transport-manager: Connections not properly rejected

[lexnv]

Warning detected after running litep2p for several days:

WARN tokio-runtime-worker litep2p::transport-manager: peer ids do not match but transport was supposed to reject connection dialed_peer=PeerId("12D3KooWAnh5dx5RuPEaoghE6ZW4HB3KyUdS3FdM5Hs6E2hzQMM8") peer=PeerId("12D3KooWAynpgAaBddL4zoRjcSfwkt3jUcdZK6beecVYsHmLzkQT") endpoint=Dialer { address: "/ip4/195.201.243.172/tcp/30385/ws/p2p/12D3KooWAynpgAaBddL4zoRjcSfwkt3jUcdZK6beecVYsHmLzkQT", connection_id: ConnectionId(199889) }

litep2p/src/transport/manager/mod.rs

Lines 768 to 775 in 58b7bf7

	if let Some(dialed_peer) = self.pending_connections.remove(&endpoint.connection_id()) {
	if dialed_peer != peer {
	tracing::warn!(
	target: LOG_TARGET,
	?dialed_peer,
	?peer,
	?endpoint,
	"peer ids do not match but transport was supposed to reject connection"

Issue:

• we have dialed a peer (A) on websocket connection
• transport manager extracted a different peer (B) from the connected endpoint
• individual transports should have rejected the connection and not propagate the event back to the transport manager

We might have a broken logic in connection ID assignment, or there is a problem at the transport layer