Follow on from CVE-2014-3571. This fixes the code that was the original source

of the crash due to p being NULL. Steve's fix prevents this situation from
occuring - however this is by no means obvious by looking at the code for
dtls1_get_record. This fix just makes things look a bit more sane.
Reviewed-by: NDr Stephen Henson <steve@openssl.org>

ssl/d1_pkt.c

@@ -679,7 +679,8 @@ again:
 		 * would be dropped unnecessarily.
 		 */
 		if (!(s->d1->listen && rr->type == SSL3_RT_HANDSHAKE &&
-		    *p == SSL3_MT_CLIENT_HELLO) &&
+		    s->packet_length > DTLS1_RT_HEADER_LENGTH &&
+		    s->packet[DTLS1_RT_HEADER_LENGTH] == SSL3_MT_CLIENT_HELLO) &&
 		    !dtls1_record_replay_check(s, bitmap))
 			{
 			rr->length = 0;