Created attachment 135174 [details]
gpu crash dump

this seems to happen occasionally when emacs runs into a "deviant" (of some sort) character (i use various non-US alphabets).  in this case, it was a "_a_", (0x5f, 0x61, 0x5f) which emacs would like to render as underlined.

this could be an duplicate of 93166.

here are the syslog messages (of two back-to-back occurrences):
----
Oct 31 07:27:47 mymachine kernel: [ 2017.901988] [drm] stuck on render ring
Oct 31 07:27:47 mymachine kernel: [ 2017.902212] [drm] GPU HANG: ecode 9:0:0x84dffff8, in Xorg [1012], reason: Engine(s) hung, action: reset
Oct 31 07:27:47 mymachine kernel: [ 2017.902216] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
Oct 31 07:27:47 mymachine kernel: [ 2017.902218] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
Oct 31 07:27:47 mymachine kernel: [ 2017.902220] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
Oct 31 07:27:47 mymachine kernel: [ 2017.902222] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
Oct 31 07:27:47 mymachine kernel: [ 2017.902225] [drm] GPU crash dump saved to /sys/class/drm/card0/error
Oct 31 07:27:47 mymachine kernel: [ 2017.904109] drm/i915: Resetting chip after gpu hang
Oct 31 07:27:48 mymachine kernel: [ 2018.910462] [drm] RC6 on
Oct 31 07:28:01 mymachine kernel: [ 2031.914708] [drm] stuck on render ring
Oct 31 07:28:01 mymachine kernel: [ 2031.914967] [drm] GPU HANG: ecode 9:0:0x84dffff8, in Xorg [1012], reason: Engine(s) hung, action: reset
Oct 31 07:28:01 mymachine kernel: [ 2031.916837] drm/i915: Resetting chip after gpu hang
Oct 31 07:28:02 mymachine kernel: [ 2032.922945] [drm] RC6 on
----