perf expr: Add source_count for aggregating events

Events like uncore_imc/cas_count_read/ on Skylake open multiple events
and then aggregate in the metric leader. To determine the average value
per event the number of these events is needed. Add a source_count
function that returns this value by counting the number of events with
the given metric leader. For most events the value is 1 but for
uncore_imc/cas_count_read/ it can yield values like 6.

Add a generic test, but manually tested with a test metric that uses
the function.

Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul A . Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Riccardo Mancini <rickyman7@gmail.com>
Cc: Song Liu <song@kernel.org>
Cc: Wan Jiabing <wanjiabing@vivo.com>
Cc: Yury Norov <yury.norov@gmail.com>
Link: https://lore.kernel.org/r/20211111002109.194172-9-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
diff --git a/tools/perf/tests/expr.c b/tools/perf/tests/expr.c
index ed95ebd..c895de4 100644
--- a/tools/perf/tests/expr.c
+++ b/tools/perf/tests/expr.c
@@ -171,6 +171,18 @@ static int test__expr(struct test_suite *t __maybe_unused, int subtest __maybe_u
 	TEST_ASSERT_VAL("#num_packages", expr__parse(&num_packages, ctx, "#num_packages") == 0);
 	TEST_ASSERT_VAL("#num_dies >= #num_packages", num_dies >= num_packages);
 
+	/*
+	 * Source count returns the number of events aggregating in a leader
+	 * event including the leader. Check parsing yields an id.
+	 */
+	expr__ctx_clear(ctx);
+	TEST_ASSERT_VAL("source count",
+			expr__find_ids("source_count(EVENT1)",
+			NULL, ctx) == 0);
+	TEST_ASSERT_VAL("source count", hashmap__size(ctx->ids) == 1);
+	TEST_ASSERT_VAL("source count", hashmap__find(ctx->ids, "EVENT1",
+							(void **)&val_ptr));
+
 	expr__ctx_free(ctx);
 
 	return 0;
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index ec967fb..a59fb2e 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -3037,3 +3037,15 @@ void evsel__set_leader(struct evsel *evsel, struct evsel *leader)
 {
 	evsel->core.leader = &leader->core;
 }
+
+int evsel__source_count(const struct evsel *evsel)
+{
+	struct evsel *pos;
+	int count = 0;
+
+	evlist__for_each_entry(evsel->evlist, pos) {
+		if (pos->metric_leader == evsel)
+			count++;
+	}
+	return count;
+}
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index 3ea6871..29d49a8 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -489,6 +489,7 @@ struct evsel *evsel__leader(struct evsel *evsel);
 bool evsel__has_leader(struct evsel *evsel, struct evsel *leader);
 bool evsel__is_leader(struct evsel *evsel);
 void evsel__set_leader(struct evsel *evsel, struct evsel *leader);
+int evsel__source_count(const struct evsel *evsel);
 
 /*
  * Macro to swap the bit-field postition and size.
diff --git a/tools/perf/util/expr.c b/tools/perf/util/expr.c
index 15af8b8..1d532b9 100644
--- a/tools/perf/util/expr.c
+++ b/tools/perf/util/expr.c
@@ -23,7 +23,10 @@ extern int expr_debug;
 
 struct expr_id_data {
 	union {
-		double val;
+		struct {
+			double val;
+			int source_count;
+		} val;
 		struct {
 			double val;
 			const char *metric_name;
@@ -141,6 +144,13 @@ int expr__add_id(struct expr_parse_ctx *ctx, const char *id)
 /* Caller must make sure id is allocated */
 int expr__add_id_val(struct expr_parse_ctx *ctx, const char *id, double val)
 {
+	return expr__add_id_val_source_count(ctx, id, val, /*source_count=*/1);
+}
+
+/* Caller must make sure id is allocated */
+int expr__add_id_val_source_count(struct expr_parse_ctx *ctx, const char *id,
+				  double val, int source_count)
+{
 	struct expr_id_data *data_ptr = NULL, *old_data = NULL;
 	char *old_key = NULL;
 	int ret;
@@ -148,7 +158,8 @@ int expr__add_id_val(struct expr_parse_ctx *ctx, const char *id, double val)
 	data_ptr = malloc(sizeof(*data_ptr));
 	if (!data_ptr)
 		return -ENOMEM;
-	data_ptr->val = val;
+	data_ptr->val.val = val;
+	data_ptr->val.source_count = source_count;
 	data_ptr->kind = EXPR_ID_DATA__VALUE;
 
 	ret = hashmap__set(ctx->ids, id, data_ptr,
@@ -244,7 +255,7 @@ int expr__resolve_id(struct expr_parse_ctx *ctx, const char *id,
 
 	switch (data->kind) {
 	case EXPR_ID_DATA__VALUE:
-		pr_debug2("lookup(%s): val %f\n", id, data->val);
+		pr_debug2("lookup(%s): val %f\n", id, data->val.val);
 		break;
 	case EXPR_ID_DATA__REF:
 		pr_debug2("lookup(%s): ref metric name %s\n", id,
@@ -255,7 +266,7 @@ int expr__resolve_id(struct expr_parse_ctx *ctx, const char *id,
 			pr_debug("%s failed to count\n", id);
 			return -1;
 		}
-		pr_debug("processing metric: %s EXIT: %f\n", id, data->val);
+		pr_debug("processing metric: %s EXIT: %f\n", id, data->ref.val);
 		break;
 	case EXPR_ID_DATA__REF_VALUE:
 		pr_debug2("lookup(%s): ref val %f metric name %s\n", id,
@@ -370,11 +381,17 @@ int expr__find_ids(const char *expr, const char *one,
 double expr_id_data__value(const struct expr_id_data *data)
 {
 	if (data->kind == EXPR_ID_DATA__VALUE)
-		return data->val;
+		return data->val.val;
 	assert(data->kind == EXPR_ID_DATA__REF_VALUE);
 	return data->ref.val;
 }
 
