[PATCH v4] drm/v3d: Reject empty multisync extension to prevent infinite loop

Ashutosh Desai posted 1 patch 2 months ago
drivers/gpu/drm/v3d/v3d_submit.c | 5 +++++
1 file changed, 5 insertions(+)
[PATCH v4] drm/v3d: Reject empty multisync extension to prevent infinite loop
Posted by Ashutosh Desai 2 months ago
v3d_get_extensions() walks a userspace-provided singly-linked list of
ioctl extensions without any bound on the chain length. A local user
can craft a self-referential extension (ext->next == &ext) with zero
in_sync_count and out_sync_count, which bypasses the existing duplicate-
extension guard:

    if (se->in_sync_count || se->out_sync_count)
            return -EINVAL;

The guard never fires because v3d_get_multisync_post_deps() returns
immediately when count is zero, leaving both fields at zero on every
iteration. The result is an infinite loop in kernel context, blocking
the calling thread and pegging a CPU core indefinitely.

Fix this by rejecting a multisync extension where both in_sync_count
and out_sync_count are zero in v3d_get_multisync_submit_deps(). An
empty multisync carries no synchronization information and serves no
useful purpose, so returning -EINVAL for such an extension is the
correct defense against this attack vector.

Fixes: 9032d5f633ed ("drm/v3d: Detach job submissions IOCTLs to a new specific file")
Cc: stable@vger.kernel.org
Signed-off-by: Ashutosh Desai <ashutoshdesai993@gmail.com>
---
V3 -> V4: fix indentation
V2 -> V3: drop depth counter; instead reject empty multisync
          (in_sync_count == 0 && out_sync_count == 0) in
          v3d_get_multisync_submit_deps()
V1 -> V2: change cap from 16 to V3D_MAX_EXTENSIONS (7), add #define

v3: https://lore.kernel.org/dri-devel/177614548527.3603641.5360701002746181082@gmail.com/
v2: https://lore.kernel.org/dri-devel/20260413055230.3349114-1-ashutoshdesai993@gmail.com/
v1: https://lore.kernel.org/dri-devel/20260410013907.2404175-1-ashutoshdesai993@gmail.com/

 drivers/gpu/drm/v3d/v3d_submit.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c
index 18f2bf1fe89f..fc74351efad5 100644
--- a/drivers/gpu/drm/v3d/v3d_submit.c
+++ b/drivers/gpu/drm/v3d/v3d_submit.c
@@ -393,6 +393,11 @@ v3d_get_multisync_submit_deps(struct drm_file *file_priv,
 	if (multisync.pad)
 		return -EINVAL;
 
+	if (!multisync.in_sync_count && !multisync.out_sync_count) {
+		drm_dbg(&v3d->drm, "Empty multisync extension\n");
+		return -EINVAL;
+	}
+
 	ret = v3d_get_multisync_post_deps(file_priv, se, multisync.out_sync_count,
 					  multisync.out_syncs);
 	if (ret)
-- 
2.34.1
Re: [PATCH v4] drm/v3d: Reject empty multisync extension to prevent infinite loop
Posted by Maíra Canal 1 month, 3 weeks ago
Hi Ashutosh,

On 15/04/26 02:00, Ashutosh Desai wrote:
> v3d_get_extensions() walks a userspace-provided singly-linked list of
> ioctl extensions without any bound on the chain length. A local user
> can craft a self-referential extension (ext->next == &ext) with zero
> in_sync_count and out_sync_count, which bypasses the existing duplicate-
> extension guard:
> 
>      if (se->in_sync_count || se->out_sync_count)
>              return -EINVAL;
> 
> The guard never fires because v3d_get_multisync_post_deps() returns
> immediately when count is zero, leaving both fields at zero on every
> iteration. The result is an infinite loop in kernel context, blocking
> the calling thread and pegging a CPU core indefinitely.
> 
> Fix this by rejecting a multisync extension where both in_sync_count
> and out_sync_count are zero in v3d_get_multisync_submit_deps(). An
> empty multisync carries no synchronization information and serves no
> useful purpose, so returning -EINVAL for such an extension is the
> correct defense against this attack vector.
> 
> Fixes: 9032d5f633ed ("drm/v3d: Detach job submissions IOCTLs to a new specific file")
> Cc: stable@vger.kernel.org
> Signed-off-by: Ashutosh Desai <ashutoshdesai993@gmail.com>

Applied to misc/kernel.git (drm-misc-fixes) with a fix: I changed the
"Fixes" tag to reflect the commit in which the issue was indeed
introduced, that is: commit e4165ae8 ("drm/v3d: add multiple syncobjs
support").

Thanks for the contribution!

Best Regards,
- Maíra

> ---
> V3 -> V4: fix indentation
> V2 -> V3: drop depth counter; instead reject empty multisync
>            (in_sync_count == 0 && out_sync_count == 0) in
>            v3d_get_multisync_submit_deps()
> V1 -> V2: change cap from 16 to V3D_MAX_EXTENSIONS (7), add #define
> 
> v3: https://lore.kernel.org/dri-devel/177614548527.3603641.5360701002746181082@gmail.com/
> v2: https://lore.kernel.org/dri-devel/20260413055230.3349114-1-ashutoshdesai993@gmail.com/
> v1: https://lore.kernel.org/dri-devel/20260410013907.2404175-1-ashutoshdesai993@gmail.com/
> 
>   drivers/gpu/drm/v3d/v3d_submit.c | 5 +++++
>   1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/gpu/drm/v3d/v3d_submit.c b/drivers/gpu/drm/v3d/v3d_submit.c
> index 18f2bf1fe89f..fc74351efad5 100644
> --- a/drivers/gpu/drm/v3d/v3d_submit.c
> +++ b/drivers/gpu/drm/v3d/v3d_submit.c
> @@ -393,6 +393,11 @@ v3d_get_multisync_submit_deps(struct drm_file *file_priv,
>   	if (multisync.pad)
>   		return -EINVAL;
>   
> +	if (!multisync.in_sync_count && !multisync.out_sync_count) {
> +		drm_dbg(&v3d->drm, "Empty multisync extension\n");
> +		return -EINVAL;
> +	}
> +
>   	ret = v3d_get_multisync_post_deps(file_priv, se, multisync.out_sync_count,
>   					  multisync.out_syncs);
>   	if (ret)