#include <sys/ksynch.h>
void sema_init(ksema_t *sp, u_int val, char *name, ksema_type_t type, void *arg);
void sema_destroy(ksema_t *sp);
void sema_p(ksema_t *sp);
void sema_v(ksema_t *sp);
int sema_p_sig(ksema_t *sp);
int sema_tryp(ksema_t *sp);
These functions implement counting semaphores as described by Dijkstra. A semaphore has a value which is atomicly decremented by sema_p() and atomicly incremented by sema_v(). The value must always be greater than or equal to zero. If sema_p() is called and the value is zero, the calling thread is blocked until another thread performs a sema_v() operation on the semaphore.
Semaphores are initialized by calling sema_init(). The argument, val, gives the initial value for the semaphore. The semaphore storage is provided by the caller but more may be dynamicly allocated, if necessary, by sema_init(). For this reason, sema_destroy() should be called before deallocating the storage containing the semaphore.
sema_p_sig() decrements the semaphore, as does sema_p(), however, if the semaphore value is zero, sema_p_sig() will return without decrementing the value if a signal (e.g. from kill(2) ) is pending for the thread.
sema_tryp() will decrement the semaphore value only if it is greater than zero, and will not block.
If sema_p() is used from interrupt context, lower-priority interrupts will not be serviced during the wait. This means that if the thread that will eventually perform the sema_v() becomes blocked on anything that requires the lower-priority interrupt, the system will hang.
For example, the thread that will perform the sema_v() may need to first allocate memory. This memory allocation may require waiting for paging I/O to complete, which may require a lower-priority disk or network interrupt to be serviced. In general, situations like this are hard to predict, so it is advisable to avoid waiting on semaphores or condition variables in an interrupt context.