mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Jiri Olsa <jolsa@redhat.com>
To: linux-kernel@vger.kernel.org
Cc: Jiri Olsa <jolsa@redhat.com>,
	Arnaldo Carvalho de Melo <acme@ghostprotocols.net>,
	Peter Zijlstra <a.p.zijlstra@chello.nl>,
	Ingo Molnar <mingo@elte.hu>, Paul Mackerras <paulus@samba.org>,
	Corey Ashford <cjashfor@linux.vnet.ibm.com>,
	Frederic Weisbecker <fweisbec@gmail.com>,
	Namhyung Kim <namhyung@kernel.org>
Subject: [PATCH 08/11] perf tool: Add 'S' event/group modifier to read sample value
Date: Sat, 20 Oct 2012 16:33:16 +0200	[thread overview]
Message-ID: <1350743599-4805-9-git-send-email-jolsa@redhat.com> (raw)
In-Reply-To: <1350743599-4805-1-git-send-email-jolsa@redhat.com>

Adding 'S' event/group modifier to specify that the event value/s
are read by PERF_SAMPLE_READ sample type processing, instead of the
period value offered by lower layers.

There's additional behaviour change for 'S' modifier being specified
on event group:

Currently all the events within a group makes samples. If user now
specifies 'S' within group modifier, only the leader will trigger
samples. The rest of events in the group will have sampling disabled.

And same as for single events, values of all events within the group
(including leader) are read by PERF_SAMPLE_READ sample type processing.

Following example will create event group with cycles and cache-misses
events, setting the cycles as group leader and the only event to actually
sample. Both cycles and cache-misses event period values are read by
PERF_SAMPLE_READ sample type processing with PERF_FORMAT_GROUP read format.

  $ perf record -e '{cycles,cache-misses}:S' ls

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/Documentation/perf-list.txt |  1 +
 tools/perf/util/evsel.c                | 25 +++++++++++++++++++++++++
 tools/perf/util/evsel.h                |  1 +
 tools/perf/util/parse-events.c         |  6 ++++++
 tools/perf/util/parse-events.l         |  2 +-
 5 files changed, 34 insertions(+), 1 deletion(-)

diff --git a/tools/perf/Documentation/perf-list.txt b/tools/perf/Documentation/perf-list.txt
index d1e39dc..7ddef65 100644
--- a/tools/perf/Documentation/perf-list.txt
+++ b/tools/perf/Documentation/perf-list.txt
@@ -29,6 +29,7 @@ counted. The following modifiers exist:
  G - guest counting (in KVM guests)
  H - host counting (not in KVM guests)
  p - precise level
+ S - read sample value (PERF_SAMPLE_READ)
 
 The 'p' modifier can be used for specifying how precise the instruction
 address should be. The 'p' modifier can be specified multiple times:
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index f37bec0..836a786 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -432,6 +432,7 @@ int perf_evsel__group_desc(struct perf_evsel *evsel, char *buf, size_t size)
 void perf_evsel__config(struct perf_evsel *evsel, struct perf_record_opts *opts,
 			struct perf_evsel *first)
 {
+	struct perf_evsel *leader = evsel->leader;
 	struct perf_event_attr *attr = &evsel->attr;
 	int track = !evsel->idx; /* only the first counter needs these */
 
@@ -442,6 +443,21 @@ void perf_evsel__config(struct perf_evsel *evsel, struct perf_record_opts *opts,
 			      PERF_FORMAT_TOTAL_TIME_RUNNING |
 			      PERF_FORMAT_ID;
 
+	if (evsel->sample_read) {
+		attr->sample_type |= PERF_SAMPLE_READ;
+
+		/*
+		 * Apply group format only if:
+		 * - we have a leader so there's a group and at
+		 *   least 2 events,
+		 * - we are a leader with more than 1 event
+		 */
+		if (evsel->leader || evsel->nr_members) {
+			attr->read_format |= PERF_FORMAT_GROUP;
+			attr->inherit = 0;
+		}
+	}
+
 	attr->sample_type  |= PERF_SAMPLE_IP | PERF_SAMPLE_TID;
 
 	/*
@@ -459,6 +475,15 @@ void perf_evsel__config(struct perf_evsel *evsel, struct perf_record_opts *opts,
 		}
 	}
 
+	/*
+	 * We are safe here, because we have a leader so there's at
+	 * least 2 events in the group.
+	 */
+	if (leader && leader->sample_read) {
+		attr->sample_freq   = 0;
+		attr->sample_period = 0;
+	}
+
 	if (opts->no_samples)
 		attr->sample_freq = 0;
 
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index 351b48e..1e34060 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -75,6 +75,7 @@ struct perf_evsel {
 	bool 			needs_swap;
 	/* parse modifier helper */
 	int			exclude_GH;
+	int			sample_read;
 	struct perf_evsel	*leader;
 	char			*group_name;
 	union {
diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
index 364e518..5a6ca8a 100644
--- a/tools/perf/util/parse-events.c
+++ b/tools/perf/util/parse-events.c
@@ -642,6 +642,7 @@ struct event_modifier {
 	int eG;
 	int precise;
 	int exclude_GH;
+	int sample_read;
 };
 
 static int get_event_modifier(struct event_modifier *mod, char *str,
@@ -653,6 +654,7 @@ static int get_event_modifier(struct event_modifier *mod, char *str,
 	int eH = evsel ? evsel->attr.exclude_host : 0;
 	int eG = evsel ? evsel->attr.exclude_guest : 0;
 	int precise = evsel ? evsel->attr.precise_ip : 0;
+	int sample_read = 0;
 
 	int exclude = eu | ek | eh;
 	int exclude_GH = evsel ? evsel->exclude_GH : 0;
@@ -690,6 +692,8 @@ static int get_event_modifier(struct event_modifier *mod, char *str,
 			eH = 0;
 		} else if (*str == 'p') {
 			precise++;
+		} else if (*str == 'S') {
+			sample_read = 1;
 		} else
 			break;
 
@@ -716,6 +720,7 @@ static int get_event_modifier(struct event_modifier *mod, char *str,
 	mod->eG = eG;
 	mod->precise = precise;
 	mod->exclude_GH = exclude_GH;
+	mod->sample_read = sample_read;
 	return 0;
 }
 
@@ -742,6 +747,7 @@ int parse_events__modifier_event(struct list_head *list, char *str, bool add)
 		evsel->attr.exclude_host   = mod.eH;
 		evsel->attr.exclude_guest  = mod.eG;
 		evsel->exclude_GH          = mod.exclude_GH;
+		evsel->sample_read         = mod.sample_read;
 	}
 
 	return 0;
diff --git a/tools/perf/util/parse-events.l b/tools/perf/util/parse-events.l
index c87efc1..391aabb 100644
--- a/tools/perf/util/parse-events.l
+++ b/tools/perf/util/parse-events.l
@@ -81,7 +81,7 @@ num_dec		[0-9]+
 num_hex		0x[a-fA-F0-9]+
 num_raw_hex	[a-fA-F0-9]+
 name		[a-zA-Z_*?][a-zA-Z0-9_*?]*
-modifier_event	[ukhpGH]{1,8}
+modifier_event	[ukhpGHS]{1,9}
 modifier_bp	[rwx]{1,3}
 
 %%
-- 
1.7.11.4


  parent reply	other threads:[~2012-10-20 14:35 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-10-20 14:33 [PATCH 00/11] perf tool: Add PERF_SAMPLE_READ sample read support Jiri Olsa
2012-10-20 14:33 ` [PATCH 01/11] perf: Add PERF_EVENT_IOC_ID ioctl to return event ID Jiri Olsa
2012-10-20 14:33 ` [PATCH 02/11] perf: Do not get values from disabled counters in group format read Jiri Olsa
2012-10-23 16:13   ` Peter Zijlstra
2012-10-23 16:50     ` Jiri Olsa
2012-10-24 12:01       ` Peter Zijlstra
2012-10-24 12:14         ` Jiri Olsa
2012-10-24 12:32           ` Peter Zijlstra
2012-10-24 16:03           ` Jiri Olsa
2012-10-20 14:33 ` [PATCH 03/11] perf tool: Use PERF_EVENT_IOC_ID perf ioctl to read event id Jiri Olsa
2012-10-22  7:57   ` Namhyung Kim
2012-10-22  8:40     ` Jiri Olsa
2012-10-20 14:33 ` [PATCH 04/11] perf tool: Add support for parsing PERF_SAMPLE_READ sample type Jiri Olsa
2012-10-22  8:56   ` Namhyung Kim
2012-10-20 14:33 ` [PATCH 05/11] perf tool: Fix event ID retrieval for group format read case Jiri Olsa
2012-10-20 14:33 ` [PATCH 06/11] perf tool: Add perf_evlist__id2sid function to get event ID related data Jiri Olsa
2012-10-20 14:33 ` [PATCH 07/11] perf tool: Add PERF_SAMPLE_READ sample related processing Jiri Olsa
2012-10-20 14:33 ` Jiri Olsa [this message]
2012-10-20 14:33 ` [PATCH 09/11] perf test: Add parse events tests for leader sampling Jiri Olsa
2012-10-22  8:58   ` Namhyung Kim
2012-10-22  9:12     ` Jiri Olsa
2012-10-20 14:33 ` [PATCH 10/11] perf tool: Display period values for all group members Jiri Olsa
2012-10-20 14:33 ` [PATCH 11/11] perf record: Fix mmap error output condition Jiri Olsa
2012-10-30 12:11   ` [tip:perf/core] " tip-bot for Jiri Olsa
2012-10-21 16:38 ` [PATCH 00/11] perf tool: Add PERF_SAMPLE_READ sample read support Ingo Molnar
2012-10-22  8:09   ` Jiri Olsa
2012-10-22  8:51     ` Namhyung Kim
2012-10-22  9:15       ` Jiri Olsa
2012-10-22  7:32 ` Namhyung Kim
2012-10-22  7:53   ` Jiri Olsa
2012-10-22  8:53     ` Namhyung Kim
2012-10-22  9:06       ` Jiri Olsa
2012-10-26  1:29 ` Namhyung Kim
2012-10-26  9:14   ` Peter Zijlstra
2012-10-26 10:23     ` Jiri Olsa
2012-10-26 15:39       ` Namhyung Kim
2012-10-26 16:14         ` David Ahern
2012-10-26 16:25           ` Namhyung Kim
2012-10-26 16:47             ` David Ahern
2012-10-26 17:00         ` Arnaldo Carvalho de Melo

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1350743599-4805-9-git-send-email-jolsa@redhat.com \
    --to=jolsa@redhat.com \
    --cc=a.p.zijlstra@chello.nl \
    --cc=acme@ghostprotocols.net \
    --cc=cjashfor@linux.vnet.ibm.com \
    --cc=fweisbec@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=namhyung@kernel.org \
    --cc=paulus@samba.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome