py/runtime: mp_stack_ctrl_init() should be called immediately on startup.

Calling it from mp_init() is too late for some ports (like Unix), and leads
to incomplete stack frame being captured, with following GC issues. So, now
each port should call mp_stack_ctrl_init() on its own, ASAP after startup,
and taking special precautions so it really was called before stack variables
get allocated (because if such variable with a pointer is missed, it may lead
to over-collecting (typical symptom is segfaulting)).

unix/main.c

@@ -376,7 +376,19 @@ STATIC void set_sys_argv(char *argv[], int argc, int start_arg) {
 #define PATHLIST_SEP_CHAR ':'
 #endif
 
+MP_NOINLINE int main_(int argc, char **argv);
+
 int main(int argc, char **argv) {
+    // We should capture stack top ASAP after start, and it should be
+    // captured guaranteedly before any other stack variables are allocated.
+    // For this, actual main (renamed main_) should not be inlined into
+    // this function. main_() itself may have other functions inlined (with
+    // their own stack variables), that's why we need this main/main_ split.
+    mp_stack_ctrl_init();
+    return main_(argc, argv);
+}
+
+MP_NOINLINE int main_(int argc, char **argv) {
     mp_stack_set_limit(40000 * (BYTES_PER_WORD / 4));
 
     pre_process_options(argc, argv);