Bug 82096 - [uxa ilkl] Blank screen after suspend, GPU hang and gnome-shell crash
Summary: [uxa ilkl] Blank screen after suspend, GPU hang and gnome-shell crash
Status: RESOLVED INVALID
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/intel (show other bugs)
Version: 7.7 (2012.06)
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Chris Wilson
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-08-03 14:56 UTC by Milan Bouchet-Valat
Modified: 2015-01-31 20:56 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
dmesg (98.38 KB, text/plain)
2014-08-03 14:56 UTC, Milan Bouchet-Valat
no flags Details
Xorg.0.log (58.34 KB, text/plain)
2014-08-03 14:57 UTC, Milan Bouchet-Valat
no flags Details
0/i915_error_state (25 bytes, text/plain)
2014-08-03 14:58 UTC, Milan Bouchet-Valat
no flags Details
64/i915_error_state (25 bytes, text/plain)
2014-08-03 14:58 UTC, Milan Bouchet-Valat
no flags Details
lspci (2.79 KB, text/plain)
2014-08-03 14:58 UTC, Milan Bouchet-Valat
no flags Details
0/opregion (8.00 KB, text/plain)
2014-08-03 14:59 UTC, Milan Bouchet-Valat
no flags Details
64/opregion (8.00 KB, text/plain)
2014-08-03 14:59 UTC, Milan Bouchet-Valat
no flags Details
rom (64.00 KB, text/plain)
2014-08-03 14:59 UTC, Milan Bouchet-Valat
no flags Details
snapshot (2.08 MB, text/plain)
2014-08-03 15:03 UTC, Milan Bouchet-Valat
no flags Details

Description Milan Bouchet-Valat 2014-08-03 14:56:58 UTC
Created attachment 103913 [details]
dmesg

About every 3 suspend/resume cycles, my screen remains blank; I can still SSH into the machine. This is with an HP dv6 with an Intel Ironlake GPU (8086:0046) on Fedora 20. When this bug does not happen on resume, the desktop works fine except that the mouse cursor is not visible and that gnome-shell crashes (before restarting fine).

Everything used to work fine when only using the VESA driver. Installing the Intel driver made the problem appear.

This is a dual-GPU machine, the other card being an ATI Radeon, which is apparently turned off according to vgaswitcheroo:
0:IGD:+:Pwr:0000:00:02.0
1:DIS: :DynOff:0000:01:00.0
2:DIS-Audio: :Off:0000:01:00.1

A weird thing that I noticed is that the GNOME Screens panel shows two integrated displays, the second one of which is not visible.

Please ask if you need any additional information.

xorg-x11-server-Xorg-1.14.4-11
xorg-x11-drv-intel-2.21.15-7
Comment 1 Milan Bouchet-Valat 2014-08-03 14:57:30 UTC
Created attachment 103914 [details]
Xorg.0.log
Comment 2 Milan Bouchet-Valat 2014-08-03 14:58:09 UTC
Created attachment 103916 [details]
0/i915_error_state
Comment 3 Milan Bouchet-Valat 2014-08-03 14:58:27 UTC
Created attachment 103917 [details]
64/i915_error_state
Comment 4 Milan Bouchet-Valat 2014-08-03 14:58:46 UTC
Created attachment 103918 [details]
lspci
Comment 5 Milan Bouchet-Valat 2014-08-03 14:59:02 UTC
Created attachment 103919 [details]
0/opregion
Comment 6 Milan Bouchet-Valat 2014-08-03 14:59:21 UTC
Created attachment 103920 [details]
64/opregion
Comment 7 Milan Bouchet-Valat 2014-08-03 14:59:42 UTC
Created attachment 103921 [details]
rom
Comment 8 Milan Bouchet-Valat 2014-08-03 15:03:15 UTC
Created attachment 103922 [details]
snapshot
Comment 9 Milan Bouchet-Valat 2014-08-03 15:03:22 UTC
Oh, and when I call poweroff, the Plymouth boot screen appears just fine.
Comment 10 Milan Bouchet-Valat 2014-08-23 12:31:15 UTC
Bump!
Comment 11 Chris Wilson 2014-09-08 07:13:59 UTC
The actual crash may have been more informative - all that is shown here is some benign oddity in the dual-gpu config.
Comment 12 Milan Bouchet-Valat 2014-09-08 08:43:32 UTC
What do you mean by "the actual crash"? I took all the information I could after the screen turned black. But maybe there's no crash at all, just a bug with backlight?
Comment 13 Chris Wilson 2014-09-08 08:53:41 UTC
"When this bug does not happen on resume, the desktop works fine except that the mouse cursor is not visible and that gnome-shell crashes"

If that is a driver related crash, it may be more meaningful than that the kernel or UXA loses a vblank across resume.
Comment 14 Milan Bouchet-Valat 2014-09-08 12:32:52 UTC
OK, I'll try to debug the gnome-shell crash, but for now I don't have access to that machine anymore.
Comment 15 Chris Wilson 2015-01-31 10:32:25 UTC
Any chance you can continue investigating with this machine?
Comment 16 Milan Bouchet-Valat 2015-01-31 13:48:36 UTC
Not at the moment. I'll reopen when/if I can get more information about the crash.
Comment 17 Chris Wilson 2015-01-31 20:56:43 UTC
Thanks for the bug report. I am sorry we never made any progress on resolving the issue, so please do reopen if you ever get the chance!


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.