header.c 59.2 KB
Newer Older
1
#include "util.h"
2
#include <sys/types.h>
3
#include <byteswap.h>
4 5 6
#include <unistd.h>
#include <stdio.h>
#include <stdlib.h>
7
#include <linux/list.h>
8
#include <linux/kernel.h>
9
#include <linux/bitops.h>
10
#include <sys/utsname.h>
11

12
#include "evlist.h"
13
#include "evsel.h"
14
#include "header.h"
15 16
#include "../perf.h"
#include "trace-event.h"
17
#include "session.h"
18
#include "symbol.h"
19
#include "debug.h"
20
#include "cpumap.h"
21
#include "pmu.h"
22
#include "vdso.h"
23
#include "strbuf.h"
24
#include "build-id.h"
25
#include "data.h"
26

27 28 29
static u32 header_argc;
static const char **header_argv;

30 31 32 33 34 35 36 37 38 39 40 41
/*
 * magic2 = "PERFILE2"
 * must be a numerical value to let the endianness
 * determine the memory layout. That way we are able
 * to detect endianness when reading the perf.data file
 * back.
 *
 * we check for legacy (PERFFILE) format.
 */
static const char *__perf_magic1 = "PERFFILE";
static const u64 __perf_magic2    = 0x32454c4946524550ULL;
static const u64 __perf_magic2_sw = 0x50455246494c4532ULL;
42

43
#define PERF_MAGIC	__perf_magic2
44 45

struct perf_file_attr {
46
	struct perf_event_attr	attr;
47 48 49
	struct perf_file_section	ids;
};

50
void perf_header__set_feat(struct perf_header *header, int feat)
51
{
52
	set_bit(feat, header->adds_features);
53 54
}

55
void perf_header__clear_feat(struct perf_header *header, int feat)
56
{
57
	clear_bit(feat, header->adds_features);
58 59
}

60
bool perf_header__has_feat(const struct perf_header *header, int feat)
61
{
62
	return test_bit(feat, header->adds_features);
63 64
}

65
static int do_write(int fd, const void *buf, size_t size)
66 67 68 69 70
{
	while (size) {
		int ret = write(fd, buf, size);

		if (ret < 0)
71
			return -errno;
72 73 74 75

		size -= ret;
		buf += ret;
	}
76 77

	return 0;
78 79
}

80
int write_padded(int fd, const void *bf, size_t count, size_t count_aligned)
81 82 83 84 85 86 87 88 89 90
{
	static const char zero_buf[NAME_ALIGN];
	int err = do_write(fd, bf, count);

	if (!err)
		err = do_write(fd, zero_buf, count_aligned - count);

	return err;
}

91 92 93 94 95 96
static int do_write_string(int fd, const char *str)
{
	u32 len, olen;
	int ret;

	olen = strlen(str) + 1;
97
	len = PERF_ALIGN(olen, NAME_ALIGN);
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112

	/* write len, incl. \0 */
	ret = do_write(fd, &len, sizeof(len));
	if (ret < 0)
		return ret;

	return write_padded(fd, str, olen, len);
}

static char *do_read_string(int fd, struct perf_header *ph)
{
	ssize_t sz, ret;
	u32 len;
	char *buf;

113
	sz = readn(fd, &len, sizeof(len));
114 115 116 117 118 119 120 121 122 123
	if (sz < (ssize_t)sizeof(len))
		return NULL;

	if (ph->needs_swap)
		len = bswap_32(len);

	buf = malloc(len);
	if (!buf)
		return NULL;

124
	ret = readn(fd, buf, len);
125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142
	if (ret == (ssize_t)len) {
		/*
		 * strings are padded by zeroes
		 * thus the actual strlen of buf
		 * may be less than len
		 */
		return buf;
	}

	free(buf);
	return NULL;
}

int
perf_header__set_cmdline(int argc, const char **argv)
{
	int i;

143 144 145 146 147 148 149 150 151
	/*
	 * If header_argv has already been set, do not override it.
	 * This allows a command to set the cmdline, parse args and
	 * then call another builtin function that implements a
	 * command -- e.g, cmd_kvm calling cmd_record.
	 */
	if (header_argv)
		return 0;

152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168
	header_argc = (u32)argc;

	/* do not include NULL termination */
	header_argv = calloc(argc, sizeof(char *));
	if (!header_argv)
		return -ENOMEM;

	/*
	 * must copy argv contents because it gets moved
	 * around during option parsing
	 */
	for (i = 0; i < argc ; i++)
		header_argv[i] = argv[i];

	return 0;
}

169
static int write_tracing_data(int fd, struct perf_header *h __maybe_unused,
170 171 172 173 174 175 176
			    struct perf_evlist *evlist)
{
	return read_tracing_data(fd, &evlist->entries);
}


static int write_build_id(int fd, struct perf_header *h,
177
			  struct perf_evlist *evlist __maybe_unused)
178 179 180 181 182 183
{
	struct perf_session *session;
	int err;

	session = container_of(h, struct perf_session, header);

184 185 186
	if (!perf_session__read_build_ids(session, true))
		return -1;

187
	err = perf_session__write_buildid_table(session, fd);
188 189 190 191
	if (err < 0) {
		pr_debug("failed to write buildid table\n");
		return err;
	}
192
	perf_session__cache_build_ids(session);
193 194 195 196

	return 0;
}

197 198
static int write_hostname(int fd, struct perf_header *h __maybe_unused,
			  struct perf_evlist *evlist __maybe_unused)
199 200 201 202 203 204 205 206 207 208 209
{
	struct utsname uts;
	int ret;

	ret = uname(&uts);
	if (ret < 0)
		return -1;

	return do_write_string(fd, uts.nodename);
}

210 211
static int write_osrelease(int fd, struct perf_header *h __maybe_unused,
			   struct perf_evlist *evlist __maybe_unused)
212 213 214 215 216 217 218 219 220 221 222
{
	struct utsname uts;
	int ret;

	ret = uname(&uts);
	if (ret < 0)
		return -1;

	return do_write_string(fd, uts.release);
}

223 224
static int write_arch(int fd, struct perf_header *h __maybe_unused,
		      struct perf_evlist *evlist __maybe_unused)
225 226 227 228 229 230 231 232 233 234 235
{
	struct utsname uts;
	int ret;

	ret = uname(&uts);
	if (ret < 0)
		return -1;

	return do_write_string(fd, uts.machine);
}

236 237
static int write_version(int fd, struct perf_header *h __maybe_unused,
			 struct perf_evlist *evlist __maybe_unused)
238 239 240 241
{
	return do_write_string(fd, perf_version_string);
}

242
static int __write_cpudesc(int fd, const char *cpuinfo_proc)
243 244 245 246
{
	FILE *file;
	char *buf = NULL;
	char *s, *p;
247
	const char *search = cpuinfo_proc;
248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263
	size_t len = 0;
	int ret = -1;

	if (!search)
		return -1;

	file = fopen("/proc/cpuinfo", "r");
	if (!file)
		return -1;

	while (getline(&buf, &len, file) > 0) {
		ret = strncmp(buf, search, strlen(search));
		if (!ret)
			break;
	}

264 265
	if (ret) {
		ret = -1;
266
		goto done;
267
	}
268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298

	s = buf;

	p = strchr(buf, ':');
	if (p && *(p+1) == ' ' && *(p+2))
		s = p + 2;
	p = strchr(s, '\n');
	if (p)
		*p = '\0';

	/* squash extra space characters (branding string) */
	p = s;
	while (*p) {
		if (isspace(*p)) {
			char *r = p + 1;
			char *q = r;
			*p = ' ';
			while (*q && isspace(*q))
				q++;
			if (q != (p+1))
				while ((*r++ = *q++));
		}
		p++;
	}
	ret = do_write_string(fd, s);
done:
	free(buf);
	fclose(file);
	return ret;
}

299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317
static int write_cpudesc(int fd, struct perf_header *h __maybe_unused,
		       struct perf_evlist *evlist __maybe_unused)
{
#ifndef CPUINFO_PROC
#define CPUINFO_PROC {"model name", }
#endif
	const char *cpuinfo_procs[] = CPUINFO_PROC;
	unsigned int i;

	for (i = 0; i < ARRAY_SIZE(cpuinfo_procs); i++) {
		int ret;
		ret = __write_cpudesc(fd, cpuinfo_procs[i]);
		if (ret >= 0)
			return ret;
	}
	return -1;
}


318 319
static int write_nrcpus(int fd, struct perf_header *h __maybe_unused,
			struct perf_evlist *evlist __maybe_unused)
320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343
{
	long nr;
	u32 nrc, nra;
	int ret;

	nr = sysconf(_SC_NPROCESSORS_CONF);
	if (nr < 0)
		return -1;

	nrc = (u32)(nr & UINT_MAX);

	nr = sysconf(_SC_NPROCESSORS_ONLN);
	if (nr < 0)
		return -1;

	nra = (u32)(nr & UINT_MAX);

	ret = do_write(fd, &nrc, sizeof(nrc));
	if (ret < 0)
		return ret;

	return do_write(fd, &nra, sizeof(nra));
}

344
static int write_event_desc(int fd, struct perf_header *h __maybe_unused,
345 346
			    struct perf_evlist *evlist)
{
347
	struct perf_evsel *evsel;
348
	u32 nre, nri, sz;
349 350
	int ret;

351
	nre = evlist->nr_entries;
352 353 354 355 356 357 358 359 360 361 362

	/*
	 * write number of events
	 */
	ret = do_write(fd, &nre, sizeof(nre));
	if (ret < 0)
		return ret;

	/*
	 * size of perf_event_attr struct
	 */
363
	sz = (u32)sizeof(evsel->attr);
364 365 366 367
	ret = do_write(fd, &sz, sizeof(sz));
	if (ret < 0)
		return ret;

368
	evlist__for_each(evlist, evsel) {
369
		ret = do_write(fd, &evsel->attr, sz);
370 371 372 373 374 375 376 377 378
		if (ret < 0)
			return ret;
		/*
		 * write number of unique id per event
		 * there is one id per instance of an event
		 *
		 * copy into an nri to be independent of the
		 * type of ids,
		 */
379
		nri = evsel->ids;
380 381 382 383 384 385 386
		ret = do_write(fd, &nri, sizeof(nri));
		if (ret < 0)
			return ret;

		/*
		 * write event string as passed on cmdline
		 */
387
		ret = do_write_string(fd, perf_evsel__name(evsel));
388 389 390 391 392
		if (ret < 0)
			return ret;
		/*
		 * write unique ids for this event
		 */
393
		ret = do_write(fd, evsel->id, evsel->ids * sizeof(u64));
394 395 396 397 398 399
		if (ret < 0)
			return ret;
	}
	return 0;
}

400 401
static int write_cmdline(int fd, struct perf_header *h __maybe_unused,
			 struct perf_evlist *evlist __maybe_unused)
402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455
{
	char buf[MAXPATHLEN];
	char proc[32];
	u32 i, n;
	int ret;

