Since GPU passthrough is involved I'd ping @Team-Hypervisor-Kernel