Skip to contentSkip to conversation

gpiolib: fix reference leaks when removing GPIO chips still in use

PATCH Superseded Bartosz Golaszewski 3 participants 6 messages Aug 10, 2023 – Aug 11, 2023 GPIO

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Bartosz Golaszewski

author
#0 lore raw

PATCH v2 1/1 gpiolib: fix reference leaks when removing GPIO chips still in use

From: Bartosz Golaszewski <[email protected]>

After we remove a GPIO chip that still has some requested descriptors, gpiod_free_commit() will fail and we will never put the references to the GPIO device and the owning module in gpiod_free().

Rework this function to:
- not warn on desc == NULL as this is a use-case on which most free functions silently return
- put the references to desc->gdev and desc->gdev->owner unconditionally so that the release callback actually gets called when the remaining references are dropped by external GPIO users

Signed-off-by Bartosz Golaszewski <[email protected]>

---
v1 -> v2:
- add a comment about why we can't use VALIDATE_DESC_VOID()

 drivers/gpio/gpiolib.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)
drivers/gpio/gpiolib.c +11 -5
index 251c875b5c34..2158067c4fd2 100644
@@ -2167,12 +2167,18 @@ static bool gpiod_free_commit(struct gpio_desc *desc)
 
 void gpiod_free(struct gpio_desc *desc)
 {
-	if (desc && desc->gdev && gpiod_free_commit(desc)) {
-		module_put(desc->gdev->owner);
-		gpio_device_put(desc->gdev);
-	} else {
+	/*
+	 * We must not use VALIDATE_DESC_VOID() as the underlying gdev->chip
+	 * may already be NULL but we still want to put the references.
+	 */
+	if (!desc)
+		return;
+
+	if (!gpiod_free_commit(desc))
 		WARN_ON(extra_checks);
-	}
+
+	gpio_device_put(desc->gdev);
+	module_put(desc->gdev->owner);
 }
 
 /**
Signature · 1 line
2.39.2
author
#1 lore raw
Quoted 7 lines · Bartosz Golaszewski

From: Bartosz Golaszewski <[email protected]>

After we remove a GPIO chip that still has some requested descriptors, gpiod_free_commit() will fail and we will never put the references to the GPIO device and the owning module in gpiod_free().

Quoted 8 lines

Rework this function to:
- not warn on desc == NULL as this is a use-case on which most free functions silently return
- put the references to desc->gdev and desc->gdev->owner unconditionally so that the release callback actually gets called when the remaining references are dropped by external GPIO users

Signed-off-by Bartosz Golaszewski <[email protected]>

I forgot to pick up Linus' tag. I'll do it when applying. I should really start using b4...

Bart

Quoted 38 lines

---
v1 -> v2:
- add a comment about why we can't use VALIDATE_DESC_VOID()

 drivers/gpio/gpiolib.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)
drivers/gpio/gpiolib.c +11 -6
index 251c875b5c34..2158067c4fd2 100644
@@ -2167,12 +2167,18 @@ static bool gpiod_free_commit(struct gpio_desc *desc)
 
 void gpiod_free(struct gpio_desc *desc)
 {
-       if (desc && desc->gdev && gpiod_free_commit(desc)) {
-               module_put(desc->gdev->owner);
-               gpio_device_put(desc->gdev);
-       } else {
+       /*
+        * We must not use VALIDATE_DESC_VOID() as the underlying gdev->chip
+        * may already be NULL but we still want to put the references.
+        */
+       if (!desc)
+               return;
+
+       if (!gpiod_free_commit(desc))
                WARN_ON(extra_checks);
-       }
+
+       gpio_device_put(desc->gdev);
+       module_put(desc->gdev->owner);
 }
 
 /**
--

2.39.2

#2 lore raw
Quoted 2 lines · Bartosz Golaszewski

I forgot to pick up Linus' tag. I'll do it when applying. I should really start using b4...

Well with b4 you still have to issue b4 trailers -u, I almost want to make a cron script that just loop over my b4 branches and do that every night when I'm asleep so tags get picked up in case I miss them.

Hm, maybe not such a bad idea.

Yours,
Linus Walleij

#3 lore raw
In reply to Bartosz Golaszewski
Quoted 12 lines · Bartosz Golaszewski

From: Bartosz Golaszewski <[email protected]>

After we remove a GPIO chip that still has some requested descriptors, gpiod_free_commit() will fail and we will never put the references to the GPIO device and the owning module in gpiod_free().

Rework this function to:
- not warn on desc == NULL as this is a use-case on which most free functions silently return
- put the references to desc->gdev and desc->gdev->owner unconditionally so that the release callback actually gets called when the remaining references are dropped by external GPIO users

...

drivers/gpio/gpiolib.c static bool gpiod_free_commit(struct gpio_desc *desc) · line 2173

Quoted 4 lines
drivers/gpio/gpiolib.c +4 -0
+	/*
+	 * We must not use VALIDATE_DESC_VOID() as the underlying gdev->chip
+	 * may already be NULL but we still want to put the references.
+	 */

Clear now!

...

drivers/gpio/gpiolib.c static bool gpiod_free_commit(struct gpio_desc *desc) · line 2178

Quoted 2 lines
drivers/gpio/gpiolib.c +1 -0
+	if (!gpiod_free_commit(desc))
 		WARN_ON(extra_checks);

	/* Always call gpiod_free_commit() */
	WARN_ON(!gpiod_free_commit(desc) && extra_checks);

?

Signature · 2 lines
With Best Regards,
Andy Shevchenko
author
#4 lore raw
In reply to Bartosz Golaszewski
Quoted 51 lines · Bartosz Golaszewski

From: Bartosz Golaszewski <[email protected]>

After we remove a GPIO chip that still has some requested descriptors, gpiod_free_commit() will fail and we will never put the references to the GPIO device and the owning module in gpiod_free().

Rework this function to:
- not warn on desc == NULL as this is a use-case on which most free functions silently return
- put the references to desc->gdev and desc->gdev->owner unconditionally so that the release callback actually gets called when the remaining references are dropped by external GPIO users

Signed-off-by Bartosz Golaszewski <[email protected]>

---
v1 -> v2:
- add a comment about why we can't use VALIDATE_DESC_VOID()

 drivers/gpio/gpiolib.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)
drivers/gpio/gpiolib.c +11 -6
index 251c875b5c34..2158067c4fd2 100644
@@ -2167,12 +2167,18 @@ static bool gpiod_free_commit(struct gpio_desc *desc)
 
 void gpiod_free(struct gpio_desc *desc)
 {
-       if (desc && desc->gdev && gpiod_free_commit(desc)) {
-               module_put(desc->gdev->owner);
-               gpio_device_put(desc->gdev);
-       } else {
+       /*
+        * We must not use VALIDATE_DESC_VOID() as the underlying gdev->chip
+        * may already be NULL but we still want to put the references.
+        */
+       if (!desc)
+               return;
+
+       if (!gpiod_free_commit(desc))
                WARN_ON(extra_checks);
-       }
+
+       gpio_device_put(desc->gdev);
+       module_put(desc->gdev->owner);
 }
 
 /**
--
Quoted 2 lines

2.39.2

Queued for fixes.

Bartosz

author
#5 lore raw
Quoted 51 lines · Bartosz Golaszewski
Quoted 49 lines · Bartosz Golaszewski

From: Bartosz Golaszewski <[email protected]>

After we remove a GPIO chip that still has some requested descriptors, gpiod_free_commit() will fail and we will never put the references to the GPIO device and the owning module in gpiod_free().

Rework this function to:
- not warn on desc == NULL as this is a use-case on which most free functions silently return
- put the references to desc->gdev and desc->gdev->owner unconditionally so that the release callback actually gets called when the remaining references are dropped by external GPIO users

Signed-off-by Bartosz Golaszewski <[email protected]>

---
v1 -> v2:
- add a comment about why we can't use VALIDATE_DESC_VOID()

 drivers/gpio/gpiolib.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)
drivers/gpio/gpiolib.c +11 -5
index 251c875b5c34..2158067c4fd2 100644
@@ -2167,12 +2167,18 @@ static bool gpiod_free_commit(struct gpio_desc *desc)
 
 void gpiod_free(struct gpio_desc *desc)
 {
-       if (desc && desc->gdev && gpiod_free_commit(desc)) {
-               module_put(desc->gdev->owner);
-               gpio_device_put(desc->gdev);
-       } else {
+       /*
+        * We must not use VALIDATE_DESC_VOID() as the underlying gdev->chip
+        * may already be NULL but we still want to put the references.
+        */
+       if (!desc)
+               return;
+
+       if (!gpiod_free_commit(desc))
                WARN_ON(extra_checks);
-       }
+
+       gpio_device_put(desc->gdev);
+       module_put(desc->gdev->owner);
 }
Quoted 8 lines
Quoted 4 lines
drivers/gpio/gpiolib.c +0 -1
 /**
--

2.39.2

Queued for fixes.

Bartosz

Eek, I need to back it out. I realized there's an issue with this patch. I'll send a v3.

Bartosz