Skip to contentSkip to conversation

gpiolib: Unify access to the device properties

PATCH In review Andy Shevchenko 4 participants 8 messages Nov 16, 2022 – Nov 28, 2022 GPIO

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 gpiolib: Unify access to the device properties

Some of the functions are using struct fwnode_handle, some struct device pointer. In the GPIO library the firmware node of the GPIO device is the same as GPIO node of the GPIO chip. Due to this fact we may use former to access properties everywhere in the code.

---
 drivers/gpio/gpiolib.c | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)
drivers/gpio/gpiolib.c +9 -8
index 51afdc6ac919..c163b354e727 100644
@@ -367,12 +367,12 @@ static int gpiochip_set_desc_names(struct gpio_chip *gc)
 static int devprop_gpiochip_set_names(struct gpio_chip *chip)
 {
 	struct gpio_device *gdev = chip->gpiodev;
-	const struct fwnode_handle *fwnode = dev_fwnode(&gdev->dev);
+	struct device *dev = &gdev->dev;
 	const char **names;
 	int ret, i;
 	int count;
 
-	count = fwnode_property_string_array_count(fwnode, "gpio-line-names");
+	count = device_property_string_array_count(dev, "gpio-line-names");
 	if (count < 0)
 		return 0;
 
@@ -385,7 +385,7 @@ static int devprop_gpiochip_set_names(struct gpio_chip *chip)
 	 * gpiochips.
 	 */
 	if (count <= chip->offset) {
-		dev_warn(&gdev->dev, "gpio-line-names too short (length %d), cannot map names for the gpiochip at offset %u\n",
+		dev_warn(dev, "gpio-line-names too short (length %d), cannot map names for the gpiochip at offset %u\n",
 			 count, chip->offset);
 		return 0;
 	}
@@ -394,10 +394,9 @@ static int devprop_gpiochip_set_names(struct gpio_chip *chip)
 	if (!names)
 		return -ENOMEM;
 
-	ret = fwnode_property_read_string_array(fwnode, "gpio-line-names",
-						names, count);
+	ret = device_property_read_string_array(dev, "gpio-line-names", names, count);
 	if (ret < 0) {
-		dev_warn(&gdev->dev, "failed to read GPIO line names\n");
+		dev_warn(dev, "failed to read GPIO line names\n");
 		kfree(names);
 		return ret;
 	}
@@ -448,10 +447,11 @@ static unsigned long *gpiochip_allocate_mask(struct gpio_chip *gc)
 
 static unsigned int gpiochip_count_reserved_ranges(struct gpio_chip *gc)
 {
+	struct device *dev = &gc->gpiodev->dev;
 	int size;
 
 	/* Format is "start, count, ..." */
-	size = fwnode_property_count_u32(gc->fwnode, "gpio-reserved-ranges");
+	size = device_property_count_u32(dev, "gpio-reserved-ranges");
 	if (size > 0 && size % 2 == 0)
 		return size;
 
@@ -472,6 +472,7 @@ static int gpiochip_alloc_valid_mask(struct gpio_chip *gc)
 
 static int gpiochip_apply_reserved_ranges(struct gpio_chip *gc)
 {
+	struct device *dev = &gc->gpiodev->dev;
 	unsigned int size;
 	u32 *ranges;
 	int ret;
@@ -484,7 +485,7 @@ static int gpiochip_apply_reserved_ranges(struct gpio_chip *gc)
 	if (!ranges)
 		return -ENOMEM;
 
-	ret = fwnode_property_read_u32_array(gc->fwnode, "gpio-reserved-ranges", ranges, size);
+	ret = device_property_read_u32_array(dev, "gpio-reserved-ranges", ranges, size);
 	if (ret) {
 		kfree(ranges);
 		return ret;

base-commit: 40059212f99c31f26c69763e560325e59eac02c6

Signature · 1 line
2.35.1
#1 lore raw
Quoted 4 lines · Andy Shevchenko

Some of the functions are using struct fwnode_handle, some struct device pointer. In the GPIO library the firmware node of the GPIO device is the same as GPIO node of the GPIO chip. Due to this fact we may use former to access properties everywhere in the code.

Citing myself from another thread to the topic:

Nevertheless, for of_gpiochip_add()/of_gpiochip_remove() and of_mm_gpiochip_add_data() I still left use of fwnode, because it feels the right thing to do: we are taking reference on the input data in such cases.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
Tested-by on v1 1/1
#3 lore raw
In reply to Andy Shevchenko
Quoted 6 lines · Andy Shevchenko

Some of the functions are using struct fwnode_handle, some struct device pointer. In the GPIO library the firmware node of the GPIO device is the same as GPIO node of the GPIO chip. Due to this fact we may use former to access properties everywhere in the code.

I presume this is a more complete version of [1], as this also happens to address the crash on MSM [1], hence:

Just in case, this seems to depend on [3] or it fails to apply cleanly because of the const change on fwnode (consider this irrelevant if it has already been applied, I'm still on -next-20221115 for this integration branch).

