[PATCH v2] leds: tca6507: Fix use-after-free on unbind

[PATCH v2] leds: tca6507: Fix use-after-free on unbind

[Danish Khateeb]

Since the LEDs are registered with devm_led_classdev_register(), they
are unregistered only after tca6507_remove() has cancelled the work.
Unregistering an LED that is on turns it off, which queues the work
again from led_assign(), and the work then runs after devres has freed
the chip:

  BUG: KASAN: slab-use-after-free in assign_work+0x414/0x5b0
  Read of size 8 at addr ffff888010d5c0c0 by task kworker/1:2/65
  ...
  Last potentially related work creation:
   ...
   queue_work_on+0xb6/0xc0
   led_assign+0x223/0x340 [leds_tca6507]
   led_classdev_unregister+0x26b/0x340
   release_nodes+0xb3/0x140
   devres_release_group+0x247/0x470
   i2c_device_remove+0xbf/0x1b0

Set up the work with devm_work_autocancel() before the LEDs are
registered, so that devres cancels it after they are gone, and drop
tca6507_remove().

Fixes: 1b5c2fa7081c ("leds: tca6507: Use devm_led_classdev_register() to simplify remove path")
Cc: stable@vger.kernel.org
Acked-by: Andrew Davis <afd@ti.com>
Assisted-by: LLM
Signed-off-by: Danish Khateeb <danishkhateeb03@gmail.com>
---

Notes:
    Changes in v2:
    - Move the new include to the top of the list, where it sorts
      alphabetically (Lee)
    - Drop the comment above devm_work_autocancel() (Lee)
    - Add Andrew's Acked-by
    - Link to v1: https://lore.kernel.org/all/20260926172115.543800-1-danishkhateeb03@gmail.com/
    
    Tested in QEMU on a next-20260925 KASAN kernel, with the driver built as
    a module and a test-only module that adds a fake I2C adapter and a
    software node for a TCA6507 with two LEDs. With the LEDs on, unbinding
    hits the report above (38 reports in 20 unbind/bind cycles), and so does
    unloading the driver in 2 of 3 runs. With this patch there are no
    reports in 20 cycles or 3 unloads, and the LEDs come back after each
    rebind. A W=1 build with GPIOLIB is clean.
    
    v2 was retested the same way on next-20260930: 39 reports in 20 cycles
    without the patch, none with it in 20 cycles or 3 unloads.

 drivers/leds/leds-tca6507.c | 15 +++++----------
 1 file changed, 5 insertions(+), 10 deletions(-)

diff --git a/drivers/leds/leds-tca6507.c b/drivers/leds/leds-tca6507.c
index 9afe2722986c..0f987e551f03 100644
--- a/drivers/leds/leds-tca6507.c
+++ b/drivers/leds/leds-tca6507.c
@@ -71,6 +71,7 @@
  * not recognised when matching.
  */
 
+#include <linux/devm-helpers.h>
 #include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/leds.h>
@@ -735,9 +736,11 @@ static int tca6507_probe(struct i2c_client *client)
 		return -ENOMEM;
 
 	tca->client = client;
-	INIT_WORK(&tca->work, tca6507_work);
 	spin_lock_init(&tca->lock);
-	i2c_set_clientdata(client, tca);
+
+	err = devm_work_autocancel(dev, &tca->work, tca6507_work);
+	if (err)
+		return err;
 
 	for (i = 0; i < NUM_LEDS; i++) {
 		struct tca6507_led *l = tca->leds + i;
@@ -766,20 +769,12 @@ static int tca6507_probe(struct i2c_client *client)
 	return 0;
 }
 
-static void tca6507_remove(struct i2c_client *client)
-{
-	struct tca6507_chip *tca = i2c_get_clientdata(client);
-
-	cancel_work_sync(&tca->work);
-}
-
 static struct i2c_driver tca6507_driver = {
 	.driver   = {
 		.name    = "leds-tca6507",
 		.of_match_table = of_match_ptr(of_tca6507_leds_match),
 	},
 	.probe    = tca6507_probe,
-	.remove   = tca6507_remove,
 	.id_table = tca6507_id,
 };
 

base-commit: 05b4738b0078f7d6f154f68068a11c8a0635e9df
-- 
2.56.0

Re: [PATCH v2] leds: tca6507: Fix use-after-free on unbind

[Lee Jones]

On Thu, 15 Oct 2026, Danish Khateeb wrote:

Please review, explain or fix the Sashiko review(s).

> Since the LEDs are registered with devm_led_classdev_register(), they
> are unregistered only after tca6507_remove() has cancelled the work.
> Unregistering an LED that is on turns it off, which queues the work
> again from led_assign(), and the work then runs after devres has freed
> the chip:
>
>   BUG: KASAN: slab-use-after-free in assign_work+0x414/0x5b0
>   Read of size 8 at addr ffff888010d5c0c0 by task kworker/1:2/65
>   ...
>   Last potentially related work creation:
>    ...
>    queue_work_on+0xb6/0xc0
>    led_assign+0x223/0x340 [leds_tca6507]
>    led_classdev_unregister+0x26b/0x340
>    release_nodes+0xb3/0x140
>    devres_release_group+0x247/0x470
>    i2c_device_remove+0xbf/0x1b0
>
> Set up the work with devm_work_autocancel() before the LEDs are
> registered, so that devres cancels it after they are gone, and drop
> tca6507_remove().
>
> Fixes: 1b5c2fa7081c ("leds: tca6507: Use devm_led_classdev_register() to simplify remove path")
> Cc: stable@vger.kernel.org
> Acked-by: Andrew Davis <afd@ti.com>
> Assisted-by: LLM
> Signed-off-by: Danish Khateeb <danishkhateeb03@gmail.com>
> ---
>
> Notes:
>     Changes in v2:
>     - Move the new include to the top of the list, where it sorts
>       alphabetically (Lee)
>     - Drop the comment above devm_work_autocancel() (Lee)
>     - Add Andrew's Acked-by
>     - Link to v1: https://lore.kernel.org/all/20260926172115.543800-1-danishkhateeb03@gmail.com/
>
>     Tested in QEMU on a next-20260925 KASAN kernel, with the driver built as
>     a module and a test-only module that adds a fake I2C adapter and a
>     software node for a TCA6507 with two LEDs. With the LEDs on, unbinding
>     hits the report above (38 reports in 20 unbind/bind cycles), and so does
>     unloading the driver in 2 of 3 runs. With this patch there are no
>     reports in 20 cycles or 3 unloads, and the LEDs come back after each
>     rebind. A W=1 build with GPIOLIB is clean.
>
>     v2 was retested the same way on next-20260930: 39 reports in 20 cycles
>     without the patch, none with it in 20 cycles or 3 unloads.
>
>  drivers/leds/leds-tca6507.c | 15 +++++----------
>  1 file changed, 5 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/leds/leds-tca6507.c b/drivers/leds/leds-tca6507.c
> index 9afe2722986c..0f987e551f03 100644
> --- a/drivers/leds/leds-tca6507.c
> +++ b/drivers/leds/leds-tca6507.c
> @@ -71,6 +71,7 @@
>   * not recognised when matching.
>   */
>
> +#include <linux/devm-helpers.h>
>  #include <linux/module.h>
>  #include <linux/slab.h>
>  #include <linux/leds.h>
> @@ -735,9 +736,11 @@ static int tca6507_probe(struct i2c_client *client)
>  		return -ENOMEM;
>
>  	tca->client = client;
> -	INIT_WORK(&tca->work, tca6507_work);
>  	spin_lock_init(&tca->lock);
> -	i2c_set_clientdata(client, tca);
> +
> +	err = devm_work_autocancel(dev, &tca->work, tca6507_work);
> +	if (err)
> +		return err;
>
>  	for (i = 0; i < NUM_LEDS; i++) {
>  		struct tca6507_led *l = tca->leds + i;
> @@ -766,20 +769,12 @@ static int tca6507_probe(struct i2c_client *client)
>  	return 0;
>  }
>
> -static void tca6507_remove(struct i2c_client *client)
> -{
> -	struct tca6507_chip *tca = i2c_get_clientdata(client);
> -
> -	cancel_work_sync(&tca->work);
> -}
> -
>  static struct i2c_driver tca6507_driver = {
>  	.driver   = {
>  		.name    = "leds-tca6507",
>  		.of_match_table = of_match_ptr(of_tca6507_leds_match),
>  	},
>  	.probe    = tca6507_probe,
> -	.remove   = tca6507_remove,
>  	.id_table = tca6507_id,
>  };
>
>
> base-commit: 05b4738b0078f7d6f154f68068a11c8a0635e9df
> -- 
> 2.56.0
>

-- 
Lee Jones

Re: [PATCH v2] leds: tca6507: Fix use-after-free on unbind

[Danish Khateeb]

On Wed, 28 Oct 2026, Lee Jones wrote:
> Please review, explain or fix the Sashiko review(s).

Sashiko is right that the "LED off" write queued on unbind gets
cancelled, but that predates 1b5c2fa7081c: remove() used to call
led_classdev_unregister() and then cancel_work_sync(), the same order
this patch restores. Since 1b5c2fa7081c that write only ran on freed
memory. If you'd like the LEDs turned off on unbind, I can send a
separate patch that flushes the work instead of cancelling it.

Thanks,
Danish

Re: [PATCH v2] leds: tca6507: Fix use-after-free on unbind

[Lee Jones]

On Wed, 28 Oct 2026, Danish Khateeb wrote:

> On Wed, 28 Oct 2026, Lee Jones wrote:
> > Please review, explain or fix the Sashiko review(s).
> 
> Sashiko is right that the "LED off" write queued on unbind gets
> cancelled, but that predates 1b5c2fa7081c: remove() used to call
> led_classdev_unregister() and then cancel_work_sync(), the same order
> this patch restores. Since 1b5c2fa7081c that write only ran on freed
> memory. If you'd like the LEDs turned off on unbind, I can send a
> separate patch that flushes the work instead of cancelling it.

Submit them as a set.

-- 
Lee Jones