Re: [PATCH] gpio: rockchip: mark the GPIO controller as sleeping

["Diederik de Haas" <diederik@cknow-tech.com>]

On Fri Jan 23, 2026 at 2:27 PM CET, Robin Murphy wrote:
> On 2026-01-12 9:08 am, Bartosz Golaszewski wrote:
>> On Sat, Jan 10, 2026 at 12:55 AM Sebastian Reichel
>> <sebastian.reichel@collabora.com> wrote:
>>> On Tue, Jan 06, 2026 at 10:00:11AM +0100, Bartosz Golaszewski wrote:
>>>> The GPIO controller is configured as non-sleeping but it uses generic
>>>> pinctrl helpers which use a mutex for synchronization.
>>>>
>>>> This can cause the following lockdep splat with shared GPIOs enabled on
>>>> boards which have multiple devices using the same GPIO:
>>>>
>> 
>> [snip]
>> 
>>>>
>>>> Fixes: 936ee2675eee ("gpio/rockchip: add driver for rockchip gpio")
>>>> Cc: stable@vger.kernel.org
>>>> Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
>>>> Closes: https://lore.kernel.org/all/d035fc29-3b03-4cd6-b8ec-001f93540bc6@samsung.com/
>>>> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
>>>> ---
>>>>   drivers/gpio/gpio-rockchip.c | 1 +
>>>>   1 file changed, 1 insertion(+)
>>>>
>>>> diff --git a/drivers/gpio/gpio-rockchip.c b/drivers/gpio/gpio-rockchip.c
>>>> index 47174eb3ba76..bae2061f15fc 100644
>>>> --- a/drivers/gpio/gpio-rockchip.c
>>>> +++ b/drivers/gpio/gpio-rockchip.c
>>>> @@ -593,6 +593,7 @@ static int rockchip_gpiolib_register(struct rockchip_pin_bank *bank)
>>>>        gc->ngpio = bank->nr_pins;
>>>>        gc->label = bank->name;
>>>>        gc->parent = bank->dev;
>>>> +     gc->can_sleep = true;
>>>
>>> This means all operations are marked as can_sleep, even though
>>> pinctrl operations are only used for the direction setting.
>>> I.e. the common get/set operations always worked in atomic mode,
>>> but now complain. See for example:
>>>
>>> https://lore.kernel.org/all/20260108-media-synopsys-hdmirx-fix-gpio-cansleep-v1-1-3570518d8bab@kernel.org/
>>>
>>> It's not a big issue for the hdmirx driver specifically, but I wonder
>>> how many more (less often tested) rockchip drivers use GPIOs from their
>>> IRQ handler.
>
> Yeah, seems this finally reached my distro kernel and now the kernel log 
> on one of my boards is totally flooded from gpio_ir_recv_irq() 
> (legitimately) calling gpio_get_value()... that's not really OK :/

Yeah, I'm getting it too on several of my boards, like on Rock64:
https://paste.sr.ht/~diederik/154c5023a3a50d77f1da2195e7bb9a96f6a88555
(that's just a fraction as dmesg ran out of its buffer ...)
Also mentioned here:
https://lore.kernel.org/all/DFOEGOTI1AQ9.175GP7V1VK1XU@cknow-tech.com/

Diederik

>
> Thanks,
> Robin.
>
>>> Considering setting or getting the GPIO from atomic context is much
>>> more common than changing the direction - is there some way to
>>> describe the sleep behavior in a more specific way in the GPIO
>>> controller?
>>>
>> 
>> No, there's no such switch at the moment. This is because there are
>> paths that we can take, where we *do* end up setting direction from
>> gpiod_set_value(). For instance:
>> 
>> gpiod_set_value()
>>    gpiod_set_value_nocheck()
>>      gpio_set_open_drain_value_commit()
>>        gpiochip_direction_output()
>> 
>> I'm afraid, for correctness, it has to be either sleeping, or not. I
>> would love - at some point - to make pinctrl mostly lockless with
>> SRCU, like we did with GPIO. That would solve this issue correctly.
>> But until then, I'm afraid we need to keep a chip-global switch for
>> sleeping.
>> 
>> Bartosz
>> 
>> _______________________________________________
>> Linux-rockchip mailing list
>> Linux-rockchip@lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/linux-rockchip
>
>
> _______________________________________________
> Linux-rockchip mailing list
> Linux-rockchip@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-rockchip