>>> On 21.09.10 at 17:17, Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx> wrote:
> On Mon, Sep 20, 2010 at 02:23:51PM +0100, Jan Beulich wrote:
>> These structures are used by Xen, and hence guests must not be able
>> to fiddle with them.
>> qemu-dm currently plays with the MSI-X table, requiring Dom0 to
>> still have write access. This is broken (explicitly allowing the guest
>> write access to the mask bit) and should be fixed in qemu-dm, at which
>> time Dom0 won't need any special casing anymore.
>> The changes are made under the assumption that p2m_mmio_direct will
>> only ever be used for order 0 pages.
>> An open question is whether dealing with pv guests (including the
>> IOMMU-less case) is necessary, as handling mappings a domain may
>> already have in place at the time the first interrupt gets set up
>> would require scanning all of the guest's L1 page table pages.
> When the PCI passthrough is utilized for PV guests we utilize
> the xc_domain_iomem_permission, xc_domain_ioport_permission, and
> xc_physdev_map_pirq before we even start the guest.
> With your patch, will the MFN regions that are specified by the
> iomem_permission still be visible to the PV domain?
Yes, just that the page(s) containing MSI-X table and PBA won't
be writeable anymore (if the guest tries to map them so, they'll
get mapped read-only). And yes, the MSI-X table should be
ignored by pv guests altogether, and the PBA (afaict) isn't being
used by Linux up to now.
Xen-devel mailing list