std::atomic_flag_test_and_set, std::atomic_flag_test_and_set_explicit

From cppreference.com
Defined in header <atomic>
bool atomic_flag_test_and_set( volatile std::atomic_flag* p );
(1) (since C++11)
bool atomic_flag_test_and_set( std::atomic_flag* p );
(2) (since C++11)
bool atomic_flag_test_and_set_explicit( volatile std::atomic_flag* p,
                                        std::memory_order order );
(3) (since C++11)
bool atomic_flag_test_and_set_explicit( std::atomic_flag* p,
                                        std::memory_order order );
(4) (since C++11)

Atomically changes the state of a std::atomic_flag pointed to by p to set (true) and returns the value it held before.

Contents

[edit] Parameters

p - pointer to std::atomic_flag to access
order - the memory synchronization order for this operation

[edit] Return value

The value previously held by the flag pointed to by p

[edit] Exceptions

noexcept specification:  
noexcept
  (since C++11)

[edit] Possible implementation

First version
bool atomic_flag_test_and_set(volatile std::atomic_flag* p)
{
    return p->test_and_set();
}
Second version
bool atomic_flag_test_and_set(std::atomic_flag* p)
{
    return p->test_and_set();
}
Third version
bool atomic_flag_test_and_set_explicit(volatile std::atomic_flag* p, 
                                       std::memory_order order)
{
    return p->test_and_set(order);
}
Fourth version
bool atomic_flag_test_and_set_explicit(std::atomic_flag* p, 
                                       std::memory_order order) 
{
    return p->test_and_set(order);
}

[edit] Example

A spinlock mutex can be implemented in userspace using an atomic_flag

#include <thread>
#include <vector>
#include <iostream>
#include <atomic>
 
std::atomic_flag lock = ATOMIC_FLAG_INIT;
 
void f(int n)
{
    for(int cnt = 0; cnt < 100; ++cnt) {
        while(std::atomic_flag_test_and_set_explicit(&lock, std::memory_order_acquire))
             ; // spin until the lock is acquired
        std::cout << "Output from thread " << n << '\n';
        std::atomic_flag_clear_explicit(&lock, std::memory_order_release);
    }
}
int main()
{
    std::vector<std::thread> v;
    for (int n = 0; n < 10; ++n) {
        v.emplace_back(f, n);
    }
    for (auto& t : v) {
        t.join();
    }
}

Output:

Output from thread 2
Output from thread 6
Output from thread 7
...<exactly 1000 lines>...

[edit] See also

(C++11)
the lock-free boolean atomic type
(class)
atomically sets the value of the flag to false
(function)
(C++11)
defines memory ordering constraints for the given atomic operation
(typedef)