On Wed, 14 Apr 2021 at 11:48, Christian König <ckoenig.leichtzumerken@gmail.com> wrote:

That is expected behavior, the application is just buggy and causing a
page fault on the GPU.

The kernel should just not crash with a backtrace.

Regards,
Christian.

If after it GPU hangs with the message "[drm:amdgpu_dm_atomic_commit_tail [amdgpu]] *ERROR* Waiting for fences timed out!" is it also expected behavior?

Kernel log: https://pastebin.com/WkhATKXX

--

Best Regards,
Mike Gavrilov.