Full Code of craflin/LockFreeQueue for AI

Repository: craflin/LockFreeQueue
Branch: master
Commit: 2d9f743bc797
Files: 16
Total size: 41.8 KB

Directory structure:
gitextract_8qognr8x/

├── .gitmodules
├── LICENSE
├── LockFreeLifoQueue.h
├── LockFreeQueue.h
├── LockFreeQueueCpp11.h
├── LockFreeQueueSlow1.h
├── LockFreeQueueSlow2.h
├── LockFreeQueueSlow3.h
├── Marefile
├── MutexLockQueue.h
├── README.md
├── SpinLockQueue.h
├── Test.cpp
├── generate
├── generate.bat
└── mpmc_bounded_queue.h

================================================
FILE CONTENTS
================================================

================================================
FILE: .gitmodules
================================================
[submodule "Ext/libnstd"]
	path = Ext/libnstd
	url = https://github.com/craflin/libnstd.git
[submodule "Ext/mare"]
	path = Ext/mare
	url = https://github.com/craflin/mare.git


================================================
FILE: LICENSE
================================================

                                 Apache License
                           Version 2.0, January 2004
                        http://www.apache.org/licenses/

   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION

   1. Definitions.

      "License" shall mean the terms and conditions for use, reproduction,
      and distribution as defined by Sections 1 through 9 of this document.

      "Licensor" shall mean the copyright owner or entity authorized by
      the copyright owner that is granting the License.

      "Legal Entity" shall mean the union of the acting entity and all
      other entities that control, are controlled by, or are under common
      control with that entity. For the purposes of this definition,
      "control" means (i) the power, direct or indirect, to cause the
      direction or management of such entity, whether by contract or
      otherwise, or (ii) ownership of fifty percent (50%) or more of the
      outstanding shares, or (iii) beneficial ownership of such entity.

      "You" (or "Your") shall mean an individual or Legal Entity
      exercising permissions granted by this License.

      "Source" form shall mean the preferred form for making modifications,
      including but not limited to software source code, documentation
      source, and configuration files.

      "Object" form shall mean any form resulting from mechanical
      transformation or translation of a Source form, including but
      not limited to compiled object code, generated documentation,
      and conversions to other media types.

      "Work" shall mean the work of authorship, whether in Source or
      Object form, made available under the License, as indicated by a
      copyright notice that is included in or attached to the work
      (an example is provided in the Appendix below).

      "Derivative Works" shall mean any work, whether in Source or Object
      form, that is based on (or derived from) the Work and for which the
      editorial revisions, annotations, elaborations, or other modifications
      represent, as a whole, an original work of authorship. For the purposes
      of this License, Derivative Works shall not include works that remain
      separable from, or merely link (or bind by name) to the interfaces of,
      the Work and Derivative Works thereof.

      "Contribution" shall mean any work of authorship, including
      the original version of the Work and any modifications or additions
      to that Work or Derivative Works thereof, that is intentionally
      submitted to Licensor for inclusion in the Work by the copyright owner
      or by an individual or Legal Entity authorized to submit on behalf of
      the copyright owner. For the purposes of this definition, "submitted"
      means any form of electronic, verbal, or written communication sent
      to the Licensor or its representatives, including but not limited to
      communication on electronic mailing lists, source code control systems,
      and issue tracking systems that are managed by, or on behalf of, the
      Licensor for the purpose of discussing and improving the Work, but
      excluding communication that is conspicuously marked or otherwise
      designated in writing by the copyright owner as "Not a Contribution."

      "Contributor" shall mean Licensor and any individual or Legal Entity
      on behalf of whom a Contribution has been received by Licensor and
      subsequently incorporated within the Work.

   2. Grant of Copyright License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      copyright license to reproduce, prepare Derivative Works of,
      publicly display, publicly perform, sublicense, and distribute the
      Work and such Derivative Works in Source or Object form.

   3. Grant of Patent License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      (except as stated in this section) patent license to make, have made,
      use, offer to sell, sell, import, and otherwise transfer the Work,
      where such license applies only to those patent claims licensable
      by such Contributor that are necessarily infringed by their
      Contribution(s) alone or by combination of their Contribution(s)
      with the Work to which such Contribution(s) was submitted. If You
      institute patent litigation against any entity (including a
      cross-claim or counterclaim in a lawsuit) alleging that the Work
      or a Contribution incorporated within the Work constitutes direct
      or contributory patent infringement, then any patent licenses
      granted to You under this License for that Work shall terminate
      as of the date such litigation is filed.

   4. Redistribution. You may reproduce and distribute copies of the
      Work or Derivative Works thereof in any medium, with or without
      modifications, and in Source or Object form, provided that You
      meet the following conditions:

      (a) You must give any other recipients of the Work or
          Derivative Works a copy of this License; and

      (b) You must cause any modified files to carry prominent notices
          stating that You changed the files; and

      (c) You must retain, in the Source form of any Derivative Works
          that You distribute, all copyright, patent, trademark, and
          attribution notices from the Source form of the Work,
          excluding those notices that do not pertain to any part of
          the Derivative Works; and

      (d) If the Work includes a "NOTICE" text file as part of its
          distribution, then any Derivative Works that You distribute must
          include a readable copy of the attribution notices contained
          within such NOTICE file, excluding those notices that do not
          pertain to any part of the Derivative Works, in at least one
          of the following places: within a NOTICE text file distributed
          as part of the Derivative Works; within the Source form or
          documentation, if provided along with the Derivative Works; or,
          within a display generated by the Derivative Works, if and
          wherever such third-party notices normally appear. The contents
          of the NOTICE file are for informational purposes only and
          do not modify the License. You may add Your own attribution
          notices within Derivative Works that You distribute, alongside
          or as an addendum to the NOTICE text from the Work, provided
          that such additional attribution notices cannot be construed
          as modifying the License.

