Gnome apps segfault in Nvidia Wayland session (noble)

[Jonathan Cave]

I was happily using a wayland session on lunar on an AMD Threadripper workstation with an NVIDIA Quadro discrete graphics card.

Since upgrade to noble I have experienced the similar symptons as https://askubuntu.com/questions/1514352/ubuntu-24-04-with-nvidia-driver-libegl-warning-egl-failed-to-create-dri2-scre except in this case there is no Intel chipset to be a root cause.

I am using nividia-driver-535

ProblemType: Bug
DistroRelease: Ubuntu 24.04
Package: gnome-shell 46.0-0ubuntu5.1
ProcVersionSignature: Ubuntu 6.8.0-31.31-generic 6.8.1
Uname: Linux 6.8.0-31-generic x86_64
NonfreeKernelModules: zfs nvidia_modeset nvidia
ApportVersion: 2.28.1-0ubuntu3
Architecture: amd64
CasperMD5CheckResult: pass
CurrentDesktop: ubuntu:GNOME
Date: Mon Jun 3 09:14:55 2024
DisplayManager: gdm3
InstallationDate: Installed on 2023-05-23 (377 days ago)
InstallationMedia: Ubuntu 23.04 "Lunar Lobster" - Release amd64 (20230417)
RelatedPackageVersions: mutter-common 46.0-1ubuntu9
SourcePackage: gnome-shell
UpgradeStatus: Upgraded to noble on 2024-04-29 (35 days ago)

[Daniel van Vugt (vanvugt)]

It looks like you have a lot of bug 1967707, but I think that's just a symptom. It also looks like mutter is trying to treat simpledrm as a secondary GPU (similar to bug 1978905), but that GPU doesn't really exist (bug 2060268).

Please try one of the workarounds in bug 2060268. If that doesn't solve it then please run:

  journalctl -b0 > journal.txt
  lspci -k > lspci.txt

and attach the resulting text files here.

[Jonathan Cave (jocave)]

I confirmed that I did have the Unknown Display problem so I blacklisted simpledrm_platform_driver_init ass suggested and the extra display went away. However, gnome apps continue to segfault while in a Wayland session.

Will attachi the other logs as requested when I've switched back to xorg session...

[Daniel van Vugt (vanvugt)]

Thanks. If there are segfaults then please use these steps to try and track down some of the crash reports:

https://wiki.ubuntu.com/Bugs/Responses#Missing_a_crash_report_or_having_a_.crash_attachment

[Jonathan Cave (jocave)]

This appears to be a relevant crash report: https://errors.ubuntu.com/oops/2d43fe25-2263-11ef-9e41-fa163ec44ecd

[Daniel van Vugt (vanvugt)]

Yes the last few crashes from that machine do seem to all be the same problem in different binaries. There's no debug information (I don't know why that's a common issue) but I can tell all the crashes seem to be using Zink from Mesa which is the wrong driver for Nvidia. This suggests the Nvidia driver is not correctly installed.

I suggest using the Additional Drivers app to remove the Nvidia driver, reboot, and then use the same app to install a fresh (possibly different?) Nvidia driver.

[Daniel van Vugt (vanvugt)]

Slightly more detail:
https://bugs.launchpad.net/ubuntu/+source/xdg-desktop-portal-gnome/+bug/2068869/comments/4

[Daniel van Vugt (vanvugt)]

Also reported in bug 2063827