You are not logged in.
Hello everyone,
Yesterday i updated my system and Gnome stopped working on my GPU (nvidia GTX1060 6GB driver nvidia-580.95.05-2)
When I rebooted the computer I got a black screen with a cursor on the top left that was constantly refreshing (preventing me to switch with ctrl+alt+F2, for instance) eventually I would land on the log-in screen (which is unusual because I have auto log-in enabled). The log-in screen does not work and freeze after one attempt. However then I am able to switch tty, log-in an disable gdm.service, reboot and now it boots properly into... tty. When I try to launch
gnome-shell --wayland
I get the following output :
libmutter-Message: 10:21:55.658: Running GNOME Shell (using mutter 49.0) as a Wayland display server
libmutter-Message: 10:21:55.706: Thread 'KMS thread' will be using high priority scheduling
libmutter-Message: 10:21:55.738: Device '/dev/dri/card1' prefers shadow buffer
libmutter-Message: 10:21:55.794: Added device '/dev/dri/card1' (nvidia-drm) using atomic mode setting.
libmutter-Message: 10:21:55.795: Failed to initialize accelerated iGPU/dGPU framebuffer sharing: An EGLDisplay argument does not name a valid EGL display connection.
libmutter-Message: 10:21:55.795: Created gbm renderer for '/dev/dri/card1'
libmutter-Message: 10:21:55.795: Boot VGA GPU /dev/dri/card1 selected as primary
Failed to setup: Unable to initialize the Clutter backend: no available drivers found.
Which I don't understand because I do have the nvidia drivers installed the drivers according to the infos on the Wiki https://wiki.archlinux.org/title/NVIDIA.
So far I tried :
- Updating the driver (new version came today, I had hope)
- Reinstalling the driver, verifying that DRM is enabled, verifying that nvidia, nvidia_modeset, nvidia_uvm, nvidia_drm modules are in /etc/mkinitcpio.conf and regenerating the initramfs
- Installing another driver (nvidia-open and nvidia_dkms)
- Reinstalling gnome and wayland using pacman
- Downgrading the driver
With no success.
I've used this system for around a year (and arch for 2 years) and never had issues with nvidia. I did not change any hardware, just updated with
sudo pacman -Syu
Usually I am able to fix any issues by myself using what I know and documentations, wikis and forums but here I feel stuck and can't find anything.
Thanks in advance for your contributions
Offline
Please post your complete system journal for the boot:
sudo journalctl -b | curl -F 'file=@-' 0x0.st
Offline
Sorry forgot the journal because I was so focused on the log from the command I mentioned in my post.
I logged in; tried to launch gnome-shell --wayland ( got the same error by the way); then tried to start gdm.service (failed like i described) and then switched tty to make the journal.
Thanks for the command that uploads the journal to an URL, it's very convenient!
Offline
First and foremost clean up your network stack - you've at least NM, networkd and dhcpcd enabled, pick one, disable the others and then post the output of
find /etc/systemd -type l -exec test -f {} \; -print | awk -F'/' '{ printf ("%-40s | %s\n", $(NF-0), $(NF-1)) }' | sort -f
for a sanity check.
GDM seems to run fine on wayland, it's only
oct. 08 14:35:56 archugo gnome-shell[1835]: Running GNOME Shell (using mutter 49.0) as a Wayland display server
oct. 08 14:35:56 archugo rtkit-daemon[973]: Supervising 7 threads of 4 processes of 1 users.
oct. 08 14:35:56 archugo rtkit-daemon[973]: Successfully made thread 1869 of process 1835 owned by '1000' high priority at nice level -15.
oct. 08 14:35:56 archugo rtkit-daemon[973]: Supervising 8 threads of 5 processes of 2 users.
oct. 08 14:35:56 archugo gnome-shell[1835]: Thread 'KMS thread' will be using high priority scheduling
oct. 08 14:35:56 archugo gnome-shell[1835]: Device '/dev/dri/card1' prefers shadow buffer
oct. 08 14:35:56 archugo gnome-shell[1835]: Added device '/dev/dri/card1' (nvidia-drm) using atomic mode setting.
oct. 08 14:35:56 archugo gnome-shell[1835]: Failed to initialize accelerated iGPU/dGPU framebuffer sharing: An EGLDisplay argument does not name a valid EGL display connection.
oct. 08 14:35:56 archugo gnome-shell[1835]: Created gbm renderer for '/dev/dri/card1'
oct. 08 14:35:56 archugo gnome-shell[1835]: Boot VGA GPU /dev/dri/card1 selected as primary
oct. 08 14:35:56 archugo gnome-shell[1835]: Failed to setup: Unable to initialize the Clutter backend: no available drivers found.
oct. 08 14:35:56 archugo systemd[1736]: org.gnome.Shell@wayland.service: Failed with result 'protocol'.
oct. 08 14:35:56 archugo systemd[1736]: Failed to start GNOME Shell on Wayland.
oct. 08 14:35:56 archugo systemd[1736]: org.gnome.Shell@wayland.service: Triggering OnFailure= dependencies.
This doesn't look like a hybrid system and there're no glaring errors w/ the nvidia GPU.
Log into a console and post
printenv | curl -F 'file=@-' 0x0.st
Can you log in from GDM into a gnome session for a fresh user?
Offline
I have disable network manager and dhcpcd. The command you send me returned this :
bluetooth.service | bluetooth.target.wants
dbus-org.bluez.service | system
dbus-org.freedesktop.network1.service | system
dbus-org.freedesktop.resolve1.service | system
dbus-org.freedesktop.timesync1.service | system
gcr-ssh-agent.socket | sockets.target.wants
getty@tty1.service | getty.target.wants
gnome-keyring-daemon.socket | sockets.target.wants
lm_sensors.service | multi-user.target.wants
nvidia-hibernate.service | systemd-hibernate.service.wants
nvidia-resume.service | systemd-hibernate.service.wants
nvidia-resume.service | systemd-suspend.service.wants
nvidia-suspend.service | systemd-suspend.service.wants
p11-kit-server.socket | sockets.target.wants
pipewire-pulse.socket | sockets.target.wants
pipewire-session-manager.service | user
pipewire.socket | sockets.target.wants
remote-fs.target | multi-user.target.wants
speech-dispatcher.socket | sockets.target.wants
systemd-networkd.service | multi-user.target.wants
systemd-networkd.socket | sockets.target.wants
systemd-networkd-varlink.socket | sockets.target.wants
systemd-networkd-wait-online.service | network-online.target.wants
systemd-network-generator.service | sysinit.target.wants
systemd-resolved.service | sysinit.target.wants
systemd-timesyncd.service | sysinit.target.wants
wireplumber.service | pipewire.service.wants
xdg-user-dirs.service | graphical-session-pre.target.wants
Log into a console and post
Can you log in from GDM into a gnome session for a fresh user?
If i make a new user and then start gdm.service and log with that user it works fine. (Which makes it even more confusing to me...) However I've noticed that it takes less time between the moment I start gdm.service and the moment the screen appears.
Edit : I wanted to point out that when I try with my user it does a black screen and put me back on the log-in screen where i can select any user. If i select my user again it says "Authentication error".
Last edited by Koraku54 (2025-10-08 15:09:31)
Offline
systemd-networkd is still enabled?
One big red flag is
ZSH=/home/hugo/.oh-my-zsh
but this might also be specific to your gnome configuration - you could try to **move** away ~/.config (eg. rename it to .config.old before trying to login)
Offline
systemd-networkd is still enabled?
Yes it is
I installed zsh way back when i configured the machine.
I changed my shell back to bash, did the printenv angain and the variable ZSH wasn't there anymore however I would still be stuck at the log-in screen when I start gdm.service I also tried to rename .config like you suggested but that didn't make a difference.
At this point I'm considering deleting the old user (after backing up my data and important configs of course... ) and reconfiguring a new one since the problem seems to come from MY user... After digging in the folders a bit there are a lot of things that belong to programs that I don't use anymore or I could redo a clean install (I've learn a lot since I did it last time... and read the wiki a bunch more) but that would take a lot of time and effort.
Last edited by Koraku54 (2025-10-08 16:28:42)
Offline
Yes it is
Disable it?
Does your old user have a ~/.profile file?
~/.cache might be a problem and recent versions of pixbuf2 seem to have problems w/ symlinked fonts (~/.local/share/fonts)
Can your old user log into a different kind of session, lxqt, xfce, or just sway or openbox?
At this point I'm considering deleting the old user (after backing up my data and important configs of course... ) and reconfiguring a new one
An alternative strategy would be to poison then new user (copy stuff from your old user there) until it breaks, too.
Offline
Disable it?
Sadly didn't make a difference.
Does your old user have a ~/.profile file?
It does not
~/.cache might be a problem
I deleted the cache and also did a lot of clean up in my home directory, sadly no change.
recent versions of pixbuf2 seem to have problems w/ symlinked fonts (~/.local/share/fonts)
Wierdly enough, i don't have ~/.local/share/fonts
Can your old user log into a different kind of session, lxqt, xfce, or just sway or openbox?
I will install one and try (will edit when I did with result)
This issue make less sense the more I think about it. The title does not seem appropriate anymore.
Offline
What's the output of "glxinfo -B" from your new test user (just to make sure that one's not running on software emulation)?
Also post the journal from such login and then try to add
module_blacklist=simpledrm initcall_blacklist=simpledrm_platform_driver_init
to the https://wiki.archlinux.org/title/Kernel_parameters - maybe something™ mistakes the simpledrm device for a hybrid setup…
Offline