(idx);
 	} else {
 		bip->bip_vec = bip->bip_inline_vecs;
+		bip->bip_max_vcnt = inline_vecs;
 	}
 
 	bip->bip_slab = idx;
@@ -114,14 +116,6 @@
 }
 EXPORT_SYMBOL(bio_integrity_free);
 
-static inline unsigned int bip_integrity_vecs(struct bio_integrity_payload *bip)
-{
-	if (bip->bip_slab == BIO_POOL_NONE)
-		return BIP_INLINE_VECS;
-
-	return bvec_nr_vecs(bip->bip_slab);
-}
-
 /**
  * bio_integrity_add_page - Attach integrity metadata
  * @bio:	bio to update
@@ -137,7 +131,7 @@
 	struct bio_integrity_payload *bip = bio->bi_integrity;
 	struct bio_vec *iv;
 
-	if (bip->bip_vcnt >= bip_integrity_vecs(bip)) {
+	if (bip->bip_vcnt >= bip->bip_max_vcnt) {
 		printk(KERN_ERR "%s: bip_vec full\n", __func__);
 		return 0;
 	}
diff --git a/block/bio.c b/block/bio.c
index 0ec61c9..3e6331d 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -112,7 +112,8 @@
 	bslab = &bio_slabs[entry];
 
 	snprintf(bslab->name, sizeof(bslab->name), "bio-%d", entry);
-	slab = kmem_cache_create(bslab->name, sz, 0, SLAB_HWCACHE_ALIGN, NULL);
+	slab = kmem_cache_create(bslab->name, sz, ARCH_KMALLOC_MINALIGN,
+				 SLAB_HWCACHE_ALIGN, NULL);
 	if (!slab)
 		goto out_unlock;
 
diff --git a/block/blk-core.c b/block/blk-core.c
index 6f8dba1..c359d72 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -438,14 +438,17 @@
  */
 void blk_queue_bypass_start(struct request_queue *q)
 {
-	bool drain;
-
 	spin_lock_irq(q->queue_lock);
-	drain = !q->bypass_depth++;
+	q->bypass_depth++;
 	queue_flag_set(QUEUE_FLAG_BYPASS, q);
 	spin_unlock_irq(q->queue_lock);
 
-	if (drain) {
+	/*
+	 * Queues start drained.  Skip actual draining till init is
+	 * complete.  This avoids lenghty delays during queue init which
+	 * can happen many times during boot.
+	 */
+	if (blk_queue_init_done(q)) {
 		spin_lock_irq(q->queue_lock);
 		__blk_drain_queue(q, false);
 		spin_unlock_irq(q->queue_lock);
@@ -511,7 +514,7 @@
 	 * prevent that q->request_fn() gets invoked after draining finished.
 	 */
 	if (q->mq_ops) {
-		blk_mq_drain_queue(q);
+		blk_mq_freeze_queue(q);
 		spin_lock_irq(lock);
 	} else {
 		spin_lock_irq(lock);
diff --git a/block/blk-mq.c b/block/blk-mq.c
index ad69ef6..5189cb1 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -78,68 +78,47 @@
 
 static int blk_mq_queue_enter(struct request_queue *q)
 {
-	int ret;
+	while (true) {
+		int ret;
 
-	__percpu_counter_add(&q->mq_usage_counter, 1, 1000000);
-	smp_wmb();
+		if (percpu_ref_tryget_live(&q->mq_usage_counter))
+			return 0;
 
-	/* we have problems freezing the queue if it's initializing */
-	if (!blk_queue_dying(q) &&
-	    (!blk_queue_bypass(q) || !blk_queue_init_done(q)))
-		return 0;
-
-	__percpu_counter_add(&q->mq_usage_counter, -1, 1000000);
-
-	spin_lock_irq(q->queue_lock);
-	ret = wait_event_interruptible_lock_irq(q->mq_freeze_wq,
-		!blk_queue_bypass(q) || blk_queue_dying(q),
-		*q->queue_lock);
-	/* inc usage with lock hold to avoid freeze_queue runs here */
-	if (!ret && !blk_queue_dying(q))
-		__percpu_counter_add(&q->mq_usage_counter, 1, 1000000);
-	else if (blk_queue_dying(q))
-		ret = -ENODEV;
-	spin_unlock_irq(q->queue_lock);
-
-	return ret;
+		ret = wait_event_interruptible(q->mq_freeze_wq,
+				!q->mq_freeze_depth || blk_queue_dying(q));
+		if (blk_queue_dying(q))
+			return -ENODEV;
+		if (ret)
+			return ret;
+	}
 }
 
 static void blk_mq_queue_exit(struct request_queue *q)
 {
-	__percpu_counter_add(&q->mq_usage_counter, -1, 1000000);
+	percpu_ref_put(&q->mq_usage_counter);
 }
 
-void blk_mq_drain_queue(struct request_queue *q)
+static void blk_mq_usage_counter_release(struct percpu_ref *ref)
 {
-	while (true) {
-		s64 count;
+	struct request_queue *q =
+		container_of(ref, struct request_queue, mq_usage_counter);
 
-		spin_lock_irq(q->queue_lock);
-		count = percpu_counter_sum(&q->mq_usage_counter);
-		spin_unlock_irq(q->queue_lock);
-
-		if (count == 0)
-			break;
-		blk_mq_start_hw_queues(q);
-		msleep(10);
-	}
+	wake_up_all(&q->mq_freeze_wq);
 }
 
 /*
  * Guarantee no request is in use, so we can change any data structure of
  * the queue afterward.
  */
-static void blk_mq_freeze_queue(struct request_queue *q)
+void blk_mq_freeze_queue(struct request_queue *q)
 {
-	bool drain;
-
 	spin_lock_irq(q->queue_lock);
-	drain = !q->bypass_depth++;
-	queue_flag_set(QUEUE_FLAG_BYPASS, q);
+	q->mq_freeze_depth++;
 	spin_unlock_irq(q->queue_lock);
 
-	if (drain)
-		blk_mq_drain_queue(q);
+	percpu_ref_kill(&q->mq_usage_counter);
+	blk_mq_run_queues(q, false);
+	wait_event(q->mq_freeze_wq, percpu_ref_is_zero(&q->mq_usage_counter));
 }
 
 static void blk_mq_unfreeze_queue(struct request_queue *q)
@@ -147,14 +126,13 @@
 	bool wake = false;
 
 	spin_lock_irq(q->queue_lock);
-	if (!--q->bypass_depth) {
-		queue_flag_clear(QUEUE_FLAG_BYPASS, q);
-		wake = true;
-	}
-	WARN_ON_ONCE(q->bypass_depth < 0);
+	wake = !--q->mq_freeze_depth;
+	WARN_ON_ONCE(q->mq_freeze_depth < 0);
 	spin_unlock_irq(q->queue_lock);
-	if (wake)
+	if (wake) {
+		percpu_ref_reinit(&q->mq_usage_counter);
 		wake_up_all(&q->mq_freeze_wq);
+	}
 }
 
 bool blk_mq_can_queue(struct blk_mq_hw_ctx *hctx)
@@ -1798,7 +1776,7 @@
 	if (!q)
 		goto err_hctxs;
 
-	if (percpu_counter_init(&q->mq_usage_counter, 0))
+	if (percpu_ref_init(&q->mq_usage_counter, blk_mq_usage_counter_release))
 		goto err_map;
 
 	setup_timer(&q->timeout, blk_mq_rq_timer, (unsigned long) q);
@@ -1891,7 +1869,7 @@
 	blk_mq_exit_hw_queues(q, set, set->nr_hw_queues);
 	blk_mq_free_hw_queues(q, set);
 
-	percpu_counter_destroy(&q->mq_usage_counter);
+	percpu_ref_exit(&q->mq_usage_counter);
 
 	free_percpu(q->queue_ctx);
 	kfree(q->queue_hw_ctx);
@@ -2050,8 +2028,7 @@
 {
 	blk_mq_cpu_init();
 
-	/* Must be called after percpu_counter_hotcpu_callback() */
-	hotcpu_notifier(blk_mq_queue_reinit_notify, -10);
+	hotcpu_notifier(blk_mq_queue_reinit_notify, 0);
 
 	return 0;
 }
diff --git a/block/blk-mq.h b/block/blk-mq.h
index 2646088..ca4964a 100644
--- a/block/blk-mq.h
+++ b/block/blk-mq.h
@@ -28,7 +28,7 @@
 void __blk_mq_complete_request(struct request *rq);
 void blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx, bool async);
 void blk_mq_init_flush(struct request_queue *q);
-void blk_mq_drain_queue(struct request_queue *q);
+void blk_mq_freeze_queue(struct request_queue *q);
 void blk_mq_free_queue(struct request_queue *q);
 void blk_mq_clone_flush_request(struct request *flush_rq,
 		struct request *orig_rq);
diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c
index 23321fb..4db5abf 100644
--- a/block/blk-sysfs.c
+++ b/block/blk-sysfs.c
@@ -554,8 +554,8 @@
 	 * Initialization must be complete by now.  Finish the initial
 	 * bypass from queue allocation.
 	 */
-	blk_queue_bypass_end(q);
 	queue_flag_set_unlocked(QUEUE_FLAG_INIT_DONE, q);
+	blk_queue_bypass_end(q);
 
 	ret = blk_trace_init_sysfs(dev);
 	if (ret)
diff --git a/block/compat_ioctl.c b/block/compat_ioctl.c
index a0926a6..18b282c 100644
--- a/block/compat_ioctl.c
+++ b/block/compat_ioctl.c
@@ -663,6 +663,7 @@
 	fmode_t mode = file->f_mode;
 	struct backing_dev_info *bdi;
 	loff_t size;
+	unsigned int max_sectors;
 
 	/*
 	 * O_NDELAY can be altered using fcntl(.., F_SETFL, ..), so we have
@@ -719,8 +720,9 @@
 	case BLKSSZGET: /* get block device hardware sector size */
 		return compat_put_int(arg, bdev_logical_block_size(bdev));
 	case BLKSECTGET:
-		return compat_put_ushort(arg,
-					 queue_max_sectors(bdev_get_queue(bdev)));
+		max_sectors = min_t(unsigned int, USHRT_MAX,
+				    queue_max_sectors(bdev_get_queue(bdev)));
+		return compat_put_ushort(arg, max_sectors);
 	case BLKROTATIONAL:
 		return compat_put_ushort(arg,
 					 !blk_queue_nonrot(bdev_get_queue(bdev)));
diff --git a/block/ioctl.c b/block/ioctl.c
index 7d5c3b2..d6cda81 100644
--- a/block/ioctl.c
+++ b/block/ioctl.c
@@ -278,6 +278,7 @@
 	struct backing_dev_info *bdi;
 	loff_t size;
 	int ret, n;
+	unsigned int max_sectors;
 
 	switch(cmd) {
 	case BLKFLSBUF:
@@ -375,7 +376,9 @@
 	case BLKDISCARDZEROES:
 		return put_uint(arg, bdev_discard_zeroes_data(bdev));
 	case BLKSECTGET:
-		return put_ushort(arg, queue_max_sectors(bdev_get_queue(bdev)));
+		max_sectors = min_t(unsigned int, USHRT_MAX,
+				    queue_max_sectors(bdev_get_queue(bdev)));
+		return put_ushort(arg, max_sectors);
 	case BLKROTATIONAL:
 		return put_ushort(arg, !blk_queue_nonrot(bdev_get_queue(bdev)));
 	case BLKRASET:
diff --git a/block/partitions/aix.c b/block/partitions/aix.c
index 43be471..f3ed7b2 100644
--- a/block/partitions/aix.c
+++ b/block/partitions/aix.c
@@ -215,7 +215,7 @@
 		numlvs = be16_to_cpu(p->numlvs);
 		put_dev_sector(sect);
 	}
-	lvip = kzalloc(sizeof(struct lv_info) * state->limit, GFP_KERNEL);
+	lvip = kcalloc(state->limit, sizeof(struct lv_info), GFP_KERNEL);
 	if (!lvip)
 		return 0;
 	if (numlvs && (d = read_part_sector(state, vgda_sector + 1, &sect))) {
@@ -253,7 +253,7 @@
 				continue;
 			}
 			lv_ix = be16_to_cpu(p->lv_ix) - 1;
-			if (lv_ix > state->limit) {
+			if (lv_ix >= state->limit) {
 				cur_lv_ix = -1;
 				continue;
 			}
diff --git a/block/partitions/amiga.c b/block/partitions/amiga.c
index 70cbf44..2b13533 100644
--- a/block/partitions/amiga.c
+++ b/block/partitions/amiga.c
@@ -7,6 +7,8 @@
  *  Re-organised Feb 1998 Russell King
  */
 
+#define pr_fmt(fmt) fmt
+
 #include <linux/types.h>
 #include <linux/affs_hardblocks.h>
 
@@ -40,7 +42,7 @@
 		data = read_part_sector(state, blk, &sect);
 		if (!data) {
 			if (warn_no_part)
-				printk("Dev %s: unable to read RDB block %d\n",
+				pr_err("Dev %s: unable to read RDB block %d\n",
 				       bdevname(state->bdev, b), blk);
 			res = -1;
 			goto rdb_done;
@@ -57,12 +59,12 @@
 		*(__be32 *)(data+0xdc) = 0;
 		if (checksum_block((__be32 *)data,
 				be32_to_cpu(rdb->rdb_SummedLongs) & 0x7F)==0) {
-			printk("Warning: Trashed word at 0xd0 in block %d "
-				"ignored in checksum calculation\n",blk);
+			pr_err("Trashed word at 0xd0 in block %d ignored in checksum calculation\n",
+			       blk);
 			break;
 		}
 
-		printk("Dev %s: RDB in block %d has bad checksum\n",
+		pr_err("Dev %s: RDB in block %d has bad checksum\n",
 		       bdevname(state->bdev, b), blk);
 	}
 
@@ -83,7 +85,7 @@
 		data = read_part_sector(state, blk, &sect);
 		if (!data) {
 			if (warn_no_part)
-				printk("Dev %s: unable to read partition block %d\n",
+				pr_err("Dev %s: unable to read partition block %d\n",
 				       bdevname(state->bdev, b), blk);
 			res = -1;
 			goto rdb_done;
diff --git a/block/partitions/efi.c b/block/partitions/efi.c
index dc51f46..56d08fd 100644
--- a/block/partitions/efi.c
+++ b/block/partitions/efi.c
@@ -121,7 +121,7 @@
 /**
  * efi_crc32() - EFI version of crc32 function
  * @buf: buffer to calculate crc32 of
- * @len - length of buf
+ * @len: length of buf
  *
  * Description: Returns EFI-style CRC32 value for @buf
  * 
@@ -240,10 +240,10 @@
 
 /**
  * read_lba(): Read bytes from disk, starting at given LBA
- * @state
- * @lba
- * @buffer
- * @size_t
+ * @state: disk parsed partitions
+ * @lba: the Logical Block Address of the partition table
+ * @buffer: destination buffer
+ * @count: bytes to read
  *
  * Description: Reads @count bytes from @state->bdev into @buffer.
  * Returns number of bytes read on success, 0 on error.
@@ -277,8 +277,8 @@
 
 /**
  * alloc_read_gpt_entries(): reads partition entries from disk
- * @state
- * @gpt - GPT header
+ * @state: disk parsed partitions
+ * @gpt: GPT header
  * 
  * Description: Returns ptes on success,  NULL on error.
  * Allocates space for PTEs based on information found in @gpt.
@@ -312,8 +312,8 @@
 
 /**
  * alloc_read_gpt_header(): Allocates GPT header, reads into it from disk
- * @state
- * @lba is the Logical Block Address of the partition table
+ * @state: disk parsed partitions
+ * @lba: the Logical Block Address of the partition table
  * 
  * Description: returns GPT header on success, NULL on error.   Allocates
  * and fills a GPT header starting at @ from @state->bdev.
@@ -340,10 +340,10 @@
 
 /**
  * is_gpt_valid() - tests one GPT header and PTEs for validity
- * @state
- * @lba is the logical block address of the GPT header to test
- * @gpt is a GPT header ptr, filled on return.
- * @ptes is a PTEs ptr, filled on return.
+ * @state: disk parsed partitions
+ * @lba: logical block address of the GPT header to test
+ * @gpt: GPT header ptr, filled on return.
+ * @ptes: PTEs ptr, filled on return.
  *
  * Description: returns 1 if valid,  0 on error.
  * If valid, returns pointers to newly allocated GPT header and PTEs.
@@ -461,8 +461,8 @@
 
 /**
  * is_pte_valid() - tests one PTE for validity
- * @pte is the pte to check
- * @lastlba is last lba of the disk
+ * @pte:pte to check
+ * @lastlba: last lba of the disk
  *
  * Description: returns 1 if valid,  0 on error.
  */
@@ -478,9 +478,10 @@
 
 /**
  * compare_gpts() - Search disk for valid GPT headers and PTEs
- * @pgpt is the primary GPT header
- * @agpt is the alternate GPT header
- * @lastlba is the last LBA number
+ * @pgpt: primary GPT header
+ * @agpt: alternate GPT header
+ * @lastlba: last LBA number
+ *
  * Description: Returns nothing.  Sanity checks pgpt and agpt fields
  * and prints warnings on discrepancies.
  * 
@@ -572,9 +573,10 @@
 
 /**
  * find_valid_gpt() - Search disk for valid GPT headers and PTEs
- * @state
- * @gpt is a GPT header ptr, filled on return.
- * @ptes is a PTEs ptr, filled on return.
+ * @state: disk parsed partitions
+ * @gpt: GPT header ptr, filled on return.
+ * @ptes: PTEs ptr, filled on return.
+ *
  * Description: Returns 1 if valid, 0 on error.
  * If valid, returns pointers to newly allocated GPT header and PTEs.
  * Validity depends on PMBR being valid (or being overridden by the
@@ -663,7 +665,7 @@
 
 /**
  * efi_partition(struct parsed_partitions *state)
- * @state
+ * @state: disk parsed partitions
  *
  * Description: called from check.c, if the disk contains GPT
  * partitions, sets up partition entries in the kernel.
diff --git a/block/partitions/msdos.c b/block/partitions/msdos.c
index 9123f25..93e7c1b 100644
--- a/block/partitions/msdos.c
+++ b/block/partitions/msdos.c
@@ -159,8 +159,9 @@
 		/*
 		 * First process the data partition(s)
 		 */
-		for (i=0; i<4; i++, p++) {
+		for (i = 0; i < 4; i++, p++) {
 			sector_t offs, size, next;
+
 			if (!nr_sects(p) || is_extended_partition(p))
 				continue;
 
@@ -194,7 +195,7 @@
 		 * It should be a link to the next logical partition.
 		 */
 		p -= 4;
-		for (i=0; i<4; i++, p++)
+		for (i = 0; i < 4; i++, p++)
 			if (nr_sects(p) && is_extended_partition(p))
 				break;
 		if (i == 4)
@@ -243,8 +244,8 @@
 		return;
 	}
 	/* Ensure we can handle previous case of VTOC with 8 entries gracefully */
-	max_nparts = le16_to_cpu (v->v_nparts) > 8 ? SOLARIS_X86_NUMSLICE : 8;
-	for (i=0; i<max_nparts && state->next<state->limit; i++) {
+	max_nparts = le16_to_cpu(v->v_nparts) > 8 ? SOLARIS_X86_NUMSLICE : 8;
+	for (i = 0; i < max_nparts && state->next < state->limit; i++) {
 		struct solaris_x86_slice *s = &v->v_slice[i];
 		char tmp[3 + 10 + 1 + 1];
 
@@ -409,7 +410,7 @@
 	/* The first sector of a Minix partition can have either
 	 * a secondary MBR describing its subpartitions, or
 	 * the normal boot sector. */
-	if (msdos_magic_present (data + 510) &&
+	if (msdos_magic_present(data + 510) &&
 	    SYS_IND(p) == MINIX_PARTITION) { /* subpartition table present */
 		char tmp[1 + BDEVNAME_SIZE + 10 + 9 + 1];
 
@@ -527,6 +528,7 @@
 	for (slot = 1 ; slot <= 4 ; slot++, p++) {
 		sector_t start = start_sect(p)*sector_size;
 		sector_t size = nr_sects(p)*sector_size;
+
 		if (!size)
 			continue;
 		if (is_extended_partition(p)) {
@@ -537,6 +539,7 @@
 			 * sector, although it may not be enough/proper.
 			 */
 			sector_t n = 2;
+
 			n = min(size, max(sector_size, n));
 			put_partition(state, slot, start, n);
 
diff --git a/block/scsi_ioctl.c b/block/scsi_ioctl.c
index 14695c6..51bf515 100644
--- a/block/scsi_ioctl.c
+++ b/block/scsi_ioctl.c
@@ -82,9 +82,18 @@
 	return err;
 }
 
+static int max_sectors_bytes(struct request_queue *q)
+{
+	unsigned int max_sectors = queue_max_sectors(q);
+
+	max_sectors = min_t(unsigned int, max_sectors, INT_MAX >> 9);
+
+	return max_sectors << 9;
+}
+
 static int sg_get_reserved_size(struct request_queue *q, int __user *p)
 {
-	unsigned val = min(q->sg_reserved_size, queue_max_sectors(q) << 9);
+	int val = min_t(int, q->sg_reserved_size, max_sectors_bytes(q));
 
 	return put_user(val, p);
 }
@@ -98,10 +107,8 @@
 
 	if (size < 0)
 		return -EINVAL;
-	if (size > (queue_max_sectors(q) << 9))
-		size = queue_max_sectors(q) << 9;
 
-	q->sg_reserved_size = size;
+	q->sg_reserved_size = min(size, max_sectors_bytes(q));
 	return 0;
 }
 
@@ -283,6 +290,7 @@
 	unsigned long start_time;
 	ssize_t ret = 0;
 	int writing = 0;
+	int at_head = 0;
 	struct request *rq;
 	char sense[SCSI_SENSE_BUFFERSIZE];
 	struct bio *bio;
@@ -306,6 +314,8 @@
 		case SG_DXFER_FROM_DEV:
 			break;
 		}
+	if (hdr->flags & SG_FLAG_Q_AT_HEAD)
+		at_head = 1;
 
 	rq = blk_get_request(q, writing ? WRITE : READ, GFP_KERNEL);
 	if (!rq)
@@ -362,7 +372,7 @@
 	 * (if he doesn't check that is his problem).
 	 * N.B. a non-zero SCSI status is _not_ necessarily an error.
 	 */
-	blk_execute_rq(q, bd_disk, rq, 0);
+	blk_execute_rq(q, bd_disk, rq, at_head);
 
 	hdr->duration = jiffies_to_msecs(jiffies - start_time);
 
diff --git a/drivers/block/drbd/Makefile b/drivers/block/drbd/Makefile
index 8b45033..4464e35 100644
--- a/drivers/block/drbd/Makefile
+++ b/drivers/block/drbd/Makefile
@@ -3,5 +3,6 @@
 drbd-y += drbd_main.o drbd_strings.o drbd_nl.o
 drbd-y += drbd_interval.o drbd_state.o
 drbd-y += drbd_nla.o
+drbd-$(CONFIG_DEBUG_FS) += drbd_debugfs.o
 
 obj-$(CONFIG_BLK_DEV_DRBD)     += drbd.o
diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c
index 05a1780..d26a3fa 100644
--- a/drivers/block/drbd/drbd_actlog.c
+++ b/drivers/block/drbd/drbd_actlog.c
@@ -92,34 +92,26 @@
 	__be32	context[AL_CONTEXT_PER_TRANSACTION];
 };
 
-struct update_odbm_work {
-	struct drbd_work w;
-	struct drbd_device *device;
-	unsigned int enr;
-};
-
-struct update_al_work {
-	struct drbd_work w;
-	struct drbd_device *device;
-	struct completion event;
-	int err;
-};
-
-
-void *drbd_md_get_buffer(struct drbd_device *device)
+void *drbd_md_get_buffer(struct drbd_device *device, const char *intent)
 {
 	int r;
 
 	wait_event(device->misc_wait,
-		   (r = atomic_cmpxchg(&device->md_io_in_use, 0, 1)) == 0 ||
+		   (r = atomic_cmpxchg(&device->md_io.in_use, 0, 1)) == 0 ||
 		   device->state.disk <= D_FAILED);
 
-	return r ? NULL : page_address(device->md_io_page);
+	if (r)
+		return NULL;
+
+	device->md_io.current_use = intent;
+	device->md_io.start_jif = jiffies;
+	device->md_io.submit_jif = device->md_io.start_jif - 1;
+	return page_address(device->md_io.page);
 }
 
 void drbd_md_put_buffer(struct drbd_device *device)
 {
-	if (atomic_dec_and_test(&device->md_io_in_use))
+	if (atomic_dec_and_test(&device->md_io.in_use))
 		wake_up(&device->misc_wait);
 }
 
@@ -145,10 +137,11 @@
 
 static int _drbd_md_sync_page_io(struct drbd_device *device,
 				 struct drbd_backing_dev *bdev,
-				 struct page *page, sector_t sector,
-				 int rw, int size)
+				 sector_t sector, int rw)
 {
 	struct bio *bio;
+	/* we do all our meta data IO in aligned 4k blocks. */
+	const int size = 4096;
 	int err;
 
 	device->md_io.done = 0;
@@ -156,15 +149,15 @@
 
 	if ((rw & WRITE) && !test_bit(MD_NO_FUA, &device->flags))
 		rw |= REQ_FUA | REQ_FLUSH;
-	rw |= REQ_SYNC;
+	rw |= REQ_SYNC | REQ_NOIDLE;
 
 	bio = bio_alloc_drbd(GFP_NOIO);
 	bio->bi_bdev = bdev->md_bdev;
 	bio->bi_iter.bi_sector = sector;
 	err = -EIO;
-	if (bio_add_page(bio, page, size, 0) != size)
+	if (bio_add_page(bio, device->md_io.page, size, 0) != size)
 		goto out;
-	bio->bi_private = &device->md_io;
+	bio->bi_private = device;
 	bio->bi_end_io = drbd_md_io_complete;
 	bio->bi_rw = rw;
 
@@ -179,7 +172,8 @@
 	}
 
 	bio_get(bio); /* one bio_put() is in the completion handler */
-	atomic_inc(&device->md_io_in_use); /* drbd_md_put_buffer() is in the completion handler */
+	atomic_inc(&device->md_io.in_use); /* drbd_md_put_buffer() is in the completion handler */
+	device->md_io.submit_jif = jiffies;
 	if (drbd_insert_fault(device, (rw & WRITE) ? DRBD_FAULT_MD_WR : DRBD_FAULT_MD_RD))
 		bio_endio(bio, -EIO);
 	else
@@ -197,9 +191,7 @@
 			 sector_t sector, int rw)
 {
 	int err;
-	struct page *iop = device->md_io_page;
-
-	D_ASSERT(device, atomic_read(&device->md_io_in_use) == 1);
+	D_ASSERT(device, atomic_read(&device->md_io.in_use) == 1);
 
 	BUG_ON(!bdev->md_bdev);
 
@@ -214,8 +206,7 @@
 		     current->comm, current->pid, __func__,
 		     (unsigned long long)sector, (rw & WRITE) ? "WRITE" : "READ");
 
-	/* we do all our meta data IO in aligned 4k blocks. */
-	err = _drbd_md_sync_page_io(device, bdev, iop, sector, rw, 4096);
+	err = _drbd_md_sync_page_io(device, bdev, sector, rw);
 	if (err) {
 		drbd_err(device, "drbd_md_sync_page_io(,%llus,%s) failed with error %d\n",
 		    (unsigned long long)sector, (rw & WRITE) ? "WRITE" : "READ", err);
@@ -297,26 +288,12 @@
 	return need_transaction;
 }
 
-static int al_write_transaction(struct drbd_device *device, bool delegate);
+static int al_write_transaction(struct drbd_device *device);
 
-/* When called through generic_make_request(), we must delegate
- * activity log I/O to the worker thread: a further request
- * submitted via generic_make_request() within the same task
- * would be queued on current->bio_list, and would only start
- * after this function returns (see generic_make_request()).
- *
- * However, if we *are* the worker, we must not delegate to ourselves.
- */
-
-/*
- * @delegate:   delegate activity log I/O to the worker thread
- */
-void drbd_al_begin_io_commit(struct drbd_device *device, bool delegate)
+void drbd_al_begin_io_commit(struct drbd_device *device)
 {
 	bool locked = false;
 
-	BUG_ON(delegate && current == first_peer_device(device)->connection->worker.task);
-
 	/* Serialize multiple transactions.
 	 * This uses test_and_set_bit, memory barrier is implicit.
 	 */
@@ -335,7 +312,7 @@
 			rcu_read_unlock();
 
 			if (write_al_updates)
-				al_write_transaction(device, delegate);
+				al_write_transaction(device);
 			spin_lock_irq(&device->al_lock);
 			/* FIXME
 			if (err)
@@ -352,12 +329,10 @@
 /*
  * @delegate:   delegate activity log I/O to the worker thread
  */
-void drbd_al_begin_io(struct drbd_device *device, struct drbd_interval *i, bool delegate)
+void drbd_al_begin_io(struct drbd_device *device, struct drbd_interval *i)
 {
-	BUG_ON(delegate && current == first_peer_device(device)->connection->worker.task);
-
 	if (drbd_al_begin_io_prepare(device, i))
-		drbd_al_begin_io_commit(device, delegate);
+		drbd_al_begin_io_commit(device);
 }
 
 int drbd_al_begin_io_nonblock(struct drbd_device *device, struct drbd_interval *i)
@@ -380,8 +355,19 @@
 	/* We want all necessary updates for a given request within the same transaction
 	 * We could first check how many updates are *actually* needed,
 	 * and use that instead of the worst-case nr_al_extents */
-	if (available_update_slots < nr_al_extents)
-		return -EWOULDBLOCK;
+	if (available_update_slots < nr_al_extents) {
+		/* Too many activity log extents are currently "hot".
+		 *
+		 * If we have accumulated pending changes already,
+		 * we made progress.
+		 *
+		 * If we cannot get even a single pending change through,
+		 * stop the fast path until we made some progress,
+		 * or requests to "cold" extents could be starved. */
+		if (!al->pending_changes)
+			__set_bit(__LC_STARVING, &device->act_log->flags);
+		return -ENOBUFS;
+	}
 
 	/* Is resync active in this area? */
 	for (enr = first; enr <= last; enr++) {
@@ -452,15 +438,6 @@
 		 (AL_EXTENT_SHIFT - BM_BLOCK_SHIFT));
 }
 
-static unsigned int rs_extent_to_bm_page(unsigned int rs_enr)
-{
-	return rs_enr >>
-		/* bit to page */
-		((PAGE_SHIFT + 3) -
-		/* resync extent number to bit */
-		 (BM_EXT_SHIFT - BM_BLOCK_SHIFT));
-}
-
 static sector_t al_tr_number_to_on_disk_sector(struct drbd_device *device)
 {
 	const unsigned int stripes = device->ldev->md.al_stripes;
@@ -479,8 +456,7 @@
 	return device->ldev->md.md_offset + device->ldev->md.al_offset + t;
 }
 
-static int
-_al_write_transaction(struct drbd_device *device)
+int al_write_transaction(struct drbd_device *device)
 {
 	struct al_transaction_on_disk *buffer;
 	struct lc_element *e;
@@ -505,7 +481,8 @@
 		return -EIO;
 	}
 
-	buffer = drbd_md_get_buffer(device); /* protects md_io_buffer, al_tr_cycle, ... */
+	/* protects md_io_buffer, al_tr_cycle, ... */
+	buffer = drbd_md_get_buffer(device, __func__);
 	if (!buffer) {
 		drbd_err(device, "disk failed while waiting for md_io buffer\n");
 		put_ldev(device);
@@ -590,38 +567,6 @@
 	return err;
 }
 
-
-static int w_al_write_transaction(struct drbd_work *w, int unused)
-{
-	struct update_al_work *aw = container_of(w, struct update_al_work, w);
-	struct drbd_device *device = aw->device;
-	int err;
-
-	err = _al_write_transaction(device);
-	aw->err = err;
-	complete(&aw->event);
-
-	return err != -EIO ? err : 0;
-}
-
-/* Calls from worker context (see w_restart_disk_io()) need to write the
-   transaction directly. Others came through generic_make_request(),
-   those need to delegate it to the worker. */
-static int al_write_transaction(struct drbd_device *device, bool delegate)
-{
-	if (delegate) {
-		struct update_al_work al_work;
-		init_completion(&al_work.event);
-		al_work.w.cb = w_al_write_transaction;
-		al_work.device = device;
-		drbd_queue_work_front(&first_peer_device(device)->connection->sender_work,
-				      &al_work.w);
-		wait_for_completion(&al_work.event);
-		return al_work.err;
-	} else
-		return _al_write_transaction(device);
-}
-
 static int _try_lc_del(struct drbd_device *device, struct lc_element *al_ext)
 {
 	int rv;
@@ -682,72 +627,56 @@
 	return 0;
 }
 
-static int w_update_odbm(struct drbd_work *w, int unused)
-{
-	struct update_odbm_work *udw = container_of(w, struct update_odbm_work, w);
-	struct drbd_device *device = udw->device;
-	struct sib_info sib = { .sib_reason = SIB_SYNC_PROGRESS, };
-
-	if (!get_ldev(device)) {
-		if (__ratelimit(&drbd_ratelimit_state))
-			drbd_warn(device, "Can not update on disk bitmap, local IO disabled.\n");
-		kfree(udw);
-		return 0;
-	}
-
-	drbd_bm_write_page(device, rs_extent_to_bm_page(udw->enr));
-	put_ldev(device);
-
-	kfree(udw);
-
-	if (drbd_bm_total_weight(device) <= device->rs_failed) {
-		switch (device->state.conn) {
-		case C_SYNC_SOURCE:  case C_SYNC_TARGET:
-		case C_PAUSED_SYNC_S: case C_PAUSED_SYNC_T:
-			drbd_resync_finished(device);
-		default:
-			/* nothing to do */
-			break;
-		}
-	}
-	drbd_bcast_event(device, &sib);
-
-	return 0;
-}
-
+static const char *drbd_change_sync_fname[] = {
+	[RECORD_RS_FAILED] = "drbd_rs_failed_io",
+	[SET_IN_SYNC] = "drbd_set_in_sync",
+	[SET_OUT_OF_SYNC] = "drbd_set_out_of_sync"
+};
 
 /* ATTENTION. The AL's extents are 4MB each, while the extents in the
  * resync LRU-cache are 16MB each.
  * The caller of this function has to hold an get_ldev() reference.
  *
+ * Adjusts the caching members ->rs_left (success) or ->rs_failed (!success),
+ * potentially pulling in (and recounting the corresponding bits)
+ * this resync extent into the resync extent lru cache.
+ *
+ * Returns whether all bits have been cleared for this resync extent,
+ * precisely: (rs_left <= rs_failed)
+ *
  * TODO will be obsoleted once we have a caching lru of the on disk bitmap
  */
-static void drbd_try_clear_on_disk_bm(struct drbd_device *device, sector_t sector,
-				      int count, int success)
+static bool update_rs_extent(struct drbd_device *device,
+		unsigned int enr, int count,
+		enum update_sync_bits_mode mode)
 {
 	struct lc_element *e;
-	struct update_odbm_work *udw;
-
-	unsigned int enr;
 
 	D_ASSERT(device, atomic_read(&device->local_cnt));
 
-	/* I simply assume that a sector/size pair never crosses
-	 * a 16 MB extent border. (Currently this is true...) */
-	enr = BM_SECT_TO_EXT(sector);
-
-	e = lc_get(device->resync, enr);
+	/* When setting out-of-sync bits,
+	 * we don't need it cached (lc_find).
+	 * But if it is present in the cache,
+	 * we should update the cached bit count.
+	 * Otherwise, that extent should be in the resync extent lru cache
+	 * already -- or we want to pull it in if necessary -- (lc_get),
+	 * then update and check rs_left and rs_failed. */
+	if (mode == SET_OUT_OF_SYNC)
+		e = lc_find(device->resync, enr);
+	else
+		e = lc_get(device->resync, enr);
 	if (e) {
 		struct bm_extent *ext = lc_entry(e, struct bm_extent, lce);
 		if (ext->lce.lc_number == enr) {
-			if (success)
+			if (mode == SET_IN_SYNC)
 				ext->rs_left -= count;
+			else if (mode == SET_OUT_OF_SYNC)
+				ext->rs_left += count;
 			else
 				ext->rs_failed += count;
 			if (ext->rs_left < ext->rs_failed) {
-				drbd_warn(device, "BAD! sector=%llus enr=%u rs_left=%d "
+				drbd_warn(device, "BAD! enr=%u rs_left=%d "
 				    "rs_failed=%d count=%d cstate=%s\n",
-				     (unsigned long long)sector,
 				     ext->lce.lc_number, ext->rs_left,
 				     ext->rs_failed, count,
 				     drbd_conn_str(device->state.conn));
@@ -781,34 +710,27 @@
 				     ext->lce.lc_number, ext->rs_failed);
 			}
 			ext->rs_left = rs_left;
-			ext->rs_failed = success ? 0 : count;
+			ext->rs_failed = (mode == RECORD_RS_FAILED) ? count : 0;
 			/* we don't keep a persistent log of the resync lru,
 			 * we can commit any change right away. */
 			lc_committed(device->resync);
 		}
-		lc_put(device->resync, &ext->lce);
+		if (mode != SET_OUT_OF_SYNC)
+			lc_put(device->resync, &ext->lce);
 		/* no race, we are within the al_lock! */
 
-		if (ext->rs_left == ext->rs_failed) {
+		if (ext->rs_left <= ext->rs_failed) {
 			ext->rs_failed = 0;
-
-			udw = kmalloc(sizeof(*udw), GFP_ATOMIC);
-			if (udw) {
-				udw->enr = ext->lce.lc_number;
-				udw->w.cb = w_update_odbm;
-				udw->device = device;
-				drbd_queue_work_front(&first_peer_device(device)->connection->sender_work,
-						      &udw->w);
-			} else {
-				drbd_warn(device, "Could not kmalloc an udw\n");
-			}
+			return true;
 		}
-	} else {
+	} else if (mode != SET_OUT_OF_SYNC) {
+		/* be quiet if lc_find() did not find it. */
 		drbd_err(device, "lc_get() failed! locked=%d/%d flags=%lu\n",
 		    device->resync_locked,
 		    device->resync->nr_elements,
 		    device->resync->flags);
 	}
+	return false;
 }
 
 void drbd_advance_rs_marks(struct drbd_device *device, unsigned long still_to_go)
@@ -827,6 +749,76 @@
 	}
 }
 
+/* It is called lazy update, so don't do write-out too often. */
+static bool lazy_bitmap_update_due(struct drbd_device *device)
+{
+	return time_after(jiffies, device->rs_last_bcast + 2*HZ);
+}
+
+static void maybe_schedule_on_disk_bitmap_update(struct drbd_device *device, bool rs_done)
+{
+	if (rs_done)
+		set_bit(RS_DONE, &device->flags);
+		/* and also set RS_PROGRESS below */
+	else if (!lazy_bitmap_update_due(device))
+		return;
+
+	drbd_device_post_work(device, RS_PROGRESS);
+}
+
+static int update_sync_bits(struct drbd_device *device,
+		unsigned long sbnr, unsigned long ebnr,
+		enum update_sync_bits_mode mode)
+{
+	/*
+	 * We keep a count of set bits per resync-extent in the ->rs_left
+	 * caching member, so we need to loop and work within the resync extent
+	 * alignment. Typically this loop will execute exactly once.
+	 */
+	unsigned long flags;
+	unsigned long count = 0;
+	unsigned int cleared = 0;
+	while (sbnr <= ebnr) {
+		/* set temporary boundary bit number to last bit number within
+		 * the resync extent of the current start bit number,
+		 * but cap at provided end bit number */
+		unsigned long tbnr = min(ebnr, sbnr | BM_BLOCKS_PER_BM_EXT_MASK);
+		unsigned long c;
+
+		if (mode == RECORD_RS_FAILED)
+			/* Only called from drbd_rs_failed_io(), bits
+			 * supposedly still set.  Recount, maybe some
+			 * of the bits have been successfully cleared
+			 * by application IO meanwhile.
+			 */
+			c = drbd_bm_count_bits(device, sbnr, tbnr);
+		else if (mode == SET_IN_SYNC)
+			c = drbd_bm_clear_bits(device, sbnr, tbnr);
+		else /* if (mode == SET_OUT_OF_SYNC) */
+			c = drbd_bm_set_bits(device, sbnr, tbnr);
+
+		if (c) {
+			spin_lock_irqsave(&device->al_lock, flags);
+			cleared += update_rs_extent(device, BM_BIT_TO_EXT(sbnr), c, mode);
+			spin_unlock_irqrestore(&device->al_lock, flags);
+			count += c;
+		}
+		sbnr = tbnr + 1;
+	}
+	if (count) {
+		if (mode == SET_IN_SYNC) {
+			unsigned long still_to_go = drbd_bm_total_weight(device);
+			bool rs_is_done = (still_to_go <= device->rs_failed);
+			drbd_advance_rs_marks(device, still_to_go);
+			if (cleared || rs_is_done)
+				maybe_schedule_on_disk_bitmap_update(device, rs_is_done);
+		} else if (mode == RECORD_RS_FAILED)
+			device->rs_failed += count;
+		wake_up(&device->al_wait);
+	}
+	return count;
+}
+
 /* clear the bit corresponding to the piece of storage in question:
  * size byte of data starting from sector.  Only clear a bits of the affected
  * one ore more _aligned_ BM_BLOCK_SIZE blocks.
@@ -834,24 +826,28 @@
  * called by worker on C_SYNC_TARGET and receiver on SyncSource.
  *
  */
-void __drbd_set_in_sync(struct drbd_device *device, sector_t sector, int size,
-		       const char *file, const unsigned int line)
+int __drbd_change_sync(struct drbd_device *device, sector_t sector, int size,
+		enum update_sync_bits_mode mode,
+		const char *file, const unsigned int line)
 {
 	/* Is called from worker and receiver context _only_ */
 	unsigned long sbnr, ebnr, lbnr;
 	unsigned long count = 0;
 	sector_t esector, nr_sectors;
-	int wake_up = 0;
-	unsigned long flags;
+
+	/* This would be an empty REQ_FLUSH, be silent. */
+	if ((mode == SET_OUT_OF_SYNC) && size == 0)
+		return 0;
 
 	if (size <= 0 || !IS_ALIGNED(size, 512) || size > DRBD_MAX_DISCARD_SIZE) {
-		drbd_err(device, "drbd_set_in_sync: sector=%llus size=%d nonsense!\n",
+		drbd_err(device, "%s: sector=%llus size=%d nonsense!\n",
+				drbd_change_sync_fname[mode],
 				(unsigned long long)sector, size);
-		return;
+		return 0;
 	}
 
 	if (!get_ldev(device))
-		return; /* no disk, no metadata, no bitmap to clear bits in */
+		return 0; /* no disk, no metadata, no bitmap to manipulate bits in */
 
 	nr_sectors = drbd_get_capacity(device->this_bdev);
 	esector = sector + (size >> 9) - 1;
@@ -863,97 +859,26 @@
 
 	lbnr = BM_SECT_TO_BIT(nr_sectors-1);
 
-	/* we clear it (in sync).
-	 * round up start sector, round down end sector.  we make sure we only
-	 * clear full, aligned, BM_BLOCK_SIZE (4K) blocks */
-	if (unlikely(esector < BM_SECT_PER_BIT-1))
-		goto out;
-	if (unlikely(esector == (nr_sectors-1)))
-		ebnr = lbnr;
-	else
-		ebnr = BM_SECT_TO_BIT(esector - (BM_SECT_PER_BIT-1));
-	sbnr = BM_SECT_TO_BIT(sector + BM_SECT_PER_BIT-1);
-
-	if (sbnr > ebnr)
-		goto out;
-
-	/*
-	 * ok, (capacity & 7) != 0 sometimes, but who cares...
-	 * we count rs_{total,left} in bits, not sectors.
-	 */
-	count = drbd_bm_clear_bits(device, sbnr, ebnr);
-	if (count) {
-		drbd_advance_rs_marks(device, drbd_bm_total_weight(device));
-		spin_lock_irqsave(&device->al_lock, flags);
-		drbd_try_clear_on_disk_bm(device, sector, count, true);
-		spin_unlock_irqrestore(&device->al_lock, flags);
-
-		/* just wake_up unconditional now, various lc_chaged(),
-		 * lc_put() in drbd_try_clear_on_disk_bm(). */
-		wake_up = 1;
-	}
-out:
-	put_ldev(device);
-	if (wake_up)
-		wake_up(&device->al_wait);
-}
-
-/*
- * this is intended to set one request worth of data out of sync.
- * affects at least 1 bit,
- * and at most 1+DRBD_MAX_BIO_SIZE/BM_BLOCK_SIZE bits.
- *
- * called by tl_clear and drbd_send_dblock (==drbd_make_request).
- * so this can be _any_ process.
- */
-int __drbd_set_out_of_sync(struct drbd_device *device, sector_t sector, int size,
-			    const char *file, const unsigned int line)
-{
-	unsigned long sbnr, ebnr, flags;
-	sector_t esector, nr_sectors;
-	unsigned int enr, count = 0;
-	struct lc_element *e;
-
-	/* this should be an empty REQ_FLUSH */
-	if (size == 0)
-		return 0;
-
-	if (size < 0 || !IS_ALIGNED(size, 512) || size > DRBD_MAX_DISCARD_SIZE) {
-		drbd_err(device, "sector: %llus, size: %d\n",
-			(unsigned long long)sector, size);
-		return 0;
+	if (mode == SET_IN_SYNC) {
+		/* Round up start sector, round down end sector.  We make sure
+		 * we only clear full, aligned, BM_BLOCK_SIZE blocks. */
+		if (unlikely(esector < BM_SECT_PER_BIT-1))
+			goto out;
+		if (unlikely(esector == (nr_sectors-1)))
+			ebnr = lbnr;
+		else
+			ebnr = BM_SECT_TO_BIT(esector - (BM_SECT_PER_BIT-1));
+		sbnr = BM_SECT_TO_BIT(sector + BM_SECT_PER_BIT-1);
+	} else {
+		/* We set it out of sync, or record resync failure.
+		 * Should not round anything here. */
+		sbnr = BM_SECT_TO_BIT(sector);
+		ebnr = BM_SECT_TO_BIT(esector);
 	}
 
-	if (!get_ldev(device))
-		return 0; /* no disk, no metadata, no bitmap to set bits in */
-
-	nr_sectors = drbd_get_capacity(device->this_bdev);
-	esector = sector + (size >> 9) - 1;
-
-	if (!expect(sector < nr_sectors))
-		goto out;
-	if (!expect(esector < nr_sectors))
-		esector = nr_sectors - 1;
-
-	/* we set it out of sync,
-	 * we do not need to round anything here */
-	sbnr = BM_SECT_TO_BIT(sector);
-	ebnr = BM_SECT_TO_BIT(esector);
-
-	/* ok, (capacity & 7) != 0 sometimes, but who cares...
-	 * we count rs_{total,left} in bits, not sectors.  */
-	spin_lock_irqsave(&device->al_lock, flags);
-	count = drbd_bm_set_bits(device, sbnr, ebnr);
-
-	enr = BM_SECT_TO_EXT(sector);
-	e = lc_find(device->resync, enr);
-	if (e)
-		lc_entry(e, struct bm_extent, lce)->rs_left += count;
-	spin_unlock_irqrestore(&device->al_lock, flags);
-
+	count = update_sync_bits(device, sbnr, ebnr, mode);
 out:
 	put_ldev(device);
-
 	return count;
 }
 
@@ -1075,6 +1000,15 @@
 	struct lc_element *e;
 	struct bm_extent *bm_ext;
 	int i;
+	bool throttle = drbd_rs_should_slow_down(device, sector, true);
+
+	/* If we need to throttle, a half-locked (only marked BME_NO_WRITES,
+	 * not yet BME_LOCKED) extent needs to be kicked out explicitly if we
+	 * need to throttle. There is at most one such half-locked extent,
+	 * which is remembered in resync_wenr. */
+
+	if (throttle && device->resync_wenr != enr)
+		return -EAGAIN;
 
 	spin_lock_irq(&device->al_lock);
 	if (device->resync_wenr != LC_FREE && device->resync_wenr != enr) {
@@ -1098,8 +1032,10 @@
 			D_ASSERT(device, test_bit(BME_NO_WRITES, &bm_ext->flags));
 			clear_bit(BME_NO_WRITES, &bm_ext->flags);
 			device->resync_wenr = LC_FREE;
-			if (lc_put(device->resync, &bm_ext->lce) == 0)
+			if (lc_put(device->resync, &bm_ext->lce) == 0) {
+				bm_ext->flags = 0;
 				device->resync_locked--;
+			}
 			wake_up(&device->al_wait);
 		} else {
 			drbd_alert(device, "LOGIC BUG\n");
@@ -1161,8 +1097,20 @@
 	return 0;
 
 try_again:
-	if (bm_ext)
-		device->resync_wenr = enr;
+	if (bm_ext) {
+		if (throttle) {
+			D_ASSERT(device, !test_bit(BME_LOCKED, &bm_ext->flags));
+			D_ASSERT(device, test_bit(BME_NO_WRITES, &bm_ext->flags));
+			clear_bit(BME_NO_WRITES, &bm_ext->flags);
+			device->resync_wenr = LC_FREE;
+			if (lc_put(device->resync, &bm_ext->lce) == 0) {
+				bm_ext->flags = 0;
+				device->resync_locked--;
+			}
+			wake_up(&device->al_wait);
+		} else
+			device->resync_wenr = enr;
+	}
 	spin_unlock_irq(&device->al_lock);
 	return -EAGAIN;
 }
@@ -1270,69 +1218,3 @@
 
 	return 0;
 }
-
-/**
- * drbd_rs_failed_io() - Record information on a failure to resync the specified blocks
- * @device:	DRBD device.
- * @sector:	The sector number.
- * @size:	Size of failed IO operation, in byte.
- */
-void drbd_rs_failed_io(struct drbd_device *device, sector_t sector, int size)
-{
-	/* Is called from worker and receiver context _only_ */
-	unsigned long sbnr, ebnr, lbnr;
-	unsigned long count;
-	sector_t esector, nr_sectors;
-	int wake_up = 0;
-
-	if (size <= 0 || !IS_ALIGNED(size, 512) || size > DRBD_MAX_DISCARD_SIZE) {
-		drbd_err(device, "drbd_rs_failed_io: sector=%llus size=%d nonsense!\n",
-				(unsigned long long)sector, size);
-		return;
-	}
-	nr_sectors = drbd_get_capacity(device->this_bdev);
-	esector = sector + (size >> 9) - 1;
-
-	if (!expect(sector < nr_sectors))
-		return;
-	if (!expect(esector < nr_sectors))
-		esector = nr_sectors - 1;
-
-	lbnr = BM_SECT_TO_BIT(nr_sectors-1);
-
-	/*
-	 * round up start sector, round down end sector.  we make sure we only
-	 * handle full, aligned, BM_BLOCK_SIZE (4K) blocks */
-	if (unlikely(esector < BM_SECT_PER_BIT-1))
-		return;
-	if (unlikely(esector == (nr_sectors-1)))
-		ebnr = lbnr;
-	else
-		ebnr = BM_SECT_TO_BIT(esector - (BM_SECT_PER_BIT-1));
-	sbnr = BM_SECT_TO_BIT(sector + BM_SECT_PER_BIT-1);
-
-	if (sbnr > ebnr)
-		return;
-
-	/*
-	 * ok, (capacity & 7) != 0 sometimes, but who cares...
-	 * we count rs_{total,left} in bits, not sectors.
-	 */
-	spin_lock_irq(&device->al_lock);
-	count = drbd_bm_count_bits(device, sbnr, ebnr);
-	if (count) {
-		device->rs_failed += count;
-
-		if (get_ldev(device)) {
-			drbd_try_clear_on_disk_bm(device, sector, count, false);
-			put_ldev(device);
-		}
-
-		/* just wake_up unconditional now, various lc_chaged(),
-		 * lc_put() in drbd_try_clear_on_disk_bm(). */
-		wake_up = 1;
-	}
-	spin_unlock_irq(&device->al_lock);
-	if (wake_up)
-		wake_up(&device->al_wait);
-}
diff --git a/drivers/block/drbd/drbd_bitmap.c b/drivers/block/drbd/drbd_bitmap.c
index 1aa29f8..426c97a 100644
--- a/drivers/block/drbd/drbd_bitmap.c
+++ b/drivers/block/drbd/drbd_bitmap.c
@@ -22,6 +22,8 @@
    the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
  */
 
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
 #include <linux/bitops.h>
 #include <linux/vmalloc.h>
 #include <linux/string.h>
@@ -353,9 +355,8 @@
 
 	for (i = 0; i < number; i++) {
 		if (!pages[i]) {
-			printk(KERN_ALERT "drbd: bm_free_pages tried to free "
-					  "a NULL pointer; i=%lu n=%lu\n",
-					  i, number);
+			pr_alert("bm_free_pages tried to free a NULL pointer; i=%lu n=%lu\n",
+				 i, number);
 			continue;
 		}
 		__free_page(pages[i]);
@@ -592,7 +593,7 @@
 	end = offset + len;
 
 	if (end > b->bm_words) {
-		printk(KERN_ALERT "drbd: bm_memset end > bm_words\n");
+		pr_alert("bm_memset end > bm_words\n");
 		return;
 	}
 
@@ -602,7 +603,7 @@
 		p_addr = bm_map_pidx(b, idx);
 		bm = p_addr + MLPP(offset);
 		if (bm+do_now > p_addr + LWPP) {
-			printk(KERN_ALERT "drbd: BUG BUG BUG! p_addr:%p bm:%p do_now:%d\n",
+			pr_alert("BUG BUG BUG! p_addr:%p bm:%p do_now:%d\n",
 			       p_addr, bm, (int)do_now);
 		} else
 			memset(bm, c, do_now * sizeof(long));
@@ -927,22 +928,14 @@
 	spin_unlock_irq(&b->bm_lock);
 }
 
-struct bm_aio_ctx {
-	struct drbd_device *device;
-	atomic_t in_flight;
-	unsigned int done;
-	unsigned flags;
-#define BM_AIO_COPY_PAGES	1
-#define BM_AIO_WRITE_HINTED	2
-#define BM_WRITE_ALL_PAGES	4
-	int error;
-	struct kref kref;
-};
-
-static void bm_aio_ctx_destroy(struct kref *kref)
+static void drbd_bm_aio_ctx_destroy(struct kref *kref)
 {
-	struct bm_aio_ctx *ctx = container_of(kref, struct bm_aio_ctx, kref);
+	struct drbd_bm_aio_ctx *ctx = container_of(kref, struct drbd_bm_aio_ctx, kref);
+	unsigned long flags;
 
+	spin_lock_irqsave(&ctx->device->resource->req_lock, flags);
+	list_del(&ctx->list);
+	spin_unlock_irqrestore(&ctx->device->resource->req_lock, flags);
 	put_ldev(ctx->device);
 	kfree(ctx);
 }
@@ -950,7 +943,7 @@
 /* bv_page may be a copy, or may be the original */
 static void bm_async_io_complete(struct bio *bio, int error)
 {
-	struct bm_aio_ctx *ctx = bio->bi_private;
+	struct drbd_bm_aio_ctx *ctx = bio->bi_private;
 	struct drbd_device *device = ctx->device;
 	struct drbd_bitmap *b = device->bitmap;
 	unsigned int idx = bm_page_to_idx(bio->bi_io_vec[0].bv_page);
@@ -993,17 +986,18 @@
 	if (atomic_dec_and_test(&ctx->in_flight)) {
 		ctx->done = 1;
 		wake_up(&device->misc_wait);
-		kref_put(&ctx->kref, &bm_aio_ctx_destroy);
+		kref_put(&ctx->kref, &drbd_bm_aio_ctx_destroy);
 	}
 }
 
-static void bm_page_io_async(struct bm_aio_ctx *ctx, int page_nr, int rw) __must_hold(local)
+static void bm_page_io_async(struct drbd_bm_aio_ctx *ctx, int page_nr) __must_hold(local)
 {
 	struct bio *bio = bio_alloc_drbd(GFP_NOIO);
 	struct drbd_device *device = ctx->device;
 	struct drbd_bitmap *b = device->bitmap;
 	struct page *page;
 	unsigned int len;
+	unsigned int rw = (ctx->flags & BM_AIO_READ) ? READ : WRITE;
 
 	sector_t on_disk_sector =
 		device->ldev->md.md_offset + device->ldev->md.bm_offset;
@@ -1049,9 +1043,9 @@
 /*
  * bm_rw: read/write the whole bitmap from/to its on disk location.
  */
-static int bm_rw(struct drbd_device *device, int rw, unsigned flags, unsigned lazy_writeout_upper_idx) __must_hold(local)
+static int bm_rw(struct drbd_device *device, const unsigned int flags, unsigned lazy_writeout_upper_idx) __must_hold(local)
 {
-	struct bm_aio_ctx *ctx;
+	struct drbd_bm_aio_ctx *ctx;
 	struct drbd_bitmap *b = device->bitmap;
 	int num_pages, i, count = 0;
 	unsigned long now;
@@ -1067,12 +1061,13 @@
 	 * as we submit copies of pages anyways.
 	 */
 
-	ctx = kmalloc(sizeof(struct bm_aio_ctx), GFP_NOIO);
+	ctx = kmalloc(sizeof(struct drbd_bm_aio_ctx), GFP_NOIO);
 	if (!ctx)
 		return -ENOMEM;
 
-	*ctx = (struct bm_aio_ctx) {
+	*ctx = (struct drbd_bm_aio_ctx) {
 		.device = device,
+		.start_jif = jiffies,
 		.in_flight = ATOMIC_INIT(1),
 		.done = 0,
 		.flags = flags,
@@ -1080,15 +1075,21 @@
 		.kref = { ATOMIC_INIT(2) },
 	};
 
-	if (!get_ldev_if_state(device, D_ATTACHING)) {  /* put is in bm_aio_ctx_destroy() */
+	if (!get_ldev_if_state(device, D_ATTACHING)) {  /* put is in drbd_bm_aio_ctx_destroy() */
 		drbd_err(device, "ASSERT FAILED: get_ldev_if_state() == 1 in bm_rw()\n");
 		kfree(ctx);
 		return -ENODEV;
 	}
+	/* Here D_ATTACHING is sufficient since drbd_bm_read() is called only from
+	   drbd_adm_attach(), after device->ldev was assigned. */
 
-	if (!ctx->flags)
+	if (0 == (ctx->flags & ~BM_AIO_READ))
 		WARN_ON(!(BM_LOCKED_MASK & b->bm_flags));
 
+	spin_lock_irq(&device->resource->req_lock);
+	list_add_tail(&ctx->list, &device->pending_bitmap_io);
+	spin_unlock_irq(&device->resource->req_lock);
+
 	num_pages = b->bm_number_of_pages;
 
 	now = jiffies;
@@ -1098,13 +1099,13 @@
 		/* ignore completely unchanged pages */
 		if (lazy_writeout_upper_idx && i == lazy_writeout_upper_idx)
 			break;
-		if (rw & WRITE) {
+		if (!(flags & BM_AIO_READ)) {
 			if ((flags & BM_AIO_WRITE_HINTED) &&
 			    !test_and_clear_bit(BM_PAGE_HINT_WRITEOUT,
 				    &page_private(b->bm_pages[i])))
 				continue;
 
-			if (!(flags & BM_WRITE_ALL_PAGES) &&
+			if (!(flags & BM_AIO_WRITE_ALL_PAGES) &&
 			    bm_test_page_unchanged(b->bm_pages[i])) {
 				dynamic_drbd_dbg(device, "skipped bm write for idx %u\n", i);
 				continue;
@@ -1118,7 +1119,7 @@
 			}
 		}
 		atomic_inc(&ctx->in_flight);
-		bm_page_io_async(ctx, i, rw);
+		bm_page_io_async(ctx, i);
 		++count;
 		cond_resched();
 	}
@@ -1134,12 +1135,12 @@
 	if (!atomic_dec_and_test(&ctx->in_flight))
 		wait_until_done_or_force_detached(device, device->ldev, &ctx->done);
 	else
-		kref_put(&ctx->kref, &bm_aio_ctx_destroy);
+		kref_put(&ctx->kref, &drbd_bm_aio_ctx_destroy);
 
 	/* summary for global bitmap IO */
 	if (flags == 0)
 		drbd_info(device, "bitmap %s of %u pages took %lu jiffies\n",
-			 rw == WRITE ? "WRITE" : "READ",
+			 (flags & BM_AIO_READ) ? "READ" : "WRITE",
 			 count, jiffies - now);
 
 	if (ctx->error) {
@@ -1152,20 +1153,18 @@
 		err = -EIO; /* Disk timeout/force-detach during IO... */
 
 	now = jiffies;
-	if (rw == WRITE) {
-		drbd_md_flush(device);
-	} else /* rw == READ */ {
+	if (flags & BM_AIO_READ) {
 		b->bm_set = bm_count_bits(b);
 		drbd_info(device, "recounting of set bits took additional %lu jiffies\n",
 		     jiffies - now);
 	}
 	now = b->bm_set;
 
-	if (flags == 0)
+	if ((flags & ~BM_AIO_READ) == 0)
 		drbd_info(device, "%s (%lu bits) marked out-of-sync by on disk bit-map.\n",
 		     ppsize(ppb, now << (BM_BLOCK_SHIFT-10)), now);
 
-	kref_put(&ctx->kref, &bm_aio_ctx_destroy);
+	kref_put(&ctx->kref, &drbd_bm_aio_ctx_destroy);
 	return err;
 }
 
@@ -1175,7 +1174,7 @@
  */
 int drbd_bm_read(struct drbd_device *device) __must_hold(local)
 {
-	return bm_rw(device, READ, 0, 0);
+	return bm_rw(device, BM_AIO_READ, 0);
 }
 
 /**
@@ -1186,7 +1185,7 @@
  */
 int drbd_bm_write(struct drbd_device *device) __must_hold(local)
 {
-	return bm_rw(device, WRITE, 0, 0);
+	return bm_rw(device, 0, 0);
 }
 
 /**
@@ -1197,7 +1196,17 @@
  */
 int drbd_bm_write_all(struct drbd_device *device) __must_hold(local)
 {
-	return bm_rw(device, WRITE, BM_WRITE_ALL_PAGES, 0);
+	return bm_rw(device, BM_AIO_WRITE_ALL_PAGES, 0);
+}
+
+/**
+ * drbd_bm_write_lazy() - Write bitmap pages 0 to @upper_idx-1, if they have changed.
+ * @device:	DRBD device.
+ * @upper_idx:	0: write all changed pages; +ve: page index to stop scanning for changed pages
+ */
+int drbd_bm_write_lazy(struct drbd_device *device, unsigned upper_idx) __must_hold(local)
+{
+	return bm_rw(device, BM_AIO_COPY_PAGES, upper_idx);
 }
 
 /**
@@ -1213,7 +1222,7 @@
  */
 int drbd_bm_write_copy_pages(struct drbd_device *device) __must_hold(local)
 {
-	return bm_rw(device, WRITE, BM_AIO_COPY_PAGES, 0);
+	return bm_rw(device, BM_AIO_COPY_PAGES, 0);
 }
 
 /**
@@ -1222,62 +1231,7 @@
  */
 int drbd_bm_write_hinted(struct drbd_device *device) __must_hold(local)
 {
-	return bm_rw(device, WRITE, BM_AIO_WRITE_HINTED | BM_AIO_COPY_PAGES, 0);
-}
-
-/**
- * drbd_bm_write_page() - Writes a PAGE_SIZE aligned piece of bitmap
- * @device:	DRBD device.
- * @idx:	bitmap page index
- *
- * We don't want to special case on logical_block_size of the backend device,
- * so we submit PAGE_SIZE aligned pieces.
- * Note that on "most" systems, PAGE_SIZE is 4k.
- *
- * In case this becomes an issue on systems with larger PAGE_SIZE,
- * we may want to change this again to write 4k aligned 4k pieces.
- */
-int drbd_bm_write_page(struct drbd_device *device, unsigned int idx) __must_hold(local)
-{
-	struct bm_aio_ctx *ctx;
-	int err;
-
-	if (bm_test_page_unchanged(device->bitmap->bm_pages[idx])) {
-		dynamic_drbd_dbg(device, "skipped bm page write for idx %u\n", idx);
-		return 0;
-	}
-
-	ctx = kmalloc(sizeof(struct bm_aio_ctx), GFP_NOIO);
-	if (!ctx)
-		return -ENOMEM;
-
-	*ctx = (struct bm_aio_ctx) {
-		.device = device,
-		.in_flight = ATOMIC_INIT(1),
-		.done = 0,
-		.flags = BM_AIO_COPY_PAGES,
-		.error = 0,
-		.kref = { ATOMIC_INIT(2) },
-	};
-
-	if (!get_ldev_if_state(device, D_ATTACHING)) {  /* put is in bm_aio_ctx_destroy() */
-		drbd_err(device, "ASSERT FAILED: get_ldev_if_state() == 1 in drbd_bm_write_page()\n");
-		kfree(ctx);
-		return -ENODEV;
-	}
-
-	bm_page_io_async(ctx, idx, WRITE_SYNC);
-	wait_until_done_or_force_detached(device, device->ldev, &ctx->done);
-
-	if (ctx->error)
-		drbd_chk_io_error(device, 1, DRBD_META_IO_ERROR);
-		/* that causes us to detach, so the in memory bitmap will be
-		 * gone in a moment as well. */
-
-	device->bm_writ_cnt++;
-	err = atomic_read(&ctx->in_flight) ? -EIO : ctx->error;
-	kref_put(&ctx->kref, &bm_aio_ctx_destroy);
-	return err;
+	return bm_rw(device, BM_AIO_WRITE_HINTED | BM_AIO_COPY_PAGES, 0);
 }
 
 /* NOTE
diff --git a/drivers/block/drbd/drbd_debugfs.c b/drivers/block/drbd/drbd_debugfs.c
new file mode 100644
index 0000000..5c20b18
--- /dev/null
+++ b/drivers/block/drbd/drbd_debugfs.c
@@ -0,0 +1,958 @@
+#define pr_fmt(fmt) "drbd debugfs: " fmt
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+#include <linux/stat.h>
+#include <linux/jiffies.h>
+#include <linux/list.h>
+
+#include "drbd_int.h"
+#include "drbd_req.h"
+#include "drbd_debugfs.h"
+
+
+/**********************************************************************
+ * Whenever you change the file format, remember to bump the version. *
+ **********************************************************************/
+
+static struct dentry *drbd_debugfs_root;
+static struct dentry *drbd_debugfs_version;
+static struct dentry *drbd_debugfs_resources;
+static struct dentry *drbd_debugfs_minors;
+
+static void seq_print_age_or_dash(struct seq_file *m, bool valid, unsigned long dt)
+{
+	if (valid)
+		seq_printf(m, "\t%d", jiffies_to_msecs(dt));
+	else
+		seq_printf(m, "\t-");
+}
+
+static void __seq_print_rq_state_bit(struct seq_file *m,
+	bool is_set, char *sep, const char *set_name, const char *unset_name)
+{
+	if (is_set && set_name) {
+		seq_putc(m, *sep);
+		seq_puts(m, set_name);
+		*sep = '|';
+	} else if (!is_set && unset_name) {
+		seq_putc(m, *sep);
+		seq_puts(m, unset_name);
+		*sep = '|';
+	}
+}
+
+static void seq_print_rq_state_bit(struct seq_file *m,
+	bool is_set, char *sep, const char *set_name)
+{
+	__seq_print_rq_state_bit(m, is_set, sep, set_name, NULL);
+}
+
+/* pretty print enum drbd_req_state_bits req->rq_state */
+static void seq_print_request_state(struct seq_file *m, struct drbd_request *req)
+{
+	unsigned int s = req->rq_state;
+	char sep = ' ';
+	seq_printf(m, "\t0x%08x", s);
+	seq_printf(m, "\tmaster: %s", req->master_bio ? "pending" : "completed");
+
+	/* RQ_WRITE ignored, already reported */
+	seq_puts(m, "\tlocal:");
+	seq_print_rq_state_bit(m, s & RQ_IN_ACT_LOG, &sep, "in-AL");
+	seq_print_rq_state_bit(m, s & RQ_POSTPONED, &sep, "postponed");
+	seq_print_rq_state_bit(m, s & RQ_COMPLETION_SUSP, &sep, "suspended");
+	sep = ' ';
+	seq_print_rq_state_bit(m, s & RQ_LOCAL_PENDING, &sep, "pending");
+	seq_print_rq_state_bit(m, s & RQ_LOCAL_COMPLETED, &sep, "completed");
+	seq_print_rq_state_bit(m, s & RQ_LOCAL_ABORTED, &sep, "aborted");
+	seq_print_rq_state_bit(m, s & RQ_LOCAL_OK, &sep, "ok");
+	if (sep == ' ')
+		seq_puts(m, " -");
+
+	/* for_each_connection ... */
+	seq_printf(m, "\tnet:");
+	sep = ' ';
+	seq_print_rq_state_bit(m, s & RQ_NET_PENDING, &sep, "pending");
+	seq_print_rq_state_bit(m, s & RQ_NET_QUEUED, &sep, "queued");
+	seq_print_rq_state_bit(m, s & RQ_NET_SENT, &sep, "sent");
+	seq_print_rq_state_bit(m, s & RQ_NET_DONE, &sep, "done");
+	seq_print_rq_state_bit(m, s & RQ_NET_SIS, &sep, "sis");
+	seq_print_rq_state_bit(m, s & RQ_NET_OK, &sep, "ok");
+	if (sep == ' ')
+		seq_puts(m, " -");
+
+	seq_printf(m, " :");
+	sep = ' ';
+	seq_print_rq_state_bit(m, s & RQ_EXP_RECEIVE_ACK, &sep, "B");
+	seq_print_rq_state_bit(m, s & RQ_EXP_WRITE_ACK, &sep, "C");
+	seq_print_rq_state_bit(m, s & RQ_EXP_BARR_ACK, &sep, "barr");
+	if (sep == ' ')
+		seq_puts(m, " -");
+	seq_printf(m, "\n");
+}
+
+static void seq_print_one_request(struct seq_file *m, struct drbd_request *req, unsigned long now)
+{
+	/* change anything here, fixup header below! */
+	unsigned int s = req->rq_state;
+
+#define RQ_HDR_1 "epoch\tsector\tsize\trw"
+	seq_printf(m, "0x%x\t%llu\t%u\t%s",
+		req->epoch,
+		(unsigned long long)req->i.sector, req->i.size >> 9,
+		(s & RQ_WRITE) ? "W" : "R");
+
+#define RQ_HDR_2 "\tstart\tin AL\tsubmit"
+	seq_printf(m, "\t%d", jiffies_to_msecs(now - req->start_jif));
+	seq_print_age_or_dash(m, s & RQ_IN_ACT_LOG, now - req->in_actlog_jif);
+	seq_print_age_or_dash(m, s & RQ_LOCAL_PENDING, now - req->pre_submit_jif);
+
+#define RQ_HDR_3 "\tsent\tacked\tdone"
+	seq_print_age_or_dash(m, s & RQ_NET_SENT, now - req->pre_send_jif);
+	seq_print_age_or_dash(m, (s & RQ_NET_SENT) && !(s & RQ_NET_PENDING), now - req->acked_jif);
+	seq_print_age_or_dash(m, s & RQ_NET_DONE, now - req->net_done_jif);
+
+#define RQ_HDR_4 "\tstate\n"
+	seq_print_request_state(m, req);
+}
+#define RQ_HDR RQ_HDR_1 RQ_HDR_2 RQ_HDR_3 RQ_HDR_4
+
+static void seq_print_minor_vnr_req(struct seq_file *m, struct drbd_request *req, unsigned long now)
+{
+	seq_printf(m, "%u\t%u\t", req->device->minor, req->device->vnr);
+	seq_print_one_request(m, req, now);
+}
+
+static void seq_print_resource_pending_meta_io(struct seq_file *m, struct drbd_resource *resource, unsigned long now)
+{
+	struct drbd_device *device;
+	unsigned int i;
+
+	seq_puts(m, "minor\tvnr\tstart\tsubmit\tintent\n");
+	rcu_read_lock();
+	idr_for_each_entry(&resource->devices, device, i) {
+		struct drbd_md_io tmp;
+		/* In theory this is racy,
+		 * in the sense that there could have been a
+		 * drbd_md_put_buffer(); drbd_md_get_buffer();
+		 * between accessing these members here.  */
+		tmp = device->md_io;
+		if (atomic_read(&tmp.in_use)) {
+			seq_printf(m, "%u\t%u\t%d\t",
+				device->minor, device->vnr,
+				jiffies_to_msecs(now - tmp.start_jif));
+			if (time_before(tmp.submit_jif, tmp.start_jif))
+				seq_puts(m, "-\t");
+			else
+				seq_printf(m, "%d\t", jiffies_to_msecs(now - tmp.submit_jif));
+			seq_printf(m, "%s\n", tmp.current_use);
+		}
+	}
+	rcu_read_unlock();
+}
+
+static void seq_print_waiting_for_AL(struct seq_file *m, struct drbd_resource *resource, unsigned long now)
+{
+	struct drbd_device *device;
+	unsigned int i;
+
+	seq_puts(m, "minor\tvnr\tage\t#waiting\n");
+	rcu_read_lock();
+	idr_for_each_entry(&resource->devices, device, i) {
+		unsigned long jif;
+		struct drbd_request *req;
+		int n = atomic_read(&device->ap_actlog_cnt);
+		if (n) {
+			spin_lock_irq(&device->resource->req_lock);
+			req = list_first_entry_or_null(&device->pending_master_completion[1],
+				struct drbd_request, req_pending_master_completion);
+			/* if the oldest request does not wait for the activity log
+			 * it is not interesting for us here */
+			if (req && !(req->rq_state & RQ_IN_ACT_LOG))
+				jif = req->start_jif;
+			else
+				req = NULL;
+			spin_unlock_irq(&device->resource->req_lock);
+		}
+		if (n) {
+			seq_printf(m, "%u\t%u\t", device->minor, device->vnr);
+			if (req)
+				seq_printf(m, "%u\t", jiffies_to_msecs(now - jif));
+			else
+				seq_puts(m, "-\t");
+			seq_printf(m, "%u\n", n);
+		}
+	}
+	rcu_read_unlock();
+}
+
+static void seq_print_device_bitmap_io(struct seq_file *m, struct drbd_device *device, unsigned long now)
+{
+	struct drbd_bm_aio_ctx *ctx;
+	unsigned long start_jif;
+	unsigned int in_flight;
+	unsigned int flags;
+	spin_lock_irq(&device->resource->req_lock);
+	ctx = list_first_entry_or_null(&device->pending_bitmap_io, struct drbd_bm_aio_ctx, list);
+	if (ctx && ctx->done)
+		ctx = NULL;
+	if (ctx) {
+		start_jif = ctx->start_jif;
+		in_flight = atomic_read(&ctx->in_flight);
+		flags = ctx->flags;
+	}
+	spin_unlock_irq(&device->resource->req_lock);
+	if (ctx) {
+		seq_printf(m, "%u\t%u\t%c\t%u\t%u\n",
+			device->minor, device->vnr,
+			(flags & BM_AIO_READ) ? 'R' : 'W',
+			jiffies_to_msecs(now - start_jif),
+			in_flight);
+	}
+}
+
+static void seq_print_resource_pending_bitmap_io(struct seq_file *m, struct drbd_resource *resource, unsigned long now)
+{
+	struct drbd_device *device;
+	unsigned int i;
+
+	seq_puts(m, "minor\tvnr\trw\tage\t#in-flight\n");
+	rcu_read_lock();
+	idr_for_each_entry(&resource->devices, device, i) {
+		seq_print_device_bitmap_io(m, device, now);
+	}
+	rcu_read_unlock();
+}
+
+/* pretty print enum peer_req->flags */
+static void seq_print_peer_request_flags(struct seq_file *m, struct drbd_peer_request *peer_req)
+{
+	unsigned long f = peer_req->flags;
+	char sep = ' ';
+
+	__seq_print_rq_state_bit(m, f & EE_SUBMITTED, &sep, "submitted", "preparing");
+	__seq_print_rq_state_bit(m, f & EE_APPLICATION, &sep, "application", "internal");
+	seq_print_rq_state_bit(m, f & EE_CALL_AL_COMPLETE_IO, &sep, "in-AL");
+	seq_print_rq_state_bit(m, f & EE_SEND_WRITE_ACK, &sep, "C");
+	seq_print_rq_state_bit(m, f & EE_MAY_SET_IN_SYNC, &sep, "set-in-sync");
+
+	if (f & EE_IS_TRIM) {
+		seq_putc(m, sep);
+		sep = '|';
+		if (f & EE_IS_TRIM_USE_ZEROOUT)
+			seq_puts(m, "zero-out");
+		else
+			seq_puts(m, "trim");
+	}
+	seq_putc(m, '\n');
+}
+
+static void seq_print_peer_request(struct seq_file *m,
+	struct drbd_device *device, struct list_head *lh,
+	unsigned long now)
+{
+	bool reported_preparing = false;
+	struct drbd_peer_request *peer_req;
+	list_for_each_entry(peer_req, lh, w.list) {
+		if (reported_preparing && !(peer_req->flags & EE_SUBMITTED))
+			continue;
+
+		if (device)
+			seq_printf(m, "%u\t%u\t", device->minor, device->vnr);
+
+		seq_printf(m, "%llu\t%u\t%c\t%u\t",
+			(unsigned long long)peer_req->i.sector, peer_req->i.size >> 9,
+			(peer_req->flags & EE_WRITE) ? 'W' : 'R',
+			jiffies_to_msecs(now - peer_req->submit_jif));
+		seq_print_peer_request_flags(m, peer_req);
+		if (peer_req->flags & EE_SUBMITTED)
+			break;
+		else
+			reported_preparing = true;
+	}
+}
+
+static void seq_print_device_peer_requests(struct seq_file *m,
+	struct drbd_device *device, unsigned long now)
+{
+	seq_puts(m, "minor\tvnr\tsector\tsize\trw\tage\tflags\n");
+	spin_lock_irq(&device->resource->req_lock);
+	seq_print_peer_request(m, device, &device->active_ee, now);
+	seq_print_peer_request(m, device, &device->read_ee, now);
+	seq_print_peer_request(m, device, &device->sync_ee, now);
+	spin_unlock_irq(&device->resource->req_lock);
+	if (test_bit(FLUSH_PENDING, &device->flags)) {
+		seq_printf(m, "%u\t%u\t-\t-\tF\t%u\tflush\n",
+			device->minor, device->vnr,
+			jiffies_to_msecs(now - device->flush_jif));
+	}
+}
+
+static void seq_print_resource_pending_peer_requests(struct seq_file *m,
+	struct drbd_resource *resource, unsigned long now)
+{
+	struct drbd_device *device;
+	unsigned int i;
+
+	rcu_read_lock();
+	idr_for_each_entry(&resource->devices, device, i) {
+		seq_print_device_peer_requests(m, device, now);
+	}
+	rcu_read_unlock();
+}
+
+static void seq_print_resource_transfer_log_summary(struct seq_file *m,
+	struct drbd_resource *resource,
+	struct drbd_connection *connection,
+	unsigned long now)
+{
+	struct drbd_request *req;
+	unsigned int count = 0;
+	unsigned int show_state = 0;
+
+	seq_puts(m, "n\tdevice\tvnr\t" RQ_HDR);
+	spin_lock_irq(&resource->req_lock);
+	list_for_each_entry(req, &connection->transfer_log, tl_requests) {
+		unsigned int tmp = 0;
+		unsigned int s;
+		++count;
+
+		/* don't disable irq "forever" */
+		if (!(count & 0x1ff)) {
+			struct drbd_request *req_next;
+			kref_get(&req->kref);
+			spin_unlock_irq(&resource->req_lock);
+			cond_resched();
+			spin_lock_irq(&resource->req_lock);
+			req_next = list_next_entry(req, tl_requests);
+			if (kref_put(&req->kref, drbd_req_destroy))
+				req = req_next;
+			if (&req->tl_requests == &connection->transfer_log)
+				break;
+		}
+
+		s = req->rq_state;
+
+		/* This is meant to summarize timing issues, to be able to tell
+		 * local disk problems from network problems.
+		 * Skip requests, if we have shown an even older request with
+		 * similar aspects already.  */
+		if (req->master_bio == NULL)
+			tmp |= 1;
+		if ((s & RQ_LOCAL_MASK) && (s & RQ_LOCAL_PENDING))
+			tmp |= 2;
+		if (s & RQ_NET_MASK) {
+			if (!(s & RQ_NET_SENT))
+				tmp |= 4;
+			if (s & RQ_NET_PENDING)
+				tmp |= 8;
+			if (!(s & RQ_NET_DONE))
+				tmp |= 16;
+		}
+		if ((tmp & show_state) == tmp)
+			continue;
+		show_state |= tmp;
+		seq_printf(m, "%u\t", count);
+		seq_print_minor_vnr_req(m, req, now);
+		if (show_state == 0x1f)
+			break;
+	}
+	spin_unlock_irq(&resource->req_lock);
+}
+
+/* TODO: transfer_log and friends should be moved to resource */
+static int in_flight_summary_show(struct seq_file *m, void *pos)
+{
+	struct drbd_resource *resource = m->private;
+	struct drbd_connection *connection;
+	unsigned long jif = jiffies;
+
+	connection = first_connection(resource);
+	/* This does not happen, actually.
+	 * But be robust and prepare for future code changes. */
+	if (!connection || !kref_get_unless_zero(&connection->kref))
+		return -ESTALE;
+
+	/* BUMP me if you change the file format/content/presentation */
+	seq_printf(m, "v: %u\n\n", 0);
+
+	seq_puts(m, "oldest bitmap IO\n");
+	seq_print_resource_pending_bitmap_io(m, resource, jif);
+	seq_putc(m, '\n');
+
+	seq_puts(m, "meta data IO\n");
+	seq_print_resource_pending_meta_io(m, resource, jif);
+	seq_putc(m, '\n');
+
+	seq_puts(m, "socket buffer stats\n");
+	/* for each connection ... once we have more than one */
+	rcu_read_lock();
+	if (connection->data.socket) {
+		/* open coded SIOCINQ, the "relevant" part */
+		struct tcp_sock *tp = tcp_sk(connection->data.socket->sk);
+		int answ = tp->rcv_nxt - tp->copied_seq;
+		seq_printf(m, "unread receive buffer: %u Byte\n", answ);
+		/* open coded SIOCOUTQ, the "relevant" part */
+		answ = tp->write_seq - tp->snd_una;
+		seq_printf(m, "unacked send buffer: %u Byte\n", answ);
+	}
+	rcu_read_unlock();
+	seq_putc(m, '\n');
+
+	seq_puts(m, "oldest peer requests\n");
+	seq_print_resource_pending_peer_requests(m, resource, jif);
+	seq_putc(m, '\n');
+
+	seq_puts(m, "application requests waiting for activity log\n");
+	seq_print_waiting_for_AL(m, resource, jif);
+	seq_putc(m, '\n');
+
+	seq_puts(m, "oldest application requests\n");
+	seq_print_resource_transfer_log_summary(m, resource, connection, jif);
+	seq_putc(m, '\n');
+
+	jif = jiffies - jif;
+	if (jif)
+		seq_printf(m, "generated in %d ms\n", jiffies_to_msecs(jif));
+	kref_put(&connection->kref, drbd_destroy_connection);
+	return 0;
+}
+
+/* simple_positive(file->f_dentry) respectively debugfs_positive(),
+ * but neither is "reachable" from here.
+ * So we have our own inline version of it above.  :-( */
+static inline int debugfs_positive(struct dentry *dentry)
+{
+        return dentry->d_inode && !d_unhashed(dentry);
+}
+
+/* make sure at *open* time that the respective object won't go away. */
+static int drbd_single_open(struct file *file, int (*show)(struct seq_file *, void *),
+		                void *data, struct kref *kref,
+				void (*release)(struct kref *))
+{
+	struct dentry *parent;
+	int ret = -ESTALE;
+
+	/* Are we still linked,
+	 * or has debugfs_remove() already been called? */
+	parent = file->f_dentry->d_parent;
+	/* not sure if this can happen: */
+	if (!parent || !parent->d_inode)
+		goto out;
+	/* serialize with d_delete() */
+	mutex_lock(&parent->d_inode->i_mutex);
+	/* Make sure the object is still alive */
+	if (debugfs_positive(file->f_dentry)
+	&& kref_get_unless_zero(kref))
+		ret = 0;
+	mutex_unlock(&parent->d_inode->i_mutex);
+	if (!ret) {
+		ret = single_open(file, show, data);
+		if (ret)
+			kref_put(kref, release);
+	}
+out:
+	return ret;
+}
+
+static int in_flight_summary_open(struct inode *inode, struct file *file)
+{
+	struct drbd_resource *resource = inode->i_private;
+	return drbd_single_open(file, in_flight_summary_show, resource,
+				&resource->kref, drbd_destroy_resource);
+}
+
+static int in_flight_summary_release(struct inode *inode, struct file *file)
+{
+	struct drbd_resource *resource = inode->i_private;
+	kref_put(&resource->kref, drbd_destroy_resource);
+	return single_release(inode, file);
+}
+
+static const struct file_operations in_flight_summary_fops = {
+	.owner		= THIS_MODULE,
+	.open		= in_flight_summary_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= in_flight_summary_release,
+};
+
+void drbd_debugfs_resource_add(struct drbd_resource *resource)
+{
+	struct dentry *dentry;
+	if (!drbd_debugfs_resources)
+		return;
+
+	dentry = debugfs_create_dir(resource->name, drbd_debugfs_resources);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	resource->debugfs_res = dentry;
+
+	dentry = debugfs_create_dir("volumes", resource->debugfs_res);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	resource->debugfs_res_volumes = dentry;
+
+	dentry = debugfs_create_dir("connections", resource->debugfs_res);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	resource->debugfs_res_connections = dentry;
+
+	dentry = debugfs_create_file("in_flight_summary", S_IRUSR|S_IRGRP,
+			resource->debugfs_res, resource,
+			&in_flight_summary_fops);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	resource->debugfs_res_in_flight_summary = dentry;
+	return;
+
+fail:
+	drbd_debugfs_resource_cleanup(resource);
+	drbd_err(resource, "failed to create debugfs dentry\n");
+}
+
+static void drbd_debugfs_remove(struct dentry **dp)
+{
+	debugfs_remove(*dp);
+	*dp = NULL;
+}
+
+void drbd_debugfs_resource_cleanup(struct drbd_resource *resource)
+{
+	/* it is ok to call debugfs_remove(NULL) */
+	drbd_debugfs_remove(&resource->debugfs_res_in_flight_summary);
+	drbd_debugfs_remove(&resource->debugfs_res_connections);
+	drbd_debugfs_remove(&resource->debugfs_res_volumes);
+	drbd_debugfs_remove(&resource->debugfs_res);
+}
+
+static void seq_print_one_timing_detail(struct seq_file *m,
+	const struct drbd_thread_timing_details *tdp,
+	unsigned long now)
+{
+	struct drbd_thread_timing_details td;
+	/* No locking...
+	 * use temporary assignment to get at consistent data. */
+	do {
+		td = *tdp;
+	} while (td.cb_nr != tdp->cb_nr);
+	if (!td.cb_addr)
+		return;
+	seq_printf(m, "%u\t%d\t%s:%u\t%ps\n",
+			td.cb_nr,
+			jiffies_to_msecs(now - td.start_jif),
+			td.caller_fn, td.line,
+			td.cb_addr);
+}
+
+static void seq_print_timing_details(struct seq_file *m,
+		const char *title,
+		unsigned int cb_nr, struct drbd_thread_timing_details *tdp, unsigned long now)
+{
+	unsigned int start_idx;
+	unsigned int i;
+
+	seq_printf(m, "%s\n", title);
+	/* If not much is going on, this will result in natural ordering.
+	 * If it is very busy, we will possibly skip events, or even see wrap
+	 * arounds, which could only be avoided with locking.
+	 */
+	start_idx = cb_nr % DRBD_THREAD_DETAILS_HIST;
+	for (i = start_idx; i < DRBD_THREAD_DETAILS_HIST; i++)
+		seq_print_one_timing_detail(m, tdp+i, now);
+	for (i = 0; i < start_idx; i++)
+		seq_print_one_timing_detail(m, tdp+i, now);
+}
+
+static int callback_history_show(struct seq_file *m, void *ignored)
+{
+	struct drbd_connection *connection = m->private;
+	unsigned long jif = jiffies;
+
+	/* BUMP me if you change the file format/content/presentation */
+	seq_printf(m, "v: %u\n\n", 0);
+
+	seq_puts(m, "n\tage\tcallsite\tfn\n");
+	seq_print_timing_details(m, "worker", connection->w_cb_nr, connection->w_timing_details, jif);
+	seq_print_timing_details(m, "receiver", connection->r_cb_nr, connection->r_timing_details, jif);
+	return 0;
+}
+
+static int callback_history_open(struct inode *inode, struct file *file)
+{
+	struct drbd_connection *connection = inode->i_private;
+	return drbd_single_open(file, callback_history_show, connection,
+				&connection->kref, drbd_destroy_connection);
+}
+
+static int callback_history_release(struct inode *inode, struct file *file)
+{
+	struct drbd_connection *connection = inode->i_private;
+	kref_put(&connection->kref, drbd_destroy_connection);
+	return single_release(inode, file);
+}
+
+static const struct file_operations connection_callback_history_fops = {
+	.owner		= THIS_MODULE,
+	.open		= callback_history_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= callback_history_release,
+};
+
+static int connection_oldest_requests_show(struct seq_file *m, void *ignored)
+{
+	struct drbd_connection *connection = m->private;
+	unsigned long now = jiffies;
+	struct drbd_request *r1, *r2;
+
+	/* BUMP me if you change the file format/content/presentation */
+	seq_printf(m, "v: %u\n\n", 0);
+
+	spin_lock_irq(&connection->resource->req_lock);
+	r1 = connection->req_next;
+	if (r1)
+		seq_print_minor_vnr_req(m, r1, now);
+	r2 = connection->req_ack_pending;
+	if (r2 && r2 != r1) {
+		r1 = r2;
+		seq_print_minor_vnr_req(m, r1, now);
+	}
+	r2 = connection->req_not_net_done;
+	if (r2 && r2 != r1)
+		seq_print_minor_vnr_req(m, r2, now);
+	spin_unlock_irq(&connection->resource->req_lock);
+	return 0;
+}
+
+static int connection_oldest_requests_open(struct inode *inode, struct file *file)
+{
+	struct drbd_connection *connection = inode->i_private;
+	return drbd_single_open(file, connection_oldest_requests_show, connection,
+				&connection->kref, drbd_destroy_connection);
+}
+
+static int connection_oldest_requests_release(struct inode *inode, struct file *file)
+{
+	struct drbd_connection *connection = inode->i_private;
+	kref_put(&connection->kref, drbd_destroy_connection);
+	return single_release(inode, file);
+}
+
+static const struct file_operations connection_oldest_requests_fops = {
+	.owner		= THIS_MODULE,
+	.open		= connection_oldest_requests_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= connection_oldest_requests_release,
+};
+
+void drbd_debugfs_connection_add(struct drbd_connection *connection)
+{
+	struct dentry *conns_dir = connection->resource->debugfs_res_connections;
+	struct dentry *dentry;
+	if (!conns_dir)
+		return;
+
+	/* Once we enable mutliple peers,
+	 * these connections will have descriptive names.
+	 * For now, it is just the one connection to the (only) "peer". */
+	dentry = debugfs_create_dir("peer", conns_dir);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	connection->debugfs_conn = dentry;
+
+	dentry = debugfs_create_file("callback_history", S_IRUSR|S_IRGRP,
+			connection->debugfs_conn, connection,
+			&connection_callback_history_fops);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	connection->debugfs_conn_callback_history = dentry;
+
+	dentry = debugfs_create_file("oldest_requests", S_IRUSR|S_IRGRP,
+			connection->debugfs_conn, connection,
+			&connection_oldest_requests_fops);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	connection->debugfs_conn_oldest_requests = dentry;
+	return;
+
+fail:
+	drbd_debugfs_connection_cleanup(connection);
+	drbd_err(connection, "failed to create debugfs dentry\n");
+}
+
+void drbd_debugfs_connection_cleanup(struct drbd_connection *connection)
+{
+	drbd_debugfs_remove(&connection->debugfs_conn_callback_history);
+	drbd_debugfs_remove(&connection->debugfs_conn_oldest_requests);
+	drbd_debugfs_remove(&connection->debugfs_conn);
+}
+
+static void resync_dump_detail(struct seq_file *m, struct lc_element *e)
+{
+       struct bm_extent *bme = lc_entry(e, struct bm_extent, lce);
+
+       seq_printf(m, "%5d %s %s %s\n", bme->rs_left,
+		  test_bit(BME_NO_WRITES, &bme->flags) ? "NO_WRITES" : "---------",
+		  test_bit(BME_LOCKED, &bme->flags) ? "LOCKED" : "------",
+		  test_bit(BME_PRIORITY, &bme->flags) ? "PRIORITY" : "--------"
+		  );
+}
+
+static int device_resync_extents_show(struct seq_file *m, void *ignored)
+{
+	struct drbd_device *device = m->private;
+
+	/* BUMP me if you change the file format/content/presentation */
+	seq_printf(m, "v: %u\n\n", 0);
+
+	if (get_ldev_if_state(device, D_FAILED)) {
+		lc_seq_printf_stats(m, device->resync);
+		lc_seq_dump_details(m, device->resync, "rs_left flags", resync_dump_detail);
+		put_ldev(device);
+	}
+	return 0;
+}
+
+static int device_act_log_extents_show(struct seq_file *m, void *ignored)
+{
+	struct drbd_device *device = m->private;
+
+	/* BUMP me if you change the file format/content/presentation */
+	seq_printf(m, "v: %u\n\n", 0);
+
+	if (get_ldev_if_state(device, D_FAILED)) {
+		lc_seq_printf_stats(m, device->act_log);
+		lc_seq_dump_details(m, device->act_log, "", NULL);
+		put_ldev(device);
+	}
+	return 0;
+}
+
+static int device_oldest_requests_show(struct seq_file *m, void *ignored)
+{
+	struct drbd_device *device = m->private;
+	struct drbd_resource *resource = device->resource;
+	unsigned long now = jiffies;
+	struct drbd_request *r1, *r2;
+	int i;
+
+	/* BUMP me if you change the file format/content/presentation */
+	seq_printf(m, "v: %u\n\n", 0);
+
+	seq_puts(m, RQ_HDR);
+	spin_lock_irq(&resource->req_lock);
+	/* WRITE, then READ */
+	for (i = 1; i >= 0; --i) {
+		r1 = list_first_entry_or_null(&device->pending_master_completion[i],
+			struct drbd_request, req_pending_master_completion);
+		r2 = list_first_entry_or_null(&device->pending_completion[i],
+			struct drbd_request, req_pending_local);
+		if (r1)
+			seq_print_one_request(m, r1, now);
+		if (r2 && r2 != r1)
+			seq_print_one_request(m, r2, now);
+	}
+	spin_unlock_irq(&resource->req_lock);
+	return 0;
+}
+
+static int device_data_gen_id_show(struct seq_file *m, void *ignored)
+{
+	struct drbd_device *device = m->private;
+	struct drbd_md *md;
+	enum drbd_uuid_index idx;
+
+	if (!get_ldev_if_state(device, D_FAILED))
+		return -ENODEV;
+
+	md = &device->ldev->md;
+	spin_lock_irq(&md->uuid_lock);
+	for (idx = UI_CURRENT; idx <= UI_HISTORY_END; idx++) {
+		seq_printf(m, "0x%016llX\n", md->uuid[idx]);
+	}
+	spin_unlock_irq(&md->uuid_lock);
+	put_ldev(device);
+	return 0;
+}
+
+#define drbd_debugfs_device_attr(name)						\
+static int device_ ## name ## _open(struct inode *inode, struct file *file)	\
+{										\
+	struct drbd_device *device = inode->i_private;				\
+	return drbd_single_open(file, device_ ## name ## _show, device,		\
+				&device->kref, drbd_destroy_device);		\
+}										\
+static int device_ ## name ## _release(struct inode *inode, struct file *file)	\
+{										\
+	struct drbd_device *device = inode->i_private;				\
+	kref_put(&device->kref, drbd_destroy_device);				\
+	return single_release(inode, file);					\
+}										\
+static const struct file_operations device_ ## name ## _fops = {		\
+	.owner		= THIS_MODULE,						\
+	.open		= device_ ## name ## _open,				\
+	.read		= seq_read,						\
+	.llseek		= seq_lseek,						\
+	.release	= device_ ## name ## _release,				\
+};
+
+drbd_debugfs_device_attr(oldest_requests)
+drbd_debugfs_device_attr(act_log_extents)
+drbd_debugfs_device_attr(resync_extents)
+drbd_debugfs_device_attr(data_gen_id)
+
+void drbd_debugfs_device_add(struct drbd_device *device)
+{
+	struct dentry *vols_dir = device->resource->debugfs_res_volumes;
+	char minor_buf[8]; /* MINORMASK, MINORBITS == 20; */
+	char vnr_buf[8];   /* volume number vnr is even 16 bit only; */
+	char *slink_name = NULL;
+
+	struct dentry *dentry;
+	if (!vols_dir || !drbd_debugfs_minors)
+		return;
+
+	snprintf(vnr_buf, sizeof(vnr_buf), "%u", device->vnr);
+	dentry = debugfs_create_dir(vnr_buf, vols_dir);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	device->debugfs_vol = dentry;
+
+	snprintf(minor_buf, sizeof(minor_buf), "%u", device->minor);
+	slink_name = kasprintf(GFP_KERNEL, "../resources/%s/volumes/%u",
+			device->resource->name, device->vnr);
+	if (!slink_name)
+		goto fail;
+	dentry = debugfs_create_symlink(minor_buf, drbd_debugfs_minors, slink_name);
+	kfree(slink_name);
+	slink_name = NULL;
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	device->debugfs_minor = dentry;
+
+#define DCF(name)	do {					\
+	dentry = debugfs_create_file(#name, S_IRUSR|S_IRGRP,	\
+			device->debugfs_vol, device,		\
+			&device_ ## name ## _fops);		\
+	if (IS_ERR_OR_NULL(dentry))				\
+		goto fail;					\
+	device->debugfs_vol_ ## name = dentry;			\
+	} while (0)
+
+	DCF(oldest_requests);
+	DCF(act_log_extents);
+	DCF(resync_extents);
+	DCF(data_gen_id);
+#undef DCF
+	return;
+
+fail:
+	drbd_debugfs_device_cleanup(device);
+	drbd_err(device, "failed to create debugfs entries\n");
+}
+
+void drbd_debugfs_device_cleanup(struct drbd_device *device)
+{
+	drbd_debugfs_remove(&device->debugfs_minor);
+	drbd_debugfs_remove(&device->debugfs_vol_oldest_requests);
+	drbd_debugfs_remove(&device->debugfs_vol_act_log_extents);
+	drbd_debugfs_remove(&device->debugfs_vol_resync_extents);
+	drbd_debugfs_remove(&device->debugfs_vol_data_gen_id);
+	drbd_debugfs_remove(&device->debugfs_vol);
+}
+
+void drbd_debugfs_peer_device_add(struct drbd_peer_device *peer_device)
+{
+	struct dentry *conn_dir = peer_device->connection->debugfs_conn;
+	struct dentry *dentry;
+	char vnr_buf[8];
+
+	if (!conn_dir)
+		return;
+
+	snprintf(vnr_buf, sizeof(vnr_buf), "%u", peer_device->device->vnr);
+	dentry = debugfs_create_dir(vnr_buf, conn_dir);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	peer_device->debugfs_peer_dev = dentry;
+	return;
+
+fail:
+	drbd_debugfs_peer_device_cleanup(peer_device);
+	drbd_err(peer_device, "failed to create debugfs entries\n");
+}
+
+void drbd_debugfs_peer_device_cleanup(struct drbd_peer_device *peer_device)
+{
+	drbd_debugfs_remove(&peer_device->debugfs_peer_dev);
+}
+
+static int drbd_version_show(struct seq_file *m, void *ignored)
+{
+	seq_printf(m, "# %s\n", drbd_buildtag());
+	seq_printf(m, "VERSION=%s\n", REL_VERSION);
+	seq_printf(m, "API_VERSION=%u\n", API_VERSION);
+	seq_printf(m, "PRO_VERSION_MIN=%u\n", PRO_VERSION_MIN);
+	seq_printf(m, "PRO_VERSION_MAX=%u\n", PRO_VERSION_MAX);
+	return 0;
+}
+
+static int drbd_version_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, drbd_version_show, NULL);
+}
+
+static struct file_operations drbd_version_fops = {
+	.owner = THIS_MODULE,
+	.open = drbd_version_open,
+	.llseek = seq_lseek,
+	.read = seq_read,
+	.release = single_release,
+};
+
+/* not __exit, may be indirectly called
+ * from the module-load-failure path as well. */
+void drbd_debugfs_cleanup(void)
+{
+	drbd_debugfs_remove(&drbd_debugfs_resources);
+	drbd_debugfs_remove(&drbd_debugfs_minors);
+	drbd_debugfs_remove(&drbd_debugfs_version);
+	drbd_debugfs_remove(&drbd_debugfs_root);
+}
+
+int __init drbd_debugfs_init(void)
+{
+	struct dentry *dentry;
+
+	dentry = debugfs_create_dir("drbd", NULL);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	drbd_debugfs_root = dentry;
+
+	dentry = debugfs_create_file("version", 0444, drbd_debugfs_root, NULL, &drbd_version_fops);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	drbd_debugfs_version = dentry;
+
+	dentry = debugfs_create_dir("resources", drbd_debugfs_root);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	drbd_debugfs_resources = dentry;
+
+	dentry = debugfs_create_dir("minors", drbd_debugfs_root);
+	if (IS_ERR_OR_NULL(dentry))
+		goto fail;
+	drbd_debugfs_minors = dentry;
+	return 0;
+
+fail:
+	drbd_debugfs_cleanup();
+	if (dentry)
+		return PTR_ERR(dentry);
+	else
+		return -EINVAL;
+}
diff --git a/drivers/block/drbd/drbd_debugfs.h b/drivers/block/drbd/drbd_debugfs.h
new file mode 100644
index 0000000..8bee213
--- /dev/null
+++ b/drivers/block/drbd/drbd_debugfs.h
@@ -0,0 +1,39 @@
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/debugfs.h>
+
+#include "drbd_int.h"
+
+#ifdef CONFIG_DEBUG_FS
+int __init drbd_debugfs_init(void);
+void drbd_debugfs_cleanup(void);
+
+void drbd_debugfs_resource_add(struct drbd_resource *resource);
+void drbd_debugfs_resource_cleanup(struct drbd_resource *resource);
+
+void drbd_debugfs_connection_add(struct drbd_connection *connection);
+void drbd_debugfs_connection_cleanup(struct drbd_connection *connection);
+
+void drbd_debugfs_device_add(struct drbd_device *device);
+void drbd_debugfs_device_cleanup(struct drbd_device *device);
+
+void drbd_debugfs_peer_device_add(struct drbd_peer_device *peer_device);
+void drbd_debugfs_peer_device_cleanup(struct drbd_peer_device *peer_device);
+#else
+
+static inline int __init drbd_debugfs_init(void) { return -ENODEV; }
+static inline void drbd_debugfs_cleanup(void) { }
+
+static inline void drbd_debugfs_resource_add(struct drbd_resource *resource) { }
+static inline void drbd_debugfs_resource_cleanup(struct drbd_resource *resource) { }
+
+static inline void drbd_debugfs_connection_add(struct drbd_connection *connection) { }
+static inline void drbd_debugfs_connection_cleanup(struct drbd_connection *connection) { }
+
+static inline void drbd_debugfs_device_add(struct drbd_device *device) { }
+static inline void drbd_debugfs_device_cleanup(struct drbd_device *device) { }
+
+static inline void drbd_debugfs_peer_device_add(struct drbd_peer_device *peer_device) { }
+static inline void drbd_debugfs_peer_device_cleanup(struct drbd_peer_device *peer_device) { }
+
+#endif
diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index a76ceb3..1a00001 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -317,7 +317,63 @@
 
 	struct list_head tl_requests; /* ring list in the transfer log */
 	struct bio *master_bio;       /* master bio pointer */
-	unsigned long start_time;
+
+	/* see struct drbd_device */
+	struct list_head req_pending_master_completion;
+	struct list_head req_pending_local;
+
+	/* for generic IO accounting */
+	unsigned long start_jif;
+
+	/* for DRBD internal statistics */
+
+	/* Minimal set of time stamps to determine if we wait for activity log
+	 * transactions, local disk or peer.  32 bit "jiffies" are good enough,
+	 * we don't expect a DRBD request to be stalled for several month.
+	 */
+
+	/* before actual request processing */
+	unsigned long in_actlog_jif;
+
+	/* local disk */
+	unsigned long pre_submit_jif;
+
+	/* per connection */
+	unsigned long pre_send_jif;
+	unsigned long acked_jif;
+	unsigned long net_done_jif;
+
+	/* Possibly even more detail to track each phase:
+	 *  master_completion_jif
+	 *      how long did it take to complete the master bio
+	 *      (application visible latency)
+	 *  allocated_jif
+	 *      how long the master bio was blocked until we finally allocated
+	 *      a tracking struct
+	 *  in_actlog_jif
+	 *      how long did we wait for activity log transactions
+	 *
+	 *  net_queued_jif
+	 *      when did we finally queue it for sending
+	 *  pre_send_jif
+	 *      when did we start sending it
+	 *  post_send_jif
+	 *      how long did we block in the network stack trying to send it
+	 *  acked_jif
+	 *      when did we receive (or fake, in protocol A) a remote ACK
+	 *  net_done_jif
+	 *      when did we receive final acknowledgement (P_BARRIER_ACK),
+	 *      or decide, e.g. on connection loss, that we do no longer expect
+	 *      anything from this peer for this request.
+	 *
+	 *  pre_submit_jif
+	 *  post_sub_jif
+	 *      when did we start submiting to the lower level device,
+	 *      and how long did we block in that submit function
+	 *  local_completion_jif
+	 *      how long did it take the lower level device to complete this request
+	 */
+
 
