[PATCH 1/3] perf bench futex: Use cpumaps

From: Davidlohr Bueso
Date: Sun Nov 26 2017 - 23:25:10 EST

Next message: Ming Lei: "[PATCH V2 0/5] dm-rq: improve sequential I/O performance"
Previous message: Davidlohr Bueso: "[PATCH 3/3] perf bench futex: sync waker threads"
In reply to: Davidlohr Bueso: "[PATCH 3/3] perf bench futex: sync waker threads"
Next in thread: Kim Phillips: "Re: [PATCH -tip 0/3] perf bench futex: Improvements"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

It was reported that the whole futex bench breaks when
dealing with non-contiguously numbered cpus.

$ echo 0 | sudo tee /sys/devices/system/cpu/cpu3/online
$ ./perf bench futex all
perf: pthread_create: Operation not permitted
Run summary [PID 14934]: 7 threads, each ....

James had implemented an approach with cpumaps that use an
in house flavor. Instead of re-inventing the wheel, I've
redone the patch such that we use the perf's util/cpumap.c
interface instead.

Applies to all futex benchmarks.

Cc: Kim Phillips <Kim.Phillips@xxxxxxx>
Originally-from: James Yang <James.Yang@xxxxxxx>
Signed-off-by: Davidlohr Bueso <dbueso@xxxxxxx>
---
tools/perf/bench/futex-hash.c | 19 ++++++++++++-------
tools/perf/bench/futex-lock-pi.c | 23 ++++++++++++++---------
tools/perf/bench/futex-requeue.c | 22 +++++++++++++---------
tools/perf/bench/futex-wake-parallel.c | 24 +++++++++++++++---------
tools/perf/bench/futex-wake.c | 18 +++++++++++-------
5 files changed, 65 insertions(+), 41 deletions(-)

diff --git a/tools/perf/bench/futex-hash.c b/tools/perf/bench/futex-hash.c
index fe16b310097f..230d7d5fc6e4 100644
--- a/tools/perf/bench/futex-hash.c
+++ b/tools/perf/bench/futex-hash.c
@@ -23,6 +23,7 @@
#include <subcmd/parse-options.h>
#include "bench.h"
#include "futex.h"
+#include "cpumap.h"

#include <err.h>
#include <sys/time.h>
@@ -117,11 +118,12 @@ static void print_summary(void)
int bench_futex_hash(int argc, const char **argv)
{
int ret = 0;
- cpu_set_t cpu;
+ cpu_set_t cpuset;
struct sigaction act;
- unsigned int i, ncpus;
+ unsigned int i;
pthread_attr_t thread_attr;
struct worker *worker = NULL;
+ struct cpu_map *cpu;

argc = parse_options(argc, argv, options, bench_futex_hash_usage, 0);
if (argc) {
@@ -129,14 +131,16 @@ int bench_futex_hash(int argc, const char **argv)
exit(EXIT_FAILURE);
}

- ncpus = sysconf(_SC_NPROCESSORS_ONLN);
+ cpu = cpu_map__new(NULL);
+ if (!cpu)
+ goto errmem;

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

if (!nthreads) /* default to the number of CPUs */
- nthreads = ncpus;
+ nthreads = cpu->nr;

worker = calloc(nthreads, sizeof(*worker));
if (!worker)
@@ -162,10 +166,10 @@ int bench_futex_hash(int argc, const char **argv)
if (!worker[i].futex)
goto errmem;

- CPU_ZERO(&cpu);
- CPU_SET(i % ncpus, &cpu);
+ CPU_ZERO(&cpuset);
+ CPU_SET(cpu->map[i % cpu->nr], &cpuset);

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

@@ -216,6 +220,7 @@ int bench_futex_hash(int argc, const char **argv)
print_summary();

free(worker);
+ free(cpu);
return ret;
errmem:
err(EXIT_FAILURE, "calloc");
diff --git a/tools/perf/bench/futex-lock-pi.c b/tools/perf/bench/futex-lock-pi.c
index 73a1c44ea63c..0868444c9e67 100644
--- a/tools/perf/bench/futex-lock-pi.c
+++ b/tools/perf/bench/futex-lock-pi.c
@@ -14,6 +14,7 @@
#include <errno.h>
#include "bench.h"
#include "futex.h"
+#include "cpumap.h"

#include <err.h>
#include <stdlib.h>
@@ -31,7 +32,7 @@ static struct worker *worker;
static unsigned int nsecs = 10;
static bool silent = false, multi = false;
static bool done = false, fshared = false;
-static unsigned int ncpus, nthreads = 0;
+static unsigned int nthreads = 0;
static int futex_flag = 0;
struct timeval start, end, runtime;
static pthread_mutex_t thread_lock;
@@ -112,9 +113,10 @@ static void *workerfn(void *arg)
return NULL;
}

-static void create_threads(struct worker *w, pthread_attr_t thread_attr)
+static void create_threads(struct worker *w, pthread_attr_t thread_attr,
+ struct cpu_map *cpu)
{
- cpu_set_t cpu;
+ cpu_set_t cpuset;
unsigned int i;

threads_starting = nthreads;
@@ -129,10 +131,10 @@ static void create_threads(struct worker *w, pthread_attr_t thread_attr)
} else
worker[i].futex = &global_futex;

