Re: [PATCH v2] migration: check magic value for deciding the mapping of channels

[Peter Xu]

On Tue, Nov 15, 2022 at 11:29:13PM +0530, manish.mishra wrote:
> > > +   while (bytes < nbytes) {
> > > +       bytes = klass->io_read_peek(ioc,
> > > +                                   buf,
> > > +                                   nbytes,
> > > +                                   errp);
> > IIUC here you need to accumulate bytes rather than directly reusing it, so
> > e.g. two reads on 2 bytes each will satisfy a 4 bytes read.
> 
> Actually peek always reads from top so even if there are multile reads it
> does not accumulate.

Fair enough.

[...]

> Thank you Peter for review. I see this patch is included by Juan too in
> some other patch series of 30 patches, so how it will work. Do i need to
> send a independent V3 for this? Or it should be reviewed in that series.

>From what I see right now, feel free to rework on new versions.  Please
also consider copying Dave (dgilbert@redhat.com) and Leonardo
(lsoaresp@redhat.com) on future posts.

Thanks,

-- 
Peter Xu