[svn-commits] mmichelson: branch group/CCSS r219894 - /team/group/CCSS/main/ccss.c

[SVN commits to the Digium repositories]

Author: mmichelson
Date: Wed Sep 23 10:41:46 2009
New Revision: 219894

URL: http://svnview.digium.com/svn/asterisk?view=rev&rev=219894
Log:
Clean up some comments.


Modified:
    team/group/CCSS/main/ccss.c

Modified: team/group/CCSS/main/ccss.c
URL: http://svnview.digium.com/svn/asterisk/team/group/CCSS/main/ccss.c?view=diff&rev=219894&r1=219893&r2=219894
==============================================================================
--- team/group/CCSS/main/ccss.c (original)
+++ team/group/CCSS/main/ccss.c Wed Sep 23 10:41:46 2009
@@ -1100,27 +1100,16 @@
  * Since this executes in the core taskprocessor thread,
  * there is no need to worry about doing any locking of
  * lists or links.
+ *
+ * Note that the monitor passed to this function has had
+ * its refcount increased by ast_cc_monitor_announce_availability,
+ * so it is essential that the refcount is decremented prior
+ * to returning.
+ *
  * \param datap pointer to the monitor reporting the state change
  */
 static int cc_devstate_change(void *datap)
 {
-	/* Let's map this out a bit.
-	   The device being monitored has
-	   had a device state change.
-
-	   According to architecture doc, the device monitor
-	   should report the state change to a parent extension
-	   monitor. In choosing which one to report the change
-	   to, it should choose the lowest weighted unsuspended
-	   link. That's easy enough to find.
-
-	   It's important that once we have picked a link, we
-	   need to continue feeding information to that link
-	   until it goes away or becomes suspended. So the
-	   monitor is going to keep a pointer to whichever
-	   link it has chosen. This will change if the link
-	   gets destroyed or becomes suspended.
-	 */
 	struct ast_cc_monitor *monitor = datap;
 	struct ast_cc_monitor_link *link = monitor->saved_link;
 
@@ -1130,12 +1119,10 @@
 		return 0;
 	}
 
-	/* Dang, we need to find the lowest weighted unsuspended
-	   link. Link weights are established at the time that CC
-	   requests are issued. Since weights increase chronologically
-	   and links are always added to the tail of the list, we
-	   know that links at the head are weighted lower than links
-	   at the end of the list.
+	/* Since links are established chronologically, and new links are
+	 * appended to the end of the list, the list is automatically sorted
+	 * with the higher priority links at the front and the lower
+	 * priority links at the back.
 	 */
 	AST_LIST_TRAVERSE(&monitor->parent_links, link, next_parent) {
 		if (!link->is_suspended) {
@@ -1144,17 +1131,14 @@
 	}
 
 	if (!link) {
-		/* Hmm, all upstream links are suspended. Nothing wrong with
-		   that. It just means that there's no one to report this
-		   change to. Since we've saved the current state on the
-		   monitor, we'll be able to report this state upstream
-		   as soon as one of the links becomes unsuspended.
+		/* Either there are no upstream links, or all upstream links
+		 * are suspended. Either way, this isn't a particularly big
+		 * deal.
 		 */
 		ao2_t_ref(monitor, -1, "Kill reference from ast_cc_monitor_announce_availabilty");
 		return 0;
 	}
 
-	/* We found a link to report on. Yay */
 	monitor->saved_link = link;
 	pass_availability_up(link->parent, link->core_id);
 	ao2_t_ref(monitor, -1, "Kill reference from ast_cc_monitor_announce_availabilty");