)]}' { "log": [ { "commit": "0ad2507d5d93f39619fc42372c347d6006b64319", "tree": "039de7928561cddf233dde5e3d91e2cd3cf87524", "parents": [ "224e74511041c2d33dcf24adf54d447eb72399c8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 14:02:44 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 14:02:44 2025 -0800" }, "message": "Linux 6.14-rc3\n" }, { "commit": "224e74511041c2d33dcf24adf54d447eb72399c8", "tree": "8a9fe82db3d6693131f3ebdc7d9429588702f9fa", "parents": [ "ae5fa8ce7ef230be50ed63d16883bada29990293", "b28fb1f2ef45eeef1cd2c23149b50d184d545a3e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 12:58:51 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 12:58:51 2025 -0800" }, "message": "Merge tag \u0027kbuild-fixes-v6.14-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild\n\nPull Kbuild fixes from Masahiro Yamada:\n\n - Fix annoying logs when building tools in parallel\n\n - Fix the Debian linux-headers package build again\n\n - Fix the target triple detection for userspace programs on Clang\n\n* tag \u0027kbuild-fixes-v6.14-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:\n modpost: Fix a few typos in a comment\n kbuild: userprogs: fix bitsize and target detection on clang\n kbuild: fix linux-headers package build when $(CC) cannot link userspace\n tools: fix annoying \"mkdir -p ...\" logs when building tools in parallel\n" }, { "commit": "ae5fa8ce7ef230be50ed63d16883bada29990293", "tree": "a550fa2c85b10b2a3a6844d6f783b47ce527dd4e", "parents": [ "56400391b1d312a14f912f36d2f04b0dba0d4a2f", "78418f300d3999f1cf8a9ac71065bf2eca61f4dd" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 12:54:42 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 12:54:42 2025 -0800" }, "message": "Merge tag \u0027driver-core-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core\n\nPull driver core api addition from Greg KH:\n \"Here is a driver core new api for 6.14-rc3 that is being added to\n allow platform devices from stop being abused.\n\n It adds a new \u0027faux_device\u0027 structure and bus and api to allow almost\n a straight or simpler conversion from platform devices that were not\n really a platform device. It also comes with a binding for rust, with\n an example driver in rust showing how it\u0027s used.\n\n I\u0027m adding this now so that the patches that convert the different\n drivers and subsystems can all start flowing into linux-next now\n through their different development trees, in time for 6.15-rc1.\n\n We have a number that are already reviewed and tested, but adding\n those conversions now doesn\u0027t seem right. For now, no one is using\n this, and it passes all build tests from 0-day and linux-next, so all\n should be good\"\n\n* tag \u0027driver-core-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:\n rust/kernel: Add faux device bindings\n driver core: add a faux bus for use when a simple device/bus is needed\n" }, { "commit": "56400391b1d312a14f912f36d2f04b0dba0d4a2f", "tree": "1669937468ef024cab27247aaf31a7241c10fef2", "parents": [ "6bfcc5fb2fcee1e914708ce55b70e707d03d8097", "9e512eaaf8f4008c44ede3dfc0fbc9d9c5118583" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 12:50:44 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 12:50:44 2025 -0800" }, "message": "Merge tag \u0027tty-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty\n\nPull serial driver fixes from Greg KH:\n \"Here are some small serial driver fixes for some reported problems.\n Nothing major, just:\n\n - sc16is7xx irq check fix\n\n - 8250 fifo underflow fix\n\n - serial_port and 8250 iotype fixes\n\n Most of these have been in linux-next already, and all have passed\n 0-day testing\"\n\n* tag \u0027tty-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:\n serial: 8250: Fix fifo underflow on flush\n serial: 8250_pnp: Remove unneeded -\u003eiotype assignment\n serial: 8250_platform: Remove unneeded -\u003eiotype assignment\n serial: 8250_of: Remove unneeded -\u003eiotype assignment\n serial: port: Make -\u003eiotype validation global in __uart_read_properties()\n serial: port: Always update -\u003eiotype in __uart_read_properties()\n serial: port: Assign -\u003eiotype correctly when -\u003eiobase is set\n serial: sc16is7xx: Fix IRQ number check behavior\n" }, { "commit": "6bfcc5fb2fcee1e914708ce55b70e707d03d8097", "tree": "1124339d471a3220d1ecd2f33a062833a804371f", "parents": [ "ba643b6d84409e8a9057d5bdd6dd99255b1a88fe", "659f5d55feb75782bd46cf130da3c1f240afe9ba" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 11:15:50 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 11:15:50 2025 -0800" }, "message": "Merge tag \u0027usb-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb\n\nPull USB fixes from Greg KH:\n \"Here are some small USB driver fixes, and new device ids, for\n 6.14-rc3. Lots of tiny stuff for reported problems, including:\n\n - new device ids and quirks\n\n - usb hub crash fix found by syzbot\n\n - dwc2 driver fix\n\n - dwc3 driver fixes\n\n - uvc gadget driver fix\n\n - cdc-acm driver fixes for a variety of different issues\n\n - other tiny bugfixes\n\n Almost all of these have been in linux-next this week, and all have\n passed 0-day testing\"\n\n* tag \u0027usb-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (25 commits)\n usb: typec: tcpm: PSSourceOffTimer timeout in PR_Swap enters ERROR_RECOVERY\n usb: roles: set switch registered flag early on\n usb: gadget: uvc: Fix unstarted kthread worker\n USB: quirks: add USB_QUIRK_NO_LPM quirk for Teclast dist\n usb: gadget: core: flush gadget workqueue after device removal\n USB: gadget: f_midi: f_midi_complete to call queue_work\n usb: core: fix pipe creation for get_bMaxPacketSize0\n usb: dwc3: Fix timeout issue during controller enter/exit from halt state\n USB: Add USB_QUIRK_NO_LPM quirk for sony xperia xz1 smartphone\n USB: cdc-acm: Fill in Renesas R-Car D3 USB Download mode quirk\n usb: cdc-acm: Fix handling of oversized fragments\n usb: cdc-acm: Check control transfer buffer size before access\n usb: xhci: Restore xhci_pci support for Renesas HCs\n USB: pci-quirks: Fix HCCPARAMS register error for LS7A EHCI\n USB: serial: option: drop MeiG Smart defines\n USB: serial: option: fix Telit Cinterion FN990A name\n USB: serial: option: add Telit Cinterion FN990B compositions\n USB: serial: option: add MeiG Smart SLM828\n usb: gadget: f_midi: fix MIDI Streaming descriptor lengths\n usb: dwc2: gadget: remove of_node reference upon udc_stop\n ...\n" }, { "commit": "ba643b6d84409e8a9057d5bdd6dd99255b1a88fe", "tree": "4759a7fff2a81679311e89798729f998b3924970", "parents": [ "ff1848d81ca63e49bce57c29ffb12c6539333ac4", "4cf7d58620bfc2ebe934e3dfa97208f13f14ab8b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:55:17 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:55:17 2025 -0800" }, "message": "Merge tag \u0027irq_urgent_for_v6.14_rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull irq Kconfig cleanup from Borislav Petkov:\n\n - Remove an unused config item GENERIC_PENDING_IRQ_CHIPFLAGS\n\n* tag \u0027irq_urgent_for_v6.14_rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n genirq: Remove unused CONFIG_GENERIC_PENDING_IRQ_CHIPFLAGS\n" }, { "commit": "ff1848d81ca63e49bce57c29ffb12c6539333ac4", "tree": "85209d457aeeaf620700d05e9ecb81b820dae1f6", "parents": [ "ff3b373ecc88098103b6ae80d96d6527210b2e64", "c631a2de7ae48d50434bdc205d901423f8577c65" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:41:50 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:41:50 2025 -0800" }, "message": "Merge tag \u0027perf_urgent_for_v6.14_rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 perf fixes from Borislav Petkov:\n\n - Explicitly clear DEBUGCTL.LBR to prevent LBRs continuing being\n enabled after handoff to the OS\n\n - Check CPUID(0x23) leaf and subleafs presence properly\n\n - Remove the PEBS-via-PT feature from being supported on hybrid systems\n\n - Fix perf record/top default commands on systems without a raw PMU\n registered\n\n* tag \u0027perf_urgent_for_v6.14_rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n perf/x86/intel: Ensure LBRs are disabled when a CPU is starting\n perf/x86/intel: Fix ARCH_PERFMON_NUM_COUNTER_LEAF\n perf/x86/intel: Clean up PEBS-via-PT on hybrid\n perf/x86/rapl: Fix the error checking order\n" }, { "commit": "ff3b373ecc88098103b6ae80d96d6527210b2e64", "tree": "edb8c7f6a97c149d1d3e39895e33c96e86469ec0", "parents": [ "592c358ea963d7227e3e7b8ace91b2b5dd81f6cb", "bcc6244e13b4d4903511a1ea84368abf925031c0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:38:24 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:38:24 2025 -0800" }, "message": "Merge tag \u0027sched_urgent_for_v6.14_rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull scheduler fix from Borislav Petkov:\n\n - Clarify what happens when a task is woken up from the wake queue and\n make clear its removal from that queue is atomic\n\n* tag \u0027sched_urgent_for_v6.14_rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n sched: Clarify wake_up_q()\u0027s write to task-\u003ewake_q.next\n" }, { "commit": "592c358ea963d7227e3e7b8ace91b2b5dd81f6cb", "tree": "c2d130b5b8eb8f8e6da0ac95fcb075214e4fc876", "parents": [ "82ff31645685559e3732f7143538c9fe88221453", "7e501637bd5b702a2fa627e903a0025654110e1e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:30:58 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:30:58 2025 -0800" }, "message": "Merge tag \u0027objtool_urgent_for_v6.14_rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull objtool fixes from Borislav Petkov:\n\n - Move a warning about a lld.ld breakage into the verbose setting as\n said breakage has been fixed in the meantime\n\n - Teach objtool to ignore dangling jump table entries added by Clang\n\n* tag \u0027objtool_urgent_for_v6.14_rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n objtool: Move dodgy linker warn to verbose\n objtool: Ignore dangling jump table entries\n" }, { "commit": "82ff31645685559e3732f7143538c9fe88221453", "tree": "78c74a67eae91032767c7d798911f364260f9e8f", "parents": [ "b878a1c072a4912e14a38e90a8b1883d4c513d9d", "d3d0b8dfe06098d6d584266c35e9a0947f5b7132" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:25:12 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:25:12 2025 -0800" }, "message": "Merge tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/virt/kvm/kvm\n\nPull kvm fixes from Paolo Bonzini:\n \"ARM:\n\n - Large set of fixes for vector handling, especially in the\n interactions between host and guest state.\n\n This fixes a number of bugs affecting actual deployments, and\n greatly simplifies the FP/SIMD/SVE handling. Thanks to Mark Rutland\n for dealing with this thankless task.\n\n - Fix an ugly race between vcpu and vgic creation/init, resulting in\n unexpected behaviours\n\n - Fix use of kernel VAs at EL2 when emulating timers with nVHE\n\n - Small set of pKVM improvements and cleanups\n\n x86:\n\n - Fix broken SNP support with KVM module built-in, ensuring the PSP\n module is initialized before KVM even when the module\n infrastructure cannot be used to order initcalls\n\n - Reject Hyper-V SEND_IPI hypercalls if the local APIC isn\u0027t being\n emulated by KVM to fix a NULL pointer dereference\n\n - Enter guest mode (L2) from KVM\u0027s perspective before initializing\n the vCPU\u0027s nested NPT MMU so that the MMU is properly tagged for\n L2, not L1\n\n - Load the guest\u0027s DR6 outside of the innermost .vcpu_run() loop, as\n the guest\u0027s value may be stale if a VM-Exit is handled in the\n fastpath\"\n\n* tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/virt/kvm/kvm: (25 commits)\n x86/sev: Fix broken SNP support with KVM module built-in\n KVM: SVM: Ensure PSP module is initialized if KVM module is built-in\n crypto: ccp: Add external API interface for PSP module initialization\n KVM: arm64: vgic: Hoist SGI/PPI alloc from vgic_init() to kvm_create_vgic()\n KVM: arm64: timer: Drop warning on failed interrupt signalling\n KVM: arm64: Fix alignment of kvm_hyp_memcache allocations\n KVM: arm64: Convert timer offset VA when accessed in HYP code\n KVM: arm64: Simplify warning in kvm_arch_vcpu_load_fp()\n KVM: arm64: Eagerly switch ZCR_EL{1,2}\n KVM: arm64: Mark some header functions as inline\n KVM: arm64: Refactor exit handlers\n KVM: arm64: Refactor CPTR trap deactivation\n KVM: arm64: Remove VHE host restore of CPACR_EL1.SMEN\n KVM: arm64: Remove VHE host restore of CPACR_EL1.ZEN\n KVM: arm64: Remove host FPSIMD saving for non-protected KVM\n KVM: arm64: Unconditionally save+flush host FPSIMD/SVE/SME state\n KVM: x86: Load DR6 with guest value only before entering .vcpu_run() loop\n KVM: nSVM: Enter guest mode before initializing nested NPT MMU\n KVM: selftests: Add CPUID tests for Hyper-V features that need in-kernel APIC\n KVM: selftests: Manage CPUID array in Hyper-V CPUID test\u0027s core helper\n ...\n" }, { "commit": "b878a1c072a4912e14a38e90a8b1883d4c513d9d", "tree": "de39433d4f367054f9d7af059ac15fe9da134d2f", "parents": [ "ad1b832bf1cf2df9304f8eb72943111625c7e5a7", "733a90561ad0a4a74035d2d627098da85d43b592" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:19:41 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Feb 16 10:19:41 2025 -0800" }, "message": "Merge tag \u0027mips-fixes_6.14_1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux\n\nPull MIPS fixes from Thomas Bogendoerfer:\n \"Fix for o32 ptrace/get_syscall_info\"\n\n* tag \u0027mips-fixes_6.14_1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:\n MIPS: fix mips_get_syscall_arg() for o32\n MIPS: Export syscall stack arguments properly for remote use\n" }, { "commit": "ad1b832bf1cf2df9304f8eb72943111625c7e5a7", "tree": "a2de268c868cef9f098298937189b66f64846cd4", "parents": [ "ad73b9a17d66366d8c9198bc90f1ea99f24a912c", "038e33fcd40e59b60cdca561c2a39998e6759e08" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 17:20:39 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 17:20:39 2025 -0800" }, "message": "Merge tag \u0027devicetree-fixes-for-6.14-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux\n\nPull devicetree fixes from Rob Herring:\n\n - Add bindings for QCom QCS8300 clocks, QCom SAR2130P qfprom, and\n powertip,{st7272|hx8238a} displays\n\n - Fix compatible for TI am62a7 dss\n\n - Add a kunit test for __of_address_resource_bounds()\n\n* tag \u0027devicetree-fixes-for-6.14-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:\n dt-bindings: display: Add powertip,{st7272|hx8238a} as DT Schema description\n dt-bindings: nvmem: qcom,qfprom: Add SAR2130P compatible\n dt-bindings: display: ti: Fix compatible for am62a7 dss\n of: address: Add kunit test for __of_address_resource_bounds()\n dt-bindings: clock: qcom: Add QCS8300 video clock controller\n dt-bindings: clock: qcom: Add CAMCC clocks for QCS8300\n dt-bindings: clock: qcom: Add GPU clocks for QCS8300\n" }, { "commit": "ad73b9a17d66366d8c9198bc90f1ea99f24a912c", "tree": "44d2a3038e20b8c103f5a04b4f09beb4177526ea", "parents": [ "5784d8c93eb0c3b26cbaa86237c3b106657fc0c1", "96178631c3f53398044ed437010f7632ad764bf8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 17:14:53 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 17:14:53 2025 -0800" }, "message": "Merge tag \u0027uml-for-linus-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/uml/linux\n\nPull UML fixes from Richard Weinberger:\n\n - Align signal stack correctly\n\n - Convert to raw spinlocks where needed (irq and virtio)\n\n - FPU related fixes\n\n* tag \u0027uml-for-linus-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/uml/linux:\n um: convert irq_lock to raw spinlock\n um: virtio_uml: use raw spinlock\n um: virt-pci: don\u0027t use kmalloc()\n um: fix execve stub execution on old host OSs\n um: properly align signal stack on x86_64\n um: avoid copying FP state from init_task\n um: add back support for FXSAVE registers\n" }, { "commit": "5784d8c93eb0c3b26cbaa86237c3b106657fc0c1", "tree": "4338be468243d771c91c5e8b09b3603184b081b6", "parents": [ "496659003dac5d08ea292c44fce9dfb36fa34691", "97937834ae876f29565415ab15f1284666dc6be3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 16:34:41 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 16:34:41 2025 -0800" }, "message": "Merge tag \u0027trace-ring-buffer-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull trace ring buffer fixes from Steven Rostedt:\n\n - Enable resize on mmap() error\n\n When a process mmaps a ring buffer, its size is locked and resizing\n is disabled. But if the user passes in a wrong parameter, the mmap()\n can fail after the resize was disabled and the mmap() exits with\n error without reenabling the ring buffer resize. This prevents the\n ring buffer from ever being resized after that. Reenable resizing of\n the ring buffer on mmap() error.\n\n - Have resizing return proper error and not always -ENOMEM\n\n If the ring buffer is mmapped by one task and another task tries to\n resize the buffer it will error with -ENOMEM. This is confusing to\n the user as there may be plenty of memory available. Have it return\n the error that actually happens (in this case -EBUSY) where the user\n can understand why the resize failed.\n\n - Test the sub-buffer array to validate persistent memory buffer\n\n On boot up, the initialization of the persistent memory buffer will\n do a validation check to see if the content of the data is valid, and\n if so, it will use the memory as is, otherwise it re-initializes it.\n There\u0027s meta data in this persistent memory that keeps track of which\n sub-buffer is the reader page and an array that states the order of\n the sub-buffers. The values in this array are indexes into the\n sub-buffers. The validator checks to make sure that all the entries\n in the array are within the sub-buffer list index, but it does not\n check for duplications.\n\n While working on this code, the array got corrupted and had\n duplicates, where not all the sub-buffers were accounted for. This\n passed the validator as all entries were valid, but the link list was\n incorrect and could have caused a crash. The corruption only produced\n incorrect data, but it could have been more severe. To fix this,\n create a bitmask that covers all the sub-buffer indexes and set it to\n all zeros. While iterating the array checking the values of the array\n content, have it set a bit corresponding to the index in the array.