My initrd also includes “usr/lib/modprobe.d/50-nvidia-default.conf”, but it seems to be ignored! I recognized that the last line of “50-nvidia-default.conf” (“install nvidia PATH=…”) does not contain a “LF” at the end, but that probably is not the issue!
For anyone on the non production drivers (having installed them the hard way), fbdev does not work, making modesetting with simpledrm buggy, which then makes wayland (specifically KDE) not working as expected. Either switch to X11 or make 6.10 default for now.
As above, we have a means to make this work fine with Wayland on 6.11 already - enabling kernel mode setting features of the driver with the appropriate parameter.
The issue here is that it should already work for everybody because this file exists and specifies that parameter is set already.
The purpose of this thread is now to figure out WHY that fails, so that we can make sure this is actually fixed.
Since we’re all stumped, I’m taking an educated guess and filing this as a bug with openSUSE’s nvidia driver packages:
https://bugzilla.opensuse.org/show_bug.cgi?id=1231067
Let’s hope we can fix this before it rears its head for the next person! 
FYI Stefan Dirsch (the excellent and very busy maintainer of the SUSE nvidia driver packages) has replied to the above issue and it has been closed as fixed upstream.
The TL;DR here is that this was never meant to work, because of the driver bug we know about (even though it has a non-matching description of symptoms, that’s our bug) It should have been broken for everybody, and required modification in order for it to work.
Why our workaround works, and why some people didn’t need one, is still a mystery.
At least Optimus hybrid systems like the one I’m writing from use the framebuffer on the integrated graphics, so the Nvidia bug has no effect:
bruno@LT-B:~> dmesg |grep -E "fbdev|framebuffer"
[ 0.309279] [ T1] [drm] Initialized simpledrm 1.0.0 for simple-framebuffer.0 on minor 0
[ 0.316780] [ T1] simple-framebuffer simple-framebuffer.0: [drm] fb0: simpledrmdrmfb frame buffer device
[ 6.794982] [ T1456] nvidia_drm: unknown parameter 'fbdev' ignored
bruno@LT-B:~>
1 Like
Thanks mate, I figured that the optimus/prime type setups likely worked around it ‘in hardware’.
I get this twice, so the one in my command line is ignored and also the one in my conf in initrd.
Still weird though, I mean, optimus aside, it should have broken every nvidia gpu user. There ought to be thousands of us not working. There are still no posts on reddit or nvidia forums, this is it, the few of us here. Weird right?
Not so many using Plasma over Wayland with other distros?
So I encountered the buggy/not working wayland with NVidia 560.35.03 and 6.11 kernel driver and put in nvidia_drm.modeset=1 nvidia_drm.fbdev=1 and that allowed me to run Wayland but my games are stuttering unlike before, actually its not stuttering, I get whole system pauses for 5-10 seconds.
I read on the nvidia developer forum about excessive vram issue with kde and resizing windows but I’m not sure if it’s the exact same thing. Particularly I’m testing it while playing world of warcraft. This post has a temporary fix but I can’t seem to load this page 560 release feedback & discussion - #477 by JamesJones - Linux - NVIDIA Developer Forums atm because I get a timeline error. That was linked from 560 release feedback & discussion - #477 by JamesJones - Linux - NVIDIA Developer Forums
Sorry it took me a while to get back to this thread. Just to clarify, I’m running KDE with X11.
The file /usr/lib/modprobe.d/50-nvidia-default.conf already has the options set for “nvidia_drm.modeset=1 nvidia_drm.fbdev=1”. However I also tried adding those to the kernel line on boot and that did not work for me. I am still seeing the same error that in the xorg log that I posted in post #2.
Since the 6.10 kernel still works fine for me I am content to just wait until the fix is released to hopefully resolve this issue.
Well to expand I put in that fix after finally getting the discussion to load on the link (which I didn’t link correctly in my post) Xwayland VRAM usage is still excessive when resizing X11 apps under wayland. · Issue #126 · NVIDIA/egl-wayland · GitHub and putting in the temporary fix that was recommended I haven’t had any more problems with the system pausing. Still not sure if it’s going to happen again but I wanted to share that link. Worth implementing just to fix the resizing memory allocation problem in kwin.
As far as other people not getting it working at all, I don’t know what to tell you. Not sure what driver version you’re running @Maverick1024 but with 560.35.03 it is working for me. It booted to sddm before I put those options in and as soon as I tried Wayland I got graphical errors and couldn’t see my cursor. I could use shortcuts to restart and logging into X11 worked so it may not be the same problem you are having.
I am running driver version 550.107.02-26.1. I just install from the repos instead of using the run file from Nvidia.
This is consistent with theory, since nvidia_drm.fbdev=1 doesn’t work with the 550.107 driver and 6.11 kernel, so you are apparently left without a framebuffer device to work with.
The puzzle is how comes that other users were successful adding that option to the boot command line, where in theory it shouldn’t work?
1 Like
You don’t need to edit grub much, so for a very few it may be because they forgot to rebuild Grub after saving and exiting:
Rebuild Grub:
sudo grub2-mkconfig -o /boot/grub2/grub.cfg
For the rest it is indeed a puzzle. It is surprising that it works for some of us. Maybe someone will figure it out…
@snoopy YaST Bootloader or from command line in a tty or terminal yast bootloader no editor required, and is saves the changes for you…
2 Likes
I didn’t know about the terminal option. That is cool 
@snoopy Yes, there is always an ncurses alternative when for example working remotely via say ssh, or stuck in a tty with no desktop 
I now removed the grub entries “nvidia_drm.modeset=1” and “nvidia_drm.fbdev=1”
and removed “fbdev=1” from “50-nvidia-default.conf” (as Stefan Dirsch suggested) and did “grub2-mkconfig -o /boot/grub2/grub.cfg” and “dracut --force”, but no luck, didn’t work. I then added just “nvidia_drm.modeset=1”, but NOT “nvidia_drm.fbdev=1” to grub and that did work. So it seems to be that “options nvidia-drm modeset=1” from “50-nvidia-default.conf” is ignored and the presence of “nvidia_drm.fbdev=1” does not matter (on my system)!
https://old.reddit.com/r/openSUSE/comments/1fsasx0/no_desktop_after_updating_to_kernel_611_snapshot/
Stefan’s theory that users are yet to update to 6.11, rings true. This is still spreading. Good that the workaround is widely known.
Am I right in guessing that most of the people that the workaround works for is running Wayland and not X11?
Just trying to figure out why I’m still having issues with X11…