Garuda Zen Mokka - apparently random System freezes - lots of errors

[Nash]

After some research, the only relevant thing I found is this thread suggesting performance mode may have an impact on this problem.
However I can’t find any documentation regarding hash tables outside of programming software. Nvidia forums have a couple threads regarding driver errors and some users report the same issue, among others, but it just gets mentioned.
From my understanding, since hash table resizing/resetting is possible, performance settings might influence the way the system handles RAM (in this case vRAM) in limiting how much can be used.
I will wait for some feedback and/or other sources before trying to change more settings.

At this moments the only things mentioned that look reasonable to me would be switching kernel and/or changing power manager and tweaking power management modes/settings. But I wouldn’t know where it would be better to start.

[LolFlasherNoobX]

IMHO
Do them as follows:

1. Change Kernels
2. Tweak power management settings
3. Change power managers

If you are gonna do these. Should you do these first disregarding @nepti 's advice? I don’t think so because he certainly is a wizard .

Should you even do these? Idk. I would probably be bald pulling my hairs at this point.

[Colin]

@LolFlasherNoobX

I would probably be bald pulling my hairs at this point.

I’ve gone bald reading this thread - yikes!

[nepti]

@Nash

Last night I switched the GPU PCI-e slot. While performing the swap, I noticed the GPU was not sitting at a right angle as it should have, and reinstalling I also noticed some screws wouldn’t properly thread all the way. So it is a possibility that the weight of the card over the years bent the slot it was sitting in. Now it’s installed properly and it does not move around. The system has been running for the last 12 hours with no freezes.

If your system now works longer without issues and you no longer get a “GPU has fallen of the bus” error message, then the correct reinstallation of the GPU should have fixed this source of error.

@Nash

mag 06 22:22:25 home-tower-station kwin_wayland[5978]: kwin_wayland_drm: Pageflip timed out! This is a bug in the nvidia-drm kernel driver
mag 06 22:22:25 home-tower-station kwin_wayland[5978]: kwin_wayland_drm: Please report this at https://forums.developer.nvidia.com/c/gpu-gra>
mag 06 22:22:25 home-tower-station kwin_wayland[5978]: kwin_wayland_drm: With the output of 'sudo dmesg' and 'journalctl --user-unit plasma->
mag 06 22:22:26 home-tower-station kernel: NVRM: _instmemAddHashEntry: Display Hash table is FULL!!
mag 06 22:22:26 home-tower-station kernel: NVRM: _instmemAddHashEntry: Display Hash table is FULL!!
mag 06 22:22:26 home-tower-station kernel: NVRM: _instmemAddHashEntry: Display Hash table is FULL!!
mag 06 22:22:26 home-tower-station kernel: NVRM: _instmemAddHashEntry: Display Hash table is FULL!!
mag 06 22:22:26 home-tower-station kernel: NVRM: _instmemAddHashEntry: Display Hash table is FULL!!
mag 06 22:22:26 home-tower-station kernel: NVRM: _instmemAddHashEntry: Display Hash table is FULL!!
mag 06 22:22:26 home-tower-station kernel: NVRM: _instmemAddHashEntry: Display Hash table is FULL!!

This is a bug in the nvidia driver, try switching to an x11 session (you can switch between wayland and x11 at the bottom left of the login screen) and check if everything works there without issues.