1 files changed, 709 insertions, 805 deletions

diff --git a/tools/perf/util/stat-display.c b/tools/perf/util/stat-display.c
index ba66bb7fc1ca..8bd8b0142630 100644
--- a/tools/perf/util/stat-display.c
+++ b/tools/perf/util/stat-display.c
@@ -25,41 +25,124 @@
 #define CNTR_NOT_SUPPORTED	"<not supported>"
 #define CNTR_NOT_COUNTED	"<not counted>"
 
-static void print_running(struct perf_stat_config *config,
-			  u64 run, u64 ena)
+#define METRIC_LEN   38
+#define EVNAME_LEN   32
+#define COUNTS_LEN   18
+#define INTERVAL_LEN 16
+#define CGROUP_LEN   16
+#define COMM_LEN     16
+#define PID_LEN       7
+#define CPUS_LEN      4
+
+static int aggr_header_lens[] = {
+	[AGGR_CORE] 	= 18,
+	[AGGR_DIE] 	= 12,
+	[AGGR_SOCKET] 	= 6,
+	[AGGR_NODE] 	= 6,
+	[AGGR_NONE] 	= 6,
+	[AGGR_THREAD] 	= 16,
+	[AGGR_GLOBAL] 	= 0,
+};
+
+static const char *aggr_header_csv[] = {
+	[AGGR_CORE] 	= 	"core,cpus,",
+	[AGGR_DIE] 	= 	"die,cpus,",
+	[AGGR_SOCKET] 	= 	"socket,cpus,",
+	[AGGR_NONE] 	= 	"cpu,",
+	[AGGR_THREAD] 	= 	"comm-pid,",
+	[AGGR_NODE] 	= 	"node,",
+	[AGGR_GLOBAL] 	=	""
+};
+
+static const char *aggr_header_std[] = {
+	[AGGR_CORE] 	= 	"core",
+	[AGGR_DIE] 	= 	"die",
+	[AGGR_SOCKET] 	= 	"socket",
+	[AGGR_NONE] 	= 	"cpu",
+	[AGGR_THREAD] 	= 	"comm-pid",
+	[AGGR_NODE] 	= 	"node",
+	[AGGR_GLOBAL] 	=	""
+};
+
+static void print_running_std(struct perf_stat_config *config, u64 run, u64 ena)
 {
+	if (run != ena)
+		fprintf(config->output, "  (%.2f%%)", 100.0 * run / ena);
+}
 
+static void print_running_csv(struct perf_stat_config *config, u64 run, u64 ena)
+{
 	double enabled_percent = 100;
 
 	if (run != ena)
 		enabled_percent = 100 * run / ena;
-	if (config->json_output)
-		fprintf(config->output,
-			"\"event-runtime\" : %" PRIu64 ", \"pcnt-running\" : %.2f, ",
-			run, enabled_percent);
-	else if (config->csv_output)
-		fprintf(config->output,
-			"%s%" PRIu64 "%s%.2f", config->csv_sep,
-			run, config->csv_sep, enabled_percent);
-	else if (run != ena)
-		fprintf(config->output, "  (%.2f%%)", 100.0 * run / ena);
+	fprintf(config->output, "%s%" PRIu64 "%s%.2f",
+		config->csv_sep, run, config->csv_sep, enabled_percent);
+}
+
+static void print_running_json(struct perf_stat_config *config, u64 run, u64 ena)
+{
+	double enabled_percent = 100;
+
+	if (run != ena)
+		enabled_percent = 100 * run / ena;
+	fprintf(config->output, "\"event-runtime\" : %" PRIu64 ", \"pcnt-running\" : %.2f, ",
+		run, enabled_percent);
+}
+
+static void print_running(struct perf_stat_config *config,
+			  u64 run, u64 ena, bool before_metric)
+{
+	if (config->json_output) {
+		if (before_metric)
+			print_running_json(config, run, ena);
+	} else if (config->csv_output) {
+		if (before_metric)
+			print_running_csv(config, run, ena);
+	} else {
+		if (!before_metric)
+			print_running_std(config, run, ena);
+	}
+}
+
+static void print_noise_pct_std(struct perf_stat_config *config,
+				double pct)
+{
+	if (pct)
+		fprintf(config->output, "  ( +-%6.2f%% )", pct);
+}
+
+static void print_noise_pct_csv(struct perf_stat_config *config,
+				double pct)
+{
+	fprintf(config->output, "%s%.2f%%", config->csv_sep, pct);
+}
+
+static void print_noise_pct_json(struct perf_stat_config *config,
+				 double pct)
+{
+	fprintf(config->output, "\"variance\" : %.2f, ", pct);
 }
 
 static void print_noise_pct(struct perf_stat_config *config,
-			    double total, double avg)
+			    double total, double avg, bool before_metric)
 {
 	double pct = rel_stddev_stats(total, avg);
 
-	if (config->json_output)
-		fprintf(config->output, "\"variance\" : %.2f, ", pct);
-	else if (config->csv_output)
-		fprintf(config->output, "%s%.2f%%", config->csv_sep, pct);
-	else if (pct)
-		fprintf(config->output, "  ( +-%6.2f%% )", pct);
+	if (config->json_output) {
+		if (before_metric)
+			print_noise_pct_json(config, pct);
+	} else if (config->csv_output) {
+		if (before_metric)
+			print_noise_pct_csv(config, pct);
+	} else {
+		if (!before_metric)
+			print_noise_pct_std(config, pct);
+	}
 }
 
 static void print_noise(struct perf_stat_config *config,
-			struct evsel *evsel, double avg)
+			struct evsel *evsel, double avg, bool before_metric)
 {
 	struct perf_stat_evsel *ps;
 
