Re: [PATCH] leds: leds-gpio: Convert gpio_blink_set() to use GPIO descriptors

Re: [PATCH] leds: leds-gpio: Convert gpio_blink_set() to use GPIO descriptors

[Geert Uytterhoeven]

Hi Mika, Rafael,

On Tue, Nov 4, 2014 at 11:50 PM, Rafael J. Wysocki <rjw@rjwysocki.net> wrote:
> On Tuesday, November 04, 2014 12:10:41 PM Alexandre Courbot wrote:
>> On 10/31/2014 08:40 PM, Mika Westerberg wrote:
>> > Commit 21f2aae91e902aad ("leds: leds-gpio: Add support for GPIO
>> > descriptors") already converted most of the driver to use GPIO descriptors.
>> > What is still missing is the platform specific hook gpio_blink_set() and
>> > board files which pass legacy GPIO numbers to this driver in platform data.
>> >
>> > In this patch we handle the former and convert gpio_blink_set() to take
>> > GPIO descriptor instead. In order to do this we convert the existing four
>> > users to accept GPIO descriptor and translate it to legacy GPIO number in
>> > the platform code. This effectively "pushes" legacy GPIO number usage from
>> > the driver to platforms.
>> >
>> > Also add comment to the remaining block describing that it is legacy code
>> > path and we are getting rid of it eventually.
>> >
>> > Suggested-by: Linus Walleij <linus.walleij@linaro.org>
>> > Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
>>
>> Acked-by: Alexandre Courbot <acourbot@nvidia.com>
>
> Patch applied, thanks everyone!

"leds: leds-gpio: Add support for GPIO descriptors" broke leds-gpio on
non-DT platforms for me:

    gpiod_direction_output: invalid GPIO
    leds-gpio: probe of leds-gpio failed with error -22

(desc is NULL in gpiod_direction_output()).

DT shmobile reference/multi-platform are fine.

I noticed the hard way, as I wanted to add some LEDs to a new platform,
but couldn't get it work. It turned out it also had stopped working on
r8a7740/armadillo-legacy, so I started bisecting...

Unfortunately the offending patch can't just be reverted.
Reverting all three of these on pm/linux-next fixed the issue, though:
c673a2b400810352 leds: leds-gpio: Convert gpio_blink_set() to use GPIO descripto
a43f2cbbb009f962 leds: leds-gpio: Make use of device property API
5c51277a9ababfa4 leds: leds-gpio: Add support for GPIO descriptors

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH] leds: leds-gpio: Convert gpio_blink_set() to use GPIO descriptors

[Mika Westerberg]

On Thu, Nov 06, 2014 at 10:52:35AM +0100, Geert Uytterhoeven wrote:
> Hi Mika, Rafael,
> 
> On Tue, Nov 4, 2014 at 11:50 PM, Rafael J. Wysocki <rjw@rjwysocki.net> wrote:
> > On Tuesday, November 04, 2014 12:10:41 PM Alexandre Courbot wrote:
> >> On 10/31/2014 08:40 PM, Mika Westerberg wrote:
> >> > Commit 21f2aae91e902aad ("leds: leds-gpio: Add support for GPIO
> >> > descriptors") already converted most of the driver to use GPIO descriptors.
> >> > What is still missing is the platform specific hook gpio_blink_set() and
> >> > board files which pass legacy GPIO numbers to this driver in platform data.
> >> >
> >> > In this patch we handle the former and convert gpio_blink_set() to take
> >> > GPIO descriptor instead. In order to do this we convert the existing four
> >> > users to accept GPIO descriptor and translate it to legacy GPIO number in
> >> > the platform code. This effectively "pushes" legacy GPIO number usage from
> >> > the driver to platforms.
> >> >
> >> > Also add comment to the remaining block describing that it is legacy code
> >> > path and we are getting rid of it eventually.
> >> >
> >> > Suggested-by: Linus Walleij <linus.walleij@linaro.org>
> >> > Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
> >>
> >> Acked-by: Alexandre Courbot <acourbot@nvidia.com>
> >
> > Patch applied, thanks everyone!
> 
> "leds: leds-gpio: Add support for GPIO descriptors" broke leds-gpio on
> non-DT platforms for me:
> 
>     gpiod_direction_output: invalid GPIO
>     leds-gpio: probe of leds-gpio failed with error -22
> 
> (desc is NULL in gpiod_direction_output()).
> 
> DT shmobile reference/multi-platform are fine.
> 
> I noticed the hard way, as I wanted to add some LEDs to a new platform,
> but couldn't get it work. It turned out it also had stopped working on
> r8a7740/armadillo-legacy, so I started bisecting...

Which board file that is?

There is a bug that gpio_to_desc() returns NULL instead if ERR_PTR() in
that patch but I wonder why gpio_is_valid() and devm_gpio_request_one()
do not complain about that prior.

Re: [PATCH] leds: leds-gpio: Convert gpio_blink_set() to use GPIO descriptors

[Geert Uytterhoeven]

Hi Mika,

On Thu, Nov 6, 2014 at 11:30 AM, Mika Westerberg
<mika.westerberg@linux.intel.com> wrote:
>> "leds: leds-gpio: Add support for GPIO descriptors" broke leds-gpio on
>> non-DT platforms for me:
>>
>>     gpiod_direction_output: invalid GPIO
>>     leds-gpio: probe of leds-gpio failed with error -22
>>
>> (desc is NULL in gpiod_direction_output()).
>>
>> DT shmobile reference/multi-platform are fine.
>>
>> I noticed the hard way, as I wanted to add some LEDs to a new platform,
>> but couldn't get it work. It turned out it also had stopped working on
>> r8a7740/armadillo-legacy, so I started bisecting...
>
> Which board file that is?
>
> There is a bug that gpio_to_desc() returns NULL instead if ERR_PTR() in
> that patch but I wonder why gpio_is_valid() and devm_gpio_request_one()
> do not complain about that prior.

arch/arm/mach-shmobile/board-armadillo800eva.c

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH] leds: leds-gpio: Convert gpio_blink_set() to use GPIO descriptors

[Mika Westerberg]

On Thu, Nov 06, 2014 at 11:32:19AM +0100, Geert Uytterhoeven wrote:
> Hi Mika,
> 
> On Thu, Nov 6, 2014 at 11:30 AM, Mika Westerberg
> <mika.westerberg@linux.intel.com> wrote:
> >> "leds: leds-gpio: Add support for GPIO descriptors" broke leds-gpio on
> >> non-DT platforms for me:
> >>
> >>     gpiod_direction_output: invalid GPIO
> >>     leds-gpio: probe of leds-gpio failed with error -22
> >>
> >> (desc is NULL in gpiod_direction_output()).
> >>
> >> DT shmobile reference/multi-platform are fine.
> >>
> >> I noticed the hard way, as I wanted to add some LEDs to a new platform,
> >> but couldn't get it work. It turned out it also had stopped working on
> >> r8a7740/armadillo-legacy, so I started bisecting...
> >
> > Which board file that is?
> >
> > There is a bug that gpio_to_desc() returns NULL instead if ERR_PTR() in
> > that patch but I wonder why gpio_is_valid() and devm_gpio_request_one()
> > do not complain about that prior.
> 
> arch/arm/mach-shmobile/board-armadillo800eva.c

Thanks.

Are you able to put some printks() to the 'if (!template->gpiod)' branch
so that it prints out gpio number and what does devm_gpio_request_one()
return?

Something like:

	if (!template->gpiod) {
		...
		ret = devm_gpio_request_one(parent, template->gpio, flags,
					    template->name);
		dev_info(parent, "GPIO %u, ret: %d\n", template->gpio, ret);
		if (ret < 0)
			...

		led_dat->gpiod = gpio_to_desc(template->gpio);
		dev_info(parent, "GPIOD: %p\n", led_dat->gpiod);
	}

Re: [PATCH] leds: leds-gpio: Convert gpio_blink_set() to use GPIO descriptors

[Geert Uytterhoeven]

