* [patch] perf stat: --repeat forever
@ 2013-02-15 22:28 Frederik Deweerdt
2013-02-18 15:04 ` Arnaldo Carvalho de Melo
2013-03-01 18:02 ` [patch v2] " Frederik Deweerdt
0 siblings, 2 replies; 5+ messages in thread
From: Frederik Deweerdt @ 2013-02-15 22:28 UTC (permalink / raw)
To: a.p.zijlstra, paulus, mingo, acme; +Cc: linux-kernel
Hi,
The following patch causes 'perf stat --repeat 0' to be interpreted as
'forever', displaying the stats for every run.
We act as if a single run was asked, and reset the stats in each
iteration. In this mode SIGINT is passed to perf to be able to stop the
loop with Ctrl+C.
Regards,
Frederik
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@gmail.com>
diff --git a/tools/perf/Documentation/perf-stat.txt b/tools/perf/Documentation/perf-stat.txt
index cf0c310..784976a 100644
--- a/tools/perf/Documentation/perf-stat.txt
+++ b/tools/perf/Documentation/perf-stat.txt
@@ -52,7 +52,7 @@ OPTIONS
-r::
--repeat=<n>::
- repeat command and print average + stddev (max: 100)
+ repeat command and print average + stddev (max: 100). 0 means forever.
-B::
--big-num::
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index c247fac..b02b3a9 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -87,6 +87,7 @@ static FILE *output = NULL;
static const char *pre_cmd = NULL;
static const char *post_cmd = NULL;
static bool sync_run = false;
+static bool forever = false;
static volatile int done = 0;
@@ -94,6 +95,11 @@ struct perf_stat {
struct stats res_stats[3];
};
+static void perf_evsel__reset_stat_priv(struct perf_evsel *evsel)
+{
+ memset(evsel->priv, 0, sizeof(struct perf_stat));
+}
+
static int perf_evsel__alloc_stat_priv(struct perf_evsel *evsel)
{
evsel->priv = zalloc(sizeof(struct perf_stat));
@@ -129,6 +135,22 @@ static struct stats runtime_itlb_cache_stats[MAX_NR_CPUS];
static struct stats runtime_dtlb_cache_stats[MAX_NR_CPUS];
static struct stats walltime_nsecs_stats;
+static void reset_stats(void)
+{
+ memset(runtime_nsecs_stats, 0, sizeof(runtime_nsecs_stats));
+ memset(runtime_cycles_stats, 0, sizeof(runtime_cycles_stats));
+ memset(runtime_stalled_cycles_front_stats, 0, sizeof(runtime_stalled_cycles_front_stats));
+ memset(runtime_stalled_cycles_back_stats, 0, sizeof(runtime_stalled_cycles_back_stats));
+ memset(runtime_branches_stats, 0, sizeof(runtime_branches_stats));
+ memset(runtime_cacherefs_stats, 0, sizeof(runtime_cacherefs_stats));
+ memset(runtime_l1_dcache_stats, 0, sizeof(runtime_l1_dcache_stats));
+ memset(runtime_l1_icache_stats, 0, sizeof(runtime_l1_icache_stats));
+ memset(runtime_ll_cache_stats, 0, sizeof(runtime_ll_cache_stats));
+ memset(runtime_itlb_cache_stats, 0, sizeof(runtime_itlb_cache_stats));
+ memset(runtime_dtlb_cache_stats, 0, sizeof(runtime_dtlb_cache_stats));
+ memset(&walltime_nsecs_stats, 0, sizeof(walltime_nsecs_stats));
+}
+
static int create_perf_stat_counter(struct perf_evsel *evsel)
{
struct perf_event_attr *attr = &evsel->attr;
@@ -1120,7 +1142,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
OPT_INCR('v', "verbose", &verbose,
"be more verbose (show counter open errors, etc)"),
OPT_INTEGER('r', "repeat", &run_count,
- "repeat command and print average + stddev (max: 100)"),
+ "repeat command and print average + stddev (max: 100, forever: 0)"),
OPT_BOOLEAN('n', "null", &null_run,
"null run - dont start any counters"),
OPT_INCR('d', "detailed", &detailed_run,
@@ -1220,8 +1242,12 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
if (!argc && !perf_target__has_task(&target))
usage_with_options(stat_usage, options);
- if (run_count <= 0)
+ if (run_count < 0) {
usage_with_options(stat_usage, options);
+ } else if (run_count == 0) {
+ forever = true;
+ run_count = 1;
+ }
/* no_aggr, cgroup are for system-wide only */
if ((no_aggr || nr_cgroups) && !perf_target__has_cpu(&target)) {
@@ -1259,21 +1285,31 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
* task, but being ignored by perf stat itself:
*/
atexit(sig_atexit);
- signal(SIGINT, skip_signal);
+ if (!forever)
+ signal(SIGINT, skip_signal);
signal(SIGALRM, skip_signal);
signal(SIGABRT, skip_signal);
status = 0;
- for (run_idx = 0; run_idx < run_count; run_idx++) {
+ for (run_idx = 0; forever || run_idx < run_count; run_idx++) {
if (run_count != 1 && verbose)
fprintf(output, "[ perf stat: executing run #%d ... ]\n",
- run_idx + 1);
+ run_idx + 1);
status = run_perf_stat(argc, argv);
+ if (forever && status != -1) {
+ print_stat(argc, argv);
+ list_for_each_entry(pos, &evsel_list->entries, node) {
+ perf_evsel__reset_stat_priv(pos);
+ perf_evsel__reset_counts(pos, perf_evsel__nr_cpus(pos));
+ }
+ reset_stats();
+ }
}
- if (status != -1)
+ if (!forever && status != -1)
print_stat(argc, argv);
+
out_free_fd:
list_for_each_entry(pos, &evsel_list->entries, node)
perf_evsel__free_stat_priv(pos);
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index 1b16dd1..0c49e79 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -580,6 +580,12 @@ int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads)
return 0;
}
+void perf_evsel__reset_counts(struct perf_evsel *evsel, int ncpus)
+{
+ memset(evsel->counts, 0, (sizeof(*evsel->counts) +
+ (ncpus * sizeof(struct perf_counts_values))));
+}
+
int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus)
{
evsel->counts = zalloc((sizeof(*evsel->counts) +
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index 3d2b801..364d820 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -114,6 +114,7 @@ const char *perf_evsel__name(struct perf_evsel *evsel);
int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads);
int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads);
int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus);
+void perf_evsel__reset_counts(struct perf_evsel *evsel, int ncpus);
void perf_evsel__free_fd(struct perf_evsel *evsel);
void perf_evsel__free_id(struct perf_evsel *evsel);
void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus, int nthreads);
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [patch] perf stat: --repeat forever
2013-02-15 22:28 [patch] perf stat: --repeat forever Frederik Deweerdt
@ 2013-02-18 15:04 ` Arnaldo Carvalho de Melo
2013-02-18 17:24 ` Frederik Deweerdt
2013-03-01 18:02 ` [patch v2] " Frederik Deweerdt
1 sibling, 1 reply; 5+ messages in thread
From: Arnaldo Carvalho de Melo @ 2013-02-18 15:04 UTC (permalink / raw)
To: Frederik Deweerdt; +Cc: a.p.zijlstra, paulus, mingo, linux-kernel
Em Fri, Feb 15, 2013 at 05:28:49PM -0500, Frederik Deweerdt escreveu:
> Hi,
>
> The following patch causes 'perf stat --repeat 0' to be interpreted as
> 'forever', displaying the stats for every run.
>
> We act as if a single run was asked, and reset the stats in each
> iteration. In this mode SIGINT is passed to perf to be able to stop the
> loop with Ctrl+C.