      You may add Your own copyright statement to Your modifications and
      may provide additional or different license terms and conditions
      for use, reproduction, or distribution of Your modifications, or
      for any such Derivative Works as a whole, provided Your use,
      reproduction, and distribution of the Work otherwise complies with
      the conditions stated in this License.

   5. Submission of Contributions. Unless You explicitly state otherwise,
      any Contribution intentionally submitted for inclusion in the Work
      by You to the Licensor shall be under the terms and conditions of
      this License, without any additional terms or conditions.
      Notwithstanding the above, nothing herein shall supersede or modify
      the terms of any separate license agreement you may have executed
      with Licensor regarding such Contributions.

   6. Trademarks. This License does not grant permission to use the trade
      names, trademarks, service marks, or product names of the Licensor,
      except as required for reasonable and customary use in describing the
      origin of the Work and reproducing the content of the NOTICE file.

   7. Disclaimer of Warranty. Unless required by applicable law or
      agreed to in writing, Licensor provides the Work (and each
      Contributor provides its Contributions) on an "AS IS" BASIS,
      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
      implied, including, without limitation, any warranties or conditions
      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
      PARTICULAR PURPOSE. You are solely responsible for determining the
      appropriateness of using or redistributing the Work and assume any
      risks associated with Your exercise of permissions under this License.

   8. Limitation of Liability. In no event and under no legal theory,
      whether in tort (including negligence), contract, or otherwise,
      unless required by applicable law (such as deliberate and grossly
      negligent acts) or agreed to in writing, shall any Contributor be
      liable to You for damages, including any direct, indirect, special,
      incidental, or consequential damages of any character arising as a
      result of this License or out of the use or inability to use the
      Work (including but not limited to damages for loss of goodwill,
      work stoppage, computer failure or malfunction, or any and all
      other commercial damages or losses), even if such Contributor
      has been advised of the possibility of such damages.

   9. Accepting Warranty or Additional Liability. While redistributing
      the Work or Derivative Works thereof, You may choose to offer,
      and charge a fee for, acceptance of support, warranty, indemnity,
      or other liability obligations and/or rights consistent with this
      License. However, in accepting such obligations, You may act only
      on Your own behalf and on Your sole responsibility, not on behalf
      of any other Contributor, and only if You agree to indemnify,
      defend, and hold each Contributor harmless for any liability
      incurred by, or claims asserted against, such Contributor by reason
      of your accepting any such warranty or additional liability.

   END OF TERMS AND CONDITIONS

   APPENDIX: How to apply the Apache License to your work.

      To apply the Apache License to your work, attach the following
      boilerplate notice, with the fields enclosed by brackets "[]"
      replaced with your own identifying information. (Don't include
      the brackets!)  The text should be enclosed in the appropriate
      comment syntax for the file format. We also recommend that a
      file or class name and description of purpose be included on the
      same "printed page" as the copyright notice for easier
      identification within third-party archives.

   Copyright [yyyy] [name of copyright owner]

   Licensed under the Apache License, Version 2.0 (the "License");
   you may not use this file except in compliance with the License.
   You may obtain a copy of the License at

       http://www.apache.org/licenses/LICENSE-2.0

   Unless required by applicable law or agreed to in writing, software
   distributed under the License is distributed on an "AS IS" BASIS,
   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
   See the License for the specific language governing permissions and
   limitations under the License.


================================================
FILE: LockFreeLifoQueue.h
================================================

#pragma once

#include <nstd/Atomic.h>
#include <nstd/Memory.h>

template <typename T> class LockFreeLifoQueue
{
public:
  explicit LockFreeLifoQueue(usize capacity)
    : _capacity(capacity)
  {
    _indexMask = capacity;
    for(usize i = 1; i <= sizeof(void*) * 4; i <<= 1)
      _indexMask |= _indexMask >> i;
    _abaOffset = _indexMask + 1;

    _queue = (Node*)Memory::alloc(sizeof(Node) * (capacity + 1));
    for(usize i = 1; i < capacity;)
    {
      Node& node = _queue[i];
      node.abaNextFree = ++i;
    }
    _queue[capacity].abaNextFree = 0;

    _abaFree = 1;
    _abaPushed = 0;
  }

  ~LockFreeLifoQueue()
  {
    for(usize abaPushed = _abaPushed;;)
    {
      usize nodeIndex = abaPushed & _indexMask;
      if(!nodeIndex)
        break;
      Node& node = _queue[nodeIndex];
      abaPushed = node.abaNextPushed;
      (&node.data)->~T();
    }

    Memory::free(_queue);
  }
  
  usize capacity() const {return _capacity;}

  usize size() const {return 0;}

  bool push(const T& data)
  {
    Node* node;
    usize abaFree;
    for(;;)
    {
      abaFree = _abaFree;
      usize nodeIndex = abaFree & _indexMask;
      if(!nodeIndex)
        return false;
      node = &_queue[nodeIndex];
      if(Atomic::compareAndSwap(_abaFree, abaFree, node->abaNextFree + _abaOffset) == abaFree)
        break;
    }

    new (&node->data)T(data);
    for(;;)
    {
      usize abaPushed = _abaPushed;
      node->abaNextPushed = abaPushed;
      if(Atomic::compareAndSwap(_abaPushed, abaPushed, abaFree) == abaPushed)
        return true;
    }
  }

  bool pop(T& result)
  {
    Node* node;
    usize abaPushed;
    for(;;)
    {
      abaPushed = _abaPushed;
      usize nodeIndex = abaPushed & _indexMask;
      if(!nodeIndex)
        return false;
      node = &_queue[nodeIndex];
      if(Atomic::compareAndSwap(_abaPushed, abaPushed, node->abaNextPushed + _abaOffset) == abaPushed)
        break;
    }

    result = node->data;
    (&node->data)->~T();
    abaPushed += _abaOffset;
    for(;;)
    {
      usize abaFree = _abaFree;
      node->abaNextFree = abaFree;
      if(Atomic::compareAndSwap(_abaFree, abaFree, abaPushed) == abaFree)
        return true;
    }
  }

private:
  struct Node
  {
    T data;
    volatile usize abaNextFree;
    volatile usize abaNextPushed;
  };

private:
  usize _indexMask;
  Node* _queue;
  usize _abaOffset;
  usize _capacity;
  char cacheLinePad1[64];
  volatile usize _abaFree;
  char cacheLinePad2[64];
  volatile usize _abaPushed;
  char cacheLinePad3[64];
};


