Bug 567429 - slapd didn't close connection and get into CLOSE_WAIT state

https://bugzilla.redhat.com/show_bug.cgi?id=567429
Resolves: bug 567429
Bug Description: slapd didn't close connection and get into CLOSE_WAIT state
Reviewed by: nhosoi (Thanks!)
Branch: HEAD
Fix Description: The JNDI code attached to the bug uses persistent search.
The connection pool code handles persistent searches differently than
regular connections.  The connection pool code was acquiring a reference
to a conn, but was not releasing it in the persistent search case, assuming
the persistent search code did not also have a reference, but it does.  This
caused connection_table_move_connection_out_of_active_list() to not move
the connection out of the active list, and therefore available for closing,
because there was an outstanding reference.  The solution is for the
connection pool code to release its reference.
Platforms tested: RHEL5 x86_64
Flag Day: no
Doc impact: no

1 files changed, 11 insertions, 0 deletions

diff --git a/ldap/servers/slapd/connection.c b/ldap/servers/slapd/connection.c
index 8686d16e..bab33453 100644
--- a/ldap/servers/slapd/connection.c
+++ b/ldap/servers/slapd/connection.c
@@ -991,6 +991,13 @@ done:
 		}
 	    PR_Unlock( conn->c_mutex );
 
+	} else { /* ps code acquires ref to conn - we need to release ours here */
+	    PR_Lock( conn->c_mutex );
+        if (connection_release_nolock (conn) != 0)
+	    {
+			return_value = -1;
+		}
+	    PR_Unlock( conn->c_mutex );
 	}
 	return return_value;
 }
@@ -2314,6 +2321,10 @@ done:
 				connection_release_nolock (conn);
 			}
 		    PR_Unlock( conn->c_mutex );
+		} else { /* the ps code acquires a ref to the conn - we need to release ours here */
+		    PR_Lock( conn->c_mutex );
+			connection_release_nolock (conn);
+		    PR_Unlock( conn->c_mutex );
 		}
 		/* Since we didn't do so earlier, we need to make a replication connection readable again here */
 		if ( ((1 == is_timedout) || (replication_connection && !thread_turbo_flag)) && !more_data)