＜atomic＞ 注释7：函 _Combine_cas_memory_orders ()，_Atomic_reinterpret_as ()，_Load_barrier ()

（19） 函 _Combine_cas_memory_orders ()，该函数似乎是为了给交换和比较操作指定一个合适的内存约束级别。源码如下：

// Finds upper bound of a compare/exchange memory order pair, according to the following partial order:
// 查找上界，根据以下部分顺序比较 / 交换内存订单对
//     seq_cst              5
//        |                 |
//     acq_rel              4
//     /     \            /   \
// acquire  release      2     3
//    |       |          |     |
// consume    |          1     |
//     \     /            \   /
//     relaxed              0
// 
// 下面函数里的二维数组里就是这些内容： _Success 是横坐标，_Failure 是纵坐标
// 0    0 1 2 3 4 5 
// 1    1 1 2 4 4 5
// 2    2 2 2 4 4 5   
// 3    3 4 4 3 4 5
// 4    4 4 4 4 4 5
// 5    5 5 5 5 5 5
// 
// enum class memory_order : int { relaxed0, consume1, acquire2, release3, acq_rel4, seq_cst5 };
inline memory_order _Combine_cas_memory_orders( const memory_order _Success, const memory_order _Failure) 
{

    static constexpr memory_order _Combined_memory_orders[6][6] = {// combined upper bounds
        {memory_order_relaxed, memory_order_consume, memory_order_acquire, 
            memory_order_release, memory_order_acq_rel, memory_order_seq_cst},

        {memory_order_consume, memory_order_consume, memory_order_acquire,
            memory_order_acq_rel, memory_order_acq_rel, memory_order_seq_cst},

        {memory_order_acquire, memory_order_acquire, memory_order_acquire, 
            memory_order_acq_rel, memory_order_acq_rel, memory_order_seq_cst},

        {memory_order_release, memory_order_acq_rel, memory_order_acq_rel, 
            memory_order_release, memory_order_acq_rel, memory_order_seq_cst},

        {memory_order_acq_rel, memory_order_acq_rel, memory_order_acq_rel, 
            memory_order_acq_rel, memory_order_acq_rel, memory_order_seq_cst},

        {memory_order_seq_cst, memory_order_seq_cst, memory_order_seq_cst, 
            memory_order_seq_cst, memory_order_seq_cst, memory_order_seq_cst}  };

    _Check_memory_order(_Success);
    _Check_load_memory_order(_Failure);

    return _Combined_memory_orders[static_cast<int>(_Success)][static_cast<int>(_Failure)];
}

（20）函数 _Atomic_reinterpret_as ()，本函数实现整数类型与指针类型到整数类型的数据转换，且只允许往大的方向转换：

++ 以及：

++ 源代码如下：

（21）函数 _Load_barrier () ，此函数看来是为了为接下来的原子操作，设置内存栅栏，以确保原子操作的实现，源代码如下：

++ 测试结果：

（22）

谢谢