1 files changed, 745 insertions, 195 deletions

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index c247faca712..65a151e3606 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -46,6 +46,7 @@
 #include "util/util.h"
 #include "util/parse-options.h"
 #include "util/parse-events.h"
+#include "util/pmu.h"
 #include "util/event.h"
 #include "util/evlist.h"
 #include "util/evsel.h"
@@ -65,19 +66,67 @@
 #define CNTR_NOT_SUPPORTED	"<not supported>"
 #define CNTR_NOT_COUNTED	"<not counted>"
 
+static void print_stat(int argc, const char **argv);
+static void print_counter_aggr(struct perf_evsel *counter, char *prefix);
+static void print_counter(struct perf_evsel *counter, char *prefix);
+static void print_aggr(char *prefix);
+
+/* Default events used for perf stat -T */
+static const char * const transaction_attrs[] = {
+	"task-clock",
+	"{"
+	"instructions,"
+	"cycles,"
+	"cpu/cycles-t/,"
+	"cpu/tx-start/,"
+	"cpu/el-start/,"
+	"cpu/cycles-ct/"
+	"}"
+};
+
+/* More limited version when the CPU does not have all events. */
+static const char * const transaction_limited_attrs[] = {
+	"task-clock",
+	"{"
+	"instructions,"
+	"cycles,"
+	"cpu/cycles-t/,"
+	"cpu/tx-start/"
+	"}"
+};
+
+/* must match transaction_attrs and the beginning limited_attrs */
+enum {
+	T_TASK_CLOCK,
+	T_INSTRUCTIONS,
+	T_CYCLES,
+	T_CYCLES_IN_TX,
+	T_TRANSACTION_START,
+	T_ELISION_START,
+	T_CYCLES_IN_TX_CP,
+};
+
 static struct perf_evlist	*evsel_list;
 
-static struct perf_target	target = {
+static struct target target = {
 	.uid	= UINT_MAX,
 };
 
+enum aggr_mode {
+	AGGR_NONE,
+	AGGR_GLOBAL,
+	AGGR_SOCKET,
+	AGGR_CORE,
+};
+
 static int			run_count			=  1;
 static bool			no_inherit			= false;
 static bool			scale				=  true;
-static bool			no_aggr				= false;
-static pid_t			child_pid			= -1;
+static enum aggr_mode		aggr_mode			= AGGR_GLOBAL;
+static volatile pid_t		child_pid			= -1;
 static bool			null_run			=  false;
 static int			detailed_run			=  0;
+static bool			transaction_run;
 static bool			big_num				=  true;
 static int			big_num_opt			=  -1;
 static const char		*csv_sep			= NULL;
@@ -87,6 +136,13 @@ static FILE			*output				= NULL;
 static const char		*pre_cmd			= NULL;
 static const char		*post_cmd			= NULL;
 static bool			sync_run			= false;
+static unsigned int		interval			= 0;
+static unsigned int		initial_delay			= 0;
+static unsigned int		unit_width			= 4; /* strlen("unit") */
+static bool			forever				= false;
+static struct timespec		ref_time;
+static struct cpu_map		*aggr_map;
+static int			(*aggr_get_id)(struct cpu_map *m, int cpu);
 
 static volatile int done = 0;
 
@@ -94,26 +150,100 @@ struct perf_stat {
 	struct stats	  res_stats[3];
 };
 
+static inline void diff_timespec(struct timespec *r, struct timespec *a,
+				 struct timespec *b)
+{
+	r->tv_sec = a->tv_sec - b->tv_sec;
+	if (a->tv_nsec < b->tv_nsec) {
+		r->tv_nsec = a->tv_nsec + 1000000000L - b->tv_nsec;
+		r->tv_sec--;
+	} else {
+		r->tv_nsec = a->tv_nsec - b->tv_nsec ;
+	}
+}
+
+static inline struct cpu_map *perf_evsel__cpus(struct perf_evsel *evsel)
+{
+	return (evsel->cpus && !target.cpu_list) ? evsel->cpus : evsel_list->cpus;
+}
+
+static inline int perf_evsel__nr_cpus(struct perf_evsel *evsel)
+{
+	return perf_evsel__cpus(evsel)->nr;
+}
+
+static void perf_evsel__reset_stat_priv(struct perf_evsel *evsel)
+{
+	int i;
+	struct perf_stat *ps = evsel->priv;
+
+	for (i = 0; i < 3; i++)
+		init_stats(&ps->res_stats[i]);
+}
+
 static int perf_evsel__alloc_stat_priv(struct perf_evsel *evsel)
 {
 	evsel->priv = zalloc(sizeof(struct perf_stat));
-	return evsel->priv == NULL ? -ENOMEM : 0;
+	if (evsel == NULL)
+		return -ENOMEM;
+	perf_evsel__reset_stat_priv(evsel);
+	return 0;
 }
 
 static void perf_evsel__free_stat_priv(struct perf_evsel *evsel)
 {
-	free(evsel->priv);
-	evsel->priv = NULL;
+	zfree(&evsel->priv);
 }
 
-static inline struct cpu_map *perf_evsel__cpus(struct perf_evsel *evsel)
+static int perf_evsel__alloc_prev_raw_counts(struct perf_evsel *evsel)
 {
-	return (evsel->cpus && !target.cpu_list) ? evsel->cpus : evsel_list->cpus;
+	void *addr;
+	size_t sz;
+
+	sz = sizeof(*evsel->counts) +
+	     (perf_evsel__nr_cpus(evsel) * sizeof(struct perf_counts_values));
+
+	addr = zalloc(sz);
+	if (!addr)
+		return -ENOMEM;
+
+	evsel->prev_raw_counts =  addr;
+
+	return 0;
 }
 