================================================
FILE: LockFreeQueue.h
================================================

#pragma once

#include <nstd/Atomic.h>
#include <nstd/Memory.h>

template <typename T> class LockFreeQueue
{
public:
  explicit LockFreeQueue(usize capacity)
  {
    _capacityMask = capacity - 1;
    for(usize i = 1; i <= sizeof(void*) * 4; i <<= 1)
      _capacityMask |= _capacityMask >> i;
    _capacity = _capacityMask + 1;

    _queue = (Node*)Memory::alloc(sizeof(Node) * _capacity);
    for(usize i = 0; i < _capacity; ++i)
    {
      _queue[i].tail = i;
      _queue[i].head = -1;
    }

    _tail = 0;
    _head = 0;
  }

  ~LockFreeQueue()
  {
    for(usize i = _head; i != _tail; ++i)
      (&_queue[i & _capacityMask].data)->~T();

    Memory::free(_queue);
  }
  
  usize capacity() const {return _capacity;}
  
  usize size() const
  {
    usize head = Atomic::load(_head);
    return _tail - head;
  }
  
  bool push(const T& data)
  {
    Node* node;
    usize next, tail = _tail;
    for(;; tail = next)
    {
      node = &_queue[tail & _capacityMask];
      if(Atomic::load(node->tail) != tail)
        return false;
      if((next = Atomic::compareAndSwap(_tail, tail, tail + 1)) == tail)
        break;
    }
    new (&node->data)T(data);
    Atomic::store(node->head, tail);
    return true;
  }

  bool pop(T& result)
  {
    Node* node;
    usize next, head = _head;
    for(;; head = next)
    {
      node = &_queue[head & _capacityMask];
      if(Atomic::load(node->head) != head)
        return false;
      if((next = Atomic::compareAndSwap(_head, head, head + 1)) == head)
        break;
    }
    result = node->data;
    (&node->data)->~T();
    Atomic::store(node->tail, head + _capacity);
    return true;
  }

private:
  struct Node
  {
    T data;
    usize tail;
    usize head;
  };

private:
  usize _capacityMask;
  Node* _queue;
  usize _capacity;
  char cacheLinePad1[64];
  usize _tail;
  char cacheLinePad2[64];
  usize _head;
  char cacheLinePad3[64];
};


================================================
FILE: LockFreeQueueCpp11.h
================================================

#pragma once

#include <atomic>
#include <cstddef>

template <typename T> class LockFreeQueueCpp11
{
public:
  explicit LockFreeQueueCpp11(size_t capacity)
  {
    _capacityMask = capacity - 1;
    for(size_t i = 1; i <= sizeof(void*) * 4; i <<= 1)
      _capacityMask |= _capacityMask >> i;
    _capacity = _capacityMask + 1;

    _queue = (Node*)new char[sizeof(Node) * _capacity];
    for(size_t i = 0; i < _capacity; ++i)
    {
      _queue[i].tail.store(i, std::memory_order_relaxed);
      _queue[i].head.store(-1, std::memory_order_relaxed);
    }

    _tail.store(0, std::memory_order_relaxed);
    _head.store(0, std::memory_order_relaxed);
  }

  ~LockFreeQueueCpp11()
  {
    for(size_t i = _head; i != _tail; ++i)
      (&_queue[i & _capacityMask].data)->~T();

    delete [] (char*)_queue;
  }
  
  size_t capacity() const {return _capacity;}
  
  size_t size() const
  {
    size_t head = _head.load(std::memory_order_acquire);
    return _tail.load(std::memory_order_relaxed) - head;
  }
  
  bool push(const T& data)
  {
    Node* node;
    size_t tail = _tail.load(std::memory_order_relaxed);
    for(;;)
    {
      node = &_queue[tail & _capacityMask];
      if(node->tail.load(std::memory_order_relaxed) != tail)
        return false;
      if((_tail.compare_exchange_weak(tail, tail + 1, std::memory_order_relaxed)))
        break;
    }
    new (&node->data)T(data);
    node->head.store(tail, std::memory_order_release);
    return true;
  }

  bool pop(T& result)
  {
    Node* node;
    size_t head = _head.load(std::memory_order_relaxed);
    for(;;)
    {
      node = &_queue[head & _capacityMask];
      if(node->head.load(std::memory_order_relaxed) != head)
        return false;
      if(_head.compare_exchange_weak(head, head + 1, std::memory_order_relaxed))
        break;
    }
    result = node->data;
    (&node->data)->~T();
    node->tail.store(head + _capacity, std::memory_order_release);
    return true;
  }

private:
  struct Node
  {
    T data;
    std::atomic<size_t> tail;
    std::atomic<size_t> head;
  };

private:
  size_t _capacityMask;
  Node* _queue;
  size_t _capacity;
  char cacheLinePad1[64];
  std::atomic<size_t> _tail;
  char cacheLinePad2[64];
  std::atomic<size_t> _head;
  char cacheLinePad3[64];
};


