xref: /linux/tools/perf/builtin-stat.c (revision c7decec2f2d2ab0366567f9e30c0e1418cece43f)

// SPDX-License-Identifier: GPL-2.0-only
/*
 * builtin-stat.c
 *
 * Builtin stat command: Give a precise performance counters summary
 * overview about any workload, CPU or specific PID.
 *
 * Sample output:

   $ perf stat ./hackbench 10

  Time: 0.118

  Performance counter stats for './hackbench 10':

       1708.761321 task-clock                #   11.037 CPUs utilized
            41,190 context-switches          #    0.024 M/sec
             6,735 CPU-migrations            #    0.004 M/sec
            17,318 page-faults               #    0.010 M/sec
     5,205,202,243 cycles                    #    3.046 GHz
     3,856,436,920 stalled-cycles-frontend   #   74.09% frontend cycles idle
     1,600,790,871 stalled-cycles-backend    #   30.75% backend  cycles idle
     2,603,501,247 instructions              #    0.50  insns per cycle
                                             #    1.48  stalled cycles per insn
       484,357,498 branches                  #  283.455 M/sec
         6,388,934 branch-misses             #    1.32% of all branches

        0.154822978  seconds time elapsed

 *
 * Copyright (C) 2008-2011, Red Hat Inc, Ingo Molnar <mingo@redhat.com>
 *
 * Improvements and fixes by:
 *
 *   Arjan van de Ven <arjan@linux.intel.com>
 *   Yanmin Zhang <yanmin.zhang@intel.com>
 *   Wu Fengguang <fengguang.wu@intel.com>
 *   Mike Galbraith <efault@gmx.de>
 *   Paul Mackerras <paulus@samba.org>
 *   Jaswinder Singh Rajput <jaswinder@kernel.org>
 */

#include "builtin.h"
#include "util/cgroup.h"
#include <subcmd/parse-options.h>
#include "util/parse-events.h"
#include "util/pmus.h"
#include "util/pmu.h"
#include "util/tool_pmu.h"
#include "util/event.h"
#include "util/evlist.h"
#include "util/evsel.h"
#include "util/debug.h"
#include "util/color.h"
#include "util/stat.h"
#include "util/header.h"
#include "util/cpumap.h"
#include "util/thread_map.h"
#include "util/counts.h"
#include "util/topdown.h"
#include "util/session.h"
#include "util/tool.h"
#include "util/string2.h"
#include "util/metricgroup.h"
#include "util/synthetic-events.h"
#include "util/target.h"
#include "util/time-utils.h"
#include "util/top.h"
#include "util/affinity.h"
#include "util/pfm.h"
#include "util/bpf_counter.h"
#include "util/iostat.h"
#include "util/util.h"
#include "util/intel-tpebs.h"
#include "asm/bug.h"

#include <linux/list_sort.h>
#include <linux/time64.h>
#include <linux/zalloc.h>
#include <api/fs/fs.h>
#include <errno.h>
#include <signal.h>
#include <stdlib.h>
#include <sys/prctl.h>
#include <inttypes.h>
#include <locale.h>
#include <math.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/wait.h>
#include <unistd.h>
#include <sys/time.h>
#include <sys/resource.h>
#include <linux/err.h>

#include <linux/ctype.h>
#include <perf/evlist.h>
#include <internal/threadmap.h>

#ifdef HAVE_BPF_SKEL
#include "util/bpf_skel/bperf_cgroup.h"
#endif

#define DEFAULT_SEPARATOR	" "
#define FREEZE_ON_SMI_PATH	"bus/event_source/devices/cpu/freeze_on_smi"

struct rusage_stats {
	struct stats ru_utime_usec_stat;
	struct stats ru_stime_usec_stat;
};

static void print_counters(struct timespec *ts, int argc, const char **argv);

static struct evlist	*evsel_list;
static struct parse_events_option_args parse_events_option_args = {
	.evlistp = &evsel_list,
};

static bool all_counters_use_bpf = true;

static struct target target;

static volatile sig_atomic_t	child_pid			= -1;
static int			detailed_run			=  0;
static bool			transaction_run;
static bool			topdown_run			= false;
static bool			smi_cost			= false;
static bool			smi_reset			= false;
static int			big_num_opt			=  -1;
static const char		*pre_cmd			= NULL;
static const char		*post_cmd			= NULL;
static bool			sync_run			= false;
static bool			forever				= false;
static bool			force_metric_only		= false;
static struct timespec		ref_time;
static bool			append_file;
static bool			interval_count;
static const char		*output_name;
static int			output_fd;
static char			*metrics;
static struct rusage_stats	ru_stats;

struct perf_stat {
	bool			 record;
	struct perf_data	 data;
	struct perf_session	*session;
	u64			 bytes_written;
	struct perf_tool	 tool;
	bool			 maps_allocated;
	struct perf_cpu_map	*cpus;
	struct perf_thread_map *threads;
	enum aggr_mode		 aggr_mode;
	u32			 aggr_level;
};

static struct perf_stat		perf_stat;
#define STAT_RECORD		perf_stat.record

static volatile sig_atomic_t done = 0;

/* Options set from the command line. */
struct opt_aggr_mode {
	bool node, socket, die, cluster, cache, core, thread, no_aggr;
};

/* Turn command line option into most generic aggregation mode setting. */
static enum aggr_mode opt_aggr_mode_to_aggr_mode(struct opt_aggr_mode *opt_mode)
{
	enum aggr_mode mode = AGGR_GLOBAL;

	if (opt_mode->node)
		mode = AGGR_NODE;
	if (opt_mode->socket)
		mode = AGGR_SOCKET;
	if (opt_mode->die)
		mode = AGGR_DIE;
	if (opt_mode->cluster)
		mode = AGGR_CLUSTER;
	if (opt_mode->cache)
		mode = AGGR_CACHE;
	if (opt_mode->core)
		mode = AGGR_CORE;
	if (opt_mode->thread)
		mode = AGGR_THREAD;
	if (opt_mode->no_aggr)
		mode = AGGR_NONE;
	return mode;
}

static void evlist__check_cpu_maps(struct evlist *evlist)
{
	struct evsel *evsel, *warned_leader = NULL;

	evlist__for_each_entry(evlist, evsel) {
		struct evsel *leader = evsel__leader(evsel);

		/* Check that leader matches cpus with each member. */
		if (leader == evsel)
			continue;
		if (perf_cpu_map__equal(leader->core.cpus, evsel->core.cpus))
			continue;

		/* If there's mismatch disable the group and warn user. */
		if (warned_leader != leader) {
			char buf[200];

			pr_warning("WARNING: grouped events cpus do not match.\n"
				"Events with CPUs not matching the leader will "
				"be removed from the group.\n");
			evsel__group_desc(leader, buf, sizeof(buf));
			pr_warning("  %s\n", buf);
			warned_leader = leader;
		}
		if (verbose > 0) {
			char buf[200];

			cpu_map__snprint(leader->core.cpus, buf, sizeof(buf));
			pr_warning("     %s: %s\n", leader->name, buf);
			cpu_map__snprint(evsel->core.cpus, buf, sizeof(buf));
			pr_warning("     %s: %s\n", evsel->name, buf);
		}

		evsel__remove_from_group(evsel, leader);
	}
}

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 + NSEC_PER_SEC - b->tv_nsec;
		r->tv_sec--;
	} else {
		r->tv_nsec = a->tv_nsec - b->tv_nsec ;
	}
}

static void perf_stat__reset_stats(void)
{
	evlist__reset_stats(evsel_list);
	memset(stat_config.walltime_nsecs_stats, 0, sizeof(*stat_config.walltime_nsecs_stats));
}

static int process_synthesized_event(const struct perf_tool *tool __maybe_unused,
				     union perf_event *event,
				     struct perf_sample *sample __maybe_unused,
				     struct machine *machine __maybe_unused)
{
	if (perf_data__write(&perf_stat.data, event, event->header.size) < 0) {
		pr_err("failed to write perf data, error: %m\n");
		return -1;
	}

	perf_stat.bytes_written += event->header.size;
	return 0;
}

static int write_stat_round_event(u64 tm, u64 type)
{
	return perf_event__synthesize_stat_round(NULL, tm, type,
						 process_synthesized_event,
						 NULL);
}

#define WRITE_STAT_ROUND_EVENT(time, interval) \
	write_stat_round_event(time, PERF_STAT_ROUND_TYPE__ ## interval)

#define SID(e, x, y) xyarray__entry(e->core.sample_id, x, y)

static int evsel__write_stat_event(struct evsel *counter, int cpu_map_idx, u32 thread,
				   struct perf_counts_values *count)
{
	struct perf_sample_id *sid = SID(counter, cpu_map_idx, thread);
	struct perf_cpu cpu = perf_cpu_map__cpu(evsel__cpus(counter), cpu_map_idx);

	return perf_event__synthesize_stat(NULL, cpu, thread, sid->id, count,
					   process_synthesized_event, NULL);
}

static int read_single_counter(struct evsel *counter, int cpu_map_idx, int thread)
{
	int err = evsel__read_counter(counter, cpu_map_idx, thread);

	/*
	 * Reading user and system time will fail when the process
	 * terminates. Use the wait4 values in that case.
	 */
	if (err && cpu_map_idx == 0 &&
	    (evsel__tool_event(counter) == TOOL_PMU__EVENT_USER_TIME ||
	     evsel__tool_event(counter) == TOOL_PMU__EVENT_SYSTEM_TIME)) {
		struct perf_counts_values *count =
			perf_counts(counter->counts, cpu_map_idx, thread);
		struct perf_counts_values *old_count = NULL;
		u64 val;

		if (counter->prev_raw_counts)
			old_count = perf_counts(counter->prev_raw_counts, cpu_map_idx, thread);

		if (evsel__tool_event(counter) == TOOL_PMU__EVENT_USER_TIME)
			val = ru_stats.ru_utime_usec_stat.mean;
		else
			val = ru_stats.ru_stime_usec_stat.mean;

		count->val = val;
		if (old_count) {
			count->run = old_count->run + 1;
			count->ena = old_count->ena + 1;
		} else {
			count->run++;
			count->ena++;
		}
		return 0;
	}
	return err;
}

/*
 * Read out the results of a single counter:
 * do not aggregate counts across CPUs in system-wide mode
 */
static int read_counter_cpu(struct evsel *counter, int cpu_map_idx)
{
	int nthreads = perf_thread_map__nr(evsel_list->core.threads);
	int thread;

	if (!counter->supported)
		return -ENOENT;

	for (thread = 0; thread < nthreads; thread++) {
		struct perf_counts_values *count;

		count = perf_counts(counter->counts, cpu_map_idx, thread);

		/*
		 * The leader's group read loads data into its group members
		 * (via evsel__read_counter()) and sets their count->loaded.
		 */
		if (!perf_counts__is_loaded(counter->counts, cpu_map_idx, thread) &&
		    read_single_counter(counter, cpu_map_idx, thread)) {
			counter->counts->scaled = -1;
			perf_counts(counter->counts, cpu_map_idx, thread)->ena = 0;
			perf_counts(counter->counts, cpu_map_idx, thread)->run = 0;
			return -1;
		}

		perf_counts__set_loaded(counter->counts, cpu_map_idx, thread, false);

		if (STAT_RECORD) {
			if (evsel__write_stat_event(counter, cpu_map_idx, thread, count)) {
				pr_err("failed to write stat event\n");
				return -1;
			}
		}

		if (verbose > 1) {
			fprintf(stat_config.output,
				"%s: %d: %" PRIu64 " %" PRIu64 " %" PRIu64 "\n",
					evsel__name(counter),
					perf_cpu_map__cpu(evsel__cpus(counter),
							  cpu_map_idx).cpu,
					count->val, count->ena, count->run);
		}
	}

	return 0;
}

static int read_counters_with_affinity(void)
{
	struct evlist_cpu_iterator evlist_cpu_itr;

	if (all_counters_use_bpf)
		return 0;

	evlist__for_each_cpu(evlist_cpu_itr, evsel_list) {
		struct evsel *counter = evlist_cpu_itr.evsel;

		if (evsel__is_bpf(counter))
			continue;

		if (evsel__is_tool(counter))
			continue;

		if (!counter->err)
			counter->err = read_counter_cpu(counter, evlist_cpu_itr.cpu_map_idx);
	}

	return 0;
}

static int read_bpf_map_counters(void)
{
	struct evsel *counter;
	int err;

	evlist__for_each_entry(evsel_list, counter) {
		if (!evsel__is_bpf(counter))
			continue;

		err = bpf_counter__read(counter);
		if (err)
			return err;
	}
	return 0;
}

static int read_tool_counters(void)
{
	struct evsel *counter;

	evlist__for_each_entry(evsel_list, counter) {
		int idx;

		if (!evsel__is_tool(counter))
			continue;

		perf_cpu_map__for_each_idx(idx, counter->core.cpus) {
			if (!counter->err)
				counter->err = read_counter_cpu(counter, idx);
		}
	}
	return 0;
}

static int read_counters(void)
{
	int ret;

	if (stat_config.stop_read_counter)
		return 0;

