Bug 99814 - GPU Hang after Resume from Hibernation (ecode 8:0:0x00200000)
Summary: GPU Hang after Resume from Hibernation (ecode 8:0:0x00200000)
Status: CLOSED DUPLICATE of bug 96526
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: XOrg git
Hardware: Other All
: medium normal
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-02-14 16:38 UTC by Kassick
Modified: 2017-07-24 22:39 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
dmesg (under xorg) (76.49 KB, text/plain)
2017-02-14 16:39 UTC, Kassick
no flags Details
/sys/class/drm/card0/error (under xorg) (565.86 KB, text/plain)
2017-02-14 16:39 UTC, Kassick
no flags Details
journalctl --system (under xorg) (262.24 KB, text/plain)
2017-02-14 16:40 UTC, Kassick
no flags Details
journalctl --user (under xorg) (89.77 KB, text/plain)
2017-02-14 16:40 UTC, Kassick
no flags Details
dmesg (under wayland) (77.71 KB, text/plain)
2017-02-14 16:40 UTC, Kassick
no flags Details
/sys/class/drm/card0/error (under wayland) (713.50 KB, text/plain)
2017-02-14 16:41 UTC, Kassick
no flags Details
journalctl --system (under wayland) (17.94 MB, text/plain)
2017-02-14 16:45 UTC, Kassick
no flags Details
journalctl --user (under wayland) (8.94 MB, text/plain)
2017-02-14 16:50 UTC, Kassick
no flags Details

Description Kassick 2017-02-14 16:38:30 UTC
After resuming from hibernation (aka suspend to disk), the GPU always hangs and get's reset.

When using Xorg, the process caught in the hang is always systemd-logind . Under Xorg, after the GPU is reset, gnome-shell dies and get's restarted. Xorg itself does not crash, so no graphical app crashes and the session continues normally. I have experienced no artifacts afterwards, the system works normally and hibernate again (to the same results) or suspend to ram (without incident).

When using Wayland, the process caught in the hang is always gnome-shell itself. After resume, it crashes with the message "intel_do_flush_locked failed: Input/output error". As shell dies, applications love their connection to the wayland compositor and the session is eventually finished. Logging again from GDM works fine and the system can hibernate or suspend, just as under Xorg.

Suspend-to-ram works without any incident. The gang occurs *only* after resuming from disk.
Comment 1 Kassick 2017-02-14 16:39:00 UTC
Created attachment 129603 [details]
dmesg (under xorg)
Comment 2 Kassick 2017-02-14 16:39:46 UTC
Created attachment 129604 [details]
/sys/class/drm/card0/error (under xorg)
Comment 3 Kassick 2017-02-14 16:40:12 UTC
Created attachment 129605 [details]
journalctl --system (under xorg)
Comment 4 Kassick 2017-02-14 16:40:32 UTC
Created attachment 129606 [details]
journalctl --user (under xorg)
Comment 5 Kassick 2017-02-14 16:40:49 UTC
Created attachment 129607 [details]
dmesg (under wayland)
Comment 6 Kassick 2017-02-14 16:41:19 UTC
Created attachment 129608 [details]
/sys/class/drm/card0/error (under wayland)
Comment 7 Kassick 2017-02-14 16:45:59 UTC
Created attachment 129609 [details]
journalctl --system (under wayland)
Comment 8 Kassick 2017-02-14 16:48:28 UTC
Fedora 25 x86_64
Kernel 4.9.8-201.fc25.x86_64 
Gnome-Shell gnome-shell-3.22.2-2
Mesa mesa-dri-drivers-13.0.3-5
Comment 9 Kassick 2017-02-14 16:50:21 UTC
Created attachment 129610 [details]
journalctl --user (under wayland)
Comment 10 Chris Wilson 2017-02-14 16:54:26 UTC

*** This bug has been marked as a duplicate of bug 96526 ***


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.