Created attachment 142867 [details]
2018-12-20-crash.txt

The last months I observed this less.
But in December it started to get more visible again.
(I am following debian testing closely.)

It shows now nearly always when e.g. browsing with firefox -
even in simple sites e.g. bugtrackers.

The chess board like pattern is still the same.

The blocked kernel threads stack changed with some kernel upgrade to:

Call Trace:
 ? __schedule+0x2b7/0x880
 ? __switch_to_asm+0x40/0x70
 ? __switch_to_asm+0x34/0x70
 schedule+0x28/0x80
 schedule_timeout+0x1ee/0x380
 ? dce110_timing_generator_get_position+0x5b/0x70 [amdgpu]
 ? dce110_timing_generator_get_crtc_scanoutpos+0x70/0xb0 [amdgpu]
 dma_fence_default_wait+0x1fd/0x280
 ? dma_fence_release+0x90/0x90
 dma_fence_wait_timeout+0x39/0xf0
 reservation_object_wait_timeout_rcu+0x10a/0x290
 amdgpu_dm_do_flip+0x112/0x340 [amdgpu]
 amdgpu_dm_atomic_commit_tail+0xbac/0xdb0 [amdgpu]
 ? __switch_to_asm+0x34/0x70
 ? __switch_to_asm+0x40/0x70
 ? __switch_to+0x16f/0x450
 commit_tail+0x3d/0x70 [drm_kms_helper]
 process_one_work+0x195/0x380
 worker_thread+0x30/0x390
 ? process_one_work+0x380/0x380
 kthread+0x113/0x130
 ? kthread_create_worker_on_cpu+0x70/0x70
 ret_from_fork+0x22/0x40

What can I do to get this bug forward?