It is not applying to my perf/core branch, please take a look at:
git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux
branch perf/core
Also please consider removing the need to specify 0, i.e. I think this
is possible and shorter:
# perf stat -a --repeat sleep 1
- Arnaldo
> Regards,
> Frederik
>
>
> Signed-off-by: Frederik Deweerdt <frederik.deweerdt@gmail.com>
>
> diff --git a/tools/perf/Documentation/perf-stat.txt b/tools/perf/Documentation/perf-stat.txt
> index cf0c310..784976a 100644
> --- a/tools/perf/Documentation/perf-stat.txt
> +++ b/tools/perf/Documentation/perf-stat.txt
> @@ -52,7 +52,7 @@ OPTIONS
>
> -r::
> --repeat=<n>::
> - repeat command and print average + stddev (max: 100)
> + repeat command and print average + stddev (max: 100). 0 means forever.
>
> -B::
> --big-num::
> diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
> index c247fac..b02b3a9 100644
> --- a/tools/perf/builtin-stat.c
> +++ b/tools/perf/builtin-stat.c
> @@ -87,6 +87,7 @@ static FILE *output = NULL;
> static const char *pre_cmd = NULL;
> static const char *post_cmd = NULL;
> static bool sync_run = false;
> +static bool forever = false;
>
> static volatile int done = 0;
>
> @@ -94,6 +95,11 @@ struct perf_stat {
> struct stats res_stats[3];
> };
>
> +static void perf_evsel__reset_stat_priv(struct perf_evsel *evsel)
> +{
> + memset(evsel->priv, 0, sizeof(struct perf_stat));
> +}
> +
> static int perf_evsel__alloc_stat_priv(struct perf_evsel *evsel)
> {
> evsel->priv = zalloc(sizeof(struct perf_stat));
> @@ -129,6 +135,22 @@ static struct stats runtime_itlb_cache_stats[MAX_NR_CPUS];
> static struct stats runtime_dtlb_cache_stats[MAX_NR_CPUS];
> static struct stats walltime_nsecs_stats;
>
> +static void reset_stats(void)
> +{
> + memset(runtime_nsecs_stats, 0, sizeof(runtime_nsecs_stats));
> + memset(runtime_cycles_stats, 0, sizeof(runtime_cycles_stats));
> + memset(runtime_stalled_cycles_front_stats, 0, sizeof(runtime_stalled_cycles_front_stats));
> + memset(runtime_stalled_cycles_back_stats, 0, sizeof(runtime_stalled_cycles_back_stats));
> + memset(runtime_branches_stats, 0, sizeof(runtime_branches_stats));
> + memset(runtime_cacherefs_stats, 0, sizeof(runtime_cacherefs_stats));
> + memset(runtime_l1_dcache_stats, 0, sizeof(runtime_l1_dcache_stats));
> + memset(runtime_l1_icache_stats, 0, sizeof(runtime_l1_icache_stats));
> + memset(runtime_ll_cache_stats, 0, sizeof(runtime_ll_cache_stats));
> + memset(runtime_itlb_cache_stats, 0, sizeof(runtime_itlb_cache_stats));
> + memset(runtime_dtlb_cache_stats, 0, sizeof(runtime_dtlb_cache_stats));
> + memset(&walltime_nsecs_stats, 0, sizeof(walltime_nsecs_stats));
> +}
> +
> static int create_perf_stat_counter(struct perf_evsel *evsel)
> {
> struct perf_event_attr *attr = &evsel->attr;
> @@ -1120,7 +1142,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
> OPT_INCR('v', "verbose", &verbose,
> "be more verbose (show counter open errors, etc)"),
> OPT_INTEGER('r', "repeat", &run_count,
> - "repeat command and print average + stddev (max: 100)"),
> + "repeat command and print average + stddev (max: 100, forever: 0)"),
> OPT_BOOLEAN('n', "null", &null_run,
> "null run - dont start any counters"),
> OPT_INCR('d', "detailed", &detailed_run,
> @@ -1220,8 +1242,12 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
>
> if (!argc && !perf_target__has_task(&target))
> usage_with_options(stat_usage, options);
> - if (run_count <= 0)
> + if (run_count < 0) {
> usage_with_options(stat_usage, options);
> + } else if (run_count == 0) {