[1]: https://lore.kernel.org/all/Y3S62i7OzocP5QrT@orome/
[2]: https://lore.kernel.org/linux-arm-msm/[email protected]/
[3]: https://lore.kernel.org/linux-gpio/[email protected]/

- Marijn

Quoted 80 lines
---
 drivers/gpio/gpiolib.c | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)
drivers/gpio/gpiolib.c +9 -8
index 51afdc6ac919..c163b354e727 100644
@@ -367,12 +367,12 @@ static int gpiochip_set_desc_names(struct gpio_chip *gc)
 static int devprop_gpiochip_set_names(struct gpio_chip *chip)
 {
 	struct gpio_device *gdev = chip->gpiodev;
-	const struct fwnode_handle *fwnode = dev_fwnode(&gdev->dev);
+	struct device *dev = &gdev->dev;
 	const char **names;
 	int ret, i;
 	int count;
 
-	count = fwnode_property_string_array_count(fwnode, "gpio-line-names");
+	count = device_property_string_array_count(dev, "gpio-line-names");
 	if (count < 0)
 		return 0;
 
@@ -385,7 +385,7 @@ static int devprop_gpiochip_set_names(struct gpio_chip *chip)
 	 * gpiochips.
 	 */
 	if (count <= chip->offset) {
-		dev_warn(&gdev->dev, "gpio-line-names too short (length %d), cannot map names for the gpiochip at offset %u\n",
+		dev_warn(dev, "gpio-line-names too short (length %d), cannot map names for the gpiochip at offset %u\n",
 			 count, chip->offset);
 		return 0;
 	}
@@ -394,10 +394,9 @@ static int devprop_gpiochip_set_names(struct gpio_chip *chip)
 	if (!names)
 		return -ENOMEM;
 
-	ret = fwnode_property_read_string_array(fwnode, "gpio-line-names",
-						names, count);
+	ret = device_property_read_string_array(dev, "gpio-line-names", names, count);
 	if (ret < 0) {
-		dev_warn(&gdev->dev, "failed to read GPIO line names\n");
+		dev_warn(dev, "failed to read GPIO line names\n");
 		kfree(names);
 		return ret;
 	}
@@ -448,10 +447,11 @@ static unsigned long *gpiochip_allocate_mask(struct gpio_chip *gc)
 
 static unsigned int gpiochip_count_reserved_ranges(struct gpio_chip *gc)
 {
+	struct device *dev = &gc->gpiodev->dev;
 	int size;
 
 	/* Format is "start, count, ..." */
-	size = fwnode_property_count_u32(gc->fwnode, "gpio-reserved-ranges");
+	size = device_property_count_u32(dev, "gpio-reserved-ranges");
 	if (size > 0 && size % 2 == 0)
 		return size;
 
@@ -472,6 +472,7 @@ static int gpiochip_alloc_valid_mask(struct gpio_chip *gc)
 
 static int gpiochip_apply_reserved_ranges(struct gpio_chip *gc)
 {
+	struct device *dev = &gc->gpiodev->dev;
 	unsigned int size;
 	u32 *ranges;
 	int ret;
@@ -484,7 +485,7 @@ static int gpiochip_apply_reserved_ranges(struct gpio_chip *gc)
 	if (!ranges)
 		return -ENOMEM;
 
-	ret = fwnode_property_read_u32_array(gc->fwnode, "gpio-reserved-ranges", ranges, size);
+	ret = device_property_read_u32_array(dev, "gpio-reserved-ranges", ranges, size);
 	if (ret) {
 		kfree(ranges);
 		return ret;

base-commit: 40059212f99c31f26c69763e560325e59eac02c6

Signature · 1 line
2.35.1
#4 lore raw
Quoted 6 lines · Marijn Suijten
Quoted 5 lines · Andy Shevchenko

Some of the functions are using struct fwnode_handle, some struct device pointer. In the GPIO library the firmware node of the GPIO device is the same as GPIO node of the GPIO chip. Due to this fact we may use former to access properties everywhere in the code.

Quoted 6 lines
Quoted 1 line

I presume this is a more complete version of [1], as this also happens to address the crash on MSM [1], hence:

Thank you!

Quoted 4 lines

Just in case, this seems to depend on [3] or it fails to apply cleanly because of the const change on fwnode (consider this irrelevant if it has already been applied, I'm still on -next-20221115 for this integration branch).

But initial issue was induced by (my) patch which is only in gpio/for-next. It's supposed to be only Linux Next issue.

Quoted 3 lines

[1]: https://lore.kernel.org/all/Y3S62i7OzocP5QrT@orome/
[2]: https://lore.kernel.org/linux-arm-msm/[email protected]/
[3]: https://lore.kernel.org/linux-gpio/[email protected]/

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#5 lore raw
In reply to Andy Shevchenko
Quoted 4 lines · Andy Shevchenko

Some of the functions are using struct fwnode_handle, some struct device pointer. In the GPIO library the firmware node of the GPIO device is the same as GPIO node of the GPIO chip. Due to this fact we may use former to access properties everywhere in the code.

Bart, can this be applied?

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#6 lore raw
Quoted 7 lines · Andy Shevchenko
Quoted 4 lines · Andy Shevchenko

Some of the functions are using struct fwnode_handle, some struct device pointer. In the GPIO library the firmware node of the GPIO device is the same as GPIO node of the GPIO chip. Due to this fact we may use former to access properties everywhere in the code.

Quoted 7 lines

Bart, can this be applied?

--
With Best Regards,
Andy Shevchenko

Sure, now applied. Got carried away with this use-after-free from userspace problem.

Bart

#7 lore raw
Quoted 3 lines · Bartosz Golaszewski
Quoted 1 line · Andy Shevchenko

On Wed, Nov 16, 2022 at 04:17:28PM +0200, Andy Shevchenko wrote:

...

Quoted 4 lines
Quoted 1 line

Bart, can this be applied?

Sure, now applied. Got carried away with this use-after-free from userspace problem.

Thank you!

Signature · 2 lines
With Best Regards,
Andy Shevchenko