GTX 1080 Ti + NVIDIA 580.x + Kernel 6.12 Crash

Message

terramotu · Post by **terramotu** » Wed Jan 28, 2026 11:55 am

Hello!
I must admit, I'm a little bit hopeless writing this post…

My Gentoo setup is:
OpenRC + dist-kernel (6.12.63) + -wayland USE flag + XFCE4 + Redshift + proprietary NVIDIA drivers (580.95.05)

My hardware setup is:
GTX 1080 Ti (Pascal) + 3 QHD screens (2560×1440, 144 Hz)

When I launch a game (Vulkan/Steam), the system works for a couple of minutes, then I notice my applications in the background getting terminated.
It starts with Steam relaunching itself, then Discord, the taskbar, the XFCE application launcher, and finally the game itself.
After that, I’m only left with my wallpaper.

I tried switching to a TTY, but it’s impossible to log in, so I’m forced to use the magic SysRq.

I also tried running the apps with only one monitor, but it just delays the issue.
I played with nvidia-drm.modeset according to the Gentoo NVIDIA wiki https://wiki.gentoo.org/wiki/NVIDIA/nvidia-drivers, but I think it’s irrelevant for Pascal GPUs.

Has anyone else experienced this issue on Pascal GPUs with 580.x drivers on Gentoo?
Are there any workarounds besides downgrading the driver or kernel? (I read that Pascal GPUs are no longer actively maintained neither officialy nor yet by the community.)

monkeyboy537 · Post by **monkeyboy537** » Thu Jan 29, 2026 2:30 am

Could you post your dmesg while this happens? I'm curious on what it says

terramotu · Post by **terramotu** » Sun Feb 01, 2026 10:54 am

Hello monkeyboy537, sorry for the delay...

I don't have right now a dmesg of the crash itself. I should do more testing and try to find how to store the dmesg to make it persistant to reboot.

I have only the initialisation:

Code: Select all

$ doas dmesg -T | grep -iE "gpu|drm|amdgpu|nvidia|i915|oom|kill"
doas (terramotu@desktop-system) password:

[Wed Jan 28 09:17:35 2026] ACPI: bus type drm_connector registered
[Wed Jan 28 09:17:35 2026] ata1.00: supports DRM functions and may not be fully accessible
[Wed Jan 28 09:17:35 2026] ata1.00: supports DRM functions and may not be fully accessible

[Wed Jan 28 09:17:59 2026] input: HDA NVidia HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:03.1/0000:0b:00.1/sound/card0/input32
[Wed Jan 28 09:17:59 2026] input: HDA NVidia HDMI/DP,pcm=7 as /devices/pci0000:00/0000:00:03.1/0000:0b:00.1/sound/card0/input33
[Wed Jan 28 09:17:59 2026] input: HDA NVidia HDMI/DP,pcm=8 as /devices/pci0000:00/0000:00:03.1/0000:0b:00.1/sound/card0/input34
[Wed Jan 28 09:17:59 2026] input: HDA NVidia HDMI/DP,pcm=9 as /devices/pci0000:00/0000:00:03.1/0000:0b:00.1/sound/card0/input35

[Wed Jan 28 09:17:59 2026] nvidia: module license 'NVIDIA' taints kernel.
[Wed Jan 28 09:17:59 2026] nvidia: module license taints kernel.

[Wed Jan 28 09:18:00 2026] nvidia-nvlink: Nvlink Core is being initialized, major device number 234
[Wed Jan 28 09:18:00 2026] nvidia 0000:0b:00.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=io+mem

[Wed Jan 28 09:18:00 2026] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 580.95.05 Tue Sep 23 10:11:16 UTC 2025
[Wed Jan 28 09:18:00 2026] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 580.95.05 Tue Sep 23 09:41:17 UTC 2025

[Wed Jan 28 09:18:00 2026] [drm] [nvidia-drm] [GPU ID 0x00000b00] Loading driver
[Wed Jan 28 09:18:01 2026] [drm] Initialized nvidia-drm 0.0.0 for 0000:0b:00.0 on minor 0

[Wed Jan 28 09:18:49 2026] nvidia_uvm: module uses symbols nvUvmInterfaceDisableAccessCntr from proprietary module nvidia, inheriting taint.

