Right now, libvirt has generic support for vhost-user network interfaces (<interface type='vhostuser'>), and for the passt back-end, but it's not possible to connect them together, that is, it's not possible to tell the passt back-end to enable vhost-user mode, and use the same control socket as the (libvirt) vhost-user interface.

A typical invocation of passt in vhost-user mode is simply: "passt --vhost-user". The corresponding QEMU command line is... wait for it... qemu-system-x86_64 ...
-chardev socket,id=chr0,path=/tmp/passt_1.socket
-netdev vhost-user,id=netdev0,chardev=chr0
-device virtio-net,netdev=netdev0
-object memory-backend-memfd,id=memfd0,share=on,size=$RAMSIZE
-numa node,memdev=memfd0.

This supports 1. improved throughput/decreasing CPU load and 2. live migration of passt context (including TCP migration). 2. would be covered entirely by QEMU and passt themselves, without the need for libvirt's intervention, but it's much easier to implement this in passt in vhost-user mode because of https://qemu-project.gitlab.io/qemu/interop/vhost-user.html#migrating-back-end-state.

vhost-user support has just been merged in passt and it will be available in a release (and distribution packages) in a couple of days, but I wanted to file this request as soon as possible. Patchset: https://archives.passt.top/passt-dev/20241122164337.3377854-1-lvivier@redhat.com/.