Simplify TLS and reduce startup latency

This change simplifies the thread-local storage support code. On Windows
and Mac OS X the startup latency of __enable_tls() has been reduced from
30ms to 1ms. On Windows, TLS memory accesses will now go much faster due
to better self-modifying code that prevents a function call and acquires
our thread information block pointer in a single instruction.

libc/thread/self.c

@@ -26,5 +26,5 @@ STATIC_YOINK("_main_thread_ctor");
  * Returns thread descriptor of the current thread.
  */
 cthread_t(cthread_self)(void) {
-  return (cthread_t)__get_tls_inline();
+  return (cthread_t)__get_tls();
 }