mutex: improve compatibility with pthread_mutex_t initializers

pthread_mutex_t has a 32-bit field, __kind, at offset 16.  Non-standard
static initializers set this field to a nonzero value, which can corrupt
fields in our implementation.

Rearrange field layout so we have a hole in that position.  To keep the
structure size small enough so that condvar will still fit in
pthread_condvar_t, we need to change the size of the _depth field to
16 bits.

core/mutex.cc

 #include <osv/mutex.h>
 #include <sched.hh>
 #include "arch.hh"
+#include <pthread.h>
+#include <cassert>
+
+static_assert(sizeof(mutex) <= sizeof(pthread_mutex_t), "mutex too big");
+static_assert(offsetof(mutex, _hole_for_pthread_compatiblity) == 16, "mutex hole in wrong place");
 
 struct waiter {
     struct waiter*	next;

include/osv/mutex.h

@@ -3,6 +3,7 @@
 
 #include <stdbool.h>
 #include <string.h>
+#include <stdint.h>
 
 #ifdef __cplusplus
 extern "C" {
@@ -42,13 +43,14 @@ void spinlock::unlock()
 // Mutex:
 
 typedef struct mutex {
-    spinlock_t _wait_lock;
-    unsigned _depth;
-    void *_owner;
     struct wait_list {
         struct waiter *first;
         struct waiter *last;
     } _wait_list;
+    uint32_t _hole_for_pthread_compatiblity; // pthread_mutex_t's __kind
+    spinlock_t _wait_lock;
+    uint16_t _depth;
+    void *_owner;
 #ifdef __cplusplus
     // additional convenience methods for C++
     inline mutex();