From: weilin.wang@intel.com
To: weilin.wang@intel.com, Ian Rogers <irogers@google.com>,
Kan Liang <kan.liang@linux.intel.com>,
Namhyung Kim <namhyung@kernel.org>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>,
Alexander Shishkin <alexander.shishkin@linux.intel.com>,
Jiri Olsa <jolsa@kernel.org>,
Adrian Hunter <adrian.hunter@intel.com>
Cc: linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org,
Perry Taylor <perry.taylor@intel.com>,
Samantha Alt <samantha.alt@intel.com>,
Caleb Biggers <caleb.biggers@intel.com>
Subject: [RFC PATCH v5 15/16] perf stat: use tool event helper function in metricgroup__build_event_string
Date: Fri, 12 Apr 2024 14:07:55 -0700 [thread overview]
Message-ID: <20240412210756.309828-16-weilin.wang@intel.com> (raw)
In-Reply-To: <20240412210756.309828-1-weilin.wang@intel.com>
From: Weilin Wang <weilin.wang@intel.com>
We want to include all used tool events in each event group to ensure
sharing events so that help improve multiplexing.
This change updates parse_id and metricgroupg__build_event_string to use
the get_tool_event_str helper function to generate tool event string
instead of inserting temporary tool event ids and generate string from
the tool event ids.
Signed-off-by: Weilin Wang <weilin.wang@intel.com>
---
tools/perf/util/metricgroup.c | 61 +++++++++++++----------------------
1 file changed, 22 insertions(+), 39 deletions(-)
diff --git a/tools/perf/util/metricgroup.c b/tools/perf/util/metricgroup.c
index 86b6528e5a44..39746d18f078 100644
--- a/tools/perf/util/metricgroup.c
+++ b/tools/perf/util/metricgroup.c
@@ -774,6 +774,18 @@ static int get_tool_event_str(struct strbuf *events,
int i = 0;
int ret;
+ /*
+ * We may fail to share events between metrics because a tool
+ * event isn't present in one metric. For example, a ratio of
+ * cache misses doesn't need duration_time but the same events
+ * may be used for a misses per second. Events without sharing
+ * implies multiplexing, that is best avoided, so place
+ * all tool events in every group.
+ * This function helps place all tool events in every group by
+ * generating the tool event strbuf that to be added in event
+ * group strings.
+ */
+
perf_tool_event__for_each_event(i) {
if (tool_events[i]) {
const char *tmp = strdup(perf_tool_event__to_str(i));
@@ -795,15 +807,18 @@ static int get_tool_event_str(struct strbuf *events,
static int metricgroup__build_event_string(struct strbuf *events,
const struct expr_parse_ctx *ctx,
const char *modifier,
- bool group_events)
+ bool group_events,
+ const bool tool_events[PERF_TOOL_MAX])
{
struct hashmap_entry *cur;
size_t bkt;
bool no_group = true, has_tool_events = false;
- bool tool_events[PERF_TOOL_MAX] = {false};
int ret = 0;
+ struct strbuf tool_event_str = STRBUF_INIT;
#define RETURN_IF_NON_ZERO(x) do { if (x) return x; } while (0)
+ ret = get_tool_event_str(&tool_event_str, tool_events, &has_tool_events);
+ RETURN_IF_NON_ZERO(ret);
hashmap__for_each_entry(ctx->ids, cur, bkt) {
const char *sep, *rsep, *id = cur->pkey;
@@ -814,8 +829,6 @@ static int metricgroup__build_event_string(struct strbuf *events,
/* Always move tool events outside of the group. */
ev = perf_tool_event__from_str(id);
if (ev != PERF_TOOL_NONE) {
- has_tool_events = true;
- tool_events[ev] = true;
continue;
}
/* Separate events with commas and open the group if necessary. */
@@ -879,19 +892,8 @@ static int metricgroup__build_event_string(struct strbuf *events,
RETURN_IF_NON_ZERO(ret);
}
if (has_tool_events) {
- int i;
-
- perf_tool_event__for_each_event(i) {
- if (tool_events[i]) {
- if (!no_group) {
- ret = strbuf_addch(events, ',');
- RETURN_IF_NON_ZERO(ret);
- }
- no_group = false;
- ret = strbuf_addstr(events, perf_tool_event__to_str(i));
- RETURN_IF_NON_ZERO(ret);
- }
- }
+ ret = strbuf_addstr(events, tool_event_str.buf);
+ RETURN_IF_NON_ZERO(ret);
}
return ret;
@@ -2421,32 +2423,13 @@ static int parse_ids(bool metric_no_merge, struct perf_pmu *fake_pmu,
*out_evlist = NULL;
if (!metric_no_merge || hashmap__size(ids->ids) == 0) {
- bool added_event = false;
- int i;
/*
- * We may fail to share events between metrics because a tool
- * event isn't present in one metric. For example, a ratio of
- * cache misses doesn't need duration_time but the same events
- * may be used for a misses per second. Events without sharing
- * implies multiplexing, that is best avoided, so place
- * all tool events in every group.
- *
- * Also, there may be no ids/events in the expression parsing
+ * There may be no ids/events in the expression parsing
* context because of constant evaluation, e.g.:
* event1 if #smt_on else 0
* Add a tool event to avoid a parse error on an empty string.
*/
- perf_tool_event__for_each_event(i) {
- if (tool_events[i]) {
- char *tmp = strdup(perf_tool_event__to_str(i));
-
- if (!tmp)
- return -ENOMEM;
- ids__insert(ids->ids, tmp);
- added_event = true;
- }
- }
- if (!added_event && hashmap__size(ids->ids) == 0) {
+ if (hashmap__size(ids->ids) == 0) {
char *tmp = strdup("duration_time");
if (!tmp)
@@ -2455,7 +2438,7 @@ static int parse_ids(bool metric_no_merge, struct perf_pmu *fake_pmu,
}
}
ret = metricgroup__build_event_string(&events, ids, modifier,
- group_events);
+ group_events, tool_events);
if (ret)
return ret;
--
2.42.0
next prev parent reply other threads:[~2024-04-12 21:08 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-12 21:07 [RFC PATCH v5 00/16] Perf stat metric grouping with hardware information weilin.wang
2024-04-12 21:07 ` [RFC PATCH v5 01/16] perf stat: Add new field in stat_config to enable hardware aware grouping weilin.wang
2024-04-17 3:49 ` Ian Rogers
2024-04-17 16:21 ` Arnaldo Carvalho de Melo
2024-04-12 21:07 ` [RFC PATCH v5 02/16] perf stat: Add basic functions for the " weilin.wang
2024-04-17 4:56 ` Ian Rogers
2024-04-17 16:24 ` Arnaldo Carvalho de Melo
2024-04-12 21:07 ` [RFC PATCH v5 03/16] perf pmu-events: Add functions in jevent.py to parse counter and event info for " weilin.wang
2024-04-17 5:42 ` Ian Rogers
2024-04-12 21:07 ` [RFC PATCH v5 04/16] find_bit: add _find_last_and_bit() to support finding the most significant set bit weilin.wang
2024-04-12 21:07 ` [RFC PATCH v5 05/16] perf stat: Add functions to set counter bitmaps for hardware-grouping method weilin.wang
2024-04-17 5:35 ` Ian Rogers
2024-04-12 21:07 ` [RFC PATCH v5 06/16] perf stat: Add functions to get counter info weilin.wang
2024-04-12 21:07 ` [RFC PATCH v5 07/16] perf stat: Add functions to create new group and assign events into groups weilin.wang
2024-04-12 21:07 ` [RFC PATCH v5 08/16] perf stat: Add build string function and topdown events handling in hardware-grouping weilin.wang
2024-04-17 5:56 ` Ian Rogers
2024-04-12 21:07 ` [RFC PATCH v5 09/16] perf stat: Add function to handle special events " weilin.wang
2024-04-17 6:12 ` Ian Rogers
2024-04-12 21:07 ` [RFC PATCH v5 10/16] perf stat: Add function to combine metrics for hardware-grouping weilin.wang
2024-04-12 21:07 ` [RFC PATCH v5 11/16] perf stat: Add partial support on MSR in hardware-grouping weilin.wang
2024-04-12 21:07 ` [RFC PATCH v5 12/16] perf stat: Handle NMI " weilin.wang
2024-04-12 21:07 ` [RFC PATCH v5 13/16] perf stat: Code refactoring " weilin.wang
2024-04-12 21:07 ` [RFC PATCH v5 14/16] perf stat: Add tool events support " weilin.wang
2024-04-12 21:07 ` weilin.wang [this message]
2024-04-17 6:36 ` [RFC PATCH v5 15/16] perf stat: use tool event helper function in metricgroup__build_event_string Ian Rogers
2024-04-12 21:07 ` [RFC PATCH v5 16/16] perf stat: Add hardware-grouping cmd option to perf stat weilin.wang
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=20240412210756.309828-16-weilin.wang@intel.com \
--to=weilin.wang@intel.com \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=caleb.biggers@intel.com \
--cc=irogers@google.com \
--cc=jolsa@kernel.org \
--cc=kan.liang@linux.intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=namhyung@kernel.org \
--cc=perry.taylor@intel.com \
--cc=peterz@infradead.org \
--cc=samantha.alt@intel.com \
/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;
as well as URLs for NNTP newsgroup(s).