System freeze but no crash since nvidia-open 575.x

Biscuit · 2025-07-07 23:03:31

Hey all,

Since the most recent Arch system update that required manual intervention - alongside the NVIDIA driver upgrade from `570.x -> 575.x` - I’ve been experiencing serious stability issues on my personal machine.

My system doesn't crash outright, but it completely freezes, becoming entirely unresponsive until I manually reset it. I feel this usually happens when dealing with Vulkan shaders, like running a game or while precompiling, but it sometimes does just happen on the desktop for seemingly no reason at all. And it doesn't appear to leave any logs of the issue.

It happens often enough that gaming and even general usage has become impossible. I initially assumed this might be some temporary issue with the new update and drivers, but after a week or two of waiting, there doesn't seem to have been an improvement. I am starting to wonder if it is specific to my system.

Here's what I've tried so far:
- Switched to `nvidia-open-dkms`, which seems to have slightly improved stability.
- Checked for broken libraries, zero-length .so files, and broken symlinks - I've repaired everything I could find.
- Attempted to downgrade the NVIDIA driver, but older versions are incompatible with the current kernel and libraries, which resulted in a failure to boot. I reinstalled the latest driver after that.

I have used RHEL-based system for years in a work context, but this is my first personal Linux machine for home hobbies and gaming, so I am still learning how to do this all on my own.

System deatils:
- Arch Linux x68_64
- Linux 6.15.5-arch1-1
- NVIDIA RTX 5080
- AMD Ryzen 7 9800X3D

I’d really appreciate any advice or pointers on how to further investigate or work around this issue. I haven’t been able to find other users reporting the same problem, so I’d love to know if it’s a known issue with the 50-series, the open driver, or something on my end.

Thank you for your help!

Last edited by Biscuit (2025-07-10 08:28:13)

euromatlox · 2025-07-08 10:55:09

Have you googled ?: RTX 5080 freeze
If you have another video card (less power hungry Nvidia card), testing with it can be beneficial (the problem might be RTX 5080).
For relatively new cards a future driver update might solve the problem.
Of course particularly for power hungry video cards PSU wattage (and possibly quality) is a factor too.

I've not had any problems with my Nvidia Quadro K620, but it's an old workhorse and does not need much 'juice' (power).
(I also have an old Xeon, much slower than the Ryzen of yours.)

Last edited by euromatlox (2025-07-08 11:01:47)

Biscuit · 2025-07-10 08:37:20

I have Googled “RTX 5080 freeze”, though none of what I found has been particularly helpful.

Sadly, I do not have another GPU to test with. I won't discount the idea that it may be the card or the power supply, but the fact that these issues arose immediately after a driver update is what leaves me suspicious. I have been using the machine for months beforehand without any issues at all, and I run the GPU pretty hard. So I am fairly certain that prior to the update, the PSU could handle the GPUs power demands.

I would try downgrading, but I think I need to downgrade the kernel along with it and I just don't know how to do that. And I don't want to pin myself to an older version either.

I will.. try test the PSU. Thanks

seth · 2025-07-10 20:43:28

Hybrid system, amdgpu + nvidia. Apparently no outputs connected to the AMD GPU, so reverse prime might become an issue.
kwin_wayland - do you get the same problem w/ plasma on X11?

qdbus org.kde.KWin /KWin supportInformation

and for X11 please post your Xorg log, https://wiki.archlinux.org/title/Xorg#General

until I manually reset it

What does that mean exactly? Do you hold the power button?
The posted journal runs from 18:22 - 20:00 - there's only one kwin_wayland and sddm process.
Would that time frame cover the freeze?

Since you're apparently not using it, can you disable the Raphael APU in the UEFI?

Biscuit · 2025-07-12 00:04:41

seth wrote:

Since you're apparently not using it, can you disable the Raphael APU in the UEFI?

Disabling the iGPU seems to have solved the issue! It hasn't frozen/crashed once since you suggested it.

That must have been introduced with the latest drivers? Should I remove my AMD drivers then and leave it disabled? It's not like I use the integrated graphics anyway..

seth wrote:

What does that mean exactly? Do you hold the power button?

Yes. I meant holding the power button.

seth wrote:

The posted journal runs from 18:22 - 20:00 - there's only one kwin_wayland and sddm process.
Would that time frame cover the freeze?

Yes, I believe that was one session.

Thank you SO much seth. This is the most progress I've made in weeks.

seth · 2025-07-12 08:21:06

Whether or not you've any drivers installed is irrelevant if the device isn't active at all anyway.
The problem is likely around https://wiki.archlinux.org/title/PRIME# … ronization
If you want to investigate the situation, we'll need to look a bit at how the malfunctioning system is operating (kwin support information and ideally X11 situation and log, because <rant about wayland compositors all having shitty useless logging here>

Yes. I meant holding the power button.

That will drop at least the tail of that boots journal (nothing can get sync'd to disk), avoid it at all costs.
Frenetically press ctrl+alt+del or activate and use https://wiki.archlinux.org/title/Keyboa … el_(SysRq) to reboot
Also http://0x0.st/8GfV.txt is from a running boot, that's unlike the frozen one then?
You might even be able to just ssh in.

Biscuit · 2025-07-25 19:09:12

Apologies for the delay, Things have been very busy!

Sadly, I cannot just SSH in after this happens, it locks up entirely. I will re-enable the igpu and try to get some logs! From what I remember I wasn't able to get to any other screen, but I'll try what you linked.

I do want to diagnose this because now with my igpu disabled I am having issues with VLC and remote streaming. So I really appreciate the help

vikigenius · 2025-08-05 15:58:40

I have the exact same issue. And I have a similar stack as well: AMD Ryzen 7 9800X3D but I have a 5070ti.

I tried void linux as well with the 570 drivers (that's the stable one there) and Arch with the 575.x

I don't seem to have good way to preserve the logs either since the entire system locks up. Although I do sometimes see this on boot.

[ 18.161468] [vkps] Update[907]: segfault at 7fca4ea0e4d0 ip 00007fca4ea0e4d0 sp 00007fca39bf8390 error 14 likely on CPU 2 (core 2, socket 0)
[ 18.161475] Code: Unable to access opcode bytes at 0x7fca4ea0e4a6.

And the vkps thing also came up with some vulkan shader issue with nvidia https://github.com/ggml-org/llama.cpp/issues/10528

I regret getting an Nvidia card now, should have gone with AMD :sobs. What's worse is this doesn't happen every time either.

Last edited by vikigenius (2025-08-05 15:59:41)

seth · 2025-08-05 20:53:36

Biscuit wrote:

Disabling the iGPU seems to have solved the issue! It hasn't frozen/crashed once since you suggested it.

Same as well?

Arch Linux

#1 2025-07-07 23:03:31

System freeze but no crash since nvidia-open 575.x

#2 2025-07-08 10:55:09

Re: System freeze but no crash since nvidia-open 575.x

#3 2025-07-10 08:37:20

Re: System freeze but no crash since nvidia-open 575.x

#4 2025-07-10 20:43:28

Re: System freeze but no crash since nvidia-open 575.x

#5 2025-07-12 00:04:41

Re: System freeze but no crash since nvidia-open 575.x

#6 2025-07-12 08:21:06

Re: System freeze but no crash since nvidia-open 575.x

#7 2025-07-25 19:09:12

Re: System freeze but no crash since nvidia-open 575.x

#8 2025-08-05 15:58:40

Re: System freeze but no crash since nvidia-open 575.x

#9 2025-08-05 20:53:36

Re: System freeze but no crash since nvidia-open 575.x

Board footer