NAME

ck_ring_enqueue_spmc — enqueue pointer into bounded FIFO

LIBRARY

Concurrency Kit (libck, -lck)

SYNOPSIS

#include <ck_ring.h>

bool
ck_ring_enqueue_spmc(ck_ring_t *ring, ck_ring_buffer_t *buffer, void *entry);

DESCRIPTION

The ck_ring_enqueue_spmc(3) function enqueues the pointer entry into the bounded buffer pointed to by ring in FIFO fashion. The buffer pointed to by buffer must be unique to ring and point to an array of ck_ring_buffer_t of sufficient length (according to the power-of-2 elements in the buffer). The decoupling of the ring from the buffer serves to address use-cases involving multiple address spaces and DMA, among others. If you are on non-POSIX platforms or wish for strict compliance with C, then it is recommended to pass a pointer of type void ** for entry. This function is safe to call without locking for UINT_MAX concurrent invocations of ck_ring_dequeue_spmc(3) or ck_ring_trydequeue_spmc(3). This function provides wait-free progress guarantees for one active invocation.

EXAMPLE

#include <ck_ring.h>

/* This ring was previously initialized with ck_ring_init. */
ck_ring_t ring;

/* The ring was initialized for 1023 elements. */
ck_ring_buffer_t buffer[1024];

void
enqueue(void)
{
	void *entry = some_object;

	/* Attempt to enqueue pointer to some_object into buffer. */
	if (ck_ring_enqueue_spmc(&ring, &buffer, &entry) == false) {
		/*
		 * The buffer was full and the enqueue operation
		 * has failed.
		 */
		return;
	}

	/* Enqueue operation completed successfully. */
	return;
}

RETURN VALUES

The function returns true if the value of entry was successfully enqueued into ring. The function will return false if the value of entry could not be enqueued which only occurs if ring was full.