Bug 62311

Summary: Screen freeze/hard lockup after DisplayPort monitor was connected
Product: Drivers Reporter: Milan Plzik (milan.plzik)
Component: Video(DRI - Intel)Assignee: intel-gfx-bugs (intel-gfx-bugs)
Status: RESOLVED CODE_FIX    
Severity: normal CC: chris, daniel, intel-gfx-bugs
Priority: P1    
Hardware: x86-64   
OS: Linux   
Kernel Version: 3.12.1 Subsystem:
Regression: No Bisected commit-id:
Attachments: lspci -v of the system
i915_error_state after stuck ring error

Description Milan Plzik 2013-09-28 23:28:21 UTC
Created attachment 109891 [details]
lspci -v of the system

After connecting and enabling display using DisplayPort, using chrome to play video/google maps/... on external screen usually causes few-second screen freeze followed by these lines in dmesg:

[163015.696253] Watchdog[2074]: segfault at 0 ip 00007f05e2a53638 sp 00007f05d057a640 error 6 in chromium[7f05e1d3c000+4c8c000]
[163017.244564] [drm:ring_stuck] *ERROR* Kicking stuck wait on render ring
[163022.245406] [drm:ring_stuck] *ERROR* Kicking stuck wait on render ring

Eventually, this might also result in hard lockup of the system (cursor does not move, magic sysrq does not work, machine does not reply to ping, nothing recorded in logs). System is Lenovo Thinkpad X220 tablet (attached lspci.txt). Kernel is 3.11.1, arch x86_64 .
Comment 1 Daniel Vetter 2013-09-29 12:48:07 UTC
You probably need to upgrade your xfree86-video-intel driver, but we need the error state to tell for sure. It's in /sys/class/drm/card0/error. You need to capture it _after_ you've reproduced the problem.
Comment 2 Chris Wilson 2013-09-29 12:50:37 UTC
The machine hang is a separate bug to the kicking the stuck wait.
Comment 3 Milan Plzik 2013-10-06 10:12:37 UTC
Unfortunately, there was no such file in /sys/class/drm/card0 . Attaching i915_error_state from debugfs, I hope it helps at least a little.
Comment 4 Milan Plzik 2013-10-06 10:13:37 UTC
Created attachment 110311 [details]
i915_error_state after stuck ring error
Comment 5 Milan Plzik 2014-01-11 12:24:05 UTC
Kernel 3.12.1 is also affected
Comment 6 Daniel Vetter 2014-01-14 10:42:16 UTC
The new error state doesn't indicate a stuck wait on render ring, so that bug is resolved (probably the ddx bug).

Now you seem to hit the mesa gen6 pipe control woes. Please retest with latest mesa (I think the latest point release should be good enough, but maybe also try git master). If the issue still occurs please file a new bug against Mesa -> drivers/i965 on bugs.freedesktop.org.