Random crash over nvidia hardware

[DT3264]

Given a journalctl on a Thinkstation P320 with an Nvidia Quadro P400, a random crash left this in the journal before puting the machine on halt:

abr 22 09:12:07 huronOS kernel: INFO: task kworker/u8:3:21879 blocked for more than 966 seconds.
abr 22 09:12:07 huronOS kernel:       Tainted: G            E      6.0.15-huronos+ #2
abr 22 09:12:07 huronOS kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
abr 22 09:12:07 huronOS kernel: task:kworker/u8:3    state:D stack:    0 pid:21879 ppid:     2 flags:0x00004000
abr 22 09:12:07 huronOS kernel: Workqueue: events_unbound nv50_disp_atomic_commit_work [nouveau]
abr 22 09:12:07 huronOS kernel: Call Trace:
abr 22 09:12:07 huronOS kernel:  <TASK>
abr 22 09:12:07 huronOS kernel:  __schedule+0x2ef/0x9c0
abr 22 09:12:07 huronOS kernel:  ? _raw_spin_unlock_irqrestore+0x23/0x40
abr 22 09:12:07 huronOS kernel:  schedule+0x5d/0xd0
abr 22 09:12:07 huronOS kernel:  schedule_timeout+0x11d/0x160
abr 22 09:12:07 huronOS kernel:  ? nouveau_fence_enable_signaling+0x2a/0x70 [nouveau]
abr 22 09:12:07 huronOS kernel:  dma_fence_default_wait+0x1e7/0x230
abr 22 09:12:07 huronOS kernel:  ? dma_fence_free+0x20/0x20
abr 22 09:12:07 huronOS kernel:  dma_fence_wait_timeout+0xe5/0x110
abr 22 09:12:07 huronOS kernel:  drm_atomic_helper_wait_for_fences+0x7f/0xf0 [drm_kms_helper]
abr 22 09:12:07 huronOS kernel:  nv50_disp_atomic_commit_tail+0x90/0x870 [nouveau]
abr 22 09:12:07 huronOS kernel:  ? _raw_spin_unlock+0x15/0x30
abr 22 09:12:07 huronOS kernel:  ? __switch_to+0x112/0x420
abr 22 09:12:07 huronOS kernel:  process_one_work+0x1e5/0x3b0
abr 22 09:12:07 huronOS kernel:  worker_thread+0x50/0x3a0
abr 22 09:12:07 huronOS kernel:  ? rescuer_thread+0x390/0x390
abr 22 09:12:07 huronOS kernel:  kthread+0xe7/0x110
abr 22 09:12:07 huronOS kernel:  ? kthread_complete_and_exit+0x20/0x20
abr 22 09:12:07 huronOS kernel:  ret_from_fork+0x22/0x30
abr 22 09:12:07 huronOS kernel:  </TASK>

And as @equetzal mentioned given this and this, it seems to be an nouveau issue.

Possible solutions at the moment include

• erase nouveau-dri.so
• use xf86-video-nouveau
• Changing the nouveau drivers to the nvidia propietary ones
• Updating the culprit (if found)
• Disabling nvidia hardware if found (discarded bc some pc's have the display output directly through the discrete graphics card)

Although before a solution, it would be awesome to have a way to reproduce this issue and identify the affected systems and test either the causes / possible solutions.

[DT3264]

After some testing with the nvidia drivers (as seen on #135 ) on both the latest (build 27) and an experimental build (using debian 12), I am quite sure that the issue has been resolved, although I think this issue would be closed once #135 has been merged.