[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH 2 of 6 V6] amd iommu: call guest_iommu_set_base from hvmloader

>>> On 15.10.12 at 15:41, Wei Wang <wei.wang2@xxxxxxx> wrote:
> My idea is to make the interface looks like that
> guest_iommu_set_base(d, iommu_base, iommu_seg)

That part is obvious; the more interesting part is how you plan
on getting this fix the HVM parameter model.

> This will allow hvmloader to choose a non-zero iommu segment.
> if iommu_seg > 0, then I will add a new iommu to guest iommu list and 
> update iommu_segment field accordingly. But I seem to see no reason to 
> add multiple guest iommus to pci segment 0.

I don't see a need for this at present too, and as I said before
there's no need to implement support for multiple IOMMUs. The
interface definition, however, has to be flexible enough so we
don't need to introduce another interface in a few years time
just because right now we don't expect multiple IOMMUs to be
needed for guests (after all, there must be reasons why real
hardware frequently comes with more than one IOMMU despite
supporting only a single PCI segment - just consider the guest
visible NUMA case, where you would quite likely want an IOMMU
per guest visible node).


Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.