[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] [PATCH 2 of 3] Ensure liveness of pages involved in a guest page table walk


  • To: xen-devel@xxxxxxxxxxxxxxxxxxx
  • From: Andres Lagar-Cavilla <andres@xxxxxxxxxxxxxxxx>
  • Date: Thu, 24 Nov 2011 08:41:32 -0500
  • Cc: andres@xxxxxxxxxxxxxx, keir.xen@xxxxxxxxx, tim@xxxxxxx, JBeulich@xxxxxxxx, adin@xxxxxxxxxxxxxx
  • Delivery-date: Thu, 24 Nov 2011 13:48:45 +0000
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=lagarcavilla.org; h=content-type :mime-version:content-transfer-encoding:subject:message-id :in-reply-to:references:date:from:to:cc; q=dns; s= lagarcavilla.org; b=WsopyjACrdeoOCkgQ9XGu3WF3TtKqBYGiCU9JE9GFMz6 G9QquZIBDmqX+o8ha6pYyaK+D59pVz58Idbes9Hsjt1zHj/IBxpukZJK36wqLEcy HRFH6ncV2ZvtrCpuGrZtb2C2KxcEscXHq1x33uKBCspN5cAJ+Mrl237YtDvV1LE=
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>

 xen/arch/x86/mm/guest_walk.c |  16 +++++++++++-----
 1 files changed, 11 insertions(+), 5 deletions(-)


Instead of risking deadlock by holding onto the gfn's acquired during
a guest page table walk, acquire an extra reference within the get_gfn/
put_gfn critical section, and drop the extra reference when done with
the map. This ensures liveness of the map, i.e. the underlying page
won't be paged out.

Signed-off-by: Andres Lagar-Cavilla <andres@xxxxxxxxxxxxxxxx>

diff -r f079cbeae77e -r b0410cb27d17 xen/arch/x86/mm/guest_walk.c
--- a/xen/arch/x86/mm/guest_walk.c
+++ b/xen/arch/x86/mm/guest_walk.c
@@ -87,7 +87,7 @@ static uint32_t set_ad_bits(void *guest_
 }
 
 /* If the map is non-NULL, we leave this function having 
- * called get_gfn, you need to call put_gfn. */
+ * acquired an extra ref on mfn_to_page(*mfn) */
 static inline void *map_domain_gfn(struct p2m_domain *p2m,
                                    gfn_t gfn, 
                                    mfn_t *mfn,
@@ -95,6 +95,7 @@ static inline void *map_domain_gfn(struc
                                    uint32_t *rc) 
 {
     p2m_access_t p2ma;
+    void *map;
 
     /* Translate the gfn, unsharing if shared */
     *mfn = get_gfn_type_access(p2m, gfn_x(gfn), p2mt, &p2ma, p2m_unshare, 
NULL);
@@ -120,7 +121,12 @@ static inline void *map_domain_gfn(struc
     }
     ASSERT(mfn_valid(mfn_x(*mfn)));
     
-    return map_domain_page(mfn_x(*mfn));
+    /* Get an extra ref to the page to ensure liveness of the map.
+     * Then we can safely put gfn */
+    page_get_owner_and_reference(mfn_to_page(mfn_x(*mfn)));
+    map = map_domain_page(mfn_x(*mfn));
+    __put_gfn(p2m, gfn_x(gfn));
+    return map;
 }
 
 
@@ -357,20 +363,20 @@ set_ad:
     if ( l3p ) 
     {
         unmap_domain_page(l3p);
-        __put_gfn(p2m, gfn_x(guest_l4e_get_gfn(gw->l4e)));
+        put_page(mfn_to_page(mfn_x(gw->l3mfn)));
     }
 #endif
 #if GUEST_PAGING_LEVELS >= 3
     if ( l2p ) 
     {
         unmap_domain_page(l2p);
-        __put_gfn(p2m, gfn_x(guest_l3e_get_gfn(gw->l3e))); 
+        put_page(mfn_to_page(mfn_x(gw->l2mfn)));
     }
 #endif
     if ( l1p ) 
     {
         unmap_domain_page(l1p);
-        __put_gfn(p2m, gfn_x(guest_l2e_get_gfn(gw->l2e)));
+        put_page(mfn_to_page(mfn_x(gw->l1mfn)));
     }
 
     return rc;

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.