bfd/syms.c

   1 /* Generic symbol-table support for the BFD library.
   2    Copyright (C) 1990, 1991, 1992, 1993 Free Software Foundation, Inc.
   3    Written by Cygnus Support.
   4
   5 This file is part of BFD, the Binary File Descriptor library.
   6
   7 This program is free software; you can redistribute it and/or modify
   8 it under the terms of the GNU General Public License as published by
   9 the Free Software Foundation; either version 2 of the License, or
  10 (at your option) any later version.
  11
  12 This program is distributed in the hope that it will be useful,
  13 but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15 GNU General Public License for more details.
  16
  17 You should have received a copy of the GNU General Public License
  18 along with this program; if not, write to the Free Software
  19 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.  */
  20
  21 /*
  22 SECTION
  23         Symbols
  24
  25         BFD tries to maintain as much symbol information as it can when
  26         it moves information from file to file. BFD passes information
  27         to applications though the <<asymbol>> structure. When the
  28         application requests the symbol table, BFD reads the table in
  29         the native form and translates parts of it into the internal
  30         format. To maintain more than the information passed to
  31         applications, some targets keep some information ``behind the
  32         scenes'' in a structure only the particular back end knows
  33         about. For example, the coff back end keeps the original
  34         symbol table structure as well as the canonical structure when
  35         a BFD is read in. On output, the coff back end can reconstruct
  36         the output symbol table so that no information is lost, even
  37         information unique to coff which BFD doesn't know or
  38         understand. If a coff symbol table were read, but were written
  39         through an a.out back end, all the coff specific information
  40         would be lost. The symbol table of a BFD
  41         is not necessarily read in until a canonicalize request is
  42         made. Then the BFD back end fills in a table provided by the
  43         application with pointers to the canonical information.  To
  44         output symbols, the application provides BFD with a table of
  45         pointers to pointers to <<asymbol>>s. This allows applications
  46         like the linker to output a symbol as it was read, since the ``behind
  47         the scenes'' information will be still available.
  48 @menu
  49 @* Reading Symbols::
  50 @* Writing Symbols::
  51 @* typedef asymbol::
  52 @* symbol handling functions::
  53 @end menu
  54
  55 INODE
  56 Reading Symbols, Writing Symbols, Symbols, Symbols
  57 SUBSECTION
  58         Reading Symbols
  59
  60         There are two stages to reading a symbol table from a BFD:
  61         allocating storage, and the actual reading process. This is an
  62         excerpt from an application which reads the symbol table:
  63
  64 |         unsigned int storage_needed;
  65 |         asymbol **symbol_table;
  66 |         unsigned int number_of_symbols;
  67 |         unsigned int i;
  68 |
  69 |         storage_needed = get_symtab_upper_bound (abfd);
  70 |
  71 |         if (storage_needed == 0) {
  72 |            return ;
  73 |         }
  74 |         symbol_table = (asymbol **) bfd_xmalloc (storage_needed);
  75 |           ...
  76 |         number_of_symbols =
  77 |            bfd_canonicalize_symtab (abfd, symbol_table);
  78 |
  79 |         for (i = 0; i < number_of_symbols; i++) {
  80 |            process_symbol (symbol_table[i]);
  81 |         }
  82
  83         All storage for the symbols themselves is in an obstack
  84         connected to the BFD; it is freed when the BFD is closed.
  85
  86
  87 INODE
  88 Writing Symbols, typedef asymbol, Reading Symbols, Symbols
  89 SUBSECTION
  90         Writing Symbols
  91
  92         Writing of a symbol table is automatic when a BFD open for
  93         writing is closed. The application attaches a vector of
  94         pointers to pointers to symbols to the BFD being written, and
  95         fills in the symbol count. The close and cleanup code reads
  96         through the table provided and performs all the necessary
  97         operations. The BFD output code must always be provided with an
  98         ``owned'' symbol: one which has come from another BFD, or one
  99         which has been created using <<bfd_make_empty_symbol>>.  Here is an
 100         example showing the creation of a symbol table with only one element:
 101
 102 |       #include "bfd.h"
 103 |       main()
 104 |       {
 105 |         bfd *abfd;
 106 |         asymbol *ptrs[2];
 107 |         asymbol *new;
 108 |
 109 |         abfd = bfd_openw("foo","a.out-sunos-big");
 110 |         bfd_set_format(abfd, bfd_object);
 111 |         new = bfd_make_empty_symbol(abfd);
 112 |         new->name = "dummy_symbol";
 113 |         new->section = bfd_make_section_old_way(abfd, ".text");
 114 |         new->flags = BSF_GLOBAL;
 115 |         new->value = 0x12345;
 116 |
 117 |         ptrs[0] = new;
 118 |         ptrs[1] = (asymbol *)0;
 119 |
 120 |         bfd_set_symtab(abfd, ptrs, 1);
 121 |         bfd_close(abfd);
 122 |       }
 123 |
 124 |       ./makesym
 125 |       nm foo
 126 |       00012345 A dummy_symbol
 127
 128         Many formats cannot represent arbitary symbol information; for
 129         instance, the <<a.out>> object format does not allow an
 130         arbitary number of sections. A symbol pointing to a section
 131         which is not one  of <<.text>>, <<.data>> or <<.bss>> cannot
 132         be described.