> + forever = true;
> + run_count = 1;
> + }
>
> /* no_aggr, cgroup are for system-wide only */
> if ((no_aggr || nr_cgroups) && !perf_target__has_cpu(&target)) {
> @@ -1259,21 +1285,31 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
> * task, but being ignored by perf stat itself:
> */
> atexit(sig_atexit);
> - signal(SIGINT, skip_signal);
> + if (!forever)
> + signal(SIGINT, skip_signal);
> signal(SIGALRM, skip_signal);
> signal(SIGABRT, skip_signal);
>
> status = 0;
> - for (run_idx = 0; run_idx < run_count; run_idx++) {
> + for (run_idx = 0; forever || run_idx < run_count; run_idx++) {
> if (run_count != 1 && verbose)
> fprintf(output, "[ perf stat: executing run #%d ... ]\n",
> - run_idx + 1);
> + run_idx + 1);
>
> status = run_perf_stat(argc, argv);
> + if (forever && status != -1) {
> + print_stat(argc, argv);
> + list_for_each_entry(pos, &evsel_list->entries, node) {
> + perf_evsel__reset_stat_priv(pos);
> + perf_evsel__reset_counts(pos, perf_evsel__nr_cpus(pos));
> + }
> + reset_stats();
> + }
> }
>
> - if (status != -1)
> + if (!forever && status != -1)
> print_stat(argc, argv);
> +
> out_free_fd:
> list_for_each_entry(pos, &evsel_list->entries, node)
> perf_evsel__free_stat_priv(pos);
> diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
> index 1b16dd1..0c49e79 100644
> --- a/tools/perf/util/evsel.c
> +++ b/tools/perf/util/evsel.c
> @@ -580,6 +580,12 @@ int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads)
> return 0;
> }
>
> +void perf_evsel__reset_counts(struct perf_evsel *evsel, int ncpus)
> +{
> + memset(evsel->counts, 0, (sizeof(*evsel->counts) +
> + (ncpus * sizeof(struct perf_counts_values))));
> +}
> +
> int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus)
> {
> evsel->counts = zalloc((sizeof(*evsel->counts) +
> diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
> index 3d2b801..364d820 100644
> --- a/tools/perf/util/evsel.h
> +++ b/tools/perf/util/evsel.h
> @@ -114,6 +114,7 @@ const char *perf_evsel__name(struct perf_evsel *evsel);
> int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads);
> int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads);
> int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus);
> +void perf_evsel__reset_counts(struct perf_evsel *evsel, int ncpus);
> void perf_evsel__free_fd(struct perf_evsel *evsel);
> void perf_evsel__free_id(struct perf_evsel *evsel);
> void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus, int nthreads);
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [patch] perf stat: --repeat forever
2013-02-18 15:04 ` Arnaldo Carvalho de Melo
@ 2013-02-18 17:24 ` Frederik Deweerdt
0 siblings, 0 replies; 5+ messages in thread
From: Frederik Deweerdt @ 2013-02-18 17:24 UTC (permalink / raw)
To: Arnaldo Carvalho de Melo; +Cc: a.p.zijlstra, paulus, mingo, linux-kernel
Hi Arnaldo,
On Mon, Feb 18, 2013 at 12:04:44PM -0300, Arnaldo Carvalho de Melo wrote:
> Em Fri, Feb 15, 2013 at 05:28:49PM -0500, Frederik Deweerdt escreveu:
> > Hi,
> >
> > The following patch causes 'perf stat --repeat 0' to be interpreted as
> > 'forever', displaying the stats for every run.
> >
> > We act as if a single run was asked, and reset the stats in each
> > iteration. In this mode SIGINT is passed to perf to be able to stop the
> > loop with Ctrl+C.
>
> It is not applying to my perf/core branch, please take a look at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux
>
> branch perf/core
>
Noted, I'll base the next version based on this branch.
> Also please consider removing the need to specify 0, i.e. I think this
> is possible and shorter:
>
> # perf stat -a --repeat sleep 1
I'm unsure how to achieve that. I've looked at the different OPT_*
defines, but I can't find one that would do the job. Any pointers?
Thanks,
Frederik
^ permalink raw reply [flat|nested] 5+ messages in thread
* [patch v2] perf stat: --repeat forever
2013-02-15 22:28 [patch] perf stat: --repeat forever Frederik Deweerdt
2013-02-18 15:04 ` Arnaldo Carvalho de Melo
@ 2013-03-01 18:02 ` Frederik Deweerdt
2013-03-21 11:48 ` [tip:perf/core] perf stat: Introduce " tip-bot for Frederik Deweerdt
1 sibling, 1 reply; 5+ messages in thread
From: Frederik Deweerdt @ 2013-03-01 18:02 UTC (permalink / raw)
To: acme; +Cc: linux-kernel, a.p.zijlstra, paulus, mingo
Hi,
[This is based on git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux perf/core]
The following patch causes 'perf stat --repeat 0' to be interpreted as
'forever', displaying the stats for every run.
We act as if a single run was asked, and reset the stats in each
iteration. In this mode SIGINT is passed to perf to be able to stop the
loop with Ctrl+C.
Regards,
Frederik
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@gmail.com>
diff --git a/tools/perf/Documentation/perf-stat.txt b/tools/perf/Documentation/perf-stat.txt
index faf4f4f..23e587a 100644
--- a/tools/perf/Documentation/perf-stat.txt
+++ b/tools/perf/Documentation/perf-stat.txt
@@ -52,7 +52,7 @@ OPTIONS
-r::
--repeat=<n>::
- repeat command and print average + stddev (max: 100)
+ repeat command and print average + stddev (max: 100). 0 means forever.
-B::
--big-num::
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 9984876..f59ab78 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -94,6 +94,7 @@ static const char *pre_cmd = NULL;
static const char *post_cmd = NULL;
static bool sync_run = false;
static unsigned int interval = 0;
+static bool forever = false;
static struct timespec ref_time;
static struct cpu_map *sock_map;
@@ -125,6 +126,11 @@ static inline int perf_evsel__nr_cpus(struct perf_evsel *evsel)
return perf_evsel__cpus(evsel)->nr;
}
+static void perf_evsel__reset_stat_priv(struct perf_evsel *evsel)
+{
+ memset(evsel->priv, 0, sizeof(struct perf_stat));
+}
+
static int perf_evsel__alloc_stat_priv(struct perf_evsel *evsel)
{
evsel->priv = zalloc(sizeof(struct perf_stat));
@@ -173,6 +179,22 @@ static struct stats runtime_itlb_cache_stats[MAX_NR_CPUS];
static struct stats runtime_dtlb_cache_stats[MAX_NR_CPUS];
static struct stats walltime_nsecs_stats;
+static void reset_stats(void)
+{
+ memset(runtime_nsecs_stats, 0, sizeof(runtime_nsecs_stats));
+ memset(runtime_cycles_stats, 0, sizeof(runtime_cycles_stats));
+ memset(runtime_stalled_cycles_front_stats, 0, sizeof(runtime_stalled_cycles_front_stats));
+ memset(runtime_stalled_cycles_back_stats, 0, sizeof(runtime_stalled_cycles_back_stats));
+ memset(runtime_branches_stats, 0, sizeof(runtime_branches_stats));
+ memset(runtime_cacherefs_stats, 0, sizeof(runtime_cacherefs_stats));
+ memset(runtime_l1_dcache_stats, 0, sizeof(runtime_l1_dcache_stats));
+ memset(runtime_l1_icache_stats, 0, sizeof(runtime_l1_icache_stats));
+ memset(runtime_ll_cache_stats, 0, sizeof(runtime_ll_cache_stats));
+ memset(runtime_itlb_cache_stats, 0, sizeof(runtime_itlb_cache_stats));
+ memset(runtime_dtlb_cache_stats, 0, sizeof(runtime_dtlb_cache_stats));
+ memset(&walltime_nsecs_stats, 0, sizeof(walltime_nsecs_stats));
+}
+
static int create_perf_stat_counter(struct perf_evsel *evsel)
{
struct perf_event_attr *attr = &evsel->attr;
@@ -1296,7 +1318,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
OPT_INCR('v', "verbose", &verbose,
"be more verbose (show counter open errors, etc)"),
OPT_INTEGER('r', "repeat", &run_count,
- "repeat command and print average + stddev (max: 100)"),
+ "repeat command and print average + stddev (max: 100, forever: 0)"),
OPT_BOOLEAN('n', "null", &null_run,
"null run - dont start any counters"),
OPT_INCR('d', "detailed", &detailed_run,
@@ -1399,8 +1421,12 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
if (!argc && !perf_target__has_task(&target))
usage_with_options(stat_usage, options);
- if (run_count <= 0)
+ if (run_count < 0) {
usage_with_options(stat_usage, options);
+ } else if (run_count == 0) {
+ forever = true;
+ run_count = 1;
+ }
/* no_aggr, cgroup are for system-wide only */
if ((no_aggr || nr_cgroups) && !perf_target__has_cpu(&target)) {
@@ -1457,21 +1483,30 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
* task, but being ignored by perf stat itself:
*/
atexit(sig_atexit);
- signal(SIGINT, skip_signal);
+ if (!forever)
+ signal(SIGINT, skip_signal);
signal(SIGCHLD, skip_signal);
signal(SIGALRM, skip_signal);
signal(SIGABRT, skip_signal);
status = 0;
- for (run_idx = 0; run_idx < run_count; run_idx++) {
+ for (run_idx = 0; forever || run_idx < run_count; run_idx++) {
if (run_count != 1 && verbose)
fprintf(output, "[ perf stat: executing run #%d ... ]\n",
run_idx + 1);
status = run_perf_stat(argc, argv);
+ if (forever && status != -1) {
+ print_stat(argc, argv);
+ list_for_each_entry(pos, &evsel_list->entries, node) {
+ perf_evsel__reset_stat_priv(pos);
+ perf_evsel__reset_counts(pos, perf_evsel__nr_cpus(pos));
+ }
+ reset_stats();
+ }
}
- if (status != -1 && !interval)
+ if (!forever && status != -1 && !interval)
print_stat(argc, argv);
out_free_fd:
list_for_each_entry(pos, &evsel_list->entries, node) {
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index dc16231..6e55a5f 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -633,6 +633,12 @@ int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads)
return 0;
}
+void perf_evsel__reset_counts(struct perf_evsel *evsel, int ncpus)
+{
+ memset(evsel->counts, 0, (sizeof(*evsel->counts) +
+ (ncpus * sizeof(struct perf_counts_values))));
+}
+
int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus)
{
evsel->counts = zalloc((sizeof(*evsel->counts) +
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index 52021c3..d1eb4da 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -120,6 +120,7 @@ int perf_evsel__group_desc(struct perf_evsel *evsel, char *buf, size_t size);
int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads);
int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads);
int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus);
+void perf_evsel__reset_counts(struct perf_evsel *evsel, int ncpus);
void perf_evsel__free_fd(struct perf_evsel *evsel);
void perf_evsel__free_id(struct perf_evsel *evsel);
void perf_evsel__free_counts(struct perf_evsel *evsel);
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [tip:perf/core] perf stat: Introduce --repeat forever
2013-03-01 18:02 ` [patch v2] " Frederik Deweerdt
@ 2013-03-21 11:48 ` tip-bot for Frederik Deweerdt
0 siblings, 0 replies; 5+ messages in thread
From: tip-bot for Frederik Deweerdt @ 2013-03-21 11:48 UTC (permalink / raw)
To: linux-tip-commits
Cc: acme, linux-kernel, paulus, mingo, frederik.deweerdt, hpa, mingo,
a.p.zijlstra, frederik.deweerdt, tglx
Commit-ID: a7e191c376fad084d9f3c7ac89a1f7c47462ebc8
Gitweb: http://git.kernel.org/tip/a7e191c376fad084d9f3c7ac89a1f7c47462ebc8
Author: Frederik Deweerdt <frederik.deweerdt@xprog.eu>
AuthorDate: Fri, 1 Mar 2013 13:02:27 -0500
Committer: Arnaldo Carvalho de Melo <acme@redhat.com>
CommitDate: Fri, 15 Mar 2013 14:01:26 -0300
perf stat: Introduce --repeat forever
The following patch causes 'perf stat --repeat 0' to be interpreted as
'forever', displaying the stats for every run.
