Unify TUN initialization and raw-IP writes, preserve socketpair directions, and bound reader delivery with queue backpressure. Make deferred packet ownership safe across close and uasync teardown, and apply per-interface MTU limits to lwIP MSS.
Validation: full check.sh passed 172 tests, 9 proxy integration checks, burst and load. TUN ASAN/UBSAN regressions passed 30 cases; TCP exchange passed with asymmetric and minimum MTUs. Sanitized NAT integration exposed an existing unaligned uint16_t read in eim_nat.c; that separate defect remains unchanged. Native Windows and FreeBSD validation remains outstanding.
Refresh both AGENTS guides and clarify module purpose, usage, ownership, threading, return values and wire formats in headers. Remove the unused timeout_heap freed_count field and unimplemented getter.
При half-close (exit закрылся первым, браузер держит соединение) pcb уходил в
TIME_WAIT, а conn_finish ждал flow.fin_sent (доставку FIN в exit) — при backpressure
это не наступало, pc->pcb оставался висячим. Таймер TIME_WAIT освобождал pcb без
err_cb, и CLOSE_ALL давал double free. Теперь pcb отсоединяется по завершении
локальной стороны, а tcp_abort/tcp_close защищены проверкой memory_pool_is_freed.
- error_cb (tcp_io): при graceful HUP (SO_ERROR==0) после уже прочитанного FIN
сокет оставался в epoll и level-triggered EPOLLHUP крутил цикл на 100% CPU.
Теперь сокет закрывается через tcp_conn_push_close.
- socks_proxy on_fin_cb: при обоих обменянных FIN локальный сокет закрывается
(send_close + push_close), симметрично exit-стороне (tcp_proxy_server).
- u_async.c: process immediate_queue BEFORE epoll_wait in uasync_poll
- stcp_link.c: defensive NULL etcp_conn/etcp_link in stcp_link_close
- utun_instance.c: move stcp_server_list_destroy_all to Phase J (before ETCP),
add deferred drain in both Phase J and L
- stcp.h/c: STCP_HS_ENC_CLIENT 6→38, STCP_HS_ENC_SERVER 7→39,
exchange ed25519_pubkey in both handshake directions
- etcp_connections.c: copy peer_ed25519 from STCP to ETCP_CONN in
etcp_link_enter_ready_tcp; guard link->etcp in tcp_link_close_cb
- chat_sync.c: fix lk->conn NULL dereference for TCP links
- test_stcp.c: update stcp_server_create/stcp_client_connect callsites
- tools/chat_tcp_test: change transport udp→tcp
- etcp_router: SVC_ROUTE_HDR (+8B), ETCP_ROUTER_CONN/TRANSIT_QUEUE хеши расширены под group_id
- Все caller'ы (proxy, routing, conn_mgr, nat, chatgui, tests) обновлены
- topo_group: глобальный реестр TOPO_NODE по node_id, ref_count=0 при создании
- BGP-обмен активируется для всех групп (а не только дефолтной)
- Протокол: group_id добавлен в WITHDRAW/TABLE_REQ/TABLE_COMPLETE/ERR_GROUP_MISMATCH
- Per-connection коллбэки заменены на instance-level conn_status
- msg_transport полностью удалён из проекта
57/57 тестов
Added #ifdef __cplusplus / extern "C" { ... } / #endif to ~65 headers
in src/ and lib/, enabling the C library to be linked into C++ code (chatgui).
Fixed packet_dump.h (added missing header guard) and etcp_bbr.h
(#pragma once guard).
waiter-based backpressure (commit ba14e31e) could starve HTTPS CONNECT
worker under concurrent 4-worker load — send_q never drained below threshold
fast enough for the waiter to fire before curl timed out (exit=28).
Add retry_timer (500ms, force=1) alongside etcp_router_on_send_ready
waiter: waiter gives instant wake when send_q drains, timer guarantees
delivery via force=1 if waiter hasn't fired in time.
Also improve socks_proxy DEBUG diagnostics in backpressure branches.
- tcp_proxy_server: drain_cb uses etcp_router_on_send_ready instead of retry_timer
- socks_proxy: on_read_cb + tx_waiter_cb use etcp_router_on_send_ready instead of waiter_register
- Result: FIFO round-robin via ll_queue waiters on send_q, all streams get fair share
- No timers — wakeup is event-driven via queue_data_get → check_waiters
- TCP backpressure: read_queue callback_suspended prevents over-reading
evgeny