bgpd: fix missing bgp_attr_flush on errors in bgp_update

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>

diff --git a/bgpd/bgp_route.c b/bgpd/bgp_route.c
index 5bc2b8737edfe40a9ce5fed603f836417133d951..ac70155f162a75b26b94a98d9b90d3d001e40204 100644 (file)
--- a/bgpd/bgp_route.c
+++ b/bgpd/bgp_route.c
@@ -3930,6 +3930,7 @@ int bgp_update(struct peer *peer, const struct prefix *p, uint32_t addpath_id,
        if (bgp_mac_entry_exists(p) || bgp_mac_exist(&attr->rmac)) {
                peer->stat_pfx_nh_invalid++;
                reason = "self mac;";
+               bgp_attr_flush(&new_attr);
                goto filtered;
        }
 
@@ -3946,13 +3947,15 @@ int bgp_update(struct peer *peer, const struct prefix *p, uint32_t addpath_id,
            && (!CHECK_FLAG(dest->flags, BGP_NODE_FIB_INSTALLED)))
                SET_FLAG(dest->flags, BGP_NODE_FIB_INSTALL_PENDING);
 
-       attr_new = bgp_attr_intern(&new_attr);
-
        /* If maximum prefix count is configured and current prefix
         * count exeed it.
         */
-       if (bgp_maximum_prefix_overflow(peer, afi, safi, 0))
+       if (bgp_maximum_prefix_overflow(peer, afi, safi, 0)) {
+               bgp_attr_flush(&new_attr);
                return -1;
+       }
+
+       attr_new = bgp_attr_intern(&new_attr);
 
        /* If the update is implicit withdraw. */
        if (pi) {