	/*
	 * actual atual path to perf binary
	 */
	sprintf(proc, "/proc/%d/exe", getpid());
	ret = readlink(proc, buf, sizeof(buf));
	if (ret <= 0)
		return -1;

	/* readlink() does not add null termination */
	buf[ret] = '\0';

	/* account for binary path */
	n = header_argc + 1;

	ret = do_write(fd, &n, sizeof(n));
	if (ret < 0)
		return ret;

	ret = do_write_string(fd, buf);
	if (ret < 0)
		return ret;

	for (i = 0 ; i < header_argc; i++) {
		ret = do_write_string(fd, header_argv[i]);
		if (ret < 0)
			return ret;
	}
	return 0;
}

#define CORE_SIB_FMT \
	"/sys/devices/system/cpu/cpu%d/topology/core_siblings_list"
#define THRD_SIB_FMT \
	"/sys/devices/system/cpu/cpu%d/topology/thread_siblings_list"

struct cpu_topo {
	u32 core_sib;
	u32 thread_sib;
	char **core_siblings;
	char **thread_siblings;
};

static int build_cpu_topo(struct cpu_topo *tp, int cpu)
{
	FILE *fp;
	char filename[MAXPATHLEN];
	char *buf = NULL, *p;
	size_t len = 0;
456
	ssize_t sret;
457 458 459 460 461 462
	u32 i = 0;
	int ret = -1;

	sprintf(filename, CORE_SIB_FMT, cpu);
	fp = fopen(filename, "r");
	if (!fp)
463
		goto try_threads;
464

465
	sret = getline(&buf, &len, fp);
466
	fclose(fp);
467 468
	if (sret <= 0)
		goto try_threads;
469 470 471 472 473 474 475 476 477 478 479 480 481 482 483

	p = strchr(buf, '\n');
	if (p)
		*p = '\0';

	for (i = 0; i < tp->core_sib; i++) {
		if (!strcmp(buf, tp->core_siblings[i]))
			break;
	}
	if (i == tp->core_sib) {
		tp->core_siblings[i] = buf;
		tp->core_sib++;
		buf = NULL;
		len = 0;
	}
484
	ret = 0;
485

486
try_threads:
487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523
	sprintf(filename, THRD_SIB_FMT, cpu);
	fp = fopen(filename, "r");
	if (!fp)
		goto done;

	if (getline(&buf, &len, fp) <= 0)
		goto done;

	p = strchr(buf, '\n');
	if (p)
		*p = '\0';

	for (i = 0; i < tp->thread_sib; i++) {
		if (!strcmp(buf, tp->thread_siblings[i]))
			break;
	}
	if (i == tp->thread_sib) {
		tp->thread_siblings[i] = buf;
		tp->thread_sib++;
		buf = NULL;
	}
	ret = 0;
done:
	if(fp)
		fclose(fp);
	free(buf);
	return ret;
}

static void free_cpu_topo(struct cpu_topo *tp)
{
	u32 i;

	if (!tp)
		return;

	for (i = 0 ; i < tp->core_sib; i++)
524
		zfree(&tp->core_siblings[i]);
525 526

	for (i = 0 ; i < tp->thread_sib; i++)
527
		zfree(&tp->thread_siblings[i]);
528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571

	free(tp);
}

static struct cpu_topo *build_cpu_topology(void)
{
	struct cpu_topo *tp;
	void *addr;
	u32 nr, i;
	size_t sz;
	long ncpus;
	int ret = -1;

	ncpus = sysconf(_SC_NPROCESSORS_CONF);
	if (ncpus < 0)
		return NULL;

	nr = (u32)(ncpus & UINT_MAX);

	sz = nr * sizeof(char *);

	addr = calloc(1, sizeof(*tp) + 2 * sz);
	if (!addr)
		return NULL;

	tp = addr;

	addr += sizeof(*tp);
	tp->core_siblings = addr;
	addr += sz;
	tp->thread_siblings = addr;

	for (i = 0; i < nr; i++) {
		ret = build_cpu_topo(tp, i);
		if (ret < 0)
			break;
	}
	if (ret) {
		free_cpu_topo(tp);
		tp = NULL;
	}
	return tp;
}

572 573
static int write_cpu_topology(int fd, struct perf_header *h __maybe_unused,
			  struct perf_evlist *evlist __maybe_unused)
574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607
{
	struct cpu_topo *tp;
	u32 i;
	int ret;

	tp = build_cpu_topology();
	if (!tp)
		return -1;

	ret = do_write(fd, &tp->core_sib, sizeof(tp->core_sib));
	if (ret < 0)
		goto done;

	for (i = 0; i < tp->core_sib; i++) {
		ret = do_write_string(fd, tp->core_siblings[i]);
		if (ret < 0)
			goto done;
	}
	ret = do_write(fd, &tp->thread_sib, sizeof(tp->thread_sib));
	if (ret < 0)
		goto done;

	for (i = 0; i < tp->thread_sib; i++) {
		ret = do_write_string(fd, tp->thread_siblings[i]);
		if (ret < 0)
			break;
	}
done:
	free_cpu_topo(tp);
	return ret;
}



608 609
static int write_total_mem(int fd, struct perf_header *h __maybe_unused,
			  struct perf_evlist *evlist __maybe_unused)
610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629
{
	char *buf = NULL;
	FILE *fp;
	size_t len = 0;
	int ret = -1, n;
	uint64_t mem;

	fp = fopen("/proc/meminfo", "r");
	if (!fp)
		return -1;

	while (getline(&buf, &len, fp) > 0) {
		ret = strncmp(buf, "MemTotal:", 9);
		if (!ret)
			break;
	}
	if (!ret) {
		n = sscanf(buf, "%*s %"PRIu64, &mem);
		if (n == 1)
			ret = do_write(fd, &mem, sizeof(mem));
630 631
	} else
		ret = -1;
632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655
	free(buf);
	fclose(fp);
	return ret;
}

static int write_topo_node(int fd, int node)
{
	char str[MAXPATHLEN];
	char field[32];
	char *buf = NULL, *p;
	size_t len = 0;
	FILE *fp;
	u64 mem_total, mem_free, mem;
	int ret = -1;

	sprintf(str, "/sys/devices/system/node/node%d/meminfo", node);
	fp = fopen(str, "r");
	if (!fp)
		return -1;

	while (getline(&buf, &len, fp) > 0) {
		/* skip over invalid lines */
		if (!strchr(buf, ':'))
			continue;
656
		if (sscanf(buf, "%*s %*d %31s %"PRIu64, field, &mem) != 2)
657 658 659 660 661 662 663 664
			goto done;
		if (!strcmp(field, "MemTotal:"))
			mem_total = mem;
		if (!strcmp(field, "MemFree:"))
			mem_free = mem;
	}

	fclose(fp);
665
	fp = NULL;
666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691

	ret = do_write(fd, &mem_total, sizeof(u64));
	if (ret)
		goto done;

	ret = do_write(fd, &mem_free, sizeof(u64));
	if (ret)
		goto done;

	ret = -1;
	sprintf(str, "/sys/devices/system/node/node%d/cpulist", node);

	fp = fopen(str, "r");
	if (!fp)
		goto done;

	if (getline(&buf, &len, fp) <= 0)
		goto done;

	p = strchr(buf, '\n');
	if (p)
		*p = '\0';

	ret = do_write_string(fd, buf);
done:
	free(buf);
692 693
	if (fp)
		fclose(fp);
694 695 696
	return ret;
}

697 698
static int write_numa_topology(int fd, struct perf_header *h __maybe_unused,
			  struct perf_evlist *evlist __maybe_unused)
699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745
{
	char *buf = NULL;
	size_t len = 0;
	FILE *fp;
	struct cpu_map *node_map = NULL;
	char *c;
	u32 nr, i, j;
	int ret = -1;

	fp = fopen("/sys/devices/system/node/online", "r");
	if (!fp)
		return -1;

	if (getline(&buf, &len, fp) <= 0)
		goto done;

	c = strchr(buf, '\n');
	if (c)
		*c = '\0';

	node_map = cpu_map__new(buf);
	if (!node_map)
		goto done;

	nr = (u32)node_map->nr;

	ret = do_write(fd, &nr, sizeof(nr));
	if (ret < 0)
		goto done;

	for (i = 0; i < nr; i++) {
		j = (u32)node_map->map[i];
		ret = do_write(fd, &j, sizeof(j));
		if (ret < 0)
			break;

		ret = write_topo_node(fd, i);
		if (ret < 0)
			break;
	}
done:
	free(buf);
	fclose(fp);
	free(node_map);
	return ret;
}

746 747 748 749 750 751 752 753 754 755 756 757
/*
 * File format:
 *
 * struct pmu_mappings {
 *	u32	pmu_num;
 *	struct pmu_map {
 *		u32	type;
 *		char	name[];
 *	}[pmu_num];
 * };
 */

758 759
static int write_pmu_mappings(int fd, struct perf_header *h __maybe_unused,
			      struct perf_evlist *evlist __maybe_unused)
760 761 762 763
{
	struct perf_pmu *pmu = NULL;
	off_t offset = lseek(fd, 0, SEEK_CUR);
	__u32 pmu_num = 0;
764
	int ret;
765 766

	/* write real pmu_num later */
767 768 769
	ret = do_write(fd, &pmu_num, sizeof(pmu_num));
	if (ret < 0)
		return ret;
770 771 772 773 774

	while ((pmu = perf_pmu__scan(pmu))) {
		if (!pmu->name)
			continue;
		pmu_num++;
775 776 777 778 779 780 781 782

		ret = do_write(fd, &pmu->type, sizeof(pmu->type));
		if (ret < 0)
			return ret;

		ret = do_write_string(fd, pmu->name);
		if (ret < 0)
			return ret;
783 784 785 786 787 788 789 790 791 792 793
	}

	if (pwrite(fd, &pmu_num, sizeof(pmu_num), offset) != sizeof(pmu_num)) {
		/* discard all */
		lseek(fd, offset, SEEK_SET);
		return -1;
	}

	return 0;
}

794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816
/*
 * File format:
 *
 * struct group_descs {
 *	u32	nr_groups;
 *	struct group_desc {
 *		char	name[];
 *		u32	leader_idx;
 *		u32	nr_members;
 *	}[nr_groups];
 * };
 */
static int write_group_desc(int fd, struct perf_header *h __maybe_unused,
			    struct perf_evlist *evlist)
{
	u32 nr_groups = evlist->nr_groups;
	struct perf_evsel *evsel;
	int ret;

	ret = do_write(fd, &nr_groups, sizeof(nr_groups));
	if (ret < 0)
		return ret;

817
	evlist__for_each(evlist, evsel) {
818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839
		if (perf_evsel__is_group_leader(evsel) &&
		    evsel->nr_members > 1) {
			const char *name = evsel->group_name ?: "{anon_group}";
			u32 leader_idx = evsel->idx;
			u32 nr_members = evsel->nr_members;

			ret = do_write_string(fd, name);
			if (ret < 0)
				return ret;

			ret = do_write(fd, &leader_idx, sizeof(leader_idx));
			if (ret < 0)
				return ret;

			ret = do_write(fd, &nr_members, sizeof(nr_members));
			if (ret < 0)
				return ret;
		}
	}
	return 0;
}

