gpio: shared: check if a reference is populated before cleaning its resources

It's possible that not all proxy entries will be set up when the device
gets removed so check if they are before trying to dereference members
which are still NULL. This can happen if some consumers never requested
their shared GPIOs.

Fixes: a060b8c511ab ("gpiolib: implement low-level, shared GPIO support")
Link: https://lore.kernel.org/r/20251206-gpio-shared-teardown-fixes-v1-2-35ac458cfce1@oss.qualcomm.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

gpio: shared: fix NULL-pointer dereference in teardown path

We need to actually store the address of the GPIO lookup table in the
reference struct before we try to free it or - worse - dereference its
members.

Fixes: a060b8c511ab ("gpiolib: implement low-level, shared GPIO support")
Link: https://lore.kernel.org/r/20251206-gpio-shared-teardown-fixes-v1-1-35ac458cfce1@oss.qualcomm.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

gpio: shared: ignore disabled nodes when traversing the device-tree

Don't consider disabled devices when traversing the device-tree looking
for shared GPIOs. Even if they do share a phandle to a pin, they can
never be instantiated and request it.

Fixes: a060b8c511ab ("gpiolib: implement low-level, shared GPIO support")
Link: https://lore.kernel.org/r/20251203092309.34737-1-bartosz.golaszewski@oss.qualcomm.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

gpiolib: acpi: Add quirk for Dell Precision 7780

Dell Precision 7780 often wakes up on its own from suspend. Sometimes
wake up happens immediately (i. e. within 7 seconds), sometimes it happens
after, say, 30 minutes.

Fixes: 1796f808e4bb ("HID: i2c-hid: acpi: Stop setting wakeup_capable")
Link: https://lore.kernel.org/linux-i2c/197ae95ffd8.dc819e60457077.7692120488609091556@zohomail.com/
Cc: stable@vger.kernel.org
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Askar Safin <safinaskar@gmail.com>
Link: https://lore.kernel.org/r/20251206180414.3183334-2-safinaskar@gmail.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

gpio: tb10x: fix OF_GPIO dependency

Selecting OF_GPIO is generally not allowed, it always gets enabled
when both GPIOLIB and OF are turned on.

The tb10x driver now warns about this after it was enabled for
compile-testing:

WARNING: unmet direct dependencies detected for OF_GPIO
  Depends on [n]: GPIOLIB [=y] && OF [=n] && HAS_IOMEM [=y]
  Selected by [y]:
  - GPIO_TB10X [=y] && GPIOLIB [=y] && HAS_IOMEM [=y] && (ARC_PLAT_TB10X || COMPILE_TEST [=y])

OF_GPIO is not required for compile-testing and is already enabled
when the driver is usable, so just drop the 'select' line.

Fixes: 682fbb18e14c ("gpio: tb10x: allow building the module with COMPILE_TEST=y")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20251205095429.1291866-1-arnd@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

gpio: qixis: select CONFIG_REGMAP_MMIO

The regmap drivers need to be selected by each user, without that there
can be configurations that fail to link:

x86_64-linux-ld: drivers/gpio/gpio-qixis-fpga.o: in function `qixis_cpld_gpio_probe':
gpio-qixis-fpga.c:(.text+0x13a): undefined reference to `__devm_regmap_init_mmio_clk'

Fixes: e88500247dc3 ("gpio: add QIXIS FPGA GPIO controller")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20251204094928.1031494-1-arnd@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

gpio: regmap: Fix memleak in error path in gpio_regmap_register()

Call gpiochip_remove() to free the resources allocated by
gpiochip_add_data() in error path.

Fixes: 553b75d4bfe9 ("gpio: regmap: Allow to allocate regmap-irq device")
Fixes: ae495810cffe ("gpio: regmap: add the .fixed_direction_output configuration parameter")
CC: stable@vger.kernel.org
Co-developed-by: WangYuli <wangyl5933@chinaunicom.cn>
Signed-off-by: WangYuli <wangyl5933@chinaunicom.cn>
Signed-off-by: Wentao Guan <guanwentao@uniontech.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20251204101303.30353-1-guanwentao@uniontech.com
[Bartosz: reworked the commit message]
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

gpio: mmio: fix bad guard conversion

A recent spinlock guard conversion consistently used the wrong guard so
that interrupts are no longer disabled while holding the chip lock
(which can cause deadlocks).

Fixes: 7e061b462b3d ("gpio: mmio: use lock guards")
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20251203105206.24453-1-johan@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

gpio: loongson: Switch 2K2000/3000 GPIO to BYTE_CTRL_MODE

The manuals of 2K2000 says both BIT_CTRL_MODE and BYTE_CTRL_MODE are
supported but the latter is recommended.  Also on 2K3000, per the ACPI
DSDT the GPIO controller is compatible with 2K2000, but it fails to
operate GPIOs 62 and 63 (and maybe others) using BIT_CTRL_MODE.
Using BYTE_CTRL_MODE also makes those 2K3000 GPIOs work.

Fixes: 3feb70a61740 ("gpio: loongson: add more gpio chip support")
Cc: stable@vger.kernel.org
Signed-off-by: Xi Ruoyao <xry111@xry111.site>
Reviewed-by: Huacai Chen <chenhuacai@loongson.cn>
Link: https://lore.kernel.org/r/20251128075033.255821-1-xry111@xry111.site
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: regmap: fix kernel-doc notation

Add a ':' to the end of struct member names to prevent kernel-doc
warnings:

Warning: include/linux/gpio/regmap.h:108 struct member 'regmap_irq_line'
 not described in 'gpio_regmap_config'
Warning: include/linux/gpio/regmap.h:108 struct member 'regmap_irq_flags'
 not described in 'gpio_regmap_config'

Fixes: 553b75d4bfe9 ("gpio: regmap: Allow to allocate regmap-irq device")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Michael Walle <mwalle@kernel.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20251128062739.845403-1-rdunlap@infradead.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: shared: fix a deadlock

It's possible that the auxiliary proxy device we add when setting up the
GPIO controller exposing shared pins, will get matched and probed
immediately. The gpio-proxy-driver will then retrieve the shared
descriptor structure. That will cause a recursive mutex locking and
a deadlock because we're already holding the gpio_shared_lock in
gpio_device_setup_shared() and try to take it again in
devm_gpiod_shared_get() like this:

[    4.298346] gpiolib_shared: GPIO 130 owned by f100000.pinctrl is shared by multiple consumers
[    4.307157] gpiolib_shared: Setting up a shared GPIO entry for speaker@0,3
[    4.314604]
[    4.316146] ============================================
[    4.321600] WARNING: possible recursive locking detected
[    4.327054] 6.18.0-rc7-next-20251125-g3f300d0674f6-dirty #3887 Not tainted
[    4.334115] --------------------------------------------
[    4.339566] kworker/u32:3/71 is trying to acquire lock:
[    4.344931] ffffda019ba71850 (gpio_shared_lock){+.+.}-{4:4}, at: devm_gpiod_shared_get+0x34/0x2e0
[    4.354057]
[    4.354057] but task is already holding lock:
[    4.360041] ffffda019ba71850 (gpio_shared_lock){+.+.}-{4:4}, at: gpio_device_setup_shared+0x30/0x268
[    4.369421]
[    4.369421] other info that might help us debug this:
[    4.376126]  Possible unsafe locking scenario:
[    4.376126]
[    4.382198]        CPU0
[    4.384711]        ----
[    4.387223]   lock(gpio_shared_lock);
[    4.390992]   lock(gpio_shared_lock);
[    4.394761]
[    4.394761]  *** DEADLOCK ***
[    4.394761]
[    4.400832]  May be due to missing lock nesting notation
[    4.400832]
[    4.407802] 5 locks held by kworker/u32:3/71:
[    4.412279]  #0: ffff000080020948 ((wq_completion)events_unbound){+.+.}-{0:0}, at: process_one_work+0x194/0x64c
[    4.422650]  #1: ffff800080963d60 (deferred_probe_work){+.+.}-{0:0}, at: process_one_work+0x1bc/0x64c
[    4.432117]  #2: ffff00008165c8f8 (&dev->mutex){....}-{4:4}, at: __device_attach+0x3c/0x198
[    4.440700]  #3: ffffda019ba71850 (gpio_shared_lock){+.+.}-{4:4}, at: gpio_device_setup_shared+0x30/0x268
[    4.450523]  #4: ffff0000810fe918 (&dev->mutex){....}-{4:4}, at: __device_attach+0x3c/0x198
[    4.459103]
[    4.459103] stack backtrace:
[    4.463581] CPU: 6 UID: 0 PID: 71 Comm: kworker/u32:3 Not tainted 6.18.0-rc7-next-20251125-g3f300d0674f6-dirty #3887 PREEMPT
[    4.463589] Hardware name: Qualcomm Technologies, Inc. Robotics RB5 (DT)
[    4.463593] Workqueue: events_unbound deferred_probe_work_func
[    4.463602] Call trace:
[    4.463604]  show_stack+0x18/0x24 (C)
[    4.463617]  dump_stack_lvl+0x70/0x98
[    4.463627]  dump_stack+0x18/0x24
[    4.463636]  print_deadlock_bug+0x224/0x238
[    4.463643]  __lock_acquire+0xe4c/0x15f0
[    4.463648]  lock_acquire+0x1cc/0x344
[    4.463653]  __mutex_lock+0xb8/0x840
[    4.463661]  mutex_lock_nested+0x24/0x30
[    4.463667]  devm_gpiod_shared_get+0x34/0x2e0
[    4.463674]  gpio_shared_proxy_probe+0x18/0x138
[    4.463682]  auxiliary_bus_probe+0x40/0x78
[    4.463688]  really_probe+0xbc/0x2c0
[    4.463694]  __driver_probe_device+0x78/0x120
[    4.463701]  driver_probe_device+0x3c/0x160
[    4.463708]  __device_attach_driver+0xb8/0x140
[    4.463716]  bus_for_each_drv+0x88/0xe8
[    4.463723]  __device_attach+0xa0/0x198
[    4.463729]  device_initial_probe+0x14/0x20
[    4.463737]  bus_probe_device+0xb4/0xc0
[    4.463743]  device_add+0x578/0x76c
[    4.463747]  __auxiliary_device_add+0x40/0xac
[    4.463752]  gpio_device_setup_shared+0x1f8/0x268
[    4.463758]  gpiochip_add_data_with_key+0xdac/0x10ac
[    4.463763]  devm_gpiochip_add_data_with_key+0x30/0x80
[    4.463768]  msm_pinctrl_probe+0x4b0/0x5e0
[    4.463779]  sm8250_pinctrl_probe+0x18/0x40
[    4.463784]  platform_probe+0x5c/0xa4
[    4.463793]  really_probe+0xbc/0x2c0
[    4.463800]  __driver_probe_device+0x78/0x120
[    4.463807]  driver_probe_device+0x3c/0x160
[    4.463814]  __device_attach_driver+0xb8/0x140
[    4.463821]  bus_for_each_drv+0x88/0xe8
[    4.463827]  __device_attach+0xa0/0x198
[    4.463834]  device_initial_probe+0x14/0x20
[    4.463841]  bus_probe_device+0xb4/0xc0
[    4.463847]  deferred_probe_work_func+0x90/0xcc
[    4.463854]  process_one_work+0x214/0x64c
[    4.463860]  worker_thread+0x1bc/0x360
[    4.463866]  kthread+0x14c/0x220
[    4.463871]  ret_from_fork+0x10/0x20
[   77.265041] random: crng init done

Fortunately, at the time of creating of the auxiliary device, we already
know the correct entry so let's store it as the device's platform data.
We don't need to hold gpio_shared_lock in devm_gpiod_shared_get() as
we're not removing the entry or traversing the list anymore but we still
need to protect it from concurrent modification of its fields so add a
more fine-grained mutex.

Fixes: a060b8c511ab ("gpiolib: implement low-level, shared GPIO support")
Reported-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Closes: https://lore.kernel.org/all/fimuvblfy2cmn7o4wzcxjzrux5mwhvlvyxfsgeqs6ore2xg75i@ax46d3sfmdux/
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Tested-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20251128-gpio-shared-deadlock-v2-1-9f3ae8ddcb09@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: shared-proxy: set suppress_bind_attrs

User-space must not fiddle with shared-proxy auxiliary devices. Disable
bind/unbind attributes in sysfs.

Link: https://lore.kernel.org/r/20251126191730.66277-1-brgl@bgdev.pl
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: shared: ignore GPIO hogs when traversing the device tree

GPIO hogs have a "gpios" property but it's not a phandle to a remote
node - it references the parent GPIO controller. We must not try to
parse it as a phandle.

Fixes: a060b8c511ab ("gpiolib: implement low-level, shared GPIO support")
Reported-by: Cosmin Tanislav <demonsingur@gmail.com>
Closes: https://lore.kernel.org/all/2d96e464-e17c-4ff5-9a08-b215b77da04f@gmail.com/
Link: https://lore.kernel.org/r/20251126-gpio-shared-fixes-v1-2-18309c0e87b5@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: shared: ignore special __symbols__ node when traversing device tree

The __symbols__ node is a special, internal node and its properties must
not be considered when scanning the device-tree for shared GPIOs.

Fixes: a060b8c511ab ("gpiolib: implement low-level, shared GPIO support")
Reported-by: Jon Hunter <jonathanh@nvidia.com>
Closes: https://lore.kernel.org/all/0829a21c-f97d-41b6-90bc-2acaec42caab@nvidia.com/
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20251126-gpio-shared-fixes-v1-1-18309c0e87b5@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: shared: handle the reset-gpios corner case

There's an unexpected interaction between the reset-gpio driver and the
shared GPIO support. The reset-gpio device is an auxiliary device that's
created dynamically and fulfills a similar role to the gpio-shared-proxy
driver but is limited in scope to just supporting the "reset-gpios"
property.

