These functions serialize the execution of lightweight processes. They are useful for ensuring that only one lightweight process can execute a critical section of code at any one time (mutual exclusion). LWP mutexes must be initialized to 0 before use.

The _lwp_mutex_lock() function locks the LWP mutex pointed to by mp. If the mutex is already locked, the calling LWP blocks until the mutex becomes available. When _lwp_mutex_lock() returns, the mutex is locked and the calling LWP is the "owner".

The _lwp_mutex_trylock() function attempts to lock the mutex. If the mutex is already locked it returns with an error. If the mutex is unlocked, it is locked and _lwp_mutex_trylock() returns.

The _lwp_mutex_unlock() function unlocks a locked mutex. The mutex must be locked and the calling LWP must be the one that last locked the mutex (the owner). If any other LWPs are waiting for the mutex to become available, one of them is unblocked.

Upon successful completion, 0 is returned. A non-zero value indicates an error.

If any of the following conditions are detected, _lwp_mutex_lock(), _lwp_mutex_trylock(), and _lwp_mutex_unlock() fail and return the corresponding value:

EINVAL

The mp argument points to an invalid LWP mutex.

EFAULT

The mp argument points to an illegal address.

If any of the following conditions occur, _lwp_mutex_trylock() fails and returns the corresponding value:

EBUSY

The mp argument points to a locked mutex.

intro(2), _lwp_cond_wait(2)