support per-cpu-threads for macOS

[hardfist]

really wanted this features in macOS, especially when debugging application using tokio, which spawns tons of threads and hard to profile

[mstange]

I don't know of a way to get information about which core a thread was running on when we interrupted it with thread_suspend.

One idea I've been playing with is to add a recording code path where Instruments is used for recording, via xctrace. The files created by Instruments probably contain this information.