Skip to contentSkip to conversation

gpio: pca953x: Better quirk for Intel Galileo Gen 2

PATCH In review Andy Shevchenko 4 participants 12 messages Feb 25, 2021 – Mar 2, 2021

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover gpio: pca953x: Better quirk for Intel Galileo Gen 2

It appears that after commit to the gpio-dwapb driver the quirk for Intel Galileo Gen 2 stopped working. I tried a few approaches but none of them worked besides the fact that they copy like a ~10% of the gpiolib-acpi functionality.

Instead I decided to come back to a mix of v1 [1] of the initial series (the v3 has been applied) and Mika's suggestion about using gpio_to_desc() API.

Unfortunately it requires to have a quirk in the core, but in comparison to v1 it much less intrusive and doesn't use any predefined numbers.

I would like to create an immutable branch and send TWIMC (GPIO and ACPI subsystems I guess). This is material for v5.12-rcX.

[1]: https://lore.kernel.org/linux-gpio/[email protected]/

Andy Shevchenko (3):
gpiolib: acpi: Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk
gpiolib: acpi: Allow to find GpioInt() resource by name and index
gpio: pca953x: Set IRQ type when handle Intel Galileo Gen 2

 drivers/gpio/gpio-pca953x.c   | 78 +++++++++++------------------------
 drivers/gpio/gpiolib-acpi.c   | 19 ++++++---
 include/linux/acpi.h          | 10 ++++-
 include/linux/gpio/consumer.h |  2 +
 4 files changed, 47 insertions(+), 62 deletions(-)
Signature · 1 line
2.30.0
#1 lore raw

PATCH 1/3 gpiolib: acpi: Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk

On some systems the ACPI tables has wrong pin number and instead of having a relative one it provides an absolute one in the global GPIO number space.

Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk to cope with such cases.

Fixes ba8c90c61847 ("gpio: pca953x: Override IRQ for one of the expanders on Galileo Gen 2")

Depends-on: 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip")

---
 drivers/gpio/gpiolib-acpi.c   | 7 ++++++-
 include/linux/gpio/consumer.h | 2 ++
 2 files changed, 8 insertions(+), 1 deletion(-)
drivers/gpio/gpiolib-acpi.c +6 -1
index e37a57d0a2f0..026ef258d4b9 100644
@@ -677,6 +677,7 @@ static int acpi_populate_gpio_lookup(struct acpi_resource *ares, void *data)
 	if (!lookup->desc) {
 		const struct acpi_resource_gpio *agpio = &ares->data.gpio;
 		bool gpioint = agpio->connection_type == ACPI_RESOURCE_GPIO_TYPE_INT;
+		struct gpio_desc *desc;
 		u16 pin_index;
 
 		if (lookup->info.quirks & ACPI_GPIO_QUIRK_ONLY_GPIOIO && gpioint)
@@ -689,8 +690,12 @@ static int acpi_populate_gpio_lookup(struct acpi_resource *ares, void *data)
 		if (pin_index >= agpio->pin_table_length)
 			return 1;
 
-		lookup->desc = acpi_get_gpiod(agpio->resource_source.string_ptr,
+		if (lookup->info.quirks & ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER)
+			desc = gpio_to_desc(agpio->pin_table[pin_index]);
+		else
+			desc = acpi_get_gpiod(agpio->resource_source.string_ptr,
 					      agpio->pin_table[pin_index]);
+		lookup->desc = desc;
 		lookup->info.pin_config = agpio->pin_config;
 		lookup->info.debounce = agpio->debounce_timeout;
 		lookup->info.gpioint = gpioint;
include/linux/gpio/consumer.h +2 -0
index ef49307611d2..c73b25bc9213 100644
@@ -674,6 +674,8 @@ struct acpi_gpio_mapping {
  * get GpioIo type explicitly, this quirk may be used.
  */
 #define ACPI_GPIO_QUIRK_ONLY_GPIOIO		BIT(1)
+/* Use given pin as an absolute GPIO number in the system */
+#define ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER		BIT(2)
 
 	unsigned int quirks;
 };
Signature · 1 line
2.30.0
Acked-by on v1 1/3
#2 lore raw

Re: [PATCH v1 1/3] gpiolib: acpi: Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk

Quoted 9 lines · Andy Shevchenko

On some systems the ACPI tables has wrong pin number and instead of having a relative one it provides an absolute one in the global GPIO number space.

Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk to cope with such cases.

Fixes ba8c90c61847 ("gpio: pca953x: Override IRQ for one of the expanders on Galileo Gen 2")

Depends-on: 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip")

OH THE HORROR!
However, we discussed it before. It is as it is.

It's the right place to fix the problem, so:

Yours,
Linus Walleij

#3 lore raw

Re: [PATCH v1 1/3] gpiolib: acpi: Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk

Quoted 7 lines · Linus Walleij
Quoted 4 lines · Andy Shevchenko

On some systems the ACPI tables has wrong pin number and instead of having a relative one it provides an absolute one in the global GPIO number space.

Quoted 8 lines
Quoted 5 lines

Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk to cope with such cases.

Fixes ba8c90c61847 ("gpio: pca953x: Override IRQ for one of the expanders on Galileo Gen 2")

Depends-on: 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip")

OH THE HORROR!
However, we discussed it before. It is as it is.

Unfortunately :-( (And recently it seems MS does something really "creative" on ARM ACPI platform)

Quoted 2 lines

It's the right place to fix the problem, so:

I am waiting for Mika, but if he keeps silent let's say to the end of the day, I will submit it as is to the v5.12-rcX fixes.

Thanks!

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#4 lore raw

Re: [PATCH v1 1/3] gpiolib: acpi: Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk

Quoted 17 lines · Andy Shevchenko
Quoted 7 lines · Linus Walleij
Quoted 4 lines · Andy Shevchenko

On some systems the ACPI tables has wrong pin number and instead of having a relative one it provides an absolute one in the global GPIO number space.

Quoted 8 lines
Quoted 5 lines

Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk to cope with such cases.

Fixes ba8c90c61847 ("gpio: pca953x: Override IRQ for one of the expanders on Galileo Gen 2")

Depends-on: 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip")

OH THE HORROR!
However, we discussed it before. It is as it is.

Quoted 8 lines

Unfortunately :-( (And recently it seems MS does something really "creative" on ARM ACPI platform)

Quoted 2 lines

It's the right place to fix the problem, so:

I am waiting for Mika, but if he keeps silent let's say to the end of the day, I will submit it as is to the v5.12-rcX fixes.

Sorry for the delay - I somehow missed this. Feel free to add my ACK too.

#5 lore raw

Re: [PATCH v1 1/3] gpiolib: acpi: Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk

Quoted 21 lines · Mika Westerberg
Quoted 20 lines · Andy Shevchenko
Quoted 7 lines · Linus Walleij
Quoted 4 lines · Andy Shevchenko

On some systems the ACPI tables has wrong pin number and instead of having a relative one it provides an absolute one in the global GPIO number space.

Quoted 8 lines
Quoted 5 lines

Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk to cope with such cases.

Fixes ba8c90c61847 ("gpio: pca953x: Override IRQ for one of the expanders on Galileo Gen 2")

Depends-on: 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip")

OH THE HORROR!
However, we discussed it before. It is as it is.

Unfortunately :-( (And recently it seems MS does something really "creative" on ARM ACPI platform)

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

It's the right place to fix the problem, so:

I am waiting for Mika, but if he keeps silent let's say to the end of the day, I will submit it as is to the v5.12-rcX fixes.

Sorry for the delay - I somehow missed this. Feel free to add my ACK too.

Thanks, Mika!

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#6 lore raw

Re: [PATCH v1 1/3] gpiolib: acpi: Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk

Quoted 29 lines · Andy Shevchenko
Quoted 27 lines · Mika Westerberg
Quoted 25 lines · Andy Shevchenko
Quoted 7 lines · Linus Walleij
Quoted 4 lines · Andy Shevchenko

On some systems the ACPI tables has wrong pin number and instead of having a relative one it provides an absolute one in the global GPIO number space.

Quoted 8 lines
Quoted 5 lines

Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk to cope with such cases.

Fixes ba8c90c61847 ("gpio: pca953x: Override IRQ for one of the expanders on Galileo Gen 2")

Depends-on: 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip")

OH THE HORROR!
However, we discussed it before. It is as it is.

Unfortunately :-( (And recently it seems MS does something really "creative" on ARM ACPI platform)

Quoted 2 lines

It's the right place to fix the problem, so:

I am waiting for Mika, but if he keeps silent let's say to the end of the day, I will submit it as is to the v5.12-rcX fixes.

Quoted 10 lines
Quoted 1 line

Sorry for the delay - I somehow missed this. Feel free to add my ACK too.

Thanks, Mika!

--
With Best Regards,
Andy Shevchenko

Hi Andy,

Do you want me to take these, or will you include them in your PR?

Bart

#7 lore raw

Re: [PATCH v1 1/3] gpiolib: acpi: Add ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk

Quoted 4 lines · Bartosz Golaszewski

On Tue, Mar 2, 2021 at 4:21 PM Andy Shevchenko
<[email protected]> wrote:

Do you want me to take these, or will you include them in your PR?

I'll send PR.
I'm waiting for vger to process the queue (it has some lags).

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#8 lore raw

PATCH 2/3 gpiolib: acpi: Allow to find GpioInt() resource by name and index

In reply to Andy Shevchenko

Currently only search by index is supported. However, in some cases we might need to pass the quirks to the acpi_dev_gpio_irq_get().

For this, split out acpi_dev_gpio_irq_get_by() and replace acpi_dev_gpio_irq_get() by calling above with NULL for name parameter.

Fixes ba8c90c61847 ("gpio: pca953x: Override IRQ for one of the expanders on Galileo Gen 2")

Depends-on: 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip")

---
 drivers/gpio/gpiolib-acpi.c | 12 ++++++++----
 include/linux/acpi.h        | 10 ++++++++--
 2 files changed, 16 insertions(+), 6 deletions(-)
drivers/gpio/gpiolib-acpi.c +8 -4
index 026ef258d4b9..495f779b2ab9 100644
@@ -945,8 +945,9 @@ struct gpio_desc *acpi_node_get_gpiod(struct fwnode_handle *fwnode,
 }
 
 /**
- * acpi_dev_gpio_irq_get() - Find GpioInt and translate it to Linux IRQ number
+ * acpi_dev_gpio_irq_get_by() - Find GpioInt and translate it to Linux IRQ number
  * @adev: pointer to a ACPI device to get IRQ from
+ * @name: optional name of GpioInt resource
  * @index: index of GpioInt resource (starting from %0)
  *
  * If the device has one or more GpioInt resources, this function can be
@@ -956,9 +957,12 @@ struct gpio_desc *acpi_node_get_gpiod(struct fwnode_handle *fwnode,
  * The function is idempotent, though each time it runs it will configure GPIO
  * pin direction according to the flags in GpioInt resource.
  *
+ * The function takes optional @name parameter. If the resource has a property
+ * name, then only those will be taken into account.
+ *
  * Return: Linux IRQ number (> %0) on success, negative errno on failure.
  */
-int acpi_dev_gpio_irq_get(struct acpi_device *adev, int index)
+int acpi_dev_gpio_irq_get_by(struct acpi_device *adev, const char *name, int index)
 {
 	int idx, i;
 	unsigned int irq_flags;
@@ -968,7 +972,7 @@ int acpi_dev_gpio_irq_get(struct acpi_device *adev, int index)
 		struct acpi_gpio_info info;
 		struct gpio_desc *desc;
 
-		desc = acpi_get_gpiod_by_index(adev, NULL, i, &info);
+		desc = acpi_get_gpiod_by_index(adev, name, i, &info);
 
 		/* Ignore -EPROBE_DEFER, it only matters if idx matches */
 		if (IS_ERR(desc) && PTR_ERR(desc) != -EPROBE_DEFER)
@@ -1013,7 +1017,7 @@ int acpi_dev_gpio_irq_get(struct acpi_device *adev, int index)
 	}
 	return -ENOENT;
 }
-EXPORT_SYMBOL_GPL(acpi_dev_gpio_irq_get);
+EXPORT_SYMBOL_GPL(acpi_dev_gpio_irq_get_by);
 
