Precisely the example about which you make sarcasm is one of the errors that we seek to avoid. The bug reported by friend Maumeloni is a bug introduced in Linux 6.10 and is not related to X, Y or Z version of the nvidia driver. Someone with less knowledge may be confused and think that there is a relationship. That’s why it’s healthier to open a new thread properly to help better troubleshooting.
Indeed. Thanks for the tip @mario156090, and apologies to everyone for adding noise.
Having downgraded to 6.1.0-23-amd64, I have no weird stacktraces in dmesg nor any crashes so far :pray: I’ll track the linux/rwsem.h issue in lkml.
As for the issue at hand, it looks reasonably solved to me, though the more users confirming it works, the better.
Can anyone verify whether 560.35 drivers solve this issue on Arch Linux? I will be switching back to Arch if the issue has been resolved for 560 as well.
i used 560.31 closed and now back to open with 560.35.03 and in terms of freezing and kernel panics i haven’t seen any issues.
I am using EndeavourOS(for months now) and the packages are the ones from the arch and cachyOS repos(e.g.: nvidia, nvidia-open, nvidia-open-dkms).
Works fine now.
The open drivers never had this issue, because these are completely different from the normal ones.
Since yesterday on my Arch Linux (EndeavourOS) I use 560.35.03 with the 6.10.6-zen1-1.1-zen kernel.
It’s to short to say if it’s finally fixed but until now these random crashes didn’t occur.
I could use standby and resume or other stuff without problems and this crash only happened randomly while doing these things or just working at my computer.
Without using nvidia at all (iGPU)…
Hopefully it’s really fixed now and I won’t experience a crash in the next days.
Sorry forgot to add that i used 560.31 without issues for a week.
I only just reverted to open because they feel pretty much as functional anyways and as well since they are gonna be the default(for pre-turing and above however).
Until now I also hat no crashes the the closed source driver (GTX 940MX Maxwell). :)
For what it’s worth, I did observe again the kernel panic issue with the 550.107.02 drivers, which are the latest available on OpenSUSE Tumbleweed, using KDE Plasma as the desktop environment and an RTX3090 as a secondary discrete GPU (primary output on the iGPU). In my case, it seems to occur almost always (albeit randomly and rarely) when the GPU is idling immediately following certain CUDA compute workloads, although never during.
Unfortunately I cannot easily install the 560 drivers, although I might eventually consider this if it gets unbearable.
It doesn’t help that obtaining logging information during these conditions seems very difficult (the system freezes, nothing gets written on system logs).
Or use the “open” drivers.
(Which are not really open. The code is now in the firmware blob.)
Hi everyone,
I’m really a linux newbie, sorry. If anyone could show me step-by-step how to solve this problem, I’ll thanks a lot.
I have a Dell laptop G15 5530, Intel i7-13650HX, RTX 4050 running both Win11 and Ubuntu.
I tryid others solutions, but I haven’t solved yet.
Thks eveone if yoy could help me. ;)
Initially, when I installed CachyOS, using the ‘nvidia.modeset=0’ kernel parameter during installation and in GRUB for the installed system helped resolve the freezing. I also removed all NVIDIA drivers. However, after some time, the freezes started occurring again, becoming more and more frequent. Eventually, the problem returned to its original state.
After this, I tried installing MX Linux (MX-23.3_x64 “ahs”). I chose to run the installation using AMD video drivers. Since then, I haven’t experienced any freezing issues. I’ve even updated the system, including the kernel, and everything continues to work normally.
While MX Linux may not be as customizable as Arch-based distributions, it’s a stable solution for my current situation. I still haven’t fully understood the root cause of the problem with other distributions.
Now I had a kernel crash on resume from standby over night.
The SHIFT KEY was blinking and I just saw a black screen and my mouse pointer.
Edit:
And it happend again after standby.
It seems it happens since upgrading the nvidia-dkms driver from 560.35.03-2 to 560.35.03-3
See upgpkg: 560.35.03-3: Use egl-gbm package instead of included binary blobs (61a16ad9) · Commits · Arch Linux / Packaging / Packages / nvidia-utils · GitLab
It uses now the egl-gbm package instead of the included binaries.
https://archlinux.org/packages/extra/x86_64/egl-gbm/
I have a supposition this or similar issue re-appeared in arch version of proprietary 560.35.03 driver with NVreg_EnableGpuFirmware=0
I am trying to debug it now. Not sure if it is the same issue. Symptoms seems similar. Had three kernel panics today. Haven’t had them for week till 560.35.03.
I found the issue on my system.
The kernel 6.10.10-zen has a standby/resume bug.
I downgraded to 6.10.9-zen and it works again.
See System restarts when waking it up from sleep. · Issue #359 · zen-kernel/zen-kernel · GitHub