Re: [PATCH 4/8] gpiolib: remove gpio_set_debounce

[Arnd Bergmann]

On Thu, Jan 26, 2023, at 14:50, Andy Shevchenko wrote:
> On Thu, Jan 26, 2023 at 02:27:57PM +0100, Arnd Bergmann wrote:
>> From: Arnd Bergmann <arnd@xxxxxxxx>
>> 
>> gpio_set_debounce() only has a single user, which is trivially
>> converted to gpiod_set_debounce().
>
> Also need to mention the gpio_cansleep() removal.

Indeed, I think that bit got lost in one of the rebases
after some of the other functions that I originally
removed here are already gone.

I've dropped the gpio_cansleep change for now, we can
revisit this later and possibly drop it at the same
time as gpio_{get,set}_value_cansleep that are still
used in a handful of places that could use gpiod
instead.

> P.S. Fun fact that I was yesterday thinking about the pretty much
> the same change, except intrusive conversion of the user.
>
> ...
>
>>  Documentation/driver-api/gpio/legacy.rst |  2 --
>
> Documentation/translations/zh_CN/driver-api/gpio/legacy.rst:222:        
>         gpio_set_debounce()
> Documentation/translations/zh_TW/gpio.txt:229:  gpio_set_debounce()

Fixed.

>>  drivers/input/touchscreen/ads7846.c      | 24 +++++++++++++-----------
>>  include/linux/gpio.h                     | 22 ----------------------
>
> ...
>
>>  #include <linux/of_gpio.h>
>
> Do we need this?

Dropped now.

>>  #include <linux/of_device.h>
>> +#include <linux/gpio/consumer.h>
>
>>  #include <linux/gpio.h>
>
> And this?

I think we still need both linux/gpio/consumer.h and linux/gpio.h
since there is still a call to devm_gpio_request_one() and
gpio_is_valid().

    Arnd