Re: [PATCH 1/3] gpiolib: add gpiod_to_gpio_device() stub for !GPIOLIB

[Krzysztof Kozlowski]

On 25/01/2024 09:36, Bartosz Golaszewski wrote:
> On Thu, Jan 25, 2024 at 9:16 AM Krzysztof Kozlowski
> <krzysztof.kozlowski@xxxxxxxxxx> wrote:
>>
>> Add empty stub of gpiod_to_gpio_device() when GPIOLIB is not enabled.
>>
>> Cc: <stable@xxxxxxxxxxxxxxx>
>> Fixes: 370232d096e3 ("gpiolib: provide gpiod_to_gpio_device()")
>> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
>>
>> ---
>>
>> Cc: Philipp Zabel <p.zabel@xxxxxxxxxxxxxx>
>> ---
>>  include/linux/gpio/driver.h | 6 ++++++
>>  1 file changed, 6 insertions(+)
>>
>> diff --git a/include/linux/gpio/driver.h b/include/linux/gpio/driver.h
>> index 9a5c6c76e653..012797e7106d 100644
>> --- a/include/linux/gpio/driver.h
>> +++ b/include/linux/gpio/driver.h
>> @@ -819,6 +819,12 @@ static inline struct gpio_chip *gpiod_to_chip(const struct gpio_desc *desc)
>>         return ERR_PTR(-ENODEV);
>>  }
>>
>> +static inline struct gpio_device *gpiod_to_gpio_device(struct gpio_desc *desc)
>> +{
>> +       WARN_ON(1);
>> +       return ERR_PTR(-ENODEV);
>> +}
>> +
>>  static inline int gpiochip_lock_as_irq(struct gpio_chip *gc,
>>                                        unsigned int offset)
>>  {
>> --
>> 2.34.1
>>
> 
> Why is this needed? Users of gpio/driver.h should select GPIOLIB.

The third patch shows you the user which will not select GPIOLIB. Why?
Because there is no hard dependency between one core framework (RESET)
on other core framework (GPIOLIB).

The first two patches are added for the same purpose, even though there
is no need currently.


Best regards,
Krzysztof