[PATCH v5 0/7] Adding general performance benchmarking subcommand to perf.

[PATCH v5 0/7] Adding general performance benchmarking subcommand to perf.

[Hitoshi Mitake]

This patch series add general benchmark subcommand to perf.

The subcommand will unify the benchmarking programs
and provide these in standardized way.

I fixed this series for latest tip tree.
My previous patch series conflicts with cmd_probe().
So please discard v4 and use this v5.

Hitoshi Mitake (7):
  Adding new directory and header for new subcommand 'bench'
  sched-messaging.c: benchmark for scheduler and IPC mechanisms based
    on hackbench
  sched-pipe.c: benchmark for pipe() system call
  builtin-bench.c: General framework for benchmark suites
  Modifying builtin.h for new prototype
  Modyfing perf.c for subcommand 'bench'
  Modyfing Makefile to build subcommand 'bench'

 tools/perf/Makefile                |    6 +
 tools/perf/bench/bench.h           |    9 +
 tools/perf/bench/sched-messaging.c |  332 ++++++++++++++++++++++++++++++++++++
 tools/perf/bench/sched-pipe.c      |  113 ++++++++++++
 tools/perf/builtin-bench.c         |  128 ++++++++++++++
 tools/perf/builtin.h               |    1 +
 tools/perf/perf.c                  |    1 +
 7 files changed, 590 insertions(+), 0 deletions(-)
 create mode 100644 tools/perf/bench/bench.h
 create mode 100644 tools/perf/bench/sched-messaging.c
 create mode 100644 tools/perf/bench/sched-pipe.c
 create mode 100644 tools/perf/builtin-bench.c

[PATCH v5 1/7] Adding new directory and header for new subcommand 'bench'

[Hitoshi Mitake]

This patch adds bench/ directory and bench/bench.h.
bench/ directory will contain modules for bench subcommand.
bench/bench.h is for listing prototypes of module functions.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
---
 tools/perf/bench/bench.h |    9 +++++++++
 1 files changed, 9 insertions(+), 0 deletions(-)
 create mode 100644 tools/perf/bench/bench.h

diff --git a/tools/perf/bench/bench.h b/tools/perf/bench/bench.h
new file mode 100644
index 0000000..59adb27
--- /dev/null
+++ b/tools/perf/bench/bench.h
@@ -0,0 +1,9 @@
+#ifndef BENCH_H
+#define BENCH_H
+
+extern int bench_sched_messaging(int argc, const char **argv,
+				 const char *prefix);
+extern int bench_sched_pipe(int argc, const char **argv,
+			    const char *prefix);
+
+#endif
-- 
1.6.5.2

[PATCH v5 2/7] sched-messaging.c: benchmark for scheduler and IPC mechanisms based on hackbench

[Hitoshi Mitake]

This patch adds bench/sched-messaging.c.
This benchmark measures performance of scheduler and IPC mechanisms,
and is based on hackbench by Rusty Russell.

Example of usage:
% perf bench sched messaging -g 20 -l 1000 -s
5.432  	  	       	    	    	     # in sec
% perf bench sched messaging                 # run with default options
(20 sender and receiver processes per group)
(10 groups == 400 processes run)

        Total time:0.308 sec
% perf bench sched messaging -t -g 20	     # # be multi-thread, with 20 groups
(20 sender and receiver threads per group)
(20 groups == 800 threads run)

        Total time:0.582 sec

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
---
 tools/perf/bench/sched-messaging.c |  332 ++++++++++++++++++++++++++++++++++++
 1 files changed, 332 insertions(+), 0 deletions(-)
 create mode 100644 tools/perf/bench/sched-messaging.c

