builtin-top.c

		event__process_mmap(event);
		break;
	default:
		break;
	}

	return 0;
}

struct mmap_data {
	int			counter;
	void			*base;
	int			mask;
	unsigned int		prev;
};

static unsigned int mmap_read_head(struct mmap_data *md)
{
	struct perf_event_mmap_page *pc = md->base;
	int head;

	head = pc->data_head;
	rmb();

	return head;
}

static void mmap_read_counter(struct mmap_data *md)
{
	unsigned int head = mmap_read_head(md);
	unsigned int old = md->prev;
	unsigned char *data = md->base + page_size;
	int diff;

	/*
	 * If we're further behind than half the buffer, there's a chance
	 * the writer will bite our tail and mess up the samples under us.
	 *
	 * If we somehow ended up ahead of the head, we got messed up.
	 *
	 * In either case, truncate and restart at head.
	 */
	diff = head - old;
	if (diff > md->mask / 2 || diff < 0) {
		fprintf(stderr, "WARNING: failed to keep up with mmap data.\n");

		/*
		 * head points to a known good entry, start there.
		 */
		old = head;
	}

	for (; old != head;) {
		event_t *event = (event_t *)&data[old & md->mask];

		event_t event_copy;

		size_t size = event->header.size;

		/*
		 * Event straddles the mmap boundary -- header should always
		 * be inside due to u64 alignment of output.
		 */
		if ((old & md->mask) + size != ((old + size) & md->mask)) {
			unsigned int offset = old;
			unsigned int len = min(sizeof(*event), size), cpy;
			void *dst = &event_copy;

			do {
				cpy = min(md->mask + 1 - (offset & md->mask), len);
				memcpy(dst, &data[offset & md->mask], cpy);
				offset += cpy;
				dst += cpy;
				len -= cpy;
			} while (len);

			event = &event_copy;
		}

		if (event->header.type == PERF_RECORD_SAMPLE)
			event__process_sample(event, md->counter);
		else
			event__process(event);
		old += size;
	}

	md->prev = old;
}

static struct pollfd event_array[MAX_NR_CPUS * MAX_COUNTERS];
static struct mmap_data mmap_array[MAX_NR_CPUS][MAX_COUNTERS];

static void mmap_read(void)
{
	int i, counter;

	for (i = 0; i < nr_cpus; i++) {
		for (counter = 0; counter < nr_counters; counter++)
			mmap_read_counter(&mmap_array[i][counter]);
	}
}

int nr_poll;
int group_fd;

static void start_counter(int i, int counter)
{
	struct perf_event_attr *attr;
	int cpu;

	cpu = profile_cpu;
	if (target_pid == -1 && profile_cpu == -1)
		cpu = i;

	attr = attrs + counter;

	attr->sample_type	= PERF_SAMPLE_IP | PERF_SAMPLE_TID;

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

	attr->inherit		= (cpu < 0) && inherit;
	attr->mmap		= 1;

try_again:
	fd[i][counter] = sys_perf_event_open(attr, target_pid, cpu, group_fd, 0);

	if (fd[i][counter] < 0) {
		int err = errno;

		if (err == EPERM || err == EACCES)
			die("No permission - are you root?\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("perfcounter syscall returned with %d (%s)\n",
			fd[i][counter], strerror(err));
		die("No CONFIG_PERF_EVENTS=y kernel support configured?\n");
		exit(-1);
	}
	assert(fd[i][counter] >= 0);
	fcntl(fd[i][counter], F_SETFL, O_NONBLOCK);

	/*
	 * First counter acts as the group leader:
	 */
	if (group && group_fd == -1)
		group_fd = fd[i][counter];

	event_array[nr_poll].fd = fd[i][counter];
	event_array[nr_poll].events = POLLIN;
	nr_poll++;

	mmap_array[i][counter].counter = counter;
	mmap_array[i][counter].prev = 0;
	mmap_array[i][counter].mask = mmap_pages*page_size - 1;
	mmap_array[i][counter].base = mmap(NULL, (mmap_pages+1)*page_size,
			PROT_READ, MAP_SHARED, fd[i][counter], 0);
	if (mmap_array[i][counter].base == MAP_FAILED)
		die("failed to mmap with %d (%s)\n", errno, strerror(errno));
}

static int __cmd_top(void)
{
	pthread_t thread;
	int i, counter;
	int ret;

	if (target_pid != -1)
		event__synthesize_thread(target_pid, event__process);
	else
		event__synthesize_threads(event__process);

	for (i = 0; i < nr_cpus; i++) {
		group_fd = -1;
		for (counter = 0; counter < nr_counters; counter++)
			start_counter(i, counter);
	}

	/* Wait for a minimal set of events before starting the snapshot */
	poll(event_array, nr_poll, 100);

	mmap_read();

	if (pthread_create(&thread, NULL, display_thread, NULL)) {
		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;

		mmap_read();

		if (hits == samples)
			ret = poll(event_array, nr_poll, 100);
	}

	return 0;
}

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

static const struct option options[] = {
	OPT_CALLBACK('e', "event", NULL, "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 pid"),
	OPT_BOOLEAN('a', "all-cpus", &system_wide,
			    "system-wide collection from all CPUs"),
	OPT_INTEGER('C', "CPU", &profile_cpu,
		    "CPU to profile on"),
	OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name,
		   "file", "vmlinux pathname"),
	OPT_BOOLEAN('K', "hide_kernel_symbols", &hide_kernel_symbols,
		    "hide kernel symbols"),
	OPT_INTEGER('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 - requires -k option"),
	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_BOOLEAN('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)
{
	int counter;

	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_pid != -1 && profile_cpu != -1) {
		printf("WARNING: PID switch overriding CPU\n");
		sleep(1);
		profile_cpu = -1;
	}

	if (!nr_counters)
		nr_counters = 1;

	symbol_conf.priv_size = (sizeof(struct sym_entry) +
				 (nr_counters + 1) * sizeof(unsigned long));
	if (symbol_conf.vmlinux_name == NULL)
		symbol_conf.try_vmlinux_path = true;
	if (symbol__init(&symbol_conf) < 0)
		return -1;

	if (delay_secs < 1)
		delay_secs = 1;

	parse_source(sym_filter_entry);

	/*
	 * 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);
	}

	/*
	 * Fill in the ones not specifically initialized via -c:
	 */
	for (counter = 0; counter < nr_counters; counter++) {
		if (attrs[counter].sample_period)
			continue;

		attrs[counter].sample_period = default_interval;
	}

	nr_cpus = sysconf(_SC_NPROCESSORS_ONLN);
	assert(nr_cpus <= MAX_NR_CPUS);
	assert(nr_cpus >= 0);

	if (target_pid != -1 || profile_cpu != -1)
		nr_cpus = 1;

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

	return __cmd_top();
}