-static inline int perf_evsel__nr_cpus(struct perf_evsel *evsel)
+static void perf_evsel__free_prev_raw_counts(struct perf_evsel *evsel)
 {
-	return perf_evsel__cpus(evsel)->nr;
+	zfree(&evsel->prev_raw_counts);
+}
+
+static void perf_evlist__free_stats(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel;
+
+	evlist__for_each(evlist, evsel) {
+		perf_evsel__free_stat_priv(evsel);
+		perf_evsel__free_counts(evsel);
+		perf_evsel__free_prev_raw_counts(evsel);
+	}
+}
+
+static int perf_evlist__alloc_stats(struct perf_evlist *evlist, bool alloc_raw)
+{
+	struct perf_evsel *evsel;
+
+	evlist__for_each(evlist, evsel) {
+		if (perf_evsel__alloc_stat_priv(evsel) < 0 ||
+		    perf_evsel__alloc_counts(evsel, perf_evsel__nr_cpus(evsel)) < 0 ||
+		    (alloc_raw && perf_evsel__alloc_prev_raw_counts(evsel) < 0))
+			goto out_free;
+	}
+
+	return 0;
+
+out_free:
+	perf_evlist__free_stats(evlist);
+	return -1;
 }
 
 static struct stats runtime_nsecs_stats[MAX_NR_CPUS];
@@ -127,13 +257,42 @@ static struct stats runtime_l1_icache_stats[MAX_NR_CPUS];
 static struct stats runtime_ll_cache_stats[MAX_NR_CPUS];
 static struct stats runtime_itlb_cache_stats[MAX_NR_CPUS];
 static struct stats runtime_dtlb_cache_stats[MAX_NR_CPUS];
+static struct stats runtime_cycles_in_tx_stats[MAX_NR_CPUS];
 static struct stats walltime_nsecs_stats;
+static struct stats runtime_transaction_stats[MAX_NR_CPUS];
+static struct stats runtime_elision_stats[MAX_NR_CPUS];
+
+static void perf_stat__reset_stats(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel;
+
+	evlist__for_each(evlist, evsel) {
+		perf_evsel__reset_stat_priv(evsel);
+		perf_evsel__reset_counts(evsel, perf_evsel__nr_cpus(evsel));
+	}
+
+	memset(runtime_nsecs_stats, 0, sizeof(runtime_nsecs_stats));
+	memset(runtime_cycles_stats, 0, sizeof(runtime_cycles_stats));
+	memset(runtime_stalled_cycles_front_stats, 0, sizeof(runtime_stalled_cycles_front_stats));
+	memset(runtime_stalled_cycles_back_stats, 0, sizeof(runtime_stalled_cycles_back_stats));
+	memset(runtime_branches_stats, 0, sizeof(runtime_branches_stats));
+	memset(runtime_cacherefs_stats, 0, sizeof(runtime_cacherefs_stats));
+	memset(runtime_l1_dcache_stats, 0, sizeof(runtime_l1_dcache_stats));
+	memset(runtime_l1_icache_stats, 0, sizeof(runtime_l1_icache_stats));
+	memset(runtime_ll_cache_stats, 0, sizeof(runtime_ll_cache_stats));
+	memset(runtime_itlb_cache_stats, 0, sizeof(runtime_itlb_cache_stats));
+	memset(runtime_dtlb_cache_stats, 0, sizeof(runtime_dtlb_cache_stats));
+	memset(runtime_cycles_in_tx_stats, 0,
+			sizeof(runtime_cycles_in_tx_stats));
+	memset(runtime_transaction_stats, 0,
+		sizeof(runtime_transaction_stats));
+	memset(runtime_elision_stats, 0, sizeof(runtime_elision_stats));
+	memset(&walltime_nsecs_stats, 0, sizeof(walltime_nsecs_stats));
+}
 
 static int create_perf_stat_counter(struct perf_evsel *evsel)
 {
 	struct perf_event_attr *attr = &evsel->attr;
-	bool exclude_guest_missing = false;
-	int ret;
 
 	if (scale)
 		attr->read_format = PERF_FORMAT_TOTAL_TIME_ENABLED |
@@ -141,38 +300,16 @@ static int create_perf_stat_counter(struct perf_evsel *evsel)
 
 	attr->inherit = !no_inherit;
 
-retry:
-	if (exclude_guest_missing)
-		evsel->attr.exclude_guest = evsel->attr.exclude_host = 0;
-
-	if (perf_target__has_cpu(&target)) {
-		ret = perf_evsel__open_per_cpu(evsel, perf_evsel__cpus(evsel));
-		if (ret)
-			goto check_ret;
-		return 0;
-	}
+	if (target__has_cpu(&target))
+		return perf_evsel__open_per_cpu(evsel, perf_evsel__cpus(evsel));
 
-	if (!perf_target__has_task(&target) &&
-	    !perf_evsel__is_group_member(evsel)) {
+	if (!target__has_task(&target) && perf_evsel__is_group_leader(evsel)) {
 		attr->disabled = 1;
-		attr->enable_on_exec = 1;
+		if (!initial_delay)
+			attr->enable_on_exec = 1;
 	}
 
-	ret = perf_evsel__open_per_thread(evsel, evsel_list->threads);
-	if (!ret)
-		return 0;
-	/* fall through */
-check_ret:
-	if (ret && errno == EINVAL) {
-		if (!exclude_guest_missing &&
-		    (evsel->attr.exclude_guest || evsel->attr.exclude_host)) {
-			pr_debug("Old kernel, cannot exclude "
-				 "guest or host samples.\n");
-			exclude_guest_missing = true;
-			goto retry;
-		}
-	}
-	return ret;
+	return perf_evsel__open_per_thread(evsel, evsel_list->threads);
 }
 
 /*
@@ -187,6 +324,29 @@ static inline int nsec_counter(struct perf_evsel *evsel)
 	return 0;
 }
 
+static struct perf_evsel *nth_evsel(int n)
+{
+	static struct perf_evsel **array;
+	static int array_len;
+	struct perf_evsel *ev;
+	int j;
+
+	/* Assumes this only called when evsel_list does not change anymore. */
+	if (!array) {
+		evlist__for_each(evsel_list, ev)
+			array_len++;
+		array = malloc(array_len * sizeof(void *));
+		if (!array)
+			exit(ENOMEM);
+		j = 0;
+		evlist__for_each(evsel_list, ev)
+			array[j++] = ev;
+	}
+	if (n < array_len)
+		return array[n];
+	return NULL;
+}
+
 /*
  * Update various tracking values we maintain to print
  * more semantic information such as miss/hit ratios,
@@ -198,6 +358,15 @@ static void update_shadow_stats(struct perf_evsel *counter, u64 *count)
 		update_stats(&runtime_nsecs_stats[0], count[0]);
 	else if (perf_evsel__match(counter, HARDWARE, HW_CPU_CYCLES))
 		update_stats(&runtime_cycles_stats[0], count[0]);
+	else if (transaction_run &&
+		 perf_evsel__cmp(counter, nth_evsel(T_CYCLES_IN_TX)))
+		update_stats(&runtime_cycles_in_tx_stats[0], count[0]);
+	else if (transaction_run &&
+		 perf_evsel__cmp(counter, nth_evsel(T_TRANSACTION_START)))
+		update_stats(&runtime_transaction_stats[0], count[0]);
+	else if (transaction_run &&
+		 perf_evsel__cmp(counter, nth_evsel(T_ELISION_START)))
+		update_stats(&runtime_elision_stats[0], count[0]);
 	else if (perf_evsel__match(counter, HARDWARE, HW_STALLED_CYCLES_FRONTEND))
 		update_stats(&runtime_stalled_cycles_front_stats[0], count[0]);
 	else if (perf_evsel__match(counter, HARDWARE, HW_STALLED_CYCLES_BACKEND))
@@ -229,7 +398,7 @@ static int read_counter_aggr(struct perf_evsel *counter)
 	int i;
 
