3 displays work fine in installer, system boots to fixed cursor - Intel Arc A380, Nvidia RTX A4500

Displays are connected to the Intel card.
After install, the system boots to a non-blinking cursor. Can’t switch to a tty. It seems to be when the display manager should start.
I’ve read posts that are a couple of years old about adding i915.force_probe=[PCI-ID], but I believe I read elsewhere that that is no longer necessary. Since the installer runs fine, I don’t think it should be needed.

Dead in the water, and would appreciate any suggestions!

Edit: installer won’t open after selecting Start installation. Error in dmesg saying Module Nvidia is blacklisted. (Which is fine, the Nvidia GPU will just be for CUDA stuff, not displays.)

Edit2: Selecting fallback nomodeset disables the Intel driver, one display works, and I can run the installer.

Latest EndeavourOS download 2024.01.25
Asus ProArt Creator WiFi
NVidia RTX A4500
Intel Arc A380
bunch of other stuff

installer has some resolving implementations for GPU and drivers… because it supports nvidia driver and nouveau in addition it forces modesetting for intel and AMD in case they are present.

Your case looks like it can not handle the intel card as default output device.
Could be you need to manually blacklist nouveau on bootup as kernel ships with this too…

nouveau.modeset=0 module_blacklist=nouveau

as kernel parameters could do that… you can also add these manually on boot menu after hitting e

this may need some extra setup i bet…

I greatly appreciate the replies @joekamprad ! I did finally get it working.
I added modprobe.blacklist=nouveau to the kernel boot parameters, and was able to boot into my installed system.
I then added that to the /etc/default/grub configuration file:

GRUB_CMDLINE_LINUX_DEFAULT='nowatchdog nvme_load=YES loglevel=3 modprobe.blacklist=nouveau'

And updated grub:

grub-mkconfig -o /boot/grub/grub.cfg 

And the system now boots normally.

To get CUDA set up, I did the following:

sudo pacman -S cudnn

And for the Nvidia driver:

sudo pacman -S nvidia-lts

If one is using the non-lts kernel, you would do:

sudo pacman -S nvidia

I now have all the Nvidia GPU memory available for running models, which was the motivation for all of this.

refs: Arch wiki, Reddit thread

1 Like

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.