smb: client: fix double-free in SMB2_open() replay
A response-bearing attempt can return a replayable error and free its
response buffer. If SMB2_open_init() fails before the next send, cleanup
retains the previous buffer type and frees that response again.
Reset response bookkeeping before each attempt to prevent the stale free.
Fixes: 4f1fffa23769 ("cifs: commands that are retried should have replay flag set") Cc: stable@vger.kernel.org Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com> Signed-off-by: Steve French <stfrench@microsoft.com>
Zhao Zhang [Thu, 18 Jun 2026 15:28:05 +0000 (23:28 +0800)]
smb: client: fix double-free in SMB2_flush() replay
SMB2_flush() keeps its response buffer bookkeeping across replay
attempts. If a replayable flush response is received and the retry then
fails before cifs_send_recv() stores a replacement response, flush_exit
will free the stale response pointer a second time.
Reinitialize resp_buftype and rsp_iov at the top of the replay loop so
cleanup only acts on response state produced by the current attempt.
This fixes a double-free without changing replay handling for successful
requests.
Fixes: 4f1fffa23769 ("cifs: commands that are retried should have replay flag set") Cc: stable@vger.kernel.org Reported-by: Yuan Tan <yuantan098@gmail.com> Reported-by: Zhengchuan Liang <zcliangcn@gmail.com> Reported-by: Xin Liu <bird@lzu.edu.cn> Assisted-by: Codex:GPT-5.4 Acked-by: Henrique Carvalho <henrique.carvalho@suse.com> Signed-off-by: Zhao Zhang <zzhan461@ucr.edu> Signed-off-by: Ren Wei <n05ec@lzu.edu.cn> Signed-off-by: Steve French <stfrench@microsoft.com>
Remove the reference to the obsolete kselftest wiki.
The kselftest wiki is marked obsolete and is no longer updated. The last
edit was in 2019, and the information is outdated, referring readers for
support to IRC networks that have not been used for years, and to kernel
versions that are no longer supported.
If there is any relevant information left in the wiki it needs to be
cleaned up and moved into the canonical kselftest documentation here.
Link: https://lore.kernel.org/r/20260115172817.7120-1-bacs@librecast.net Signed-off-by: Brett A C Sheffield <bacs@librecast.net> Acked-by: Bagas Sanjaya <bagasdotme@gmail.com> Acked-by: SeongJae Park <sj@kernel.org> Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Linus Torvalds [Thu, 18 Jun 2026 22:03:21 +0000 (15:03 -0700)]
Merge tag 'pinctrl-v7.2-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl
Pull pin control updates from Linus Walleij:
"Core changes:
- Add new generic callbacks to populate per-pin pin controllers
creating groups and functions from the device tree building out
pinctrl_generic_to_map() and move the Spacemit driver over to use
this
- Generic board-level pin control driver using the mux framework
New pin controller drivers:
- Amlogic (meson) A9 SoC
- Aspeed AST2700 SoC0 and SoC1
- nVidia Tegra264 and Tegra238
- Qualcomm Nord TLMM, Shikra TLMM, SM6350 LPASS LPI, and IPQ9650 TLMM
- Renesas RZ/G3L SoC
- UltraRISC DP1000
Improvements:
- Handle pull up/pull down properly in the Renesas RZG2L driver
- Fix up nVidia Tegra 234 DT bindings
- Fix up pin definitions in the Qualcomm Eliza driver
- Qualcomm PM8010 GPIO support in the PM8010
- Qualcomm SM6115 EGPIO support in the SM6115
- Switch Qualcomm LPASS LPI drivers to use runtime PM for power
management
- Clean up the Qualcomm Kconfig business a bit to include the
necessary drivers for each subarch
- Fix output glitch in the Amlogic (meson) A4 pin controller
- Move the Airoha driver from the Mediatek directory to its own
directory. It is too different from other Mediatek hardware
- A slew of fixes to the Airoha AN7581 and AN7583 drivers"
* tag 'pinctrl-v7.2-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl: (151 commits)
pinctrl: Export pinctrl_get_group_selector()
pinctrl: Match DT helper types
pinctrl: qcom: Register functions before enabling pinctrl
pinctrl: meson: amlogic-a4: use nolock get range
pinctrl: ultrarisc: Add UltraRISC DP1000 pinctrl driver
dt-bindings: pinctrl: Add UltraRISC DP1000 pinctrl controller
pinctrl: qcom: Remove unused macro definitions
pinctrl: tegra: PINCTRL_TEGRA264 should depend on ARCH_TEGRA
pinctrl: tegra: PINCTRL_TEGRA238 should depend on ARCH_TEGRA
pinctrl: tegra238: add missing AON pin groups
dt-bindings: pinctrl: tegra238: add missing AON pin groups
pinctrl: airoha: an7583: remove undefined groups from pcm_spi pin function
pinctrl: airoha: an7583: fix phy1_led1 pin function
pinctrl: airoha: an7583: add missed gpio22 pin group
pinctrl: airoha: an7583: fix gpio21 pin group
pinctrl: airoha: fix pwm pin function for an7581 and an7583
pinctrl: airoha: an7583: fix incorrect led mapping in phy4_led1 pin function
pinctrl: airoha: an7581: fix incorrect led mapping in phy4_led1 pin function
pinctrl: airoha: an7583: fix misprint in gpio19 pinconf
pinctrl: airoha: an7581: fix misprint in gpio19 pinconf
...
Linus Torvalds [Thu, 18 Jun 2026 21:55:20 +0000 (14:55 -0700)]
Merge tag 'backlight-next-7.2' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/backlight
Pull backlight updates from Lee Jones:
"New Support & Features:
- Maxim MAX25014: Add support for the Maxim MAX25014 4-channel
automotive grade backlight driver IC
Improvements & Fixes:
- Maintainers: Add the Congatec Board Controller backlight driver to
its corresponding entry
- Congatec Board Controller: Remove redundant X86 dependency from the
backlight driver
- Kinetic KTD2801: Enable BL_CORE_SUSPENDRESUME to ensure the chip is
powered off during suspend
Cleanups & Refactoring:
- Core: Use named initializers for i2c_device_id arrays to improve
readability and robustness
Device Tree Binding Updates:
- Maxim MAX25014: Add device tree bindings for the Maxim MAX25014
backlight controller"
* tag 'backlight-next-7.2' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/backlight:
backlight: Use named initializers for arrays of i2c_device_data
MAINTAINERS: Add cgbc backlight driver
backlight: cgbc: Remove redundant X86 dependency
backlight: Add max25014atg backlight
dt-bindings: backlight: Add max25014 support
backlight: ktd2801: Enable BL_CORE_SUSPENDRESUME
Linus Torvalds [Thu, 18 Jun 2026 21:45:08 +0000 (14:45 -0700)]
Merge tag 'leds-next-7.2' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/leds
Pull LED updates from Lee Jones:
"New Support & Features:
- Samsung S2MU005: Add support for the Samsung S2MU005 PMIC which
includes flash and RGB LED controllers
- Texas Instruments:
- LP5812: Add support for the TI LP5812 LED driver
- LP5860: Add support for the Texas Instruments LP5860 LED matrix
driver via SPI
Improvements & Fixes:
- Core:
- Adjust the brightness sysfs node documentation to clarify that
only decimal values are accepted
- Fix a race condition in the software blink logic when stopping
blinking and setting brightness simultaneously
- Introduce the `multi_max_intensity` sysfs attribute for
multicolor LEDs to support hardware-based global brightness
control
- Replace OF-based device lookup with firmware node equivalents
to support ACPI and software nodes
- Return `ENODATA` when reading brightness from
hardware-controlled LEDs
- Set the coherent DMA mask to zero for the Samsung PMIC device
to suppress unnecessary "DMA mask not set" messages
- ams OSRAM AS3668: Fix a Kconfig symbol name mismatch in the
Makefile that prevented the driver from being built
- BlinkM: Fix spelling and comment style issues in the driver
- DAC124S085: Declare the SPI command word as `__le16` to ensure
correct endianness and pass sparse checks
- GPIO Trigger: Use `GPIOD_FLAGS_BIT_NONEXCLUSIVE` to allow sharing
GPIOs between the LED trigger and other drivers
- NXP PCA9532: Fix an issue where the LED would stop blinking when
changing brightness to a non-zero value
- Qualcomm: Unify the user-visible company name to "Qualcomm" across
flash LED config options
- Qualcomm LPG: Optimize memory allocation by combining main
structure and channels into a single allocation using flexible
array members
- Texas Instruments
- LP5860: Add missing `CONFIG_OF` dependency to prevent build
warnings
- TPS6131x: Increase the overvoltage protection threshold to 6V
to avoid false triggers with 5V input supplies
- Userspace LEDs (uLEDs):
- Fix a potential buffer overread by using `strnchr()` for name
string validation
- Return `-EFAULT` on `copy_to_user()` failure to properly handle
read errors
Cleanups & Refactoring:
- Core:
- Convert various `i2c_device_id` arrays to use named
initializers for improved robustness and readability
- Multi-color: Fix incorrect `KernelVersion` and `Date` tags for
the `multi_max_intensity` ABI
- Broadcom BCM63138 / ChromeOS EC: Move `MODULE_DEVICE_TABLE`
declarations next to the ID tables for consistency
- LP5812: Fix a sysfs ABI reference in the documentation
- ST1202: Remove an unused legacy GPIO header include
Device Tree Binding Updates:
- Class: Document the keyboard backlight LED class naming
conventions, including a new scheme for zoned backlights
- Core: Dual-license the common LED bindings header under GPLv2
and BSD-2-Clause
- IR SPI LED: Add a new 30% duty-cycle value for the IR transmitter
used in Xiaomi Redmi Note 8
- Samsung S2M series:
- Document the flash LED device bindings for Samsung S2M series
PMICs
- Document the pattern behavior for Samsung S2M series PMIC RGB
- S2MU005: Add device tree bindings for the S2MU005 PMIC,
including its flash and RGB LED sub-devices
- TI LM3560: Document the TI LM3559 and LM3560 synchronous boost
flash drivers"
* tag 'leds-next-7.2' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/leds: (30 commits)
leds: tps6131x: Increase overvoltage protection threshold to 6V
leds: Fix sysfs ABI date
leds: Fix CONFIG_OF dependency for LEDS_LP5860_CORE
leds: uleds: Fix potential buffer overread
leds: Use named initializers for arrays of i2c_device_data
leds: uleds: Return -EFAULT on copy_to_user() failure
leds: core: Report ENODATA for brightness of hardware controlled LED
leds: class: Use firmware nodes for device lookup
Documentation: leds: Document pattern behavior of Samsung S2M series PMIC RGB LEDs
leds: rgb: Add support for Samsung S2M series PMIC RGB LED device
leds: flash: Add support for Samsung S2M series PMIC flash LED device
dt-bindings: leds: Document Samsung S2M series PMIC flash LED device
leds: core: Fix race condition for software blink
leds: Adjust documentation of brightness sysfs node
leds: dac124s085: Declare SPI command word as __le16
leds: Introduce the multi_max_intensity sysfs attribute
dt-bindings: leds: Document TI LM3560 Synchronous Boost Flash Driver
leds: bcm63138/cros_ec: Move MODULE_DEVICE_TABLE next to the table itself
leds: Add support for TI LP5860 LED driver chip
Documentation: leds: leds-class: Document keyboard backlight LED class naming
...
