什么是锁定多个std :: mutex'es的最好的方法？ [英] What's the best way to lock multiple std::mutex'es?

问题描述

当我们要锁定多个 std :: mutex '时，我们使用 std :: lock（） 。但 std :: lock（）不提供RAII功能。

When we want to lock multiple std::mutex'es, we use std::lock(). But std::lock() does not provide RAII feature.

当我们要锁定 std :: mutex 在RAII方式，我们使用 std :: lock_guard 。但 std :: lock_guard 不能安全地锁定多个 std :: mutex 。

When we want to lock a std::mutex in RAII way, we use std::lock_guard. But std::lock_guard cannot lock multiple std::mutex'es safely.

有什么方法可以利用这两种方法的优点，以RAII方式锁定多个 std :: mutex '？

Is there any way to take the advantages of both methods, to lock multiple std::mutex'es in RAII way?

推荐答案

是的，您可以使用 std :: unique_lock 与 std :: defer_lock 。它告诉unique_lock不立即锁定互斥体，而是要构建RAII包装器。

Yes, you can use a std::unique_lock with std::defer_lock. It tells the unique_lock to not lock the mutex immediately, but to build the RAII wrapper.

std::unique_lock<std::mutex> lk1(mutex1, std::defer_lock);
std::unique_lock<std::mutex> lk2(mutex2, std::defer_lock);
std::lock(lk1, lk2);

由于其变化性质 std :: lock 不仅绑定到两个参数，但可以使用与您的编译器支持的参数一样多。

Due to its variadic nature std::lock is not bound to only two arguments but can be used with as many arguments as your compiler has support for.

Howard Hinnant也指出了一个关于效能的有趣事实，您可以检查这个链接，如果你有兴趣。他解决了性能问题，并显示 std :: lock 可以有效地实现，我也可以推荐阅读该帖子中的所有注释。

Howard Hinnant also pointed out an interesting fact about performance, you can check this link if you are interested. He addresses performance concerns and shows that std::lock can be implemented efficiently, I can also recommend to read all the comments in that post.

这篇关于什么是锁定多个std :: mutex'es的最好的方法？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！