symbol.c 28.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10
#include <dirent.h>
#include <errno.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/param.h>
#include <fcntl.h>
#include <unistd.h>
11
#include <inttypes.h>
12
#include "build-id.h"
13
#include "util.h"
14
#include "debug.h"
15
#include "machine.h"
16
#include "symbol.h"
17
#include "strlist.h"
18 19

#include <elf.h>
20
#include <limits.h>
21
#include <sys/utsname.h>
P
Peter Zijlstra 已提交
22

23
#ifndef KSYM_NAME_LEN
24
#define KSYM_NAME_LEN 256
25 26
#endif

27
static int dso__load_kernel_sym(struct dso *dso, struct map *map,
28
				symbol_filter_t filter);
29
static int dso__load_guest_kernel_sym(struct dso *dso, struct map *map,
30
			symbol_filter_t filter);
31 32
int vmlinux_path__nr_entries;
char **vmlinux_path;
33

34
struct symbol_conf symbol_conf = {
35 36
	.use_modules	  = true,
	.try_vmlinux_path = true,
37
	.annotate_src	  = true,
38
	.demangle	  = true,
39
	.symfs            = "",
40 41
};

42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
static enum dso_binary_type binary_type_symtab[] = {
	DSO_BINARY_TYPE__KALLSYMS,
	DSO_BINARY_TYPE__GUEST_KALLSYMS,
	DSO_BINARY_TYPE__JAVA_JIT,
	DSO_BINARY_TYPE__DEBUGLINK,
	DSO_BINARY_TYPE__BUILD_ID_CACHE,
	DSO_BINARY_TYPE__FEDORA_DEBUGINFO,
	DSO_BINARY_TYPE__UBUNTU_DEBUGINFO,
	DSO_BINARY_TYPE__BUILDID_DEBUGINFO,
	DSO_BINARY_TYPE__SYSTEM_PATH_DSO,
	DSO_BINARY_TYPE__GUEST_KMODULE,
	DSO_BINARY_TYPE__SYSTEM_PATH_KMODULE,
	DSO_BINARY_TYPE__NOT_FOUND,
};

57
#define DSO_BINARY_TYPE__SYMTAB_CNT ARRAY_SIZE(binary_type_symtab)
58

59
bool symbol_type__is_a(char symbol_type, enum map_type map_type)
60
{
61 62
	symbol_type = toupper(symbol_type);

63 64 65
	switch (map_type) {
	case MAP__FUNCTION:
		return symbol_type == 'T' || symbol_type == 'W';
66
	case MAP__VARIABLE:
67
		return symbol_type == 'D';
68 69 70 71 72
	default:
		return false;
	}
}

73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89
static int prefix_underscores_count(const char *str)
{
	const char *tail = str;

	while (*tail == '_')
		tail++;

	return tail - str;
}

#define SYMBOL_A 0
#define SYMBOL_B 1

static int choose_best_symbol(struct symbol *syma, struct symbol *symb)
{
	s64 a;
	s64 b;
90
	size_t na, nb;
91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123

	/* Prefer a symbol with non zero length */
	a = syma->end - syma->start;
	b = symb->end - symb->start;
	if ((b == 0) && (a > 0))
		return SYMBOL_A;
	else if ((a == 0) && (b > 0))
		return SYMBOL_B;

	/* Prefer a non weak symbol over a weak one */
	a = syma->binding == STB_WEAK;
	b = symb->binding == STB_WEAK;
	if (b && !a)
		return SYMBOL_A;
	if (a && !b)
		return SYMBOL_B;

	/* Prefer a global symbol over a non global one */
	a = syma->binding == STB_GLOBAL;
	b = symb->binding == STB_GLOBAL;
	if (a && !b)
		return SYMBOL_A;
	if (b && !a)
		return SYMBOL_B;

	/* Prefer a symbol with less underscores */
	a = prefix_underscores_count(syma->name);
	b = prefix_underscores_count(symb->name);
	if (b > a)
		return SYMBOL_A;
	else if (a > b)
		return SYMBOL_B;

124 125 126 127
	/* Choose the symbol with the longest name */
	na = strlen(syma->name);
	nb = strlen(symb->name);
	if (na > nb)
128
		return SYMBOL_A;
129
	else if (na < nb)
130
		return SYMBOL_B;
131 132 133 134 135 136 137 138

	/* Avoid "SyS" kernel syscall aliases */
	if (na >= 3 && !strncmp(syma->name, "SyS", 3))
		return SYMBOL_B;
	if (na >= 10 && !strncmp(syma->name, "compat_SyS", 10))
		return SYMBOL_B;

	return SYMBOL_A;
139 140
}

141
void symbols__fixup_duplicate(struct rb_root *symbols)
142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169
{
	struct rb_node *nd;
	struct symbol *curr, *next;

	nd = rb_first(symbols);

	while (nd) {
		curr = rb_entry(nd, struct symbol, rb_node);
again:
		nd = rb_next(&curr->rb_node);
		next = rb_entry(nd, struct symbol, rb_node);

		if (!nd)
			break;

		if (curr->start != next->start)
			continue;

		if (choose_best_symbol(curr, next) == SYMBOL_A) {
			rb_erase(&next->rb_node, symbols);
			goto again;
		} else {
			nd = rb_next(&curr->rb_node);
			rb_erase(&curr->rb_node, symbols);
		}
	}
}

