On 02/13/2011 06:40 AM, Alexander Graf wrote:
Ah, yeah. I'm still not sure what to do about it. I was going to
fold the dynamic hcall registration into the patch set before
upstreaming. But then something paulus said made me rethink whether
the dynamic registration was a good idea. Still need to sort this out
before the series is really ready.
We can surely move it to dynamic later on. I think the "proper" way would be to
populate a qdev bus and have the individual hypercall receivers register themselves
through -device creations.
Ignore the qdev bit for a moment. Hypercalls could be plausibly
implemented as another I/O space from a processor so the thing to model
off of would be PIO dispatch (cpu_outb and friends).
From a qdev perspective, having a VIO bus makes sense. The details of
which I/O spaces are uses are not as important from a device tree
perspective.
Regards,
Anthony Liguori
Alex
