You are not logged in.
This is a quite special setup, i have a r9 380 and a gtx 1080. All displays are connected to the amd and i want to run games on the nvidia gpu. This works fine using prime so far.
BUT this does only work when my gpu is binded to xorg, so that it shows up in xrandr --listproviders.
Because i want to use the nvidia gpu in a windows 10 vm, i need to rebind it to vfio driver. Is it possible to achive gpu offloading in my main xorg session and still be able to rebind the driver?
Things i tried: nvidia-xrun and bumblebee, because these tools should be able to create a "backgound" xorg session. But bumblebee complains about no igpu. Could i use bumblebee on the amd gpu or are there better options?
Offline
Bumblebee is intended for Hybrid graphics systems with an integrated gpu (normally intel) and 1 discrete nvidia videocard.
Your description suggests you have a system with 2 discrete videocards.
In theory it should be possible to switch the kernel module binding of the nvidia card from vfio to nvidia and vv when it's not in use.
I don't recall someone doing that successfully though.
The redhat vfio-users mailing list is/has been the goto place for compilcated passthrough setups for a long time, Try asking there .
https://www.redhat.com/mailman/listinfo/vfio-users .
Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.
clean chroot building not flexible enough ?
Try clean chroot manager by graysky
Online