170
void symbols__fixup_end(struct rb_root *symbols)
171
{
172
	struct rb_node *nd, *prevnd = rb_first(symbols);
173
	struct symbol *curr, *prev;
174 175 176 177

	if (prevnd == NULL)
		return;

178 179
	curr = rb_entry(prevnd, struct symbol, rb_node);

180
	for (nd = rb_next(prevnd); nd; nd = rb_next(nd)) {
181 182
		prev = curr;
		curr = rb_entry(nd, struct symbol, rb_node);
183

184
		if (prev->end == prev->start && prev->end != curr->start)
185 186
			prev->end = curr->start - 1;
	}
187 188 189 190

	/* Last entry */
	if (curr->end == curr->start)
		curr->end = roundup(curr->start, 4096);
191 192
}

193
void __map_groups__fixup_end(struct map_groups *mg, enum map_type type)
194 195
{
	struct map *prev, *curr;
196
	struct rb_node *nd, *prevnd = rb_first(&mg->maps[type]);
197 198 199 200 201 202 203 204 205 206

	if (prevnd == NULL)
		return;

	curr = rb_entry(prevnd, struct map, rb_node);

	for (nd = rb_next(prevnd); nd; nd = rb_next(nd)) {
		prev = curr;
		curr = rb_entry(nd, struct map, rb_node);
		prev->end = curr->start - 1;
207
	}
208 209 210 211 212

	/*
	 * We still haven't the actual symbols, so guess the
	 * last map final address.
	 */
213
	curr->end = ~0ULL;
214 215
}

216
struct symbol *symbol__new(u64 start, u64 len, u8 binding, const char *name)
217
{
218
	size_t namelen = strlen(name) + 1;
219 220 221
	struct symbol *sym = calloc(1, (symbol_conf.priv_size +
					sizeof(*sym) + namelen));
	if (sym == NULL)
222 223
		return NULL;

224
	if (symbol_conf.priv_size)
225
		sym = ((void *)sym) + symbol_conf.priv_size;
226

227 228 229 230
	sym->start   = start;
	sym->end     = len ? start + len - 1 : start;
	sym->binding = binding;
	sym->namelen = namelen - 1;
231

232 233 234
	pr_debug4("%s: %s %#" PRIx64 "-%#" PRIx64 "\n",
		  __func__, name, start, sym->end);
	memcpy(sym->name, name, namelen);
235

236
	return sym;
237 238
}

239
void symbol__delete(struct symbol *sym)
240
{
241
	free(((void *)sym) - symbol_conf.priv_size);
242 243
}

244
size_t symbol__fprintf(struct symbol *sym, FILE *fp)
245
{
246
	return fprintf(fp, " %" PRIx64 "-%" PRIx64 " %c %s\n",
247 248 249 250
		       sym->start, sym->end,
		       sym->binding == STB_GLOBAL ? 'g' :
		       sym->binding == STB_LOCAL  ? 'l' : 'w',
		       sym->name);
251 252
}

253 254
size_t symbol__fprintf_symname_offs(const struct symbol *sym,
				    const struct addr_location *al, FILE *fp)
255
{
256 257 258 259 260 261 262 263 264 265 266 267 268
	unsigned long offset;
	size_t length;

	if (sym && sym->name) {
		length = fprintf(fp, "%s", sym->name);
		if (al) {
			offset = al->addr - sym->start;
			length += fprintf(fp, "+0x%lx", offset);
		}
		return length;
	} else
		return fprintf(fp, "[unknown]");
}
269

270 271 272
size_t symbol__fprintf_symname(const struct symbol *sym, FILE *fp)
{
	return symbol__fprintf_symname_offs(sym, NULL, fp);
273 274
}

275
void symbols__delete(struct rb_root *symbols)
276 277
{
	struct symbol *pos;
278
	struct rb_node *next = rb_first(symbols);
279 280 281 282

	while (next) {
		pos = rb_entry(next, struct symbol, rb_node);
		next = rb_next(&pos->rb_node);
283
		rb_erase(&pos->rb_node, symbols);
284
		symbol__delete(pos);
285 286 287
	}
}

288
void symbols__insert(struct rb_root *symbols, struct symbol *sym)
289
{
290
	struct rb_node **p = &symbols->rb_node;
291
	struct rb_node *parent = NULL;
292
	const u64 ip = sym->start;
293 294 295 296 297 298 299 300 301 302 303
	struct symbol *s;

	while (*p != NULL) {
		parent = *p;
		s = rb_entry(parent, struct symbol, rb_node);
		if (ip < s->start)
			p = &(*p)->rb_left;
		else
			p = &(*p)->rb_right;
	}
	rb_link_node(&sym->rb_node, parent, p);
304
	rb_insert_color(&sym->rb_node, symbols);
305 306
}

307
static struct symbol *symbols__find(struct rb_root *symbols, u64 ip)
308 309 310
{
	struct rb_node *n;

311
	if (symbols == NULL)
312 313
		return NULL;

314
	n = symbols->rb_node;
315 316 317 318 319 320 321 322 323 324 325 326 327 328 329

	while (n) {
		struct symbol *s = rb_entry(n, struct symbol, rb_node);

		if (ip < s->start)
			n = n->rb_left;
		else if (ip > s->end)
			n = n->rb_right;
		else
			return s;
	}

	return NULL;
}

330 331 332 333 334
struct symbol_name_rb_node {
	struct rb_node	rb_node;
	struct symbol	sym;
};

335
static void symbols__insert_by_name(struct rb_root *symbols, struct symbol *sym)
336
{
337
	struct rb_node **p = &symbols->rb_node;
338
	struct rb_node *parent = NULL;
339 340 341
	struct symbol_name_rb_node *symn, *s;

	symn = container_of(sym, struct symbol_name_rb_node, sym);
342 343 344 345 346 347 348 349 350 351

	while (*p != NULL) {
		parent = *p;
		s = rb_entry(parent, struct symbol_name_rb_node, rb_node);
		if (strcmp(sym->name, s->sym.name) < 0)
			p = &(*p)->rb_left;
		else
			p = &(*p)->rb_right;
	}
	rb_link_node(&symn->rb_node, parent, p);
352
	rb_insert_color(&symn->rb_node, symbols);
353 354
}

355 356
static void symbols__sort_by_name(struct rb_root *symbols,
				  struct rb_root *source)
357 358 359 360 361
{
	struct rb_node *nd;

	for (nd = rb_first(source); nd; nd = rb_next(nd)) {
		struct symbol *pos = rb_entry(nd, struct symbol, rb_node);
362
		symbols__insert_by_name(symbols, pos);
363 364 365
	}
}

366 367
static struct symbol *symbols__find_by_name(struct rb_root *symbols,
					    const char *name)
368 369 370
{
	struct rb_node *n;

371
	if (symbols == NULL)
372 373
		return NULL;

374
	n = symbols->rb_node;
375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393

	while (n) {
		struct symbol_name_rb_node *s;
		int cmp;

		s = rb_entry(n, struct symbol_name_rb_node, rb_node);
		cmp = strcmp(name, s->sym.name);

		if (cmp < 0)
			n = n->rb_left;
		else if (cmp > 0)
			n = n->rb_right;
		else
			return &s->sym;
	}

	return NULL;
}

394
struct symbol *dso__find_symbol(struct dso *dso,
395
				enum map_type type, u64 addr)
396
{
397
	return symbols__find(&dso->symbols[type], addr);
398 399
}

400
struct symbol *dso__find_symbol_by_name(struct dso *dso, enum map_type type,
401 402
					const char *name)
{
403
	return symbols__find_by_name(&dso->symbol_names[type], name);
404 405
}

406
void dso__sort_by_name(struct dso *dso, enum map_type type)
407
{
408 409 410
	dso__set_sorted_by_name(dso, type);
	return symbols__sort_by_name(&dso->symbol_names[type],
				     &dso->symbols[type]);
411 412
}

413 414
size_t dso__fprintf_symbols_by_name(struct dso *dso,
				    enum map_type type, FILE *fp)
415 416 417 418 419
{
	size_t ret = 0;
	struct rb_node *nd;
	struct symbol_name_rb_node *pos;

420
	for (nd = rb_first(&dso->symbol_names[type]); nd; nd = rb_next(nd)) {
421 422 423 424 425 426 427
		pos = rb_entry(nd, struct symbol_name_rb_node, rb_node);
		fprintf(fp, "%s\n", pos->sym.name);
	}

	return ret;
}

428 429
int kallsyms__parse(const char *filename, void *arg,
		    int (*process_symbol)(void *arg, const char *name,
430
					  char type, u64 start))
431 432 433
{
	char *line = NULL;
	size_t n;
434
	int err = -1;
435
	FILE *file = fopen(filename, "r");
436 437 438 439

	if (file == NULL)
		goto out_failure;

440 441
	err = 0;

442
	while (!feof(file)) {
443
		u64 start;
444 445
		int line_len, len;
		char symbol_type;
446
		char *symbol_name;
447 448

		line_len = getline(&line, &n, file);
449
		if (line_len < 0 || !line)
450 451 452 453
			break;

		line[--line_len] = '\0'; /* \n */

454
		len = hex2u64(line, &start);
455 456 457 458 459

		len++;
		if (len + 2 >= line_len)
			continue;

460
		symbol_type = line[len];
461 462 463
		len += 2;
		symbol_name = line + len;
		len = line_len - len;
464

465 466
		if (len >= KSYM_NAME_LEN) {
			err = -1;
467
			break;
468 469
		}

470
		err = process_symbol(arg, symbol_name,
471
				     symbol_type, start);
472 473
		if (err)
			break;
474 475 476 477
	}

	free(line);
	fclose(file);
478
	return err;
479 480 481 482 483

out_failure:
	return -1;
}

484 485 486 487 488
struct process_kallsyms_args {
	struct map *map;
	struct dso *dso;
};

489 490 491 492 493 494 495 496
static u8 kallsyms2elf_type(char type)
{
	if (type == 'W')
		return STB_WEAK;

	return isupper(type) ? STB_GLOBAL : STB_LOCAL;
}

497
static int map__process_kallsym_symbol(void *arg, const char *name,
498
				       char type, u64 start)
499 500 501 502 503 504 505 506
{
	struct symbol *sym;
	struct process_kallsyms_args *a = arg;
	struct rb_root *root = &a->dso->symbols[a->map->type];

	if (!symbol_type__is_a(type, a->map->type))
		return 0;

