[Qemu-devel] Re: [SeaBIOS] seabios: acpi: add _RMV control method for PCI devices

[Marcelo Tosatti <mtosatti@redhat.com>]

On Mon, Dec 20, 2010 at 11:15:40AM +0200, Avi Kivity wrote:
> On 12/20/2010 10:49 AM, Marcelo Tosatti wrote:
> >On Sun, Dec 12, 2010 at 02:57:45PM -0500, Kevin O'Connor wrote:
> >>  On Sun, Dec 12, 2010 at 09:49:16PM +0200, Gleb Natapov wrote:
> >>  >  On Sun, Dec 12, 2010 at 02:11:29PM -0500, Kevin O'Connor wrote:
> >>  >  >  On Wed, Dec 08, 2010 at 03:08:59PM -0200, Marcelo Tosatti wrote:
> >>  >  >  >  Use _RMV method to indicate whether device can be removed.
> >>  >  >  >
> >>  >  >  >  Data is retrieved from QEMU via I/O port 0xae0c.
> >>  >  >  >
> >>  >  >  >  Signed-off-by: Marcelo Tosatti<mtosatti@redhat.com>
> >>  >  >
> >>  >  >  Thanks Marcelo,
> >>  >  >
> >>  >  >  Can you add acked-bys from the qemu/kvm maintaners?
> >>  >  >
> >>  >  Marcelo is kvm maintainer ;) FWIW I tested this with Windows 7&  XP
> >>  >  and it works as expected. The patch relies on patch not yes excepted
> >>  >  to qemu though.
> >>
> >>  Yeah - I know - but I figured I'd ask for concensus before committing.
> >>  :-)
> >>
> >>  The committing of seabios patches that have a dependency on qemu/kvm
> >>  is an area that I think could be better clarified.  I'm thinking that
> >>  if the seabios part depends on something in qemu/kvm then we should
> >>  have the corresponding kvm/qemu maintaners "ack" it.
> >>
> >>  -Kevin
> >
> >Avi, are you OK with this patch ?
> >
> >Yes, perhaps it all (including PCI hotplug controller) should be using
> >something else than hardcoded IO ports, but thats what we have now.
> 
> At least it should be documented.
> 
> What's the behaviour with a qemu that doesn't support the new port?
> We don't strictly support it, but let's do so if we can.

All slots will be marked as hotpluggable (since the "removability" IO port will not be
registered and return all 1's).