"Dr. David Alan Gilbert" <dgilb...@redhat.com> writes: > Yep, that's probably safest; although if you recorded the features used > in the migration stream you could check for those on the destination and > if they mismatch complain then. >
There is no clear use-case for hv-all other than development at this moment; as Daniel previously stated we may never support it in libvirt. I decided to take the easiest path first and think about migration later, when we understand why we would want to migrate such guests. In theory, yes, we may compare Hyper-V feature words on source and destination and allow migration when the former is a subset of the later. -- Vitaly
