scoped_lock *next;
//! The local spin-wait variable
- /** Inverted (0 - blocked, 1 - acquired the mutex) for the sake of
+ /** Inverted (0 - blocked, 1 - acquired the mutex) for the sake of
zero-initialization. Defining it as an entire word instead of
a byte seems to help performance slightly. */
uintptr_t going;
static const bool is_rw_mutex = false;
static const bool is_recursive_mutex = false;
static const bool is_fair_mutex = true;
-
- friend class scoped_lock;
private:
//! The last competitor requesting the lock
atomic<scoped_lock*> q_tail;