gnttab: add preemption check to gnttab_release_mappings()

A guest may die with many grant mappings still in place, or simply with
a large maptrack table. Iterating through this may take more time than
is reasonable without intermediate preemption (to run softirqs and
perhaps the scheduler).

Move the invocation of the function to the section where other
restartable functions get invoked, and have the function itself check
for preemption every once in a while. Have it iterate the table
backwards, such that decreasing the maptrack limit is all it takes to
convey restart information.

In domain_teardown() introduce PROG_none such that inserting at the
front will be easier going forward.

This is part of CVE-2021-28698 / XSA-380.

Reported-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Julien Grall <jgrall@amazon.com>
master commit: b1ee10be5625b7d502cef1e6ee3818610ab0d29c
master date: 2021-08-25 14:18:18 +0200

diff --git a/xen/common/domain.c b/xen/common/domain.c
index 0a7f193e43869569bf45e821f02c9cd6ef4c9342..4d42caf9a9c1106a4d36d411cd5c28475ebbc070 100644 (file)
--- a/xen/common/domain.c
+++ b/xen/common/domain.c
@@ -646,13 +646,15 @@ int domain_kill(struct domain *d)
         if ( d->is_dying != DOMDYING_alive )
             return domain_kill(d);
         d->is_dying = DOMDYING_dying;
-        gnttab_release_mappings(d);
         tmem_destroy(d->tmem_client);
         vnuma_destroy(d->vnuma);
         domain_set_outstanding_pages(d, 0);
         d->tmem_client = NULL;
         /* fallthrough */
     case DOMDYING_dying:
+        rc = gnttab_release_mappings(d);
+        if ( rc )
+            break;
         rc = evtchn_destroy(d);
         if ( rc )
             break;

diff --git a/xen/common/grant_table.c b/xen/common/grant_table.c
index 7d464a96415730c5eec88eb0111e06503b5312b1..33a6c6591a640cb8afe4ddd06daa147b82724b71 100644 (file)
--- a/xen/common/grant_table.c
+++ b/xen/common/grant_table.c
@@ -62,7 +62,13 @@ struct grant_table {
     unsigned int          nr_grant_frames;
     /* Number of grant status frames shared with guest (for version 2) */
     unsigned int          nr_status_frames;
-    /* Number of available maptrack entries. */
+    /*
+     * Number of available maptrack entries.  For cleanup purposes it is
+     * important to realize that this field and @maptrack further down will
+     * only ever be accessed by the local domain.  Thus it is okay to clean
+     * up early, and to shrink the limit for the purpose of tracking cleanup
+     * progress.
+     */
     unsigned int          maptrack_limit;
     /* Shared grant table (see include/public/grant_table.h). */
     union {
@@ -3618,9 +3624,7 @@ grant_table_create(
     return ret;
 }
 
-void
-gnttab_release_mappings(
-    struct domain *d)
+int gnttab_release_mappings(struct domain *d)
 {
     struct grant_table   *gt = d->grant_table, *rgt;
     struct grant_mapping *map;
@@ -3634,8 +3638,32 @@ gnttab_release_mappings(
 
     BUG_ON(!d->is_dying);
 
-    for ( handle = 0; handle < gt->maptrack_limit; handle++ )
+    if ( !gt || !gt->maptrack )
+        return 0;
+
+    for ( handle = gt->maptrack_limit; handle; )
     {
+        /*
+         * Deal with full pages such that their freeing (in the body of the
+         * if()) remains simple.
+         */
+        if ( handle < gt->maptrack_limit && !(handle % MAPTRACK_PER_PAGE) )
+        {
+            /*
+             * Changing maptrack_limit alters nr_maptrack_frames()'es return
+             * value. Free the then excess trailing page right here, rather
+             * than leaving it to grant_table_destroy() (and in turn requiring
+             * to leave gt->maptrack_limit unaltered).
+             */
+            gt->maptrack_limit = handle;
+            FREE_XENHEAP_PAGE(gt->maptrack[nr_maptrack_frames(gt)]);
+
+            if ( hypercall_preempt_check() )
+                return -ERESTART;
+        }
+
+        --handle;
+
         map = &maptrack_entry(gt, handle);
         if ( !(map->flags & (GNTMAP_device_map|GNTMAP_host_map)) )
             continue;
@@ -3723,6 +3751,11 @@ gnttab_release_mappings(
 
         map->flags = 0;
     }
+
+    gt->maptrack_limit = 0;
+    FREE_XENHEAP_PAGE(gt->maptrack[0]);
+
+    return 0;
 }
 
 void grant_table_warn_active_grants(struct domain *d)
@@ -3785,8 +3818,7 @@ grant_table_destroy(
         free_xenheap_page(t->shared_raw[i]);
     xfree(t->shared_raw);
 
-    for ( i = 0; i < nr_maptrack_frames(t); i++ )
-        free_xenheap_page(t->maptrack[i]);
+    ASSERT(!t->maptrack_limit);
     vfree(t->maptrack);
 
     for ( i = 0; i < nr_active_grant_frames(t); i++ )

diff --git a/xen/include/xen/grant_table.h b/xen/include/xen/grant_table.h
index b3a95fda5860a5171da3723b1f2bc7a191f3d218..543bfcc7234573866a7d0abb5a99673698bf6194 100644 (file)
--- a/xen/include/xen/grant_table.h
+++ b/xen/include/xen/grant_table.h
@@ -46,9 +46,7 @@ int grant_table_set_limits(struct domain *d, unsigned int grant_frames,
 void grant_table_warn_active_grants(struct domain *d);
 
 /* Domain death release of granted mappings of other domains' memory. */
-void
-gnttab_release_mappings(
-    struct domain *d);
+int gnttab_release_mappings(struct domain *d);
 
 int mem_sharing_gref_to_gfn(struct grant_table *gt, grant_ref_t ref,
                             gfn_t *gfn, uint16_t *status);