[Xen-devel] Re: Illegal PV kernel pfm/pfn translations on PROT_NONE ioremaps

["Stephen C. Tweedie" <sct@xxxxxxxxxx>]

Hi,

On Wed, 2008-03-19 at 16:42 +0000, Keir Fraser wrote:

> > It might be easier to do this at the pte_machine_to_phys level instead,
> > where we can potentially take advantage of other bits of the pte to
> > encode the special casing.
> 
> Oh yes, the PAGE_IO type of trick I mentioned in my other email just now.

Yep.  There are a number of bits that could be used: for example, I
don't think PROT_WRITE is ever going to be set on a PROT_NONE page,
either.

But...

> SIF_PRIVILEGED is no longer used. It's still set for dom0, but not for
> hardware-capable domUs. It's tricky anyway, since a domU can be given
> hardware capabilities after it has booted through mechanisms like
> pciback-pcifront PCI hotplug.

Right; but I still think it's reasonable to expect this sort of thing to
be declared in advance for a domain.  We clearly cannot enter a non-
translating mode dynamically, because of the risk of already having
untranslated PROT_NONE ptes in the page tables.  But becoming unable to
save or migrate a guest is a Big Thing, and it would seem not
unreasonable to ask for a startup flag such as SIF_PRIVILEGED to be
declared in advance before we permit that.

I'll probably do a quick patch based in SIF_INITDOMAIN for now, simply
to be able to test whether it does in fact fix the symptoms we're
seeing; but that might not be the best solution to merge long term.

Cheers,
 Stephen



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