Linus Torvalds [Thu, 18 Jun 2026 21:26:29 +0000 (14:26 -0700)]
Merge tag 'mfd-next-7.2' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd
Pull MFD updates from Lee Jones:
"New Support & Features:
- Renesas RSMU: Add support for the IDT 8a34002 Clock Matrix
- Samsung S2MU005: Add support for the Samsung S2MU005 PMIC which
includes charger, MUIC, flash and RGB LED controllers
- SpacemiT P1: Add a reboot cell for the SpacemiT P1 chip
- Texas Instruments BQ25792: Add support for the TI BQ25792 charger
manager
Improvements & Fixes:
- Core: Unify the user-visible company name to "Qualcomm" across
various config options
- ChromeOS EC:
- Delay `dev_set_drvdata()` until the probe process has
successfully completed to avoid use-after-free issues
- Prevent adding `cros_ec_ucsi` as an MFD sub-device if it is
already defined in Device Tree or ACPI
- Cirrus Logic CS42L43: Add a sanity check for firmware size to
prevent out-of-bounds memory access during firmware loading
- Cirrus Logic CS5535: Associate the GPIO cell with a dedicated
software node to support board files requesting GPIOs
- Maxim MAX77620: Modernize poweroff handling by converting to the
sys-off API
- Qualcomm RPM: Add the missing QDSS clock resource for the MSM8960
SoC
- Renesas RSMU: Fix page register setup for the 8A3400x family by
correctly calculating the page address
- Renesas RZ/MTU3:
- Make the reset line optional to support newer SoC variants
(RZ/T2H, RZ/N2H)
- Modernize the driver by using device-managed APIs for reset
control and device addition
- Samsung Core: Set the coherent DMA mask to zero for the Samsung
PMIC device to suppress unnecessary "DMA mask not set" messages
- Silicon Motion SM501: Fix a reference leak on failed device
registration by properly dropping the platform device reference
- Texas Instruments:
- TPS65219: Make poweroff handler registration conditional on the
"system-power-controller" Device Tree property
- TPS6586x: Fix Device Tree node reference counting by manually
bumping the refcount for sub-devices
- TPS65910: Add return value checking for the dummy I2C transfer
used to work around silicon erratum SWCZ010
- TWL4030: Update board-specific checks to use Device Tree
compatibles instead of legacy machine IDs
Cleanups & Refactoring:
- Core: Consistently define `pci_device_id` arrays using named
initializers across various Intel and Silicon Motion drivers
- Maintainers: Shift maintenance of Samsung PMIC drivers to André
Draszik
- Maxim MAX77759: Improve code style by reformatting the IRQ table
and refining macro comments
- MEN MENF21BMC / Texas Instruments TWL: Correctly treat
`i2c_check_functionality()` as returning a boolean status
- Rohm BD72720: Drop the non-existent BUCK11 ID to improve code
clarity
- Silicon Labs Si476x: Fix various spelling mistakes in driver
comments
- Spreadtrum SC27xx: Transition to `devm_mfd_add_devices()` and
separate MFD cell tables for each PMIC model
- Timberdale: Move GPIO pin definitions into the driver and
transition to using a software node for the GPIO cell
- Wolfson WM8994: Remove dead legacy-GPIO code and its associated
`irq_gpio` member
Device Tree Binding Updates:
- Aspeed AST2x00: Document the AST2700 SCU0 and add support for its
SoC0/SoC1 pin controllers
- Hisilicon Hi655x: Convert the Hi655x PMIC binding from text format
to YAML DT schema
- Khadas MCU: Add a new compatible and fan-supply property for the
Khadas VIM4 MCU
- MediaTek MT6397: Add support for the MT6365 PMIC and document
regulator supplies for the MT6359 variant
- Qualcomm TCSR: Add compatibles for Nord and IPQ5210 TCSR blocks
- Renesas RZ/G3L: Revert the addition of the
`renesas,r9a08g046-lvds-cmn` compatible string due to documentation
errors
- Samsung S2MU005: Document the S2MU005 PMIC and its sub-devices
(charger, MUIC, flash and RGB LEDs)
- Spreadtrum SC2731: Include regulator bindings for the SC2730
variant
- STMPE: Fix the schema by marking 'compatible' and '#pwm-cells' as
required for the PWM subnode
- Texas Instruments BQ257xx: Expand the BQ25703A binding to include
the BQ25792 variant
Removals:
- Motorola EZX PCAP: Remove the unused and non-functional driver for
Motorola EZX phones"
* tag 'mfd-next-7.2' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd: (47 commits)
dt-bindings: mfd: syscon: Revert renesas,r9a08g046-lvds-cmn
dt-bindings: mfd: st,stmpe: Add missing properties for PWM subnode
mfd: rz-mtu3: Make reset optional
mfd: rz-mtu3: Store &pdev->dev in local variable
mfd: rz-mtu3: Use local variable for reset
mfd: rz-mtu3: Use device-managed APIs
dt-bindings: mfd: aspeed,ast2x00-scu: Support AST2700 SoC1 pinctrl
mfd: tps6586x: Fix OF node refcount
dt-bindings: mfd: sprd,sc2731: Include SC2730 regulator bindings
mfd: twl4030-power: Update checks for specific boards to use the DT
dt-bindings: mfd: qcom,tcsr: Document the IPQ5210 TCSR block
mfd: qcom_rpm: Add msm8960 QDSS clock resource
mfd: si476x-i2c: Fix spelling mistakes in comments
mfd: max77620: Convert poweroff support to sys-off API
mfd: dt-bindings: mt6397: Add regulator supplies
dt-bindings: mfd: mediatek: mt6397: Add MT6365 PMIC support
dt-bindings: mfd: mediatek: mt6397: Add rtc for MT6359
mfd: cs42l43: Sanity check firmware size
mfd: qcom: Unify user-visible "Qualcomm" name
mfd: cros_ec: Delay dev_set_drvdata() until probe success
...
Linus Torvalds [Thu, 18 Jun 2026 21:20:39 +0000 (14:20 -0700)]
Merge tag 'hid-for-linus-2026061601' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid
Pull HID updates from Jiri Kosina:
"Core:
- semantic cleanup fixes for 'hid_device_id::driver_data' (Pawel
Zalewski)
Multitouch:
- UX improvement fixes for Yoga Book 9 (Dave Carey)
Logitech:
- fix for high resolution scrolling for Logitech HID++ 2.0 devices
(Lauri Saurus)
CP2112:
- fix for cp2112 firmware-based speed configuration, if available
(Danny Kaehn)
Wacom:
- memory corruption and scheduling while atomic and error fixes and
error handling fixes (Jinmo Yang, Myeonghun Pak)
New device support:
- OneXPlayer (Derek J. Clark)
- HORI Wireless Switch Pad (Hector Zelaya)
- Rakk Dasig X (Karl Cayme)
And other assorted small fixes and device ID additions"
* tag 'hid-for-linus-2026061601' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid: (39 commits)
HID: hidpp: fix potential UAF in hidpp_connect_event()
HID: logitech-hidpp: sync wheel multiplier on wheel mode changes
HID: intel-thc-hid: intel-quickspi: reset touch IC on system resume
HID: uhid: convert to hid_safe_input_report()
HID: hid-goodix-spi: validate report size to prevent stack buffer overflow
HID: nintendo: add support for HORI Wireless Switch Pad
HID: multitouch: Honor ContactCount for Yoga Book 9 to suppress ghost contacts
HID: pidff: Use correct effect type in effect update
HID: wacom: stop hardware after post-start probe failures
HID: core: demote warning to debug level
HID: lenovo: Use KEY_PERFORMANCE capability for ThinkPad X12 Tab Gen 2
HID: lenovo: Add support for ThinkPad X13 Folio keyboard
HID: cp2112: Configure I2C bus speed from firmware
HID: cp2112: Add fwnode support
HID: hid-lenovo-go: cancel cfg_setup work in hid_go_cfg_remove()
HID: Input: Add battery list cleanup with devm action
HID: logitech-hidpp: remove excess kernel-doc member in hidpp_scroll_counter
HID: wacom: use cleanup.h for wacom_wac_queue_flush() buffer management
HID: wacom: use GFP_ATOMIC in wacom_wac_queue_flush()
HID: wacom: fix slab-out-of-bounds write in wacom_wac_queue_insert
...
Pengpeng Hou [Thu, 11 Jun 2026 06:00:00 +0000 (14:00 +0800)]
kconfig: add optional warnings for changed input values
When reading .config input, Kconfig stores user-provided values first
and then resolves the final value after applying dependencies, ranges,
and other constraints.
If the final value differs from the user input, Kconfig already tracks
that state internally, but it does not provide a focused diagnostic to
show which explicit inputs were adjusted. This is particularly confusing
for requested values that get forced down by unmet dependencies or
clamped by ranges.
Add an opt-in diagnostic controlled by KCONFIG_WARN_CHANGED_INPUT. Emit
the warnings from conf_write() and conf_write_defconfig() after value
resolution. Print the diagnostic to stderr directly, not through the
normal message callback, so it remains visible when conf is run with -s,
such as from make -s.
Keep the diagnostic out of the conf_message() formatting buffer so long
warning lists are not truncated, and mark processed symbols as written
before the SYMBOL_WRITE check so duplicate menu nodes cannot emit
duplicate warnings.
Document the new environment variable and add tests for olddefconfig,
savedefconfig, and the silent-conf path.
Rong Xu [Wed, 17 Jun 2026 22:46:23 +0000 (15:46 -0700)]
modpost: Ignore Clang LTO suffixes in symbol matching
When building the kernel with Clang ThinLTO enabled, the compiler
can mangle static variable names by appending suffixes such as
".llvm.<hash>" to prevent naming collisions across translation units.
This name mangling breaks the section mismatch whitelisting in modpost.
modpost relies on glob patterns (e.g., "*_ops" or "*_probe") to identify
safe references between permanent data and initialization code. Because
the LTO suffix modifies the end of the symbol name, legitimately
whitelisted structures fail the match, resulting in false positive
warnings.
For example, a static pernet_operations struct triggers the following:
Quang Nguyen [Thu, 18 Jun 2026 08:19:30 +0000 (09:19 +0100)]
spi: rpc-if: Use correct device for hardware reinitialization on resume
rpcif_spi_resume() currently passes the SPI controller device to
rpcif_hw_init(), but the function should be called with the RPC
interface device.
Retrieve the rpcif private data from the SPI controller and pass
rpc->dev instead. Also propagate the return value of rpcif_hw_init() so
that a failure during resume is properly reported rather than silently
ignored.
Fixes: ad4728740bd6 ("spi: rpc-if: Add resume support for RZ/G3E") Signed-off-by: Quang Nguyen <quang.nguyen.wx@renesas.com> Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be> Link: https://patch.msgid.link/20260618081932.172168-1-biju.das.jz@bp.renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
PCI: host-common: Request bus reassignment when not probe-only
pci_host_common_init() is used by several generic ECAM host drivers.
After PCI core changes around pci_flags and preserve_config, these hosts
no longer opted into full bus number reassignment the way they did
before, which broke enumeration of devices on a Marvell CN106XX board.
When PCI_PROBE_ONLY is not set, add PCI_REASSIGN_ALL_BUS so
pci_scan_bridge_extend() takes the reassignment path: bus numbers can be
assigned from firmware EA data (e.g. pci_ea_fixed_busnrs()). Skip the
flag in probe-only mode so existing assignments are not overridden.
Fixes: 7246a4520b4b ("PCI: Use preserve_config in place of pci_flags") Closes: https://lore.kernel.org/all/abkqm_LCd9zAM8cW@rkannoth-OptiPlex-7090/ Signed-off-by: Ratheesh Kannoth <rkannoth@marvell.com>
[mani: added stable tag] Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
[bhelgaas: add problem report link] Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Cc: stable@vger.kernel.org Cc: Vidya Sagar <vidyas@nvidia.com> Link: https://patch.msgid.link/20260414081730.3864372-1-rkannoth@marvell.com
Hans Zhang [Thu, 21 May 2026 16:18:20 +0000 (00:18 +0800)]
PCI: rockchip: Protect root bus removal with rescan lock
Hold the pci_rescan_remove_lock lock while stopping and removing a root bus
to avoid racing with concurrent rescan or hotplug operations triggered via
sysfs. Such races may lead to use-after-free issues or system crashes.
Hans Zhang [Thu, 21 May 2026 16:18:22 +0000 (00:18 +0800)]
PCI: plda: Protect root bus removal with rescan lock
Hold the pci_rescan_remove_lock lock while stopping and removing a root bus
to avoid racing with concurrent rescan or hotplug operations triggered via
sysfs. Such races may lead to use-after-free issues or system crashes.
Hans Zhang [Thu, 21 May 2026 16:18:19 +0000 (00:18 +0800)]
PCI: mediatek: Protect root bus removal with rescan lock
Hold the pci_rescan_remove_lock lock while stopping and removing a root bus
to avoid racing with concurrent rescan or hotplug operations triggered via
sysfs. Such races may lead to use-after-free issues or system crashes.
Hans Zhang [Thu, 21 May 2026 16:18:18 +0000 (00:18 +0800)]
PCI: iproc: Protect root bus removal with rescan lock
Hold the pci_rescan_remove_lock lock while stopping and removing a root bus
to avoid racing with concurrent rescan or hotplug operations triggered via
sysfs. Such races may lead to use-after-free issues or system crashes.
Hans Zhang [Thu, 21 May 2026 16:18:15 +0000 (00:18 +0800)]
PCI: dwc: Protect root bus removal with rescan lock
Hold the pci_rescan_remove_lock lock while stopping and removing a root bus
to avoid racing with concurrent rescan or hotplug operations triggered via
sysfs. Such races may lead to use-after-free issues or system crashes.
Hans Zhang [Thu, 21 May 2026 16:18:14 +0000 (00:18 +0800)]
PCI: cadence: Protect root bus removal with rescan lock
Hold the pci_rescan_remove_lock lock while stopping and removing a root bus
to avoid racing with concurrent rescan or hotplug operations triggered via
sysfs. Such races may lead to use-after-free issues or system crashes.
Hans Zhang [Thu, 21 May 2026 16:18:17 +0000 (00:18 +0800)]
PCI: brcmstb: Protect root bus removal with rescan lock
Hold the pci_rescan_remove_lock lock while stopping and removing a root bus
to avoid racing with concurrent rescan or hotplug operations triggered via
sysfs. Such races may lead to use-after-free issues or system crashes.
Hans Zhang [Thu, 21 May 2026 16:18:16 +0000 (00:18 +0800)]
PCI: altera: Protect root bus removal with rescan lock
Hold the pci_rescan_remove_lock lock while stopping and removing a root bus
to avoid racing with concurrent rescan or hotplug operations triggered via
sysfs. Such races may lead to use-after-free issues or system crashes.
Linus Torvalds [Thu, 18 Jun 2026 16:42:02 +0000 (09:42 -0700)]
Merge tag 'ext4_for_linus-7.2-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
Pull ext4 updates from Ted Ts'o:
- A major rework of the fast commit mechanism to avoid lock contention
and deadlocks. We also export snapshot statistics in
/proc/fs/ext4/*/fc_info
- Performance optimization for directory hash computation by processing
input in 4-byte chunks and removing function pointers, along with new
KUnit tests for directory hash
- Cleanups in JBD2 to remove special slabs and use kmalloc() instead
- Various bug fixes, including:
- Early validation of donor superblock in EXT4_IOC_MOVE_EXT to
avoid cross-fs deadlock
- Fix for a kernel BUG in ext4_write_inline_data_end under
data=journal
- Fix for a NULL dereference in jbd2_journal_dirty_metadata when
handle is aborted
- Fix for an underflow in JBD2 fast commit block initialization
check
- Fix for LOGFLUSH shutdown ordering to ensure ordered data
writeback
- Miscellaneous fixes for error path return values and KUnit
assertions
* tag 'ext4_for_linus-7.2-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
ext4: validate donor file superblock early in EXT4_IOC_MOVE_EXT
ext4: fix kernel BUG in ext4_write_inline_data_end
ext4: fix ERR_PTR(0) in ext4_mkdir()
jbd2: remove special jbd2 slabs
ext4: remove mention of PageWriteback
ext4: improve str2hashbuf by processing 4-byte chunks and removing function pointers
ext4: add Kunit coverage for directory hash computation
ext4: fast commit: export snapshot stats in fc_info
ext4: fast commit: add lock_updates tracepoint
ext4: fast commit: avoid i_data_sem by dropping ext4_map_blocks() in snapshots
ext4: fast commit: avoid self-deadlock in inode snapshotting
ext4: fast commit: avoid waiting for FC_COMMITTING
ext4: lockdep: handle i_data_sem subclassing for special inodes
ext4: fast commit: snapshot inode state before writing log
jbd2: fix integer underflow in jbd2_journal_initialize_fast_commit()
ext4: fix fast commit wait/wake bit mapping on 64-bit
jbd2: check for aborted handle in jbd2_journal_dirty_metadata()
ext4: Use %pe to print PTR_ERR()
ext4: fix LOGFLUSH shutdown ordering to allow ordered-mode data writeback
ext4: replace KUnit tests for memcmp() with KUNIT_ASSERT_MEMEQ()
Linus Torvalds [Thu, 18 Jun 2026 16:21:50 +0000 (09:21 -0700)]
Merge tag 'v7.2-rc-part1-ksmbd-fixes' of git://git.samba.org/ksmbd
Pull smb server updates from Steve French:
- Use after free fixes
- Out of bounds read fix
- Add SMB compression support both at rest and over the wire: support
decompression of compressed SMB2 requests, initially allow compressed
SMB2 READ responses, and implement get/set compression operations for
per-file compression state.
- Credentials fixes: for various FSCTLs, setinfo, delete on close and
for alternate data streams
- Fix access checks and permission checks in DUPLICAT_EXTENTS and
SET_ZERO_DATA fsctls, find_file_posix_info, FILE_LINK_INFORMATION and
smb2_set_info_sec
- Reject non valid session in compound request
- Serialize QUERY_DIRECTORY
- Prevent path traversal bypass by restricting caseless retry
- Path lookup fix
- Two minor cleanup fixes
* tag 'v7.2-rc-part1-ksmbd-fixes' of git://git.samba.org/ksmbd: (31 commits)
ksmbd: fix path resolution in ksmbd_vfs_kern_path_create
ksmbd: use opener credentials for FSCTL mutations
ksmbd: use opener credentials for ADS I/O
ksmbd: require source read access for duplicate extents
ksmbd: run set info with opener credentials
ksmbd: use opener credentials for delete-on-close
ksmbd: serialize QUERY_DIRECTORY requests per file
ksmbd: add permission checks for FSCTL_DUPLICATE_EXTENTS_TO_FILE
ksmbd: enforce FILE_READ_ATTRIBUTES on SMB_FIND_FILE_POSIX_INFORMATION
ksmbd: reject non-VALID session in compound request branch
ksmbd: compress SMB2 READ responses
ksmbd: negotiate and decode SMB2 compression
cifs: negotiate chained SMB2 compression capabilities
smb: add common SMB2 compression transform helpers
smb: move LZ77 compression into common code
ksmbd: add per-handle permission check to FILE_LINK_INFORMATION
ksmbd: add a permission check for FSCTL_SET_ZERO_DATA
ksmbd: add a WRITE_DAC/WRITE_OWNER check to SMB2 SET_INFO SECURITY
ksmbd: fix use-after-free of a deferred file_lock on SMB2_CLOSE then SMB2_CANCEL
smb: server: remove code guarded by nonexistent config option
...
Linus Torvalds [Thu, 18 Jun 2026 16:18:00 +0000 (09:18 -0700)]
Merge tag 'v7.2-rc-part1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Pull smb client updates from Steve French:
- Three cleanup patches
- Fix error return value in smb2_aead_req_alloc
- Three compression fixes
- Update i_blocks after write (fixes various xfstests)
- Fix races in cifsd thread creation
- Fix potential out of bounds read parsing security descriptors
- Witness protocol fix
- Fix umount bug
- Mount fix
- Fix cached directory entries on unlink/rmdir/rename
* tag 'v7.2-rc-part1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
smb: client: Use more common code in SMB2_tcon()
smb: client: Use more common error handling code in smb3_reconfigure()
smb/client: Fix error code in smb2_aead_req_alloc()
smb/client: clean up a type issue in cifs_xattr_get()
smb/client: allow FS_IOC_SETFLAGS to clear compression
smb/client: use writable handle for FS_IOC_SETFLAGS compression
smb/client: always return a value for FS_IOC_GETFLAGS
smb/client: update i_blocks after contiguous writes
smb: client: fix races in cifsd thread creation
cifs: validate full SID length in security descriptors
smb: client: resolve SWN tcon from live registrations
cifs: remove all cifs files before kill super
smb: client: fix conflicting option validation for new mount API
cifs: invalidate cfid on unlink/rename/rmdir
Linus Torvalds [Thu, 18 Jun 2026 16:14:15 +0000 (09:14 -0700)]
Merge tag 'nfsd-7.2' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux
Pull nfsd updates from Chuck Lever:
"Jeff Layton wired up netlink upcalls for the auth.unix.ip and
auth.unix.gid caches in SunRPC and the svc_export and nfsd.fh caches
in NFSD. The new kernel-user API is more extensible and lays the
groundwork for retiring the old pipe interface.
The default NFS r/w block size rises to 4MB on hosts with at least
16GB of RAM, reducing per-RPC overhead on fast networks. Smaller
machines keep their previously computed default, and the value remains
tunable through /proc/fs/nfsd/max_block_size.
Chuck Lever converted the server's RPCSEC GSS Kerberos code to the
kernel's shared crypto/krb5 library. The conversion retires and
removes SunRPC's bespoke implementation of Kerberos v5, but keeps
RPCSEC GSS-API.
Continuing the xdrgen migration that converted the NLMv4 server XDR
layer in v7.1, Chuck Lever converted the NLM version 3 server-side XDR
layer from hand-written C to xdrgen-generated code. As with the NLMv4
conversion in v7.1, the goals are improved memory safety, lower
maintenance burden, and groundwork for generation of Rust code for
this layer instead of C.
Chuck Lever fixed an issue where lingering NFSv4 state pins a mounted
file system after it is unexported. A new netlink-based mechanism can
now release NLM locks and NFSv4 state by client address, by
filesystem, and by export. Now an administrator can quiesce an export
cleanly before unmounting it.
The remaining patches are bug fixes, clean-ups, and minor
optimizations, including a batch of memory-leak and use-after-free
fixes in the ACL, lockd, and TLS handshake paths, many of them
reported by Chris Mason. Sincere thanks to all contributors,
reviewers, testers, and bug reporters who participated in the v7.2
NFSD development cycle"
* tag 'nfsd-7.2' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux: (106 commits)
svcrdma: wake sq waiters when the transport closes
nfsd: reset write verifier on deferred writeback errors
nfsd: avoid leaking pre-allocated openowner on unconfirmed retry race
sunrpc: wait for in-flight TLS handshake callback when cancel loses race
sunrpc: pin svc_xprt across the asynchronous TLS handshake callback
nfsd: fix posix_acl leak on SETACL decode failure
nfsd: fix posix_acl leak and ignored error in nfsd4_create_file
nfsd: check get_user() return when reading princhashlen
nfsd: fix inverted cp_ttl check in async copy reaper
nfsd: fix dead ACL conflict guard in nfsd4_create
NFSD: Fix SECINFO_NO_NAME decode error cleanup
sunrpc: harden rq_procinfo lifecycle to prevent double-free
SUNRPC: Return an error from xdr_buf_to_bvec() on overflow
SUNRPC: Bound-check xdr_buf_to_bvec() stores before writing
nfsd: release layout stid on setlease failure
lockd: Avoid hashing uninitialized bytes in nlm4svc_lookup_file()
lockd: Plug nlm_file refcount leak on cached nlm_do_fopen() failure
lockd: Plug nlm_file leak when nlm_do_fopen() fails
Revert "NFSD: Defer sub-object cleanup in export put callbacks"
Revert "svcrdma: Use contiguous pages for RDMA Read sink buffers"
...
Linus Torvalds [Thu, 18 Jun 2026 15:50:52 +0000 (08:50 -0700)]
Merge tag 'fuse-update-7.2' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse
Pull fuse updates from Miklos Szeredi:
- Fix lots of bugs, most from the late 6.x era, but some going back
to 2.6.x
- Add subsystems (io-uring, passthrough) and respective maintainers
(Bernd, Joanne and Amir)
- Separate transport and fs layers (Miklos)
- Don't block on cat /dev/fuse (Joanne)
- Perform some refactoring in fuse-uring (Joanne)
- Don't use bounce-buffer for READDIR reply in virtio-fs (Matthew Ochs)
- Clean up documentation (Randy)
- Improve tracing (Amir)
- Extend page cache invalidation after DIO (Cheng Ding)
- Invalidate readdir cache on epoch change (Jun Wu)
- Misc cleanups
* tag 'fuse-update-7.2' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse: (81 commits)
fuse-uring: clear ent->fuse_req in commit_fetch error path
fuse-uring: use named constants for io-uring iovec indices
fuse-uring: refactor setting up copy state for payload copying
fuse-uring: use enum types for header copying
fuse-uring: refactor io-uring header copying from ring
fuse-uring: refactor io-uring header copying to ring
fuse-uring: separate next request fetching from sending logic
fuse: invalidate readdir cache on epoch bump
virtio-fs: avoid double-free on failed queue setup
fuse: invalidate page cache after DIO and async DIO writes
fuse: set ff->flock only on success
fuse: clean up interrupt reading
fuse: remove stray newline in fuse_dev_do_read()
fuse: use READ_ONCE in fuse_chan_num_background()
fuse: dax: Move long delayed work on system_dfl_long_wq
fuse: add fuse_request_sent tracepoint
fuse: Add SPDX ID lines to some files
fuse: use QSTR() instead of QSTR_INIT() in fuse_get_dentry
fuse: convert page array allocation to kcalloc()
fuse: use current creds for backing files
...
This quirk is identical to the one for the Yamaha Steinberg UR22, here
applied to a CD player that also uses the Steinberg USB interface.
This quirk is necessary to avoid sporadic "clic" noise when using the DAC
of the player.
Mark Tomlinson [Thu, 30 Apr 2026 02:16:28 +0000 (14:16 +1200)]
PCI: iproc: Restore .map_irq() for the platform bus driver
Commit b64aa11eb2dd ("PCI: Set bridge map_irq and swizzle_irq to default
functions") moved the assignment of default .map_irq() callback to
devm_of_pci_bridge_init() and removed the initialization of
'iproc_pcie::map_irq' in platform bus driver. This led to the callback
getting assigned the NULL pointer for platform bus driver, thereby breaking
the INTx functionality, since 'iproc_pcie::map_irq' overrides the
'pci_host_bridge::map_irq' callback in iproc_pcie_setup().
This issue only affected the iproc platform bus driver as this driver
relies on the default callback for non-PAXC controllers. iproc-brcm driver
was already providing the custom mapping function, so it was unaffected.
Restore the original (and intended) behaviour to use the default map_irq
function by removing the local 'iproc_pcie::map_irq' pointer and directly
assigning the 'pci_host_bridge::map_irq' callback in iproc-bcma driver.
This ensures that the default 'map_irq' callback is used for platform bus
driver and only iproc-brcm driver overrides it with a custom one.
Fixes: b64aa11eb2dd ("PCI: Set bridge map_irq and swizzle_irq to default functions") Signed-off-by: Mark Tomlinson <mark.tomlinson@alliedtelesis.co.nz>
[mani: commit log] Signed-off-by: Manivannan Sadhasivam <mani@kernel.org> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Acked-by: Ray Jui <ray.jui@broadcom.com> Link: https://patch.msgid.link/20260430021628.1343154-1-mark.tomlinson@alliedtelesis.co.nz
Linus Torvalds [Thu, 18 Jun 2026 15:16:21 +0000 (08:16 -0700)]
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Pull rdma updates from Jason Gunthorpe:
"Many AI driven bug fixes, and several big driver API cleanups
- Driver bug fixes and minor cleanups in mlx5, hns, rxe, efa, siw,
rtrs, mana, irdma, mlx4. Commonly error path flows, integer
arithmetic overflows on unsafe data, out of bounds access, and use
after free issues under races.