 133
 134 */
 135
 136
 137
 138 /*
 139 DOCDD
 140 INODE
 141 typedef asymbol, symbol handling functions, Writing Symbols, Symbols
 142
 143 */
 144 /*
 145 SUBSECTION
 146         typedef asymbol
 147
 148         An <<asymbol>> has the form:
 149
 150 */
 151
 152 /*
 153 CODE_FRAGMENT
 154
 155 .
 156 .typedef struct symbol_cache_entry
 157 .{
 158 .       {* A pointer to the BFD which owns the symbol. This information
 159 .          is necessary so that a back end can work out what additional
 160 .          information (invisible to the application writer) is carried
 161 .          with the symbol.
 162 .
 163 .          This field is *almost* redundant, since you can use section->owner
 164 .          instead, except that some symbols point to the global sections
 165 .          bfd_{abs,com,und}_section.  This could be fixed by making
 166 .          these globals be per-bfd (or per-target-flavor).  FIXME. *}
 167 .
 168 .  struct _bfd *the_bfd; {* Use bfd_asymbol_bfd(sym) to access this field. *}
 169 .
 170 .       {* The text of the symbol. The name is left alone, and not copied; the
 171 .          application may not alter it. *}
 172 .  CONST char *name;
 173 .
 174 .       {* The value of the symbol.  This really should be a union of a
 175 .          numeric value with a pointer, since some flags indicate that
 176 .          a pointer to another symbol is stored here.  *}
 177 .  symvalue value;
 178 .
 179 .       {* Attributes of a symbol: *}
 180 .
 181 .#define BSF_NO_FLAGS    0x00
 182 .
 183 .       {* The symbol has local scope; <<static>> in <<C>>. The value
 184 .          is the offset into the section of the data. *}
 185 .#define BSF_LOCAL      0x01
 186 .
 187 .       {* The symbol has global scope; initialized data in <<C>>. The
 188 .          value is the offset into the section of the data. *}
 189 .#define BSF_GLOBAL     0x02
 190 .
 191 .       {* The symbol has global scope and is exported. The value is
 192 .          the offset into the section of the data. *}
 193 .#define BSF_EXPORT     BSF_GLOBAL {* no real difference *}
 194 .
 195 .       {* A normal C symbol would be one of:
 196 .          <<BSF_LOCAL>>, <<BSF_FORT_COMM>>,  <<BSF_UNDEFINED>> or
 197 .          <<BSF_GLOBAL>> *}
 198 .
 199 .       {* The symbol is a debugging record. The value has an arbitary
 200 .          meaning. *}
 201 .#define BSF_DEBUGGING  0x08
 202 .
 203 .       {* The symbol denotes a function entry point.  Used in ELF,
 204 .          perhaps others someday.  *}
 205 .#define BSF_FUNCTION    0x10
 206 .
 207 .       {* Used by the linker. *}
 208 .#define BSF_KEEP        0x20
 209 .#define BSF_KEEP_G      0x40
 210 .
 211 .       {* A weak global symbol, overridable without warnings by
 212 .          a regular global symbol of the same name.  *}
 213 .#define BSF_WEAK        0x80
 214 .
 215 .       {* This symbol was created to point to a section, e.g. ELF's
 216 .          STT_SECTION symbols.  *}
 217 .#define BSF_SECTION_SYM 0x100
 218 .
 219 .       {* The symbol used to be a common symbol, but now it is
 220 .          allocated. *}
 221 .#define BSF_OLD_COMMON  0x200
 222 .
 223 .       {* The default value for common data. *}
 224 .#define BFD_FORT_COMM_DEFAULT_VALUE 0
 225 .
 226 .       {* In some files the type of a symbol sometimes alters its
 227 .          location in an output file - ie in coff a <<ISFCN>> symbol
 228 .          which is also <<C_EXT>> symbol appears where it was
 229 .          declared and not at the end of a section.  This bit is set
 230 .          by the target BFD part to convey this information. *}
 231 .
 232 .#define BSF_NOT_AT_END    0x400
 233 .
 234 .       {* Signal that the symbol is the label of constructor section. *}
 235 .#define BSF_CONSTRUCTOR   0x800
 236 .
 237 .       {* Signal that the symbol is a warning symbol. If the symbol
 238 .          is a warning symbol, then the value field (I know this is
 239 .          tacky) will point to the asymbol which when referenced will
 240 .          cause the warning. *}
 241 .#define BSF_WARNING       0x1000
 242 .
