Hello-
I hate asking for tech support on this stuff, but I’ve RTFM-d every resource I could find, and have gotten nowhere. Also I apologize if I’ve confused any names like libvirt, qemu, etc.
I am running Arch Linux, and trying to do the (in)famous gpu passthrough VM setup. I am following this guide on the glorious Arch wiki “PCI Passthrough via OVMF” (can’t post a link as a new user)
Everything has gone fine until section 4.1, configuring libvirt. It instructs me to edit a config file to enable EFI support with the virtual machine, but when I edit it as instructed, and restart libvirt, it prevents virt-manager from making any kvm connections. (Specifically it says Error: No active connection to install on.)
Nearly every resource I found told me to add the same line to the nvram section, so I have no idea what I am supposed to do instead.
The comments in the config file mention that the nvram option is obsolete, so I am wondering if the guides (including the Arch wiki) are just all outdated.
I’ve been searching for a solution for a while, and have only found the same info.
Something that is not mentioned in that article is the fact that libvirtd needs additional packages for the virtual network. If libvirtd isn’t starting try installing the following packages: dmidecode, dnmasq, and ebtables see if libvirtd starts then.
Also, as gordonthree mentioned, make sure your user account is in the right group, I personally have my user in the kvm group for virtualization and it works for me. These are the settings I would use in /etc/libvirt/qemu.conf:
FFS I not only made a syntax error in the config file, but I forgot to uncomment a line too!
…Everything works now, I was stupid.
Sorry about that.
Last time I make a forum post at 11pm…
Thanks for the help, sorry if I wasted your time.