Core/Networking: Rewrite networking threading model

Each network thread has its own io_service - this means that all operations on a given socket except queueing packets run from a single thread, removing the need for locking
Sending packets now writes to a lockfree intermediate queue directly, encryption is applied in network thread if it was required at the time of sending the packet

(cherry picked from commit 97a79af4701621ec04b88c8b548dbc35d120e99e)

1 files changed, 83 insertions, 0 deletions

diff --git a/src/common/Threading/MPSCQueue.h b/src/common/Threading/MPSCQueue.h
new file mode 100644
index 00000000000..09648b844be
--- /dev/null
+++ b/src/common/Threading/MPSCQueue.h
@@ -0,0 +1,83 @@
+/*
+ * Copyright (C) 2008-2016 TrinityCore <http://www.trinitycore.org/>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; either version 2 of the License, or (at your
+ * option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef MPSCQueue_h__
+#define MPSCQueue_h__
+
+#include <atomic>
+#include <utility>
+
+// C++ implementation of Dmitry Vyukov's lock free MPSC queue
+// http://www.1024cores.net/home/lock-free-algorithms/queues/non-intrusive-mpsc-node-based-queue
+template<typename T>
+class MPSCQueue
+{
+public:
+    MPSCQueue() : _head(new Node()), _tail(_head.load(std::memory_order_relaxed))
+    {
+        Node* front = _head.load(std::memory_order_relaxed);
+        front->Next.store(nullptr, std::memory_order_relaxed);
+    }
+
+    ~MPSCQueue()
+    {
+        T* output;
+        while (this->Dequeue(output))
+            ;
+
+        Node* front = _head.load(std::memory_order_relaxed);
+        delete front;
+    }
+
+    void Enqueue(T* input)
+    {
+        Node* node = new Node(input);
+        Node* prevHead = _head.exchange(node, std::memory_order_acq_rel);
+        prevHead->Next.store(node, std::memory_order_release);
+    }
+
+    bool Dequeue(T*& result)
+    {
+        Node* tail = _tail.load(std::memory_order_relaxed);
+        Node* next = tail->Next.load(std::memory_order_acquire);
+        if (!next)
+            return false;
+
+        result = next->Data;
+        _tail.store(next, std::memory_order_release);
+        delete tail;
+        return true;
+    }
+
+private:
+    struct Node
+    {
+        Node() = default;
+        explicit Node(T* data) : Data(data) { Next.store(nullptr, std::memory_order_relaxed); }
+
+        T* Data;
+        std::atomic<Node*> Next;
+    };
+
+    std::atomic<Node*> _head;
+    std::atomic<Node*> _tail;
+
+    MPSCQueue(MPSCQueue const&) = delete;
+    MPSCQueue& operator=(MPSCQueue const&) = delete;
+};
+
+#endif // MPSCQueue_h__