diff options
| author | Davidlohr Bueso <dave@stgolabs.net> | 2017-11-27 07:20:59 +0300 |
|---|---|---|
| committer | Arnaldo Carvalho de Melo <acme@redhat.com> | 2017-11-30 20:02:05 +0300 |
| commit | 3b2323c2c1c4acf8961cfcdddcee9889daaa21e3 (patch) | |
| tree | a366e8b5137d89f10885ff7d0e6ade8ed0cbf047 /tools/perf/bench/futex-lock-pi.c | |
| parent | 17a68b835921c4664929376447cae5e3374ce2a7 (diff) | |
| download | linux-3b2323c2c1c4acf8961cfcdddcee9889daaa21e3.tar.xz | |
perf bench futex: Use cpumaps
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.
Suggested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Originally-from: James Yang <james.yang@arm.com>
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Davidlohr Bueso <dbueso@suse.de>
Cc: Kim Phillips <kim.phillips@arm.com>
Link: http://lkml.kernel.org/r/20171127042101.3659-2-dave@stgolabs.net
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Diffstat (limited to 'tools/perf/bench/futex-lock-pi.c')
| -rw-r--r-- | tools/perf/bench/futex-lock-pi.c | 23 |
1 files changed, 14 insertions, 9 deletions
diff --git a/tools/perf/bench/futex-lock-pi.c b/tools/perf/bench/futex-lock-pi.c index 08653ae8a8c4..8e9c4753e304 100644 --- a/tools/perf/bench/futex-lock-pi.c +++ b/tools/perf/bench/futex-lock-pi.c @@ -15,6 +15,7 @@ #include <errno.h> #include "bench.h" #include "futex.h" +#include "cpumap.h" #include <err.h> #include <stdlib.h> @@ -32,7 +33,7 @@ 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; @@ -113,9 +114,10 @@ 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; @@ -130,10 +132,10 @@ 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])) @@ -147,19 +149,22 @@ 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) @@ -180,7 +185,7 @@ 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); |