================================================
FILE: LockFreeQueueSlow1.h
================================================

#pragma once

#include <nstd/Atomic.h>
#include <nstd/Memory.h>

template <typename T> class LockFreeQueueSlow1
{
public:
  explicit LockFreeQueueSlow1(usize capacity)
  {
    _capacityMask = capacity - 1;
    for(usize i = 1; i <= sizeof(void*) * 4; i <<= 1)
      _capacityMask |= _capacityMask >> i;
    _capacity = _capacityMask + 1;

    _queue = (Node*)Memory::alloc(sizeof(Node) * _capacity);
    for(Node* node = _queue, * end = _queue + _capacity; node < end; ++node)
      node->state = Node::free;

    _freeNodes = _capacity;
    _occupiedNodes = 0;
    _writeIndex = -1;
    _safeWriteIndex = -1;
    _readIndex = -1;
    _safeReadIndex = -1;
  }

  ~LockFreeQueueSlow1()
  {
    for(Node* node = _queue, * end = _queue + _capacity; node < end; ++node)
      switch(node->state)
      {
      case Node::set:
      case Node::occupied:
        (&node->data)->~T();
        break;
      default:
        break;
      }
    Memory::free(_queue);
  }
  
  usize capacity() const {return _capacity;}
  
  usize size() const {return _capacity - _freeNodes;}
  
  bool push(const T& data)
  {
  begin:
    usize freeNodes = _freeNodes;
    if(freeNodes == 0)
      return false; // queue is full
    if(Atomic::compareAndSwap(_freeNodes, freeNodes, freeNodes - 1) != freeNodes)
      goto begin;
    usize writeIndex = Atomic::increment(_writeIndex);
    Node* node = &_queue[writeIndex & _capacityMask];
    ASSERT(node->state == Node::free);
    new (&node->data)T(data);
    Atomic::store(node->state, Node::set);
  commit:
    usize safeWriteIndex = _safeWriteIndex;
    usize nextSafeWriteIndex = safeWriteIndex + 1;
  commitNext:
    node = &_queue[nextSafeWriteIndex & _capacityMask];
    if(node->state == Node::set && Atomic::compareAndSwap(node->state, Node::set, Node::occupied) == Node::set)
    {
      if(Atomic::compareAndSwap(_safeWriteIndex, safeWriteIndex, nextSafeWriteIndex) == safeWriteIndex)
      {
        Atomic::increment(_occupiedNodes);
        safeWriteIndex = nextSafeWriteIndex;
        ++nextSafeWriteIndex;
        goto commitNext;
      }
      else
        node->state = Node::set;
      goto commit;
    }
    return true;
  }
  
  bool pop(T& result)
  {
  begin:
    usize occupiedNodes = _occupiedNodes;
    if(occupiedNodes == 0)
      return false; // queue is empty
    if(Atomic::compareAndSwap(_occupiedNodes, occupiedNodes, occupiedNodes - 1) != occupiedNodes)
      goto begin;
    usize readIndex = Atomic::increment(_readIndex);
    Node* node = &_queue[readIndex & _capacityMask];
    ASSERT(node->state == Node::occupied);
    result = node->data;
    (&node->data)->~T();
    Atomic::store(node->state, Node::unset);
  release:
    usize safeReadIndex = _safeReadIndex;
    usize nextSafeReadIndex = safeReadIndex + 1;
  releaseNext:
    node = &_queue[nextSafeReadIndex & _capacityMask];
    if(node->state == Node::unset && Atomic::compareAndSwap(node->state, Node::unset, Node::free) == Node::unset)
    {
      if(Atomic::compareAndSwap(_safeReadIndex, safeReadIndex, nextSafeReadIndex) == safeReadIndex)
      {
        Atomic::increment(_freeNodes);
        safeReadIndex = nextSafeReadIndex;
        ++nextSafeReadIndex;
        goto releaseNext;
      }
      else
        node->state = Node::unset;
      goto release;
    }
    return true;
  }

private:
  struct Node
  {
    T data;
    enum State
    {
      free,
      set,
      occupied,
      unset,
    };
    volatile int state;
  };

private:
  usize _capacityMask;
  Node* _queue;
  usize _capacity;
  volatile usize _freeNodes;
  volatile usize _occupiedNodes;
  volatile usize _writeIndex;
  volatile usize _safeWriteIndex;
  volatile usize _readIndex;
  volatile usize _safeReadIndex;
};


================================================
FILE: LockFreeQueueSlow2.h
================================================

#pragma once

#include <nstd/Atomic.h>
#include <nstd/Memory.h>

template <typename T> class LockFreeQueueSlow2
{
public:
  explicit LockFreeQueueSlow2(usize capacity)
  {
    _capacityMask = capacity - 1;
    for(usize i = 1; i <= sizeof(void*) * 4; i <<= 1)
      _capacityMask |= _capacityMask >> i;
    _capacity = _capacityMask + 1;

    _queue = (Node*)Memory::alloc(sizeof(Node) * _capacity);
    for(usize i = 0; i < _capacity; ++i)
    {
      _queue[i].tail = i;
      _queue[i].head = i - 1;
    }

    _tail = 0;
    _head = 0;
  }

  ~LockFreeQueueSlow2()
  {
    for(usize i = _head; i != _tail; ++i)
      (&_queue[i & _capacityMask].data)->~T();