507 508 509 510 511 512
	/*
	 * module symbols are not sorted so we add all
	 * symbols, setting length to 0, and rely on
	 * symbols__fixup_end() to fix it up.
	 */
	sym = symbol__new(start, 0, kallsyms2elf_type(type), name);
513 514 515 516 517 518 519
	if (sym == NULL)
		return -ENOMEM;
	/*
	 * We will pass the symbols to the filter later, in
	 * map__split_kallsyms, when we have split the maps per module
	 */
	symbols__insert(root, sym);
520

521 522 523 524 525 526 527 528
	return 0;
}

/*
 * Loads the function entries in /proc/kallsyms into kernel_map->dso,
 * so that we can in the next step set the symbol ->end address and then
 * call kernel_maps__split_kallsyms.
 */
529
static int dso__load_all_kallsyms(struct dso *dso, const char *filename,
530
				  struct map *map)
531
{
532
	struct process_kallsyms_args args = { .map = map, .dso = dso, };
533
	return kallsyms__parse(filename, &args, map__process_kallsym_symbol);
534 535
}

536 537 538 539 540
/*
 * Split the symbols into maps, making sure there are no overlaps, i.e. the
 * kernel range is broken in several maps, named [kernel].N, as we don't have
 * the original ELF section names vmlinux have.
 */
541
static int dso__split_kallsyms(struct dso *dso, struct map *map,
542
			       symbol_filter_t filter)
543
{
544
	struct map_groups *kmaps = map__kmap(map)->kmaps;
545
	struct machine *machine = kmaps->machine;
546
	struct map *curr_map = map;
547
	struct symbol *pos;
548
	int count = 0, moved = 0;	
549
	struct rb_root *root = &dso->symbols[map->type];
550
	struct rb_node *next = rb_first(root);
551 552 553 554 555 556 557 558 559 560
	int kernel_range = 0;

	while (next) {
		char *module;

		pos = rb_entry(next, struct symbol, rb_node);
		next = rb_next(&pos->rb_node);

		module = strchr(pos->name, '\t');
		if (module) {
561
			if (!symbol_conf.use_modules)
562 563
				goto discard_symbol;

564 565
			*module++ = '\0';

566
			if (strcmp(curr_map->dso->short_name, module)) {
567
				if (curr_map != map &&
568
				    dso->kernel == DSO_TYPE_GUEST_KERNEL &&
569
				    machine__is_default_guest(machine)) {
570 571 572 573 574 575 576 577 578 579 580 581 582
					/*
					 * We assume all symbols of a module are
					 * continuous in * kallsyms, so curr_map
					 * points to a module and all its
					 * symbols are in its kmap. Mark it as
					 * loaded.
					 */
					dso__set_loaded(curr_map->dso,
							curr_map->type);
				}

				curr_map = map_groups__find_by_name(kmaps,
							map->type, module);
583
				if (curr_map == NULL) {
584
					pr_debug("%s/proc/{kallsyms,modules} "
585
					         "inconsistency while looking "
586
						 "for \"%s\" module!\n",
587
						 machine->root_dir, module);
588 589
					curr_map = map;
					goto discard_symbol;
590
				}
591

592
				if (curr_map->dso->loaded &&
593
				    !machine__is_default_guest(machine))
594
					goto discard_symbol;
595
			}
596 597 598 599
			/*
			 * So that we look just like we get from .ko files,
			 * i.e. not prelinked, relative to map->start.
			 */
600 601 602
			pos->start = curr_map->map_ip(curr_map, pos->start);
			pos->end   = curr_map->map_ip(curr_map, pos->end);
		} else if (curr_map != map) {
603
			char dso_name[PATH_MAX];
604
			struct dso *ndso;
605

606 607 608 609 610
			if (count == 0) {
				curr_map = map;
				goto filter_symbol;
			}

611
			if (dso->kernel == DSO_TYPE_GUEST_KERNEL)
612 613 614 615 616 617 618
				snprintf(dso_name, sizeof(dso_name),
					"[guest.kernel].%d",
					kernel_range++);
			else
				snprintf(dso_name, sizeof(dso_name),
					"[kernel].%d",
					kernel_range++);
619

620 621
			ndso = dso__new(dso_name);
			if (ndso == NULL)
622 623
				return -1;

624
			ndso->kernel = dso->kernel;
625

626
			curr_map = map__new2(pos->start, ndso, map->type);
627
			if (curr_map == NULL) {
628
				dso__delete(ndso);
629 630
				return -1;
			}
631

632
			curr_map->map_ip = curr_map->unmap_ip = identity__map_ip;
633
			map_groups__insert(kmaps, curr_map);
634 635
			++kernel_range;
		}
636
filter_symbol:
637
		if (filter && filter(curr_map, pos)) {
638
discard_symbol:		rb_erase(&pos->rb_node, root);
639
			symbol__delete(pos);
640
		} else {
641 642 643
			if (curr_map != map) {
				rb_erase(&pos->rb_node, root);
				symbols__insert(&curr_map->dso->symbols[curr_map->type], pos);
644 645 646
				++moved;
			} else
				++count;
647
		}
648 649
	}

650
	if (curr_map != map &&
651
	    dso->kernel == DSO_TYPE_GUEST_KERNEL &&
652
	    machine__is_default_guest(kmaps->machine)) {
653 654 655
		dso__set_loaded(curr_map->dso, curr_map->type);
	}