 	if (__perf_evsel__read(counter, perf_evsel__nr_cpus(counter),
-			       evsel_list->threads->nr, scale) < 0)
+			       thread_map__nr(evsel_list->threads), scale) < 0)
 		return -1;
 
 	for (i = 0; i < 3; i++)
@@ -269,70 +438,128 @@ static int read_counter(struct perf_evsel *counter)
 	return 0;
 }
 
-static int __run_perf_stat(int argc __maybe_unused, const char **argv)
+static void print_interval(void)
 {
-	unsigned long long t0, t1;
+	static int num_print_interval;
 	struct perf_evsel *counter;
-	int status = 0;
-	int child_ready_pipe[2], go_pipe[2];
-	const bool forks = (argc > 0);
-	char buf;
+	struct perf_stat *ps;
+	struct timespec ts, rs;
+	char prefix[64];
 
-	if (forks && (pipe(child_ready_pipe) < 0 || pipe(go_pipe) < 0)) {
-		perror("failed to create pipes");
-		return -1;
+	if (aggr_mode == AGGR_GLOBAL) {
+		evlist__for_each(evsel_list, counter) {
+			ps = counter->priv;
+			memset(ps->res_stats, 0, sizeof(ps->res_stats));
+			read_counter_aggr(counter);
+		}
+	} else	{
+		evlist__for_each(evsel_list, counter) {
+			ps = counter->priv;
+			memset(ps->res_stats, 0, sizeof(ps->res_stats));
+			read_counter(counter);
+		}
 	}
 
-	if (forks) {
-		if ((child_pid = fork()) < 0)
-			perror("failed to fork");
+	clock_gettime(CLOCK_MONOTONIC, &ts);
+	diff_timespec(&rs, &ts, &ref_time);
+	sprintf(prefix, "%6lu.%09lu%s", rs.tv_sec, rs.tv_nsec, csv_sep);
+
+	if (num_print_interval == 0 && !csv_output) {
+		switch (aggr_mode) {
+		case AGGR_SOCKET:
+			fprintf(output, "#           time socket cpus             counts %*s events\n", unit_width, "unit");
+			break;
+		case AGGR_CORE:
+			fprintf(output, "#           time core         cpus             counts %*s events\n", unit_width, "unit");
+			break;
+		case AGGR_NONE:
+			fprintf(output, "#           time CPU                counts %*s events\n", unit_width, "unit");
+			break;
+		case AGGR_GLOBAL:
+		default:
+			fprintf(output, "#           time             counts %*s events\n", unit_width, "unit");
+		}
+	}
 
-		if (!child_pid) {
-			close(child_ready_pipe[0]);
-			close(go_pipe[1]);
-			fcntl(go_pipe[0], F_SETFD, FD_CLOEXEC);
+	if (++num_print_interval == 25)
+		num_print_interval = 0;
+
+	switch (aggr_mode) {
+	case AGGR_CORE:
+	case AGGR_SOCKET:
+		print_aggr(prefix);
+		break;
+	case AGGR_NONE:
+		evlist__for_each(evsel_list, counter)
+			print_counter(counter, prefix);
+		break;
+	case AGGR_GLOBAL:
+	default:
+		evlist__for_each(evsel_list, counter)
+			print_counter_aggr(counter, prefix);
+	}
 
-			/*
-			 * Do a dummy execvp to get the PLT entry resolved,
-			 * so we avoid the resolver overhead on the real
-			 * execvp call.
-			 */
-			execvp("", (char **)argv);
+	fflush(output);
+}
 
-			/*
-			 * Tell the parent we're ready to go
-			 */
-			close(child_ready_pipe[1]);
+static void handle_initial_delay(void)
+{
+	struct perf_evsel *counter;
 
-			/*
-			 * Wait until the parent tells us to go.
-			 */
-			if (read(go_pipe[0], &buf, 1) == -1)
-				perror("unable to read pipe");
+	if (initial_delay) {
+		const int ncpus = cpu_map__nr(evsel_list->cpus),
+			nthreads = thread_map__nr(evsel_list->threads);
+
+		usleep(initial_delay * 1000);
+		evlist__for_each(evsel_list, counter)
+			perf_evsel__enable(counter, ncpus, nthreads);
+	}
+}
 
-			execvp(argv[0], (char **)argv);
+static volatile int workload_exec_errno;
 
-			perror(argv[0]);
-			exit(-1);
-		}
+/*
+ * perf_evlist__prepare_workload will send a SIGUSR1
+ * if the fork fails, since we asked by setting its
+ * want_signal to true.
+ */
+static void workload_exec_failed_signal(int signo __maybe_unused, siginfo_t *info,
+					void *ucontext __maybe_unused)
+{
+	workload_exec_errno = info->si_value.sival_int;
+}
+
+static int __run_perf_stat(int argc, const char **argv)
+{
+	char msg[512];
+	unsigned long long t0, t1;
+	struct perf_evsel *counter;
+	struct timespec ts;
+	size_t l;
+	int status = 0;
+	const bool forks = (argc > 0);
 
