On Wed, 2007-09-26 at 17:12 -0700, Jeremy Fitzhardinge wrote:
> rusty@xxxxxxxxxxxxxxx wrote:
> > +/* We walk down the guest page tables to get a guest-physical address */
> > +unsigned long guest_pa(struct lguest *lg, unsigned long vaddr)
> > +{
> > + pgd_t gpgd;
> > + pte_t gpte;
> > +
> > + /* First step: get the top-level Guest page table entry. */
> > + gpgd = __pgd(lgread_u32(lg, gpgd_addr(lg, vaddr)));
> > + /* Toplevel not present? We can't map it in. */
> > + if (!(pgd_flags(gpgd) & _PAGE_PRESENT))
> > + kill_guest(lg, "Bad address %#lx", vaddr);
> > +
> > + gpte = __pte(lgread_u32(lg, gpte_addr(lg, gpgd, vaddr)));
> > + if (!(pte_flags(gpte) & _PAGE_PRESENT))
> > + kill_guest(lg, "Bad address %#lx", vaddr);
> > +
> > + return pte_pfn(gpte) * PAGE_SIZE | (vaddr & ~PAGE_MASK);
> > +}
>
> That's nice and general, but is it necessary? Is this just to handle
> the two cases of booting with P=V, then switching to the normal kernel
> offset?
Hi Jeremy,
It's probably overkill, but it has the advantage of being
straightforward and futureproof.
Cheers,
Rusty.
_______________________________________________
Virtualization mailing list
Virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linux-foundation.org/mailman/listinfo/virtualization
