[PATCH] pinctrl: meson: amlogic-a4: mark the GPIO controller as sleeping

[PATCH] pinctrl: meson: amlogic-a4: mark the GPIO controller as sleeping

[Bartosz Golaszewski]

The GPIO controller is configured as non-sleeping but it uses generic
pinctrl helpers which use a mutex for synchronization. This will cause
lockdep splats when used together with shared GPIOs going through the
GPIO shared proxy driver.

Fixes: 6e9be3abb78c ("pinctrl: Add driver support for Amlogic SoCs")
Cc: stable@vger.kernel.org
Reported-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Closes: https://lore.kernel.org/all/CAFBinCAc7CO8gfNQakCu3LfkYXuyTd2iRpMRm8EKXSL0mwOnJw@mail.gmail.com/
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
---
 drivers/pinctrl/meson/pinctrl-amlogic-a4.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/pinctrl/meson/pinctrl-amlogic-a4.c b/drivers/pinctrl/meson/pinctrl-amlogic-a4.c
index d9e3a8d5932a..ded7b218e2ec 100644
--- a/drivers/pinctrl/meson/pinctrl-amlogic-a4.c
+++ b/drivers/pinctrl/meson/pinctrl-amlogic-a4.c
@@ -893,7 +893,7 @@ static const struct gpio_chip aml_gpio_template = {
 	.direction_input	= aml_gpio_direction_input,
 	.direction_output	= aml_gpio_direction_output,
 	.get_direction		= aml_gpio_get_direction,
-	.can_sleep		= false,
+	.can_sleep		= true,
 };
 
 static void init_bank_register_bit(struct aml_pinctrl *info,
-- 
2.47.3

Re: [PATCH] pinctrl: meson: amlogic-a4: mark the GPIO controller as sleeping

[Neil Armstrong]

On 1/6/26 09:52, Bartosz Golaszewski wrote:
> The GPIO controller is configured as non-sleeping but it uses generic
> pinctrl helpers which use a mutex for synchronization. This will cause
> lockdep splats when used together with shared GPIOs going through the
> GPIO shared proxy driver.
> 
> Fixes: 6e9be3abb78c ("pinctrl: Add driver support for Amlogic SoCs")
> Cc: stable@vger.kernel.org
> Reported-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
> Closes: https://lore.kernel.org/all/CAFBinCAc7CO8gfNQakCu3LfkYXuyTd2iRpMRm8EKXSL0mwOnJw@mail.gmail.com/
> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
> ---
>   drivers/pinctrl/meson/pinctrl-amlogic-a4.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/pinctrl/meson/pinctrl-amlogic-a4.c b/drivers/pinctrl/meson/pinctrl-amlogic-a4.c
> index d9e3a8d5932a..ded7b218e2ec 100644
> --- a/drivers/pinctrl/meson/pinctrl-amlogic-a4.c
> +++ b/drivers/pinctrl/meson/pinctrl-amlogic-a4.c
> @@ -893,7 +893,7 @@ static const struct gpio_chip aml_gpio_template = {
>   	.direction_input	= aml_gpio_direction_input,
>   	.direction_output	= aml_gpio_direction_output,
>   	.get_direction		= aml_gpio_get_direction,
> -	.can_sleep		= false,
> +	.can_sleep		= true,
>   };
>   
>   static void init_bank_register_bit(struct aml_pinctrl *info,

Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>

Thanks,
Neil

Re: [PATCH] pinctrl: meson: amlogic-a4: mark the GPIO controller as sleeping

[Xianwei Zhao]

Hi Bartosz,
    Could you explain the reproduction method for this problem ?

On 2026/1/6 16:55, Neil Armstrong wrote:
> The GPIO controller is configured as non-sleeping but it uses generic
> pinctrl helpers which use a mutex for synchronization. This will cause
> lockdep splats when used together with shared GPIOs going through the
> GPIO shared proxy driver.
> 
> Fixes: 6e9be3abb78c ("pinctrl: Add driver support for Amlogic SoCs")
> Cc: stable@vger.kernel.org
> Reported-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
> Closes: 
> https://lore.kernel.org/all/CAFBinCAc7CO8gfNQakCu3LfkYXuyTd2iRpMRm8EKXSL0mwOnJw@mail.gmail.com/
> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
> ---
>    drivers/pinctrl/meson/pinctrl-amlogic-a4.c | 2 +-
>    1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/pinctrl/meson/pinctrl-amlogic-a4.c 
> b/drivers/pinctrl/meson/pinctrl-amlogic-a4.c
> index d9e3a8d5932a..ded7b218e2ec 100644
> --- a/drivers/pinctrl/meson/pinctrl-amlogic-a4.c
> +++ b/drivers/pinctrl/meson/pinctrl-amlogic-a4.c
> @@ -893,7 +893,7 @@ static const struct gpio_chip aml_gpio_template = {
>        .direction_input        = aml_gpio_direction_input,
>        .direction_output       = aml_gpio_direction_output,
>        .get_direction          = aml_gpio_get_direction,
> -     .can_sleep              = false,
> +     .can_sleep              = true,
>    };
> 
>    static void init_bank_register_bit(struct aml_pinctrl *info,

Re: [PATCH] pinctrl: meson: amlogic-a4: mark the GPIO controller as sleeping

[Bartosz Golaszewski]

On Tue, Jan 6, 2026 at 10:32 AM Xianwei Zhao <xianwei.zhao@amlogic.com> wrote:
>
> Hi Bartosz,
>     Could you explain the reproduction method for this problem ?
>

Please don't top-post on the mailing list.

If you have a pin shared by multiple devices in device-tree, it will
now go through a dedicated gpio-shared-proxy driver. It will use a
locking mechanism depending on the can_sleep flag of the GPIO
controller device - spinlock or mutex. If the flag is wrong - like is
the case here - it may happen that we take a spinlock and then descend
into pinctrl where a mutex is taken resulting in a lockdep splat.

Bart

Re: [PATCH] pinctrl: meson: amlogic-a4: mark the GPIO controller as sleeping

[Martin Blumenstingl]

On Tue, Jan 6, 2026 at 9:53 AM Bartosz Golaszewski
<bartosz.golaszewski@oss.qualcomm.com> wrote:
>
> The GPIO controller is configured as non-sleeping but it uses generic
> pinctrl helpers which use a mutex for synchronization. This will cause
> lockdep splats when used together with shared GPIOs going through the
> GPIO shared proxy driver.
>
> Fixes: 6e9be3abb78c ("pinctrl: Add driver support for Amlogic SoCs")
> Cc: stable@vger.kernel.org
> Reported-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
> Closes: https://lore.kernel.org/all/CAFBinCAc7CO8gfNQakCu3LfkYXuyTd2iRpMRm8EKXSL0mwOnJw@mail.gmail.com/
> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>

Re: [PATCH] pinctrl: meson: amlogic-a4: mark the GPIO controller as sleeping

[Linus Walleij]

On Tue, Jan 6, 2026 at 9:53 AM Bartosz Golaszewski
<bartosz.golaszewski@oss.qualcomm.com> wrote:

> The GPIO controller is configured as non-sleeping but it uses generic
> pinctrl helpers which use a mutex for synchronization. This will cause
> lockdep splats when used together with shared GPIOs going through the
> GPIO shared proxy driver.
>
> Fixes: 6e9be3abb78c ("pinctrl: Add driver support for Amlogic SoCs")
> Cc: stable@vger.kernel.org
> Reported-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
> Closes: https://lore.kernel.org/all/CAFBinCAc7CO8gfNQakCu3LfkYXuyTd2iRpMRm8EKXSL0mwOnJw@mail.gmail.com/
> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

Patch applied for fixes!

Yours,
Linus Walleij

Re: [PATCH] pinctrl: meson: amlogic-a4: mark the GPIO controller as sleeping

[Linus Walleij]

On Tue, Jan 6, 2026 at 9:53 AM Bartosz Golaszewski
<bartosz.golaszewski@oss.qualcomm.com> wrote:

> The GPIO controller is configured as non-sleeping but it uses generic
> pinctrl helpers which use a mutex for synchronization. This will cause
> lockdep splats when used together with shared GPIOs going through the
> GPIO shared proxy driver.

BTW it is a very *welcome* side-effect of the new shared GPIO rework
that it just happens to find these old bugs, I'm very pleased with this :)

Yours,
Linus Walleij

Re: [PATCH] pinctrl: meson: amlogic-a4: mark the GPIO controller as sleeping

[Bartosz Golaszewski]

On Fri, Jan 9, 2026 at 10:35 AM Linus Walleij <linusw@kernel.org> wrote:
>
> On Tue, Jan 6, 2026 at 9:53 AM Bartosz Golaszewski
> <bartosz.golaszewski@oss.qualcomm.com> wrote:
>
> > The GPIO controller is configured as non-sleeping but it uses generic
> > pinctrl helpers which use a mutex for synchronization. This will cause
> > lockdep splats when used together with shared GPIOs going through the
> > GPIO shared proxy driver.
>
> BTW it is a very *welcome* side-effect of the new shared GPIO rework
> that it just happens to find these old bugs, I'm very pleased with this :)
>

Actually, I was surprised I was getting so many bug reports and then
it turned out, there was a bug that caused a lot of false-positives to
turn up when scanning the devicetree for shared pins. I just queued a
fix for that so it should slow down now.

Bart