 static acpi_status
 acpi_gpio_adr_space_handler(u32 function, acpi_physical_address address,
include/linux/acpi.h +8 -2
index 053bf05fb1f7..b20568c44001 100644
@@ -1072,19 +1072,25 @@ void __acpi_handle_debug(struct _ddebug *descriptor, acpi_handle handle, const c
 #if defined(CONFIG_ACPI) && defined(CONFIG_GPIOLIB)
 bool acpi_gpio_get_irq_resource(struct acpi_resource *ares,
 				struct acpi_resource_gpio **agpio);
-int acpi_dev_gpio_irq_get(struct acpi_device *adev, int index);
+int acpi_dev_gpio_irq_get_by(struct acpi_device *adev, const char *name, int index);
 #else
 static inline bool acpi_gpio_get_irq_resource(struct acpi_resource *ares,
 					      struct acpi_resource_gpio **agpio)
 {
 	return false;
 }
-static inline int acpi_dev_gpio_irq_get(struct acpi_device *adev, int index)
+static inline int acpi_dev_gpio_irq_get_by(struct acpi_device *adev,
+					   const char *name, int index)
 {
 	return -ENXIO;
 }
 #endif
 
+static inline int acpi_dev_gpio_irq_get(struct acpi_device *adev, int index)
+{
+	return acpi_dev_gpio_irq_get_by(adev, NULL, index);
+}
+
 /* Device properties */
 
 #ifdef CONFIG_ACPI
Signature · 1 line
2.30.0
Acked-by on v1 2/3
#9 lore raw

Re: [PATCH v1 2/3] gpiolib: acpi: Allow to find GpioInt() resource by name and index

Quoted 9 lines · Andy Shevchenko

Currently only search by index is supported. However, in some cases we might need to pass the quirks to the acpi_dev_gpio_irq_get().

For this, split out acpi_dev_gpio_irq_get_by() and replace acpi_dev_gpio_irq_get() by calling above with NULL for name parameter.

Fixes ba8c90c61847 ("gpio: pca953x: Override IRQ for one of the expanders on Galileo Gen 2")

Depends-on: 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip")

Yours,
Linus Walleij

#10 lore raw

PATCH 3/3 gpio: pca953x: Set IRQ type when handle Intel Galileo Gen 2

In reply to Andy Shevchenko

The commit 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip") indeliberately made a regression on how IRQ line from GPIO I²C expander is handled. I.e. it reveals that the quirk for Intel Galileo Gen 2 misses the part of setting IRQ type which previously was predefined by gpio-dwapb driver. Now, we have to reorganize the approach to call necessary parts, which can be done via ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk.

Without this fix and with above mentioned change the kernel hangs on the first IRQ event with:

    gpio gpiochip3: Persistence not supported for GPIO 1
    irq 32, desc: 62f8fb50, depth: 0, count: 0, unhandled: 0
    ->handle_irq():  41c7b0ab, handle_bad_irq+0x0/0x40
    ->irq_data.chip(): e03f1e72, 0xc2539218
    ->action(): 0ecc7e6f
    ->action->handler(): 8a3db21e, irq_default_primary_handler+0x0/0x10
       IRQ_NOPROBE set
    unexpected IRQ trap at vector 20
Fixes ba8c90c61847 ("gpio: pca953x: Override IRQ for one of the expanders on Galileo Gen 2")

Depends-on: 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip")

---
 drivers/gpio/gpio-pca953x.c | 78 +++++++++++--------------------------
 1 file changed, 23 insertions(+), 55 deletions(-)
drivers/gpio/gpio-pca953x.c +23 -55
index 825b362eb4b7..6898c27f71f8 100644
@@ -112,8 +112,29 @@ MODULE_DEVICE_TABLE(i2c, pca953x_id);
 #ifdef CONFIG_GPIO_PCA953X_IRQ
 
