diff options
Diffstat (limited to 'tools/perf/builtin-stat.c')
| -rw-r--r-- | tools/perf/builtin-stat.c | 1821 |
1 files changed, 1078 insertions, 743 deletions
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 9f3e4b257516..ab40d85fb125 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -41,14 +41,14 @@ */ #include "builtin.h" -#include "perf.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/evlist-hybrid.h" #include "util/evsel.h" #include "util/debug.h" #include "util/color.h" @@ -70,9 +70,11 @@ #include "util/pfm.h" #include "util/bpf_counter.h" #include "util/iostat.h" -#include "util/pmu-hybrid.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> @@ -95,84 +97,28 @@ #include <perf/evlist.h> #include <internal/threadmap.h> -#define DEFAULT_SEPARATOR " " -#define FREEZE_ON_SMI_PATH "devices/cpu/freeze_on_smi" - -static void print_counters(struct timespec *ts, int argc, const char **argv); - -/* Default events used for perf stat -T */ -static const char *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 * transaction_limited_attrs = { - "task-clock," - "{" - "instructions," - "cycles," - "cpu/cycles-t/," - "cpu/tx-start/" - "}" -}; +#ifdef HAVE_BPF_SKEL +#include "util/bpf_skel/bperf_cgroup.h" +#endif -static const char * topdown_attrs[] = { - "topdown-total-slots", - "topdown-slots-retired", - "topdown-recovery-bubbles", - "topdown-fetch-bubbles", - "topdown-slots-issued", - NULL, -}; +#define DEFAULT_SEPARATOR " " +#define FREEZE_ON_SMI_PATH "bus/event_source/devices/cpu/freeze_on_smi" -static const char *topdown_metric_attrs[] = { - "slots", - "topdown-retiring", - "topdown-bad-spec", - "topdown-fe-bound", - "topdown-be-bound", - NULL, +struct rusage_stats { + struct stats ru_utime_usec_stat; + struct stats ru_stime_usec_stat; }; -static const char *topdown_metric_L2_attrs[] = { - "slots", - "topdown-retiring", - "topdown-bad-spec", - "topdown-fe-bound", - "topdown-be-bound", - "topdown-heavy-ops", - "topdown-br-mispredict", - "topdown-fetch-lat", - "topdown-mem-bound", - NULL, -}; - -#define TOPDOWN_MAX_LEVEL 2 +static void print_counters(struct timespec *ts, int argc, const char **argv); -static const char *smi_cost_attrs = { - "{" - "msr/aperf/," - "msr/smi/," - "cycles" - "}" +static struct evlist *evsel_list; +static struct parse_events_option_args parse_events_option_args = { + .evlistp = &evsel_list, }; -static struct evlist *evsel_list; static bool all_counters_use_bpf = true; -static struct target target = { - .uid = UINT_MAX, -}; - -#define METRIC_ONLY_LEN 20 +static struct target target; static volatile sig_atomic_t child_pid = -1; static int detailed_run = 0; @@ -192,6 +138,7 @@ 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; @@ -203,6 +150,7 @@ struct perf_stat { struct perf_cpu_map *cpus; struct perf_thread_map *threads; enum aggr_mode aggr_mode; + u32 aggr_level; }; static struct perf_stat perf_stat; @@ -210,71 +158,69 @@ static struct perf_stat perf_stat; static volatile sig_atomic_t done = 0; -static struct perf_stat_config stat_config = { - .aggr_mode = AGGR_GLOBAL, - .scale = true, - .unit_width = 4, /* strlen("unit") */ - .run_count = 1, - .metric_only_len = METRIC_ONLY_LEN, - .walltime_nsecs_stats = &walltime_nsecs_stats, - .ru_stats = &ru_stats, - .big_num = true, - .ctl_fd = -1, - .ctl_fd_ack = -1, - .iostat_run = false, +/* Options set from the command line. */ +struct opt_aggr_mode { + bool node, socket, die, cluster, cache, core, thread, no_aggr; }; -static bool cpus_map_matched(struct evsel *a, struct evsel *b) -{ - if (!a->core.cpus && !b->core.cpus) - return true; - - if (!a->core.cpus || !b->core.cpus) - return false; - - if (perf_cpu_map__nr(a->core.cpus) != perf_cpu_map__nr(b->core.cpus)) - return false; - - for (int i = 0; i < perf_cpu_map__nr(a->core.cpus); i++) { - if (perf_cpu_map__cpu(a->core.cpus, i).cpu != - perf_cpu_map__cpu(b->core.cpus, i).cpu) - return false; - } - - return true; +/* 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, *pos, *leader; - char buf[1024]; - - if (evlist__has_hybrid(evlist)) - evlist__warn_hybrid_group(evlist); + struct evsel *evsel, *warned_leader = NULL; evlist__for_each_entry(evlist, evsel) { - leader = evsel__leader(evsel); + struct evsel *leader = evsel__leader(evsel); /* Check that leader matches cpus with each member. */ if (leader == evsel) continue; - if (cpus_map_matched(leader, evsel)) + if (perf_cpu_map__equal(leader->core.cpus, evsel->core.cpus)) continue; /* If there's mismatch disable the group and warn user. */ - WARN_ONCE(1, "WARNING: grouped events cpus do not match, disabling group:\n"); - evsel__group_desc(leader, buf, sizeof(buf)); - pr_warning(" %s\n", buf); - + 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); } - for_each_group_evsel(pos, leader) - evsel__remove_from_group(pos, leader); + evsel__remove_from_group(evsel, leader); } } @@ -293,10 +239,10 @@ static inline void diff_timespec(struct timespec *r, struct timespec *a, static