evlist.c 34.6 KB
Newer Older
1 2 3 4 5 6 7 8
/*
 * Copyright (C) 2011, Red Hat Inc, Arnaldo Carvalho de Melo <acme@redhat.com>
 *
 * Parts came from builtin-{top,stat,record}.c, see those files for further
 * copyright notes.
 *
 * Released under the GPL v2. (and only v2, not any later version)
 */
9
#include "util.h"
10
#include <api/fs/debugfs.h>
11
#include <poll.h>
12 13
#include "cpumap.h"
#include "thread_map.h"
14
#include "target.h"
15 16
#include "evlist.h"
#include "evsel.h"
Adrian Hunter's avatar
Adrian Hunter committed
17
#include "debug.h"
18
#include <unistd.h>
19

20
#include "parse-events.h"
21
#include "parse-options.h"
22

23 24
#include <sys/mman.h>

25 26 27
#include <linux/bitops.h>
#include <linux/hash.h>

28 29 30
static void perf_evlist__mmap_put(struct perf_evlist *evlist, int idx);
static void __perf_evlist__munmap(struct perf_evlist *evlist, int idx);

31
#define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y))
32
#define SID(e, x, y) xyarray__entry(e->sample_id, x, y)
33

34 35
void perf_evlist__init(struct perf_evlist *evlist, struct cpu_map *cpus,
		       struct thread_map *threads)
36 37 38 39 40 41
{
	int i;

	for (i = 0; i < PERF_EVLIST__HLIST_SIZE; ++i)
		INIT_HLIST_HEAD(&evlist->heads[i]);
	INIT_LIST_HEAD(&evlist->entries);
42
	perf_evlist__set_maps(evlist, cpus, threads);
43
	fdarray__init(&evlist->pollfd, 64);
44
	evlist->workload.pid = -1;
45 46
}

47
struct perf_evlist *perf_evlist__new(void)
48 49 50
{
	struct perf_evlist *evlist = zalloc(sizeof(*evlist));

51
	if (evlist != NULL)
52
		perf_evlist__init(evlist, NULL, NULL);
53 54 55 56

	return evlist;
}

57 58 59 60 61 62 63 64 65 66 67 68
struct perf_evlist *perf_evlist__new_default(void)
{
	struct perf_evlist *evlist = perf_evlist__new();

	if (evlist && perf_evlist__add_default(evlist)) {
		perf_evlist__delete(evlist);
		evlist = NULL;
	}

	return evlist;
}

69 70 71 72 73 74 75 76 77 78 79 80 81 82 83
/**
 * perf_evlist__set_id_pos - set the positions of event ids.
 * @evlist: selected event list
 *
 * Events with compatible sample types all have the same id_pos
 * and is_pos.  For convenience, put a copy on evlist.
 */
void perf_evlist__set_id_pos(struct perf_evlist *evlist)
{
	struct perf_evsel *first = perf_evlist__first(evlist);

	evlist->id_pos = first->id_pos;
	evlist->is_pos = first->is_pos;
}

84 85 86 87
static void perf_evlist__update_id_pos(struct perf_evlist *evlist)
{
	struct perf_evsel *evsel;

88
	evlist__for_each(evlist, evsel)
89 90 91 92 93
		perf_evsel__calc_id_pos(evsel);

	perf_evlist__set_id_pos(evlist);
}

94 95 96 97
static void perf_evlist__purge(struct perf_evlist *evlist)
{
	struct perf_evsel *pos, *n;

98
	evlist__for_each_safe(evlist, n, pos) {
99 100 101 102 103 104 105
		list_del_init(&pos->node);
		perf_evsel__delete(pos);
	}

	evlist->nr_entries = 0;
}

106
void perf_evlist__exit(struct perf_evlist *evlist)
107
{
108
	zfree(&evlist->mmap);
109
	fdarray__exit(&evlist->pollfd);
110 111 112 113
}

void perf_evlist__delete(struct perf_evlist *evlist)
{
114
	perf_evlist__munmap(evlist);
115
	perf_evlist__close(evlist);
116 117 118 119
	cpu_map__delete(evlist->cpus);
	thread_map__delete(evlist->threads);
	evlist->cpus = NULL;
	evlist->threads = NULL;
120 121
	perf_evlist__purge(evlist);
	perf_evlist__exit(evlist);
122 123 124 125 126 127
	free(evlist);
}

void perf_evlist__add(struct perf_evlist *evlist, struct perf_evsel *entry)
{
	list_add_tail(&entry->node, &evlist->entries);
128
	entry->idx = evlist->nr_entries;
129
	entry->tracking = !entry->idx;
130

131 132
	if (!evlist->nr_entries++)
		perf_evlist__set_id_pos(evlist);
133 134
}

135 136 137
void perf_evlist__splice_list_tail(struct perf_evlist *evlist,
				   struct list_head *list,
				   int nr_entries)
138
{
139 140
	bool set_id_pos = !evlist->nr_entries;

141 142
	list_splice_tail(list, &evlist->entries);
	evlist->nr_entries += nr_entries;
143 144
	if (set_id_pos)
		perf_evlist__set_id_pos(evlist);
145 146
}

147 148 149 150 151
void __perf_evlist__set_leader(struct list_head *list)
{
	struct perf_evsel *evsel, *leader;

	leader = list_entry(list->next, struct perf_evsel, node);
152 153 154
	evsel = list_entry(list->prev, struct perf_evsel, node);

	leader->nr_members = evsel->idx - leader->idx + 1;
155

156
	__evlist__for_each(list, evsel) {
157
		evsel->leader = leader;
158 159 160 161
	}
}

void perf_evlist__set_leader(struct perf_evlist *evlist)
162
{
163 164
	if (evlist->nr_entries) {
		evlist->nr_groups = evlist->nr_entries > 1 ? 1 : 0;
165
		__perf_evlist__set_leader(&evlist->entries);
166
	}
167 168
}

169 170 171 172 173 174
int perf_evlist__add_default(struct perf_evlist *evlist)
{
	struct perf_event_attr attr = {
		.type = PERF_TYPE_HARDWARE,
		.config = PERF_COUNT_HW_CPU_CYCLES,
	};
175 176 177
	struct perf_evsel *evsel;

	event_attr_init(&attr);
178

179
	evsel = perf_evsel__new(&attr);
180
	if (evsel == NULL)
181 182 183 184 185 186
		goto error;

	/* use strdup() because free(evsel) assumes name is allocated */
	evsel->name = strdup("cycles");
	if (!evsel->name)
		goto error_free;
187 188 189

	perf_evlist__add(evlist, evsel);
	return 0;
190 191 192 193
error_free:
	perf_evsel__delete(evsel);
error:
	return -ENOMEM;
194
}
195

196 197
static int perf_evlist__add_attrs(struct perf_evlist *evlist,
				  struct perf_event_attr *attrs, size_t nr_attrs)
198 199 200 201 202 203
{
	struct perf_evsel *evsel, *n;
	LIST_HEAD(head);
	size_t i;

	for (i = 0; i < nr_attrs; i++) {
204
		evsel = perf_evsel__new_idx(attrs + i, evlist->nr_entries + i);
205 206 207 208 209 210 211 212 213 214
		if (evsel == NULL)
			goto out_delete_partial_list;
		list_add_tail(&evsel->node, &head);
	}

	perf_evlist__splice_list_tail(evlist, &head, nr_attrs);

	return 0;

out_delete_partial_list:
215
	__evlist__for_each_safe(&head, n, evsel)
216 217 218 219
		perf_evsel__delete(evsel);
	return -1;
}

220 221 222 223 224 225 226 227 228 229 230
int __perf_evlist__add_default_attrs(struct perf_evlist *evlist,
				     struct perf_event_attr *attrs, size_t nr_attrs)
{
	size_t i;

	for (i = 0; i < nr_attrs; i++)
		event_attr_init(attrs + i);

	return perf_evlist__add_attrs(evlist, attrs, nr_attrs);
}

231 232
struct perf_evsel *
perf_evlist__find_tracepoint_by_id(struct perf_evlist *evlist, int id)
233 234 235
{
	struct perf_evsel *evsel;

236
	evlist__for_each(evlist, evsel) {
237 238 239 240 241 242 243 244
		if (evsel->attr.type   == PERF_TYPE_TRACEPOINT &&
		    (int)evsel->attr.config == id)
			return evsel;
	}

	return NULL;
}

245 246 247 248 249 250
struct perf_evsel *
perf_evlist__find_tracepoint_by_name(struct perf_evlist *evlist,
				     const char *name)
{
	struct perf_evsel *evsel;

251
	evlist__for_each(evlist, evsel) {
252 253 254 255 256 257 258 259
		if ((evsel->attr.type == PERF_TYPE_TRACEPOINT) &&
		    (strcmp(evsel->name, name) == 0))
			return evsel;
	}

	return NULL;
}

260 261 262
int perf_evlist__add_newtp(struct perf_evlist *evlist,
			   const char *sys, const char *name, void *handler)
{
263
	struct perf_evsel *evsel = perf_evsel__newtp(sys, name);
264 265 266 267

	if (evsel == NULL)
		return -1;

268
	evsel->handler = handler;
269 270 271 272
	perf_evlist__add(evlist, evsel);
	return 0;
}

273 274 275 276 277 278 279 280 281
static int perf_evlist__nr_threads(struct perf_evlist *evlist,
				   struct perf_evsel *evsel)
{
	if (evsel->system_wide)
		return 1;
	else
		return thread_map__nr(evlist->threads);
}

282 283 284 285
void perf_evlist__disable(struct perf_evlist *evlist)
{
	int cpu, thread;
	struct perf_evsel *pos;
286
	int nr_cpus = cpu_map__nr(evlist->cpus);
287
	int nr_threads;
288

289
	for (cpu = 0; cpu < nr_cpus; cpu++) {
290
		evlist__for_each(evlist, pos) {
291
			if (!perf_evsel__is_group_leader(pos) || !pos->fd)
292
				continue;
293
			nr_threads = perf_evlist__nr_threads(evlist, pos);
294
			for (thread = 0; thread < nr_threads; thread++)
295 296
				ioctl(FD(pos, cpu, thread),
				      PERF_EVENT_IOC_DISABLE, 0);
297 298 299 300
		}
	}
}

301 302 303 304
void perf_evlist__enable(struct perf_evlist *evlist)
{
	int cpu, thread;
	struct perf_evsel *pos;
305
	int nr_cpus = cpu_map__nr(evlist->cpus);
306
	int nr_threads;
307

308
	for (cpu = 0; cpu < nr_cpus; cpu++) {
309
		evlist__for_each(evlist, pos) {
310
			if (!perf_evsel__is_group_leader(pos) || !pos->fd)
311
				continue;
312
			nr_threads = perf_evlist__nr_threads(evlist, pos);
313
			for (thread = 0; thread < nr_threads; thread++)
314 315
				ioctl(FD(pos, cpu, thread),
				      PERF_EVENT_IOC_ENABLE, 0);
316 317 318 319
		}
	}
}

320 321 322 323
int perf_evlist__disable_event(struct perf_evlist *evlist,
			       struct perf_evsel *evsel)
{
	int cpu, thread, err;
324 325
	int nr_cpus = cpu_map__nr(evlist->cpus);
	int nr_threads = perf_evlist__nr_threads(evlist, evsel);
326 327 328 329

	if (!evsel->fd)
		return 0;

330 331
	for (cpu = 0; cpu < nr_cpus; cpu++) {
		for (thread = 0; thread < nr_threads; thread++) {
332 333 334 335 336 337 338 339 340 341 342 343 344
			err = ioctl(FD(evsel, cpu, thread),
				    PERF_EVENT_IOC_DISABLE, 0);
			if (err)
				return err;
		}
	}
	return 0;
}

int perf_evlist__enable_event(struct perf_evlist *evlist,
			      struct perf_evsel *evsel)
{
	int cpu, thread, err;
345 346
	int nr_cpus = cpu_map__nr(evlist->cpus);
	int nr_threads = perf_evlist__nr_threads(evlist, evsel);
347 348 349 350

	if (!evsel->fd)
		return -EINVAL;

351 352
	for (cpu = 0; cpu < nr_cpus; cpu++) {
		for (thread = 0; thread < nr_threads; thread++) {
353 354 355 356 357 358 359 360 361
			err = ioctl(FD(evsel, cpu, thread),
				    PERF_EVENT_IOC_ENABLE, 0);
			if (err)
				return err;
		}
	}
	return 0;
}