SCR4MBL3 · Post by **SCR4MBL3** » Sun Feb 01, 2026 12:53 pm

Hello! Are you sure that those drivers are the ones your GTX needs? Maybe you should update.

Another way to solve the problem could be to add the nvidia-drm.fbdev=1

Post by **NeddySeagoon** » Sun Feb 01, 2026 2:34 pm

terramotu,

If its really a crash, the kernel cannot write anything anywhere, so you cannot save dmesg.

The way it fails, with things being killed sounds like a memory leak somewhere.
You should be able to spot the OOM in dmesg after the first thing disappears.
You may even have time to save dmesg.

That only tells what was killed, not where the problem is.

top, or htop, can show you the memory hog.

The community cannot maintain nvidia-drivers ever. The code is proprietary.
Maybe its time to move to the open source nouveau driver?

Downgrading is not future proof. You will need to find a way ahead sometime.

monkeyboy537 · Post by **monkeyboy537** » Sun Feb 01, 2026 3:43 pm

How much RAM and SWAP (if you have that enabled) do you have? I used to have a GTX 1060 6GB vram but that was a while ago (before the 580 driver series came out) and I never had this issue. I would try to update the Nvidia drivers as there is a newer one out on the gentoo repository.

terramotu · Post by **terramotu** » Mon Mar 30, 2026 5:02 pm

Hello, and thanks for all your answers.
I took some time to reply because I was quite clueless about how to identify, reproduce, and solve my issue (or find a workaround).

I do have new information and answers I’d like to share with you:

What drivers are you running ?

I’m running version 580, which seems to be the latest version that supports my 1080 Ti.

Code: Select all

+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.126.18             Driver Version: 580.126.18     CUDA Version: 13.0     |
+-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce GTX 1080 Ti     Off |   00000000:0B:00.0  On |                  N/A |
| 28%   54C    P0             70W /  250W |    1064MiB /  11264MiB |      2%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A            3732      G   /usr/bin/X                              608MiB |
|    0   N/A  N/A            4126      G   xfwm4                                     4MiB |
|    0   N/A  N/A            4494      G   gnome-calendar                           16MiB |
|    0   N/A  N/A            4630      G   /usr/lib64/firefox/firefox              325MiB |
|    0   N/A  N/A            9646      G   /opt/discord/Discord                     77MiB |
|    0   N/A  N/A           11547      G   alacritty                                 9MiB |
+-----------------------------------------------------------------------------------------+

Why not migrate to nouveau drivers ?
Honestly, I would have loved to, because those drivers are also a pain when it comes to Wayland. I wanted to run Sway, but I got messages saying that it’s not supported.
The reason I didn’t do it is because of what’s written on the wiki: https://wiki.gentoo.org/wiki/NVIDIA

It shows a board and says that Pascal graphics cards are stuck on version 4.6 of the driver.
I have three QHD screens, and even if it costs me in terms of gaming, I need drivers that can handle three screens. (Version 580 is not the latest, but I was confident it could handle multi-screen setups.)

How much RAM and SWAP?
I have 32 GB of RAM. I haven’t implemented ZRAM. I have two swap partitions of 40 GB each. The top-priority swap partition is on an SSD, and the second is reserved for hibernation (not implemented yet).

What games provoque those crash
World of Tanks crashes my computer during the first game I play. A month ago, it was working, and I experienced similar crashes, but they were rarer (I didn’t pay attention to what I had changed).
Raft crashes my computer after 30 minutes to 1 hour of gameplay.
CS:GO 2 crashes my computer instantly.

What did you try
I tried launching game while my other screen are turned off, I tried running them without redshift and nothing worked.
Next test will be to launch those game in "windowed" mode.

I tried monitoring with htop during the crashes, but I didn’t notice anything unusual (maybe I’m not skilled enough to detect it).
I will take a look at nvidia-drm.fbdev=1, although I think I’ve already implemented it.
I’ve thought about buying a new AMD graphics card, but I need to consider it carefully and be sure it will solve my issue before spending a lot of money on a problem that only occurs while gaming (sometimes my left screen is blinking and I can see my wallpaper behing the apps but it's not that violent).

I'm eager to read from you
Thanks a lot, everyone, for your help.