diff --git a/tools/perf/bench/sched-messaging.c b/tools/perf/bench/sched-messaging.c
new file mode 100644
index 0000000..36b62c5
--- /dev/null
+++ b/tools/perf/bench/sched-messaging.c
@@ -0,0 +1,332 @@
+/*
+ *
+ * builtin-bench-messaging.c
+ *
+ * messaging: Benchmark for scheduler and IPC mechanisms
+ *
+ * Based on hackbench by Rusty Russell <rusty@rustcorp.com.au>
+ * Ported to perf by Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
+ *
+ */
+
+#include "../perf.h"
+#include "../util/util.h"
+#include "../util/parse-options.h"
+#include "../builtin.h"
+#include "bench.h"
+
+/* Test groups of 20 processes spraying to 20 receivers */
+#include <pthread.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <errno.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <sys/wait.h>
+#include <sys/time.h>
+#include <sys/poll.h>
+#include <limits.h>
+
+#define DATASIZE 100
+
+static int use_pipes = 0;
+static unsigned int loops = 100;
+static unsigned int thread_mode = 0;
+static unsigned int num_groups = 10;
+static int simple = 0;
+
+struct sender_context {
+	unsigned int num_fds;
+	int ready_out;
+	int wakefd;
+	int out_fds[0];
+};
+
+struct receiver_context {
+	unsigned int num_packets;
+	int in_fds[2];
+	int ready_out;
+	int wakefd;
+};
+
+static void barf(const char *msg)
+{
+	fprintf(stderr, "%s (error: %s)\n", msg, strerror(errno));
+	exit(1);
+}
+
+static void fdpair(int fds[2])
+{
+	if (use_pipes) {
+		if (pipe(fds) == 0)
+			return;
+	} else {
+		if (socketpair(AF_UNIX, SOCK_STREAM, 0, fds) == 0)
+			return;
+	}
+
+	barf(use_pipes ? "pipe()" : "socketpair()");
+}
+
+/* Block until we're ready to go */
+static void ready(int ready_out, int wakefd)
+{
+	char dummy;
+	struct pollfd pollfd = { .fd = wakefd, .events = POLLIN };
+
+	/* Tell them we're ready. */
+	if (write(ready_out, &dummy, 1) != 1)
+		barf("CLIENT: ready write");
+
+	/* Wait for "GO" signal */
+	if (poll(&pollfd, 1, -1) != 1)
+		barf("poll");
+}
+
+/* Sender sprays loops messages down each file descriptor */
+static void *sender(struct sender_context *ctx)
+{
+	char data[DATASIZE];
+	unsigned int i, j;
+
+	ready(ctx->ready_out, ctx->wakefd);
+
+	/* Now pump to every receiver. */
+	for (i = 0; i < loops; i++) {
+		for (j = 0; j < ctx->num_fds; j++) {
+			int ret, done = 0;
+
+again:
+			ret = write(ctx->out_fds[j], data + done,
+				    sizeof(data)-done);
+			if (ret < 0)
+				barf("SENDER: write");
+			done += ret;
+			if (done < DATASIZE)
+				goto again;
+		}
+	}
+
+	return NULL;
+}
+
+
+/* One receiver per fd */
+static void *receiver(struct receiver_context* ctx)
+{
+	unsigned int i;
+
+	if (!thread_mode)
+		close(ctx->in_fds[1]);
+
+	/* Wait for start... */
+	ready(ctx->ready_out, ctx->wakefd);
+
+	/* Receive them all */
+	for (i = 0; i < ctx->num_packets; i++) {
+		char data[DATASIZE];
+		int ret, done = 0;
+
+again:
+		ret = read(ctx->in_fds[0], data + done, DATASIZE - done);
+		if (ret < 0)
+			barf("SERVER: read");
+		done += ret;
+		if (done < DATASIZE)
+			goto again;
+	}
+
+	return NULL;
+}
+
+static pthread_t create_worker(void *ctx, void *(*func)(void *))
+{
+	pthread_attr_t attr;
+	pthread_t childid;
+	int err;
+
+	if (!thread_mode) {
+		/* process mode */
+		/* Fork the receiver. */
+		switch (fork()) {
+		case -1:
+			barf("fork()");
+			break;
+		case 0:
+			(*func) (ctx);
+			exit(0);
+			break;
+		default:
+			break;
+		}
+
+		return (pthread_t)0;
+	}
+
+	if (pthread_attr_init(&attr) != 0)
+		barf("pthread_attr_init:");
+
+#ifndef __ia64__
+	if (pthread_attr_setstacksize(&attr, PTHREAD_STACK_MIN) != 0)
+		barf("pthread_attr_setstacksize");
+#endif
+
+	err = pthread_create(&childid, &attr, func, ctx);
+	if (err != 0) {
+		fprintf(stderr, "pthread_create failed: %s (%d)\n",
+			strerror(err), err);
+		exit(-1);
+	}
+	return childid;
+}
+
+static void reap_worker(pthread_t id)
+{
+	int proc_status;
+	void *thread_status;
+
+	if (!thread_mode) {
+		/* process mode */
+		wait(&proc_status);
+		if (!WIFEXITED(proc_status))
+			exit(1);
+	} else {
+		pthread_join(id, &thread_status);
+	}
+}
+
+/* One group of senders and receivers */
+static unsigned int group(pthread_t *pth,
+		unsigned int num_fds,
+		int ready_out,
+		int wakefd)
+{
+	unsigned int i;
+	struct sender_context *snd_ctx = malloc(sizeof(struct sender_context)
+			+ num_fds * sizeof(int));
+
+	if (!snd_ctx)
+		barf("malloc()");
+
+	for (i = 0; i < num_fds; i++) {
+		int fds[2];
+		struct receiver_context *ctx = malloc(sizeof(*ctx));
+
+		if (!ctx)
+			barf("malloc()");
+
+
+		/* Create the pipe between client and server */
+		fdpair(fds);
+
+		ctx->num_packets = num_fds * loops;
+		ctx->in_fds[0] = fds[0];
+		ctx->in_fds[1] = fds[1];
+		ctx->ready_out = ready_out;
+		ctx->wakefd = wakefd;
+
+		pth[i] = create_worker(ctx, (void *)receiver);
+
+		snd_ctx->out_fds[i] = fds[1];
+		if (!thread_mode)
+			close(fds[0]);
+	}
+
+	/* Now we have all the fds, fork the senders */
+	for (i = 0; i < num_fds; i++) {
+		snd_ctx->ready_out = ready_out;
+		snd_ctx->wakefd = wakefd;
+		snd_ctx->num_fds = num_fds;
+
+		pth[num_fds+i] = create_worker(snd_ctx, (void *)sender);
+	}
+
+	/* Close the fds we have left */
+	if (!thread_mode)
+		for (i = 0; i < num_fds; i++)
+			close(snd_ctx->out_fds[i]);
+
+	/* Return number of children to reap */
+	return num_fds * 2;
+}
+
+static const struct option options[] = {
+	OPT_BOOLEAN('p', "pipe", &use_pipes,
+		    "Use pipe() instead of socketpair()"),
+	OPT_BOOLEAN('t', "thread", &thread_mode,
+		    "Be multi thread instead of multi process"),
+	OPT_INTEGER('g', "group", &num_groups,
+		    "Specify number of groups"),
+	OPT_INTEGER('l', "loop", &loops,
+		    "Specify number of loops"),
+	OPT_BOOLEAN('s', "simple-output", &simple,
+		    "Do simple output (this maybe useful for"
+		    "processing by scripts or graph tools like gnuplot)"),
+	OPT_END()
+};
+
+static const char * const bench_sched_message_usage[] = {
+	"perf bench sched messaging <options>",
+	NULL
+};
+
+int bench_sched_messaging(int argc, const char **argv,
+		    const char *prefix __used)
+{
+	unsigned int i, total_children;
+	struct timeval start, stop, diff;
+	unsigned int num_fds = 20;
+	int readyfds[2], wakefds[2];
+	char dummy;
+	pthread_t *pth_tab;
+
+	argc = parse_options(argc, argv, options,
+			     bench_sched_message_usage, 0);
+
+	pth_tab = malloc(num_fds * 2 * num_groups * sizeof(pthread_t));
+	if (!pth_tab)
+		barf("main:malloc()");
+
+	fdpair(readyfds);
+	fdpair(wakefds);
+
+	total_children = 0;
+	for (i = 0; i < num_groups; i++)
+		total_children += group(pth_tab+total_children, num_fds,
+					readyfds[1], wakefds[0]);
+
+	/* Wait for everyone to be ready */
+	for (i = 0; i < total_children; i++)
+		if (read(readyfds[0], &dummy, 1) != 1)
+			barf("Reading for readyfds");
+
+	gettimeofday(&start, NULL);
+
+	/* Kick them off */
+	if (write(wakefds[1], &dummy, 1) != 1)
+		barf("Writing to start them");
+
+	/* Reap them all */
+	for (i = 0; i < total_children; i++)
+		reap_worker(pth_tab[i]);
+
+	gettimeofday(&stop, NULL);
+
+	timersub(&stop, &start, &diff);
+
+	if (simple)
+		printf("%lu.%03lu\n", diff.tv_sec, diff.tv_usec/1000);
+	else {
+		printf("(%d sender and receiver %s per group)\n",
+		       num_fds, thread_mode ? "threads" : "processes");
+		printf("(%d groups == %d %s run)\n\n",
+		       num_groups, num_groups * 2 * num_fds,
+		       thread_mode ? "threads" : "processes");
+		printf("\tTotal time:%lu.%03lu sec\n",
+		       diff.tv_sec, diff.tv_usec/1000);
+	}
+
+	return 0;
+}
-- 
1.6.5.2

[PATCH v5 3/7] sched-pipe.c: benchmark for pipe() system call

[Hitoshi Mitake]

This patch adds bench/sched-pipe.c.

bench/sched-pipe.c is a benchmark program
to measure performance of pipe() system call.
This benchmark is based on pipe-test-1m.c by Ingo Molnar.
http://people.redhat.com/mingo/cfs-scheduler/tools/pipe-test-1m.c

Example of use:
% perf bench sched pipe
(executing 1000000 pipe operations between two tasks)

        Total time:4.499 sec
                4.499179 usecs/op
                222262 ops/sec
% perf bench sched pipe -s -l 1000
0.015

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
---
 tools/perf/bench/sched-pipe.c |  113 +++++++++++++++++++++++++++++++++++++++++
 1 files changed, 113 insertions(+), 0 deletions(-)
 create mode 100644 tools/perf/bench/sched-pipe.c

diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c
new file mode 100644
index 0000000..3214ed2
--- /dev/null
+++ b/tools/perf/bench/sched-pipe.c
@@ -0,0 +1,113 @@
+/*
+ *
+ * builtin-bench-pipe.c
+ *
+ * pipe: Benchmark for pipe()
+ *
+ * Based on pipe-test-1m.c by Ingo Molnar <mingo@redhat.com>
+ *  http://people.redhat.com/mingo/cfs-scheduler/tools/pipe-test-1m.c
+ * Ported to perf by Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
+ *
+ */
+
+#include "../perf.h"
+#include "../util/util.h"
+#include "../util/parse-options.h"
+#include "../builtin.h"
+#include "bench.h"
+
+#include <unistd.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <sys/wait.h>
+#include <linux/unistd.h>
+#include <string.h>
+#include <errno.h>
+#include <assert.h>
+#include <sys/time.h>
+
+#define LOOPS_DEFAULT 1000000
+static int loops = LOOPS_DEFAULT;
+static int simple = 0;
+
+static const struct option options[] = {
+	OPT_INTEGER('l', "loop", &loops,
+		    "Specify number of loops"),
+	OPT_BOOLEAN('s', "simple-output", &simple,
+		    "Do simple output (this maybe useful for"
+		    "processing by scripts or graph tools like gnuplot)"),
+	OPT_END()
+};
+
+static const char * const bench_sched_pipe_usage[] = {
+	"perf bench sched pipe <options>",
+	NULL
+};
+
+int bench_sched_pipe(int argc, const char **argv,
+		     const char *prefix __used)
+{
+	int pipe_1[2], pipe_2[2];
+	int m = 0, i;
+	struct timeval start, stop, diff;
+	unsigned long long result_usec = 0;
+
+	/*
+	 * why does "ret" exist?
+	 * discarding returned value of read(), write()
+	 * causes error in building environment for perf
+	 */
+	int ret;
+	pid_t pid;
+
+	argc = parse_options(argc, argv, options,
+			     bench_sched_pipe_usage, 0);
+
+	assert(!pipe(pipe_1));
+	assert(!pipe(pipe_2));
+
+	pid = fork();
+	assert(pid >= 0);
+
+	gettimeofday(&start, NULL);
+
+	if (!pid) {
+		for (i = 0; i < loops; i++) {
+			ret = read(pipe_1[0], &m, sizeof(int));
+			ret = write(pipe_2[1], &m, sizeof(int));
+		}
+	} else {
+		for (i = 0; i < loops; i++) {
+			ret = write(pipe_1[1], &m, sizeof(int));
+			ret = read(pipe_2[0], &m, sizeof(int));
+		}
+	}
+
+	gettimeofday(&stop, NULL);
+	timersub(&stop, &start, &diff);
+
+	if (pid)
+		return 0;
+
+	if (simple)
+		printf("%lu.%03lu\n",
+		       diff.tv_sec, diff.tv_usec / 1000);
+	else {
+		printf("(executing %d pipe operations between two tasks)\n\n",
+			loops);
+
+		result_usec = diff.tv_sec * 1000000;
+		result_usec += diff.tv_usec;
+
+		printf("\tTotal time:%lu.%03lu sec\n",
+		       diff.tv_sec, diff.tv_usec / 1000);
+		printf("\t\t%lf usecs/op\n",
+		       (double)result_usec / (double)loops);
+		printf("\t\t%d ops/sec\n",
+		       (int)((double)loops /
+			     ((double)result_usec / (double)1000000)));
+	}
+
+	return 0;
+}
-- 
1.6.5.2

