Bug 94358 - Display sleep mode causes session crash
Summary: Display sleep mode causes session crash
Status: RESOLVED INVALID
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/intel (show other bugs)
Version: unspecified
Hardware: Other All
: medium normal
Assignee: Chris Wilson
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-03-01 23:51 UTC by Andreas Kloeckner
Modified: 2016-03-03 07:00 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
Crash log (19.69 KB, text/plain)
2016-03-01 23:51 UTC, Andreas Kloeckner
no flags Details
Debian package info (148.68 KB, text/plain)
2016-03-01 23:51 UTC, Andreas Kloeckner
no flags Details

Description Andreas Kloeckner 2016-03-01 23:51:06 UTC
Created attachment 122068 [details]
Crash log

It's a little tricky to pinpoint what's causing this, so I'm starting here.

I recently got a shiny new 4k monitor (Dell P2715Q), which I'm running off the (mini) DisplayPort of my ThinkPad X250. Now it seems that whenever display power saving kicks in, my X11 session dies, with the attached log from "journalctl". As you can probably imagine, this is highly irritating. Any help in tracking this down would be much appreciated. For what it's worth, this does not seem to happen with lower-resolution monitors (various Dell 2560x1440 monitors for example).
Comment 1 Andreas Kloeckner 2016-03-01 23:51:29 UTC
Created attachment 122069 [details]
Debian package info
Comment 2 Chris Wilson 2016-03-02 07:54:27 UTC
logind vs wayland. It's not clear who started the trouble, but logind is certainly part of the problem. Anyway not the ddx or Xserver or kernel and I have no idea where to redirect the bug.
Comment 3 Andreas Kloeckner 2016-03-02 18:20:43 UTC
I'm not really using Wayland though. I know that the X server said something about Wayland events, but I'm not sure where that came from.

Occasionally (but not every time) I get these messages as part of the crash, which does implicate the kernel:

Mar  2 08:52:01 bolt kernel: [  558.921225] [drm:intel_dp_start_link_train [i915]] *ERROR* failed to enable link training
Mar  2 08:52:01 bolt kernel: [  558.923915] [drm:intel_dp_complete_link_train [i915]] *ERROR* failed to start channel equalization

This only seems to happen with my 4K monitor.

(and then all hell breaks loose:)
Mar  2 08:52:01 bolt /usr/lib/gdm3/gdm-x-session[1909]: (II) systemd-logind: got pause for 13:87
Mar  2 08:52:01 bolt /usr/lib/gdm3/gdm-x-session[1909]: (II) systemd-logind: got pause for 13:75
Mar  2 08:52:01 bolt /usr/lib/gdm3/gdm-x-session[1909]: (II) systemd-logind: got pause for 13:88
Mar  2 08:52:01 bolt /usr/lib/gdm3/gdm-x-session[1909]: (II) systemd-logind: got pause for 13:68
Mar  2 08:52:01 bolt /usr/lib/gdm3/gdm-x-session[1909]: (II) systemd-logind: got pause for 13:67
Mar  2 08:52:01 bolt gdm3: Child process -1136 was already dead.
Mar  2 08:52:01 bolt systemd[1]: Stopping User Manager for UID 119...
Mar  2 08:52:01 bolt systemd[1127]: Reached target Shutdown.
Comment 4 Chris Wilson 2016-03-02 18:29:38 UTC
(In reply to Andreas Kloeckner from comment #3)
> I'm not really using Wayland though. I know that the X server said something
> about Wayland events, but I'm not sure where that came from.
> 
> Occasionally (but not every time) I get these messages as part of the crash,
> which does implicate the kernel:
> 
> Mar  2 08:52:01 bolt kernel: [  558.921225] [drm:intel_dp_start_link_train
> [i915]] *ERROR* failed to enable link training
> Mar  2 08:52:01 bolt kernel: [  558.923915]
> [drm:intel_dp_complete_link_train [i915]] *ERROR* failed to start channel
> equalization

That would only result in a black screen. If that is still happening with the latest kernel, it is worth filing a separate bug for.

> This only seems to happen with my 4K monitor.
> 
> (and then all hell breaks loose:)
> Mar  2 08:52:01 bolt /usr/lib/gdm3/gdm-x-session[1909]: (II) systemd-logind:
> got pause for 13:87
> Mar  2 08:52:01 bolt /usr/lib/gdm3/gdm-x-session[1909]: (II) systemd-logind:
> got pause for 13:75
> Mar  2 08:52:01 bolt /usr/lib/gdm3/gdm-x-session[1909]: (II) systemd-logind:
> got pause for 13:88
> Mar  2 08:52:01 bolt /usr/lib/gdm3/gdm-x-session[1909]: (II) systemd-logind:
> got pause for 13:68
> Mar  2 08:52:01 bolt /usr/lib/gdm3/gdm-x-session[1909]: (II) systemd-logind:
> got pause for 13:67
> Mar  2 08:52:01 bolt gdm3: Child process -1136 was already dead.
> Mar  2 08:52:01 bolt systemd[1]: Stopping User Manager for UID 119...
> Mar  2 08:52:01 bolt systemd[1127]: Reached target Shutdown.

And

är 01 15:21:42 bolt /usr/lib/gdm3/gdm-x-session[1706]: (II) intel(0): switch to mode 1920x1080@60.0 on eDP1 using pipe 0, position (0, 0), rotation normal, reflection none
Mär 01 15:21:42 bolt /usr/lib/gdm3/gdm-x-session[1706]: (EE) intel(0): failed to set mode: Permission denied [13]
Mär 01 15:21:44 bolt gnome-session-binary[2015]: WARNING: Application 'gnome-shell.desktop' killed by signal 11
Mär 01 15:21:44 bolt gnome-session[2015]: libGL error: failed to authenticate magic 1
Mär 01 15:21:44 bolt gnome-session[2015]: libGL error: failed to load driver: i965

tells me that logind tried to revoke the active display fd.
Comment 5 Andreas Kloeckner 2016-03-03 07:00:22 UTC
Switching from gdm to lightdm seems to resolve this, so it's likely that this has nothing to do with you guys. The DP errors seem to have been a red herring. Thanks for the help!


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.