std::atomic_fetch_sub, std::atomic_fetch_sub_explicit

Defined in header `<atomic>`
	(1)
template< class T > T atomic_fetch_sub( std::atomic<T>* obj, typename std::atomic<T>::difference_type arg ) noexcept;
template< class T > T atomic_fetch_sub( volatile std::atomic<T>* obj, typename std::atomic<T>::difference_type arg ) noexcept;
	(2)
template< class T > T atomic_fetch_sub_explicit( std::atomic<T>* obj, typename std::atomic<T>::difference_type arg, std::memory_order order ) noexcept;
template< class T > T atomic_fetch_sub_explicit( volatile std::atomic<T>* obj, typename std::atomic<T>::difference_type arg, std::memory_order order ) noexcept;

Performs atomic subtraction. Atomically subtracts arg from the value pointed to by obj and returns the value obj held previously. The operation is performed as if the following was executed:

1) obj->fetch_sub(arg)

2) obj->fetch_sub(arg, order)

Parameters

obj	-	pointer to the atomic object to modify
arg	-	the value to subtract from the value stored in the atomic object
order	-	the memory sycnhronization ordering for this operation: all values are permitted.

Return value

The value immediately preceding the effects of this function in the modification order of *obj.

Possible implementation

template< class T >
T atomic_fetch_sub( std::atomic<T>* obj,
                    typename std::atomic<T>::difference_type arg ) noexcept
{
    return obj->fetch_sub(arg);
}

Example

Multiple threads may use fetch_sub to concurrently process an indexed container.

#include <string>
#include <thread>
#include <vector>
#include <iostream>
#include <atomic>
#include <numeric>
 
const int N = 50;
std::atomic<int> cnt;
std::vector<int> data(N);
 
void reader(int id) 
{
    for (;;) {
        int idx = atomic_fetch_sub_explicit(&cnt, 1, std::memory_order_relaxed);
        if (idx >= 0) {
            std::cout << "reader " << std::to_string(id) << " processed item "
                      << std::to_string(data[idx]) << '\n';
        } else {
            std::cout << "reader " << std::to_string(id) << " done\n";
            break;
        }
    }
}
 
int main()
{
    std::iota(data.begin(), data.end(), 1);
    cnt = data.size() - 1;
 
    std::vector<std::thread> v;
    for (int n = 0; n < 5; ++n) {
        v.emplace_back(reader, n);
    }
    for (auto& t : v) {
        t.join();
    }
}

Output:

reader 2 processed item 50
reader 1 processed item 44
reader 4 processed item 46
<....>
reader 0 done
reader 4 done
reader 3 done

Defect reports

The following behavior-changing defect reports were applied retroactively to previously published C++ standards.

DR	Applied to	Behavior as published	Correct behavior
P0558R1	C++11	exact type match required because `T` is deduced from multiple arguments	`T` is deduced from the `atomic` argument only

fetch_sub	atomically subtracts the argument from the value stored in the atomic object and obtains the value held previously (public member function of `std::atomic<T>`)
atomic_fetch_addatomic_fetch_add_explicit (C++11)(C++11)	adds a non-atomic value to an atomic object and obtains the previous value of the atomic (function template)