- Second half of the new udata API for drivers focusing on uAPI
response
- bnxt_re supports more options for QP creation that will allow a dv
path in rdma-core
- Untangle the module dependencies so drivers don't link to
ib_uverbs.ko as was originall intended
- Provide a new way to handle umems with a consistent simplified uAPI
and update several drivers to use it. This brings dmabuf support to
more places and more drivers
- Support for mlx5 rate limit and packet pacing for UD and UC
- A batch of fixes for the new shared FRMR pools infrastructure"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (148 commits)
RDMA/irdma: Replace waitqueue and flag with completion
RDMA/hns: Fix memory leak of bonding resources
RDMA/rtrs-srv: Bound RDMA-Write length to chunk size in rdma_write_sg
docs: infiniband: correct name of option to enable the ib_uverbs module
RDMA/bnxt_re: Reject GET_TOGGLE_MEM when toggle page was not allocated
RDMA/bnxt_re: Fail DBR related page allocation UAPIs if the feature is disabled
RDMA/bnxt_re: Avoid repeated requests to allocate WC pages
RDMA/bnxt_re: Proper rollback if the ioremap fails
RDMA/bnxt_re: Add a max slot check for SQ
RDMA/bnxt_re: Avoid displaying the kernel pointer
RDMA/bnxt_re: Free CQ toggle page after firmware teardown
RDMA/bnxt_re: Free SRQ toggle page after firmware teardown
RDMA/bnxt_re: Initialize dpi variable to zero
ABI: sysfs-class-infiniband: minor cleanup
RDMA/mlx5: Release the HW‑provided UAR index rather than the SW one
RDMA/mlx5: Fix undefined shift of user RQ WQE size
RDMA/mlx5: Remove raw RSS QP restrack tracking
RDMA/mlx5: Remove DCT restrack tracking
RDMA/mlx5: Drop FRMR pool handle on UMR revoke failure
RDMA/core: Add ib_frmr_pool_drop for unrecoverable handles
...
Rosen Penev [Tue, 26 May 2026 04:40:16 +0000 (21:40 -0700)]
PCI: mvebu: Use fixed-width interrupt masks to avoid truncation in 64-bit builds
Use u32-typed BIT and GENMASK helpers for PCIe interrupt register
masks. This keeps inverted masks in the same width as the registers
and avoids truncation warnings on 64-bit compile-test builds.
Fixes below and similar warnings:
drivers/pci/controller/pci-mvebu.c:316:21: error: implicit conversion from 'unsigned long' to 'u32' (aka 'unsigned int') changes value from 18446744069414584320 to 0 [-Werror,-Wconstant-conversion]
mvebu_writel(port, ~PCIE_INT_ALL_MASK, PCIE_INT_UNMASK_OFF);
Linus Torvalds [Thu, 18 Jun 2026 15:12:50 +0000 (08:12 -0700)]
Merge tag 'ata-7.2-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/libata/linux
Pull ata updates from Niklas Cassel:
- Bump required Clang version to 23 (Marco), and add Clang context
analysis annotations (Bart)
- Use the ahci_nr_ports() helper in libahci (me)
- Fail to probe the ahci driver if the BAR size is smaller than the
required size to support CAP.NP (Number of Ports) (liyouhong)
- Move EXPORT_SYMBOL_GPL(ahci_do_softreset) to be just below the
function definition (Bart)
- Make ata_scsi_scan_host() schedule hotplug work on the
system_dfl_long_wq workqueue so that it can benefit from scheduler
task placement (Marco)
- Make ata_scsi_port_error_handler() schedule hotplug work on the
system_dfl_long_wq workqueue, such that the work always uses the same
workqueue (me)
- Use devm_platform_get_and_ioremap_resource() in pata_arasan_cf driver
(Rosen)
- Fix ata_exec_internal() to only release and acquire the EH mutex if
the calling function is the one holding the EH mutex (Bart)
- Use hweight_long() to count the port_map bits (TanZheng)
- Add COMPILE_TEST support for pata_ep93xx driver (Rosen)
- Drop unused assignments from pata_isapnp driver (Uwe)
- Extend existing JMicron PMP quirk to include JMicron JMS562 (Xu)
- Drop unused assignments of pci_device_id driver data (Uwe)
- Use named initializers for pci_device_id arrays (Uwe)
* tag 'ata-7.2-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/libata/linux:
ata: Use named initializers for pci_device_id arrays
ata: Drop unused assignments of pci_device_id driver data
ata: libata-pmp: add JMicron JMS562 quirk
ata: pata_isapnp: Drop unused assignments from pnp_device_id array
ata: pata_ep93xx: add COMPILE_TEST support
ata: pata_ep93xx: use unsigned long for data
ata: pata_ep93xx: avoid asm on non ARM
ata: Annotate functions in the issuing path with __must_hold()
ata: libata: Pass ap parameter directly to functions in the issuing path
ata: libata: Document when host->eh_mutex should be held
ata: libata: Add an argument to ata_eh_reset()
ata: ahci: use hweight_long() to count port_map bits
ata: libata: Fix ata_exec_internal()
ata: pata_arasan_cf: simplify ioremap
ata: libata-eh: queue hotplug work on the system_dfl_long_wq workqueue
ata: libata-scsi: Move long delayed work on system_dfl_long_wq
ata: ahci: Move EXPORT_SYMBOL_GPL(ahci_do_softreset)
ata: ahci: fail probe if BAR too small for claimed ports
ata: libahci: use ahci_nr_ports() helper
Linus Torvalds [Thu, 18 Jun 2026 15:09:57 +0000 (08:09 -0700)]
Merge tag 'for-7.2/io_uring-epoll-20260616' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux
Pull io_uring epoll update from Jens Axboe:
"As discussed a few months ago, this pull request gets rid of allowing
nested epoll notification contexts via io_uring.
Nested contexts have been a source of issues on the epoll side, and
there should not be a need to support them from io_uring. The epoll
io_uring side exists mainly to facilitate a gradual migration from a
notification based epoll setup to an io_uring ditto"
* tag 'for-7.2/io_uring-epoll-20260616' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux:
io_uring/epoll: disallow adding an epoll file to an epoll context
io_uring/epoll: switch to using do_epoll_ctl_file() interface
Alvin Šipraga [Wed, 17 Jun 2026 14:55:08 +0000 (16:55 +0200)]
ASoC: audio-graph-card2: Drop warning for manually selected DAI formats
Sometimes the DAI format must be specified in the audio-graph-card2
device tree, so emitting a warning can be misleading. Revert back to
emitting no warning.
A few examples where automatic format selection might not be applicable:
- For DPCM, where the other side of the DAI link is not apparent, no
proper selection can actually be made. This can lead to disagreeing
formats.
- Due to hardware peculiarities, some ostensibly supported formats might
not work in practice.
In either case, the only correct solution is for the sound card to set
the format
Heiko Carstens [Wed, 17 Jun 2026 14:53:13 +0000 (16:53 +0200)]
s390/idle: Add missing EXPORT_SYMBOL_GPL()
Uwe Kleine-König reported this build breakage caused by a recent commit
which provides arch specific kcpustat_field_idle()/kcpustat_field_iowait()
functions:
ALSA: usb-audio: qcom: clear opened when stream enable fails
On enable, subs->opened is set before the service_interval is validated;
an invalid interval jumps to the response label without clearing it, so
the substream is wedged at -EBUSY until a disable or disconnect.
ALSA: usb-audio: qcom: reject stream disable with no active interface
handle_uaudio_stream_req() resolves an interface index with
info_idx_from_ifnum(), which returns -EINVAL when no interface matches.
The enable branch and the response: cleanup label both guard against a
negative index, but the disable branch does not: it forms
info = &uadev[pcm_card_num].info[info_idx] and dereferences it.
uadev[].info is a pointer allocated only when a stream is first enabled,
so a negative info_idx on the disable path is unsafe in two ways:
- If the card was never enabled, .info is NULL and &info[-EINVAL] is a
wild pointer; reading info->data_ep_pipe faults (kernel oops).
- If the card was enabled at least once (.info allocated) and the
disable names an interface that does not match, &info[-EINVAL] points
before the allocation; info->data_ep_pipe / info->sync_ep_pipe are an
out-of-bounds slab read and, when non-zero, an out-of-bounds 4-byte
write (both pipe fields are cleared to 0). That is memory corruption,
not just a NULL dereference.
The request is reachable from unprivileged local userspace over
AF_QIPCRTR. Reject a disable request with no resolved interface, matching
the guard the enable path already has.
Haowen Tu [Thu, 18 Jun 2026 07:54:31 +0000 (15:54 +0800)]
ALSA: hda/realtek: Add headset mic quirk for Acer S40-54
Acer S40-54 with ALC256 does not restore headset mic detection properly
after S4 resume. After resume, headset plug events may no longer update
the headset mic state, leaving the headset microphone unavailable.
The system uses subsystem ID 1025:161f. Applying the existing
ALC256_FIXUP_ACER_MIC_NO_PRESENCE fixup restores headset mic detection
on this machine.
Maoyi Xie [Thu, 18 Jun 2026 06:03:15 +0000 (14:03 +0800)]
ALSA: caiaq: bound the length in the EP1 input parsers
snd_caiaq_input_read_erp() and snd_caiaq_input_read_io() can be reached
from snd_usb_caiaq_input_dispatch(). They read fixed byte offsets from
the reply buffer without checking the reported length. On a short reply
they decode stale bytes left from a previous, longer report and feed them
to the input layer.
This is not an out-of-bounds access. Every offset is a compile-time
driver constant. The largest is buf[21] in the Maschine ERP case. The
EP1 transfer buffer ep1_in_buf is EP1_BUFSIZE (64) bytes, and the USB
core caps actual_length at 64, so a short reply only reads in-bounds
stale data. Acting on data the device did not send is still wrong, so
bail out per usb_id case when the reply is shorter than the bytes that
case consumes.
read_erp: AK1 needs 2 bytes, Kore needs 16, Maschine needs 22.
read_io: the Kore case needs 5 bytes (buf[4]) and the Traktor Kontrol
X1 case needs 7 (buf[5]/buf[6]). The preceding key bit loop
is already bounded by "i < len * 8" and is left untouched.
snd_caiaq_input_read_analog() and snd_usb_caiaq_maschine_dispatch() are
not changed. Their callers already floor the reply length.
Maoyi Xie [Thu, 18 Jun 2026 06:03:15 +0000 (14:03 +0800)]
ALSA: caiaq: fix out-of-bounds read in the Traktor Kontrol S4 input parser
snd_usb_caiaq_tks4_dispatch() decodes the Traktor Kontrol S4 input
stream in fixed 16-byte (TKS4_MSGBLOCK_SIZE) message blocks. On every
iteration it advances buf and subtracts the block size while looping on
"while (len)".
len is urb->actual_length. That value is supplied by the device and is
not guaranteed to be a multiple of 16. When a final short block leaves
len between 1 and 15, the loop runs once more, reads up to buf[15], and
then does "len -= TKS4_MSGBLOCK_SIZE". As len is unsigned this underflows
to a huge value. The loop then keeps iterating and walking buf far past
the end of the 512-byte ep4_in_buf, reading out of bounds until a bogus
block id happens to be hit.
Iterate only while a full message block is available. This stops the
unsigned underflow and silently drops any trailing partial block, which
carries no complete control value anyway.
The sibling endpoint-4 parsers are not affected. The Traktor Kontrol X1
and Maschine arms in snd_usb_caiaq_ep4_reply_dispatch() floor
urb->actual_length before dispatching.
Ben Dooks [Wed, 17 Jun 2026 10:29:43 +0000 (11:29 +0100)]
ALSA: pcm: fix __le32 cast warning in snd_pcm_set_sync_per_card
In snd_pcm_set_sync_per_card() the le32 value is written to an u32
instead of an __le32 pointer. Fix the following warning by fixing
the type:
sound/soc/soc-pcm.c:2166:9: warning: incorrect type in argument 7 (different base types)
sound/soc/soc-pcm.c:2166:9: expected int
sound/soc/soc-pcm.c:2166:9: got restricted snd_pcm_format_t
i2c: pxa: Use named initializers for the platform_device_id array
Named initializers are better readable and more robust to changes of the
struct definition. This robustness is relevant for a planned change to
struct platform_device_id replacing .driver_data by an anonymous union.
Randy Dunlap [Wed, 17 Jun 2026 16:31:25 +0000 (09:31 -0700)]
slab: recognize @GFP parameter as optional in kernel-doc
Since the @GFP parameter in kmalloc_obj() etc. is now optional, change
the kernel-doc to indicate that it is optional. This avoids kernel-doc
warnings:
WARNING: include/linux/slab.h:1101 Excess function parameter 'GFP' description in 'kmalloc_obj'
WARNING: include/linux/slab.h:1113 Excess function parameter 'GFP' description in 'kmalloc_objs'
WARNING: include/linux/slab.h:1128 Excess function parameter 'GFP' description in 'kmalloc_flex'
Fixes: e19e1b480ac7 ("add default_gfp() helper macro and use it in the new *alloc_obj() helpers") Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Harry Yoo (Oracle) <harry@kernel.org> Link: https://patch.msgid.link/20260617163125.2716279-1-rdunlap@infradead.org Signed-off-by: Vlastimil Babka (SUSE) <vbabka@kernel.org>
mm/slab: replace __GFP_NO_OBJ_EXT with SLAB_ALLOC_NO_RECURSE for sheaves
Finish the switch away from __GFP_NO_OBJ_EXT by replacing it with
SLAB_ALLOC_NO_RECURSE when allocating empty sheaves. Pass alloc_flags to
[__]alloc_empty_sheaf(). Callers that can't be part of a recursive
kmalloc() chain simply pass SLAB_ALLOC_DEFAULT. Use kmalloc_flags()
instead of kzalloc() for allocating the sheaf.
With that we can finalize the removal the __GFP_NO_OBJ_EXT handling from
obj_ext allocations as well, leaving only SLAB_ALLOC_NO_RECURSE in
place.
This leaves __GFP_NO_OBJ_EXT with no users in slab, so stop allowing the
flag in kmalloc_nolock().
mm/slab: remove __GFP_NO_OBJ_EXT usage from alloc_slab_obj_exts()
__GFP_NO_OBJ_EXT has limited scope within the slab allocator itself and
gfp flags are a scarce resource, unlike slab's alloc_flags.
