> That's a problem indeed. Perhaps we need a global "fork lock" shared
> between subprocess and multiprocessing?

I did an atfork patch which included a (recursive) fork lock.  See

    http://bugs.python.org/review/6721/show

The patch included changes to multiprocessing and subprocess.  (Being able to acquire the lock when doing fd manipulation is quite useful.  For instance, the creation of Process.sentinel currently has a race which can mean than another process inherits the write end of the pipe.  That would cause Process.join() to wait till both processes terminate.)

Actually, for Finalizers I think it would be easier to just record and check the pid.