P2628 Extend barrier APIs with memory_order

[wg21bot]

P2628R0 Extend barrier APIs with memory_order (Gonzalo Brito Gadeschi)

[ogiroux]

Poll: We like synchronization primitives that don't always synchronize
in C++ (Be it with memory_order arguments or as separate types.)
(Don't always establish "synchronize-with" into the memory model.)
SF F N A SA
1 2 6 5 1

Consensus we do not like

Poll: We can use memory_order on APIs beyond atomic<> in C++
(It's not precluded, we have to see proposals.)
SF F N A SA
1 6 2 6 0

No consensus