binutils/strings.c

   1 /* strings -- print the strings of printable characters in files
   2    Copyright (C) 1993-2018 Free Software Foundation, Inc.
   3
   4    This program is free software; you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation; either version 3, or (at your option)
   7    any later version.
   8
   9    This program is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12    GNU General Public License for more details.
  13
  14    You should have received a copy of the GNU General Public License
  15    along with this program; if not, write to the Free Software
  16    Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA
  17    02110-1301, USA.  */
  18 \f
  19 /* Usage: strings [options] file...
  20
  21    Options:
  22    --all
  23    -a
  24    -            Scan each file in its entirety.
  25
  26    --data
  27    -d           Scan only the initialized data section(s) of object files.
  28
  29    --print-file-name
  30    -f           Print the name of the file before each string.
  31
  32    --bytes=min-len
  33    -n min-len
  34    -min-len     Print graphic char sequences, MIN-LEN or more bytes long,
  35                 that are followed by a NUL or a newline.  Default is 4.
  36
  37    --radix={o,x,d}
  38    -t {o,x,d}   Print the offset within the file before each string,
  39                 in octal/hex/decimal.
  40
  41   --include-all-whitespace
  42   -w            By default tab and space are the only whitepace included in graphic
  43                 char sequences.  This option considers all of isspace() valid.
  44
  45    -o           Like -to.  (Some other implementations have -o like -to,
  46                 others like -td.  We chose one arbitrarily.)
  47
  48    --encoding={s,S,b,l,B,L}
  49    -e {s,S,b,l,B,L}
  50                 Select character encoding: 7-bit-character, 8-bit-character,
  51                 bigendian 16-bit, littleendian 16-bit, bigendian 32-bit,
  52                 littleendian 32-bit.
  53
  54    --target=BFDNAME
  55    -T {bfdname}
  56                 Specify a non-default object file format.
  57
  58   --output-separator=sep_string
  59   -s sep_string String used to separate parsed strings in output.
  60                 Default is newline.
  61
  62    --help
  63    -h           Print the usage message on the standard output.
  64
  65    --version
  66    -V
  67    -v           Print the program version number.
  68
  69    Written by Richard Stallman <rms@gnu.ai.mit.edu>
  70    and David MacKenzie <djm@gnu.ai.mit.edu>.  */
  71
  72 #include "sysdep.h"
  73 #include "bfd.h"
  74 #include "getopt.h"
  75 #include "libiberty.h"
  76 #include "safe-ctype.h"
  77 #include "bucomm.h"
  78
  79 #define STRING_ISGRAPHIC(c) \
  80       (   (c) >= 0 \
  81        && (c) <= 255 \
  82        && ((c) == '\t' || ISPRINT (c) || (encoding == 'S' && (c) > 127) \
  83            || (include_all_whitespace && ISSPACE (c))) \
  84       )
  85
  86 #ifndef errno
  87 extern int errno;
  88 #endif
  89
  90 /* The BFD section flags that identify an initialized data section.  */
  91 #define DATA_FLAGS (SEC_ALLOC | SEC_LOAD | SEC_HAS_CONTENTS)
  92
  93 /* Radix for printing addresses (must be 8, 10 or 16).  */
  94 static int address_radix;
  95
  96 /* Minimum length of sequence of graphic chars to trigger output.  */
  97 static int string_min;
  98
  99 /* Whether or not we include all whitespace as a graphic char.   */
 100 static bfd_boolean include_all_whitespace;
 101
 102 /* TRUE means print address within file for each string.  */
 103 static bfd_boolean print_addresses;
 104
 105 /* TRUE means print filename for each string.  */
 106 static bfd_boolean print_filenames;
 107
 108 /* TRUE means for object files scan only the data section.  */
 109 static bfd_boolean datasection_only;
 110
 111 /* The BFD object file format.  */
 112 static char *target;
 113
 114 /* The character encoding format.  */
 115 static char encoding;
 116 static int encoding_bytes;
 117
 118 /* Output string used to separate parsed strings  */
 119 static char *output_separator;
 120
 121 static struct option long_options[] =
 122 {
 123   {"all", no_argument, NULL, 'a'},
 124   {"data", no_argument, NULL, 'd'},
 125   {"print-file-name", no_argument, NULL, 'f'},
 126   {"bytes", required_argument, NULL, 'n'},
 127   {"radix", required_argument, NULL, 't'},
 128   {"include-all-whitespace", no_argument, NULL, 'w'},
 129   {"encoding", required_argument, NULL, 'e'},
 130   {"target", required_argument, NULL, 'T'},
 131   {"output-separator", required_argument, NULL, 's'},
 132   {"help", no_argument, NULL, 'h'},
 133   {"version", no_argument, NULL, 'v'},
 134   {NULL, 0, NULL, 0}
 135 };
 136
 137 static bfd_boolean strings_file (char *);
 138 static void print_strings (const char *, FILE *, file_ptr, int, int, char *);
 139 static void usage (FILE *, int) ATTRIBUTE_NORETURN;
 140 \f
 141 int main (int, char **);
 142
 143 int
 144 main (int argc, char **argv)
 145 {
 146   int optc;
 147   int exit_status = 0;
 148   bfd_boolean files_given = FALSE;
 149   char *s;
 150   int numeric_opt = 0;
 151
 152 #if defined (HAVE_SETLOCALE)
 153   setlocale (LC_ALL, "");
 154 #endif
 155   bindtextdomain (PACKAGE, LOCALEDIR);
 156   textdomain (PACKAGE);
 157
 158   program_name = argv[0];
 159   xmalloc_set_program_name (program_name);
 160   bfd_set_error_program_name (program_name);
 161
 162   expandargv (&argc, &argv);
 163
 164   string_min = 4;
 165   include_all_whitespace = FALSE;
 166   print_addresses = FALSE;
 167   print_filenames = FALSE;
 168   if (DEFAULT_STRINGS_ALL)
 169     datasection_only = FALSE;
 170   else
 171     datasection_only = TRUE;
 172   target = NULL;
 173   encoding = 's';
 174   output_separator = NULL;
 175
 176   while ((optc = getopt_long (argc, argv, "adfhHn:wot:e:T:s:Vv0123456789",
 177                               long_options, (int *) 0)) != EOF)
 178     {
 179       switch (optc)
 180         {
 181         case 'a':
 182           datasection_only = FALSE;
 183           break;
 184
 185         case 'd':
 186           datasection_only = TRUE;
 187           break;
 188
 189         case 'f':
 190           print_filenames = TRUE;
 191           break;
 192
 193         case 'H':
 194         case 'h':
 195           usage (stdout, 0);
 196
 197         case 'n':
 198           string_min = (int) strtoul (optarg, &s, 0);
 199           if (s != NULL && *s != 0)
 200             fatal (_("invalid integer argument %s"), optarg);
 201           break;
 202
 203         case 'w':
 204           include_all_whitespace = TRUE;
 205           break;
 206
 207         case 'o':
 208           print_addresses = TRUE;
 209           address_radix = 8;
 210           break;
 211
 212         case 't':
 213           print_addresses = TRUE;
 214           if (optarg[1] != '\0')
 215             usage (stderr, 1);
 216           switch (optarg[0])
 217             {
 218             case 'o':
 219               address_radix = 8;
 220               break;
 221
 222             case 'd':
 223               address_radix = 10;
 224               break;
 225
 226             case 'x':
 227               address_radix = 16;
 228               break;
 229
 230             default:
 231               usage (stderr, 1);
 232             }
 233           break;
 234
 235         case 'T':
 236           target = optarg;
 237           break;
 238
 239         case 'e':
 240           if (optarg[1] != '\0')
 241             usage (stderr, 1);
 242           encoding = optarg[0];
 243           break;
 244
 245         case 's':
 246           output_separator = optarg;
 247           break;
 248
 249         case 'V':
 250         case 'v':
 251           print_version ("strings");
 252           break;
 253
 254         case '?':
 255           usage (stderr, 1);
 256
 257         default:
 258           numeric_opt = optind;
 259           break;
 260         }
 261     }
 262
 263   if (numeric_opt != 0)
 264     {
 265       string_min = (int) strtoul (argv[numeric_opt - 1] + 1, &s, 0);
 266       if (s != NULL && *s != 0)
 267         fatal (_("invalid integer argument %s"), argv[numeric_opt - 1] + 1);
 268     }
 269   if (string_min < 1)
 270     fatal (_("invalid minimum string length %d"), string_min);
 271
 272   switch (encoding)
 273     {
 274     case 'S':
 275     case 's':
 276       encoding_bytes = 1;
 277       break;
 278     case 'b':
 279     case 'l':
 280       encoding_bytes = 2;
 281       break;
 282     case 'B':
 283     case 'L':
 284       encoding_bytes = 4;
 285       break;
 286     default:
 287       usage (stderr, 1);
 288     }
 289
 290   bfd_init ();
 291   set_default_bfd_target ();
 292
 293   if (optind >= argc)
 294     {
