Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1344865 > unrolled thread

[PATCH 1/6] perf, tools, stat: Check existence of frontend/backed stalled cycles

Started byAndi Kleen <andi@firstfloor.org>
First post2016-02-27 01:30 +0100
Last post2016-02-29 15:50 +0100
Articles 3 — 3 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  [PATCH 1/6] perf, tools, stat: Check existence of frontend/backed stalled cycles Andi Kleen <andi@firstfloor.org> - 2016-02-27 01:30 +0100
    Re: [PATCH 1/6] perf, tools, stat: Check existence of  frontend/backed stalled cycles Jiri Olsa <jolsa@redhat.com> - 2016-02-29 11:20 +0100
      Re: [PATCH 1/6] perf, tools, stat: Check existence of  frontend/backed stalled cycles Arnaldo Carvalho de Melo <acme@kernel.org> - 2016-02-29 15:50 +0100

#1344865 — [PATCH 1/6] perf, tools, stat: Check existence of frontend/backed stalled cycles

FromAndi Kleen <andi@firstfloor.org>
Date2016-02-27 01:30 +0100
Subject[PATCH 1/6] perf, tools, stat: Check existence of frontend/backed stalled cycles
Message-ID<r6AVI-2Yh-15@gated-at.bofh.it>
From: Andi Kleen <ak@linux.intel.com>

Only put the frontend/backend stalled cycles into the default
perf stat events when the CPU actually supports them.

This avoids empty columns with --metric-only on newer Intel CPUs.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
 tools/perf/builtin-stat.c | 22 ++++++++++++++++++++--
 1 file changed, 20 insertions(+), 2 deletions(-)

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 8c0bc0f..24f222d 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -1441,7 +1441,7 @@ static int perf_stat_init_aggr_mode_file(struct perf_stat *st)
  */
 static int add_default_attributes(void)
 {
-	struct perf_event_attr default_attrs[] = {
+	struct perf_event_attr default_attrs0[] = {
 
   { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_TASK_CLOCK		},
   { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_CONTEXT_SWITCHES	},
@@ -1449,8 +1449,14 @@ static int add_default_attributes(void)
   { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_PAGE_FAULTS		},
 
   { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_CPU_CYCLES		},
+};
+	struct perf_event_attr frontend_attrs[] = {
   { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_STALLED_CYCLES_FRONTEND	},
+};
+	struct perf_event_attr backend_attrs[] = {
   { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_STALLED_CYCLES_BACKEND	},
+};
+	struct perf_event_attr default_attrs1[] = {
   { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_INSTRUCTIONS		},
   { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_BRANCH_INSTRUCTIONS	},
   { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_BRANCH_MISSES		},
@@ -1567,7 +1573,19 @@ static int add_default_attributes(void)
 	}
 
 	if (!evsel_list->nr_entries) {
-		if (perf_evlist__add_default_attrs(evsel_list, default_attrs) < 0)
+		if (perf_evlist__add_default_attrs(evsel_list, default_attrs0) < 0)
+			return -1;
+		if (pmu_have_event("cpu", "stalled-cycles-frontend")) {
+			if (perf_evlist__add_default_attrs(evsel_list,
+						frontend_attrs) < 0)
+				return -1;
+		}
+		if (pmu_have_event("cpu", "stalled-cycles-backend")) {
+			if (perf_evlist__add_default_attrs(evsel_list,
+						backend_attrs) < 0)
+				return -1;
+		}
+		if (perf_evlist__add_default_attrs(evsel_list, default_attrs1) < 0)
 			return -1;
 	}
 
-- 
2.5.0

[toc] | [next] | [standalone]


#1345623 — Re: [PATCH 1/6] perf, tools, stat: Check existence of frontend/backed stalled cycles

FromJiri Olsa <jolsa@redhat.com>
Date2016-02-29 11:20 +0100
SubjectRe: [PATCH 1/6] perf, tools, stat: Check existence of frontend/backed stalled cycles
Message-ID<r7t5M-1ES-17@gated-at.bofh.it>
In reply to#1344865
On Fri, Feb 26, 2016 at 04:27:56PM -0800, Andi Kleen wrote:
> From: Andi Kleen <ak@linux.intel.com>
> 
> Only put the frontend/backend stalled cycles into the default
> perf stat events when the CPU actually supports them.
> 
> This avoids empty columns with --metric-only on newer Intel CPUs.
> 
> Signed-off-by: Andi Kleen <ak@linux.intel.com>

Acked-by: Jiri Olsa <jolsa@kernel.org>

thanks,
jirka

> ---
>  tools/perf/builtin-stat.c | 22 ++++++++++++++++++++--
>  1 file changed, 20 insertions(+), 2 deletions(-)
> 
> diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
> index 8c0bc0f..24f222d 100644
> --- a/tools/perf/builtin-stat.c
> +++ b/tools/perf/builtin-stat.c
> @@ -1441,7 +1441,7 @@ static int perf_stat_init_aggr_mode_file(struct perf_stat *st)
>   */
>  static int add_default_attributes(void)
>  {
> -	struct perf_event_attr default_attrs[] = {
> +	struct perf_event_attr default_attrs0[] = {
>  
>    { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_TASK_CLOCK		},
>    { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_CONTEXT_SWITCHES	},
> @@ -1449,8 +1449,14 @@ static int add_default_attributes(void)
>    { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_PAGE_FAULTS		},
>  
>    { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_CPU_CYCLES		},
> +};
> +	struct perf_event_attr frontend_attrs[] = {
>    { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_STALLED_CYCLES_FRONTEND	},
> +};
> +	struct perf_event_attr backend_attrs[] = {
>    { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_STALLED_CYCLES_BACKEND	},
> +};
> +	struct perf_event_attr default_attrs1[] = {
>    { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_INSTRUCTIONS		},
>    { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_BRANCH_INSTRUCTIONS	},
>    { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_BRANCH_MISSES		},
> @@ -1567,7 +1573,19 @@ static int add_default_attributes(void)
>  	}
>  
>  	if (!evsel_list->nr_entries) {
> -		if (perf_evlist__add_default_attrs(evsel_list, default_attrs) < 0)
> +		if (perf_evlist__add_default_attrs(evsel_list, default_attrs0) < 0)
> +			return -1;
> +		if (pmu_have_event("cpu", "stalled-cycles-frontend")) {
> +			if (perf_evlist__add_default_attrs(evsel_list,
> +						frontend_attrs) < 0)
> +				return -1;
> +		}
> +		if (pmu_have_event("cpu", "stalled-cycles-backend")) {
> +			if (perf_evlist__add_default_attrs(evsel_list,
> +						backend_attrs) < 0)
> +				return -1;
> +		}
> +		if (perf_evlist__add_default_attrs(evsel_list, default_attrs1) < 0)
>  			return -1;
>  	}
>  
> -- 
> 2.5.0
> 

[toc] | [prev] | [next] | [standalone]


#1345873 — Re: [PATCH 1/6] perf, tools, stat: Check existence of frontend/backed stalled cycles

FromArnaldo Carvalho de Melo <acme@kernel.org>
Date2016-02-29 15:50 +0100
SubjectRe: [PATCH 1/6] perf, tools, stat: Check existence of frontend/backed stalled cycles
Message-ID<r7xj4-4ik-11@gated-at.bofh.it>
In reply to#1345623
Em Mon, Feb 29, 2016 at 11:10:35AM +0100, Jiri Olsa escreveu:
> On Fri, Feb 26, 2016 at 04:27:56PM -0800, Andi Kleen wrote:
> > From: Andi Kleen <ak@linux.intel.com>
> > 
> > Only put the frontend/backend stalled cycles into the default
> > perf stat events when the CPU actually supports them.
> > 
> > This avoids empty columns with --metric-only on newer Intel CPUs.
> > 
> > Signed-off-by: Andi Kleen <ak@linux.intel.com>
> 
> Acked-by: Jiri Olsa <jolsa@kernel.org>

Applied and tested, thanks!

- Arnaldo

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web