The shared GPIO core code does not take into account that the machine
lookup entry we create when scanning the device-tree must connect the
reset-gpio device - that is the actual consumer of the GPIO and not the
consumer defined on the device tree, which in turn consumes the shared
reset control exposed by the reset-gpio device - to the GPIO controller.

We also must not skip the gpio-shared-proxy driver as it's possible that
a shared GPIO may be used by one consumer as a reset-gpios going through
the reset-gpio device and another that uses GPIOLIB.

We need to make it a special case handled in gpiolib-shared.c. Add a new
function - gpio_shared_dev_is_reset_gpio() - whose role it is to verify
if a non-matching consumer of a shared pin is a reset-gpio device and
make sure it's the right one for this pin. To that end make sure that
its parent is the GPIO controller in question and that the fwnode we
identified as sharing the pin references that controller via the
"reset-gpios" property.

Only include that code if the reset-gpio driver is enabled.

Fixes: a060b8c511ab ("gpiolib: implement low-level, shared GPIO support")
Reported-by: Val Packett <val@packett.cool>
Closes: https://lore.kernel.org/all/3b5d9df5-934d-4591-8827-6c9573a6f7ba@packett.cool/
Tested-by: Val Packett <val@packett.cool>
Tested-by: Abel Vesa <abel.vesa@linaro.org>
Link: https://lore.kernel.org/r/20251125-gpiolib-shared-reset-gpio-fix-v2-1-4eb6fa41f1dd@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: zynq: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use __maybe_unused

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-15-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: xilinx: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use __maybe_unused

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-14-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: xgene: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use __maybe_unused

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-13-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: uniphier: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use __maybe_unused

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-12-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: tqmx86: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use __maybe_unused

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20251124002105.25429-11-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: pch: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use __maybe_unused

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Andy Shevchenko <andy@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-10-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: omap: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use __maybe_unused

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-9-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: msc313: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use __maybe_unused

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-8-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: mlxbf2: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use __maybe_unused

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-7-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: ml-ioh: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use __maybe_unused

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Andy Shevchenko <andy@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-6-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: pl061: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use #ifdef guards.

This has the advantage of always compiling these functions in,
independently of any Kconfig option. Thanks to that, bugs and other
regressions are subsequently easier to catch.

The pl061_context_save_regs structure is always embedded into struct
pl061 to simplify code, so this brings a tiny 8 bytes memory overhead
for !CONFIG_PM_SLEEP.

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-5-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: htc-egpio: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use #ifdef guards.

This has the advantage of always compiling these functions in,
independently of any Kconfig option. Thanks to that, bugs and other
regressions are subsequently easier to catch.

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-4-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: brcmstb: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use #ifdef guards.

This has the advantage of always compiling these functions in,
independently of any Kconfig option. Thanks to that, bugs and other
regressions are subsequently easier to catch.

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Doug Berger <opendmb@gmail.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20251124002105.25429-3-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: dwapb: Use modern PM macros

Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or
CONFIG_PM_SLEEP are disabled, without having to use #ifdef guards.

This has the advantage of always compiling these functions in,
independently of any Kconfig option. Thanks to that, bugs and other
regressions are subsequently easier to catch.

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Link: https://lore.kernel.org/r/20251124002105.25429-2-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

Merge tag 'intel-gpio-v6.19-1' of git://git.kernel.org/pub/scm/linux/kernel/git/andy/linux-gpio-intel into gpio/for-next

intel-gpio for v6.19-1

* Replace min_t() by min() to avoid cutting upper bits and do type checking

gpiolib: acpi: use min() instead of min_t()

string: use __attribute__((nonnull())) in strends()

The arguments of strends() must not be NULL so annotate the function
with the nonnull attribute.

Suggested-by: Kees Cook <kees@kernel.org>
Link: https://lore.kernel.org/r/20251118-strends-follow-up-v1-2-d3f8ef750f59@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

Merge tag 'reset-gpio-for-v6.19-2' of https://git.pengutronix.de/git/pza/linux into gpio/for-next

Reset/GPIO/swnode changes for v6.19 (v2)

* Fix chtwc_int33fe build issue since commit d7cdbbc93c56 ("software
  node: allow referencing firmware nodes").

platform/x86: intel: chtwc_int33fe: don't dereference swnode args

Members of struct software_node_ref_args should not be dereferenced
directly but set using the provided macros. Commit d7cdbbc93c56
("software node: allow referencing firmware nodes") changed the name of
the software node member and caused a build failure. Remove all direct
dereferences of the ref struct as a fix.

However, this driver also seems to abuse the software node interface by
waiting for a node with an arbitrary name "intel-xhci-usb-sw" to appear
in the system before setting up the reference for the I2C device, while
the actual software node already exists in the intel-xhci-usb-role-switch
module and should be used to set up a static reference. Add a FIXME for
a future improvement.

Fixes: d7cdbbc93c56 ("software node: allow referencing firmware nodes")
Fixes: 53c24c2932e5 ("platform/x86: intel_cht_int33fe: use inline reference properties")
Cc: stable@vger.kernel.org
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Closes: https://lore.kernel.org/all/20251121111534.7cdbfe5c@canb.auug.org.au/
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Hans de Goede <johannes.goede@oss.qualcomm.com>
Acked-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

Merge tag 'reset-gpio-for-v6.19' of https://git.pengutronix.de/git/pza/linux into gpio/for-next

Reset/GPIO/swnode changes for v6.19

* Extend software node implementation, allowing its properties to reference
  existing firmware nodes.
* Update the GPIO property interface to use reworked swnode macros.
* Rework reset-gpio code to use GPIO lookup via swnode.
* Fix spi-cs42l43 driver to work with swnode changes.

reset: gpio: use software nodes to setup the GPIO lookup

GPIO machine lookup is a nice mechanism for associating GPIOs with
consumers if we don't know what kind of device the GPIO provider is or
when it will become available. However in the case of the reset-gpio, we
are already holding a reference to the device and so can reference its
firmware node. Let's setup a software node that references the relevant
GPIO and attach it to the auxiliary device we're creating.

Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

reset: gpio: convert the driver to using the auxiliary bus

As the reset-gpio devices are purely virtual and never instantiated from
real firmware nodes, let's convert the driver to using the - more
fitting - auxiliary bus.

Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

reset: make the provider of reset-gpios the parent of the reset device

Auxiliary devices really do need a parent so ahead of converting the
reset-gpios driver to registering on the auxiliary bus, make the GPIO
device that provides the reset GPIO the parent of the reset-gpio device.
To that end move the lookup of the GPIO device by fwnode to the
beginning of __reset_add_reset_gpio_device() which has the added benefit
of bailing out earlier, before allocating resources for the virtual
device, if the chip is not up yet.

Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

reset: order includes alphabetically in reset/core.c

For better readability and easier maintenance order the includes
alphabetically.

Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

gpio: swnode: allow referencing GPIO chips by firmware nodes

When doing a software node lookup, we require both the fwnode that
references a GPIO chip as well as the node associated with that chip to
be software nodes. However, we now allow referencing generic firmware
nodes from software nodes in driver core so we should allow the same in
GPIO core. Make the software node name check optional and dependent on
whether the referenced firmware node is a software node. If it's not,
just continue with the lookup.

Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

spi: cs42l43: Use actual ACPI firmware node for chip selects

On some systems the cs42l43 has amplifiers attached to its SPI
controller that are not properly defined in ACPI. Currently
software nodes are added to support this case, however, the chip
selects for these devices are specified using a hack. A software
node is added with the same name as the pinctrl driver, as the
look up was name based, this allowed the GPIO look up to return
the pinctrl driver even though the swnode was not owned by it.
This was necessary as the swnodes did not support directly
linking to real firmware nodes.

Since commit e5d527be7e69 ("gpio: swnode: don't use the swnode's
name as the key for GPIO lookup") changed the lookup to be
fwnode based this hack will no longer find the pinctrl driver,
resulting in the driver not probing. There is no pinctrl driver
attached to the swnode itself. But other patches did add support
for linking a swnode to a real fwnode node [1]. As such the hack
is no longer needed, so switch over to just passing the real
fwnode for the pinctrl property to avoid any issues.

[Bartosz:
  - remove unneeded Fixes: tag,
  - use PROPERTY_ENTRY_REF_ARRAY() instead of PROPERTY_ENTRY_REF_ARRAY_LEN()]

Link: https://lore.kernel.org/linux-gpio/20251106-reset-gpios-swnodes-v6-0-69aa852de9e4@linaro.org/
Fixes: 439fbc97502a ("spi: cs42l43: Add bridged cs35l56 amplifiers")
Cc: stable+noautosel@kernel.org # Don't backport, previous approach works, fix relies on swnode changes
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Acked-by: Mark Brown <broonie@kernel.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

software node: allow referencing firmware nodes

At the moment software nodes can only reference other software nodes.
This is a limitation for devices created, for instance, on the auxiliary
bus with a dynamic software node attached which cannot reference devices
the firmware node of which is "real" (as an OF node or otherwise).

Make it possible for a software node to reference all firmware nodes in
addition to static software nodes. To that end: add a second pointer to
struct software_node_ref_args of type struct fwnode_handle. The core
swnode code will first check the swnode pointer and if it's NULL, it
will assume the fwnode pointer should be set.

Software node graphs remain the same, as in: the remote endpoints still
have to be software nodes.

Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Tested-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

software node: increase the reference of the swnode by its fwnode

Once we allow software nodes to reference other kinds of firmware nodes,
the node in args will no longer necessarily be a software node so bump
its reference count using its fwnode interface.

Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Tested-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

software node: read the reference args via the fwnode API

Once we allow software nodes to reference all kinds of firmware nodes,
the refnode here will no longer necessarily be a software node so read
its proprties going through its fwnode implementation.

Acked-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Tested-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

string: fix kerneldoc formatting in strends()

strends() kernel doc should have used `@str:` format for arguments
instead of `@str -`.

Fixes: 197b3f3c70d6 ("string: provide strends()")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Closes: https://lore.kernel.org/all/20251118134748.40f03b9c@canb.auug.org.au/
Link: https://lore.kernel.org/r/20251118-strends-follow-up-v1-1-d3f8ef750f59@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: fxl6408: Add suspend/resume support

Currently, during suspend, do nothing; during resume, just sync the
regmap cache to hw regs.

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Link: https://lore.kernel.org/r/20251119140455.10096-1-jszhang@kernel.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: menz127: add support for 16Z034 and 16Z037 GPIO controllers

The 16Z034 and 16Z037 are 8 bits GPIO controllers that share the
same registers and features of the 16Z127 GPIO controller.

Signed-off-by: Jose Javier Rodriguez Barbarin <dev-josejavier.rodriguez@duagon.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20251118083115.9545-1-dev-josejavier.rodriguez@duagon.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpiolib: acpi: use min() instead of min_t()

min_t(u16, a, b) casts an 'unsigned long' to 'u16'.
Use min(a, b) instead as it promotes the both values to int
and so cannot discard significant bits.

In this case the values should be ok.

Detected by an extra check added to min_t().

Signed-off-by: David Laight <david.laight.linux@gmail.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

gpio: elkhartlake: Convert to auxiliary driver

Since PCI device should not be abusing platform device, MFD parent to
platform child path is no longer being pursued for this driver. Convert
it to auxiliary driver, which will be used by EHL PSE auxiliary device.

Signed-off-by: Raag Jadav <raag.jadav@intel.com>
Acked-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/r/20251112034040.457801-3-raag.jadav@intel.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

platform/x86/intel: Introduce Intel Elkhart Lake PSE I/O

Intel Elkhart Lake Programmable Service Engine (PSE) includes two PCI
devices that expose two different capabilities of GPIO and Timed I/O
as a single PCI function through shared MMIO with below layout.

GPIO: 0x0000 - 0x1000
TIO:  0x1000 - 0x2000

This driver enumerates the PCI parent device and creates auxiliary child
devices for these capabilities. The actual functionalities are provided
by their respective auxiliary drivers.

Signed-off-by: Raag Jadav <raag.jadav@intel.com>
Acked-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/r/20251112034040.457801-2-raag.jadav@intel.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

Documentation: gpio: Add a compatibility and feature list for PCA953x

I went through all the datasheets and created this note listing
chip functions and register layouts.

Signed-off-by: Levente Révész <levente.revesz@eilabs.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Bagas Sanjaya <bagasdotme@gmail.com>
Link: https://lore.kernel.org/r/20251112224924.2091880-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: shared: extend the ifdef guard to gpio_shared_find_entry()

While this function is supposed to be used by all scanning functions, so
far we only have a single one for OF trees. Once we add support for ACPI
and software nodes, we'll drop the CONFIG_OF guard around this routine
but in order to avoid build warnings, let's extend it to cover it in the
meantime.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202511180232.EItKeYjY-lkp@intel.com/
Link: https://lore.kernel.org/r/20251118-gpiolib-shared-of-guard-v1-1-e4ef149a2e0b@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: shared: fix a NULL-pointer dereference

The fact that CONFIG_OF is enabled does not mean that the device tree is
populated and that of_root points to a valid device node. Check if it's
NULL before trying to traverse the tree.

Fixes: a060b8c511ab ("gpiolib: implement low-level, shared GPIO support")
Reported-by: Mark Brown <broonie@kernel.org>
Closes: https://lore.kernel.org/all/dbe20642-9662-40af-a593-c1263baea73b@sirena.org.uk/
Tested-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20251118200459.13969-1-brgl@bgdev.pl
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

dt-bindings: gpio: mpfs-gpio: Add pic64gx GPIO compatibility

pic64gx GPIO is compatible with mpfs-gpio controller, add it with a
fallback.