    Memory::free(_queue);
  }
  
  usize capacity() const {return _capacity;}
  
  usize size() const
  {
    usize head = Atomic::load(_head);
    return _tail - head;
  }
  
  bool push(const T& data)
  {
  begin:
    usize tail = _tail;
    Node* node = &_queue[tail & _capacityMask];
    usize newTail = tail + 1;
    usize nodeTail = node->tail;
    if(nodeTail == tail)
    {
      nodeTail = Atomic::compareAndSwap(node->tail, tail, newTail);
      if(nodeTail == tail)
      {
        Atomic::compareAndSwap(_tail, tail, newTail);
        new (&node->data)T(data);
        Atomic::store(node->head, tail);
        return true;
      }
    }
    if(nodeTail == newTail)
    {
      Atomic::compareAndSwap(_tail, tail, newTail);
      goto begin;
    }
    else
      return false;
  }

  bool pop(T& result)
  {
  begin:
    usize head = _head;
    Node* node = &_queue[head & _capacityMask];
    usize newHead = head + 1;
    usize nodeHead = node->head;
    if(nodeHead == head)
    {
      nodeHead = Atomic::compareAndSwap(node->head, head, newHead);
      if(nodeHead == head)
      {
        Atomic::compareAndSwap(_head, head, newHead);
        result = node->data;
        (&node->data)->~T();
        Atomic::store(node->tail, head + _capacity);
        return true;
      }
    }
    if(nodeHead == newHead)
    {
      Atomic::compareAndSwap(_head, head, newHead);
      goto begin;
    }
    else
      return false;
  }

private:
  struct Node
  {
    T data;
    volatile usize head;
    volatile usize tail;
  };

private:
  usize _capacityMask;
  Node* _queue;
  usize _capacity;
  char cacheLinePad1[64];
  volatile usize _head;
  char cacheLinePad2[64];
  volatile usize _tail;
  char cacheLinePad3[64];
};


================================================
FILE: LockFreeQueueSlow3.h
================================================

#pragma once

#include <nstd/Atomic.h>
#include <nstd/Memory.h>

template <typename T> class LockFreeQueueSlow3
{
public:
  explicit LockFreeQueueSlow3(usize capacity)
  {
    _capacityMask = capacity - 1;
    for(usize i = 1; i <= sizeof(void*) * 4; i <<= 1)
      _capacityMask |= _capacityMask >> i;
    _capacity = _capacityMask + 1;

    _queue = (Node*)Memory::alloc(sizeof(Node) * _capacity);
    for(usize i = 0; i < _capacity; ++i)
      _queue[i].state = 0;

    _tail = 0;
    _head = 0;
  }

  ~LockFreeQueueSlow3()
  {
    for(usize i = _head; i != _tail; ++i)
      (&_queue[i & _capacityMask].data)->~T();

    Memory::free(_queue);
  }
  
  usize capacity() const {return _capacity;}
  
  usize size() const
  {
    usize head = Atomic::load(_head);
    return _tail - head;
  }
  
  bool push(const T& data)
  {
    for(;;)
    {
      usize tail = _tail;
      Node* node = &_queue[tail & _capacityMask];
      switch(Atomic::compareAndSwap(node->state, 0, 2))
      {
      case 2:
        continue;
      case 0:
        break;
      default:
        return false;
      }
      if(Atomic::compareAndSwap(_tail, tail, tail + 1) == tail)
      {
        new (&node->data)T(data);
        Atomic::store(node->state, 1);
        return true;
      }
      else
        node->state = 0;
    }
  }

  bool pop(T& result)
  {
    for(;;)
    {
      usize head = _head;
      Node* node = &_queue[head & _capacityMask];
      switch(Atomic::compareAndSwap(node->state, 1, 3))
      {
      case 3:
        continue;
      case 1:
        break;
      default:
        return false;
      }
      if(Atomic::compareAndSwap(_head, head, head + 1) == head)
      {
        result = node->data;
        (&node->data)->~T();
        Atomic::store(node->state, 0);
        return true;
      }
      else
        node->state = 1;
    }
  }

private:
  struct Node
  {
    T data;
    volatile int state;
  };

private:
  usize _capacityMask;
  Node* _queue;
  usize _capacity;
  char cacheLinePad1[64];
  volatile usize _tail;
  char cacheLinePad2[64];
  volatile usize _head;
  char cacheLinePad3[64];
};


================================================
FILE: Marefile
================================================


if(tool == "vcxproj") {
  platforms = { "Win32", "x64" }
}

buildDir = "Build/$(configuration)/.$(target)"

targets = {

  LockFreeQueue = cppApplication + {
    dependencies = { "libnstd" }
    includePaths = { "Ext/libnstd/include" }
    libPaths = { "Build/$(configuration)/.libnstd" }
    libs = { "nstd" }
    root = "Src"
    files = {
      "*.cpp" = cppSource
      "*.h"
    }
    if tool == "vcxproj" {
      linkFlags += { "/SUBSYSTEM:CONSOLE" }
    }
    if platform == "Linux" {
      libs += { "pthread", "rt" }
      cppFlags += { "-std=c++11" }
    }
    defines -= "NDEBUG"
  }
  
  include "Ext/libnstd/libnstd.mare"
  libnstd += {
    folder = "Ext"
  }
}


================================================
FILE: MutexLockQueue.h
================================================

#pragma once

#include <nstd/Mutex.h>
#include <nstd/Memory.h>

template <typename T> class MutexLockQueue
{
public:
  explicit MutexLockQueue(usize capacity)
  {
    _capacityMask = capacity - 1;
    for(usize i = 1; i <= sizeof(void*) * 4; i <<= 1)
      _capacityMask |= _capacityMask >> i;
    _capacity = _capacityMask + 1;

    _queue = (Node*)Memory::alloc(sizeof(Node) * _capacity);

