[xen stable-4.5] x86/mm: don't wrongly set page ownership

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

[xen stable-4.5] x86/mm: don't wrongly set page ownership

patchbot
commit ab5ca36994f68c6a401eaf6f92d04babff4c322e
Author:     Jan Beulich <[hidden email]>
AuthorDate: Tue Dec 12 15:12:24 2017 +0100
Commit:     Jan Beulich <[hidden email]>
CommitDate: Tue Dec 12 15:12:24 2017 +0100

    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 <[hidden email]>
    Reviewed-by: Tim Deegan <[hidden email]>
    Reviewed-by: George Dunlap <[hidden email]>
    master commit: ff2a793e15bb0b6254bc849ef8e83e1c284c3583
    master date: 2017-12-12 14:28:36 +0100
---
 xen/arch/x86/mm/hap/hap.c       | 23 ++++++++++-----------
 xen/arch/x86/mm/shadow/common.c | 44 +++++++++++++++++++++++++++++------------
 2 files changed, 43 insertions(+), 24 deletions(-)

diff --git a/xen/arch/x86/mm/hap/hap.c b/xen/arch/x86/mm/hap/hap.c
index 84531b1..7b240bf 100644
--- 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_page(struct domain *d)
     {
         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_page(struct domain *d)
 
 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));
diff --git a/xen/arch/x86/mm/shadow/common.c b/xen/arch/x86/mm/shadow/common.c
index 4fb3de8..dea2b61 100644
--- 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 mode)
                     | _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 mode)
     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)
--
generated by git-patchbot for /home/xen/git/xen.git#stable-4.5

_______________________________________________
Xen-changelog mailing list
[hidden email]
https://lists.xenproject.org/xen-changelog