bgpd: Show the reason when the session is killed due to RTT

Simulated latency with:

```
tc qdisc add dev eth3 root netem delay 100ms
```

```
donatas-laptop# sh ip bgp summary failed

IPv4 Unicast Summary (VRF default):
BGP router identifier 192.0.2.252, local AS number 65000 vrf-id 0
BGP table version 28
RIB entries 0, using 0 bytes of memory
Peers 1, using 724 KiB of memory

Neighbor        EstdCnt DropCnt ResetTime Reason
192.168.10.65         2       2  00:00:17 Admin. shutdown (RTT)

Displayed neighbors 1
Total number of neighbors 1
donatas-laptop#
```

Another end received:

```
%NOTIFICATION: received from neighbor 192.168.10.17 6/2 (Cease/Administrative Shutdown) "shutdown due to high round-trip-time (104ms > 5ms, hit 21 times)"
```

Signed-off-by: Donatas Abraitis <donatas@opensourcerouting.org>

diff --git a/bgpd/bgp_fsm.c b/bgpd/bgp_fsm.c
index 01c61615a3312c2ba5eec8b75316880d386e272e..f110005e49f28d3b05c900719c8709534f3880d0 100644 (file)
--- a/bgpd/bgp_fsm.c
+++ b/bgpd/bgp_fsm.c
@@ -634,7 +634,8 @@ const char *const peer_down_str[] = {"",
                                     "AS Set config change",
                                     "Waiting for peer OPEN",
                                     "Reached received prefix count",
-                                    "Socket Error"};
+                                    "Socket Error",
+                                    "Admin. shutdown (RTT)"};
 
 static void bgp_graceful_restart_timer_off(struct peer *peer)
 {
@@ -1832,7 +1833,9 @@ int bgp_start(struct peer *peer)
                        flog_err(EC_BGP_FSM,
                                 "%s [FSM] Trying to start suppressed peer - this is never supposed to happen!",
                                 peer->host);
-               if (CHECK_FLAG(peer->flags, PEER_FLAG_SHUTDOWN))
+               if (CHECK_FLAG(peer->sflags, PEER_STATUS_RTT_SHUTDOWN))
+                       peer->last_reset = PEER_DOWN_RTT_SHUTDOWN;
+               else if (CHECK_FLAG(peer->flags, PEER_FLAG_SHUTDOWN))
                        peer->last_reset = PEER_DOWN_USER_SHUTDOWN;
                else if (CHECK_FLAG(peer->bgp->flags, BGP_FLAG_SHUTDOWN))
                        peer->last_reset = PEER_DOWN_USER_SHUTDOWN;

diff --git a/bgpd/bgp_packet.c b/bgpd/bgp_packet.c
index 72d6a9231756538a102fb18831fb9da2988d8ed2..f3ca3bba0a0651299f63bed039b01a011cd735e6 100644 (file)
--- a/bgpd/bgp_packet.c
+++ b/bgpd/bgp_packet.c
@@ -1011,9 +1011,12 @@ static void bgp_notify_send_internal(struct peer *peer, uint8_t code,
        if (code == BGP_NOTIFY_CEASE) {
                if (sub_code == BGP_NOTIFY_CEASE_ADMIN_RESET)
                        peer->last_reset = PEER_DOWN_USER_RESET;
-               else if (sub_code == BGP_NOTIFY_CEASE_ADMIN_SHUTDOWN)
-                       peer->last_reset = PEER_DOWN_USER_SHUTDOWN;
-               else
+               else if (sub_code == BGP_NOTIFY_CEASE_ADMIN_SHUTDOWN) {
+                       if (CHECK_FLAG(peer->sflags, PEER_STATUS_RTT_SHUTDOWN))
+                               peer->last_reset = PEER_DOWN_RTT_SHUTDOWN;
+                       else
+                               peer->last_reset = PEER_DOWN_USER_SHUTDOWN;
+               } else
                        peer->last_reset = PEER_DOWN_NOTIFY_SEND;
        } else
                peer->last_reset = PEER_DOWN_NOTIFY_SEND;
@@ -1749,15 +1752,24 @@ static int bgp_keepalive_receive(struct peer *peer, bgp_size_t size)
        /* If the peer's RTT is higher than expected, shutdown
         * the peer automatically.
         */
-       if (CHECK_FLAG(peer->flags, PEER_FLAG_RTT_SHUTDOWN)
-           && peer->rtt > peer->rtt_expected) {
+       if (!CHECK_FLAG(peer->flags, PEER_FLAG_RTT_SHUTDOWN))
+               return Receive_KEEPALIVE_message;
 
+       if (peer->rtt > peer->rtt_expected) {
                peer->rtt_keepalive_rcv++;
 
                if (peer->rtt_keepalive_rcv > peer->rtt_keepalive_conf) {
-                       zlog_warn(
-                               "%s shutdown due to high round-trip-time (%dms > %dms)",
-                               peer->host, peer->rtt, peer->rtt_expected);
+                       char rtt_shutdown_reason[BUFSIZ] = {};
+
+                       snprintfrr(
+                               rtt_shutdown_reason,
+                               sizeof(rtt_shutdown_reason),
+                               "shutdown due to high round-trip-time (%dms > %dms, hit %u times)",
+                               peer->rtt, peer->rtt_expected,
+                               peer->rtt_keepalive_rcv);
+                       zlog_warn("%s %s", peer->host, rtt_shutdown_reason);
+                       SET_FLAG(peer->sflags, PEER_STATUS_RTT_SHUTDOWN);
+                       peer_tx_shutdown_message_set(peer, rtt_shutdown_reason);
                        peer_flag_set(peer, PEER_FLAG_SHUTDOWN);
                }
        } else {

diff --git a/bgpd/bgp_vty.c b/bgpd/bgp_vty.c
index dcbd157e7533313d01e3b8fbacb71c3b6a7a7f14..2228e0d4367be9e745b53c2ac59abb6b8cab568c 100644 (file)
--- a/bgpd/bgp_vty.c
+++ b/bgpd/bgp_vty.c
@@ -5272,8 +5272,10 @@ static int peer_flag_modify_vty(struct vty *vty, const char *ip_str,
                return CMD_WARNING_CONFIG_FAILED;
        }
 
-       if (!set && flag == PEER_FLAG_SHUTDOWN)
+       if (!set && flag == PEER_FLAG_SHUTDOWN) {
                peer_tx_shutdown_message_unset(peer);
+               UNSET_FLAG(peer->sflags, PEER_STATUS_RTT_SHUTDOWN);
+       }
 
        if (set)
                ret = peer_flag_set(peer, flag);

diff --git a/bgpd/bgpd.h b/bgpd/bgpd.h
index 8daf8aeb6924e2a6c9d4e967c86e3a1acc06172d..c957a407514af8408408ed9fa3a41bd75443b6b0 100644 (file)
--- a/bgpd/bgpd.h
+++ b/bgpd/bgpd.h
@@ -1522,6 +1522,7 @@ struct peer {
 /* LLGR aware peer */
 #define PEER_STATUS_LLGR_WAIT (1U << 11)
 #define PEER_STATUS_REFRESH_PENDING (1U << 12) /* refresh request from peer */
+#define PEER_STATUS_RTT_SHUTDOWN (1U << 13) /* In shutdown state due to RTT */
 
        /* Configured timer values. */
        _Atomic uint32_t holdtime;
@@ -1734,6 +1735,7 @@ struct peer {
 #define PEER_DOWN_WAITING_OPEN          32U /* Waiting for open to succeed */
 #define PEER_DOWN_PFX_COUNT             33U /* Reached received prefix count */
 #define PEER_DOWN_SOCKET_ERROR          34U /* Some socket error happened */
+#define PEER_DOWN_RTT_SHUTDOWN          35U /* Automatically shutdown due to RTT */
        /*
         * Remember to update peer_down_str in bgp_fsm.c when you add
         * a new value to the last_reset reason