[mypyc] Enable partial, unsafe support for free-threading

[JukkaL]

Enable multi-phase init when using a free-threaded (no-GIL) CPython build so we can enable proper multihreading.

Work on mypyc/mypyc#1104. Work on mypyc/mypyc#1038.

The implementation is still quite incomplete. We are missing synchronization in various places, so race conditions can cause segfaults. Only single-module compilation units are supported for now.

Here's a toy benchmark I used to check that free threading works and can improve performance:

import sys
import threading
import time

def fib(n: int) -> int:
    if n <= 1:
        return n
    else:
        return fib(n - 1) + fib(n - 2)

NTHREADS = 6
print(f"Using {NTHREADS} threads")
print(f"{sys._is_gil_enabled()=}")

t0 = time.time()

threads = []
for i in range(NTHREADS):
    t = threading.Thread(target=lambda: fib(36))
    t.start()
    threads.append(t)

for t in threads:
    t.join()

print()
print('elapsed time:', time.time() - t0)