Fix race condition in HttpSys shutdown request draining logic

[Copilot]

Fix race condition in HttpSys shutdown request draining logic

• You've read the Contributor Guide and Code of Conduct.
• You've included unit or integration tests for your change, where applicable.
• You've included inline docs for your change, where applicable.
• There's an open issue for the PR that you are making. If you'd like to propose a new feature or change, please open an issue to discuss the change or find an existing issue.

Addresses potential hang in StopAsync due to memory visibility issues when checking shutdown state.

Description

Without proper synchronization, DecrementOutstandingRequest could observe stale _acceptLoopsCompleted.Task.IsCompleted state after the last request completes, causing _requestsDrained to never signal and StopAsync to hang.

Changes:

• Take _shutdownLock in DecrementOutstandingRequest when count == 0 before checking _acceptLoopsCompleted state
• Ensures memory visibility with AcceptLoopCompleted which also holds the lock when setting completion
• Lock only acquired once during server lifetime (when final request completes during shutdown)

internal int DecrementOutstandingRequest()
{
    var count = Interlocked.Decrement(ref _outstandingRequests);

    if (count == 0)
    {
        lock (_shutdownLock)
        {
            if (_acceptLoopsCompleted.Task.IsCompleted)
            {
                _requestsDrained.TrySetResult();
            }
        }
    }

    return count;
}

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.