)]}' { "log": [ { "commit": "26260251022fbc2f248a3d747a9b2b961b18d2d8", "tree": "1b3b1225ba9ad4d1489f641e8ea0ea701a9cf7c3", "parents": [ "2f0f6b0773be0a1ec475097ae54848eea42adc7d", "e77f165d26f72c280654cff248f5106cb9a53951" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 13:11:56 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 13:11:56 2026 -0700" }, "message": "Merge tag \u0027livepatching-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/livepatching/livepatching\n\nPull livepatching updates from Petr Mladek:\n\n - Move consistency checks to catch missing func-\u003eold_name before the\n first access\n\n - Allow to run livepatching selftests from top-level directory\n\n* tag \u0027livepatching-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/livepatching/livepatching:\n kbuild: unset sub_make_done before calling kselftest build system\n livepatch: Fix NULL pointer dereference in klp_find_func()\n" }, { "commit": "2f0f6b0773be0a1ec475097ae54848eea42adc7d", "tree": "63dc254caea8a8a76d0889305317cab33119289d", "parents": [ "27a59c0251e5d1f41a3a6fbae7c4dc478c34d919", "3dfaae04243cde460d82dfc2a7dd0bb6664d20ae" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:47:25 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:47:25 2026 -0700" }, "message": "Merge tag \u0027modules-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/modules/linux\n\nPull module updates from Petr Pavlu:\n\n - Remove unnecessary module::args. Nowadays, no parameter-handling code\n points into the module::args buffer. The last user of module::args in\n xtensa/simdisk is updated and the data is then removed\n\n - Add Rust support for boolean parameters. This will initially be used\n by the Rust null block driver\n\n - Fix clearing the current charp parameter value when setting a new one\n fails due to an allocation failure\n\n - Improve the debugging code for kmod (request_module()) duplicates.\n Fix a potential use-after-free when waiting on a duplicate request\n and make several general improvements to the code\n\n - Fix the symbol size returned when looking up a data symbol through\n kallsyms\n\n - Smaller fixes and cleanups\n\n* tag \u0027modules-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/modules/linux:\n params: fix charp corruption on allocation failure\n module: validate string table section types\n module/dups: Clean up includes\n module/dups: Use strcmp() to compare module names\n module/dups: Use scope-based cleanup helpers\n module/dups: Avoid unnecessary kmod_dup_req allocations\n module/dups: Fix use-after-free in kmod_dup_req lifetime handling\n module/dups: Inform duplicate requests about the result directly\n rust: module_param: support bool parameters\n rust: module_param: return value by copy from `value`\n module: Remove unnecessary module::args\n xtensa/simdisk: Avoid referring to module::args\n module: Remove unused DISCARD_EH_FRAME definition from module.lds.S\n module: procfs: use matching type for accumulator in module_total_size()\n module: use strscpy() to copy module names in stats and dup tracking\n params: fix path of /sys/module/XYZ/parameters/ in comment\n module/kallsyms: fix nextval for data symbol lookup\n" }, { "commit": "27a59c0251e5d1f41a3a6fbae7c4dc478c34d919", "tree": "ed4658b35c2d8311efd174f443ef15bfdcc7359f", "parents": [ "af33c5a2a9fd52f07ffb428255b7f060da1de49d", "3b7711be02612845e44b5e9ef7c0b9f203cf29b6" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:39:49 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:39:49 2026 -0700" }, "message": "Merge tag \u0027mips_7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux\n\nPull MIPS updates from Thomas Bogendoerfer:\n\n - switch gpio code to use swnodes\n\n - rework of TXX9 gpio code\n\n - enable multi-vpe for econet\n\n - cleanups and fixes\n\n* tag \u0027mips_7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:\n mips: dts: econet: Describe dual-VPE 34Kc processor\n mips: econet: add multi-vpe capability to EN751221\n MIPS: ptrace: Fix syscall skipping via PTRACE_SYSCALL\n mips: remove dead select\n MIPS: BCM47XX: Convert buttons to software nodes\n ssb: gpio: Add and register software node for GPIO controller\n bcma: gpio: Add and register software node for GPIO controller\n MIPS: ip22-gio: Drop #include of \u003clinux/mod_devicetable.h\u003e\n MIPS: TXX9: Clean up txx9_iocled_init()\n MIPS: TXX9: Convert gpio_txx9 to dynamic GPIO base allocation\n MIPS: TXX9: Drop GPIOLIB_LEGACY select\n MIPS: TXX9: Use GPIO lookup table for iocled LEDs\n MIPS: TXX9: Reduce TXX9_IOCLED_MAXLEDS to 3\n MIPS: TXX9: rbtx4927: Use GPIO lookup table for TXx9 LEDs\n MIPS: TXX9: rbtx4927: Use GPIO lookup table for SIO DTR\n MIPS: TXX9: Remove txx9_7segled_*() forward declarations\n MIPS: TXX9: Remove tx4938_spi_init() and txx9_spi_init()\n MIPS: kernel: proc: Use two seq_putc() calls in show_cpuinfo()\n" }, { "commit": "af33c5a2a9fd52f07ffb428255b7f060da1de49d", "tree": "ac99937cc7d0e451aeb2b9182dd7034aaca9470e", "parents": [ "7199989f3f3194d653b024ce8e79cea6b15e38b9", "f81cb44f9a4b88d73ee5dec4a1ccdb0232fd2e3f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:32:01 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:32:01 2026 -0700" }, "message": "Merge tag \u0027ecryptfs-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tyhicks/ecryptfs\n\nPull eCryptfs updates from Tyler Hicks:\n\n - Hardening and fixes for maliciously crafted eCryptfs metadata in the\n lower encrypted file\n\n - Hardening and fixes for maliciously crafted userspace \u003c-\u003e kernel\n miscdev communications\n\n - Locking fixes for userspace \u003c-\u003e kernel miscdev communications\n\n - Fix to display encrypted filename related mount options\n\n - Clean up address_space_operations and reduce build dependencies by\n moving to filemap_dirty_folio()\n\n - Get rid of an unnecessary memory allocation in the inode update path\n\n - Kernel-doc formatting corrections\n\n* tag \u0027ecryptfs-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tyhicks/ecryptfs:\n ecryptfs: ecryptfs_kernel.h: clean up kernel-doc comments\n ecryptfs: use filemap_dirty_folio for address space operations\n ecryptfs: avoid heap allocation for inode size write\n ecryptfs: show filename encryption options\n eCryptfs: bound the packet-length peek to the user buffer\n ecryptfs: reject too-small tag 70 packets\n ecryptfs: fix tag 11 packet exact-fit size check\n ecryptfs: pass packet set buffer size to parser\n ecryptfs: hold msg ctx list lock when cleaning daemon queue\n ecryptfs: release message context on send failure\n ecryptfs: reject oversized encrypted_key_size in parse_tag_3_packet\n" }, { "commit": "7199989f3f3194d653b024ce8e79cea6b15e38b9", "tree": "00801d41a37df6bdd737ede8798f68006be3ca88", "parents": [ "614b9fb585f143d65162f17f1a4b4ec4cbdf5794", "172b6a6d8463562b0cbebfd66f770b078f81966b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:28:35 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:28:35 2026 -0700" }, "message": "Merge tag \u0027landlock-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mic/linux\n\nPull Landlock update from Mickaël Salaün:\n \"This improves observability with Landlock tracepoints support, which\n required some refactoring for dedicated domain types and common\n helpers shared with audit code.\n\n A LANDLOCK_RESTRICT_SELF_NO_NEW_PRIVS flag is also added to improve\n process-wide domain enforcement consistency.\n\n Whiteout files are now correctly handled and tested, and a few other\n fixes\"\n\n* tag \u0027landlock-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mic/linux: (34 commits)\n landlock: Document tracepoints\n selftests/landlock: Add landlock_enforce_domain trace tests\n selftests/landlock: Add scope and ptrace tracepoint tests\n selftests/landlock: Add network tracepoint tests\n selftests/landlock: Add filesystem tracepoint tests\n selftests/landlock: Add trace event test infrastructure and tests\n landlock: Add tracepoints for ptrace and scope denials\n landlock: Add landlock_deny_access_fs and landlock_deny_access_net\n landlock: Add tracepoints for rule checking\n landlock: Add landlock_enforce_domain tracepoint\n landlock: Add create_domain and free_domain tracepoints\n landlock: Add landlock_add_rule_fs and landlock_add_rule_net tracepoints\n landlock: Add create_ruleset and free_ruleset tracepoints\n landlock: Consolidate access-right and scope names in a shared header\n landlock: Decouple the per-denial logging decision from CONFIG_AUDIT\n landlock: Split denial logging from audit into common framework\n landlock: Split struct landlock_domain from struct landlock_ruleset\n landlock: Move domain query functions to domain.c\n landlock: Prepare ruleset and domain type split\n samples/landlock: Add LANDLOCK_RESTRICT_SELF_NO_NEW_PRIVS to sampler\n ...\n" }, { "commit": "614b9fb585f143d65162f17f1a4b4ec4cbdf5794", "tree": "44ceff9d1560674730ebf2bb2d9755e2159d9e13", "parents": [ "2be02a7c996aa733bb36e29e07715621b0de9736", "934753d26f1ed7fadeedf1b9e7df6fe75de303e8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:16:31 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:16:31 2026 -0700" }, "message": "Merge tag \u0027cifs-maintainer-switch-7.3-rc1\u0027 of https://git.manguebit.org/linux\n\nPull cifs maintainer update from Paulo Alcantara.\n\n* tag \u0027cifs-maintainer-switch-7.3-rc1\u0027 of https://git.manguebit.org/linux:\n MAINTAINERS: Replace Steve French as CIFS maintainer\n" }, { "commit": "2be02a7c996aa733bb36e29e07715621b0de9736", "tree": "8bb6dd627780d735d1d8a3fabcfaa83c525e3d59", "parents": [ "283955dfacfe9e4716b13aa7cd360544717f00cb", "2da28b059e0ddcd2e1956eeae383246207965573" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:05:10 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:05:10 2026 -0700" }, "message": "Merge tag \u0027for-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-power-supply\n\nPull power supply and reset updates from Sebastian Reichel:\n \"Power-supply core:\n - Add PbAc, NiZn, RAM, and ZnAr battery chemistry types\n - Create LED triggers based on properties instead of device type\n - Provide power_supply_get_system_batteries() for usage with USB-C\n - Add registration init callback for race-free device setup\n\n Power-supply drivers:\n - new TI BQ25630 charger driver\n - new SG Micro sgm41542 charger driver\n - bq257xx: Add support for BQ25792\n - max8903: add DC and USB input current-limit controls\n - max17042_battery: Initialize MAX17055 from battery info\n - sbs-battery: map newly introduced battery chemistries\n - drop extra error messages for IRQ request failures\n - lot\u0027s of driver removal race condition fixes\n - misc small cleanups and fixes\n\n Reset drivers:\n - add MCF5441x RCM power-on reason driver\n - misc small cleanups and fixes\"\n\n* tag \u0027for-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-power-supply: (115 commits)\n power: supply: bq27xxx: bq27z561: fix invalid AverageEnergy address\n power: supply: bq27xxx: bq28z610: fix invalid AverageEnergy address\n power: supply: bq27xxx: bq27520g4: fix REG_TTES address\n power: supply: max17040: synchronize work cancellation on suspend\n power: supply: lp8727: fix use-after-free in lp8727_release_irq()\n power: supply: bq256xx: drain usb_work before freeing the charger\n power: supply: qcom_battmgr: fix battery chemistry strncmp length\n power: supply: bd99954: Drop bad register fields\n power: supply: bd71828: Do not hide errors\n power: supply: bd71828: Drop duplicate power-supply property\n power: supply: bd71828: Fix current direction\n power: supply: bd71815: Fix temperature reading\n power: supply: add stubs for notifier registration helpers\n power: supply: ucs1002: fix use-after-free on remove\n power: supply: lp8788-charger: fix use-after-free on remove\n power: supply: ab8500_fg: fix use-after-free on remove\n power: supply: bq24257: fix use-after-free on remove\n power: supply: qcom_battmgr: fix use-after-free\n power: supply: max17040: drop incorrect I2C functionality check\n power: supply: charger-manager: register regulators before exposing sysfs\n ...\n" }, { "commit": "283955dfacfe9e4716b13aa7cd360544717f00cb", "tree": "59efe0e39253b006ffef61be7fbc491f17816cad", "parents": [ "f96718fc1f6142038a3f836b83b12e5062576a2f", "e81250ec6b69248b00d38c523dc6a13efaf38aab" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:01:23 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 12:01:23 2026 -0700" }, "message": "Merge tag \u0027hsi-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-hsi\n\nPull HSI updates from Sebastian Reichel:\n\n - omap_ssi_core: fix missing DMA mask\n\n - misc small cleanups\n\n* tag \u0027hsi-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-hsi:\n hsi: omap_ssi_core: fix missing DMA mask setup for SSI controller device\n HSI: omap_ssi: Remove redundant dev_err()\n HSI: nokia-modem: Remove redundant dev_err()\n hsi: omap_ssi: remove debugfs on port creation failure\n" }, { "commit": "f96718fc1f6142038a3f836b83b12e5062576a2f", "tree": "1d2b1d61f1b4ac294bdcd2f9f6d6786544852d45", "parents": [ "21bd0802cd3f58b656065f1be236694c40588c3a", "35e689fb79dcbaa8d69cc4a58a7ad4d0d7346e4a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 11:58:08 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 11:58:08 2026 -0700" }, "message": "Merge tag \u0027mmc-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc\n\nPull MMC updates from Ulf Hansson:\n \"MMC core:\n - Reject invalid perdev_minors for the block device before division\n - Document DT fixed-layout NVMEM provider support for eMMC cards\n\n MMC host:\n - Convert a couple of plain text DT bindings to the yaml format\n - bcm2835: DMA mapping improvements\n - cqhci: Fix sparse warnings for endian conversions\n - dw_mmc: Stop and complete DMA also in busy state\n - dw_mmc-rockchip: Add support for the RV1106 variant\n - litex_mmc: Add dynamic bus width support\n - moxart: Propagate error for DMA completion timeout\n - pxamci: Remove PXA remnants for the PXA93x support\n - rtsx_usb_sdmmc: Avoid USB I/O in runtime autosuspend\n - rtsx_usb_sdmmc: Suppress false CD after init timeout\n - sdhci_am654: Add Judith Mendez as maintainer\n - sdhci-esdhc-mcf: Do not use readl()/writel() on ColdFire\n - sdhci-msm: Extend the DT bindings for ICE\n - sdhci-tegra: Add support for Tegra264/Tegra238 variants\n - sunxi: Add support for the Allwinner A733 variant\n - via-sdmmc: Clean up card detect work at remove and in probe error path\"\n\n* tag \u0027mmc-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc: (35 commits)\n ARM: PXA: Fix build error for PXA93x\n dt-bindings: mmc: sunxi: add compatible string for Allwinner A733 MMC0/1/2\n dt-bindings: mmc: Document fixed-layout NVMEM provider support\n mmc: sdhci-tegra: Add Tegra264 SoC data\n dt-bindings: mmc: tegra: Document Tegra264 SDHCI\n mmc: sdhci-tegra: Add Tegra238 SoC data\n dt-bindings: mmc: tegra: Document Tegra238 SDHCI\n mmc: omap_hsmmc: use platform_get_irq_optional for wake IRQ\n mmc: via-sdmmc: cancel card-detect work on remove\n mmc: via-sdmmc: stop card-detect handling on probe failure\n mmc: moxart: use platform helpers for resource and IRQ\n mmc: host: Remove redundant dev_err()/dev_err_probe()\n mmc: bcm2835: DMA mapping improvements\n mmc: dw_mmc: move declaration of dw_mci_pmops\n dt-bindings: mmc: rockchip-dw-mshc: Add RV1106 compatible\n mmc: rtsx_usb_sdmmc: suppress false CD after init timeout\n misc: rtsx_usb: avoid USB I/O in runtime autosuspend\n mmc: sdhci-of-dwcmshc: Log eMMC reset calls\n mmc: block: reject invalid perdev_minors before division\n mmc: sdhci-of-ma35d1: add missing MODULE_DEVICE_TABLE()\n ...\n" }, { "commit": "21bd0802cd3f58b656065f1be236694c40588c3a", "tree": "08f3479623310d779023b5a8cd4290e53d1e37d7", "parents": [ "35748ddd3b2c91555bd86b8cf88edc90e7317e57", "8049741ac93acd3a590dac070e12571fddf0e294" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 11:48:54 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 11:48:54 2026 -0700" }, "message": "Merge tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma\n\nPull RDMA updates from Jason Gunthorpe:\n \"About the normal size, still a lot of AI bug fixes and so on, but some\n interesting new functionality too:\n\n - Assorted locking, bounds-checking, cleanup, and error-path fixes\n across UCMA/CMA, bng_re, bnxt_re, cxgb4, EFA, ERDMA, HFI1, HNS,\n ionic, iRDMA, mlx4/mlx5, RXE, SIW, SRP/SRPT, and iSER target.\n\n - netlink report for max # of supported resources\n\n - get_zeroed_page()/etc removal\n\n - Robust udata for ionic\n\n - Allow unique RDMA device names per network namespace\n\n - Completion counters and v2 admit queue support for EFA\n\n - UC QP support for MANA\n\n - Completion timestamps for ionic\n\n - Harden uverbs data validation and resource lifetime handling,\n fixing several core use-after-free conditions.\n\n - bnxt_re toggle-page ownership and lifetime bug fixes\n\n - dmabuf SRQ support for mlx5\"\n\n* tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (160 commits)\n RDMA/ucma: Allow path records to exactly fit the output buffer\n RDMA/uverbs: Guard legacy bundles without method_elm\n RDMA/efa: Add support for 128B admin v2 SQ entry\n RDMA/efa: Generalize the admin SQ\n RDMA/efa: Decouple admin command payload from admin header\n RDMA/rxe: Fix OOB in free_rd_atomic_resources()\n RDMA/cma: Fix WARNING in res_to_rt\n RDMA/cxgb4: Free debugfs on registration failure\n RDMA/cxgb4: Cancel reg_work before freeing device on remove\n RDMA/ucma: Lock the handler in ucma_set_ib_path()\n RDMA/ucma: Lock the handler in ucma_write_cm_event()\n RDMA/erdma: restrict the driver to little-endian systems\n RDMA/ionic: Embed counter driver data in rdma_counter allocation\n RDMA/ionic: Cap eq_count to the eth driver\u0027s interrupt vector budget\n RDMA/siw: Fix use-after-free in siw_accept()\n IB/isert: post the full-feature receive buffers after session registration\n IB/isert: delay the final Login Response until the session is registered\n RDMA/srp: fix heap information leak on a truncated SRP_CRED_REQ\n RDMA/erdma: Hold QP references for AE and CM processing\n RDMA/erdma: Hold CQ references when processing EQ events\n ...\n" }, { "commit": "35748ddd3b2c91555bd86b8cf88edc90e7317e57", "tree": "df47244436f2a14f38f06bd21856523a5572bef0", "parents": [ "531ed942bb0df04f6747983fecdedce76a22d07f", "376a3960e5efe85ff765abfb5b5b7e4655ad6aed" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 11:23:20 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 11:23:20 2026 -0700" }, "message": "Merge tag \u0027scsi-misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi\n\nPull SCSI updates from James Bottomley:\n \"One new driver: leapraid (similar to mpi3mr but OK\u0027d by Broadcom). The\n usual suspects for driver updates (ufs, qla2xxx, smartpqi, zfcp, fnic,\n ibmvfc) plus a few small core updates: a fix for an uninitialized sg\n list pad bytes plus the removal of the dma mask check for max sectors.\n\n The big update in the sd driver is mostly code refactoring for obscure\n error leg handling\"\n\n* tag \u0027scsi-misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (195 commits)\n scsi: fnic: Fix built-in NVMe/FC build\n scsi: fnic: Fix invalid comparison for error\n scsi: core: Fill in DMA padding bytes in scsi_alloc_sgtables()\n scsi: zfcp: Enable CONTEXT_ANALYSIS\n scsi: zfcp: Add __must_hold() attribute to zfcp_qdio_sbal_get()\n scsi: fnic: Use GFP_ATOMIC for VLAN alloc under spinlock\n scsi: storvsc: Support manual scans for all Hyper-V targets\n scsi: sd: Fix sd_done() sense handling condition\n scsi: sd: Fix special_vec mempool leak when scsi_alloc_sgtables() fails\n scsi: sd: Fix error handling in sd_probe() after large pool creation failure\n scsi: leapraid: Add driver documentation\n scsi: leapraid: Add new SCSI driver\n scsi: ufs: Add support for the aggregated read query opcode\n scsi: ufs: Use unsigned types for the BSG query\n scsi: ibmvfc: Fix spelling mistake \"Deleteing\" -\u003e \"Deleting\"\n scsi: qla2xxx: Update version to 12.00.00.2607b2\n scsi: qla2xxx: Bound i2c-\u003elength in I2C bsg handlers\n scsi: qla2xxx: Zero SFP DMA buffer in FRU/I2C bsg handlers\n scsi: qla2xxx: Validate BSG request_len before reading vendor_cmd[]\n scsi: qla2xxx: Zero-init bsg stack buffers to avoid info leak\n ...\n" }, { "commit": "531ed942bb0df04f6747983fecdedce76a22d07f", "tree": "90b0e9425a115c48f32d2f65ba02aa1574d7c47d", "parents": [ "4e69c1856bfd9ffb7e9d335a25842fa211628929", "d463633d63e6aa10384683bf39971a4b00780c56" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 09:49:54 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 09:49:54 2026 -0700" }, "message": "Merge tag \u0027fbdev-for-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/deller/linux-fbdev\n\nPull fbdev updates from Helge Deller:\n \"The usual bunch of many small fixes in various fbdev drivers, but\n more interestingly the Vodoo3/4/5 cards will now be initialized even\n without PC BIOS support and vintage Atari computers gain more color\n depths and console acceleration on SuperVidel\u0027s SuperBlitter chips.\n\n New features:\n - Allow Vodoo3/4/5 card to be initialized without PC-BIOS (Daniel\n Palmer)\n - Add support for SuperVidel\u0027s SuperBlitter (Miro Kropacek)\n - Add further video bit depths (Miro Kropacek)\n - Detect default graphics card for console output on sticon/parisc\n (Helge Deller)\n\n Fixes:\n - kyro: Validate overlay viewport coordinates (Danila Chernetsov)\n - platinumfb: add error checking for ioremap calls (BingKun Yue)\n - ssd1307fb: damage callback fixes (Hui Su)\n - maxine: fix 64-bit build error and code cleanups (Randy Dunlap)\n - omapfb: panel-dsi-cm: initialize lock before registering display\n (Runyu Xiao)\n - core: Clamp total_size to smem_len in read/write functions\n (Mingyu Wang)\n - uvesafb, tdxfb: failure path cleanups (Myeonghun Pak)\n - udlfb: validate DisplayLink vendor descriptor items before usage\n (Pengpeng Hou)\n\n Cleanups:\n - Convert multiple drivers to managed PCI and ioremap API (Shixiong Ou)\n - Remove redundant dev_err() from multiple drivers (Pan Chuang)\n - omap2: do not copy isr table (Andreas Kemnade)\n - pvr2fb: correct user pointer annotation and sentinel initializer\n (Florian Fuchs)\n - mb862xxfb: silence possibly unused functions (Helge Deller)\n - au1100fb: drop unneeded semicolon (Julia Lawall)\n - clps711x-fb: remove unreachable code (Karl Mehltretter)\n - viafb: refactor strcpy call (Ajith P V)\n - sstfb: add missing MODULE_DEVICE_TABLE() (Pengpeng Hou)\n - mb862xx: replace dead Kconfig select with dependency (Julian Braha)\n\n Documentation fixes:\n - fonts: fixup font.h kernel-doc warnings (Randy Dunlap)\"\n\n* tag \u0027fbdev-for-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/deller/linux-fbdev: (43 commits)\n fbdev: atafb: Add support for SuperVidel\u0027s SuperBlitter\n fbdev: atafb: Add support for further video bit depths on atafb:external\n fbdev: atafb: Give atafb proper parent\n fbdev: omapfb: panel-dsi-cm: initialize lock before registering display\n fbdev: viafb: refactor strcpy and viafb_name\n fbdev: platinumfb: add error checking for ioremap calls\n fbdev: maxine: use MODULE_LICENSE() unconditionally\n fbdev: maxine: fix maxinefb_init() return value\n fbdev: maxine: fix 64-bit build error\n fbdev: maxine: elide an unused function\n fbdev: maxine: make functions static\n fbdev: atyfb: Convert to managed PCI and ioremap API\n fbdev: matrox: Convert to managed PCI and ioremap API\n fbdev: savage: Convert to managed PCI and ioremap API\n fbdev: nvidia: Convert to managed PCI and ioremap API\n fbdev: aty128fb: Convert to managed PCI and ioremap API\n fbdev: clps711x-fb: Remove unreachable unregister_framebuffer() call\n fbdev: mb862xxfb: Silence possibly unused functions\n sticon/parisc: Detect default STI graphics card for console output\n fbdev: ssd1307fb: defer I2C transfers from damage callbacks\n ...\n" }, { "commit": "4e69c1856bfd9ffb7e9d335a25842fa211628929", "tree": "bc47d84baf013634463f55ba26a554cbfcb7c908", "parents": [ "818bebeb63dd6bf5f4e07e145f6cdbace520a34c", "c44e278ce02efd0c4be79a8eda1ea6885c1ce5ec" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 08:41:00 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 21 08:41:00 2026 -0700" }, "message": "Merge tag \u0027drm-next-2026-08-20\u0027 of https://gitlab.freedesktop.org/drm/kernel\n\nPull drm updates from Dave Airlie:\n \"Highlights:\n\n - dmemcg eviction support is good for low VRAM things like Steam\n Machine\n\n - AMD adds gfx6-8 modifier support for older GPUs that enables a\n bunch of wayland stuff\n\n - i915/xe has some new hw support but also a lot of display\n refactoring\n\n Everything:\n\n perf:\n - export perf_allow_ APIs for xe\n\n udmabuf:\n - remove default size limit of 64MB\n\n rust:\n - i/o rework (signed tag from driver-core tree)\n - add registration guard and registration data\n - fix unbounded lifetimes in ioctl handler args\n - fix a drm_dev_register race\n - gem_shmem: add DmaResvGuard helper\n - gpuvm: require send/sync for driver data\n - implement send/sync for GpuVaAlloc and GpuVmBo\n - add SmContext lifetime\n - rename dma_handle to dma_address\n - change pci_sriov_get_totalvfs return to unsigned int\n\n core:\n - create drm_of_get_panel_orientation\n - send per-connector hotplug events\n - add thunderbolt UBHR tunneling support\n\n connector:\n - add color format property\n\n dmem:\n - introduce a peak file\n - accept one region per limit\n - add dmemcg support for eviction\n\n gpusvm:\n - reorg code to give drivers more flexibility\n\n atomic:\n - add create_state callback and helper\n - add documentation on atomic commit lifetime\n\n buddy:\n - add per-order free\n - add used block scoreboard\n - fix UAF\n - test buffer clearance on resume\n - add phys_addr-\u003eblock helper\n\n gem:\n - drop DRIVER_GEM_GPUVA flag\n\n ttm:\n - be more aggressive allocating below protection limit\n\n sched:\n - add test suite for concurrent job submissions\n\n hdmi:\n - hook the color format property in helpers\n\n mipi-dsi:\n - add MIPI_DSI_MODE_DSC_ALL_SLICES_IN_PKT\n\n bridge:\n - add atomic create callbacks\n - drop atomic reset\n - display-connector: don\u0027t autoenable HPD IRQ\n - trigger initial HPD for DP\n - ti-sn65dsi83: remove NO_HFP and NO_HBP mode flags\n - analogix_dp: switch to DP link training helpers\n\n dp:\n - add support for DSC max delta BPP\n\n edid:\n - parse panel type from DisplayID 2.x Display Parameters\n\n sysfb:\n - improve panel, stride, framebuffer size validation\n\n panel:\n - implement ref counting for struct drm_panel\n - himax-hx83121a: add backlight regulator support\n - novatek-nt36672a: Inline panel init sequences\n - visionox-vtdr6130: enable DSC\n - novatek-nt37801: Use mipi_dsi_*_multi() functions\n - samsung-s6d16d0: Fix prepare error handling\n - support Novatek NT36536 plus DT bindings\n - sofef00: fix backlight updates\n - osd101t2587: use mipi_dsi_*_multi interface\n - panel-edp: adjust timing for AUO displays\n - panel-lvds: support Opto Logic SCX1001511GGC49\n - panel-simple: support Kyocera tcg070wvlq\n - panel-edp: quirks\n - AUO B116XAT04.3, CMN N116BCP-EA2, CSW MNB601LS1-8\n - BOE NV116WH2-M30, BOE NT116WHM-N21, BOE NV116FH1-M31\n - BOE NV116FH1-M30, NV140FHM-N5B, TM156VDXP25\n - BOE NE160QDM-NY1, MB116AS01\n - new:\n - Samsung ATNA40HQ08-0, Anbernic TD4310\n - Chipone ICNA35XX, Ilitek ILI9488\n - Ilitek ILI7807S, Renesas R63419\n - MNE001BS6-2, MNF601BS4-1, Sharp LQ120P1JX51\n\n virtio:\n - add support for save/restore virtio_gpu_objects\n - abort vq wait on device removal\n\n amdgpu:\n - add color format DRM property\n - initial compute pipe reset support\n - add GFX 6-8 modifier support\n - initial DCN 6.0.0 support\n - dmemcg eviction support\n - improved boundary checking for bios parsing\n - RAS updates and rework\n - VCN secure submission fixes\n - 8K panel fix\n - Display KUNIT tests\n - parse panel type from DisplayID\n - Align IP discovery to pci device lifetime\n - SOC15 register macro cleanups\n - UVD memory placement fixes\n - GFX9 mode2 reset fixes\n - drop unnecessary BUG/BUG_ON\n - GFX8 soft reset rework\n - enable soft reset on GFX8\n - PSP/SMU 15.0.9 update\n - VI ASPM fix\n - userq fixes\n - amdgpu_vm_get_task_info_pasid lifetime fix\n - DC CACP support\n - change system_unbound_wq with system_dfl_wq\n - Loosen VFCT bios parsing to deal with pci\u003drealloc\n - SI/SMU7 AC/DC switch fix\n - VM fence handling fix\n - GEM close optimisation\n - Apple Studio Display fixes\n - DC FRL fixes\n\n amdkfd:\n - initial compute pipe reset support\n - allow applications to opt out of sigbus on fatal errors\n - improve CRIU boundary checks\n - MQD handling rework\n - move TBA/TMA from system to device memory\n - avoid topology-lock in kfd_mmap\n - SVM eviction fixes\n\n radeon:\n - fix unset CONFIG_ACPI build\n\n i915:\n - Novalake (NVL display version 35) timing generator enabling\n - NVL DC3CO enabling\n - enable UBHR link rates on thunderbolt tunnels\n - Reduce Xe3+ PM demand peak bandwidth\n - enable pipe DMC error interrupts for display 30+\n - add kunit tests for DP link config selection\n - refactor and document DP link recovery\n - i915/xe driver display probe/remove/suspend/resume/shutdown cleanup\n and unification\n - i915/xe display runtime PM unified\n - Break i915 and xe panic dependency on struct intel_framebuffer\n - Streamline Pre/Post-CSC LUT loops\n - drop TGL DC3DO support\n - CDCLK santization\n - fix HDMI scrambling enable\n - fix phys bo pread/pwrite with offset\n - add missing nospec on parallel submit slot\n - fix some NULL derefs\n\n xe:\n - drop force_execlist module param\n - gate observation streams with perf_allow_cpu\n - skip FORCE_WC and vm_bound check for external dma-bufs\n - dmemcg eviction support\n - remove unused NVL-S GuC\n - TLB invalidation improvements\n - NVL-S updated PCI-IDs and w/a\n - madvise: optimise invalidation path\n - fix infinite gt-reset loop in timeout recovery\n - update TTM device benefical_order\n - wait on external BO kernel fences in exec ioctl\n - add/use more KLV helpers\n - sriov: disable display in admin only PF mode\n - add RAS GPU health indicator\n - optimise TTM populate for DONTNEED BO\n - drop force_probe for NVL-s\n - add debugfs for pcode info\n\n amdxdna:\n - disable device buffer export\n\n nova:\n - build nova-core/nova-drm from drivers/gpu\n - export nova-core rust symbols (workaround)\n - GSP boot process consolidation\n - Boot GSP with vGPU enabled\n - TLV firmware image format support\n - Hopper/Blackwell fixes and cleanups\n - I/O projection adoption\n\n tyr:\n - firmware loading and MCU boot\n - add generic slot manager + MMU\n - GPU VM support ARM64 LPAE page tables\n - add kernel buffer object for internal allocations\n - add parser for Mali CSF\n - add MCU booting\n\n nouveau:\n - race fixes\n - check instmem iomapping at first use\n - add dmemcg support\n - expose NVDEC channels\n - add scanline position/head state support for GSP\n\n qxl:\n - convert simple encoder to regular\n\n ethosu:\n - add perf counter support\n\n etnaviv:\n - force flush on power register ops\n\n msm:\n - support DSC configuration with slice_per_pkt \u003e 1\n\n mxsfb:\n - fix disable sequence\n\n panthor:\n - support sparse mappings\n\n rockchip:\n - switch away from simple helpers\n - support YUV background color\n - fix layer config timeout\n - add edp support for rk3576\n - add batch command submission function\n\n rocket:\n - error handling and NULL ptr deref fixes\n\n sun4i:\n - switch away from simple helpers\n\n imagination:\n - mark BXM-4-64 MC1 as support\n\n host1x:\n - support tegra264\n\n tegra:\n - add DSI for tegra 20/30\n\n v3d:\n - reduce PM runtime autosuspend delay\n - scheduler fixes and refactoring\n - deprecate v3d 3.3 and 4.1\n - validate CPU job query boundaries\n\n hibmc:\n - improve plane format handling\n - switch to gem shmem\n\n mediatek:\n - cec: correct compat for mt7623-8167?\n\n exynos:\n - remove simple dependency\n - add error handling to encoder paths\n - take i2c adapter module reference\"\n\n* tag \u0027drm-next-2026-08-20\u0027 of https://gitlab.freedesktop.org/drm/kernel: (2074 commits)\n drm/xe/mcr: Take vcs1/vecs1 into account for first media slice\n drm/xe: Fix a bug in pc_adjust_freq_bounds()\n drm/xe: Fix xe_device_probe() failure\n drm/xe/drm_ras: Move has_drm_ras check to drm_ras layer\n drm/xe/ras: Fix boot-time ras error processing\n drm/amd/display: make DC_RUN_WITH_PREEMPTION_ENABLED misuse a build error\n drm/amd/pm: silence uninitialized variable warnings\n drm/amdgpu: skip BOs being torn down during GTT recovery\n drm/amdgpu: Reject UVD message with invalid number of h265 refs\n drm/amdgpu: keep PRT mappings off the vm_bo state lists\n drm/amdgpu: fix nbif 6.3.1 l1 low power not functional\n drm/amd/display: fix BT.2020 YCbCr output CSC matrices for DCE\n drm/amd/display: fix BT.2020 YCbCr limited output CSC matrix\n drm/amdgpu: Implement insert_end for VCE 3\n drm/amdgpu: Fix UVD min buffer sizes\n drm/amdgpu: Fix UVD decode image min size calculation\n drm/amdgpu: Fix UVD dpb min size calculation for H264\n drm/amdgpu: Reject UVD message with dimensions above 4096\n drm/amdgpu: check ASPM on the dGPU host link\n drm/radeon: fix autosuspend cleanup during teardown\n ...\n" }, { "commit": "818bebeb63dd6bf5f4e07e145f6cdbace520a34c", "tree": "2bf0949016cf5bb61adae8561a5afa62bb938bb6", "parents": [ "8dbbc7e188949b6e1f6be4e82a44dd1d8b1d6c8b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 20:12:39 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 20:12:39 2026 -0700" }, "message": "drm/xe: Don\u0027t hand out the flat CCS storage as usable VRAM\n\nget_flat_ccs_offset() reads the base of the flat CCS storage from the\nhardware, scales it by the number of enabled L3 nodes, and rounds the\nresult up to 128K. Everything below that offset is then handed to the\nVRAM allocator as usable memory.\n\nRounding a limit that means \"usable memory ends here\" upwards publishes\nwhatever lies between the real base and the rounded one as free memory,\nand that memory belongs to the compression hardware. The scaled value\nhas no reason to be 128K aligned, and on a Battlemage G21 with 16 GiB it\nis not:\n\n\tflat CCS base: raw 0x3fafff800, rounded 0x3fb000000\n\nso the last 2 KiB of page 0x3fafff000 is CCS storage, in the allocator\u0027s\npool. Whatever is allocated there gets that tail overwritten by the\ncompression hardware, which needs no page-table entry, no buffer object\nand no GPU submission to do it, and does it before userspace exists.\n\nOn this machine a Mesa VM\u0027s level-3 page table landed on that page on\nevery cold boot. It lost the entry covering the compositor\u0027s\nbatch-buffer heap, so the compositor\u0027s first submission faulted fetching\nits batch and gdm restarted it forever: a black screen on an otherwise\nworking machine. Restarting gdm cleared it because the next VM\u0027s page\ntables were allocated somewhere else.\n\nRound down instead, to the page size the allocator works in. On this\nmachine that excludes exactly one page.\n\nReading the reserved page afterwards shows what had been writing it:\n\n\t[369] 0xcccc000000000000\n\t[371] 0xcc77000000000000\n\t[373] 0xcccc000000000000\n\t[375] 0xcc77000000000000\n\ncompression metadata, two bytes per sixteen, sitting where the driver\nused to hand out memory.\n\nThe assertion that should have caught this compares the offset against\nGSMBASE - ccs_size for equality. That value is 128K aligned, so it\nagrees with the rounded-up offset precisely when the base is not\naligned - the check cannot fail in the case it exists to catch, and is\ncompiled out unless CONFIG_DRM_XE_DEBUG is set. Replace it with one\nthat can fail: CCS storage must not run into GSM.\n\n[ And this was a debug session from hell, enormously helped by an AI\n doing much of the grunt-work.\n\n I\u0027d like to call it my tireless helper, but the AI several times\n stated flat out that this was impossible and unsolvable and that we\n should just write a report about it.\n\n I suspect those things have been trained by people who may not be\n quite as stubborn as I am.\n\n But while the AI was ready to give up several times, it did keep\n adding debug code and analyzing it faithfully when I pushed. So credit\n where credit is due and I let the AI write the commit message above.\n\n This is basically a one-liner fixing a bogus \"round_up()\" to a\n \"round_down()\", but there were 24 patches adding more and more debug\n information to this, and 18 kernel boot to finally narrow it down to\n this. - Linus ]\n\nFixes: 37173392741c (\"drm/xe/vram: fix ccs offset calculation\")\nCc: stable@kernel.org\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "8dbbc7e188949b6e1f6be4e82a44dd1d8b1d6c8b", "tree": "efc33f6ebb6b7081c54f2a7dc4c0efea50e5516a", "parents": [ "ed3b875bea55a3ec4837113356df2ead11115af9", "aedf2efd18977e0cef7eb963166e2b1fcc0aa321" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 20:05:59 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 20:05:59 2026 -0700" }, "message": "Merge tag \u0027mm-hotfixes-stable-2026-08-19-21-33\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm\n\nPull hotfixes from Andrew Morton:\n \"8 hotfixes. 5 are cc:stable. 5 are for MM.\n\n All are singletons, please see their changelogs for details\"\n\n* tag \u0027mm-hotfixes-stable-2026-08-19-21-33\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:\n mm/pagewalk: fix stale walk-\u003eaction escaping walk_pmd_range()\n mm, swap: don\u0027t free a hibernation slot that is in the swap cache\n mm: memcg-v1: fix memsw and TCP failcnt accounting\n mm/vmscan: report RCU-tasks quiescent states in shrink_lruvec()\n mailmap: add entries for Guodong Xu\n MAINTAINERS, mailmap: update email address for JP Kobryn\n MAINTAINERS: remove git URL for Squashfs\n memcg: keep folio\u0027s objcg same as its node\n" }, { "commit": "ed3b875bea55a3ec4837113356df2ead11115af9", "tree": "fd38aec3d0b15bc5db711df2ed20e2df2cf307d8", "parents": [ "7f063b2f17eaba2a35e251aa53627f2a70d536e2", "08745c62350126bc31b09548137be87e2866f628" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 18:17:08 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 18:17:08 2026 -0700" }, "message": "Merge tag \u0027mm-stable-2026-08-18-18-39\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm\n\nPull MM updates from Andrew Morton:\n\n - \"mm: drop \"sub\" prefix from various places\" (Dev Jain)\n\n page-\u003efolio conversion and a naming cleanup\n\n - \"mm/kasan: remove redundant initialization for kasan_flag_write_only\"\n (Igor Putko)\n\n KASAN cleanup work\n\n - \"mm/filemap: reduce unnecessary xarray lookups\" (Chi Zhiling)\n\n Small speedup in the pagecaache read code\n\n - \"mm/percpu: Fix possible NOFS/NOIO reclaim recursion\" (Kaitao Cheng)\n\n Improve the vmalloc code - mainly the avoidance of GFP_KERNEL\n allocations when the caller asked for GFP_NOFS or GFP_NOIO\n\n - \"mm/kmemleak: avoid soft lockup when scanning task stacks\" (Breno\n Leitao)\n\n Avoid a soft lockup watchdog trigger from the kmemleak scanning code\n in extreme situations\n\n - \"mm/page_owner: misc cleanups\" (Ye Liu)\n\n Cleanups to the page_owner code. For some reason lots of people have\n been working on the page_owner code this cycle.\n\n - \"mm: convert to walk_page_range_vma() to eliminate find_vma()\"\n (Kefeng Wang)\n\n Simplify and accelerate the page walking library function\n\n - \"mm/migrate: preparatory cleanups for batch copy and offload\"\n (Shivank Garg)\n\n Cleanups in the migration code\n\n - \"mm/page_owner: add per-fd filter infrastructure for print_mode and\n NUMA filtering\" (Zhen Ni)\n\n Per-fd filtering to page_owner in order to reduce the sometimes vast\n amount of output it can produce\n\n - \"mm: Refactor bootmem gigantic hugepage allocation\" (Muchun Song)\n\n Fixes and preparatory cleanups around bootmem HugeTLB handling,\n sparse initialization ordering, and related vmemmap setup\n\n - \"mm/zsmalloc: reduce lock contention in zs_free()\" (Wenchao Hao)\n\n Reduce lock contention in zs_free(), which dominates the unmap path\n under memory pressure on Android (LMK kills) and on x86 servers\n running zswap-heavy workloads.\n\n Up to 1.83x improvement in microbenchmarking.\n\n - \"move alloc_tag.c file under mm/\" (Suren Baghdasaryan)\n\n - \"samples/damon: handle damon_{start,stop}() failures\" (SJ Park)\n\n Fix improper handling of damon_start(), damon_stop(), and\n damon_call() failures across DAMON sample modules to prevent\n potential memory leaks, operation disruptions and use-after-free\n bugs\n\n - \"mm/damon/sysfs: kobject_del() directories that users can\n create/remove\" (SJ Park)\n\n Fix delayed sysfs directory removal under DEBUG_KOBJECT_RELEASE\n causeing creation failures due to duplicate directory names by adding\n missing kobject_del() calls before creating new directories\n\n - \"mm: cleanup clear_not_present_full_ptes()\" (David Hildenbrand)\n\n Clean up the core pte handling code\n\n - \"selftests/damon: misc fixes for test bugs\" (Kunwu Chan)\n\n Fix several bugs in the DAMON selftests\n\n - \"selftests/damon: fix memcg_path staging handling\" (Cheng Nie)\n\n Fix a bug in _damon_sysfs.py for damos_filter memcg_path setup, and\n add a test case for it in sysfs.py.\n\n - \"selftests/damon: test kdamond refresh_ms\" (Ruslan Valiyev)\n\n Selftest coverage for DAMON\u0027s refresh_ms sysfs feature by updating\n the test control module and verifying that scheme stats update\n automatically without manual intervention\n\n - \"mm/damon: five misc fixups\" (Akinobu Mita)\n\n Miscellaneous DAMON fixups.\n\n - \"mm/damon/core: detect internal variation above max_nr_regions/2\"\n (Jiayuan Chen)\n\n Fix DAMON\u0027s region splitting behavior when region counts exceed half\n the maximum budget by dynamically scaling down the split fraction as\n the limit approaches, preventing large regions from staying un-split,\n and add corresponding KUnit test coverage\n\n - \"mm: preparatory patches for PMD level swap entries\" (Usama Arif)\n\n Refactor and clean up PMD softleaf helpers, call sites, and\n architecture flags to lay the groundwork for a follow-up series that\n introduces PMD page table swap entries\n\n - \"mm/damon: update, optimize, and clean up doc, tests, and code\" (SJ\n Park)\n\n Update DAMON design and ABI documentation, expands unit and selftest\n coverage, optimize damon_commit_target_regions(), and clean up\n recently added sysfs interface code for better readability\n\n - \"mm/vmpressure: reduce CPU, memory and code overhead on cgroup v2\"\n (Usama Arif)\n\n Optimize vmpressure() by skipping unnecessary work on cgroup v2 for\n userspace event notifications and refactor v1-only eventfd handling\n into mm/memcontrol-v1.c to reduce memory overhead and code complexity\n\n - \"selftests/mm: refactor pkey helpers and fix mmap error handling\"\n (Hongfu Li)\n\n Refactor pkeys shared tracing and assertion helpers into a common\n file, unify protection key selftests to use consistent diagnostic\n logging and assertions, and enforce standardized MAP_FAILED return\n checks for mmap() calls across the tests\n\n - \"mm/damon: optimize out nr_accesses_bp\" (SJ Park)\n\n Replace the error-prone, continuously updated nr_accesses_bp field in\n damon_region with an on-demand moving sum function, reducing\n structure memory overhead and avoiding state corruption bugs\n\n - \"Open HugeTLB allocation routine for more generic use\" (Ackerley Tng)\n\n Decouple HugeTLB folio allocation from VMA dependencies by\n introducing hugetlb_alloc_folio(), enabling subsystems like\n guest_memfd to allocate HugeTLB folios without standard VMA\n reservations or pseudo-VMAs\n\n - \"mm/damon: provide pseudo moving sum probe_hits\" (SJ Park)\n\n Integrate DAMON\u0027s probe_hits attribute counter into the pseudo moving\n sum infrastructure, enabling real-time, online monitoring without\n waiting for full aggregation intervals\n\n - \"mm: Some cleanups for page allocator APIs\" (Brendan Jackman)\n\n Simplify and refactor the page allocator entry points and flags by\n unifying allocation paths, adding internal alloc_flags arguments, and\n eliminating redundant __ prefixed alloc_pages variants.\n\n - \"Fix incorrect access of hugetlb pte entries\" (Dev Jain)\n\n Enforce the consistent use of huge_ptep_get() instead of ptep_get()\n for HugeTLB entries and fixes an unaligned address issue in arm64\u0027s\n huge_ptep_get() implementation\n\n - \"mm/damon: validate all parameters in the core\" (SJ Park)\n\n Consolidate parameter validation into the DAMON core specifically\n within damon_start() and damon_commit_ctx() to centralize error\n checking, eliminate caller-side redundant checks and to improve\n maintenance efficiency\n\n - \"tools/mm/page_owner_sort: fix filtering and cleanup issues\" (Yichong\n Chen)\n\n Rename is_need() to filter_record() for clearer return semantics, fix\n per-record allocation memory leaks and bound output copies in\n search_pattern() to address an existing buffer issue\n\n - \"memcg: bail out reclaim when memcg is dying\" (Jiayuan Chen)\n\n Mitigate a system-wide stall which occurs when a cgroup is removed\n while one of its memory control files is doing synchronous reclaim\n\n - \"mm/memory-failure: add panic option for unrecoverable pages\" (Breno\n Leitao)\n\n Introduce an opt-in vm.panic_on_unrecoverable_memory_failure sysctl\n that immediately panics the kernel on unrecoverable memory errors in\n kernel-owned pages to preserve error context and prevent delayed,\n silent data corruption\n\n - \"mm/damon: refactor damon_{start,stop,commit}() for simple error\n handling\" (SJ Park)\n\n Refactor the DAMON core API functions to guarantee that all contexts\n are fully stopped when damon_start(), damon_stop(), or damon_commit()\n fail, eliminating the need for complex and error-prone caller-side\n cleanup code\n\n - \"Keep tail page private zero at free and folio split\" (Zi Yan)\n\n Add checks to ensure tail_page-\u003eprivate is zero when freeing compound\n or high-order pages and when promoting tail pages during large folio\n splits. By validating these fields at free and split time, it allows\n the removal of redundant private field clearing inside\n prep_compound_tail()\n\n - \"mm: drop redundant lru_add_drain in anon folio reuse paths\" (Barry\n Song)\n\n Eliminate redundant lru_add_drain() calls in\n wp_can_reuse_anon_folio() and do_swap_page() to reduce LRU lock\n contention and system overhead\n\n By validating folio refcounts against the LRU cache before draining\n and removing unnecessary drains in the swap path, it achieves up to a\n 30.5% reduction in drain calls during heavy swap workloads\n\n - \"mm: clean up folio LRU and swap declarations\" (Jianyue Wu)\n\n Reorganize folio LRU and swap code by relocating page-cluster state\n to mm/swap_state.c, renaming mm/swap.c to mm/folio.c, and moving\n MM-internal reclaim declarations into mm/internal.h.\n\n - \"userfaultfd: working set tracking for VM guest memory\" (Kiryl\n Shutsemau)\n\n Add userfaultfd support for tracking the working set of VM guest\n memory, so a VMM can identify hot pages and reclaim cold ones to\n tiered or remote storage\n\n - \"mm: remove CONFIG_HAVE_BOOTMEM_INFO_NODE (Part 2)\" (David\n Hildenbrand)\n\n Remove the remaining pieces of CONFIG_HAVE_BOOTMEM_INFO_NODE,\n performing some smaller cleanups around freeing of reserved vmemmap\n pages on the way.\n\n - \"mm/damon: update probe hits for runtime parameter commits\" (SJ Park)\n\n Ensure that DAMON\u0027s probe_hits attribute counter is properly updated\n when monitoring intervals are changed at runtime, matching the\n behavior of nr_accesses. To achieve this, it refactors and renames\n existing helper functions for shared use, applies the updates to\n probe_hits, and handles edge cases in damon_probe_hits_mvsum() to\n maintain measurement accuracy.\n\n - \"KSM: performance optimizations for rmap_walk_ksm\" (xu xin)\n\n Resolve a severe KSM reverse-mapping performance bottleneck where\n thousands of split VMAs sharing a single anon_vma cause extended lock\n contention.\n\n By adding an interval-filtering check during the rmap walk, it\n reduces worst-case anon_vma lock hold times from over 500ms down to\n under 2ms, preventing application freezes and latency spikes under\n memory pressure.\n\n - \"mm: split a couple of headers from internal.h\" (Mike Rapoport)\n\n Split declarations related to mm_init, memblock, vmalloc and sparse\n into new headers\n\n - \"KSM: use linear_page_index in collect_procs_ksm()\" (xu xin)\n\n Apply the interval tree optimization from rmap_walk_ksm() to\n collect_procs_ksm() to avoid iterating over non-matching VMAs during\n KSM memory error handling.\n\n It hoists loop-invariant address initialization and restricts the\n anon_vma_interval_tree_foreach walk to a targeted page offset range,\n reducing redundant checks and improving lookup efficiency.\n\n - \"selftests/mm: avoid false failures in hugetlb and KSM tests\" (Sayali\n Patil)\n\n Fix issues in the hugetlb and KSM MM selftest categories that can\n report failures when the prerequisites for the tests are not\n satisfied\n\n - \"mm/damon: introduce data attributes only monitoring\" (SJ Park)\n\n Introduce attribute-weighted region management in DAMON, allowing\n users to prioritize specific data attributes (such as page sizes or\n cgroups) over or instead of access monitoring.\n\n By assigning weights to attribute probes, DAMON can completely\n disable access tracking and adjust monitoring regions based on\n weighted probe-hit counters to optimize monitoring quality for\n attribute-focused workloads.\n\n - \"mm/hmm: Add mmap lock-drop support for userfaultfd-backed mappings\"\n (Stanislav Kinsburskii)\n\n Extend hmm_range_fault() to support userfaultfd-backed regions by\n allowing the mmap lock to be dropped during fault handling via a new\n hmm_range_fault_locked() helper.\n\n By accepting a locked pointer and signaling retry status when lock\n release occurs, it enables page fault resolution in userfaultfd\n regions while preserving backward compatibility for existing callers.\n\n - \"mm: make VMA page offset handling more consistent\" (Lorenzo Stoakes)\n\n Clean up and standardize how vma-\u003evm_pgoff is accessed and\n manipulated across file-backed and anonymous mappings in the kernel\n\n It introduces dedicated helper functions such as vma_start_pgoff(),\n vma_end_pgoff(), vma_set_pgoff() and linear_page_delta() while\n renaming rmap interval tree helpers to better reflect their\n functionality.\n\n These changes establish a cleaner foundation for future work that\n will unify virtual page offset indexing for all anonymous and CoW\u0027d\n folios.\n\n - \"mm: handle device-private PMDs in walk callbacks\" (Usama Arif)\n\n Address kernel panics and state corruption caused by MM walk\n callbacks reaching non-present device-private PMD swap entries\n created during HMM migrations\n\n It ensures that functions which acquire pmd_trans_huge_lock()\n properly recognize device-private PMDs instead of assuming a present\n THP or a standard migration entry.\n\n - \"mm/rmap: Refactor try_to_unmap_one\" (Dev Jain)\n\n Refactor try_to_unmap_one by modularizing Hugetlb,\n anonymous-lazyfree, and anonymous-swapbacked logic into dedicated\n functions, laying the structural groundwork for batched anonymous\n large folio unmapping.\n\n - \"Docs/ABI/damon: sysfs ABI document fixes and additions\" (Song Hu)\n\n Fix typos and fills in missing entries in the DAMON sysfs ABI\n document\n\n - \"dax/kmem: atomic whole-device hotplug via sysfs\" (Gregory Price)\n\n Introduce an atomic sysfs state attribute and supporting DAX/MM\n infrastructure to prevent userland races when offlining and removing\n entire memory regions\n\n By adding an unplugged state alongside standard online modes, it\n enables whole-device atomic hotplug control while preserving backward\n compatibility.\n\n - \"mm: convert more vm_flags_t users to vma_flags_t\" (Lorenzo Stoakes)\n\n Continue transitioning the kernel from the deprecated vm_flags_t type\n to vma_flags_t across core memory management infrastructure.\n\n It replaces legacy type usage in core functions such as do_mmap(),\n unmapped area allocation, mm-\u003edef_vma_flags, and VMA operations like\n mlock, mprotect, and mremap.\n\n - \"Two small patches to clean up mm/mm_slot.h\" (xu xin)\n\n Refactor mm_slot.h by introducing mm_slot_remove() to unify duplicate\n slot deletion sequences in khugepaged and KSM. It also adds code\n documentation explaining why mm_slot_lookup and mm_slot_insert must\n remain as preprocessor macros rather than static inline functions.\n\n - \"mm/damon/core: hide core-private struct fields\" (SJ Park)\n\n Clean up DAMON core structures by consistently marking internal-only\n fields with private: comment tags to prevent improper direct access\n from outer layers.\n\n It enforces encapsulation across core structures including\n damon_region, damon_target, and damon_ctx and updates DAMON_SYSFS to\n interact through approved access APIs instead of exposing raw struct\n members.\n\n - \"mm/damon: unurgent fixes for infinite loop, NULL de-ref and races\"\n (SJ Park)\n\n Address potential infinite loops, NULL dereferences, and race\n conditions identified in DAMON\n\n It fixes an infinite loop triggered by extreme user configurations, a\n NULL pointer dereference within unit tests and minor monitoring\n accuracy degradation caused by subtle runtime races.\n\n - \"mm/page_alloc: fixes for free_pages_nolock() on RT/UP\" (Brendan\n Jackman)\n\n Fix an NMI safety flaw in __free_frozen_pages() where freeing pages\n on non-SMP or PREEMPT_RT kernels can bypass can_spin_trylock() checks\n via non-PCP or isolated migration paths.\n\n It also resolves potential kernel crashes and privilege escalation\n risks triggered when BPF tracing runs in NMI context alongside memory\n hotplug or large allocation frees.\n\n - \"mm/page_alloc: couple of followups for recent cleanups\" (Brendan\n Jackman)\n\n Clean up and update page allocator nomenclature, documentation, and\n debug assertions.\n\n It aligns internal FPI_ flags with the public \"nolock\" naming\n convention, removes outdated internal implementation details from\n high-level page allocator comments, and eliminates obsolete\n VM_BUG_ON() assertions in allocation paths.\n\n - \"mm/mseal: further cleanups\" (Lorenzo Stoakes)\n\n Refactor and simplify the mseal implementation by clarifying API\n boundaries and removing unnecessary code complexity.\n\n It replaces generic do_mseal() usage outside the syscall with a\n dedicated mseal_mmap_page_zero() helper for MMAP_PAGE_ZERO,\n eliminates mm_struct parameters to enforce that sealing applies only\n to current-\u003emm, and streamlines overall logic and comments with no\n functional changes intended.\n\n - \"mm/vmscan: fix swappiness\u003dmax and clean up per-node proactive\n reclaim\" (Ridong Chen)\n\n Resolve reclaim behavior bugs and clean up function parameters across\n memory reclaim paths\n\n It fixes swappiness\u003dmax in both standard reclaim and MGLRU so\n unswappable anonymous memory no longer falls back to evicting page\n cache, ensures reclaim_store() returns accurate error codes instead\n of collapsing all failures into -EAGAIN, and removes the obsolete\n gfp_mask parameter from __node_reclaim().\n\n - \"mm: mincore: misc cleanups\" (Kefeng Wang)\n\n Clean up and simplifies the mincore code. Most importantly, it\n removes the historical special behavior that always reports VM_PFNMAP\n pages as non-resident.\n\n - \"mm/huge_memory: drop dead split helper variants\" (Kiryl Shutsemau)\n\n Two trivial cleanups in the folio split API\n\n - \"mm/damon: fix uninitialized DAMOS field and kunit exec expectation\n bugs\" (SJ Park)\n\n Resolve minor operational and testing bugs in DAMON identified by\n Sashiko. It initializes the damos-\u003elast_applied field to prevent\n occasional efficiency degradation and fixes invalid memory accesses\n in DAMON KUnit tests during test failure handling.\n\n - \"cleanup for stable_page_flags()\" (Jinjiang Tu)\n\n Clean up and refactor stable_page_flags() used by /proc/kpageflags\n without altering functionality.\n\n It uses BIT_ULL() to prevent shift-overflow warnings on 64-bit flag\n bits, converts folio-specific flag checks to standard folio_test_*()\n helpers, and removes redundant CONFIG_PAGE_IDLE_FLAG handling.\n\n - \"Batch unmap of uffd-wp file folios\" (Dev Jain)\n\n Extend batched folio unmapping support to file folios within\n userfaultfd write-protect (uffd-wp) VMAs by adding batching\n capabilities to pte_install_uffd_wp_if_needed().\n\n This removes special-case restrictions on uffd-wp VMAs in\n try_to_unmap_one(), significantly simplifying the function\u0027s control\n flow and complexity.\n\n - \"mm/early_ioremap: clarify and clean up early_ioremap_reset()\"\n (Sang-Heon Jeon)\n\n Clarify and clean up the architecture-specific usage of\n __late_set_fixmap() and __late_clear_fixmap() after\n early_ioremap_reset()\n\n It adds explicit documentation regarding when early_ioremap_reset()\n must be called and removes redundant macro definitions and reset\n calls in the RISC-V and ARM64 architectures.\n\n - \"mm: fix reclaim storms in defrag_mode\" (Johannes Weiner)\n\n Address severe performance regressions, swap storms, and spurious\n OOMs caused by vm.defrag_mode\u003d1 under high memory pressure in Meta\n production\n\n It updates the page allocator slowpath so non-movable allocation\n requests actively trigger direct reclaim and direct compaction at\n pageblock_order scale, allowing them to claim whole pageblocks rather\n than spinning unproductively.\n\n - \"zram: lockmap tweaks\" (Sebastian Siewior)\n\n Optimize and fix lockdep tracking for zram devices by consolidating\n per-entry lockmaps and isolate lock classes across multiple instances\n\n This reduces memory overhead by replacing per-entry lockdep_map\n instances with a single map per struct zram, and assigns a dynamic\n lock_class_key to each instance to prevent false deadlock reports\n when different zram devices are backed by distinct filesystems.\n\n* tag \u0027mm-stable-2026-08-18-18-39\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (501 commits)\n selftests/mm: thuge-gen: fix test_shmget() for PAGE_SIZE check\n selftests/mm: unpoison pages in memory-failure teardown\n mm/shmem: downgrade final i_blocks check in shmem_evict_inode() to pr_warn()\n mm/khugepaged: replace mutex_lock/mutex_unlock usage with guard macro\n mm/zsmalloc: fix release order of locks in zs_page_migrate()\n Documentation: zram: remove sections numbering\n ksm: stop iterating VMAs when ksm_test_exit returns true\n mm: fold userfaultfd_rwp() to false without CONFIG_ARCH_HAS_PTE_PROTNONE\n mm/migrate: report RCU-tasks quiescent states in migrate_pages_batch()\n zram: use a custom key for each zram object\n zram: move lockmap to be per-zram instead per table\n selftests/mm: fix gup_longterm EINVAL error message\n mm: page_alloc: fix non-movable reclaim storm in defrag_mode\n mm: page_alloc: move capture_control to the page allocator\n mm: compaction: support non-movable compaction for pageblock requests\n mm: page_alloc: __GFP_FS lockdep annotation for direct compaction\n hugetlb: evaluate subpool free state while locked\n mm/damon: remove trailing semicolons after function definitions\n mm/damon/ops-common: prevent migration fallback to non-target nodes\n mm/damon: update outdated comment about DAMOS filter handling\n ...\n" }, { "commit": "934753d26f1ed7fadeedf1b9e7df6fe75de303e8", "tree": "3c3d6f9cdb553351a42ee81cdfe371d754584853", "parents": [ "8d3ae59288f1e7d58d76558a6ee96d533bc5019f" ], "author": { "name": "Paulo Alcantara", "email": "pc@manguebit.org", "time": "Tue Aug 18 11:07:38 2026 -0300" }, "committer": { "name": "Paulo Alcantara", "email": "pc@manguebit.org", "time": "Thu Aug 20 21:32:00 2026 -0300" }, "message": "MAINTAINERS: Replace Steve French as CIFS maintainer\n\nUnfortunately, due to health reasons, Steve French can no longer\ncontinue as CIFS maintainer; therefore, I am taking over the role with\nNamjae Jeon as secondary maintainer.\n\nAcked-by: Namjae Jeon \u003clinkinjeon@kernel.org\u003e\nReviewed-by: David Howells \u003cdhowells@redhat.com\u003e\nSigned-off-by: Namjae Jeon \u003clinkinjeon@kernel.org\u003e\nCc: Steve French \u003csmfrench@gmail.com\u003e\nCc: Steve French \u003csfrench@samba.org\u003e\nCc: Greg Kroah-Hartman \u003cgregkh@linuxfoundation.org\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nCc: Christian Brauner \u003cbrauner@kernel.org\u003e\nCc: Ronnie Sahlberg \u003cronniesahlberg@gmail.com\u003e\nCc: Shyam Prasad N \u003csprasad@microsoft.com\u003e\nCc: Tom Talpey \u003ctom@talpey.com\u003e\nCc: Bharath SM \u003cbharathsm@microsoft.com\u003e\nCc: samba-technical@lists.samba.org\nCc: linux-cifs@vger.kernel.org\nCc: linux-fsdevel@vger.kernel.org\nCc: linux-kernel@vger.kernel.org\nAcked-by: Tom Talpey \u003ctom@talpey.com\u003e\nSigned-off-by: Paulo Alcantara \u003cpc@manguebit.org\u003e\n" }, { "commit": "7f063b2f17eaba2a35e251aa53627f2a70d536e2", "tree": "82676aefc15baa911c835e69c840184ee45607c7", "parents": [ "a625b2a387628df94e385faf5c81bf252f304ed9", "b282418bc366194677eafd1dad180d92254586ac" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 17:07:50 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 17:07:50 2026 -0700" }, "message": "Merge tag \u0027for_linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost\n\nPull vhost,vdpa,virtio updates from Michael Tsirkin:\n\n - transport v3 support in virtio-mmio\n\n - suspend support in vduse\n\n - fixes, cleanups all over the place\n\n* tag \u0027for_linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (54 commits)\n vduse: Add suspend\n vduse: do not take rwsem at reset work flush\n vduse: add F_QUEUE_READY feature\n vduse: add VDUSE_SET_FEATURES ioctl\n vduse: add VDUSE_GET_FEATURES ioctl\n vduse: store control device pointer\n tools/virtio: Fix control typo in trace agent comment\n tools/virtio: Fix userspace typo in vringh test comment\n vhost: reject zero-size IOTLB INVALIDATE\n vdpa: Remove redundant dev_err()\n virtio_ring: fix infinite loop in virtnet_poll_cleantx when device is broken\n vdpa/mlx5: roll back MR update after VQ setup failure\n MAINTAINERS: remove Gabriel from LiteX and fw-cfg drivers\n virtio_mem: fix typo in comment\n vdpa/solidrun: fix typos in snet_ctrl comments\n virtio: fix article before virtio in dma-buf comment\n vhost: fix inaccurate kdoc in iotlb helpers\n virtio: rtc: time out alarm requests\n vdpa/mlx5: fix wrong MLX5_ADDR_OF struct type in alloc_inout()\n vdpa: octeon_ep: add missing MODULE_DEVICE_TABLE()\n ...\n" }, { "commit": "a625b2a387628df94e385faf5c81bf252f304ed9", "tree": "e3be1a5a3d57d45aea64e977827d2fd9c6adfd7a", "parents": [ "77ae27fd98f3b548797c9f22c10ab5cf1c4ada53", "4e3c1fc8abcb8eff062150b4340fa4569696d645" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 16:57:59 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 16:57:59 2026 -0700" }, "message": "Merge tag \u0027vfio-v7.3-rc1\u0027 of https://github.com/awilliam/linux-vfio\n\nPull VFIO updates from Alex Williamson:\n\n - Add nv_falcon vfio selftest driver.\n\n The Falcon is a general-purpose microcontroller embedded within\n NVIDIA GPUs, presenting a relatively simple DMA programming\n interface.\n\n This adds another selftest target for vfio-pci with real DMA\n transfers (Rubin Du, Alex Williamson)\n\n - Add allocation assertion helpers to vfio selftests and use them to\n avoid variable length arrays and the compiler errors they generate\n (Alex Mastro)\n\n - Fix use-after-free hazard where an init path error in MSI support\n leaves a stray pointer that can later be reused or double-freed\n (Xiang Mei)\n\n - Fix previous refactor of PCI BAR mappings to honor non_mappable_bars\n flag, which otherwise generates a warning when trying to pci_iomap()\n a 256TiB BAR on ISM devices on s390 (Farhan Ali)\n\n - Add igb vfio selftest driver.\n\n Like nv_falcon, this provides another target for DMA testing with\n vfio selftests, but importantly this driver supports both physical\n 82576 NICs and the emulation model in QEMU.\n\n This therefore enables a vfio selftest vector with no physical\n hardware requirements (Josh Hilke, Alex Williamson)\n\n - Mark selftest fixture objects __maybe_unused to accommodate builds\n with clang -Wunused-but-set-global (David Matlack)\n\n - Add error recovery for vfio-pci devices on s390x.\n\n This expands devices which expose the existing error eventfd and\n introduces a device feature for reporting firmware defined error\n state information to the user, allowing recovery through hypervisor\n channels (Farhan Ali)\n\n* tag \u0027vfio-v7.3-rc1\u0027 of https://github.com/awilliam/linux-vfio:\n vfio/pci: Remove the pcie check for VFIO_PCI_ERR_IRQ_INDEX\n vfio-pci/zdev: Add a device feature for error information\n s390/pci: Store PCI error information for passthrough devices\n PCI/MSI: Enable memory decoding before restoring MSI-X messages\n PCI: Fail FLR when config space is inaccessible\n PCI: Avoid saving config space state if inaccessible\n PCI: Allow per function PCI slots to fix slot reset on s390\n PCI: Introduce PCI_SLOT_PLACEHOLDER constant for slot_nr placeholder value\n selftests: harness: Mark test fixture objects __maybe_unused\n vfio: selftests: Retry on EAGAIN during device reset\n vfio: selftests: igb: Add driver for Intel 82576 device\n vfio: selftests: Add helpers to re-enable interrupts\n vfio/pci: Avoid mapping BARs for devices with non-mappable BARs\n vfio/pci: clear vdev-\u003emsi_perm after freeing it on init failure\n vfio: selftests: Avoid VLAs\n vfio: selftests: Add allocation assert helpers\n vfio: selftests: Add NVIDIA Falcon driver for DMA testing\n vfio: selftests: Allow drivers without send_msi() support\n vfio: selftests: Add generic PCI command register helpers\n vfio: selftests: Add memcpy chunking to vfio_pci_driver_memcpy()\n" }, { "commit": "77ae27fd98f3b548797c9f22c10ab5cf1c4ada53", "tree": "e34b7bfc2811f8ff82742f10d5dee2d2fa68f7cd", "parents": [ "6439079365dcb33c6701f5f0e480ac2c17312b6b", "3783364ce6f883b5020b031e65cab6d3cd450b82" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 15:34:31 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 15:34:31 2026 -0700" }, "message": "Merge tag \u0027printk-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/printk/linux\n\nPull printk updates from Petr Mladek:\n\n - Prevent a possible out-of-bound access and a use-after-free in rather\n theoretical situations\n\n - Make no_hash_pointers take effect early\n\n - Some fixes and clean up of the ratelimit KUnit test\n\n* tag \u0027printk-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/printk/linux:\n printk: Handle pre-enabled consoles in the top-level register_console()\n printk: Fix possible console use-after-free\n lib/tests: test_ratelimit: fix stress test thread lifecycle and leak\n lib/vsprintf: Make no_hash_pointers take effect early\n" }, { "commit": "6439079365dcb33c6701f5f0e480ac2c17312b6b", "tree": "e40d568d716a7d01e5cc05f9baed32be8d22464b", "parents": [ "920f27122cbacfd3b540a3f2f67b0145203d5581", "29a86c5e6361caffa7e75e891169e813f82ff688" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 15:11:52 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 15:11:52 2026 -0700" }, "message": "Merge tag \u0027probes-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull probes updates from Masami Hiramatsu:\n \"BTF typecasting and variable fetch enhancements:\n\n - Typecast support across probe events: Extended BTF typecasting\n syntax (e.g., (STRUCT)PARAM-\u003eMEMBER) to kprobes, uprobes, and\n fprobes on function entry and return\n\n - Nested typecasts: Added support for chaining and nesting typecasts\n up to 3 levels, including casting registers and stack variables\n\n - Field specifier option: Added (STRUCT,FIELD) syntax to emulate\n container_of(), allowing retrieval of parent structures from member\n pointers\n\n - $current variable support: Introduced $current special variable to\n access the running task_struct via BTF dereferencing\n\n - Per-CPU variable access: Added this_cpu_read() and this_cpu_ptr()\n fetcharg methods to trace CPU-local data safely\n\n - Fetcharg bytecode dumper: Added CONFIG_PROBE_EVENTS_DUMP_FETCHARG\n to dump the compiled fetcharg bytecode instructions as comments in\n dynamic_events\n\n - Extended symbol name handling: Removed the MAX_COMMON_HEAD_LEN\n limit and extended MAX_ARGSTR_LEN to 256 bytes, enabling probing of\n long symbols, mangled Rust symbols and complex BTF expressions\n\n - eprobe variable syntax: Allowed eprobes to reference event fields\n directly without requiring a \u0027$\u0027 prefix\n\n - Cleanup unused parameters, redundant codes, duplicate macros and\n pointer arithmetic\n\n - Use a ternary operator for simplifying fetch_type_from_btf_type()\n\n Expanded boot time dynamic probe support:\n\n - Add boot-time tracing configuration support for event probes\n (eprobes), function probes (fprobes), and tracepoint probes\n (tprobes)\n\n - Allow comment lines (\u0027#\u0027) in dynamic_events file\n\n Optimization, robustness, and cleanups:\n\n - Simplify fprobe_remove_ips() by reusing graph and ftrace helpers\n\n - Remove __packed attribute from struct __fprobe_header to avoid\n unaligned memory access penalties on RISC architectures\n\n - Remove redundant memset() calls in perf event probe handlers\n\n - Replace legacy __ASSEMBLY__ with __ASSEMBLER__ in header files\n\n Selftests \u0026 refactoring:\n\n - Refactor parse_probe_arg() and parse_probe_vars(), and eliminate\n recursion in probe argument parsing to protect kernel stack depth\n\n - Add selftests for BTF typecasts and module probing without module\n prefixes\n\n - Force LC_ALL\u003dC in ftracetest to prevent test failures on localized\n systems\n\n - Refactor btf_type_skip_modifiers() to remove ignored id parameter\n\n - Sort ERRORS list in trace_probe.h alphabetically\n\n - Fix typo in fprobe docs, and trace_fprobe function name\n\n - Rename FETCH_OP_DATA to FETCH_OP_IMMSTR\n\n - Make file offset error message probe-agnostic\"\n\n* tag \u0027probes-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (37 commits)\n fprobe: Simplify fprobe_remove_ips() by reusing existing helpers\n tracing/boot: Add support for eprobe, fprobe, and tprobe events\n selftests/ftrace: Force C locale in ftracetest\n tracing/probes: Treating longer symbol name on event comparation\n docs: trace: fprobe: fix \u0027thos\u0027 spelling\n tracing/probes: Fix extra whitespace in trace_probe_kernel.h\n tracing/kprobe: Remove redundant memset in kprobe_perf_func()\n tracing/fprobe: Remove redundant memset in fentry_perf_func()\n tracing/fprobe: Remove redundant snprintf in trace_fprobe_match_command_head()\n tracing/probes: Simplify BTF_KIND_PTR case in fetch_type_from_btf_type()\n tracing/probes: Cleanup pointer arithmetic in store_trace_entry_data()\n tracing/probes: Remove unused parameter from parse_probe_var_retval()\n tracing/probes: Remove redundant bounds check in trace_probe_compare_arg_type()\n tracing/probes: Remove redundant boolean conversion in trace_probe_has_single_file()\n tracing/probes: Remove duplicate MAX_ARRAY_LEN macro definition\n selftests/ftrace: Add test case for a symbol in a module without module name\n tracing/probes: Eliminate recursion in parse_probe_arg()\n tracing/probes: Extend max length of argument string\n tracing/probes: Sort ERRORS list in trace_probe.h alphabetically\n tracing/probes: Refactor parse_probe_arg()\n ...\n" }, { "commit": "d463633d63e6aa10384683bf39971a4b00780c56", "tree": "ff8298374b5da3a1363b7a9ad3f69468935861cb", "parents": [ "51df976c3268b96a926d8f389fd9e91c0b5392e1" ], "author": { "name": "Miro Kropacek", "email": "miro.kropacek@gmail.com", "time": "Thu Aug 20 10:28:37 2026 +1200" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Fri Aug 21 00:02:41 2026 +0200" }, "message": "fbdev: atafb: Add support for SuperVidel\u0027s SuperBlitter\n\nThe SuperVidel graphics FPGA includes a hardware blitter (bit block\ntransfer engine) operating within the SuperVidel DDR2 video RAM.\n\nTwo versions of this blitter exist. Later versions (\u003e\u003d 9) of the\nSuperVidel firmware support an asynchronous command FIFO, older\nversions must be polled for command completion.\n\nAdd hardware-accelerated copyarea, fillrect and imageblit fb\noperations (falling back to the non-accelerated versions for\nanything that exceeds blitter capabilities).\n\nSigned-off-by: Miro Kropacek \u003cmiro.kropacek@gmail.com\u003e\nReviewed-by: Michael Schmitz \u003cschmitzmic@gmail.com\u003e\nSigned-off-by: Michael Schmitz \u003cschmitzmic@gmail.com\u003e\nLink: https://lists.debian.org/debian-68k/2026/08/msg00000.html\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "51df976c3268b96a926d8f389fd9e91c0b5392e1", "tree": "3110eb808166a1604710bfdfa3d9bac37fb5d689", "parents": [ "d7dda89ad05173f9bbb795cd3cfc38262381f7f4" ], "author": { "name": "Miro Kropacek", "email": "miro.kropacek@gmail.com", "time": "Thu Aug 20 10:28:36 2026 +1200" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Fri Aug 21 00:02:41 2026 +0200" }, "message": "fbdev: atafb: Add support for further video bit depths on atafb:external\n\nSupervidel offers additional video bit depths: 8-bit chunky,\n16-bit RGB565 (also on the original Videl) and ARGB888.\n\nAdd code to support these bit depths.\n\nSigned-off-by: Miro Kropacek \u003cmiro.kropacek@gmail.com\u003e\nReviewed-by: Michael Schmitz \u003cschmitzmic@gmail.com\u003e\nSigned-off-by: Michael Schmitz \u003cschmitzmic@gmail.com\u003e\nLink: https://lists.debian.org/debian-68k/2026/08/msg00000.html\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "d7dda89ad05173f9bbb795cd3cfc38262381f7f4", "tree": "93c0349e5821415a479542391489788e74c7ae34", "parents": [ "f8e43fe0f22b7137ce456e6fe3581d3098174f74" ], "author": { "name": "Miro Kropacek", "email": "miro.kropacek@gmail.com", "time": "Thu Aug 20 10:28:35 2026 +1200" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Fri Aug 21 00:02:41 2026 +0200" }, "message": "fbdev: atafb: Give atafb proper parent\n\nThe atafb fb device registers no parent, causing a missing symlink\n(/sys/class/graphics/fb0/device). Xorg fbdevhw driver looks for that\nsymlink when scanning for devices, so add a parent node for atafb.\n\nThe proper way to do that is use of framebuffer_alloc(), which\nrequries refactoring atafb to move from static fb_info to allocation\nof fb_info by framebuffer_alloc().\nAbout the only location where a fb_info pointer cannot be passed\nis monspec setup in atafb_setup(). Store static monspecs there\nand copy into place after framebuffer_alloc().\n\nPass fb_info everywhere it\u0027s needed (detect, decode_var and\ndo_fb_set_var functions), and remove the static fb_info struct.\n\nSigned-off-by: Miro Kropacek \u003cmiro.kropacek@gmail.com\u003e\nReviewed-by: Michael Schmitz \u003cschmitzmic@gmail.com\u003e\nSigned-off-by: Michael Schmitz \u003cschmitzmic@gmail.com\u003e\nLink: https://lists.debian.org/debian-68k/2026/08/msg00000.html\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "920f27122cbacfd3b540a3f2f67b0145203d5581", "tree": "3f7121a1b7255f12736ed2c96225ce29dea9a788", "parents": [ "073e62fd33fe9cec754cb89e60c0ebbab781a50a", "7098e9cd98a05c0c5de2fae0c2465f9d966fdd07" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 15:02:09 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 15:02:09 2026 -0700" }, "message": "Merge tag \u0027cxl-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl\n\nPull CXL updates from Dave Jiang:\n \"This includes the remaining patches to complete the basic support for\n CXL type2 (accelerators) drivers by the CXL core. A set of unit\n testing support to catch regressions for the CXL core type2 exported\n functions has also been added. The actual CXL type 2 driver support\n for the AMD Solarflare NIC driver has been picked up by net-dev.\n\n A set of preparatory patches has been picked up from the CXL port\n error handling series in order to reduce the number of patches in that\n series for review. The rest of that series is getting close to be\n merged but was not ready for the 7.3 merge window.\n\n There are also a number of misc patches and a large number of bug\n fixes against pre-existing issues flagged by sashiko reviews as\n contributors post new patches on the mailing list. We will continue to\n work through the sashiko raised issues as they show up.\"\n\n* tag \u0027cxl-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl: (35 commits)\n cxl/Documentation: Spell out globbed sysfs attribute names\n cxl: Rename find_cxl_port() to find_cxl_port_by_dport()\n cxl: Tighten CPER kfifo registration API and symbol visibility\n acpi/apei/ghes: Use raw_spinlock_t for CXL CPER work locks\n cxl/ras: Fix cxl_rch_get_aer_severity() wrong severity register\n cxl/ras: Fix cxl_rch_get_aer_info() out-of-bounds AER register read\n cxl/test: Assign one mock memdev a full-width serial number\n cxl/core: Format the memdev serial number as unsigned in TP_printk\n cxl/pmem: Format the nvdimm serial number as unsigned decimal\n cxl: Use %pe to print error pointers\n cxl/region: Use __free(put_device) in find_pos_and_ways()\n cxl/region: Fix use-after-free in find_pos_and_ways() error path\n cxl: Deny Features commands on the RAW mailbox path\n cxl/features: Clamp Get Feature output size to the remaining buffer\n cxl/features: Reject Set Features output buffer smaller than the header\n cxl/features: Reject Get Feature count larger than the output buffer\n cxl/port: Restart port enumeration when a sibling adds the dport first\n cxl/features: Serialize multi-part Get/Set Feature transfers\n cxl/pci: Honor -EPROBE_DEFER from component register setup\n cxl/mbox: Break poison list loop on an empty payload\n ...\n" }, { "commit": "073e62fd33fe9cec754cb89e60c0ebbab781a50a", "tree": "796305cc43819b294ef8877eb7bdb0d27fe567bd", "parents": [ "73cc4da3857bb2970b3e349ed9608ca37861a84e", "8da45c93dfa98f25b148512d92ecfa75d027a5cc" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 14:53:54 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 14:53:54 2026 -0700" }, "message": "Merge tag \u0027riscv-for-linus-7.3-mw1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux\n\nPull RISC-V updates from Paul Walmsley\n\n - Add initial definitions and discovery for the Smcsrind, Sscsrind,\n Smcntrpmf, Ssccfg, Smcdeleg, Zicclsm, Ziccamoa, Ziccif, Ziccrse,\n Za64rs, and Ssqosid RISC-V ISA extensions\n\n - Improve the RISC-V update_mmu_cache_range() implementation by using\n flush-by-ASID, enabling performance improvements on\n microarchitectures that support related optimizations; and by taking\n advantage of the Svinval RISC-V ISA extension on microarchitectures\n that support it\n\n - Shrink CFI shadow stack allocation further (to 512MB) to save virtual\n address space (and physical memory on systems with strict overcommit\n policies)\n\n - Add initial CPU context switch support for QoS tagging (Ssqosid)\n\n - Change our vector misaligned access speed test code to be\n synchronous, avoiding a bunch of bugs related to unnecessary\n asynchronicity\n\n - Enable ARCH_HAS_ACPI_TABLE_UPGRADE for RISC-V to help with ACPI table\n debugging (similar to ARM64 and x86)\n\n - Extract vDSO section offsets at build time, rather than run time, to\n avoid the boot time overhead\n\n - Use assembler directives to control the use of instructions from the\n RISC-V ISA extensions Zacas and Zabha, rather than compiler -march\n flags that could affect the whole kernel (similar to ARM64)\n\n - Add a kselftest for kprobes support for the c.jal instruction on\n RISC-V\n\n - When UEFI runtime services are available, use them to restart and\n power off\n\n - Fix ARCH_HAS_ACPI_TABLE_UPGRADE for RISC-V by adding a missing\n include file to the ACPI code\n\n - Move contiguous DMA memory reservation later in the boot, to ensure\n that the correct global and per-node pools are allocated when\n CONFIG_CMA_SIZE_PERNUMA is enabled\n\n - Disallow probes on breakpoint handlers to avoid recursing\n indefinitely into do_trap_break()\n\n - Patch the compat vDSO during runtime alternatives processing, not\n only the standard vDSO\n\n - Remove some leftover XIP support elements missed by commit\n 9b3a2be84803 (\"riscv: Remove support for XIP kernel\")\n\n - Fix broadcast timer switching in ACPI LPI power states by\n implementing a RISC-V version of arch_get_idle_state_flags()\n\n - Miscellaneous fixes, including: vDSO makefile simplification; marking\n our default_power_off() as __noreturn; error path fixes in the PMU\n SBI perf driver; RISC-V extension capitalization consistency; the use\n of BIT() macros in one of our include files; and some documentation\n fixes\n\n* tag \u0027riscv-for-linus-7.3-mw1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux: (39 commits)\n RISC-V: hwprobe: Use BIT macro to avoid warnings\n riscv: alternative: Also patch the compat vDSO\n riscv: alternative: Use the statically extracted vDSO section offsets\n riscv: vdso: Add symbols for the alternative section boundaries\n riscv: Add Ziccamoa, Ziccif, Ziccrse, and Za64rs to cpufeature and hwprobe\n riscv: Add Zicclsm to cpufeature and hwprobe\n riscv: cpufeature: Clarify ISA spec version for canonical order\n riscv: Standardize extension capitalization\n riscv: kprobes: Prevent probes in breakpoint handlers\n riscv: cmpxchg: Use .option arch for Zacas and Zabha\n riscv: acpi: Handle LPI architectural context loss flags\n riscv: further remove XIP\n riscv: Add support for srmcfg CSR from Ssqosid extension\n riscv: Detect the Ssqosid extension\n dt-bindings: riscv: Add Ssqosid extension description\n dt-bindings: riscv: add Counter delegation ISA extensions description\n RISC-V: Add Ssccfg/Smcdeleg ISA extension definition and parsing\n RISC-V: Add Ssccfg extension CSR definition\n dt-bindings: riscv: add Smcntrpmf ISA extension description\n RISC-V: Add Smcntrpmf extension parsing\n ...\n" }, { "commit": "73cc4da3857bb2970b3e349ed9608ca37861a84e", "tree": "22c32246b2341fb9d6d5df5b6921a25423d28c97", "parents": [ "caf325ab81be0bc22adf75248ecd22f52f01ad21", "2f8e3cad53b5c36ab0ed5d3195bfc55c59ea61a5" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 14:49:34 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 14:49:34 2026 -0700" }, "message": "Merge tag \u0027m68k-for-v7.3-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k\n\nPull m68k updates from Geert Uytterhoeven:\n\n - Fix backtraces for non-running tasks\n\n - Fix deadlock in show_cons_active()\n\n - defconfig updates\n\n* tag \u0027m68k-for-v7.3-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:\n m68k: nfcon: Do not call console_is_registered() in nfcon_device()\n m68k: Fix backtraces for non-running tasks\n m68k: defconfig: Update defconfigs for v7.2-rc1\n m68k: amiga: Remove redundant amiga_reset() prototype\n" }, { "commit": "caf325ab81be0bc22adf75248ecd22f52f01ad21", "tree": "f568f39418e0141411fad0f79f986782fe4b7d2f", "parents": [ "55ab7e14222e5f0b0fd9f7711ca391d2924b35e3", "6afb03599653619e661121628611f808de1f7c7a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 14:46:50 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 14:46:50 2026 -0700" }, "message": "Merge tag \u0027loongarch-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson\n\nPull LoongArch updates from Huacai Chen:\n\n - Add DIRECT_MAP_PHYSMEM_END definition\n\n - Expand module virtual address space to 2GB\n\n - Use current_stack_pointer in current_pt_regs()\n\n - BPF JIT enhancements such as kptr_xchg and timed may_goto\n\n - Some bug fixes and other small changes\n\n* tag \u0027loongarch-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson:\n selftests/bpf: Enable kptr_xchg_inline test on LoongArch\n LoongArch: BPF: Add arch_bpf_stack_walk() implementation\n LoongArch: BPF: Add timed may_goto implementation\n LoongArch: BPF: Resolve per-CPU addrs for internal-only MOV\n LoongArch: BPF: Advertise JIT support for kptr xchg inline\n LoongArch: BPF: Align value-returning atomics with LKMM\n LoongArch: BPF: Split unconditional branch JA paths statically\n LoongArch: BPF: Remove dead move_imm() call in BPF_NEG path\n LoongArch: BPF: Remove redundant zext jumping in move_imm()\n LoongArch: BPF: Implement branchless conditional move for TCC\n LoongArch: BPF: Refactor jump offset calculation in tail call\n LoongArch: BPF: Move arena register slot below TCC context\n LoongArch: BPF: Optimize redundant TCC loads in epilogue\n LoongArch: Use current_stack_pointer in current_pt_regs()\n LoongArch: Use generic cmp_int() instead of custom cmp_3way()\n LoongArch: Expand module virtual address space to 2GB\n LoongArch: Add DIRECT_MAP_PHYSMEM_END definition\n LoongArch: Fix acpi_package_ids[] array overflow\n" }, { "commit": "f8e43fe0f22b7137ce456e6fe3581d3098174f74", "tree": "b516bcc60181f982386139edee0eba40273c18a2", "parents": [ "f74cf70e5e78a43bd316e5c8e371caea0a49c207" ], "author": { "name": "Runyu Xiao", "email": "runyu.xiao@seu.edu.cn", "time": "Tue Aug 18 21:53:18 2026 +0800" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Thu Aug 20 23:45:07 2026 +0200" }, "message": "fbdev: omapfb: panel-dsi-cm: initialize lock before registering display\n\ndsicm_probe() registers the display before initializing ddata-\u003elock.\nOnce omapdss_register_display() publishes the display, another consumer\ncan reach a dsicm callback that takes this mutex while it is still\nuninitialized.\n\nInitialize the mutex before registering the display so the published\ncallbacks always see a valid lock.\n\nFixes: f76ee892a99e (\"omapfb: copy omapdss \u0026 displays for omapfb\")\nCc: stable@vger.kernel.org\nSigned-off-by: Runyu Xiao \u003crunyu.xiao@seu.edu.cn\u003e\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "f74cf70e5e78a43bd316e5c8e371caea0a49c207", "tree": "201d47d3a41893056bdbbedbec9d0a603a15e48e", "parents": [ "ba13226fc966004713aed47931db8922b84d1960" ], "author": { "name": "Ajith P V", "email": "ajithpv.linux@gmail.com", "time": "Tue Aug 18 11:17:22 2026 +0000" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Thu Aug 20 23:45:07 2026 +0200" }, "message": "fbdev: viafb: refactor strcpy and viafb_name\n\nEliminate the single-use file-scope variable `viafb_name` and pass the\n\"Via\" string literal directly to strcpy().\n\nSince \"Via\" is a literal constant, the compiler safely executes\ncompile-time bounds checking during fortify verification. This satisfies\nthe security initiative requirements while minimizing code complexity.\n\nSigned-off-by: Ajith P V \u003cajithpv.linux@gmail.com\u003e\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "ba13226fc966004713aed47931db8922b84d1960", "tree": "876285f76821e0a08d3ecd11bcda954f5dc722aa", "parents": [ "772bd1a84288e07d2621108473c96602c3d6b6bb" ], "author": { "name": "BingKun Yue", "email": "yuebingkun@kylinos.cn", "time": "Mon Aug 17 14:39:42 2026 +0800" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Thu Aug 20 23:45:07 2026 +0200" }, "message": "fbdev: platinumfb: add error checking for ioremap calls\n\nThe ioremap() and ioremap_wt() calls in platinumfb_probe() were not\nchecked for failure. If any of these mappings fail, the driver would\ndereference NULL pointers, leading to a kernel panic.\n\nAdd proper error checking and use goto-based cleanup to avoid code\nduplication across the error paths.\n\nSigned-off-by: BingKun Yue \u003cyuebingkun@kylinos.cn\u003e\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "772bd1a84288e07d2621108473c96602c3d6b6bb", "tree": "8a8cb873d0b453ec38ba82bdd09b7458863c4a55", "parents": [ "c761c0400fb7785672a5df5b21518f1dfd8fc81a" ], "author": { "name": "Randy Dunlap", "email": "rdunlap@infradead.org", "time": "Fri Aug 14 11:42:34 2026 -0700" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Thu Aug 20 23:45:07 2026 +0200" }, "message": "fbdev: maxine: use MODULE_LICENSE() unconditionally\n\nThis driver cannot be built as a loadable module so testing for\n\"#ifdef MODULE\" is not appropriate here.\nAlso, MODULE_LICENSE() is always available.\n\nSigned-off-by: Randy Dunlap \u003crdunlap@infradead.org\u003e\nAcked-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "c761c0400fb7785672a5df5b21518f1dfd8fc81a", "tree": "4d6985f39cfeee9d8826e32d865cd9e4ee0b4c74", "parents": [ "3adec5c640dc09dca5605043e138412fd5a0629d" ], "author": { "name": "Randy Dunlap", "email": "rdunlap@infradead.org", "time": "Thu Aug 13 19:56:40 2026 -0700" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Thu Aug 20 23:45:07 2026 +0200" }, "message": "fbdev: maxine: fix maxinefb_init() return value\n\nA driver should return a negative error code on failure of its\nmodule_init() function so that the system recognizes the failure.\nChange the \"return 1\" to \"return -ENODEV\".\n\nSuggested-by: sashiko-bot@kernel.org\nLink: https://sashiko.dev/#/patchset/20260809234810.982500-1-rdunlap@infradead.org?part\u003d1\nSigned-off-by: Randy Dunlap \u003crdunlap@infradead.org\u003e\nAcked-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "3adec5c640dc09dca5605043e138412fd5a0629d", "tree": "39b5ea051953cb0b9f1e75ded4b4cd76333bae72", "parents": [ "47eb05e731fbef11dc36e9a9a9a05e61d43480ae" ], "author": { "name": "Randy Dunlap", "email": "rdunlap@infradead.org", "time": "Thu Aug 13 16:29:34 2026 -0700" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Thu Aug 20 23:45:07 2026 +0200" }, "message": "fbdev: maxine: fix 64-bit build error\n\nThe KSEG1ADDR() macro is only defined for non-64BIT builds.\nUse the CKSEG1ADDR() macro instead.\n\nIn file included from ../drivers/video/fbdev/maxinefb.c:34:\n../drivers/video/fbdev/maxinefb.c: In function \u0027maxinefb_ims332_write_register\u0027:\n../include/video/maxinefb.h:16:41: error: implicit declaration of function \u0027KSEG1ADDR\u0027; did you mean \u0027CKSEG1ADDR\u0027? [-Wimplicit-function-declaration]\n 16 | #define MAXINEFB_IMS332_ADDRESS KSEG1ADDR(0x1c140000)\n../drivers/video/fbdev/maxinefb.c:66:49: note: in expansion of macro \u0027MAXINEFB_IMS332_ADDRESS\u0027\n 66 | register unsigned char *regs \u003d (char *) MAXINEFB_IMS332_ADDRESS;\n../drivers/video/fbdev/maxinefb.c:66:40: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]\n 66 | register unsigned char *regs \u003d (char *) MAXINEFB_IMS332_ADDRESS;\n\nSigned-off-by: Randy Dunlap \u003crdunlap@infradead.org\u003e\nAcked-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "47eb05e731fbef11dc36e9a9a9a05e61d43480ae", "tree": "006e6afd2206e9f18cb2eff0d8e06a4d7732f28c", "parents": [ "311f2c770d6786a1b835d544196c345b0f9d7767" ], "author": { "name": "Randy Dunlap", "email": "rdunlap@infradead.org", "time": "Thu Aug 13 16:29:33 2026 -0700" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Thu Aug 20 23:45:06 2026 +0200" }, "message": "fbdev: maxine: elide an unused function\n\nmaxinefb_ims332_read_register() is not used, but since it describes a\nhardware interface, leave it in the source file as documentation\nand surround it inside an #if 0/#endif block.\n\n../drivers/video/fbdev/maxinefb.c:74:21: warning: \u0027maxinefb_ims332_read_register\u0027 defined but not used [-Wunused-function]\n 74 | static unsigned int maxinefb_ims332_read_register(int regno)\n\nSigned-off-by: Randy Dunlap \u003crdunlap@infradead.org\u003e\nAcked-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "311f2c770d6786a1b835d544196c345b0f9d7767", "tree": "7854e58c2d9b8eea749e1760385ad37c112a528b", "parents": [ "27c97be7f3c6b62cf872502374fb77d973950648" ], "author": { "name": "Randy Dunlap", "email": "rdunlap@infradead.org", "time": "Thu Aug 13 16:29:32 2026 -0700" }, "committer": { "name": "Helge Deller", "email": "deller@gmx.de", "time": "Thu Aug 20 23:45:06 2026 +0200" }, "message": "fbdev: maxine: make functions static\n\nMake 3 functions static so that they don\u0027t need to be declared in a\nheader file.\n\n../drivers/video/fbdev/maxinefb.c:64:6: warning: no previous prototype for \u0027maxinefb_ims332_write_register\u0027 [-Wmissing-prototypes]\n 64 | void maxinefb_ims332_write_register(int regno, register unsigned int val)\n../drivers/video/fbdev/maxinefb.c:74:14: warning: no previous prototype for \u0027maxinefb_ims332_read_register\u0027 [-Wmissing-prototypes]\n 74 | unsigned int maxinefb_ims332_read_register(int regno)\n../drivers/video/fbdev/maxinefb.c:114:12: warning: no previous prototype for \u0027maxinefb_init\u0027 [-Wmissing-prototypes]\n 114 | int __init maxinefb_init(void)\n\nSigned-off-by: Randy Dunlap \u003crdunlap@infradead.org\u003e\nAcked-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nSigned-off-by: Helge Deller \u003cdeller@gmx.de\u003e\n" }, { "commit": "55ab7e14222e5f0b0fd9f7711ca391d2924b35e3", "tree": "5ad69fb3647de784748f8cfa93fddaa0dc70ea22", "parents": [ "f5437ff7299e47e76e52d37a2937a4b0f04e399f", "326d49039c10b65522ac7277b19b9b5c42ed1aeb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 13:55:16 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 13:55:16 2026 -0700" }, "message": "Merge tag \u0027for-7.3/block-20260819\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux\n\nPull block updates from Jens Axboe:\n\n - NVMe updates via Keith:\n - Enable Clang context analysis for the nvme host driver, adding\n context annotations across core, fabrics, rdma, tcp and pci\n - nvmet reservation state exposed through a new namespace-level\n debugfs directory, plus ABI documentation for the host sysfs and\n target configfs interfaces\n - nvme-tcp host memory disclosure fixes on the read path: reject a\n read that transferred too few bytes, don\u0027t accept C2HData based\n on blk_rq_payload_bytes() alone, and fix the R2T case for a read\n command\n - Parallelize nvme-rdma I/O queue allocation and startup (Surabhi)\n - Apple nvme fixes and quirks: page aligned admin queue buffers,\n destroy the admin queue on removal, and various DMA/NVMMU\n correctness fixes\n - A large pile of nvmet and host fixes for out-of-bounds reads,\n refcount/resource leaks, and NULL derefs across auth, zns,\n passthru, pci-epf, rdma and configfs\n - Various other fixes and cleanups\n\n - MD updates via Yu Kuai:\n - llbitmap reshape support, the large series wiring exact bitmap\n mapping and reshape lifecycle through raid5 and raid10, growing\n the page cache in place, and remapping checkpointed bits as\n reshape progresses\n - raid5 fixes for lockless max_nr_stripes and recovery_offset\n accesses, a reshape deadlock with more failed devices than max\n degraded, and bitmap batch counter consistency\n - Atomic write handling for raid1/raid10, and removal of the\n REQ_NOWAIT support from raid1/10/456\n - raid5-ppl use-after-free fix in ppl_do_flush()\n - A batch of smaller fixes across md core and the bitmap code\n\n - s390/dasd ESE full-track write support and the surrounding\n infrastructure, plus enabling CONTEXT_ANALYSIS for s390/block\n\n - RWF_DONTCACHE support for block devices, built on new task-context\n bio completion infrastructure, and wiring it up for the iomap and\n buffer dropbehind writeback paths\n\n - Async io_uring zone reset all, plus zone management command cleanups\n allowing REQ_NOWAIT and tightening conventional zone rejection\n\n - Block integrity refactoring: lift BIP_CHECK_FLAGS to the shared\n header, handle nogenerate/noverify properly in fs-integrity, and drop\n the blk-integrity.h include from bdev.c\n\n - Split out a new blk_plug.h header\n\n - ublk improvements: add UBLK_F_IO_DESC_SIZE, split request validation\n from io_desc init, reject non-power-of-2 zone sizes in SET_PARAMS,\n and a series of hardening fixes around map/unmap and auto buf reg\n\n - null_blk cleanups and configfs serialization fixes\n\n - nbd queue freeze removal on the setup paths, and a new\n pre_defined_connections module parameter for pre-created devices\n\n - blk-cgroup fixes for the race between policy activation and blkg\n destruction, and accounting per-cpu stats over possible CPUs across\n blk-stat, iolatency, iocost and kyber\n\n - Various dio fixes: leak on metadata mapping error, validate user\n space vectors during extraction, and set dma_alignment from the\n backing file for loop and zloop direct I/O\n\n - bio cleanups\n\n - Various other fixes and cleanups all over\n\n* tag \u0027for-7.3/block-20260819\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux: (241 commits)\n nbd: add pre_defined_connections module parameter for pre-created devices\n nbd: remove queue freeze for newly created nbd from netlink path\n nbd: factor out a nbd_genl_foreach_sock\n nbd: skip queue freeze when setting size at device startup\n nbd: remove queue freeze in nbd_add_socket\n nbd: clear queue limits on disconnect\n nbd: disallow NBD_SET_SOCK on an active device\n nbd: simplify find_fallback() by removing redundant logic\n blk-mq: add missing call to srcu_barrier() in blk_mq_free_tag_set()\n block: mtip32xx: synchronize ioctls with device removal\n ublk: avoid teardown retry loop on xarray allocation failure\n null_blk: fix UBSAN shift-out-of-bounds when zone_size is 0 or overflows\n block: don\u0027t include blk-integrity.h in bdev.c\n xfs: avoid double deferrals for RWF_DONTCACHE writes\n loop: Fix recently introduced lock inversion\n block: set QUEUE_FLAG_DYING unconditionally in blk_mark_disk_dead()\n swim3: Add missing MODULE_DESCRIPTION\n selftests: ublk: add SET_PARAMS validation test\n selftests: ublk: add helper for SET_PARAMS\n ublk: reject non-power-of-2 zone sizes in SET_PARAMS\n ...