@@ -67,139 +150,166 @@ static void print_noise(struct perf_stat_config *config,
 		return;
 
 	ps = evsel->stats;
-	print_noise_pct(config, stddev_stats(&ps->res_stats), avg);
+	print_noise_pct(config, stddev_stats(&ps->res_stats), avg, before_metric);
 }
 
-static void print_cgroup(struct perf_stat_config *config, struct evsel *evsel)
+static void print_cgroup_std(struct perf_stat_config *config, const char *cgrp_name)
 {
-	if (nr_cgroups) {
-		const char *cgrp_name = evsel->cgrp ? evsel->cgrp->name  : "";
+	fprintf(config->output, " %-*s", CGROUP_LEN, cgrp_name);
+}
+
+static void print_cgroup_csv(struct perf_stat_config *config, const char *cgrp_name)
+{
+	fprintf(config->output, "%s%s", config->csv_sep, cgrp_name);
+}
+
+static void print_cgroup_json(struct perf_stat_config *config, const char *cgrp_name)
+{
+	fprintf(config->output, "\"cgroup\" : \"%s\", ", cgrp_name);
+}
+
+static void print_cgroup(struct perf_stat_config *config, struct cgroup *cgrp)
+{
+	if (nr_cgroups || config->cgroup_list) {
+		const char *cgrp_name = cgrp ? cgrp->name  : "";
 
 		if (config->json_output)
-			fprintf(config->output, "\"cgroup\" : \"%s\", ", cgrp_name);
+			print_cgroup_json(config, cgrp_name);
+		else if (config->csv_output)
+			print_cgroup_csv(config, cgrp_name);
 		else
-			fprintf(config->output, "%s%s", config->csv_sep, cgrp_name);
+			print_cgroup_std(config, cgrp_name);
 	}
 }
 
-
-static void aggr_printout(struct perf_stat_config *config,
-			  struct evsel *evsel, struct aggr_cpu_id id, int nr)
+static void print_aggr_id_std(struct perf_stat_config *config,
+			      struct evsel *evsel, struct aggr_cpu_id id, int nr)
 {
+	FILE *output = config->output;
+	int idx = config->aggr_mode;
+	char buf[128];
+
+	switch (config->aggr_mode) {
+	case AGGR_CORE:
+		snprintf(buf, sizeof(buf), "S%d-D%d-C%d", id.socket, id.die, id.core);
+		break;
+	case AGGR_DIE:
+		snprintf(buf, sizeof(buf), "S%d-D%d", id.socket, id.die);
+		break;
+	case AGGR_SOCKET:
+		snprintf(buf, sizeof(buf), "S%d", id.socket);
+		break;
+	case AGGR_NODE:
+		snprintf(buf, sizeof(buf), "N%d", id.node);
+		break;
+	case AGGR_NONE:
+		if (evsel->percore && !config->percore_show_thread) {
+			snprintf(buf, sizeof(buf), "S%d-D%d-C%d ",
+				id.socket, id.die, id.core);
+			fprintf(output, "%-*s ",
+				aggr_header_lens[AGGR_CORE], buf);
+		} else if (id.cpu.cpu > -1) {
+			fprintf(output, "CPU%-*d ",
+				aggr_header_lens[AGGR_NONE] - 3, id.cpu.cpu);
+		}
+		return;
+	case AGGR_THREAD:
+		fprintf(output, "%*s-%-*d ",
+			COMM_LEN, perf_thread_map__comm(evsel->core.threads, id.thread_idx),
+			PID_LEN, perf_thread_map__pid(evsel->core.threads, id.thread_idx));
+		return;
+	case AGGR_GLOBAL:
+	case AGGR_UNSET:
+	case AGGR_MAX:
+	default:
+		return;
+	}
 
+	fprintf(output, "%-*s %*d ", aggr_header_lens[idx], buf, 4, nr);
+}
 
-	if (config->json_output && !config->interval)
-		fprintf(config->output, "{");
+static void print_aggr_id_csv(struct perf_stat_config *config,
+			      struct evsel *evsel, struct aggr_cpu_id id, int nr)
+{
+	FILE *output = config->output;
+	const char *sep = config->csv_sep;
 
 	switch (config->aggr_mode) {
 	case AGGR_CORE:
-		if (config->json_output) {
-			fprintf(config->output,
-				"\"core\" : \"S%d-D%d-C%d\", \"aggregate-number\" : %d, ",
-				id.socket,
-				id.die,
-				id.core,
-				nr);
-		} else {
-			fprintf(config->output, "S%d-D%d-C%*d%s%*d%s",
-				id.socket,
-				id.die,
-				config->csv_output ? 0 : -8,
-				id.core,
-				config->csv_sep,
-				config->csv_output ? 0 : 4,
-				nr,
-				config->csv_sep);
-		}
+		fprintf(output, "S%d-D%d-C%d%s%d%s",
+			id.socket, id.die, id.core, sep, nr, sep);
 		break;
 	case AGGR_DIE:
-		if (config->json_output) {
-			fprintf(config->output,
-				"\"die\" : \"S%d-D%d\", \"aggregate-number\" : %d, ",
-				id.socket,
-				id.die,
-				nr);
-		} else {
-			fprintf(config->output, "S%d-D%*d%s%*d%s",
-				id.socket,
-				config->csv_output ? 0 : -8,
-				id.die,
-				config->csv_sep,
-				config->csv_output ? 0 : 4,
-				nr,
-				config->csv_sep);
-		}
+		fprintf(output, "S%d-D%d%s%d%s",
+			id.socket, id.die, sep, nr, sep);
 		break;
 	case AGGR_SOCKET:
-		if (config->json_output) {
-			fprintf(config->output,
-				"\"socket\" : \"S%d\", \"aggregate-number\" : %d, ",
-				id.socket,
-				nr);
-		} else {
-			fprintf(config->output, "S%*d%s%*d%s",
-				config->csv_output ? 0 : -5,
-				id.socket,
-				config->csv_sep,
-				config->csv_output ? 0 : 4,
-				nr,
-				config->csv_sep);
-		}
+		fprintf(output, "S%d%s%d%s",
+			id.socket, sep, nr, sep);
 		break;
 	case AGGR_NODE:
-		if (config->json_output) {
-			fprintf(config->output, "\"node\" : \"N%d\", \"aggregate-number\" : %d, ",
-				id.node,
-				nr);
-		} else {
-			fprintf(config->output, "N%*d%s%*d%s",
-				config->csv_output ? 0 : -5,
-				id.node,
-				config->csv_sep,
-				config->csv_output ? 0 : 4,
-				nr,
-				config->csv_sep);
-		}
+		fprintf(output, "N%d%s%d%s",
+			id.node, sep, nr, sep);
 		break;
 	case AGGR_NONE:
-		if (config->json_output) {
-			if (evsel->percore && !config->percore_show_thread) {
-				fprintf(config->output, "\"core\" : \"S%d-D%d-C%d\"",
-					id.socket,
-					id.die,
-					id.core);
-			} else if (id.cpu.cpu > -1) {
-				fprintf(config->output, "\"cpu\" : \"%d\", ",
-					id.cpu.cpu);
-			}
-		} else {
-			if (evsel->percore && !config->percore_show_thread) {
-				fprintf(config->output, "S%d-D%d-C%*d%s",
-					id.socket,
-					id.die,
-					config->csv_output ? 0 : -3,
-					id.core, config->csv_sep);
-			} else if (id.cpu.cpu > -1) {
-				fprintf(config->output, "CPU%*d%s",
-					config->csv_output ? 0 : -7,
-					id.cpu.cpu, config->csv_sep);
-			}
+		if (evsel->percore && !config->percore_show_thread) {
+			fprintf(output, "S%d-D%d-C%d%s",
+				id.socket, id.die, id.core, sep);
+		} else if (id.cpu.cpu > -1) {
+			fprintf(output, "CPU%d%s",
+				id.cpu.cpu, sep);
 		}
 		break;
 	case AGGR_THREAD:
-		if (config->json_output) {
-			fprintf(config->output, "\"thread\" : \"%s-%d\", ",
-				perf_thread_map__comm(evsel->core.threads, id.thread_idx),
-				perf_thread_map__pid(evsel->core.threads, id.thread_idx));
-		} else {
-			fprintf(config->output, "%*s-%*d%s",
-				config->csv_output ? 0 : 16,
-				perf_thread_map__comm(evsel->core.threads, id.thread_idx),
-				config->csv_output ? 0 : -8,
-				perf_thread_map__pid(evsel->core.threads, id.thread_idx),
-				config->csv_sep);
+		fprintf(output, "%s-%d%s",
+			perf_thread_map__comm(evsel->core.threads, id.thread_idx),
+			perf_thread_map__pid(evsel->core.threads, id.thread_idx),
+			sep);
+		break;
+	case AGGR_GLOBAL:
+	case AGGR_UNSET:
+	case AGGR_MAX:
+	default:
+		break;
+	}
+}
+
+static void print_aggr_id_json(struct perf_stat_config *config,
+			       struct evsel *evsel, struct aggr_cpu_id id, int nr)
+{
+	FILE *output = config->output;
+
+	switch (config->aggr_mode) {
+	case AGGR_CORE:
+		fprintf(output, "\"core\" : \"S%d-D%d-C%d\", \"aggregate-number\" : %d, ",
+			id.socket, id.die, id.core, nr);
+		break;
+	case AGGR_DIE:
+		fprintf(output, "\"die\" : \"S%d-D%d\", \"aggregate-number\" : %d, ",
+			id.socket, id.die, nr);
+		break;
+	case AGGR_SOCKET:
+		fprintf(output, "\"socket\" : \"S%d\", \"aggregate-number\" : %d, ",
+			id.socket, nr);
+		break;
+	case AGGR_NODE:
+		fprintf(output, "\"node\" : \"N%d\", \"aggregate-number\" : %d, ",
+			id.node, nr);
+		break;
+	case AGGR_NONE:
+		if (evsel->percore && !config->percore_show_thread) {
+			fprintf(output, "\"core\" : \"S%d-D%d-C%d\"",
+				id.socket, id.die, id.core);
+		} else if (id.cpu.cpu > -1) {
+			fprintf(output, "\"cpu\" : \"%d\", ",
+				id.cpu.cpu);
 		}
 		break;
+	case AGGR_THREAD:
+		fprintf(output, "\"thread\" : \"%s-%d\", ",
+			perf_thread_map__comm(evsel->core.threads, id.thread_idx),
+			perf_thread_map__pid(evsel->core.threads, id.thread_idx));
+		break;
 	case AGGR_GLOBAL:
 	case AGGR_UNSET:
 	case AGGR_MAX:
@@ -208,18 +318,29 @@ static void aggr_printout(struct perf_stat_config *config,
 	}
 }
 
