KDE - Slow Login and automatic Logout

Issues & Assistance
1

Hello,

since a few weeks I have a problem with my KDE on Cachyos.

First the Login takes around 30 seconds after I entered my password and pressed Enter.
Then around 90-120 seconds later it shows me a new Login-Screen. I guess it’s a new session, because when I press control + alt and F1 oder F2 on my keyboard, I’m back in me logged in session. I was hopping that there was a fix in the repo but still with the newest packages I still have this error. And also I can’t run “reboot” or “shutdown -h now” as a normal user. I have to do this as root with sudo. Call to Reboot failed: Interactive authentication required.

I’m using:
CachyOS 24.12
Kernel Linux 6.16.0-5-cachyos
DE: KDE Plasma 6.4.4
WM: KWin (Wayland)
Bugreport: wastebin

Thanks
haruzept

2

This is due to Polkit permissions requiring authentication.
To allow shutdown/reboot without entering a password for users in the wheel group:
sudo nano /etc/polkit-1/rules.d/49-nopasswd-shutdown.rules
add

polkit.addRule(function(action, subject) {
    if ((action.id == "org.freedesktop.login1.reboot" ||
         action.id == "org.freedesktop.login1.power-off") &&
        subject.isInGroup("wheel")) {
        return polkit.Result.YES;
    }
});

save and reboot

3

Use systemd-analyze blame to see which services take the longest during login

4

❯ systemd-analyze blame
3.935s plymouth-quit-wait.service
3.935s plymouth-quit.service
1.110s plocate-updatedb.service
1.048s mnt-backup.mount
325ms upower.service
325ms NetworkManager.service
281ms dev-nvme0n1p2.device
250ms ufw.service
152ms systemd-udev-trigger.service
134ms dev-zram0.swap
100ms user@965.service
95ms user@1000.service
89ms systemd-journald.service
88ms lvm2-monitor.service
85ms systemd-modules-load.service
82ms udisks2.service
68ms systemd-tmpfiles-setup.service
59ms polkit.service
58ms plymouth-start.service
54ms systemd-tmpfiles-clean.service
52ms systemd-remount-fs.service
50ms systemd-tmpfiles-setup-dev-early.service
50ms systemd-udevd.service
47ms logrotate.service
43ms systemd-boot-random-seed.service
42ms systemd-resolved.service
40ms boot.mount
40ms systemd-hostnamed.service
40ms systemd-logind.service
40ms systemd-zram-setup@zram0.service
35ms power-profiles-daemon.service
34ms avahi-daemon.service
34ms plymouth-read-write.service
33ms systemd-fsck@dev-disk-by\x2duuid-B6CE\x2d9FE0.service
33ms home.mount
32ms root.mount
31ms alsa-restore.service
31ms srv.mount
31ms systemd-vconsole-setup.service
30ms systemd-binfmt.service
30ms systemd-udev-load-credentials.service
27ms user-runtime-dir@965.service
27ms systemd-update-utmp.service
27ms user-runtime-dir@1000.service
26ms modprobe@dm_mod.service
25ms modprobe@loop.service
23ms systemd-user-sessions.service
23ms bluetooth.service
22ms dbus-broker.service
22ms systemd-tmpfiles-setup-dev.service
21ms systemd-timesyncd.service
20ms systemd-journal-flush.service
14ms sys-fs-fuse-connections.mount
14ms systemd-userdbd.service
14ms sys-kernel-config.mount
11ms dev-hugepages.mount
11ms dev-mqueue.mount
10ms sys-kernel-debug.mount
10ms systemd-rfkill.service
10ms sys-kernel-tracing.mount
9ms kmod-static-nodes.service
9ms modprobe@configfs.service
8ms modprobe@drm.service
8ms modprobe@fuse.service
7ms rtkit-daemon.service
5ms systemd-sysctl.service
5ms systemd-random-seed.service
4ms var-log.mount
3ms var-tmp.mount
3ms var-cache.mount
3ms proc-sys-fs-binfmt_misc.mount
3ms tmp.mount

5

Part of the delay is caused by Plymouth holding up the session:

3.935s plymouth-quit-wait.service
3.935s plymouth-quit.service

That’s almost 8 seconds waiting for the splash screen to exit
You can disable Plymouth to test:

sudo systemctl disable plymouth-quit-wait.service plymouth-quit.service plymouth-start.service
sudo mkinitcpio -P

or edit /etc/plymouth/plymouthd.conf to reduce timeout

Also check startup apps in System Settings > Startup and Shutdown > Autostart

6

journalctl shows:

kwin_wayland: kwin_scene_opengl: Invalid framebuffer status: "GL_FRAMEBUFFER_INCOMPLETE_ATTACHMENT"

Maybe this is the problem? This seems to be a KWin Wayland crash.
The problem starts after the last big cachyos changes on the kernel/nvidia-driver

I don’t have any apps in Autostart

7

That’s a KWin Wayland crash caused by the NVIDIA driver after the recent kernel/driver update
try switch to X11

Log out so you see the login screen (SDDM).
Look for the gear icon / session menu near your username.
Click it and choose Plasma (X11)
Log in normally.
If the auto-logout stops, the bug is 100% in NVIDIA’s Wayland driver, not your settings

8

if you want to keep Wayland , Roll back NVIDIA driver

  1. Open Konsole (terminal).
  2. Run:
ls -1 /var/cache/pacman/pkg/ | grep nvidia

and Tell me what you see I’ll give you the exact downgrade command.

9

I tried this and it seems to fix my problem:

for c in /sys/class/drm/card*; do
  printf "%s -> vendor=%s driver=%s\n" \
    "$c" "$(cat "$c/device/vendor" 2>/dev/null)" \
    "$(basename "$(readlink -f "$c/device/driver")")"
done

Output:

/sys/class/drm/card0 -> vendor=0x10de driver=nvidia
/sys/class/drm/card0-DP-1 -> vendor= driver=driver
/sys/class/drm/card0-DP-2 -> vendor= driver=driver
/sys/class/drm/card0-DP-3 -> vendor= driver=driver
/sys/class/drm/card0-HDMI-A-1 -> vendor= driver=driver
/sys/class/drm/card0-HDMI-A-2 -> vendor= driver=driver
/sys/class/drm/card1 -> vendor=0x1002 driver=amdgpu
/sys/class/drm/card1-DP-4 -> vendor= driver=driver
/sys/class/drm/card1-HDMI-A-3 -> vendor= driver=driver
/sys/class/drm/card1-Writeback-1 -> vendor= driver=driver

and then:

mkdir -p ~/.config/plasma-workspace/env
printf 'export KWIN_DRM_DEVICES=/dev/dri/card0:/dev/dri/card1\n' > ~/.config/plasma-workspace/env/kwin-env.sh
chmod +x ~/.config/plasma-workspace/env/kwin-env.sh

And then reboot. Login is fast and no crash after login.