Signed-off-by: Pierre-Henry Moussay <pierre-henry.moussay@microchip.com>
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Link: https://lore.kernel.org/r/20251117-grumbly-oversized-2215fe887181@spud
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: tegra186: Fix GPIO name collisions for Tegra410

On Tegra410, Compute and System GPIOs have same port names. This
results in the same GPIO names for both Compute and System GPIOs
during initialization in `tegra186_gpio_probe()`, which results in
following warnings:

  kernel: gpio gpiochip1: Detected name collision for GPIO name 'PA.00'
  kernel: gpio gpiochip1: Detected name collision for GPIO name 'PA.01'
  kernel: gpio gpiochip1: Detected name collision for GPIO name 'PA.02'
  kernel: gpio gpiochip1: Detected name collision for GPIO name 'PB.00'
  kernel: gpio gpiochip1: Detected name collision for GPIO name 'PB.01'
  ...

Add GPIO name prefix in the SoC data and use it to initialize the GPIO
name.

Port names remain unchanged for previous SoCs. On Tegra410, Compute
GPIOs are named COMPUTE-P<PORT>.GPIO, and System GPIOs are named
SYSTEM-P<PORT>.GPIO.

Fixes: 9631a10083d8 ("gpio: tegra186: Add support for Tegra410")
Signed-off-by: Kartik Rajput <kkartik@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20251113163112.885900-1-kkartik@nvidia.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpiolib: legacy: Allow to kill devm_gpio_request_one() independently

Allow to kill devm_gpio_request_one() independently by converting it
to use legacy APIs that will be alive a bit longer.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20251112093608.1481030-3-andriy.shevchenko@linux.intel.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpiolib: legacy: Make sure we kill gpio_request_one() first

Make sure we kill gpio_request_one() first by converting it to
use legacy APIs that will be alive a bit longer. In particular,
this also shows the code we will use in another function to make
it die independently.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20251112093608.1481030-2-andriy.shevchenko@linux.intel.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

Merge tag 'gpio/shared-gpios-for-v6.19-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux.git into gpio/for-next

Immutable branch between the GPIO, ASoC and regulator trees for v6.19-rc1

Add better support for GPIOs shared by multiple consumers.

arm64: select HAVE_SHARED_GPIOS for ARCH_QCOM

Some qualcomm platforms use shared GPIOs. Enable support for them by
selecting the Kconfig switch provided by GPIOLIB.

Acked-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Bjorn Andersson <andersson@kernel.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20251112-gpio-shared-v4-7-b51f97b1abd8@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: provide gpiod_is_shared()

Provide an interface allowing consumers to check if a GPIO descriptor
represents a GPIO that can potentially be shared by multiple consumers
at the same time. This is exposed to allow subsystems that already
work around the limitations of the current non-exclusive GPIO handling
in some ways, to gradually convert to relying on the new shared GPIO
feature of GPIOLIB.

Extend the gpiolib-shared module to mark the GPIO shared proxy
descriptors with a flag checked by the new interface.

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20251112-gpio-shared-v4-6-b51f97b1abd8@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpiolib: support shared GPIOs in core subsystem code

As the final step in adding official support for shared GPIOs, enable
the previously added elements in core GPIO subsystem code. Set-up shared
GPIOs when adding a GPIO chip, tear it down on removal and check if a
GPIO descriptor looked up during the firmware-node stage is shared and
fall-back to machine lookup in this case.

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20251112-gpio-shared-v4-5-b51f97b1abd8@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpio: shared-proxy: implement the shared GPIO proxy driver

Add a virtual GPIO proxy driver which arbitrates access to a single
shared GPIO by multiple users. It works together with the core shared
GPIO support from GPIOLIB and functions by acquiring a reference to a
shared GPIO descriptor exposed by gpiolib-shared and making sure that
the state of the GPIO stays consistent.

In general: if there's only one user at the moment: allow it to do
anything as if this was a normal GPIO (in essence: just propagate calls
to the underlying real hardware driver). If there are more users: don't
allow to change the direction set by the initial user, allow to change
configuration options but warn about possible conflicts and finally:
treat the output-high value as a reference counted, logical "GPIO
enabled" setting, meaning: the GPIO value is set to high when the first
user requests it to be high and back to low once the last user stops
"voting" for high.

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20251112-gpio-shared-v4-4-b51f97b1abd8@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpiolib: implement low-level, shared GPIO support

This module scans the device tree (for now only OF nodes are supported
but care is taken to make other fwnode implementations easy to
integrate) and determines which GPIO lines are shared by multiple users.
It stores that information in memory. When the GPIO chip exposing shared
lines is registered, the shared GPIO descriptors it exposes are marked
as shared and virtual "proxy" devices that mediate access to the shared
lines are created. When a consumer of a shared GPIO looks it up, its
fwnode lookup is redirected to a just-in-time machine lookup that points
to this proxy device.

This code can be compiled out on platforms which don't use shared GPIOs.

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20251112-gpio-shared-v4-3-b51f97b1abd8@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

gpiolib: define GPIOD_FLAG_SHARED

Define a new GPIO descriptor flag for marking pins that are shared by
multiple consumer. This flag will be used in several places so we need
to do it in advance and separately from other changes.

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20251112-gpio-shared-v4-2-b51f97b1abd8@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

string: provide strends()

Implement a function for checking if a string ends with a different
string and add its kunit test cases.

Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20251112-gpio-shared-v4-1-b51f97b1abd8@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

Linux 6.18-rc5

Merge tag 'i2c-for-6.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux

Pull i2c fix from Wolfram Sang:
 "Two reverts merged into one commit to handle a regression caused by a
  wrong cleanup because the underlying implications were unclear"

* tag 'i2c-for-6.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: muxes: pca954x: Fix broken reset-gpio usage

Merge tag 'kbuild-fixes-6.18-3' of git://git.kernel.org/pub/scm/linux/kernel/git/kbuild/linux

Pull Kbuild fixes from Nathan Chancellor:

 - Strip trailing padding bytes from modules.builtin.modinfo to fix
   error during modules_install with certain versions of kmod

 - Drop unused static inline function warning in .c files with clang
   from W=1 to W=2

 - Ensure kernel-doc.py invocations use the PYTHON3 make variable to
   ensure user's choice of Python interpreter is always respected

* tag 'kbuild-fixes-6.18-3' of git://git.kernel.org/pub/scm/linux/kernel/git/kbuild/linux:
  kbuild: Let kernel-doc.py use PYTHON3 override
  compiler_types: Move unused static inline functions warning to W=2
  kbuild: Strip trailing padding bytes from modules.builtin.modinfo

kbuild: Let kernel-doc.py use PYTHON3 override

It is possible to force a specific version of python to be used when
building the kernel by passing PYTHON3= on the make command line.
However kernel-doc.py is currently called with python3 hard-coded and
thus ignores this setting.

Use $(PYTHON3) to run $(KERNELDOC) so that the desired version of
python is used.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Reviewed-by: Nicolas Schier <nsc@kernel.org>
Reviewed-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Link: https://patch.msgid.link/20251107192933.2bfe9e57@endymion
Signed-off-by: Nathan Chancellor <nathan@kernel.org>

Merge tag 'drm-fixes-2025-11-09' of https://gitlab.freedesktop.org/drm/kernel

Pull drm fix from Dave Airlie:
 "Brown paper bag, the dma mask fix which I applied and actually looked
  through for bad things, actually broke newer GPUs, there might be some
  latent part in the boot path that is assuming 32-bit still, but we
  will figure that out elsewhere.

  nouveau:
   - revert DMA mask change"

* tag 'drm-fixes-2025-11-09' of https://gitlab.freedesktop.org/drm/kernel:
  Revert "drm/nouveau: set DMA mask before creating the flush page"

Merge tag 'rtc-6.18-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux

Pull RTC fixes from Alexandre Belloni:
 "The two reverts are for patches that I shouldn't have applied. The
  rx8025 patch fixes an issue present since 2022:

   - cpcap, tps6586x: revert incorrect irq enable/disable balance fix

   - rx8025: fix incorrect register reference"

* tag 'rtc-6.18-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux:
  rtc: rx8025: fix incorrect register reference
  Revert "rtc: cpcap: Fix initial enable_irq/disable_irq balance"
  Revert "rtc: tps6586x: Fix initial enable_irq/disable_irq balance"

rtc: rx8025: fix incorrect register reference

This code is intended to operate on the CTRL1 register, but ctrl[1] is
actually CTRL2. Correctly, ctrl[0] is CTRL1.

Signed-off-by: Yuta Hayama <hayama@lineo.co.jp>
Fixes: 71af91565052 ("rtc: rx8025: fix 12/24 hour mode detection on RX-8035")
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/eae5f479-5d28-4a37-859d-d54794e7628c@lineo.co.jp
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Merge tag 'v6.18rc4-SMB-client-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull smb client fixes from Steve French:

 - Fix change notify packet validation check

 - Refcount fix (e.g. rename error paths)

 - Fix potential UAF due to missing locks on directory lease refcount

* tag 'v6.18rc4-SMB-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  smb: client: validate change notify buffer before copy
  smb: client: fix refcount leak in smb2_set_path_attr
  smb: client: fix potential UAF in smb2_close_cached_fid()

Merge tag 'x86-urgent-2025-11-08' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 fixes from Ingo Molnar:

 - Fix AMD PCI root device caching regression that triggers
   on certain firmware variants

 - Fix the zen5_rdseed_microcode[] array to be NULL-terminated

 - Add more AMD models to microcode signature checking

* tag 'x86-urgent-2025-11-08' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/microcode/AMD: Add more known models to entry sign checking
  x86/CPU/AMD: Add missing terminator for zen5_rdseed_microcode
  x86/amd_node: Fix AMD root device caching

Merge tag 'sched-urgent-2025-11-08' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull scheduler fix from Ingo Molnar:
 "Fix a group-throttling bug in the fair scheduler"

* tag 'sched-urgent-2025-11-08' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/fair: Prevent cfs_rq from being unthrottled with zero runtime_remaining

Merge tag 'perf-urgent-2025-11-08' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull perf event fix from Ingo Molnar:
 "Fix a system hang caused by cpu-clock events deadlock"

* tag 'perf-urgent-2025-11-08' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/core: Fix system hang caused by cpu-clock usage

Merge tag 'locking-urgent-2025-11-08' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull locking fix from Ingo Molnar:
 "Fix (well, cut in half) a futex performance regression on PowerPC"

* tag 'locking-urgent-2025-11-08' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  futex: Optimize per-cpu reference counting

Merge tag 'io_uring-6.18-20251107' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux

Pull io_uring fix from Jens Axboe:
 "Single fix in there, fixing an overflow in calculating the needed
  segments for converting into a bvec array"

* tag 'io_uring-6.18-20251107' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux:
  io_uring: fix regbuf vector size truncation

Merge tag 'xfs-fixes-6.18-rc5' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux

Pull xfs fixes from Carlos Maiolino:
 "This contain fixes for the RT and zoned allocator, and a few fixes for
  atomic writes"

* tag 'xfs-fixes-6.18-rc5' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  xfs: free xfs_busy_extents structure when no RT extents are queued
  xfs: fix zone selection in xfs_select_open_zone_mru
  xfs: fix a rtgroup leak when xfs_init_zone fails
  xfs: fix various problems in xfs_atomic_write_cow_iomap_begin
  xfs: fix delalloc write failures in software-provided atomic writes

Revert "drm/nouveau: set DMA mask before creating the flush page"

This reverts commit ebe755605082eddff80eafe0c50915b1366ee98f.

Tested the latest kernel on my GB203 and this seems to break it somehow.

Nov 09 04:16:14 bighp kernel: nouveau 0000:02:00.0: gsp: GSP-FMC boot failed (mbox: 0x0000000b)
Nov 09 04:16:14 bighp kernel: nouveau 0000:02:00.0: gsp: init failed, -5
Nov 09 04:16:14 bighp kernel: nouveau 0000:02:00.0: init failed with -5
Nov 09 04:16:14 bighp kernel: nouveau: drm:00000000:00000080: init failed with -5
Nov 09 04:16:14 bighp kernel: nouveau 0000:02:00.0: drm: Device allocation failed: -5
Nov 09 04:16:14 bighp kernel: nouveau 0000:02:00.0: probe with driver nouveau failed with error -5

Not sure why, I went over the patch and thought it should have worked, but there must be some
32-bit problem maybe in the FMC boot path.

Signed-off-by: Dave Airlie <airlied@redhat.com>

io_uring: fix regbuf vector size truncation

There is a report of io_estimate_bvec_size() truncating the calculated
number of segments that leads to corruption issues. Check it doesn't
overflow "int"s used later. Rough but simple, can be improved on top.

Cc: stable@vger.kernel.org
Fixes: 9ef4cbbcb4ac3 ("io_uring: add infra for importing vectored reg buffers")
Reported-by: Google Big Sleep <big-sleep-vuln-reports+bigsleep-458654612@google.com>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Reviewed-by: Günther Noack <gnoack@google.com>
Tested-by: Günther Noack <gnoack@google.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

Merge tag 'drm-fixes-2025-11-08' of https://gitlab.freedesktop.org/drm/kernel

Pull drm fixes from Dave Airlie:
 "Back from travel, thanks to Simona for handling things. regular fixes,
  seems about the right size, but spread out a bit.

  amdgpu has the usual range of fixes, xe has a few fixes, and nouveau
  has a couple of fixes, one for blackwell modifiers on 8/16 bit
  surfaces.

  Otherwise a few small fixes for mediatek, sched, imagination and
  pixpaper.

  sched:
   - Fix deadlock

  amdgpu:
   - Reset fixes
   - Misc fixes
   - Panel scaling fixes
   - HDMI fix
   - S0ix fixes
   - Hibernation fix
   - Secure display fix
   - Suspend fix
   - MST fix

  amdkfd:
   - Process cleanup fix

  xe:
   - Fix missing  synchronization on unbind
   - Fix device shutdown when doing FLR
   - Fix user fence signaling order

  i915:
   - Avoid lock inversion when pinning to GGTT on CHV/BXT+VTD
   - Fix conversion between clock ticks and nanoseconds

  mediatek:
   - Disable AFBC support on Mediatek DRM driver
   - Add pm_runtime support for GCE power control

  imagination:
   - kconfig: Fix dependencies

  nouveau:
   - Set DMA mask earlier
   - Advertize correct modifiers for GB20x

  pixpaper:
   - kconfig: Fix dependencies"

* tag 'drm-fixes-2025-11-08' of https://gitlab.freedesktop.org/drm/kernel: (26 commits)
  drm/xe: Enforce correct user fence signaling order using
  drm/xe: Do clean shutdown also when using flr
  drm/xe: Move declarations under conditional branch
  drm/xe/guc: Synchronize Dead CT worker with unbind
  drm/amd/display: Enable mst when it's detected but yet to be initialized
  drm/amdgpu: Fix wait after reset sequence in S3
  drm/amd: Fix suspend failure with secure display TA
  drm/amdgpu: fix gpu page fault after hibernation on PF passthrough
  drm/tiny: pixpaper: add explicit dependency on MMU
  drm/nouveau: Advertise correct modifiers on GB20x
  drm: define NVIDIA DRM format modifiers for GB20x
  drm/nouveau: set DMA mask before creating the flush page
  drm/sched: Fix deadlock in drm_sched_entity_kill_jobs_cb
  drm/amd/display: Fix NULL deref in debugfs odm_combine_segments
  drm/amdkfd: Don't clear PT after process killed
  drm/amdgpu/smu: Handle S0ix for vangogh
  drm/amdgpu: Drop PMFW RLC notifier from amdgpu_device_suspend()
  drm/amd/display: Fix black screen with HDMI outputs
  drm/amd/display: Don't stretch non-native images by default in eDP
  drm/amd/pm: fix missing device_attr cleanup in amdgpu_pm_sysfs_init()
  ...

Merge tag 'drm-xe-fixes-2025-11-07' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes

Driver Changes:
 - Fix missing  synchronization on unbind (Balasubramani Vivekanandan)
 - Fix device shutdown when doing FLR (Jouni Högander)
 - Fix user fence signaling order (Matthew Brost)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patch.msgid.link/mvfyflloncy76a7nmkatpj6f2afddavwsibz3y4u4wo6gznro5@rdulkuh5wvje

Merge tag 'parisc-for-6.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux

Pull parisc fix from Helge Deller:

 - fix crash triggered by unaligned access in parisc unwinder

* tag 'parisc-for-6.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: Avoid crash due to unaligned access in unwinder

Merge tag 'for-linus-iommufd' of git://git.kernel.org/pub/scm/linux/kernel/git/jgg/iommufd

Pull iommufd fixes from Jason Gunthorpe:

 - Syzkaller found a case where maths overflows can cause divide by 0

 - Typo in a compiler bug warning fix in the selftests broke the
   selftests

 - type1 compatability had a mismatch when unmapping an already unmapped
   range, it should succeed

* tag 'for-linus-iommufd' of git://git.kernel.org/pub/scm/linux/kernel/git/jgg/iommufd:
  iommufd: Make vfio_compat's unmap succeed if the range is already empty
  iommufd/selftest: Fix ioctl return value in _test_cmd_trigger_vevents()
  iommufd: Don't overflow during division for dirty tracking

compiler_types: Move unused static inline functions warning to W=2

Per Nathan, clang catches unused "static inline" functions in C files
since commit 6863f5643dd7 ("kbuild: allow Clang to find unused static
inline functions for W=1 build").

Linus said:

> So I entirely ignore W=1 issues, because I think so many of the extra
> warnings are bogus.
>
> But if this one in particular is causing more problems than most -
> some teams do seem to use W=1 as part of their test builds - it's fine
> to send me a patch that just moves bad warnings to W=2.
>
> And if anybody uses W=2 for their test builds, that's THEIR problem..

Here is the change to bump the warning from W=1 to W=2.

Fixes: 6863f5643dd7 ("kbuild: allow Clang to find unused static inline functions for W=1 build")
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://patch.msgid.link/20251106105000.2103276-1-andriy.shevchenko@linux.intel.com
[nathan: Adjust comment as well]
Signed-off-by: Nathan Chancellor <nathan@kernel.org>

smb: client: validate change notify buffer before copy

SMB2_change_notify called smb2_validate_iov() but ignored the return
code, then kmemdup()ed using server provided OutputBufferOffset/Length.

Check the return of smb2_validate_iov() and bail out on error.

Discovered with help from the ZeroPath security tooling.

Signed-off-by: Joshua Rogers <linux@joshua.hu>
Reviewed-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Cc: stable@vger.kernel.org
Fixes: e3e9463414f61 ("smb3: improve SMB3 change notification support")
Signed-off-by: Steve French <stfrench@microsoft.com>

Merge tag 'gpio-fixes-for-v6.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux

Pull gpio fixes from Bartosz Golaszewski:

 - use the firmware node of the GPIO chip, not its label for software
   node lookup

 - fix invalid pointer access in GPIO debugfs

 - drop unused functions from gpio-tb10x

 - fix a regression in gpio-aggregator: restore the set_config()
   callback in the driver

 - correct schema $id path in ti,twl4030 DT bindings

* tag 'gpio-fixes-for-v6.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
  gpio: tb10x: Drop unused tb10x_set_bits() function
  gpio: aggregator: restore the set_config operation
  gpiolib: fix invalid pointer access in debugfs
  gpio: swnode: don't use the swnode's name as the key for GPIO lookup
  dt-bindings: gpio: ti,twl4030: Correct the schema $id path

Merge tag 'trace-v6.18-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace

Pull tracing fixes from Steven Rostedt:

 - Check for reader catching up in ring_buffer_map_get_reader()

   If the reader catches up to the writer in the memory mapped ring
   buffer then calling rb_get_reader_page() will return NULL as there's
   no pages left. But this isn't checked for before calling
   rb_get_reader_page() and the return of NULL causes a warning.

   If it is detected that the reader caught up to the writer, then
   simply exit the routine

 - Fix memory leak in histogram create_field_var()

   The couple of the error paths in create_field_var() did not properly
   clean up what was allocated. Make sure everything is freed properly
   on error

 - Fix help message of tools latency_collector

   The help message incorrectly stated that "-t" was the same as
   "--threads" whereas "--threads" is actually represented by "-e"

* tag 'trace-v6.18-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  tracing/tools: Fix incorrcet short option in usage text for --threads
  tracing: Fix memory leaks in create_field_var()
  ring-buffer: Do not warn in ring_buffer_map_get_reader() when reader catches up

Merge tag 'slab-for-6.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/vbabka/slab

Pull slab fix from Vlastimil Babka:

 - Fix for potential infinite loop in kmalloc_nolock() when debugging
   is enabled for the cache (Vlastimil Babka)

* tag 'slab-for-6.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/vbabka/slab:
  slab: prevent infinite loop in kmalloc_nolock() with debugging

Merge tag 'io_uring-6.18-20251106' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux

Pull io_uring fixes from Jens Axboe:

 - Remove the sync refill API that was added in this release, in
   anticipation of doing it in a better way for the next release

 - Fix type extension for calculating size off nr_pages, like we do
   in other spots

* tag 'io_uring-6.18-20251106' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux:
  io_uring: fix types for region size calulation
  io_uring/zcrx: remove sync refill uapi

Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "All fixes in the UFS driver.

  The big contributor to the diffstats is the Intel controller S0ix/S3
  fix which has to special case the suspend/resume patch for intel
  controllers in ufshcd-pci.c"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: ufs: core: Fix invalid probe error return value
  scsi: ufs: ufs-pci: Set UFSHCD_QUIRK_PERFORM_LINK_STARTUP_ONCE for Intel ADL
  scsi: ufs: core: Add a quirk to suppress link_startup_again
  scsi: ufs: ufs-pci: Fix S0ix/S3 for Intel controllers
  scsi: ufs: core: Revert "Make HID attributes visible"
  scsi: ufs: core: Reduce link startup failure logging
  scsi: ufs: core: Fix a race condition related to the "hid" attribute group
  scsi: ufs: ufs-qcom: Fix UFS OCP issue during UFS power down (PC=3)

Merge tag 'v6.18-rc4-smb-server-fixes' of git://git.samba.org/ksmbd

Pull smb server fixes from Steve French:

 - More safely detect RDMA capable devices correctly

* tag 'v6.18-rc4-smb-server-fixes' of git://git.samba.org/ksmbd:
  ksmbd: detect RDMA capable netdevs include IPoIB
  ksmbd: detect RDMA capable lower devices when bridge and vlan netdev is used

tracing/tools: Fix incorrcet short option in usage text for --threads

The help message incorrectly listed '-t' as the short option for
--threads, but the actual getopt_long configuration uses '-e'.
This mismatch can confuse users and lead to incorrect command-line
usage. This patch updates the usage string to correctly show:
"-e, --threads NRTHR"
to match the implementation.

Note: checkpatch.pl reports a false-positive spelling warning on
'Run', which is intentional.

Link: https://patch.msgid.link/20251106031040.1869-1-zhangchujun@cmss.chinamobile.com
Signed-off-by: Zhang Chujun <zhangchujun@cmss.chinamobile.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>

drm/xe: Enforce correct user fence signaling order using

Prevent application hangs caused by out-of-order fence signaling when
user fences are attached. Use drm_syncobj (via dma-fence-chain) to
guarantee that each user fence signals in order, regardless of the
signaling order of the attached fences. Ensure user fence writebacks to
user space occur in the correct sequence.

v7:
 - Skip drm_syncbj create of error (CI)

Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Link: https://patch.msgid.link/20251031234050.3043507-2-matthew.brost@intel.com
(cherry picked from commit adda4e855ab6409a3edaa585293f1f2069ab7299)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

x86/microcode/AMD: Add more known models to entry sign checking

Two Zen5 systems are missing from need_sha_check(). Add them.

Fixes: 50cef76d5cb0 ("x86/microcode/AMD: Load only SHA256-checksummed patches")
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: <stable@kernel.org>
Link: https://patch.msgid.link/20251106182904.4143757-1-superm1@kernel.org

drm/xe: Do clean shutdown also when using flr

Currently Xe driver is triggering flr without any clean-up on
shutdown. This is causing random warnings from pending related works as the
underlying hardware is reset in the middle of their execution.

Fix this by performing clean shutdown also when using flr.

Fixes: 501d799a47e2 ("drm/xe: Wire up device shutdown handler")
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Jouni Högander <jouni.hogander@intel.com>
Reviewed-by: Maarten Lankhorst <dev@lankhorst.se>
Link: https://patch.msgid.link/20251031122312.1836534-1-jouni.hogander@intel.com
Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
(cherry picked from commit a4ff26b7c8ef38e4dd34f77cbcd73576fdde6dd4)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

drm/xe: Move declarations under conditional branch

The xe_device_shutdown() function was needing a few declarations
that were only required under a specific condition. This change
moves those declarations to be within that conditional branch
to avoid unnecessary declarations.

Reviewed-by: Nitin Gote <nitin.r.gote@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20251007100208.1407021-1-tejas.upadhyay@intel.com
Signed-off-by: Tejas Upadhyay <tejas.upadhyay@intel.com>
(cherry picked from commit 15b3036045188f4da4ca62b2ed01b0f160252e9b)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

drm/xe/guc: Synchronize Dead CT worker with unbind

Cancel and wait for any Dead CT worker to complete before continuing
with device unbinding. Else the worker will end up using resources freed
by the undind operation.

Cc: Zhanjun Dong <zhanjun.dong@intel.com>
Fixes: d2c5a5a926f4 ("drm/xe/guc: Dead CT helper")
Signed-off-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Reviewed-by: Stuart Summers <stuart.summers@intel.com>
Link: https://patch.msgid.link/20251103123144.3231829-6-balasubramani.vivekanandan@intel.com
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
(cherry picked from commit 492671339114e376aaa38626d637a2751cdef263)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

Merge tag 'mediatek-drm-fixes-20251105' of https://git.kernel.org/pub/scm/linux/kernel/git/chunkuang.hu/linux into drm-fixes

Mediatek DRM Fixes - 20251105

1. Disable AFBC support on Mediatek DRM driver
2. Add pm_runtime support for GCE power control

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Chun-Kuang Hu <chunkuang.hu@kernel.org>
Link: https://patch.msgid.link/20251105151443.3909-1-chunkuang.hu@kernel.org

tracing: Fix memory leaks in create_field_var()

The function create_field_var() allocates memory for 'val' through
create_hist_field() inside parse_atom(), and for 'var' through
create_var(), which in turn allocates var->type and var->var.name
internally. Simply calling kfree() to release these structures will
result in memory leaks.

Use destroy_hist_field() to properly free 'val', and explicitly release
the memory of var->type and var->var.name before freeing 'var' itself.

Link: https://patch.msgid.link/20251106120132.3639920-1-zilin@seu.edu.cn
Fixes: 02205a6752f22 ("tracing: Add support for 'field variables'")
Signed-off-by: Zilin Guan <zilin@seu.edu.cn>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>