That would appear to be very different from what we're trying to report
here:

1. It's not hanging. It's presumably exiting, forcing a logout, and it
all happens before the monitors are awake enough to show a picture.
We're not in a hung state at any point.

2. It's not only suspend, it's wake from only display-sleep too. The
rest of the computer isn't suspended or sleeping in any manner. It
*also* failed, in the exact same manner, on wake from suspend, but this
suggests the problem isn't in suspending or waking from that, but just
about the display sleep/wake.

3. It's only affecting nVidia (or at least, for me, it seems not to be
also affecting Intel, but I haven't tried such a long sleep under
Intel/Xorg and I should - tomorrow, as I'm in bed now). I bet there's no
problem. :-)

4. It appears to be only affecting Xorg, although of course with nVidia
there's too much else broken when trying to force it to use Wayland
(which is, after all, not enabled by default for these reasons).
Certainly no such problems on Intel/Wayland which *does* regularly get
to sleep for long periods of time between uses, and I've never seen it
hang on wake.

(Hardware of my Intel graphics machine is an Early 2015 13" Retina
Macbook Pro. This specifically is a Broadwell Core-i5 with Intel Iris
6100, and I've only attempted to use it on its own screen, not with any
external screens connected. Generally, it runs Linux brilliantly, as far
as I've so far tried, and is very happy with Wayland. All hardware just
works OOTB.)

I'm not using the default ubuntu session (with or without wayland) on
either machine, but the GNOME session - wayland on the intel, xorg on
the nvidia. Basically just because I wanted vanilla dash-to-dock. I'm
not running many extensions, dash to dock, a wallpaper changer, ubuntu
appindicators (which isn't working on Xorg on either machine btw, to be
the subject of another bug report when I get a round tuit)... and
nothing else that I remember right now while I'm not at it. Come to
think of it, I do need to test using the default gnome shell theme
rather than Arc's. I will tomorrow.

-- 
You received this bug notification because you are a member of Desktop
Packages, which is subscribed to gnome-shell in Ubuntu.
https://bugs.launchpad.net/bugs/1720149

Title:
  gnome-shell segv on wake from display or system sleep

Status in gnome-shell package in Ubuntu:
  Confirmed

Bug description:
  The observable symptom is that if I go away and let the computer go
  either to display sleep or suspend, when I come back and try to wake
  it up again, I find myself, after sliding up the lock screen, at the
  gdm login screen. I'm logged out, and logging in gives me a fresh new
  session. Anything I had open in the previous session is lost. (I'm
  getting used to saving stuff before I step away from the computer!)

  The only trace I've been able to find is that in /var/log/syslog I see
  a line like:

  Sep 28 14:38:44 fleetfoot kernel: [10621.432586] gnome-shell[1863]:
  segfault at 2c ip 00007f6200e82434 sp 00007ffc46f981c8 error 4 in
  libmutter-1.so.0.0.0[7f6200de4000+140000]

  This happens at the time of attempted wake, not the time of going to
  sleep. It's preceded by a lot of activity from gdm-x-session that I'm
  not sure is indicating any error, just the process of waking up. I
  will attach a portion of syslog surrounding the whole event in the
  hope it helps. (In fact attaching the whole current syslog file - it
  has two such events occuring in it, first display sleep/wake at 09:02
  this morning, and secondly (because I was experimenting) a suspend-
  wake at 14:38 this afternoon.

  This is only affecting Xorg sessions, and may possibly only be
  affecting where nvidia is in use, I'm not sure about that. But the
  fact it occurs on display wake alone, not needing the system itself to
  have suspended (I was trying suspend to see if it *avoided* the
  problem, which it doesn't), does point the finger in that direction I
  guess. But ultimately I'm guessing it's gnome-shell itself segfaulting
  that's causing the login session to end?

  Not observed on my Wayland system.

  This is not new as of the current version of gnome-shell but I think
  it is relatively recent. Now my hackintosh partition is wrecked by a
  failed upgrade to High Sierra I'm using my Linux system more in
  earnest than before, so I'm hitting this often enough for it to
  provoke a bug report (and much of the time, to just turn all auto
  suspend/display-sleep off and instead shut down the computer when I'm
  leaving it for any length of time).

  The ubuntu bug reporter is not picking this up by itself, so I presume
  a crash report of the sort that uses is not being saved.

  ProblemType: Bug
  DistroRelease: Ubuntu 17.10
  Package: gnome-shell 3.26.0-0ubuntu2
  ProcVersionSignature: Ubuntu 4.13.0-12.13-generic 4.13.3
  Uname: Linux 4.13.0-12-generic x86_64
  NonfreeKernelModules: nvidia_uvm nvidia_drm nvidia_modeset nvidia
  ApportVersion: 2.20.7-0ubuntu1
  Architecture: amd64
  CurrentDesktop: GNOME
  Date: Thu Sep 28 14:41:07 2017
  DisplayManager: gdm3
  InstallationDate: Installed on 2017-07-30 (59 days ago)
  InstallationMedia: Ubuntu-GNOME 17.04 "Zesty Zapus" - Release amd64 (20170412)
  SourcePackage: gnome-shell
  UpgradeStatus: Upgraded to artful on 2017-08-22 (37 days ago)

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/gnome-shell/+bug/1720149/+subscriptions

-- 
Mailing list: https://launchpad.net/~desktop-packages
Post to     : [email protected]
Unsubscribe : https://launchpad.net/~desktop-packages
More help   : https://help.launchpad.net/ListHelp

Reply via email to