Thanks to a report by Marko Cupac a mistake could be identified when
generating the data shown in 'bgpctl show rib out nei $FOO'.
The nexthop shown is actually the nexthop of the Loc-RIB and not from
Adj-RIB-Out and so the effect of a filter rule like 'match to $FOO set
nexthop 192.0.2.17' was not visible. Instead of using the nexthop from
struct prefix the nexthop from the filter-state needs to be used.
The following diff adjusts this problem.
--
:wq Claudio
Index: rde.c
===================================================================
RCS file: /cvs/src/usr.sbin/bgpd/rde.c,v
retrieving revision 1.438
diff -u -p -r1.438 rde.c
--- rde.c 22 Oct 2018 07:46:55 -0000 1.438
+++ rde.c 22 Oct 2018 13:35:57 -0000
@@ -69,14 +69,6 @@ void rde_update_log(const char *, u_in
void rde_as4byte_fixup(struct rde_peer *, struct rde_aspath *);
void rde_reflector(struct rde_peer *, struct rde_aspath *);
-void rde_dump_rib_as(struct prefix *, struct rde_aspath *, pid_t,
- int);
-void rde_dump_filter(struct prefix *,
- struct ctl_show_rib_request *);
-void rde_dump_filterout(struct rde_peer *, struct prefix *,
- struct ctl_show_rib_request *);
-void rde_dump_upcall(struct rib_entry *, void *);
-void rde_dump_prefix_upcall(struct rib_entry *, void *);
void rde_dump_ctx_new(struct ctl_show_rib_request *, pid_t,
enum imsg_type);
void rde_dump_ctx_throttle(pid_t pid, int throttle);
@@ -2148,8 +2140,9 @@ rde_reflector(struct rde_peer *peer, str
/*
* control specific functions
*/
-void
-rde_dump_rib_as(struct prefix *p, struct rde_aspath *asp, pid_t pid, int flags)
+static void
+rde_dump_rib_as(struct prefix *p, struct rde_aspath *asp,
+ struct nexthop *nexthop, pid_t pid, int flags)
{
struct ctl_show_rib rib;
struct ibuf *wbuf;
@@ -2167,10 +2160,10 @@ rde_dump_rib_as(struct prefix *p, struct
memcpy(&rib.remote_addr, &prefix_peer(p)->remote_addr,
sizeof(rib.remote_addr));
rib.remote_id = prefix_peer(p)->remote_bgpid;
- if (prefix_nexthop(p) != NULL) {
- memcpy(&rib.true_nexthop, &prefix_nexthop(p)->true_nexthop,
+ if (nexthop != NULL) {
+ memcpy(&rib.true_nexthop, &nexthop->true_nexthop,
sizeof(rib.true_nexthop));
- memcpy(&rib.exit_nexthop, &prefix_nexthop(p)->exit_nexthop,
+ memcpy(&rib.exit_nexthop, &nexthop->exit_nexthop,
sizeof(rib.exit_nexthop));
} else {
/* announced network may have a NULL nexthop */
@@ -2190,8 +2183,7 @@ rde_dump_rib_as(struct prefix *p, struct
rib.flags |= F_PREF_INTERNAL;
if (asp->flags & F_PREFIX_ANNOUNCED)
rib.flags |= F_PREF_ANNOUNCE;
- if (prefix_nexthop(p) == NULL ||
- prefix_nexthop(p)->state == NEXTHOP_REACH)
+ if (nexthop == NULL || nexthop->state == NEXTHOP_REACH)
rib.flags |= F_PREF_ELIGIBLE;
if (asp->flags & F_ATTR_LOOP)
rib.flags &= ~F_PREF_ELIGIBLE;
@@ -2232,7 +2224,7 @@ rde_dump_rib_as(struct prefix *p, struct
}
}
-void
+static void
rde_dump_filterout(struct rde_peer *peer, struct prefix *p,
struct ctl_show_rib_request *req)
{
@@ -2247,12 +2239,13 @@ rde_dump_filterout(struct rde_peer *peer
a = rde_filter(out_rules, peer, p, &state);
if (a == ACTION_ALLOW)
- rde_dump_rib_as(p, &state.aspath, req->pid, req->flags);
+ rde_dump_rib_as(p, &state.aspath, state.nexthop, req->pid,
+ req->flags);
rde_filterstate_clean(&state);
}
-void
+static void
rde_dump_filter(struct prefix *p, struct ctl_show_rib_request *req)
{
struct rde_peer *peer;
@@ -2293,11 +2286,12 @@ rde_dump_filter(struct prefix *p, struct
return;
if (!ovs_match(p, req->flags))
return;
- rde_dump_rib_as(p, asp, req->pid, req->flags);
+ rde_dump_rib_as(p, asp, prefix_nexthop(p), req->pid,
+ req->flags);
}
}
-void
+static void
rde_dump_upcall(struct rib_entry *re, void *ptr)
{
struct prefix *p;
@@ -2307,7 +2301,7 @@ rde_dump_upcall(struct rib_entry *re, vo
rde_dump_filter(p, &ctx->req);
}
-void
+static void
rde_dump_prefix_upcall(struct rib_entry *re, void *ptr)
{
struct rde_dump_ctx *ctx = ptr;
