Fix Crash Caused By One-way Audio With auto_* NAT Settings Fix

The prior code committed, r385473, failed to take into consideration that not
all outgoing calls will be to a peer.  My fault.

This patch does the following:

* Check if there is a related peer involved.  If there is, check and set NAT 
  settings according to the peer's settings.

* Fix a problem with realtime peers.  If the global setting has auto_force_rport
  set and we issued a "sip reload" while a peer is still registered, the peer's
  flags for NAT are reset to off.  When this happens, we were always setting the
  contact address of the peer to that of the full contact info that we had.

(closes issue ASTERISK-21374)
Reported by: jmls
Tested by: Michael L. Young
Patches:
   asterisk-21374-fix-crash-and-rt-peers.diff by Michael L. Young (license 5026)

Review: https://reviewboard.asterisk.org/r/2524/
........

Merged revisions 388601 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@388602 65c4cc65-6c06-0410-ace0-fbb531ad65f3

1 files changed, 19 insertions, 11 deletions

diff --git a/channels/chan_sip.c b/channels/chan_sip.c
index 9b7c14be7..7c49f8841 100644
--- a/channels/chan_sip.c
+++ b/channels/chan_sip.c
@@ -30108,18 +30108,25 @@ static struct ast_channel *sip_request_call(const char *type, struct ast_format_
 		ast_string_field_set(p, peername, ext);
 	/* Recalculate our side, and recalculate Call ID */
 	ast_sip_ouraddrfor(&p->sa, &p->ourip, p);
-	/* When chan_sip is first loaded, we may have a peer entry but it hasn't re-registered yet.
-	   If the peer hasn't re-registered, we have not checked for NAT yet.  With the new
-	   auto_* settings, we need to check for NAT so we do not have one-way audio. */
-	check_for_nat(&p->ourip, p);
-	set_peer_nat(p, p->relatedpeer);
+	/* When chan_sip is first loaded or reloaded, we need to check for NAT and set the appropiate flags
+	   now that we have the auto_* settings. */
+	check_for_nat(&p->sa, p);
+	/* If there is a peer related to this outgoing call and it hasn't re-registered after
+	   a reload, we need to set the peer's NAT flags accordingly. */
+	if (p->relatedpeer) {
 
-	if (p->natdetected && ast_test_flag(&p->flags[2], SIP_PAGE3_NAT_AUTO_RPORT)) {
-		ast_copy_flags(&p->flags[0], &p->relatedpeer->flags[0], SIP_NAT_FORCE_RPORT);
-	}
+		if (!ast_strlen_zero(p->relatedpeer->fullcontact) && !p->natdetected &&
+			(ast_test_flag(&p->flags[2], SIP_PAGE3_NAT_AUTO_RPORT) && !ast_test_flag(&p->flags[0], SIP_NAT_FORCE_RPORT))) {
+			/* We need to make an attempt to determine if a peer is behind NAT
+			   if the peer has the auto_force_rport flag set. */
+			struct ast_sockaddr tmpaddr;
+
+			__set_address_from_contact(p->relatedpeer->fullcontact, &tmpaddr, 0);
+
+			check_for_nat(&tmpaddr, p);
+		}
 
-	if (p->natdetected && ast_test_flag(&p->flags[2], SIP_PAGE3_NAT_AUTO_COMEDIA)) {
-		ast_copy_flags(&p->flags[1], &p->relatedpeer->flags[1], SIP_PAGE2_SYMMETRICRTP);
+		set_peer_nat(p, p->relatedpeer);
 	}
 
 	do_setnat(p);
@@ -31340,7 +31347,8 @@ static struct sip_peer *build_peer(const char *name, struct ast_variable *v, str
 		 * specified, use that address instead. */
 		/* XXX May need to revisit the final argument; does the realtime DB store whether
 		 * the original contact was over TLS or not? XXX */
-		if (!ast_test_flag(&peer->flags[0], SIP_NAT_FORCE_RPORT) || ast_sockaddr_isnull(&peer->addr)) {
+		if ((!ast_test_flag(&peer->flags[2],  SIP_PAGE3_NAT_AUTO_RPORT) && !ast_test_flag(&peer->flags[0], SIP_NAT_FORCE_RPORT))
+		    || ast_sockaddr_isnull(&peer->addr)) {
 			__set_address_from_contact(ast_str_buffer(fullcontact), &peer->addr, 0);
 		}
 	}