\n" }, { "commit": "f5437ff7299e47e76e52d37a2937a4b0f04e399f", "tree": "29c5fec2e09c24835ede4c62a5a5edd4113d27f8", "parents": [ "c84d3e3130dfe1058cb27dc78e7ad8bd36f0545a", "881dc9dd66d2367e7fe1be0f953a8837cb3b26a9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 13:47:56 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 13:47:56 2026 -0700" }, "message": "Merge tag \u0027for-7.3/io_uring-20260819\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux\n\nPull io_uring update from Jens Axboe:\n \"On top of the usual cleanups and fixes, the bigger items in here are:\n\n - zcrx work, most of it centered around adding dynamic area\n provisioning, plus a bunch of prep and cleanups leading up to it:\n scale refilling with large pages, coalesce same-niov RQEs on\n refill, separate the RQ head/tail cache lines and cache the RQ\n tail, and rework the area creation locking.\n\n - Fix the futex inflight accounting so that only private futex waits\n are marked inflight, and don\u0027t mark wake requests as inflight at\n all.\n\n - Drop the custom iov copy in the buffer select prep and msg header\n copy paths, using the generic helpers instead.\n\n - Fix a folio size overflow in io_vec_fill_bvec(), and account the\n pages a compound region really uses in the memmap path.\n\n - Fix an iovec leak in uring_cmd when the async cmd isn\u0027t recycled,\n skip the blocking task work for io_uring_cmd_issue_blocking(), and\n don\u0027t skip completion for a synchronous multishot cmd\n\n - Defer eventfd signaling when queued from a wakeup handler\n\n - Fix io-wq worker accounting when canceling creation callbacks\n\n - Annotate remote tasks for kcoverage\"\n\n* tag \u0027for-7.3/io_uring-20260819\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux: (30 commits)\n io_uring: Add missing include for ITER_SOURCE and ITER_DEST\n io_uring/uring_cmd: don\u0027t skip completion for a synchronous multishot cmd\n io_uring/memmap: account the pages a compound region really uses\n io_uring/zcrx: add dynamic area provisioning\n io_uring/zcrx: lock area creation with pp_lock\n io_uring/zcrx: keep array of areas\n io_uring/zcrx: move freelist lock to struct zcrx\n io_uring/zcrx: unmap under netdev lock\n io_uring/zcrx: split dmabuf unmap and release\n io_uring/zcrx: don\u0027t pass ifq_reg to area creation\n io_uring/zcrx: add helper for deriving area token\n io_uring/zcrx: don\u0027t reload skb_shinfo\n io_urint/zcrx: narrow var scope in io_zcrx_recv_skb()\n io_uring/zcrx: constify area_reg on import\n io_uring/zcrx: coalesce same-niov RQEs on refill\n io_uring/zcrx: cache RQ tail\n io_uring/zcrx: add RQ iterator\n io_uring/zcrx: move RQ head/tail to separate cache lines\n io_uring/zcrx: scale refilling with large pages\n io_uring/io-wq: fix worker accounting when canceling creation callbacks\n ...\n" }, { "commit": "c84d3e3130dfe1058cb27dc78e7ad8bd36f0545a", "tree": "0adeaef5227fcaf5b95e8831b5f33af00cec50ae", "parents": [ "50c44fea13ec339d0d457079b254e8c8420d6511", "9091c97be34083587a75db174aab51551d8e8543" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 13:17:11 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 13:17:11 2026 -0700" }, "message": "Merge tag \u0027ext4_for_linus-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4\n\nPull ext4 updates from Ted Ts\u0027o:\n\n - Improve performance by allowing parallel DIO writes when we were\n previously being overly conservative when checking whether it was\n safe to avoid requiring an exclusive lock\n\n - Improve the performance of ext4_mb_prefetch() used by fallocate() by\n avoiding work when it is not needed\n\n - Remove the unnecessary custom end_io function\n ext4_end_buffer_io_sync()\n\n - Improve performance when performing an overwrite to an already\n uptodate folio\n\n - Clean up how we handle deallocating EA inodes to avoid a potential\n lock ordering issue when there is a failed mount while an EA inode is\n still being evicted\n\n - Use str_plural() instead of a custom macro\n\n - Avoid soft lockups or RCU stalls if there are many busy buffers\n (caused by heavy I/O) while checkpointing\n\n - Use scoped NOFS when starting a handle in nojournal mode\n\n - Align fields in handle structure to optimize setting and getting the\n h_type and h_line_no fields\n\n - Fix documentation of the meta_bg block group layout\n\n - Bug fixes:\n - Fix a potential out-of-bounds read in ext4_read_inline_dir()\n - Fix a potential deadlock when concurrent xattr operations are\n racing with each other when some of the xattrs are using the\n ea_inode feature\n - Fix a spurious warning with data\u003djournal that can be triggered\n when writeback races with remounting the file system read-only\n - Fix a potential deadlock when EXT4_IOC_MIGRATE races with a file\n system freeze operation\n - Make sure all in-flight direct I/O operations are complete before\n falling back to buffered I/O\n - Handle IOCB_NOWAIT properly when performing a extending DAX write\n - Prevent potentially sleeping on a block allocation when\n IOCB_NOWAIT is set\n - Fix potential races when racing an inline data write with a page\n fault\n - Propagate errors when adding or removing extent ranges during a\n fast commit replay\n - Avoid trying to expand an inode\u0027s extra size when it is being\n evicted to avoid a number of corner case or deadlocks\n - Avoid spurious error when retrying inode extra size expansion\n - Fix corner cases where we underestimate the number of journal\n credits needed\n - Avoid hangs/crashes/WARNINGS caused by maliciously corrupted file\n systems\n - Don\u0027t issue spurious orphan clean message on RO file systems\n - Avoid leaving the file system in an inconsistent state after a\n crash when a WRITE_ZEROS in progress converting an unwritten\n extent to a written extent\n - Handle WRITE_ZEROS correctly when there are some partially dirtied\n regions in the page cache\n - Pass errors during zero-rage, truncate, or punch hole to the\n caller if ext4_get_block() fails\n - Wait for writeback to finish when triggered by zero-range or\n zero-range for those devices that require stable writes\n - If the reserved gid superblock field is set, set the reserved gid\n instead of the reserved uid\n\n* tag \u0027ext4_for_linus-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (56 commits)\n ext4: fix estimate extent index blocks in ext4_ext_index_trans_blocks()\n ext4: fix transaction overflow during writeback\n ext4: teach ext4_meta_trans_blocks() about number of allocated extents\n ext4: guard against NULL s_group_info in ext4_get_group_info\n ext4: fix spurious message about orphan cleanup on RO fs\n ext4: stop retrying saturated xattr cache entries\n ext4: don\u0027t enable DAX on new encrypted files\n ext4: protect WRITE_ZEROES written extents with orphan list\n ext4: export converted block count from ext4_convert_unwritten_extents()\n ext4: fix incorrect function call when initializing s_resgid\n ext4: validate EA inode i_nlink in ext4_xattr_inode_iget\n jbd2: align h_type and h_line_no in the handle structure on byte boundaries\n ext4: enable scoped NOFS when starting a handle in nojournal mode\n ext4: write back partial-zeroed edges in WRITE_ZEROES\n ext4: zero out whole block for clean edges in WRITE_ZEROES\n ext4: track partial-zero outcome per edge in ext4_zero_partial_blocks()\n ext4: clarify return semantics of ext4_load_tail_bh()\n ext4: move partial block zeroing earlier in ext4_zero_range()\n ext4: check return value of ext4_get_block() in ext4_load_tail_bh()\n ext4: skip tail block zeroing for inline data files\n ...\n" }, { "commit": "50c44fea13ec339d0d457079b254e8c8420d6511", "tree": "f5b6454b2b3dd3a4da37f7dab6f07ae363e28fa5", "parents": [ "9328b3b03bdce05f660dbf33a4183716a64e304f", "4096e2a06f43fae209d6472d395f3219e02ab2e6" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 13:00:40 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 13:00:40 2026 -0700" }, "message": "Merge tag \u0027for-7.3-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux\n\nPull btrfs updates from David Sterba:\n \"This is the summer edition of btrfs changes, smaller than usual. Yet,\n there are performance improvements in various areas or for specific\n workloads and some notable changes like removing space cache v1 code\n or mount option reduction.\n\n User visible changes:\n\n - free space v1 disabled by default; the v2 (free space tree) is mkfs\n default since 5.15, filesystems with v1 still work but could be\n slightly slower due to lack of block group caching\n\n - mount option \u0027rescue\u003dusebackuproot\u0027 requires read-only mount, it\u0027s\n too risky to allow writable mount\n\n - remove standalone mount option \u0027usebackuproot\u0027, deprecated in 5.9\n\n - preserve constraints of NODATASUM and NODATACOW when chattr and\n mount options may change the attributes\n\n - remove arbitrary limitation of 4KiB for page size when allowing\n block sizes smaller than page\n\n - print messages when pinned block groups affect swap activation\n\n Performance improvements:\n\n - use iomap bounce buffer for direct io instead of a fall back to\n buffered io; past correctness vs speed trade-offs dropped\n performance to ~50% of theoretical maximum, now it\u0027s ~95%,\n effectively doubled\n\n - replace xarray with local LRU list for tracking inhibited\n extent buffers, restored performance to pre-inhibition state\n (relatively ~3x)\n\n - remove unnecessary 1 jiffy delay in \"non-SSD\" mode with multiple\n logging tasks, decrease latency, throughput increased ~5x on sample\n workload\n\n - skip hole detection during full fsync for files without holes\n and lots of extents, reduce run time ~5x on sample workload\n (microsecond ranges)\n\n - reduce locking around extent readahead so it does not slow down\n other tasks using an overlapping range\n\n - enhance extent buffer allocation modes to allow NOWAIT semantics\n in some cases\n\n Notable fixes:\n\n - write-protect folios during writeback, prevent concurrent mmap\n and compress/checksumming/etc undesired interactions\n\n - in zoned mode, handle transient overcommit full instead of going\n read-only\n\n - fix possible deadlock between defragmentation and delayed\n allocation reservations\n\n - handle remaining iputs at umount time\n\n - fix lockdep warning between device scan locking and log mutex\n\n - add workaround for degenerate RAID56 device count modes (2 and 3)\n not supported by the parity calculation library\n\n - restore check that subvolume is not read-only when changing ACLs\n\n - retry reading verity data colliding with up-to-date status changes\n\n Core:\n\n - simplify raid56 stripe handling by using contiguous virtual\n allocations\n\n - in zoned mode, fix various metadata write issues in writeback or\n unmount\n\n - space reservation fixes\n\n - remove unused data structure members\n\n - more auto-freeing conversions\n\n - error pointer values are printed using %pe format\"\n\n* tag \u0027for-7.3-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (72 commits)\n btrfs: skip hole detection during full fsync for files without holes\n btrfs: add extra ASSERT()s to make sure the folio size is correct\n btrfs: use GFP_NOWAIT for tree block readahead\n btrfs: enable unlocked NOFAIL retry for eb allocations\n btrfs: add struct btrfs_eb_prealloc\n btrfs: factor init_extent_buffer from __alloc_extent_buffer\n btrfs: qgroup: fix a wrong length calculation in qgroup_free_reserved_data()\n btrfs: add validation for extent states\n btrfs: use aligned range for locking in reflink\n btrfs: use aligned range for locking in extent_fiemap()\n btrfs: zoned: don\u0027t clobber the extent buffer when zeroing it out\n btrfs: zoned: drop stranded dirty metadata buffers at unmount\n btrfs: zoned: drop stranded dirty metadata on transaction abort\n btrfs: zoned: flush active metadata block group at btree_writepages() start\n btrfs: convert reflink.c to use btrfs_inode as parameters\n btrfs: use simple booleans for log_commit field in struct btrfs_root\n btrfs: check for exit condition after waking in wait_log_commit()\n btrfs: move condition for log commit wait into wait_log_commit()\n btrfs: remove log batch counter use for fsync\n btrfs: stop sleeping for one jiffy in non-ssd mounts during log commit\n ...\n" }, { "commit": "9328b3b03bdce05f660dbf33a4183716a64e304f", "tree": "9c89bb513fbbf0108c9194f9dd258f4935e14791", "parents": [ "81e51378b78a11b168ffec08d80ff4b47e5b0227", "62333e480d12ab186f89fe2725b372d12f72d5eb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:52:35 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:52:35 2026 -0700" }, "message": "Merge tag \u0027fs_for_v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs\n\nPull ext2, udf, isofs, and quota updates from Jan Kara:\n\n - Remove deprecated quota code printing warnings about exceeded quota\n directly to console\n\n - Various udf \u0026 isofs hardening for handling of corrupted filesystems\n\n - Fix a possible data loss in udf when converting files from inline to\n out-of-line format\n\n - Simplify EIO error handling in ext2 xattr code\n\n* tag \u0027fs_for_v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:\n udf: Fix data loss when converting inline inodes to out of line\n udf: Move udf_map_block() up\n ext2: Simplify error handling of IO error when adding xattr\n isofs: Drop support of directory entries straddling blocks\n isofs: validate directory records consistently\n quota: remove CONFIG_PRINT_QUOTA_WARNING code\n udf: Fix i_lenExtents truncation on 32-bit kernels\n isofs: release zisofs block pointer buffer head\n udf: Fix bh leak for unallocated space entries\n udf: bound lengthAllocDescs from unallocated space entry\n UDF symlink pathComponent header OOB read\n isofs: fix out-of-bounds page array access on empty zisofs block\n udf: reject VAT indexes equal to the entry count\n udf: Mark LVID buffer as uptodate before marking it dirty\n udf: avoid recursive s_alloc_mutex deadlock when freeing AED blocks\n udf: validate extent partition references in udf_current_aext()\n" }, { "commit": "81e51378b78a11b168ffec08d80ff4b47e5b0227", "tree": "358bd6594d453ff9fb816f290cb2ce6383195220", "parents": [ "d141ec2825b4d3ec52f27c43bdd864090159273a", "e422777fdd4746de1109575c51e65038d4c5c1be" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:47:41 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:47:41 2026 -0700" }, "message": "Merge tag \u0027fsnotify_for_v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs\n\nPull fsnotify updates from Jan Kara:\n \"A couple of assorted fixes (mostly stuff spotted by Sashiko) for\n fsnotify subsystem.\n\n I\u0027m also removing Matt as a reviewer because he was not active in\n fsnotify in last years and after he stopped working for Google I don\u0027t\n have a working contact to him\"\n\n* tag \u0027fsnotify_for_v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:\n fsnotify: Fix stale object mask after concurrent mark updates\n fanotify: report full event length for FIONREAD\n fanotify: fix use-after-free of file range info\n fanotify: stop permission watchdog when timeout is zero\n fsnotify: Remove Matt Bobrowski as a reviewer\n fanotify: initialize permission event watchdog state\n" }, { "commit": "d141ec2825b4d3ec52f27c43bdd864090159273a", "tree": "1a51bb2c3692f0ab5f609c1a9e64374bcbb30490", "parents": [ "ada9ccfb81eca6943c3cd42d23b9ddd446e75342", "46db3c8a1be96a354758b44b1d4fbb4b70d09a20" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:39:56 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:39:56 2026 -0700" }, "message": "Merge tag \u0027nfsd-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux\n\nPull nfsd updates from Chuck Lever:\n\n - CB_NOTIFY support for NFSD\u0027s NFSv4.1 directory delegations\n\n The server used to recall a delegation as soon as the directory\n changed. NFSD now watches delegated directories through fsnotify and\n reports adds, removes, renames, and directory attribute changes,\n carrying the filehandle and attributes of the affected entry, so\n clients can keep their caches. Some of the NOTIFY4 flags come from\n RFC 8881bis (Jeff Layton)\n\n - Continued netlink work\n\n A new server-stats-get operation reports what /proc/net/rpc/nfsd\n publishes, plus NFSv4 callback counts, and SUNRPC now keeps its\n per-procedure call counts per network namespace, so a container sees\n its own numbers. nfsstat reads all of this over netlink, with a\n procfs fallback for older kernels (Jeff Layton)\n\n - Remove SUNRPC service thread pool mode selection\n\n Per node is the right choice on any host we run today, so the auto,\n global, and percpu modes have been removed. A single-node host still\n gets one pool. A multi-NUMA host now gets a pool per node.\n\n sunrpc.pool_mode accepts the old names but no longer selects\n anything.\n\n - Bug fixes, clean-ups, and small optimizations:\n - async COPY offload rework (Jeff Layton)\n - more use-after-free fixes in the NFSv4 state revocation paths\n - percpu counter contention removed from the reply cache and IO\n accounting\n - a long list of hardening fixes (Chris Mason)\n\nSincere thanks to all contributors, reviewers, testers, and bug\nreporters who participated in the v7.3 NFSD development cycle.\n\n* tag \u0027nfsd-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux: (182 commits)\n nfsd: export NFSv4 callback op stats via netlink\n nfsd: count NFSv4 callback operations per netns\n sunrpc: remove unused svc_version vs_count field\n nfsd: implement server-stats-get netlink handler\n sunrpc: use per-net counts in svc_seq_show()\n sunrpc: add per-netns per-procedure call counts to svc_stat\n NFSD: Document reply_cache_stats ABI\n NFSD: Eliminate percpu counter contention in IO byte accounting\n NFSD: Eliminate percpu counter contention in reply cache statistics\n NFSD: Eliminate percpu counter contention in DRC memory accounting\n NFSD: Fix off-by-one in DRC bucket pruning limit\n NFSD: Relocate NFSv4 \"supported attributes\" to new header\n NFSD: Relocate nfsd4_set_netaddr()\n NFSD: Relocate nfsd_user_namespace()\n NFSD: Move struct readdir_cd\n NFSD: Move the export.h include from nfsd.h to auth.c\n NFSD: Remove \u0027#include \"nfsd.h\"\u0027 from fs/nfsd/cache.h\n NFSD: include \"netns.h\"\n NFSD: Explicitly include \"stats.h\"\n NFSD: Make \"stats.h\" self-contained\n ...\n" }, { "commit": "ada9ccfb81eca6943c3cd42d23b9ddd446e75342", "tree": "963550660d97dcc84378b6682ec1b367fd86fd67", "parents": [ "dc83d18cdd90482c70fa4320160bba70ec5c9ef8", "d410cd5303ec59c7cf23dd61423752ce8e9ecb59" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:36:26 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:36:26 2026 -0700" }, "message": "Merge tag \u0027for-linus-7.3-ofs1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/hubcap/linux\n\nPull orangefs updates from Mike Marshall:\n \"Fixes:\n - fix double-free of trailer_buf\n - skip leading spaces before parsing client debug masks\n\n Cleanup:\n - Remove commented out code\n\n New:\n - use folio_pos() and folio_size() in orangefs_page_mkwrite()\"\n\n* tag \u0027for-linus-7.3-ofs1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/hubcap/linux:\n orangefs: skip leading spaces before parsing client debug masks\n orangefs: Remove commented out code in find_cached_xattr\n orangefs: use folio_pos() and folio_size() in orangefs_page_mkwrite()\n orangefs: fix double-free of trailer_buf on readdir copy failure\n" }, { "commit": "dc83d18cdd90482c70fa4320160bba70ec5c9ef8", "tree": "e50648d8d52da494ba1a90c8876e0d6389b0d914", "parents": [ "8f421dfec2347c67f5f8c261fbfdee407ac5e20f", "c22f91d82cb9a29d22bdffdce6c803467984ad0c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:27:02 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:27:02 2026 -0700" }, "message": "Merge tag \u0027ntfs3_for_7.3\u0027 of https://github.com/Paragon-Software-Group/linux-ntfs3\n\nPull ntfs3 updates from Konstantin Komarov:\n \"Added:\n load ATTR_BITMAP run extents from $MFT extension records\n initialize err in attr_wof_frame_info\n reserve NUL byte when converting UTF-16 names\n reject restart table growth beyond U16_MAX entries\n validate dirty page table on log replay\n basic support for alternative data streams\n validate ef-\u003esize covers the record\u0027s name and value\n\n Fixed:\n slab-out-of-bounds write in ni_create_attr_list()\n out-of-bounds read of INDEX_ROOT in reparse/objid init\n boundary check in ntfs_dir_count()\n info-leak in ntfs_rename()\n lseek EINVAL on sparse/compressed files with 64-bit clusters\n info-leak on partial LZNT decompress in ni_read_frame()\n bound page_lcns[] index by the log record\n memory leak in indx_find_sort()\n integer overflow in MFT cluster validation\n reject out-of-range evcn in mi_enum_attr()\n out-of-bounds read in read_log_rec_buf()\n\n Changed;\n widen inode/record number storage to u64\n cosmetic fixes and improvements\n rename \u0027err\u0027 to \u0027ret\u0027 in read paths\"\n\n* tag \u0027ntfs3_for_7.3\u0027 of https://github.com/Paragon-Software-Group/linux-ntfs3: (21 commits)\n fs/ntfs3: validate ef-\u003esize covers the record\u0027s name and value\n fs/ntfs3: fix out-of-bounds read in read_log_rec_buf()\n fs/ntfs3: reject out-of-range evcn in mi_enum_attr()\n fs/ntfs3: fix integer overflow in MFT cluster validation\n fs/ntfs3: Add basic support for alternative data streams\n fs/ntfs3: Rename \u0027err\u0027 to \u0027ret\u0027 in read paths\n fs/ntfs3: Fix memory leak in indx_find_sort()\n fs/ntfs3: bound page_lcns[] index by the log record\n fs/ntfs3: validate dirty page table on log replay\n fs/ntfs3: reject restart table growth beyond U16_MAX entries\n fs/ntfs3: fix info-leak on partial LZNT decompress in ni_read_frame()\n fs/ntfs3: reserve NUL byte when converting UTF-16 names\n ntfs3: initialize err in attr_wof_frame_info\n fs/ntfs3: fix lseek EINVAL on sparse/compressed files with 64-bit clusters\n fs/ntfs3: load ATTR_BITMAP run extents from $MFT extension records\n ntfs3: fix info-leak in ntfs_rename()\n ntfs3: fix boundary check in ntfs_dir_count()\n fs/ntfs3: fix out-of-bounds read of INDEX_ROOT in reparse/objid init\n fs/ntfs3: fix slab-out-of-bounds write in ni_create_attr_list()\n fs/ntfs3: cosmetic fixes and improvements\n ...\n" }, { "commit": "8f421dfec2347c67f5f8c261fbfdee407ac5e20f", "tree": "fef9493a5f70c05a32af8775c02caad4d56d7606", "parents": [ "62abb6cffd6bab44d430627043778ec157c32cce", "988c3c186c730590a71df8adc1424c5578ef8bae" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:24:48 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:24:48 2026 -0700" }, "message": "Merge tag \u0027erofs-for-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs\n\nPull erofs updates from Gao Xiang:\n \"The most notable enhancement is to allow passing source fds via\n fsconfig() for composefs. The others are all various fixes:\n\n - Allow source fds via fsconfig(), in addition to source paths\n\n - Use dedicated metadata inodes for file-backed mounts\n\n - Disallow invalid interlaced ztailpacking pclusters\n\n - Validate on-disk compression algorithm IDs against supported ones\n\n - Fix unused pcluster pools on higher page-size platforms\"\n\n* tag \u0027erofs-for-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs:\n erofs: fix unused pcluster_pools for higher page sizes\n erofs: guard on-disk algorithm IDs against Z_EROFS_COMPRESSION_MAX\n erofs: fix interlaced ztailpacking pclusters\n erofs: use dedicated meta inodes for file-backed mounts\n erofs: accept source file descriptor via fsconfig\n erofs: fix typo in error messages\n" }, { "commit": "62abb6cffd6bab44d430627043778ec157c32cce", "tree": "cdfc069c86df7223ad48cb0d47ae3bc3f7d86b91", "parents": [ "cf62c7084acfa9f1ef98ce81b38c934c7ea10f93", "e99cb3ecd8334ca21e01ff9a79a916693f58f1fb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:17:36 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:17:36 2026 -0700" }, "message": "Merge tag \u0027libnvdimm-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm\n\nPull NVDIMM and DAX updates from Alison Schofield:\n \"Most are DAX preparatory patches for FAMFS support, along with a few\n NVDIMM fixes and documentation cleanups.\n\n - Documentation cleanup, removing kernel-doc warnings\n\n - preparing DAX for FAMFS\n\n - misc NVDIMM fixups with cleanups for issues reported by Coccinelle\"\n\n* tag \u0027libnvdimm-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:\n nvdimm-btt: clean up kernel-doc warnings\n libnvdimm: nd.h: clean up kernel-doc warnings\n dax: fsdev.c minor formatting cleanup\n dax: fix holder_ops race in fs_put_dax()\n dax: read holder_ops once in dax_holder_notify_failure()\n dax/fsdev: fail probe on invalid pgmap offset\n dax/fsdev: use __va(phys) for kaddr in direct_access\n dax/fsdev: clear pgmap ops and owner on unbind\n dax/fsdev: don\u0027t leave a dangling dev_dax-\u003epgmap on probe failure\n dax/fsdev: clear vmemmap_shift when binding static pgmap\n dax/fsdev: fix multi-range offset in memory_failure handler\n dax: fix misleading comment about share/index union in dax_folio_reset_order()\n nvdimm/btt: reject an arena whose nfree is below the lane count\n libnvdimm/labels: Bound the on-media label size before the shift\n libnvdimm/labels: Prevent integer overflow in __nd_label_validate()\n nvdimm: ndtest: remove redundant NULL check before vfree()\n nvdimm: nfit: remove redundant NULL check before vfree()\n" }, { "commit": "cf62c7084acfa9f1ef98ce81b38c934c7ea10f93", "tree": "db3440210585fd7eb6305a381730a9eac544622d", "parents": [ "ae814200e8393fa504dd246e98fcba8f5493de28", "ab11176bd3a76058ecafd066d0f9c718dc80f389" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:03:45 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 12:03:45 2026 -0700" }, "message": "Merge tag \u0027execve-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux\n\nPull execve update from Kees Cook:\n\n - x86/elf: Correct comment for STACK_RND_MASK() (Lorenzo Stoakes)\n\n* tag \u0027execve-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:\n x86/elf: Correct comment for STACK_RND_MASK()\n" }, { "commit": "ae814200e8393fa504dd246e98fcba8f5493de28", "tree": "9e49a70e51ba3be1d2c36ffc3cc9d391f1d0ad68", "parents": [ "216b3f432a36549767ca750e64badd71340b8c0f", "f4806cc63cc65bd752fd72d84937614dca6504ec" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 11:41:46 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 11:41:46 2026 -0700" }, "message": "Merge tag \u0027bitmap-for-7.3\u0027 of https://github.com/norov/linux\n\nPull bitmap updates from Yury Norov:\n \"The usual set of fixes, cleanups and performance improvements together\n with a couple of new tests:\n\n - bitmap_find_next_zero_area_off() optimization (Sunyi)\n\n - bitmap_find_next_zero_area_off(): return size when no zero area is\n found (Yury)\n\n - bitmap vs IDA vs Maple Tree performance test (Yury)\n\n - get rid of cpumap_print_to_pagebuf() (Yury)\n\n - use nr_node_ids in __nodemask_pr_numnodes() (Li RongQing)\n\n - bitops: make the *_bit_le functions use unsigned long (Benjamin)\n\n - bitmap scatter \u0026 gather test fix (Christophe)\n\n - use __ASSEMBLER__ in bitmap header files (Thomas)\"\n\n* tag \u0027bitmap-for-7.3\u0027 of https://github.com/norov/linux: (25 commits)\n lib: test bitmap vs IDA vs Maple Tree performance for region allocations\n bitmap: Return size when no zero area is found\n media: s5p-mfc: Treat bitmap size as allocation failure\n crypto: ccp: Treat bitmap size as allocation failure\n powerpc/msi: Treat bitmap size as allocation failure\n ARM: dma-mapping: Treat bitmap size as allocation failure\n bitmap: drop bitmap_next_set_region()\n nodemask: reduce bitmap width to nr_node_ids in __nodemask_pr_numnodes()\n bitmap: Properly initialise destination bitmap for scatter \u0026 gather test\n lib/bitmap-str: get rid of cpumap_print_to_pagebuf()\n perf: Use sysfs_emit() for cpumask show callbacks\n PCI/sysfs: Use sysfs_emit() for cpumask show callbacks\n RDMA/hfi1: Use sysfs_emit() for cpumask show helper\n hwtracing: hisi_ptt: Use sysfs_emit() for cpumask show\n fpga: dfl-fme-perf: Use sysfs_emit() for cpumask show\n devfreq: Use sysfs_emit() for cpumask show callbacks\n cpu: Use sysfs_emit() for cpumask show callback\n x86/events: Use sysfs_emit() for cpumask show callbacks\n powerpc: Use sysfs_emit() for cpumask show callbacks\n arm: Use sysfs_emit() for cpumask show callbacks\n ...\n" }, { "commit": "216b3f432a36549767ca750e64badd71340b8c0f", "tree": "b97cc8d35aad7be48845a4eaa3aec1dd3fc014f2", "parents": [ "11260c335ec6071af5543aef73000b28f041c124", "a8488ecbd7ba44d65b912dfe88a73f438eba2447" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 11:32:27 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 11:32:27 2026 -0700" }, "message": "Merge tag \u0027kcsan-20260817-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/melver/linux\n\nPull Kernel Concurrency Sanitizer (KCSAN) update from Marco Elver:\n\n - Avoid unintended access checking in NMIs\n\n* tag \u0027kcsan-20260817-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/melver/linux:\n kcsan: avoid unintended access checking in NMIs\n" }, { "commit": "11260c335ec6071af5543aef73000b28f041c124", "tree": "72eda88ab241d491ec3b6ec52f31b5f55a4324eb", "parents": [ "40d8c81577db09b71ee5402ba336b642d32d6a82", "fab183d632628381b466a41479489541ac0e29a0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 11:01:37 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 11:01:37 2026 -0700" }, "message": "Merge tag \u0027sched_ext-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext\n\nPull sched_ext updates from Tejun Heo:\n \"Most of this cycle completes the enqueue-path support for hierarchical\n sub-scheduling, which makes sub-scheduler support feature complete: a\n root BPF scheduler can now hand a cgroup subtree over to a nested\n sub-scheduler together with revocable CPU grants, and the\n sub-scheduler owns all scheduling decisions for its tasks on those\n CPUs.\n\n Development volume was high and a number of changes plugging holes in\n the new support landed late in the cycle. Also included are core\n scheduling fixes that were completed too late for the v7.2 release and\n are routed through this pull request.\n\n Sub-scheduler CPU delegation:\n\n - Parent schedulers now grant and revoke per-CPU capabilities\n (enqueueing, preemption, CPU frequency control) on their children,\n enforced on every path a scheduler can reach a CPU through.\n Previously only dispatching could be delegated; this lets\n sub-schedulers fully schedule their CPUs.\n\n - Rescue execution: a task whose scheduler doesn\u0027t have access to the\n CPUs the task needs to run on starved until the watchdog ejected\n the whole scheduler. The kernel now runs such tasks directly on a\n small bandwidth budget, turning a scheduler-killing failure into\n bounded degradation.\n\n - Cgroup integration: tasks migrating across a sub-scheduler boundary\n weren\u0027t re-homed to the new owner, causing wrong-scheduler\n scheduling and a use-after-free. Sub-schedulers now take over their\n cgroup subtree and receive its cgroup callbacks.\n\n - Arena objects now cross the kernel/BPF boundary as typed pointer\n arguments, translated transparently by the BPF tree\u0027s new arena\n argument support, replacing untyped arguments with manual\n translation.\n\n - scx_qmap now demonstrates full hierarchical sub-scheduling.\n\n Other fixes and updates:\n\n - Robustness improvements: the abort path is now NMI-safe, fixing\n deadlocks when errors are raised from NMI context and making\n hardlockup recovery direct. Reenqueue loops that could monopolize a\n CPU ahead of the watchdog now eject the offending scheduler, and\n stalls are blamed on the scheduler actually responsible.\n\n - Hardening: BPF-writable arena memory is validated before kernel\n use, and task slice and vtime writes got explicit synchronization\n rules, closing corruption vectors open to buggy or malicious\n schedulers.\n\n - Core scheduling: sched_ext dispatching can drop the rq lock inside\n the core-wide pick, which let interleaving selections corrupt each\n other\u0027s state and hard-hang the machine. The selection now restarts\n when the lock was released. The task ordering callback was also\n invoked with its arguments swapped, and the default ordering is\n updated to work across sub-scheduler boundaries. The fixes are\n marked for stable.\n\n - Other fixes headed for stable: a task init leak on fork failure\n during enable, tooling compat macros that silently failed to detect\n newer kernels, and a crash on reenqueueing against a destroyed\n dispatch queue.\n\n - Tooling: scx_pair moves off deprecated callbacks, and the\n deprecated scx_bpf_cpu_rq() kfunc is removed\"\n\n* tag \u0027sched_ext-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext: (144 commits)\n sched_ext: Drop the dead SCX_DEQ_CORE_SCHED_EXEC test in dequeue_task_scx()\n sched_ext: Make core-sched task ordering hierarchy-aware\n sched_ext: Use runnable_at for the default core-sched task ordering\n sched_ext: Fix inverted ops.core_sched_before() invocation\n sched_ext: Move the config-off sub-cap kfunc stubs into sub.c\n sched_ext: Rename balance-era identifiers to dispatch terms\n sched_ext: Drop the stale keep_prev fixup in dispatch_pick()\n sched_ext: Keep kick_sync waiting on the rq\u0027s own CPU\n sched_ext: Make SCHED_CLASS_EXT select GENERIC_ALLOCATOR\n sched_ext/scx_flatcg: Fix cvtime true-up on slice expiry\n sched_ext: Don\u0027t BUG_ON a destroyed DSQ in process_deferred_reenq_users\n sched_ext: Fix scx_bpf_dsq_move_to_local___v2 compat detection\n sched_ext: Make scx_bpf_events() read the calling scheduler\u0027s counters\n sched_ext: Drop unlocked scx_rq_clock_invalidate() from scx_root_disable()\n selftests/sched_ext: Fix flaky ddsp failure tests on busy systems\n selftests/sched_ext: Make numa idle validation race-free\n sched_ext: Fix scx_bpf_dsq_reenq___compat kfunc extern prototype\n sched_ext/scx_flatcg: expire cached hweights on weight changes\n sched_ext: Fix exit_task leak on fork failure during enable\n sched_ext: fix stale references in doc comments\n ...\n" }, { "commit": "40d8c81577db09b71ee5402ba336b642d32d6a82", "tree": "8bee29f278de583c3345b19b5e627fc522c45395", "parents": [ "39e34e88ec0dc7dbe3668dd54a0a9346a8323a8f", "2d19207f3fc8e08bab3270af2e3835358ab1473a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 10:37:42 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 10:37:42 2026 -0700" }, "message": "Merge tag \u0027cgroup-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup\n\nPull cgroup updates from Tejun Heo:\n\n - Attach path bug fixes: migrations spanning multiple source or\n destination cpusets were mishandled, most visibly leaving thread\n affinities stale when the controller is disabled in a threaded\n subtree. Configuration writes could also race an in-flight attach and\n apply stale state, and the deadline task count could get corrupted by\n concurrent updates, skewing SCHED_DEADLINE admission decisions.\n\n - Memory binding bug fixes: which node masks get applied differed\n between the binding update paths, and tasks cloned with\n CLONE_INTO_CGROUP skipped rebinding entirely. Rebinding also now runs\n once per process instead of repeating for every thread sharing the\n mm.\n\n - Overhead removals with no behavior change: CPU hotplug iterated tasks\n of cpusets that just inherit the parent\u0027s effective masks, and the\n slab-spreading task flag was still being maintained although the SLAB\n allocator that consumed it is long gone.\n\n - Data-race annotations for benign races so that KCSAN reports stay\n meaningful, selftest coverage for the fixes above along with\n flakiness and portability fixes, and documentation corrections.\n\n* tag \u0027cgroup-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup: (34 commits)\n selftests/cgroup: Remove redundant chown in test_cgcore_lesser_ns_open\n selftests/cgroup: Preserve CPU hotplug write errors\n cgroup/cpuset: Add test for partition root invalidation returning wrong CPUs\n cgroup/cpuset: Remove obsolete PFA_SPREAD_SLAB task flag\n docs: cgroup-v2: fix stale \"io\" controller introduction\n selftests/cgroup: Avoid awk -e in cpuset tests\n cgroup/cpuset: Use WRITE_ONCE() for shared prs_err updates\n selftests/cgroup: add user_usec sanity check in test_cpucg_nice\n cgroup: drop unneeded semicolon\n docs: cgroup-v2: mark memory.pressure and io.pressure as read-write\n selftests/cgroup: Fix minor defects in test_cpuset\n Docs/admin-guide/cgroup-v2: fix delay_nsec unit in io.latency doc\n selftests/cgroup: Remove redundant cg_enter_current() call in test_core\n selftests/cgroup: Add test for cpuset affinity on controller disable\n cgroup/cpuset: Handle the special case of non-moving tasks in cpuset_can_attach()\n cgroup/cpuset: Support multiple destination cpusets for cpuset_*attach()\n selftests/cgroup: fix missing TAP output in test_hugetlb_memcg\n cgroup/cpuset: Support multiple source cpusets for cpuset_*attach()\n cgroup/cpuset: Move mpol_rebind_mm/cpuset_migrate_mm() calls inside cpuset_attach_task()\n cgroup/cpuset: Make attach_ctx.old_cs track task group leader\n ...\n" }, { "commit": "39e34e88ec0dc7dbe3668dd54a0a9346a8323a8f", "tree": "86a1ed3c6dc6ca731acf190d065e7f3b492e7ae4", "parents": [ "72fdff1416e280e2baaa3cca69574defb998437e", "20a80e7f6b71bd664c98e95589f0cbc68804d200" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 10:34:25 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 10:34:25 2026 -0700" }, "message": "Merge tag \u0027wq-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq\n\nPull workqueue updates from Tejun Heo:\n\n - Worker wakeups moved out of pool-\u003elock on the hot paths, shortening\n lock hold times. The wakeup can be expensive on arm64 due to the idle\n wakeup IPI, and the change improves the workqueue microbenchmark\n there by up to 10%\n\n - Stall diagnostics now report pools stuck with no running worker, with\n a backtrace of what the CPU is executing and the likely culprit\n worker, instead of dumping every in-flight worker\n\n - Preparation for turning per-cpu workqueues into an affinity scope of\n unbound workqueues instead of a separate backend\n\n - Race annotations for KCSAN and sparse warnings, and doc and\n monitoring script fixes\n\n* tag \u0027wq-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq: (23 commits)\n workqueue: annotate racy p-\u003ewake_cpu accesses in kick_pool_pick()\n workqueue: BUG_ON() instead of returning NULL in wq_node_nr_active()\n workqueue: use RCU accessors when populating wq-\u003ecpu_pwq\n workqueue: use rcu_dereference_sched() in workqueue_congested()\n workqueue: skip the node_nr_active update for non-unbound workqueues\n workqueue: rename alloc_unbound_pwq() to alloc_pwq()\n workqueue: allocate attrs for all workqueues\n workqueue: rename wq-\u003eunbound_attrs to wq-\u003eattrs\n workqueue: test WQ_UNBOUND explicitly in the hotplug loops\n workqueue: account nr_active by the backing pool\n workqueue: release pwq pools by pool type\n workqueue: factor out alloc_and_link_percpu_pwqs()\n workqueue: factor out get_percpu_pool()\n docs: workqueue: Fix bracket\n workqueue: annotate racy sum_exec_runtime reads for CPU-intensive detection\n workqueue: annotate racy PWQ_STAT_CPU_TIME update in wq_worker_tick()\n workqueue: dump the last woken worker for stalled pools\n workqueue: trigger a single-CPU backtrace for stalled pools\n workqueue: only show running workers in stall diagnostics\n workqueue: defer the worker wakeup outside pool-\u003elock in process_one_work()\n ...\n" }, { "commit": "72fdff1416e280e2baaa3cca69574defb998437e", "tree": "23eb03cbd38f7b1642d78219b80b1452b9f597f6", "parents": [ "9d2ed026f031f764e9450ac9aada2f46bc977397", "791e420360669d55b7f90ca0c6d61b10e4992aec" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 10:31:50 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 10:31:50 2026 -0700" }, "message": "Merge tag \u0027docs-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/docs/linux\n\nPull documentation updates from Jonathan Corbet:\n \"It has been a not-too-busy cycle for docs; here\u0027s the highlights:\n\n - A (hopefully) consensus change to our LLM-attribution requirements,\n removing the specific model name from the Assisted-by tag\n\n - A couple of new realtime documents\n\n - Various docs-build-system fixes\n\n - Ongoing work with the Chinese, Portuguese, and Japanese\n translations\n\n ...and lots of typo fixes, grammar tweaks, etc\"\n\n* tag \u0027docs-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/docs/linux: (85 commits)\n Doc: admin-guide: pm: Remove unnecessary backticks and fix a spell\n Documentation: Extend the real-time hardware bits with some firmware bits\n docs: pt_BR: Reorganize process/index.rst to follow english structure\n docs: conf.py: fix the \u0027utf-8\u0027 typo\n doc tools: fix \u0027path\u0027 typos\n Documentation: real-time: Add kernel configuration guide\n docs: python: abi_regex: convert adjacent index placeholders\n docs: python: abi_regex: catch the right exception for a bad regex\n docs: sphinx-build-wrapper: include localversion in kernel version string\n Documentation: html: adjust sidebar section titles styling\n Documentation: html: show sections in the sidebar\n checkpatch.pl: adapt to new Assisted-by: format\n MAINTAINERS: update Traditional Chinese documentation maintainers\n docs: pt_BR: process: Translate CVE documentation\n docs: pt_BR: translate the management-style.rst to Brazilian Portuguese\n docs: xforms_lists: support DEFINE_IDTENTRY_IRQ()\n coding-assistants: simplify attribution\n docs: translations: pt_BR: translate email-clients.rst\n docs: pt_BR: process: Translate the security-bugs.rst\n doc:it_IT: align doc-guide translation\n ...\n" }, { "commit": "9d2ed026f031f764e9450ac9aada2f46bc977397", "tree": "62f07957703256ec389c3975269b0c7127afcf77", "parents": [ "91ec2035134982b98fab0609a9fd8480e8217dc1", "8d75c338f0bcecaa6c9af67f86c176b67b6acf3e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 08:46:41 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 08:46:41 2026 -0700" }, "message": "Merge tag \u0027sysctl-7.03-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/sysctl/sysctl\n\nPull sysctl updates from Joel Granados:\n\n - Fix kernel-doc warnings by adjusting in file documentation\n\n - Consolidate do_proc_* function into do_proc_vec\n\n Consolidate three slightly different implementations of applying a\n converter on all elements of a vector. Fixes to this function now\n propagate to the three types.\n\n - Replace CONFIG_PROC_SYSCTL with CONFIG_SYSCTL (they were the same)\n and restrict cad_pid modifications to global root (GLOBAL_ROOT_UID)\n\n* tag \u0027sysctl-7.03-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/sysctl/sysctl:\n sysctl: remove CONFIG_PROC_SYSCTL, it just mirrors CONFIG_SYSCTL\n sysctl: move the \"cad_pid\" entry from pid_table[] to kern_reboot_table[]\n sysctl: repair some kernel-doc comments\n sysctl: add Returns: kernel-doc for all functions\n sysctl: Update API function documentation\n sysctl: Rename proc_doulongvec_minmax_conv to proc_doulongvec_conv\n sysctl: Group proc_handler declarations and document\n sysctl: Replace do_proc_do{int,ulong,uint}vec with do_proc_vec\n sysctl: Add negp parameter to douintvec converter functions\n sysctl: Move default converter assignment out of do_proc_dointvec\n" }, { "commit": "91ec2035134982b98fab0609a9fd8480e8217dc1", "tree": "fed1207d0ad121a52ea1da4b4b8a2c52e3f2c15b", "parents": [ "5a8cd539ac19f7a68e68e1d25ef9ca2ff55b8500", "61eb236c41c2a4717015dff18016a75a5eb90052" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 08:16:04 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 08:16:04 2026 -0700" }, "message": "Merge tag \u0027net-next-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next\n\nPull networking updates from Jakub Kicinski:\n \"One of the \u0027small improvements all over the place\u0027 releases for us.\n\n It\u0027s hard to draw any direct comparisons because summer vacations\n disrupted our patch processing (and presumably - generation) quite a\n bit.\n\n Quick and dirty count suggests we (Paolo and I) merged a very similar\n number of net (632) and net-next (648) patches. This is not telling\n the full story either because 1/3 to 1/2 of the net-next patches also\n *seem* like AI-driven low priority fixes, cleanups and clarifications.\n\n We are completely overwhelmed, of course. The glimmer of hope is that\n we secured sufficient LLM budget and access (thank you Meta!) to run\n reviews with multiple frontier models on each patch. This eliminates\n some hallucinations. That said, in terms of review, the LLMs can only\n do so much.\n\n The sad truth is that our APIs (especially for rare events like PCIe\n errors, timeouts etc) have always been racy, and now LLMs don\u0027t let us\n ignore that. I expect our direction for the next release will be to\n tweak the reviews a little bit more, but start shifting focus to\n letting the LLMs take care of the busy work - managing patchwork,\n automating common process complaints, editing commit messages, and\n maybe applying patches which already got \"reviewed-by\" tags from\n people we trust...\n\n Core \u0026 protocols:\n\n - A few steps lowering rtnl_lock dependence:\n - per-netns netdev unregistration for select SW drivers (e.g.\n veth, ipvlan, tunnels)\n - rtnl_lock-less FIB rule changes (RTM_NEWRULE and RTM_DELRULE)\n - prepare software drivers and TC qdiscs for rtnl_lock-less GET\n\n - Support BIG TCP (\u003e64kB TSO) in UDP tunnels (vxlan, geneve)\n\n - Support buffers larger than PAGE_SIZE in devmem zero-copy API\n\n - Improve MPTCP handling of extreme memory pressure handling, when\n out-of-order queue had to be pruned\n\n - Report the per-group user count via RTM_GETMULTICAST\n\n - Expose the route deletion reason in RTM_DELROUTE\n\n - Add a SO_RIGHTS_NOTRUNC option to UNIX sockets to enable more\n useful handling of LSM denials when receiving SCM_RIGHTS messages:\n instead of truncating the message at the first blocked fd, keep\n every fd slot and store the LSM errno in the blocked slot\n\n - IPv6 Segment Routing - support looking up the post-encap SID\n (address) in a different/specified routing table\n\n - Support PRP RedBox (interlink) creation\n\n - Support per-nexthop UDP dst port in VXLAN\n\n - Continue converting getsockopt callbacks in a number of protocols\n to iov_iter\n\n Ethernet:\n\n - Merge initial CXL support for AMD/Solarflare NICs (shared branch\n with the CXL tree)\n\n - New drivers:\n - ADIN1140 10BASE-T1S MACPHY\n - Initial skeleton of Intel iXD and ZTE Dinghai drivers\n\n - High-speed NICs:\n - AMD/Pensando:\n - support firmware flashing\n - Cisco (enic):\n - SR-IOV V2 admin channel and MBOX protocol\n - Huawei (hns3):\n - support for ethtool pfc_prevention_tout\n - nVidia/Mellanox:\n - support sharing bandwidth control across interfaces\n of the same device\n - Marvell (octeontx2-pf):\n - link RQ page pools to netdev for Netlink stats\n - Google vNIC:\n - XDP metadata support for DQ RDA\n - Microsoft vNIC:\n - support forcing full-page RX buffers\n\n - Other NICs:\n - Synopsys IP:\n - eic7700: support for eth1\n - Microchip (lan743x):\n - support for RMII interface\n - Wangxun:\n - support for ethtool -G and -C for VFs\n - add Tx timeout and PCIe error handling\n - Intel (igb/igc):\n - RSS key get/set support\n - support for forcing link speed without auto-negotiation\n\n - Switches:\n - NXP (dpaa2):\n - support bonding/LAG offload\n - Mediatek:\n - mt7530: EN7528 support\n - initial support for MT7628\n - Micrel (ksz8/9):\n - refactoring work to move towards library model\n - PTP support for KSZ8463\n - nVidia/Mellanox:\n - support rtnl-lock-less ethtool callbacks\n - Realtek:\n - rtl8366rb: use generic RTL83xx code\n - support SGMII and HSGMII for RTL8367S\n\n - PHYs:\n - Airoha:\n - EcoNet EN7528 PHY support\n - DAPU Telecom\n - DAPU Telecom DAP8211R(I) Gigabit PHY support\n - Realtek:\n - support RTL8261C_CG\n - support RTL8261D\n\n Wireless:\n\n - nl80211: per-link statistics support for multi-link operation\n\n - mac80211: AQL/airtime-fairness support for multicast\n\n - Merge Peripheral Authentication Service (PAS) / TEE support for\n ath12k (shared branch with the firmware/qcom tree)\n\n - New drivers:\n - mm81x for Morse Micro Long-Range S1G devices\n - nxpwifi for NXP devices (mostly forked off from mwifiex)\n\n - Driver changes:\n - Broadcom (brcmfmac):\n - DPP support, some Cypress part update\n - MediaTek (mt76):\n - mt7928 support\n - mt7925 NAN support\n - mt7996 AP powersave improvements\n - Qualcomm (ath12k):\n - much kernel infrastructure integration work\n - AHB platform MultiPD support\n - Realtek (rt89):\n - LED support\n - RTL8922DE support\n - dual-BT coex for RTL8922D\n - Intel:\n - new FW version support\n\n Bluetooth:\n\n - HCI: add support for Shorter Connection Interval (SCI) feature\n\n - af_bluetooth: add minimal context analysis annotations\n\n - Driver changes:\n - Intel:\n - add Bluetooth SAR revision 2 support\n - add vendor_reset PCI sysfs for PLDR\n - Mediatek:\n - add USB IDs for MT7902 and MT7922 devices\n - Realtek:\n - add USB IDs for 8761CU and 8852BE devices\n - NXP:\n - add M.2 Bluetooth device support using pwrseq\n\n Misc:\n\n - DPLL support for manual/numerical oscillator control (NCO)\n (implement in zl3073x)\n\n - MCTP support for MCTP over USB v1.1 (DMTF DSP0283)\n\n - Power-over-Ethernet: support Realtek PSE controllers\n\n - Remove the IBM EHEA driver\n\n - Remove tulip/xircom_cb driver\"\n\n* tag \u0027net-next-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1433 commits)\n net/mlx5e: do not HW-GRO coalesce small frames\n net: openvswitch: fix nf_connlabels leak in ovs_ct_init\n net: add missing ref_tracker_dir_exit() to alloc_netdev_mqs()\n net: openvswitch: fix flow mask use-after-free on flow deletion\n sctp: stop processing a packet once its association is deleted\n dpll: zl3073x: add PTP clock support\n dpll: zl3073x: add channel ToD, phase step and TIE operations\n dpll: zl3073x: scale poll interval proportionally to timeout\n ptp: vmclock: prevent read-only mappings from becoming writable\n ipv4: reject undersized MTUs in ip_do_fragment()\n bonding: initialize err for empty target lists\n net: dsa: initial support for MT7628 embedded switch\n net: dsa: initial MT7628 tagging driver\n net: phy: mediatek: add phy driver for MT7628 built-in Fast Ethernet PHYs\n dt-bindings: net: dsa: add MT7628 ESW\n net: pse-pd: realtek-pse-mcu: add UART transport\n net: pse-pd: realtek-pse-mcu: add I2C transport\n net: pse-pd: add Realtek PSE MCU core\n dt-bindings: net: pse-pd: add bindings for Realtek PSE MCU\n vsock: use sock_error() to consume sk_err after a failed connect\n ...\n" }, { "commit": "5a8cd539ac19f7a68e68e1d25ef9ca2ff55b8500", "tree": "49b7f303dc09e1806777ebdfc64a30ba31d73720", "parents": [ "a4ff2be345d0abc943da8dd8da98151843b750dc", "f79066c784022fda83f5936559a1af414e41b603" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 07:36:20 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 20 07:36:20 2026 -0700" }, "message": "Merge tag \u0027bpf-next-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next\n\nPull bpf updates from Daniel Borkmann:\n \"Major changes:\n\n - Redesign the verifier error reporting: failures now carry source\n and instruction annotations along with the causal event history\n that led to them, making program rejections far easier to debug and\n repair (Kumar Kartikeya Dwivedi)\n\n - Add arena argument support to kfuncs and struct_ops through the new\n __arena and __arena__nullable suffixes (Tejun Heo, Puranjay Mohan,\n Kumar Kartikeya Dwivedi, Ihor Solodrai)\n\n - Signed BPF program loader rework to accommodate both BPF and\n security community needs where the kernel runs the signature\n verification at BPF_PROG_LOAD time before the LSM admission hook\n (Daniel Borkmann)\n\n - Add a set of ksock kfuncs which let BPF LSM and syscall programs\n create, connect and send on UDP sockets in order to emit telemetry\n data (Mahe Tardy)\n\n - Unify helper and kfunc call argument verification and classify\n kfunc arguments purely from BTF into a generated bpf_func_proto\n which is computed once at add-call time (Amery Hung)\n\n Other features and fixes:\n\n - Enable EXECMEM_ROX_CACHE for BPF allocations on x86 (Mike Rapoport)\n\n - Add bidirectional VLAN support to bpf_fib_lookup() through the new\n BPF_FIB_LOOKUP_VLAN and BPF_FIB_LOOKUP_VLAN_INPUT flags (Avinash\n Duduskar)\n\n - Infer zext_dst from static register liveness analysis to fix 32-bit\n zero-extension semantics, and remove the artificial limitations on\n pointer types eligible for spilling (Eduard Zingerman)\n\n - Inline the numeric open-coded iterator kfuncs so that bpf_for()\n loops no longer pay a kfunc call on every iteration (Puranjay\n Mohan)\n\n - Add an arena-based bitmap data structure to libarena along with\n serial and parallel selftests (Emil Tsalapatis)\n\n - Teach resolve_btfids to discover kfuncs from the kernel\u0027s BTF ID\n sets and to emit kfunc BTF decl tags, reducing the kernel build\u0027s\n dependency on pahole features (Ihor Solodrai)\n\n - Add BPF_F_ADJ_ROOM_DECAP_* flags to bpf_skb_adjust_room() so that\n tunnel decapsulation can update the GSO and encapsulation state of\n the skb (Nick Hudson)\n\n - Fix the ring buffer pending_pos walk and the available-data\n accounting on 32-bit position wrap (Israel Téllez García)\n\n - Add memory usage accounting for arena maps and fix an mmap_lock\n deadlock on arena lock failure (Jiayuan Chen)\n\n - Add tracing_multi link info support to the kernel UAPI and bpftool,\n and refactor the stack map code to run with preemption disabled\n (Jiri Olsa)\n\n - Support BPF_F_EGRESS in bpf_redirect_peer() to emit the skb in the\n egress direction of the target\u0027s peer device (Jordan Rife)\n\n - Add a KF_SPINLOCK_SAFE kfunc flag so that providers, in particular\n modules, can declare kfuncs safe to call under bpf_spin_lock\n instead of relying on the verifier\u0027s hard-coded allowlist (Kaitao\n Cheng)\n\n - Introduce global percpu data for BPF programs with libbpf probing\n and bpftool skeleton support, and stop exposing uninitialized\n kernel heap memory when copying per-CPU map values (Leon Hwang)\n\n - Add s390 JIT support for load-acquire and store-release\n instructions (Maxim Khmelevskii)\n\n - Fix a CFI mismatch in the task work callback and an arm64 KASAN\n false positive after bpf_throw() (Mykyta Yatsenko)\n\n - Reject writes through untrusted BTF pointers and bound the\n rdonly/rdwr_buf_size kfunc arguments (Nicholas Dudar)\n\n - Invalidate RCU pointers only after the final spin unlock and\n account for preempt and IRQ disabled regions as overlapping RCU\n protection (Ning Ding)\n\n - Support mixing bpf2bpf calls and tail calls on RV64, add signed\n operations and 32-bit atomics to the RV32 JIT, and add timed\n may_goto support (Pu Lehui, Kuan-Wei Chiu, Feng Jiang)\n\n - Fix a use-after-free on mm_struct in bpf_find_vma() for foreign\n tasks and an mmap_lock leak in the irq_work path (Sanghyun Park)\n\n - Populate mmap-able BPF array map memory lazily which makes mmap()\n O(1) instead of proportional to the map size (Song Liu)\n\n - Introduce a jit_required flag and reject programs with inlined\n helpers when no JIT is available, where the interpreter would\n otherwise jump into an invalid address (Tiezhu Yang)\n\n - Fix the x86 JIT per-CPU address resolution into an extended\n register where the REX prefix dropped the high destination register\n bit (Vineet Gupta)\n\n - Reject MEM_ALLOC BTF accesses past object bounds, arena frees below\n the arena base, and mixed arena and ordinary atomic paths (Yiyang\n Chen)\n\n - Fix the trampoline handling of 128-bit arguments and of return\n values larger than 8 bytes (Yonghong Song)\n\n - Ensure that any fault prone load is rewritten with exception table\n handling, and fix the arena load-acquire and atomic fetch handling\n in the x86, arm64, riscv and s390 JITs (Daniel Borkmann)\n\n - Many more fixes and cleanups across the verifier, arena,\n trampolines, sockmap, cgroup, ring buffer, x86/arm64/riscv/s390\n JITs, libbpf, bpftool, resolve_btfids and selftests\"\n\n* tag \u0027bpf-next-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (373 commits)\n selftests/bpf: Add tests for a store on a fault prone qdisc pointer\n selftests/bpf: Add tests for fault prone loads out of RCU pointers\n selftests/bpf: Add tests for pointer type merge at a shared load\n selftests/bpf: Remove duplicate copies of the arena spinlock qnodes\n selftests/bpf: Retry stat generation in cgroup_iter_memcg\n selftests/bpf: Test pseudo-function policy diagnostics\n bpf: Distinguish function references in policy diagnostics\n bpf: Preserve source attribution without source text\n selftests/bpf: Test kfunc argument diagnostics\n bpf: Correct kfunc argument diagnostics\n bpf: Use canonical stack argument names in diagnostics\n bpf: Preserve R0 lineage across helper calls\n selftests/bpf: Exercise negative optlen in cgroup getsockopt hook\n bpf: Reject negative optlen in cgroup getsockopt hook\n selftests/bpf: tc_tunnel - validate decap GSO and encapsulation state\n bpf: Clear decap state on skb_adjust_room shrink path\n bpf: Allow new DECAP flags and add guard rails\n bpf: Add BPF_F_ADJ_ROOM_DECAP_* flags for tunnel decapsulation\n bpf: Refactor masks for ADJ_ROOM flags and encap validation\n bpf: Name the enum for BPF_FUNC_skb_adjust_room flags\n ...\n" }, { "commit": "8049741ac93acd3a590dac070e12571fddf0e294", "tree": "af118e523d532a3edaeb54ef6f674f2dedb26e7f", "parents": [ "60a42d510113f46de47e86a84bf5758597644487" ], "author": { "name": "Serhat Kumral", "email": "serhatkumral1@gmail.com", "time": "Thu Aug 06 23:13:58 2026 +0300" }, "committer": { "name": "Jason Gunthorpe", "email": "jgg@nvidia.com", "time": "Thu Aug 20 08:46:15 2026 -0300" }, "message": "RDMA/ucma: Allow path records to exactly fit the output buffer\n\nucma_query_path() emits a path record only when the remaining output\nbuffer is strictly larger than struct ib_path_rec_data. A buffer sized\nexactly for the response header and N complete records therefore gets\nonly N - 1 records, while resp-\u003enum_paths still advertises N. A caller\nsizing its buffer for a single record gets a header claiming one path\nand no path data at all.\n\nucma_query_ib_service() in the same file computes the record count with\na plain division and so accepts an exact fit; make ucma_query_path()\nbehave the same way.\n\nCurrent librdmacm is unaffected because it always sizes the response for\nsix records while the kernel currently reports at most two paths. Other\nusers of the UAPI that provide an exactly sized buffer can observe the\ntruncated response.\n\nFixes: ac53b264b2f3 (\"RDMA/ucma: Support querying when IB paths are not reversible\")\nSigned-off-by: Serhat Kumral \u003cserhatkumral1@gmail.com\u003e\nLink: https://patch.msgid.link/20260806201358.147478-1-serhatkumral1@gmail.com\nSigned-off-by: Leon Romanovsky \u003cleon@kernel.org\u003e\nSigned-off-by: Jason Gunthorpe \u003cjgg@nvidia.com\u003e\n" }, { "commit": "c44e278ce02efd0c4be79a8eda1ea6885c1ce5ec", "tree": "b6b7324a56fb1869a094da177d6a41f4d2efe8e0", "parents": [ "65b92bc671db7c222f8de1146c88e40fe8764d95", "8d3ae59288f1e7d58d76558a6ee96d533bc5019f" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Thu Aug 20 10:58:44 2026 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Thu Aug 20 10:58:44 2026 +1000" }, "message": "BackMerge tag \u0027v7.2\u0027 into drm-next\n\nLinux 7.2\n\nThere was a lot of conflicts this round between fixes and next,\nand I\u0027d like to get the merge resolutions that we have in drm-tip.\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "a4ff2be345d0abc943da8dd8da98151843b750dc", "tree": "da64a462ec69df460299cfec5e0d61534daa8c7c", "parents": [ "a51ec5e8e5dae80824239f0344210060cb92a4b0", "7537036a2e6fe96f8ed82034f755c54714a0e417" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 17:25:42 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 17:25:42 2026 -0700" }, "message": "Merge tag \u0027v7.3-p1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6\n\nPull crypto update from Herbert Xu:\n \"API:\n - Add af_alg_restrict sysctl and white list\n - Fix potential suspend/resume races in hwrng\n\n Algorithms:\n - Optimize vli additive operations using compiler builtins in ecc\n\n Drivers:\n - Remove unsafe/deprecated algorithms from qce\n - Mark qce as BROKEN\n - Add runtime PM and interconnect bandwidth scaling support to qce\n - Remove crypto_rng from qcom, sun8i and caam\n - Fix SG list issues in iaa\n - Fix SEV init path bugs in ccp\"\n\n* tag \u0027v7.3-p1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (122 commits)\n crypto: lskcipher - propagate errors from unaligned crypt\n crypto: keembay - use crypto_memneq() to compare CCM AEAD tags\n crypto: keembay - use crypto_memneq() to compare GCM AEAD tags\n crypto: sa2ul - use crypto_memneq() to compare AEAD tag\n hwrng: drivers - use named initializers for acpi_device_id\n crypto: qce - fix CCM AAD buffer underallocation\n crypto: iaa - unmap dst before software fallback on decompress\n crypto: iaa - use bounce buffer for multi-sg decompress input\n crypto: iaa - avoid counting fallback decompression bytes\n crypto: iaa - fall back to software for multi-entry scatterlists\n hwrng: core - Stop/start hwrng_fillfn() kthread before/after suspend-resume\n crypto: hisilicon/sec2 - fix CCM algorithm long packet failure\n crypto: eip93 - use struct_size() and flexible array for ring allocation\n crypto: krb5 - use kfree_sensitive() for derived key buffers\n crypto: af_alg - Stop after finding name in allowlist\n crypto: af_alg - Replace \u0027bool privileged\u0027 with flags\n crypto: af_alg - Make cbc(paes) privileged-only\n hwrng: imx-rngc - Disable clock on registration failure\n crypto: qat - remove dead ADF_HEX code\n crypto: qce - simplify qce_handle_request\n ...\n" }, { "commit": "a51ec5e8e5dae80824239f0344210060cb92a4b0", "tree": "a74a19a16a8238d46a2892dae8520eb5148b4d34", "parents": [ "cbad8981fa3b1fc726e6e5652e92fd76c01e543c", "f1e10b10874051e4d99911dae0dd7b75a9f8ae66" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 17:17:17 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 17:17:17 2026 -0700" }, "message": "Merge tag \u0027integrity-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/zohar/linux-integrity\n\nPull integrity updates from Mimi Zohar:\n\n - TPM initialization is sometimes delayed until deferred_probe_initcall\n\n Since ordering is not guaranteed within the same initcall level, IMA\n may initialize before the TPM and fall back to TPM-bypass mode. A new\n config option, CONFIG_IMA_INIT_LATE_SYNC, allows those building the\n kernel to defer IMA initialization to late_initcall_sync, accepting\n the integrity risk of missing early measurements in exchange for\n avoiding TPM-bypass mode.\n\n - The raw policy rules are now measured, as well as the complete\n policy, closing a gap in integrity measurement coverage\n\n* tag \u0027integrity-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/zohar/linux-integrity:\n ima: measure userspace policy writes before parsing\n ima: add critical data measurement for loaded policy\n security: ima: rename boot_aggregate when ima is initialised at late_sync\n security: ima: introduce IMA_INIT_LATE_SYNC option\n security: lsm: allow LSMs to register for late_initcall_sync init\n" }, { "commit": "cbad8981fa3b1fc726e6e5652e92fd76c01e543c", "tree": "11e6aa60b1f407a5276fe1cc9b88f323ad663a21", "parents": [ "09005a63988521f74111fae344aecb3f63306168", "fedc88e38ce979a720cd2de042578cb5df3dc8de" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 16:51:39 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 16:51:39 2026 -0700" }, "message": "Merge tag \u0027Smack-for-7.3\u0027 of https://github.com/cschaufler/smack-next\n\nPull smack updates from Casey Schaufler:\n\n - Spelling fix\n\n - Code optimization in smackfs\n\n - Fix credential mis-uses\n\n - Place limits on two of the smackfs interfaces\n\n* tag \u0027Smack-for-7.3\u0027 of https://github.com/cschaufler/smack-next:\n smack: fix cred UAF in smack_file_send_sigiotask()\n smack: restrict smackfs/{direct,mapped} values to 0-255\n smack: deduplicate smackfs/{direct,mapped} file_operations\n smack: show msgrcv() subject task in audit\n smack: fix incorrect task context in smack_msg_queue_msgrcv\n security: smack: fix spelling mistake\n smack: simplify write handlers of sysfs entries\n Smack: Fix error in capability bypass\n" }, { "commit": "09005a63988521f74111fae344aecb3f63306168", "tree": "d77ec023ab384ca1ab073e707dd830b3d3117907", "parents": [ "4253eb09d25bcfe44f6987b6f67c09f09d80a966", "0cee720cfd51402cfcb14d96cb326a36c13b823a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 16:28:23 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 16:28:23 2026 -0700" }, "message": "Merge tag \u0027lsm-pr-20260814\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm\n\nPull LSM updates from Paul Moore:\n\n - Remove task_euid()\n\n The task_euid(), and Rust counterpart, was never widely used, for\n good reason, and now that the only user is gone we\u0027re removing it to\n rid ourselves of both dead and funky code.\n\n - Documentation improvements\n\n Correct some of the kdoc comments for security_task_prctl() and\n clarify the rust comments on task UID accessors.\n\n - Fix a memory leak in the LSM syscall selftests\n\n* tag \u0027lsm-pr-20260814\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm:\n selftests/lsm: Fix memory leak in attr_lsm_count\n cred: delete task_euid()\n rust: task: clarify comments on task UID accessors\n lsm: clarify security_task_prctl() hook documentation\n" }, { "commit": "4253eb09d25bcfe44f6987b6f67c09f09d80a966", "tree": "a87c01d2183ea31387ef8489a8a7840111e1621b", "parents": [ "83453b6f5131a83af7b2a4df28bc776353ac56c5", "fd6e2388a3ea55e58cbbbef840c1d8aa2067dbb3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 16:24:46 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 16:24:46 2026 -0700" }, "message": "Merge tag \u0027selinux-pr-20260814\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux\n\nPull selinux updates from Paul Moore:\n\n - Convert a __get_free_page() call into a kmalloc() call\n\n We had some very old code that called out to __get_free_page() for\n allocating a pathname. There is no reason this couldn\u0027t be done with\n a call to kmalloc() so we\u0027ve done the conversion and now there is one\n less __get_free_page() caller in the kernel.\n\n - Limit the number of retired/unknown DCCP netlink messages\n\n While DCCP is gone from the kernel, there are still userspace tools\n which try to talk to the kernel about DCCP sockets which were\n generating SELinux related log noise (unrecognized netlink message).\n This pull request both limits the log messages to just the first\n instance and also explains to the user that DCCP support has been\n removed.\n\n - Convert the SELinux strlcat() calls to seq_buf_XXX() calls\n\n As part of the effort to drop the strlcat() API from the kernel, the\n SELinux/IMA code was converted over to using seq_buf_XXX() calls.\n\n - Only calculate the SELinux IMA configuration string length once\n\n Previously each call to generate a SELinux configuration string for\n IMA would have to calculate the length of the string. While the\n contents of the string will likely change over the lifetime of the\n system, the length of the string will not. Calculate the string\n length once at boot and reuse the length value throughout the\n lifetime of the system.\n\n - Further validation of the SELinux policy at policy load time\n\n Perform additional sanity checks on the policy constraints and types.\n\n - Proper cleanup and error handling for selinuxfs init failures\n\n We were not properly cleaning up some state in the case where\n selinuxfs fails to initialize properly. It\u0027s somewhat of an academic\n exercise as a failure to initialize selinuxfs will cause the system\n to fail on boot, but it\u0027s arguably better to make sure we do things\n the proper way.\n\n - Various code cleanups\n\n Convert integer flags to boolean types and drop an uncessary goto\n from the SELinux code.\n\n* tag \u0027selinux-pr-20260814\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux:\n selinux: validate constraint expression attr and op at load time\n selinux: compute the IMA configuration settings string length once at boot\n selinux: replace strlcat() with seq_buf in selinux_ima_collect_state()\n selinux: suppress warning flood for retired DCCP netlink messages\n selinux: tighten type validation during policy load\n selinux: drop unnecessary goto and label from avc_alloc_node()\n selinux: convert int flags to bool flags in ss/services.c\n selinux: clean up selinuxfs resources on init failure\n selinux: hooks: use kmalloc() to allocate path buffer\n" }, { "commit": "83453b6f5131a83af7b2a4df28bc776353ac56c5", "tree": "88c1aa2a799b8aea425fe5e084a97f4cfb22a18a", "parents": [ "cb8a75eec0877810b50aa1c5a833f929525cd2ee", "783f0f0974c156aca630f4ffff248671082a098d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 16:21:32 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 16:21:32 2026 -0700" }, "message": "Merge tag \u0027audit-pr-20260814\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/audit\n\nPull audit updates from Paul Moore:\n\n - Drop BUG_ON() assertions from two functions\n\n While I don\u0027t recall any bug reports from either of these assertions\n in recent memory, neither of these checks warrant the kernel panic\n that could result from BUG_ON(). One of the BUG_ON() calls is\n converted to a WARN_ON_ONCE() and the other to a lockdep assertion.\n\n - Fix an audit tree reference counting problem\n\n Fix a corner case where audit could end up unintentionally dropping\n the last reference to an audit tree while the tree was still in use.\n\n We should probably revisit the audit tree handling code in full, but\n this patch works, and should be easy to backport to stable trees and\n downstream kernels.\n\n - Update the audit syscall classification tables\n\n Add some missing syscalls to the PERM class\n\n* tag \u0027audit-pr-20260814\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/audit:\n audit: avoid dropping live tree ref on fsnotify rule autoremove\n audit: drop BUG_ON() from audit_signal_info_syscall()\n audit: drop BUG_ON() from audit_add_to_parent()\n audit: add missing syscalls to PERM class tables\n" }, { "commit": "cb8a75eec0877810b50aa1c5a833f929525cd2ee", "tree": "bdb5ca7e1a09eabe313d9c09acefdaaea985b539", "parents": [ "1484625c5968cad5e26b653d13b4b7a18c07834d", "8b502bf6eb3da15f4b954ad3632335ff10ed746a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 14:22:07 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 14:22:07 2026 -0700" }, "message": "Merge tag \u0027trace-ringbuffer-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull ring-buffer updates from Steven Rostedt:\n\n - Remove unneeded semicolon\n\n A macro ended with a semicolon that wasn\u0027t needed.\n\n - Fix freeing cpu_buffer extra subbuffer with order greater than zero\n\n When the cpu_buffer was being freed, its \"free\" page, was using\n free_page() to free it when it could be more than one page.\n\n - Hold the cpu_buffer lock when resizing the subbuffer\n\n The freeing of the \"free\" page of the cpu_buffer was done without\n locking. The order of the data was being saved and then the \"free\"\n page was set to NULL. But there is a race that the \"free\" page could\n have been updated between those two operations. Add locking around it\n to prevent the race.\n\n - Save the order of the data along with the data in the free page\n\n The cpu_buffer would store just the data portion of the subbuffer\n page in its descriptor. But it did not store the order of the data\n pages. The order was being saved in the global buffer descriptor. But\n this leads to races.\n\n Have the cpu_buffer save the subbuf data along with its metadata\n (which includes the order of the page) to make sure when it frees it,\n it frees the correct order along with it.\n\n - Remove the subbuf_size and use the order directly when needed\n\n Having a size field for the size of the subbufer along with its order\n allowed for races to have them get out of sync. Remove the\n subbuf_size and use the order from the subbuf meta data directly\n under locks.\n\n Use the subbuf_order for other calculations in the ring buffer.\n\n - Remove the useless \"cpus\" field of trace_buffer\n\n The code has been restructured and the \"cpus\" field is no longer\n used. Remove it.\n\n - Remove the \"mapped\" field of the ring buffer and use a helper\n function instead.\n\n The \"mapped\" field has become a bit overused and made the code come\n complex in using a counter for what is denoted as being mapped or\n not. There are other fields that are set when the ring buffer is\n considered mapped. Add a helper function to check those fields and\n use that instead of keeping track of a counter.\n\n* tag \u0027trace-ringbuffer-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n ring-buffer: Remove ring_buffer_per_cpu::mapped\n ring-buffer: Remove trace_buffer::cpus\n ring-buffer: Dynamically calculate max_data_size\n ring-buffer: Fix subbuf resize race with ring_buffer_alloc_read_page()\n ring-buffer: Fix subbuf resize race with ring buffer readers\n ring-buffer: Make cpu_buffer::free_page a buffer_data_read_page\n ring-buffer: Hold cpu_buffer::lock when resizing a subbuf\n ring-buffer: Free cpu_buffer::free_page with subbuf_order\n ring-buffer: drop unneeded semicolon\n" }, { "commit": "1484625c5968cad5e26b653d13b4b7a18c07834d", "tree": "51106d968c224bd4ac58b403ad2c6158363a4905", "parents": [ "081e5bf2a9d941da60cd70c97c5a704b29e47f7f", "43a17abc83c6402bffac80a6dd434f8555f457e8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 14:18:45 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 14:18:45 2026 -0700" }, "message": "Merge tag \u0027tracefs-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull tracefs updates from Steven Rostedt:\n\n - Define event fields before directory creation\n\n Move the event_define_fields() call in event_create_dir() before the\n eventfs directory creation. Previously, a failure after directory\n creation wouldn\u0027t clean up eventfs_inode because the error path\n didn\u0027t call eventfs_remove_dir(). This eliminates the need to clean\n up the eventfs directories if event_define_fields() fails.\n\n - Add warning for out of bounds pos in __eventfs_iterate()\n\n Sashiko complains about the ctx-\u003epos causing issues if it is less\n than 2 or greater than MAX_INT in __eventfs_iterate(). The thing is,\n the logic prevents that from happening. But to make Sashiko happy,\n add a WARN_ON() and exit safely if the function ever does get input\n that is out of the range the function expects.\n\n* tag \u0027tracefs-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n eventfs: Add warning for out of bounds pos in __eventfs_iterate()\n eventfs: Define event fields before directory creation\n" }, { "commit": "081e5bf2a9d941da60cd70c97c5a704b29e47f7f", "tree": "b3c0df90abf9b3e64ec54358a778d46b20a95568", "parents": [ "00d66b29a66ce18e417a8436076c629d03186a27", "ae70b04ab9c7f6162a8c0fdd18a62a945c133142" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 14:06:14 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 14:06:14 2026 -0700" }, "message": "Merge tag \u0027trace-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull tracing updates from Steven Rostedt:\n\n - Expose btf_ids to trace events\n\n In order to allow BPF programs to attach to system call trace events\n (which are actually pseudo trace events built on top of raw_syscall\n events), expose the BTF ID of the events. This will allow BPF\n programs better precision in attaching to events.\n\n - Use \"u64\" to assign to hist_field-\u003etype\n\n Instead of using kstrdup(\"u64\", GFP_KERNEL) to assign the\n hist_field-\u003etype, just point it to \"u64\" instead. The\n hist_field-\u003etype is freed via kfree_const().\n\n - Replace kmalloc()/strcpy() with kstrdup() for trace_printk\n\n Instead of having two calls to copy the module format string, just\n use kstrdup().\n\n - Use __free() in trace event histograms and triggres where possible\n\n - Use seq_buf in trace event code instead of strcat()\n\n Instead of calculating the size of the buffer to use and filling it\n with strcat(), use the seq_buf infrastructure that takes care of\n making sure not to overflow the string size.\n\n - Reject invalid preemptirq_delay_test CPU affinity\n\n The preempt_delay_test module can take an invalid CPU affinity mask\n and create confusing output. Simply have the module reject invalid\n affinity masks.\n\n - Prevent division by zero in ftrace_ops sample module code\n\n If the ftrace_ops sample module code receives the module parameter\n nr_function_calls set to zero, it can cause a division by zero error.\n\n - Warn when an event dereferences a parameter in TP_printk()\n\n On boot up and module load, the trace event TP_printk() is scanned\n for possible bugs. As the TP_printk() code is executed when the user\n reads the \"trace\" file and processes the data written when the\n trace_event executed, the data it reads can be literally days old.\n The scan currently checks for dereferencing printk formats like\n \"%pI6\". But it does not check if the parameters themselves have a\n dereference like:\n\n\tTP_printk(\"offset %08x: value %08x\",\n\t\t(u32)(__entry-\u003eaddr - __entry-\u003eedma-\u003emembase), __entry-\u003evalue)\n\n __entry represents the pointer to the event on the ring buffer. The\n __entry-\u003eedma-\u003emembase is dereferencing a pointer on the ring buffer\n to find membase, but the __entry-\u003eedma may no longer be a valid\n pointer.\n\n Warn on this case too.\n\n - Replace some strcpy() with strscpy()\n\n - Clean up mmiotrace events to use assign_type() macro\n\n The assign_type() macro makes sure the event type is indeed the type\n that is being parsed. The mmiotrace trace was written before that\n macro was created so it just simply typecasted the pointer.\n\n Replace the typecasting with the macro.\n\n - Have the ENUM processing to numbers only process what is added\n\n The code that converts ENUMs to their numbers in the trace events\n scanned all events to do the processing. This was true when a module\n was loaded too. That is, instead of processing just the events for\n the module, it processed *all* events. Even the builtin ones that\n were processed at boot up.\n\n Add a check for the event-\u003emodule matching mod if it is a module\n before processing it.\n\n* tag \u0027trace-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (21 commits)\n tracing: Have trace_event_update_all() only handle module that is loading\n tracing: Cleanup event_enable_trigger_parse() by using __free()\n tracing: Report every TP_printk double dereference\n tracing/mmiotrace: Use trace_assign_type() in mmio_print_mark()\n tracing: Make per-template BTF id lists file-local\n tracing: Use seq_buf for string concatenation\n tracing: Use strscpy() instead of strcpy() in trace_sched_switch\n tracing: Warn when an event dereferences a pointer in TP_printk()\n samples/ftrace: Prevent division by zero when nr_function_calls is zero\n tracing: Reject invalid preemptirq_delay_test CPU affinity\n fgraph: Use trace_seq_putc() in print_graph_return()\n tracing/user_events: Replace a seq_printf() call by seq_puts() in user_seq_show()\n tracing/user_events: Use seq_putc() in two functions\n tracing: Bound histogram expression strings with seq_buf\n tracing: Return ERR_PTR() from expr_str()\n tracing: Use __free() for expr_str() buffer\n kernel/trace/trace_printk: Use kstrdup() instead of kmalloc() and strcpy()\n tracing: Point constant hist field type to string literal\n selftests/bpf: Add test for tracepoint btf_ids tracefs file\n tracing: Expose tracepoint BTF ids via tracefs\n ...\n" }, { "commit": "00d66b29a66ce18e417a8436076c629d03186a27", "tree": "b4f5d8d1ea2734b61f866c4f00b13dea9161a8e8", "parents": [ "3793b558ff3b1cae64a3eddb497ab9e5636bf022", "88b3e7fedc07d940c32eb5c0733a780e944b6b47" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 14:01:58 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 14:01:58 2026 -0700" }, "message": "Merge tag \u0027ftrace-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull ftrace updates from Steven Rostedt:\n\n - Deprecrate ftrace_enabled in disabling ftrace\n\n The file /proc/sys/kernel/ftrace_enabled was created when ftrace was\n first introduced back in 2008. It was to be a \"kill switch\" if\n something was to go wrong. It was also used as a way to turn off\n function tracing for the latency tracers that would have it on by\n default. But in 2013 (Linux 3.10) the option \"function-trace\" was\n introduced to disable function tracing for the latency tracers as the\n \"ftrace_enabled\" file was considered too big of a hammer and caused\n too many side effects.\n\n When live kernel patching came along, disabling ftrace via the\n ftrace_enabled file would put the system into an unstable state if a\n live kernel patch was installed. This created the need to mark some\n function hooks as \"PERMANENT\".\n\n Now there\u0027s a need for BPF usage marked as PERMANENT for the same\n reasons.\n\n The file \"ftrace_enabled\" usage is no longer viable. It doesn\u0027t do\n what it says it does and there is no reason to use it.\n\n Make writing \u00270\u0027 to it a nop and print a message saying its usage is\n deprecated. The return value of writing \u00270\u0027 is -EOPNOTSUPP so that\n user space will error on that write (hopefully to inform any\n developer that it no longer works).\n\n Eventually the file should be removed completely, but for now just\n making it not do anything is the path forward to that.\n\n - Update the livepatch tests to handle ftrace_enabled being disabled\n\n Because in the past, livepatch was broken by ftrace_enabled being\n turned off, there\u0027s a test case that checks to make sure it still\n doesn\u0027t break. But having the write of \u00270\u0027 return an error caused\n that test to break. Updated the test to handle the new change.\n\n* tag \u0027ftrace-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n selftests/livepatch: update test-ftrace.sh for deprecated ftrace_enabled\n ftrace: deprecate disabling via ftrace_enabled sysctl\n" }, { "commit": "3793b558ff3b1cae64a3eddb497ab9e5636bf022", "tree": "f63886c002647346e7dfbf78e10ed9d03baff282", "parents": [ "55ee4b931a7ffedc886175d265dd6e6d08fd4151", "0ad45018ace734e3a21c8b15280df65b429135d1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 14:00:15 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 14:00:15 2026 -0700" }, "message": "Merge tag \u0027trace-tools-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull RTLA updates from Steven Rostedt:\n\n - Extend support for unsetting CLI options\n\n libsubcmd auto-generates \"--no-\u003coption\u003e\" to unset options, provided\n the option callback supports it. Implement this for RTLA CLI beyond\n boolean options, and document the few exceptions that are left out.\n\n - Test all tracer options in runtime tests\n\n Verify that RTLA sets osnoise/timerlat options correctly by reading\n them from tracefs during runtime tests.\n\n - Improve range validation for option arguments\n\n Make CLI range validation consistent with the kernel limits and unify\n implementation and error messages between options.\n\n - Improve invalid option argument parsing\n\n Consistently reject invalid values for numeric option arguments with\n a unified error message for all options.\n\n* tag \u0027trace-tools-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n rtla/cli: Unify and improve handling of invalid option arguments\n rtla/cli: Unify and improve range validation logic\n rtla/tests: Test all tracer options in runtime tests\n Documentation/rtla: Document unsetting options\n rtla: Add unit tests for CLI with unset\n rtla: Add unit tests for unset in opt callbacks\n rtla: Allow unsetting non-list custom-callback CLI options\n" }, { "commit": "55ee4b931a7ffedc886175d265dd6e6d08fd4151", "tree": "23501b461d58dd52edec64fc6f1cb995e7cc5862", "parents": [ "98f21c54f99519329c18e2625b0ea6db14524d09", "7d81675d1bb2cc6db61a2d93b1e0dc7fb0929f9c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 13:29:22 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 13:29:22 2026 -0700" }, "message": "Merge tag \u0027trace-rv-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull Real-time Verifier updates from Steven Rostedt:\n\n - Switch LTL and DOT parsers to Lark in code generation tool\n\n The rvgen code generation tool originally parsed DOT files and LTL\n specifications using custom string parsing and Ply, which is no\n longer maintained. The DOT parser was fragile and prone to failure on\n minor format variations. Both LTL and DOT parsers have been rewritten\n to use the Lark parsing library.\n\n - Simplify Hybrid Automata clock variables\n\n The clock variables in hybrid automata monitors now use a single\n representation of the elapsed time since the clock was reset, rather\n than converting between invariant and guard representations. This\n allows simpler code generation for the newly refactored parser.\n\n - Generate cleanup hook for per-obj monitor\n\n The code generation scripts now adds a cleanup function to per-obj\n monitors for the user to wire to the appropriate event (e.g.\n sched_process_exit for tasks).\n\n - Reduce read_lock scope during per-task cleanup\n\n Take the tasklist_lock only when necessary, that is when iterating\n over for_each_process_thread().\n\n - Simplify task monitor slot management\n\n Only rely on the slot array for per-task slot management to avoid\n inconsistency with the unused counter.\n\n - Improve rvgen code robustness and templates\n\n Use pathlib in rvgen and improve kernel path discovery. Also improve\n consistency across templates when generating code (e.g. author\n placeholder and monitor struct name).\n\n - Update rtapp sleep monitor\n\n Simplify the sleep monitor by excluding kernel threads and updating\n the nanosleep check to focus only on CLOCK_REALTIME. Also switch to\n use the sched_exit tracepoint to run in the context of the offending\n (wakee) task.\n\n - Add wakeup monitor\n\n Add the new rtapp/wakeup monitor to detect when lower-priority tasks\n wake up higher-priority ones, complementing the existing sleep\n monitor by running in the waker context and capturing its stack\n trace.\n\n - Fix tools/rv exit status on failure\n\n Ensure the rv tool returns a failure exit code when a monitor fails\n to start because it was already running.\n\n - Add automated selftests for tools/rv and rvgen\n\n Introduced automated bash selftests to validate rv monitor listing\n and execution under different configurations. Added tests for the\n rvgen code generator, validating generated files against expected\n output (golden). Tests are reachable via make check.\n\n - Add KUnit test coverage for verification monitors\n\n Added comprehensive KUnit tests to validate the functionality of\n deterministic, hybrid, and LTL monitors by emulating event sequences\n and timing in a mock environment without affecting the running kernel\n while expecting mock reactions to fire. Ensure real RV monitors\n cannot run during KUnit tests to avoid state corruption.\n\n - Mock current in rv monitors\n\n Mock the call to current in rv monitors when the KUnit tests are\n built to allow them to run the test on dummy tasks. No overhead is\n expected when KUnit tests aren\u0027t running.\n\n - Introduce rvgen kunit subcommand\n\n Added a new \u0027kunit\u0027 subcommand to rvgen to automatically patch an\n already generated monitor with KUnit integration templates by parsing\n its event handlers and creating the required mock structures and\n initializations.\n\n - Refine kernel verification selftests\n\n Added new selftests for the deadline and stall monitors and\n rearranged the existing wwnr_printk test to resolve flakiness.\n Additionally, fixed an issue in the selftests framework where\n negative assertion failures were not correctly propagated due to\n shell rules.\n\n - Fix 32-bit build of nomiss KUnit test\n\n A previous commit introduced a division between an u64 and a constant\n value and that doesn\u0027t build on 32-bit systems. Use div_u64()\n instead.\n\n - Document changes in sleep monitor\n\n The sleep monitor introduced some changes in the past like allowing\n epoll_wait() as a valid sleep and a task going to runnable before\n scheduling as a valid wakeup. Document both.\n\n* tag \u0027trace-rv-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (40 commits)\n Documentation/rv: Explain epoll and aborted sleeps\n rv: Fix 32-bit build of nomiss KUnit test\n selftests/verification: Add selftests for deadline and stall monitors\n selftests/verification: Rearrange the wwnr_printk test\n selftests/verification: Fix wrong errexit assumption\n rv: Add KUnit tests for some LTL monitors\n rv: Add KUnit mock for current\n rv: Add KUnit tests for some DA/HA monitors\n rv: Export task monitor slot and react symbols\n verification/rvgen: Add selftests for rvgen kunit\n verification/rvgen: Add the rvgen kunit subcommand\n verification/rvgen: Add selftests\n verification/rvgen: Add golden and spec folders for tests\n tools/rv: Add selftests\n verification/rvgen: Improve consistency in template files\n verification/rvgen: Use pathlib instead of os.path\n verification/rvgen: Improve rv_dir discovery in RVGenerator\n tools/rv: Fix exit status when monitor execution fails\n rv: Use generic rv_this for the rv_monitor variable in LTL\n rv/rtapp: Add wakeup monitor\n ...\n" }, { "commit": "98f21c54f99519329c18e2625b0ea6db14524d09", "tree": "cc8926009720125d50c5ffb0ca184befa23ed549", "parents": [ "85e0d1de369cdf827dae8c759f40696de81b7ca7", "e052daab94ee8c4081c91de3549772bd1c235729" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 12:51:36 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 12:51:36 2026 -0700" }, "message": "Merge tag \u0027for-linus-fwctl\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/fwctl/fwctl\n\nPull fwctl updates from Jason Gunthorpe:\n\n - Support more commands in bnxt, this completes what they originally\n wanted to do\n\n - Rust bindings for fwctl. The Nova GPU is expected to use them next\n cycle\n\n* tag \u0027for-linus-fwctl\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/fwctl/fwctl:\n rust: introduce abstractions for fwctl\n fwctl/bnxt: Add DMA buffer support for HWRM commands\n bnxt_en: Update bnxt firmware spec\n" }, { "commit": "85e0d1de369cdf827dae8c759f40696de81b7ca7", "tree": "6bf63080b2b06426360182c4146c9ed3b34305b5", "parents": [ "4994ef0fe01e0c58a9af187f98a030c35dfb1a93", "8c07df7cdfcf52f1ff276c588612aabc6c6b8399" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 12:42:58 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 12:42:58 2026 -0700" }, "message": "Merge tag \u0027for-linus-iommufd\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jgg/iommufd\n\nPull iommufd updates from Jason Gunthorpe:\n \"One small feature this cycle, the noiommu mode is useful in\n single-purpose VMs running something like DPDK. It avoids the double\n translation overhead and it seems to be commonly used with some hacks.\n\n Summary:\n\n - Formal API for \"no iommu\" mode in VFIO. iommufd now works in this\n environment and provides page pinning and phyiscal address services\n to userspace. This avoids nasty fragile tricks with mprotect and\n pgmap\n\n - Fix sykzaller crash racing change_process with map_pages\n\n - Various skyzkaller/AI fixes for the selftests\"\n\n* tag \u0027for-linus-iommufd\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jgg/iommufd:\n iommufd: Fix UAF in selftest IOPF reporting\n iommu/iommufd: Fix NULL pointer deref in iommufd_ioas_change_process when racing with iopt_map_file_pages\n Documentation: Update VFIO NOIOMMU mode\n vfio: Enable cdev noiommu mode under iommufd\n iommufd: Add an ioctl to query PA from IOVA for noiommu mode\n iommufd: Allow binding to a noiommu device\n iommufd: Move igroup allocation to a function\n iommufd: Support a HWPT without an iommu driver for noiommu\n iommufd: Simplify iommufd_device_remove_vdev()\n iommufd: Fix grammar and spelling in comments\n iommu: Fix dev_iommu memory leak when device_add fails in iommu_mock_device_add\n iommufd/selftest: Fix dmabuf leak in iommufd_test_dmabuf_get()\n iommufd/selftest: Avoid selftest dirty bitmap size wrap\n" }, { "commit": "4994ef0fe01e0c58a9af187f98a030c35dfb1a93", "tree": "2fc94108d9a04eba290fc6e9e07c0e6a2204cd18", "parents": [ "7fa7d4c6038bb4b394478ba1b31a6e8d89ed4f60", "3c6a2bac15247bc4f28125247ff89165324612fe" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 12:38:26 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 12:38:26 2026 -0700" }, "message": "Merge tag \u0027iommu-updates-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/iommu/linux\n\nPull iommu updates from Joerg Roedel:\n \"ARM SMMUv2:\n - Device-tree binding updates for Qualcomm Eliza, Maili, Shikra and\n IPQ9650 SoCs\n - Add support for Qualcomm SM8450\n - Numerous fixes for lifetime and ordering issues found by Sashiko in\n the Qualcomm driver\n\n ARM SMMUv3:\n - Fix interrupt type in device-tree binding example for NVIDIA CMDQV\n - Numerous fixes for issues identified by Sashiko in the NVIDIA CMDQV\n driver\n - Work around TLB erratum T264-SMMU-3 on Tegra264 by repeating the\n invalidation sequence\n - Add support for HAFT (hardware access flag in table entries) when\n using SVA\n - Probe for 52-bit addressing with a page size smaller than 64k\n (\u0027DS\u0027) but don\u0027t do anything with it for now\n - Minor driver improvements (remove sort_nonatomic(), use\n readl_relaxed_poll_timeout_atomic(), fix IOPF teardown ordering)\n\n Intel VT-d:\n - Consolidation of complex enablement logic into a clean,\n priority-based state machine\n - Support for the DMA_REMAP_OPT_OUT flag from the VT-d v5.2\n specification\n - An update to cache_tag_flush_devtlb_psi() to use full-range\n constants instead of modifying shared variables for\n CACHE_TAG_NESTING_DEVTLB\n - A fix for the UCTP context-table slot when copying root entries\n - Fixes for several pre-existing issues reported by Sashiko\n - General code cleanup and refinement\n\n AMD IOMMU:\n - Add SNP page-mode-0 support, enabling passthrough, v2 DMA page\n tables and host SVA on supporting systems\n - Fix invalid PPR handling, COMPLETE_PPR responses and guest-mode\n reporting\n - Improve Southbridge IOAPIC validation and remove the dependency on\n hard-coded device IDs\n - Fix PCI-device lifetime, debugfs and diagnostic issues\n\n IOMMU core and IOMMUFD:\n - Restore serialization of the shared MSI-page list\n - Fix SVA-handle publication and several IOMMUFD reference and error\n path leaks\n - Return the expected zero result for invalid generic page-table\n translations\n - Allocate per-CPU IOVA magazines lazily to reduce memory use on\n large systems\n\n PCI ATS:\n - Make VF support checks account for the associated PF and validate\n that VF and PF Smallest Translation Unit settings agree\n\n Platform drivers:\n - Fix Qualcomm runtime-PM, probe unwind, fault reporting and page\n table initialization races\n - Rework Rockchip state handling and fix clock, probe and stale-fault\n handling\n - Fix smaller issues in the MSM and MediaTek drivers\n\n Device-tree bindings:\n - Add new Qualcomm SMMU compatibles, convert the OMAP IOMMU binding\n to YAML, and fix the Tegra264 CMDQV interrupt example\n\n Various smaller cleanups, documentation fixes and a Rust IOMMU\n safety/readability improvement\"\n\n* tag \u0027iommu-updates-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/iommu/linux: (93 commits)\n iommu/amd: Add SNP page mode 0 support\n iommu/amd: Fix GN bit setting in COMPLETE_PPR_REQUEST command\n iommu/amd: Rate limit INVALID_PPR_REQUEST error logging\n iommu/amd: Fix missing CMD_COMPLETE_PPR response for invalid PPR requests\n iommu/amd: Introduce PPR_TAG_LAST_PAGE() macro\n iommu/amd: Fix incorrect device ID in invalid PASID error message\n iommu/vt-d: Flush context cache with correct SID when tearing down aliases\n iommu/vt-d: Tear down scalable-mode context on probe failure\n iommu/vt-d: Fix iopf_refcount leak on RID domain replacement\n iommu/vt-d: Clear Present bit before tearing down copied context entry\n iommu/vt-d: Fix copied_tables bitmap leak on error in copy_translation_tables\n iommu/vt-d: Cache max domain ID to avoid redundant calculation\n iommu/vt-d: Support the new DMA_REMAP_OPT_OUT flag bit\n iommu/vt-d: Remove dmar_disabled\n iommu/vt-d: Remove the \u0027force_on\u0027 variable\n iommu/vt-d: Call dmar_can_force_on() for tboot opt-in\n iommu/vt-d: Use dmar_can_force_on() for platform opt-in\n iommu/vt-d: Consolidate dmar policy management and force_on logic\n iommu/vt-d: Remove dead code when CONFIG_INTEL_IOMMU is not set\n iommu/vt-d: Force requesting ACS when tboot is enabled\n ...