System freezes after waking up from sleep

[righN]

Hi, so after KDE Plasma 6 update, one strange thing started to happen. I noticed that sometimes my RAM usage for some reason is way to high, but seems like it solved itself, because I was having issues with stuttering at first. But right now, there are times when I would wake my laptop from sleep, the whole desktop just freezes after a few minutes of usage and the only solution is to restart my laptop. Journalctl shows this

Mar 11 18:33:21 righn-omen kernel: BUG: unable to handle page fault for address: 000000000002168f
Mar 11 18:33:21 righn-omen kernel: #PF: supervisor read access in kernel mode
Mar 11 18:33:21 righn-omen kernel: #PF: error_code(0x0000) - not-present page

I doubt that the RAM is faulty, because this never happened on Plasma 5. And the freezes somehow only happen after waking up from sleep.

[righN]

I have been runing Memtest86+ for about an hour now, one pass has been made. No errors yet, I’ll keep it running for the time being. Freeze happened again today, I had enabled SysRq before and clean shutdown after freeze wasn’t possible (clicking on power button one time did nothing), had to do a hard shutdown. Should I report it to Bugzilla or does anyone else has any other suggestions?

[righN]

If anyone else is experiencing similar issues, here’s your answer - https://forums.developer.nvidia.com/t/series-550-freezes-laptop/284772

[Atomic-no-80]

Thank you for the link. I’m just trying to identify whether it is a kernel issue or nvidia issue

To solve this i tried this:

i donwngraded these 3 packages :

• nvidia-dkms
• nvidia-utils
• lib32-nvidia-utils

Version downgraded : 550.54.14 → 545.29.06

But seems like linux 6.8.1.arch1-1 is not working with nvidia 545.29.06

The following logs are being generated :

Mar 22 00:39:14 endeavour systemd-modules-load[222]: Failed to find module 'nvidia-uvm'
Mar 22 00:39:19 endeavour systemd-modules-load[469]: Failed to find module 'nvidia-uvm'
Mar 22 00:39:19 endeavour nvidia-persistenced[674]: Failed to query NVIDIA devices. Please ensure that the NVIDIA device files (/dev/nvidia*) exist, and that user 143 has read and write permissions for those fil>
Mar 22 00:39:19 endeavour systemd[1]: Failed to start NVIDIA Persistence Daemon.

What worked : Latest lts kernel is working fine with 545. This is my current working setup

I want to ask you 3 questions ?

• Are you using external monitor
• Whenever you saw these error, did you plugged or unplugged your device
• What did you do to solve this ?

[Atomic-no-80]

For someone who might know what is happening, these were my issues :

System Info : ASUS TUF Gaming A15 FA507RE_FA577RE

• CPU: AMD Ryzen 7 6800H with Radeon Graphics (16) @ 4.785GHz
• CPU: AMD Ryzen 7 6800H with Radeon Graphics (16) @ 4.785GHz

1. A lot of times, system doesn’t restart/shutdown normally (I’ve to force shutdown it), it doesn’t happen everytime

2. During shutdown/restart these errors occurs and nothing happened

3. Problems with waking up from sleep (Hence i decided to manually block sleep and screen locking)

[righN]

1. I am, but the issue happened with the external monitor disconnected too.
2. There were times when it happened with it plugged in and unplugged, so I don’t think it matters.
3. Downgraded to kernel 6.7.9 and 545.29.06 driver. But as far as I know, you can downgrade to 535 version with 6.8 kernel. Or either use nvidia-all. But in my case, I’ll just wait until NVIDIA fixes their mess.

[Atomic-no-80]

I’ll stick to lts-kernel with 545 for now, until it causes any problems.

This topic might be closed till nvidia fixes it, if you see the issue fixed please leave a message

[Atomic-no-80]

I downgraded to 6.7.9 for backup kernel

Are you sure you’re using this combination : kernel 6.7.9.arch1-1 and nvidia 545.29.06-1. Because i’m still getting these error in my journal on this kernel:

Mar 22 02:27:38 endeavour systemd-modules-load[221]: Failed to find module 'nvidia-uvm'
Mar 22 02:27:42 endeavour systemd-modules-load[481]: Failed to find module 'nvidia-uvm'
Mar 22 02:27:43 endeavour nvidia-persistenced[731]: Failed to query NVIDIA devices. Please ensure that the NVIDIA device files (/dev/nvidia*) exist, and that user 143 has read and write permissions for those files.
Mar 22 02:27:44 endeavour systemd[1]: Failed to start NVIDIA Persistence Daemon.

While rebooting into lts kernel doesn’t have them

[righN]

I definitely am on 6.7.9.arch1-1, but on nvidia 545.29.06-4. But I don’t think your issue and mine are related to be honest. I checked journalctl a few times after those kernel panics/faults happened and I didn’t get any of those errors you’re getting. And a lot of people in the mentioned NVIDIA post don’t getting anything related to NVIDIA either in the logs, but are still experiencing the same issue as me.

[Atomic-no-80]

The errors that i gave you on previous reply were occuring after the downgrading nvidia to 545 (with both linux 6.7.9 and 6.8). That’s why i asked

On 550 i was getting these errors :

Mar 22 00:09:20 endeavour kernel: BUG: unable to handle page fault for address: 0000000000002328
Mar 22 00:09:20 endeavour kernel: #PF: supervisor read access in kernel mode
Mar 22 00:09:20 endeavour kernel: #PF: error_code(0x0000) - not-present page

And i actually searched the first error on this forum to reach this post

[righN]

Maybe for some reason you need to rebuild kernel image sudo reinstall-kernels? Because I was on cachyos kernel when doing the downgrade. But Arch kernel was kept as a backup and that’s the one I downgraded. So maybe if you did the downgrade of NVIDIA drivers while on 6.8 kernel, it caused issues, as 545 is not supported on 6.8? It’s just an idea, but can’t say more, as I hadn’t experienced the same errors as you do.