ospf6d: CVE-2011-3324 (DD LSA assertion)

This vulnerability (CERT-FI #514839) was reported by CROSS project.

When Database Description LSA header list contains trailing zero octets,
ospf6d tries to process this data as an LSA header. This triggers an
assertion in the code and ospf6d shuts down.

* ospf6_lsa.c
  * ospf6_lsa_is_changed(): handle header-only argument(s)
    appropriately, do not treat LSA length underrun as a fatal error.

1 files changed, 11 insertions, 1 deletions

diff --git a/ospf6d/ospf6_lsa.c b/ospf6d/ospf6_lsa.c
index 588b9462..e65752d8 100644
--- a/ospf6d/ospf6_lsa.c
+++ b/ospf6d/ospf6_lsa.c
@@ -163,9 +163,19 @@ ospf6_lsa_is_changed (struct ospf6_lsa *lsa1,
     return 1;
   if (ntohs (lsa1->header->length) != ntohs (lsa2->header->length))
     return 1;
+  /* Going beyond LSA headers to compare the payload only makes sense, when both LSAs aren't header-only. */
+  if (CHECK_FLAG (lsa1->flag, OSPF6_LSA_HEADERONLY) != CHECK_FLAG (lsa2->flag, OSPF6_LSA_HEADERONLY))
+  {
+    zlog_warn ("%s: only one of two (%s, %s) LSAs compared is header-only", __func__, lsa1->name, lsa2->name);
+    return 1;
+  }
+  if (CHECK_FLAG (lsa1->flag, OSPF6_LSA_HEADERONLY))
+    return 0;
 
   length = OSPF6_LSA_SIZE (lsa1->header) - sizeof (struct ospf6_lsa_header);
-  assert (length > 0);
+  /* Once upper layer verifies LSAs received, length underrun should become a warning. */
+  if (length <= 0)
+    return 0;
 
   return memcmp (OSPF6_LSA_HEADER_END (lsa1->header),
                  OSPF6_LSA_HEADER_END (lsa2->header), length);