Adjustments to scheduling of Routing Engine work.

These changes mean that Quagga copes a little better when there are
very large numbers of updates/withrawal messages arriving all at
once...

...it is not possible to cure the problem of overloading Quagga by
throwing too much at it.  However, these changes at least mean that
when BGP sessions drop, the Routing Engine will notice that in a
reasonable time, and can clear up all routes associated with
the session (throwing away any updates/withdraws already received,
but not yet dealt with.)

Amonst these changes are Chris Caputo's patches for bgp_node locking
issues -- see quagga-dev 7960 mailing list message.

1 files changed, 18 insertions, 4 deletions

diff --git a/bgpd/bgp_nexthop.c b/bgpd/bgp_nexthop.c
index 7e8c82ee..c72ca09f 100644
--- a/bgpd/bgp_nexthop.c
+++ b/bgpd/bgp_nexthop.c
@@ -276,6 +276,8 @@ bgp_nexthop_lookup_ipv6 (struct peer *peer, struct bgp_info *ri, int *changed,
 
 		  if (bnc->metric != oldbnc->metric)
 		    bnc->metricchanged = 1;
+
+		  bgp_unlock_node (oldrn);
 		}
 	    }
 	}
@@ -365,6 +367,8 @@ bgp_nexthop_lookup (afi_t afi, struct peer *peer, struct bgp_info *ri,
 
 		  if (bnc->metric != oldbnc->metric)
 		    bnc->metricchanged = 1;
+
+                  bgp_unlock_node (oldrn);
 		}
 	    }
 	}
@@ -571,7 +575,7 @@ bgp_connected_add (struct connected *ifc)
 	}
       else
 	{
-	  bc = XCALLOC (0, sizeof (struct bgp_connected_ref));
+	  bc = XCALLOC (MTYPE_BGP_CONN, sizeof (struct bgp_connected_ref));
 	  bc->refcnt = 1;
 	  rn->info = bc;
 	}
@@ -596,7 +600,7 @@ bgp_connected_add (struct connected *ifc)
 	}
       else
 	{
-	  bc = XCALLOC (0, sizeof (struct bgp_connected_ref));
+	  bc = XCALLOC (MTYPE_BGP_CONN, sizeof (struct bgp_connected_ref));
 	  bc->refcnt = 1;
 	  rn->info = bc;
 	}
@@ -636,7 +640,7 @@ bgp_connected_delete (struct connected *ifc)
       bc->refcnt--;
       if (bc->refcnt == 0)
 	{
-	  XFREE (0, bc);
+	  XFREE (MTYPE_BGP_CONN, bc);
 	  rn->info = NULL;
 	}
       bgp_unlock_node (rn);
@@ -662,7 +666,7 @@ bgp_connected_delete (struct connected *ifc)
       bc->refcnt--;
       if (bc->refcnt == 0)
 	{
-	  XFREE (0, bc);
+	  XFREE (MTYPE_BGP_CONN, bc);
 	  rn->info = NULL;
 	}
       bgp_unlock_node (rn);
@@ -1113,11 +1117,15 @@ bgp_multiaccess_check_v4 (struct in_addr nexthop, char *peer)
   rn1 = bgp_node_match (bgp_connected_table[AFI_IP], &p1);
   if (! rn1)
     return 0;
+  bgp_unlock_node (rn1);
 
   rn2 = bgp_node_match (bgp_connected_table[AFI_IP], &p2);
   if (! rn2)
     return 0;
+  bgp_unlock_node (rn2);
 
+  /* This is safe, even with above unlocks, since we are just
+     comparing pointers to the objects, not the objects themselves. */
   if (rn1 == rn2)
     return 1;
 
@@ -1296,6 +1304,9 @@ bgp_scan_init (void)
 void
 bgp_scan_finish (void)
 {
+  /* Only the current one needs to be reset. */
+  bgp_nexthop_cache_reset (bgp_nexthop_cache_table[AFI_IP]);
+
   bgp_table_unlock (cache1_table[AFI_IP]);
   cache1_table[AFI_IP] = NULL;
 
@@ -1306,6 +1317,9 @@ bgp_scan_finish (void)
   bgp_connected_table[AFI_IP] = NULL;
 
 #ifdef HAVE_IPV6
+  /* Only the current one needs to be reset. */
+  bgp_nexthop_cache_reset (bgp_nexthop_cache_table[AFI_IP6]);
+
   bgp_table_unlock (cache1_table[AFI_IP6]);
   cache1_table[AFI_IP6] = NULL;