    _head = 0;
    _tail = 0;
  }

  ~MutexLockQueue()
  {
    for(usize i = _head; i != _tail; ++i)
      (&_queue[i & _capacityMask].data)->~T();
    Memory::free(_queue);
  }
  
  usize capacity() const {return _capacity;}
  
  usize size() const
  {
    usize result;
    _mutex.lock();
    result = _tail - _head;
    _mutex.unlock();
    return result;
  }
  
  bool push(const T& data)
  {
    _mutex.lock();
    if(_tail - _head == _capacity)
    {
      _mutex.unlock();
      return false; // queue is full
    }
    Node& node = _queue[(_tail++) & _capacityMask];
    new (&node.data)T(data);
    _mutex.unlock();
    return true;
  }
  
  bool pop(T& result)
  {
    _mutex.lock();
    if(_head == _tail)
    {
      _mutex.unlock();
      return false; // queue is empty
    }
    Node& node = _queue[(_head++) & _capacityMask];
    result = node.data;
    (&node.data)->~T();
    _mutex.unlock();
    return true;
  }

private:
  struct Node
  {
    T data;
  };

private:
  usize _capacityMask;
  Node* _queue;
  usize _capacity;
  usize _head;
  usize _tail;
  mutable Mutex _mutex;
};


================================================
FILE: README.md
================================================
LockFreeQueue
=============

Here, I am testing some multi-producer multi-consumer bounded ring buffer FIFO queue implementations for fun.

* [LockFreeQueueCpp11.h](LockFreeQueueCpp11.h) - The fastest lock free queue I have managed to implement.
* [LockFreeQueue.h](LockFreeQueue.h) - The fastest lock free queue I have managed to implement without c++11. It is equally fast as LockFreeQueueCpp11.h.
* [mpmc_bounded_queue.h](mpmc_bounded_queue.h) - Bounded MPMC queue by [Dmitry Vyukov, 2011]
* [LockFreeQueueSlow1.h](LockFreeQueueSlow1.h) - My first attempt at implementing a lock free queue. It is working correctly, but it is a lot slower than LockFreeQueue.h.
* [LockFreeQueueSlow2.h](LockFreeQueueSlow2.h) - A lock free queue based on [John D. Valois, 1994]. The queue uses Valois' algorithm adapted to a ring buffer structure with some modifications to tackle the ABA-Problem.
* [LockFreeQueueSlow3.h](LockFreeQueueSlow3.h) - Another lock free queue almost as fast as LockFreeQueue.h.
* [MutexLockQueue.h](MutexLockQueue.h) - A naive queue implementation that uses a conventional mutex lock (CriticalSecion / pthread-Mutex).
* [SpinLockQueue.h](SpinLockQueue.h) - A naive queue implementation that uses an atomic TestAndSet-lock.

And for the fun of it, here is a multi-producer multi-consumer LIFO queue:

* [LockFreeLifoQueue.h](LockFreeLifoQueue.h) - A lock free multi-producer multi-consumer bounded LIFO queue.

#### References

[John D. Valois, 1994] - Implementing Lock-Free Queues<br/>
[Dmitry Vyukov, 2011] - [Bounded MPMC queue](http://www.1024cores.net/home/lock-free-algorithms/queues/bounded-mpmc-queue)

================================================
FILE: SpinLockQueue.h
================================================

#pragma once

#include <nstd/Atomic.h>
#include <nstd/Memory.h>

template <typename T> class SpinLockQueue
{
public:
  explicit SpinLockQueue(usize capacity)
  {
    _capacityMask = capacity - 1;
    for(usize i = 1; i <= sizeof(void*) * 4; i <<= 1)
      _capacityMask |= _capacityMask >> i;
    _capacity = _capacityMask + 1;

    _queue = (Node*)Memory::alloc(sizeof(Node) * _capacity);

    _lock = 0;
    _head = 0;
    _tail = 0;
  }

  ~SpinLockQueue()
  {
    for(usize i = _head; i != _tail; ++i)
      (&_queue[i & _capacityMask].data)->~T();
    Memory::free(_queue);
  }
  
  usize capacity() const {return _capacity;}
  
  usize size() const
  {
    usize result;
    while(Atomic::testAndSet(_lock) != 0);
    result = _tail - _head;
    Atomic::store(_lock, 0);
    return result;
  }
  
  bool push(const T& data)
  {
    while(Atomic::testAndSet(_lock) != 0);
    if(_tail - _head == _capacity)
    {
      _lock = 0;
      return false; // queue is full
    }
    Node& node = _queue[(_tail++) & _capacityMask];
    new (&node.data)T(data);
    Atomic::store(_lock, 0);
    return true;
  }
  
  bool pop(T& result)
  {
    while(Atomic::testAndSet(_lock) != 0);
    if(_head == _tail)
    {
      _lock = 0;
      return false; // queue is empty
    }
    Node& node = _queue[(_head++) & _capacityMask];
    result = node.data;
    (&node.data)->~T();
    Atomic::store(_lock, 0);
    return true;
  }

private:
  struct Node
  {
    T data;
  };

private:
  usize _capacityMask;
  Node* _queue;
  usize _capacity;
  usize _head;
  usize _tail;
  mutable volatile int32 _lock;
};


================================================
FILE: Test.cpp
================================================

#include <nstd/Console.h>
#include <nstd/Debug.h>
#include <nstd/Thread.h>
#include <nstd/List.h>
#include <nstd/Time.h>

#include "LockFreeQueueCpp11.h"
#include "LockFreeQueue.h"
#include "LockFreeQueueSlow1.h"
#include "LockFreeQueueSlow2.h"
#include "LockFreeQueueSlow3.h"
#include "MutexLockQueue.h"
#include "SpinLockQueue.h"
#include "mpmc_bounded_queue.h"
#include "LockFreeLifoQueue.h"

static const int testItems = 250000 * 64 / 3 * 10;
static const int testThreadConsumerThreads = 8;
static const int testThreadProducerThreads = 8;
static const int testItemsPerConsumerThread = testItems / testThreadConsumerThreads;
static const int testItemsPerProducerThread = testItems / testThreadProducerThreads;

template<typename T> class IQueue
{
public:
  virtual usize size() const = 0;
  virtual usize capacity() const = 0;
  virtual bool push(const T& data) = 0;
  virtual bool pop(T& result) = 0;
};

template<typename T, class Q> class TestQueue : public IQueue<T>
{
public:
  TestQueue(usize size) : queue(size) {}
  usize size() const {return queue.size();}
  usize capacity() const {return queue.capacity();}
  bool push(const T& data) {return queue.push(data);}
  bool pop(T& result) {return queue.pop(result);}
private:
  Q queue;
};

volatile usize producerSum;
volatile usize consumerSum;
volatile int64 maxPushDuration;
volatile int64 maxPopDuration;

uint producerThread(void* param)
{
  IQueue<int>* queue = (IQueue<int>*)param;
  for(int i = 0; i < testItemsPerProducerThread; ++i)
  {
    for(;;)
    {
      int64 startTime = Time::microTicks();
      while(!queue->push(i))
      {
        Thread::yield();
        startTime = Time::microTicks();
      }
      {
        int64 duration = Time::microTicks() - startTime;
        for(;;)
        {
          int64 lmaxPushDuration = maxPushDuration;
          if(duration <= lmaxPushDuration || Atomic::compareAndSwap(maxPushDuration, lmaxPushDuration, duration) == lmaxPushDuration)
            break;
        }
        break;
      }
    }
    Atomic::fetchAndAdd(producerSum, i);
  }
  return 0;
}

uint consumerThread(void* param)
{
  IQueue<int>* queue = (IQueue<int>*)param;
  int val;
  for(int i = 0; i < testItemsPerConsumerThread; ++i)
  {
    for(;;)
    {
      int64 startTime = Time::microTicks();
      while(!queue->pop(val))
      {
        Thread::yield();
        startTime = Time::microTicks();
      }
      {
        int64 duration = Time::microTicks() - startTime;
        for(;;)
        {
          int64 lmaxPopDuration = maxPopDuration;
          if(duration <= lmaxPopDuration || Atomic::compareAndSwap(maxPopDuration, lmaxPopDuration, duration) == lmaxPopDuration)
            break;
        }
        break;
      }
    }
    Atomic::fetchAndAdd(consumerSum, val);
  }
  return 0;
}

template<class Q> void testQueue(const String& name, bool fifo = false)
{
  Console::printf(_T("Testing %s... \n"), (const tchar*)name);

  volatile int32 int32_ = 0;
  volatile uint32 uint32_ = 0;
  ASSERT(Atomic::compareAndSwap(int32_, 0, 1) == 0);
  ASSERT(int32_ == 1);
  ASSERT(Atomic::compareAndSwap(uint32_, 0, 1) == 0);
  ASSERT(uint32_ == 1);

  volatile int64 int64_ = 0;
  volatile uint64 uint64_ = 0;
  ASSERT(Atomic::compareAndSwap(int64_, 0, 1) == 0);
  ASSERT(int64_ == 1);
  ASSERT(Atomic::compareAndSwap(uint64_, 0, 1) == 0);
  ASSERT(uint64_ == 1);

  {
    TestQueue<int, Q> queue(10000);
    int result;
    ASSERT(queue.capacity() >= 10000);
    ASSERT(!queue.pop(result));
    ASSERT(queue.push(42));
    ASSERT(queue.pop(result));
    ASSERT(result == 42);
    ASSERT(!queue.pop(result));
  }

  {
    TestQueue<int, Q> queue(2);
    int result;
    ASSERT(queue.capacity() >= 2);
    ASSERT(!queue.pop(result));
    ASSERT(queue.push(42));
    ASSERT(queue.push(43));
    ASSERT(queue.pop(result));
    ASSERT(result == (fifo ? 43 : 42));
    ASSERT(queue.pop(result));
    ASSERT(result == (fifo ? 42 : 43));
    ASSERT(!queue.pop(result));
    ASSERT(queue.push(44));
    ASSERT(queue.push(45));
    ASSERT(queue.pop(result));
    ASSERT(result == (fifo ? 45 : 44));
    ASSERT(queue.push(47));
  }

  producerSum = 0;
  consumerSum = 0;
  maxPushDuration = 0;
  maxPopDuration = 0;

  int64 microStartTime = Time::microTicks();
  {
    TestQueue<int, Q> queue(100);
    List<Thread*> threads;
    for(int i = 0; i < testThreadProducerThreads; ++i)
    {
      Thread* thread = new Thread;
      thread->start(producerThread, &queue);
      threads.append(thread);
    }
    for(int i = 0; i < testThreadConsumerThreads; ++i)
    {
      Thread* thread = new Thread;
      thread->start(consumerThread, &queue);
      threads.append(thread);
    }
    for(List<Thread*>::Iterator i = threads.begin(), end = threads.end(); i != end; ++i)
    {
      Thread* thread = *i;
      thread->join();
      delete thread;
    }
    ASSERT(queue.size() == 0);
    ASSERT(producerSum == consumerSum);
  }
  int64 microDuration = Time::microTicks() - microStartTime;
  Console::printf(_T("%lld ms, maxPush: %lld microseconds, maxPop: %lld microseconds\n"), microDuration / 1000, maxPushDuration, maxPopDuration);
}

int main(int argc, char* argv[])
{
  for(int i = 0; i < 3; ++i)
  {
    Console::printf(_T("--- Run %d ---\n"), i);
    testQueue<LockFreeQueueCpp11<int> >("LockFreeQueueCpp11");
    testQueue<mpmc_bounded_queue<int> >("mpmc_bounded_queue");
    testQueue<LockFreeQueue<int> >("LockFreeQueue");
    testQueue<LockFreeQueueSlow1<int> >("LockFreeQueueSlow1");
    testQueue<LockFreeQueueSlow2<int> >("LockFreeQueueSlow2");
    testQueue<LockFreeQueueSlow3<int> >("LockFreeQueueSlow3");
    testQueue<MutexLockQueue<int> >("MutexLockQueue");
    testQueue<SpinLockQueue<int> >("SpinLockQueue");
    testQueue<LockFreeLifoQueue<int> >("LockFreeLifoQueue", true);
  }

  return 0;
}


================================================
FILE: generate
================================================
#!/bin/sh

