[PATCH] pwm-backlight: Pinctrl-fy

[PATCH] pwm-backlight: Pinctrl-fy

[Pantelis Antoniou]

Enable pinctrl for pwm-backlight.

Signed-off-by: Pantelis Antoniou <panto@antoniou-consulting.com>
---
 drivers/video/backlight/pwm_bl.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/video/backlight/pwm_bl.c b/drivers/video/backlight/pwm_bl.c
index 0c91023..f3b6194 100644
--- a/drivers/video/backlight/pwm_bl.c
+++ b/drivers/video/backlight/pwm_bl.c
@@ -20,6 +20,8 @@
 #include <linux/pwm.h>
 #include <linux/pwm_backlight.h>
 #include <linux/slab.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/err.h>
 
 struct pwm_bl_data {
 	struct pwm_device	*pwm;
@@ -180,9 +182,14 @@ static int pwm_backlight_probe(struct platform_device *pdev)
 	struct backlight_properties props;
 	struct backlight_device *bl;
 	struct pwm_bl_data *pb;
+	struct pinctrl *pinctrl;
 	unsigned int max;
 	int ret;
 
+	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
+	if (IS_ERR(pinctrl))
+		dev_warn(&pdev->dev, "unable to select pin group\n");
+
 	if (!data) {
 		ret = pwm_backlight_parse_dt(&pdev->dev, &defdata);
 		if (ret < 0) {
-- 
1.7.12

Re: [PATCH] pwm-backlight: Pinctrl-fy

[Thierry Reding]

[-- Attachment #1: Type: text/plain, Size: 387 bytes --]

On Wed, Oct 31, 2012 at 05:57:27PM +0200, Pantelis Antoniou wrote:
[...]
> diff --git a/drivers/video/backlight/pwm_bl.c b/drivers/video/backlight/pwm_bl.c
[...]
> @@ -20,6 +20,8 @@
>  #include <linux/pwm.h>
>  #include <linux/pwm_backlight.h>
>  #include <linux/slab.h>
> +#include <linux/pinctrl/consumer.h>
> +#include <linux/err.h>

linux/err.h is already included earlier.

Thierry

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH] pwm-backlight: Pinctrl-fy

[Thierry Reding]

[-- Attachment #1: Type: text/plain, Size: 1923 bytes --]

On Wed, Oct 31, 2012 at 05:57:27PM +0200, Pantelis Antoniou wrote:
> Enable pinctrl for pwm-backlight.
> 
> Signed-off-by: Pantelis Antoniou <panto@antoniou-consulting.com>
> ---
>  drivers/video/backlight/pwm_bl.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/drivers/video/backlight/pwm_bl.c b/drivers/video/backlight/pwm_bl.c
> index 0c91023..f3b6194 100644
> --- a/drivers/video/backlight/pwm_bl.c
> +++ b/drivers/video/backlight/pwm_bl.c
> @@ -20,6 +20,8 @@
>  #include <linux/pwm.h>
>  #include <linux/pwm_backlight.h>
>  #include <linux/slab.h>
> +#include <linux/pinctrl/consumer.h>
> +#include <linux/err.h>
>  
>  struct pwm_bl_data {
>  	struct pwm_device	*pwm;
> @@ -180,9 +182,14 @@ static int pwm_backlight_probe(struct platform_device *pdev)
>  	struct backlight_properties props;
>  	struct backlight_device *bl;
>  	struct pwm_bl_data *pb;
> +	struct pinctrl *pinctrl;
>  	unsigned int max;
>  	int ret;
>  
> +	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
> +	if (IS_ERR(pinctrl))
> +		dev_warn(&pdev->dev, "unable to select pin group\n");
> +

I just saw this done in a similar way in some other driver and then
remembered your patch. When I reviewed this I wasn't sure if a warning
was good enough in this case. I've checked the kernel tree and it seems
like a majority handled this as failure instead of a warning. I took a
look at the pinctrl core and it seems like indeed if neither pinctrl is
enabled or if there isn't actually a pinmux configuration for a device,
then devm_pinctrl_get_select_default() will actually not return an
error, so in those cases where an error is returned it should actually
be handled as a fatal error.

I'm Cc'ing Linus Walleij, maybe he knows what to do.

Thierry

>  	if (!data) {
>  		ret = pwm_backlight_parse_dt(&pdev->dev, &defdata);
>  		if (ret < 0) {
> -- 
> 1.7.12
> 
> 
> 

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH] pwm-backlight: Pinctrl-fy

[Linus Walleij]

On Fri, Nov 9, 2012 at 9:06 AM, Thierry Reding
<thierry.reding@avionic-design.de> wrote:
> On Wed, Oct 31, 2012 at 05:57:27PM +0200, Pantelis Antoniou wrote:

>> +     pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
>> +     if (IS_ERR(pinctrl))
>> +             dev_warn(&pdev->dev, "unable to select pin group\n");
>> +
>
> I just saw this done in a similar way in some other driver and then
> remembered your patch. When I reviewed this I wasn't sure if a warning
> was good enough in this case. I've checked the kernel tree and it seems
> like a majority handled this as failure instead of a warning. I took a
> look at the pinctrl core and it seems like indeed if neither pinctrl is
> enabled or if there isn't actually a pinmux configuration for a device,
> then devm_pinctrl_get_select_default() will actually not return an
> error, so in those cases where an error is returned it should actually
> be handled as a fatal error.

So it depends.

One good reason to just error out and return the error code is
if the error returned is -EPROBE_DEFER, right?

Because then you're pinctrl driver is not up yet, and you
need to bail out and be revisited later.

On a related key we have this debate going on with some
subsystem maintainers as to whether we should try do
centralize boilerplate like this, the lates suggestion is:
http://marc.info/?l=linux-kernel&m=135263661110528&w=2

The fun never ends...

Yours,
Linus Walleij

Re: [PATCH] pwm-backlight: Pinctrl-fy

[Pantelis Antoniou]

Hi Linus,

On Nov 11, 2012, at 7:26 PM, Linus Walleij wrote:

> On Fri, Nov 9, 2012 at 9:06 AM, Thierry Reding
> <thierry.reding@avionic-design.de> wrote:
>> On Wed, Oct 31, 2012 at 05:57:27PM +0200, Pantelis Antoniou wrote:
> 
>>> +     pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
>>> +     if (IS_ERR(pinctrl))
>>> +             dev_warn(&pdev->dev, "unable to select pin group\n");
>>> +
>> 
>> I just saw this done in a similar way in some other driver and then
>> remembered your patch. When I reviewed this I wasn't sure if a warning
>> was good enough in this case. I've checked the kernel tree and it seems
>> like a majority handled this as failure instead of a warning. I took a
>> look at the pinctrl core and it seems like indeed if neither pinctrl is
>> enabled or if there isn't actually a pinmux configuration for a device,
>> then devm_pinctrl_get_select_default() will actually not return an
>> error, so in those cases where an error is returned it should actually
>> be handled as a fatal error.
> 
> So it depends.
> 
> One good reason to just error out and return the error code is
> if the error returned is -EPROBE_DEFER, right?
> 
> Because then you're pinctrl driver is not up yet, and you
> need to bail out and be revisited later.
> 
> On a related key we have this debate going on with some
> subsystem maintainers as to whether we should try do
> centralize boilerplate like this, the lates suggestion is:
> http://marc.info/?l=linux-kernel&m=135263661110528&w=2
> 

Interesting... This is certainly much nicer than peppering all the
devices with pinctrl calls.

> The fun never ends...
> 

Indeed.

> Yours,
> Linus Walleij

Regards

-- Pantelis