)]}' { "log": [ { "commit": "bd5f485f3f026225b86573e559af0b7254ef4184", "tree": "92b9cabbbefe9b8cee634769676b9b97238cdfc7", "parents": [ "6eb1ea5ff48d647deb67c58b7a0e97bbf4de088a", "acf1efa9485fda84f332c8fbbe7b4453dbec183a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:50:40 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:50:40 2026 -0700" }, "message": "Merge tag \u0027soc-arm-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc\n\nPull ARM SoC platform updates from Arnd Bergmann:\n \"The 32-bit Arm platforms are a bit more interesting this time: I\n refreshed an earlier series to mark code as deprecated that does have\n the tendency of getting in the way of cleanups and new features but\n has close to zero users. Among these are:\n\n - 22 of the remaining 28 legacy board files that predate the current\n devicetree based descriptions, using old chips from Intel and\n Marvell. The remaining six board files are for TI OMAP1 and Samsung\n s3c64xx chips and all still have known users.\n\n - support for Cortex-M3/M4/M7 and ARM1136r0 CPU cores and the 25\n machines based on these. These all use devicetree but the CPU\n support causes disproportional work. Most of them are just\n reference boards, the notable exceptions being the Nokia N800/N810\n tablet and the Buglabs BUG platform.\n\n - be8, be32, oabi and iwmmxt userspace binaries, which were mostly\n associated with the platforms now scheduled for removal and are\n increasingly problematic to support with modern toolchains.\n\n Nothing is actually removed at this point, to ensure that any\n remaining users continue to have the 7.3-LTS for a while longer.\n Patches for removal are currently being tested.\n\n Other updates include a continued work to convert GPIO number based\n interfaces to descriptors, a patch to restore little-endian mode on\n the one Arm platform (ixp4xx) that only worked in big-endian mode\n recently, and some minor cleanups and bugfixes\"\n\n* tag \u0027soc-arm-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (41 commits)\n MAINTAINERS: Drop redundant lists from various Samsung entries\n ARM: tegra: Replace __ASSEMBLY__ with __ASSEMBLER__\n ARM: tegra: Fix OF node reference leaks in IRQ init\n ARM: lpc32xx: remove a few manually populated OF devices\n ARM: lpc32xx: only run SoC init on LPC32xx hardware\n firmware: imx: scu: manage mailbox channels and global handle\n ARM: sa1100: h3xxx: convert gpio-keys to use software nodes\n ARM: sa1100: collie: convert gpio-keys to use software nodes\n ARM: sa1100: assabet: convert gpio-keys to use software nodes\n gpio: sa1100: register software node for GPIO controller\n ARM: ixp4xx: Relax endianness\n ARM: replace linux/gpio.h inclusions\n soc: imx9: devm_kasprintf error handling\n ARM: mark mv78xx0 support as deprecated\n ARM: mark axxia platform as deprecated\n ARM: mark Cortex-M3/M4/M7 based boards as deprecated\n ARM: mark footbridge as deprecated\n ARM: mark RiscPC as deprecated\n ARM: mark mach-sa1100 as deprecated\n ARM: orion5x: mark all board files as deprecated\n ...\n" }, { "commit": "6eb1ea5ff48d647deb67c58b7a0e97bbf4de088a", "tree": "390fe17f867aa76760466c2b4e25cb8abc7fd057", "parents": [ "368cf60c36a3a311474de08e52dc6314df3b06ce", "373eab9eec316c315e884e9cb520c6d8c3236e0f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:47:08 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:47:08 2026 -0700" }, "message": "Merge tag \u0027soc-defconfig-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc\n\nPull SoC defconfig updates from Arnd Bergmann:\n \"Just the usual updates to the main defconfig files as well as the\n omap2 specific one, to enable more loadable modules for better default\n hardware support\"\n\n* tag \u0027soc-defconfig-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc:\n riscv: defconfig: thead: enable PCA953X GPIO driver\n arm64: defconfig: Enable drivers for BeagleBadge\n arm64: defconfig: Enable Qualcomm BAM-DMUX WWAN driver\n arm64: defconfig: Enable Sound DMIC driver\n arm: omap2plus_defconfig: Enable multi-LED\n arm64: defconfig: Enable Qualcomm reference device EC driver\n ARM: omap2plus_defconfig: enable things required by iwd\n arm: multi_v7_defconfig: Enable BRIDGE and DP83848_PHY for TI AM57xx, AM437x and AM335x\n arm64: defconfig: Enable Allwinner LRADC input driver\n" }, { "commit": "368cf60c36a3a311474de08e52dc6314df3b06ce", "tree": "769debd6b400d2cd7d6f9f9266a00d2331582371", "parents": [ "cbe8aadf1551f3f4a853f24b3b96d9f0eea49c53", "aed89a05a736d2570a514f00c2ab83388662b271" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:40:29 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:40:29 2026 -0700" }, "message": "Merge tag \u0027soc-drivers-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc\n\nPull SoC driver updates from Arnd Bergmann:\n \"The SoC driver changes once more consist of many small fixes and\n cleanups, that are to a large part the result of automated testing.\n\n On platform specific drivers, this includes SoC specific code for\n xilinx, freescale/nxp, qualcomm, TI, aspeed, omap, tegra, samsung,\n rockchip, renesas, ixp4xx. In firmware drivers, we see a similar\n picture for SCMI and qcomtee.\n\n Aside from these, we see actual new hardware support in a few areas:\n\n - The Apple platform gets a new driver for low power states\n\n - Updates to Qualcomm platform drivers add several new hardware\n specific features and additional SoCs.\n\n - Amlogic SoC support for A1 and T7 is added\n\n - The Mediatek MMSYS driver is refactored as a cleanup\"\n\n* tag \u0027soc-drivers-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (157 commits)\n soc: qcom: make QCOM_PDR_MSG selectable\n soc: qcom: ubwc: Fix missing include\n soc: qcom: ubwc: Fix link error when QCOM_SMEM\u003dn\n media: iris: Guard the QCOM_UBWC_CONFIG select with QCOM_SMEM\n drm/msm: Guard the QCOM_UBWC_CONFIG select with QCOM_SMEM\n dt-bindings: arm: qcom,ids: Add SoC ID for Snapdragon SDM 850\n firmware: xilinx: Clear firmware notifiers across kexec transitions\n firmware: xilinx: Release all peripheral devices from firmware\n firmware: xilinx: Add support to clear EL3 PM state\n firmware: xilinx: Propagate actual error from feature check\n firmware: xilinx: Use TF-A feature check for TF-A-specific APIs\n bus: fsl-mc: drop unused assignment of acpi_device_id::driver_data\n soc: fsl: qe: check platform_driver_register() in qe_ic_of_init()\n phy: lynx-10g: use RCW override procedure for dynamic protocol change\n soc: fsl: guts: implement the RCW override procedure\n dt-bindings: fsl: layerscape-dcfg: define DCFG_DCSR region\n soc: fsl: guts: make fsl_soc_data available after fsl_guts_init()\n soc: fsl: guts: make it easier to determine on which SoC we are running\n soc: fsl: guts: add a central fsl_guts_read() function\n soc: fsl: guts: add a global structure to hold state\n ...\n" }, { "commit": "cbe8aadf1551f3f4a853f24b3b96d9f0eea49c53", "tree": "9a5dd456ffd84791902b1b89734aca578149bb3e", "parents": [ "ca58a3dde621067333f481547db4afc587d7708a", "dfda289ad743d037f33f9e3796e3fd40ca23374c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:36:32 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:36:32 2026 -0700" }, "message": "Merge tag \u0027soc-dt-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc\n\nPull SoC devicetree updates from Arnd Bergmann:\n \"There are two new mystery SoCs with very little public information\n about them so far, Qualcomm\u0027s Cortex-A78C based \"Shikra\" and Altera\u0027s\n Cortex-A720 based Agilex72 SOCFPGA.\n\n We have also gained support for a couple of SoCs from the 2023/2024\n timeframe that have been in the making for a while:\n\n - The Apple platform gains support for M3 Pro, Max and Ultra SoC,\n following the basic M3 support from 7.2.\n\n - Samsung Exynos 1580 is a high-end mobile phone SoC from 2024\n\n - Canaan K230 is a RISC-V based 64-bit AI SoC, based on the earlier\n K210 chip\n\n - Sophgo SG2000 is a mixed Arm/RISC-V chip that was already supported\n using the Arm core but is now also added for RISC-V along with\n several other variants of the cv18xx series\n\n In terms of newly added machines we have reference platforms for all\n the chips above, plus\n\n - Only four 32-bit Arm boards: two older phones and two older\n industrial/embedded boards; using Allwinner, Qualcomm and Rockchip\n SoCs\n\n - Three laptops und three phones using Qualcomm SoCs\n\n - Ten 64-bit Rockchips based single-board computers, along with one\n NAS box and a game console.\n\n - Seven industrial/embedded boards and modules using NXP i.MX8/9\n SoCs.\n\n - A Lenovo desktop box based on NVIDIA Jetson Xavier NX\n\n - A few more single-board computers based on Allwinnner A133,\n Spacemit K1 and TI AM62.\n\n As usual, there is a constant stream of minor cleanups and fixes\n towards addressing the \u0027dt-check-style --mode strict\u0027 warnings, and\n everyone is expected to address those warnings for new submissions\n now\"\n\n* tag \u0027soc-dt-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (864 commits)\n Revert \"riscv: dts: spacemit: k3: add i2s0-i2s5 nodes\"\n Documentation/process: maintainer-soc: Mention expectation about dt-check-style\n ARM: dts: helios4: add SATA regulator supplies\n ARM: dts: helios4: add vcc-supply to GPIO expander\n ARM: dts: helios4: add vcc-supply to EEPROM\n arm64: dts: turris-mox: fix usb3 phys\n arm64: dts: ti: Correct white-space style\n arm64: dts: ti: k3-am64: Fix MDIO clock reference for ICSSG0 node\n arm64: dts: qcom: talos-evk: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: sa8155p-adp: Fix swapped USB and UFS QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: eliza-mtp: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: lemans: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: glymur: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: kaanapali: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: sar2130p: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: sm8750: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: sm8650: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: purwa: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: hamoa: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n arm64: dts: qcom: sc8180x: Fix swapped USB QMP PHY vdda-phy/vdda-pll supplies\n ...\n" }, { "commit": "ca58a3dde621067333f481547db4afc587d7708a", "tree": "bc17c2c49ea1cdc163fd2f2b8aaa2fae21d64c32", "parents": [ "d10e148b27cd746cc5c27e30686153eb97e8d8d0", "b39c748d9db0c6f3ada5e7cef7f56415827d5f81" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:33:32 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:33:32 2026 -0700" }, "message": "Merge tag \u0027alpha-for-v7.3-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/lindholm/alpha\n\nPull alpha updates from Magnus Lindholm:\n \"This contains two fixes for Alpha floating-point exception handling,\n two clang-related fixes, the preparatory changes from the\n generic-entry series, an interrupt-entry lockdep fix, two Marvel/EV7\n IRQ fixes, an RTC fix, and one header cleanup.\n\n The generic-entry preparation adds regset-based ptrace and core dumps,\n ARCH_STACKWALK and lockdep hardirq-state tracking. These changes are\n useful independently and enable previously missing debugging\n facilities on Alpha.\n\n The final patch switching Alpha to GENERIC_ENTRY is intentionally not\n included in this pull request. I am deferring that change to allow\n further testing and to reduce the risk of conflicts with ongoing\n entry-path work elsewhere in the kernel\"\n\n* tag \u0027alpha-for-v7.3-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/lindholm/alpha:\n alpha: read $gp and $sp explicitly for clang\n alpha: pass -Wa,-mev6 only when using GNU as\n alpha: annotate hardirqs-off on IPL 7 interrupt entry\n alpha: run the remote RTC access in a worker, not an IPI callback\n alpha: don\u0027t leak hardware-fabricated FP exception bits to user space\n alpha: fix ieee_swcr_to_fpcr setting FPCR_DNOD unconditionally\n alpha: enable lockdep hardirq state tracking\n alpha: use raw spinlocks for low-level platform locks\n alpha: provide ftrace return address support for lockdep\n alpha: make irqflags helpers operate on IPL state\n alpha: add ARCH_STACKWALK-based stacktrace support\n alpha: enable regset-based ptrace and core dumps\n alpha: marvel: Fix lock ordering in init_io7_irqs()\n alpha: marvel: Fix irq_set_status_flags to use correct IRQ number\n alpha: remove unnecessary architecture-specific \u003casm/device.h\u003e\n" }, { "commit": "d10e148b27cd746cc5c27e30686153eb97e8d8d0", "tree": "bd094b772a0a05ee51db0da3d0bf6621f0c7d3c7", "parents": [ "5808ac188949d7b49ed9e51f1ffcd096cac57006", "abb81e5ce7d995baa41556b8125fa59e28ba3be8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:32:34 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:32:34 2026 -0700" }, "message": "Merge tag \u0027csky-for-linus-7.3\u0027 of https://github.com/c-sky/csky-linux\n\nPull csky update from Guo Ren:\n\n - Fix a4/a5 restoration in syscall trace path\n\n* tag \u0027csky-for-linus-7.3\u0027 of https://github.com/c-sky/csky-linux:\n csky: Fix a4/a5 restoration in syscall trace path\n" }, { "commit": "5808ac188949d7b49ed9e51f1ffcd096cac57006", "tree": "6a2a5d5b58e5bfce7a36abe098a26641bcedcd07", "parents": [ "bbd0571269b5a17bb0685fc30ea0b89b950801f5", "6cc3a72caa6eb118c38f887d4608b9adb80e90e8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:09:42 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 19:09:42 2026 -0700" }, "message": "Merge tag \u0027x86_cpu_for_v7.3_rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 cpuid updates from Borislav Petkov:\n\n - Get rid of static_cpu_has() - one less API to care about testing CPU\n features\n\n - Unify the handling of CPU core types (performance, efficient, etc) by\n mapping the vendor-specific types to Linux ones\n\n - Continuation of the work of Ahmed Darwish to centralize CPUID leaf\n representation\n\n* tag \u0027x86_cpu_for_v7.3_rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/CPU: Rename struct cpuid_read_output to struct cpuid_output\n x86/cpu/scattered: Sort it properly\n x86/cpu: Use parsed CPUID(0x1)\n x86/lib: Add CPUID(0x1) family and model calculation\n x86/cpu: Use parsed CPUID(0x0)\n x86/cpu/transmeta: Rescan CPUID(0x1) after modifying capabilities\n x86/topology: Add TOPO_CPU_TYPE_LOW_POWER\n x86/topology: Name the AMD core-type values\n x86/topo: Map vendor CPU types to generic Linux such types\n x86/bugs: Don\u0027t use cpu-type matching in cpu_vuln_blacklist\n x86/cpu: Hide and rename static_cpu_has()\n" }, { "commit": "bbd0571269b5a17bb0685fc30ea0b89b950801f5", "tree": "0b6c88b0e743036c83cf6caa942b43cd9fefe378", "parents": [ "94845e2929701018100389b3554dc54d1eee0c70", "10b45e63a5b010be59ecbd2e437744adb12f838a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 18:34:58 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 18:34:58 2026 -0700" }, "message": "Merge tag \u0027x86_cleanups_for_v7.3_rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 cleanups from Borislav Petkov:\n\n - The usual pile of smallish cleanups and fixlets all over the place\n\n* tag \u0027x86_cleanups_for_v7.3_rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/cpu: Remove unnecessary __maybe_unused annotations\n x86/msr: Document the I/O-like write semantics in the msr driver\n x86/apic: Ensure ICR register write value is handled as 32 bits\n x86/boot/compressed/head_64.S: Clean up SEV-related comments\n Documentation/arch/x86/amd-memory-encryption.rst: Fix typo\n x86/platform/quark: Fix kernel-doc warnings in imr.c\n x86/ras: Move contents from arch/x86/ras/Kconfig into drivers/ras/Kconfig\n x86/cpu: Move intel_get_platform_id() to cpu/intel.c\n x86/mm: Fix typo in comment\n x86/fpu: Fix kernel-doc formatting above fpu_enable_guest_xfd_features()\n x86/cfi: Use symmetric SYM_START and SYM_END in __CFI_TYPE()\n x86/cfi: Add __init_or_module annotations for fineibt\n" }, { "commit": "94845e2929701018100389b3554dc54d1eee0c70", "tree": "2d5e3dce3c03ee956f47b89edf25e42280a83515", "parents": [ "ccb9331e132a0554f6efb18140eae78bc32b0ee4", "854db793d4073087fa799d886ed84a342e3f484a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 18:32:24 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 18:32:24 2026 -0700" }, "message": "Merge tag \u0027x86_cache_for_v7.3_rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 resource control updates from Borislav Petkov:\n\n - How refreshing: no new features but a whole pile of fixes to more or\n less serious issues reported by Sashiko along with miscellaneous\n cleanups all over the place. All except one by Reinette Chatre, the\n one by Tony Luck.\n\n* tag \u0027x86_cache_for_v7.3_rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n fs/resctrl: Inform user space when status buffer overflowed\n fs/resctrl: Communicate resource group deleted error via last_cmd_status\n fs/resctrl: Add last_cmd_status support for writes to max_threshold_occupancy\n fs/resctrl: Change last_cmd_status custom during input parsing\n fs/resctrl: Use accurate and symmetric exit flows\n fs/resctrl: Pass error reading event through to user space\n fs/resctrl: Use accurate type for rdt_resource::rid\n fs/resctrl: Change pattern used to track number of entries in enum resctrl_conf_type\n x86/resctrl: Protect against bad shift\n fs/resctrl: Use correct format specifier for printing error pointers\n fs/resctrl: Fix UAF from worker threads when domains are removed\n x86/resctrl: Ensure domain fully initialized before placed on RCU list\n fs/resctrl: Prevent deadlock and use-after-free in info file handlers\n fs/resctrl: Prevent use-after-free in rdtgroup_kn_put()\n fs/resctrl: Fix deadlock on errors during mount\n fs/resctrl: Move functions to avoid forward references in subsequent fixes\n x86,fs/resctrl: Document safe RCU list traversal\n" }, { "commit": "ccb9331e132a0554f6efb18140eae78bc32b0ee4", "tree": "9af27985432a7ac6214fecb66cc5eb911eec0ad1", "parents": [ "b960327ab340b54f8e20f3af5d0751e2941b6675", "4a73b93ce489a58c9a2aac4a5f5cd8cfa58ced64" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 18:11:25 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 18:11:25 2026 -0700" }, "message": "Merge tag \u0027x86_alternatives_for_v7.3_rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 alternatives update from Borislav Petkov:\n\n - Remove the smp_locks alternatives machinery which was used to patch\n out lock prefixes when running a SMP kernel on a uniprocessor machine\n\n* tag \u0027x86_alternatives_for_v7.3_rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/alternative: Drop smp_locks glue\n" }, { "commit": "b960327ab340b54f8e20f3af5d0751e2941b6675", "tree": "0468f46f04b86b9d40f6080e5b78faaf01037d89", "parents": [ "b126f6f5940f614cc06bba61c77dec2d75e8ceaf", "b0fa123835c147e77b9a967de76dc899062bd405" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 18:09:42 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 18:09:42 2026 -0700" }, "message": "Merge tag \u0027ras_core_for_v7.3_rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 MCE update from Borislav Petkov:\n\n - Add mce\u003dprint_all to the mce\u003d kernel cmdline params documentation\n\n* tag \u0027ras_core_for_v7.3_rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/mce: Document the mce\u003dprint_all command line parameter\n" }, { "commit": "b126f6f5940f614cc06bba61c77dec2d75e8ceaf", "tree": "7c52ad5a911a4634d20e7639ccc7c9285cbd7765", "parents": [ "2385c8b47b7f64ccf24404191cef30bdd8663ed7", "dee87e09b0dd63da9b1e1876167ccae37842dfd0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 17:44:09 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 17:44:09 2026 -0700" }, "message": "Merge tag \u0027x86_mm_for_7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 mm updates from Dave Hansen:\n \"There are two things in here of note. First, there are fixes to reduce\n the time that the TLB flushing code sits around with interrupts\n disabled. It should be well tested, but it\u0027s certainly something to\n keep an eye on.\n\n Second is a little back-and-forth in the interactions between the\n set_memory*() locking and DEBUG_PAGEALLOC. There was a slightly\n painful re-discovery of why DEBUG_PAGEALLOC has special locking rules,\n but the code did end up better for it.\n\n Summary:\n\n - Fix pkey_alloc() return value when pkeys are not supported\n\n - Allow preemption during IPI completion waiting to improve wakeup\n latency\n\n - Convert more x86 page table code to ptdescs\n\n - Fixes and cleanups for set_memory*() locking, mostly around large\n pages\"\n\n* tag \u0027x86_mm_for_7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/pkeys: Fix pkey_alloc() return value when pkeys are not supported\n x86/mm: Fix and document DEBUG_PAGEALLOC\n x86/mm: Use guard() for pgd_lock\n x86/mm: Use guard() in cpa_collapse_large_pages()\n x86/mm: Re-enable preemption before flush_tlb_multi()\n x86/kvm: Disable preemption in kvm_flush_tlb_multi()\n x86/mm: Move flush_tlb_info back to the stack\n x86/mm: Cap flush_tlb_info alignment at 64 bytes\n x86/mm: Factor out flush_tlb_info initialization\n x86/mm/pat: Take cpa_lock around large-page collapse\n x86/mm/pat: Don\u0027t gate cpa_lock on debug_pagealloc_enabled()\n x86/xen: Convert xen_mm_unpin_all() to ptdescs\n x86/xen: Convert xen_mm_pin_all() to ptdescs\n x86/mm: Convert pgd_page_get_mm() to ptdescs\n x86/mm: Convert sync_global_pgds_l4() to ptdescs\n x86/mm: Convert sync_global_pgds_l5() to ptdescs\n x86/mm: Convert arch_sync_kernel_mappings() to ptdescs\n x86/mm/pat: Convert collapse_pmd_page() to ptdescs\n x86/mm/pat: Convert __set_pmd_pte() to ptdescs\n x86/mm/pat: Use IS_ENABLED() instead of ifdef\n" }, { "commit": "2385c8b47b7f64ccf24404191cef30bdd8663ed7", "tree": "d240a8ebd291cc6955823acf4b9d0181f21d7e00", "parents": [ "058f2c4b7533365d99d4868fb29a68a5b62ecf5a", "941370fc93cc3474e26811f4d3b062903eefe2cf" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 17:41:36 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 17:41:36 2026 -0700" }, "message": "Merge tag \u0027x86_tdx_for_7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 TDX updates from Dave Hansen:\n \"TDX guests cause #VE exceptions whenever they do port I/O. The guest\n then does some instruction decoding and makes a call up to the host\n for help. That decoding had a couple of bugs.\n\n Fix those bugs. Use an existing and now shared KVM helper for one of\n them\"\n\n* tag \u0027x86_tdx_for_7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/tdx: Fix zero-extension for 32-bit port I/O\n x86/insn-eval: Move assign_register() out of KVM as insn_assign_reg()\n x86/tdx: Fix off-by-one in port I/O handling\n" }, { "commit": "058f2c4b7533365d99d4868fb29a68a5b62ecf5a", "tree": "1862e2e30990f7471b6ce508d3e75c4ed8ca99b3", "parents": [ "23eb7901811a06805ec7d6e766dd82f083d03204", "977991fdc43dbe2f3f473699bbe45207969f020f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 17:40:06 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 17:40:06 2026 -0700" }, "message": "Merge tag \u0027x86_misc_for_7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull misc x86 updates from Dave Hansen:\n \"Just adding some documentation and removing a super stale comment\n about an LED driver that was moved long ago:\n\n - Document the intricacies of GS context switching\n\n - Remove old TODO message about Geode LED driver\"\n\n* tag \u0027x86_misc_for_7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/platform/geode: Remove old TODO message about leds-net5501.c\n Documentation/x86: Document the intricacies of GS context switching\n" }, { "commit": "23eb7901811a06805ec7d6e766dd82f083d03204", "tree": "30d78d8dd2aabfcb753746a3f774a249ae38748c", "parents": [ "104a813376837da3b9651457d6fdc99596257fba", "dab01c597f6bd40e0efe7da967b8374ca1971b79" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 17:38:30 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 17:38:30 2026 -0700" }, "message": "Merge tag \u0027x86_entry_for_7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 entry update from Dave Hansen:\n \"Fix frame pointer unwinder when it encounters FRED stack frames\"\n\n* tag \u0027x86_entry_for_7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/entry/fred: Encode frame pointer on entry\n" }, { "commit": "104a813376837da3b9651457d6fdc99596257fba", "tree": "282c3904ed84cd479c3d7d472c77979938bb0967", "parents": [ "3b4128b9f374b4219eb716f4ad8a307bc7eb3d84", "92e785d887a8a61a897289f2b342f754cf02551b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 16:56:25 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 16:56:25 2026 -0700" }, "message": "Merge tag \u0027timers-vdso-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull VDSO updates from Thomas Gleixner:\n\n - Consolidate the VDSO datastore further and provide support for\n mlock_all() and prefaulting.\n\n - Provide 32-bit legacy time related functionality only if\n CONFIG_COMPAT_32BIT_TIME is enabled. The config switch exists,\n but architecture code still exposes the legacy functionality even\n disabled.\n\n Clean this up by adding the missing guards and validating at build\n time that the VDSO is legacy free if disabled.\n\n - Consolidate the VDSO related config options in core and drivers,\n which removes some non-sensical dependencies and quite an amount of\n #ifdeffery.\n\n - Clean up the PAGE_SIZE definition maze\n\n* tag \u0027timers-vdso-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (30 commits)\n random: vDSO: Drop custom PAGE_SIZE definitions\n LoongArch: Remove CONFIG_GENERIC_GETTIMEOFDAY ifdeffery\n clocksource/drivers/timer-riscv: Remove CONFIG_GENERIC_GETTIMEOFDAY ifdeffery\n clocksource/drivers/arm_arch_timer: Remove CONFIG_GENERIC_GETTIMEOFDAY ifdeffery\n clocksource/drivers/mips-gic-timer: Remove CONFIG_GENERIC_GETTIMEOFDAY ifdeffery\n MIPS: csrc-r4k: Remove CONFIG_GENERIC_GETTIMEOFDAY ifdeffery\n vDSO: Make clockmode constants available without CONFIG_GENERIC_GETTIMEOFDAY\n kbuild: Support generated asm-headers in subdirectories\n vdso: Rename HAVE_GENERIC_VDSO to VDSO_DATASTORE\n vdso: Drop HAVE_GENERIC_VDSO from architecture kconfig files\n vdso: Automatically select HAVE_GENERIC_VDSO if necessary\n MIPS: vdso: Stop using CONFIG_HAVE_GENERIC_VDSO\n vdso: Remove the dependency on HAVE_GENERIC_VDSO from ARCH_HAS_VDSO_ARCH_DATA\n futex: Remove dependency on HAVE_GENERIC_VDSO from FUTEX_ROBUST_UNLOCK\n vdso/gettimeofday: Verify COMPAT_32BIT_TIME interactions\n sparc: vdso: Respect COMPAT_32BIT_TIME\n MIPS: VDSO: Respect COMPAT_32BIT_TIME\n powerpc/vdso: Respect COMPAT_32BIT_TIME\n ARM: VDSO: Respect COMPAT_32BIT_TIME\n arm64: vdso32: Respect COMPAT_32BIT_TIME\n ...\n" }, { "commit": "3b4128b9f374b4219eb716f4ad8a307bc7eb3d84", "tree": "e9489db7a4ab524e7849fb70a3cc47a08700e1ad", "parents": [ "b0239dd672306ad242545f793132938847f17e53", "4fa377c19e111c539a530a8200996b911ceff9ff" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 16:23:56 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 16:23:56 2026 -0700" }, "message": "Merge tag \u0027timers-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull timer and timekeeping core updates from Thomas Gleixner:\n\n - Fix a subtly inconsistency in the timekeeping code, which fails to\n account for the monotonicity adjustment in ntp_error.\n\n For small changes of the clocksource multiplicator (+/-1) which are\n typically used by the NTP PLL this is hard to observe. But for larger\n adjustments, e.g. caused by a direct frequency setting through\n adjtimex() the one-time uncompensated offset is significant.\n\n Cure this by adjusting ntp_error with the resulting offset so that\n the discrepancy is smoothed away over time\n\n - Make tick length calculations correct in NTP.\n\n The timekeeping core takes the quantisation of the clocksource into\n account when calculating the tick length to compensate for the\n deviation of the nominal NTP_INTERVAL_LENGTH.\n\n While timekeeping gets this right, NTP is not aware of that, which\n means it operates on the nominal value and not on the actual value\n which is determined by the clock source frequency. The rounding of a\n coarse clocksource like the ACPI PM timer results in a +127 PPM\n deviation.\n\n Cure this by exposing the deviation to the NTP code so that it can\n operate on the same data as the timekeeping core. This is purely\n kernel internal. User space still sees the nominal tick lenght via\n adjtimex().\n\n - The accuracy of the NTP adjustments is fairly approximate as the code\n assumes that the invocations are precisely in NTP interval frequency\n ticks and the final adjustment can over and under-run.\n\n Cure this by adjusting ntp_error by the intended skew on each tick to\n achieve the desired rate.\n\n - Handle the two competing skews of time offset and time adjustment\n correctly by calculating the conflict portion between the skews and\n adjusting both accordingly.\n\n - A set of updates and improvements for the selftests\n\n - The usual small fixes and improvements all over the place\n\n* tag \u0027timers-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (58 commits)\n selftests: timers: nsleep-lat: Check all calls to clock_nanosleep() and clock_gettime()\n selftests: timers: nsleep-lat: Reuse kselftest error numbers\n selftests: timers: nsleep-lat: Explicitly list the tested clocks\n selftests: timers: nsleep-lat: Use NSEC_PER_MSEC define for unreasonable latency\n selftests: timers: nanosleep: Report each test separately\n selftests: timers: nanosleep: Explicitly handle timer_delete() failure\n selftests: timers: nanosleep: Move all single clock tests out of the loop in main()\n selftests: timers: nanosleep: Reuse kselftest error numbers\n selftests: timers: nanosleep: Explicitly list the tested clocks\n selftests: timers: nanosleep: Drop output alignment\n selftests: timers: Use clock_name() and constants from clock-helpers.h\n selftests: Add clock-helpers.h\n timer_list: Use ktime_t over nanoseconds\n timer_list: Use standard \u0027long long\u0027 format placeholders\n hrtimer: Add a lockdep assertion to hrtimer_update_base()\n timekeeping: Use u32 for clock_was_set_seq\n timekeeping: Rename clockid_aux_valid() to clockid_is_aux_clock()\n hrtimer: Account nr_retries on recovered interrupt retries\n timers/itimer: Zero-init old itimerval before copy to userspace\n nohz: Replace dead select with choice default\n ...\n" }, { "commit": "b0239dd672306ad242545f793132938847f17e53", "tree": "62c21cd60ad99120fa21855831ef7a1f466cd33b", "parents": [ "030c9f813b8e48d2b066983c94cf6294968f9496", "0eaed89c18aeedf0898baf2dbf5ff027c6795152" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 16:20:25 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 16:20:25 2026 -0700" }, "message": "Merge tag \u0027timers-clocksource-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull clocksource/event driver updates from Thomas Gleixner:\n\n - Remove redundant dev_err() and dev_err_probe() invocations in various\n drivers\n\n - Handle SWAP_IO_SPACE correctly in the rtl-otto driver\n\n - Make the Samsumg PWM timer driver PREEMPT_RT compatible\n\n - Ensure that the SUN4I timer is programmed with a delta larger than\n zero as a zero delta causes the the timer to be disabled\n\n - The usual fixes and improvements all over the place\n\n* tag \u0027timers-clocksource-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n clocksource/drivers/armada: Unwind timer clock on init failure\n clocksource/drivers/rtl-otto: Change driver to use __raw reads and writes\n clocksource/drivers/samsung_pwm: Switch to raw_spinlock_t type\n clocksource/drivers/clps711x: Do not unmap clocksource MMIO\n clocksource/drivers/nxp-pit: Fix IRQ leak on cpuhp_setup_state error path\n clocksource/drivers/timer-sun4i: Advertise a real minimum delta\n clocksource: Remove redundant dev_err()/dev_err_probe()\n clocksource/drivers/sh_cmt: Use named initializers for platform_device_id arrays\n clocksource/drivers/sh_mtu2: Drop unused assignment of platform_device_id\n" }, { "commit": "030c9f813b8e48d2b066983c94cf6294968f9496", "tree": "9a9fc6f2fe2e9cc3b0558c5a92ebeb6e6b8d361a", "parents": [ "0dd1a54f44348d9cf6bae57a2b5cb0b53826a2c7", "dfc256dac54c8b692110bf905c64cb130e15963d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:58:29 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:58:29 2026 -0700" }, "message": "Merge tag \u0027timers-cleanups-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull treewide timer related cleanups from Thomas Gleixner:\n\n - Remove the leftover CLOCK_TICK_RATE which has been scheduled for\n removal more than a decade ago along with some now empty asm/timex.h\n files.\n\n - Consolidate delay timer calibration\n\n The construct of having a define in a header requires that\n architectures provided asm/timex.h for no reason. Also the function\n name for reading the delay timer is confusing at best.\n\n Use a config switch to enable that functionality and rename the\n function to delay_read_timer() to make the purpose clear.\n\n This removes some more now empty asm/timex.h files as well.\n\n* tag \u0027timers-cleanups-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n calibrate: Rework delay timer calibration\n treewide: Remove CLOCK_TICK_RATE\n x86: Use PIT_TICK_RATE instead of CLOCK_TICK_RATE\n" }, { "commit": "0dd1a54f44348d9cf6bae57a2b5cb0b53826a2c7", "tree": "cc7ca2ad4a92fb593ffdf72fe6ebc0dff75155c0", "parents": [ "b844715e8aca2929c0a97a32b3a5650496b5872f", "99b49e02f9488335156c896c24aab0785623eb67" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:29:53 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:29:53 2026 -0700" }, "message": "Merge tag \u0027smp-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull SMP core updates from Thomas Gleixner:\n\n - Reduce the preemption disabled sections in smp_call_function*().\n\n The various smp call functions keep preemption disabled accross the\n full operation which includes the wait for completion. Especially the\n latter can take some time when one of the target CPUs is not\n immediately responding to the IPI, which can result in large latency\n spikes.\n\n To improve this provide a per task CPU mask to track the CPUs to wait\n for. That makes the information required for the wait task local and\n therefore allows to reenable preemption before the wait.\n\n While this comes with moderate extra memory cost this reduces SMP\n function call induced latency measured in a fleet for high priority\n tasks from ~17ms to ~1.5ms (~90%).\n\n - Reduce the overhead of the CSD debug code by replacing the heavy\n memory barriers with smp_store_release()/acquire()\n\n - Remove obsolute unused hotplug states\n\n* tag \u0027smp-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n scftorture: Remove preempt_disable() in scftorture_invoke_one()\n smp: Remove preempt_disable() from on_each_cpu_cond_mask()\n smp: Remove preempt_disable() from smp_call_function()\n smp: Enable preemption early in smp_call_function_many_cond()\n smp: Alloc percpu csd data in smpcfd_prepare_cpu() only once\n smp: Use task-local IPI cpumask in smp_call_function_many_cond()\n smp: Refactor remote CPU selection in smp_call_function_any()\n smp: Enable preemption early in smp_call_function_single()\n smp: Disable preemption explicitly in __csd_lock_wait()\n cpu/hotplug: Remove CPUHP_AP_ARM_CORESIGHT_CTI_STARTING\n smp: Use release stores for csd_lock_record() state\n" }, { "commit": "b844715e8aca2929c0a97a32b3a5650496b5872f", "tree": "99ef07873c05289aa5e296b060139dd29eb01058", "parents": [ "8b5d31123f41957fe265deeb03ee87fe62f155a8", "d1c656df36d00aad9c810e563f918dbbaf41a7b9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:26:35 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:26:35 2026 -0700" }, "message": "Merge tag \u0027locking-futex-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull futex updates from Thomas Gleixner:\n\n - Improvements to various futex self tests:\n\n - Conversion to the selftest harness\n\n - Provide and use thread creation and synchronization helpers to\n reduce the dependency on delays, which tend to fail on loaded test\n systems\n\n - New tests for validating owner exit scenarios for robust and PI\n futexes\n\n - Runtime detect supported features and skip the tests if the kernel\n has no support\n\n - A few minor fixes\n\n* tag \u0027locking-futex-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n selftests/futex: Give circular-list nodes static storage\n selftests/futex: Use thread synchronization helpers instead of usleep()\n selftests/futex: Provide thread creation and synchronization helpers\n selftests/futex: Dynamically skip unsupported tests\n selftests/futex: Add FUTEX_LOCK_PI owner-exiting coverage\n selftests/futex: Migrate robust_list to harness\n selftests/futex: Migrate futex_priv_hash to harness\n selftests/futex: Migrate futex_numa_mpol to harness\n selftests/futex: Migrate futex_requeue_pi_signal_restart to harness\n selftests/futex: Migrate futex_requeue_pi_mismatched_ops to harness\n selftests/futex: Migrate futex_requeue_pi to harness\n selftests/futex: Migrate futex_requeue to harness\n selftests/futex: Migrate futex_wait_uninitialized_heap to harness\n selftests/futex: Migrate futex_wait_private_mapped_file to harness\n selftests/futex: Migrate futex_wait to harness\n selftests/futex: Correct validation logic in waitv\n selftests/futex: Migrate functional tests to harness\n selftests/futex: Remove static keyword from \u0027head\u0027\n futex: Remove unnecessary NULL check before kvfree()\n selftests/rseq: Replace glibc-specific __GNUC_PREREQ with portable check\n" }, { "commit": "8b5d31123f41957fe265deeb03ee87fe62f155a8", "tree": "55c7871554aee125124bd9beaface3f327706775", "parents": [ "762fffa407d8d24288513538cf5d17c2c5425258", "7f328162a98ee4e10aed42b51a7f8125c6e73384" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:24:18 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:24:18 2026 -0700" }, "message": "Merge tag \u0027irq-drivers-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull interrupt chip driver updates from Thomas Gleixner:\n\n - Update the realtek-rtl driver to support multiple interrupt domains\n and remove hardcoded assumptions about the underlying demultiplex\n interrupt\n\n - Remove the unused IMGPDC driver, which got orphaned when the metag\n architecture was removed long ago.\n\n - Update the Qualcom PDC interrupt chip driver to support differnent\n versions of the IP block, to properly differentiable between direct\n and GPIO based interrupts and to utilize pass through mode to the\n underlying GIC interrupt.\n\n - Remove redundant dev_err() and dev_err_probe() invocations in various\n drivers as that error logging has been moved to devm_request_*_irq().\n\n - Support dynamic MSI-X allocations in the GIC-V3-ITS MSI parent domain\n driver. That\u0027s important for VFIO pass through devices to avoid\n reinitialization of the MSI-X space when extending the vector range.\n Such a reinitialization can result is lost interrupts and stale\n devices.\n\n - Fix a bunch of resource leaks in various driver error paths\n\n - The usual small improvements all over the place\n\n* tag \u0027irq-drivers-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (21 commits)\n irqchip/gic-v3-its: Fix grammar and replace a bit number with its symbol\n irqchip/gic-v3-its: Prevent leak in its_vpe_irq_domain_alloc()\n irqchip/renesas-irqc: Fix generic interrupt chip leak on remove\n irqchip/gic-v3-its: Enable dynamic MSI-X allocation\n irqchip: Remove redundant dev_err()/dev_err_probe()\n irqchip/gic-v5/iwb: Fix stray verb in comment\n irqchip/qcom-pdc: Fix kernel doc for qcom_pdc_gic_secondary_set_type()\n irqchip/gic-v3-its: Fix its node leak in gic_acpi_parse_madt_its()\n irqchip/gic-v3-its: Fix memleak in its_probe_one()\n irqchip/qcom-pdc: Configure PDC to pass through mode\n irqchip/qcom-pdc: Differentiate between direct SPI and GPIO as SPI\n irqchip/qcom-pdc: Move all static variables to struct pdc_desc\n irqchip/qcom-pdc: Restructure version support\n irqchip/irq-imgpdc: Remove unused driver\n irqchip/irq-realtek-rtl: Activate multiple parents\n irqchip/irq-realtek-rtl: Allow shuffled interrupt order\n irqchip/irq-realtek-rtl: Add a select function\n irqchip/irq-realtek-rtl: Add mask for interrupt handling\n irqchip/irq-realtek-rtl: Add interrupt data structure\n irqchip/irq-realtek-rtl: Split out parent setup code\n ...\n" }, { "commit": "762fffa407d8d24288513538cf5d17c2c5425258", "tree": "563dba12f6fe1caeea545dfa23c847c7d25723e1", "parents": [ "6138016752322261e26ac88eaeb64c0c890f98a2", "1ad6d4a722f509128038f9c85c6a122a382179fe" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:19:50 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:19:50 2026 -0700" }, "message": "Merge tag \u0027irq-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull generic interrupt subsystem updates from Thomas Gleixner:\n\n - Remove pointless NULL checks of the kstats_irqs field. That\u0027s a\n historical left over and not longer required.\n\n - Add Radu Rendec as reviewer. Radu thankfully stepped up to help\n reviewing the interrupt core and the related drivers code.\n\n - The usual small improvements and fixes\n\n* tag \u0027irq-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n MAINTAINERS: Add Radu Rendec as reviewer for the interrupt subsystem\n genirq/msi: Move misplaced EXPORT_SYMBOL_GPL for msi_domain_free_irqs_all()\n parisc: Remove unnecessary NULL check of the kstat_irqs field\n genirq: Remove unnecessary NULL check of the kstat_irqs field\n irqdomain: Remove unnedded NULL check in __irq_domain_[de]activate_irq()\n genirq/manage: Use irqd_get_parent_data() helper in __irq_get_irqchip_state()\n irqdomain: Plug leak in irq_domain_alloc_irqs_locked() error path\n" }, { "commit": "6138016752322261e26ac88eaeb64c0c890f98a2", "tree": "ca5c9f66f9fdcbc9c79325f767fcd731fac810e0", "parents": [ "3424d8c18a7da1010d03391a22e728b857d0d5c5", "2a98ffac7733e77a178af7240f9fa9392ca45aad" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:16:10 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:16:10 2026 -0700" }, "message": "Merge tag \u0027core-rseq-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull rseq updates from Thomas Gleixner:\n \"Two trivial updates for RSEQ:\n\n - Add the recently added new test binaries to .gitignore\n\n - Fix a trivial typo in a comment\"\n\n* tag \u0027core-rseq-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n selftests/rseq: Add missing test binaries to .gitignore\n selftests/rseq: Fix spelling of accommodate\n" }, { "commit": "3424d8c18a7da1010d03391a22e728b857d0d5c5", "tree": "3d27ce37b7517ee422064b462e93e1a99302986c", "parents": [ "3a0dd7ba4f44cdc116d83712f61e7c1a95be3588", "05c033db7e9ad3c34f6968ec568cb6ee01051c57" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:00:56 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 15:00:56 2026 -0700" }, "message": "Merge tag \u0027core-entry-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull generic entry code updates from Thomas Gleixner:\n\n - Make syscall user dispatching configurable\n\n Not all architectures can makes use of syscall user dispatching.\n Allow them to disable the feature completely.\n\n - Consolidate stack randomization for the generic entry code and the\n architectures using it.\n\n Stack randomization on syscall entry was sprinkled throughout the\n architecture specific low level entry code and in some cases at the\n wrong points, e.g. before establishing state, which violates the\n non-instrumentable constraints of that code.\n\n Clean this up by integrating stack randomization into the generic\n entry code helpers so that it is invoked at the earliest possible\n point right after establishing state and converting all generic entry\n code using architecture over.\n\n - Clean up the syscall number handling in the generic entry code. It\n works correctly for architectures which have a separate return value\n storage in pt_regs, but fails to distinguish the case where user\n space handed in -1 as syscall number from the case where the entry\n code rejects it by returning -1 to the callers. Aside of that the\n return value functionality of those interfaces is not really\n intuitive.\n\n Fix this by separating the decision to reject a syscall (user\n dispatch, ptrace, seccomp ...) from the potential modification of the\n syscall number through these mechanisms.\n\n This solves most of the problems for architectures which do not have\n a separate return value storage in pt_regs except for the case where\n a tracepoint has a BPF script or a probe attached which overwrite\n both the syscall number and the return value. But that\u0027s a problem\n which cannot be solved in the generic code, that only can be\n addressed by separating the storage model in the affected\n architectures.\n\n* tag \u0027core-entry-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (23 commits)\n entry, treewide: Make syscall_enter_from_user_mode[_work]() indicate syscall execution\n entry: Make return type of syscall_trace_enter() bool\n entry: Rework trace_syscall_enter()\n entry: Rework syscall_audit_enter()\n syscall_user_dispatch: Introduce ARCH_SUPPORTS_SYSCALL_USER_DISPATCH\n entry: Fix seccomp bypass after ptrace with TSYNC\n x86/entry: Simplify the syscall number logic\n x86/entry: Get rid of the sys_ni_syscall() indirection\n x86/entry: Make syscall functions static\n ptrace, treewide: Rename ptrace_report_syscall_entry() to ptrace_report_syscall_permit_entry()\n seccomp, treewide: Rename and convert __secure_computing() to return boolean\n entry: Use syscall number instead of rereading it\n entry: Remove syscall_enter_from_user_mode()\n x86/syscall: Use [syscall_]enter_from_user_mode_randomize_stack()\n s390/syscall: Use enter_from_user_mode_randomize_stack()\n riscv/syscall: Use syscall_enter_from_user_mode_randomize_stack()\n powerpc/syscall: Use syscall_enter_from_user_mode_randomize_stack()\n loongarch/syscall: Use syscall_enter_from_user_mode_randomize_stack()\n entry: Provide [syscall_]enter_from_user_mode_randomize_stack()\n randomize_kstack: Provide add_random_kstack_offset_irqsoff()\n ...\n" }, { "commit": "3a0dd7ba4f44cdc116d83712f61e7c1a95be3588", "tree": "fed16cf3c32b6121fcd9314eb091d5c6d3ebe397", "parents": [ "d5b550edadab6810653f287715a0d696306b6ae0", "33a581d800670e5d91e99203536d15ef7451ad69" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 14:13:43 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 14:13:43 2026 -0700" }, "message": "Merge tag \u0027x86-documentation-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 documentation updates from Ingo Molnar:\n\n - Misc x86 documentation updates (Randy Dunlap, Daniel Gibson)\n\n* tag \u0027x86-documentation-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n Documentation/arch/x86/amd-debugging: Add example for reset register\n x86/intel/quark: Clean up function kernel-doc\n x86/speculation: Drop Excess function parameter descriptions\n x86/lib: checksum_64.h: Eliminate all kernel-doc warnings\n x86/barrier: Use correct parameter names in kernel-doc\n" }, { "commit": "d5b550edadab6810653f287715a0d696306b6ae0", "tree": "be38aaf138393478dd90567260351605c35100f0", "parents": [ "3dd1f7447f4f989b3a348cdc347b15397d03bebc", "d824ed1307680dd482f607b0e707c575f70668c4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 14:08:57 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 14:08:57 2026 -0700" }, "message": "Merge tag \u0027x86-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 core updates from Ingo Molnar:\n\n - Optimize the kcfi call sequence (Peter Zijlstra)\n\n - Use sfence for wmb() if SSE is available (Yao Zi)\n\n* tag \u0027x86-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/locking: Use sfence for wmb() if SSE is available\n x86/kcfi: Optimize call sequence\n" }, { "commit": "3dd1f7447f4f989b3a348cdc347b15397d03bebc", "tree": "d13c71ec4a9234af54cfc321a6a94ed21c24fd06", "parents": [ "09d639f1f1f6347eb28802799a2187ae540b00ba", "f6cff820877ff34839920ff9de9b912359cfb607" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 14:06:52 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 14:06:52 2026 -0700" }, "message": "Merge tag \u0027x86-build-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 build update from Ingo Molnar:\n\n - Only align ENTRY_TEXT to PMD_SIZE if necessary (Hamza Mahfooz)\n\n* tag \u0027x86-build-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/build: Only align ENTRY_TEXT to PMD_SIZE if necessary\n" }, { "commit": "09d639f1f1f6347eb28802799a2187ae540b00ba", "tree": "6c3f547e9442ba39626117917878b804490d11b1", "parents": [ "8dcef8882aad8f1b8668d1c39968cde99312aa3c", "2082fec6d4c65b4031dc82daa10cc9a5d35654c9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 14:04:48 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 14:04:48 2026 -0700" }, "message": "Merge tag \u0027x86-boot-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 boot updates from Ingo Molnar:\n\n - Misc simplifications and cleanups (Thorsten Blum)\n\n* tag \u0027x86-boot-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/boot: Use bool and IS_ENABLED() to simplify query_edd()\n x86/boot: Use IS_ENABLED() to simplify built-in-or-module checks\n x86/purgatory: Return bool from verify_sha256_digest()\n" }, { "commit": "8dcef8882aad8f1b8668d1c39968cde99312aa3c", "tree": "6c47988a06817c054b53abf061ff1de2883b129d", "parents": [ "e2457a664ea02c414c6b9828bff3a0df4c300f63", "a7187ca1e72505fd6785b0ce537ae066773a5532" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 14:02:15 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 14:02:15 2026 -0700" }, "message": "Merge tag \u0027x86-msr-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 MSR updates from Ingo Molnar:\n\n - Streamline the x86 MSR handling APIs along the 64-bit variants,\n simplifying the interfaces.\n\n Removal of the old APIs is planned for the next cycle, to reduce\n churn \u0026 integration pain (Juergen Gross)\n\n* tag \u0027x86-msr-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (21 commits)\n x86/mce: Work around build warning after MSR-interface switch\n cpufreq: Stop using 32-bit MSR interfaces\n x86/featctl: Stop using 32-bit MSR interfaces\n KVM/x86: Stop using 32-bit MSR interfaces\n x86/mtrr: Stop using 32-bit MSR interfaces\n acpi: Stop using 32-bit MSR interfaces\n powercap: Stop using 32-bit MSR interfaces\n thermal/intel: Stop using 32-bit MSR interfaces\n x86/olpc: Stop using 32-bit MSR interfaces\n x86/hyperv: Stop using 32-bit MSR interfaces\n hwmon: Stop using 32-bit MSR interfaces\n EDAC: Stop using 32-bit MSR interfaces\n x86/cpu: Stop using 32-bit MSR interfaces\n x86/apic: Stop using 32-bit MSR interfaces\n x86/resctrl: Stop using 32-bit MSR interfaces\n x86/tsc: Stop using 32-bit MSR interfaces\n x86/amd: Stop using 32-bit MSR interfaces\n x86/pci: Stop using 32-bit MSR interfaces\n x86/hygon: Stop using 32-bit MSR interfaces\n x86/mce: Stop using 32-bit MSR interfaces\n ...\n" }, { "commit": "e2457a664ea02c414c6b9828bff3a0df4c300f63", "tree": "36ce10cf26cb39782a0be00b4e1c7f7ec683c0e6", "parents": [ "dfa35434d7f20142fedd7120277b1044a0a2bb64", "68e37487810a3da43c48340fab7a55b3b6efdae3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 13:39:48 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 13:39:48 2026 -0700" }, "message": "Merge tag \u0027sched-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull scheduler updates from Ingo Molnar:\n \"Load-balancing updates:\n\n - \u0027flatten the pick\u0027: improve cgroup scheduling, which has always\n been problematic and painful, which has caused various scheduling\n misbehavior such as the mishandling of reniced tasks et al.\n\n Add various cgroup weight distribution methods via cgroup_mode:\n \u0027up\u0027, \u0027max\u0027, \u0027concur\u0027 and \u0027tasks\u0027 - with the default being \u0027concur\u0027\n which is the most precise yet also most expensive version.\n\n Finally, change cgroup scheduling to a single runqueue (Peter\n Zijlstra)\n\n - Series to improve the scheduling latency of short slice tasks\n (Vincent Guittot)\n\n - Series to fix cluster scheduling in the presence of asymmetric\n capacity (Ricardo Neri)\n\n - Prefer fully idle cores for NOHZ balancing (Andrea Righi)\n\n - Don\u0027t trigger active load-balancing if src_rq-\u003ecurr is not on_rq\n (Xin Zhao)\n\n PSI updates:\n\n - Skip irqtime accounting when no new irq time has elapsed (Usama\n Arif)\n\n Scheduler debugging updates:\n\n - Remove unused schedstats (Shrikanth Hegde)\n\n - Defer freeing of cpumask memblock memory to initcall (Waiman Long)\n\n Misc fixes and updates by Yu C Chen, K Prateek Nayak, Peter Zijlstra,\n Vincent Guittot, Xin Zhao, Yury Norov, Zhan Xusheng\"\n\n* tag \u0027sched-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (33 commits)\n sched/fair: Fix flat hierarchy\n sched/isolation: Defer freeing of cpumask memblock memory to initcall\n sched/topology: Restore SD_PREFER_SIBLING in domains with asymmetric capacity\n sched/fair: Allow load balancing between CPUs of identical capacity\n sched/fair: Skip misfit load accounting when the destination CPU cannot help\n sched/fair: Check CPU capacity before comparing group types during load balance\n sched/fair: Also gate overloaded status update for SD_ASYM_CPUCAPACITY\n sched/fair: Do not skip CPUs of similar capacity with busy SMT siblings\n sched/fair: Prefer fully idle cores for NOHZ balancing\n stop_machine: Make stop_one_cpu_nowait() return void\n sched/eevdf: Delayed dequeue task can\u0027t preempt\n sched/fair: Fix stale comments referring to removed CFS concepts\n sched/debug: Remove unused schedstats\n sched/psi: skip irqtime accounting when no new irq time has elapsed\n sched/fair: Reflow sched_balance_rq()\n sched/fair: Simplify balance_interval reset logic in sched_balance_rq()\n sched/fair: Don\u0027t trigger active lb if src_rq-\u003ecurr is not on_rq\n sched/eevdf: Speedup short slice task scheduling\n sched/eevdf: Always update slice protection\n sched/eevdf: Cancel slice protection if short slice task is eligible\n ...\n" }, { "commit": "dfa35434d7f20142fedd7120277b1044a0a2bb64", "tree": "a9507732ba5b93ff16acf8eebb6fcfbc86f50f2b", "parents": [ "8915457146a11d20a6c0786396376afda65eec40", "ec4fad7c2bdc80c62fa73f799b77ad299f73dcc3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 13:07:17 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 13:07:17 2026 -0700" }, "message": "Merge tag \u0027locking-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull locking updates from Ingo Molnar:\n \"Futexes:\n\n - Use runtime constants for futex_hash computation (K Prateek Nayak,\n Peter Zijlstra)\n\n - Optimise the size check get_futex_key() (Sebastian Andrzej Siewior)\n\n - Avoid private hash use-after-free on final put (Felix Hoffmann)\n\n - Tell kmemleak we\u0027re not leaking __futex_queues (Peter Zijlstra)\n\n Rust integration updates:\n\n - Implement refcounted interrupt disable and SpinLockIrq for Rust\n (Boqun Feng, Heiko Carstens, Joel Fernandes, Lyude Paul)\n\n - Rust sync: add helpers for mb, dma_mb and friends; add generic\n memory barriers and use LKMM atomics instead of Rust atomics in the\n revocable code (Gary Guo)\n\n - Add abstraction and integrate synchronize_rcu() (Philipp Stanner)\n\n Lock debugging:\n\n - Add qspinlock contended_release tracepoint (Dmitry Ilvokhin, Peter\n Zijlstra)\n\n - Enable the printing of held locks of remote running tasks and print\n task CPU (Ingo Molnar)\n\n - percpu-rwsem: Annotate intentional data race in readers_active_check()\n (Sun Shaojie)\n\n Misc fixes and updates by Boqun Feng, Peter Zijlstra, Fangrui Song,\n Naveen Kumar Chaudhary and Thomas Huth\"\n\n* tag \u0027locking-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (44 commits)\n rust: sync: Introduce SpinLockIrq::lock_with() and friends\n rust: sync: Add SpinLockIrq\n rust: sync: Use super::* in spinlock.rs\n rust: helper: Add spin_{un,}lock_irq_{enable,disable}() helpers\n rust: Introduce interrupt module\n s390/preempt: Enable HAS_SEPARATE_PREEMPT_RESCHED_BITS\n arm64: sched/preempt: Enable HAS_SEPARATE_PREEMPT_RESCHED_BITS\n preempt: Introduce HAS_SEPARATE_PREEMPT_RESCHED_BITS\n sched: Avoid signed comparison of preempt_count() in __cant_migrate()\n sched: Remove the unused preempt_offset parameter of __cant_sleep()\n locking: Switch to _irq_{disable,enable}() variants in cleanup guards\n irq: Add KUnit test for refcounted interrupt enable/disable\n irq,spin_lock: Add counted interrupt disabling/enabling\n openrisc: Include \u003clinux/cpumask.h\u003e in smp.h\n preempt: Introduce __preempt_count_{sub,add}_return()\n preempt: Introduce HARDIRQ_DISABLE_BITS\n preempt: Track NMI nesting to separate per-CPU counter\n futex: Tell kmemleak we\u0027re not leaking __futex_queues\n x86/paravirt: Trace contended_release on unlock\n tracing/lock: Use TRACE_EVENT_FN() for contended_release\n ...\n" }, { "commit": "8915457146a11d20a6c0786396376afda65eec40", "tree": "75a8c5625e800896bb4d977dede53590e60cce95", "parents": [ "a0acd94e3819fcd8346ae3c16df987e0fabb3128", "917d558b151cad5b05991e5eaee22efab33525ca" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 12:31:07 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 12:31:07 2026 -0700" }, "message": "Merge tag \u0027perf-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull performance events updates from Ingo Molnar:\n \"uprobes updates:\n\n - Fix a category of bugs with optimized uprobes that can clobber the\n redzone area with call instruction storing return address on stack\n where user code may keep temporary data without adjusting RSP.\n\n Fix this by moving the optimized uprobes on top of 10-bytes NOP\n instruction, so we can squeeze another instruction to escape the\n redzone area before doing the call (Jiri Olsa, Andrii Nakryiko)\n\n - Switch uretprobes_srcu to SRCU-fast-updown, to improve performance\n (Puranjay Mohan)\n\n Intel CPU PMU driver updates:\n\n - Optimize ACR handling in match_prev_assignment() (Dapeng Mi)\n\n - Fix various PMU driver bugs and data leaks (Dapeng Mi)\n\n - Fix Intel PT stop/start with no update (Adrian Hunter)\n\n Intel uncore PMU driver updates:\n\n - Fix various uncore PMU setup robustness bugs (Zide Chen)\n\n AMD uncore PMU driver updates:\n\n - Add group validation (Sandipan Das)\n\n .. and misc fixes and updates by Dapeng Mi, Randy Dunlap and Zide Chen\"\n\n* tag \u0027perf-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (42 commits)\n perf/x86: Optimize ACR handling in match_prev_assignment()\n perf/x86/intel: Fix intel_cap handling on hybrid PMUs\n perf/x86: Remove stale fixed counter helper and fix hybrid PMU access\n perf/x86/intel: Unwind cpuc state if PEBS buffer setup fails\n perf/x86: Guard intel_pmu_cpu_dead() against invalid hybrid PMU casts\n perf/x86: Free hybrid state on PMU init failure\n perf/x86: Unregister PMI handler on PMU init failure\n perf/x86/intel/pt: Fix stop/start with no update\n perf/x86/intel/pt: Use bitwise access for PERF_HES_STOPPED\n perf/x86/intel/pt: Factor out pt_config_enable()\n uprobes: Switch uretprobes_srcu to SRCU-fast-updown\n srcu: Add lock guard for srcu_fast_updown flavor\n perf/x86/intel/pt: Drop kernel-doc for deleted struct members\n perf/x86/amd/uncore: Add group validation\n selftests/bpf: Add tests for forked/cloned optimized uprobes\n selftests/bpf: Add tests for uprobe nop10 red zone clobbering\n selftests/bpf: Add reattach tests for uprobe syscall\n selftests/bpf: Change uprobe/usdt trigger bench code to use nop10\n selftests/bpf: Change uprobe syscall tests to use nop10\n selftests/bpf: Emit nop,nop10 instructions combo for x86_64 arch\n ...\n" }, { "commit": "a0acd94e3819fcd8346ae3c16df987e0fabb3128", "tree": "42b1fd69eac025cc5883778d63cbe055393b182c", "parents": [ "1200d84f4c0a929a0780180d25063d93773be79c", "d8a2860b4a366bfa8acb3d64da2c546ea26d2091" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 12:18:33 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 12:18:33 2026 -0700" }, "message": "Merge tag \u0027objtool-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull objtool updates from Ingo Molnar:\n\n - Fix various klp-build bugs reported by Joe Lawrence (Josh Poimboeuf,\n Joe Lawrence)\n\n - Misc fixes and cleanups (Puranjay Mohan, Thomas Huth and Ingo Molnar)\n\n* tag \u0027objtool-core-2026-08-17\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n objtool/klp: Fix vmlinux klp relocations for EXPORT_SYMBOL_FOR_MODULES()\n objtool/klp: Fix .kcfi_traps special section extraction\n objtool/klp: Fix vmlinux .klp.symid link error for .exitcall.exit symbols\n objtool/klp: Fix line numbers in Module.symvers parse errors\n objtool/klp: Fix relocations for EXPORT_SYMBOL_FOR_MODULES() symbols\n objtool/klp: Allow new references to module exports\n objtool/klp: Don\u0027t match local symbols against exports\n objtool/klp: Fix cross-module klp relocation section naming\n objtool/klp: Explicitly disallow patching or referencing init code/data\n objtool/klp: Ignore replacement offset of empty x86 alternatives\n objtool/klp: Fix size of empty special section entries\n objtool/klp: Fix vmlinux .klp.symid link error for .no_trim_symbol symbols\n objtool/headers: Sync tools/include/linux/objtool_types.h with include/linux/objtool_types.h\n objtool: Replace __ASSEMBLY__ with __ASSEMBLER__ in header files\n objtool/klp: Fix symbol resolution for duplicate data symbols\n objtool/klp: Add .klp.symid for sympos disambiguation\n objtool/klp: Skip hidden directories when finding objects\n objtool/klp: Fix false module dependencies caused by dead relocs\n objtool/klp: Normalize Module.symvers paths to module names\n objtool/klp: Fix module name normalization for paths with dots\n" }, { "commit": "1200d84f4c0a929a0780180d25063d93773be79c", "tree": "27d498f910603e4f15d587aeb9384969dd9c1774", "parents": [ "85cdaca6970028bf6f544c355c90035586836ddf", "1304643a1c20badbb91b86a5084dd76cb7620c05" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 12:08:13 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 12:08:13 2026 -0700" }, "message": "Merge tag \u0027powerpc-7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux\n\nPull powerpc updates from Madhavan Srinivasan:\n\n - Enable Rust for ppc64le\n\n - ppc4xx gpio driver updates\n\n - Add power12 base enablement support\n\n - Validate arch_compat against host compatibility mode\n\n - Simplify bootx_scan_dt_build_struct() in powermac platform\n\n - Implement get_direction() in cpm2\n\n - Use cpu_relax() in ps3_create_spu()\n\n - Add NULL guard for cause_ipi in smp_muxed_ipi_message_pass\n\n - Fixes to handle pseries watchdogs in kdump path\n\n - Fix missing r2 clobber in PCREL inline assembly\n\n - Set GPIO chip parent on ppc44x\n\n - KVM: Introduce KVM_CAP_PPC_COMPAT_CAPS and wire up ioctl\n\n - KVM: Use generic xfer to guest work function\n\n - Enable to run posix cpu timers in task context\n\n - Misc fixes and cleanups\n\nThanks to Aditya Gupta, Alice Ryhl, Amit Machhiwal, Andrew Morton,\nAnushree Mathur, Athira Rajeev, Bartosz Golaszewski, Cédric Le Goater,\nChristian König, Christophe Leroy (CS GROUP), Gary Guo, Gaurav Batra,\nGautam Menghani, Gou Hao, Hari Bathini, Harsh Prateek Bora, jiazhenyuan,\nJinjie Ruan, Link Mauve, Linus Walleij, Mahesh Kumar G Mahesh\nSalgaonkar, Michael Walle, Michal Suchánek, Mukesh Kumar Chaurasiya\n(IBM), Nicholas Piggin, Nikhil Kumar Singh, Praveen K Pandey, Ritesh\nHarjani (IBM), Rosen Penev, Saket Kumar Bhaskar, Shrikanth Hegde,\nSourabh Jain, Thorsten Blum, Vaibhav Jain, Venkat Rao Bagalkote, Vishal\nChourasia, Wentao Guan, and Yanfei Xu.\n\n* tag \u0027powerpc-7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (62 commits)\n powerpc/pseries/iommu: switch to Default DMA window during kdump\n powerpc/configs: enable CONFIG_RAS to fix EDAC support\n KVM: PPC: Document KVM_PPC_GET_COMPAT_CAPS ioctl\n KVM: PPC: Book3S HV: Add support for compat CPU capabilities for KVM on PowerNV\n KVM: PPC: Book3S HV: Implement compat CPU capability retrieval for KVM on PowerVM\n KVM: PPC: Introduce KVM_CAP_PPC_COMPAT_CAPS and wire up ioctl\n gpio: ppc44x: use dev_name() for chip label\n gpio: ppc44x: fix undefined behavior in GPIO_MASK2 macro\n gpio: ppc44x: drop PPC-specific IO helpers\n gpio: ppc44x: Convert GPIO to generic MMIO\n gpio: ppc44x: Use platform resource helper for GPIO MMIO\n gpio: ppc44x: Use module platform driver helper for GPIO\n gpio: ppc44x: update all 4xx to 44x\n gpio: move ppc4xx gpio driver from arch/powerpc to drivers/gpio\n KVM: PPC: Use min() in kvm_vm_ioctl_check_extension()\n KVM: PPC: booke: Use min() in watchdog_next_timeout()\n powerpc/perf: Add power12 Base Performance Monitoring support\n powerpc: Add Power12 architected mode\n powerpc: Add Power12 raw mode\n powerpc/pseries: Limit PVR list to 16 entries for CAS negotiation\n ...\n" }, { "commit": "85cdaca6970028bf6f544c355c90035586836ddf", "tree": "77f5b379507d6f0e410daa5862065d5a6409a9ea", "parents": [ "d24f5cdbeff8b6a063fca92d0a1f94122a799b59", "26b77009ee74f6122a067a4f9926ff6e09fffb94" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 11:57:17 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 11:57:17 2026 -0700" }, "message": "Merge tag \u0027arm64-upstream\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux\n\nPull arm64 updates from Will Deacon:\n \"There\u0027s a reasonable amount of stuff here, including a bunch of\n updates to the perf PMU drivers and some MPAM updates to expose the\n memory bandwidth counters via resctrl.\n\n On the architecture side, some highlights include support for BBML3\n and steps towards support for an architectural NMI solution, all\n wrapped up in a web of fixes for latent issues identified by Sashiko.\n\n ACPI:\n\n - Combine reads of AMU counters into a single FFH feedback counter op\n\n Confidential computing:\n\n - Fix smp_processor_id() in preemptible context when retrieving an\n attestation token inside a realm\n\n - Convert pKVM over to a \"CC platform\"\n\n - Clean-up our SWIOTLB configuration in preparation for reworking the\n handling of encrypted/decryped DMA buffers in the dma-mapping tree\n\n CPU errata handling:\n\n - Work around broken device memory ordering on NVIDIA Olympus cores\n\n - Fix broken \u0027nospectre_bhb\u0027 command-line option\n\n - Select the idle loop backend instruction on the command-line\n\n CPU features:\n\n - Replace our BBML2-noabort feature with the new architectural BBML3\n feature\n\n - Disable in-kernel BTI for recent versions of Clang due to issues\n with livepatch that are still being investigated\n\n - Clean-up documentation describing which ID register fields are\n exposed to userspace\n\n Interrupts:\n\n - Preliminary work towards supporting FEAT_NMI, which cleans up our\n IRQ entry code and fixes some latent issues with pseudo-NMI\n\n - Support for an SDEI backend to trigger an NMI backtrace\n\n Memory management:\n\n - Treat all devices as coherent when CLIDR_EL1.LoC \u003d\u003d 0\n\n - Fix no-map handling of sub-page-sized regions\n\n - Second attempt at unmapping the linear aliases of the kernel data\n and bss sections\n\n - Fix EFI runtime calls when software-PAN is enabled\n\n Miscellaneous:\n\n - Add Mark Rutland as a reviewer!\n\n - Tidy-up our futex cmpxchg logic when using the new LSUI\n instructions\n\n - Drop the requirement on DYNAMIC_FTRACE_WITH_CALL_OPS when\n selecting HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS\n\n - Fix a false-positive KCSCAN splat in the delay loop\n\n - Use a portable typedef for 128-bit scalar types in our UAPI headers\n\n - Non-critical fixes for Sashiko reports all over\n\n MPAM:\n\n - Hook MPAM memory bandwidth counters into resctrl\u0027s counter\n assignment interface\n\n - Fix a quirk in the MPAM bandwidth counting on Nvidia T241 so that\n it also applies to 63 bit counters\n\n Perf:\n\n - Workarounds for hardware issues in the CMN-S3 PMU (Graviton 5) and\n CPU PMU (NVIDIA Olympus again!)\n\n - Add support for the DDR PMU on Marvell CN20K SoCs\n\n - Add support for Picoheart implementations of the DCW PCIe PMU\n\n - Add support for Channel/Rank/Bank filtering in the CXL PMU driver\n\n - Add support for 64-bit counters in the CSPMU device\n\n - Add support for revision 2 of the CMN S3 PMU\n\n Ptrace:\n\n - Fix a decade-old bug in our handling of seccomp and tracing on\n syscall entry\n\n - Fix regset handling for inactive SVE and SSVE registers\n\n Selftests\n\n - Add some tests for the decade-old bug that we just tried to fix in\n our syscall entry path\n\n - Fix SVE test crash on SME-only CPUs\"\n\n* tag \u0027arm64-upstream\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux: (95 commits)\n arm64/efi: Avoid voluntary preemption with efi_mm installed\n arm64: bti: Disable in-kernel BTI with recent versions of Clang\n arm64: entry: Avoid unnecessary local_irq_disable() on kernel exit\n irqchip/gic-v3: make the unmasking of pseudo-NMIs explicit when handling IRQs\n arm64: Disable KCSAN instrumentation in delay.o\n arm_mpam: Disable driver unbind to avoid UAF\n arm_mpam: Fix a NULL pointer dereference on unbinding after an error interrupt\n perf: arm_pmuv3: Zero initialize hw_id branch stack field\n arm64: mm: Unmap kernel data/bss entirely from the linear map\n iommu/arm-smmu-v3-sva: Use system_supports_bbml3() to detect CPU feature\n perf/arm-cmn: Support CMN S3 r2\n perf/arm-cmn: Plumb in new filter types\n perf/arm-cmn: Refactor event filter data\n perf/arm-cmn: Refactor event filter programming\n perf/arm-cmn: Rename filter variables for clarity\n arm64: mm: fix accidental linear mapping of no-map reserved memory\n tools: Ensure tools copy of linux/filter.h exports the UAPI\n kselftest/arm64: Fix abi test compilation errors\n arch: arm64: add early_param idle\u003d\u003cwfi|yield|nop\u003e\n arm64: entry: mask DAIF before returning from C EL1 handlers\n ...\n" }, { "commit": "d24f5cdbeff8b6a063fca92d0a1f94122a799b59", "tree": "1c0c344db3c18572d7176c7caec4a8abdd295647", "parents": [ "2df813c9a6b6704b75aec7538d2225292e6b6fe6", "47f27155f17498fccb1f222f79089642337498a9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 11:25:09 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 11:25:09 2026 -0700" }, "message": "Merge tag \u0027rust-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux\n\nPull Rust updates from Miguel Ojeda:\n \"Toolchain and infrastructure:\n\n - Warn when using \u0027bindgen\u0027 \u003c 0.72.1 with \u0027libclang\u0027 \u003e\u003d 22, since\n that combination may fail to build. It includes a probe for the bug\n in case \u0027bindgen\u0027 happens to be patched, and tests\n\n In parallel, Nathan updated the instructions for the kernel.org\n LLVM+Rust toolchains so that the latest version of \u0027bindgen\u0027 is\n installed, which should avoid some of these situations\n\n - Support testing \u0027rust_is_available.sh\u0027 with \u0027bash\u0027 as \u0027/bin/sh\u0027\n\n - Fix an objtool warning by adding one more \u0027noreturn\u0027 function for\n Rust 1.99.0 (expected 2026-10-01)\n\n - Fix build error in the \u0027rusttest\u0027 target due to ambiguity when the\n \u0027rustc-dev\u0027 component is installed, which was uncovered by the work\n to support Rust\u0027s GCC backend (\u0027rustc_codegen_gcc\u0027)\n\n - Fix future Clang warnings in the upcoming powerpc support due to\n macro redefinitions in the UAPI helper header by including the\n arch-aware \u0027ioctl.h\u0027 header\n\n \u0027kernel\u0027 crate:\n\n - Rework module ownership support:\n\n - Move the module-related types into a new \u0027module\u0027 module and\n make the \u0027THIS_MODULE\u0027 pointer a constant of \u0027ModuleMetadata\u0027\n so that modules can provide the pointer in const contexts, and\n add a \u0027this_module\u0027 \u0027const fn\u0027 to retrieve it\n\n This was enabled by upstream Rust\u0027s work on the \u0027const_mut_refs\u0027\n and \u0027const_refs_to_static\u0027 features which were stabilized back\n in Rust 1.83.0\n\n - Teach \u0027#[vtable]\u0027 to associate implementations with their\n owning module, defaulting to the local one, including fallbacks\n for doctests, uses within the \u0027kernel\u0027 crate (like upcoming\n KUnit \u0027#[test]\u0027s for DRM) and \u0027rusttest\u0027\n\n - Set \u0027fops.owner\u0027 from the module pointer for DRM and\n miscdevice\n\n - Migrate Rust Binder and configfs away from the old\n \u0027THIS_MODULE\u0027 \u0027static\u0027 and finally remove it from the \u0027module!\u0027\n macro\n\n - \u0027num\u0027 module:\n\n - Add the new \u0027casts\u0027 module for lossless integer conversions\n\n Rust\u0027s \u0027core\u0027 library\u0027s \u0027From\u0027 implementations do not cover\n conversions that are not portable or future-proof. However, the\n kernel supports a narrower set of architectures, which makes it\n helpful to provide more infallible conversions, instead of\n having developers use \u0027as\u0027 casts, which carry the risk of\n silently losing data\n\n This goes along with previous work we did to avoid casts in\n Rust kernel code since they are more powerful than needed\n\n Thus, provide safe \u0027const\u0027 conversion functions (e.g.\n \u0027usize_as_u64\u0027 and \u0027u64_into_u8\u0027), as well as the\n \u0027FromSafeCast\u0027 and \u0027IntoSafeCast\u0027 extension traits that provide\n conversions that are known to be lossless in the kernel, and an\n \u0027arch\u0027 submodule defining conversions that are known to be\n lossless on particular architectures (e.g. 64-bit platforms).\n For instance:\n\n // Conversion in const context.\n const USIZED_CONST: usize \u003d u8_as_usize(255u8);\n\n // Non-const conversions.\n let a \u003d u64::from_safe_cast(4096usize);\n let b: u64 \u003d 4096usize.into_safe_cast();\n\n - Add \u0027Bounded::shr_exact\u0027 method in the vein of \u0027try_shrink\u0027\n which shifts a bounded right only if it loses no set bits\n\n - Fix unsoundness issue in the \u0027Bounded::shr\u0027 method by\n rejecting, at compile-time, shifts of at least the type\u0027s bit\n width\n\n - \u0027fmt\u0027 module:\n\n - Route \u0027{:p}\u0027 raw pointer formatting through the kernel\u0027s hashed\n \u0027%p\u0027 format to prevent address leaks, including support for\n width and padding. Include tests for both \u0027no_hash_pointers\u0027\n case and the default (hashed) one\n\n - Fix the \u0027{:p}\u0027 forwarding implementation, which could print the\n address of a temporary stack variable\n\n - \u0027time\u0027 module:\n\n - Make \u0027Delta\u0027 generic over its time unit, with a default unit of\n nanoseconds (\u0027Nsec\u0027), preserving the existing behavior. Then,\n add a \u0027Jiffy\u0027 time unit\n\n - Add the \u0027Delta::as_millis_ceil()\u0027 method\n\n - Fix \u0027as_micros_ceil()\u0027 rounding near \u0027i64::MAX\u0027, which could\n yield a result one microsecond too small\n\n - \u0027sync\u0027 module:\n\n - Implement \u0027ForeignOwnable\u0027 for \u0027ARef\u003cT\u003e\u0027, allowing C code to\n own an \u0027ARef\u003cT\u003e\u0027\n\n - Add a safe abstraction for \u0027rcu_barrier()\u0027\n\n - \u0027error\u0027 module: add all of the remaining error codes, except the\n deprecated compatibility aliases\n\n - \u0027bug\u0027 module:\n\n - Fix build error on UML in \u0027warn_on!\u0027 for callers from within\n the \u0027kernel\u0027 crate\n\n - Fix future \u0027dead_code\u0027 warning on arm and loongarch64 and under\n \u0027CONFIG_BUG\u003dn\u0027 in \u0027warn_on!\u0027, which would trigger with the\n upcoming SRCU abstractions\n\n - Fix future build error in \u0027rusttest\u0027 on cross-compilation\n cases, which would trigger when \u0027warn_on!\u0027 has callers inside\n the \u0027kernel\u0027 crate\n\n - \u0027bitfield\u0027 module: fix build error for the upcoming support for\n Rust\u0027s GCC backend (\u0027rustc_codegen_gcc\u0027) by always inlining a\n couple conversions used in tests\n\n \u0027pin-init\u0027 crate:\n\n - User-visible changes:\n\n - Merge the \u0027__pinned_init\u0027 and \u0027__init\u0027 methods and make \u0027Init\u0027\n a marker trait\n\n - Introduce public APIs \u0027raw_init\u0027 and \u0027raw_try_init\u0027 to prevent\n users from needing to invoke the internal \u0027__pinned_init\u0027 and\n \u0027__init\u0027 methods\n\n - Emit errors for duplicate \u0027#[pin]\u0027 attributes\n\n - Link \u0027Zeroable::zeroed\u0027 and \u0027pin_init::zeroed\u0027 in documentation\n\n - Other changes:\n\n - Fix unwind safety issues\n\n - Clean up lint \u0027allow\u0027 and \u0027expect\u0027s\n\n - Overhaul \u0027#[cfg]\u0027 handling to pave the way for tuple structs\n and self-referential structs\n\n - Mark many functions as \u0027#[inline]\u0027 for better codegen with \u0027-C\n opt-level\u003ds\u0027 (\u0027CC_OPTIMIZE_FOR_SIZE\u0027)\n\n \u0027MAINTAINERS\u0027:\n\n - Update \u0027MODULE SUPPORT\u0027 to cover the new \u0027module\u0027 module\n\n And some other fixes, cleanups and improvements\"\n\n* tag \u0027rust-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux: (54 commits)\n rust: add functions and traits for lossless integer conversions\n rust: kernel: add `LocalModule` fallback for `#[vtable]` `impl`s\n rust: fmt: route {:p} through HashedPtr to prevent address leaks\n rust: fmt: fix {:p} printing stack addresses\n rust: module: update MAINTAINERS to cover module.rs\n rust: macros: remove `THIS_MODULE` static from `module!`\n rust_binder: use `LocalModule` for `THIS_MODULE`\n rust: configfs: use `LocalModule` for `THIS_MODULE`\n rust: miscdevice: set fops.owner from driver module pointer\n rust: drm: set fops.owner from driver module pointer\n rust: macros: auto-insert OwnerModule in #[vtable]\n rust: doctest: add LocalModule fallback for #[vtable] ThisModule\n rust: module: add `THIS_MODULE` const to `ModuleMetadata` trait\n rust: module: move module types into `module.rs`\n rust: num: add Bounded::shr_exact\n rust: num: reject Bounded::shr overshifts at build time\n rust: num: use const_assert! in Bounded\n rust: uapi: replace direct asm-generic/ioctl.h include with linux/ioctl.h\n rust: time: add Delta::as_millis_ceil()\n rust: time: add jiffies time unit for Delta\n ...\n" }, { "commit": "2df813c9a6b6704b75aec7538d2225292e6b6fe6", "tree": "72d9ebe17b67200d98f7fe68223504fb4b4f508c", "parents": [ "2063dd9d0b5a495d93f466f698bf0329809647ee", "d330fb86a7170f845123ae82d95df440fad9b707" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 11:19:37 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 11:19:37 2026 -0700" }, "message": "Merge tag \u0027for-linus-7.3-rc1-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip\n\nPull xen updates from Juergen Gross:\n\n - Small cleanups for the Xen ACPI pad driver and the gnttab driver\n\n - Fix an issue with Xen PV device initialization seen with QubesOS\n tests\n\n - Fixes for the Xen balloon driver and the xenbus driver\n\n - Simplify Xen related kernel configuration\n\n* tag \u0027for-linus-7.3-rc1-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:\n xenbus: Unregister reboot notifier on init failure\n x86/xen: Drop CONFIG_XEN_PVHVM_SMP\n xen: Drop CONFIG_XEN_AUTO_XLATE\n xen: Drop CONFIG_XEN_PVHVM\n x86/xen: Remove redundant config dependency on X86_LOCAL_APIC\n x86/xen: fix init of balloon stats again\n xen/xenbus: check otherend_id only after it has been initialized\n xen/xenbus: log more information when device state got reset\n Xen/gnttab: adjust two uses of sizeof()\n ACPI: PAD: xen: Stop setting acpi_device_name/class()\n" }, { "commit": "2063dd9d0b5a495d93f466f698bf0329809647ee", "tree": "b7ea54e4ed2f2a5fec477d376d13ff66929e48d5", "parents": [ "7b24dd46a70e94d8db83b50a04d46690fac216d0", "b58afc8d1cf85688601b8d7e271125fff666f661" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 10:50:44 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 10:50:44 2026 -0700" }, "message": "Merge tag \u0027nolibc-20260814-for-7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/nolibc/linux-nolibc\n\nPull nolibc updates from Thomas Weißschuh:\n\n - New architecture: Alpha\n\n - New library functionality: readlink(), getcwd()\n\n - Various bugfixes and cleanups\n\n* tag \u0027nolibc-20260814-for-7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/nolibc/linux-nolibc:\n tools/nolibc: add support for Alpha\n tools/nolibc/powerpc: mark ctr and xer as clobbered by system call\n tools/nolibc: remove dead __ARCH_WANT_SYS_OLD_SELECT\n selftests/nolibc: add debug information\n tools/nolibc: mark arg1 operand in __nolibc_syscall0() as write-only\n selftests/nolibc: Add test for getcwd() and readlink()\n tools/nolibc: unistd: Add readlink()\n tools/nolibc: unistd: Add getcwd()\n" }, { "commit": "7b24dd46a70e94d8db83b50a04d46690fac216d0", "tree": "b03a5f0521d65701a9e0dc5717af94ca38755baf", "parents": [ "ba24659b1dbee90d4ce7ffc7577e299f085533a1", "bcf2a0b55e322a663c9602c9e773770b87d3e629" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 10:43:01 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 10:43:01 2026 -0700" }, "message": "Merge tag \u0027liveupdate-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/liveupdate/linux\n\nPull liveupdate updates from Mike Rapoport:\n \"Kexec Handover:\n\n - Fix size calculation in kho_preserved_memory_reserve() for\n preservations larger than 2 GiB\n\n Live Update Orchestrator:\n\n - move liveupdate selftest utilities into a library so that selftests\n of subsystems participating in liveupdate, e.g. PCI and VFIO, can\n use them and drop direct ioctl calls from the tests\n\n - add end to end liveupdate test infrastructure that allows running\n the tests across a kexec in QEMU\n\n - remove redundant INIT_LIST_HEAD in luo_session_alloc()\n\n - remember the error status of an FLB retrieve() and return it on\n subsequent attempts rather than retrying retrieve() with an FLB in\n an unexpected state\n\n - reference count the outgoing FLB so that it cannot be freed while a\n caller is using it, the same way it\u0027s done for the incoming FLB\n\n - reject nonzero reserved field in LIVEUPDATE_SESSION_FINISH so that\n it can be reused by a future extension\"\n\n* tag \u0027liveupdate-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/liveupdate/linux:\n kho: fix size calculation in kho_preserved_memory_reserve()\n selftests/liveupdate: Move luo_test_utils.* into a reusable library\n selftests/liveupdate: Use luo_test_utils.c for liveupdate ioctl APIs\n liveupdate: Remember FLB retrieve() status\n liveupdate: Reference count outgoing FLB data\n liveupdate: reject nonzero reserved value for SESSION_FINISH\n liveupdate: Remove redundant INIT_LIST_HEAD in luo_session_alloc\n selftests/liveupdate: add end to end test infrastructure and scripts\n" }, { "commit": "ba24659b1dbee90d4ce7ffc7577e299f085533a1", "tree": "6b928f169d4cd020133f0ad13f8a29ae6cbc90f8", "parents": [ "38fda1d9d2f5df55bd1c095aec07de3699091316", "9d50f7c5ad5b4b031c9b85f9a43c626d8268ab51" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 10:28:28 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 10:28:28 2026 -0700" }, "message": "Merge tag \u0027kexec-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/liveupdate/linux\n\nPull kexec updates from Mike Rapoport:\n\n - Deduplicate crash memory allocation and the exclusion of reserved\n crash kernel regions from architecture specific code into a generic\n crash_prepare_headers() and enable crashkernel CMA reservation on\n arm64 and riscv reservation on arm64 and riscv.\n\n - Skip purgatory checksum verification when the kexec segments cannot\n be corrupted by DMA, which saves about 250ms on kexec.\n\n - Replace __ASSEMBLY__ with the compiler provided __ASSEMBLER__ in\n include/linux/kexec.h.\n\n - Fix a keyring refcount imbalance in the kdump kernel\u0027s dm-crypt key\n restore path, which over-dropped the user keyring reference when\n more than one key was restored.\n\n* tag \u0027kexec-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/liveupdate/linux:\n crash_dump: release keyring reference at the correct time\n kexec: Replace __ASSEMBLY__ with __ASSEMBLER__ in header file\n kexec_file: skip checksum verification when safe\n riscv: kexec_file: Add support for crashkernel CMA reservation\n arm64: kexec_file: Add support for crashkernel CMA reservation\n powerpc/kexec_file: Use crash_exclude_core_ranges() helper\n LoongArch: kexec_file: Use crash_prepare_headers() helper to simplify code\n riscv: kexec_file: Use crash_prepare_headers() helper to simplify code\n x86/crash: Use crash_prepare_headers() helper to simplify code\n arm64: kexec_file: Use crash_prepare_headers() helper to simplify code\n crash: Add crash_prepare_headers() to exclude crash kernel memory\n powerpc/crash: sort crash memory ranges before preparing elfcorehdr\n riscv: kexec_file: Fix crashk_low_res not exclude bug\n" }, { "commit": "38fda1d9d2f5df55bd1c095aec07de3699091316", "tree": "378f5915d09f73e573941934067f29289dfb4cc0", "parents": [ "3abe3d0223dd1ceb3af6543b71db3856762603a8", "b3c1ec1a9ad84ec2f14d343d28c6032feaec5b79" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 10:03:54 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 10:03:54 2026 -0700" }, "message": "Merge tag \u0027memblock-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rppt/memblock\n\nPull memblock updates from Mike Rapoport:\n \"Non-urgent fixes:\n\n - Fix calculation of node_spanned_pages when running\n with \u0027kernelcore\u003dmirror\u0027\n\n - Properly handle failure to allocate per_cpu_nodestats\n in free_area_init_core_hotplug()\n\n - Fix deferred initialization of the memory map for\n configurations where node\u0027s RAM end is not aligned\n on PAGES_PER_SECTION\n\n Cleanups:\n\n - Remove redundant pageblock_align() call in free_unused_memmap()\n\n - Remove unnecessary invalid range checks in users of memblock\n iterators. Some users of for_each_mem_range() and\n for_each_mem_pfn_range() verify that start \u003c end for each range.\n\n This is redundant because memblock iterators guarantee to never\n return an invalid range\n\n - Stop overlapping zones with \u0027kernelcore\u003dmirror\u0027 and align behaviour\n of \u0027kernelcore\u003dmirror\u0027 with other variants of kernelcore and\n movablecore\n\n - Remove redundant updates of numa_nodes_parsed mask in the callers\n of numa_add_memblk(), the latter always updates the mask anyway\n\n - Remove unnecessary initialization of pgdat-\u003eper_cpu_nodestats to\n NULL, the variable is reset to the actual value a few lines below\"\n\n* tag \u0027memblock-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rppt/memblock: (25 commits)\n mm/mm_init: deferred_grow_zone(): fix out-of-range first_deferred_pfn\n mm/mm_init: remove unnecessary initialization of pgdat-\u003eper_cpu_nodestats\n mm/mm_init: remove redundant memset in free_area_init()\n mm: numa_memblks: use numa_add_reserved_memblk() in numa_cleanup_meminfo()\n arch_numa: remove redundant node_possible_map assignment\n mm: numa_memblks: remove redundant numa_nodemask_from_meminfo()\n LoongArch: remove redundant numa_nodes_parsed node_set()\n arch_numa: remove redundant numa_nodes_parsed node_set()\n x86/numa: remove redundant numa_nodes_parsed node_set()\n of/numa: remove redundant numa_nodes_parsed node_set()\n ACPI: NUMA: remove redundant numa_nodes_parsed node_set()\n mm: numa_memblks: set numa_nodes_parsed in numa_add_memblk()\n mm/mm_init: handle alloc_percpu failure in free_area_init_core_hotplug\n mm/mm_init: drop overlap_memmap_init()\n mm/mm_init: don\u0027t overlap NORMAL and MOVABLE zones with kernelcore\u003dmirror\n mm/hugetlb: remove unnecessary empty range check in hugetlb_bootmem_set_nodes()\n mm: remove unnecessary empty range check in early_calculate_totalpages()\n powerpc64/kasan: Remove unreachable invalid range check in kasan_init_phys_region()\n ARM: remove unreachable invalid range check in kasan_init()\n riscv: remove unreachable invalid range check in kasan_init()\n ...\n" }, { "commit": "3abe3d0223dd1ceb3af6543b71db3856762603a8", "tree": "9355bcb2b4508890c2e4ae94b71edb1f72a81d54", "parents": [ "c5c7a47af8d93059f45b6f656a5115e020dec477", "d900723d78adec229996aefbab0dcaa66a177b77" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 09:26:31 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 09:26:31 2026 -0700" }, "message": "Merge tag \u0027kbuild-7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kbuild/linux\n\nPull Kbuild/Kconfig updates from Nicolas Schier:\n \"Kbuild updates:\n\n - Use --force-group-allocation when linking modules\n\n Have the linker resolve the COMDAT groups and place their members\n as regular sections instead of possibly leaving multiple copies in\n the resulting modules and unnecessary group metadata.\n\n - UAPI header files: Canonicalize __ASSEMBLER__ / __ASSEMBLY__ mixed\n use to __ASSEMBLER__\n\n There is an ongoing effort to change __ASSEMBLY__ to __ASSEMBLER__\n treewide. For consistency, UAPI headers are normalised to use\n __ASSEMBLER__ only. Normalisation is done in two subsequent patches\n to simplify a revert in the unexpected case of a regression report.\n\n - link-vmlinux.sh: Improve detection of third pass requirement\n\n - modpost: Canonicalize format of warnings and errors\n\n - Minor changes:\n - Remove srctree path from CHECK output\n - Set the initial value of subdir-rustflags-y\n - Remove broken and unused modules.builtin(.modinfo) targets from\n the top-level Makefile\n - Add symbol size for kallsyms symbols that can change size\n - modpost: Prevent leak when early return no suffix .o in\n read_symbols()\n - scripts/config: Update usage of POSIX sed\n - \u0027make tags\u0027: Add support for rust source files and prevent\n binary files from being analysed\n - Several spelling mistakes and rephrasing\n\n Kconfig updates:\n\n - Add Julian Braha as Kconfig reviewer\n\n - Fix submenu rendering of negative dependencies\n\n - Minor changes:\n - merge_config.sh: Keep temp file in the output dir\n - Abort rather than loop for ever on EOF\"\n\n* tag \u0027kbuild-7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kbuild/linux: (23 commits)\n modpost: use mod_warn() and mod_error(), clean up logging\n modpost: add module as parameter to modpost_log()\n kconfig: fix submenu rendering of negative dependencies\n kbuild: link-vmlinux.sh: improve detection of third pass requirement\n kallsyms: add symbol size for kallsyms symbols that can change size\n kbuild: fix modules.builtin(.modinfo) targets in the top-level Makefile\n kbuild: set the initial value of subdir-rustflags-y\n scripts/config: Use in-place editing (-i) in sed portably\n scripts/config: Use POSIX standard ERE (-E) in sed\n modpost: prevent leak when early return no suffix .o in read_symbols()\n usr: Correct a spelling by changing a letter\n fixdep: make gendered language gender-neutral\n kconfig: fix minor typos in comments\n scripts: fix spelling mistakes\n kconfig: abort rather than loop for ever on EOF\n scripts/tags.sh: Add support for rust source files\n scripts/tags.sh: Prevent binary files appearing in cscope.files\n MAINTAINERS: add Julian Braha as Kconfig reviewer\n scripts: headers_install.sh: Normalize __ASSEMBLY__ to __ASSEMBLER__\n scripts: headers_install.sh: Normalize __ASSEMBLER__ to __ASSEMBLY__\n ...\n" }, { "commit": "c5c7a47af8d93059f45b6f656a5115e020dec477", "tree": "ec1042ba611b4081ab195a32289f6bd7761d3d82", "parents": [ "a5778046a02570fd23bdf0b48dce330237d1996c", "2d5667b6c3d7adf0f8551bbac7c707fb87155611" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 08:59:40 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 08:59:40 2026 -0700" }, "message": "Merge tag \u0027thermal-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm\n\nPull thermal control updates from Rafael Wysocki:\n \"These include an introduction of Intel Directed Package Thermal\n Interrupt support into the thermal throttling driver for Intel\n processors, probe failure code path fixes and code cleanups in Intel\n thermal drivers, a thermal core fix related to hwmon, a sysfs-related\n cleanup of that code, and a thermometer utility fix:\n\n - Add support for the Directed Package-level Thermal Interrupt to the\n Intel thermal throttling driver to allow package-level thermal\n interrupts to go to one specific CPU in a processor package instead\n of going to all of the CPUs in it (Ricardo Neri)\n\n - Remove hwmon class devices created for thermal zones when the\n thermal zone devices holding them are removed (Rafael Wysocki)\n\n - Use sysfs_emit_at() in trans_table_show() (Thorsten Blum)\n\n - Clean up RFIM groups on DVFS failure and clean up ODVP on probe\n failures in the int340x thermal driver (Pengpeng Hou)\n\n - Remove redundant dev_err() from the int340x thermal driver and the\n bxt_pmic driver (Pan Chuang)\n\n - Simplify ptc_temperature_write() in the int340x thermal driver by\n using kstrtou32_from_user() (Dmitry Antipov)\n\n - Close fd on realloc() failure in the thermometer utility (Amarjeet)\"\n\n* tag \u0027thermal-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:\n thermal: hwmon: Remove hwmon class device along with its parent\n thermal: sysfs: Use sysfs_emit_at() in trans_table_show()\n tools/thermal/thermometer: close fd on realloc() failure\n thermal: intel: int340x: simplify ptc_temperature_write()\n thermal: intel: bxt_pmic: Remove redundant dev_err()\n thermal: intel: int340x: Remove redundant dev_err()\n thermal: intel: int3400: clean up ODVP on probe failures\n thermal: intel: int340x: clean up RFIM groups on DVFS failure\n thermal: intel: Add a syscore shutdown callback for kexec reboot\n thermal: intel: Add syscore callbacks for suspend and resume\n thermal: intel: Enable the Directed Package-level Thermal Interrupt\n thermal: intel: Add resources to handle directed package-level thermal interrupts\n x86/thermal: Add bit definitions for Intel Directed Package Thermal Interrupt\n" }, { "commit": "a5778046a02570fd23bdf0b48dce330237d1996c", "tree": "4c9d59146cd9a292b63777f803cee958a05e3f59", "parents": [ "0d508f1745b57272d4d4a3d50dc35aa2a3682238", "7544a367adc1ce5e9cc62315b4c32dc72b509451" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 08:47:09 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 08:47:09 2026 -0700" }, "message": "Merge tag \u0027pm-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm\n\nPull power management updates from Rafael Wysocki:\n \"As has been the case for quite some time, this set of changes is\n dominated by cpufreq updates including intel-pstate and amd-pstate\n driver updates, minor fixes and cleanups of other assorted cpufreq\n drivers, schedutil governor updates, fixes of the Rust bindings, new\n hardware support (IPQ5210 in qcom-nvmem), and some updates of self\n tests related to cpufreq.\n\n The second largest group of changes are cpuidle updates consisting of\n intel_idle driver updates and ACPI processor idle driver updates, both\n mostly related to ACPI _LPI support.\n\n There are also updates related to system sleep, mostly in the\n hibernation core code, two operating performance points (OPP) updates,\n one runtime PM framework update, one power capping update, and some\n tools updates including the addition of ACPI CPPC support to cpupower.\n\n Specifics:\n\n - Minor fixes and cleanups in assorted cpufreq drivers (Dan\n Carpenter, Guru Das Srinagesh, Haoxiang Li, Karl Mehltretter, Sasha\n Finkelstein, and Pan Chuang)\n\n - Fix cpufreq table creation and bios_limits() callback in the Rust\n bindings (Priya Bala Govindasamy)\n\n - Add IPQ5210 support to qcom-nvmem driver (Varadarajan Narayanan)\n\n - Adjust the .adjust_perf() cpufreq driver callback to allow the\n maximum performance value to be passed to drivers and update the\n intel_pstate driver to use it (Rafael Wysocki)\n\n - Set policy-\u003ecur to the actual requested frequency in the\n intel_pstate driver when the performance policy is used (Rafael\n Wysocki)\n\n - Simplify HWP handling on Broadwell processors in intel_pstate\n (Rafael Wysocki)\n\n - Fix setting minimum P-state at init time in intel_pstate (Rafael\n Wysocki)\n\n - Consolidate frequency values computation in intel_pstate and clean\n up code in that driver (Rafael Wysocki)\n\n - Add missing kernel-doc descriptions for structure and union members\n in the amd-pstate driver (David Vernet)\n\n - Handle missing policy in dynamic EPP callbacks in the amd-pstate\n driver (EDAMAMEX)\n\n - Introduce EXPORT_SYMBOL_FOR_PSTATE_UT() to export amd-pstate driver\n symbols to the amd-pstate-ut subdriver (K Prateek Nayak)\n\n - Add dynamic EPP as an \"energy_performance_preference\" mode in\n amd-pstate, remove the \"amd_dynamic_epp\" kernel command line option\n and the \"dynamic_epp\" sysfs attribute, and update the dynamic_epp\n documentation accordingly (K Prateek Nayak)\n\n - Add unit tests for CPPC Performance Priority and the \"dynamic\" EPP\n mode in the amd-pstate driver (K Prateek Nayak)\n\n - Set min_limit_freq based on bios_min_perf in amd-pstate and remove\n the defensive check for bios_min_perf from it (K Prateek Nayak)\n\n - Fix EPP return type and handle errors in amd-pstate during\n initialization, toggle auto_sel in active mode on shared memory\n systems, and cache the firmware programmed EPP value (Marco\n Scardovi)\n\n - Skip tests in amd-pstate-ut if the amd-pstate driver is not in\n active use (Qianheng Peng)\n\n - Replace sprintf() with sysfs_emit() in sysfs show in the cpufreq\n schedutil governor and fix a self-contradictory comment in\n sugov_iowait_apply() (Zhongqiu Han)\n\n - Fix the usage example for the sampling_rate tunable of the ondemand\n cpufreq governor in admin-guide (wangxiaodong)\n\n - Avoid using deep idle states during initialization in the\n intel_idle driver to work around device handling issues (Rafael\n Wysocki)\n\n - Fix and refactor the ACPI processor driver code related to ACPI\n _LPI support and add ACPI _LPI support to intel_idle based on that\n ACPI processor driver update (Rafael Wysocki)\n\n - Backup and restore governor for cpufreq sptests (Yiwei Lin)\n\n - Remove unnecessary sudo from quick_shuffle() and remove unused\n local variables from switch_show_governor() in cpufreq selftests\n (Jinseok Kim)\n\n - Rename the PM core module parameter prefix to \"pm\" and allow the PM\n transition (DPM) watchdog to be disabled by default (Tzung-Bi Shih)\n\n - Fix off-by-one in wakelocks number limit check in the system sleep\n sysfs interface (Haowen Tu)\n\n - Remove kernel-doc markings from helper descriptions in the core\n hibernation code (Adi Nata)\n\n - Use %pe to print error pointer values in the hibernation core\n (Ronan Marchal)\n\n - Fix memory leak in snapshot_write_next() error path (Malaya Kumar\n Rout)\n\n - Delay allocating and linking the next swap_map_page in the\n hibernation image saving code until another image page actually\n needs to be recorded (Haesung Kim)\n\n - Fix cleanup ordering around scope-based pointers in OPP (Gregor\n Herburger).\n\n - Use clk_get_optional() for optional clocks in OPP (Praveen Talari).\n\n - Stop setting runtime_error on runtime resume callback failures to\n allow drivers to recover from resume issues (Praveen Talari)\n\n - Handle PMU registration failure during probe in the intel_rapl_tpmi\n driver (Sumeet Pawnikar)\n\n - Avoid optional imports in intel_pstate_tracer unless they are\n really needed (Yousef Alhouseen)\n\n - Add generic CPPC performance display to the cpupower utility, build\n and call CPPC information on non-AMD processors, make cpupower\n print kernel and hardware frequency information, and add libm to\n cpupower for generic CPPC view (Jeremy Linton)\n\n - Remove conditional return with no effect from cpupower (Sang-Heon\n Jeon)\"\n\n* tag \u0027pm-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (76 commits)\n cpufreq: imx6q: fix out-of-bounds write when probed more than once\n cpufreq: imx6q: fix devres accumulation across driver rebind\n rust: cpufreq: Fix temporary write in Registration::bios_limit_callback\n rust: cpufreq: Add CPUFREQ_TABLE_END as last table entry in TableBuilder::to_table\n opp: Use clk_get_optional() to avoid leaving opp_table-\u003eclk as an error pointer\n intel_idle: Avoid using deep idle states during initialization\n cpupower: remove conditional return with no effect\n cpufreq: intel_pstate: Adjust policy-\u003ecur in active mode to policy\n cpufreq/amd-pstate: Document missing kernel-doc members\n cpufreq/amd-pstate-ut: Add unit test for CPPC Performance Priority\n cpufreq/amd-pstate-ut: Add unit test for \"dynamic\" EPP mode\n cpufreq/amd-pstate: Reduce the scope of exported symbols\n Documentation/amd-pstate: Update dynamic_epp documentation with new behavior\n cpufreq/amd-pstate: Remove \"amd_dynamic_epp\" cmdline and \"dynamic_epp\" sysfs\n cpufreq/amd-pstate: Add dynamic EPP as an \"energy_performance_preference\" mode\n cpufreq/amd-pstate: Extract platform profile to EPP conversion into a helper\n cpufreq/amd-pstate: Remove the defensive check for bios_min_perf\n cpufreq/amd-pstate: Set min_limit_freq based on bios_min_perf\n powercap: intel_rapl_tpmi: Handle PMU registration failure during probe\n PM: sleep: Allow disabling DPM watchdog by default\n ...\n" }, { "commit": "0d508f1745b57272d4d4a3d50dc35aa2a3682238", "tree": "1240ab8a8a1c02556dafe32b6cb09d646be07988", "parents": [ "0f23d56f17fdfc7db69d51f64c8b91bbab947aa9", "79ba234b396b0fdcb3a82dd8aa872348f4069aae" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 08:29:30 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 18 08:29:30 2026 -0700" }, "message": "Merge tag \u0027acpi-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm\n\nPull ACPI support updates from Rafael Wysocki:\n \"The most significant change here is the elimination of struct\n acpi_driver that has no more users in the tree now along with some\n documentation related to it, and a follow-up update to set the \"no PM\"\n flag for all ACPI devices that are now only going to play the role of\n other devices\u0027 \"companions\" (in analogy with DT nodes).\n\n There is also a significant update of irqchip code related to ACPI\n done in order to enable GICv5 IWB ACPI probe ordering detection on\n ARM, which involves RISC-V interrupt controller management code\n refactoring to extract generic code from it into the common ACPI IRQ\n code.\n\n The rest is mostly fixes, including some fallout of the _OSC handling\n rework in 7.0, ACPI CPPC library fixes, a workaround for registering\n ACPI platform devices with overlapping I/O or memory resources, an\n ACPI EC driver fix related to probe deferral on platforms using\n HW-reduced ACPI, two ACPI battery driver fixes and a workaround for\n handling model numbers with unprintable characters in it, probe error\n cleanup and driver unload code path fixes, hardware error reporting\n fixes, documentation fixes, and assorted code cleanups all over.\n\n Specifics:\n\n - Eliminate struct acpi_driver whose users have all been converted to\n bind to platform devices or auxiliary devices and set the \"no power\n management\" flag for all struct acpi_device objects (Rafael\n Wysocki)\n\n - Avoid complaints regarding missing _OSC features on platforms where\n OSC_CAPABILITIES_MASK_ERROR is set in _OSC error bits even though\n all of the requested features are actually acknowledged (Rafael\n Wysocki)\n\n - Avoid printing confusing _OSC messages for non-PCIe host bridges\n without _OSC which is a valid configuration (Kazuma Kondo)\n\n - Use correct region struct for BERT region size check and properly\n map BERT and CCEL data to their ACPI tables (Thomas Renninger)\n\n - Add acpi_device_clear_deps(), refactor RISC-V interrupt controller\n management code to extract generic code from it into the common\n ACPI IRQ code, and enable GICv5 IWB ACPI probe ordering detection\n on ARM on top of that (Lorenzo Pieralisi)\n\n - Stop using acpi_device_name() in the PNP core, stop setting\n acpi_device_name/class() in the Xen variant of the ACPI PAD\n (Processor Aggregator Device) driver, and make the Loongarch laptop\n driver stop setting acpi_device_class() (Rafael Wysocki)\n\n - Fix issues related to the desired_perf register access in the ACPI\n CPPC library and update it to avoid unnecessary overhead (Christian\n Loehle)\n\n - Simplify acpi_get_pci_dev() with the help of a mutex guard,\n introduce acpi_dev_get_pci_dev() for code that has a struct ACPI\n device for which it wants to get the struct pci_dev pointer of the\n associated PCI device, and use it in the ACPI video bus driver\n (Rafael Wysocki)\n\n - Avoid registering platform devices with resource overlaps in the\n ACPI core device enumeration code (Rafael Wysocki)\n\n - Clean up the list of included header files in the NHLT table parser\n and validate the table and record lengths in the FPDT parser (Andy\n Shevchenko and Pengpeng Hou)\n\n - Unregister the cpufreq notifier on init failure in the ACPI\n processor driver (Can Peng)\n\n - Validate MADT IOAPIC entry bounds during IOAPIC hotplug lookup in\n the ACPI processor driver (Pengpeng Hou)\n\n - Avoid _REG disconnect on probe deferrals related to GPIO IRQ in the\n ACPI EC driver (Zhu Ling)\n\n - Update kerneldoc comments of two structures in the ACPI bus type\n code to use correct struct member names to avoid warnings (Randy\n Dunlap)\n\n - Use a correct function parameter name in kernel-doc in the ACPI fan\n driver (Randy Dunlap)\n\n - Update ACPI fan IDs to follow modern style and clean up header file\n inclusions in the ACPI fan driver (Andy Shevchenko)\n\n - Use devm_acpi_install_notify_handler() to replace a custom\n open-coded devres-based management of an ACPI notify handler in the\n ACPI fan driver (Rafael Wysocki)\n\n - Adjust charging status validation check in the ACPI battery driver\n to avoid incorrect status reporting (Rafael Wysocki)\n\n - Merge consecutive battery notifications in the ACPI battery driver\n to reduce the pressure on STA, _BST and _BIX/_BIF ACPI control\n methods and make that driver use kstrtoul() instead of\n sscanf(\"%lu\\n\") (Rong Zhang)\n\n - Sanitise model_number in the ACPI battery driver by dropping\n unprintable characters (Kate Hsuan)\n\n - Remove a node_set() call that is redundant from\n acpi_parse_memory_affinity() (Sang-Heon Jeon)\n\n - Prevent kernel-doc warnings by converting 2 function description\n comments to kernel-doc format (Randy Dunlap)\n\n - Fix docs build error in the ACPI admin-guide documentation (Randy\n Dunlap)\n\n - Replace __get_free_page() with kmalloc() in the code handling ACPI\n NVS memory during system suspend/resume (Mike Rapoport)\n\n - Fix card device cleanup on registration failure in the core PNP\n code (Yuho Choi)\n\n - Drop an unused assignment of pnp_device_id driver data (Uwe\n Kleine-König)\n\n - Clear driver_data on all paths that free acpi_pci_root in\n acpi_pci_root_add() (Chen Pei)\n\n - Add locking around evaluation of ACPI control methods in the ACPI\n TAD driver to avoid race conditions (Rafael Wysocki)\n\n - Handle repeated SEA error storms in APEI (Junhao He)\n\n - Fix ERST timeout unit conversion in APEI (Nirmoy Das)\n\n - Fix ARM section length accounting after header in the ACPI APEI\n GHES driver (TanZheng)\n\n - Mark ghes_in_nmi_spool_from_list() as maybe unused (Rui Qi)\n\n - Introduce helper function acpi_dev_is_video_device() and use it in\n the core ACPI device enumeration code, in the ACPI video bus\n driver, in the ACPI support code for I2C, in the PCI VGA driver,\n and in the x86 platform thinkpad_acpi driver (Andy Shevchenko)\n\n - Add a quirk to use the native backlight on Acer Nitro AN515-46 to\n the ACPI video bus driver (Marcos Paulo Medeiros)\n\n - Release PCI device reference after lookup in\n video_detect_portege_r100() in the ACPI video bus driver (Yuho\n Choi)\"\n\n* tag \u0027acpi-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (61 commits)\n ACPI: scan: Avoid registering platform devices with resource overlaps\n ACPI: APEI: Handle repeated SEA error storms\n ACPI: APEI: Fix ERST timeout unit conversion\n ACPI: APEI: GHES: fix ARM section length accounting after header\n ACPI: video: Release PCI device reference after lookup\n ACPI: PCI: Avoid misleading _OSC messages for non-PCIe host bridges without _OSC\n ACPI: TAD: Add locking around AML evaluations\n ACPI: video: force native backlight on Acer Nitro AN515-46\n ACPI: CPPC: Evaluate performance-control PCC use once\n ACPI: CPPC: Avoid locking standalone full-width registers\n ACPI: CPPC: Avoid unnecessary reads for full-width writes\n ACPI: CPPC: Stop reading desired_perf in cppc_get_perf()\n ACPI: CPPC: Skip desired_perf read in cppc_get_perf()\n ACPI: CPPC: Reject desired_perf reads on _CPC revision 4+\n ACPI: processor: Unregister cpufreq notifier on init failure\n ACPI: bus: Avoid confusing complaints regarding missing _OSC features\n ACPI: battery: Adjust charging status validation check\n ACPI: pmtmr: Convert to kernel-doc format\n ACPI: bus: Use correct struct member names\n ACPI: fan: Use correct function parameter name in kernel-doc\n ...\n" }, { "commit": "aed89a05a736d2570a514f00c2ab83388662b271", "tree": "074aff2ffc5d9351e6de399f06e67874b95685e4", "parents": [ "c6441becf4d064bb1dcbccbd3e2276833804cff5" ], "author": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Tue Aug 18 09:47:43 2026 +0200" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Tue Aug 18 11:41:56 2026 +0200" }, "message": "soc: qcom: make QCOM_PDR_MSG selectable\n\nSelecting QCOM_PDR_HELPERS from outside of a CONFIG_QCOM_SOC\nblock causes a build warning:\n\nWARNING: unmet direct dependencies detected for QCOM_PDR_MSG\n Depends on [n]: QCOM_SOC [\u003dn]\n Selected by [y]:\n - QCOM_PDR_HELPERS [\u003dy] \u0026\u0026 NET [\u003dy]\n\nAvoid this by allowing QCOM_PDR_MSG to be selected as well.\n\nFixes: f2866e6a27f7 (\"soc: qcom: Hide all drivers behind selectable menu\")\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "c6441becf4d064bb1dcbccbd3e2276833804cff5", "tree": "41e65cf71487a989d4c62fa189ce0524d00c2ebb", "parents": [ "f24cfafd1e11feca224d085bff515f5ba45c55a2" ], "author": { "name": "Daniel Baluta", "email": "daniel.baluta@nxp.com", "time": "Mon Aug 17 15:35:41 2026 +0300" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Tue Aug 18 11:17:52 2026 +0200" }, "message": "soc: qcom: ubwc: Fix missing include\n\nWhen CONFIG_QCOM_UBWC_CONFIG\u003dn, compiler needs to know the definition\nof ERR_PTR otherwise there will be a compilation error:\n\nIn file included from drivers/gpu/drm/msm/disp/dpu1/dpu_hw_sspp_v13.c:7:\n./include/linux/soc/qcom/ubwc.h: In function ‘qcom_ubwc_config_get_data’:\n./include/linux/soc/qcom/ubwc.h:45:16: error: implicit declaration of\nfunction ‘ERR_PTR’ [-Wimplicit-function-declaration]\n\nFix this by including \u003clinux/err.h\u003e\n\nFixes: 1924272b9ce1 (\"soc: qcom: Add UBWC config provider\")\nReviewed-by: Dmitry Baryshkov \u003cdmitry.baryshkov@oss.qualcomm.com\u003e\nReviewed-by: Konrad Dybcio \u003ckonrad.dybcio@oss.qualcomm.com\u003e\nSigned-off-by: Daniel Baluta \u003cdaniel.baluta@nxp.com\u003e\nTested-by: Nathan Chancellor \u003cnathan@kernel.org\u003e # build\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "f24cfafd1e11feca224d085bff515f5ba45c55a2", "tree": "900f1faa8e81791e5ca64d5bf6c30d3749ab03c2", "parents": [ "8afb928301700aedc9a19e9f42a0be55ee03234a" ], "author": { "name": "Daniel Baluta", "email": "daniel.baluta@nxp.com", "time": "Mon Aug 17 15:35:40 2026 +0300" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Tue Aug 18 11:17:50 2026 +0200" }, "message": "soc: qcom: ubwc: Fix link error when QCOM_SMEM\u003dn\n\nubwc_config.c calls qcom_smem_is_available() and qcom_smem_dram_get_hbb()\nunconditionally, so QCOM_UBWC_CONFIG\u003dy with QCOM_SMEM\u003dn fails to link:\n\n arm-linux-gnueabihf-ld: ubwc_config.c:(.text+0x2c): undefined\n reference to \u0027qcom_smem_is_available\u0027\n\nAdd the missing \u0027depends on QCOM_SMEM\u0027.\n\nFixes: 1b445022d1d0 (\"soc: qcom: ubwc: Get HBB from SMEM\")\nReviewed-by: Konrad Dybcio \u003ckonrad.dybcio@oss.qualcomm.com\u003e\nSigned-off-by: Daniel Baluta \u003cdaniel.baluta@nxp.com\u003e\nTested-by: Nathan Chancellor \u003cnathan@kernel.org\u003e # build\nReviewed-by: Dmitry Baryshkov \u003cdmitry.baryshkov@oss.qualcomm.com\u003e\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "8afb928301700aedc9a19e9f42a0be55ee03234a", "tree": "33067bcd168e420e535cba98c432799c8e42620a", "parents": [ "a1ee4bb25b3bd97c1f46fe2968263b54f0be3c19" ], "author": { "name": "Daniel Baluta", "email": "daniel.baluta@nxp.com", "time": "Mon Aug 17 15:35:39 2026 +0300" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Tue Aug 18 11:17:48 2026 +0200" }, "message": "media: iris: Guard the QCOM_UBWC_CONFIG select with QCOM_SMEM\n\nQCOM_UBWC_CONFIG is about to depend on QCOM_SMEM. Guard the\nVIDEO_QCOM_IRIS select with \u0027if ARCH_QCOM \u0026\u0026 QCOM_SMEM\u0027, and add\n\u0027depends on QCOM_SMEM || QCOM_SMEM\u003dn\u0027 so a built-in driver is never\nselected against a missing or modular SMEM.\n\nFixes: c43207553867 (\"media: iris: retrieve UBWC platform configuration\")\nReviewed-by: Konrad Dybcio \u003ckonrad.dybcio@oss.qualcomm.com\u003e\nSigned-off-by: Daniel Baluta \u003cdaniel.baluta@nxp.com\u003e\nTested-by: Nathan Chancellor \u003cnathan@kernel.org\u003e # build\nReviewed-by: Dmitry Baryshkov \u003cdmitry.baryshkov@oss.qualcomm.com\u003e\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "a1ee4bb25b3bd97c1f46fe2968263b54f0be3c19", "tree": "019c2bdb0ef62eab03f1645307ce60fd73b667e2", "parents": [ "4e926dc3d6ad35089ca45af6d1cee36bc24a80fb" ], "author": { "name": "Daniel Baluta", "email": "daniel.baluta@nxp.com", "time": "Mon Aug 17 15:35:38 2026 +0300" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Tue Aug 18 11:17:45 2026 +0200" }, "message": "drm/msm: Guard the QCOM_UBWC_CONFIG select with QCOM_SMEM\n\nQCOM_UBWC_CONFIG is about to depend on QCOM_SMEM. Guard the DRM_MSM\nselect with \u0027if ARCH_QCOM \u0026\u0026 QCOM_SMEM\u0027, and add\n\u0027depends on QCOM_SMEM || QCOM_SMEM\u003dn\u0027 so a built-in driver is never\nselected against a missing or modular SMEM.\n\nFixes: 1b445022d1d0 (\"soc: qcom: ubwc: Get HBB from SMEM\")\nReviewed-by: Konrad Dybcio \u003ckonrad.dybcio@oss.qualcomm.com\u003e\nAcked-by: Dmitry Baryshkov \u003cdmitry.baryshkov@oss.qualcomm.com\u003e\nReviewed-by: Dmitry Baryshkov \u003cdmitry.baryshkov@oss.qualcomm.com\u003e\nSigned-off-by: Daniel Baluta \u003cdaniel.baluta@nxp.com\u003e\nTested-by: Nathan Chancellor \u003cnathan@kernel.org\u003e # build\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "0f23d56f17fdfc7db69d51f64c8b91bbab947aa9", "tree": "41b30de106e7e62710d5fc06994551fae9e442d1", "parents": [ "fd89b0be5503dbbbdbd53c8158ea6ba0e57357fc", "649ba27dfac784427a01f9c95c09ecbcb88900d8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 20:52:03 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 20:52:03 2026 -0700" }, "message": "Merge tag \u0027linux_kselftest-next-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest\n\nPull kselftest update from Shuah Khan:\n \"Fix zram test failure in kernel_gte() when using dash and a spelling\n error in ftrace poll test comment\"\n\n* tag \u0027linux_kselftest-next-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:\n selftests/zram: fix kernel_gte() for POSIX sh\n selftests/ftrace: fix spelling error in poll test comment\n" }, { "commit": "fd89b0be5503dbbbdbd53c8158ea6ba0e57357fc", "tree": "07f204cfef318924a07eeca0959abb70fe1d2970", "parents": [ "fc8c78bce3335860ff4ae9fcfd3b2eb1f674efbb", "dea754ded9518b51740c417d2c1e02ff540784c6" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 19:44:06 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 19:44:06 2026 -0700" }, "message": "Merge tag \u0027linux_kselftest-kunit-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest\n\nPull kunit updates from Shuah Khan:\n \"Fixes and new kunit and tools, enable new configs:\n\n - configs: enable GPIO kunit test cases in all_tests.config\n\n - string-stream: Replace strlcat() with strscpy() and seq_buf\n\n - configs: enable GPIO kunit test cases in all_tests.config\n\n Documentation:\n\n - Test config entries shouldn\u0027t select other configs\n\n - Fix outdated FAQ entries\n\n Add the ability to skip entire test suites and an example test suite\n that can be skipped at runtime:\n\n - Add ability to skip entire test suites\n\n - Add example of test suite that can be skipped at runtime\"\n\n* tag \u0027linux_kselftest-kunit-7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:\n kunit: tool: fix _list_tests filtering wrong variable when list has TAP prefix\n kunit: configs: enable GPIO kunit test cases in all_tests.config\n kunit: string-stream: Replace strlcat() with strscpy() and seq_buf\n Documentation: kunit: Fix outdated FAQ entries\n Documentation: kunit: Test Kconfig entries shouldn\u0027t select other configs\n kunit: Add example of test suite that can be skipped at runtime\n kunit,rust: Add ability to skip entire test suites\n" }, { "commit": "fc8c78bce3335860ff4ae9fcfd3b2eb1f674efbb", "tree": "5997e1747226eb6d6bb16b78ea64d4a0c7f739c6", "parents": [ "d47db9bf50d28647689e6743ee93c45cb755ffc3", "561131a27f5533e6e63520b4bc43d9c832a27c09" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 19:29:41 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 19:29:41 2026 -0700" }, "message": "Merge tag \u0027libcrypto-tests-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux\n\nPull crypto library test updates from Eric Biggers:\n\n - Add comprehensive KUnit test suites for the new AES-GCM and AES-CCM\n library APIs\n\n - Add FIPS self-tests for all the AES encryption modes. This is needed\n for parity with the traditional crypto API\n\n - Fix a couple more issues in the IRQ test helper\n\n* tag \u0027libcrypto-tests-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux:\n lib/crypto: aes-cmac: Use __cleanup() instead of memzero_explicit()\n kunit: irq: Unregister on-stack timer and work from debugobjects\n kunit: irq: Continue increasing hrtimer interval for longer\n lib/crypto: tests: Add KUnit test suite for AES-GCM\n lib/crypto: tests: Add KUnit test suite for AES-CCM\n lib/crypto: tests: Add aead-test-template.h\n lib/crypto: tests: Use per-test-case buffers in hash tests\n lib/crypto: tests: Create test-utils.h\n lib/crypto: aes: Add FIPS self-tests for GCM and CCM\n lib/crypto: aes: Add FIPS self-tests for unauthenticated modes\n lib/crypto: fips: Split fips.h into fips-aes.h and fips-sha.h\n" }, { "commit": "d47db9bf50d28647689e6743ee93c45cb755ffc3", "tree": "16d422e07b769f6e720102489021f55b09567c41", "parents": [ "1d7443e4dca1e8637930f3baf64e2fe82033e669", "ff6ac629076fb3f369c6ad35d6ead1c666469b95" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 19:16:42 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 19:16:42 2026 -0700" }, "message": "Merge tag \u0027libcrypto-updates-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux\n\nPull crypto library updates from Eric Biggers:\n \"Add library APIs for most AES encryption modes that are used in the\n kernel (ECB, CBC, CBC-CTS, CTR, XCTR, XTS, GCM, CCM).\n\n These AES modes have many in-kernel users that are currently using the\n crypto_skcipher or crypto_aead APIs. These existing APIs are difficult\n to use and inefficient. Until now, the lack of proper library support\n for these has been the main gap in the crypto library.\n\n This set of changes is the next stage of addressing it:\n\n - Implement the new APIs on top of the existing support for\n single-block AES in the library.\n\n - Fully document the new APIs.\n\n - Migrate the only user of the old AES-GCM library API to the new,\n more flexible API; then remove the old API and its implementation.\n\n - Wire up the new APIs to the traditional crypto API by adding\n crypto_skcipher and crypto_aead algorithms.\n\n This makes the new APIs be covered by the traditional crypto API\u0027s\n self-tests. It also makes them be already used for real on systems\n that don\u0027t have architecture-optimized code for these modes.\n\n But most importantly, this is a prerequisite for migrating the\n architecture-optimized code for these AES modes (i.e.\n arch/*/crypto/aes*) into the library, which as usual will eliminate\n a lot of redundant \"glue\" code.\n\n Note that unlike some of the other algorithms that have been migrated\n to the library, e.g. SHA-512, for these AES modes there was too much\n to get done in one cycle. Nor did it make sense to handle these modes\n one at a time, because they tend to be coupled together or depend on\n each other, especially in the architecture-optimized AES code.\n\n Thus, most of the benefits (reductions in lines of code, performance\n improvements, etc.) will follow in later cycles when\n architecture-optimized code is migrated into the library and users of\n crypto_skcipher and crypto_aead are updated to use the new APIs.\n\n The design of the new APIs was informed by writing proof-of-concept\n patches for many kernel subsystems currently accessing these same\n algorithms via crypto_skcipher or crypto_aead (patches 18-33 of\n https://lore.kernel.org/r/20260707053503.209874-1-ebiggers@kernel.org/).\n\n While those patches will be resent for real later, the total diffstat\n for them was negative 1905 lines. So clearly the new APIs are quite a\n bit easier to use and align better with what users actually need.\n\n Besides the new AES encryption APIs, there are also a few changes for\n improved AES-CMAC key and context zeroization\"\n\n* tag \u0027libcrypto-updates-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux:\n mac80211: fils_aead: Use __cleanup() instead of memzero_explicit()\n Bluetooth: SMP: clear the aes_cmac_key when done\n smb: clear the aes_cmac_key and aes_cmac_ctx when done\n lib/crypto: aes-cmac: Add zeroization functions\n lib/crypto: aesgcm: Remove old AES-GCM library\n x86/sev: Remove obsolete virtual address check\n x86/sev: Use new AES-GCM library\n crypto: aes - Add CCM support using library\n crypto: aes - Add GCM support using library\n crypto: aes - Add XTS support using library\n crypto: aes - Add CTR and XCTR support using library\n crypto: aes - Add CBC and CBC-CTS support using library\n crypto: aes - Add ECB support using library\n lib/crypto: aes: Add CCM support\n lib/crypto: aes: Add GCM support\n lib/crypto: aes: Add XTS support\n lib/crypto: aes: Add CTR and XCTR support\n lib/crypto: aes: Add CBC and CBC-CTS support\n lib/crypto: aes: Add ECB support\n crypto: xts - Split out __xts_verify_key() helper\n" }, { "commit": "1d7443e4dca1e8637930f3baf64e2fe82033e669", "tree": "5800e3a841ff00d6f547d9c23f02e9afc5650fda", "parents": [ "63c070cba0c6a586d33730ddbf63065cf28276b7", "7dd38d9dd7a05329825fe2324d4d8e27ad4b3cec" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 19:04:16 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 19:04:16 2026 -0700" }, "message": "Merge tag \u0027fscrypt-for-linus\u0027 of git://git.kernel.org/pub/scm/fs/fscrypt/linux\n\nPull fscrypt updates from Eric Biggers:\n \"The main change this cycle is a significant simplification that\u0027s been\n overdue for a while now: standardizing on a single file contents\n encryption implementation in ext4 and f2fs, instead of having two.\n\n Specifically, the original filesystem-layer file contents encryption\n implementation is removed, and the blk-crypto implementation is now\n used unconditionally. blk-crypto delegates either to inline crypto\n hardware or to the CPU via blk-crypto-fallback. The latter is\n functionally equivalent to the original filesystem-layer code.\n\n The blk-crypto implementation already existed, but previously it was\n used only when the filesystem was mounted with \"-o inlinecrypt\". Now,\n \"-o inlinecrypt\" just selects whether inline crypto hardware is used.\n\n To allow maintaining that user control over hardware use, the\n blk-crypto API is extended with a new flag BLK_CRYPTO_CFG_ALLOW_HW.\n\n Overall, this removes quite a bit of redundant code from ext4, f2fs,\n and fs/crypto/. It should make things easier for ongoing filesystem\n efforts such as iomap support, large folios, and btrfs encryption\n (btrfs had already been planning to use blk-crypto exclusively.)\n\n There are two small behavior changes of note:\n\n - Direct I/O now works on encrypted files even without \"-o inlinecrypt\",\n rather than falling back to buffered I/O. This is effectively a\n bugfix, though I\u0027ll continue to keep an eye out for any user that\n may have been depending on the buffered I/O fallback.\n\n - IV_INO_LBLK_32 policies are no longer supported in certain cases\n that didn\u0027t make sense and have no known uses.\n\n This has been in linux-next since July 22 with no reported issues. All\n encryption xfstests pass on ext4 and f2fs. As usual I\u0027ve also been\n using it on a system with an fscrypt-encrypted home directory. Of\n course, the blk-crypto code paths also aren\u0027t new and were already\n being used on many systems via the inlinecrypt mount option.\n\n In addition to the main change described above, there are a few other\n cleanups such as using lock guards for mutexes, improving\n documentation, and removing a workaround for outdated gcc versions\"\n\n* tag \u0027fscrypt-for-linus\u0027 of git://git.kernel.org/pub/scm/fs/fscrypt/linux: (29 commits)\n blk-crypto: Update docs for blk-crypto-fallback motivation\n blk-crypto: Remove unused function blk_crypto_config_supported()\n fscrypt: Update docs for data path\n fscrypt: Remove unused function fscrypt_finalize_bounce_page()\n f2fs: Update outdated comment in f2fs_write_begin()\n fs: Update outdated comment for SB_INLINECRYPT\n fscrypt: Update encryption policy version docs\n fscrypt: Replace some variable-size memsets with fixed-size\n fscrypt: Add safety checks to non-block-based en/decryption\n fscrypt: Merge bio.c and inline_crypt.c into block.c\n fscrypt: Remove unused functions and workqueue\n fscrypt: Remove fs-layer zeroout code\n fscrypt: Remove fscrypt_dio_supported()\n fscrypt: Replace calls to fscrypt_inode_uses_inline_crypto()\n fs/buffer: Remove fs-layer decryption code\n f2fs: Remove fs-layer file contents en/decryption code\n ext4: Further de-generalize the bio postprocessing code\n ext4: Make ext4_bio_write_folio() return void\n ext4: Remove fs-layer file contents en/decryption code\n Documentation: fscrypt: Update docs for inlinecrypt\n ...\n" }, { "commit": "63c070cba0c6a586d33730ddbf63065cf28276b7", "tree": "456207ef242efa3884158c36d896965ba4cbafe0", "parents": [ "0de672c7e1920acc5675d0dd22718c4f6e6b4447", "24a098541a765ed017274f8094004404a83b0daf" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 18:34:43 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 18:34:43 2026 -0700" }, "message": "Merge tag \u0027nilfs2-v7.3-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/nilfs2\n\nPull nilfs2 updates from Viacheslav Dubeyko:\n \"This contains fixes of syzbot reported issue and various fixes in\n NILFS2 functionality:\n\n - Reject super-root inode sizes whose computed on-disk footprint\n exceeds the filesystem block size (David Lee)\n\n - Replace WARN_ON() in nilfs_cpfile_delete_checkpoints() with\n returning -EIO and reporting a filesystem error via nilfs_error()\n in the case of corrupted checkpoint count on the storage medium\n (Igor Putko)\n\n - Fixed a potential infinite loop in nilfs_clean_segments() reported\n by syzbot (Joshua Crofts)\n\n In nilfs_clean_segments(), if err is non-zero, logic logs the error\n and sleeps but doesn\u0027t abort when it encounters a terminal error\n like -EROFS. This causes the thread to loop forever.\n\n Fix this by breaking out of the loop if nilfs_segctor_construct()\n returns -EROFS.\n\n - Fix small grammar mistake in the description for nilfs2 recovery\n code (Manoj K M)\n\n - Multiple fixes by Ryusuke Konishi:\n - fix the list corruption issue recently detected by syzbot, that\n can occur when out-of-range values are intentionally passed to\n certain GC ioctl parameters\n - fix a flaw in the original B-tree implementation related to\n truncation and resolves the reported out-of-bounds memory\n access issue\n - fix an issue reported by syzbot where a kernel BUG could be\n triggered depending on timing after filesystem corruption is\n detected\n - fix an issue where a WARN_ON check is triggered by sufile\n functions within the log writer after the filesystem degrades\n to read-only mode\n\n - Check for sorted keys when reading btree node blocks into the cache\n (Wang Jianjian)\n\n This prevents unexpected errors during the block number assignment\n phase in log writing caused by key order inconsistencies, as well\n as the kernel warnings reported by syzbot\"\n\n* tag \u0027nilfs2-v7.3-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/nilfs2:\n nilfs2: standardize the inode number type to u64\n nilfs2: enhance btree node keys check\n nilfs2: suppress false positive WARN_ONs for sufile after an FS error\n nilfs2: fix BUG in nilfs_copy_dirty_pages() on dirty state mismatch\n nilfs2: prevent out-of-bounds read in super root block parsing\n nilfs2: fix infinite loop in nilfs_clean_segments()\n nilfs2: fix slab-out-of-bounds in nilfs_direct_propagate after truncation\n Documentation: fix grammar in description of nilfs2 recovery code\n nilfs2: handle corrupted checkpoint count gracefully during deletion\n nilfs2: reject invalid block index in GC ioctl\n" }, { "commit": "0de672c7e1920acc5675d0dd22718c4f6e6b4447", "tree": "3785783eb0647633fae8a742d53a02d1b62bd85e", "parents": [ "4bb187d6f379c5490e3034b153fe476127308519", "627b7865c062ff642c0000f3a4775f1a44b986a4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 18:14:32 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 18:14:32 2026 -0700" }, "message": "Merge tag \u0027hfs-v7.3-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/hfs\n\nPull HFS updates from Viacheslav Dubeyko:\n \"This contains several fixes in HFS/HFS+ of syzbot reported issues and\n HFS/HFS+ fixes of xfstests failures.\n\n - b-tree bitmap corruption check (Aditya Prakash Srivastava)\n\n During b-tree open (hfs_btree_open()), the code verifies that the\n allocation map bit for the tree header (node 0) is set. If not, it\n indicates a corrupted map record/bitmap and mounts the volume as\n read-only (SB_RDONLY) to prevent further damage.\n\n - Validate catalog CNIDs before instantiating inodes (David\n Maximiliano Hermitte)\n\n The hfs_cat_find_brec() first resolves a catalog thread record by\n CNID and then looks up the corresponding catalog record by\n parent/name. On a corrupted filesystem image, the second lookup may\n find a record whose CNID does not match the CNID that was\n requested. Finally, corrupted catalog records are rejected.\n\n - Validate B-tree record offset table (Jiaming Zhang)\n\n A crafted HFS+ image can contain a corrupted B-tree node. The node\n descriptor may contain a record count that does not fit in the\n node, and record offsets may be unordered, unaligned, outside the\n node, or point into the offset table itself. Validate num_recs\n against the node size before walking the record offset table.\n Reject record ranges that are unordered, unaligned, outside the\n node, or overlapping the offset table. Reject invalid record\n indexes before reading their offset entries, and avoid decrementing\n an already-zero leaf_count.\n\n - Refactoring of hfsplus_delete_cat() logic (Kyle Zeng).\n\n The hfsplus_delete_cat() is called with str \u003d\u003d NULL when the last\n open reference to an unlinked HFS+ hardlink backing inode is\n closed. In that case, the function finds the catalog thread by CNID\n and rebuilds the catalog key from thread.nodeName. A corrupted\n image can therefore provide an oversized thread name length and\n make hfs_bnode_read() write past the catalog search-key allocation.\n Read the CNID record through hfsplus_brec_read_cat(), which bounds\n the record read to sizeof(hfsplus_cat_entry) and verifies that a\n thread record\u0027s size exactly matches nodeName.length.\n\n - Cleanup in KUnit test (Mohammad Shahid)\n\n The kfree() safely handles NULL pointers, so the explicit NULL\n check in free_mock_str_env() before calling kfree() is unnecessary.\n\n The rest contain fixes of generic/564 xfstests\u0027 test-case failure\n for the case of HFS+ file system, syzbot reported issue in\n hfs_mdb_commit() and hfs_mdb_close() methods of HFS file system,\n and reworking the MDB locking scheme in HFS file system\"\n\n* tag \u0027hfs-v7.3-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/hfs:\n hfsplus: validate extent record length before writing it back\n hfsplus: validate B-tree record offset table\n hfs: rework MDB locking scheme\n fs: hfsplus: remove redundant NULL check before kfree()\n hfs: port HFS+ b-tree bitmap corruption check\n hfs: don\u0027t re-dirty MDB buffers after a write failure\n hfsplus: fix error code when writing beyond volume capacity\n hfs: fix error code when writing beyond volume capacity\n hfsplus: validate thread record before delete key rebuild\n hfs: validate catalog CNIDs before instantiating inodes\n" }, { "commit": "4bb187d6f379c5490e3034b153fe476127308519", "tree": "0c3091e0f3291a69a9a8261dd00821a5f3a7a7f8", "parents": [ "cf07e82984895a06a7cbfadee1b13d83805bb41b", "657e5af4fe0b9696a6c7232da1727a9f37eaa04d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 17:03:34 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 17:03:34 2026 -0700" }, "message": "Merge tag \u0027gfs2-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2\n\nPull gfs2 updates from Andreas Gruenbacher:\n\n - Don\u0027t cache unreferenced glocks: when a glock is no longer referenced\n (for example, because the inode it protects is evicted), it is now\n released as soon as possible instead of leaving it around until\n memory pressure or an unmount forces it out.\n\n For some workloads, this saves a lot of memory and speeds up unmounts\n significantly.\n\n - Harden gfs2_glock_hold() by making sure the caller holds a reference\n and fix a related race in checking for the liveliness of glocks\n between gdlm_bast() and gfs2_glock_cb().\n\n* tag \u0027gfs2-for-7.3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2:\n gfs2: harden gfs2_glock_hold\n gfs2: Remove the glock lru list and shrinker\n gfs2: Skip dlm unlocks earlier\n gfs2: Don\u0027t cache unreferenced glocks\n gfs2: Enable automatic glock hash table shrinking\n" }, { "commit": "cf07e82984895a06a7cbfadee1b13d83805bb41b", "tree": "3d8406033a3e3c9eee42cbc9093d753bcc56dde9", "parents": [ "ff68e5f557f69a08fdcfa4ce8b1b809d63bd4f45", "b7eea80be25f3334f131d52982b3131aba77b97d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 16:41:57 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 16:41:57 2026 -0700" }, "message": "Merge tag \u0027xfs-merge-7.3\u0027 of git://git.kernel.org:/pub/scm/fs/xfs/xfs-linux\n\nPull xfs updates from Carlos Maiolino:\n \"There are no big standing out features on this window, so this\n mostly consists on bug fixes and code refactoring.\n\n The only user visible change that stands out is the support for\n FALLOC_FL_WRITE_ZEROES added to this\"\n\n* tag \u0027xfs-merge-7.3\u0027 of git://git.kernel.org:/pub/scm/fs/xfs/xfs-linux: (23 commits)\n xfs: validate attr entry pointer before field access\n xfs: check split_sectors validity before bio_split call\n xfs: use file target for post-log fsync fallback flush\n xfs: restore nofs context unconditionally in xfs_trans_roll\n xfs: add lockless xfs_buf_readahead_map fast path\n xfs: move buffer locking out of xfs_find_get_buf\n xfs: merge xfs_buf_reverify into xfs_buf_read_map\n xfs: use goto based error unwinding in xfs_buf_read_map\n xfs: don\u0027t reverify buffers in xfs_buf_readahead_map\n xfs: use WRITE_ONCE to update b_flags\n xfs: hide b_flags manipulation from code outside of xfs_buf.c\n xfs: remove _XBF_LOGRECOVERY\n xfs: remove spurious XBF_DONE clearing on readahead validation failure\n xfs: split out a lower-level xfs_buf_get_map helper from xfs_find_get_buf\n xfs: consolidate buffer locking in xfs_buf_get_map\n xfs: don\u0027t get a pag reference in xfs_buf_get_map\n xfs: use kmalloc_objs() instead of kmalloc() in xfs_da_grow_inode_int\n xfs: mark internal metadir file creation helpers static\n xfs: create rtgroup metadir inodes using xfs_metadir_create_file\n xfs: create quota metadir inodes using xfs_metadir_create_file\n ...\n" }, { "commit": "ff68e5f557f69a08fdcfa4ce8b1b809d63bd4f45", "tree": "a0f66fa4118692cdc7bc58cc42f040b3d7f38421", "parents": [ "1781f0b3d75caa22376cf7fa0224f9aca696580d", "974d0be0cb8e48d63b9d413a2e1a8fba16cd2583" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 14:21:00 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 14:21:00 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.sync\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull vfs writeback updates from Christian Brauner:\n \"This makes sync_inode_metadata() and writeback_single_inode() persist\n not only the inode but all metadata associated with it.\n\n A new .sync_inode_metadata superblock operation is called from\n __writeback_single_inode(). Alongside it a new I_METADATA_WRITEBACK\n state flag is added.\n\n Filesystems no longer need their own mmb_fsync() implementations and\n can just use simple_fsync(). All metadata is now written for IS_SYNC\n and IS_DIRSYNC inodes. Races where several fsyncs raced and mmb_sync()\n could return before all buffers were really persisted are fixed since\n I_SYNC now serializes properly.\n\n The I_METADATA_WRITEBACK scheme also fixes the case where a\n WB_SYNC_NONE writeback landing between write(2) and fsync(2) left\n fsync(2) failing to persist the inode. That problem is not specific to\n filesystems using the generic metadata bh tracking, and the ones that\n do not are left alone.\n\n ext2, udf, bfs, minix, fat and ext4 in nojournal mode have their data\n integrity writeout fixed and are converted. affs drops metadata bh\n tracking and mmb_fsync() is removed.\n\n A few other fixes came out of this:\n\n - a UAF in mark_buffer_write_io_error()\n\n - missed inode writeback when racing with __writeback_single_inode()\n\n - ext4 allocating the mapping_metadata_bhs struct on demand\n\n - three fat fixes: a lost inode update in do_msdos_rename() with\n DIRSYNC, inode buffer write errors not propagating out of\n fat_sync_inode_metadata() and directory entries not being\n persisted on fsync(2) of the root directory\"\n\n* tag \u0027vfs-7.3-rc1.sync\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (24 commits)\n writeback: Export __inode_attach_wb()\n fat: Fix persisting directory entries on fsync(2) of the root directory\n fat: Propagate inode buffer write errors from fat_sync_inode_metadata()\n fat: Fix lost inode update in do_msdos_rename() with DIRSYNC\n vfs: Remove mmb_fsync()\n fat: Replace fat_sync_inode() with sync_inode_metadata()\n fat: Fix missed inode writeback during fsync(2)\n ext4: Fix data integrity writeout issues in nojournal mode\n minix: Fix data integrity writeout issues\n bfs: Fix data integrity writeout issues\n udf: Fold udf_update_inode() into udf_write_inode()\n udf: Use sync_inode_metadata() in udf_evict_inode()\n udf: Drop udf_sync_inode()\n udf: Use sync_inode_metadata() to writeout IS_SYNC inode\n udf: Fix data integrity writeout issues\n ext2: Fix data integrity writeout issues\n ext2: Avoid unnecessary inode buffer writeback for sync(2)\n ext2: Drop __ext2_write_inode()\n ext2: Fix lost inode updates for IS_SYNC inodes\n fs: Provide way for filesystem to wait for metadata writeback\n ...\n" }, { "commit": "1781f0b3d75caa22376cf7fa0224f9aca696580d", "tree": "042a9acf224b96254f38829630c05ee7dd7b2970", "parents": [ "aaed66fadba2d2de8fe0daa0aa3eac827d2076b9", "9ac8fd831252e52aa78399eaccc11a72f6c2af1a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 13:57:04 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 13:57:04 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.super\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull vfs superblock updates from Christian Brauner:\n\n - Make it possible to share a block device between multiple\n filesystems.\n\n erofs can mount read-only blob devices shared between many\n superblocks, but because we only tracked a single superblock a\n freeze, thaw, removal or sync on such a device was never propagated\n to all the superblocks using it, and there was no way to find them.\n\n Add an efficient table to lookup all superblocks using a given block\n device.\n\n - A bunch of pre-existing fixes fell out of this work:\n\n A block-device freeze racing a btrfs device change could leave the\n whole filesystem stuck frozen. A bdev_freeze() issued by \"dmsetup\n suspend\" or an LVM snapshot resolves that holder to freeze the\n filesystem. and bdev_thaw() resolves it again to thaw. A freeze\n landing while btrfs is adding, removing or replacing a device freezes\n the filesystem. The membership change then drops that link. So the\n matching thaw could no longer find the superblock.\n\n Forbid freezing a device for the duration of a membership change,\n modelled on deny_write_access()/allow_write_access().\n\n* tag \u0027vfs-7.3-rc1.super\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (24 commits)\n super: fix dying superblock warning messages\n block: reject block device inodes with i_rdev \u003d\u003d 0 in lookup_bdev()\n selftests/filesystems: add ustat() coverage\n fs: look up the superblock via the device table in user_get_super()\n super: make fs_holder_ops private\n f2fs: open via dedicated fs bdev helpers\n erofs: open via dedicated fs bdev helpers\n fs: tolerate per-superblock freeze errors on shared devices\n fs: look up superblocks via the device table in fs_holder_ops\n ext4: open via dedicated fs bdev helpers\n btrfs: open via dedicated fs bdev helpers\n xfs: port to fs_bdev_file_open_by_path()\n fs: add dedicated block device open helpers for filesystems\n fs: maintain a global device-to-superblock table\n ocfs2: don\u0027t reset s_dev on dismount\n ext4: use anonymous devices for KUnit test superblocks\n fs, block: move blk_mode_t and fop_flags_t into \u003clinux/types.h\u003e\n super: take lock after last reference count\n super: convert s_count to refcount_t s_passive\n btrfs: deny freezing devices undergoing a replace\n ...\n" }, { "commit": "aaed66fadba2d2de8fe0daa0aa3eac827d2076b9", "tree": "13bf879302b369bc58162772b1b27ff043804dad", "parents": [ "55668d04e3a821281f885e59ac951a22525b78e3", "077ab8985ee278c3d8618182d335b0f0cd919e16" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 13:38:36 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 13:38:36 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.ovl\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull overlayfs updates from Christian Brauner:\n \"This lets the merged overlayfs mount itself be idmapped through\n mount_setattr(MOUNT_ATTR_IDMAP), in addition to the already supported\n idmapped lower and upper layers. The same overlay tree can then be\n exposed under a different ownership view.\n\n Overlayfs already normalizes every underlying id through the relevant\n layer idmap when ovl_copyattr() copies attributes into the overlay\n inode. So the overlay inode\u0027s i_uid and i_gid are overlay-final ids.\n\n The overlay mount idmap composes on top of that and is applied at the\n overlay-inode boundary only while the underlying layers keep being\n accessed with the mounter\u0027s credentials through their own (possibly\n idmapped) mounts.\n\n So this only changes how the caller sees the overlay inode and never\n widens the mounter\u0027s access to the layers. The second,\n mounter-credential check in ovl_permission() against the real inode\n stays on the layer idmap.\n\n Most paths need no change because the VFS applies the mount idmap to\n the overlay inode before overlayfs runs or after it returns at the\n syscall boundary. Overlayfs only has to change where it bypasses the\n generic path.\n\n This also included is a fix for a double end_creating() on the\n overlayfs casefold-mismatch path\"\n\n* tag \u0027vfs-7.3-rc1.ovl\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n ovl: fix double end_creating() on the casefold-mismatch path\n ovl: document security.capability idmapping on the xattr forward paths\n selftests/filesystems/overlayfs: test idmapped overlay mounts\n selftests/filesystems/overlayfs: fix set_layers_via_fds link error\n docs: document idmapped overlay mounts\n ovl: allow idmapping overlay mounts\n ovl: handle idmapped mounts in ovl_set_acl()\n ovl: handle idmapped mounts in ovl_getattr()\n ovl: handle idmapped mounts in ovl_setattr()\n ovl: handle idmapped mounts in ovl_permission()\n ovl: handle idmapped mounts in ovl_create_object() and ovl_tmpfile()\n" }, { "commit": "55668d04e3a821281f885e59ac951a22525b78e3", "tree": "686b81d64350cb496821476d4aa0b73d125e72cb", "parents": [ "c3d6d6dde3a3e11b421839497793da71ad6faa44", "949eb95d5bcc530e6ce74cc95897c1849c53ea89" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 13:32:50 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 13:32:50 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.netfs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull cachefiles ondemand removal from Christian Brauner:\n \"This sunsets cachefiles ondemand mode.\n\n It was an effort to make fscache usable as a kernel cache for lazy\n pulling. EROFS over fscache was its only in-tree user. fscache has\n since become netfslib-oriented while EROFS never acts as a network\n filesystem and EROFS over fscache has been removed.\n\n So this cleans up the netfs, fscache and cachefiles side as well\"\n\n* tag \u0027vfs-7.3-rc1.netfs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n cachefiles,netfs: sunset ondemand mode\n" }, { "commit": "c3d6d6dde3a3e11b421839497793da71ad6faa44", "tree": "0f7777c5204b74fee33e101adf1ef796ce2104c7", "parents": [ "1c3e8cef79ea5f1415cff0d3c507e2e07b71ade8", "083e8742e301a24f0c458696b45b481345888b85" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 13:26:40 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 13:26:40 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.mount\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull mount updates from Christian Brauner:\n\n - Make the legacy mount API notify pollers of propagation changes.\n\n Changing propagation via change_mnt_propagation() or with\n MOVE_MOUNT_SET_GROUP update the propagation relationship of the\n target mount. But unlike mount_setattr() neither path touched the\n affected mount namespace. So pollers of /proc/\u003cpid\u003e/mountinfo were\n never woken.\n\n - Also remove a redundant panic() in mnt_init()\n\n* tag \u0027vfs-7.3-rc1.mount\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n mount: remove redundant panic() in mnt_init()\n fs/namespace: notify pollers of legacy propagation changes\n" }, { "commit": "1c3e8cef79ea5f1415cff0d3c507e2e07b71ade8", "tree": "250813712d5f13fe5cb38a9923d8a70ee0043aab", "parents": [ "ab5ed08f2d8396fb8e3942569bbbd5cd569a753e", "dcacab904fe78d60840ba947a104993ee9ded887" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 12:56:12 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 12:56:12 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull misc vfs updates from Christian Brauner:\n \"Bigger cleanups:\n\n - The lockref dead-count handling is tidied up.\n\n The open-coded check for a count below zero as the dead marker\n relies on information the caller should not have.\n\n - make put_mnt_ns() leave mounts connected. Destroying a mount\n namespace disconnected its mounts from their mount points. So a\n file descriptor still open on the parent of a mount point could be\n used to peek under it.\n\n Locked mounts were already kept connected to prevent exactly that.\n But a mount is only locked when its tree is copied across a user\n namespace boundary. So a mount namespace set up by a privileged\n component had no locked mounts and its mounts were disconnected.\n Passing UMOUNT_CONNECTED keeps every mount connected and prevents\n that bug.\n\n - vfs_prepare_mode() passes S_IFDIR for directories. I meant to fix\n that ago but didn\u0027t get to it. So now someone finally did it.\n\n This kills the exception where the mode could be 0 when a directory\n was created whereas every other creation operation passed it\n explicitly already.\n\n - move long delayed work for ufs, jffs2, hfsplus, hfs and affs from\n the per-cpu system_long_wq to the new unbound system_dfl_long_wq.\n\n None of that work relies on per-cpu state and the work item is\n enqueued with queue_delayed_work() whose timer is global anyway. So\n it may as well benefit from scheduler task placement.\n\n Smaller fixes and cleanups:\n\n - unlock_buffer() and journal_end_buffer_io_sync() use\n clear_and_wake_up_bit()\n\n - the pipe page pools are unified into a single per-pipe pool and the\n extra wake_up(rd_wait) is limited to EPOLLET consumers\n\n - eventpoll now computes its timer slack lazily in ep_poll()\n\n - shrink_dcache_for_umount() keeps making progress on busy roots\n\n - excess xarray nodes are freed in clear_inode()\n\n - romfs detects hard link cycles\n\n - the user path of nested backing files is fixed\n\n - pidfd holds exec_update_lock around the namespace ioctl\n\n - non-memcg-aware nr_cached_objects is skipped during memcg slab\n shrink\n\n - iomap_write_iter() always returns status\n\n - mangle_path() is renamed to seq_mangle_path()\n\n - inode timestamp accessors are annotated\n\n - new regression test for pipe-\u003epoll_usage.\n\n - a few documentation, kernel-doc and selftest fixes\"\n\n* tag \u0027vfs-7.3-rc1.misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (67 commits)\n selftests/namespaces: Fix racy pipe handshake in timens and pidns_separate\n selftests/epoll: add a regression test for pipe-\u003epoll_usage\n pipe: only enable the extra wake_up(rd_wait) for EPOLLET consumers\n pidfd: hold exec_update_lock around namespace ioctl\n fs: fix user path of nested backing files\n fs: remove stale inode_insert5() kernel-doc parameter\n fs: fix switch/case indentation in sysfs() syscall\n fs: document semantics of kstat::{uid,gid} fields\n dcache: keep shrink_dcache_for_umount() making progress on busy roots\n seq_file: rename mangle_path to seq_mangle_path\n nstree: add/fix struct ns_id_req kernel-doc member fields\n dcache: use lockref routines for dead count checks\n lockref: tidy up dead count handling\n initramfs: fix typo in reserve_initrd_mem comment\n fs/pipe: unify the page pools into a single per-pipe pool\n fs: annotate inode timestamp accessors\n eventpoll: compute timer slack lazily in ep_poll()\n selftests/filesystems: add mntns cleanup test\n put_mnt_ns(): leave mounts connected\n affs: Move long delayed work on system_dfl_long_wq\n ...\n" }, { "commit": "ab5ed08f2d8396fb8e3942569bbbd5cd569a753e", "tree": "36323a501886e146e044e0a4c8cb25523abffb42", "parents": [ "fff0150b0299f834fe2c335ce0eb5c68bb414cbd", "895741772f6ca989da565bcbe4673f12d3d0ac31" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 12:03:34 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 12:03:34 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.lookup\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull vfs lookup updates from Christian Brauner:\n \"This refactors lookup_open() and adds vfs_lookup_open() for nfsd.\n\n mnt_want_write() and parent locking are moved into lookup_open()\n itself.\n\n audit_inode_child() is also now called in lookup_open() on failure.\n That is the calling convention in vfs_create() and vfs_mkdir(), but\n lookup_open() made no such call when atomic_open() should have created\n a file and did not. And neither did the regular -\u003ecreate() path fwiw.\n\n This also contains work to remove the unneeded excl argument from the\n -\u003ecreate() inode op\"\n\n* tag \u0027vfs-7.3-rc1.lookup\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n fs/namei.c: fix coding style in atomic_open() and lookup_open()\n fs/namei.c: fix kerneldoc of atomic_open() and vfs_lookup_open()\n fs/namei.c: update stale comments in lookup_open()\n Remove excl arg to -\u003ecreate inode_operation\n fs/namei.c: update kerneldoc of atomic_open()\n vfs: call audit_inode_child() in lookup_open() on failure\n vfs: move create error \u0026\u0026 negative dentry case in lookup_open() up\n VFS: add vfs_lookup_open() for nfsd\n VFS: move delegated_inode retry loop into lookup_open()\n VFS: move mnt_want_write() and locking into lookup_open()\n" }, { "commit": "fff0150b0299f834fe2c335ce0eb5c68bb414cbd", "tree": "87b252e40d51b3d033926fee45cea2220fcb87f7", "parents": [ "de03b17ec046d81c4713bc54306579b9024fc513", "b4343aebd3a4dd255b15b2e5b1363d6399da302f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 11:35:12 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 11:35:12 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.kthread\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull kthread vfs updates from Christian Brauner:\n \"This stops kernel threads from sharing filesystem state with\n userspace. This work is about 3 cycles old and has been in -next\n for about that time.\n\n When the kernel boots init_task creates PID 1 and then kthreadd. From\n that point every kthread and PID 1 share the same fs_struct. That is\n why pivot_root() has to rewrite the fs_struct of all kthreads. The\n rewriting exists so that kthreads can use init\u0027s filesystem state when\n they want to. It also means userspace can move the ground out from\n under the kernel.\n\n PID 1 now gets a completely separate fs_struct. All kthreads are\n anchored in a private SB_KERNMOUNT instance of nullfs that cannot be\n mounted on and cannot be used to follow other mounts. Userspace init\n can no longer affect kthread filesystem state and kthreads can no\n longer affect userspace fs state without explicit opting in to that.\n\n Path lookup from a kthread now fails by default. It makes it\n deliberately hard to offload security sensitive operations into init\u0027s\n filesystem state from a kthread.\n\n Places that legitimately need to look something up there opt in\n through the new scoped_with_init_fs() which temporarily overrides the\n caller\u0027s fs_struct with init\u0027s. usermodehelpers remain the only kernel\n tasks that genuinely share init\u0027s filesystem state, since they execute\n random binaries in the root filesystem (excellent...).\n\n The visible result is that /proc/2/root is a nullfs with an empty\n mountinfo while /proc/1/root is the real root\"\n\n* tag \u0027vfs-7.3-rc1.kthread\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (26 commits)\n initramfs_test: use test init/exit hooks to override init fs\n fs: stop rewriting paths for PF_EXITING | PF_DUMPCORE\n fs: stop rewriting kthread fs structs\n fs: start all kthreads in nullfs\n nullfs: make nullfs multi-instance\n devtmpfs: create private mount namespace\n fs: add umh argument to struct kernel_clone_args\n fs: stop sharing fs_struct between init_task and pid 1\n af_unix: use scoped_with_init_fs() for coredump socket lookup\n initramfs: use scoped_with_init_fs() for rootfs unpacking\n pnfs/blocklayout: use scoped_with_init_fs() for SCSI device lookup\n ksmbd: use scoped_with_init_fs() for VFS path operations\n ksmbd: use scoped_with_init_fs() for filesystem info path lookup\n ksmbd: use scoped_with_init_fs() for share path resolution\n fs: use scoped_with_init_fs() for kernel_read_file_from_path_initns()\n coredump: use scoped_with_init_fs() for coredump path resolution\n btrfs: use scoped_with_init_fs() for update_dev_time()\n scsi: target: use scoped_with_init_fs() for APTPL metadata\n scsi: target: use scoped_with_init_fs() for ALUA metadata\n crypto: ccp: use scoped_with_init_fs() for SEV file access\n ...\n" }, { "commit": "de03b17ec046d81c4713bc54306579b9024fc513", "tree": "1a15783aa1f0662c6563719f2394eecf2ff6a688", "parents": [ "9ea8d6197d9f43a15ccb9c0dce601ec535d5da7e", "8abc1c8badc11c859bb4ba43211611347e59c55c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 11:02:10 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 11:02:10 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.kfunc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull vfs bpf access updates from Christian Brauner:\n \"This adds a bpf_sock_read_xattr() kfunc so a BPF LSM program can read\n a user.* extended attribute from a socket\u0027s sockfs inode locklessly.\n\n userspace already uses user.* xattrs on sockets to implement socket\n rate limiting and to tag sockets for other purposes such as a varlink\n registry. There has been no efficient way for a BPF program to read\n those labels back. With this a listening socket marked from userspace\n with fsetxattr() can be read back during bind or connect and acted\n upon on the connecting socket. That lets userspace mark sockets and\n later rediscover them or implement policy on them\"\n\n* tag \u0027vfs-7.3-rc1.kfunc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n selftests/bpf: Add test for bpf_sock_read_xattr() kfunc\n fs: Add bpf_sock_read_xattr() kfunc to read socket xattrs\n" }, { "commit": "9ea8d6197d9f43a15ccb9c0dce601ec535d5da7e", "tree": "dd135e062b99a2b84130acb3de31a5d1d79ed19c", "parents": [ "3d1f95267769caf5c3eb71a2c9586213ea0a3ef2", "9948bc9aa1c2e5e77ab989a8e5d5eec829e967d5" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 10:03:55 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 10:03:55 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.iomap\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull iomap updates from Christian Brauner:\n \"The bulk of this is the conversion of iomap to a single -\u003eiomap_next()\n callback and thus finishing the move to an iterator model.\n\n Every iomap operation drove its iteration through a struct iomap_ops\n holding -\u003eiomap_begin() and -\u003eiomap_end(). iomap_iter() only ever sees\n those as pointers. That means every step of every iteration is an\n indirect call.\n\n This collapses both into one -\u003eiomap_next() callback that finishes the\n previous mapping and produces the next one. This lets callers inline\n the iteration loop and pass its -\u003eiomap_next() as a compile time\n constant. That means the compiler can turn it into a direct and hence\n inlineable call.\n\n This also allows future callers to express custom logic to drive the\n iteration forward better. xfs, btrfs, ext4, ext2, erofs, f2fs, gfs2,\n hpfs, fuse, exfat, zonefs, ntfs, ntfs3 and the block device mapping\n are all converted. No functional changes are intended.\n\n This also adds a simple direct I/O path for small reads. On Gen5 NVMe\n the __iomap_dio_rw() dominates 4K random reads. The same single-core\n io_uring poll mode workload reaches ~3.2M IOPS against the raw block\n device but only ~1.92M through ext4 or XFS.\n\n __iomap_dio_rw(), iomap_iter(), iomap_dio_bio_iter() and kfree() were\n at the top of the profile. The new path is very lightweight if no\n special behavior is requested. The bio comes from a dedicated bioset\n and laid out so the whole request is a single cacheline aligned\n allocation. Completion runs inline.\n\n That takes ext4 from 1.92M to 2.19M IOPS in the original workload. fio\n shows around:\n - 4% at libaio queue depths of 64 and up\n - around 5% for io_uring\n - up to 10% for io_uring poll mode at depth 256\n on both ext4 and xfs.\n\n A few other patches:\n\n - iomap_folio_mark_uptodate() lets a filesystem that writes into the\n page cache outside the iomap read and write paths keep iomap\u0027s\n internal uptodate bitmap in sync, which fuse needs for\n server-pushed notify stores before it can enable large folios;\n\n - two fixes for iomap_bio_read_folio_range_sync(): a potential crash\n when device integrity behavior is changed and a missing\n bio_uninit().\n\n - a folio batch release fix on iomap callback failures\n\n - FGP_NOFS is dropped from iomap_get_folio()\n\n - documentation fix\"\n\n* tag \u0027vfs-7.3-rc1.iomap\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (29 commits)\n iomap: iomap_bio_read_folio_range_sync is missing a call to bio_uninit\n iomap: don\u0027t free integrity payload that doesn\u0027t exist\n docs: fix grammatical error in iomap docs\n exfat: convert iomap ops to -\u003eiomap_next()\n fuse: convert iomap ops to -\u003eiomap_next()\n hpfs: convert iomap ops to -\u003eiomap_next()\n gfs2: convert iomap ops to -\u003eiomap_next()\n f2fs: convert iomap ops to -\u003eiomap_next()\n block: convert iomap ops to -\u003eiomap_next()\n ext2: convert iomap ops to -\u003eiomap_next()\n zonefs: convert iomap ops to -\u003eiomap_next()\n erofs: convert iomap ops to -\u003eiomap_next()\n ext4: convert iomap ops to -\u003eiomap_next()\n ntfs: convert iomap ops to -\u003eiomap_next()\n ntfs3: convert iomap ops to -\u003eiomap_next()\n btrfs: convert iomap ops to -\u003eiomap_next()\n xfs: convert iomap ops to -\u003eiomap_next()\n iomap: add -\u003eiomap_next()\n iomap: use GFP_NOWAIT when application for iomap_dio_simple allocations\n iomap: decouple simple direct I/O reads from iomap_dio_rw\n ...\n" }, { "commit": "3d1f95267769caf5c3eb71a2c9586213ea0a3ef2", "tree": "dc0c9f9454152051dfb5daffe9d7010ed2646b4f", "parents": [ "cd051cfe1e35a471fc2cdf6d32fae6ee23305ecb", "9d7ed813ee5ff0d469bd99630828ed6fdef4e8da" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 09:30:42 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 09:30:42 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.fat\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull FAT update from Christian Brauner:\n \"This rejects names longer than NAME_MAX in msdos_format_name().\n\n The VFS only enforces PATH_MAX rather than the length of an individual\n component. open() on such a path component reported success for a name\n far longer than NAME_MAX\"\n\n* tag \u0027vfs-7.3-rc1.fat\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n fat: reject name longer than NAME_MAX in msdos_format_name()\n" }, { "commit": "cd051cfe1e35a471fc2cdf6d32fae6ee23305ecb", "tree": "b54ff317457566bb10bd8e2e8345a440907ef7da", "parents": [ "d31a688a49a82a48e0434aee9d851e094270a6ee", "1d38e750a389e919c1608dbc61cd5c93cd4915dc" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 09:15:52 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 09:15:52 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.failfs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull failfs filesystem from Christian Brauner:\n \"Add failfs and expose a FD_FAILFS_ROOT sentinel.\n\n This allows userspace to shed their filesystem state completely. A\n process with its root or working directory in failfs must anchor every\n path lookup at an explicit file descriptor. Absolute paths, absolute\n symlinks and AT_FDCWD-relative lookups simply fail.\n\n Failfs is the counterpart to nullfs. nullfs says adds a permanently\n empty, immutable directory whose lookups fail with ENOENT but which\n can be opened, read, stat\u0027d and mounted upon. Failfs on the other hand\n fails every operation. The root cannot be opened at all. A single\n instance is mounted during early boot via kern_mount(), which makes it\n logically distinct from every mount namespace.\n\n This is accompanied by a new fchroot() system call which makes\n chrooting via a file descriptor a first class concept. It\u0027s possible\n to chroot into failfs as an unprivileged user provided the task has no\n new privileges set\"\n\n* tag \u0027vfs-7.3-rc1.failfs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n Documentation: add failfs documentation\n selftests/filesystems: add failfs selftests\n arch: hookup fchroot() system call\n fs: support FD_FAILFS_ROOT in fchroot()\n fs: add fchroot()\n fs: support FD_FAILFS_ROOT in fchdir()\n fs: add failfs\n" }, { "commit": "d31a688a49a82a48e0434aee9d851e094270a6ee", "tree": "4db4f4f757807f639127fbcbe9e830f7724ddd63", "parents": [ "b9cba7ebfe539f3e4bbdd03a1e0efa3b30b3f592", "bf2fbbca5d49d07917236604837145ee9e662c41", "a3ed4fcd04b95347ad04cf4409b394bb7fbd0c02" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 08:55:58 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 08:55:58 2026 -0700" }, "message": "Merge tags \u0027vfs-7.3-rc1.efs\u0027 and \u0027vfs-7.3-rc1.freevxfs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull efs and freevxfs removal from Christian Brauner:\n \"This removes the EFS and freevxfs filesystems:\n\n - EFS was the read-only on-disk format SGI used on IRIX before XFS\n\n - freevxfs provided compatibility with various old-school Unix\n systems from the 1990s and was fun 25 years ago. Today it mostly\n serves as fodder for automated bug checkers. There has been only\n one known user and contributor in the last 15 years\"\n\n* tag \u0027vfs-7.3-rc1.efs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n efs: Remove EFS\n\n* tag \u0027vfs-7.3-rc1.freevxfs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n freevxfs: remove the driver\n" }, { "commit": "b9cba7ebfe539f3e4bbdd03a1e0efa3b30b3f592", "tree": "aff2220d3b5e523e5c0c08c5e67eb471dc5f2f26", "parents": [ "043d7a2b4045df5eebc699c39bc3a571afb9f1ae", "68aabd01ddd26ced458a9e5716a640eaf8e4b7a6" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 08:35:25 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 08:35:25 2026 -0700" }, "message": "Merge tag \u0027vfs-7.3-rc1.binfmt\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull binfmt updates from Christian Brauner:\n \"This contains a bunch of work for binfmt_misc. It fixes a bunch of\n old bugs, reworks the locking, and then extends the format registry\n so a binary type can be matched programmatically and its interpreter\n computed per exec instead of being a fixed string recorded at\n registration time.\n\n This allows nixos and other to e.g., implement relocatable binaries\n meaning the interpreter/dynamic loader can be determined\n programatically, say found relative to the binary. The mechanism is\n flexible and can support other policies:\n\n - Handler lookup is now an rcu walk. An exec that matches no\n binfmt_misc entry should now never write to a shared cacheline\n\n - remove the VERBOSE_STATUS and USE_DEBUG compile time toggles\n\n - convert the entry file to a seq_file which simplifies things quite\n a bit and kills a lot of custom logic\n\n - make flags proper enums\n\n - rename struct Node to binfmt_misc_entry\n\n - allow entries to be removed with unlink(2)\n\n - Add the ability to attach bpf programs to binfmt_misc entries so\n it\u0027s possible to dynamically choose the execution environment such\n as the loader or interpreter on a per binary basis.\n\n A handler is an instance of a binfmt_misc_ops struct_ops with a\n -\u003ematch() and a -\u003eload() program. match() decides from the entry\n lookup walk whether the handler applies under the same\n registration-order. It can read file content as needed not only the\n prefetched 256 bytes in bprm-\u003ebuf.\n\n load() then selects the interpreter and stages it through the new\n bpf_binprm_set_interp(), bpf_binprm_set_interp_arg() and\n bpf_binprm_set_flags() kfuncs.\n\n Handlers are published in a registry keyed by the registering\n task\u0027s user namespace and activated through the existing text\n interface with a new \u0027B\u0027 type carrying the handler name:\n\n\techo \u0027:origin:B::::nix:\u0027 \u003e /proc/sys/fs/binfmt_misc/register\n\n The permission and namespacing model is unchanged. Activating a\n handler requires the same write access to an instance as any other\n registration. A container mounting its own instance escapes the\n host\u0027s entries exactly as before. The computed interpreter is\n opened with open_exec() under the caller\u0027s credentials and goes\n through full LSM vetting as the next binprm level. A program can\n only ever redirect the caller to something the caller could exec\n anyway.\n\n - Two dispatch modes are added. So far the chosen interpreter owns\n the whole process identity (argv[0], /proc/pid/cmdline,\n /proc/self/exe all name interpreter information). So relocatable\n find the dynamic linker instead. Also a binary passed to execveat()\n as an inaccessible O_CLOEXEC fd cannot run at all and gdb trips\n because AT_ENTRY and AT_PHDR do not match the exe file. So PIE\n symbols are unrelocated.\n\n This adds transparent dispatch which allows the interpreter to load\n the binary through AT_EXECFD and leaves the argument vector exactly\n as the caller built it and labels mm-\u003eexe_file and comm with the\n binary. It also raises the AT_FLAGS_TRANSPARENT_INTERP aux vector\n bit. The interpreter keeps control of mapping the binary.\n\n The second mode is loader substitution. This allows a binary to be\n executed natively and only the interpreter to be changed.\n\n - Last, interpreters can be bound at registration time. Each\n interpreter is opened by its own write with the credentials the\n entry file was opened with. The program picks one per exec with\n bpf_binprm_select_interp().\n\n Ucounts are used to properly account for pre-opened interpreters\n via /proc/sys/user/max_binfmt_misc_interpreters\"\n\n* tag \u0027vfs-7.3-rc1.binfmt\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (63 commits)\n binfmt_misc: document the pre-opened interpreter limit\n selftests/exec: test the pre-opened interpreter limit\n binfmt_misc: correctly account pre-opened interpreters\n binfmt_misc: document interpreters bound by a \u0027B\u0027 entry\n selftests/exec: test interpreters bound to a \u0027B\u0027 entry\n binfmt_misc: let a \u0027B\u0027 entry bind its interpreters\n binfmt_misc: carry pre-opened interpreters in struct binfmt_misc_interp\n selftests/exec: share the bpf handler preconditions\n binfmt_misc: document registering an entry disabled\n selftests/exec: test registering an entry disabled\n selftests/exec: let binfmt_flag_supported() return a bool\n selftests/exec: check that a binfmt_misc instance cannot be pinned\n binfmt_misc: let a register string create an entry disabled\n binfmt_misc: document loader substitution\n selftests/exec: test binfmt_misc loader substitution\n binfmt_misc: let a bpf handler request loader substitution\n binfmt_misc: add the \u0027L\u0027 loader substitution flag\n binfmt_elf_fdpic: consume a stashed PT_INTERP substitute\n binfmt_elf: consume a stashed PT_INTERP substitute\n exec: carry a PT_INTERP substitute in struct linux_binprm\n ...\n" }, { "commit": "043d7a2b4045df5eebc699c39bc3a571afb9f1ae", "tree": "687a0028970bc4f0f9794378e7c4517662111cb9", "parents": [ "8d3ae59288f1e7d58d76558a6ee96d533bc5019f", "1184f5e8200902cae3e098ac68d8600ecdf6fe28", "d7337cad4dbe6caf9535d3539daed353dd425dc1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 08:27:30 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 17 08:27:30 2026 -0700" }, "message": "Merge tags \u0027ipc-7.3-rc1.misc\u0027 and \u0027kernel-7.3-rc1.misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull misc ipc and core updates from Christian Brauner:\n\n - reject mq_notify() with a zero signal number\n\n - fix coding style in the exit path\n\n* tag \u0027ipc-7.3-rc1.misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n mqueue: reject mq_notify with signo 0\n\n* tag \u0027kernel-7.3-rc1.misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n kernel: exit: fix coding style missing spaces\n" }, { "commit": "4e926dc3d6ad35089ca45af6d1cee36bc24a80fb", "tree": "73d0b0dd22837d9751d1a0467f61241f9c12f280", "parents": [ "75fe716d07702e687718f743c6655bbe464e337a" ], "author": { "name": "David Heidelberg", "email": "david@ixit.cz", "time": "Sat Jun 27 19:50:15 2026 +0200" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Mon Aug 17 16:38:28 2026 +0200" }, "message": "dt-bindings: arm: qcom,ids: Add SoC ID for Snapdragon SDM 850\n\nAdd SoC ID for Qualcomm Snapdragon SDM850.\n\nSigned-off-by: David Heidelberg \u003cdavid@ixit.cz\u003e\nReviewed-by: Konrad Dybcio \u003ckonrad.dybcio@oss.qualcomm.com\u003e\nAcked-by: Krzysztof Kozlowski \u003ckrzysztof.kozlowski@oss.qualcomm.com\u003e\nLink: https://lore.kernel.org/r/20260627-sda850-v2-1-44bf46ade42e@ixit.cz\nSigned-off-by: Bjorn Andersson \u003candersson@kernel.org\u003e\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\n" }, { "commit": "dfda289ad743d037f33f9e3796e3fd40ca23374c", "tree": "cbd9860465badcbf42237e8d982499a91d3f8c8c", "parents": [ "63dbd330b5952a017abfd6ea86663bff29573187", "0c2a8eed95160e41b367a1fa605a2c1b24a9639a" ], "author": { "name": "Alexandre Belloni", "email": "alexandre.belloni@bootlin.com", "time": "Mon Aug 17 15:08:09 2026 +0200" }, "committer": { "name": "Alexandre Belloni", "email": "alexandre.belloni@bootlin.com", "time": "Mon Aug 17 15:08:09 2026 +0200" }, "message": "Merge tag \u0027mvebu-dt64-7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gclement/mvebu into soc/dt\n\nmvebu dt64 for 7.3 (part 1)\n\nFix usb3 phys declaration on Turris MOX an Armada 3700 based Soc\n\n* tag \u0027mvebu-dt64-7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gclement/mvebu:\n arm64: dts: turris-mox: fix usb3 phys\n\nSigned-off-by: Alexandre Belloni \u003calexandre.belloni@bootlin.com\u003e\n" }, { "commit": "63dbd330b5952a017abfd6ea86663bff29573187", "tree": "c8a2d77eac025be68aaad33ced5fe224e55fbea5", "parents": [ "6fdec78001b3b94f1b6000bbb2f9ec3498f1c73e", "7a90520e8c9a8f48fe6f3e741cf34d8c2b6dbc61" ], "author": { "name": "Alexandre Belloni", "email": "alexandre.belloni@bootlin.com", "time": "Mon Aug 17 15:06:22 2026 +0200" }, "committer": { "name": "Alexandre Belloni", "email": "alexandre.belloni@bootlin.com", "time": "Mon Aug 17 15:06:22 2026 +0200" }, "message": "Merge tag \u0027mvebu-dt-7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gclement/mvebu into soc/dt\n\nmvebu dt for 7.3 (part 1)\n\nAdd regulator supplies for helios4 (Armada 388 based)\nCorrect indentation on dts files\n\n* tag \u0027mvebu-dt-7.3-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/gclement/mvebu:\n ARM: dts: helios4: add SATA regulator supplies\n ARM: dts: helios4: add vcc-supply to GPIO expander\n ARM: dts: helios4: add vcc-supply to EEPROM\n ARM: dts: marvell: orion5x: Correct indentation\n ARM: dts: marvell: kirkwood: Correct indentation\n ARM: dts: marvell: armada: Correct indentation\n\nSigned-off-by: Alexandre Belloni \u003calexandre.belloni@bootlin.com\u003e\n" }, { "commit": "d8a2860b4a366bfa8acb3d64da2c546ea26d2091", "tree": "f9e31b6fc545f42de9093831602b1215f6689657", "parents": [ "7df1638df97b2aaaff4731b72d4940053257c952" ], "author": { "name": "Josh Poimboeuf", "email": "jpoimboe@kernel.org", "time": "Fri Aug 14 19:36:34 2026 -0700" }, "committer": { "name": "Ingo Molnar", "email": "mingo@kernel.org", "time": "Mon Aug 17 10:53:55 2026 +0200" }, "message": "objtool/klp: Fix vmlinux klp relocations for EXPORT_SYMBOL_FOR_MODULES()\n\nWhen a module function references a vmlinux symbol which is exported\nwith EXPORT_SYMBOL_FOR_MODULES(), a patch to that function needs to use\na klp reloc.\n\nCurrently, livepatch fails to load such a module:\n\n livepatch: invalid access to vmlinux symbol \u0027get_task_policy\u0027 from module-specific livepatch relocation section\n livepatch: failed to initialize patch \u0027livepatch_test\u0027 for module \u0027testmod\u0027 (-22)\n livepatch: patch \u0027livepatch_test\u0027 failed for module \u0027testmod\u0027, refusing to load module \u0027testmod\u0027\n\nklp diff puts all klp relocs in __klp_relocs.\u003cpatched object\u003e, so\npost-link names the section .klp.rela.\u003cpatched object\u003e.\u003csecname\u003e, which the\nkernel rejects for vmlinux symbols.\n\nCommit 07f14d6af9d77 (\"objtool/klp: Fix cross-module klp relocation\nsection naming\") changed the meaning of objname in the klp rela section\nname to be where the referenced symbol is referenced rather than where\nit lives. That premise only holds for symbols in a module: the relocs\nget applied when the patched module gets patched, and the module\ndependency guarantees the referenced module is loaded by then.\n\nA vmlinux symbol needs the opposite. It\u0027s always resolvable, and it has\nto be applied when the patch module loads, before the module loader\ninitializes the patch module\u0027s special sections, which may reference it.\nThat\u0027s why livepatch rejects vmlinux symbols in module-specific\nsections.