Re: [Xen-ia64-devel] paravirt_ops and its alternatives

[Isaku Yamahata <yamahata@xxxxxxxxxxxxx>]

On Wed, Jan 30, 2008 at 08:23:54PM -0700, Alex Williamson wrote:

>    We discussed this a little on the phone conference today.  It's going
> to be a bit of a learning curve (at least for me), but it should
> probably be in git.  This will allow us to better coordinate with the
> other upstream linux trees and we might even be able to get other
> maintainers to pull from it.

Yes the git rep. is usefull for other maintainers.
So git rep + patch queues sound reasonable.


> Isaku, I think you might be pretty close
> to getting domU booting on 2.6.24, but if it might be a while, perhaps
> you could send out a snapshot.

I see. Anyway please wait until the next week.


>    Yes, I hope the xen machine vector will be useful, but dom0 currently
> still runs with the bare metal machine vector.  I don't know if we can
> get away with that, or somehow make a paravirt wrapper for the machine
> vectors.  For DMA, we've taken the approach of turning the swiotlb and
> sba_iommu drivers into transparently paravirtualized drivers.  These
> don't feel clean enough for upstream as they are, so I hope we can
> abstract the interfaces further.  I would guess for interrupts, we
> probably need to transform the upstream code to create a more modular
> interrupt controller infrastructure.  Then the Xen virtual irq chip
> would just be a driver that plugs into that interface.

Yes, what I mind are also irq chip and dma api.

-- 
yamahata

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel