No connection created through x/xnet had one. Neither NewTCPPool nor the
dial path in StackGo set ConnConfig.LossRecovery and ConnConfig.Nanotime,
so tcp.Handler ran with loss recovery disabled: nothing noticed a lost
segment, and the connection simply stopped: the sender waiting for an
ACK that cannot arrive, the receiver for data nobody will resend.
TCPPoolConfig.NanoTime already documents itself as "passed to each
tcp.Conn for retransmission timing (RFC 6298)", including the
time.Now fallback when it is nil; this makes that true. Each connection
gets its own tcp.RTO, which shadows that connection's send sequence space
and so cannot be shared. That is 80 bytes per connection, allocated once
at pool construction, against buffers measured in kilobytes.
The test drops exactly one data segment and requires the byte to arrive
anyway. It depends on the FIN-WAIT-1 retransmission fix, since the server
closes after writing.
* tcp: remove retransmit logic entirely; add duplicate ack counting to ControlBlock
* retransmit implemented in nice simple straightforward way
* narrow down retransmission cases
* remove old timing tests
* add ControlBlock retransmit test
* add failing handler test
* reworking payload length semantic meaning in code
* fix establish conn logic
* clean up tests and add TCB dupack generation and test it
* catch pending retransmit satisfy in test
* add fuzz test for control block
* bugfix: be more strict in what is considered dupack
* add IncomingIsDupACK docs
* limit queue of retransmits
* protect retransmit overflow from incorrectly updating nxt
* pcap: reuse Frame memory
* slog: reduce heap allocations of addresses; also prevent heap alloc of dhcp options in pcap
* dns: heapless improvement; add StackAsync buffer for more heapless operation; start thinking of errors
* errors: begin standardise errors in lneto
* errors: finish standardization of errors
* fix merge issues
* add more lneto errors to rest of package
* format errors.go
* do not rely on unspecified Go behaviour; fix TCP ring buffer bug
* even more stricter error handling on tcp connections
* stricter lock copying in tcp.Conn, even though likely not a problem
* add listener and tcp pool logging
* forgot reqAddr unused
* add logging to TCPConn and friends
* begin reviewing TCP listener
* begin adding listener tests
* adding TestExchange implementation
* split out legacy and Step tests
* move listener to tcp package
* fix up tcp tests taking very long
* add xnet.TCPPool and work on TCPPool semantics
* xnet: TCPPool only accepts established connections
* xnet: getting to bottom of panic in xnet.Listener
* xnet: improve listener tests
* fix several data races in testing fixtures
* fix more synchronization things in listener test
* finalize tcplistener test
Derek den Haas
Pat Whittingslow