[ ! -f "Build/Debug/.mare/mare" ] && Ext/mare/compile --buildDir=Build/Debug/.mare --outputDir=Build/Debug/.mare --sourceDir=Ext/mare/src
Build/Debug/.mare/mare $@



================================================
FILE: generate.bat
================================================
@echo off

if not exist Build\Debug\.mare\mare.exe call Ext\mare\compile.bat --buildDir=Build/Debug/.mare --outputDir=Build/Debug/.mare --sourceDir=Ext/mare/src
if not "%1"=="" (Build\Debug\.mare\mare.exe %*) else Build\Debug\.mare\mare.exe --vcxproj=2013


================================================
FILE: mpmc_bounded_queue.h
================================================
/*  Multi-producer/multi-consumer bounded queue.
 *  Copyright (c) 2010-2011, Dmitry Vyukov. All rights reserved.
 *  Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
 *     1. Redistributions of source code must retain the above copyright notice, this list of
 *        conditions and the following disclaimer.
 *     2. Redistributions in binary form must reproduce the above copyright notice, this list
 *        of conditions and the following disclaimer in the documentation and/or other materials
 *        provided with the distribution.
 *  THIS SOFTWARE IS PROVIDED BY DMITRY VYUKOV "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
 *  THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
 *  DMITRY VYUKOV OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 *  (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 *  HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
 *  ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *  The views and conclusions contained in the software and documentation are those of the authors and should not be interpreted
 *  as representing official policies, either expressed or implied, of Dmitry Vyukov.
 */ 

#include <atomic>
#include <cassert>

template<typename T>
class mpmc_bounded_queue
{
private:
  static inline size_t nextPowerOfTwo(size_t buffer_size)
  {
    size_t result = buffer_size - 1;
    for(size_t i = 1; i <= sizeof(void*) * 4; i <<= 1)
      result |= result >> i;
    return result + 1;
  }
public:
  mpmc_bounded_queue(size_t buffer_size)
    : buffer_(new cell_t [nextPowerOfTwo(buffer_size)])
    , buffer_mask_(nextPowerOfTwo(buffer_size) - 1)
  {
    buffer_size = buffer_mask_ + 1;
    assert((buffer_size >= 2) &&
      ((buffer_size & (buffer_size - 1)) == 0));
    for (size_t i = 0; i != buffer_size; i += 1)
      buffer_[i].sequence_.store(i, std::memory_order_relaxed);
    enqueue_pos_.store(0, std::memory_order_relaxed);
    dequeue_pos_.store(0, std::memory_order_relaxed);
  }

  ~mpmc_bounded_queue()
  {
    delete [] buffer_;
  }

  size_t size() const
  {
    size_t head = dequeue_pos_.load(std::memory_order_acquire);;
    return enqueue_pos_.load(std::memory_order_relaxed) - head;
  }

  size_t capacity() const
  {
    return buffer_mask_ + 1;
  }

  bool push(T const& data)
  {
    cell_t* cell;
    size_t pos = enqueue_pos_.load(std::memory_order_relaxed);
    for (;;)
    {
      cell = &buffer_[pos & buffer_mask_];
      size_t seq = 
        cell->sequence_.load(std::memory_order_acquire);
      intptr_t dif = (intptr_t)seq - (intptr_t)pos;
      if (dif == 0)
      {
        if (enqueue_pos_.compare_exchange_weak
            (pos, pos + 1, std::memory_order_relaxed))
          break;
      }
      else if (dif < 0)
        return false;
      else
        pos = enqueue_pos_.load(std::memory_order_relaxed);
    }
    cell->data_ = data;
    cell->sequence_.store(pos + 1, std::memory_order_release);
    return true;
  }

  bool pop(T& data)
  {
    cell_t* cell;
    size_t pos = dequeue_pos_.load(std::memory_order_relaxed);
    for (;;)
    {
      cell = &buffer_[pos & buffer_mask_];
      size_t seq = 
        cell->sequence_.load(std::memory_order_acquire);
      intptr_t dif = (intptr_t)seq - (intptr_t)(pos + 1);
      if (dif == 0)
      {
        if (dequeue_pos_.compare_exchange_weak
            (pos, pos + 1, std::memory_order_relaxed))
          break;
      }
      else if (dif < 0)
        return false;
      else
        pos = dequeue_pos_.load(std::memory_order_relaxed);
    }
    data = cell->data_;
    cell->sequence_.store
      (pos + buffer_mask_ + 1, std::memory_order_release);
    return true;
  }

private:
  struct cell_t
  {
    std::atomic<size_t>   sequence_;
    T                     data_;
  };
  static size_t const     cacheline_size = 64;
  typedef char            cacheline_pad_t [cacheline_size];
  cacheline_pad_t         pad0_;
  cell_t* const           buffer_;
  size_t const            buffer_mask_;
  cacheline_pad_t         pad1_;
  std::atomic<size_t>     enqueue_pos_;
  cacheline_pad_t         pad2_;
  std::atomic<size_t>     dequeue_pos_;
  cacheline_pad_t         pad3_;
  mpmc_bounded_queue(mpmc_bounded_queue const&);
  void operator = (mpmc_bounded_queue const&);
};