Re: [PATCH 2/5] migration/postcopy: magic value for postcopy channel

[Peter Xu <peterx@redhat.com>]

On Mon, Nov 04, 2024 at 06:02:33PM +0530, Prasad Pandit wrote:
> On Fri, 1 Nov 2024 at 20:04, Peter Xu <peterx@redhat.com> wrote:
> > As we discussed internally, we can't do this unconditionally.  We at least
> > some compat properties.
> 
> * ie. we define a new compat property to check if postcopy sends a
> magic value or not?
> 
> >  Or we need to see whether Fabiano's handshake can
> > simplify this, because the handshake will also re-design the channel
> > establishment protocol.
> 
> * May I know more about this handshake change? Is there a repository?
> OR a page/document that describes what is being planned? Is it just
> channel establishment change or there's more to it?

After a 2nd thought, maybe we don't need a new compat property, and this
should work even before handshake ready.

Firstly, we'll need a way to tell mgmt that the new qemu binary supports
enablement of both multifd + postcopy feature.  That can be done with a

  "features": [ "postcopy-with-multifd-precopy" ]

Flag attached to the QMP "migrate" command.

Then, I think we don't need a magic for preempt channel, because new qemu
binaries (after 7.2) have no issue on out-of-order connections between main
/ preempt channel.  Preempt channel is always connected later than main.

It means in the test logic of "which channel is which", it should be:

  - If it's a multifd channel (check multifd header match), it's a multifd
    channel, otherwise,

    - if main channel is not present yet, it must be the main channel (and
      we can double check the main channel magic), otherwise,

    - it's the preempt channel

With that, I think we can drop the new magic sent here.

Thanks,

-- 
Peter Xu