View previous topic :: View next topic |
Author |
Message |
Evilguru Guru
Joined: 16 Aug 2005 Posts: 390 Location: England
|
Posted: Thu Sep 02, 2021 2:02 am Post subject: SDDM with NVIDIA Drivers and nvidia-drm.modeset=1 |
|
|
Recently I switched my ~AMD64 system over from an integrated Intel GPU to an NVIDIA GPU. As such I emerged the nvidia-drivers and proceeded through https://wiki.gentoo.org/wiki/NVIDIA/nvidia-drivers
However, upon rebooting and trying to start sddm, I found nothing happened; apparently hitting the following bug: https://bugs.gentoo.org/697634. Working through the bugzilla thread I checked that my kernel (which is essentially a livecd config) had the required DRM options enabled, and added "nvidia-drm.modeset=1" to my command line. Rebooting, however, I found that nothing had changed. Moreover, going through dmesg:
Code: |
[ 18.710913] nvidia_drm: loading out-of-tree module taints kernel.
[ 18.710915] nvidia_drm: module license 'MIT' taints kernel.
[ 18.710941] nvidia_drm: unknown parameter 'modeset' ignored
[ 18.717511] nvidia-gpu 0000:01:00.3: enabling device (0000 -> 0002)
[ 18.775824] nvidia-nvlink: Nvlink Core is being initialized, major device number 246
[ 18.778203] nvidia 0000:01:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=none
|
with the nvidia_drm module essentially saying that it has never heard of the modeset parameter. However, according to the bugzilla thread this argument is essential. As such, I am at something of a loss with regards to how to proceed.
Regards, Freddie. |
|
Back to top |
|
|
psycho Guru
Joined: 22 Jun 2007 Posts: 534 Location: New Zealand
|
Posted: Tue Sep 07, 2021 10:35 am Post subject: |
|
|
Hi Evilguru.
That's weird: I've never seen that "unknown parameter" message before. Are you sure your GPU supports nvidia modesetting...it's not some kind of hybrid (optimus or whatever) thing that needs the intel chipset for the framebuffer? Actually the whole thing seems weird to me: I don't use sddm, but it looks to me as though it sits on x11 rather than launching it, in which case I can't see what difference kms would make: the X server should be setting its own graphics mode regardless. I'm running lightdm with nvidia_drm.modeset=0 (just a plain old VGA console) and everything works perfectly. |
|
Back to top |
|
|
Section_8 l33t
Joined: 22 May 2004 Posts: 627
|
Posted: Tue Sep 07, 2021 1:15 pm Post subject: |
|
|
This is probably way off the mark, but in my dmesg, I see:
Code: | [ +0.001164] [drm] [nvidia-drm] [GPU ID 0x00000500] Loading driver
[ +0.890966] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:05:00.0 on minor 0 |
IOW, nvidia(dash)drm instead of nvidia(underscore)drm as the OP quoted. This isn't just a typo on the kernel command parm is it? |
|
Back to top |
|
|
Ionen Developer
Joined: 06 Dec 2018 Posts: 2716
|
Posted: Tue Sep 07, 2021 1:30 pm Post subject: |
|
|
I usually use - myself, but think _ is fine too -- you could also use module params, /etc/modprobe.d/nvidia.conf has this bit: Code: | # Kernel Mode Setting (notably needed for EGLStream/Wayland)
# Enabling may possibly cause issues with SLI and Reverse PRIME.
#options nvidia-drm modeset=1 | It's # commented-out by default unless USE=wayland (some configurations need this, while others it breaks things), but you can change that manually being a config file if you need it.
What version if nvidia-drivers are you using though, I see no mention of this. The above config line may notably missing if e.g. you're using nvidia-390 or some out of tree older drivers (option does exist on 390.xx too, just not in the .conf given not sure if I can recommend it -- lot of things work badly with 390 and modern packages, notably most dual-gpu/offloading setups which forces to use a single one unless go considerably out of your way). |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|