linux/tools/perf/tests/event-times.c
Andi Kleen 4804e01116 perf stat: Use affinity for opening events
Restructure the event opening in perf stat to cycle through the events
by CPU after setting affinity to that CPU.

This eliminates IPI overhead in the perf API.

We have to loop through the CPU in the outter builtin-stat code instead
of leaving that to low level functions.

It has to change the weak group fallback strategy slightly.  Since we
cannot easily undo the opens for other CPUs move the weak group retry to
a separate loop.

Before with a large test case with 94 CPUs:

  % time     seconds  usecs/call     calls    errors syscall
  ------ ----------- ----------- --------- --------- ----------------
   42.75    4.050910          67     60046       110 perf_event_open

After:

   26.86    0.944396          16     58069       110 perf_event_open

(the number changes slightly because the weak group retries
work differently and the test case relies on weak groups)

Committer notes:

Added one of the hunks in a patch provided by Andi after I noticed that
the "event times" 'perf test' entry was segfaulting.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: http://lore.kernel.org/lkml/20191121001522.180827-10-andi@firstfloor.org
Link: http://lore.kernel.org/lkml/20191127232657.GL84886@tassilo.jf.intel.com # Fix
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2019-11-29 12:20:45 -03:00

242 lines
5.4 KiB
C

// SPDX-License-Identifier: GPL-2.0
#include <linux/compiler.h>
#include <linux/string.h>
#include <errno.h>
#include <inttypes.h>
#include <string.h>
#include <sys/wait.h>
#include <perf/cpumap.h>
#include "tests.h"
#include "evlist.h"
#include "evsel.h"
#include "debug.h"
#include "parse-events.h"
#include "thread_map.h"
#include "target.h"
static int attach__enable_on_exec(struct evlist *evlist)
{
struct evsel *evsel = evlist__last(evlist);
struct target target = {
.uid = UINT_MAX,
};
const char *argv[] = { "true", NULL, };
char sbuf[STRERR_BUFSIZE];
int err;
pr_debug("attaching to spawned child, enable on exec\n");
err = perf_evlist__create_maps(evlist, &target);
if (err < 0) {
pr_debug("Not enough memory to create thread/cpu maps\n");
return err;
}
err = perf_evlist__prepare_workload(evlist, &target, argv, false, NULL);
if (err < 0) {
pr_debug("Couldn't run the workload!\n");
return err;
}
evsel->core.attr.enable_on_exec = 1;
err = evlist__open(evlist);
if (err < 0) {
pr_debug("perf_evlist__open: %s\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
return err;
}
return perf_evlist__start_workload(evlist) == 1 ? TEST_OK : TEST_FAIL;
}
static int detach__enable_on_exec(struct evlist *evlist)
{
waitpid(evlist->workload.pid, NULL, 0);
return 0;
}
static int attach__current_disabled(struct evlist *evlist)
{
struct evsel *evsel = evlist__last(evlist);
struct perf_thread_map *threads;
int err;
pr_debug("attaching to current thread as disabled\n");
threads = thread_map__new(-1, getpid(), UINT_MAX);
if (threads == NULL) {
pr_debug("thread_map__new\n");
return -1;
}
evsel->core.attr.disabled = 1;
err = perf_evsel__open_per_thread(evsel, threads);
if (err) {
pr_debug("Failed to open event cpu-clock:u\n");
return err;
}
perf_thread_map__put(threads);
return evsel__enable(evsel) == 0 ? TEST_OK : TEST_FAIL;
}
static int attach__current_enabled(struct evlist *evlist)
{
struct evsel *evsel = evlist__last(evlist);
struct perf_thread_map *threads;
int err;
pr_debug("attaching to current thread as enabled\n");
threads = thread_map__new(-1, getpid(), UINT_MAX);
if (threads == NULL) {
pr_debug("failed to call thread_map__new\n");
return -1;
}
err = perf_evsel__open_per_thread(evsel, threads);
perf_thread_map__put(threads);
return err == 0 ? TEST_OK : TEST_FAIL;
}
static int detach__disable(struct evlist *evlist)
{
struct evsel *evsel = evlist__last(evlist);
return evsel__enable(evsel);
}
static int attach__cpu_disabled(struct evlist *evlist)
{
struct evsel *evsel = evlist__last(evlist);
struct perf_cpu_map *cpus;
int err;
pr_debug("attaching to CPU 0 as enabled\n");
cpus = perf_cpu_map__new("0");
if (cpus == NULL) {
pr_debug("failed to call perf_cpu_map__new\n");
return -1;
}
evsel->core.attr.disabled = 1;
err = perf_evsel__open_per_cpu(evsel, cpus, -1);
if (err) {
if (err == -EACCES)
return TEST_SKIP;
pr_debug("Failed to open event cpu-clock:u\n");
return err;
}
perf_cpu_map__put(cpus);
return evsel__enable(evsel);
}
static int attach__cpu_enabled(struct evlist *evlist)
{
struct evsel *evsel = evlist__last(evlist);
struct perf_cpu_map *cpus;
int err;
pr_debug("attaching to CPU 0 as enabled\n");
cpus = perf_cpu_map__new("0");
if (cpus == NULL) {
pr_debug("failed to call perf_cpu_map__new\n");
return -1;
}
err = perf_evsel__open_per_cpu(evsel, cpus, -1);
if (err == -EACCES)
return TEST_SKIP;
perf_cpu_map__put(cpus);
return err ? TEST_FAIL : TEST_OK;
}
static int test_times(int (attach)(struct evlist *),
int (detach)(struct evlist *))
{
struct perf_counts_values count;
struct evlist *evlist = NULL;
struct evsel *evsel;
int err = -1, i;
evlist = evlist__new();
if (!evlist) {
pr_debug("failed to create event list\n");
goto out_err;
}
err = parse_events(evlist, "cpu-clock:u", NULL);
if (err) {
pr_debug("failed to parse event cpu-clock:u\n");
goto out_err;
}
evsel = evlist__last(evlist);
evsel->core.attr.read_format |=
PERF_FORMAT_TOTAL_TIME_ENABLED |
PERF_FORMAT_TOTAL_TIME_RUNNING;
err = attach(evlist);
if (err == TEST_SKIP) {
pr_debug(" SKIP : not enough rights\n");
return err;
}
TEST_ASSERT_VAL("failed to attach", !err);
for (i = 0; i < 100000000; i++) { }
TEST_ASSERT_VAL("failed to detach", !detach(evlist));
perf_evsel__read(&evsel->core, 0, 0, &count);
err = !(count.ena == count.run);
pr_debug(" %s: ena %" PRIu64", run %" PRIu64"\n",
!err ? "OK " : "FAILED",
count.ena, count.run);
out_err:
evlist__delete(evlist);
return !err ? TEST_OK : TEST_FAIL;
}
/*
* This test creates software event 'cpu-clock'
* attaches it in several ways (explained below)
* and checks that enabled and running times
* match.
*/
int test__event_times(struct test *test __maybe_unused, int subtest __maybe_unused)
{
int err, ret = 0;
#define _T(attach, detach) \
err = test_times(attach, detach); \
if (err && (ret == TEST_OK || ret == TEST_SKIP)) \
ret = err;
/* attach on newly spawned process after exec */
_T(attach__enable_on_exec, detach__enable_on_exec)
/* attach on current process as enabled */
_T(attach__current_enabled, detach__disable)
/* attach on current process as disabled */
_T(attach__current_disabled, detach__disable)
/* attach on cpu as disabled */
_T(attach__cpu_disabled, detach__disable)
/* attach on cpu as enabled */
_T(attach__cpu_enabled, detach__disable)
#undef _T
return ret;
}