Kernel panic, system freezes, no BSOD

i have an RTX 3050 Mobile on an Asus Tuf A17 with R7-4800H and 16G

Do you mean the build error with the 6.10 kernel? Are you perhaps setting boot.kernelPackages = pkgs.linuxPackages_latest or something similar in your config?

Afaik, boot.kernelPackages should default to pkgs.linuxPackages, which is an alias for pkgs.linuxKernel.packageAliases.linux_default which should be kernel version 6.6, not 6.10.

If that isn’t the issue you are experiencing, please provide more information beyond “some vague error i couldnt fix”.

Your GPU is fairly recent, so it shouldn’t be a problem. What’s your current hardware configuration?

Also, can you collect more information about this? Perhaps use the nvidia-bug-report.sh or get the previous system logs with journalctl -b -1

i did set the kernel to _6_9 as you suggested, it just got built but i was unable to login to gnome --_–

i have added the journalctl -b --log-level=3 output to gist:13bfdd54660a8f51ba47408c5ab210f5 · GitHub
this is after the direct gdm freeze on login, (went to tty3 to get the log)

this is the config i have currently.

I didn’t suggest setting the kernel to _6_9. I was suggesting not setting it to _latest, instead keeping it as the default (which should currently be an alias to _6_6). But that is beside the point)))

“unable to login to gnome” isn’t too specific. Are you getting a kernel panic when trying to login? Or does it just throw you back to the login screen? Or is there some other error/crash you are getting?

P.S. This thread in discourse is already closed, because I was able to fix my problem. I recommend that you create your own topic/thread, so that your issue gets more visibility.

this is working now, it didnt when all those environment variables weren’t commented,
but now without these i wont be able to run firefox or wezterm over the gpu.

( i am sorry if i feel crude or rude, english is not my first language) i really appreciate all the help you all give.

Try removing this line:

  boot.kernelParams = [ "nvidia-drm.fbdev=1"];

According to the nvidia module, it should be set only for drivers 545 and above.

Can you pinpoint which one exactly is causing this?

offo, it is still working rn with it still set. i just commented all the env vars and it worked. is there anything you can find… (after your reply). i’ll probably trial-and-error them, just uncommenting them one by one

I’d try the Wayland-related ones first, since those are the ones that seem to be crashing gdm.

1 Like

!! i was trying to explain it again but this issue is opened somewhere else as well, i wont waste your time.

i am new to this discourse thing and forums, i felt like my issue aligned with yours so i continued. would happily open a new issue if i couldnt get it fixed. would write a blog post on it when i fix.

1 Like

My point is that since this is a closed topic, your questions are only getting responses from the people who originally participated in this thread. If you create your own thread, it will be seen by some of the people that monitor and answer questions in the Help topic.

Also, your problems might not even be related to the nvidia kernel panics at this point since you are getting core dumps in journalctl which indicates a regular userland crash, not a kernel panic.

Good luck

2 Likes

GBM_BACKEND = "nvidia-drm";
is the culprit,
no kernel panics yet, working fine still, would create a new thread if anything goes haywire now.