\n If the bit was already set, then it is a duplicate and mark the\n buffer as invalid and reset it.\n\n - Prevent mmap()ing persistent ring buffer\n\n The persistent ring buffer uses vmap() to map the persistent memory.\n Currently, the mmap() logic only uses virt_to_page() to get the page\n from the ring buffer memory and use that to map to user space. This\n works because a normal ring buffer uses alloc_page() to allocate its\n memory. But because the persistent ring buffer use vmap() it causes a\n kernel crash.\n\n Fixing this to work with vmap() is not hard, but since mmap() on\n persistent memory buffers never worked, just have the mmap() return\n -ENODEV (what was returned before mmap() for persistent memory ring\n buffers, as they never supported mmap. Normal buffers will still\n allow mmap(). Implementing mmap() for persistent memory ring buffers\n can wait till the next merge window.\n\n - Fix polling on persistent ring buffers\n\n There\u0027s a \"buffer_percent\" option (default set to 50), that is used\n to have reads of the ring buffer binary data block until the buffer\n fills to that percentage. The field \"pages_touched\" is incremented\n every time a new sub-buffer has content added to it. This field is\n used in the calculations to determine the amount of content is in the\n buffer and if it exceeds the \"buffer_percent\" then it will wake the\n task polling on the buffer.\n\n As persistent ring buffers can be created by the content from a\n previous boot, the \"pages_touched\" field was not updated. This means\n that if a task were to poll on the persistent buffer, it would block\n even if the buffer was completely full. It would block even if the\n \"buffer_percent\" was zero, because with \"pages_touched\" as zero, it\n would be calculated as the buffer having no content. Update\n pages_touched when initializing the persistent ring buffer from a\n previous boot.\n\n* tag \u0027trace-ring-buffer-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n ring-buffer: Update pages_touched to reflect persistent buffer content\n tracing: Do not allow mmap() of persistent ring buffer\n ring-buffer: Validate the persistent meta data subbuf array\n tracing: Have the error of __tracing_resize_ring_buffer() passed to user\n ring-buffer: Unlock resize on mmap error\n" }, { "commit": "97937834ae876f29565415ab15f1284666dc6be3", "tree": "482d8ee9c864db647e7fc5d5e8af85474d0557da", "parents": [ "129fe718819cc5e24ea2f489db9ccd4371f0c6f6" ], "author": { "name": "Steven Rostedt", "email": "rostedt@goodmis.org", "time": "Fri Feb 14 12:35:12 2025 -0500" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Sat Feb 15 14:00:59 2025 -0500" }, "message": "ring-buffer: Update pages_touched to reflect persistent buffer content\n\nThe pages_touched field represents the number of subbuffers in the ring\nbuffer that have content that can be read. This is used in accounting of\n\"dirty_pages\" and \"buffer_percent\" to allow the user to wait for the\nbuffer to be filled to a certain amount before it reads the buffer in\nblocking mode.\n\nThe persistent buffer never updated this value so it was set to zero, and\nthis accounting would take it as it had no content. This would cause user\nspace to wait for content even though there\u0027s enough content in the ring\nbuffer that satisfies the buffer_percent.\n\nCc: stable@vger.kernel.org\nCc: Masami Hiramatsu \u003cmhiramat@kernel.org\u003e\nCc: Mathieu Desnoyers \u003cmathieu.desnoyers@efficios.com\u003e\nCc: Vincent Donnefort \u003cvdonnefort@google.com\u003e\nLink: https://lore.kernel.org/20250214123512.0631436e@gandalf.local.home\nFixes: 5f3b6e839f3ce (\"ring-buffer: Validate boot range memory events\")\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "129fe718819cc5e24ea2f489db9ccd4371f0c6f6", "tree": "811004274e11f15c74d529d08d5573e3e6eabf0c", "parents": [ "f5b95f1fa2ef3a03f49eeec658ba97e721412b32" ], "author": { "name": "Steven Rostedt", "email": "rostedt@goodmis.org", "time": "Fri Feb 14 11:55:47 2025 -0500" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Sat Feb 15 13:59:52 2025 -0500" }, "message": "tracing: Do not allow mmap() of persistent ring buffer\n\nWhen trying to mmap a trace instance buffer that is attached to\nreserve_mem, it would crash:\n\n BUG: unable to handle page fault for address: ffffe97bd00025c8\n #PF: supervisor read access in kernel mode\n #PF: error_code(0x0000) - not-present page\n PGD 2862f3067 P4D 2862f3067 PUD 0\n Oops: Oops: 0000 [#1] PREEMPT_RT SMP PTI\n CPU: 4 UID: 0 PID: 981 Comm: mmap-rb Not tainted 6.14.0-rc2-test-00003-g7f1a5e3fbf9e-dirty #233\n Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2 04/01/2014\n RIP: 0010:validate_page_before_insert+0x5/0xb0\n Code: e2 01 89 d0 c3 cc cc cc cc 66 66 2e 0f 1f 84 00 00 00 00 00 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 0f 1f 44 00 00 \u003c48\u003e 8b 46 08 a8 01 75 67 66 90 48 89 f0 8b 50 34 85 d2 74 76 48 89\n RSP: 0018:ffffb148c2f3f968 EFLAGS: 00010246\n RAX: ffff9fa5d3322000 RBX: ffff9fa5ccff9c08 RCX: 00000000b879ed29\n RDX: ffffe97bd00025c0 RSI: ffffe97bd00025c0 RDI: ffff9fa5ccff9c08\n RBP: ffffb148c2f3f9f0 R08: 0000000000000004 R09: 0000000000000004\n R10: 0000000000000000 R11: 0000000000000200 R12: 0000000000000000\n R13: 00007f16a18d5000 R14: ffff9fa5c48db6a8 R15: 0000000000000000\n FS: 00007f16a1b54740(0000) GS:ffff9fa73df00000(0000) knlGS:0000000000000000\n CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033\n CR2: ffffe97bd00025c8 CR3: 00000001048c6006 CR4: 0000000000172ef0\n Call Trace:\n \u003cTASK\u003e\n ? __die_body.cold+0x19/0x1f\n ? __die+0x2e/0x40\n ? page_fault_oops+0x157/0x2b0\n ? search_module_extables+0x53/0x80\n ? validate_page_before_insert+0x5/0xb0\n ? kernelmode_fixup_or_oops.isra.0+0x5f/0x70\n ? __bad_area_nosemaphore+0x16e/0x1b0\n ? bad_area_nosemaphore+0x16/0x20\n ? do_kern_addr_fault+0x77/0x90\n ? exc_page_fault+0x22b/0x230\n ? asm_exc_page_fault+0x2b/0x30\n ? validate_page_before_insert+0x5/0xb0\n ? vm_insert_pages+0x151/0x400\n __rb_map_vma+0x21f/0x3f0\n ring_buffer_map+0x21b/0x2f0\n tracing_buffers_mmap+0x70/0xd0\n __mmap_region+0x6f0/0xbd0\n mmap_region+0x7f/0x130\n do_mmap+0x475/0x610\n vm_mmap_pgoff+0xf2/0x1d0\n ksys_mmap_pgoff+0x166/0x200\n __x64_sys_mmap+0x37/0x50\n x64_sys_call+0x1670/0x1d70\n do_syscall_64+0xbb/0x1d0\n entry_SYSCALL_64_after_hwframe+0x77/0x7f\n\nThe reason was that the code that maps the ring buffer pages to user space\nhas:\n\n\tpage \u003d virt_to_page((void *)cpu_buffer-\u003esubbuf_ids[s]);\n\nAnd uses that in:\n\n\tvm_insert_pages(vma, vma-\u003evm_start, pages, \u0026nr_pages);\n\nBut virt_to_page() does not work with vmap()\u0027d memory which is what the\npersistent ring buffer has. It is rather trivial to allow this, but for\nnow just disable mmap() of instances that have their ring buffer from the\nreserve_mem option.\n\nIf an mmap() is performed on a persistent buffer it will return -ENODEV\njust like it would if the .mmap field wasn\u0027t defined in the\nfile_operations structure.\n\nCc: stable@vger.kernel.org\nCc: Masami Hiramatsu \u003cmhiramat@kernel.org\u003e\nCc: Mathieu Desnoyers \u003cmathieu.desnoyers@efficios.com\u003e\nCc: Vincent Donnefort \u003cvdonnefort@google.com\u003e\nLink: https://lore.kernel.org/20250214115547.0d7287d3@gandalf.local.home\nFixes: 9b7bdf6f6ece6 (\"tracing: Have trace_printk not use binary prints if boot buffer\")\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "496659003dac5d08ea292c44fce9dfb36fa34691", "tree": "a94fd76e161b38f4ed96fbbbe8b8ff05f0ed9c70", "parents": [ "f3d8b0ebaec4e755317e2e8b863a4004d6505288", "f85478fb3fecad01927935c51fe7e9dd5731d0b7" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 10:20:47 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 10:20:47 2025 -0800" }, "message": "Merge tag \u0027i2c-for-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux\n\nPull i2c fixes from Wolfram Sang:\n \"MAINTAINERS maintenance.\n\n Changed email, added entry, deleted entry falling back to a generic\n one\"\n\n* tag \u0027i2c-for-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:\n MAINTAINERS: Add maintainer for Qualcomm\u0027s I2C GENI driver\n MAINTAINERS: delete entry for AXXIA I2C\n MAINTAINERS: Use my kernel.org address for I2C ACPI work\n" }, { "commit": "f3d8b0ebaec4e755317e2e8b863a4004d6505288", "tree": "3372f79f056095fb30db86250d7ab16df8a0ed61", "parents": [ "243899076c3efdf98d8e922a802896424a597580", "2844ddbd540fc84d7571cca65d6c43088e4d6952" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 10:15:24 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 10:15:24 2025 -0800" }, "message": "Merge tag \u0027s390-6.14-4\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux\n\nPull s390 fixes from Vasily Gorbik:\n\n - Fix isolated VFs handling by verifying that a VF’s parent PF is\n locally owned before registering it in an existing PCI domain\n\n - Disable arch_test_bit() optimization for PROFILE_ALL_BRANCHES to\n workaround gcc failure in handling __builtin_constant_p() in this\n case\n\n - Fix CHPID \"configure\" attribute caching in CIO by not updating the\n cache when SCLP returns no data, ensuring consistent sysfs output\n\n - Remove CONFIG_LSM from default configs and rely on defaults, which\n enables BPF LSM hook\n\n* tag \u0027s390-6.14-4\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:\n s390/pci: Fix handling of isolated VFs\n s390/pci: Pull search for parent PF out of zpci_iov_setup_virtfn()\n s390/bitops: Disable arch_test_bit() optimization for PROFILE_ALL_BRANCHES\n s390/cio: Fix CHPID \"configure\" attribute caching\n s390/configs: Remove CONFIG_LSM\n" }, { "commit": "b28fb1f2ef45eeef1cd2c23149b50d184d545a3e", "tree": "6a87be1084b08f590f5546cb36c36c210c16a298", "parents": [ "1b71c2fb04e7a713abc6edde4a412416ff3158f2" ], "author": { "name": "Uwe Kleine-König", "email": "u.kleine-koenig@baylibre.com", "time": "Thu Feb 13 17:04:29 2025 +0100" }, "committer": { "name": "Masahiro Yamada", "email": "masahiroy@kernel.org", "time": "Sun Feb 16 03:10:58 2025 +0900" }, "message": "modpost: Fix a few typos in a comment\n\nNamely: s/becasue/because/ and s/wiht/with/ plus an added article.\n\nSigned-off-by: Uwe Kleine-König \u003cu.kleine-koenig@baylibre.com\u003e\nSigned-off-by: Masahiro Yamada \u003cmasahiroy@kernel.org\u003e\n" }, { "commit": "1b71c2fb04e7a713abc6edde4a412416ff3158f2", "tree": "6963531f6f9ab645e747c644c27e10b01eb900dc", "parents": [ "140332b6ed727a4ec2e5722a1ccda28b52d45771" ], "author": { "name": "Thomas Weißschuh", "email": "thomas.weissschuh@linutronix.de", "time": "Thu Feb 13 15:55:17 2025 +0100" }, "committer": { "name": "Masahiro Yamada", "email": "masahiroy@kernel.org", "time": "Sun Feb 16 03:10:58 2025 +0900" }, "message": "kbuild: userprogs: fix bitsize and target detection on clang\n\nscripts/Makefile.clang was changed in the linked commit to move --target from\nKBUILD_CFLAGS to KBUILD_CPPFLAGS, as that generally has a broader scope.\nHowever that variable is not inspected by the userprogs logic,\nbreaking cross compilation on clang.\n\nUse both variables to detect bitsize and target arguments for userprogs.\n\nFixes: feb843a469fb (\"kbuild: add $(CLANG_FLAGS) to KBUILD_CPPFLAGS\")\nCc: stable@vger.kernel.org\nSigned-off-by: Thomas Weißschuh \u003cthomas.weissschuh@linutronix.de\u003e\nReviewed-by: Nathan Chancellor \u003cnathan@kernel.org\u003e\nSigned-off-by: Masahiro Yamada \u003cmasahiroy@kernel.org\u003e\n" }, { "commit": "243899076c3efdf98d8e922a802896424a597580", "tree": "195f3dd0c38d605b666cf72b9c2292a9d9a704e0", "parents": [ "d440148418f4816b4973ec6723bf63821793a0a7", "2e4f982cf392af2f1282b5537a72144e064799e3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 09:54:46 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 09:54:46 2025 -0800" }, "message": "Merge tag \u0027rust-fixes-6.14-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux\n\nPull rust fixes from Miguel Ojeda:\n\n - Fix objtool warning due to future Rust 1.85.0 (to be released in a\n few days)\n\n - Clean future Rust 1.86.0 (to be released 2025-04-03) Clippy warning\n\n* tag \u0027rust-fixes-6.14-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux:\n rust: rbtree: fix overindented list item\n objtool/rust: add one more `noreturn` Rust function\n" }, { "commit": "d440148418f4816b4973ec6723bf63821793a0a7", "tree": "ec2c16e52c4ad1d869757ebabdfc6870a3fe52f2", "parents": [ "6452feaf29a6a1cc1c904520c4b1b4cd90133fac" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 09:28:55 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 09:28:55 2025 -0800" }, "message": "tegra210-adma: fix 32-bit x86 build\n\nThe Tegra210 Audio DMA controller driver did a plain divide:\n\n\tpage_no \u003d (res_page-\u003estart - res_base-\u003estart) / cdata-\u003ech_base_offset;\n\nwhich causes problems on 32-bit x86 configurations that have 64-bit\nresource sizes:\n\n x86_64-linux-ld: drivers/dma/tegra210-adma.o: in function `tegra_adma_probe\u0027:\n tegra210-adma.c:(.text+0x1322): undefined reference to `__udivdi3\u0027\n\nbecause gcc doesn\u0027t generate the trivial code for a 64-by-32 divide,\nturning it into a function call to do a full 64-by-64 divide. And the\nkernel intentionally doesn\u0027t provide that helper function, because 99%\nof the time all you want is the narrower version.\n\nOf course, tegra210 is a 64-bit architecture and the 32-bit x86 build is\npurely for build testing, so this really is just about build coverage\nfailure.\n\nBut build coverage is good.\n\nSide note: div_u64() would be suboptimal if you actually have a 32-bit\nresource_t, so our \"helper\" for divides are admittedly making it harder\nthan it should be to generate good code for all the possible cases.\n\nAt some point, I\u0027ll consider 32-bit x86 so entirely legacy that I can\u0027t\nfind it in myself to care any more, and we\u0027ll just add the __udivdi3\nlibrary function.\n\nBut for now, the right thing to do is to use \"div_u64()\" to show that\nyou know that you are doing the simpler divide with a 32-bit number.\nAnd the build error enforces that.\n\nWhile fixing the build issue, also check for division-by-zero, and for\noverflow. Which hopefully cannot happen on real production hardware,\nbut the value of \u0027ch_base_offset\u0027 can definitely be zero in other\nplaces.\n\nReported-by: Guenter Roeck \u003clinux@roeck-us.net\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6452feaf29a6a1cc1c904520c4b1b4cd90133fac", "tree": "fa09e14a05d9f9b2818737bdd732f9d23ab1c757", "parents": [ "7ff71e6d923969d933e1ba7e0db857782d36cd19", "7b4aebeecbbd5b5fe73e35fad3f62ed21aa7ef44" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 08:13:45 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 15 08:13:45 2025 -0800" }, "message": "Merge tag \u0027gpio-fixes-for-v6.14-rc3-take2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux\n\nPull gpio fixes from Bartosz Golaszewski:\n\n - fix interrupt handling issues in gpio-bcm-kona\n\n - add an ACPI quirk for Acer Nitro ANV14 fixing an issue with spurious\n wake up events\n\n - add missing return value checks to gpio-stmpe\n\n - fix a crash in error path in gpiochip_get_ngpios()\n\n* tag \u0027gpio-fixes-for-v6.14-rc3-take2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:\n gpiolib: Fix crash on error in gpiochip_get_ngpios()\n gpio: stmpe: Check return value of stmpe_reg_read in stmpe_gpio_irq_sync_unlock\n gpiolib: acpi: Add a quirk for Acer Nitro ANV14\n gpio: bcm-kona: Add missing newline to dev_err format string\n gpio: bcm-kona: Make sure GPIO bits are unlocked when requesting IRQ\n gpio: bcm-kona: Fix GPIO lock/unlock for banks above bank 0\n" }, { "commit": "140332b6ed727a4ec2e5722a1ccda28b52d45771", "tree": "e60702d352de0c101cd1a76c107c5cd4497cfcdd", "parents": [ "d1d0963121769d8d16150b913fe886e48efefa51" ], "author": { "name": "Masahiro Yamada", "email": "masahiroy@kernel.org", "time": "Thu Feb 13 15:26:44 2025 +0900" }, "committer": { "name": "Masahiro Yamada", "email": "masahiroy@kernel.org", "time": "Sat Feb 15 22:40:52 2025 +0900" }, "message": "kbuild: fix linux-headers package build when $(CC) cannot link userspace\n\nSince commit 5f73e7d0386d (\"kbuild: refactor cross-compiling\nlinux-headers package\"), the linux-headers Debian package fails to\nbuild when $(CC) cannot build userspace applications, for example,\nwhen using toolchains installed by the 0day bot.\n\nThe host programs in the linux-headers package should be rebuilt using\nthe disto\u0027s cross-compiler, ${DEB_HOST_GNU_TYPE}-gcc instead of $(CC).\nHence, the variable \u0027CC\u0027 must be expanded in this shell script instead\nof in the top-level Makefile.\n\nCommit f354fc88a72a (\"kbuild: install-extmod-build: add missing\nquotation marks for CC variable\") was not a correct fix because\nCC\u003d\"ccache gcc\" should be unrelated when rebuilding userspace tools.\n\nFixes: 5f73e7d0386d (\"kbuild: refactor cross-compiling linux-headers package\")\nReported-by: Jeff Johnson \u003cjeff.johnson@oss.qualcomm.com\u003e\nCloses: https://lore.kernel.org/linux-kbuild/CAK7LNARb3xO3ptBWOMpwKcyf3\u003dzkfhMey5H2KnB1dOmUwM79dA@mail.gmail.com/T/#t\nSigned-off-by: Masahiro Yamada \u003cmasahiroy@kernel.org\u003e\nTested-by: Jeff Johnson \u003cjeff.johnson@oss.qualcomm.com\u003e\n" }, { "commit": "d1d0963121769d8d16150b913fe886e48efefa51", "tree": "1218b50aee86f38b6e8c48d2e0729defb3823be0", "parents": [ "a64dcfb451e254085a7daee5fe51bf22959d52d3" ], "author": { "name": "Masahiro Yamada", "email": "masahiroy@kernel.org", "time": "Tue Feb 11 09:29:06 2025 +0900" }, "committer": { "name": "Masahiro Yamada", "email": "masahiroy@kernel.org", "time": "Sat Feb 15 22:36:10 2025 +0900" }, "message": "tools: fix annoying \"mkdir -p ...\" logs when building tools in parallel\n\nWhen CONFIG_OBJTOOL\u003dy or CONFIG_DEBUG_INFO_BTF\u003dy, parallel builds\nshow awkward \"mkdir -p ...\" logs.