On Mon, 03 Feb 2014 13:50:02 -0700
Eric Blake <ebl...@redhat.com> wrote:
> On 02/03/2014 01:38 PM, Igor Mammedov wrote:
>
> >>> +static inline void qbus_set_hotplug_handler(BusState *bus, DeviceState
> >>> *handler,
> >>> + Error **errp)
> >>> +{
> >>> + object_property_set_link(OBJECT(bus), OBJECT(handler),
> >>> + QDEV_HOTPLUG_HANDLER_PROPERTY, errp);
> >>> + bus->allow_hotplug = 1;
> >>
> >> Should we convert allow_hotplug to bool over the course of this series?
> > There isn't much point in touching it, since it's going to be removed once
> > all hotplug-able buses are converted to new interface.
>
> After finishing my read through your series, I noticed that you were
> using allow_hotplug less and less and instead using a new bool in most
> code you added. Does that just mean your series didn't touch every last
> client of the old interface, and that there still remains code to be
> converted, where this series serves as the example of how?
Yes, beside adding new hotplug mechanism, it converts only PCI hotplug
workflow as example.
Paolo had plans to convert scsi hotplug once this series is in.
>
> --
> Eric Blake eblake redhat com +1-919-301-3266
> Libvirt virtualization library http://libvirt.org
>
--
Regards,
Igor
