co VS libco

Yeah, at some point, would just say if you want stackful coroutine, go to Go and stackless one, use C++.

BTW, here is my macros for stackless coroutine in C: https://github.com/liuliu/co/blob/master/example.c

Higan's coroutines library libco is low-overhead cooperative multi-tasking that avoids OS context switches. As it says in the readme:

Don't know about other systems, but bsnes/higan is probably the most widely known SNES emulator that operates at the cycle level. It uses libco to implement green threads. More details here, but afaik the basic idea is to emulate one subsystem for as long as possible until it touches another subsystem (like the CPU reading a register of the PPU), and at that point the other subsystem catches up.

Not sure exactly what you're looking for/what you mean by 'asynchronous/coroutine', but you probably want either libco or libuv (or both?)