Comment # 6 on bug 104345 from bernhardu

Created attachment 136709 [details]
kernel output captured from the serial line. (4.14.10+lockdep)

Attaching another example with a vanilla kernel 4.14.10 with lockdep enabled.

Shows again such "GPU fault detected:"
followed by a "task amdgpu_cs:0:1014 blocked for more than 120 seconds.

I might add that it may be related to one or more suspend and resume cycles.

Is there anything I can attach or test to make
identification of the issue easier?