2005-05-01 Paul Brook <paul@codesourcery.com>
[binutils.git] / gprof / corefile.c
blobbd594f5a23b6d6a88f011095ec43d5057b37c297
1 /* corefile.c
3 Copyright 1999, 2000, 2001, 2002, 2003, 2004
4 Free Software Foundation, Inc.
6 This file is part of GNU Binutils.
8 This program is free software; you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation; either version 2 of the License, or
11 (at your option) any later version.
13 This program is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with this program; if not, write to the Free Software
20 Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
21 02111-1307, USA. */
23 #include "libiberty.h"
24 #include "gprof.h"
25 #include "search_list.h"
26 #include "source.h"
27 #include "symtab.h"
28 #include "corefile.h"
30 bfd *core_bfd;
31 int core_num_syms;
32 asymbol **core_syms;
33 asection *core_text_sect;
34 PTR core_text_space;
36 int min_insn_size;
37 int offset_to_code;
39 /* For mapping symbols to specific .o files during file ordering. */
40 struct function_map *symbol_map;
41 unsigned int symbol_map_count;
43 static void read_function_mappings (const char *);
44 static int core_sym_class (asymbol *);
45 static bfd_boolean get_src_info
46 (bfd_vma, const char **, const char **, int *);
48 extern void i386_find_call (Sym *, bfd_vma, bfd_vma);
49 extern void alpha_find_call (Sym *, bfd_vma, bfd_vma);
50 extern void vax_find_call (Sym *, bfd_vma, bfd_vma);
51 extern void tahoe_find_call (Sym *, bfd_vma, bfd_vma);
52 extern void sparc_find_call (Sym *, bfd_vma, bfd_vma);
53 extern void mips_find_call (Sym *, bfd_vma, bfd_vma);
55 static void
56 read_function_mappings (const char *filename)
58 FILE *file = fopen (filename, "r");
59 char dummy[1024];
60 int count = 0;
62 if (!file)
64 fprintf (stderr, _("%s: could not open %s.\n"), whoami, filename);
65 done (1);
68 /* First parse the mapping file so we know how big we need to
69 make our tables. We also do some sanity checks at this
70 time. */
71 while (!feof (file))
73 int matches;
75 matches = fscanf (file, "%[^\n:]", dummy);
76 if (!matches)
78 fprintf (stderr, _("%s: unable to parse mapping file %s.\n"),
79 whoami, filename);
80 done (1);
83 /* Just skip messages about files with no symbols. */
84 if (!strncmp (dummy, "No symbols in ", 14))
86 fscanf (file, "\n");
87 continue;
90 /* Don't care what else is on this line at this point. */
91 fscanf (file, "%[^\n]\n", dummy);
92 count++;
95 /* Now we know how big we need to make our table. */
96 symbol_map = ((struct function_map *)
97 xmalloc (count * sizeof (struct function_map)));
99 /* Rewind the input file so we can read it again. */
100 rewind (file);
102 /* Read each entry and put it into the table. */
103 count = 0;
104 while (!feof (file))
106 int matches;
107 char *tmp;
109 matches = fscanf (file, "%[^\n:]", dummy);
110 if (!matches)
112 fprintf (stderr, _("%s: unable to parse mapping file %s.\n"),
113 whoami, filename);
114 done (1);
117 /* Just skip messages about files with no symbols. */
118 if (!strncmp (dummy, "No symbols in ", 14))
120 fscanf (file, "\n");
121 continue;
124 /* dummy has the filename, go ahead and copy it. */
125 symbol_map[count].file_name = xmalloc (strlen (dummy) + 1);
126 strcpy (symbol_map[count].file_name, dummy);
128 /* Now we need the function name. */
129 fscanf (file, "%[^\n]\n", dummy);
130 tmp = strrchr (dummy, ' ') + 1;
131 symbol_map[count].function_name = xmalloc (strlen (tmp) + 1);
132 strcpy (symbol_map[count].function_name, tmp);
133 count++;
136 /* Record the size of the map table for future reference. */
137 symbol_map_count = count;
141 void
142 core_init (const char *aout_name)
144 int core_sym_bytes;
145 core_bfd = bfd_openr (aout_name, 0);
147 if (!core_bfd)
149 perror (aout_name);
150 done (1);
153 if (!bfd_check_format (core_bfd, bfd_object))
155 fprintf (stderr, _("%s: %s: not in a.out format\n"), whoami, aout_name);
156 done (1);
159 /* Get core's text section. */
160 core_text_sect = bfd_get_section_by_name (core_bfd, ".text");
161 if (!core_text_sect)
163 core_text_sect = bfd_get_section_by_name (core_bfd, "$CODE$");
164 if (!core_text_sect)
166 fprintf (stderr, _("%s: can't find .text section in %s\n"),
167 whoami, aout_name);
168 done (1);
172 /* Read core's symbol table. */
174 /* This will probably give us more than we need, but that's ok. */
175 core_sym_bytes = bfd_get_symtab_upper_bound (core_bfd);
176 if (core_sym_bytes < 0)
178 fprintf (stderr, "%s: %s: %s\n", whoami, aout_name,
179 bfd_errmsg (bfd_get_error ()));
180 done (1);
183 core_syms = (asymbol **) xmalloc (core_sym_bytes);
184 core_num_syms = bfd_canonicalize_symtab (core_bfd, core_syms);
186 if (core_num_syms < 0)
188 fprintf (stderr, "%s: %s: %s\n", whoami, aout_name,
189 bfd_errmsg (bfd_get_error ()));
190 done (1);
193 min_insn_size = 1;
194 offset_to_code = 0;
196 switch (bfd_get_arch (core_bfd))
198 case bfd_arch_vax:
199 case bfd_arch_tahoe:
200 offset_to_code = 2;
201 break;
203 case bfd_arch_alpha:
204 min_insn_size = 4;
205 break;
207 default:
208 break;
211 if (function_mapping_file)
212 read_function_mappings (function_mapping_file);
215 /* Read in the text space of an a.out file. */
217 void
218 core_get_text_space (bfd *cbfd)
220 core_text_space = malloc (bfd_get_section_size (core_text_sect));
222 if (!core_text_space)
224 fprintf (stderr, _("%s: ran out room for %lu bytes of text space\n"),
225 whoami, (unsigned long) bfd_get_section_size (core_text_sect));
226 done (1);
229 if (!bfd_get_section_contents (cbfd, core_text_sect, core_text_space,
230 0, bfd_get_section_size (core_text_sect)))
232 bfd_perror ("bfd_get_section_contents");
233 free (core_text_space);
234 core_text_space = 0;
237 if (!core_text_space)
238 fprintf (stderr, _("%s: can't do -c\n"), whoami);
242 void
243 find_call (Sym *parent, bfd_vma p_lowpc, bfd_vma p_highpc)
245 switch (bfd_get_arch (core_bfd))
247 case bfd_arch_i386:
248 i386_find_call (parent, p_lowpc, p_highpc);
249 break;
251 case bfd_arch_alpha:
252 alpha_find_call (parent, p_lowpc, p_highpc);
253 break;
255 case bfd_arch_vax:
256 vax_find_call (parent, p_lowpc, p_highpc);
257 break;
259 case bfd_arch_sparc:
260 sparc_find_call (parent, p_lowpc, p_highpc);
261 break;
263 case bfd_arch_tahoe:
264 tahoe_find_call (parent, p_lowpc, p_highpc);
265 break;
267 case bfd_arch_mips:
268 mips_find_call (parent, p_lowpc, p_highpc);
269 break;
271 default:
272 fprintf (stderr, _("%s: -c not supported on architecture %s\n"),
273 whoami, bfd_printable_name(core_bfd));
275 /* Don't give the error more than once. */
276 ignore_direct_calls = FALSE;
280 /* Return class of symbol SYM. The returned class can be any of:
281 0 -> symbol is not interesting to us
282 'T' -> symbol is a global name
283 't' -> symbol is a local (static) name. */
285 static int
286 core_sym_class (asymbol *sym)
288 symbol_info syminfo;
289 const char *name;
290 char sym_prefix;
291 int i;
293 if (sym->section == NULL || (sym->flags & BSF_DEBUGGING) != 0)
294 return 0;
296 /* Must be a text symbol, and static text symbols
297 don't qualify if ignore_static_funcs set. */
298 if (ignore_static_funcs && (sym->flags & BSF_LOCAL))
300 DBG (AOUTDEBUG, printf ("[core_sym_class] %s: not a function\n",
301 sym->name));
302 return 0;
305 bfd_get_symbol_info (core_bfd, sym, &syminfo);
306 i = syminfo.type;
308 if (i == 'T')
309 return i; /* It's a global symbol. */
311 if (i == 'W')
312 /* Treat weak symbols as text symbols. FIXME: a weak symbol may
313 also be a data symbol. */
314 return 'T';
316 if (i != 't')
318 /* Not a static text symbol. */
319 DBG (AOUTDEBUG, printf ("[core_sym_class] %s is of class %c\n",
320 sym->name, i));
321 return 0;
324 /* Do some more filtering on static function-names. */
325 if (ignore_static_funcs)
326 return 0;
328 /* Can't zero-length name or funny characters in name, where
329 `funny' includes: `.' (.o file names) and `$' (Pascal labels). */
330 if (!sym->name || sym->name[0] == '\0')
331 return 0;
333 for (name = sym->name; *name; ++name)
335 if (*name == '.' || *name == '$')
336 return 0;
339 /* On systems where the C compiler adds an underscore to all
340 names, static names without underscores seem usually to be
341 labels in hand written assembler in the library. We don't want
342 these names. This is certainly necessary on a Sparc running
343 SunOS 4.1 (try profiling a program that does a lot of
344 division). I don't know whether it has harmful side effects on
345 other systems. Perhaps it should be made configurable. */
346 sym_prefix = bfd_get_symbol_leading_char (core_bfd);
348 if ((sym_prefix && sym_prefix != sym->name[0])
349 /* GCC may add special symbols to help gdb figure out the file
350 language. We want to ignore these, since sometimes they mask
351 the real function. (dj@ctron) */
352 || !strncmp (sym->name, "__gnu_compiled", 14)
353 || !strncmp (sym->name, "___gnu_compiled", 15))
355 return 0;
358 /* If the object file supports marking of function symbols, then
359 we can zap anything that doesn't have BSF_FUNCTION set. */
360 if (ignore_non_functions && (sym->flags & BSF_FUNCTION) == 0)
361 return 0;
363 return 't'; /* It's a static text symbol. */
366 /* Get whatever source info we can get regarding address ADDR. */
368 static bfd_boolean
369 get_src_info (bfd_vma addr, const char **filename, const char **name, int *line_num)
371 const char *fname = 0, *func_name = 0;
372 int l = 0;
374 if (bfd_find_nearest_line (core_bfd, core_text_sect, core_syms,
375 addr - core_text_sect->vma,
376 &fname, &func_name, (unsigned int *) &l)
377 && fname && func_name && l)
379 DBG (AOUTDEBUG, printf ("[get_src_info] 0x%lx -> %s:%d (%s)\n",
380 (unsigned long) addr, fname, l, func_name));
381 *filename = fname;
382 *name = func_name;
383 *line_num = l;
384 return TRUE;
386 else
388 DBG (AOUTDEBUG, printf ("[get_src_info] no info for 0x%lx (%s:%d,%s)\n",
389 (long) addr, fname ? fname : "<unknown>", l,
390 func_name ? func_name : "<unknown>"));
391 return FALSE;
395 /* Read in symbol table from core.
396 One symbol per function is entered. */
398 void
399 core_create_function_syms ()
401 bfd_vma min_vma = ~(bfd_vma) 0;
402 bfd_vma max_vma = 0;
403 int class;
404 long i, found, skip;
405 unsigned int j;
407 /* Pass 1 - determine upper bound on number of function names. */
408 symtab.len = 0;
410 for (i = 0; i < core_num_syms; ++i)
412 if (!core_sym_class (core_syms[i]))
413 continue;
415 /* This should be replaced with a binary search or hashed
416 search. Gross.
418 Don't create a symtab entry for a function that has
419 a mapping to a file, unless it's the first function
420 in the file. */
421 skip = 0;
422 for (j = 0; j < symbol_map_count; j++)
423 if (!strcmp (core_syms[i]->name, symbol_map[j].function_name))
425 if (j > 0 && ! strcmp (symbol_map [j].file_name,
426 symbol_map [j - 1].file_name))
427 skip = 1;
428 break;
431 if (!skip)
432 ++symtab.len;
435 if (symtab.len == 0)
437 fprintf (stderr, _("%s: file `%s' has no symbols\n"), whoami, a_out_name);
438 done (1);
441 /* The "+ 2" is for the sentinels. */
442 symtab.base = (Sym *) xmalloc ((symtab.len + 2) * sizeof (Sym));
444 /* Pass 2 - create symbols. */
445 symtab.limit = symtab.base;
447 for (i = 0; i < core_num_syms; ++i)
449 asection *sym_sec;
451 class = core_sym_class (core_syms[i]);
453 if (!class)
455 DBG (AOUTDEBUG,
456 printf ("[core_create_function_syms] rejecting: 0x%lx %s\n",
457 (unsigned long) core_syms[i]->value,
458 core_syms[i]->name));
459 continue;
462 /* This should be replaced with a binary search or hashed
463 search. Gross. */
464 skip = 0;
465 found = 0;
467 for (j = 0; j < symbol_map_count; j++)
468 if (!strcmp (core_syms[i]->name, symbol_map[j].function_name))
470 if (j > 0 && ! strcmp (symbol_map [j].file_name,
471 symbol_map [j - 1].file_name))
472 skip = 1;
473 else
474 found = j;
475 break;
478 if (skip)
479 continue;
481 sym_init (symtab.limit);
483 /* Symbol offsets are always section-relative. */
484 sym_sec = core_syms[i]->section;
485 symtab.limit->addr = core_syms[i]->value;
486 if (sym_sec)
487 symtab.limit->addr += bfd_get_section_vma (sym_sec->owner, sym_sec);
489 if (symbol_map_count
490 && !strcmp (core_syms[i]->name, symbol_map[found].function_name))
492 symtab.limit->name = symbol_map[found].file_name;
493 symtab.limit->mapped = 1;
495 else
497 symtab.limit->name = core_syms[i]->name;
498 symtab.limit->mapped = 0;
501 /* Lookup filename and line number, if we can. */
503 const char *filename, *func_name;
505 if (get_src_info (symtab.limit->addr, &filename, &func_name,
506 &symtab.limit->line_num))
508 symtab.limit->file = source_file_lookup_path (filename);
510 /* FIXME: Checking __osf__ here does not work with a cross
511 gprof. */
512 #ifdef __osf__
513 /* Suppress symbols that are not function names. This is
514 useful to suppress code-labels and aliases.
516 This is known to be useful under DEC's OSF/1. Under SunOS 4.x,
517 labels do not appear in the symbol table info, so this isn't
518 necessary. */
520 if (strcmp (symtab.limit->name, func_name) != 0)
522 /* The symbol's address maps to a different name, so
523 it can't be a function-entry point. This happens
524 for labels, for example. */
525 DBG (AOUTDEBUG,
526 printf ("[core_create_function_syms: rej %s (maps to %s)\n",
527 symtab.limit->name, func_name));
528 continue;
530 #endif
534 symtab.limit->is_func = TRUE;
535 symtab.limit->is_bb_head = TRUE;
537 if (class == 't')
538 symtab.limit->is_static = TRUE;
540 /* Keep track of the minimum and maximum vma addresses used by all
541 symbols. When computing the max_vma, use the ending address of the
542 section containing the symbol, if available. */
543 min_vma = MIN (symtab.limit->addr, min_vma);
544 if (sym_sec)
545 max_vma = MAX (bfd_get_section_vma (sym_sec->owner, sym_sec)
546 + bfd_section_size (sym_sec->owner, sym_sec) - 1,
547 max_vma);
548 else
549 max_vma = MAX (symtab.limit->addr, max_vma);
551 /* If we see "main" without an initial '_', we assume names
552 are *not* prefixed by '_'. */
553 if (symtab.limit->name[0] == 'm' && discard_underscores
554 && strcmp (symtab.limit->name, "main") == 0)
555 discard_underscores = 0;
557 DBG (AOUTDEBUG, printf ("[core_create_function_syms] %ld %s 0x%lx\n",
558 (long) (symtab.limit - symtab.base),
559 symtab.limit->name,
560 (unsigned long) symtab.limit->addr));
561 ++symtab.limit;
564 /* Create sentinels. */
565 sym_init (symtab.limit);
566 symtab.limit->name = "<locore>";
567 symtab.limit->addr = 0;
568 symtab.limit->end_addr = min_vma - 1;
569 ++symtab.limit;
571 sym_init (symtab.limit);
572 symtab.limit->name = "<hicore>";
573 symtab.limit->addr = max_vma + 1;
574 symtab.limit->end_addr = ~(bfd_vma) 0;
575 ++symtab.limit;
577 symtab.len = symtab.limit - symtab.base;
578 symtab_finalize (&symtab);
581 /* Read in symbol table from core.
582 One symbol per line of source code is entered. */
584 void
585 core_create_line_syms ()
587 char *prev_name, *prev_filename;
588 unsigned int prev_name_len, prev_filename_len;
589 bfd_vma vma, min_vma = ~(bfd_vma) 0, max_vma = 0;
590 Sym *prev, dummy, *sentinel, *sym;
591 const char *filename;
592 int prev_line_num;
593 Sym_Table ltab;
594 bfd_vma vma_high;
596 /* Create symbols for functions as usual. This is necessary in
597 cases where parts of a program were not compiled with -g. For
598 those parts we still want to get info at the function level. */
599 core_create_function_syms ();
601 /* Pass 1: count the number of symbols. */
603 /* To find all line information, walk through all possible
604 text-space addresses (one by one!) and get the debugging
605 info for each address. When the debugging info changes,
606 it is time to create a new symbol.
608 Of course, this is rather slow and it would be better if
609 BFD would provide an iterator for enumerating all line infos. */
610 prev_name_len = PATH_MAX;
611 prev_filename_len = PATH_MAX;
612 prev_name = xmalloc (prev_name_len);
613 prev_filename = xmalloc (prev_filename_len);
614 ltab.len = 0;
615 prev_line_num = 0;
617 vma_high = core_text_sect->vma + bfd_get_section_size (core_text_sect);
618 for (vma = core_text_sect->vma; vma < vma_high; vma += min_insn_size)
620 unsigned int len;
622 if (!get_src_info (vma, &filename, &dummy.name, &dummy.line_num)
623 || (prev_line_num == dummy.line_num
624 && prev_name != NULL
625 && strcmp (prev_name, dummy.name) == 0
626 && strcmp (prev_filename, filename) == 0))
627 continue;
629 ++ltab.len;
630 prev_line_num = dummy.line_num;
632 len = strlen (dummy.name);
633 if (len >= prev_name_len)
635 prev_name_len = len + 1024;
636 free (prev_name);
637 prev_name = xmalloc (prev_name_len);
640 strcpy (prev_name, dummy.name);
641 len = strlen (filename);
643 if (len >= prev_filename_len)
645 prev_filename_len = len + 1024;
646 free (prev_filename);
647 prev_filename = xmalloc (prev_filename_len);
650 strcpy (prev_filename, filename);
652 min_vma = MIN (vma, min_vma);
653 max_vma = MAX (vma, max_vma);
656 free (prev_name);
657 free (prev_filename);
659 /* Make room for function symbols, too. */
660 ltab.len += symtab.len;
661 ltab.base = (Sym *) xmalloc (ltab.len * sizeof (Sym));
662 ltab.limit = ltab.base;
664 /* Pass 2 - create symbols. */
666 /* We now set is_static as we go along, rather than by running
667 through the symbol table at the end.
669 The old way called symtab_finalize before the is_static pass,
670 causing a problem since symtab_finalize uses is_static as part of
671 its address conflict resolution algorithm. Since global symbols
672 were prefered over static symbols, and all line symbols were
673 global at that point, static function names that conflicted with
674 their own line numbers (static, but labeled as global) were
675 rejected in favor of the line num.
677 This was not the desired functionality. We always want to keep
678 our function symbols and discard any conflicting line symbols.
679 Perhaps symtab_finalize should be modified to make this
680 distinction as well, but the current fix works and the code is a
681 lot cleaner now. */
682 prev = 0;
684 for (vma = core_text_sect->vma; vma < vma_high; vma += min_insn_size)
686 sym_init (ltab.limit);
688 if (!get_src_info (vma, &filename, &ltab.limit->name, &ltab.limit->line_num)
689 || (prev && prev->line_num == ltab.limit->line_num
690 && strcmp (prev->name, ltab.limit->name) == 0
691 && strcmp (prev->file->name, filename) == 0))
692 continue;
694 /* Make name pointer a malloc'ed string. */
695 ltab.limit->name = xstrdup (ltab.limit->name);
696 ltab.limit->file = source_file_lookup_path (filename);
698 ltab.limit->addr = vma;
700 /* Set is_static based on the enclosing function, using either:
701 1) the previous symbol, if it's from the same function, or
702 2) a symtab lookup. */
703 if (prev && ltab.limit->file == prev->file &&
704 strcmp (ltab.limit->name, prev->name) == 0)
706 ltab.limit->is_static = prev->is_static;
708 else
710 sym = sym_lookup(&symtab, ltab.limit->addr);
711 ltab.limit->is_static = sym->is_static;
714 prev = ltab.limit;
716 /* If we see "main" without an initial '_', we assume names
717 are *not* prefixed by '_'. */
718 if (ltab.limit->name[0] == 'm' && discard_underscores
719 && strcmp (ltab.limit->name, "main") == 0)
720 discard_underscores = 0;
722 DBG (AOUTDEBUG, printf ("[core_create_line_syms] %lu %s 0x%lx\n",
723 (unsigned long) (ltab.limit - ltab.base),
724 ltab.limit->name,
725 (unsigned long) ltab.limit->addr));
726 ++ltab.limit;
729 /* Update sentinels. */
730 sentinel = sym_lookup (&symtab, (bfd_vma) 0);
732 if (sentinel
733 && strcmp (sentinel->name, "<locore>") == 0
734 && min_vma <= sentinel->end_addr)
735 sentinel->end_addr = min_vma - 1;
737 sentinel = sym_lookup (&symtab, ~(bfd_vma) 0);
739 if (sentinel
740 && strcmp (sentinel->name, "<hicore>") == 0
741 && max_vma >= sentinel->addr)
742 sentinel->addr = max_vma + 1;
744 /* Copy in function symbols. */
745 memcpy (ltab.limit, symtab.base, symtab.len * sizeof (Sym));
746 ltab.limit += symtab.len;
748 if ((unsigned int) (ltab.limit - ltab.base) != ltab.len)
750 fprintf (stderr,
751 _("%s: somebody miscounted: ltab.len=%d instead of %ld\n"),
752 whoami, ltab.len, (long) (ltab.limit - ltab.base));
753 done (1);
756 /* Finalize ltab and make it symbol table. */
757 symtab_finalize (&ltab);
758 free (symtab.base);
759 symtab = ltab;