bgpd: fix withdraw stale routes

Incoming MPLS VPN staled routes are not flushing imported entries, upon expiration of the stalepath timer. The incoming MPLS VPN staled routes that are received from a given peer have a sub_type set to BGP_ROUTE_NORMAL. The original patch is wrong as it filters out the locally exported routes marked with the BGP_ROUTE_IMPORTED subtype. Fix this by extending the test to all MPLS VPN routes received from this peer. Fixes: 641065d ("bgpd: VRF-Lite fix to clear stale leaked routes") Signed-off-by: Philippe Guibert <[email protected]>
pguibert6WIND · Dec 4, 2023 · 5b41a0e · 5b41a0e
1 parent 6be9452
commit 5b41a0e
Showing 1 changed file with 1 addition and 2 deletions.
diff --git a/bgpd/bgp_route.c b/bgpd/bgp_route.c
@@ -5858,8 +5858,7 @@ void bgp_clear_stale_route(struct peer *peer, afi_t afi, safi_t safi)
 					 * If this is VRF leaked route
 					 * process for withdraw.
 					 */
-					if (pi->sub_type ==
-						    BGP_ROUTE_IMPORTED &&
+					if (safi == SAFI_MPLS_VPN &&
 					    peer->bgp->inst_type ==
 						    BGP_INSTANCE_TYPE_DEFAULT)
 						vpn_leak_to_vrf_withdraw(pi);