+double expr_id_data__source_count(const struct expr_id_data *data)
+{
+	assert(data->kind == EXPR_ID_DATA__VALUE);
+	return data->val.source_count;
+}
+
 double expr__get_literal(const char *literal)
 {
 	static struct cpu_topology *topology;
diff --git a/tools/perf/util/expr.h b/tools/perf/util/expr.h
index a6ab7f2..bd21169 100644
--- a/tools/perf/util/expr.h
+++ b/tools/perf/util/expr.h
@@ -40,6 +40,8 @@ void expr__ctx_free(struct expr_parse_ctx *ctx);
 void expr__del_id(struct expr_parse_ctx *ctx, const char *id);
 int expr__add_id(struct expr_parse_ctx *ctx, const char *id);
 int expr__add_id_val(struct expr_parse_ctx *ctx, const char *id, double val);
+int expr__add_id_val_source_count(struct expr_parse_ctx *ctx, const char *id,
+				double val, int source_count);
 int expr__add_ref(struct expr_parse_ctx *ctx, struct metric_ref *ref);
 int expr__get_id(struct expr_parse_ctx *ctx, const char *id,
 		 struct expr_id_data **data);
@@ -55,6 +57,7 @@ int expr__find_ids(const char *expr, const char *one,
 		   struct expr_parse_ctx *ids);
 
 double expr_id_data__value(const struct expr_id_data *data);
+double expr_id_data__source_count(const struct expr_id_data *data);
 double expr__get_literal(const char *literal);
 
 #endif
diff --git a/tools/perf/util/expr.l b/tools/perf/util/expr.l
index cf6e3c7..0a13eb2 100644
--- a/tools/perf/util/expr.l
+++ b/tools/perf/util/expr.l
@@ -107,6 +107,7 @@
 min		{ return MIN; }
 if		{ return IF; }
 else		{ return ELSE; }
+source_count	{ return SOURCE_COUNT; }
 {literal}	{ return literal(yyscanner); }
 {number}	{ return value(yyscanner); }
 {symbol}	{ return str(yyscanner, ID, sctx->runtime); }
diff --git a/tools/perf/util/expr.y b/tools/perf/util/expr.y
index 1ec9c9b..a30b825 100644
--- a/tools/perf/util/expr.y
+++ b/tools/perf/util/expr.y
@@ -37,7 +37,7 @@
 	} ids;
 }
 
-%token ID NUMBER MIN MAX IF ELSE LITERAL D_RATIO EXPR_ERROR
+%token ID NUMBER MIN MAX IF ELSE LITERAL D_RATIO SOURCE_COUNT EXPR_ERROR
 %left MIN MAX IF
 %left '|'
 %left '^'
@@ -84,7 +84,7 @@
 }
 
 static struct ids handle_id(struct expr_parse_ctx *ctx, char *id,
-			    bool compute_ids)
+			    bool compute_ids, bool source_count)
 {
 	struct ids result;
 
@@ -96,9 +96,11 @@
 		struct expr_id_data *data;
 
 		result.val = NAN;
-		if (expr__resolve_id(ctx, id, &data) == 0)
-			result.val = expr_id_data__value(data);
-
+		if (expr__resolve_id(ctx, id, &data) == 0) {
+			result.val = source_count
+				? expr_id_data__source_count(data)
+				: expr_id_data__value(data);
+		}
 		result.ids = NULL;
 		free(id);
 	} else {
@@ -201,7 +203,8 @@
 	$$.val = $1;
 	$$.ids = NULL;
 }
-| ID		{ $$ = handle_id(ctx, $1, compute_ids); }
+| ID				{ $$ = handle_id(ctx, $1, compute_ids, /*source_count=*/false); }
+| SOURCE_COUNT '(' ID ')'	{ $$ = handle_id(ctx, $3, compute_ids, /*source_count=*/true); }
 | expr '|' expr { BINARY_LONG_OP($$, |, $1, $3); }
 | expr '&' expr { BINARY_LONG_OP($$, &, $1, $3); }
 | expr '^' expr { BINARY_LONG_OP($$, ^, $1, $3); }
diff --git a/tools/perf/util/stat-shadow.c b/tools/perf/util/stat-shadow.c
index e4fb02b..5c7308e 100644
--- a/tools/perf/util/stat-shadow.c
+++ b/tools/perf/util/stat-shadow.c
@@ -829,10 +829,12 @@ static int prepare_metric(struct evsel **metric_events,
 		struct saved_value *v;
 		struct stats *stats;
 		u64 metric_total = 0;
+		int source_count;
 
 		if (!strcmp(metric_events[i]->name, "duration_time")) {
 			stats = &walltime_nsecs_stats;
 			scale = 1e-9;
+			source_count = 1;
 		} else {
 			v = saved_value_lookup(metric_events[i], cpu, false,
 					       STAT_NONE, 0, st,
@@ -841,6 +843,7 @@ static int prepare_metric(struct evsel **metric_events,
 				break;
 			stats = &v->stats;
 			scale = 1.0;
+			source_count = evsel__source_count(metric_events[i]);
 
 			if (v->metric_other)
 				metric_total = v->metric_total;
@@ -849,7 +852,9 @@ static int prepare_metric(struct evsel **metric_events,
 		if (!n)
 			return -ENOMEM;
 
-		expr__add_id_val(pctx, n, metric_total ? : avg_stats(stats) * scale);
+		expr__add_id_val_source_count(pctx, n,
+					metric_total ? : avg_stats(stats) * scale,
+					source_count);
 	}
 
 	for (j = 0; metric_refs && metric_refs[j].metric_name; j++) {