- CPU_ZERO(&cpu);
- CPU_SET(i % ncpus, &cpu);
+ CPU_ZERO(&cpuset);
+ CPU_SET(cpu->map[i % cpu->nr], &cpuset);

- if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu))
+ if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");

if (pthread_create(&w[i].thread, &thread_attr, workerfn, &worker[i]))
@@ -146,19 +148,22 @@ int bench_futex_lock_pi(int argc, const char **argv)
unsigned int i;
struct sigaction act;
pthread_attr_t thread_attr;
+ struct cpu_map *cpu;

argc = parse_options(argc, argv, options, bench_futex_lock_pi_usage, 0);
if (argc)
goto err;

- ncpus = sysconf(_SC_NPROCESSORS_ONLN);
+ cpu = cpu_map__new(NULL);
+ if (!cpu)
+ err(EXIT_FAILURE, "calloc");

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

if (!nthreads)
- nthreads = ncpus;
+ nthreads = cpu->nr;

worker = calloc(nthreads, sizeof(*worker));
if (!worker)
@@ -179,7 +184,7 @@ int bench_futex_lock_pi(int argc, const char **argv)
pthread_attr_init(&thread_attr);
gettimeofday(&start, NULL);

- create_threads(worker, thread_attr);
+ create_threads(worker, thread_attr, cpu);
pthread_attr_destroy(&thread_attr);

pthread_mutex_lock(&thread_lock);
diff --git a/tools/perf/bench/futex-requeue.c b/tools/perf/bench/futex-requeue.c
index 41786cbea24c..69f3bbc765d1 100644
--- a/tools/perf/bench/futex-requeue.c
+++ b/tools/perf/bench/futex-requeue.c
@@ -21,6 +21,7 @@
#include <errno.h>
#include "bench.h"
#include "futex.h"
+#include "cpumap.h"

#include <err.h>
#include <stdlib.h>
@@ -39,7 +40,7 @@ static bool done = false, silent = false, fshared = false;
static pthread_mutex_t thread_lock;
static pthread_cond_t thread_parent, thread_worker;
static struct stats requeuetime_stats, requeued_stats;
-static unsigned int ncpus, threads_starting, nthreads = 0;
+static unsigned int threads_starting, nthreads = 0;
static int futex_flag = 0;

static const struct option options[] = {
@@ -82,19 +83,19 @@ static void *workerfn(void *arg __maybe_unused)
}

static void block_threads(pthread_t *w,
- pthread_attr_t thread_attr)
+ pthread_attr_t thread_attr, struct cpu_map *cpu)
{
- cpu_set_t cpu;
+ cpu_set_t cpuset;
unsigned int i;

threads_starting = nthreads;

/* create and block all threads */
for (i = 0; i < nthreads; i++) {
- CPU_ZERO(&cpu);
- CPU_SET(i % ncpus, &cpu);
+ CPU_ZERO(&cpuset);
+ CPU_SET(cpu->map[i % cpu->nr], &cpuset);

- if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu))
+ if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");

if (pthread_create(&w[i], &thread_attr, workerfn, NULL))
@@ -115,19 +116,22 @@ int bench_futex_requeue(int argc, const char **argv)
unsigned int i, j;
struct sigaction act;
pthread_attr_t thread_attr;
+ struct cpu_map *cpu;

argc = parse_options(argc, argv, options, bench_futex_requeue_usage, 0);
if (argc)
goto err;

- ncpus = sysconf(_SC_NPROCESSORS_ONLN);
+ cpu = cpu_map__new(NULL);
+ if (!cpu)
+ err(EXIT_FAILURE, "cpu_map__new");

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

if (!nthreads)
- nthreads = ncpus;
+ nthreads = cpu->nr;

worker = calloc(nthreads, sizeof(*worker));
if (!worker)
@@ -155,7 +159,7 @@ int bench_futex_requeue(int argc, const char **argv)
struct timeval start, end, runtime;

/* create, launch & block all threads */
- block_threads(worker, thread_attr);
+ block_threads(worker, thread_attr, cpu);

/* make sure all threads are already blocked */
pthread_mutex_lock(&thread_lock);
diff --git a/tools/perf/bench/futex-wake-parallel.c b/tools/perf/bench/futex-wake-parallel.c
index 4ab12c8e016a..979e303e4797 100644
--- a/tools/perf/bench/futex-wake-parallel.c
+++ b/tools/perf/bench/futex-wake-parallel.c
@@ -20,6 +20,7 @@
#include <errno.h>
#include "bench.h"
#include "futex.h"
+#include "cpumap.h"

#include <err.h>
#include <stdlib.h>
@@ -42,7 +43,7 @@ static unsigned int nblocked_threads = 0, nwaking_threads = 0;
static pthread_mutex_t thread_lock;
static pthread_cond_t thread_parent, thread_worker;
static struct stats waketime_stats, wakeup_stats;
-static unsigned int ncpus, threads_starting;
+static unsigned int threads_starting;
static int futex_flag = 0;

static const struct option options[] = {
@@ -118,19 +119,20 @@ static void *blocked_workerfn(void *arg __maybe_unused)
return NULL;
}