656
	return count + moved;
657
}
658

659 660
bool symbol__restricted_filename(const char *filename,
				 const char *restricted_filename)
661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676
{
	bool restricted = false;

	if (symbol_conf.kptr_restrict) {
		char *r = realpath(filename, NULL);

		if (r != NULL) {
			restricted = strcmp(r, restricted_filename) == 0;
			free(r);
			return restricted;
		}
	}

	return restricted;
}

677
int dso__load_kallsyms(struct dso *dso, const char *filename,
678
		       struct map *map, symbol_filter_t filter)
679
{
680 681 682
	if (symbol__restricted_filename(filename, "/proc/kallsyms"))
		return -1;

683
	if (dso__load_all_kallsyms(dso, filename, map) < 0)
684 685
		return -1;

686
	symbols__fixup_duplicate(&dso->symbols[map->type]);
687 688
	symbols__fixup_end(&dso->symbols[map->type]);

689
	if (dso->kernel == DSO_TYPE_GUEST_KERNEL)
690
		dso->symtab_type = DSO_BINARY_TYPE__GUEST_KALLSYMS;
691
	else
692
		dso->symtab_type = DSO_BINARY_TYPE__KALLSYMS;
693

694
	return dso__split_kallsyms(dso, map, filter);
695 696
}

697
static int dso__load_perf_map(struct dso *dso, struct map *map,
698
			      symbol_filter_t filter)
699 700 701 702 703 704
{
	char *line = NULL;
	size_t n;
	FILE *file;
	int nr_syms = 0;

705
	file = fopen(dso->long_name, "r");
706 707 708 709
	if (file == NULL)
		goto out_failure;

	while (!feof(file)) {
710
		u64 start, size;
711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734
		struct symbol *sym;
		int line_len, len;

		line_len = getline(&line, &n, file);
		if (line_len < 0)
			break;

		if (!line)
			goto out_failure;

		line[--line_len] = '\0'; /* \n */

		len = hex2u64(line, &start);

		len++;
		if (len + 2 >= line_len)
			continue;

		len += hex2u64(line + len, &size);

		len++;
		if (len + 2 >= line_len)
			continue;

735
		sym = symbol__new(start, size, STB_GLOBAL, line + len);
736 737 738 739

		if (sym == NULL)
			goto out_delete_line;

740
		if (filter && filter(map, sym))
741
			symbol__delete(sym);
742
		else {
743
			symbols__insert(&dso->symbols[map->type], sym);
744 745 746 747 748 749 750 751 752 753 754 755 756 757 758
			nr_syms++;
		}
	}

	free(line);
	fclose(file);

	return nr_syms;

out_delete_line:
	free(line);
out_failure:
	return -1;
}

759
int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter)
760
{
761
	char *name;
762
	int ret = -1;
763
	u_int i;
764
	struct machine *machine;
765
	char *root_dir = (char *) "";
766 767 768
	int ss_pos = 0;
	struct symsrc ss_[2];
	struct symsrc *syms_ss = NULL, *runtime_ss = NULL;
769

770
	dso__set_loaded(dso, map->type);
771

772 773 774 775
	if (dso->kernel == DSO_TYPE_KERNEL)
		return dso__load_kernel_sym(dso, map, filter);
	else if (dso->kernel == DSO_TYPE_GUEST_KERNEL)
		return dso__load_guest_kernel_sym(dso, map, filter);
776

777 778
	if (map->groups && map->groups->machine)
		machine = map->groups->machine;
779
	else
780
		machine = NULL;
781

782
	dso->adjust_symbols = 0;
783

784
	if (strncmp(dso->name, "/tmp/perf-", 10) == 0) {
785 786
		struct stat st;

787
		if (lstat(dso->name, &st) < 0)
788 789 790 791 792 793 794 795
			return -1;

		if (st.st_uid && (st.st_uid != geteuid())) {
			pr_warning("File %s not owned by current user or root, "
				"ignoring it.\n", dso->name);
			return -1;
		}

796
		ret = dso__load_perf_map(dso, map, filter);
797 798
		dso->symtab_type = ret > 0 ? DSO_BINARY_TYPE__JAVA_JIT :
					     DSO_BINARY_TYPE__NOT_FOUND;
799 800 801
		return ret;
	}

802 803 804
	if (machine)
		root_dir = machine->root_dir;

805 806 807 808
	name = malloc(PATH_MAX);
	if (!name)
		return -1;

809
	/* Iterate over candidate debug images.
810 811
	 * Keep track of "interesting" ones (those which have a symtab, dynsym,
	 * and/or opd section) for processing.
812
	 */
813
	for (i = 0; i < DSO_BINARY_TYPE__SYMTAB_CNT; i++) {
814 815
		struct symsrc *ss = &ss_[ss_pos];
		bool next_slot = false;
816

817
		enum dso_binary_type symtab_type = binary_type_symtab[i];
818

819
		if (dso__binary_type_file(dso, symtab_type,
820 821
					  root_dir, name, PATH_MAX))
			continue;
822 823

		/* Name is now the name of the next image to try */
824
		if (symsrc__init(ss, dso, name, symtab_type) < 0)
825
			continue;
826

827 828 829
		if (!syms_ss && symsrc__has_symtab(ss)) {
			syms_ss = ss;
			next_slot = true;
830 831
		}

832 833 834
		if (!runtime_ss && symsrc__possibly_runtime(ss)) {
			runtime_ss = ss;
			next_slot = true;
835
		}
836

837 838
		if (next_slot) {
			ss_pos++;
839

840 841
			if (syms_ss && runtime_ss)
				break;
842
		}
843

844
	}
845

846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861
	if (!runtime_ss && !syms_ss)
		goto out_free;

	if (runtime_ss && !syms_ss) {
		syms_ss = runtime_ss;
	}

	/* We'll have to hope for the best */
	if (!runtime_ss && syms_ss)
		runtime_ss = syms_ss;

	if (syms_ss)
		ret = dso__load_sym(dso, map, syms_ss, runtime_ss, filter, 0);
	else
		ret = -1;

862
	if (ret > 0) {
863 864 865 866 867
		int nr_plt;

		nr_plt = dso__synthesize_plt_symbols(dso, runtime_ss, map, filter);
		if (nr_plt > 0)
			ret += nr_plt;
868 869
	}

870 871 872
	for (; ss_pos > 0; ss_pos--)
		symsrc__destroy(&ss_[ss_pos - 1]);
out_free:
873
	free(name);
874
	if (ret < 0 && strstr(dso->name, " (deleted)") != NULL)
875
		return 0;
876 877 878
	return ret;
}

879
struct map *map_groups__find_by_name(struct map_groups *mg,
880
				     enum map_type type, const char *name)
881 882 883
{
	struct rb_node *nd;

884
	for (nd = rb_first(&mg->maps[type]); nd; nd = rb_next(nd)) {
885 886
		struct map *map = rb_entry(nd, struct map, rb_node);

887
		if (map->dso && strcmp(map->dso->short_name, name) == 0)
888 889 890 891 892 893
			return map;
	}

	return NULL;
}

894
int dso__load_vmlinux(struct dso *dso, struct map *map,
895
		      const char *vmlinux, symbol_filter_t filter)
896
{
897 898
	int err = -1;
	struct symsrc ss;
899
	char symfs_vmlinux[PATH_MAX];
900
	enum dso_binary_type symtab_type;
901

902 903 904 905 906
	if (vmlinux[0] == '/')
		snprintf(symfs_vmlinux, sizeof(symfs_vmlinux), "%s", vmlinux);
	else
		snprintf(symfs_vmlinux, sizeof(symfs_vmlinux), "%s%s",
			 symbol_conf.symfs, vmlinux);
907

908
	if (dso->kernel == DSO_TYPE_GUEST_KERNEL)
909
		symtab_type = DSO_BINARY_TYPE__GUEST_VMLINUX;
910
	else
911
		symtab_type = DSO_BINARY_TYPE__VMLINUX;
912

913
	if (symsrc__init(&ss, dso, symfs_vmlinux, symtab_type))
914 915
		return -1;

916
	err = dso__load_sym(dso, map, &ss, &ss, filter, 0);
917
	symsrc__destroy(&ss);
918

919
	if (err > 0) {
920 921 922 923
		if (dso->kernel == DSO_TYPE_GUEST_KERNEL)
			dso->data_type = DSO_BINARY_TYPE__GUEST_VMLINUX;
		else
			dso->data_type = DSO_BINARY_TYPE__VMLINUX;
924 925
		dso__set_long_name(dso, (char *)vmlinux);
		dso__set_loaded(dso, map->type);
926
		pr_debug("Using %s for symbols\n", symfs_vmlinux);
927
	}
928

929 930 931
	return err;
}

932
int dso__load_vmlinux_path(struct dso *dso, struct map *map,
933
			   symbol_filter_t filter)
934 935
{
	int i, err = 0;
936
	char *filename;
937 938

	pr_debug("Looking at the vmlinux_path (%d entries long)\n",
939 940
		 vmlinux_path__nr_entries + 1);

941
	filename = dso__build_id_filename(dso, NULL, 0);
942
	if (filename != NULL) {
943
		err = dso__load_vmlinux(dso, map, filename, filter);
944 945
		if (err > 0) {
			dso->lname_alloc = 1;
946
			goto out;
947
		}
948 949
		free(filename);
	}
950 951

	for (i = 0; i < vmlinux_path__nr_entries; ++i) {
952
		err = dso__load_vmlinux(dso, map, vmlinux_path[i], filter);
953
		if (err > 0) {
954
			dso__set_long_name(dso, strdup(vmlinux_path[i]));
955
			dso->lname_alloc = 1;
956 957 958
			break;
		}
	}
959
out:
960 961 962
	return err;
}

963
static int dso__load_kernel_sym(struct dso *dso, struct map *map,
964
				symbol_filter_t filter)
965
{
966
	int err;
967 968
	const char *kallsyms_filename = NULL;
	char *kallsyms_allocated_filename = NULL;
969
	/*
970 971
	 * Step 1: if the user specified a kallsyms or vmlinux filename, use
	 * it and only it, reporting errors to the user if it cannot be used.
972 973 974 975 976 977 978 979 980 981 982 983
	 *
	 * For instance, try to analyse an ARM perf.data file _without_ a
	 * build-id, or if the user specifies the wrong path to the right
	 * vmlinux file, obviously we can't fallback to another vmlinux (a
	 * x86_86 one, on the machine where analysis is being performed, say),
	 * or worse, /proc/kallsyms.
	 *
	 * If the specified file _has_ a build-id and there is a build-id
	 * section in the perf.data file, we will still do the expected
	 * validation in dso__load_vmlinux and will bail out if they don't
	 * match.
	 */
984 985 986 987 988
	if (symbol_conf.kallsyms_name != NULL) {
		kallsyms_filename = symbol_conf.kallsyms_name;
		goto do_kallsyms;
	}

989
	if (symbol_conf.vmlinux_name != NULL) {
990
		err = dso__load_vmlinux(dso, map,
991
					symbol_conf.vmlinux_name, filter);
992
		if (err > 0) {
993
			dso__set_long_name(dso,
994
					   strdup(symbol_conf.vmlinux_name));
995
			dso->lname_alloc = 1;
996
			return err;
997 998
		}
		return err;
999
	}
1000 1001

	if (vmlinux_path != NULL) {
1002
		err = dso__load_vmlinux_path(dso, map, filter);
1003
		if (err > 0)
1004
			return err;
1005 1006
	}

1007 1008 1009 1010
	/* do not try local files if a symfs was given */
	if (symbol_conf.symfs[0] != 0)
		return -1;

1011 1012 1013 1014 1015
	/*
	 * Say the kernel DSO was created when processing the build-id header table,
	 * we have a build-id, so check if it is the same as the running kernel,
	 * using it if it is.
	 */
1016
	if (dso->has_build_id) {
1017
		u8 kallsyms_build_id[BUILD_ID_SIZE];
1018
		char sbuild_id[BUILD_ID_SIZE * 2 + 1];
1019 1020

		if (sysfs__read_build_id("/sys/kernel/notes", kallsyms_build_id,
1021
					 sizeof(kallsyms_build_id)) == 0) {
1022
			if (dso__build_id_equal(dso, kallsyms_build_id)) {
1023
				kallsyms_filename = "/proc/kallsyms";
1024
				goto do_kallsyms;
1025
			}
1026
		}
1027 1028 1029 1030
		/*
		 * Now look if we have it on the build-id cache in
		 * $HOME/.debug/[kernel.kallsyms].
		 */
1031
		build_id__sprintf(dso->build_id, sizeof(dso->build_id),
1032 1033 1034 1035
				  sbuild_id);

		if (asprintf(&kallsyms_allocated_filename,
			     "%s/.debug/[kernel.kallsyms]/%s",
1036 1037
			     getenv("HOME"), sbuild_id) == -1) {
			pr_err("Not enough memory for kallsyms file lookup\n");
1038
			return -1;
1039
		}
1040

1041 1042
		kallsyms_filename = kallsyms_allocated_filename;

1043
		if (access(kallsyms_filename, F_OK)) {
1044 1045
			pr_err("No kallsyms or vmlinux with build-id %s "
			       "was found\n", sbuild_id);
1046
			free(kallsyms_allocated_filename);
1047
			return -1;
1048
		}
1049 1050 1051 1052 1053
	} else {
		/*
		 * Last resort, if we don't have a build-id and couldn't find
		 * any vmlinux file, try the running kernel kallsyms table.
		 */
1054 1055
		kallsyms_filename = "/proc/kallsyms";
	}
1056

1057
do_kallsyms:
1058
	err = dso__load_kallsyms(dso, kallsyms_filename, map, filter);
1059 1060
	if (err > 0)
		pr_debug("Using %s for symbols\n", kallsyms_filename);
1061
	free(kallsyms_allocated_filename);
1062 1063

	if (err > 0) {
1064
		dso__set_long_name(dso, strdup("[kernel.kallsyms]"));
1065 1066
		map__fixup_start(map);
		map__fixup_end(map);
1067
	}
1068

1069 1070 1071
	return err;
}

1072 1073
static int dso__load_guest_kernel_sym(struct dso *dso, struct map *map,
				      symbol_filter_t filter)
1074 1075 1076
{
	int err;
	const char *kallsyms_filename = NULL;
1077
	struct machine *machine;
1078 1079 1080 1081 1082 1083
	char path[PATH_MAX];

	if (!map->groups) {
		pr_debug("Guest kernel map hasn't the point to groups\n");
		return -1;
	}
1084
	machine = map->groups->machine;
1085

1086
	if (machine__is_default_guest(machine)) {
1087 1088 1089 1090 1091 1092
		/*
		 * if the user specified a vmlinux filename, use it and only
		 * it, reporting errors to the user if it cannot be used.
		 * Or use file guest_kallsyms inputted by user on commandline
		 */
		if (symbol_conf.default_guest_vmlinux_name != NULL) {
1093
			err = dso__load_vmlinux(dso, map,
1094
				symbol_conf.default_guest_vmlinux_name, filter);
1095
			return err;
1096 1097 1098 1099 1100 1101
		}

		kallsyms_filename = symbol_conf.default_guest_kallsyms;
		if (!kallsyms_filename)
			return -1;
	} else {
1102
		sprintf(path, "%s/proc/kallsyms", machine->root_dir);
1103 1104 1105
		kallsyms_filename = path;
	}

