NVIDIA driver affecting Chrome, Slack apps

English Hardware
21

Removed and installed mentioned packages. Thanks for confirming about reboot. I had already done it after earlier response to be totally sure.

Outputs pasted here due to character limit.

I wonder why anything related to nouveau exists at all:

-rw-r--r--   1 root     root      1529207 Apr 17 11:10 usr/lib/modules/6.15.0-1-default/kernel/drivers/gpu/drm/nouveau/nouveau.ko.zst
-rw-r--r--   1 root     root        35112 Apr 17 11:10 usr/lib/modules/6.15.0-1-default/kernel/drivers/hid/hid-nvidia-shield.ko.zst
22

@dharmit you also have some bbswitch stuff there as well usr/lib/modprobe.d/09-nvidia-modprobe-bbswitch-G04.conf and modprobe-pm-G05.conf all old stuff?

Can you please check if anything bumblebee or bbswitch is installed and remove.

Once it’s cleaned up rebuild initrd with dracut -f --regenerate-all reboot and check that lsinitrd output again to check the nouveau and other stuff is gone.

Also can you check which Nvidia kmp is installed with zypper se -si kmp

1 Like
23

Which DE are you using? I run GNOME here, and I think I’m seeing the same issue that you are. From the overview view, it looks like you are also using GNOME.

I have found that restarting gnome-shell (alt+f2 and then ‘r’) temporarily resolves it.

24

@hendersj restart is not available in Wayland… :wink: only X11

Likewise graphics is controlled via the Intel GPU using modesetting and no nvidia there, it should be;

Display: wayland server: X.org v: 1.21.1.15 with: Xwayland v: 24.1.6
    compositor: gnome-shell v: 48.2 driver: X: loaded: modesetting,nvidia

@dharmit’s output is;

Display: wayland server: X.org v: 1.21.1.15 with: Xwayland v: 24.1.6
    compositor: gnome-shell driver: X: loaded: modesetting

Also as root user can you update inxi with inxi -U

Edit, I would also suggest adding your user to the video and renderer groups, you need to logout/login.

25

After searching around it looks like it might be a electron app/gpu acceleration thing … maybe try chrome --disable-gpu-vsync ?

26

@dart364 I doubt it, I run google-chrome on three systems with intel/nvidia setups, on my primary desktop slack. No issues… It’s something tied in with vulkan and lack of nvidia in the output from what I can see…

27

Ah, right. Similar issue, but I’m on X11, not Wayland.

So it does seem that the issue is present on both servers. In my case, I’m just nvidia (3090ti), no Intel at all here.

If it’s helpful, I’m on 20250604, prior to that update, this issue wasn’t present. I see OP is on 20250406.

I can actually force it to happen - if that’s useful info (and my issue appears to be the same to you), I can provide some details.

28

@malcolmlewis hopefully I’m using the right command to find the packages installed from the long disabled bumblebee repo (z is aliased to sudo zypper):

$ z --plus-content home_Bumblebee-Project_nVidia_latest se -i -r home_Bumblebee-Project_nVidia_latest
Temporarily enabling repository 'Downloader and installer for the nVidia driver package (latest) (openSUSE_Tumbleweed)'. [--plus-content]
Looking for gpg keys in repository Downloader and installer for the nVidia driver package (latest) (openSUSE_Tumbleweed).
  gpgkey=https://download.opensuse.org/repositories/home:/Bumblebee-Project:/nVidia:/latest/openSUSE_Tumbleweed/repodata/repomd.xml.key
Building repository 'Downloader and installer for the nVidia driver package (latest) (openSUSE_Tumbleweed)' cache ..................................................[done]
Loading repository data...
Reading installed packages...
No matching items found.

If it’s correct, there’s nothing else. For the two conf files you mentioned, I removed the suse-prime package. I don’t remember using it anyway.

$ z se -si kmp
Loading repository data...
Reading installed packages...

S  | Name                          | Type    | Version                                  | Arch   | Repository
---+-------------------------------+---------+------------------------------------------+--------+------------------
i+ | nvidia-driver-G06-kmp-default | package | 570.153.02_k6.14.6_1-36.1                | x86_64 | repo-non-free
i+ | nvidia-driver-G06-kmp-default | package | 570.153.02_k6.14.6_1-36.1                | x86_64 | nvidia
i  | ovpn-dco-kmp-default          | package | 0.2.20241216~git0.a08b2fd_k6.14.6_2-2.41 | x86_64 | (System Packages)
i  | ovpn-dco-kmp-default          | package | 0.2.20241216~git0.a08b2fd_k6.14.6_1-2.38 | x86_64 | (System Packages)
i  | ovpn-dco-kmp-default          | package | 0.2.20241216~git0.a08b2fd_k6.14.5_1-2.35 | x86_64 | (System Packages)
i  | ovpn-dco-kmp-default          | package | 0.2.20241216~git0.a08b2fd_k6.14.4_1-2.27 | x86_64 | (System Packages)
i  | ovpn-dco-kmp-default          | package | 0.2.20241216~git0.a08b2fd_k6.14.3_1-2.26 | x86_64 | (System Packages)
i  | ovpn-dco-kmp-default          | package | 0.2.20241216~git0.a08b2fd_k6.15.0_1-3.5  | x86_64 | repo-oss

Isn’t repo-non-free and nvidia the same thing? Why is zypper listing it twice? :thinking:

Regenerating the initrd with dracut command you provided helped solve the issue.

A request: please give me a day to mark this issue as solved because in the past couple of days, booting into different kernel versions has “sometimes” (not reliably) helped me start Slack. I’ll try again a few times tomorrow (already checked twice so far, and it works) and mark the issue solved if Chrome and Slack open up just fine each time. :crossed_fingers:

29

Because you added the nvidia repo manually. It is repository number 6. As it is a duplicate of repo number 1 (which is managed by a service) you can safely remove repo 6.

30

Sharing the relevant output after latest changes:

  Display: wayland server: X.org v: 1.21.1.15 with: Xwayland v: 24.1.6
    compositor: gnome-shell driver: X: loaded: modesetting unloaded: vesa
    alternate: fbdev,intel dri: iris gpu: i915,nvidia,nvidia-nvswitch
    display-ID: 0

I tried a few flags like below before opening issue here, but it didn’t help. However, I didn’t use the exact flag that you’ve shared:

$ google-chrome --disable-gpu --disable-gpu-compositing

I have a (bad?) habit of running z dup almost everyday. Last week’s mirror issue bite me real bad. E.g., I started facing mouse freeze with Slack. Especially Slack. And had to restart it for things to work. And Firefox downgrade to 137 made things a nightmare too. But I think things are resolved on that front.

Thank you, @hui. Removed.

31

@malcolmlewis thanks a lot for your help! Can’t thank you enough. :pray:
I had missed doing inxi -U and adding my user to audio and video groups. Have done that now. Can you help understand why that (adding to groups) would be helpful in my particular case?

@hui thanks for the tips. :+1:

32

@dharmit Hi for the intel gpu to perform with the level-zero libraries and Nvidia also takes advantage of being in the video group.

You can check with nvidia-smi for nvidia, but nvtop gives a better overview of graphics and compute in use for both gpus.

Also ensure switcherooctl is installed and enabled SDB:NVIDIA Switcheroo Control - openSUSE Wiki to take advantage of right-click → launch with discrete gpu.

1 Like
33

I think adding my user to audio and video groups (inclined more towards video) has helped fix another longstanding issue. When I was on Fedora, I would power on the laptop and close the lid because the external monitor is my main and only screen. After login, I would land on the desktop without any issues.

With Tumbleweed, it was a lucky draw where I used to be mostly unlucky and had to force power off the system. I learned to close the lid only after desktop had loaded. But since fixing the issue on this thread and following the recommendations, I have won the TW lucky draw 100% of the times. :crossed_fingers: :nazar_amulet: