You are not logged in.
I have two graphics cards in my computer, one AMD and one Nvidia, each connected to their own screen. Nouveau is giving me no issue, but when I install the proprietary ones, the nvidia-connected monitor freezes before the TTY login show up.
I'm getting two error codes in journalctl, and one in dmesg.
Journalctl:
systemd-udevd[270]: nvidia: Process '/usr/bin/bash -c '/usr/bin/mknod -Z -m 666 /dev/nvidiactl c $(grep nvidia-frontend /proc/devices | cut -d \ -f 1) 255'' failed with exit code 1.
systemd-udevd[281]: nvidia: Process '/usr/bin/bash -c 'for i in $(cat /proc/driver/nvidia/gpus/*/information | grep Minor | cut -d \ -f 4); do /usr/bin/mknod -Z -m 666 /dev/nvidia${i} c $(grep nvidia-frontend /proc/devices | cut -d \ -f 1) ${i}; done'' failed with exit code 1.
Dmesg:
[ 3.700881] platform regulatory.0: Direct firmware load for regulatory.db failed with error -2
The device shows up in lspci and
nvidia 45395968 35 nvidia_uvm,nvidia_modeset
shows up in lsmod so I assume the driver is loaded? I've tried with both the standard kernel and with zen (with their respective driver-version) but it's the same issue. It's a 1070 so it should be supported in the non-legacy nvidia package.
Does anyone know that the issue might be?
Last edited by Horses eat meat (2022-09-25 01:27:22)
Offline
Do you use EFI boot or legacy BIOS boot?
Isn't this the same issue: https://bbs.archlinux.org/viewtopic.php?id=278351 ?
systemd-udevd[270]: nvidia: Process '/usr/bin/bash -c '/usr/bin/mknod -Z -m 666 /dev/nvidiactl c $(grep nvidia-frontend /proc/devices | cut -d \ -f 1) 255'' failed with exit code 1.
systemd-udevd[281]: nvidia: Process '/usr/bin/bash -c 'for i in $(cat /proc/driver/nvidia/gpus/*/information | grep Minor | cut -d \ -f 4); do /usr/bin/mknod -Z -m 666 /dev/nvidia${i} c $(grep nvidia-frontend /proc/devices | cut -d \ -f 1) ${i}; done'' failed
These messages aren't related to the issue.
Offline
I use UEFI and systemd-boot. Using the kernel parameter nomodeset works as expected, nvidia-drm.modeset=0 doesn't do anything, but nvidia-drm.modeset=1 freezes like it usually does, but when I launch sway that screen starts to work (although without a mouse cursor) while my other screen dies.
Offline
Do you assign different VTs to each GPU?
Eg. "fbcon=map:01", see https://raw.githubusercontent.com/torva … /fbcon.rst
Online
Do you assign different VTs to each GPU?
Eg. "fbcon=map:01", see https://raw.githubusercontent.com/torva … /fbcon.rst
No, I haven't touched that at all. Without the driver, amdgpudrmfb and nouveau's equivalent shows up under /proc/fb, but with the proprietary drivers there is just AMD. Forcing it to be only mapped to 1 did just turn off the one good screen I have.
Enabling nvidia-drm.modeset doesn't change anything with the /proc/fb.
Last edited by Horses eat meat (2022-09-23 10:40:59)
Offline
The nvidia driver does not provide a framebuffer https://github.com/NVIDIA/open-gpu-kern … 1201748789
Offline
The nvidia driver does not provide a framebuffer https://github.com/NVIDIA/open-gpu-kern … 1201748789
I guess that would explain it, thanks.
Offline