[PATCH v3] gpiolib: acpi: Program debounce when finding GPIO

[PATCH v3] gpiolib: acpi: Program debounce when finding GPIO

[Mario Limonciello (AMD)]

When soc-button-array looks up the GPIO to use it calls acpi_find_gpio()
which will parse _CRS.

acpi_find_gpio.cold (drivers/gpio/gpiolib-acpi-core.c:953)
gpiod_find_and_request (drivers/gpio/gpiolib.c:4598 drivers/gpio/gpiolib.c:4625)
gpiod_get_index (drivers/gpio/gpiolib.c:4877)

The GPIO is setup basically, but the debounce information is discarded.
The platform will assert what debounce should be in _CRS, so program it
at the time it's available.

As this is considered non fatal if it fails, introduce a helper for
programming debounce and show a warning when failing to program.

Reviewed-by: Hans de Goede <hansg@kernel.org>
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
---
v3:
 * squash patches 1 and 2 together and drop the extra fatal hunk
v2:
 * https://lore.kernel.org/all/20250625181342.3175969-1-superm1@kernel.org/
---
 drivers/gpio/gpiolib-acpi-core.c | 25 ++++++++++++++-----------
 1 file changed, 14 insertions(+), 11 deletions(-)

diff --git a/drivers/gpio/gpiolib-acpi-core.c b/drivers/gpio/gpiolib-acpi-core.c
index 12b24a717e43..15222bfc25bb 100644
--- a/drivers/gpio/gpiolib-acpi-core.c
+++ b/drivers/gpio/gpiolib-acpi-core.c
@@ -291,6 +291,17 @@ acpi_gpio_to_gpiod_flags(const struct acpi_resource_gpio *agpio, int polarity)
 	return GPIOD_ASIS;
 }
 
+static void acpi_set_debounce_timeout(struct gpio_desc *desc, unsigned int timeout)
+{
+	int ret;
+
+	/* ACPI uses hundredths of milliseconds units */
+	ret = gpio_set_debounce_timeout(desc, timeout * 10);
+	if (ret)
+		dev_warn(&desc->gdev->dev,
+			 "Failed to set debounce-timeout: %d\n", ret);
+}
+
 static struct gpio_desc *acpi_request_own_gpiod(struct gpio_chip *chip,
 						struct acpi_resource_gpio *agpio,
 						unsigned int index,
@@ -300,18 +311,12 @@ static struct gpio_desc *acpi_request_own_gpiod(struct gpio_chip *chip,
 	enum gpiod_flags flags = acpi_gpio_to_gpiod_flags(agpio, polarity);
 	unsigned int pin = agpio->pin_table[index];
 	struct gpio_desc *desc;
-	int ret;
 
 	desc = gpiochip_request_own_desc(chip, pin, label, polarity, flags);
 	if (IS_ERR(desc))
 		return desc;
 
-	/* ACPI uses hundredths of milliseconds units */
-	ret = gpio_set_debounce_timeout(desc, agpio->debounce_timeout * 10);
-	if (ret)
-		dev_warn(chip->parent,
-			 "Failed to set debounce-timeout for pin 0x%04X, err %d\n",
-			 pin, ret);
+	acpi_set_debounce_timeout(desc, agpio->debounce_timeout);
 
 	return desc;
 }
@@ -957,6 +962,7 @@ struct gpio_desc *acpi_find_gpio(struct fwnode_handle *fwnode,
 
 	acpi_gpio_update_gpiod_flags(dflags, &info);
 	acpi_gpio_update_gpiod_lookup_flags(lookupflags, &info);
+	acpi_set_debounce_timeout(desc, info.debounce);
 	return desc;
 }
 
@@ -1025,10 +1031,7 @@ int acpi_dev_gpio_irq_wake_get_by(struct acpi_device *adev, const char *con_id,
 			if (ret < 0)
 				return ret;
 
-			/* ACPI uses hundredths of milliseconds units */
-			ret = gpio_set_debounce_timeout(desc, info.debounce * 10);
-			if (ret)
-				return ret;
+			acpi_set_debounce_timeout(desc, info.debounce);
 
 			irq_flags = acpi_dev_get_irq_type(info.triggering,
 							  info.polarity);

base-commit: 8f5ae30d69d7543eee0d70083daf4de8fe15d585
-- 
2.50.1

Re: [PATCH v3] gpiolib: acpi: Program debounce when finding GPIO

[Andy Shevchenko]

On Mon, Aug 11, 2025 at 11:43:56AM -0500, Mario Limonciello (AMD) wrote:
> When soc-button-array looks up the GPIO to use it calls acpi_find_gpio()
> which will parse _CRS.
> 
> acpi_find_gpio.cold (drivers/gpio/gpiolib-acpi-core.c:953)
> gpiod_find_and_request (drivers/gpio/gpiolib.c:4598 drivers/gpio/gpiolib.c:4625)
> gpiod_get_index (drivers/gpio/gpiolib.c:4877)
> 
> The GPIO is setup basically, but the debounce information is discarded.
> The platform will assert what debounce should be in _CRS, so program it
> at the time it's available.
> 
> As this is considered non fatal if it fails, introduce a helper for
> programming debounce and show a warning when failing to program.

I think I already commented on this previously. Let me do that below anyway.

...


> +static void acpi_set_debounce_timeout(struct gpio_desc *desc, unsigned int timeout)
> +{
> +	int ret;
> +
> +	/* ACPI uses hundredths of milliseconds units */
> +	ret = gpio_set_debounce_timeout(desc, timeout * 10);
> +	if (ret)
> +		dev_warn(&desc->gdev->dev,
> +			 "Failed to set debounce-timeout: %d\n", ret);
> +}

I would make it still return the code to the caller. See below why.


...

> -			/* ACPI uses hundredths of milliseconds units */
> -			ret = gpio_set_debounce_timeout(desc, info.debounce * 10);
> -			if (ret)
> -				return ret;
> +			acpi_set_debounce_timeout(desc, info.debounce);

The commit message fails to explain why we do relax the condition here. This is
about GpioInt() resource and so far I haven't heard about misused debounce
values there. If we drop the fatality, it has to be a separate patch explaining
why. But only if we have practical use cases. AS long as there no failed
platforms, I can't agree on this piece of change.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v3] gpiolib: acpi: Program debounce when finding GPIO

[Mario Limonciello]

On 8/11/25 3:42 PM, Andy Shevchenko wrote:
> On Mon, Aug 11, 2025 at 11:43:56AM -0500, Mario Limonciello (AMD) wrote:
>> When soc-button-array looks up the GPIO to use it calls acpi_find_gpio()
>> which will parse _CRS.
>>
>> acpi_find_gpio.cold (drivers/gpio/gpiolib-acpi-core.c:953)
>> gpiod_find_and_request (drivers/gpio/gpiolib.c:4598 drivers/gpio/gpiolib.c:4625)
>> gpiod_get_index (drivers/gpio/gpiolib.c:4877)
>>
>> The GPIO is setup basically, but the debounce information is discarded.
>> The platform will assert what debounce should be in _CRS, so program it
>> at the time it's available.
>>
>> As this is considered non fatal if it fails, introduce a helper for
>> programming debounce and show a warning when failing to program.
> 
> I think I already commented on this previously. Let me do that below anyway.
> 
> ...
> 
> 
>> +static void acpi_set_debounce_timeout(struct gpio_desc *desc, unsigned int timeout)
>> +{
>> +	int ret;
>> +
>> +	/* ACPI uses hundredths of milliseconds units */
>> +	ret = gpio_set_debounce_timeout(desc, timeout * 10);
>> +	if (ret)
>> +		dev_warn(&desc->gdev->dev,
>> +			 "Failed to set debounce-timeout: %d\n", ret);
>> +}
> 
> I would make it still return the code to the caller. See below why.
> 
> 
> ...
> 
>> -			/* ACPI uses hundredths of milliseconds units */
>> -			ret = gpio_set_debounce_timeout(desc, info.debounce * 10);
>> -			if (ret)
>> -				return ret;
>> +			acpi_set_debounce_timeout(desc, info.debounce);
> 
> The commit message fails to explain why we do relax the condition here. This is
> about GpioInt() resource and so far I haven't heard about misused debounce
> values there. If we drop the fatality, it has to be a separate patch explaining
> why. But only if we have practical use cases. AS long as there no failed
> platforms, I can't agree on this piece of change.
> 

Thanks for the feedback.  I thought that I got your feedback the first 
time when I originally squashed, but I must have failed.

In that case I don't think we really need a helper at all.  I'll change 
it for v4 to just add the extra call without the use of a helper.