840 841 842 843
/*
 * default get_cpuid(): nothing gets recorded
 * actual implementation must be in arch/$(ARCH)/util/header.c
 */
844 845
int __attribute__ ((weak)) get_cpuid(char *buffer __maybe_unused,
				     size_t sz __maybe_unused)
846 847 848 849
{
	return -1;
}

850 851
static int write_cpuid(int fd, struct perf_header *h __maybe_unused,
		       struct perf_evlist *evlist __maybe_unused)
852 853 854 855 856 857 858 859 860 861 862 863 864
{
	char buffer[64];
	int ret;

	ret = get_cpuid(buffer, sizeof(buffer));
	if (!ret)
		goto write_it;

	return -1;
write_it:
	return do_write_string(fd, buffer);
}

865 866 867
static int write_branch_stack(int fd __maybe_unused,
			      struct perf_header *h __maybe_unused,
		       struct perf_evlist *evlist __maybe_unused)
868 869 870 871
{
	return 0;
}

872 873
static void print_hostname(struct perf_header *ph, int fd __maybe_unused,
			   FILE *fp)
874
{
875
	fprintf(fp, "# hostname : %s\n", ph->env.hostname);
876 877
}

878 879
static void print_osrelease(struct perf_header *ph, int fd __maybe_unused,
			    FILE *fp)
880
{
881
	fprintf(fp, "# os release : %s\n", ph->env.os_release);
882 883
}

884
static void print_arch(struct perf_header *ph, int fd __maybe_unused, FILE *fp)
885
{
886
	fprintf(fp, "# arch : %s\n", ph->env.arch);
887 888
}

889 890
static void print_cpudesc(struct perf_header *ph, int fd __maybe_unused,
			  FILE *fp)
891
{
892
	fprintf(fp, "# cpudesc : %s\n", ph->env.cpu_desc);
893 894
}

895 896
static void print_nrcpus(struct perf_header *ph, int fd __maybe_unused,
			 FILE *fp)
897
{
898 899
	fprintf(fp, "# nrcpus online : %u\n", ph->env.nr_cpus_online);
	fprintf(fp, "# nrcpus avail : %u\n", ph->env.nr_cpus_avail);
900 901
}

902 903
static void print_version(struct perf_header *ph, int fd __maybe_unused,
			  FILE *fp)
904
{
905
	fprintf(fp, "# perf version : %s\n", ph->env.version);
906 907
}

908 909
static void print_cmdline(struct perf_header *ph, int fd __maybe_unused,
			  FILE *fp)
910
{
911
	int nr, i;
912 913
	char *str;

914 915
	nr = ph->env.nr_cmdline;
	str = ph->env.cmdline;
916 917 918 919 920

	fprintf(fp, "# cmdline : ");

	for (i = 0; i < nr; i++) {
		fprintf(fp, "%s ", str);
921
		str += strlen(str) + 1;
922 923 924 925
	}
	fputc('\n', fp);
}

926 927
static void print_cpu_topology(struct perf_header *ph, int fd __maybe_unused,
			       FILE *fp)
928
{
929
	int nr, i;
930 931
	char *str;

932 933
	nr = ph->env.nr_sibling_cores;
	str = ph->env.sibling_cores;
934 935 936

	for (i = 0; i < nr; i++) {
		fprintf(fp, "# sibling cores   : %s\n", str);
937
		str += strlen(str) + 1;
938 939
	}

940 941
	nr = ph->env.nr_sibling_threads;
	str = ph->env.sibling_threads;
942 943 944

	for (i = 0; i < nr; i++) {
		fprintf(fp, "# sibling threads : %s\n", str);
945
		str += strlen(str) + 1;
946 947 948
	}
}

949
static void free_event_desc(struct perf_evsel *events)
950
{
951 952 953 954 955 956
	struct perf_evsel *evsel;

	if (!events)
		return;

	for (evsel = events; evsel->attr.size; evsel++) {
957 958
		zfree(&evsel->name);
		zfree(&evsel->id);
959 960 961 962 963 964 965 966 967 968
	}

	free(events);
}

static struct perf_evsel *
read_event_desc(struct perf_header *ph, int fd)
{
	struct perf_evsel *evsel, *events = NULL;
	u64 *id;
969
	void *buf = NULL;
970 971 972
	u32 nre, sz, nr, i, j;
	ssize_t ret;
	size_t msz;
973 974

	/* number of events */
975
	ret = readn(fd, &nre, sizeof(nre));
976 977 978 979 980 981
	if (ret != (ssize_t)sizeof(nre))
		goto error;

	if (ph->needs_swap)
		nre = bswap_32(nre);

982
	ret = readn(fd, &sz, sizeof(sz));
983 984 985 986 987 988
	if (ret != (ssize_t)sizeof(sz))
		goto error;

	if (ph->needs_swap)
		sz = bswap_32(sz);

989
	/* buffer to hold on file attr struct */
990 991 992 993
	buf = malloc(sz);
	if (!buf)
		goto error;

994 995 996 997 998 999
	/* the last event terminates with evsel->attr.size == 0: */
	events = calloc(nre + 1, sizeof(*events));
	if (!events)
		goto error;

	msz = sizeof(evsel->attr);
1000
	if (sz < msz)
1001 1002
		msz = sz;

1003 1004
	for (i = 0, evsel = events; i < nre; evsel++, i++) {
		evsel->idx = i;
1005

1006 1007 1008 1009
		/*
		 * must read entire on-file attr struct to
		 * sync up with layout.
		 */
1010
		ret = readn(fd, buf, sz);
1011 1012 1013 1014 1015 1016
		if (ret != (ssize_t)sz)
			goto error;

		if (ph->needs_swap)
			perf_event__attr_swap(buf);

1017
		memcpy(&evsel->attr, buf, msz);
1018

1019
		ret = readn(fd, &nr, sizeof(nr));
1020 1021 1022
		if (ret != (ssize_t)sizeof(nr))
			goto error;

1023
		if (ph->needs_swap) {
1024
			nr = bswap_32(nr);
1025 1026
			evsel->needs_swap = true;
		}
1027

1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039
		evsel->name = do_read_string(fd, ph);

		if (!nr)
			continue;

		id = calloc(nr, sizeof(*id));
		if (!id)
			goto error;
		evsel->ids = nr;
		evsel->id = id;

		for (j = 0 ; j < nr; j++) {
1040
			ret = readn(fd, id, sizeof(*id));
1041 1042 1043 1044 1045 1046 1047 1048
			if (ret != (ssize_t)sizeof(*id))
				goto error;
			if (ph->needs_swap)
				*id = bswap_64(*id);
			id++;
		}
	}
out:
1049
	free(buf);
1050 1051 1052 1053 1054 1055 1056 1057
	return events;
error:
	if (events)
		free_event_desc(events);
	events = NULL;
	goto out;
}

1058 1059 1060 1061 1062 1063
static int __desc_attr__fprintf(FILE *fp, const char *name, const char *val,
				void *priv __attribute__((unused)))
{
	return fprintf(fp, ", %s = %s", name, val);
}

1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076
static void print_event_desc(struct perf_header *ph, int fd, FILE *fp)
{
	struct perf_evsel *evsel, *events = read_event_desc(ph, fd);
	u32 j;
	u64 *id;

	if (!events) {
		fprintf(fp, "# event desc: not available or unable to read\n");
		return;
	}

	for (evsel = events; evsel->attr.size; evsel++) {
		fprintf(fp, "# event : name = %s, ", evsel->name);
1077

1078
		if (evsel->ids) {
1079
			fprintf(fp, ", id = {");
1080 1081 1082 1083 1084
			for (j = 0, id = evsel->id; j < evsel->ids; j++, id++) {
				if (j)
					fputc(',', fp);
				fprintf(fp, " %"PRIu64, *id);
			}
1085
			fprintf(fp, " }");
1086
		}
1087

1088
		perf_event_attr__fprintf(fp, &evsel->attr, __desc_attr__fprintf, NULL);
1089

1090 1091
		fputc('\n', fp);
	}
1092 1093

	free_event_desc(events);
1094 1095
}

1096
static void print_total_mem(struct perf_header *ph, int fd __maybe_unused,
1097
			    FILE *fp)
1098
{
1099
	fprintf(fp, "# total memory : %Lu kB\n", ph->env.total_mem);
1100 1101
}

1102
static void print_numa_topology(struct perf_header *ph, int fd __maybe_unused,
1103
				FILE *fp)
1104 1105
{
	u32 nr, c, i;
1106
	char *str, *tmp;
1107 1108 1109
	uint64_t mem_total, mem_free;

	/* nr nodes */
1110 1111
	nr = ph->env.nr_numa_nodes;
	str = ph->env.numa_nodes;
1112 1113 1114

	for (i = 0; i < nr; i++) {
		/* node number */
1115 1116
		c = strtoul(str, &tmp, 0);
		if (*tmp != ':')
1117 1118
			goto error;

1119 1120 1121
		str = tmp + 1;
		mem_total = strtoull(str, &tmp, 0);
		if (*tmp != ':')
1122 1123
			goto error;

1124 1125 1126
		str = tmp + 1;
		mem_free = strtoull(str, &tmp, 0);
		if (*tmp != ':')
1127 1128 1129 1130
			goto error;

		fprintf(fp, "# node%u meminfo  : total = %"PRIu64" kB,"
			    " free = %"PRIu64" kB\n",
1131
			c, mem_total, mem_free);
1132

1133
		str = tmp + 1;
1134
		fprintf(fp, "# node%u cpu list : %s\n", c, str);
1135 1136

		str += strlen(str) + 1;
1137 1138 1139 1140 1141 1142
	}
	return;
error:
	fprintf(fp, "# numa topology : not available\n");
}

1143
static void print_cpuid(struct perf_header *ph, int fd __maybe_unused, FILE *fp)
1144
{
1145
	fprintf(fp, "# cpuid : %s\n", ph->env.cpuid);
1146 1147
}

1148
static void print_branch_stack(struct perf_header *ph __maybe_unused,
1149
			       int fd __maybe_unused, FILE *fp)
1150 1151 1152 1153
{
	fprintf(fp, "# contains samples with branch stack\n");
}

1154 1155
static void print_pmu_mappings(struct perf_header *ph, int fd __maybe_unused,
			       FILE *fp)
1156 1157
{
	const char *delimiter = "# pmu mappings: ";
1158
	char *str, *tmp;
1159 1160 1161
	u32 pmu_num;
	u32 type;

1162
	pmu_num = ph->env.nr_pmu_mappings;
1163 1164 1165 1166 1167
	if (!pmu_num) {
		fprintf(fp, "# pmu mappings: not available\n");
		return;
	}

1168 1169
	str = ph->env.pmu_mappings;