Introduce SLAB_ALLOC_NO_RECURSE alloc flag that has the same intent as
__GFP_NO_OBJ_EXT but a more generic name, meaning that a kmalloc()
family function should not recurse into another kmalloc*() for the
purposes of allocating auxiliary structures (obj_ext arrays or sheaves).
First, replace the __GFP_NO_OBJ_EXT for allocating obj_ext arrays in
alloc_slab_obj_exts(). Make use of the newly added kmalloc_flags()
function, where we can pass alloc_flags with SLAB_ALLOC_NO_RECURSE
added. This will also pass through SLAB_ALLOC_NOLOCK so we don't need
to special case kmalloc_nolock() anymore.
Note that until now the kmalloc_nolock() ignored the incoming gfp flags
and hardcoded __GFP_ZERO | __GFP_NO_OBJ_EXT. But it's correct to pass on
the incoming gfp flags (only augmented with __GFP_ZERO), because if
alloc_flags contain SLAB_ALLOC_NOLOCK, the incoming gfp flags have to
be also compatible with it. However, we might have added __GFP_THISNODE
for opportunistic slab allocation, as pointed out by Hao Li, and
__GFP_COMP by allocate_slab() as pointed out by Shengming Hu. Solve this
by adding both flags to OBJCGS_CLEAR_MASK as it makes sense to strip
them anyway for non-kmalloc_nolock() allocations of sheaves or obj_ext
arrays as well.
To avoid recursion of sheaf -> obj_ext -> sheaf -> ... allocations at
this patch, until the next patch converts sheaves to
SLAB_ALLOC_NO_RECURSE, use both gfp and alloc_flags for obj_ext. The
next patch will remove the gfp part.
Zhan Xusheng [Wed, 17 Jun 2026 07:43:34 +0000 (15:43 +0800)]
erofs: add folio order to trace_erofs_read_folio
erofs supports large folios for reads, but the actual folio order
instantiated in the page cache may be lower due to allocation
constraints such as memory fragmentation.
trace_erofs_read_folio already receives the folio being read but
currently records only its index. Add folio_order() to the tracepoint
so that users can observe the realized folio order and verify the
effectiveness of large folio reads.
Also drop the uptodate field: read_folio() is only called for a
non-uptodate folio, so it is always 0.
Runyu Xiao [Wed, 17 Jun 2026 15:40:35 +0000 (23:40 +0800)]
gpio: eic-sprd: use raw_spinlock_t in the irq startup path
sprd_eic_irq_unmask() enables the GPIO IRQ and then updates controller
state through sprd_eic_update(), which takes sprd_eic->lock with
spin_lock_irqsave(). The callback can be reached from irq_startup()
while setting up a requested IRQ. That path is not sleepable, but on
PREEMPT_RT a regular spinlock_t becomes a sleeping lock.
This issue was found by our static analysis tool and then manually
reviewed against the current tree.
The grounded PoC kept the request_threaded_irq() -> __setup_irq() ->
irq_startup() -> sprd_eic_irq_unmask() -> sprd_eic_update() carrier and
used the original spin_lock_irqsave(&sprd_eic->lock) edge. Lockdep
reported:
BUG: sleeping function called from invalid context
hardirqs last disabled at ... __setup_irq.constprop.0 ... [vuln_msv]
sprd_rt_spin_lock_irqsave+0x1c/0x30 [vuln_msv]
sprd_eic_update.constprop.0+0x48/0x90 [vuln_msv]
sprd_eic_irq_unmask.constprop.0+0x35/0x50 [vuln_msv]
__setup_irq.constprop.0+0xd/0x30 [vuln_msv]
Convert the Spreadtrum EIC controller lock to raw_spinlock_t. The
locked section only serializes MMIO register updates and does not contain
sleepable operations, so keeping it non-sleeping is appropriate for the
irqchip callbacks.
Runyu Xiao [Wed, 17 Jun 2026 15:40:34 +0000 (23:40 +0800)]
gpio: sch: use raw_spinlock_t in the irq startup path
sch_irq_unmask() enables the GPIO IRQ and then updates the controller
state through sch_irq_mask_unmask(), which takes sch->lock with
spin_lock_irqsave(). The callback can be reached from irq_startup()
while setting up a requested IRQ. That path is not sleepable, but on
PREEMPT_RT a regular spinlock_t becomes a sleeping lock.
This issue was found by our static analysis tool and then manually
reviewed against the current tree.
The grounded PoC kept the request_threaded_irq() -> __setup_irq() ->
irq_startup() -> sch_irq_unmask() -> sch_irq_mask_unmask() carrier and
used the original spin_lock_irqsave(&sch->lock) edge. Lockdep reported:
BUG: sleeping function called from invalid context
hardirqs last disabled at ... __setup_irq.constprop.0 ... [vuln_msv]
sch_rt_spin_lock_irqsave+0x1c/0x30 [vuln_msv]
sch_irq_mask_unmask.constprop.0+0x31/0x70 [vuln_msv]
__setup_irq.constprop.0+0xd/0x30 [vuln_msv]
Convert the SCH controller lock to raw_spinlock_t. The same lock is
also used by the GPIO direction and value callbacks, but those critical
sections only update MMIO-backed GPIO registers and do not contain
sleepable operations. Keeping this register lock non-sleeping is
therefore appropriate for the irqchip callbacks and does not change the
GPIO-side locking contract.
Fixes: 7a81638485c1 ("gpio: sch: Add edge event support") Cc: stable@vger.kernel.org Signed-off-by: Runyu Xiao <runyu.xiao@seu.edu.cn> Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com> Link: https://patch.msgid.link/20260617154035.1199948-2-runyu.xiao@seu.edu.cn Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Xin Long [Mon, 15 Jun 2026 19:36:30 +0000 (15:36 -0400)]
sctp: hold socket lock when dumping endpoints in sctp_diag
SCTP_DIAG endpoint dumping was traversing endpoint address lists without
holding lock_sock(), while those lists could change concurrently via
socket operations (e.g., bindx changes). This creates a race where
nla_reserve() counts addresses under RCU protection, but the subsequent
copy may see fewer entries, potentially leaking uninitialized memory to
userspace.
Fix this by:
- Taking a reference on each endpoint during hash traversal
- Moving socket operations (lock_sock()) outside read_lock_bh()
- Serializing address list access during dump
- Reworking sctp_for_each_endpoint() to support restart-based traversal
with (net, pos) tracking
Also:
- Add WARN_ON_ONCE() for inconsistent address counts
- Fix idiag_states filtering for LISTEN vs association cases
- Skip dumping endpoints being freed (ep->base.dead)
- Move dump position tracking into iterator, removing cb->args[4] and
its comment for sctp_ep_dump().,
- Update the comment for cb->args[4] and remove the comment for unused
cb->args[5] for sctp_sock_dump().
Note: traversal is restart-based and may re-scan buckets multiple times,
but this is acceptable due to small bucket sizes and required to support
sleeping-safe callbacks.
This issue was reported by Nico Yip (@_cyeaa_) working with TrendAI Zero
Day Initiative.
Jakub Kicinski [Tue, 16 Jun 2026 00:30:46 +0000 (17:30 -0700)]
net: psample: fix info leak in PSAMPLE_ATTR_DATA
psample open codes nla_put() presumably to avoid wiping
the data with 0s just to override it with packet data.
This open coding is missing clearing the pad, however,
each netlink attr is padded to 4B and data_len may
not be divisible by 4B.
Fixes: 6ae0a6286171 ("net: Introduce psample, a new genetlink channel for packet sampling") Reported-by: Weiming Shi <bestswngs@gmail.com> Reviewed-by: Jiri Pirko <jiri@nvidia.com> Link: https://patch.msgid.link/20260616003046.1099490-1-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Li RongQing [Mon, 15 Jun 2026 14:04:06 +0000 (22:04 +0800)]
net/mlx5: Remove broken and unused mlx5_query_mtppse()
mlx5_query_mtppse() reads the Event Trigger Pin (MTPPSE) register but
reads the returned arm and mode values from the input buffer 'in'
instead of the output buffer 'out', so it always returns the values
that were written rather than the actual hardware state, making the
query useless.
The function has no in-tree callers. Remove it rather than fix it.
Wyatt Feng [Mon, 15 Jun 2026 10:31:18 +0000 (18:31 +0800)]
net: ipv4: bound TCP reordering sysctl writes and MTU probe sizes
Reject invalid `net.ipv4.tcp_reordering` values before they reach TCP
socket state. The sysctl is stored as an `int` but copied into the
`u32` `tp->reordering` field for new sockets, so negative writes wrap
to large values.
With `tcp_mtu_probing=2`, the wrapped value can overflow the
`tcp_mtu_probe()` size calculation and drive the MTU probing path into
an out-of-bounds read. Route `tcp_reordering` writes through
`proc_dointvec_minmax()` and require it to be at least 1. Also require
`tcp_max_reordering` to be at least 1 so the configured maximum cannot
become negative either.
When registering the table for a non-init network namespace, relocate
`extra2` pointers that refer into `init_net.ipv4` so the
`tcp_reordering` upper bound follows that namespace's
`tcp_max_reordering`.
Harden `tcp_mtu_probe()` itself by computing `size_needed` as `u64`.
This keeps the send queue and window checks from being bypassed through
signed integer overflow.
Fixes: 91cc17c0e5e5 ("[TCP]: MTUprobe: receiver window & data available checks fixed") Cc: stable@vger.kernel.org Reported-by: Yuan Tan <yuantan098@gmail.com> Reported-by: Zhengchuan Liang <zcliangcn@gmail.com> Reported-by: Xin Liu <bird@lzu.edu.cn> Suggested-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Wyatt Feng <bronzed_45_vested@icloud.com> Signed-off-by: Ren Wei <n05ec@lzu.edu.cn> Reviewed-by: Eric Dumazet <edumazet@google.com> Link: https://patch.msgid.link/1a5b7e1ef4d70fbad8c8ee0b82d8405f3c964a3d.1781395200.git.bronzed_45_vested@icloud.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pengpeng Hou [Mon, 15 Jun 2026 07:00:31 +0000 (15:00 +0800)]
net: ehea: unwind probe_port sysfs file on failure
ehea_create_device_sysfs() creates probe_port and then remove_port. If
the second device_create_file() fails, the helper returns the error but
leaves probe_port installed even though probe treats the sysfs setup as
failed.
Remove probe_port on the remove_port creation failure path so the helper
leaves no partial sysfs state behind.
Ratheesh Kannoth [Mon, 15 Jun 2026 03:31:57 +0000 (09:01 +0530)]
octeontx2-af: npc: Log successful MCAM drop-on-non-hit install at debug level
npc_install_mcam_drop_rule() used dev_err() after a successful
rvu_mbox_handler_npc_mcam_write_entry() call, so normal installs appeared
as errors in dmesg. Use dev_dbg() for the success path and keep dev_err()
for real failures.
Fixes: 3571fe07a090 ("octeontx2-af: Drop rules for NPC MCAM") Signed-off-by: Ratheesh Kannoth <rkannoth@marvell.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20260615033157.535237-1-rkannoth@marvell.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Ratheesh Kannoth [Mon, 15 Jun 2026 03:07:04 +0000 (08:37 +0530)]
octeontx2-pf: Fix leak of SQ timestamp buffer on teardown
The send-queue timestamp ring is allocated with qmem_alloc() when
timestamping is used, but otx2_free_sq_res() never freed sq->timestamps,
leaking that memory across ifdown and device removal. Add the missing
qmem_free() alongside the other SQ companion buffers.
Fixes: c9c12d339d93 ("octeontx2-pf: Add support for PTP clock") Cc: Aleksey Makarov <amakarov@marvell.com> Signed-off-by: Ratheesh Kannoth <rkannoth@marvell.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20260615030704.504536-1-rkannoth@marvell.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Tianchen Ding [Mon, 1 Jun 2026 02:32:51 +0000 (10:32 +0800)]
selftests/ftrace: Fix trace_marker_raw test on 64K page kernels
On ARM64 kernels with 64K pages, the trace_marker_raw test fails because
bash's printf builtin uses stdio buffering which splits output into
multiple small write() calls to the tracefs file. Since each individual
write is within TRACE_MARKER_MAX_SIZE (4096), they all succeed, causing
the "too big" write test to incorrectly pass.
Fix by writing through dd with iflag=fullblock to guarantee a single
atomic write() syscall to trace_marker_raw.
Link: https://lore.kernel.org/r/20260601023251.1916483-1-dtcccc@linux.alibaba.com Fixes: 37f46601383a ("selftests/tracing: Add basic test for trace_marker_raw file") Signed-off-by: Tianchen Ding <dtcccc@linux.alibaba.com> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Maoyi Xie [Mon, 15 Jun 2026 17:17:36 +0000 (01:17 +0800)]
netdev-genl: report NAPI thread PID in the caller's pid namespace
netdev_nl_napi_fill_one() reports the NAPI kthread PID in NETDEV_A_NAPI_PID
using task_pid_nr(), which returns the PID in the initial pid namespace.
NETDEV_CMD_NAPI_GET does not have GENL_ADMIN_PERM and the netdev genl family
is netnsok, so a caller in a child pid namespace can issue it. That caller
then sees the kthread's global PID, even though the kthread is not visible
in its pid namespace, where the value should be 0.
Translate the PID through the caller's pid namespace, the same way commit 3799c2570982 ("io_uring/fdinfo: translate SqThread PID through caller's
pid_ns") did for the io_uring SQPOLL thread. The doit and dumpit paths both
run synchronously in the caller's context, so task_active_pid_ns(current) is
the caller's pid namespace.
Fixes: db4704f4e4df ("netdev-genl: Add PID for the NAPI thread") Cc: stable@vger.kernel.org Signed-off-by: Maoyi Xie <maoyixie.tju@gmail.com> Reviewed-by: Joe Damato <joe@dama.to> Reviewed-by: Samiullah Khawaja <skhawaja@google.com> Link: https://patch.msgid.link/20260615171736.1709318-1-maoyixie.tju@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
net: ethernet: mtk_eth_soc: fix supported_interface set after phylink_create
Everything configured in phylink_config it's assumed to be set before
calling phylink_create() to permit correct parsing of all the different
modes and capabilities.
Commit 51cf06ddafc9 ("net: ethernet: mtk_eth_soc: add support for MT7988
internal 2.5G PHY") while introducing support for 2.5G phy for MT7988,
probably due to an auto-rebase, placed the configuration of the INTERNAL
interface mode for the supported_interfaces for phylink_config right after
phylink_create() introducing a possible problem with supported interfaces
parsing.
While this doesn't currently create any problem/bug, move setting this bit
before phylink_create() to prevent any possible regression in future code
change in phylink core.
Fixes: 51cf06ddafc9 ("net: ethernet: mtk_eth_soc: add support for MT7988 internal 2.5G PHY") Signed-off-by: Christian Marangi <ansuelsmth@gmail.com> Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Reviewed-by: Daniel Golle <daniel@makrotopia.org> Link: https://patch.msgid.link/20260615151106.15438-1-ansuelsmth@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
====================
net: require CAP_NET_ADMIN in the device netns for tunnel changelink
A tunnel changelink() operates on at most two netns, dev_net(dev) and
the tunnel link netns t->net. They differ once the device is created in
or moved to a netns other than the one the request runs in. The rtnl
changelink path checks CAP_NET_ADMIN only against dev_net(dev), so a
caller privileged there but not in the link netns can rewrite a tunnel
that lives in the link netns. Commit 8b484efd5cb4 ("ip6: vti: Use
ip6_tnl.net in vti6_siocdevprivate().") added the same check on the
ioctl path. This series adds it on the RTM_NEWLINK path.
Each changelink is gated at the top of the op, before any attribute is
parsed, because the per-type parsers can update live tunnel fields first
(for example ipgre_netlink_parms() sets t->collect_md). The check is
skipped when the link netns is dev_net(dev), where the rtnl path already
checked the cap.
The check goes through a new helper, rtnl_dev_link_net_capable(), added
in patch 1 next to rtnl_get_net_ns_capable() in net/core/rtnetlink.c.
Tested on net/main. For every tunnel type in the series a migrated
fake-root changelink is rejected with EPERM. For vti6 SIOCGETTUNNEL
confirms the link netns hash is left unchanged. Legit non-migrated
changelinks still succeed.
Maoyi Xie [Fri, 12 Jun 2026 08:59:41 +0000 (16:59 +0800)]
xfrm: xfrm_interface: require CAP_NET_ADMIN in the device netns for changelink
xfrmi_changelink() operates on at most two netns, dev_net(dev) and the
interface link netns xi->net. They differ once the device is created in
or moved to a netns other than the one the request runs in. The rtnl
changelink path checks CAP_NET_ADMIN only against dev_net(dev), so a
caller privileged there but not in xi->net can rewrite an interface that
lives in xi->net.
Gate xfrmi_changelink() on rtnl_dev_link_net_capable() at its top,
before any attribute is parsed.
Maoyi Xie [Fri, 12 Jun 2026 08:59:40 +0000 (16:59 +0800)]
net: ip6_vti: require CAP_NET_ADMIN in the device netns for changelink
vti6_changelink() operates on at most two netns, dev_net(dev) and the
tunnel link netns t->net. They differ once the device is created in or
moved to a netns other than the one the request runs in. The rtnl
changelink path checks CAP_NET_ADMIN only against dev_net(dev), so a
caller privileged there but not in t->net can rewrite a tunnel that
lives in t->net.
Gate vti6_changelink() on rtnl_dev_link_net_capable() at its top,
before any attribute is parsed.
Maoyi Xie [Fri, 12 Jun 2026 08:59:39 +0000 (16:59 +0800)]
net: ip6_gre: require CAP_NET_ADMIN in the device netns for changelink
ip6gre_changelink() and ip6erspan_changelink() operate on at most two
netns, dev_net(dev) and the tunnel link netns t->net. They differ once
the device is created in or moved to a netns other than the one the
request runs in. The rtnl changelink path checks CAP_NET_ADMIN only
against dev_net(dev), so a caller privileged there but not in t->net can
rewrite a tunnel that lives in t->net.
Gate both ops on rtnl_dev_link_net_capable() at their top, before any
attribute is parsed.
Maoyi Xie [Fri, 12 Jun 2026 08:59:38 +0000 (16:59 +0800)]
net: ip6_tunnel: require CAP_NET_ADMIN in the device netns for changelink
ip6_tnl_changelink() operates on at most two netns, dev_net(dev) and the
tunnel link netns t->net. They differ once the device is created in or
moved to a netns other than the one the request runs in. The rtnl
changelink path checks CAP_NET_ADMIN only against dev_net(dev), so a
caller privileged there but not in t->net can rewrite a tunnel that
lives in t->net.
Gate ip6_tnl_changelink() on rtnl_dev_link_net_capable() at its top,
before any attribute is parsed.
Maoyi Xie [Fri, 12 Jun 2026 08:59:37 +0000 (16:59 +0800)]
net: ip_vti: require CAP_NET_ADMIN in the device netns for changelink
vti_changelink() operates on at most two netns, dev_net(dev) and the
tunnel link netns t->net. They differ once the device is created in or
moved to a netns other than the one the request runs in. The rtnl
changelink path checks CAP_NET_ADMIN only against dev_net(dev), so a
caller privileged there but not in t->net can rewrite a tunnel that
lives in t->net.
Gate vti_changelink() on rtnl_dev_link_net_capable() at its top,
before any attribute is parsed.
Maoyi Xie [Fri, 12 Jun 2026 08:59:36 +0000 (16:59 +0800)]
net: ipip: require CAP_NET_ADMIN in the device netns for changelink
ipip_changelink() operates on at most two netns, dev_net(dev) and the
tunnel link netns t->net. They differ once the device is created in or
moved to a netns other than the one the request runs in. The rtnl
changelink path checks CAP_NET_ADMIN only against dev_net(dev), so a
caller privileged there but not in t->net can rewrite a tunnel that
lives in t->net.
Gate ipip_changelink() on rtnl_dev_link_net_capable() at its top,
before any attribute is parsed.
Maoyi Xie [Fri, 12 Jun 2026 08:59:35 +0000 (16:59 +0800)]
net: ip_gre: require CAP_NET_ADMIN in the device netns for changelink
A tunnel changelink() operates on at most two netns, dev_net(dev) and
the tunnel link netns t->net. They differ once the device is created in
or moved to a netns other than the one the request runs in. The rtnl
changelink path checks CAP_NET_ADMIN only against dev_net(dev), so a
caller privileged there but not in t->net can rewrite a tunnel that
lives in t->net.
Add rtnl_dev_link_net_capable() next to rtnl_get_net_ns_capable() in
net/core/rtnetlink.c. It requires CAP_NET_ADMIN in the link netns and is
skipped when the link netns is dev_net(dev), where the rtnl path already
checked it. The other patches in this series use the same helper.
Gate ipgre_changelink() and erspan_changelink() with it, at the top of
the op before any attribute is parsed, because the parsers update live
tunnel fields first. ipgre_netlink_parms() sets t->collect_md before
ip_tunnel_changelink() runs.
Commit 8b484efd5cb4 ("ip6: vti: Use ip6_tnl.net in
vti6_siocdevprivate().") added the same check on the ioctl path. This
adds it on RTM_NEWLINK.
Robert Marko [Thu, 11 Jun 2026 10:24:49 +0000 (12:24 +0200)]
net: pse-pd: set user byte command SUB2 field
The Set User Byte to Save command has three subject bytes.
The PD692x0 protocol guides defines SUB2 with value 0x4e, while SUB1
carries the NVM user byte.
Template only initialized SUB and SUB1.
Fill SUB2 explicitly so the command matches the documented layout.
Aleksandr Nogikh [Fri, 12 Jun 2026 11:50:20 +0000 (11:50 +0000)]
ocfs2: fix circular locking dependency in ocfs2_dio_end_io_write
A circular locking dependency involves INODE_ALLOC_SYSTEM_INODE,
EXTENT_ALLOC_SYSTEM_INODE, and ORPHAN_DIR_SYSTEM_INODE.
1. ocfs2_mknod() acquires INODE_ALLOC then EXTENT_ALLOC.
2. ocfs2_dio_end_io_write() acquires EXTENT_ALLOC for unwritten
extents, then ORPHAN_DIR via ocfs2_del_inode_from_orphan() while still
holding EXTENT_ALLOC.
3. ocfs2_wipe_inode() acquires ORPHAN_DIR then INODE_ALLOC via
ocfs2_remove_inode.
Break the cycle in ocfs2_dio_end_io_write() by freeing the allocation
contexts (releasing EXTENT_ALLOC) before acquiring ORPHAN_DIR.
WARNING: possible circular locking dependency detected
------------------------------------------------------
is trying to acquire lock: ffff8881e78b33a0
(&ocfs2_sysfile_lock_key[INODE_ALLOC_SYSTEM_INODE]){+.+.}-{4:4}, at:
ocfs2_evict_inode+0x1539/0x43b0 fs/ocfs2/inode.c:1299
but task is already holding lock: ffff8881e78b4fa0
(&ocfs2_sysfile_lock_key[ORPHAN_DIR_SYSTEM_INODE]){+.+.}-{4:4}, at:
ocfs2_evict_inode+0xe97/0x43b0 fs/ocfs2/inode.c:1299
the existing dependency chain (in reverse order) is:
Ian Bridges [Thu, 11 Jun 2026 14:46:38 +0000 (09:46 -0500)]
ocfs2: fix NULL h_transaction deref in ocfs2_assure_trans_credits
[BUG]
A direct write over unwritten extents can panic the kernel in
ocfs2_assure_trans_credits() when the journal aborts during DIO
completion. The crash is a general protection fault from a NULL pointer
dereference.
[CAUSE]
ocfs2_dio_end_io_write() loops over a direct write's unwritten extents,
marking each written under a single journal handle. If the journal
aborts (for example after an I/O error) while the extent tree is being
updated, the handle is left aborted with its transaction pointer
cleared. The extent merge treats that failure as not critical and
reports success, so the loop keeps using the handle.
ocfs2_assure_trans_credits() reads the handle's remaining credits
without first checking whether the handle is aborted, and that read
dereferences the cleared transaction pointer.
[FIX]
A journal abort is recorded in the handle itself, so callers are
expected to test the handle rather than rely on a returned error.
Make ocfs2_assure_trans_credits() do that, as the other ocfs2 journal
helpers already do, and return -EROFS when the handle is aborted.
Link: https://lore.kernel.org/airKTsM1fRVN-Wj7@dev Fixes: be346c1a6eeb ("ocfs2: fix DIO failure due to insufficient transaction credits") Signed-off-by: Ian Bridges <icb@fastmail.org> Reported-by: syzbot+e9c15ff790cea6a0cfae@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=e9c15ff790cea6a0cfae Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com> Cc: Mark Fasheh <mark@fasheh.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Junxiao Bi <junxiao.bi@oracle.com> Cc: Changwei Ge <gechangwei@live.cn> Cc: Jun Piao <piaojun@huawei.com> Cc: Heming Zhao <heming.zhao@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The interval tree runtime test accepts module parameters that are later
used as divisors while generating randomized intervals and while reporting
average timings. For example, max_endpoint=1 makes the generated interval
end value zero and the next modulo operation divides by that zero value.
Reject non-positive counts and require max_endpoint to provide at least
one non-zero generated endpoint before the test allocates state or starts
the benchmark.
Kyle Zeng [Thu, 11 Jun 2026 21:35:10 +0000 (14:35 -0700)]
ocfs2: avoid moving extents to occupied clusters
For non-auto OCFS2_IOC_MOVE_EXT operations, userspace supplies a physical
me_goal. ocfs2_move_extent() initializes new_phys_cpos from that goal and
expects ocfs2_probe_alloc_group() to replace it with a free run in the
target block group.
The probe currently leaves *phys_cpos unchanged if the scan reaches the
end of the group without finding a free run. An occupied goal at the last
bit can therefore survive the probe and be passed to
__ocfs2_move_extent(), which copies file data into a cluster still owned
by another inode before the bitmap is updated.
When the probe does find a free run, it also subtracts move_len from the
ending bit. The start of an N-bit run ending at i is i - N + 1, so the
current calculation can report the bit immediately before the free run.
Clear *phys_cpos before scanning and use the correct free-run start.
Callers already treat a zero result as -ENOSPC, so failed probes no longer
continue with an occupied caller-controlled goal.
Link: https://lore.kernel.org/20260611213510.16956-1-kylebot@openai.com Fixes: e6b5859cccfa ("Ocfs2/move_extents: helper to probe a proper region to move in an alloc group.") Fixes: 236b9254f8d1 ("ocfs2: fix non-auto defrag path not working issue") Assisted-by: Codex:gpt-5.5 Signed-off-by: Kyle Zeng <kylebot@openai.com> Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com> Cc: Mark Fasheh <mark@fasheh.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Junxiao Bi <junxiao.bi@oracle.com> Cc: Changwei Ge <gechangwei@live.cn> Cc: Jun Piao <piaojun@huawei.com> Cc: Heming Zhao <heming.zhao@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ian Bridges [Thu, 11 Jun 2026 00:23:11 +0000 (19:23 -0500)]
ocfs2: fix UBSAN array-index-out-of-bounds in ocfs2_sum_rightmost_rec
[BUG]
On-disk corruption setting l_next_free_rec to 0 in an inode's embedded
extent list triggers a UBSAN panic on the next write to that file.
