Created attachment 121306 [details]
[Sij26 01:40] [drm] GPU HANG: ecode 6:-1:0x00000000, reason: Kicking stuck semap
hore on render ring, action: continue
[ +0,000004] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stac
k, including userspace.
[ +0,000001] [drm] Please file a _new_ bug report on bugs.freedesktop.org again
st DRI -> DRM/Intel
[ +0,000000] [drm] drm/i915 developers can then reassign to the right component
if it's not a kernel issue.
[ +0,000001] [drm] The gpu crash dump is required to analyze gpu hangs, so plea
se always attach it.
[ +0,000001] [drm] GPU crash dump saved to /sys/class/drm/card0/error
$ uname -r
Seems to get better with every kernel update and got to the point that dmesg is not filled with messages for couple of days by running with a new kernel.
The biggest recent change would be read-read not requiring the semaphore. Other than that there still haven't been any major breakthroughs on how to make snb behave.
*** This bug has been marked as a duplicate of bug 54226 ***