\n\nUse \"vmlinux\" as the section objname when the referenced symbol lives in\nvmlinux. This moves such klp relocs from .klp.rela.kvm..text to\n.klp.rela.vmlinux..text.\n\nFixes: 07f14d6af9d77 (\"objtool/klp: Fix cross-module klp relocation section naming\")\nReported-by: Dylan Hatch \u003cdylanbhatch@google.com\u003e\nSigned-off-by: Josh Poimboeuf \u003cjpoimboe@kernel.org\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@kernel.org\u003e\nAcked-by: Song Liu \u003csong@kernel.org\u003e\nLink: https://patch.msgid.link/f8e3b9fae109903a6aafb2a33310e4afdcebf58e.1786761327.git.jpoimboe@kernel.org\nCloses: https://lore.kernel.org/CADBMgpz7iWC0\u003dt\u003d_gE-tfvv0mTPq4kg0qQ2zgPH8DVPE6eQ9Kw@mail.gmail.com\n" }, { "commit": "0eaed89c18aeedf0898baf2dbf5ff027c6795152", "tree": "4422cb10597581af05704ba6510018aabf25b268", "parents": [ "c66494c79ede1af529dbf67f9ed6fdbf42e05ef3", "8b4127f6db40381229f3564d34ac35f36311c201" ], "author": { "name": "Thomas Gleixner", "email": "tglx@kernel.org", "time": "Mon Aug 17 10:29:52 2026 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@kernel.org", "time": "Mon Aug 17 10:29:52 2026 +0200" }, "message": "Merge tag \u0027timers-v7.3-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/daniel.lezcano/linux into timers/clocksource\n\n - Use designated initializers for sh_mtu2, sh_cmt, and sh_tmu, and\n drop the unused initializer in the platform_device_id table for\n sh_mtu2 (Uwe Kleine-König)\n\n - Remove redundant dev_err()/dev_err_probe() messages when\n devm_request_*_irq() fails, as the helper already logs an error\n message (Pan Chuang)\n\n - Fix a boot hang on Allwinner D1 when a forced minimum delta is used\n with the sun4i timer (Felix Yan)\n\n - Fix an IRQ leak in the cpuhp_setup_state() error path by freeing the\n IRQ on failure in the NXP PIT driver (WenTao Liang)\n\n - Fix incorrect unmapping of shared MMIO between the clocksource and\n clockevent drivers. If one of them fails to initialize, the error\n path unmaps the shared MMIO region, leaving the other driver with an\n invalid mapping on clps711x (Guangshuo Li)\n\n - Make the samsung_pwm driver compatible with PREEMPT_RT by replacing\n regular spinlocks with raw_spinlock_t in atomic contexts (Marek\n Szyprowski)\n\n - Use __raw_readl() and __raw_writel() instead of ioread32() and\n iowrite32() to support SWAP_IO_SPACE in the rtl-otto driver (Rustam\n Adilov)\n\n - Fix a missing clk_disable_unprepare() call in the timer\n initialization error path of the Armada driver (Yuho Choi)\n\nLink: https://lore.kernel.org/lkml/75feea31-683d-45a1-87f4-ab045e0152ae@oss.qualcomm.com\n" }, { "commit": "8d3ae59288f1e7d58d76558a6ee96d533bc5019f", "tree": "a25d6a94ad99aa059aa6dc728c45052fb2968352", "parents": [ "fd923b32d7614047c8b2acecae3915ec94f7afab" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 14:32:26 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 14:32:26 2026 -0700" }, "message": "Linux 7.2\n" }, { "commit": "fd923b32d7614047c8b2acecae3915ec94f7afab", "tree": "30d7b3414c9b907d554b504e8a5a90a98a801a5d", "parents": [ "240de1acf318ba53b6d34094030822797c65154a", "333238da9a193ffc58792995f3e951e4cb87bfd2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 11:15:23 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 11:15:23 2026 -0700" }, "message": "Merge tag \u0027sched_urgent_for_v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull scheduler fix from Borislav Petkov:\n\n - Make sure a delayed sched entity\u0027s runtime stats are updated at the\n right time so that it receives the proper lag compensation\n\n* tag \u0027sched_urgent_for_v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n sched: Update time before requeueing delayed entities\n" }, { "commit": "240de1acf318ba53b6d34094030822797c65154a", "tree": "b04879ba4642b741300e00d96c39f90b8bf5c60e", "parents": [ "7820dd4a127ae83b530e177faa8e213c2d5717e1", "94f39e4c017e66130e476268bdaa0bf61e914fa2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 11:12:13 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 11:12:13 2026 -0700" }, "message": "Merge tag \u0027timers_urgent_for_v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull timer fixes from Borislav Petkov:\n\n - Detect a broken EL2 virtual timer in the bcm2712 SoC boards (RPi5)\n and fallback to the physical one instead\n\n - Fix a build error with ARM rpc_defconfig and function tracer enabled\n\n* tag \u0027timers_urgent_for_v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n clocksource/drivers/arm_arch_timer: Workaround bcm2712 broken EL2 virtual timer\n tick: Include ktime.h and jiffies.h in linux/tick.h\n" }, { "commit": "7820dd4a127ae83b530e177faa8e213c2d5717e1", "tree": "e799beeb2700da6fe49cd0bedb119c6cdcb4e6f8", "parents": [ "d6e7d57ed967def9c964a58328ffba409f7efb64", "ada54c2ba652348c590d1ace6a2f4ff77cbbf809" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 11:09:37 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 11:09:37 2026 -0700" }, "message": "Merge tag \u0027core_urgent_for_v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull rseq fix from Borislav Petkov:\n\n - Prevent a lockup when rseq grants a timeslice extension\n\n* tag \u0027core_urgent_for_v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n rseq: Prevent hard lockup on granted time slice extension\n" }, { "commit": "d6e7d57ed967def9c964a58328ffba409f7efb64", "tree": "045c036568c4947493f72a01e8ee40e36eb5a5ef", "parents": [ "d5b95e612cde33f9def1a7a6c3242d03d3bbde3a" ], "author": { "name": "Charlie-cy Wu", "email": "charlie-cy.wu@mediatek.corp-partner.google.com", "time": "Mon Jun 29 16:35:43 2026 +0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 11:07:18 2026 -0700" }, "message": "wifi: mt76: mt7921: refactor regd update to fix recursive mutex deadlock\n\nSplit mt7921_mcu_regd_update() into two functions to prevent recursive\nmutex acquisition. Introduce __mt7921_mcu_regd_update() as the internal\nimplementation that assumes the mutex is already held by the caller,\nwhile mt7921_mcu_regd_update() remains as the external interface that\nhandles mutex acquisition and release.\n\nThis fixes a deadlock issue when mt7921_regd_set_6ghz_power_type() is\ncalled with the device mutex already held. Without this change, calling\nmt7921_mcu_regd_update() would attempt to acquire the same mutex again,\ncausing a recursive lock deadlock.\n\nThe __mt7921_mcu_regd_update() function can be safely called when the\ncaller has already acquired the device mutex, avoiding the deadlock\nwhile maintaining proper synchronization for regulatory domain updates.\n\nFixes: dc2608cf5224 (\"wifi: mt76: mt7921: refactor regulatory notifier flow\")\nSigned-off-by: Charlie-cy Wu \u003cCharlie-cy.Wu@mediatek.com\u003e\nTested-by: Mikhail Gavrilov \u003cmikhail.v.gavrilov@gmail.com\u003e\nTested-by: Devin Wittmayer \u003clucid_duck@justthetip.ca\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d5b95e612cde33f9def1a7a6c3242d03d3bbde3a", "tree": "2466a37891ec5c469ee3785090cf861ed2d0d3f9", "parents": [ "9da3fc37f5fe8b5adc0c6dd798d2caa3855dac4c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 10:40:14 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 10:40:14 2026 -0700" }, "message": "Revert \"i2c: designware: defer probe if child GpioInt controllers are not bound\"\n\nThis reverts commit 0a4bb2abc3e56d7be6e69b050c88ba52c87e22bf.\n\nThis was reported to break the touchpad on at least some Thinkpads, and\nwhile the revert has hit the i2c tree, it hasn\u0027t hit mine. So I\u0027m\nreverting it directly just to have this resolved for the imminent 7.2\nrelease.\n\nReported-by: Thorsten Leemhuis \u003clinux@leemhuis.info\u003e\nLink: https://lore.kernel.org/all/b4a4eadb-282f-464c-843a-19d415a34d0c@leemhuis.info/\nCc: Mario Limonciello \u003cmario.limonciello@amd.com\u003e\nCC: Hardik Prakash \u003chardikprakash.official@gmail.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "9da3fc37f5fe8b5adc0c6dd798d2caa3855dac4c", "tree": "625f8a4ca1eec67fce40f733cadc6d9d13117a62", "parents": [ "16429bb371999e26b243f6462234d841d271c5f1", "42c5ca1f0a288a52878bd72a5595b08261057438" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 10:31:05 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 10:31:05 2026 -0700" }, "message": "Merge tag \u0027perf_urgent_for_v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull perf fixes from Borislav Petkov:\n\n - Prevent the use of exited events as group leaders\n\n - Avoid use-after-free of an event\u0027s group leader by promoting detached\n sibling events to standalone entities and correct related accounting\n and state transitions\n\n* tag \u0027perf_urgent_for_v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n perf/core: Fix group leader use-after-free after sibling detach\n perf: Reject exited events as group leaders\n" }, { "commit": "16429bb371999e26b243f6462234d841d271c5f1", "tree": "5219b4f4f0afc783c3630ca9c43927795b047a21", "parents": [ "dcb68831eac76dbfda1cf5930d3003d938890d34", "abe7c8b09bd72a9c726016257c6281f129b4c02d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 10:28:31 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 10:28:31 2026 -0700" }, "message": "Merge tag \u0027x86_urgent_for_v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 fix from Borislav Petkov:\n\n - Add a proper kernel cmdline option to control the TLB invalidation\n method on x86 prompted mainly by a recent finding on AMD related to\n INVLPGB/TYLBSYNC invalidations.\n\n Having the command line option is simply another way to alleviate\n the situation short-term\n\n* tag \u0027x86_urgent_for_v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/CPU: Add a tlbi\u003d cmdline switch\n" }, { "commit": "dcb68831eac76dbfda1cf5930d3003d938890d34", "tree": "eb0f406a90219e1387b57e483f3d56f17e327d81", "parents": [ "0bae94aab8208b7107a2dda5de6ce046466663fd", "c71bf113dfdf426bdaf106636f573ef87b6613a0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 07:00:40 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 07:00:40 2026 -0700" }, "message": "Merge tag \u0027block-7.2-20260815\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux\n\nPull block fix from Jens Axboe:\n \"A single fix for a regression in this cycle, where drbd would leak\n shared secrets over netlink. This restores the behavior to match\n what we had before\"\n\n* tag \u0027block-7.2-20260815\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux:\n drbd: don\u0027t leak the shared secret to unprivileged netlink dumps\n" }, { "commit": "b39c748d9db0c6f3ada5e7cef7f56415827d5f81", "tree": "67bf78715dfd83fc0d17ff4a4d1e3d065a75933d", "parents": [ "89679d7908e7f09dfa45cf9fe1ab59795455d197" ], "author": { "name": "Matt Turner", "email": "mattst88@gmail.com", "time": "Mon Aug 03 13:08:42 2026 -0400" }, "committer": { "name": "Magnus Lindholm", "email": "linmag7@gmail.com", "time": "Sun Aug 16 16:00:00 2026 +0200" }, "message": "alpha: read $gp and $sp explicitly for clang\n\nclang honors a local `register unsigned long x __asm__(\"$N\")` variable\nonly where it appears as an inline-asm operand; merely reading it does\nnot produce the contents of that register. So trap_init() passed an\nundefined global pointer to PAL_wrkgp, and load_PCB() stored an undefined\nstack pointer into the PCB that swpctx then loaded. Either one wedges an\nearly boot.\n\nRead the registers explicitly instead: an inline mov for $gp in\ntrap_init(), and the file-scope current_stack_pointer for $sp in\nload_PCB(). A file-scope register-asm variable is the form clang does\nsupport.\n\nSigned-off-by: Matt Turner \u003cmattst88@gmail.com\u003e\nReviewed-by: Maciej W. Rozycki \u003cmacro@orcam.me.uk\u003e\nReviewed-by: Magnus Lindholm \u003clinmag7@gmail.com\u003e\nTested-by: Magnus Lindholm \u003clinmag7@gmail.com\u003e\nLink: https://lore.kernel.org/r/20260803-alpha-clang-v1-2-1c4ba5ba7a64@gmail.com\nSigned-off-by: Magnus Lindholm \u003clinmag7@gmail.com\u003e\n" }, { "commit": "0bae94aab8208b7107a2dda5de6ce046466663fd", "tree": "8af82ec16610bc9b52027d099309969a266aab20", "parents": [ "3eb40771c00a8488fa6ed2cc1fe203477908bf38", "6ca662cc71df7eb4eaf1b4bcb07cd3f188ad19f2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 06:58:27 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Aug 16 06:58:27 2026 -0700" }, "message": "Merge tag \u0027io_uring-7.2-20260815\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux\n\nPull io_uring fix from Jens Axboe:\n \"Just a single fix for a potential issue on 32-bit x86 with PAE\"\n\n* tag \u0027io_uring-7.2-20260815\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux:\n io_uring/rsrc: reject overflowing regvec bvec byte counts\n" }, { "commit": "c71bf113dfdf426bdaf106636f573ef87b6613a0", "tree": "c785ed2ff6057c06adb5eba5c252c0ac3a37a3a3", "parents": [ "7f40b346462f563a0d6e841a77b5163d2a882a04" ], "author": { "name": "Christoph Böhmwalder", "email": "christoph.boehmwalder@linbit.com", "time": "Fri Aug 14 17:16:17 2026 +0200" }, "committer": { "name": "Jens Axboe", "email": "axboe@kernel.dk", "time": "Sat Aug 15 20:00:10 2026 -0600" }, "message": "drbd: don\u0027t leak the shared secret to unprivileged netlink dumps\n\nThe conversion to explicit netlink serialization dropped the\nexclude_sensitive parameter from net_conf_to_skb(), so each caller has\nto sanitize by hand. Two dump paths were missed:\ndrbd_nl_get_connections_dumpit() and the volume-less connection branch\nof get_one_status(). Neither op carries GENL_ADMIN_PERM, so any\nunprivileged local user could read the CRAM-HMAC secret.\n\nAdd a net_conf_to_skb_sanitized() wrapper and route all three callers\nthrough it.\n\nFixes: 8098eeb693c4 (\"drbd: replace genl_magic with explicit netlink serialization\")\nReported-by: Vivek Parikh \u003cvivek.parikh@breachx.ai\u003e\nSigned-off-by: Christoph Böhmwalder \u003cchristoph.boehmwalder@linbit.com\u003e\nLink: https://patch.msgid.link/20260814151617.73752-1-christoph.boehmwalder@linbit.com\nSigned-off-by: Jens Axboe \u003caxboe@kernel.dk\u003e\n" }, { "commit": "d900723d78adec229996aefbab0dcaa66a177b77", "tree": "e11ad53ca4129204e2e98521f472dc41a115c03b", "parents": [ "a765d3c6cd88696fd233b9596b721400c6b25396" ], "author": { "name": "Jani Nikula", "email": "jani.nikula@intel.com", "time": "Fri Aug 07 19:29:41 2026 +0300" }, "committer": { "name": "Nicolas Schier", "email": "nsc@kernel.org", "time": "Sat Aug 15 21:01:51 2026 +0200" }, "message": "modpost: use mod_warn() and mod_error(), clean up logging\n\nConvert all module name logging to use the mod_warn() and mod_error()\nhelpers, and pass the module to modpost_log() where used directly, to\nalways have the module name prefixed in the log message, with .ko suffix\nfor modules.\n\nPass struct module *mod around in a few places instead of just\nmod-\u003ename.\n\nFurther unify the logging while at it. Use single quotes instead of\ndouble quotes for symbols, sections, and namespaces. Explicitly state\nit\u0027s a \"symbol\" when referencing symbols.\n\nSigned-off-by: Jani Nikula \u003cjani.nikula@intel.com\u003e\nLink: https://patch.msgid.link/17ed1bce5d54fb32533ba83bc83c429cb71adcb0.1786120005.git.jani.nikula@intel.com\nReviewed-by: Nicolas Schier \u003cnsc@kernel.org\u003e\nReviewed-by: Nathan Chancellor \u003cnathan@kernel.org\u003e\nSigned-off-by: Nicolas Schier \u003cnsc@kernel.org\u003e\n" }, { "commit": "3eb40771c00a8488fa6ed2cc1fe203477908bf38", "tree": "922aec7a266981d35502a328027e5f5536d0bde5", "parents": [ "5e060ff9d18748dbb21b12b821fcfc738823ba93", "e36c0670d5ebebd7dba493f14966051f354fbb34" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Aug 15 08:36:26 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Aug 15 08:36:26 2026 -0700" }, "message": "Merge tag \u0027soc-fixes-7.2-3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc\n\nPull SoC fixes from Arnd Bergmann:\n \"These are three last-minute fixes for the 7.2 release, though nothing\n alarming:\n\n - one error handling fix for optee firmware\n\n - incorrect i2c data for the apple M3 that was added in 7.2\n\n - a boot time warning fix for nvidia tegra\"\n\n* tag \u0027soc-fixes-7.2-3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc:\n arm64: tegra: Add EL2 virtual timer interrupt for Tegra194\n arm64: dts: apple: t8122: Fix I2C resources\n optee: ffa: Add NULL check in optee_ffa_lend_protmem\n" }, { "commit": "5e060ff9d18748dbb21b12b821fcfc738823ba93", "tree": "868901d5e925aba97300d5bb18d9bac9b3bd6610", "parents": [ "15ef2f78c49d20d53ec7c0f1c9b40b02e089f2d6", "32ef1b30ad736519f7a207bcc2986f3d4129d972" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Aug 15 08:05:58 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Aug 15 08:05:58 2026 -0700" }, "message": "Merge tag \u0027for-linus\u0027 of https://github.com/openrisc/linux\n\nPull OpenRISC fix from Stafford Horne:\n \"A bug fix found by researchers:\n\n - mask all privileged bits when restoring the supervisor register\n from sigreturn\"\n\n* tag \u0027for-linus\u0027 of https://github.com/openrisc/linux:\n openrisc: signal: do not restore privileged SR bits on sigreturn\n" }, { "commit": "32ef1b30ad736519f7a207bcc2986f3d4129d972", "tree": "dcdcde418e1ce69a8631620374b00e23571ae795", "parents": [ "db2ddb87143519e20a95aa36c60b36107b736a58" ], "author": { "name": "Ali Ahmet Memis", "email": "ali@iusegentoo.com", "time": "Fri Aug 07 23:42:30 2026 +0000" }, "committer": { "name": "Stafford Horne", "email": "shorne@gmail.com", "time": "Sat Aug 15 07:14:55 2026 +0100" }, "message": "openrisc: signal: do not restore privileged SR bits on sigreturn\n\nrestore_sigcontext() copies the whole supervision register (SR) from the\nsignal frame and only clears SPR_SR_SM before the value is reloaded into\nthe hardware SR (through ESR and l.rfe) on the return to user space. All\nother SR bits are left under user control.\n\nAn unprivileged task can thus return from a signal handler through a\ncrafted sigframe that clears SPR_SR_DME. With the data MMU disabled the\nCPU performs no translation or protection on data accesses, so the task\ngains read and write access to arbitrary physical memory, a local\nprivilege escalation. SPR_SR_IME, SPR_SR_SUMRA, SPR_SR_LEE, SPR_SR_EPH\nand the cache-enable bits are exposed the same way. The ptrace GPR regset\nalready refuses any change to SR for exactly this reason.\n\nRestore only the arithmetic flag bits (F, CY, OV) from the signal frame\nand take every privileged control bit from the SR the kernel saved on\nsignal entry.\n\nVerified with qemu-system-or1k -M or1k-sim: before this change an\nunprivileged PoC clears SPR_SR_DME in rt_sigreturn and writes a marker to\nphysical address 0x03000000 (beyond the kernel\u0027s mem\u003d32M); afterwards the\nsame PoC receives SIGSEGV and physical memory is unchanged.\n\nFixes: ac689eb7f9d4 (\"OpenRISC: Signal handling\")\nCc: stable@vger.kernel.org\nSigned-off-by: Ali Ahmet Memis \u003cali@iusegentoo.com\u003e\nSigned-off-by: Stafford Horne \u003cshorne@gmail.com\u003e\n" }, { "commit": "15ef2f78c49d20d53ec7c0f1c9b40b02e089f2d6", "tree": "4bfdb0f0823fd9195932486565e1c78bfae25863", "parents": [ "dac3e89a2c90c2feeb471e1f22a2512ad424b792", "062dc4693e2c10d77de06f61e6f3faf37c0a8383" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 14 21:51:26 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 14 21:51:26 2026 -0700" }, "message": "Merge tag \u0027input-for-v7.2-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input\n\nPull input fixes from Dmitry Torokhov:\n\n - A couple of fixes to the sur40 touchscreen driver to correct\n registration and teardown ordering, and to fix error path\n unwinding when video device registration fails.\n\n* tag \u0027input-for-v7.2-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:\n Input: sur40 - fix V4L error path cleanup\n Input: sur40 - fix input device registration ordering\n" }, { "commit": "dac3e89a2c90c2feeb471e1f22a2512ad424b792", "tree": "7ba50dba561b413288b6329fd6d37ec95ba56ace", "parents": [ "b26d316aaa8e8b7ff6f4f77006a9b7c20ee848c8", "b2601e783a2e54f6963d65f0d94f96d96c146a3a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 14 15:48:05 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 14 15:48:05 2026 -0700" }, "message": "Merge tag \u0027drm-fixes-2026-08-15\u0027 of https://gitlab.freedesktop.org/drm/kernel\n\nPull drm fixes from Dave Airlie:\n \"While this is large for rc8 time but also AI driven fixes is a lot of\n it, we had a more traditional screw up, and a regression was just\n found in the fair scheduling patches that went in back in rc1. This\n reverts the fair scheduler back to an option and sets the default back\n to what it should have been. We might have been a bit overly zealous\n in switching over, but at least it feels more normal than the AI\n driven fixes.\n\n Apart from the scheduler, it\u0027s mostly amdgpu and xe fixes, with some\n misc fixes to the log code and connector code.\n\n scheduler:\n - revert fair scheduler patches due to regression\n - mark fair as experimental\n\n connector:\n - fix OOB read in hdmi audio infoframe\n\n log:\n - fix divide by 0 if module param is set to 0\n - fix OOB read on empty message\n - fix infinite loop for too large scale\n\n xe:\n - Fix DPT Allocation paths\n - Fixes around UM queue BO\n - Order ring writes before ring tail updates\n - Add termination on resume for PXP\n - Document Sentinel and make CTX_TIMESTAMP read TOCTOU-safe\n - Fix sync entry leak on OA config emit failure\n - Check managed mutex initilization errors\n - Fix min frequency setting\n - Fix xe_device_probe error path\n\n amdgpu:\n - Bounds checking fix in CS IOCTL\n - Bounds checking fix in GEM IOCTL\n - Display fixes\n - GPUVM fix\n - ASPM fix\n - UVD bounds checking fixes\n - VCE 3 fix\n - BT.2020 fixes\n - NBIF 6.3.1 fix\n - IP discovery fix\n\n radeon:\n - Runtime pm fix\n\n amdxdna:\n - skip attempting to populate unmapped pages\"\n\n* tag \u0027drm-fixes-2026-08-15\u0027 of https://gitlab.freedesktop.org/drm/kernel: (51 commits)\n drm/log: Fix infinite loop when scale is too large for display\n drm/log: Fix out-of-bounds read on empty message length\n drm/log: Fix division by zero when scale module parameter is 0\n drm/xe: Fix xe_device_probe() failure\n drm/xe: Fix a bug in pc_adjust_freq_bounds()\n drm/xe/oa: Check managed mutex initialization errors\n drm/xe/oa: Fix sync entry leak on OA config emit failure\n drm/xe/lrc: document sentinel and make CTX_TIMESTAMP read TOCTOU-safe\n drm/xe/pxp: add termination on resume\n drm/xe: Order ring writes before ring tail updates\n drm/xe/guc_ads: use uncached mapping for UM queue BO\n drm/xe/guc_ads: allocate UM queues in VRAM on dGFX\n drm/xe/guc_ads: allocate UM queues in a separate BO\n drm/xe: Fix DPT allocation paths.\n accel/amdxdna: Skip unmapped range in aie2_populate_range()\n drm/amdgpu: Prefer default discovery offset\n drm/amdgpu: Reject UVD message with invalid number of h265 refs\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 ...\n" }, { "commit": "b26d316aaa8e8b7ff6f4f77006a9b7c20ee848c8", "tree": "f3627e2f13ee153b67b9b320851471c5e8e5d13e", "parents": [ "06d9a86e0222dcf6858544ba7257994bfe9e63c7", "7ae8acad2d5c415ecb754409a4b64f09169c082e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 14 15:34:14 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 14 15:34:14 2026 -0700" }, "message": "Merge tag \u0027clk-fixes-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux\n\nPull clk fixes from Stephen Boyd:\n \"Fixes for the Qualcomm, Rockchip, and SpacemiT clk drivers:\n\n - Keep audio working on Rockchip rk3588 by skipping disabling unused\n clks\n\n - Fix SpacemiT USB2 clk data so they actually work and keep the HDMA\n bus clk enabled to avoid system hangs\n\n - Avoid clk hangs on Qualcomm Eliza display hardware and revert a\n patch that breaks PCIe on some Qualcomm platforms\"\n\n* tag \u0027clk-fixes-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux:\n dt-bindings: clock: Replace bouncing emails\n Revert \"clk: qcom: regmap-phy-mux: Rework the implementation\"\n clk: spacemit: k3: set hdma clock as critical\n clk: spacemit: k3: fix USB2 bus clock\n clk: qcom: dispcc-eliza: Fix disp_cc_mdss_mdp_clk_src RCG stall on Eliza EVK\n clk: rockchip: rk3588: don\u0027t disable unused I2S MCLK output gates\n" }, { "commit": "06d9a86e0222dcf6858544ba7257994bfe9e63c7", "tree": "e1c284b8c553e1b66cf8fc8f2eb57ed5b98e195f", "parents": [ "53313bf327bbc7fa6580d9be314a135513e46fa9", "11058bd3d47d57eb3473935feae53868d6d168b7" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 14 15:28:29 2026 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Aug 14 15:28:29 2026 -0700" }, "message": "Merge tag \u0027spi-fix-v7.2-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi\n\nPull spi fixes from Mark Brown:\n \"A couple of relatively minor (but as ever important if you\u0027re hitting\n them) and straightforward driver specific fixes, plus one new device\n ID documented in the DT bindings for the DesignWare controller\"\n\n* tag \u0027spi-fix-v7.2-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:\n spi: virtio: mark device ready before registering the controller\n spi: dw: fix wrong RX_SAMPLE_DLY setting after resume\n spi: dt-bindings: snps,dw-apb-ssi: Document Axiado AX3005\n" } ], "next": "53313bf327bbc7fa6580d9be314a135513e46fa9" }