 	/* once it hits 0, we may complete the master_bio */
 	atomic_t completion_ref;
@@ -366,6 +422,7 @@
 	struct drbd_interval i;
 	/* see comments on ee flag bits below */
 	unsigned long flags;
+	unsigned long submit_jif;
 	union {
 		u64 block_id;
 		struct digest_info *digest;
@@ -408,6 +465,17 @@
 
 	/* Is set when net_conf had two_primaries set while creating this peer_req */
 	__EE_IN_INTERVAL_TREE,
+
+	/* for debugfs: */
+	/* has this been submitted, or does it still wait for something else? */
+	__EE_SUBMITTED,
+
+	/* this is/was a write request */
+	__EE_WRITE,
+
+	/* this originates from application on peer
+	 * (not some resync or verify or other DRBD internal request) */
+	__EE_APPLICATION,
 };
 #define EE_CALL_AL_COMPLETE_IO (1<<__EE_CALL_AL_COMPLETE_IO)
 #define EE_MAY_SET_IN_SYNC     (1<<__EE_MAY_SET_IN_SYNC)
@@ -419,6 +487,9 @@
 #define EE_RESTART_REQUESTS	(1<<__EE_RESTART_REQUESTS)
 #define EE_SEND_WRITE_ACK	(1<<__EE_SEND_WRITE_ACK)
 #define EE_IN_INTERVAL_TREE	(1<<__EE_IN_INTERVAL_TREE)
+#define EE_SUBMITTED		(1<<__EE_SUBMITTED)
+#define EE_WRITE		(1<<__EE_WRITE)
+#define EE_APPLICATION		(1<<__EE_APPLICATION)
 
 /* flag bits per device */
 enum {
@@ -433,11 +504,11 @@
 	CONSIDER_RESYNC,
 
 	MD_NO_FUA,		/* Users wants us to not use FUA/FLUSH on meta data dev */
+
 	SUSPEND_IO,		/* suspend application io */
 	BITMAP_IO,		/* suspend application io;
 				   once no more io in flight, start bitmap io */
 	BITMAP_IO_QUEUED,       /* Started bitmap IO */
-	GO_DISKLESS,		/* Disk is being detached, on io-error or admin request. */
 	WAS_IO_ERROR,		/* Local disk failed, returned IO error */
 	WAS_READ_ERROR,		/* Local disk READ failed (set additionally to the above) */
 	FORCE_DETACH,		/* Force-detach from local disk, aborting any pending local IO */
@@ -450,6 +521,20 @@
 	B_RS_H_DONE,		/* Before resync handler done (already executed) */
 	DISCARD_MY_DATA,	/* discard_my_data flag per volume */
 	READ_BALANCE_RR,
+
+	FLUSH_PENDING,		/* if set, device->flush_jif is when we submitted that flush
+				 * from drbd_flush_after_epoch() */
+
+	/* cleared only after backing device related structures have been destroyed. */
+	GOING_DISKLESS,		/* Disk is being detached, because of io-error, or admin request. */
+
+	/* to be used in drbd_device_post_work() */
+	GO_DISKLESS,		/* tell worker to schedule cleanup before detach */
+	DESTROY_DISK,		/* tell worker to close backing devices and destroy related structures. */
+	MD_SYNC,		/* tell worker to call drbd_md_sync() */
+	RS_START,		/* tell worker to start resync/OV */
+	RS_PROGRESS,		/* tell worker that resync made significant progress */
+	RS_DONE,		/* tell worker that resync is done */
 };
 
 struct drbd_bitmap; /* opaque for drbd_device */
@@ -531,6 +616,11 @@
 };
 
 struct drbd_md_io {
+	struct page *page;
+	unsigned long start_jif;	/* last call to drbd_md_get_buffer */
+	unsigned long submit_jif;	/* last _drbd_md_sync_page_io() submit */
+	const char *current_use;
+	atomic_t in_use;
 	unsigned int done;
 	int error;
 };
@@ -577,10 +667,18 @@
 				 * and potentially deadlock on, this drbd worker.
 				 */
 	DISCONNECT_SENT,
+
+	DEVICE_WORK_PENDING,	/* tell worker that some device has pending work */
 };
 
 struct drbd_resource {
 	char *name;
+#ifdef CONFIG_DEBUG_FS
+	struct dentry *debugfs_res;
+	struct dentry *debugfs_res_volumes;
+	struct dentry *debugfs_res_connections;
+	struct dentry *debugfs_res_in_flight_summary;
+#endif
 	struct kref kref;
 	struct idr devices;		/* volume number to device mapping */
 	struct list_head connections;
@@ -594,12 +692,28 @@
 	unsigned susp_nod:1;		/* IO suspended because no data */
 	unsigned susp_fen:1;		/* IO suspended because fence peer handler runs */
 
+	enum write_ordering_e write_ordering;
+
 	cpumask_var_t cpu_mask;
 };
 
+struct drbd_thread_timing_details
+{
+	unsigned long start_jif;
+	void *cb_addr;
+	const char *caller_fn;
+	unsigned int line;
+	unsigned int cb_nr;
+};
+
 struct drbd_connection {
 	struct list_head connections;
 	struct drbd_resource *resource;
+#ifdef CONFIG_DEBUG_FS
+	struct dentry *debugfs_conn;
+	struct dentry *debugfs_conn_callback_history;
+	struct dentry *debugfs_conn_oldest_requests;
+#endif
 	struct kref kref;
 	struct idr peer_devices;	/* volume number to peer device mapping */
 	enum drbd_conns cstate;		/* Only C_STANDALONE to C_WF_REPORT_PARAMS */
@@ -636,7 +750,6 @@
 	struct drbd_epoch *current_epoch;
 	spinlock_t epoch_lock;
 	unsigned int epochs;
-	enum write_ordering_e write_ordering;
 	atomic_t current_tle_nr;	/* transfer log epoch number */
 	unsigned current_tle_writes;	/* writes seen within this tl epoch */
 
@@ -645,9 +758,22 @@
 	struct drbd_thread worker;
 	struct drbd_thread asender;
 
+	/* cached pointers,
+	 * so we can look up the oldest pending requests more quickly.
+	 * protected by resource->req_lock */
+	struct drbd_request *req_next; /* DRBD 9: todo.req_next */
+	struct drbd_request *req_ack_pending;
+	struct drbd_request *req_not_net_done;
+
 	/* sender side */
 	struct drbd_work_queue sender_work;
 
+#define DRBD_THREAD_DETAILS_HIST	16
+	unsigned int w_cb_nr; /* keeps counting up */
+	unsigned int r_cb_nr; /* keeps counting up */
+	struct drbd_thread_timing_details w_timing_details[DRBD_THREAD_DETAILS_HIST];
+	struct drbd_thread_timing_details r_timing_details[DRBD_THREAD_DETAILS_HIST];
+
 	struct {
 		/* whether this sender thread
 		 * has processed a single write yet. */
@@ -663,11 +789,22 @@
 	} send;
 };
 
+void __update_timing_details(
+		struct drbd_thread_timing_details *tdp,
+		unsigned int *cb_nr,
+		void *cb,
+		const char *fn, const unsigned int line);
+
+#define update_worker_timing_details(c, cb) \
+	__update_timing_details(c->w_timing_details, &c->w_cb_nr, cb, __func__ , __LINE__ )
+#define update_receiver_timing_details(c, cb) \
+	__update_timing_details(c->r_timing_details, &c->r_cb_nr, cb, __func__ , __LINE__ )
+
 struct submit_worker {
 	struct workqueue_struct *wq;
 	struct work_struct worker;
 
-	spinlock_t lock;
+	/* protected by ..->resource->req_lock */
 	struct list_head writes;
 };
 
@@ -675,12 +812,29 @@
 	struct list_head peer_devices;
 	struct drbd_device *device;
 	struct drbd_connection *connection;
+#ifdef CONFIG_DEBUG_FS
+	struct dentry *debugfs_peer_dev;
+#endif
 };
 
 struct drbd_device {
 	struct drbd_resource *resource;
 	struct list_head peer_devices;
-	int vnr;			/* volume number within the connection */
+	struct list_head pending_bitmap_io;
+
+	unsigned long flush_jif;
+#ifdef CONFIG_DEBUG_FS
+	struct dentry *debugfs_minor;
+	struct dentry *debugfs_vol;
+	struct dentry *debugfs_vol_oldest_requests;
+	struct dentry *debugfs_vol_act_log_extents;
+	struct dentry *debugfs_vol_resync_extents;
+	struct dentry *debugfs_vol_data_gen_id;
+#endif
+
+	unsigned int vnr;	/* volume number within the connection */
+	unsigned int minor;	/* device minor number */
+
 	struct kref kref;
 
 	/* things that are stored as / read from meta data on disk */
@@ -697,19 +851,10 @@
 	unsigned long last_reattach_jif;
 	struct drbd_work resync_work;
 	struct drbd_work unplug_work;
-	struct drbd_work go_diskless;
-	struct drbd_work md_sync_work;
-	struct drbd_work start_resync_work;
 	struct timer_list resync_timer;
 	struct timer_list md_sync_timer;
 	struct timer_list start_resync_timer;
 	struct timer_list request_timer;
-#ifdef DRBD_DEBUG_MD_SYNC
-	struct {
-		unsigned int line;
-		const char* func;
-	} last_md_mark_dirty;
-#endif
 
 	/* Used after attach while negotiating new disk state. */
 	union drbd_state new_state_tmp;
@@ -724,6 +869,7 @@
 	unsigned int al_writ_cnt;
 	unsigned int bm_writ_cnt;
 	atomic_t ap_bio_cnt;	 /* Requests we need to complete */
+	atomic_t ap_actlog_cnt;  /* Requests waiting for activity log */
 	atomic_t ap_pending_cnt; /* AP data packets on the wire, ack expected */
 	atomic_t rs_pending_cnt; /* RS request/data packets on the wire */
 	atomic_t unacked_cnt;	 /* Need to send replies for */
@@ -733,6 +879,13 @@
 	struct rb_root read_requests;
 	struct rb_root write_requests;
 
+	/* for statistics and timeouts */
+	/* [0] read, [1] write */
+	struct list_head pending_master_completion[2];
+	struct list_head pending_completion[2];
+
+	/* use checksums for *this* resync */
+	bool use_csums;
 	/* blocks to resync in this run [unit BM_BLOCK_SIZE] */
 	unsigned long rs_total;
 	/* number of resync blocks that failed in this run */
@@ -788,9 +941,7 @@
 	atomic_t pp_in_use;		/* allocated from page pool */
 	atomic_t pp_in_use_by_net;	/* sendpage()d, still referenced by tcp */
 	wait_queue_head_t ee_wait;
-	struct page *md_io_page;	/* one page buffer for md_io */
 	struct drbd_md_io md_io;
-	atomic_t md_io_in_use;		/* protects the md_io, md_io_page and md_io_tmpp */
 	spinlock_t al_lock;
 	wait_queue_head_t al_wait;
 	struct lru_cache *act_log;	/* activity log */
@@ -800,7 +951,6 @@
 	atomic_t packet_seq;
 	unsigned int peer_seq;
 	spinlock_t peer_seq_lock;
-	unsigned int minor;
 	unsigned long comm_bm_set; /* communicated number of set bits. */
 	struct bm_io_work bm_io_work;
 	u64 ed_uuid; /* UUID of the exposed data */
@@ -824,6 +974,21 @@
 	struct submit_worker submit;
 };
 
+struct drbd_bm_aio_ctx {
+	struct drbd_device *device;
+	struct list_head list; /* on device->pending_bitmap_io */;
+	unsigned long start_jif;
+	atomic_t in_flight;
+	unsigned int done;
+	unsigned flags;
+#define BM_AIO_COPY_PAGES	1
+#define BM_AIO_WRITE_HINTED	2
+#define BM_AIO_WRITE_ALL_PAGES	4
+#define BM_AIO_READ		8
+	int error;
+	struct kref kref;
+};
+
 struct drbd_config_context {
 	/* assigned from drbd_genlmsghdr */
 	unsigned int minor;
@@ -949,7 +1114,7 @@
 extern int drbd_send_bitmap(struct drbd_device *device);
 extern void drbd_send_sr_reply(struct drbd_peer_device *, enum drbd_state_rv retcode);
 extern void conn_send_sr_reply(struct drbd_connection *connection, enum drbd_state_rv retcode);
-extern void drbd_free_bc(struct drbd_backing_dev *ldev);
+extern void drbd_free_ldev(struct drbd_backing_dev *ldev);
 extern void drbd_device_cleanup(struct drbd_device *device);
 void drbd_print_uuids(struct drbd_device *device, const char *text);
 
@@ -966,13 +1131,7 @@
 extern void drbd_md_set_flag(struct drbd_device *device, int flags) __must_hold(local);
 extern void drbd_md_clear_flag(struct drbd_device *device, int flags)__must_hold(local);
 extern int drbd_md_test_flag(struct drbd_backing_dev *, int);
-#ifndef DRBD_DEBUG_MD_SYNC
 extern void drbd_md_mark_dirty(struct drbd_device *device);
-#else
-#define drbd_md_mark_dirty(m)	drbd_md_mark_dirty_(m, __LINE__ , __func__ )
-extern void drbd_md_mark_dirty_(struct drbd_device *device,
-		unsigned int line, const char *func);
-#endif
 extern void drbd_queue_bitmap_io(struct drbd_device *device,
 				 int (*io_fn)(struct drbd_device *),
 				 void (*done)(struct drbd_device *, int),
@@ -983,9 +1142,8 @@
 extern int drbd_bitmap_io_from_worker(struct drbd_device *device,
 		int (*io_fn)(struct drbd_device *),
 		char *why, enum bm_flag flags);
-extern int drbd_bmio_set_n_write(struct drbd_device *device);
-extern int drbd_bmio_clear_n_write(struct drbd_device *device);
-extern void drbd_ldev_destroy(struct drbd_device *device);
+extern int drbd_bmio_set_n_write(struct drbd_device *device) __must_hold(local);
+extern int drbd_bmio_clear_n_write(struct drbd_device *device) __must_hold(local);
 
 /* Meta data layout
  *
@@ -1105,17 +1263,21 @@
 /* in which _bitmap_ extent (resp. sector) the bit for a certain
  * _storage_ sector is located in */
 #define BM_SECT_TO_EXT(x)   ((x)>>(BM_EXT_SHIFT-9))
+#define BM_BIT_TO_EXT(x)    ((x) >> (BM_EXT_SHIFT - BM_BLOCK_SHIFT))
 
-/* how much _storage_ sectors we have per bitmap sector */
+/* first storage sector a bitmap extent corresponds to */
 #define BM_EXT_TO_SECT(x)   ((sector_t)(x) << (BM_EXT_SHIFT-9))
+/* how much _storage_ sectors we have per bitmap extent */
 #define BM_SECT_PER_EXT     BM_EXT_TO_SECT(1)
+/* how many bits are covered by one bitmap extent (resync extent) */
+#define BM_BITS_PER_EXT     (1UL << (BM_EXT_SHIFT - BM_BLOCK_SHIFT))
+
+#define BM_BLOCKS_PER_BM_EXT_MASK  (BM_BITS_PER_EXT - 1)
+
 
 /* in one sector of the bitmap, we have this many activity_log extents. */
 #define AL_EXT_PER_BM_SECT  (1 << (BM_EXT_SHIFT - AL_EXTENT_SHIFT))
 
-#define BM_BLOCKS_PER_BM_EXT_B (BM_EXT_SHIFT - BM_BLOCK_SHIFT)
-#define BM_BLOCKS_PER_BM_EXT_MASK  ((1<<BM_BLOCKS_PER_BM_EXT_B) - 1)
-
 /* the extent in "PER_EXTENT" below is an activity log extent
  * we need that many (long words/bytes) to store the bitmap
  *		     of one AL_EXTENT_SIZE chunk of storage.
@@ -1195,11 +1357,11 @@
 		const unsigned long s, const unsigned long e);
 extern int  drbd_bm_test_bit(struct drbd_device *device, unsigned long bitnr);
 extern int  drbd_bm_e_weight(struct drbd_device *device, unsigned long enr);
-extern int  drbd_bm_write_page(struct drbd_device *device, unsigned int idx) __must_hold(local);
 extern int  drbd_bm_read(struct drbd_device *device) __must_hold(local);
 extern void drbd_bm_mark_for_writeout(struct drbd_device *device, int page_nr);
 extern int  drbd_bm_write(struct drbd_device *device) __must_hold(local);
 extern int  drbd_bm_write_hinted(struct drbd_device *device) __must_hold(local);
+extern int  drbd_bm_write_lazy(struct drbd_device *device, unsigned upper_idx) __must_hold(local);
 extern int drbd_bm_write_all(struct drbd_device *device) __must_hold(local);
 extern int  drbd_bm_write_copy_pages(struct drbd_device *device) __must_hold(local);
 extern size_t	     drbd_bm_words(struct drbd_device *device);
@@ -1213,7 +1375,6 @@
 extern unsigned long _drbd_bm_find_next_zero(struct drbd_device *device, unsigned long bm_fo);
 extern unsigned long _drbd_bm_total_weight(struct drbd_device *device);
 extern unsigned long drbd_bm_total_weight(struct drbd_device *device);
-extern int drbd_bm_rs_done(struct drbd_device *device);
 /* for receive_bitmap */
 extern void drbd_bm_merge_lel(struct drbd_device *device, size_t offset,
 		size_t number, unsigned long *buffer);
@@ -1312,7 +1473,7 @@
 extern enum determine_dev_size
 drbd_determine_dev_size(struct drbd_device *, enum dds_flags, struct resize_parms *) __must_hold(local);
 extern void resync_after_online_grow(struct drbd_device *);
-extern void drbd_reconsider_max_bio_size(struct drbd_device *device);
+extern void drbd_reconsider_max_bio_size(struct drbd_device *device, struct drbd_backing_dev *bdev);
 extern enum drbd_state_rv drbd_set_role(struct drbd_device *device,
 					enum drbd_role new_role,
 					int force);
@@ -1333,7 +1494,7 @@
 extern void suspend_other_sg(struct drbd_device *device);
 extern int drbd_resync_finished(struct drbd_device *device);
 /* maybe rather drbd_main.c ? */
-extern void *drbd_md_get_buffer(struct drbd_device *device);
+extern void *drbd_md_get_buffer(struct drbd_device *device, const char *intent);
 extern void drbd_md_put_buffer(struct drbd_device *device);
 extern int drbd_md_sync_page_io(struct drbd_device *device,
 		struct drbd_backing_dev *bdev, sector_t sector, int rw);
@@ -1380,7 +1541,8 @@
 extern int drbd_receiver(struct drbd_thread *thi);
 extern int drbd_asender(struct drbd_thread *thi);
 extern bool drbd_rs_c_min_rate_throttle(struct drbd_device *device);
-extern bool drbd_rs_should_slow_down(struct drbd_device *device, sector_t sector);
+extern bool drbd_rs_should_slow_down(struct drbd_device *device, sector_t sector,
+		bool throttle_if_app_is_waiting);
 extern int drbd_submit_peer_request(struct drbd_device *,
 				    struct drbd_peer_request *, const unsigned,
 				    const int);
@@ -1464,10 +1626,7 @@
 {
 	__release(local);
 	if (!bio->bi_bdev) {
-		printk(KERN_ERR "drbd%d: drbd_generic_make_request: "
-				"bio->bi_bdev == NULL\n",
-		       device_to_minor(device));
-		dump_stack();
+		drbd_err(device, "drbd_generic_make_request: bio->bi_bdev == NULL\n");
 		bio_endio(bio, -ENODEV);
 		return;
 	}
@@ -1478,7 +1637,8 @@
 		generic_make_request(bio);
 }
 
-void drbd_bump_write_ordering(struct drbd_connection *connection, enum write_ordering_e wo);
+void drbd_bump_write_ordering(struct drbd_resource *resource, struct drbd_backing_dev *bdev,
+			      enum write_ordering_e wo);
 
 /* drbd_proc.c */
 extern struct proc_dir_entry *drbd_proc;
@@ -1489,9 +1649,9 @@
 /* drbd_actlog.c */
 extern bool drbd_al_begin_io_prepare(struct drbd_device *device, struct drbd_interval *i);
 extern int drbd_al_begin_io_nonblock(struct drbd_device *device, struct drbd_interval *i);
-extern void drbd_al_begin_io_commit(struct drbd_device *device, bool delegate);
+extern void drbd_al_begin_io_commit(struct drbd_device *device);
 extern bool drbd_al_begin_io_fastpath(struct drbd_device *device, struct drbd_interval *i);
-extern void drbd_al_begin_io(struct drbd_device *device, struct drbd_interval *i, bool delegate);
+extern void drbd_al_begin_io(struct drbd_device *device, struct drbd_interval *i);
 extern void drbd_al_complete_io(struct drbd_device *device, struct drbd_interval *i);
 extern void drbd_rs_complete_io(struct drbd_device *device, sector_t sector);
 extern int drbd_rs_begin_io(struct drbd_device *device, sector_t sector);
@@ -1501,14 +1661,17 @@
 extern void drbd_rs_failed_io(struct drbd_device *device,
 		sector_t sector, int size);
 extern void drbd_advance_rs_marks(struct drbd_device *device, unsigned long still_to_go);
-extern void __drbd_set_in_sync(struct drbd_device *device, sector_t sector,
-		int size, const char *file, const unsigned int line);
+
+enum update_sync_bits_mode { RECORD_RS_FAILED, SET_OUT_OF_SYNC, SET_IN_SYNC };
+extern int __drbd_change_sync(struct drbd_device *device, sector_t sector, int size,
+		enum update_sync_bits_mode mode,
+		const char *file, const unsigned int line);
 #define drbd_set_in_sync(device, sector, size) \
-	__drbd_set_in_sync(device, sector, size, __FILE__, __LINE__)
-extern int __drbd_set_out_of_sync(struct drbd_device *device, sector_t sector,
-		int size, const char *file, const unsigned int line);
+	__drbd_change_sync(device, sector, size, SET_IN_SYNC, __FILE__, __LINE__)
 #define drbd_set_out_of_sync(device, sector, size) \
-	__drbd_set_out_of_sync(device, sector, size, __FILE__, __LINE__)
+	__drbd_change_sync(device, sector, size, SET_OUT_OF_SYNC, __FILE__, __LINE__)
+#define drbd_rs_failed_io(device, sector, size) \
+	__drbd_change_sync(device, sector, size, RECORD_RS_FAILED, __FILE__, __LINE__)
 extern void drbd_al_shrink(struct drbd_device *device);
 extern int drbd_initialize_al(struct drbd_device *, void *);
 
@@ -1764,16 +1927,6 @@
 }
 
 static inline void
-drbd_queue_work_front(struct drbd_work_queue *q, struct drbd_work *w)
-{
-	unsigned long flags;
-	spin_lock_irqsave(&q->q_lock, flags);
-	list_add(&w->list, &q->q);
-	spin_unlock_irqrestore(&q->q_lock, flags);
-	wake_up(&q->q_wait);
-}
-
-static inline void
 drbd_queue_work(struct drbd_work_queue *q, struct drbd_work *w)
 {
 	unsigned long flags;
@@ -1783,6 +1936,29 @@
 	wake_up(&q->q_wait);
 }
 
+static inline void
+drbd_queue_work_if_unqueued(struct drbd_work_queue *q, struct drbd_work *w)
+{
+	unsigned long flags;
+	spin_lock_irqsave(&q->q_lock, flags);
+	if (list_empty_careful(&w->list))
+		list_add_tail(&w->list, &q->q);
+	spin_unlock_irqrestore(&q->q_lock, flags);
+	wake_up(&q->q_wait);
+}
+
+static inline void
+drbd_device_post_work(struct drbd_device *device, int work_bit)
+{
+	if (!test_and_set_bit(work_bit, &device->flags)) {
+		struct drbd_connection *connection =
+			first_peer_device(device)->connection;
+		struct drbd_work_queue *q = &connection->sender_work;
+		if (!test_and_set_bit(DEVICE_WORK_PENDING, &connection->flags))
+			wake_up(&q->q_wait);
+	}
+}
+
 extern void drbd_flush_workqueue(struct drbd_work_queue *work_queue);
 
 static inline void wake_asender(struct drbd_connection *connection)
@@ -1859,7 +2035,7 @@
 			func, line,					\
 			atomic_read(&device->which))
 
-#define dec_ap_pending(device) _dec_ap_pending(device, __FUNCTION__, __LINE__)
+#define dec_ap_pending(device) _dec_ap_pending(device, __func__, __LINE__)
 static inline void _dec_ap_pending(struct drbd_device *device, const char *func, int line)
 {
 	if (atomic_dec_and_test(&device->ap_pending_cnt))
@@ -1878,7 +2054,7 @@
 	atomic_inc(&device->rs_pending_cnt);
 }
 
-#define dec_rs_pending(device) _dec_rs_pending(device, __FUNCTION__, __LINE__)
+#define dec_rs_pending(device) _dec_rs_pending(device, __func__, __LINE__)
 static inline void _dec_rs_pending(struct drbd_device *device, const char *func, int line)
 {
 	atomic_dec(&device->rs_pending_cnt);
@@ -1899,20 +2075,29 @@
 	atomic_inc(&device->unacked_cnt);
 }
 
-#define dec_unacked(device) _dec_unacked(device, __FUNCTION__, __LINE__)
+#define dec_unacked(device) _dec_unacked(device, __func__, __LINE__)
 static inline void _dec_unacked(struct drbd_device *device, const char *func, int line)
 {
 	atomic_dec(&device->unacked_cnt);
 	ERR_IF_CNT_IS_NEGATIVE(unacked_cnt, func, line);
 }
 
-#define sub_unacked(device, n) _sub_unacked(device, n, __FUNCTION__, __LINE__)
+#define sub_unacked(device, n) _sub_unacked(device, n, __func__, __LINE__)
 static inline void _sub_unacked(struct drbd_device *device, int n, const char *func, int line)
 {
 	atomic_sub(n, &device->unacked_cnt);
 	ERR_IF_CNT_IS_NEGATIVE(unacked_cnt, func, line);
 }
 
+static inline bool is_sync_state(enum drbd_conns connection_state)
+{
+	return
+	   (connection_state == C_SYNC_SOURCE
+	||  connection_state == C_SYNC_TARGET
+	||  connection_state == C_PAUSED_SYNC_S
+	||  connection_state == C_PAUSED_SYNC_T);
+}
+
 /**
  * get_ldev() - Increase the ref count on device->ldev. Returns 0 if there is no ldev
  * @M:		DRBD device.
@@ -1924,6 +2109,11 @@
 
 static inline void put_ldev(struct drbd_device *device)
 {
+	enum drbd_disk_state ds = device->state.disk;
+	/* We must check the state *before* the atomic_dec becomes visible,
+	 * or we have a theoretical race where someone hitting zero,
+	 * while state still D_FAILED, will then see D_DISKLESS in the
+	 * condition below and calling into destroy, where he must not, yet. */
 	int i = atomic_dec_return(&device->local_cnt);
 
 	/* This may be called from some endio handler,
@@ -1932,15 +2122,13 @@
 	__release(local);
 	D_ASSERT(device, i >= 0);
 	if (i == 0) {
-		if (device->state.disk == D_DISKLESS)
+		if (ds == D_DISKLESS)
 			/* even internal references gone, safe to destroy */
-			drbd_ldev_destroy(device);
-		if (device->state.disk == D_FAILED) {
+			drbd_device_post_work(device, DESTROY_DISK);
+		if (ds == D_FAILED)
 			/* all application IO references gone. */
-			if (!test_and_set_bit(GO_DISKLESS, &device->flags))
-				drbd_queue_work(&first_peer_device(device)->connection->sender_work,
-						&device->go_diskless);
-		}
+			if (!test_and_set_bit(GOING_DISKLESS, &device->flags))
+				drbd_device_post_work(device, GO_DISKLESS);
 		wake_up(&device->misc_wait);
 	}
 }
@@ -1964,54 +2152,6 @@
 extern int _get_ldev_if_state(struct drbd_device *device, enum drbd_disk_state mins);
 #endif
 
-/* you must have an "get_ldev" reference */
-static inline void drbd_get_syncer_progress(struct drbd_device *device,
-		unsigned long *bits_left, unsigned int *per_mil_done)
-{
-	/* this is to break it at compile time when we change that, in case we
-	 * want to support more than (1<<32) bits on a 32bit arch. */
-	typecheck(unsigned long, device->rs_total);
-
-	/* note: both rs_total and rs_left are in bits, i.e. in
-	 * units of BM_BLOCK_SIZE.
-	 * for the percentage, we don't care. */
-
-	if (device->state.conn == C_VERIFY_S || device->state.conn == C_VERIFY_T)
-		*bits_left = device->ov_left;
-	else
-		*bits_left = drbd_bm_total_weight(device) - device->rs_failed;
-	/* >> 10 to prevent overflow,
-	 * +1 to prevent division by zero */
-	if (*bits_left > device->rs_total) {
-		/* doh. maybe a logic bug somewhere.
-		 * may also be just a race condition
-		 * between this and a disconnect during sync.
-		 * for now, just prevent in-kernel buffer overflow.
-		 */
-		smp_rmb();
-		drbd_warn(device, "cs:%s rs_left=%lu > rs_total=%lu (rs_failed %lu)\n",
-				drbd_conn_str(device->state.conn),
-				*bits_left, device->rs_total, device->rs_failed);
-		*per_mil_done = 0;
-	} else {
-		/* Make sure the division happens in long context.
-		 * We allow up to one petabyte storage right now,
-		 * at a granularity of 4k per bit that is 2**38 bits.
-		 * After shift right and multiplication by 1000,
-		 * this should still fit easily into a 32bit long,
-		 * so we don't need a 64bit division on 32bit arch.
-		 * Note: currently we don't support such large bitmaps on 32bit
-		 * arch anyways, but no harm done to be prepared for it here.
-		 */
-		unsigned int shift = device->rs_total > UINT_MAX ? 16 : 10;
-		unsigned long left = *bits_left >> shift;
-		unsigned long total = 1UL + (device->rs_total >> shift);
-		unsigned long tmp = 1000UL - left * 1000UL/total;
-		*per_mil_done = tmp;
-	}
-}
-
-
 /* this throttles on-the-fly application requests
  * according to max_buffers settings;
  * maybe re-implement using semaphores? */
@@ -2201,25 +2341,6 @@
 	return QUEUE_ORDERED_NONE;
 }
 
-static inline void drbd_md_flush(struct drbd_device *device)
-{
-	int r;
-
-	if (device->ldev == NULL) {
-		drbd_warn(device, "device->ldev == NULL in drbd_md_flush\n");
-		return;
-	}
-
-	if (test_bit(MD_NO_FUA, &device->flags))
-		return;
-
-	r = blkdev_issue_flush(device->ldev->md_bdev, GFP_NOIO, NULL);
-	if (r) {
-		set_bit(MD_NO_FUA, &device->flags);
-		drbd_err(device, "meta data flush failed with status %d, disabling md-flushes\n", r);
-	}
-}
-
 static inline struct drbd_connection *first_connection(struct drbd_resource *resource)
 {
 	return list_first_entry_or_null(&resource->connections,
diff --git a/drivers/block/drbd/drbd_interval.h b/drivers/block/drbd/drbd_interval.h
index f38fcb0..f210543 100644
--- a/drivers/block/drbd/drbd_interval.h
+++ b/drivers/block/drbd/drbd_interval.h
@@ -10,7 +10,9 @@
 	unsigned int size;	/* size in bytes */
 	sector_t end;		/* highest interval end in subtree */
 	int local:1		/* local or remote request? */;
-	int waiting:1;
+	int waiting:1;		/* someone is waiting for this to complete */
+	int completed:1;	/* this has been completed already;
+				 * ignore for conflict detection */
 };
 
 static inline void drbd_clear_interval(struct drbd_interval *i)
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 960645c..9b465bb 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -26,7 +26,10 @@
 
  */
 
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
 #include <linux/module.h>
+#include <linux/jiffies.h>
 #include <linux/drbd.h>
 #include <asm/uaccess.h>
 #include <asm/types.h>
@@ -54,16 +57,14 @@
 #include "drbd_int.h"
 #include "drbd_protocol.h"
 #include "drbd_req.h" /* only for _req_mod in tl_release and tl_clear */
-
 #include "drbd_vli.h"
+#include "drbd_debugfs.h"
 
 static DEFINE_MUTEX(drbd_main_mutex);
 static int drbd_open(struct block_device *bdev, fmode_t mode);
 static void drbd_release(struct gendisk *gd, fmode_t mode);
-static int w_md_sync(struct drbd_work *w, int unused);
 static void md_sync_timer_fn(unsigned long data);
 static int w_bitmap_io(struct drbd_work *w, int unused);
-static int w_go_diskless(struct drbd_work *w, int unused);
 
 MODULE_AUTHOR("Philipp Reisner <phil@linbit.com>, "
 	      "Lars Ellenberg <lars@linbit.com>");
@@ -264,7 +265,7 @@
 
 /**
  * _tl_restart() - Walks the transfer log, and applies an action to all requests
- * @device:	DRBD device.
+ * @connection:	DRBD connection to operate on.
  * @what:       The action/event to perform with all request objects
  *
  * @what might be one of CONNECTION_LOST_WHILE_PENDING, RESEND, FAIL_FROZEN_DISK_IO,
@@ -662,6 +663,11 @@
 			    msg_flags);
 	if (data && !err)
 		err = drbd_send_all(connection, sock->socket, data, size, 0);
+	/* DRBD protocol "pings" are latency critical.
+	 * This is supposed to trigger tcp_push_pending_frames() */
+	if (!err && (cmd == P_PING || cmd == P_PING_ACK))
+		drbd_tcp_nodelay(sock->socket);
+
 	return err;
 }
 
@@ -1636,7 +1642,10 @@
 	if (peer_device->connection->agreed_pro_version >= 100) {
 		if (req->rq_state & RQ_EXP_RECEIVE_ACK)
 			dp_flags |= DP_SEND_RECEIVE_ACK;
-		if (req->rq_state & RQ_EXP_WRITE_ACK)
+		/* During resync, request an explicit write ack,
+		 * even in protocol != C */
+		if (req->rq_state & RQ_EXP_WRITE_ACK
+		|| (dp_flags & DP_MAY_SET_IN_SYNC))
 			dp_flags |= DP_SEND_WRITE_ACK;
 	}
 	p->dp_flags = cpu_to_be32(dp_flags);
@@ -1900,6 +1909,7 @@
 	drbd_set_defaults(device);
 
 	atomic_set(&device->ap_bio_cnt, 0);
+	atomic_set(&device->ap_actlog_cnt, 0);
 	atomic_set(&device->ap_pending_cnt, 0);
 	atomic_set(&device->rs_pending_cnt, 0);
 	atomic_set(&device->unacked_cnt, 0);
@@ -1908,7 +1918,7 @@
 	atomic_set(&device->rs_sect_in, 0);
 	atomic_set(&device->rs_sect_ev, 0);
 	atomic_set(&device->ap_in_flight, 0);
-	atomic_set(&device->md_io_in_use, 0);
+	atomic_set(&device->md_io.in_use, 0);
 
 	mutex_init(&device->own_state_mutex);
 	device->state_mutex = &device->own_state_mutex;
@@ -1924,17 +1934,15 @@
 	INIT_LIST_HEAD(&device->resync_reads);
 	INIT_LIST_HEAD(&device->resync_work.list);
 	INIT_LIST_HEAD(&device->unplug_work.list);
-	INIT_LIST_HEAD(&device->go_diskless.list);
-	INIT_LIST_HEAD(&device->md_sync_work.list);
-	INIT_LIST_HEAD(&device->start_resync_work.list);
 	INIT_LIST_HEAD(&device->bm_io_work.w.list);
+	INIT_LIST_HEAD(&device->pending_master_completion[0]);
+	INIT_LIST_HEAD(&device->pending_master_completion[1]);
+	INIT_LIST_HEAD(&device->pending_completion[0]);
+	INIT_LIST_HEAD(&device->pending_completion[1]);
 
 	device->resync_work.cb  = w_resync_timer;
 	device->unplug_work.cb  = w_send_write_hint;
-	device->go_diskless.cb  = w_go_diskless;
-	device->md_sync_work.cb = w_md_sync;
 	device->bm_io_work.w.cb = w_bitmap_io;
-	device->start_resync_work.cb = w_start_resync;
 
 	init_timer(&device->resync_timer);
 	init_timer(&device->md_sync_timer);
@@ -1992,7 +2000,7 @@
 		drbd_bm_cleanup(device);
 	}
 
-	drbd_free_bc(device->ldev);
+	drbd_free_ldev(device->ldev);
 	device->ldev = NULL;
 
 	clear_bit(AL_SUSPENDED, &device->flags);
@@ -2006,7 +2014,6 @@
 	D_ASSERT(device, list_empty(&first_peer_device(device)->connection->sender_work.q));
 	D_ASSERT(device, list_empty(&device->resync_work.list));
 	D_ASSERT(device, list_empty(&device->unplug_work.list));
-	D_ASSERT(device, list_empty(&device->go_diskless.list));
 
 	drbd_set_defaults(device);
 }
@@ -2129,20 +2136,6 @@
 	return -ENOMEM;
 }
 
-static int drbd_notify_sys(struct notifier_block *this, unsigned long code,
-	void *unused)
-{
-	/* just so we have it.  you never know what interesting things we
-	 * might want to do here some day...
-	 */
-
-	return NOTIFY_DONE;
-}
-
-static struct notifier_block drbd_notifier = {
-	.notifier_call = drbd_notify_sys,
-};
-
 static void drbd_release_all_peer_reqs(struct drbd_device *device)
 {
 	int rr;
@@ -2173,7 +2166,7 @@
 {
 	struct drbd_device *device = container_of(kref, struct drbd_device, kref);
 	struct drbd_resource *resource = device->resource;
-	struct drbd_connection *connection;
+	struct drbd_peer_device *peer_device, *tmp_peer_device;
 
 	del_timer_sync(&device->request_timer);
 
@@ -2187,7 +2180,7 @@
 	if (device->this_bdev)
 		bdput(device->this_bdev);
 
-	drbd_free_bc(device->ldev);
+	drbd_free_ldev(device->ldev);
 	device->ldev = NULL;
 
 	drbd_release_all_peer_reqs(device);
@@ -2200,15 +2193,20 @@
 
 	if (device->bitmap) /* should no longer be there. */
 		drbd_bm_cleanup(device);
-	__free_page(device->md_io_page);
+	__free_page(device->md_io.page);
 	put_disk(device->vdisk);
 	blk_cleanup_queue(device->rq_queue);
 	kfree(device->rs_plan_s);
-	kfree(first_peer_device(device));
-	kfree(device);
 
-	for_each_connection(connection, resource)
-		kref_put(&connection->kref, drbd_destroy_connection);
+	/* not for_each_connection(connection, resource):
+	 * those may have been cleaned up and disassociated already.
+	 */
+	for_each_peer_device_safe(peer_device, tmp_peer_device, device) {
+		kref_put(&peer_device->connection->kref, drbd_destroy_connection);
+		kfree(peer_device);
+	}
+	memset(device, 0xfd, sizeof(*device));
+	kfree(device);
 	kref_put(&resource->kref, drbd_destroy_resource);
 }
 
@@ -2236,7 +2234,7 @@
 	list_for_each_entry_safe(req, tmp, &writes, tl_requests) {
 		struct drbd_device *device = req->device;
 		struct bio *bio = req->master_bio;
-		unsigned long start_time = req->start_time;
+		unsigned long start_jif = req->start_jif;
 		bool expected;
 
 		expected =
@@ -2271,10 +2269,12 @@
 		/* We are not just doing generic_make_request(),
 		 * as we want to keep the start_time information. */
 		inc_ap_bio(device);
-		__drbd_make_request(device, bio, start_time);
+		__drbd_make_request(device, bio, start_jif);
 	}
 }
 
+/* called via drbd_req_put_completion_ref(),
+ * holds resource->req_lock */
 void drbd_restart_request(struct drbd_request *req)
 {
 	unsigned long flags;
@@ -2298,6 +2298,7 @@
 	idr_destroy(&resource->devices);
 	free_cpumask_var(resource->cpu_mask);
 	kfree(resource->name);
+	memset(resource, 0xf2, sizeof(*resource));
 	kfree(resource);
 }
 
@@ -2307,8 +2308,10 @@
 
 	for_each_connection_safe(connection, tmp, resource) {
 		list_del(&connection->connections);
+		drbd_debugfs_connection_cleanup(connection);
 		kref_put(&connection->kref, drbd_destroy_connection);
 	}
+	drbd_debugfs_resource_cleanup(resource);
 	kref_put(&resource->kref, drbd_destroy_resource);
 }
 
@@ -2318,8 +2321,6 @@
 	struct drbd_device *device;
 	struct drbd_resource *resource, *tmp;
 
-	unregister_reboot_notifier(&drbd_notifier);
-
 	/* first remove proc,
 	 * drbdsetup uses it's presence to detect
 	 * whether DRBD is loaded.
@@ -2335,6 +2336,7 @@
 		destroy_workqueue(retry.wq);
 
 	drbd_genl_unregister();
+	drbd_debugfs_cleanup();
 
 	idr_for_each_entry(&drbd_devices, device, i)
 		drbd_delete_device(device);
@@ -2350,7 +2352,7 @@
 
 	idr_destroy(&drbd_devices);
 
-	printk(KERN_INFO "drbd: module cleanup done.\n");
+	pr_info("module cleanup done.\n");
 }
 
 /**
@@ -2539,6 +2541,20 @@
 	if (nr_cpu_ids > 1 && res_opts->cpu_mask[0] != 0) {
 		err = bitmap_parse(res_opts->cpu_mask, DRBD_CPU_MASK_SIZE,
 				   cpumask_bits(new_cpu_mask), nr_cpu_ids);
+		if (err == -EOVERFLOW) {
+			/* So what. mask it out. */
+			cpumask_var_t tmp_cpu_mask;
+			if (zalloc_cpumask_var(&tmp_cpu_mask, GFP_KERNEL)) {
+				cpumask_setall(tmp_cpu_mask);
+				cpumask_and(new_cpu_mask, new_cpu_mask, tmp_cpu_mask);
+				drbd_warn(resource, "Overflow in bitmap_parse(%.12s%s), truncating to %u bits\n",
+					res_opts->cpu_mask,
+					strlen(res_opts->cpu_mask) > 12 ? "..." : "",
+					nr_cpu_ids);
+				free_cpumask_var(tmp_cpu_mask);
+				err = 0;
+			}
+		}
 		if (err) {
 			drbd_warn(resource, "bitmap_parse() failed with %d\n", err);
 			/* retcode = ERR_CPU_MASK_PARSE; */
@@ -2579,10 +2595,12 @@
 	kref_init(&resource->kref);
 	idr_init(&resource->devices);
 	INIT_LIST_HEAD(&resource->connections);
+	resource->write_ordering = WO_bdev_flush;
 	list_add_tail_rcu(&resource->resources, &drbd_resources);
 	mutex_init(&resource->conf_update);
 	mutex_init(&resource->adm_mutex);
 	spin_lock_init(&resource->req_lock);
+	drbd_debugfs_resource_add(resource);
 	return resource;
 
 fail_free_name:
@@ -2593,7 +2611,7 @@
 	return NULL;
 }
 
-/* caller must be under genl_lock() */
+/* caller must be under adm_mutex */
 struct drbd_connection *conn_create(const char *name, struct res_opts *res_opts)
 {
 	struct drbd_resource *resource;
@@ -2617,7 +2635,6 @@
 	INIT_LIST_HEAD(&connection->current_epoch->list);
 	connection->epochs = 1;
 	spin_lock_init(&connection->epoch_lock);
-	connection->write_ordering = WO_bdev_flush;
 
 	connection->send.seen_any_write_yet = false;
 	connection->send.current_epoch_nr = 0;
@@ -2652,6 +2669,7 @@
 
 	kref_get(&resource->kref);
 	list_add_tail_rcu(&connection->connections, &resource->connections);
+	drbd_debugfs_connection_add(connection);
 	return connection;
 
 fail_resource:
@@ -2680,6 +2698,7 @@
 	drbd_free_socket(&connection->data);
 	kfree(connection->int_dig_in);
 	kfree(connection->int_dig_vv);
+	memset(connection, 0xfc, sizeof(*connection));
 	kfree(connection);
 	kref_put(&resource->kref, drbd_destroy_resource);
 }
@@ -2694,7 +2713,6 @@
 		return -ENOMEM;
 
 	INIT_WORK(&device->submit.worker, do_submit);
-	spin_lock_init(&device->submit.lock);
 	INIT_LIST_HEAD(&device->submit.writes);
 	return 0;
 }
@@ -2764,8 +2782,8 @@
 	blk_queue_merge_bvec(q, drbd_merge_bvec);
 	q->queue_lock = &resource->req_lock;
 
-	device->md_io_page = alloc_page(GFP_KERNEL);
-	if (!device->md_io_page)
+	device->md_io.page = alloc_page(GFP_KERNEL);
+	if (!device->md_io.page)
 		goto out_no_io_page;
 
 	if (drbd_bm_init(device))
@@ -2794,6 +2812,7 @@
 	kref_get(&device->kref);
 
 	INIT_LIST_HEAD(&device->peer_devices);
+	INIT_LIST_HEAD(&device->pending_bitmap_io);
 	for_each_connection(connection, resource) {
 		peer_device = kzalloc(sizeof(struct drbd_peer_device), GFP_KERNEL);
 		if (!peer_device)
@@ -2829,7 +2848,10 @@
 		for_each_peer_device(peer_device, device)
 			drbd_connected(peer_device);
 	}
-
+	/* move to create_peer_device() */
+	for_each_peer_device(peer_device, device)
+		drbd_debugfs_peer_device_add(peer_device);
+	drbd_debugfs_device_add(device);
 	return NO_ERROR;
 
 out_idr_remove_vol:
@@ -2853,7 +2875,7 @@
 out_no_minor_idr:
 	drbd_bm_cleanup(device);
 out_no_bitmap:
-	__free_page(device->md_io_page);
+	__free_page(device->md_io.page);
 out_no_io_page:
 	put_disk(disk);
 out_no_disk:
@@ -2868,8 +2890,13 @@
 {
 	struct drbd_resource *resource = device->resource;
 	struct drbd_connection *connection;
+	struct drbd_peer_device *peer_device;
 	int refs = 3;
 
+	/* move to free_peer_device() */
+	for_each_peer_device(peer_device, device)
+		drbd_debugfs_peer_device_cleanup(peer_device);
+	drbd_debugfs_device_cleanup(device);
 	for_each_connection(connection, resource) {
 		idr_remove(&connection->peer_devices, device->vnr);
 		refs++;
@@ -2881,13 +2908,12 @@
 	kref_sub(&device->kref, refs, drbd_destroy_device);
 }
 
-int __init drbd_init(void)
+static int __init drbd_init(void)
 {
 	int err;
 
 	if (minor_count < DRBD_MINOR_COUNT_MIN || minor_count > DRBD_MINOR_COUNT_MAX) {
-		printk(KERN_ERR
-		       "drbd: invalid minor_count (%d)\n", minor_count);
+		pr_err("invalid minor_count (%d)\n", minor_count);
 #ifdef MODULE
 		return -EINVAL;
 #else
@@ -2897,14 +2923,11 @@
 
 	err = register_blkdev(DRBD_MAJOR, "drbd");
 	if (err) {
-		printk(KERN_ERR
-		       "drbd: unable to register block device major %d\n",
+		pr_err("unable to register block device major %d\n",
 		       DRBD_MAJOR);
 		return err;
 	}
 
-	register_reboot_notifier(&drbd_notifier);
-
 	/*
 	 * allocate all necessary structs
 	 */
@@ -2918,7 +2941,7 @@
 
 	err = drbd_genl_register();
 	if (err) {
-		printk(KERN_ERR "drbd: unable to register generic netlink family\n");
+		pr_err("unable to register generic netlink family\n");
 		goto fail;
 	}
 
@@ -2929,38 +2952,39 @@
 	err = -ENOMEM;
 	drbd_proc = proc_create_data("drbd", S_IFREG | S_IRUGO , NULL, &drbd_proc_fops, NULL);
 	if (!drbd_proc)	{
-		printk(KERN_ERR "drbd: unable to register proc file\n");
+		pr_err("unable to register proc file\n");
 		goto fail;
 	}
 
 	retry.wq = create_singlethread_workqueue("drbd-reissue");
 	if (!retry.wq) {
-		printk(KERN_ERR "drbd: unable to create retry workqueue\n");
+		pr_err("unable to create retry workqueue\n");
 		goto fail;
 	}
 	INIT_WORK(&retry.worker, do_retry);
 	spin_lock_init(&retry.lock);
 	INIT_LIST_HEAD(&retry.writes);
 
-	printk(KERN_INFO "drbd: initialized. "
+	if (drbd_debugfs_init())
+		pr_notice("failed to initialize debugfs -- will not be available\n");
+
+	pr_info("initialized. "
 	       "Version: " REL_VERSION " (api:%d/proto:%d-%d)\n",
 	       API_VERSION, PRO_VERSION_MIN, PRO_VERSION_MAX);
-	printk(KERN_INFO "drbd: %s\n", drbd_buildtag());
-	printk(KERN_INFO "drbd: registered as block device major %d\n",
-		DRBD_MAJOR);
-
+	pr_info("%s\n", drbd_buildtag());
+	pr_info("registered as block device major %d\n", DRBD_MAJOR);
 	return 0; /* Success! */
 
 fail:
 	drbd_cleanup();
 	if (err == -ENOMEM)
-		printk(KERN_ERR "drbd: ran out of memory\n");
+		pr_err("ran out of memory\n");
 	else
-		printk(KERN_ERR "drbd: initialization failure\n");
+		pr_err("initialization failure\n");
 	return err;
 }
 
-void drbd_free_bc(struct drbd_backing_dev *ldev)
+void drbd_free_ldev(struct drbd_backing_dev *ldev)
 {
 	if (ldev == NULL)
 		return;
@@ -2972,22 +2996,27 @@
 	kfree(ldev);
 }
 
+static void drbd_free_one_sock(struct drbd_socket *ds)
+{
+	struct socket *s;
+	mutex_lock(&ds->mutex);
+	s = ds->socket;
+	ds->socket = NULL;
+	mutex_unlock(&ds->mutex);
+	if (s) {
+		/* so debugfs does not need to mutex_lock() */
+		synchronize_rcu();
+		kernel_sock_shutdown(s, SHUT_RDWR);
+		sock_release(s);
+	}
+}
+
 void drbd_free_sock(struct drbd_connection *connection)
 {
-	if (connection->data.socket) {
-		mutex_lock(&connection->data.mutex);
-		kernel_sock_shutdown(connection->data.socket, SHUT_RDWR);
-		sock_release(connection->data.socket);
-		connection->data.socket = NULL;
-		mutex_unlock(&connection->data.mutex);
-	}
-	if (connection->meta.socket) {
-		mutex_lock(&connection->meta.mutex);
-		kernel_sock_shutdown(connection->meta.socket, SHUT_RDWR);
-		sock_release(connection->meta.socket);
-		connection->meta.socket = NULL;
-		mutex_unlock(&connection->meta.mutex);
-	}
+	if (connection->data.socket)
+		drbd_free_one_sock(&connection->data);
+	if (connection->meta.socket)
+		drbd_free_one_sock(&connection->meta);
 }
 
 /* meta data management */
@@ -3093,7 +3122,7 @@
 	if (!get_ldev_if_state(device, D_FAILED))
 		return;
 
-	buffer = drbd_md_get_buffer(device);
+	buffer = drbd_md_get_buffer(device, __func__);
 	if (!buffer)
 		goto out;
 
@@ -3253,7 +3282,7 @@
 	if (device->state.disk != D_DISKLESS)
 		return ERR_DISK_CONFIGURED;
 
-	buffer = drbd_md_get_buffer(device);
+	buffer = drbd_md_get_buffer(device, __func__);
 	if (!buffer)
 		return ERR_NOMEM;
 
@@ -3466,23 +3495,19 @@
  *
  * Sets all bits in the bitmap and writes the whole bitmap to stable storage.
  */
-int drbd_bmio_set_n_write(struct drbd_device *device)
+int drbd_bmio_set_n_write(struct drbd_device *device) __must_hold(local)
 {
 	int rv = -EIO;
 
-	if (get_ldev_if_state(device, D_ATTACHING)) {
-		drbd_md_set_flag(device, MDF_FULL_SYNC);
+	drbd_md_set_flag(device, MDF_FULL_SYNC);
+	drbd_md_sync(device);
+	drbd_bm_set_all(device);
+
+	rv = drbd_bm_write(device);
+
+	if (!rv) {
+		drbd_md_clear_flag(device, MDF_FULL_SYNC);
 		drbd_md_sync(device);
-		drbd_bm_set_all(device);
-
-		rv = drbd_bm_write(device);
-
-		if (!rv) {
-			drbd_md_clear_flag(device, MDF_FULL_SYNC);
-			drbd_md_sync(device);
-		}
-
-		put_ldev(device);
 	}
 
 	return rv;
@@ -3494,18 +3519,11 @@
  *
  * Clears all bits in the bitmap and writes the whole bitmap to stable storage.
  */
-int drbd_bmio_clear_n_write(struct drbd_device *device)
+int drbd_bmio_clear_n_write(struct drbd_device *device) __must_hold(local)
 {
-	int rv = -EIO;
-
 	drbd_resume_al(device);
-	if (get_ldev_if_state(device, D_ATTACHING)) {
-		drbd_bm_clear_all(device);
-		rv = drbd_bm_write(device);
-		put_ldev(device);
-	}
-
-	return rv;
+	drbd_bm_clear_all(device);
+	return drbd_bm_write(device);
 }
 
 static int w_bitmap_io(struct drbd_work *w, int unused)
@@ -3537,61 +3555,6 @@
 	return 0;
 }
 
-void drbd_ldev_destroy(struct drbd_device *device)
-{
-	lc_destroy(device->resync);
-	device->resync = NULL;
-	lc_destroy(device->act_log);
-	device->act_log = NULL;
-	__no_warn(local,
-		drbd_free_bc(device->ldev);
-		device->ldev = NULL;);
-
-	clear_bit(GO_DISKLESS, &device->flags);
-}
-
-static int w_go_diskless(struct drbd_work *w, int unused)
-{
-	struct drbd_device *device =
-		container_of(w, struct drbd_device, go_diskless);
-
-	D_ASSERT(device, device->state.disk == D_FAILED);
-	/* we cannot assert local_cnt == 0 here, as get_ldev_if_state will
-	 * inc/dec it frequently. Once we are D_DISKLESS, no one will touch
-	 * the protected members anymore, though, so once put_ldev reaches zero
-	 * again, it will be safe to free them. */
-
-	/* Try to write changed bitmap pages, read errors may have just
-	 * set some bits outside the area covered by the activity log.
-	 *
-	 * If we have an IO error during the bitmap writeout,
-	 * we will want a full sync next time, just in case.
-	 * (Do we want a specific meta data flag for this?)
-	 *
-	 * If that does not make it to stable storage either,
-	 * we cannot do anything about that anymore.
-	 *
-	 * We still need to check if both bitmap and ldev are present, we may
-	 * end up here after a failed attach, before ldev was even assigned.
-	 */
-	if (device->bitmap && device->ldev) {
-		/* An interrupted resync or similar is allowed to recounts bits
-		 * while we detach.
-		 * Any modifications would not be expected anymore, though.
-		 */
-		if (drbd_bitmap_io_from_worker(device, drbd_bm_write,
-					"detach", BM_LOCKED_TEST_ALLOWED)) {
-			if (test_bit(WAS_READ_ERROR, &device->flags)) {
-				drbd_md_set_flag(device, MDF_FULL_SYNC);
-				drbd_md_sync(device);
-			}
-		}
-	}
-
-	drbd_force_state(device, NS(disk, D_DISKLESS));
-	return 0;
-}
-
 /**
  * drbd_queue_bitmap_io() - Queues an IO operation on the whole bitmap
  * @device:	DRBD device.
@@ -3603,6 +3566,9 @@
  * that drbd_set_out_of_sync() can not be called. This function MAY ONLY be
  * called from worker context. It MUST NOT be used while a previous such
  * work is still pending!
+ *
+ * Its worker function encloses the call of io_fn() by get_ldev() and
+ * put_ldev().
  */
 void drbd_queue_bitmap_io(struct drbd_device *device,
 			  int (*io_fn)(struct drbd_device *),
@@ -3685,25 +3651,7 @@
 static void md_sync_timer_fn(unsigned long data)
 {
 	struct drbd_device *device = (struct drbd_device *) data;
-
-	/* must not double-queue! */
-	if (list_empty(&device->md_sync_work.list))
-		drbd_queue_work_front(&first_peer_device(device)->connection->sender_work,
-				      &device->md_sync_work);
-}
-
-static int w_md_sync(struct drbd_work *w, int unused)
-{
-	struct drbd_device *device =
-		container_of(w, struct drbd_device, md_sync_work);
-
-	drbd_warn(device, "md_sync_timer expired! Worker calls drbd_md_sync().\n");
-#ifdef DEBUG
-	drbd_warn(device, "last md_mark_dirty: %s:%u\n",
-		device->last_md_mark_dirty.func, device->last_md_mark_dirty.line);
-#endif
-	drbd_md_sync(device);
-	return 0;
+	drbd_device_post_work(device, MD_SYNC);
 }
 
 const char *cmdname(enum drbd_packet cmd)
diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index 3f2e167..1cd47df 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -23,6 +23,8 @@
 
  */
 
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
 #include <linux/module.h>
 #include <linux/drbd.h>
 #include <linux/in.h>
@@ -85,7 +87,7 @@
 {
 	genlmsg_end(skb, genlmsg_data(nlmsg_data(nlmsg_hdr(skb))));
 	if (genlmsg_reply(skb, info))
-		printk(KERN_ERR "drbd: error sending genl reply\n");
+		pr_err("error sending genl reply\n");
 }
 
 /* Used on a fresh "drbd_adm_prepare"d reply_skb, this cannot fail: The only
@@ -558,8 +560,10 @@
 }
 
 enum drbd_state_rv
-drbd_set_role(struct drbd_device *device, enum drbd_role new_role, int force)
+drbd_set_role(struct drbd_device *const device, enum drbd_role new_role, int force)
 {
+	struct drbd_peer_device *const peer_device = first_peer_device(device);
+	struct drbd_connection *const connection = peer_device ? peer_device->connection : NULL;
 	const int max_tries = 4;
 	enum drbd_state_rv rv = SS_UNKNOWN_ERROR;
 	struct net_conf *nc;
@@ -607,7 +611,7 @@
 		    device->state.disk == D_CONSISTENT && mask.pdsk == 0) {
 			D_ASSERT(device, device->state.pdsk == D_UNKNOWN);
 
-			if (conn_try_outdate_peer(first_peer_device(device)->connection)) {
+			if (conn_try_outdate_peer(connection)) {
 				val.disk = D_UP_TO_DATE;
 				mask.disk = D_MASK;
 			}
@@ -617,7 +621,7 @@
 		if (rv == SS_NOTHING_TO_DO)
 			goto out;
 		if (rv == SS_PRIMARY_NOP && mask.pdsk == 0) {
-			if (!conn_try_outdate_peer(first_peer_device(device)->connection) && force) {
+			if (!conn_try_outdate_peer(connection) && force) {
 				drbd_warn(device, "Forced into split brain situation!\n");
 				mask.pdsk = D_MASK;
 				val.pdsk  = D_OUTDATED;
@@ -630,7 +634,7 @@
 			   retry at most once more in this case. */
 			int timeo;
 			rcu_read_lock();
-			nc = rcu_dereference(first_peer_device(device)->connection->net_conf);
+			nc = rcu_dereference(connection->net_conf);
 			timeo = nc ? (nc->ping_timeo + 1) * HZ / 10 : 1;
 			rcu_read_unlock();
 			schedule_timeout_interruptible(timeo);
@@ -659,19 +663,17 @@
 	/* FIXME also wait for all pending P_BARRIER_ACK? */
 
 	if (new_role == R_SECONDARY) {
-		set_disk_ro(device->vdisk, true);
 		if (get_ldev(device)) {
 			device->ldev->md.uuid[UI_CURRENT] &= ~(u64)1;
 			put_ldev(device);
 		}
 	} else {
-		/* Called from drbd_adm_set_role only.
-		 * We are still holding the conf_update mutex. */
-		nc = first_peer_device(device)->connection->net_conf;
+		mutex_lock(&device->resource->conf_update);
+		nc = connection->net_conf;
 		if (nc)
 			nc->discard_my_data = 0; /* without copy; single bit op is atomic */
+		mutex_unlock(&device->resource->conf_update);
 
-		set_disk_ro(device->vdisk, false);
 		if (get_ldev(device)) {
 			if (((device->state.conn < C_CONNECTED ||
 			       device->state.pdsk <= D_FAILED)
@@ -689,12 +691,12 @@
 	if (device->state.conn >= C_WF_REPORT_PARAMS) {
 		/* if this was forced, we should consider sync */
 		if (forced)
-			drbd_send_uuids(first_peer_device(device));
-		drbd_send_current_state(first_peer_device(device));
+			drbd_send_uuids(peer_device);
+		drbd_send_current_state(peer_device);
 	}
 
 	drbd_md_sync(device);
-
+	set_disk_ro(device->vdisk, new_role == R_SECONDARY);
 	kobject_uevent(&disk_to_dev(device->vdisk)->kobj, KOBJ_CHANGE);
 out:
 	mutex_unlock(device->state_mutex);
@@ -891,7 +893,7 @@
 	 * still lock the act_log to not trigger ASSERTs there.
 	 */
 	drbd_suspend_io(device);
-	buffer = drbd_md_get_buffer(device); /* Lock meta-data IO */
+	buffer = drbd_md_get_buffer(device, __func__); /* Lock meta-data IO */
 	if (!buffer) {
 		drbd_resume_io(device);
 		return DS_ERROR;
@@ -971,6 +973,10 @@
 	if (la_size_changed || md_moved || rs) {
 		u32 prev_flags;
 
+		/* We do some synchronous IO below, which may take some time.
+		 * Clear the timer, to avoid scary "timer expired!" messages,
+		 * "Superblock" is written out at least twice below, anyways. */
+		del_timer(&device->md_sync_timer);
 		drbd_al_shrink(device); /* All extents inactive. */
 
 		prev_flags = md->flags;
@@ -1116,15 +1122,16 @@
 	return 0;
 }
 
-static void drbd_setup_queue_param(struct drbd_device *device, unsigned int max_bio_size)
+static void drbd_setup_queue_param(struct drbd_device *device, struct drbd_backing_dev *bdev,
+				   unsigned int max_bio_size)
 {
 	struct request_queue * const q = device->rq_queue;
 	unsigned int max_hw_sectors = max_bio_size >> 9;
 	unsigned int max_segments = 0;
 	struct request_queue *b = NULL;
 
-	if (get_ldev_if_state(device, D_ATTACHING)) {
-		b = device->ldev->backing_bdev->bd_disk->queue;
+	if (bdev) {
+		b = bdev->backing_bdev->bd_disk->queue;
 
 		max_hw_sectors = min(queue_max_hw_sectors(b), max_bio_size >> 9);
 		rcu_read_lock();
@@ -1169,11 +1176,10 @@
 				 b->backing_dev_info.ra_pages);
 			q->backing_dev_info.ra_pages = b->backing_dev_info.ra_pages;
 		}
-		put_ldev(device);
 	}
 }
 
-void drbd_reconsider_max_bio_size(struct drbd_device *device)
+void drbd_reconsider_max_bio_size(struct drbd_device *device, struct drbd_backing_dev *bdev)
 {
 	unsigned int now, new, local, peer;
 
@@ -1181,10 +1187,9 @@
 	local = device->local_max_bio_size; /* Eventually last known value, from volatile memory */
 	peer = device->peer_max_bio_size; /* Eventually last known value, from meta data */
 
-	if (get_ldev_if_state(device, D_ATTACHING)) {
-		local = queue_max_hw_sectors(device->ldev->backing_bdev->bd_disk->queue) << 9;
+	if (bdev) {
+		local = queue_max_hw_sectors(bdev->backing_bdev->bd_disk->queue) << 9;
 		device->local_max_bio_size = local;
-		put_ldev(device);
 	}
 	local = min(local, DRBD_MAX_BIO_SIZE);
 
@@ -1217,7 +1222,7 @@
 	if (new != now)
 		drbd_info(device, "max BIO size = %u\n", new);
 
-	drbd_setup_queue_param(device, new);
+	drbd_setup_queue_param(device, bdev, new);
 }
 
 /* Starts the worker thread */
@@ -1299,6 +1304,13 @@
 	return (al_size_4k - 1) * AL_CONTEXT_PER_TRANSACTION;
 }
 
+static bool write_ordering_changed(struct disk_conf *a, struct disk_conf *b)
+{
+	return	a->disk_barrier != b->disk_barrier ||
+		a->disk_flushes != b->disk_flushes ||
+		a->disk_drain != b->disk_drain;
+}
+
 int drbd_adm_disk_opts(struct sk_buff *skb, struct genl_info *info)
 {
 	struct drbd_config_context adm_ctx;
@@ -1405,7 +1417,8 @@
 	else
 		set_bit(MD_NO_FUA, &device->flags);
 
-	drbd_bump_write_ordering(first_peer_device(device)->connection, WO_bdev_flush);
+	if (write_ordering_changed(old_disk_conf, new_disk_conf))
+		drbd_bump_write_ordering(device->resource, NULL, WO_bdev_flush);
 
 	drbd_md_sync(device);
 
@@ -1440,6 +1453,8 @@
 {
 	struct drbd_config_context adm_ctx;
 	struct drbd_device *device;
+	struct drbd_peer_device *peer_device;
+	struct drbd_connection *connection;
 	int err;
 	enum drbd_ret_code retcode;
 	enum determine_dev_size dd;
@@ -1462,7 +1477,9 @@
 
 	device = adm_ctx.device;
 	mutex_lock(&adm_ctx.resource->adm_mutex);
-	conn_reconfig_start(first_peer_device(device)->connection);
+	peer_device = first_peer_device(device);
+	connection = peer_device ? peer_device->connection : NULL;
+	conn_reconfig_start(connection);
 
 	/* if you want to reconfigure, please tear down first */
 	if (device->state.disk > D_DISKLESS) {
@@ -1473,7 +1490,7 @@
 	 * drbd_ldev_destroy is done already, we may end up here very fast,
 	 * e.g. if someone calls attach from the on-io-error handler,
 	 * to realize a "hot spare" feature (not that I'd recommend that) */
-	wait_event(device->misc_wait, !atomic_read(&device->local_cnt));
+	wait_event(device->misc_wait, !test_bit(GOING_DISKLESS, &device->flags));
 
 	/* make sure there is no leftover from previous force-detach attempts */
 	clear_bit(FORCE_DETACH, &device->flags);
@@ -1529,7 +1546,7 @@
 		goto fail;
 
 	rcu_read_lock();
-	nc = rcu_dereference(first_peer_device(device)->connection->net_conf);
+	nc = rcu_dereference(connection->net_conf);
 	if (nc) {
 		if (new_disk_conf->fencing == FP_STONITH && nc->wire_protocol == DRBD_PROT_A) {
 			rcu_read_unlock();
@@ -1649,7 +1666,7 @@
 	 */
 	wait_event(device->misc_wait, !atomic_read(&device->ap_pending_cnt) || drbd_suspended(device));
 	/* and for any other previously queued work */
-	drbd_flush_workqueue(&first_peer_device(device)->connection->sender_work);
+	drbd_flush_workqueue(&connection->sender_work);
 
 	rv = _drbd_request_state(device, NS(disk, D_ATTACHING), CS_VERBOSE);
 	retcode = rv;  /* FIXME: Type mismatch. */
@@ -1710,7 +1727,7 @@
 	new_disk_conf = NULL;
 	new_plan = NULL;
 
-	drbd_bump_write_ordering(first_peer_device(device)->connection, WO_bdev_flush);
+	drbd_bump_write_ordering(device->resource, device->ldev, WO_bdev_flush);
 
 	if (drbd_md_test_flag(device->ldev, MDF_CRASHED_PRIMARY))
 		set_bit(CRASHED_PRIMARY, &device->flags);
@@ -1726,7 +1743,7 @@
 	device->read_cnt = 0;
 	device->writ_cnt = 0;
 
-	drbd_reconsider_max_bio_size(device);
+	drbd_reconsider_max_bio_size(device, device->ldev);
 
 	/* If I am currently not R_PRIMARY,
 	 * but meta data primary indicator is set,
@@ -1845,7 +1862,7 @@
 
 	kobject_uevent(&disk_to_dev(device->vdisk)->kobj, KOBJ_CHANGE);
 	put_ldev(device);
-	conn_reconfig_done(first_peer_device(device)->connection);
+	conn_reconfig_done(connection);
 	mutex_unlock(&adm_ctx.resource->adm_mutex);
 	drbd_adm_finish(&adm_ctx, info, retcode);
 	return 0;
@@ -1856,7 +1873,7 @@
 	drbd_force_state(device, NS(disk, D_DISKLESS));
 	drbd_md_sync(device);
  fail:
-	conn_reconfig_done(first_peer_device(device)->connection);
+	conn_reconfig_done(connection);
 	if (nbc) {
 		if (nbc->backing_bdev)
 			blkdev_put(nbc->backing_bdev,
@@ -1888,7 +1905,7 @@
 	}
 
 	drbd_suspend_io(device); /* so no-one is stuck in drbd_al_begin_io */
-	drbd_md_get_buffer(device); /* make sure there is no in-flight meta-data IO */
+	drbd_md_get_buffer(device, __func__); /* make sure there is no in-flight meta-data IO */
 	retcode = drbd_request_state(device, NS(disk, D_FAILED));
 	drbd_md_put_buffer(device);
 	/* D_FAILED will transition to DISKLESS. */
@@ -2654,8 +2671,13 @@
 	if (retcode != NO_ERROR)
 		goto out;
 
-	mutex_lock(&adm_ctx.resource->adm_mutex);
 	device = adm_ctx.device;
+	if (!get_ldev(device)) {
+		retcode = ERR_NO_DISK;
+		goto out;
+	}
+
+	mutex_lock(&adm_ctx.resource->adm_mutex);
 
 	/* If there is still bitmap IO pending, probably because of a previous
 	 * resync just being finished, wait for it before requesting a new resync.
@@ -2679,6 +2701,7 @@
 		retcode = drbd_request_state(device, NS(conn, C_STARTING_SYNC_T));
 	drbd_resume_io(device);
 	mutex_unlock(&adm_ctx.resource->adm_mutex);
+	put_ldev(device);
 out:
 	drbd_adm_finish(&adm_ctx, info, retcode);
 	return 0;
@@ -2704,7 +2727,7 @@
 	return 0;
 }
 
-static int drbd_bmio_set_susp_al(struct drbd_device *device)
+static int drbd_bmio_set_susp_al(struct drbd_device *device) __must_hold(local)
 {
 	int rv;
 
@@ -2725,8 +2748,13 @@
 	if (retcode != NO_ERROR)
 		goto out;
 
-	mutex_lock(&adm_ctx.resource->adm_mutex);
 	device = adm_ctx.device;
+	if (!get_ldev(device)) {
+		retcode = ERR_NO_DISK;
+		goto out;
+	}
+
+	mutex_lock(&adm_ctx.resource->adm_mutex);
 
 	/* If there is still bitmap IO pending, probably because of a previous
 	 * resync just being finished, wait for it before requesting a new resync.
@@ -2753,6 +2781,7 @@
 		retcode = drbd_request_state(device, NS(conn, C_STARTING_SYNC_S));
 	drbd_resume_io(device);
 	mutex_unlock(&adm_ctx.resource->adm_mutex);
+	put_ldev(device);
 out:
 	drbd_adm_finish(&adm_ctx, info, retcode);
 	return 0;
@@ -2892,7 +2921,7 @@
 	return list_first_entry(&resource->connections, struct drbd_connection, connections);
 }
 
-int nla_put_status_info(struct sk_buff *skb, struct drbd_device *device,
+static int nla_put_status_info(struct sk_buff *skb, struct drbd_device *device,
 		const struct sib_info *sib)
 {
 	struct drbd_resource *resource = device->resource;
@@ -3622,13 +3651,6 @@
 	unsigned seq;
 	int err = -ENOMEM;
 
-	if (sib->sib_reason == SIB_SYNC_PROGRESS) {
-		if (time_after(jiffies, device->rs_last_bcast + HZ))
-			device->rs_last_bcast = jiffies;
-		else
-			return;
-	}
-
 	seq = atomic_inc_return(&drbd_genl_seq);
 	msg = genlmsg_new(NLMSG_GOODSIZE, GFP_NOIO);
 	if (!msg)
diff --git a/drivers/block/drbd/drbd_proc.c b/drivers/block/drbd/drbd_proc.c
index 89736bd..06e6147 100644
--- a/drivers/block/drbd/drbd_proc.c
+++ b/drivers/block/drbd/drbd_proc.c
@@ -60,20 +60,65 @@
 		seq_printf(seq, "%ld", v);
 }
 
+static void drbd_get_syncer_progress(struct drbd_device *device,
+		union drbd_dev_state state, unsigned long *rs_total,
+		unsigned long *bits_left, unsigned int *per_mil_done)
+{
+	/* this is to break it at compile time when we change that, in case we
+	 * want to support more than (1<<32) bits on a 32bit arch. */
+	typecheck(unsigned long, device->rs_total);
+	*rs_total = device->rs_total;
+
+	/* note: both rs_total and rs_left are in bits, i.e. in
+	 * units of BM_BLOCK_SIZE.
+	 * for the percentage, we don't care. */
+
+	if (state.conn == C_VERIFY_S || state.conn == C_VERIFY_T)
+		*bits_left = device->ov_left;
+	else
+		*bits_left = drbd_bm_total_weight(device) - device->rs_failed;
+	/* >> 10 to prevent overflow,
+	 * +1 to prevent division by zero */
+	if (*bits_left > *rs_total) {
+		/* D'oh. Maybe a logic bug somewhere.  More likely just a race
+		 * between state change and reset of rs_total.
+		 */
+		*bits_left = *rs_total;
+		*per_mil_done = *rs_total ? 0 : 1000;
+	} else {
+		/* Make sure the division happens in long context.
+		 * We allow up to one petabyte storage right now,
+		 * at a granularity of 4k per bit that is 2**38 bits.
+		 * After shift right and multiplication by 1000,
+		 * this should still fit easily into a 32bit long,
+		 * so we don't need a 64bit division on 32bit arch.
+		 * Note: currently we don't support such large bitmaps on 32bit
+		 * arch anyways, but no harm done to be prepared for it here.
+		 */
+		unsigned int shift = *rs_total > UINT_MAX ? 16 : 10;
+		unsigned long left = *bits_left >> shift;
+		unsigned long total = 1UL + (*rs_total >> shift);
+		unsigned long tmp = 1000UL - left * 1000UL/total;
+		*per_mil_done = tmp;
+	}
+}
+
+
 /*lge
  * progress bars shamelessly adapted from driver/md/md.c
  * output looks like
  *	[=====>..............] 33.5% (23456/123456)
  *	finish: 2:20:20 speed: 6,345 (6,456) K/sec
  */
-static void drbd_syncer_progress(struct drbd_device *device, struct seq_file *seq)
+static void drbd_syncer_progress(struct drbd_device *device, struct seq_file *seq,
+		union drbd_dev_state state)
 {
-	unsigned long db, dt, dbdt, rt, rs_left;
+	unsigned long db, dt, dbdt, rt, rs_total, rs_left;
 	unsigned int res;
 	int i, x, y;
 	int stalled = 0;
 
-	drbd_get_syncer_progress(device, &rs_left, &res);
+	drbd_get_syncer_progress(device, state, &rs_total, &rs_left, &res);
 
 	x = res/50;
 	y = 20-x;
@@ -85,21 +130,21 @@
 		seq_printf(seq, ".");
 	seq_printf(seq, "] ");
 
-	if (device->state.conn == C_VERIFY_S || device->state.conn == C_VERIFY_T)
+	if (state.conn == C_VERIFY_S || state.conn == C_VERIFY_T)
 		seq_printf(seq, "verified:");
 	else
 		seq_printf(seq, "sync'ed:");
 	seq_printf(seq, "%3u.%u%% ", res / 10, res % 10);
 
 	/* if more than a few GB, display in MB */
-	if (device->rs_total > (4UL << (30 - BM_BLOCK_SHIFT)))
+	if (rs_total > (4UL << (30 - BM_BLOCK_SHIFT)))
 		seq_printf(seq, "(%lu/%lu)M",
 			    (unsigned long) Bit2KB(rs_left >> 10),
-			    (unsigned long) Bit2KB(device->rs_total >> 10));
+			    (unsigned long) Bit2KB(rs_total >> 10));
 	else
 		seq_printf(seq, "(%lu/%lu)K\n\t",
 			    (unsigned long) Bit2KB(rs_left),
-			    (unsigned long) Bit2KB(device->rs_total));
+			    (unsigned long) Bit2KB(rs_total));
 
 	/* see drivers/md/md.c
 	 * We do not want to overflow, so the order of operands and
@@ -150,13 +195,13 @@
 	dt = (jiffies - device->rs_start - device->rs_paused) / HZ;
 	if (dt == 0)
 		dt = 1;
-	db = device->rs_total - rs_left;
+	db = rs_total - rs_left;
 	dbdt = Bit2KB(db/dt);
 	seq_printf_with_thousands_grouping(seq, dbdt);
 	seq_printf(seq, ")");
 
-	if (device->state.conn == C_SYNC_TARGET ||
-	    device->state.conn == C_VERIFY_S) {
+	if (state.conn == C_SYNC_TARGET ||
+	    state.conn == C_VERIFY_S) {
 		seq_printf(seq, " want: ");
 		seq_printf_with_thousands_grouping(seq, device->c_sync_rate);
 	}
@@ -168,8 +213,8 @@
 		unsigned long bm_bits = drbd_bm_bits(device);
 		unsigned long bit_pos;
 		unsigned long long stop_sector = 0;
-		if (device->state.conn == C_VERIFY_S ||
-		    device->state.conn == C_VERIFY_T) {
+		if (state.conn == C_VERIFY_S ||
+		    state.conn == C_VERIFY_T) {
 			bit_pos = bm_bits - device->ov_left;
 			if (verify_can_do_stop_sector(device))
 				stop_sector = device->ov_stop_sector;
@@ -188,22 +233,13 @@
 	}
 }
 
-static void resync_dump_detail(struct seq_file *seq, struct lc_element *e)
-{
-	struct bm_extent *bme = lc_entry(e, struct bm_extent, lce);
-
-	seq_printf(seq, "%5d %s %s\n", bme->rs_left,
-		   bme->flags & BME_NO_WRITES ? "NO_WRITES" : "---------",
-		   bme->flags & BME_LOCKED ? "LOCKED" : "------"
-		   );
-}
-
 static int drbd_seq_show(struct seq_file *seq, void *v)
 {
 	int i, prev_i = -1;
 	const char *sn;
 	struct drbd_device *device;
 	struct net_conf *nc;
+	union drbd_dev_state state;
 	char wp;
 
 	static char write_ordering_chars[] = {
@@ -241,11 +277,12 @@
 			seq_printf(seq, "\n");
 		prev_i = i;
 
-		sn = drbd_conn_str(device->state.conn);
+		state = device->state;
+		sn = drbd_conn_str(state.conn);
 
-		if (device->state.conn == C_STANDALONE &&
-		    device->state.disk == D_DISKLESS &&
-		    device->state.role == R_SECONDARY) {
+		if (state.conn == C_STANDALONE &&
+		    state.disk == D_DISKLESS &&
+		    state.role == R_SECONDARY) {
 			seq_printf(seq, "%2d: cs:Unconfigured\n", i);
 		} else {
 			/* reset device->congestion_reason */
@@ -258,15 +295,15 @@
 			   "    ns:%u nr:%u dw:%u dr:%u al:%u bm:%u "
 			   "lo:%d pe:%d ua:%d ap:%d ep:%d wo:%c",
 			   i, sn,
-			   drbd_role_str(device->state.role),
-			   drbd_role_str(device->state.peer),
-			   drbd_disk_str(device->state.disk),
-			   drbd_disk_str(device->state.pdsk),
+			   drbd_role_str(state.role),
+			   drbd_role_str(state.peer),
+			   drbd_disk_str(state.disk),
+			   drbd_disk_str(state.pdsk),
 			   wp,
 			   drbd_suspended(device) ? 's' : 'r',
-			   device->state.aftr_isp ? 'a' : '-',
-			   device->state.peer_isp ? 'p' : '-',
-			   device->state.user_isp ? 'u' : '-',
+			   state.aftr_isp ? 'a' : '-',
+			   state.peer_isp ? 'p' : '-',
+			   state.user_isp ? 'u' : '-',
 			   device->congestion_reason ?: '-',
 			   test_bit(AL_SUSPENDED, &device->flags) ? 's' : '-',
 			   device->send_cnt/2,
@@ -281,17 +318,17 @@
 			   atomic_read(&device->unacked_cnt),
 			   atomic_read(&device->ap_bio_cnt),
 			   first_peer_device(device)->connection->epochs,
-			   write_ordering_chars[first_peer_device(device)->connection->write_ordering]
+			   write_ordering_chars[device->resource->write_ordering]
 			);
 			seq_printf(seq, " oos:%llu\n",
 				   Bit2KB((unsigned long long)
 					   drbd_bm_total_weight(device)));
 		}
-		if (device->state.conn == C_SYNC_SOURCE ||
-		    device->state.conn == C_SYNC_TARGET ||
-		    device->state.conn == C_VERIFY_S ||
-		    device->state.conn == C_VERIFY_T)
-			drbd_syncer_progress(device, seq);
+		if (state.conn == C_SYNC_SOURCE ||
+		    state.conn == C_SYNC_TARGET ||
+		    state.conn == C_VERIFY_S ||
+		    state.conn == C_VERIFY_T)
+			drbd_syncer_progress(device, seq, state);
 
 		if (proc_details >= 1 && get_ldev_if_state(device, D_FAILED)) {
 			lc_seq_printf_stats(seq, device->resync);
@@ -299,12 +336,8 @@
 			put_ldev(device);
 		}
 
-		if (proc_details >= 2) {
-			if (device->resync) {
-				lc_seq_dump_details(seq, device->resync, "rs_left",
-					resync_dump_detail);
-			}
-		}
+		if (proc_details >= 2)
+			seq_printf(seq, "\tblocked on activity log: %d\n", atomic_read(&device->ap_actlog_cnt));
 	}
 	rcu_read_unlock();
 
@@ -316,7 +349,7 @@
 	int err;
 
 	if (try_module_get(THIS_MODULE)) {
-		err = single_open(file, drbd_seq_show, PDE_DATA(inode));
+		err = single_open(file, drbd_seq_show, NULL);
 		if (err)
 			module_put(THIS_MODULE);
 		return err;
diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index 5b17ec8..9342b8d 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -362,17 +362,14 @@
 			goto fail;
 	}
 
+	memset(peer_req, 0, sizeof(*peer_req));
+	INIT_LIST_HEAD(&peer_req->w.list);
 	drbd_clear_interval(&peer_req->i);
 	peer_req->i.size = data_size;
 	peer_req->i.sector = sector;
-	peer_req->i.local = false;
-	peer_req->i.waiting = false;
-
-	peer_req->epoch = NULL;
+	peer_req->submit_jif = jiffies;
 	peer_req->peer_device = peer_device;
 	peer_req->pages = page;
-	atomic_set(&peer_req->pending_bios, 0);
-	peer_req->flags = 0;
 	/*
 	 * The block_id is opaque to the receiver.  It is not endianness
 	 * converted, and sent back to the sender unchanged.
@@ -389,11 +386,16 @@
 void __drbd_free_peer_req(struct drbd_device *device, struct drbd_peer_request *peer_req,
 		       int is_net)
 {
+	might_sleep();
 	if (peer_req->flags & EE_HAS_DIGEST)
 		kfree(peer_req->digest);
 	drbd_free_pages(device, peer_req->pages, is_net);
 	D_ASSERT(device, atomic_read(&peer_req->pending_bios) == 0);
 	D_ASSERT(device, drbd_interval_empty(&peer_req->i));
+	if (!expect(!(peer_req->flags & EE_CALL_AL_COMPLETE_IO))) {
+		peer_req->flags &= ~EE_CALL_AL_COMPLETE_IO;
+		drbd_al_complete_io(device, &peer_req->i);
+	}
 	mempool_free(peer_req, drbd_ee_mempool);
 }
 
@@ -791,8 +793,18 @@
 {
 	unsigned int header_size = drbd_header_size(connection);
 	struct packet_info pi;
+	struct net_conf *nc;
 	int err;
 
+	rcu_read_lock();
+	nc = rcu_dereference(connection->net_conf);
+	if (!nc) {
+		rcu_read_unlock();
+		return -EIO;
+	}
+	sock->sk->sk_rcvtimeo = nc->ping_timeo * 4 * HZ / 10;
+	rcu_read_unlock();
+
 	err = drbd_recv_short(sock, connection->data.rbuf, header_size, 0);
 	if (err != header_size) {
 		if (err >= 0)
@@ -809,7 +821,7 @@
  * drbd_socket_okay() - Free the socket if its connection is not okay
  * @sock:	pointer to the pointer to the socket.
  */
-static int drbd_socket_okay(struct socket **sock)
+static bool drbd_socket_okay(struct socket **sock)
 {
 	int rr;
 	char tb[4];
@@ -827,6 +839,30 @@
 		return false;
 	}
 }
+
+static bool connection_established(struct drbd_connection *connection,
+				   struct socket **sock1,
+				   struct socket **sock2)
+{
+	struct net_conf *nc;
+	int timeout;
+	bool ok;
+
+	if (!*sock1 || !*sock2)
+		return false;
+
+	rcu_read_lock();
+	nc = rcu_dereference(connection->net_conf);
+	timeout = (nc->sock_check_timeo ?: nc->ping_timeo) * HZ / 10;
+	rcu_read_unlock();
+	schedule_timeout_interruptible(timeout);
+
+	ok = drbd_socket_okay(sock1);
+	ok = drbd_socket_okay(sock2) && ok;
+
+	return ok;
+}
+
 /* Gets called if a connection is established, or if a new minor gets created
    in a connection */
 int drbd_connected(struct drbd_peer_device *peer_device)
@@ -868,8 +904,8 @@
 	struct drbd_socket sock, msock;
 	struct drbd_peer_device *peer_device;
 	struct net_conf *nc;
-	int vnr, timeout, h, ok;
-	bool discard_my_data;
+	int vnr, timeout, h;
+	bool discard_my_data, ok;
 	enum drbd_state_rv rv;
 	struct accept_wait_data ad = {
 		.connection = connection,
@@ -913,17 +949,8 @@
 			}
 		}
 
-		if (sock.socket && msock.socket) {
-			rcu_read_lock();
-			nc = rcu_dereference(connection->net_conf);
-			timeout = nc->ping_timeo * HZ / 10;
-			rcu_read_unlock();
-			schedule_timeout_interruptible(timeout);
-			ok = drbd_socket_okay(&sock.socket);
-			ok = drbd_socket_okay(&msock.socket) && ok;
-			if (ok)
-				break;
-		}
+		if (connection_established(connection, &sock.socket, &msock.socket))
+			break;
 
 retry:
 		s = drbd_wait_for_connect(connection, &ad);
@@ -969,8 +996,7 @@
 				goto out_release_sockets;
 		}
 
-		ok = drbd_socket_okay(&sock.socket);
-		ok = drbd_socket_okay(&msock.socket) && ok;
+		ok = connection_established(connection, &sock.socket, &msock.socket);
 	} while (!ok);
 
 	if (ad.s_listen)
@@ -1151,7 +1177,7 @@
 	struct drbd_peer_device *peer_device;
 	int vnr;
 
-	if (connection->write_ordering >= WO_bdev_flush) {
+	if (connection->resource->write_ordering >= WO_bdev_flush) {
 		rcu_read_lock();
 		idr_for_each_entry(&connection->peer_devices, peer_device, vnr) {
 			struct drbd_device *device = peer_device->device;
@@ -1161,14 +1187,22 @@
 			kref_get(&device->kref);
 			rcu_read_unlock();
 
+			/* Right now, we have only this one synchronous code path
+			 * for flushes between request epochs.
+			 * We may want to make those asynchronous,
+			 * or at least parallelize the flushes to the volume devices.
+			 */
+			device->flush_jif = jiffies;
+			set_bit(FLUSH_PENDING, &device->flags);
 			rv = blkdev_issue_flush(device->ldev->backing_bdev,
 					GFP_NOIO, NULL);
+			clear_bit(FLUSH_PENDING, &device->flags);
 			if (rv) {
 				drbd_info(device, "local disk flush failed with status %d\n", rv);
 				/* would rather check on EOPNOTSUPP, but that is not reliable.
 				 * don't try again for ANY return value != 0
 				 * if (rv == -EOPNOTSUPP) */
-				drbd_bump_write_ordering(connection, WO_drain_io);
+				drbd_bump_write_ordering(connection->resource, NULL, WO_drain_io);
 			}
 			put_ldev(device);
 			kref_put(&device->kref, drbd_destroy_device);
@@ -1257,15 +1291,30 @@
 	return rv;
 }
 
+static enum write_ordering_e
+max_allowed_wo(struct drbd_backing_dev *bdev, enum write_ordering_e wo)
+{
+	struct disk_conf *dc;
+
+	dc = rcu_dereference(bdev->disk_conf);
+
+	if (wo == WO_bdev_flush && !dc->disk_flushes)
+		wo = WO_drain_io;
+	if (wo == WO_drain_io && !dc->disk_drain)
+		wo = WO_none;
+
+	return wo;
+}
+
 /**
  * drbd_bump_write_ordering() - Fall back to an other write ordering method
  * @connection:	DRBD connection.
  * @wo:		Write ordering method to try.
  */
-void drbd_bump_write_ordering(struct drbd_connection *connection, enum write_ordering_e wo)
+void drbd_bump_write_ordering(struct drbd_resource *resource, struct drbd_backing_dev *bdev,
+			      enum write_ordering_e wo)
 {
-	struct disk_conf *dc;
-	struct drbd_peer_device *peer_device;
+	struct drbd_device *device;
 	enum write_ordering_e pwo;
 	int vnr;
 	static char *write_ordering_str[] = {
@@ -1274,26 +1323,27 @@
 		[WO_bdev_flush] = "flush",
 	};
 
-	pwo = connection->write_ordering;
-	wo = min(pwo, wo);
+	pwo = resource->write_ordering;
+	if (wo != WO_bdev_flush)
+		wo = min(pwo, wo);
 	rcu_read_lock();
-	idr_for_each_entry(&connection->peer_devices, peer_device, vnr) {
-		struct drbd_device *device = peer_device->device;
-
-		if (!get_ldev_if_state(device, D_ATTACHING))
-			continue;
-		dc = rcu_dereference(device->ldev->disk_conf);
-
-		if (wo == WO_bdev_flush && !dc->disk_flushes)
-			wo = WO_drain_io;
-		if (wo == WO_drain_io && !dc->disk_drain)
-			wo = WO_none;
-		put_ldev(device);
+	idr_for_each_entry(&resource->devices, device, vnr) {
+		if (get_ldev(device)) {
+			wo = max_allowed_wo(device->ldev, wo);
+			if (device->ldev == bdev)
+				bdev = NULL;
+			put_ldev(device);
+		}
 	}
+
+	if (bdev)
+		wo = max_allowed_wo(bdev, wo);
+
 	rcu_read_unlock();
-	connection->write_ordering = wo;
-	if (pwo != connection->write_ordering || wo == WO_bdev_flush)
-		drbd_info(connection, "Method to ensure write ordering: %s\n", write_ordering_str[connection->write_ordering]);
+
+	resource->write_ordering = wo;
+	if (pwo != resource->write_ordering || wo == WO_bdev_flush)
+		drbd_info(resource, "Method to ensure write ordering: %s\n", write_ordering_str[resource->write_ordering]);
 }
 
 /**
@@ -1330,6 +1380,13 @@
 		/* wait for all pending IO completions, before we start
 		 * zeroing things out. */
 		conn_wait_active_ee_empty(first_peer_device(device)->connection);
+		/* add it to the active list now,
+		 * so we can find it to present it in debugfs */
+		peer_req->submit_jif = jiffies;
+		peer_req->flags |= EE_SUBMITTED;
+		spin_lock_irq(&device->resource->req_lock);
+		list_add_tail(&peer_req->w.list, &device->active_ee);
+		spin_unlock_irq(&device->resource->req_lock);
 		if (blkdev_issue_zeroout(device->ldev->backing_bdev,
 			sector, ds >> 9, GFP_NOIO))
 			peer_req->flags |= EE_WAS_ERROR;
@@ -1398,6 +1455,9 @@
 	D_ASSERT(device, page == NULL);
 
 	atomic_set(&peer_req->pending_bios, n_bios);
+	/* for debugfs: update timestamp, mark as submitted */
+	peer_req->submit_jif = jiffies;
+	peer_req->flags |= EE_SUBMITTED;
 	do {
 		bio = bios;
 		bios = bios->bi_next;
@@ -1471,7 +1531,7 @@
 	 * R_PRIMARY crashes now.
 	 * Therefore we must send the barrier_ack after the barrier request was
 	 * completed. */
-	switch (connection->write_ordering) {
+	switch (connection->resource->write_ordering) {
 	case WO_none:
 		if (rv == FE_RECYCLED)
 			return 0;
@@ -1498,7 +1558,8 @@
 
 		return 0;
 	default:
-		drbd_err(connection, "Strangeness in connection->write_ordering %d\n", connection->write_ordering);
+		drbd_err(connection, "Strangeness in connection->write_ordering %d\n",
+			 connection->resource->write_ordering);
 		return -EIO;
 	}
 
@@ -1531,7 +1592,7 @@
 	struct drbd_peer_request *peer_req;
 	struct page *page;
 	int dgs, ds, err;
-	int data_size = pi->size;
+	unsigned int data_size = pi->size;
 	void *dig_in = peer_device->connection->int_dig_in;
 	void *dig_vv = peer_device->connection->int_dig_vv;
 	unsigned long *data;
@@ -1578,6 +1639,7 @@
 	if (!peer_req)
 		return NULL;
 
+	peer_req->flags |= EE_WRITE;
 	if (trim)
 		return peer_req;
 
@@ -1734,9 +1796,10 @@
 	 * respective _drbd_clear_done_ee */
 
 	peer_req->w.cb = e_end_resync_block;
+	peer_req->submit_jif = jiffies;
 
 	spin_lock_irq(&device->resource->req_lock);
-	list_add(&peer_req->w.list, &device->sync_ee);
+	list_add_tail(&peer_req->w.list, &device->sync_ee);
 	spin_unlock_irq(&device->resource->req_lock);
 
 	atomic_add(pi->size >> 9, &device->rs_sect_ev);
@@ -1889,6 +1952,7 @@
 		}
 		dec_unacked(device);
 	}
+
 	/* we delete from the conflict detection hash _after_ we sent out the
 	 * P_WRITE_ACK / P_NEG_ACK, to get the sequence number right.  */
 	if (peer_req->flags & EE_IN_INTERVAL_TREE) {
@@ -2115,6 +2179,8 @@
 	drbd_for_each_overlap(i, &device->write_requests, sector, size) {
 		if (i == &peer_req->i)
 			continue;
+		if (i->completed)
+			continue;
 
 		if (!i->local) {
 			/*
@@ -2147,7 +2213,6 @@
 					  (unsigned long long)sector, size,
 					  superseded ? "local" : "remote");
 
-			inc_unacked(device);
 			peer_req->w.cb = superseded ? e_send_superseded :
 						   e_send_retry_write;
 			list_add_tail(&peer_req->w.list, &device->done_ee);
@@ -2206,6 +2271,7 @@
 {
 	struct drbd_peer_device *peer_device;
 	struct drbd_device *device;
+	struct net_conf *nc;
 	sector_t sector;
 	struct drbd_peer_request *peer_req;
 	struct p_data *p = pi->data;
@@ -2245,6 +2311,8 @@
 	}
 
 	peer_req->w.cb = e_end_block;
+	peer_req->submit_jif = jiffies;
+	peer_req->flags |= EE_APPLICATION;
 
 	dp_flags = be32_to_cpu(p->dp_flags);
 	rw |= wire_flags_to_bio(dp_flags);
@@ -2271,9 +2339,36 @@
 	spin_unlock(&connection->epoch_lock);
 
 	rcu_read_lock();
-	tp = rcu_dereference(peer_device->connection->net_conf)->two_primaries;
+	nc = rcu_dereference(peer_device->connection->net_conf);
+	tp = nc->two_primaries;
+	if (peer_device->connection->agreed_pro_version < 100) {
+		switch (nc->wire_protocol) {
+		case DRBD_PROT_C:
+			dp_flags |= DP_SEND_WRITE_ACK;
+			break;
+		case DRBD_PROT_B:
+			dp_flags |= DP_SEND_RECEIVE_ACK;
+			break;
+		}
+	}
 	rcu_read_unlock();
+
+	if (dp_flags & DP_SEND_WRITE_ACK) {
+		peer_req->flags |= EE_SEND_WRITE_ACK;
+		inc_unacked(device);
+		/* corresponding dec_unacked() in e_end_block()
+		 * respective _drbd_clear_done_ee */
+	}
+
+	if (dp_flags & DP_SEND_RECEIVE_ACK) {
+		/* I really don't like it that the receiver thread
+		 * sends on the msock, but anyways */
+		drbd_send_ack(first_peer_device(device), P_RECV_ACK, peer_req);
+	}
+
 	if (tp) {
+		/* two primaries implies protocol C */
+		D_ASSERT(device, dp_flags & DP_SEND_WRITE_ACK);
 		peer_req->flags |= EE_IN_INTERVAL_TREE;
 		err = wait_for_and_update_peer_seq(peer_device, peer_seq);
 		if (err)
@@ -2297,44 +2392,18 @@
 	 * active_ee to become empty in drbd_submit_peer_request();
 	 * better not add ourselves here. */
 	if ((peer_req->flags & EE_IS_TRIM_USE_ZEROOUT) == 0)
-		list_add(&peer_req->w.list, &device->active_ee);
+		list_add_tail(&peer_req->w.list, &device->active_ee);
 	spin_unlock_irq(&device->resource->req_lock);
 
 	if (device->state.conn == C_SYNC_TARGET)
 		wait_event(device->ee_wait, !overlapping_resync_write(device, peer_req));
 
-	if (peer_device->connection->agreed_pro_version < 100) {
-		rcu_read_lock();
-		switch (rcu_dereference(peer_device->connection->net_conf)->wire_protocol) {
-		case DRBD_PROT_C:
-			dp_flags |= DP_SEND_WRITE_ACK;
-			break;
-		case DRBD_PROT_B:
-			dp_flags |= DP_SEND_RECEIVE_ACK;
-			break;
-		}
-		rcu_read_unlock();
-	}
-
-	if (dp_flags & DP_SEND_WRITE_ACK) {
-		peer_req->flags |= EE_SEND_WRITE_ACK;
-		inc_unacked(device);
-		/* corresponding dec_unacked() in e_end_block()
-		 * respective _drbd_clear_done_ee */
-	}
-
-	if (dp_flags & DP_SEND_RECEIVE_ACK) {
-		/* I really don't like it that the receiver thread
-		 * sends on the msock, but anyways */
-		drbd_send_ack(first_peer_device(device), P_RECV_ACK, peer_req);
-	}
-
 	if (device->state.pdsk < D_INCONSISTENT) {
 		/* In case we have the only disk of the cluster, */
 		drbd_set_out_of_sync(device, peer_req->i.sector, peer_req->i.size);
-		peer_req->flags |= EE_CALL_AL_COMPLETE_IO;
 		peer_req->flags &= ~EE_MAY_SET_IN_SYNC;
-		drbd_al_begin_io(device, &peer_req->i, true);
+		drbd_al_begin_io(device, &peer_req->i);
+		peer_req->flags |= EE_CALL_AL_COMPLETE_IO;
 	}
 
 	err = drbd_submit_peer_request(device, peer_req, rw, DRBD_FAULT_DT_WR);
@@ -2347,8 +2416,10 @@
 	list_del(&peer_req->w.list);
 	drbd_remove_epoch_entry_interval(device, peer_req);
 	spin_unlock_irq(&device->resource->req_lock);
-	if (peer_req->flags & EE_CALL_AL_COMPLETE_IO)
+	if (peer_req->flags & EE_CALL_AL_COMPLETE_IO) {
+		peer_req->flags &= ~EE_CALL_AL_COMPLETE_IO;
 		drbd_al_complete_io(device, &peer_req->i);
+	}
 
 out_interrupted:
 	drbd_may_finish_epoch(connection, peer_req->epoch, EV_PUT + EV_CLEANUP);
@@ -2368,13 +2439,14 @@
  * The current sync rate used here uses only the most recent two step marks,
  * to have a short time average so we can react faster.
  */
-bool drbd_rs_should_slow_down(struct drbd_device *device, sector_t sector)
+bool drbd_rs_should_slow_down(struct drbd_device *device, sector_t sector,
+		bool throttle_if_app_is_waiting)
 {
 	struct lc_element *tmp;
-	bool throttle = true;
+	bool throttle = drbd_rs_c_min_rate_throttle(device);
 
