Once logged in, GDM Shell process CPU usage spiking at 100%

[Didier Roche-Tolomelli]

This happens on some nvidia card, with the nvidia blob driver (fallbacked to Xorg for gdm).

1. Boot your laptop
2. Log into gdm
-> once your user session is loaded, gdm GNOME Shell process goes crazy, up to 100%, continously.

GNOME Shell (for user gdm) restarts in loops due to not be able to paint (context lost). More info on the upstream bug: https://gitlab.gnome.org/GNOME/gdm/issues/429#note_344826.

#gnome-shell IRC logs:
   jadahl | so its the context-lost handling going crazy then
 didrocks | looks like so…
 didrocks | jadahl: FYI, I have a gdb shell opened right now, tell me if you have time or if I can help for anything
   jadahl | didrocks: I suppose what we need is a wayt to tell if its a context lost because of gdm switch away, or if its a context lost that should be handled
   jadahl | can't say I know at hand how to make that distinction
 didrocks | makes sense (surprised though that it doesn't seem to hit everyone fallbacking to Xorg for gdm). Is it something that should be discussed with halfline?
halfline1 | check if its foreground vt i guess...
   jadahl | didrocks: only nvidia has the context lost thing
 didrocks | jadahl: Laney has a nvidia card as well, he checked that he's fallback to Xorg as well but doesn't have that issue apparently, which is puzzling…
halfline1 | i wouldnt be surprised if theres a driver bug involved
 didrocks | yeah, I think that I'm on the "unlucky card triggering this nvidia driver bug" case :p

ProblemType: Bug
DistroRelease: Ubuntu 18.10
Package: mutter 3.30.1-1
ProcVersionSignature: Ubuntu 4.18.0-8.9-generic 4.18.7
Uname: Linux 4.18.0-8-generic x86_64
NonfreeKernelModules: nvidia_modeset nvidia
ApportVersion: 2.20.10-0ubuntu11
Architecture: amd64
CurrentDesktop: ubuntu:GNOME
Date: Thu Oct 11 12:19:41 2018
InstallationDate: Installed on 2018-05-24 (139 days ago)
InstallationMedia: Ubuntu 18.04 LTS "Bionic Beaver" - Release amd64 (20180426)
ProcEnviron:
 TERM=xterm-256color
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=fr_FR.UTF-8
 SHELL=/bin/bash
SourcePackage: mutter
UpgradeStatus: No upgrade log present (probably fresh install)

[Didier Roche-Tolomelli (didrocks)]

didrocks | yes ;) I wonder what would happen if, while this is worked on, we remove the udev rule to not regress some of our nvidia users for 18.10 release next week (as
         | a workaround)
halfline | didrocks: yea the udev rule makes less sense for you indeed
halfline | since you default to Xorg for user sessions
didrocks | right, so if we keep gdm / wayland and user sessions -> Xorg, that could still be an option
didrocks | (until a proper fix is found)
halfline | makes sense to me

[Didier Roche-Tolomelli (didrocks)]

Assigning it to gdm, even if the loop is in mutter, as we are going to workaround it there.

[Brian Curtis (bcurtiswx)]

NVIDIA GeForce GTX 970 using driver 390.87 64Bit architecture

I know that my windows NVIDIA drivers are in the 400's. Would it be worth testing a newer NVIDIA driver?

[Daniel van Vugt (vanvugt)]

See also bug 1764137 and its upstream link.

[Launchpad Janitor (janitor)]

This bug was fixed in the package gdm3 - 3.30.1-1ubuntu3

---------------
gdm3 (3.30.1-1ubuntu3) cosmic; urgency=medium

  * Build-Depend on gjs so that we don't produce uninstallable packages
    on s390x which doesn't have gjs (LP: #1794721)

 -- Jeremy Bicha <email address hidden> Thu, 11 Oct 2018 13:52:55 -0400