RTX 5090 (GB202): Flip event timeout and sync FD semaphore failure under GPU load — 590.48.01

Graphics / Linux Linux
1

I’m running into an intermittent crash happening mostly under heavy load, but occasionally not. journalctl -b 1 -p err shows

...
Feb 26 09:34:03 cygnus kernel: [drm:__nv_drm_semsurf_wait_fence_work_cb [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to register auto-value-update on pre-wait value for sync FD semaphore surface
...
Feb 26 09:34:13 cygnus kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
...

Here’s the complete log, along with the usual bug report:

err.log (137.4 KB)

nvidia-bug-report.log.gz (889.3 KB)

Here’s my system info:

  • GPU: NVIDIA GB202 [GeForce RTX 5090]
  • Driver: nvidia-open-dkms 590.48.01-4
  • Kernel: Linux 6.12.74-1-lts x86_64
  • Compositor: Hyprland

Can anyone tell me if this is a hardware issue, or a problem with drivers?

Nvidia-drm: Stub EDID cached on wake for slow-waking DisplayPort monitor — semsurf error and pageflip timeout on head 1 (RTX 5080, 590.48.01)
2

I have same issue with the same this messsage, if you using Hyprland with hybrid GPU this might be fixing it. To fix just setting environment AQ_DRM_DEVICES to use iGPU, then dGPU on hyprland.conf or any your environment set or just delete the AQ_DRM_DEVICES so hyprland will pick automatically

3

Hi All,

I have filed a bug 5981743 internally for tracking purpose.

Team will try to duplicate issue in house and will get back if required any additional information.