Section
std::atomic_flag
std::atomic_flag is an atomic boolean type. Unlike all specializations of std::atomic, it is guaranteed to be lock-free. Unlike std::atomic
# Declarations
class atomic_flag;
(since C++11)
# Example
#include <atomic>
#include <iostream>
#include <mutex>
#include <thread>
#include <vector>
class mutex
{
std::atomic_flag m_{};
public:
void lock() noexcept
{
while (m_.test_and_set(std::memory_order_acquire))
#if defined(__cpp_lib_atomic_wait) && __cpp_lib_atomic_wait >= 201907L
// Since C++20, locks can be acquired only after notification in the unlock,
// avoiding any unnecessary spinning.
// Note that even though wait guarantees it returns only after the value has
// changed, the lock is acquired after the next condition check.
m_.wait(true, std::memory_order_relaxed)
#endif
;
}
bool try_lock() noexcept
{
return !m_.test_and_set(std::memory_order_acquire);
}
void unlock() noexcept
{
m_.clear(std::memory_order_release);
#if defined(__cpp_lib_atomic_wait) && __cpp_lib_atomic_wait >= 201907L
m_.notify_one();
#endif
}
};
static mutex m;
static int out{};
void f(std::size_t n)
{
for (std::size_t cnt{}; cnt < 40; ++cnt)
{
std::lock_guard lock{m};
std::cout << n << ((++out % 40) == 0 ? '\n' : ' ');
}
}
int main()
{
std::vector<std::thread> v;
for (std::size_t n{}; n < 10; ++n)
v.emplace_back(f, n);
for (auto &t : v)
t.join();
}