362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408
static int perf_evlist__enable_event_cpu(struct perf_evlist *evlist,
					 struct perf_evsel *evsel, int cpu)
{
	int thread, err;
	int nr_threads = perf_evlist__nr_threads(evlist, evsel);

	if (!evsel->fd)
		return -EINVAL;

	for (thread = 0; thread < nr_threads; thread++) {
		err = ioctl(FD(evsel, cpu, thread),
			    PERF_EVENT_IOC_ENABLE, 0);
		if (err)
			return err;
	}
	return 0;
}

static int perf_evlist__enable_event_thread(struct perf_evlist *evlist,
					    struct perf_evsel *evsel,
					    int thread)
{
	int cpu, err;
	int nr_cpus = cpu_map__nr(evlist->cpus);

	if (!evsel->fd)
		return -EINVAL;

	for (cpu = 0; cpu < nr_cpus; cpu++) {
		err = ioctl(FD(evsel, cpu, thread), PERF_EVENT_IOC_ENABLE, 0);
		if (err)
			return err;
	}
	return 0;
}

int perf_evlist__enable_event_idx(struct perf_evlist *evlist,
				  struct perf_evsel *evsel, int idx)
{
	bool per_cpu_mmaps = !cpu_map__empty(evlist->cpus);

	if (per_cpu_mmaps)
		return perf_evlist__enable_event_cpu(evlist, evsel, idx);
	else
		return perf_evlist__enable_event_thread(evlist, evsel, idx);
}

409
int perf_evlist__alloc_pollfd(struct perf_evlist *evlist)
410
{
411 412
	int nr_cpus = cpu_map__nr(evlist->cpus);
	int nr_threads = thread_map__nr(evlist->threads);
413 414 415
	int nfds = 0;
	struct perf_evsel *evsel;

416
	evlist__for_each(evlist, evsel) {
417 418 419 420 421 422
		if (evsel->system_wide)
			nfds += nr_cpus;
		else
			nfds += nr_cpus * nr_threads;
	}

423 424
	if (fdarray__available_entries(&evlist->pollfd) < nfds &&
	    fdarray__grow(&evlist->pollfd, nfds) < 0)
425 426 427
		return -ENOMEM;

	return 0;
428
}
429

430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445
static int __perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd, int idx)
{
	int pos = fdarray__add(&evlist->pollfd, fd, POLLIN | POLLERR | POLLHUP);
	/*
	 * Save the idx so that when we filter out fds POLLHUP'ed we can
	 * close the associated evlist->mmap[] entry.
	 */
	if (pos >= 0) {
		evlist->pollfd.priv[pos].idx = idx;

		fcntl(fd, F_SETFL, O_NONBLOCK);
	}

	return pos;
}

446
int perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd)
447
{
448 449 450 451 452 453
	return __perf_evlist__add_pollfd(evlist, fd, -1);
}

static void perf_evlist__munmap_filtered(struct fdarray *fda, int fd)
{
	struct perf_evlist *evlist = container_of(fda, struct perf_evlist, pollfd);
454

455
	perf_evlist__mmap_put(evlist, fda->priv[fd].idx);
456
}
457

458 459
int perf_evlist__filter_pollfd(struct perf_evlist *evlist, short revents_and_mask)
{
460 461
	return fdarray__filter(&evlist->pollfd, revents_and_mask,
			       perf_evlist__munmap_filtered);
462 463
}

464 465
int perf_evlist__poll(struct perf_evlist *evlist, int timeout)
{
466
	return fdarray__poll(&evlist->pollfd, timeout);
467 468
}

469 470 471
static void perf_evlist__id_hash(struct perf_evlist *evlist,
				 struct perf_evsel *evsel,
				 int cpu, int thread, u64 id)
472 473 474 475 476 477 478 479 480 481
{
	int hash;
	struct perf_sample_id *sid = SID(evsel, cpu, thread);

	sid->id = id;
	sid->evsel = evsel;
	hash = hash_64(sid->id, PERF_EVLIST__HLIST_BITS);
	hlist_add_head(&sid->node, &evlist->heads[hash]);
}

482 483 484 485 486 487 488 489 490 491
void perf_evlist__id_add(struct perf_evlist *evlist, struct perf_evsel *evsel,
			 int cpu, int thread, u64 id)
{
	perf_evlist__id_hash(evlist, evsel, cpu, thread, id);
	evsel->id[evsel->ids++] = id;
}

static int perf_evlist__id_add_fd(struct perf_evlist *evlist,
				  struct perf_evsel *evsel,
				  int cpu, int thread, int fd)
492 493
{
	u64 read_data[4] = { 0, };
494
	int id_idx = 1; /* The first entry is the counter value */
495 496 497 498 499 500 501 502 503 504 505
	u64 id;
	int ret;

	ret = ioctl(fd, PERF_EVENT_IOC_ID, &id);
	if (!ret)
		goto add;

	if (errno != ENOTTY)
		return -1;

	/* Legacy way to get event id.. All hail to old kernels! */
506

507 508 509 510 511 512 513
	/*
	 * This way does not work with group format read, so bail
	 * out in that case.
	 */
	if (perf_evlist__read_format(evlist) & PERF_FORMAT_GROUP)
		return -1;

514 515 516 517 518 519 520 521 522
	if (!(evsel->attr.read_format & PERF_FORMAT_ID) ||
	    read(fd, &read_data, sizeof(read_data)) == -1)
		return -1;

	if (evsel->attr.read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
		++id_idx;
	if (evsel->attr.read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
		++id_idx;

523 524 525 526
	id = read_data[id_idx];

 add:
	perf_evlist__id_add(evlist, evsel, cpu, thread, id);
527 528 529
	return 0;
}

Adrian Hunter's avatar
Adrian Hunter committed
530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545
static void perf_evlist__set_sid_idx(struct perf_evlist *evlist,
				     struct perf_evsel *evsel, int idx, int cpu,
				     int thread)
{
	struct perf_sample_id *sid = SID(evsel, cpu, thread);
	sid->idx = idx;
	if (evlist->cpus && cpu >= 0)
		sid->cpu = evlist->cpus->map[cpu];
	else
		sid->cpu = -1;
	if (!evsel->system_wide && evlist->threads && thread >= 0)
		sid->tid = evlist->threads->map[thread];
	else
		sid->tid = -1;
}

546
struct perf_sample_id *perf_evlist__id2sid(struct perf_evlist *evlist, u64 id)
547 548 549 550 551 552 553 554
{
	struct hlist_head *head;
	struct perf_sample_id *sid;
	int hash;

	hash = hash_64(id, PERF_EVLIST__HLIST_BITS);
	head = &evlist->heads[hash];

555
	hlist_for_each_entry(sid, head, node)
556
		if (sid->id == id)
557 558 559 560 561 562 563 564 565 566 567 568 569 570 571
			return sid;

	return NULL;
}

struct perf_evsel *perf_evlist__id2evsel(struct perf_evlist *evlist, u64 id)
{
	struct perf_sample_id *sid;

	if (evlist->nr_entries == 1)
		return perf_evlist__first(evlist);

	sid = perf_evlist__id2sid(evlist, id);
	if (sid)
		return sid->evsel;
572 573

	if (!perf_evlist__sample_id_all(evlist))
574
		return perf_evlist__first(evlist);
575

576 577
	return NULL;
}
578

579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602
static int perf_evlist__event2id(struct perf_evlist *evlist,
				 union perf_event *event, u64 *id)
{
	const u64 *array = event->sample.array;
	ssize_t n;

	n = (event->header.size - sizeof(event->header)) >> 3;

	if (event->header.type == PERF_RECORD_SAMPLE) {
		if (evlist->id_pos >= n)
			return -1;
		*id = array[evlist->id_pos];
	} else {
		if (evlist->is_pos > n)
			return -1;
		n -= evlist->is_pos;
		*id = array[n];
	}
	return 0;
}

static struct perf_evsel *perf_evlist__event2evsel(struct perf_evlist *evlist,
						   union perf_event *event)
{
603
	struct perf_evsel *first = perf_evlist__first(evlist);
604 605 606 607 608 609
	struct hlist_head *head;
	struct perf_sample_id *sid;
	int hash;
	u64 id;

	if (evlist->nr_entries == 1)
610 611 612 613 614
		return first;

	if (!first->attr.sample_id_all &&
	    event->header.type != PERF_RECORD_SAMPLE)
		return first;
615 616 617 618 619 620

	if (perf_evlist__event2id(evlist, event, &id))
		return NULL;

	/* Synthesized events have an id of zero */
	if (!id)
621
		return first;
622 623 624 625 626 627 628 629 630 631 632

	hash = hash_64(id, PERF_EVLIST__HLIST_BITS);
	head = &evlist->heads[hash];

	hlist_for_each_entry(sid, head, node) {
		if (sid->id == id)
			return sid->evsel;
	}
	return NULL;
}

633
union perf_event *perf_evlist__mmap_read(struct perf_evlist *evlist, int idx)
634
{
635
	struct perf_mmap *md = &evlist->mmap[idx];
636 637 638
	unsigned int head = perf_mmap__read_head(md);
	unsigned int old = md->prev;
	unsigned char *data = md->base + page_size;
639
	union perf_event *event = NULL;
640

641
	if (evlist->overwrite) {
642
		/*
643 644 645 646 647 648
		 * 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.
649
		 */
650 651 652 653 654 655 656 657 658
		int 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;
		}
659 660 661 662 663
	}

	if (old != head) {
		size_t size;

664
		event = (union perf_event *)&data[old & md->mask];
665 666 667 668 669 670 671 672 673
		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;
674
			void *dst = md->event_copy;
675 676 677 678 679 680 681 682 683

			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);

684
			event = (union perf_event *) md->event_copy;
685 686 687 688 689 690
		}

		old += size;
	}

	md->prev = old;
691

692 693
	return event;
}
694

695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712
static bool perf_mmap__empty(struct perf_mmap *md)
{
	return perf_mmap__read_head(md) != md->prev;
}

static void perf_evlist__mmap_get(struct perf_evlist *evlist, int idx)
{
	++evlist->mmap[idx].refcnt;
}

static void perf_evlist__mmap_put(struct perf_evlist *evlist, int idx)
{
	BUG_ON(evlist->mmap[idx].refcnt == 0);

	if (--evlist->mmap[idx].refcnt == 0)
		__perf_evlist__munmap(evlist, idx);
}

713 714
void perf_evlist__mmap_consume(struct perf_evlist *evlist, int idx)
{
715 716
	struct perf_mmap *md = &evlist->mmap[idx];

717 718 719 720 721
	if (!evlist->overwrite) {
		unsigned int old = md->prev;

		perf_mmap__write_tail(md, old);
	}
722 723 724

	if (md->refcnt == 1 && perf_mmap__empty(md))
		perf_evlist__mmap_put(evlist, idx);
725 726
}

727 728 729 730 731
static void __perf_evlist__munmap(struct perf_evlist *evlist, int idx)
{
	if (evlist->mmap[idx].base != NULL) {
		munmap(evlist->mmap[idx].base, evlist->mmap_len);
		evlist->mmap[idx].base = NULL;
732
		evlist->mmap[idx].refcnt = 0;
733 734 735
	}
}

736
void perf_evlist__munmap(struct perf_evlist *evlist)
737
{
738
	int i;
739

740 741 742
	if (evlist->mmap == NULL)
		return;

743 744
	for (i = 0; i < evlist->nr_mmaps; i++)
		__perf_evlist__munmap(evlist, i);
745

746
	zfree(&evlist->mmap);
747 748
}

749
static int perf_evlist__alloc_mmap(struct perf_evlist *evlist)
750
{
751
	evlist->nr_mmaps = cpu_map__nr(evlist->cpus);
752
	if (cpu_map__empty(evlist->cpus))
753
		evlist->nr_mmaps = thread_map__nr(evlist->threads);
754
	evlist->mmap = zalloc(evlist->nr_mmaps * sizeof(struct perf_mmap));
755 756 757
	return evlist->mmap != NULL ? 0 : -ENOMEM;
}

758 759 760 761 762 763 764
struct mmap_params {
	int prot;
	int mask;
};

static int __perf_evlist__mmap(struct perf_evlist *evlist, int idx,
			       struct mmap_params *mp, int fd)
765
{
766 767 768 769 770 771 772 773 774 775 776 777 778 779
	/*
	 * The last one will be done at perf_evlist__mmap_consume(), so that we
	 * make sure we don't prevent tools from consuming every last event in
	 * the ring buffer.
	 *
	 * I.e. we can get the POLLHUP meaning that the fd doesn't exist
	 * anymore, but the last events for it are still in the ring buffer,
	 * waiting to be consumed.
	 *
	 * Tools can chose to ignore this at their own discretion, but the
	 * evlist layer can't just drop it when filtering events in
	 * perf_evlist__filter_pollfd().
	 */
	evlist->mmap[idx].refcnt = 2;
780
	evlist->mmap[idx].prev = 0;
781 782
	evlist->mmap[idx].mask = mp->mask;
	evlist->mmap[idx].base = mmap(NULL, evlist->mmap_len, mp->prot,
783
				      MAP_SHARED, fd, 0);
784
	if (evlist->mmap[idx].base == MAP_FAILED) {
785 786
		pr_debug2("failed to mmap perf event ring buffer, error %d\n",
			  errno);
787
		evlist->mmap[idx].base = NULL;
788
		return -1;
789
	}
790

791 792 793
	return 0;
}

794
static int perf_evlist__mmap_per_evsel(struct perf_evlist *evlist, int idx,
795 796
				       struct mmap_params *mp, int cpu,
				       int thread, int *output)
797 798
{
	struct perf_evsel *evsel;
799

800
	evlist__for_each(evlist, evsel) {
801 802 803 804 805 806
		int fd;

		if (evsel->system_wide && thread)
			continue;

		fd = FD(evsel, cpu, thread);
807 808 809

		if (*output == -1) {
			*output = fd;
810
			if (__perf_evlist__mmap(evlist, idx, mp, *output) < 0)
811 812 813 814
				return -1;
		} else {
			if (ioctl(fd, PERF_EVENT_IOC_SET_OUTPUT, *output) != 0)
				return -1;
815 816

			perf_evlist__mmap_get(evlist, idx);
817 818
		}

819 820 821 822 823 824 825 826 827
		/*
		 * The system_wide flag causes a selected event to be opened
		 * always without a pid.  Consequently it will never get a
		 * POLLHUP, but it is used for tracking in combination with
		 * other events, so it should not need to be polled anyway.
		 * Therefore don't add it for polling.
		 */
		if (!evsel->system_wide &&
		    __perf_evlist__add_pollfd(evlist, fd, idx) < 0) {
828
			perf_evlist__mmap_put(evlist, idx);
829
			return -1;
830
		}
831

Adrian Hunter's avatar
Adrian Hunter committed
832 833 834 835 836 837 838
		if (evsel->attr.read_format & PERF_FORMAT_ID) {
			if (perf_evlist__id_add_fd(evlist, evsel, cpu, thread,
						   fd) < 0)
				return -1;
			perf_evlist__set_sid_idx(evlist, evsel, idx, cpu,
						 thread);
		}
839 840 841 842 843
	}

	return 0;
}

844 845
static int perf_evlist__mmap_per_cpu(struct perf_evlist *evlist,
				     struct mmap_params *mp)
846
{
847
	int cpu, thread;
848 849
	int nr_cpus = cpu_map__nr(evlist->cpus);
	int nr_threads = thread_map__nr(evlist->threads);
850

Adrian Hunter's avatar
Adrian Hunter committed
851
	pr_debug2("perf event ring buffer mmapped per cpu\n");
852
	for (cpu = 0; cpu < nr_cpus; cpu++) {
853 854
		int output = -1;

855
		for (thread = 0; thread < nr_threads; thread++) {
856 857
			if (perf_evlist__mmap_per_evsel(evlist, cpu, mp, cpu,
							thread, &output))
858
				goto out_unmap;
859 860 861 862 863 864
		}
	}

	return 0;

out_unmap:
865 866
	for (cpu = 0; cpu < nr_cpus; cpu++)
		__perf_evlist__munmap(evlist, cpu);
867 868 869
	return -1;
}

870 871
static int perf_evlist__mmap_per_thread(struct perf_evlist *evlist,
					struct mmap_params *mp)
872 873
{
	int thread;
874
	int nr_threads = thread_map__nr(evlist->threads);
875

Adrian Hunter's avatar
Adrian Hunter committed
876
	pr_debug2("perf event ring buffer mmapped per thread\n");
877
	for (thread = 0; thread < nr_threads; thread++) {
878 879
		int output = -1;

880 881
		if (perf_evlist__mmap_per_evsel(evlist, thread, mp, 0, thread,
						&output))
882
			goto out_unmap;
883 884 885 886 887
	}

	return 0;

out_unmap:
888 889
	for (thread = 0; thread < nr_threads; thread++)
		__perf_evlist__munmap(evlist, thread);
890 891 892
	return -1;
}

893 894 895 896 897 898 899 900 901 902 903
static size_t perf_evlist__mmap_size(unsigned long pages)
{
	/* 512 kiB: default amount of unprivileged mlocked memory */
	if (pages == UINT_MAX)
		pages = (512 * 1024) / page_size;
	else if (!is_power_of_2(pages))
		return 0;

	return (pages + 1) * page_size;
}

904 905
static long parse_pages_arg(const char *str, unsigned long min,
			    unsigned long max)
906
{
907
	unsigned long pages, val;
908 909 910 911 912 913 914
	static struct parse_tag tags[] = {
		{ .tag  = 'B', .mult = 1       },
		{ .tag  = 'K', .mult = 1 << 10 },
		{ .tag  = 'M', .mult = 1 << 20 },
		{ .tag  = 'G', .mult = 1 << 30 },
		{ .tag  = 0 },
	};
915

916
	if (str == NULL)
917
		return -EINVAL;
918

919
	val = parse_tag_value(str, tags);
920
	if (val != (unsigned long) -1) {
921 922 923 924 925 926
		/* we got file size value */
		pages = PERF_ALIGN(val, page_size) / page_size;
	} else {
		/* we got pages count value */
		char *eptr;
		pages = strtoul(str, &eptr, 10);
927 928
		if (*eptr != '\0')
			return -EINVAL;
929 930
	}

931
	if (pages == 0 && min == 0) {
932
		/* leave number of pages at 0 */
933
	} else if (!is_power_of_2(pages)) {
934
		/* round pages up to next power of 2 */
935 936 937
		pages = next_pow2_l(pages);
		if (!pages)
			return -EINVAL;
938 939
		pr_info("rounding mmap pages size to %lu bytes (%lu pages)\n",
			pages * page_size, pages);
940 941
	}

942 943 944 945 946 947 948 949 950 951 952 953 954
	if (pages > max)
		return -EINVAL;

	return pages;
}

int perf_evlist__parse_mmap_pages(const struct option *opt, const char *str,
				  int unset __maybe_unused)
{
	unsigned int *mmap_pages = opt->value;
	unsigned long max = UINT_MAX;
	long pages;

955
	if (max > SIZE_MAX / page_size)
956 957 958 959 960
		max = SIZE_MAX / page_size;

	pages = parse_pages_arg(str, 1, max);
	if (pages < 0) {
		pr_err("Invalid argument for --mmap_pages/-m\n");
961 962 963 964 965 966 967
		return -1;
	}

	*mmap_pages = pages;
	return 0;
}

968 969 970 971 972
/**
 * perf_evlist__mmap - Create mmaps to receive events.
 * @evlist: list of events
 * @pages: map length in pages
 * @overwrite: overwrite older events?
973
 *
974 975 976
 * If @overwrite is %false the user needs to signal event consumption using
 * perf_mmap__write_tail().  Using perf_evlist__mmap_read() does this
 * automatically.
977
 *
978
 * Return: %0 on success, negative error code otherwise.
979
 */
980 981
int perf_evlist__mmap(struct perf_evlist *evlist, unsigned int pages,
		      bool overwrite)
982
{
983
	struct perf_evsel *evsel;
984 985
	const struct cpu_map *cpus = evlist->cpus;
	const struct thread_map *threads = evlist->threads;
986 987 988
	struct mmap_params mp = {
		.prot = PROT_READ | (overwrite ? 0 : PROT_WRITE),
	};
989

990
	if (evlist->mmap == NULL && perf_evlist__alloc_mmap(evlist) < 0)
991 992
		return -ENOMEM;

993
	if (evlist->pollfd.entries == NULL && perf_evlist__alloc_pollfd(evlist) < 0)
994 995 996
		return -ENOMEM;

	evlist->overwrite = overwrite;
997
	evlist->mmap_len = perf_evlist__mmap_size(pages);
998
	pr_debug("mmap size %zuB\n", evlist->mmap_len);
999
	mp.mask = evlist->mmap_len - page_size - 1;
1000

1001
	evlist__for_each(evlist, evsel) {
1002
		if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
1003
		    evsel->sample_id == NULL &&
1004
		    perf_evsel__alloc_id(evsel, cpu_map__nr(cpus), threads->nr) < 0)
1005 1006 1007
			return -ENOMEM;
	}

1008
	if (cpu_map__empty(cpus))
1009
		return perf_evlist__mmap_per_thread(evlist, &mp);
1010

1011
	return perf_evlist__mmap_per_cpu(evlist, &mp);
1012
}
1013

1014
int perf_evlist__create_maps(struct perf_evlist *evlist, struct target *target)
1015
{
1016 1017
	evlist->threads = thread_map__new_str(target->pid, target->tid,
					      target->uid);
1018 1019 1020 1021

	if (evlist->threads == NULL)
		return -1;

1022
	if (target__uses_dummy_map(target))
1023
		evlist->cpus = cpu_map__dummy_new();
1024 1025
	else
		evlist->cpus = cpu_map__new(target->cpu_list);
1026 1027 1028 1029 1030 1031 1032 1033

	if (evlist->cpus == NULL)
		goto out_delete_threads;

	return 0;

out_delete_threads:
	thread_map__delete(evlist->threads);
1034
	evlist->threads = NULL;
1035 1036 1037
	return -1;
}

1038
int perf_evlist__apply_filters(struct perf_evlist *evlist)
1039 1040
{
	struct perf_evsel *evsel;
1041 1042
	int err = 0;
	const int ncpus = cpu_map__nr(evlist->cpus),
1043
		  nthreads = thread_map__nr(evlist->threads);
1044

1045
	evlist__for_each(evlist, evsel) {
1046
		if (evsel->filter == NULL)
1047
			continue;
1048 1049 1050 1051

		err = perf_evsel__set_filter(evsel, ncpus, nthreads, evsel->filter);
		if (err)
			break;
1052 1053
	}

1054 1055 1056 1057 1058 1059 1060 1061
	return err;
}

int perf_evlist__set_filter(struct perf_evlist *evlist, const char *filter)
{
	struct perf_evsel *evsel;
	int err = 0;
	const int ncpus = cpu_map__nr(evlist->cpus),
1062
		  nthreads = thread_map__nr(evlist->threads);
1063

1064
	evlist__for_each(evlist, evsel) {
1065 1066 1067 1068 1069 1070
		err = perf_evsel__set_filter(evsel, ncpus, nthreads, filter);
		if (err)
			break;
	}

	return err;
1071
}
1072

1073
bool perf_evlist__valid_sample_type(struct perf_evlist *evlist)
1074
{
1075
	struct perf_evsel *pos;
1076

1077 1078 1079 1080 1081 1082
	if (evlist->nr_entries == 1)
		return true;

	if (evlist->id_pos < 0 || evlist->is_pos < 0)
		return false;

1083
	evlist__for_each(evlist, pos) {
1084 1085
		if (pos->id_pos != evlist->id_pos ||
		    pos->is_pos != evlist->is_pos)
1086
			return false;
1087 1088
	}

1089
	return true;
1090 1091
}

1092
u64 __perf_evlist__combined_sample_type(struct perf_evlist *evlist)
1093
{
1094 1095 1096 1097 1098
	struct perf_evsel *evsel;

	if (evlist->combined_sample_type)
		return evlist->combined_sample_type;

1099
	evlist__for_each(evlist, evsel)
1100 1101 1102 1103 1104 1105 1106 1107 1108
		evlist->combined_sample_type |= evsel->attr.sample_type;

	return evlist->combined_sample_type;
}

u64 perf_evlist__combined_sample_type(struct perf_evlist *evlist)
{
	evlist->combined_sample_type = 0;
	return __perf_evlist__combined_sample_type(evlist);
1109 1110
}

1111 1112 1113 1114 1115 1116
bool perf_evlist__valid_read_format(struct perf_evlist *evlist)
{
	struct perf_evsel *first = perf_evlist__first(evlist), *pos = first;
	u64 read_format = first->attr.read_format;
	u64 sample_type = first->attr.sample_type;

1117
	evlist__for_each(evlist, pos) {
1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136
		if (read_format != pos->attr.read_format)
			return false;
	}

	/* PERF_SAMPLE_READ imples PERF_FORMAT_ID. */
	if ((sample_type & PERF_SAMPLE_READ) &&
	    !(read_format & PERF_FORMAT_ID)) {
		return false;
	}

	return true;
}

u64 perf_evlist__read_format(struct perf_evlist *evlist)
{
	struct perf_evsel *first = perf_evlist__first(evlist);
	return first->attr.read_format;
}

1137
u16 perf_evlist__id_hdr_size(struct perf_evlist *evlist)
1138
{
1139
	struct perf_evsel *first = perf_evlist__first(evlist);
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162
	struct perf_sample *data;
	u64 sample_type;
	u16 size = 0;

	if (!first->attr.sample_id_all)
		goto out;

	sample_type = first->attr.sample_type;

	if (sample_type & PERF_SAMPLE_TID)
		size += sizeof(data->tid) * 2;

       if (sample_type & PERF_SAMPLE_TIME)
		size += sizeof(data->time);

	if (sample_type & PERF_SAMPLE_ID)
		size += sizeof(data->id);

	if (sample_type & PERF_SAMPLE_STREAM_ID)
		size += sizeof(data->stream_id);

	if (sample_type & PERF_SAMPLE_CPU)
		size += sizeof(data->cpu) * 2;
1163 1164 1165

	if (sample_type & PERF_SAMPLE_IDENTIFIER)
		size += sizeof(data->id);
1166 1167 1168 1169
out:
	return size;
}

1170
bool perf_evlist__valid_sample_id_all(struct perf_evlist *evlist)
1171
{
1172
	struct perf_evsel *first = perf_evlist__first(evlist), *pos = first;
1173

1174
	evlist__for_each_continue(evlist, pos) {
1175 1176
		if (first->attr.sample_id_all != pos->attr.sample_id_all)
			return false;
1177 1178
	}

1179 1180 1181
	return true;
}

1182
bool perf_evlist__sample_id_all(struct perf_evlist *evlist)
1183
{
1184
	struct perf_evsel *first = perf_evlist__first(evlist);
1185
	return first->attr.sample_id_all;
1186
}
1187 1188 1189 1190 1191 1192

void perf_evlist__set_selected(struct perf_evlist *evlist,
			       struct perf_evsel *evsel)
{
	evlist->selected = evsel;
}
1193

1194 1195 1196 1197 1198
void perf_evlist__close(struct perf_evlist *evlist)
{
	struct perf_evsel *evsel;
	int ncpus = cpu_map__nr(evlist->cpus);
	int nthreads = thread_map__nr(evlist->threads);
1199
	int n;
1200

1201 1202 1203 1204
	evlist__for_each_reverse(evlist, evsel) {
		n = evsel->cpus ? evsel->cpus->nr : ncpus;
		perf_evsel__close(evsel, n, nthreads);
	}
1205 1206
}

1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236
static int perf_evlist__create_syswide_maps(struct perf_evlist *evlist)
{
	int err = -ENOMEM;

	/*
	 * Try reading /sys/devices/system/cpu/online to get
	 * an all cpus map.
	 *
	 * FIXME: -ENOMEM is the best we can do here, the cpu_map
	 * code needs an overhaul to properly forward the
	 * error, and we may not want to do that fallback to a
	 * default cpu identity map :-\
	 */
	evlist->cpus = cpu_map__new(NULL);
	if (evlist->cpus == NULL)
		goto out;

	evlist->threads = thread_map__new_dummy();
	if (evlist->threads == NULL)
		goto out_free_cpus;

	err = 0;
out:
	return err;
out_free_cpus:
	cpu_map__delete(evlist->cpus);
	evlist->cpus = NULL;
	goto out;
}

1237
int perf_evlist__open(struct perf_evlist *evlist)
1238
{
1239
	struct perf_evsel *evsel;
1240
	int err;
1241

1242 1243 1244 1245 1246 1247 1248 1249 1250 1251
	/*
	 * Default: one fd per CPU, all threads, aka systemwide
	 * as sys_perf_event_open(cpu = -1, thread = -1) is EINVAL
	 */
	if (evlist->threads == NULL && evlist->cpus == NULL) {
		err = perf_evlist__create_syswide_maps(evlist);
		if (err < 0)
			goto out_err;
	}

1252 1253
	perf_evlist__update_id_pos(evlist);

1254
	evlist__for_each(evlist, evsel) {
1255
		err = perf_evsel__open(evsel, evlist->cpus, evlist->threads);
1256 1257 1258 1259 1260 1261
		if (err < 0)
			goto out_err;
	}

	return 0;
out_err:
1262
	perf_evlist__close(evlist);
1263
	errno = -err;
1264 1265
	return err;
}
1266

1267
int perf_evlist__prepare_workload(struct perf_evlist *evlist, struct target *target,
1268
				  const char *argv[], bool pipe_output,
1269
				  void (*exec_error)(int signo, siginfo_t *info, void *ucontext))
1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
{
	int child_ready_pipe[2], go_pipe[2];
	char bf;

	if (pipe(child_ready_pipe) < 0) {
		perror("failed to create 'ready' pipe");
		return -1;
	}

	if (pipe(go_pipe) < 0) {
		perror("failed to create 'go' pipe");
		goto out_close_ready_pipe;
	}

	evlist->workload.pid = fork();
	if (evlist->workload.pid < 0) {
		perror("failed to fork");
		goto out_close_pipes;
	}

	if (!evlist->workload.pid) {
1291 1292
		int ret;

1293
		if (pipe_output)
1294 1295
			dup2(2, 1);

1296 1297
		signal(SIGTERM, SIG_DFL);

1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309
		close(child_ready_pipe[0]);
		close(go_pipe[1]);
		fcntl(go_pipe[0], F_SETFD, FD_CLOEXEC);

		/*
		 * Tell the parent we're ready to go
		 */
		close(child_ready_pipe[1]);

		/*
		 * Wait until the parent tells us to go.
		 */
1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325
		ret = read(go_pipe[0], &bf, 1);
		/*
		 * The parent will ask for the execvp() to be performed by
		 * writing exactly one byte, in workload.cork_fd, usually via
		 * perf_evlist__start_workload().
		 *
		 * For cancelling the workload without actuallin running it,
		 * the parent will just close workload.cork_fd, without writing
		 * anything, i.e. read will return zero and we just exit()
		 * here.
		 */
		if (ret != 1) {
			if (ret == -1)
				perror("unable to read pipe");
			exit(ret);
		}
1326 1327 1328

		execvp(argv[0], (char **)argv);

1329
		if (exec_error) {
1330 1331 1332 1333 1334 1335 1336
			union sigval val;

			val.sival_int = errno;
			if (sigqueue(getppid(), SIGUSR1, val))
				perror(argv[0]);
		} else
			perror(argv[0]);
1337 1338 1339
		exit(-1);
	}

1340 1341 1342 1343 1344 1345 1346 1347
	if (exec_error) {
		struct sigaction act = {
			.sa_flags     = SA_SIGINFO,
			.sa_sigaction = exec_error,
		};
		sigaction(SIGUSR1, &act, NULL);
	}

1348 1349 1350 1351 1352 1353
	if (target__none(target)) {
		if (evlist->threads == NULL) {
			fprintf(stderr, "FATAL: evlist->threads need to be set at this point (%s:%d).\n",
				__func__, __LINE__);
			goto out_close_pipes;
		}
1354
		evlist->threads->map[0] = evlist->workload.pid;
1355
	}
1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366

	close(child_ready_pipe[1]);
	close(go_pipe[0]);
	/*
	 * wait for child to settle
	 */
	if (read(child_ready_pipe[0], &bf, 1) == -1) {
		perror("unable to read pipe");
		goto out_close_pipes;
	}

1367
	fcntl(go_pipe[1], F_SETFD, FD_CLOEXEC);
1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383
	evlist->workload.cork_fd = go_pipe[1];
	close(child_ready_pipe[0]);
	return 0;

out_close_pipes:
	close(go_pipe[0]);
	close(go_pipe[1]);
out_close_ready_pipe:
	close(child_ready_pipe[0]);
	close(child_ready_pipe[1]);
	return -1;
}

int perf_evlist__start_workload(struct perf_evlist *evlist)
{
	if (evlist->workload.cork_fd > 0) {
1384
		char bf = 0;
1385
		int ret;
1386 1387 1388
		/*
		 * Remove the cork, let it rip!
		 */
1389 1390 1391 1392 1393 1394
		ret = write(evlist->workload.cork_fd, &bf, 1);
		if (ret < 0)
			perror("enable to write to pipe");

		close(evlist->workload.cork_fd);
		return ret;
1395 1396 1397 1398
	}

	return 0;
}
1399

1400
int perf_evlist__parse_sample(struct perf_evlist *evlist, union perf_event *event,
1401
			      struct perf_sample *sample)
1402
{
1403 1404 1405 1406
	struct perf_evsel *evsel = perf_evlist__event2evsel(evlist, event);

	if (!evsel)
		return -EFAULT;
1407
	return perf_evsel__parse_sample(evsel, event, sample);
1408
}
1409 1410 1411 1412 1413 1414

size_t perf_evlist__fprintf(struct perf_evlist *evlist, FILE *fp)
{
	struct perf_evsel *evsel;
	size_t printed = 0;

1415
	evlist__for_each(evlist, evsel) {
1416 1417 1418 1419
		printed += fprintf(fp, "%s%s", evsel->idx ? ", " : "",
				   perf_evsel__name(evsel));
	}

1420
	return printed + fprintf(fp, "\n");
1421
}
1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448

int perf_evlist__strerror_tp(struct perf_evlist *evlist __maybe_unused,
			     int err, char *buf, size_t size)
{
	char sbuf[128];

	switch (err) {
	case ENOENT:
		scnprintf(buf, size, "%s",
			  "Error:\tUnable to find debugfs\n"
			  "Hint:\tWas your kernel was compiled with debugfs support?\n"
			  "Hint:\tIs the debugfs filesystem mounted?\n"
			  "Hint:\tTry 'sudo mount -t debugfs nodev /sys/kernel/debug'");
		break;
	case EACCES:
		scnprintf(buf, size,
			  "Error:\tNo permissions to read %s/tracing/events/raw_syscalls\n"
			  "Hint:\tTry 'sudo mount -o remount,mode=755 %s'\n",
			  debugfs_mountpoint, debugfs_mountpoint);
		break;
	default:
		scnprintf(buf, size, "%s", strerror_r(err, sbuf, sizeof(sbuf)));
		break;
	}

	return 0;
}
1449 1450 1451 1452 1453

int perf_evlist__strerror_open(struct perf_evlist *evlist __maybe_unused,
			       int err, char *buf, size_t size)
{
	int printed, value;
1454
	char sbuf[STRERR_BUFSIZE], *emsg = strerror_r(err, sbuf, sizeof(sbuf));
1455 1456 1457 1458 1459 1460 1461 1462

	switch (err) {
	case EACCES:
	case EPERM:
		printed = scnprintf(buf, size,
				    "Error:\t%s.\n"
				    "Hint:\tCheck /proc/sys/kernel/perf_event_paranoid setting.", emsg);

1463
		value = perf_event_paranoid();
1464 1465 1466 1467 1468 1469 1470 1471

		printed += scnprintf(buf + printed, size - printed, "\nHint:\t");

		if (value >= 2) {
			printed += scnprintf(buf + printed, size - printed,
					     "For your workloads it needs to be <= 1\nHint:\t");
		}
		printed += scnprintf(buf + printed, size - printed,
1472
				     "For system wide tracing it needs to be set to -1.\n");
1473 1474

		printed += scnprintf(buf + printed, size - printed,
1475 1476
				    "Hint:\tTry: 'sudo sh -c \"echo -1 > /proc/sys/kernel/perf_event_paranoid\"'\n"
				    "Hint:\tThe current value is %d.", value);
1477 1478 1479 1480 1481 1482 1483 1484
		break;
	default:
		scnprintf(buf, size, "%s", emsg);
		break;
	}

	return 0;
}
1485 1486 1487 1488 1489 1490 1491 1492 1493 1494

void perf_evlist__to_front(struct perf_evlist *evlist,
			   struct perf_evsel *move_evsel)
{
	struct perf_evsel *evsel, *n;
	LIST_HEAD(move);

	if (move_evsel == perf_evlist__first(evlist))
		return;

1495
	evlist__for_each_safe(evlist, n, evsel) {
1496 1497 1498 1499 1500 1501
		if (evsel->leader == move_evsel->leader)
			list_move_tail(&evsel->node, &move);
	}

	list_splice(&move, &evlist->entries);
}
1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517

void perf_evlist__set_tracking_event(struct perf_evlist *evlist,
				     struct perf_evsel *tracking_evsel)
{
	struct perf_evsel *evsel;

	if (tracking_evsel->tracking)
		return;

	evlist__for_each(evlist, evsel) {
		if (evsel != tracking_evsel)
			evsel->tracking = false;
	}

	tracking_evsel->tracking = true;
}