Re: [Qemu-devel] [PATCH for-2.12 REPOST] migration: convert socket server to QIONetListener

[Dr. David Alan Gilbert]

* Paolo Bonzini (address@hidden) wrote:
> On 13/03/2018 12:06, Daniel P. Berrangé wrote:
> > On Tue, Mar 13, 2018 at 11:57:17AM +0100, Paolo Bonzini wrote:
> >> On 13/03/2018 10:20, Dr. David Alan Gilbert wrote:
> >>>>  static void socket_start_incoming_migration(SocketAddress *saddr,
> >>>>                                              Error **errp)
> >>>>  {
> >>>> -    QIOChannelSocket *listen_ioc = qio_channel_socket_new();
> >>>> +    QIONetListener *listener = qio_net_listener_new();
> >>>>  
> >>>> -    qio_channel_set_name(QIO_CHANNEL(listen_ioc),
> >>>> -                         "migration-socket-listener");
> >>>> +    qio_net_listener_set_name(listener, "migration-socket-listener");
> >>>>  
> >>>> -    if (qio_channel_socket_listen_sync(listen_ioc, saddr, errp) < 0) {
> >>>> -        object_unref(OBJECT(listen_ioc));
> >>>> +    if (qio_net_listener_open_sync(listener, saddr, errp) < 0) {
> >>>> +        object_unref(OBJECT(listener));
> >>>>          return;
> >>>>      }
> >>>>  
> >>>> -    qio_channel_add_watch(QIO_CHANNEL(listen_ioc),
> >>>> -                          G_IO_IN,
> >>>> -                          socket_accept_incoming_migration,
> >>>> -                          listen_ioc,
> >>>> -                          (GDestroyNotify)object_unref);
> >>>> +    qio_net_listener_set_client_func(listener,
> >>>> +                                     socket_accept_incoming_migration,
> >>>> +                                     NULL, NULL);
> >>>>  }
> >>>>  
> >>>>  void tcp_start_incoming_migration(const char *host_port, Error **errp)
> >>>> -- 
> >>>> 2.14.3
> >>>>
> >>> --
> >>> Dr. David Alan Gilbert / address@hidden / Manchester, UK
> >>>
> >>
> >> Could this get in 2.12 as a bugfix, even after soft freeze?  It's the
> >> last user of qio_channel_listen_sync that blocks full IPv4/IPv6 support.
> >>  All other users are for vsock/AF_UNIX.
> > 
> > I had been thinking of all this conversion as feature, but I guess if we
> > squint our eyes a bit we could claim it is just a bugfix on the grounds
> > that it is fixing broken IPv4/6 dual stack support.
> 
> It is certainly a feature, but when only one listener remains broken
> then it becomes more of a bugfix.

Sounds reasonable to me.

Dave

> Paolo
--
Dr. David Alan Gilbert / address@hidden / Manchester, UK