[deliverable/linux.git] / tools / perf / bench / futex-hash.c

/*
 * Copyright (C) 2013  Davidlohr Bueso <davidlohr@hp.com>
 *
 * futex-hash: Stress the hell out of the Linux kernel futex uaddr hashing.
 *
 * This program is particularly useful for measuring the kernel's futex hash
 * table/function implementation. In order for it to make sense, use with as
 * many threads and futexes as possible.
 */

/* For the CLR_() macros */
#include <pthread.h>

#include <errno.h>
#include <signal.h>
#include <stdlib.h>
#include <linux/compiler.h>
#include <linux/kernel.h>
#include <sys/time.h>

#include "../util/stat.h"
#include <subcmd/parse-options.h>
#include "bench.h"
#include "futex.h"

#include <err.h>
#include <sys/time.h>

static unsigned int nthreads = 0;
static unsigned int nsecs    = 10;
/* amount of futexes per thread */
static unsigned int nfutexes = 1024;
static bool fshared = false, done = false, silent = false;
static int futex_flag = 0;

struct timeval start, end, runtime;
static pthread_mutex_t thread_lock;
static unsigned int threads_starting;
static struct stats throughput_stats;
static pthread_cond_t thread_parent, thread_worker;

struct worker {
	int tid;
	u_int32_t *futex;
	pthread_t thread;
	unsigned long ops;
};

static const struct option options[] = {
	OPT_UINTEGER('t', "threads", &nthreads, "Specify amount of threads"),
	OPT_UINTEGER('r', "runtime", &nsecs,    "Specify runtime (in seconds)"),
	OPT_UINTEGER('f', "futexes", &nfutexes, "Specify amount of futexes per threads"),
	OPT_BOOLEAN( 's', "silent",  &silent,   "Silent mode: do not display data/details"),
	OPT_BOOLEAN( 'S', "shared",  &fshared,  "Use shared futexes instead of private ones"),
	OPT_END()
};

static const char * const bench_futex_hash_usage[] = {
	"perf bench futex hash <options>",
	NULL
};

static void *workerfn(void *arg)
{
	int ret;
	unsigned int i;
	struct worker *w = (struct worker *) arg;

	pthread_mutex_lock(&thread_lock);
	threads_starting--;
	if (!threads_starting)
		pthread_cond_signal(&thread_parent);
	pthread_cond_wait(&thread_worker, &thread_lock);
	pthread_mutex_unlock(&thread_lock);

	do {
		for (i = 0; i < nfutexes; i++, w->ops++) {
			/*
			 * We want the futex calls to fail in order to stress
			 * the hashing of uaddr and not measure other steps,
			 * such as internal waitqueue handling, thus enlarging
			 * the critical region protected by hb->lock.
			 */
			ret = futex_wait(&w->futex[i], 1234, NULL, futex_flag);
			if (!silent &&
			    (!ret || errno != EAGAIN || errno != EWOULDBLOCK))
				warn("Non-expected futex return call");
		}
	}  while (!done);

	return NULL;
}

static void toggle_done(int sig __maybe_unused,
			siginfo_t *info __maybe_unused,
			void *uc __maybe_unused)
{
	/* inform all threads that we're done for the day */
	done = true;
	gettimeofday(&end, NULL);
	timersub(&end, &start, &runtime);
}

static void print_summary(void)
{
	unsigned long avg = avg_stats(&throughput_stats);
	double stddev = stddev_stats(&throughput_stats);

	printf("%sAveraged %ld operations/sec (+- %.2f%%), total secs = %d\n",
	       !silent ? "\n" : "", avg, rel_stddev_stats(stddev, avg),
	       (int) runtime.tv_sec);
}

int bench_futex_hash(int argc, const char **argv,
		     const char *prefix __maybe_unused)
{
	int ret = 0;
	cpu_set_t cpu;
	struct sigaction act;
	unsigned int i, ncpus;
	pthread_attr_t thread_attr;
	struct worker *worker = NULL;

	argc = parse_options(argc, argv, options, bench_futex_hash_usage, 0);
	if (argc) {
		usage_with_options(bench_futex_hash_usage, options);
		exit(EXIT_FAILURE);
	}

	ncpus = sysconf(_SC_NPROCESSORS_ONLN);

	sigfillset(&act.sa_mask);
	act.sa_sigaction = toggle_done;
	sigaction(SIGINT, &act, NULL);

	if (!nthreads) /* default to the number of CPUs */
		nthreads = ncpus;

	worker = calloc(nthreads, sizeof(*worker));
	if (!worker)
		goto errmem;

	if (!fshared)
		futex_flag = FUTEX_PRIVATE_FLAG;

	printf("Run summary [PID %d]: %d threads, each operating on %d [%s] futexes for %d secs.\n\n",
	       getpid(), nthreads, nfutexes, fshared ? "shared":"private", nsecs);

	init_stats(&throughput_stats);
	pthread_mutex_init(&thread_lock, NULL);
	pthread_cond_init(&thread_parent, NULL);
	pthread_cond_init(&thread_worker, NULL);

	threads_starting = nthreads;
	pthread_attr_init(&thread_attr);
	gettimeofday(&start, NULL);
	for (i = 0; i < nthreads; i++) {
		worker[i].tid = i;
		worker[i].futex = calloc(nfutexes, sizeof(*worker[i].futex));
		if (!worker[i].futex)
			goto errmem;

		CPU_ZERO(&cpu);
		CPU_SET(i % ncpus, &cpu);

		ret = pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu);
		if (ret)
			err(EXIT_FAILURE, "pthread_attr_setaffinity_np");

		ret = pthread_create(&worker[i].thread, &thread_attr, workerfn,
				     (void *)(struct worker *) &worker[i]);
		if (ret)
			err(EXIT_FAILURE, "pthread_create");

	}
	pthread_attr_destroy(&thread_attr);

	pthread_mutex_lock(&thread_lock);
	while (threads_starting)
		pthread_cond_wait(&thread_parent, &thread_lock);
	pthread_cond_broadcast(&thread_worker);
	pthread_mutex_unlock(&thread_lock);

	sleep(nsecs);
	toggle_done(0, NULL, NULL);

	for (i = 0; i < nthreads; i++) {
		ret = pthread_join(worker[i].thread, NULL);
		if (ret)
			err(EXIT_FAILURE, "pthread_join");
	}

	/* cleanup & report results */
	pthread_cond_destroy(&thread_parent);
	pthread_cond_destroy(&thread_worker);
	pthread_mutex_destroy(&thread_lock);

	for (i = 0; i < nthreads; i++) {
		unsigned long t = worker[i].ops/runtime.tv_sec;
		update_stats(&throughput_stats, t);
		if (!silent) {
			if (nfutexes == 1)
				printf("[thread %2d] futex: %p [ %ld ops/sec ]\n",
				       worker[i].tid, &worker[i].futex[0], t);
			else
				printf("[thread %2d] futexes: %p ... %p [ %ld ops/sec ]\n",
				       worker[i].tid, &worker[i].futex[0],
				       &worker[i].futex[nfutexes-1], t);
		}

		free(worker[i].futex);
	}

	print_summary();

	free(worker);
	return ret;
errmem:
	err(EXIT_FAILURE, "calloc");
}
Commit	Line	Data
a0439711 DB	1	/*
	2	* Copyright (C) 2013 Davidlohr Bueso <davidlohr@hp.com>
	3	*
	4	* futex-hash: Stress the hell out of the Linux kernel futex uaddr hashing.
	5	*
	6	* This program is particularly useful for measuring the kernel's futex hash
	7	* table/function implementation. In order for it to make sense, use with as
	8	* many threads and futexes as possible.
	9	*/
	10
8a158589 ACM	11	/* For the CLR_() macros */
	12	#include <pthread.h>
	13
9c304f6c ACM	14	#include <errno.h>
	15	#include <signal.h>
	16	#include <stdlib.h>
86695f59	17	#include <linux/compiler.h>
9c304f6c ACM	18	#include <linux/kernel.h>
	19	#include <sys/time.h>
	20
a0439711	21	#include "../util/stat.h"
4b6ab94e	22	#include <subcmd/parse-options.h>
a0439711 DB	23	#include "bench.h"
	24	#include "futex.h"
	25
	26	#include <err.h>
a0439711	27	#include <sys/time.h>
a0439711 DB	28
	29	static unsigned int nthreads = 0;
	30	static unsigned int nsecs = 10;
	31	/* amount of futexes per thread */
	32	static unsigned int nfutexes = 1024;
	33	static bool fshared = false, done = false, silent = false;
86c87e13	34	static int futex_flag = 0;
a0439711 DB	35
	36	struct timeval start, end, runtime;
	37	static pthread_mutex_t thread_lock;
	38	static unsigned int threads_starting;
	39	static struct stats throughput_stats;
	40	static pthread_cond_t thread_parent, thread_worker;
	41
	42	struct worker {
	43	int tid;
	44	u_int32_t *futex;
	45	pthread_t thread;
	46	unsigned long ops;
	47	};
	48
	49	static const struct option options[] = {
	50	OPT_UINTEGER('t', "threads", &nthreads, "Specify amount of threads"),
	51	OPT_UINTEGER('r', "runtime", &nsecs, "Specify runtime (in seconds)"),
	52	OPT_UINTEGER('f', "futexes", &nfutexes, "Specify amount of futexes per threads"),
	53	OPT_BOOLEAN( 's', "silent", &silent, "Silent mode: do not display data/details"),
	54	OPT_BOOLEAN( 'S', "shared", &fshared, "Use shared futexes instead of private ones"),
	55	OPT_END()
	56	};
	57
	58	static const char * const bench_futex_hash_usage[] = {
	59	"perf bench futex hash <options>",
	60	NULL
	61	};
	62
	63	static void workerfn(void arg)
	64	{
	65	int ret;
	66	unsigned int i;
	67	struct worker w = (struct worker ) arg;
	68
	69	pthread_mutex_lock(&thread_lock);
	70	threads_starting--;
	71	if (!threads_starting)
	72	pthread_cond_signal(&thread_parent);
	73	pthread_cond_wait(&thread_worker, &thread_lock);
	74	pthread_mutex_unlock(&thread_lock);
	75
	76	do {
	77	for (i = 0; i < nfutexes; i++, w->ops++) {
	78	/*
	79	* We want the futex calls to fail in order to stress
	80	* the hashing of uaddr and not measure other steps,
	81	* such as internal waitqueue handling, thus enlarging
	82	* the critical region protected by hb->lock.
	83	*/
86c87e13	84	ret = futex_wait(&w->futex[i], 1234, NULL, futex_flag);
a0439711 DB	85	if (!silent &&
	86	(!ret \|\| errno != EAGAIN \|\| errno != EWOULDBLOCK))
	87	warn("Non-expected futex return call");
	88	}
	89	} while (!done);
	90
	91	return NULL;
	92	}
	93
	94	static void toggle_done(int sig __maybe_unused,
	95	siginfo_t *info __maybe_unused,
	96	void *uc __maybe_unused)
	97	{
	98	/* inform all threads that we're done for the day */
	99	done = true;
	100	gettimeofday(&end, NULL);
	101	timersub(&end, &start, &runtime);
	102	}
	103
	104	static void print_summary(void)
	105	{
	106	unsigned long avg = avg_stats(&throughput_stats);
	107	double stddev = stddev_stats(&throughput_stats);
	108
	109	printf("%sAveraged %ld operations/sec (+- %.2f%%), total secs = %d\n",
	110	!silent ? "\n" : "", avg, rel_stddev_stats(stddev, avg),
	111	(int) runtime.tv_sec);
	112	}
	113
	114	int bench_futex_hash(int argc, const char **argv,
	115	const char *prefix __maybe_unused)
	116	{
	117	int ret = 0;
	118	cpu_set_t cpu;
	119	struct sigaction act;
	120	unsigned int i, ncpus;
	121	pthread_attr_t thread_attr;
	122	struct worker *worker = NULL;
	123
	124	argc = parse_options(argc, argv, options, bench_futex_hash_usage, 0);
	125	if (argc) {
	126	usage_with_options(bench_futex_hash_usage, options);
	127	exit(EXIT_FAILURE);
	128	}
	129
	130	ncpus = sysconf(_SC_NPROCESSORS_ONLN);
	131
	132	sigfillset(&act.sa_mask);
	133	act.sa_sigaction = toggle_done;
	134	sigaction(SIGINT, &act, NULL);
	135
	136	if (!nthreads) /* default to the number of CPUs */
	137	nthreads = ncpus;
	138
	139	worker = calloc(nthreads, sizeof(*worker));
	140	if (!worker)
	141	goto errmem;
	142
86c87e13 DB	143	if (!fshared)
	144	futex_flag = FUTEX_PRIVATE_FLAG;
	145
a0439711 DB	146	printf("Run summary [PID %d]: %d threads, each operating on %d [%s] futexes for %d secs.\n\n",
	147	getpid(), nthreads, nfutexes, fshared ? "shared":"private", nsecs);
	148
	149	init_stats(&throughput_stats);
	150	pthread_mutex_init(&thread_lock, NULL);
	151	pthread_cond_init(&thread_parent, NULL);
	152	pthread_cond_init(&thread_worker, NULL);
	153
	154	threads_starting = nthreads;
	155	pthread_attr_init(&thread_attr);
	156	gettimeofday(&start, NULL);
	157	for (i = 0; i < nthreads; i++) {
	158	worker[i].tid = i;
	159	worker[i].futex = calloc(nfutexes, sizeof(*worker[i].futex));
	160	if (!worker[i].futex)
	161	goto errmem;
	162
	163	CPU_ZERO(&cpu);
	164	CPU_SET(i % ncpus, &cpu);
	165
	166	ret = pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu);
	167	if (ret)
	168	err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
	169
	170	ret = pthread_create(&worker[i].thread, &thread_attr, workerfn,
	171	(void )(struct worker ) &worker[i]);
	172	if (ret)
	173	err(EXIT_FAILURE, "pthread_create");
	174
	175	}
	176	pthread_attr_destroy(&thread_attr);
	177
	178	pthread_mutex_lock(&thread_lock);
	179	while (threads_starting)
	180	pthread_cond_wait(&thread_parent, &thread_lock);
	181	pthread_cond_broadcast(&thread_worker);
	182	pthread_mutex_unlock(&thread_lock);
	183
	184	sleep(nsecs);
	185	toggle_done(0, NULL, NULL);
	186
	187	for (i = 0; i < nthreads; i++) {
	188	ret = pthread_join(worker[i].thread, NULL);
	189	if (ret)
	190	err(EXIT_FAILURE, "pthread_join");
	191	}
	192
	193	/* cleanup & report results */
	194	pthread_cond_destroy(&thread_parent);
	195	pthread_cond_destroy(&thread_worker);
	196	pthread_mutex_destroy(&thread_lock);
	197
	198	for (i = 0; i < nthreads; i++) {
	199	unsigned long t = worker[i].ops/runtime.tv_sec;
	200	update_stats(&throughput_stats, t);
	201	if (!silent) {
	202	if (nfutexes == 1)
	203	printf("[thread %2d] futex: %p [ %ld ops/sec ]\n",
	204	worker[i].tid, &worker[i].futex[0], t);
	205	else
	206	printf("[thread %2d] futexes: %p ... %p [ %ld ops/sec ]\n",
	207	worker[i].tid, &worker[i].futex[0],
	208	&worker[i].futex[nfutexes-1], t);
	209	}
210
211	free(worker[i].futex);
212	}
213
214	print_summary();
215
216	free(worker);
217	return ret;
218	errmem:
219	err(EXIT_FAILURE, "calloc");
220	}