How is your graphics card pass-through? Is graphics card pass-through related to the brand or the generation of the graphics card?

I don’t quite understand why qubes’ graphics card passthrough is so difficult. I have a few questions:

Completed goals: First, I want to pass the RTX4090 graphics card directly to appvm, so the display of dom0 must be handed over to the nuclear display

First: If I use a monitor cable to install by plugging in the RTX4090, can I still achieve my goal by modifying grub parameters in the future? I tried many ways but still failed, is there anyone who has succeeded? Can you give me the complete steps?

Second: Secondly, if you must first remove the graphics card and use the only core display to install Qubes, after the installation is completed, I plug the RTX4090 back into the computer, will Dom0 still use the RTX4090 first? If not, what steps should I take to let the graphics card pass through after plugging the RTX4090 back into the computer?

My question: When the monitor cable is plugged into the core display port

BIOS

GRUB :white_check_mark: There is a screen

Qubes start

LUKS password interface :white_check_mark:/There is a screen in some cases

Enter password

Black screen/no signal on monitor

I dont really need any, I want reliable iGPU virtualization :frowning:

Answering your question from the title:

Is graphics card pass-through related to the brand or the generation of the graphics card?

Yes. GPU PCI passthrough may or may not work depending on your:

  1. CPU (iGPU)
  2. GPU
  3. Motherboard
  4. BIOS/UEFI version/type (proprietary/FOSS)

First: If I use a monitor cable to install by plugging in the RTX4090, can I still achieve my goal by modifying grub parameters in the future?

If I understand you correctly then yes, GPU passthrough can be done at any point after the installation, doesn’t have to be immediately after.

Can you give me the complete steps?

https://neowutran.ovh/qubes/articles/gaming_linux_hvm.html

what steps should I take to let the graphics card pass through after plugging the RTX4090 back into the computer?

In most situations, if you plug your monitor into the iGPU HDMI/DP ports on your motherboard and mask your dGPU, dom0 should use the iGPU automatically. Sometimes that does not happen and requires a small xorg.conf snippet to be added to dom0.

Thank you for your help, but I still hope you can give me detailed steps. For example, can you write out the detailed configuration of xorg.conf? I have also updated my question. If you look at the steps of my question, there is no signal on that interface. I don’t know if it is a hardware problem or a problem with the qubes system settings? ? ? @Atrate

If I recall correctly I used this xorg.conf.d drop-in:

Section "Device"
    Identifier  "IntelGPU"
    Driver	"modesetting"
    BusID	"PCI:0:2:0"
EndSection

I have a question, is this a solution when installing the qubes system or a method of using it after installing the system? If you need to edit this document during installation, how should you edit it? Because only grub parameters can be modified during installation, you cannot enter the dom0 shell and enter commands to create and modify this file. @Atrate

After installation.