kernel/drivers/gpio/gpiolib.c, branch linux-rolling-lts

gpiolib: normalize the return value of gc->get() on behalf of buggy drivers

2026-03-19T15:08:33Z

commit ec2cceadfae72304ca19650f9cac4b2a97b8a2fc upstream. Commit 86ef402d805d ("gpiolib: sanitize the return value of gpio_chip::get()") started checking the return value of the .get() callback in struct gpio_chip. Now - almost a year later - it turns out that there are quite a few drivers in tree that can break with this change. Partially revert it: normalize the return value in GPIO core but also emit a warning. Cc: stable@vger.kernel.org Fixes: 86ef402d805d ("gpiolib: sanitize the return value of gpio_chip::get()") Reported-by: Dmitry Torokhov Closes: https://lore.kernel.org/all/aZSkqGTqMp_57qC7@google.com/ Reviewed-by: Linus Walleij Reviewed-by: Dmitry Torokhov Link: https://patch.msgid.link/20260219-gpiolib-set-normalize-v2-1-f84630e45796@oss.qualcomm.com Signed-off-by: Bartosz Golaszewski Signed-off-by: Greg Kroah-Hartman

gpiolib: fix race condition for gdev->srcu

2026-01-17T15:35:22Z

[ Upstream commit a7ac22d53d0990152b108c3f4fe30df45fcb0181 ] If two drivers were calling gpiochip_add_data_with_key(), one may be traversing the srcu-protected list in gpio_name_to_desc(), meanwhile other has just added its gdev in gpiodev_add_to_list_unlocked(). This creates a non-mutexed and non-protected timeframe, when one instance is dereferencing and using &gdev->srcu, before the other has initialized it, resulting in crash: [ 4.935481] Unable to handle kernel paging request at virtual address ffff800272bcc000 [ 4.943396] Mem abort info: [ 4.943400] ESR = 0x0000000096000005 [ 4.943403] EC = 0x25: DABT (current EL), IL = 32 bits [ 4.943407] SET = 0, FnV = 0 [ 4.943410] EA = 0, S1PTW = 0 [ 4.943413] FSC = 0x05: level 1 translation fault [ 4.943416] Data abort info: [ 4.943418] ISV = 0, ISS = 0x00000005, ISS2 = 0x00000000 [ 4.946220] CM = 0, WnR = 0, TnD = 0, TagAccess = 0 [ 4.955261] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0 [ 4.955268] swapper pgtable: 4k pages, 48-bit VAs, pgdp=0000000038e6c000 [ 4.961449] [ffff800272bcc000] pgd=0000000000000000 [ 4.969203] , p4d=1000000039739003 [ 4.979730] , pud=0000000000000000 [ 4.980210] phandle (CPU): 0x0000005e, phandle (BE): 0x5e000000 for node "reset" [ 4.991736] Internal error: Oops: 0000000096000005 [#1] PREEMPT SMP ... [ 5.121359] pc : __srcu_read_lock+0x44/0x98 [ 5.131091] lr : gpio_name_to_desc+0x60/0x1a0 [ 5.153671] sp : ffff8000833bb430 [ 5.298440] [ 5.298443] Call trace: [ 5.298445] __srcu_read_lock+0x44/0x98 [ 5.309484] gpio_name_to_desc+0x60/0x1a0 [ 5.320692] gpiochip_add_data_with_key+0x488/0xf00 5.946419] ---[ end trace 0000000000000000 ]--- Move initialization code for gdev fields before it is added to gpio_devices, with adjacent initialization code. Adjust goto statements to reflect modified order of operations Fixes: 47d8b4c1d868 ("gpio: add SRCU infrastructure to struct gpio_device") Reviewed-by: Jakub Lewalski Signed-off-by: Paweł Narewski [Bartosz: fixed a build issue, removed stray newline] Link: https://lore.kernel.org/r/20251224082641.10769-1-bartosz.golaszewski@oss.qualcomm.com Signed-off-by: Bartosz Golaszewski Signed-off-by: Sasha Levin

gpiolib: rename GPIO chip printk macros

2026-01-17T15:35:21Z

[ Upstream commit d4f335b410ddbe3e99f48f8b5ea78a25041274f1 ] The chip_$level() macros take struct gpio_chip as argument so make it follow the convention of using the 'gpiochip_' prefix. Reviewed-by: Linus Walleij Signed-off-by: Bartosz Golaszewski Stable-dep-of: a7ac22d53d09 ("gpiolib: fix race condition for gdev->srcu") Signed-off-by: Sasha Levin

gpiolib: remove unnecessary 'out of memory' messages

2026-01-17T15:35:21Z

[ Upstream commit 0ba6f1ed3808b1f095fbdb490006f0ecd00f52bd ] We don't need to add additional logs when returning -ENOMEM so remove unnecessary error messages. Reviewed-by: Linus Walleij Signed-off-by: Bartosz Golaszewski Stable-dep-of: a7ac22d53d09 ("gpiolib: fix race condition for gdev->srcu") Signed-off-by: Sasha Levin

gpiolib: fix invalid pointer access in debugfs

2025-11-03T15:11:01Z

If the memory allocation in gpiolib_seq_start() fails, the s->private field remains uninitialized and is later dereferenced without checking in gpiolib_seq_stop(). Initialize s->private to NULL before calling kzalloc() and check it before dereferencing it. Fixes: e348544f7994 ("gpio: protect the list of GPIO devices with SRCU") Reviewed-by: Linus Walleij Reviewed-by: Andy Shevchenko Link: https://lore.kernel.org/r/20251103141132.53471-1-brgl@bgdev.pl Signed-off-by: Bartosz Golaszewski

Merge tag 'gpio-updates-for-v6.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux

2025-10-01T18:34:12Z

Pull gpio updates from Bartosz Golaszewski: "There are two new drivers and support for more models in existing ones. The generic GPIO API has been reworked and all users converted which allowed us to move the fields specific to the generic GPIO implementation out of the high-level struct gpio_chip into its own structure that wraps the gpio_chip. Other than that, there's nothing too exciting. Mostly minor tweaks and fixes all over the place, some refactoring and some small new features in helper modules. GPIO core: - add support for sparse pin ranges to the glue between GPIO and pinctrl - use a common prefix across all GPIO descriptor flags for improved namespacing New drivers: - add new GPIO driver for the Nuvoton NCT6694 - add new GPIO driver for MAX7360 Driver improvements: - add support for Tegra 256 to the gpio-tegra186 driver - add support for Loongson-2K0300 to the gpio-loongson-64bit driver - refactor the gpio-aggregator module to expose its GPIO forwarder API to other in-kernel users (to enable merging of a new pinctrl driver that uses it) - convert all remaining drivers to using the modernized generic GPIO chip API and remove the old interface - stop displaying global GPIO numbers in debugfs output of controller drivers - extend the gpio-regmap helper with a new config option and improve its support for GPIO interrupts - remove redundant fast_io parameter from regmap configs in GPIO drivers that already use MMIO regmaps which imply it - add support for a new model in gpio-mmio: ixp4xx expansion bus - order includes alphabetically in a few drivers for better readability - use generic device properties where applicable - use devm_mutex_init() where applicable - extend build coverage of drivers by enabling more to be compiled with COMPILE_TEST enabled - allow building gpio-stmpe as a module - use dev_err_probe() where it makes sense in drivers Late driver fixes: - fix setting GPIO direction to output in gpio-mpfs Documentation: - document the usage of software nodes with GPIO chips Device-tree bindings: - Add DT bindings documents for new hardware: Tegra256, MAX7360 - Document a new model in Loongson bindings: LS2K0300 - Document a new model using the generic GPIO binding: IXP4xx - Convert the DT binding for fsl,mxs-pinctrl to YAML - fix the schema ID in the "trivial" GPIO schema - describe GPIO hogs in the generic GPIO binding" * tag 'gpio-updates-for-v6.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux: (122 commits) gpio: mpfs: fix setting gpio direction to output gpio: generic: move GPIO_GENERIC_ flags to the correct header gpio: generic: rename BGPIOF_ flags to GPIO_GENERIC_ gpio: nomadik: fix the debugfs helper stub MAINTAINERS: Add entry on MAX7360 driver input: misc: Add support for MAX7360 rotary input: keyboard: Add support for MAX7360 keypad gpio: max7360: Add MAX7360 gpio support gpio: regmap: Allow to provide init_valid_mask callback gpio: regmap: Allow to allocate regmap-irq device pwm: max7360: Add MAX7360 PWM support pinctrl: Add MAX7360 pinctrl driver mfd: Add max7360 support dt-bindings: mfd: gpio: Add MAX7360 rtc: Add Nuvoton NCT6694 RTC support hwmon: Add Nuvoton NCT6694 HWMON support watchdog: Add Nuvoton NCT6694 WDT support can: Add Nuvoton NCT6694 CANFD support i2c: Add Nuvoton NCT6694 I2C support gpio: Add Nuvoton NCT6694 GPIO support ...

gpiolib: Extend software-node support to support secondary software-nodes

2025-09-22T14:23:43Z

When a software-node gets added to a device which already has another fwnode as primary node it will become the secondary fwnode for that device. Currently if a software-node with GPIO properties ends up as the secondary fwnode then gpiod_find_by_fwnode() will fail to find the GPIOs. Add a new gpiod_fwnode_lookup() helper which falls back to calling gpiod_find_by_fwnode() with the secondary fwnode if the GPIO was not found in the primary fwnode. Fixes: e7f9ff5dc90c ("gpiolib: add support for software nodes") Cc: stable@vger.kernel.org Signed-off-by: Hans de Goede Reviewed-by: Dmitry Torokhov Link: https://lore.kernel.org/r/20250920200955.20403-1-hansg@kernel.org Signed-off-by: Bartosz Golaszewski

gpiolib: add a common prefix to GPIO descriptor flags

2025-09-12T07:20:38Z

While these flags are private within drivers/gpio/, when looking at the code, it's not really clear they are GPIO-specific. Since these are GPIO descriptor flags, prepend their names with a common "GPIOD" prefix. While at it: update the flags' docs: make spelling consistent, correct outdated information, etc. Reviewed-by: Linus Walleij Link: https://lore.kernel.org/r/20250909-rename-gpio-flags-v1-1-bda208a40856@linaro.org Signed-off-by: Bartosz Golaszewski

gpiolib: add support to register sparse pin range

2025-08-11T13:39:31Z

Add support to register for GPIO<->pin mapping using a list of non consecutive pins. The core already supports sparse pin range (pins member of struct pinctrl_gpio_range), but it was not possible to register one. If pins is not NULL the core uses it, otherwise it assumes that a consecutive pin range was registered and it uses pin_base. The function gpiochip_add_pin_range() which allocates and fills the struct pinctrl_gpio_range was renamed to gpiochip_add_pin_range_with_pins() and the pins parameter was added. Two new functions were added, gpiochip_add_pin_range() and gpiochip_add_sparse_pin_range() to register a consecutive or sparse pins range. Both use gpiochip_add_pin_range_with_pins(). Reviewed-by: Linus Walleij Reviewed-by: Andy Shevchenko Acked-by: Linus Walleij Signed-off-by: Thomas Richard Link: https://lore.kernel.org/r/20250811-aaeon-up-board-pinctrl-support-v9-1-29f0cbbdfb30@bootlin.com Signed-off-by: Bartosz Golaszewski

treewide: rename GPIO set callbacks back to their original names

2025-08-07T08:07:06Z

The conversion of all GPIO drivers to using the .set_rv() and .set_multiple_rv() callbacks from struct gpio_chip (which - unlike their predecessors - return an integer and allow the controller drivers to indicate failures to users) is now complete and the legacy ones have been removed. Rename the new callbacks back to their original names in one sweeping change. Signed-off-by: Bartosz Golaszewski