Comment 51 for bug 1301125

I had the issue that Emmanuel describes, with a fresh install of 16.04.1 LTS 64-bit on a self-built machine with Athlon 64 X2 CPU and IGP ATI RS690 (X1200) and the open source radeon driver and kernels from 4.4.0-31 to -38.

As per Jonas's comment #50, cannot switch TTY, but as I've enabled REISUB I can Alt-SysRq-R then E before I can regain control, but then systemd keeps restarting the desktop. A more stable workaround was to set up ssh and ssh in from another box and restart lightdm. Setting the idle time for the screen saver to Never was only partially satisfactory as manually locking the screen would also trigger the screen fade-out and thus the bug. Interestingly, manually blanking the screen using "xset s activate" would blank the screen and not trigger the bug, so it doesn't appear to be a DPMS problem.

While sshing in from another box, at the point of triggering the bug there was no output to Xorg.0.log, nor syslog, nor kern.log. However, after some attempted keyboard & mouse activity on the affected system, Xorg.0.log did show an MIEQ overflow backtrace:

(EE) [mi] EQ overflowing. Additional events will be discarded until existing events are processed.
(EE)
(EE) Backtrace:
(EE) 0: /usr/lib/xorg/Xorg (xorg_backtrace+0x4e) [0x558ad06155ce]
(EE) 1: /usr/lib/xorg/Xorg (mieqEnqueue+0x253) [0x558ad05f7083]
(EE) 2: /usr/lib/xorg/Xorg (QueuePointerEvents+0x52) [0x558ad04cf662]
(EE) 3: /usr/lib/xorg/modules/input/evdev_drv.so (0x7f1b3560f000+0x61f3) [0x7f1b356151f3]
(EE) 4: /usr/lib/xorg/modules/input/evdev_drv.so (0x7f1b3560f000+0x6a5d) [0x7f1b35615a5d]
(EE) 5: /usr/lib/xorg/Xorg (0x558ad0463000+0x94228) [0x558ad04f7228]
(EE) 6: /usr/lib/xorg/Xorg (0x558ad0463000+0xb96f2) [0x558ad051c6f2]
(EE) 7: /lib/x86_64-linux-gnu/libc.so.6 (0x7f1b3f2e6000+0x354a0) [0x7f1b3f31b4a0]
(EE) 8: /lib/x86_64-linux-gnu/libc.so.6 (__read+0x2b) [0x7f1b3f3dc9cb]
(EE) 9: /usr/lib/x86_64-linux-gnu/libdrm.so.2 (drmHandleEvent+0x37) [0x7f1b406cdde7]
(EE) 10: /usr/lib/xorg/modules/drivers/radeon_drv.so (0x7f1b3b3d4000+0x4f9ec) [0x7f1b3b4239ec]
(EE) 11: /usr/lib/xorg/modules/drivers/radeon_drv.so (0x7f1b3b3d4000+0x498e6) [0x7f1b3b41d8e6]
(EE) 12: /usr/lib/xorg/modules/drivers/radeon_drv.so (0x7f1b3b3d4000+0x499af) [0x7f1b3b41d9af]
(EE) 13: /usr/lib/xorg/Xorg (dixSaveScreens+0x29e) [0x558ad04e84de]
(EE) 14: /usr/lib/xorg/Xorg (mieqProcessInputEvents+0x1e6) [0x558ad05f7706]
(EE) 15: /usr/lib/xorg/Xorg (ProcessInputEvents+0x19) [0x558ad04f7479]
(EE) 16: /usr/lib/xorg/Xorg (0x558ad0463000+0x539df) [0x558ad04b69df]
(EE) 17: /usr/lib/xorg/Xorg (0x558ad0463000+0x57c33) [0x558ad04bac33]
(EE) 18: /lib/x86_64-linux-gnu/libc.so.6 (__libc_start_main+0xf0) [0x7f1b3f306830]
(EE) 19: /usr/lib/xorg/Xorg (_start+0x29) [0x558ad04a4f59]
(EE)
(EE) [mi] These backtraces from mieqEnqueue may point to a culprit higher up the stack.
(EE) [mi] mieq is *NOT* the cause. It is a victim.

As this suggested a problem with the graphics stack, I tried the latest drivers from the Oibaf PPA, and so far I haven't been able to reproduce the bug. I don't know which component/version is responsible for the apparent fix.