Re: [V2 PATCH 7/9] x86/hvm: pkeys, add pkeys support for guest_walk_tables

["Han, Huaitong" <huaitong.han@intel.com>]

On Tue, 2015-12-01 at 20:30 +0000, Andrew Cooper wrote:
> > +#include <asm/i387.h>
> 
> I can see why you need xstate.h, but I why do you need i387.h ?
Use vcpu_save_fpu functions.

> > +
> >      if ( pse2M )
> >      {
> >          /* Special case: this guest VA is in a PSE superpage, so
> > there's
> > @@ -330,6 +390,11 @@ guest_walk_tables(struct vcpu *v, struct
> > p2m_domain *p2m,
> >              goto out;
> >          }
> >          rc |= ((gflags & mflags) ^ mflags);
> > +#if GUEST_PAGING_LEVELS >= 4
> > +        pkeys = guest_l1e_get_pkeys(gw->l1e);
> > +        if (leaf_pte_pkeys_check(v, pfec, gflags, pkeys))
> > +            rc |= _PAGE_PKEY_BIT;
> > +#endif
> 
> As I identified in v1, the fact that you do not modify the callers of
> guest_walk_tables() proves that this change is buggy.  You must
> modify
> the callers to cope with the new error of _PAGE_PKEY_BIT.
_PAGE_PKEY_BIT just a flag to tell
 hap_p2m_ga_to_gfn(GUEST_PAGING_LEVELS) there is a page fault, PK is
real physical fault which is included in pfec, the funtion of
_PAGE_PKEY_BIT is that hap_p2m_ga_to_gfn returns INVALID_GFN, you can
find the pfec modification on PATCH 9/9.

> 
> ~Andrew