Re: [PATCH v15 13/14] migration: Implement MigrateChannelList to hmp migration flow.

[Peter Xu]

On Tue, Oct 31, 2023 at 06:42:43PM +0100, Juan Quintela wrote:
> Fabiano Rosas <farosas@suse.de> wrote:
> > From: Het Gala <het.gala@nutanix.com>
> >
> > Integrate MigrateChannelList with all transport backends
> > (socket, exec and rdma) for both src and dest migration
> > endpoints for hmp migration.
> >
> > Suggested-by: Aravind Retnakaran <aravind.retnakaran@nutanix.com>
> > Signed-off-by: Het Gala <het.gala@nutanix.com>
> > Signed-off-by: Fabiano Rosas <farosas@suse.de>
> 
> 
> >      bool resume = qdict_get_try_bool(qdict, "resume", false);
> >      const char *uri = qdict_get_str(qdict, "uri");
> >      Error *err = NULL;
> > +    MigrationChannelList *caps = NULL;
> > +    g_autoptr(MigrationChannel) channel = NULL;
> >  
> > -    qmp_migrate(uri, false, NULL, !!blk, blk, !!inc, inc,
> > +    if (!migrate_uri_parse(uri, &channel, &err)) {
> > +        goto end;
> > +    }
> > +    QAPI_LIST_PREPEND(caps, g_steal_pointer(&channel));
> > +
> > +    qmp_migrate(NULL, true, caps, !!blk, blk, !!inc, inc,
> >                   false, false, true, resume, &err);
> 
> 
> > -    if (hmp_handle_error(mon, err)) {
> > -        return;
> > -    }
> 
> I think that dropping this chunk is wrong.  What assures that
> qmp_migrate will not give an error?
> 
> > +    qapi_free_MigrationChannelList(caps);
> 
> 
> 
> 
> 
> >  
> >      if (!detach) {
> >          HMPMigrationStatus *status;
> > @@ -766,6 +780,9 @@ void hmp_migrate(Monitor *mon, const QDict *qdict)
> >                                            status);
> >          timer_mod(status->timer, qemu_clock_get_ms(QEMU_CLOCK_REALTIME));
> >      }
> > +
> > +end:
> > +    hmp_handle_error(mon, err);
> 
> 
> Oh, you put it here, but you enter in the detach case even if there is
> one error.
> 
> I think it is easier to just repeat the hmp_mhandle_error() inplace of
> the goto.

Agreed.  We don't want to kick the timer even if error..

-- 
Peter Xu