builtin-top.c

{
	u64 ip = event->ip.ip;
	struct sym_entry *syme;
	struct addr_location al;
	struct machine *machine;
	u8 origin = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;

	++samples;

	switch (origin) {
	case PERF_RECORD_MISC_USER:
		++us_samples;
		if (hide_user_symbols)
			return;
		machine = perf_session__find_host_machine(session);
		break;
	case PERF_RECORD_MISC_KERNEL:
		++kernel_samples;
		if (hide_kernel_symbols)
			return;
		machine = perf_session__find_host_machine(session);
		break;
	case PERF_RECORD_MISC_GUEST_KERNEL:
		++guest_kernel_samples;
		machine = perf_session__find_machine(session, event->ip.pid);
		break;
	case PERF_RECORD_MISC_GUEST_USER:
		++guest_us_samples;
		/*
		 * TODO: we don't process guest user from host side
		 * except simple counting.
		 */
		return;
	default:
		return;
	}

	if (!machine && perf_guest) {
		pr_err("Can't find guest [%d]'s kernel information\n",
			event->ip.pid);
		return;
	}

	if (event->header.misc & PERF_RECORD_MISC_EXACT_IP)
		exact_samples++;

	if (perf_event__preprocess_sample(event, session, &al, sample,
					  symbol_filter) < 0 ||
	    al.filtered)
		return;

	if (al.sym == NULL) {
		/*
		 * As we do lazy loading of symtabs we only will know if the
		 * specified vmlinux file is invalid when we actually have a
		 * hit in kernel space and then try to load it. So if we get
		 * here and there are _no_ symbols in the DSO backing the
		 * kernel map, bail out.
		 *
		 * We may never get here, for instance, if we use -K/
		 * --hide-kernel-symbols, even if the user specifies an
		 * invalid --vmlinux ;-)
		 */
		if (al.map == machine->vmlinux_maps[MAP__FUNCTION] &&
		    RB_EMPTY_ROOT(&al.map->dso->symbols[MAP__FUNCTION])) {
			pr_err("The %s file can't be used\n",
			       symbol_conf.vmlinux_name);
			exit(1);
		}

		return;
	}

	/* let's see, whether we need to install initial sym_filter_entry */
	if (sym_filter_entry_sched) {
		sym_filter_entry = sym_filter_entry_sched;
		sym_filter_entry_sched = NULL;
		if (parse_source(sym_filter_entry) < 0) {
			struct symbol *sym = sym_entry__symbol(sym_filter_entry);

			pr_err("Can't annotate %s", sym->name);
			if (sym_filter_entry->map->dso->origin == DSO__ORIG_KERNEL) {
				pr_err(": No vmlinux file was found in the path:\n");
				machine__fprintf_vmlinux_path(machine, stderr);
			} else
				pr_err(".\n");
			exit(1);
		}
	}

	syme = symbol__priv(al.sym);
	if (!syme->skip) {
		struct perf_evsel *evsel;

		syme->origin = origin;
		evsel = perf_evlist__id2evsel(evsel_list, sample->id);
		assert(evsel != NULL);
		syme->count[evsel->idx]++;
		record_precise_ip(syme, evsel->idx, ip);
		pthread_mutex_lock(&active_symbols_lock);
		if (list_empty(&syme->node) || !syme->node.next)
			__list_insert_active_sym(syme);
		pthread_mutex_unlock(&active_symbols_lock);
	}
}

static void perf_session__mmap_read_cpu(struct perf_session *self, int cpu)
{
	struct perf_sample sample;
	union perf_event *event;

	while ((event = perf_evlist__read_on_cpu(evsel_list, cpu)) != NULL) {
		perf_session__parse_sample(self, event, &sample);

		if (event->header.type == PERF_RECORD_SAMPLE)
			perf_event__process_sample(event, &sample, self);
		else
			perf_event__process(event, &sample, self);
	}
}

static void perf_session__mmap_read(struct perf_session *self)
{
	int i;

	for (i = 0; i < evsel_list->cpus->nr; i++)
		perf_session__mmap_read_cpu(self, i);
}

static void start_counters(struct perf_evlist *evlist)
{
	struct perf_evsel *counter;

	list_for_each_entry(counter, &evlist->entries, node) {
		struct perf_event_attr *attr = &counter->attr;

		attr->sample_type = PERF_SAMPLE_IP | PERF_SAMPLE_TID;

		if (freq) {
			attr->sample_type |= PERF_SAMPLE_PERIOD;
			attr->freq	  = 1;
			attr->sample_freq = freq;
		}

		if (evlist->nr_entries > 1) {
			attr->sample_type |= PERF_SAMPLE_ID;
			attr->read_format |= PERF_FORMAT_ID;
		}

		attr->mmap = 1;
try_again:
		if (perf_evsel__open(counter, evsel_list->cpus,
				     evsel_list->threads, group, inherit) < 0) {
			int err = errno;

			if (err == EPERM || err == EACCES)
				die("Permission error - are you root?\n"
					"\t Consider tweaking"
					" /proc/sys/kernel/perf_event_paranoid.\n");
			/*
			 * If it's cycles then fall back to hrtimer
			 * based cpu-clock-tick sw counter, which
			 * is always available even if no PMU support:
			 */
			if (attr->type == PERF_TYPE_HARDWARE &&
			    attr->config == PERF_COUNT_HW_CPU_CYCLES) {

				if (verbose)
					warning(" ... trying to fall back to cpu-clock-ticks\n");

				attr->type = PERF_TYPE_SOFTWARE;
				attr->config = PERF_COUNT_SW_CPU_CLOCK;
				goto try_again;
			}
			printf("\n");
			error("sys_perf_event_open() syscall returned with %d "
			      "(%s).  /bin/dmesg may provide additional information.\n",
			      err, strerror(err));
			die("No CONFIG_PERF_EVENTS=y kernel support configured?\n");
			exit(-1);
		}
	}

	if (perf_evlist__mmap(evlist, mmap_pages, false) < 0)
		die("failed to mmap with %d (%s)\n", errno, strerror(errno));
}

static int __cmd_top(void)
{
	pthread_t thread;
	struct perf_evsel *first;
	int ret;
	/*
	 * FIXME: perf_session__new should allow passing a O_MMAP, so that all this
	 * mmap reading, etc is encapsulated in it. Use O_WRONLY for now.
	 */
	struct perf_session *session = perf_session__new(NULL, O_WRONLY, false, false, NULL);
	if (session == NULL)
		return -ENOMEM;

	if (target_tid != -1)
		perf_event__synthesize_thread(target_tid, perf_event__process,
					      session);
	else
		perf_event__synthesize_threads(perf_event__process, session);

	start_counters(evsel_list);
	first = list_entry(evsel_list->entries.next, struct perf_evsel, node);
	perf_session__set_sample_type(session, first->attr.sample_type);

	/* Wait for a minimal set of events before starting the snapshot */
	poll(evsel_list->pollfd, evsel_list->nr_fds, 100);

	perf_session__mmap_read(session);

	if (pthread_create(&thread, NULL, display_thread, session)) {
		printf("Could not create display thread.\n");
		exit(-1);
	}

	if (realtime_prio) {
		struct sched_param param;

		param.sched_priority = realtime_prio;
		if (sched_setscheduler(0, SCHED_FIFO, &param)) {
			printf("Could not set realtime priority.\n");
			exit(-1);
		}
	}

	while (1) {
		int hits = samples;

		perf_session__mmap_read(session);

		if (hits == samples)
			ret = poll(evsel_list->pollfd, evsel_list->nr_fds, 100);
	}

	return 0;
}

static const char * const top_usage[] = {
	"perf top [<options>]",
	NULL
};