+static void aggr_printout(struct perf_stat_config *config,
+			  struct evsel *evsel, struct aggr_cpu_id id, int nr)
+{
+	if (config->json_output)
+		print_aggr_id_json(config, evsel, id, nr);
+	else if (config->csv_output)
+		print_aggr_id_csv(config, evsel, id, nr);
+	else
+		print_aggr_id_std(config, evsel, id, nr);
+}
+
 struct outstate {
 	FILE *fh;
 	bool newline;
+	bool first;
 	const char *prefix;
 	int  nfields;
 	int  nr;
 	struct aggr_cpu_id id;
 	struct evsel *evsel;
+	struct cgroup *cgrp;
 };
 
-#define METRIC_LEN  35
-
 static void new_line_std(struct perf_stat_config *config __maybe_unused,
 			 void *ctx)
 {
@@ -232,7 +353,8 @@ static void do_new_line_std(struct perf_stat_config *config,
 			    struct outstate *os)
 {
 	fputc('\n', os->fh);
-	fputs(os->prefix, os->fh);
+	if (os->prefix)
+		fputs(os->prefix, os->fh);
 	aggr_printout(config, os->evsel, os->id, os->nr);
 	if (config->aggr_mode == AGGR_NONE)
 		fprintf(os->fh, "        ");
@@ -319,7 +441,7 @@ static void new_line_json(struct perf_stat_config *config, void *ctx)
 {
 	struct outstate *os = ctx;
 
-	fputc('\n', os->fh);
+	fputs("\n{", os->fh);
 	if (os->prefix)
 		fprintf(os->fh, "%s", os->prefix);
 	aggr_printout(config, os->evsel, os->id, os->nr);
@@ -368,6 +490,7 @@ static void print_metric_only(struct perf_stat_config *config,
 
 	color_snprintf(str, sizeof(str), color ?: "", fmt, val);
 	fprintf(out, "%*s ", mlen, str);
+	os->first = false;
 }
 
 static void print_metric_only_csv(struct perf_stat_config *config __maybe_unused,
@@ -389,6 +512,7 @@ static void print_metric_only_csv(struct perf_stat_config *config __maybe_unused
 		ends++;
 	*ends = 0;
 	fprintf(out, "%s%s", vals, config->csv_sep);
+	os->first = false;
 }
 
 static void print_metric_only_json(struct perf_stat_config *config __maybe_unused,
@@ -409,7 +533,10 @@ static void print_metric_only_json(struct perf_stat_config *config __maybe_unuse
 	while (isdigit(*ends) || *ends == '.')
 		ends++;
 	*ends = 0;
-	fprintf(out, "{\"metric-value\" : \"%s\"}", vals);
+	if (!unit[0] || !vals[0])
+		return;
+	fprintf(out, "%s\"%s\" : \"%s\"", os->first ? "" : ", ", unit, vals);
+	os->first = false;
 }
 
 static void new_line_metric(struct perf_stat_config *config __maybe_unused,
@@ -430,84 +557,100 @@ static void print_metric_header(struct perf_stat_config *config,
 	    os->evsel->priv != os->evsel->evlist->selected->priv)
 		return;
 
-	if (!valid_only_metric(unit) && !config->json_output)
+	if (os->evsel->cgrp != os->cgrp)
+		return;
+
+	if (!valid_only_metric(unit))
 		return;
 	unit = fixunit(tbuf, os->evsel, unit);
 
 	if (config->json_output)
-		fprintf(os->fh, "\"unit\" : \"%s\"", unit);
+		return;
 	else if (config->csv_output)
 		fprintf(os->fh, "%s%s", unit, config->csv_sep);
 	else
 		fprintf(os->fh, "%*s ", config->metric_only_len, unit);
 }
 
-static int first_shadow_map_idx(struct perf_stat_config *config,
-				struct evsel *evsel, const struct aggr_cpu_id *id)
+static void print_counter_value_std(struct perf_stat_config *config,
+				    struct evsel *evsel, double avg, bool ok)
 {
-	struct perf_cpu_map *cpus = evsel__cpus(evsel);
-	struct perf_cpu cpu;
-	int idx;
-
-	if (config->aggr_mode == AGGR_NONE)
-		return perf_cpu_map__idx(cpus, id->cpu);
+	FILE *output = config->output;
+	double sc =  evsel->scale;
+	const char *fmt;
+	const char *bad_count = evsel->supported ? CNTR_NOT_COUNTED : CNTR_NOT_SUPPORTED;
 
-	if (config->aggr_mode == AGGR_THREAD)
-		return id->thread_idx;
+	if (config->big_num)
+		fmt = floor(sc) != sc ? "%'*.2f " : "%'*.0f ";
+	else
+		fmt = floor(sc) != sc ? "%*.2f " : "%*.0f ";
 
-	if (!config->aggr_get_id)
-		return 0;
+	if (ok)
+		fprintf(output, fmt, COUNTS_LEN, avg);
+	else
+		fprintf(output, "%*s ", COUNTS_LEN, bad_count);
 
-	perf_cpu_map__for_each_cpu(cpu, idx, cpus) {
-		struct aggr_cpu_id cpu_id = config->aggr_get_id(config, cpu);
+	if (evsel->unit)
+		fprintf(output, "%-*s ", config->unit_width, evsel->unit);
 
-		if (aggr_cpu_id__equal(&cpu_id, id))
-			return idx;
-	}
-	return 0;
+	fprintf(output, "%-*s", EVNAME_LEN, evsel__name(evsel));
 }
 
-static void abs_printout(struct perf_stat_config *config,
-			 struct aggr_cpu_id id, int nr, struct evsel *evsel, double avg)
+static void print_counter_value_csv(struct perf_stat_config *config,
+				    struct evsel *evsel, double avg, bool ok)
 {
 	FILE *output = config->output;
 	double sc =  evsel->scale;
-	const char *fmt;
+	const char *sep = config->csv_sep;
+	const char *fmt = floor(sc) != sc ? "%.2f%s" : "%.0f%s";
+	const char *bad_count = evsel->supported ? CNTR_NOT_COUNTED : CNTR_NOT_SUPPORTED;
 
-	if (config->csv_output) {
-		fmt = floor(sc) != sc ?  "%.2f%s" : "%.0f%s";
-	} else {
-		if (config->big_num)
-			fmt = floor(sc) != sc ? "%'18.2f%s" : "%'18.0f%s";
-		else
-			fmt = floor(sc) != sc ? "%18.2f%s" : "%18.0f%s";
-	}
+	if (ok)
+		fprintf(output, fmt, avg, sep);
+	else
+		fprintf(output, "%s%s", bad_count, sep);
 
-	aggr_printout(config, evsel, id, nr);
+	if (evsel->unit)
+		fprintf(output, "%s%s", evsel->unit, sep);
 
-	if (config->json_output)
+	fprintf(output, "%s", evsel__name(evsel));
+}
+
+static void print_counter_value_json(struct perf_stat_config *config,
+				     struct evsel *evsel, double avg, bool ok)
+{
+	FILE *output = config->output;
+	const char *bad_count = evsel->supported ? CNTR_NOT_COUNTED : CNTR_NOT_SUPPORTED;
+
+	if (ok)
 		fprintf(output, "\"counter-value\" : \"%f\", ", avg);
 	else
-		fprintf(output, fmt, avg, config->csv_sep);
+		fprintf(output, "\"counter-value\" : \"%s\", ", bad_count);
 
-	if (config->json_output) {
-		if (evsel->unit) {
-			fprintf(output, "\"unit\" : \"%s\", ",
-				evsel->unit);
-		}
-	} else {
-		if (evsel->unit)
-			fprintf(output, "%-*s%s",
-				config->csv_output ? 0 : config->unit_width,
-				evsel->unit, config->csv_sep);
-	}
+	if (evsel->unit)
+		fprintf(output, "\"unit\" : \"%s\", ", evsel->unit);
 
+	fprintf(output, "\"event\" : \"%s\", ", evsel__name(evsel));
+}
+
+static void print_counter_value(struct perf_stat_config *config,
+				struct evsel *evsel, double avg, bool ok)
+{
 	if (config->json_output)
-		fprintf(output, "\"event\" : \"%s\", ", evsel__name(evsel));
+		print_counter_value_json(config, evsel, avg, ok);
+	else if (config->csv_output)
+		print_counter_value_csv(config, evsel, avg, ok);
 	else
-		fprintf(output, "%-*s", config->csv_output ? 0 : 32, evsel__name(evsel));
+		print_counter_value_std(config, evsel, avg, ok);
+}
 
-	print_cgroup(config, evsel);
+static void abs_printout(struct perf_stat_config *config,
+			 struct aggr_cpu_id id, int nr,
+			 struct evsel *evsel, double avg, bool ok)
+{
+	aggr_printout(config, evsel, id, nr);
+	print_counter_value(config, evsel, avg, ok);
+	print_cgroup(config, evsel->cgrp);
 }
 
 static bool is_mixed_hw_group(struct evsel *counter)
@@ -534,37 +677,19 @@ static bool is_mixed_hw_group(struct evsel *counter)
 	return false;
 }
 
-static void printout(struct perf_stat_config *config, struct aggr_cpu_id id, int nr,
-		     struct evsel *counter, double uval,
-		     char *prefix, u64 run, u64 ena, double noise,
-		     struct runtime_stat *st)
+static void printout(struct perf_stat_config *config, struct outstate *os,
+		     double uval, u64 run, u64 ena, double noise, int map_idx)
 {
 	struct perf_stat_output_ctx out;
-	struct outstate os = {
-		.fh = config->output,
-		.prefix = prefix ? prefix : "",
-		.id = id,
-		.nr = nr,
-		.evsel = counter,
-	};
 	print_metric_t pm;
 	new_line_t nl;
+	bool ok = true;
+	struct evsel *counter = os->evsel;
 
 	if (config->csv_output) {
-		static const int aggr_fields[AGGR_MAX] = {
-			[AGGR_NONE] = 1,
-			[AGGR_GLOBAL] = 0,
-			[AGGR_SOCKET] = 2,
-			[AGGR_DIE] = 2,
-			[AGGR_CORE] = 2,
-			[AGGR_THREAD] = 1,
-			[AGGR_UNSET] = 0,
-			[AGGR_NODE] = 1,
-		};
-
 		pm = config->metric_only ? print_metric_only_csv : print_metric_csv;
 		nl = config->metric_only ? new_line_metric : new_line_csv;
-		os.nfields = 3 + aggr_fields[config->aggr_mode] + (counter->cgrp ? 1 : 0);
+		os->nfields = 4 + (counter->cgrp ? 1 : 0);
 	} else if (config->json_output) {
 		pm = config->metric_only ? print_metric_only_json : print_metric_json;
 		nl = config->metric_only ? new_line_metric : new_line_json;
@@ -573,27 +698,13 @@ static void printout(struct perf_stat_config *config, struct aggr_cpu_id id, int
 		nl = config->metric_only ? new_line_metric : new_line_std;
 	}
 
-	if (!config->no_csv_summary && config->csv_output &&
-	    config->summary && !config->interval) {
-		fprintf(config->output, "%16s%s", "summary", config->csv_sep);
-	}
-
 	if (run == 0 || ena == 0 || counter->counts->scaled == -1) {
 		if (config->metric_only) {
-			pm(config, &os, NULL, "", "", 0);
+			pm(config, os, NULL, "", "", 0);
 			return;
 		}
-		aggr_printout(config, counter, id, nr);
 
-		if (config->json_output) {
-			fprintf(config->output, "\"counter-value\" : \"%s\", ",
-					counter->supported ? CNTR_NOT_COUNTED : CNTR_NOT_SUPPORTED);
-		} else {
-			fprintf(config->output, "%*s%s",
-				config->csv_output ? 0 : 18,
-				counter->supported ? CNTR_NOT_COUNTED : CNTR_NOT_SUPPORTED,
-				config->csv_sep);
-		}
+		ok = false;
 
 		if (counter->supported) {
 			if (!evlist__has_hybrid(counter->evlist)) {
@@ -602,86 +713,30 @@ static void printout(struct perf_stat_config *config, struct aggr_cpu_id id, int
 					config->print_mixed_hw_group_error = 1;
 			}
 		}
-
-		if (config->json_output) {
-			fprintf(config->output, "\"unit\" : \"%s\", ", counter->unit);
-		} else {
-			fprintf(config->output, "%-*s%s",
-				config->csv_output ? 0 : config->unit_width,
-				counter->unit, config->csv_sep);
-		}
-
-		if (config->json_output) {
-			fprintf(config->output, "\"event\" : \"%s\", ",
-				evsel__name(counter));
-		} else {
-			fprintf(config->output, "%*s",
-				 config->csv_output ? 0 : -25, evsel__name(counter));
-		}
-
-		print_cgroup(config, counter);
-
-		if (!config->csv_output && !config->json_output)
-			pm(config, &os, NULL, NULL, "", 0);
-		print_noise(config, counter, noise);
-		print_running(config, run, ena);
-		if (config->csv_output)
-			pm(config, &os, NULL, NULL, "", 0);
-		else if (config->json_output)
-			pm(config, &os, NULL, NULL, "", 0);
-		return;
 	}
 
-	if (!config->metric_only)
-		abs_printout(config, id, nr, counter, uval);
-
 	out.print_metric = pm;
 	out.new_line = nl;
-	out.ctx = &os;
+	out.ctx = os;
 	out.force_header = false;
 
-	if (config->csv_output && !config->metric_only) {
-		print_noise(config, counter, noise);
-		print_running(config, run, ena);
-	} else if (config->json_output && !config->metric_only) {
-		print_noise(config, counter, noise);
-		print_running(config, run, ena);
-	}
+	if (!config->metric_only) {
+		abs_printout(config, os->id, os->nr, counter, uval, ok);
 
-	perf_stat__print_shadow_stats(config, counter, uval,
-				first_shadow_map_idx(config, counter, &id),
-				&out, &config->metric_events, st);
-	if (!config->csv_output && !config->metric_only && !config->json_output) {
-		print_noise(config, counter, noise);
-		print_running(config, run, ena);
+		print_noise(config, counter, noise, /*before_metric=*/true);
+		print_running(config, run, ena, /*before_metric=*/true);
 	}
-}
 
-static void aggr_update_shadow(struct perf_stat_config *config,
-			       struct evlist *evlist)
-{
-	int idx, s;
-	struct perf_cpu cpu;
-	struct aggr_cpu_id s2, id;
-	u64 val;
-	struct evsel *counter;
-	struct perf_cpu_map *cpus;
+	if (ok) {
+		perf_stat__print_shadow_stats(config, counter, uval, map_idx,
+					      &out, &config->metric_events, &rt_stat);
+	} else {
+		pm(config, os, /*color=*/NULL, /*format=*/NULL, /*unit=*/"", /*val=*/0);
+	}
 
