fix(modbus): transport self-heals with one throttled in-gate connect; PDU length check

P3: a torn-down socket is re-dialled inside the gate (one attempt per backoff
window) instead of failing fast with 'Transport not connected' until the 90 s
comms-loss watchdog re-initialised the driver.
L3: ConnectAsync disposes the previous/failed TcpClient, DNS runs under the
per-op timeout, and the reconnect never sleeps inside the gate (the old
10-attempt backoff held it for ~81 s).
H9: an MBAP response with no PDU, or an exception PDU with no code, is a
TruncatedFrame desync rather than an IndexOutOfRangeException.

Refs #741
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
This commit is contained in:
Joseph Doherty
2026-10-22 15:02:11 -04:00
parent 7bcb6331ed
commit 530f1b2f0b
8 changed files with 655 additions and 70 deletions
@@ -0,0 +1,160 @@
using System.Net;
using System.Net.Sockets;
namespace ZB.MOM.WW.OtOpcUa.Driver.Modbus.Tests;
/// <summary>
/// Loopback Modbus-TCP stub whose reply to each request is scripted by the test, and whose listener can
/// be stopped (connection refused) and restarted on the SAME port. Used by the transport self-heal tests
/// that need a device which disappears and comes back.
/// </summary>
internal sealed class ScriptedModbusTcpServer : IAsyncDisposable
{
/// <summary>What the server does with one request.</summary>
public enum Action
{
/// <summary>Answer with an FC03-shaped response of zeroed registers.</summary>
Answer,
/// <summary>Read the request and never answer it (the client's per-op timeout fires).</summary>
Stall,
/// <summary>Close the socket without answering.</summary>
Drop,
}
private TcpListener? _listener;
private CancellationTokenSource _stop = new();
private int _requests;
private int _accepted;
/// <summary>Optional scripted reply. Receives the 0-based request index and the request PDU; returns
/// the raw bytes to write AFTER the MBAP TxId (i.e. the full ADU is built by the test), or null to fall
/// through to <see cref="Script"/>.</summary>
public Func<int, byte[], byte[], byte[]?>? RawReply { get; set; }
/// <summary>Per-request action; defaults to <see cref="Action.Answer"/>.</summary>
public Func<int, Action> Script { get; set; } = _ => Action.Answer;
/// <summary>Gets the port the server listens on (stable across <see cref="Stop"/>/<see cref="Start"/>).</summary>
public int Port { get; }
/// <summary>Gets the number of accepted TCP connections.</summary>
public int AcceptedConnections => Volatile.Read(ref _accepted);
/// <summary>Gets the number of requests received.</summary>
public int Requests => Volatile.Read(ref _requests);
public ScriptedModbusTcpServer()
{
var probe = new TcpListener(IPAddress.Loopback, 0);
probe.Start();
Port = ((IPEndPoint)probe.LocalEndpoint).Port;
probe.Stop();
Start();
}
/// <summary>(Re)starts listening on <see cref="Port"/>.</summary>
public void Start()
{
_stop = new CancellationTokenSource();
_listener = new TcpListener(IPAddress.Loopback, Port);
_listener.Server.SetSocketOption(SocketOptionLevel.Socket, SocketOptionName.ReuseAddress, true);
_listener.Start();
var listener = _listener;
var token = _stop.Token;
_ = Task.Run(() => AcceptLoopAsync(listener, token));
}
/// <summary>Stops listening and closes every live connection — new connects are refused.</summary>
public void Stop()
{
try { _stop.Cancel(); } catch { }
try { _listener?.Stop(); } catch { }
_listener = null;
}
private async Task AcceptLoopAsync(TcpListener listener, CancellationToken token)
{
while (!token.IsCancellationRequested)
{
TcpClient client;
try { client = await listener.AcceptTcpClientAsync(token); }
catch { return; }
Interlocked.Increment(ref _accepted);
_ = Task.Run(() => ServeAsync(client, token));
}
}
private async Task ServeAsync(TcpClient client, CancellationToken token)
{
using var _ = client;
using var reg = token.Register(() => { try { client.Close(); } catch { } });
try
{
var stream = client.GetStream();
while (!token.IsCancellationRequested)
{
var header = new byte[7];
if (!await ReadExactly(stream, header, token)) return;
var len = (ushort)((header[4] << 8) | header[5]);
var pdu = new byte[len - 1];
if (!await ReadExactly(stream, pdu, token)) return;
var idx = Interlocked.Increment(ref _requests) - 1;
var raw = RawReply?.Invoke(idx, header, pdu);
if (raw is not null)
{
await stream.WriteAsync(raw, token);
await stream.FlushAsync(token);
continue;
}
switch (Script(idx))
{
case Action.Stall:
continue;
case Action.Drop:
client.Client.Shutdown(SocketShutdown.Both);
return;
}
var qty = (ushort)((pdu[3] << 8) | pdu[4]);
var respPdu = new byte[2 + qty * 2];
respPdu[0] = pdu[0];
respPdu[1] = (byte)(qty * 2);
await stream.WriteAsync(Adu(header, respPdu), token);
await stream.FlushAsync(token);
}
}
catch { /* connection torn down */ }
}
/// <summary>Builds a reply ADU echoing the request's TxId/unit around <paramref name="respPdu"/>.</summary>
public static byte[] Adu(byte[] requestHeader, byte[] respPdu)
{
var respLen = (ushort)(1 + respPdu.Length);
var adu = new byte[7 + respPdu.Length];
adu[0] = requestHeader[0]; adu[1] = requestHeader[1];
adu[4] = (byte)(respLen >> 8); adu[5] = (byte)(respLen & 0xFF);
adu[6] = requestHeader[6];
Buffer.BlockCopy(respPdu, 0, adu, 7, respPdu.Length);
return adu;
}
private static async Task<bool> ReadExactly(Stream s, byte[] buf, CancellationToken token)
{
var read = 0;
while (read < buf.Length)
{
var n = await s.ReadAsync(buf.AsMemory(read), token);
if (n == 0) return false;
read += n;
}
return true;
}
public ValueTask DisposeAsync()
{
Stop();
return ValueTask.CompletedTask;
}
}