Refactor some code

- Write tests for cthreads
- Fix bugs in pe2.com tool
- Fix ASAN issue with GetDosEnviron()
- Consolidate the cthread header files
- Some code size optimizations for MODE=
- Attempted to squash a tls linker warning
- Attempted to get futexes working on FreeBSD

libc/runtime/pthread.c

@@ -16,29 +16,56 @@
 │ TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR             │
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
+#include "libc/bits/atomic.h"
+#include "libc/calls/calls.h"
 #include "libc/errno.h"
 #include "libc/intrin/lockcmpxchg.h"
+#include "libc/intrin/spinlock.h"
 #include "libc/runtime/runtime.h"
 
 typedef void *pthread_t;
-typedef bool pthread_once_t;
+typedef int pthread_once_t;
 typedef int pthread_mutex_t;
 
 int pthread_once(pthread_once_t *once, void init(void)) {
-  if (_lockcmpxchg(once, 0, 1)) init();
+  int x;
+  unsigned tries;
+  switch ((x = atomic_load(once))) {
+    case 0:
+      if (atomic_compare_exchange_strong(once, &x, 1)) {
+        init();
+        atomic_store(once, 2);
+        break;
+      }
+      // fallthrough
+    case 1:
+      tries = 0;
+      do {
+        if (++tries & 7) {
+          __builtin_ia32_pause();
+        } else {
+          sched_yield();
+        }
+      } while (atomic_load(once) == 1);
+      break;
+    default:
+      break;
+  }
   return 0;
 }
 
 int pthread_mutex_lock(pthread_mutex_t *mutex) {
-  return EINVAL;
+  _spinlock(mutex);
+  return 0;
 }
 
 int pthread_mutex_trylock(pthread_mutex_t *mutex) {
-  return EINVAL;
+  return _trylock(mutex);
 }
 
 int pthread_mutex_unlock(pthread_mutex_t *mutex) {
-  return EPERM;
+  _spunlock(mutex);
+  return 0;
 }
 
 int pthread_cancel(pthread_t thread) {