[PATCH V2 0/2] ARM: dts: bcm2711-rpi-400: Fix delete-node of led_act

[PATCH V2 0/2] ARM: dts: bcm2711-rpi-400: Fix delete-node of led_act

[Stefan Wahren]

This series fixes the probing of leds-gpio on the Raspberry Pi 400.
Also try to improve the error logging of leds-gpio.

Changes in V2:
- replace dev_err with dev_err_probe in patch 1 to handle EPROBE_DEFER
  properly

Stefan Wahren (2):
  leds: gpio: Add kernel log if devm_fwnode_gpiod_get fails
  ARM: dts: bcm2711-rpi-400: Fix delete-node of led_act

 arch/arm/boot/dts/broadcom/bcm2711-rpi-400.dts | 4 +---
 drivers/leds/leds-gpio.c                       | 2 ++
 2 files changed, 3 insertions(+), 3 deletions(-)

--
2.34.1

[PATCH V2 1/2] leds: gpio: Add kernel log if devm_fwnode_gpiod_get fails

[Stefan Wahren]

In case leds-gpio fails to get the GPIO from the DT (e.g. the GPIO is
already requested) the driver doesn't provide any helpful error log:

    leds-gpio: probe of leds failed with error -16

So add a new error log in case devm_fwnode_gpiod_get() fails.

Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
---
 drivers/leds/leds-gpio.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/leds/leds-gpio.c b/drivers/leds/leds-gpio.c
index 710c319ad312..0159cedffa9e 100644
--- a/drivers/leds/leds-gpio.c
+++ b/drivers/leds/leds-gpio.c
@@ -172,6 +172,8 @@ static struct gpio_leds_priv *gpio_leds_create(struct device *dev)
 		led.gpiod = devm_fwnode_gpiod_get(dev, child, NULL, GPIOD_ASIS,
 						  NULL);
 		if (IS_ERR(led.gpiod)) {
+			dev_err_probe(dev, PTR_ERR(led.gpiod), "Failed to get gpio '%pfw'\n",
+				      child);
 			fwnode_handle_put(child);
 			return ERR_CAST(led.gpiod);
 		}
--
2.34.1

Re: [PATCH V2 1/2] leds: gpio: Add kernel log if devm_fwnode_gpiod_get fails

[Andy Shevchenko]

On Sat, Nov 18, 2023 at 01:42:51PM +0100, Stefan Wahren wrote:
> In case leds-gpio fails to get the GPIO from the DT (e.g. the GPIO is
> already requested) the driver doesn't provide any helpful error log:
> 
>     leds-gpio: probe of leds failed with error -16
> 
> So add a new error log in case devm_fwnode_gpiod_get() fails.

...

>  		led.gpiod = devm_fwnode_gpiod_get(dev, child, NULL, GPIOD_ASIS,
>  						  NULL);
>  		if (IS_ERR(led.gpiod)) {
> +			dev_err_probe(dev, PTR_ERR(led.gpiod), "Failed to get gpio '%pfw'\n",
> +				      child);
>  			fwnode_handle_put(child);
>  			return ERR_CAST(led.gpiod);
>  		}

Thinking more about it. GPIO library already issues bunch of messages.

	"using DT ... for ... GPIO lookup"
	"using lookup tables for GPIO lookup"
	"No GPIO consumer ... found"

Isn't it enough?

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH V2 1/2] leds: gpio: Add kernel log if devm_fwnode_gpiod_get fails

[Stefan Wahren]

Hi Andy,

[add Linus and Bartosz]

Am 20.11.23 um 12:47 schrieb Andy Shevchenko:
> On Sat, Nov 18, 2023 at 01:42:51PM +0100, Stefan Wahren wrote:
>> In case leds-gpio fails to get the GPIO from the DT (e.g. the GPIO is
>> already requested) the driver doesn't provide any helpful error log:
>>
>>      leds-gpio: probe of leds failed with error -16
>>
>> So add a new error log in case devm_fwnode_gpiod_get() fails.
> ...
>
>>   		led.gpiod = devm_fwnode_gpiod_get(dev, child, NULL, GPIOD_ASIS,
>>   						  NULL);
>>   		if (IS_ERR(led.gpiod)) {
>> +			dev_err_probe(dev, PTR_ERR(led.gpiod), "Failed to get gpio '%pfw'\n",
>> +				      child);
>>   			fwnode_handle_put(child);
>>   			return ERR_CAST(led.gpiod);
>>   		}
> Thinking more about it. GPIO library already issues bunch of messages.
>
> 	"using DT ... for ... GPIO lookup"
> 	"using lookup tables for GPIO lookup"
> 	"No GPIO consumer ... found"
are these errors or debug messages?

I cannot remember that i saw any of them on info level in my case of an
already allocated pin (see patch 2).

I'm open to place the log within gpiolib, if this a better place.

Best regards
>
> Isn't it enough?
>

Re: [PATCH V2 1/2] leds: gpio: Add kernel log if devm_fwnode_gpiod_get fails

[Andy Shevchenko]

On Mon, Nov 20, 2023 at 01:02:55PM +0100, Stefan Wahren wrote:
> Am 20.11.23 um 12:47 schrieb Andy Shevchenko:
> > On Sat, Nov 18, 2023 at 01:42:51PM +0100, Stefan Wahren wrote:
> > > In case leds-gpio fails to get the GPIO from the DT (e.g. the GPIO is
> > > already requested) the driver doesn't provide any helpful error log:
> > > 
> > >      leds-gpio: probe of leds failed with error -16
> > > 
> > > So add a new error log in case devm_fwnode_gpiod_get() fails.

...

> > >   		led.gpiod = devm_fwnode_gpiod_get(dev, child, NULL, GPIOD_ASIS,
> > >   						  NULL);
> > >   		if (IS_ERR(led.gpiod)) {
> > > +			dev_err_probe(dev, PTR_ERR(led.gpiod), "Failed to get gpio '%pfw'\n",
> > > +				      child);
> > >   			fwnode_handle_put(child);
> > >   			return ERR_CAST(led.gpiod);
> > >   		}
> > Thinking more about it. GPIO library already issues bunch of messages.
> > 
> > 	"using DT ... for ... GPIO lookup"
> > 	"using lookup tables for GPIO lookup"
> > 	"No GPIO consumer ... found"
> are these errors or debug messages?

Indeed they are on debug level.

> I cannot remember that i saw any of them on info level in my case of an
> already allocated pin (see patch 2).
> 
> I'm open to place the log within gpiolib, if this a better place.

I'm not sure, let's hear GPIO maintainers for that.

> > Isn't it enough?

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH V2 1/2] leds: gpio: Add kernel log if devm_fwnode_gpiod_get fails

[Bartosz Golaszewski]

On Mon, Nov 20, 2023 at 1:36 PM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
>
> On Mon, Nov 20, 2023 at 01:02:55PM +0100, Stefan Wahren wrote:
> > Am 20.11.23 um 12:47 schrieb Andy Shevchenko:
> > > On Sat, Nov 18, 2023 at 01:42:51PM +0100, Stefan Wahren wrote:
> > > > In case leds-gpio fails to get the GPIO from the DT (e.g. the GPIO is
> > > > already requested) the driver doesn't provide any helpful error log:
> > > >
> > > >      leds-gpio: probe of leds failed with error -16
> > > >
> > > > So add a new error log in case devm_fwnode_gpiod_get() fails.
>
> ...
>
> > > >                   led.gpiod = devm_fwnode_gpiod_get(dev, child, NULL, GPIOD_ASIS,
> > > >                                                     NULL);
> > > >                   if (IS_ERR(led.gpiod)) {
> > > > +                 dev_err_probe(dev, PTR_ERR(led.gpiod), "Failed to get gpio '%pfw'\n",
> > > > +                               child);
> > > >                           fwnode_handle_put(child);
> > > >                           return ERR_CAST(led.gpiod);
> > > >                   }
> > > Thinking more about it. GPIO library already issues bunch of messages.
> > >
> > >     "using DT ... for ... GPIO lookup"
> > >     "using lookup tables for GPIO lookup"
> > >     "No GPIO consumer ... found"
> > are these errors or debug messages?
>
> Indeed they are on debug level.
>
> > I cannot remember that i saw any of them on info level in my case of an
> > already allocated pin (see patch 2).
> >
> > I'm open to place the log within gpiolib, if this a better place.
>
> I'm not sure, let's hear GPIO maintainers for that.
>

Hard to tell which method is preferred among all the subsystems.
Personally I'm more inclined towards letting drivers decide whether to
emit an error message and only emit our own when an error cannot be
propagated down the stack.

Linus: Any thoughts?

Bart

> > > Isn't it enough?
>
> --
> With Best Regards,
> Andy Shevchenko
>
>

Re: [PATCH V2 1/2] leds: gpio: Add kernel log if devm_fwnode_gpiod_get fails

[Stefan Wahren]

Am 22.11.23 um 11:52 schrieb Bartosz Golaszewski:
> On Mon, Nov 20, 2023 at 1:36 PM Andy Shevchenko
> <andriy.shevchenko@linux.intel.com> wrote:
>> On Mon, Nov 20, 2023 at 01:02:55PM +0100, Stefan Wahren wrote:
>>> Am 20.11.23 um 12:47 schrieb Andy Shevchenko:
>>>> On Sat, Nov 18, 2023 at 01:42:51PM +0100, Stefan Wahren wrote:
>>>>> In case leds-gpio fails to get the GPIO from the DT (e.g. the GPIO is
>>>>> already requested) the driver doesn't provide any helpful error log:
>>>>>
>>>>>       leds-gpio: probe of leds failed with error -16
>>>>>
>>>>> So add a new error log in case devm_fwnode_gpiod_get() fails.
>> ...
>>
>>>>>                    led.gpiod = devm_fwnode_gpiod_get(dev, child, NULL, GPIOD_ASIS,
>>>>>                                                      NULL);
>>>>>                    if (IS_ERR(led.gpiod)) {
>>>>> +                 dev_err_probe(dev, PTR_ERR(led.gpiod), "Failed to get gpio '%pfw'\n",
>>>>> +                               child);
>>>>>                            fwnode_handle_put(child);
>>>>>                            return ERR_CAST(led.gpiod);
>>>>>                    }
>>>> Thinking more about it. GPIO library already issues bunch of messages.
>>>>
>>>>      "using DT ... for ... GPIO lookup"
>>>>      "using lookup tables for GPIO lookup"
>>>>      "No GPIO consumer ... found"
>>> are these errors or debug messages?
>> Indeed they are on debug level.
>>
>>> I cannot remember that i saw any of them on info level in my case of an
>>> already allocated pin (see patch 2).
>>>
>>> I'm open to place the log within gpiolib, if this a better place.
>> I'm not sure, let's hear GPIO maintainers for that.
>>
> Hard to tell which method is preferred among all the subsystems.
> Personally I'm more inclined towards letting drivers decide whether to
> emit an error message and only emit our own when an error cannot be
> propagated down the stack.
>
> Linus: Any thoughts?
gentle ping ...
>
> Bart
>
>>>> Isn't it enough?
>> --
>> With Best Regards,
>> Andy Shevchenko
>>
>>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH V2 1/2] leds: gpio: Add kernel log if devm_fwnode_gpiod_get fails

[Linus Walleij]

On Wed, Nov 22, 2023 at 11:53 AM Bartosz Golaszewski <brgl@bgdev.pl> wrote:

> > > I cannot remember that i saw any of them on info level in my case of an
> > > already allocated pin (see patch 2).
> > >
> > > I'm open to place the log within gpiolib, if this a better place.
> >
> > I'm not sure, let's hear GPIO maintainers for that.
>
> Hard to tell which method is preferred among all the subsystems.
> Personally I'm more inclined towards letting drivers decide whether to
> emit an error message and only emit our own when an error cannot be
> propagated down the stack.
>
> Linus: Any thoughts?

I never managed to get it right so I can't give any good advice.

Usually I tend to think better one more error message than one too little.

Then again I'm a dmesg maximalist who just want it to scroll on forever
also with positive messages...

Yours,
Linus Walleij

Re: [PATCH V2 1/2] leds: gpio: Add kernel log if devm_fwnode_gpiod_get fails

[Stefan Wahren]

Am 29.11.23 um 15:03 schrieb Linus Walleij:
> On Wed, Nov 22, 2023 at 11:53 AM Bartosz Golaszewski <brgl@bgdev.pl> wrote:
>
>>>> I cannot remember that i saw any of them on info level in my case of an
>>>> already allocated pin (see patch 2).
>>>>
>>>> I'm open to place the log within gpiolib, if this a better place.
>>> I'm not sure, let's hear GPIO maintainers for that.
>> Hard to tell which method is preferred among all the subsystems.
>> Personally I'm more inclined towards letting drivers decide whether to
>> emit an error message and only emit our own when an error cannot be
>> propagated down the stack.
>>
>> Linus: Any thoughts?
> I never managed to get it right so I can't give any good advice.
>
> Usually I tend to think better one more error message than one too little.
>
> Then again I'm a dmesg maximalist who just want it to scroll on forever
> also with positive messages...
Okay, based on the feedback this sounds like nobody is against this patch?
>
> Yours,
> Linus Walleij
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH V2 1/2] leds: gpio: Add kernel log if devm_fwnode_gpiod_get fails

[Lee Jones]

On Sat, 18 Nov 2023, Stefan Wahren wrote:

> In case leds-gpio fails to get the GPIO from the DT (e.g. the GPIO is
> already requested) the driver doesn't provide any helpful error log:
> 
>     leds-gpio: probe of leds failed with error -16
> 
> So add a new error log in case devm_fwnode_gpiod_get() fails.
> 
> Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
> ---
>  drivers/leds/leds-gpio.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/leds/leds-gpio.c b/drivers/leds/leds-gpio.c
> index 710c319ad312..0159cedffa9e 100644
> --- a/drivers/leds/leds-gpio.c
> +++ b/drivers/leds/leds-gpio.c
> @@ -172,6 +172,8 @@ static struct gpio_leds_priv *gpio_leds_create(struct device *dev)
>  		led.gpiod = devm_fwnode_gpiod_get(dev, child, NULL, GPIOD_ASIS,
>  						  NULL);
>  		if (IS_ERR(led.gpiod)) {
> +			dev_err_probe(dev, PTR_ERR(led.gpiod), "Failed to get gpio '%pfw'\n",

Nit: GPIO

> +				      child);
>  			fwnode_handle_put(child);
>  			return ERR_CAST(led.gpiod);
>  		}
> --
> 2.34.1
> 

-- 
Lee Jones [李琼斯]

[PATCH V2 2/2] ARM: dts: bcm2711-rpi-400: Fix delete-node of led_act

[Stefan Wahren]

The LED ACT which is included from bcm2711-rpi-4-b doesn't exists
on the Raspberry Pi 400. So the bcm2711-rpi-400.dts tries to
use the delete-node directive in order to remove the complete
node. Unfortunately the usage get broken in commit 1156e3a78bcc
("ARM: dts: bcm283x: Move ACT LED into separate dtsi")
and now ACT and PWR LED using the same GPIO and this prevent
probing of led-gpios on Raspberry Pi 400:

    leds-gpio: probe of leds failed with error -16

So fix the delete-node directive.

Fixes: 1156e3a78bcc ("ARM: dts: bcm283x: Move ACT LED into separate dtsi")
Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
---
 arch/arm/boot/dts/broadcom/bcm2711-rpi-400.dts | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/arch/arm/boot/dts/broadcom/bcm2711-rpi-400.dts b/arch/arm/boot/dts/broadcom/bcm2711-rpi-400.dts
index 1ab8184302db..5a2869a18bd5 100644
--- a/arch/arm/boot/dts/broadcom/bcm2711-rpi-400.dts
+++ b/arch/arm/boot/dts/broadcom/bcm2711-rpi-400.dts
@@ -36,9 +36,7 @@ &led_pwr {
 	gpios = <&gpio 42 GPIO_ACTIVE_HIGH>;
 };

-&leds {
-	/delete-node/ led_act;
-};
+/delete-node/ &led_act;

 &pm {
 	/delete-property/ system-power-controller;
--
2.34.1

Re: [PATCH V2 2/2] ARM: dts: bcm2711-rpi-400: Fix delete-node of led_act

[Florian Fainelli]

From: Florian Fainelli <f.fainelli@gmail.com>

On Sat, 18 Nov 2023 13:42:52 +0100, Stefan Wahren <wahrenst@gmx.net> wrote:
> The LED ACT which is included from bcm2711-rpi-4-b doesn't exists
> on the Raspberry Pi 400. So the bcm2711-rpi-400.dts tries to
> use the delete-node directive in order to remove the complete
> node. Unfortunately the usage get broken in commit 1156e3a78bcc
> ("ARM: dts: bcm283x: Move ACT LED into separate dtsi")
> and now ACT and PWR LED using the same GPIO and this prevent
> probing of led-gpios on Raspberry Pi 400:
> 
>     leds-gpio: probe of leds failed with error -16
> 
> So fix the delete-node directive.
> 
> Fixes: 1156e3a78bcc ("ARM: dts: bcm283x: Move ACT LED into separate dtsi")
> Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
> ---

Applied to https://github.com/Broadcom/stblinux/commits/devicetree/fixes, thanks!
--
Florian