Hard freezes

[matthew]

Hi everyone, I’ve installed Endeavouros on my Lenovo Legion 5 Pro laptop a week ago. Since then I had around 3 hard freezes. First this happened randomly when watching a YouTube video on Firefox, second and third times it happened when i tried to restart my system using the KDE interface, so that’s (systemctl reboot) as far as i know. Now what happens is that the system is in a weird state, I could open konsole but the shell literally didn’t respond, the shell froze, I couldn’t even run echo and I had to use the power button. Looking at the logs after the boot it seems that there was a Null pointer de-reference in the kernel. I’m not really sure what the problem. Not sure if it’s related, but today I’ve tried to install postman-bin from the AUR and it constantly froze on ‘Copying source files needed for debug symbols’ with no CPU usage or signs of life, after which I decided to restart my system and it hard froze.

Please do note, I’ve been using windows 11 on my laptop for a few months after buying and it worked fine so this is not a hardware failure, at least I assume.

My laptop’s specs are:
i7 13700HX, RTX 4070 Mobile, 32GB of RAM.

uname -a
Linux ******* 6.6.24-1-lts #1 SMP PREEMPT_DYNAMIC Wed, 03 Apr 2024 14:50:23 +0000 x86_64 GNU/Linux

[BluishHumility]

@matthew

RTX 4070 Mobile

Take a look at this thread:

Issues with nvidia 550 drivers

There seem to be some issues with the current nvidia drivers causing crashing and hard freezes in certain circumstances. Notably, it can cause a freeze during the update process leaving the system in a broken state. I believe we have seen a few people have this issue on the forum recently though I don’t think anyone attributed it nvidia. There are a couple of references to the issue below though they are a bit hard to follow. It is something to be aware of if you are using these drivers.

It is possible the hard freezes you are experiencing are related to the kernel memory corruption issue introduced by the 550 drivers.

[matthew]

Thanks for pointing that out. I’m on driver version 550.67 running the DKMS version. What can I do to fix this issue? Can i rollback somehow? Or what shall I currently do?

[lavavex]

how long till next driver series…

[BluishHumility]

@matthew

Can i rollback somehow?

There are ways of reverting to a previous version of the drivers. For example, you can use the downgrade utility if you have the previous version in your package cache.

Help! Latest update broke Optimus+nVidia

I was able to get the 525xx drivers back but only in the 6.1 LTS kernel using: sudo downgrade nvidia-dkms nvidia-utils lib32-nvidia-utils However, if I boot into the 6.5 kernel and run the same command, it still only changes it for the 6.1 LTS kernel, not the 6.5. How can I sudo downgrade to 525 in the 6.5 kernel?

Or you can uninstall the packages providing the 550 version and directly install the previous version from the AUR.

Downgrade Nvidia driver

For example: sudo pacman -R nvidia-dkms nvidia-settings yay -S nvidia-470xx-dkms nvidia-470xx-settings There’s not a specific how-to, because it’s the same process as any package management.

Or you can try using nvidia-all to manage the driver versions.

GitHub - Frogging-Family/nvidia-all: Nvidia driver latest to 396 series AIO...

Other possibilities to consider would be trying different kernels (for example the LTS kernel) which may produce a different result, or try uninstalling the Nvidia drivers altogether and use the Nouveau driver instead if your use case is not overly demanding.

@lavavex

how long till next driver series

I have no idea, I do not follow Nvidia development very closely. I don’t own any Nvidia hardware, and I probably never will because it seems stressful to deal with.

[Melways]

Like @BluishHumility said you can downgrade.

I think the downgrade command works like this:

sudo downgrade nvidia

It will generate a list of driver versions. Just choose a 545 version and hit enter.

Lets hope nvidia fix it soon.

Edit: See @dalto post below

[dalto]

@BluishHumility

you can use the downgrade utility if you have the previous version in your package cache.

@Melways

I think the downgrade command also works with the nvidia driver?

Be careful since the older versions aren’t compatible with Linux 6.8.

If you want to run the older driver either use the lts kernel or nvidia-all as suggested above.