[svn-commits] mmichelson: trunk r430715 - in /trunk: ./ channels/ res/

SVN commits to the Digium repositories svn-commits at lists.digium.com
Fri Jan 16 16:13:25 CST 2015


Author: mmichelson
Date: Fri Jan 16 16:13:23 2015
New Revision: 430715

URL: http://svnview.digium.com/svn/asterisk?view=rev&rev=430715
Log:
Fix problem where a hung channel could occur on a failed blind transfer.

Different clients react differently to being told that a blind transfer
has failed. Some will simply send a BYE and be done with it. Others will
attempt to reinvite themselves back onto the call.

In the latter case, we were creating a new channel and then leaving it to
sit forever doing nothing. With this code change, that new channel will
not be created and the dialog with the transferring channel will be cleaned
up properly.

ASTERISK-24624 #close
Reported by Zane Conkle

Review: https://reviewboard.asterisk.org/r/4339
........

Merged revisions 430714 from http://svn.asterisk.org/svn/asterisk/branches/13

Modified:
    trunk/   (props changed)
    trunk/channels/chan_pjsip.c
    trunk/res/res_pjsip_session.c

Propchange: trunk/
------------------------------------------------------------------------------
Binary property 'branch-13-merged' - no diff available.

Modified: trunk/channels/chan_pjsip.c
URL: http://svnview.digium.com/svn/asterisk/trunk/channels/chan_pjsip.c?view=diff&rev=430715&r1=430714&r2=430715
==============================================================================
--- trunk/channels/chan_pjsip.c (original)
+++ trunk/channels/chan_pjsip.c Fri Jan 16 16:13:23 2015
@@ -2063,6 +2063,21 @@
 		return 0;
 	}
 
+	if (session->inv_session->state >= PJSIP_INV_STATE_CONFIRMED) {
+		/* Weird case. We've received a reinvite but we don't have a channel. The most
+		 * typical case for this happening is that a blind transfer fails, and so the
+		 * transferer attempts to reinvite himself back into the call. We already got
+		 * rid of that channel, and the other side of the call is unrecoverable.
+		 *
+		 * We treat this as a failure, so our best bet is to just hang this call
+		 * up and not create a new channel. Clearing defer_terminate here ensures that
+		 * calling ast_sip_session_terminate() can result in a BYE being sent ASAP.
+		 */
+		session->defer_terminate = 0;
+		ast_sip_session_terminate(session, 400);
+		return -1;
+	}
+
 	datastore = ast_sip_session_alloc_datastore(&transport_info, "transport_info");
 	if (!datastore) {
 		return -1;

Modified: trunk/res/res_pjsip_session.c
URL: http://svnview.digium.com/svn/asterisk/trunk/res/res_pjsip_session.c?view=diff&rev=430715&r1=430714&r2=430715
==============================================================================
--- trunk/res/res_pjsip_session.c (original)
+++ trunk/res/res_pjsip_session.c Fri Jan 16 16:13:23 2015
@@ -779,7 +779,8 @@
 
 	if (rdata->msg_info.msg->line.req.method.id != PJSIP_INVITE_METHOD ||
 		!(dlg = pjsip_ua_find_dialog(&rdata->msg_info.cid->id, &rdata->msg_info.to->tag, &rdata->msg_info.from->tag, PJ_FALSE)) ||
-		!(session = ast_sip_dialog_get_session(dlg))) {
+		!(session = ast_sip_dialog_get_session(dlg)) ||
+		!session->channel) {
 		return PJ_FALSE;
 	}
 




More information about the svn-commits mailing list