Hi Mika,

On Thu, Nov 6, 2014 at 11:58 AM, Mika Westerberg
<mika.westerberg@linux.intel.com> wrote:
> On Thu, Nov 06, 2014 at 11:32:19AM +0100, Geert Uytterhoeven wrote:
>> On Thu, Nov 6, 2014 at 11:30 AM, Mika Westerberg
>> <mika.westerberg@linux.intel.com> wrote:
>> >> "leds: leds-gpio: Add support for GPIO descriptors" broke leds-gpio on
>> >> non-DT platforms for me:
>> >>
>> >>     gpiod_direction_output: invalid GPIO
>> >>     leds-gpio: probe of leds-gpio failed with error -22
>> >>
>> >> (desc is NULL in gpiod_direction_output()).
>> >>
>> >> DT shmobile reference/multi-platform are fine.
>> >>
>> >> I noticed the hard way, as I wanted to add some LEDs to a new platform,
>> >> but couldn't get it work. It turned out it also had stopped working on
>> >> r8a7740/armadillo-legacy, so I started bisecting...
>> >
>> > Which board file that is?
>> >
>> > There is a bug that gpio_to_desc() returns NULL instead if ERR_PTR() in
>> > that patch but I wonder why gpio_is_valid() and devm_gpio_request_one()
>> > do not complain about that prior.
>>
>> arch/arm/mach-shmobile/board-armadillo800eva.c
>
> Are you able to put some printks() to the 'if (!template->gpiod)' branch
> so that it prints out gpio number and what does devm_gpio_request_one()
> return?
>
> Something like:
>
>         if (!template->gpiod) {
>                 ...
>                 ret = devm_gpio_request_one(parent, template->gpio, flags,
>                                             template->name);
>                 dev_info(parent, "GPIO %u, ret: %d\n", template->gpio, ret);
>                 if (ret < 0)
>                         ...
>
>                 led_dat->gpiod = gpio_to_desc(template->gpio);
>                 dev_info(parent, "GPIOD: %p\n", led_dat->gpiod);

Sure:

    leds-gpio leds-gpio: GPIO 102, ret: 0
    leds-gpio leds-gpio: GPIOD: c050e970

So led_dat is non-NULL. But it's overwritten by NULL later:

    led_dat->gpiod = template->gpiod;

Whitespace damaged fix below, to fold into the original.
If you prefer a proper separate patch, let me know.

diff --git a/drivers/leds/leds-gpio.c b/drivers/leds/leds-gpio.c
index ba4698c32bb04bde..b3c5d9d6a42bcd8b 100644
--- a/drivers/leds/leds-gpio.c
+++ b/drivers/leds/leds-gpio.c
@@ -92,7 +92,8 @@ static int create_gpio_led(const struct gpio_led *template,
 {
        int ret, state;

-       if (!template->gpiod) {
+       led_dat->gpiod = template->gpiod;
+       if (!led_dat->gpiod) {
                /*
                 * This is the legacy code path for platform code that
                 * still uses GPIO numbers. Ultimately we would like to get
@@ -122,8 +123,7 @@ static int create_gpio_led(const struct gpio_led *template,

        led_dat->cdev.name = template->name;
        led_dat->cdev.default_trigger = template->default_trigger;
-       led_dat->gpiod = template->gpiod;
-       led_dat->can_sleep = gpiod_cansleep(template->gpiod);
+       led_dat->can_sleep = gpiod_cansleep(led_dat->gpiod);
        led_dat->blinking = 0;
        if (blink_set) {
                led_dat->platform_gpio_blink_set = blink_set;

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH] leds: leds-gpio: Convert gpio_blink_set() to use GPIO descriptors

[Mika Westerberg]

On Thu, Nov 06, 2014 at 12:12:04PM +0100, Geert Uytterhoeven wrote:
> Hi Mika,
> 
> On Thu, Nov 6, 2014 at 11:58 AM, Mika Westerberg
> <mika.westerberg@linux.intel.com> wrote:
> > On Thu, Nov 06, 2014 at 11:32:19AM +0100, Geert Uytterhoeven wrote:
> >> On Thu, Nov 6, 2014 at 11:30 AM, Mika Westerberg
> >> <mika.westerberg@linux.intel.com> wrote:
> >> >> "leds: leds-gpio: Add support for GPIO descriptors" broke leds-gpio on
> >> >> non-DT platforms for me:
> >> >>
> >> >>     gpiod_direction_output: invalid GPIO
> >> >>     leds-gpio: probe of leds-gpio failed with error -22
> >> >>
> >> >> (desc is NULL in gpiod_direction_output()).
> >> >>
> >> >> DT shmobile reference/multi-platform are fine.
> >> >>
> >> >> I noticed the hard way, as I wanted to add some LEDs to a new platform,
> >> >> but couldn't get it work. It turned out it also had stopped working on
> >> >> r8a7740/armadillo-legacy, so I started bisecting...
> >> >
> >> > Which board file that is?
> >> >
> >> > There is a bug that gpio_to_desc() returns NULL instead if ERR_PTR() in
> >> > that patch but I wonder why gpio_is_valid() and devm_gpio_request_one()
> >> > do not complain about that prior.
> >>
> >> arch/arm/mach-shmobile/board-armadillo800eva.c
> >
> > Are you able to put some printks() to the 'if (!template->gpiod)' branch
> > so that it prints out gpio number and what does devm_gpio_request_one()
> > return?
> >
> > Something like:
> >
> >         if (!template->gpiod) {
> >                 ...
> >                 ret = devm_gpio_request_one(parent, template->gpio, flags,
> >                                             template->name);
> >                 dev_info(parent, "GPIO %u, ret: %d\n", template->gpio, ret);
> >                 if (ret < 0)
> >                         ...
> >
> >                 led_dat->gpiod = gpio_to_desc(template->gpio);
> >                 dev_info(parent, "GPIOD: %p\n", led_dat->gpiod);
> 
> Sure:
> 
>     leds-gpio leds-gpio: GPIO 102, ret: 0
>     leds-gpio leds-gpio: GPIOD: c050e970
> 
> So led_dat is non-NULL. But it's overwritten by NULL later:
> 
>     led_dat->gpiod = template->gpiod;

Ah, that's it. Nice catch!

> Whitespace damaged fix below, to fold into the original.
> If you prefer a proper separate patch, let me know.

It is up to Rafael. I think he wants to stabilize this branch so in that
case separate patch on top would work better.

Feel free to add,

Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>

to the patch.

> 
> diff --git a/drivers/leds/leds-gpio.c b/drivers/leds/leds-gpio.c
> index ba4698c32bb04bde..b3c5d9d6a42bcd8b 100644
> --- a/drivers/leds/leds-gpio.c
> +++ b/drivers/leds/leds-gpio.c
> @@ -92,7 +92,8 @@ static int create_gpio_led(const struct gpio_led *template,
>  {
>         int ret, state;
> 
> -       if (!template->gpiod) {
> +       led_dat->gpiod = template->gpiod;
> +       if (!led_dat->gpiod) {
>                 /*
>                  * This is the legacy code path for platform code that
>                  * still uses GPIO numbers. Ultimately we would like to get
> @@ -122,8 +123,7 @@ static int create_gpio_led(const struct gpio_led *template,
> 
>         led_dat->cdev.name = template->name;
>         led_dat->cdev.default_trigger = template->default_trigger;
> -       led_dat->gpiod = template->gpiod;
> -       led_dat->can_sleep = gpiod_cansleep(template->gpiod);
> +       led_dat->can_sleep = gpiod_cansleep(led_dat->gpiod);
>         led_dat->blinking = 0;
>         if (blink_set) {
>                 led_dat->platform_gpio_blink_set = blink_set;
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
> 
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds

Re: [PATCH] leds: leds-gpio: Convert gpio_blink_set() to use GPIO descriptors

[Geert Uytterhoeven]

On Thu, Nov 6, 2014 at 12:22 PM, Mika Westerberg
<mika.westerberg@linux.intel.com> wrote:
> On Thu, Nov 06, 2014 at 12:12:04PM +0100, Geert Uytterhoeven wrote:
>> On Thu, Nov 6, 2014 at 11:58 AM, Mika Westerberg
>> <mika.westerberg@linux.intel.com> wrote:
>> > On Thu, Nov 06, 2014 at 11:32:19AM +0100, Geert Uytterhoeven wrote:
>> >> On Thu, Nov 6, 2014 at 11:30 AM, Mika Westerberg
>> >> <mika.westerberg@linux.intel.com> wrote:
>> >> >> "leds: leds-gpio: Add support for GPIO descriptors" broke leds-gpio on
>> >> >> non-DT platforms for me:
>> >> >>
>> >> >>     gpiod_direction_output: invalid GPIO
>> >> >>     leds-gpio: probe of leds-gpio failed with error -22
>> >> >>
>> >> >> (desc is NULL in gpiod_direction_output()).

>>     leds-gpio leds-gpio: GPIO 102, ret: 0
>>     leds-gpio leds-gpio: GPIOD: c050e970
>>
>> So led_dat is non-NULL. But it's overwritten by NULL later:
>>
>>     led_dat->gpiod = template->gpiod;
>
> Ah, that's it. Nice catch!
>
>> Whitespace damaged fix below, to fold into the original.
>> If you prefer a proper separate patch, let me know.
>
> It is up to Rafael. I think he wants to stabilize this branch so in that
> case separate patch on top would work better.

I've already sent a patch, as I have to publish the fix in my git tree anyway.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds