Atomic Modeset Error -22 after latest Nvidia Update on Tumbleweed

After the latest Nvidia update on Tumbleweed I can no longer boot into X. It enters a continuous loop trying to start X and repeatedly gives the error message ‘atomic modeset error -22’. Secure boot is turned off so it shouldn’t be a driver registration issue. I did reinstall the Nvidia drivers without any change. Zypper dup reports nothing to update so all packages are current. Any suggestions? I would really like to get X back up and running. In X I am using Plasma 6 in Wayland.

A second question is how do you boot to runlevel 3 from the boot screen. A long time ago you entered init=init 3 but no variation of that worked. I also tried just the number 3 without success. It would really help in situations like this when the machine is looping trying to establish the mode lines.

Does it work if you rollback to the snapshot with the previous NVIDIA driver?

No, you have never entered it. You simply added 3 to the kernel command line as a separate word. And it still works as before.

Never tried rolling back to a previous snapshot, will give that a try. As far as init 3 goes I was talking way back. I started using SuSE back in 1998 and was refering to the lilo boot loader. Haven’t had a need for it since. I did try adding 3 to the end of the kernel command line and it still goes to the atomic modeset routine. Not much fun to get to runlevel 3 from there. I will let you know what happens.

KMS is used by Plymouth. Try adding plymouth.enable=0

is not the same as

The first is, as @arvidjaar posted, nonsense.
The last will probably still function, but is deprecated by

systemctl isolate multi-user.target

init=init 3 is from the lilo days. needed it because back then as we had to write our own modelines. init 3 didn’t work. neither did just 3.

tried the earliest snapshot available from 2024-06-02 with the same result. the nvidia update was from around then so the snapshot might have been made after the update. the error stems from nv_drm_atomic_commit (nvidia_drm). This started immediately after the nvidia update and I believe the hardware is fine as I am typing this from windows on the same machine.

I will try adding plymouth.enable=0 and advise.

@diacobel.1 what is your GPU hardware, what driver is in use? inxi -Gxxz

adding plymouth.enable=0 worked. I am typing this from linux. my gpu is a GeForce RTX 4060 Ti. the driver version is 550.78-22.1 G06. Everything was working quite well until this update. not a good time for this to happen being the first of the month.

The 550.67-20.1 is still available.

how do i get it? not visible in yast and not sure how to do it in zypper

Index of /opensuse/tumbleweed/x86_64 (nvidia.com)

In YaST2-Software look in the “versions” tab.
With zypper try:

zypper in -D --oldpackage nvidia_driver=550.67

the -D stands for “dry run”, answer two questions to resolve conflicts, when you are satisfied with the dry run result repeat the command without the -D

thanks. i will try the zypper command.

Using zypper I installed the nvidia drivers 550.67. There is no change from the 550.78 drivers. I cannot get into an X session without adding plymouth.enable=o to the kernel boot parameters. I noticed there was a massive update to a lot of the core packages and the kernel. I installed the updates allowing the Nvidia drivers to be updated back to 550.78 since reverting back to 550.68 did not solve the issue.

Same problem after the update. Cannot boot into an X session without adding the above line to the kernel parameters at boot.

Am I the only one with this issue? Online searches are quiet on this issue. You would think people would post if they could no longer boot to X.

I am open to trying any ideas you come up with.

typo. the actual code added is plymouth.enable=0, not o

@diacobel.1 all good here with the beta driver (installed the hard way)…

pinxi -Gxz
Graphics:
  Device-1: NVIDIA GM107GL [Quadro K620] vendor: Hewlett-Packard
    driver: nvidia v: 555.42.02 arch: Maxwell bus-ID: 02:00.0
  Device-2: NVIDIA TU117GLM [Quadro T400 Mobile] driver: nvidia v: 555.42.02
    arch: Turing bus-ID: 03:00.0
  Display: x11 server: X.Org v: 21.1.12 with: Xwayland v: 24.1.0 driver: X:
    loaded: nvidia gpu: nvidia,nvidia-nvswitch resolution: 1: 1920x1080~60Hz
    2: 1920x1080~60Hz 3: 1920x1080~60Hz
  API: EGL v: 1.5 drivers: nvidia,swrast platforms:
    active: gbm,x11,surfaceless,device inactive: wayland,device-2,device-3
  API: OpenGL v: 4.6.0 compat-v: 4.5 vendor: nvidia mesa v: 555.42.02
    glx-v: 1.4 direct-render: yes renderer: NVIDIA T400/PCIe/SSE2
  API: Vulkan v: 1.3.283 drivers: N/A surfaces: xcb,xlib devices: 2

I don’t use plymouth…

So this is a discrete GPU, not a laptop with dual gpus?

This is a recent desktop build with one GPU. I normally go with the stable leap version but my hardware was too new for the older kernel used in leap so I had to go with Tumbleweed.

Just waiting for leap to catch up to my hardware. There was talk of a slowroll release but it is still not an option on the website.

@diacobel.1 and there is no integrated GPU in the cpu? It hasn’t installed the open driver nvidia-open-driver-G06-signed-kmp-default?

You have the kernel boot options fbdev=1 nvidia_drm.modeset=1 set?

Never even thought about the cpu/gpu. I will check the boot options. This is as straight an install as Tumbleweed gets. I stopped tweaking years ago and just use now. This says a lot about how far Linux has come that Iam comfortable doing that.

Here are the optional boot options according to yast2-bootloader

Blockquote
splash=silent quiet security=apparmor nosimplefb=1
Blockquote

didn’t see anything different when I checked /etc/default/grub.

Where else can I look?