On Fri, 13 Jul 2012 08:59:29 +0100, Peter Maydell
<peter.maydell at linaro.org> wrote:
> On 13 July 2012 02:06, Rusty Russell <rusty.russell at linaro.org> wrote:
>> What do you want to do about the API breakage? Should we just insist
on
>> a lockstep qemu/kernel update, or some temporary bandaid in qemu? The
>> latter should be possible, but is it worth the hassle?
>
> I'm not fussed personally. Lockstep would be fine with me. I think
> we should do whatever's easiest unless there are active developers
> here that would be seriously inconvinienced by that...
The sooner we switch to what we expect to be a stable API, the better.
There is only a handful of KVM/ARM developers, and I'm sure we can manage a
flag day switch.
Same goes for the VGIC interrupt injection API. And if we can combine all
breakages in one go, it's even better ;-).
M.
--
Fast, cheap, reliable. Pick two.
