multithreading - pthread_cond_wait 在等待时自动原子地解锁互斥锁

Question

从这里开始：https ://computing.llnl.gov/tutorials/pthreads/#ConVarSignal

请注意，pthread_cond_wait 例程将在等待时自动且原子地解锁互斥锁。

以下子代码来自同一链接（由我格式化）：

pthread_mutex_lock(&count_mutex);

  while (count<COUNT_LIMIT) 
  {
      pthread_cond_wait(&count_threshold_cv, &count_mutex);
      printf("watch_count(): thread %ld Condition signal received.\n", my_id);
      count += 125;
      printf("watch_count(): thread %ld count now = %d.\n", my_id, count);
  }

pthread_mutex_unlock(&count_mutex);  

问题：
当它说pthread_cond_wait在等待时会自动解锁互斥锁，那么为什么我们必须在上面的代码末尾显式指定函数？pthread_mutex_unlock

我错过了什么？

Answer 1

当pthread_cond_wait解除阻塞时，它再次持有锁。例如，假设您绕过循环两次，您会在互斥体上获得以下锁定/解锁序列：

lock

# Around loop twice:
    wait (unlock)
    awaken (holding lock)
    wait (unlock)
    awaken (holding lock)

# loop done, still holding lock

unlock

如果您没有最后一次解锁，那么下次其他人想要获得锁时，您将陷入死锁。