[prev in list] [next in list] [prev in thread] [next in thread]
List: oss-security
Subject: [oss-security] Xen Security Advisory 248 - x86 PV guests may gain access to internally used pages
From: Xen.org security team <security () xen ! org>
Date: 2017-12-12 12:00:06
Message-ID: E1eOjEE-0002Bc-Qe () xenbits ! xenproject ! org
[Download RAW message or body]
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
Xen Security Advisory XSA-248
version 2
x86 PV guests may gain access to internally used pages
UPDATES IN VERSION 2
====================
Public release.
Provide metadata file.
ISSUE DESCRIPTION
=================
Memory management for PV guests builds on page ownership and page
attributes. A domain can always map, at least r/o, pages of which it
is the owner. Certain fields in the control structure of a page are
used for different purposes in the main PV memory management code and
in code handling shadow paging.
When a guest is running in shadow mode (which for PV guests is necessary
e.g. for live migration), certain auxiliary pages used by Xen internally
had their owner set to the guest itself. When the PV guest maps such a
page, shadow code and PV memory management code will disagree on the
meaning of said multi-purpose fields, generally leading to a crash of
the hypervisor.
IMPACT
======
A malicious or buggy PV guest may cause a hypervisor crash, resulting in
a Denial of Service (DoS) affecting the entire host, or cause hypervisor
memory corruption. We cannot rule out a guest being able to escalate
its privilege.
VULNERABLE SYSTEMS
==================
All versions of Xen are vulnerable.
Only x86 systems are affected. ARM systems are not vulnerable.
x86 HVM guests cannot exploit this vulnerability.
Only x86 PV guests can exploit this vulnerability, and only when being
run in shadow mode. PV guests are typically run in shadow mode for live
migration, as well as for features like VM snapshot.
Note that save / restore does *not* use shadow mode, and so does not
expose this vulnerability. Some downstreams also include a "non-live
migration" feature, which also does not use shadow mode (and thus does
not expose this vulnerability).
MITIGATION
==========
Running only HVM guests avoids the vulnerability.
Avoiding live migration of x86 PV guests also avoids the vulnerability.
CREDITS
=======
This issue was discovered by Jan Beulich of SUSE.
RESOLUTION
==========
Applying the appropriate attached patch resolves this issue.
xsa248.patch xen-unstable, Xen 4.9.x
xsa248-4.8.patch Xen 4.8.x, Xen 4.7.x, Xen 4.6.x
xsa248-4.5.patch Xen 4.5.x
$ sha256sum xsa248*
f0ac5c5ff956118f52821e111c6e27416f788cea6e98cc54cb051c42b793357e xsa248.meta
20bcfb1890d90bd74f52e45a1e8aa020a8991e3a0db37eecf53ce48b16e602bf xsa248.patch
ec4227633df18f76fbd8cb12e367879470b63fb5236f10b2a971dccef9f83172 xsa248-4.5.patch
3bbd9fd92e5ffab1ddd7ff804bfbab09c1c654af3aa7f80f742f321da120b715 xsa248-4.8.patch
$
DEPLOYMENT DURING EMBARGO
=========================
Deployment of the patches and/or mitigations described above (or
others which are substantially similar) is permitted during the
embargo, even on public-facing systems with untrusted guest users and
administrators.
But: Distribution of updated software is prohibited (except to other
members of the predisclosure list).
Predisclosure list members who wish to deploy significantly different
patches and/or mitigations, please contact the Xen Project Security
Team.
(Note: this during-embargo deployment notice is retained in
post-embargo publicly released Xen Project advisories, even though it
is then no longer applicable. This is to enable the community to have
oversight of the Xen Project Security Team's decisionmaking.)
For more information about permissible uses of embargoed information,
consult the Xen Project community's agreed Security Policy:
http://www.xenproject.org/security-policy.html
-----BEGIN PGP SIGNATURE-----
iQFABAEBCAAqFiEEI+MiLBRfRHX6gGCng/4UyVfoK9kFAlovuMMMHHBncEB4ZW4u
b3JnAAoJEIP+FMlX6CvZy20H/0z0FHbgG05yJtRNRZZ6p/YuygBi7yiuddoa1BTX
jVHYBd6TBw577qVYCGv77+ta4RhZdmg+qjVayzMQy0r07maq8jbNse7bTfjkbS8L
levYk0Yjr03jDRWW4//WurV9vlxgoTRGSjz3RlbqPPC/ugpZbj8VrrqOxqhV5dhR
umZSXIFQroZrDsHeAl+or84h+psSvGYcUnuMI/ML96hBqUjVi/owLIPhJw3OzK1i
VBQTuBAVWYR1CVIGq1KQCEpKD0NbRC3M0+cxDibdhwH+Md0O947m0W8fA5p3hYdg
e8CwuA0pk6CVKkFummEcC5FvwQStyVZjgG+X7aOwXobsMlA=
=aj+p
-----END PGP SIGNATURE-----
["xsa248.meta" (application/octet-stream)]
["xsa248.patch" (application/octet-stream)]
From: Jan Beulich <jbeulich@suse.com>
Subject: x86/mm: don't wrongly set page ownership
PV domains can obtain mappings of any pages owned by the correct domain,
including ones that aren't actually assigned as "normal" RAM, but used
by Xen internally. At the moment such "internal" pages marked as owned
by a guest include pages used to track logdirty bits, as well as p2m
pages and the "unpaged pagetable" for HVM guests. Since the PV memory
management and shadow code conflict in their use of struct page_info
fields, and since shadow code is being used for log-dirty handling for
PV domains, pages coming from the shadow pool must, for PV domains, not
have the domain set as their owner.
While the change could be done conditionally for just the PV case in
shadow code, do it unconditionally (and for consistency also for HAP),
just to be on the safe side.
There's one special case though for shadow code: The page table used for
running a HVM guest in unpaged mode is subject to get_page() (in
set_shadow_status()) and hence must have its owner set.
This is XSA-248.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Tim Deegan <tim@xen.org>
Reviewed-by: George Dunlap <george.dunlap@citrix.com>
---
v2: Drop PGC_page_table related pieces.
--- a/xen/arch/x86/mm/hap/hap.c
+++ b/xen/arch/x86/mm/hap/hap.c
@@ -286,8 +286,7 @@ static struct page_info *hap_alloc_p2m_p
{
d->arch.paging.hap.total_pages--;
d->arch.paging.hap.p2m_pages++;
- page_set_owner(pg, d);
- pg->count_info |= 1;
+ ASSERT(!page_get_owner(pg) && !(pg->count_info & PGC_count_mask));
}
else if ( !d->arch.paging.p2m_alloc_failed )
{
@@ -302,21 +301,23 @@ static struct page_info *hap_alloc_p2m_p
static void hap_free_p2m_page(struct domain *d, struct page_info *pg)
{
+ struct domain *owner = page_get_owner(pg);
+
/* This is called both from the p2m code (which never holds the
* paging lock) and the log-dirty code (which always does). */
paging_lock_recursive(d);
- ASSERT(page_get_owner(pg) == d);
- /* Should have just the one ref we gave it in alloc_p2m_page() */
- if ( (pg->count_info & PGC_count_mask) != 1 ) {
- HAP_ERROR("Odd p2m page %p count c=%#lx t=%"PRtype_info"\n",
- pg, pg->count_info, pg->u.inuse.type_info);
+ /* Should still have no owner and count zero. */
+ if ( owner || (pg->count_info & PGC_count_mask) )
+ {
+ HAP_ERROR("d%d: Odd p2m page %"PRI_mfn" d=%d c=%lx t=%"PRtype_info"\n",
+ d->domain_id, mfn_x(page_to_mfn(pg)),
+ owner ? owner->domain_id : DOMID_INVALID,
+ pg->count_info, pg->u.inuse.type_info);
WARN();
+ pg->count_info &= ~PGC_count_mask;
+ page_set_owner(pg, NULL);
}
- pg->count_info &= ~PGC_count_mask;
- /* Free should not decrement domain's total allocation, since
- * these pages were allocated without an owner. */
- page_set_owner(pg, NULL);
d->arch.paging.hap.p2m_pages--;
d->arch.paging.hap.total_pages++;
hap_free(d, page_to_mfn(pg));
--- a/xen/arch/x86/mm/shadow/common.c
+++ b/xen/arch/x86/mm/shadow/common.c
@@ -1503,32 +1503,29 @@ shadow_alloc_p2m_page(struct domain *d)
pg = mfn_to_page(shadow_alloc(d, SH_type_p2m_table, 0));
d->arch.paging.shadow.p2m_pages++;
d->arch.paging.shadow.total_pages--;
+ ASSERT(!page_get_owner(pg) && !(pg->count_info & PGC_count_mask));
paging_unlock(d);
- /* Unlike shadow pages, mark p2m pages as owned by the domain.
- * Marking the domain as the owner would normally allow the guest to
- * create mappings of these pages, but these p2m pages will never be
- * in the domain's guest-physical address space, and so that is not
- * believed to be a concern. */
- page_set_owner(pg, d);
- pg->count_info |= 1;
return pg;
}
static void
shadow_free_p2m_page(struct domain *d, struct page_info *pg)
{
- ASSERT(page_get_owner(pg) == d);
- /* Should have just the one ref we gave it in alloc_p2m_page() */
- if ( (pg->count_info & PGC_count_mask) != 1 )
+ struct domain *owner = page_get_owner(pg);
+
+ /* Should still have no owner and count zero. */
+ if ( owner || (pg->count_info & PGC_count_mask) )
{
- SHADOW_ERROR("Odd p2m page count c=%#lx t=%"PRtype_info"\n",
+ SHADOW_ERROR("d%d: Odd p2m page %"PRI_mfn" d=%d c=%lx t=%"PRtype_info"\n",
+ d->domain_id, mfn_x(page_to_mfn(pg)),
+ owner ? owner->domain_id : DOMID_INVALID,
pg->count_info, pg->u.inuse.type_info);
+ pg->count_info &= ~PGC_count_mask;
+ page_set_owner(pg, NULL);
}
- pg->count_info &= ~PGC_count_mask;
pg->u.sh.type = SH_type_p2m_table; /* p2m code reuses type-info */
- page_set_owner(pg, NULL);
/* This is called both from the p2m code (which never holds the
* paging lock) and the log-dirty code (which always does). */
@@ -3132,7 +3129,9 @@ int shadow_enable(struct domain *d, u32
e = __map_domain_page(pg);
write_32bit_pse_identmap(e);
unmap_domain_page(e);
+ pg->count_info = 1;
pg->u.inuse.type_info = PGT_l2_page_table | 1 | PGT_validated;
+ page_set_owner(pg, d);
}
paging_lock(d);
@@ -3170,7 +3169,11 @@ int shadow_enable(struct domain *d, u32
if ( rv != 0 && !pagetable_is_null(p2m_get_pagetable(p2m)) )
p2m_teardown(p2m);
if ( rv != 0 && pg != NULL )
+ {
+ pg->count_info &= ~PGC_count_mask;
+ page_set_owner(pg, NULL);
shadow_free_p2m_page(d, pg);
+ }
domain_unpause(d);
return rv;
}
@@ -3279,7 +3282,22 @@ out:
/* Must be called outside the lock */
if ( unpaged_pagetable )
+ {
+ if ( page_get_owner(unpaged_pagetable) == d &&
+ (unpaged_pagetable->count_info & PGC_count_mask) == 1 )
+ {
+ unpaged_pagetable->count_info &= ~PGC_count_mask;
+ page_set_owner(unpaged_pagetable, NULL);
+ }
+ /* Complain here in cases where shadow_free_p2m_page() won't. */
+ else if ( !page_get_owner(unpaged_pagetable) &&
+ !(unpaged_pagetable->count_info & PGC_count_mask) )
+ SHADOW_ERROR("d%d: Odd unpaged pt %"PRI_mfn" c=%lx t=%"PRtype_info"\n",
+ d->domain_id, mfn_x(page_to_mfn(unpaged_pagetable)),
+ unpaged_pagetable->count_info,
+ unpaged_pagetable->u.inuse.type_info);
shadow_free_p2m_page(d, unpaged_pagetable);
+ }
}
void shadow_final_teardown(struct domain *d)
["xsa248-4.5.patch" (application/octet-stream)]
From: Jan Beulich <jbeulich@suse.com>
Subject: x86/mm: don't wrongly set page ownership
PV domains can obtain mappings of any pages owned by the correct domain,
including ones that aren't actually assigned as "normal" RAM, but used
by Xen internally. At the moment such "internal" pages marked as owned
by a guest include pages used to track logdirty bits, as well as p2m
pages and the "unpaged pagetable" for HVM guests. Since the PV memory
management and shadow code conflict in their use of struct page_info
fields, and since shadow code is being used for log-dirty handling for
PV domains, pages coming from the shadow pool must, for PV domains, not
have the domain set as their owner.
While the change could be done conditionally for just the PV case in
shadow code, do it unconditionally (and for consistency also for HAP),
just to be on the safe side.
There's one special case though for shadow code: The page table used for
running a HVM guest in unpaged mode is subject to get_page() (in
set_shadow_status()) and hence must have its owner set.
This is XSA-248.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Tim Deegan <tim@xen.org>
Reviewed-by: George Dunlap <george.dunlap@citrix.com>
--- a/xen/arch/x86/mm/hap/hap.c
+++ b/xen/arch/x86/mm/hap/hap.c
@@ -264,8 +264,7 @@ static struct page_info *hap_alloc_p2m_p
{
d->arch.paging.hap.total_pages--;
d->arch.paging.hap.p2m_pages++;
- page_set_owner(pg, d);
- pg->count_info |= 1;
+ ASSERT(!page_get_owner(pg) && !(pg->count_info & PGC_count_mask));
}
else if ( !d->arch.paging.p2m_alloc_failed )
{
@@ -280,21 +279,23 @@ static struct page_info *hap_alloc_p2m_p
static void hap_free_p2m_page(struct domain *d, struct page_info *pg)
{
+ struct domain *owner = page_get_owner(pg);
+
/* This is called both from the p2m code (which never holds the
* paging lock) and the log-dirty code (which always does). */
paging_lock_recursive(d);
- ASSERT(page_get_owner(pg) == d);
- /* Should have just the one ref we gave it in alloc_p2m_page() */
- if ( (pg->count_info & PGC_count_mask) != 1 ) {
- HAP_ERROR("Odd p2m page %p count c=%#lx t=%"PRtype_info"\n",
- pg, pg->count_info, pg->u.inuse.type_info);
+ /* Should still have no owner and count zero. */
+ if ( owner || (pg->count_info & PGC_count_mask) )
+ {
+ HAP_ERROR("d%d: Odd p2m page %"PRI_mfn" d=%d c=%lx t=%"PRtype_info"\n",
+ d->domain_id, mfn_x(page_to_mfn(pg)),
+ owner ? owner->domain_id : DOMID_INVALID,
+ pg->count_info, pg->u.inuse.type_info);
WARN();
+ pg->count_info &= ~PGC_count_mask;
+ page_set_owner(pg, NULL);
}
- pg->count_info &= ~PGC_count_mask;
- /* Free should not decrement domain's total allocation, since
- * these pages were allocated without an owner. */
- page_set_owner(pg, NULL);
d->arch.paging.hap.p2m_pages--;
d->arch.paging.hap.total_pages++;
hap_free(d, page_to_mfn(pg));
--- a/xen/arch/x86/mm/shadow/common.c
+++ b/xen/arch/x86/mm/shadow/common.c
@@ -1622,32 +1622,29 @@ shadow_alloc_p2m_page(struct domain *d)
pg = mfn_to_page(shadow_alloc(d, SH_type_p2m_table, 0));
d->arch.paging.shadow.p2m_pages++;
d->arch.paging.shadow.total_pages--;
+ ASSERT(!page_get_owner(pg) && !(pg->count_info & PGC_count_mask));
paging_unlock(d);
- /* Unlike shadow pages, mark p2m pages as owned by the domain.
