Files
ferrix/crates/integration-tests/tests/basic_connection.rs
T
Rayhaan JaufeerallyandClaude Fable 5.1 41090966e4 Stop treating a dead RIB as an empty one
The RIB handles swallowed every failure of their actor: an UPDATE applied
to a stopped Adj-RIB-In vanished, and a snapshot of it came back empty.
Sessions stayed up, peers' routes went nowhere, and the status page showed
a healthy router with no routes. An actor stops only by panicking, so this
is the state the server would be in after a bug.

- Every operation on AdjRibIn and LocalRib returns Result<_, RibStopped>.
  LocalRibError is replaced by it.
- The server waits on both actors, the advertiser and the dialers. If one
  stops it sends every peer Cease / Administrative Shutdown and returns an
  error, so the process exits and can be restarted. Before, the tasks'
  handles were held and never looked at.
- A session that cannot store an UPDATE ends with the same Cease.
- The HTTP API answers 503 with the reason for anything that needs a table
  it cannot read, including /metricz: a scrape that fails is noticed, last
  known sizes are not.

Also:
- The routes of a session are now removed even if its task is torn down
  without running its cleanup (a panic), instead of staying in the table
  as routes of a peer that is gone.
- Table sizes for the status page and metrics are asked for as a count
  instead of copying the whole table to take its length.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
2026-10-01 23:18:59 +02:00

559 lines
19 KiB
Rust

use std::net::Ipv4Addr;
use std::time::Duration;
use bgp_integration_tests::test_peer::TestPeer;
use bgp_integration_tests::test_server::TestServer;
use bgp_packet::errors::{BgpError, CeaseSubcode, FsmSubcode, MsgHeaderSubcode, UpdateMsgSubcode};
use bgp_packet::message::{
AsPathAttribute, Message, MpReachNlriPathAttribute, MpUnreachNlriPathAttribute,
NextHopPathAttribute, NlriNextHop, NotificationMessage, OriginPathAttribute, PathAttribute,
UpdateMessage,
};
use bgp_primitives::constants::{AddressFamilyId, Afi, Safi};
use bgp_primitives::ip_prefix::IpPrefix;
use bgp_rib::RibEvent;
use eyre::Result;
const TIMEOUT: Duration = Duration::from_secs(5);
/// The attributes every announcement of IPv4 routes must carry. An UPDATE
/// without them has its routes withdrawn instead (RFC 7606).
fn mandatory_attributes() -> Vec<PathAttribute> {
let mut attributes = with_origin_and_as_path(PathAttribute::NextHop(NextHopPathAttribute(
Ipv4Addr::new(192, 0, 2, 1),
)));
// In order of type code: ORIGIN, AS_PATH, NEXT_HOP.
attributes.rotate_left(1);
attributes
}
/// `first` followed by the attributes every announcement must carry.
/// Routes in MP_REACH_NLRI have their next hop there.
fn with_origin_and_as_path(first: PathAttribute) -> Vec<PathAttribute> {
vec![
first,
PathAttribute::Origin(OriginPathAttribute::IGP),
AsPathAttribute::from_asns(vec![65001]),
]
}
// ---------------------------------------------------------------------------
// Connectivity
// ---------------------------------------------------------------------------
#[tokio::test]
async fn test_server_starts() -> Result<()> {
let server = TestServer::start(65000).await?;
assert_ne!(server.addr.port(), 0);
Ok(())
}
#[tokio::test]
async fn test_server_accepts_tcp_connection() -> Result<()> {
let server = TestServer::start(65000).await?;
let _peer = TestPeer::connect(server.addr).await?;
Ok(())
}
#[tokio::test]
async fn test_server_accepts_multiple_connections() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut peers = Vec::new();
for _ in 0..5 {
peers.push(TestPeer::connect(server.addr).await?);
}
assert_eq!(peers.len(), 5);
Ok(())
}
// ---------------------------------------------------------------------------
// OpenSent state
// ---------------------------------------------------------------------------
/// Server sends an OPEN immediately after TCP is established.
#[tokio::test]
async fn test_server_sends_open() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut peer = TestPeer::connect(server.addr).await?;
let open = peer.recv_open().await?;
assert_eq!(open.version, 4);
assert_eq!(open.asn, 65000);
Ok(())
}
/// Server sends a NOTIFICATION and closes if we send garbage instead of OPEN.
#[tokio::test]
async fn test_bad_marker_does_not_crash_server() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut bad = TestPeer::connect(server.addr).await?;
// Discard the server's OPEN, then send garbage.
let _ = bad.recv_open().await?;
bad.send_raw(&[0x00; 19]).await?;
drop(bad);
// New connections are still accepted.
let _good = TestPeer::connect(server.addr).await?;
Ok(())
}
/// Reads until a NOTIFICATION arrives, skipping KEEPALIVEs.
async fn expect_notification(peer: &mut TestPeer) -> Result<NotificationMessage> {
tokio::time::timeout(TIMEOUT, async {
loop {
match peer.recv().await? {
Message::Notification(n) => return Ok(n),
Message::KeepAlive => {}
other => eyre::bail!("expected NOTIFICATION, got {other:?}"),
}
}
})
.await?
}
/// A header with the given length and type, followed by `body`.
fn framed(length: u16, msg_type: u8, body: &[u8]) -> Vec<u8> {
let mut buf = vec![0xFF; 16];
buf.extend_from_slice(&length.to_be_bytes());
buf.push(msg_type);
buf.extend_from_slice(body);
buf
}
/// What the peer sent decides the NOTIFICATION it gets back (RFC 4271 §6),
/// in every state, and the session ends.
#[tokio::test]
async fn test_malformed_messages_are_answered_with_the_matching_notification() -> Result<()> {
let server = TestServer::start(65000).await?;
let header = |subcode| BgpError::MsgHeader(subcode);
let mut bad_marker = framed(19, 4, &[]);
bad_marker[0] = 0;
let cases: Vec<(&str, Vec<u8>, BgpError, Vec<u8>)> = vec![
(
"length beyond the maximum",
framed(5000, 2, &[]),
header(MsgHeaderSubcode::BadMessageLength),
5000u16.to_be_bytes().to_vec(),
),
(
"length below a header",
framed(5, 4, &[]),
header(MsgHeaderSubcode::BadMessageLength),
vec![0, 5],
),
(
"unknown message type",
framed(19, 9, &[]),
header(MsgHeaderSubcode::BadMessageType),
vec![9],
),
(
"marker not all ones",
bad_marker,
header(MsgHeaderSubcode::ConnNotSynchronized),
vec![],
),
(
"UPDATE with attributes running past its end",
framed(23, 2, &[0, 0, 0, 40]),
BgpError::UpdateMsg(UpdateMsgSubcode::MalformedAttrs),
vec![],
),
];
for (i, (what, bytes, error, data)) in cases.into_iter().enumerate() {
// Established: a distinct router ID each time, so that a session
// still closing does not collide with the next one.
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 1, i as u8 + 1), 90)
.await?;
peer.send_raw(&bytes).await?;
let notification = expect_notification(&mut peer).await?;
assert_eq!(
(notification.error, notification.data),
(error.clone(), data.clone()),
"{what}, Established"
);
// OpenSent: the same bytes instead of an OPEN.
let mut peer = TestPeer::connect(server.addr).await?;
let _ = peer.recv_open().await?;
peer.send_raw(&bytes).await?;
let notification = expect_notification(&mut peer).await?;
assert_eq!(
(notification.error, notification.data),
(error, data),
"{what}, OpenSent"
);
}
// A well-formed message at the wrong time is an FSM error, not a
// message error.
let mut peer = TestPeer::connect(server.addr).await?;
let _ = peer.recv_open().await?;
peer.send(&Message::KeepAlive).await?;
let notification = expect_notification(&mut peer).await?;
assert_eq!(
notification.error,
BgpError::FsmError(FsmSubcode::UnexpectedOpenSent)
);
Ok(())
}
/// A message may arrive in pieces, with the server's own timers firing in
/// between. The server once threw away the piece it had already read when
/// that happened, so the message was silently lost (and the stream was only
/// back in step if the remainder happened to be thrown away as well).
#[tokio::test]
async fn test_message_split_across_a_keepalive_tick_is_still_one_message() -> Result<()> {
// Hold time 3 s: the server sends a KEEPALIVE every second.
let server = TestServer::start_with_hold_time(65000, 3).await?;
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 3).await?;
let prefix = IpPrefix::new(AddressFamilyId::Ipv4, vec![10], 8)?;
let mut update = bytes::BytesMut::new();
Message::Update(UpdateMessage {
withdrawn_routes: vec![],
path_attributes: mandatory_attributes(),
announced_routes: vec![prefix.clone()],
})
.to_wire(&peer.ctx, &mut update)?;
// The first half, a pause long enough for the server's keepalive timer
// to fire, then the rest.
peer.send_raw(&update[..10]).await?;
tokio::time::sleep(Duration::from_millis(1300)).await;
peer.send_raw(&update[10..]).await?;
tokio::time::timeout(TIMEOUT, async {
while !server
.adj_rib_in
.snapshot()
.await
.unwrap()
.iter()
.any(|e| e.prefix == prefix)
{
tokio::time::sleep(Duration::from_millis(20)).await;
}
})
.await
.map_err(|_| eyre::eyre!("the split UPDATE never reached the RIB"))?;
assert_eq!(server.registry.list().len(), 1, "session was dropped");
Ok(())
}
// ---------------------------------------------------------------------------
// OpenConfirm → Established
// ---------------------------------------------------------------------------
/// Full OPEN/KEEPALIVE handshake reaches Established.
#[tokio::test]
async fn test_session_established() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut peer = TestPeer::connect(server.addr).await?;
let server_open = peer
.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 90)
.await?;
assert_eq!(server_open.version, 4);
Ok(())
}
/// After Established, the server sends periodic KEEPALIVEs.
#[tokio::test]
async fn test_server_sends_keepalives() -> Result<()> {
// Use a short hold time so the keepalive interval (hold/3 = 1 s) fires quickly.
let server = TestServer::start_with_hold_time(65000, 3).await?;
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 3).await?;
// We should receive a KEEPALIVE within ~1 second (hold/3).
let msg = tokio::time::timeout(Duration::from_secs(3), peer.recv()).await??;
assert!(
matches!(msg, Message::KeepAlive),
"Expected KEEPALIVE, got {:?}",
msg
);
Ok(())
}
/// Server closes with a NOTIFICATION when the hold timer expires.
#[tokio::test]
async fn test_hold_timer_expires() -> Result<()> {
// hold_time = 3 s → server expects a KEEPALIVE within 3 s.
let server = TestServer::start_with_hold_time(65000, 3).await?;
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 3).await?;
// Don't send any keepalives. The server will send keepalives until its own
// hold timer fires and it sends a NOTIFICATION. Drain messages until we
// see it (or the outer timeout fires).
let notif = tokio::time::timeout(Duration::from_secs(6), async {
loop {
match peer.recv().await? {
Message::Notification(n) => return Ok::<_, eyre::Error>(n),
Message::KeepAlive => {} // server keepalive — keep waiting
other => eyre::bail!("Unexpected message: {:?}", other),
}
}
})
.await??;
assert!(
matches!(notif.error, BgpError::HoldTimer),
"Expected HoldTimer NOTIFICATION, got {:?}",
notif
);
Ok(())
}
/// A peer-sent NOTIFICATION is handled gracefully (session closes cleanly).
