Make more threading improvements

- ASAN memory morgue is now lockless
- Make C11 atomics header more portable
- Rewrote pthread keys support to be lockless
- Simplify Python's unicode table unpacking code
- Make crash report write(2) closer to being atomic
- Make it possible to strace/ftrace a single thread
- ASAN now checks nul-terminated strings fast and properly
- Windows fork() now restores TLS memory of calling thread

third_party/dlmalloc/init.inc

@@ -1,5 +1,7 @@
 // clang-format off
 
+STATIC_YOINK("_init_dlmalloc");
+
 /* ---------------------------- setting mparams -------------------------- */
 
 #if LOCK_AT_FORK
@@ -9,13 +11,13 @@ static void dlmalloc_post_fork_child(void)  { INITIAL_LOCK(&(gm)->mutex); }
 #endif /* LOCK_AT_FORK */
 
 /* Initialize mparams */
-static int init_mparams(void) {
+int init_mparams(void) {
 #ifdef NEED_GLOBAL_LOCK_INIT
   if (malloc_global_mutex_status <= 0)
     init_malloc_global_mutex();
 #endif
 
-  ACQUIRE_MALLOC_GLOBAL_LOCK();
+  // ACQUIRE_MALLOC_GLOBAL_LOCK();
   if (mparams.magic == 0) {
     size_t magic;
     size_t psize;
@@ -94,7 +96,7 @@ static int init_mparams(void) {
     }
   }
 
-  RELEASE_MALLOC_GLOBAL_LOCK();
+  // RELEASE_MALLOC_GLOBAL_LOCK();
 
   return 1;
 }