[Qemu-devel] Re: [PATCH v2 3/9] provide in-kernel ioapic

[Glauber Costa <glommer@redhat.com>]

On Fri, Oct 09, 2009 at 09:55:13PM +0200, Juan Quintela wrote:
> Jamie Lokier <jamie@shareable.org> wrote:
> > Glauber Costa wrote:
> >> On Thu, Oct 08, 2009 at 06:22:48PM +0200, Gleb Natapov wrote:
> >> > On Thu, Oct 08, 2009 at 06:17:57PM +0200, Avi Kivity wrote:
> >> > > On 10/08/2009 06:07 PM, Jamie Lokier wrote:
> >> > > >Haven't we already confirmed that it *isn't* just an ioapic accelerator
> >> > > >because you can't migrate between in-kernel iopic and qemu's ioapic?
> >> > > 
> >> > > We haven't confirmed it.  Both implement the same spec, and if you
> >> > > can't migrate between them, one of them has a bug (for example, qemu
> >> > > ioapic doesn't implement polarity - but it's still just a bug).
> >> > > 
> >> > Are you saying that HW spec (that only describes software visible behavior)
> >> > completely defines implementation? No other internal state is needed
> >> > that may be done differently by different implementations?
> >> Most specifications leaves a lot as implementation specific.
> >> 
> >> It's not hard to imagine a case in which both devices will follow
> >> the spec correctly, (no bugs involved), and yet differ in the
> >> implementation.
> >
> > Avi's not saying the implementations won't differ.  I believe he's
> > saying that implementation-specific states don't need to be saved if
> > they have no effect on guest visible behaviour.
> 
> Just to re-state.  I would also prefer to have a single device.  Reasons
> (majority already told in the thread):
> - We can switch between devices more easily
> - They are emulating the same device.
> - At the moment that you have two different devices, one of them will
>   rot :(
> - Adding state to the save/load format that is used only from one device
>   is not a problem.
> 
> I notice that discussion is going nowhere, basically we are in the
> state:
> - people that want one device
>   * they emulate the same hardware
>   * lots of code is shared
>   * they should be interchageable
>   * if they are not interchageable, it is a bug
>   * once that they are split, it is basically imposible to join then
>     again.
> - people that want 2 devices:
>   * The devices can more easily diverge if they are two devices
>   * They are not interchageable now
>   * It allows you more freedom in changing any of them if they are
>     separate.
> 
> As you can see, none of the proposals is a clear winner.  And what is
> worse, we have the two maintainers (avi and anthony), the two with more
> experience having to deal with this kind of situation disagreeing.
> 
> How to fix the impass?
a deathmatch?