On Mon, May 20, 2019 at 12:44:29PM +0200, Paolo Bonzini wrote:
> On 20/05/19 05:08, Peter Xu wrote:
> > Expose the new capability via "query-kvm" QMP command too so we know
> > whether that's turned on on the source VM when we want.
> > 
> > Signed-off-by: Peter Xu <pet...@redhat.com>
> 
> Is this useful?  We could I guess make a migration capability in order
> to benchmark with the old code, but otherwise I would just make this a
> "hidden" optimization just like many others (same for patch 14).
> 
> In other words, there are many other capabilities that we could inform
> the user about, I don't see what makes manual_dirty_log_protect special.

Yes this is mostly used for me to make sure the new capability is
enabled when comparing with the old code.  I added QMP part too
because otherwise I'll need to justify why I only add HMP...

But I agree with above - let's drop these two QMP/HMP patches.

Thanks,

-- 
Peter Xu

Reply via email to