\n\n $ make -j16\n [ snip ]\n mkdir -p /home/masahiro/ref/linux/tools/objtool \u0026\u0026 make O\u003d/home/masahiro/ref/linux subdir\u003dtools/objtool --no-print-directory -C objtool\n mkdir -p /home/masahiro/ref/linux/tools/bpf/resolve_btfids \u0026\u0026 make O\u003d/home/masahiro/ref/linux subdir\u003dtools/bpf/resolve_btfids --no-print-directory -C bpf/resolve_btfids\n\nDefining MAKEFLAGS\u003d\u003cvalue\u003e on the command line wipes out command line\nswitches from the resultant MAKEFLAGS definition, even though the command\nline switches are active. [1]\n\nMAKEFLAGS puts all single-letter options into the first word, and that\nword will be empty if no single-letter options were given. [2]\nHowever, this breaks if MAKEFLAGS\u003d\u003cvalue\u003e is given on the command line.\n\nThe tools/ and tools/% targets set MAKEFLAGS\u003d\u003cvalue\u003e on the command\nline, which breaks the following code in tools/scripts/Makefile.include:\n\n short-opts :\u003d $(firstword -$(MAKEFLAGS))\n\nIf MAKEFLAGS really needs modification, it should be done through the\nenvironment variable, as follows:\n\n MAKEFLAGS\u003d\u003cvalue\u003e $(MAKE) ...\n\nThat said, I question whether modifying MAKEFLAGS is necessary here.\nThe only flag we might want to exclude is --no-print-directory, as the\ntools build system changes the working directory. However, people might\nfind the \"Entering/Leaving directory\" logs annoying.\n\nI simply removed the offending MAKEFLAGS\u003d\u003cvalue\u003e.\n\n[1]: https://savannah.gnu.org/bugs/?62469\n[2]: https://www.gnu.org/software/make/manual/make.html#Testing-Flags\n\nFixes: ea01fa9f63ae (\"tools: Connect to the kernel build system\")\nFixes: a50e43332756 (\"perf tools: Honor parallel jobs\")\nSigned-off-by: Masahiro Yamada \u003cmasahiroy@kernel.org\u003e\nTested-by: Daniel Xu \u003cdxu@dxuuu.xyz\u003e\n" }, { "commit": "7ff71e6d923969d933e1ba7e0db857782d36cd19", "tree": "de60ad08b12c8402283d0901492e0c6d239db50b", "parents": [ "78a632a2086c5d5468b0e088a97b26e47c569567", "1523226edda566057bdd3264ceb56631ddf5f6f7" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 19:56:12 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 19:56:12 2025 -0800" }, "message": "Merge tag \u0027alpha-fixes-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mattst88/alpha\n\nPull alpha fixes from Matt Turner:\n \"A few changes for alpha, including some important fixes for kernel\n stack alignment\"\n\n* tag \u0027alpha-fixes-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mattst88/alpha:\n alpha: Use str_yes_no() helper in pci_dac_dma_supported()\n alpha: Replace one-element array with flexible array member\n alpha: align stack for page fault and user unaligned trap handlers\n alpha: make stack 16-byte aligned (most cases)\n alpha: replace hardcoded stack offsets with autogenerated ones\n" }, { "commit": "78a632a2086c5d5468b0e088a97b26e47c569567", "tree": "f0f667e5cbbf524afa022970e6979bbf4ffd4d28", "parents": [ "3f2ca7b8b33db7d094ef729e6b0707b5534159b0", "81f64e925c29fe6e99f04b131fac1935ac931e81" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 16:49:07 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 16:49:07 2025 -0800" }, "message": "Merge tag \u0027pci-v6.14-fixes-3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci\n\nPull pci fixes from Bjorn Helgaas:\n\n - Update a BUILD_BUG_ON() usage that works on current compilers, but\n breaks compilation on gcc 5.3.1 (Alex Williamson)\n\n - Avoid use of FLR for Mediatek MT7922 WiFi; the device previously\n worked after a long timeout and fallback to SBR, but after a recent\n RRS change it doesn\u0027t work at all after FLR (Bjorn Helgaas)\n\n* tag \u0027pci-v6.14-fixes-3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci:\n PCI: Avoid FLR for Mediatek MT7922 WiFi\n PCI: Fix BUILD_BUG_ON usage for old gcc\n" }, { "commit": "d3d0b8dfe06098d6d584266c35e9a0947f5b7132", "tree": "2c38950a3489a5c0e6ac98a711cf1f5384c5b7f5", "parents": [ "409f45387c937145adeeeebc6d6032c2ec232b35", "c2fee09fc167c74a64adb08656cb993ea475197e" ], "author": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Fri Feb 14 19:08:35 2025 -0500" }, "committer": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Fri Feb 14 19:08:35 2025 -0500" }, "message": "Merge tag \u0027kvm-x86-fixes-6.14-rcN\u0027 of https://github.com/kvm-x86/linux into HEAD\n\nKVM fixes for 6.14 part 1\n\n - Reject Hyper-V SEND_IPI hypercalls if the local APIC isn\u0027t being emulated\n by KVM to fix a NULL pointer dereference.\n\n - Enter guest mode (L2) from KVM\u0027s perspective before initializing the vCPU\u0027s\n nested NPT MMU so that the MMU is properly tagged for L2, not L1.\n\n - Load the guest\u0027s DR6 outside of the innermost .vcpu_run() loop, as the\n guest\u0027s value may be stale if a VM-Exit is handled in the fastpath.\n" }, { "commit": "409f45387c937145adeeeebc6d6032c2ec232b35", "tree": "0fc67341ec47f6e7b510280e536025d8ca504ba8", "parents": [ "44e70718df4fc2fadf1665eb9374df71aeda1f03" ], "author": { "name": "Ashish Kalra", "email": "ashish.kalra@amd.com", "time": "Mon Feb 10 22:54:18 2025 +0000" }, "committer": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Fri Feb 14 18:39:19 2025 -0500" }, "message": "x86/sev: Fix broken SNP support with KVM module built-in\n\nFix issues with enabling SNP host support and effectively SNP support\nwhich is broken with respect to the KVM module being built-in.\n\nSNP host support is enabled in snp_rmptable_init() which is invoked as\ndevice_initcall(). SNP check on IOMMU is done during IOMMU PCI init\n(IOMMU_PCI_INIT stage). And for that reason snp_rmptable_init() is\ncurrently invoked via device_initcall() and cannot be invoked via\nsubsys_initcall() as core IOMMU subsystem gets initialized via\nsubsys_initcall().\n\nNow, if kvm_amd module is built-in, it gets initialized before SNP host\nsupport is enabled in snp_rmptable_init() :\n\n[ 10.131811] kvm_amd: TSC scaling supported\n[ 10.136384] kvm_amd: Nested Virtualization enabled\n[ 10.141734] kvm_amd: Nested Paging enabled\n[ 10.146304] kvm_amd: LBR virtualization supported\n[ 10.151557] kvm_amd: SEV enabled (ASIDs 100 - 509)\n[ 10.156905] kvm_amd: SEV-ES enabled (ASIDs 1 - 99)\n[ 10.162256] kvm_amd: SEV-SNP enabled (ASIDs 1 - 99)\n[ 10.171508] kvm_amd: Virtual VMLOAD VMSAVE supported\n[ 10.177052] kvm_amd: Virtual GIF supported\n...\n...\n[ 10.201648] kvm_amd: in svm_enable_virtualization_cpu\n\nAnd then svm_x86_ops-\u003eenable_virtualization_cpu()\n(svm_enable_virtualization_cpu) programs MSR_VM_HSAVE_PA as following:\nwrmsrl(MSR_VM_HSAVE_PA, sd-\u003esave_area_pa);\n\nSo VM_HSAVE_PA is non-zero before SNP support is enabled on all CPUs.\n\nsnp_rmptable_init() gets invoked after svm_enable_virtualization_cpu()\nas following :\n...\n[ 11.256138] kvm_amd: in svm_enable_virtualization_cpu\n...\n[ 11.264918] SEV-SNP: in snp_rmptable_init\n\nThis triggers a #GP exception in snp_rmptable_init() when snp_enable()\nis invoked to set SNP_EN in SYSCFG MSR:\n\n[ 11.294289] unchecked MSR access error: WRMSR to 0xc0010010 (tried to write 0x0000000003fc0000) at rIP: 0xffffffffaf5d5c28 (native_write_msr+0x8/0x30)\n...\n[ 11.294404] Call Trace:\n[ 11.294482] \u003cIRQ\u003e\n[ 11.294513] ? show_stack_regs+0x26/0x30\n[ 11.294522] ? ex_handler_msr+0x10f/0x180\n[ 11.294529] ? search_extable+0x2b/0x40\n[ 11.294538] ? fixup_exception+0x2dd/0x340\n[ 11.294542] ? exc_general_protection+0x14f/0x440\n[ 11.294550] ? asm_exc_general_protection+0x2b/0x30\n[ 11.294557] ? __pfx_snp_enable+0x10/0x10\n[ 11.294567] ? native_write_msr+0x8/0x30\n[ 11.294570] ? __snp_enable+0x5d/0x70\n[ 11.294575] snp_enable+0x19/0x20\n[ 11.294578] __flush_smp_call_function_queue+0x9c/0x3a0\n[ 11.294586] generic_smp_call_function_single_interrupt+0x17/0x20\n[ 11.294589] __sysvec_call_function+0x20/0x90\n[ 11.294596] sysvec_call_function+0x80/0xb0\n[ 11.294601] \u003c/IRQ\u003e\n[ 11.294603] \u003cTASK\u003e\n[ 11.294605] asm_sysvec_call_function+0x1f/0x30\n...\n[ 11.294631] arch_cpu_idle+0xd/0x20\n[ 11.294633] default_idle_call+0x34/0xd0\n[ 11.294636] do_idle+0x1f1/0x230\n[ 11.294643] ? complete+0x71/0x80\n[ 11.294649] cpu_startup_entry+0x30/0x40\n[ 11.294652] start_secondary+0x12d/0x160\n[ 11.294655] common_startup_64+0x13e/0x141\n[ 11.294662] \u003c/TASK\u003e\n\nThis #GP exception is getting triggered due to the following errata for\nAMD family 19h Models 10h-1Fh Processors:\n\nProcessor may generate spurious #GP(0) Exception on WRMSR instruction:\nDescription:\nThe Processor will generate a spurious #GP(0) Exception on a WRMSR\ninstruction if the following conditions are all met:\n- the target of the WRMSR is a SYSCFG register.\n- the write changes the value of SYSCFG.SNPEn from 0 to 1.\n- One of the threads that share the physical core has a non-zero\nvalue in the VM_HSAVE_PA MSR.\n\nThe document being referred to above:\nhttps://www.amd.com/content/dam/amd/en/documents/processor-tech-docs/revision-guides/57095-PUB_1_01.pdf\n\nTo summarize, with kvm_amd module being built-in, KVM/SVM initialization\nhappens before host SNP is enabled and this SVM initialization\nsets VM_HSAVE_PA to non-zero, which then triggers a #GP when\nSYSCFG.SNPEn is being set and this will subsequently cause\nSNP_INIT(_EX) to fail with INVALID_CONFIG error as SYSCFG[SnpEn] is not\nset on all CPUs.\n\nEssentially SNP host enabling code should be invoked before KVM\ninitialization, which is currently not the case when KVM is built-in.\n\nAdd fix to call snp_rmptable_init() early from iommu_snp_enable()\ndirectly and not invoked via device_initcall() which enables SNP host\nsupport before KVM initialization with kvm_amd module built-in.\n\nAdd additional handling for `iommu\u003doff` or `amd_iommu\u003doff` options.\n\nNote that IOMMUs need to be enabled for SNP initialization, therefore,\nif host SNP support is enabled but late IOMMU initialization fails\nthen that will cause PSP driver\u0027s SNP_INIT to fail as IOMMU SNP sanity\nchecks in SNP firmware will fail with invalid configuration error as\nbelow:\n\n[ 9.723114] ccp 0000:23:00.1: sev enabled\n[ 9.727602] ccp 0000:23:00.1: psp enabled\n[ 9.732527] ccp 0000:a2:00.1: enabling device (0000 -\u003e 0002)\n[ 9.739098] ccp 0000:a2:00.1: no command queues available\n[ 9.745167] ccp 0000:a2:00.1: psp enabled\n[ 9.805337] ccp 0000:23:00.1: SEV-SNP: failed to INIT rc -5, error 0x3\n[ 9.866426] ccp 0000:23:00.1: SEV API:1.53 build:5\n\nFixes: c3b86e61b756 (\"x86/cpufeatures: Enable/unmask SEV-SNP CPU feature\")\nCo-developed-by: Sean Christopherson \u003cseanjc@google.com\u003e\nSigned-off-by: Sean Christopherson \u003cseanjc@google.com\u003e\nCo-developed-by: Vasant Hegde \u003cvasant.hegde@amd.com\u003e\nSigned-off-by: Vasant Hegde \u003cvasant.hegde@amd.com\u003e\nCc: \u003cStable@vger.kernel.org\u003e\nSigned-off-by: Ashish Kalra \u003cashish.kalra@amd.com\u003e\nAcked-by: Joerg Roedel \u003cjroedel@suse.de\u003e\nMessage-ID: \u003c138b520fb83964782303b43ade4369cd181fdd9c.1739226950.git.ashish.kalra@amd.com\u003e\nSigned-off-by: Paolo Bonzini \u003cpbonzini@redhat.com\u003e\n" }, { "commit": "44e70718df4fc2fadf1665eb9374df71aeda1f03", "tree": "5fbf93fdc419fb54b09c686ae9996618afe02294", "parents": [ "435b344a7042e91fb4719d589f18310e8919e39f" ], "author": { "name": "Sean Christopherson", "email": "seanjc@google.com", "time": "Mon Feb 10 22:54:02 2025 +0000" }, "committer": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Fri Feb 14 18:39:19 2025 -0500" }, "message": "KVM: SVM: Ensure PSP module is initialized if KVM module is built-in\n\nThe kernel\u0027s initcall infrastructure lacks the ability to express\ndependencies between initcalls, whereas the modules infrastructure\nautomatically handles dependencies via symbol loading. Ensure the\nPSP SEV driver is initialized before proceeding in sev_hardware_setup()\nif KVM is built-in as the dependency isn\u0027t handled by the initcall\ninfrastructure.\n\nSigned-off-by: Sean Christopherson \u003cseanjc@google.com\u003e\nReviewed-by: Tom Lendacky \u003cthomas.lendacky@amd.com\u003e\nSigned-off-by: Ashish Kalra \u003cashish.kalra@amd.com\u003e\nMessage-ID: \u003cf78ddb64087df27e7bcb1ae0ab53f55aa0804fab.1739226950.git.ashish.kalra@amd.com\u003e\nSigned-off-by: Paolo Bonzini \u003cpbonzini@redhat.com\u003e\n" }, { "commit": "435b344a7042e91fb4719d589f18310e8919e39f", "tree": "2237aa4dae8e8577897cbb6a6bd631d4652f3bd0", "parents": [ "3bb7dcebd022fea8a696185592d21bceed7b426c" ], "author": { "name": "Sean Christopherson", "email": "seanjc@google.com", "time": "Mon Feb 10 22:53:47 2025 +0000" }, "committer": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Fri Feb 14 18:39:19 2025 -0500" }, "message": "crypto: ccp: Add external API interface for PSP module initialization\n\nKVM is dependent on the PSP SEV driver and PSP SEV driver needs to be\nloaded before KVM module. In case of module loading any dependent\nmodules are automatically loaded but in case of built-in modules there\nis no inherent mechanism available to specify dependencies between\nmodules and ensure that any dependent modules are loaded implicitly.\n\nAdd a new external API interface for PSP module initialization which\nallows PSP SEV driver to be loaded explicitly if KVM is built-in.\n\nSigned-off-by: Sean Christopherson \u003cseanjc@google.com\u003e\nCo-developed-by: Ashish Kalra \u003cashish.kalra@amd.com\u003e\nSigned-off-by: Ashish Kalra \u003cashish.kalra@amd.com\u003e\nReviewed-by: Tom Lendacky \u003cthomas.lendacky@amd.com\u003e\nMessage-ID: \u003c15279ca0cad56a07cf12834ec544310f85ff5edc.1739226950.git.ashish.kalra@amd.com\u003e\nSigned-off-by: Paolo Bonzini \u003cpbonzini@redhat.com\u003e\n" }, { "commit": "3bb7dcebd022fea8a696185592d21bceed7b426c", "tree": "589e669d7aa8de1277d9c6d5305b832d74ff2e45", "parents": [ "43fb96ae78551d7bfa4ecca956b258f085d67c40", "b3aa9283c0c505b5cfd25f7d6cfd720de2adc807" ], "author": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Fri Feb 14 18:32:47 2025 -0500" }, "committer": { "name": "Paolo Bonzini", "email": "pbonzini@redhat.com", "time": "Fri Feb 14 18:32:47 2025 -0500" }, "message": "Merge tag \u0027kvmarm-fixes-6.14-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm into HEAD\n\nKVM/arm64 fixes for 6.14, take #2\n\n- Large set of fixes for vector handling, specially in the interactions\n between host and guest state. This fixes a number of bugs affecting\n actual deployments, and greatly simplifies the FP/SIMD/SVE handling.\n Thanks to Mark Rutland for dealing with this thankless task.\n\n- Fix an ugly race between vcpu and vgic creation/init, resulting in\n unexpected behaviours.\n\n- Fix use of kernel VAs at EL2 when emulating timers with nVHE.\n\n- Small set of pKVM improvements and cleanups.\n" }, { "commit": "3f2ca7b8b33db7d094ef729e6b0707b5534159b0", "tree": "fde3e8a58a05976f2c6206c3d46b85abdc007b86", "parents": [ "f92a80bdb445e4e73e810ea7e94bebb10717f40a", "a6768c4f92e152265590371975d44c071a5279c7" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 15:07:11 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 15:07:11 2025 -0800" }, "message": "Merge tag \u0027thermal-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm\n\nPull thermal control fixes from Rafael Wysocki:\n \"Fix a regression caused by an inadvertent change of the\n THERMAL_GENL_ATTR_CPU_CAPABILITY value in one of the recent thermal\n commits (Zhang Rui) and drop a stale piece of documentation (Daniel\n Lezcano)\"\n\n* tag \u0027thermal-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:\n thermal/cpufreq_cooling: Remove structure member documentation\n thermal/netlink: Prevent userspace segmentation fault by adjusting UAPI header\n" }, { "commit": "f92a80bdb445e4e73e810ea7e94bebb10717f40a", "tree": "0bf3877de51ad6f635787cda292b26d2988e40ba", "parents": [ "cabb162eb5ab02bf10809fe5025d437e99e119e3", "3e68abf2b9cebe76c6cd4b1aca8e95cd671035a3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 14:55:50 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 14:55:50 2025 -0800" }, "message": "Merge tag \u0027mmc-v6.14-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc\n\nPull MMC host fixes from Ulf Hansson:\n\n - mtk-sd: Fix register settings for hs400(es) mode\n\n - sdhci_am654: Revert patch for start-signal-voltage-switch\n\n* tag \u0027mmc-v6.14-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc:\n mmc: mtk-sd: Fix register settings for hs400(es) mode\n Revert \"mmc: sdhci_am654: Add sdhci_am654_start_signal_voltage_switch\"\n" }, { "commit": "cabb162eb5ab02bf10809fe5025d437e99e119e3", "tree": "8b4c258752d930af3e963796e3190eb0dbb42244", "parents": [ "c7ab7b2a18f386b37a289a00ff3d92f00bb16934", "f1bf10d7e909fe898a112f5cae1e97ce34d6484d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 14:42:52 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 14:42:52 2025 -0800" }, "message": "Merge tag \u0027v6.14-rc2-smb3-client-fixes\u0027 of git://git.samba.org/sfrench/cifs-2.6\n\nPull smb client fix from Steve French:\n \"SMB3 client multichannel fix\"\n\n* tag \u0027v6.14-rc2-smb3-client-fixes\u0027 of git://git.samba.org/sfrench/cifs-2.6:\n cifs: pick channels for individual subrequests\n" }, { "commit": "c7ab7b2a18f386b37a289a00ff3d92f00bb16934", "tree": "54256bc258399c336a440483263e21027eee756d", "parents": [ "1b8c8cdad1749d68130147f187008a368d564933", "bbc4578537e350d5bf8a7a2c7d054d6b163b3c41" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 13:56:04 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 13:56:04 2025 -0800" }, "message": "Merge tag \u0027efi-fixes-for-v6.14-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/efi/efi\n\nPull EFI fixes from Ard Biesheuvel:\n \"Take the newly introduced EFI_MEMORY_HOT_PLUGGABLE memory attribute\n into account when placing the kernel image in memory at boot.\n\n Otherwise, the presence of the kernel image could prevent such a\n memory region from being unplugged at runtime if it was \u0027cold\n plugged\u0027, i.e., already plugged in at boot time (and exposed via the\n EFI memory map).\n\n This should ensure that the new EFI_MEMORY_HOT_PLUGGABLE memory\n attribute is used consistently by Linux before it ever turns up in\n production, ensuring that we can make meaningful use of it without\n running the risk of regressing existing users\"\n\n* tag \u0027efi-fixes-for-v6.14-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/efi/efi:\n efi: Use BIT_ULL() constants for memory attributes\n efi: Avoid cold plugged memory for placing the kernel\n" }, { "commit": "1b8c8cdad1749d68130147f187008a368d564933", "tree": "2607026dfd9039f59cae273a2dbdb8f04e4477d1", "parents": [ "ea717324741471665110b4475a52c08a56026a9e", "80e648042e512d5a767da251d44132553fe04ae0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 11:40:59 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 11:40:59 2025 -0800" }, "message": "Merge tag \u0027block-6.14-20250214\u0027 of git://git.kernel.dk/linux\n\nPull block fixes from Jens Axboe:\n\n - Fix for request rejection for batch addition\n\n - Fix a few issues for bogus mac partition tables\n\n* tag \u0027block-6.14-20250214\u0027 of git://git.kernel.dk/linux:\n partitions: mac: fix handling of bogus partition table\n block: cleanup and fix batch completion adding conditions\n" }, { "commit": "ea717324741471665110b4475a52c08a56026a9e", "tree": "42dacb310f49e126339993f1d905358e17a651d8", "parents": [ "04f41cbf03ec7221ab0b179e336f4c805ee55520", "d6211ebbdaa541af197b50b8dd8f22642ce0b87f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 11:30:53 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 11:30:53 2025 -0800" }, "message": "Merge tag \u0027io_uring-6.14-20250214\u0027 of git://git.kernel.dk/linux\n\nPull io_uring fixes from Jens Axboe:\n\n - fixes for a potential data corruption issue with IORING_OP_URING_CMD,\n where not all the SQE data is stable. Will be revisited in the\n future, for now it ends up with just always copying it beyond prep to\n provide the same guarantees as all other opcodes\n\n - make the waitid opcode setup async data like any other opcodes (no\n real fix here, just a consistency thing)\n\n - fix for waitid io_tw_state abuse\n\n - when a buffer group is type is changed, do so by allocating a new\n buffer group entry and discard the old one, rather than migrating\n\n* tag \u0027io_uring-6.14-20250214\u0027 of git://git.kernel.dk/linux:\n io_uring/uring_cmd: unconditionally copy SQEs at prep time\n io_uring/waitid: setup async data in the prep handler\n io_uring/uring_cmd: remove dead req_has_async_data() check\n io_uring/uring_cmd: switch sqe to async_data on EAGAIN\n io_uring/uring_cmd: don\u0027t assume io_uring_cmd_data layout\n io_uring/kbuf: reallocate buf lists on upgrade\n io_uring/waitid: don\u0027t abuse io_tw_state\n" }, { "commit": "04f41cbf03ec7221ab0b179e336f4c805ee55520", "tree": "3602f0ee06188739b4500f2f769982972104f19c", "parents": [ "80868f5d3dca7230a87ee7947d0c7f821ce85b3e", "f5717c93a1b999970f3a64d771a1a9ee68cc37d0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 11:14:24 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 11:14:24 2025 -0800" }, "message": "Merge tag \u0027sched_ext-for-6.14-rc2-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext\n\nPull sched_ext fixes from Tejun Heo:\n\n - Fix lock imbalance in a corner case of dispatch_to_local_dsq()\n\n - Migration disabled tasks were confusing some BPF schedulers and its\n handling had a bug. Fix it and simplify the default behavior by\n dispatching them automatically\n\n - ops.tick(), ops.disable() and ops.exit_task() were incorrectly\n disallowing kfuncs that require the task argument to be the rq\n operation is currently operating on and thus is rq-locked.\n Allow them.\n\n - Fix autogroup migration handling bug which was occasionally\n triggering a warning in the cgroup migration path\n\n - tools/sched_ext, selftest and other misc updates\n\n* tag \u0027sched_ext-for-6.14-rc2-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext:\n sched_ext: Use SCX_CALL_OP_TASK in task_tick_scx\n sched_ext: Fix the incorrect bpf_list kfunc API in common.bpf.h.\n sched_ext: selftests: Fix grammar in tests description\n sched_ext: Fix incorrect assumption about migration disabled tasks in task_can_run_on_remote_rq()\n sched_ext: Fix migration disabled handling in targeted dispatches\n sched_ext: Implement auto local dispatching of migration disabled tasks\n sched_ext: Fix incorrect time delta calculation in time_delta()\n sched_ext: Fix lock imbalance in dispatch_to_local_dsq()\n sched_ext: selftests/dsp_local_on: Fix selftest on UP systems\n tools/sched_ext: Add helper to check task migration state\n sched_ext: Fix incorrect autogroup migration detection\n sched_ext: selftests/dsp_local_on: Fix sporadic failures\n selftests/sched_ext: Fix enum resolution\n sched_ext: Include task weight in the error state dump\n sched_ext: Fixes typos in comments\n" }, { "commit": "1523226edda566057bdd3264ceb56631ddf5f6f7", "tree": "c1573827972f30874b01f7f2c33279cfbd168fc8", "parents": [ "757f051a506198186d796dff4ba696adb7bda54c" ], "author": { "name": "Thorsten Blum", "email": "thorsten.blum@linux.dev", "time": "Wed Feb 12 12:14:47 2025 +0100" }, "committer": { "name": "Matt Turner", "email": "mattst88@gmail.com", "time": "Fri Feb 14 14:06:41 2025 -0500" }, "message": "alpha: Use str_yes_no() helper in pci_dac_dma_supported()\n\nRemove hard-coded strings by using the str_yes_no() helper function.\n\nReviewed-by: Geert Uytterhoeven \u003cgeert@linux-m68k.org\u003e\nSigned-off-by: Thorsten Blum \u003cthorsten.blum@linux.dev\u003e\nSigned-off-by: Matt Turner \u003cmattst88@gmail.com\u003e\n" }, { "commit": "757f051a506198186d796dff4ba696adb7bda54c", "tree": "a170923dc87d7ddc3935b6438a312aa74418e11d", "parents": [ "3b35a171060f846b08b48646b38c30b5d57d17ff" ], "author": { "name": "Thorsten Blum", "email": "thorsten.blum@linux.dev", "time": "Tue Jan 07 11:43:42 2025 +0100" }, "committer": { "name": "Matt Turner", "email": "mattst88@gmail.com", "time": "Fri Feb 14 14:06:15 2025 -0500" }, "message": "alpha: Replace one-element array with flexible array member\n\nReplace the deprecated one-element array with a modern flexible array\nmember in the struct crb_struct.\n\nReviewed-by: Kees Cook \u003ckees@kernel.org\u003e\nSigned-off-by: Thorsten Blum \u003cthorsten.blum@linux.dev\u003e\nSigned-off-by: Matt Turner \u003cmattst88@gmail.com\u003e\n" }, { "commit": "3b35a171060f846b08b48646b38c30b5d57d17ff", "tree": "c0f5bb6997edc44a881b2aee103842f963ee151c", "parents": [ "0a0f7362b0367634a2d5cb7c96226afc116f19c9" ], "author": { "name": "Ivan Kokshaysky", "email": "ink@unseen.parts", "time": "Tue Feb 04 23:35:24 2025 +0100" }, "committer": { "name": "Matt Turner", "email": "mattst88@gmail.com", "time": "Fri Feb 14 14:06:04 2025 -0500" }, "message": "alpha: align stack for page fault and user unaligned trap handlers\n\ndo_page_fault() and do_entUna() are special because they use\nnon-standard stack frame layout. Fix them manually.\n\nCc: stable@vger.kernel.org\nTested-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nTested-by: Magnus Lindholm \u003clinmag7@gmail.com\u003e\nTested-by: Matt Turner \u003cmattst88@gmail.com\u003e\nReviewed-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nSuggested-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nSigned-off-by: Ivan Kokshaysky \u003cink@unseen.parts\u003e\nSigned-off-by: Matt Turner \u003cmattst88@gmail.com\u003e\n" }, { "commit": "0a0f7362b0367634a2d5cb7c96226afc116f19c9", "tree": "1e03eab3ce588b49d3251007f022901c72f51465", "parents": [ "77b823fa619f97d16409ca37ad4f7936e28c5f83" ], "author": { "name": "Ivan Kokshaysky", "email": "ink@unseen.parts", "time": "Tue Feb 04 23:35:23 2025 +0100" }, "committer": { "name": "Matt Turner", "email": "mattst88@gmail.com", "time": "Fri Feb 14 14:05:47 2025 -0500" }, "message": "alpha: make stack 16-byte aligned (most cases)\n\nThe problem is that GCC expects 16-byte alignment of the incoming stack\nsince early 2004, as Maciej found out [1]:\n Having actually dug speculatively I can see that the psABI was changed in\n GCC 3.5 with commit e5e10fb4a350 (\"re PR target/14539 (128-bit long double\n improperly aligned)\") back in Mar 2004, when the stack pointer alignment\n was increased from 8 bytes to 16 bytes, and arch/alpha/kernel/entry.S has\n various suspicious stack pointer adjustments, starting with SP_OFF which\n is not a whole multiple of 16.\n\nAlso, as Magnus noted, \"ALPHA Calling Standard\" [2] required the same:\n D.3.1 Stack Alignment\n This standard requires that stacks be octaword aligned at the time a\n new procedure is invoked.\n\nHowever:\n- the \"normal\" kernel stack is always misaligned by 8 bytes, thanks to\n the odd number of 64-bit words in \u0027struct pt_regs\u0027, which is the very\n first thing pushed onto the kernel thread stack;\n- syscall, fault, interrupt etc. handlers may, or may not, receive aligned\n stack depending on numerous factors.\n\nSomehow we got away with it until recently, when we ended up with\na stack corruption in kernel/smp.c:smp_call_function_single() due to\nits use of 32-byte aligned local data and the compiler doing clever\nthings allocating it on the stack.\n\nThis adds padding between the PAL-saved and kernel-saved registers\nso that \u0027struct pt_regs\u0027 have an even number of 64-bit words.\nThis makes the stack properly aligned for most of the kernel\ncode, except two handlers which need special threatment.\n\nNote: struct pt_regs doesn\u0027t belong in uapi/asm; this should be fixed,\nbut let\u0027s put this off until later.\n\nLink: https://lore.kernel.org/rcu/alpine.DEB.2.21.2501130248010.18889@angie.orcam.me.uk/ [1]\nLink: https://bitsavers.org/pdf/dec/alpha/Alpha_Calling_Standard_Rev_2.0_19900427.pdf [2]\n\nCc: stable@vger.kernel.org\nTested-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nTested-by: Magnus Lindholm \u003clinmag7@gmail.com\u003e\nTested-by: Matt Turner \u003cmattst88@gmail.com\u003e\nReviewed-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nSigned-off-by: Ivan Kokshaysky \u003cink@unseen.parts\u003e\nSigned-off-by: Matt Turner \u003cmattst88@gmail.com\u003e\n" }, { "commit": "77b823fa619f97d16409ca37ad4f7936e28c5f83", "tree": "ee04cfd3bfd530743790f0f7872e75c578e3523e", "parents": [ "128c8f96eb8638c060cd3532dc394d046ce64fe1" ], "author": { "name": "Ivan Kokshaysky", "email": "ink@unseen.parts", "time": "Tue Feb 04 23:35:22 2025 +0100" }, "committer": { "name": "Matt Turner", "email": "mattst88@gmail.com", "time": "Fri Feb 14 14:03:40 2025 -0500" }, "message": "alpha: replace hardcoded stack offsets with autogenerated ones\n\nThis allows the assembly in entry.S to automatically keep in sync with\nchanges in the stack layout (struct pt_regs and struct switch_stack).\n\nCc: stable@vger.kernel.org\nTested-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nTested-by: Matt Turner \u003cmattst88@gmail.com\u003e\nReviewed-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nSigned-off-by: Ivan Kokshaysky \u003cink@unseen.parts\u003e\nSigned-off-by: Matt Turner \u003cmattst88@gmail.com\u003e\n" }, { "commit": "80868f5d3dca7230a87ee7947d0c7f821ce85b3e", "tree": "35b13e74e884a190d4baee575bf08f66894e91cb", "parents": [ "f4d4680965aa3be2b2bf392a54fabb244fdce46a", "db5fd3cf8bf41b84b577b8ad5234ea95f327c9be" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 11:00:42 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 11:00:42 2025 -0800" }, "message": "Merge tag \u0027cgroup-for-6.14-rc2-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup\n\nPull cgroup fixes from Tejun Heo:\n\n - Fix a race window where a newly forked task could escape cgroup.kill\n\n - Remove incorrectly included steal time from cpu.stat::usage_usec\n\n - Minor update in selftest\n\n* tag \u0027cgroup-for-6.14-rc2-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:\n cgroup: Remove steal time from usage_usec\n selftests/cgroup: use bash in test_cpuset_v1_hp.sh\n cgroup: fix race between fork and cgroup.kill\n" }, { "commit": "f4d4680965aa3be2b2bf392a54fabb244fdce46a", "tree": "5bf260a6ced50b8d9ba616379b6593f5c24ef541", "parents": [ "111b29599c189247919a514b6c3a93e828db27c2", "e76946110137703c16423baf6ee177b751a34b7e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 10:58:03 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 10:58:03 2025 -0800" }, "message": "Merge tag \u0027wq-for-6.14-rc2-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq\n\nPull workqueue fix from Tejun Heo:\n\n - Fix a regression where a worker pool can be freed before rescuer\n workers are done with it leading to user-after-free\n\n* tag \u0027wq-for-6.14-rc2-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:\n workqueue: Put the pwq after detaching the rescuer from the pool\n" }, { "commit": "111b29599c189247919a514b6c3a93e828db27c2", "tree": "903e6236fcfb2c44be34d3ec9c31eb0d953f7dfb", "parents": [ "93bd481b6076299256c96aa21362f09cea121c69", "446a8351f160d65a1c5df7097f31c74102ed2bb1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 09:55:17 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 09:55:17 2025 -0800" }, "message": "Merge tag \u0027arm64-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux\n\nPull arm64 fixes from Will Deacon:\n\n - Fix kexec and hibernation when using 5-level page-table configuration\n\n - Remove references to non-existent SF8MM4 and SF8MM8 ID register\n fields, hooking up hwcaps for the FPRCVT, F8MM4 and F8MM8 fields\n instead\n\n - Drop unused .ARM.attributes ELF sections\n\n - Fix array indexing when probing CPU cache topology from firmware\n\n - Fix potential use-after-free in AMU initialisation code\n\n - Work around broken GTDT entries by tolerating excessively large timer\n arrays\n\n - Force use of Rust\u0027s \"softfloat\" target to avoid a threatening warning\n about the NEON target feature\n\n - Typo fix in GCS documentation and removal of duplicate Kconfig select\n\n* tag \u0027arm64-fixes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:\n arm64: rust: clean Rust 1.85.0 warning using softfloat target\n arm64: Add missing registrations of hwcaps\n ACPI: GTDT: Relax sanity checking on Platform Timers array count\n arm64: amu: Delay allocating cpumask for AMU FIE support\n arm64: cacheinfo: Avoid out-of-bounds write to cacheinfo array\n arm64: Handle .ARM.attributes section in linker scripts\n arm64/hwcap: Remove stray references to SF8MMx\n arm64/gcs: Fix documentation for HWCAP\n arm64: Kconfig: Remove selecting replaced HAVE_FUNCTION_GRAPH_RETVAL\n arm64: Fix 5-level paging support in kexec/hibernate trampoline\n" }, { "commit": "f5b95f1fa2ef3a03f49eeec658ba97e721412b32", "tree": "1d0e29b21ee1285966cef14e60041e3eb6107618", "parents": [ "60b8f711143de7cd9c0f55be0fe7eb94b19eb5c7" ], "author": { "name": "Steven Rostedt", "email": "rostedt@goodmis.org", "time": "Fri Feb 14 10:28:20 2025 -0500" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Fri Feb 14 12:50:51 2025 -0500" }, "message": "ring-buffer: Validate the persistent meta data subbuf array\n\nThe meta data for a mapped ring buffer contains an array of indexes of all\nthe subbuffers. The first entry is the reader page, and the rest of the\nentries lay out the order of the subbuffers in how the ring buffer link\nlist is to be created.\n\nThe validator currently makes sure that all the entries are within the\nrange of 0 and nr_subbufs. But it does not check if there are any\nduplicates.\n\nWhile working on the ring buffer, I corrupted this array, where I added\nduplicates. The validator did not catch it and created the ring buffer\nlink list on top of it. Luckily, the corruption was only that the reader\npage was also in the writer path and only presented corrupted data but did\nnot crash the kernel. But if there were duplicates in the writer side,\nthen it could corrupt the ring buffer link list and cause a crash.\n\nCreate a bitmask array with the size of the number of subbuffers. Then\nclear it. When walking through the subbuf array checking to see if the\nentries are within the range, test if its bit is already set in the\nsubbuf_mask. If it is, then there is duplicates and fail the validation.\nIf not, set the corresponding bit and continue.\n\nCc: stable@vger.kernel.org\nCc: Masami Hiramatsu \u003cmhiramat@kernel.org\u003e\nCc: Mathieu Desnoyers \u003cmathieu.desnoyers@efficios.com\u003e\nCc: Vincent Donnefort \u003cvdonnefort@google.com\u003e\nLink: https://lore.kernel.org/20250214102820.7509ddea@gandalf.local.home\nFixes: c76883f18e59b (\"ring-buffer: Add test if range of boot buffer is valid\")\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "60b8f711143de7cd9c0f55be0fe7eb94b19eb5c7", "tree": "5980b90cf6461c5879637808da5eb6427489a1e0", "parents": [ "9ba0e1755a40f9920ad0f4168031291b3eb58d7b" ], "author": { "name": "Steven Rostedt", "email": "rostedt@goodmis.org", "time": "Thu Feb 13 13:41:32 2025 -0500" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Fri Feb 14 12:50:27 2025 -0500" }, "message": "tracing: Have the error of __tracing_resize_ring_buffer() passed to user\n\nCurrently if __tracing_resize_ring_buffer() returns an error, the\ntracing_resize_ringbuffer() returns -ENOMEM. But it may not be a memory\nissue that caused the function to fail. If the ring buffer is memory\nmapped, then the resizing of the ring buffer will be disabled. But if the\nuser tries to resize the buffer, it will get an -ENOMEM returned, which is\nconfusing because there is plenty of memory. The actual error returned was\n-EBUSY, which would make much more sense to the user.\n\nCc: stable@vger.kernel.org\nCc: Mathieu Desnoyers \u003cmathieu.desnoyers@efficios.com\u003e\nCc: Vincent Donnefort \u003cvdonnefort@google.com\u003e\nLink: https://lore.kernel.org/20250213134132.7e4505d7@gandalf.local.home\nFixes: 117c39200d9d7 (\"ring-buffer: Introducing ring-buffer mapping functions\")\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\nReviewed-by: Masami Hiramatsu (Google) \u003cmhiramat@kernel.org\u003e\n" }, { "commit": "9ba0e1755a40f9920ad0f4168031291b3eb58d7b", "tree": "393bd09ffdcf607f26d7c60dc54de0c258ab1e02", "parents": [ "a64dcfb451e254085a7daee5fe51bf22959d52d3" ], "author": { "name": "Steven Rostedt", "email": "rostedt@goodmis.org", "time": "Thu Feb 13 13:19:57 2025 -0500" }, "committer": { "name": "Steven Rostedt (Google)", "email": "rostedt@goodmis.org", "time": "Fri Feb 14 12:50:05 2025 -0500" }, "message": "ring-buffer: Unlock resize on mmap error\n\nMemory mapping the tracing ring buffer will disable resizing the buffer.\nBut if there\u0027s an error in the memory mapping like an invalid parameter,\nthe function exits out without re-enabling the resizing of the ring\nbuffer, preventing the ring buffer from being resized after that.\n\nCc: stable@vger.kernel.org\nCc: Masami Hiramatsu \u003cmhiramat@kernel.org\u003e\nCc: Mathieu Desnoyers \u003cmathieu.desnoyers@efficios.com\u003e\nCc: Vincent Donnefort \u003cvdonnefort@google.com\u003e\nLink: https://lore.kernel.org/20250213131957.530ec3c5@gandalf.local.home\nFixes: 117c39200d9d7 (\"ring-buffer: Introducing ring-buffer mapping functions\")\nSigned-off-by: Steven Rostedt (Google) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "93bd481b6076299256c96aa21362f09cea121c69", "tree": "f5c254414efe1f061e55de7dbde3a0da4b7528a7", "parents": [ "fd31a1bea3c94e01cb7b998485d2d7b14bdc8101", "add43c4fbc92f8b48c1acd64e953af3b1be4cd9c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 09:46:51 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 09:46:51 2025 -0800" }, "message": "Merge tag \u0027iommu-fixes-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/iommu/linux\n\nPull iommu fixes from Joerg Roedel:\n\n - core: fix potential memory leak in iopf_queue_remove_device()\n\n - Intel VT-d: handle faults correctly in intel_iommu_drain_pasid_prq()\n\n - AMD-Vi: fix faults happening in resume path\n\n - typo and spelling fixes\n\n* tag \u0027iommu-fixes-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/iommu/linux:\n iommu/vt-d: Make intel_iommu_drain_pasid_prq() cover faults for RID\n iommu/exynos: Fix typos\n iommu: Fix a spelling error\n iommu/amd: Expicitly enable CNTRL.EPHEn bit in resume path\n iommu: Fix potential memory leak in iopf_queue_remove_device()\n" }, { "commit": "f85478fb3fecad01927935c51fe7e9dd5731d0b7", "tree": "4b89fd36bb561138f35a31871d30af8056eea9eb", "parents": [ "6fe9116dd6bebee570406ec3f00a50388a62ccb3", "7422c319fd805b956aab5ba93e0274517a8e3650" ], "author": { "name": "Wolfram Sang", "email": "wsa+renesas@sang-engineering.com", "time": "Fri Feb 14 17:23:54 2025 +0100" }, "committer": { "name": "Wolfram Sang", "email": "wsa+renesas@sang-engineering.com", "time": "Fri Feb 14 17:23:54 2025 +0100" }, "message": "Merge tag \u0027i2c-host-fixes-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/andi.shyti/linux into i2c/for-current\n\ni2c-host-fixes for v6.14-rc3\n\n- Mukesh and Viken take over maintainership of the Qualcomm I2C\n driver.\n- Krzysztof Adamski is removed as maintainer of the Axxia I2C\n driver.\n" }, { "commit": "fd31a1bea3c94e01cb7b998485d2d7b14bdc8101", "tree": "dd17f1bb05698a7f6d79efc3659271e597f91bea", "parents": [ "128c8f96eb8638c060cd3532dc394d046ce64fe1", "75ad02318af2e4ae669e26a79f001bd5e1f97472" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 08:15:17 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Feb 14 08:15:17 2025 -0800" }, "message": "Merge tag \u0027for-linus-6.14-rc3-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip\n\nPull xen fixes from Juergen Gross:\n \"Three fixes to xen-swiotlb driver:\n\n - two fixes for issues coming up due to another fix in 6.12\n\n - addition of an __init annotation\"\n\n* tag \u0027for-linus-6.14-rc3-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:\n Xen/swiotlb: mark xen_swiotlb_fixup() __init\n x86/xen: allow larger contiguous memory regions in PV guests\n xen/swiotlb: relax alignment requirements\n" }, { "commit": "80e648042e512d5a767da251d44132553fe04ae0", "tree": "4516f9009fa27b77f790cb09c7a44dd6a955ab9e", "parents": [ "1f47ed294a2bd577d5ae43e6e28e1c9a3be4a833" ], "author": { "name": "Jann Horn", "email": "jannh@google.com", "time": "Fri Feb 14 02:39:50 2025 +0100" }, "committer": { "name": "Jens Axboe", "email": "axboe@kernel.dk", "time": "Fri Feb 14 08:38:28 2025 -0700" }, "message": "partitions: mac: fix handling of bogus partition table\n\nFix several issues in partition probing:\n\n - The bailout for a bad partoffset must use put_dev_sector(), since the\n preceding read_part_sector() succeeded.\n - If the partition table claims a silly sector size like 0xfff bytes\n (which results in partition table entries straddling sector boundaries),\n bail out instead of accessing out-of-bounds memory.\n - We must not assume that the partition table contains proper NUL\n termination - use strnlen() and strncmp() instead of strlen() and\n strcmp().\n\nCc: stable@vger.kernel.org\nSigned-off-by: Jann Horn \u003cjannh@google.com\u003e\nLink: https://lore.kernel.org/r/20250214-partition-mac-v1-1-c1c626dffbd5@google.com\nSigned-off-by: Jens Axboe \u003caxboe@kernel.dk\u003e\n" }, { "commit": "9e512eaaf8f4008c44ede3dfc0fbc9d9c5118583", "tree": "04cce12ef6dcbab25eaf97edcdd962076ae7f109", "parents": [ "0f3fd9cf6491f5beecbb65abb41556c56135340c" ], "author": { "name": "John Keeping", "email": "jkeeping@inmusicbrands.com", "time": "Sat Feb 08 12:41:44 2025 +0000" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:50:55 2025 +0100" }, "message": "serial: 8250: Fix fifo underflow on flush\n\nWhen flushing the serial port\u0027s buffer, uart_flush_buffer() calls\nkfifo_reset() but if there is an outstanding DMA transfer then the\ncompletion function will consume data from the kfifo via\nuart_xmit_advance(), underflowing and leading to ongoing DMA as the\ndriver tries to transmit another 2^32 bytes.\n\nThis is readily reproduced with serial-generic and amidi sending even\nshort messages as closing the device on exit will wait for the fifo to\ndrain and in the underflow case amidi hangs for 30 seconds on exit in\ntty_wait_until_sent(). A trace of that gives:\n\n kworker/1:1-84 [001] 51.769423: bprint: serial8250_tx_dma: tx_size\u003d3 fifo_len\u003d3\n amidi-763 [001] 51.769460: bprint: uart_flush_buffer: resetting fifo\n irq/21-fe530000-76 [000] 51.769474: bprint: __dma_tx_complete: tx_size\u003d3\n irq/21-fe530000-76 [000] 51.769479: bprint: serial8250_tx_dma: tx_size\u003d4096 fifo_len\u003d4294967293\n irq/21-fe530000-76 [000] 51.781295: bprint: __dma_tx_complete: tx_size\u003d4096\n irq/21-fe530000-76 [000] 51.781301: bprint: serial8250_tx_dma: tx_size\u003d4096 fifo_len\u003d4294963197\n irq/21-fe530000-76 [000] 51.793131: bprint: __dma_tx_complete: tx_size\u003d4096\n irq/21-fe530000-76 [000] 51.793135: bprint: serial8250_tx_dma: tx_size\u003d4096 fifo_len\u003d4294959101\n irq/21-fe530000-76 [000] 51.804949: bprint: __dma_tx_complete: tx_size\u003d4096\n\nSince the port lock is held in when the kfifo is reset in\nuart_flush_buffer() and in __dma_tx_complete(), adding a flush_buffer\nhook to adjust the outstanding DMA byte count is sufficient to avoid the\nkfifo underflow.\n\nFixes: 9ee4b83e51f74 (\"serial: 8250: Add support for dmaengine\")\nCc: stable \u003cstable@kernel.org\u003e\nSigned-off-by: John Keeping \u003cjkeeping@inmusicbrands.com\u003e\nLink: https://lore.kernel.org/r/20250208124148.1189191-1-jkeeping@inmusicbrands.com\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "659f5d55feb75782bd46cf130da3c1f240afe9ba", "tree": "d321148200d5d69357836b21fbcee38385e9b27d", "parents": [ "634775a752a86784511018a108f3b530cc3399a7" ], "author": { "name": "Jos Wang", "email": "joswang@lenovo.com", "time": "Thu Feb 13 21:49:21 2025 +0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:26:44 2025 +0100" }, "message": "usb: typec: tcpm: PSSourceOffTimer timeout in PR_Swap enters ERROR_RECOVERY\n\nAs PD2.0 spec (\"6.5.6.2 PSSourceOffTimer\")，the PSSourceOffTimer is\nused by the Policy Engine in Dual-Role Power device that is currently\nacting as a Sink to timeout on a PS_RDY Message during a Power Role\nSwap sequence. This condition leads to a Hard Reset for USB Type-A and\nType-B Plugs and Error Recovery for Type-C plugs and return to USB\nDefault Operation.\n\nTherefore, after PSSourceOffTimer timeout, the tcpm state machine should\nswitch from PR_SWAP_SNK_SRC_SINK_OFF to ERROR_RECOVERY. This can also\nsolve the test items in the USB power delivery compliance test:\nTEST.PD.PROT.SNK.12 PR_Swap – PSSourceOffTimer Timeout\n\n[1] https://usb.org/document-library/usb-power-delivery-compliance-test-specification-0/USB_PD3_CTS_Q4_2025_OR.zip\n\nFixes: f0690a25a140 (\"staging: typec: USB Type-C Port Manager (tcpm)\")\nCc: stable \u003cstable@kernel.org\u003e\nSigned-off-by: Jos Wang \u003cjoswang@lenovo.com\u003e\nReviewed-by: Heikki Krogerus \u003cheikki.krogerus@linux.intel.com\u003e\nTested-by: Amit Sunil Dhamne \u003camitsd@google.com\u003e\nLink: https://lore.kernel.org/r/20250213134921.3798-1-joswang1221@gmail.com\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "634775a752a86784511018a108f3b530cc3399a7", "tree": "c70a2ffd00a52ff154193115881a8c272fdcd070", "parents": [ "e5644be4079750a0a0a5a7068fd90b97bf6fac55" ], "author": { "name": "Elson Roy Serrao", "email": "quic_eserrao@quicinc.com", "time": "Thu Feb 06 11:39:50 2025 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:26:25 2025 +0100" }, "message": "usb: roles: set switch registered flag early on\n\nThe role switch registration and set_role() can happen in parallel as they\nare invoked independent of each other. There is a possibility that a driver\nmight spend significant amount of time in usb_role_switch_register() API\ndue to the presence of time intensive operations like component_add()\nwhich operate under common mutex. This leads to a time window after\nallocating the switch and before setting the registered flag where the set\nrole notifications are dropped. Below timeline summarizes this behavior\n\nThread1\t\t\t\t|\tThread2\nusb_role_switch_register()\t|\n\t|\t\t\t|\n\t---\u003e allocate switch\t|\n\t|\t\t\t|\n\t---\u003e component_add()\t|\tusb_role_switch_set_role()\n\t|\t\t\t|\t|\n\t|\t\t\t|\t--\u003e Drop role notifications\n\t|\t\t\t|\t since sw-\u003eregistered\n\t|\t\t\t|\t flag is not set.\n\t|\t\t\t|\n\t---\u003eSet registered flag.|\n\nTo avoid this, set the registered flag early on in the switch register\nAPI.\n\nFixes: b787a3e78175 (\"usb: roles: don\u0027t get/set_role() when usb_role_switch is unregistered\")\nCc: stable \u003cstable@kernel.org\u003e\nSigned-off-by: Elson Roy Serrao \u003cquic_eserrao@quicinc.com\u003e\nReviewed-by: Heikki Krogerus \u003cheikki.krogerus@linux.intel.com\u003e\nLink: https://lore.kernel.org/r/20250206193950.22421-1-quic_eserrao@quicinc.com\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "e5644be4079750a0a0a5a7068fd90b97bf6fac55", "tree": "b5eb732b994f1fdd9206a5f8db62cefcf46f4347", "parents": [ "e169d96eecd447ff7fd7542ca5fa0911f5622054" ], "author": { "name": "Frederic Weisbecker", "email": "frederic@kernel.org", "time": "Wed Feb 12 14:55:14 2025 +0100" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:26:00 2025 +0100" }, "message": "usb: gadget: uvc: Fix unstarted kthread worker\n\nThe behaviour of kthread_create_worker() was recently changed to align\nwith the one of kthread_create(). The kthread worker is created but not\nawaken by default. This is to allow the use of kthread_affine_preferred()\nand kthread_bind[_mask]() with kthread workers. In order to keep the\nold behaviour and wake the kthread up, kthread_run_worker() must be\nused. All the pre-existing users have been converted, except for UVC\nthat was introduced in the same merge window as the API change.\n\nThis results in hangs:\n\n\tINFO: task UVCG:82 blocked for more than 491 seconds.\n\tTainted: G T 6.13.0-rc2-00014-gb04e317b5226 #1\n\ttask:UVCG state:D stack:0 pid:82\n\t Call Trace:\n\t __schedule\n\t schedule\n\t schedule_preempt_disabled\n\t kthread\n\t ? kthread_flush_work\n\t ret_from_fork\n\t ret_from_fork_asm\n\t entry_INT80_32\n\nFix this with converting UVCG kworker to the new API.\n\nReported-by: kernel test robot \u003coliver.sang@intel.com\u003e\nCloses: https://lore.kernel.org/oe-lkp/202502121025.55bfa801-lkp@intel.com\nFixes: f0bbfbd16b3b (\"usb: gadget: uvc: rework to enqueue in pump worker from encoded queue\")\nCc: stable \u003cstable@kernel.org\u003e\nCc: Michael Grzeschik \u003cm.grzeschik@pengutronix.de\u003e\nSigned-off-by: Frederic Weisbecker \u003cfrederic@kernel.org\u003e\nLink: https://lore.kernel.org/r/20250212135514.30539-1-frederic@kernel.org\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "e169d96eecd447ff7fd7542ca5fa0911f5622054", "tree": "d381f1efc8274410d774a2b0d3451a84794dff65", "parents": [ "399a45e5237ca14037120b1b895bd38a3b4492ea" ], "author": { "name": "Lei Huang", "email": "huanglei@kylinos.cn", "time": "Wed Feb 12 17:38:29 2025 +0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:25:41 2025 +0100" }, "message": "USB: quirks: add USB_QUIRK_NO_LPM quirk for Teclast dist\n\nTeclast disk used on Huawei hisi platforms doesn\u0027t work well,\nlosing connectivity intermittently if LPM is enabled.\nAdd quirk disable LPM to resolve the issue.\n\nSigned-off-by: Lei Huang \u003chuanglei@kylinos.cn\u003e\nCc: stable \u003cstable@kernel.org\u003e\nLink: https://lore.kernel.org/r/20250212093829.7379-1-huanglei814@163.com\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "399a45e5237ca14037120b1b895bd38a3b4492ea", "tree": "5ceb98b74bff8c0cfa001dcbb219374c32a4b158", "parents": [ "4ab37fcb42832cdd3e9d5e50653285ca84d6686f" ], "author": { "name": "Roy Luo", "email": "royluo@google.com", "time": "Tue Feb 04 23:36:42 2025 +0000" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:25:12 2025 +0100" }, "message": "usb: gadget: core: flush gadget workqueue after device removal\n\ndevice_del() can lead to new work being scheduled in gadget-\u003ework\nworkqueue. This is observed, for example, with the dwc3 driver with the\nfollowing call stack:\n device_del()\n gadget_unbind_driver()\n usb_gadget_disconnect_locked()\n dwc3_gadget_pullup()\n\t dwc3_gadget_soft_disconnect()\n\t usb_gadget_set_state()\n\t schedule_work(\u0026gadget-\u003ework)\n\nMove flush_work() after device_del() to ensure the workqueue is cleaned\nup.\n\nFixes: 5702f75375aa9 (\"usb: gadget: udc-core: move sysfs_notify() to a workqueue\")\nCc: stable \u003cstable@kernel.org\u003e\nSigned-off-by: Roy Luo \u003croyluo@google.com\u003e\nReviewed-by: Alan Stern \u003cstern@rowland.harvard.edu\u003e\nReviewed-by: Thinh Nguyen \u003cThinh.Nguyen@synopsys.com\u003e\nLink: https://lore.kernel.org/r/20250204233642.666991-1-royluo@google.com\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "4ab37fcb42832cdd3e9d5e50653285ca84d6686f", "tree": "938b6af7fdda3ae99977e6101128446314748df2", "parents": [ "4aac0db5a0ebc599d4ad9bf5ebab78afa1f33e10" ], "author": { "name": "Jill Donahue", "email": "jilliandonahue58@gmail.com", "time": "Tue Feb 11 10:48:05 2025 -0700" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:24:20 2025 +0100" }, "message": "USB: gadget: f_midi: f_midi_complete to call queue_work\n\nWhen using USB MIDI, a lock is attempted to be acquired twice through a\nre-entrant call to f_midi_transmit, causing a deadlock.\n\nFix it by using queue_work() to schedule the inner f_midi_transmit() via\na high priority work queue from the completion handler.\n\nLink: https://lore.kernel.org/all/CAArt\u003dLjxU0fUZOj06X+5tkeGT+6RbXzpWg1h4t4Fwa_KGVAX6g@mail.gmail.com/\nFixes: d5daf49b58661 (\"USB: gadget: midi: add midi function driver\")\nCc: stable \u003cstable@kernel.org\u003e\nSigned-off-by: Jill Donahue \u003cjilliandonahue58@gmail.com\u003e\nLink: https://lore.kernel.org/r/20250211174805.1369265-1-jdonahue@fender.com\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "4aac0db5a0ebc599d4ad9bf5ebab78afa1f33e10", "tree": "e05e5c5a3b8c68ca9a6719df5277beb83f490f7b", "parents": [ "d3a8c28426fc1fb3252753a9f1db0d691ffc21b0" ], "author": { "name": "Stefan Eichenberger", "email": "stefan.eichenberger@toradex.com", "time": "Mon Feb 03 11:58:24 2025 +0100" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:23:35 2025 +0100" }, "message": "usb: core: fix pipe creation for get_bMaxPacketSize0\n\nWhen usb_control_msg is used in the get_bMaxPacketSize0 function, the\nUSB pipe does not include the endpoint device number. This can cause\nfailures when a usb hub port is reinitialized after encountering a bad\ncable connection. As a result, the system logs the following error\nmessages:\nusb usb2-port1: cannot reset (err \u003d -32)\nusb usb2-port1: Cannot enable. Maybe the USB cable is bad?\nusb usb2-port1: attempt power cycle\nusb 2-1: new high-speed USB device number 5 using ci_hdrc\nusb 2-1: device descriptor read/8, error -71\n\nThe problem began after commit 85d07c556216 (\"USB: core: Unite old\nscheme and new scheme descriptor reads\"). There\nusb_get_device_descriptor was replaced with get_bMaxPacketSize0. Unlike\nusb_get_device_descriptor, the get_bMaxPacketSize0 function uses the\nmacro usb_rcvaddr0pipe, which does not include the endpoint device\nnumber. usb_get_device_descriptor, on the other hand, used the macro\nusb_rcvctrlpipe, which includes the endpoint device number.\n\nBy modifying the get_bMaxPacketSize0 function to use usb_rcvctrlpipe\ninstead of usb_rcvaddr0pipe, the issue can be resolved. This change will\nensure that the endpoint device number is included in the USB pipe,\npreventing reinitialization failures. If the endpoint has not set the\ndevice number yet, it will still work because the device number is 0 in\nudev.\n\nCc: stable \u003cstable@kernel.org\u003e\nFixes: 85d07c556216 (\"USB: core: Unite old scheme and new scheme descriptor reads\")\nSigned-off-by: Stefan Eichenberger \u003cstefan.eichenberger@toradex.com\u003e\nReviewed-by: Alan Stern \u003cstern@rowland.harvard.edu\u003e\nLink: https://lore.kernel.org/r/20250203105840.17539-1-eichest@gmail.com\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "d3a8c28426fc1fb3252753a9f1db0d691ffc21b0", "tree": "b518ff8bd609cf4617dc2a53bacae38b67fad582", "parents": [ "159daf1258227f44b26b5d38f4aa8f37b8cca663" ], "author": { "name": "Selvarasu Ganesan", "email": "selvarasu.g@samsung.com", "time": "Sat Feb 01 22:09:02 2025 +0530" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:23:23 2025 +0100" }, "message": "usb: dwc3: Fix timeout issue during controller enter/exit from halt state\n\nThere is a frequent timeout during controller enter/exit from halt state\nafter toggling the run_stop bit by SW. This timeout occurs when\nperforming frequent role switches between host and device, causing\ndevice enumeration issues due to the timeout. This issue was not present\nwhen USB2 suspend PHY was disabled by passing the SNPS quirks\n(snps,dis_u2_susphy_quirk and snps,dis_enblslpm_quirk) from the DTS.\nHowever, there is a requirement to enable USB2 suspend PHY by setting of\nGUSB2PHYCFG.ENBLSLPM and GUSB2PHYCFG.SUSPHY bits when controller starts\nin gadget or host mode results in the timeout issue.\n\nThis commit addresses this timeout issue by ensuring that the bits\nGUSB2PHYCFG.ENBLSLPM and GUSB2PHYCFG.SUSPHY are cleared before starting\nthe dwc3_gadget_run_stop sequence and restoring them after the\ndwc3_gadget_run_stop sequence is completed.\n\nFixes: 72246da40f37 (\"usb: Introduce DesignWare USB3 DRD Driver\")\nCc: stable \u003cstable@kernel.org\u003e\nSigned-off-by: Selvarasu Ganesan \u003cselvarasu.g@samsung.com\u003e\nAcked-by: Thinh Nguyen \u003cThinh.Nguyen@synopsys.com\u003e\nLink: https://lore.kernel.org/r/20250201163903.459-1-selvarasu.g@samsung.com\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "159daf1258227f44b26b5d38f4aa8f37b8cca663", "tree": "90955e1f64deaa9983188adf6814fe503bd9ab68", "parents": [ "7284922f3e4fa285dff1b8bb593aa9a0b8458f30" ], "author": { "name": "Mathias Nyman", "email": "mathias.nyman@linux.intel.com", "time": "Thu Feb 06 17:18:36 2025 +0200" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:23:02 2025 +0100" }, "message": "USB: Add USB_QUIRK_NO_LPM quirk for sony xperia xz1 smartphone\n\nThe fastboot tool for communicating with Android bootloaders does not\nwork reliably with this device if USB 2 Link Power Management (LPM)\nis enabled.\n\nVarious fastboot commands are affected, including the\nfollowing, which usually reproduces the problem within two tries:\n\n fastboot getvar kernel\n getvar:kernel FAILED (remote: \u0027GetVar Variable Not found\u0027)\n\nThis issue was hidden on many systems up until commit 63a1f8454962\n(\"xhci: stored cached port capability values in one place\") as the xhci\ndriver failed to detect USB 2 LPM support if USB 3 ports were listed\nbefore USB 2 ports in the \"supported protocol capabilities\".\n\nAdding the quirk resolves the issue. No drawbacks are expected since\nthe device uses different USB product IDs outside of fastboot mode, and\nsince fastboot commands worked before, until LPM was enabled on the\ntested system by the aforementioned commit.\n\nBased on a patch from Forest \u003cforestix@nom.one\u003e from which most of the\ncode and commit message is taken.\n\nCc: stable \u003cstable@kernel.org\u003e\nReported-by: Forest \u003cforestix@nom.one\u003e\nCloses: https://lore.kernel.org/hk8umj9lv4l4qguftdq1luqtdrpa1gks5l@sonic.net\nTested-by: Forest \u003cforestix@nom.one\u003e\nSigned-off-by: Mathias Nyman \u003cmathias.nyman@linux.intel.com\u003e\nLink: https://lore.kernel.org/r/20250206151836.51742-1-mathias.nyman@linux.intel.com\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "7284922f3e4fa285dff1b8bb593aa9a0b8458f30", "tree": "4c4b9c6d3e8f09394cc8d18152addff8718541d1", "parents": [ "12e712964f41d05ae034989892de445781c46730" ], "author": { "name": "Marek Vasut", "email": "marek.vasut+renesas@mailbox.org", "time": "Sun Feb 09 15:56:11 2025 +0100" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:22:36 2025 +0100" }, "message": "USB: cdc-acm: Fill in Renesas R-Car D3 USB Download mode quirk\n\nAdd Renesas R-Car D3 USB Download mode quirk and update comments\non all the other Renesas R-Car USB Download mode quirks to discern\nthem from each other. This follows R-Car Series, 3rd Generation\nreference manual Rev.2.00 chapter 19.2.8 USB download mode .\n\nFixes: 6d853c9e4104 (\"usb: cdc-acm: Add DISABLE_ECHO for Renesas USB Download mode\")\nCc: stable \u003cstable@kernel.org\u003e\nSigned-off-by: Marek Vasut \u003cmarek.vasut+renesas@mailbox.org\u003e\nReviewed-by: Geert Uytterhoeven \u003cgeert+renesas@glider.be\u003e\nLink: https://lore.kernel.org/r/20250209145708.106914-1-marek.vasut+renesas@mailbox.org\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "12e712964f41d05ae034989892de445781c46730", "tree": "d7eab502904e2d60a03df0a20e6aebd7694e5795", "parents": [ "e563b01208f4d1f609bcab13333b6c0e24ce6a01" ], "author": { "name": "Jann Horn", "email": "jannh@google.com", "time": "Wed Feb 12 19:15:16 2025 +0100" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:22:15 2025 +0100" }, "message": "usb: cdc-acm: Fix handling of oversized fragments\n\nIf we receive an initial fragment of size 8 bytes which specifies a wLength\nof 1 byte (so the reassembled message is supposed to be 9 bytes long), and\nwe then receive a second fragment of size 9 bytes (which is not supposed to\nhappen), we currently wrongly bypass the fragment reassembly code but still\npass the pointer to the acm-\u003enotification_buffer to\nacm_process_notification().\n\nMake this less wrong by always going through fragment reassembly when we\nexpect more fragments.\n\nBefore this patch, receiving an overlong fragment could lead to `newctrl`\nin acm_process_notification() being uninitialized data (instead of data\ncoming from the device).\n\nCc: stable \u003cstable@kernel.org\u003e\nFixes: ea2583529cd1 (\"cdc-acm: reassemble fragmented notifications\")\nSigned-off-by: Jann Horn \u003cjannh@google.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "e563b01208f4d1f609bcab13333b6c0e24ce6a01", "tree": "6f4b453316081ee4221fb3b06d8f7e4875f408aa", "parents": [ "c81d9fcd5b9402166048f377d4e5e0ee6f9ef26d" ], "author": { "name": "Jann Horn", "email": "jannh@google.com", "time": "Wed Feb 12 19:15:15 2025 +0100" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:21:56 2025 +0100" }, "message": "usb: cdc-acm: Check control transfer buffer size before access\n\nIf the first fragment is shorter than struct usb_cdc_notification, we can\u0027t\ncalculate an expected_size. Log an error and discard the notification\ninstead of reading lengths from memory outside the received data, which can\nlead to memory corruption when the expected_size decreases between\nfragments, causing `expected_size - acm-\u003enb_index` to wrap.\n\nThis issue has been present since the beginning of git history; however,\nit only leads to memory corruption since commit ea2583529cd1\n(\"cdc-acm: reassemble fragmented notifications\").\n\nA mitigating factor is that acm_ctrl_irq() can only execute after userspace\nhas opened /dev/ttyACM*; but if ModemManager is running, ModemManager will\ndo that automatically depending on the USB device\u0027s vendor/product IDs and\nits other interfaces.\n\nCc: stable \u003cstable@kernel.org\u003e\nFixes: 1da177e4c3f4 (\"Linux-2.6.12-rc2\")\nSigned-off-by: Jann Horn \u003cjannh@google.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "c81d9fcd5b9402166048f377d4e5e0ee6f9ef26d", "tree": "15028a709ce1e1e0b8f2bdc595e7232c6e684b85", "parents": [ "e71f7f42e3c874ac3314b8f250e8416a706165af" ], "author": { "name": "Michal Pecio", "email": "michal.pecio@gmail.com", "time": "Tue Jan 28 10:45:29 2025 +0100" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:21:24 2025 +0100" }, "message": "usb: xhci: Restore xhci_pci support for Renesas HCs\n\nSome Renesas HCs require firmware upload to work, this is handled by the\nxhci_pci_renesas driver. Other variants of those chips load firmware from\na SPI flash and are ready to work with xhci_pci alone.\n\nA refactor merged in v6.12 broke the latter configuration so that users\nare finding their hardware ignored by the normal driver and are forced to\nenable the firmware loader which isn\u0027t really necessary on their systems.\n\nLet xhci_pci work with those chips as before when the firmware loader is\ndisabled by kernel configuration.\n\nFixes: 25f51b76f90f (\"xhci-pci: Make xhci-pci-renesas a proper modular driver\")\nCc: stable \u003cstable@kernel.org\u003e\nCloses: https://bugzilla.kernel.org/show_bug.cgi?id\u003d219616\nCloses: https://bugzilla.kernel.org/show_bug.cgi?id\u003d219726\nSigned-off-by: Michal Pecio \u003cmichal.pecio@gmail.com\u003e\nTested-by: Nicolai Buchwitz \u003cnb@tipi-net.de\u003e\nLink: https://lore.kernel.org/r/20250128104529.58a79bfc@foxbook\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "e71f7f42e3c874ac3314b8f250e8416a706165af", "tree": "69cca98c3e3836a563b574733b9596c45e0e6052", "parents": [ "71db7b9a019b76df43512639c282c03733ba3eeb" ], "author": { "name": "Huacai Chen", "email": "chenhuacai@loongson.cn", "time": "Sun Feb 02 20:49:35 2025 +0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@linuxfoundation.org", "time": "Fri Feb 14 09:18:14 2025 +0100" }, "message": "USB: pci-quirks: Fix HCCPARAMS register error for LS7A EHCI\n\nLS7A EHCI controller doesn\u0027t have extended capabilities, so the EECP\n(EHCI Extended Capabilities Pointer) field of HCCPARAMS register should\nbe 0x0, but it reads as 0xa0 now. This is a hardware flaw and will be\nfixed in future, now just clear the EECP field to avoid error messages\non boot:\n\n......\n[ 0.581675] pci 0000:00:04.1: EHCI: unrecognized capability ff\n[ 0.581699] pci 0000:00:04.1: EHCI: unrecognized capability ff\n[ 0.581716] pci 0000:00:04.1: EHCI: unrecognized capability ff\n[ 0.581851] pci 0000:00:04.1: EHCI: unrecognized capability ff\n......\n[ 0.581916] pci 0000:00:05.1: EHCI: unrecognized capability ff\n[ 0.581951] pci 0000:00:05.1: EHCI: unrecognized capability ff\n[ 0.582704] pci 0000:00:05.1: EHCI: unrecognized capability ff\n[ 0.582799] pci 0000:00:05.1: EHCI: unrecognized capability ff\n......\n\nCc: stable \u003cstable@kernel.org\u003e\nSigned-off-by: Baoqi Zhang \u003czhangbaoqi@loongson.cn\u003e\nSigned-off-by: Huacai Chen \u003cchenhuacai@loongson.cn\u003e\nLink: https://lore.kernel.org/r/20250202124935.480500-1-chenhuacai@loongson.cn\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\n" }, { "commit": "add43c4fbc92f8b48c1acd64e953af3b1be4cd9c", "tree": "c0768c2eb8eb73b531f8c46eac647a158e35d973", "parents": [ "4a8991fe9cd0b6a509bab3d056700d3520601d86" ], "author": { "name": "Lu Baolu", "email": "baolu.lu@linux.intel.com", "time": "Tue Feb 11 08:55:12 2025 +0800" }, "committer": { "name": "Joerg Roedel", "email": "jroedel@suse.de", "time": "Fri Feb 14 09:12:47 2025 +0100" }, "message": "iommu/vt-d: Make intel_iommu_drain_pasid_prq() cover faults for RID\n\nThis driver supports page faults on PCI RID since commit \u003c9f831c16c69e\u003e\n(\"iommu/vt-d: Remove the pasid present check in prq_event_thread\") by\nallowing the reporting of page faults with the pasid_present field cleared\nto the upper layer for further handling. The fundamental assumption here\nis that the detach or replace operations act as a fence for page faults.\nThis implies that all pending page faults associated with a specific RID\nor PASID are flushed when a domain is detached or replaced from a device\nRID or PASID.\n\nHowever, the intel_iommu_drain_pasid_prq() helper does not correctly\nhandle faults for RID. This leads to faults potentially remaining pending\nin the iommu hardware queue even after the domain is detached, thereby\nviolating the aforementioned assumption.\n\nFix this issue by extending intel_iommu_drain_pasid_prq() to cover faults\nfor RID.\n\nFixes: 9f831c16c69e (\"iommu/vt-d: Remove the pasid present check in prq_event_thread\")\nCc: stable@vger.kernel.org\nSuggested-by: Kevin Tian \u003ckevin.tian@intel.com\u003e\nSigned-off-by: Lu Baolu \u003cbaolu.lu@linux.intel.com\u003e\nReviewed-by: Kevin Tian \u003ckevin.tian@intel.com\u003e\nLink: https://lore.kernel.org/r/20250121023150.815972-1-baolu.lu@linux.intel.com\nReviewed-by: Yi Liu \u003cyi.l.liu@intel.com\u003e\nLink: https://lore.kernel.org/r/20250211005512.985563-2-baolu.lu@linux.intel.com\nSigned-off-by: Joerg Roedel \u003cjroedel@suse.de\u003e\n" }, { "commit": "4a8991fe9cd0b6a509bab3d056700d3520601d86", "tree": "a39c36699b2e7ed7f3e9c65e31edb7d3aa5e8691", "parents": [ "78be7f04537fa35f6cc694879e9a475ca1984936" ], "author": { "name": "Andrew Kreimer", "email": "algonell@gmail.com", "time": "Mon Feb 10 13:20:04 2025 +0200" }, "committer": { "name": "Joerg Roedel", "email": "jroedel@suse.de", "time": "Fri Feb 14 09:12:46 2025 +0100" }, "message": "iommu/exynos: Fix typos\n\nThere are some typos in comments/messages:\n - modyfying -\u003e modifying\n - Unabled -\u003e Unable\n\nFix them via codespell.\n\nSigned-off-by: Andrew Kreimer \u003calgonell@gmail.com\u003e\nLink: https://lore.kernel.org/r/20250210112027.29791-1-algonell@gmail.com\nSigned-off-by: Joerg Roedel \u003cjroedel@suse.de\u003e\n" }, { "commit": "78be7f04537fa35f6cc694879e9a475ca1984936", "tree": "416dce106db97def8f15c1e6495b9d583a120a96", "parents": [ "ef75966abf950c0539534effa4960caa29fb7167" ], "author": { "name": "Easwar Hariharan", "email": "eahariha@linux.microsoft.com", "time": "Tue Jan 28 19:05:21 2025 +0000" }, "committer": { "name": "Joerg Roedel", "email": "jroedel@suse.de", "time": "Fri Feb 14 09:12:46 2025 +0100" }, "message": "iommu: Fix a spelling error\n\nFix spelling error IDENITY -\u003e IDENTITY in\ndrivers/iommu/iommu.c.\n\nSigned-off-by: Easwar Hariharan \u003ceahariha@linux.microsoft.com\u003e\nReviewed-by: Jason Gunthorpe \u003cjgg@nvidia.com\u003e\nLink: https://lore.kernel.org/r/20250128190522.70800-1-eahariha@linux.microsoft.com\n[ joro: Add commit message ]\nSigned-off-by: Joerg Roedel \u003cjroedel@suse.de\u003e\n" }, { "commit": "ef75966abf950c0539534effa4960caa29fb7167", "tree": "e6a21648a367598a03a2bc31d0846c0932c71f9b", "parents": [ "9759ae2cee7cd42b95f1c48aa3749bd02b5ddb08" ], "author": { "name": "Vasant Hegde", "email": "vasant.hegde@amd.com", "time": "Mon Jan 27 09:44:11 2025 +0000" }, "committer": { "name": "Joerg Roedel", "email": "jroedel@suse.de", "time": "Fri Feb 14 09:11:58 2025 +0100" }, "message": "iommu/amd: Expicitly enable CNTRL.EPHEn bit in resume path\n\nWith recent kernel, AMDGPU failed to resume after suspend on certain laptop.\n\nSample log:\n-----------\nNov 14 11:52:19 Thinkbook kernel: iommu ivhd0: AMD-Vi: Event logged [ILLEGAL_DEV_TABLE_ENTRY device\u003d0000:06:00.0 pasid\u003d0x00000 address\u003d0x135300000 flags\u003d0x0080]\nNov 14 11:52:19 Thinkbook kernel: AMD-Vi: DTE[0]: 7d90000000000003\nNov 14 11:52:19 Thinkbook kernel: AMD-Vi: DTE[1]: 0000100103fc0009\nNov 14 11:52:19 Thinkbook kernel: AMD-Vi: DTE[2]: 2000000117840013\nNov 14 11:52:19 Thinkbook kernel: AMD-Vi: DTE[3]: 0000000000000000\n\nThis is because in resume path, CNTRL[EPHEn] is not set. Fix this by\nsetting CNTRL[EPHEn] to 1 in resume path if EFR[EPHSUP] is set.\n\nNote\n May be better approach is to save the control register in suspend path\n and restore it in resume path instead of trying to set indivisual\n bits. We will have separate patch for that.\n\nCloses: https://bugzilla.kernel.org/show_bug.cgi?id\u003d219499\nFixes: c4cb23111103 (\"iommu/amd: Add support for enable/disable IOPF\")\nTested-by: Hamish McIntyre-Bhatty \u003ckernel-bugzilla@regd.hamishmb.com\u003e\nSigned-off-by: Vasant Hegde \u003cvasant.hegde@amd.com\u003e\nLink: https://lore.kernel.org/r/20250127094411.5931-1-vasant.hegde@amd.com\nSigned-off-by: Joerg Roedel \u003cjroedel@suse.de\u003e\n" }, { "commit": "128c8f96eb8638c060cd3532dc394d046ce64fe1", "tree": "cfba5f395ada6310c488e9ab473b92b508cebc68", "parents": [ "68763b29e0a6441f57f9ee652bbf8e7bc59183e5", "04485cc34868892a823b909c6d5468ba21b63569" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 20:04:43 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 20:04:43 2025 -0800" }, "message": "Merge tag \u0027drm-fixes-2025-02-14\u0027 of https://gitlab.freedesktop.org/drm/kernel\n\nPull drm fixes from Dave Airlie:\n \"Weekly drm fixes pull request, nothing too unusual, the hdmi tests\n needs a bit of refactoring after lockdep shouted at them, otherwise\n amdgpu and xe lead and a few misc otherwise.\n\n amdgpu:\n - Fix shutdown regression on old APUs\n - Fix compute queue hang on gfx9 APUs\n - Fix possible invalid access in PSP failure path\n - Avoid possible buffer overflow in pptable override\n\n amdkfd:\n - Properly free gang bo in failure path\n - GFX12 trap handler fix\n\n i915:\n - selftest fix: avoid using uninitialized context\n\n xe:\n - Remove bo-\u003eclients out of bos_lock area\n - Carve out wopcm portion from the stolen memory\n\n tests:\n - fix lockdep with hdmi infrastructure tests\n\n host1x:\n - fix uninitialised mutex usage\n\n panthor:\n - fix uninit variable\n\n hibmc:\n - fix missing Kconfig select\"\n\n* tag \u0027drm-fixes-2025-02-14\u0027 of https://gitlab.freedesktop.org/drm/kernel:\n drm: Fix DSC BPP increment decoding\n drm/amdgpu: avoid buffer overflow attach in smu_sys_set_pp_table()\n drm/amdkfd: Ensure consistent barrier state saved in gfx12 trap handler\n drm/amdgpu: bail out when failed to load fw in psp_init_cap_microcode()\n amdkfd: properly free gang_ctx_bo when failed to init user queue\n drm/amdgpu: bump version for RV/PCO compute fix\n drm/amdgpu/gfx9: manually control gfxoff for CS on RV\n drm/amdgpu/pm: fix UVD handing in amdgpu_dpm_set_powergating_by_smu()\n drm/xe: Carve out wopcm portion from the stolen memory\n drm/i915/selftests: avoid using uninitialized context\n drm/xe/client: bo-\u003eclient does not need bos_lock\n drm/hisilicon/hibmc: select CONFIG_DRM_DISPLAY_DP_HELPER\n drm/panthor: avoid garbage value in panthor_ioctl_dev_query()\n gpu: host1x: Fix a use of uninitialized mutex\n drm/tests: hdmi: Fix recursive locking\n drm/tests: hdmi: Reorder DRM entities variables assignment\n drm/tests: hdmi: Remove redundant assignments\n drm/tests: hdmi: Fix WW_MUTEX_SLOWPATH failures\n" }, { "commit": "04485cc34868892a823b909c6d5468ba21b63569", "tree": "6c0c974a097b8456c7886ba937b0b66016b7083d", "parents": [ "d70c6ae53d790d3528a14092ae208d0ed125d7a5", "e977499820782ab1c69f354d9f41b6d9ad1f43d9" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri Feb 14 12:15:53 2025 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri Feb 14 12:15:59 2025 +1000" }, "message": "Merge tag \u0027drm-xe-fixes-2025-02-13\u0027 of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes\n\n- Remove bo-\u003eclients out of bos_lock area (Tejas)\n- Carve out wopcm portion from the stolen memory (Nirmoy)\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n\nFrom: Rodrigo Vivi \u003crodrigo.vivi@intel.com\u003e\nLink: https://patchwork.freedesktop.org/patch/msgid/Z64rCicgpBe_t5GY@intel.com\n" }, { "commit": "d70c6ae53d790d3528a14092ae208d0ed125d7a5", "tree": "5a8000b3805ba99de3b528b5a3e21ed46ba49a09", "parents": [ "981724b463141cf828744320ee8c93468d5dbe01", "53139b3f9998ea07289e7b70b909fea2264a0de9" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri Feb 14 12:11:19 2025 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri Feb 14 12:11:28 2025 +1000" }, "message": "Merge tag \u0027drm-intel-fixes-2025-02-13\u0027 of https://gitlab.freedesktop.org/drm/i915/kernel into drm-fixes\n\n- Selftest fix: avoid using uninitialized context\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n\nFrom: Rodrigo Vivi \u003crodrigo.vivi@intel.com\u003e\nLink: https://patchwork.freedesktop.org/patch/msgid/Z64qg13R_72iN3_X@intel.com\n" }, { "commit": "981724b463141cf828744320ee8c93468d5dbe01", "tree": "86c3ee8553d0be55bd9ef5f4b4fc02c2581d0e6a", "parents": [ "79f9efa085f98353d0a7452806b82d6fe0923194", "1abb2648698bf10783d2236a6b4a7ca5e8021699" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri Feb 14 11:54:19 2025 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri Feb 14 11:54:24 2025 +1000" }, "message": "Merge tag \u0027amd-drm-fixes-6.14-2025-02-13\u0027 of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes\n\namd-drm-fixes-6.14-2025-02-13:\n\namdgpu:\n- Fix shutdown regression on old APUs\n- Fix compute queue hang on gfx9 APUs\n- Fix possible invalid access in PSP failure path\n- Avoid possible buffer overflow in pptable override\n\namdkfd:\n- Properly free gang bo in failure path\n- GFX12 trap handler fix\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n\nFrom: Alex Deucher \u003calexander.deucher@amd.com\u003e\nLink: https://patchwork.freedesktop.org/patch/msgid/20250213153843.242640-1-alexander.deucher@amd.com\n" }, { "commit": "79f9efa085f98353d0a7452806b82d6fe0923194", "tree": "6a7748b15b84d3204219abfbbd62e5e7dbcff74d", "parents": [ "a64dcfb451e254085a7daee5fe51bf22959d52d3", "e00a2e5d485faf53c7a24b9d1b575a642227947f" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri Feb 14 10:32:31 2025 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri Feb 14 10:32:35 2025 +1000" }, "message": "Merge tag \u0027drm-misc-fixes-2025-02-13\u0027 of https://gitlab.freedesktop.org/drm/misc/kernel into drm-fixes\n\nSome locking fixes for the HDMI infrastructure tests, an unitialized\nmutex fix for host1x, an unitialized variable fix for panthor, and a\nconfig selection fix for hibmc.\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n\nFrom: Maxime Ripard \u003cmripard@redhat.com\u003e\nLink: https://patchwork.freedesktop.org/patch/msgid/20250213-brilliant-terrier-from-hell-d06dd5@houat\n" }, { "commit": "7422c319fd805b956aab5ba93e0274517a8e3650", "tree": "3905c22bb893445cab30c023d59b9b8d626f340e", "parents": [ "320702a76186222426e5dc8efb9d68ba9d4ed0ab" ], "author": { "name": "Mukesh Kumar Savaliya", "email": "quic_msavaliy@quicinc.com", "time": "Thu Jan 23 14:11:47 2025 +0530" }, "committer": { "name": "Andi Shyti", "email": "andi.shyti@kernel.org", "time": "Thu Feb 13 23:23:55 2025 +0100" }, "message": "MAINTAINERS: Add maintainer for Qualcomm\u0027s I2C GENI driver\n\nAdd a new entry for the I2C QCOM GENI driver to the MAINTAINERS file.\nThis entry includes the maintainer\u0027s name and contact information,\nensuring proper maintainership and communication for the i2c-qcom-geni\ndriver file.\n\nSigned-off-by: Mukesh Kumar Savaliya \u003cquic_msavaliy@quicinc.com\u003e\nLink: https://lore.kernel.org/r/20250123084147.3632023-1-quic_msavaliy@quicinc.com\nSigned-off-by: Andi Shyti \u003candi.shyti@kernel.org\u003e\n" }, { "commit": "320702a76186222426e5dc8efb9d68ba9d4ed0ab", "tree": "8f6cbd69794a56f3166bfc1af97a83f63ee62410", "parents": [ "a64dcfb451e254085a7daee5fe51bf22959d52d3" ], "author": { "name": "Wolfram Sang", "email": "wsa+renesas@sang-engineering.com", "time": "Thu Feb 13 17:29:51 2025 +0100" }, "committer": { "name": "Andi Shyti", "email": "andi.shyti@kernel.org", "time": "Thu Feb 13 23:20:09 2025 +0100" }, "message": "MAINTAINERS: delete entry for AXXIA I2C\n\nThe maintainer\u0027s email address bounced and he wasn\u0027t active for 4 years.\nDelete this entry and fall back to the generic I2C host drivers entry.\n\nSigned-off-by: Wolfram Sang \u003cwsa+renesas@sang-engineering.com\u003e\nLink: https://lore.kernel.org/r/20250213162950.45596-2-wsa+renesas@sang-engineering.com\nSigned-off-by: Andi Shyti \u003candi.shyti@kernel.org\u003e\n" }, { "commit": "68763b29e0a6441f57f9ee652bbf8e7bc59183e5", "tree": "12c3ef58c04a2954588b1a5a8da321ec20336c69", "parents": [ "85cc5751db7cc3211945bc380c944de6fe6979d1", "3588b1c0fde2f58d166e3f94a5a58d64b893526c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 13:13:37 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 13:13:37 2025 -0800" }, "message": "Merge tag \u0027spi-fix-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi\n\nPull spi fixes from Mark Brown:\n \"A small collection of driver specific fixes, none standing out in\n particular\"\n\n* tag \u0027spi-fix-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:\n spi: sn-f-ospi: Fix division by zero\n spi: pxa2xx: Fix regression when toggling chip select on LPSS devices\n spi: atmel-quadspi: Fix warning in doc-comment\n" }, { "commit": "85cc5751db7cc3211945bc380c944de6fe6979d1", "tree": "f122ca800fc63cfad856a1d0a04a9ea4901c0925", "parents": [ "f9af8e7718904f0b77595d1b610c96dd7796c655", "35e21de48e693af1dcfdbf2dc3d73dcfa3c8f2d9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 13:09:01 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 13:09:01 2025 -0800" }, "message": "Merge tag \u0027regulator-fix-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator\n\nPull regulator fixes from Mark Brown:\n \"The main change here is a revert for a cleanup that was done in the\n core, attempting to resolve some confusion about how we handle systems\n where we\u0027ve somehow managed to end up with both platform data and\n device tree data for the same device. Unfortunately it turns out there\n are actually a few systems that deliberately do this and were broken\n by the change so we\u0027ve just reverted it.\n\n There\u0027s also a new Qualcomm device ID\"\n\n* tag \u0027regulator-fix-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:\n regulator: core: let dt properties override driver init_data\n regulator: qcom_smd: Add l2, l5 sub-node to mp5496 regulator\n" }, { "commit": "f9af8e7718904f0b77595d1b610c96dd7796c655", "tree": "cf782d50302b4e155dea40628aa81f82780e2a8a", "parents": [ "348f968b89bfeec0bb53dd82dba58b94d97fbd34", "32ffed055dcee17f6705f545b069e44a66067808" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 13:07:30 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 13:07:30 2025 -0800" }, "message": "Merge tag \u0027regmap-fix-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap\n\nPull regmap fix from Mark Brown:\n \"A simple fix for memory leaks when deallocating regmap-irq\n controllers\"\n\n* tag \u0027regmap-fix-v6.14-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:\n regmap-irq: Add missing kfree()\n" }, { "commit": "348f968b89bfeec0bb53dd82dba58b94d97fbd34", "tree": "6346da22168bc50cc5d077405aad86f982b962ca", "parents": [ "945ce413ac14388219afe09de84ee08994f05e53", "488fb6effe03e20f38d34da7425de77bbd3e2665" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 12:17:04 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 12:17:04 2025 -0800" }, "message": "Merge tag \u0027net-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net\n\nPull networking fixes from Jakub Kicinski:\n \"Including fixes from netfilter, wireless and bluetooth.\n\n Kalle Valo steps down after serving as the WiFi driver maintainer for\n over a decade.\n\n Current release - fix to a fix:\n\n - vsock: orphan socket after transport release, avoid null-deref\n\n - Bluetooth: L2CAP: fix corrupted list in hci_chan_del\n\n Current release - regressions:\n\n - eth:\n - stmmac: correct Rx buffer layout when SPH is enabled\n - iavf: fix a locking bug in an error path\n\n - rxrpc: fix alteration of headers whilst zerocopy pending\n\n - s390/qeth: move netif_napi_add_tx() and napi_enable() from under BH\n\n - Revert \"netfilter: flowtable: teardown flow if cached mtu is stale\"\n\n Current release - new code bugs:\n\n - rxrpc: fix ipv6 path MTU discovery, only ipv4 worked\n\n - pse-pd: fix deadlock in current limit functions\n\n Previous releases - regressions:\n\n - rtnetlink: fix netns refleak with rtnl_setlink()\n\n - wifi: brcmfmac: use random seed flag for BCM4355 and BCM4364\n firmware\n\n Previous releases - always broken:\n\n - add missing RCU protection of struct net throughout the stack\n\n - can: rockchip: bail out if skb cannot be allocated\n\n - eth: ti: am65-cpsw: base XDP support fixes\n\n Misc:\n\n - ethtool: tsconfig: update the format of hwtstamp flags, changes the\n uAPI but this uAPI was not in any release yet\"\n\n* tag \u0027net-6.14-rc3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (72 commits)\n net: pse-pd: Fix deadlock in current limit functions\n rxrpc: Fix ipv6 path MTU discovery\n Reapply \"net: skb: introduce and use a single page frag cache\"\n s390/qeth: move netif_napi_add_tx() and napi_enable() from under BH\n mlxsw: Add return value check for mlxsw_sp_port_get_stats_raw()\n ipv6: mcast: add RCU protection to mld_newpack()\n team: better TEAM_OPTION_TYPE_STRING validation\n Bluetooth: L2CAP: Fix corrupted list in hci_chan_del\n Bluetooth: btintel_pcie: Fix a potential race condition\n Bluetooth: L2CAP: Fix slab-use-after-free Read in l2cap_send_cmd\n net: ethernet: ti: am65_cpsw: fix tx_cleanup for XDP case\n net: ethernet: ti: am65-cpsw: fix RX \u0026 TX statistics for XDP_TX case\n net: ethernet: ti: am65-cpsw: fix memleak in certain XDP cases\n vsock/test: Add test for SO_LINGER null ptr deref\n vsock: Orphan socket after transport release\n MAINTAINERS: Add sctp headers to the general netdev entry\n Revert \"netfilter: flowtable: teardown flow if cached mtu is stale\"\n iavf: Fix a locking bug in an error path\n rxrpc: Fix alteration of headers whilst zerocopy pending\n net: phylink: make configuring clock-stop dependent on MAC support\n ...\n" }, { "commit": "945ce413ac14388219afe09de84ee08994f05e53", "tree": "04ff4367e7403406337a83a847c9a1b255b654c7", "parents": [ "1854c7f79dcaaba9f1c0b131445ace03f9fd532d", "da2dccd7451de62b175fb8f0808d644959e964c7" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 12:06:29 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 12:06:29 2025 -0800" }, "message": "Merge tag \u0027for-6.14-rc2-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux\n\nPull btrfs fixes from David Sterba:\n\n - fix stale page cache after race between readahead and direct IO write\n\n - fix hole expansion when writing at an offset beyond EOF, the range\n will not be zeroed\n\n - use proper way to calculate offsets in folio ranges\n\n* tag \u0027for-6.14-rc2-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:\n btrfs: fix hole expansion when writing at an offset beyond EOF\n btrfs: fix stale page cache after race between readahead and direct IO write\n btrfs: fix two misuses of folio_shift()\n" }, { "commit": "1854c7f79dcaaba9f1c0b131445ace03f9fd532d", "tree": "329fba56f58c788616b13f4019a815b0d129ab25", "parents": [ "ab68d7eb7b1a64f3f4710da46cc5f93c6c154942", "406e445b3c6be65ab0ee961145e74bfd7ef6c9e1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 11:58:11 2025 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Feb 13 11:58:11 2025 -0800" }, "message": "Merge tag \u0027bcachefs-2025-02-12\u0027 of git://evilpiepirate.org/bcachefs\n\nPull bcachefs fixes from Kent Overstreet:\n \"Just small stuff.\n\n As a general announcement, on disk format is now frozen in my master\n branch - future on disk format changes will be optional, not required.\n\n - More fixes for going read-only: the previous fix was insufficient,\n but with more work on ordering journal reclaim flushing (and a\n btree node accounting fix so we don\u0027t split until we have to) the\n tiering_replication test now consistently goes read-only in less\n than a second.\n\n - fix for fsck when we have reflink pointers to missing indirect\n extents\n\n - some transaction restart handling fixes from Alan; the \"Pass\n _orig_restart_count to trans_was_restarted\" likely fixes some rare\n undefined behaviour heisenbugs\"\n\n* tag \u0027bcachefs-2025-02-12\u0027 of git://evilpiepirate.org/bcachefs:\n bcachefs: Reuse transaction\n bcachefs: Pass _orig_restart_count to trans_was_restarted\n bcachefs: CONFIG_BCACHEFS_INJECT_TRANSACTION_RESTARTS\n bcachefs: Fix want_new_bset() so we write until the end of the btree node\n bcachefs: Split out journal pins by btree level\n bcachefs: Fix use after free\n bcachefs: Fix marking reflink pointers to missing indirect extents\n" }, { "commit": "b3aa9283c0c505b5cfd25f7d6cfd720de2adc807", "tree": "753ed87f4933a91d92ecf166c8cbd5b4c1be005b", "parents": [ "e6e3e0022ef8f1d584ee4d5b89dca02472c5eb1f" ], "author": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Wed Feb 12 18:25:58 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 18:03:54 2025 +0000" }, "message": "KVM: arm64: vgic: Hoist SGI/PPI alloc from vgic_init() to kvm_create_vgic()\n\nIf userspace creates vcpus, then a vgic, we end-up in a situation\nwhere irqchip_in_kernel() will return true, but no private interrupt\nhas been allocated for these vcpus. This situation will continue\nuntil userspace initialises the vgic, at which point we fix the\nearly vcpus. Should a vcpu run or be initialised in the interval,\nbad things may happen.\n\nAn obvious solution is to move this fix-up phase to the point where\nthe vgic is created. This ensures that from that point onwards,\nall vcpus have their private interrupts, as new vcpus will directly\nallocate them.\n\nWith that, we have the invariant that when irqchip_in_kernel() is\ntrue, all vcpus have their private interrupts.\n\nReported-by: Alexander Potapenko \u003cglider@google.com\u003e\nReviewed-by: Oliver Upton \u003coliver.upton@linux.dev\u003e\nLink: https://lore.kernel.org/r/20250212182558.2865232-3-maz@kernel.org\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "e6e3e0022ef8f1d584ee4d5b89dca02472c5eb1f", "tree": "98d5151a02128ba690b14dd892d32b172e434f41", "parents": [ "b938731ed2d4eea8e268a27bfc600581fedae2a9" ], "author": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Wed Feb 12 18:25:57 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 18:03:54 2025 +0000" }, "message": "KVM: arm64: timer: Drop warning on failed interrupt signalling\n\nWe currently spit out a warning if making a timer interrupt pending\nfails. But not only this is loud and easy to trigger from userspace,\nwe also fail to do anything useful with that information.\n\nDropping the warning is the easiest thing to do for now. We can\nalways add error reporting if we really want in the future.\n\nReported-by: Alexander Potapenko \u003cglider@google.com\u003e\nReviewed-by: Oliver Upton \u003coliver.upton@linux.dev\u003e\nLink: https://lore.kernel.org/r/20250212182558.2865232-2-maz@kernel.org\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "488fb6effe03e20f38d34da7425de77bbd3e2665", "tree": "4d3eaee7d33827df98a9b993c2f73a26afedbd17", "parents": [ "540cda75884a6ba4c289980c84392261b1f61a9c" ], "author": { "name": "Kory Maincent", "email": "kory.maincent@bootlin.com", "time": "Wed Feb 12 16:17:51 2025 +0100" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Feb 13 10:00:39 2025 -0800" }, "message": "net: pse-pd: Fix deadlock in current limit functions\n\nFix a deadlock in pse_pi_get_current_limit and pse_pi_set_current_limit\ncaused by consecutive mutex_lock calls. One in the function itself and\nanother in pse_pi_get_voltage.\n\nResolve the issue by using the unlocked version of pse_pi_get_voltage\ninstead.\n\nFixes: e0a5e2bba38a (\"net: pse-pd: Use power limit at driver side instead of current limit\")\nSigned-off-by: Kory Maincent \u003ckory.maincent@bootlin.com\u003e\nLink: https://patch.msgid.link/20250212151751.1515008-1-kory.maincent@bootlin.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "540cda75884a6ba4c289980c84392261b1f61a9c", "tree": "76c56bce8738059d57a1036fb5584ac33f49c5d8", "parents": [ "82c260c8806bf248ee11a053dd69665b8b207531" ], "author": { "name": "David Howells", "email": "dhowells@redhat.com", "time": "Wed Feb 12 11:21:24 2025 +0000" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Feb 13 10:00:12 2025 -0800" }, "message": "rxrpc: Fix ipv6 path MTU discovery\n\nrxrpc path MTU discovery currently only makes use of ICMPv4, but not\nICMPv6, which means that pmtud for IPv6 doesn\u0027t work correctly. Fix it to\ncheck for ICMPv6 messages also.\n\nFixes: eeaedc5449d9 (\"rxrpc: Implement path-MTU probing using padded PING ACKs (RFC8899)\")\nSigned-off-by: David Howells \u003cdhowells@redhat.com\u003e\ncc: Marc Dionne \u003cmarc.dionne@auristor.com\u003e\ncc: Simon Horman \u003chorms@kernel.org\u003e\ncc: linux-afs@lists.infradead.org\nLink: https://patch.msgid.link/3517283.1739359284@warthog.procyon.org.uk\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "b938731ed2d4eea8e268a27bfc600581fedae2a9", "tree": "51f03a7fb0c8fd13a2c27783dfc1e9b6b8fe8f59", "parents": [ "65729da9ce37f5a2c62e2542ef03bc9ac6775a7d" ], "author": { "name": "Quentin Perret", "email": "qperret@google.com", "time": "Thu Feb 13 15:36:14 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 17:57:27 2025 +0000" }, "message": "KVM: arm64: Fix alignment of kvm_hyp_memcache allocations\n\nWhen allocating guest stage-2 page-table pages at EL2, pKVM can consume\npages from the host-provided kvm_hyp_memcache. As pgtable.c expects\nzeroed pages, guest_s2_zalloc_page() actively implements this zeroing\nwith a PAGE_SIZE memset. Unfortunately, we don\u0027t check the page\nalignment of the host-provided address before doing so, which could\nlead to the memset overrunning the page if the host was malicious.\n\nFix this by simply force-aligning all kvm_hyp_memcache allocations to\npage boundaries.\n\nFixes: 60dfe093ec13 (\"KVM: arm64: Instantiate guest stage-2 page-tables at EL2\")\nReported-by: Ben Simner \u003cben.simner@cl.cam.ac.uk\u003e\nSigned-off-by: Quentin Perret \u003cqperret@google.com\u003e\nLink: https://lore.kernel.org/r/20250213153615.3642515-1-qperret@google.com\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "65729da9ce37f5a2c62e2542ef03bc9ac6775a7d", "tree": "3a63dee64a913d3b054f95acfe0677ad425718dd", "parents": [ "332b7e6d62b7a3a988017f5184e547aa20e3a19a" ], "author": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Wed Feb 12 17:34:54 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 17:55:13 2025 +0000" }, "message": "KVM: arm64: Convert timer offset VA when accessed in HYP code\n\nNow that EL2 has gained some early timer emulation, it accesses\nthe offsets pointed to by the timer structure, both of which\nlive in the KVM structure.