Re: [RFC v2 4/5] gpio: add pinctrl based generic gpio driver

[AKASHI Takahiro <takahiro.akashi@xxxxxxxxxx>]

Hi Linus and Oleksii,

On Tue, Oct 10, 2023 at 02:00:40PM +0200, Linus Walleij wrote:
> On Thu, Oct 5, 2023 at 4:59???AM AKASHI Takahiro
> <takahiro.akashi@xxxxxxxxxx> wrote:
> 
> 
> > Some pin controllers provide not only a method to set up lines but
> > also gpio function. With this commit, a new generic gpio driver will
> > be provided. It is implemented purely by using pinctrl interfaces.
> > One of such pin controllers is Arm's SCMI.
> >
> > Signed-off-by: AKASHI Takahiro <takahiro.akashi@xxxxxxxxxx>
> > ---
> > RFC v2 (Oct 5, 2023)
> 
> RFC v2 looks very good to me, definitely something that can be merged
> as a starting point once the hardware has been tested.

Thank you for your support.
I think the easiest and best way to test the code is that Oleskii will try
my patch on his platform, r-car, on which I believe that SCMI FW for pin
controller is already available since he tested his pinctrl driver.

@Oleskii, can you please take a time for the test?
(I will assist you in case of any error.)

> > +static int pin_control_gpio_direction_input(struct gpio_chip *chip,
> > +                                           unsigned int offset)
> > +{
> > +       return pinctrl_gpio_direction_input(chip->gpiodev->base + offset);
> > +}
> > +
> > +static int pin_control_gpio_direction_output(struct gpio_chip *chip,
> > +                                            unsigned int offset, int val)
> > +{
> > +       return pinctrl_gpio_direction_output(chip->gpiodev->base + offset);
> > +}
> 
> IIRC Bartosz is working on a patch set getting rid of this kludge having to
> call with base + offset in every driver, replacing it with generic calls that
> you can just assign in the gpio_chip.
> 
> When this gets applied these changes will likely be in place so you will
> get rid of this too.

I will try to keep eyes on Bartosz's patch.

Thanks,
-Takahiro Akashi


> Yours,
> Linus Walleij