A player's progress is persisted from two unrelated flows: action handlers triggered by incoming packets (sequential per connection) and the periodic save, which runs on an independent timer. Action handlers mutate tracked entities with plain field and collection writes - for example crafting toggles item.ItemOptions, and item stacking and NPC selling delete item rows - which bypass the persistence context's own lock. When such a mutation runs while SaveChangesAsync enumerates the change tracker, the save throws (collection-modified, or a DbUpdateConcurrency "affected 0 rows" since the context has no concurrency tokens). SaveChanges is atomic, so every following save fails too and the whole session never persists: on relog the player rolls back, losing progress and items. Add a per-player re-entrant persistence lock and acquire it around both the packet-handling funnel and SaveProgressAsync, so a player's mutations and saves can never overlap. The lock is re-entrant per asynchronous flow (an instance AsyncLocal), so an inline save inside an already-serialized handler does not deadlock; and it is per player, so a trade still acquires the trading partner's lock separately. Add regression tests covering mutual exclusion and re-entrancy of the lock.
115 lines
4.6 KiB
C#
115 lines
4.6 KiB
C#
// <copyright file="PersistenceLockTest.cs" company="MUnique">
|
|
// Licensed under the MIT License. See LICENSE file in the project root for full license information.
|
|
// </copyright>
|
|
|
|
namespace MUnique.OpenMU.Tests;
|
|
|
|
using System.Threading;
|
|
|
|
/// <summary>
|
|
/// Tests for the per-player persistence lock (<see cref="MUnique.OpenMU.GameLogic.Player.RunPersistenceExclusiveAsync(System.Func{System.Threading.Tasks.ValueTask},System.Threading.CancellationToken)"/>),
|
|
/// which serializes a player's context mutations against its periodic/disconnect progress saves so they
|
|
/// can never run concurrently. Without it, a mutation running during <c>SaveChangesAsync</c> corrupts the
|
|
/// change tracker and rolls the whole session back.
|
|
/// </summary>
|
|
[TestFixture]
|
|
public class PersistenceLockTest
|
|
{
|
|
/// <summary>
|
|
/// Verifies that concurrent exclusive operations for the same player never overlap.
|
|
/// </summary>
|
|
[Test]
|
|
public async Task ConcurrentAccessIsSerializedAsync()
|
|
{
|
|
var player = await PlayerTestHelper.CreatePlayerAsync().ConfigureAwait(false);
|
|
var concurrent = 0;
|
|
var overlapDetected = false;
|
|
|
|
async ValueTask BodyAsync()
|
|
{
|
|
if (Interlocked.Increment(ref concurrent) > 1)
|
|
{
|
|
overlapDetected = true;
|
|
}
|
|
|
|
await Task.Delay(1).ConfigureAwait(false);
|
|
Interlocked.Decrement(ref concurrent);
|
|
}
|
|
|
|
var tasks = Enumerable.Range(0, 50)
|
|
.Select(_ => player.RunPersistenceExclusiveAsync(BodyAsync).AsTask())
|
|
.ToArray();
|
|
await Task.WhenAll(tasks).ConfigureAwait(false);
|
|
|
|
Assert.That(overlapDetected, Is.False, "Two exclusive operations for the same player ran at the same time.");
|
|
}
|
|
|
|
/// <summary>
|
|
/// Verifies that re-entering the lock from within an already-held exclusive scope does not deadlock
|
|
/// (an inline save inside a packet handler is exactly this case).
|
|
/// </summary>
|
|
[Test]
|
|
public async Task ReentrantAccessDoesNotDeadlockAsync()
|
|
{
|
|
var player = await PlayerTestHelper.CreatePlayerAsync().ConfigureAwait(false);
|
|
var executed = 0;
|
|
|
|
var run = player.RunPersistenceExclusiveAsync(async () =>
|
|
{
|
|
Interlocked.Increment(ref executed);
|
|
await player.RunPersistenceExclusiveAsync(async () =>
|
|
{
|
|
Interlocked.Increment(ref executed);
|
|
await Task.Yield();
|
|
}).ConfigureAwait(false);
|
|
}).AsTask();
|
|
|
|
// If reentrancy deadlocked, this would hang; fail fast instead of blocking the suite.
|
|
await run.WaitAsync(TimeSpan.FromSeconds(5)).ConfigureAwait(false);
|
|
|
|
Assert.That(executed, Is.EqualTo(2));
|
|
}
|
|
|
|
/// <summary>
|
|
/// Verifies that a re-entrant exclusive operation still runs while another flow holds the lock:
|
|
/// the outer flow keeps the lock, an independent flow must wait, and the re-entrant call inside the
|
|
/// outer flow proceeds without waiting for itself.
|
|
/// </summary>
|
|
[Test]
|
|
public async Task IndependentFlowWaitsWhileLockIsHeldAsync()
|
|
{
|
|
var player = await PlayerTestHelper.CreatePlayerAsync().ConfigureAwait(false);
|
|
var otherEntered = false;
|
|
var holderHasLock = new TaskCompletionSource();
|
|
var mayRelease = new TaskCompletionSource();
|
|
|
|
// Holder runs on its own flow and keeps the lock until signalled.
|
|
var holder = Task.Run(() => player.RunPersistenceExclusiveAsync(async () =>
|
|
{
|
|
holderHasLock.SetResult();
|
|
|
|
// A re-entrant call from the holding flow must NOT block on the lock we already hold.
|
|
await player.RunPersistenceExclusiveAsync(() => ValueTask.CompletedTask).ConfigureAwait(false);
|
|
|
|
await mayRelease.Task.ConfigureAwait(false);
|
|
}).AsTask());
|
|
|
|
await holderHasLock.Task.WaitAsync(TimeSpan.FromSeconds(5)).ConfigureAwait(false);
|
|
|
|
// Competing flow started from an INDEPENDENT context (does not inherit the reentrancy flag).
|
|
var other = Task.Run(() => player.RunPersistenceExclusiveAsync(() =>
|
|
{
|
|
otherEntered = true;
|
|
return ValueTask.CompletedTask;
|
|
}).AsTask());
|
|
|
|
await Task.Delay(50).ConfigureAwait(false);
|
|
Assert.That(otherEntered, Is.False, "An independent flow entered while the lock was held.");
|
|
|
|
mayRelease.SetResult();
|
|
await holder.WaitAsync(TimeSpan.FromSeconds(5)).ConfigureAwait(false);
|
|
await other.WaitAsync(TimeSpan.FromSeconds(5)).ConfigureAwait(false);
|
|
Assert.That(otherEntered, Is.True, "The competing flow never ran after the lock was released.");
|
|
}
|
|
}
|