Created attachment 138153 [details]
dump of gpu error state
Just seen this GPU-hanger once, will report back if it occurs again. The application was in the system tray and when I tried to open the window, the system seemed frozen for several seconds.
$ uname -a
Linux ch10dd405 4.9.76-realtime-1-rt61 #1 SMP PREEMPT RT Wed Jan 17 14:36:24 CET 2018 x86_64 x86_64 x86_64 GNU/Linux
$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04.5 LTS
$ lspci -v | grep -i vga
00:02.0 VGA compatible controller: Intel Corporation Device 5912 (rev 04) (prog-if 00 [VGA controller])
$ cat /proc/cpuinfo | grep "model name" -m 1
model name : Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz
[145347.575536] [drm] GPU HANG: ecode 9:0:0x00775757, in Worklog Assista , reason: Hang on render ring, action: reset
[145347.575540] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[145347.575540] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[145347.575540] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[145347.575541] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
[145347.575543] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[145347.575581] drm/i915: Resetting chip after gpu hang
[145347.575621] [drm] RC6 on
[145347.593831] [drm] GuC firmware load skipped
[145359.543380] drm/i915: Resetting chip after gpu hang
[145359.543422] [drm] RC6 on
[145359.561485] [drm] GuC firmware load skipped
Note it doesn't look like a very mesa-ish batch buffer. At the very least this needs confirming with an uptodate graphics driver stack.