[PATCH v2] drm/panthor: Fix TOCTOU race between GROUP_REGISTERED check and erase

[PATCH v2] drm/panthor: Fix TOCTOU race between GROUP_REGISTERED check and erase

[Osama Abdelkader]

panthor_group_destroy() checks the GROUP_REGISTERED mark and then calls
xa_erase() without holding the xarray lock across both. If the handle is
destroyed by another thread in between, and a concurrent GROUP_CREATE
reuses the freed ID, the first destroy erases and releases the new,
still-initializing group, leading to a use-after-free in
panthor_group_create().

Do the mark check and erase atomically under xa_lock().

Fixes: eec7e23d848d ("drm/panthor: Prevent potential UAF in group creation")
Assisted-by: Claude:claude-opus-5.5
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Liviu Dudau <liviu.dudau@arm.com>
Cc: stable@vger.kernel.org
Signed-off-by: Osama Abdelkader <osama.abdelkader@gmail.com>
---
v2:
- added Assisted-by tag, no code changes.
---
 drivers/gpu/drm/panthor/panthor_sched.c | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/panthor/panthor_sched.c b/drivers/gpu/drm/panthor/panthor_sched.c
index 0493858e55d7..a343c5180e01 100644
--- a/drivers/gpu/drm/panthor/panthor_sched.c
+++ b/drivers/gpu/drm/panthor/panthor_sched.c
@@ -3770,12 +3770,18 @@ int panthor_group_destroy(struct panthor_file *pfile, u32 group_handle)
 	struct panthor_group_pool *gpool = pfile->groups;
 	struct panthor_device *ptdev = pfile->ptdev;
 	struct panthor_scheduler *sched = ptdev->scheduler;
-	struct panthor_group *group;
+	struct panthor_group *group = NULL;
 
-	if (!xa_get_mark(&gpool->xa, group_handle, GROUP_REGISTERED))
-		return -EINVAL;
+	/*
+	 * Check the mark and erase the entry atomically, so a concurrent
+	 * destroy + create can't make us erase a group that's still being
+	 * initialized and happens to reuse the same handle.
+	 */
+	xa_lock(&gpool->xa);
+	if (xa_get_mark(&gpool->xa, group_handle, GROUP_REGISTERED))
+		group = __xa_erase(&gpool->xa, group_handle);
+	xa_unlock(&gpool->xa);
 
-	group = xa_erase(&gpool->xa, group_handle);
 	if (!group)
 		return -EINVAL;
 
-- 
2.53.0

Re: [PATCH v2] drm/panthor: Fix TOCTOU race between GROUP_REGISTERED check and erase

[Steven Price]

On 08/10/2026 11:09, Osama Abdelkader wrote:
> panthor_group_destroy() checks the GROUP_REGISTERED mark and then calls
> xa_erase() without holding the xarray lock across both. If the handle is
> destroyed by another thread in between, and a concurrent GROUP_CREATE
> reuses the freed ID, the first destroy erases and releases the new,
> still-initializing group, leading to a use-after-free in
> panthor_group_create().
> 
> Do the mark check and erase atomically under xa_lock().
> 
> Fixes: eec7e23d848d ("drm/panthor: Prevent potential UAF in group creation")
> Assisted-by: Claude:claude-opus-5.5
> Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
> Reviewed-by: Liviu Dudau <liviu.dudau@arm.com>
> Cc: stable@vger.kernel.org
> Signed-off-by: Osama Abdelkader <osama.abdelkader@gmail.com>

Reviewed-by: Steven Price <steven.price@arm.com>

> ---
> v2:
> - added Assisted-by tag, no code changes.
> ---
>  drivers/gpu/drm/panthor/panthor_sched.c | 14 ++++++++++----
>  1 file changed, 10 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/gpu/drm/panthor/panthor_sched.c b/drivers/gpu/drm/panthor/panthor_sched.c
> index 0493858e55d7..a343c5180e01 100644
> --- a/drivers/gpu/drm/panthor/panthor_sched.c
> +++ b/drivers/gpu/drm/panthor/panthor_sched.c
> @@ -3770,12 +3770,18 @@ int panthor_group_destroy(struct panthor_file *pfile, u32 group_handle)
>  	struct panthor_group_pool *gpool = pfile->groups;
>  	struct panthor_device *ptdev = pfile->ptdev;
>  	struct panthor_scheduler *sched = ptdev->scheduler;
> -	struct panthor_group *group;
> +	struct panthor_group *group = NULL;
>  
> -	if (!xa_get_mark(&gpool->xa, group_handle, GROUP_REGISTERED))
> -		return -EINVAL;
> +	/*
> +	 * Check the mark and erase the entry atomically, so a concurrent
> +	 * destroy + create can't make us erase a group that's still being
> +	 * initialized and happens to reuse the same handle.
> +	 */
> +	xa_lock(&gpool->xa);
> +	if (xa_get_mark(&gpool->xa, group_handle, GROUP_REGISTERED))
> +		group = __xa_erase(&gpool->xa, group_handle);
> +	xa_unlock(&gpool->xa);
>  
> -	group = xa_erase(&gpool->xa, group_handle);
>  	if (!group)
>  		return -EINVAL;
>