-	if (!drbd_rs_c_min_rate_throttle(device))
-		return false;
+	if (!throttle || throttle_if_app_is_waiting)
+		return throttle;
 
 	spin_lock_irq(&device->al_lock);
 	tmp = lc_find(device->resync, BM_SECT_TO_EXT(sector));
@@ -2382,7 +2454,8 @@
 		struct bm_extent *bm_ext = lc_entry(tmp, struct bm_extent, lce);
 		if (test_bit(BME_PRIORITY, &bm_ext->flags))
 			throttle = false;
-		/* Do not slow down if app IO is already waiting for this extent */
+		/* Do not slow down if app IO is already waiting for this extent,
+		 * and our progress is necessary for application IO to complete. */
 	}
 	spin_unlock_irq(&device->al_lock);
 
@@ -2407,7 +2480,9 @@
 	curr_events = (int)part_stat_read(&disk->part0, sectors[0]) +
 		      (int)part_stat_read(&disk->part0, sectors[1]) -
 			atomic_read(&device->rs_sect_ev);
-	if (!device->rs_last_events || curr_events - device->rs_last_events > 64) {
+
+	if (atomic_read(&device->ap_actlog_cnt)
+	    || !device->rs_last_events || curr_events - device->rs_last_events > 64) {
 		unsigned long rs_left;
 		int i;
 
@@ -2508,6 +2583,7 @@
 		peer_req->w.cb = w_e_end_data_req;
 		fault_type = DRBD_FAULT_DT_RD;
 		/* application IO, don't drbd_rs_begin_io */
+		peer_req->flags |= EE_APPLICATION;
 		goto submit;
 
 	case P_RS_DATA_REQUEST:
@@ -2538,6 +2614,8 @@
 			peer_req->w.cb = w_e_end_csum_rs_req;
 			/* used in the sector offset progress display */
 			device->bm_resync_fo = BM_SECT_TO_BIT(sector);
+			/* remember to report stats in drbd_resync_finished */
+			device->use_csums = true;
 		} else if (pi->cmd == P_OV_REPLY) {
 			/* track progress, we may need to throttle */
 			atomic_add(size >> 9, &device->rs_sect_in);
@@ -2595,8 +2673,20 @@
 	 * we would also throttle its application reads.
 	 * In that case, throttling is done on the SyncTarget only.
 	 */
-	if (device->state.peer != R_PRIMARY && drbd_rs_should_slow_down(device, sector))
+
+	/* Even though this may be a resync request, we do add to "read_ee";
+	 * "sync_ee" is only used for resync WRITEs.
+	 * Add to list early, so debugfs can find this request
+	 * even if we have to sleep below. */
+	spin_lock_irq(&device->resource->req_lock);
+	list_add_tail(&peer_req->w.list, &device->read_ee);
+	spin_unlock_irq(&device->resource->req_lock);
+
+	update_receiver_timing_details(connection, drbd_rs_should_slow_down);
+	if (device->state.peer != R_PRIMARY
+	&& drbd_rs_should_slow_down(device, sector, false))
 		schedule_timeout_uninterruptible(HZ/10);
+	update_receiver_timing_details(connection, drbd_rs_begin_io);
 	if (drbd_rs_begin_io(device, sector))
 		goto out_free_e;
 
@@ -2604,22 +2694,20 @@
 	atomic_add(size >> 9, &device->rs_sect_ev);
 
 submit:
+	update_receiver_timing_details(connection, drbd_submit_peer_request);
 	inc_unacked(device);
-	spin_lock_irq(&device->resource->req_lock);
-	list_add_tail(&peer_req->w.list, &device->read_ee);
-	spin_unlock_irq(&device->resource->req_lock);
-
 	if (drbd_submit_peer_request(device, peer_req, READ, fault_type) == 0)
 		return 0;
 
 	/* don't care for the reason here */
 	drbd_err(device, "submit failed, triggering re-connect\n");
+
+out_free_e:
 	spin_lock_irq(&device->resource->req_lock);
 	list_del(&peer_req->w.list);
 	spin_unlock_irq(&device->resource->req_lock);
 	/* no drbd_rs_complete_io(), we are dropping the connection anyways */
 
-out_free_e:
 	put_ldev(device);
 	drbd_free_peer_req(device, peer_req);
 	return -EIO;
@@ -2842,8 +2930,10 @@
 -1091   requires proto 91
 -1096   requires proto 96
  */
-static int drbd_uuid_compare(struct drbd_device *device, int *rule_nr) __must_hold(local)
+static int drbd_uuid_compare(struct drbd_device *const device, int *rule_nr) __must_hold(local)
 {
+	struct drbd_peer_device *const peer_device = first_peer_device(device);
+	struct drbd_connection *const connection = peer_device ? peer_device->connection : NULL;
 	u64 self, peer;
 	int i, j;
 
@@ -2869,7 +2959,7 @@
 
 		if (device->p_uuid[UI_BITMAP] == (u64)0 && device->ldev->md.uuid[UI_BITMAP] != (u64)0) {
 
-			if (first_peer_device(device)->connection->agreed_pro_version < 91)
+			if (connection->agreed_pro_version < 91)
 				return -1091;
 
 			if ((device->ldev->md.uuid[UI_BITMAP] & ~((u64)1)) == (device->p_uuid[UI_HISTORY_START] & ~((u64)1)) &&
@@ -2892,7 +2982,7 @@
 
 		if (device->ldev->md.uuid[UI_BITMAP] == (u64)0 && device->p_uuid[UI_BITMAP] != (u64)0) {
 
-			if (first_peer_device(device)->connection->agreed_pro_version < 91)
+			if (connection->agreed_pro_version < 91)
 				return -1091;
 
 			if ((device->ldev->md.uuid[UI_HISTORY_START] & ~((u64)1)) == (device->p_uuid[UI_BITMAP] & ~((u64)1)) &&
@@ -2925,7 +3015,7 @@
 		case 1: /*  self_pri && !peer_pri */ return 1;
 		case 2: /* !self_pri &&  peer_pri */ return -1;
 		case 3: /*  self_pri &&  peer_pri */
-			dc = test_bit(RESOLVE_CONFLICTS, &first_peer_device(device)->connection->flags);
+			dc = test_bit(RESOLVE_CONFLICTS, &connection->flags);
 			return dc ? -1 : 1;
 		}
 	}
@@ -2938,14 +3028,14 @@
 	*rule_nr = 51;
 	peer = device->p_uuid[UI_HISTORY_START] & ~((u64)1);
 	if (self == peer) {
-		if (first_peer_device(device)->connection->agreed_pro_version < 96 ?
+		if (connection->agreed_pro_version < 96 ?
 		    (device->ldev->md.uuid[UI_HISTORY_START] & ~((u64)1)) ==
 		    (device->p_uuid[UI_HISTORY_START + 1] & ~((u64)1)) :
 		    peer + UUID_NEW_BM_OFFSET == (device->p_uuid[UI_BITMAP] & ~((u64)1))) {
 			/* The last P_SYNC_UUID did not get though. Undo the last start of
 			   resync as sync source modifications of the peer's UUIDs. */
 
-			if (first_peer_device(device)->connection->agreed_pro_version < 91)
+			if (connection->agreed_pro_version < 91)
 				return -1091;
 
 			device->p_uuid[UI_BITMAP] = device->p_uuid[UI_HISTORY_START];
@@ -2975,14 +3065,14 @@
 	*rule_nr = 71;
 	self = device->ldev->md.uuid[UI_HISTORY_START] & ~((u64)1);
 	if (self == peer) {
-		if (first_peer_device(device)->connection->agreed_pro_version < 96 ?
+		if (connection->agreed_pro_version < 96 ?
 		    (device->ldev->md.uuid[UI_HISTORY_START + 1] & ~((u64)1)) ==
 		    (device->p_uuid[UI_HISTORY_START] & ~((u64)1)) :
 		    self + UUID_NEW_BM_OFFSET == (device->ldev->md.uuid[UI_BITMAP] & ~((u64)1))) {
 			/* The last P_SYNC_UUID did not get though. Undo the last start of
 			   resync as sync source modifications of our UUIDs. */
 
-			if (first_peer_device(device)->connection->agreed_pro_version < 91)
+			if (connection->agreed_pro_version < 91)
 				return -1091;
 
 			__drbd_uuid_set(device, UI_BITMAP, device->ldev->md.uuid[UI_HISTORY_START]);
@@ -3352,8 +3442,7 @@
  * return: NULL (alg name was "")
  *         ERR_PTR(error) if something goes wrong
  *         or the crypto hash ptr, if it worked out ok. */
-static
-struct crypto_hash *drbd_crypto_alloc_digest_safe(const struct drbd_device *device,
+static struct crypto_hash *drbd_crypto_alloc_digest_safe(const struct drbd_device *device,
 		const char *alg, const char *name)
 {
 	struct crypto_hash *tfm;
@@ -3639,7 +3728,7 @@
 	struct drbd_device *device;
 	struct p_sizes *p = pi->data;
 	enum determine_dev_size dd = DS_UNCHANGED;
-	sector_t p_size, p_usize, my_usize;
+	sector_t p_size, p_usize, p_csize, my_usize;
 	int ldsc = 0; /* local disk size changed */
 	enum dds_flags ddsf;
 
@@ -3650,6 +3739,7 @@
 
 	p_size = be64_to_cpu(p->d_size);
 	p_usize = be64_to_cpu(p->u_size);
+	p_csize = be64_to_cpu(p->c_size);
 
 	/* just store the peer's disk size for now.
 	 * we still need to figure out whether we accept that. */
@@ -3710,7 +3800,6 @@
 	}
 
 	device->peer_max_bio_size = be32_to_cpu(p->max_bio_size);
-	drbd_reconsider_max_bio_size(device);
 	/* Leave drbd_reconsider_max_bio_size() before drbd_determine_dev_size().
 	   In case we cleared the QUEUE_FLAG_DISCARD from our queue in
 	   drbd_reconsider_max_bio_size(), we can be sure that after
@@ -3718,14 +3807,28 @@
 
 	ddsf = be16_to_cpu(p->dds_flags);
 	if (get_ldev(device)) {
+		drbd_reconsider_max_bio_size(device, device->ldev);
 		dd = drbd_determine_dev_size(device, ddsf, NULL);
 		put_ldev(device);
 		if (dd == DS_ERROR)
 			return -EIO;
 		drbd_md_sync(device);
 	} else {
-		/* I am diskless, need to accept the peer's size. */
-		drbd_set_my_capacity(device, p_size);
+		/*
+		 * I am diskless, need to accept the peer's *current* size.
+		 * I must NOT accept the peers backing disk size,
+		 * it may have been larger than mine all along...
+		 *
+		 * At this point, the peer knows more about my disk, or at
+		 * least about what we last agreed upon, than myself.
+		 * So if his c_size is less than his d_size, the most likely
+		 * reason is that *my* d_size was smaller last time we checked.
+		 *
+		 * However, if he sends a zero current size,
+		 * take his (user-capped or) backing disk size anyways.
+		 */
+		drbd_reconsider_max_bio_size(device, NULL);
+		drbd_set_my_capacity(device, p_csize ?: p_usize ?: p_size);
 	}
 
 	if (get_ldev(device)) {
@@ -4501,6 +4604,7 @@
 		struct data_cmd *cmd;
 
 		drbd_thread_current_set_cpu(&connection->receiver);
+		update_receiver_timing_details(connection, drbd_recv_header);
 		if (drbd_recv_header(connection, &pi))
 			goto err_out;
 
@@ -4519,12 +4623,14 @@
 		}
 
 		if (shs) {
+			update_receiver_timing_details(connection, drbd_recv_all_warn);
 			err = drbd_recv_all_warn(connection, pi.data, shs);
 			if (err)
 				goto err_out;
 			pi.size -= shs;
 		}
 
+		update_receiver_timing_details(connection, cmd->fn);
 		err = cmd->fn(connection, &pi);
 		if (err) {
 			drbd_err(connection, "error receiving %s, e: %d l: %d!\n",
diff --git a/drivers/block/drbd/drbd_req.c b/drivers/block/drbd/drbd_req.c
index 09803d0..c67717d 100644
--- a/drivers/block/drbd/drbd_req.c
+++ b/drivers/block/drbd/drbd_req.c
@@ -52,7 +52,7 @@
 static void _drbd_end_io_acct(struct drbd_device *device, struct drbd_request *req)
 {
 	int rw = bio_data_dir(req->master_bio);
-	unsigned long duration = jiffies - req->start_time;
+	unsigned long duration = jiffies - req->start_jif;
 	int cpu;
 	cpu = part_stat_lock();
 	part_stat_add(cpu, &device->vdisk->part0, ticks[rw], duration);
@@ -66,7 +66,7 @@
 {
 	struct drbd_request *req;
 
-	req = mempool_alloc(drbd_request_mempool, GFP_NOIO);
+	req = mempool_alloc(drbd_request_mempool, GFP_NOIO | __GFP_ZERO);
 	if (!req)
 		return NULL;
 
@@ -84,6 +84,8 @@
 
 	INIT_LIST_HEAD(&req->tl_requests);
 	INIT_LIST_HEAD(&req->w.list);
+	INIT_LIST_HEAD(&req->req_pending_master_completion);
+	INIT_LIST_HEAD(&req->req_pending_local);
 
 	/* one reference to be put by __drbd_make_request */
 	atomic_set(&req->completion_ref, 1);
@@ -92,6 +94,19 @@
 	return req;
 }
 
+static void drbd_remove_request_interval(struct rb_root *root,
+					 struct drbd_request *req)
+{
+	struct drbd_device *device = req->device;
+	struct drbd_interval *i = &req->i;
+
+	drbd_remove_interval(root, i);
+
+	/* Wake up any processes waiting for this request to complete.  */
+	if (i->waiting)
+		wake_up(&device->misc_wait);
+}
+
 void drbd_req_destroy(struct kref *kref)
 {
 	struct drbd_request *req = container_of(kref, struct drbd_request, kref);
@@ -107,14 +122,30 @@
 		return;
 	}
 
-	/* remove it from the transfer log.
-	 * well, only if it had been there in the first
-	 * place... if it had not (local only or conflicting
-	 * and never sent), it should still be "empty" as
-	 * initialized in drbd_req_new(), so we can list_del() it
-	 * here unconditionally */
+	/* If called from mod_rq_state (expected normal case) or
+	 * drbd_send_and_submit (the less likely normal path), this holds the
+	 * req_lock, and req->tl_requests will typicaly be on ->transfer_log,
+	 * though it may be still empty (never added to the transfer log).
+	 *
+	 * If called from do_retry(), we do NOT hold the req_lock, but we are
+	 * still allowed to unconditionally list_del(&req->tl_requests),
+	 * because it will be on a local on-stack list only. */
 	list_del_init(&req->tl_requests);
 
+	/* finally remove the request from the conflict detection
+	 * respective block_id verification interval tree. */
+	if (!drbd_interval_empty(&req->i)) {
+		struct rb_root *root;
+
+		if (s & RQ_WRITE)
+			root = &device->write_requests;
+		else
+			root = &device->read_requests;
+		drbd_remove_request_interval(root, req);
+	} else if (s & (RQ_NET_MASK & ~RQ_NET_DONE) && req->i.size != 0)
+		drbd_err(device, "drbd_req_destroy: Logic BUG: interval empty, but: rq_state=0x%x, sect=%llu, size=%u\n",
+			s, (unsigned long long)req->i.sector, req->i.size);
+
 	/* if it was a write, we may have to set the corresponding
 	 * bit(s) out-of-sync first. If it had a local part, we need to
 	 * release the reference to the activity log. */
@@ -188,19 +219,6 @@
 }
 
 
-static void drbd_remove_request_interval(struct rb_root *root,
-					 struct drbd_request *req)
-{
-	struct drbd_device *device = req->device;
-	struct drbd_interval *i = &req->i;
-
-	drbd_remove_interval(root, i);
-
-	/* Wake up any processes waiting for this request to complete.  */
-	if (i->waiting)
-		wake_up(&device->misc_wait);
-}
-
 /* Helper for __req_mod().
  * Set m->bio to the master bio, if it is fit to be completed,
  * or leave it alone (it is initialized to NULL in __req_mod),
@@ -254,18 +272,6 @@
 	ok = (s & RQ_LOCAL_OK) || (s & RQ_NET_OK);
 	error = PTR_ERR(req->private_bio);
 
-	/* remove the request from the conflict detection
-	 * respective block_id verification hash */
-	if (!drbd_interval_empty(&req->i)) {
-		struct rb_root *root;
-
-		if (rw == WRITE)
-			root = &device->write_requests;
-		else
-			root = &device->read_requests;
-		drbd_remove_request_interval(root, req);
-	}
-
 	/* Before we can signal completion to the upper layers,
 	 * we may need to close the current transfer log epoch.
 	 * We are within the request lock, so we can simply compare
@@ -301,9 +307,24 @@
 		m->error = ok ? 0 : (error ?: -EIO);
 		m->bio = req->master_bio;
 		req->master_bio = NULL;
+		/* We leave it in the tree, to be able to verify later
+		 * write-acks in protocol != C during resync.
+		 * But we mark it as "complete", so it won't be counted as
+		 * conflict in a multi-primary setup. */
+		req->i.completed = true;
 	}
+
+	if (req->i.waiting)
+		wake_up(&device->misc_wait);
+
+	/* Either we are about to complete to upper layers,
+	 * or we will restart this request.
+	 * In either case, the request object will be destroyed soon,
+	 * so better remove it from all lists. */
+	list_del_init(&req->req_pending_master_completion);
 }
 
+/* still holds resource->req_lock */
 static int drbd_req_put_completion_ref(struct drbd_request *req, struct bio_and_error *m, int put)
 {
 	struct drbd_device *device = req->device;
@@ -324,12 +345,91 @@
 	return 1;
 }
 
+static void set_if_null_req_next(struct drbd_peer_device *peer_device, struct drbd_request *req)
+{
+	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
+	if (!connection)
+		return;
+	if (connection->req_next == NULL)
+		connection->req_next = req;
+}
+
+static void advance_conn_req_next(struct drbd_peer_device *peer_device, struct drbd_request *req)
+{
+	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
+	if (!connection)
+		return;
+	if (connection->req_next != req)
+		return;
+	list_for_each_entry_continue(req, &connection->transfer_log, tl_requests) {
+		const unsigned s = req->rq_state;
+		if (s & RQ_NET_QUEUED)
+			break;
+	}
+	if (&req->tl_requests == &connection->transfer_log)
+		req = NULL;
+	connection->req_next = req;
+}
+
+static void set_if_null_req_ack_pending(struct drbd_peer_device *peer_device, struct drbd_request *req)
+{
+	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
+	if (!connection)
+		return;
+	if (connection->req_ack_pending == NULL)
+		connection->req_ack_pending = req;
+}
+
+static void advance_conn_req_ack_pending(struct drbd_peer_device *peer_device, struct drbd_request *req)
+{
+	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
+	if (!connection)
+		return;
+	if (connection->req_ack_pending != req)
+		return;
+	list_for_each_entry_continue(req, &connection->transfer_log, tl_requests) {
+		const unsigned s = req->rq_state;
+		if ((s & RQ_NET_SENT) && (s & RQ_NET_PENDING))
+			break;
+	}
+	if (&req->tl_requests == &connection->transfer_log)
+		req = NULL;
+	connection->req_ack_pending = req;
+}
+
+static void set_if_null_req_not_net_done(struct drbd_peer_device *peer_device, struct drbd_request *req)
+{
+	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
+	if (!connection)
+		return;
+	if (connection->req_not_net_done == NULL)
+		connection->req_not_net_done = req;
+}
+
+static void advance_conn_req_not_net_done(struct drbd_peer_device *peer_device, struct drbd_request *req)
+{
+	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
+	if (!connection)
+		return;
+	if (connection->req_not_net_done != req)
+		return;
+	list_for_each_entry_continue(req, &connection->transfer_log, tl_requests) {
+		const unsigned s = req->rq_state;
+		if ((s & RQ_NET_SENT) && !(s & RQ_NET_DONE))
+			break;
+	}
+	if (&req->tl_requests == &connection->transfer_log)
+		req = NULL;
+	connection->req_not_net_done = req;
+}
+
 /* I'd like this to be the only place that manipulates
  * req->completion_ref and req->kref. */
 static void mod_rq_state(struct drbd_request *req, struct bio_and_error *m,
 		int clear, int set)
 {
 	struct drbd_device *device = req->device;
+	struct drbd_peer_device *peer_device = first_peer_device(device);
 	unsigned s = req->rq_state;
 	int c_put = 0;
 	int k_put = 0;
@@ -356,14 +456,23 @@
 		atomic_inc(&req->completion_ref);
 	}
 
-	if (!(s & RQ_NET_QUEUED) && (set & RQ_NET_QUEUED))
+	if (!(s & RQ_NET_QUEUED) && (set & RQ_NET_QUEUED)) {
 		atomic_inc(&req->completion_ref);
+		set_if_null_req_next(peer_device, req);
+	}
 
 	if (!(s & RQ_EXP_BARR_ACK) && (set & RQ_EXP_BARR_ACK))
 		kref_get(&req->kref); /* wait for the DONE */
 
-	if (!(s & RQ_NET_SENT) && (set & RQ_NET_SENT))
-		atomic_add(req->i.size >> 9, &device->ap_in_flight);
+	if (!(s & RQ_NET_SENT) && (set & RQ_NET_SENT)) {
+		/* potentially already completed in the asender thread */
+		if (!(s & RQ_NET_DONE)) {
+			atomic_add(req->i.size >> 9, &device->ap_in_flight);
+			set_if_null_req_not_net_done(peer_device, req);
+		}
+		if (s & RQ_NET_PENDING)
+			set_if_null_req_ack_pending(peer_device, req);
+	}
 
 	if (!(s & RQ_COMPLETION_SUSP) && (set & RQ_COMPLETION_SUSP))
 		atomic_inc(&req->completion_ref);
@@ -386,20 +495,34 @@
 			++k_put;
 		else
 			++c_put;
+		list_del_init(&req->req_pending_local);
 	}
 
 	if ((s & RQ_NET_PENDING) && (clear & RQ_NET_PENDING)) {
 		dec_ap_pending(device);
 		++c_put;
+		req->acked_jif = jiffies;
+		advance_conn_req_ack_pending(peer_device, req);
 	}
 
-	if ((s & RQ_NET_QUEUED) && (clear & RQ_NET_QUEUED))
+	if ((s & RQ_NET_QUEUED) && (clear & RQ_NET_QUEUED)) {
 		++c_put;
+		advance_conn_req_next(peer_device, req);
+	}
 
-	if ((s & RQ_EXP_BARR_ACK) && !(s & RQ_NET_DONE) && (set & RQ_NET_DONE)) {
-		if (req->rq_state & RQ_NET_SENT)
+	if (!(s & RQ_NET_DONE) && (set & RQ_NET_DONE)) {
+		if (s & RQ_NET_SENT)
 			atomic_sub(req->i.size >> 9, &device->ap_in_flight);
-		++k_put;
+		if (s & RQ_EXP_BARR_ACK)
+			++k_put;
+		req->net_done_jif = jiffies;
+
+		/* in ahead/behind mode, or just in case,
+		 * before we finally destroy this request,
+		 * the caching pointers must not reference it anymore */
+		advance_conn_req_next(peer_device, req);
+		advance_conn_req_ack_pending(peer_device, req);
+		advance_conn_req_not_net_done(peer_device, req);
 	}
 
 	/* potentially complete and destroy */
@@ -439,6 +562,19 @@
 			bdevname(device->ldev->backing_bdev, b));
 }
 
+/* Helper for HANDED_OVER_TO_NETWORK.
+ * Is this a protocol A write (neither WRITE_ACK nor RECEIVE_ACK expected)?
+ * Is it also still "PENDING"?
+ * --> If so, clear PENDING and set NET_OK below.
+ * If it is a protocol A write, but not RQ_PENDING anymore, neg-ack was faster
+ * (and we must not set RQ_NET_OK) */
+static inline bool is_pending_write_protocol_A(struct drbd_request *req)
+{
+	return (req->rq_state &
+		   (RQ_WRITE|RQ_NET_PENDING|RQ_EXP_WRITE_ACK|RQ_EXP_RECEIVE_ACK))
+		== (RQ_WRITE|RQ_NET_PENDING);
+}
+
 /* obviously this could be coded as many single functions
  * instead of one huge switch,
  * or by putting the code directly in the respective locations
@@ -454,7 +590,9 @@
 int __req_mod(struct drbd_request *req, enum drbd_req_event what,
 		struct bio_and_error *m)
 {
-	struct drbd_device *device = req->device;
+	struct drbd_device *const device = req->device;
+	struct drbd_peer_device *const peer_device = first_peer_device(device);
+	struct drbd_connection *const connection = peer_device ? peer_device->connection : NULL;
 	struct net_conf *nc;
 	int p, rv = 0;
 
@@ -477,7 +615,7 @@
 		 * and from w_read_retry_remote */
 		D_ASSERT(device, !(req->rq_state & RQ_NET_MASK));
 		rcu_read_lock();
-		nc = rcu_dereference(first_peer_device(device)->connection->net_conf);
+		nc = rcu_dereference(connection->net_conf);
 		p = nc->wire_protocol;
 		rcu_read_unlock();
 		req->rq_state |=
@@ -549,7 +687,7 @@
 		D_ASSERT(device, (req->rq_state & RQ_LOCAL_MASK) == 0);
 		mod_rq_state(req, m, 0, RQ_NET_QUEUED);
 		req->w.cb = w_send_read_req;
-		drbd_queue_work(&first_peer_device(device)->connection->sender_work,
+		drbd_queue_work(&connection->sender_work,
 				&req->w);
 		break;
 
@@ -585,23 +723,23 @@
 		D_ASSERT(device, req->rq_state & RQ_NET_PENDING);
 		mod_rq_state(req, m, 0, RQ_NET_QUEUED|RQ_EXP_BARR_ACK);
 		req->w.cb =  w_send_dblock;
-		drbd_queue_work(&first_peer_device(device)->connection->sender_work,
+		drbd_queue_work(&connection->sender_work,
 				&req->w);
 
 		/* close the epoch, in case it outgrew the limit */
 		rcu_read_lock();
-		nc = rcu_dereference(first_peer_device(device)->connection->net_conf);
+		nc = rcu_dereference(connection->net_conf);
 		p = nc->max_epoch_size;
 		rcu_read_unlock();
-		if (first_peer_device(device)->connection->current_tle_writes >= p)
-			start_new_tl_epoch(first_peer_device(device)->connection);
+		if (connection->current_tle_writes >= p)
+			start_new_tl_epoch(connection);
 
 		break;
 
 	case QUEUE_FOR_SEND_OOS:
 		mod_rq_state(req, m, 0, RQ_NET_QUEUED);
 		req->w.cb =  w_send_out_of_sync;
-		drbd_queue_work(&first_peer_device(device)->connection->sender_work,
+		drbd_queue_work(&connection->sender_work,
 				&req->w);
 		break;
 
@@ -615,18 +753,16 @@
 
 	case HANDED_OVER_TO_NETWORK:
 		/* assert something? */
-		if (bio_data_dir(req->master_bio) == WRITE &&
-		    !(req->rq_state & (RQ_EXP_RECEIVE_ACK | RQ_EXP_WRITE_ACK))) {
+		if (is_pending_write_protocol_A(req))
 			/* this is what is dangerous about protocol A:
 			 * pretend it was successfully written on the peer. */
-			if (req->rq_state & RQ_NET_PENDING)
-				mod_rq_state(req, m, RQ_NET_PENDING, RQ_NET_OK);
-			/* else: neg-ack was faster... */
-			/* it is still not yet RQ_NET_DONE until the
-			 * corresponding epoch barrier got acked as well,
-			 * so we know what to dirty on connection loss */
-		}
-		mod_rq_state(req, m, RQ_NET_QUEUED, RQ_NET_SENT);
+			mod_rq_state(req, m, RQ_NET_QUEUED|RQ_NET_PENDING,
+						RQ_NET_SENT|RQ_NET_OK);
+		else
+			mod_rq_state(req, m, RQ_NET_QUEUED, RQ_NET_SENT);
+		/* It is still not yet RQ_NET_DONE until the
+		 * corresponding epoch barrier got acked as well,
+		 * so we know what to dirty on connection loss. */
 		break;
 
 	case OOS_HANDED_TO_NETWORK:
@@ -658,12 +794,13 @@
 	case WRITE_ACKED_BY_PEER_AND_SIS:
 		req->rq_state |= RQ_NET_SIS;
 	case WRITE_ACKED_BY_PEER:
-		D_ASSERT(device, req->rq_state & RQ_EXP_WRITE_ACK);
-		/* protocol C; successfully written on peer.
+		/* Normal operation protocol C: successfully written on peer.
+		 * During resync, even in protocol != C,
+		 * we requested an explicit write ack anyways.
+		 * Which means we cannot even assert anything here.
 		 * Nothing more to do here.
 		 * We want to keep the tl in place for all protocols, to cater
 		 * for volatile write-back caches on lower level devices. */
-
 		goto ack_common;
 	case RECV_ACKED_BY_PEER:
 		D_ASSERT(device, req->rq_state & RQ_EXP_RECEIVE_ACK);
@@ -671,7 +808,6 @@
 		 * see also notes above in HANDED_OVER_TO_NETWORK about
 		 * protocol != C */
 	ack_common:
-		D_ASSERT(device, req->rq_state & RQ_NET_PENDING);
 		mod_rq_state(req, m, RQ_NET_PENDING, RQ_NET_OK);
 		break;
 
@@ -714,7 +850,7 @@
 
 		get_ldev(device); /* always succeeds in this call path */
 		req->w.cb = w_restart_disk_io;
-		drbd_queue_work(&first_peer_device(device)->connection->sender_work,
+		drbd_queue_work(&connection->sender_work,
 				&req->w);
 		break;
 
@@ -736,7 +872,8 @@
 
 			mod_rq_state(req, m, RQ_COMPLETION_SUSP, RQ_NET_QUEUED|RQ_NET_PENDING);
 			if (req->w.cb) {
-				drbd_queue_work(&first_peer_device(device)->connection->sender_work,
+				/* w.cb expected to be w_send_dblock, or w_send_read_req */
+				drbd_queue_work(&connection->sender_work,
 						&req->w);
 				rv = req->rq_state & RQ_WRITE ? MR_WRITE : MR_READ;
 			} /* else: FIXME can this happen? */
@@ -769,7 +906,7 @@
 		break;
 
 	case QUEUE_AS_DRBD_BARRIER:
-		start_new_tl_epoch(first_peer_device(device)->connection);
+		start_new_tl_epoch(connection);
 		mod_rq_state(req, m, 0, RQ_NET_OK|RQ_NET_DONE);
 		break;
 	};
@@ -886,6 +1023,9 @@
 	    connection->agreed_pro_version < 96)
 		return;
 
+	if (on_congestion == OC_PULL_AHEAD && device->state.conn == C_AHEAD)
+		return; /* nothing to do ... */
+
 	/* If I don't even have good local storage, we can not reasonably try
 	 * to pull ahead of the peer. We also need the local reference to make
 	 * sure device->act_log is there.
@@ -1021,6 +1161,7 @@
 	 * stable storage, and this is a WRITE, we may not even submit
 	 * this bio. */
 	if (get_ldev(device)) {
+		req->pre_submit_jif = jiffies;
 		if (drbd_insert_fault(device,
 				      rw == WRITE ? DRBD_FAULT_DT_WR
 				    : rw == READ  ? DRBD_FAULT_DT_RD
@@ -1035,10 +1176,14 @@
 
 static void drbd_queue_write(struct drbd_device *device, struct drbd_request *req)
 {
-	spin_lock(&device->submit.lock);
+	spin_lock_irq(&device->resource->req_lock);
 	list_add_tail(&req->tl_requests, &device->submit.writes);
-	spin_unlock(&device->submit.lock);
+	list_add_tail(&req->req_pending_master_completion,
+			&device->pending_master_completion[1 /* WRITE */]);
+	spin_unlock_irq(&device->resource->req_lock);
 	queue_work(device->submit.wq, &device->submit.worker);
+	/* do_submit() may sleep internally on al_wait, too */
+	wake_up(&device->al_wait);
 }
 
 /* returns the new drbd_request pointer, if the caller is expected to
@@ -1047,7 +1192,7 @@
  * Returns ERR_PTR(-ENOMEM) if we cannot allocate a drbd_request.
  */
 static struct drbd_request *
-drbd_request_prepare(struct drbd_device *device, struct bio *bio, unsigned long start_time)
+drbd_request_prepare(struct drbd_device *device, struct bio *bio, unsigned long start_jif)
 {
 	const int rw = bio_data_dir(bio);
 	struct drbd_request *req;
@@ -1062,7 +1207,7 @@
 		bio_endio(bio, -ENOMEM);
 		return ERR_PTR(-ENOMEM);
 	}
-	req->start_time = start_time;
+	req->start_jif = start_jif;
 
 	if (!get_ldev(device)) {
 		bio_put(req->private_bio);
@@ -1075,10 +1220,12 @@
 	if (rw == WRITE && req->private_bio && req->i.size
 	&& !test_bit(AL_SUSPENDED, &device->flags)) {
 		if (!drbd_al_begin_io_fastpath(device, &req->i)) {
+			atomic_inc(&device->ap_actlog_cnt);
 			drbd_queue_write(device, req);
 			return NULL;
 		}
 		req->rq_state |= RQ_IN_ACT_LOG;
+		req->in_actlog_jif = jiffies;
 	}
 
 	return req;
@@ -1086,11 +1233,13 @@
 
 static void drbd_send_and_submit(struct drbd_device *device, struct drbd_request *req)
 {
+	struct drbd_resource *resource = device->resource;
 	const int rw = bio_rw(req->master_bio);
 	struct bio_and_error m = { NULL, };
 	bool no_remote = false;
+	bool submit_private_bio = false;
 
-	spin_lock_irq(&device->resource->req_lock);
+	spin_lock_irq(&resource->req_lock);
 	if (rw == WRITE) {
 		/* This may temporarily give up the req_lock,
 		 * but will re-aquire it before it returns here.
@@ -1148,13 +1297,18 @@
 			no_remote = true;
 	}
 
+	/* If it took the fast path in drbd_request_prepare, add it here.
+	 * The slow path has added it already. */
+	if (list_empty(&req->req_pending_master_completion))
+		list_add_tail(&req->req_pending_master_completion,
+			&device->pending_master_completion[rw == WRITE]);
 	if (req->private_bio) {
 		/* needs to be marked within the same spinlock */
+		list_add_tail(&req->req_pending_local,
+			&device->pending_completion[rw == WRITE]);
 		_req_mod(req, TO_BE_SUBMITTED);
 		/* but we need to give up the spinlock to submit */
-		spin_unlock_irq(&device->resource->req_lock);
-		drbd_submit_req_private_bio(req);
-		spin_lock_irq(&device->resource->req_lock);
+		submit_private_bio = true;
 	} else if (no_remote) {
 nodata:
 		if (__ratelimit(&drbd_ratelimit_state))
@@ -1167,15 +1321,23 @@
 out:
 	if (drbd_req_put_completion_ref(req, &m, 1))
 		kref_put(&req->kref, drbd_req_destroy);
-	spin_unlock_irq(&device->resource->req_lock);
+	spin_unlock_irq(&resource->req_lock);
 
+	/* Even though above is a kref_put(), this is safe.
+	 * As long as we still need to submit our private bio,
+	 * we hold a completion ref, and the request cannot disappear.
+	 * If however this request did not even have a private bio to submit
+	 * (e.g. remote read), req may already be invalid now.
+	 * That's why we cannot check on req->private_bio. */
+	if (submit_private_bio)
+		drbd_submit_req_private_bio(req);
 	if (m.bio)
 		complete_master_bio(device, &m);
 }
 
-void __drbd_make_request(struct drbd_device *device, struct bio *bio, unsigned long start_time)
+void __drbd_make_request(struct drbd_device *device, struct bio *bio, unsigned long start_jif)
 {
-	struct drbd_request *req = drbd_request_prepare(device, bio, start_time);
+	struct drbd_request *req = drbd_request_prepare(device, bio, start_jif);
 	if (IS_ERR_OR_NULL(req))
 		return;
 	drbd_send_and_submit(device, req);
@@ -1194,6 +1356,8 @@
 				continue;
 
 			req->rq_state |= RQ_IN_ACT_LOG;
+			req->in_actlog_jif = jiffies;
+			atomic_dec(&device->ap_actlog_cnt);
 		}
 
 		list_del_init(&req->tl_requests);
@@ -1203,7 +1367,8 @@
 
 static bool prepare_al_transaction_nonblock(struct drbd_device *device,
 					    struct list_head *incoming,
-					    struct list_head *pending)
+					    struct list_head *pending,
+					    struct list_head *later)
 {
 	struct drbd_request *req, *tmp;
 	int wake = 0;
@@ -1212,45 +1377,105 @@
 	spin_lock_irq(&device->al_lock);
 	list_for_each_entry_safe(req, tmp, incoming, tl_requests) {
 		err = drbd_al_begin_io_nonblock(device, &req->i);
+		if (err == -ENOBUFS)
+			break;
 		if (err == -EBUSY)
 			wake = 1;
 		if (err)
-			continue;
-		req->rq_state |= RQ_IN_ACT_LOG;
-		list_move_tail(&req->tl_requests, pending);
+			list_move_tail(&req->tl_requests, later);
+		else
+			list_move_tail(&req->tl_requests, pending);
 	}
 	spin_unlock_irq(&device->al_lock);
 	if (wake)
 		wake_up(&device->al_wait);
-
 	return !list_empty(pending);
 }
 
+void send_and_submit_pending(struct drbd_device *device, struct list_head *pending)
+{
+	struct drbd_request *req, *tmp;
+
+	list_for_each_entry_safe(req, tmp, pending, tl_requests) {
+		req->rq_state |= RQ_IN_ACT_LOG;
+		req->in_actlog_jif = jiffies;
+		atomic_dec(&device->ap_actlog_cnt);
+		list_del_init(&req->tl_requests);
+		drbd_send_and_submit(device, req);
+	}
+}
+
 void do_submit(struct work_struct *ws)
 {
 	struct drbd_device *device = container_of(ws, struct drbd_device, submit.worker);
-	LIST_HEAD(incoming);
-	LIST_HEAD(pending);
-	struct drbd_request *req, *tmp;
+	LIST_HEAD(incoming);	/* from drbd_make_request() */
+	LIST_HEAD(pending);	/* to be submitted after next AL-transaction commit */
+	LIST_HEAD(busy);	/* blocked by resync requests */
+
+	/* grab new incoming requests */
+	spin_lock_irq(&device->resource->req_lock);
+	list_splice_tail_init(&device->submit.writes, &incoming);
+	spin_unlock_irq(&device->resource->req_lock);
 
 	for (;;) {
-		spin_lock(&device->submit.lock);
-		list_splice_tail_init(&device->submit.writes, &incoming);
-		spin_unlock(&device->submit.lock);
+		DEFINE_WAIT(wait);
 
+		/* move used-to-be-busy back to front of incoming */
+		list_splice_init(&busy, &incoming);
 		submit_fast_path(device, &incoming);
 		if (list_empty(&incoming))
 			break;
 
-skip_fast_path:
-		wait_event(device->al_wait, prepare_al_transaction_nonblock(device, &incoming, &pending));
-		/* Maybe more was queued, while we prepared the transaction?
-		 * Try to stuff them into this transaction as well.
-		 * Be strictly non-blocking here, no wait_event, we already
-		 * have something to commit.
-		 * Stop if we don't make any more progres.
-		 */
 		for (;;) {
+			prepare_to_wait(&device->al_wait, &wait, TASK_UNINTERRUPTIBLE);
+
+			list_splice_init(&busy, &incoming);
+			prepare_al_transaction_nonblock(device, &incoming, &pending, &busy);
+			if (!list_empty(&pending))
+				break;
+
+			schedule();
+
+			/* If all currently "hot" activity log extents are kept busy by
+			 * incoming requests, we still must not totally starve new
+			 * requests to "cold" extents.
+			 * Something left on &incoming means there had not been
+			 * enough update slots available, and the activity log
+			 * has been marked as "starving".
+			 *
+			 * Try again now, without looking for new requests,
+			 * effectively blocking all new requests until we made
+			 * at least _some_ progress with what we currently have.
+			 */
+			if (!list_empty(&incoming))
+				continue;
+
+			/* Nothing moved to pending, but nothing left
+			 * on incoming: all moved to busy!
+			 * Grab new and iterate. */
+			spin_lock_irq(&device->resource->req_lock);
+			list_splice_tail_init(&device->submit.writes, &incoming);
+			spin_unlock_irq(&device->resource->req_lock);
+		}
+		finish_wait(&device->al_wait, &wait);
+
+		/* If the transaction was full, before all incoming requests
+		 * had been processed, skip ahead to commit, and iterate
+		 * without splicing in more incoming requests from upper layers.
+		 *
+		 * Else, if all incoming have been processed,
+		 * they have become either "pending" (to be submitted after
+		 * next transaction commit) or "busy" (blocked by resync).
+		 *
+		 * Maybe more was queued, while we prepared the transaction?
+		 * Try to stuff those into this transaction as well.
+		 * Be strictly non-blocking here,
+		 * we already have something to commit.
+		 *
+		 * Commit if we don't make any more progres.
+		 */
+
+		while (list_empty(&incoming)) {
 			LIST_HEAD(more_pending);
 			LIST_HEAD(more_incoming);
 			bool made_progress;
@@ -1260,55 +1485,32 @@
 			if (list_empty(&device->submit.writes))
 				break;
 
-			spin_lock(&device->submit.lock);
+			spin_lock_irq(&device->resource->req_lock);
 			list_splice_tail_init(&device->submit.writes, &more_incoming);
-			spin_unlock(&device->submit.lock);
+			spin_unlock_irq(&device->resource->req_lock);
 
 			if (list_empty(&more_incoming))
 				break;
 
-			made_progress = prepare_al_transaction_nonblock(device, &more_incoming, &more_pending);
+			made_progress = prepare_al_transaction_nonblock(device, &more_incoming, &more_pending, &busy);
 
 			list_splice_tail_init(&more_pending, &pending);
 			list_splice_tail_init(&more_incoming, &incoming);
-
 			if (!made_progress)
 				break;
 		}
-		drbd_al_begin_io_commit(device, false);
 
-		list_for_each_entry_safe(req, tmp, &pending, tl_requests) {
-			list_del_init(&req->tl_requests);
-			drbd_send_and_submit(device, req);
-		}
-
-		/* If all currently hot activity log extents are kept busy by
-		 * incoming requests, we still must not totally starve new
-		 * requests to cold extents. In that case, prepare one request
-		 * in blocking mode. */
-		list_for_each_entry_safe(req, tmp, &incoming, tl_requests) {
-			list_del_init(&req->tl_requests);
-			req->rq_state |= RQ_IN_ACT_LOG;
-			if (!drbd_al_begin_io_prepare(device, &req->i)) {
-				/* Corresponding extent was hot after all? */
-				drbd_send_and_submit(device, req);
-			} else {
-				/* Found a request to a cold extent.
-				 * Put on "pending" list,
-				 * and try to cumulate with more. */
-				list_add(&req->tl_requests, &pending);
-				goto skip_fast_path;
-			}
-		}
+		drbd_al_begin_io_commit(device);
+		send_and_submit_pending(device, &pending);
 	}
 }
 
 void drbd_make_request(struct request_queue *q, struct bio *bio)
 {
 	struct drbd_device *device = (struct drbd_device *) q->queuedata;
-	unsigned long start_time;
+	unsigned long start_jif;
 
-	start_time = jiffies;
+	start_jif = jiffies;
 
 	/*
 	 * what we "blindly" assume:
@@ -1316,7 +1518,7 @@
 	D_ASSERT(device, IS_ALIGNED(bio->bi_iter.bi_size, 512));
 
 	inc_ap_bio(device);
-	__drbd_make_request(device, bio, start_time);
+	__drbd_make_request(device, bio, start_jif);
 }
 
 /* This is called by bio_add_page().
@@ -1353,36 +1555,13 @@
 	return limit;
 }
 
-static void find_oldest_requests(
-		struct drbd_connection *connection,
-		struct drbd_device *device,
-		struct drbd_request **oldest_req_waiting_for_peer,
-		struct drbd_request **oldest_req_waiting_for_disk)
-{
-	struct drbd_request *r;
-	*oldest_req_waiting_for_peer = NULL;
-	*oldest_req_waiting_for_disk = NULL;
-	list_for_each_entry(r, &connection->transfer_log, tl_requests) {
-		const unsigned s = r->rq_state;
-		if (!*oldest_req_waiting_for_peer
-		&& ((s & RQ_NET_MASK) && !(s & RQ_NET_DONE)))
-			*oldest_req_waiting_for_peer = r;
-
-		if (!*oldest_req_waiting_for_disk
-		&& (s & RQ_LOCAL_PENDING) && r->device == device)
-			*oldest_req_waiting_for_disk = r;
-
-		if (*oldest_req_waiting_for_peer && *oldest_req_waiting_for_disk)
-			break;
-	}
-}
-
 void request_timer_fn(unsigned long data)
 {
 	struct drbd_device *device = (struct drbd_device *) data;
 	struct drbd_connection *connection = first_peer_device(device)->connection;
-	struct drbd_request *req_disk, *req_peer; /* oldest request */
+	struct drbd_request *req_read, *req_write, *req_peer; /* oldest request */
 	struct net_conf *nc;
+	unsigned long oldest_submit_jif;
 	unsigned long ent = 0, dt = 0, et, nt; /* effective timeout = ko_count * timeout */
 	unsigned long now;
 
@@ -1403,14 +1582,31 @@
 		return; /* Recurring timer stopped */
 
 	now = jiffies;
+	nt = now + et;
 
 	spin_lock_irq(&device->resource->req_lock);
-	find_oldest_requests(connection, device, &req_peer, &req_disk);
-	if (req_peer == NULL && req_disk == NULL) {
-		spin_unlock_irq(&device->resource->req_lock);
-		mod_timer(&device->request_timer, now + et);
-		return;
-	}
+	req_read = list_first_entry_or_null(&device->pending_completion[0], struct drbd_request, req_pending_local);
+	req_write = list_first_entry_or_null(&device->pending_completion[1], struct drbd_request, req_pending_local);
+	req_peer = connection->req_not_net_done;
+	/* maybe the oldest request waiting for the peer is in fact still
+	 * blocking in tcp sendmsg */
+	if (!req_peer && connection->req_next && connection->req_next->pre_send_jif)
+		req_peer = connection->req_next;
+
+	/* evaluate the oldest peer request only in one timer! */
+	if (req_peer && req_peer->device != device)
+		req_peer = NULL;
+
+	/* do we have something to evaluate? */
+	if (req_peer == NULL && req_write == NULL && req_read == NULL)
+		goto out;
+
+	oldest_submit_jif =
+		(req_write && req_read)
+		? ( time_before(req_write->pre_submit_jif, req_read->pre_submit_jif)
+		  ? req_write->pre_submit_jif : req_read->pre_submit_jif )
+		: req_write ? req_write->pre_submit_jif
+		: req_read ? req_read->pre_submit_jif : now;
 
 	/* The request is considered timed out, if
 	 * - we have some effective timeout from the configuration,
@@ -1429,13 +1625,13 @@
 	 * to expire twice (worst case) to become effective. Good enough.
 	 */
 	if (ent && req_peer &&
-		 time_after(now, req_peer->start_time + ent) &&
+		 time_after(now, req_peer->pre_send_jif + ent) &&
 		!time_in_range(now, connection->last_reconnect_jif, connection->last_reconnect_jif + ent)) {
 		drbd_warn(device, "Remote failed to finish a request within ko-count * timeout\n");
 		_drbd_set_state(_NS(device, conn, C_TIMEOUT), CS_VERBOSE | CS_HARD, NULL);
 	}
-	if (dt && req_disk &&
-		 time_after(now, req_disk->start_time + dt) &&
+	if (dt && oldest_submit_jif != now &&
+		 time_after(now, oldest_submit_jif + dt) &&
 		!time_in_range(now, device->last_reattach_jif, device->last_reattach_jif + dt)) {
 		drbd_warn(device, "Local backing device failed to meet the disk-timeout\n");
 		__drbd_chk_io_error(device, DRBD_FORCE_DETACH);
@@ -1443,11 +1639,12 @@
 
 	/* Reschedule timer for the nearest not already expired timeout.
 	 * Fallback to now + min(effective network timeout, disk timeout). */
-	ent = (ent && req_peer && time_before(now, req_peer->start_time + ent))
-		? req_peer->start_time + ent : now + et;
-	dt = (dt && req_disk && time_before(now, req_disk->start_time + dt))
-		? req_disk->start_time + dt : now + et;
+	ent = (ent && req_peer && time_before(now, req_peer->pre_send_jif + ent))
+		? req_peer->pre_send_jif + ent : now + et;
+	dt = (dt && oldest_submit_jif != now && time_before(now, oldest_submit_jif + dt))
+		? oldest_submit_jif + dt : now + et;
 	nt = time_before(ent, dt) ? ent : dt;
+out:
 	spin_unlock_irq(&connection->resource->req_lock);
 	mod_timer(&device->request_timer, nt);
 }
diff --git a/drivers/block/drbd/drbd_req.h b/drivers/block/drbd/drbd_req.h
index 8566cd5..9f6a040 100644
--- a/drivers/block/drbd/drbd_req.h
+++ b/drivers/block/drbd/drbd_req.h
@@ -288,6 +288,7 @@
 extern void request_timer_fn(unsigned long data);
 extern void tl_restart(struct drbd_connection *connection, enum drbd_req_event what);
 extern void _tl_restart(struct drbd_connection *connection, enum drbd_req_event what);
+extern void tl_abort_disk_io(struct drbd_device *device);
 
 /* this is in drbd_main.c */
 extern void drbd_restart_request(struct drbd_request *req);
diff --git a/drivers/block/drbd/drbd_state.c b/drivers/block/drbd/drbd_state.c
index a5d8aae..c35c0f0 100644
--- a/drivers/block/drbd/drbd_state.c
+++ b/drivers/block/drbd/drbd_state.c
@@ -410,7 +410,7 @@
 	return rv;
 }
 
-static void print_st(struct drbd_device *device, char *name, union drbd_state ns)
+static void print_st(struct drbd_device *device, const char *name, union drbd_state ns)
 {
 	drbd_err(device, " %s = { cs:%s ro:%s/%s ds:%s/%s %c%c%c%c%c%c }\n",
 	    name,
@@ -952,11 +952,12 @@
 __drbd_set_state(struct drbd_device *device, union drbd_state ns,
 	         enum chg_state_flags flags, struct completion *done)
 {
+	struct drbd_peer_device *peer_device = first_peer_device(device);
+	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
 	union drbd_state os;
 	enum drbd_state_rv rv = SS_SUCCESS;
 	enum sanitize_state_warnings ssw;
 	struct after_state_chg_work *ascw;
-	bool did_remote, should_do_remote;
 
 	os = drbd_read_state(device);
 
@@ -978,9 +979,9 @@
 			   this happen...*/
 
 			if (is_valid_state(device, os) == rv)
-				rv = is_valid_soft_transition(os, ns, first_peer_device(device)->connection);
+				rv = is_valid_soft_transition(os, ns, connection);
 		} else
-			rv = is_valid_soft_transition(os, ns, first_peer_device(device)->connection);
+			rv = is_valid_soft_transition(os, ns, connection);
 	}
 
 	if (rv < SS_SUCCESS) {
@@ -997,7 +998,7 @@
 	   sanitize_state(). Only display it here if we where not called from
 	   _conn_request_state() */
 	if (!(flags & CS_DC_SUSP))
-		conn_pr_state_change(first_peer_device(device)->connection, os, ns,
+		conn_pr_state_change(connection, os, ns,
 				     (flags & ~CS_DC_MASK) | CS_DC_SUSP);
 
 	/* if we are going -> D_FAILED or D_DISKLESS, grab one extra reference
@@ -1008,28 +1009,35 @@
 	    (os.disk != D_DISKLESS && ns.disk == D_DISKLESS))
 		atomic_inc(&device->local_cnt);
 
-	did_remote = drbd_should_do_remote(device->state);
+	if (!is_sync_state(os.conn) && is_sync_state(ns.conn))
+		clear_bit(RS_DONE, &device->flags);
+
+	/* changes to local_cnt and device flags should be visible before
+	 * changes to state, which again should be visible before anything else
+	 * depending on that change happens. */
+	smp_wmb();
 	device->state.i = ns.i;
-	should_do_remote = drbd_should_do_remote(device->state);
 	device->resource->susp = ns.susp;
 	device->resource->susp_nod = ns.susp_nod;
 	device->resource->susp_fen = ns.susp_fen;
+	smp_wmb();
 
 	/* put replicated vs not-replicated requests in seperate epochs */
-	if (did_remote != should_do_remote)
-		start_new_tl_epoch(first_peer_device(device)->connection);
+	if (drbd_should_do_remote((union drbd_dev_state)os.i) !=
+	    drbd_should_do_remote((union drbd_dev_state)ns.i))
+		start_new_tl_epoch(connection);
 
 	if (os.disk == D_ATTACHING && ns.disk >= D_NEGOTIATING)
 		drbd_print_uuids(device, "attached to UUIDs");
 
 	/* Wake up role changes, that were delayed because of connection establishing */
 	if (os.conn == C_WF_REPORT_PARAMS && ns.conn != C_WF_REPORT_PARAMS &&
-	    no_peer_wf_report_params(first_peer_device(device)->connection))
-		clear_bit(STATE_SENT, &first_peer_device(device)->connection->flags);
+	    no_peer_wf_report_params(connection))
+		clear_bit(STATE_SENT, &connection->flags);
 
 	wake_up(&device->misc_wait);
 	wake_up(&device->state_wait);
-	wake_up(&first_peer_device(device)->connection->ping_wait);
+	wake_up(&connection->ping_wait);
 
 	/* Aborted verify run, or we reached the stop sector.
 	 * Log the last position, unless end-of-device. */
@@ -1118,21 +1126,21 @@
 
 	/* Receiver should clean up itself */
 	if (os.conn != C_DISCONNECTING && ns.conn == C_DISCONNECTING)
-		drbd_thread_stop_nowait(&first_peer_device(device)->connection->receiver);
+		drbd_thread_stop_nowait(&connection->receiver);
 
 	/* Now the receiver finished cleaning up itself, it should die */
 	if (os.conn != C_STANDALONE && ns.conn == C_STANDALONE)
-		drbd_thread_stop_nowait(&first_peer_device(device)->connection->receiver);
+		drbd_thread_stop_nowait(&connection->receiver);
 
 	/* Upon network failure, we need to restart the receiver. */
 	if (os.conn > C_WF_CONNECTION &&
 	    ns.conn <= C_TEAR_DOWN && ns.conn >= C_TIMEOUT)
-		drbd_thread_restart_nowait(&first_peer_device(device)->connection->receiver);
+		drbd_thread_restart_nowait(&connection->receiver);
 
 	/* Resume AL writing if we get a connection */
 	if (os.conn < C_CONNECTED && ns.conn >= C_CONNECTED) {
 		drbd_resume_al(device);
-		first_peer_device(device)->connection->connect_cnt++;
+		connection->connect_cnt++;
 	}
 
 	/* remember last attach time so request_timer_fn() won't
@@ -1150,7 +1158,7 @@
 		ascw->w.cb = w_after_state_ch;
 		ascw->device = device;
 		ascw->done = done;
-		drbd_queue_work(&first_peer_device(device)->connection->sender_work,
+		drbd_queue_work(&connection->sender_work,
 				&ascw->w);
 	} else {
 		drbd_err(device, "Could not kmalloc an ascw\n");
@@ -1222,13 +1230,16 @@
 			   union drbd_state ns, enum chg_state_flags flags)
 {
 	struct drbd_resource *resource = device->resource;
+	struct drbd_peer_device *peer_device = first_peer_device(device);
+	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
 	struct sib_info sib;
 
 	sib.sib_reason = SIB_STATE_CHANGE;
 	sib.os = os;
 	sib.ns = ns;
 
-	if (os.conn != C_CONNECTED && ns.conn == C_CONNECTED) {
+	if ((os.disk != D_UP_TO_DATE || os.pdsk != D_UP_TO_DATE)
+	&&  (ns.disk == D_UP_TO_DATE && ns.pdsk == D_UP_TO_DATE)) {
 		clear_bit(CRASHED_PRIMARY, &device->flags);
 		if (device->p_uuid)
 			device->p_uuid[UI_FLAGS] &= ~((u64)2);
@@ -1245,7 +1256,6 @@
 	   state change. This function might sleep */
 
 	if (ns.susp_nod) {
-		struct drbd_connection *connection = first_peer_device(device)->connection;
 		enum drbd_req_event what = NOTHING;
 
 		spin_lock_irq(&device->resource->req_lock);
@@ -1267,8 +1277,6 @@
 	}
 
 	if (ns.susp_fen) {
-		struct drbd_connection *connection = first_peer_device(device)->connection;
-
 		spin_lock_irq(&device->resource->req_lock);
 		if (resource->susp_fen && conn_lowest_conn(connection) >= C_CONNECTED) {
 			/* case2: The connection was established again: */
@@ -1294,8 +1302,8 @@
 	 * which is unexpected. */
 	if ((os.conn != C_SYNC_SOURCE && os.conn != C_PAUSED_SYNC_S) &&
 	    (ns.conn == C_SYNC_SOURCE || ns.conn == C_PAUSED_SYNC_S) &&
-	    first_peer_device(device)->connection->agreed_pro_version >= 96 && get_ldev(device)) {
-		drbd_gen_and_send_sync_uuid(first_peer_device(device));
+	    connection->agreed_pro_version >= 96 && get_ldev(device)) {
+		drbd_gen_and_send_sync_uuid(peer_device);
 		put_ldev(device);
 	}
 
@@ -1309,8 +1317,8 @@
 		atomic_set(&device->rs_pending_cnt, 0);
 		drbd_rs_cancel_all(device);
 
-		drbd_send_uuids(first_peer_device(device));
-		drbd_send_state(first_peer_device(device), ns);
+		drbd_send_uuids(peer_device);
+		drbd_send_state(peer_device, ns);
 	}
 	/* No point in queuing send_bitmap if we don't have a connection
 	 * anymore, so check also the _current_ state, not only the new state
@@ -1335,7 +1343,7 @@
 					set_bit(NEW_CUR_UUID, &device->flags);
 				} else {
 					drbd_uuid_new_current(device);
-					drbd_send_uuids(first_peer_device(device));
+					drbd_send_uuids(peer_device);
 				}
 			}
 			put_ldev(device);
@@ -1346,7 +1354,7 @@
 		if (os.peer == R_SECONDARY && ns.peer == R_PRIMARY &&
 		    device->ldev->md.uuid[UI_BITMAP] == 0 && ns.disk >= D_UP_TO_DATE) {
 			drbd_uuid_new_current(device);
-			drbd_send_uuids(first_peer_device(device));
+			drbd_send_uuids(peer_device);
 		}
 		/* D_DISKLESS Peer becomes secondary */
 		if (os.peer == R_PRIMARY && ns.peer == R_SECONDARY)
@@ -1373,16 +1381,16 @@
 	/* Last part of the attaching process ... */
 	if (ns.conn >= C_CONNECTED &&
 	    os.disk == D_ATTACHING && ns.disk == D_NEGOTIATING) {
-		drbd_send_sizes(first_peer_device(device), 0, 0);  /* to start sync... */
-		drbd_send_uuids(first_peer_device(device));
-		drbd_send_state(first_peer_device(device), ns);
+		drbd_send_sizes(peer_device, 0, 0);  /* to start sync... */
+		drbd_send_uuids(peer_device);
+		drbd_send_state(peer_device, ns);
 	}
 
 	/* We want to pause/continue resync, tell peer. */
 	if (ns.conn >= C_CONNECTED &&
 	     ((os.aftr_isp != ns.aftr_isp) ||
 	      (os.user_isp != ns.user_isp)))
-		drbd_send_state(first_peer_device(device), ns);
+		drbd_send_state(peer_device, ns);
 
 	/* In case one of the isp bits got set, suspend other devices. */
 	if ((!os.aftr_isp && !os.peer_isp && !os.user_isp) &&
@@ -1392,10 +1400,10 @@
 	/* Make sure the peer gets informed about eventual state
 	   changes (ISP bits) while we were in WFReportParams. */
 	if (os.conn == C_WF_REPORT_PARAMS && ns.conn >= C_CONNECTED)
-		drbd_send_state(first_peer_device(device), ns);
+		drbd_send_state(peer_device, ns);
 
 	if (os.conn != C_AHEAD && ns.conn == C_AHEAD)
-		drbd_send_state(first_peer_device(device), ns);
+		drbd_send_state(peer_device, ns);
 
 	/* We are in the progress to start a full sync... */
 	if ((os.conn != C_STARTING_SYNC_T && ns.conn == C_STARTING_SYNC_T) ||
@@ -1449,7 +1457,7 @@
 					drbd_disk_str(device->state.disk));
 
 			if (ns.conn >= C_CONNECTED)
-				drbd_send_state(first_peer_device(device), ns);
+				drbd_send_state(peer_device, ns);
 
 			drbd_rs_cancel_all(device);
 
@@ -1473,7 +1481,7 @@
 				 drbd_disk_str(device->state.disk));
 
 		if (ns.conn >= C_CONNECTED)
-			drbd_send_state(first_peer_device(device), ns);
+			drbd_send_state(peer_device, ns);
 		/* corresponding get_ldev in __drbd_set_state
 		 * this may finally trigger drbd_ldev_destroy. */
 		put_ldev(device);
@@ -1481,7 +1489,7 @@
 
 	/* Notify peer that I had a local IO error, and did not detached.. */
 	if (os.disk == D_UP_TO_DATE && ns.disk == D_INCONSISTENT && ns.conn >= C_CONNECTED)
-		drbd_send_state(first_peer_device(device), ns);
+		drbd_send_state(peer_device, ns);
 
 	/* Disks got bigger while they were detached */
 	if (ns.disk > D_NEGOTIATING && ns.pdsk > D_NEGOTIATING &&
@@ -1499,14 +1507,14 @@
 	/* sync target done with resync.  Explicitly notify peer, even though
 	 * it should (at least for non-empty resyncs) already know itself. */
 	if (os.disk < D_UP_TO_DATE && os.conn >= C_SYNC_SOURCE && ns.conn == C_CONNECTED)
-		drbd_send_state(first_peer_device(device), ns);
+		drbd_send_state(peer_device, ns);
 
 	/* Verify finished, or reached stop sector.  Peer did not know about
 	 * the stop sector, and we may even have changed the stop sector during
 	 * verify to interrupt/stop early.  Send the new state. */
 	if (os.conn == C_VERIFY_S && ns.conn == C_CONNECTED
 	&& verify_can_do_stop_sector(device))
-		drbd_send_state(first_peer_device(device), ns);
+		drbd_send_state(peer_device, ns);
 
 	/* This triggers bitmap writeout of potentially still unwritten pages
 	 * if the resync finished cleanly, or aborted because of peer disk
@@ -1563,7 +1571,7 @@
 		old_conf = connection->net_conf;
 		connection->my_addr_len = 0;
 		connection->peer_addr_len = 0;
-		rcu_assign_pointer(connection->net_conf, NULL);
+		RCU_INIT_POINTER(connection->net_conf, NULL);
 		conn_free_crypto(connection);
 		mutex_unlock(&connection->resource->conf_update);
 
@@ -1599,7 +1607,7 @@
 	return 0;
 }
 
-void conn_old_common_state(struct drbd_connection *connection, union drbd_state *pcs, enum chg_state_flags *pf)
+static void conn_old_common_state(struct drbd_connection *connection, union drbd_state *pcs, enum chg_state_flags *pf)
 {
 	enum chg_state_flags flags = ~0;
 	struct drbd_peer_device *peer_device;
@@ -1688,7 +1696,7 @@
 	return rv;
 }
 
-void
+static void
 conn_set_state(struct drbd_connection *connection, union drbd_state mask, union drbd_state val,
 	       union drbd_state *pns_min, union drbd_state *pns_max, enum chg_state_flags flags)
 {
diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c
index d8f57b6..50776b3 100644
--- a/drivers/block/drbd/drbd_worker.c
+++ b/drivers/block/drbd/drbd_worker.c
@@ -67,13 +67,10 @@
  */
 void drbd_md_io_complete(struct bio *bio, int error)
 {
-	struct drbd_md_io *md_io;
 	struct drbd_device *device;
 
-	md_io = (struct drbd_md_io *)bio->bi_private;
-	device = container_of(md_io, struct drbd_device, md_io);
-
-	md_io->error = error;
+	device = bio->bi_private;
+	device->md_io.error = error;
 
 	/* We grabbed an extra reference in _drbd_md_sync_page_io() to be able
 	 * to timeout on the lower level device, and eventually detach from it.
@@ -87,7 +84,7 @@
 	 * ASSERT(atomic_read(&device->md_io_in_use) == 1) there.
 	 */
 	drbd_md_put_buffer(device);
-	md_io->done = 1;
+	device->md_io.done = 1;
 	wake_up(&device->misc_wait);
 	bio_put(bio);
 	if (device->ldev) /* special case: drbd_md_read() during drbd_adm_attach() */
@@ -135,6 +132,7 @@
 	i = peer_req->i;
 	do_al_complete_io = peer_req->flags & EE_CALL_AL_COMPLETE_IO;
 	block_id = peer_req->block_id;
+	peer_req->flags &= ~EE_CALL_AL_COMPLETE_IO;
 
 	spin_lock_irqsave(&device->resource->req_lock, flags);
 	device->writ_cnt += peer_req->i.size >> 9;
@@ -398,9 +396,6 @@
 	if (!get_ldev(device))
 		return -EIO;
 
-	if (drbd_rs_should_slow_down(device, sector))
-		goto defer;
-
 	/* GFP_TRY, because if there is no memory available right now, this may
 	 * be rescheduled for later. It is "only" background resync, after all. */
 	peer_req = drbd_alloc_peer_req(peer_device, ID_SYNCER /* unused */, sector,
@@ -410,7 +405,7 @@
 
 	peer_req->w.cb = w_e_send_csum;
 	spin_lock_irq(&device->resource->req_lock);
-	list_add(&peer_req->w.list, &device->read_ee);
+	list_add_tail(&peer_req->w.list, &device->read_ee);
 	spin_unlock_irq(&device->resource->req_lock);
 
 	atomic_add(size >> 9, &device->rs_sect_ev);
@@ -452,9 +447,9 @@
 {
 	struct drbd_device *device = (struct drbd_device *) data;
 
-	if (list_empty(&device->resync_work.list))
-		drbd_queue_work(&first_peer_device(device)->connection->sender_work,
-				&device->resync_work);
+	drbd_queue_work_if_unqueued(
+		&first_peer_device(device)->connection->sender_work,
+		&device->resync_work);
 }
 
 static void fifo_set(struct fifo_buffer *fb, int value)
@@ -504,9 +499,9 @@
 static int drbd_rs_controller(struct drbd_device *device, unsigned int sect_in)
 {
 	struct disk_conf *dc;
-	unsigned int want;     /* The number of sectors we want in the proxy */
+	unsigned int want;     /* The number of sectors we want in-flight */
 	int req_sect; /* Number of sectors to request in this turn */
-	int correction; /* Number of sectors more we need in the proxy*/
+	int correction; /* Number of sectors more we need in-flight */
 	int cps; /* correction per invocation of drbd_rs_controller() */
 	int steps; /* Number of time steps to plan ahead */
 	int curr_corr;
@@ -577,20 +572,27 @@
 	 * potentially causing a distributed deadlock on congestion during
 	 * online-verify or (checksum-based) resync, if max-buffers,
 	 * socket buffer sizes and resync rate settings are mis-configured. */
-	if (mxb - device->rs_in_flight < number)
-		number = mxb - device->rs_in_flight;
+
+	/* note that "number" is in units of "BM_BLOCK_SIZE" (which is 4k),
+	 * mxb (as used here, and in drbd_alloc_pages on the peer) is
+	 * "number of pages" (typically also 4k),
+	 * but "rs_in_flight" is in "sectors" (512 Byte). */
+	if (mxb - device->rs_in_flight/8 < number)
+		number = mxb - device->rs_in_flight/8;
 
 	return number;
 }
 
-static int make_resync_request(struct drbd_device *device, int cancel)
+static int make_resync_request(struct drbd_device *const device, int cancel)
 {
+	struct drbd_peer_device *const peer_device = first_peer_device(device);
+	struct drbd_connection *const connection = peer_device ? peer_device->connection : NULL;
 	unsigned long bit;
 	sector_t sector;
 	const sector_t capacity = drbd_get_capacity(device->this_bdev);
 	int max_bio_size;
 	int number, rollback_i, size;
-	int align, queued, sndbuf;
+	int align, requeue = 0;
 	int i = 0;
 
 	if (unlikely(cancel))
@@ -617,17 +619,22 @@
 		goto requeue;
 
 	for (i = 0; i < number; i++) {
-		/* Stop generating RS requests, when half of the send buffer is filled */
-		mutex_lock(&first_peer_device(device)->connection->data.mutex);
-		if (first_peer_device(device)->connection->data.socket) {
-			queued = first_peer_device(device)->connection->data.socket->sk->sk_wmem_queued;
-			sndbuf = first_peer_device(device)->connection->data.socket->sk->sk_sndbuf;
-		} else {
-			queued = 1;
-			sndbuf = 0;
-		}
-		mutex_unlock(&first_peer_device(device)->connection->data.mutex);
-		if (queued > sndbuf / 2)
+		/* Stop generating RS requests when half of the send buffer is filled,
+		 * but notify TCP that we'd like to have more space. */
+		mutex_lock(&connection->data.mutex);
+		if (connection->data.socket) {
+			struct sock *sk = connection->data.socket->sk;
+			int queued = sk->sk_wmem_queued;
+			int sndbuf = sk->sk_sndbuf;
+			if (queued > sndbuf / 2) {
+				requeue = 1;
+				if (sk->sk_socket)
+					set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
+			}
+		} else
+			requeue = 1;
+		mutex_unlock(&connection->data.mutex);
+		if (requeue)
 			goto requeue;
 
 next_sector:
@@ -642,8 +649,7 @@
 
 		sector = BM_BIT_TO_SECT(bit);
 
-		if (drbd_rs_should_slow_down(device, sector) ||
-		    drbd_try_rs_begin_io(device, sector)) {
+		if (drbd_try_rs_begin_io(device, sector)) {
 			device->bm_resync_fo = bit;
 			goto requeue;
 		}
@@ -696,9 +702,9 @@
 		/* adjust very last sectors, in case we are oddly sized */
 		if (sector + (size>>9) > capacity)
 			size = (capacity-sector)<<9;
-		if (first_peer_device(device)->connection->agreed_pro_version >= 89 &&
-		    first_peer_device(device)->connection->csums_tfm) {
-			switch (read_for_csum(first_peer_device(device), sector, size)) {
+
+		if (device->use_csums) {
+			switch (read_for_csum(peer_device, sector, size)) {
 			case -EIO: /* Disk failure */
 				put_ldev(device);
 				return -EIO;
@@ -717,7 +723,7 @@
 			int err;
 
 			inc_rs_pending(device);
-			err = drbd_send_drequest(first_peer_device(device), P_RS_DATA_REQUEST,
+			err = drbd_send_drequest(peer_device, P_RS_DATA_REQUEST,
 						 sector, size, ID_SYNCER);
 			if (err) {
 				drbd_err(device, "drbd_send_drequest() failed, aborting...\n");
@@ -774,8 +780,7 @@
 
 		size = BM_BLOCK_SIZE;
 
-		if (drbd_rs_should_slow_down(device, sector) ||
-		    drbd_try_rs_begin_io(device, sector)) {
+		if (drbd_try_rs_begin_io(device, sector)) {
 			device->ov_position = sector;
 			goto requeue;
 		}
@@ -911,7 +916,7 @@
 		if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T)
 			khelper_cmd = "after-resync-target";
 
-		if (first_peer_device(device)->connection->csums_tfm && device->rs_total) {
+		if (device->use_csums && device->rs_total) {
 			const unsigned long s = device->rs_same_csum;
 			const unsigned long t = device->rs_total;
 			const int ratio =
@@ -1351,13 +1356,15 @@
 {
 	struct drbd_request *req = container_of(w, struct drbd_request, w);
 	struct drbd_device *device = req->device;
-	struct drbd_connection *connection = first_peer_device(device)->connection;
+	struct drbd_peer_device *const peer_device = first_peer_device(device);
+	struct drbd_connection *const connection = peer_device->connection;
 	int err;
 
 	if (unlikely(cancel)) {
 		req_mod(req, SEND_CANCELED);
 		return 0;
 	}
+	req->pre_send_jif = jiffies;
 
 	/* this time, no connection->send.current_epoch_writes++;
 	 * If it was sent, it was the closing barrier for the last
@@ -1365,7 +1372,7 @@
 	 * No more barriers will be sent, until we leave AHEAD mode again. */
 	maybe_send_barrier(connection, req->epoch);
 
-	err = drbd_send_out_of_sync(first_peer_device(device), req);
+	err = drbd_send_out_of_sync(peer_device, req);
 	req_mod(req, OOS_HANDED_TO_NETWORK);
 
 	return err;
@@ -1380,19 +1387,21 @@
 {
 	struct drbd_request *req = container_of(w, struct drbd_request, w);
 	struct drbd_device *device = req->device;
-	struct drbd_connection *connection = first_peer_device(device)->connection;
+	struct drbd_peer_device *const peer_device = first_peer_device(device);
+	struct drbd_connection *connection = peer_device->connection;
 	int err;
 
 	if (unlikely(cancel)) {
 		req_mod(req, SEND_CANCELED);
 		return 0;
 	}
+	req->pre_send_jif = jiffies;
 
 	re_init_if_first_write(connection, req->epoch);
 	maybe_send_barrier(connection, req->epoch);
 	connection->send.current_epoch_writes++;
 
-	err = drbd_send_dblock(first_peer_device(device), req);
+	err = drbd_send_dblock(peer_device, req);
 	req_mod(req, err ? SEND_FAILED : HANDED_OVER_TO_NETWORK);
 
 	return err;
@@ -1407,19 +1416,21 @@
 {
 	struct drbd_request *req = container_of(w, struct drbd_request, w);
 	struct drbd_device *device = req->device;
-	struct drbd_connection *connection = first_peer_device(device)->connection;
+	struct drbd_peer_device *const peer_device = first_peer_device(device);
+	struct drbd_connection *connection = peer_device->connection;
 	int err;
 
 	if (unlikely(cancel)) {
 		req_mod(req, SEND_CANCELED);
 		return 0;
 	}
+	req->pre_send_jif = jiffies;
 
 	/* Even read requests may close a write epoch,
 	 * if there was any yet. */
 	maybe_send_barrier(connection, req->epoch);
 
-	err = drbd_send_drequest(first_peer_device(device), P_DATA_REQUEST, req->i.sector, req->i.size,
+	err = drbd_send_drequest(peer_device, P_DATA_REQUEST, req->i.sector, req->i.size,
 				 (unsigned long)req);
 
 	req_mod(req, err ? SEND_FAILED : HANDED_OVER_TO_NETWORK);
@@ -1433,7 +1444,7 @@
 	struct drbd_device *device = req->device;
 
 	if (bio_data_dir(req->master_bio) == WRITE && req->rq_state & RQ_IN_ACT_LOG)
-		drbd_al_begin_io(device, &req->i, false);
+		drbd_al_begin_io(device, &req->i);
 
 	drbd_req_make_private_bio(req, req->master_bio);
 	req->private_bio->bi_bdev = device->ldev->backing_bdev;
@@ -1601,26 +1612,32 @@
 void start_resync_timer_fn(unsigned long data)
 {
 	struct drbd_device *device = (struct drbd_device *) data;
-
-	drbd_queue_work(&first_peer_device(device)->connection->sender_work,
-			&device->start_resync_work);
+	drbd_device_post_work(device, RS_START);
 }
 
-int w_start_resync(struct drbd_work *w, int cancel)
+static void do_start_resync(struct drbd_device *device)
 {
-	struct drbd_device *device =
-		container_of(w, struct drbd_device, start_resync_work);
-
 	if (atomic_read(&device->unacked_cnt) || atomic_read(&device->rs_pending_cnt)) {
-		drbd_warn(device, "w_start_resync later...\n");
+		drbd_warn(device, "postponing start_resync ...\n");
 		device->start_resync_timer.expires = jiffies + HZ/10;
 		add_timer(&device->start_resync_timer);
-		return 0;
+		return;
 	}
 
 	drbd_start_resync(device, C_SYNC_SOURCE);
 	clear_bit(AHEAD_TO_SYNC_SOURCE, &device->flags);
-	return 0;
+}
+
+static bool use_checksum_based_resync(struct drbd_connection *connection, struct drbd_device *device)
+{
+	bool csums_after_crash_only;
+	rcu_read_lock();
+	csums_after_crash_only = rcu_dereference(connection->net_conf)->csums_after_crash_only;
+	rcu_read_unlock();
+	return connection->agreed_pro_version >= 89 &&		/* supported? */
+		connection->csums_tfm &&			/* configured? */
+		(csums_after_crash_only == 0			/* use for each resync? */
+		 || test_bit(CRASHED_PRIMARY, &device->flags));	/* or only after Primary crash? */
 }
 
 /**
@@ -1633,6 +1650,8 @@
  */
 void drbd_start_resync(struct drbd_device *device, enum drbd_conns side)
 {
+	struct drbd_peer_device *peer_device = first_peer_device(device);
+	struct drbd_connection *connection = peer_device ? peer_device->connection : NULL;
 	union drbd_state ns;
 	int r;
 
@@ -1651,7 +1670,7 @@
 			if (r > 0) {
 				drbd_info(device, "before-resync-target handler returned %d, "
 					 "dropping connection.\n", r);
-				conn_request_state(first_peer_device(device)->connection, NS(conn, C_DISCONNECTING), CS_HARD);
+				conn_request_state(connection, NS(conn, C_DISCONNECTING), CS_HARD);
 				return;
 			}
 		} else /* C_SYNC_SOURCE */ {
@@ -1664,7 +1683,7 @@
 				} else {
 					drbd_info(device, "before-resync-source handler returned %d, "
 						 "dropping connection.\n", r);
-					conn_request_state(first_peer_device(device)->connection,
+					conn_request_state(connection,
 							   NS(conn, C_DISCONNECTING), CS_HARD);
 					return;
 				}
@@ -1672,7 +1691,7 @@
 		}
 	}
 
-	if (current == first_peer_device(device)->connection->worker.task) {
+	if (current == connection->worker.task) {
 		/* The worker should not sleep waiting for state_mutex,
 		   that can take long */
 		if (!mutex_trylock(device->state_mutex)) {
@@ -1733,11 +1752,20 @@
 			device->rs_mark_time[i] = now;
 		}
 		_drbd_pause_after(device);
+		/* Forget potentially stale cached per resync extent bit-counts.
+		 * Open coded drbd_rs_cancel_all(device), we already have IRQs
+		 * disabled, and know the disk state is ok. */
+		spin_lock(&device->al_lock);
+		lc_reset(device->resync);
+		device->resync_locked = 0;
+		device->resync_wenr = LC_FREE;
+		spin_unlock(&device->al_lock);
 	}
 	write_unlock(&global_state_lock);
 	spin_unlock_irq(&device->resource->req_lock);
 
 	if (r == SS_SUCCESS) {
+		wake_up(&device->al_wait); /* for lc_reset() above */
 		/* reset rs_last_bcast when a resync or verify is started,
 		 * to deal with potential jiffies wrap. */
 		device->rs_last_bcast = jiffies - HZ;
@@ -1746,8 +1774,12 @@
 		     drbd_conn_str(ns.conn),
 		     (unsigned long) device->rs_total << (BM_BLOCK_SHIFT-10),
 		     (unsigned long) device->rs_total);
-		if (side == C_SYNC_TARGET)
+		if (side == C_SYNC_TARGET) {
 			device->bm_resync_fo = 0;
+			device->use_csums = use_checksum_based_resync(connection, device);
+		} else {
+			device->use_csums = 0;
+		}
 
 		/* Since protocol 96, we must serialize drbd_gen_and_send_sync_uuid
 		 * with w_send_oos, or the sync target will get confused as to
@@ -1756,12 +1788,10 @@
 		 * drbd_resync_finished from here in that case.
 		 * We drbd_gen_and_send_sync_uuid here for protocol < 96,
 		 * and from after_state_ch otherwise. */
-		if (side == C_SYNC_SOURCE &&
-		    first_peer_device(device)->connection->agreed_pro_version < 96)
-			drbd_gen_and_send_sync_uuid(first_peer_device(device));
+		if (side == C_SYNC_SOURCE && connection->agreed_pro_version < 96)
+			drbd_gen_and_send_sync_uuid(peer_device);
 
-		if (first_peer_device(device)->connection->agreed_pro_version < 95 &&
-		    device->rs_total == 0) {
+		if (connection->agreed_pro_version < 95 && device->rs_total == 0) {
 			/* This still has a race (about when exactly the peers
 			 * detect connection loss) that can lead to a full sync
 			 * on next handshake. In 8.3.9 we fixed this with explicit
@@ -1777,7 +1807,7 @@
 				int timeo;
 
 				rcu_read_lock();
-				nc = rcu_dereference(first_peer_device(device)->connection->net_conf);
+				nc = rcu_dereference(connection->net_conf);
 				timeo = nc->ping_int * HZ + nc->ping_timeo * HZ / 9;
 				rcu_read_unlock();
 				schedule_timeout_interruptible(timeo);
@@ -1799,10 +1829,165 @@
 	mutex_unlock(device->state_mutex);
 }
 
+static void update_on_disk_bitmap(struct drbd_device *device, bool resync_done)
+{
+	struct sib_info sib = { .sib_reason = SIB_SYNC_PROGRESS, };
+	device->rs_last_bcast = jiffies;
+
+	if (!get_ldev(device))
+		return;
+
+	drbd_bm_write_lazy(device, 0);
+	if (resync_done && is_sync_state(device->state.conn))
+		drbd_resync_finished(device);
+
+	drbd_bcast_event(device, &sib);
+	/* update timestamp, in case it took a while to write out stuff */
+	device->rs_last_bcast = jiffies;
+	put_ldev(device);
+}
+
+static void drbd_ldev_destroy(struct drbd_device *device)
+{
+	lc_destroy(device->resync);
+	device->resync = NULL;
+	lc_destroy(device->act_log);
+	device->act_log = NULL;
+	__no_warn(local,
+		drbd_free_ldev(device->ldev);
+		device->ldev = NULL;);
+	clear_bit(GOING_DISKLESS, &device->flags);
+	wake_up(&device->misc_wait);
+}
+
+static void go_diskless(struct drbd_device *device)
+{
+	D_ASSERT(device, device->state.disk == D_FAILED);
+	/* we cannot assert local_cnt == 0 here, as get_ldev_if_state will
+	 * inc/dec it frequently. Once we are D_DISKLESS, no one will touch
+	 * the protected members anymore, though, so once put_ldev reaches zero
+	 * again, it will be safe to free them. */
+
+	/* Try to write changed bitmap pages, read errors may have just
+	 * set some bits outside the area covered by the activity log.
+	 *
+	 * If we have an IO error during the bitmap writeout,
+	 * we will want a full sync next time, just in case.
+	 * (Do we want a specific meta data flag for this?)
+	 *
+	 * If that does not make it to stable storage either,
+	 * we cannot do anything about that anymore.
+	 *
+	 * We still need to check if both bitmap and ldev are present, we may
+	 * end up here after a failed attach, before ldev was even assigned.
+	 */
+	if (device->bitmap && device->ldev) {
+		/* An interrupted resync or similar is allowed to recounts bits
+		 * while we detach.
+		 * Any modifications would not be expected anymore, though.
+		 */
+		if (drbd_bitmap_io_from_worker(device, drbd_bm_write,
+					"detach", BM_LOCKED_TEST_ALLOWED)) {
+			if (test_bit(WAS_READ_ERROR, &device->flags)) {
+				drbd_md_set_flag(device, MDF_FULL_SYNC);
+				drbd_md_sync(device);
+			}
+		}
+	}
+
+	drbd_force_state(device, NS(disk, D_DISKLESS));
+}
+
+static int do_md_sync(struct drbd_device *device)
+{
+	drbd_warn(device, "md_sync_timer expired! Worker calls drbd_md_sync().\n");
+	drbd_md_sync(device);
+	return 0;
+}
+
+/* only called from drbd_worker thread, no locking */
+void __update_timing_details(
+		struct drbd_thread_timing_details *tdp,
+		unsigned int *cb_nr,
+		void *cb,
+		const char *fn, const unsigned int line)
+{
+	unsigned int i = *cb_nr % DRBD_THREAD_DETAILS_HIST;
+	struct drbd_thread_timing_details *td = tdp + i;
+
+	td->start_jif = jiffies;
+	td->cb_addr = cb;
+	td->caller_fn = fn;
+	td->line = line;
+	td->cb_nr = *cb_nr;
+
+	i = (i+1) % DRBD_THREAD_DETAILS_HIST;
+	td = tdp + i;
+	memset(td, 0, sizeof(*td));
+
+	++(*cb_nr);
+}
+
+#define WORK_PENDING(work_bit, todo)	(todo & (1UL << work_bit))
+static void do_device_work(struct drbd_device *device, const unsigned long todo)
+{
+	if (WORK_PENDING(MD_SYNC, todo))
+		do_md_sync(device);
+	if (WORK_PENDING(RS_DONE, todo) ||
+	    WORK_PENDING(RS_PROGRESS, todo))
+		update_on_disk_bitmap(device, WORK_PENDING(RS_DONE, todo));
+	if (WORK_PENDING(GO_DISKLESS, todo))
+		go_diskless(device);
+	if (WORK_PENDING(DESTROY_DISK, todo))
+		drbd_ldev_destroy(device);
+	if (WORK_PENDING(RS_START, todo))
+		do_start_resync(device);
+}
+
+#define DRBD_DEVICE_WORK_MASK	\
+	((1UL << GO_DISKLESS)	\
+	|(1UL << DESTROY_DISK)	\
+	|(1UL << MD_SYNC)	\
+	|(1UL << RS_START)	\
+	|(1UL << RS_PROGRESS)	\
+	|(1UL << RS_DONE)	\
+	)
+
+static unsigned long get_work_bits(unsigned long *flags)
+{
+	unsigned long old, new;
+	do {
+		old = *flags;
+		new = old & ~DRBD_DEVICE_WORK_MASK;
+	} while (cmpxchg(flags, old, new) != old);
+	return old & DRBD_DEVICE_WORK_MASK;
+}
+
+static void do_unqueued_work(struct drbd_connection *connection)
+{
+	struct drbd_peer_device *peer_device;
+	int vnr;
+
+	rcu_read_lock();
+	idr_for_each_entry(&connection->peer_devices, peer_device, vnr) {
+		struct drbd_device *device = peer_device->device;
+		unsigned long todo = get_work_bits(&device->flags);
+		if (!todo)
+			continue;
+
+		kref_get(&device->kref);
+		rcu_read_unlock();
+		do_device_work(device, todo);
+		kref_put(&device->kref, drbd_destroy_device);
+		rcu_read_lock();
+	}
+	rcu_read_unlock();
+}
+
 static bool dequeue_work_batch(struct drbd_work_queue *queue, struct list_head *work_list)
 {
 	spin_lock_irq(&queue->q_lock);
-	list_splice_init(&queue->q, work_list);
+	list_splice_tail_init(&queue->q, work_list);
 	spin_unlock_irq(&queue->q_lock);
 	return !list_empty(work_list);
 }
@@ -1851,7 +2036,7 @@
 		/* dequeue single item only,
 		 * we still use drbd_queue_work_front() in some places */
 		if (!list_empty(&connection->sender_work.q))
-			list_move(connection->sender_work.q.next, work_list);
+			list_splice_tail_init(&connection->sender_work.q, work_list);
 		spin_unlock(&connection->sender_work.q_lock);	/* FIXME get rid of this one? */
 		if (!list_empty(work_list) || signal_pending(current)) {
 			spin_unlock_irq(&connection->resource->req_lock);
@@ -1873,6 +2058,14 @@
 		if (send_barrier)
 			maybe_send_barrier(connection,
 					connection->send.current_epoch_nr + 1);
+
+		if (test_bit(DEVICE_WORK_PENDING, &connection->flags))
+			break;
+
+		/* drbd_send() may have called flush_signals() */
+		if (get_t_state(&connection->worker) != RUNNING)
+			break;
+
 		schedule();
 		/* may be woken up for other things but new work, too,
 		 * e.g. if the current epoch got closed.
@@ -1906,10 +2099,15 @@
 	while (get_t_state(thi) == RUNNING) {
 		drbd_thread_current_set_cpu(thi);
 
-		/* as long as we use drbd_queue_work_front(),
-		 * we may only dequeue single work items here, not batches. */
-		if (list_empty(&work_list))
+		if (list_empty(&work_list)) {
+			update_worker_timing_details(connection, wait_for_work);
 			wait_for_work(connection, &work_list);
+		}
+
+		if (test_and_clear_bit(DEVICE_WORK_PENDING, &connection->flags)) {
+			update_worker_timing_details(connection, do_unqueued_work);
+			do_unqueued_work(connection);
+		}
 
 		if (signal_pending(current)) {
 			flush_signals(current);
@@ -1926,6 +2124,7 @@
 		while (!list_empty(&work_list)) {
 			w = list_first_entry(&work_list, struct drbd_work, list);
 			list_del_init(&w->list);
+			update_worker_timing_details(connection, w->cb);
 			if (w->cb(w, connection->cstate < C_WF_REPORT_PARAMS) == 0)
 				continue;
 			if (connection->cstate >= C_WF_REPORT_PARAMS)
@@ -1934,13 +2133,18 @@
 	}
 
 	do {
+		if (test_and_clear_bit(DEVICE_WORK_PENDING, &connection->flags)) {
+			update_worker_timing_details(connection, do_unqueued_work);
+			do_unqueued_work(connection);
+		}
 		while (!list_empty(&work_list)) {
 			w = list_first_entry(&work_list, struct drbd_work, list);
 			list_del_init(&w->list);
+			update_worker_timing_details(connection, w->cb);
 			w->cb(w, 1);
 		}
 		dequeue_work_batch(&connection->sender_work, &work_list);
-	} while (!list_empty(&work_list));
+	} while (!list_empty(&work_list) || test_bit(DEVICE_WORK_PENDING, &connection->flags));
 
 	rcu_read_lock();
 	idr_for_each_entry(&connection->peer_devices, peer_device, vnr) {
diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
index f63d358..0a58140 100644
--- a/drivers/block/virtio_blk.c
+++ b/drivers/block/virtio_blk.c
@@ -15,17 +15,22 @@
 #include <linux/numa.h>
 
 #define PART_BITS 4
+#define VQ_NAME_LEN 16
 
 static int major;
 static DEFINE_IDA(vd_index_ida);
 
 static struct workqueue_struct *virtblk_wq;
 
+struct virtio_blk_vq {
+	struct virtqueue *vq;
+	spinlock_t lock;
+	char name[VQ_NAME_LEN];
+} ____cacheline_aligned_in_smp;
+
 struct virtio_blk
 {
 	struct virtio_device *vdev;
-	struct virtqueue *vq;
-	spinlock_t vq_lock;
 
 	/* The disk structure for the kernel. */
 	struct gendisk *disk;
@@ -47,6 +52,10 @@
 
 	/* Ida index - used to track minor number allocations. */
 	int index;
+
+	/* num of vqs */
+	int num_vqs;
+	struct virtio_blk_vq *vqs;
 };
 
 struct virtblk_req
@@ -133,14 +142,15 @@
 {
 	struct virtio_blk *vblk = vq->vdev->priv;
 	bool req_done = false;
+	int qid = vq->index;
 	struct virtblk_req *vbr;
 	unsigned long flags;
 	unsigned int len;
 
-	spin_lock_irqsave(&vblk->vq_lock, flags);
+	spin_lock_irqsave(&vblk->vqs[qid].lock, flags);
 	do {
 		virtqueue_disable_cb(vq);
-		while ((vbr = virtqueue_get_buf(vblk->vq, &len)) != NULL) {
+		while ((vbr = virtqueue_get_buf(vblk->vqs[qid].vq, &len)) != NULL) {
 			blk_mq_complete_request(vbr->req);
 			req_done = true;
 		}
@@ -151,7 +161,7 @@
 	/* In case queue is stopped waiting for more buffers. */
 	if (req_done)
 		blk_mq_start_stopped_hw_queues(vblk->disk->queue, true);
-	spin_unlock_irqrestore(&vblk->vq_lock, flags);
+	spin_unlock_irqrestore(&vblk->vqs[qid].lock, flags);
 }
 
 static int virtio_queue_rq(struct blk_mq_hw_ctx *hctx, struct request *req)
@@ -160,6 +170,7 @@
 	struct virtblk_req *vbr = blk_mq_rq_to_pdu(req);
 	unsigned long flags;
 	unsigned int num;
+	int qid = hctx->queue_num;
 	const bool last = (req->cmd_flags & REQ_END) != 0;
 	int err;
 	bool notify = false;
@@ -202,12 +213,12 @@
 			vbr->out_hdr.type |= VIRTIO_BLK_T_IN;
 	}
 
-	spin_lock_irqsave(&vblk->vq_lock, flags);
-	err = __virtblk_add_req(vblk->vq, vbr, vbr->sg, num);
+	spin_lock_irqsave(&vblk->vqs[qid].lock, flags);
+	err = __virtblk_add_req(vblk->vqs[qid].vq, vbr, vbr->sg, num);
 	if (err) {
-		virtqueue_kick(vblk->vq);
+		virtqueue_kick(vblk->vqs[qid].vq);
 		blk_mq_stop_hw_queue(hctx);
-		spin_unlock_irqrestore(&vblk->vq_lock, flags);
+		spin_unlock_irqrestore(&vblk->vqs[qid].lock, flags);
 		/* Out of mem doesn't actually happen, since we fall back
 		 * to direct descriptors */
 		if (err == -ENOMEM || err == -ENOSPC)
@@ -215,12 +226,12 @@
 		return BLK_MQ_RQ_QUEUE_ERROR;
 	}
 
-	if (last && virtqueue_kick_prepare(vblk->vq))
+	if (last && virtqueue_kick_prepare(vblk->vqs[qid].vq))
 		notify = true;
-	spin_unlock_irqrestore(&vblk->vq_lock, flags);
+	spin_unlock_irqrestore(&vblk->vqs[qid].lock, flags);
 
 	if (notify)
-		virtqueue_notify(vblk->vq);
+		virtqueue_notify(vblk->vqs[qid].vq);
 	return BLK_MQ_RQ_QUEUE_OK;
 }
 
@@ -377,12 +388,64 @@
 static int init_vq(struct virtio_blk *vblk)
 {
 	int err = 0;
+	int i;
+	vq_callback_t **callbacks;
+	const char **names;
+	struct virtqueue **vqs;
+	unsigned short num_vqs;
+	struct virtio_device *vdev = vblk->vdev;
 
-	/* We expect one virtqueue, for output. */
-	vblk->vq = virtio_find_single_vq(vblk->vdev, virtblk_done, "requests");
-	if (IS_ERR(vblk->vq))
-		err = PTR_ERR(vblk->vq);
+	err = virtio_cread_feature(vdev, VIRTIO_BLK_F_MQ,
+				   struct virtio_blk_config, num_queues,
+				   &num_vqs);
+	if (err)
+		num_vqs = 1;
 
+	vblk->vqs = kmalloc(sizeof(*vblk->vqs) * num_vqs, GFP_KERNEL);
+	if (!vblk->vqs) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	names = kmalloc(sizeof(*names) * num_vqs, GFP_KERNEL);
+	if (!names)
+		goto err_names;
+
+	callbacks = kmalloc(sizeof(*callbacks) * num_vqs, GFP_KERNEL);
+	if (!callbacks)
+		goto err_callbacks;
+
+	vqs = kmalloc(sizeof(*vqs) * num_vqs, GFP_KERNEL);
+	if (!vqs)
+		goto err_vqs;
+
+	for (i = 0; i < num_vqs; i++) {
+		callbacks[i] = virtblk_done;
+		snprintf(vblk->vqs[i].name, VQ_NAME_LEN, "req.%d", i);
+		names[i] = vblk->vqs[i].name;
+	}
+
+	/* Discover virtqueues and write information to configuration.  */
+	err = vdev->config->find_vqs(vdev, num_vqs, vqs, callbacks, names);
+	if (err)
+		goto err_find_vqs;
+
+	for (i = 0; i < num_vqs; i++) {
+		spin_lock_init(&vblk->vqs[i].lock);
+		vblk->vqs[i].vq = vqs[i];
+	}
+	vblk->num_vqs = num_vqs;
+
+ err_find_vqs:
+	kfree(vqs);
+ err_vqs:
+	kfree(callbacks);
+ err_callbacks:
+	kfree(names);
+ err_names:
+	if (err)
+		kfree(vblk->vqs);
+ out:
 	return err;
 }
 
@@ -551,7 +614,6 @@
 	err = init_vq(vblk);
 	if (err)
 		goto out_free_vblk;
-	spin_lock_init(&vblk->vq_lock);
 
 	/* FIXME: How many partitions?  How long is a piece of string? */
 	vblk->disk = alloc_disk(1 << PART_BITS);
@@ -562,7 +624,7 @@
 
 	/* Default queue sizing is to fill the ring. */
 	if (!virtblk_queue_depth) {
-		virtblk_queue_depth = vblk->vq->num_free;
+		virtblk_queue_depth = vblk->vqs[0].vq->num_free;
 		/* ... but without indirect descs, we use 2 descs per req */
 		if (!virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC))
 			virtblk_queue_depth /= 2;
@@ -570,7 +632,6 @@
 
 	memset(&vblk->tag_set, 0, sizeof(vblk->tag_set));
 	vblk->tag_set.ops = &virtio_mq_ops;
-	vblk->tag_set.nr_hw_queues = 1;
 	vblk->tag_set.queue_depth = virtblk_queue_depth;
 	vblk->tag_set.numa_node = NUMA_NO_NODE;
 	vblk->tag_set.flags = BLK_MQ_F_SHOULD_MERGE;
@@ -578,6 +639,7 @@
 		sizeof(struct virtblk_req) +
 		sizeof(struct scatterlist) * sg_elems;
 	vblk->tag_set.driver_data = vblk;
+	vblk->tag_set.nr_hw_queues = vblk->num_vqs;
 
 	err = blk_mq_alloc_tag_set(&vblk->tag_set);
 	if (err)
@@ -727,6 +789,7 @@
 	refc = atomic_read(&disk_to_dev(vblk->disk)->kobj.kref.refcount);
 	put_disk(vblk->disk);
 	vdev->config->del_vqs(vdev);
+	kfree(vblk->vqs);
 	kfree(vblk);
 
 	/* Only free device id if we don't have any users */
@@ -777,7 +840,8 @@
 static unsigned int features[] = {
 	VIRTIO_BLK_F_SEG_MAX, VIRTIO_BLK_F_SIZE_MAX, VIRTIO_BLK_F_GEOMETRY,
 	VIRTIO_BLK_F_RO, VIRTIO_BLK_F_BLK_SIZE, VIRTIO_BLK_F_SCSI,
-	VIRTIO_BLK_F_WCE, VIRTIO_BLK_F_TOPOLOGY, VIRTIO_BLK_F_CONFIG_WCE
+	VIRTIO_BLK_F_WCE, VIRTIO_BLK_F_TOPOLOGY, VIRTIO_BLK_F_CONFIG_WCE,
+	VIRTIO_BLK_F_MQ,
 };
 
 static struct virtio_driver virtio_blk = {
diff --git a/drivers/cpufreq/pmac64-cpufreq.c b/drivers/cpufreq/pmac64-cpufreq.c
index 8bc4229..4ff8687 100644
--- a/drivers/cpufreq/pmac64-cpufreq.c
+++ b/drivers/cpufreq/pmac64-cpufreq.c
@@ -499,8 +499,7 @@
 	}
 
 	/* Lookup the i2c hwclock */
-	for (hwclock = NULL;
-	     (hwclock = of_find_node_by_name(hwclock, "i2c-hwclock")) != NULL;){
+	for_each_node_by_name(hwclock, "i2c-hwclock") {
 		const char *loc = of_get_property(hwclock,
 				"hwctrl-location", NULL);
 		if (loc == NULL)
diff --git a/drivers/crypto/nx/nx-842.c b/drivers/crypto/nx/nx-842.c
index 544f6d3..061407d 100644
--- a/drivers/crypto/nx/nx-842.c
+++ b/drivers/crypto/nx/nx-842.c
@@ -936,28 +936,14 @@
 		goto error_out;
 	}
 
-	/* Set ptr to new property if provided */
-	if (new_prop) {
-		/* Single property */
-		if (!strncmp(new_prop->name, "status", new_prop->length)) {
-			status = new_prop;
-
-		} else if (!strncmp(new_prop->name, "ibm,max-sg-len",
-					new_prop->length)) {
-			maxsglen = new_prop;
-
-		} else if (!strncmp(new_prop->name, "ibm,max-sync-cop",
-					new_prop->length)) {
-			maxsyncop = new_prop;
-
-		} else {
-			/*
-			 * Skip the update, the property being updated
-			 * has no impact.
-			 */
-			goto out;
-		}
-	}
+	/*
+	 * If this is a property update, there are only certain properties that
+	 * we care about. Bail if it isn't in the below list
+	 */
+	if (new_prop && (strncmp(new_prop->name, "status", new_prop->length) ||
+		         strncmp(new_prop->name, "ibm,max-sg-len", new_prop->length) ||
+		         strncmp(new_prop->name, "ibm,max-sync-cop", new_prop->length)))
+		goto out;
 
 	/* Perform property updates */
 	ret = nx842_OF_upd_status(new_devdata, status);
diff --git a/drivers/edac/cell_edac.c b/drivers/edac/cell_edac.c
index 374b57f..a12c855 100644
--- a/drivers/edac/cell_edac.c
+++ b/drivers/edac/cell_edac.c
@@ -134,8 +134,7 @@
 	int				j;
 	u32				nr_pages;
 
-	for (np = NULL;
-	     (np = of_find_node_by_name(np, "memory")) != NULL;) {
+	for_each_node_by_name(np, "memory") {
 		struct resource r;
 
 		/* We "know" that the Cell firmware only creates one entry
diff --git a/drivers/hwmon/adm1025.c b/drivers/hwmon/adm1025.c
index d3d0e8c..d6c767a 100644
--- a/drivers/hwmon/adm1025.c
+++ b/drivers/hwmon/adm1025.c
@@ -382,6 +382,9 @@
 	if (err)
 		return err;
 
+	if (val > 255)
+		return -EINVAL;
+
 	data->vrm = val;
 	return count;
 }
diff --git a/drivers/hwmon/adm1026.c b/drivers/hwmon/adm1026.c
index ca8430f..e67b9a5 100644
--- a/drivers/hwmon/adm1026.c
+++ b/drivers/hwmon/adm1026.c
@@ -1085,6 +1085,9 @@
 	if (err)
 		return err;
 
+	if (val > 255)
+		return -EINVAL;
+
 	data->vrm = val;
 	return count;
 }
diff --git a/drivers/hwmon/ads1015.c b/drivers/hwmon/ads1015.c
index 22e0c92..1265164 100644
--- a/drivers/hwmon/ads1015.c
+++ b/drivers/hwmon/ads1015.c
@@ -212,6 +212,7 @@
 				dev_err(&client->dev,
 					"invalid gain on %s\n",
 					node->full_name);
+				return -EINVAL;
 			}
 		}
 
@@ -222,6 +223,7 @@
 				dev_err(&client->dev,
 					"invalid data_rate on %s\n",
 					node->full_name);
+				return -EINVAL;
 			}
 		}
 
diff --git a/drivers/hwmon/asb100.c b/drivers/hwmon/asb100.c
index f960636..272fcc8 100644
--- a/drivers/hwmon/asb100.c
+++ b/drivers/hwmon/asb100.c
@@ -510,6 +510,10 @@
 	err = kstrtoul(buf, 10, &val);
 	if (err)
 		return err;
+
+	if (val > 255)
+		return -EINVAL;
+
 	data->vrm = val;
 	return count;
 }
diff --git a/drivers/hwmon/dme1737.c b/drivers/hwmon/dme1737.c
index 4ae3fff..bea0a34 100644
--- a/drivers/hwmon/dme1737.c
+++ b/drivers/hwmon/dme1737.c
@@ -247,8 +247,8 @@
 	u8  pwm_acz[3];
 	u8  pwm_freq[6];
 	u8  pwm_rr[2];
-	u8  zone_low[3];
-	u8  zone_abs[3];
+	s8  zone_low[3];
+	s8  zone_abs[3];
 	u8  zone_hyst[2];
 	u32 alarms;
 };
@@ -277,7 +277,7 @@
 	return (reg * nominal + (3 << (res - 3))) / (3 << (res - 2));
 }
 
-static inline int IN_TO_REG(int val, int nominal)
+static inline int IN_TO_REG(long val, int nominal)
 {
 	return clamp_val((val * 192 + nominal / 2) / nominal, 0, 255);
 }
@@ -293,7 +293,7 @@
 	return (reg * 1000) >> (res - 8);
 }
 
-static inline int TEMP_TO_REG(int val)
+static inline int TEMP_TO_REG(long val)
 {
 	return clamp_val((val < 0 ? val - 500 : val + 500) / 1000, -128, 127);
 }
@@ -308,7 +308,7 @@
 	return TEMP_RANGE[(reg >> 4) & 0x0f];
 }
 
-static int TEMP_RANGE_TO_REG(int val, int reg)
+static int TEMP_RANGE_TO_REG(long val, int reg)
 {
 	int i;
 
@@ -331,7 +331,7 @@
 	return (((ix == 1) ? reg : reg >> 4) & 0x0f) * 1000;
 }
 
-static inline int TEMP_HYST_TO_REG(int val, int ix, int reg)
+static inline int TEMP_HYST_TO_REG(long val, int ix, int reg)
 {
 	int hyst = clamp_val((val + 500) / 1000, 0, 15);
 
@@ -347,7 +347,7 @@
 		return (reg == 0 || reg == 0xffff) ? 0 : 90000 * 60 / reg;
 }
 
-static inline int FAN_TO_REG(int val, int tpc)
+static inline int FAN_TO_REG(long val, int tpc)
 {
 	if (tpc) {
 		return clamp_val(val / tpc, 0, 0xffff);
@@ -379,7 +379,7 @@
 	return (edge > 0) ? 1 << (edge - 1) : 0;
 }
 
-static inline int FAN_TYPE_TO_REG(int val, int reg)
+static inline int FAN_TYPE_TO_REG(long val, int reg)
 {
 	int edge = (val == 4) ? 3 : val;
 
@@ -402,7 +402,7 @@
 	return 1000 + i * 500;
 }
 
-static int FAN_MAX_TO_REG(int val)
+static int FAN_MAX_TO_REG(long val)
 {
 	int i;
 
@@ -460,7 +460,7 @@
 	return acz[(reg >> 5) & 0x07];
 }
 
-static inline int PWM_ACZ_TO_REG(int val, int reg)
+static inline int PWM_ACZ_TO_REG(long val, int reg)
 {
 	int acz = (val == 4) ? 2 : val - 1;
 
@@ -476,7 +476,7 @@
 	return PWM_FREQ[reg & 0x0f];
 }
 
-static int PWM_FREQ_TO_REG(int val, int reg)
+static int PWM_FREQ_TO_REG(long val, int reg)
 {
 	int i;
 
@@ -510,7 +510,7 @@
 	return (rr & 0x08) ? PWM_RR[rr & 0x07] : 0;
 }
 
-static int PWM_RR_TO_REG(int val, int ix, int reg)
+static int PWM_RR_TO_REG(long val, int ix, int reg)
 {
 	int i;
 
@@ -528,7 +528,7 @@
 	return PWM_RR_FROM_REG(reg, ix) ? 1 : 0;
 }
 
-static inline int PWM_RR_EN_TO_REG(int val, int ix, int reg)
+static inline int PWM_RR_EN_TO_REG(long val, int ix, int reg)
 {
 	int en = (ix == 1) ? 0x80 : 0x08;
 
@@ -1481,13 +1481,16 @@
 		       const char *buf, size_t count)
 {
 	struct dme1737_data *data = dev_get_drvdata(dev);
-	long val;
+	unsigned long val;
 	int err;
 
-	err = kstrtol(buf, 10, &val);
+	err = kstrtoul(buf, 10, &val);
 	if (err)
 		return err;
 
+	if (val > 255)
+		return -EINVAL;
+
 	data->vrm = val;
 	return count;
 }
diff --git a/drivers/hwmon/emc6w201.c b/drivers/hwmon/emc6w201.c
index e87da90..ada9071 100644
--- a/drivers/hwmon/emc6w201.c
+++ b/drivers/hwmon/emc6w201.c
@@ -252,12 +252,12 @@
 	if (err < 0)
 		return err;
 
-	val /= 1000;
+	val = DIV_ROUND_CLOSEST(val, 1000);
 	reg = (sf == min) ? EMC6W201_REG_TEMP_LOW(nr)
 			  : EMC6W201_REG_TEMP_HIGH(nr);
 
 	mutex_lock(&data->update_lock);
-	data->temp[sf][nr] = clamp_val(val, -127, 128);
+	data->temp[sf][nr] = clamp_val(val, -127, 127);
 	err = emc6w201_write8(client, reg, data->temp[sf][nr]);
 	mutex_unlock(&data->update_lock);
 
diff --git a/drivers/hwmon/hih6130.c b/drivers/hwmon/hih6130.c
index 0e01c4e..7b73d20 100644
--- a/drivers/hwmon/hih6130.c
+++ b/drivers/hwmon/hih6130.c
@@ -238,6 +238,9 @@
 	hih6130->client = client;
 	mutex_init(&hih6130->lock);
 
+	if (!i2c_check_functionality(client->adapter, I2C_FUNC_SMBUS_QUICK))
+		hih6130->write_length = 1;
+
 	hwmon_dev = devm_hwmon_device_register_with_groups(dev, client->name,
 							   hih6130,
 							   hih6130_groups);
diff --git a/drivers/hwmon/lm87.c b/drivers/hwmon/lm87.c
index ba1d83d..a5e2958 100644
--- a/drivers/hwmon/lm87.c
+++ b/drivers/hwmon/lm87.c
@@ -617,6 +617,10 @@
 	err = kstrtoul(buf, 10, &val);
 	if (err)
 		return err;
+
+	if (val > 255)
+		return -EINVAL;
+
 	data->vrm = val;
 	return count;
 }
diff --git a/drivers/hwmon/lm92.c b/drivers/hwmon/lm92.c
index d2060e2..cfaf70b 100644
--- a/drivers/hwmon/lm92.c
+++ b/drivers/hwmon/lm92.c
@@ -74,12 +74,9 @@
 	return reg / 8 * 625 / 10;
 }
 
-static inline s16 TEMP_TO_REG(int val)
+static inline s16 TEMP_TO_REG(long val)
 {
-	if (val <= -60000)
-		return -60000 * 10 / 625 * 8;
-	if (val >= 160000)
-		return 160000 * 10 / 625 * 8;
+	val = clamp_val(val, -60000, 160000);
 	return val * 10 / 625 * 8;
 }
 
@@ -206,10 +203,12 @@
 	if (err)
 		return err;
 
+	val = clamp_val(val, -120000, 220000);
 	mutex_lock(&data->update_lock);
-	data->temp[t_hyst] = TEMP_FROM_REG(data->temp[attr->index]) - val;
+	 data->temp[t_hyst] =
+		TEMP_TO_REG(TEMP_FROM_REG(data->temp[attr->index]) - val);
 	i2c_smbus_write_word_swapped(client, LM92_REG_TEMP_HYST,
-				     TEMP_TO_REG(data->temp[t_hyst]));
+				     data->temp[t_hyst]);
 	mutex_unlock(&data->update_lock);
 	return count;
 }
diff --git a/drivers/hwmon/pc87360.c b/drivers/hwmon/pc87360.c
index 988181e..145f674 100644
--- a/drivers/hwmon/pc87360.c
+++ b/drivers/hwmon/pc87360.c
@@ -615,6 +615,9 @@
 	if (err)
 		return err;
 
+	if (val > 255)
+		return -EINVAL;
+
 	data->vrm = val;
 	return count;
 }
diff --git a/drivers/hwmon/tmp103.c b/drivers/hwmon/tmp103.c
index c74d2da..e42964f 100644
--- a/drivers/hwmon/tmp103.c
+++ b/drivers/hwmon/tmp103.c
@@ -131,13 +131,6 @@
 	struct regmap *regmap;
 	int ret;
 
-	if (!i2c_check_functionality(client->adapter,
-				     I2C_FUNC_SMBUS_BYTE_DATA)) {
-		dev_err(&client->dev,
-			"adapter doesn't support SMBus byte transactions\n");
-		return -ENODEV;
-	}
-
 	regmap = devm_regmap_init_i2c(client, &tmp103_regmap_config);
 	if (IS_ERR(regmap)) {
 		dev_err(dev, "failed to allocate register map\n");
diff --git a/drivers/hwmon/vt1211.c b/drivers/hwmon/vt1211.c
index 344b22e..3ea57c3 100644
--- a/drivers/hwmon/vt1211.c
+++ b/drivers/hwmon/vt1211.c
@@ -879,6 +879,9 @@
 	if (err)
 		return err;
 
+	if (val > 255)
+		return -EINVAL;
+
 	data->vrm = val;
 
 	return count;
diff --git a/drivers/hwmon/w83627hf.c b/drivers/hwmon/w83627hf.c
index c1726be..2f55973 100644
--- a/drivers/hwmon/w83627hf.c
+++ b/drivers/hwmon/w83627hf.c
@@ -820,6 +820,9 @@
 	err = kstrtoul(buf, 10, &val);
 	if (err)
 		return err;
+
+	if (val > 255)
+		return -EINVAL;
 	data->vrm = val;
 
 	return count;
diff --git a/drivers/hwmon/w83791d.c b/drivers/hwmon/w83791d.c
index cb3765f..001df85 100644
--- a/drivers/hwmon/w83791d.c
+++ b/drivers/hwmon/w83791d.c
@@ -1181,6 +1181,9 @@
 	if (err)
 		return err;
 
+	if (val > 255)
+		return -EINVAL;
+
 	data->vrm = val;
 	return count;
 }
diff --git a/drivers/hwmon/w83793.c b/drivers/hwmon/w83793.c
index 9d63d71..816aa6c 100644
--- a/drivers/hwmon/w83793.c
+++ b/drivers/hwmon/w83793.c
@@ -353,6 +353,9 @@
 	if (err)
 		return err;
 
+	if (val > 255)
+		return -EINVAL;
+
 	data->vrm = val;
 	return count;
 }
diff --git a/drivers/hwspinlock/Kconfig b/drivers/hwspinlock/Kconfig
index 70637d2..3612cb5 100644
--- a/drivers/hwspinlock/Kconfig
+++ b/drivers/hwspinlock/Kconfig
@@ -10,7 +10,7 @@
 
 config HWSPINLOCK_OMAP
 	tristate "OMAP Hardware Spinlock device"
-	depends on ARCH_OMAP4 || SOC_OMAP5
+	depends on ARCH_OMAP4 || SOC_OMAP5 || SOC_DRA7XX || SOC_AM33XX || SOC_AM43XX
 	select HWSPINLOCK
 	help
 	  Say y here to support the OMAP Hardware Spinlock device (firstly
diff --git a/drivers/hwspinlock/omap_hwspinlock.c b/drivers/hwspinlock/omap_hwspinlock.c
index 292869c..c1e2cd4 100644
--- a/drivers/hwspinlock/omap_hwspinlock.c
+++ b/drivers/hwspinlock/omap_hwspinlock.c
@@ -98,10 +98,29 @@
 	if (!io_base)
 		return -ENOMEM;
 
+	/*
+	 * make sure the module is enabled and clocked before reading
+	 * the module SYSSTATUS register
+	 */
+	pm_runtime_enable(&pdev->dev);
+	ret = pm_runtime_get_sync(&pdev->dev);
+	if (ret < 0) {
+		pm_runtime_put_noidle(&pdev->dev);
+		goto iounmap_base;
+	}
+
 	/* Determine number of locks */
 	i = readl(io_base + SYSSTATUS_OFFSET);
 	i >>= SPINLOCK_NUMLOCKS_BIT_OFFSET;
 
+	/*
+	 * runtime PM will make sure the clock of this module is
+	 * enabled again iff at least one lock is requested
+	 */
+	ret = pm_runtime_put(&pdev->dev);
+	if (ret < 0)
+		goto iounmap_base;
+
 	/* one of the four lsb's must be set, and nothing else */
 	if (hweight_long(i & 0xf) != 1 || i > 8) {
 		ret = -EINVAL;
@@ -121,12 +140,6 @@
 	for (i = 0, hwlock = &bank->lock[0]; i < num_locks; i++, hwlock++)
 		hwlock->priv = io_base + LOCK_BASE_OFFSET + sizeof(u32) * i;
 
-	/*
-	 * runtime PM will make sure the clock of this module is
-	 * enabled iff at least one lock is requested
-	 */
-	pm_runtime_enable(&pdev->dev);
-
 	ret = hwspin_lock_register(bank, &pdev->dev, &omap_hwspinlock_ops,
 						pdata->base_id, num_locks);
 	if (ret)
@@ -135,9 +148,9 @@
 	return 0;
 
 reg_fail:
-	pm_runtime_disable(&pdev->dev);
 	kfree(bank);
 iounmap_base:
+	pm_runtime_disable(&pdev->dev);
 	iounmap(io_base);
 	return ret;
 }
diff --git a/drivers/infiniband/core/agent.c b/drivers/infiniband/core/agent.c
index 2bc7f5a..f6d2961 100644
--- a/drivers/infiniband/core/agent.c
+++ b/drivers/infiniband/core/agent.c
@@ -94,14 +94,14 @@
 		port_priv = ib_get_agent_port(device, port_num);
 
 	if (!port_priv) {
-		printk(KERN_ERR SPFX "Unable to find port agent\n");
+		dev_err(&device->dev, "Unable to find port agent\n");
 		return;
 	}
 
 	agent = port_priv->agent[qpn];
 	ah = ib_create_ah_from_wc(agent->qp->pd, wc, grh, port_num);
 	if (IS_ERR(ah)) {
-		printk(KERN_ERR SPFX "ib_create_ah_from_wc error %ld\n",
+		dev_err(&device->dev, "ib_create_ah_from_wc error %ld\n",
 			PTR_ERR(ah));
 		return;
 	}
@@ -110,7 +110,7 @@
 				      IB_MGMT_MAD_HDR, IB_MGMT_MAD_DATA,
 				      GFP_KERNEL);
 	if (IS_ERR(send_buf)) {
-		printk(KERN_ERR SPFX "ib_create_send_mad error\n");
+		dev_err(&device->dev, "ib_create_send_mad error\n");
 		goto err1;
 	}
 
@@ -125,7 +125,7 @@
 	}
 
 	if (ib_post_send_mad(send_buf, NULL)) {
-		printk(KERN_ERR SPFX "ib_post_send_mad error\n");
+		dev_err(&device->dev, "ib_post_send_mad error\n");
 		goto err2;
 	}
 	return;
@@ -151,7 +151,7 @@
 	/* Create new device info */
 	port_priv = kzalloc(sizeof *port_priv, GFP_KERNEL);
 	if (!port_priv) {
-		printk(KERN_ERR SPFX "No memory for ib_agent_port_private\n");
+		dev_err(&device->dev, "No memory for ib_agent_port_private\n");
 		ret = -ENOMEM;
 		goto error1;
 	}
@@ -161,7 +161,7 @@
 		port_priv->agent[0] = ib_register_mad_agent(device, port_num,
 							    IB_QPT_SMI, NULL, 0,
 							    &agent_send_handler,
-							    NULL, NULL);
+							    NULL, NULL, 0);
 		if (IS_ERR(port_priv->agent[0])) {
 			ret = PTR_ERR(port_priv->agent[0]);
 			goto error2;
@@ -172,7 +172,7 @@
 	port_priv->agent[1] = ib_register_mad_agent(device, port_num,
 						    IB_QPT_GSI, NULL, 0,
 						    &agent_send_handler,
-						    NULL, NULL);
+						    NULL, NULL, 0);
 	if (IS_ERR(port_priv->agent[1])) {
 		ret = PTR_ERR(port_priv->agent[1]);
 		goto error3;
@@ -202,7 +202,7 @@
 	port_priv = __ib_get_agent_port(device, port_num);
 	if (port_priv == NULL) {
 		spin_unlock_irqrestore(&ib_agent_port_list_lock, flags);
-		printk(KERN_ERR SPFX "Port %d not found\n", port_num);
+		dev_err(&device->dev, "Port %d not found\n", port_num);
 		return -ENODEV;
 	}
 	list_del(&port_priv->port_list);
diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c
index c323917..e28a494 100644
--- a/drivers/infiniband/core/cm.c
+++ b/drivers/infiniband/core/cm.c
@@ -3753,7 +3753,7 @@
 	struct cm_port *port;
 	struct ib_mad_reg_req reg_req = {
 		.mgmt_class = IB_MGMT_CLASS_CM,
-		.mgmt_class_version = IB_CM_CLASS_VERSION
+		.mgmt_class_version = IB_CM_CLASS_VERSION,
 	};
 	struct ib_port_modify port_modify = {
 		.set_port_cap_mask = IB_PORT_CM_SUP
@@ -3801,7 +3801,8 @@
 							0,
 							cm_send_handler,
 							cm_recv_handler,
-							port);
+							port,
+							0);
 		if (IS_ERR(port->mad_agent))
 			goto error2;
 
diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c
index 3d2e489..ff9163d 100644
--- a/drivers/infiniband/core/iwcm.c
+++ b/drivers/infiniband/core/iwcm.c
@@ -46,6 +46,7 @@
 #include <linux/completion.h>
 #include <linux/slab.h>
 #include <linux/module.h>
+#include <linux/sysctl.h>
 
 #include <rdma/iw_cm.h>
 #include <rdma/ib_addr.h>
@@ -65,6 +66,20 @@
 	struct list_head free_list;
 };
 
+static unsigned int default_backlog = 256;
+
+static struct ctl_table_header *iwcm_ctl_table_hdr;
+static struct ctl_table iwcm_ctl_table[] = {
+	{
+		.procname	= "default_backlog",
+		.data		= &default_backlog,
+		.maxlen		= sizeof(default_backlog),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec,
+	},
+	{ }
+};
+
 /*
  * The following services provide a mechanism for pre-allocating iwcm_work
  * elements.  The design pre-allocates them  based on the cm_id type:
@@ -425,6 +440,9 @@
 
 	cm_id_priv = container_of(cm_id, struct iwcm_id_private, id);
 
+	if (!backlog)
+		backlog = default_backlog;
+
 	ret = alloc_work_entries(cm_id_priv, backlog);
 	if (ret)
 		return ret;
@@ -1030,11 +1048,20 @@
 	if (!iwcm_wq)
 		return -ENOMEM;
 
+	iwcm_ctl_table_hdr = register_net_sysctl(&init_net, "net/iw_cm",
+						 iwcm_ctl_table);
+	if (!iwcm_ctl_table_hdr) {
+		pr_err("iw_cm: couldn't register sysctl paths\n");
+		destroy_workqueue(iwcm_wq);
+		return -ENOMEM;
+	}
+
 	return 0;
 }
 
 static void __exit iw_cm_cleanup(void)
 {
+	unregister_net_sysctl_table(iwcm_ctl_table_hdr);
 	destroy_workqueue(iwcm_wq);
 }
 
diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c
index ab31f13..74c30f4 100644
--- a/drivers/infiniband/core/mad.c
+++ b/drivers/infiniband/core/mad.c
@@ -33,6 +33,9 @@
  * SOFTWARE.
  *
  */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
 #include <linux/dma-mapping.h>
 #include <linux/slab.h>
 #include <linux/module.h>
@@ -195,7 +198,8 @@
 					   u8 rmpp_version,
 					   ib_mad_send_handler send_handler,
 					   ib_mad_recv_handler recv_handler,
-					   void *context)
+					   void *context,
+					   u32 registration_flags)
 {
 	struct ib_mad_port_private *port_priv;
 	struct ib_mad_agent *ret = ERR_PTR(-EINVAL);
@@ -211,68 +215,109 @@
 
 	/* Validate parameters */
 	qpn = get_spl_qp_index(qp_type);
-	if (qpn == -1)
+	if (qpn == -1) {
+		dev_notice(&device->dev,
+			   "ib_register_mad_agent: invalid QP Type %d\n",
+			   qp_type);
 		goto error1;
+	}
 
-	if (rmpp_version && rmpp_version != IB_MGMT_RMPP_VERSION)
+	if (rmpp_version && rmpp_version != IB_MGMT_RMPP_VERSION) {
+		dev_notice(&device->dev,
+			   "ib_register_mad_agent: invalid RMPP Version %u\n",
+			   rmpp_version);
 		goto error1;
+	}
 
 	/* Validate MAD registration request if supplied */
 	if (mad_reg_req) {
-		if (mad_reg_req->mgmt_class_version >= MAX_MGMT_VERSION)
+		if (mad_reg_req->mgmt_class_version >= MAX_MGMT_VERSION) {
+			dev_notice(&device->dev,
+				   "ib_register_mad_agent: invalid Class Version %u\n",
+				   mad_reg_req->mgmt_class_version);
 			goto error1;
-		if (!recv_handler)
+		}
+		if (!recv_handler) {
+			dev_notice(&device->dev,
+				   "ib_register_mad_agent: no recv_handler\n");
 			goto error1;
+		}
 		if (mad_reg_req->mgmt_class >= MAX_MGMT_CLASS) {
 			/*
 			 * IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE is the only
 			 * one in this range currently allowed
 			 */
 			if (mad_reg_req->mgmt_class !=
-			    IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)
+			    IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) {
+				dev_notice(&device->dev,
+					   "ib_register_mad_agent: Invalid Mgmt Class 0x%x\n",
+					   mad_reg_req->mgmt_class);
 				goto error1;
+			}
 		} else if (mad_reg_req->mgmt_class == 0) {
 			/*
 			 * Class 0 is reserved in IBA and is used for
 			 * aliasing of IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE
 			 */
+			dev_notice(&device->dev,
+				   "ib_register_mad_agent: Invalid Mgmt Class 0\n");
 			goto error1;
 		} else if (is_vendor_class(mad_reg_req->mgmt_class)) {
 			/*
 			 * If class is in "new" vendor range,
 			 * ensure supplied OUI is not zero
 			 */
-			if (!is_vendor_oui(mad_reg_req->oui))
+			if (!is_vendor_oui(mad_reg_req->oui)) {
+				dev_notice(&device->dev,
+					   "ib_register_mad_agent: No OUI specified for class 0x%x\n",
+					   mad_reg_req->mgmt_class);
 				goto error1;
+			}
 		}
 		/* Make sure class supplied is consistent with RMPP */
 		if (!ib_is_mad_class_rmpp(mad_reg_req->mgmt_class)) {
-			if (rmpp_version)
+			if (rmpp_version) {
+				dev_notice(&device->dev,
+					   "ib_register_mad_agent: RMPP version for non-RMPP class 0x%x\n",
+					   mad_reg_req->mgmt_class);
 				goto error1;
+			}
 		}
+
 		/* Make sure class supplied is consistent with QP type */
 		if (qp_type == IB_QPT_SMI) {
 			if ((mad_reg_req->mgmt_class !=
 					IB_MGMT_CLASS_SUBN_LID_ROUTED) &&
 			    (mad_reg_req->mgmt_class !=
-					IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE))
+					IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)) {
+				dev_notice(&device->dev,
+					   "ib_register_mad_agent: Invalid SM QP type: class 0x%x\n",
+					   mad_reg_req->mgmt_class);
 				goto error1;
+			}
 		} else {
 			if ((mad_reg_req->mgmt_class ==
 					IB_MGMT_CLASS_SUBN_LID_ROUTED) ||
 			    (mad_reg_req->mgmt_class ==
-					IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE))
+					IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)) {
+				dev_notice(&device->dev,
+					   "ib_register_mad_agent: Invalid GS QP type: class 0x%x\n",
+					   mad_reg_req->mgmt_class);
 				goto error1;
+			}
 		}
 	} else {
 		/* No registration request supplied */
 		if (!send_handler)
 			goto error1;
+		if (registration_flags & IB_MAD_USER_RMPP)
+			goto error1;
 	}
 
 	/* Validate device and port */
 	port_priv = ib_get_mad_port(device, port_num);
 	if (!port_priv) {
+		dev_notice(&device->dev, "ib_register_mad_agent: Invalid port\n");
 		ret = ERR_PTR(-ENODEV);
 		goto error1;
 	}
@@ -280,6 +325,8 @@
 	/* Verify the QP requested is supported.  For example, Ethernet devices
 	 * will not have QP0 */
 	if (!port_priv->qp_info[qpn].qp) {
+		dev_notice(&device->dev,
+			   "ib_register_mad_agent: QP %d not supported\n", qpn);
 		ret = ERR_PTR(-EPROTONOSUPPORT);
 		goto error1;
 	}
@@ -316,6 +363,7 @@
 	mad_agent_priv->agent.context = context;
 	mad_agent_priv->agent.qp = port_priv->qp_info[qpn].qp;
 	mad_agent_priv->agent.port_num = port_num;
+	mad_agent_priv->agent.flags = registration_flags;
 	spin_lock_init(&mad_agent_priv->lock);
 	INIT_LIST_HEAD(&mad_agent_priv->send_list);
 	INIT_LIST_HEAD(&mad_agent_priv->wait_list);
@@ -706,7 +754,7 @@
 	     smi_handle_dr_smp_send(smp, device->node_type, port_num) ==
 	     IB_SMI_DISCARD) {
 		ret = -EINVAL;
-		printk(KERN_ERR PFX "Invalid directed route\n");
+		dev_err(&device->dev, "Invalid directed route\n");
 		goto out;
 	}
 
@@ -718,7 +766,7 @@
 	local = kmalloc(sizeof *local, GFP_ATOMIC);
 	if (!local) {
 		ret = -ENOMEM;
-		printk(KERN_ERR PFX "No memory for ib_mad_local_private\n");
+		dev_err(&device->dev, "No memory for ib_mad_local_private\n");
 		goto out;
 	}
 	local->mad_priv = NULL;
@@ -726,7 +774,7 @@
 	mad_priv = kmem_cache_alloc(ib_mad_cache, GFP_ATOMIC);
 	if (!mad_priv) {
 		ret = -ENOMEM;
-		printk(KERN_ERR PFX "No memory for local response MAD\n");
+		dev_err(&device->dev, "No memory for local response MAD\n");
 		kfree(local);
 		goto out;
 	}
@@ -837,9 +885,9 @@
 	for (left = send_buf->data_len + pad; left > 0; left -= seg_size) {
 		seg = kmalloc(sizeof (*seg) + seg_size, gfp_mask);
 		if (!seg) {
-			printk(KERN_ERR "alloc_send_rmpp_segs: RMPP mem "
-			       "alloc failed for len %zd, gfp %#x\n",
-			       sizeof (*seg) + seg_size, gfp_mask);
+			dev_err(&send_buf->mad_agent->device->dev,
+				"alloc_send_rmpp_segs: RMPP mem alloc failed for len %zd, gfp %#x\n",
+				sizeof (*seg) + seg_size, gfp_mask);
 			free_send_rmpp_list(send_wr);
 			return -ENOMEM;
 		}
@@ -862,6 +910,12 @@
 	return 0;
 }
 
+int ib_mad_kernel_rmpp_agent(struct ib_mad_agent *agent)
+{
+	return agent->rmpp_version && !(agent->flags & IB_MAD_USER_RMPP);
+}
+EXPORT_SYMBOL(ib_mad_kernel_rmpp_agent);
+
 struct ib_mad_send_buf * ib_create_send_mad(struct ib_mad_agent *mad_agent,
 					    u32 remote_qpn, u16 pkey_index,
 					    int rmpp_active,
@@ -878,10 +932,12 @@
 	pad = get_pad_size(hdr_len, data_len);
 	message_size = hdr_len + data_len + pad;
 
-	if ((!mad_agent->rmpp_version &&
-	     (rmpp_active || message_size > sizeof(struct ib_mad))) ||
-	    (!rmpp_active && message_size > sizeof(struct ib_mad)))
-		return ERR_PTR(-EINVAL);
+	if (ib_mad_kernel_rmpp_agent(mad_agent)) {
+		if (!rmpp_active && message_size > sizeof(struct ib_mad))
+			return ERR_PTR(-EINVAL);
+	} else
+		if (rmpp_active || message_size > sizeof(struct ib_mad))
+			return ERR_PTR(-EINVAL);
 
 	size = rmpp_active ? hdr_len : sizeof(struct ib_mad);
 	buf = kzalloc(sizeof *mad_send_wr + size, gfp_mask);
@@ -1135,7 +1191,7 @@
 			      &mad_agent_priv->send_list);
 		spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
 
-		if (mad_agent_priv->agent.rmpp_version) {
+		if (ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent)) {
 			ret = ib_send_rmpp_mad(mad_send_wr);
 			if (ret >= 0 && ret != IB_RMPP_RESULT_CONSUMED)
 				ret = ib_send_mad(mad_send_wr);
@@ -1199,7 +1255,8 @@
 int ib_process_mad_wc(struct ib_mad_agent *mad_agent,
 		      struct ib_wc *wc)
 {
-	printk(KERN_ERR PFX "ib_process_mad_wc() not implemented yet\n");
+	dev_err(&mad_agent->device->dev,
+		"ib_process_mad_wc() not implemented yet\n");
 	return 0;
 }
 EXPORT_SYMBOL(ib_process_mad_wc);
@@ -1211,7 +1268,7 @@
 
 	for_each_set_bit(i, mad_reg_req->method_mask, IB_MGMT_MAX_METHODS) {
 		if ((*method)->agent[i]) {
-			printk(KERN_ERR PFX "Method %d already in use\n", i);
+			pr_err("Method %d already in use\n", i);
 			return -EINVAL;
 		}
 	}
@@ -1223,8 +1280,7 @@
 	/* Allocate management method table */
 	*method = kzalloc(sizeof **method, GFP_ATOMIC);
 	if (!*method) {
-		printk(KERN_ERR PFX "No memory for "
-		       "ib_mad_mgmt_method_table\n");
+		pr_err("No memory for ib_mad_mgmt_method_table\n");
 		return -ENOMEM;
 	}
 
@@ -1319,8 +1375,8 @@
 		/* Allocate management class table for "new" class version */
 		*class = kzalloc(sizeof **class, GFP_ATOMIC);
 		if (!*class) {
-			printk(KERN_ERR PFX "No memory for "
-			       "ib_mad_mgmt_class_table\n");
+			dev_err(&agent_priv->agent.device->dev,
+				"No memory for ib_mad_mgmt_class_table\n");
 			ret = -ENOMEM;
 			goto error1;
 		}
@@ -1386,8 +1442,8 @@
 		/* Allocate mgmt vendor class table for "new" class version */
 		vendor = kzalloc(sizeof *vendor, GFP_ATOMIC);
 		if (!vendor) {
-			printk(KERN_ERR PFX "No memory for "
-			       "ib_mad_mgmt_vendor_class_table\n");
+			dev_err(&agent_priv->agent.device->dev,
+				"No memory for ib_mad_mgmt_vendor_class_table\n");
 			goto error1;
 		}
 
@@ -1397,8 +1453,8 @@
 		/* Allocate table for this management vendor class */
 		vendor_class = kzalloc(sizeof *vendor_class, GFP_ATOMIC);
 		if (!vendor_class) {
-			printk(KERN_ERR PFX "No memory for "
-			       "ib_mad_mgmt_vendor_class\n");
+			dev_err(&agent_priv->agent.device->dev,
+				"No memory for ib_mad_mgmt_vendor_class\n");
 			goto error2;
 		}
 
@@ -1429,7 +1485,7 @@
 			goto check_in_use;
 		}
 	}
-	printk(KERN_ERR PFX "All OUI slots in use\n");
+	dev_err(&agent_priv->agent.device->dev, "All OUI slots in use\n");
 	goto error3;
 
 check_in_use:
@@ -1640,9 +1696,9 @@
 		if (mad_agent->agent.recv_handler)
 			atomic_inc(&mad_agent->refcount);
 		else {
-			printk(KERN_NOTICE PFX "No receive handler for client "
-			       "%p on port %d\n",
-			       &mad_agent->agent, port_priv->port_num);
+			dev_notice(&port_priv->device->dev,
+				   "No receive handler for client %p on port %d\n",
+				   &mad_agent->agent, port_priv->port_num);
 			mad_agent = NULL;
 		}
 	}
@@ -1658,8 +1714,8 @@
 
 	/* Make sure MAD base version is understood */
 	if (mad->mad_hdr.base_version != IB_MGMT_BASE_VERSION) {
-		printk(KERN_ERR PFX "MAD received with unsupported base "
-		       "version %d\n", mad->mad_hdr.base_version);
+		pr_err("MAD received with unsupported base version %d\n",
+			mad->mad_hdr.base_version);
 		goto out;
 	}
 
@@ -1685,6 +1741,7 @@
 
 	rmpp_mad = (struct ib_rmpp_mad *)mad_hdr;
 	return !mad_agent_priv->agent.rmpp_version ||
+		!ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent) ||
 		!(ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) &
 				    IB_MGMT_RMPP_FLAG_ACTIVE) ||
 		(rmpp_mad->rmpp_hdr.rmpp_type == IB_MGMT_RMPP_TYPE_DATA);
@@ -1812,7 +1869,7 @@
 
 	INIT_LIST_HEAD(&mad_recv_wc->rmpp_list);
 	list_add(&mad_recv_wc->recv_buf.list, &mad_recv_wc->rmpp_list);
-	if (mad_agent_priv->agent.rmpp_version) {
+	if (ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent)) {
 		mad_recv_wc = ib_process_rmpp_recv_wc(mad_agent_priv,
 						      mad_recv_wc);
 		if (!mad_recv_wc) {
@@ -1827,23 +1884,39 @@
 		mad_send_wr = ib_find_send_mad(mad_agent_priv, mad_recv_wc);
 		if (!mad_send_wr) {
 			spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
-			ib_free_recv_mad(mad_recv_wc);
-			deref_mad_agent(mad_agent_priv);
-			return;
+			if (!ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent)
+			   && ib_is_mad_class_rmpp(mad_recv_wc->recv_buf.mad->mad_hdr.mgmt_class)
+			   && (ib_get_rmpp_flags(&((struct ib_rmpp_mad *)mad_recv_wc->recv_buf.mad)->rmpp_hdr)
+					& IB_MGMT_RMPP_FLAG_ACTIVE)) {
+				/* user rmpp is in effect
+				 * and this is an active RMPP MAD
+				 */
+				mad_recv_wc->wc->wr_id = 0;
+				mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent,
+								   mad_recv_wc);
+				atomic_dec(&mad_agent_priv->refcount);
+			} else {
+				/* not user rmpp, revert to normal behavior and
+				 * drop the mad */
+				ib_free_recv_mad(mad_recv_wc);
+				deref_mad_agent(mad_agent_priv);
+				return;
+			}
+		} else {
+			ib_mark_mad_done(mad_send_wr);
+			spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
+
+			/* Defined behavior is to complete response before request */
+			mad_recv_wc->wc->wr_id = (unsigned long) &mad_send_wr->send_buf;
+			mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent,
+							   mad_recv_wc);
+			atomic_dec(&mad_agent_priv->refcount);
+
+			mad_send_wc.status = IB_WC_SUCCESS;
+			mad_send_wc.vendor_err = 0;
+			mad_send_wc.send_buf = &mad_send_wr->send_buf;
+			ib_mad_complete_send_wr(mad_send_wr, &mad_send_wc);
 		}
-		ib_mark_mad_done(mad_send_wr);
-		spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
-
-		/* Defined behavior is to complete response before request */
-		mad_recv_wc->wc->wr_id = (unsigned long) &mad_send_wr->send_buf;
-		mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent,
-						   mad_recv_wc);
-		atomic_dec(&mad_agent_priv->refcount);
-
-		mad_send_wc.status = IB_WC_SUCCESS;
-		mad_send_wc.vendor_err = 0;
-		mad_send_wc.send_buf = &mad_send_wr->send_buf;
-		ib_mad_complete_send_wr(mad_send_wr, &mad_send_wc);
 	} else {
 		mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent,
 						   mad_recv_wc);
@@ -1911,8 +1984,8 @@
 
 	response = kmem_cache_alloc(ib_mad_cache, GFP_KERNEL);
 	if (!response) {
-		printk(KERN_ERR PFX "ib_mad_recv_done_handler no memory "
-		       "for response buffer\n");
+		dev_err(&port_priv->device->dev,
+			"ib_mad_recv_done_handler no memory for response buffer\n");
 		goto out;
 	}
 
@@ -2083,7 +2156,7 @@
 
 	mad_agent_priv = mad_send_wr->mad_agent_priv;
 	spin_lock_irqsave(&mad_agent_priv->lock, flags);
-	if (mad_agent_priv->agent.rmpp_version) {
+	if (ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent)) {
 		ret = ib_process_rmpp_send_wc(mad_send_wr, mad_send_wc);
 		if (ret == IB_RMPP_RESULT_CONSUMED)
 			goto done;
@@ -2176,7 +2249,8 @@
 		ret = ib_post_send(qp_info->qp, &queued_send_wr->send_wr,
 				   &bad_send_wr);
 		if (ret) {
-			printk(KERN_ERR PFX "ib_post_send failed: %d\n", ret);
+			dev_err(&port_priv->device->dev,
+				"ib_post_send failed: %d\n", ret);
 			mad_send_wr = queued_send_wr;
 			wc->status = IB_WC_LOC_QP_OP_ERR;
 			goto retry;
@@ -2248,8 +2322,9 @@
 					   IB_QP_STATE | IB_QP_CUR_STATE);
 			kfree(attr);
 			if (ret)
-				printk(KERN_ERR PFX "mad_error_handler - "
-				       "ib_modify_qp to RTS : %d\n", ret);
+				dev_err(&port_priv->device->dev,
+					"mad_error_handler - ib_modify_qp to RTS : %d\n",
+					ret);
 			else
 				mark_sends_for_retry(qp_info);
 		}
@@ -2408,7 +2483,8 @@
 		if (local->mad_priv) {
 			recv_mad_agent = local->recv_mad_agent;
 			if (!recv_mad_agent) {
-				printk(KERN_ERR PFX "No receive MAD agent for local completion\n");
+				dev_err(&mad_agent_priv->agent.device->dev,
+					"No receive MAD agent for local completion\n");
 				free_mad = 1;
 				goto local_send_completion;
 			}
@@ -2476,7 +2552,7 @@
 
 	mad_send_wr->timeout = msecs_to_jiffies(mad_send_wr->send_buf.timeout_ms);
 
-	if (mad_send_wr->mad_agent_priv->agent.rmpp_version) {
+	if (ib_mad_kernel_rmpp_agent(&mad_send_wr->mad_agent_priv->agent)) {
 		ret = ib_retry_rmpp(mad_send_wr);
 		switch (ret) {
 		case IB_RMPP_RESULT_UNHANDLED:
@@ -2589,7 +2665,8 @@
 		} else {
 			mad_priv = kmem_cache_alloc(ib_mad_cache, GFP_KERNEL);
 			if (!mad_priv) {
-				printk(KERN_ERR PFX "No memory for receive buffer\n");
+				dev_err(&qp_info->port_priv->device->dev,
+					"No memory for receive buffer\n");
 				ret = -ENOMEM;
 				break;
 			}
@@ -2625,7 +2702,8 @@
 					      sizeof mad_priv->header,
 					    DMA_FROM_DEVICE);
 			kmem_cache_free(ib_mad_cache, mad_priv);
-			printk(KERN_ERR PFX "ib_post_recv failed: %d\n", ret);
+			dev_err(&qp_info->port_priv->device->dev,
+				"ib_post_recv failed: %d\n", ret);
 			break;
 		}
 	} while (post);
@@ -2681,7 +2759,8 @@
 
 	attr = kmalloc(sizeof *attr, GFP_KERNEL);
 	if (!attr) {
-		printk(KERN_ERR PFX "Couldn't kmalloc ib_qp_attr\n");
+		dev_err(&port_priv->device->dev,
+			"Couldn't kmalloc ib_qp_attr\n");
 		return -ENOMEM;
 	}
 
@@ -2705,16 +2784,18 @@
 		ret = ib_modify_qp(qp, attr, IB_QP_STATE |
 					     IB_QP_PKEY_INDEX | IB_QP_QKEY);
 		if (ret) {
-			printk(KERN_ERR PFX "Couldn't change QP%d state to "
-			       "INIT: %d\n", i, ret);
+			dev_err(&port_priv->device->dev,
+				"Couldn't change QP%d state to INIT: %d\n",
+				i, ret);
 			goto out;
 		}
 
 		attr->qp_state = IB_QPS_RTR;
 		ret = ib_modify_qp(qp, attr, IB_QP_STATE);
 		if (ret) {
-			printk(KERN_ERR PFX "Couldn't change QP%d state to "
-			       "RTR: %d\n", i, ret);
+			dev_err(&port_priv->device->dev,
+				"Couldn't change QP%d state to RTR: %d\n",
+				i, ret);
 			goto out;
 		}
 
@@ -2722,16 +2803,18 @@
 		attr->sq_psn = IB_MAD_SEND_Q_PSN;
 		ret = ib_modify_qp(qp, attr, IB_QP_STATE | IB_QP_SQ_PSN);
 		if (ret) {
-			printk(KERN_ERR PFX "Couldn't change QP%d state to "
-			       "RTS: %d\n", i, ret);
+			dev_err(&port_priv->device->dev,
+				"Couldn't change QP%d state to RTS: %d\n",
+				i, ret);
 			goto out;
 		}
 	}
 
 	ret = ib_req_notify_cq(port_priv->cq, IB_CQ_NEXT_COMP);
 	if (ret) {
-		printk(KERN_ERR PFX "Failed to request completion "
-		       "notification: %d\n", ret);
+		dev_err(&port_priv->device->dev,
+			"Failed to request completion notification: %d\n",
+			ret);
 		goto out;
 	}
 
@@ -2741,7 +2824,8 @@
 
 		ret = ib_mad_post_receive_mads(&port_priv->qp_info[i], NULL);
 		if (ret) {
-			printk(KERN_ERR PFX "Couldn't post receive WRs\n");
+			dev_err(&port_priv->device->dev,
+				"Couldn't post receive WRs\n");
 			goto out;
 		}
 	}
@@ -2755,7 +2839,8 @@
 	struct ib_mad_qp_info	*qp_info = qp_context;
 
 	/* It's worse than that! He's dead, Jim! */
-	printk(KERN_ERR PFX "Fatal error (%d) on MAD QP (%d)\n",
+	dev_err(&qp_info->port_priv->device->dev,
+		"Fatal error (%d) on MAD QP (%d)\n",
 		event->event, qp_info->qp->qp_num);
 }
 
@@ -2801,8 +2886,9 @@
 	qp_init_attr.event_handler = qp_event_handler;
 	qp_info->qp = ib_create_qp(qp_info->port_priv->pd, &qp_init_attr);
 	if (IS_ERR(qp_info->qp)) {
-		printk(KERN_ERR PFX "Couldn't create ib_mad QP%d\n",
-		       get_spl_qp_index(qp_type));
+		dev_err(&qp_info->port_priv->device->dev,
+			"Couldn't create ib_mad QP%d\n",
+			get_spl_qp_index(qp_type));
 		ret = PTR_ERR(qp_info->qp);
 		goto error;
 	}
@@ -2840,7 +2926,7 @@
 	/* Create new device info */
 	port_priv = kzalloc(sizeof *port_priv, GFP_KERNEL);
 	if (!port_priv) {
-		printk(KERN_ERR PFX "No memory for ib_mad_port_private\n");
+		dev_err(&device->dev, "No memory for ib_mad_port_private\n");
 		return -ENOMEM;
 	}
 
@@ -2860,21 +2946,21 @@
 				     ib_mad_thread_completion_handler,
 				     NULL, port_priv, cq_size, 0);
 	if (IS_ERR(port_priv->cq)) {
-		printk(KERN_ERR PFX "Couldn't create ib_mad CQ\n");
+		dev_err(&device->dev, "Couldn't create ib_mad CQ\n");
 		ret = PTR_ERR(port_priv->cq);
 		goto error3;
 	}
 
 	port_priv->pd = ib_alloc_pd(device);
 	if (IS_ERR(port_priv->pd)) {
-		printk(KERN_ERR PFX "Couldn't create ib_mad PD\n");
+		dev_err(&device->dev, "Couldn't create ib_mad PD\n");
 		ret = PTR_ERR(port_priv->pd);
 		goto error4;
 	}
 
 	port_priv->mr = ib_get_dma_mr(port_priv->pd, IB_ACCESS_LOCAL_WRITE);
 	if (IS_ERR(port_priv->mr)) {
-		printk(KERN_ERR PFX "Couldn't get ib_mad DMA MR\n");
+		dev_err(&device->dev, "Couldn't get ib_mad DMA MR\n");
 		ret = PTR_ERR(port_priv->mr);
 		goto error5;
 	}
@@ -2902,7 +2988,7 @@
 
 	ret = ib_mad_port_start(port_priv);
 	if (ret) {
-		printk(KERN_ERR PFX "Couldn't start port\n");
+		dev_err(&device->dev, "Couldn't start port\n");
 		goto error9;
 	}
 
@@ -2946,7 +3032,7 @@
 	port_priv = __ib_get_mad_port(device, port_num);
 	if (port_priv == NULL) {
 		spin_unlock_irqrestore(&ib_mad_port_list_lock, flags);
-		printk(KERN_ERR PFX "Port %d not found\n", port_num);
+		dev_err(&device->dev, "Port %d not found\n", port_num);
 		return -ENODEV;
 	}
 	list_del_init(&port_priv->port_list);
@@ -2984,14 +3070,12 @@
 
 	for (i = start; i <= end; i++) {
 		if (ib_mad_port_open(device, i)) {
-			printk(KERN_ERR PFX "Couldn't open %s port %d\n",
-			       device->name, i);
+			dev_err(&device->dev, "Couldn't open port %d\n", i);
 			goto error;
 		}
 		if (ib_agent_port_open(device, i)) {
-			printk(KERN_ERR PFX "Couldn't open %s port %d "
-			       "for agents\n",
-			       device->name, i);
+			dev_err(&device->dev,
+				"Couldn't open port %d for agents\n", i);
 			goto error_agent;
 		}
 	}
@@ -2999,20 +3083,17 @@
 
 error_agent:
 	if (ib_mad_port_close(device, i))
-		printk(KERN_ERR PFX "Couldn't close %s port %d\n",
-		       device->name, i);
+		dev_err(&device->dev, "Couldn't close port %d\n", i);
 
 error:
 	i--;
 
 	while (i >= start) {
 		if (ib_agent_port_close(device, i))
-			printk(KERN_ERR PFX "Couldn't close %s port %d "
-			       "for agents\n",
-			       device->name, i);
+			dev_err(&device->dev,
+				"Couldn't close port %d for agents\n", i);
 		if (ib_mad_port_close(device, i))
-			printk(KERN_ERR PFX "Couldn't close %s port %d\n",
-			       device->name, i);
+			dev_err(&device->dev, "Couldn't close port %d\n", i);
 		i--;
 	}
 }
@@ -3033,12 +3114,12 @@
 	}
 	for (i = 0; i < num_ports; i++, cur_port++) {
 		if (ib_agent_port_close(device, cur_port))
-			printk(KERN_ERR PFX "Couldn't close %s port %d "
-			       "for agents\n",
-			       device->name, cur_port);
+			dev_err(&device->dev,
+				"Couldn't close port %d for agents\n",
+				cur_port);
 		if (ib_mad_port_close(device, cur_port))
-			printk(KERN_ERR PFX "Couldn't close %s port %d\n",
-			       device->name, cur_port);
+			dev_err(&device->dev, "Couldn't close port %d\n",
+				cur_port);
 	}
 }
 
@@ -3064,7 +3145,7 @@
 					 SLAB_HWCACHE_ALIGN,
 					 NULL);
 	if (!ib_mad_cache) {
-		printk(KERN_ERR PFX "Couldn't create ib_mad cache\n");
+		pr_err("Couldn't create ib_mad cache\n");
 		ret = -ENOMEM;
 		goto error1;
 	}
@@ -3072,7 +3153,7 @@
 	INIT_LIST_HEAD(&ib_mad_port_list);
 
 	if (ib_register_client(&mad_client)) {
-		printk(KERN_ERR PFX "Couldn't register ib_mad client\n");
+		pr_err("Couldn't register ib_mad client\n");
 		ret = -EINVAL;
 		goto error2;
 	}
diff --git a/drivers/infiniband/core/mad_priv.h b/drivers/infiniband/core/mad_priv.h
index 9430ab4..d1a0b0e 100644
--- a/drivers/infiniband/core/mad_priv.h
+++ b/drivers/infiniband/core/mad_priv.h
@@ -42,9 +42,6 @@
 #include <rdma/ib_mad.h>
 #include <rdma/ib_smi.h>
 
-
-#define PFX "ib_mad: "
-
 #define IB_MAD_QPS_CORE		2 /* Always QP0 and QP1 as a minimum */
 
 /* QP and CQ parameters */
diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c
index 233eaf5..c38f030 100644
--- a/drivers/infiniband/core/sa_query.c
+++ b/drivers/infiniband/core/sa_query.c
@@ -1184,7 +1184,7 @@
 		sa_dev->port[i].agent =
 			ib_register_mad_agent(device, i + s, IB_QPT_GSI,
 					      NULL, 0, send_handler,
-					      recv_handler, sa_dev);
+					      recv_handler, sa_dev, 0);
 		if (IS_ERR(sa_dev->port[i].agent))
 			goto err;
 
diff --git a/drivers/infiniband/core/user_mad.c b/drivers/infiniband/core/user_mad.c
index 1acb991..928cdd2 100644
--- a/drivers/infiniband/core/user_mad.c
+++ b/drivers/infiniband/core/user_mad.c
@@ -33,6 +33,8 @@
  * SOFTWARE.
  */
 
+#define pr_fmt(fmt) "user_mad: " fmt
+
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/device.h>
@@ -504,13 +506,15 @@
 
 	rmpp_mad = (struct ib_rmpp_mad *) packet->mad.data;
 	hdr_len = ib_get_mad_data_offset(rmpp_mad->mad_hdr.mgmt_class);
-	if (!ib_is_mad_class_rmpp(rmpp_mad->mad_hdr.mgmt_class)) {
-		copy_offset = IB_MGMT_MAD_HDR;
-		rmpp_active = 0;
-	} else {
+
+	if (ib_is_mad_class_rmpp(rmpp_mad->mad_hdr.mgmt_class)
+	    && ib_mad_kernel_rmpp_agent(agent)) {
 		copy_offset = IB_MGMT_RMPP_HDR;
 		rmpp_active = ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) &
-			      IB_MGMT_RMPP_FLAG_ACTIVE;
+						IB_MGMT_RMPP_FLAG_ACTIVE;
+	} else {
+		copy_offset = IB_MGMT_MAD_HDR;
+		rmpp_active = 0;
 	}
 
 	data_len = count - hdr_size(file) - hdr_len;
@@ -556,14 +560,22 @@
 		rmpp_mad->mad_hdr.tid = *tid;
 	}
 
-	spin_lock_irq(&file->send_lock);
-	ret = is_duplicate(file, packet);
-	if (!ret)
+	if (!ib_mad_kernel_rmpp_agent(agent)
+	   && ib_is_mad_class_rmpp(rmpp_mad->mad_hdr.mgmt_class)
+	   && (ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) & IB_MGMT_RMPP_FLAG_ACTIVE)) {
+		spin_lock_irq(&file->send_lock);
 		list_add_tail(&packet->list, &file->send_list);
-	spin_unlock_irq(&file->send_lock);
-	if (ret) {
-		ret = -EINVAL;
-		goto err_msg;
+		spin_unlock_irq(&file->send_lock);
+	} else {
+		spin_lock_irq(&file->send_lock);
+		ret = is_duplicate(file, packet);
+		if (!ret)
+			list_add_tail(&packet->list, &file->send_list);
+		spin_unlock_irq(&file->send_lock);
+		if (ret) {
+			ret = -EINVAL;
+			goto err_msg;
+		}
 	}
 
 	ret = ib_post_send_mad(packet->msg, NULL);
@@ -614,6 +626,8 @@
 	mutex_lock(&file->mutex);
 
 	if (!file->port->ib_dev) {
+		dev_notice(file->port->dev,
+			   "ib_umad_reg_agent: invalid device\n");
 		ret = -EPIPE;
 		goto out;
 	}
@@ -624,6 +638,9 @@
 	}
 
 	if (ureq.qpn != 0 && ureq.qpn != 1) {
+		dev_notice(file->port->dev,
+			   "ib_umad_reg_agent: invalid QPN %d specified\n",
+			   ureq.qpn);
 		ret = -EINVAL;
 		goto out;
 	}
@@ -632,11 +649,15 @@
 		if (!__get_agent(file, agent_id))
 			goto found;
 
+	dev_notice(file->port->dev,
+		   "ib_umad_reg_agent: Max Agents (%u) reached\n",
+		   IB_UMAD_MAX_AGENTS);
 	ret = -ENOMEM;
 	goto out;
 
 found:
 	if (ureq.mgmt_class) {
+		memset(&req, 0, sizeof(req));
 		req.mgmt_class         = ureq.mgmt_class;
 		req.mgmt_class_version = ureq.mgmt_class_version;
 		memcpy(req.oui, ureq.oui, sizeof req.oui);
@@ -657,7 +678,7 @@
 				      ureq.qpn ? IB_QPT_GSI : IB_QPT_SMI,
 				      ureq.mgmt_class ? &req : NULL,
 				      ureq.rmpp_version,
-				      send_handler, recv_handler, file);
+				      send_handler, recv_handler, file, 0);
 	if (IS_ERR(agent)) {
 		ret = PTR_ERR(agent);
 		agent = NULL;
@@ -673,10 +694,11 @@
 	if (!file->already_used) {
 		file->already_used = 1;
 		if (!file->use_pkey_index) {
-			printk(KERN_WARNING "user_mad: process %s did not enable "
-			       "P_Key index support.\n", current->comm);
-			printk(KERN_WARNING "user_mad:   Documentation/infiniband/user_mad.txt "
-			       "has info on the new ABI.\n");
+			dev_warn(file->port->dev,
+				"process %s did not enable P_Key index support.\n",
+				current->comm);
+			dev_warn(file->port->dev,
+				"   Documentation/infiniband/user_mad.txt has info on the new ABI.\n");
 		}
 	}
 
@@ -694,6 +716,119 @@
 	return ret;
 }
 
+static int ib_umad_reg_agent2(struct ib_umad_file *file, void __user *arg)
+{
+	struct ib_user_mad_reg_req2 ureq;
+	struct ib_mad_reg_req req;
+	struct ib_mad_agent *agent = NULL;
+	int agent_id;
+	int ret;
+
+	mutex_lock(&file->port->file_mutex);
+	mutex_lock(&file->mutex);
+
+	if (!file->port->ib_dev) {
+		dev_notice(file->port->dev,
+			   "ib_umad_reg_agent2: invalid device\n");
+		ret = -EPIPE;
+		goto out;
+	}
+
+	if (copy_from_user(&ureq, arg, sizeof(ureq))) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	if (ureq.qpn != 0 && ureq.qpn != 1) {
+		dev_notice(file->port->dev,
+			   "ib_umad_reg_agent2: invalid QPN %d specified\n",
+			   ureq.qpn);
+		ret = -EINVAL;
+		goto out;
+	}
+
+	if (ureq.flags & ~IB_USER_MAD_REG_FLAGS_CAP) {
+		dev_notice(file->port->dev,
+			   "ib_umad_reg_agent2 failed: invalid registration flags specified 0x%x; supported 0x%x\n",
+			   ureq.flags, IB_USER_MAD_REG_FLAGS_CAP);
+		ret = -EINVAL;
+
+		if (put_user((u32)IB_USER_MAD_REG_FLAGS_CAP,
+				(u32 __user *) (arg + offsetof(struct
+				ib_user_mad_reg_req2, flags))))
+			ret = -EFAULT;
+
+		goto out;
+	}
+
+	for (agent_id = 0; agent_id < IB_UMAD_MAX_AGENTS; ++agent_id)
+		if (!__get_agent(file, agent_id))
+			goto found;
+
+	dev_notice(file->port->dev,
+		   "ib_umad_reg_agent2: Max Agents (%u) reached\n",
+		   IB_UMAD_MAX_AGENTS);
+	ret = -ENOMEM;
+	goto out;
+
+found:
+	if (ureq.mgmt_class) {
+		memset(&req, 0, sizeof(req));
+		req.mgmt_class         = ureq.mgmt_class;
+		req.mgmt_class_version = ureq.mgmt_class_version;
+		if (ureq.oui & 0xff000000) {
+			dev_notice(file->port->dev,
+				   "ib_umad_reg_agent2 failed: oui invalid 0x%08x\n",
+				   ureq.oui);
+			ret = -EINVAL;
+			goto out;
+		}
+		req.oui[2] =  ureq.oui & 0x0000ff;
+		req.oui[1] = (ureq.oui & 0x00ff00) >> 8;
+		req.oui[0] = (ureq.oui & 0xff0000) >> 16;
+		memcpy(req.method_mask, ureq.method_mask,
+			sizeof(req.method_mask));
+	}
+
+	agent = ib_register_mad_agent(file->port->ib_dev, file->port->port_num,
+				      ureq.qpn ? IB_QPT_GSI : IB_QPT_SMI,
+				      ureq.mgmt_class ? &req : NULL,
+				      ureq.rmpp_version,
+				      send_handler, recv_handler, file,
+				      ureq.flags);
+	if (IS_ERR(agent)) {
+		ret = PTR_ERR(agent);
+		agent = NULL;
+		goto out;
+	}
+
+	if (put_user(agent_id,
+		     (u32 __user *)(arg +
+				offsetof(struct ib_user_mad_reg_req2, id)))) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	if (!file->already_used) {
+		file->already_used = 1;
+		file->use_pkey_index = 1;
+	}
+
+	file->agent[agent_id] = agent;
+	ret = 0;
+
+out:
+	mutex_unlock(&file->mutex);
+
+	if (ret && agent)
+		ib_unregister_mad_agent(agent);
+
+	mutex_unlock(&file->port->file_mutex);
+
+	return ret;
+}
+
+
 static int ib_umad_unreg_agent(struct ib_umad_file *file, u32 __user *arg)
 {
 	struct ib_mad_agent *agent = NULL;
@@ -749,6 +884,8 @@
 		return ib_umad_unreg_agent(filp->private_data, (__u32 __user *) arg);
 	case IB_USER_MAD_ENABLE_PKEY:
 		return ib_umad_enable_pkey(filp->private_data);
+	case IB_USER_MAD_REGISTER_AGENT2:
+		return ib_umad_reg_agent2(filp->private_data, (void __user *) arg);
 	default:
 		return -ENOIOCTLCMD;
 	}
@@ -765,6 +902,8 @@
 		return ib_umad_unreg_agent(filp->private_data, compat_ptr(arg));
 	case IB_USER_MAD_ENABLE_PKEY:
 		return ib_umad_enable_pkey(filp->private_data);
+	case IB_USER_MAD_REGISTER_AGENT2:
+		return ib_umad_reg_agent2(filp->private_data, compat_ptr(arg));
 	default:
 		return -ENOIOCTLCMD;
 	}
@@ -983,7 +1122,7 @@
 
 static dev_t overflow_maj;
 static DECLARE_BITMAP(overflow_map, IB_UMAD_MAX_PORTS);
-static int find_overflow_devnum(void)
+static int find_overflow_devnum(struct ib_device *device)
 {
 	int ret;
 
@@ -991,7 +1130,8 @@
 		ret = alloc_chrdev_region(&overflow_maj, 0, IB_UMAD_MAX_PORTS * 2,
 					  "infiniband_mad");
 		if (ret) {
-			printk(KERN_ERR "user_mad: couldn't register dynamic device number\n");
+			dev_err(&device->dev,
+				"couldn't register dynamic device number\n");
 			return ret;
 		}
 	}
@@ -1014,7 +1154,7 @@
 	devnum = find_first_zero_bit(dev_map, IB_UMAD_MAX_PORTS);
 	if (devnum >= IB_UMAD_MAX_PORTS) {
 		spin_unlock(&port_lock);
-		devnum = find_overflow_devnum();
+		devnum = find_overflow_devnum(device);
 		if (devnum < 0)
 			return -1;
 
@@ -1200,14 +1340,14 @@
 	ret = register_chrdev_region(base_dev, IB_UMAD_MAX_PORTS * 2,
 				     "infiniband_mad");
 	if (ret) {
-		printk(KERN_ERR "user_mad: couldn't register device number\n");
+		pr_err("couldn't register device number\n");
 		goto out;
 	}
 
 	umad_class = class_create(THIS_MODULE, "infiniband_mad");
 	if (IS_ERR(umad_class)) {
 		ret = PTR_ERR(umad_class);
-		printk(KERN_ERR "user_mad: couldn't create class infiniband_mad\n");
+		pr_err("couldn't create class infiniband_mad\n");
 		goto out_chrdev;
 	}
 
@@ -1215,13 +1355,13 @@
 
 	ret = class_create_file(umad_class, &class_attr_abi_version.attr);
 	if (ret) {
-		printk(KERN_ERR "user_mad: couldn't create abi_version attribute\n");
+		pr_err("couldn't create abi_version attribute\n");
 		goto out_class;
 	}
 
 	ret = ib_register_client(&umad_client);
 	if (ret) {
-		printk(KERN_ERR "user_mad: couldn't register ib_umad client\n");
+		pr_err("couldn't register ib_umad client\n");
 		goto out_class;
 	}
 
diff --git a/drivers/infiniband/core/uverbs.h b/drivers/infiniband/core/uverbs.h
index a283274..643c08a 100644
--- a/drivers/infiniband/core/uverbs.h
+++ b/drivers/infiniband/core/uverbs.h
@@ -221,6 +221,7 @@
 IB_UVERBS_DECLARE_CMD(alloc_pd);
 IB_UVERBS_DECLARE_CMD(dealloc_pd);
 IB_UVERBS_DECLARE_CMD(reg_mr);
+IB_UVERBS_DECLARE_CMD(rereg_mr);
 IB_UVERBS_DECLARE_CMD(dereg_mr);
 IB_UVERBS_DECLARE_CMD(alloc_mw);
 IB_UVERBS_DECLARE_CMD(dealloc_mw);
diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c
index ea6203e..0600c50 100644
--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -1002,6 +1002,99 @@
 	return ret;
 }
 
+ssize_t ib_uverbs_rereg_mr(struct ib_uverbs_file *file,
+			   const char __user *buf, int in_len,
+			   int out_len)
+{
+	struct ib_uverbs_rereg_mr      cmd;
+	struct ib_uverbs_rereg_mr_resp resp;
+	struct ib_udata              udata;
+	struct ib_pd                *pd = NULL;
+	struct ib_mr                *mr;
+	struct ib_pd		    *old_pd;
+	int                          ret;
+	struct ib_uobject	    *uobj;
+
+	if (out_len < sizeof(resp))
+		return -ENOSPC;
+
+	if (copy_from_user(&cmd, buf, sizeof(cmd)))
+		return -EFAULT;
+
+	INIT_UDATA(&udata, buf + sizeof(cmd),
+		   (unsigned long) cmd.response + sizeof(resp),
+		   in_len - sizeof(cmd), out_len - sizeof(resp));
+
+	if (cmd.flags & ~IB_MR_REREG_SUPPORTED || !cmd.flags)
+		return -EINVAL;
+
+	if ((cmd.flags & IB_MR_REREG_TRANS) &&
+	    (!cmd.start || !cmd.hca_va || 0 >= cmd.length ||
+	     (cmd.start & ~PAGE_MASK) != (cmd.hca_va & ~PAGE_MASK)))
+			return -EINVAL;
+
+	uobj = idr_write_uobj(&ib_uverbs_mr_idr, cmd.mr_handle,
+			      file->ucontext);
+
+	if (!uobj)
+		return -EINVAL;
+
+	mr = uobj->object;
+
+	if (cmd.flags & IB_MR_REREG_ACCESS) {
+		ret = ib_check_mr_access(cmd.access_flags);
+		if (ret)
+			goto put_uobjs;
+	}
+
+	if (cmd.flags & IB_MR_REREG_PD) {
+		pd = idr_read_pd(cmd.pd_handle, file->ucontext);
+		if (!pd) {
+			ret = -EINVAL;
+			goto put_uobjs;
+		}
+	}
+
+	if (atomic_read(&mr->usecnt)) {
+		ret = -EBUSY;
+		goto put_uobj_pd;
+	}
+
+	old_pd = mr->pd;
+	ret = mr->device->rereg_user_mr(mr, cmd.flags, cmd.start,
+					cmd.length, cmd.hca_va,
+					cmd.access_flags, pd, &udata);
+	if (!ret) {
+		if (cmd.flags & IB_MR_REREG_PD) {
+			atomic_inc(&pd->usecnt);
+			mr->pd = pd;
+			atomic_dec(&old_pd->usecnt);
+		}
+	} else {
+		goto put_uobj_pd;
+	}
+
+	memset(&resp, 0, sizeof(resp));
+	resp.lkey      = mr->lkey;
+	resp.rkey      = mr->rkey;
+
+	if (copy_to_user((void __user *)(unsigned long)cmd.response,
+			 &resp, sizeof(resp)))
+		ret = -EFAULT;
+	else
+		ret = in_len;
+
+put_uobj_pd:
+	if (cmd.flags & IB_MR_REREG_PD)
+		put_pd_read(pd);
+
+put_uobjs:
+
+	put_uobj_write(mr->uobject);
+
+	return ret;
+}
+
 ssize_t ib_uverbs_dereg_mr(struct ib_uverbs_file *file,
 			   const char __user *buf, int in_len,
 			   int out_len)
diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c
index 08219fb..c73b22a 100644
--- a/drivers/infiniband/core/uverbs_main.c
+++ b/drivers/infiniband/core/uverbs_main.c
@@ -87,6 +87,7 @@
 	[IB_USER_VERBS_CMD_ALLOC_PD]		= ib_uverbs_alloc_pd,
 	[IB_USER_VERBS_CMD_DEALLOC_PD]		= ib_uverbs_dealloc_pd,
 	[IB_USER_VERBS_CMD_REG_MR]		= ib_uverbs_reg_mr,
+	[IB_USER_VERBS_CMD_REREG_MR]		= ib_uverbs_rereg_mr,
 	[IB_USER_VERBS_CMD_DEREG_MR]		= ib_uverbs_dereg_mr,
 	[IB_USER_VERBS_CMD_ALLOC_MW]		= ib_uverbs_alloc_mw,
 	[IB_USER_VERBS_CMD_DEALLOC_MW]		= ib_uverbs_dealloc_mw,
diff --git a/drivers/infiniband/hw/amso1100/c2_cq.c b/drivers/infiniband/hw/amso1100/c2_cq.c
index 49e0e85..1b63185 100644
--- a/drivers/infiniband/hw/amso1100/c2_cq.c
+++ b/drivers/infiniband/hw/amso1100/c2_cq.c
@@ -260,11 +260,14 @@
 			  mq->msg_pool.host, dma_unmap_addr(mq, mapping));
 }
 
-static int c2_alloc_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq, int q_size,
-			   int msg_size)
+static int c2_alloc_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq,
+			   size_t q_size, size_t msg_size)
 {
 	u8 *pool_start;
 
+	if (q_size > SIZE_MAX / msg_size)
+		return -EINVAL;
+
 	pool_start = dma_alloc_coherent(&c2dev->pcidev->dev, q_size * msg_size,
 					&mq->host_dma, GFP_KERNEL);
 	if (!pool_start)
diff --git a/drivers/infiniband/hw/cxgb4/ev.c b/drivers/infiniband/hw/cxgb4/ev.c
index fbe6051..c9df054 100644
--- a/drivers/infiniband/hw/cxgb4/ev.c
+++ b/drivers/infiniband/hw/cxgb4/ev.c
@@ -227,6 +227,7 @@
 
 	chp = get_chp(dev, qid);
 	if (chp) {
+		t4_clear_cq_armed(&chp->cq);
 		spin_lock_irqsave(&chp->comp_handler_lock, flag);
 		(*chp->ibcq.comp_handler)(&chp->ibcq, chp->ibcq.cq_context);
 		spin_unlock_irqrestore(&chp->comp_handler_lock, flag);
diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c
index c158fcc..41cd688 100644
--- a/drivers/infiniband/hw/cxgb4/qp.c
+++ b/drivers/infiniband/hw/cxgb4/qp.c
@@ -1105,7 +1105,7 @@
 		       struct c4iw_cq *schp)
 {
 	int count;
-	int flushed;
+	int rq_flushed, sq_flushed;
 	unsigned long flag;
 
 	PDBG("%s qhp %p rchp %p schp %p\n", __func__, qhp, rchp, schp);
@@ -1123,27 +1123,40 @@
 
 	c4iw_flush_hw_cq(rchp);
 	c4iw_count_rcqes(&rchp->cq, &qhp->wq, &count);
-	flushed = c4iw_flush_rq(&qhp->wq, &rchp->cq, count);
+	rq_flushed = c4iw_flush_rq(&qhp->wq, &rchp->cq, count);
 	spin_unlock(&qhp->lock);
 	spin_unlock_irqrestore(&rchp->lock, flag);
-	if (flushed) {
-		spin_lock_irqsave(&rchp->comp_handler_lock, flag);
-		(*rchp->ibcq.comp_handler)(&rchp->ibcq, rchp->ibcq.cq_context);
-		spin_unlock_irqrestore(&rchp->comp_handler_lock, flag);
-	}
 
 	/* locking hierarchy: cq lock first, then qp lock. */
 	spin_lock_irqsave(&schp->lock, flag);
 	spin_lock(&qhp->lock);
 	if (schp != rchp)
 		c4iw_flush_hw_cq(schp);
-	flushed = c4iw_flush_sq(qhp);
+	sq_flushed = c4iw_flush_sq(qhp);
 	spin_unlock(&qhp->lock);
 	spin_unlock_irqrestore(&schp->lock, flag);
-	if (flushed) {
-		spin_lock_irqsave(&schp->comp_handler_lock, flag);
-		(*schp->ibcq.comp_handler)(&schp->ibcq, schp->ibcq.cq_context);
-		spin_unlock_irqrestore(&schp->comp_handler_lock, flag);
+
+	if (schp == rchp) {
+		if (t4_clear_cq_armed(&rchp->cq) &&
+		    (rq_flushed || sq_flushed)) {
+			spin_lock_irqsave(&rchp->comp_handler_lock, flag);
+			(*rchp->ibcq.comp_handler)(&rchp->ibcq,
+						   rchp->ibcq.cq_context);
+			spin_unlock_irqrestore(&rchp->comp_handler_lock, flag);
+		}
+	} else {
+		if (t4_clear_cq_armed(&rchp->cq) && rq_flushed) {
+			spin_lock_irqsave(&rchp->comp_handler_lock, flag);
+			(*rchp->ibcq.comp_handler)(&rchp->ibcq,
+						   rchp->ibcq.cq_context);
+			spin_unlock_irqrestore(&rchp->comp_handler_lock, flag);
+		}
+		if (t4_clear_cq_armed(&schp->cq) && sq_flushed) {
+			spin_lock_irqsave(&schp->comp_handler_lock, flag);
+			(*schp->ibcq.comp_handler)(&schp->ibcq,
+						   schp->ibcq.cq_context);
+			spin_unlock_irqrestore(&schp->comp_handler_lock, flag);
+		}
 	}
 }
 
diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h
index df5edfa..c04e513 100644
--- a/drivers/infiniband/hw/cxgb4/t4.h
+++ b/drivers/infiniband/hw/cxgb4/t4.h
@@ -524,6 +524,10 @@
 	return !wq->rq.queue[wq->rq.size].status.db_off;
 }
 
+enum t4_cq_flags {
+	CQ_ARMED	= 1,
+};
+
 struct t4_cq {
 	struct t4_cqe *queue;
 	dma_addr_t dma_addr;
@@ -544,12 +548,19 @@
 	u16 cidx_inc;
 	u8 gen;
 	u8 error;
+	unsigned long flags;
 };
 
+static inline int t4_clear_cq_armed(struct t4_cq *cq)
+{
+	return test_and_clear_bit(CQ_ARMED, &cq->flags);
+}
+
 static inline int t4_arm_cq(struct t4_cq *cq, int se)
 {
 	u32 val;
 
+	set_bit(CQ_ARMED, &cq->flags);
 	while (cq->cidx_inc > CIDXINC_MASK) {
 		val = SEINTARM(0) | CIDXINC(CIDXINC_MASK) | TIMERREG(7) |
 		      INGRESSQID(cq->cqid);
diff --git a/drivers/infiniband/hw/ipath/ipath_mad.c b/drivers/infiniband/hw/ipath/ipath_mad.c
index 43f2d04..e890e5b 100644
--- a/drivers/infiniband/hw/ipath/ipath_mad.c
+++ b/drivers/infiniband/hw/ipath/ipath_mad.c
@@ -726,7 +726,7 @@
  * @dd: the infinipath device
  * @pkeys: the PKEY table
  */
-static int set_pkeys(struct ipath_devdata *dd, u16 *pkeys)
+static int set_pkeys(struct ipath_devdata *dd, u16 *pkeys, u8 port)
 {
 	struct ipath_portdata *pd;
 	int i;
@@ -759,6 +759,7 @@
 	}
 	if (changed) {
 		u64 pkey;
+		struct ib_event event;
 
 		pkey = (u64) dd->ipath_pkeys[0] |
 			((u64) dd->ipath_pkeys[1] << 16) |
@@ -768,12 +769,17 @@
 			   (unsigned long long) pkey);
 		ipath_write_kreg(dd, dd->ipath_kregs->kr_partitionkey,
 				 pkey);
+
+		event.event = IB_EVENT_PKEY_CHANGE;
+		event.device = &dd->verbs_dev->ibdev;
+		event.element.port_num = port;
+		ib_dispatch_event(&event);
 	}
 	return 0;
 }
 
 static int recv_subn_set_pkeytable(struct ib_smp *smp,
-				   struct ib_device *ibdev)
+				   struct ib_device *ibdev, u8 port)
 {
 	u32 startpx = 32 * (be32_to_cpu(smp->attr_mod) & 0xffff);
 	__be16 *p = (__be16 *) smp->data;
@@ -784,7 +790,7 @@
 	for (i = 0; i < n; i++)
 		q[i] = be16_to_cpu(p[i]);
 
-	if (startpx != 0 || set_pkeys(dev->dd, q) != 0)
+	if (startpx != 0 || set_pkeys(dev->dd, q, port) != 0)
 		smp->status |= IB_SMP_INVALID_FIELD;
 
 	return recv_subn_get_pkeytable(smp, ibdev);
@@ -1342,7 +1348,7 @@
 			ret = recv_subn_set_portinfo(smp, ibdev, port_num);
 			goto bail;
 		case IB_SMP_ATTR_PKEY_TABLE:
-			ret = recv_subn_set_pkeytable(smp, ibdev);
+			ret = recv_subn_set_pkeytable(smp, ibdev, port_num);
 			goto bail;
 		case IB_SMP_ATTR_SM_INFO:
 			if (dev->port_cap_flags & IB_PORT_SM_DISABLED) {
diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c
index 287ad05..82a7dd8 100644
--- a/drivers/infiniband/hw/mlx4/mad.c
+++ b/drivers/infiniband/hw/mlx4/mad.c
@@ -891,7 +891,7 @@
 				agent = ib_register_mad_agent(&dev->ib_dev, p + 1,
 							      q ? IB_QPT_GSI : IB_QPT_SMI,
 							      NULL, 0, send_handler,
-							      NULL, NULL);
+							      NULL, NULL, 0);
 				if (IS_ERR(agent)) {
 					ret = PTR_ERR(agent);
 					goto err;
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 0f7027e..e1e558a 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -910,8 +910,7 @@
 	const struct default_rules *pdefault_rules = default_table;
 	u8 link_layer = rdma_port_get_link_layer(qp->device, flow_attr->port);
 
-	for (i = 0; i < sizeof(default_table)/sizeof(default_table[0]); i++,
-	     pdefault_rules++) {
+	for (i = 0; i < ARRAY_SIZE(default_table); i++, pdefault_rules++) {
 		__u32 field_types[IB_FLOW_SPEC_SUPPORT_LAYERS];
 		memset(&field_types, 0, sizeof(field_types));
 
@@ -965,8 +964,7 @@
 	int size = 0;
 	int i;
 
-	for (i = 0; i < sizeof(pdefault_rules->rules_create_list)/
-			sizeof(pdefault_rules->rules_create_list[0]); i++) {
+	for (i = 0; i < ARRAY_SIZE(pdefault_rules->rules_create_list); i++) {
 		int ret;
 		union ib_flow_spec ib_spec;
 		switch (pdefault_rules->rules_create_list[i]) {
@@ -2007,6 +2005,7 @@
 		(1ull << IB_USER_VERBS_CMD_ALLOC_PD)		|
 		(1ull << IB_USER_VERBS_CMD_DEALLOC_PD)		|
 		(1ull << IB_USER_VERBS_CMD_REG_MR)		|
+		(1ull << IB_USER_VERBS_CMD_REREG_MR)		|
 		(1ull << IB_USER_VERBS_CMD_DEREG_MR)		|
 		(1ull << IB_USER_VERBS_CMD_CREATE_COMP_CHANNEL)	|
 		(1ull << IB_USER_VERBS_CMD_CREATE_CQ)		|
@@ -2059,6 +2058,7 @@
 	ibdev->ib_dev.req_notify_cq	= mlx4_ib_arm_cq;
 	ibdev->ib_dev.get_dma_mr	= mlx4_ib_get_dma_mr;
 	ibdev->ib_dev.reg_user_mr	= mlx4_ib_reg_user_mr;
+	ibdev->ib_dev.rereg_user_mr	= mlx4_ib_rereg_user_mr;
 	ibdev->ib_dev.dereg_mr		= mlx4_ib_dereg_mr;
 	ibdev->ib_dev.alloc_fast_reg_mr = mlx4_ib_alloc_fast_reg_mr;
 	ibdev->ib_dev.alloc_fast_reg_page_list = mlx4_ib_alloc_fast_reg_page_list;
diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h
index 369da3c..e8cad39 100644
--- a/drivers/infiniband/hw/mlx4/mlx4_ib.h
+++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h
@@ -788,5 +788,9 @@
 void mlx4_ib_steer_qp_free(struct mlx4_ib_dev *dev, u32 qpn, int count);
 int mlx4_ib_steer_qp_reg(struct mlx4_ib_dev *mdev, struct mlx4_ib_qp *mqp,
 			 int is_attach);
+int mlx4_ib_rereg_user_mr(struct ib_mr *mr, int flags,
+			  u64 start, u64 length, u64 virt_addr,
+			  int mr_access_flags, struct ib_pd *pd,
+			  struct ib_udata *udata);
 
 #endif /* MLX4_IB_H */
diff --git a/drivers/infiniband/hw/mlx4/mr.c b/drivers/infiniband/hw/mlx4/mr.c
index cb2a872..9b0e80e 100644
--- a/drivers/infiniband/hw/mlx4/mr.c
+++ b/drivers/infiniband/hw/mlx4/mr.c
@@ -144,8 +144,10 @@
 	if (!mr)
 		return ERR_PTR(-ENOMEM);
 
+	/* Force registering the memory as writable. */
+	/* Used for memory re-registeration. HCA protects the access */
 	mr->umem = ib_umem_get(pd->uobject->context, start, length,
-			       access_flags, 0);
+			       access_flags | IB_ACCESS_LOCAL_WRITE, 0);
 	if (IS_ERR(mr->umem)) {
 		err = PTR_ERR(mr->umem);
 		goto err_free;
@@ -183,6 +185,90 @@
 	return ERR_PTR(err);
 }
 
+int mlx4_ib_rereg_user_mr(struct ib_mr *mr, int flags,
+			  u64 start, u64 length, u64 virt_addr,
+			  int mr_access_flags, struct ib_pd *pd,
+			  struct ib_udata *udata)
+{
+	struct mlx4_ib_dev *dev = to_mdev(mr->device);
+	struct mlx4_ib_mr *mmr = to_mmr(mr);
+	struct mlx4_mpt_entry *mpt_entry;
+	struct mlx4_mpt_entry **pmpt_entry = &mpt_entry;
+	int err;
+
+	/* Since we synchronize this call and mlx4_ib_dereg_mr via uverbs,
+	 * we assume that the calls can't run concurrently. Otherwise, a
+	 * race exists.
+	 */
+	err =  mlx4_mr_hw_get_mpt(dev->dev, &mmr->mmr, &pmpt_entry);
+
+	if (err)
+		return err;
+
+	if (flags & IB_MR_REREG_PD) {
+		err = mlx4_mr_hw_change_pd(dev->dev, *pmpt_entry,
+					   to_mpd(pd)->pdn);
+
+		if (err)
+			goto release_mpt_entry;
+	}
+
+	if (flags & IB_MR_REREG_ACCESS) {
+		err = mlx4_mr_hw_change_access(dev->dev, *pmpt_entry,
+					       convert_access(mr_access_flags));
+
+		if (err)
+			goto release_mpt_entry;
+	}
+
+	if (flags & IB_MR_REREG_TRANS) {
+		int shift;
+		int err;
+		int n;
+
+		mlx4_mr_rereg_mem_cleanup(dev->dev, &mmr->mmr);
+		ib_umem_release(mmr->umem);
+		mmr->umem = ib_umem_get(mr->uobject->context, start, length,
+					mr_access_flags |
+					IB_ACCESS_LOCAL_WRITE,
+					0);
+		if (IS_ERR(mmr->umem)) {
+			err = PTR_ERR(mmr->umem);
+			mmr->umem = NULL;
+			goto release_mpt_entry;
+		}
+		n = ib_umem_page_count(mmr->umem);
+		shift = ilog2(mmr->umem->page_size);
+
+		mmr->mmr.iova       = virt_addr;
+		mmr->mmr.size       = length;
+		err = mlx4_mr_rereg_mem_write(dev->dev, &mmr->mmr,
+					      virt_addr, length, n, shift,
+					      *pmpt_entry);
+		if (err) {
+			ib_umem_release(mmr->umem);
+			goto release_mpt_entry;
+		}
+
+		err = mlx4_ib_umem_write_mtt(dev, &mmr->mmr.mtt, mmr->umem);
+		if (err) {
+			mlx4_mr_rereg_mem_cleanup(dev->dev, &mmr->mmr);
+			ib_umem_release(mmr->umem);
+			goto release_mpt_entry;
+		}
+	}
+
+	/* If we couldn't transfer the MR to the HCA, just remember to
+	 * return a failure. But dereg_mr will free the resources.
+	 */
+	err = mlx4_mr_hw_write_mpt(dev->dev, &mmr->mmr, pmpt_entry);
+
+release_mpt_entry:
+	mlx4_mr_hw_put_mpt(dev->dev, pmpt_entry);
+
+	return err;
+}
+
 int mlx4_ib_dereg_mr(struct ib_mr *ibmr)
 {
 	struct mlx4_ib_mr *mr = to_mmr(ibmr);
diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index 7efe6e3..8c574b6 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -2501,7 +2501,7 @@
 	spin_lock_irqsave(&qp->sq.lock, flags);
 
 	for (nreq = 0; wr; nreq++, wr = wr->next) {
-		if (unlikely(wr->opcode >= sizeof(mlx5_ib_opcode) / sizeof(mlx5_ib_opcode[0]))) {
+		if (unlikely(wr->opcode >= ARRAY_SIZE(mlx5_ib_opcode))) {
 			mlx5_ib_warn(dev, "\n");
 			err = -EINVAL;
 			*bad_wr = wr;
diff --git a/drivers/infiniband/hw/mthca/mthca_mad.c b/drivers/infiniband/hw/mthca/mthca_mad.c
index b6f7f45..8881fa3 100644
--- a/drivers/infiniband/hw/mthca/mthca_mad.c
+++ b/drivers/infiniband/hw/mthca/mthca_mad.c
@@ -294,7 +294,7 @@
 			agent = ib_register_mad_agent(&dev->ib_dev, p + 1,
 						      q ? IB_QPT_GSI : IB_QPT_SMI,
 						      NULL, 0, send_handler,
-						      NULL, NULL);
+						      NULL, NULL, 0);
 			if (IS_ERR(agent)) {
 				ret = PTR_ERR(agent);
 				goto err;
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma.h b/drivers/infiniband/hw/ocrdma/ocrdma.h
index 19011db..b43456a 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma.h
+++ b/drivers/infiniband/hw/ocrdma/ocrdma.h
@@ -40,7 +40,7 @@
 #include <be_roce.h>
 #include "ocrdma_sli.h"
 
-#define OCRDMA_ROCE_DRV_VERSION "10.2.145.0u"
+#define OCRDMA_ROCE_DRV_VERSION "10.2.287.0u"
 
 #define OCRDMA_ROCE_DRV_DESC "Emulex OneConnect RoCE Driver"
 #define OCRDMA_NODE_DESC "Emulex OneConnect RoCE HCA"
@@ -137,6 +137,7 @@
 	u16 cqe_status;
 	u16 ext_status;
 	bool cmd_done;
+	bool fw_error_state;
 };
 
 struct ocrdma_hw_mr {
@@ -235,7 +236,10 @@
 	struct list_head entry;
 	struct rcu_head rcu;
 	int id;
-	u64 stag_arr[OCRDMA_MAX_STAG];
+	u64 *stag_arr;
+	u8 sl; /* service level */
+	bool pfc_state;
+	atomic_t update_sl;
 	u16 pvid;
 	u32 asic_id;
 
@@ -518,4 +522,22 @@
 				OCRDMA_SLI_ASIC_GEN_NUM_SHIFT;
 }
 
+static inline u8 ocrdma_get_pfc_prio(u8 *pfc, u8 prio)
+{
+	return *(pfc + prio);
+}
+
+static inline u8 ocrdma_get_app_prio(u8 *app_prio, u8 prio)
+{
+	return *(app_prio + prio);
+}
+
+static inline u8 ocrdma_is_enabled_and_synced(u32 state)
+{	/* May also be used to interpret TC-state, QCN-state
+	 * Appl-state and Logical-link-state in future.
+	 */
+	return (state & OCRDMA_STATE_FLAG_ENABLED) &&
+		(state & OCRDMA_STATE_FLAG_SYNC);
+}
+
 #endif
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_ah.c b/drivers/infiniband/hw/ocrdma/ocrdma_ah.c
index d4cc01f..40f8536 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_ah.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_ah.c
@@ -35,6 +35,8 @@
 #include "ocrdma_ah.h"
 #include "ocrdma_hw.h"
 
+#define OCRDMA_VID_PCP_SHIFT	0xD
+
 static inline int set_av_attr(struct ocrdma_dev *dev, struct ocrdma_ah *ah,
 				struct ib_ah_attr *attr, int pdid)
 {
@@ -55,7 +57,7 @@
 	if (vlan_tag && (vlan_tag < 0x1000)) {
 		eth.eth_type = cpu_to_be16(0x8100);
 		eth.roce_eth_type = cpu_to_be16(OCRDMA_ROCE_ETH_TYPE);
-		vlan_tag |= (attr->sl & 7) << 13;
+		vlan_tag |= (dev->sl & 0x07) << OCRDMA_VID_PCP_SHIFT;
 		eth.vlan_tag = cpu_to_be16(vlan_tag);
 		eth_sz = sizeof(struct ocrdma_eth_vlan);
 		vlan_enabled = true;
@@ -100,6 +102,8 @@
 	if (!(attr->ah_flags & IB_AH_GRH))
 		return ERR_PTR(-EINVAL);
 
+	if (atomic_cmpxchg(&dev->update_sl, 1, 0))
+		ocrdma_init_service_level(dev);
 	ah = kzalloc(sizeof(*ah), GFP_ATOMIC);
 	if (!ah)
 		return ERR_PTR(-ENOMEM);
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_hw.c b/drivers/infiniband/hw/ocrdma/ocrdma_hw.c
index 3bbf201..dd35ae5 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_hw.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_hw.c
@@ -525,7 +525,7 @@
 
 	cmd->ev_cnt_flags = OCRDMA_CREATE_CQ_DEF_FLAGS;
 	cmd->eqn = eq->id;
-	cmd->cqe_count = cq->size / sizeof(struct ocrdma_mcqe);
+	cmd->pdid_cqecnt = cq->size / sizeof(struct ocrdma_mcqe);
 
 	ocrdma_build_q_pages(&cmd->pa[0], cq->size / OCRDMA_MIN_Q_PAGE_SIZE,
 			     cq->dma, PAGE_SIZE_4K);
@@ -661,7 +661,7 @@
 {
 	struct ocrdma_qp *qp = NULL;
 	struct ocrdma_cq *cq = NULL;
-	struct ib_event ib_evt = { 0 };
+	struct ib_event ib_evt;
 	int cq_event = 0;
 	int qp_event = 1;
 	int srq_event = 0;
@@ -674,6 +674,8 @@
 	if (cqe->cqvalid_cqid & OCRDMA_AE_MCQE_CQVALID)
 		cq = dev->cq_tbl[cqe->cqvalid_cqid & OCRDMA_AE_MCQE_CQID_MASK];
 
+	memset(&ib_evt, 0, sizeof(ib_evt));
+
 	ib_evt.device = &dev->ibdev;
 
 	switch (type) {
@@ -771,6 +773,10 @@
 					OCRDMA_AE_PVID_MCQE_TAG_MASK) >>
 					OCRDMA_AE_PVID_MCQE_TAG_SHIFT);
 		break;
+
+	case OCRDMA_ASYNC_EVENT_COS_VALUE:
+		atomic_set(&dev->update_sl, 1);
+		break;
 	default:
 		/* Not interested evts. */
 		break;
@@ -962,8 +968,12 @@
 				    msecs_to_jiffies(30000));
 	if (status)
 		return 0;
-	else
+	else {
+		dev->mqe_ctx.fw_error_state = true;
+		pr_err("%s(%d) mailbox timeout: fw not responding\n",
+		       __func__, dev->id);
 		return -1;
+	}
 }
 
 /* issue a mailbox command on the MQ */
@@ -975,6 +985,8 @@
 	struct ocrdma_mbx_rsp *rsp = NULL;
 
 	mutex_lock(&dev->mqe_ctx.lock);
+	if (dev->mqe_ctx.fw_error_state)
+		goto mbx_err;
 	ocrdma_post_mqe(dev, mqe);
 	status = ocrdma_wait_mqe_cmpl(dev);
 	if (status)
@@ -1078,7 +1090,8 @@
 	    OCRDMA_MBX_QUERY_CFG_CA_ACK_DELAY_SHIFT;
 	attr->max_mw = rsp->max_mw;
 	attr->max_mr = rsp->max_mr;
-	attr->max_mr_size = ~0ull;
+	attr->max_mr_size = ((u64)rsp->max_mr_size_hi << 32) |
+			      rsp->max_mr_size_lo;
 	attr->max_fmr = 0;
 	attr->max_pages_per_frmr = rsp->max_pages_per_frmr;
 	attr->max_num_mr_pbl = rsp->max_num_mr_pbl;
@@ -1252,7 +1265,9 @@
 		ctrl_attr_rsp = (struct ocrdma_get_ctrl_attribs_rsp *)dma.va;
 		hba_attribs = &ctrl_attr_rsp->ctrl_attribs.hba_attribs;
 
-		dev->hba_port_num = hba_attribs->phy_port;
+		dev->hba_port_num = (hba_attribs->ptpnum_maxdoms_hbast_cv &
+					OCRDMA_HBA_ATTRB_PTNUM_MASK)
+					>> OCRDMA_HBA_ATTRB_PTNUM_SHIFT;
 		strncpy(dev->model_number,
 			hba_attribs->controller_model_number, 31);
 	}
@@ -1302,7 +1317,8 @@
 		goto mbx_err;
 
 	rsp = (struct ocrdma_get_link_speed_rsp *)cmd;
-	*lnk_speed = rsp->phys_port_speed;
+	*lnk_speed = (rsp->pflt_pps_ld_pnum & OCRDMA_PHY_PS_MASK)
+			>> OCRDMA_PHY_PS_SHIFT;
 
 mbx_err:
 	kfree(cmd);
@@ -1328,11 +1344,16 @@
 		goto mbx_err;
 
 	rsp = (struct ocrdma_get_phy_info_rsp *)cmd;
-	dev->phy.phy_type = le16_to_cpu(rsp->phy_type);
+	dev->phy.phy_type =
+			(rsp->ityp_ptyp & OCRDMA_PHY_TYPE_MASK);
+	dev->phy.interface_type =
+			(rsp->ityp_ptyp & OCRDMA_IF_TYPE_MASK)
+				>> OCRDMA_IF_TYPE_SHIFT;
 	dev->phy.auto_speeds_supported  =
-			le16_to_cpu(rsp->auto_speeds_supported);
+			(rsp->fspeed_aspeed & OCRDMA_ASPEED_SUPP_MASK);
 	dev->phy.fixed_speeds_supported =
-			le16_to_cpu(rsp->fixed_speeds_supported);
+			(rsp->fspeed_aspeed & OCRDMA_FSPEED_SUPP_MASK)
+				>> OCRDMA_FSPEED_SUPP_SHIFT;
 mbx_err:
 	kfree(cmd);
 	return status;
@@ -1457,8 +1478,8 @@
 
 	pbes = (struct ocrdma_pbe *)dev->av_tbl.pbl.va;
 	for (i = 0; i < dev->av_tbl.size / OCRDMA_MIN_Q_PAGE_SIZE; i++) {
-		pbes[i].pa_lo = (u32) (pa & 0xffffffff);
-		pbes[i].pa_hi = (u32) upper_32_bits(pa);
+		pbes[i].pa_lo = (u32)cpu_to_le32(pa & 0xffffffff);
+		pbes[i].pa_hi = (u32)cpu_to_le32(upper_32_bits(pa));
 		pa += PAGE_SIZE;
 	}
 	cmd->tbl_addr[0].lo = (u32)(dev->av_tbl.pbl.pa & 0xFFFFFFFF);
@@ -1501,6 +1522,7 @@
 	ocrdma_mbx_cmd(dev, (struct ocrdma_mqe *)cmd);
 	dma_free_coherent(&pdev->dev, dev->av_tbl.size, dev->av_tbl.va,
 			  dev->av_tbl.pa);
+	dev->av_tbl.va = NULL;
 	dma_free_coherent(&pdev->dev, PAGE_SIZE, dev->av_tbl.pbl.va,
 			  dev->av_tbl.pbl.pa);
 	kfree(cmd);
@@ -1624,14 +1646,16 @@
 			cmd->cmd.pgsz_pgcnt |= OCRDMA_CREATE_CQ_DPP <<
 				OCRDMA_CREATE_CQ_TYPE_SHIFT;
 		cq->phase_change = false;
-		cmd->cmd.cqe_count = (cq->len / cqe_size);
+		cmd->cmd.pdid_cqecnt = (cq->len / cqe_size);
 	} else {
-		cmd->cmd.cqe_count = (cq->len / cqe_size) - 1;
+		cmd->cmd.pdid_cqecnt = (cq->len / cqe_size) - 1;
 		cmd->cmd.ev_cnt_flags |= OCRDMA_CREATE_CQ_FLAGS_AUTO_VALID;
 		cq->phase_change = true;
 	}
 
-	cmd->cmd.pd_id = pd_id; /* valid only for v3 */
+	/* pd_id valid only for v3 */
+	cmd->cmd.pdid_cqecnt |= (pd_id <<
+		OCRDMA_CREATE_CQ_CMD_PDID_SHIFT);
 	ocrdma_build_q_pages(&cmd->cmd.pa[0], hw_pages, cq->pa, page_size);
 	status = ocrdma_mbx_cmd(dev, (struct ocrdma_mqe *)cmd);
 	if (status)
@@ -2206,7 +2230,8 @@
 				OCRDMA_CREATE_QP_REQ_RQ_CQID_MASK;
 	qp->rq_cq = cq;
 
-	if (pd->dpp_enabled && pd->num_dpp_qp) {
+	if (pd->dpp_enabled && attrs->cap.max_inline_data && pd->num_dpp_qp &&
+	    (attrs->cap.max_inline_data <= dev->attr.max_inline_data)) {
 		ocrdma_set_create_qp_dpp_cmd(cmd, pd, qp, enable_dpp_cq,
 					     dpp_cq_id);
 	}
@@ -2264,6 +2289,8 @@
 
 	if ((ah_attr->ah_flags & IB_AH_GRH) == 0)
 		return -EINVAL;
+	if (atomic_cmpxchg(&qp->dev->update_sl, 1, 0))
+		ocrdma_init_service_level(qp->dev);
 	cmd->params.tclass_sq_psn |=
 	    (ah_attr->grh.traffic_class << OCRDMA_QP_PARAMS_TCLASS_SHIFT);
 	cmd->params.rnt_rc_sl_fl |=
@@ -2297,6 +2324,8 @@
 		cmd->params.vlan_dmac_b4_to_b5 |=
 		    vlan_id << OCRDMA_QP_PARAMS_VLAN_SHIFT;
 		cmd->flags |= OCRDMA_QP_PARA_VLAN_EN_VALID;
+		cmd->params.rnt_rc_sl_fl |=
+			(qp->dev->sl & 0x07) << OCRDMA_QP_PARAMS_SL_SHIFT;
 	}
 	return 0;
 }
@@ -2604,6 +2633,168 @@
 	return status;
 }
 
+static int ocrdma_mbx_get_dcbx_config(struct ocrdma_dev *dev, u32 ptype,
+				      struct ocrdma_dcbx_cfg *dcbxcfg)
+{
+	int status = 0;
+	dma_addr_t pa;
+	struct ocrdma_mqe cmd;
+
+	struct ocrdma_get_dcbx_cfg_req *req = NULL;
+	struct ocrdma_get_dcbx_cfg_rsp *rsp = NULL;
+	struct pci_dev *pdev = dev->nic_info.pdev;
+	struct ocrdma_mqe_sge *mqe_sge = cmd.u.nonemb_req.sge;
+
+	memset(&cmd, 0, sizeof(struct ocrdma_mqe));
+	cmd.hdr.pyld_len = max_t (u32, sizeof(struct ocrdma_get_dcbx_cfg_rsp),
+					sizeof(struct ocrdma_get_dcbx_cfg_req));
+	req = dma_alloc_coherent(&pdev->dev, cmd.hdr.pyld_len, &pa, GFP_KERNEL);
+	if (!req) {
+		status = -ENOMEM;
+		goto mem_err;
+	}
+
+	cmd.hdr.spcl_sge_cnt_emb |= (1 << OCRDMA_MQE_HDR_SGE_CNT_SHIFT) &
+					OCRDMA_MQE_HDR_SGE_CNT_MASK;
+	mqe_sge->pa_lo = (u32) (pa & 0xFFFFFFFFUL);
+	mqe_sge->pa_hi = (u32) upper_32_bits(pa);
+	mqe_sge->len = cmd.hdr.pyld_len;
+
+	memset(req, 0, sizeof(struct ocrdma_get_dcbx_cfg_req));
+	ocrdma_init_mch(&req->hdr, OCRDMA_CMD_GET_DCBX_CONFIG,
+			OCRDMA_SUBSYS_DCBX, cmd.hdr.pyld_len);
+	req->param_type = ptype;
+
+	status = ocrdma_mbx_cmd(dev, &cmd);
+	if (status)
+		goto mbx_err;
+
+	rsp = (struct ocrdma_get_dcbx_cfg_rsp *)req;
+	ocrdma_le32_to_cpu(rsp, sizeof(struct ocrdma_get_dcbx_cfg_rsp));
+	memcpy(dcbxcfg, &rsp->cfg, sizeof(struct ocrdma_dcbx_cfg));
+
+mbx_err:
+	dma_free_coherent(&pdev->dev, cmd.hdr.pyld_len, req, pa);
+mem_err:
+	return status;
+}
+
+#define OCRDMA_MAX_SERVICE_LEVEL_INDEX	0x08
+#define OCRDMA_DEFAULT_SERVICE_LEVEL	0x05
+
+static int ocrdma_parse_dcbxcfg_rsp(struct ocrdma_dev *dev, int ptype,
+				    struct ocrdma_dcbx_cfg *dcbxcfg,
+				    u8 *srvc_lvl)
+{
+	int status = -EINVAL, indx, slindx;
+	int ventry_cnt;
+	struct ocrdma_app_parameter *app_param;
+	u8 valid, proto_sel;
+	u8 app_prio, pfc_prio;
+	u16 proto;
+
+	if (!(dcbxcfg->tcv_aev_opv_st & OCRDMA_DCBX_STATE_MASK)) {
+		pr_info("%s ocrdma%d DCBX is disabled\n",
+			dev_name(&dev->nic_info.pdev->dev), dev->id);
+		goto out;
+	}
+
+	if (!ocrdma_is_enabled_and_synced(dcbxcfg->pfc_state)) {
+		pr_info("%s ocrdma%d priority flow control(%s) is %s%s\n",
+			dev_name(&dev->nic_info.pdev->dev), dev->id,
+			(ptype > 0 ? "operational" : "admin"),
+			(dcbxcfg->pfc_state & OCRDMA_STATE_FLAG_ENABLED) ?
+			"enabled" : "disabled",
+			(dcbxcfg->pfc_state & OCRDMA_STATE_FLAG_SYNC) ?
+			"" : ", not sync'ed");
+		goto out;
+	} else {
+		pr_info("%s ocrdma%d priority flow control is enabled and sync'ed\n",
+			dev_name(&dev->nic_info.pdev->dev), dev->id);
+	}
+
+	ventry_cnt = (dcbxcfg->tcv_aev_opv_st >>
+				OCRDMA_DCBX_APP_ENTRY_SHIFT)
+				& OCRDMA_DCBX_STATE_MASK;
+
+	for (indx = 0; indx < ventry_cnt; indx++) {
+		app_param = &dcbxcfg->app_param[indx];
+		valid = (app_param->valid_proto_app >>
+				OCRDMA_APP_PARAM_VALID_SHIFT)
+				& OCRDMA_APP_PARAM_VALID_MASK;
+		proto_sel = (app_param->valid_proto_app
+				>>  OCRDMA_APP_PARAM_PROTO_SEL_SHIFT)
+				& OCRDMA_APP_PARAM_PROTO_SEL_MASK;
+		proto = app_param->valid_proto_app &
+				OCRDMA_APP_PARAM_APP_PROTO_MASK;
+
+		if (
+			valid && proto == OCRDMA_APP_PROTO_ROCE &&
+			proto_sel == OCRDMA_PROTO_SELECT_L2) {
+			for (slindx = 0; slindx <
+				OCRDMA_MAX_SERVICE_LEVEL_INDEX; slindx++) {
+				app_prio = ocrdma_get_app_prio(
+						(u8 *)app_param->app_prio,
+						slindx);
+				pfc_prio = ocrdma_get_pfc_prio(
+						(u8 *)dcbxcfg->pfc_prio,
+						slindx);
+
+				if (app_prio && pfc_prio) {
+					*srvc_lvl = slindx;
+					status = 0;
+					goto out;
+				}
+			}
+			if (slindx == OCRDMA_MAX_SERVICE_LEVEL_INDEX) {
+				pr_info("%s ocrdma%d application priority not set for 0x%x protocol\n",
+					dev_name(&dev->nic_info.pdev->dev),
+					dev->id, proto);
+			}
+		}
+	}
+
+out:
+	return status;
+}
+
+void ocrdma_init_service_level(struct ocrdma_dev *dev)
+{
+	int status = 0, indx;
+	struct ocrdma_dcbx_cfg dcbxcfg;
+	u8 srvc_lvl = OCRDMA_DEFAULT_SERVICE_LEVEL;
+	int ptype = OCRDMA_PARAMETER_TYPE_OPER;
+
+	for (indx = 0; indx < 2; indx++) {
+		status = ocrdma_mbx_get_dcbx_config(dev, ptype, &dcbxcfg);
+		if (status) {
+			pr_err("%s(): status=%d\n", __func__, status);
+			ptype = OCRDMA_PARAMETER_TYPE_ADMIN;
+			continue;
+		}
+
+		status = ocrdma_parse_dcbxcfg_rsp(dev, ptype,
+						  &dcbxcfg, &srvc_lvl);
+		if (status) {
+			ptype = OCRDMA_PARAMETER_TYPE_ADMIN;
+			continue;
+		}
+
+		break;
+	}
+
+	if (status)
+		pr_info("%s ocrdma%d service level default\n",
+			dev_name(&dev->nic_info.pdev->dev), dev->id);
+	else
+		pr_info("%s ocrdma%d service level %d\n",
+			dev_name(&dev->nic_info.pdev->dev), dev->id,
+			srvc_lvl);
+
+	dev->pfc_state = ocrdma_is_enabled_and_synced(dcbxcfg.pfc_state);
+	dev->sl = srvc_lvl;
+}
+
 int ocrdma_alloc_av(struct ocrdma_dev *dev, struct ocrdma_ah *ah)
 {
 	int i;
@@ -2709,13 +2900,15 @@
 		goto conf_err;
 	status = ocrdma_mbx_get_phy_info(dev);
 	if (status)
-		goto conf_err;
+		goto info_attrb_err;
 	status = ocrdma_mbx_get_ctrl_attribs(dev);
 	if (status)
-		goto conf_err;
+		goto info_attrb_err;
 
 	return 0;
 
+info_attrb_err:
+	ocrdma_mbx_delete_ah_tbl(dev);
 conf_err:
 	ocrdma_destroy_mq(dev);
 mq_err:
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_hw.h b/drivers/infiniband/hw/ocrdma/ocrdma_hw.h
index e513f72..6eed8f19 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_hw.h
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_hw.h
@@ -135,4 +135,6 @@
 
 int ocrdma_mbx_rdma_stats(struct ocrdma_dev *, bool reset);
 char *port_speed_string(struct ocrdma_dev *dev);
+void ocrdma_init_service_level(struct ocrdma_dev *);
+
 #endif				/* __OCRDMA_HW_H__ */
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_main.c b/drivers/infiniband/hw/ocrdma/ocrdma_main.c
index 7c504e0..256a06b 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_main.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_main.c
@@ -324,6 +324,11 @@
 		if (!dev->qp_tbl)
 			goto alloc_err;
 	}
+
+	dev->stag_arr = kzalloc(sizeof(u64) * OCRDMA_MAX_STAG, GFP_KERNEL);
+	if (dev->stag_arr == NULL)
+		goto alloc_err;
+
 	spin_lock_init(&dev->av_tbl.lock);
 	spin_lock_init(&dev->flush_q_lock);
 	return 0;
@@ -334,6 +339,7 @@
 
 static void ocrdma_free_resources(struct ocrdma_dev *dev)
 {
+	kfree(dev->stag_arr);
 	kfree(dev->qp_tbl);
 	kfree(dev->cq_tbl);
 	kfree(dev->sgid_tbl);
@@ -353,15 +359,25 @@
 {
 	struct ocrdma_dev *dev = dev_get_drvdata(device);
 
-	return scnprintf(buf, PAGE_SIZE, "%s", &dev->attr.fw_ver[0]);
+	return scnprintf(buf, PAGE_SIZE, "%s\n", &dev->attr.fw_ver[0]);
+}
+
+static ssize_t show_hca_type(struct device *device,
+			     struct device_attribute *attr, char *buf)
+{
+	struct ocrdma_dev *dev = dev_get_drvdata(device);
+
+	return scnprintf(buf, PAGE_SIZE, "%s\n", &dev->model_number[0]);
 }
 
 static DEVICE_ATTR(hw_rev, S_IRUGO, show_rev, NULL);
 static DEVICE_ATTR(fw_ver, S_IRUGO, show_fw_ver, NULL);
+static DEVICE_ATTR(hca_type, S_IRUGO, show_hca_type, NULL);
 
 static struct device_attribute *ocrdma_attributes[] = {
 	&dev_attr_hw_rev,
-	&dev_attr_fw_ver
+	&dev_attr_fw_ver,
+	&dev_attr_hca_type
 };
 
 static void ocrdma_remove_sysfiles(struct ocrdma_dev *dev)
@@ -372,6 +388,58 @@
 		device_remove_file(&dev->ibdev.dev, ocrdma_attributes[i]);
 }
 
+static void ocrdma_init_ipv4_gids(struct ocrdma_dev *dev,
+				  struct net_device *net)
+{
+	struct in_device *in_dev;
+	union ib_gid gid;
+	in_dev = in_dev_get(net);
+	if (in_dev) {
+		for_ifa(in_dev) {
+			ipv6_addr_set_v4mapped(ifa->ifa_address,
+					       (struct in6_addr *)&gid);
+			ocrdma_add_sgid(dev, &gid);
+		}
+		endfor_ifa(in_dev);
+		in_dev_put(in_dev);
+	}
+}
+
+static void ocrdma_init_ipv6_gids(struct ocrdma_dev *dev,
+				  struct net_device *net)
+{
+#if IS_ENABLED(CONFIG_IPV6)
+	struct inet6_dev *in6_dev;
+	union ib_gid  *pgid;
+	struct inet6_ifaddr *ifp;
+	in6_dev = in6_dev_get(net);
+	if (in6_dev) {
+		read_lock_bh(&in6_dev->lock);
+		list_for_each_entry(ifp, &in6_dev->addr_list, if_list) {
+			pgid = (union ib_gid *)&ifp->addr;
+			ocrdma_add_sgid(dev, pgid);
+		}
+		read_unlock_bh(&in6_dev->lock);
+		in6_dev_put(in6_dev);
+	}
+#endif
+}
+
+static void ocrdma_init_gid_table(struct ocrdma_dev *dev)
+{
+	struct  net_device *net_dev;
+
+	for_each_netdev(&init_net, net_dev) {
+		struct net_device *real_dev = rdma_vlan_dev_real_dev(net_dev) ?
+				rdma_vlan_dev_real_dev(net_dev) : net_dev;
+
+		if (real_dev == dev->nic_info.netdev) {
+			ocrdma_init_ipv4_gids(dev, net_dev);
+			ocrdma_init_ipv6_gids(dev, net_dev);
+		}
+	}
+}
+
 static struct ocrdma_dev *ocrdma_add(struct be_dev_info *dev_info)
 {
 	int status = 0, i;
@@ -399,6 +467,8 @@
 	if (status)
 		goto alloc_err;
 
+	ocrdma_init_service_level(dev);
+	ocrdma_init_gid_table(dev);
 	status = ocrdma_register_device(dev);
 	if (status)
 		goto alloc_err;
@@ -508,6 +578,12 @@
 	return 0;
 }
 
+static void ocrdma_shutdown(struct ocrdma_dev *dev)
+{
+	ocrdma_close(dev);
+	ocrdma_remove(dev);
+}
+
 /* event handling via NIC driver ensures that all the NIC specific
  * initialization done before RoCE driver notifies
  * event to stack.
@@ -521,6 +597,9 @@
 	case BE_DEV_DOWN:
 		ocrdma_close(dev);
 		break;
+	case BE_DEV_SHUTDOWN:
+		ocrdma_shutdown(dev);
+		break;
 	}
 }
 
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_sli.h b/drivers/infiniband/hw/ocrdma/ocrdma_sli.h
index 96c9ee6..904989e 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_sli.h
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_sli.h
@@ -44,35 +44,39 @@
 #define OCRDMA_SUBSYS_ROCE 10
 enum {
 	OCRDMA_CMD_QUERY_CONFIG = 1,
-	OCRDMA_CMD_ALLOC_PD,
-	OCRDMA_CMD_DEALLOC_PD,
+	OCRDMA_CMD_ALLOC_PD = 2,
+	OCRDMA_CMD_DEALLOC_PD = 3,
 
-	OCRDMA_CMD_CREATE_AH_TBL,
-	OCRDMA_CMD_DELETE_AH_TBL,
+	OCRDMA_CMD_CREATE_AH_TBL = 4,
+	OCRDMA_CMD_DELETE_AH_TBL = 5,
 
-	OCRDMA_CMD_CREATE_QP,
-	OCRDMA_CMD_QUERY_QP,
-	OCRDMA_CMD_MODIFY_QP,
-	OCRDMA_CMD_DELETE_QP,
+	OCRDMA_CMD_CREATE_QP = 6,
+	OCRDMA_CMD_QUERY_QP = 7,
+	OCRDMA_CMD_MODIFY_QP = 8 ,
+	OCRDMA_CMD_DELETE_QP = 9,
 
-	OCRDMA_CMD_RSVD1,
-	OCRDMA_CMD_ALLOC_LKEY,
-	OCRDMA_CMD_DEALLOC_LKEY,
-	OCRDMA_CMD_REGISTER_NSMR,
-	OCRDMA_CMD_REREGISTER_NSMR,
-	OCRDMA_CMD_REGISTER_NSMR_CONT,
-	OCRDMA_CMD_QUERY_NSMR,
-	OCRDMA_CMD_ALLOC_MW,
-	OCRDMA_CMD_QUERY_MW,
+	OCRDMA_CMD_RSVD1 = 10,
+	OCRDMA_CMD_ALLOC_LKEY = 11,
+	OCRDMA_CMD_DEALLOC_LKEY = 12,
+	OCRDMA_CMD_REGISTER_NSMR = 13,
+	OCRDMA_CMD_REREGISTER_NSMR = 14,
+	OCRDMA_CMD_REGISTER_NSMR_CONT = 15,
+	OCRDMA_CMD_QUERY_NSMR = 16,
+	OCRDMA_CMD_ALLOC_MW = 17,
+	OCRDMA_CMD_QUERY_MW = 18,
 
-	OCRDMA_CMD_CREATE_SRQ,
-	OCRDMA_CMD_QUERY_SRQ,
-	OCRDMA_CMD_MODIFY_SRQ,
-	OCRDMA_CMD_DELETE_SRQ,
+	OCRDMA_CMD_CREATE_SRQ = 19,
+	OCRDMA_CMD_QUERY_SRQ = 20,
+	OCRDMA_CMD_MODIFY_SRQ = 21,
+	OCRDMA_CMD_DELETE_SRQ = 22,
 
-	OCRDMA_CMD_ATTACH_MCAST,
-	OCRDMA_CMD_DETACH_MCAST,
-	OCRDMA_CMD_GET_RDMA_STATS,
+	OCRDMA_CMD_ATTACH_MCAST = 23,
+	OCRDMA_CMD_DETACH_MCAST = 24,
+
+	OCRDMA_CMD_CREATE_RBQ = 25,
+	OCRDMA_CMD_DESTROY_RBQ = 26,
+
+	OCRDMA_CMD_GET_RDMA_STATS = 27,
 
 	OCRDMA_CMD_MAX
 };
@@ -103,7 +107,7 @@
 
 #define OCRDMA_MAX_QP    2048
 #define OCRDMA_MAX_CQ    2048
-#define OCRDMA_MAX_STAG  8192
+#define OCRDMA_MAX_STAG 16384
 
 enum {
 	OCRDMA_DB_RQ_OFFSET		= 0xE0,
@@ -422,7 +426,12 @@
 
 #define OCRDMA_ASYNC_RDMA_EVE_CODE 0x14
 #define OCRDMA_ASYNC_GRP5_EVE_CODE 0x5
-#define OCRDMA_ASYNC_EVENT_PVID_STATE 0x3
+
+enum ocrdma_async_grp5_events {
+	OCRDMA_ASYNC_EVENT_QOS_VALUE	= 0x01,
+	OCRDMA_ASYNC_EVENT_COS_VALUE	= 0x02,
+	OCRDMA_ASYNC_EVENT_PVID_STATE	= 0x03
+};
 
 enum OCRDMA_ASYNC_EVENT_TYPE {
 	OCRDMA_CQ_ERROR			= 0x00,
@@ -525,8 +534,8 @@
 	u32 max_ird_ord_per_qp;
 	u32 max_shared_ird_ord;
 	u32 max_mr;
-	u32 max_mr_size_lo;
 	u32 max_mr_size_hi;
+	u32 max_mr_size_lo;
 	u32 max_num_mr_pbl;
 	u32 max_mw;
 	u32 max_fmr;
@@ -580,17 +589,26 @@
 	OCRDMA_FN_MODE_RDMA	= 0x4
 };
 
+enum {
+	OCRDMA_IF_TYPE_MASK		= 0xFFFF0000,
+	OCRDMA_IF_TYPE_SHIFT		= 0x10,
+	OCRDMA_PHY_TYPE_MASK		= 0x0000FFFF,
+	OCRDMA_FUTURE_DETAILS_MASK	= 0xFFFF0000,
+	OCRDMA_FUTURE_DETAILS_SHIFT	= 0x10,
+	OCRDMA_EX_PHY_DETAILS_MASK	= 0x0000FFFF,
+	OCRDMA_FSPEED_SUPP_MASK		= 0xFFFF0000,
+	OCRDMA_FSPEED_SUPP_SHIFT	= 0x10,
+	OCRDMA_ASPEED_SUPP_MASK		= 0x0000FFFF
+};
+
 struct ocrdma_get_phy_info_rsp {
 	struct ocrdma_mqe_hdr hdr;
 	struct ocrdma_mbx_rsp rsp;
 
-	u16 phy_type;
-	u16 interface_type;
+	u32 ityp_ptyp;
 	u32 misc_params;
-	u16 ext_phy_details;
-	u16 rsvd;
-	u16 auto_speeds_supported;
-	u16 fixed_speeds_supported;
+	u32 ftrdtl_exphydtl;
+	u32 fspeed_aspeed;
 	u32 future_use[2];
 };
 
@@ -603,19 +621,34 @@
 	OCRDMA_PHY_SPEED_40GBPS = 0x20
 };
 
+enum {
+	OCRDMA_PORT_NUM_MASK	= 0x3F,
+	OCRDMA_PT_MASK		= 0xC0,
+	OCRDMA_PT_SHIFT		= 0x6,
+	OCRDMA_LINK_DUP_MASK	= 0x0000FF00,
+	OCRDMA_LINK_DUP_SHIFT	= 0x8,
+	OCRDMA_PHY_PS_MASK	= 0x00FF0000,
+	OCRDMA_PHY_PS_SHIFT	= 0x10,
+	OCRDMA_PHY_PFLT_MASK	= 0xFF000000,
+	OCRDMA_PHY_PFLT_SHIFT	= 0x18,
+	OCRDMA_QOS_LNKSP_MASK	= 0xFFFF0000,
+	OCRDMA_QOS_LNKSP_SHIFT	= 0x10,
+	OCRDMA_LLST_MASK	= 0xFF,
+	OCRDMA_PLFC_MASK	= 0x00000400,
+	OCRDMA_PLFC_SHIFT	= 0x8,
+	OCRDMA_PLRFC_MASK	= 0x00000200,
+	OCRDMA_PLRFC_SHIFT	= 0x8,
+	OCRDMA_PLTFC_MASK	= 0x00000100,
+	OCRDMA_PLTFC_SHIFT	= 0x8
+};
 
 struct ocrdma_get_link_speed_rsp {
 	struct ocrdma_mqe_hdr hdr;
 	struct ocrdma_mbx_rsp rsp;
 
-	u8 pt_port_num;
-	u8 link_duplex;
-	u8 phys_port_speed;
-	u8 phys_port_fault;
-	u16 rsvd1;
-	u16 qos_lnk_speed;
-	u8 logical_lnk_status;
-	u8 rsvd2[3];
+	u32 pflt_pps_ld_pnum;
+	u32 qos_lsp;
+	u32 res_lls;
 };
 
 enum {
@@ -666,8 +699,7 @@
 	u32 pgsz_pgcnt;
 	u32 ev_cnt_flags;
 	u32 eqn;
-	u16 cqe_count;
-	u16 pd_id;
+	u32 pdid_cqecnt;
 	u32 rsvd6;
 	struct ocrdma_pa pa[OCRDMA_CREATE_CQ_MAX_PAGES];
 };
@@ -678,6 +710,10 @@
 };
 
 enum {
+	OCRDMA_CREATE_CQ_CMD_PDID_SHIFT	= 0x10
+};
+
+enum {
 	OCRDMA_CREATE_CQ_RSP_CQ_ID_MASK	= 0xFFFF
 };
 
@@ -1231,7 +1267,6 @@
 
 enum {
 	OCRDMA_ALLOC_PD_ENABLE_DPP	= BIT(16),
-	OCRDMA_PD_MAX_DPP_ENABLED_QP	= 8,
 	OCRDMA_DPP_PAGE_SIZE		= 4096
 };
 
@@ -1896,12 +1931,62 @@
 	struct ocrdma_rx_dbg_stats	rx_dbg_stats;
 } __packed;
 
+enum {
+	OCRDMA_HBA_ATTRB_EPROM_VER_LO_MASK	= 0xFF,
+	OCRDMA_HBA_ATTRB_EPROM_VER_HI_MASK	= 0xFF00,
+	OCRDMA_HBA_ATTRB_EPROM_VER_HI_SHIFT	= 0x08,
+	OCRDMA_HBA_ATTRB_CDBLEN_MASK		= 0xFFFF,
+	OCRDMA_HBA_ATTRB_ASIC_REV_MASK		= 0xFF0000,
+	OCRDMA_HBA_ATTRB_ASIC_REV_SHIFT		= 0x10,
+	OCRDMA_HBA_ATTRB_GUID0_MASK		= 0xFF000000,
+	OCRDMA_HBA_ATTRB_GUID0_SHIFT		= 0x18,
+	OCRDMA_HBA_ATTRB_GUID13_MASK		= 0xFF,
+	OCRDMA_HBA_ATTRB_GUID14_MASK		= 0xFF00,
+	OCRDMA_HBA_ATTRB_GUID14_SHIFT		= 0x08,
+	OCRDMA_HBA_ATTRB_GUID15_MASK		= 0xFF0000,
+	OCRDMA_HBA_ATTRB_GUID15_SHIFT		= 0x10,
+	OCRDMA_HBA_ATTRB_PCNT_MASK		= 0xFF000000,
+	OCRDMA_HBA_ATTRB_PCNT_SHIFT		= 0x18,
+	OCRDMA_HBA_ATTRB_LDTOUT_MASK		= 0xFFFF,
+	OCRDMA_HBA_ATTRB_ISCSI_VER_MASK		= 0xFF0000,
+	OCRDMA_HBA_ATTRB_ISCSI_VER_SHIFT	= 0x10,
+	OCRDMA_HBA_ATTRB_MFUNC_DEV_MASK		= 0xFF000000,
+	OCRDMA_HBA_ATTRB_MFUNC_DEV_SHIFT	= 0x18,
+	OCRDMA_HBA_ATTRB_CV_MASK		= 0xFF,
+	OCRDMA_HBA_ATTRB_HBA_ST_MASK		= 0xFF00,
+	OCRDMA_HBA_ATTRB_HBA_ST_SHIFT		= 0x08,
+	OCRDMA_HBA_ATTRB_MAX_DOMS_MASK		= 0xFF0000,
+	OCRDMA_HBA_ATTRB_MAX_DOMS_SHIFT		= 0x10,
+	OCRDMA_HBA_ATTRB_PTNUM_MASK		= 0x3F000000,
+	OCRDMA_HBA_ATTRB_PTNUM_SHIFT		= 0x18,
+	OCRDMA_HBA_ATTRB_PT_MASK		= 0xC0000000,
+	OCRDMA_HBA_ATTRB_PT_SHIFT		= 0x1E,
+	OCRDMA_HBA_ATTRB_ISCSI_FET_MASK		= 0xFF,
+	OCRDMA_HBA_ATTRB_ASIC_GEN_MASK		= 0xFF00,
+	OCRDMA_HBA_ATTRB_ASIC_GEN_SHIFT		= 0x08,
+	OCRDMA_HBA_ATTRB_PCI_VID_MASK		= 0xFFFF,
+	OCRDMA_HBA_ATTRB_PCI_DID_MASK		= 0xFFFF0000,
+	OCRDMA_HBA_ATTRB_PCI_DID_SHIFT		= 0x10,
+	OCRDMA_HBA_ATTRB_PCI_SVID_MASK		= 0xFFFF,
+	OCRDMA_HBA_ATTRB_PCI_SSID_MASK		= 0xFFFF0000,
+	OCRDMA_HBA_ATTRB_PCI_SSID_SHIFT		= 0x10,
+	OCRDMA_HBA_ATTRB_PCI_BUSNUM_MASK	= 0xFF,
+	OCRDMA_HBA_ATTRB_PCI_DEVNUM_MASK	= 0xFF00,
+	OCRDMA_HBA_ATTRB_PCI_DEVNUM_SHIFT	= 0x08,
+	OCRDMA_HBA_ATTRB_PCI_FUNCNUM_MASK	= 0xFF0000,
+	OCRDMA_HBA_ATTRB_PCI_FUNCNUM_SHIFT	= 0x10,
+	OCRDMA_HBA_ATTRB_IF_TYPE_MASK		= 0xFF000000,
+	OCRDMA_HBA_ATTRB_IF_TYPE_SHIFT		= 0x18,
+	OCRDMA_HBA_ATTRB_NETFIL_MASK		=0xFF
+};
 
 struct mgmt_hba_attribs {
 	u8 flashrom_version_string[32];
 	u8 manufacturer_name[32];
 	u32 supported_modes;
-	u32 rsvd0[3];
+	u32 rsvd_eprom_verhi_verlo;
+	u32 mbx_ds_ver;
+	u32 epfw_ds_ver;
 	u8 ncsi_ver_string[12];
 	u32 default_extended_timeout;
 	u8 controller_model_number[32];
@@ -1914,34 +1999,26 @@
 	u8 driver_version_string[32];
 	u8 fw_on_flash_version_string[32];
 	u32 functionalities_supported;
-	u16 max_cdblength;
-	u8 asic_revision;
-	u8 generational_guid[16];
-	u8 hba_port_count;
-	u16 default_link_down_timeout;
-	u8 iscsi_ver_min_max;
-	u8 multifunction_device;
-	u8 cache_valid;
-	u8 hba_status;
-	u8 max_domains_supported;
-	u8 phy_port;
+	u32 guid0_asicrev_cdblen;
+	u8 generational_guid[12];
+	u32 portcnt_guid15;
+	u32 mfuncdev_iscsi_ldtout;
+	u32 ptpnum_maxdoms_hbast_cv;
 	u32 firmware_post_status;
 	u32 hba_mtu[8];
-	u32 rsvd1[4];
+	u32 res_asicgen_iscsi_feaures;
+	u32 rsvd1[3];
 };
 
 struct mgmt_controller_attrib {
 	struct mgmt_hba_attribs hba_attribs;
-	u16 pci_vendor_id;
-	u16 pci_device_id;
-	u16 pci_sub_vendor_id;
-	u16 pci_sub_system_id;
-	u8 pci_bus_number;
-	u8 pci_device_number;
-	u8 pci_function_number;
-	u8 interface_type;
-	u64 unique_identifier;
-	u32 rsvd0[5];
+	u32 pci_did_vid;
+	u32 pci_ssid_svid;
+	u32 ityp_fnum_devnum_bnum;
+	u32 uid_hi;
+	u32 uid_lo;
+	u32 res_nnetfil;
+	u32 rsvd0[4];
 };
 
 struct ocrdma_get_ctrl_attribs_rsp {
@@ -1949,5 +2026,79 @@
 	struct mgmt_controller_attrib ctrl_attribs;
 };
 
+#define OCRDMA_SUBSYS_DCBX 0x10
+
+enum OCRDMA_DCBX_OPCODE {
+	OCRDMA_CMD_GET_DCBX_CONFIG = 0x01
+};
+
+enum OCRDMA_DCBX_PARAM_TYPE {
+	OCRDMA_PARAMETER_TYPE_ADMIN	= 0x00,
+	OCRDMA_PARAMETER_TYPE_OPER	= 0x01,
+	OCRDMA_PARAMETER_TYPE_PEER	= 0x02
+};
+
+enum OCRDMA_DCBX_APP_PROTO {
+	OCRDMA_APP_PROTO_ROCE	= 0x8915
+};
+
+enum OCRDMA_DCBX_PROTO {
+	OCRDMA_PROTO_SELECT_L2	= 0x00,
+	OCRDMA_PROTO_SELECT_L4	= 0x01
+};
+
+enum OCRDMA_DCBX_APP_PARAM {
+	OCRDMA_APP_PARAM_APP_PROTO_MASK = 0xFFFF,
+	OCRDMA_APP_PARAM_PROTO_SEL_MASK = 0xFF,
+	OCRDMA_APP_PARAM_PROTO_SEL_SHIFT = 0x10,
+	OCRDMA_APP_PARAM_VALID_MASK	= 0xFF,
+	OCRDMA_APP_PARAM_VALID_SHIFT	= 0x18
+};
+
+enum OCRDMA_DCBX_STATE_FLAGS {
+	OCRDMA_STATE_FLAG_ENABLED	= 0x01,
+	OCRDMA_STATE_FLAG_ADDVERTISED	= 0x02,
+	OCRDMA_STATE_FLAG_WILLING	= 0x04,
+	OCRDMA_STATE_FLAG_SYNC		= 0x08,
+	OCRDMA_STATE_FLAG_UNSUPPORTED	= 0x40000000,
+	OCRDMA_STATE_FLAG_NEG_FAILD	= 0x80000000
+};
+
+enum OCRDMA_TCV_AEV_OPV_ST {
+	OCRDMA_DCBX_TC_SUPPORT_MASK	= 0xFF,
+	OCRDMA_DCBX_TC_SUPPORT_SHIFT	= 0x18,
+	OCRDMA_DCBX_APP_ENTRY_SHIFT	= 0x10,
+	OCRDMA_DCBX_OP_PARAM_SHIFT	= 0x08,
+	OCRDMA_DCBX_STATE_MASK		= 0xFF
+};
+
+struct ocrdma_app_parameter {
+	u32 valid_proto_app;
+	u32 oui;
+	u32 app_prio[2];
+};
+
+struct ocrdma_dcbx_cfg {
+	u32 tcv_aev_opv_st;
+	u32 tc_state;
+	u32 pfc_state;
+	u32 qcn_state;
+	u32 appl_state;
+	u32 ll_state;
+	u32 tc_bw[2];
+	u32 tc_prio[8];
+	u32 pfc_prio[2];
+	struct ocrdma_app_parameter app_param[15];
+};
+
+struct ocrdma_get_dcbx_cfg_req {
+	struct ocrdma_mbx_hdr hdr;
+	u32 param_type;
+} __packed;
+
+struct ocrdma_get_dcbx_cfg_rsp {
+	struct ocrdma_mbx_rsp hdr;
+	struct ocrdma_dcbx_cfg cfg;
+} __packed;
 
 #endif				/* __OCRDMA_SLI_H__ */
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
index edf6211..acb434d 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
@@ -69,11 +69,11 @@
 	memcpy(&attr->fw_ver, &dev->attr.fw_ver[0],
 	       min(sizeof(dev->attr.fw_ver), sizeof(attr->fw_ver)));
 	ocrdma_get_guid(dev, (u8 *)&attr->sys_image_guid);
-	attr->max_mr_size = ~0ull;
+	attr->max_mr_size = dev->attr.max_mr_size;
 	attr->page_size_cap = 0xffff000;
 	attr->vendor_id = dev->nic_info.pdev->vendor;
 	attr->vendor_part_id = dev->nic_info.pdev->device;
-	attr->hw_ver = 0;
+	attr->hw_ver = dev->asic_id;
 	attr->max_qp = dev->attr.max_qp;
 	attr->max_ah = OCRDMA_MAX_AH;
 	attr->max_qp_wr = dev->attr.max_wqe;
@@ -268,7 +268,8 @@
 		pd->dpp_enabled =
 			ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R;
 		pd->num_dpp_qp =
-			pd->dpp_enabled ? OCRDMA_PD_MAX_DPP_ENABLED_QP : 0;
+			pd->dpp_enabled ? (dev->nic_info.db_page_size /
+					   dev->attr.wqe_size) : 0;
 	}
 
 retry:
@@ -328,7 +329,10 @@
 	struct ocrdma_pd *pd = uctx->cntxt_pd;
 	struct ocrdma_dev *dev = get_ocrdma_dev(pd->ibpd.device);
 
-	BUG_ON(uctx->pd_in_use);
+	if (uctx->pd_in_use) {
+		pr_err("%s(%d) Freeing in use pdid=0x%x.\n",
+		       __func__, dev->id, pd->id);
+	}
 	uctx->cntxt_pd = NULL;
 	status = _ocrdma_dealloc_pd(dev, pd);
 	return status;
@@ -843,6 +847,13 @@
 	if (mr->umem)
 		ib_umem_release(mr->umem);
 	kfree(mr);
+
+	/* Don't stop cleanup, in case FW is unresponsive */
+	if (dev->mqe_ctx.fw_error_state) {
+		status = 0;
+		pr_err("%s(%d) fw not responding.\n",
+		       __func__, dev->id);
+	}
 	return status;
 }
 
@@ -2054,6 +2065,13 @@
 	}
 
 	while (wr) {
+		if (qp->qp_type == IB_QPT_UD &&
+		    (wr->opcode != IB_WR_SEND &&
+		     wr->opcode != IB_WR_SEND_WITH_IMM)) {
+			*bad_wr = wr;
+			status = -EINVAL;
+			break;
+		}
 		if (ocrdma_hwq_free_cnt(&qp->sq) == 0 ||
 		    wr->num_sge > qp->sq.max_sges) {
 			*bad_wr = wr;
@@ -2488,6 +2506,11 @@
 			*stop = true;
 			expand = false;
 		}
+	} else if (is_hw_sq_empty(qp)) {
+		/* Do nothing */
+		expand = false;
+		*polled = false;
+		*stop = false;
 	} else {
 		*polled = true;
 		expand = ocrdma_update_err_scqe(ibwc, cqe, qp, status);
@@ -2593,6 +2616,11 @@
 			*stop = true;
 			expand = false;
 		}
+	} else if (is_hw_rq_empty(qp)) {
+		/* Do nothing */
+		expand = false;
+		*polled = false;
+		*stop = false;
 	} else {
 		*polled = true;
 		expand = ocrdma_update_err_rcqe(ibwc, cqe, qp, status);
diff --git a/drivers/infiniband/hw/qib/qib_mad.c b/drivers/infiniband/hw/qib/qib_mad.c
index 22c720e..636be11 100644
--- a/drivers/infiniband/hw/qib/qib_mad.c
+++ b/drivers/infiniband/hw/qib/qib_mad.c
@@ -2476,7 +2476,7 @@
 		ibp = &dd->pport[p].ibport_data;
 		agent = ib_register_mad_agent(&dev->ibdev, p + 1, IB_QPT_SMI,
 					      NULL, 0, send_handler,
-					      NULL, NULL);
+					      NULL, NULL, 0);
 		if (IS_ERR(agent)) {
 			ret = PTR_ERR(agent);
 			goto err;
diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h
index c639f90..3edce61 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib.h
+++ b/drivers/infiniband/ulp/ipoib/ipoib.h
@@ -86,7 +86,6 @@
 	IPOIB_FLAG_INITIALIZED	  = 1,
 	IPOIB_FLAG_ADMIN_UP	  = 2,
 	IPOIB_PKEY_ASSIGNED	  = 3,
-	IPOIB_PKEY_STOP		  = 4,
 	IPOIB_FLAG_SUBINTERFACE	  = 5,
 	IPOIB_MCAST_RUN		  = 6,
 	IPOIB_STOP_REAPER	  = 7,
@@ -312,7 +311,6 @@
 	struct list_head multicast_list;
 	struct rb_root multicast_tree;
 
-	struct delayed_work pkey_poll_task;
 	struct delayed_work mcast_task;
 	struct work_struct carrier_on_task;
 	struct work_struct flush_light;
@@ -473,10 +471,11 @@
 void ipoib_pkey_event(struct work_struct *work);
 void ipoib_ib_dev_cleanup(struct net_device *dev);
 
-int ipoib_ib_dev_open(struct net_device *dev);
+int ipoib_ib_dev_open(struct net_device *dev, int flush);
 int ipoib_ib_dev_up(struct net_device *dev);
 int ipoib_ib_dev_down(struct net_device *dev, int flush);
 int ipoib_ib_dev_stop(struct net_device *dev, int flush);
+void ipoib_pkey_dev_check_presence(struct net_device *dev);
 
 int ipoib_dev_init(struct net_device *dev, struct ib_device *ca, int port);
 void ipoib_dev_cleanup(struct net_device *dev);
@@ -532,8 +531,7 @@
 
 void ipoib_setup(struct net_device *dev);
 
-void ipoib_pkey_poll(struct work_struct *work);
-int ipoib_pkey_dev_delay_open(struct net_device *dev);
+void ipoib_pkey_open(struct ipoib_dev_priv *priv);
 void ipoib_drain_cq(struct net_device *dev);
 
 void ipoib_set_ethtool_ops(struct net_device *dev);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_fs.c b/drivers/infiniband/ulp/ipoib/ipoib_fs.c
index 5006185..6bd5740 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_fs.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_fs.c
@@ -281,10 +281,8 @@
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 
-	if (priv->mcg_dentry)
-		debugfs_remove(priv->mcg_dentry);
-	if (priv->path_dentry)
-		debugfs_remove(priv->path_dentry);
+	debugfs_remove(priv->mcg_dentry);
+	debugfs_remove(priv->path_dentry);
 }
 
 int ipoib_register_debugfs(void)
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index 6a7003d..72626c34 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -664,17 +664,18 @@
 	drain_tx_cq((struct net_device *)ctx);
 }
 
-int ipoib_ib_dev_open(struct net_device *dev)
+int ipoib_ib_dev_open(struct net_device *dev, int flush)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	int ret;
 
-	if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &priv->pkey_index)) {
-		ipoib_warn(priv, "P_Key 0x%04x not found\n", priv->pkey);
-		clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
+	ipoib_pkey_dev_check_presence(dev);
+
+	if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags)) {
+		ipoib_warn(priv, "P_Key 0x%04x is %s\n", priv->pkey,
+			   (!(priv->pkey & 0x7fff) ? "Invalid" : "not found"));
 		return -1;
 	}
-	set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
 
 	ret = ipoib_init_qp(dev);
 	if (ret) {
@@ -705,16 +706,17 @@
 dev_stop:
 	if (!test_and_set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags))
 		napi_enable(&priv->napi);
-	ipoib_ib_dev_stop(dev, 1);
+	ipoib_ib_dev_stop(dev, flush);
 	return -1;
 }
 
-static void ipoib_pkey_dev_check_presence(struct net_device *dev)
+void ipoib_pkey_dev_check_presence(struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
-	u16 pkey_index = 0;
 
-	if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &pkey_index))
+	if (!(priv->pkey & 0x7fff) ||
+	    ib_find_pkey(priv->ca, priv->port, priv->pkey,
+			 &priv->pkey_index))
 		clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
 	else
 		set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
@@ -745,14 +747,6 @@
 	clear_bit(IPOIB_FLAG_OPER_UP, &priv->flags);
 	netif_carrier_off(dev);
 
-	/* Shutdown the P_Key thread if still active */
-	if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags)) {
-		mutex_lock(&pkey_mutex);
-		set_bit(IPOIB_PKEY_STOP, &priv->flags);
-		cancel_delayed_work_sync(&priv->pkey_poll_task);
-		mutex_unlock(&pkey_mutex);
-	}
-
 	ipoib_mcast_stop_thread(dev, flush);
 	ipoib_mcast_dev_flush(dev);
 
@@ -924,7 +918,7 @@
 		    (unsigned long) dev);
 
 	if (dev->flags & IFF_UP) {
-		if (ipoib_ib_dev_open(dev)) {
+		if (ipoib_ib_dev_open(dev, 1)) {
 			ipoib_transport_dev_cleanup(dev);
 			return -ENODEV;
 		}
@@ -966,13 +960,27 @@
 
 	return 1;
 }
+/*
+ * returns 0 if pkey value was found in a different slot.
+ */
+static inline int update_child_pkey(struct ipoib_dev_priv *priv)
+{
+	u16 old_index = priv->pkey_index;
+
+	priv->pkey_index = 0;
+	ipoib_pkey_dev_check_presence(priv->dev);
+
+	if (test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags) &&
+	    (old_index == priv->pkey_index))
+		return 1;
+	return 0;
+}
 
 static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv,
 				enum ipoib_flush_level level)
 {
 	struct ipoib_dev_priv *cpriv;
 	struct net_device *dev = priv->dev;
-	u16 new_index;
 	int result;
 
 	down_read(&priv->vlan_rwsem);
@@ -986,16 +994,20 @@
 
 	up_read(&priv->vlan_rwsem);
 
-	if (!test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags)) {
-		/* for non-child devices must check/update the pkey value here */
-		if (level == IPOIB_FLUSH_HEAVY &&
-		    !test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags))
-			update_parent_pkey(priv);
+	if (!test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags) &&
+	    level != IPOIB_FLUSH_HEAVY) {
 		ipoib_dbg(priv, "Not flushing - IPOIB_FLAG_INITIALIZED not set.\n");
 		return;
 	}
 
 	if (!test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)) {
+		/* interface is down. update pkey and leave. */
+		if (level == IPOIB_FLUSH_HEAVY) {
+			if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags))
+				update_parent_pkey(priv);
+			else
+				update_child_pkey(priv);
+		}
 		ipoib_dbg(priv, "Not flushing - IPOIB_FLAG_ADMIN_UP not set.\n");
 		return;
 	}
@@ -1005,20 +1017,13 @@
 		 * (parent) devices should always takes what present in pkey index 0
 		 */
 		if (test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) {
-			if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &new_index)) {
-				clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
-				ipoib_ib_dev_down(dev, 0);
-				ipoib_ib_dev_stop(dev, 0);
-				if (ipoib_pkey_dev_delay_open(dev))
-					return;
-			}
-			/* restart QP only if P_Key index is changed */
-			if (test_and_set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags) &&
-			    new_index == priv->pkey_index) {
+			result = update_child_pkey(priv);
+			if (result) {
+				/* restart QP only if P_Key index is changed */
 				ipoib_dbg(priv, "Not flushing - P_Key index not changed.\n");
 				return;
 			}
-			priv->pkey_index = new_index;
+
 		} else {
 			result = update_parent_pkey(priv);
 			/* restart QP only if P_Key value changed */
@@ -1038,8 +1043,12 @@
 		ipoib_ib_dev_down(dev, 0);
 
 	if (level == IPOIB_FLUSH_HEAVY) {
-		ipoib_ib_dev_stop(dev, 0);
-		ipoib_ib_dev_open(dev);
+		if (test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags))
+			ipoib_ib_dev_stop(dev, 0);
+		if (ipoib_ib_dev_open(dev, 0) != 0)
+			return;
+		if (netif_queue_stopped(dev))
+			netif_start_queue(dev);
 	}
 
 	/*
@@ -1094,54 +1103,4 @@
 	ipoib_transport_dev_cleanup(dev);
 }
 
-/*
- * Delayed P_Key Assigment Interim Support
- *
- * The following is initial implementation of delayed P_Key assigment
- * mechanism. It is using the same approach implemented for the multicast
- * group join. The single goal of this implementation is to quickly address
- * Bug #2507. This implementation will probably be removed when the P_Key
- * change async notification is available.
- */
 
-void ipoib_pkey_poll(struct work_struct *work)
-{
-	struct ipoib_dev_priv *priv =
-		container_of(work, struct ipoib_dev_priv, pkey_poll_task.work);
-	struct net_device *dev = priv->dev;
-
-	ipoib_pkey_dev_check_presence(dev);
-
-	if (test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags))
-		ipoib_open(dev);
-	else {
-		mutex_lock(&pkey_mutex);
-		if (!test_bit(IPOIB_PKEY_STOP, &priv->flags))
-			queue_delayed_work(ipoib_workqueue,
-					   &priv->pkey_poll_task,
-					   HZ);
-		mutex_unlock(&pkey_mutex);
-	}
-}
-
-int ipoib_pkey_dev_delay_open(struct net_device *dev)
-{
-	struct ipoib_dev_priv *priv = netdev_priv(dev);
-
-	/* Look for the interface pkey value in the IB Port P_Key table and */
-	/* set the interface pkey assigment flag                            */
-	ipoib_pkey_dev_check_presence(dev);
-
-	/* P_Key value not assigned yet - start polling */
-	if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags)) {
-		mutex_lock(&pkey_mutex);
-		clear_bit(IPOIB_PKEY_STOP, &priv->flags);
-		queue_delayed_work(ipoib_workqueue,
-				   &priv->pkey_poll_task,
-				   HZ);
-		mutex_unlock(&pkey_mutex);
-		return 1;
-	}
-
-	return 0;
-}
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 4e675f4..1310acf 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -108,11 +108,11 @@
 
 	set_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags);
 
-	if (ipoib_pkey_dev_delay_open(dev))
-		return 0;
-
-	if (ipoib_ib_dev_open(dev))
+	if (ipoib_ib_dev_open(dev, 1)) {
+		if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags))
+			return 0;
 		goto err_disable;
+	}
 
 	if (ipoib_ib_dev_up(dev))
 		goto err_stop;
@@ -1379,7 +1379,6 @@
 	INIT_LIST_HEAD(&priv->dead_ahs);
 	INIT_LIST_HEAD(&priv->multicast_list);
 
-	INIT_DELAYED_WORK(&priv->pkey_poll_task, ipoib_pkey_poll);
 	INIT_DELAYED_WORK(&priv->mcast_task,   ipoib_mcast_join_task);
 	INIT_WORK(&priv->carrier_on_task, ipoib_mcast_carrier_on_task);
 	INIT_WORK(&priv->flush_light,   ipoib_ib_dev_flush_light);
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c b/drivers/infiniband/ulp/iser/iscsi_iser.c
index eb79739..61ee91d 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.c
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.c
@@ -596,20 +596,28 @@
 	struct iser_conn *ib_conn;
 	struct iscsi_endpoint *ep;
 
-	ep = iscsi_create_endpoint(sizeof(*ib_conn));
+	ep = iscsi_create_endpoint(0);
 	if (!ep)
 		return ERR_PTR(-ENOMEM);
 
-	ib_conn = ep->dd_data;
+	ib_conn = kzalloc(sizeof(*ib_conn), GFP_KERNEL);
+	if (!ib_conn) {
+		err = -ENOMEM;
+		goto failure;
+	}
+
+	ep->dd_data = ib_conn;
 	ib_conn->ep = ep;
 	iser_conn_init(ib_conn);
 
-	err = iser_connect(ib_conn, NULL, (struct sockaddr_in *)dst_addr,
-			   non_blocking);
+	err = iser_connect(ib_conn, NULL, dst_addr, non_blocking);
 	if (err)
-		return ERR_PTR(err);
+		goto failure;
 
 	return ep;
+failure:
+	iscsi_destroy_endpoint(ep);
+	return ERR_PTR(err);
 }
 
 static int
@@ -619,15 +627,16 @@
 	int rc;
 
 	ib_conn = ep->dd_data;
-	rc = wait_event_interruptible_timeout(ib_conn->wait,
-			     ib_conn->state == ISER_CONN_UP,
-			     msecs_to_jiffies(timeout_ms));
-
+	rc = wait_for_completion_interruptible_timeout(&ib_conn->up_completion,
+						       msecs_to_jiffies(timeout_ms));
 	/* if conn establishment failed, return error code to iscsi */
-	if (!rc &&
-	    (ib_conn->state == ISER_CONN_TERMINATING ||
-	     ib_conn->state == ISER_CONN_DOWN))
-		rc = -1;
+	if (rc == 0) {
+		mutex_lock(&ib_conn->state_mutex);
+		if (ib_conn->state == ISER_CONN_TERMINATING ||
+		    ib_conn->state == ISER_CONN_DOWN)
+			rc = -1;
+		mutex_unlock(&ib_conn->state_mutex);
+	}
 
 	iser_info("ib conn %p rc = %d\n", ib_conn, rc);
 
@@ -646,19 +655,25 @@
 
 	ib_conn = ep->dd_data;
 	iser_info("ep %p ib conn %p state %d\n", ep, ib_conn, ib_conn->state);
+	mutex_lock(&ib_conn->state_mutex);
 	iser_conn_terminate(ib_conn);
 
 	/*
-	 * if iser_conn and iscsi_conn are bound, we must wait iscsi_conn_stop
-	 * call and ISER_CONN_DOWN state before freeing the iser resources.
-	 * otherwise we are safe to free resources immediately.
+	 * if iser_conn and iscsi_conn are bound, we must wait for
+	 * iscsi_conn_stop and flush errors completion before freeing
+	 * the iser resources. Otherwise we are safe to free resources
+	 * immediately.
 	 */
 	if (ib_conn->iscsi_conn) {
 		INIT_WORK(&ib_conn->release_work, iser_release_work);
 		queue_work(release_wq, &ib_conn->release_work);
+		mutex_unlock(&ib_conn->state_mutex);
 	} else {
+		ib_conn->state = ISER_CONN_DOWN;
+		mutex_unlock(&ib_conn->state_mutex);
 		iser_conn_release(ib_conn);
 	}
+	iscsi_destroy_endpoint(ep);
 }
 
 static umode_t iser_attr_is_visible(int param_type, int param)
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h
index 97cd385..c877dad 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.h
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.h
@@ -326,7 +326,6 @@
 	struct iser_device           *device;       /* device context          */
 	struct rdma_cm_id            *cma_id;       /* CMA ID		       */
 	struct ib_qp	             *qp;           /* QP 		       */
-	wait_queue_head_t	     wait;          /* waitq for conn/disconn  */
 	unsigned		     qp_max_recv_dtos; /* num of rx buffers */
 	unsigned		     qp_max_recv_dtos_mask; /* above minus 1 */
 	unsigned		     min_posted_rx; /* qp_max_recv_dtos >> 2 */
@@ -335,6 +334,9 @@
 	char 			     name[ISER_OBJECT_NAME_SIZE];
 	struct work_struct	     release_work;
 	struct completion	     stop_completion;
+	struct mutex		     state_mutex;
+	struct completion	     flush_completion;
+	struct completion	     up_completion;
 	struct list_head	     conn_list;       /* entry in ig conn list */
 
 	char  			     *login_buf;
@@ -448,8 +450,8 @@
 			       enum iser_data_dir cmd_dir);
 
 int  iser_connect(struct iser_conn   *ib_conn,
-		  struct sockaddr_in *src_addr,
-		  struct sockaddr_in *dst_addr,
+		  struct sockaddr    *src_addr,
+		  struct sockaddr    *dst_addr,
 		  int                non_blocking);
 
 int  iser_reg_page_vec(struct iser_conn     *ib_conn,
diff --git a/drivers/infiniband/ulp/iser/iser_verbs.c b/drivers/infiniband/ulp/iser/iser_verbs.c
index ea01075..3ef167f 100644
--- a/drivers/infiniband/ulp/iser/iser_verbs.c
+++ b/drivers/infiniband/ulp/iser/iser_verbs.c
@@ -491,10 +491,9 @@
 }
 
 /**
- * releases the QP objects, returns 0 on success,
- * -1 on failure
+ * releases the QP object
  */
-static int iser_free_ib_conn_res(struct iser_conn *ib_conn)
+static void iser_free_ib_conn_res(struct iser_conn *ib_conn)
 {
 	int cq_index;
 	BUG_ON(ib_conn == NULL);
@@ -513,8 +512,6 @@
 	}
 
 	ib_conn->qp	  = NULL;
-
-	return 0;
 }
 
 /**
@@ -568,31 +565,40 @@
 	mutex_unlock(&ig.device_list_mutex);
 }
 
+/**
+ * Called with state mutex held
+ **/
 static int iser_conn_state_comp_exch(struct iser_conn *ib_conn,
 				     enum iser_ib_conn_state comp,
 				     enum iser_ib_conn_state exch)
 {
 	int ret;
 
-	spin_lock_bh(&ib_conn->lock);
 	if ((ret = (ib_conn->state == comp)))
 		ib_conn->state = exch;
-	spin_unlock_bh(&ib_conn->lock);
 	return ret;
 }
 
 void iser_release_work(struct work_struct *work)
 {
 	struct iser_conn *ib_conn;
+	int rc;
 
 	ib_conn = container_of(work, struct iser_conn, release_work);
 
 	/* wait for .conn_stop callback */
-	wait_for_completion(&ib_conn->stop_completion);
+	rc = wait_for_completion_timeout(&ib_conn->stop_completion, 30 * HZ);
+	WARN_ON(rc == 0);
 
 	/* wait for the qp`s post send and post receive buffers to empty */
-	wait_event_interruptible(ib_conn->wait,
-				 ib_conn->state == ISER_CONN_DOWN);
+	rc = wait_for_completion_timeout(&ib_conn->flush_completion, 30 * HZ);
+	WARN_ON(rc == 0);
+
+	ib_conn->state = ISER_CONN_DOWN;
+
+	mutex_lock(&ib_conn->state_mutex);
+	ib_conn->state = ISER_CONN_DOWN;
+	mutex_unlock(&ib_conn->state_mutex);
 
 	iser_conn_release(ib_conn);
 }
@@ -604,23 +610,27 @@
 {
 	struct iser_device  *device = ib_conn->device;
 
-	BUG_ON(ib_conn->state == ISER_CONN_UP);
-
 	mutex_lock(&ig.connlist_mutex);
 	list_del(&ib_conn->conn_list);
 	mutex_unlock(&ig.connlist_mutex);
+
+	mutex_lock(&ib_conn->state_mutex);
+	BUG_ON(ib_conn->state != ISER_CONN_DOWN);
+
 	iser_free_rx_descriptors(ib_conn);
 	iser_free_ib_conn_res(ib_conn);
 	ib_conn->device = NULL;
 	/* on EVENT_ADDR_ERROR there's no device yet for this conn */
 	if (device != NULL)
 		iser_device_try_release(device);
+	mutex_unlock(&ib_conn->state_mutex);
+
 	/* if cma handler context, the caller actually destroy the id */
 	if (ib_conn->cma_id != NULL) {
 		rdma_destroy_id(ib_conn->cma_id);
 		ib_conn->cma_id = NULL;
 	}
-	iscsi_destroy_endpoint(ib_conn->ep);
+	kfree(ib_conn);
 }
 
 /**
@@ -642,22 +652,31 @@
 			 ib_conn,err);
 }
 
+/**
+ * Called with state mutex held
+ **/
 static void iser_connect_error(struct rdma_cm_id *cma_id)
 {
 	struct iser_conn *ib_conn;
 
 	ib_conn = (struct iser_conn *)cma_id->context;
-
 	ib_conn->state = ISER_CONN_DOWN;
-	wake_up_interruptible(&ib_conn->wait);
 }
 
+/**
+ * Called with state mutex held
+ **/
 static void iser_addr_handler(struct rdma_cm_id *cma_id)
 {
 	struct iser_device *device;
 	struct iser_conn   *ib_conn;
 	int    ret;
 
+	ib_conn = (struct iser_conn *)cma_id->context;
+	if (ib_conn->state != ISER_CONN_PENDING)
+		/* bailout */
+		return;
+
 	device = iser_device_find_by_ib_device(cma_id);
 	if (!device) {
 		iser_err("device lookup/creation failed\n");
@@ -665,7 +684,6 @@
 		return;
 	}
 
-	ib_conn = (struct iser_conn *)cma_id->context;
 	ib_conn->device = device;
 
 	/* connection T10-PI support */
@@ -689,18 +707,27 @@
 	}
 }
 
+/**
+ * Called with state mutex held
+ **/
 static void iser_route_handler(struct rdma_cm_id *cma_id)
 {
 	struct rdma_conn_param conn_param;
 	int    ret;
 	struct iser_cm_hdr req_hdr;
+	struct iser_conn *ib_conn = (struct iser_conn *)cma_id->context;
+	struct iser_device *device = ib_conn->device;
+
+	if (ib_conn->state != ISER_CONN_PENDING)
+		/* bailout */
+		return;
 
 	ret = iser_create_ib_conn_res((struct iser_conn *)cma_id->context);
 	if (ret)
 		goto failure;
 
 	memset(&conn_param, 0, sizeof conn_param);
-	conn_param.responder_resources = 4;
+	conn_param.responder_resources = device->dev_attr.max_qp_rd_atom;
 	conn_param.initiator_depth     = 1;
 	conn_param.retry_count	       = 7;
 	conn_param.rnr_retry_count     = 6;
@@ -728,12 +755,16 @@
 	struct ib_qp_attr attr;
 	struct ib_qp_init_attr init_attr;
 
+	ib_conn = (struct iser_conn *)cma_id->context;
+	if (ib_conn->state != ISER_CONN_PENDING)
+		/* bailout */
+		return;
+
 	(void)ib_query_qp(cma_id->qp, &attr, ~0, &init_attr);
 	iser_info("remote qpn:%x my qpn:%x\n", attr.dest_qp_num, cma_id->qp->qp_num);
 
-	ib_conn = (struct iser_conn *)cma_id->context;
-	if (iser_conn_state_comp_exch(ib_conn, ISER_CONN_PENDING, ISER_CONN_UP))
-		wake_up_interruptible(&ib_conn->wait);
+	ib_conn->state = ISER_CONN_UP;
+	complete(&ib_conn->up_completion);
 }
 
 static void iser_disconnected_handler(struct rdma_cm_id *cma_id)
@@ -752,19 +783,25 @@
 			iser_err("iscsi_iser connection isn't bound\n");
 	}
 
-	/* Complete the termination process if no posts are pending */
+	/* Complete the termination process if no posts are pending. This code
+	 * block also exists in iser_handle_comp_error(), but it is needed here
+	 * for cases of no flushes at all, e.g. discovery over rdma.
+	 */
 	if (ib_conn->post_recv_buf_count == 0 &&
 	    (atomic_read(&ib_conn->post_send_buf_count) == 0)) {
-		ib_conn->state = ISER_CONN_DOWN;
-		wake_up_interruptible(&ib_conn->wait);
+		complete(&ib_conn->flush_completion);
 	}
 }
 
 static int iser_cma_handler(struct rdma_cm_id *cma_id, struct rdma_cm_event *event)
 {
+	struct iser_conn *ib_conn;
+
+	ib_conn = (struct iser_conn *)cma_id->context;
 	iser_info("event %d status %d conn %p id %p\n",
 		  event->event, event->status, cma_id->context, cma_id);
 
+	mutex_lock(&ib_conn->state_mutex);
 	switch (event->event) {
 	case RDMA_CM_EVENT_ADDR_RESOLVED:
 		iser_addr_handler(cma_id);
@@ -785,24 +822,28 @@
 	case RDMA_CM_EVENT_DISCONNECTED:
 	case RDMA_CM_EVENT_DEVICE_REMOVAL:
 	case RDMA_CM_EVENT_ADDR_CHANGE:
+	case RDMA_CM_EVENT_TIMEWAIT_EXIT:
 		iser_disconnected_handler(cma_id);
 		break;
 	default:
 		iser_err("Unexpected RDMA CM event (%d)\n", event->event);
 		break;
 	}
+	mutex_unlock(&ib_conn->state_mutex);
 	return 0;
 }
 
 void iser_conn_init(struct iser_conn *ib_conn)
 {
 	ib_conn->state = ISER_CONN_INIT;
-	init_waitqueue_head(&ib_conn->wait);
 	ib_conn->post_recv_buf_count = 0;
 	atomic_set(&ib_conn->post_send_buf_count, 0);
 	init_completion(&ib_conn->stop_completion);
+	init_completion(&ib_conn->flush_completion);
+	init_completion(&ib_conn->up_completion);
 	INIT_LIST_HEAD(&ib_conn->conn_list);
 	spin_lock_init(&ib_conn->lock);
+	mutex_init(&ib_conn->state_mutex);
 }
 
  /**
@@ -810,22 +851,21 @@
  * sleeps until the connection is established or rejected
  */
 int iser_connect(struct iser_conn   *ib_conn,
-		 struct sockaddr_in *src_addr,
-		 struct sockaddr_in *dst_addr,
+		 struct sockaddr    *src_addr,
+		 struct sockaddr    *dst_addr,
 		 int                 non_blocking)
 {
-	struct sockaddr *src, *dst;
 	int err = 0;
 
-	sprintf(ib_conn->name, "%pI4:%d",
-		&dst_addr->sin_addr.s_addr, dst_addr->sin_port);
+	mutex_lock(&ib_conn->state_mutex);
+
+	sprintf(ib_conn->name, "%pISp", dst_addr);
+
+	iser_info("connecting to: %s\n", ib_conn->name);
 
 	/* the device is known only --after-- address resolution */
 	ib_conn->device = NULL;
 
-	iser_info("connecting to: %pI4, port 0x%x\n",
-		  &dst_addr->sin_addr, dst_addr->sin_port);
-
 	ib_conn->state = ISER_CONN_PENDING;
 
 	ib_conn->cma_id = rdma_create_id(iser_cma_handler,
@@ -837,23 +877,21 @@
 		goto id_failure;
 	}
 
-	src = (struct sockaddr *)src_addr;
-	dst = (struct sockaddr *)dst_addr;
-	err = rdma_resolve_addr(ib_conn->cma_id, src, dst, 1000);
+	err = rdma_resolve_addr(ib_conn->cma_id, src_addr, dst_addr, 1000);
 	if (err) {
 		iser_err("rdma_resolve_addr failed: %d\n", err);
 		goto addr_failure;
 	}
 
 	if (!non_blocking) {
-		wait_event_interruptible(ib_conn->wait,
-					 (ib_conn->state != ISER_CONN_PENDING));
+		wait_for_completion_interruptible(&ib_conn->up_completion);
 
 		if (ib_conn->state != ISER_CONN_UP) {
 			err =  -EIO;
 			goto connect_failure;
 		}
 	}
+	mutex_unlock(&ib_conn->state_mutex);
 
 	mutex_lock(&ig.connlist_mutex);
 	list_add(&ib_conn->conn_list, &ig.connlist);
@@ -865,6 +903,7 @@
 addr_failure:
 	ib_conn->state = ISER_CONN_DOWN;
 connect_failure:
+	mutex_unlock(&ib_conn->state_mutex);
 	iser_conn_release(ib_conn);
 	return err;
 }
@@ -1049,18 +1088,19 @@
 
 	if (ib_conn->post_recv_buf_count == 0 &&
 	    atomic_read(&ib_conn->post_send_buf_count) == 0) {
-		/* getting here when the state is UP means that the conn is *
-		 * being terminated asynchronously from the iSCSI layer's   *
-		 * perspective.                                             */
-		if (iser_conn_state_comp_exch(ib_conn, ISER_CONN_UP,
-		    ISER_CONN_TERMINATING))
+		/**
+		 * getting here when the state is UP means that the conn is
+		 * being terminated asynchronously from the iSCSI layer's
+		 * perspective. It is safe to peek at the connection state
+		 * since iscsi_conn_failure is allowed to be called twice.
+		 **/
+		if (ib_conn->state == ISER_CONN_UP)
 			iscsi_conn_failure(ib_conn->iscsi_conn,
 					   ISCSI_ERR_CONN_FAILED);
 
 		/* no more non completed posts to the QP, complete the
 		 * termination process w.o worrying on disconnect event */
-		ib_conn->state = ISER_CONN_DOWN;
-		wake_up_interruptible(&ib_conn->wait);
+		complete(&ib_conn->flush_completion);
 	}
 }
 
diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c
index e3c2c5b..62d2a18 100644
--- a/drivers/infiniband/ulp/srp/ib_srp.c
+++ b/drivers/infiniband/ulp/srp/ib_srp.c
@@ -130,6 +130,7 @@
 static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event);
 
 static struct scsi_transport_template *ib_srp_transport_template;
+static struct workqueue_struct *srp_remove_wq;
 
 static struct ib_client srp_client = {
 	.name   = "srp",
@@ -731,7 +732,7 @@
 	spin_unlock_irq(&target->lock);
 
 	if (changed)
-		queue_work(system_long_wq, &target->remove_work);
+		queue_work(srp_remove_wq, &target->remove_work);
 
 	return changed;
 }
@@ -1643,10 +1644,14 @@
 				     SCSI_SENSE_BUFFERSIZE));
 		}
 
-		if (rsp->flags & (SRP_RSP_FLAG_DOOVER | SRP_RSP_FLAG_DOUNDER))
-			scsi_set_resid(scmnd, be32_to_cpu(rsp->data_out_res_cnt));
-		else if (rsp->flags & (SRP_RSP_FLAG_DIOVER | SRP_RSP_FLAG_DIUNDER))
+		if (unlikely(rsp->flags & SRP_RSP_FLAG_DIUNDER))
 			scsi_set_resid(scmnd, be32_to_cpu(rsp->data_in_res_cnt));
+		else if (unlikely(rsp->flags & SRP_RSP_FLAG_DIOVER))
+			scsi_set_resid(scmnd, -be32_to_cpu(rsp->data_in_res_cnt));
+		else if (unlikely(rsp->flags & SRP_RSP_FLAG_DOUNDER))
+			scsi_set_resid(scmnd, be32_to_cpu(rsp->data_out_res_cnt));
+		else if (unlikely(rsp->flags & SRP_RSP_FLAG_DOOVER))
+			scsi_set_resid(scmnd, -be32_to_cpu(rsp->data_out_res_cnt));
 
 		srp_free_req(target, req, scmnd,
 			     be32_to_cpu(rsp->req_lim_delta));
@@ -3261,9 +3266,10 @@
 		spin_unlock(&host->target_lock);
 
 		/*
-		 * Wait for target port removal tasks.
+		 * Wait for tl_err and target port removal tasks.
 		 */
 		flush_workqueue(system_long_wq);
+		flush_workqueue(srp_remove_wq);
 
 		kfree(host);
 	}
@@ -3313,16 +3319,22 @@
 		indirect_sg_entries = cmd_sg_entries;
 	}
 
+	srp_remove_wq = create_workqueue("srp_remove");
+	if (!srp_remove_wq) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = -ENOMEM;
 	ib_srp_transport_template =
 		srp_attach_transport(&ib_srp_transport_functions);
 	if (!ib_srp_transport_template)
-		return -ENOMEM;
+		goto destroy_wq;
 
 	ret = class_register(&srp_class);
 	if (ret) {
 		pr_err("couldn't register class infiniband_srp\n");
-		srp_release_transport(ib_srp_transport_template);
-		return ret;
+		goto release_tr;
 	}
 
 	ib_sa_register_client(&srp_sa_client);
@@ -3330,13 +3342,22 @@
 	ret = ib_register_client(&srp_client);
 	if (ret) {
 		pr_err("couldn't register IB client\n");
-		srp_release_transport(ib_srp_transport_template);
-		ib_sa_unregister_client(&srp_sa_client);
-		class_unregister(&srp_class);
-		return ret;
+		goto unreg_sa;
 	}
 
-	return 0;
+out:
+	return ret;
+
+unreg_sa:
+	ib_sa_unregister_client(&srp_sa_client);
+	class_unregister(&srp_class);
+
+release_tr:
+	srp_release_transport(ib_srp_transport_template);
+
+destroy_wq:
+	destroy_workqueue(srp_remove_wq);
+	goto out;
 }
 
 static void __exit srp_cleanup_module(void)
@@ -3345,6 +3366,7 @@
 	ib_sa_unregister_client(&srp_sa_client);
 	class_unregister(&srp_class);
 	srp_release_transport(ib_srp_transport_template);
+	destroy_workqueue(srp_remove_wq);
 }
 
 module_init(srp_init_module);
diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index fe09f27..d28a8c2 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -198,6 +198,7 @@
 	case IB_EVENT_PKEY_CHANGE:
 	case IB_EVENT_SM_CHANGE:
 	case IB_EVENT_CLIENT_REREGISTER:
+	case IB_EVENT_GID_CHANGE:
 		/* Refresh port data asynchronously. */
 		if (event->element.port_num <= sdev->device->phys_port_cnt) {
 			sport = &sdev->port[event->element.port_num - 1];
@@ -563,7 +564,7 @@
 							 &reg_req, 0,
 							 srpt_mad_send_handler,
 							 srpt_mad_recv_handler,
-							 sport);
+							 sport, 0);
 		if (IS_ERR(sport->mad_agent)) {
 			ret = PTR_ERR(sport->mad_agent);
 			sport->mad_agent = NULL;
diff --git a/drivers/md/bcache/alloc.c b/drivers/md/bcache/alloc.c
index 443d03f..8eeab72 100644
--- a/drivers/md/bcache/alloc.c
+++ b/drivers/md/bcache/alloc.c
@@ -331,7 +331,7 @@
 				mutex_unlock(&ca->set->bucket_lock);
 				blkdev_issue_discard(ca->bdev,
 					bucket_to_sector(ca->set, bucket),
-					ca->sb.block_size, GFP_KERNEL, 0);
+					ca->sb.bucket_size, GFP_KERNEL, 0);
 				mutex_lock(&ca->set->bucket_lock);
 			}
 
diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h
index d2ebcf3..04f7bc2 100644
--- a/drivers/md/bcache/bcache.h
+++ b/drivers/md/bcache/bcache.h
@@ -477,9 +477,13 @@
  * CACHE_SET_STOPPING always gets set first when we're closing down a cache set;
  * we'll continue to run normally for awhile with CACHE_SET_STOPPING set (i.e.
  * flushing dirty data).
+ *
+ * CACHE_SET_RUNNING means all cache devices have been registered and journal
+ * replay is complete.
  */
 #define CACHE_SET_UNREGISTERING		0
 #define	CACHE_SET_STOPPING		1
+#define	CACHE_SET_RUNNING		2
 
 struct cache_set {
 	struct closure		cl;
diff --git a/drivers/md/bcache/bset.c b/drivers/md/bcache/bset.c
index 5454164..646fe85 100644
--- a/drivers/md/bcache/bset.c
+++ b/drivers/md/bcache/bset.c
@@ -1182,7 +1182,7 @@
 {
 	uint64_t start_time;
 	bool used_mempool = false;
-	struct bset *out = (void *) __get_free_pages(__GFP_NOWARN|GFP_NOIO,
+	struct bset *out = (void *) __get_free_pages(__GFP_NOWARN|GFP_NOWAIT,
 						     order);
 	if (!out) {
 		struct page *outp;
diff --git a/drivers/md/bcache/bset.h b/drivers/md/bcache/bset.h
index 5f6728d..ae96462 100644
--- a/drivers/md/bcache/bset.h
+++ b/drivers/md/bcache/bset.h
@@ -453,7 +453,7 @@
 {
 	return (KEY_DIRTY(l) == KEY_DIRTY(r) &&
 		KEY_PTRS(l) == KEY_PTRS(r) &&
-		KEY_CSUM(l) == KEY_CSUM(l));
+		KEY_CSUM(l) == KEY_CSUM(r));
 }
 
 /* Keylists */
diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index 7347b61..00cde40 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -117,9 +117,9 @@
 ({									\
 	int _r, l = (b)->level - 1;					\
 	bool _w = l <= (op)->lock;					\
-	struct btree *_child = bch_btree_node_get((b)->c, op, key, l, _w);\
+	struct btree *_child = bch_btree_node_get((b)->c, op, key, l,	\
+						  _w, b);		\
 	if (!IS_ERR(_child)) {						\
-		_child->parent = (b);					\
 		_r = bch_btree_ ## fn(_child, op, ##__VA_ARGS__);	\
 		rw_unlock(_w, _child);					\
 	} else								\
@@ -142,7 +142,6 @@
 		rw_lock(_w, _b, _b->level);				\
 		if (_b == (c)->root &&					\
 		    _w == insert_lock(op, _b)) {			\
-			_b->parent = NULL;				\
 			_r = bch_btree_ ## fn(_b, op, ##__VA_ARGS__);	\
 		}							\
 		rw_unlock(_w, _b);					\
@@ -202,7 +201,7 @@
 	struct bset *i = btree_bset_first(b);
 	struct btree_iter *iter;
 
-	iter = mempool_alloc(b->c->fill_iter, GFP_NOWAIT);
+	iter = mempool_alloc(b->c->fill_iter, GFP_NOIO);
 	iter->size = b->c->sb.bucket_size / b->c->sb.block_size;
 	iter->used = 0;
 
@@ -421,7 +420,7 @@
 	SET_PTR_OFFSET(&k.key, 0, PTR_OFFSET(&k.key, 0) +
 		       bset_sector_offset(&b->keys, i));
 
-	if (!bio_alloc_pages(b->bio, GFP_NOIO)) {
+	if (!bio_alloc_pages(b->bio, __GFP_NOWARN|GFP_NOWAIT)) {
 		int j;
 		struct bio_vec *bv;
 		void *base = (void *) ((unsigned long) i & ~(PAGE_SIZE - 1));
@@ -967,7 +966,8 @@
  * level and op->lock.
  */
 struct btree *bch_btree_node_get(struct cache_set *c, struct btree_op *op,
-				 struct bkey *k, int level, bool write)
+				 struct bkey *k, int level, bool write,
+				 struct btree *parent)
 {
 	int i = 0;
 	struct btree *b;
@@ -1002,6 +1002,7 @@
 		BUG_ON(b->level != level);
 	}
 
+	b->parent = parent;
 	b->accessed = 1;
 
 	for (; i <= b->keys.nsets && b->keys.set[i].size; i++) {
@@ -1022,15 +1023,16 @@
 	return b;
 }
 
-static void btree_node_prefetch(struct cache_set *c, struct bkey *k, int level)
+static void btree_node_prefetch(struct btree *parent, struct bkey *k)
 {
 	struct btree *b;
 
-	mutex_lock(&c->bucket_lock);
-	b = mca_alloc(c, NULL, k, level);
-	mutex_unlock(&c->bucket_lock);
+	mutex_lock(&parent->c->bucket_lock);
+	b = mca_alloc(parent->c, NULL, k, parent->level - 1);
+	mutex_unlock(&parent->c->bucket_lock);
 
 	if (!IS_ERR_OR_NULL(b)) {
+		b->parent = parent;
 		bch_btree_node_read(b);
 		rw_unlock(true, b);
 	}
@@ -1060,15 +1062,16 @@
 	mutex_unlock(&b->c->bucket_lock);
 }
 
-struct btree *bch_btree_node_alloc(struct cache_set *c, struct btree_op *op,
-				   int level)
+struct btree *__bch_btree_node_alloc(struct cache_set *c, struct btree_op *op,
+				     int level, bool wait,
+				     struct btree *parent)
 {
 	BKEY_PADDED(key) k;
 	struct btree *b = ERR_PTR(-EAGAIN);
 
 	mutex_lock(&c->bucket_lock);
 retry:
-	if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, op != NULL))
+	if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, wait))
 		goto err;
 
 	bkey_put(c, &k.key);
@@ -1085,6 +1088,7 @@
 	}
 
 	b->accessed = 1;
+	b->parent = parent;
 	bch_bset_init_next(&b->keys, b->keys.set->data, bset_magic(&b->c->sb));
 
 	mutex_unlock(&c->bucket_lock);
@@ -1096,14 +1100,21 @@
 err:
 	mutex_unlock(&c->bucket_lock);
 
-	trace_bcache_btree_node_alloc_fail(b);
+	trace_bcache_btree_node_alloc_fail(c);
 	return b;
 }
 
+static struct btree *bch_btree_node_alloc(struct cache_set *c,
+					  struct btree_op *op, int level,
+					  struct btree *parent)
+{
+	return __bch_btree_node_alloc(c, op, level, op != NULL, parent);
+}
+
 static struct btree *btree_node_alloc_replacement(struct btree *b,
 						  struct btree_op *op)
 {
-	struct btree *n = bch_btree_node_alloc(b->c, op, b->level);
+	struct btree *n = bch_btree_node_alloc(b->c, op, b->level, b->parent);
 	if (!IS_ERR_OR_NULL(n)) {
 		mutex_lock(&n->write_lock);
 		bch_btree_sort_into(&b->keys, &n->keys, &b->c->sort);
@@ -1403,6 +1414,7 @@
 	BUG_ON(btree_bset_first(new_nodes[0])->keys);
 	btree_node_free(new_nodes[0]);
 	rw_unlock(true, new_nodes[0]);
+	new_nodes[0] = NULL;
 
 	for (i = 0; i < nodes; i++) {
 		if (__bch_keylist_realloc(&keylist, bkey_u64s(&r[i].b->key)))
@@ -1516,7 +1528,7 @@
 		k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad);
 		if (k) {
 			r->b = bch_btree_node_get(b->c, op, k, b->level - 1,
-						  true);
+						  true, b);
 			if (IS_ERR(r->b)) {
 				ret = PTR_ERR(r->b);
 				break;
@@ -1811,7 +1823,7 @@
 			k = bch_btree_iter_next_filter(&iter, &b->keys,
 						       bch_ptr_bad);
 			if (k)
-				btree_node_prefetch(b->c, k, b->level - 1);
+				btree_node_prefetch(b, k);
 
 			if (p)
 				ret = btree(check_recurse, p, b, op);
@@ -1976,12 +1988,12 @@
 
 		trace_bcache_btree_node_split(b, btree_bset_first(n1)->keys);
 
-		n2 = bch_btree_node_alloc(b->c, op, b->level);
+		n2 = bch_btree_node_alloc(b->c, op, b->level, b->parent);
 		if (IS_ERR(n2))
 			goto err_free1;
 
 		if (!b->parent) {
-			n3 = bch_btree_node_alloc(b->c, op, b->level + 1);
+			n3 = bch_btree_node_alloc(b->c, op, b->level + 1, NULL);
 			if (IS_ERR(n3))
 				goto err_free2;
 		}
diff --git a/drivers/md/bcache/btree.h b/drivers/md/bcache/btree.h
index 91dfa5e..5c391fa 100644
--- a/drivers/md/bcache/btree.h
+++ b/drivers/md/bcache/btree.h
@@ -242,9 +242,10 @@
 void bch_btree_node_write(struct btree *, struct closure *);
 
 void bch_btree_set_root(struct btree *);
-struct btree *bch_btree_node_alloc(struct cache_set *, struct btree_op *, int);
+struct btree *__bch_btree_node_alloc(struct cache_set *, struct btree_op *,
+				     int, bool, struct btree *);
 struct btree *bch_btree_node_get(struct cache_set *, struct btree_op *,
-				 struct bkey *, int, bool);
+				 struct bkey *, int, bool, struct btree *);
 
 int bch_btree_insert_check_key(struct btree *, struct btree_op *,
 			       struct bkey *);
diff --git a/drivers/md/bcache/extents.c b/drivers/md/bcache/extents.c
index 3a0de4c..243de0bf 100644
--- a/drivers/md/bcache/extents.c
+++ b/drivers/md/bcache/extents.c
@@ -474,9 +474,8 @@
 	return false;
 }
 
-static bool bch_extent_invalid(struct btree_keys *bk, const struct bkey *k)
+bool __bch_extent_invalid(struct cache_set *c, const struct bkey *k)
 {
-	struct btree *b = container_of(bk, struct btree, keys);
 	char buf[80];
 
 	if (!KEY_SIZE(k))
@@ -485,16 +484,22 @@
 	if (KEY_SIZE(k) > KEY_OFFSET(k))
 		goto bad;
 
-	if (__ptr_invalid(b->c, k))
+	if (__ptr_invalid(c, k))
 		goto bad;
 
 	return false;
 bad:
 	bch_extent_to_text(buf, sizeof(buf), k);
-	cache_bug(b->c, "spotted extent %s: %s", buf, bch_ptr_status(b->c, k));
+	cache_bug(c, "spotted extent %s: %s", buf, bch_ptr_status(c, k));
 	return true;
 }
 
+static bool bch_extent_invalid(struct btree_keys *bk, const struct bkey *k)
+{
+	struct btree *b = container_of(bk, struct btree, keys);
+	return __bch_extent_invalid(b->c, k);
+}
+
 static bool bch_extent_bad_expensive(struct btree *b, const struct bkey *k,
 				     unsigned ptr)
 {
diff --git a/drivers/md/bcache/extents.h b/drivers/md/bcache/extents.h
index e4e2340..e2ed540 100644
--- a/drivers/md/bcache/extents.h
+++ b/drivers/md/bcache/extents.h
@@ -9,5 +9,6 @@
 
 void bch_extent_to_text(char *, size_t, const struct bkey *);
 bool __bch_btree_ptr_invalid(struct cache_set *, const struct bkey *);
+bool __bch_extent_invalid(struct cache_set *, const struct bkey *);
 
 #endif /* _BCACHE_EXTENTS_H */
diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c
index 59e8202..fe080ad 100644
--- a/drivers/md/bcache/journal.c
+++ b/drivers/md/bcache/journal.c
@@ -7,6 +7,7 @@
 #include "bcache.h"
 #include "btree.h"
 #include "debug.h"
+#include "extents.h"
 
 #include <trace/events/bcache.h>
 
@@ -189,11 +190,15 @@
 			if (read_bucket(l))
 				goto bsearch;
 
-		if (list_empty(list))
+		/* no journal entries on this device? */
+		if (l == ca->sb.njournal_buckets)
 			continue;
 bsearch:
+		BUG_ON(list_empty(list));
+
 		/* Binary search */
-		m = r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1);
+		m = l;
+		r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1);
 		pr_debug("starting binary search, l %u r %u", l, r);
 
 		while (l + 1 < r) {
@@ -291,15 +296,16 @@
 
 		for (k = i->j.start;
 		     k < bset_bkey_last(&i->j);
-		     k = bkey_next(k)) {
-			unsigned j;
+		     k = bkey_next(k))
+			if (!__bch_extent_invalid(c, k)) {
+				unsigned j;
 
-			for (j = 0; j < KEY_PTRS(k); j++)
-				if (ptr_available(c, k, j))
-					atomic_inc(&PTR_BUCKET(c, k, j)->pin);
+				for (j = 0; j < KEY_PTRS(k); j++)
+					if (ptr_available(c, k, j))
+						atomic_inc(&PTR_BUCKET(c, k, j)->pin);
 
-			bch_initial_mark_key(c, 0, k);
-		}
+				bch_initial_mark_key(c, 0, k);
+			}
 	}
 }
 
diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index 15fff4f..62e6e98 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -311,7 +311,8 @@
 {
 	struct data_insert_op *op = container_of(cl, struct data_insert_op, cl);
 
-	trace_bcache_write(op->bio, op->writeback, op->bypass);
+	trace_bcache_write(op->c, op->inode, op->bio,
+			   op->writeback, op->bypass);
 
 	bch_keylist_init(&op->insert_keys);
 	bio_get(op->bio);
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 926ded8..d4713d0 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -733,8 +733,6 @@
 static void bcache_device_attach(struct bcache_device *d, struct cache_set *c,
 				 unsigned id)
 {
-	BUG_ON(test_bit(CACHE_SET_STOPPING, &c->flags));
-
 	d->id = id;
 	d->c = c;
 	c->devices[id] = d;
@@ -927,6 +925,7 @@
 	list_move(&dc->list, &uncached_devices);
 
 	clear_bit(BCACHE_DEV_DETACHING, &dc->disk.flags);
+	clear_bit(BCACHE_DEV_UNLINK_DONE, &dc->disk.flags);
 
 	mutex_unlock(&bch_register_lock);
 
@@ -1041,6 +1040,9 @@
 	 */
 	atomic_set(&dc->count, 1);
 
+	if (bch_cached_dev_writeback_start(dc))
+		return -ENOMEM;
+
 	if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) {
 		bch_sectors_dirty_init(dc);
 		atomic_set(&dc->has_dirty, 1);
@@ -1070,7 +1072,8 @@
 	struct cached_dev *dc = container_of(cl, struct cached_dev, disk.cl);
 
 	cancel_delayed_work_sync(&dc->writeback_rate_update);
-	kthread_stop(dc->writeback_thread);
+	if (!IS_ERR_OR_NULL(dc->writeback_thread))
+		kthread_stop(dc->writeback_thread);
 
 	mutex_lock(&bch_register_lock);
 
@@ -1081,12 +1084,8 @@
 
 	mutex_unlock(&bch_register_lock);
 
-	if (!IS_ERR_OR_NULL(dc->bdev)) {
-		if (dc->bdev->bd_disk)
-			blk_sync_queue(bdev_get_queue(dc->bdev));
-
+	if (!IS_ERR_OR_NULL(dc->bdev))
 		blkdev_put(dc->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
-	}
 
 	wake_up(&unregister_wait);
 
@@ -1213,7 +1212,9 @@
 static void flash_dev_free(struct closure *cl)
 {
 	struct bcache_device *d = container_of(cl, struct bcache_device, cl);
+	mutex_lock(&bch_register_lock);
 	bcache_device_free(d);
+	mutex_unlock(&bch_register_lock);
 	kobject_put(&d->kobj);
 }
 
@@ -1221,7 +1222,9 @@
 {
 	struct bcache_device *d = container_of(cl, struct bcache_device, cl);
 
+	mutex_lock(&bch_register_lock);
 	bcache_device_unlink(d);
+	mutex_unlock(&bch_register_lock);
 	kobject_del(&d->kobj);
 	continue_at(cl, flash_dev_free, system_wq);
 }
@@ -1277,6 +1280,9 @@
 	if (test_bit(CACHE_SET_STOPPING, &c->flags))
 		return -EINTR;
 
+	if (!test_bit(CACHE_SET_RUNNING, &c->flags))
+		return -EPERM;
+
 	u = uuid_find_empty(c);
 	if (!u) {
 		pr_err("Can't create volume, no room for UUID");
@@ -1346,8 +1352,11 @@
 	bch_journal_free(c);
 
 	for_each_cache(ca, c, i)
-		if (ca)
+		if (ca) {
+			ca->set = NULL;
+			c->cache[ca->sb.nr_this_dev] = NULL;
 			kobject_put(&ca->kobj);
+		}
 
 	bch_bset_sort_state_free(&c->sort);
 	free_pages((unsigned long) c->uuids, ilog2(bucket_pages(c)));
@@ -1405,9 +1414,11 @@
 		if (ca->alloc_thread)
 			kthread_stop(ca->alloc_thread);
 
-	cancel_delayed_work_sync(&c->journal.work);
-	/* flush last journal entry if needed */
-	c->journal.work.work.func(&c->journal.work.work);
+	if (c->journal.cur) {
+		cancel_delayed_work_sync(&c->journal.work);
+		/* flush last journal entry if needed */
+		c->journal.work.work.func(&c->journal.work.work);
+	}
 
 	closure_return(cl);
 }
@@ -1586,7 +1597,7 @@
 			goto err;
 
 		err = "error reading btree root";
-		c->root = bch_btree_node_get(c, NULL, k, j->btree_level, true);
+		c->root = bch_btree_node_get(c, NULL, k, j->btree_level, true, NULL);
 		if (IS_ERR_OR_NULL(c->root))
 			goto err;
 
@@ -1661,7 +1672,7 @@
 			goto err;
 
 		err = "cannot allocate new btree root";
-		c->root = bch_btree_node_alloc(c, NULL, 0);
+		c->root = __bch_btree_node_alloc(c, NULL, 0, true, NULL);
 		if (IS_ERR_OR_NULL(c->root))
 			goto err;
 
@@ -1697,6 +1708,7 @@
 
 	flash_devs_run(c);
 
+	set_bit(CACHE_SET_RUNNING, &c->flags);
 	return;
 err:
 	closure_sync(&cl);
@@ -1760,6 +1772,7 @@
 		pr_debug("set version = %llu", c->sb.version);
 	}
 
+	kobject_get(&ca->kobj);
 	ca->set = c;
 	ca->set->cache[ca->sb.nr_this_dev] = ca;
 	c->cache_by_alloc[c->caches_loaded++] = ca;
@@ -1780,8 +1793,10 @@
 	struct cache *ca = container_of(kobj, struct cache, kobj);
 	unsigned i;
 
-	if (ca->set)
+	if (ca->set) {
+		BUG_ON(ca->set->cache[ca->sb.nr_this_dev] != ca);
 		ca->set->cache[ca->sb.nr_this_dev] = NULL;
+	}
 
 	bio_split_pool_free(&ca->bio_split_hook);
 
@@ -1798,10 +1813,8 @@
 	if (ca->sb_bio.bi_inline_vecs[0].bv_page)
 		put_page(ca->sb_bio.bi_io_vec[0].bv_page);
 
-	if (!IS_ERR_OR_NULL(ca->bdev)) {
-		blk_sync_queue(bdev_get_queue(ca->bdev));
+	if (!IS_ERR_OR_NULL(ca->bdev))
 		blkdev_put(ca->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
-	}
 
 	kfree(ca);
 	module_put(THIS_MODULE);
@@ -1844,7 +1857,7 @@
 }
 
 static void register_cache(struct cache_sb *sb, struct page *sb_page,
-				  struct block_device *bdev, struct cache *ca)
+				struct block_device *bdev, struct cache *ca)
 {
 	char name[BDEVNAME_SIZE];
 	const char *err = "cannot allocate memory";
@@ -1877,10 +1890,12 @@
 		goto err;
 
 	pr_info("registered cache device %s", bdevname(bdev, name));
+out:
+	kobject_put(&ca->kobj);
 	return;
 err:
 	pr_notice("error opening %s: %s", bdevname(bdev, name), err);
-	kobject_put(&ca->kobj);
+	goto out;
 }
 
 /* Global interfaces/init */
@@ -1945,10 +1960,12 @@
 	if (IS_ERR(bdev)) {
 		if (bdev == ERR_PTR(-EBUSY)) {
 			bdev = lookup_bdev(strim(path));
+			mutex_lock(&bch_register_lock);
 			if (!IS_ERR(bdev) && bch_is_open(bdev))
 				err = "device already registered";
 			else
 				err = "device busy";
+			mutex_unlock(&bch_register_lock);
 		}
 		goto err;
 	}
diff --git a/drivers/md/bcache/util.h b/drivers/md/bcache/util.h
index ac7d0d1..98df757 100644
--- a/drivers/md/bcache/util.h
+++ b/drivers/md/bcache/util.h
@@ -416,8 +416,8 @@
 			  average_frequency,	frequency_units);	\
 	__print_time_stat(stats, name,					\
 			  average_duration,	duration_units);	\
-	__print_time_stat(stats, name,					\
-			  max_duration,		duration_units);	\
+	sysfs_print(name ## _ ##max_duration ## _ ## duration_units,	\
+			div_u64((stats)->max_duration, NSEC_PER_ ## duration_units));\
 									\
 	sysfs_print(name ## _last_ ## frequency_units, (stats)->last	\
 		    ? div_s64(local_clock() - (stats)->last,		\
diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c
index f4300e4..f1986bc 100644
--- a/drivers/md/bcache/writeback.c
+++ b/drivers/md/bcache/writeback.c
@@ -239,7 +239,7 @@
 		if (KEY_START(&w->key) != dc->last_read ||
 		    jiffies_to_msecs(delay) > 50)
 			while (!kthread_should_stop() && delay)
-				delay = schedule_timeout_uninterruptible(delay);
+				delay = schedule_timeout_interruptible(delay);
 
 		dc->last_read	= KEY_OFFSET(&w->key);
 
@@ -436,7 +436,7 @@
 			while (delay &&
 			       !kthread_should_stop() &&
 			       !test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags))
-				delay = schedule_timeout_uninterruptible(delay);
+				delay = schedule_timeout_interruptible(delay);
 		}
 	}
 
@@ -478,7 +478,7 @@
 	dc->disk.sectors_dirty_last = bcache_dev_sectors_dirty(&dc->disk);
 }
 
-int bch_cached_dev_writeback_init(struct cached_dev *dc)
+void bch_cached_dev_writeback_init(struct cached_dev *dc)
 {
 	sema_init(&dc->in_flight, 64);
 	init_rwsem(&dc->writeback_lock);
@@ -494,14 +494,20 @@
 	dc->writeback_rate_d_term	= 30;
 	dc->writeback_rate_p_term_inverse = 6000;
 
+	INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate);
+}
+
+int bch_cached_dev_writeback_start(struct cached_dev *dc)
+{
 	dc->writeback_thread = kthread_create(bch_writeback_thread, dc,
 					      "bcache_writeback");
 	if (IS_ERR(dc->writeback_thread))
 		return PTR_ERR(dc->writeback_thread);
 
-	INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate);
 	schedule_delayed_work(&dc->writeback_rate_update,
 			      dc->writeback_rate_update_seconds * HZ);
 
+	bch_writeback_queue(dc);
+
 	return 0;
 }
diff --git a/drivers/md/bcache/writeback.h b/drivers/md/bcache/writeback.h
index e2f8598..0a9dab1 100644
--- a/drivers/md/bcache/writeback.h
+++ b/drivers/md/bcache/writeback.h
@@ -85,6 +85,7 @@
 void bcache_dev_sectors_dirty_add(struct cache_set *, unsigned, uint64_t, int);
 
 void bch_sectors_dirty_init(struct cached_dev *dc);
-int bch_cached_dev_writeback_init(struct cached_dev *);
+void bch_cached_dev_writeback_init(struct cached_dev *);
+int bch_cached_dev_writeback_start(struct cached_dev *);
 
 #endif
diff --git a/drivers/md/dm-cache-metadata.c b/drivers/md/dm-cache-metadata.c
index d2899e7..0670925 100644
--- a/drivers/md/dm-cache-metadata.c
+++ b/drivers/md/dm-cache-metadata.c
@@ -330,7 +330,7 @@
 	disk_super->discard_root = cpu_to_le64(cmd->discard_root);
 	disk_super->discard_block_size = cpu_to_le64(cmd->discard_block_size);
 	disk_super->discard_nr_blocks = cpu_to_le64(from_oblock(cmd->discard_nr_blocks));
-	disk_super->metadata_block_size = cpu_to_le32(DM_CACHE_METADATA_BLOCK_SIZE >> SECTOR_SHIFT);
+	disk_super->metadata_block_size = cpu_to_le32(DM_CACHE_METADATA_BLOCK_SIZE);
 	disk_super->data_block_size = cpu_to_le32(cmd->data_block_size);
 	disk_super->cache_blocks = cpu_to_le32(0);
 
@@ -478,7 +478,7 @@
 					    bool may_format_device)
 {
 	int r;
-	cmd->bm = dm_block_manager_create(cmd->bdev, DM_CACHE_METADATA_BLOCK_SIZE,
+	cmd->bm = dm_block_manager_create(cmd->bdev, DM_CACHE_METADATA_BLOCK_SIZE << SECTOR_SHIFT,
 					  CACHE_METADATA_CACHE_SIZE,
 					  CACHE_MAX_CONCURRENT_LOCKS);
 	if (IS_ERR(cmd->bm)) {
diff --git a/drivers/md/dm-cache-metadata.h b/drivers/md/dm-cache-metadata.h
index cd70a78..7383c90 100644
--- a/drivers/md/dm-cache-metadata.h
+++ b/drivers/md/dm-cache-metadata.h
@@ -9,19 +9,17 @@
 
 #include "dm-cache-block-types.h"
 #include "dm-cache-policy-internal.h"
+#include "persistent-data/dm-space-map-metadata.h"
 
 /*----------------------------------------------------------------*/
 
-#define DM_CACHE_METADATA_BLOCK_SIZE 4096
+#define DM_CACHE_METADATA_BLOCK_SIZE DM_SM_METADATA_BLOCK_SIZE
 
 /* FIXME: remove this restriction */
 /*
  * The metadata device is currently limited in size.
- *
- * We have one block of index, which can hold 255 index entries.  Each
- * index entry contains allocation info about 16k metadata blocks.
  */
-#define DM_CACHE_METADATA_MAX_SECTORS (255 * (1 << 14) * (DM_CACHE_METADATA_BLOCK_SIZE / (1 << SECTOR_SHIFT)))
+#define DM_CACHE_METADATA_MAX_SECTORS DM_SM_METADATA_MAX_SECTORS
 
 /*
  * A metadata device larger than 16GB triggers a warning.
diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c
index 2c63326..1af40ee 100644
--- a/drivers/md/dm-cache-target.c
+++ b/drivers/md/dm-cache-target.c
@@ -718,6 +718,22 @@
 	return bio->bi_rw & (REQ_FLUSH | REQ_FUA);
 }
 
+/*
+ * You must increment the deferred set whilst the prison cell is held.  To
+ * encourage this, we ask for 'cell' to be passed in.
+ */
+static void inc_ds(struct cache *cache, struct bio *bio,
+		   struct dm_bio_prison_cell *cell)
+{
+	size_t pb_data_size = get_per_bio_data_size(cache);
+	struct per_bio_data *pb = get_per_bio_data(bio, pb_data_size);
+
+	BUG_ON(!cell);
+	BUG_ON(pb->all_io_entry);
+
+	pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
+}
+
 static void issue(struct cache *cache, struct bio *bio)
 {
 	unsigned long flags;
@@ -737,6 +753,12 @@
 	spin_unlock_irqrestore(&cache->lock, flags);
 }
 
+static void inc_and_issue(struct cache *cache, struct bio *bio, struct dm_bio_prison_cell *cell)
+{
+	inc_ds(cache, bio, cell);
+	issue(cache, bio);
+}
+
 static void defer_writethrough_bio(struct cache *cache, struct bio *bio)
 {
 	unsigned long flags;
@@ -1015,6 +1037,11 @@
 
 	dm_hook_bio(&pb->hook_info, bio, overwrite_endio, mg);
 	remap_to_cache_dirty(mg->cache, bio, mg->new_oblock, mg->cblock);
+
+	/*
+	 * No need to inc_ds() here, since the cell will be held for the
+	 * duration of the io.
+	 */
 	generic_make_request(bio);
 }
 
@@ -1115,8 +1142,7 @@
 		return;
 
 	INIT_LIST_HEAD(&work);
-	if (pb->all_io_entry)
-		dm_deferred_entry_dec(pb->all_io_entry, &work);
+	dm_deferred_entry_dec(pb->all_io_entry, &work);
 
 	if (!list_empty(&work))
 		queue_quiesced_migrations(cache, &work);
@@ -1252,6 +1278,11 @@
 	else
 		remap_to_cache(cache, bio, 0);
 
+	/*
+	 * REQ_FLUSH is not directed at any particular block so we don't
+	 * need to inc_ds().  REQ_FUA's are split into a write + REQ_FLUSH
+	 * by dm-core.
+	 */
 	issue(cache, bio);
 }
 
@@ -1301,15 +1332,6 @@
 		   &cache->stats.read_miss : &cache->stats.write_miss);
 }
 
-static void issue_cache_bio(struct cache *cache, struct bio *bio,
-			    struct per_bio_data *pb,
-			    dm_oblock_t oblock, dm_cblock_t cblock)
-{
-	pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
-	remap_to_cache_dirty(cache, bio, oblock, cblock);
-	issue(cache, bio);
-}
-
 static void process_bio(struct cache *cache, struct prealloc *structs,
 			struct bio *bio)
 {
@@ -1318,8 +1340,6 @@
 	dm_oblock_t block = get_bio_block(cache, bio);
 	struct dm_bio_prison_cell *cell_prealloc, *old_ocell, *new_ocell;
 	struct policy_result lookup_result;
-	size_t pb_data_size = get_per_bio_data_size(cache);
-	struct per_bio_data *pb = get_per_bio_data(bio, pb_data_size);
 	bool discarded_block = is_discarded_oblock(cache, block);
 	bool passthrough = passthrough_mode(&cache->features);
 	bool can_migrate = !passthrough && (discarded_block || spare_migration_bandwidth(cache));
@@ -1359,9 +1379,8 @@
 
 			} else {
 				/* FIXME: factor out issue_origin() */
-				pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
 				remap_to_origin_clear_discard(cache, bio, block);
-				issue(cache, bio);
+				inc_and_issue(cache, bio, new_ocell);
 			}
 		} else {
 			inc_hit_counter(cache, bio);
@@ -1369,20 +1388,21 @@
 			if (bio_data_dir(bio) == WRITE &&
 			    writethrough_mode(&cache->features) &&
 			    !is_dirty(cache, lookup_result.cblock)) {
-				pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
 				remap_to_origin_then_cache(cache, bio, block, lookup_result.cblock);
-				issue(cache, bio);
-			} else
-				issue_cache_bio(cache, bio, pb, block, lookup_result.cblock);
+				inc_and_issue(cache, bio, new_ocell);
+
+			} else  {
+				remap_to_cache_dirty(cache, bio, block, lookup_result.cblock);
+				inc_and_issue(cache, bio, new_ocell);
+			}
 		}
 
 		break;
 
 	case POLICY_MISS:
 		inc_miss_counter(cache, bio);
-		pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
 		remap_to_origin_clear_discard(cache, bio, block);
-		issue(cache, bio);
+		inc_and_issue(cache, bio, new_ocell);
 		break;
 
 	case POLICY_NEW:
@@ -1501,6 +1521,9 @@
 	bio_list_init(&cache->deferred_flush_bios);
 	spin_unlock_irqrestore(&cache->lock, flags);
 
+	/*
+	 * These bios have already been through inc_ds()
+	 */
 	while ((bio = bio_list_pop(&bios)))
 		submit_bios ? generic_make_request(bio) : bio_io_error(bio);
 }
@@ -1518,6 +1541,9 @@
 	bio_list_init(&cache->deferred_writethrough_bios);
 	spin_unlock_irqrestore(&cache->lock, flags);
 
+	/*
+	 * These bios have already been through inc_ds()
+	 */
 	while ((bio = bio_list_pop(&bios)))
 		generic_make_request(bio);
 }
@@ -1694,6 +1720,7 @@
 
 		if (commit_if_needed(cache)) {
 			process_deferred_flush_bios(cache, false);
+			process_migrations(cache, &cache->need_commit_migrations, migration_failure);
 
 			/*
 			 * FIXME: rollback metadata or just go into a
@@ -2406,16 +2433,13 @@
 	return r;
 }
 
-static int cache_map(struct dm_target *ti, struct bio *bio)
+static int __cache_map(struct cache *cache, struct bio *bio, struct dm_bio_prison_cell **cell)
 {
-	struct cache *cache = ti->private;
-
 	int r;
 	dm_oblock_t block = get_bio_block(cache, bio);
 	size_t pb_data_size = get_per_bio_data_size(cache);
 	bool can_migrate = false;
 	bool discarded_block;
-	struct dm_bio_prison_cell *cell;
 	struct policy_result lookup_result;
 	struct per_bio_data *pb = init_per_bio_data(bio, pb_data_size);
 
@@ -2437,15 +2461,15 @@
 	/*
 	 * Check to see if that block is currently migrating.
 	 */
-	cell = alloc_prison_cell(cache);
-	if (!cell) {
+	*cell = alloc_prison_cell(cache);
+	if (!*cell) {
 		defer_bio(cache, bio);
 		return DM_MAPIO_SUBMITTED;
 	}
 
-	r = bio_detain(cache, block, bio, cell,
+	r = bio_detain(cache, block, bio, *cell,
 		       (cell_free_fn) free_prison_cell,
-		       cache, &cell);
+		       cache, cell);
 	if (r) {
 		if (r < 0)
 			defer_bio(cache, bio);
@@ -2458,11 +2482,12 @@
 	r = policy_map(cache->policy, block, false, can_migrate, discarded_block,
 		       bio, &lookup_result);
 	if (r == -EWOULDBLOCK) {
-		cell_defer(cache, cell, true);
+		cell_defer(cache, *cell, true);
 		return DM_MAPIO_SUBMITTED;
 
 	} else if (r) {
 		DMERR_LIMIT("Unexpected return from cache replacement policy: %d", r);
+		cell_defer(cache, *cell, false);
 		bio_io_error(bio);
 		return DM_MAPIO_SUBMITTED;
 	}
@@ -2476,52 +2501,44 @@
 				 * We need to invalidate this block, so
 				 * defer for the worker thread.
 				 */
-				cell_defer(cache, cell, true);
+				cell_defer(cache, *cell, true);
 				r = DM_MAPIO_SUBMITTED;
 
 			} else {
-				pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
 				inc_miss_counter(cache, bio);
 				remap_to_origin_clear_discard(cache, bio, block);
-
-				cell_defer(cache, cell, false);
 			}
 
 		} else {
 			inc_hit_counter(cache, bio);
-			pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
-
 			if (bio_data_dir(bio) == WRITE && writethrough_mode(&cache->features) &&
 			    !is_dirty(cache, lookup_result.cblock))
 				remap_to_origin_then_cache(cache, bio, block, lookup_result.cblock);
 			else
 				remap_to_cache_dirty(cache, bio, block, lookup_result.cblock);
-
-			cell_defer(cache, cell, false);
 		}
 		break;
 
 	case POLICY_MISS:
 		inc_miss_counter(cache, bio);
-		pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
-
 		if (pb->req_nr != 0) {
 			/*
 			 * This is a duplicate writethrough io that is no
 			 * longer needed because the block has been demoted.
 			 */
 			bio_endio(bio, 0);
-			cell_defer(cache, cell, false);
-			return DM_MAPIO_SUBMITTED;
-		} else {
+			cell_defer(cache, *cell, false);
+			r = DM_MAPIO_SUBMITTED;
+
+		} else
 			remap_to_origin_clear_discard(cache, bio, block);
-			cell_defer(cache, cell, false);
-		}
+
 		break;
 
 	default:
 		DMERR_LIMIT("%s: erroring bio: unknown policy op: %u", __func__,
 			    (unsigned) lookup_result.op);
+		cell_defer(cache, *cell, false);
 		bio_io_error(bio);
 		r = DM_MAPIO_SUBMITTED;
 	}
@@ -2529,6 +2546,21 @@
 	return r;
 }
 
+static int cache_map(struct dm_target *ti, struct bio *bio)
+{
+	int r;
+	struct dm_bio_prison_cell *cell;
+	struct cache *cache = ti->private;
+
+	r = __cache_map(cache, bio, &cell);
+	if (r == DM_MAPIO_REMAPPED) {
+		inc_ds(cache, bio, cell);
+		cell_defer(cache, cell, false);
+	}
+
+	return r;
+}
+
 static int cache_end_io(struct dm_target *ti, struct bio *bio, int error)
 {
 	struct cache *cache = ti->private;
@@ -2808,7 +2840,7 @@
 		residency = policy_residency(cache->policy);
 
 		DMEMIT("%u %llu/%llu %u %llu/%llu %u %u %u %u %u %u %lu ",
-		       (unsigned)(DM_CACHE_METADATA_BLOCK_SIZE >> SECTOR_SHIFT),
+		       (unsigned)DM_CACHE_METADATA_BLOCK_SIZE,
 		       (unsigned long long)(nr_blocks_metadata - nr_free_blocks_metadata),
 		       (unsigned long long)nr_blocks_metadata,
 		       cache->sectors_per_block,
@@ -3062,7 +3094,7 @@
 	 */
 	if (io_opt_sectors < cache->sectors_per_block ||
 	    do_div(io_opt_sectors, cache->sectors_per_block)) {
-		blk_limits_io_min(limits, 0);
+		blk_limits_io_min(limits, cache->sectors_per_block << SECTOR_SHIFT);
 		blk_limits_io_opt(limits, cache->sectors_per_block << SECTOR_SHIFT);
 	}
 	set_discard_limits(cache, limits);
@@ -3072,7 +3104,7 @@
 
 static struct target_type cache_target = {
 	.name = "cache",
-	.version = {1, 4, 0},
+	.version = {1, 5, 0},
 	.module = THIS_MODULE,
 	.ctr = cache_ctr,
 	.dtr = cache_dtr,
diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c
index 4cba2d8..2785007 100644
--- a/drivers/md/dm-crypt.c
+++ b/drivers/md/dm-crypt.c
@@ -59,7 +59,7 @@
 	int error;
 	sector_t sector;
 	struct dm_crypt_io *base_io;
-};
+} CRYPTO_MINALIGN_ATTR;
 
 struct dm_crypt_request {
 	struct convert_context *ctx;
@@ -162,6 +162,8 @@
 	 */
 	unsigned int dmreq_start;
 
+	unsigned int per_bio_data_size;
+
 	unsigned long flags;
 	unsigned int key_size;
 	unsigned int key_parts;      /* independent parts in key buffer */
@@ -895,6 +897,15 @@
 	    kcryptd_async_done, dmreq_of_req(cc, ctx->req));
 }
 
+static void crypt_free_req(struct crypt_config *cc,
+			   struct ablkcipher_request *req, struct bio *base_bio)
+{
+	struct dm_crypt_io *io = dm_per_bio_data(base_bio, cc->per_bio_data_size);
+
+	if ((struct ablkcipher_request *)(io + 1) != req)
+		mempool_free(req, cc->req_pool);
+}
+
 /*
  * Encrypt / decrypt data from one bio to another one (can be the same one)
  */
@@ -1008,12 +1019,9 @@
 	}
 }
 
-static struct dm_crypt_io *crypt_io_alloc(struct crypt_config *cc,
-					  struct bio *bio, sector_t sector)
+static void crypt_io_init(struct dm_crypt_io *io, struct crypt_config *cc,
+			  struct bio *bio, sector_t sector)
 {
-	struct dm_crypt_io *io;
-
-	io = mempool_alloc(cc->io_pool, GFP_NOIO);
 	io->cc = cc;
 	io->base_bio = bio;
 	io->sector = sector;
@@ -1021,8 +1029,6 @@
 	io->base_io = NULL;
 	io->ctx.req = NULL;
 	atomic_set(&io->io_pending, 0);
-
-	return io;
 }
 
 static void crypt_inc_pending(struct dm_crypt_io *io)
@@ -1046,8 +1052,9 @@
 		return;
 
 	if (io->ctx.req)
-		mempool_free(io->ctx.req, cc->req_pool);
-	mempool_free(io, cc->io_pool);
+		crypt_free_req(cc, io->ctx.req, base_bio);
+	if (io != dm_per_bio_data(base_bio, cc->per_bio_data_size))
+		mempool_free(io, cc->io_pool);
 
 	if (likely(!base_io))
 		bio_endio(base_bio, error);
@@ -1255,8 +1262,8 @@
 		 * between fragments, so switch to a new dm_crypt_io structure.
 		 */
 		if (unlikely(!crypt_finished && remaining)) {
-			new_io = crypt_io_alloc(io->cc, io->base_bio,
-						sector);
+			new_io = mempool_alloc(cc->io_pool, GFP_NOIO);
+			crypt_io_init(new_io, io->cc, io->base_bio, sector);
 			crypt_inc_pending(new_io);
 			crypt_convert_init(cc, &new_io->ctx, NULL,
 					   io->base_bio, sector);
@@ -1325,7 +1332,7 @@
 	if (error < 0)
 		io->error = -EIO;
 
-	mempool_free(req_of_dmreq(cc, dmreq), cc->req_pool);
+	crypt_free_req(cc, req_of_dmreq(cc, dmreq), io->base_bio);
 
 	if (!atomic_dec_and_test(&ctx->cc_pending))
 		return;
@@ -1728,6 +1735,10 @@
 		goto bad;
 	}
 
+	cc->per_bio_data_size = ti->per_bio_data_size =
+				sizeof(struct dm_crypt_io) + cc->dmreq_start +
+				sizeof(struct dm_crypt_request) + cc->iv_size;
+
 	cc->page_pool = mempool_create_page_pool(MIN_POOL_PAGES, 0);
 	if (!cc->page_pool) {
 		ti->error = "Cannot allocate page mempool";
@@ -1824,7 +1835,9 @@
 		return DM_MAPIO_REMAPPED;
 	}
 
-	io = crypt_io_alloc(cc, bio, dm_target_offset(ti, bio->bi_iter.bi_sector));
+	io = dm_per_bio_data(bio, cc->per_bio_data_size);
+	crypt_io_init(io, cc, bio, dm_target_offset(ti, bio->bi_iter.bi_sector));
+	io->ctx.req = (struct ablkcipher_request *)(io + 1);
 
 	if (bio_data_dir(io->base_bio) == READ) {
 		if (kcryptd_io_read(io, GFP_NOWAIT))
diff --git a/drivers/md/dm-io.c b/drivers/md/dm-io.c
index db404a0..c09359d 100644
--- a/drivers/md/dm-io.c
+++ b/drivers/md/dm-io.c
@@ -33,7 +33,6 @@
 struct io {
 	unsigned long error_bits;
 	atomic_t count;
-	struct completion *wait;
 	struct dm_io_client *client;
 	io_notify_fn callback;
 	void *context;
@@ -112,28 +111,27 @@
  * We need an io object to keep track of the number of bios that
  * have been dispatched for a particular io.
  *---------------------------------------------------------------*/
+static void complete_io(struct io *io)
+{
+	unsigned long error_bits = io->error_bits;
+	io_notify_fn fn = io->callback;
+	void *context = io->context;
+
+	if (io->vma_invalidate_size)
+		invalidate_kernel_vmap_range(io->vma_invalidate_address,
+					     io->vma_invalidate_size);
+
+	mempool_free(io, io->client->pool);
+	fn(error_bits, context);
+}
+
 static void dec_count(struct io *io, unsigned int region, int error)
 {
 	if (error)
 		set_bit(region, &io->error_bits);
 
-	if (atomic_dec_and_test(&io->count)) {
-		if (io->vma_invalidate_size)
-			invalidate_kernel_vmap_range(io->vma_invalidate_address,
-						     io->vma_invalidate_size);
-
-		if (io->wait)
-			complete(io->wait);
-
-		else {
-			unsigned long r = io->error_bits;
-			io_notify_fn fn = io->callback;
-			void *context = io->context;
-
-			mempool_free(io, io->client->pool);
-			fn(r, context);
-		}
-	}
+	if (atomic_dec_and_test(&io->count))
+		complete_io(io);
 }
 
 static void endio(struct bio *bio, int error)
@@ -376,41 +374,51 @@
 	dec_count(io, 0, 0);
 }
 
+struct sync_io {
+	unsigned long error_bits;
+	struct completion wait;
+};
+
+static void sync_io_complete(unsigned long error, void *context)
+{
+	struct sync_io *sio = context;
+
+	sio->error_bits = error;
+	complete(&sio->wait);
+}
+
 static int sync_io(struct dm_io_client *client, unsigned int num_regions,
 		   struct dm_io_region *where, int rw, struct dpages *dp,
 		   unsigned long *error_bits)
 {
-	/*
-	 * gcc <= 4.3 can't do the alignment for stack variables, so we must
-	 * align it on our own.
-	 * volatile prevents the optimizer from removing or reusing
-	 * "io_" field from the stack frame (allowed in ANSI C).
-	 */
-	volatile char io_[sizeof(struct io) + __alignof__(struct io) - 1];
-	struct io *io = (struct io *)PTR_ALIGN(&io_, __alignof__(struct io));
-	DECLARE_COMPLETION_ONSTACK(wait);
+	struct io *io;
+	struct sync_io sio;
 
 	if (num_regions > 1 && (rw & RW_MASK) != WRITE) {
 		WARN_ON(1);
 		return -EIO;
 	}
 
+	init_completion(&sio.wait);
+
+	io = mempool_alloc(client->pool, GFP_NOIO);
 	io->error_bits = 0;
 	atomic_set(&io->count, 1); /* see dispatch_io() */
-	io->wait = &wait;
 	io->client = client;
+	io->callback = sync_io_complete;
+	io->context = &sio;
 
 	io->vma_invalidate_address = dp->vma_invalidate_address;
 	io->vma_invalidate_size = dp->vma_invalidate_size;
 
 	dispatch_io(rw, num_regions, where, dp, io, 1);
 
-	wait_for_completion_io(&wait);
+	wait_for_completion_io(&sio.wait);
 
 	if (error_bits)
-		*error_bits = io->error_bits;
+		*error_bits = sio.error_bits;
 
-	return io->error_bits ? -EIO : 0;
+	return sio.error_bits ? -EIO : 0;
 }
 
 static int async_io(struct dm_io_client *client, unsigned int num_regions,
@@ -428,7 +436,6 @@
 	io = mempool_alloc(client->pool, GFP_NOIO);
 	io->error_bits = 0;
 	atomic_set(&io->count, 1); /* see dispatch_io() */
-	io->wait = NULL;
 	io->client = client;
 	io->callback = fn;
 	io->context = context;
@@ -481,9 +488,9 @@
  * New collapsed (a)synchronous interface.
  *
  * If the IO is asynchronous (i.e. it has notify.fn), you must either unplug
- * the queue with blk_unplug() some time later or set REQ_SYNC in
-io_req->bi_rw. If you fail to do one of these, the IO will be submitted to
- * the disk after q->unplug_delay, which defaults to 3ms in blk-settings.c.
+ * the queue with blk_unplug() some time later or set REQ_SYNC in io_req->bi_rw.
+ * If you fail to do one of these, the IO will be submitted to the disk after
+ * q->unplug_delay, which defaults to 3ms in blk-settings.c.
  */
 int dm_io(struct dm_io_request *io_req, unsigned num_regions,
 	  struct dm_io_region *where, unsigned long *sync_error_bits)
diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
index f4167b0..833d7e7 100644
--- a/drivers/md/dm-mpath.c
+++ b/drivers/md/dm-mpath.c
@@ -373,8 +373,6 @@
 		 dm_noflush_suspending(m->ti)));
 }
 
-#define pg_ready(m) (!(m)->queue_io && !(m)->pg_init_required)
-
 /*
  * Map cloned requests
  */
@@ -402,11 +400,11 @@
 		if (!__must_push_back(m))
 			r = -EIO;	/* Failed */
 		goto out_unlock;
-	}
-	if (!pg_ready(m)) {
+	} else if (m->queue_io || m->pg_init_required) {
 		__pg_init_all_paths(m);
 		goto out_unlock;
 	}
+
 	if (set_mapinfo(m, map_context) < 0)
 		/* ENOMEM, requeue */
 		goto out_unlock;
diff --git a/drivers/md/dm-switch.c b/drivers/md/dm-switch.c
index 09a688b..50fca46 100644
--- a/drivers/md/dm-switch.c
+++ b/drivers/md/dm-switch.c
@@ -137,13 +137,23 @@
 	*bit *= sctx->region_table_entry_bits;
 }
 
+static unsigned switch_region_table_read(struct switch_ctx *sctx, unsigned long region_nr)
+{
+	unsigned long region_index;
+	unsigned bit;
+
+	switch_get_position(sctx, region_nr, &region_index, &bit);
+
+	return (ACCESS_ONCE(sctx->region_table[region_index]) >> bit) &
+		((1 << sctx->region_table_entry_bits) - 1);
+}
+
 /*
  * Find which path to use at given offset.
  */
 static unsigned switch_get_path_nr(struct switch_ctx *sctx, sector_t offset)
 {
-	unsigned long region_index;
-	unsigned bit, path_nr;
+	unsigned path_nr;
 	sector_t p;
 
 	p = offset;
@@ -152,9 +162,7 @@
 	else
 		sector_div(p, sctx->region_size);
 
-	switch_get_position(sctx, p, &region_index, &bit);
-	path_nr = (ACCESS_ONCE(sctx->region_table[region_index]) >> bit) &
-	       ((1 << sctx->region_table_entry_bits) - 1);
+	path_nr = switch_region_table_read(sctx, p);
 
 	/* This can only happen if the processor uses non-atomic stores. */
 	if (unlikely(path_nr >= sctx->nr_paths))
@@ -363,7 +371,7 @@
 }
 
 static int process_set_region_mappings(struct switch_ctx *sctx,
-			     unsigned argc, char **argv)
+				       unsigned argc, char **argv)
 {
 	unsigned i;
 	unsigned long region_index = 0;
@@ -372,6 +380,51 @@
 		unsigned long path_nr;
 		const char *string = argv[i];
 
+		if ((*string & 0xdf) == 'R') {
+			unsigned long cycle_length, num_write;
+
+			string++;
+			if (unlikely(*string == ',')) {
+				DMWARN("invalid set_region_mappings argument: '%s'", argv[i]);
+				return -EINVAL;
+			}
+			cycle_length = parse_hex(&string);
+			if (unlikely(*string != ',')) {
+				DMWARN("invalid set_region_mappings argument: '%s'", argv[i]);
+				return -EINVAL;
+			}
+			string++;
+			if (unlikely(!*string)) {
+				DMWARN("invalid set_region_mappings argument: '%s'", argv[i]);
+				return -EINVAL;
+			}
+			num_write = parse_hex(&string);
+			if (unlikely(*string)) {
+				DMWARN("invalid set_region_mappings argument: '%s'", argv[i]);
+				return -EINVAL;
+			}
+
+			if (unlikely(!cycle_length) || unlikely(cycle_length - 1 > region_index)) {
+				DMWARN("invalid set_region_mappings cycle length: %lu > %lu",
+				       cycle_length - 1, region_index);
+				return -EINVAL;
+			}
+			if (unlikely(region_index + num_write < region_index) ||
+			    unlikely(region_index + num_write >= sctx->nr_regions)) {
+				DMWARN("invalid set_region_mappings region number: %lu + %lu >= %lu",
+				       region_index, num_write, sctx->nr_regions);
+				return -EINVAL;
+			}
+
+			while (num_write--) {
+				region_index++;
+				path_nr = switch_region_table_read(sctx, region_index - cycle_length);
+				switch_region_table_write(sctx, region_index, path_nr);
+			}
+
+			continue;
+		}
+
 		if (*string == ':')
 			region_index++;
 		else {
@@ -500,7 +553,7 @@
 
 static struct target_type switch_target = {
 	.name = "switch",
-	.version = {1, 0, 0},
+	.version = {1, 1, 0},
 	.module = THIS_MODULE,
 	.ctr = switch_ctr,
 	.dtr = switch_dtr,
diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index 5f59f1e..f9c6cb8 100644
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -1386,6 +1386,14 @@
 	return q && !blk_queue_add_random(q);
 }
 
+static int queue_supports_sg_merge(struct dm_target *ti, struct dm_dev *dev,
+				   sector_t start, sector_t len, void *data)
+{
+	struct request_queue *q = bdev_get_queue(dev->bdev);
+
+	return q && !test_bit(QUEUE_FLAG_NO_SG_MERGE, &q->queue_flags);
+}
+
 static bool dm_table_all_devices_attribute(struct dm_table *t,
 					   iterate_devices_callout_fn func)
 {
@@ -1430,6 +1438,43 @@
 	return true;
 }
 
+static int device_discard_capable(struct dm_target *ti, struct dm_dev *dev,
+				  sector_t start, sector_t len, void *data)
+{
+	struct request_queue *q = bdev_get_queue(dev->bdev);
+
+	return q && blk_queue_discard(q);
+}
+
+static bool dm_table_supports_discards(struct dm_table *t)
+{
+	struct dm_target *ti;
+	unsigned i = 0;
+
+	/*
+	 * Unless any target used by the table set discards_supported,
+	 * require at least one underlying device to support discards.
+	 * t->devices includes internal dm devices such as mirror logs
+	 * so we need to use iterate_devices here, which targets
+	 * supporting discard selectively must provide.
+	 */
+	while (i < dm_table_get_num_targets(t)) {
+		ti = dm_table_get_target(t, i++);
+
+		if (!ti->num_discard_bios)
+			continue;
+
+		if (ti->discards_supported)
+			return 1;
+
+		if (ti->type->iterate_devices &&
+		    ti->type->iterate_devices(ti, device_discard_capable, NULL))
+			return 1;
+	}
+
+	return 0;
+}
+
 void dm_table_set_restrictions(struct dm_table *t, struct request_queue *q,
 			       struct queue_limits *limits)
 {
@@ -1464,6 +1509,11 @@
 	if (!dm_table_supports_write_same(t))
 		q->limits.max_write_same_sectors = 0;
 
+	if (dm_table_all_devices_attribute(t, queue_supports_sg_merge))
+		queue_flag_clear_unlocked(QUEUE_FLAG_NO_SG_MERGE, q);
+	else
+		queue_flag_set_unlocked(QUEUE_FLAG_NO_SG_MERGE, q);
+
 	dm_table_set_integrity(t);
 
 	/*
@@ -1636,39 +1686,3 @@
 }
 EXPORT_SYMBOL(dm_table_run_md_queue_async);
 
-static int device_discard_capable(struct dm_target *ti, struct dm_dev *dev,
-				  sector_t start, sector_t len, void *data)
-{
-	struct request_queue *q = bdev_get_queue(dev->bdev);
-
-	return q && blk_queue_discard(q);
-}
-
-bool dm_table_supports_discards(struct dm_table *t)
-{
-	struct dm_target *ti;
-	unsigned i = 0;
-
-	/*
-	 * Unless any target used by the table set discards_supported,
-	 * require at least one underlying device to support discards.
-	 * t->devices includes internal dm devices such as mirror logs
-	 * so we need to use iterate_devices here, which targets
-	 * supporting discard selectively must provide.
-	 */
-	while (i < dm_table_get_num_targets(t)) {
-		ti = dm_table_get_target(t, i++);
-
-		if (!ti->num_discard_bios)
-			continue;
-
-		if (ti->discards_supported)
-			return 1;
-
-		if (ti->type->iterate_devices &&
-		    ti->type->iterate_devices(ti, device_discard_capable, NULL))
-			return 1;
-	}
-
-	return 0;
-}
diff --git a/drivers/md/dm-thin.c b/drivers/md/dm-thin.c
index fc9c848..4843801 100644
--- a/drivers/md/dm-thin.c
+++ b/drivers/md/dm-thin.c
@@ -227,6 +227,7 @@
 	struct list_head list;
 	struct dm_dev *pool_dev;
 	struct dm_dev *origin_dev;
+	sector_t origin_size;
 	dm_thin_id dev_id;
 
 	struct pool *pool;
@@ -554,11 +555,16 @@
 struct dm_thin_new_mapping {
 	struct list_head list;
 
-	bool quiesced:1;
-	bool prepared:1;
 	bool pass_discard:1;
 	bool definitely_not_shared:1;
 
+	/*
+	 * Track quiescing, copying and zeroing preparation actions.  When this
+	 * counter hits zero the block is prepared and can be inserted into the
+	 * btree.
+	 */
+	atomic_t prepare_actions;
+
 	int err;
 	struct thin_c *tc;
 	dm_block_t virt_block;
@@ -575,43 +581,41 @@
 	bio_end_io_t *saved_bi_end_io;
 };
 
-static void __maybe_add_mapping(struct dm_thin_new_mapping *m)
+static void __complete_mapping_preparation(struct dm_thin_new_mapping *m)
 {
 	struct pool *pool = m->tc->pool;
 
-	if (m->quiesced && m->prepared) {
+	if (atomic_dec_and_test(&m->prepare_actions)) {
 		list_add_tail(&m->list, &pool->prepared_mappings);
 		wake_worker(pool);
 	}
 }
 
-static void copy_complete(int read_err, unsigned long write_err, void *context)
+static void complete_mapping_preparation(struct dm_thin_new_mapping *m)
 {
 	unsigned long flags;
-	struct dm_thin_new_mapping *m = context;
 	struct pool *pool = m->tc->pool;
 
-	m->err = read_err || write_err ? -EIO : 0;
-
 	spin_lock_irqsave(&pool->lock, flags);
-	m->prepared = true;
-	__maybe_add_mapping(m);
+	__complete_mapping_preparation(m);
 	spin_unlock_irqrestore(&pool->lock, flags);
 }
 
+static void copy_complete(int read_err, unsigned long write_err, void *context)
+{
+	struct dm_thin_new_mapping *m = context;
+
+	m->err = read_err || write_err ? -EIO : 0;
+	complete_mapping_preparation(m);
+}
+
 static void overwrite_endio(struct bio *bio, int err)
 {
-	unsigned long flags;
 	struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
 	struct dm_thin_new_mapping *m = h->overwrite_mapping;
-	struct pool *pool = m->tc->pool;
 
 	m->err = err;
-
-	spin_lock_irqsave(&pool->lock, flags);
-	m->prepared = true;
-	__maybe_add_mapping(m);
-	spin_unlock_irqrestore(&pool->lock, flags);
+	complete_mapping_preparation(m);
 }
 
 /*----------------------------------------------------------------*/
@@ -821,10 +825,31 @@
 	return m;
 }
 
+static void ll_zero(struct thin_c *tc, struct dm_thin_new_mapping *m,
+		    sector_t begin, sector_t end)
+{
+	int r;
+	struct dm_io_region to;
+
+	to.bdev = tc->pool_dev->bdev;
+	to.sector = begin;
+	to.count = end - begin;
+
+	r = dm_kcopyd_zero(tc->pool->copier, 1, &to, 0, copy_complete, m);
+	if (r < 0) {
+		DMERR_LIMIT("dm_kcopyd_zero() failed");
+		copy_complete(1, 1, m);
+	}
+}
+
+/*
+ * A partial copy also needs to zero the uncopied region.
+ */
 static void schedule_copy(struct thin_c *tc, dm_block_t virt_block,
 			  struct dm_dev *origin, dm_block_t data_origin,
 			  dm_block_t data_dest,
-			  struct dm_bio_prison_cell *cell, struct bio *bio)
+			  struct dm_bio_prison_cell *cell, struct bio *bio,
+			  sector_t len)
 {
 	int r;
 	struct pool *pool = tc->pool;
@@ -835,8 +860,15 @@
 	m->data_block = data_dest;
 	m->cell = cell;
 
+	/*
+	 * quiesce action + copy action + an extra reference held for the
+	 * duration of this function (we may need to inc later for a
+	 * partial zero).
+	 */
+	atomic_set(&m->prepare_actions, 3);
+
 	if (!dm_deferred_set_add_work(pool->shared_read_ds, &m->list))
-		m->quiesced = true;
+		complete_mapping_preparation(m); /* already quiesced */
 
 	/*
 	 * IO to pool_dev remaps to the pool target's data_dev.
@@ -857,20 +889,38 @@
 
 		from.bdev = origin->bdev;
 		from.sector = data_origin * pool->sectors_per_block;
-		from.count = pool->sectors_per_block;
+		from.count = len;
 
 		to.bdev = tc->pool_dev->bdev;
 		to.sector = data_dest * pool->sectors_per_block;
-		to.count = pool->sectors_per_block;
+		to.count = len;
 
 		r = dm_kcopyd_copy(pool->copier, &from, 1, &to,
 				   0, copy_complete, m);
 		if (r < 0) {
-			mempool_free(m, pool->mapping_pool);
 			DMERR_LIMIT("dm_kcopyd_copy() failed");
-			cell_error(pool, cell);
+			copy_complete(1, 1, m);
+
+			/*
+			 * We allow the zero to be issued, to simplify the
+			 * error path.  Otherwise we'd need to start
+			 * worrying about decrementing the prepare_actions
+			 * counter.
+			 */
+		}
+
+		/*
+		 * Do we need to zero a tail region?
+		 */
+		if (len < pool->sectors_per_block && pool->pf.zero_new_blocks) {
+			atomic_inc(&m->prepare_actions);
+			ll_zero(tc, m,
+				data_dest * pool->sectors_per_block + len,
+				(data_dest + 1) * pool->sectors_per_block);
 		}
 	}
+
+	complete_mapping_preparation(m); /* drop our ref */
 }
 
 static void schedule_internal_copy(struct thin_c *tc, dm_block_t virt_block,
@@ -878,15 +928,8 @@
 				   struct dm_bio_prison_cell *cell, struct bio *bio)
 {
 	schedule_copy(tc, virt_block, tc->pool_dev,
-		      data_origin, data_dest, cell, bio);
-}
-
-static void schedule_external_copy(struct thin_c *tc, dm_block_t virt_block,
-				   dm_block_t data_dest,
-				   struct dm_bio_prison_cell *cell, struct bio *bio)
-{
-	schedule_copy(tc, virt_block, tc->origin_dev,
-		      virt_block, data_dest, cell, bio);
+		      data_origin, data_dest, cell, bio,
+		      tc->pool->sectors_per_block);
 }
 
 static void schedule_zero(struct thin_c *tc, dm_block_t virt_block,
@@ -896,8 +939,7 @@
 	struct pool *pool = tc->pool;
 	struct dm_thin_new_mapping *m = get_next_mapping(pool);
 
-	m->quiesced = true;
-	m->prepared = false;
+	atomic_set(&m->prepare_actions, 1); /* no need to quiesce */
 	m->tc = tc;
 	m->virt_block = virt_block;
 	m->data_block = data_block;
@@ -919,21 +961,33 @@
 		save_and_set_endio(bio, &m->saved_bi_end_io, overwrite_endio);
 		inc_all_io_entry(pool, bio);
 		remap_and_issue(tc, bio, data_block);
-	} else {
-		int r;
-		struct dm_io_region to;
 
-		to.bdev = tc->pool_dev->bdev;
-		to.sector = data_block * pool->sectors_per_block;
-		to.count = pool->sectors_per_block;
+	} else
+		ll_zero(tc, m,
+			data_block * pool->sectors_per_block,
+			(data_block + 1) * pool->sectors_per_block);
+}
 
-		r = dm_kcopyd_zero(pool->copier, 1, &to, 0, copy_complete, m);
-		if (r < 0) {
-			mempool_free(m, pool->mapping_pool);
-			DMERR_LIMIT("dm_kcopyd_zero() failed");
-			cell_error(pool, cell);
-		}
-	}
+static void schedule_external_copy(struct thin_c *tc, dm_block_t virt_block,
+				   dm_block_t data_dest,
+				   struct dm_bio_prison_cell *cell, struct bio *bio)
+{
+	struct pool *pool = tc->pool;
+	sector_t virt_block_begin = virt_block * pool->sectors_per_block;
+	sector_t virt_block_end = (virt_block + 1) * pool->sectors_per_block;
+
+	if (virt_block_end <= tc->origin_size)
+		schedule_copy(tc, virt_block, tc->origin_dev,
+			      virt_block, data_dest, cell, bio,
+			      pool->sectors_per_block);
+
+	else if (virt_block_begin < tc->origin_size)
+		schedule_copy(tc, virt_block, tc->origin_dev,
+			      virt_block, data_dest, cell, bio,
+			      tc->origin_size - virt_block_begin);
+
+	else
+		schedule_zero(tc, virt_block, data_dest, cell, bio);
 }
 
 /*
@@ -1315,7 +1369,18 @@
 			inc_all_io_entry(pool, bio);
 			cell_defer_no_holder(tc, cell);
 
-			remap_to_origin_and_issue(tc, bio);
+			if (bio_end_sector(bio) <= tc->origin_size)
+				remap_to_origin_and_issue(tc, bio);
+
+			else if (bio->bi_iter.bi_sector < tc->origin_size) {
+				zero_fill_bio(bio);
+				bio->bi_iter.bi_size = (tc->origin_size - bio->bi_iter.bi_sector) << SECTOR_SHIFT;
+				remap_to_origin_and_issue(tc, bio);
+
+			} else {
+				zero_fill_bio(bio);
+				bio_endio(bio, 0);
+			}
 		} else
 			provision_block(tc, bio, block, cell);
 		break;
@@ -3112,7 +3177,7 @@
 	 */
 	if (io_opt_sectors < pool->sectors_per_block ||
 	    do_div(io_opt_sectors, pool->sectors_per_block)) {
-		blk_limits_io_min(limits, 0);
+		blk_limits_io_min(limits, pool->sectors_per_block << SECTOR_SHIFT);
 		blk_limits_io_opt(limits, pool->sectors_per_block << SECTOR_SHIFT);
 	}
 
@@ -3141,7 +3206,7 @@
 	.name = "thin-pool",
 	.features = DM_TARGET_SINGLETON | DM_TARGET_ALWAYS_WRITEABLE |
 		    DM_TARGET_IMMUTABLE,
-	.version = {1, 12, 0},
+	.version = {1, 13, 0},
 	.module = THIS_MODULE,
 	.ctr = pool_ctr,
 	.dtr = pool_dtr,
@@ -3361,8 +3426,7 @@
 		spin_lock_irqsave(&pool->lock, flags);
 		list_for_each_entry_safe(m, tmp, &work, list) {
 			list_del(&m->list);
-			m->quiesced = true;
-			__maybe_add_mapping(m);
+			__complete_mapping_preparation(m);
 		}
 		spin_unlock_irqrestore(&pool->lock, flags);
 	}
@@ -3401,6 +3465,16 @@
 	noflush_work(tc, do_noflush_stop);
 }
 
+static int thin_preresume(struct dm_target *ti)
+{
+	struct thin_c *tc = ti->private;
+
+	if (tc->origin_dev)
+		tc->origin_size = get_dev_size(tc->origin_dev->bdev);
+
+	return 0;
+}
+
 /*
  * <nr mapped sectors> <highest mapped sector>
  */
@@ -3483,12 +3557,13 @@
 
 static struct target_type thin_target = {
 	.name = "thin",
-	.version = {1, 12, 0},
+	.version = {1, 13, 0},
 	.module	= THIS_MODULE,
 	.ctr = thin_ctr,
 	.dtr = thin_dtr,
 	.map = thin_map,
 	.end_io = thin_endio,
+	.preresume = thin_preresume,
 	.presuspend = thin_presuspend,
 	.postsuspend = thin_postsuspend,
 	.status = thin_status,
diff --git a/drivers/md/dm.h b/drivers/md/dm.h
index ed76126..e81d215 100644
--- a/drivers/md/dm.h
+++ b/drivers/md/dm.h
@@ -72,7 +72,6 @@
 unsigned dm_table_get_type(struct dm_table *t);
 struct target_type *dm_table_get_immutable_target_type(struct dm_table *t);
 bool dm_table_request_based(struct dm_table *t);
-bool dm_table_supports_discards(struct dm_table *t);
 void dm_table_free_md_mempools(struct dm_table *t);
 struct dm_md_mempools *dm_table_get_md_mempools(struct dm_table *t);
 
diff --git a/drivers/mfd/rtsx_usb.c b/drivers/mfd/rtsx_usb.c
index 6352bec..71f387c 100644
--- a/drivers/mfd/rtsx_usb.c
+++ b/drivers/mfd/rtsx_usb.c
@@ -744,6 +744,7 @@
 	{ USB_DEVICE(0x0BDA, 0x0140) },
 	{ }
 };
+MODULE_DEVICE_TABLE(usb, rtsx_usb_usb_ids);
 
 static struct usb_driver rtsx_usb_driver = {
 	.name			= "rtsx_usb",
diff --git a/drivers/mmc/card/block.c b/drivers/mmc/card/block.c
index 452782b..ede41f0 100644
--- a/drivers/mmc/card/block.c
+++ b/drivers/mmc/card/block.c
@@ -2028,8 +2028,7 @@
 		/* complete ongoing async transfer before issuing discard */
 		if (card->host->areq)
 			mmc_blk_issue_rw_rq(mq, NULL);
-		if (req->cmd_flags & REQ_SECURE &&
-			!(card->quirks & MMC_QUIRK_SEC_ERASE_TRIM_BROKEN))
+		if (req->cmd_flags & REQ_SECURE)
 			ret = mmc_blk_issue_secdiscard_rq(mq, req);
 		else
 			ret = mmc_blk_issue_discard_rq(mq, req);
@@ -2432,6 +2431,8 @@
 	if (!(card->csd.cmdclass & CCC_BLOCK_READ))
 		return -ENODEV;
 
+	mmc_fixup_device(card, blk_fixups);
+
 	md = mmc_blk_alloc(card);
 	if (IS_ERR(md))
 		return PTR_ERR(md);
@@ -2446,7 +2447,6 @@
 		goto out;
 
 	mmc_set_drvdata(card, md);
-	mmc_fixup_device(card, blk_fixups);
 
 	if (mmc_add_disk(md))
 		goto out;
diff --git a/drivers/mmc/core/bus.c b/drivers/mmc/core/bus.c
index d2dbf02..8a1f124 100644
--- a/drivers/mmc/core/bus.c
+++ b/drivers/mmc/core/bus.c
@@ -180,7 +180,6 @@
 #endif
 
 #ifdef CONFIG_PM_RUNTIME
-
 static int mmc_runtime_suspend(struct device *dev)
 {
 	struct mmc_card *card = mmc_dev_to_card(dev);
@@ -196,17 +195,10 @@
 
 	return host->bus_ops->runtime_resume(host);
 }
-
-static int mmc_runtime_idle(struct device *dev)
-{
-	return 0;
-}
-
 #endif /* !CONFIG_PM_RUNTIME */
 
 static const struct dev_pm_ops mmc_bus_pm_ops = {
-	SET_RUNTIME_PM_OPS(mmc_runtime_suspend, mmc_runtime_resume,
-			mmc_runtime_idle)
+	SET_RUNTIME_PM_OPS(mmc_runtime_suspend, mmc_runtime_resume, NULL)
 	SET_SYSTEM_SLEEP_PM_OPS(mmc_bus_suspend, mmc_bus_resume)
 };
 
diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
index 7dc0c85..d03a080 100644
--- a/drivers/mmc/core/core.c
+++ b/drivers/mmc/core/core.c
@@ -2102,7 +2102,8 @@
 
 int mmc_can_secure_erase_trim(struct mmc_card *card)
 {
-	if (card->ext_csd.sec_feature_support & EXT_CSD_SEC_ER_EN)
+	if ((card->ext_csd.sec_feature_support & EXT_CSD_SEC_ER_EN) &&
+	    !(card->quirks & MMC_QUIRK_SEC_ERASE_TRIM_BROKEN))
 		return 1;
 	return 0;
 }
diff --git a/drivers/mmc/core/mmc.c b/drivers/mmc/core/mmc.c
index 793c6f7..1eda8dd 100644
--- a/drivers/mmc/core/mmc.c
+++ b/drivers/mmc/core/mmc.c
@@ -324,13 +324,12 @@
 		}
 	}
 
+	/*
+	 * The EXT_CSD format is meant to be forward compatible. As long
+	 * as CSD_STRUCTURE does not change, all values for EXT_CSD_REV
+	 * are authorized, see JEDEC JESD84-B50 section B.8.
+	 */
 	card->ext_csd.rev = ext_csd[EXT_CSD_REV];
-	if (card->ext_csd.rev > 7) {
-		pr_err("%s: unrecognised EXT_CSD revision %d\n",
-			mmc_hostname(card->host), card->ext_csd.rev);
-		err = -EINVAL;
-		goto out;
-	}
 
 	card->ext_csd.raw_sectors[0] = ext_csd[EXT_CSD_SEC_CNT + 0];
 	card->ext_csd.raw_sectors[1] = ext_csd[EXT_CSD_SEC_CNT + 1];
diff --git a/drivers/mmc/core/quirks.c b/drivers/mmc/core/quirks.c
index 6c36fcc..dd1d1e0 100644
--- a/drivers/mmc/core/quirks.c
+++ b/drivers/mmc/core/quirks.c
@@ -91,7 +91,7 @@
 		    (f->cis_device == card->cis.device ||
 		     f->cis_device == (u16) SDIO_ANY_ID) &&
 		    rev >= f->rev_start && rev <= f->rev_end) {
-			dev_dbg(&card->dev, "calling %pF\n", f->vendor_fixup);
+			dev_dbg(&card->dev, "calling %pf\n", f->vendor_fixup);
 			f->vendor_fixup(card, f->data);
 		}
 	}
diff --git a/drivers/mmc/core/sd_ops.c b/drivers/mmc/core/sd_ops.c
index 274ef00..48d0c93 100644
--- a/drivers/mmc/core/sd_ops.c
+++ b/drivers/mmc/core/sd_ops.c
@@ -184,6 +184,9 @@
 		mmc_delay(10);
 	}
 
+	if (!i)
+		pr_err("%s: card never left busy state\n", mmc_hostname(host));
+
 	if (rocr && !mmc_host_is_spi(host))
 		*rocr = cmd.resp[0];
 
diff --git a/drivers/mmc/host/Kconfig b/drivers/mmc/host/Kconfig
index a565254..4511358 100644
--- a/drivers/mmc/host/Kconfig
+++ b/drivers/mmc/host/Kconfig
@@ -290,6 +290,18 @@
 	  be found on some embedded hardware such as UC-7112-LX.
 	  If you have a controller with this interface, say Y here.
 
+config MMC_SDHCI_ST
+	tristate "SDHCI support on STMicroelectronics SoC"
+	depends on ARCH_STI
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This selects the Secure Digital Host Controller Interface in
+	  STMicroelectronics SoCs.
+
+	  If you have a controller with this interface, say Y or M here.
+	  If unsure, say N.
+
 config MMC_OMAP
 	tristate "TI OMAP Multimedia Card Interface support"
 	depends on ARCH_OMAP
@@ -303,6 +315,7 @@
 
 config MMC_OMAP_HS
 	tristate "TI OMAP High Speed Multimedia Card Interface support"
+	depends on HAS_DMA
 	depends on ARCH_OMAP2PLUS || COMPILE_TEST
 	help
 	  This selects the TI OMAP High Speed Multimedia card Interface.
@@ -343,7 +356,7 @@
 
 config MMC_SDHCI_MSM
 	tristate "Qualcomm SDHCI Controller Support"
-	depends on ARCH_QCOM
+	depends on ARCH_QCOM || (ARM && COMPILE_TEST)
 	depends on MMC_SDHCI_PLTFM
 	help
 	  This selects the Secure Digital Host Controller Interface (SDHCI)
@@ -440,6 +453,7 @@
 config MMC_S3C
 	tristate "Samsung S3C SD/MMC Card Interface support"
 	depends on ARCH_S3C24XX
+	depends on S3C24XX_DMAC
 	help
 	  This selects a driver for the MCI interface found in
           Samsung's S3C2410, S3C2412, S3C2440, S3C2442 CPUs.
@@ -477,15 +491,6 @@
 	  working properly and needs to be debugged before this
 	  option is useful.
 
-config MMC_S3C_PIODMA
-	bool "Support for both PIO and DMA"
-	help
-	  Compile both the PIO and DMA transfer routines into the
-	  driver and let the platform select at run-time which one
-	  is best.
-
-	  See notes for the DMA option.
-
 endchoice
 
 config MMC_SDRICOH_CS
@@ -623,7 +628,7 @@
 
 config MMC_SH_MMCIF
 	tristate "SuperH Internal MMCIF support"
-	depends on MMC_BLOCK
+	depends on MMC_BLOCK && HAS_DMA
 	depends on SUPERH || ARCH_SHMOBILE || COMPILE_TEST
 	help
 	  This selects the MMC Host Interface controller (MMCIF).
@@ -697,6 +702,7 @@
 
 config MMC_USDHI6ROL0
 	tristate "Renesas USDHI6ROL0 SD/SDIO Host Controller support"
+	depends on HAS_DMA
 	help
 	  This selects support for the Renesas USDHI6ROL0 SD/SDIO
 	  Host Controller
diff --git a/drivers/mmc/host/Makefile b/drivers/mmc/host/Makefile
index 7f81ddf..f211eed 100644
--- a/drivers/mmc/host/Makefile
+++ b/drivers/mmc/host/Makefile
@@ -68,6 +68,7 @@
 obj-$(CONFIG_MMC_SDHCI_BCM_KONA)	+= sdhci-bcm-kona.o
 obj-$(CONFIG_MMC_SDHCI_BCM2835)		+= sdhci-bcm2835.o
 obj-$(CONFIG_MMC_SDHCI_MSM)		+= sdhci-msm.o
+obj-$(CONFIG_MMC_SDHCI_ST)		+= sdhci-st.o
 
 ifeq ($(CONFIG_CB710_DEBUG),y)
 	CFLAGS-cb710-mmc	+= -DDEBUG
diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
index 1ac227c..8f216ed 100644
--- a/drivers/mmc/host/dw_mmc.c
+++ b/drivers/mmc/host/dw_mmc.c
@@ -111,8 +111,7 @@
 	0xff, 0x77, 0x77, 0xff, 0x77, 0xbb, 0xdd, 0xee,
 };
 
-static inline bool dw_mci_fifo_reset(struct dw_mci *host);
-static inline bool dw_mci_ctrl_all_reset(struct dw_mci *host);
+static bool dw_mci_reset(struct dw_mci *host);
 
 #if defined(CONFIG_DEBUG_FS)
 static int dw_mci_req_show(struct seq_file *s, void *v)
@@ -997,7 +996,8 @@
 	int gpio_ro = mmc_gpio_get_ro(mmc);
 
 	/* Use platform get_ro function, else try on board write protect */
-	if (slot->quirks & DW_MCI_SLOT_QUIRK_NO_WRITE_PROTECT)
+	if ((slot->quirks & DW_MCI_SLOT_QUIRK_NO_WRITE_PROTECT) ||
+			(slot->host->quirks & DW_MCI_QUIRK_NO_WRITE_PROTECT))
 		read_only = 0;
 	else if (!IS_ERR_VALUE(gpio_ro))
 		read_only = gpio_ro;
@@ -1235,7 +1235,7 @@
 		 * After an error, there may be data lingering
 		 * in the FIFO
 		 */
-		dw_mci_fifo_reset(host);
+		dw_mci_reset(host);
 	} else {
 		data->bytes_xfered = data->blocks * data->blksz;
 		data->error = 0;
@@ -1352,7 +1352,7 @@
 
 			/* CMD error in data command */
 			if (mrq->cmd->error && mrq->data)
-				dw_mci_fifo_reset(host);
+				dw_mci_reset(host);
 
 			host->cmd = NULL;
 			host->data = NULL;
@@ -1963,14 +1963,8 @@
 			}
 
 			/* Power down slot */
-			if (present == 0) {
-				/* Clear down the FIFO */
-				dw_mci_fifo_reset(host);
-#ifdef CONFIG_MMC_DW_IDMAC
-				dw_mci_idmac_reset(host);
-#endif
-
-			}
+			if (present == 0)
+				dw_mci_reset(host);
 
 			spin_unlock_bh(&host->lock);
 
@@ -2021,8 +2015,11 @@
 
 	/* get quirks */
 	for (idx = 0; idx < ARRAY_SIZE(of_slot_quirks); idx++)
-		if (of_get_property(np, of_slot_quirks[idx].quirk, NULL))
+		if (of_get_property(np, of_slot_quirks[idx].quirk, NULL)) {
+			dev_warn(dev, "Slot quirk %s is deprecated\n",
+					of_slot_quirks[idx].quirk);
 			quirks |= of_slot_quirks[idx].id;
+		}
 
 	return quirks;
 }
@@ -2208,8 +2205,11 @@
 	return false;
 }
 
-static inline bool dw_mci_fifo_reset(struct dw_mci *host)
+static bool dw_mci_reset(struct dw_mci *host)
 {
+	u32 flags = SDMMC_CTRL_RESET | SDMMC_CTRL_FIFO_RESET;
+	bool ret = false;
+
 	/*
 	 * Reseting generates a block interrupt, hence setting
 	 * the scatter-gather pointer to NULL.
@@ -2219,15 +2219,60 @@
 		host->sg = NULL;
 	}
 
-	return dw_mci_ctrl_reset(host, SDMMC_CTRL_FIFO_RESET);
-}
+	if (host->use_dma)
+		flags |= SDMMC_CTRL_DMA_RESET;
 
-static inline bool dw_mci_ctrl_all_reset(struct dw_mci *host)
-{
-	return dw_mci_ctrl_reset(host,
-				 SDMMC_CTRL_FIFO_RESET |
-				 SDMMC_CTRL_RESET |
-				 SDMMC_CTRL_DMA_RESET);
+	if (dw_mci_ctrl_reset(host, flags)) {
+		/*
+		 * In all cases we clear the RAWINTS register to clear any
+		 * interrupts.
+		 */
+		mci_writel(host, RINTSTS, 0xFFFFFFFF);
+
+		/* if using dma we wait for dma_req to clear */
+		if (host->use_dma) {
+			unsigned long timeout = jiffies + msecs_to_jiffies(500);
+			u32 status;
+			do {
+				status = mci_readl(host, STATUS);
+				if (!(status & SDMMC_STATUS_DMA_REQ))
+					break;
+				cpu_relax();
+			} while (time_before(jiffies, timeout));
+
+			if (status & SDMMC_STATUS_DMA_REQ) {
+				dev_err(host->dev,
+					"%s: Timeout waiting for dma_req to "
+					"clear during reset\n", __func__);
+				goto ciu_out;
+			}
+
+			/* when using DMA next we reset the fifo again */
+			if (!dw_mci_ctrl_reset(host, SDMMC_CTRL_FIFO_RESET))
+				goto ciu_out;
+		}
+	} else {
+		/* if the controller reset bit did clear, then set clock regs */
+		if (!(mci_readl(host, CTRL) & SDMMC_CTRL_RESET)) {
+			dev_err(host->dev, "%s: fifo/dma reset bits didn't "
+				"clear but ciu was reset, doing clock update\n",
+				__func__);
+			goto ciu_out;
+		}
+	}
+
+#if IS_ENABLED(CONFIG_MMC_DW_IDMAC)
+	/* It is also recommended that we reset and reprogram idmac */
+	dw_mci_idmac_reset(host);
+#endif
+
+	ret = true;
+
+ciu_out:
+	/* After a CTRL reset we need to have CIU set clock registers  */
+	mci_send_cmd(host->cur_slot, SDMMC_CMD_UPD_CLK, 0);
+
+	return ret;
 }
 
 #ifdef CONFIG_OF
@@ -2238,6 +2283,9 @@
 	{
 		.quirk	= "broken-cd",
 		.id	= DW_MCI_QUIRK_BROKEN_CARD_DETECTION,
+	}, {
+		.quirk	= "disable-wp",
+		.id	= DW_MCI_QUIRK_NO_WRITE_PROTECT,
 	},
 };
 
@@ -2425,7 +2473,7 @@
 	}
 
 	/* Reset all blocks */
-	if (!dw_mci_ctrl_all_reset(host))
+	if (!dw_mci_ctrl_reset(host, SDMMC_CTRL_ALL_RESET_FLAGS))
 		return -ENODEV;
 
 	host->dma_ops = host->pdata->dma_ops;
@@ -2612,7 +2660,7 @@
 		}
 	}
 
-	if (!dw_mci_ctrl_all_reset(host)) {
+	if (!dw_mci_ctrl_reset(host, SDMMC_CTRL_ALL_RESET_FLAGS)) {
 		ret = -ENODEV;
 		return ret;
 	}
diff --git a/drivers/mmc/host/dw_mmc.h b/drivers/mmc/host/dw_mmc.h
index 738fa24..08fd956 100644
--- a/drivers/mmc/host/dw_mmc.h
+++ b/drivers/mmc/host/dw_mmc.h
@@ -129,6 +129,7 @@
 #define SDMMC_CMD_INDX(n)		((n) & 0x1F)
 /* Status register defines */
 #define SDMMC_GET_FCNT(x)		(((x)>>17) & 0x1FFF)
+#define SDMMC_STATUS_DMA_REQ		BIT(31)
 /* FIFOTH register defines */
 #define SDMMC_SET_FIFOTH(m, r, t)	(((m) & 0x7) << 28 | \
 					 ((r) & 0xFFF) << 16 | \
@@ -150,6 +151,10 @@
 /* Card read threshold */
 #define SDMMC_SET_RD_THLD(v, x)		(((v) & 0x1FFF) << 16 | (x))
 
+/* All ctrl reset bits */
+#define SDMMC_CTRL_ALL_RESET_FLAGS \
+	(SDMMC_CTRL_RESET | SDMMC_CTRL_FIFO_RESET | SDMMC_CTRL_DMA_RESET)
+
 /* Register access macros */
 #define mci_readl(dev, reg)			\
 	__raw_readl((dev)->regs + SDMMC_##reg)
diff --git a/drivers/mmc/host/mmci.c b/drivers/mmc/host/mmci.c
index 7ad463e..e4d4707 100644
--- a/drivers/mmc/host/mmci.c
+++ b/drivers/mmc/host/mmci.c
@@ -52,34 +52,53 @@
  * struct variant_data - MMCI variant-specific quirks
  * @clkreg: default value for MCICLOCK register
  * @clkreg_enable: enable value for MMCICLOCK register
+ * @clkreg_8bit_bus_enable: enable value for 8 bit bus
+ * @clkreg_neg_edge_enable: enable value for inverted data/cmd output
  * @datalength_bits: number of bits in the MMCIDATALENGTH register
  * @fifosize: number of bytes that can be written when MMCI_TXFIFOEMPTY
  *	      is asserted (likewise for RX)
  * @fifohalfsize: number of bytes that can be written when MCI_TXFIFOHALFEMPTY
  *		  is asserted (likewise for RX)
+ * @data_cmd_enable: enable value for data commands.
  * @sdio: variant supports SDIO
  * @st_clkdiv: true if using a ST-specific clock divider algorithm
+ * @datactrl_mask_ddrmode: ddr mode mask in datactrl register.
  * @blksz_datactrl16: true if Block size is at b16..b30 position in datactrl register
+ * @blksz_datactrl4: true if Block size is at b4..b16 position in datactrl
+ *		     register
  * @pwrreg_powerup: power up value for MMCIPOWER register
+ * @f_max: maximum clk frequency supported by the controller.
  * @signal_direction: input/out direction of bus signals can be indicated
  * @pwrreg_clkgate: MMCIPOWER register must be used to gate the clock
  * @busy_detect: true if busy detection on dat0 is supported
  * @pwrreg_nopower: bits in MMCIPOWER don't controls ext. power supply
+ * @explicit_mclk_control: enable explicit mclk control in driver.
+ * @qcom_fifo: enables qcom specific fifo pio read logic.
+ * @reversed_irq_handling: handle data irq before cmd irq.
  */
 struct variant_data {
 	unsigned int		clkreg;
 	unsigned int		clkreg_enable;
+	unsigned int		clkreg_8bit_bus_enable;
+	unsigned int		clkreg_neg_edge_enable;
 	unsigned int		datalength_bits;
 	unsigned int		fifosize;
 	unsigned int		fifohalfsize;
+	unsigned int		data_cmd_enable;
+	unsigned int		datactrl_mask_ddrmode;
 	bool			sdio;
 	bool			st_clkdiv;
 	bool			blksz_datactrl16;
+	bool			blksz_datactrl4;
 	u32			pwrreg_powerup;
+	u32			f_max;
 	bool			signal_direction;
 	bool			pwrreg_clkgate;
 	bool			busy_detect;
 	bool			pwrreg_nopower;
+	bool			explicit_mclk_control;
+	bool			qcom_fifo;
+	bool			reversed_irq_handling;
 };
 
 static struct variant_data variant_arm = {
@@ -87,6 +106,8 @@
 	.fifohalfsize		= 8 * 4,
 	.datalength_bits	= 16,
 	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 100000000,
+	.reversed_irq_handling	= true,
 };
 
 static struct variant_data variant_arm_extended_fifo = {
@@ -94,6 +115,7 @@
 	.fifohalfsize		= 64 * 4,
 	.datalength_bits	= 16,
 	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 100000000,
 };
 
 static struct variant_data variant_arm_extended_fifo_hwfc = {
@@ -102,15 +124,18 @@
 	.clkreg_enable		= MCI_ARM_HWFCEN,
 	.datalength_bits	= 16,
 	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 100000000,
 };
 
 static struct variant_data variant_u300 = {
 	.fifosize		= 16 * 4,
 	.fifohalfsize		= 8 * 4,
 	.clkreg_enable		= MCI_ST_U300_HWFCEN,
+	.clkreg_8bit_bus_enable = MCI_ST_8BIT_BUS,
 	.datalength_bits	= 16,
 	.sdio			= true,
 	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
 	.signal_direction	= true,
 	.pwrreg_clkgate		= true,
 	.pwrreg_nopower		= true,
@@ -124,6 +149,7 @@
 	.sdio			= true,
 	.st_clkdiv		= true,
 	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
 	.signal_direction	= true,
 	.pwrreg_clkgate		= true,
 	.pwrreg_nopower		= true,
@@ -134,10 +160,13 @@
 	.fifohalfsize		= 8 * 4,
 	.clkreg			= MCI_CLK_ENABLE,
 	.clkreg_enable		= MCI_ST_UX500_HWFCEN,
+	.clkreg_8bit_bus_enable = MCI_ST_8BIT_BUS,
+	.clkreg_neg_edge_enable	= MCI_ST_UX500_NEG_EDGE,
 	.datalength_bits	= 24,
 	.sdio			= true,
 	.st_clkdiv		= true,
 	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
 	.signal_direction	= true,
 	.pwrreg_clkgate		= true,
 	.busy_detect		= true,
@@ -149,17 +178,38 @@
 	.fifohalfsize		= 8 * 4,
 	.clkreg			= MCI_CLK_ENABLE,
 	.clkreg_enable		= MCI_ST_UX500_HWFCEN,
+	.clkreg_8bit_bus_enable = MCI_ST_8BIT_BUS,
+	.clkreg_neg_edge_enable	= MCI_ST_UX500_NEG_EDGE,
+	.datactrl_mask_ddrmode	= MCI_ST_DPSM_DDRMODE,
 	.datalength_bits	= 24,
 	.sdio			= true,
 	.st_clkdiv		= true,
 	.blksz_datactrl16	= true,
 	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
 	.signal_direction	= true,
 	.pwrreg_clkgate		= true,
 	.busy_detect		= true,
 	.pwrreg_nopower		= true,
 };
 
+static struct variant_data variant_qcom = {
+	.fifosize		= 16 * 4,
+	.fifohalfsize		= 8 * 4,
+	.clkreg			= MCI_CLK_ENABLE,
+	.clkreg_enable		= MCI_QCOM_CLK_FLOWENA |
+				  MCI_QCOM_CLK_SELECT_IN_FBCLK,
+	.clkreg_8bit_bus_enable = MCI_QCOM_CLK_WIDEBUS_8,
+	.datactrl_mask_ddrmode	= MCI_QCOM_CLK_SELECT_IN_DDR_MODE,
+	.data_cmd_enable	= MCI_QCOM_CSPM_DATCMD,
+	.blksz_datactrl4	= true,
+	.datalength_bits	= 24,
+	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 208000000,
+	.explicit_mclk_control	= true,
+	.qcom_fifo		= true,
+};
+
 static int mmci_card_busy(struct mmc_host *mmc)
 {
 	struct mmci_host *host = mmc_priv(mmc);
@@ -260,7 +310,9 @@
 	host->cclk = 0;
 
 	if (desired) {
-		if (desired >= host->mclk) {
+		if (variant->explicit_mclk_control) {
+			host->cclk = host->mclk;
+		} else if (desired >= host->mclk) {
 			clk = MCI_CLK_BYPASS;
 			if (variant->st_clkdiv)
 				clk |= MCI_ST_UX500_NEG_EDGE;
@@ -299,11 +351,11 @@
 	if (host->mmc->ios.bus_width == MMC_BUS_WIDTH_4)
 		clk |= MCI_4BIT_BUS;
 	if (host->mmc->ios.bus_width == MMC_BUS_WIDTH_8)
-		clk |= MCI_ST_8BIT_BUS;
+		clk |= variant->clkreg_8bit_bus_enable;
 
 	if (host->mmc->ios.timing == MMC_TIMING_UHS_DDR50 ||
 	    host->mmc->ios.timing == MMC_TIMING_MMC_DDR52)
-		clk |= MCI_ST_UX500_NEG_EDGE;
+		clk |= variant->clkreg_neg_edge_enable;
 
 	mmci_write_clkreg(host, clk);
 }
@@ -719,7 +771,7 @@
 	data->bytes_xfered = 0;
 
 	clks = (unsigned long long)data->timeout_ns * host->cclk;
-	do_div(clks, 1000000000UL);
+	do_div(clks, NSEC_PER_SEC);
 
 	timeout = data->timeout_clks + (unsigned int)clks;
 
@@ -732,6 +784,8 @@
 
 	if (variant->blksz_datactrl16)
 		datactrl = MCI_DPSM_ENABLE | (data->blksz << 16);
+	else if (variant->blksz_datactrl4)
+		datactrl = MCI_DPSM_ENABLE | (data->blksz << 4);
 	else
 		datactrl = MCI_DPSM_ENABLE | blksz_bits << 4;
 
@@ -767,7 +821,7 @@
 
 	if (host->mmc->ios.timing == MMC_TIMING_UHS_DDR50 ||
 	    host->mmc->ios.timing == MMC_TIMING_MMC_DDR52)
-		datactrl |= MCI_ST_DPSM_DDRMODE;
+		datactrl |= variant->datactrl_mask_ddrmode;
 
 	/*
 	 * Attempt to use DMA operation mode, if this
@@ -812,7 +866,7 @@
 
 	if (readl(base + MMCICOMMAND) & MCI_CPSM_ENABLE) {
 		writel(0, base + MMCICOMMAND);
-		udelay(1);
+		mmci_reg_delay(host);
 	}
 
 	c |= cmd->opcode | MCI_CPSM_ENABLE;
@@ -824,6 +878,9 @@
 	if (/*interrupt*/0)
 		c |= MCI_CPSM_INTERRUPT;
 
+	if (mmc_cmd_type(cmd) == MMC_CMD_ADTC)
+		c |= host->variant->data_cmd_enable;
+
 	host->cmd = cmd;
 
 	writel(cmd->arg, base + MMCIARGUMENT);
@@ -834,6 +891,10 @@
 mmci_data_irq(struct mmci_host *host, struct mmc_data *data,
 	      unsigned int status)
 {
+	/* Make sure we have data to handle */
+	if (!data)
+		return;
+
 	/* First check for errors */
 	if (status & (MCI_DATACRCFAIL|MCI_DATATIMEOUT|MCI_STARTBITERR|
 		      MCI_TXUNDERRUN|MCI_RXOVERRUN)) {
@@ -902,9 +963,17 @@
 	     unsigned int status)
 {
 	void __iomem *base = host->base;
-	bool sbc = (cmd == host->mrq->sbc);
-	bool busy_resp = host->variant->busy_detect &&
-			(cmd->flags & MMC_RSP_BUSY);
+	bool sbc, busy_resp;
+
+	if (!cmd)
+		return;
+
+	sbc = (cmd == host->mrq->sbc);
+	busy_resp = host->variant->busy_detect && (cmd->flags & MMC_RSP_BUSY);
+
+	if (!((status|host->busy_status) & (MCI_CMDCRCFAIL|MCI_CMDTIMEOUT|
+		MCI_CMDSENT|MCI_CMDRESPEND)))
+		return;
 
 	/* Check if we need to wait for busy completion. */
 	if (host->busy_status && (status & MCI_ST_CARDBUSY))
@@ -957,15 +1026,34 @@
 	}
 }
 
+static int mmci_get_rx_fifocnt(struct mmci_host *host, u32 status, int remain)
+{
+	return remain - (readl(host->base + MMCIFIFOCNT) << 2);
+}
+
+static int mmci_qcom_get_rx_fifocnt(struct mmci_host *host, u32 status, int r)
+{
+	/*
+	 * on qcom SDCC4 only 8 words are used in each burst so only 8 addresses
+	 * from the fifo range should be used
+	 */
+	if (status & MCI_RXFIFOHALFFULL)
+		return host->variant->fifohalfsize;
+	else if (status & MCI_RXDATAAVLBL)
+		return 4;
+
+	return 0;
+}
+
 static int mmci_pio_read(struct mmci_host *host, char *buffer, unsigned int remain)
 {
 	void __iomem *base = host->base;
 	char *ptr = buffer;
-	u32 status;
+	u32 status = readl(host->base + MMCISTATUS);
 	int host_remain = host->size;
 
 	do {
-		int count = host_remain - (readl(base + MMCIFIFOCNT) << 2);
+		int count = host->get_rx_fifocnt(host, status, host_remain);
 
 		if (count > remain)
 			count = remain;
@@ -1132,9 +1220,6 @@
 	spin_lock(&host->lock);
 
 	do {
-		struct mmc_command *cmd;
-		struct mmc_data *data;
-
 		status = readl(host->base + MMCISTATUS);
 
 		if (host->singleirq) {
@@ -1154,16 +1239,13 @@
 
 		dev_dbg(mmc_dev(host->mmc), "irq0 (data+cmd) %08x\n", status);
 
-		cmd = host->cmd;
-		if ((status|host->busy_status) & (MCI_CMDCRCFAIL|MCI_CMDTIMEOUT|
-			MCI_CMDSENT|MCI_CMDRESPEND) && cmd)
-			mmci_cmd_irq(host, cmd, status);
-
-		data = host->data;
-		if (status & (MCI_DATACRCFAIL|MCI_DATATIMEOUT|MCI_STARTBITERR|
-			      MCI_TXUNDERRUN|MCI_RXOVERRUN|MCI_DATAEND|
-			      MCI_DATABLOCKEND) && data)
-			mmci_data_irq(host, data, status);
+		if (host->variant->reversed_irq_handling) {
+			mmci_data_irq(host, host->data, status);
+			mmci_cmd_irq(host, host->cmd, status);
+		} else {
+			mmci_cmd_irq(host, host->cmd, status);
+			mmci_data_irq(host, host->data, status);
+		}
 
 		/* Don't poll for busy completion in irq context. */
 		if (host->busy_status)
@@ -1296,6 +1378,17 @@
 	if (!ios->clock && variant->pwrreg_clkgate)
 		pwr &= ~MCI_PWR_ON;
 
+	if (host->variant->explicit_mclk_control &&
+	    ios->clock != host->clock_cache) {
+		ret = clk_set_rate(host->clk, ios->clock);
+		if (ret < 0)
+			dev_err(mmc_dev(host->mmc),
+				"Error setting clock rate (%d)\n", ret);
+		else
+			host->mclk = clk_get_rate(host->clk);
+	}
+	host->clock_cache = ios->clock;
+
 	spin_lock_irqsave(&host->lock, flags);
 
 	mmci_set_clkreg(host, ios->clock);
@@ -1443,6 +1536,11 @@
 	if (ret)
 		goto host_free;
 
+	if (variant->qcom_fifo)
+		host->get_rx_fifocnt = mmci_qcom_get_rx_fifocnt;
+	else
+		host->get_rx_fifocnt = mmci_get_rx_fifocnt;
+
 	host->plat = plat;
 	host->variant = variant;
 	host->mclk = clk_get_rate(host->clk);
@@ -1451,8 +1549,8 @@
 	 * so we try to adjust the clock down to this,
 	 * (if possible).
 	 */
-	if (host->mclk > 100000000) {
-		ret = clk_set_rate(host->clk, 100000000);
+	if (host->mclk > variant->f_max) {
+		ret = clk_set_rate(host->clk, variant->f_max);
 		if (ret < 0)
 			goto clk_disable;
 		host->mclk = clk_get_rate(host->clk);
@@ -1471,9 +1569,12 @@
 	 * The ARM and ST versions of the block have slightly different
 	 * clock divider equations which means that the minimum divider
 	 * differs too.
+	 * on Qualcomm like controllers get the nearest minimum clock to 100Khz
 	 */
 	if (variant->st_clkdiv)
 		mmc->f_min = DIV_ROUND_UP(host->mclk, 257);
+	else if (variant->explicit_mclk_control)
+		mmc->f_min = clk_round_rate(host->clk, 100000);
 	else
 		mmc->f_min = DIV_ROUND_UP(host->mclk, 512);
 	/*
@@ -1483,9 +1584,14 @@
 	 * the block, of course.
 	 */
 	if (mmc->f_max)
-		mmc->f_max = min(host->mclk, mmc->f_max);
+		mmc->f_max = variant->explicit_mclk_control ?
+				min(variant->f_max, mmc->f_max) :
+				min(host->mclk, mmc->f_max);
 	else
-		mmc->f_max = min(host->mclk, fmax);
+		mmc->f_max = variant->explicit_mclk_control ?
+				fmax : min(host->mclk, fmax);
+
+
 	dev_dbg(mmc_dev(mmc), "clocking block at %u Hz\n", mmc->f_max);
 
 	/* Get regulators and the supported OCR mask */
@@ -1752,6 +1858,12 @@
 		.mask   = 0xf0ffffff,
 		.data	= &variant_ux500v2,
 	},
+	/* Qualcomm variants */
+	{
+		.id     = 0x00051180,
+		.mask	= 0x000fffff,
+		.data	= &variant_qcom,
+	},
 	{ 0, 0 },
 };
 
diff --git a/drivers/mmc/host/mmci.h b/drivers/mmc/host/mmci.h
index 347d942..a1f5e4f 100644
--- a/drivers/mmc/host/mmci.h
+++ b/drivers/mmc/host/mmci.h
@@ -41,6 +41,15 @@
 /* Modified PL180 on Versatile Express platform */
 #define MCI_ARM_HWFCEN		(1 << 12)
 
+/* Modified on Qualcomm Integrations */
+#define MCI_QCOM_CLK_WIDEBUS_8	(BIT(10) | BIT(11))
+#define MCI_QCOM_CLK_FLOWENA	BIT(12)
+#define MCI_QCOM_CLK_INVERTOUT	BIT(13)
+
+/* select in latch data and command in */
+#define MCI_QCOM_CLK_SELECT_IN_FBCLK	BIT(15)
+#define MCI_QCOM_CLK_SELECT_IN_DDR_MODE	(BIT(14) | BIT(15))
+
 #define MMCIARGUMENT		0x008
 #define MMCICOMMAND		0x00c
 #define MCI_CPSM_RESPONSE	(1 << 6)
@@ -54,6 +63,14 @@
 #define MCI_ST_NIEN		(1 << 13)
 #define MCI_ST_CE_ATACMD	(1 << 14)
 
+/* Modified on Qualcomm Integrations */
+#define MCI_QCOM_CSPM_DATCMD		BIT(12)
+#define MCI_QCOM_CSPM_MCIABORT		BIT(13)
+#define MCI_QCOM_CSPM_CCSENABLE		BIT(14)
+#define MCI_QCOM_CSPM_CCSDISABLE	BIT(15)
+#define MCI_QCOM_CSPM_AUTO_CMD19	BIT(16)
+#define MCI_QCOM_CSPM_AUTO_CMD21	BIT(21)
+
 #define MMCIRESPCMD		0x010
 #define MMCIRESPONSE0		0x014
 #define MMCIRESPONSE1		0x018
@@ -191,6 +208,8 @@
 	spinlock_t		lock;
 
 	unsigned int		mclk;
+	/* cached value of requested clk in set_ios */
+	unsigned int		clock_cache;
 	unsigned int		cclk;
 	u32			pwr_reg;
 	u32			pwr_reg_add;
@@ -210,6 +229,7 @@
 	/* pio stuff */
 	struct sg_mapping_iter	sg_miter;
 	unsigned int		size;
+	int (*get_rx_fifocnt)(struct mmci_host *h, u32 status, int remain);
 
 #ifdef CONFIG_DMA_ENGINE
 	/* DMA stuff */
diff --git a/drivers/mmc/host/moxart-mmc.c b/drivers/mmc/host/moxart-mmc.c
index 74924a0..b4b1efb 100644
--- a/drivers/mmc/host/moxart-mmc.c
+++ b/drivers/mmc/host/moxart-mmc.c
@@ -13,7 +13,6 @@
  * warranty of any kind, whether express or implied.
  */
 
-#include <linux/version.h>
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/platform_device.h>
diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c
index babfea0..140885a 100644
--- a/drivers/mmc/host/mxs-mmc.c
+++ b/drivers/mmc/host/mxs-mmc.c
@@ -86,7 +86,8 @@
 	if (ret >= 0)
 		return ret;
 
-	present = !(readl(ssp->base + HW_SSP_STATUS(ssp)) &
+	present = mmc->caps & MMC_CAP_NEEDS_POLL ||
+		!(readl(ssp->base + HW_SSP_STATUS(ssp)) &
 			BM_SSP_STATUS_CARD_DETECT);
 
 	if (mmc->caps2 & MMC_CAP2_CD_ACTIVE_HIGH)
diff --git a/drivers/mmc/host/omap_hsmmc.c b/drivers/mmc/host/omap_hsmmc.c
index 6b7b755..9656726 100644
--- a/drivers/mmc/host/omap_hsmmc.c
+++ b/drivers/mmc/host/omap_hsmmc.c
@@ -29,6 +29,7 @@
 #include <linux/timer.h>
 #include <linux/clk.h>
 #include <linux/of.h>
+#include <linux/of_irq.h>
 #include <linux/of_gpio.h>
 #include <linux/of_device.h>
 #include <linux/omap-dmaengine.h>
@@ -36,6 +37,7 @@
 #include <linux/mmc/core.h>
 #include <linux/mmc/mmc.h>
 #include <linux/io.h>
+#include <linux/irq.h>
 #include <linux/gpio.h>
 #include <linux/regulator/consumer.h>
 #include <linux/pinctrl/consumer.h>
@@ -54,6 +56,7 @@
 #define OMAP_HSMMC_RSP54	0x0118
 #define OMAP_HSMMC_RSP76	0x011C
 #define OMAP_HSMMC_DATA		0x0120
+#define OMAP_HSMMC_PSTATE	0x0124
 #define OMAP_HSMMC_HCTL		0x0128
 #define OMAP_HSMMC_SYSCTL	0x012C
 #define OMAP_HSMMC_STAT		0x0130
@@ -91,7 +94,10 @@
 #define BCE			(1 << 1)
 #define FOUR_BIT		(1 << 1)
 #define HSPE			(1 << 2)
+#define IWE			(1 << 24)
 #define DDR			(1 << 19)
+#define CLKEXTFREE		(1 << 16)
+#define CTPL			(1 << 11)
 #define DW8			(1 << 5)
 #define OD			0x1
 #define STAT_CLEAR		0xFFFFFFFF
@@ -101,11 +107,15 @@
 #define SRD			(1 << 26)
 #define SOFTRESET		(1 << 1)
 
+/* PSTATE */
+#define DLEV_DAT(x)		(1 << (20 + (x)))
+
 /* Interrupt masks for IE and ISE register */
 #define CC_EN			(1 << 0)
 #define TC_EN			(1 << 1)
 #define BWR_EN			(1 << 4)
 #define BRR_EN			(1 << 5)
+#define CIRQ_EN			(1 << 8)
 #define ERR_EN			(1 << 15)
 #define CTO_EN			(1 << 16)
 #define CCRC_EN			(1 << 17)
@@ -140,7 +150,6 @@
 #define VDD_3V0			3000000		/* 300000 uV */
 #define VDD_165_195		(ffs(MMC_VDD_165_195) - 1)
 
-#define AUTO_CMD23		(1 << 1)	/* Auto CMD23 support */
 /*
  * One controller can have multiple slots, like on some omap boards using
  * omap.c controller driver. Luckily this is not currently done on any known
@@ -194,6 +203,7 @@
 	u32			sysctl;
 	u32			capa;
 	int			irq;
+	int			wake_irq;
 	int			use_dma, dma_ch;
 	struct dma_chan		*tx_chan;
 	struct dma_chan		*rx_chan;
@@ -206,6 +216,9 @@
 	int			req_in_progress;
 	unsigned long		clk_rate;
 	unsigned int		flags;
+#define AUTO_CMD23		(1 << 0)        /* Auto CMD23 support */
+#define HSMMC_SDIO_IRQ_ENABLED	(1 << 1)        /* SDIO irq enabled */
+#define HSMMC_WAKE_IRQ_ENABLED	(1 << 2)
 	struct omap_hsmmc_next	next_data;
 	struct	omap_mmc_platform_data	*pdata;
 };
@@ -510,27 +523,40 @@
 static void omap_hsmmc_enable_irq(struct omap_hsmmc_host *host,
 				  struct mmc_command *cmd)
 {
-	unsigned int irq_mask;
+	u32 irq_mask = INT_EN_MASK;
+	unsigned long flags;
 
 	if (host->use_dma)
-		irq_mask = INT_EN_MASK & ~(BRR_EN | BWR_EN);
-	else
-		irq_mask = INT_EN_MASK;
+		irq_mask &= ~(BRR_EN | BWR_EN);
 
 	/* Disable timeout for erases */
 	if (cmd->opcode == MMC_ERASE)
 		irq_mask &= ~DTO_EN;
 
+	spin_lock_irqsave(&host->irq_lock, flags);
 	OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
 	OMAP_HSMMC_WRITE(host->base, ISE, irq_mask);
+
+	/* latch pending CIRQ, but don't signal MMC core */
+	if (host->flags & HSMMC_SDIO_IRQ_ENABLED)
+		irq_mask |= CIRQ_EN;
 	OMAP_HSMMC_WRITE(host->base, IE, irq_mask);
+	spin_unlock_irqrestore(&host->irq_lock, flags);
 }
 
 static void omap_hsmmc_disable_irq(struct omap_hsmmc_host *host)
 {
-	OMAP_HSMMC_WRITE(host->base, ISE, 0);
-	OMAP_HSMMC_WRITE(host->base, IE, 0);
+	u32 irq_mask = 0;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->irq_lock, flags);
+	/* no transfer running but need to keep cirq if enabled */
+	if (host->flags & HSMMC_SDIO_IRQ_ENABLED)
+		irq_mask |= CIRQ_EN;
+	OMAP_HSMMC_WRITE(host->base, ISE, irq_mask);
+	OMAP_HSMMC_WRITE(host->base, IE, irq_mask);
 	OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
+	spin_unlock_irqrestore(&host->irq_lock, flags);
 }
 
 /* Calculate divisor for the given clock frequency */
@@ -667,6 +693,9 @@
 		capa = VS18;
 	}
 
+	if (host->mmc->caps & MMC_CAP_SDIO_IRQ)
+		hctl |= IWE;
+
 	OMAP_HSMMC_WRITE(host->base, HCTL,
 			OMAP_HSMMC_READ(host->base, HCTL) | hctl);
 
@@ -681,7 +710,9 @@
 		&& time_before(jiffies, timeout))
 		;
 
-	omap_hsmmc_disable_irq(host);
+	OMAP_HSMMC_WRITE(host->base, ISE, 0);
+	OMAP_HSMMC_WRITE(host->base, IE, 0);
+	OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
 
 	/* Do not initialize card-specific things if the power is off */
 	if (host->power_mode == MMC_POWER_OFF)
@@ -1118,8 +1149,12 @@
 	int status;
 
 	status = OMAP_HSMMC_READ(host->base, STAT);
-	while (status & INT_EN_MASK && host->req_in_progress) {
-		omap_hsmmc_do_irq(host, status);
+	while (status & (INT_EN_MASK | CIRQ_EN)) {
+		if (host->req_in_progress)
+			omap_hsmmc_do_irq(host, status);
+
+		if (status & CIRQ_EN)
+			mmc_signal_sdio_irq(host->mmc);
 
 		/* Flush posted write */
 		status = OMAP_HSMMC_READ(host->base, STAT);
@@ -1128,6 +1163,22 @@
 	return IRQ_HANDLED;
 }
 
+static irqreturn_t omap_hsmmc_wake_irq(int irq, void *dev_id)
+{
+	struct omap_hsmmc_host *host = dev_id;
+
+	/* cirq is level triggered, disable to avoid infinite loop */
+	spin_lock(&host->irq_lock);
+	if (host->flags & HSMMC_WAKE_IRQ_ENABLED) {
+		disable_irq_nosync(host->wake_irq);
+		host->flags &= ~HSMMC_WAKE_IRQ_ENABLED;
+	}
+	spin_unlock(&host->irq_lock);
+	pm_request_resume(host->dev); /* no use counter */
+
+	return IRQ_HANDLED;
+}
+
 static void set_sd_bus_power(struct omap_hsmmc_host *host)
 {
 	unsigned long i;
@@ -1639,6 +1690,103 @@
 		mmc_slot(host).init_card(card);
 }
 
+static void omap_hsmmc_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+	u32 irq_mask, con;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->irq_lock, flags);
+
+	con = OMAP_HSMMC_READ(host->base, CON);
+	irq_mask = OMAP_HSMMC_READ(host->base, ISE);
+	if (enable) {
+		host->flags |= HSMMC_SDIO_IRQ_ENABLED;
+		irq_mask |= CIRQ_EN;
+		con |= CTPL | CLKEXTFREE;
+	} else {
+		host->flags &= ~HSMMC_SDIO_IRQ_ENABLED;
+		irq_mask &= ~CIRQ_EN;
+		con &= ~(CTPL | CLKEXTFREE);
+	}
+	OMAP_HSMMC_WRITE(host->base, CON, con);
+	OMAP_HSMMC_WRITE(host->base, IE, irq_mask);
+
+	/*
+	 * if enable, piggy back detection on current request
+	 * but always disable immediately
+	 */
+	if (!host->req_in_progress || !enable)
+		OMAP_HSMMC_WRITE(host->base, ISE, irq_mask);
+
+	/* flush posted write */
+	OMAP_HSMMC_READ(host->base, IE);
+
+	spin_unlock_irqrestore(&host->irq_lock, flags);
+}
+
+static int omap_hsmmc_configure_wake_irq(struct omap_hsmmc_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	int ret;
+
+	/*
+	 * For omaps with wake-up path, wakeirq will be irq from pinctrl and
+	 * for other omaps, wakeirq will be from GPIO (dat line remuxed to
+	 * gpio). wakeirq is needed to detect sdio irq in runtime suspend state
+	 * with functional clock disabled.
+	 */
+	if (!host->dev->of_node || !host->wake_irq)
+		return -ENODEV;
+
+	/* Prevent auto-enabling of IRQ */
+	irq_set_status_flags(host->wake_irq, IRQ_NOAUTOEN);
+	ret = devm_request_irq(host->dev, host->wake_irq, omap_hsmmc_wake_irq,
+			       IRQF_TRIGGER_LOW | IRQF_ONESHOT,
+			       mmc_hostname(mmc), host);
+	if (ret) {
+		dev_err(mmc_dev(host->mmc), "Unable to request wake IRQ\n");
+		goto err;
+	}
+
+	/*
+	 * Some omaps don't have wake-up path from deeper idle states
+	 * and need to remux SDIO DAT1 to GPIO for wake-up from idle.
+	 */
+	if (host->pdata->controller_flags & OMAP_HSMMC_SWAKEUP_MISSING) {
+		struct pinctrl *p = devm_pinctrl_get(host->dev);
+		if (!p) {
+			ret = -ENODEV;
+			goto err_free_irq;
+		}
+		if (IS_ERR(pinctrl_lookup_state(p, PINCTRL_STATE_DEFAULT))) {
+			dev_info(host->dev, "missing default pinctrl state\n");
+			devm_pinctrl_put(p);
+			ret = -EINVAL;
+			goto err_free_irq;
+		}
+
+		if (IS_ERR(pinctrl_lookup_state(p, PINCTRL_STATE_IDLE))) {
+			dev_info(host->dev, "missing idle pinctrl state\n");
+			devm_pinctrl_put(p);
+			ret = -EINVAL;
+			goto err_free_irq;
+		}
+		devm_pinctrl_put(p);
+	}
+
+	OMAP_HSMMC_WRITE(host->base, HCTL,
+			 OMAP_HSMMC_READ(host->base, HCTL) | IWE);
+	return 0;
+
+err_free_irq:
+	devm_free_irq(host->dev, host->wake_irq, host);
+err:
+	dev_warn(host->dev, "no SDIO IRQ support, falling back to polling\n");
+	host->wake_irq = 0;
+	return ret;
+}
+
 static void omap_hsmmc_conf_bus_power(struct omap_hsmmc_host *host)
 {
 	u32 hctl, capa, value;
@@ -1691,7 +1839,7 @@
 	.get_cd = omap_hsmmc_get_cd,
 	.get_ro = omap_hsmmc_get_ro,
 	.init_card = omap_hsmmc_init_card,
-	/* NYET -- enable_sdio_irq */
+	.enable_sdio_irq = omap_hsmmc_enable_sdio_irq,
 };
 
 #ifdef CONFIG_DEBUG_FS
@@ -1701,13 +1849,23 @@
 	struct mmc_host *mmc = s->private;
 	struct omap_hsmmc_host *host = mmc_priv(mmc);
 
-	seq_printf(s, "mmc%d:\n ctx_loss:\t%d\n\nregs:\n",
-			mmc->index, host->context_loss);
+	seq_printf(s, "mmc%d:\n", mmc->index);
+	seq_printf(s, "sdio irq mode\t%s\n",
+		   (mmc->caps & MMC_CAP_SDIO_IRQ) ? "interrupt" : "polling");
+
+	if (mmc->caps & MMC_CAP_SDIO_IRQ) {
+		seq_printf(s, "sdio irq \t%s\n",
+			   (host->flags & HSMMC_SDIO_IRQ_ENABLED) ?  "enabled"
+			   : "disabled");
+	}
+	seq_printf(s, "ctx_loss:\t%d\n", host->context_loss);
 
 	pm_runtime_get_sync(host->dev);
-
+	seq_puts(s, "\nregs:\n");
 	seq_printf(s, "CON:\t\t0x%08x\n",
 			OMAP_HSMMC_READ(host->base, CON));
+	seq_printf(s, "PSTATE:\t\t0x%08x\n",
+		   OMAP_HSMMC_READ(host->base, PSTATE));
 	seq_printf(s, "HCTL:\t\t0x%08x\n",
 			OMAP_HSMMC_READ(host->base, HCTL));
 	seq_printf(s, "SYSCTL:\t\t0x%08x\n",
@@ -1761,6 +1919,10 @@
 static const struct omap_mmc_of_data omap4_mmc_of_data = {
 	.reg_offset = 0x100,
 };
+static const struct omap_mmc_of_data am33xx_mmc_of_data = {
+	.reg_offset = 0x100,
+	.controller_flags = OMAP_HSMMC_SWAKEUP_MISSING,
+};
 
 static const struct of_device_id omap_mmc_of_match[] = {
 	{
@@ -1777,6 +1939,10 @@
 		.compatible = "ti,omap4-hsmmc",
 		.data = &omap4_mmc_of_data,
 	},
+	{
+		.compatible = "ti,am33xx-hsmmc",
+		.data = &am33xx_mmc_of_data,
+	},
 	{},
 };
 MODULE_DEVICE_TABLE(of, omap_mmc_of_match);
@@ -1850,7 +2016,6 @@
 	const struct of_device_id *match;
 	dma_cap_mask_t mask;
 	unsigned tx_req, rx_req;
-	struct pinctrl *pinctrl;
 	const struct omap_mmc_of_data *data;
 	void __iomem *base;
 
@@ -1913,6 +2078,9 @@
 
 	platform_set_drvdata(pdev, host);
 
+	if (pdev->dev.of_node)
+		host->wake_irq = irq_of_parse_and_map(pdev->dev.of_node, 1);
+
 	mmc->ops	= &omap_hsmmc_ops;
 
 	mmc->f_min = OMAP_MMC_MIN_CLOCK;
@@ -2061,10 +2229,17 @@
 
 	omap_hsmmc_disable_irq(host);
 
-	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
-	if (IS_ERR(pinctrl))
-		dev_warn(&pdev->dev,
-			"pins are not configured from the driver\n");
+	/*
+	 * For now, only support SDIO interrupt if we have a separate
+	 * wake-up interrupt configured from device tree. This is because
+	 * the wake-up interrupt is needed for idle state and some
+	 * platforms need special quirks. And we don't want to add new
+	 * legacy mux platform init code callbacks any longer as we
+	 * are moving to DT based booting anyways.
+	 */
+	ret = omap_hsmmc_configure_wake_irq(host);
+	if (!ret)
+		mmc->caps |= MMC_CAP_SDIO_IRQ;
 
 	omap_hsmmc_protect_card(host);
 
@@ -2170,11 +2345,18 @@
 	pm_runtime_get_sync(host->dev);
 
 	if (!(host->mmc->pm_flags & MMC_PM_KEEP_POWER)) {
-		omap_hsmmc_disable_irq(host);
+		OMAP_HSMMC_WRITE(host->base, ISE, 0);
+		OMAP_HSMMC_WRITE(host->base, IE, 0);
+		OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
 		OMAP_HSMMC_WRITE(host->base, HCTL,
 				OMAP_HSMMC_READ(host->base, HCTL) & ~SDBP);
 	}
 
+	/* do not wake up due to sdio irq */
+	if ((host->mmc->caps & MMC_CAP_SDIO_IRQ) &&
+	    !(host->mmc->pm_flags & MMC_PM_WAKE_SDIO_IRQ))
+		disable_irq(host->wake_irq);
+
 	if (host->dbclk)
 		clk_disable_unprepare(host->dbclk);
 
@@ -2200,6 +2382,10 @@
 
 	omap_hsmmc_protect_card(host);
 
+	if ((host->mmc->caps & MMC_CAP_SDIO_IRQ) &&
+	    !(host->mmc->pm_flags & MMC_PM_WAKE_SDIO_IRQ))
+		enable_irq(host->wake_irq);
+
 	pm_runtime_mark_last_busy(host->dev);
 	pm_runtime_put_autosuspend(host->dev);
 	return 0;
@@ -2215,22 +2401,77 @@
 static int omap_hsmmc_runtime_suspend(struct device *dev)
 {
 	struct omap_hsmmc_host *host;
+	unsigned long flags;
+	int ret = 0;
 
 	host = platform_get_drvdata(to_platform_device(dev));
 	omap_hsmmc_context_save(host);
 	dev_dbg(dev, "disabled\n");
 
-	return 0;
+	spin_lock_irqsave(&host->irq_lock, flags);
+	if ((host->mmc->caps & MMC_CAP_SDIO_IRQ) &&
+	    (host->flags & HSMMC_SDIO_IRQ_ENABLED)) {
+		/* disable sdio irq handling to prevent race */
+		OMAP_HSMMC_WRITE(host->base, ISE, 0);
+		OMAP_HSMMC_WRITE(host->base, IE, 0);
+
+		if (!(OMAP_HSMMC_READ(host->base, PSTATE) & DLEV_DAT(1))) {
+			/*
+			 * dat1 line low, pending sdio irq
+			 * race condition: possible irq handler running on
+			 * multi-core, abort
+			 */
+			dev_dbg(dev, "pending sdio irq, abort suspend\n");
+			OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
+			OMAP_HSMMC_WRITE(host->base, ISE, CIRQ_EN);
+			OMAP_HSMMC_WRITE(host->base, IE, CIRQ_EN);
+			pm_runtime_mark_last_busy(dev);
+			ret = -EBUSY;
+			goto abort;
+		}
+
+		pinctrl_pm_select_idle_state(dev);
+
+		WARN_ON(host->flags & HSMMC_WAKE_IRQ_ENABLED);
+		enable_irq(host->wake_irq);
+		host->flags |= HSMMC_WAKE_IRQ_ENABLED;
+	} else {
+		pinctrl_pm_select_idle_state(dev);
+	}
+
+abort:
+	spin_unlock_irqrestore(&host->irq_lock, flags);
+	return ret;
 }
 
 static int omap_hsmmc_runtime_resume(struct device *dev)
 {
 	struct omap_hsmmc_host *host;
+	unsigned long flags;
 
 	host = platform_get_drvdata(to_platform_device(dev));
 	omap_hsmmc_context_restore(host);
 	dev_dbg(dev, "enabled\n");
 
+	spin_lock_irqsave(&host->irq_lock, flags);
+	if ((host->mmc->caps & MMC_CAP_SDIO_IRQ) &&
+	    (host->flags & HSMMC_SDIO_IRQ_ENABLED)) {
+		/* sdio irq flag can't change while in runtime suspend */
+		if (host->flags & HSMMC_WAKE_IRQ_ENABLED) {
+			disable_irq_nosync(host->wake_irq);
+			host->flags &= ~HSMMC_WAKE_IRQ_ENABLED;
+		}
+
+		pinctrl_pm_select_default_state(host->dev);
+
+		/* irq lost, if pinmux incorrect */
+		OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
+		OMAP_HSMMC_WRITE(host->base, ISE, CIRQ_EN);
+		OMAP_HSMMC_WRITE(host->base, IE, CIRQ_EN);
+	} else {
+		pinctrl_pm_select_default_state(host->dev);
+	}
+	spin_unlock_irqrestore(&host->irq_lock, flags);
 	return 0;
 }
 
diff --git a/drivers/mmc/host/s3cmci.c b/drivers/mmc/host/s3cmci.c
index f237826..e5516a2 100644
--- a/drivers/mmc/host/s3cmci.c
+++ b/drivers/mmc/host/s3cmci.c
@@ -12,6 +12,7 @@
  */
 
 #include <linux/module.h>
+#include <linux/dmaengine.h>
 #include <linux/dma-mapping.h>
 #include <linux/clk.h>
 #include <linux/mmc/host.h>
@@ -27,6 +28,7 @@
 #include <mach/dma.h>
 #include <mach/gpio-samsung.h>
 
+#include <linux/platform_data/dma-s3c24xx.h>
 #include <linux/platform_data/mmc-s3cmci.h>
 
 #include "s3cmci.h"
@@ -140,10 +142,6 @@
 		dev_dbg(&host->pdev->dev, args);  \
 	} while (0)
 
-static struct s3c2410_dma_client s3cmci_dma_client = {
-	.name		= "s3c-mci",
-};
-
 static void finalize_request(struct s3cmci_host *host);
 static void s3cmci_send_request(struct mmc_host *mmc);
 static void s3cmci_reset(struct s3cmci_host *host);
@@ -256,25 +254,8 @@
 {
 #ifdef CONFIG_MMC_S3C_PIO
 	return false;
-#elif defined(CONFIG_MMC_S3C_DMA)
+#else /* CONFIG_MMC_S3C_DMA */
 	return true;
-#else
-	return host->dodma;
-#endif
-}
-
-/**
- * s3cmci_host_canpio - return true if host has pio code available
- *
- * Return true if the driver has been compiled with the PIO support code
- * available.
- */
-static inline bool s3cmci_host_canpio(void)
-{
-#ifdef CONFIG_MMC_S3C_PIO
-	return true;
-#else
-	return false;
 #endif
 }
 
@@ -841,60 +822,24 @@
 	return IRQ_HANDLED;
 }
 
-static void s3cmci_dma_done_callback(struct s3c2410_dma_chan *dma_ch,
-				     void *buf_id, int size,
-				     enum s3c2410_dma_buffresult result)
+static void s3cmci_dma_done_callback(void *arg)
 {
-	struct s3cmci_host *host = buf_id;
+	struct s3cmci_host *host = arg;
 	unsigned long iflags;
-	u32 mci_csta, mci_dsta, mci_fsta, mci_dcnt;
-
-	mci_csta = readl(host->base + S3C2410_SDICMDSTAT);
-	mci_dsta = readl(host->base + S3C2410_SDIDSTA);
-	mci_fsta = readl(host->base + S3C2410_SDIFSTA);
-	mci_dcnt = readl(host->base + S3C2410_SDIDCNT);
 
 	BUG_ON(!host->mrq);
 	BUG_ON(!host->mrq->data);
-	BUG_ON(!host->dmatogo);
 
 	spin_lock_irqsave(&host->complete_lock, iflags);
 
-	if (result != S3C2410_RES_OK) {
-		dbg(host, dbg_fail, "DMA FAILED: csta=0x%08x dsta=0x%08x "
-			"fsta=0x%08x dcnt:0x%08x result:0x%08x toGo:%u\n",
-			mci_csta, mci_dsta, mci_fsta,
-			mci_dcnt, result, host->dmatogo);
-
-		goto fail_request;
-	}
-
-	host->dmatogo--;
-	if (host->dmatogo) {
-		dbg(host, dbg_dma, "DMA DONE  Size:%i DSTA:[%08x] "
-			"DCNT:[%08x] toGo:%u\n",
-			size, mci_dsta, mci_dcnt, host->dmatogo);
-
-		goto out;
-	}
-
-	dbg(host, dbg_dma, "DMA FINISHED Size:%i DSTA:%08x DCNT:%08x\n",
-		size, mci_dsta, mci_dcnt);
+	dbg(host, dbg_dma, "DMA FINISHED\n");
 
 	host->dma_complete = 1;
 	host->complete_what = COMPLETION_FINALIZE;
 
-out:
 	tasklet_schedule(&host->pio_tasklet);
 	spin_unlock_irqrestore(&host->complete_lock, iflags);
-	return;
 
-fail_request:
-	host->mrq->data->error = -EINVAL;
-	host->complete_what = COMPLETION_FINALIZE;
-	clear_imask(host);
-
-	goto out;
 }
 
 static void finalize_request(struct s3cmci_host *host)
@@ -966,7 +911,7 @@
 	 * DMA channel and the fifo to clear out any garbage. */
 	if (mrq->data->error != 0) {
 		if (s3cmci_host_usedma(host))
-			s3c2410_dma_ctrl(host->dma, S3C2410_DMAOP_FLUSH);
+			dmaengine_terminate_all(host->dma);
 
 		if (host->is2440) {
 			/* Clear failure register and reset fifo. */
@@ -992,29 +937,6 @@
 	mmc_request_done(host->mmc, mrq);
 }
 
-static void s3cmci_dma_setup(struct s3cmci_host *host,
-			     enum dma_data_direction source)
-{
-	static enum dma_data_direction last_source = -1;
-	static int setup_ok;
-
-	if (last_source == source)
-		return;
-
-	last_source = source;
-
-	s3c2410_dma_devconfig(host->dma, source,
-			      host->mem->start + host->sdidata);
-
-	if (!setup_ok) {
-		s3c2410_dma_config(host->dma, 4);
-		s3c2410_dma_set_buffdone_fn(host->dma,
-					    s3cmci_dma_done_callback);
-		s3c2410_dma_setflags(host->dma, S3C2410_DMAF_AUTOSTART);
-		setup_ok = 1;
-	}
-}
-
 static void s3cmci_send_command(struct s3cmci_host *host,
 					struct mmc_command *cmd)
 {
@@ -1162,43 +1084,45 @@
 
 static int s3cmci_prepare_dma(struct s3cmci_host *host, struct mmc_data *data)
 {
-	int dma_len, i;
 	int rw = data->flags & MMC_DATA_WRITE;
+	struct dma_async_tx_descriptor *desc;
+	struct dma_slave_config conf = {
+		.src_addr = host->mem->start + host->sdidata,
+		.dst_addr = host->mem->start + host->sdidata,
+		.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+		.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+	};
 
 	BUG_ON((data->flags & BOTH_DIR) == BOTH_DIR);
 
-	s3cmci_dma_setup(host, rw ? DMA_TO_DEVICE : DMA_FROM_DEVICE);
-	s3c2410_dma_ctrl(host->dma, S3C2410_DMAOP_FLUSH);
+	/* Restore prescaler value */
+	writel(host->prescaler, host->base + S3C2410_SDIPRE);
 
-	dma_len = dma_map_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+	if (!rw)
+		conf.direction = DMA_DEV_TO_MEM;
+	else
+		conf.direction = DMA_MEM_TO_DEV;
+
+	dma_map_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
 			     rw ? DMA_TO_DEVICE : DMA_FROM_DEVICE);
 
-	if (dma_len == 0)
-		return -ENOMEM;
-
-	host->dma_complete = 0;
-	host->dmatogo = dma_len;
-
-	for (i = 0; i < dma_len; i++) {
-		int res;
-
-		dbg(host, dbg_dma, "enqueue %i: %08x@%u\n", i,
-		    sg_dma_address(&data->sg[i]),
-		    sg_dma_len(&data->sg[i]));
-
-		res = s3c2410_dma_enqueue(host->dma, host,
-					  sg_dma_address(&data->sg[i]),
-					  sg_dma_len(&data->sg[i]));
-
-		if (res) {
-			s3c2410_dma_ctrl(host->dma, S3C2410_DMAOP_FLUSH);
-			return -EBUSY;
-		}
-	}
-
-	s3c2410_dma_ctrl(host->dma, S3C2410_DMAOP_START);
+	dmaengine_slave_config(host->dma, &conf);
+	desc = dmaengine_prep_slave_sg(host->dma, data->sg, data->sg_len,
+		conf.direction,
+		DMA_CTRL_ACK | DMA_PREP_INTERRUPT);
+	if (!desc)
+		goto unmap_exit;
+	desc->callback = s3cmci_dma_done_callback;
+	desc->callback_param = host;
+	dmaengine_submit(desc);
+	dma_async_issue_pending(host->dma);
 
 	return 0;
+
+unmap_exit:
+	dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+			     rw ? DMA_TO_DEVICE : DMA_FROM_DEVICE);
+	return -ENOMEM;
 }
 
 static void s3cmci_send_request(struct mmc_host *mmc)
@@ -1676,10 +1600,6 @@
 	host->complete_what 	= COMPLETION_NONE;
 	host->pio_active 	= XFER_NONE;
 
-#ifdef CONFIG_MMC_S3C_PIODMA
-	host->dodma		= host->pdata->use_dma;
-#endif
-
 	host->mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	if (!host->mem) {
 		dev_err(&pdev->dev,
@@ -1765,17 +1685,17 @@
 	/* depending on the dma state, get a dma channel to use. */
 
 	if (s3cmci_host_usedma(host)) {
-		host->dma = s3c2410_dma_request(DMACH_SDI, &s3cmci_dma_client,
-						host);
-		if (host->dma < 0) {
+		dma_cap_mask_t mask;
+
+		dma_cap_zero(mask);
+		dma_cap_set(DMA_SLAVE, mask);
+
+		host->dma = dma_request_slave_channel_compat(mask,
+			s3c24xx_dma_filter, (void *)DMACH_SDI, &pdev->dev, "rx-tx");
+		if (!host->dma) {
 			dev_err(&pdev->dev, "cannot get DMA channel.\n");
-			if (!s3cmci_host_canpio()) {
-				ret = -EBUSY;
-				goto probe_free_gpio_wp;
-			} else {
-				dev_warn(&pdev->dev, "falling back to PIO.\n");
-				host->dodma = 0;
-			}
+			ret = -EBUSY;
+			goto probe_free_gpio_wp;
 		}
 	}
 
@@ -1787,7 +1707,7 @@
 		goto probe_free_dma;
 	}
 
-	ret = clk_enable(host->clk);
+	ret = clk_prepare_enable(host->clk);
 	if (ret) {
 		dev_err(&pdev->dev, "failed to enable clock source.\n");
 		goto clk_free;
@@ -1816,7 +1736,7 @@
 	mmc->max_segs		= 128;
 
 	dbg(host, dbg_debug,
-	    "probe: mode:%s mapped mci_base:%p irq:%u irq_cd:%u dma:%u.\n",
+	    "probe: mode:%s mapped mci_base:%p irq:%u irq_cd:%u dma:%p.\n",
 	    (host->is2440?"2440":""),
 	    host->base, host->irq, host->irq_cd, host->dma);
 
@@ -1845,14 +1765,14 @@
 	s3cmci_cpufreq_deregister(host);
 
  free_dmabuf:
-	clk_disable(host->clk);
+	clk_disable_unprepare(host->clk);
 
  clk_free:
 	clk_put(host->clk);
 
  probe_free_dma:
 	if (s3cmci_host_usedma(host))
-		s3c2410_dma_free(host->dma, &s3cmci_dma_client);
+		dma_release_channel(host->dma);
 
  probe_free_gpio_wp:
 	if (!host->pdata->no_wprotect)
@@ -1897,7 +1817,7 @@
 	s3cmci_debugfs_remove(host);
 	s3cmci_cpufreq_deregister(host);
 	mmc_remove_host(mmc);
-	clk_disable(host->clk);
+	clk_disable_unprepare(host->clk);
 }
 
 static int s3cmci_remove(struct platform_device *pdev)
@@ -1914,7 +1834,7 @@
 	tasklet_disable(&host->pio_tasklet);
 
 	if (s3cmci_host_usedma(host))
-		s3c2410_dma_free(host->dma, &s3cmci_dma_client);
+		dma_release_channel(host->dma);
 
 	free_irq(host->irq, host);
 
diff --git a/drivers/mmc/host/s3cmci.h b/drivers/mmc/host/s3cmci.h
index c76b53d..cc2e46c 100644
--- a/drivers/mmc/host/s3cmci.h
+++ b/drivers/mmc/host/s3cmci.h
@@ -26,7 +26,7 @@
 	void __iomem		*base;
 	int			irq;
 	int			irq_cd;
-	int			dma;
+	struct dma_chan		*dma;
 
 	unsigned long		clk_rate;
 	unsigned long		clk_div;
@@ -36,8 +36,6 @@
 	int			is2440;
 	unsigned		sdiimsk;
 	unsigned		sdidata;
-	int			dodma;
-	int			dmatogo;
 
 	bool			irq_disabled;
 	bool			irq_enabled;
diff --git a/drivers/mmc/host/sdhci-acpi.c b/drivers/mmc/host/sdhci-acpi.c
index 8ce3c28..8c53370 100644
--- a/drivers/mmc/host/sdhci-acpi.c
+++ b/drivers/mmc/host/sdhci-acpi.c
@@ -124,9 +124,11 @@
 
 static const struct sdhci_acpi_slot sdhci_acpi_slot_int_emmc = {
 	.chip    = &sdhci_acpi_chip_int,
-	.caps    = MMC_CAP_8_BIT_DATA | MMC_CAP_NONREMOVABLE | MMC_CAP_HW_RESET,
+	.caps    = MMC_CAP_8_BIT_DATA | MMC_CAP_NONREMOVABLE |
+		   MMC_CAP_HW_RESET | MMC_CAP_1_8V_DDR,
 	.caps2   = MMC_CAP2_HC_ERASE_SZ,
 	.flags   = SDHCI_ACPI_RUNTIME_PM,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
 };
 
 static const struct sdhci_acpi_slot sdhci_acpi_slot_int_sdio = {
diff --git a/drivers/mmc/host/sdhci-msm.c b/drivers/mmc/host/sdhci-msm.c
index 40573a5..1a6661e 100644
--- a/drivers/mmc/host/sdhci-msm.c
+++ b/drivers/mmc/host/sdhci-msm.c
@@ -16,7 +16,6 @@
 
 #include <linux/module.h>
 #include <linux/of_device.h>
-#include <linux/regulator/consumer.h>
 #include <linux/delay.h>
 #include <linux/mmc/mmc.h>
 #include <linux/slab.h>
diff --git a/drivers/mmc/host/sdhci-pci.c b/drivers/mmc/host/sdhci-pci.c
index 52c42fc..c3a1deb 100644
--- a/drivers/mmc/host/sdhci-pci.c
+++ b/drivers/mmc/host/sdhci-pci.c
@@ -103,6 +103,10 @@
 			  SDHCI_QUIRK_BROKEN_TIMEOUT_VAL,
 };
 
