Re: [Qemu-devel] [RFC PATCH v1 06/13] spapr: CPU hotplug support

From: Michael Roth <mdroth@linux.vnet.ibm.com>
To: bharata@linux.vnet.ibm.comBharata B Rao <bharata@linux.vnet.ibm.com>
Cc: imammedo@redhat.com, qemu-devel@nongnu.org, agraf@suse.de
Subject: Re: [Qemu-devel] [RFC PATCH v1 06/13] spapr: CPU hotplug support
Date: Tue, 27 Jan 2015 23:41:19 -0600	[thread overview]
Message-ID: <20150128054119.11140.24166@loki> (raw)
In-Reply-To: <20150128041956.GA3393@in.ibm.com>

Quoting Bharata B Rao (2015-01-27 22:19:56)
> On Thu, Jan 22, 2015 at 04:16:01PM -0600, Michael Roth wrote:
> > Quoting Bharata B Rao (2015-01-08 00:10:13)
> > > +static void spapr_machine_device_plug(HotplugHandler *hotplug_dev,
> > > +                                      DeviceState *dev, Error **errp)
> > > +{
> > > +    if (object_dynamic_cast(OBJECT(dev), TYPE_CPU)) {
> > > +        if (dev->hotplugged) {
> > 
> > Maybe just
> > 
> >            if (dev->hotplugged && spapr->dr_cpu_enabled) {
> >                ...
> > 
> > Would do it
> 
> This is a common ->plug() handler and would be used for memory too. Hence
> there is a need to identify the type of object (CPU or memory) and handle it
> differently.

I mean in terms of the "/* TODO: Check if DR is enabled ? */". Adding this
check here, as well as during spapr_dr_connector_new(), should cover all
the cases where the value needs to be checked AFAICT.

> 
> Regards,
> Bharata.