gnome-shell crashes on Lenovo Thinkpad P50 after upgrade to 22.04.1

[maarten]

I upgrade to Ubuntu 22.04 recently, and while upgrading and updating I also got a new bios version (1.69) for my Lenovo Thinkpad P50 laptop. Since then gnome-shell has been very unstable, crashing regularly on minor changes, like plugging in a monitor or going to suspend mode.

Update: I just found out that the bug is not reproducable when logged in with Xorg instead of Wayland (https://askubuntu.com/questions/1428898/lenovo-p50-crashes-gnome-shell-regularly-after-upgrade-to-ubuntu-22-04-1-using)

So we got a workaround and a culprit: Wayland ... in conjunction with something quite typical for my hardware maybe(?), Lenovo Thinkpad P50 (20ENCTO1WW) with Quadro M2000M discrete graphics ...

ProblemType: Bug
DistroRelease: Ubuntu 22.04
Package: gnome-shell 42.4-0ubuntu0.22.04.1
ProcVersionSignature: Ubuntu 5.15.0-47.51-generic 5.15.46
Uname: Linux 5.15.0-47-generic x86_64
ApportVersion: 2.20.11-0ubuntu82.1
Architecture: amd64
CasperMD5CheckResult: unknown
CurrentDesktop: ubuntu:GNOME
Date: Mon Sep 12 09:49:57 2022
DisplayManager: gdm3
InstallationDate: Installed on 2022-09-09 (2 days ago)
InstallationMedia: Ubuntu 20.04.1 LTS "Focal Fossa" - Release amd64 (20200731)
ProcEnviron:
 TERM=xterm-256color
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=nl_NL.UTF-8
 SHELL=/bin/bash
RelatedPackageVersions: mutter-common 42.2-0ubuntu1
SourcePackage: gnome-shell
UpgradeStatus: Upgraded to jammy on 2022-09-09 (2 days ago)

[Daniel van Vugt (vanvugt)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. It sounds like some part of the system has crashed. To help us find the cause of the crash please follow these steps:

1. Look in /var/crash for crash files and if found run:
    ubuntu-bug YOURFILE.crash
Then tell us the ID of the newly-created bug.

2. If step 1 failed then look at https://errors.ubuntu.com/user/ID where ID is the content of file /var/lib/whoopsie/whoopsie-id on the machine. Do you find any links to recent problems on that page? If so then please send the links to us.

Please take care to avoid attaching .crash files to bugs as we are unable to process them as file attachments. It would also be a security risk for yourself.

[maarten (info-maartenabbring)]

1) did that, but don't know where to look for the bug-id
2) used the whoopsie-id but that didn't give me new insights

[Daniel van Vugt (vanvugt)]

The web page https://errors.ubuntu.com/user/... will initially show no results but updates after a few seconds. If you still see no results after 10 seconds then next please provide a log by:

1. Wait for the crash to occur again.

2. Wait 10 seconds.

3. Reboot.

4. Run:

   journalctl -b-1 > prevboot.txt

5. Attach the resulting text file here.

[maarten (info-maartenabbring)]

I'm sorry, but have spend lots of time working around this bug and need to focus on this weird phenomenon called work first, but will try to reproduce as soon as I have time

[maarten (info-maartenabbring)]

Still haven't got time to reproduce, but I do think as long as this hardware combination doesn't work out of the box with wayland it should be disabled automatically for this hardware after installation or upgrade.

This script is already disabling Wayland for lots of hardware configurations:

/usr/lib/udev/rules.d/61-gdm.rules

First step would be to find the most relevant values for my hardware that is proven not to work. More will follow. Would be so nice if I could make sure more Ubuntu users have a smooth experience with an out-of-the-box installation

[Daniel van Vugt (vanvugt)]

The most likely cause of this crash is locally installed extensions and you seem to have a few. Please start by deleting them:

  cd ~/.local/share/gnome-shell/
  rm -rf extensions

and then log in again.

[Daniel van Vugt (vanvugt)]

Another common cause of crashes like this is the "nouveau" graphics driver. It's not very reliable so we recommend opening the "Additional Drivers" app and using it to install the NVIDIA driver instead.

[maarten (info-maartenabbring)]

@vanvugt: I've tried disabling all extensions but that did not have any correlation with crashes. It kept on crashing regardless, also before they were installed at all.

It would be worrying that something quite high level can crash the entire window manager. It's disturbing. The culprit is Wayland itself, because a Window manager should never crash in total, regardless of what a single application or module is trying to do. It is interesting what triggers it, so we can get to the cause within Wayland

@the nvidia drivers have all kinds of other issues in my hybrid setup. The nouveau driver has proven to be much more reliable out-of-the-box for this setup. It has been with other hardware for decades and still. Nvidia can't seem to get it right, even after version 1 million + 1. For that reason, I prefer my next laptop to not have a discrete graphics card at all

[Daniel van Vugt (vanvugt)]

The only definite crash I can see here happened during GPU startup:

[ 3020.788168] lapis gnome-shell[13120]: Created gbm renderer for '/dev/dri/card1'
[ 3021.413765] lapis gnome-shell[13120]: GNOME Shell crashed with signal 11
[ 3021.413765] lapis gnome-shell[13120]: == Stack trace for context 0x555a549a04b0 ==

And then it restarted successfully:

[ 3037.624313] lapis gnome-shell[22361]: Running GNOME Shell (using mutter 42.2) as a Wayland display server
[ 3037.648641] lapis gnome-shell[22361]: Device '/dev/dri/card1' prefers shadow buffer
[ 3037.951835] lapis gnome-shell[22361]: Added device '/dev/dri/card1' (nouveau) using non-atomic mode setting.
[ 3037.953804] lapis gnome-shell[22361]: Device '/dev/dri/card0' prefers shadow buffer
[ 3037.956366] lapis gnome-shell[22361]: Added device '/dev/dri/card0' (i915) using atomic mode setting.
[ 3037.995604] lapis gnome-shell[22361]: Created gbm renderer for '/dev/dri/card1'
[ 3038.005524] lapis gnome-shell[22361]: Created gbm renderer for '/dev/dri/card0'
[ 3038.005728] lapis gnome-shell[22361]: Boot VGA GPU /dev/dri/card0 selected as primary

So I believe nouveau may be to blame (it usually is). If you can then please disable the discrete GPU in the BIOS and retest. Please also check for crash reports per comment #2.

[maarten (info-maartenabbring)]

Well, I have the choice in my BIOS between hybrid graphics and discrete graphics, no option to disable the discrete graphics card entirely. And before you ask: I will not run on the discrete graphics card alone, because of the battery drain that comes with it

[Daniel van Vugt (vanvugt)]

OK, next time (each time) the crash happens, please:

1. Wait 10 seconds.

2. Reboot.

3. Run:

   journalctl -b-1 > prevboot.txt

4. Attach the resulting text file here.

5. Redo the steps in comment #2 at the top of the page.

[Launchpad Janitor (janitor)]

[Expired for gnome-shell (Ubuntu) because there has been no activity for 60 days.]