-		if (perf_target__none(&target))
-			evsel_list->threads->map[0] = child_pid;
-
-		/*
-		 * Wait for the child to be ready to exec.
-		 */
-		close(child_ready_pipe[1]);
-		close(go_pipe[0]);
-		if (read(child_ready_pipe[0], &buf, 1) == -1)
-			perror("unable to read pipe");
-		close(child_ready_pipe[0]);
+	if (interval) {
+		ts.tv_sec  = interval / 1000;
+		ts.tv_nsec = (interval % 1000) * 1000000;
+	} else {
+		ts.tv_sec  = 1;
+		ts.tv_nsec = 0;
+	}
+
+	if (forks) {
+		if (perf_evlist__prepare_workload(evsel_list, &target, argv, false,
+						  workload_exec_failed_signal) < 0) {
+			perror("failed to prepare workload");
+			return -1;
+		}
+		child_pid = evsel_list->workload.pid;
 	}
 
 	if (group)
 		perf_evlist__set_leader(evsel_list);
 
-	list_for_each_entry(counter, &evsel_list->entries, node) {
+	evlist__for_each(evsel_list, counter) {
 		if (create_perf_stat_counter(counter) < 0) {
 			/*
 			 * PPC returns ENXIO for HW counters until 2.6.37
@@ -348,23 +575,20 @@ static int __run_perf_stat(int argc __maybe_unused, const char **argv)
 				continue;
 			}
 
-			if (errno == EPERM || errno == EACCES) {
-				error("You may not have permission to collect %sstats.\n"
-				      "\t Consider tweaking"
-				      " /proc/sys/kernel/perf_event_paranoid or running as root.",
-				      target.system_wide ? "system-wide " : "");
-			} else {
-				error("open_counter returned with %d (%s). "
-				      "/bin/dmesg may provide additional information.\n",
-				       errno, strerror(errno));
-			}
+			perf_evsel__open_strerror(counter, &target,
+						  errno, msg, sizeof(msg));
+			ui__error("%s\n", msg);
+
 			if (child_pid != -1)
 				kill(child_pid, SIGTERM);
 
-			pr_err("Not all events could be opened.\n");
 			return -1;
 		}
 		counter->supported = true;
+
+		l = strlen(counter->unit);
+		if (l > unit_width)
+			unit_width = l;
 	}
 
 	if (perf_evlist__apply_filters(evsel_list)) {
@@ -377,37 +601,58 @@ static int __run_perf_stat(int argc __maybe_unused, const char **argv)
 	 * Enable counters and exec the command:
 	 */
 	t0 = rdclock();
+	clock_gettime(CLOCK_MONOTONIC, &ref_time);
 
 	if (forks) {
-		close(go_pipe[1]);
+		perf_evlist__start_workload(evsel_list);
+		handle_initial_delay();
+
+		if (interval) {
+			while (!waitpid(child_pid, &status, WNOHANG)) {
+				nanosleep(&ts, NULL);
+				print_interval();
+			}
+		}
 		wait(&status);
+
+		if (workload_exec_errno) {
+			const char *emsg = strerror_r(workload_exec_errno, msg, sizeof(msg));
+			pr_err("Workload failed: %s\n", emsg);
+			return -1;
+		}
+
 		if (WIFSIGNALED(status))
 			psignal(WTERMSIG(status), argv[0]);
 	} else {
-		while(!done) sleep(1);
+		handle_initial_delay();
+		while (!done) {
+			nanosleep(&ts, NULL);
+			if (interval)
+				print_interval();
+		}
 	}
 
 	t1 = rdclock();
 
 	update_stats(&walltime_nsecs_stats, t1 - t0);
 
-	if (no_aggr) {
-		list_for_each_entry(counter, &evsel_list->entries, node) {
-			read_counter(counter);
-			perf_evsel__close_fd(counter, perf_evsel__nr_cpus(counter), 1);
-		}
-	} else {
-		list_for_each_entry(counter, &evsel_list->entries, node) {
+	if (aggr_mode == AGGR_GLOBAL) {
+		evlist__for_each(evsel_list, counter) {
 			read_counter_aggr(counter);
 			perf_evsel__close_fd(counter, perf_evsel__nr_cpus(counter),
-					     evsel_list->threads->nr);
+					     thread_map__nr(evsel_list->threads));
+		}
+	} else {
+		evlist__for_each(evsel_list, counter) {
+			read_counter(counter);
+			perf_evsel__close_fd(counter, perf_evsel__nr_cpus(counter), 1);
 		}
 	}
 
 	return WEXITSTATUS(status);
 }
 
-static int run_perf_stat(int argc __maybe_unused, const char **argv)
+static int run_perf_stat(int argc, const char **argv)
 {
 	int ret;
 
@@ -454,23 +699,66 @@ static void print_noise(struct perf_evsel *evsel, double avg)
 	print_noise_pct(stddev_stats(&ps->res_stats[0]), avg);
 }
 
-static void nsec_printout(int cpu, struct perf_evsel *evsel, double avg)
+static void aggr_printout(struct perf_evsel *evsel, int id, int nr)
+{
+	switch (aggr_mode) {
+	case AGGR_CORE:
+		fprintf(output, "S%d-C%*d%s%*d%s",
+			cpu_map__id_to_socket(id),
+			csv_output ? 0 : -8,
+			cpu_map__id_to_cpu(id),
+			csv_sep,
+			csv_output ? 0 : 4,
+			nr,
+			csv_sep);
+		break;
+	case AGGR_SOCKET:
+		fprintf(output, "S%*d%s%*d%s",
+			csv_output ? 0 : -5,
+			id,
+			csv_sep,
+			csv_output ? 0 : 4,
+			nr,
+			csv_sep);
+			break;
+	case AGGR_NONE:
+		fprintf(output, "CPU%*d%s",
+			csv_output ? 0 : -4,
+			perf_evsel__cpus(evsel)->map[id], csv_sep);
+		break;
+	case AGGR_GLOBAL:
+	default:
+		break;
+	}
+}
+
+static void nsec_printout(int cpu, int nr, struct perf_evsel *evsel, double avg)
 {
 	double msecs = avg / 1e6;
-	char cpustr[16] = { '\0', };
-	const char *fmt = csv_output ? "%s%.6f%s%s" : "%s%18.6f%s%-25s";
+	const char *fmt_v, *fmt_n;
+	char name[25];
 
-	if (no_aggr)
-		sprintf(cpustr, "CPU%*d%s",
-			csv_output ? 0 : -4,
-			perf_evsel__cpus(evsel)->map[cpu], csv_sep);
+	fmt_v = csv_output ? "%.6f%s" : "%18.6f%s";
+	fmt_n = csv_output ? "%s" : "%-25s";
 
