wandering1
7efa58d9a1
分两段（用户裁定 2026-11-02）：先做地址迁移，再把同类字面量收敛成
单一真源，以后再改地址只动一处。
一、IP 迁移 192.168.1.100 → 192.168.31.100（gw 同步 → .31.1）
- 工厂默认：Drivers/BSP/NETCFG/net_cfg.c 的 NET_CFG_DEF_IP/_GW
- 注释/文档示例、主机与板级测试、台架 py 脚本、两份 sim_config.ini
- 字节序教学点同步换算 hex：0x6401A8C0 → 0x641FA8C0（内存
C0 A8 1F 64），旧版示例引文与历史 Trap 记录按冻结纪律不动
- UI 黄金基准：网口设置 4 态 + shot_info 共 5 张 .bin 由
check_content_pages.py --update 重刷（渲染值 IP 变了，属预期变更）
- 保留判定：客户点表引用（IEC104/继电保护规划）、芯片手册摘录、
日期冻结的变更记录行、已退役台架档案 MDK-ARM/retracted_bench_*
二、单一真源收口（A+B+C+D，用户批准全做）
- 新建 test/host/bench_defaults.py：台架 py 默认值唯一定义点
（DEVICE_IP/DEVICE_GW/PC_IP/NET_BCAST），22 个脚本改为引用；
MDK-ARM 两探针与 scripts/board_ui_nav.py 经 sys.path shim 引入
- test/board/net_test_task.h 新增 TEST_PC_PEER_IP，板测三处
PHASE*_REMOTE_IP 改引该宏（保留 -D 覆盖）；Phase1 断言改
memcmp 数组比较，数组各只出现一次
- simulator/tests/check_content_pages.py 从 net_cfg.c 正则解析
NET_CFG_DEF_* 推导期望值（解析失败即报错、不静默兜底），
脚本内台架 IP 字面量清零
- 装饰性注释去具体值（net_cfg.h/net_socket.h/net_socket.c/
net_socket_sim.c）；字节序教学 hex 对刻意保留为唯一写值点
- 刻意不动：主机单测夹具 IP（逐文件钉住的任意值是正确测试
实践）、文档正文 IP（那是文档内容本身）
顺带：scripts/gen_task_table.py 规范串清掉「，B5」——编号清退提交
72b4b094 改了受管区没重跑生成器，task-table 门禁因此一直红
验证：产品 @build 0/0；CH395F 套件开关构建 0/0 后还原产品态；
6 套主机单测 rc=0；content_pages（含宏推导路径）/菜单 8 态/
推流位比对全 PASS；bom-drift 99 无漂移；doc-links/coding-standard/
cppcheck 绿；py_compile 23 文件 + --help 冒烟（含 shim 绝对路径）
89 lines
3.5 KiB
Python
89 lines
3.5 KiB
Python
# -*- coding: utf-8 -*-
|
||
# TRAP-NET-17 复现探针：快速重连到设备监听器: rapid reconnect to a device listener.
|
||
#
|
||
# Failure under test (original report): after a rapid reconnect (<~1s) on a
|
||
# relistened socket, the chip completes the handshake but the MCU never gets
|
||
# standalone accept -> the socket stays LISTENING -> later the chip auto-closes it
|
||
# -> THE LISTENER DIES PERMANENTLY (all later SYNs refused, no self-recovery).
|
||
#
|
||
# Judgement is purely client-side (no firmware probes):
|
||
# * every round must (re)connect successfully within --timeout
|
||
# * on failure the probe keeps retrying for --recover-s to distinguish
|
||
# "transient" from "permanently dead listener"
|
||
#
|
||
# Usage: tcp_reconnect_probe.py [--host <设备IP>] [--port 21] [--rounds 50]
|
||
# [--gap-ms 600] [--timeout 3] [--recover-s 10]
|
||
# [--banner] (read one line after connect, e.g. FTP 220)
|
||
# Output:
|
||
# TS <epoch>
|
||
# ROUND <n> gap=<ms> connect=<OK|FAIL:reason> recovered=<na|yes|no>
|
||
# SUMMARY rounds=<n> ok=<n> fail=<n> unrecovered=<n>
|
||
# Exit: 0 all rounds connected / 1 any failure / 2 usage
|
||
import argparse, os, socket, sys, time
|
||
|
||
# 台架连接默认值单一真源在 test/host/bench_defaults.py（跨目录：补 sys.path 后 import）
|
||
_sys_path = os.path.join(os.path.dirname(os.path.abspath(__file__)),
|
||
"..", "test", "host")
|
||
if _sys_path not in sys.path:
|
||
sys.path.insert(0, _sys_path)
|
||
import bench_defaults # noqa: E402
|
||
|
||
def log(m):
|
||
sys.stdout.write(m + "\n"); sys.stdout.flush()
|
||
|
||
def try_connect(host, port, timeout, banner):
|
||
s = socket.create_connection((host, port), timeout=timeout)
|
||
if banner:
|
||
s.settimeout(timeout)
|
||
try:
|
||
s.recv(256) # e.g. FTP "220 ..." greeting
|
||
except socket.timeout:
|
||
pass
|
||
return s
|
||
|
||
def main():
|
||
ap = argparse.ArgumentParser()
|
||
ap.add_argument("--host", default=bench_defaults.DEVICE_IP)
|
||
ap.add_argument("--port", type=int, default=21)
|
||
ap.add_argument("--rounds", type=int, default=50)
|
||
ap.add_argument("--gap-ms", type=int, default=600)
|
||
ap.add_argument("--timeout", type=float, default=3.0)
|
||
ap.add_argument("--recover-s", type=float, default=10.0)
|
||
ap.add_argument("--banner", action="store_true")
|
||
a = ap.parse_args()
|
||
|
||
nok = nfail = nunrec = 0
|
||
for n in range(1, a.rounds + 1):
|
||
log("TS %d" % int(time.time()))
|
||
res = "OK"
|
||
recovered = "na"
|
||
try:
|
||
s = try_connect(a.host, a.port, a.timeout, a.banner)
|
||
s.close()
|
||
except Exception as e:
|
||
res = "FAIL:%s" % type(e).__name__
|
||
# keep probing: transient (relisten window) or permanently dead?
|
||
t0 = time.time()
|
||
recovered = "no"
|
||
while time.time() - t0 < a.recover_s:
|
||
time.sleep(0.5)
|
||
try:
|
||
s = try_connect(a.host, a.port, a.timeout, a.banner)
|
||
s.close()
|
||
recovered = "yes"
|
||
break
|
||
except Exception:
|
||
pass
|
||
ok = (res == "OK")
|
||
nok += 1 if ok else 0
|
||
nfail += 0 if ok else 1
|
||
if not ok and recovered != "yes":
|
||
nunrec += 1
|
||
log("ROUND %d gap=%d connect=%s recovered=%s" % (n, a.gap_ms, res, recovered))
|
||
if a.gap_ms > 0 and n < a.rounds:
|
||
time.sleep(a.gap_ms / 1000.0)
|
||
log("SUMMARY rounds=%d ok=%d fail=%d unrecovered=%d" % (a.rounds, nok, nfail, nunrec))
|
||
return 0 if nfail == 0 else 1
|
||
|
||
sys.exit(main())
|