- * Marking the domain as the owner would normally allow the guest to
- * create mappings of these pages, but these p2m pages will never be
- * in the domain's guest-physical address space, and so that is not
- * believed to be a concern. */
- page_set_owner(pg, d);
- pg->count_info |= 1;
return pg;
}
static void
shadow_free_p2m_page(struct domain *d, struct page_info *pg)
{
- ASSERT(page_get_owner(pg) == d);
- /* Should have just the one ref we gave it in alloc_p2m_page() */
- if ( (pg->count_info & PGC_count_mask) != 1 )
+ struct domain *owner = page_get_owner(pg);
+
+ /* Should still have no owner and count zero. */
+ if ( owner || (pg->count_info & PGC_count_mask) )
{
- SHADOW_ERROR("Odd p2m page count c=%#lx t=%"PRtype_info"\n",
+ SHADOW_ERROR("d%d: Odd p2m page %"PRI_mfn" d=%d c=%lx t=%"PRtype_info"\n",
+ d->domain_id, mfn_x(page_to_mfn(pg)),
+ owner ? owner->domain_id : DOMID_INVALID,
pg->count_info, pg->u.inuse.type_info);
+ pg->count_info &= ~PGC_count_mask;
+ page_set_owner(pg, NULL);
}
- pg->count_info &= ~PGC_count_mask;
pg->u.sh.type = SH_type_p2m_table; /* p2m code reuses type-info */
- page_set_owner(pg, NULL);
/* This is called both from the p2m code (which never holds the
* paging lock) and the log-dirty code (which always does). */
@@ -3012,7 +3009,9 @@ int shadow_enable(struct domain *d, u32
| _PAGE_PRESENT | _PAGE_RW | _PAGE_USER
| _PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_PSE);
sh_unmap_domain_page(e);
+ pg->count_info = 1;
pg->u.inuse.type_info = PGT_l2_page_table | 1 | PGT_validated;
+ page_set_owner(pg, d);
}
paging_lock(d);
@@ -3050,7 +3049,11 @@ int shadow_enable(struct domain *d, u32
if ( rv != 0 && !pagetable_is_null(p2m_get_pagetable(p2m)) )
p2m_teardown(p2m);
if ( rv != 0 && pg != NULL )
+ {
+ pg->count_info &= ~PGC_count_mask;
+ page_set_owner(pg, NULL);
shadow_free_p2m_page(d, pg);
+ }
domain_unpause(d);
return rv;
}
@@ -3161,7 +3164,22 @@ out:
/* Must be called outside the lock */
if ( unpaged_pagetable )
+ {
+ if ( page_get_owner(unpaged_pagetable) == d &&
+ (unpaged_pagetable->count_info & PGC_count_mask) == 1 )
+ {
+ unpaged_pagetable->count_info &= ~PGC_count_mask;
+ page_set_owner(unpaged_pagetable, NULL);
+ }
+ /* Complain here in cases where shadow_free_p2m_page() won't. */
+ else if ( !page_get_owner(unpaged_pagetable) &&
+ !(unpaged_pagetable->count_info & PGC_count_mask) )
+ SHADOW_ERROR("d%d: Odd unpaged pt %"PRI_mfn" c=%lx t=%"PRtype_info"\n",
+ d->domain_id, mfn_x(page_to_mfn(unpaged_pagetable)),
+ unpaged_pagetable->count_info,
+ unpaged_pagetable->u.inuse.type_info);
shadow_free_p2m_page(d, unpaged_pagetable);
+ }
}
void shadow_final_teardown(struct domain *d)
["xsa248-4.8.patch" (application/octet-stream)]
From: Jan Beulich <jbeulich@suse.com>
Subject: x86/mm: don't wrongly set page ownership
PV domains can obtain mappings of any pages owned by the correct domain,
including ones that aren't actually assigned as "normal" RAM, but used
by Xen internally. At the moment such "internal" pages marked as owned
by a guest include pages used to track logdirty bits, as well as p2m
pages and the "unpaged pagetable" for HVM guests. Since the PV memory
management and shadow code conflict in their use of struct page_info
fields, and since shadow code is being used for log-dirty handling for
PV domains, pages coming from the shadow pool must, for PV domains, not
have the domain set as their owner.
While the change could be done conditionally for just the PV case in
shadow code, do it unconditionally (and for consistency also for HAP),
just to be on the safe side.
There's one special case though for shadow code: The page table used for
running a HVM guest in unpaged mode is subject to get_page() (in
set_shadow_status()) and hence must have its owner set.
This is XSA-248.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Tim Deegan <tim@xen.org>
Reviewed-by: George Dunlap <george.dunlap@citrix.com>
--- a/xen/arch/x86/mm/hap/hap.c
+++ b/xen/arch/x86/mm/hap/hap.c
@@ -283,8 +283,7 @@ static struct page_info *hap_alloc_p2m_p
{
d->arch.paging.hap.total_pages--;
d->arch.paging.hap.p2m_pages++;
- page_set_owner(pg, d);
- pg->count_info |= 1;
+ ASSERT(!page_get_owner(pg) && !(pg->count_info & PGC_count_mask));
}
else if ( !d->arch.paging.p2m_alloc_failed )
{
@@ -299,21 +298,23 @@ static struct page_info *hap_alloc_p2m_p
static void hap_free_p2m_page(struct domain *d, struct page_info *pg)
{
+ struct domain *owner = page_get_owner(pg);
+
/* This is called both from the p2m code (which never holds the
* paging lock) and the log-dirty code (which always does). */
paging_lock_recursive(d);
- ASSERT(page_get_owner(pg) == d);
- /* Should have just the one ref we gave it in alloc_p2m_page() */
- if ( (pg->count_info & PGC_count_mask) != 1 ) {
- HAP_ERROR("Odd p2m page %p count c=%#lx t=%"PRtype_info"\n",
- pg, pg->count_info, pg->u.inuse.type_info);
+ /* Should still have no owner and count zero. */
+ if ( owner || (pg->count_info & PGC_count_mask) )
+ {
+ HAP_ERROR("d%d: Odd p2m page %"PRI_mfn" d=%d c=%lx t=%"PRtype_info"\n",
+ d->domain_id, mfn_x(page_to_mfn(pg)),
+ owner ? owner->domain_id : DOMID_INVALID,
+ pg->count_info, pg->u.inuse.type_info);
WARN();
+ pg->count_info &= ~PGC_count_mask;
+ page_set_owner(pg, NULL);
}
- pg->count_info &= ~PGC_count_mask;
- /* Free should not decrement domain's total allocation, since
- * these pages were allocated without an owner. */
- page_set_owner(pg, NULL);
d->arch.paging.hap.p2m_pages--;
d->arch.paging.hap.total_pages++;
hap_free(d, page_to_mfn(pg));
--- a/xen/arch/x86/mm/shadow/common.c
+++ b/xen/arch/x86/mm/shadow/common.c
@@ -1573,32 +1573,29 @@ shadow_alloc_p2m_page(struct domain *d)
pg = mfn_to_page(shadow_alloc(d, SH_type_p2m_table, 0));
d->arch.paging.shadow.p2m_pages++;
d->arch.paging.shadow.total_pages--;
+ ASSERT(!page_get_owner(pg) && !(pg->count_info & PGC_count_mask));
paging_unlock(d);
- /* Unlike shadow pages, mark p2m pages as owned by the domain.
- * Marking the domain as the owner would normally allow the guest to
- * create mappings of these pages, but these p2m pages will never be
- * in the domain's guest-physical address space, and so that is not
- * believed to be a concern. */
- page_set_owner(pg, d);
- pg->count_info |= 1;
return pg;
}
static void
shadow_free_p2m_page(struct domain *d, struct page_info *pg)
{
- ASSERT(page_get_owner(pg) == d);
- /* Should have just the one ref we gave it in alloc_p2m_page() */
- if ( (pg->count_info & PGC_count_mask) != 1 )
+ struct domain *owner = page_get_owner(pg);
+
+ /* Should still have no owner and count zero. */
+ if ( owner || (pg->count_info & PGC_count_mask) )
{
- SHADOW_ERROR("Odd p2m page count c=%#lx t=%"PRtype_info"\n",
+ SHADOW_ERROR("d%d: Odd p2m page %"PRI_mfn" d=%d c=%lx t=%"PRtype_info"\n",
+ d->domain_id, mfn_x(page_to_mfn(pg)),
+ owner ? owner->domain_id : DOMID_INVALID,
pg->count_info, pg->u.inuse.type_info);
+ pg->count_info &= ~PGC_count_mask;
+ page_set_owner(pg, NULL);
}
- pg->count_info &= ~PGC_count_mask;
pg->u.sh.type = SH_type_p2m_table; /* p2m code reuses type-info */
- page_set_owner(pg, NULL);
/* This is called both from the p2m code (which never holds the
* paging lock) and the log-dirty code (which always does). */
@@ -3216,7 +3213,9 @@ int shadow_enable(struct domain *d, u32
| _PAGE_PRESENT | _PAGE_RW | _PAGE_USER
| _PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_PSE);
unmap_domain_page(e);
+ pg->count_info = 1;
pg->u.inuse.type_info = PGT_l2_page_table | 1 | PGT_validated;
+ page_set_owner(pg, d);
}
paging_lock(d);
@@ -3254,7 +3253,11 @@ int shadow_enable(struct domain *d, u32
if ( rv != 0 && !pagetable_is_null(p2m_get_pagetable(p2m)) )
p2m_teardown(p2m);
if ( rv != 0 && pg != NULL )
+ {
+ pg->count_info &= ~PGC_count_mask;
+ page_set_owner(pg, NULL);
shadow_free_p2m_page(d, pg);
+ }
domain_unpause(d);
return rv;
}
@@ -3363,7 +3366,22 @@ out:
/* Must be called outside the lock */
if ( unpaged_pagetable )
+ {
+ if ( page_get_owner(unpaged_pagetable) == d &&
+ (unpaged_pagetable->count_info & PGC_count_mask) == 1 )
+ {
+ unpaged_pagetable->count_info &= ~PGC_count_mask;
+ page_set_owner(unpaged_pagetable, NULL);
+ }
+ /* Complain here in cases where shadow_free_p2m_page() won't. */
+ else if ( !page_get_owner(unpaged_pagetable) &&
+ !(unpaged_pagetable->count_info & PGC_count_mask) )
+ SHADOW_ERROR("d%d: Odd unpaged pt %"PRI_mfn" c=%lx t=%"PRtype_info"\n",
+ d->domain_id, mfn_x(page_to_mfn(unpaged_pagetable)),
+ unpaged_pagetable->count_info,
+ unpaged_pagetable->u.inuse.type_info);
shadow_free_p2m_page(d, unpaged_pagetable);
+ }
}
void shadow_final_teardown(struct domain *d)
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic