fix(direct): validate A-role accepted connection, skip stale ones

The Acceptor's accept() on the shared signal endpoint can dequeue
a stale QUIC connection from a previous call that the Dialer has
already dropped. This results in "connection lost" errors when
media datagrams are sent — 100% drops on both sides.

Fix: after accepting a connection, check close_reason(). If the
connection is already closed, log a warning and re-accept. Also
verify max_datagram_size() is available before returning.

Additionally: emit transport details (remote addr, max_datagram,
close_reason) in the call_engine_starting debug event so stale
connection issues are visible in the user-facing debug log.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
Siavash Sameni
2026-04-12 13:50:21 +04:00
parent 40955bd11c
commit 1904b19d05
2 changed files with 50 additions and 31 deletions

View File

@@ -194,43 +194,53 @@ pub async fn race(
let ep_for_fut = ep.clone();
let v6_ep_for_accept = ipv6_endpoint.clone();
direct_fut = Box::pin(async move {
// Phase 7: accept on both IPv4 and IPv6 endpoints.
// First incoming connection on either wins.
match v6_ep_for_accept {
Some(v6_ep) => {
tracing::debug!("dual_path: A-role accepting on both v4 + v6 endpoints");
tokio::select! {
v4 = wzp_transport::accept(&ep_for_fut) => {
let conn = v4.map_err(|e| anyhow::anyhow!("v4 accept: {e}"))?;
tracing::info!(
remote = %conn.remote_address(),
stable_id = conn.stable_id(),
"dual_path: A-role accepted on IPv4 endpoint"
);
Ok(QuinnTransport::new(conn))
}
v6 = wzp_transport::accept(&v6_ep) => {
let conn = v6.map_err(|e| anyhow::anyhow!("v6 accept: {e}"))?;
tracing::info!(
remote = %conn.remote_address(),
stable_id = conn.stable_id(),
"dual_path: A-role accepted on IPv6 endpoint"
);
Ok(QuinnTransport::new(conn))
// Accept loop: retry if we get a stale/closed
// connection from a previous call. Between rapid
// successive calls, quinn's accept queue may
// contain connections that the peer has already
// dropped. Verify the connection is alive via
// max_datagram_size() before returning it.
loop {
let conn = match &v6_ep_for_accept {
Some(v6_ep) => {
tokio::select! {
v4 = wzp_transport::accept(&ep_for_fut) => {
v4.map_err(|e| anyhow::anyhow!("v4 accept: {e}"))?
}
v6 = wzp_transport::accept(v6_ep) => {
v6.map_err(|e| anyhow::anyhow!("v6 accept: {e}"))?
}
}
}
}
None => {
let conn = wzp_transport::accept(&ep_for_fut)
.await
.map_err(|e| anyhow::anyhow!("direct accept: {e}"))?;
tracing::info!(
None => {
wzp_transport::accept(&ep_for_fut)
.await
.map_err(|e| anyhow::anyhow!("direct accept: {e}"))?
}
};
// Validate the connection is alive. A stale
// connection from a previous call will report
// close_reason = Some(...) immediately.
if let Some(reason) = conn.close_reason() {
tracing::warn!(
remote = %conn.remote_address(),
stable_id = conn.stable_id(),
"dual_path: A-role accepted (v4-only)"
?reason,
"dual_path: A-role skipping stale connection, re-accepting"
);
Ok(QuinnTransport::new(conn))
continue;
}
let has_dgram = conn.max_datagram_size().is_some();
tracing::info!(
remote = %conn.remote_address(),
stable_id = conn.stable_id(),
has_dgram,
"dual_path: A-role accepted direct connection"
);
break Ok(QuinnTransport::new(conn));
}
});
direct_ep = ep;

View File

@@ -588,8 +588,17 @@ async fn connect(
}
let app_clone = app.clone();
// Log transport details for debugging direct P2P media issues
let transport_info = pre_connected_transport.as_ref().map(|t| {
serde_json::json!({
"remote": t.remote_address().to_string(),
"max_datagram": t.max_datagram_size(),
"close_reason": t.connection().close_reason().map(|r| format!("{r:?}")),
})
});
emit_call_debug(&app, "connect:call_engine_starting", serde_json::json!({
"is_direct_p2p": is_direct_p2p_agreed,
"transport": transport_info,
}));
let app_for_engine = app.clone();
match CallEngine::start(relay, room, alias, os_aec, quality, reuse_endpoint, pre_connected_transport, is_direct_p2p_agreed, app_for_engine, move |event_kind, message| {