#[tokio::test]
async fn test_peer_sends_notification() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 90)
.await?;
peer.send(&Message::Notification(NotificationMessage {
error: BgpError::Cease(CeaseSubcode::AdminShutdown),
data: vec![],
}))
.await?;
// Server should close the connection cleanly.
tokio::time::sleep(Duration::from_millis(100)).await;
assert!(peer.is_closed().await);
Ok(())
}
// ---------------------------------------------------------------------------
// UPDATE
// ---------------------------------------------------------------------------
/// An UPDATE message is accepted without crashing the server.
#[tokio::test]
async fn test_server_accepts_update() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 90)
.await?;
let update = Message::Update(UpdateMessage {
withdrawn_routes: vec![],
path_attributes: mandatory_attributes(),
// 10.0.0.0/8 — one byte of prefix for a /8.
announced_routes: vec![IpPrefix::new(AddressFamilyId::Ipv4, vec![10], 8)?],
});
peer.send(&update).await?;
// Server should stay up — we can still exchange a keepalive.
peer.send(&Message::KeepAlive).await?;
Ok(())
}
/// An IPv6 UPDATE via MP_REACH_NLRI is accepted without crashing the server.
#[tokio::test]
async fn test_server_accepts_ipv6_update() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 90)
.await?;
// 2001:db8::/32
let prefix = IpPrefix::new(AddressFamilyId::Ipv6, vec![0x20, 0x01, 0x0d, 0xb8], 32)?;
let update = Message::Update(UpdateMessage {
withdrawn_routes: vec![],
path_attributes: with_origin_and_as_path(PathAttribute::MpReachNlri(
MpReachNlriPathAttribute {
afi: Afi::IPV6,
safi: Safi::UNICAST,
next_hop: NlriNextHop::Ipv6("2001:db8::1".parse().unwrap()),
prefixes: vec![prefix],
},
)),
announced_routes: vec![],
});
peer.send(&update).await?;
// Server should stay up — we can still exchange a keepalive.
peer.send(&Message::KeepAlive).await?;
Ok(())
}
// ---------------------------------------------------------------------------
// Adj-RIB-In introspection
// ---------------------------------------------------------------------------
/// IPv4 prefix announced via the legacy NLRI field appears in the RIB.
#[tokio::test]
async fn test_rib_records_ipv4_announce() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut events = server.adj_rib_in.subscribe();
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 90)
.await?;
let prefix = IpPrefix::new(AddressFamilyId::Ipv4, vec![10], 8)?;
peer.send(&Message::Update(UpdateMessage {
withdrawn_routes: vec![],
path_attributes: mandatory_attributes(),
announced_routes: vec![prefix.clone()],
}))
.await?;
let event = tokio::time::timeout(Duration::from_secs(1), events.recv()).await??;
assert!(
matches!(&event, RibEvent::Announced { prefix: p, .. } if p == &prefix),
"Expected Announced for 10.0.0.0/8, got {event:?}",
);
let snapshot = server.adj_rib_in.snapshot().await.unwrap();
assert!(
snapshot.iter().any(|e| e.prefix == prefix),
"10.0.0.0/8 missing from RIB snapshot",
);
Ok(())
}
/// Withdrawn IPv4 prefix is removed from the RIB.
#[tokio::test]
async fn test_rib_records_ipv4_withdraw() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut events = server.adj_rib_in.subscribe();
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 90)
.await?;
let prefix = IpPrefix::new(AddressFamilyId::Ipv4, vec![10], 8)?;
// Announce first.
peer.send(&Message::Update(UpdateMessage {
withdrawn_routes: vec![],
path_attributes: mandatory_attributes(),
announced_routes: vec![prefix.clone()],
}))
.await?;
tokio::time::timeout(Duration::from_secs(1), events.recv()).await??;
// Now withdraw.
peer.send(&Message::Update(UpdateMessage {
withdrawn_routes: vec![prefix.clone()],
path_attributes: vec![],
announced_routes: vec![],
}))
.await?;
let event = tokio::time::timeout(Duration::from_secs(1), events.recv()).await??;
assert!(
matches!(&event, RibEvent::Withdrawn { prefix: p, .. } if p == &prefix),
"Expected Withdrawn for 10.0.0.0/8, got {event:?}",
);
let snapshot = server.adj_rib_in.snapshot().await.unwrap();
assert!(
!snapshot.iter().any(|e| e.prefix == prefix),
"10.0.0.0/8 should have been removed from RIB snapshot",
);
Ok(())
}
/// IPv6 prefix announced via MP_REACH_NLRI appears in the RIB.
#[tokio::test]
async fn test_rib_records_ipv6_announce() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut events = server.adj_rib_in.subscribe();
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 90)
.await?;
let prefix = IpPrefix::new(AddressFamilyId::Ipv6, vec![0x20, 0x01, 0x0d, 0xb8], 32)?;
peer.send(&Message::Update(UpdateMessage {
withdrawn_routes: vec![],
path_attributes: with_origin_and_as_path(PathAttribute::MpReachNlri(
MpReachNlriPathAttribute {
afi: Afi::IPV6,
safi: Safi::UNICAST,
next_hop: NlriNextHop::Ipv6("2001:db8::1".parse().unwrap()),
prefixes: vec![prefix.clone()],
},
)),
announced_routes: vec![],
}))
.await?;
let event = tokio::time::timeout(Duration::from_secs(1), events.recv()).await??;
assert!(
matches!(&event, RibEvent::Announced { prefix: p, .. } if p == &prefix),
"Expected Announced for 2001:db8::/32, got {event:?}",
);
let snapshot = server.adj_rib_in.snapshot().await.unwrap();
assert!(
snapshot.iter().any(|e| e.prefix == prefix),
"2001:db8::/32 missing from RIB snapshot",
);
Ok(())
}
/// IPv6 prefix withdrawn via MP_UNREACH_NLRI is removed from the RIB.
#[tokio::test]
async fn test_rib_records_ipv6_withdraw() -> Result<()> {
let server = TestServer::start(65000).await?;
let mut events = server.adj_rib_in.subscribe();
let mut peer = TestPeer::connect(server.addr).await?;
peer.handshake(65001, Ipv4Addr::new(10, 0, 0, 1), 90)
.await?;
let prefix = IpPrefix::new(AddressFamilyId::Ipv6, vec![0x20, 0x01, 0x0d, 0xb8], 32)?;
// Announce first.
peer.send(&Message::Update(UpdateMessage {
withdrawn_routes: vec![],
path_attributes: with_origin_and_as_path(PathAttribute::MpReachNlri(
MpReachNlriPathAttribute {
afi: Afi::IPV6,
safi: Safi::UNICAST,
next_hop: NlriNextHop::Ipv6("2001:db8::1".parse().unwrap()),
prefixes: vec![prefix.clone()],
},
)),
announced_routes: vec![],
}))
.await?;
tokio::time::timeout(Duration::from_secs(1), events.recv()).await??;
// Now withdraw via MP_UNREACH_NLRI.
peer.send(&Message::Update(UpdateMessage {
withdrawn_routes: vec![],
path_attributes: vec![PathAttribute::MpUnreachNlri(MpUnreachNlriPathAttribute {
afi: Afi::IPV6,
safi: Safi::UNICAST,
prefixes: vec![prefix.clone()],
})],
announced_routes: vec![],
}))
.await?;
let event = tokio::time::timeout(Duration::from_secs(1), events.recv()).await??;
assert!(
matches!(&event, RibEvent::Withdrawn { prefix: p, .. } if p == &prefix),
"Expected Withdrawn for 2001:db8::/32, got {event:?}",
);
let snapshot = server.adj_rib_in.snapshot().await.unwrap();
assert!(
!snapshot.iter().any(|e| e.prefix == prefix),
"2001:db8::/32 should have been removed from RIB snapshot",
);
Ok(())
}