1170
	while (pmu_num) {
1171 1172 1173 1174 1175 1176
		type = strtoul(str, &tmp, 0);
		if (*tmp != ':')
			goto error;

		str = tmp + 1;
		fprintf(fp, "%s%s = %" PRIu32, delimiter, str, type);
1177

1178
		delimiter = ", ";
1179 1180
		str += strlen(str) + 1;
		pmu_num--;
1181 1182 1183 1184 1185 1186 1187 1188 1189 1190
	}

	fprintf(fp, "\n");

	if (!pmu_num)
		return;
error:
	fprintf(fp, "# pmu mappings: unable to read\n");
}

1191 1192 1193 1194 1195 1196 1197 1198 1199
static void print_group_desc(struct perf_header *ph, int fd __maybe_unused,
			     FILE *fp)
{
	struct perf_session *session;
	struct perf_evsel *evsel;
	u32 nr = 0;

	session = container_of(ph, struct perf_session, header);

1200
	evlist__for_each(session->evlist, evsel) {
1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215
		if (perf_evsel__is_group_leader(evsel) &&
		    evsel->nr_members > 1) {
			fprintf(fp, "# group: %s{%s", evsel->group_name ?: "",
				perf_evsel__name(evsel));

			nr = evsel->nr_members - 1;
		} else if (nr) {
			fprintf(fp, ",%s", perf_evsel__name(evsel));

			if (--nr == 0)
				fprintf(fp, "}\n");
		}
	}
}

1216 1217 1218 1219 1220
static int __event_process_build_id(struct build_id_event *bev,
				    char *filename,
				    struct perf_session *session)
{
	int err = -1;
1221
	struct dsos *dsos;
1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235
	struct machine *machine;
	u16 misc;
	struct dso *dso;
	enum dso_kernel_type dso_type;

	machine = perf_session__findnew_machine(session, bev->pid);
	if (!machine)
		goto out;

	misc = bev->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;

	switch (misc) {
	case PERF_RECORD_MISC_KERNEL:
		dso_type = DSO_TYPE_KERNEL;
1236
		dsos = &machine->kernel_dsos;
1237 1238 1239
		break;
	case PERF_RECORD_MISC_GUEST_KERNEL:
		dso_type = DSO_TYPE_GUEST_KERNEL;
1240
		dsos = &machine->kernel_dsos;
1241 1242 1243 1244
		break;
	case PERF_RECORD_MISC_USER:
	case PERF_RECORD_MISC_GUEST_USER:
		dso_type = DSO_TYPE_USER;
1245
		dsos = &machine->user_dsos;
1246 1247 1248 1249 1250
		break;
	default:
		goto out;
	}

1251
	dso = __dsos__findnew(dsos, filename);
1252 1253 1254 1255 1256
	if (dso != NULL) {
		char sbuild_id[BUILD_ID_SIZE * 2 + 1];

		dso__set_build_id(dso, &bev->build_id);

1257
		if (!is_kernel_module(filename))
1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276
			dso->kernel = dso_type;

		build_id__sprintf(dso->build_id, sizeof(dso->build_id),
				  sbuild_id);
		pr_debug("build id event received for %s: %s\n",
			 dso->long_name, sbuild_id);
	}

	err = 0;
out:
	return err;
}

static int perf_header__read_build_ids_abi_quirk(struct perf_header *header,
						 int input, u64 offset, u64 size)
{
	struct perf_session *session = container_of(header, struct perf_session, header);
	struct {
		struct perf_event_header   header;
1277
		u8			   build_id[PERF_ALIGN(BUILD_ID_SIZE, sizeof(u64))];
1278 1279 1280 1281 1282 1283 1284 1285 1286
		char			   filename[0];
	} old_bev;
	struct build_id_event bev;
	char filename[PATH_MAX];
	u64 limit = offset + size;

	while (offset < limit) {
		ssize_t len;

1287
		if (readn(input, &old_bev, sizeof(old_bev)) != sizeof(old_bev))
1288 1289 1290 1291 1292 1293
			return -1;

		if (header->needs_swap)
			perf_event_header__bswap(&old_bev.header);

		len = old_bev.header.size - sizeof(old_bev);
1294
		if (readn(input, filename, len) != len)
1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328
			return -1;

		bev.header = old_bev.header;

		/*
		 * As the pid is the missing value, we need to fill
		 * it properly. The header.misc value give us nice hint.
		 */
		bev.pid	= HOST_KERNEL_ID;
		if (bev.header.misc == PERF_RECORD_MISC_GUEST_USER ||
		    bev.header.misc == PERF_RECORD_MISC_GUEST_KERNEL)
			bev.pid	= DEFAULT_GUEST_KERNEL_ID;

		memcpy(bev.build_id, old_bev.build_id, sizeof(bev.build_id));
		__event_process_build_id(&bev, filename, session);

		offset += bev.header.size;
	}

	return 0;
}

static int perf_header__read_build_ids(struct perf_header *header,
				       int input, u64 offset, u64 size)
{
	struct perf_session *session = container_of(header, struct perf_session, header);
	struct build_id_event bev;
	char filename[PATH_MAX];
	u64 limit = offset + size, orig_offset = offset;
	int err = -1;

	while (offset < limit) {
		ssize_t len;

1329
		if (readn(input, &bev, sizeof(bev)) != sizeof(bev))
1330 1331 1332 1333 1334 1335
			goto out;

		if (header->needs_swap)
			perf_event_header__bswap(&bev.header);

		len = bev.header.size - sizeof(bev);
1336
		if (readn(input, filename, len) != len)
1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365
			goto out;
		/*
		 * The a1645ce1 changeset:
		 *
		 * "perf: 'perf kvm' tool for monitoring guest performance from host"
		 *
		 * Added a field to struct build_id_event that broke the file
		 * format.
		 *
		 * Since the kernel build-id is the first entry, process the
		 * table using the old format if the well known
		 * '[kernel.kallsyms]' string for the kernel build-id has the
		 * first 4 characters chopped off (where the pid_t sits).
		 */
		if (memcmp(filename, "nel.kallsyms]", 13) == 0) {
			if (lseek(input, orig_offset, SEEK_SET) == (off_t)-1)
				return -1;
			return perf_header__read_build_ids_abi_quirk(header, input, offset, size);
		}

		__event_process_build_id(&bev, filename, session);

		offset += bev.header.size;
	}
	err = 0;
out:
	return err;
}

1366 1367 1368
static int process_tracing_data(struct perf_file_section *section __maybe_unused,
				struct perf_header *ph __maybe_unused,
				int fd, void *data)
1369
{
1370 1371
	ssize_t ret = trace_report(fd, data, false);
	return ret < 0 ? -1 : 0;
1372 1373 1374
}

static int process_build_id(struct perf_file_section *section,
1375
			    struct perf_header *ph, int fd,
1376
			    void *data __maybe_unused)
1377 1378 1379 1380 1381 1382
{
	if (perf_header__read_build_ids(ph, fd, section->offset, section->size))
		pr_debug("Failed to read buildids, continuing...\n");
	return 0;
}

1383
static int process_hostname(struct perf_file_section *section __maybe_unused,
1384 1385
			    struct perf_header *ph, int fd,
			    void *data __maybe_unused)
1386 1387 1388 1389 1390 1391
{
	ph->env.hostname = do_read_string(fd, ph);
	return ph->env.hostname ? 0 : -ENOMEM;
}

static int process_osrelease(struct perf_file_section *section __maybe_unused,
1392 1393
			     struct perf_header *ph, int fd,
			     void *data __maybe_unused)
1394 1395 1396 1397 1398 1399
{
	ph->env.os_release = do_read_string(fd, ph);
	return ph->env.os_release ? 0 : -ENOMEM;
}

static int process_version(struct perf_file_section *section __maybe_unused,
1400 1401
			   struct perf_header *ph, int fd,
			   void *data __maybe_unused)
1402 1403 1404 1405 1406 1407
{
	ph->env.version = do_read_string(fd, ph);
	return ph->env.version ? 0 : -ENOMEM;
}

static int process_arch(struct perf_file_section *section __maybe_unused,
1408 1409
			struct perf_header *ph,	int fd,
			void *data __maybe_unused)
1410 1411 1412 1413 1414 1415
{
	ph->env.arch = do_read_string(fd, ph);
	return ph->env.arch ? 0 : -ENOMEM;
}

static int process_nrcpus(struct perf_file_section *section __maybe_unused,
1416 1417
			  struct perf_header *ph, int fd,
			  void *data __maybe_unused)
1418
{
1419
	ssize_t ret;
1420 1421
	u32 nr;

1422
	ret = readn(fd, &nr, sizeof(nr));
1423 1424 1425 1426 1427 1428 1429 1430
	if (ret != sizeof(nr))
		return -1;

	if (ph->needs_swap)
		nr = bswap_32(nr);

	ph->env.nr_cpus_online = nr;

1431
	ret = readn(fd, &nr, sizeof(nr));
1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442
	if (ret != sizeof(nr))
		return -1;

	if (ph->needs_swap)
		nr = bswap_32(nr);

	ph->env.nr_cpus_avail = nr;
	return 0;
}

static int process_cpudesc(struct perf_file_section *section __maybe_unused,
1443 1444
			   struct perf_header *ph, int fd,
			   void *data __maybe_unused)
1445 1446 1447 1448 1449 1450
{
	ph->env.cpu_desc = do_read_string(fd, ph);
	return ph->env.cpu_desc ? 0 : -ENOMEM;
}

static int process_cpuid(struct perf_file_section *section __maybe_unused,
1451 1452
			 struct perf_header *ph,  int fd,
			 void *data __maybe_unused)
1453 1454 1455 1456 1457 1458
{
	ph->env.cpuid = do_read_string(fd, ph);
	return ph->env.cpuid ? 0 : -ENOMEM;
}

static int process_total_mem(struct perf_file_section *section __maybe_unused,
1459 1460
			     struct perf_header *ph, int fd,
			     void *data __maybe_unused)
1461 1462
{
	uint64_t mem;
1463
	ssize_t ret;
1464

1465
	ret = readn(fd, &mem, sizeof(mem));
1466 1467 1468 1469 1470 1471 1472 1473 1474 1475
	if (ret != sizeof(mem))
		return -1;

	if (ph->needs_swap)
		mem = bswap_64(mem);

	ph->env.total_mem = mem;
	return 0;
}

1476 1477 1478 1479 1480
static struct perf_evsel *
perf_evlist__find_by_index(struct perf_evlist *evlist, int idx)
{
	struct perf_evsel *evsel;

1481
	evlist__for_each(evlist, evsel) {
1482 1483 1484 1485 1486 1487 1488 1489
		if (evsel->idx == idx)
			return evsel;
	}

	return NULL;
}

static void
1490 1491
perf_evlist__set_event_name(struct perf_evlist *evlist,
			    struct perf_evsel *event)
1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508
{
	struct perf_evsel *evsel;

	if (!event->name)
		return;

	evsel = perf_evlist__find_by_index(evlist, event->idx);
	if (!evsel)
		return;

	if (evsel->name)
		return;

	evsel->name = strdup(event->name);
}

static int
1509
process_event_desc(struct perf_file_section *section __maybe_unused,
1510
		   struct perf_header *header, int fd,
1511
		   void *data __maybe_unused)
1512
{
1513
	struct perf_session *session;
1514 1515 1516 1517 1518
	struct perf_evsel *evsel, *events = read_event_desc(header, fd);

	if (!events)
		return 0;

1519
	session = container_of(header, struct perf_session, header);
1520 1521 1522 1523 1524 1525 1526 1527
	for (evsel = events; evsel->attr.size; evsel++)
		perf_evlist__set_event_name(session->evlist, evsel);

	free_event_desc(events);

	return 0;
}

1528
static int process_cmdline(struct perf_file_section *section __maybe_unused,
1529 1530
			   struct perf_header *ph, int fd,
			   void *data __maybe_unused)
1531
{
1532
	ssize_t ret;
1533 1534 1535 1536
	char *str;
	u32 nr, i;
	struct strbuf sb;

1537
	ret = readn(fd, &nr, sizeof(nr));
1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564
	if (ret != sizeof(nr))
		return -1;

	if (ph->needs_swap)
		nr = bswap_32(nr);

	ph->env.nr_cmdline = nr;
	strbuf_init(&sb, 128);

	for (i = 0; i < nr; i++) {
		str = do_read_string(fd, ph);
		if (!str)
			goto error;

		/* include a NULL character at the end */
		strbuf_add(&sb, str, strlen(str) + 1);
		free(str);
	}
	ph->env.cmdline = strbuf_detach(&sb, NULL);
	return 0;

error:
	strbuf_release(&sb);
	return -1;
}

static int process_cpu_topology(struct perf_file_section *section __maybe_unused,
1565 1566
				struct perf_header *ph, int fd,
				void *data __maybe_unused)
1567
{
1568
	ssize_t ret;
1569 1570 1571 1572
	u32 nr, i;
	char *str;
	struct strbuf sb;

1573
	ret = readn(fd, &nr, sizeof(nr));
1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593
	if (ret != sizeof(nr))
		return -1;

	if (ph->needs_swap)
		nr = bswap_32(nr);

	ph->env.nr_sibling_cores = nr;
	strbuf_init(&sb, 128);

	for (i = 0; i < nr; i++) {
		str = do_read_string(fd, ph);
		if (!str)
			goto error;

		/* include a NULL character at the end */
		strbuf_add(&sb, str, strlen(str) + 1);
		free(str);
	}
	ph->env.sibling_cores = strbuf_detach(&sb, NULL);

1594
	ret = readn(fd, &nr, sizeof(nr));
1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620
	if (ret != sizeof(nr))
		return -1;

	if (ph->needs_swap)
		nr = bswap_32(nr);

	ph->env.nr_sibling_threads = nr;

	for (i = 0; i < nr; i++) {
		str = do_read_string(fd, ph);
		if (!str)
			goto error;

		/* include a NULL character at the end */
		strbuf_add(&sb, str, strlen(str) + 1);
		free(str);
	}
	ph->env.sibling_threads = strbuf_detach(&sb, NULL);
	return 0;

error:
	strbuf_release(&sb);
	return -1;
}

static int process_numa_topology(struct perf_file_section *section __maybe_unused,
1621 1622
				 struct perf_header *ph, int fd,
				 void *data __maybe_unused)
1623
{
1624
	ssize_t ret;
1625 1626 1627 1628 1629 1630
	u32 nr, node, i;
	char *str;
	uint64_t mem_total, mem_free;
	struct strbuf sb;

	/* nr nodes */
1631
	ret = readn(fd, &nr, sizeof(nr));
1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642
	if (ret != sizeof(nr))
		goto error;

	if (ph->needs_swap)
		nr = bswap_32(nr);

	ph->env.nr_numa_nodes = nr;
	strbuf_init(&sb, 256);

	for (i = 0; i < nr; i++) {
		/* node number */
1643
		ret = readn(fd, &node, sizeof(node));
1644 1645 1646
		if (ret != sizeof(node))
			goto error;

1647
		ret = readn(fd, &mem_total, sizeof(u64));
1648 1649 1650
		if (ret != sizeof(u64))
			goto error;

1651
		ret = readn(fd, &mem_free, sizeof(u64));
1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680
		if (ret != sizeof(u64))
			goto error;

		if (ph->needs_swap) {
			node = bswap_32(node);
			mem_total = bswap_64(mem_total);
			mem_free = bswap_64(mem_free);
		}

		strbuf_addf(&sb, "%u:%"PRIu64":%"PRIu64":",
			    node, mem_total, mem_free);

		str = do_read_string(fd, ph);
		if (!str)
			goto error;

		/* include a NULL character at the end */
		strbuf_add(&sb, str, strlen(str) + 1);
		free(str);
	}
	ph->env.numa_nodes = strbuf_detach(&sb, NULL);
	return 0;

error:
	strbuf_release(&sb);
	return -1;
}

static int process_pmu_mappings(struct perf_file_section *section __maybe_unused,
1681 1682
				struct perf_header *ph, int fd,
				void *data __maybe_unused)
1683
{
1684
	ssize_t ret;
1685 1686 1687 1688 1689
	char *name;
	u32 pmu_num;
	u32 type;
	struct strbuf sb;

1690
	ret = readn(fd, &pmu_num, sizeof(pmu_num));
1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705
	if (ret != sizeof(pmu_num))
		return -1;

	if (ph->needs_swap)
		pmu_num = bswap_32(pmu_num);

	if (!pmu_num) {
		pr_debug("pmu mappings not available\n");
		return 0;
	}

	ph->env.nr_pmu_mappings = pmu_num;
	strbuf_init(&sb, 128);

	while (pmu_num) {
1706
		if (readn(fd, &type, sizeof(type)) != sizeof(type))
1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729
			goto error;
		if (ph->needs_swap)
			type = bswap_32(type);

		name = do_read_string(fd, ph);
		if (!name)
			goto error;

		strbuf_addf(&sb, "%u:%s", type, name);
		/* include a NULL character at the end */
		strbuf_add(&sb, "", 1);

		free(name);
		pmu_num--;
	}
	ph->env.pmu_mappings = strbuf_detach(&sb, NULL);
	return 0;

error:
	strbuf_release(&sb);
	return -1;
}

1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783
static int process_group_desc(struct perf_file_section *section __maybe_unused,
			      struct perf_header *ph, int fd,
			      void *data __maybe_unused)
{
	size_t ret = -1;
	u32 i, nr, nr_groups;
	struct perf_session *session;
	struct perf_evsel *evsel, *leader = NULL;
	struct group_desc {
		char *name;
		u32 leader_idx;
		u32 nr_members;
	} *desc;

	if (readn(fd, &nr_groups, sizeof(nr_groups)) != sizeof(nr_groups))
		return -1;

	if (ph->needs_swap)
		nr_groups = bswap_32(nr_groups);

	ph->env.nr_groups = nr_groups;
	if (!nr_groups) {
		pr_debug("group desc not available\n");
		return 0;
	}

	desc = calloc(nr_groups, sizeof(*desc));
	if (!desc)
		return -1;

	for (i = 0; i < nr_groups; i++) {
		desc[i].name = do_read_string(fd, ph);
		if (!desc[i].name)
			goto out_free;

		if (readn(fd, &desc[i].leader_idx, sizeof(u32)) != sizeof(u32))
			goto out_free;

		if (readn(fd, &desc[i].nr_members, sizeof(u32)) != sizeof(u32))
			goto out_free;

		if (ph->needs_swap) {
			desc[i].leader_idx = bswap_32(desc[i].leader_idx);
			desc[i].nr_members = bswap_32(desc[i].nr_members);
		}
	}

	/*
	 * Rebuild group relationship based on the group_desc
	 */
	session = container_of(ph, struct perf_session, header);
	session->evlist->nr_groups = nr_groups;

	i = nr = 0;
1784
	evlist__for_each(session->evlist, evsel) {
1785 1786 1787
		if (evsel->idx == (int) desc[i].leader_idx) {
			evsel->leader = evsel;
			/* {anon_group} is a dummy name */
N
Namhyung Kim 已提交
1788
			if (strcmp(desc[i].name, "{anon_group}")) {
1789
				evsel->group_name = desc[i].name;
N
Namhyung Kim 已提交
1790 1791
				desc[i].name = NULL;
			}
1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816
			evsel->nr_members = desc[i].nr_members;

			if (i >= nr_groups || nr > 0) {
				pr_debug("invalid group desc\n");
				goto out_free;
			}

			leader = evsel;
			nr = evsel->nr_members - 1;
			i++;
		} else if (nr) {
			/* This is a group member */
			evsel->leader = leader;

			nr--;
		}
	}

	if (i != nr_groups || nr != 0) {
		pr_debug("invalid group desc\n");
		goto out_free;
	}

	ret = 0;
out_free:
1817
	for (i = 0; i < nr_groups; i++)
1818
		zfree(&desc[i].name);
1819 1820 1821 1822 1823
	free(desc);

	return ret;
}

1824 1825 1826
struct feature_ops {
	int (*write)(int fd, struct perf_header *h, struct perf_evlist *evlist);
	void (*print)(struct perf_header *h, int fd, FILE *fp);
1827
	int (*process)(struct perf_file_section *section,
1828
		       struct perf_header *h, int fd, void *data);
1829 1830 1831 1832
	const char *name;
	bool full_only;
};

1833 1834
#define FEAT_OPA(n, func) \
	[n] = { .name = #n, .write = write_##func, .print = print_##func }
1835 1836 1837
#define FEAT_OPP(n, func) \
	[n] = { .name = #n, .write = write_##func, .print = print_##func, \
		.process = process_##func }
1838
#define FEAT_OPF(n, func) \
1839
	[n] = { .name = #n, .write = write_##func, .print = print_##func, \
1840
		.process = process_##func, .full_only = true }
1841 1842

/* feature_ops not implemented: */
1843 1844
#define print_tracing_data	NULL
#define print_build_id		NULL
1845 1846

static const struct feature_ops feat_ops[HEADER_LAST_FEATURE] = {
1847
	FEAT_OPP(HEADER_TRACING_DATA,	tracing_data),
1848
	FEAT_OPP(HEADER_BUILD_ID,	build_id),
1849 1850 1851 1852 1853 1854
	FEAT_OPP(HEADER_HOSTNAME,	hostname),
	FEAT_OPP(HEADER_OSRELEASE,	osrelease),
	FEAT_OPP(HEADER_VERSION,	version),
	FEAT_OPP(HEADER_ARCH,		arch),
	FEAT_OPP(HEADER_NRCPUS,		nrcpus),
	FEAT_OPP(HEADER_CPUDESC,	cpudesc),
1855
	FEAT_OPP(HEADER_CPUID,		cpuid),
1856
	FEAT_OPP(HEADER_TOTAL_MEM,	total_mem),
1857
	FEAT_OPP(HEADER_EVENT_DESC,	event_desc),
1858
	FEAT_OPP(HEADER_CMDLINE,	cmdline),
1859 1860
	FEAT_OPF(HEADER_CPU_TOPOLOGY,	cpu_topology),
	FEAT_OPF(HEADER_NUMA_TOPOLOGY,	numa_topology),
1861
	FEAT_OPA(HEADER_BRANCH_STACK,	branch_stack),
1862
	FEAT_OPP(HEADER_PMU_MAPPINGS,	pmu_mappings),
1863
	FEAT_OPP(HEADER_GROUP_DESC,	group_desc),
1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881
};

struct header_print_data {
	FILE *fp;
	bool full; /* extended list of headers */
};

static int perf_file_section__fprintf_info(struct perf_file_section *section,
					   struct perf_header *ph,
					   int feat, int fd, void *data)
{
	struct header_print_data *hd = data;

	if (lseek(fd, section->offset, SEEK_SET) == (off_t)-1) {
		pr_debug("Failed to lseek to %" PRIu64 " offset for feature "
				"%d, continuing...\n", section->offset, feat);
		return 0;
	}
1882
	if (feat >= HEADER_LAST_FEATURE) {
1883
		pr_warning("unknown feature %d\n", feat);
1884
		return 0;
1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901
	}
	if (!feat_ops[feat].print)
		return 0;

	if (!feat_ops[feat].full_only || hd->full)
		feat_ops[feat].print(ph, fd, hd->fp);
	else
		fprintf(hd->fp, "# %s info available, use -I to display\n",
			feat_ops[feat].name);

	return 0;
}

int perf_header__fprintf_info(struct perf_session *session, FILE *fp, bool full)
{
	struct header_print_data hd;
	struct perf_header *header = &session->header;
1902
	int fd = perf_data_file__fd(session->file);
1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918
	hd.fp = fp;
	hd.full = full;

	perf_header__process_sections(header, fd, &hd,
				      perf_file_section__fprintf_info);
	return 0;
}

static int do_write_feat(int fd, struct perf_header *h, int type,
			 struct perf_file_section **p,
			 struct perf_evlist *evlist)
{
	int err;
	int ret = 0;

	if (perf_header__has_feat(h, type)) {
1919 1920
		if (!feat_ops[type].write)
			return -1;
1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938

		(*p)->offset = lseek(fd, 0, SEEK_CUR);

		err = feat_ops[type].write(fd, h, evlist);
		if (err < 0) {
			pr_debug("failed to write feature %d\n", type);

			/* undo anything written */
			lseek(fd, (*p)->offset, SEEK_SET);

			return -1;
		}
		(*p)->size = lseek(fd, 0, SEEK_CUR) - (*p)->offset;
		(*p)++;
	}
	return ret;
}

1939
static int perf_header__adds_write(struct perf_header *header,
1940
				   struct perf_evlist *evlist, int fd)
1941
{
1942
	int nr_sections;
1943
	struct perf_file_section *feat_sec, *p;
1944 1945
	int sec_size;
	u64 sec_start;
1946
	int feat;
1947
	int err;
1948

1949
	nr_sections = bitmap_weight(header->adds_features, HEADER_FEAT_BITS);
1950
	if (!nr_sections)
1951
		return 0;
1952

1953
	feat_sec = p = calloc(nr_sections, sizeof(*feat_sec));
1954 1955
	if (feat_sec == NULL)
		return -ENOMEM;
1956 1957 1958

	sec_size = sizeof(*feat_sec) * nr_sections;

1959
	sec_start = header->feat_offset;
1960
	lseek(fd, sec_start + sec_size, SEEK_SET);
1961

1962 1963 1964 1965
	for_each_set_bit(feat, header->adds_features, HEADER_FEAT_BITS) {
		if (do_write_feat(fd, header, feat, &p, evlist))
			perf_header__clear_feat(header, feat);
	}
1966

1967
	lseek(fd, sec_start, SEEK_SET);
1968 1969 1970 1971
	/*
	 * may write more than needed due to dropped feature, but
	 * this is okay, reader will skip the mising entries
	 */
1972 1973 1974
	err = do_write(fd, feat_sec, sec_size);
	if (err < 0)
		pr_debug("failed to write feature section\n");
1975
	free(feat_sec);
1976
	return err;
1977
}
1978

1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997
int perf_header__write_pipe(int fd)
{
	struct perf_pipe_file_header f_header;
	int err;

	f_header = (struct perf_pipe_file_header){
		.magic	   = PERF_MAGIC,
		.size	   = sizeof(f_header),
	};

	err = do_write(fd, &f_header, sizeof(f_header));
	if (err < 0) {
		pr_debug("failed to write perf pipe header\n");
		return err;
	}

	return 0;
}

1998 1999 2000
int perf_session__write_header(struct perf_session *session,
			       struct perf_evlist *evlist,
			       int fd, bool at_exit)
2001 2002 2003
{
	struct perf_file_header f_header;
	struct perf_file_attr   f_attr;
2004
	struct perf_header *header = &session->header;
2005
	struct perf_evsel *evsel;
2006
	u64 attr_offset;
2007
	int err;
2008 2009 2010

	lseek(fd, sizeof(f_header), SEEK_SET);

2011
	evlist__for_each(session->evlist, evsel) {
2012 2013
		evsel->id_offset = lseek(fd, 0, SEEK_CUR);
		err = do_write(fd, evsel->id, evsel->ids * sizeof(u64));
2014 2015 2016 2017
		if (err < 0) {
			pr_debug("failed to write perf header\n");
			return err;
		}
2018 2019
	}

2020
	attr_offset = lseek(fd, 0, SEEK_CUR);
2021

2022
	evlist__for_each(evlist, evsel) {
2023
		f_attr = (struct perf_file_attr){
2024
			.attr = evsel->attr,
2025
			.ids  = {
2026 2027
				.offset = evsel->id_offset,
				.size   = evsel->ids * sizeof(u64),
2028 2029
			}
		};
2030 2031 2032 2033 2034
		err = do_write(fd, &f_attr, sizeof(f_attr));
		if (err < 0) {
			pr_debug("failed to write perf header attribute\n");
			return err;
		}
2035 2036
	}

2037 2038
	if (!header->data_offset)
		header->data_offset = lseek(fd, 0, SEEK_CUR);
2039
	header->feat_offset = header->data_offset + header->data_size;
2040

2041
	if (at_exit) {
2042
		err = perf_header__adds_write(header, evlist, fd);
2043 2044 2045
		if (err < 0)
			return err;
	}
2046

2047 2048 2049 2050 2051
	f_header = (struct perf_file_header){
		.magic	   = PERF_MAGIC,
		.size	   = sizeof(f_header),
		.attr_size = sizeof(f_attr),
		.attrs = {
2052
			.offset = attr_offset,
2053
			.size   = evlist->nr_entries * sizeof(f_attr),
2054 2055
		},
		.data = {
2056 2057
			.offset = header->data_offset,
			.size	= header->data_size,
2058
		},
2059
		/* event_types is ignored, store zeros */
2060 2061
	};

2062
	memcpy(&f_header.adds_features, &header->adds_features, sizeof(header->adds_features));
2063

2064
	lseek(fd, 0, SEEK_SET);
2065 2066 2067 2068 2069
	err = do_write(fd, &f_header, sizeof(f_header));
	if (err < 0) {
		pr_debug("failed to write perf header\n");
		return err;
	}
2070
	lseek(fd, header->data_offset + header->data_size, SEEK_SET);
2071

2072
	return 0;
2073 2074
}

2075
static int perf_header__getbuffer64(struct perf_header *header,
2076 2077
				    int fd, void *buf, size_t size)
{
2078
	if (readn(fd, buf, size) <= 0)
2079 2080
		return -1;

2081
	if (header->needs_swap)
2082 2083 2084 2085 2086
		mem_bswap_64(buf, size);

	return 0;
}

2087
int perf_header__process_sections(struct perf_header *header, int fd,
2088
				  void *data,
2089
				  int (*process)(struct perf_file_section *section,
2090 2091
						 struct perf_header *ph,
						 int feat, int fd, void *data))
2092
{
2093
	struct perf_file_section *feat_sec, *sec;
2094 2095
	int nr_sections;
	int sec_size;
2096 2097
	int feat;
	int err;
2098

2099
	nr_sections = bitmap_weight(header->adds_features, HEADER_FEAT_BITS);
2100
	if (!nr_sections)
2101
		return 0;
2102

2103
	feat_sec = sec = calloc(nr_sections, sizeof(*feat_sec));
2104
	if (!feat_sec)
2105
		return -1;
2106 2107 2108

	sec_size = sizeof(*feat_sec) * nr_sections;

2109
	lseek(fd, header->feat_offset, SEEK_SET);
2110

2111 2112
	err = perf_header__getbuffer64(header, fd, feat_sec, sec_size);
	if (err < 0)
2113
		goto out_free;
2114

2115 2116 2117 2118
	for_each_set_bit(feat, header->adds_features, HEADER_LAST_FEATURE) {
		err = process(sec++, header, feat, fd, data);
		if (err < 0)
			goto out_free;
2119
	}
2120
	err = 0;
2121
out_free:
2122 2123
	free(feat_sec);
	return err;
2124
}
2125

2126 2127 2128
static const int attr_file_abi_sizes[] = {
	[0] = PERF_ATTR_SIZE_VER0,
	[1] = PERF_ATTR_SIZE_VER1,
2129
	[2] = PERF_ATTR_SIZE_VER2,
2130
	[3] = PERF_ATTR_SIZE_VER3,
2131
	[4] = PERF_ATTR_SIZE_VER4,
2132 2133 2134 2135 2136 2137 2138 2139 2140 2141
	0,
};

/*
 * In the legacy file format, the magic number is not used to encode endianness.
 * hdr_sz was used to encode endianness. But given that hdr_sz can vary based
 * on ABI revisions, we need to try all combinations for all endianness to
 * detect the endianness.
 */
static int try_all_file_abis(uint64_t hdr_sz, struct perf_header *ph)
2142
{
2143 2144
	uint64_t ref_size, attr_size;
	int i;
2145

2146 2147 2148 2149 2150 2151 2152
	for (i = 0 ; attr_file_abi_sizes[i]; i++) {
		ref_size = attr_file_abi_sizes[i]
			 + sizeof(struct perf_file_section);
		if (hdr_sz != ref_size) {
			attr_size = bswap_64(hdr_sz);
			if (attr_size != ref_size)
				continue;
2153

2154 2155 2156 2157 2158 2159 2160 2161 2162 2163
			ph->needs_swap = true;
		}
		pr_debug("ABI%d perf.data file detected, need_swap=%d\n",
			 i,
			 ph->needs_swap);
		return 0;
	}
	/* could not determine endianness */
	return -1;
}
2164

2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188
#define PERF_PIPE_HDR_VER0	16

static const size_t attr_pipe_abi_sizes[] = {
	[0] = PERF_PIPE_HDR_VER0,
	0,
};

/*
 * In the legacy pipe format, there is an implicit assumption that endiannesss
 * between host recording the samples, and host parsing the samples is the
 * same. This is not always the case given that the pipe output may always be
 * redirected into a file and analyzed on a different machine with possibly a
 * different endianness and perf_event ABI revsions in the perf tool itself.
 */
static int try_all_pipe_abis(uint64_t hdr_sz, struct perf_header *ph)
{
	u64 attr_size;
	int i;

	for (i = 0 ; attr_pipe_abi_sizes[i]; i++) {
		if (hdr_sz != attr_pipe_abi_sizes[i]) {
			attr_size = bswap_64(hdr_sz);
			if (attr_size != hdr_sz)
				continue;
2189 2190 2191

			ph->needs_swap = true;
		}
2192
		pr_debug("Pipe ABI%d perf.data file detected\n", i);
2193 2194
		return 0;
	}
2195 2196 2197
	return -1;
}

F
Feng Tang 已提交
2198 2199 2200 2201 2202 2203 2204 2205 2206 2207
bool is_perf_magic(u64 magic)
{
	if (!memcmp(&magic, __perf_magic1, sizeof(magic))
		|| magic == __perf_magic2
		|| magic == __perf_magic2_sw)
		return true;

	return false;
}

2208 2209 2210 2211 2212 2213 2214 2215
static int check_magic_endian(u64 magic, uint64_t hdr_sz,
			      bool is_pipe, struct perf_header *ph)
{
	int ret;

	/* check for legacy format */
	ret = memcmp(&magic, __perf_magic1, sizeof(magic));
	if (ret == 0) {
2216
		ph->version = PERF_HEADER_VERSION_1;
2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227
		pr_debug("legacy perf.data format\n");
		if (is_pipe)
			return try_all_pipe_abis(hdr_sz, ph);

		return try_all_file_abis(hdr_sz, ph);
	}
	/*
	 * the new magic number serves two purposes:
	 * - unique number to identify actual perf.data files
	 * - encode endianness of file
	 */
2228
	ph->version = PERF_HEADER_VERSION_2;
2229

2230 2231
	/* check magic number with one endianness */
	if (magic == __perf_magic2)
2232 2233
		return 0;

2234 2235
	/* check magic number with opposite endianness */
	if (magic != __perf_magic2_sw)
2236 2237 2238 2239 2240 2241 2242
		return -1;

	ph->needs_swap = true;

	return 0;
}

2243
int perf_file_header__read(struct perf_file_header *header,
2244 2245
			   struct perf_header *ph, int fd)
{
2246
	ssize_t ret;
2247

2248 2249
	lseek(fd, 0, SEEK_SET);

2250 2251
	ret = readn(fd, header, sizeof(*header));
	if (ret <= 0)
2252 2253
		return -1;

2254 2255 2256
	if (check_magic_endian(header->magic,
			       header->attr_size, false, ph) < 0) {
		pr_debug("magic/endian check failed\n");
2257
		return -1;
2258
	}
2259

2260
	if (ph->needs_swap) {
2261
		mem_bswap_64(header, offsetof(struct perf_file_header,
2262
			     adds_features));
2263 2264
	}

2265
	if (header->size != sizeof(*header)) {
2266
		/* Support the previous format */
2267 2268
		if (header->size == offsetof(typeof(*header), adds_features))
			bitmap_zero(header->adds_features, HEADER_FEAT_BITS);
2269 2270
		else
			return -1;
2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286
	} else if (ph->needs_swap) {
		/*
		 * feature bitmap is declared as an array of unsigned longs --
		 * not good since its size can differ between the host that
		 * generated the data file and the host analyzing the file.
		 *
		 * We need to handle endianness, but we don't know the size of
		 * the unsigned long where the file was generated. Take a best
		 * guess at determining it: try 64-bit swap first (ie., file
		 * created on a 64-bit host), and check if the hostname feature
		 * bit is set (this feature bit is forced on as of fbe96f2).
		 * If the bit is not, undo the 64-bit swap and try a 32-bit
		 * swap. If the hostname bit is still not set (e.g., older data
		 * file), punt and fallback to the original behavior --
		 * clearing all feature bits and setting buildid.
		 */
2287 2288
		mem_bswap_64(&header->adds_features,
			    BITS_TO_U64(HEADER_FEAT_BITS));
2289 2290

		if (!test_bit(HEADER_HOSTNAME, header->adds_features)) {
2291 2292 2293 2294 2295 2296 2297
			/* unswap as u64 */
			mem_bswap_64(&header->adds_features,
				    BITS_TO_U64(HEADER_FEAT_BITS));

			/* unswap as u32 */
			mem_bswap_32(&header->adds_features,
				    BITS_TO_U32(HEADER_FEAT_BITS));
2298 2299 2300 2301 2302 2303
		}

		if (!test_bit(HEADER_HOSTNAME, header->adds_features)) {
			bitmap_zero(header->adds_features, HEADER_FEAT_BITS);
			set_bit(HEADER_BUILD_ID, header->adds_features);
		}
2304
	}
2305

2306
	memcpy(&ph->adds_features, &header->adds_features,
2307
	       sizeof(ph->adds_features));
2308

2309 2310
	ph->data_offset  = header->data.offset;
	ph->data_size	 = header->data.size;
2311
	ph->feat_offset  = header->data.offset + header->data.size;
2312 2313 2314
	return 0;
}

2315
static int perf_file_section__process(struct perf_file_section *section,
2316
				      struct perf_header *ph,
2317
				      int feat, int fd, void *data)
2318
{
2319
	if (lseek(fd, section->offset, SEEK_SET) == (off_t)-1) {
2320
		pr_debug("Failed to lseek to %" PRIu64 " offset for feature "
2321
			  "%d, continuing...\n", section->offset, feat);
2322 2323 2324
		return 0;
	}

2325 2326 2327 2328 2329
	if (feat >= HEADER_LAST_FEATURE) {
		pr_debug("unknown feature %d, continuing...\n", feat);
		return 0;
	}

2330 2331
	if (!feat_ops[feat].process)
		return 0;
2332

2333
	return feat_ops[feat].process(section, ph, fd, data);
2334
}
2335

2336
static int perf_file_header__read_pipe(struct perf_pipe_file_header *header,
T
Tom Zanussi 已提交
2337 2338
				       struct perf_header *ph, int fd,
				       bool repipe)
2339
{
2340
	ssize_t ret;
2341 2342 2343 2344 2345

	ret = readn(fd, header, sizeof(*header));
	if (ret <= 0)
		return -1;

2346 2347
	if (check_magic_endian(header->magic, header->size, true, ph) < 0) {
		pr_debug("endian/magic failed\n");
2348
		return -1;
2349 2350 2351 2352
	}

	if (ph->needs_swap)
		header->size = bswap_64(header->size);
2353

2354
	if (repipe && do_write(STDOUT_FILENO, header, sizeof(*header)) < 0)
T
Tom Zanussi 已提交
2355 2356
		return -1;

2357 2358 2359
	return 0;
}

2360
static int perf_header__read_pipe(struct perf_session *session)
2361
{
2362
	struct perf_header *header = &session->header;
2363 2364
	struct perf_pipe_file_header f_header;

2365 2366
	if (perf_file_header__read_pipe(&f_header, header,
					perf_data_file__fd(session->file),
T
Tom Zanussi 已提交
2367
					session->repipe) < 0) {
2368 2369 2370 2371 2372 2373 2374
		pr_debug("incompatible file format\n");
		return -EINVAL;
	}

	return 0;
}

2375 2376 2377 2378 2379 2380
static int read_attr(int fd, struct perf_header *ph,
		     struct perf_file_attr *f_attr)
{
	struct perf_event_attr *attr = &f_attr->attr;
	size_t sz, left;
	size_t our_sz = sizeof(f_attr->attr);
2381
	ssize_t ret;
2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394

	memset(f_attr, 0, sizeof(*f_attr));

	/* read minimal guaranteed structure */
	ret = readn(fd, attr, PERF_ATTR_SIZE_VER0);
	if (ret <= 0) {
		pr_debug("cannot read %d bytes of header attr\n",
			 PERF_ATTR_SIZE_VER0);
		return -1;
	}

	/* on file perf_event_attr size */
	sz = attr->size;
2395

2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420
	if (ph->needs_swap)
		sz = bswap_32(sz);

	if (sz == 0) {
		/* assume ABI0 */
		sz =  PERF_ATTR_SIZE_VER0;
	} else if (sz > our_sz) {
		pr_debug("file uses a more recent and unsupported ABI"
			 " (%zu bytes extra)\n", sz - our_sz);
		return -1;
	}
	/* what we have not yet read and that we know about */
	left = sz - PERF_ATTR_SIZE_VER0;
	if (left) {
		void *ptr = attr;
		ptr += PERF_ATTR_SIZE_VER0;

		ret = readn(fd, ptr, left);
	}
	/* read perf_file_section, ids are read in caller */
	ret = readn(fd, &f_attr->ids, sizeof(f_attr->ids));

	return ret <= 0 ? -1 : 0;
}

2421 2422
static int perf_evsel__prepare_tracepoint_event(struct perf_evsel *evsel,
						struct pevent *pevent)
2423
{
2424
	struct event_format *event;
2425 2426
	char bf[128];

2427 2428 2429 2430
	/* already prepared */
	if (evsel->tp_format)
		return 0;

2431 2432 2433 2434 2435
	if (pevent == NULL) {
		pr_debug("broken or missing trace data\n");
		return -1;
	}

2436
	event = pevent_find_event(pevent, evsel->attr.config);
2437 2438 2439
	if (event == NULL)
		return -1;

2440 2441 2442 2443 2444 2445
	if (!evsel->name) {
		snprintf(bf, sizeof(bf), "%s:%s", event->system, event->name);
		evsel->name = strdup(bf);
		if (evsel->name == NULL)
			return -1;
	}
2446

2447
	evsel->tp_format = event;
2448 2449 2450
	return 0;
}

2451 2452
static int perf_evlist__prepare_tracepoint_events(struct perf_evlist *evlist,
						  struct pevent *pevent)
2453 2454 2455
{
	struct perf_evsel *pos;

2456
	evlist__for_each(evlist, pos) {
2457 2458
		if (pos->attr.type == PERF_TYPE_TRACEPOINT &&
		    perf_evsel__prepare_tracepoint_event(pos, pevent))
2459 2460 2461 2462 2463 2464
			return -1;
	}

	return 0;
}

2465
int perf_session__read_header(struct perf_session *session)
2466
{
2467
	struct perf_data_file *file = session->file;
2468
	struct perf_header *header = &session->header;
2469
	struct perf_file_header	f_header;
2470 2471 2472
	struct perf_file_attr	f_attr;
	u64			f_id;
	int nr_attrs, nr_ids, i, j;
2473
	int fd = perf_data_file__fd(file);
2474

2475
	session->evlist = perf_evlist__new();
2476 2477 2478
	if (session->evlist == NULL)
		return -ENOMEM;

2479
	if (perf_data_file__is_pipe(file))
2480
		return perf_header__read_pipe(session);
2481

2482
	if (perf_file_header__read(&f_header, header, fd) < 0)
2483
		return -EINVAL;
2484

2485 2486 2487 2488 2489 2490 2491 2492 2493
	/*
	 * Sanity check that perf.data was written cleanly; data size is
	 * initialized to 0 and updated only if the on_exit function is run.
	 * If data size is still 0 then the file contains only partial
	 * information.  Just warn user and process it as much as it can.
	 */
	if (f_header.data.size == 0) {
		pr_warning("WARNING: The %s file's data size field is 0 which is unexpected.\n"
			   "Was the 'perf record' command properly terminated?\n",
2494
			   file->path);
2495 2496
	}

2497
	nr_attrs = f_header.attrs.size / f_header.attr_size;
2498 2499 2500
	lseek(fd, f_header.attrs.offset, SEEK_SET);

	for (i = 0; i < nr_attrs; i++) {
2501
		struct perf_evsel *evsel;
2502
		off_t tmp;
2503

2504
		if (read_attr(fd, header, &f_attr) < 0)
2505
			goto out_errno;
2506

2507 2508 2509
		if (header->needs_swap)
			perf_event__attr_swap(&f_attr.attr);

2510
		tmp = lseek(fd, 0, SEEK_CUR);
2511
		evsel = perf_evsel__new(&f_attr.attr);
2512

2513 2514
		if (evsel == NULL)
			goto out_delete_evlist;
2515 2516

		evsel->needs_swap = header->needs_swap;
2517 2518 2519 2520 2521
		/*
		 * Do it before so that if perf_evsel__alloc_id fails, this
		 * entry gets purged too at perf_evlist__delete().
		 */
		perf_evlist__add(session->evlist, evsel);
2522 2523

		nr_ids = f_attr.ids.size / sizeof(u64);
2524 2525 2526 2527 2528 2529 2530 2531
		/*
		 * We don't have the cpu and thread maps on the header, so
		 * for allocating the perf_sample_id table we fake 1 cpu and
		 * hattr->ids threads.
		 */
		if (perf_evsel__alloc_id(evsel, 1, nr_ids))
			goto out_delete_evlist;

2532 2533 2534
		lseek(fd, f_attr.ids.offset, SEEK_SET);

		for (j = 0; j < nr_ids; j++) {
2535
			if (perf_header__getbuffer64(header, fd, &f_id, sizeof(f_id)))
2536
				goto out_errno;
2537

2538
			perf_evlist__id_add(session->evlist, evsel, 0, j, f_id);
2539
		}
2540

2541 2542 2543
		lseek(fd, tmp, SEEK_SET);
	}

2544 2545
	symbol_conf.nr_events = nr_attrs;

J
Jiri Olsa 已提交
2546
	perf_header__process_sections(header, fd, &session->tevent,
2547
				      perf_file_section__process);
2548

2549
	if (perf_evlist__prepare_tracepoint_events(session->evlist,
J
Jiri Olsa 已提交
2550
						   session->tevent.pevent))
2551 2552
		goto out_delete_evlist;

2553
	return 0;
2554 2555
out_errno:
	return -errno;
2556 2557 2558 2559 2560

out_delete_evlist:
	perf_evlist__delete(session->evlist);
	session->evlist = NULL;
	return -ENOMEM;
2561
}
2562

2563
int perf_event__synthesize_attr(struct perf_tool *tool,
2564
				struct perf_event_attr *attr, u32 ids, u64 *id,
2565
				perf_event__handler_t process)
2566
{
2567
	union perf_event *ev;
2568 2569 2570 2571
	size_t size;
	int err;

	size = sizeof(struct perf_event_attr);
2572
	size = PERF_ALIGN(size, sizeof(u64));
2573 2574 2575 2576 2577
	size += sizeof(struct perf_event_header);
	size += ids * sizeof(u64);

	ev = malloc(size);

2578 2579 2580
	if (ev == NULL)
		return -ENOMEM;

2581 2582 2583 2584
	ev->attr.attr = *attr;
	memcpy(ev->attr.id, id, ids * sizeof(u64));

	ev->attr.header.type = PERF_RECORD_HEADER_ATTR;
2585
	ev->attr.header.size = (u16)size;
2586

2587 2588 2589 2590
	if (ev->attr.header.size == size)
		err = process(tool, ev, NULL, NULL);
	else
		err = -E2BIG;
2591 2592 2593 2594 2595 2596

	free(ev);

	return err;
}

2597
int perf_event__synthesize_attrs(struct perf_tool *tool,
2598
				   struct perf_session *session,
2599
				   perf_event__handler_t process)
2600
{
2601
	struct perf_evsel *evsel;
2602
	int err = 0;
2603

2604
	evlist__for_each(session->evlist, evsel) {
2605 2606
		err = perf_event__synthesize_attr(tool, &evsel->attr, evsel->ids,
						  evsel->id, process);
2607 2608 2609 2610 2611 2612 2613 2614 2615
		if (err) {
			pr_debug("failed to create perf header attribute\n");
			return err;
		}
	}

	return err;
}

2616 2617
int perf_event__process_attr(struct perf_tool *tool __maybe_unused,
			     union perf_event *event,
2618
			     struct perf_evlist **pevlist)
2619
{
2620
	u32 i, ids, n_ids;
2621
	struct perf_evsel *evsel;
2622
	struct perf_evlist *evlist = *pevlist;
2623

2624
	if (evlist == NULL) {
2625
		*pevlist = evlist = perf_evlist__new();
2626
		if (evlist == NULL)
2627 2628 2629
			return -ENOMEM;
	}

2630
	evsel = perf_evsel__new(&event->attr.attr);
2631
	if (evsel == NULL)
2632 2633
		return -ENOMEM;

2634
	perf_evlist__add(evlist, evsel);
2635

2636 2637
	ids = event->header.size;
	ids -= (void *)&event->attr.id - (void *)event;
2638
	n_ids = ids / sizeof(u64);
2639 2640 2641 2642 2643 2644 2645
	/*
	 * We don't have the cpu and thread maps on the header, so
	 * for allocating the perf_sample_id table we fake 1 cpu and
	 * hattr->ids threads.
	 */
	if (perf_evsel__alloc_id(evsel, 1, n_ids))
		return -ENOMEM;
2646 2647

	for (i = 0; i < n_ids; i++) {
2648
		perf_evlist__id_add(evlist, evsel, 0, i, event->attr.id[i]);
2649 2650
	}

2651 2652
	symbol_conf.nr_events = evlist->nr_entries;

2653 2654
	return 0;
}
2655

2656
int perf_event__synthesize_tracing_data(struct perf_tool *tool, int fd,
2657
					struct perf_evlist *evlist,
2658
					perf_event__handler_t process)
2659
{
2660
	union perf_event ev;
J
Jiri Olsa 已提交
2661
	struct tracing_data *tdata;
2662
	ssize_t size = 0, aligned_size = 0, padding;
2663
	int err __maybe_unused = 0;
2664

J
Jiri Olsa 已提交
2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679
	/*
	 * We are going to store the size of the data followed
	 * by the data contents. Since the fd descriptor is a pipe,
	 * we cannot seek back to store the size of the data once
	 * we know it. Instead we:
	 *
	 * - write the tracing data to the temp file
	 * - get/write the data size to pipe
	 * - write the tracing data from the temp file
	 *   to the pipe
	 */
	tdata = tracing_data_get(&evlist->entries, fd, true);
	if (!tdata)
		return -1;

2680 2681 2682
	memset(&ev, 0, sizeof(ev));

	ev.tracing_data.header.type = PERF_RECORD_HEADER_TRACING_DATA;
J
Jiri Olsa 已提交
2683
	size = tdata->size;
2684
	aligned_size = PERF_ALIGN(size, sizeof(u64));
2685 2686 2687 2688
	padding = aligned_size - size;
	ev.tracing_data.header.size = sizeof(ev.tracing_data);
	ev.tracing_data.size = aligned_size;

2689
	process(tool, &ev, NULL, NULL);
2690

J
Jiri Olsa 已提交
2691 2692 2693 2694 2695 2696
	/*
	 * The put function will copy all the tracing data
	 * stored in temp file to the pipe.
	 */
	tracing_data_put(tdata);

2697 2698 2699 2700 2701
	write_padded(fd, NULL, 0, padding);

	return aligned_size;
}

2702 2703
int perf_event__process_tracing_data(struct perf_tool *tool __maybe_unused,
				     union perf_event *event,
2704
				     struct perf_session *session)
2705
{
2706
	ssize_t size_read, padding, size = event->tracing_data.size;
2707 2708
	int fd = perf_data_file__fd(session->file);
	off_t offset = lseek(fd, 0, SEEK_CUR);
2709 2710 2711
	char buf[BUFSIZ];

	/* setup for reading amidst mmap */
2712
	lseek(fd, offset + sizeof(struct tracing_data_event),
2713 2714
	      SEEK_SET);

J
Jiri Olsa 已提交
2715
	size_read = trace_report(fd, &session->tevent,
2716
				 session->repipe);
2717
	padding = PERF_ALIGN(size_read, sizeof(u64)) - size_read;
2718

2719
	if (readn(fd, buf, padding) < 0) {
2720 2721 2722
		pr_err("%s: reading input file", __func__);
		return -1;
	}
T
Tom Zanussi 已提交
2723 2724
	if (session->repipe) {
		int retw = write(STDOUT_FILENO, buf, padding);
2725 2726 2727 2728
		if (retw <= 0 || retw != padding) {
			pr_err("%s: repiping tracing data padding", __func__);
			return -1;
		}
T
Tom Zanussi 已提交
2729
	}
2730

2731 2732 2733 2734
	if (size_read + padding != size) {
		pr_err("%s: tracing data size mismatch", __func__);
		return -1;
	}
2735

2736
	perf_evlist__prepare_tracepoint_events(session->evlist,
J
Jiri Olsa 已提交
2737
					       session->tevent.pevent);
2738

2739 2740
	return size_read + padding;
}
2741

2742
int perf_event__synthesize_build_id(struct perf_tool *tool,
2743
				    struct dso *pos, u16 misc,
2744
				    perf_event__handler_t process,
2745
				    struct machine *machine)
2746
{
2747
	union perf_event ev;
2748 2749 2750 2751 2752 2753 2754 2755 2756
	size_t len;
	int err = 0;

	if (!pos->hit)
		return err;

	memset(&ev, 0, sizeof(ev));

	len = pos->long_name_len + 1;
2757
	len = PERF_ALIGN(len, NAME_ALIGN);
2758 2759 2760
	memcpy(&ev.build_id.build_id, pos->build_id, sizeof(pos->build_id));
	ev.build_id.header.type = PERF_RECORD_HEADER_BUILD_ID;
	ev.build_id.header.misc = misc;
2761
	ev.build_id.pid = machine->pid;
2762 2763 2764
	ev.build_id.header.size = sizeof(ev.build_id) + len;
	memcpy(&ev.build_id.filename, pos->long_name, pos->long_name_len);

2765
	err = process(tool, &ev, NULL, machine);
2766 2767 2768 2769

	return err;
}

2770
int perf_event__process_build_id(struct perf_tool *tool __maybe_unused,
2771
				 union perf_event *event,
2772
				 struct perf_session *session)
2773
{
2774 2775
	__event_process_build_id(&event->build_id,
				 event->build_id.filename,
2776
				 session);
2777 2778
	return 0;
}