On 02/06/2018 12:20 PM, Cornelia Huck wrote:
> On Tue, 6 Feb 2018 12:10:20 +0100
> Halil Pasic <> wrote:
>
>> On 02/06/2018 01:05 AM, Michael S. Tsirkin wrote:
>>> On Mon, Feb 05, 2018 at 11:54:52PM +0100, Halil Pasic wrote:
[..]
>> In the CIO transport we have an explicit mention of virtio 1.0 (explains
>> revision 1). I wonder if that is still appropriate. Shouldn't that just
>> be virtio 1?
>
> We can certainly do s/1.0/1/ in the ccw transport (mind doing a patch?)
>
Will do.
> For the greater picture: As I see it, any implementation conforming to
> 1.1 is also conforming to 1.0. To conform to 1.1, it only needs the
> split layout. I think we can continue with that for any further
> iterations of 1.n (and keep VERSION_1 as it is now.)
>
Agreed. I will think a bit if there is an elegant way to codify this in
the spec.
> What we want to do with version 2.n (should that ever come up) is up
> for discussion (can an implementation conform to version 2 but not to
> version 1?). But I prefer to cross that bridge when we come to it.
>
I agree. There is no reason to ponder about this right now.
Regards,
Halil