-static void block_threads(pthread_t *w, pthread_attr_t thread_attr)
+static void block_threads(pthread_t *w, pthread_attr_t thread_attr,
+ struct cpu_map *cpu)
{
- cpu_set_t cpu;
+ cpu_set_t cpuset;
unsigned int i;

threads_starting = nblocked_threads;

/* create and block all threads */
for (i = 0; i < nblocked_threads; i++) {
- CPU_ZERO(&cpu);
- CPU_SET(i % ncpus, &cpu);
+ CPU_ZERO(&cpuset);
+ CPU_SET(cpu->map[i % cpu->nr], &cpuset);

- if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu))
+ if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");

if (pthread_create(&w[i], &thread_attr, blocked_workerfn, NULL))
@@ -204,6 +206,7 @@ int bench_futex_wake_parallel(int argc, const char **argv)
struct sigaction act;
pthread_attr_t thread_attr;
struct thread_data *waking_worker;
+ struct cpu_map *cpu;

argc = parse_options(argc, argv, options,
bench_futex_wake_parallel_usage, 0);
@@ -216,9 +219,12 @@ int bench_futex_wake_parallel(int argc, const char **argv)
act.sa_sigaction = toggle_done;
sigaction(SIGINT, &act, NULL);

- ncpus = sysconf(_SC_NPROCESSORS_ONLN);
+ cpu = cpu_map__new(NULL);
+ if (!cpu)
+ err(EXIT_FAILURE, "calloc");
+
if (!nblocked_threads)
- nblocked_threads = ncpus;
+ nblocked_threads = cpu->nr;

/* some sanity checks */
if (nwaking_threads > nblocked_threads || !nwaking_threads)
@@ -258,7 +264,7 @@ int bench_futex_wake_parallel(int argc, const char **argv)
err(EXIT_FAILURE, "calloc");

/* create, launch & block all threads */
- block_threads(blocked_worker, thread_attr);
+ block_threads(blocked_worker, thread_attr, cpu);

/* make sure all threads are already blocked */
pthread_mutex_lock(&thread_lock);
diff --git a/tools/perf/bench/futex-wake.c b/tools/perf/bench/futex-wake.c
index 2fa49222ef8d..61877b83719a 100644
--- a/tools/perf/bench/futex-wake.c
+++ b/tools/perf/bench/futex-wake.c
@@ -21,6 +21,7 @@
#include <errno.h>
#include "bench.h"
#include "futex.h"
+#include "cpumap.h"

#include <err.h>
#include <stdlib.h>
@@ -88,19 +89,19 @@ static void print_summary(void)
}

static void block_threads(pthread_t *w,
- pthread_attr_t thread_attr)
+ pthread_attr_t thread_attr, struct cpu_map *cpu)
{
- cpu_set_t cpu;
+ cpu_set_t cpuset;
unsigned int i;

threads_starting = nthreads;

/* create and block all threads */
for (i = 0; i < nthreads; i++) {
- CPU_ZERO(&cpu);
- CPU_SET(i % ncpus, &cpu);
+ CPU_ZERO(&cpuset);
+ CPU_SET(cpu->map[i % cpu->nr], &cpuset);

- if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu))
+ if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");

if (pthread_create(&w[i], &thread_attr, workerfn, NULL))
@@ -121,6 +122,7 @@ int bench_futex_wake(int argc, const char **argv)
unsigned int i, j;
struct sigaction act;
pthread_attr_t thread_attr;
+ struct cpu_map *cpu;

argc = parse_options(argc, argv, options, bench_futex_wake_usage, 0);
if (argc) {
@@ -128,7 +130,9 @@ int bench_futex_wake(int argc, const char **argv)
exit(EXIT_FAILURE);
}

- ncpus = sysconf(_SC_NPROCESSORS_ONLN);
+ cpu = cpu_map__new(NULL);
+ if (!cpu)
+ err(EXIT_FAILURE, "calloc");

sigfillset(&act.sa_mask);
act.sa_sigaction = toggle_done;
@@ -160,7 +164,7 @@ int bench_futex_wake(int argc, const char **argv)
struct timeval start, end, runtime;

/* create, launch & block all threads */
- block_threads(worker, thread_attr);
+ block_threads(worker, thread_attr, cpu);

/* make sure all threads are already blocked */
pthread_mutex_lock(&thread_lock);
--
2.13.6

Next message: Ming Lei: "[PATCH V2 0/5] dm-rq: improve sequential I/O performance"
Previous message: Davidlohr Bueso: "[PATCH 3/3] perf bench futex: sync waker threads"
In reply to: Davidlohr Bueso: "[PATCH 3/3] perf bench futex: sync waker threads"
Next in thread: Kim Phillips: "Re: [PATCH -tip 0/3] perf bench futex: Improvements"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]