 243 .       {* Signal that the symbol is indirect. The value of the symbol
 244 .          is a pointer to an undefined asymbol which contains the
 245 .          name to use instead. *}
 246 .#define BSF_INDIRECT      0x2000
 247 .
 248 .       {* BSF_FILE marks symbols that contain a file name.  This is used
 249 .          for ELF STT_FILE symbols.  *}
 250 .#define BSF_FILE          0x4000
 251 .
 252 .  flagword flags;
 253 .
 254 .       {* A pointer to the section to which this symbol is
 255 .          relative.  This will always be non NULL, there are special
 256 .          sections for undefined and absolute symbols *}
 257 .  struct sec *section;
 258 .
 259 .       {* Back end special data. This is being phased out in favour
 260 .          of making this a union. *}
 261 .  PTR udata;
 262 .
 263 .} asymbol;
 264 */
 265
 266 #include "bfd.h"
 267 #include "sysdep.h"
 268
 269 #include "libbfd.h"
 270 #include "aout/stab_gnu.h"
 271
 272 /*
 273 DOCDD
 274 INODE
 275 symbol handling functions,  , typedef asymbol, Symbols
 276 SUBSECTION
 277         Symbol Handling Functions
 278 */
 279
 280 /*
 281 FUNCTION
 282         get_symtab_upper_bound
 283
 284 DESCRIPTION
 285         Return the number of bytes required to store a vector of pointers
 286         to <<asymbols>> for all the symbols in the BFD @var{abfd},
 287         including a terminal NULL pointer. If there are no symbols in
 288         the BFD, then return 0.
 289
 290 .#define get_symtab_upper_bound(abfd) \
 291 .     BFD_SEND (abfd, _get_symtab_upper_bound, (abfd))
 292
 293 */
 294
 295 /*
 296 FUNCTION
 297         bfd_canonicalize_symtab
 298
 299 DESCRIPTION
 300         Read the symbols from the BFD @var{abfd}, and fills in
 301         the vector @var{location} with pointers to the symbols and
 302         a trailing NULL.
 303         Return the actual number of symbol pointers, not
 304         including the NULL.
 305
 306
 307 .#define bfd_canonicalize_symtab(abfd, location) \
 308 .     BFD_SEND (abfd, _bfd_canonicalize_symtab,\
 309 .                  (abfd, location))
 310
 311 */
 312
 313
 314 /*
 315 FUNCTION
 316         bfd_set_symtab
 317
 318 SYNOPSIS
 319         boolean bfd_set_symtab (bfd *abfd, asymbol **location, unsigned int count);
 320
 321 DESCRIPTION
 322         Arrange that when the output BFD @var{abfd} is closed,
 323         the table @var{location} of @var{count} pointers to symbols
 324         will be written.
 325 */
 326
 327 boolean
 328 bfd_set_symtab (abfd, location, symcount)
 329      bfd *abfd;
 330      asymbol **location;
 331      unsigned int symcount;
 332 {
 333   if ((abfd->format != bfd_object) || (bfd_read_p (abfd))) {
 334     bfd_error = invalid_operation;
 335     return false;
 336   }
 337
 338   bfd_get_outsymbols (abfd) = location;
 339   bfd_get_symcount (abfd) = symcount;
 340   return true;
 341 }
 342
 343 /*
 344 FUNCTION
 345         bfd_print_symbol_vandf
 346
 347 SYNOPSIS
 348         void bfd_print_symbol_vandf(PTR file, asymbol *symbol);
 349
 350 DESCRIPTION
 351         Print the value and flags of the @var{symbol} supplied to the
 352         stream @var{file}.
 353 */
 354 void
 355 DEFUN(bfd_print_symbol_vandf,(file, symbol),
 356 PTR file AND
 357 asymbol *symbol)
 358 {
 359   flagword type = symbol->flags;
 360   if (symbol->section != (asection *)NULL)
 361       {
 362         fprintf_vma(file, symbol->value+symbol->section->vma);
 363       }
 364   else
 365       {
 366         fprintf_vma(file, symbol->value);
 367       }
 368   fprintf(file," %c%c%c%c%c%c%c",
 369           (type & BSF_LOCAL)  ? 'l':' ',
 370           (type & BSF_GLOBAL) ? 'g' : ' ',
 371           (type & BSF_WEAK) ? 'w' : ' ',
 372           (type & BSF_CONSTRUCTOR) ? 'C' : ' ',
 373           (type & BSF_WARNING) ? 'W' : ' ',
 374           (type & BSF_INDIRECT) ? 'I' : ' ',
 375           (type & BSF_DEBUGGING) ? 'd' :' ');
 376
 377 }
 378
 379
 380 /*
 381 FUNCTION
 382         bfd_make_empty_symbol
 383
 384 DESCRIPTION
 385         Create a new <<asymbol>> structure for the BFD @var{abfd}
 386         and return a pointer to it.
 387
 388         This routine is necessary because each back end has private
 389         information surrounding the <<asymbol>>. Building your own
 390         <<asymbol>> and pointing to it will not create the private
 391         information, and will cause problems later on.
 392
 393 .#define bfd_make_empty_symbol(abfd) \
 394 .     BFD_SEND (abfd, _bfd_make_empty_symbol, (abfd))
 395 */
 396
 397 /*
 398 FUNCTION
 399         bfd_make_debug_symbol
 400
 401 DESCRIPTION
 402         Create a new <<asymbol>> structure for the BFD @var{abfd},
 403         to be used as a debugging symbol.  Further details of its use have
 404         yet to be worked out.
 405
 406 .#define bfd_make_debug_symbol(abfd,ptr,size) \
 407 .        BFD_SEND (abfd, _bfd_make_debug_symbol, (abfd, ptr, size))
 408 */
 409
 410 struct section_to_type
 411 {
 412   CONST char *section;
 413   char type;
 414 };
 415
 416 /* Map section names to POSIX/BSD single-character symbol types.
 417    This table is probably incomplete.  It is sorted for convenience of
 418    adding entries.  Since it is so short, a linear search is used.  */
 419 static CONST struct section_to_type stt[] = {
 420   {"*DEBUG*", 'N'},
 421   {".bss", 'b'},
 422   {".data", 'd'},
 423   {".sbss", 's'},               /* Small BSS (uninitialized data) */
 424   {".scommon", 'c'},            /* Small common */
 425   {".sdata", 'g'},              /* Small initialized data */
 426   {".text", 't'},
 427   {0, 0}
 428 };
 429
 430 /* Return the single-character symbol type corresponding to
 431    section S, or '?' for an unknown COFF section.  */
 432
 433 static char
 434 coff_section_type (s)
 435      char *s;
 436 {
 437   CONST struct section_to_type *t;
 438
 439   for (t = &stt[0]; t->section; t++)
 440     if (!strcmp (s, t->section))
 441       return t->type;
 442   return '?';
 443 }
 444
 445 #ifndef islower
 446 #define islower(c) ((c) >= 'a' && (c) <= 'z')
 447 #endif
 448 #ifndef toupper
 449 #define toupper(c) (islower(c) ? ((c) & ~0x20) : (c))
 450 #endif
 451
 452 /*
 453 FUNCTION
 454         bfd_decode_symclass
 455
 456 DESCRIPTION
 457         Return a character corresponding to the symbol
 458         class of @var{symbol}, or '?' for an unknown class.
 459
 460 SYNOPSIS
 461         int bfd_decode_symclass(asymbol *symbol);
 462 */
 463 int
 464 DEFUN(bfd_decode_symclass,(symbol),
 465 asymbol *symbol)
 466 {
 467   char c;
 468
 469   if (bfd_is_com_section (symbol->section))
 470     return 'C';
 471   if (symbol->section == &bfd_und_section)
 472     return 'U';
 473   if (symbol->section == &bfd_ind_section)
 474     return 'I';
 475   if (!(symbol->flags & (BSF_GLOBAL|BSF_LOCAL)))
 476     return '?';
 477
 478   if (symbol->section == &bfd_abs_section)
 479     c = 'a';
 480   else if (symbol->section)
 481     c = coff_section_type (symbol->section->name);
 482   else
 483     return '?';
 484   if (symbol->flags & BSF_GLOBAL)
 485     c = toupper (c);
 486   return c;
 487
 488   /* We don't have to handle these cases just yet, but we will soon:
 489      N_SETV: 'v';
 490      N_SETA: 'l';
 491      N_SETT: 'x';
 492      N_SETD: 'z';
 493      N_SETB: 's';
 494      N_INDR: 'i';
 495      */
 496 }
 497
 498 /*
 499 FUNCTION
 500         bfd_symbol_info
 501
 502 DESCRIPTION
 503         Fill in the basic info about symbol that nm needs.
 504         Additional info may be added by the back-ends after
 505         calling this function.
 506
 507 SYNOPSIS
 508         void bfd_symbol_info(asymbol *symbol, symbol_info *ret);
 509 */
 510
 511 void
 512 DEFUN(bfd_symbol_info,(symbol, ret),
 513       asymbol *symbol AND
 514       symbol_info *ret)
 515 {
 516   ret->type = bfd_decode_symclass (symbol);
 517   if (ret->type != 'U')
 518     ret->value = symbol->value+symbol->section->vma;
 519   else
 520     ret->value = 0;
 521   ret->name = symbol->name;
 522 }
 523
 524 void
 525 bfd_symbol_is_absolute()
 526 {
 527   abort();
 528 }
 529