-	fprintf(output, fmt, cpustr, msecs, csv_sep, perf_evsel__name(evsel));
+	aggr_printout(evsel, cpu, nr);
+
+	scnprintf(name, sizeof(name), "%s%s",
+		  perf_evsel__name(evsel), csv_output ? "" : " (msec)");
+
+	fprintf(output, fmt_v, msecs, csv_sep);
+
+	if (csv_output)
+		fprintf(output, "%s%s", evsel->unit, csv_sep);
+	else
+		fprintf(output, "%-*s%s", unit_width, evsel->unit, csv_sep);
+
+	fprintf(output, fmt_n, name);
 
 	if (evsel->cgrp)
 		fprintf(output, "%s%s", csv_sep, evsel->cgrp->name);
 
-	if (csv_output)
+	if (csv_output || interval)
 		return;
 
 	if (perf_evsel__match(evsel, SOFTWARE, SW_TASK_CLOCK))
@@ -659,48 +947,56 @@ static void print_ll_cache_misses(int cpu,
 	fprintf(output, " of all LL-cache hits   ");
 }
 
-static void abs_printout(int cpu, struct perf_evsel *evsel, double avg)
+static void abs_printout(int cpu, int nr, struct perf_evsel *evsel, double avg)
 {
-	double total, ratio = 0.0;
-	char cpustr[16] = { '\0', };
+	double total, ratio = 0.0, total2;
+	double sc =  evsel->scale;
 	const char *fmt;
 
-	if (csv_output)
-		fmt = "%s%.0f%s%s";
-	else if (big_num)
-		fmt = "%s%'18.0f%s%-25s";
-	else
-		fmt = "%s%18.0f%s%-25s";
+	if (csv_output) {
+		fmt = sc != 1.0 ?  "%.2f%s" : "%.0f%s";
+	} else {
+		if (big_num)
+			fmt = sc != 1.0 ? "%'18.2f%s" : "%'18.0f%s";
+		else
+			fmt = sc != 1.0 ? "%18.2f%s" : "%18.0f%s";
+	}
 
-	if (no_aggr)
-		sprintf(cpustr, "CPU%*d%s",
-			csv_output ? 0 : -4,
-			perf_evsel__cpus(evsel)->map[cpu], csv_sep);
-	else
+	aggr_printout(evsel, cpu, nr);
+
+	if (aggr_mode == AGGR_GLOBAL)
 		cpu = 0;
 
-	fprintf(output, fmt, cpustr, avg, csv_sep, perf_evsel__name(evsel));
+	fprintf(output, fmt, avg, csv_sep);
+
+	if (evsel->unit)
+		fprintf(output, "%-*s%s",
+			csv_output ? 0 : unit_width,
+			evsel->unit, csv_sep);
+
+	fprintf(output, "%-*s", csv_output ? 0 : 25, perf_evsel__name(evsel));
 
 	if (evsel->cgrp)
 		fprintf(output, "%s%s", csv_sep, evsel->cgrp->name);
 
-	if (csv_output)
+	if (csv_output || interval)
 		return;
 
 	if (perf_evsel__match(evsel, HARDWARE, HW_INSTRUCTIONS)) {
 		total = avg_stats(&runtime_cycles_stats[cpu]);
-
-		if (total)
+		if (total) {
 			ratio = avg / total;
-
-		fprintf(output, " #   %5.2f  insns per cycle        ", ratio);
-
+			fprintf(output, " #   %5.2f  insns per cycle        ", ratio);
+		}
 		total = avg_stats(&runtime_stalled_cycles_front_stats[cpu]);
 		total = max(total, avg_stats(&runtime_stalled_cycles_back_stats[cpu]));
 
 		if (total && avg) {
 			ratio = total / avg;
-			fprintf(output, "\n                                             #   %5.2f  stalled cycles per insn", ratio);
+			fprintf(output, "\n");
+			if (aggr_mode == AGGR_NONE)
+				fprintf(output, "        ");
+			fprintf(output, "                                                  #   %5.2f  stalled cycles per insn", ratio);
 		}
 
 	} else if (perf_evsel__match(evsel, HARDWARE, HW_BRANCH_MISSES) &&
@@ -757,10 +1053,47 @@ static void abs_printout(int cpu, struct perf_evsel *evsel, double avg)
 	} else if (perf_evsel__match(evsel, HARDWARE, HW_CPU_CYCLES)) {
 		total = avg_stats(&runtime_nsecs_stats[cpu]);
 
+		if (total) {
+			ratio = avg / total;
+			fprintf(output, " # %8.3f GHz                    ", ratio);
+		}
+	} else if (transaction_run &&
+		   perf_evsel__cmp(evsel, nth_evsel(T_CYCLES_IN_TX))) {
+		total = avg_stats(&runtime_cycles_stats[cpu]);
 		if (total)
-			ratio = 1.0 * avg / total;
+			fprintf(output,
+				" #   %5.2f%% transactional cycles   ",
+				100.0 * (avg / total));
+	} else if (transaction_run &&
+		   perf_evsel__cmp(evsel, nth_evsel(T_CYCLES_IN_TX_CP))) {
+		total = avg_stats(&runtime_cycles_stats[cpu]);
+		total2 = avg_stats(&runtime_cycles_in_tx_stats[cpu]);
+		if (total2 < avg)
+			total2 = avg;
+		if (total)
+			fprintf(output,
+				" #   %5.2f%% aborted cycles         ",
+				100.0 * ((total2-avg) / total));
+	} else if (transaction_run &&
+		   perf_evsel__cmp(evsel, nth_evsel(T_TRANSACTION_START)) &&
+		   avg > 0 &&
+		   runtime_cycles_in_tx_stats[cpu].n != 0) {
+		total = avg_stats(&runtime_cycles_in_tx_stats[cpu]);
 
-		fprintf(output, " # %8.3f GHz                    ", ratio);
+		if (total)
+			ratio = total / avg;
+
+		fprintf(output, " # %8.0f cycles / transaction   ", ratio);
+	} else if (transaction_run &&
+		   perf_evsel__cmp(evsel, nth_evsel(T_ELISION_START)) &&
+		   avg > 0 &&
+		   runtime_cycles_in_tx_stats[cpu].n != 0) {
+		total = avg_stats(&runtime_cycles_in_tx_stats[cpu]);
+
+		if (total)
+			ratio = total / avg;
+
+		fprintf(output, " # %8.0f cycles / elision       ", ratio);
 	} else if (runtime_nsecs_stats[cpu].n != 0) {
 		char unit = 'M';
 
@@ -779,22 +1112,100 @@ static void abs_printout(int cpu, struct perf_evsel *evsel, double avg)
 	}
 }
 
+static void print_aggr(char *prefix)
+{
+	struct perf_evsel *counter;
+	int cpu, cpu2, s, s2, id, nr;
+	double uval;
+	u64 ena, run, val;
+
+	if (!(aggr_map || aggr_get_id))
+		return;
+
+	for (s = 0; s < aggr_map->nr; s++) {
+		id = aggr_map->map[s];
+		evlist__for_each(evsel_list, counter) {
+			val = ena = run = 0;
+			nr = 0;
+			for (cpu = 0; cpu < perf_evsel__nr_cpus(counter); cpu++) {
+				cpu2 = perf_evsel__cpus(counter)->map[cpu];
+				s2 = aggr_get_id(evsel_list->cpus, cpu2);
+				if (s2 != id)
+					continue;
+				val += counter->counts->cpu[cpu].val;
+				ena += counter->counts->cpu[cpu].ena;
+				run += counter->counts->cpu[cpu].run;
+				nr++;
+			}
+			if (prefix)
+				fprintf(output, "%s", prefix);
+
+			if (run == 0 || ena == 0) {
+				aggr_printout(counter, id, nr);
+
+				fprintf(output, "%*s%s",
+					csv_output ? 0 : 18,
+					counter->supported ? CNTR_NOT_COUNTED : CNTR_NOT_SUPPORTED,
+					csv_sep);
+
+				fprintf(output, "%-*s%s",
+					csv_output ? 0 : unit_width,
+					counter->unit, csv_sep);
+
+				fprintf(output, "%*s",
+					csv_output ? 0 : -25,
+					perf_evsel__name(counter));
+
+				if (counter->cgrp)
+					fprintf(output, "%s%s",
+						csv_sep, counter->cgrp->name);
+
+				fputc('\n', output);
+				continue;
+			}
+			uval = val * counter->scale;
+
+			if (nsec_counter(counter))
+				nsec_printout(id, nr, counter, uval);
+			else
+				abs_printout(id, nr, counter, uval);
+
+			if (!csv_output) {
+				print_noise(counter, 1.0);
+
+				if (run != ena)
+					fprintf(output, "  (%.2f%%)",
+						100.0 * run / ena);
+			}
+			fputc('\n', output);
+		}
+	}
+}
+
 /*
  * Print out the results of a single counter:
  * aggregated counts in system-wide mode
  */
-static void print_counter_aggr(struct perf_evsel *counter)
+static void print_counter_aggr(struct perf_evsel *counter, char *prefix)
 {
 	struct perf_stat *ps = counter->priv;
 	double avg = avg_stats(&ps->res_stats[0]);
 	int scaled = counter->counts->scaled;
+	double uval;
+
+	if (prefix)
+		fprintf(output, "%s", prefix);
 
 	if (scaled == -1) {
-		fprintf(output, "%*s%s%*s",
+		fprintf(output, "%*s%s",
 			csv_output ? 0 : 18,
 			counter->supported ? CNTR_NOT_COUNTED : CNTR_NOT_SUPPORTED,
-			csv_sep,
-			csv_output ? 0 : -24,
+			csv_sep);
+		fprintf(output, "%-*s%s",
+			csv_output ? 0 : unit_width,
+			counter->unit, csv_sep);
+		fprintf(output, "%*s",
+			csv_output ? 0 : -25,
 			perf_evsel__name(counter));
 
 		if (counter->cgrp)
@@ -804,10 +1215,12 @@ static void print_counter_aggr(struct perf_evsel *counter)
 		return;
 	}
 
+	uval = avg * counter->scale;
+
 	if (nsec_counter(counter))
-		nsec_printout(-1, counter, avg);
+		nsec_printout(-1, 0, counter, uval);
 	else
-		abs_printout(-1, counter, avg);
+		abs_printout(-1, 0, counter, uval);
 
 	print_noise(counter, avg);
 
@@ -831,24 +1244,35 @@ static void print_counter_aggr(struct perf_evsel *counter)
  * Print out the results of a single counter:
  * does not use aggregated count in system-wide
  */
-static void print_counter(struct perf_evsel *counter)
+static void print_counter(struct perf_evsel *counter, char *prefix)
 {
 	u64 ena, run, val;
+	double uval;
 	int cpu;
 
 	for (cpu = 0; cpu < perf_evsel__nr_cpus(counter); cpu++) {
 		val = counter->counts->cpu[cpu].val;
 		ena = counter->counts->cpu[cpu].ena;
 		run = counter->counts->cpu[cpu].run;
+
+		if (prefix)
+			fprintf(output, "%s", prefix);
+
 		if (run == 0 || ena == 0) {
-			fprintf(output, "CPU%*d%s%*s%s%*s",
+			fprintf(output, "CPU%*d%s%*s%s",
 				csv_output ? 0 : -4,
 				perf_evsel__cpus(counter)->map[cpu], csv_sep,
 				csv_output ? 0 : 18,
 				counter->supported ? CNTR_NOT_COUNTED : CNTR_NOT_SUPPORTED,
-				csv_sep,
-				csv_output ? 0 : -24,
-				perf_evsel__name(counter));
+				csv_sep);
+
+				fprintf(output, "%-*s%s",
+					csv_output ? 0 : unit_width,
+					counter->unit, csv_sep);
+
+				fprintf(output, "%*s",
+					csv_output ? 0 : -25,
+					perf_evsel__name(counter));
 
 			if (counter->cgrp)
 				fprintf(output, "%s%s",
@@ -858,10 +1282,12 @@ static void print_counter(struct perf_evsel *counter)
 			continue;
 		}
 
+		uval = val * counter->scale;
+
 		if (nsec_counter(counter))
-			nsec_printout(cpu, counter, val);
+			nsec_printout(cpu, 0, counter, uval);
 		else
-			abs_printout(cpu, counter, val);
+			abs_printout(cpu, 0, counter, uval);
 
 		if (!csv_output) {
 			print_noise(counter, 1.0);
@@ -884,7 +1310,11 @@ static void print_stat(int argc, const char **argv)
 	if (!csv_output) {
 		fprintf(output, "\n");
 		fprintf(output, " Performance counter stats for ");
-		if (!perf_target__has_task(&target)) {
+		if (target.system_wide)
+			fprintf(output, "\'system wide");
+		else if (target.cpu_list)
+			fprintf(output, "\'CPU(s) %s", target.cpu_list);
+		else if (!target__has_task(&target)) {
 			fprintf(output, "\'%s", argv[0]);
 			for (i = 1; i < argc; i++)
 				fprintf(output, " %s", argv[i]);
@@ -899,12 +1329,21 @@ static void print_stat(int argc, const char **argv)
 		fprintf(output, ":\n\n");
 	}
 
-	if (no_aggr) {
-		list_for_each_entry(counter, &evsel_list->entries, node)
-			print_counter(counter);
-	} else {
-		list_for_each_entry(counter, &evsel_list->entries, node)
-			print_counter_aggr(counter);
+	switch (aggr_mode) {
+	case AGGR_CORE:
+	case AGGR_SOCKET:
+		print_aggr(NULL);
+		break;
+	case AGGR_GLOBAL:
+		evlist__for_each(evsel_list, counter)
+			print_counter_aggr(counter, NULL);
+		break;
+	case AGGR_NONE:
+		evlist__for_each(evsel_list, counter)
+			print_counter(counter, NULL);
+		break;
+	default:
+		break;
 	}
 
 	if (!csv_output) {
@@ -925,17 +1364,38 @@ static volatile int signr = -1;
 
 static void skip_signal(int signo)
 {
-	if(child_pid == -1)
+	if ((child_pid == -1) || interval)
 		done = 1;
 
 	signr = signo;
+	/*
+	 * render child_pid harmless
+	 * won't send SIGTERM to a random
+	 * process in case of race condition
+	 * and fast PID recycling
+	 */
+	child_pid = -1;
 }
 
 static void sig_atexit(void)
 {
+	sigset_t set, oset;
+
+	/*
+	 * avoid race condition with SIGCHLD handler
+	 * in skip_signal() which is modifying child_pid
+	 * goal is to avoid send SIGTERM to a random
+	 * process
+	 */
+	sigemptyset(&set);
+	sigaddset(&set, SIGCHLD);
+	sigprocmask(SIG_BLOCK, &set, &oset);
+
 	if (child_pid != -1)
 		kill(child_pid, SIGTERM);
 
+	sigprocmask(SIG_SETMASK, &oset, NULL);
+
 	if (signr == -1)
 		return;
 
@@ -950,6 +1410,42 @@ static int stat__set_big_num(const struct option *opt __maybe_unused,
 	return 0;
 }
 
+static int perf_stat_init_aggr_mode(void)
+{
+	switch (aggr_mode) {
+	case AGGR_SOCKET:
+		if (cpu_map__build_socket_map(evsel_list->cpus, &aggr_map)) {
+			perror("cannot build socket map");
+			return -1;
+		}
+		aggr_get_id = cpu_map__get_socket;
+		break;
+	case AGGR_CORE:
+		if (cpu_map__build_core_map(evsel_list->cpus, &aggr_map)) {
+			perror("cannot build core map");
+			return -1;
+		}
+		aggr_get_id = cpu_map__get_core;
+		break;
+	case AGGR_NONE:
+	case AGGR_GLOBAL:
+	default:
+		break;
+	}
+	return 0;
+}
+
+static int setup_events(const char * const *attrs, unsigned len)
+{
+	unsigned i;
+
+	for (i = 0; i < len; i++) {
+		if (parse_events(evsel_list, attrs[i]))
+			return -1;
+	}
+	return 0;
+}
+
 /*
  * Add default attributes, if there were no attributes specified or
  * if -d/--detailed, -d -d or -d -d -d is used:
@@ -1067,6 +1563,22 @@ static int add_default_attributes(void)
 	if (null_run)
 		return 0;
 
+	if (transaction_run) {
+		int err;
+		if (pmu_have_event("cpu", "cycles-ct") &&
+		    pmu_have_event("cpu", "el-start"))
+			err = setup_events(transaction_attrs,
+					ARRAY_SIZE(transaction_attrs));
+		else
+			err = setup_events(transaction_limited_attrs,
+				 ARRAY_SIZE(transaction_limited_attrs));
+		if (err < 0) {
+			fprintf(stderr, "Cannot set up transaction events\n");
+			return -1;
+		}
+		return 0;
+	}
+
 	if (!evsel_list->nr_entries) {
 		if (perf_evlist__add_default_attrs(evsel_list, default_attrs) < 0)
 			return -1;
@@ -1101,6 +1613,8 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 	int output_fd = 0;
 	const char *output_name	= NULL;
 	const struct option options[] = {
+	OPT_BOOLEAN('T', "transaction", &transaction_run,
+		    "hardware transaction statistics"),
 	OPT_CALLBACK('e', "event", &evsel_list, "event",
 		     "event selector. use 'perf list' to list available events",
 		     parse_events_option),
@@ -1120,7 +1634,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 	OPT_INCR('v', "verbose", &verbose,
 		    "be more verbose (show counter open errors, etc)"),
 	OPT_INTEGER('r', "repeat", &run_count,
-		    "repeat command and print average + stddev (max: 100)"),
+		    "repeat command and print average + stddev (max: 100, forever: 0)"),
 	OPT_BOOLEAN('n', "null", &null_run,
 		    "null run - dont start any counters"),
 	OPT_INCR('d', "detailed", &detailed_run,
@@ -1132,7 +1646,8 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 			   stat__set_big_num),
 	OPT_STRING('C', "cpu", &target.cpu_list, "cpu",
 		    "list of cpus to monitor in system-wide"),
-	OPT_BOOLEAN('A', "no-aggr", &no_aggr, "disable CPU count aggregation"),
+	OPT_SET_UINT('A', "no-aggr", &aggr_mode,
+		    "disable CPU count aggregation", AGGR_NONE),
 	OPT_STRING('x', "field-separator", &csv_sep, "separator",
 		   "print counts with custom separator"),
 	OPT_CALLBACK('G', "cgroup", &evsel_list, "name",
@@ -1145,19 +1660,26 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 			"command to run prior to the measured command"),
 	OPT_STRING(0, "post", &post_cmd, "command",
 			"command to run after to the measured command"),
+	OPT_UINTEGER('I', "interval-print", &interval,
+		    "print counts at regular interval in ms (>= 100)"),
+	OPT_SET_UINT(0, "per-socket", &aggr_mode,
+		     "aggregate counts per processor socket", AGGR_SOCKET),
+	OPT_SET_UINT(0, "per-core", &aggr_mode,
+		     "aggregate counts per physical processor core", AGGR_CORE),
+	OPT_UINTEGER('D', "delay", &initial_delay,
+		     "ms to wait before starting measurement after program start"),
 	OPT_END()
 	};
 	const char * const stat_usage[] = {
 		"perf stat [<options>] [<command>]",
 		NULL
 	};
-	struct perf_evsel *pos;
-	int status = -ENOMEM, run_idx;
+	int status = -EINVAL, run_idx;
 	const char *mode;
 
 	setlocale(LC_ALL, "");
 