	// Read all BPF counters first.
	ret = read_bpf_map_counters();
	if (ret)
		return ret;

	// Read non-BPF and non-tool counters next.
	ret = read_counters_with_affinity();
	if (ret)
		return ret;

	// Read the tool counters last. This way the duration_time counter
	// should always be greater than any other counter's enabled time.
	return read_tool_counters();
}

static void process_counters(void)
{
	struct evsel *counter;

	evlist__for_each_entry(evsel_list, counter) {
		if (counter->err)
			pr_debug("failed to read counter %s\n", counter->name);
		if (counter->err == 0 && perf_stat_process_counter(&stat_config, counter))
			pr_warning("failed to process counter %s\n", counter->name);
		counter->err = 0;
	}

	perf_stat_merge_counters(&stat_config, evsel_list);
	perf_stat_process_percore(&stat_config, evsel_list);
}

static void process_interval(void)
{
	struct timespec ts, rs;

	clock_gettime(CLOCK_MONOTONIC, &ts);
	diff_timespec(&rs, &ts, &ref_time);

	evlist__reset_aggr_stats(evsel_list);

	if (read_counters() == 0)
		process_counters();

	if (STAT_RECORD) {
		if (WRITE_STAT_ROUND_EVENT(rs.tv_sec * NSEC_PER_SEC + rs.tv_nsec, INTERVAL))
			pr_err("failed to write stat round event\n");
	}

	init_stats(stat_config.walltime_nsecs_stats);
	update_stats(stat_config.walltime_nsecs_stats, stat_config.interval * 1000000ULL);
	print_counters(&rs, 0, NULL);
}

static bool handle_interval(unsigned int interval, int *times)
{
	if (interval) {
		process_interval();
		if (interval_count && !(--(*times)))
			return true;
	}
	return false;
}

static int enable_counters(void)
{
	struct evsel *evsel;
	int err;

	evlist__for_each_entry(evsel_list, evsel) {
		if (!evsel__is_bpf(evsel))
			continue;

		err = bpf_counter__enable(evsel);
		if (err)
			return err;
	}

	if (!target__enable_on_exec(&target)) {
		if (!all_counters_use_bpf)
			evlist__enable(evsel_list);
	}
	return 0;
}

static void disable_counters(void)
{
	struct evsel *counter;

	/*
	 * If we don't have tracee (attaching to task or cpu), counters may
	 * still be running. To get accurate group ratios, we must stop groups
	 * from counting before reading their constituent counters.
	 */
	if (!target__none(&target)) {
		evlist__for_each_entry(evsel_list, counter)
			bpf_counter__disable(counter);
		if (!all_counters_use_bpf)
			evlist__disable(evsel_list);
	}
}

static volatile sig_atomic_t workload_exec_errno;

/*
 * 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 bool evsel__should_store_id(struct evsel *counter)
{
	return STAT_RECORD || counter->core.attr.read_format & PERF_FORMAT_ID;
}

static bool is_target_alive(struct target *_target,
			    struct perf_thread_map *threads)
{
	struct stat st;
	int i;

	if (!target__has_task(_target))
		return true;

	for (i = 0; i < threads->nr; i++) {
		char path[PATH_MAX];

		scnprintf(path, PATH_MAX, "%s/%d", procfs__mountpoint(),
			  threads->map[i].pid);

		if (!stat(path, &st))
			return true;
	}

	return false;
}

static void process_evlist(struct evlist *evlist, unsigned int interval)
{
	enum evlist_ctl_cmd cmd = EVLIST_CTL_CMD_UNSUPPORTED;

	if (evlist__ctlfd_process(evlist, &cmd) > 0) {
		switch (cmd) {
		case EVLIST_CTL_CMD_ENABLE:
			fallthrough;
		case EVLIST_CTL_CMD_DISABLE:
			if (interval)
				process_interval();
			break;
		case EVLIST_CTL_CMD_SNAPSHOT:
		case EVLIST_CTL_CMD_ACK:
		case EVLIST_CTL_CMD_UNSUPPORTED:
		case EVLIST_CTL_CMD_EVLIST:
		case EVLIST_CTL_CMD_STOP:
		case EVLIST_CTL_CMD_PING:
		default:
			break;
		}
	}
}

static void compute_tts(struct timespec *time_start, struct timespec *time_stop,
			int *time_to_sleep)
{
	int tts = *time_to_sleep;
	struct timespec time_diff;

	diff_timespec(&time_diff, time_stop, time_start);

	tts -= time_diff.tv_sec * MSEC_PER_SEC +
	       time_diff.tv_nsec / NSEC_PER_MSEC;

	if (tts < 0)
		tts = 0;

	*time_to_sleep = tts;
}

static int dispatch_events(bool forks, int timeout, int interval, int *times)
{
	int child_exited = 0, status = 0;
	int time_to_sleep, sleep_time;
	struct timespec time_start, time_stop;

	if (interval)
		sleep_time = interval;
	else if (timeout)
		sleep_time = timeout;
	else
		sleep_time = 1000;

	time_to_sleep = sleep_time;

	while (!done) {
		if (forks)
			child_exited = waitpid(child_pid, &status, WNOHANG);
		else
			child_exited = !is_target_alive(&target, evsel_list->core.threads) ? 1 : 0;

		if (child_exited)
			break;

		clock_gettime(CLOCK_MONOTONIC, &time_start);
		if (!(evlist__poll(evsel_list, time_to_sleep) > 0)) { /* poll timeout or EINTR */
			if (timeout || handle_interval(interval, times))
				break;
			time_to_sleep = sleep_time;
		} else { /* fd revent */
			process_evlist(evsel_list, interval);
			clock_gettime(CLOCK_MONOTONIC, &time_stop);
			compute_tts(&time_start, &time_stop, &time_to_sleep);
		}
	}

	return status;
}

enum counter_recovery {
	COUNTER_SKIP,
	COUNTER_RETRY,
};

static enum counter_recovery stat_handle_error(struct evsel *counter, int err)
{
	char msg[BUFSIZ];

	assert(!counter->supported);

	/*
	 * PPC returns ENXIO for HW counters until 2.6.37
	 * (behavior changed with commit b0a873e).
	 */
	if (err == EINVAL || err == ENOSYS || err == ENOENT || err == ENXIO) {
		if (verbose > 0) {
			evsel__open_strerror(counter, &target, err, msg, sizeof(msg));
			ui__warning("%s event is not supported by the kernel.\n%s\n",
				    evsel__name(counter), msg);
		}
		return COUNTER_SKIP;
	}
	if (evsel__fallback(counter, &target, err, msg, sizeof(msg))) {
		if (verbose > 0)
			ui__warning("%s\n", msg);
		counter->supported = true;
		return COUNTER_RETRY;
	}
	if (target__has_per_thread(&target) && err != EOPNOTSUPP &&
	    evsel_list->core.threads && evsel_list->core.threads->err_thread != -1) {
		/*
		 * For global --per-thread case, skip current
		 * error thread.
		 */
		if (!thread_map__remove(evsel_list->core.threads,
					evsel_list->core.threads->err_thread)) {
			evsel_list->core.threads->err_thread = -1;
			counter->supported = true;
			return COUNTER_RETRY;
		}
	}
	if (verbose > 0) {
		evsel__open_strerror(counter, &target, err, msg, sizeof(msg));
		ui__warning(err == EOPNOTSUPP
			? "%s event is not supported by the kernel.\n%s\n"
			: "skipping event %s that kernel failed to open.\n%s\n",
			evsel__name(counter), msg);
	}
	return COUNTER_SKIP;
}

static int create_perf_stat_counter(struct evsel *evsel,
				    struct perf_stat_config *config,
				    int cpu_map_idx)
{
	struct perf_event_attr *attr = &evsel->core.attr;
	struct evsel *leader = evsel__leader(evsel);

	/* Reset supported flag as creating a stat counter is retried. */
	attr->read_format = PERF_FORMAT_TOTAL_TIME_ENABLED |
			    PERF_FORMAT_TOTAL_TIME_RUNNING;

	/*
	 * The event is part of non trivial group, let's enable
	 * the group read (for leader) and ID retrieval for all
	 * members.
	 */
	if (leader->core.nr_members > 1)
		attr->read_format |= PERF_FORMAT_ID|PERF_FORMAT_GROUP;

	attr->inherit = !config->no_inherit && list_empty(&evsel->bpf_counter_list);

	/*
	 * Some events get initialized with sample_(period/type) set,
	 * like tracepoints. Clear it up for counting.
	 */
	attr->sample_period = 0;

	if (config->identifier)
		attr->sample_type = PERF_SAMPLE_IDENTIFIER;

	if (config->all_user) {
		attr->exclude_kernel = 1;
		attr->exclude_user   = 0;
	}

	if (config->all_kernel) {
		attr->exclude_kernel = 0;
		attr->exclude_user   = 1;
	}

	/*
	 * Disabling all counters initially, they will be enabled
	 * either manually by us or by kernel via enable_on_exec
	 * set later.
	 */
	if (evsel__is_group_leader(evsel)) {
		attr->disabled = 1;

		if (target__enable_on_exec(&target))
			attr->enable_on_exec = 1;
	}

	return evsel__open_per_cpu_and_thread(evsel, evsel__cpus(evsel), cpu_map_idx,
					      evsel->core.threads);
}

static void update_rusage_stats(const struct rusage *rusage)
{
	const u64 us_to_ns = 1000;
	const u64 s_to_ns = 1000000000;

	update_stats(&ru_stats.ru_utime_usec_stat,
		(rusage->ru_utime.tv_usec * us_to_ns + rusage->ru_utime.tv_sec * s_to_ns));
	update_stats(&ru_stats.ru_stime_usec_stat,
		(rusage->ru_stime.tv_usec * us_to_ns + rusage->ru_stime.tv_sec * s_to_ns));
}

static int __run_perf_stat(int argc, const char **argv, int run_idx)
{
	int interval = stat_config.interval;
	int times = stat_config.times;
	int timeout = stat_config.timeout;
	char msg[BUFSIZ];
	unsigned long long t0, t1;
	struct evsel *counter;
	size_t l;
	int status = 0;
	const bool forks = (argc > 0);
	bool is_pipe = STAT_RECORD ? perf_stat.data.is_pipe : false;
	struct evlist_cpu_iterator evlist_cpu_itr;
	int err, open_err = 0;
	bool second_pass = false, has_supported_counters;

	if (forks) {
		if (evlist__prepare_workload(evsel_list, &target, argv, is_pipe, workload_exec_failed_signal) < 0) {
			perror("failed to prepare workload");
			return -1;
		}
		child_pid = evsel_list->workload.pid;
	}

	evlist__for_each_entry(evsel_list, counter) {
		counter->reset_group = false;
		if (bpf_counter__load(counter, &target)) {
			err = -1;
			goto err_out;
		}
		if (!(evsel__is_bperf(counter)))
			all_counters_use_bpf = false;
	}

	evlist__reset_aggr_stats(evsel_list);

	/*
	 * bperf calls evsel__open_per_cpu() in bperf__load(), so
	 * no need to call it again here.
	 */
	if (!target.use_bpf) {
		evlist__for_each_cpu(evlist_cpu_itr, evsel_list) {
			counter = evlist_cpu_itr.evsel;

			if (counter->reset_group || !counter->supported)
				continue;
			if (evsel__is_bperf(counter))
				continue;

			while (true) {
				if (create_perf_stat_counter(counter, &stat_config,
							      evlist_cpu_itr.cpu_map_idx) == 0)
					break;

				open_err = errno;
				/*
				 * Weak group failed. We cannot just undo this
				 * here because earlier CPUs might be in group
				 * mode, and the kernel doesn't support mixing
				 * group and non group reads. Defer it to later.
				 * Don't close here because we're in the wrong
				 * affinity.
				 */
				if ((open_err == EINVAL || open_err == EBADF) &&
					evsel__leader(counter) != counter &&
					counter->weak_group) {
					evlist__reset_weak_group(evsel_list, counter, false);
					assert(counter->reset_group);
					counter->supported = true;
					second_pass = true;
					break;
				}

				if (stat_handle_error(counter, open_err) != COUNTER_RETRY)
					break;
			}
		}
	}
	if (second_pass) {
		/*
		 * Now redo all the weak group after closing them,
		 * and also close errored counters.
		 */

		/* First close errored or weak retry */
		evlist__for_each_cpu(evlist_cpu_itr, evsel_list) {
			counter = evlist_cpu_itr.evsel;

			if (!counter->reset_group && counter->supported)
				continue;

			perf_evsel__close_cpu(&counter->core, evlist_cpu_itr.cpu_map_idx);
		}
		/* Now reopen weak */
		evlist__for_each_cpu(evlist_cpu_itr, evsel_list) {
			counter = evlist_cpu_itr.evsel;

			if (!counter->reset_group)
				continue;

			while (true) {
				pr_debug2("reopening weak %s\n", evsel__name(counter));
				if (create_perf_stat_counter(counter, &stat_config,
							     evlist_cpu_itr.cpu_map_idx) == 0) {
					evlist_cpu_iterator__exit(&evlist_cpu_itr);
					break;
				}
				open_err = errno;
				if (stat_handle_error(counter, open_err) != COUNTER_RETRY) {
					evlist_cpu_iterator__exit(&evlist_cpu_itr);
					break;
				}
			}
		}
	}

	has_supported_counters = false;
	evlist__for_each_entry(evsel_list, counter) {
		if (!counter->supported) {
			perf_evsel__free_fd(&counter->core);
			continue;
		}
		has_supported_counters = true;

		l = strlen(counter->unit);
		if (l > stat_config.unit_width)
			stat_config.unit_width = l;

		if (evsel__should_store_id(counter) &&
		    evsel__store_ids(counter, evsel_list)) {
			err = -1;
			goto err_out;
		}
	}
	if (!has_supported_counters && !stat_config.null_run) {
		if (open_err) {
			evsel__open_strerror(evlist__first(evsel_list), &target, open_err,
					     msg, sizeof(msg));
		}
		ui__error("No supported events found.\n%s\n", msg);

		if (child_pid != -1)
			kill(child_pid, SIGTERM);
		err = -1;
		goto err_out;
	}

	if (evlist__apply_filters(evsel_list, &counter, &target)) {
		pr_err("failed to set filter \"%s\" on event %s: %m\n",
			counter->filter, evsel__name(counter));
		return -1;
	}

	if (STAT_RECORD) {
		int fd = perf_data__fd(&perf_stat.data);

		if (is_pipe) {
			err = perf_header__write_pipe(perf_data__fd(&perf_stat.data));
		} else {
			err = perf_session__write_header(perf_stat.session, evsel_list,
							 fd, false);
		}

		if (err < 0)
			goto err_out;

		err = perf_event__synthesize_stat_events(&stat_config, NULL, evsel_list,
							 process_synthesized_event, is_pipe);
		if (err < 0)
			goto err_out;

	}

	if (target.initial_delay) {
		pr_info(EVLIST_DISABLED_MSG);
	} else {
		err = enable_counters();
		if (err) {
			err = -1;
			goto err_out;
		}
	}

	/* Exec the command, if any */
	if (forks)
		evlist__start_workload(evsel_list);

	if (target.initial_delay > 0) {
		usleep(target.initial_delay * USEC_PER_MSEC);
		err = enable_counters();
		if (err) {
			err = -1;
			goto err_out;
		}

		pr_info(EVLIST_ENABLED_MSG);
	}

	t0 = rdclock();
	clock_gettime(CLOCK_MONOTONIC, &ref_time);

	if (forks) {
		if (interval || timeout || evlist__ctlfd_initialized(evsel_list))
			status = dispatch_events(forks, timeout, interval, &times);
		if (child_pid != -1) {
			if (timeout)
				kill(child_pid, SIGTERM);
			wait4(child_pid, &status, 0, &stat_config.ru_data);
		}

		if (workload_exec_errno) {
			errno = workload_exec_errno;
			pr_err("Workload failed: %m\n");
			err = -1;
			goto err_out;
		}

		if (WIFSIGNALED(status)) {
			/*
			 * We want to indicate failure to stop a repeat run,
			 * hence negative. We want the value to be the exit code
			 * of perf, which for termination by a signal is 128
			 * plus the signal number.
			 */
			err = 0 - (128 + WTERMSIG(status));
			psignal(WTERMSIG(status), argv[0]);
		} else {
			err = WEXITSTATUS(status);
		}
	} else {
		err = dispatch_events(forks, timeout, interval, &times);
	}

	disable_counters();

	t1 = rdclock();

	if (stat_config.walltime_run_table)
		stat_config.walltime_run[run_idx] = t1 - t0;

	if (interval && stat_config.summary) {
		stat_config.interval = 0;
		stat_config.stop_read_counter = true;
		init_stats(stat_config.walltime_nsecs_stats);
		update_stats(stat_config.walltime_nsecs_stats, t1 - t0);

		evlist__copy_prev_raw_counts(evsel_list);
		evlist__reset_prev_raw_counts(evsel_list);
		evlist__reset_aggr_stats(evsel_list);
	} else {
		update_stats(stat_config.walltime_nsecs_stats, t1 - t0);
		update_rusage_stats(&stat_config.ru_data);
	}

	/*
	 * Closing a group leader splits the group, and as we only disable
	 * group leaders, results in remaining events becoming enabled. To
	 * avoid arbitrary skew, we must read all counters before closing any
	 * group leaders.
	 */
	if (read_counters() == 0)
		process_counters();

	/*
	 * We need to keep evsel_list alive, because it's processed
	 * later the evsel_list will be closed after.
	 */
	if (!STAT_RECORD)
		evlist__close(evsel_list);

	return err;

err_out:
	if (forks)
		evlist__cancel_workload(evsel_list);

	return err;
}

/*
 * Returns -1 for fatal errors which signifies to not continue
 * when in repeat mode.
 *
 * Returns < -1 error codes when stat record is used. These
 * result in the stat information being displayed, but writing
 * to the file fails and is non fatal.
 */
static int run_perf_stat(int argc, const char **argv, int run_idx)
{
	int ret;

	if (pre_cmd) {
		ret = system(pre_cmd);
		if (ret)
			return ret;
	}

	if (sync_run)
		sync();

	ret = __run_perf_stat(argc, argv, run_idx);
	if (ret)
		return ret;

	if (post_cmd) {
		ret = system(post_cmd);
		if (ret)
			return ret;
	}

	return ret;
}

static void print_counters(struct timespec *ts, int argc, const char **argv)
{
	/* Do not print anything if we record to the pipe. */
	if (STAT_RECORD && perf_stat.data.is_pipe)
		return;
	if (quiet)
		return;

	evlist__print_counters(evsel_list, &stat_config, &target, ts, argc, argv);
}

static volatile sig_atomic_t signr = -1;

static void skip_signal(int signo)
{
	if ((child_pid == -1) || stat_config.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;

	signal(signr, SIG_DFL);
	kill(getpid(), signr);
}

static int stat__set_big_num(const struct option *opt __maybe_unused,
			     const char *s __maybe_unused, int unset)
{
	big_num_opt = unset ? 0 : 1;
	perf_stat__set_big_num(!unset);
	return 0;
}

static int enable_metric_only(const struct option *opt __maybe_unused,
			      const char *s __maybe_unused, int unset)
{
	force_metric_only = true;
	stat_config.metric_only = !unset;
	return 0;
}

static int append_metric_groups(const struct option *opt __maybe_unused,
			       const char *str,
			       int unset __maybe_unused)
{
	if (metrics) {
		char *tmp;

		if (asprintf(&tmp, "%s,%s", metrics, str) < 0)
			return -ENOMEM;
		free(metrics);
		metrics = tmp;
	} else {
		metrics = strdup(str);
		if (!metrics)
			return -ENOMEM;
	}
	return 0;
}

static int parse_control_option(const struct option *opt,
				const char *str,
				int unset __maybe_unused)
{
	struct perf_stat_config *config = opt->value;

	return evlist__parse_control(str, &config->ctl_fd, &config->ctl_fd_ack, &config->ctl_fd_close);
}

static int parse_stat_cgroups(const struct option *opt,
			      const char *str, int unset)
{
	if (stat_config.cgroup_list) {
		pr_err("--cgroup and --for-each-cgroup cannot be used together\n");
		return -1;
	}

	return parse_cgroups(opt, str, unset);
}

static int parse_cputype(const struct option *opt,
			     const char *str,
			     int unset __maybe_unused)
{
	const struct perf_pmu *pmu;
	struct evlist *evlist = *(struct evlist **)opt->value;

	if (!list_empty(&evlist->core.entries)) {
		fprintf(stderr, "Must define cputype before events/metrics\n");
		return -1;
	}

	pmu = perf_pmus__pmu_for_pmu_filter(str);
	if (!pmu) {
		fprintf(stderr, "--cputype %s is not supported!\n", str);
		return -1;
	}
	parse_events_option_args.pmu_filter = pmu->name;

	return 0;
}

static int parse_cache_level(const struct option *opt,
			     const char *str,
			     int unset __maybe_unused)
{
	int level;
	struct opt_aggr_mode *opt_aggr_mode = (struct opt_aggr_mode *)opt->value;
	u32 *aggr_level = (u32 *)opt->data;

	/*
	 * If no string is specified, aggregate based on the topology of
	 * Last Level Cache (LLC). Since the LLC level can change from
	 * architecture to architecture, set level greater than
	 * MAX_CACHE_LVL which will be interpreted as LLC.
	 */
	if (str == NULL) {
		level = MAX_CACHE_LVL + 1;
		goto out;
	}

	/*
	 * The format to specify cache level is LX or lX where X is the
	 * cache level.
	 */
	if (strlen(str) != 2 || (str[0] != 'l' && str[0] != 'L')) {
		pr_err("Cache level must be of form L[1-%d], or l[1-%d]\n",
		       MAX_CACHE_LVL,
		       MAX_CACHE_LVL);
		return -EINVAL;
	}

	level = atoi(&str[1]);
	if (level < 1) {
		pr_err("Cache level must be of form L[1-%d], or l[1-%d]\n",
		       MAX_CACHE_LVL,
		       MAX_CACHE_LVL);
		return -EINVAL;
	}

	if (level > MAX_CACHE_LVL) {
		pr_err("perf only supports max cache level of %d.\n"
		       "Consider increasing MAX_CACHE_LVL\n", MAX_CACHE_LVL);
		return -EINVAL;
	}
out:
	opt_aggr_mode->cache = true;
	*aggr_level = level;
	return 0;
}

/**
 * Calculate the cache instance ID from the map in
 * /sys/devices/system/cpu/cpuX/cache/indexY/shared_cpu_list
 * Cache instance ID is the first CPU reported in the shared_cpu_list file.
 */
static int cpu__get_cache_id_from_map(struct perf_cpu cpu, char *map)
{
	int id;
	struct perf_cpu_map *cpu_map = perf_cpu_map__new(map);

	/*
	 * If the map contains no CPU, consider the current CPU to
	 * be the first online CPU in the cache domain else use the
	 * first online CPU of the cache domain as the ID.
	 */
	id = perf_cpu_map__min(cpu_map).cpu;
	if (id == -1)
		id = cpu.cpu;

	/* Free the perf_cpu_map used to find the cache ID */
	perf_cpu_map__put(cpu_map);

	return id;
}

/**
 * cpu__get_cache_id - Returns 0 if successful in populating the
 * cache level and cache id. Cache level is read from
 * /sys/devices/system/cpu/cpuX/cache/indexY/level where as cache instance ID
 * is the first CPU reported by
 * /sys/devices/system/cpu/cpuX/cache/indexY/shared_cpu_list
 */
static int cpu__get_cache_details(struct perf_cpu cpu, struct perf_cache *cache)
{
	int ret = 0;
	u32 cache_level = stat_config.aggr_level;
	struct cpu_cache_level caches[MAX_CACHE_LVL];
	u32 i = 0, caches_cnt = 0;

	cache->cache_lvl = (cache_level > MAX_CACHE_LVL) ? 0 : cache_level;
	cache->cache = -1;

	ret = build_caches_for_cpu(cpu.cpu, caches, &caches_cnt);
	if (ret) {
		/*
		 * If caches_cnt is not 0, cpu_cache_level data
		 * was allocated when building the topology.
		 * Free the allocated data before returning.
		 */
		if (caches_cnt)
			goto free_caches;

		return ret;
	}

	if (!caches_cnt)
		return -1;

	/*
	 * Save the data for the highest level if no
	 * level was specified by the user.
	 */
	if (cache_level > MAX_CACHE_LVL) {
		int max_level_index = 0;

		for (i = 1; i < caches_cnt; ++i) {
			if (caches[i].level > caches[max_level_index].level)
				max_level_index = i;
		}

		cache->cache_lvl = caches[max_level_index].level;
		cache->cache = cpu__get_cache_id_from_map(cpu, caches[max_level_index].map);

		/* Reset i to 0 to free entire caches[] */
		i = 0;
		goto free_caches;
	}

	for (i = 0; i < caches_cnt; ++i) {
		if (caches[i].level == cache_level) {
			cache->cache_lvl = cache_level;
			cache->cache = cpu__get_cache_id_from_map(cpu, caches[i].map);
		}

		cpu_cache_level__free(&caches[i]);
	}

free_caches:
	/*
	 * Free all the allocated cpu_cache_level data.
	 */
	while (i < caches_cnt)
		cpu_cache_level__free(&caches[i++]);

	return ret;
}

/**
 * aggr_cpu_id__cache - Create an aggr_cpu_id with cache instache ID, cache
 * level, die and socket populated with the cache instache ID, cache level,
 * die and socket for cpu. The function signature is compatible with
 * aggr_cpu_id_get_t.
 */
static struct aggr_cpu_id aggr_cpu_id__cache(struct perf_cpu cpu, void *data)
{
	int ret;
	struct aggr_cpu_id id;
	struct perf_cache cache;

	id = aggr_cpu_id__die(cpu, data);
	if (aggr_cpu_id__is_empty(&id))
		return id;

	ret = cpu__get_cache_details(cpu, &cache);
	if (ret)
		return id;

	id.cache_lvl = cache.cache_lvl;
	id.cache = cache.cache;
	return id;
}

static const char *const aggr_mode__string[] = {
	[AGGR_CORE] = "core",
	[AGGR_CACHE] = "cache",
	[AGGR_CLUSTER] = "cluster",
	[AGGR_DIE] = "die",
	[AGGR_GLOBAL] = "global",
	[AGGR_NODE] = "node",
	[AGGR_NONE] = "none",
	[AGGR_SOCKET] = "socket",
	[AGGR_THREAD] = "thread",
	[AGGR_UNSET] = "unset",
};

static struct aggr_cpu_id perf_stat__get_socket(struct perf_stat_config *config __maybe_unused,
						struct perf_cpu cpu)
{
	return aggr_cpu_id__socket(cpu, /*data=*/NULL);
}

static struct aggr_cpu_id perf_stat__get_die(struct perf_stat_config *config __maybe_unused,
					     struct perf_cpu cpu)
{
	return aggr_cpu_id__die(cpu, /*data=*/NULL);
}

static struct aggr_cpu_id perf_stat__get_cache_id(struct perf_stat_config *config __maybe_unused,
						  struct perf_cpu cpu)
{
	return aggr_cpu_id__cache(cpu, /*data=*/NULL);
}

static struct aggr_cpu_id perf_stat__get_cluster(struct perf_stat_config *config __maybe_unused,
						 struct perf_cpu cpu)
{
	return aggr_cpu_id__cluster(cpu, /*data=*/NULL);
}

static struct aggr_cpu_id perf_stat__get_core(struct perf_stat_config *config __maybe_unused,
					      struct perf_cpu cpu)
{
	return aggr_cpu_id__core(cpu, /*data=*/NULL);
}

static struct aggr_cpu_id perf_stat__get_node(struct perf_stat_config *config __maybe_unused,
					      struct perf_cpu cpu)
{
	return aggr_cpu_id__node(cpu, /*data=*/NULL);
}

static struct aggr_cpu_id perf_stat__get_global(struct perf_stat_config *config __maybe_unused,
						struct perf_cpu cpu)
{
	return aggr_cpu_id__global(cpu, /*data=*/NULL);
}

static struct aggr_cpu_id perf_stat__get_cpu(struct perf_stat_config *config __maybe_unused,
					     struct perf_cpu cpu)
{
	return aggr_cpu_id__cpu(cpu, /*data=*/NULL);
}

static struct aggr_cpu_id perf_stat__get_aggr(struct perf_stat_config *config,
					      aggr_get_id_t get_id, struct perf_cpu cpu)
{
	struct aggr_cpu_id id;

	/* per-process mode - should use global aggr mode */
	if (cpu.cpu == -1 || cpu.cpu >= config->cpus_aggr_map->nr)
		return get_id(config, cpu);

	if (aggr_cpu_id__is_empty(&config->cpus_aggr_map->map[cpu.cpu]))
		config->cpus_aggr_map->map[cpu.cpu] = get_id(config, cpu);

	id = config->cpus_aggr_map->map[cpu.cpu];
	return id;
}

static struct aggr_cpu_id perf_stat__get_socket_cached(struct perf_stat_config *config,
						       struct perf_cpu cpu)
{
	return perf_stat__get_aggr(config, perf_stat__get_socket, cpu);
}

static struct aggr_cpu_id perf_stat__get_die_cached(struct perf_stat_config *config,
						    struct perf_cpu cpu)
{
	return perf_stat__get_aggr(config, perf_stat__get_die, cpu);
}

static struct aggr_cpu_id perf_stat__get_cluster_cached(struct perf_stat_config *config,
							struct perf_cpu cpu)
{
	return perf_stat__get_aggr(config, perf_stat__get_cluster, cpu);
}

static struct aggr_cpu_id perf_stat__get_cache_id_cached(struct perf_stat_config *config,
							 struct perf_cpu cpu)
{
	return perf_stat__get_aggr(config, perf_stat__get_cache_id, cpu);
}

static struct aggr_cpu_id perf_stat__get_core_cached(struct perf_stat_config *config,
						     struct perf_cpu cpu)
{
	return perf_stat__get_aggr(config, perf_stat__get_core, cpu);
}

static struct aggr_cpu_id perf_stat__get_node_cached(struct perf_stat_config *config,
						     struct perf_cpu cpu)
{
	return perf_stat__get_aggr(config, perf_stat__get_node, cpu);
}

static struct aggr_cpu_id perf_stat__get_global_cached(struct perf_stat_config *config,
						       struct perf_cpu cpu)
{
	return perf_stat__get_aggr(config, perf_stat__get_global, cpu);
}

static struct aggr_cpu_id perf_stat__get_cpu_cached(struct perf_stat_config *config,
						    struct perf_cpu cpu)
{
	return perf_stat__get_aggr(config, perf_stat__get_cpu, cpu);
}

static aggr_cpu_id_get_t aggr_mode__get_aggr(enum aggr_mode aggr_mode)
{
	switch (aggr_mode) {
	case AGGR_SOCKET:
		return aggr_cpu_id__socket;
	case AGGR_DIE:
		return aggr_cpu_id__die;
	case AGGR_CLUSTER:
		return aggr_cpu_id__cluster;
	case AGGR_CACHE:
		return aggr_cpu_id__cache;
	case AGGR_CORE:
		return aggr_cpu_id__core;
	case AGGR_NODE:
		return aggr_cpu_id__node;
	case AGGR_NONE:
		return aggr_cpu_id__cpu;
	case AGGR_GLOBAL:
		return aggr_cpu_id__global;
	case AGGR_THREAD:
	case AGGR_UNSET:
	case AGGR_MAX:
	default:
		return NULL;
	}
}

static aggr_get_id_t aggr_mode__get_id(enum aggr_mode aggr_mode)
{
	switch (aggr_mode) {
	case AGGR_SOCKET:
		return perf_stat__get_socket_cached;
	case AGGR_DIE:
		return perf_stat__get_die_cached;
	case AGGR_CLUSTER:
		return perf_stat__get_cluster_cached;
	case AGGR_CACHE:
		return perf_stat__get_cache_id_cached;
	case AGGR_CORE:
		return perf_stat__get_core_cached;
	case AGGR_NODE:
		return perf_stat__get_node_cached;
	case AGGR_NONE:
		return perf_stat__get_cpu_cached;
	case AGGR_GLOBAL:
		return perf_stat__get_global_cached;
	case AGGR_THREAD:
	case AGGR_UNSET:
	case AGGR_MAX:
	default:
		return NULL;
	}
}

static int perf_stat_init_aggr_mode(void)
{
	int nr;
	aggr_cpu_id_get_t get_id = aggr_mode__get_aggr(stat_config.aggr_mode);

	if (get_id) {
		bool needs_sort = stat_config.aggr_mode != AGGR_NONE;
		stat_config.aggr_map = cpu_aggr_map__new(evsel_list->core.user_requested_cpus,
							 get_id, /*data=*/NULL, needs_sort);
		if (!stat_config.aggr_map) {
			pr_err("cannot build %s map\n", aggr_mode__string[stat_config.aggr_mode]);
			return -1;
		}
		stat_config.aggr_get_id = aggr_mode__get_id(stat_config.aggr_mode);
	}

	if (stat_config.aggr_mode == AGGR_THREAD) {
		nr = perf_thread_map__nr(evsel_list->core.threads);
		stat_config.aggr_map = cpu_aggr_map__empty_new(nr);
		if (stat_config.aggr_map == NULL)
			return -ENOMEM;

		for (int s = 0; s < nr; s++) {
			struct aggr_cpu_id id = aggr_cpu_id__empty();

			id.thread_idx = s;
			stat_config.aggr_map->map[s] = id;
		}
		return 0;
	}

	/*
	 * The evsel_list->cpus is the base we operate on,
	 * taking the highest cpu number to be the size of
	 * the aggregation translate cpumap.
	 */
	nr = perf_cpu_map__max(evsel_list->core.all_cpus).cpu + 1;
	stat_config.cpus_aggr_map = cpu_aggr_map__empty_new(nr);
	return stat_config.cpus_aggr_map ? 0 : -ENOMEM;
}

static void cpu_aggr_map__delete(struct cpu_aggr_map *map)
{
	free(map);
}

static void perf_stat__exit_aggr_mode(void)
{
	cpu_aggr_map__delete(stat_config.aggr_map);
	cpu_aggr_map__delete(stat_config.cpus_aggr_map);
	stat_config.aggr_map = NULL;
	stat_config.cpus_aggr_map = NULL;
}

static struct aggr_cpu_id perf_env__get_socket_aggr_by_cpu(struct perf_cpu cpu, void *data)
{
	struct perf_env *env = data;
	struct aggr_cpu_id id = aggr_cpu_id__empty();

	if (cpu.cpu != -1)
		id.socket = env->cpu[cpu.cpu].socket_id;

	return id;
}

static struct aggr_cpu_id perf_env__get_die_aggr_by_cpu(struct perf_cpu cpu, void *data)
{
	struct perf_env *env = data;
	struct aggr_cpu_id id = aggr_cpu_id__empty();

	if (cpu.cpu != -1) {
		/*
		 * die_id is relative to socket, so start
		 * with the socket ID and then add die to
		 * make a unique ID.
		 */
		id.socket = env->cpu[cpu.cpu].socket_id;
		id.die = env->cpu[cpu.cpu].die_id;
	}

	return id;
}

static void perf_env__get_cache_id_for_cpu(struct perf_cpu cpu, struct perf_env *env,
					   u32 cache_level, struct aggr_cpu_id *id)
{
	int i;
	int caches_cnt = env->caches_cnt;
	struct cpu_cache_level *caches = env->caches;

	id->cache_lvl = (cache_level > MAX_CACHE_LVL) ? 0 : cache_level;
	id->cache = -1;

	if (!caches_cnt)
		return;

	for (i = caches_cnt - 1; i > -1; --i) {
		struct perf_cpu_map *cpu_map;
		int map_contains_cpu;

		/*
		 * If user has not specified a level, find the fist level with
		 * the cpu in the map. Since building the map is expensive, do
		 * this only if levels match.
		 */
		if (cache_level <= MAX_CACHE_LVL && caches[i].level != cache_level)
			continue;

		cpu_map = perf_cpu_map__new(caches[i].map);
		map_contains_cpu = perf_cpu_map__idx(cpu_map, cpu);
		perf_cpu_map__put(cpu_map);

		if (map_contains_cpu != -1) {
			id->cache_lvl = caches[i].level;
			id->cache = cpu__get_cache_id_from_map(cpu, caches[i].map);
			return;
		}
	}
}

static struct aggr_cpu_id perf_env__get_cache_aggr_by_cpu(struct perf_cpu cpu,
							  void *data)
{
	struct perf_env *env = data;
	struct aggr_cpu_id id = aggr_cpu_id__empty();

	if (cpu.cpu != -1) {
		u32 cache_level = (perf_stat.aggr_level) ?: stat_config.aggr_level;

		id.socket = env->cpu[cpu.cpu].socket_id;
		id.die = env->cpu[cpu.cpu].die_id;
		perf_env__get_cache_id_for_cpu(cpu, env, cache_level, &id);
	}

	return id;
}

static struct aggr_cpu_id perf_env__get_cluster_aggr_by_cpu(struct perf_cpu cpu,
							    void *data)
{
	struct perf_env *env = data;
	struct aggr_cpu_id id = aggr_cpu_id__empty();

	if (cpu.cpu != -1) {
		id.socket = env->cpu[cpu.cpu].socket_id;
		id.die = env->cpu[cpu.cpu].die_id;
		id.cluster = env->cpu[cpu.cpu].cluster_id;
	}

	return id;
}

static struct aggr_cpu_id perf_env__get_core_aggr_by_cpu(struct perf_cpu cpu, void *data)
{
	struct perf_env *env = data;
	struct aggr_cpu_id id = aggr_cpu_id__empty();

	if (cpu.cpu != -1) {
		/*
		 * core_id is relative to socket, die and cluster, we need a
		 * global id. So we set socket, die id, cluster id and core id.
		 */
		id.socket = env->cpu[cpu.cpu].socket_id;
		id.die = env->cpu[cpu.cpu].die_id;
		id.cluster = env->cpu[cpu.cpu].cluster_id;
		id.core = env->cpu[cpu.cpu].core_id;
	}

	return id;
}

static struct aggr_cpu_id perf_env__get_cpu_aggr_by_cpu(struct perf_cpu cpu, void *data)
{
	struct perf_env *env = data;
	struct aggr_cpu_id id = aggr_cpu_id__empty();

	if (cpu.cpu != -1) {
		/*
		 * core_id is relative to socket and die,
		 * we need a global id. So we set
		 * socket, die id and core id
		 */
		id.socket = env->cpu[cpu.cpu].socket_id;
		id.die = env->cpu[cpu.cpu].die_id;
		id.core = env->cpu[cpu.cpu].core_id;
		id.cpu = cpu;
	}

	return id;
}

static struct aggr_cpu_id perf_env__get_node_aggr_by_cpu(struct perf_cpu cpu, void *data)
{
	struct aggr_cpu_id id = aggr_cpu_id__empty();

	id.node = perf_env__numa_node(data, cpu);
	return id;
}

static struct aggr_cpu_id perf_env__get_global_aggr_by_cpu(struct perf_cpu cpu __maybe_unused,
							   void *data __maybe_unused)
{
	struct aggr_cpu_id id = aggr_cpu_id__empty();

	/* it always aggregates to the cpu 0 */
	id.cpu = (struct perf_cpu){ .cpu = 0 };
	return id;
}

static struct aggr_cpu_id perf_stat__get_socket_file(struct perf_stat_config *config __maybe_unused,
						     struct perf_cpu cpu)
{
	return perf_env__get_socket_aggr_by_cpu(cpu, perf_session__env(perf_stat.session));
}
static struct aggr_cpu_id perf_stat__get_die_file(struct perf_stat_config *config __maybe_unused,
						  struct perf_cpu cpu)
{
	return perf_env__get_die_aggr_by_cpu(cpu, perf_session__env(perf_stat.session));
}

static struct aggr_cpu_id perf_stat__get_cluster_file(struct perf_stat_config *config __maybe_unused,
						      struct perf_cpu cpu)
{
	return perf_env__get_cluster_aggr_by_cpu(cpu, perf_session__env(perf_stat.session));
}

static struct aggr_cpu_id perf_stat__get_cache_file(struct perf_stat_config *config __maybe_unused,
						    struct perf_cpu cpu)
{
	return perf_env__get_cache_aggr_by_cpu(cpu, perf_session__env(perf_stat.session));
}

static struct aggr_cpu_id perf_stat__get_core_file(struct perf_stat_config *config __maybe_unused,
						   struct perf_cpu cpu)
{
	return perf_env__get_core_aggr_by_cpu(cpu, perf_session__env(perf_stat.session));
}

static struct aggr_cpu_id perf_stat__get_cpu_file(struct perf_stat_config *config __maybe_unused,
						  struct perf_cpu cpu)
{
	return perf_env__get_cpu_aggr_by_cpu(cpu, perf_session__env(perf_stat.session));
}

static struct aggr_cpu_id perf_stat__get_node_file(struct perf_stat_config *config __maybe_unused,
						   struct perf_cpu cpu)
{
	return perf_env__get_node_aggr_by_cpu(cpu, perf_session__env(perf_stat.session));
}

static struct aggr_cpu_id perf_stat__get_global_file(struct perf_stat_config *config __maybe_unused,
						     struct perf_cpu cpu)
{
	return perf_env__get_global_aggr_by_cpu(cpu, perf_session__env(perf_stat.session));
}

static aggr_cpu_id_get_t aggr_mode__get_aggr_file(enum aggr_mode aggr_mode)
{
	switch (aggr_mode) {
	case AGGR_SOCKET:
		return perf_env__get_socket_aggr_by_cpu;
	case AGGR_DIE:
		return perf_env__get_die_aggr_by_cpu;
	case AGGR_CLUSTER:
		return perf_env__get_cluster_aggr_by_cpu;
	case AGGR_CACHE:
		return perf_env__get_cache_aggr_by_cpu;
	case AGGR_CORE:
		return perf_env__get_core_aggr_by_cpu;
	case AGGR_NODE:
		return perf_env__get_node_aggr_by_cpu;
	case AGGR_GLOBAL:
		return perf_env__get_global_aggr_by_cpu;
	case AGGR_NONE:
		return perf_env__get_cpu_aggr_by_cpu;
	case AGGR_THREAD:
	case AGGR_UNSET:
	case AGGR_MAX:
	default:
		return NULL;
	}
}

static aggr_get_id_t aggr_mode__get_id_file(enum aggr_mode aggr_mode)
{
	switch (aggr_mode) {
	case AGGR_SOCKET:
		return perf_stat__get_socket_file;
	case AGGR_DIE:
		return perf_stat__get_die_file;
	case AGGR_CLUSTER:
		return perf_stat__get_cluster_file;
	case AGGR_CACHE:
		return perf_stat__get_cache_file;
	case AGGR_CORE:
		return perf_stat__get_core_file;
	case AGGR_NODE:
		return perf_stat__get_node_file;
	case AGGR_GLOBAL:
		return perf_stat__get_global_file;
	case AGGR_NONE:
		return perf_stat__get_cpu_file;
	case AGGR_THREAD:
	case AGGR_UNSET:
	case AGGR_MAX:
	default:
		return NULL;
	}
}

static int perf_stat_init_aggr_mode_file(struct perf_stat *st)
{
	struct perf_env *env = perf_session__env(st->session);
	aggr_cpu_id_get_t get_id = aggr_mode__get_aggr_file(stat_config.aggr_mode);
	bool needs_sort = stat_config.aggr_mode != AGGR_NONE;

	if (stat_config.aggr_mode == AGGR_THREAD) {
		int nr = perf_thread_map__nr(evsel_list->core.threads);

		stat_config.aggr_map = cpu_aggr_map__empty_new(nr);
		if (stat_config.aggr_map == NULL)
			return -ENOMEM;

		for (int s = 0; s < nr; s++) {
			struct aggr_cpu_id id = aggr_cpu_id__empty();

			id.thread_idx = s;
			stat_config.aggr_map->map[s] = id;
		}
		return 0;
	}

	if (!get_id)
		return 0;

	stat_config.aggr_map = cpu_aggr_map__new(evsel_list->core.user_requested_cpus,
						 get_id, env, needs_sort);
	if (!stat_config.aggr_map) {
		pr_err("cannot build %s map\n", aggr_mode__string[stat_config.aggr_mode]);
		return -1;
	}
	stat_config.aggr_get_id = aggr_mode__get_id_file(stat_config.aggr_mode);
	return 0;
}

static int default_evlist_evsel_cmp(void *priv __maybe_unused,
				    const struct list_head *l,
				    const struct list_head *r)
{
	const struct perf_evsel *lhs_core = container_of(l, struct perf_evsel, node);
	const struct evsel *lhs = container_of(lhs_core, struct evsel, core);
	const struct perf_evsel *rhs_core = container_of(r, struct perf_evsel, node);
	const struct evsel *rhs = container_of(rhs_core, struct evsel, core);

	if (evsel__leader(lhs) == evsel__leader(rhs)) {
		/* Within the same group, respect the original order. */
		return lhs_core->idx - rhs_core->idx;
	}

	/* Sort default metrics evsels first, and default show events before those. */
	if (lhs->default_metricgroup != rhs->default_metricgroup)
		return lhs->default_metricgroup ? -1 : 1;

	if (lhs->default_show_events != rhs->default_show_events)
		return lhs->default_show_events ? -1 : 1;

	/* Sort by PMU type (prefers legacy types first). */
	if (lhs->pmu != rhs->pmu)
		return lhs->pmu->type - rhs->pmu->type;

	/* Sort by name. */
	return strcmp(evsel__name((struct evsel *)lhs), evsel__name((struct evsel *)rhs));
}

/*
 * Add default events, if there were no attributes specified or
 * if -d/--detailed, -d -d or -d -d -d is used:
 */
static int add_default_events(void)
{
	const char *pmu = parse_events_option_args.pmu_filter ?: "all";
	struct parse_events_error err;
	struct evlist *evlist = evlist__new();
	struct evsel *evsel;
	int ret = 0;

	if (!evlist)
		return -ENOMEM;

	parse_events_error__init(&err);

	/* Set attrs if no event is selected and !null_run: */
	if (stat_config.null_run)
		goto out;

	if (transaction_run) {
		/* Handle -T as -M transaction. Once platform specific metrics
		 * support has been added to the json files, all architectures
		 * will use this approach. To determine transaction support
		 * on an architecture test for such a metric name.
		 */
		if (!metricgroup__has_metric_or_groups(pmu, "transaction")) {
			pr_err("Missing transaction metrics\n");
			ret = -1;
			goto out;
		}
		ret = metricgroup__parse_groups(evlist, pmu, "transaction",
						stat_config.metric_no_group,
						stat_config.metric_no_merge,
						stat_config.metric_no_threshold,
						stat_config.user_requested_cpu_list,
						stat_config.system_wide,
						stat_config.hardware_aware_grouping);
		goto out;
	}

	if (smi_cost) {
		int smi;

		if (sysfs__read_int(FREEZE_ON_SMI_PATH, &smi) < 0) {
			pr_err("freeze_on_smi is not supported.\n");
			ret = -1;
			goto out;
		}

		if (!smi) {
			if (sysfs__write_int(FREEZE_ON_SMI_PATH, 1) < 0) {
				pr_err("Failed to set freeze_on_smi.\n");
				ret = -1;
				goto out;
			}
			smi_reset = true;
		}

		if (!metricgroup__has_metric_or_groups(pmu, "smi")) {
			pr_err("Missing smi metrics\n");
			ret = -1;
			goto out;
		}

		if (!force_metric_only)
			stat_config.metric_only = true;

		ret = metricgroup__parse_groups(evlist, pmu, "smi",
						stat_config.metric_no_group,
						stat_config.metric_no_merge,
						stat_config.metric_no_threshold,
						stat_config.user_requested_cpu_list,
						stat_config.system_wide,
						stat_config.hardware_aware_grouping);
		goto out;
	}

	if (topdown_run) {
		unsigned int max_level = metricgroups__topdown_max_level();
		char str[] = "TopdownL1";

		if (!force_metric_only)
			stat_config.metric_only = true;

		if (!max_level) {
			pr_err("Topdown requested but the topdown metric groups aren't present.\n"
				"(See perf list the metric groups have names like TopdownL1)\n");
			ret = -1;
			goto out;
		}
		if (stat_config.topdown_level > max_level) {
			pr_err("Invalid top-down metrics level. The max level is %u.\n", max_level);
			ret = -1;
			goto out;
		} else if (!stat_config.topdown_level) {
			stat_config.topdown_level = 1;
		}
		if (!stat_config.interval && !stat_config.metric_only) {
			fprintf(stat_config.output,
				"Topdown accuracy may decrease when measuring long periods.\n"
				"Please print the result regularly, e.g. -I1000\n");
		}
		str[8] = stat_config.topdown_level + '0';
		if (metricgroup__parse_groups(evlist,
						pmu, str,
						/*metric_no_group=*/false,
						/*metric_no_merge=*/false,
						/*metric_no_threshold=*/true,
						stat_config.user_requested_cpu_list,
						stat_config.system_wide,
						stat_config.hardware_aware_grouping) < 0) {
			ret = -1;
			goto out;
		}
	}

	if (!stat_config.topdown_level)
		stat_config.topdown_level = 1;

	if (!evlist->core.nr_entries && !evsel_list->core.nr_entries) {
		/*
		 * Add Default metrics. To minimize multiplexing, don't request
		 * threshold computation, but it will be computed if the events
		 * are present.
		 */
		const char *default_metricgroup_names[] = {
			"Default", "Default2", "Default3", "Default4",
		};

		for (size_t i = 0; i < ARRAY_SIZE(default_metricgroup_names); i++) {
			struct evlist *metric_evlist;

			if (!metricgroup__has_metric_or_groups(pmu, default_metricgroup_names[i]))
				continue;

			if ((int)i > detailed_run)
				break;

			metric_evlist = evlist__new();
			if (!metric_evlist) {
				ret = -ENOMEM;
				break;
			}
			if (metricgroup__parse_groups(metric_evlist, pmu, default_metricgroup_names[i],
							/*metric_no_group=*/false,
							/*metric_no_merge=*/false,
							/*metric_no_threshold=*/true,
							stat_config.user_requested_cpu_list,
							stat_config.system_wide,
							stat_config.hardware_aware_grouping) < 0) {
				evlist__delete(metric_evlist);
				ret = -1;
				break;
			}

			evlist__for_each_entry(metric_evlist, evsel)
				evsel->default_metricgroup = true;

			evlist__splice_list_tail(evlist, &metric_evlist->core.entries);
			metricgroup__copy_metric_events(evlist, /*cgrp=*/NULL,
							&evlist->metric_events,
							&metric_evlist->metric_events);
			evlist__delete(metric_evlist);
		}
		list_sort(/*priv=*/NULL, &evlist->core.entries, default_evlist_evsel_cmp);

	}
out:
	if (!ret) {
		evlist__for_each_entry(evlist, evsel) {
			/*
			 * Make at least one event non-skippable so fatal errors are visible.
			 * 'cycles' always used to be default and non-skippable, so use that.
			 */
			if (!evsel__match(evsel, HARDWARE, HW_CPU_CYCLES))
				evsel->skippable = true;
		}
	}
	parse_events_error__exit(&err);
	evlist__splice_list_tail(evsel_list, &evlist->core.entries);
	metricgroup__copy_metric_events(evsel_list, /*cgrp=*/NULL,
					&evsel_list->metric_events,
					&evlist->metric_events);
	evlist__delete(evlist);
	return ret;
}

static const char * const stat_record_usage[] = {
	"perf stat record [<options>]",
	NULL,
};

static void init_features(struct perf_session *session)
{
	int feat;

	for (feat = HEADER_FIRST_FEATURE; feat < HEADER_LAST_FEATURE; feat++)
		perf_header__set_feat(&session->header, feat);

	perf_header__clear_feat(&session->header, HEADER_DIR_FORMAT);
	perf_header__clear_feat(&session->header, HEADER_BUILD_ID);
	perf_header__clear_feat(&session->header, HEADER_TRACING_DATA);
	perf_header__clear_feat(&session->header, HEADER_BRANCH_STACK);
	perf_header__clear_feat(&session->header, HEADER_AUXTRACE);
}

static int __cmd_record(const struct option stat_options[], struct opt_aggr_mode *opt_mode,
			int argc, const char **argv)
{
	struct perf_session *session;
	struct perf_data *data = &perf_stat.data;

	argc = parse_options(argc, argv, stat_options, stat_record_usage,
			     PARSE_OPT_STOP_AT_NON_OPTION);
	stat_config.aggr_mode = opt_aggr_mode_to_aggr_mode(opt_mode);

	if (output_name)
		data->path = output_name;

	if (stat_config.run_count != 1 || forever) {
		pr_err("Cannot use -r option with perf stat record.\n");
		return -1;
	}

	session = perf_session__new(data, NULL);
	if (IS_ERR(session)) {
		pr_err("Perf session creation failed\n");
		return PTR_ERR(session);
	}

	init_features(session);

	session->evlist   = evsel_list;
	perf_stat.session = session;
	perf_stat.record  = true;
	return argc;
}

static int process_stat_round_event(const struct perf_tool *tool __maybe_unused,
				    struct perf_session *session,
				    union perf_event *event)
{
	struct perf_record_stat_round *stat_round = &event->stat_round;
	struct timespec tsh, *ts = NULL;
	struct perf_env *env = perf_session__env(session);
	const char **argv = env->cmdline_argv;
	int argc = env->nr_cmdline;

	process_counters();

	if (stat_round->type == PERF_STAT_ROUND_TYPE__FINAL)
		update_stats(stat_config.walltime_nsecs_stats, stat_round->time);

	if (stat_config.interval && stat_round->time) {
		tsh.tv_sec  = stat_round->time / NSEC_PER_SEC;
		tsh.tv_nsec = stat_round->time % NSEC_PER_SEC;
		ts = &tsh;
	}

	print_counters(ts, argc, argv);
	return 0;
}

static
int process_stat_config_event(const struct perf_tool *tool,
			      struct perf_session *session,
			      union perf_event *event)
{
	struct perf_stat *st = container_of(tool, struct perf_stat, tool);

	perf_event__read_stat_config(&stat_config, &event->stat_config);

	if (perf_cpu_map__is_empty(st->cpus)) {
		if (st->aggr_mode != AGGR_UNSET)
			pr_warning("warning: processing task data, aggregation mode not set\n");
	} else if (st->aggr_mode != AGGR_UNSET) {
		stat_config.aggr_mode = st->aggr_mode;
	}

	if (perf_stat.data.is_pipe)
		perf_stat_init_aggr_mode();
	else
		perf_stat_init_aggr_mode_file(st);

	if (stat_config.aggr_map) {
		int nr_aggr = stat_config.aggr_map->nr;

		if (evlist__alloc_aggr_stats(session->evlist, nr_aggr) < 0) {
			pr_err("cannot allocate aggr counts\n");
			return -1;
		}
	}
	return 0;
}

static int set_maps(struct perf_stat *st)
{
	if (!st->cpus || !st->threads)
		return 0;

	if (WARN_ONCE(st->maps_allocated, "stats double allocation\n"))
		return -EINVAL;

	perf_evlist__set_maps(&evsel_list->core, st->cpus, st->threads);

	if (evlist__alloc_stats(&stat_config, evsel_list, /*alloc_raw=*/true))
		return -ENOMEM;

	st->maps_allocated = true;
	return 0;
}

static
int process_thread_map_event(const struct perf_tool *tool,
			     struct perf_session *session __maybe_unused,
			     union perf_event *event)
{
	struct perf_stat *st = container_of(tool, struct perf_stat, tool);

	if (st->threads) {
		pr_warning("Extra thread map event, ignoring.\n");
		return 0;
	}

	st->threads = thread_map__new_event(&event->thread_map);
	if (!st->threads)
		return -ENOMEM;

	return set_maps(st);
}

static
int process_cpu_map_event(const struct perf_tool *tool,
			  struct perf_session *session __maybe_unused,
			  union perf_event *event)
{
	struct perf_stat *st = container_of(tool, struct perf_stat, tool);
	struct perf_cpu_map *cpus;

	if (st->cpus) {
		pr_warning("Extra cpu map event, ignoring.\n");
		return 0;
	}

	cpus = cpu_map__new_data(&event->cpu_map.data);
	if (!cpus)
		return -ENOMEM;

	st->cpus = cpus;
	return set_maps(st);
}

static const char * const stat_report_usage[] = {
	"perf stat report [<options>]",
	NULL,
};

static struct perf_stat perf_stat = {
	.aggr_mode	= AGGR_UNSET,
	.aggr_level	= 0,
};

static int __cmd_report(int argc, const char **argv)
{
	struct perf_session *session;
	const struct option options[] = {
	OPT_STRING('i', "input", &input_name, "file", "input file name"),
	OPT_SET_UINT(0, "per-socket", &perf_stat.aggr_mode,
		     "aggregate counts per processor socket", AGGR_SOCKET),
	OPT_SET_UINT(0, "per-die", &perf_stat.aggr_mode,
		     "aggregate counts per processor die", AGGR_DIE),
	OPT_SET_UINT(0, "per-cluster", &perf_stat.aggr_mode,
		     "aggregate counts perf processor cluster", AGGR_CLUSTER),
	OPT_CALLBACK_OPTARG(0, "per-cache", &perf_stat.aggr_mode, &perf_stat.aggr_level,
			    "cache level",
			    "aggregate count at this cache level (Default: LLC)",
			    parse_cache_level),
	OPT_SET_UINT(0, "per-core", &perf_stat.aggr_mode,
		     "aggregate counts per physical processor core", AGGR_CORE),
	OPT_SET_UINT(0, "per-node", &perf_stat.aggr_mode,
		     "aggregate counts per numa node", AGGR_NODE),
	OPT_SET_UINT('A', "no-aggr", &perf_stat.aggr_mode,
		     "disable CPU count aggregation", AGGR_NONE),
	OPT_END()
	};
	struct stat st;
	int ret;

	argc = parse_options(argc, argv, options, stat_report_usage, 0);

	if (!input_name || !strlen(input_name)) {
		if (!fstat(STDIN_FILENO, &st) && S_ISFIFO(st.st_mode))
			input_name = "-";
		else
			input_name = "perf.data";
	}

	perf_stat.data.path = input_name;
	perf_stat.data.mode = PERF_DATA_MODE_READ;

	perf_tool__init(&perf_stat.tool, /*ordered_events=*/false);
	perf_stat.tool.attr		= perf_event__process_attr;
	perf_stat.tool.event_update	= perf_event__process_event_update;
	perf_stat.tool.thread_map	= process_thread_map_event;
	perf_stat.tool.cpu_map		= process_cpu_map_event;
	perf_stat.tool.stat_config	= process_stat_config_event;
	perf_stat.tool.stat		= perf_event__process_stat_event;
	perf_stat.tool.stat_round	= process_stat_round_event;

	session = perf_session__new(&perf_stat.data, &perf_stat.tool);
	if (IS_ERR(session))
		return PTR_ERR(session);

	perf_stat.session  = session;
	stat_config.output = stderr;
	evlist__delete(evsel_list);
	evsel_list         = session->evlist;

	ret = perf_session__process_events(session);
	if (ret)
		return ret;

	perf_session__delete(session);
	return 0;
}

static void setup_system_wide(int forks)
{
	/*
	 * Make system wide (-a) the default target if
	 * no target was specified and one of following
	 * conditions is met:
	 *
	 *   - there's no workload specified
	 *   - there is workload specified but all requested
	 *     events are system wide events
	 */
	if (!target__none(&target))
		return;

	if (!forks)
		target.system_wide = true;
	else {
		struct evsel *counter;

		evlist__for_each_entry(evsel_list, counter) {
			if (!counter->core.requires_cpu &&
			    !evsel__name_is(counter, "duration_time")) {
				return;
			}
		}

		if (evsel_list->core.nr_entries)
			target.system_wide = true;
	}
}

#ifdef HAVE_ARCH_X86_64_SUPPORT
static int parse_tpebs_mode(const struct option *opt, const char *str,
			    int unset __maybe_unused)
{
	enum tpebs_mode *mode = opt->value;

	if (!strcasecmp("mean", str)) {
		*mode = TPEBS_MODE__MEAN;
		return 0;
	}
	if (!strcasecmp("min", str)) {
		*mode = TPEBS_MODE__MIN;
		return 0;
	}
	if (!strcasecmp("max", str)) {
		*mode = TPEBS_MODE__MAX;
		return 0;
	}
	if (!strcasecmp("last", str)) {
		*mode = TPEBS_MODE__LAST;
		return 0;
	}
	return -1;
}
#endif // HAVE_ARCH_X86_64_SUPPORT

int cmd_stat(int argc, const char **argv)
{
	struct opt_aggr_mode opt_mode = {};
	bool affinity = true, affinity_set = false;
	struct option stat_options[] = {
		OPT_BOOLEAN('T', "transaction", &transaction_run,
			"hardware transaction statistics"),
		OPT_CALLBACK('e', "event", &parse_events_option_args, "event",
			"event selector. use 'perf list' to list available events",
			parse_events_option),
		OPT_CALLBACK(0, "filter", &evsel_list, "filter",
			"event filter", parse_filter),
		OPT_BOOLEAN('i', "no-inherit", &stat_config.no_inherit,
			"child tasks do not inherit counters"),
		OPT_STRING('p', "pid", &target.pid, "pid",
			"stat events on existing process id"),
		OPT_STRING('t', "tid", &target.tid, "tid",
			"stat events on existing thread id"),
#ifdef HAVE_BPF_SKEL
		OPT_STRING('b', "bpf-prog", &target.bpf_str, "bpf-prog-id",
			"stat events on existing bpf program id"),
		OPT_BOOLEAN(0, "bpf-counters", &target.use_bpf,
			"use bpf program to count events"),
		OPT_STRING(0, "bpf-attr-map", &target.attr_map, "attr-map-path",
			"path to perf_event_attr map"),
#endif
		OPT_BOOLEAN('a', "all-cpus", &target.system_wide,
			"system-wide collection from all CPUs"),
		OPT_BOOLEAN(0, "scale", &stat_config.scale,
			"Use --no-scale to disable counter scaling for multiplexing"),
		OPT_INCR('v', "verbose", &verbose,
			"be more verbose (show counter open errors, etc)"),
		OPT_INTEGER('r', "repeat", &stat_config.run_count,
			"repeat command and print average + stddev (max: 100, forever: 0)"),
		OPT_BOOLEAN(0, "table", &stat_config.walltime_run_table,
			"display details about each run (only with -r option)"),
		OPT_BOOLEAN('n', "null", &stat_config.null_run,
			"null run - dont start any counters"),
		OPT_INCR('d', "detailed", &detailed_run,
			"detailed run - start a lot of events"),
		OPT_BOOLEAN('S', "sync", &sync_run,
			"call sync() before starting a run"),
		OPT_CALLBACK_NOOPT('B', "big-num", NULL, NULL,
				"print large numbers with thousands\' separators",
				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", &opt_mode.no_aggr,
			"disable aggregation across CPUs or PMUs"),
		OPT_BOOLEAN(0, "no-merge", &opt_mode.no_aggr,
			"disable aggregation the same as -A or -no-aggr"),
		OPT_BOOLEAN(0, "hybrid-merge", &stat_config.hybrid_merge,
			"Merge identical named hybrid events"),
		OPT_STRING('x', "field-separator", &stat_config.csv_sep, "separator",
			"print counts with custom separator"),
		OPT_BOOLEAN('j', "json-output", &stat_config.json_output,
			"print counts in JSON format"),
		OPT_CALLBACK('G', "cgroup", &evsel_list, "name",
			"monitor event in cgroup name only", parse_stat_cgroups),
		OPT_STRING(0, "for-each-cgroup", &stat_config.cgroup_list, "name",
			"expand events for each cgroup"),
		OPT_STRING('o', "output", &output_name, "file", "output file name"),
		OPT_BOOLEAN(0, "append", &append_file, "append to the output file"),
		OPT_INTEGER(0, "log-fd", &output_fd,
			"log output to fd, instead of stderr"),
		OPT_STRING(0, "pre", &pre_cmd, "command",
			"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", &stat_config.interval,
			"print counts at regular interval in ms "
			"(overhead is possible for values <= 100ms)"),
		OPT_INTEGER(0, "interval-count", &stat_config.times,
			"print counts for fixed number of times"),
		OPT_BOOLEAN(0, "interval-clear", &stat_config.interval_clear,
			"clear screen in between new interval"),
		OPT_UINTEGER(0, "timeout", &stat_config.timeout,
			"stop workload and print counts after a timeout period in ms (>= 10ms)"),
		OPT_BOOLEAN(0, "per-socket", &opt_mode.socket,
			"aggregate counts per processor socket"),
		OPT_BOOLEAN(0, "per-die", &opt_mode.die, "aggregate counts per processor die"),
		OPT_BOOLEAN(0, "per-cluster", &opt_mode.cluster,
			"aggregate counts per processor cluster"),
		OPT_CALLBACK_OPTARG(0, "per-cache", &opt_mode, &stat_config.aggr_level,
				"cache level", "aggregate count at this cache level (Default: LLC)",
				parse_cache_level),
		OPT_BOOLEAN(0, "per-core", &opt_mode.core,
			"aggregate counts per physical processor core"),
		OPT_BOOLEAN(0, "per-thread", &opt_mode.thread, "aggregate counts per thread"),
		OPT_BOOLEAN(0, "per-node", &opt_mode.node, "aggregate counts per numa node"),
		OPT_INTEGER('D', "delay", &target.initial_delay,
			"ms to wait before starting measurement after program start (-1: start with events disabled)"),
		OPT_CALLBACK_NOOPT(0, "metric-only", &stat_config.metric_only, NULL,
				"Only print computed metrics. No raw values", enable_metric_only),
		OPT_BOOLEAN(0, "metric-no-group", &stat_config.metric_no_group,
			"don't group metric events, impacts multiplexing"),
		OPT_BOOLEAN(0, "metric-no-merge", &stat_config.metric_no_merge,
			"don't try to share events between metrics in a group"),
		OPT_BOOLEAN(0, "metric-no-threshold", &stat_config.metric_no_threshold,
			"disable adding events for the metric threshold calculation"),
		OPT_BOOLEAN(0, "topdown", &topdown_run,
			"measure top-down statistics"),
#ifdef HAVE_ARCH_X86_64_SUPPORT
		OPT_BOOLEAN(0, "record-tpebs", &tpebs_recording,
			"enable recording for tpebs when retire_latency required"),
		OPT_CALLBACK(0, "tpebs-mode", &tpebs_mode, "tpebs-mode",
			"Mode of TPEBS recording: mean, min or max",
			parse_tpebs_mode),
#endif
		OPT_UINTEGER(0, "td-level", &stat_config.topdown_level,
			"Set the metrics level for the top-down statistics (0: max level)"),
		OPT_BOOLEAN(0, "smi-cost", &smi_cost,
			"measure SMI cost"),
		OPT_CALLBACK('M', "metrics", &evsel_list, "metric/metric group list",
			"monitor specified metrics or metric groups (separated by ,)",
			append_metric_groups),
		OPT_BOOLEAN_FLAG(0, "all-kernel", &stat_config.all_kernel,
				"Configure all used events to run in kernel space.",
				PARSE_OPT_EXCLUSIVE),
		OPT_BOOLEAN_FLAG(0, "all-user", &stat_config.all_user,
				"Configure all used events to run in user space.",
				PARSE_OPT_EXCLUSIVE),
		OPT_BOOLEAN(0, "percore-show-thread", &stat_config.percore_show_thread,
			"Use with 'percore' event qualifier to show the event "
			"counts of one hardware thread by sum up total hardware "
			"threads of same physical core"),
		OPT_BOOLEAN(0, "summary", &stat_config.summary,
			"print summary for interval mode"),
		OPT_BOOLEAN(0, "no-csv-summary", &stat_config.no_csv_summary,
			"don't print 'summary' for CSV summary output"),
		OPT_BOOLEAN(0, "quiet", &quiet,
			"don't print any output, messages or warnings (useful with record)"),
		OPT_BOOLEAN_SET(0, "affinity", &affinity, &affinity_set,
			"enable (default) or disable affinity optimizations to reduce IPIs"),
		OPT_CALLBACK(0, "cputype", &evsel_list, "hybrid cpu type",
			"Only enable events on applying cpu with this type "
			"for hybrid platform (e.g. core or atom)",
			parse_cputype),
#ifdef HAVE_LIBPFM
		OPT_CALLBACK(0, "pfm-events", &evsel_list, "event",
			"libpfm4 event selector. use 'perf list' to list available events",
			parse_libpfm_events_option),
#endif
		OPT_CALLBACK(0, "control", &stat_config, "fd:ctl-fd[,ack-fd] or fifo:ctl-fifo[,ack-fifo]",
			"Listen on ctl-fd descriptor for command to control measurement ('enable': enable events, 'disable': disable events).\n"
			"\t\t\t  Optionally send control command completion ('ack\\n') to ack-fd descriptor.\n"
			"\t\t\t  Alternatively, ctl-fifo / ack-fifo will be opened and used as ctl-fd / ack-fd.",
			parse_control_option),
		OPT_CALLBACK_OPTARG(0, "iostat", &evsel_list, &stat_config, "default",
				"measure I/O performance metrics provided by arch/platform",
				iostat_parse),
		OPT_END()
	};
	const char * const stat_usage[] = {
		"perf stat [<options>] [<command>]",
		NULL
	};
	int status = -EINVAL, run_idx, err;
	const char *mode;
	FILE *output = stderr;
	unsigned int interval, timeout;
	const char * const stat_subcommands[] = { "record", "report" };
	char errbuf[BUFSIZ];
	struct evsel *counter;

	setlocale(LC_ALL, "");

	evsel_list = evlist__new();
	if (evsel_list == NULL)
		return -ENOMEM;

	parse_events__shrink_config_terms();

	/* String-parsing callback-based options would segfault when negated */
	set_option_flag(stat_options, 'e', "event", PARSE_OPT_NONEG);
	set_option_flag(stat_options, 'M', "metrics", PARSE_OPT_NONEG);
	set_option_flag(stat_options, 'G', "cgroup", PARSE_OPT_NONEG);

	argc = parse_options_subcommand(argc, argv, stat_options, stat_subcommands,
					(const char **) stat_usage,
					PARSE_OPT_STOP_AT_NON_OPTION);

	stat_config.aggr_mode = opt_aggr_mode_to_aggr_mode(&opt_mode);

	if (stat_config.csv_sep) {
		stat_config.csv_output = true;
		if (!strcmp(stat_config.csv_sep, "\\t"))
			stat_config.csv_sep = "\t";
	} else
		stat_config.csv_sep = DEFAULT_SEPARATOR;

	if (affinity_set)
		evsel_list->no_affinity = !affinity;

	if (argc && strlen(argv[0]) > 2 && strstarts("record", argv[0])) {
		argc = __cmd_record(stat_options, &opt_mode, argc, argv);
		if (argc < 0)
			return -1;
	} else if (argc && strlen(argv[0]) > 2 && strstarts("report", argv[0]))
		return __cmd_report(argc, argv);

	interval = stat_config.interval;
	timeout = stat_config.timeout;

	/*
	 * For record command the -o is already taken care of.
	 */
	if (!STAT_RECORD && output_name && strcmp(output_name, "-"))
		output = NULL;

	if (output_name && output_fd) {
		fprintf(stderr, "cannot use both --output and --log-fd\n");
		parse_options_usage(stat_usage, stat_options, "o", 1);
		parse_options_usage(NULL, stat_options, "log-fd", 0);
		goto out;
	}

	if (stat_config.metric_only && stat_config.aggr_mode == AGGR_THREAD) {
		fprintf(stderr, "--metric-only is not supported with --per-thread\n");
		goto out;
	}

	if (stat_config.metric_only && stat_config.run_count > 1) {
		fprintf(stderr, "--metric-only is not supported with -r\n");
		goto out;
	}

	if (stat_config.csv_output || (stat_config.metric_only && stat_config.json_output)) {
		/*
		 * Current CSV and metric-only JSON output doesn't display the
		 * metric threshold so don't compute it.
		 */
		stat_config.metric_no_threshold = true;
	}

	if (stat_config.walltime_run_table && stat_config.run_count <= 1) {
		fprintf(stderr, "--table is only supported with -r\n");
		parse_options_usage(stat_usage, stat_options, "r", 1);
		parse_options_usage(NULL, stat_options, "table", 0);
		goto out;
	}

	if (output_fd < 0) {
		fprintf(stderr, "argument to --log-fd must be a > 0\n");
		parse_options_usage(stat_usage, stat_options, "log-fd", 0);
		goto out;
	}

	if (!output && !quiet) {
		struct timespec tm;
		mode = append_file ? "a" : "w";

		output = fopen(output_name, mode);
		if (!output) {
			perror("failed to create output file");
			return -1;
		}
		if (!stat_config.json_output) {
			clock_gettime(CLOCK_REALTIME, &tm);
			fprintf(output, "# started on %s\n", ctime(&tm.tv_sec));
		}
	} else if (output_fd > 0) {
		mode = append_file ? "a" : "w";
		output = fdopen(output_fd, mode);
		if (!output) {
			perror("Failed opening logfd");
			return -errno;
		}
	}

	if (stat_config.interval_clear && !isatty(fileno(output))) {
		fprintf(stderr, "--interval-clear does not work with output\n");
		parse_options_usage(stat_usage, stat_options, "o", 1);
		parse_options_usage(NULL, stat_options, "log-fd", 0);
		parse_options_usage(NULL, stat_options, "interval-clear", 0);
		return -1;
	}

	stat_config.output = output;

	/*
	 * let the spreadsheet do the pretty-printing
	 */
	if (stat_config.csv_output) {
		/* User explicitly passed -B? */
		if (big_num_opt == 1) {
			fprintf(stderr, "-B option not supported with -x\n");
			parse_options_usage(stat_usage, stat_options, "B", 1);
			parse_options_usage(NULL, stat_options, "x", 1);
			goto out;
		} else /* Nope, so disable big number formatting */
			stat_config.big_num = false;
	} else if (big_num_opt == 0) /* User passed --no-big-num */
		stat_config.big_num = false;

	target.inherit = !stat_config.no_inherit;
	err = target__validate(&target);
	if (err) {
		target__strerror(&target, err, errbuf, BUFSIZ);
		pr_warning("%s\n", errbuf);
	}

	setup_system_wide(argc);

	/*
	 * Display user/system times only for single
	 * run and when there's specified tracee.
	 */
	if ((stat_config.run_count == 1) && target__none(&target))
		stat_config.ru_display = true;

	if (stat_config.run_count < 0) {
		pr_err("Run count must be a positive number\n");
		parse_options_usage(stat_usage, stat_options, "r", 1);
		goto out;
	} else if (stat_config.run_count == 0) {
		forever = true;
		stat_config.run_count = 1;
	}

	if (stat_config.walltime_run_table) {
		stat_config.walltime_run = zalloc(stat_config.run_count * sizeof(stat_config.walltime_run[0]));
		if (!stat_config.walltime_run) {
			pr_err("failed to setup -r option");
			goto out;
		}
	}

	if ((stat_config.aggr_mode == AGGR_THREAD) &&
		!target__has_task(&target)) {
		if (!target.system_wide || target.cpu_list) {
			fprintf(stderr, "The --per-thread option is only "
				"available when monitoring via -p -t -a "
				"options or only --per-thread.\n");
			parse_options_usage(NULL, stat_options, "p", 1);
			parse_options_usage(NULL, stat_options, "t", 1);
			goto out;
		}
	}

	/*
	 * no_aggr, cgroup are for system-wide only
	 * --per-thread is aggregated per thread, we dont mix it with cpu mode
	 */
	if (((stat_config.aggr_mode != AGGR_GLOBAL &&
	      stat_config.aggr_mode != AGGR_THREAD) ||
	     (nr_cgroups || stat_config.cgroup_list)) &&
	    !target__has_cpu(&target)) {
		fprintf(stderr, "both cgroup and no-aggregation "
			"modes only available in system-wide mode\n");

		parse_options_usage(stat_usage, stat_options, "G", 1);
		parse_options_usage(NULL, stat_options, "A", 1);
		parse_options_usage(NULL, stat_options, "a", 1);
		parse_options_usage(NULL, stat_options, "for-each-cgroup", 0);
		goto out;
	}

	if (stat_config.iostat_run) {
		status = iostat_prepare(evsel_list, &stat_config);
		if (status)
			goto out;
		if (iostat_mode == IOSTAT_LIST) {
			iostat_list(evsel_list, &stat_config);
			goto out;
		} else if (verbose > 0)
			iostat_list(evsel_list, &stat_config);
		if (iostat_mode == IOSTAT_RUN && !target__has_cpu(&target))
			target.system_wide = true;
	}

	if ((stat_config.aggr_mode == AGGR_THREAD) && (target.system_wide))
		target.per_thread = true;

	stat_config.system_wide = target.system_wide;
	if (target.cpu_list) {
		stat_config.user_requested_cpu_list = strdup(target.cpu_list);
		if (!stat_config.user_requested_cpu_list) {
			status = -ENOMEM;
			goto out;
		}
	}

	/*
	 * Metric parsing needs to be delayed as metrics may optimize events
	 * knowing the target is system-wide.
	 */
	if (metrics) {
		const char *pmu = parse_events_option_args.pmu_filter ?: "all";
		int ret = metricgroup__parse_groups(evsel_list, pmu, metrics,
						stat_config.metric_no_group,
						stat_config.metric_no_merge,
						stat_config.metric_no_threshold,
						stat_config.user_requested_cpu_list,
						stat_config.system_wide,
						stat_config.hardware_aware_grouping);

		zfree(&metrics);
		if (ret) {
			status = ret;
			goto out;
		}
	}

	if (add_default_events())
		goto out;

	if (stat_config.cgroup_list) {
		if (nr_cgroups > 0) {
			pr_err("--cgroup and --for-each-cgroup cannot be used together\n");
			parse_options_usage(stat_usage, stat_options, "G", 1);
			parse_options_usage(NULL, stat_options, "for-each-cgroup", 0);
			goto out;
		}

		if (evlist__expand_cgroup(evsel_list, stat_config.cgroup_list, true) < 0) {
			parse_options_usage(stat_usage, stat_options,
					    "for-each-cgroup", 0);
			goto out;
		}
	}
#ifdef HAVE_BPF_SKEL
	if (target.use_bpf && nr_cgroups &&
	    (evsel_list->core.nr_entries / nr_cgroups) > BPERF_CGROUP__MAX_EVENTS) {
		pr_warning("Disabling BPF counters due to more events (%d) than the max (%d)\n",
			   evsel_list->core.nr_entries / nr_cgroups, BPERF_CGROUP__MAX_EVENTS);
		target.use_bpf = false;
	}
#endif // HAVE_BPF_SKEL
	evlist__warn_user_requested_cpus(evsel_list, target.cpu_list);

	evlist__for_each_entry(evsel_list, counter) {
		/*
		 * Setup BPF counters to require CPUs as any(-1) isn't
		 * supported. evlist__create_maps below will propagate this
		 * information to the evsels. Note, evsel__is_bperf isn't yet
		 * set up, and this change must happen early, so directly use
		 * the bpf_counter variable and target information.
		 */
		if ((counter->bpf_counter || target.use_bpf) && !target__has_cpu(&target))
			counter->core.requires_cpu = true;
	}

	if (evlist__create_maps(evsel_list, &target) < 0) {
		if (target__has_task(&target)) {
			pr_err("Problems finding threads of monitor\n");
			parse_options_usage(stat_usage, stat_options, "p", 1);
			parse_options_usage(NULL, stat_options, "t", 1);
		} else if (target__has_cpu(&target)) {
			perror("failed to parse CPUs map");
			parse_options_usage(stat_usage, stat_options, "C", 1);
			parse_options_usage(NULL, stat_options, "a", 1);
		}
		goto out;
	}

	evlist__check_cpu_maps(evsel_list);

	/*
	 * Initialize thread_map with comm names,
	 * so we could print it out on output.
	 */
	if (stat_config.aggr_mode == AGGR_THREAD) {
		thread_map__read_comms(evsel_list->core.threads);
	}

	if (stat_config.aggr_mode == AGGR_NODE)
		cpu__setup_cpunode_map();

	if (stat_config.times && interval)
		interval_count = true;
	else if (stat_config.times && !interval) {
		pr_err("interval-count option should be used together with "
				"interval-print.\n");
		parse_options_usage(stat_usage, stat_options, "interval-count", 0);
		parse_options_usage(stat_usage, stat_options, "I", 1);
		goto out;
	}

	if (timeout && timeout < 100) {
		if (timeout < 10) {
			pr_err("timeout must be >= 10ms.\n");
			parse_options_usage(stat_usage, stat_options, "timeout", 0);
			goto out;
		} else
			pr_warning("timeout < 100ms. "
				   "The overhead percentage could be high in some cases. "
				   "Please proceed with caution.\n");
	}
	if (timeout && interval) {
		pr_err("timeout option is not supported with interval-print.\n");
		parse_options_usage(stat_usage, stat_options, "timeout", 0);
		parse_options_usage(stat_usage, stat_options, "I", 1);
		goto out;
	}

	if (perf_stat_init_aggr_mode())
		goto out;

	if (evlist__alloc_stats(&stat_config, evsel_list, interval))
		goto out;

	/*
	 * Set sample_type to PERF_SAMPLE_IDENTIFIER, which should be harmless
	 * while avoiding that older tools show confusing messages.
	 *
	 * However for pipe sessions we need to keep it zero,
	 * because script's perf_evsel__check_attr is triggered
	 * by attr->sample_type != 0, and we can't run it on
	 * stat sessions.
	 */
	stat_config.identifier = !(STAT_RECORD && perf_stat.data.is_pipe);

	/*
	 * We dont want to block the signals - that would cause
	 * child tasks to inherit that and Ctrl-C would not work.
	 * What we want is for Ctrl-C to work in the exec()-ed
	 * task, but being ignored by perf stat itself:
	 */
	atexit(sig_atexit);
	if (!forever)
		signal(SIGINT,  skip_signal);
	signal(SIGCHLD, skip_signal);
	signal(SIGALRM, skip_signal);
	signal(SIGABRT, skip_signal);

	if (evlist__initialize_ctlfd(evsel_list, stat_config.ctl_fd, stat_config.ctl_fd_ack))
		goto out;

	/* Enable ignoring missing threads when -p option is defined. */
	evlist__first(evsel_list)->ignore_missing_thread = target.pid;
	status = 0;
	for (run_idx = 0; forever || run_idx < stat_config.run_count; run_idx++) {
		if (stat_config.run_count != 1 && verbose > 0)
			fprintf(output, "[ perf stat: executing run #%d ... ]\n",
				run_idx + 1);

		if (run_idx != 0)
			evlist__reset_prev_raw_counts(evsel_list);

		status = run_perf_stat(argc, argv, run_idx);
		if (status < 0)
			break;

		if (forever && !interval) {
			print_counters(NULL, argc, argv);
			perf_stat__reset_stats();
		}
	}

	if (!forever && status != -1 && (!interval || stat_config.summary)) {
		if (stat_config.run_count > 1)
			evlist__copy_res_stats(&stat_config, evsel_list);
		print_counters(NULL, argc, argv);
	}

	evlist__finalize_ctlfd(evsel_list);

	if (STAT_RECORD) {
		/*
		 * We synthesize the kernel mmap record just so that older tools
		 * don't emit warnings about not being able to resolve symbols
		 * due to /proc/sys/kernel/kptr_restrict settings and instead provide
		 * a saner message about no samples being in the perf.data file.
		 *
		 * This also serves to suppress a warning about f_header.data.size == 0
		 * in header.c at the moment 'perf stat record' gets introduced, which
		 * is not really needed once we start adding the stat specific PERF_RECORD_
		 * records, but the need to suppress the kptr_restrict messages in older
		 * tools remain  -acme
		 */
		int fd = perf_data__fd(&perf_stat.data);

		err = perf_event__synthesize_kernel_mmap((void *)&perf_stat,
							 process_synthesized_event,
							 &perf_stat.session->machines.host);
		if (err) {
			pr_warning("Couldn't synthesize the kernel mmap record, harmless, "
				   "older tools may produce warnings about this file\n.");
		}

		if (!interval) {
			if (WRITE_STAT_ROUND_EVENT(stat_config.walltime_nsecs_stats->max, FINAL))
				pr_err("failed to write stat round event\n");
		}

		if (!perf_stat.data.is_pipe) {
			perf_stat.session->header.data_size += perf_stat.bytes_written;
			perf_session__write_header(perf_stat.session, evsel_list, fd, true);
		}

		evlist__close(evsel_list);
		perf_session__delete(perf_stat.session);
	}

	perf_stat__exit_aggr_mode();
	evlist__free_stats(evsel_list);
out:
	if (stat_config.iostat_run)
		iostat_release(evsel_list);

	zfree(&stat_config.walltime_run);
	zfree(&stat_config.user_requested_cpu_list);

	if (smi_cost && smi_reset)
		sysfs__write_int(FREEZE_ON_SMI_PATH, 0);

	evlist__delete(evsel_list);

	evlist__close_control(stat_config.ctl_fd, stat_config.ctl_fd_ack, &stat_config.ctl_fd_close);

	/* Only the low byte of status becomes the exit code. */
	return abs(status);
}