* [PATCH] perf: Generate EXIT event only once per task context
@ 2013-03-15 7:27 Namhyung Kim
2013-03-15 8:59 ` Namhyung Kim
2013-03-18 11:07 ` [tip:perf/urgent] perf: Generate EXIT event only once per task context tip-bot for Namhyung Kim
0 siblings, 2 replies; 4+ messages in thread
From: Namhyung Kim @ 2013-03-15 7:27 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar
Cc: Arnaldo Carvalho de Melo, Jiri Olsa, LKML, Namhyung Kim
From: Namhyung Kim <namhyung.kim@lge.com>
perf_event_task_event() iterates pmu list and generate events for each
eligible pmu context. But if task_event has task_ctx like in EXIT
it'll generate events even though the pmu doesn't have an eligible one.
Fix it by moving the code to proper places.
Before this patch:
$ perf record -n true
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.006 MB perf.data (~248 samples) ]
$ perf report -D | tail
Aggregated stats:
TOTAL events: 73
MMAP events: 67
COMM events: 2
EXIT events: 4
cycles stats:
TOTAL events: 73
MMAP events: 67
COMM events: 2
EXIT events: 4
After this patch:
$ perf report -D | tail
Aggregated stats:
TOTAL events: 70
MMAP events: 67
COMM events: 2
EXIT events: 1
cycles stats:
TOTAL events: 70
MMAP events: 67
COMM events: 2
EXIT events: 1
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
kernel/events/core.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 53ac21cc4012..73ea6fcacf52 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -4432,12 +4432,15 @@ static void perf_event_task_event(struct perf_task_event *task_event)
if (ctxn < 0)
goto next;
ctx = rcu_dereference(current->perf_event_ctxp[ctxn]);
+ if (ctx)
+ perf_event_task_ctx(ctx, task_event);
}
- if (ctx)
- perf_event_task_ctx(ctx, task_event);
next:
put_cpu_ptr(pmu->pmu_cpu_context);
}
+ if (task_event->task_ctx)
+ perf_event_task_ctx(task_event->task_ctx, task_event);
+
rcu_read_unlock();
}
--
1.7.11.7
^ permalink raw reply related [flat|nested] 4+ messages in thread* Re: [PATCH] perf: Generate EXIT event only once per task context 2013-03-15 7:27 [PATCH] perf: Generate EXIT event only once per task context Namhyung Kim @ 2013-03-15 8:59 ` Namhyung Kim 2013-03-21 11:47 ` [tip:perf/core] perf test: Add test case for checking number of EXIT events tip-bot for Namhyung Kim 2013-03-18 11:07 ` [tip:perf/urgent] perf: Generate EXIT event only once per task context tip-bot for Namhyung Kim 1 sibling, 1 reply; 4+ messages in thread From: Namhyung Kim @ 2013-03-15 8:59 UTC (permalink / raw) To: Peter Zijlstra Cc: Ingo Molnar, Arnaldo Carvalho de Melo, Jiri Olsa, LKML, Namhyung Kim Hi, I added a test case for this. Please take a look on it too. This patch is based on my previous evlist cleanup patchset: https://lkml.org/lkml/2013/3/15/47 Thanks, Namhyung >From df86ff1984a2ab89331a10f5e1fc93f09e260ed9 Mon Sep 17 00:00:00 2001 From: Namhyung Kim <namhyung.kim@lge.com> Date: Fri, 15 Mar 2013 14:58:11 +0900 Subject: [PATCH] perf test: Add test case for checking number of EXIT events The test__task_exit() runs a simple "/usr/bin/true" workload and then checks whether the number of EXIT event is 1 or not. Cc: Jiri Olsa <jolsa@redhat.com> Signed-off-by: Namhyung Kim <namhyung@kernel.org> --- tools/perf/Makefile | 1 + tools/perf/tests/builtin-test.c | 4 ++ tools/perf/tests/task-exit.c | 123 ++++++++++++++++++++++++++++++++++++++++ tools/perf/tests/tests.h | 1 + 4 files changed, 129 insertions(+) create mode 100644 tools/perf/tests/task-exit.c diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 3dcd6273a90b..0145df853af4 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -511,6 +511,7 @@ LIB_OBJS += $(OUTPUT)tests/evsel-tp-sched.o LIB_OBJS += $(OUTPUT)tests/pmu.o LIB_OBJS += $(OUTPUT)tests/hists_link.o LIB_OBJS += $(OUTPUT)tests/python-use.o +LIB_OBJS += $(OUTPUT)tests/task-exit.o BUILTIN_OBJS += $(OUTPUT)builtin-annotate.o BUILTIN_OBJS += $(OUTPUT)builtin-bench.o diff --git a/tools/perf/tests/builtin-test.c b/tools/perf/tests/builtin-test.c index acb98e0e39f2..dd00e5e444f8 100644 --- a/tools/perf/tests/builtin-test.c +++ b/tools/perf/tests/builtin-test.c @@ -78,6 +78,10 @@ static struct test { .func = test__python_use, }, { + .desc = "Test number of exit event of a simple workload", + .func = test__task_exit, + }, + { .func = NULL, }, }; diff --git a/tools/perf/tests/task-exit.c b/tools/perf/tests/task-exit.c new file mode 100644 index 000000000000..28fe5894b061 --- /dev/null +++ b/tools/perf/tests/task-exit.c @@ -0,0 +1,123 @@ +#include "evlist.h" +#include "evsel.h" +#include "thread_map.h" +#include "cpumap.h" +#include "tests.h" + +#include <signal.h> + +static int exited; +static int nr_exit; + +static void sig_handler(int sig) +{ + exited = 1; + + if (sig == SIGUSR1) + nr_exit = -1; +} + +/* + * This test will start a workload that does nothing then it checks + * if the number of exit event reported by the kernel is 1 or not + * in order to check the kernel returns correct number of event. + */ +int test__task_exit(void) +{ + int err = -1; + union perf_event *event; + struct perf_evsel *evsel; + struct perf_evlist *evlist; + struct perf_target target = { + .uid = UINT_MAX, + .uses_mmap = true, + }; + const char *argv[] = { "true", NULL }; + + signal(SIGCHLD, sig_handler); + signal(SIGUSR1, sig_handler); + + evlist = perf_evlist__new(); + if (evlist == NULL) { + pr_debug("perf_evlist__new\n"); + return -1; + } + /* + * We need at least one evsel in the evlist, use the default + * one: "cycles". + */ + err = perf_evlist__add_default(evlist); + if (err < 0) { + pr_debug("Not enough memory to create evsel\n"); + goto out_free_evlist; + } + + /* + * Create maps of threads and cpus to monitor. In this case + * we start with all threads and cpus (-1, -1) but then in + * perf_evlist__prepare_workload we'll fill in the only thread + * we're monitoring, the one forked there. + */ + evlist->cpus = cpu_map__dummy_new(); + evlist->threads = thread_map__new_by_tid(-1); + if (!evlist->cpus || !evlist->threads) { + err = -ENOMEM; + pr_debug("Not enough memory to create thread/cpu maps\n"); + goto out_delete_maps; + } + + err = perf_evlist__prepare_workload(evlist, &target, argv, false, true); + if (err < 0) { + pr_debug("Couldn't run the workload!\n"); + goto out_delete_maps; + } + + evsel = perf_evlist__first(evlist); + evsel->attr.task = 1; + evsel->attr.sample_freq = 0; + evsel->attr.inherit = 0; + evsel->attr.watermark = 0; + evsel->attr.wakeup_events = 1; + evsel->attr.exclude_kernel = 1; + + err = perf_evlist__open(evlist); + if (err < 0) { + pr_debug("Couldn't open the evlist: %s\n", strerror(-err)); + goto out_delete_maps; + } + + if (perf_evlist__mmap(evlist, 128, true) < 0) { + pr_debug("failed to mmap events: %d (%s)\n", errno, + strerror(errno)); + goto out_close_evlist; + } + + perf_evlist__start_workload(evlist); + +retry: + while ((event = perf_evlist__mmap_read(evlist, 0)) != NULL) { + if (event->header.type != PERF_RECORD_EXIT) + continue; + + nr_exit++; + } + + if (!exited || !nr_exit) { + poll(evlist->pollfd, evlist->nr_fds, -1); + goto retry; + } + + if (nr_exit != 1) { + pr_debug("received %d EXIT records\n", nr_exit); + err = -1; + } + + perf_evlist__munmap(evlist); +out_close_evlist: + perf_evlist__close(evlist); +out_delete_maps: + perf_evlist__delete_maps(evlist); +out_free_evlist: + perf_evlist__delete(evlist); + return err; +} diff --git a/tools/perf/tests/tests.h b/tools/perf/tests/tests.h index 5de0be1ff4b6..4590c10b7e65 100644 --- a/tools/perf/tests/tests.h +++ b/tools/perf/tests/tests.h @@ -23,5 +23,6 @@ int test__dso_data(void); int test__parse_events(void); int test__hists_link(void); int test__python_use(void); +int test__task_exit(void); #endif /* TESTS_H */ -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 4+ messages in thread
* [tip:perf/core] perf test: Add test case for checking number of EXIT events 2013-03-15 8:59 ` Namhyung Kim @ 2013-03-21 11:47 ` tip-bot for Namhyung Kim 0 siblings, 0 replies; 4+ messages in thread From: tip-bot for Namhyung Kim @ 2013-03-21 11:47 UTC (permalink / raw) To: linux-tip-commits Cc: acme, linux-kernel, hpa, mingo, a.p.zijlstra, namhyung.kim, namhyung, jolsa, tglx Commit-ID: d723a55096b81a13c319485f01994e0a539efcf9 Gitweb: http://git.kernel.org/tip/d723a55096b81a13c319485f01994e0a539efcf9 Author: Namhyung Kim <namhyung.kim@lge.com> AuthorDate: Fri, 15 Mar 2013 14:58:11 +0900 Committer: Arnaldo Carvalho de Melo <acme@redhat.com> CommitDate: Fri, 15 Mar 2013 13:06:12 -0300 perf test: Add test case for checking number of EXIT events The new test__task_exit() test runs a simple "/usr/bin/true" workload and then checks whether the number of EXIT event is 1 or not. Signed-off-by: Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/87obeljax4.fsf@sejong.aot.lge.com [ committer note: Fixup conflicts with f4c66b4 ( bp overflow tests ) ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> --- tools/perf/Makefile | 1 + tools/perf/tests/builtin-test.c | 4 ++ tools/perf/tests/task-exit.c | 123 ++++++++++++++++++++++++++++++++++++++++ tools/perf/tests/tests.h | 1 + 4 files changed, 129 insertions(+) diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 990e9a1..8e1bba3 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -513,6 +513,7 @@ LIB_OBJS += $(OUTPUT)tests/hists_link.o LIB_OBJS += $(OUTPUT)tests/python-use.o LIB_OBJS += $(OUTPUT)tests/bp_signal.o LIB_OBJS += $(OUTPUT)tests/bp_signal_overflow.o +LIB_OBJS += $(OUTPUT)tests/task-exit.o BUILTIN_OBJS += $(OUTPUT)builtin-annotate.o BUILTIN_OBJS += $(OUTPUT)builtin-bench.o diff --git a/tools/perf/tests/builtin-test.c b/tools/perf/tests/builtin-test.c index 45d9ad4..9b5c70a 100644 --- a/tools/perf/tests/builtin-test.c +++ b/tools/perf/tests/builtin-test.c @@ -86,6 +86,10 @@ static struct test { .func = test__bp_signal_overflow, }, { + .desc = "Test number of exit event of a simple workload", + .func = test__task_exit, + }, + { .func = NULL, }, }; diff --git a/tools/perf/tests/task-exit.c b/tools/perf/tests/task-exit.c new file mode 100644 index 0000000..28fe589 --- /dev/null +++ b/tools/perf/tests/task-exit.c @@ -0,0 +1,123 @@ +#include "evlist.h" +#include "evsel.h" +#include "thread_map.h" +#include "cpumap.h" +#include "tests.h" + +#include <signal.h> + +static int exited; +static int nr_exit; + +static void sig_handler(int sig) +{ + exited = 1; + + if (sig == SIGUSR1) + nr_exit = -1; +} + +/* + * This test will start a workload that does nothing then it checks + * if the number of exit event reported by the kernel is 1 or not + * in order to check the kernel returns correct number of event. + */ +int test__task_exit(void) +{ + int err = -1; + union perf_event *event; + struct perf_evsel *evsel; + struct perf_evlist *evlist; + struct perf_target target = { + .uid = UINT_MAX, + .uses_mmap = true, + }; + const char *argv[] = { "true", NULL }; + + signal(SIGCHLD, sig_handler); + signal(SIGUSR1, sig_handler); + + evlist = perf_evlist__new(); + if (evlist == NULL) { + pr_debug("perf_evlist__new\n"); + return -1; + } + /* + * We need at least one evsel in the evlist, use the default + * one: "cycles". + */ + err = perf_evlist__add_default(evlist); + if (err < 0) { + pr_debug("Not enough memory to create evsel\n"); + goto out_free_evlist; + } + + /* + * Create maps of threads and cpus to monitor. In this case + * we start with all threads and cpus (-1, -1) but then in + * perf_evlist__prepare_workload we'll fill in the only thread + * we're monitoring, the one forked there. + */ + evlist->cpus = cpu_map__dummy_new(); + evlist->threads = thread_map__new_by_tid(-1); + if (!evlist->cpus || !evlist->threads) { + err = -ENOMEM; + pr_debug("Not enough memory to create thread/cpu maps\n"); + goto out_delete_maps; + } + + err = perf_evlist__prepare_workload(evlist, &target, argv, false, true); + if (err < 0) { + pr_debug("Couldn't run the workload!\n"); + goto out_delete_maps; + } + + evsel = perf_evlist__first(evlist); + evsel->attr.task = 1; + evsel->attr.sample_freq = 0; + evsel->attr.inherit = 0; + evsel->attr.watermark = 0; + evsel->attr.wakeup_events = 1; + evsel->attr.exclude_kernel = 1; + + err = perf_evlist__open(evlist); + if (err < 0) { + pr_debug("Couldn't open the evlist: %s\n", strerror(-err)); + goto out_delete_maps; + } + + if (perf_evlist__mmap(evlist, 128, true) < 0) { + pr_debug("failed to mmap events: %d (%s)\n", errno, + strerror(errno)); + goto out_close_evlist; + } + + perf_evlist__start_workload(evlist); + +retry: + while ((event = perf_evlist__mmap_read(evlist, 0)) != NULL) { + if (event->header.type != PERF_RECORD_EXIT) + continue; + + nr_exit++; + } + + if (!exited || !nr_exit) { + poll(evlist->pollfd, evlist->nr_fds, -1); + goto retry; + } + + if (nr_exit != 1) { + pr_debug("received %d EXIT records\n", nr_exit); + err = -1; + } + + perf_evlist__munmap(evlist); +out_close_evlist: + perf_evlist__close(evlist); +out_delete_maps: + perf_evlist__delete_maps(evlist); +out_free_evlist: + perf_evlist__delete(evlist); + return err; +} diff --git a/tools/perf/tests/tests.h b/tools/perf/tests/tests.h index 6cf1ec4..b33b328 100644 --- a/tools/perf/tests/tests.h +++ b/tools/perf/tests/tests.h @@ -25,5 +25,6 @@ int test__hists_link(void); int test__python_use(void); int test__bp_signal(void); int test__bp_signal_overflow(void); +int test__task_exit(void); #endif /* TESTS_H */ ^ permalink raw reply related [flat|nested] 4+ messages in thread
* [tip:perf/urgent] perf: Generate EXIT event only once per task context 2013-03-15 7:27 [PATCH] perf: Generate EXIT event only once per task context Namhyung Kim 2013-03-15 8:59 ` Namhyung Kim @ 2013-03-18 11:07 ` tip-bot for Namhyung Kim 1 sibling, 0 replies; 4+ messages in thread From: tip-bot for Namhyung Kim @ 2013-03-18 11:07 UTC (permalink / raw) To: linux-tip-commits Cc: linux-kernel, hpa, mingo, a.p.zijlstra, acme, namhyung.kim, namhyung, jolsa, tglx Commit-ID: d610d98b5de6860feb21539726e9af7c9094151c Gitweb: http://git.kernel.org/tip/d610d98b5de6860feb21539726e9af7c9094151c Author: Namhyung Kim <namhyung.kim@lge.com> AuthorDate: Fri, 15 Mar 2013 16:27:13 +0900 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Mon, 18 Mar 2013 09:47:33 +0100 perf: Generate EXIT event only once per task context perf_event_task_event() iterates pmu list and generate events for each eligible pmu context. But if task_event has task_ctx like in EXIT it'll generate events even though the pmu doesn't have an eligible one. Fix it by moving the code to proper places. Before this patch: $ perf record -n true [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.006 MB perf.data (~248 samples) ] $ perf report -D | tail Aggregated stats: TOTAL events: 73 MMAP events: 67 COMM events: 2 EXIT events: 4 cycles stats: TOTAL events: 73 MMAP events: 67 COMM events: 2 EXIT events: 4 After this patch: $ perf report -D | tail Aggregated stats: TOTAL events: 70 MMAP events: 67 COMM events: 2 EXIT events: 1 cycles stats: TOTAL events: 70 MMAP events: 67 COMM events: 2 EXIT events: 1 Signed-off-by: Namhyung Kim <namhyung@kernel.org> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Namhyung Kim <namhyung.kim@lge.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1363332433-7637-1-git-send-email-namhyung@kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org> --- kernel/events/core.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/kernel/events/core.c b/kernel/events/core.c index fa79c37..59412d0 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -4434,12 +4434,15 @@ static void perf_event_task_event(struct perf_task_event *task_event) if (ctxn < 0) goto next; ctx = rcu_dereference(current->perf_event_ctxp[ctxn]); + if (ctx) + perf_event_task_ctx(ctx, task_event); } - if (ctx) - perf_event_task_ctx(ctx, task_event); next: put_cpu_ptr(pmu->pmu_cpu_context); } + if (task_event->task_ctx) + perf_event_task_ctx(task_event->task_ctx, task_event); + rcu_read_unlock(); } ^ permalink raw reply related [flat|nested] 4+ messages in thread
end of thread, other threads:[~2013-03-21 11:49 UTC | newest] Thread overview: 4+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2013-03-15 7:27 [PATCH] perf: Generate EXIT event only once per task context Namhyung Kim 2013-03-15 8:59 ` Namhyung Kim 2013-03-21 11:47 ` [tip:perf/core] perf test: Add test case for checking number of EXIT events tip-bot for Namhyung Kim 2013-03-18 11:07 ` [tip:perf/urgent] perf: Generate EXIT event only once per task context tip-bot for Namhyung Kim
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox