1. Yes, you can JIT a callback, but without a GC that just leaves you with lots of callbacks scattered about your memory. I.e. qsort() calls lots of times, then returns and never refers to the object again; libevent functions may cause any number of calls (including zero, if the event is deleted); if you have a function pointer void (put)(int key, void value, void (cb)(void cb_arg), void *cb_arg) to store data, it may either call back before returning (e.g. data is stored in memory) or after (e.g. data is stored on a slow disk), or even call back multiple times (to provide progress reports). I see no way to handle this in a sane fashion.

4. If you are going to need threads by default (e.g. for your GC thread) you are breaking lots of UNIX, notably fork(), and require far too much infrastructure for some kernel/embedded programming. Yes, a good concurrency story is needed, but your program should be able to run without threads.