[svn-commits] russell: branch 1.6.1 r140063 - in /branches/1.6.1: ./ channels/chan_sip.c

[SVN commits to the Digium repositories]

Author: russell
Date: Tue Aug 26 11:11:23 2008
New Revision: 140063

URL: http://svn.digium.com/view/asterisk?view=rev&rev=140063
Log:
Merged revisions 140061 via svnmerge from 
https://origsvn.digium.com/svn/asterisk/trunk

................
r140061 | russell | 2008-08-26 11:10:06 -0500 (Tue, 26 Aug 2008) | 14 lines

Merged revisions 140060 via svnmerge from 
https://origsvn.digium.com/svn/asterisk/branches/1.4

........
r140060 | russell | 2008-08-26 11:07:58 -0500 (Tue, 26 Aug 2008) | 6 lines

Fix some bogus scheduler usage in chan_sip.  This code used the return value
of a completely unrelated function to determine whether the scheduler should
be run or not.  This would have caused the scheduler to not run in cases where
it should have.  Also, leave a note about another scheduler issue that needs
to be addressed at some point.

........

................

Modified:
    branches/1.6.1/   (props changed)
    branches/1.6.1/channels/chan_sip.c

Propchange: branches/1.6.1/
------------------------------------------------------------------------------
Binary property 'trunk-merged' - no diff available.

Modified: branches/1.6.1/channels/chan_sip.c
URL: http://svn.digium.com/view/asterisk/branches/1.6.1/channels/chan_sip.c?view=diff&rev=140063&r1=140062&r2=140063
==============================================================================
--- branches/1.6.1/channels/chan_sip.c (original)
+++ branches/1.6.1/channels/chan_sip.c Tue Aug 26 11:11:23 2008
@@ -19546,6 +19546,12 @@
 		   dialog that was found and destroyed, probably because the list contents would change,
 		   so we'd need to restart. This isn't the best thing to do with callbacks. */
 
+		/* XXX TODO The scheduler usage in this module does not have sufficient 
+		 * synchronization being done between running the scheduler and places 
+		 * scheduling tasks.  As it is written, any scheduled item may not run 
+		 * any sooner than about  1 second, regardless of whether a sooner time 
+		 * was asked for. */
+
 		pthread_testcancel();
 		/* Wait for sched or io */
 		res = ast_sched_wait(sched);
@@ -19555,11 +19561,9 @@
 		if (res > 20)
 			ast_debug(1, "chan_sip: ast_io_wait ran %d all at once\n", res);
 		ast_mutex_lock(&monlock);
-		if (res >= 0)  {
-			res = ast_sched_runq(sched);
-			if (res >= 20)
-				ast_debug(1, "chan_sip: ast_sched_runq ran %d all at once\n", res);
-		}
+		res = ast_sched_runq(sched);
+		if (res >= 20)
+			ast_debug(1, "chan_sip: ast_sched_runq ran %d all at once\n", res);
 		ast_mutex_unlock(&monlock);
 	}