-	evsel_list = perf_evlist__new(NULL, NULL);
+	evsel_list = perf_evlist__new();
 	if (evsel_list == NULL)
 		return -ENOMEM;
 
@@ -1170,12 +1692,15 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 
 	if (output_name && output_fd) {
 		fprintf(stderr, "cannot use both --output and --log-fd\n");
-		usage_with_options(stat_usage, options);
+		parse_options_usage(stat_usage, options, "o", 1);
+		parse_options_usage(NULL, options, "log-fd", 0);
+		goto out;
 	}
 
 	if (output_fd < 0) {
 		fprintf(stderr, "argument to --log-fd must be a > 0\n");
-		usage_with_options(stat_usage, options);
+		parse_options_usage(stat_usage, options, "log-fd", 0);
+		goto out;
 	}
 
 	if (!output) {
@@ -1212,46 +1737,67 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 		/* User explicitly passed -B? */
 		if (big_num_opt == 1) {
 			fprintf(stderr, "-B option not supported with -x\n");
-			usage_with_options(stat_usage, options);
+			parse_options_usage(stat_usage, options, "B", 1);
+			parse_options_usage(NULL, options, "x", 1);
+			goto out;
 		} else /* Nope, so disable big number formatting */
 			big_num = false;
 	} else if (big_num_opt == 0) /* User passed --no-big-num */
 		big_num = false;
 
-	if (!argc && !perf_target__has_task(&target))
-		usage_with_options(stat_usage, options);
-	if (run_count <= 0)
+	if (!argc && target__none(&target))
 		usage_with_options(stat_usage, options);
 
+	if (run_count < 0) {
+		pr_err("Run count must be a positive number\n");
+		parse_options_usage(stat_usage, options, "r", 1);
+		goto out;
+	} else if (run_count == 0) {
+		forever = true;
+		run_count = 1;
+	}
+
 	/* no_aggr, cgroup are for system-wide only */
-	if ((no_aggr || nr_cgroups) && !perf_target__has_cpu(&target)) {
+	if ((aggr_mode != AGGR_GLOBAL || nr_cgroups) &&
+	    !target__has_cpu(&target)) {
 		fprintf(stderr, "both cgroup and no-aggregation "
 			"modes only available in system-wide mode\n");
 
-		usage_with_options(stat_usage, options);
+		parse_options_usage(stat_usage, options, "G", 1);
+		parse_options_usage(NULL, options, "A", 1);
+		parse_options_usage(NULL, options, "a", 1);
+		goto out;
 	}
 
 	if (add_default_attributes())
 		goto out;
 
-	perf_target__validate(&target);
+	target__validate(&target);
 
 	if (perf_evlist__create_maps(evsel_list, &target) < 0) {
-		if (perf_target__has_task(&target))
+		if (target__has_task(&target)) {
 			pr_err("Problems finding threads of monitor\n");
-		if (perf_target__has_cpu(&target))
+			parse_options_usage(stat_usage, options, "p", 1);
+			parse_options_usage(NULL, options, "t", 1);
+		} else if (target__has_cpu(&target)) {
 			perror("failed to parse CPUs map");
-
-		usage_with_options(stat_usage, options);
-		return -1;
+			parse_options_usage(stat_usage, options, "C", 1);
+			parse_options_usage(NULL, options, "a", 1);
+		}
+		goto out;
 	}
-
-	list_for_each_entry(pos, &evsel_list->entries, node) {
-		if (perf_evsel__alloc_stat_priv(pos) < 0 ||
-		    perf_evsel__alloc_counts(pos, perf_evsel__nr_cpus(pos)) < 0)
-			goto out_free_fd;
+	if (interval && interval < 100) {
+		pr_err("print interval must be >= 100ms\n");
+		parse_options_usage(stat_usage, options, "I", 1);
+		goto out;
 	}
 
+	if (perf_evlist__alloc_stats(evsel_list, interval))
+		goto out;
+
+	if (perf_stat_init_aggr_mode())
+		goto out;
+
 	/*
 	 * We dont want to block the signals - that would cause
 	 * child tasks to inherit that and Ctrl-C would not work.
@@ -1259,25 +1805,29 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 	 * task, but being ignored by perf stat itself:
 	 */
 	atexit(sig_atexit);
-	signal(SIGINT,  skip_signal);
+	if (!forever)
+		signal(SIGINT,  skip_signal);
+	signal(SIGCHLD, skip_signal);
 	signal(SIGALRM, skip_signal);
 	signal(SIGABRT, skip_signal);
 
 	status = 0;
-	for (run_idx = 0; run_idx < run_count; run_idx++) {
+	for (run_idx = 0; forever || run_idx < run_count; run_idx++) {
 		if (run_count != 1 && verbose)
 			fprintf(output, "[ perf stat: executing run #%d ... ]\n",
 				run_idx + 1);
 
 		status = run_perf_stat(argc, argv);
+		if (forever && status != -1) {
+			print_stat(argc, argv);
+			perf_stat__reset_stats(evsel_list);
+		}
 	}
 
-	if (status != -1)
+	if (!forever && status != -1 && !interval)
 		print_stat(argc, argv);
-out_free_fd:
-	list_for_each_entry(pos, &evsel_list->entries, node)
-		perf_evsel__free_stat_priv(pos);
-	perf_evlist__delete_maps(evsel_list);
+
+	perf_evlist__free_stats(evsel_list);
 out:
 	perf_evlist__delete(evsel_list);
 	return status;