\n" }, { "commit": "7fa7d4c6038bb4b394478ba1b31a6e8d89ed4f60", "tree": "ecc0fbcf2d4190389535bf37058ba09d17b3dcdb", "parents": [ "faabe2db712e8310dbe38fb81d33e5313ff0f059", "39c5aa3bd8ec3912d2cd0b3fe092642b0d2b0713" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 12:35:15 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 12:35:15 2026 -0700" }, "message": "Merge tag \u0027for-7.3/dm-changes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm\n\nPull device mapper updates from Mikulas Patocka:\n\n - minor cleanups found by Claude Opus 4.6\n\n - small cleanups in dm core, dm-cache, dm-switch, dm-inlinecrypt,\n dm-vdo\n\n - improve validation of metadata in dm-pcache\n\n - fix resume-vs-remove ioctl race condition\n\n - fix race condition when issuing table load ioctls concurrently\n\n - fix dm-raid1 and dm-io, so that they work with unaligned bio vectors\n\n - dm-integrity: use keyed markers as discard fillers\n\n - improve metadata validation in dm-array\n\n - fix dm-stats crash on memory allocation failure\n\n - fix dm-dust, so that it works if it is not the first target in a\n table\n\n - dm-era: fix superblock refcount leak on snapshot failure\n\n* tag \u0027for-7.3/dm-changes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm: (46 commits)\n dm-era: fix shadowed superblock leak on take-snap failure\n dm dust: make badblock messages target-relative\n dm-stats: fix a crash if allocation of per-cpu data fails\n dm array: reject an array block whose value size is not the caller\u0027s\n dm array: validate array block headers on read\n dm-integrity: replace forgeable discard filler with a keyed sector marker\n dm vdo indexer: embed geometry in parent structures\n dm vdo indexer: simplify sub-index parameter calculations\n dm-pcache: remove unused \u0027cache\u0027 parameter from cache_key_gc()\n docs: device-mapper: dm-inlinecrypt: fix \u0027bellow\u0027 spelling\n dm-pcache: remove unused miss_read_end_work_fn declaration\n dm-io: report non-retryable errors separatedly\n dm-io: clone the source bio instead of copying its biovec\n dm: fix race when loading and unloading a table\n dm: fix resume-vs-remove race\n dm-pcache: remove unused \u0027allocated\u0027 variable in cache_data_alloc()\n dm-pcache: replace tabs with spaces in comments to fix ASCII diagram alignment\n dm-pcache: fix use-after-free and invalid seg operations in kset_replay()\n dm-pcache: fix implicit u8 truncation of gc_percent in message handler\n dm raid1: reserve space for NUL-terminator in build_constructor_string()\n ...\n" }, { "commit": "faabe2db712e8310dbe38fb81d33e5313ff0f059", "tree": "8f917cb0126f652d041e8b2ed03f5aa6fa5a1e54", "parents": [ "59e6295fac26b8e85c1ea859cdd89fa1e47519d7", "73b6bf3c314ad34a9fd389f84de6e8dbfc204880" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 12:31:50 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 12:31:50 2026 -0700" }, "message": "Merge tag \u0027ata-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/libata/linux\n\nPull ata updates from Damien Le Moal:\n\n - Some code cleanups to rename the function used to identify ZAC\n devices and declare some local functions static (me)\n\n - Refactoring and improvement of the translation of the SCSI REPORT\n SUPPORTED OPCODES command to allow users access to the entire list of\n supported commands (me)\n\n - Fix the translation of the WRITE SAME command with UNMAP bit set (DSM\n TRIM) for devices with a sector size larger than 2K and devices that\n support multiple TRIM segments (Niklas)\n\n - Add support detecting support for and translating the SCSI commands\n related to the storage elements depopulation feature (GET PHYSICAL\n ELEMENT STATUS, REMOVE ELEMENT AND TRUCATE, REMOVE ELEMENT AND MODIFY\n ZONES and RESTORE ELEMENTS AND REBUILD) (me)\n\n - Improvements to the sata_mv driver probe code (clocks and IRQ\n initialization) (Rosen)\n\n - Improve resource initialization in the pata_rb532_cf, pata_pxa,\n sata_highbank and ahci_da850 drivers (Rosen)\n\n - Improve PIO data-in command completions to better hndle slow devices,\n e.g. CF cards (Richard)\n\n - Improve the DMA channel management using device resources in the\n pata_pxa driver (Rosen)\n\n - Fix the pata_ep93xx driver to correctly fallback to PIO mode if DMA\n initialization fails (Rosen)\n\n - Use named initializers to define the match tables of the ahci_xgene,\n ahci_qoriq and ahci_platform drivers (Pawel)\n\n* tag \u0027ata-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/libata/linux: (28 commits)\n ata: use named initializers for acpi_device_id\n ata: pata_ep93xx: fix PIO fallback when DMA init fails\n ata: pata_pxa: use devres for DMA channel management\n ata: libata-sff: don\u0027t busy-wait for PIO data-in command completion\n ata: ahci_da850: use devm_platform_ioremap_resource()\n ata: sata_highbank: use devm_platform_ioremap_resource\n ata: pata_pxa: use devm_platform_ioremap_resource\n ata: pata_rb532_cf: use devm_platform_ioremap_resource()\n ata: sata_mv: use devm clock helpers\n ata: sata_mv: Use platform_get_irq() to get interrupt\n ata: pata_mpc52xx: Remove redundant dev_err()\n ata: libata-eh: make ata_eh_qc_complete() and ata_eh_qc_retry() static\n ata: libata-scsi: add support for the REMOVE ELEMENT AND MODIFY ZONES command\n ata: libata-scsi: add support for the RESTORE ELEMENTS AND REBUILD command\n ata: libata-scsi: add support for the REMOVE ELEMENT AND TRUNCATE command\n ata: libata-scsi: add support for the GET PHYSICAL ELEMENT STATUS command\n ata: libata-core: detect support for depopulation capabilities\n ata: libata-scsi: improve ata_get_xlat_func\n ata: libata: improve the definition of device flags\n scsi: define depopulation capabilities related service actions\n ...\n" }, { "commit": "59e6295fac26b8e85c1ea859cdd89fa1e47519d7", "tree": "7c071f8bab440eddbb569e84e9f31c80f8a4e529", "parents": [ "c6cf4441a3a05bb7273ed022f3e56c4fc591da08", "3b5ea0f078e1b72276e99c237f3dfc2fd72938bb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 10:42:18 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 10:42:18 2026 -0700" }, "message": "Merge tag \u0027driver-core-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/driver-core/driver-core\n\nPull driver core updates from Danilo Krummrich:\n \"container_of:\n\n - Apply typeof_member(), remove the local __mptr variable to\n eliminate variable shadowing warnings on nested container_of()\n calls, and remove unnecessary parentheses\n\n core:\n\n - Add driver name to probe debug print for initcall_debug\n\n - Avoid repeatedly printing the same \u0027Fixed dependency cycle\u0027 log\n\n - Unwind device_add() on attribute creation failure in\n attribute_container_add_class_device()\n\n - Remove statistics group if encryption group creation fails in\n transport_add_class_device()\n\n debugfs:\n\n - Fix lockdown check for mmap_prepare()\n\n - Warn if file creation failed due to uninitialized debugfs\n\n device property:\n\n - Implement fw_devlink support for software nodes by adding\n software_node_add_links(), which creates fwnode links from\n DEV_PROP_REF properties to enable automatic probe ordering. Add\n kunit-managed fwnode helpers and test coverage\n\n - Fix infinite loop in fwnode_for_each_child_node() when the\n secondary fwnode has more than one child. Add test cases\n\n - Fix out-of-bounds access in software_node_get_reference_args() when\n called with index -1 (UINT_MAX)\n\n - Refactor to use RAII approach with __free()\n\n - Add Bartosz Golaszewski as software node reviewer\n\n firmware loader:\n\n - Fix race where a sysfs fallback request can complete before being\n queued as pending, leading to a use-after-free on the next fallback\n request\n\n - Reject 0-size built-in firmware and fail the build on empty\n firmware files in CONFIG_EXTRA_FIRMWARE\n\n kobject:\n\n - Provide __KOBJ_ATTR() and __KOBJ_ATTR_RO/WO() initialization macros\n and allow the constification of kobject attributes, enabling them\n to reside in read-only memory\n\n platform:\n\n - Provide platform_device_set_of_node(), platform_device_set_fwnode(),\n and platform_device_set_of_node_from_dev() helpers that encapsulate\n firmware node reference counting for dynamically allocated platform\n devices\n\n Convert all in-tree users that manually assigned dev.of_node or\n dev.fwnode, fixing a pre-existing refcount bug in powermac. Switch\n to counting references of all firmware node types, not only OF\n nodes\n\n - Unify the release path for dynamically allocated platform devices\n by removing platform_device_release_full(). Amend the fwnode setter\n API contract to warn if a primary software node is overwritten. Add\n KUnit tests for correct software node removal on device\n unregistration\n\n Rust:\n\n - Auxiliary:\n - Add registration_data_with() closure-based API for invariant\n ForLt types\n\n - Debugfs:\n - Migrate BinaryWriter and BinaryReaderMut trait requirements\n from kernel::transmute traits to zerocopy traits\n\n - Device:\n - Add BoundInternal device context and InternalBoundContext trait\n for bus abstractions that need internal access to a bound\n device.\n - Make the lifetime on Core and CoreInternal invariant to prevent\n coercion to shorter lifetimes\n\n - Devres:\n - Fix race between concurrent revokers where the losing revoker\n could return before the winning revoker finished dropping the\n inner data, causing use-after-free.\n - Ensure revocation is complete before the device finishes\n unbinding by making the synchronization bidirectional.\n - Add DevresLt\u003cF: ForLt\u003e, a wrapper around Devres that shortens\n \u0027static back to the caller\u0027s borrow scope. Implement ForLt and\n CovariantForLt for Bar, IoMem, and ExclusiveIoMem\n\n - Driver:\n - Switch from index-based to pointer-based device ID info lookup,\n storing static references in driver_data. Centralize device ID\n handling in device_id.rs, removing the open-coded ACPI/OF\n matching logic and duplicate ID table from driver.rs\n\n - I/O:\n - Make I/O regions typed (with a dynamically-sized Region type\n for the existing untyped case), create view types representing\n subregions of a mapped I/O region, and add io_project!() for\n safely creating subviews.\n - Split Io into a base trait (IoBase) and an extension trait (Io)\n with a blanket implementation, preventing implementers from\n overriding provided methods that unsafe code relies on.\n - Add a SysMem backend for shared system memory with volatile\n access, and make Coherent implement Io via an I/O view type.\n Add IoSysMap as sum type of Mmio and SysMem. Add copying\n methods (memcpy_{from,to}io()) and read_val()/write_val() for\n typed access.\n - Replace dma_read!()/dma_write!() with io_read!()/io_write!()\n for primitives and copying methods for aggregates; drop the old\n macros. Convert nova-core to use I/O projection.\n - Fix internal shortcut rule dispatch in the register!() macro,\n remove unused rule arguments, and use path fragments for alias\n destinations\n\n - IRQ:\n - Make irq::Registration compatible with lifetime-bound drivers\n by removing the \u0027static bound on Handler/ThreadedHandler and\n replacing Devres\u003cRegistrationInner\u003e with direct\n request_irq()/free_irq() calls. Handlers can now directly own\n lifetime-bound device resources\n\n - PCI:\n - Convert IrqVectorRegistration to a lifetime-annotated owning\n type, giving drivers explicit control over the allocation\n lifetime. IrqVector embeds a resolved IrqRequest, making the\n conversion infallible. Remove the redundant\n request_irq()/request_threaded_irq() wrappers from pci::Device.\n - Add pci_irq_type() C helper and expose it via irq_type() on\n IrqVectorRegistration and IrqVector, returning PCI_IRQ_MSIX,\n PCI_IRQ_MSI, or PCI_IRQ_INTX.\n - Mark pci::Device refcount methods inline\n\n - Serdev:\n - Add Rust abstractions for the serial device bus, including\n serdev::Driver trait, serdev::Device wrapping struct\n serdev_device, and serdev::Adapter implementing\n RegistrationOps. Includes a sample driver. Markus Probst takes\n over as serdev maintainer for both C and Rust code\n\n - Misc:\n - Split ForLt into a base trait (providing the Of\u003c\u0027a\u003e GAT) and an\n unsafe CovariantForLt subtrait guaranteeing covariance,\n enabling invariant types (e.g. those containing Mutex\u003c\u0026\u0027bound T\u003e)\n to participate in the ForLt abstraction.\n - Fix Coherent read past EOF returning -ERANGE instead of zero.\n - Fix firmware example UB by avoiding null-pointer ARef\n\n misc:\n - Avoid iattr allocation in kernfs listxattr by using\n kernfs_iattrs_noalloc().\n - Unregister SoC bus on early device registration failure.\n - Remove unused DMA_FENCE_TRACE Kconfig symbol.\n - Fix /sys/module path in comment.\n - Refactor ISA bus init to remove nested blocks.\n - Remove redundant nodemask clears in numa_init().\n - Add kernel-doc for fwnode_operations and sys_soc.h, mark\n internal property data as private for kernel-doc, and add\n property.h/fwnode.h to driver-api infrastructure docs.\n - Add MAINTAINERS entry for sys_soc.h\"\n\n* tag \u0027driver-core-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/driver-core/driver-core: (129 commits)\n rust: pci: expose the allocated interrupt type\n PCI: Add pci_irq_type() to query the allocated interrupt type\n rust: pci: remove request_irq() and request_threaded_irq() from Device\n rust: pci: resolve IRQ in index() and embed IrqRequest in IrqVector\n rust: pci: convert IrqVectorRegistration to a lifetime-managed owning type\n kernfs: avoid iattr allocation in listxattr\n rust: serdev: use ThisModule::as_ptr() instead of field access\n ACPI/IORT: use platform_device_set_fwnode()\n ACPI/APMT: use platform_device_set_fwnode()\n firmware_loader: do not queue completed sysfs fallback requests\n rust: pci: Mark Device refcount methods inline\n rust: irq: make Registration compatible with lifetime-bound drivers\n rust: net/phy: remove expansion from doc\n rust: dma: return zero for Coherent reads past EOF\n rust: io: register: use path fragment for alias destination\n rust: io: register: remove unused rule arguments\n rust: io: register: dispatch shortcut rules internally\n MAINTAINERS: add sys_soc.h to DRIVER CORE\n rust: debugfs: remove unsafe blocks from traits impl for Vec\n rust: debugfs: migrate debugfs traits requirements to zerocopy\n ...\n" }, { "commit": "c6cf4441a3a05bb7273ed022f3e56c4fc591da08", "tree": "c7671ca74940ec7ab0a524bb96e2dedb333d76dd", "parents": [ "f4cdf7ca9a1fdcca413157df19753f388a5a224e", "e1c13aed4a1641550e8c9518c96f016e1d52d09b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 10:32:34 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 10:32:34 2026 -0700" }, "message": "Merge tag \u0027devicetree-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux\n\nPull devicetree updates from Rob Herring:\n\n - Add a DT maintainer profile document\n\n - Various dt-check-style improvements\n\n - Add a devres managed reserved memory region init function\n\n - Print node name on any skipped reserved memory regions\n\n - Correctly handle optional argument in\n of_parse_phandle_with_args_map()\n\n - Convert ti,keystone-reset, ti,da850-vpif, TI L4 interconnect, TI\n SmartReflex, microchip,pic32mzda-dmt, microchip,pic32mzda-wdt, TI\n DA8XX MSTPRI bus, and Xen VM bindings to DT schema format\n\n - Add bindings for StarFive JHB100 plic, Allwinner A733 NMI controller,\n MediaTek MT8173 GPU, QCom Shikra, Eliza, and Maili cpu-bwmon, and\n QCom Shikra SCM firmware\n\n - A couple of syntax fixes found using PoC Rust implementation of\n dtschema tools\n\n - Clean-ups for typos, brackets, incorrect \"::\" usages, and\n white-space style\n\n* tag \u0027devicetree-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux: (40 commits)\n dt-bindings: interconnect: qcom-bwmon: Add Maili cpu-bwmon compatible\n dtc: dt-check-style: Simplify setting depth of DtsLine\n dtc: dt-check-style: Add missing /dts-v1/ to few test cases\n dt-bindings: power: reset: ti,keystone-reset: Convert to DT schema\n media: dt-bindings: ti,da850-vpif: Convert to dt-schema\n dt-bindings: devfreq: samsung,exynos-ppmu: Use standard regex syntax\n dt-bindings: interrupt-controller: mediatek,mt6577-sysirq: Drop invalid JSON pointer\n dt-bindings: arm: omap: Convert L4 interconnect to DT schema\n dt-bindings: power: Convert TI SmartReflex to DT schema\n of: reserved_mem: Introduce devres-managed initialization function\n dt-bindings: interrupt-controller: Add StarFive JHB100 plic\n dt-bindings: irq: sun7i-nmi: Document the Allwinner A733 NMI controller\n dt-bindings: Correct white-space style\n dt-bindings: fix typos and brackets\n docs: dt: submitting-patches: Mention expectation about dt-check-style\n docs: dt: maintainer: Add Devicetree and OF maintainer profile document\n docs: dt: writing-schema: Extend expectations about example part of binding\n dt-bindings: gpu: powervr-rogue: Add MediaTek MT8173 GPU\n of: base: Handle optional argument in of_parse_phandle_with_args_map()\n dt-bindings: update Sudeep Holla\u0027s email address\n ...\n" }, { "commit": "f4cdf7ca9a1fdcca413157df19753f388a5a224e", "tree": "60a31895b2446cd7fa27514e8c2843534a0de0da", "parents": [ "e5c91aac491def6ab3f90c4cc246e3fcb0f8f058", "4900cad020c0580dfb1be27776ff10a4ef110cfa" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 10:09:22 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 10:09:22 2026 -0700" }, "message": "Merge tag \u0027media/v7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media\n\nPull media updates from Mauro Carvalho Chehab:\n\n - v4l2-core: added ISP statistics support and per-block validation\n\n - v4l2-core: Allow unknown HDR10 white point and luminance\n\n - New camera sensors: Sony IMX678 and IMX471m, Himax HM1092 IR sensor\n\n - New codec: Milos: VPU v2.0 codec support\n\n - isp driver: gained support for Dreamchip RPPX1 ISP framework\n\n - vsp1 driver: gained support for RZ/T2H and RZ/N2H\n\n - Novalake driver: gained CVS support for new NVL hardware\n\n - dvb-core: fix feed leak on failed DMX_ADD_PID\n\n - several driver fixes, cleanups and minor improvements\n\n* tag \u0027media/v7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media: (308 commits)\n media: ipu-bridge: check all DMI entries when overriding sensor rotation\n media: v4l2-async: avoid deleting unlinked ASC entry on link error\n media: rzg2l-cru: Align bytesperline to hardware DMA stride requirement\n media: intel/ipu6: fix async notifier cleanup leak on parse error\n media: staging/ipu7: fix async notifier UAF on probe error path\n media: amd: isp4: fix self-deadlock in isp4sd_pwron_and_init() error path\n media: amd: isp4: release partial allocations in isp4if_alloc_fw_gpumem()\n media: rcar-isp: Fix VSPX reference leaks\n media: rcar-isp: Release ISPCORE resources\n media: i2c: imx415: Release runtime PM reference on VBLANK error\n media: i2c: imx415: Return test pattern write errors\n media: renesas: vsp1: Declare index variables in for loop statement\n media: renesas: vsp1: Make reset control optional to support platforms without a reset line\n media: dt-bindings: media: renesas,vsp1: Document RZ/T2H and RZ/N2H SoCs\n media: dt-bindings: media: renesas,fcp: Document RZ/T2H and RZ/N2H SoCs\n media: nxp: imx8-isi: Add additional 32-bit RGB format support\n media: nxp: imx8-isi: Add 16-bit raw Bayer format support\n media: nxp: imx8-isi: Implement per-stream reference counting for multiplexed streams\n media: nxp: imx8-isi: Use BIT_ULL() for 64-bit stream masks\n media: nxp: imx8-isi: Correct color map between V4L2 and ISI\n ...\n" }, { "commit": "e5c91aac491def6ab3f90c4cc246e3fcb0f8f058", "tree": "20ce0b23e5e56efa3707c7ff4244fe4da65547c7", "parents": [ "a93f3bf4e1d60777b1659b812c9e818cfc53b449", "c139e7e44f58a6f8ddc9d850ea9924d34963b5da" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 10:04:59 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 10:04:59 2026 -0700" }, "message": "Merge tag \u0027sound-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound\n\nPull sound updates from Takashi Iwai:\n \"It was a fairly busy development cycle - the changes spread over from\n the core side to leaf drivers, with lots of cleanups and enhancements.\n Here we go, some highlights:\n\n ALSA core:\n - Extension of ALSA control component list ABI\n - Locking optimization and RCU conversion of ALSA sequencer core\n - A few hardening fixes for UMP and sequencer core\n - Drop __bitwise and __force prefix from UAPI definitions\n\n ASoC:\n - Automatic DAI format selection code deployment across many drivers\n - Sorting of register default tables to prevent ordering issues in\n many drivers\n - Lots of code cleanups and refactoring\n - Updates in Qualcomm driver stack\n - New platforms: AMD ACP7.B/F, Cirrus Logic CS35L62, Loongson\n 2K0300, Meson GX, Qualcomm LPI MI2S, SM8475, WSA855X, Realtek\n RT1321 VA1/2 and RT766/7\n\n HD-audio:\n - Support for AW88399 HD-audio side codec for Lenovo Legion laptops\n - Support for Hygon and Lisuan HDMI controllers\n - Robustness fixes for wild device binding\n - Lots of quirks/fixups: Realtek and Conexant codecs for ASUS,\n Lenovo, Acer, etc\n\n USB-audio:\n - Support for Pioneer DJ DJM-S11\n - Scarlett2/FCP private URB notification fixes\n - Extended quirk_flags to 64bit\n - Hardening fixes for 6fire, bcd2000, usx2y\n - Device-specific quirks for Mackie, Valeton, SPACETOUCH\n\n General:\n - Auto-cleanup for put_device() and firmware loading across multiple\n platforms\"\n\n* tag \u0027sound-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound: (791 commits)\n ALSA: hda: Fix connection list comparison in proc output\n ALSA: docs: fix dead link to Intel HD-audio spec\n ALSA: usb-audio: Add delay quirk for SPACETOUCH USB Audio\n ALSA: hda: Add Lisuan HDMI controller and codec support\n ALSA: hda/realtek: Fix Lenovo Yoga Slim 7 14AKP10 quirk ordering\n ALSA: hda/tas2781: Add hardware stabilization delay during firmware load retries\n ALSA: hda/realtek: Fix mute LED for HP Victus 15-fa1xxx (MB 8C3F)\n ALSA: hda/realtek: Add micmute LED quirk for Acer Aspire A515-57\n ASoC: tas2783-sdw: do not treat read-only Controls as writable\n ASoC: SOF: validate topology volume range before allocation\n ASoC: cs35l56: Use IRQ provided by the SoundWire core\n soundwire: bus_type: Create IRQ mapping before calling driver probe()\n ASoC: cs35l56: Move cs35l56_irq_request() after cs35l56_irq()\n ASoC: cs35l56: Request IRQ in cs35l56_common_probe()\n ALSA: core: Fix use-after-free in snd_card_do_free()\n ALSA: hda/realtek: Drop duplicate quirk for Lenovo 0x17aa:0x38df\n ALSA: usb-audio: Rename the Audient iD14 monitor mix volume control\n ASoC: tas2781: Refactor calibration start kcontrol creation to separate helper\n ASoC: dt-bindings: es8316: Fix supply property constraints\n ALSA: seq: midi: Serialize input teardown with event_input\n ...\n" }, { "commit": "a93f3bf4e1d60777b1659b812c9e818cfc53b449", "tree": "b72ace6a8daab0659264d38399cc618f01c0ec2b", "parents": [ "7acf90feab8b009fde7def08ff2c622d0f10e99f", "d89f04ac41e3b83a25b02c8d818ce09d7afc74c9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 10:00:31 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 10:00:31 2026 -0700" }, "message": "Merge tag \u0027hid-for-linus-2026081901\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid\n\nPull HID updates from Jiri Kosina:\n \"Core:\n - fix long-standing force-feedback initialization race across the\n subsystem (Dmitry Torokhov)\n - switch to system_dfl_wq (Marco Crivellari)\n\n AMD-SFH:\n - support for tablet-mode switch for AMD SFH-based systems (Basavaraj\n Natikar)\n\n HyperX:\n - support for HyperX QuadCast 2 (Benjamin Blume)\n\n I2C-HID:\n - support for devices that provide HID descriptor solely through\n the ACPI _DSM method (XIE Zhibang)\n\n Intel-THC-HID:\n - support for full I2C bus config parameters (Even Xu)\n\n Logitech:\n - HID++ 2.0 repogrammable button support (Elliot Douglas)\n - Bolt receiver support for HID++ devices (Erik Håkansson)\n\n MSI:\n - support for MSI Claw (Derek J. Clark)\n\n Steam:\n - initial support for 2026 Steam Controller (Vicki Pfau)\n - support for sensor events on the 2025 Steam Controller (Vicki Pfau)\n\n And many, many other fixes for various long standing issues that were\n found by new modern tools, and quite a few device ID additions\"\n\n* tag \u0027hid-for-linus-2026081901\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid: (146 commits)\n HID: tmff: Use 64-bit arithmetic for force feedback scaling\n HID: multitouch: reclassify HTIX5288 to WIN_8_FORCE_MULTI_INPUT_NSMU\n HID: sensor: custom: Fix field sysfs group cleanup on failure\n HID: sensor: custom: Fix use-after-free in enable_sensor\n HID: intel-thc-hid: intel-quickspi: bound GET_REPORT response to the caller buffer\n HID: haptic: don\u0027t write an uninitialized value to unhandled usages\n HID: intel-thc-hid: intel-quickspi: fix autosuspend cleanup during teardown\n HID: intel-thc-hid: intel-quicki2c: fix autosuspend cleanup during teardown\n HID: steam: Zero out inputs when disabling gamepad mode\n HID: steam: Clean up locking\n HID: steam: Don\u0027t set feature reports when disconnecting\n HID: steam: Fix wording of connect/disconnect logs\n HID: steam: Initial 2026 Steam Controller support\n HID: steam: Refactor registration\n HID: logitech: add Bolt receiver support for Logitech HID++ devices\n HID: sensor-hub: Fix out-of-bounds write in sensor_hub_get_feature\n HID: universal-pidff: stop the device when force-feedback init fails\n HID: haptic: move FF initialization into .input_configured()\n HID: logitech-hidpp: move FF initialization to .input_configured()\n HID: megaworld: move FF initialization to .input_configured()\n ...\n" }, { "commit": "7acf90feab8b009fde7def08ff2c622d0f10e99f", "tree": "7d35a67ff5ece640499d3ab73200afae3bfead56", "parents": [ "4e1b759c06721bc4e0fdfe5c179b58e7e20d3d13", "75f2c0b3690702c90863c2e138cb5520670845ea" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:56:28 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:56:28 2026 -0700" }, "message": "Merge tag \u0027hwmon-for-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging\n\nPull hwmon updates from Guenter Roeck:\n \"New drivers:\n - Kandou KB9002 retimer\n - PolarFire SoC temp/voltage sensor\n - Eswin EIC7700 PVT sensor\n - PMBus:\n - Analog Devices MAX16545/MAX16550 and Volterra VT7505\n - Monolithic MPQ82D00 and MPQ8646\n - Silergy SQ24860\n\n Added support to existing drivers:\n - asus-ec-sensors: Support for ROG STRIX Z390-E GAMING, ProArt\n Z690-CREATOR WIFI, ROG STRIX X870E-E GAMING WIFI7 R2, ROG CROSSHAIR\n X870E HERO, and ROG Maximus Z790 Hero\n - asus_rog_ryujin: Siupport for ROG Ryujin III\n - ina2xx: Support for INA232\n - k10temp: Per-CCD temperature monitoring for Zen5 Turin\n - nct6775: List NCT5585D as supported chip\n - nzxt-kraken3: Support for NZXT Kraken 2024 Elite\n - sht3x: Support for GXCAS GXHT30\n - tmp102: Add device IDs for TMP110 and TMP113\n - yogafan: Support for LOQ 15IAX9, XiaoXin Pro 13ARE 2020, IdeaPad 3\n 15ALC6, Legion Pro 7 16AFR10H, Yoga Pro 7 14IAH10, Yoga 7 16ARP8,\n and Lenovo LOQ 15IAX9\n - PMBus:\n - max20830: Support for max20830c and max20840c\n - max34440: Support for MAX34452, and support for newer version of\n max34451\n - adm1275: Support for ROHM BD12780 and BD12790\n\n Other notable changes:\n - Constify various device attributes\n - Remove redundant dev_err() and dev_err_probe() from various drivers\n - applesmc: Convert to hwmon_device_register_with_info\n - adt7470: Add thermal zone sensor support\n - coretemp: Fix core_data leak on CPUs without PTS\n - emc1403: Drop hysteresis for low limit temperature\n - max6621: Fix various over- and underflow problems\n - PMBus:\n - Introduce pmbus_read_smbus_i2c_block_data() and use it in\n various drivers\n - Export and use pmbus_check_and_notify_faults()\n - Let PMBus drivers report the supported PMBus revision\n\n Various other minor fixes and improvements\"\n\n* tag \u0027hwmon-for-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging: (110 commits)\n hwmon: (emc1403) Drop hysteresis for low limit temperature\n hwmon: (coretemp) Fix core_data leak on CPUs without PTS\n hwmon: (max6621) fix negative temperature offset and crit readings\n hwmon: (max6621) fix temperature clamp range\n hwmon: (asus_rog_ryujin) Add ROG Ryujin III White Edition\n hwmon: (asus_rog_ryujin) Add ROG Ryujin III support\n hwmon: (asus_rog_ryujin) Add per-device configuration\n hwmon: (k10temp) Add per-CCD temperature monitoring for Zen5 Turin\n hwmon: (tmp102) Add TMP113 device ID\n hwmon: (tmp102) Add TMP110 device ID\n hwmon: (nct6775) Add NCT5585D to list of supported chips\n Documentation: hwmon: (nct6775) Add missing NCT6797D and NCT6798D\n hwmon: (emc1403) Add regulator support\n hwmon: (emc1403) Convert to use OF bindings\n dt-bindings: hwmon: Document SMSC EMC1402/1403/1404/1428\n hwmon: (asus-ec-sensors) add ROG STRIX Z390-E GAMING\n hwmon: (sysfs) Allow drivers to register const attributes\n hwmon: (corsair-psu) Update documentation\n hwmon: (core) Use const APIs for the dynamically allocated sysfs attributes\n hwmon: (core) Constify device attributes\n ...\n" }, { "commit": "4e1b759c06721bc4e0fdfe5c179b58e7e20d3d13", "tree": "972227ab15ef36ecd7ce58eb3dfb93ce4dcc95a3", "parents": [ "307b9ddbbcf987db77d52da6f9ff5b4096ac9599", "3bbbeb7fddf54a7d7116b2defe0a63a9b583c84e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:53:35 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:53:35 2026 -0700" }, "message": "Merge tag \u0027watchdog-for-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging\n\nPull watchdog updates from Guenter Roeck:\n \"New Drivers:\n - Nuvoton MA35D1\n - Lenovo SE30G2 and SE60\n\n Added support to existing drivers:\n - snps,dw-wdt: Add RV1106 compatible\n - apple,wdt: Add t6030, t6031, and t8132 compatibles\n\n Other notable changes:\n - New \"dump\" pretimeout governor\n - Propagate errors from optional IRQ lookup\n - Remove redundant dev_err() and dev_err_probe() messages\n - npcm, qcom: Improved bootstatus reports\n - realtek-otto: Change to use regmap API\n - w83627hf_wdt: Report running watchdog, identify NCT6126\n\n Various other minor fixes and improvements\"\n\n* tag \u0027watchdog-for-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging: (40 commits)\n watchdog: orion_wdt: Propagate errors from optional IRQ lookup\n watchdog: qcom: Propagate errors from optional IRQ lookup\n watchdog: aspeed: Propagate errors from optional IRQ lookup\n watchdog: stm32_iwdg: Propagate errors from optional IRQ lookup\n watchdog: dw_wdt: Propagate errors from optional IRQ lookup\n watchdog: mediatek: Propagate errors from optional IRQ lookup\n watchdog: apple: Constify some structures\n watchdog: pretimeout: Convert dump pretimeout governor to tristate\n nmi: Export CPU backtrace APIs for loadable modules\n watchdog: booke_wdt: Document unused parameter of __booke_wdt_disable()\n watchdog: wdat_wdt: map registers that fall inside ACPI NVS\n watchdog: Add Nuvoton MA35D1 watchdog driver support\n dt-bindings: watchdog: Add MA35D1 Watchdog\n watchdog: qcom: report bootstatus on IPQ9574 and IPQ5332\n watchdog: qcom: report WDIOF_POWERUNDER in bootstatus\n watchdog: sprd: Remove redundant dev_err()\n watchdog: sama5d4: Remove redundant dev_err()\n watchdog: realtek_otto: Remove redundant dev_err_probe()\n watchdog: orion: Remove redundant dev_err()\n watchdog: marvell_gti: Remove redundant dev_err_probe()\n ...\n" }, { "commit": "307b9ddbbcf987db77d52da6f9ff5b4096ac9599", "tree": "1980cc69ac47eeb544e0c58079b88ebcd273932b", "parents": [ "b3438e5ca785565a65ef231bc03cd5a05c3be5c7", "23688febe4b44b10f4b454eb1cde0ba379a84119" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:47:41 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:47:41 2026 -0700" }, "message": "Merge tag \u0027spi-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi\n\nPull spi updates from Mark Brown:\n \"Along with a lot of driver specific work we\u0027ve got a couple of core\n features here. The bigger one is that we\u0027ve now got support for\n instantiating devices from sysfs similarly to how it\u0027s already done\n for I2C, this is used with development boards with non-enumerable\n expansion headers since SPI devices need to be manually specified. We\n also have support for the DQS signal on higher end flash devices.\n\n - Support for instantiating devices from sysfs, useful for\n development boards with non-enumerable plugin modules, from\n Vishwaroop A.\n\n - Support for DQS in spi-mem, an additional signal used by flash\n devices to avoid clock skew from Miquel Raynal.\n\n - Support for more advanced SPI modes on DesignWare controllers from\n Sudip Mukherjee.\n\n - Changes from Jisheng Zhang to update to modern methods of\n specifying the PM callbacks.\n\n - Fixes for DMA mapping error handling, plus KUnit tests for this,\n from Honghui Jiang.\n\n - Substantial cleanup and performance work in the nxp-spi driver.\n\n - Support for Microchip LAN969x, Nuvoton MA35D1 QSPI, Qualcomm\n SA8255p and SA8797P, and StarFive JHB100 SFC\"\n\n* tag \u0027spi-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi: (132 commits)\n spi: Add KUnit coverage for DMA mapping error paths\n spi: Clear current DMA devices when unmapping a message\n spi: Move __spi_unmap_msg() before __spi_map_msg()\n spi: Fix DMA mapping ownership on partial map failure\n spi: dt-bindings: sun6i: Add compatibles for A733\u0027s SPI controllers\n spi: ma35d1-qspi: Use the existing update helper\n spi: ma35d1-qspi: Add DTR support\n spi: ma35d1-qspi: Allow several command bytes\n spi: ma35d1-qspi: Move speed setting to bus configuration\n spi: ma35d1-qspi: Remove redundant reset operation\n spi: dw: Remove shadowed dws in dw_spi_setup()\n spi: img-spfi: don\u0027t disable runtime PM on DMA deferred probe\n spi: mtk-nor: Propagate errors from IRQ request\n spi: mtk-nor: Propagate errors from optional IRQ lookup\n spi: spi-qpic-snand: Handle Macronix quad read opcode 0x6b\n spi: spi-qpic-snand: add quad mode support\n spi: spi-qpic-snand: move command mapping helper\n spi: hisi-sfc-v3xx: Propagate errors from optional IRQ lookup\n spi: meson-spifc: use devm_pm_runtime_set_active_enabled\n spi: sprd-adi: Fix probe succeeding without registering the controller\n ...\n" }, { "commit": "b3438e5ca785565a65ef231bc03cd5a05c3be5c7", "tree": "af71b6874f05d24c45e0104c0c0cfa87ecc38807", "parents": [ "259c4f8e775cf25069c30e806b037b154147720d", "e739acbe05b06de78ef7089470f122432f651faa" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:36:58 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:36:58 2026 -0700" }, "message": "Merge tag \u0027regulator-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator\n\nPull regulator updates from Mark Brown:\n \"This is a relatively quiet release for the regulator API, we\u0027ve had no\n major core work and not really that much driver work either. There\u0027s a\n bunch of activity, including several new devices, but nothing hugely\n remarkable here.\n\n - Reworking of the mode handling in the max14577 driver to fix issues\n with collisions with enables\n\n - Support for onsemi FAN53555BUC23X, Qualcomm IPQ9650, PM4125 VBUS\n and PM8150B and Unisoc SC2730\"\n\n* tag \u0027regulator-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator: (36 commits)\n regulator: fan53555: Add support for FAN53555BUC23X type\n regulator: qcom-rpmh: Fix coding style issues\n regulator: qcom-rpmh: readback voltage/bypass/mode set during bootup\n regulator: qcom-rpmh: Fix PMIC5 BOB bypass mode handling\n soc: qcom: rpmh: Add support to read back resource settings\n regulator: dt-bindings: ti,pbias-omap: Convert to DT schema\n regulator: ab8500: Remove stale expand_register kernel-doc entry\n regulator: dt-bindings: Correct white-space style\n regulator: pfuze100: add set_suspend_disable for LDO ops\n regulator: core: use system_freezable_wq for init complete work\n regulator: rt6245: Restore state on enable failure\n regulator: tps65185: handle gpiod_get_value_cansleep() error returns\n regulator: fan53555: Add support for mode operations on Silergy devices\n regulator: dt-bindings: Add fan53555 allowed modes\n regulator: wm831x-isink: remove conditional return with no effect\n regulator: dt-bindings: Convert ltc3589.txt to yaml format\n regulator: dt-bindings: tps51632: Convert to DT schema\n regulator: mcp16502: Convert to dev_err_probe() in mcp16502_probe()\n regulator: adp5055: Fix error code in adp5055_of_parse_cb()\n regulator: qcom_usb_vbus: add support for qcom,pm4125-vbus-reg\n ...\n" }, { "commit": "259c4f8e775cf25069c30e806b037b154147720d", "tree": "0ca236c1df4d66197c913ae37c0daca01b55ae43", "parents": [ "ee1b6365f07a14e987dd0f994fc7d46966037ea9", "ef4f2357df6e9dd6213ce00d01253f3c458d5df0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:32:57 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:32:57 2026 -0700" }, "message": "Merge tag \u0027regmap-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap\n\nPull regmap updates from Mark Brown:\n \"This is a relatively busy release, though it\u0027s mostly cleanup work. We\n did add some new hooks for regmap-irq to support some driver work,\n that should also come in as part of a shared branch with the relevant\n driver work in the GPIO subsystem\"\n\n* tag \u0027regmap-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:\n regmap: clean up kernel-doc comments\n regcache: Validate cache_only state in regcache_sync_region()\n regcache: Warn if regcache_sync() is called in cache_only mode\n regcache: Mark cache dirty if selector register rewrite fails\n regcache: Preserve cache synchronization errors in regcache_sync()\n regmap: maple: Workaround for another false-positive compiler warning\n regcache: Make -\u003eexit() callback return void\n" }, { "commit": "ee1b6365f07a14e987dd0f994fc7d46966037ea9", "tree": "891d2f0fbbd37e52667a46c21226f862f00e665e", "parents": [ "abea5c349368a683e40fdc2e745099aaaacc89d7", "3ca7b513aec1db32b75b909395be1111c1198afb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:28:46 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:28:46 2026 -0700" }, "message": "Merge tag \u0027pmdomain-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/linux-pm\n\nPull pmdomain updates from Ulf Hansson:\n - amlogic: Add support for A9 power domains\n - bcm: Raise ASB poll timeout to 100us for bcm2835-power\n - imx: Allow building power domain drivers as a modules\n - mediatek:\n - Add support for the MT6858 power domains\n - Add support for the MT8196 HFRP DirectCTL power domains\n - qcom:\n - Add support for RPMh power domains for Maili\n - Skip retention by default for rpmhpd\n - renesas: Add support for R-Car X5H Module Controller\n - rockchip: Add a regulator to the RK3568 NPU power domain\n - tegra: Add support for multi-socket platforms\n\n* tag \u0027pmdomain-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/linux-pm: (24 commits)\n pmdomain: renesas: Add R-Car X5H MDLC driver\n dt-bindings: power: Document Renesas R-Car X5H Module Controller\n pmdomain: amlogic: Add support for A9 power domains controller\n dt-bindings: power: Add Amlogic A9 power domains\n clk: imx: imx8qxp: add soft dependency on SCU power domain driver\n pmdomain: imx: scu-pd: allow building as a module\n of: export of_stdout symbol\n pmdomain: imx8m{p,}-blk-ctrl: Add MODULE_DESCRIPTION\n pmdomain: mediatek: Add support for MT6858 SoC\n pmdomain: mediatek: Add support for secure modem power domain control\n dt-bindings: power: Add MediaTek MT6858 power domain controller\n pmdomain: rockchip: Add a regulator to the RK3568 NPU power domain\n pmdomain: imx: Make IMX8M/IMX9 BLK_CTRL tristate\n dt-bindings: power: qcom,rpmpd: document RPMh power domain for Maili\n pmdomain: tegra: Add support for multi-socket platforms\n pmdomain: bcm: bcm2835-power: Raise ASB poll timeout to 100us\n pmdomain: mediatek: Add support for MT8196 HFRP DirectCTL domains\n pmdomain: mediatek: Add support for Direct CTL simple power sequence\n pmdomain: mediatek: Respect PD relationships during error cleanup\n dt-bindings: power: mediatek: Add support for MT8196 direct HFRP\n ...\n" }, { "commit": "abea5c349368a683e40fdc2e745099aaaacc89d7", "tree": "dd2f7c3b8283159ae75b859e6b1a4c71e41947ac", "parents": [ "7711f4417f1870e4bfbee1a7d501f789255bc7aa", "18ca2ce160ffff977f28456039f2475a3b500bc0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:23:13 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:23:13 2026 -0700" }, "message": "Merge tag \u0027i2c-7.3-part1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/andi.shyti/linux\n\nPull i2c updates from Andi Shyti:\n \"The main changes are support for shared SCL lines in i2c-gpio, a\n larger qcom-geni update covering tracing and transfer recovery and\n support for R-Car Gen5.\n\n The rest is mostly smaller driver, core and DT binding updates.\n\n Core and helpers:\n - support bus recovery with single-ended GPIOs\n - acpi: clean up resource handling\n - acpi: force ELAN1300 to 100 kHz\n - algo-bit: allow consumers to skip the optional bus test\n\n Drivers:\n - use generic bus frequency definitions in nomadik, octeon-core,\n microchip-corei2c, k1, davinci and pnx\n - i2c-gpio: support multiple buses sharing the same SCL line\n - qup: propagate clock enable failures\n - spacemit: configure SCL timing and clean up clock handling\n - amd-asf: guard against oversized firmware length\n\n qcom-geni:\n - add tracepoints for bus setup, interrupts and errors\n - use dedicated completion events for abort and reset\n - distinguish address and data NACK handling\n - cancel transfers before falling back to abort\n - simplify runtime PM and resource management\n - refactor resource and serial engine initialization\n\n DT bindings:\n - convert Altera bindings to DT schema\n - convert Axxia bindings to DT schema\n\n New support:\n - R-Car Gen5 and R-Car X5H\n - Axiado AX3005\n - Qualcomm Nord SA8797P\n - Qualcomm SA8255p\"\n\n* tag \u0027i2c-7.3-part1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/andi.shyti/linux: (33 commits)\n i2c: core: support recovery for single-ended GPIOs\n i2c: rcar: add R-Car Gen5 support\n dt-bindings: i2c: rcar-i2c: Document R-Car X5H support\n i2c: i2c-gpio: Enhance driver for buses with shared SCL\n i2c: algo: bit: Allow to skip bit test\n i2c: qcom-geni: Add trace events for Qualcomm GENI I2C driver\n i2c: qcom-geni: trace: Add trace events for Qualcomm GENI I2C\n i2c: qup: Propagate clock enable failures\n i2c: qcom-geni: distinguish address-phase and data-phase NACK\n i2c: qcom-geni: use dedicated completions for abort and reset events\n i2c: qcom-geni: use cancel command before abort on transfer timeout\n dt-bindings: i2c: cdns: add Axiado AX3005 I2C variant\n i2c: qcom-geni: Use devm_pm_runtime_enable() for PM management\n dt-bindings: i2c: qcom,sa8255p-geni-i2c: Add compatible for Nord SA8797P\n i2c: nomadik: Use generic definitions for bus frequencies\n i2c: octeon-core: Use generic definitions for bus frequencies\n i2c: microchip-corei2c: Use generic definitions for bus frequencies\n i2c: k1: Use generic definitions for bus frequencies\n i2c: davinci: Use generic definitions for bus frequencies\n i2c: pnx: Use generic definitions for bus frequencies\n ...\n" }, { "commit": "7711f4417f1870e4bfbee1a7d501f789255bc7aa", "tree": "5044643e85b9049adc5675310a37d709d1423797", "parents": [ "d0d82a84c965915ac13d08b726b4cb440d90b122", "146cc263e457ff6055fe7829e4f4f4b0b5d5dd86" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:10:07 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:10:07 2026 -0700" }, "message": "Merge tag \u0027gpio-updates-for-v7.3-rc1-v2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux\n\nPull gpio updates from Bartosz Golaszewski:\n \"GPIO core:\n - extend the gpio-regmap abstraction layer with more features\n allowing users to override configuration setting, translate\n register values and masks and enable/disable interrupts\n - extend GPIO kunit tests with suites verifying probe ordering by\n software node devlink support and software node hogs\n - shrink GPIO kunit initialization code\n - coding style updates (remove commas from sentinels where\n applicable)\n - with all users now converted treewide to using real firmware node\n links for software node GPIO lookup: remove the deprecated\n label-matching mechanism from from GPIO core\n - drop redundant return value check of nonseekable_open() in\n gpiolib-cdev\n - use IRQ trigger helpers where applicable\n\n Driver updates:\n - refactor error paths and logging in gpio-nomadik\n - use more modern interfaces for getting resources in gpio-rockchip,\n gpio-bt8xx and gpio-pca9570\n - add missing MODULE_DEVICE_TABLE() to gpio-sifive and gpio-vf610\n - drop unused FILONOFF macro from gpio-rcar\n - extend build coverage of ioport GPIO drivers with COMPILE_TEST\u003dy\n - only enable the gpio-rtd driver by default with ARCH_REALTEK\u003dy to\n avoid bloating the build\n - refactor coding style in several drivers\n - use correct endianess translation in gpio-pcf85x\n - add wake-up interrupt support to gpio-mvebu\n - apply initial value in direction output setter in gpio-by-pinctrl\n\n Misc:\n - replace linux/gpio.h inclusions treewide with linux/gpio/legacy.h\n which now exports all the deprecated APIs\n - select GPIOLIB_LEGACY in Kconfig where required treewide\n - use software nodes for gpio-keys in MFD drivers\n\n Devicetree bindings:\n - describe the realtek rtd1625 GPIO controller\n - document new models for gpio-pca95xx and gpio-cadence\n - document new property in gpio-rockchip\"\n\n* tag \u0027gpio-updates-for-v7.3-rc1-v2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux: (61 commits)\n gpio: gpio-by-pinctrl: Apply initial value in direction output wrapper\n dt-bindings: gpio: rockchip,gpio-bank: Add rockchip,grf property\n gpio: Use IRQ trigger mask helpers\n gpio: allow COMPILE_TEST for IOPORT drivers\n gpio: realtek: Add driver for Realtek DHC RTD1625 SoC\n gpio: regmap: Add IRQ enable/disable helpers\n gpio: regmap: Add set_config callback\n gpio: regmap: Add value_xlate callback\n gpio: regmap: Add gpio_regmap_operation to extend reg_mask_xlate callback\n gpio: regmap: Order kernel-doc descriptions with the actual appearance\n gpio: regmap: Apply default resource callbacks for regmap IRQ chip\n gpio: regmap: Provide default IRQ resource request and release callbacks\n Revert \"gpio: realtek: Add driver for Realtek DHC RTD1625 SoC\"\n gpib: gpio: replace linux/gpio.h inclusion\n Input: matrix_keyboard - replace linux/gpio.h inclusion\n phy: replace linux/gpio.h inclusions\n pcmcia: replace linux/gpio.h inclusions\n ASoC: replace linux/gpio.h inclusions\n mfd: replace linux/gpio.h inclusions\n sh: replace linux/gpio.h inclusions\n ...\n" }, { "commit": "d0d82a84c965915ac13d08b726b4cb440d90b122", "tree": "4138287f6fcb5deaa141eeb4b57576d289133cc3", "parents": [ "a288cd69f7dea3cd232b721f935c379b74a69814", "a39ac4651e3b9d2ca78b6d3dc2a81a6bd25fa616" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:08:06 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:08:06 2026 -0700" }, "message": "Merge tag \u0027pwrseq-updates-for-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux\n\nPull power sequencing updates from Bartosz Golaszewski:\n \"This a very tiny pull for v7.3 from the power sequencing tree. It only\n contains a handful of updates to the pwrseq-pcie-m2 driver:\n\n - add support for new devices to pwrseq-pcie-m2\n\n - make device matching more fine-grained for cases where the same\n combo chips are wired differently on the M.2 card (and - for\n instance - don\u0027t require serial device creation because they expose\n BT over USB) in pwrseq-pcie-m2\"\n\n* tag \u0027pwrseq-updates-for-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:\n power: sequencing: pcie-m2: Match WCN6855 and WCN7851 UART BT variants by subdevice ID\n power: sequencing: pcie-m2: Add QCA2066 (QCNFA765) BT serdev ID\n power: sequencing: pcie-m2: Add PCI ID for NXP 88W9098 and AW693 Bluetooth\n" }, { "commit": "a288cd69f7dea3cd232b721f935c379b74a69814", "tree": "3e79016fd27060305d9fc71bf68128a8bcb23e5e", "parents": [ "1be05c6afbb2c4052d0a13fc91ddcd89aaee9614", "b4e94b0418ad6dbe62b6ca53369159ed3d9f0d8c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:03:59 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 09:03:59 2026 -0700" }, "message": "Merge tag \u0027pwm/for-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ukleinek/linux\n\nPull pwm updates from Uwe Kleine-König:\n \"A bunch of cleanups - in C and Rust - and a devicetree and driver\n extension for a new SoC variant.\n\n Thanks to Biju Das, Francis Laniel, Guru Das Srinagesh, Markus\n Elfring, Mikko Perttunen, Thierry Reding, and Yi-Wei Wang for their\n changes and further Alexandre Courbot, Benno Lossin, Chen Wang, Geert\n Uytterhoeven, Jon Hunter, Laurent Pinchart, Michal Wilczynski, Mikko\n Perttunen, and Rob Herring for valuable review feedback\"\n\n* tag \u0027pwm/for-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ukleinek/linux:\n pwm: th1520: use vertical import style\n rust: pwm: replace `core::mem::zeroed` with `pin_init::zeroed`\n pwm: rzg2l-gpt: Drop unused rzg2l_gpt_chip parameter from rzg2l_gpt_calculate_prescale()\n pwm: Use seq_putc() calls in pwm_dbg_show()\n pwm: tegra: Add support for Tegra264\n pwm: tegra: Parametrize duty and scale field widths\n pwm: tegra: Modify read/write accessors for multi-register channel\n pwm: tegra: Avoid hard-coded max clock frequency\n pwm: tegra: Prefix driver-local macros and functions\n dt-bindings: pwm: Document Tegra264 controller\n pwm: lpss-pci: Unify coding style of pci_device_id array\n pwm: Unify coding style of of_device_id arrays\n pwm: Unify coding style of acpi_device_id arrays\n pwm: Use named initializers for arrays of acpi_device_id\n pwm: pca9685: Drop unused assignment of acpi_device_id driver data\n pwm: pxa: Depend on OF and simplify accordingly\n pwm: Use named initializers for platform_device_id arrays\n pwm: mc33xs2410: Initialize spi_device_id arrays using member names\n" }, { "commit": "1be05c6afbb2c4052d0a13fc91ddcd89aaee9614", "tree": "86f6afe38778cdb96e85da662dc34510cb73d40b", "parents": [ "e8bf40d154020dd323596933ffaebda7111828fa", "b4d85f863e5a6f1fa01b186001af0f68f4f5239e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 08:59:24 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 08:59:24 2026 -0700" }, "message": "Merge tag \u0027input-for-v7.3-rc0\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input\n\nPull input updates from Dmitry Torokhov:\n\n - A new driver and device tree binding for Imagis ISA1200 haptic motor\n controller\n\n - Improvements to input core opening, closing and inhibiting devices,\n ensuring devices are fully ready before delivering events, deferring\n handler start() until the device is opened, resyncing state on\n uninhibit, and rejecting inhibit requests during unregistration\n\n - Updates to cap11xx capacitive touch driver to support Microchip\n CAP1114, optional hardware reset GPIO handling, and per-chip LED\n constraints\n\n - Fixes for MELFAS MMS114 touchscreen driver hardening incoming data\n parsing, endianness fixes for I2C packet layout, Y-resolution\n configuration, and refactoring to use chip variant descriptors\n\n - Updates for psmouse driver resolving a potential UAF during protocol\n disconnect, cleaning up PNP ID matching, and making use of guard()\n\n - Fix for FocalTech PS/2 protocol to prevent coordinate underflow and\n cursor jumps at boundaries\n\n - A change to Synaptics driver to enable InterTouch (SMBus) mode on\n Dell Inspiron 3521\n\n - Refactoring of PA-RISC keyboard support in gscps2 to supply keymaps\n via software node device properties, removing architecture-specific\n tables from the generic atkbd driver\n\n - Updates to Samsung keypad driver to keep interrupts disabled while\n device is closed, along with wakeup logic cleanups and use of\n pm_runtime_active guards\n\n - Updates to NXP i.MX SNVS power key driver to report press events\n during resume to avoid lost events, and error handling cleanups\n\n - Updated TCA8418 keypad driver enabling overflow mode per hardware\n errata\n\n - Conversion of ROHM BD718x7 and BD71828 PMIC drivers to instantiate\n gpio-keys child devices using software nodes instead of platform data\n (coming from MFD immutable branch)\n\n - Updates to Synaptics RMI4 driver to use touchscreen dimensions from\n platform data when specified\n\n - Firmware update speed optimization for IC Type 0x19 in ELAN I2C\n driver\n\n - A fix to Azoteq IQS5xx driver to validate firmware record spans\n against programmable map size\n\n - Update to Samsung SUR40 contact count based on PixelSense\n specification\n\n - A number of updates to device tree bindings, including TI TPS65217\n power button schema conversion and new compatibles for FocalTech\n FT3D81 and Synaptics S3706B\n\n - Other assorted driver cleanups, style fixes, and conversions to\n modern string and cleanup helpers\n\n* tag \u0027input-for-v7.3-rc0\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input: (61 commits)\n Input: rmi4 - use platform data instead of query, when available\n Input: elan_i2c - sort include statements\n Input: elan_i2c - optimize update speed for IC Type 0x19.\n Input: elan_i2c - use device-id/acpi.h for ACPI IDs\n Input: reject inhibit and uninhibit requests on unregistering devices\n Input: defer handler\u0027s start() until device is opened\n Input: call handler-\u003estart() when uninhibiting device\n Input: clear inhibited flag before re-opening device on uninhibit\n Input: ensure device is ready before delivering events\n Input: gscps2 - supply PA-RISC keyboard keymap via device property\n Input: synaptics_i2c - return 0 explicitly on success\n Input: rmi_smbus - remove conditional return with no effect\n Input: pmic8xxx-keypad - remove conditional return with no effect\n Input: focaltech - use signed coordinates to prevent underflow\n Input: psmouse - use guard() for resource management\n Input: psmouse - modernize PNP ID parsing\n Input: psmouse - clean up locking around disable_work_sync()\n Input: psmouse - fix use-after-free during protocol disconnect\n Input: samsung-keypad - use pm_runtime_active guard\n Input: samsung-keypad - keep interrupt disabled while closed\n ...\n" }, { "commit": "e8bf40d154020dd323596933ffaebda7111828fa", "tree": "b2c80654528fbe8d338111c18c128267d2c69263", "parents": [ "c36a4991e231242c04141536718d47255ec80887", "a58a57a1076f8c5dae0327e3710899478c3be901" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 08:49:43 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 08:49:43 2026 -0700" }, "message": "Merge tag \u0027chrome-platform-firmware-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux\n\nPull chrome platform firmware updates from Tzung-Bi Shih:\n \"Fixes:\n - Don\u0027t map no-map memory regions for CBMEM entries\n - Check bound of coreboot table entries\n\n Cleanups:\n - Fix typo in docs\"\n\n* tag \u0027chrome-platform-firmware-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux:\n firmware: coreboot: Validate table bounds\n firmware: coreboot: Skip no-map CBMEM entries\n docs: ABI: testing: Fix typo\n" }, { "commit": "c36a4991e231242c04141536718d47255ec80887", "tree": "9fa08a84624b5b2a7bdd7690f1b11e51f17747e6", "parents": [ "bd5f485f3f026225b86573e559af0b7254ef4184", "81c1c0cb68a13d14dd1bbffcf4c49e26f9a6478c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 08:27:06 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 19 08:27:06 2026 -0700" }, "message": "Merge tag \u0027chrome-platform-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux\n\nPull chrome platform updates from Tzung-Bi Shih:\n \"New:\n - Add post_power_on_delay_ms for Hana in of_hw_prober\n\n Improvements:\n - Use dumb trackpad prober for Spherion in of_hw_prober\n\n Fixes:\n - Check bound of firmware-reported data in cros_ec_sensorhub and\n cros_ec_typec\n - Fix memory overread in cros_ec_sensorhub\n - Fix resource leak in cros_ec_debugfs\n - Clamp payload length for LIGHTBAR_CMD_SET_PROGRAM_EX in\n cros_ec_lightbar\n\n Cleanups:\n - Drop unused platform_device_id driver data\n - Remove redundant log\"\n\n* tag \u0027chrome-platform-v7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux:\n platform/chrome: of_hw_prober: Add delay for hana trackpads\n platform/chrome: lightbar: Limit payload to max packet size\n platform/chrome: cros_ec_debugfs: Unregister panic notifier\n platform/chrome: cros_ec_debugfs: Clean up console log on probe failure\n platform/chrome: cros_ec: Remove redundant dev_err()\n platform/chrome: sensorhub: Fix dropped timestamp events and log spam\n platform/chrome: sensorhub: Fix memory overread in ring handler\n platform/chrome: cros_ec_typec: Reject out-of-bounds PD cap count\n platform/chrome: of_hw_prober: Use dumb trackpad prober for Spherion\n platform/chrome: Drop unused assignment of platform_device_id driver data\n platform/chrome: sensorhub: Bound the EC-reported sensor number\n" }, { "commit": "b282418bc366194677eafd1dad180d92254586ac", "tree": "78da76aea2cdc21f2bcef31c73992032fa7ab6e6", "parents": [ "675087c762f95c498524164de7417f9f77d3ed15" ], "author": { "name": "Eugenio Pérez", "email": "eperezma@redhat.com", "time": "Tue Jul 07 14:33:44 2026 +0200" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed Aug 19 06:38:48 2026 -0400" }, "message": "vduse: Add suspend\n\nImplement suspend operation for vduse devices, so vhost-vdpa will offer\nthat backend feature and userspace can effectively suspend the device.\n\nThis is a must before get virtqueue indexes (base) for live migration,\nsince the device could modify them after userland gets them.\n\nThis patch does not implement resume, so VMM resets the whole device\nto recover from a live migration failure. Resume optimization can be\nimplemented on top of these patches, as other vDPA devices have done in\nthe past.\n\nSigned-off-by: Eugenio Pérez \u003ceperezma@redhat.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nMessage-ID: \u003c20260707123344.244575-3-eperezma@redhat.com\u003e\n" }, { "commit": "675087c762f95c498524164de7417f9f77d3ed15", "tree": "67eb6bccfdc8758c39d0ed88097cfee5c5cc8011", "parents": [ "4c318d91cc60a0c2c94bcb6db2630baed43e9387" ], "author": { "name": "Eugenio Pérez", "email": "eperezma@redhat.com", "time": "Tue Jul 07 14:33:43 2026 +0200" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed Aug 19 06:38:48 2026 -0400" }, "message": "vduse: do not take rwsem at reset work flush\n\nNext patches need to check suspend flag at this work item, and the\nrwlock is used to protect the suspend flag update. If the work takes\nthe rwlock too it will produce a deadlock.\n\nMake flushing work do nothing when called by de-initializing everything:\nvq-\u003eready, vq-\u003ekickfd, vq-\u003ecb.callback.\n\nSigned-off-by: Eugenio Pérez \u003ceperezma@redhat.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nMessage-ID: \u003c20260707123344.244575-2-eperezma@redhat.com\u003e\n" }, { "commit": "4c318d91cc60a0c2c94bcb6db2630baed43e9387", "tree": "9e1830f489602f71aa9e0799aa1ac3709722b2b1", "parents": [ "3b441820cb61ac1137dd4b336adfb4892cda4233" ], "author": { "name": "Eugenio Pérez", "email": "eperezma@redhat.com", "time": "Tue Jul 07 14:25:02 2026 +0200" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed Aug 19 06:38:48 2026 -0400" }, "message": "vduse: add F_QUEUE_READY feature\n\nAdd the VDUSE_F_QUEUE_READY feature flag. This allows the kernel module\nto explicitly signal userspace when a specific virtqueue has been\nenabled.\n\nIn scenarios like Live Migration of VirtIO net devices, the dataplane\nstarts after the control virtqueue allowing QEMU to apply configuration\nin the destination device.\n\nSigned-off-by: Eugenio Pérez \u003ceperezma@redhat.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nMessage-ID: \u003c20260707122502.239022-5-eperezma@redhat.com\u003e\n" }, { "commit": "3b441820cb61ac1137dd4b336adfb4892cda4233", "tree": "786354464435abb1e31580fb40256f6403de67cd", "parents": [ "a596238c2ab6944861404dc597133ffd4ad062d5" ], "author": { "name": "Eugenio Pérez", "email": "eperezma@redhat.com", "time": "Tue Jul 07 14:25:01 2026 +0200" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed Aug 19 06:38:48 2026 -0400" }, "message": "vduse: add VDUSE_SET_FEATURES ioctl\n\nAdd an ioctl to allow VDUSE instances to set the VDUSE features\nsupported by the userland VDUSE instance.\n\nSigned-off-by: Eugenio Pérez \u003ceperezma@redhat.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nMessage-ID: \u003c20260707122502.239022-4-eperezma@redhat.com\u003e\n" }, { "commit": "a596238c2ab6944861404dc597133ffd4ad062d5", "tree": "0e7d5dfabba4005ddf7e6a29bd886ecd64b97406", "parents": [ "0ff906166f00ac7b2ccc36c57be6ad4be0dd9e90" ], "author": { "name": "Eugenio Pérez", "email": "eperezma@redhat.com", "time": "Tue Jul 07 14:25:00 2026 +0200" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed Aug 19 06:38:48 2026 -0400" }, "message": "vduse: add VDUSE_GET_FEATURES ioctl\n\nAdd an ioctl to allow VDUSE instances to query the available features\nsupported by the kernel module.\n\nSigned-off-by: Eugenio Pérez \u003ceperezma@redhat.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nMessage-ID: \u003c20260707122502.239022-3-eperezma@redhat.com\u003e\n" }, { "commit": "0ff906166f00ac7b2ccc36c57be6ad4be0dd9e90", "tree": "76266f6d6bca89fb27666fe78b29b59dc2d9f9c9", "parents": [ "05d32474ab547fb097cfe3f3eb00a2aa7cdf4067" ], "author": { "name": "Eugenio Pérez", "email": "eperezma@redhat.com", "time": "Tue Jul 07 14:24:59 2026 +0200" }, "committer": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Wed Aug 19 06:38:48 2026 -0400" }, "message": "vduse: store control device pointer\n\nThis helps log the errors in next patches. The alternative is to\nperform a linear search for it with class_find_device_by_devt(class, devt),\nas device_destroy do for cleaning.\n\nSigned-off-by: Eugenio Pérez \u003ceperezma@redhat.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nMessage-ID: \u003c20260707122502.239022-2-eperezma@redhat.com\u003e\n" } ], "next": "05d32474ab547fb097cfe3f3eb00a2aa7cdf4067" }