.\"
.\" Copyright 2013 Brendon Scheinman.
.\" All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\"    notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\"    notice, this list of conditions and the following disclaimer in the
.\"    documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\"
.Dd February 24, 2013.
.Dt ck_cohort 3
.Sh NAME
.Nm ck_cohort
.Nd Generalized interface for lock cohorts
.Sh LIBRARY
Concurrency Kit (libck, \-lck)
.Sh SYNOPSIS
.In ck_cohort.h
.Fn CK_COHORT_PROTOTYPE "COHORT_NAME cohort_name" "LOCK_FXN global_lock_method" \
"LOCK_FXN global_unlock_method" "LOCK_FXN local_lock_method" "LOCK_FXN local_unlock_method"
.Fn CK_COHORT_TRYLOCK_PROTOTYPE "COHORT_NAME cohort_name" \
"LOCK_FXN global_lock_method" "LOCK_FXN global_unlock_method" "TRYLOCK_FXN global_trylock_method" \
"LOCK_FXN local_lock_method" "LOCK_FXN local_unlock_method" "TRYLOCK_FXN local_trylock_method"
.Fn CK_COHORT_INSTANCE "COHORT_NAME cohort_name"
.Fn CK_COHORT_INIT "COHORT_NAME cohort_name" "ck_cohort *cohort" \
"void *global_lock" "void *local_lock" "unsigned int pass_limit"
.Fn CK_COHORT_LOCK "COHORT_NAME cohort_name" "ck_cohort *cohort" \
"void *global_context" "void *local_context"
.Fn CK_COHORT_UNLOCK "COHORT_NAME cohort_name" "ck_cohort *cohort" \
"void *global_context" "void *local_context"
.Pp
Where LOCK_FXN refers to a method with the signature
.br
void(void *lock, void *context)
.br
and TRYLOCK_FXN refers to a method with the signature
.br
bool(void *lock, void *context)
.Pp
The
.Fa context
argument in each signature is used to pass along any additional information that
the lock might need for its lock, unlock and trylock methods.  The values for this
argument are provided to each call to
.Xr CK_COHORT_LOCK 3
,
.Xr CK_COHORT_UNLOCK 3
and
.Xr CK_COHORT_TRYLOCK 3
.
.Sh DESCRIPTION
ck_cohort.h provides an interface for defining lock cohorts with
arbitrary lock types.  Cohorts are a mechanism for coordinating
threads on NUMA architectures in order to reduce the frequency
with which a lock is passed between threads on different clusters.
.Pp
Before using a cohort, the user must define a cohort type using
either the
.Fn CK_COHORT_PROTOTYPE
or the
.Fn CK_COHORT_TRYLOCK_PROTOTYPE
macros.  These macros allow the user to specify the lock methods that
they would like the cohort to use.  See the
.Xr CK_COHORT_PROTOTYPE 3
and
.Xr CK_COHORT_TRYLOCK_PROTOTYPE 3
man pages for more details.
.Pp
.Sh EXAMPLE
.Bd -literal -offset indent

#include <stdlib.h>
#include <pthread.h>

#include <ck_pr.h>
#include <ck_cohort.h>
#include <ck_spinlock.h>

/* 
 * Create lock/unlock methods with signatures that match
 * the required signature
 */
static void
ck_spinlock_lock_with_context(ck_spinlock_t *lock, void *context)
{
	(void)context;
	ck_spinlock_lock(lock);
	return;
}

static void
ck_spinlock_unlock_with_context(ck_spinlock_t *lock, void *context)
{
	(void)context;
	ck_spinlock_unlock(lock);
	return;
}

/*
 * define a cohort type named "test_cohort" that will use
 * the above methods for both its global and local locks
 */
CK_COHORT_PROTOTYPE(test_cohort,
	ck_spinlock_lock_with_context, ck_spinlock_unlock_with_context,
	ck_spinlock_lock_with_context, ck_spinlock_unlock_with_context)

static ck_spinlock_t global_lock = CK_SPINLOCK_INITIALIZER;
static unsigned int ready;

static void*
function(void *context)
{
	CK_COHORT_INSTANCE(test_cohort) *cohort = context;

	while (ready == 0);

	while (ready > 0) {
		/*
		 * acquire the cohort lock before performing critical section.
		 * note that we pass NULL for both the global and local context
		 * arguments because neither the lock nor unlock functions
		 * will use them.
		 */
		CK_COHORT_LOCK(test_cohort, cohort, NULL, NULL);

		/* perform critical section */

		/* relinquish cohort lock */
		CK_COHORT_UNLOCK(test_cohort, cohort, NULL, NULL);
	}

	return NULL;
}

int
main(void)
{
	unsigned int nthr = 4;
	unsigned int n_cohorts = 2;
	unsigned int i;

	/* allocate 2 cohorts of the defined type */
	CK_COHORT_INSTANCE(test_cohort) *cohorts =
	    calloc(n_cohorts, sizeof(CK_COHORT_INSTANCE(test_cohort)));

	/* create local locks to use with each cohort */
	ck_spinlock_t *local_locks = 
		calloc(n_cohorts, sizeof(ck_spinlock_t));

	pthread_t *threads =
		calloc(nthr, sizeof(pthread_t));

	/* initialize each of the cohorts before using them */
	for (i = 0 ; i < n_cohorts ; ++i) {
		CK_COHORT_INIT(test_cohort, cohorts + i, &global_lock, local_locks + i,
			CK_COHORT_DEFAULT_LOCAL_PASS_LIMIT);
	}

	/* start each thread and assign cohorts equally */
	for (i = 0 ; i < nthr ; ++i) {
		pthread_create(threads + i, NULL, function, cohorts + (i % n_cohorts));
	}

	ck_pr_store_uint(&ready, 1);
	sleep(10);
	ck_pr_store_uint(&ready, 0);

	for (i = 0 ; i < nthr ; ++i) {
		pthread_join(threads[i], NULL);
	}

	return 0;
}

.Sh SEE ALSO
.Xr CK_COHORT_PROTOTYPE 3 ,
.Xr CK_COHORT_TRYLOCK_PROTOTYPE 3 ,
.Xr CK_COHORT_INSTANCE 3 ,
.Xr CK_COHORT_INITIALIZER 3 ,
.Xr CK_COHORT_INIT 3 ,
.Xr CK_COHORT_LOCK 3 ,
.Xr CK_COHORT_UNLOCK 3 ,
.Xr CK_COHORT_TRYLOCK 3 ,
.Pp
Additional information available at http://concurrencykit.org/