[PATCH v5 4/7] builtin-bench.c: General framework for benchmark suites

[Hitoshi Mitake]

This patch adds builtin-bench.c
builtin-bench.c is a general framework for benchmark suites.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
---
 tools/perf/builtin-bench.c |  128 ++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 128 insertions(+), 0 deletions(-)
 create mode 100644 tools/perf/builtin-bench.c

diff --git a/tools/perf/builtin-bench.c b/tools/perf/builtin-bench.c
new file mode 100644
index 0000000..31f4164
--- /dev/null
+++ b/tools/perf/builtin-bench.c
@@ -0,0 +1,128 @@
+/*
+ *
+ * builtin-bench.c
+ *
+ * General benchmarking subsystem provided by perf
+ *
+ * Copyright (C) 2009, Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
+ *
+ */
+
+/*
+ *
+ * Available subsystem list:
+ *  sched ... scheduler and IPC mechanism
+ *
+ */
+
+#include "perf.h"
+#include "util/util.h"
+#include "util/parse-options.h"
+#include "builtin.h"
+#include "bench/bench.h"
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+struct bench_suite {
+	const char *name;
+	const char *summary;
+	int (*fn)(int, const char **, const char *);
+};
+
+static struct bench_suite sched_suites[] = {
+	{ "messaging",
+	  "Benchmark for scheduler and IPC mechanisms",
+	  bench_sched_messaging },
+	{ "pipe",
+	  "Flood of communication over pipe() between two processes",
+	  bench_sched_pipe      },
+	{ NULL,
+	  NULL,
+	  NULL                  }
+};
+
+struct bench_subsys {
+	const char *name;
+	const char *summary;
+	struct bench_suite *suites;
+};
+
+static struct bench_subsys subsystems[] = {
+	{ "sched",
+	  "scheduler and IPC mechanism",
+	  sched_suites },
+	{ NULL,
+	  NULL,
+	  NULL         }
+};
+
+static void dump_suites(int subsys_index)
+{
+	int i;
+
+	printf("List of available suites for %s...\n\n",
+	       subsystems[subsys_index].name);
+
+	for (i = 0; subsystems[subsys_index].suites[i].name; i++)
+		printf("\t%s: %s\n",
+		       subsystems[subsys_index].suites[i].name,
+		       subsystems[subsys_index].suites[i].summary);
+
+	printf("\n");
+	return;
+}
+
+int cmd_bench(int argc, const char **argv, const char *prefix __used)
+{
+	int i, j, status = 0;
+
+	if (argc < 2) {
+		/* No subsystem specified. */
+		printf("Usage: perf bench <subsystem> <suite> [<options>]\n\n");
+		printf("List of available subsystems...\n\n");
+
+		for (i = 0; subsystems[i].name; i++)
+			printf("\t%s: %s\n",
+			       subsystems[i].name, subsystems[i].summary);
+		printf("\n");
+
+		goto end;
+	}
+
+	for (i = 0; subsystems[i].name; i++) {
+		if (strcmp(subsystems[i].name, argv[1]))
+			continue;
+
+		if (argc < 3) {
+			/* No suite specified. */
+			dump_suites(i);
+			goto end;
+		}
+
+		for (j = 0; subsystems[i].suites[j].name; j++) {
+			if (strcmp(subsystems[i].suites[j].name, argv[2]))
+				continue;
+
+			status = subsystems[i].suites[j].fn(argc - 2,
+							    argv + 2, prefix);
+			goto end;
+		}
+
+		if (!strcmp(argv[2], "-h") || !strcmp(argv[2], "--help")) {
+			dump_suites(i);
+			goto end;
+		}
+
+		printf("Unknown suite:%s for %s\n", argv[2], argv[1]);
+		status = 1;
+		goto end;
+	}
+
+	printf("Unknown subsystem:%s\n", argv[1]);
+	status = 1;
+
+end:
+	return status;
+}
-- 
1.6.5.2

[PATCH v5 5/7] Modifying builtin.h for new prototype

[Hitoshi Mitake]

This patch modifies builtin.h to add prototype of cmd_bench().

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
---
 tools/perf/builtin.h |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/tools/perf/builtin.h b/tools/perf/builtin.h
index ad5f0f4..e84f63e 100644
--- a/tools/perf/builtin.h
+++ b/tools/perf/builtin.h
@@ -26,5 +26,6 @@ extern int cmd_top(int argc, const char **argv, const char *prefix);
 extern int cmd_trace(int argc, const char **argv, const char *prefix);
 extern int cmd_version(int argc, const char **argv, const char *prefix);
 extern int cmd_probe(int argc, const char **argv, const char *prefix);
+extern int cmd_bench(int argc, const char **argv, const char *prefix);
 
 #endif
-- 
1.6.5.2

[PATCH v5 6/7] Modyfing perf.c for subcommand 'bench'

[Hitoshi Mitake]

This patch modifies perf.c for invoking 'bench' subcommand.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
---
 tools/perf/perf.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/tools/perf/perf.c b/tools/perf/perf.c
index 9cafe54..5877b01 100644
--- a/tools/perf/perf.c
+++ b/tools/perf/perf.c
@@ -296,6 +296,7 @@ static void handle_internal_command(int argc, const char **argv)
 		{ "trace", cmd_trace, 0 },
 		{ "sched", cmd_sched, 0 },
 		{ "probe", cmd_probe, 0 },
+		{ "bench", cmd_bench, 0 },
 	};
 	unsigned int i;
 	static const char ext[] = STRIP_EXTENSION;
-- 
1.6.5.2

[PATCH v5 7/7] Modyfing Makefile to build subcommand 'bench'

[Hitoshi Mitake]

This patch modifies Makefile for new files related to 'bench' subcommand.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
---
 tools/perf/Makefile |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)

diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index a130134..17f8e9d 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -427,6 +427,12 @@ BUILTIN_OBJS += builtin-timechart.o
 BUILTIN_OBJS += builtin-top.o
 BUILTIN_OBJS += builtin-trace.o
 BUILTIN_OBJS += builtin-probe.o
+BUILTIN_OBJS += builtin-bench.o
+
+# Benchmark modules
+BUILTIN_OBJS += bench/sched-messaging.o
+BUILTIN_OBJS += bench/sched-pipe.o
+
 
 PERFLIBS = $(LIB_FILE)
 
-- 
1.6.5.2

[tip:perf/bench] perf bench: Add subcommand 'bench' to the Makefile

[tip-bot for Hitoshi Mitake]

Commit-ID:  bfde82ef51e3ea6ab8634d0fdbf5adcdd1b429cb
Gitweb:     http://git.kernel.org/tip/bfde82ef51e3ea6ab8634d0fdbf5adcdd1b429cb
Author:     Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
AuthorDate: Thu, 5 Nov 2009 09:31:37 +0900
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Sun, 8 Nov 2009 10:19:20 +0100

perf bench: Add subcommand 'bench' to the Makefile

This patch modifies Makefile for new files related to 'bench'
subcommand. The new code is active from this point on.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: fweisbec@gmail.com
Cc: Jiri Kosina <jkosina@suse.cz>
LKML-Reference: <1257381097-4743-8-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 tools/perf/Makefile |    7 +++++++
 1 files changed, 7 insertions(+), 0 deletions(-)

diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index 542b29e..0a25428 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -416,6 +416,13 @@ LIB_OBJS += util/hist.o
 LIB_OBJS += util/data_map.o
 
 BUILTIN_OBJS += builtin-annotate.o
+
+BUILTIN_OBJS += builtin-bench.o
+
+# Benchmark modules
+BUILTIN_OBJS += bench/sched-messaging.o
+BUILTIN_OBJS += bench/sched-pipe.o
+
 BUILTIN_OBJS += builtin-help.o
 BUILTIN_OBJS += builtin-sched.o
 BUILTIN_OBJS += builtin-list.o

[tip:perf/bench] perf bench: Add new subcommand 'bench' to perf.c

[tip-bot for Hitoshi Mitake]

Commit-ID:  dcba8848d3bc83ec9ee0858b9ae6e4f1c1fa7fa3
Gitweb:     http://git.kernel.org/tip/dcba8848d3bc83ec9ee0858b9ae6e4f1c1fa7fa3
Author:     Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
AuthorDate: Thu, 5 Nov 2009 09:31:36 +0900
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Sun, 8 Nov 2009 10:19:19 +0100

perf bench: Add new subcommand 'bench' to perf.c

This patch modifies perf.c for invoking 'bench' subcommand.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: fweisbec@gmail.com
Cc: Jiri Kosina <jkosina@suse.cz>
LKML-Reference: <1257381097-4743-7-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 tools/perf/perf.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/tools/perf/perf.c b/tools/perf/perf.c
index 624e62d..f90ca5e 100644
--- a/tools/perf/perf.c
+++ b/tools/perf/perf.c
@@ -288,6 +288,7 @@ static void handle_internal_command(int argc, const char **argv)
 		{ "list", cmd_list, 0 },
 		{ "record", cmd_record, 0 },
 		{ "report", cmd_report, 0 },
+		{ "bench", cmd_bench, 0 },
 		{ "stat", cmd_stat, 0 },
 		{ "timechart", cmd_timechart, 0 },
 		{ "top", cmd_top, 0 },

[tip:perf/bench] perf bench: Modify builtin.h for new prototype

[tip-bot for Hitoshi Mitake]

Commit-ID:  11bd341c043348ecb7462d3bd8e1ad6d00f6892a
Gitweb:     http://git.kernel.org/tip/11bd341c043348ecb7462d3bd8e1ad6d00f6892a
Author:     Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
AuthorDate: Thu, 5 Nov 2009 09:31:35 +0900
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Sun, 8 Nov 2009 10:19:19 +0100

perf bench: Modify builtin.h for new prototype

This patch modifies builtin.h to add prototype of cmd_bench().

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: fweisbec@gmail.com
Cc: Jiri Kosina <jkosina@suse.cz>
LKML-Reference: <1257381097-4743-6-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 tools/perf/builtin.h |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/tools/perf/builtin.h b/tools/perf/builtin.h
index e11d8d2..f0cd5b1 100644
--- a/tools/perf/builtin.h
+++ b/tools/perf/builtin.h
@@ -15,6 +15,7 @@ extern int read_line_with_nul(char *buf, int size, FILE *file);
 extern int check_pager_config(const char *cmd);
 
 extern int cmd_annotate(int argc, const char **argv, const char *prefix);
+extern int cmd_bench(int argc, const char **argv, const char *prefix);
 extern int cmd_help(int argc, const char **argv, const char *prefix);
 extern int cmd_sched(int argc, const char **argv, const char *prefix);
 extern int cmd_list(int argc, const char **argv, const char *prefix);

[tip:perf/bench] perf bench: Add builtin-bench.c: General framework for benchmark suites

[tip-bot for Hitoshi Mitake]

Commit-ID:  629cc356653719c206a05f4dee5c5e242edb6546
Gitweb:     http://git.kernel.org/tip/629cc356653719c206a05f4dee5c5e242edb6546
Author:     Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
AuthorDate: Thu, 5 Nov 2009 09:31:34 +0900
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Sun, 8 Nov 2009 10:19:18 +0100

perf bench: Add builtin-bench.c: General framework for benchmark suites

This patch adds builtin-bench.c
builtin-bench.c is a general framework for benchmark suites.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: fweisbec@gmail.com
Cc: Jiri Kosina <jkosina@suse.cz>
LKML-Reference: <1257381097-4743-5-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 tools/perf/builtin-bench.c |  128 ++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 128 insertions(+), 0 deletions(-)

diff --git a/tools/perf/builtin-bench.c b/tools/perf/builtin-bench.c
new file mode 100644
index 0000000..31f4164
--- /dev/null
+++ b/tools/perf/builtin-bench.c
@@ -0,0 +1,128 @@
+/*
+ *
+ * builtin-bench.c
+ *
+ * General benchmarking subsystem provided by perf
+ *
+ * Copyright (C) 2009, Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
+ *
+ */
+
+/*
+ *
+ * Available subsystem list:
+ *  sched ... scheduler and IPC mechanism
+ *
+ */
+
+#include "perf.h"
+#include "util/util.h"
+#include "util/parse-options.h"
+#include "builtin.h"
+#include "bench/bench.h"
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+struct bench_suite {
+	const char *name;
+	const char *summary;
+	int (*fn)(int, const char **, const char *);
+};
+
+static struct bench_suite sched_suites[] = {
+	{ "messaging",
+	  "Benchmark for scheduler and IPC mechanisms",
+	  bench_sched_messaging },
+	{ "pipe",
+	  "Flood of communication over pipe() between two processes",
+	  bench_sched_pipe      },
+	{ NULL,
+	  NULL,
+	  NULL                  }
+};
+
+struct bench_subsys {
+	const char *name;
+	const char *summary;
+	struct bench_suite *suites;
+};
+
+static struct bench_subsys subsystems[] = {
+	{ "sched",
+	  "scheduler and IPC mechanism",
+	  sched_suites },
+	{ NULL,
+	  NULL,
+	  NULL         }
+};
+
+static void dump_suites(int subsys_index)
+{
+	int i;
+
+	printf("List of available suites for %s...\n\n",
+	       subsystems[subsys_index].name);
+
+	for (i = 0; subsystems[subsys_index].suites[i].name; i++)
+		printf("\t%s: %s\n",
+		       subsystems[subsys_index].suites[i].name,
+		       subsystems[subsys_index].suites[i].summary);
+
+	printf("\n");
+	return;
+}
+
+int cmd_bench(int argc, const char **argv, const char *prefix __used)
+{
+	int i, j, status = 0;
+
+	if (argc < 2) {
+		/* No subsystem specified. */
+		printf("Usage: perf bench <subsystem> <suite> [<options>]\n\n");
+		printf("List of available subsystems...\n\n");
+
+		for (i = 0; subsystems[i].name; i++)
+			printf("\t%s: %s\n",
+			       subsystems[i].name, subsystems[i].summary);
+		printf("\n");
+
+		goto end;
+	}
+
+	for (i = 0; subsystems[i].name; i++) {
+		if (strcmp(subsystems[i].name, argv[1]))
+			continue;
+
+		if (argc < 3) {
+			/* No suite specified. */
+			dump_suites(i);
+			goto end;
+		}
+
+		for (j = 0; subsystems[i].suites[j].name; j++) {
+			if (strcmp(subsystems[i].suites[j].name, argv[2]))
+				continue;
+
+			status = subsystems[i].suites[j].fn(argc - 2,
+							    argv + 2, prefix);
+			goto end;
+		}
+
+		if (!strcmp(argv[2], "-h") || !strcmp(argv[2], "--help")) {
+			dump_suites(i);
+			goto end;
+		}
+
+		printf("Unknown suite:%s for %s\n", argv[2], argv[1]);
+		status = 1;
+		goto end;
+	}
+
+	printf("Unknown subsystem:%s\n", argv[1]);
+	status = 1;
+
+end:
+	return status;
+}

[tip:perf/bench] perf bench: Add sched-pipe.c: Benchmark for pipe() system call

[tip-bot for Hitoshi Mitake]

Commit-ID:  c7d9300f367f480aee4663a0e3695c5b48859a1a
Gitweb:     http://git.kernel.org/tip/c7d9300f367f480aee4663a0e3695c5b48859a1a
Author:     Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
AuthorDate: Thu, 5 Nov 2009 09:31:33 +0900
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Sun, 8 Nov 2009 10:19:18 +0100

perf bench: Add sched-pipe.c: Benchmark for pipe() system call

This patch adds bench/sched-pipe.c.

bench/sched-pipe.c is a benchmark program
to measure performance of pipe() system call.
This benchmark is based on pipe-test-1m.c by Ingo Molnar:

   http://people.redhat.com/mingo/cfs-scheduler/tools/pipe-test-1m.c

Example of use:

% perf bench sched pipe
  (executing 1000000 pipe operations between two tasks)

          Total time:4.499 sec
                  4.499179 usecs/op
                  222262 ops/sec

% perf bench sched pipe -s -l 1000
0.015

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: fweisbec@gmail.com
Cc: Jiri Kosina <jkosina@suse.cz>
LKML-Reference: <1257381097-4743-4-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 tools/perf/bench/sched-pipe.c |  113 +++++++++++++++++++++++++++++++++++++++++
 1 files changed, 113 insertions(+), 0 deletions(-)

diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c
new file mode 100644
index 0000000..3214ed2
--- /dev/null
+++ b/tools/perf/bench/sched-pipe.c
@@ -0,0 +1,113 @@
+/*
+ *
+ * builtin-bench-pipe.c
+ *
+ * pipe: Benchmark for pipe()
+ *
+ * Based on pipe-test-1m.c by Ingo Molnar <mingo@redhat.com>
+ *  http://people.redhat.com/mingo/cfs-scheduler/tools/pipe-test-1m.c
+ * Ported to perf by Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
+ *
+ */
+
+#include "../perf.h"
+#include "../util/util.h"
+#include "../util/parse-options.h"
+#include "../builtin.h"
+#include "bench.h"
+
+#include <unistd.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <sys/wait.h>
+#include <linux/unistd.h>
+#include <string.h>
+#include <errno.h>
+#include <assert.h>
+#include <sys/time.h>
+
+#define LOOPS_DEFAULT 1000000
+static int loops = LOOPS_DEFAULT;
+static int simple = 0;
+
+static const struct option options[] = {
+	OPT_INTEGER('l', "loop", &loops,
+		    "Specify number of loops"),
+	OPT_BOOLEAN('s', "simple-output", &simple,
+		    "Do simple output (this maybe useful for"
+		    "processing by scripts or graph tools like gnuplot)"),
+	OPT_END()
+};
+
+static const char * const bench_sched_pipe_usage[] = {
+	"perf bench sched pipe <options>",
+	NULL
+};
+
+int bench_sched_pipe(int argc, const char **argv,
+		     const char *prefix __used)
+{
+	int pipe_1[2], pipe_2[2];
+	int m = 0, i;
+	struct timeval start, stop, diff;
+	unsigned long long result_usec = 0;
+
+	/*
+	 * why does "ret" exist?
+	 * discarding returned value of read(), write()
+	 * causes error in building environment for perf
+	 */
+	int ret;
+	pid_t pid;
+
+	argc = parse_options(argc, argv, options,
+			     bench_sched_pipe_usage, 0);
+
+	assert(!pipe(pipe_1));
+	assert(!pipe(pipe_2));
+
+	pid = fork();
+	assert(pid >= 0);
+
+	gettimeofday(&start, NULL);
+
+	if (!pid) {
+		for (i = 0; i < loops; i++) {
+			ret = read(pipe_1[0], &m, sizeof(int));
+			ret = write(pipe_2[1], &m, sizeof(int));
+		}
+	} else {
+		for (i = 0; i < loops; i++) {
+			ret = write(pipe_1[1], &m, sizeof(int));
+			ret = read(pipe_2[0], &m, sizeof(int));
+		}
+	}
+
+	gettimeofday(&stop, NULL);
+	timersub(&stop, &start, &diff);
+
+	if (pid)
+		return 0;
+
+	if (simple)
+		printf("%lu.%03lu\n",
+		       diff.tv_sec, diff.tv_usec / 1000);
+	else {
+		printf("(executing %d pipe operations between two tasks)\n\n",
+			loops);
+
+		result_usec = diff.tv_sec * 1000000;
+		result_usec += diff.tv_usec;
+
+		printf("\tTotal time:%lu.%03lu sec\n",
+		       diff.tv_sec, diff.tv_usec / 1000);
+		printf("\t\t%lf usecs/op\n",
+		       (double)result_usec / (double)loops);
+		printf("\t\t%d ops/sec\n",
+		       (int)((double)loops /
+			     ((double)result_usec / (double)1000000)));
+	}
+
+	return 0;
+}

[tip:perf/bench] perf bench: Add sched-messaging.c: Benchmark for scheduler and IPC mechanisms based on hackbench

[tip-bot for Hitoshi Mitake]

Commit-ID:  e27454cc6352c4226ddc76f5e3a5dedd7dff456a
Gitweb:     http://git.kernel.org/tip/e27454cc6352c4226ddc76f5e3a5dedd7dff456a
Author:     Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
AuthorDate: Thu, 5 Nov 2009 09:31:32 +0900
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Sun, 8 Nov 2009 10:19:17 +0100

perf bench: Add sched-messaging.c: Benchmark for scheduler and IPC mechanisms based on hackbench

This patch adds bench/sched-messaging.c.

This benchmark measures performance of scheduler and IPC
mechanisms, and is based on hackbench by Rusty Russell.

Example of usage:

  % perf bench sched messaging -g 20 -l 1000 -s
  5.432  	  	       	    	    	     # in sec

  % perf bench sched messaging                 # run with default
  options (20 sender and receiver processes per group)
  (10 groups == 400 processes run)

        Total time:0.308 sec

  % perf bench sched messaging -t -g 20	     # # be multi-thread,
  with 20 groups (20 sender and receiver threads per group)
  (20 groups == 800 threads run)

        Total time:0.582 sec

( Rusty is the original author of hackbench.c and he said the code is
  and was under the GPLv2 so fine to be merged. )

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: fweisbec@gmail.com
Cc: Jiri Kosina <jkosina@suse.cz>
LKML-Reference: <1257381097-4743-3-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 tools/perf/bench/sched-messaging.c |  332 ++++++++++++++++++++++++++++++++++++
 1 files changed, 332 insertions(+), 0 deletions(-)

diff --git a/tools/perf/bench/sched-messaging.c b/tools/perf/bench/sched-messaging.c
new file mode 100644
index 0000000..36b62c5
--- /dev/null
+++ b/tools/perf/bench/sched-messaging.c
@@ -0,0 +1,332 @@
+/*
+ *
+ * builtin-bench-messaging.c
+ *
+ * messaging: Benchmark for scheduler and IPC mechanisms
+ *
+ * Based on hackbench by Rusty Russell <rusty@rustcorp.com.au>
+ * Ported to perf by Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
+ *
+ */
+
+#include "../perf.h"
+#include "../util/util.h"
+#include "../util/parse-options.h"
+#include "../builtin.h"
+#include "bench.h"
+
+/* Test groups of 20 processes spraying to 20 receivers */
+#include <pthread.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <errno.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <sys/wait.h>
+#include <sys/time.h>
+#include <sys/poll.h>
+#include <limits.h>
+
+#define DATASIZE 100
+
+static int use_pipes = 0;
+static unsigned int loops = 100;
+static unsigned int thread_mode = 0;
+static unsigned int num_groups = 10;
+static int simple = 0;
+
+struct sender_context {
+	unsigned int num_fds;
+	int ready_out;
+	int wakefd;
+	int out_fds[0];
+};
+
+struct receiver_context {
+	unsigned int num_packets;
+	int in_fds[2];
+	int ready_out;
+	int wakefd;
+};
+
+static void barf(const char *msg)
+{
+	fprintf(stderr, "%s (error: %s)\n", msg, strerror(errno));
+	exit(1);
+}
+
+static void fdpair(int fds[2])
+{
+	if (use_pipes) {
+		if (pipe(fds) == 0)
+			return;
+	} else {
+		if (socketpair(AF_UNIX, SOCK_STREAM, 0, fds) == 0)
+			return;
+	}
+
+	barf(use_pipes ? "pipe()" : "socketpair()");
+}
+
+/* Block until we're ready to go */
+static void ready(int ready_out, int wakefd)
+{
+	char dummy;
+	struct pollfd pollfd = { .fd = wakefd, .events = POLLIN };
+
+	/* Tell them we're ready. */
+	if (write(ready_out, &dummy, 1) != 1)
+		barf("CLIENT: ready write");
+
+	/* Wait for "GO" signal */
+	if (poll(&pollfd, 1, -1) != 1)
+		barf("poll");
+}
+
+/* Sender sprays loops messages down each file descriptor */
+static void *sender(struct sender_context *ctx)
+{
+	char data[DATASIZE];
+	unsigned int i, j;
+
+	ready(ctx->ready_out, ctx->wakefd);
+
+	/* Now pump to every receiver. */
+	for (i = 0; i < loops; i++) {
+		for (j = 0; j < ctx->num_fds; j++) {
+			int ret, done = 0;
+
+again:
+			ret = write(ctx->out_fds[j], data + done,
+				    sizeof(data)-done);
+			if (ret < 0)
+				barf("SENDER: write");
+			done += ret;
+			if (done < DATASIZE)
+				goto again;
+		}
+	}
+
+	return NULL;
+}
+
+
+/* One receiver per fd */
+static void *receiver(struct receiver_context* ctx)
+{
+	unsigned int i;
+
+	if (!thread_mode)
+		close(ctx->in_fds[1]);
+
+	/* Wait for start... */
+	ready(ctx->ready_out, ctx->wakefd);
+
+	/* Receive them all */
+	for (i = 0; i < ctx->num_packets; i++) {
+		char data[DATASIZE];
+		int ret, done = 0;
+
+again:
+		ret = read(ctx->in_fds[0], data + done, DATASIZE - done);
+		if (ret < 0)
+			barf("SERVER: read");
+		done += ret;
+		if (done < DATASIZE)
+			goto again;
+	}
+
+	return NULL;
+}
+
+static pthread_t create_worker(void *ctx, void *(*func)(void *))
+{
+	pthread_attr_t attr;
+	pthread_t childid;
+	int err;
+
+	if (!thread_mode) {
+		/* process mode */
+		/* Fork the receiver. */
+		switch (fork()) {
+		case -1:
+			barf("fork()");
+			break;
+		case 0:
+			(*func) (ctx);
+			exit(0);
+			break;
+		default:
+			break;
+		}
+
+		return (pthread_t)0;
+	}
+
+	if (pthread_attr_init(&attr) != 0)
+		barf("pthread_attr_init:");
+
+#ifndef __ia64__
+	if (pthread_attr_setstacksize(&attr, PTHREAD_STACK_MIN) != 0)
+		barf("pthread_attr_setstacksize");
+#endif
+
+	err = pthread_create(&childid, &attr, func, ctx);
+	if (err != 0) {
+		fprintf(stderr, "pthread_create failed: %s (%d)\n",
+			strerror(err), err);
+		exit(-1);
+	}
+	return childid;
+}
+
+static void reap_worker(pthread_t id)
+{
+	int proc_status;
+	void *thread_status;
+
+	if (!thread_mode) {
+		/* process mode */
+		wait(&proc_status);
+		if (!WIFEXITED(proc_status))
+			exit(1);
+	} else {
+		pthread_join(id, &thread_status);
+	}
+}
+
+/* One group of senders and receivers */
+static unsigned int group(pthread_t *pth,
+		unsigned int num_fds,
+		int ready_out,
+		int wakefd)
+{
+	unsigned int i;
+	struct sender_context *snd_ctx = malloc(sizeof(struct sender_context)
+			+ num_fds * sizeof(int));
+
+	if (!snd_ctx)
+		barf("malloc()");
+
+	for (i = 0; i < num_fds; i++) {
+		int fds[2];
+		struct receiver_context *ctx = malloc(sizeof(*ctx));
+
+		if (!ctx)
+			barf("malloc()");
+
+
+		/* Create the pipe between client and server */
+		fdpair(fds);
+
+		ctx->num_packets = num_fds * loops;
+		ctx->in_fds[0] = fds[0];
+		ctx->in_fds[1] = fds[1];
+		ctx->ready_out = ready_out;
+		ctx->wakefd = wakefd;
+
+		pth[i] = create_worker(ctx, (void *)receiver);
+
+		snd_ctx->out_fds[i] = fds[1];
+		if (!thread_mode)
+			close(fds[0]);
+	}
+
+	/* Now we have all the fds, fork the senders */
+	for (i = 0; i < num_fds; i++) {
+		snd_ctx->ready_out = ready_out;
+		snd_ctx->wakefd = wakefd;
+		snd_ctx->num_fds = num_fds;
+
+		pth[num_fds+i] = create_worker(snd_ctx, (void *)sender);
+	}
+
+	/* Close the fds we have left */
+	if (!thread_mode)
+		for (i = 0; i < num_fds; i++)
+			close(snd_ctx->out_fds[i]);
+
+	/* Return number of children to reap */
+	return num_fds * 2;
+}
+
+static const struct option options[] = {
+	OPT_BOOLEAN('p', "pipe", &use_pipes,
+		    "Use pipe() instead of socketpair()"),
+	OPT_BOOLEAN('t', "thread", &thread_mode,
+		    "Be multi thread instead of multi process"),
+	OPT_INTEGER('g', "group", &num_groups,
+		    "Specify number of groups"),
+	OPT_INTEGER('l', "loop", &loops,
+		    "Specify number of loops"),
+	OPT_BOOLEAN('s', "simple-output", &simple,
+		    "Do simple output (this maybe useful for"
+		    "processing by scripts or graph tools like gnuplot)"),
+	OPT_END()
+};
+
+static const char * const bench_sched_message_usage[] = {
+	"perf bench sched messaging <options>",
+	NULL
+};
+
+int bench_sched_messaging(int argc, const char **argv,
+		    const char *prefix __used)
+{
+	unsigned int i, total_children;
+	struct timeval start, stop, diff;
+	unsigned int num_fds = 20;
+	int readyfds[2], wakefds[2];
+	char dummy;
+	pthread_t *pth_tab;
+
+	argc = parse_options(argc, argv, options,
+			     bench_sched_message_usage, 0);
+
+	pth_tab = malloc(num_fds * 2 * num_groups * sizeof(pthread_t));
+	if (!pth_tab)
+		barf("main:malloc()");
+
+	fdpair(readyfds);
+	fdpair(wakefds);
+
+	total_children = 0;
+	for (i = 0; i < num_groups; i++)
+		total_children += group(pth_tab+total_children, num_fds,
+					readyfds[1], wakefds[0]);
+
+	/* Wait for everyone to be ready */
+	for (i = 0; i < total_children; i++)
+		if (read(readyfds[0], &dummy, 1) != 1)
+			barf("Reading for readyfds");
+
+	gettimeofday(&start, NULL);
+
+	/* Kick them off */
+	if (write(wakefds[1], &dummy, 1) != 1)
+		barf("Writing to start them");
+
+	/* Reap them all */
+	for (i = 0; i < total_children; i++)
+		reap_worker(pth_tab[i]);
+
+	gettimeofday(&stop, NULL);
+
+	timersub(&stop, &start, &diff);
+
+	if (simple)
+		printf("%lu.%03lu\n", diff.tv_sec, diff.tv_usec/1000);
+	else {
+		printf("(%d sender and receiver %s per group)\n",
+		       num_fds, thread_mode ? "threads" : "processes");
+		printf("(%d groups == %d %s run)\n\n",
+		       num_groups, num_groups * 2 * num_fds,
+		       thread_mode ? "threads" : "processes");
+		printf("\tTotal time:%lu.%03lu sec\n",
+		       diff.tv_sec, diff.tv_usec/1000);
+	}
+
+	return 0;
+}