 295       datasection_only = FALSE;
 296       SET_BINARY (fileno (stdin));
 297       print_strings ("{standard input}", stdin, 0, 0, 0, (char *) NULL);
 298       files_given = TRUE;
 299     }
 300   else
 301     {
 302       for (; optind < argc; ++optind)
 303         {
 304           if (strcmp (argv[optind], "-") == 0)
 305             datasection_only = FALSE;
 306           else
 307             {
 308               files_given = TRUE;
 309               exit_status |= !strings_file (argv[optind]);
 310             }
 311         }
 312     }
 313
 314   if (!files_given)
 315     usage (stderr, 1);
 316
 317   return (exit_status);
 318 }
 319 \f
 320 /* Scan section SECT of the file ABFD, whose printable name is
 321    FILENAME.  If it contains initialized data set GOT_A_SECTION and
 322    print the strings in it.  */
 323
 324 static void
 325 strings_a_section (bfd *abfd, asection *sect, const char *filename,
 326                    bfd_boolean *got_a_section)
 327 {
 328   bfd_size_type sectsize;
 329   bfd_byte *mem;
 330
 331   if ((sect->flags & DATA_FLAGS) != DATA_FLAGS)
 332     return;
 333
 334   sectsize = bfd_get_section_size (sect);
 335   if (sectsize == 0)
 336     return;
 337
 338   if (!bfd_malloc_and_get_section (abfd, sect, &mem))
 339     {
 340       non_fatal (_("%s: Reading section %s failed: %s"),
 341                  filename, sect->name, bfd_errmsg (bfd_get_error ()));
 342       return;
 343     }
 344
 345   *got_a_section = TRUE;
 346   print_strings (filename, NULL, sect->filepos, 0, sectsize, (char *) mem);
 347   free (mem);
 348 }
 349
 350 /* Scan all of the sections in FILE, and print the strings
 351    in the initialized data section(s).
 352
 353    Return TRUE if successful,
 354    FALSE if not (such as if FILE is not an object file).  */
 355
 356 static bfd_boolean
 357 strings_object_file (const char *file)
 358 {
 359   bfd *abfd;
 360   asection *s;
 361   bfd_boolean got_a_section;
 362
 363   abfd = bfd_openr (file, target);
 364
 365   if (abfd == NULL)
 366     /* Treat the file as a non-object file.  */
 367     return FALSE;
 368
 369   /* This call is mainly for its side effect of reading in the sections.
 370      We follow the traditional behavior of `strings' in that we don't
 371      complain if we don't recognize a file to be an object file.  */
 372   if (!bfd_check_format (abfd, bfd_object))
 373     {
 374       bfd_close (abfd);
 375       return FALSE;
 376     }
 377
 378   got_a_section = FALSE;
 379   for (s = abfd->sections; s != NULL; s = s->next)
 380     strings_a_section (abfd, s, file, &got_a_section);
 381
 382   if (!bfd_close (abfd))
 383     {
 384       bfd_nonfatal (file);
 385       return FALSE;
 386     }
 387
 388   return got_a_section;
 389 }
 390
 391 /* Print the strings in FILE.  Return TRUE if ok, FALSE if an error occurs.  */
 392
 393 static bfd_boolean
 394 strings_file (char *file)
 395 {
 396   struct stat st;
 397
 398   /* get_file_size does not support non-S_ISREG files.  */
 399
 400   if (stat (file, &st) < 0)
 401     {
 402       if (errno == ENOENT)
 403         non_fatal (_("'%s': No such file"), file);
 404       else
 405         non_fatal (_("Warning: could not locate '%s'.  reason: %s"),
 406                    file, strerror (errno));
 407       return FALSE;
 408     }
 409   else if (S_ISDIR (st.st_mode))
 410     {
 411       non_fatal (_("Warning: '%s' is a directory"), file);
 412       return FALSE;
 413     }
 414
 415   /* If we weren't told to scan the whole file,
 416      try to open it as an object file and only look at
 417      initialized data sections.  If that fails, fall back to the
 418      whole file.  */
 419   if (!datasection_only || !strings_object_file (file))
 420     {
 421       FILE *stream;
 422
 423       stream = fopen (file, FOPEN_RB);
 424       if (stream == NULL)
 425         {
 426           fprintf (stderr, "%s: ", program_name);
 427           perror (file);
 428           return FALSE;
 429         }
 430
 431       print_strings (file, stream, (file_ptr) 0, 0, 0, (char *) 0);
 432
 433       if (fclose (stream) == EOF)
 434         {
 435           fprintf (stderr, "%s: ", program_name);
 436           perror (file);
 437           return FALSE;
 438         }
 439     }
 440
 441   return TRUE;
 442 }
 443 \f
 444 /* Read the next character, return EOF if none available.