static const struct option options[] = {
	OPT_CALLBACK('e', "event", &evsel_list, "event",
		     "event selector. use 'perf list' to list available events",
		     parse_events),
	OPT_INTEGER('c', "count", &default_interval,
		    "event period to sample"),
	OPT_INTEGER('p', "pid", &target_pid,
		    "profile events on existing process id"),
	OPT_INTEGER('t', "tid", &target_tid,
		    "profile events on existing thread id"),
	OPT_BOOLEAN('a', "all-cpus", &system_wide,
			    "system-wide collection from all CPUs"),
	OPT_STRING('C', "cpu", &cpu_list, "cpu",
		    "list of cpus to monitor"),
	OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name,
		   "file", "vmlinux pathname"),
	OPT_BOOLEAN('K', "hide_kernel_symbols", &hide_kernel_symbols,
		    "hide kernel symbols"),
	OPT_UINTEGER('m', "mmap-pages", &mmap_pages, "number of mmap data pages"),
	OPT_INTEGER('r', "realtime", &realtime_prio,
		    "collect data with this RT SCHED_FIFO priority"),
	OPT_INTEGER('d', "delay", &delay_secs,
		    "number of seconds to delay between refreshes"),
	OPT_BOOLEAN('D', "dump-symtab", &dump_symtab,
			    "dump the symbol table used for profiling"),
	OPT_INTEGER('f', "count-filter", &count_filter,
		    "only display functions with more events than this"),
	OPT_BOOLEAN('g', "group", &group,
			    "put the counters into a counter group"),
	OPT_BOOLEAN('i', "inherit", &inherit,
		    "child tasks inherit counters"),
	OPT_STRING('s', "sym-annotate", &sym_filter, "symbol name",
		    "symbol to annotate"),
	OPT_BOOLEAN('z', "zero", &zero,
		    "zero history across updates"),
	OPT_INTEGER('F', "freq", &freq,
		    "profile at this frequency"),
	OPT_INTEGER('E', "entries", &print_entries,
		    "display this many functions"),
	OPT_BOOLEAN('U', "hide_user_symbols", &hide_user_symbols,
		    "hide user symbols"),
	OPT_INCR('v', "verbose", &verbose,
		    "be more verbose (show counter open errors, etc)"),
	OPT_END()
};

int cmd_top(int argc, const char **argv, const char *prefix __used)
{
	struct perf_evsel *pos;
	int status = -ENOMEM;

	evsel_list = perf_evlist__new(NULL, NULL);
	if (evsel_list == NULL)
		return -ENOMEM;

	page_size = sysconf(_SC_PAGE_SIZE);

	argc = parse_options(argc, argv, options, top_usage, 0);
	if (argc)
		usage_with_options(top_usage, options);

	/* CPU and PID are mutually exclusive */
	if (target_tid > 0 && cpu_list) {
		printf("WARNING: PID switch overriding CPU\n");
		sleep(1);
		cpu_list = NULL;
	}

	if (target_pid != -1)
		target_tid = target_pid;

	if (perf_evlist__create_maps(evsel_list, target_pid,
				     target_tid, cpu_list) < 0)
		usage_with_options(top_usage, options);

	if (!evsel_list->nr_entries &&
	    perf_evlist__add_default(evsel_list) < 0) {
		pr_err("Not enough memory for event selector list\n");
		return -ENOMEM;
	}

	if (delay_secs < 1)
		delay_secs = 1;

	/*
	 * User specified count overrides default frequency.
	 */
	if (default_interval)
		freq = 0;
	else if (freq) {
		default_interval = freq;
	} else {
		fprintf(stderr, "frequency and count are zero, aborting\n");
		exit(EXIT_FAILURE);
	}

	list_for_each_entry(pos, &evsel_list->entries, node) {
		if (perf_evsel__alloc_fd(pos, evsel_list->cpus->nr,
					 evsel_list->threads->nr) < 0)
			goto out_free_fd;
		/*
		 * Fill in the ones not specifically initialized via -c:
		 */
		if (pos->attr.sample_period)
			continue;

		pos->attr.sample_period = default_interval;
	}

	if (perf_evlist__alloc_pollfd(evsel_list) < 0 ||
	    perf_evlist__alloc_mmap(evsel_list) < 0)
		goto out_free_fd;

	sym_evsel = list_entry(evsel_list->entries.next, struct perf_evsel, node);

	symbol_conf.priv_size = (sizeof(struct sym_entry) +
				 (evsel_list->nr_entries + 1) * sizeof(unsigned long));

	symbol_conf.try_vmlinux_path = (symbol_conf.vmlinux_name == NULL);
	if (symbol__init() < 0)
		return -1;

	get_term_dimensions(&winsize);
	if (print_entries == 0) {
		update_print_entries(&winsize);
		signal(SIGWINCH, sig_winch_handler);
	}

	status = __cmd_top();
out_free_fd:
	perf_evlist__delete(evsel_list);

	return status;
}