Hi Thomas,
Thank you Thomas for suggestions!
On 10/24/25 06:58, Thomas Zimmermann wrote:
Hi
Am 03.10.25 um 05:23 schrieb Chintan Patel:
When wait_event_timeout() in drm_wait_one_vblank() times out, the
current WARN can cause unnecessary kernel panics in environments
with panic_on_warn set (e.g. CI, fuzzing). These timeouts can happen
under heavy scheduling pressure or in rare cases of delayed vblank
handling, and are not always a kernel bug.
Replace the WARN with drm_err() messages that report the timeout
without crashing the system. Developers can still enable drm.debug
to diagnose genuine problems.
Reported-by: [email protected]
Closes: https://syzkaller.appspot.com/bug?extid=147ba789658184f0ce04
Tested-by: [email protected]
Signed-off-by: Chintan Patel <[email protected]>
v2:
- Drop unnecessary in-code comment (suggested by Thomas Zimmermann)
- Remove else branch, only log timeout case
v3:
- Use drm_err() instead of drm_dbg_kms() (suggested by Ville Syrjälä)
- Remove unnecessary curr = drm_vblank_count() (suggested by Thomas
Zimmermann)
- Fix commit message wording (“invalid userspace calls” → “delayed
vblank handling”)
---
drivers/gpu/drm/drm_vblank.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/drm_vblank.c b/drivers/gpu/drm/drm_vblank.c
index 46f59883183d..0664aea1b924 100644
--- a/drivers/gpu/drm/drm_vblank.c
+++ b/drivers/gpu/drm/drm_vblank.c
@@ -1305,7 +1305,8 @@ void drm_wait_one_vblank(struct drm_device *dev,
unsigned int pipe)
last != drm_vblank_count(dev, pipe),
msecs_to_jiffies(100));
Instead of replacing the drm_WARN(), could you please try to increase
the timeout? Let's say 1000 msec to be on the safe side.
I tried it locally and also tested with syzbot after increasing the
timeout to 1000 msec. The issue no longer reproduces with this change.
I’ll send v4 shortly with the updated timeout.