 445    Assume that STREAM is positioned so that the next byte read
 446    is at address ADDRESS in the file.
 447
 448    If STREAM is NULL, do not read from it.
 449    The caller can supply a buffer of characters
 450    to be processed before the data in STREAM.
 451    MAGIC is the address of the buffer and
 452    MAGICCOUNT is how many characters are in it.  */
 453
 454 static long
 455 get_char (FILE *stream, file_ptr *address, int *magiccount, char **magic)
 456 {
 457   int c, i;
 458   long r = 0;
 459
 460   for (i = 0; i < encoding_bytes; i++)
 461     {
 462       if (*magiccount)
 463         {
 464           (*magiccount)--;
 465           c = *(*magic)++;
 466         }
 467       else
 468         {
 469           if (stream == NULL)
 470             return EOF;
 471
 472           /* Only use getc_unlocked if we found a declaration for it.
 473              Otherwise, libc is not thread safe by default, and we
 474              should not use it.  */
 475
 476 #if defined(HAVE_GETC_UNLOCKED) && HAVE_DECL_GETC_UNLOCKED
 477           c = getc_unlocked (stream);
 478 #else
 479           c = getc (stream);
 480 #endif
 481           if (c == EOF)
 482             return EOF;
 483         }
 484
 485       (*address)++;
 486       r = (r << 8) | (c & 0xff);
 487     }
 488
 489   switch (encoding)
 490     {
 491     default:
 492       break;
 493     case 'l':
 494       r = ((r & 0xff) << 8) | ((r & 0xff00) >> 8);
 495       break;
 496     case 'L':
 497       r = (((r & 0xff) << 24) | ((r & 0xff00) << 8)
 498            | ((r & 0xff0000) >> 8) | ((r & 0xff000000) >> 24));
 499       break;
 500     }
 501
 502   return r;
 503 }
 504 \f
 505 /* Find the strings in file FILENAME, read from STREAM.
 506    Assume that STREAM is positioned so that the next byte read
 507    is at address ADDRESS in the file.
 508    Stop reading at address STOP_POINT in the file, if nonzero.
 509
 510    If STREAM is NULL, do not read from it.
 511    The caller can supply a buffer of characters
 512    to be processed before the data in STREAM.
 513    MAGIC is the address of the buffer and
 514    MAGICCOUNT is how many characters are in it.
 515    Those characters come at address ADDRESS and the data in STREAM follow.  */
 516
 517 static void
 518 print_strings (const char *filename, FILE *stream, file_ptr address,
 519                int stop_point, int magiccount, char *magic)
 520 {
 521   char *buf = (char *) xmalloc (sizeof (char) * (string_min + 1));
 522
 523   while (1)
 524     {
 525       file_ptr start;
 526       int i;
 527       long c;
 528
 529       /* See if the next `string_min' chars are all graphic chars.  */
 530     tryline:
 531       if (stop_point && address >= stop_point)
 532         break;
 533       start = address;
 534       for (i = 0; i < string_min; i++)
 535         {
 536           c = get_char (stream, &address, &magiccount, &magic);
 537           if (c == EOF)
 538             {
 539               free (buf);
 540               return;
 541             }
 542           if (! STRING_ISGRAPHIC (c))
 543             /* Found a non-graphic.  Try again starting with next char.  */
 544             goto tryline;
 545           buf[i] = c;
 546         }
 547
 548       /* We found a run of `string_min' graphic characters.  Print up
 549          to the next non-graphic character.  */
 550
 551       if (print_filenames)
 552         printf ("%s: ", filename);
 553       if (print_addresses)
 554         switch (address_radix)
 555           {
 556           case 8:
 557 #ifdef HAVE_LONG_LONG
 558             if (sizeof (start) > sizeof (long))
 559               {
 560 # ifndef __MSVCRT__
 561                 printf ("%7llo ", (unsigned long long) start);
 562 # else
 563                 printf ("%7I64o ", (unsigned long long) start);
 564 # endif
 565               }
 566             else
 567 #elif !BFD_HOST_64BIT_LONG
 568             if (start != (unsigned long) start)
 569               printf ("++%7lo ", (unsigned long) start);
 570             else
 571 #endif
 572               printf ("%7lo ", (unsigned long) start);
 573             break;
 574
 575           case 10:
 576 #ifdef HAVE_LONG_LONG
 577             if (sizeof (start) > sizeof (long))
 578               {
 579 # ifndef __MSVCRT__
 580                 printf ("%7lld ", (unsigned long long) start);
 581 # else
 582                 printf ("%7I64d ", (unsigned long long) start);
 583 # endif
 584               }
 585             else
 586 #elif !BFD_HOST_64BIT_LONG
 587             if (start != (unsigned long) start)
 588               printf ("++%7ld ", (unsigned long) start);
 589             else
 590 #endif
 591               printf ("%7ld ", (long) start);
 592             break;
 593
 594           case 16:
 595 #ifdef HAVE_LONG_LONG
 596             if (sizeof (start) > sizeof (long))
 597               {
 598 # ifndef __MSVCRT__
 599                 printf ("%7llx ", (unsigned long long) start);
 600 # else
 601                 printf ("%7I64x ", (unsigned long long) start);
 602 # endif
 603               }
 604             else
 605 #elif !BFD_HOST_64BIT_LONG
 606             if (start != (unsigned long) start)
 607               printf ("%lx%8.8lx ", (unsigned long) (start >> 32),
 608                       (unsigned long) (start & 0xffffffff));
 609             else
 610 #endif
 611               printf ("%7lx ", (unsigned long) start);
 612             break;
 613           }
 614
 615       buf[i] = '\0';
 616       fputs (buf, stdout);
 617
 618       while (1)
 619         {
 620           c = get_char (stream, &address, &magiccount, &magic);
 621           if (c == EOF)
 622             break;
 623           if (! STRING_ISGRAPHIC (c))
 624             break;
 625           putchar (c);
 626         }
 627
 628       if (output_separator)
 629         fputs (output_separator, stdout);
 630       else
 631         putchar ('\n');
 632     }
 633   free (buf);
 634 }
 635 \f
 636 static void
 637 usage (FILE *stream, int status)
 638 {
 639   fprintf (stream, _("Usage: %s [option(s)] [file(s)]\n"), program_name);
 640   fprintf (stream, _(" Display printable strings in [file(s)] (stdin by default)\n"));
 641   fprintf (stream, _(" The options are:\n"));
 642
 643   if (DEFAULT_STRINGS_ALL)
 644     fprintf (stream, _("\
 645   -a - --all                Scan the entire file, not just the data section [default]\n\
 646   -d --data                 Only scan the data sections in the file\n"));
 647   else
 648     fprintf (stream, _("\
 649   -a - --all                Scan the entire file, not just the data section\n\
 650   -d --data                 Only scan the data sections in the file [default]\n"));
 651
 652   fprintf (stream, _("\
 653   -f --print-file-name      Print the name of the file before each string\n\
 654   -n --bytes=[number]       Locate & print any NUL-terminated sequence of at\n\
 655   -<number>                   least [number] characters (default 4).\n\
 656   -t --radix={o,d,x}        Print the location of the string in base 8, 10 or 16\n\
 657   -w --include-all-whitespace Include all whitespace as valid string characters\n\
 658   -o                        An alias for --radix=o\n\
 659   -T --target=<BFDNAME>     Specify the binary file format\n\
 660   -e --encoding={s,S,b,l,B,L} Select character size and endianness:\n\
 661                             s = 7-bit, S = 8-bit, {b,l} = 16-bit, {B,L} = 32-bit\n\
 662   -s --output-separator=<string> String used to separate strings in output.\n\
 663   @<file>                   Read options from <file>\n\
 664   -h --help                 Display this information\n\
 665   -v -V --version           Print the program's version number\n"));
 666   list_supported_targets (program_name, stream);
 667   if (REPORT_BUGS_TO[0] && status == 0)
 668     fprintf (stream, _("Report bugs to %s\n"), REPORT_BUGS_TO);
 669   exit (status);
 670 }