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 **) 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 .       {* Symbol is from dynamic linking information.  *}
 253 .#define BSF_DYNAMIC       0x8000
 254 .
 255 .  flagword flags;
 256 .
 257 .       {* A pointer to the section to which this symbol is
 258 .          relative.  This will always be non NULL, there are special
 259 .          sections for undefined and absolute symbols *}
 260 .  struct sec *section;
 261 .
 262 .       {* Back end special data. This is being phased out in favour
 263 .          of making this a union. *}
 264 .  PTR udata;
 265 .
 266 .} asymbol;
 267 */
 268
 269 #include "bfd.h"
 270 #include "sysdep.h"
 271
 272 #include "libbfd.h"
 273 #include "aout/stab_gnu.h"
 274
 275 /*
 276 DOCDD
 277 INODE
 278 symbol handling functions,  , typedef asymbol, Symbols
 279 SUBSECTION
 280         Symbol handling functions
 281 */
 282
 283 /*
 284 FUNCTION
 285         get_symtab_upper_bound
 286
 287 DESCRIPTION
 288         Return the number of bytes required to store a vector of pointers
 289         to <<asymbols>> for all the symbols in the BFD @var{abfd},
 290         including a terminal NULL pointer. If there are no symbols in
 291         the BFD, then return 0.
 292
 293 .#define get_symtab_upper_bound(abfd) \
 294 .     BFD_SEND (abfd, _get_symtab_upper_bound, (abfd))
 295
 296 */
 297
 298 /*
 299 FUNCTION
 300         bfd_canonicalize_symtab
 301
 302 DESCRIPTION
 303         Read the symbols from the BFD @var{abfd}, and fills in
 304         the vector @var{location} with pointers to the symbols and
 305         a trailing NULL.
 306         Return the actual number of symbol pointers, not
 307         including the NULL.
 308
 309
 310 .#define bfd_canonicalize_symtab(abfd, location) \
 311 .     BFD_SEND (abfd, _bfd_canonicalize_symtab,\
 312 .                  (abfd, location))
 313
 314 */
 315
 316
 317 /*
 318 FUNCTION
 319         bfd_set_symtab
 320
 321 SYNOPSIS
 322         boolean bfd_set_symtab (bfd *abfd, asymbol **location, unsigned int count);
 323
 324 DESCRIPTION
 325         Arrange that when the output BFD @var{abfd} is closed,
 326         the table @var{location} of @var{count} pointers to symbols
 327         will be written.
 328 */
 329
 330 boolean
 331 bfd_set_symtab (abfd, location, symcount)
 332      bfd *abfd;
 333      asymbol **location;
 334      unsigned int symcount;
 335 {
 336   if ((abfd->format != bfd_object) || (bfd_read_p (abfd)))
 337     {
 338       bfd_set_error (bfd_error_invalid_operation);
 339       return false;
 340     }
 341
 342   bfd_get_outsymbols (abfd) = location;
 343   bfd_get_symcount (abfd) = symcount;
 344   return true;
 345 }
 346
 347 /*
 348 FUNCTION
 349         bfd_print_symbol_vandf
 350
 351 SYNOPSIS
 352         void bfd_print_symbol_vandf(PTR file, asymbol *symbol);
 353
 354 DESCRIPTION
 355         Print the value and flags of the @var{symbol} supplied to the
 356         stream @var{file}.
 357 */
 358 void
 359 bfd_print_symbol_vandf (arg, symbol)
 360      PTR arg;
 361      asymbol *symbol;
 362 {
 363   FILE *file = (FILE *) arg;
 364   flagword type = symbol->flags;
 365   if (symbol->section != (asection *) NULL)
 366     {
 367       fprintf_vma (file, symbol->value + symbol->section->vma);
 368     }
 369   else
 370     {
 371       fprintf_vma (file, symbol->value);
 372     }
 373
 374   /* This presumes that a symbol can not be both BSF_DEBUGGING and
 375      BSF_DYNAMIC.  */
 376   fprintf (file, " %c%c%c%c%c%c%c",
 377            (type & BSF_LOCAL) ? 'l' : ' ',
 378            (type & BSF_GLOBAL) ? 'g' : ' ',
 379            (type & BSF_WEAK) ? 'w' : ' ',
 380            (type & BSF_CONSTRUCTOR) ? 'C' : ' ',
 381            (type & BSF_WARNING) ? 'W' : ' ',
 382            (type & BSF_INDIRECT) ? 'I' : ' ',
 383            (type & BSF_DEBUGGING) ? 'd'
 384            : (type & BSF_DYNAMIC) ? 'D' : ' ');
 385 }
 386
 387
 388 /*
 389 FUNCTION
 390         bfd_make_empty_symbol
 391
 392 DESCRIPTION
 393         Create a new <<asymbol>> structure for the BFD @var{abfd}
 394         and return a pointer to it.
 395
 396         This routine is necessary because each back end has private
 397         information surrounding the <<asymbol>>. Building your own
 398         <<asymbol>> and pointing to it will not create the private
 399         information, and will cause problems later on.
 400
 401 .#define bfd_make_empty_symbol(abfd) \
 402 .     BFD_SEND (abfd, _bfd_make_empty_symbol, (abfd))
 403 */
 404
 405 /*
 406 FUNCTION
 407         bfd_make_debug_symbol
 408
 409 DESCRIPTION
 410         Create a new <<asymbol>> structure for the BFD @var{abfd},
 411         to be used as a debugging symbol.  Further details of its use have
 412         yet to be worked out.
 413
 414 .#define bfd_make_debug_symbol(abfd,ptr,size) \
 415 .        BFD_SEND (abfd, _bfd_make_debug_symbol, (abfd, ptr, size))
 416 */
 417
 418 struct section_to_type
 419 {
 420   CONST char *section;
 421   char type;
 422 };
 423
 424 /* Map section names to POSIX/BSD single-character symbol types.
 425    This table is probably incomplete.  It is sorted for convenience of
 426    adding entries.  Since it is so short, a linear search is used.  */
 427 static CONST struct section_to_type stt[] =
 428 {
 429   {"*DEBUG*", 'N'},
 430   {".bss", 'b'},
 431   {".data", 'd'},
 432   {".sbss", 's'},               /* Small BSS (uninitialized data) */
 433   {".scommon", 'c'},            /* Small common */
 434   {".sdata", 'g'},              /* Small initialized data */
 435   {".text", 't'},
 436   {0, 0}
 437 };
 438
 439 /* Return the single-character symbol type corresponding to
 440    section S, or '?' for an unknown COFF section.  */
 441
 442 static char
 443 coff_section_type (s)
 444      char *s;
 445 {
 446   CONST struct section_to_type *t;
 447
 448   for (t = &stt[0]; t->section; t++)
 449     if (!strcmp (s, t->section))
 450       return t->type;
 451   return '?';
 452 }
 453
 454 #ifndef islower
 455 #define islower(c) ((c) >= 'a' && (c) <= 'z')
 456 #endif
 457 #ifndef toupper
 458 #define toupper(c) (islower(c) ? ((c) & ~0x20) : (c))
 459 #endif
 460
 461 /*
 462 FUNCTION
 463         bfd_decode_symclass
 464
 465 DESCRIPTION
 466         Return a character corresponding to the symbol
 467         class of @var{symbol}, or '?' for an unknown class.
 468
 469 SYNOPSIS
 470         int bfd_decode_symclass(asymbol *symbol);
 471 */
 472 int
 473 bfd_decode_symclass (symbol)
 474      asymbol *symbol;
 475 {
 476   char c;
 477
 478   if (bfd_is_com_section (symbol->section))
 479     return 'C';
 480   if (symbol->section == &bfd_und_section)
 481     return 'U';
 482   if (symbol->section == &bfd_ind_section)
 483     return 'I';
 484   if (!(symbol->flags & (BSF_GLOBAL | BSF_LOCAL)))
 485     return '?';
 486
 487   if (symbol->section == &bfd_abs_section)
 488     c = 'a';
 489   else if (symbol->section)
 490     c = coff_section_type (symbol->section->name);
 491   else
 492     return '?';
 493   if (symbol->flags & BSF_GLOBAL)
 494     c = toupper (c);
 495   return c;
 496
 497   /* We don't have to handle these cases just yet, but we will soon:
 498      N_SETV: 'v';
 499      N_SETA: 'l';
 500      N_SETT: 'x';
 501      N_SETD: 'z';
 502      N_SETB: 's';
 503      N_INDR: 'i';
 504      */
 505 }
 506
 507 /*
 508 FUNCTION
 509         bfd_symbol_info
 510
 511 DESCRIPTION
 512         Fill in the basic info about symbol that nm needs.
 513         Additional info may be added by the back-ends after
 514         calling this function.
 515
 516 SYNOPSIS
 517         void bfd_symbol_info(asymbol *symbol, symbol_info *ret);
 518 */
 519
 520 void
 521 bfd_symbol_info (symbol, ret)
 522      asymbol *symbol;
 523      symbol_info *ret;
 524 {
 525   ret->type = bfd_decode_symclass (symbol);
 526   if (ret->type != 'U')
 527     ret->value = symbol->value + symbol->section->vma;
 528   else
 529     ret->value = 0;
 530   ret->name = symbol->name;
 531 }
 532
 533 void
 534 bfd_symbol_is_absolute ()
 535 {
 536   abort ();
 537 }