-	for (s = 0; s < config->aggr_map->nr; s++) {
-		id = config->aggr_map->map[s];
-		evlist__for_each_entry(evlist, counter) {
-			cpus = evsel__cpus(counter);
-			val = 0;
-			perf_cpu_map__for_each_cpu(cpu, idx, cpus) {
-				s2 = config->aggr_get_id(config, cpu);
-				if (!aggr_cpu_id__equal(&s2, &id))
-					continue;
-				val += perf_counts(counter->counts, idx, 0)->val;
-			}
-			perf_stat__update_shadow_stats(counter, val,
-					first_shadow_map_idx(config, counter, &id),
-					&rt_stat);
-		}
+	if (!config->metric_only) {
+		print_noise(config, counter, noise, /*before_metric=*/false);
+		print_running(config, run, ena, /*before_metric=*/false);
 	}
 }
 
@@ -704,7 +759,7 @@ static void uniquify_event_name(struct evsel *counter)
 			counter->name = new_name;
 		}
 	} else {
-		if (perf_pmu__has_hybrid()) {
+		if (evsel__is_hybrid(counter)) {
 			ret = asprintf(&new_name, "%s/%s/",
 				       counter->pmu_name, counter->name);
 		} else {
@@ -721,366 +776,180 @@ static void uniquify_event_name(struct evsel *counter)
 	counter->uniquified_name = true;
 }
 
-static void collect_all_aliases(struct perf_stat_config *config, struct evsel *counter,
-			    void (*cb)(struct perf_stat_config *config, struct evsel *counter, void *data,
-				       bool first),
-			    void *data)
-{
-	struct evlist *evlist = counter->evlist;
-	struct evsel *alias;
-
-	alias = list_prepare_entry(counter, &(evlist->core.entries), core.node);
-	list_for_each_entry_continue (alias, &evlist->core.entries, core.node) {
-		/* Merge events with the same name, etc. but on different PMUs. */
-		if (!strcmp(evsel__name(alias), evsel__name(counter)) &&
-			alias->scale == counter->scale &&
-			alias->cgrp == counter->cgrp &&
-			!strcmp(alias->unit, counter->unit) &&
-			evsel__is_clock(alias) == evsel__is_clock(counter) &&
-			strcmp(alias->pmu_name, counter->pmu_name)) {
-			alias->merged_stat = true;
-			cb(config, alias, data, false);
-		}
-	}
-}
-
-static bool is_uncore(struct evsel *evsel)
+static bool hybrid_uniquify(struct evsel *evsel, struct perf_stat_config *config)
 {
-	struct perf_pmu *pmu = evsel__find_pmu(evsel);
-
-	return pmu && pmu->is_uncore;
+	return evsel__is_hybrid(evsel) && !config->hybrid_merge;
 }
 
-static bool hybrid_uniquify(struct evsel *evsel)
+static void uniquify_counter(struct perf_stat_config *config, struct evsel *counter)
 {
-	return perf_pmu__has_hybrid() && !is_uncore(evsel);
+	if (config->no_merge || hybrid_uniquify(counter, config))
+		uniquify_event_name(counter);
 }
 
-static bool hybrid_merge(struct evsel *counter, struct perf_stat_config *config,
-			 bool check)
+static void print_counter_aggrdata(struct perf_stat_config *config,
+				   struct evsel *counter, int s,
+				   struct outstate *os)
 {
-	if (hybrid_uniquify(counter)) {
-		if (check)
-			return config && config->hybrid_merge;
-		else
-			return config && !config->hybrid_merge;
-	}
+	FILE *output = config->output;
+	u64 ena, run, val;
+	double uval;
+	struct perf_stat_evsel *ps = counter->stats;
+	struct perf_stat_aggr *aggr = &ps->aggr[s];
+	struct aggr_cpu_id id = config->aggr_map->map[s];
+	double avg = aggr->counts.val;
+	bool metric_only = config->metric_only;
 
-	return false;
-}
+	os->id = id;
+	os->nr = aggr->nr;
+	os->evsel = counter;
 
-static bool collect_data(struct perf_stat_config *config, struct evsel *counter,
-			    void (*cb)(struct perf_stat_config *config, struct evsel *counter, void *data,
-				       bool first),
-			    void *data)
-{
+	/* Skip already merged uncore/hybrid events */
 	if (counter->merged_stat)
-		return false;
-	cb(config, counter, data, true);
-	if (config->no_merge || hybrid_merge(counter, config, false))
-		uniquify_event_name(counter);
-	else if (counter->auto_merge_stats || hybrid_merge(counter, config, true))
-		collect_all_aliases(config, counter, cb, data);
-	return true;
-}
+		return;
 
-struct aggr_data {
-	u64 ena, run, val;
-	struct aggr_cpu_id id;
-	int nr;
-	int cpu_map_idx;
-};
+	uniquify_counter(config, counter);
 
-static void aggr_cb(struct perf_stat_config *config,
-		    struct evsel *counter, void *data, bool first)
-{
-	struct aggr_data *ad = data;
-	int idx;
-	struct perf_cpu cpu;
-	struct perf_cpu_map *cpus;
-	struct aggr_cpu_id s2;
+	val = aggr->counts.val;
+	ena = aggr->counts.ena;
+	run = aggr->counts.run;
 