\n\nOf course, these are *kernel* pointers, so the dereferencing\nof these pointers in non-kernel code must be itself be offset.\n\nGiven switch.h its own version of timer_get_offset() and use that\ninstead.\n\nFixes: b86fc215dc26d (\"KVM: arm64: Handle counter access early in non-HYP context\")\nReported-by: Linux Kernel Functional Testing \u003clkft@linaro.org\u003e\nReviewed-by: Oliver Upton \u003coliver.upton@linux.dev\u003e\nTested-by: Anders Roxell \u003canders.roxell@linaro.org\u003e\nLink: https://lore.kernel.org/r/20250212173454.2864462-1-maz@kernel.org\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "332b7e6d62b7a3a988017f5184e547aa20e3a19a", "tree": "0ab6038475abb599eebce85851ed45c1fa793d7e", "parents": [ "59419f10045bc955d2229819c7cf7a8b0b9c5b59" ], "author": { "name": "Mark Rutland", "email": "mark.rutland@arm.com", "time": "Thu Feb 13 09:15:31 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 17:55:13 2025 +0000" }, "message": "KVM: arm64: Simplify warning in kvm_arch_vcpu_load_fp()\n\nAt the end of kvm_arch_vcpu_load_fp() we check that no bits are set in\nSVCR. We only check this for protected mode despite this mattering\nequally for non-protected mode, and the comment above this is confusing.\n\nRemove the comment and simplify the check, moving from WARN_ON() to\nWARN_ON_ONCE() to avoid spamming the log.\n\nSigned-off-by: Mark Rutland \u003cmark.rutland@arm.com\u003e\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "59419f10045bc955d2229819c7cf7a8b0b9c5b59", "tree": "8dd38ee26f5847d567e9a387505a1093c7a8d9b3", "parents": [ "f9dd00de1e53a47763dfad601635d18542c3836d" ], "author": { "name": "Mark Rutland", "email": "mark.rutland@arm.com", "time": "Mon Feb 10 19:52:26 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 17:55:06 2025 +0000" }, "message": "KVM: arm64: Eagerly switch ZCR_EL{1,2}\n\nIn non-protected KVM modes, while the guest FPSIMD/SVE/SME state is live on the\nCPU, the host\u0027s active SVE VL may differ from the guest\u0027s maximum SVE VL:\n\n* For VHE hosts, when a VM uses NV, ZCR_EL2 contains a value constrained\n by the guest hypervisor, which may be less than or equal to that\n guest\u0027s maximum VL.\n\n Note: in this case the value of ZCR_EL1 is immaterial due to E2H.\n\n* For nVHE/hVHE hosts, ZCR_EL1 contains a value written by the guest,\n which may be less than or greater than the guest\u0027s maximum VL.\n\n Note: in this case hyp code traps host SVE usage and lazily restores\n ZCR_EL2 to the host\u0027s maximum VL, which may be greater than the\n guest\u0027s maximum VL.\n\nThis can be the case between exiting a guest and kvm_arch_vcpu_put_fp().\nIf a softirq is taken during this period and the softirq handler tries\nto use kernel-mode NEON, then the kernel will fail to save the guest\u0027s\nFPSIMD/SVE state, and will pend a SIGKILL for the current thread.\n\nThis happens because kvm_arch_vcpu_ctxsync_fp() binds the guest\u0027s live\nFPSIMD/SVE state with the guest\u0027s maximum SVE VL, and\nfpsimd_save_user_state() verifies that the live SVE VL is as expected\nbefore attempting to save the register state:\n\n| if (WARN_ON(sve_get_vl() !\u003d vl)) {\n| force_signal_inject(SIGKILL, SI_KERNEL, 0, 0);\n| return;\n| }\n\nFix this and make this a bit easier to reason about by always eagerly\nswitching ZCR_EL{1,2} at hyp during guest\u003c-\u003ehost transitions. With this\nhappening, there\u0027s no need to trap host SVE usage, and the nVHE/nVHE\n__deactivate_cptr_traps() logic can be simplified to enable host access\nto all present FPSIMD/SVE/SME features.\n\nIn protected nVHE/hVHE modes, the host\u0027s state is always saved/restored\nby hyp, and the guest\u0027s state is saved prior to exit to the host, so\nfrom the host\u0027s PoV the guest never has live FPSIMD/SVE/SME state, and\nthe host\u0027s ZCR_EL1 is never clobbered by hyp.\n\nFixes: 8c8010d69c132273 (\"KVM: arm64: Save/restore SVE state for nVHE\")\nFixes: 2e3cf82063a00ea0 (\"KVM: arm64: nv: Ensure correct VL is loaded before saving SVE state\")\nSigned-off-by: Mark Rutland \u003cmark.rutland@arm.com\u003e\nReviewed-by: Mark Brown \u003cbroonie@kernel.org\u003e\nTested-by: Mark Brown \u003cbroonie@kernel.org\u003e\nCc: Catalin Marinas \u003ccatalin.marinas@arm.com\u003e\nCc: Fuad Tabba \u003ctabba@google.com\u003e\nCc: Marc Zyngier \u003cmaz@kernel.org\u003e\nCc: Oliver Upton \u003coliver.upton@linux.dev\u003e\nCc: Will Deacon \u003cwill@kernel.org\u003e\nReviewed-by: Oliver Upton \u003coliver.upton@linux.dev\u003e\nLink: https://lore.kernel.org/r/20250210195226.1215254-9-mark.rutland@arm.com\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "f9dd00de1e53a47763dfad601635d18542c3836d", "tree": "ddda9e7b9b4a6855d7d3a747f83e7b188109ac66", "parents": [ "9b66195063c5a145843547b1d692bd189be85287" ], "author": { "name": "Mark Rutland", "email": "mark.rutland@arm.com", "time": "Mon Feb 10 19:52:25 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 17:55:03 2025 +0000" }, "message": "KVM: arm64: Mark some header functions as inline\n\nThe shared hyp switch header has a number of static functions which\nmight not be used by all files that include the header, and when unused\nthey will provoke compiler warnings, e.g.\n\n| In file included from arch/arm64/kvm/hyp/nvhe/hyp-main.c:8:\n| ./arch/arm64/kvm/hyp/include/hyp/switch.h:703:13: warning: \u0027kvm_hyp_handle_dabt_low\u0027 defined but not used [-Wunused-function]\n| 703 | static bool kvm_hyp_handle_dabt_low(struct kvm_vcpu *vcpu, u64 *exit_code)\n| | ^~~~~~~~~~~~~~~~~~~~~~~\n| ./arch/arm64/kvm/hyp/include/hyp/switch.h:682:13: warning: \u0027kvm_hyp_handle_cp15_32\u0027 defined but not used [-Wunused-function]\n| 682 | static bool kvm_hyp_handle_cp15_32(struct kvm_vcpu *vcpu, u64 *exit_code)\n| | ^~~~~~~~~~~~~~~~~~~~~~\n| ./arch/arm64/kvm/hyp/include/hyp/switch.h:662:13: warning: \u0027kvm_hyp_handle_sysreg\u0027 defined but not used [-Wunused-function]\n| 662 | static bool kvm_hyp_handle_sysreg(struct kvm_vcpu *vcpu, u64 *exit_code)\n| | ^~~~~~~~~~~~~~~~~~~~~\n| ./arch/arm64/kvm/hyp/include/hyp/switch.h:458:13: warning: \u0027kvm_hyp_handle_fpsimd\u0027 defined but not used [-Wunused-function]\n| 458 | static bool kvm_hyp_handle_fpsimd(struct kvm_vcpu *vcpu, u64 *exit_code)\n| | ^~~~~~~~~~~~~~~~~~~~~\n| ./arch/arm64/kvm/hyp/include/hyp/switch.h:329:13: warning: \u0027kvm_hyp_handle_mops\u0027 defined but not used [-Wunused-function]\n| 329 | static bool kvm_hyp_handle_mops(struct kvm_vcpu *vcpu, u64 *exit_code)\n| | ^~~~~~~~~~~~~~~~~~~\n\nMark these functions as \u0027inline\u0027 to suppress this warning. This\nshouldn\u0027t result in any functional change.\n\nAt the same time, avoid the use of __alias() in the header and alias\nkvm_hyp_handle_iabt_low() and kvm_hyp_handle_watchpt_low() to\nkvm_hyp_handle_memory_fault() using CPP, matching the style in the rest\nof the kernel. For consistency, kvm_hyp_handle_memory_fault() is also\nmarked as \u0027inline\u0027.\n\nSigned-off-by: Mark Rutland \u003cmark.rutland@arm.com\u003e\nReviewed-by: Mark Brown \u003cbroonie@kernel.org\u003e\nTested-by: Mark Brown \u003cbroonie@kernel.org\u003e\nAcked-by: Will Deacon \u003cwill@kernel.org\u003e\nCc: Catalin Marinas \u003ccatalin.marinas@arm.com\u003e\nCc: Fuad Tabba \u003ctabba@google.com\u003e\nCc: Marc Zyngier \u003cmaz@kernel.org\u003e\nCc: Oliver Upton \u003coliver.upton@linux.dev\u003e\nReviewed-by: Oliver Upton \u003coliver.upton@linux.dev\u003e\nLink: https://lore.kernel.org/r/20250210195226.1215254-8-mark.rutland@arm.com\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "9b66195063c5a145843547b1d692bd189be85287", "tree": "64589dd5759d4ce7848810bd76d21abea66a9a83", "parents": [ "ee14db31a9c84e65f5adfd45598760d851f1d817" ], "author": { "name": "Mark Rutland", "email": "mark.rutland@arm.com", "time": "Mon Feb 10 19:52:24 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 17:55:00 2025 +0000" }, "message": "KVM: arm64: Refactor exit handlers\n\nThe hyp exit handling logic is largely shared between VHE and nVHE/hVHE,\nwith common logic in arch/arm64/kvm/hyp/include/hyp/switch.h. The code\nin the header depends on function definitions provided by\narch/arm64/kvm/hyp/vhe/switch.c and arch/arm64/kvm/hyp/nvhe/switch.c\nwhen they include the header.\n\nThis is an unusual header dependency, and prevents the use of\narch/arm64/kvm/hyp/include/hyp/switch.h in other files as this would\nresult in compiler warnings regarding missing definitions, e.g.\n\n| In file included from arch/arm64/kvm/hyp/nvhe/hyp-main.c:8:\n| ./arch/arm64/kvm/hyp/include/hyp/switch.h:733:31: warning: \u0027kvm_get_exit_handler_array\u0027 used but never defined\n| 733 | static const exit_handler_fn *kvm_get_exit_handler_array(struct kvm_vcpu *vcpu);\n| | ^~~~~~~~~~~~~~~~~~~~~~~~~~\n| ./arch/arm64/kvm/hyp/include/hyp/switch.h:735:13: warning: \u0027early_exit_filter\u0027 used but never defined\n| 735 | static void early_exit_filter(struct kvm_vcpu *vcpu, u64 *exit_code);\n| | ^~~~~~~~~~~~~~~~~\n\nRefactor the logic such that the header doesn\u0027t depend on anything from\nthe C files. There should be no functional change as a result of this\npatch.\n\nSigned-off-by: Mark Rutland \u003cmark.rutland@arm.com\u003e\nReviewed-by: Mark Brown \u003cbroonie@kernel.org\u003e\nTested-by: Mark Brown \u003cbroonie@kernel.org\u003e\nAcked-by: Will Deacon \u003cwill@kernel.org\u003e\nCc: Catalin Marinas \u003ccatalin.marinas@arm.com\u003e\nCc: Fuad Tabba \u003ctabba@google.com\u003e\nCc: Marc Zyngier \u003cmaz@kernel.org\u003e\nCc: Oliver Upton \u003coliver.upton@linux.dev\u003e\nReviewed-by: Oliver Upton \u003coliver.upton@linux.dev\u003e\nLink: https://lore.kernel.org/r/20250210195226.1215254-7-mark.rutland@arm.com\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "ee14db31a9c84e65f5adfd45598760d851f1d817", "tree": "607be1f156b3273a95c7708f9d037747f17eed7f", "parents": [ "407a99c4654e8ea65393f412c421a55cac539f5b" ], "author": { "name": "Mark Rutland", "email": "mark.rutland@arm.com", "time": "Mon Feb 10 19:52:23 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 17:54:57 2025 +0000" }, "message": "KVM: arm64: Refactor CPTR trap deactivation\n\nFor historical reasons, the VHE and nVHE/hVHE implementations of\n__activate_cptr_traps() pair with a common implementation of\n__kvm_reset_cptr_el2(), which ideally would be named\n__deactivate_cptr_traps().\n\nRename __kvm_reset_cptr_el2() to __deactivate_cptr_traps(), and split it\ninto separate VHE and nVHE/hVHE variants so that each can be paired with\nits corresponding implementation of __activate_cptr_traps().\n\nAt the same time, fold kvm_write_cptr_el2() into its callers. This\nmakes it clear in-context whether a write is made to the CPACR_EL1\nencoding or the CPTR_EL2 encoding, and removes the possibility of\nconfusion as to whether kvm_write_cptr_el2() reformats the sysreg fields\nas cpacr_clear_set() does.\n\nIn the nVHE/hVHE implementation of __activate_cptr_traps(), placing the\nsysreg writes within the if-else blocks requires that the call to\n__activate_traps_fpsimd32() is moved earlier, but as this was always\ncalled before writing to CPTR_EL2/CPACR_EL1, this should not result in a\nfunctional change.\n\nSigned-off-by: Mark Rutland \u003cmark.rutland@arm.com\u003e\nReviewed-by: Mark Brown \u003cbroonie@kernel.org\u003e\nTested-by: Mark Brown \u003cbroonie@kernel.org\u003e\nAcked-by: Will Deacon \u003cwill@kernel.org\u003e\nCc: Catalin Marinas \u003ccatalin.marinas@arm.com\u003e\nCc: Fuad Tabba \u003ctabba@google.com\u003e\nCc: Marc Zyngier \u003cmaz@kernel.org\u003e\nCc: Oliver Upton \u003coliver.upton@linux.dev\u003e\nReviewed-by: Oliver Upton \u003coliver.upton@linux.dev\u003e\nLink: https://lore.kernel.org/r/20250210195226.1215254-6-mark.rutland@arm.com\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "407a99c4654e8ea65393f412c421a55cac539f5b", "tree": "ba83031a6a0e5b31fa59ed5e135b6215bddd73b2", "parents": [ "459f059be702056d91537b99a129994aa6ccdd35" ], "author": { "name": "Mark Rutland", "email": "mark.rutland@arm.com", "time": "Mon Feb 10 19:52:22 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 17:54:54 2025 +0000" }, "message": "KVM: arm64: Remove VHE host restore of CPACR_EL1.SMEN\n\nWhen KVM is in VHE mode, the host kernel tries to save and restore the\nconfiguration of CPACR_EL1.SMEN (i.e. CPTR_EL2.SMEN when HCR_EL2.E2H\u003d1)\nacross kvm_arch_vcpu_load_fp() and kvm_arch_vcpu_put_fp(), since the\nconfiguration may be clobbered by hyp when running a vCPU. This logic\nhas historically been broken, and is currently redundant.\n\nThis logic was originally introduced in commit:\n\n 861262ab86270206 (\"KVM: arm64: Handle SME host state when running guests\")\n\nAt the time, the VHE hyp code would reset CPTR_EL2.SMEN to 0b00 when\nreturning to the host, trapping host access to SME state. Unfortunately,\nthis was unsafe as the host could take a softirq before calling\nkvm_arch_vcpu_put_fp(), and if a softirq handler were to use kernel mode\nNEON the resulting attempt to save the live FPSIMD/SVE/SME state would\nresult in a fatal trap.\n\nThat issue was limited to VHE mode. For nVHE/hVHE modes, KVM always\nsaved/restored the host kernel\u0027s CPACR_EL1 value, and configured\nCPTR_EL2.TSM to 0b0, ensuring that host usage of SME would not be\ntrapped.\n\nThe issue above was incidentally fixed by commit:\n\n 375110ab51dec5dc (\"KVM: arm64: Fix resetting SME trap values on reset for (h)VHE\")\n\nThat commit changed the VHE hyp code to configure CPTR_EL2.SMEN to 0b01\nwhen returning to the host, permitting host kernel usage of SME,\navoiding the issue described above. At the time, this was not identified\nas a fix for commit 861262ab86270206.\n\nNow that the host eagerly saves and unbinds its own FPSIMD/SVE/SME\nstate, there\u0027s no need to save/restore the state of the EL0 SME trap.\nThe kernel can safely save/restore state without trapping, as described\nabove, and will restore userspace state (including trap controls) before\nreturning to userspace.\n\nRemove the redundant logic.\n\nSigned-off-by: Mark Rutland \u003cmark.rutland@arm.com\u003e\nReviewed-by: Mark Brown \u003cbroonie@kernel.org\u003e\nTested-by: Mark Brown \u003cbroonie@kernel.org\u003e\nAcked-by: Will Deacon \u003cwill@kernel.org\u003e\nCc: Catalin Marinas \u003ccatalin.marinas@arm.com\u003e\nCc: Fuad Tabba \u003ctabba@google.com\u003e\nCc: Marc Zyngier \u003cmaz@kernel.org\u003e\nCc: Oliver Upton \u003coliver.upton@linux.dev\u003e\nReviewed-by: Oliver Upton \u003coliver.upton@linux.dev\u003e\nLink: https://lore.kernel.org/r/20250210195226.1215254-5-mark.rutland@arm.com\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "459f059be702056d91537b99a129994aa6ccdd35", "tree": "5ce77b65667b646f6f56666f5164ca0654b58e7f", "parents": [ "8eca7f6d5100b6997df4f532090bc3f7e0203bef" ], "author": { "name": "Mark Rutland", "email": "mark.rutland@arm.com", "time": "Mon Feb 10 19:52:21 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 17:54:51 2025 +0000" }, "message": "KVM: arm64: Remove VHE host restore of CPACR_EL1.ZEN\n\nWhen KVM is in VHE mode, the host kernel tries to save and restore the\nconfiguration of CPACR_EL1.ZEN (i.e. CPTR_EL2.ZEN when HCR_EL2.E2H\u003d1)\nacross kvm_arch_vcpu_load_fp() and kvm_arch_vcpu_put_fp(), since the\nconfiguration may be clobbered by hyp when running a vCPU. This logic is\ncurrently redundant.\n\nThe VHE hyp code unconditionally configures CPTR_EL2.ZEN to 0b01 when\nreturning to the host, permitting host kernel usage of SVE.\n\nNow that the host eagerly saves and unbinds its own FPSIMD/SVE/SME\nstate, there\u0027s no need to save/restore the state of the EL0 SVE trap.\nThe kernel can safely save/restore state without trapping, as described\nabove, and will restore userspace state (including trap controls) before\nreturning to userspace.\n\nRemove the redundant logic.\n\nSigned-off-by: Mark Rutland \u003cmark.rutland@arm.com\u003e\nReviewed-by: Mark Brown \u003cbroonie@kernel.org\u003e\nTested-by: Mark Brown \u003cbroonie@kernel.org\u003e\nAcked-by: Will Deacon \u003cwill@kernel.org\u003e\nCc: Catalin Marinas \u003ccatalin.marinas@arm.com\u003e\nCc: Fuad Tabba \u003ctabba@google.com\u003e\nCc: Marc Zyngier \u003cmaz@kernel.org\u003e\nCc: Oliver Upton \u003coliver.upton@linux.dev\u003e\nReviewed-by: Oliver Upton \u003coliver.upton@linux.dev\u003e\nLink: https://lore.kernel.org/r/20250210195226.1215254-4-mark.rutland@arm.com\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" }, { "commit": "8eca7f6d5100b6997df4f532090bc3f7e0203bef", "tree": "ab0aecc9f8c9f386bce3084b3027bb56ba8848fd", "parents": [ "fbc7e61195e23f744814e78524b73b59faa54ab4" ], "author": { "name": "Mark Rutland", "email": "mark.rutland@arm.com", "time": "Mon Feb 10 19:52:20 2025 +0000" }, "committer": { "name": "Marc Zyngier", "email": "maz@kernel.org", "time": "Thu Feb 13 17:54:44 2025 +0000" }, "message": "KVM: arm64: Remove host FPSIMD saving for non-protected KVM\n\nNow that the host eagerly saves its own FPSIMD/SVE/SME state,\nnon-protected KVM never needs to save the host FPSIMD/SVE/SME state,\nand the code to do this is never used. Protected KVM still needs to\nsave/restore the host FPSIMD/SVE state to avoid leaking guest state to\nthe host (and to avoid revealing to the host whether the guest used\nFPSIMD/SVE/SME), and that code needs to be retained.\n\nRemove the unused code and data structures.\n\nTo avoid the need for a stub copy of kvm_hyp_save_fpsimd_host() in the\nVHE hyp code, the nVHE/hVHE version is moved into the shared switch\nheader, where it is only invoked when KVM is in protected mode.\n\nSigned-off-by: Mark Rutland \u003cmark.rutland@arm.com\u003e\nReviewed-by: Mark Brown \u003cbroonie@kernel.org\u003e\nTested-by: Mark Brown \u003cbroonie@kernel.org\u003e\nAcked-by: Will Deacon \u003cwill@kernel.org\u003e\nCc: Catalin Marinas \u003ccatalin.marinas@arm.com\u003e\nCc: Fuad Tabba \u003ctabba@google.com\u003e\nCc: Marc Zyngier \u003cmaz@kernel.org\u003e\nCc: Oliver Upton \u003coliver.upton@linux.dev\u003e\nReviewed-by: Oliver Upton \u003coliver.upton@linux.dev\u003e\nLink: https://lore.kernel.org/r/20250210195226.1215254-3-mark.rutland@arm.com\nSigned-off-by: Marc Zyngier \u003cmaz@kernel.org\u003e\n" } ], "next": "fbc7e61195e23f744814e78524b73b59faa54ab4" }