 #include <linux/dmi.h>
-#include <linux/gpio.h>
-#include <linux/list.h>
+
+static const struct acpi_gpio_params pca953x_irq_gpios = { 0, 0, true };
+
+static const struct acpi_gpio_mapping pca953x_acpi_irq_gpios[] = {
+	{ "irq-gpios", &pca953x_irq_gpios, 1, ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER },
+	{ }
+};
+
+static int pca953x_acpi_get_irq(struct device *dev)
+{
+	int ret;
+
+	ret = devm_acpi_dev_add_driver_gpios(dev, pca953x_acpi_irq_gpios);
+	if (ret)
+		dev_warn(dev, "can't add GPIO ACPI mapping\n");
+
+	ret = acpi_dev_gpio_irq_get_by(ACPI_COMPANION(dev), "irq-gpios", 0);
+	if (ret < 0)
+		return ret;
+
+	dev_info(dev, "ACPI interrupt quirk (IRQ %d)\n", ret);
+	return ret;
+}
 
 static const struct dmi_system_id pca953x_dmi_acpi_irq_info[] = {
 	{
@@ -132,59 +153,6 @@ static const struct dmi_system_id pca953x_dmi_acpi_irq_info[] = {
 	},
 	{}
 };
-
-#ifdef CONFIG_ACPI
-static int pca953x_acpi_get_pin(struct acpi_resource *ares, void *data)
-{
-	struct acpi_resource_gpio *agpio;
-	int *pin = data;
-
-	if (acpi_gpio_get_irq_resource(ares, &agpio))
-		*pin = agpio->pin_table[0];
-	return 1;
-}
-
-static int pca953x_acpi_find_pin(struct device *dev)
-{
-	struct acpi_device *adev = ACPI_COMPANION(dev);
-	int pin = -ENOENT, ret;
-	LIST_HEAD(r);
-
-	ret = acpi_dev_get_resources(adev, &r, pca953x_acpi_get_pin, &pin);
-	acpi_dev_free_resource_list(&r);
-	if (ret < 0)
-		return ret;
-
-	return pin;
-}
-#else
-static inline int pca953x_acpi_find_pin(struct device *dev) { return -ENXIO; }
-#endif
-
-static int pca953x_acpi_get_irq(struct device *dev)
-{
-	int pin, ret;
-
-	pin = pca953x_acpi_find_pin(dev);
-	if (pin < 0)
-		return pin;
-
-	dev_info(dev, "Applying ACPI interrupt quirk (GPIO %d)\n", pin);
-
-	if (!gpio_is_valid(pin))
-		return -EINVAL;
-
-	ret = gpio_request(pin, "pca953x interrupt");
-	if (ret)
-		return ret;
-
-	ret = gpio_to_irq(pin);
-
-	/* When pin is used as an IRQ, no need to keep it requested */
-	gpio_free(pin);
-
-	return ret;
-}
 #endif
 
 static const struct acpi_device_id pca953x_acpi_ids[] = {
Signature · 1 line
2.30.0
Reviewed-by on v1 3/3
#11 lore raw

Re: [PATCH v1 3/3] gpio: pca953x: Set IRQ type when handle Intel Galileo Gen 2

Quoted 20 lines · Andy Shevchenko

The commit 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip") indeliberately made a regression on how IRQ line from GPIO I²C expander is handled. I.e. it reveals that the quirk for Intel Galileo Gen 2 misses the part of setting IRQ type which previously was predefined by gpio-dwapb driver. Now, we have to reorganize the approach to call necessary parts, which can be done via ACPI_GPIO_QUIRK_ABSOLUTE_NUMBER quirk.

Without this fix and with above mentioned change the kernel hangs on the first IRQ event with:

    gpio gpiochip3: Persistence not supported for GPIO 1
    irq 32, desc: 62f8fb50, depth: 0, count: 0, unhandled: 0
    ->handle_irq():  41c7b0ab, handle_bad_irq+0x0/0x40
    ->irq_data.chip(): e03f1e72, 0xc2539218
    ->action(): 0ecc7e6f
    ->action->handler(): 8a3db21e, irq_default_primary_handler+0x0/0x10
       IRQ_NOPROBE set
    unexpected IRQ trap at vector 20
Quoted 2 lines
Fixes ba8c90c61847 ("gpio: pca953x: Override IRQ for one of the expanders on Galileo Gen 2")

Depends-on: 0ea683931adb ("gpio: dwapb: Convert driver to using the GPIO-lib-based IRQ-chip")

I never saw that before, seems useful!

Quoted 1 line

Not only does it fix the bug, it also looks so much better by separation of concerns.

Yours,
Linus Walleij