-	cpus = evsel__cpus(counter);
-	perf_cpu_map__for_each_cpu(cpu, idx, cpus) {
-		struct perf_counts_values *counts;
+	/*
+	 * Skip value 0 when enabling --per-thread globally, otherwise it will
+	 * have too many 0 output.
+	 */
+	if (val == 0 && config->aggr_mode == AGGR_THREAD && config->system_wide)
+		return;
 
-		s2 = config->aggr_get_id(config, cpu);
-		if (!aggr_cpu_id__equal(&s2, &ad->id))
-			continue;
-		if (first)
-			ad->nr++;
-		counts = perf_counts(counter->counts, idx, 0);
-		/*
-		 * When any result is bad, make them all to give
-		 * consistent output in interval mode.
-		 */
-		if (counts->ena == 0 || counts->run == 0 ||
-		    counter->counts->scaled == -1) {
-			ad->ena = 0;
-			ad->run = 0;
-			break;
-		}
-		ad->val += counts->val;
-		ad->ena += counts->ena;
-		ad->run += counts->run;
+	if (!metric_only) {
+		if (config->json_output)
+			fputc('{', output);
+		if (os->prefix)
+			fprintf(output, "%s", os->prefix);
+		else if (config->summary && config->csv_output &&
+			 !config->no_csv_summary && !config->interval)
+			fprintf(output, "%s%s", "summary", config->csv_sep);
 	}
+
+	uval = val * counter->scale;
+
+	printout(config, os, uval, run, ena, avg, s);
+
+	if (!metric_only)
+		fputc('\n', output);
 }
 
-static void print_counter_aggrdata(struct perf_stat_config *config,
-				   struct evsel *counter, int s,
-				   char *prefix, bool metric_only,
-				   bool *first, struct perf_cpu cpu)
+static void print_metric_begin(struct perf_stat_config *config,
+			       struct evlist *evlist,
+			       struct outstate *os, int aggr_idx)
 {
-	struct aggr_data ad;
-	FILE *output = config->output;
-	u64 ena, run, val;
-	int nr;
+	struct perf_stat_aggr *aggr;
 	struct aggr_cpu_id id;
-	double uval;
+	struct evsel *evsel;
 
-	ad.id = id = config->aggr_map->map[s];
-	ad.val = ad.ena = ad.run = 0;
-	ad.nr = 0;
-	if (!collect_data(config, counter, aggr_cb, &ad))
+	os->first = true;
+	if (!config->metric_only)
 		return;
 
-	if (perf_pmu__has_hybrid() && ad.ena == 0)
-		return;
+	if (config->json_output)
+		fputc('{', config->output);
+	if (os->prefix)
+		fprintf(config->output, "%s", os->prefix);
 
-	nr = ad.nr;
-	ena = ad.ena;
-	run = ad.run;
-	val = ad.val;
-	if (*first && metric_only) {
-		*first = false;
-		aggr_printout(config, counter, id, nr);
-	}
-	if (prefix && !metric_only)
-		fprintf(output, "%s", prefix);
+	evsel = evlist__first(evlist);
+	id = config->aggr_map->map[aggr_idx];
+	aggr = &evsel->stats->aggr[aggr_idx];
+	aggr_printout(config, evsel, id, aggr->nr);
 
-	uval = val * counter->scale;
-	if (cpu.cpu != -1)
-		id = aggr_cpu_id__cpu(cpu, /*data=*/NULL);
+	print_cgroup(config, os->cgrp ? : evsel->cgrp);
+}
 
-	printout(config, id, nr, counter, uval,
-		 prefix, run, ena, 1.0, &rt_stat);
-	if (!metric_only)
-		fputc('\n', output);
+static void print_metric_end(struct perf_stat_config *config, struct outstate *os)
+{
+	FILE *output = config->output;
+
+	if (!config->metric_only)
+		return;
+
+	if (config->json_output) {
+		if (os->first)
+			fputs("\"metric-value\" : \"none\"", output);
+		fputc('}', output);
+	}
+	fputc('\n', output);
 }
 
 static void print_aggr(struct perf_stat_config *config,
 		       struct evlist *evlist,
-		       char *prefix)
+		       struct outstate *os)
 {
-	bool metric_only = config->metric_only;
-	FILE *output = config->output;
 	struct evsel *counter;
 	int s;
-	bool first;
 
 	if (!config->aggr_map || !config->aggr_get_id)
 		return;
 
-	aggr_update_shadow(config, evlist);
-
 	/*
 	 * With metric_only everything is on a single line.
 	 * Without each counter has its own line.
 	 */
 	for (s = 0; s < config->aggr_map->nr; s++) {
-		if (prefix && metric_only)
-			fprintf(output, "%s", prefix);
+		print_metric_begin(config, evlist, os, s);
 
-		first = true;
 		evlist__for_each_entry(evlist, counter) {
-			print_counter_aggrdata(config, counter, s,
-					prefix, metric_only,
-					&first, (struct perf_cpu){ .cpu = -1 });
+			print_counter_aggrdata(config, counter, s, os);
 		}
-		if (metric_only)
-			fputc('\n', output);
+		print_metric_end(config, os);
 	}
 }
 
-static int cmp_val(const void *a, const void *b)
+static void print_aggr_cgroup(struct perf_