diff mbox series

[v4] media: v4l2-core: fix a use-after-free bug of sd->devnode

Message ID 20200219152554.25222-1-dafna.hirschfeld@collabora.com (mailing list archive)
State New, archived
Headers show
Series [v4] media: v4l2-core: fix a use-after-free bug of sd->devnode | expand

Commit Message

Dafna Hirschfeld Feb. 19, 2020, 3:25 p.m. UTC
sd->devnode is released after calling
v4l2_subdev_release. Therefore it should be set
to NULL so that the subdev won't hold a pointer
to a released object. This fixes a reference
after free bug in function
v4l2_device_unregister_subdev

Cc: stable@vger.kernel.org
Fixes: 0e43734d4c46e ("media: v4l2-subdev: add release() internal op")
Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@collabora.com>
---
changes since v2:
- since this is a regresion fix, I added Fixes and Cc to stable tags,
- change the commit title and log to be more clear.

changes since v3:
move the sd->devnode = NULL; line below the call to the release cb
so that it can still use it.

 drivers/media/v4l2-core/v4l2-device.c | 1 +
 1 file changed, 1 insertion(+)

Comments

Dafna Hirschfeld Feb. 26, 2020, 2:37 p.m. UTC | #1
Hi,
When a subdevice is registered with
v4l2_async_register_subdev_sensor_common and the main driver is
unloaded and reloaded, it does not register a node for the subdevice
since the devonde is not set to null.
in v4l2_device_register_subdev_nodes the sd is skipped if sd->devnode.

This patch solves this issue as well.

Dafna

On 19.02.20 16:25, Dafna Hirschfeld wrote:
> sd->devnode is released after calling
> v4l2_subdev_release. Therefore it should be set
> to NULL so that the subdev won't hold a pointer
> to a released object. This fixes a reference
> after free bug in function
> v4l2_device_unregister_subdev
> 
> Cc: stable@vger.kernel.org
> Fixes: 0e43734d4c46e ("media: v4l2-subdev: add release() internal op")
> Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
> Reviewed-by: Ezequiel Garcia <ezequiel@collabora.com>
> ---
> changes since v2:
> - since this is a regresion fix, I added Fixes and Cc to stable tags,
> - change the commit title and log to be more clear.
> 
> changes since v3:
> move the sd->devnode = NULL; line below the call to the release cb
> so that it can still use it.
> 
>   drivers/media/v4l2-core/v4l2-device.c | 1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/media/v4l2-core/v4l2-device.c b/drivers/media/v4l2-core/v4l2-device.c
> index 63d6b147b21e..41da73ce2e98 100644
> --- a/drivers/media/v4l2-core/v4l2-device.c
> +++ b/drivers/media/v4l2-core/v4l2-device.c
> @@ -179,6 +179,7 @@ static void v4l2_subdev_release(struct v4l2_subdev *sd)
>   
>   	if (sd->internal_ops && sd->internal_ops->release)
>   		sd->internal_ops->release(sd);
> +	sd->devnode = NULL;
>   	module_put(owner);
>   }
>   
>
diff mbox series

Patch

diff --git a/drivers/media/v4l2-core/v4l2-device.c b/drivers/media/v4l2-core/v4l2-device.c
index 63d6b147b21e..41da73ce2e98 100644
--- a/drivers/media/v4l2-core/v4l2-device.c
+++ b/drivers/media/v4l2-core/v4l2-device.c
@@ -179,6 +179,7 @@  static void v4l2_subdev_release(struct v4l2_subdev *sd)
 
 	if (sd->internal_ops && sd->internal_ops->release)
 		sd->internal_ops->release(sd);
+	sd->devnode = NULL;
 	module_put(owner);
 }