[CAUSE]
ocfs2_sum_rightmost_rec() computes
i = le16_to_cpu(el->l_next_free_rec) - 1
and accesses el->l_recs[i] without validating i. When l_next_free_rec
is 0, i becomes -1; when l_next_free_rec exceeds l_count, i falls
past the end of the array. Either case violates the
__counted_by_le(l_count) annotation on l_recs[] and triggers UBSAN.
[FIX]
Validate the inode's embedded extent list when the inode is read, in
ocfs2_validate_inode_block(): l_count must be non-zero and no larger
than the inode block can hold, and l_next_free_rec must not exceed
l_count. A corrupt list is rejected at read time, before the b-tree
code can index l_recs[] out of bounds.
Link: https://lore.kernel.org/ain_780qc0P4ypNd@dev Signed-off-by: Ian Bridges <icb@fastmail.org> Reported-by: syzbot+be16e33db01e6644db7a@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=be16e33db01e6644db7a Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com> Cc: Mark Fasheh <mark@fasheh.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Junxiao Bi <junxiao.bi@oracle.com> Cc: Changwei Ge <gechangwei@live.cn> Cc: Jun Piao <piaojun@huawei.com> Cc: Heming Zhao <heming.zhao@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Samuel Moelius [Fri, 5 Jun 2026 15:52:15 +0000 (15:52 +0000)]
fat: reject BPB volumes whose data area starts beyond total sectors
fat_fill_super() subtracts sbi->data_start from the BPB total sector count
before computing the number of clusters. A malformed image can declare a
total sector count smaller than data_start, causing the subtraction to
underflow and the mount code to derive a plausible cluster count from the
FAT length instead.
Reject such images before the subtraction. In QEMU, a crafted FAT image
with total_sectors=2 and data_start=3 mounted successfully before the fix
and reading a file returned bytes stored past the BPB-declared end of the
volume. With this change, the same image is rejected during mount.
Assisted-by: Codex:gpt-5.5-cyber-preview Link: https://lore.kernel.org/20260605155216.2126545-1-sam.moelius@trailofbits.com Signed-off-by: Samuel Moelius <sam.moelius@trailofbits.com> Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Cc: Christian Brauner <brauner@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Guixin Liu [Mon, 8 Jun 2026 05:39:34 +0000 (13:39 +0800)]
selftests/uevent: increase __UEVENT_BUFFER_SIZE to avoid ENOBUFS on busy systems
The kselftests case uevent.uevent_filtering fails reproducibly on busy
systems (e.g. Intel EMR / AMD servers) with:
No buffer space available - Failed to receive uevent
The listener binds the NETLINK_KOBJECT_UEVENT socket to all 32 multicast
groups (nl_groups = -1) but only sets SO_RCVBUF to 4 KiB
(__UEVENT_BUFFER_SIZE = 2048 * 2). On hosts with many devices, the kernel
and userspace daemons (udev/systemd) constantly emit uevents on multiple
groups, plus the test itself triggers 10 add events in a row. The 4 KiB
receive buffer overflows before the listener can drain it, recvmsg()
returns -ENOBUFS, and the test bails out as failure.
Increase __UEVENT_BUFFER_SIZE to 1 MiB so the receive buffer is large
enough to absorb the burst of uevents on busy systems. After this change
the test passes consistently across dozens of runs on Intel EMR and AMD
platforms.
Link: https://lore.kernel.org/20260608053934.4059533-1-kanie@linux.alibaba.com Signed-off-by: Guixin Liu <kanie@linux.alibaba.com> Cc: Christian Brauner <brauner@kernel.org> Cc: Shuah Khan <shuah@kernel.org> Cc: Wei Yang <richard.weiyang@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Timur Kristóf [Wed, 13 May 2026 17:08:49 +0000 (19:08 +0200)]
drm/amdgpu: Use system unbound workqueue for soft IH ring
Allow the kernel to dispatch the soft IH work on other CPUs.
Otherwise it can happen that the soft IH ring fills up
before it actually starts processing anything, which
can easily happen with retry page faults, in which case
the CP repeatedly spams the CPU with a lot of interrupts.
This significantly improves retry page fault handling on
GPUs that don't have the filter CAM and must rely on
software based filtering.
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com> Signed-off-by: Timur Kristóf <timur.kristof@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 3cdff3c8b93c2834977224d9c2b201fc334dd184)
drm/amdkfd: Use exclusive bounds for SVM split alignment checks
SVM ranges use inclusive page indices: prange->last is the last page in
the range. The split-remap logic introduced by commit 448ee45353ef
("drm/amdkfd: Use huge page size to check split svm range alignment")
uses ALIGN_DOWN(prange->last, 512) to determine whether the original
range can contain a 2MB huge-page mapping.
That aligns the last page itself down. Thus a range ending one page
before the next 2MB boundary is classified as if the final 2MB block did
not exist. When such a range is split inside that final block, the
split head or tail can be left off the remap list even though it was
derived from an original range that may have PMD mappings.
Use prange->last + 1 as the exclusive upper bound when computing the
original range's last 2MB-aligned boundary. Then use the actual split
boundary for the head and tail alignment checks: tail->start for a tail
split, and new_start for a head split. new_start is equivalent to
head->last + 1 and directly names the exclusive end of the split head.
Using head->last for the head-side check can both remap a head that ends
exactly one page before a 2MB boundary and miss a head whose split
boundary is one page after such a boundary. Philip Yang pointed out in
the review of the original change that this condition should use
head->last + 1 or new_start.
Xiaogang Chen identified the inclusive-last cause and posted the
candidate fix in the regression thread. With the culprit change active
and the local revert not applied, the unchanged C/HSA reproducer
completes 10/10 runs with this change on an RX 7600 XT.
Jiqian Chen [Thu, 4 Jun 2026 10:30:23 +0000 (18:30 +0800)]
drm/amdgpu/gfx9: Fix Ring and IB test fail after mode2
For Renior APU with gfx9, in some test scenarios with disabling
ring_reset, like accessing an unmapped invalid address, it can
trigger a gpu job timeout event, then driver uses Mode2 reset
to reset GPU, but after Mode2 compute Ring test and IB test fail
randomly. It because the HQDs of MECs are always active before or
after Mode2, that causes MECs use stale HQDs when MECs are unhalted
before driver restore MQDs, and causes CPC and CPF are still stuck
after Mode2, then causes compute Ring and IB tests fail.
So, add sequences to deactivate HQDs of MECs in suspend IP function
of the resetting process.
v2: Move all sequences into a new function gfx_v9_0_cp_mode2_clear_state (Ray Huang)
To check reset Mode2 method in the if condition (Ray Huang)
v3: Move all sequences before Mode2 instead of after Mode2 (Timur Kristóf)
v4: Call amdgpu_gfx_rlc_enter/exit_safe_mode int the begin and end of
gfx_v9_0_deactivate_kcq_hqd (Alex Deucher)
Signed-off-by: Jiqian Chen <Jiqian.Chen@amd.com> Reviewed-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Timur Kristóf <timur.kristof@gmail.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit c3988a7ad4799514447294f04f063b422e0551df) Cc: stable@vger.kernel.org
Timur Kristóf [Mon, 25 May 2026 11:33:20 +0000 (13:33 +0200)]
drm/amdgpu/uvd: Fix forcing MSG, FB BOs into VCPU segment when it isn't at 0 (v2)
UVD 4.x and older can only access MSG, FEEDBACK buffers from a
specific 256M VRAM segment that the VCPU BO is also located in.
We already modify all placements of the given BO to ensure
the BO is placed within this segment.
Previously, it always assumed that the VCPU segment is
the first 256M of VRAM, even though under some conditions
the VCPU BO could be allocated outside this segment,
which made UVD non-functional as the BOs were
not inside the same segment as the UVD VCPU BO.
Solve that by using the segment where the VCPU BO actually is.
This fixes an issue with UVD failing to initialize on SI/CIK
when resizable BAR is enabled and the VCPU BO is allocated
in a different segment.
v2:
- For other BOs, keep using the same UVD segment as before.
Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/3851 Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Timur Kristóf <timur.kristof@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit cbfd4d3fc2061a1ec8e9d36e65973ac3e813358a) Cc: stable@vger.kernel.org
Timur Kristóf [Mon, 25 May 2026 11:33:19 +0000 (13:33 +0200)]
drm/amdgpu/uvd: Place VCPU BO only in VRAM for UVD 4.x and older
These UVD versions don't fully support GPUVM and are only
validated to work when their VCPU BO is placed in VRAM.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 01b8dfc0660db5d6cdd62c22dc20f774a26ce853) Cc: stable@vger.kernel.org
Timur Kristóf [Mon, 25 May 2026 11:33:18 +0000 (13:33 +0200)]
drm/amdgpu: Fix amdgpu_bo_move() when old_mem and new_mem are both GTT
The UVD code relies on GTT to GTT moves in order to ensure
that its BOs don't cross 256M segments.
Fixes: bfe5e585b44f ("drm/ttm: move last binding into the drivers.") Signed-off-by: Timur Kristóf <timur.kristof@gmail.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 21fd45e5e2628d00b478590bcc3d14d3de5d45b6) Cc: stable@vger.kernel.org
Timur Kristóf [Mon, 25 May 2026 11:33:17 +0000 (13:33 +0200)]
drm/amdgpu: Respect placement requirements in amdgpu_gtt_mgr functions
When testing intersection and compatibility, respect
the actual placement requirements. This is a pre-requisite
for ensuring that UVD CS BOs do not cross 256M segments.
Fixes: ded910f368a5 ("drm/amdgpu: Implement intersect/compatible functions") Suggested-by: Christian König <christian.koenig@amd.com> Signed-off-by: Timur Kristóf <timur.kristof@gmail.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit bc06579ca29dee9c245a41b12e39c7bb6938af5d) Cc: stable@vger.kernel.org
Cao Ruichuang [Tue, 7 Apr 2026 10:26:13 +0000 (18:26 +0800)]
selftests/ftrace: Drop invalid top-level local in test_ownership
test_ownership.tc is sourced by ftracetest under /bin/sh.
The script currently declares mount_point with local at file scope,
which makes /bin/sh abort with "local: not in a function" before the
test can reach the eventfs ownership checks.
Replace the top-level local declaration with a normal shell variable so
kernels that support the gid= tracefs mount option can run the test at
all.
Link: https://lore.kernel.org/r/20260407102613.81419-1-create0818@163.com Fixes: 8b55572e51805 ("tracing/selftests: Add tracefs mount options test") Signed-off-by: Cao Ruichuang <create0818@163.com> Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org> Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org> Cc: stable@vger.kernel.org Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Tvrtko Ursulin [Mon, 1 Jun 2026 14:08:22 +0000 (15:08 +0100)]
drm/amdgpu: Fix context pstate override handling
There are several problems in the context pstate handling code.
The most serious ones are potential use-after-free and NULL pointer
dereferences at context initialization time. Both are due
amdgpu_ctx_init() not holding the adev->pm.stable_pstate_ctx_lock, which
is otherwise used from both sysfs and the context code itself for
modifying and clearing the stored context pointer.
Second issue is that context fini can trample over the pstate
configuration set via sysfs. This is due the restore state
(ctx->stable_pstate) being saved at context init time, and not if, or when
the context actually changes the pstate. As the context exits it will
therefore incorrectly restore to what was set before the sysfs override
was requested.
The simplest fix is to drastically simplify how the state is tracked, by
clearly defining the points at which pstate ownership is taken and
released, and to handle all transitions under the correct lock.
Instead of at context init time, the previous state is saved only at the
point the context overrides the current state, and is restored on context
exit only if the context is still the owner of the current override state.
Xiaogang Chen [Tue, 16 Jun 2026 18:25:56 +0000 (13:25 -0500)]
drm/amdkfd: Use memdup_array_user to copy data from/to user space at kfd ioctls
Several kfd ioctls need transfer array data from/to user space. Kfd driver
uses kmalloc_array with user provided size. That can oversize alloc or 32-bit
wrap with hostile value. Replace it by memdup_array_user that does overflow
checking and allocates through dedicated slab caches, also physical continuous
as kmalloc.
Signed-off-by: Xiaogang Chen <xiaogang.chen@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 4eca4742eb215951f9739ffe0122d179d545a7a4)
Xiaogang Chen [Wed, 27 May 2026 03:50:02 +0000 (22:50 -0500)]
drm/amdkfd: Let driver decide buffer size at AMDKFD_IOC_GET_DMABUF_INFO ioctl
amdkfd driver needs allocate buffer to return bo metadata to user space. The
buffer size is controlled by user currently. It is a potential security issue
that hostile value (e.g. 2 GiB) lets any render-group user trigger order-MAX
allocation/OOM in kernel context.
This patch first finds bo metadata size. If the size is smaller than user
provided value drive can safely allocate buffer in kernel space and copy to
user space buffer. If not, driver will let user know, not allocate and copy.
User will redo with new buffer in user space.
This patch lets driver decide buffer allocation size to avoid potential hostile
size from user space.