void perf_stat__reset_stats(void) { evlist__reset_stats(evsel_list); - perf_stat__reset_shadow_stats(); + memset(stat_config.walltime_nsecs_stats, 0, sizeof(*stat_config.walltime_nsecs_stats)); } -static int process_synthesized_event(struct perf_tool *tool __maybe_unused, +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) @@ -332,45 +278,48 @@ static int evsel__write_stat_event(struct evsel *counter, int cpu_map_idx, u32 t process_synthesized_event, NULL); } -static int read_single_counter(struct evsel *counter, int cpu_map_idx, - int thread, struct timespec *rs) +static int read_single_counter(struct evsel *counter, int cpu_map_idx, int thread) { - switch(counter->tool_event) { - case PERF_TOOL_DURATION_TIME: { - u64 val = rs->tv_nsec + rs->tv_sec*1000000000ULL; - struct perf_counts_values *count = - perf_counts(counter->counts, cpu_map_idx, thread); - count->ena = count->run = val; - count->val = val; - return 0; - } - case PERF_TOOL_USER_TIME: - case PERF_TOOL_SYSTEM_TIME: { - u64 val; - struct perf_counts_values *count = - perf_counts(counter->counts, cpu_map_idx, thread); - if (counter->tool_event == PERF_TOOL_USER_TIME) - val = ru_stats.ru_utime_usec_stat.mean; - else - val = ru_stats.ru_stime_usec_stat.mean; - count->ena = count->run = val; - count->val = val; - return 0; + 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++; } - default: - case PERF_TOOL_NONE: - return evsel__read_counter(counter, cpu_map_idx, thread); - case PERF_TOOL_MAX: - /* This should never be reached */ - return 0; + 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, struct timespec *rs, int cpu_map_idx) +static int read_counter_cpu(struct evsel *counter, int cpu_map_idx) { int nthreads = perf_thread_map__nr(evsel_list->core.threads); int thread; @@ -388,7 +337,7 @@ static int read_counter_cpu(struct evsel *counter, struct timespec *rs, int cpu_ * (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, rs)) { + 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; @@ -417,7 +366,7 @@ static int read_counter_cpu(struct evsel *counter, struct timespec *rs, int cpu_ return 0; } -static int read_affinity_counters(struct timespec *rs) +static int read_counters_with_affinity(void) { struct evlist_cpu_iterator evlist_cpu_itr; struct affinity saved_affinity, *affinity; @@ -438,10 +387,11 @@ static int read_affinity_counters(struct timespec *rs) if (evsel__is_bpf(counter)) continue; - if (!counter->err) { - counter->err = read_counter_cpu(counter, rs, - evlist_cpu_itr.cpu_map_idx); - } + if (evsel__is_tool(counter)) + continue; + + if (!counter->err) + counter->err = read_counter_cpu(counter, evlist_cpu_itr.cpu_map_idx); } if (affinity) affinity__cleanup(&saved_affinity); @@ -465,16 +415,46 @@ static int read_bpf_map_counters(void) return 0; } -static int read_counters(struct timespec *rs) +static int read_tool_counters(void) { - if (!stat_config.stop_read_counter) { - if (read_bpf_map_counters() || - read_affinity_counters(rs)) - return -1; + 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; @@ -489,7 +469,6 @@ static void process_counters(void) perf_stat_merge_counters(&stat_config, evsel_list); perf_stat_process_percore(&stat_config, evsel_list); - perf_stat_process_shadow_stats(&stat_config, evsel_list); } static void process_interval(void) @@ -499,10 +478,9 @@ static void process_interval(void) clock_gettime(CLOCK_MONOTONIC, &ts); diff_timespec(&rs, &ts, &ref_time); - perf_stat__reset_shadow_per_stat(&rt_stat); evlist__reset_aggr_stats(evsel_list); - if (read_counters(&rs) == 0) + if (read_counters() == 0) process_counters(); if (STAT_RECORD) { @@ -510,8 +488,8 @@ static void process_interval(void) pr_err("failed to write stat round event\n"); } - init_stats(&walltime_nsecs_stats); - update_stats(&walltime_nsecs_stats, stat_config.interval * 1000000ULL); + init_stats(stat_config.walltime_nsecs_stats); + update_stats(stat_config.walltime_nsecs_stats, stat_config.interval * 1000000ULL); print_counters(&rs, 0, NULL); } @@ -539,12 +517,7 @@ static int enable_counters(void) return err; } - /* - * We need to enable counters only if: - * - we don't have tracee (attaching to task or cpu) - * - we have initial delay configured - */ - if (!target__none(&target)) { + if (!target__enable_on_exec(&target)) { if (!all_counters_use_bpf) evlist__enable(evsel_list); } @@ -615,7 +588,7 @@ static void process_evlist(struct evlist *evlist, unsigned int interval) if (evlist__ctlfd_process(evlist, &cmd) > 0) { switch (cmd) { case EVLIST_CTL_CMD_ENABLE: - __fallthrough; + fallthrough; case EVLIST_CTL_CMD_DISABLE: if (interval) process_interval(); @@ -691,39 +664,34 @@ static int dispatch_events(bool forks, int timeout, int interval, int *times) enum counter_recovery { COUNTER_SKIP, COUNTER_RETRY, - COUNTER_FATAL, }; -static enum counter_recovery stat_handle_error(struct evsel *counter) +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 (errno == EINVAL || errno == ENOSYS || - errno == ENOENT || errno == EOPNOTSUPP || - errno == ENXIO) { - if (verbose > 0) - ui__warning("%s event is not supported by the kernel.\n", - evsel__name(counter)); - counter->supported = false; - /* - * errored is a sticky flag that means one of the counter's - * cpu event had a problem and needs to be reexamined. - */ - counter->errored = true; - - if ((evsel__leader(counter) != counter) || - !(counter->core.leader->nr_members > 1)) - return COUNTER_SKIP; - } else if (evsel__fallback(counter, errno, msg, sizeof(msg))) { + 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; - } else if (target__has_per_thread(&target) && - evsel_list->core.threads && - evsel_list->core.threads->err_thread != -1) { + } + 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. @@ -731,16 +699,85 @@ static enum counter_recovery stat_handle_error(struct evsel *counter) 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; +} - evsel__open_strerror(counter, &target, errno, msg, sizeof(msg)); - ui__error("%s\n", msg); +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); - if (child_pid != -1) - kill(child_pid, SIGTERM); - return COUNTER_FATAL; + /* 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) @@ -757,8 +794,8 @@ static int __run_perf_stat(int argc, const char **argv, int run_idx) bool is_pipe = STAT_RECORD ? perf_stat.data.is_pipe : false; struct evlist_cpu_iterator evlist_cpu_itr; struct affinity saved_affinity, *affinity = NULL; - int err; - bool second_pass = false; + 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) { @@ -769,19 +806,25 @@ static int __run_perf_stat(int argc, const char **argv, int run_idx) } if (!cpu_map__is_dummy(evsel_list->core.user_requested_cpus)) { - if (affinity__setup(&saved_affinity) < 0) - return -1; + if (affinity__setup(&saved_affinity) < 0) { + err = -1; + goto err_out; + } affinity = &saved_affinity; } evlist__for_each_entry(evsel_list, counter) { counter->reset_group = false; - if (bpf_counter__load(counter, &target)) - return -1; - if (!evsel__is_bpf(counter)) + 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); + evlist__for_each_cpu(evlist_cpu_itr, evsel_list, affinity) { counter = evlist_cpu_itr.evsel; @@ -792,14 +835,17 @@ static int __run_perf_stat(int argc, const char **argv, int run_idx) if (target.use_bpf) break; - if (counter->reset_group || counter->errored) + if (counter->reset_group || !counter->supported) continue; - if (evsel__is_bpf(counter)) + if (evsel__is_bperf(counter)) continue; -try_again: - if (create_perf_stat_counter(counter, &stat_config, &target, - evlist_cpu_itr.cpu_map_idx) < 0) { + 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 @@ -807,28 +853,19 @@ try_again: * it to later. * Don't close here because we're in the wrong affinity. */ - if ((errno == EINVAL || errno == EBADF) && + 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; - continue; - } - - switch (stat_handle_error(counter)) { - case COUNTER_FATAL: - return -1; - case COUNTER_RETRY: - goto try_again; - case COUNTER_SKIP: - continue; - default: break; } + if (stat_handle_error(counter, open_err) != COUNTER_RETRY) + break; } - counter->supported = true; } if (second_pass) { @@ -841,7 +878,7 @@ try_again: evlist__for_each_cpu(evlist_cpu_itr, evsel_list, affinity) { counter = evlist_cpu_itr.evsel; - if (!counter->reset_group && !counter->errored) + if (!counter->reset_group && counter->supported) continue; perf_evsel__close_cpu(&counter->core, evlist_cpu_itr.cpu_map_idx); @@ -852,43 +889,54 @@ try_again: if (!counter->reset_group) continue; -try_again_reset: - pr_debug2("reopening weak %s\n", evsel__name(counter)); - if (create_perf_stat_counter(counter, &stat_config, &target, - evlist_cpu_itr.cpu_map_idx) < 0) { - - switch (stat_handle_error(counter)) { - case COUNTER_FATAL: - return -1; - case COUNTER_RETRY: - goto try_again_reset; - case COUNTER_SKIP: - continue; - default: + + 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) + break; + + open_err = errno; + if (stat_handle_error(counter, open_err) != COUNTER_RETRY) break; - } } - counter->supported = true; } } affinity__cleanup(affinity); + affinity = NULL; + 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)) - return -1; + 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)) { + if (evlist__apply_filters(evsel_list, &counter, &target)) { pr_err("failed to set filter \"%s\" on event %s with %d (%s)\n", counter->filter, evsel__name(counter), errno, str_error_r(errno, msg, sizeof(msg))); @@ -906,31 +954,36 @@ try_again_reset: } if (err < 0) - return err; + goto err_out; err = perf_event__synthesize_stat_events(&stat_config, NULL, evsel_list, process_synthesized_event, is_pipe); if (err < 0) - return err; + goto err_out; + } - if (stat_config.initial_delay) { + if (target.initial_delay) { pr_info(EVLIST_DISABLED_MSG); } else { err = enable_counters(); - if (err) - return -1; + if (err) { + err = -1; + goto err_out; + } } /* Exec the command, if any */ if (forks) evlist__start_workload(evsel_list); - if (stat_config.initial_delay > 0) { - usleep(stat_config.initial_delay * USEC_PER_MSEC); + if (target.initial_delay > 0) { + usleep(target.initial_delay * USEC_PER_MSEC); err = enable_counters(); - if (err) - return -1; + if (err) { + err = -1; + goto err_out; + } pr_info(EVLIST_ENABLED_MSG); } @@ -950,13 +1003,24 @@ try_again_reset: if (workload_exec_errno) { const char *emsg = str_error_r(workload_exec_errno, msg, sizeof(msg)); pr_err("Workload failed: %s\n", emsg); - return -1; + err = -1; + goto err_out; } - if (WIFSIGNALED(status)) + 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 { - status = dispatch_events(forks, timeout, interval, ×); + err = dispatch_events(forks, timeout, interval, ×); } disable_counters(); @@ -969,16 +1033,15 @@ try_again_reset: if (interval && stat_config.summary) { stat_config.interval = 0; stat_config.stop_read_counter = true; - init_stats(&walltime_nsecs_stats); - update_stats(&walltime_nsecs_stats, t1 - t0); + 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); - perf_stat__reset_shadow_per_stat(&rt_stat); } else { - update_stats(&walltime_nsecs_stats, t1 - t0); - update_rusage_stats(&ru_stats, &stat_config.ru_data); + update_stats(stat_config.walltime_nsecs_stats, t1 - t0); + update_rusage_stats(&stat_config.ru_data); } /* @@ -987,7 +1050,7 @@ try_again_reset: * avoid arbitrary skew, we must read all counters before closing any * group leaders. */ - if (read_counters(&(struct timespec) { .tv_nsec = t1-t0 }) == 0) + if (read_counters() == 0) process_counters(); /* @@ -997,9 +1060,24 @@ try_again_reset: if (!STAT_RECORD) evlist__close(evsel_list); - return WEXITSTATUS(status); + return err; + +err_out: + if (forks) + evlist__cancel_workload(evsel_list); + + affinity__cleanup(affinity); + 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; @@ -1080,16 +1158,6 @@ static void sig_atexit(void) kill(getpid(), signr); } -void perf_stat__set_big_num(int set) -{ - stat_config.big_num = (set != 0); -} - -void perf_stat__set_no_csv_summary(int set) -{ - stat_config.no_csv_summary = (set != 0); -} - static int stat__set_big_num(const struct option *opt __maybe_unused, const char *s __maybe_unused, int unset) { @@ -1145,10 +1213,11 @@ static int parse_stat_cgroups(const struct option *opt, return parse_cgroups(opt, str, unset); } -static int parse_hybrid_type(const struct option *opt, +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)) { @@ -1156,153 +1225,191 @@ static int parse_hybrid_type(const struct option *opt, return -1; } - evlist->hybrid_pmu_name = perf_pmu__hybrid_type_to_pmu(str); - if (!evlist->hybrid_pmu_name) { + 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 struct option stat_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), - 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_SET_UINT('A', "no-aggr", &stat_config.aggr_mode, - "disable CPU count aggregation", AGGR_NONE), - OPT_BOOLEAN(0, "no-merge", &stat_config.no_merge, "Do not merge identical named events"), - 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_SET_UINT(0, "per-socket", &stat_config.aggr_mode, - "aggregate counts per processor socket", AGGR_SOCKET), - OPT_SET_UINT(0, "per-die", &stat_config.aggr_mode, - "aggregate counts per processor die", AGGR_DIE), - OPT_SET_UINT(0, "per-core", &stat_config.aggr_mode, - "aggregate counts per physical processor core", AGGR_CORE), - OPT_SET_UINT(0, "per-thread", &stat_config.aggr_mode, - "aggregate counts per thread", AGGR_THREAD), - OPT_SET_UINT(0, "per-node", &stat_config.aggr_mode, - "aggregate counts per numa node", AGGR_NODE), - OPT_INTEGER('D', "delay", &stat_config.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, "topdown", &topdown_run, - "measure top-down statistics"), - 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_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_hybrid_type), -#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() -}; +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", @@ -1324,6 +1431,18 @@ static struct aggr_cpu_id perf_stat__get_die(struct perf_stat_config *config __m 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) { @@ -1354,7 +1473,7 @@ static struct aggr_cpu_id perf_stat__get_aggr(struct perf_stat_config *config, struct aggr_cpu_id id; /* per-process mode - should use global aggr mode */ - if (cpu.cpu == -1) + 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])) @@ -1376,6 +1495,18 @@ static struct aggr_cpu_id perf_stat__get_die_cached(struct perf_stat_config *con 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) { @@ -1407,6 +1538,10 @@ static aggr_cpu_id_get_t aggr_mode__get_aggr(enum aggr_mode aggr_mode) 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: @@ -1430,6 +1565,10 @@ static aggr_get_id_t aggr_mode__get_id(enum aggr_mode aggr_mode) 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: @@ -1456,7 +1595,7 @@ static int perf_stat_init_aggr_mode(void) 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", aggr_mode__string[stat_config.aggr_mode]); + 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); @@ -1482,33 +1621,20 @@ static int perf_stat_init_aggr_mode(void) * taking the highest cpu number to be the size of * the aggregation translate cpumap. */ - if (evsel_list->core.user_requested_cpus) - nr = perf_cpu_map__max(evsel_list->core.user_requested_cpus).cpu; - else - nr = 0; - stat_config.cpus_aggr_map = cpu_aggr_map__empty_new(nr + 1); + 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) { - if (map) { - WARN_ONCE(refcount_read(&map->refcnt) != 0, - "cpu_aggr_map refcnt unbalanced\n"); - free(map); - } -} - -static void cpu_aggr_map__put(struct cpu_aggr_map *map) -{ - if (map && refcount_dec_and_test(&map->refcnt)) - cpu_aggr_map__delete(map); + free(map); } static void perf_stat__exit_aggr_mode(void) { - cpu_aggr_map__put(stat_config.aggr_map); - cpu_aggr_map__put(stat_config.cpus_aggr_map); + 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; } @@ -1542,6 +1668,75 @@ static struct aggr_cpu_id perf_env__get_die_aggr_by_cpu(struct perf_cpu cpu, voi 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; @@ -1549,12 +1744,12 @@ static struct aggr_cpu_id perf_env__get_core_aggr_by_cpu(struct perf_cpu cpu, vo 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 + * 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; } @@ -1602,36 +1797,48 @@ static struct aggr_cpu_id perf_env__get_global_aggr_by_cpu(struct perf_cpu cpu _ 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_stat.session->header.env); + 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_stat.session->header.env); + 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_stat.session->header.env); + 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_stat.session->header.env); + 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_stat.session->header.env); + 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_stat.session->header.env); + 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) @@ -1641,6 +1848,10 @@ static aggr_cpu_id_get_t aggr_mode__get_aggr_file(enum aggr_mode aggr_mode) 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: @@ -1664,6 +1875,10 @@ static aggr_get_id_t aggr_mode__get_id_file(enum aggr_mode aggr_mode) 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: @@ -1682,7 +1897,7 @@ static aggr_get_id_t aggr_mode__get_id_file(enum aggr_mode aggr_mode) static int perf_stat_init_aggr_mode_file(struct perf_stat *st) { - struct perf_env *env = &st->session->header.env; + 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; @@ -1708,332 +1923,229 @@ static int perf_stat_init_aggr_mode_file(struct perf_stat *st) 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", aggr_mode__string[stat_config.aggr_mode]); + 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; } -/* - * Add default attributes, if there were no attributes specified or - * if -d/--detailed, -d -d or -d -d -d is used: - */ -static int add_default_attributes(void) +static int default_evlist_evsel_cmp(void *priv __maybe_unused, + const struct list_head *l, + const struct list_head *r) { - int err; - struct perf_event_attr default_attrs0[] = { + 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); - { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_TASK_CLOCK }, - { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_CONTEXT_SWITCHES }, - { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_CPU_MIGRATIONS }, - { .type = PERF_TYPE_SOFTWARE, .config = PERF_COUNT_SW_PAGE_FAULTS }, + if (evsel__leader(lhs) == evsel__leader(rhs)) { + /* Within the same group, respect the original order. */ + return lhs_core->idx - rhs_core->idx; + } - { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_CPU_CYCLES }, -}; - struct perf_event_attr frontend_attrs[] = { - { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_STALLED_CYCLES_FRONTEND }, -}; - struct perf_event_attr backend_attrs[] = { - { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_STALLED_CYCLES_BACKEND }, -}; - struct perf_event_attr default_attrs1[] = { - { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_INSTRUCTIONS }, - { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_BRANCH_INSTRUCTIONS }, - { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_BRANCH_MISSES }, + /* 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; -/* - * Detailed stats (-d), covering the L1 and last level data caches: - */ - struct perf_event_attr detailed_attrs[] = { - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_L1D << 0 | - (PERF_COUNT_HW_CACHE_OP_READ << 8) | - (PERF_COUNT_HW_CACHE_RESULT_ACCESS << 16) }, - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_L1D << 0 | - (PERF_COUNT_HW_CACHE_OP_READ << 8) | - (PERF_COUNT_HW_CACHE_RESULT_MISS << 16) }, - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_LL << 0 | - (PERF_COUNT_HW_CACHE_OP_READ << 8) | - (PERF_COUNT_HW_CACHE_RESULT_ACCESS << 16) }, - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_LL << 0 | - (PERF_COUNT_HW_CACHE_OP_READ << 8) | - (PERF_COUNT_HW_CACHE_RESULT_MISS << 16) }, -}; + /* Sort by PMU type (prefers legacy types first). */ + if (lhs->pmu != rhs->pmu) + return lhs->pmu->type - rhs->pmu->type; -/* - * Very detailed stats (-d -d), covering the instruction cache and the TLB caches: - */ - struct perf_event_attr very_detailed_attrs[] = { - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_L1I << 0 | - (PERF_COUNT_HW_CACHE_OP_READ << 8) | - (PERF_COUNT_HW_CACHE_RESULT_ACCESS << 16) }, - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_L1I << 0 | - (PERF_COUNT_HW_CACHE_OP_READ << 8) | - (PERF_COUNT_HW_CACHE_RESULT_MISS << 16) }, - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_DTLB << 0 | - (PERF_COUNT_HW_CACHE_OP_READ << 8) | - (PERF_COUNT_HW_CACHE_RESULT_ACCESS << 16) }, - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_DTLB << 0 | - (PERF_COUNT_HW_CACHE_OP_READ << 8) | - (PERF_COUNT_HW_CACHE_RESULT_MISS << 16) }, - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_ITLB << 0 | - (PERF_COUNT_HW_CACHE_OP_READ << 8) | - (PERF_COUNT_HW_CACHE_RESULT_ACCESS << 16) }, - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_ITLB << 0 | - (PERF_COUNT_HW_CACHE_OP_READ << 8) | - (PERF_COUNT_HW_CACHE_RESULT_MISS << 16) }, - -}; + /* Sort by name. */ + return strcmp(evsel__name((struct evsel *)lhs), evsel__name((struct evsel *)rhs)); +} /* - * Very, very detailed stats (-d -d -d), adding prefetch events: + * Add default events, if there were no attributes specified or + * if -d/--detailed, -d -d or -d -d -d is used: */ - struct perf_event_attr very_very_detailed_attrs[] = { - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_L1D << 0 | - (PERF_COUNT_HW_CACHE_OP_PREFETCH << 8) | - (PERF_COUNT_HW_CACHE_RESULT_ACCESS << 16) }, - - { .type = PERF_TYPE_HW_CACHE, - .config = - PERF_COUNT_HW_CACHE_L1D << 0 | - (PERF_COUNT_HW_CACHE_OP_PREFETCH << 8) | - (PERF_COUNT_HW_CACHE_RESULT_MISS << 16) }, -}; +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; - struct perf_event_attr default_null_attrs[] = {}; + parse_events_error__init(&err); /* Set attrs if no event is selected and !null_run: */ if (stat_config.null_run) - return 0; + goto out; if (transaction_run) { - struct parse_events_error errinfo; /* 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("transaction")) { - return metricgroup__parse_groups(evsel_list, "transaction", - stat_config.metric_no_group, - stat_config.metric_no_merge, - stat_config.user_requested_cpu_list, - stat_config.system_wide, - &stat_config.metric_events); - } - - parse_events_error__init(&errinfo); - if (pmu_have_event("cpu", "cycles-ct") && - pmu_have_event("cpu", "el-start")) - err = parse_events(evsel_list, transaction_attrs, - &errinfo); - else - err = parse_events(evsel_list, - transaction_limited_attrs, - &errinfo); - if (err) { - fprintf(stderr, "Cannot set up transaction events\n"); - parse_events_error__print(&errinfo, transaction_attrs); + if (!metricgroup__has_metric_or_groups(pmu, "transaction")) { + pr_err("Missing transaction metrics\n"); + ret = -1; + goto out; } - parse_events_error__exit(&errinfo); - return err ? -1 : 0; + 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) { - struct parse_events_error errinfo; int smi; if (sysfs__read_int(FREEZE_ON_SMI_PATH, &smi) < 0) { - fprintf(stderr, "freeze_on_smi is not supported.\n"); - return -1; + 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) { - fprintf(stderr, "Failed to set freeze_on_smi.\n"); - return -1; + pr_err("Failed to set freeze_on_smi.\n"); + ret = -1; + goto out; } smi_reset = true; } - if (!pmu_have_event("msr", "aperf") || - !pmu_have_event("msr", "smi")) { - fprintf(stderr, "To measure SMI cost, it needs " - "msr/aperf/, msr/smi/ and cpu/cycles/ support\n"); - return -1; + 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; - parse_events_error__init(&errinfo); - err = parse_events(evsel_list, smi_cost_attrs, &errinfo); - if (err) { - parse_events_error__print(&errinfo, smi_cost_attrs); - fprintf(stderr, "Cannot set up SMI cost events\n"); - } - parse_events_error__exit(&errinfo); - return err ? -1 : 0; + 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) { - const char **metric_attrs = topdown_metric_attrs; - unsigned int max_level = 1; - char *str = NULL; - bool warn = false; - const char *pmu_name = arch_get_topdown_pmu_name(evsel_list, true); + unsigned int max_level = metricgroups__topdown_max_level(); + char str[] = "TopdownL1"; if (!force_metric_only) stat_config.metric_only = true; - if (pmu_have_event(pmu_name, topdown_metric_L2_attrs[5])) { - metric_attrs = topdown_metric_L2_attrs; - max_level = 2; + 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); - return -1; - } else if (!stat_config.topdown_level) - stat_config.topdown_level = max_level; - - if (topdown_filter_events(metric_attrs, &str, 1, pmu_name) < 0) { - pr_err("Out of memory\n"); - return -1; - } - - if (metric_attrs[0] && str) { - 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"); - } - goto setup_metrics; - } - - zfree(&str); - - if (stat_config.aggr_mode != AGGR_GLOBAL && - stat_config.aggr_mode != AGGR_CORE) { - pr_err("top down event configuration requires --per-core mode\n"); - return -1; - } - stat_config.aggr_mode = AGGR_CORE; - if (nr_cgroups || !target__has_cpu(&target)) { - pr_err("top down event configuration requires system-wide mode (-a)\n"); - return -1; + ret = -1; + goto out; + } else if (!stat_config.topdown_level) { + stat_config.topdown_level = 1; } - - if (topdown_filter_events(topdown_attrs, &str, - arch_topdown_check_group(&warn), - pmu_name) < 0) { - pr_err("Out of memory\n"); - return -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"); } - - if (topdown_attrs[0] && str) { - struct parse_events_error errinfo; - if (warn) - arch_topdown_group_warn(); -setup_metrics: - parse_events_error__init(&errinfo); - err = parse_events(evsel_list, str, &errinfo); - if (err) { - fprintf(stderr, - "Cannot set up top down events %s: %d\n", - str, err); - parse_events_error__print(&errinfo, str); - parse_events_error__exit(&errinfo); - free(str); - return -1; - } - parse_events_error__exit(&errinfo); - } else { - fprintf(stderr, "System does not support topdown\n"); - return -1; + 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; } - free(str); } if (!stat_config.topdown_level) - stat_config.topdown_level = TOPDOWN_MAX_LEVEL; + stat_config.topdown_level = 1; - if (!evsel_list->core.nr_entries) { - if (target__has_cpu(&target)) - default_attrs0[0].config = PERF_COUNT_SW_CPU_CLOCK; + 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", + }; - if (evlist__add_default_attrs(evsel_list, default_attrs0) < 0) - return -1; - if (pmu_have_event("cpu", "stalled-cycles-frontend")) { - if (evlist__add_default_attrs(evsel_list, frontend_attrs) < 0) - return -1; - } - if (pmu_have_event("cpu", "stalled-cycles-backend")) { - if (evlist__add_default_attrs(evsel_list, backend_attrs) < 0) - return -1; - } - if (evlist__add_default_attrs(evsel_list, default_attrs1) < 0) - return -1; - /* Platform specific attrs */ - if (evlist__add_default_attrs(evsel_list, default_null_attrs) < 0) - return -1; - } + for (size_t i = 0; i < ARRAY_SIZE(default_metricgroup_names); i++) { + struct evlist *metric_evlist; - /* Detailed events get appended to the event list: */ + if (!metricgroup__has_metric_or_groups(pmu, default_metricgroup_names[i])) + continue; - if (detailed_run < 1) - return 0; + if ((int)i > detailed_run) + break; - /* Append detailed run extra attributes: */ - if (evlist__add_default_attrs(evsel_list, detailed_attrs) < 0) - return -1; + 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; + } - if (detailed_run < 2) - return 0; + evlist__for_each_entry(metric_evlist, evsel) + evsel->default_metricgroup = true; - /* Append very detailed run extra attributes: */ - if (evlist__add_default_attrs(evsel_list, very_detailed_attrs) < 0) - return -1; - - if (detailed_run < 3) - return 0; + 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); - /* Append very, very detailed run extra attributes: */ - return evlist__add_default_attrs(evsel_list, very_very_detailed_attrs); + } +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[] = { @@ -2055,13 +2167,15 @@ static void init_features(struct perf_session *session) perf_header__clear_feat(&session->header, HEADER_AUXTRACE); } -static int __cmd_record(int argc, const char **argv) +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; @@ -2085,18 +2199,20 @@ static int __cmd_record(int argc, const char **argv) return argc; } -static int process_stat_round_event(struct perf_session *session, +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; - const char **argv = session->header.env.cmdline_argv; - int argc = session->header.env.nr_cmdline; + 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(&walltime_nsecs_stats, stat_round->time); + 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; @@ -2109,15 +2225,15 @@ static int process_stat_round_event(struct perf_session *session, } static -int process_stat_config_event(struct perf_session *session, +int process_stat_config_event(const struct perf_tool *tool, + struct perf_session *session, union perf_event *event) { - struct perf_tool *tool = session->tool; 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__empty(st->cpus)) { + 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) { @@ -2158,10 +2274,10 @@ static int set_maps(struct perf_stat *st) } static -int process_thread_map_event(struct perf_session *session, +int process_thread_map_event(const struct perf_tool *tool, + struct perf_session *session __maybe_unused, union perf_event *event) { - struct perf_tool *tool = session->tool; struct perf_stat *st = container_of(tool, struct perf_stat, tool); if (st->threads) { @@ -2177,10 +2293,10 @@ int process_thread_map_event(struct perf_session *session, } static -int process_cpu_map_event(struct perf_session *session, +int process_cpu_map_event(const struct perf_tool *tool, + struct perf_session *session __maybe_unused, union perf_event *event) { - struct perf_tool *tool = session->tool; struct perf_stat *st = container_of(tool, struct perf_stat, tool); struct perf_cpu_map *cpus; @@ -2203,16 +2319,8 @@ static const char * const stat_report_usage[] = { }; static struct perf_stat perf_stat = { - .tool = { - .attr = perf_event__process_attr, - .event_update = perf_event__process_event_update, - .thread_map = process_thread_map_event, - .cpu_map = process_cpu_map_event, - .stat_config = process_stat_config_event, - .stat = perf_event__process_stat_event, - .stat_round = process_stat_round_event, - }, - .aggr_mode = AGGR_UNSET, + .aggr_mode = AGGR_UNSET, + .aggr_level = 0, }; static int __cmd_report(int argc, const char **argv) @@ -2224,6 +2332,12 @@ static int __cmd_report(int argc, const char **argv) "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, @@ -2244,17 +2358,25 @@ static int __cmd_report(int argc, const char **argv) input_name = "perf.data"; } - perf_stat__init_shadow_stats(); - 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); @@ -2286,7 +2408,7 @@ static void setup_system_wide(int forks) evlist__for_each_entry(evsel_list, counter) { if (!counter->core.requires_cpu && - strcmp(counter->name, "duration_time")) { + !evsel__name_is(counter, "duration_time")) { return; } } @@ -2296,8 +2418,182 @@ static void setup_system_wide(int forks) } } +#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 = {}; + 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_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 @@ -2308,6 +2604,7 @@ int cmd_stat(int argc, const char **argv) unsigned int interval, timeout; const char * const stat_subcommands[] = { "record", "report" }; char errbuf[BUFSIZ]; + struct evsel *counter; setlocale(LC_ALL, ""); @@ -2326,6 +2623,8 @@ int cmd_stat(int argc, const char **argv) (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")) @@ -2334,7 +2633,7 @@ int cmd_stat(int argc, const char **argv) stat_config.csv_sep = DEFAULT_SEPARATOR; if (argc && strlen(argv[0]) > 2 && strstarts("record", argv[0])) { - argc = __cmd_record(argc, argv); + 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])) @@ -2366,6 +2665,14 @@ int cmd_stat(int argc, const char **argv) 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); @@ -2388,8 +2695,10 @@ int cmd_stat(int argc, const char **argv) perror("failed to create output file"); return -1; } - clock_gettime(CLOCK_REALTIME, &tm); - fprintf(output, "# started on %s\n", ctime(&tm.tv_sec)); + 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); @@ -2424,6 +2733,7 @@ int cmd_stat(int argc, const char **argv) } 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); @@ -2516,18 +2826,23 @@ int cmd_stat(int argc, const char **argv) * knowing the target is system-wide. */ if (metrics) { - metricgroup__parse_groups(evsel_list, metrics, - stat_config.metric_no_group, - stat_config.metric_no_merge, - stat_config.user_requested_cpu_list, - stat_config.system_wide, - &stat_config.metric_events); + 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; + } } - perf_stat__collect_metric_expr(evsel_list); - perf_stat__init_shadow_stats(); - if (add_default_attributes()) + if (add_default_events()) goto out; if (stat_config.cgroup_list) { @@ -2538,20 +2853,34 @@ int cmd_stat(int argc, const char **argv) goto out; } - if (evlist__expand_cgroup(evsel_list, stat_config.cgroup_list, - &stat_config.metric_events, true) < 0) { + 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); - if (evlist__fix_hybrid_cpus(evsel_list, target.cpu_list)) { - pr_err("failed to use cpu list %s\n", target.cpu_list); - goto out; + 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; } - target.hybrid = perf_pmu__has_hybrid(); if (evlist__create_maps(evsel_list, &target) < 0) { if (target__has_task(&target)) { pr_err("Problems finding threads of monitor\n"); @@ -2650,14 +2979,20 @@ int cmd_stat(int argc, const char **argv) evlist__reset_prev_raw_counts(evsel_list); status = run_perf_stat(argc, argv, run_idx); - if (forever && status != -1 && !interval) { + 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 (!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); @@ -2685,7 +3020,7 @@ int cmd_stat(int argc, const char **argv) } if (!interval) { - if (WRITE_STAT_ROUND_EVENT(walltime_nsecs_stats.max, FINAL)) + if (WRITE_STAT_ROUND_EVENT(stat_config.walltime_nsecs_stats->max, FINAL)) pr_err("failed to write stat round event\n"); } @@ -2712,8 +3047,8 @@ out: evlist__delete(evsel_list); - metricgroup__rblist_exit(&stat_config.metric_events); evlist__close_control(stat_config.ctl_fd, stat_config.ctl_fd_ack, &stat_config.ctl_fd_close); - return status; + /* Only the low byte of status becomes the exit code. */ + return abs(status); } |