+static const struct sdhci_pci_fixes sdhci_intel_qrk = {
+	.quirks		= SDHCI_QUIRK_NO_HISPD_BIT,
+};
+
 static int mrst_hc_probe_slot(struct sdhci_pci_slot *slot)
 {
 	slot->host->mmc->caps |= MMC_CAP_8_BIT_DATA;
@@ -264,7 +268,7 @@
 static int byt_emmc_probe_slot(struct sdhci_pci_slot *slot)
 {
 	slot->host->mmc->caps |= MMC_CAP_8_BIT_DATA | MMC_CAP_NONREMOVABLE |
-				 MMC_CAP_HW_RESET;
+				 MMC_CAP_HW_RESET | MMC_CAP_1_8V_DDR;
 	slot->host->mmc->caps2 |= MMC_CAP2_HC_ERASE_SZ;
 	slot->hw_reset = sdhci_pci_int_hw_reset;
 	return 0;
@@ -279,6 +283,7 @@
 static const struct sdhci_pci_fixes sdhci_intel_byt_emmc = {
 	.allow_runtime_pm = true,
 	.probe_slot	= byt_emmc_probe_slot,
+	.quirks2	= SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
 };
 
 static const struct sdhci_pci_fixes sdhci_intel_byt_sdio = {
@@ -753,6 +758,14 @@
 
 	{
 		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_QRK_SD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_qrk,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
 		.device		= PCI_DEVICE_ID_INTEL_MRST_SD0,
 		.subvendor	= PCI_ANY_ID,
 		.subdevice	= PCI_ANY_ID,
@@ -1130,18 +1143,13 @@
 			goto err_pci_suspend;
 	}
 
-	pci_save_state(pdev);
 	if (pm_flags & MMC_PM_KEEP_POWER) {
-		if (pm_flags & MMC_PM_WAKE_SDIO_IRQ) {
-			pci_pme_active(pdev, true);
-			pci_enable_wake(pdev, PCI_D3hot, 1);
-		}
-		pci_set_power_state(pdev, PCI_D3hot);
-	} else {
-		pci_enable_wake(pdev, PCI_D3hot, 0);
-		pci_disable_device(pdev);
-		pci_set_power_state(pdev, PCI_D3hot);
-	}
+		if (pm_flags & MMC_PM_WAKE_SDIO_IRQ)
+			device_init_wakeup(dev, true);
+		else
+			device_init_wakeup(dev, false);
+	} else
+		device_init_wakeup(dev, false);
 
 	return 0;
 
@@ -1162,12 +1170,6 @@
 	if (!chip)
 		return 0;
 
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-	ret = pci_enable_device(pdev);
-	if (ret)
-		return ret;
-
 	if (chip->fixes && chip->fixes->resume) {
 		ret = chip->fixes->resume(chip);
 		if (ret)
diff --git a/drivers/mmc/host/sdhci-pci.h b/drivers/mmc/host/sdhci-pci.h
index 6d71871..c101477 100644
--- a/drivers/mmc/host/sdhci-pci.h
+++ b/drivers/mmc/host/sdhci-pci.h
@@ -17,6 +17,7 @@
 #define PCI_DEVICE_ID_INTEL_CLV_SDIO2	0x08fb
 #define PCI_DEVICE_ID_INTEL_CLV_EMMC0	0x08e5
 #define PCI_DEVICE_ID_INTEL_CLV_EMMC1	0x08e6
+#define PCI_DEVICE_ID_INTEL_QRK_SD	0x08A7
 
 /*
  * PCI registers
diff --git a/drivers/mmc/host/sdhci-pxav3.c b/drivers/mmc/host/sdhci-pxav3.c
index f4f1289..6f842fb 100644
--- a/drivers/mmc/host/sdhci-pxav3.c
+++ b/drivers/mmc/host/sdhci-pxav3.c
@@ -288,15 +288,13 @@
 	int ret;
 	struct clk *clk;
 
-	pxa = kzalloc(sizeof(struct sdhci_pxa), GFP_KERNEL);
+	pxa = devm_kzalloc(&pdev->dev, sizeof(struct sdhci_pxa), GFP_KERNEL);
 	if (!pxa)
 		return -ENOMEM;
 
 	host = sdhci_pltfm_init(pdev, &sdhci_pxav3_pdata, 0);
-	if (IS_ERR(host)) {
-		kfree(pxa);
+	if (IS_ERR(host))
 		return PTR_ERR(host);
-	}
 
 	if (of_device_is_compatible(np, "marvell,armada-380-sdhci")) {
 		ret = mv_conf_mbus_windows(pdev, mv_mbus_dram_info());
@@ -308,7 +306,7 @@
 	pltfm_host = sdhci_priv(host);
 	pltfm_host->priv = pxa;
 
-	clk = clk_get(dev, NULL);
+	clk = devm_clk_get(dev, NULL);
 	if (IS_ERR(clk)) {
 		dev_err(dev, "failed to get io clock\n");
 		ret = PTR_ERR(clk);
@@ -389,11 +387,9 @@
 	pm_runtime_put_sync(&pdev->dev);
 	pm_runtime_disable(&pdev->dev);
 	clk_disable_unprepare(clk);
-	clk_put(clk);
 err_clk_get:
 err_mbus_win:
 	sdhci_pltfm_free(pdev);
-	kfree(pxa);
 	return ret;
 }
 
@@ -401,17 +397,14 @@
 {
 	struct sdhci_host *host = platform_get_drvdata(pdev);
 	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
-	struct sdhci_pxa *pxa = pltfm_host->priv;
 
 	pm_runtime_get_sync(&pdev->dev);
 	sdhci_remove_host(host, 1);
 	pm_runtime_disable(&pdev->dev);
 
 	clk_disable_unprepare(pltfm_host->clk);
-	clk_put(pltfm_host->clk);
 
 	sdhci_pltfm_free(pdev);
-	kfree(pxa);
 
 	return 0;
 }
diff --git a/drivers/mmc/host/sdhci-st.c b/drivers/mmc/host/sdhci-st.c
new file mode 100644
index 0000000..328f348
--- /dev/null
+++ b/drivers/mmc/host/sdhci-st.c
@@ -0,0 +1,176 @@
+/*
+ * Support for SDHCI on STMicroelectronics SoCs
+ *
+ * Copyright (C) 2014 STMicroelectronics Ltd
+ * Author: Giuseppe Cavallaro <peppe.cavallaro@st.com>
+ * Contributors: Peter Griffin <peter.griffin@linaro.org>
+ *
+ * Based on sdhci-cns3xxx.c
+ *
+ * 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 distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#include <linux/io.h>
+#include <linux/of.h>
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/mmc/host.h>
+
+#include "sdhci-pltfm.h"
+
+static u32 sdhci_st_readl(struct sdhci_host *host, int reg)
+{
+	u32 ret;
+
+	switch (reg) {
+	case SDHCI_CAPABILITIES:
+		ret = readl_relaxed(host->ioaddr + reg);
+		/* Support 3.3V and 1.8V */
+		ret &= ~SDHCI_CAN_VDD_300;
+		break;
+	default:
+		ret = readl_relaxed(host->ioaddr + reg);
+	}
+	return ret;
+}
+
+static const struct sdhci_ops sdhci_st_ops = {
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.read_l = sdhci_st_readl,
+	.reset = sdhci_reset,
+};
+
+static const struct sdhci_pltfm_data sdhci_st_pdata = {
+	.ops = &sdhci_st_ops,
+	.quirks = SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+	    SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+};
+
+
+static int sdhci_st_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct clk *clk;
+	int ret = 0;
+	u16 host_version;
+
+	clk =  devm_clk_get(&pdev->dev, "mmc");
+	if (IS_ERR(clk)) {
+		dev_err(&pdev->dev, "Peripheral clk not found\n");
+		return PTR_ERR(clk);
+	}
+
+	host = sdhci_pltfm_init(pdev, &sdhci_st_pdata, 0);
+	if (IS_ERR(host)) {
+		dev_err(&pdev->dev, "Failed sdhci_pltfm_init\n");
+		return PTR_ERR(host);
+	}
+
+	ret = mmc_of_parse(host->mmc);
+
+	if (ret) {
+		dev_err(&pdev->dev, "Failed mmc_of_parse\n");
+		return ret;
+	}
+
+	clk_prepare_enable(clk);
+
+	pltfm_host = sdhci_priv(host);
+	pltfm_host->clk = clk;
+
+	ret = sdhci_add_host(host);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed sdhci_add_host\n");
+		goto err_out;
+	}
+
+	platform_set_drvdata(pdev, host);
+
+	host_version = readw_relaxed((host->ioaddr + SDHCI_HOST_VERSION));
+
+	dev_info(&pdev->dev, "SDHCI ST Initialised: Host Version: 0x%x Vendor Version 0x%x\n",
+		((host_version & SDHCI_SPEC_VER_MASK) >> SDHCI_SPEC_VER_SHIFT),
+		((host_version & SDHCI_VENDOR_VER_MASK) >>
+		SDHCI_VENDOR_VER_SHIFT));
+
+	return 0;
+
+err_out:
+	clk_disable_unprepare(clk);
+	sdhci_pltfm_free(pdev);
+
+	return ret;
+}
+
+static int sdhci_st_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	clk_disable_unprepare(pltfm_host->clk);
+
+	return sdhci_pltfm_unregister(pdev);
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_st_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	int ret = sdhci_suspend_host(host);
+
+	if (ret)
+		goto out;
+
+	clk_disable_unprepare(pltfm_host->clk);
+out:
+	return ret;
+}
+
+static int sdhci_st_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	clk_prepare_enable(pltfm_host->clk);
+
+	return sdhci_resume_host(host);
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(sdhci_st_pmops, sdhci_st_suspend, sdhci_st_resume);
+
+static const struct of_device_id st_sdhci_match[] = {
+	{ .compatible = "st,sdhci" },
+	{},
+};
+
+MODULE_DEVICE_TABLE(of, st_sdhci_match);
+
+static struct platform_driver sdhci_st_driver = {
+	.probe = sdhci_st_probe,
+	.remove = sdhci_st_remove,
+	.driver = {
+		   .name = "sdhci-st",
+		   .pm = &sdhci_st_pmops,
+		   .of_match_table = of_match_ptr(st_sdhci_match),
+		  },
+};
+
+module_platform_driver(sdhci_st_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for STMicroelectronics SoCs");
+MODULE_AUTHOR("Giuseppe Cavallaro <peppe.cavallaro@st.com>");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:st-sdhci");
diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
index d93a063..33100d1 100644
--- a/drivers/mmc/host/sdhci-tegra.c
+++ b/drivers/mmc/host/sdhci-tegra.c
@@ -26,8 +26,6 @@
 #include <linux/mmc/host.h>
 #include <linux/mmc/slot-gpio.h>
 
-#include <asm/gpio.h>
-
 #include "sdhci-pltfm.h"
 
 /* Tegra SDHOST controller vendor register definitions */
diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
index 47055f3..37b2a9a 100644
--- a/drivers/mmc/host/sdhci.c
+++ b/drivers/mmc/host/sdhci.c
@@ -1223,8 +1223,16 @@
 static void sdhci_set_power(struct sdhci_host *host, unsigned char mode,
 			    unsigned short vdd)
 {
+	struct mmc_host *mmc = host->mmc;
 	u8 pwr = 0;
 
+	if (!IS_ERR(mmc->supply.vmmc)) {
+		spin_unlock_irq(&host->lock);
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, vdd);
+		spin_lock_irq(&host->lock);
+		return;
+	}
+
 	if (mode != MMC_POWER_OFF) {
 		switch (1 << vdd) {
 		case MMC_VDD_165_195:
@@ -1283,12 +1291,6 @@
 		if (host->quirks & SDHCI_QUIRK_DELAY_AFTER_POWER)
 			mdelay(10);
 	}
-
-	if (host->vmmc) {
-		spin_unlock_irq(&host->lock);
-		mmc_regulator_set_ocr(host->mmc, host->vmmc, vdd);
-		spin_lock_irq(&host->lock);
-	}
 }
 
 /*****************************************************************************\
@@ -1440,13 +1442,15 @@
 {
 	unsigned long flags;
 	u8 ctrl;
+	struct mmc_host *mmc = host->mmc;
 
 	spin_lock_irqsave(&host->lock, flags);
 
 	if (host->flags & SDHCI_DEVICE_DEAD) {
 		spin_unlock_irqrestore(&host->lock, flags);
-		if (host->vmmc && ios->power_mode == MMC_POWER_OFF)
-			mmc_regulator_set_ocr(host->mmc, host->vmmc, 0);
+		if (!IS_ERR(mmc->supply.vmmc) &&
+		    ios->power_mode == MMC_POWER_OFF)
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
 		return;
 	}
 
@@ -1530,7 +1534,6 @@
 			host->ops->set_clock(host, host->clock);
 		}
 
-
 		/* Reset SD Clock Enable */
 		clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
 		clk &= ~SDHCI_CLOCK_CARD_EN;
@@ -1707,6 +1710,7 @@
 static int sdhci_do_start_signal_voltage_switch(struct sdhci_host *host,
 						struct mmc_ios *ios)
 {
+	struct mmc_host *mmc = host->mmc;
 	u16 ctrl;
 	int ret;
 
@@ -1725,11 +1729,12 @@
 		ctrl &= ~SDHCI_CTRL_VDD_180;
 		sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
 
-		if (host->vqmmc) {
-			ret = regulator_set_voltage(host->vqmmc, 2700000, 3600000);
+		if (!IS_ERR(mmc->supply.vqmmc)) {
+			ret = regulator_set_voltage(mmc->supply.vqmmc, 2700000,
+						    3600000);
 			if (ret) {
 				pr_warning("%s: Switching to 3.3V signalling voltage "
-						" failed\n", mmc_hostname(host->mmc));
+						" failed\n", mmc_hostname(mmc));
 				return -EIO;
 			}
 		}
@@ -1742,16 +1747,16 @@
 			return 0;
 
 		pr_warning("%s: 3.3V regulator output did not became stable\n",
-				mmc_hostname(host->mmc));
+				mmc_hostname(mmc));
 
 		return -EAGAIN;
 	case MMC_SIGNAL_VOLTAGE_180:
-		if (host->vqmmc) {
-			ret = regulator_set_voltage(host->vqmmc,
+		if (!IS_ERR(mmc->supply.vqmmc)) {
+			ret = regulator_set_voltage(mmc->supply.vqmmc,
 					1700000, 1950000);
 			if (ret) {
 				pr_warning("%s: Switching to 1.8V signalling voltage "
-						" failed\n", mmc_hostname(host->mmc));
+						" failed\n", mmc_hostname(mmc));
 				return -EIO;
 			}
 		}
@@ -1763,24 +1768,22 @@
 		ctrl |= SDHCI_CTRL_VDD_180;
 		sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
 
-		/* Wait for 5ms */
-		usleep_range(5000, 5500);
-
 		/* 1.8V regulator output should be stable within 5 ms */
 		ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
 		if (ctrl & SDHCI_CTRL_VDD_180)
 			return 0;
 
 		pr_warning("%s: 1.8V regulator output did not became stable\n",
-				mmc_hostname(host->mmc));
+				mmc_hostname(mmc));
 
 		return -EAGAIN;
 	case MMC_SIGNAL_VOLTAGE_120:
-		if (host->vqmmc) {
-			ret = regulator_set_voltage(host->vqmmc, 1100000, 1300000);
+		if (!IS_ERR(mmc->supply.vqmmc)) {
+			ret = regulator_set_voltage(mmc->supply.vqmmc, 1100000,
+						    1300000);
 			if (ret) {
 				pr_warning("%s: Switching to 1.2V signalling voltage "
-						" failed\n", mmc_hostname(host->mmc));
+						" failed\n", mmc_hostname(mmc));
 				return -EIO;
 			}
 		}
@@ -2643,7 +2646,6 @@
 int sdhci_runtime_suspend_host(struct sdhci_host *host)
 {
 	unsigned long flags;
-	int ret = 0;
 
 	/* Disable tuning since we are suspending */
 	if (host->flags & SDHCI_USING_RETUNING_TIMER) {
@@ -2663,14 +2665,14 @@
 	host->runtime_suspended = true;
 	spin_unlock_irqrestore(&host->lock, flags);
 
-	return ret;
+	return 0;
 }
 EXPORT_SYMBOL_GPL(sdhci_runtime_suspend_host);
 
 int sdhci_runtime_resume_host(struct sdhci_host *host)
 {
 	unsigned long flags;
-	int ret = 0, host_flags = host->flags;
+	int host_flags = host->flags;
 
 	if (host_flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA)) {
 		if (host->ops->enable_dma)
@@ -2709,7 +2711,7 @@
 
 	spin_unlock_irqrestore(&host->lock, flags);
 
-	return ret;
+	return 0;
 }
 EXPORT_SYMBOL_GPL(sdhci_runtime_resume_host);
 
@@ -2820,12 +2822,12 @@
 		 * (128) and potentially one alignment transfer for
 		 * each of those entries.
 		 */
-		host->adma_desc = dma_alloc_coherent(mmc_dev(host->mmc),
+		host->adma_desc = dma_alloc_coherent(mmc_dev(mmc),
 						     ADMA_SIZE, &host->adma_addr,
 						     GFP_KERNEL);
 		host->align_buffer = kmalloc(128 * 4, GFP_KERNEL);
 		if (!host->adma_desc || !host->align_buffer) {
-			dma_free_coherent(mmc_dev(host->mmc), ADMA_SIZE,
+			dma_free_coherent(mmc_dev(mmc), ADMA_SIZE,
 					  host->adma_desc, host->adma_addr);
 			kfree(host->align_buffer);
 			pr_warning("%s: Unable to allocate ADMA "
@@ -2838,7 +2840,7 @@
 			pr_warning("%s: unable to allocate aligned ADMA descriptor\n",
 				   mmc_hostname(mmc));
 			host->flags &= ~SDHCI_USE_ADMA;
-			dma_free_coherent(mmc_dev(host->mmc), ADMA_SIZE,
+			dma_free_coherent(mmc_dev(mmc), ADMA_SIZE,
 					  host->adma_desc, host->adma_addr);
 			kfree(host->align_buffer);
 			host->adma_desc = NULL;
@@ -2853,7 +2855,7 @@
 	 */
 	if (!(host->flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA))) {
 		host->dma_mask = DMA_BIT_MASK(64);
-		mmc_dev(host->mmc)->dma_mask = &host->dma_mask;
+		mmc_dev(mmc)->dma_mask = &host->dma_mask;
 	}
 
 	if (host->version >= SDHCI_SPEC_300)
@@ -2959,28 +2961,25 @@
 		mmc->caps |= MMC_CAP_SD_HIGHSPEED | MMC_CAP_MMC_HIGHSPEED;
 
 	if ((host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) &&
-	    !(host->mmc->caps & MMC_CAP_NONREMOVABLE))
+	    !(mmc->caps & MMC_CAP_NONREMOVABLE))
 		mmc->caps |= MMC_CAP_NEEDS_POLL;
 
+	/* If there are external regulators, get them */
+	if (mmc_regulator_get_supply(mmc) == -EPROBE_DEFER)
+		return -EPROBE_DEFER;
+
 	/* If vqmmc regulator and no 1.8V signalling, then there's no UHS */
-	host->vqmmc = regulator_get_optional(mmc_dev(mmc), "vqmmc");
-	if (IS_ERR_OR_NULL(host->vqmmc)) {
-		if (PTR_ERR(host->vqmmc) < 0) {
-			pr_info("%s: no vqmmc regulator found\n",
-				mmc_hostname(mmc));
-			host->vqmmc = NULL;
-		}
-	} else {
-		ret = regulator_enable(host->vqmmc);
-		if (!regulator_is_supported_voltage(host->vqmmc, 1700000,
-			1950000))
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		ret = regulator_enable(mmc->supply.vqmmc);
+		if (!regulator_is_supported_voltage(mmc->supply.vqmmc, 1700000,
+						    1950000))
 			caps[1] &= ~(SDHCI_SUPPORT_SDR104 |
 					SDHCI_SUPPORT_SDR50 |
 					SDHCI_SUPPORT_DDR50);
 		if (ret) {
 			pr_warn("%s: Failed to enable vqmmc regulator: %d\n",
 				mmc_hostname(mmc), ret);
-			host->vqmmc = NULL;
+			mmc->supply.vqmmc = NULL;
 		}
 	}
 
@@ -3041,34 +3040,6 @@
 
 	ocr_avail = 0;
 
-	host->vmmc = regulator_get_optional(mmc_dev(mmc), "vmmc");
-	if (IS_ERR_OR_NULL(host->vmmc)) {
-		if (PTR_ERR(host->vmmc) < 0) {
-			pr_info("%s: no vmmc regulator found\n",
-				mmc_hostname(mmc));
-			host->vmmc = NULL;
-		}
-	}
-
-#ifdef CONFIG_REGULATOR
-	/*
-	 * Voltage range check makes sense only if regulator reports
-	 * any voltage value.
-	 */
-	if (host->vmmc && regulator_get_voltage(host->vmmc) > 0) {
-		ret = regulator_is_supported_voltage(host->vmmc, 2700000,
-			3600000);
-		if ((ret <= 0) || (!(caps[0] & SDHCI_CAN_VDD_330)))
-			caps[0] &= ~SDHCI_CAN_VDD_330;
-		if ((ret <= 0) || (!(caps[0] & SDHCI_CAN_VDD_300)))
-			caps[0] &= ~SDHCI_CAN_VDD_300;
-		ret = regulator_is_supported_voltage(host->vmmc, 1700000,
-			1950000);
-		if ((ret <= 0) || (!(caps[0] & SDHCI_CAN_VDD_180)))
-			caps[0] &= ~SDHCI_CAN_VDD_180;
-	}
-#endif /* CONFIG_REGULATOR */
-
 	/*
 	 * According to SD Host Controller spec v3.00, if the Host System
 	 * can afford more than 150mA, Host Driver should set XPC to 1. Also
@@ -3077,8 +3048,8 @@
 	 * value.
 	 */
 	max_current_caps = sdhci_readl(host, SDHCI_MAX_CURRENT);
-	if (!max_current_caps && host->vmmc) {
-		u32 curr = regulator_get_current_limit(host->vmmc);
+	if (!max_current_caps && !IS_ERR(mmc->supply.vmmc)) {
+		u32 curr = regulator_get_current_limit(mmc->supply.vmmc);
 		if (curr > 0) {
 
 			/* convert to SDHCI_MAX_CURRENT format */
@@ -3118,8 +3089,12 @@
 				   SDHCI_MAX_CURRENT_MULTIPLIER;
 	}
 
+	/* If OCR set by external regulators, use it instead */
+	if (mmc->ocr_avail)
+		ocr_avail = mmc->ocr_avail;
+
 	if (host->ocr_mask)
-		ocr_avail = host->ocr_mask;
+		ocr_avail &= host->ocr_mask;
 
 	mmc->ocr_avail = ocr_avail;
 	mmc->ocr_avail_sdio = ocr_avail;
@@ -3273,6 +3248,7 @@
 
 void sdhci_remove_host(struct sdhci_host *host, int dead)
 {
+	struct mmc_host *mmc = host->mmc;
 	unsigned long flags;
 
 	if (dead) {
@@ -3282,7 +3258,7 @@
 
 		if (host->mrq) {
 			pr_err("%s: Controller removed during "
-				" transfer!\n", mmc_hostname(host->mmc));
+				" transfer!\n", mmc_hostname(mmc));
 
 			host->mrq->cmd->error = -ENOMEDIUM;
 			tasklet_schedule(&host->finish_tasklet);
@@ -3293,7 +3269,7 @@
 
 	sdhci_disable_card_detection(host);
 
-	mmc_remove_host(host->mmc);
+	mmc_remove_host(mmc);
 
 #ifdef SDHCI_USE_LEDS_CLASS
 	led_classdev_unregister(&host->led);
@@ -3310,18 +3286,14 @@
 
 	tasklet_kill(&host->finish_tasklet);
 
-	if (host->vmmc) {
-		regulator_disable(host->vmmc);
-		regulator_put(host->vmmc);
-	}
+	if (!IS_ERR(mmc->supply.vmmc))
+		regulator_disable(mmc->supply.vmmc);
 
-	if (host->vqmmc) {
-		regulator_disable(host->vqmmc);
-		regulator_put(host->vqmmc);
-	}
+	if (!IS_ERR(mmc->supply.vqmmc))
+		regulator_disable(mmc->supply.vqmmc);
 
 	if (host->adma_desc)
-		dma_free_coherent(mmc_dev(host->mmc), ADMA_SIZE,
+		dma_free_coherent(mmc_dev(mmc), ADMA_SIZE,
 				  host->adma_desc, host->adma_addr);
 	kfree(host->align_buffer);
 
diff --git a/drivers/mmc/host/sh_mmcif.c b/drivers/mmc/host/sh_mmcif.c
index 656fbba..d11708c 100644
--- a/drivers/mmc/host/sh_mmcif.c
+++ b/drivers/mmc/host/sh_mmcif.c
@@ -386,7 +386,7 @@
 			 struct sh_mmcif_plat_data *pdata,
 			 enum dma_transfer_direction direction)
 {
-	struct dma_slave_config cfg;
+	struct dma_slave_config cfg = { 0, };
 	struct dma_chan *chan;
 	unsigned int slave_id;
 	struct resource *res;
@@ -417,8 +417,15 @@
 	/* In the OF case the driver will get the slave ID from the DT */
 	cfg.slave_id = slave_id;
 	cfg.direction = direction;
-	cfg.dst_addr = res->start + MMCIF_CE_DATA;
-	cfg.src_addr = 0;
+
+	if (direction == DMA_DEV_TO_MEM) {
+		cfg.src_addr = res->start + MMCIF_CE_DATA;
+		cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	} else {
+		cfg.dst_addr = res->start + MMCIF_CE_DATA;
+		cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	}
+
 	ret = dmaengine_slave_config(chan, &cfg);
 	if (ret < 0) {
 		dma_release_channel(chan);
@@ -1378,26 +1385,19 @@
 		dev_err(&pdev->dev, "Get irq error\n");
 		return -ENXIO;
 	}
+
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	if (!res) {
-		dev_err(&pdev->dev, "platform_get_resource error.\n");
-		return -ENXIO;
-	}
-	reg = ioremap(res->start, resource_size(res));
-	if (!reg) {
-		dev_err(&pdev->dev, "ioremap error.\n");
-		return -ENOMEM;
-	}
+	reg = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(reg))
+		return PTR_ERR(reg);
 
 	mmc = mmc_alloc_host(sizeof(struct sh_mmcif_host), &pdev->dev);
-	if (!mmc) {
-		ret = -ENOMEM;
-		goto ealloch;
-	}
+	if (!mmc)
+		return -ENOMEM;
 
 	ret = mmc_of_parse(mmc);
 	if (ret < 0)
-		goto eofparse;
+		goto err_host;
 
 	host		= mmc_priv(mmc);
 	host->mmc	= mmc;
@@ -1427,19 +1427,19 @@
 	pm_runtime_enable(&pdev->dev);
 	host->power = false;
 
-	host->hclk = clk_get(&pdev->dev, NULL);
+	host->hclk = devm_clk_get(&pdev->dev, NULL);
 	if (IS_ERR(host->hclk)) {
 		ret = PTR_ERR(host->hclk);
 		dev_err(&pdev->dev, "cannot get clock: %d\n", ret);
-		goto eclkget;
+		goto err_pm;
 	}
 	ret = sh_mmcif_clk_update(host);
 	if (ret < 0)
-		goto eclkupdate;
+		goto err_pm;
 
 	ret = pm_runtime_resume(&pdev->dev);
 	if (ret < 0)
-		goto eresume;
+		goto err_clk;
 
 	INIT_DELAYED_WORK(&host->timeout_work, mmcif_timeout_work);
 
@@ -1447,65 +1447,55 @@
 	sh_mmcif_writel(host->addr, MMCIF_CE_INT_MASK, MASK_ALL);
 
 	name = irq[1] < 0 ? dev_name(&pdev->dev) : "sh_mmc:error";
-	ret = request_threaded_irq(irq[0], sh_mmcif_intr, sh_mmcif_irqt, 0, name, host);
+	ret = devm_request_threaded_irq(&pdev->dev, irq[0], sh_mmcif_intr,
+					sh_mmcif_irqt, 0, name, host);
 	if (ret) {
 		dev_err(&pdev->dev, "request_irq error (%s)\n", name);
-		goto ereqirq0;
+		goto err_clk;
 	}
 	if (irq[1] >= 0) {
-		ret = request_threaded_irq(irq[1], sh_mmcif_intr, sh_mmcif_irqt,
-					   0, "sh_mmc:int", host);
+		ret = devm_request_threaded_irq(&pdev->dev, irq[1],
+						sh_mmcif_intr, sh_mmcif_irqt,
+						0, "sh_mmc:int", host);
 		if (ret) {
 			dev_err(&pdev->dev, "request_irq error (sh_mmc:int)\n");
-			goto ereqirq1;
+			goto err_clk;
 		}
 	}
 
 	if (pd && pd->use_cd_gpio) {
 		ret = mmc_gpio_request_cd(mmc, pd->cd_gpio, 0);
 		if (ret < 0)
-			goto erqcd;
+			goto err_clk;
 	}
 
 	mutex_init(&host->thread_lock);
 
-	clk_disable_unprepare(host->hclk);
 	ret = mmc_add_host(mmc);
 	if (ret < 0)
-		goto emmcaddh;
+		goto err_clk;
 
 	dev_pm_qos_expose_latency_limit(&pdev->dev, 100);
 
-	dev_info(&pdev->dev, "driver version %s\n", DRIVER_VERSION);
-	dev_dbg(&pdev->dev, "chip ver H'%04x\n",
-		sh_mmcif_readl(host->addr, MMCIF_CE_VERSION) & 0x0000ffff);
+	dev_info(&pdev->dev, "Chip version 0x%04x, clock rate %luMHz\n",
+		 sh_mmcif_readl(host->addr, MMCIF_CE_VERSION) & 0xffff,
+		 clk_get_rate(host->hclk) / 1000000UL);
+
+	clk_disable_unprepare(host->hclk);
 	return ret;
 
-emmcaddh:
-erqcd:
-	if (irq[1] >= 0)
-		free_irq(irq[1], host);
-ereqirq1:
-	free_irq(irq[0], host);
-ereqirq0:
-	pm_runtime_suspend(&pdev->dev);
-eresume:
+err_clk:
 	clk_disable_unprepare(host->hclk);
-eclkupdate:
-	clk_put(host->hclk);
-eclkget:
+err_pm:
 	pm_runtime_disable(&pdev->dev);
-eofparse:
+err_host:
 	mmc_free_host(mmc);
-ealloch:
-	iounmap(reg);
 	return ret;
 }
 
 static int sh_mmcif_remove(struct platform_device *pdev)
 {
 	struct sh_mmcif_host *host = platform_get_drvdata(pdev);
-	int irq[2];
 
 	host->dying = true;
 	clk_prepare_enable(host->hclk);
@@ -1523,16 +1513,6 @@
 	 */
 	cancel_delayed_work_sync(&host->timeout_work);
 
-	if (host->addr)
-		iounmap(host->addr);
-
-	irq[0] = platform_get_irq(pdev, 0);
-	irq[1] = platform_get_irq(pdev, 1);
-
-	free_irq(irq[0], host);
-	if (irq[1] >= 0)
-		free_irq(irq[1], host);
-
 	clk_disable_unprepare(host->hclk);
 	mmc_free_host(host->mmc);
 	pm_runtime_put_sync(&pdev->dev);
diff --git a/drivers/mmc/host/tmio_mmc_dma.c b/drivers/mmc/host/tmio_mmc_dma.c
index 03e7b28..eb8f1d5 100644
--- a/drivers/mmc/host/tmio_mmc_dma.c
+++ b/drivers/mmc/host/tmio_mmc_dma.c
@@ -294,6 +294,7 @@
 			cfg.slave_id = pdata->dma->slave_id_tx;
 		cfg.direction = DMA_MEM_TO_DEV;
 		cfg.dst_addr = res->start + (CTL_SD_DATA_PORT << host->pdata->bus_shift);
+		cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
 		cfg.src_addr = 0;
 		ret = dmaengine_slave_config(host->chan_tx, &cfg);
 		if (ret < 0)
@@ -312,6 +313,7 @@
 			cfg.slave_id = pdata->dma->slave_id_rx;
 		cfg.direction = DMA_DEV_TO_MEM;
 		cfg.src_addr = cfg.dst_addr;
+		cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
 		cfg.dst_addr = 0;
 		ret = dmaengine_slave_config(host->chan_rx, &cfg);
 		if (ret < 0)
diff --git a/drivers/mmc/host/wmt-sdmmc.c b/drivers/mmc/host/wmt-sdmmc.c
index 282891a..54181b4 100644
--- a/drivers/mmc/host/wmt-sdmmc.c
+++ b/drivers/mmc/host/wmt-sdmmc.c
@@ -72,7 +72,6 @@
 #define BM_SPI_CS			0x20
 #define BM_SD_POWER			0x40
 #define BM_SOFT_RESET			0x80
-#define BM_ONEBIT_MASK			0xFD
 
 /* SDMMC_BLKLEN bit fields */
 #define BLKL_CRCERR_ABORT		0x0800
@@ -120,6 +119,8 @@
 #define STS2_DATARSP_BUSY		0x20
 #define STS2_DIS_FORCECLK		0x80
 
+/* SDMMC_EXTCTRL bit fields */
+#define EXT_EIGHTBIT			0x04
 
 /* MMC/SD DMA Controller Registers */
 #define SDDMA_GCR			0x100
@@ -672,7 +673,7 @@
 static void wmt_mci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
 {
 	struct wmt_mci_priv *priv;
-	u32 reg_tmp;
+	u32 busmode, extctrl;
 
 	priv = mmc_priv(mmc);
 
@@ -687,28 +688,26 @@
 	if (ios->clock != 0)
 		clk_set_rate(priv->clk_sdmmc, ios->clock);
 
+	busmode = readb(priv->sdmmc_base + SDMMC_BUSMODE);
+	extctrl = readb(priv->sdmmc_base + SDMMC_EXTCTRL);
+
+	busmode &= ~(BM_EIGHTBIT_MODE | BM_FOURBIT_MODE);
+	extctrl &= ~EXT_EIGHTBIT;
+
 	switch (ios->bus_width) {
 	case MMC_BUS_WIDTH_8:
-		reg_tmp = readb(priv->sdmmc_base + SDMMC_EXTCTRL);
-		writeb(reg_tmp | 0x04, priv->sdmmc_base + SDMMC_EXTCTRL);
+		busmode |= BM_EIGHTBIT_MODE;
+		extctrl |= EXT_EIGHTBIT;
 		break;
 	case MMC_BUS_WIDTH_4:
-		reg_tmp = readb(priv->sdmmc_base + SDMMC_BUSMODE);
-		writeb(reg_tmp | BM_FOURBIT_MODE, priv->sdmmc_base +
-		       SDMMC_BUSMODE);
-
-		reg_tmp = readb(priv->sdmmc_base + SDMMC_EXTCTRL);
-		writeb(reg_tmp & 0xFB, priv->sdmmc_base + SDMMC_EXTCTRL);
+		busmode |= BM_FOURBIT_MODE;
 		break;
 	case MMC_BUS_WIDTH_1:
-		reg_tmp = readb(priv->sdmmc_base + SDMMC_BUSMODE);
-		writeb(reg_tmp & BM_ONEBIT_MASK, priv->sdmmc_base +
-		       SDMMC_BUSMODE);
-
-		reg_tmp = readb(priv->sdmmc_base + SDMMC_EXTCTRL);
-		writeb(reg_tmp & 0xFB, priv->sdmmc_base + SDMMC_EXTCTRL);
 		break;
 	}
+
+	writeb(busmode, priv->sdmmc_base + SDMMC_BUSMODE);
+	writeb(extctrl, priv->sdmmc_base + SDMMC_EXTCTRL);
 }
 
 static int wmt_mci_get_ro(struct mmc_host *mmc)
@@ -830,7 +829,7 @@
 		goto fail3;
 	}
 
-	ret = request_irq(dma_irq, wmt_mci_dma_isr, 32, "sdmmc", priv);
+	ret = request_irq(dma_irq, wmt_mci_dma_isr, 0, "sdmmc", priv);
 	if (ret) {
 		dev_err(&pdev->dev, "Register DMA IRQ fail\n");
 		goto fail4;
diff --git a/drivers/net/ethernet/emulex/benet/be.h b/drivers/net/ethernet/emulex/benet/be.h
index 811f135..43e08d0 100644
--- a/drivers/net/ethernet/emulex/benet/be.h
+++ b/drivers/net/ethernet/emulex/benet/be.h
@@ -897,5 +897,6 @@
  */
 void be_roce_dev_open(struct be_adapter *);
 void be_roce_dev_close(struct be_adapter *);
+void be_roce_dev_shutdown(struct be_adapter *);
 
 #endif				/* BE_H */
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index db4ff14..9cdeda5 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -5014,6 +5014,7 @@
 	if (!adapter)
 		return;
 
+	be_roce_dev_shutdown(adapter);
 	cancel_delayed_work_sync(&adapter->work);
 	cancel_delayed_work_sync(&adapter->func_recovery_work);
 
diff --git a/drivers/net/ethernet/emulex/benet/be_roce.c b/drivers/net/ethernet/emulex/benet/be_roce.c
index 5bf1660..ef4672d 100644
--- a/drivers/net/ethernet/emulex/benet/be_roce.c
+++ b/drivers/net/ethernet/emulex/benet/be_roce.c
@@ -120,7 +120,8 @@
 {
 	if (ocrdma_drv && adapter->ocrdma_dev &&
 	    ocrdma_drv->state_change_handler)
-		ocrdma_drv->state_change_handler(adapter->ocrdma_dev, 0);
+		ocrdma_drv->state_change_handler(adapter->ocrdma_dev,
+						 BE_DEV_UP);
 }
 
 void be_roce_dev_open(struct be_adapter *adapter)
@@ -136,7 +137,8 @@
 {
 	if (ocrdma_drv && adapter->ocrdma_dev &&
 	    ocrdma_drv->state_change_handler)
-		ocrdma_drv->state_change_handler(adapter->ocrdma_dev, 1);
+		ocrdma_drv->state_change_handler(adapter->ocrdma_dev,
+						 BE_DEV_DOWN);
 }
 
 void be_roce_dev_close(struct be_adapter *adapter)
@@ -148,6 +150,18 @@
 	}
 }
 
+void be_roce_dev_shutdown(struct be_adapter *adapter)
+{
+	if (be_roce_supported(adapter)) {
+		mutex_lock(&be_adapter_list_lock);
+		if (ocrdma_drv && adapter->ocrdma_dev &&
+		    ocrdma_drv->state_change_handler)
+			ocrdma_drv->state_change_handler(adapter->ocrdma_dev,
+							 BE_DEV_SHUTDOWN);
+		mutex_unlock(&be_adapter_list_lock);
+	}
+}
+
 int be_roce_register_driver(struct ocrdma_driver *drv)
 {
 	struct be_adapter *dev;
diff --git a/drivers/net/ethernet/emulex/benet/be_roce.h b/drivers/net/ethernet/emulex/benet/be_roce.h
index a3d9e96..e6f7eb1 100644
--- a/drivers/net/ethernet/emulex/benet/be_roce.h
+++ b/drivers/net/ethernet/emulex/benet/be_roce.h
@@ -62,7 +62,8 @@
 
 enum {
 	BE_DEV_UP	= 0,
-	BE_DEV_DOWN	= 1
+	BE_DEV_DOWN	= 1,
+	BE_DEV_SHUTDOWN = 2
 };
 
 /* APIs for RoCE driver to register callback handlers,
diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index 5d940a2..65a4a0f 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -1311,6 +1311,15 @@
 		.wrapper = mlx4_MAD_IFC_wrapper
 	},
 	{
+		.opcode = MLX4_CMD_MAD_DEMUX,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CMD_EPERM_wrapper
+	},
+	{
 		.opcode = MLX4_CMD_QUERY_IF_STAT,
 		.has_inbox = false,
 		.has_outbox = true,
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
index 688e1ea..494753e 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -136,7 +136,8 @@
 		[7] = "FSM (MAC anti-spoofing) support",
 		[8] = "Dynamic QP updates support",
 		[9] = "Device managed flow steering IPoIB support",
-		[10] = "TCP/IP offloads/flow-steering for VXLAN support"
+		[10] = "TCP/IP offloads/flow-steering for VXLAN support",
+		[11] = "MAD DEMUX (Secure-Host) support"
 	};
 	int i;
 
@@ -571,6 +572,7 @@
 #define QUERY_DEV_CAP_MAX_ICM_SZ_OFFSET		0xa0
 #define QUERY_DEV_CAP_FW_REASSIGN_MAC		0x9d
 #define QUERY_DEV_CAP_VXLAN			0x9e
+#define QUERY_DEV_CAP_MAD_DEMUX_OFFSET		0xb0
 
 	dev_cap->flags2 = 0;
 	mailbox = mlx4_alloc_cmd_mailbox(dev);
@@ -748,6 +750,11 @@
 		MLX4_GET(dev_cap->max_counters, outbox,
 			 QUERY_DEV_CAP_MAX_COUNTERS_OFFSET);
 
+	MLX4_GET(field32, outbox,
+		 QUERY_DEV_CAP_MAD_DEMUX_OFFSET);
+	if (field32 & (1 << 0))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_MAD_DEMUX;
+
 	MLX4_GET(field32, outbox, QUERY_DEV_CAP_EXT_2_FLAGS_OFFSET);
 	if (field32 & (1 << 16))
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_UPDATE_QP;
@@ -2016,3 +2023,85 @@
 out:
 	mlx4_free_cmd_mailbox(dev, mailbox);
 }
+
+static int mlx4_check_smp_firewall_active(struct mlx4_dev *dev,
+					  struct mlx4_cmd_mailbox *mailbox)
+{
+#define MLX4_CMD_MAD_DEMUX_SET_ATTR_OFFSET		0x10
+#define MLX4_CMD_MAD_DEMUX_GETRESP_ATTR_OFFSET		0x20
+#define MLX4_CMD_MAD_DEMUX_TRAP_ATTR_OFFSET		0x40
+#define MLX4_CMD_MAD_DEMUX_TRAP_REPRESS_ATTR_OFFSET	0x70
+
+	u32 set_attr_mask, getresp_attr_mask;
+	u32 trap_attr_mask, traprepress_attr_mask;
+
+	MLX4_GET(set_attr_mask, mailbox->buf,
+		 MLX4_CMD_MAD_DEMUX_SET_ATTR_OFFSET);
+	mlx4_dbg(dev, "SMP firewall set_attribute_mask = 0x%x\n",
+		 set_attr_mask);
+
+	MLX4_GET(getresp_attr_mask, mailbox->buf,
+		 MLX4_CMD_MAD_DEMUX_GETRESP_ATTR_OFFSET);
+	mlx4_dbg(dev, "SMP firewall getresp_attribute_mask = 0x%x\n",
+		 getresp_attr_mask);
+
+	MLX4_GET(trap_attr_mask, mailbox->buf,
+		 MLX4_CMD_MAD_DEMUX_TRAP_ATTR_OFFSET);
+	mlx4_dbg(dev, "SMP firewall trap_attribute_mask = 0x%x\n",
+		 trap_attr_mask);
+
+	MLX4_GET(traprepress_attr_mask, mailbox->buf,
+		 MLX4_CMD_MAD_DEMUX_TRAP_REPRESS_ATTR_OFFSET);
+	mlx4_dbg(dev, "SMP firewall traprepress_attribute_mask = 0x%x\n",
+		 traprepress_attr_mask);
+
+	if (set_attr_mask && getresp_attr_mask && trap_attr_mask &&
+	    traprepress_attr_mask)
+		return 1;
+
+	return 0;
+}
+
+int mlx4_config_mad_demux(struct mlx4_dev *dev)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	int secure_host_active;
+	int err;
+
+	/* Check if mad_demux is supported */
+	if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_MAD_DEMUX))
+		return 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		mlx4_warn(dev, "Failed to allocate mailbox for cmd MAD_DEMUX");
+		return -ENOMEM;
+	}
+
+	/* Query mad_demux to find out which MADs are handled by internal sma */
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0x01 /* subn mgmt class */,
+			   MLX4_CMD_MAD_DEMUX_QUERY_RESTR, MLX4_CMD_MAD_DEMUX,
+			   MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+	if (err) {
+		mlx4_warn(dev, "MLX4_CMD_MAD_DEMUX: query restrictions failed (%d)\n",
+			  err);
+		goto out;
+	}
+
+	secure_host_active = mlx4_check_smp_firewall_active(dev, mailbox);
+
+	/* Config mad_demux to handle all MADs returned by the query above */
+	err = mlx4_cmd(dev, mailbox->dma, 0x01 /* subn mgmt class */,
+		       MLX4_CMD_MAD_DEMUX_CONFIG, MLX4_CMD_MAD_DEMUX,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+	if (err) {
+		mlx4_warn(dev, "MLX4_CMD_MAD_DEMUX: configure failed (%d)\n", err);
+		goto out;
+	}
+
+	if (secure_host_active)
+		mlx4_warn(dev, "HCA operating in secure-host mode. SMP firewall activated.\n");
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index 80b8c5f..0158689 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -1853,6 +1853,11 @@
 			mlx4_err(dev, "Failed to initialize multicast group table, aborting\n");
 			goto err_mr_table_free;
 		}
+		err = mlx4_config_mad_demux(dev);
+		if (err) {
+			mlx4_err(dev, "Failed in config_mad_demux, aborting\n");
+			goto err_mcg_table_free;
+		}
 	}
 
 	err = mlx4_init_eq_table(dev);
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index 13fbcd0..b508c78 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -274,6 +274,8 @@
 #define MLX4_MPT_FLAG_PHYSICAL	    (1 <<  9)
 #define MLX4_MPT_FLAG_REGION	    (1 <<  8)
 
+#define MLX4_MPT_PD_MASK	    (0x1FFFFUL)
+#define MLX4_MPT_PD_VF_MASK	    (0xFE0000UL)
 #define MLX4_MPT_PD_FLAG_FAST_REG   (1 << 27)
 #define MLX4_MPT_PD_FLAG_RAE	    (1 << 28)
 #define MLX4_MPT_PD_FLAG_EN_INV	    (3 << 24)
@@ -1306,5 +1308,6 @@
 int mlx4_get_slave_num_gids(struct mlx4_dev *dev, int slave, int port);
 /* Returns the VF index of slave */
 int mlx4_get_vf_indx(struct mlx4_dev *dev, int slave);
+int mlx4_config_mad_demux(struct mlx4_dev *dev);
 
 #endif /* MLX4_H */
diff --git a/drivers/net/ethernet/mellanox/mlx4/mr.c b/drivers/net/ethernet/mellanox/mlx4/mr.c
index 2839abb..7d717ec 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mr.c
+++ b/drivers/net/ethernet/mellanox/mlx4/mr.c
@@ -298,6 +298,131 @@
 			    MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
 }
 
+int mlx4_mr_hw_get_mpt(struct mlx4_dev *dev, struct mlx4_mr *mmr,
+		       struct mlx4_mpt_entry ***mpt_entry)
+{
+	int err;
+	int key = key_to_hw_index(mmr->key) & (dev->caps.num_mpts - 1);
+	struct mlx4_cmd_mailbox *mailbox = NULL;
+
+	/* Make sure that at this point we have single-threaded access only */
+
+	if (mmr->enabled != MLX4_MPT_EN_HW)
+		return -EINVAL;
+
+	err = mlx4_HW2SW_MPT(dev, NULL, key);
+
+	if (err) {
+		mlx4_warn(dev, "HW2SW_MPT failed (%d).", err);
+		mlx4_warn(dev, "Most likely the MR has MWs bound to it.\n");
+		return err;
+	}
+
+	mmr->enabled = MLX4_MPT_EN_SW;
+
+	if (!mlx4_is_mfunc(dev)) {
+		**mpt_entry = mlx4_table_find(
+				&mlx4_priv(dev)->mr_table.dmpt_table,
+				key, NULL);
+	} else {
+		mailbox = mlx4_alloc_cmd_mailbox(dev);
+		if (IS_ERR_OR_NULL(mailbox))
+			return PTR_ERR(mailbox);
+
+		err = mlx4_cmd_box(dev, 0, mailbox->dma, key,
+				   0, MLX4_CMD_QUERY_MPT,
+				   MLX4_CMD_TIME_CLASS_B,
+				   MLX4_CMD_WRAPPED);
+
+		if (err)
+			goto free_mailbox;
+
+		*mpt_entry = (struct mlx4_mpt_entry **)&mailbox->buf;
+	}
+
+	if (!(*mpt_entry) || !(**mpt_entry)) {
+		err = -ENOMEM;
+		goto free_mailbox;
+	}
+
+	return 0;
+
+free_mailbox:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_hw_get_mpt);
+
+int mlx4_mr_hw_write_mpt(struct mlx4_dev *dev, struct mlx4_mr *mmr,
+			 struct mlx4_mpt_entry **mpt_entry)
+{
+	int err;
+
+	if (!mlx4_is_mfunc(dev)) {
+		/* Make sure any changes to this entry are flushed */
+		wmb();
+
+		*(u8 *)(*mpt_entry) = MLX4_MPT_STATUS_HW;
+
+		/* Make sure the new status is written */
+		wmb();
+
+		err = mlx4_SYNC_TPT(dev);
+	} else {
+		int key = key_to_hw_index(mmr->key) & (dev->caps.num_mpts - 1);
+
+		struct mlx4_cmd_mailbox *mailbox =
+			container_of((void *)mpt_entry, struct mlx4_cmd_mailbox,
+				     buf);
+
+		err = mlx4_SW2HW_MPT(dev, mailbox, key);
+	}
+
+	mmr->pd = be32_to_cpu((*mpt_entry)->pd_flags) & MLX4_MPT_PD_MASK;
+	if (!err)
+		mmr->enabled = MLX4_MPT_EN_HW;
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_hw_write_mpt);
+
+void mlx4_mr_hw_put_mpt(struct mlx4_dev *dev,
+			struct mlx4_mpt_entry **mpt_entry)
+{
+	if (mlx4_is_mfunc(dev)) {
+		struct mlx4_cmd_mailbox *mailbox =
+			container_of((void *)mpt_entry, struct mlx4_cmd_mailbox,
+				     buf);
+		mlx4_free_cmd_mailbox(dev, mailbox);
+	}
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_hw_put_mpt);
+
+int mlx4_mr_hw_change_pd(struct mlx4_dev *dev, struct mlx4_mpt_entry *mpt_entry,
+			 u32 pdn)
+{
+	u32 pd_flags = be32_to_cpu(mpt_entry->pd_flags);
+	/* The wrapper function will put the slave's id here */
+	if (mlx4_is_mfunc(dev))
+		pd_flags &= ~MLX4_MPT_PD_VF_MASK;
+	mpt_entry->pd_flags = cpu_to_be32((pd_flags &  ~MLX4_MPT_PD_MASK) |
+					  (pdn & MLX4_MPT_PD_MASK)
+					  | MLX4_MPT_PD_FLAG_EN_INV);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_hw_change_pd);
+
+int mlx4_mr_hw_change_access(struct mlx4_dev *dev,
+			     struct mlx4_mpt_entry *mpt_entry,
+			     u32 access)
+{
+	u32 flags = (be32_to_cpu(mpt_entry->flags) & ~MLX4_PERM_MASK) |
+		    (access & MLX4_PERM_MASK);
+
+	mpt_entry->flags = cpu_to_be32(flags);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_hw_change_access);
+
 static int mlx4_mr_alloc_reserved(struct mlx4_dev *dev, u32 mridx, u32 pd,
 			   u64 iova, u64 size, u32 access, int npages,
 			   int page_shift, struct mlx4_mr *mr)
@@ -463,6 +588,41 @@
 }
 EXPORT_SYMBOL_GPL(mlx4_mr_free);
 
+void mlx4_mr_rereg_mem_cleanup(struct mlx4_dev *dev, struct mlx4_mr *mr)
+{
+	mlx4_mtt_cleanup(dev, &mr->mtt);
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_rereg_mem_cleanup);
+
+int mlx4_mr_rereg_mem_write(struct mlx4_dev *dev, struct mlx4_mr *mr,
+			    u64 iova, u64 size, int npages,
+			    int page_shift, struct mlx4_mpt_entry *mpt_entry)
+{
+	int err;
+
+	mpt_entry->start       = cpu_to_be64(mr->iova);
+	mpt_entry->length      = cpu_to_be64(mr->size);
+	mpt_entry->entity_size = cpu_to_be32(mr->mtt.page_shift);
+
+	err = mlx4_mtt_init(dev, npages, page_shift, &mr->mtt);
+	if (err)
+		return err;
+
+	if (mr->mtt.order < 0) {
+		mpt_entry->flags |= cpu_to_be32(MLX4_MPT_FLAG_PHYSICAL);
+		mpt_entry->mtt_addr = 0;
+	} else {
+		mpt_entry->mtt_addr = cpu_to_be64(mlx4_mtt_addr(dev,
+						  &mr->mtt));
+		if (mr->mtt.page_shift == 0)
+			mpt_entry->mtt_sz    = cpu_to_be32(1 << mr->mtt.order);
+	}
+	mr->enabled = MLX4_MPT_EN_SW;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_rereg_mem_write);
+
 int mlx4_mr_enable(struct mlx4_dev *dev, struct mlx4_mr *mr)
 {
 	struct mlx4_cmd_mailbox *mailbox;
diff --git a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
index 0efc136..1089367 100644
--- a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
+++ b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
@@ -2613,12 +2613,34 @@
 	if (err)
 		return err;
 
-	if (mpt->com.from_state != RES_MPT_HW) {
+	if (mpt->com.from_state == RES_MPT_MAPPED) {
+		/* In order to allow rereg in SRIOV, we need to alter the MPT entry. To do
+		 * that, the VF must read the MPT. But since the MPT entry memory is not
+		 * in the VF's virtual memory space, it must use QUERY_MPT to obtain the
+		 * entry contents. To guarantee that the MPT cannot be changed, the driver
+		 * must perform HW2SW_MPT before this query and return the MPT entry to HW
+		 * ownership fofollowing the change. The change here allows the VF to
+		 * perform QUERY_MPT also when the entry is in SW ownership.
+		 */
+		struct mlx4_mpt_entry *mpt_entry = mlx4_table_find(
+					&mlx4_priv(dev)->mr_table.dmpt_table,
+					mpt->key, NULL);
+
+		if (NULL == mpt_entry || NULL == outbox->buf) {
+			err = -EINVAL;
+			goto out;
+		}
+
+		memcpy(outbox->buf, mpt_entry, sizeof(*mpt_entry));
+
+		err = 0;
+	} else if (mpt->com.from_state == RES_MPT_HW) {
+		err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	} else {
 		err = -EBUSY;
 		goto out;
 	}
 
-	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
 
 out:
 	put_res(dev, slave, id, RES_MPT);
diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig
index 2dcb054..5160c4e 100644
--- a/drivers/of/Kconfig
+++ b/drivers/of/Kconfig
@@ -9,7 +9,8 @@
 
 config OF_SELFTEST
 	bool "Device Tree Runtime self tests"
-	depends on OF_IRQ
+	depends on OF_IRQ && OF_EARLY_FLATTREE
+	select OF_DYNAMIC
 	help
 	  This option builds in test cases for the device tree infrastructure
 	  that are executed once at boot time, and the results dumped to the
diff --git a/drivers/of/Makefile b/drivers/of/Makefile
index 099b1fb..2b6a7b1 100644
--- a/drivers/of/Makefile
+++ b/drivers/of/Makefile
@@ -1,11 +1,13 @@
 obj-y = base.o device.o platform.o
+obj-$(CONFIG_OF_DYNAMIC) += dynamic.o
 obj-$(CONFIG_OF_FLATTREE) += fdt.o
 obj-$(CONFIG_OF_EARLY_FLATTREE) += fdt_address.o
 obj-$(CONFIG_OF_PROMTREE) += pdt.o
 obj-$(CONFIG_OF_ADDRESS)  += address.o
 obj-$(CONFIG_OF_IRQ)    += irq.o
 obj-$(CONFIG_OF_NET)	+= of_net.o
-obj-$(CONFIG_OF_SELFTEST) += selftest.o
+obj-$(CONFIG_OF_SELFTEST) += of_selftest.o
+of_selftest-objs := selftest.o testcase-data/testcases.dtb.o
 obj-$(CONFIG_OF_MDIO)	+= of_mdio.o
 obj-$(CONFIG_OF_PCI)	+= of_pci.o
 obj-$(CONFIG_OF_PCI_IRQ)  += of_pci_irq.o
diff --git a/drivers/of/base.c b/drivers/of/base.c
index b986480..d8574ad 100644
--- a/drivers/of/base.c
+++ b/drivers/of/base.c
@@ -17,6 +17,7 @@
  *      as published by the Free Software Foundation; either version
  *      2 of the License, or (at your option) any later version.
  */
+#include <linux/console.h>
 #include <linux/ctype.h>
 #include <linux/cpu.h>
 #include <linux/module.h>
@@ -35,15 +36,17 @@
 EXPORT_SYMBOL(of_allnodes);
 struct device_node *of_chosen;
 struct device_node *of_aliases;
-static struct device_node *of_stdout;
+struct device_node *of_stdout;
 
-static struct kset *of_kset;
+struct kset *of_kset;
 
 /*
- * Used to protect the of_aliases; but also overloaded to hold off addition of
- * nodes to sysfs
+ * Used to protect the of_aliases, to hold off addition of nodes to sysfs.
+ * This mutex must be held whenever modifications are being made to the
+ * device tree. The of_{attach,detach}_node() and
+ * of_{add,remove,update}_property() helpers make sure this happens.
  */
-DEFINE_MUTEX(of_aliases_mutex);
+DEFINE_MUTEX(of_mutex);
 
 /* use when traversing tree through the allnext, child, sibling,
  * or parent members of struct device_node.
@@ -89,79 +92,7 @@
 }
 #endif
 
-#if defined(CONFIG_OF_DYNAMIC)
-/**
- *	of_node_get - Increment refcount of a node
- *	@node:	Node to inc refcount, NULL is supported to
- *		simplify writing of callers
- *
- *	Returns node.
- */
-struct device_node *of_node_get(struct device_node *node)
-{
-	if (node)
-		kobject_get(&node->kobj);
-	return node;
-}
-EXPORT_SYMBOL(of_node_get);
-
-static inline struct device_node *kobj_to_device_node(struct kobject *kobj)
-{
-	return container_of(kobj, struct device_node, kobj);
-}
-
-/**
- *	of_node_release - release a dynamically allocated node
- *	@kref:  kref element of the node to be released
- *
- *	In of_node_put() this function is passed to kref_put()
- *	as the destructor.
- */
-static void of_node_release(struct kobject *kobj)
-{
-	struct device_node *node = kobj_to_device_node(kobj);
-	struct property *prop = node->properties;
-
-	/* We should never be releasing nodes that haven't been detached. */
-	if (!of_node_check_flag(node, OF_DETACHED)) {
-		pr_err("ERROR: Bad of_node_put() on %s\n", node->full_name);
-		dump_stack();
-		return;
-	}
-
-	if (!of_node_check_flag(node, OF_DYNAMIC))
-		return;
-
-	while (prop) {
-		struct property *next = prop->next;
-		kfree(prop->name);
-		kfree(prop->value);
-		kfree(prop);
-		prop = next;
-
-		if (!prop) {
-			prop = node->deadprops;
-			node->deadprops = NULL;
-		}
-	}
-	kfree(node->full_name);
-	kfree(node->data);
-	kfree(node);
-}
-
-/**
- *	of_node_put - Decrement refcount of a node
- *	@node:	Node to dec refcount, NULL is supported to
- *		simplify writing of callers
- *
- */
-void of_node_put(struct device_node *node)
-{
-	if (node)
-		kobject_put(&node->kobj);
-}
-EXPORT_SYMBOL(of_node_put);
-#else
+#ifndef CONFIG_OF_DYNAMIC
 static void of_node_release(struct kobject *kobj)
 {
 	/* Without CONFIG_OF_DYNAMIC, no nodes gets freed */
@@ -200,13 +131,16 @@
 	return name;
 }
 
-static int __of_add_property_sysfs(struct device_node *np, struct property *pp)
+int __of_add_property_sysfs(struct device_node *np, struct property *pp)
 {
 	int rc;
 
 	/* Important: Don't leak passwords */
 	bool secure = strncmp(pp->name, "security-", 9) == 0;
 
+	if (!of_kset || !of_node_is_attached(np))
+		return 0;
+
 	sysfs_bin_attr_init(&pp->attr);
 	pp->attr.attr.name = safe_name(&np->kobj, pp->name);
 	pp->attr.attr.mode = secure ? S_IRUSR : S_IRUGO;
@@ -218,12 +152,15 @@
 	return rc;
 }
 
-static int __of_node_add(struct device_node *np)
+int __of_attach_node_sysfs(struct device_node *np)
 {
 	const char *name;
 	struct property *pp;
 	int rc;
 
+	if (!of_kset)
+		return 0;
+
 	np->kobj.kset = of_kset;
 	if (!np->parent) {
 		/* Nodes without parents are new top level trees */
@@ -245,59 +182,20 @@
 	return 0;
 }
 
-int of_node_add(struct device_node *np)
-{
-	int rc = 0;
-
-	BUG_ON(!of_node_is_initialized(np));
-
-	/*
-	 * Grab the mutex here so that in a race condition between of_init() and
-	 * of_node_add(), node addition will still be consistent.
-	 */
-	mutex_lock(&of_aliases_mutex);
-	if (of_kset)
-		rc = __of_node_add(np);
-	else
-		/* This scenario may be perfectly valid, but report it anyway */
-		pr_info("of_node_add(%s) before of_init()\n", np->full_name);
-	mutex_unlock(&of_aliases_mutex);
-	return rc;
-}
-
-#if defined(CONFIG_OF_DYNAMIC)
-static void of_node_remove(struct device_node *np)
-{
-	struct property *pp;
-
-	BUG_ON(!of_node_is_initialized(np));
-
-	/* only remove properties if on sysfs */
-	if (of_node_is_attached(np)) {
-		for_each_property_of_node(np, pp)
-			sysfs_remove_bin_file(&np->kobj, &pp->attr);
-		kobject_del(&np->kobj);
-	}
-
-	/* finally remove the kobj_init ref */
-	of_node_put(np);
-}
-#endif
-
 static int __init of_init(void)
 {
 	struct device_node *np;
 
 	/* Create the kset, and register existing nodes */
-	mutex_lock(&of_aliases_mutex);
+	mutex_lock(&of_mutex);
 	of_kset = kset_create_and_add("devicetree", NULL, firmware_kobj);
 	if (!of_kset) {
-		mutex_unlock(&of_aliases_mutex);
+		mutex_unlock(&of_mutex);
 		return -ENOMEM;
 	}
 	for_each_of_allnodes(np)
-		__of_node_add(np);
-	mutex_unlock(&of_aliases_mutex);
+		__of_attach_node_sysfs(np);
+	mutex_unlock(&of_mutex);
 
 	/* Symlink in /proc as required by userspace ABI */
 	if (of_allnodes)
@@ -369,8 +267,8 @@
  * Find a property with a given name for a given node
  * and return the value.
  */
-static const void *__of_get_property(const struct device_node *np,
-				     const char *name, int *lenp)
+const void *__of_get_property(const struct device_node *np,
+			      const char *name, int *lenp)
 {
 	struct property *pp = __of_find_property(np, name, lenp);
 
@@ -1748,32 +1646,10 @@
 }
 EXPORT_SYMBOL(of_count_phandle_with_args);
 
-#if defined(CONFIG_OF_DYNAMIC)
-static int of_property_notify(int action, struct device_node *np,
-			      struct property *prop)
-{
-	struct of_prop_reconfig pr;
-
-	/* only call notifiers if the node is attached */
-	if (!of_node_is_attached(np))
-		return 0;
-
-	pr.dn = np;
-	pr.prop = prop;
-	return of_reconfig_notify(action, &pr);
-}
-#else
-static int of_property_notify(int action, struct device_node *np,
-			      struct property *prop)
-{
-	return 0;
-}
-#endif
-
 /**
  * __of_add_property - Add a property to a node without lock operations
  */
-static int __of_add_property(struct device_node *np, struct property *prop)
+int __of_add_property(struct device_node *np, struct property *prop)
 {
 	struct property **next;
 
@@ -1799,22 +1675,49 @@
 	unsigned long flags;
 	int rc;
 
-	rc = of_property_notify(OF_RECONFIG_ADD_PROPERTY, np, prop);
-	if (rc)
-		return rc;
+	mutex_lock(&of_mutex);
 
 	raw_spin_lock_irqsave(&devtree_lock, flags);
 	rc = __of_add_property(np, prop);
 	raw_spin_unlock_irqrestore(&devtree_lock, flags);
-	if (rc)
-		return rc;
 
-	if (of_node_is_attached(np))
+	if (!rc)
 		__of_add_property_sysfs(np, prop);
 
+	mutex_unlock(&of_mutex);
+
+	if (!rc)
+		of_property_notify(OF_RECONFIG_ADD_PROPERTY, np, prop, NULL);
+
 	return rc;
 }
 
+int __of_remove_property(struct device_node *np, struct property *prop)
+{
+	struct property **next;
+
+	for (next = &np->properties; *next; next = &(*next)->next) {
+		if (*next == prop)
+			break;
+	}
+	if (*next == NULL)
+		return -ENODEV;
+
+	/* found the node */
+	*next = prop->next;
+	prop->next = np->deadprops;
+	np->deadprops = prop;
+
+	return 0;
+}
+
+void __of_remove_property_sysfs(struct device_node *np, struct property *prop)
+{
+	/* at early boot, bail here and defer setup to of_init() */
+	if (of_kset && of_node_is_attached(np))
+		sysfs_remove_bin_file(&np->kobj, &prop->attr);
+}
+
 /**
  * of_remove_property - Remove a property from a node.
  *
@@ -1825,42 +1728,64 @@
  */
 int of_remove_property(struct device_node *np, struct property *prop)
 {
-	struct property **next;
 	unsigned long flags;
-	int found = 0;
 	int rc;
 
-	rc = of_property_notify(OF_RECONFIG_REMOVE_PROPERTY, np, prop);
-	if (rc)
-		return rc;
+	mutex_lock(&of_mutex);
 
 	raw_spin_lock_irqsave(&devtree_lock, flags);
-	next = &np->properties;
-	while (*next) {
-		if (*next == prop) {
-			/* found the node */
-			*next = prop->next;
-			prop->next = np->deadprops;
-			np->deadprops = prop;
-			found = 1;
-			break;
-		}
-		next = &(*next)->next;
-	}
+	rc = __of_remove_property(np, prop);
 	raw_spin_unlock_irqrestore(&devtree_lock, flags);
 
-	if (!found)
-		return -ENODEV;
+	if (!rc)
+		__of_remove_property_sysfs(np, prop);
 
-	/* at early boot, bail hear and defer setup to of_init() */
-	if (!of_kset)
-		return 0;
+	mutex_unlock(&of_mutex);
 
-	sysfs_remove_bin_file(&np->kobj, &prop->attr);
+	if (!rc)
+		of_property_notify(OF_RECONFIG_REMOVE_PROPERTY, np, prop, NULL);
+
+	return rc;
+}
+
+int __of_update_property(struct device_node *np, struct property *newprop,
+		struct property **oldpropp)
+{
+	struct property **next, *oldprop;
+
+	for (next = &np->properties; *next; next = &(*next)->next) {
+		if (of_prop_cmp((*next)->name, newprop->name) == 0)
+			break;
+	}
+	*oldpropp = oldprop = *next;
+
+	if (oldprop) {
+		/* replace the node */
+		newprop->next = oldprop->next;
+		*next = newprop;
+		oldprop->next = np->deadprops;
+		np->deadprops = oldprop;
+	} else {
+		/* new node */
+		newprop->next = NULL;
+		*next = newprop;
+	}
 
 	return 0;
 }
 
+void __of_update_property_sysfs(struct device_node *np, struct property *newprop,
+		struct property *oldprop)
+{
+	/* At early boot, bail out and defer setup to of_init() */
+	if (!of_kset)
+		return;
+
+	if (oldprop)
+		sysfs_remove_bin_file(&np->kobj, &oldprop->attr);
+	__of_add_property_sysfs(np, newprop);
+}
+
 /*
  * of_update_property - Update a property in a node, if the property does
  * not exist, add it.
@@ -1872,164 +1797,29 @@
  */
 int of_update_property(struct device_node *np, struct property *newprop)
 {
-	struct property **next, *oldprop;
+	struct property *oldprop;
 	unsigned long flags;
 	int rc;
 
-	rc = of_property_notify(OF_RECONFIG_UPDATE_PROPERTY, np, newprop);
-	if (rc)
-		return rc;
-
 	if (!newprop->name)
 		return -EINVAL;
 
-	raw_spin_lock_irqsave(&devtree_lock, flags);
-	next = &np->properties;
-	oldprop = __of_find_property(np, newprop->name, NULL);
-	if (!oldprop) {
-		/* add the new node */
-		rc = __of_add_property(np, newprop);
-	} else while (*next) {
-		/* replace the node */
-		if (*next == oldprop) {
-			newprop->next = oldprop->next;
-			*next = newprop;
-			oldprop->next = np->deadprops;
-			np->deadprops = oldprop;
-			break;
-		}
-		next = &(*next)->next;
-	}
-	raw_spin_unlock_irqrestore(&devtree_lock, flags);
-	if (rc)
-		return rc;
-
-	/* At early boot, bail out and defer setup to of_init() */
-	if (!of_kset)
-		return 0;
-
-	/* Update the sysfs attribute */
-	if (oldprop)
-		sysfs_remove_bin_file(&np->kobj, &oldprop->attr);
-	__of_add_property_sysfs(np, newprop);
-
-	return 0;
-}
-
-#if defined(CONFIG_OF_DYNAMIC)
-/*
- * Support for dynamic device trees.
- *
- * On some platforms, the device tree can be manipulated at runtime.
- * The routines in this section support adding, removing and changing
- * device tree nodes.
- */
-
-static BLOCKING_NOTIFIER_HEAD(of_reconfig_chain);
-
-int of_reconfig_notifier_register(struct notifier_block *nb)
-{
-	return blocking_notifier_chain_register(&of_reconfig_chain, nb);
-}
-EXPORT_SYMBOL_GPL(of_reconfig_notifier_register);
-
-int of_reconfig_notifier_unregister(struct notifier_block *nb)
-{
-	return blocking_notifier_chain_unregister(&of_reconfig_chain, nb);
-}
-EXPORT_SYMBOL_GPL(of_reconfig_notifier_unregister);
-
-int of_reconfig_notify(unsigned long action, void *p)
-{
-	int rc;
-
-	rc = blocking_notifier_call_chain(&of_reconfig_chain, action, p);
-	return notifier_to_errno(rc);
-}
-
-/**
- * of_attach_node - Plug a device node into the tree and global list.
- */
-int of_attach_node(struct device_node *np)
-{
-	unsigned long flags;
-	int rc;
-
-	rc = of_reconfig_notify(OF_RECONFIG_ATTACH_NODE, np);
-	if (rc)
-		return rc;
+	mutex_lock(&of_mutex);
 
 	raw_spin_lock_irqsave(&devtree_lock, flags);
-	np->sibling = np->parent->child;
-	np->allnext = np->parent->allnext;
-	np->parent->allnext = np;
-	np->parent->child = np;
-	of_node_clear_flag(np, OF_DETACHED);
+	rc = __of_update_property(np, newprop, &oldprop);
 	raw_spin_unlock_irqrestore(&devtree_lock, flags);
 
-	of_node_add(np);
-	return 0;
-}
+	if (!rc)
+		__of_update_property_sysfs(np, newprop, oldprop);
 
-/**
- * of_detach_node - "Unplug" a node from the device tree.
- *
- * The caller must hold a reference to the node.  The memory associated with
- * the node is not freed until its refcount goes to zero.
- */
-int of_detach_node(struct device_node *np)
-{
-	struct device_node *parent;
-	unsigned long flags;
-	int rc = 0;
+	mutex_unlock(&of_mutex);
 
-	rc = of_reconfig_notify(OF_RECONFIG_DETACH_NODE, np);
-	if (rc)
-		return rc;
+	if (!rc)
+		of_property_notify(OF_RECONFIG_UPDATE_PROPERTY, np, newprop, oldprop);
 
-	raw_spin_lock_irqsave(&devtree_lock, flags);
-
-	if (of_node_check_flag(np, OF_DETACHED)) {
-		/* someone already detached it */
-		raw_spin_unlock_irqrestore(&devtree_lock, flags);
-		return rc;
-	}
-
-	parent = np->parent;
-	if (!parent) {
-		raw_spin_unlock_irqrestore(&devtree_lock, flags);
-		return rc;
-	}
-
-	if (of_allnodes == np)
-		of_allnodes = np->allnext;
-	else {
-		struct device_node *prev;
-		for (prev = of_allnodes;
-		     prev->allnext != np;
-		     prev = prev->allnext)
-			;
-		prev->allnext = np->allnext;
-	}
-
-	if (parent->child == np)
-		parent->child = np->sibling;
-	else {
-		struct device_node *prevsib;
-		for (prevsib = np->parent->child;
-		     prevsib->sibling != np;
-		     prevsib = prevsib->sibling)
-			;
-		prevsib->sibling = np->sibling;
-	}
-
-	of_node_set_flag(np, OF_DETACHED);
-	raw_spin_unlock_irqrestore(&devtree_lock, flags);
-
-	of_node_remove(np);
 	return rc;
 }
-#endif /* defined(CONFIG_OF_DYNAMIC) */
 
 static void of_alias_add(struct alias_prop *ap, struct device_node *np,
 			 int id, const char *stem, int stem_len)
@@ -2062,9 +1852,12 @@
 		of_chosen = of_find_node_by_path("/chosen@0");
 
 	if (of_chosen) {
+		/* linux,stdout-path and /aliases/stdout are for legacy compatibility */
 		const char *name = of_get_property(of_chosen, "stdout-path", NULL);
 		if (!name)
 			name = of_get_property(of_chosen, "linux,stdout-path", NULL);
+		if (IS_ENABLED(CONFIG_PPC) && !name)
+			name = of_get_property(of_aliases, "stdout", NULL);
 		if (name)
 			of_stdout = of_find_node_by_path(name);
 	}
@@ -2122,7 +1915,7 @@
 	struct alias_prop *app;
 	int id = -ENODEV;
 
-	mutex_lock(&of_aliases_mutex);
+	mutex_lock(&of_mutex);
 	list_for_each_entry(app, &aliases_lookup, link) {
 		if (strcmp(app->stem, stem) != 0)
 			continue;
@@ -2132,7 +1925,7 @@
 			break;
 		}
 	}
-	mutex_unlock(&of_aliases_mutex);
+	mutex_unlock(&of_mutex);
 
 	return id;
 }
@@ -2180,20 +1973,22 @@
 EXPORT_SYMBOL_GPL(of_prop_next_string);
 
 /**
- * of_device_is_stdout_path - check if a device node matches the
- *                            linux,stdout-path property
+ * of_console_check() - Test and setup console for DT setup
+ * @dn - Pointer to device node
+ * @name - Name to use for preferred console without index. ex. "ttyS"
+ * @index - Index to use for preferred console.
  *
- * Check if this device node matches the linux,stdout-path property
- * in the chosen node. return true if yes, false otherwise.
+ * Check if the given device node matches the stdout-path property in the
+ * /chosen node. If it does then register it as the preferred console and return
+ * TRUE. Otherwise return FALSE.
  */
-int of_device_is_stdout_path(struct device_node *dn)
+bool of_console_check(struct device_node *dn, char *name, int index)
 {
-	if (!of_stdout)
+	if (!dn || dn != of_stdout || console_set_on_cmdline)
 		return false;
-
-	return of_stdout == dn;
+	return add_preferred_console(name, index, NULL);
 }
-EXPORT_SYMBOL_GPL(of_device_is_stdout_path);
+EXPORT_SYMBOL_GPL(of_console_check);
 
 /**
  *	of_find_next_cache_node - Find a node's subsidiary cache
diff --git a/drivers/of/device.c b/drivers/of/device.c
index dafb973..46d6c75c 100644
--- a/drivers/of/device.c
+++ b/drivers/of/device.c
@@ -160,7 +160,7 @@
 	add_uevent_var(env, "OF_COMPATIBLE_N=%d", seen);
 
 	seen = 0;
-	mutex_lock(&of_aliases_mutex);
+	mutex_lock(&of_mutex);
 	list_for_each_entry(app, &aliases_lookup, link) {
 		if (dev->of_node == app->np) {
 			add_uevent_var(env, "OF_ALIAS_%d=%s", seen,
@@ -168,7 +168,7 @@
 			seen++;
 		}
 	}
-	mutex_unlock(&of_aliases_mutex);
+	mutex_unlock(&of_mutex);
 }
 
 int of_device_uevent_modalias(struct device *dev, struct kobj_uevent_env *env)
diff --git a/drivers/of/dynamic.c b/drivers/of/dynamic.c
new file mode 100644
index 0000000..54fecc4
--- /dev/null
+++ b/drivers/of/dynamic.c
@@ -0,0 +1,660 @@
+/*
+ * Support for dynamic device trees.
+ *
+ * On some platforms, the device tree can be manipulated at runtime.
+ * The routines in this section support adding, removing and changing
+ * device tree nodes.
+ */
+
+#include <linux/of.h>
+#include <linux/spinlock.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include <linux/proc_fs.h>
+
+#include "of_private.h"
+
+/**
+ * of_node_get() - Increment refcount of a node
+ * @node:	Node to inc refcount, NULL is supported to simplify writing of
+ *		callers
+ *
+ * Returns node.
+ */
+struct device_node *of_node_get(struct device_node *node)
+{
+	if (node)
+		kobject_get(&node->kobj);
+	return node;
+}
+EXPORT_SYMBOL(of_node_get);
+
+/**
+ * of_node_put() - Decrement refcount of a node
+ * @node:	Node to dec refcount, NULL is supported to simplify writing of
+ *		callers
+ */
+void of_node_put(struct device_node *node)
+{
+	if (node)
+		kobject_put(&node->kobj);
+}
+EXPORT_SYMBOL(of_node_put);
+
+void __of_detach_node_sysfs(struct device_node *np)
+{
+	struct property *pp;
+
+	BUG_ON(!of_node_is_initialized(np));
+	if (!of_kset)
+		return;
+
+	/* only remove properties if on sysfs */
+	if (of_node_is_attached(np)) {
+		for_each_property_of_node(np, pp)
+			sysfs_remove_bin_file(&np->kobj, &pp->attr);
+		kobject_del(&np->kobj);
+	}
+
+	/* finally remove the kobj_init ref */
+	of_node_put(np);
+}
+
+static BLOCKING_NOTIFIER_HEAD(of_reconfig_chain);
+
+int of_reconfig_notifier_register(struct notifier_block *nb)
+{
+	return blocking_notifier_chain_register(&of_reconfig_chain, nb);
+}
+EXPORT_SYMBOL_GPL(of_reconfig_notifier_register);
+
+int of_reconfig_notifier_unregister(struct notifier_block *nb)
+{
+	return blocking_notifier_chain_unregister(&of_reconfig_chain, nb);
+}
+EXPORT_SYMBOL_GPL(of_reconfig_notifier_unregister);
+
+int of_reconfig_notify(unsigned long action, void *p)
+{
+	int rc;
+
+	rc = blocking_notifier_call_chain(&of_reconfig_chain, action, p);
+	return notifier_to_errno(rc);
+}
+
+int of_property_notify(int action, struct device_node *np,
+		       struct property *prop, struct property *oldprop)
+{
+	struct of_prop_reconfig pr;
+
+	/* only call notifiers if the node is attached */
+	if (!of_node_is_attached(np))
+		return 0;
+
+	pr.dn = np;
+	pr.prop = prop;
+	pr.old_prop = oldprop;
+	return of_reconfig_notify(action, &pr);
+}
+
+void __of_attach_node(struct device_node *np)
+{
+	const __be32 *phandle;
+	int sz;
+
+	np->name = __of_get_property(np, "name", NULL) ? : "<NULL>";
+	np->type = __of_get_property(np, "device_type", NULL) ? : "<NULL>";
+
+	phandle = __of_get_property(np, "phandle", &sz);
+	if (!phandle)
+		phandle = __of_get_property(np, "linux,phandle", &sz);
+	if (IS_ENABLED(PPC_PSERIES) && !phandle)
+		phandle = __of_get_property(np, "ibm,phandle", &sz);
+	np->phandle = (phandle && (sz >= 4)) ? be32_to_cpup(phandle) : 0;
+
+	np->child = NULL;
+	np->sibling = np->parent->child;
+	np->allnext = np->parent->allnext;
+	np->parent->allnext = np;
+	np->parent->child = np;
+	of_node_clear_flag(np, OF_DETACHED);
+}
+
+/**
+ * of_attach_node() - Plug a device node into the tree and global list.
+ */
+int of_attach_node(struct device_node *np)
+{
+	unsigned long flags;
+
+	mutex_lock(&of_mutex);
+	raw_spin_lock_irqsave(&devtree_lock, flags);
+	__of_attach_node(np);
+	raw_spin_unlock_irqrestore(&devtree_lock, flags);
+
+	__of_attach_node_sysfs(np);
+	mutex_unlock(&of_mutex);
+
+	of_reconfig_notify(OF_RECONFIG_ATTACH_NODE, np);
+
+	return 0;
+}
+
+void __of_detach_node(struct device_node *np)
+{
+	struct device_node *parent;
+
+	if (WARN_ON(of_node_check_flag(np, OF_DETACHED)))
+		return;
+
+	parent = np->parent;
+	if (WARN_ON(!parent))
+		return;
+
+	if (of_allnodes == np)
+		of_allnodes = np->allnext;
+	else {
+		struct device_node *prev;
+		for (prev = of_allnodes;
+		     prev->allnext != np;
+		     prev = prev->allnext)
+			;
+		prev->allnext = np->allnext;
+	}
+
+	if (parent->child == np)
+		parent->child = np->sibling;
+	else {
+		struct device_node *prevsib;
+		for (prevsib = np->parent->child;
+		     prevsib->sibling != np;
+		     prevsib = prevsib->sibling)
+			;
+		prevsib->sibling = np->sibling;
+	}
+
+	of_node_set_flag(np, OF_DETACHED);
+}
+
+/**
+ * of_detach_node() - "Unplug" a node from the device tree.
+ *
+ * The caller must hold a reference to the node.  The memory associated with
+ * the node is not freed until its refcount goes to zero.
+ */
+int of_detach_node(struct device_node *np)
+{
+	unsigned long flags;
+	int rc = 0;
+
+	mutex_lock(&of_mutex);
+	raw_spin_lock_irqsave(&devtree_lock, flags);
+	__of_detach_node(np);
+	raw_spin_unlock_irqrestore(&devtree_lock, flags);
+
+	__of_detach_node_sysfs(np);
+	mutex_unlock(&of_mutex);
+
+	of_reconfig_notify(OF_RECONFIG_DETACH_NODE, np);
+
+	return rc;
+}
+
+/**
+ * of_node_release() - release a dynamically allocated node
+ * @kref: kref element of the node to be released
+ *
+ * In of_node_put() this function is passed to kref_put() as the destructor.
+ */
+void of_node_release(struct kobject *kobj)
+{
+	struct device_node *node = kobj_to_device_node(kobj);
+	struct property *prop = node->properties;
+
+	/* We should never be releasing nodes that haven't been detached. */
+	if (!of_node_check_flag(node, OF_DETACHED)) {
+		pr_err("ERROR: Bad of_node_put() on %s\n", node->full_name);
+		dump_stack();
+		return;
+	}
+
+	if (!of_node_check_flag(node, OF_DYNAMIC))
+		return;
+
+	while (prop) {
+		struct property *next = prop->next;
+		kfree(prop->name);
+		kfree(prop->value);
+		kfree(prop);
+		prop = next;
+
+		if (!prop) {
+			prop = node->deadprops;
+			node->deadprops = NULL;
+		}
+	}
+	kfree(node->full_name);
+	kfree(node->data);
+	kfree(node);
+}
+
+/**
+ * __of_prop_dup - Copy a property dynamically.
+ * @prop:	Property to copy
+ * @allocflags:	Allocation flags (typically pass GFP_KERNEL)
+ *
+ * Copy a property by dynamically allocating the memory of both the
+ * property stucture and the property name & contents. The property's
+ * flags have the OF_DYNAMIC bit set so that we can differentiate between
+ * dynamically allocated properties and not.
+ * Returns the newly allocated property or NULL on out of memory error.
+ */
+struct property *__of_prop_dup(const struct property *prop, gfp_t allocflags)
+{
+	struct property *new;
+
+	new = kzalloc(sizeof(*new), allocflags);
+	if (!new)
+		return NULL;
+
+	/*
+	 * NOTE: There is no check for zero length value.
+	 * In case of a boolean property, this will allocate a value
+	 * of zero bytes. We do this to work around the use
+	 * of of_get_property() calls on boolean values.
+	 */
+	new->name = kstrdup(prop->name, allocflags);
+	new->value = kmemdup(prop->value, prop->length, allocflags);
+	new->length = prop->length;
+	if (!new->name || !new->value)
+		goto err_free;
+
+	/* mark the property as dynamic */
+	of_property_set_flag(new, OF_DYNAMIC);
+
+	return new;
+
+ err_free:
+	kfree(new->name);
+	kfree(new->value);
+	kfree(new);
+	return NULL;
+}
+
+/**
+ * __of_node_alloc() - Create an empty device node dynamically.
+ * @full_name:	Full name of the new device node
+ * @allocflags:	Allocation flags (typically pass GFP_KERNEL)
+ *
+ * Create an empty device tree node, suitable for further modification.
+ * The node data are dynamically allocated and all the node flags
+ * have the OF_DYNAMIC & OF_DETACHED bits set.
+ * Returns the newly allocated node or NULL on out of memory error.
+ */
+struct device_node *__of_node_alloc(const char *full_name, gfp_t allocflags)
+{
+	struct device_node *node;
+
+	node = kzalloc(sizeof(*node), allocflags);
+	if (!node)
+		return NULL;
+
+	node->full_name = kstrdup(full_name, allocflags);
+	of_node_set_flag(node, OF_DYNAMIC);
+	of_node_set_flag(node, OF_DETACHED);
+	if (!node->full_name)
+		goto err_free;
+
+	of_node_init(node);
+
+	return node;
+
+ err_free:
+	kfree(node->full_name);
+	kfree(node);
+	return NULL;
+}
+
+static void __of_changeset_entry_destroy(struct of_changeset_entry *ce)
+{
+	of_node_put(ce->np);
+	list_del(&ce->node);
+	kfree(ce);
+}
+
+#ifdef DEBUG
+static void __of_changeset_entry_dump(struct of_changeset_entry *ce)
+{
+	switch (ce->action) {
+	case OF_RECONFIG_ADD_PROPERTY:
+		pr_debug("%p: %s %s/%s\n",
+			ce, "ADD_PROPERTY   ", ce->np->full_name,
+			ce->prop->name);
+		break;
+	case OF_RECONFIG_REMOVE_PROPERTY:
+		pr_debug("%p: %s %s/%s\n",
+			ce, "REMOVE_PROPERTY", ce->np->full_name,
+			ce->prop->name);
+		break;
+	case OF_RECONFIG_UPDATE_PROPERTY:
+		pr_debug("%p: %s %s/%s\n",
+			ce, "UPDATE_PROPERTY", ce->np->full_name,
+			ce->prop->name);
+		break;
+	case OF_RECONFIG_ATTACH_NODE:
+		pr_debug("%p: %s %s\n",
+			ce, "ATTACH_NODE    ", ce->np->full_name);
+		break;
+	case OF_RECONFIG_DETACH_NODE:
+		pr_debug("%p: %s %s\n",
+			ce, "DETACH_NODE    ", ce->np->full_name);
+		break;
+	}
+}
+#else
+static inline void __of_changeset_entry_dump(struct of_changeset_entry *ce)
+{
+	/* empty */
+}
+#endif
+
+static void __of_changeset_entry_invert(struct of_changeset_entry *ce,
+					  struct of_changeset_entry *rce)
+{
+	memcpy(rce, ce, sizeof(*rce));
+
+	switch (ce->action) {
+	case OF_RECONFIG_ATTACH_NODE:
+		rce->action = OF_RECONFIG_DETACH_NODE;
+		break;
+	case OF_RECONFIG_DETACH_NODE:
+		rce->action = OF_RECONFIG_ATTACH_NODE;
+		break;
+	case OF_RECONFIG_ADD_PROPERTY:
+		rce->action = OF_RECONFIG_REMOVE_PROPERTY;
+		break;
+	case OF_RECONFIG_REMOVE_PROPERTY:
+		rce->action = OF_RECONFIG_ADD_PROPERTY;
+		break;
+	case OF_RECONFIG_UPDATE_PROPERTY:
+		rce->old_prop = ce->prop;
+		rce->prop = ce->old_prop;
+		break;
+	}
+}
+
+static void __of_changeset_entry_notify(struct of_changeset_entry *ce, bool revert)
+{
+	struct of_changeset_entry ce_inverted;
+	int ret;
+
+	if (revert) {
+		__of_changeset_entry_invert(ce, &ce_inverted);
+		ce = &ce_inverted;
+	}
+
+	switch (ce->action) {
+	case OF_RECONFIG_ATTACH_NODE:
+	case OF_RECONFIG_DETACH_NODE:
+		ret = of_reconfig_notify(ce->action, ce->np);
+		break;
+	case OF_RECONFIG_ADD_PROPERTY:
+	case OF_RECONFIG_REMOVE_PROPERTY:
+	case OF_RECONFIG_UPDATE_PROPERTY:
+		ret = of_property_notify(ce->action, ce->np, ce->prop, ce->old_prop);
+		break;
+	default:
+		pr_err("%s: invalid devicetree changeset action: %i\n", __func__,
+			(int)ce->action);
+		return;
+	}
+
+	if (ret)
+		pr_err("%s: notifier error @%s\n", __func__, ce->np->full_name);
+}
+
+static int __of_changeset_entry_apply(struct of_changeset_entry *ce)
+{
+	struct property *old_prop, **propp;
+	unsigned long flags;
+	int ret = 0;
+
+	__of_changeset_entry_dump(ce);
+
+	raw_spin_lock_irqsave(&devtree_lock, flags);
+	switch (ce->action) {
+	case OF_RECONFIG_ATTACH_NODE:
+		__of_attach_node(ce->np);
+		break;
+	case OF_RECONFIG_DETACH_NODE:
+		__of_detach_node(ce->np);
+		break;
+	case OF_RECONFIG_ADD_PROPERTY:
+		/* If the property is in deadprops then it must be removed */
+		for (propp = &ce->np->deadprops; *propp; propp = &(*propp)->next) {
+			if (*propp == ce->prop) {
+				*propp = ce->prop->next;
+				ce->prop->next = NULL;
+				break;
+			}
+		}
+
+		ret = __of_add_property(ce->np, ce->prop);
+		if (ret) {
+			pr_err("%s: add_property failed @%s/%s\n",
+				__func__, ce->np->full_name,
+				ce->prop->name);
+			break;
+		}
+		break;
+	case OF_RECONFIG_REMOVE_PROPERTY:
+		ret = __of_remove_property(ce->np, ce->prop);
+		if (ret) {
+			pr_err("%s: remove_property failed @%s/%s\n",
+				__func__, ce->np->full_name,
+				ce->prop->name);
+			break;
+		}
+		break;
+
+	case OF_RECONFIG_UPDATE_PROPERTY:
+		/* If the property is in deadprops then it must be removed */
+		for (propp = &ce->np->deadprops; *propp; propp = &(*propp)->next) {
+			if (*propp == ce->prop) {
+				*propp = ce->prop->next;
+				ce->prop->next = NULL;
+				break;
+			}
+		}
+
+		ret = __of_update_property(ce->np, ce->prop, &old_prop);
+		if (ret) {
+			pr_err("%s: update_property failed @%s/%s\n",
+				__func__, ce->np->full_name,
+				ce->prop->name);
+			break;
+		}
+		break;
+	default:
+		ret = -EINVAL;
+	}
+	raw_spin_unlock_irqrestore(&devtree_lock, flags);
+
+	if (ret)
+		return ret;
+
+	switch (ce->action) {
+	case OF_RECONFIG_ATTACH_NODE:
+		__of_attach_node_sysfs(ce->np);
+		break;
+	case OF_RECONFIG_DETACH_NODE:
+		__of_detach_node_sysfs(ce->np);
+		break;
+	case OF_RECONFIG_ADD_PROPERTY:
+		/* ignore duplicate names */
+		__of_add_property_sysfs(ce->np, ce->prop);
+		break;
+	case OF_RECONFIG_REMOVE_PROPERTY:
+		__of_remove_property_sysfs(ce->np, ce->prop);
+		break;
+	case OF_RECONFIG_UPDATE_PROPERTY:
+		__of_update_property_sysfs(ce->np, ce->prop, ce->old_prop);
+		break;
+	}
+
+	return 0;
+}
+
+static inline int __of_changeset_entry_revert(struct of_changeset_entry *ce)
+{
+	struct of_changeset_entry ce_inverted;
+
+	__of_changeset_entry_invert(ce, &ce_inverted);
+	return __of_changeset_entry_apply(&ce_inverted);
+}
+
+/**
+ * of_changeset_init - Initialize a changeset for use
+ *
+ * @ocs:	changeset pointer
+ *
+ * Initialize a changeset structure
+ */
+void of_changeset_init(struct of_changeset *ocs)
+{
+	memset(ocs, 0, sizeof(*ocs));
+	INIT_LIST_HEAD(&ocs->entries);
+}
+
+/**
+ * of_changeset_destroy - Destroy a changeset
+ *
+ * @ocs:	changeset pointer
+ *
+ * Destroys a changeset. Note that if a changeset is applied,
+ * its changes to the tree cannot be reverted.
+ */
+void of_changeset_destroy(struct of_changeset *ocs)
+{
+	struct of_changeset_entry *ce, *cen;
+
+	list_for_each_entry_safe_reverse(ce, cen, &ocs->entries, node)
+		__of_changeset_entry_destroy(ce);
+}
+
+/**
+ * of_changeset_apply - Applies a changeset
+ *
+ * @ocs:	changeset pointer
+ *
+ * Applies a changeset to the live tree.
+ * Any side-effects of live tree state changes are applied here on
+ * sucess, like creation/destruction of devices and side-effects
+ * like creation of sysfs properties and directories.
+ * Returns 0 on success, a negative error value in case of an error.
+ * On error the partially applied effects are reverted.
+ */
+int of_changeset_apply(struct of_changeset *ocs)
+{
+	struct of_changeset_entry *ce;
+	int ret;
+
+	/* perform the rest of the work */
+	pr_debug("of_changeset: applying...\n");
+	list_for_each_entry(ce, &ocs->entries, node) {
+		ret = __of_changeset_entry_apply(ce);
+		if (ret) {
+			pr_err("%s: Error applying changeset (%d)\n", __func__, ret);
+			list_for_each_entry_continue_reverse(ce, &ocs->entries, node)
+				__of_changeset_entry_revert(ce);
+			return ret;
+		}
+	}
+	pr_debug("of_changeset: applied, emitting notifiers.\n");
+
+	/* drop the global lock while emitting notifiers */
+	mutex_unlock(&of_mutex);
+	list_for_each_entry(ce, &ocs->entries, node)
+		__of_changeset_entry_notify(ce, 0);
+	mutex_lock(&of_mutex);
+	pr_debug("of_changeset: notifiers sent.\n");
+
+	return 0;
+}
+
+/**
+ * of_changeset_revert - Reverts an applied changeset
+ *
+ * @ocs:	changeset pointer
+ *
+ * Reverts a changeset returning the state of the tree to what it
+ * was before the application.
+ * Any side-effects like creation/destruction of devices and
+ * removal of sysfs properties and directories are applied.
+ * Returns 0 on success, a negative error value in case of an error.
+ */
+int of_changeset_revert(struct of_changeset *ocs)
+{
+	struct of_changeset_entry *ce;
+	int ret;
+
+	pr_debug("of_changeset: reverting...\n");
+	list_for_each_entry_reverse(ce, &ocs->entries, node) {
+		ret = __of_changeset_entry_revert(ce);
+		if (ret) {
+			pr_err("%s: Error reverting changeset (%d)\n", __func__, ret);
+			list_for_each_entry_continue(ce, &ocs->entries, node)
+				__of_changeset_entry_apply(ce);
+			return ret;
+		}
+	}
+	pr_debug("of_changeset: reverted, emitting notifiers.\n");
+
+	/* drop the global lock while emitting notifiers */
+	mutex_unlock(&of_mutex);
+	list_for_each_entry_reverse(ce, &ocs->entries, node)
+		__of_changeset_entry_notify(ce, 1);
+	mutex_lock(&of_mutex);
+	pr_debug("of_changeset: notifiers sent.\n");
+
+	return 0;
+}
+
+/**
+ * of_changeset_action - Perform a changeset action
+ *
+ * @ocs:	changeset pointer
+ * @action:	action to perform
+ * @np:		Pointer to device node
+ * @prop:	Pointer to property
+ *
+ * On action being one of:
+ * + OF_RECONFIG_ATTACH_NODE
+ * + OF_RECONFIG_DETACH_NODE,
+ * + OF_RECONFIG_ADD_PROPERTY
+ * + OF_RECONFIG_REMOVE_PROPERTY,
+ * + OF_RECONFIG_UPDATE_PROPERTY
+ * Returns 0 on success, a negative error value in case of an error.
+ */
+int of_changeset_action(struct of_changeset *ocs, unsigned long action,
+		struct device_node *np, struct property *prop)
+{
+	struct of_changeset_entry *ce;
+
+	ce = kzalloc(sizeof(*ce), GFP_KERNEL);
+	if (!ce) {
+		pr_err("%s: Failed to allocate\n", __func__);
+		return -ENOMEM;
+	}
+	/* get a reference to the node */
+	ce->action = action;
+	ce->np = of_node_get(np);
+	ce->prop = prop;
+
+	if (action == OF_RECONFIG_UPDATE_PROPERTY && prop)
+		ce->old_prop = of_find_property(np, prop->name, NULL);
+
+	/* add it to the list */
+	list_add_tail(&ce->node, &ocs->entries);
+	return 0;
+}
diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c
index 9aa012e..f46a24f 100644
--- a/drivers/of/fdt.c
+++ b/drivers/of/fdt.c
@@ -923,24 +923,24 @@
 }
 
 #ifdef CONFIG_HAVE_MEMBLOCK
+#define MAX_PHYS_ADDR	((phys_addr_t)~0)
+
 void __init __weak early_init_dt_add_memory_arch(u64 base, u64 size)
 {
 	const u64 phys_offset = __pa(PAGE_OFFSET);
 	base &= PAGE_MASK;
 	size &= PAGE_MASK;
 
-	if (sizeof(phys_addr_t) < sizeof(u64)) {
-		if (base > ULONG_MAX) {
-			pr_warning("Ignoring memory block 0x%llx - 0x%llx\n",
-					base, base + size);
-			return;
-		}
+	if (base > MAX_PHYS_ADDR) {
+		pr_warning("Ignoring memory block 0x%llx - 0x%llx\n",
+				base, base + size);
+		return;
+	}
 
-		if (base + size > ULONG_MAX) {
-			pr_warning("Ignoring memory range 0x%lx - 0x%llx\n",
-					ULONG_MAX, base + size);
-			size = ULONG_MAX - base;
-		}
+	if (base + size > MAX_PHYS_ADDR) {
+		pr_warning("Ignoring memory range 0x%lx - 0x%llx\n",
+				ULONG_MAX, base + size);
+		size = MAX_PHYS_ADDR - base;
 	}
 
 	if (base + size < phys_offset) {
diff --git a/drivers/of/of_private.h b/drivers/of/of_private.h
index ff350c8..858e0a5 100644
--- a/drivers/of/of_private.h
+++ b/drivers/of/of_private.h
@@ -31,6 +31,63 @@
 	char stem[0];
 };
 
-extern struct mutex of_aliases_mutex;
+extern struct mutex of_mutex;
 extern struct list_head aliases_lookup;
+extern struct kset *of_kset;
+
+
+static inline struct device_node *kobj_to_device_node(struct kobject *kobj)
+{
+	return container_of(kobj, struct device_node, kobj);
+}
+
+#if defined(CONFIG_OF_DYNAMIC)
+extern int of_property_notify(int action, struct device_node *np,
+			      struct property *prop, struct property *old_prop);
+extern void of_node_release(struct kobject *kobj);
+#else /* CONFIG_OF_DYNAMIC */
+static inline int of_property_notify(int action, struct device_node *np,
+				     struct property *prop, struct property *old_prop)
+{
+	return 0;
+}
+#endif /* CONFIG_OF_DYNAMIC */
+
+/**
+ * General utilities for working with live trees.
+ *
+ * All functions with two leading underscores operate
+ * without taking node references, so you either have to
+ * own the devtree lock or work on detached trees only.
+ */
+struct property *__of_prop_dup(const struct property *prop, gfp_t allocflags);
+struct device_node *__of_node_alloc(const char *full_name, gfp_t allocflags);
+
+extern const void *__of_get_property(const struct device_node *np,
+				     const char *name, int *lenp);
+extern int __of_add_property(struct device_node *np, struct property *prop);
+extern int __of_add_property_sysfs(struct device_node *np,
+		struct property *prop);
+extern int __of_remove_property(struct device_node *np, struct property *prop);
+extern void __of_remove_property_sysfs(struct device_node *np,
+		struct property *prop);
+extern int __of_update_property(struct device_node *np,
+		struct property *newprop, struct property **oldprop);
+extern void __of_update_property_sysfs(struct device_node *np,
+		struct property *newprop, struct property *oldprop);
+
+extern void __of_attach_node(struct device_node *np);
+extern int __of_attach_node_sysfs(struct device_node *np);
+extern void __of_detach_node(struct device_node *np);
+extern void __of_detach_node_sysfs(struct device_node *np);
+
+/* iterators for transactions, used for overlays */
+/* forward iterator */
+#define for_each_transaction_entry(_oft, _te) \
+	list_for_each_entry(_te, &(_oft)->te_list, node)
+
+/* reverse iterator */
+#define for_each_transaction_entry_reverse(_oft, _te) \
+	list_for_each_entry_reverse(_te, &(_oft)->te_list, node)
+
 #endif /* _LINUX_OF_PRIVATE_H */
diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
index 632aae8..59fb12e 100644
--- a/drivers/of/of_reserved_mem.c
+++ b/drivers/of/of_reserved_mem.c
@@ -206,8 +206,16 @@
 	for (i = 0; i < reserved_mem_count; i++) {
 		struct reserved_mem *rmem = &reserved_mem[i];
 		unsigned long node = rmem->fdt_node;
+		int len;
+		const __be32 *prop;
 		int err = 0;
 
+		prop = of_get_flat_dt_prop(node, "phandle", &len);
+		if (!prop)
+			prop = of_get_flat_dt_prop(node, "linux,phandle", &len);
+		if (prop)
+			rmem->phandle = of_read_number(prop, len/4);
+
 		if (rmem->size == 0)
 			err = __reserved_mem_alloc_size(node, rmem->name,
 						 &rmem->base, &rmem->size);
@@ -215,3 +223,65 @@
 			__reserved_mem_init_node(rmem);
 	}
 }
+
+static inline struct reserved_mem *__find_rmem(struct device_node *node)
+{
+	unsigned int i;
+
+	if (!node->phandle)
+		return NULL;
+
+	for (i = 0; i < reserved_mem_count; i++)
+		if (reserved_mem[i].phandle == node->phandle)
+			return &reserved_mem[i];
+	return NULL;
+}
+
+/**
+ * of_reserved_mem_device_init() - assign reserved memory region to given device
+ *
+ * This function assign memory region pointed by "memory-region" device tree
+ * property to the given device.
+ */
+void of_reserved_mem_device_init(struct device *dev)
+{
+	struct reserved_mem *rmem;
+	struct device_node *np;
+
+	np = of_parse_phandle(dev->of_node, "memory-region", 0);
+	if (!np)
+		return;
+
+	rmem = __find_rmem(np);
+	of_node_put(np);
+
+	if (!rmem || !rmem->ops || !rmem->ops->device_init)
+		return;
+
+	rmem->ops->device_init(rmem, dev);
+	dev_info(dev, "assigned reserved memory node %s\n", rmem->name);
+}
+
+/**
+ * of_reserved_mem_device_release() - release reserved memory device structures
+ *
+ * This function releases structures allocated for memory region handling for
+ * the given device.
+ */
+void of_reserved_mem_device_release(struct device *dev)
+{
+	struct reserved_mem *rmem;
+	struct device_node *np;
+
+	np = of_parse_phandle(dev->of_node, "memory-region", 0);
+	if (!np)
+		return;
+
+	rmem = __find_rmem(np);
+	of_node_put(np);
+
+	if (!rmem || !rmem->ops || !rmem->ops->device_release)
+		return;
+
+	rmem->ops->device_release(rmem, dev);
+}
diff --git a/drivers/of/platform.c b/drivers/of/platform.c
index 500436f..0197725 100644
--- a/drivers/of/platform.c
+++ b/drivers/of/platform.c
@@ -422,6 +422,7 @@
 			break;
 		}
 	}
+	of_node_set_flag(bus, OF_POPULATED_BUS);
 	return rc;
 }
 
@@ -508,19 +509,13 @@
 
 static int of_platform_device_destroy(struct device *dev, void *data)
 {
-	bool *children_left = data;
-
 	/* Do not touch devices not populated from the device tree */
-	if (!dev->of_node || !of_node_check_flag(dev->of_node, OF_POPULATED)) {
-		*children_left = true;
+	if (!dev->of_node || !of_node_check_flag(dev->of_node, OF_POPULATED))
 		return 0;
-	}
 
-	/* Recurse, but don't touch this device if it has any children left */
-	if (of_platform_depopulate(dev) != 0) {
-		*children_left = true;
-		return 0;
-	}
+	/* Recurse for any nodes that were treated as busses */
+	if (of_node_check_flag(dev->of_node, OF_POPULATED_BUS))
+		device_for_each_child(dev, NULL, of_platform_device_destroy);
 
 	if (dev->bus == &platform_bus_type)
 		platform_device_unregister(to_platform_device(dev));
@@ -528,19 +523,15 @@
 	else if (dev->bus == &amba_bustype)
 		amba_device_unregister(to_amba_device(dev));
 #endif
-	else {
-		*children_left = true;
-		return 0;
-	}
 
 	of_node_clear_flag(dev->of_node, OF_POPULATED);
-
+	of_node_clear_flag(dev->of_node, OF_POPULATED_BUS);
 	return 0;
 }
 
 /**
  * of_platform_depopulate() - Remove devices populated from device tree
- * @parent: device which childred will be removed
+ * @parent: device which children will be removed
  *
  * Complementary to of_platform_populate(), this function removes children
  * of the given device (and, recurrently, their children) that have been
@@ -550,14 +541,9 @@
  * Returns 0 when all children devices have been removed or
  * -EBUSY when some children remained.
  */
-int of_platform_depopulate(struct device *parent)
+void of_platform_depopulate(struct device *parent)
 {
-	bool children_left = false;
-
-	device_for_each_child(parent, &children_left,
-			      of_platform_device_destroy);
-
-	return children_left ? -EBUSY : 0;
+	device_for_each_child(parent, NULL, of_platform_device_destroy);
 }
 EXPORT_SYMBOL_GPL(of_platform_depopulate);
 
diff --git a/drivers/of/selftest.c b/drivers/of/selftest.c
index 077314e..d410026 100644
--- a/drivers/of/selftest.c
+++ b/drivers/of/selftest.c
@@ -9,6 +9,7 @@
 #include <linux/errno.h>
 #include <linux/module.h>
 #include <linux/of.h>
+#include <linux/of_fdt.h>
 #include <linux/of_irq.h>
 #include <linux/of_platform.h>
 #include <linux/list.h>
@@ -16,11 +17,17 @@
 #include <linux/slab.h>
 #include <linux/device.h>
 
+#include "of_private.h"
+
 static struct selftest_results {
 	int passed;
 	int failed;
 } selftest_results;
 
+#define NO_OF_NODES 2
+static struct device_node *nodes[NO_OF_NODES];
+static int last_node_index;
+
 #define selftest(result, fmt, ...) { \
 	if (!(result)) { \
 		selftest_results.failed++; \
@@ -266,6 +273,81 @@
 	selftest(rc == -EILSEQ, "unterminated string; rc=%i", rc);
 }
 
+#define propcmp(p1, p2) (((p1)->length == (p2)->length) && \
+			(p1)->value && (p2)->value && \
+			!memcmp((p1)->value, (p2)->value, (p1)->length) && \
+			!strcmp((p1)->name, (p2)->name))
+static void __init of_selftest_property_copy(void)
+{
+#ifdef CONFIG_OF_DYNAMIC
+	struct property p1 = { .name = "p1", .length = 0, .value = "" };
+	struct property p2 = { .name = "p2", .length = 5, .value = "abcd" };
+	struct property *new;
+
+	new = __of_prop_dup(&p1, GFP_KERNEL);
+	selftest(new && propcmp(&p1, new), "empty property didn't copy correctly\n");
+	kfree(new->value);
+	kfree(new->name);
+	kfree(new);
+
+	new = __of_prop_dup(&p2, GFP_KERNEL);
+	selftest(new && propcmp(&p2, new), "non-empty property didn't copy correctly\n");
+	kfree(new->value);
+	kfree(new->name);
+	kfree(new);
+#endif
+}
+
+static void __init of_selftest_changeset(void)
+{
+#ifdef CONFIG_OF_DYNAMIC
+	struct property *ppadd, padd = { .name = "prop-add", .length = 0, .value = "" };
+	struct property *ppupdate, pupdate = { .name = "prop-update", .length = 5, .value = "abcd" };
+	struct property *ppremove;
+	struct device_node *n1, *n2, *n21, *nremove, *parent;
+	struct of_changeset chgset;
+
+	of_changeset_init(&chgset);
+	n1 = __of_node_alloc("/testcase-data/changeset/n1", GFP_KERNEL);
+	selftest(n1, "testcase setup failure\n");
+	n2 = __of_node_alloc("/testcase-data/changeset/n2", GFP_KERNEL);
+	selftest(n2, "testcase setup failure\n");
+	n21 = __of_node_alloc("/testcase-data/changeset/n2/n21", GFP_KERNEL);
+	selftest(n21, "testcase setup failure %p\n", n21);
+	nremove = of_find_node_by_path("/testcase-data/changeset/node-remove");
+	selftest(nremove, "testcase setup failure\n");
+	ppadd = __of_prop_dup(&padd, GFP_KERNEL);
+	selftest(ppadd, "testcase setup failure\n");
+	ppupdate = __of_prop_dup(&pupdate, GFP_KERNEL);
+	selftest(ppupdate, "testcase setup failure\n");
+	parent = nremove->parent;
+	n1->parent = parent;
+	n2->parent = parent;
+	n21->parent = n2;
+	n2->child = n21;
+	ppremove = of_find_property(parent, "prop-remove", NULL);
+	selftest(ppremove, "failed to find removal prop");
+
+	of_changeset_init(&chgset);
+	selftest(!of_changeset_attach_node(&chgset, n1), "fail attach n1\n");
+	selftest(!of_changeset_attach_node(&chgset, n2), "fail attach n2\n");
+	selftest(!of_changeset_detach_node(&chgset, nremove), "fail remove node\n");
+	selftest(!of_changeset_attach_node(&chgset, n21), "fail attach n21\n");
+	selftest(!of_changeset_add_property(&chgset, parent, ppadd), "fail add prop\n");
+	selftest(!of_changeset_update_property(&chgset, parent, ppupdate), "fail update prop\n");
+	selftest(!of_changeset_remove_property(&chgset, parent, ppremove), "fail remove prop\n");
+	mutex_lock(&of_mutex);
+	selftest(!of_changeset_apply(&chgset), "apply failed\n");
+	mutex_unlock(&of_mutex);
+
+	mutex_lock(&of_mutex);
+	selftest(!of_changeset_revert(&chgset), "revert failed\n");
+	mutex_unlock(&of_mutex);
+
+	of_changeset_destroy(&chgset);
+#endif
+}
+
 static void __init of_selftest_parse_interrupts(void)
 {
 	struct device_node *np;
@@ -517,9 +599,156 @@
 	}
 }
 
+/**
+ *	update_node_properties - adds the properties
+ *	of np into dup node (present in live tree) and
+ *	updates parent of children of np to dup.
+ *
+ *	@np:	node already present in live tree
+ *	@dup:	node present in live tree to be updated
+ */
+static void update_node_properties(struct device_node *np,
+					struct device_node *dup)
+{
+	struct property *prop;
+	struct device_node *child;
+
+	for_each_property_of_node(np, prop)
+		of_add_property(dup, prop);
+
+	for_each_child_of_node(np, child)
+		child->parent = dup;
+}
+
+/**
+ *	attach_node_and_children - attaches nodes
+ *	and its children to live tree
+ *
+ *	@np:	Node to attach to live tree
+ */
+static int attach_node_and_children(struct device_node *np)
+{
+	struct device_node *next, *root = np, *dup;
+
+	if (!np) {
+		pr_warn("%s: No tree to attach; not running tests\n",
+			__func__);
+		return -ENODATA;
+	}
+
+
+	/* skip root node */
+	np = np->child;
+	/* storing a copy in temporary node */
+	dup = np;
+
+	while (dup) {
+		nodes[last_node_index++] = dup;
+		dup = dup->sibling;
+	}
+	dup = NULL;
+
+	while (np) {
+		next = np->allnext;
+		dup = of_find_node_by_path(np->full_name);
+		if (dup)
+			update_node_properties(np, dup);
+		else {
+			np->child = NULL;
+			if (np->parent == root)
+				np->parent = of_allnodes;
+			of_attach_node(np);
+		}
+		np = next;
+	}
+
+	return 0;
+}
+
+/**
+ *	selftest_data_add - Reads, copies data from
+ *	linked tree and attaches it to the live tree
+ */
+static int __init selftest_data_add(void)
+{
+	void *selftest_data;
+	struct device_node *selftest_data_node;
+	extern uint8_t __dtb_testcases_begin[];
+	extern uint8_t __dtb_testcases_end[];
+	const int size = __dtb_testcases_end - __dtb_testcases_begin;
+
+	if (!size || !of_allnodes) {
+		pr_warn("%s: No testcase data to attach; not running tests\n",
+			__func__);
+		return -ENODATA;
+	}
+
+	/* creating copy */
+	selftest_data = kmemdup(__dtb_testcases_begin, size, GFP_KERNEL);
+
+	if (!selftest_data) {
+		pr_warn("%s: Failed to allocate memory for selftest_data; "
+			"not running tests\n", __func__);
+		return -ENOMEM;
+	}
+	of_fdt_unflatten_tree(selftest_data, &selftest_data_node);
+
+	/* attach the sub-tree to live tree */
+	return attach_node_and_children(selftest_data_node);
+}
+
+/**
+ *	detach_node_and_children - detaches node
+ *	and its children from live tree
+ *
+ *	@np:	Node to detach from live tree
+ */
+static void detach_node_and_children(struct device_node *np)
+{
+	while (np->child)
+		detach_node_and_children(np->child);
+
+	while (np->sibling)
+		detach_node_and_children(np->sibling);
+
+	of_detach_node(np);
+}
+
+/**
+ *	selftest_data_remove - removes the selftest data
+ *	nodes from the live tree
+ */
+static void selftest_data_remove(void)
+{
+	struct device_node *np;
+	struct property *prop;
+
+	while (last_node_index >= 0) {
+		if (nodes[last_node_index]) {
+			np = of_find_node_by_path(nodes[last_node_index]->full_name);
+			if (strcmp(np->full_name, "/aliases") != 0) {
+				detach_node_and_children(np->child);
+				of_detach_node(np);
+			} else {
+				for_each_property_of_node(np, prop) {
+					if (strcmp(prop->name, "testcase-alias") == 0)
+						of_remove_property(np, prop);
+				}
+			}
+		}
+		last_node_index--;
+	}
+}
+
 static int __init of_selftest(void)
 {
 	struct device_node *np;
+	int res;
+
+	/* adding data for selftest */
+	res = selftest_data_add();
+	if (res)
+		return res;
 
 	np = of_find_node_by_path("/testcase-data/phandle-tests/consumer-a");
 	if (!np) {
@@ -533,12 +762,18 @@
 	of_selftest_dynamic();
 	of_selftest_parse_phandle_with_args();
 	of_selftest_property_match_string();
+	of_selftest_property_copy();
+	of_selftest_changeset();
 	of_selftest_parse_interrupts();
 	of_selftest_parse_interrupts_extended();
 	of_selftest_match_node();
 	of_selftest_platform_populate();
 	pr_info("end of selftest - %i passed, %i failed\n",
 		selftest_results.passed, selftest_results.failed);
+
+	/* removing selftest data from live tree */
+	selftest_data_remove();
+
 	return 0;
 }
 late_initcall(of_selftest);
diff --git a/drivers/of/testcase-data/testcases.dts b/drivers/of/testcase-data/testcases.dts
new file mode 100644
index 0000000..219ef93
--- /dev/null
+++ b/drivers/of/testcase-data/testcases.dts
@@ -0,0 +1,15 @@
+/dts-v1/;
+/ {
+	testcase-data {
+		changeset {
+			prop-update = "hello";
+			prop-remove = "world";
+			node-remove {
+			};
+		};
+	};
+};
+#include "tests-phandle.dtsi"
+#include "tests-interrupts.dtsi"
+#include "tests-match.dtsi"
+#include "tests-platform.dtsi"
diff --git a/drivers/of/testcase-data/testcases.dtsi b/drivers/of/testcase-data/testcases.dtsi
deleted file mode 100644
index 6d8d980a..0000000
--- a/drivers/of/testcase-data/testcases.dtsi
+++ /dev/null
@@ -1,4 +0,0 @@
-#include "tests-phandle.dtsi"
-#include "tests-interrupts.dtsi"
-#include "tests-match.dtsi"
-#include "tests-platform.dtsi"
diff --git a/drivers/pci/hotplug/rpaphp_core.c b/drivers/pci/hotplug/rpaphp_core.c
index 93aa29f..f2945fa 100644
--- a/drivers/pci/hotplug/rpaphp_core.c
+++ b/drivers/pci/hotplug/rpaphp_core.c
@@ -375,11 +375,11 @@
 
 static int __init rpaphp_init(void)
 {
-	struct device_node *dn = NULL;
+	struct device_node *dn;
 
 	info(DRIVER_DESC " version: " DRIVER_VERSION "\n");
 
-	while ((dn = of_find_node_by_name(dn, "pci")))
+	for_each_node_by_name(dn, "pci")
 		rpaphp_add_slot(dn);
 
 	return 0;
diff --git a/drivers/scsi/scsi_transport_srp.c b/drivers/scsi/scsi_transport_srp.c
index 43fea22..ae45bd9 100644
--- a/drivers/scsi/scsi_transport_srp.c
+++ b/drivers/scsi/scsi_transport_srp.c
@@ -472,7 +472,8 @@
 	if (delay > 0)
 		queue_delayed_work(system_long_wq, &rport->reconnect_work,
 				   1UL * delay * HZ);
-	if (srp_rport_set_state(rport, SRP_RPORT_BLOCKED) == 0) {
+	if ((fast_io_fail_tmo >= 0 || dev_loss_tmo >= 0) &&
+	    srp_rport_set_state(rport, SRP_RPORT_BLOCKED) == 0) {
 		pr_debug("%s new state: %d\n", dev_name(&shost->shost_gendev),
 			 rport->state);
 		scsi_target_block(&shost->shost_gendev);
diff --git a/drivers/tty/ehv_bytechan.c b/drivers/tty/ehv_bytechan.c
index 0419b69..4f485e8 100644
--- a/drivers/tty/ehv_bytechan.c
+++ b/drivers/tty/ehv_bytechan.c
@@ -108,55 +108,23 @@
  *
  * The byte channel to be used for the console is specified via a "stdout"
  * property in the /chosen node.
- *
- * For compatible with legacy device trees, we also look for a "stdout" alias.
  */
 static int find_console_handle(void)
 {
-	struct device_node *np, *np2;
+	struct device_node *np = of_stdout;
 	const char *sprop = NULL;
 	const uint32_t *iprop;
 
-	np = of_find_node_by_path("/chosen");
-	if (np)
-		sprop = of_get_property(np, "stdout-path", NULL);
-
-	if (!np || !sprop) {
-		of_node_put(np);
-		np = of_find_node_by_name(NULL, "aliases");
-		if (np)
-			sprop = of_get_property(np, "stdout", NULL);
-	}
-
-	if (!sprop) {
-		of_node_put(np);
-		return 0;
-	}
-
 	/* We don't care what the aliased node is actually called.  We only
 	 * care if it's compatible with "epapr,hv-byte-channel", because that
-	 * indicates that it's a byte channel node.  We use a temporary
-	 * variable, 'np2', because we can't release 'np' until we're done with
-	 * 'sprop'.
+	 * indicates that it's a byte channel node.
 	 */
-	np2 = of_find_node_by_path(sprop);
-	of_node_put(np);
-	np = np2;
-	if (!np) {
-		pr_warning("ehv-bc: stdout node '%s' does not exist\n", sprop);
+	if (!np || !of_device_is_compatible(np, "epapr,hv-byte-channel"))
 		return 0;
-	}
-
-	/* Is it a byte channel? */
-	if (!of_device_is_compatible(np, "epapr,hv-byte-channel")) {
-		of_node_put(np);
-		return 0;
-	}
 
 	stdout_irq = irq_of_parse_and_map(np, 0);
 	if (stdout_irq == NO_IRQ) {
-		pr_err("ehv-bc: no 'interrupts' property in %s node\n", sprop);
-		of_node_put(np);
+		pr_err("ehv-bc: no 'interrupts' property in %s node\n", np->full_name);
 		return 0;
 	}
 
@@ -167,12 +135,9 @@
 	if (!iprop) {
 		pr_err("ehv-bc: no 'hv-handle' property in %s node\n",
 		       np->name);
-		of_node_put(np);
 		return 0;
 	}
 	stdout_bc = be32_to_cpu(*iprop);
-
-	of_node_put(np);
 	return 1;
 }
 
diff --git a/drivers/tty/hvc/hvc_opal.c b/drivers/tty/hvc/hvc_opal.c
index a585079..a2cc5f8 100644
--- a/drivers/tty/hvc/hvc_opal.c
+++ b/drivers/tty/hvc/hvc_opal.c
@@ -342,22 +342,13 @@
 
 void __init hvc_opal_init_early(void)
 {
-	struct device_node *stdout_node = NULL;
+	struct device_node *stdout_node = of_node_get(of_stdout);
 	const __be32 *termno;
-	const char *name = NULL;
 	const struct hv_ops *ops;
 	u32 index;
 
-	/* find the boot console from /chosen/stdout */
-	if (of_chosen)
-		name = of_get_property(of_chosen, "linux,stdout-path", NULL);
-	if (name) {
-		stdout_node = of_find_node_by_path(name);
-		if (!stdout_node) {
-			pr_err("hvc_opal: Failed to locate default console!\n");
-			return;
-		}
-	} else {
+	/* If the console wasn't in /chosen, try /ibm,opal */
+	if (!stdout_node) {
 		struct device_node *opal, *np;
 
 		/* Current OPAL takeover doesn't provide the stdout
diff --git a/drivers/tty/hvc/hvc_vio.c b/drivers/tty/hvc/hvc_vio.c
index b594abf..5618b5f 100644
--- a/drivers/tty/hvc/hvc_vio.c
+++ b/drivers/tty/hvc/hvc_vio.c
@@ -404,42 +404,35 @@
 
 void __init hvc_vio_init_early(void)
 {
-	struct device_node *stdout_node;
 	const __be32 *termno;
 	const char *name;
 	const struct hv_ops *ops;
 
 	/* find the boot console from /chosen/stdout */
-	if (!of_chosen)
+	if (!of_stdout)
 		return;
-	name = of_get_property(of_chosen, "linux,stdout-path", NULL);
-	if (name == NULL)
-		return;
-	stdout_node = of_find_node_by_path(name);
-	if (!stdout_node)
-		return;
-	name = of_get_property(stdout_node, "name", NULL);
+	name = of_get_property(of_stdout, "name", NULL);
 	if (!name) {
 		printk(KERN_WARNING "stdout node missing 'name' property!\n");
-		goto out;
+		return;
 	}
 
 	/* Check if it's a virtual terminal */
 	if (strncmp(name, "vty", 3) != 0)
-		goto out;
-	termno = of_get_property(stdout_node, "reg", NULL);
+		return;
+	termno = of_get_property(of_stdout, "reg", NULL);
 	if (termno == NULL)
-		goto out;
+		return;
 	hvterm_priv0.termno = of_read_number(termno, 1);
 	spin_lock_init(&hvterm_priv0.buf_lock);
 	hvterm_privs[0] = &hvterm_priv0;
 
 	/* Check the protocol */
-	if (of_device_is_compatible(stdout_node, "hvterm1")) {
+	if (of_device_is_compatible(of_stdout, "hvterm1")) {
 		hvterm_priv0.proto = HV_PROTOCOL_RAW;
 		ops = &hvterm_raw_ops;
 	}
-	else if (of_device_is_compatible(stdout_node, "hvterm-protocol")) {
+	else if (of_device_is_compatible(of_stdout, "hvterm-protocol")) {
 		hvterm_priv0.proto = HV_PROTOCOL_HVSI;
 		ops = &hvterm_hvsi_ops;
 		hvsilib_init(&hvterm_priv0.hvsi, hvc_get_chars, hvc_put_chars,
@@ -447,7 +440,7 @@
 		/* HVSI, perform the handshake now */
 		hvsilib_establish(&hvterm_priv0.hvsi);
 	} else
-		goto out;
+		return;
 	udbg_putc = udbg_hvc_putc;
 	udbg_getc = udbg_hvc_getc;
 	udbg_getc_poll = udbg_hvc_getc_poll;
@@ -456,14 +449,12 @@
 	 * backend for HVSI, only do udbg
 	 */
 	if (hvterm_priv0.proto == HV_PROTOCOL_HVSI)
-		goto out;
+		return;
 #endif
 	/* Check whether the user has requested a different console. */
 	if (!strstr(cmd_line, "console="))
 		add_preferred_console("hvc", 0, NULL);
 	hvc_instantiate(0, 0, ops);
-out:
-	of_node_put(stdout_node);
 }
 
 /* call this from early_init() for a working debug console on
diff --git a/drivers/tty/serial/pmac_zilog.c b/drivers/tty/serial/pmac_zilog.c
index f7ad5b9..abbfedb 100644
--- a/drivers/tty/serial/pmac_zilog.c
+++ b/drivers/tty/serial/pmac_zilog.c
@@ -1653,8 +1653,7 @@
 	/*
 	 * Find all escc chips in the system
 	 */
-	node_p = of_find_node_by_name(NULL, "escc");
-	while (node_p) {
+	for_each_node_by_name(node_p, "escc") {
 		/*
 		 * First get channel A/B node pointers
 		 * 
@@ -1672,7 +1671,7 @@
 			of_node_put(node_b);
 			printk(KERN_ERR "pmac_zilog: missing node %c for escc %s\n",
 				(!node_a) ? 'a' : 'b', node_p->full_name);
-			goto next;
+			continue;
 		}
 
 		/*
@@ -1699,11 +1698,9 @@
 			of_node_put(node_b);
 			memset(&pmz_ports[count], 0, sizeof(struct uart_pmac_port));
 			memset(&pmz_ports[count+1], 0, sizeof(struct uart_pmac_port));
-			goto next;
+			continue;
 		}
 		count += 2;
-next:
-		node_p = of_find_node_by_name(node_p, "escc");
 	}
 	pmz_ports_count = count;
 
diff --git a/drivers/tty/serial/serial_core.c b/drivers/tty/serial/serial_core.c
index 8bb19da..29a7be4 100644
--- a/drivers/tty/serial/serial_core.c
+++ b/drivers/tty/serial/serial_core.c
@@ -26,6 +26,7 @@
 #include <linux/slab.h>
 #include <linux/init.h>
 #include <linux/console.h>
+#include <linux/of.h>
 #include <linux/proc_fs.h>
 #include <linux/seq_file.h>
 #include <linux/device.h>
@@ -2611,6 +2612,8 @@
 		spin_lock_init(&uport->lock);
 		lockdep_set_class(&uport->lock, &port_lock_key);
 	}
+	if (uport->cons && uport->dev)
+		of_console_check(uport->dev->of_node, uport->cons->name, uport->line);
 
 	uart_configure_port(drv, state, uport);
 
diff --git a/drivers/vfio/Kconfig b/drivers/vfio/Kconfig
index af7b204..d8c5763 100644
--- a/drivers/vfio/Kconfig
+++ b/drivers/vfio/Kconfig
@@ -8,11 +8,17 @@
 	depends on VFIO && SPAPR_TCE_IOMMU
 	default n
 
+config VFIO_SPAPR_EEH
+	tristate
+	depends on EEH && VFIO_IOMMU_SPAPR_TCE
+	default n
+
 menuconfig VFIO
 	tristate "VFIO Non-Privileged userspace driver framework"
 	depends on IOMMU_API
 	select VFIO_IOMMU_TYPE1 if X86
 	select VFIO_IOMMU_SPAPR_TCE if (PPC_POWERNV || PPC_PSERIES)
+	select VFIO_SPAPR_EEH if (PPC_POWERNV || PPC_PSERIES)
 	select ANON_INODES
 	help
 	  VFIO provides a framework for secure userspace device drivers.
diff --git a/drivers/vfio/Makefile b/drivers/vfio/Makefile
index 50e30bc..0b035b1 100644
--- a/drivers/vfio/Makefile
+++ b/drivers/vfio/Makefile
@@ -1,5 +1,5 @@
 obj-$(CONFIG_VFIO) += vfio.o
 obj-$(CONFIG_VFIO_IOMMU_TYPE1) += vfio_iommu_type1.o
 obj-$(CONFIG_VFIO_IOMMU_SPAPR_TCE) += vfio_iommu_spapr_tce.o
-obj-$(CONFIG_EEH) += vfio_spapr_eeh.o
+obj-$(CONFIG_VFIO_SPAPR_EEH) += vfio_spapr_eeh.o
 obj-$(CONFIG_VFIO_PCI) += pci/
diff --git a/drivers/vfio/pci/vfio_pci.c b/drivers/vfio/pci/vfio_pci.c
index e2ee80f..f782533 100644
--- a/drivers/vfio/pci/vfio_pci.c
+++ b/drivers/vfio/pci/vfio_pci.c
@@ -37,6 +37,10 @@
 MODULE_PARM_DESC(nointxmask,
 		  "Disable support for PCI 2.3 style INTx masking.  If this resolves problems for specific devices, report lspci -vvvxxx to linux-pci@vger.kernel.org so the device can be fixed automatically via the broken_intx_masking flag.");
 
+static DEFINE_MUTEX(driver_lock);
+
+static void vfio_pci_try_bus_reset(struct vfio_pci_device *vdev);
+
 static int vfio_pci_enable(struct vfio_pci_device *vdev)
 {
 	struct pci_dev *pdev = vdev->pdev;
@@ -44,6 +48,9 @@
 	u16 cmd;
 	u8 msix_pos;
 
+	/* Don't allow our initial saved state to include busmaster */
+	pci_clear_master(pdev);
+
 	ret = pci_enable_device(pdev);
 	if (ret)
 		return ret;
@@ -99,7 +106,8 @@
 	struct pci_dev *pdev = vdev->pdev;
 	int bar;
 
-	pci_disable_device(pdev);
+	/* Stop the device from further DMA */
+	pci_clear_master(pdev);
 
 	vfio_pci_set_irqs_ioctl(vdev, VFIO_IRQ_SET_DATA_NONE |
 				VFIO_IRQ_SET_ACTION_TRIGGER,
@@ -117,6 +125,8 @@
 		vdev->barmap[bar] = NULL;
 	}
 
+	vdev->needs_reset = true;
+
 	/*
 	 * If we have saved state, restore it.  If we can reset the device,
 	 * even better.  Resetting with current state seems better than
@@ -128,7 +138,7 @@
 			__func__, dev_name(&pdev->dev));
 
 		if (!vdev->reset_works)
-			return;
+			goto out;
 
 		pci_save_state(pdev);
 	}
@@ -148,46 +158,55 @@
 		if (ret)
 			pr_warn("%s: Failed to reset device %s (%d)\n",
 				__func__, dev_name(&pdev->dev), ret);
+		else
+			vdev->needs_reset = false;
 	}
 
 	pci_restore_state(pdev);
+out:
+	pci_disable_device(pdev);
+
+	vfio_pci_try_bus_reset(vdev);
 }
 
 static void vfio_pci_release(void *device_data)
 {
 	struct vfio_pci_device *vdev = device_data;
 
-	if (atomic_dec_and_test(&vdev->refcnt)) {
+	mutex_lock(&driver_lock);
+
+	if (!(--vdev->refcnt)) {
 		vfio_spapr_pci_eeh_release(vdev->pdev);
 		vfio_pci_disable(vdev);
 	}
 
+	mutex_unlock(&driver_lock);
+
 	module_put(THIS_MODULE);
 }
 
 static int vfio_pci_open(void *device_data)
 {
 	struct vfio_pci_device *vdev = device_data;
-	int ret;
+	int ret = 0;
 
 	if (!try_module_get(THIS_MODULE))
 		return -ENODEV;
 
-	if (atomic_inc_return(&vdev->refcnt) == 1) {
+	mutex_lock(&driver_lock);
+
+	if (!vdev->refcnt) {
 		ret = vfio_pci_enable(vdev);
 		if (ret)
 			goto error;
 
-		ret = vfio_spapr_pci_eeh_open(vdev->pdev);
-		if (ret) {
-			vfio_pci_disable(vdev);
-			goto error;
-		}
+		vfio_spapr_pci_eeh_open(vdev->pdev);
 	}
-
-	return 0;
+	vdev->refcnt++;
 error:
-	module_put(THIS_MODULE);
+	mutex_unlock(&driver_lock);
+	if (ret)
+		module_put(THIS_MODULE);
 	return ret;
 }
 
@@ -843,7 +862,6 @@
 	vdev->irq_type = VFIO_PCI_NUM_IRQS;
 	mutex_init(&vdev->igate);
 	spin_lock_init(&vdev->irqlock);
-	atomic_set(&vdev->refcnt, 0);
 
 	ret = vfio_add_group_dev(&pdev->dev, &vfio_pci_ops, vdev);
 	if (ret) {
@@ -858,12 +876,15 @@
 {
 	struct vfio_pci_device *vdev;
 
-	vdev = vfio_del_group_dev(&pdev->dev);
-	if (!vdev)
-		return;
+	mutex_lock(&driver_lock);
 
-	iommu_group_put(pdev->dev.iommu_group);
-	kfree(vdev);
+	vdev = vfio_del_group_dev(&pdev->dev);
+	if (vdev) {
+		iommu_group_put(pdev->dev.iommu_group);
+		kfree(vdev);
+	}
+
+	mutex_unlock(&driver_lock);
 }
 
 static pci_ers_result_t vfio_pci_aer_err_detected(struct pci_dev *pdev,
@@ -906,6 +927,110 @@
 	.err_handler	= &vfio_err_handlers,
 };
 
+/*
+ * Test whether a reset is necessary and possible.  We mark devices as
+ * needs_reset when they are released, but don't have a function-local reset
+ * available.  If any of these exist in the affected devices, we want to do
+ * a bus/slot reset.  We also need all of the affected devices to be unused,
+ * so we abort if any device has a non-zero refcnt.  driver_lock prevents a
+ * device from being opened during the scan or unbound from vfio-pci.
+ */
+static int vfio_pci_test_bus_reset(struct pci_dev *pdev, void *data)
+{
+	bool *needs_reset = data;
+	struct pci_driver *pci_drv = ACCESS_ONCE(pdev->driver);
+	int ret = -EBUSY;
+
+	if (pci_drv == &vfio_pci_driver) {
+		struct vfio_device *device;
+		struct vfio_pci_device *vdev;
+
+		device = vfio_device_get_from_dev(&pdev->dev);
+		if (!device)
+			return ret;
+
+		vdev = vfio_device_data(device);
+		if (vdev) {
+			if (vdev->needs_reset)
+				*needs_reset = true;
+
+			if (!vdev->refcnt)
+				ret = 0;
+		}
+
+		vfio_device_put(device);
+	}
+
+	/*
+	 * TODO: vfio-core considers groups to be viable even if some devices
+	 * are attached to known drivers, like pci-stub or pcieport.  We can't
+	 * freeze devices from being unbound to those drivers like we can
+	 * here though, so it would be racy to test for them.  We also can't
+	 * use device_lock() to prevent changes as that would interfere with
+	 * PCI-core taking device_lock during bus reset.  For now, we require
+	 * devices to be bound to vfio-pci to get a bus/slot reset on release.
+	 */
+
+	return ret;
+}
+
+/* Clear needs_reset on all affected devices after successful bus/slot reset */
+static int vfio_pci_clear_needs_reset(struct pci_dev *pdev, void *data)
+{
+	struct pci_driver *pci_drv = ACCESS_ONCE(pdev->driver);
+
+	if (pci_drv == &vfio_pci_driver) {
+		struct vfio_device *device;
+		struct vfio_pci_device *vdev;
+
+		device = vfio_device_get_from_dev(&pdev->dev);
+		if (!device)
+			return 0;
+
+		vdev = vfio_device_data(device);
+		if (vdev)
+			vdev->needs_reset = false;
+
+		vfio_device_put(device);
+	}
+
+	return 0;
+}
+
+/*
+ * Attempt to do a bus/slot reset if there are devices affected by a reset for
+ * this device that are needs_reset and all of the affected devices are unused
+ * (!refcnt).  Callers of this function are required to hold driver_lock such
+ * that devices can not be unbound from vfio-pci or opened by a user while we
+ * test for and perform a bus/slot reset.
+ */
+static void vfio_pci_try_bus_reset(struct vfio_pci_device *vdev)
+{
+	bool needs_reset = false, slot = false;
+	int ret;
+
+	if (!pci_probe_reset_slot(vdev->pdev->slot))
+		slot = true;
+	else if (pci_probe_reset_bus(vdev->pdev->bus))
+		return;
+
+	if (vfio_pci_for_each_slot_or_bus(vdev->pdev,
+					  vfio_pci_test_bus_reset,
+					  &needs_reset, slot) || !needs_reset)
+		return;
+
+	if (slot)
+		ret = pci_try_reset_slot(vdev->pdev->slot);
+	else
+		ret = pci_try_reset_bus(vdev->pdev->bus);
+
+	if (ret)
+		return;
+
+	vfio_pci_for_each_slot_or_bus(vdev->pdev,
+				      vfio_pci_clear_needs_reset, NULL, slot);
+}
+
 static void __exit vfio_pci_cleanup(void)
 {
 	pci_unregister_driver(&vfio_pci_driver);
diff --git a/drivers/vfio/pci/vfio_pci_private.h b/drivers/vfio/pci/vfio_pci_private.h
index 9c6d5d0..671c17a 100644
--- a/drivers/vfio/pci/vfio_pci_private.h
+++ b/drivers/vfio/pci/vfio_pci_private.h
@@ -54,8 +54,9 @@
 	bool			extended_caps;
 	bool			bardirty;
 	bool			has_vga;
+	bool			needs_reset;
 	struct pci_saved_state	*pci_saved_state;
-	atomic_t		refcnt;
+	int			refcnt;
 	struct eventfd_ctx	*err_trigger;
 };
 
diff --git a/drivers/vfio/vfio_spapr_eeh.c b/drivers/vfio/vfio_spapr_eeh.c
index f834b4c..86dfceb 100644
--- a/drivers/vfio/vfio_spapr_eeh.c
+++ b/drivers/vfio/vfio_spapr_eeh.c
@@ -9,20 +9,27 @@
  * published by the Free Software Foundation.
  */
 
+#include <linux/module.h>
 #include <linux/uaccess.h>
 #include <linux/vfio.h>
 #include <asm/eeh.h>
 
+#define DRIVER_VERSION	"0.1"
+#define DRIVER_AUTHOR	"Gavin Shan, IBM Corporation"
+#define DRIVER_DESC	"VFIO IOMMU SPAPR EEH"
+
 /* We might build address mapping here for "fast" path later */
-int vfio_spapr_pci_eeh_open(struct pci_dev *pdev)
+void vfio_spapr_pci_eeh_open(struct pci_dev *pdev)
 {
-	return eeh_dev_open(pdev);
+	eeh_dev_open(pdev);
 }
+EXPORT_SYMBOL_GPL(vfio_spapr_pci_eeh_open);
 
 void vfio_spapr_pci_eeh_release(struct pci_dev *pdev)
 {
 	eeh_dev_release(pdev);
 }
+EXPORT_SYMBOL_GPL(vfio_spapr_pci_eeh_release);
 
 long vfio_spapr_iommu_eeh_ioctl(struct iommu_group *group,
 				unsigned int cmd, unsigned long arg)
@@ -85,3 +92,9 @@
 
 	return ret;
 }
+EXPORT_SYMBOL(vfio_spapr_iommu_eeh_ioctl);
+
+MODULE_VERSION(DRIVER_VERSION);
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR(DRIVER_AUTHOR);
+MODULE_DESCRIPTION(DRIVER_DESC);
diff --git a/include/linux/bio.h b/include/linux/bio.h
index d2633ee..b39e500 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -308,6 +308,7 @@
 
 	unsigned short		bip_slab;	/* slab the bip came from */
 	unsigned short		bip_vcnt;	/* # of integrity bio_vecs */
+	unsigned short		bip_max_vcnt;	/* integrity bio_vec slots */
 	unsigned		bip_owns_buf:1;	/* should free bip_buf */
 
 	struct work_struct	bip_work;	/* I/O completion */
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 8699bcf..518b465 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -21,6 +21,7 @@
 #include <linux/bsg.h>
 #include <linux/smp.h>
 #include <linux/rcupdate.h>
+#include <linux/percpu-refcount.h>
 
 #include <asm/scatterlist.h>
 
@@ -470,6 +471,7 @@
 	struct mutex		sysfs_lock;
 
 	int			bypass_depth;
+	int			mq_freeze_depth;
 
 #if defined(CONFIG_BLK_DEV_BSG)
 	bsg_job_fn		*bsg_job_fn;
@@ -483,7 +485,7 @@
 #endif
 	struct rcu_head		rcu_head;
 	wait_queue_head_t	mq_freeze_wq;
-	struct percpu_counter	mq_usage_counter;
+	struct percpu_ref	mq_usage_counter;
 	struct list_head	all_q_node;
 
 	struct blk_mq_tag_set	*tag_set;
diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index 3dbe9bd..debb70d 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -52,7 +52,7 @@
 #endif
 
 extern const char *drbd_buildtag(void);
-#define REL_VERSION "8.4.3"
+#define REL_VERSION "8.4.5"
 #define API_VERSION 1
 #define PRO_VERSION_MIN 86
 #define PRO_VERSION_MAX 101
@@ -245,7 +245,7 @@
 	D_DISKLESS,
 	D_ATTACHING,      /* In the process of reading the meta-data */
 	D_FAILED,         /* Becomes D_DISKLESS as soon as we told it the peer */
-			/* when >= D_FAILED it is legal to access mdev->bc */
+			  /* when >= D_FAILED it is legal to access mdev->ldev */
 	D_NEGOTIATING,    /* Late attaching state, we need to talk to the peer */
 	D_INCONSISTENT,
 	D_OUTDATED,
diff --git a/include/linux/drbd_genl.h b/include/linux/drbd_genl.h
index 4193f5f..7b131ed 100644
--- a/include/linux/drbd_genl.h
+++ b/include/linux/drbd_genl.h
@@ -171,6 +171,10 @@
 	__flg_field(28, DRBD_GENLA_F_MANDATORY | DRBD_F_INVARIANT,	tentative)
 	__flg_field_def(29,	DRBD_GENLA_F_MANDATORY,	use_rle, DRBD_USE_RLE_DEF)
 	/* 9: __u32_field_def(30,	DRBD_GENLA_F_MANDATORY,	fencing_policy, DRBD_FENCING_DEF) */
+	/* 9: __str_field_def(31,     DRBD_GENLA_F_MANDATORY, name, SHARED_SECRET_MAX) */
+	/* 9: __u32_field(32,         DRBD_F_REQUIRED | DRBD_F_INVARIANT,     peer_node_id) */
+	__flg_field_def(33, 0 /* OPTIONAL */,	csums_after_crash_only, DRBD_CSUMS_AFTER_CRASH_ONLY_DEF)
+	__u32_field_def(34, 0 /* OPTIONAL */, sock_check_timeo, DRBD_SOCKET_CHECK_TIMEO_DEF)
 )
 
 GENL_struct(DRBD_NLA_SET_ROLE_PARMS, 6, set_role_parms,
diff --git a/include/linux/drbd_limits.h b/include/linux/drbd_limits.h
index 17e50bb..8ac8c5d 100644
--- a/include/linux/drbd_limits.h
+++ b/include/linux/drbd_limits.h
@@ -214,6 +214,7 @@
 #define DRBD_ALLOW_TWO_PRIMARIES_DEF	0
 #define DRBD_ALWAYS_ASBP_DEF	0
 #define DRBD_USE_RLE_DEF	1
+#define DRBD_CSUMS_AFTER_CRASH_ONLY_DEF 0
 
 #define DRBD_AL_STRIPES_MIN     1
 #define DRBD_AL_STRIPES_MAX     1024
@@ -224,4 +225,9 @@
 #define DRBD_AL_STRIPE_SIZE_MAX   16777216
 #define DRBD_AL_STRIPE_SIZE_DEF   32
 #define DRBD_AL_STRIPE_SIZE_SCALE 'k' /* kilobytes */
+
+#define DRBD_SOCKET_CHECK_TIMEO_MIN 0
+#define DRBD_SOCKET_CHECK_TIMEO_MAX DRBD_PING_TIMEO_MAX
+#define DRBD_SOCKET_CHECK_TIMEO_DEF 0
+#define DRBD_SOCKET_CHECK_TIMEO_SCALE '1'
 #endif
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index c845036..379c026 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -116,6 +116,7 @@
 	/* special QP and management commands */
 	MLX4_CMD_CONF_SPECIAL_QP = 0x23,
 	MLX4_CMD_MAD_IFC	 = 0x24,
+	MLX4_CMD_MAD_DEMUX	 = 0x203,
 
 	/* multicast commands */
 	MLX4_CMD_READ_MCG	 = 0x25,
@@ -186,6 +187,12 @@
 };
 
 enum {
+	MLX4_CMD_MAD_DEMUX_CONFIG	= 0,
+	MLX4_CMD_MAD_DEMUX_QUERY_STATE	= 1,
+	MLX4_CMD_MAD_DEMUX_QUERY_RESTR	= 2, /* Query mad demux restrictions */
+};
+
+enum {
 	MLX4_CMD_WRAPPED,
 	MLX4_CMD_NATIVE
 };
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index e15b154..071f6b2 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -183,6 +183,7 @@
 	MLX4_DEV_CAP_FLAG2_UPDATE_QP		= 1LL <<  8,
 	MLX4_DEV_CAP_FLAG2_DMFS_IPOIB		= 1LL <<  9,
 	MLX4_DEV_CAP_FLAG2_VXLAN_OFFLOADS	= 1LL <<  10,
+	MLX4_DEV_CAP_FLAG2_MAD_DEMUX		= 1LL <<  11,
 };
 
 enum {
@@ -273,6 +274,7 @@
 	MLX4_PERM_REMOTE_WRITE	= 1 << 13,
 	MLX4_PERM_ATOMIC	= 1 << 14,
 	MLX4_PERM_BIND_MW	= 1 << 15,
+	MLX4_PERM_MASK		= 0xFC00
 };
 
 enum {
@@ -1254,6 +1256,21 @@
 int mlx4_vf_get_enable_smi_admin(struct mlx4_dev *dev, int slave, int port);
 int mlx4_vf_set_enable_smi_admin(struct mlx4_dev *dev, int slave, int port,
 				 int enable);
+int mlx4_mr_hw_get_mpt(struct mlx4_dev *dev, struct mlx4_mr *mmr,
+		       struct mlx4_mpt_entry ***mpt_entry);
+int mlx4_mr_hw_write_mpt(struct mlx4_dev *dev, struct mlx4_mr *mmr,
+			 struct mlx4_mpt_entry **mpt_entry);
+int mlx4_mr_hw_change_pd(struct mlx4_dev *dev, struct mlx4_mpt_entry *mpt_entry,
+			 u32 pdn);
+int mlx4_mr_hw_change_access(struct mlx4_dev *dev,
+			     struct mlx4_mpt_entry *mpt_entry,
+			     u32 access);
+void mlx4_mr_hw_put_mpt(struct mlx4_dev *dev,
+			struct mlx4_mpt_entry **mpt_entry);
+void mlx4_mr_rereg_mem_cleanup(struct mlx4_dev *dev, struct mlx4_mr *mr);
+int mlx4_mr_rereg_mem_write(struct mlx4_dev *dev, struct mlx4_mr *mr,
+			    u64 iova, u64 size, int npages,
+			    int page_shift, struct mlx4_mpt_entry *mpt_entry);
 
 /* Returns true if running in low memory profile (kdump kernel) */
 static inline bool mlx4_low_memory_profile(void)
diff --git a/include/linux/mmc/dw_mmc.h b/include/linux/mmc/dw_mmc.h
index babaea9..29ce014 100644
--- a/include/linux/mmc/dw_mmc.h
+++ b/include/linux/mmc/dw_mmc.h
@@ -213,6 +213,8 @@
 #define DW_MCI_QUIRK_HIGHSPEED			BIT(2)
 /* Unreliable card detection */
 #define DW_MCI_QUIRK_BROKEN_CARD_DETECTION	BIT(3)
+/* No write protect */
+#define DW_MCI_QUIRK_NO_WRITE_PROTECT		BIT(4)
 
 /* Slot level quirks */
 /* This slot has no write protect */
diff --git a/include/linux/mmc/sdhci.h b/include/linux/mmc/sdhci.h
index 08abe99..09ebe57 100644
--- a/include/linux/mmc/sdhci.h
+++ b/include/linux/mmc/sdhci.h
@@ -104,9 +104,6 @@
 
 	const struct sdhci_ops *ops;	/* Low level hw interface */
 
-	struct regulator *vmmc;		/* Power regulator (vmmc) */
-	struct regulator *vqmmc;	/* Signaling regulator (vccq) */
-
 	/* Internal data */
 	struct mmc_host *mmc;	/* MMC structure */
 	u64 dma_mask;		/* custom DMA mask */
diff --git a/include/linux/of.h b/include/linux/of.h
index 196b34c..6c4363b 100644
--- a/include/linux/of.h
+++ b/include/linux/of.h
@@ -74,8 +74,6 @@
 	uint32_t args[MAX_PHANDLE_ARGS];
 };
 
-extern int of_node_add(struct device_node *node);
-
 /* initialize a node */
 extern struct kobj_type of_node_ktype;
 static inline void of_node_init(struct device_node *node)
@@ -113,6 +111,7 @@
 extern struct device_node *of_allnodes;
 extern struct device_node *of_chosen;
 extern struct device_node *of_aliases;
+extern struct device_node *of_stdout;
 extern raw_spinlock_t devtree_lock;
 
 static inline bool of_have_populated_dt(void)
@@ -204,6 +203,7 @@
 #define OF_DYNAMIC	1 /* node and properties were allocated via kmalloc */
 #define OF_DETACHED	2 /* node has been detached from the device tree */
 #define OF_POPULATED	3 /* device already created for the node */
+#define OF_POPULATED_BUS	4 /* of_platform_populate recursed to children of this node */
 
 #define OF_IS_DYNAMIC(x) test_bit(OF_DYNAMIC, &x->_flags)
 #define OF_MARK_DYNAMIC(x) set_bit(OF_DYNAMIC, &x->_flags)
@@ -322,6 +322,7 @@
 struct of_prop_reconfig {
 	struct device_node	*dn;
 	struct property		*prop;
+	struct property		*old_prop;
 };
 
 extern int of_reconfig_notifier_register(struct notifier_block *);
@@ -352,7 +353,7 @@
  */
 const char *of_prop_next_string(struct property *prop, const char *cur);
 
-int of_device_is_stdout_path(struct device_node *dn);
+bool of_console_check(struct device_node *dn, char *name, int index);
 
 #else /* CONFIG_OF */
 
@@ -564,9 +565,9 @@
 	return 0;
 }
 
-static inline int of_device_is_stdout_path(struct device_node *dn)
+static inline bool of_console_check(const struct device_node *dn, const char *name, int index)
 {
-	return 0;
+	return false;
 }
 
 static inline const __be32 *of_prop_next_u32(struct property *prop,
@@ -786,4 +787,80 @@
 #define OF_DECLARE_2(table, name, compat, fn) \
 		_OF_DECLARE(table, name, compat, fn, of_init_fn_2)
 
+/**
+ * struct of_changeset_entry	- Holds a changeset entry
+ *
+ * @node:	list_head for the log list
+ * @action:	notifier action
+ * @np:		pointer to the device node affected
+ * @prop:	pointer to the property affected
+ * @old_prop:	hold a pointer to the original property
+ *
+ * Every modification of the device tree during a changeset
+ * is held in a list of of_changeset_entry structures.
+ * That way we can recover from a partial application, or we can
+ * revert the changeset
+ */
+struct of_changeset_entry {
+	struct list_head node;
+	unsigned long action;
+	struct device_node *np;
+	struct property *prop;
+	struct property *old_prop;
+};
+
+/**
+ * struct of_changeset - changeset tracker structure
+ *
+ * @entries:	list_head for the changeset entries
+ *
+ * changesets are a convenient way to apply bulk changes to the
+ * live tree. In case of an error, changes are rolled-back.
+ * changesets live on after initial application, and if not
+ * destroyed after use, they can be reverted in one single call.
+ */
+struct of_changeset {
+	struct list_head entries;
+};
+
+#ifdef CONFIG_OF_DYNAMIC
+extern void of_changeset_init(struct of_changeset *ocs);
+extern void of_changeset_destroy(struct of_changeset *ocs);
+extern int of_changeset_apply(struct of_changeset *ocs);
+extern int of_changeset_revert(struct of_changeset *ocs);
+extern int of_changeset_action(struct of_changeset *ocs,
+		unsigned long action, struct device_node *np,
+		struct property *prop);
+
+static inline int of_changeset_attach_node(struct of_changeset *ocs,
+		struct device_node *np)
+{
+	return of_changeset_action(ocs, OF_RECONFIG_ATTACH_NODE, np, NULL);
+}
+
+static inline int of_changeset_detach_node(struct of_changeset *ocs,
+		struct device_node *np)
+{
+	return of_changeset_action(ocs, OF_RECONFIG_DETACH_NODE, np, NULL);
+}
+
+static inline int of_changeset_add_property(struct of_changeset *ocs,
+		struct device_node *np, struct property *prop)
+{
+	return of_changeset_action(ocs, OF_RECONFIG_ADD_PROPERTY, np, prop);
+}
+
+static inline int of_changeset_remove_property(struct of_changeset *ocs,
+		struct device_node *np, struct property *prop)
+{
+	return of_changeset_action(ocs, OF_RECONFIG_REMOVE_PROPERTY, np, prop);
+}
+
+static inline int of_changeset_update_property(struct of_changeset *ocs,
+		struct device_node *np, struct property *prop)
+{
+	return of_changeset_action(ocs, OF_RECONFIG_UPDATE_PROPERTY, np, prop);
+}
+#endif
+
 #endif /* _LINUX_OF_H */
diff --git a/include/linux/of_platform.h b/include/linux/of_platform.h
index d96e1ba..c2b0627 100644
--- a/include/linux/of_platform.h
+++ b/include/linux/of_platform.h
@@ -72,7 +72,7 @@
 				const struct of_device_id *matches,
 				const struct of_dev_auxdata *lookup,
 				struct device *parent);
-extern int of_platform_depopulate(struct device *parent);
+extern void of_platform_depopulate(struct device *parent);
 #else
 static inline int of_platform_populate(struct device_node *root,
 					const struct of_device_id *matches,
@@ -81,10 +81,7 @@
 {
 	return -ENODEV;
 }
-static inline int of_platform_depopulate(struct device *parent)
-{
-	return -ENODEV;
-}
+static inline void of_platform_depopulate(struct device *parent) { }
 #endif
 
 #endif	/* _LINUX_OF_PLATFORM_H */
diff --git a/include/linux/of_reserved_mem.h b/include/linux/of_reserved_mem.h
index 4669ddf..5b5efae 100644
--- a/include/linux/of_reserved_mem.h
+++ b/include/linux/of_reserved_mem.h
@@ -8,6 +8,7 @@
 struct reserved_mem {
 	const char			*name;
 	unsigned long			fdt_node;
+	unsigned long			phandle;
 	const struct reserved_mem_ops	*ops;
 	phys_addr_t			base;
 	phys_addr_t			size;
@@ -27,10 +28,16 @@
 	_OF_DECLARE(reservedmem, name, compat, init, reservedmem_of_init_fn)
 
 #ifdef CONFIG_OF_RESERVED_MEM
+void of_reserved_mem_device_init(struct device *dev);
+void of_reserved_mem_device_release(struct device *dev);
+
 void fdt_init_reserved_mem(void);
 void fdt_reserved_mem_save_node(unsigned long node, const char *uname,
 			       phys_addr_t base, phys_addr_t size);
 #else
+static inline void of_reserved_mem_device_init(struct device *dev) { }
+static inline void of_reserved_mem_device_release(struct device *pdev) { }
+
 static inline void fdt_init_reserved_mem(void) { }
 static inline void fdt_reserved_mem_save_node(unsigned long node,
 		const char *uname, phys_addr_t base, phys_addr_t size) { }
diff --git a/include/linux/platform_data/mmc-omap.h b/include/linux/platform_data/mmc-omap.h
index 2bf1b30..51e70cf 100644
--- a/include/linux/platform_data/mmc-omap.h
+++ b/include/linux/platform_data/mmc-omap.h
@@ -28,6 +28,7 @@
  */
 #define OMAP_HSMMC_SUPPORTS_DUAL_VOLT		BIT(0)
 #define OMAP_HSMMC_BROKEN_MULTIBLOCK_READ	BIT(1)
+#define OMAP_HSMMC_SWAKEUP_MISSING		BIT(2)
 
 struct mmc_card;
 
diff --git a/include/linux/printk.h b/include/linux/printk.h
index 0990997..d78125f 100644
--- a/include/linux/printk.h
+++ b/include/linux/printk.h
@@ -10,6 +10,9 @@
 extern const char linux_banner[];
 extern const char linux_proc_banner[];
 
+extern char *log_buf_addr_get(void);
+extern u32 log_buf_len_get(void);
+
 static inline int printk_get_level(const char *buffer)
 {
 	if (buffer[0] == KERN_SOH_ASCII && buffer[1]) {
diff --git a/include/linux/vfio.h b/include/linux/vfio.h
index 25a0fbd..d320411 100644
--- a/include/linux/vfio.h
+++ b/include/linux/vfio.h
@@ -98,16 +98,16 @@
 extern long vfio_external_check_extension(struct vfio_group *group,
 					  unsigned long arg);
 
+struct pci_dev;
 #ifdef CONFIG_EEH
-extern int vfio_spapr_pci_eeh_open(struct pci_dev *pdev);
+extern void vfio_spapr_pci_eeh_open(struct pci_dev *pdev);
 extern void vfio_spapr_pci_eeh_release(struct pci_dev *pdev);
 extern long vfio_spapr_iommu_eeh_ioctl(struct iommu_group *group,
 				       unsigned int cmd,
 				       unsigned long arg);
 #else
-static inline int vfio_spapr_pci_eeh_open(struct pci_dev *pdev)
+static inline void vfio_spapr_pci_eeh_open(struct pci_dev *pdev)
 {
-	return 0;
 }
 
 static inline void vfio_spapr_pci_eeh_release(struct pci_dev *pdev)
diff --git a/include/rdma/ib_mad.h b/include/rdma/ib_mad.h
index 3d81b90..9bb99e9 100644
--- a/include/rdma/ib_mad.h
+++ b/include/rdma/ib_mad.h
@@ -40,6 +40,7 @@
 #include <linux/list.h>
 
 #include <rdma/ib_verbs.h>
+#include <uapi/rdma/ib_user_mad.h>
 
 /* Management base version */
 #define IB_MGMT_BASE_VERSION			1
@@ -355,9 +356,13 @@
  * @hi_tid: Access layer assigned transaction ID for this client.
  *   Unsolicited MADs sent by this client will have the upper 32-bits
  *   of their TID set to this value.
+ * @flags: registration flags
  * @port_num: Port number on which QP is registered
  * @rmpp_version: If set, indicates the RMPP version used by this agent.
  */
+enum {
+	IB_MAD_USER_RMPP = IB_USER_MAD_USER_RMPP,
+};
 struct ib_mad_agent {
 	struct ib_device	*device;
 	struct ib_qp		*qp;
@@ -367,6 +372,7 @@
 	ib_mad_snoop_handler	snoop_handler;
 	void			*context;
 	u32			hi_tid;
+	u32			flags;
 	u8			port_num;
 	u8			rmpp_version;
 };
@@ -426,6 +432,7 @@
  *   in the range from 0x30 to 0x4f. Otherwise not used.
  * @method_mask: The caller will receive unsolicited MADs for any method
  *   where @method_mask = 1.
+ *
  */
 struct ib_mad_reg_req {
 	u8	mgmt_class;
@@ -451,6 +458,7 @@
  * @recv_handler: The completion callback routine invoked for a received
  *   MAD.
  * @context: User specified context associated with the registration.
+ * @registration_flags: Registration flags to set for this agent
  */
 struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 					   u8 port_num,
@@ -459,7 +467,8 @@
 					   u8 rmpp_version,
 					   ib_mad_send_handler send_handler,
 					   ib_mad_recv_handler recv_handler,
-					   void *context);
+					   void *context,
+					   u32 registration_flags);
 
 enum ib_mad_snoop_flags {
 	/*IB_MAD_SNOOP_POSTED_SENDS	   = 1,*/
@@ -661,4 +670,11 @@
  */
 void ib_free_send_mad(struct ib_mad_send_buf *send_buf);
 
+/**
+ * ib_mad_kernel_rmpp_agent - Returns if the agent is performing RMPP.
+ * @agent: the agent in question
+ * @return: true if agent is performing rmpp, false otherwise.
+ */
+int ib_mad_kernel_rmpp_agent(struct ib_mad_agent *agent);
+
 #endif /* IB_MAD_H */
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 7ccef34..ed44cc0 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -1097,7 +1097,8 @@
 enum ib_mr_rereg_flags {
 	IB_MR_REREG_TRANS	= 1,
 	IB_MR_REREG_PD		= (1<<1),
-	IB_MR_REREG_ACCESS	= (1<<2)
+	IB_MR_REREG_ACCESS	= (1<<2),
+	IB_MR_REREG_SUPPORTED	= ((IB_MR_REREG_ACCESS << 1) - 1)
 };
 
 /**
@@ -1547,6 +1548,13 @@
 						  u64 virt_addr,
 						  int mr_access_flags,
 						  struct ib_udata *udata);
+	int			   (*rereg_user_mr)(struct ib_mr *mr,
+						    int flags,
+						    u64 start, u64 length,
+						    u64 virt_addr,
+						    int mr_access_flags,
+						    struct ib_pd *pd,
+						    struct ib_udata *udata);
 	int                        (*query_mr)(struct ib_mr *mr,
 					       struct ib_mr_attr *mr_attr);
 	int                        (*dereg_mr)(struct ib_mr *mr);
diff --git a/include/scsi/sg.h b/include/scsi/sg.h
index 9859355..750e5db 100644
--- a/include/scsi/sg.h
+++ b/include/scsi/sg.h
@@ -86,7 +86,9 @@
 #define SG_FLAG_MMAP_IO 4       /* request memory mapped IO */
 #define SG_FLAG_NO_DXFER 0x10000 /* no transfer of kernel buffers to/from */
 				/* user space (debug indirect IO) */
-#define SG_FLAG_Q_AT_TAIL 0x10  /* default is Q_AT_HEAD */
+/* defaults:: for sg driver: Q_AT_HEAD; for block layer: Q_AT_TAIL */
+#define SG_FLAG_Q_AT_TAIL 0x10
+#define SG_FLAG_Q_AT_HEAD 0x20
 
 /* following 'info' values are "or"-ed together */
 #define SG_INFO_OK_MASK 0x1
diff --git a/include/trace/events/bcache.h b/include/trace/events/bcache.h
index c9c3c04..981acf7 100644
--- a/include/trace/events/bcache.h
+++ b/include/trace/events/bcache.h
@@ -148,11 +148,13 @@
 );
 
 TRACE_EVENT(bcache_write,
-	TP_PROTO(struct bio *bio, bool writeback, bool bypass),
-	TP_ARGS(bio, writeback, bypass),
+	TP_PROTO(struct cache_set *c, u64 inode, struct bio *bio,
+		bool writeback, bool bypass),
+	TP_ARGS(c, inode, bio, writeback, bypass),
 
 	TP_STRUCT__entry(
-		__field(dev_t,		dev			)
+		__array(char,		uuid,	16		)
+		__field(u64,		inode			)
 		__field(sector_t,	sector			)
 		__field(unsigned int,	nr_sector		)
 		__array(char,		rwbs,	6		)
@@ -161,7 +163,8 @@
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev->bd_dev;
+		memcpy(__entry->uuid, c->sb.set_uuid, 16);
+		__entry->inode		= inode;
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->nr_sector	= bio->bi_iter.bi_size >> 9;
 		blk_fill_rwbs(__entry->rwbs, bio->bi_rw, bio->bi_iter.bi_size);
@@ -169,8 +172,8 @@
 		__entry->bypass = bypass;
 	),
 
-	TP_printk("%d,%d  %s %llu + %u hit %u bypass %u",
-		  MAJOR(__entry->dev), MINOR(__entry->dev),
+	TP_printk("%pU inode %llu  %s %llu + %u hit %u bypass %u",
+		  __entry->uuid, __entry->inode,
 		  __entry->rwbs, (unsigned long long)__entry->sector,
 		  __entry->nr_sector, __entry->writeback, __entry->bypass)
 );
@@ -258,9 +261,9 @@
 	TP_ARGS(b)
 );
 
-DEFINE_EVENT(btree_node, bcache_btree_node_alloc_fail,
-	TP_PROTO(struct btree *b),
-	TP_ARGS(b)
+DEFINE_EVENT(cache_set, bcache_btree_node_alloc_fail,
+	TP_PROTO(struct cache_set *c),
+	TP_ARGS(c)
 );
 
 DEFINE_EVENT(btree_node, bcache_btree_node_free,
diff --git a/include/trace/events/thp.h b/include/trace/events/thp.h
new file mode 100644
index 0000000..b59b065
--- /dev/null
+++ b/include/trace/events/thp.h
@@ -0,0 +1,88 @@
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM thp
+
+#if !defined(_TRACE_THP_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_THP_H
+
+#include <linux/types.h>
+#include <linux/tracepoint.h>
+
+TRACE_EVENT(hugepage_invalidate,
+
+	    TP_PROTO(unsigned long addr, unsigned long pte),
+	    TP_ARGS(addr, pte),
+	    TP_STRUCT__entry(
+		    __field(unsigned long, addr)
+		    __field(unsigned long, pte)
+		    ),
+
+	    TP_fast_assign(
+		    __entry->addr = addr;
+		    __entry->pte = pte;
+		    ),
+
+	    TP_printk("hugepage invalidate at addr 0x%lx and pte = 0x%lx",
+		      __entry->addr, __entry->pte)
+);
+
+TRACE_EVENT(hugepage_set_pmd,
+
+	    TP_PROTO(unsigned long addr, unsigned long pmd),
+	    TP_ARGS(addr, pmd),
+	    TP_STRUCT__entry(
+		    __field(unsigned long, addr)
+		    __field(unsigned long, pmd)
+		    ),
+
+	    TP_fast_assign(
+		    __entry->addr = addr;
+		    __entry->pmd = pmd;
+		    ),
+
+	    TP_printk("Set pmd with 0x%lx with 0x%lx", __entry->addr, __entry->pmd)
+);
+
+
+TRACE_EVENT(hugepage_update,
+
+	    TP_PROTO(unsigned long addr, unsigned long pte, unsigned long clr, unsigned long set),
+	    TP_ARGS(addr, pte, clr, set),
+	    TP_STRUCT__entry(
+		    __field(unsigned long, addr)
+		    __field(unsigned long, pte)
+		    __field(unsigned long, clr)
+		    __field(unsigned long, set)
+		    ),
+
+	    TP_fast_assign(
+		    __entry->addr = addr;
+		    __entry->pte = pte;
+		    __entry->clr = clr;
+		    __entry->set = set;
+
+		    ),
+
+	    TP_printk("hugepage update at addr 0x%lx and pte = 0x%lx clr = 0x%lx, set = 0x%lx", __entry->addr, __entry->pte, __entry->clr, __entry->set)
+);
+TRACE_EVENT(hugepage_splitting,
+
+	    TP_PROTO(unsigned long addr, unsigned long pte),
+	    TP_ARGS(addr, pte),
+	    TP_STRUCT__entry(
+		    __field(unsigned long, addr)
+		    __field(unsigned long, pte)
+		    ),
+
+	    TP_fast_assign(
+		    __entry->addr = addr;
+		    __entry->pte = pte;
+		    ),
+
+	    TP_printk("hugepage splitting at addr 0x%lx and pte = 0x%lx",
+		      __entry->addr, __entry->pte)
+);
+
+#endif /* _TRACE_THP_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>
diff --git a/include/uapi/linux/bsg.h b/include/uapi/linux/bsg.h
index 7a12e1c..02986cf 100644
--- a/include/uapi/linux/bsg.h
+++ b/include/uapi/linux/bsg.h
@@ -10,12 +10,13 @@
 #define BSG_SUB_PROTOCOL_SCSI_TRANSPORT	2
 
 /*
- * For flags member below
- * sg.h sg_io_hdr also has bits defined for it's flags member. However
- * none of these bits are implemented/used by bsg. The bits below are
- * allocated to not conflict with sg.h ones anyway.
+ * For flag constants below:
+ * sg.h sg_io_hdr also has bits defined for it's flags member. These
+ * two flag values (0x10 and 0x20) have the same meaning in sg.h . For
+ * bsg the BSG_FLAG_Q_AT_HEAD flag is ignored since it is the deafult.
  */
-#define BSG_FLAG_Q_AT_TAIL 0x10 /* default, == 0 at this bit, is Q_AT_HEAD */
+#define BSG_FLAG_Q_AT_TAIL 0x10 /* default is Q_AT_HEAD */
+#define BSG_FLAG_Q_AT_HEAD 0x20
 
 struct sg_io_v4 {
 	__s32 guard;		/* [i] 'Q' to differentiate from v3 */
diff --git a/include/uapi/linux/virtio_blk.h b/include/uapi/linux/virtio_blk.h
index 6d8e61c..9ad67b2 100644
--- a/include/uapi/linux/virtio_blk.h
+++ b/include/uapi/linux/virtio_blk.h
@@ -40,6 +40,7 @@
 #define VIRTIO_BLK_F_WCE	9	/* Writeback mode enabled after reset */
 #define VIRTIO_BLK_F_TOPOLOGY	10	/* Topology information is available */
 #define VIRTIO_BLK_F_CONFIG_WCE	11	/* Writeback mode available in config */
+#define VIRTIO_BLK_F_MQ		12	/* support more than one vq */
 
 #ifndef __KERNEL__
 /* Old (deprecated) name for VIRTIO_BLK_F_WCE. */
@@ -77,6 +78,10 @@
 
 	/* writeback mode (if VIRTIO_BLK_F_CONFIG_WCE) */
 	__u8 wce;
+	__u8 unused;
+
+	/* number of vqs, only available when VIRTIO_BLK_F_MQ is set */
+	__u16 num_queues;
 } __attribute__((packed));
 
 /*
diff --git a/include/uapi/rdma/ib_user_mad.h b/include/uapi/rdma/ib_user_mad.h
index d6fce1c..09f809f 100644
--- a/include/uapi/rdma/ib_user_mad.h
+++ b/include/uapi/rdma/ib_user_mad.h
@@ -191,6 +191,45 @@
 	__u8	rmpp_version;
 };
 
+/**
+ * ib_user_mad_reg_req2 - MAD registration request
+ *
+ * @id                 - Set by the _kernel_; used by userspace to identify the
+ *                       registered agent in future requests.
+ * @qpn                - Queue pair number; must be 0 or 1.
+ * @mgmt_class         - Indicates which management class of MADs should be
+ *                       receive by the caller.  This field is only required if
+ *                       the user wishes to receive unsolicited MADs, otherwise
+ *                       it should be 0.
+ * @mgmt_class_version - Indicates which version of MADs for the given
+ *                       management class to receive.
+ * @res                - Ignored.
+ * @flags              - additional registration flags; Must be in the set of
+ *                       flags defined in IB_USER_MAD_REG_FLAGS_CAP
+ * @method_mask        - The caller wishes to receive unsolicited MADs for the
+ *                       methods whose bit(s) is(are) set.
+ * @oui                - Indicates IEEE OUI to use when mgmt_class is a vendor
+ *                       class in the range from 0x30 to 0x4f. Otherwise not
+ *                       used.
+ * @rmpp_version       - If set, indicates the RMPP version to use.
+ */
+enum {
+	IB_USER_MAD_USER_RMPP = (1 << 0),
+};
+#define IB_USER_MAD_REG_FLAGS_CAP (IB_USER_MAD_USER_RMPP)
+struct ib_user_mad_reg_req2 {
+	__u32	id;
+	__u32	qpn;
+	__u8	mgmt_class;
+	__u8	mgmt_class_version;
+	__u16   res;
+	__u32   flags;
+	__u64   method_mask[2];
+	__u32   oui;
+	__u8	rmpp_version;
+	__u8	reserved[3];
+};
+
 #define IB_IOCTL_MAGIC		0x1b
 
 #define IB_USER_MAD_REGISTER_AGENT	_IOWR(IB_IOCTL_MAGIC, 1, \
@@ -200,4 +239,7 @@
 
 #define IB_USER_MAD_ENABLE_PKEY		_IO(IB_IOCTL_MAGIC, 3)
 
+#define IB_USER_MAD_REGISTER_AGENT2     _IOWR(IB_IOCTL_MAGIC, 4, \
+					      struct ib_user_mad_reg_req2)
+
 #endif /* IB_USER_MAD_H */
diff --git a/include/uapi/rdma/ib_user_verbs.h b/include/uapi/rdma/ib_user_verbs.h
index cbfdd4c..26daf55 100644
--- a/include/uapi/rdma/ib_user_verbs.h
+++ b/include/uapi/rdma/ib_user_verbs.h
@@ -276,6 +276,22 @@
 	__u32 rkey;
 };
 
+struct ib_uverbs_rereg_mr {
+	__u64 response;
+	__u32 mr_handle;
+	__u32 flags;
+	__u64 start;
+	__u64 length;
+	__u64 hca_va;
+	__u32 pd_handle;
+	__u32 access_flags;
+};
+
+struct ib_uverbs_rereg_mr_resp {
+	__u32 lkey;
+	__u32 rkey;
+};
+
 struct ib_uverbs_dereg_mr {
 	__u32 mr_handle;
 };
diff --git a/include/uapi/rdma/rdma_user_cm.h b/include/uapi/rdma/rdma_user_cm.h
index 99b80ab..3066718 100644
--- a/include/uapi/rdma/rdma_user_cm.h
+++ b/include/uapi/rdma/rdma_user_cm.h
@@ -34,6 +34,7 @@
 #define RDMA_USER_CM_H
 
 #include <linux/types.h>
+#include <linux/socket.h>
 #include <linux/in6.h>
 #include <rdma/ib_user_verbs.h>
 #include <rdma/ib_user_sa.h>
diff --git a/init/Kconfig b/init/Kconfig
index 44f9ed3..e84c642 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -268,7 +268,7 @@
 	help
 	  Enabling this option adds the system calls process_vm_readv and
 	  process_vm_writev which allow a process with the correct privileges
-	  to directly read from or write to to another process's address space.
+	  to directly read from or write to another process' address space.
 	  See the man page for more details.
 
 config FHANDLE
diff --git a/kernel/fork.c b/kernel/fork.c
index 1380d8a..0cf9cdb 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1105,7 +1105,7 @@
 	 * needed because this new task is not yet running and cannot
 	 * be racing exec.
 	 */
-	BUG_ON(!spin_is_locked(&current->sighand->siglock));
+	assert_spin_locked(&current->sighand->siglock);
 
 	/* Ref-count the new filter user, and assign it. */
 	get_seccomp_filter(current);
diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
index de1a6bb..e04c455 100644
--- a/kernel/printk/printk.c
+++ b/kernel/printk/printk.c
@@ -272,6 +272,18 @@
 static char *log_buf = __log_buf;
 static u32 log_buf_len = __LOG_BUF_LEN;
 
+/* Return log buffer address */
+char *log_buf_addr_get(void)
+{
+	return log_buf;
+}
+
+/* Return log buffer size */
+u32 log_buf_len_get(void)
+{
+	return log_buf_len;
+}
+
 /* human readable text of the record */
 static char *log_text(const struct printk_log *msg)
 {
diff --git a/kernel/seccomp.c b/kernel/seccomp.c
index 25b0043..44eb005 100644
--- a/kernel/seccomp.c
+++ b/kernel/seccomp.c
@@ -203,7 +203,7 @@
 
 static inline bool seccomp_may_assign_mode(unsigned long seccomp_mode)
 {
-	BUG_ON(!spin_is_locked(&current->sighand->siglock));
+	assert_spin_locked(&current->sighand->siglock);
 
 	if (current->seccomp.mode && current->seccomp.mode != seccomp_mode)
 		return false;
@@ -214,7 +214,7 @@
 static inline void seccomp_assign_mode(struct task_struct *task,
 				       unsigned long seccomp_mode)
 {
-	BUG_ON(!spin_is_locked(&task->sighand->siglock));
+	assert_spin_locked(&task->sighand->siglock);
 
 	task->seccomp.mode = seccomp_mode;
 	/*
@@ -253,7 +253,7 @@
 	struct task_struct *thread, *caller;
 
 	BUG_ON(!mutex_is_locked(&current->signal->cred_guard_mutex));
-	BUG_ON(!spin_is_locked(&current->sighand->siglock));
+	assert_spin_locked(&current->sighand->siglock);
 
 	/* Validate all threads being eligible for synchronization. */
 	caller = current;
@@ -294,7 +294,7 @@
 	struct task_struct *thread, *caller;
 
 	BUG_ON(!mutex_is_locked(&current->signal->cred_guard_mutex));
-	BUG_ON(!spin_is_locked(&current->sighand->siglock));
+	assert_spin_locked(&current->sighand->siglock);
 
 	/* Synchronize all threads. */
 	caller = current;
@@ -464,7 +464,7 @@
 	unsigned long total_insns;
 	struct seccomp_filter *walker;
 
-	BUG_ON(!spin_is_locked(&current->sighand->siglock));
+	assert_spin_locked(&current->sighand->siglock);
 
 	/* Validate resulting filter length. */
 	total_insns = filter->prog->len;
diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c
index f36b028..fb4a9c2 100644
--- a/kernel/time/timekeeping.c
+++ b/kernel/time/timekeeping.c
@@ -338,10 +338,11 @@
 
 static inline void update_vsyscall(struct timekeeper *tk)
 {
-	struct timespec xt;
+	struct timespec xt, wm;
 
 	xt = timespec64_to_timespec(tk_xtime(tk));
-	update_vsyscall_old(&xt, &tk->wall_to_monotonic, tk->tkr.clock, tk->tkr.mult,
+	wm = timespec64_to_timespec(tk->wall_to_monotonic);
+	update_vsyscall_old(&xt, &wm, tk->tkr.clock, tk->tkr.mult,
 			    tk->tkr.cycle_last);
 }
 
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index cb45f59..07c2832 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -143,6 +143,30 @@
 	  DEBUG_INFO build and compile times are reduced too.
 	  Only works with newer gcc versions.
 
+config DEBUG_INFO_SPLIT
+	bool "Produce split debuginfo in .dwo files"
+	depends on DEBUG_INFO
+	help
+	  Generate debug info into separate .dwo files. This significantly
+	  reduces the build directory size for builds with DEBUG_INFO,
+	  because it stores the information only once on disk in .dwo
+	  files instead of multiple times in object files and executables.
+	  In addition the debug information is also compressed.
+
+	  Requires recent gcc (4.7+) and recent gdb/binutils.
+	  Any tool that packages or reads debug information would need
+	  to know about the .dwo files and include them.
+	  Incompatible with older versions of ccache.
+
+config DEBUG_INFO_DWARF4
+	bool "Generate dwarf4 debuginfo"
+	depends on DEBUG_INFO
+	help
+	  Generate dwarf4 debug info. This requires recent versions
+	  of gcc and gdb. It makes the debug information larger.
+	  But it significantly improves the success of resolving
+	  variables in gdb on optimized code.
+
 config ENABLE_WARN_DEPRECATED
 	bool "Enable __deprecated logic"
 	default y
diff --git a/lib/lru_cache.c b/lib/lru_cache.c
index 4a83ecd..852c81e 100644
--- a/lib/lru_cache.c
+++ b/lib/lru_cache.c
@@ -169,7 +169,7 @@
 	return NULL;
 }
 
-void lc_free_by_index(struct lru_cache *lc, unsigned i)
+static void lc_free_by_index(struct lru_cache *lc, unsigned i)
 {
 	void *p = lc->lc_element[i];
 	WARN_ON(!p);
@@ -643,9 +643,10 @@
  * lc_dump - Dump a complete LRU cache to seq in textual form.
  * @lc: the lru cache to operate on
  * @seq: the &struct seq_file pointer to seq_printf into
- * @utext: user supplied "heading" or other info
+ * @utext: user supplied additional "heading" or other info
  * @detail: function pointer the user may provide to dump further details
- * of the object the lc_element is embedded in.
+ * of the object the lc_element is embedded in. May be NULL.
+ * Note: a leading space ' ' and trailing newline '\n' is implied.
  */
 void lc_seq_dump_details(struct seq_file *seq, struct lru_cache *lc, char *utext,
 	     void (*detail) (struct seq_file *, struct lc_element *))
@@ -654,16 +655,18 @@
 	struct lc_element *e;
 	int i;
 
-	seq_printf(seq, "\tnn: lc_number refcnt %s\n ", utext);
+	seq_printf(seq, "\tnn: lc_number (new nr) refcnt %s\n ", utext);
 	for (i = 0; i < nr_elements; i++) {
 		e = lc_element_by_index(lc, i);
-		if (e->lc_number == LC_FREE) {
-			seq_printf(seq, "\t%2d: FREE\n", i);
-		} else {
-			seq_printf(seq, "\t%2d: %4u %4u    ", i,
-				   e->lc_number, e->refcnt);
+		if (e->lc_number != e->lc_new_number)
+			seq_printf(seq, "\t%5d: %6d %8d %6d ",
+				i, e->lc_number, e->lc_new_number, e->refcnt);
+		else
+			seq_printf(seq, "\t%5d: %6d %-8s %6d ",
+				i, e->lc_number, "-\"-", e->refcnt);
+		if (detail)
 			detail(seq, e);
-		}
+		seq_putc(seq, '\n');
 	}
 }
 
diff --git a/mm/hugetlb_cgroup.c b/mm/hugetlb_cgroup.c
index 9aae6f4..9eebfad 100644
--- a/mm/hugetlb_cgroup.c
+++ b/mm/hugetlb_cgroup.c
@@ -275,6 +275,7 @@
 		ret = res_counter_memparse_write_strategy(buf, &val);
 		if (ret)
 			break;
+		val = ALIGN(val, 1ULL << huge_page_shift(&hstates[idx]));
 		ret = res_counter_set_limit(&h_cg->hugepage[idx], val);
 		break;
 	default:
diff --git a/scripts/Kbuild.include b/scripts/Kbuild.include
index 122f95c..8a9a4e1 100644
--- a/scripts/Kbuild.include
+++ b/scripts/Kbuild.include
@@ -215,11 +215,13 @@
 arg-check = $(if $(strip $(cmd_$@)),,1)
 endif
 
-# >'< substitution is for echo to work,
-# >$< substitution to preserve $ when reloading .cmd file
-# note: when using inline perl scripts [perl -e '...$$t=1;...']
-# in $(cmd_xxx) double $$ your perl vars
-make-cmd = $(subst \\,\\\\,$(subst \#,\\\#,$(subst $$,$$$$,$(call escsq,$(cmd_$(1))))))
+# Replace >$< with >$$< to preserve $ when reloading the .cmd file
+# (needed for make)
+# Replace >#< with >\#< to avoid starting a comment in the .cmd file
+# (needed for make)
+# Replace >'< with >'\''< to be able to enclose the whole string in '...'
+# (needed for the shell)
+make-cmd = $(call escsq,$(subst \#,\\\#,$(subst $$,$$$$,$(cmd_$(1)))))
 
 # Find any prerequisites that is newer than target or that does not exist.
 # PHONY targets skipped in both cases.
@@ -230,7 +232,7 @@
 if_changed = $(if $(strip $(any-prereq) $(arg-check)),                       \
 	@set -e;                                                             \
 	$(echo-cmd) $(cmd_$(1));                                             \
-	echo 'cmd_$@ := $(make-cmd)' > $(dot-target).cmd)
+	printf '%s\n' 'cmd_$@ := $(make-cmd)' > $(dot-target).cmd)
 
 # Execute the command and also postprocess generated .d dependencies file.
 if_changed_dep = $(if $(strip $(any-prereq) $(arg-check) ),                  \
diff --git a/scripts/Makefile.clean b/scripts/Makefile.clean
index 686cb0d..a651cee 100644
--- a/scripts/Makefile.clean
+++ b/scripts/Makefile.clean
@@ -40,8 +40,8 @@
 # build a list of files to remove, usually relative to the current
 # directory
 
-__clean-files	:= $(extra-y) $(always)                  \
-		   $(targets) $(clean-files)             \
+__clean-files	:= $(extra-y) $(extra-m) $(extra-)       \
+		   $(always) $(targets) $(clean-files)   \
 		   $(host-progs)                         \
 		   $(hostprogs-y) $(hostprogs-m) $(hostprogs-)
 
diff --git a/scripts/Makefile.extrawarn b/scripts/Makefile.extrawarn
index 6564350..f734033 100644
--- a/scripts/Makefile.extrawarn
+++ b/scripts/Makefile.extrawarn
@@ -26,16 +26,6 @@
 warning-1 += $(call cc-option, -Wunused-but-set-variable)
 warning-1 += $(call cc-disable-warning, missing-field-initializers)
 
-# Clang
-warning-1 += $(call cc-disable-warning, initializer-overrides)
-warning-1 += $(call cc-disable-warning, unused-value)
-warning-1 += $(call cc-disable-warning, format)
-warning-1 += $(call cc-disable-warning, unknown-warning-option)
-warning-1 += $(call cc-disable-warning, sign-compare)
-warning-1 += $(call cc-disable-warning, format-zero-length)
-warning-1 += $(call cc-disable-warning, uninitialized)
-warning-1 += $(call cc-option, -fcatch-undefined-behavior)
-
 warning-2 := -Waggregate-return
 warning-2 += -Wcast-align
 warning-2 += -Wdisabled-optimization
@@ -64,4 +54,15 @@
 endif
 
 KBUILD_CFLAGS += $(warning)
+else
+
+ifeq ($(COMPILER),clang)
+KBUILD_CFLAGS += $(call cc-disable-warning, initializer-overrides)
+KBUILD_CFLAGS += $(call cc-disable-warning, unused-value)
+KBUILD_CFLAGS += $(call cc-disable-warning, format)
+KBUILD_CFLAGS += $(call cc-disable-warning, unknown-warning-option)
+KBUILD_CFLAGS += $(call cc-disable-warning, sign-compare)
+KBUILD_CFLAGS += $(call cc-disable-warning, format-zero-length)
+KBUILD_CFLAGS += $(call cc-disable-warning, uninitialized)
+endif
 endif
diff --git a/scripts/Makefile.host b/scripts/Makefile.host
index 6689364..ab5980f 100644
--- a/scripts/Makefile.host
+++ b/scripts/Makefile.host
@@ -20,21 +20,12 @@
 # Will compile qconf as a C++ program, and menu as a C program.
 # They are linked as C++ code to the executable qconf
 
-# hostprogs-y := conf
-# conf-objs  := conf.o libkconfig.so
-# libkconfig-objs := expr.o type.o
-# Will create a shared library named libkconfig.so that consists of
-# expr.o and type.o (they are both compiled as C code and the object files
-# are made as position independent code).
-# conf.c is compiled as a C program, and conf.o is linked together with
-# libkconfig.so as the executable conf.
-# Note: Shared libraries consisting of C++ files are not supported
-
 __hostprogs := $(sort $(hostprogs-y) $(hostprogs-m))
 
 # C code
 # Executables compiled from a single .c file
-host-csingle	:= $(foreach m,$(__hostprogs),$(if $($(m)-objs),,$(m)))
+host-csingle	:= $(foreach m,$(__hostprogs), \
+			$(if $($(m)-objs)$($(m)-cxxobjs),,$(m)))
 
 # C executables linked based on several .o files
 host-cmulti	:= $(foreach m,$(__hostprogs),\
@@ -44,33 +35,17 @@
 host-cobjs	:= $(sort $(foreach m,$(__hostprogs),$($(m)-objs)))
 
 # C++ code
-# C++ executables compiled from at least on .cc file
+# C++ executables compiled from at least one .cc file
 # and zero or more .c files
 host-cxxmulti	:= $(foreach m,$(__hostprogs),$(if $($(m)-cxxobjs),$(m)))
 
 # C++ Object (.o) files compiled from .cc files
 host-cxxobjs	:= $(sort $(foreach m,$(host-cxxmulti),$($(m)-cxxobjs)))
 
-# Shared libaries (only .c supported)
-# Shared libraries (.so) - all .so files referenced in "xxx-objs"
-host-cshlib	:= $(sort $(filter %.so, $(host-cobjs)))
-# Remove .so files from "xxx-objs"
-host-cobjs	:= $(filter-out %.so,$(host-cobjs))
-
-#Object (.o) files used by the shared libaries
-host-cshobjs	:= $(sort $(foreach m,$(host-cshlib),$($(m:.so=-objs))))
-
 # output directory for programs/.o files
-# hostprogs-y := tools/build may have been specified. Retrieve directory
-host-objdirs := $(foreach f,$(__hostprogs), $(if $(dir $(f)),$(dir $(f))))
-# directory of .o files from prog-objs notation
-host-objdirs += $(foreach f,$(host-cmulti),                  \
-                    $(foreach m,$($(f)-objs),                \
-                        $(if $(dir $(m)),$(dir $(m)))))
-# directory of .o files from prog-cxxobjs notation
-host-objdirs += $(foreach f,$(host-cxxmulti),                  \
-                    $(foreach m,$($(f)-cxxobjs),                \
-                        $(if $(dir $(m)),$(dir $(m)))))
+# hostprogs-y := tools/build may have been specified.
+# Retrieve also directory of .o files from prog-objs or prog-cxxobjs notation
+host-objdirs := $(dir $(__hostprogs) $(host-cobjs) $(host-cxxobjs))
 
 host-objdirs := $(strip $(sort $(filter-out ./,$(host-objdirs))))
 
@@ -81,8 +56,6 @@
 host-cobjs	:= $(addprefix $(obj)/,$(host-cobjs))
 host-cxxmulti	:= $(addprefix $(obj)/,$(host-cxxmulti))
 host-cxxobjs	:= $(addprefix $(obj)/,$(host-cxxobjs))
-host-cshlib	:= $(addprefix $(obj)/,$(host-cshlib))
-host-cshobjs	:= $(addprefix $(obj)/,$(host-cshobjs))
 host-objdirs    := $(addprefix $(obj)/,$(host-objdirs))
 
 obj-dirs += $(host-objdirs)
@@ -123,7 +96,7 @@
       cmd_host-cmulti	= $(HOSTCC) $(HOSTLDFLAGS) -o $@ \
 			  $(addprefix $(obj)/,$($(@F)-objs)) \
 			  $(HOST_LOADLIBES) $(HOSTLOADLIBES_$(@F))
-$(host-cmulti): $(obj)/%: $(host-cobjs) $(host-cshlib) FORCE
+$(host-cmulti): $(obj)/%: $(host-cobjs) FORCE
 	$(call if_changed,host-cmulti)
 
 # Create .o file from a single .c file
@@ -140,7 +113,7 @@
 			  $(foreach o,objs cxxobjs,\
 			  $(addprefix $(obj)/,$($(@F)-$(o)))) \
 			  $(HOST_LOADLIBES) $(HOSTLOADLIBES_$(@F))
-$(host-cxxmulti): $(obj)/%: $(host-cobjs) $(host-cxxobjs) $(host-cshlib) FORCE
+$(host-cxxmulti): $(obj)/%: $(host-cobjs) $(host-cxxobjs) FORCE
 	$(call if_changed,host-cxxmulti)
 
 # Create .o file from a single .cc (C++) file
@@ -149,21 +122,5 @@
 $(host-cxxobjs): $(obj)/%.o: $(src)/%.cc FORCE
 	$(call if_changed_dep,host-cxxobjs)
 
-# Compile .c file, create position independent .o file
-# host-cshobjs -> .o
-quiet_cmd_host-cshobjs	= HOSTCC  -fPIC $@
-      cmd_host-cshobjs	= $(HOSTCC) $(hostc_flags) -fPIC -c -o $@ $<
-$(host-cshobjs): $(obj)/%.o: $(src)/%.c FORCE
-	$(call if_changed_dep,host-cshobjs)
-
-# Link a shared library, based on position independent .o files
-# *.o -> .so shared library (host-cshlib)
-quiet_cmd_host-cshlib	= HOSTLLD -shared $@
-      cmd_host-cshlib	= $(HOSTCC) $(HOSTLDFLAGS) -shared -o $@ \
-			  $(addprefix $(obj)/,$($(@F:.so=-objs))) \
-			  $(HOST_LOADLIBES) $(HOSTLOADLIBES_$(@F))
-$(host-cshlib): $(obj)/%: $(host-cshobjs) FORCE
-	$(call if_changed,host-cshlib)
-
 targets += $(host-csingle)  $(host-cmulti) $(host-cobjs)\
-	   $(host-cxxmulti) $(host-cxxobjs) $(host-cshlib) $(host-cshobjs)
+	   $(host-cxxmulti) $(host-cxxobjs)
diff --git a/scripts/coccinelle/api/alloc/alloc_cast.cocci b/scripts/coccinelle/api/alloc/alloc_cast.cocci
new file mode 100644
index 0000000..6c308ee
--- /dev/null
+++ b/scripts/coccinelle/api/alloc/alloc_cast.cocci
@@ -0,0 +1,72 @@
+/// Remove casting the values returned by memory allocation functions
+/// like kmalloc, kzalloc, kmem_cache_alloc, kmem_cache_zalloc etc.
+///
+//# This makes an effort to find cases of casting of values returned by
+//# kmalloc, kzalloc, kcalloc, kmem_cache_alloc, kmem_cache_zalloc,
+//# kmem_cache_alloc_node, kmalloc_node and kzalloc_node and removes
+//# the casting as it is not required. The result in the patch case may
+//#need some reformatting.
+//
+// Confidence: High
+// Copyright: 2014, Himangi Saraogi  GPLv2.
+// Comments:
+// Options: --no-includes --include-headers
+//
+
+virtual context
+virtual patch
+virtual org
+virtual report
+
+//----------------------------------------------------------
+//  For context mode
+//----------------------------------------------------------
+
+@depends on context@
+type T;
+@@
+
+* (T *)
+  \(kmalloc\|kzalloc\|kcalloc\|kmem_cache_alloc\|kmem_cache_zalloc\|
+   kmem_cache_alloc_node\|kmalloc_node\|kzalloc_node\)(...)
+
+//----------------------------------------------------------
+//  For patch mode
+//----------------------------------------------------------
+
+@depends on patch@
+type T;
+@@
+
+- (T *)
+  (\(kmalloc\|kzalloc\|kcalloc\|kmem_cache_alloc\|kmem_cache_zalloc\|
+   kmem_cache_alloc_node\|kmalloc_node\|kzalloc_node\)(...))
+
+//----------------------------------------------------------
+//  For org and report mode
+//----------------------------------------------------------
+
+@r depends on org || report@
+type T;
+position p;
+@@
+
+ (T@p *)\(kmalloc\|kzalloc\|kcalloc\|kmem_cache_alloc\|kmem_cache_zalloc\|
+   kmem_cache_alloc_node\|kmalloc_node\|kzalloc_node\)(...)
+
+@script:python depends on org@
+p << r.p;
+t << r.T;
+@@
+
+coccilib.org.print_safe_todo(p[0], t)
+
+@script:python depends on report@
+p << r.p;
+t << r.T;
+@@
+
+msg="WARNING: casting value returned by memory allocation function to (%s *) is useless." % (t)
+coccilib.report.print_report(p[0], msg)
+
+
diff --git a/scripts/coccinelle/api/alloc/drop_kmalloc_cast.cocci b/scripts/coccinelle/api/alloc/drop_kmalloc_cast.cocci
deleted file mode 100644
index bd5d08b..0000000
--- a/scripts/coccinelle/api/alloc/drop_kmalloc_cast.cocci
+++ /dev/null
@@ -1,67 +0,0 @@
-///
-/// Casting (void *) value returned by kmalloc is useless
-/// as mentioned in Documentation/CodingStyle, Chap 14.
-///
-// Confidence: High
-// Copyright: 2009,2010 Nicolas Palix, DIKU.  GPLv2.
-// URL: http://coccinelle.lip6.fr/
-// Options: --no-includes --include-headers
-//
-// Keywords: kmalloc, kzalloc, kcalloc
-// Version min: < 2.6.12 kmalloc
-// Version min: < 2.6.12 kcalloc
-// Version min:   2.6.14 kzalloc
-//
-
-virtual context
-virtual patch
-virtual org
-virtual report
-
-//----------------------------------------------------------
-//  For context mode
-//----------------------------------------------------------
-
-@depends on context@
-type T;
-@@
-
-* (T *)
-  \(kmalloc\|kzalloc\|kcalloc\)(...)
-
-//----------------------------------------------------------
-//  For patch mode
-//----------------------------------------------------------
-
-@depends on patch@
-type T;
-@@
-
-- (T *)
-  \(kmalloc\|kzalloc\|kcalloc\)(...)
-
-//----------------------------------------------------------
-//  For org and report mode
-//----------------------------------------------------------
-
-@r depends on org || report@
-type T;
-position p;
-@@
-
- (T@p *)\(kmalloc\|kzalloc\|kcalloc\)(...)
-
-@script:python depends on org@
-p << r.p;
-t << r.T;
-@@
-
-coccilib.org.print_safe_todo(p[0], t)
-
-@script:python depends on report@
-p << r.p;
-t << r.T;
-@@
-
-msg="WARNING: casting value returned by k[cmz]alloc to (%s *) is useless." % (t)
-coccilib.report.print_report(p[0], msg)
diff --git a/scripts/coccinelle/misc/array_size.cocci b/scripts/coccinelle/misc/array_size.cocci
new file mode 100644
index 0000000..81e279c
--- /dev/null
+++ b/scripts/coccinelle/misc/array_size.cocci
@@ -0,0 +1,87 @@
+/// Use ARRAY_SIZE instead of dividing sizeof array with sizeof an element
+///
+//# This makes an effort to find cases where ARRAY_SIZE can be used such as
+//# where there is a division of sizeof the array by the sizeof its first
+//# element or by any indexed element or the element type. It replaces the
+//# division of the two sizeofs by ARRAY_SIZE.
+//
+// Confidence: High
+// Copyright: (C) 2014 Himangi Saraogi.  GPLv2.
+// Comments:
+// Options: --no-includes --include-headers
+
+virtual patch
+virtual context
+virtual org
+virtual report
+
+@i@
+@@
+
+#include <linux/kernel.h>
+
+//----------------------------------------------------------
+//  For context mode
+//----------------------------------------------------------
+
+@depends on i&&context@
+type T;
+T[] E;
+@@
+(
+* (sizeof(E)/sizeof(*E))
+|
+* (sizeof(E)/sizeof(E[...]))
+|
+* (sizeof(E)/sizeof(T))
+)
+
+//----------------------------------------------------------
+//  For patch mode
+//----------------------------------------------------------
+
+@depends on i&&patch@
+type T;
+T[] E;
+@@
+(
+- (sizeof(E)/sizeof(*E))
++ ARRAY_SIZE(E)
+|
+- (sizeof(E)/sizeof(E[...]))
++ ARRAY_SIZE(E)
+|
+- (sizeof(E)/sizeof(T))
++ ARRAY_SIZE(E)
+)
+
+//----------------------------------------------------------
+//  For org and report mode
+//----------------------------------------------------------
+
+@r@
+type T;
+T[] E;
+position p;
+@@
+(
+ (sizeof(E)@p /sizeof(*E))
+|
+ (sizeof(E)@p /sizeof(E[...]))
+|
+ (sizeof(E)@p /sizeof(T))
+)
+
+@script:python depends on i&&org@
+p << r.p;
+@@
+
+coccilib.org.print_todo(p[0], "WARNING should use ARRAY_SIZE")
+
+@script:python depends on i&&report@
+p << r.p;
+@@
+
+msg="WARNING: Use ARRAY_SIZE"
+coccilib.report.print_report(p[0], msg)
+
diff --git a/scripts/coccinelle/misc/badty.cocci b/scripts/coccinelle/misc/badty.cocci
new file mode 100644
index 0000000..2fc06fc
--- /dev/null
+++ b/scripts/coccinelle/misc/badty.cocci
@@ -0,0 +1,76 @@
+/// Use ARRAY_SIZE instead of dividing sizeof array with sizeof an element
+///
+//# This makes an effort to find cases where the argument to sizeof is wrong
+//# in memory allocation functions by checking the type of the allocated memory
+//# when it is a double pointer and ensuring the sizeof argument takes a pointer
+//# to the the memory being allocated. There are false positives in cases the
+//# sizeof argument is not used in constructing the return value. The result
+//# may need some reformatting.
+//
+// Confidence: Moderate
+// Copyright: (C) 2014 Himangi Saraogi.  GPLv2.
+// Comments:
+// Options:
+
+virtual patch
+virtual context
+virtual org
+virtual report
+
+//----------------------------------------------------------
+//  For context mode
+//----------------------------------------------------------
+
+@depends on context disable sizeof_type_expr@
+type T;
+T **x;
+@@
+
+  x =
+  <+...sizeof(
+* T
+  )...+>
+
+//----------------------------------------------------------
+//  For patch mode
+//----------------------------------------------------------
+
+@depends on patch disable sizeof_type_expr@
+type T;
+T **x;
+@@
+
+  x =
+  <+...sizeof(
+- T
++ *x
+  )...+>
+
+//----------------------------------------------------------
+//  For org and report mode
+//----------------------------------------------------------
+
+@r disable sizeof_type_expr@
+type T;
+T **x;
+position p;
+@@
+
+  x =
+  <+...sizeof(
+  T@p
+  )...+>
+
+@script:python depends on org@
+p << r.p;
+@@
+
+coccilib.org.print_todo(p[0], "WARNING sizeof argument should be pointer type, not structure type")
+
+@script:python depends on report@
+p << r.p;
+@@
+
+msg="WARNING: Use correct pointer type argument for sizeof"
+coccilib.report.print_report(p[0], msg)
+
diff --git a/scripts/coccinelle/misc/bugon.cocci b/scripts/coccinelle/misc/bugon.cocci
new file mode 100644
index 0000000..556456c
--- /dev/null
+++ b/scripts/coccinelle/misc/bugon.cocci
@@ -0,0 +1,62 @@
+/// Use BUG_ON instead of a if condition followed by BUG.
+///
+//# This makes an effort to find cases where BUG() follows an if
+//# condition on an expression and replaces the if condition and BUG()
+//# with a BUG_ON having the conditional expression of the if statement
+//# as argument.
+//
+// Confidence: High
+// Copyright: (C) 2014 Himangi Saraogi.  GPLv2.
+// Comments:
+// Options: --no-includes, --include-headers
+
+virtual patch
+virtual context
+virtual org
+virtual report
+
+//----------------------------------------------------------
+//  For context mode
+//----------------------------------------------------------
+
+@depends on context@
+expression e;
+@@
+
+*if (e) BUG();
+
+//----------------------------------------------------------
+//  For patch mode
+//----------------------------------------------------------
+
+@depends on patch@
+expression e;
+@@
+
+-if (e) BUG();
++BUG_ON(e);
+
+//----------------------------------------------------------
+//  For org and report mode
+//----------------------------------------------------------
+
+@r@
+expression e;
+position p;
+@@
+
+ if (e) BUG@p ();
+
+@script:python depends on org@
+p << r.p;
+@@
+
+coccilib.org.print_todo(p[0], "WARNING use BUG_ON")
+
+@script:python depends on report@
+p << r.p;
+@@
+
+msg="WARNING: Use BUG_ON"
+coccilib.report.print_report(p[0], msg)
+
diff --git a/scripts/coccinelle/null/badzero.cocci b/scripts/coccinelle/null/badzero.cocci
index d79baf7..5551da2 100644
--- a/scripts/coccinelle/null/badzero.cocci
+++ b/scripts/coccinelle/null/badzero.cocci
@@ -10,7 +10,7 @@
 // Copyright: (C) 2012 Julia Lawall, INRIA/LIP6.  GPLv2.
 // Copyright: (C) 2012 Gilles Muller, INRIA/LiP6.  GPLv2.
 // URL: http://coccinelle.lip6.fr/
-// Comments:
+// Comments: Requires Coccinelle version 1.0.0-rc20 or later
 // Options:
 
 virtual patch
@@ -19,6 +19,7 @@
 virtual report
 
 @initialize:ocaml@
+@@
 let negtable = Hashtbl.create 101
 
 @depends on patch@
diff --git a/sound/ppc/pmac.c b/sound/ppc/pmac.c
index 7a43c0c..8a431bc 100644
--- a/sound/ppc/pmac.c
+++ b/sound/ppc/pmac.c
@@ -992,9 +992,9 @@
 		return -ENODEV;
 
 	if (!sound) {
-		sound = of_find_node_by_name(NULL, "sound");
-		while (sound && sound->parent != chip->node)
-			sound = of_find_node_by_name(sound, "sound");
+		for_each_node_by_name(sound, "sound")
+			if (sound->parent == chip->node)
+				break;
 	}
 	if (! sound) {
 		of_node_put(chip->node);