Signed-off-by: Xiaogang Chen <xiaogang.chen@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit f54ce9e8cbd3abe0eda3a285f54dc4f572fe589a)
Mikhail Gavrilov [Fri, 29 May 2026 06:47:39 +0000 (11:47 +0500)]
drm/amdgpu: fix recursive ww_mutex acquire in amdgpu_devcoredump_format
When dumping IB contents from a hung job, amdgpu_devcoredump_format()
acquired the VM root PD's reservation via amdgpu_vm_lock_by_pasid() and
then, for each IB, called amdgpu_bo_reserve() on the BO backing the IB.
Both reservations are reservation_ww_class_mutex objects and neither
used a ww_acquire_ctx, which trips lockdep:
WARNING: possible recursive locking detected
--------------------------------------------
kworker/u128:0 is trying to acquire lock: ffff88838b16e1f0 (reservation_ww_class_mutex){+.+.}-{4:4},
at: amdgpu_devcoredump_format+0x1594/0x23f0 [amdgpu]
but task is already holding lock: ffff8882f82681f0 (reservation_ww_class_mutex){+.+.}-{4:4},
at: amdgpu_devcoredump_format+0x1594/0x23f0 [amdgpu]
Possible unsafe locking scenario:
CPU0
----
lock(reservation_ww_class_mutex);
lock(reservation_ww_class_mutex);
*** DEADLOCK ***
May be due to missing lock nesting notation
The two reservations are on different BOs in the captured trace, so the
splat is a lockdep-correctness warning, not an observed deadlock. It
becomes a real self-deadlock whenever the IB BO shares its dma_resv with
the root PD (the always-valid case, see amdgpu_vm_is_bo_always_valid()):
amdgpu_bo_reserve(abo) re-acquires the same ww_mutex without a ticket
and blocks forever. With amdgpu.gpu_recovery=0 the timeout handler
refires every ~2 s and each invocation produces this splat, drowning the
kernel ring buffer.
Now that amdgpu_vm_lock_by_pasid() takes a drm_exec context, move the IB
dumping into a separate helper that locks the root PD and every IB BO
together in a single drm_exec ticket. DRM_EXEC_IGNORE_DUPLICATES handles
IB BOs that share a dma_resv (e.g. always-valid BOs, or two IBs backed
by the same BO). Every lock is now a top-level acquire under one
ww_acquire_ctx, so the recursive ww_mutex condition is gone, and the
per-IB amdgpu_bo_reserve()/amdgpu_bo_unref() dance -- including a BO
refcount leak on the amdgpu_bo_reserve() failure path -- is removed.
Fixes: 7b15fc2d1f1a ("drm/amdgpu: dump job ibs in the devcoredump") Suggested-by: Christian König <christian.koenig@amd.com> Signed-off-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit d6bf4242731219ee08ce54c365631e395486651e)
Mikhail Gavrilov [Fri, 29 May 2026 06:47:38 +0000 (11:47 +0500)]
drm/amdgpu: convert amdgpu_vm_lock_by_pasid() to drm_exec
amdgpu_vm_lock_by_pasid() looks up a VM by PASID and reserves its root
PD with a bare amdgpu_bo_reserve(), returning the still-reserved root to
the caller. A caller that then needs to reserve further BOs (for example
the devcoredump IB dump) ends up nesting reservation_ww_class_mutex
acquires without a ww_acquire_ctx, which lockdep flags as recursive
locking.
Convert the helper to take a drm_exec context and lock the root PD with
drm_exec_lock_obj(). Callers now run it inside a
drm_exec_until_all_locked() loop and can lock additional BOs in the same
ww ticket, so there is no nested ww_mutex acquire.
The drm_exec context holds its own reference on the locked root BO, so
the helper no longer hands a root reference back to the caller: the
root output parameter is dropped, and the transient reference taken
across the PASID lookup is released before returning.
The only existing caller, amdgpu_vm_handle_fault(), is updated
accordingly. Its is_compute_context path, which previously dropped the
root reservation around svm_range_restore_pages() and re-took it, now
finalises the drm_exec context and re-initialises a fresh one; behaviour
is otherwise unchanged.
No functional change intended for the page-fault path.
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 14682de8ad377bf13ea66e47c26dcfea0b19a21d)
UTS_RELEASE evaluates to a static string and changes quite easily (e.g.
uncommitted changes in the source tree or new commits). So when checking
if a patch introduces changes to the resulting binary each usage of
UTS_RELEASE is source of annoyance.
Instead of using UTS_RELEASE directly use init_utsname()->release which
evaluates to the same string but with that a change of UTS_RELEASE
doesn't affect amdgpu_dev_coredump.o.
Geoffrey McRae [Mon, 1 Jun 2026 13:55:53 +0000 (23:55 +1000)]
drm/amdkfd: Fix NULL deref during sysfs teardown
Move kfd_process_remove_sysfs() earlier in kfd_process_wq_release() so
that all sysfs/procfs entries are removed before tearing down PDDs and
dropping lead_thread. The per-process sysfs attributes are backed by
struct kfd_process_device, and their show/store callbacks dereference
PDD fields. Since sysfs removal waits for active callbacks to complete,
removing these entries first closes a race where userspace reads sdma_*
and stats_* files after PDD teardown.
Previously this cleanup ran after kfd_process_destroy_pdds(), which
resets p->n_pdds to 0. This meant kfd_process_remove_sysfs() could no
longer walk the PDD array, so the per-PDD sysfs cleanup did not run as
intended.
This race caused NULL pointer dereferences observed in
kfd_sdma_activity_worker and kfd_procfs_stats_show.
Also harden kfd_process_remove_sysfs() against partially
initialized or already-freed objects:
- Check kobj_queues before removing PASID and deleting it
- Guard kobj_stats and kobj_counters before use
These checks prevent invalid dereferences during cleanup.
Cc: Felix Kuehling <Felix.Kuehling@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Geoffrey McRae <geoffrey.mcrae@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 674c692702341fed321720b4b92036c5934fb485)
drm/amdgpu: validate CP_GFX_SHADOW chunk size in CS pass1
Add a minimum-length check for the AMDGPU_CHUNK_ID_CP_GFX_SHADOW chunk in
amdgpu_cs_pass1(), matching the gate already present for the IB, FENCE and
BO_HANDLES chunk types.
The CP_GFX_SHADOW case previously shared a bare break with the dependency
and syncobj chunk types, which do not dereference a fixed-size struct. When
userspace submits this chunk with length_dw == 0, vmemdup_array_user() is
called with size 0 and returns ZERO_SIZE_PTR, which passes the IS_ERR()
check. amdgpu_cs_p2_shadow() then dereferences chunk->kdata as a struct
drm_amdgpu_cs_chunk_cp_gfx_shadow (reading shadow->flags), faulting on the
ZERO_SIZE_PTR and causing a NULL-pointer dereference.
This is reachable by an unprivileged process in the render group. Reject
undersized chunks with -EINVAL during pass1 so the bad submission is
rejected before pass2 ever dereferences the data.
Fixes: ac9287055ff1 ("drm/amdgpu: add gfx shadow CS IOCTL support") Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 7f61b2eef7415eccdb40850aca0de94211948657) Cc: stable@vger.kernel.org
drm/amdgpu: check amdgpu_vm_bo_find() result in GET_MAPPING_INFO
The AMDGPU_GEM_OP_GET_MAPPING_INFO path of amdgpu_gem_op_ioctl() looks
up the bo_va for the buffer object in the caller's VM via
amdgpu_vm_bo_find(), but uses the returned pointer without checking it.
amdgpu_vm_bo_find() returns NULL when the BO has no bo_va in that VM,
which is the normal case for a BO that has never been mapped. The result
is fed straight into amdgpu_vm_bo_va_for_each_valid_mapping(), which
expands to list_for_each_entry(mapping, &(bo_va)->valids, list) and
dereferences bo_va, causing a NULL pointer dereference.
This is reachable by any process able to issue the ioctl (render group)
simply by requesting mapping info for an unmapped BO.
Return -ENOENT when no bo_va is found, jumping to out_exec so the
drm_exec context and GEM object reference are released.
Fixes: 4d82724f7f2b ("drm/amdgpu: Add mapping info option for GEM_OP ioctl") Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 528b19377affc1cc7362a70a254c1dda793595f9) Cc: stable@vger.kernel.org
If there is an early failure during amdgpu probe, like missing firmware, it
will end up calling amdgpu_irq_disable_all, which takes irq.lock spinlock
without it being initialized.
Initializing irq.lock earlier at amdgpu_device_init fixes the issue.
Fixes: 9950cda2a018 ("drm/amdgpu: drop the drm irq pre/post/un install callbacks") Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com> Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 7dba3e10ecdeec85208e255853fcd3890880b10e)
drm/amdkfd: fix list_del corruption in kfd_criu_resume_svm
The cleanup tail of kfd_criu_resume_svm() walks
svms->criu_svm_metadata_list and kfree()s each struct criu_svm_metadata
without removing it from the list. The list head is left pointing at
freed kmalloc-96 objects.
A second AMDKFD_IOC_CRIU_OP from the same process re-enters: list_empty()
reads the dangling ->next (use-after-free), the loop walks freed entries,
and each is kfree()'d again (double-free). This is reachable by an
unprivileged render-group user via /dev/kfd with no capabilities required.
Add list_del() before the kfree() so the list is properly emptied. The
list_for_each_entry_safe() iterator already caches the next pointer, so
unlinking during the walk is safe.
Fixes: 2a909ae71871 ("drm/amdkfd: CRIU resume shared virtual memory ranges") Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 6322d278a298e2c1430b9d2697743d3a04b788b1)
Pavel Ondračka [Wed, 10 Jun 2026 08:32:45 +0000 (10:32 +0200)]
drm/radeon: fix r100_copy_blit for large BOs
r100_copy_blit() copies BOs as 1024-pixel-wide ARGB8888 blits, so one
GPU page becomes one blit row. Large copies are split into chunks of at
most 8191 rows.
The kernel register header names the packet coordinate dwords SRC_Y_X
and DST_Y_X. In the BITBLT_MULTI description in
R5xx_Acceleration_v1.5.pdf docs, these correspond to [SRC_X1 | SRC_Y1]
and [DST_X1 | DST_Y1], which are signed 13-bit coordinates in the
-8192..8191 range. The old code kept SRC/DST_PITCH_OFFSET at the BO base
and used SRC_Y_X/DST_Y_X as the chunk address, so large BO moves could
exceed that coordinate range.
Compute per-chunk SRC/DST_PITCH_OFFSET bases and emit zero source and
destination coordinates. r100_copy_blit() already packs
SRC/DST_PITCH_OFFSET as pitch plus base offset, so large chunk addresses
belong there rather than in the coordinate fields.
This fixes Prison Architect corruption with 4096x4096 mipped textures
after they are evicted to GTT under memory pressure on RV530.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/work_items/6716 Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 87be26aee76239c6da03e599f238a426897f78ad) Cc: stable@vger.kernel.org
Matthew Schwartz [Thu, 11 Jun 2026 15:44:38 +0000 (08:44 -0700)]
drm/amd/display: Fix mem_type change detection for async flips
[Why]
amdgpu_dm_crtc_mem_type_changed() fetches the "old" and "new" plane state
with two drm_atomic_get_plane_state() calls, which both return the new
state. It compares a state against itself, so it never detects a mem_type
change and never rejects the async flip.
On DCN 3.0.1, this shows up as intermittent corruption when a single DCC
plane is scanned out with immediate flips under gamescope and its buffer
moves between the VRAM carveout and GTT.
[How]
Use drm_atomic_get_old_plane_state() and drm_atomic_get_new_plane_state()
to compare the actual old and new states. These return NULL rather than
an error pointer for a plane that is not part of the commit, so the
IS_ERR() check becomes a NULL check that skips those planes, such as an
unmodified cursor still in the CRTC's plane_mask.
Fixes: 4caacd1671b7 ("drm/amd/display: Do not elevate mem_type change to full update") Reviewed-by: Harry Wentland <harry.wentland@amd.com> Reviewed-by: Melissa Wen <mwen@igalia.com> Signed-off-by: Matthew Schwartz <matthew.schwartz@linux.dev> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 13158e5dbd896281f3e9982b5437cffa5fd621b2)
James Lin [Fri, 12 Jun 2026 14:05:29 +0000 (10:05 -0400)]
drm/amd/display: Add IN_FORMATS_ASYNC support for planes
[Why]
The DRM core exposes an IN_FORMATS_ASYNC plane property describing the
set of format/modifier pairs that are valid for asynchronous (immediate)
page flips. amdgpu already advertises async page flip support via
mode_config.async_page_flip = true, but never implemented the
.format_mod_supported_async plane callback, so the IN_FORMATS_ASYNC
property was not created.
This inconsistency (advertising async flips while exposing IN_FORMATS but
no IN_FORMATS_ASYNC) causes userspace, such as igt-gpu-tools, to emit a
repeated warning during plane initialization, which in turn demotes many
otherwise passing KMS subtests to a WARN result.
[How]
Wire up .format_mod_supported_async to the existing
amdgpu_dm_plane_format_mod_supported callback so the async format list is
populated. amdgpu does not restrict async flips at the format/modifier
level: the async flip constraints are enforced at atomic check and commit
time and only require a fast update (no change to FB pitch, DCC state,
rotation or memory type) between the old and new buffers. Therefore the
set of formats/modifiers valid for async flips is identical to the
regular IN_FORMATS set, and the same callback can be reused.
Reviewed-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: James Lin <PingLei.Lin@amd.com> Signed-off-by: Ivan Lipski <ivan.lipski@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 8e2d7bbd6b184c0c1b0fe7cb404c9b5214d89931)
projects / thirdparty / linux.git / log
