Skip to contentSkip to conversation

gpiolib: acpi: Program debounce when finding GPIO

PATCH Superseded Mario Limonciello (AMD) 2 participants 3 messages Aug 11, 2025

Subsystems (4)

v3 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 gpiolib: acpi: Program debounce when finding GPIO

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.

---
v3:
* squash patches 1 and 2 together and drop the extra fatal hunk
v2:
* https://lore.kernel.org/all/[email protected]/

---
 drivers/gpio/gpiolib-acpi-core.c | 25 ++++++++++++++-----------
 1 file changed, 14 insertions(+), 11 deletions(-)
drivers/gpio/gpiolib-acpi-core.c +14 -11
index 12b24a717e43..15222bfc25bb 100644
@@ -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

Signature · 1 line
2.50.1
maintainer
#1 lore raw
Quoted 7 lines · 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)

Quoted 6 lines

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.

...

drivers/gpio/gpiolib-acpi-core.c acpi_gpio_to_gpiod_flags(const struct acpi_resource_gpio *agpio, int polarity) · line 303

Quoted 10 lines
drivers/gpio/gpiolib-acpi-core.c +10 -0
+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.

...

drivers/gpio/gpiolib-acpi-core.c int acpi_dev_gpio_irq_wake_get_by(struct acpi_device *adev, const char *con_id, · line 1034

Quoted 5 lines
drivers/gpio/gpiolib-acpi-core.c +1 -4
-			/* 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.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#2 lore raw
Quoted 36 lines · Andy Shevchenko
Quoted 7 lines · 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)

Quoted 6 lines

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.

...

drivers/gpio/gpiolib-acpi-core.c acpi_gpio_to_gpiod_flags(const struct acpi_resource_gpio *agpio, int polarity) · line 303

Quoted 10 lines
drivers/gpio/gpiolib-acpi-core.c +10 -0
+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.

...

Quoted 12 lines

drivers/gpio/gpiolib-acpi-core.c int acpi_dev_gpio_irq_wake_get_by(struct acpi_device *adev, const char *con_id, · line 1034

Quoted 5 lines
drivers/gpio/gpiolib-acpi-core.c +1 -4
-			/* 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.