[v1,3/6] leds: gpio: Avoid using GPIOF_ACTIVE_LOW

Avoid using GPIOF_ACTIVE_LOW as it's deprecated and subject to remove.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 drivers/leds/leds-gpio.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

On Mon, Nov 4, 2024 at 10:36 AM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:

> Avoid using GPIOF_ACTIVE_LOW as it's deprecated and subject to remove.
>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

Yours,
Linus Walleij

Hi Andy,

On Mon, Nov 4, 2024 at 10:37 AM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
> Avoid using GPIOF_ACTIVE_LOW as it's deprecated and subject to remove.
>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Thanks for your patch, which is now commit e6a2f0ea519fd247
("leds: gpio: Avoid using GPIOF_ACTIVE_LOW") in gpio/gpio/for-next.

> --- a/drivers/leds/leds-gpio.c
> +++ b/drivers/leds/leds-gpio.c
> @@ -212,7 +212,6 @@ static struct gpio_desc *gpio_led_get_gpiod(struct device *dev, int idx,
>                                             const struct gpio_led *template)
>  {
>         struct gpio_desc *gpiod;
> -       unsigned long flags = GPIOF_OUT_INIT_LOW;
>         int ret;
>
>         /*
> @@ -239,10 +238,7 @@ static struct gpio_desc *gpio_led_get_gpiod(struct device *dev, int idx,
>         if (!gpio_is_valid(template->gpio))
>                 return ERR_PTR(-ENOENT);
>
> -       if (template->active_low)
> -               flags |= GPIOF_ACTIVE_LOW;
> -
> -       ret = devm_gpio_request_one(dev, template->gpio, flags,
> +       ret = devm_gpio_request_one(dev, template->gpio, GPIOF_OUT_INIT_LOW,
>                                     template->name);

Just wondering, as I am not 100% sure: can this change change the
initial state of the GPIO?

>         if (ret < 0)
>                 return ERR_PTR(ret);
> @@ -251,6 +247,9 @@ static struct gpio_desc *gpio_led_get_gpiod(struct device *dev, int idx,
>         if (!gpiod)
>                 return ERR_PTR(-EINVAL);
>
> +       if (template->active_low ^ gpiod_is_active_low(gpiod))
> +               gpiod_toggle_active_low(gpiod);
> +
>         return gpiod;
>  }

Gr{oetje,eeting}s,

                        Geert

On Mon, Nov 11, 2024 at 10:45:13AM +0100, Geert Uytterhoeven wrote:
> On Mon, Nov 4, 2024 at 10:37 AM Andy Shevchenko
> <andriy.shevchenko@linux.intel.com> wrote:

...

> > -       if (template->active_low)
> > -               flags |= GPIOF_ACTIVE_LOW;
> > -
> > -       ret = devm_gpio_request_one(dev, template->gpio, flags,
> > +       ret = devm_gpio_request_one(dev, template->gpio, GPIOF_OUT_INIT_LOW,
> >                                     template->name);
> 
> Just wondering, as I am not 100% sure: can this change change the
> initial state of the GPIO?

You probably wonder how ACTIVE_LOW affects the OUT_INIT_LOW given above.
I have an answer to you, however I might be mistaken as well, but I spent some
time to investigate.

The above mentioned call ends up in the gpiod_direction_output_raw_commit() which
uses the value (low in this case) as an absolute value. It does not include the
ACTIVE_LOW in the value calculations. Hence, setting ACTIVE_LOW before or afterwards
has no effect on the existing flow.

If you notice a mistake, please elaborate this, so I can fix the approach!

> >         if (ret < 0)
> >                 return ERR_PTR(ret);

...

> > +       if (template->active_low ^ gpiod_is_active_low(gpiod))
> > +               gpiod_toggle_active_low(gpiod);
> > +

Hi Andy,

On Mon, Nov 11, 2024 at 10:57 AM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
> On Mon, Nov 11, 2024 at 10:45:13AM +0100, Geert Uytterhoeven wrote:
> > On Mon, Nov 4, 2024 at 10:37 AM Andy Shevchenko
> > <andriy.shevchenko@linux.intel.com> wrote:
>
> ...
>
> > > -       if (template->active_low)
> > > -               flags |= GPIOF_ACTIVE_LOW;
> > > -
> > > -       ret = devm_gpio_request_one(dev, template->gpio, flags,
> > > +       ret = devm_gpio_request_one(dev, template->gpio, GPIOF_OUT_INIT_LOW,
> > >                                     template->name);
> >
> > Just wondering, as I am not 100% sure: can this change change the
> > initial state of the GPIO?
>
> You probably wonder how ACTIVE_LOW affects the OUT_INIT_LOW given above.
> I have an answer to you, however I might be mistaken as well, but I spent some
> time to investigate.
>
> The above mentioned call ends up in the gpiod_direction_output_raw_commit() which
> uses the value (low in this case) as an absolute value. It does not include the
> ACTIVE_LOW in the value calculations. Hence, setting ACTIVE_LOW before or afterwards
> has no effect on the existing flow.
>
> If you notice a mistake, please elaborate this, so I can fix the approach!

Thanks, I had discovered the same, but wanted to double-check!

Gr{oetje,eeting}s,

                        Geert