Mehrere Bildschirme in KDE Debian 12 mit NVIDIA einrichten?

So, I also asked the question on another forum, the whole thing is clear. But thanks for the help. For everyone who needs the answer:
In the comments, you said the

lspci -d ::0300

output was:

00:02.0 VGA compatible controller: Intel Corporation HD Graphics 630 (rev 04) 
01:00.0 VGA compatible controller: NVIDIA Corporation GP106M [GeForce GTX 1060 Mobile] (rev a1)

indicating that your system has two GPUs: the CPU-integrated Intel GPU (iGPU for short) and the NVIDIA discrete GPU (dGPU for short), and the

xrandr --listproviders

output was:

Providers: number : 1 
Provider 0: id: 0x1b7 cap: 0x1, Source Output crtcs: 4 outputs: 3 associated providers: 0 name:NVIDIA-0

that only the NVIDIA GPU is currently being used.

The problem

You have a dual-GPU laptop: NVIDIA calls this Optimus Technology. Unfortunately, the most common implementation of this does not include a multiplexer switch that would allow either GPU a physical connection to each display output connector. This is known as “muxless Optimus”. When using a combination of display connections that are not directly weed to the active GPU, the missing connections must be substituted by passing rendered output data from one GPU to the other, which can then output the data to the displays it’s physically connected to.

In laptops, a common muxless Optimus solution is as follows:

• the laptop’s own display and VGA output is wired to the iGPU (allowing power-efficient operation when no external displays are available, or when using the VGA output for presentation on a video projector)
• any digital display outputs search as HDMI and/or DisplayPort are usually wired to the dGPU, for maximum performance when connected to external desktop display (and presumably external power supply is readily available, so being economical with battery power is not an issue).

In your specific case, it seems that only the HDMI output of the NVIDIA GPU chip is actually wired to a physical output connector.

The solution

First, you should ensure that you have the

xserver-xorg-video-intel

package installed. It contains the X11 driver for your Intel iGPU, which is required even just to pass display data through it. That driver, in turn, requires that the kernel module ‘i915’ is loaded, and that any firmware files it requires are present, i.e. the

firmware-misc-nonfree

package is installed.

Then, you should make a backup of your current

/etc/X11/xorg.conf

file (if it exists), and then change it to contain the following (creating the file if it doesn’t exist):

Section "ServerLayout"
    Identifier "layout"
    Screen 0 "nvidia"
    Inactive "intel"
EndSection

Section "Device"
    Identifier "nvidia"
    Driver "nvidia"
    BusID "PCI:1:0:0"
EndSection

Section "Screen"
    Identifier "nvidia"
    Device "nvidia"
    Option "AllowEmptyInitialConfiguration"
EndSection

Section "Device"
    Identifier "intel"
    Driver "modesetting"
EndSection

Section "Screen"
    Identifier "intel"
    Device "intel"
EndSection

(This configuration snippet is from directlyNVIDIA GPU driver README file, with the BusID for your hardware already written.)

This will tell the X server to start both drivers (even if no displays seem to be connected), and tell it to use the NVIDIA GPU as the primary one.

With modern enough version of the Xorg X server, this alone might be enough to make your extra displays available for use.

But if not, you might need two commands to specifically feed display data from NVIDIA GPU to the Intel one:

xrandr --setprovideroutputsource modesetting NVIDIA-0
xrandr --auto

The first command tells the Intel GPU base driver (

modesetting

) to act as extra outputs for display data whose source is the NVIDIA GPU.

The second command makes the X server activate any displays that might have become available as the result of the previous command.

If you find the two

xrandr

commands are needed, you could automate the process by writing them into e.g.

/etc/X11/Xsession.d/92custom_xrandr

.

The only xorg.conf file I find apart from my backup is xorg.conf.nvidia-xconfig-original and this is empty.

I had deleted the Config, before creating a backup of course, then completely reinstalled it and automatically configured it. After a restart, it brings everything back to life. I’m on the stand right now. I had previously installed the open source variant of Nvidia with Cuda, I try without. By the way, I have had this problem every time on several computers, so a new installation does not help much in my experience.

So, driver plus xorg conf removed and xorg conf reconfigured. Everything works, but I don’t use the Graka. I install the open source drivers and immediately start the problems again. The LG in the middle no longer works and nvidia-smi does not work either. Also nvidia-settings doesn’t matter.