[PATCH v2 2/4] perf parse-events: Avoid scanning PMUs that can't match a wildcard

[Ian Rogers]

Add perf_pmus__scan_matching_wildcard that only reads sysfs for pmus
that could wildcard match. For example, if searching for the event
'inst_retired.any/period=1000000/' first 'inst_retired.any' will try
to be directly looked up as a PMU, then 'inst_retired.any' will try to
be wildcard matched against all PMUs, finally it is assumed
'inst_retired.any' is an event name and opened on all PMUs avoiding
reading PMUs that can't have such an event. However, before this patch
such a case would read all PMUs for the wildcard matching. Now if the
wildcard can't be a hwmon PMU wildcard, reading hwmon PMUs is
avoided. On my laptop this reduces 'perf stat -e inst_retired.any
true' from performing 253 openat system calls down to 234.

Signed-off-by: Ian Rogers <irogers@xxxxxxxxxx>
---
 tools/perf/util/parse-events.c | 26 +++++++++++----------
 tools/perf/util/pmus.c         | 42 ++++++++++++++++++++++++++++++++++
 tools/perf/util/pmus.h         |  1 +
 3 files changed, 57 insertions(+), 12 deletions(-)

diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
index d77af1d24985..19caf1ed9f5d 100644
--- a/tools/perf/util/parse-events.c
+++ b/tools/perf/util/parse-events.c
@@ -1664,18 +1664,20 @@ int parse_events_multi_pmu_add_or_add_pmu(struct parse_events_state *parse_state
 
 	pmu = NULL;
 	/* Failed to add, try wildcard expansion of event_or_pmu as a PMU name. */
-	while ((pmu = perf_pmus__scan(pmu)) != NULL) {
-		if (!parse_events__filter_pmu(parse_state, pmu) &&
-		    perf_pmu__wildcard_match(pmu, event_or_pmu)) {
-			bool auto_merge_stats = perf_pmu__auto_merge_stats(pmu);
-
-			if (!parse_events_add_pmu(parse_state, *listp, pmu,
-						  const_parsed_terms,
-						  auto_merge_stats,
-						  /*alternate_hw_config=*/PERF_COUNT_HW_MAX)) {
-				ok++;
-				parse_state->wild_card_pmus = true;
-			}
+	while ((pmu = perf_pmus__scan_matching_wildcard(pmu, event_or_pmu)) != NULL) {
+		bool auto_merge_stats;
+
+		if (parse_events__filter_pmu(parse_state, pmu))
+			continue;
+
+		auto_merge_stats = perf_pmu__auto_merge_stats(pmu);
+
+		if (!parse_events_add_pmu(parse_state, *listp, pmu,
+					  const_parsed_terms,
+					  auto_merge_stats,
+					  /*alternate_hw_config=*/PERF_COUNT_HW_MAX)) {
+			ok++;
+			parse_state->wild_card_pmus = true;
 		}
 	}
 	if (ok)
diff --git a/tools/perf/util/pmus.c b/tools/perf/util/pmus.c
index f2706c395509..93d5db1c89b4 100644
--- a/tools/perf/util/pmus.c
+++ b/tools/perf/util/pmus.c
@@ -19,6 +19,7 @@
 #include "tool_pmu.h"
 #include "print-events.h"
 #include "strbuf.h"
+#include "string2.h"
 
 /*
  * core_pmus:  A PMU belongs to core_pmus if it's name is "cpu" or it's sysfs
@@ -385,6 +386,47 @@ struct perf_pmu *perf_pmus__scan_for_event(struct perf_pmu *pmu, const char *eve
 	return NULL;
 }
 
+struct perf_pmu *perf_pmus__scan_matching_wildcard(struct perf_pmu *pmu, const char *wildcard)
+{
+	bool use_core_pmus = !pmu || pmu->is_core;
+
+	if (!pmu) {
+		/*
+		 * Core PMUs, other sysfs PMUs and tool PMU can have any name or
+		 * aren't wother optimizing for.
+		 */
+		unsigned int to_read_pmus =  PERF_TOOL_PMU_TYPE_PE_CORE_MASK |
+			PERF_TOOL_PMU_TYPE_PE_OTHER_MASK |
+			PERF_TOOL_PMU_TYPE_TOOL_MASK;
+
+		/*
+		 * Hwmon PMUs have an alias from a sysfs name like hwmon0,
+		 * hwmon1, etc. or have a name of hwmon_<name>. They therefore
+		 * can only have a wildcard match if the wildcard begins with
+		 * "hwmon".
+		 */
+		if (strisglob(wildcard) ||
+		    (strlen(wildcard) >= 5 && strncmp("hwmon", wildcard, 5) == 0))
+			to_read_pmus |= PERF_TOOL_PMU_TYPE_HWMON_MASK;
+
+		pmu_read_sysfs(to_read_pmus);
+		pmu = list_prepare_entry(pmu, &core_pmus, list);
+	}
+	if (use_core_pmus) {
+		list_for_each_entry_continue(pmu, &core_pmus, list) {
+			if (perf_pmu__wildcard_match(pmu, wildcard))
+				return pmu;
+		}
+		pmu = NULL;
+		pmu = list_prepare_entry(pmu, &other_pmus, list);
+	}
+	list_for_each_entry_continue(pmu, &other_pmus, list) {
+		if (perf_pmu__wildcard_match(pmu, wildcard))
+			return pmu;
+	}
+	return NULL;
+}
+
 static struct perf_pmu *perf_pmus__scan_skip_duplicates(struct perf_pmu *pmu)
 {
 	bool use_core_pmus = !pmu || pmu->is_core;
diff --git a/tools/perf/util/pmus.h b/tools/perf/util/pmus.h
index 213ee65306d6..2794d8c3a466 100644
--- a/tools/perf/util/pmus.h
+++ b/tools/perf/util/pmus.h
@@ -20,6 +20,7 @@ struct perf_pmu *perf_pmus__find_by_type(unsigned int type);
 struct perf_pmu *perf_pmus__scan(struct perf_pmu *pmu);
 struct perf_pmu *perf_pmus__scan_core(struct perf_pmu *pmu);
 struct perf_pmu *perf_pmus__scan_for_event(struct perf_pmu *pmu, const char *event);
+struct perf_pmu *perf_pmus__scan_matching_wildcard(struct perf_pmu *pmu, const char *wildcard);
 
 const struct perf_pmu *perf_pmus__pmu_for_pmu_filter(const char *str);
 
-- 
2.49.0.395.g12beb8f557-goog