We act as if a single run was asked, and reset the stats in each
iteration. In this mode SIGINT is passed to perf to be able to stop the
loop with Ctrl+C.
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20130301180227.GA24385@ks398093.ip-192-95-24.net
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
tools/perf/Documentation/perf-stat.txt | 2 +-
tools/perf/builtin-stat.c | 45 ++++++++++++++++++++++++++++++----
tools/perf/util/evsel.c | 6 +++++
tools/perf/util/evsel.h | 1 +
4 files changed, 48 insertions(+), 6 deletions(-)
diff --git a/tools/perf/Documentation/perf-stat.txt b/tools/perf/Documentation/perf-stat.txt
index faf4f4f..23e587a 100644
--- a/tools/perf/Documentation/perf-stat.txt
+++ b/tools/perf/Documentation/perf-stat.txt
@@ -52,7 +52,7 @@ OPTIONS
-r::
--repeat=<n>::
- repeat command and print average + stddev (max: 100)
+ repeat command and print average + stddev (max: 100). 0 means forever.
-B::
--big-num::
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 69fe6ed..021783a 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -94,6 +94,7 @@ static const char *pre_cmd = NULL;
static const char *post_cmd = NULL;
static bool sync_run = false;
static unsigned int interval = 0;
+static bool forever = false;
static struct timespec ref_time;
static struct cpu_map *sock_map;
@@ -125,6 +126,11 @@ static inline int perf_evsel__nr_cpus(struct perf_evsel *evsel)
return perf_evsel__cpus(evsel)->nr;
}
+static void perf_evsel__reset_stat_priv(struct perf_evsel *evsel)
+{
+ memset(evsel->priv, 0, sizeof(struct perf_stat));
+}
+
static int perf_evsel__alloc_stat_priv(struct perf_evsel *evsel)
{
evsel->priv = zalloc(sizeof(struct perf_stat));
@@ -173,6 +179,22 @@ static struct stats runtime_itlb_cache_stats[MAX_NR_CPUS];
static struct stats runtime_dtlb_cache_stats[MAX_NR_CPUS];
static struct stats walltime_nsecs_stats;
+static void reset_stats(void)
+{
+ memset(runtime_nsecs_stats, 0, sizeof(runtime_nsecs_stats));
+ memset(runtime_cycles_stats, 0, sizeof(runtime_cycles_stats));
+ memset(runtime_stalled_cycles_front_stats, 0, sizeof(runtime_stalled_cycles_front_stats));
+ memset(runtime_stalled_cycles_back_stats, 0, sizeof(runtime_stalled_cycles_back_stats));
+ memset(runtime_branches_stats, 0, sizeof(runtime_branches_stats));
+ memset(runtime_cacherefs_stats, 0, sizeof(runtime_cacherefs_stats));
+ memset(runtime_l1_dcache_stats, 0, sizeof(runtime_l1_dcache_stats));
+ memset(runtime_l1_icache_stats, 0, sizeof(runtime_l1_icache_stats));
+ memset(runtime_ll_cache_stats, 0, sizeof(runtime_ll_cache_stats));
+ memset(runtime_itlb_cache_stats, 0, sizeof(runtime_itlb_cache_stats));
+ memset(runtime_dtlb_cache_stats, 0, sizeof(runtime_dtlb_cache_stats));
+ memset(&walltime_nsecs_stats, 0, sizeof(walltime_nsecs_stats));
+}
+
static int create_perf_stat_counter(struct perf_evsel *evsel)
{
struct perf_event_attr *attr = &evsel->attr;
@@ -1252,7 +1274,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
OPT_INCR('v', "verbose", &verbose,
"be more verbose (show counter open errors, etc)"),
OPT_INTEGER('r', "repeat", &run_count,
- "repeat command and print average + stddev (max: 100)"),
+ "repeat command and print average + stddev (max: 100, forever: 0)"),
OPT_BOOLEAN('n', "null", &null_run,
"null run - dont start any counters"),
OPT_INCR('d', "detailed", &detailed_run,
@@ -1355,8 +1377,12 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
if (!argc && !perf_target__has_task(&target))
usage_with_options(stat_usage, options);
- if (run_count <= 0)
+ if (run_count < 0) {
usage_with_options(stat_usage, options);
+ } else if (run_count == 0) {
+ forever = true;
+ run_count = 1;
+ }
/* no_aggr, cgroup are for system-wide only */
if ((no_aggr || nr_cgroups) && !perf_target__has_cpu(&target)) {
@@ -1413,21 +1439,30 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
* task, but being ignored by perf stat itself:
*/
atexit(sig_atexit);
- signal(SIGINT, skip_signal);
+ if (!forever)
+ signal(SIGINT, skip_signal);
signal(SIGCHLD, skip_signal);
signal(SIGALRM, skip_signal);
signal(SIGABRT, skip_signal);
status = 0;
- for (run_idx = 0; run_idx < run_count; run_idx++) {
+ for (run_idx = 0; forever || run_idx < run_count; run_idx++) {
if (run_count != 1 && verbose)
fprintf(output, "[ perf stat: executing run #%d ... ]\n",
run_idx + 1);
status = run_perf_stat(argc, argv);
+ if (forever && status != -1) {
+ print_stat(argc, argv);
+ list_for_each_entry(pos, &evsel_list->entries, node) {
+ perf_evsel__reset_stat_priv(pos);
+ perf_evsel__reset_counts(pos, perf_evsel__nr_cpus(pos));
+ }
+ reset_stats();
+ }
}
- if (status != -1 && !interval)
+ if (!forever && status != -1 && !interval)
print_stat(argc, argv);
out_free_fd:
list_for_each_entry(pos, &evsel_list->entries, node) {
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index 7fde9fb..1adb824 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -633,6 +633,12 @@ int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads)
return 0;
}
+void perf_evsel__reset_counts(struct perf_evsel *evsel, int ncpus)
+{
+ memset(evsel->counts, 0, (sizeof(*evsel->counts) +
+ (ncpus * sizeof(struct perf_counts_values))));
+}
+
int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus)
{
evsel->counts = zalloc((sizeof(*evsel->counts) +
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index bf758e5..3f156cc 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -121,6 +121,7 @@ int perf_evsel__group_desc(struct perf_evsel *evsel, char *buf, size_t size);
int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads);
int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads);
int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus);
+void perf_evsel__reset_counts(struct perf_evsel *evsel, int ncpus);
void perf_evsel__free_fd(struct perf_evsel *evsel);
void perf_evsel__free_id(struct perf_evsel *evsel);
void perf_evsel__free_counts(struct perf_evsel *evsel);
^ permalink raw reply related [flat|nested] 5+ messages in thread
end of thread, other threads:[~2013-03-21 11:50 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-02-15 22:28 [patch] perf stat: --repeat forever Frederik Deweerdt
2013-02-18 15:04 ` Arnaldo Carvalho de Melo
2013-02-18 17:24 ` Frederik Deweerdt
2013-03-01 18:02 ` [patch v2] " Frederik Deweerdt
2013-03-21 11:48 ` [tip:perf/core] perf stat: Introduce " tip-bot for Frederik Deweerdt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).