1106
	err = dso__load_kallsyms(dso, kallsyms_filename, map, filter);
1107
	if (err > 0) {
1108 1109 1110
		pr_debug("Using %s for symbols\n", kallsyms_filename);
		machine__mmap_name(machine, path, sizeof(path));
		dso__set_long_name(dso, strdup(path));
1111 1112 1113 1114 1115 1116
		map__fixup_start(map);
		map__fixup_end(map);
	}

	return err;
}
1117

1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
static void vmlinux_path__exit(void)
{
	while (--vmlinux_path__nr_entries >= 0) {
		free(vmlinux_path[vmlinux_path__nr_entries]);
		vmlinux_path[vmlinux_path__nr_entries] = NULL;
	}

	free(vmlinux_path);
	vmlinux_path = NULL;
}

static int vmlinux_path__init(void)
{
	struct utsname uts;
	char bf[PATH_MAX];

	vmlinux_path = malloc(sizeof(char *) * 5);
	if (vmlinux_path == NULL)
		return -1;

	vmlinux_path[vmlinux_path__nr_entries] = strdup("vmlinux");
	if (vmlinux_path[vmlinux_path__nr_entries] == NULL)
		goto out_fail;
	++vmlinux_path__nr_entries;
	vmlinux_path[vmlinux_path__nr_entries] = strdup("/boot/vmlinux");
	if (vmlinux_path[vmlinux_path__nr_entries] == NULL)
		goto out_fail;
	++vmlinux_path__nr_entries;
1146 1147 1148 1149 1150 1151 1152 1153

	/* only try running kernel version if no symfs was given */
	if (symbol_conf.symfs[0] != 0)
		return 0;

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

1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177
	snprintf(bf, sizeof(bf), "/boot/vmlinux-%s", uts.release);
	vmlinux_path[vmlinux_path__nr_entries] = strdup(bf);
	if (vmlinux_path[vmlinux_path__nr_entries] == NULL)
		goto out_fail;
	++vmlinux_path__nr_entries;
	snprintf(bf, sizeof(bf), "/lib/modules/%s/build/vmlinux", uts.release);
	vmlinux_path[vmlinux_path__nr_entries] = strdup(bf);
	if (vmlinux_path[vmlinux_path__nr_entries] == NULL)
		goto out_fail;
	++vmlinux_path__nr_entries;
	snprintf(bf, sizeof(bf), "/usr/lib/debug/lib/modules/%s/vmlinux",
		 uts.release);
	vmlinux_path[vmlinux_path__nr_entries] = strdup(bf);
	if (vmlinux_path[vmlinux_path__nr_entries] == NULL)
		goto out_fail;
	++vmlinux_path__nr_entries;

	return 0;

out_fail:
	vmlinux_path__exit();
	return -1;
}

1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191
static int setup_list(struct strlist **list, const char *list_str,
		      const char *list_name)
{
	if (list_str == NULL)
		return 0;

	*list = strlist__new(true, list_str);
	if (!*list) {
		pr_err("problems parsing %s list\n", list_name);
		return -1;
	}
	return 0;
}

1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210
static bool symbol__read_kptr_restrict(void)
{
	bool value = false;

	if (geteuid() != 0) {
		FILE *fp = fopen("/proc/sys/kernel/kptr_restrict", "r");
		if (fp != NULL) {
			char line[8];

			if (fgets(line, sizeof(line), fp) != NULL)
				value = atoi(line) != 0;

			fclose(fp);
		}
	}

	return value;
}

1211
int symbol__init(void)
1212
{
1213 1214
	const char *symfs;

1215 1216 1217
	if (symbol_conf.initialized)
		return 0;

1218
	symbol_conf.priv_size = PERF_ALIGN(symbol_conf.priv_size, sizeof(u64));
1219

1220 1221
	symbol__elf_init();

1222 1223 1224
	if (symbol_conf.sort_by_name)
		symbol_conf.priv_size += (sizeof(struct symbol_name_rb_node) -
					  sizeof(struct symbol));
1225

1226
	if (symbol_conf.try_vmlinux_path && vmlinux_path__init() < 0)
1227 1228
		return -1;

1229 1230 1231 1232 1233
	if (symbol_conf.field_sep && *symbol_conf.field_sep == '.') {
		pr_err("'.' is the only non valid --field-separator argument\n");
		return -1;
	}

1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245
	if (setup_list(&symbol_conf.dso_list,
		       symbol_conf.dso_list_str, "dso") < 0)
		return -1;

	if (setup_list(&symbol_conf.comm_list,
		       symbol_conf.comm_list_str, "comm") < 0)
		goto out_free_dso_list;

	if (setup_list(&symbol_conf.sym_list,
		       symbol_conf.sym_list_str, "symbol") < 0)
		goto out_free_comm_list;

1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257
	/*
	 * A path to symbols of "/" is identical to ""
	 * reset here for simplicity.
	 */
	symfs = realpath(symbol_conf.symfs, NULL);
	if (symfs == NULL)
		symfs = symbol_conf.symfs;
	if (strcmp(symfs, "/") == 0)
		symbol_conf.symfs = "";
	if (symfs != symbol_conf.symfs)
		free((void *)symfs);

1258 1259
	symbol_conf.kptr_restrict = symbol__read_kptr_restrict();

1260
	symbol_conf.initialized = true;
1261
	return 0;
1262 1263 1264

out_free_comm_list:
	strlist__delete(symbol_conf.comm_list);
1265 1266
out_free_dso_list:
	strlist__delete(symbol_conf.dso_list);
1267
	return -1;
1268 1269
}

1270 1271
void symbol__exit(void)
{
1272 1273
	if (!symbol_conf.initialized)
		return;
1274 1275 1276 1277 1278
	strlist__delete(symbol_conf.sym_list);
	strlist__delete(symbol_conf.dso_list);
	strlist__delete(symbol_conf.comm_list);
	vmlinux_path__exit();
	symbol_conf.sym_list = symbol_conf.dso_list = symbol_conf.comm_list = NULL;
1279
	symbol_conf.initialized = false;
1280
}