On Thu, 2007-11-01 at 16:43 +0000, Tim Deegan wrote:
> At 16:36 +0000 on 01 Nov (1193934965), Kieran Mansley wrote:
> > One question in my mind is whether the tests of iomem_access_permitted()
> > in _sh_propogate() would be better replaced with the more general
> > shadow_mode_translate(). The former seemed less risky to me as
>
> You could do that. I think the correct test is shadow_mode_refcounts();
> as Keir pointed out, guests whose refcounting isn't done by the shadow
> code will have had their pagetables validated bby the PV MMU interface.
>
> The test around the log-dirty code can just be on mfn_valid().
Attached is a new patch with those changes. Seems much more sensible to
me that way too, and it works just as well.
Signed-off-by: Kieran Mansley <kmansley@xxxxxxxxxxxxxx>
Thanks
Kieran
fix iomem frame shadow propogation to allow live migration
diff -r b28fa67d3940 xen/arch/x86/mm/shadow/multi.c
--- a/xen/arch/x86/mm/shadow/multi.c Thu Nov 01 11:52:34 2007 +0000
+++ b/xen/arch/x86/mm/shadow/multi.c Fri Nov 02 09:27:43 2007 +0000
@@ -696,7 +696,8 @@ _sh_propagate(struct vcpu *v,
/* N.B. For pass-through MMIO, either this test needs to be relaxed,
* and shadow_set_l1e() trained to handle non-valid MFNs (ugh), or the
* MMIO areas need to be added to the frame-table to make them "valid". */
- if ( !mfn_valid(target_mfn) && (p2mt != p2m_mmio_direct) )
+ if ( shadow_mode_refcounts(d) &&
+ !mfn_valid(target_mfn) && (p2mt != p2m_mmio_direct) )
{
ASSERT((ft == ft_prefetch));
*sp = shadow_l1e_empty();
@@ -712,6 +713,8 @@ _sh_propagate(struct vcpu *v,
_PAGE_RW | _PAGE_PRESENT);
if ( guest_supports_nx(v) )
pass_thru_flags |= _PAGE_NX_BIT;
+ if ( !shadow_mode_refcounts(d) && !mfn_valid(target_mfn) )
+ pass_thru_flags |= _PAGE_PCD | _PAGE_PWT;
sflags = gflags & pass_thru_flags;
/* Only change memory caching type for pass-through domain */
@@ -758,10 +761,12 @@ _sh_propagate(struct vcpu *v,
// p2m_ram_logdirty p2m type: only HAP uses that.)
if ( unlikely((level == 1) && shadow_mode_log_dirty(d)) )
{
- if ( ft & FETCH_TYPE_WRITE )
- paging_mark_dirty(d, mfn_x(target_mfn));
- else if ( !sh_mfn_is_dirty(d, target_mfn) )
- sflags &= ~_PAGE_RW;
+ if ( mfn_valid(target_mfn) ) {
+ if ( ft & FETCH_TYPE_WRITE )
+ paging_mark_dirty(d, mfn_x(target_mfn));
+ else if ( !sh_mfn_is_dirty(d, target_mfn) )
+ sflags &= ~_PAGE_RW;
+ }
}
/* Read-only memory */
@@ -2836,7 +2841,8 @@ static int sh_page_fault(struct vcpu *v,
gfn = guest_l1e_get_gfn(gw.eff_l1e);
gmfn = gfn_to_mfn(d, gfn, &p2mt);
- if ( !p2m_is_valid(p2mt) || (!p2m_is_mmio(p2mt) && !mfn_valid(gmfn)) )
+ if ( !shadow_mode_translate(d) &&
+ (!p2m_is_valid(p2mt) || (!p2m_is_mmio(p2mt) && !mfn_valid(gmfn))) )
{
perfc_incr(shadow_fault_bail_bad_gfn);
SHADOW_PRINTK("BAD gfn=%"SH_PRI_gfn" gmfn=%"PRI_mfn"\n",
iomem_shadow
Description: Text document
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
Home