Re: [Qemu-devel] [PATCH v2 01/10] qapi: add Visitor interfaces for uint*_t and int*_t

[Paolo Bonzini]

On 12/20/2011 09:56 PM, Anthony Liguori wrote:
> > As always, you can implement that in many ways. However, I think the
> > point of using Visitors is not to remove QEMUFile.
>
> Yes, it is.

The point of using Visitors is to provide a standard representation of 
device state.  QEMUFile is but one consumer of that representation, 
along with any other migration filter.

> We need something like a migration filter capability where we can
> encapsulate this kind of logic such that we can ween VMState away
> from these things (and ultimately switch to an IDL compiler).
>
> We can't do a migration filter until we have something like Michael's
> series.

Agreed.  I'm saying that the QEMUFile serialization should be the first 
example of a migration filter.  Mike's patch add a layer of indirection 
in the code, but not in the data.  We can't do a migration filter until 
we have device state represented as introspectable data.

> > Yes, this is accurate, but I see the goals differently. We should:
> >
> > (1) First and foremost, provide a backend-independent representation
> > of device state so that we can add other backends later.
>
> And Mike's series does this, no?

No.

> > (2) Serialize this with QEMUFile, both for backwards-compatibility and
> > to ensure that the whole thing works.
>
> Mike's series also does this, no?

It serializes device state directly, without going through a 
backend-independent representation.  It doesn't provide a place in which 
you can plug a filter.

Paolo