Re: [Qemu-devel] [0/4] RFC: Preparations for VFIO and guest IOMMUs (v2)

[David Gibson]

On Fri, Apr 26, 2013 at 09:42:31AM -0600, Alex Williamson wrote:
> On Fri, 2013-04-26 at 16:02 +1000, David Gibson wrote:
> > This patch series represents a second attempt at better integration of
> > the vfio code with qemu's handling of guest IOMMUs.  It is based on
> > Paolo Bonzini's tree at git://github.com/bonzini/qemu.git (iommu
> > branch).
> > 
> > ddThis series should open the way for using VFIO with a guest system
> > containing an IOMMU by passing guest IOMMU operations through to the
> > host IOMMU via VFIO.  That's opposed to the present model of having no
> > IOMMU in the guest, and simply mapping all guest RAM into the host
> > IOMMU.
> > 
> > Patch 1 is just a trivial fix for a build problem introduced in the
> > iommu tree.
> 
> I support what this series is trying to achieve, but not how it does it.
> vfio needs to be just another driver,

That's a desirable goal, but given vfio has inherently unique
constraints, I'm not sure it's achievable.

> the ugliness and heavy lifting
> should be done in vfio or the common should should be extended to
> generically support the hooks that vfio needs.  We can't just drop vfio
> pointers and callbacks into unrelated code.  Thanks,

Well, suggest me a way to do it..

-- 
David Gibson                    | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au  | minimalist, thank you.  NOT _the_ _other_
                                | _way_ _around_!
http://www.ozlabs.org/~dgibson

signature.asc
Description: Digital signature