[tip:perf/bench] perf bench: Add new directory and header for new subcommand 'bench'

[tip-bot for Hitoshi Mitake]

Commit-ID:  c426bba069e65ea438880a04aa4e7c5b880e1728
Gitweb:     http://git.kernel.org/tip/c426bba069e65ea438880a04aa4e7c5b880e1728
Author:     Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
AuthorDate: Thu, 5 Nov 2009 09:31:31 +0900
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Sun, 8 Nov 2009 10:19:15 +0100

perf bench: Add new directory and header for new subcommand 'bench'

This patch adds bench/ directory and bench/bench.h.

bench/ directory will contain modules for bench subcommand.
bench/bench.h is for listing prototypes of module functions.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: fweisbec@gmail.com
Cc: Jiri Kosina <jkosina@suse.cz>
LKML-Reference: <1257381097-4743-2-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 tools/perf/bench/bench.h |    9 +++++++++
 1 files changed, 9 insertions(+), 0 deletions(-)

diff --git a/tools/perf/bench/bench.h b/tools/perf/bench/bench.h
new file mode 100644
index 0000000..59adb27
--- /dev/null
+++ b/tools/perf/bench/bench.h
@@ -0,0 +1,9 @@
+#ifndef BENCH_H
+#define BENCH_H
+
+extern int bench_sched_messaging(int argc, const char **argv,
+				 const char *prefix);
+extern int bench_sched_pipe(int argc, const char **argv,
+			    const char *prefix);
+
+#endif

Re: [PATCH v5 0/7] Adding general performance benchmarking subcommand to perf.

[Ingo Molnar]

* Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> wrote:

> This patch series add general benchmark subcommand to perf.
> 
> The subcommand will unify the benchmarking programs
> and provide these in standardized way.
> 
> I fixed this series for latest tip tree.
> My previous patch series conflicts with cmd_probe().
> So please discard v4 and use this v5.
> 
> Hitoshi Mitake (7):
>   Adding new directory and header for new subcommand 'bench'
>   sched-messaging.c: benchmark for scheduler and IPC mechanisms based
>     on hackbench
>   sched-pipe.c: benchmark for pipe() system call
>   builtin-bench.c: General framework for benchmark suites
>   Modifying builtin.h for new prototype
>   Modyfing perf.c for subcommand 'bench'
>   Modyfing Makefile to build subcommand 'bench'
> 
>  tools/perf/Makefile                |    6 +
>  tools/perf/bench/bench.h           |    9 +
>  tools/perf/bench/sched-messaging.c |  332 ++++++++++++++++++++++++++++++++++++
>  tools/perf/bench/sched-pipe.c      |  113 ++++++++++++
>  tools/perf/builtin-bench.c         |  128 ++++++++++++++
>  tools/perf/builtin.h               |    1 +
>  tools/perf/perf.c                  |    1 +
>  7 files changed, 590 insertions(+), 0 deletions(-)
>  create mode 100644 tools/perf/bench/bench.h
>  create mode 100644 tools/perf/bench/sched-messaging.c
>  create mode 100644 tools/perf/bench/sched-pipe.c
>  create mode 100644 tools/perf/builtin-bench.c

Looks good - i've applied the patches to tip:perf/bench, thanks!

There's one small bug i noticed:

$ ./perf bench sched pipe
(executing 1000000 pipe operations between two tasks)

	Total time:4.898 sec
$ 		4.898586 usecs/op
		204140 ops/sec


the shell prompt came back before the usecs/op and ops/sec line was 
printed. Process teardown race, lack of wait() or so?

	Ingo

Re: [PATCH v5 0/7] Adding general performance benchmarking subcommand to perf.

[Hitoshi Mitake]

From: Ingo Molnar <mingo@elte.hu>
Subject: Re: [PATCH v5 0/7] Adding general performance benchmarking subcommand to perf.
Date: Sun, 8 Nov 2009 10:21:21 +0100

> 
> * Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> wrote:
> 
> > This patch series add general benchmark subcommand to perf.
> > 
> > The subcommand will unify the benchmarking programs
> > and provide these in standardized way.
> > 
> > I fixed this series for latest tip tree.
> > My previous patch series conflicts with cmd_probe().
> > So please discard v4 and use this v5.
> > 
> > Hitoshi Mitake (7):
> >   Adding new directory and header for new subcommand 'bench'
> >   sched-messaging.c: benchmark for scheduler and IPC mechanisms based
> >     on hackbench
> >   sched-pipe.c: benchmark for pipe() system call
> >   builtin-bench.c: General framework for benchmark suites
> >   Modifying builtin.h for new prototype
> >   Modyfing perf.c for subcommand 'bench'
> >   Modyfing Makefile to build subcommand 'bench'
> > 
> >  tools/perf/Makefile                |    6 +
> >  tools/perf/bench/bench.h           |    9 +
> >  tools/perf/bench/sched-messaging.c |  332 ++++++++++++++++++++++++++++++++++++
> >  tools/perf/bench/sched-pipe.c      |  113 ++++++++++++
> >  tools/perf/builtin-bench.c         |  128 ++++++++++++++
> >  tools/perf/builtin.h               |    1 +
> >  tools/perf/perf.c                  |    1 +
> >  7 files changed, 590 insertions(+), 0 deletions(-)
> >  create mode 100644 tools/perf/bench/bench.h
> >  create mode 100644 tools/perf/bench/sched-messaging.c
> >  create mode 100644 tools/perf/bench/sched-pipe.c
> >  create mode 100644 tools/perf/builtin-bench.c
> 
> Looks good - i've applied the patches to tip:perf/bench, thanks!

Thanks! Then I'll work on documentation and merging other
benchmarks such like lmbench or copybench.

> 
> There's one small bug i noticed:
> 
> $ ./perf bench sched pipe
> (executing 1000000 pipe operations between two tasks)
> 
> 	Total time:4.898 sec
> $ 		4.898586 usecs/op
> 		204140 ops/sec
> 
> 
> the shell prompt came back before the usecs/op and ops/sec line was 
> printed. Process teardown race, lack of wait() or so?

Thanks for your notify. I prepared a patch to fix this problem.
I'll send it later.

[PATCH] perf bench:Fix bench/sched-pipe.c to wait child process

[Hitoshi Mitake]

Fixed bench/sched-pipe.c to avoid the below case
notified by Ingo Molnar,

---

$ ./perf bench sched pipe
(executing 1000000 pipe operations between two tasks)

	Total time:4.898 sec
$ 		4.898586 usecs/op
		204140 ops/sec

the shell prompt came back before the usecs/op and ops/sec line was
printed. Process teardown race, lack of wait() or so?

---

This caused by lack of calling waitpid() by parent process,
so I added it.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Kosina <jkosina@suse.cz>
---
 tools/perf/bench/sched-pipe.c |   10 +++++++---
 1 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c
index 3214ed2..6a29100 100644
--- a/tools/perf/bench/sched-pipe.c
+++ b/tools/perf/bench/sched-pipe.c
@@ -26,6 +26,7 @@
 #include <errno.h>
 #include <assert.h>
 #include <sys/time.h>
+#include <sys/types.h>
 
 #define LOOPS_DEFAULT 1000000
 static int loops = LOOPS_DEFAULT;
@@ -58,8 +59,8 @@ int bench_sched_pipe(int argc, const char **argv,
 	 * discarding returned value of read(), write()
 	 * causes error in building environment for perf
 	 */
-	int ret;
-	pid_t pid;
+	int ret, wait_stat;
+	pid_t pid, retpid;
 
 	argc = parse_options(argc, argv, options,
 			     bench_sched_pipe_usage, 0);
@@ -87,8 +88,11 @@ int bench_sched_pipe(int argc, const char **argv,
 	gettimeofday(&stop, NULL);
 	timersub(&stop, &start, &diff);
 
-	if (pid)
+	if (pid) {
+		retpid = waitpid(pid, &wait_stat, 0);
+		assert((retpid == pid) && WIFEXITED(wait_stat));
 		return 0;
+	}
 
 	if (simple)
 		printf("%lu.%03lu\n",
-- 
1.6.5.2

Re: [PATCH] perf bench:Fix bench/sched-pipe.c to wait child process

[Hitoshi Mitake]

From: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Subject: [PATCH] perf bench:Fix bench/sched-pipe.c to wait child process
Date: Mon,  9 Nov 2009 12:13:08 +0900

Sorry, the title lacks space after "perf bench:".
I'll send new fixed one, please discard this.

> Fixed bench/sched-pipe.c to avoid the below case
> notified by Ingo Molnar,
> 
> ---
> 
> $ ./perf bench sched pipe
> (executing 1000000 pipe operations between two tasks)
> 
> 	Total time:4.898 sec
> $ 		4.898586 usecs/op
> 		204140 ops/sec
> 
> the shell prompt came back before the usecs/op and ops/sec line was
> printed. Process teardown race, lack of wait() or so?
> 
> ---
> 
> This caused by lack of calling waitpid() by parent process,
> so I added it.
> 
> Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
> Cc: Rusty Russell <rusty@rustcorp.com.au>
> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
> Cc: Paul Mackerras <paulus@samba.org>
> Cc: Mike Galbraith <efault@gmx.de>
> Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
> Cc: Frederic Weisbecker <fweisbec@gmail.com>
> Cc: Jiri Kosina <jkosina@suse.cz>
> ---
>  tools/perf/bench/sched-pipe.c |   10 +++++++---
>  1 files changed, 7 insertions(+), 3 deletions(-)
> 
> diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c
> index 3214ed2..6a29100 100644
> --- a/tools/perf/bench/sched-pipe.c
> +++ b/tools/perf/bench/sched-pipe.c
> @@ -26,6 +26,7 @@
>  #include <errno.h>
>  #include <assert.h>
>  #include <sys/time.h>
> +#include <sys/types.h>
>  
>  #define LOOPS_DEFAULT 1000000
>  static int loops = LOOPS_DEFAULT;
> @@ -58,8 +59,8 @@ int bench_sched_pipe(int argc, const char **argv,
>  	 * discarding returned value of read(), write()
>  	 * causes error in building environment for perf
>  	 */
> -	int ret;
> -	pid_t pid;
> +	int ret, wait_stat;
> +	pid_t pid, retpid;
>  
>  	argc = parse_options(argc, argv, options,
>  			     bench_sched_pipe_usage, 0);
> @@ -87,8 +88,11 @@ int bench_sched_pipe(int argc, const char **argv,
>  	gettimeofday(&stop, NULL);
>  	timersub(&stop, &start, &diff);
>  
> -	if (pid)
> +	if (pid) {
> +		retpid = waitpid(pid, &wait_stat, 0);
> +		assert((retpid == pid) && WIFEXITED(wait_stat));
>  		return 0;
> +	}
>  
>  	if (simple)
>  		printf("%lu.%03lu\n",
> -- 
> 1.6.5.2
> 
> 

[PATCH] perf bench: Fix bench/sched-pipe.c to wait child process

[Hitoshi Mitake]

Fixed bench/sched-pipe.c to avoid the below case
notified by Ingo Molnar,

---

$ ./perf bench sched pipe
(executing 1000000 pipe operations between two tasks)

	Total time:4.898 sec
$ 		4.898586 usecs/op
		204140 ops/sec

the shell prompt came back before the usecs/op and ops/sec line was
printed. Process teardown race, lack of wait() or so?

---

This caused by lack of calling waitpid() by parent process,
so I added it.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Kosina <jkosina@suse.cz>
---
 tools/perf/bench/sched-pipe.c |   10 +++++++---
 1 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c
index 3214ed2..6a29100 100644
--- a/tools/perf/bench/sched-pipe.c
+++ b/tools/perf/bench/sched-pipe.c
@@ -26,6 +26,7 @@
 #include <errno.h>
 #include <assert.h>
 #include <sys/time.h>
+#include <sys/types.h>
 
 #define LOOPS_DEFAULT 1000000
 static int loops = LOOPS_DEFAULT;
@@ -58,8 +59,8 @@ int bench_sched_pipe(int argc, const char **argv,
 	 * discarding returned value of read(), write()
 	 * causes error in building environment for perf
 	 */
-	int ret;
-	pid_t pid;
+	int ret, wait_stat;
+	pid_t pid, retpid;
 
 	argc = parse_options(argc, argv, options,
 			     bench_sched_pipe_usage, 0);
@@ -87,8 +88,11 @@ int bench_sched_pipe(int argc, const char **argv,
 	gettimeofday(&stop, NULL);
 	timersub(&stop, &start, &diff);
 
-	if (pid)
+	if (pid) {
+		retpid = waitpid(pid, &wait_stat, 0);
+		assert((retpid == pid) && WIFEXITED(wait_stat));
 		return 0;
+	}
 
 	if (simple)
 		printf("%lu.%03lu\n",
-- 
1.6.5.2

[tip:perf/bench] perf bench: Fix bench/sched-pipe.c to wait for child process

[tip-bot for Hitoshi Mitake]

Commit-ID:  5ff0cfc67f00fe0feaa1da0b2359232ea4aa0ee7
Gitweb:     http://git.kernel.org/tip/5ff0cfc67f00fe0feaa1da0b2359232ea4aa0ee7
Author:     Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
AuthorDate: Mon, 9 Nov 2009 12:31:05 +0900
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Mon, 9 Nov 2009 08:14:30 +0100

perf bench: Fix bench/sched-pipe.c to wait for child process

Ingo reported this small 'perf bench sched pipe' output problem:

 | $ ./perf bench sched pipe
 | (executing 1000000 pipe operations between two tasks)
 |
 |	Total time:4.898 sec
 | $		4.898586 usecs/op
 |		204140 ops/sec
 |
 | the shell prompt came back before the usecs/op and ops/sec line
 | was printed. Process teardown race, lack of wait() or so?

This caused by lack of calling waitpid() by parent process,
so I added it.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Kosina <jkosina@suse.cz>
LKML-Reference: <1257737465-7546-1-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 tools/perf/bench/sched-pipe.c |   10 +++++++---
 1 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c
index 3214ed2..6a29100 100644
--- a/tools/perf/bench/sched-pipe.c
+++ b/tools/perf/bench/sched-pipe.c
@@ -26,6 +26,7 @@
 #include <errno.h>
 #include <assert.h>
 #include <sys/time.h>
+#include <sys/types.h>
 
 #define LOOPS_DEFAULT 1000000
 static int loops = LOOPS_DEFAULT;
@@ -58,8 +59,8 @@ int bench_sched_pipe(int argc, const char **argv,
 	 * discarding returned value of read(), write()
 	 * causes error in building environment for perf
 	 */
-	int ret;
-	pid_t pid;
+	int ret, wait_stat;
+	pid_t pid, retpid;
 
 	argc = parse_options(argc, argv, options,
 			     bench_sched_pipe_usage, 0);
@@ -87,8 +88,11 @@ int bench_sched_pipe(int argc, const char **argv,
 	gettimeofday(&stop, NULL);
 	timersub(&stop, &start, &diff);
 
-	if (pid)
+	if (pid) {
+		retpid = waitpid(pid, &wait_stat, 0);
+		assert((retpid == pid) && WIFEXITED(wait_stat));
 		return 0;
+	}
 
 	if (simple)
 		printf("%lu.%03lu\n",