binutils/strings.c

   1 /* strings -- print the strings of printable characters in files
   2    Copyright (C) 1993-2014 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    --help
  59    -h           Print the usage message on the standard output.
  60
  61    --version
  62    -V
  63    -v           Print the program version number.
  64
  65    Written by Richard Stallman <rms@gnu.ai.mit.edu>
  66    and David MacKenzie <djm@gnu.ai.mit.edu>.  */
  67
  68 #include "sysdep.h"
  69 #include "bfd.h"
  70 #include "getopt.h"
  71 #include "libiberty.h"
  72 #include "safe-ctype.h"
  73 #include "bucomm.h"
  74
  75 #define STRING_ISGRAPHIC(c) \
  76       (   (c) >= 0 \
  77        && (c) <= 255 \
  78        && ((c) == '\t' || ISPRINT (c) || (encoding == 'S' && (c) > 127) \
  79            || (include_all_whitespace == TRUE && ISSPACE (c))) \
  80       )
  81
  82 #ifndef errno
  83 extern int errno;
  84 #endif
  85
  86 /* The BFD section flags that identify an initialized data section.  */
  87 #define DATA_FLAGS (SEC_ALLOC | SEC_LOAD | SEC_HAS_CONTENTS)
  88
  89 /* Radix for printing addresses (must be 8, 10 or 16).  */
  90 static int address_radix;
  91
  92 /* Minimum length of sequence of graphic chars to trigger output.  */
  93 static int string_min;
  94
  95 /* Whether or not we include all whitespace as a graphic char.   */
  96 static bfd_boolean include_all_whitespace;
  97
  98 /* TRUE means print address within file for each string.  */
  99 static bfd_boolean print_addresses;
 100
 101 /* TRUE means print filename for each string.  */
 102 static bfd_boolean print_filenames;
 103
 104 /* TRUE means for object files scan only the data section.  */
 105 static bfd_boolean datasection_only;
 106
 107 /* TRUE if we found an initialized data section in the current file.  */
 108 static bfd_boolean got_a_section;
 109
 110 /* The BFD object file format.  */
 111 static char *target;
 112
 113 /* The character encoding format.  */
 114 static char encoding;
 115 static int encoding_bytes;
 116
 117 static struct option long_options[] =
 118 {
 119   {"all", no_argument, NULL, 'a'},
 120   {"data", no_argument, NULL, 'd'},
 121   {"print-file-name", no_argument, NULL, 'f'},
 122   {"bytes", required_argument, NULL, 'n'},
 123   {"radix", required_argument, NULL, 't'},
 124   {"include-all-whitespace", required_argument, NULL, 'w'},
 125   {"encoding", required_argument, NULL, 'e'},
 126   {"target", required_argument, NULL, 'T'},
 127   {"help", no_argument, NULL, 'h'},
 128   {"version", no_argument, NULL, 'v'},
 129   {NULL, 0, NULL, 0}
 130 };
 131
 132 /* Records the size of a named file so that we
 133    do not repeatedly run bfd_stat() on it.  */
 134
 135 typedef struct
 136 {
 137   const char *  filename;
 138   bfd_size_type filesize;
 139 } filename_and_size_t;
 140
 141 static void strings_a_section (bfd *, asection *, void *);
 142 static bfd_boolean strings_object_file (const char *);
 143 static bfd_boolean strings_file (char *);
 144 static void print_strings (const char *, FILE *, file_ptr, int, int, char *);
 145 static void usage (FILE *, int);
 146 static long get_char (FILE *, file_ptr *, int *, char **);
 147 \f
 148 int main (int, char **);
 149
 150 int
 151 main (int argc, char **argv)
 152 {
 153   int optc;
 154   int exit_status = 0;
 155   bfd_boolean files_given = FALSE;
 156   char *s;
 157   int numeric_opt = 0;
 158
 159 #if defined (HAVE_SETLOCALE)
 160   setlocale (LC_ALL, "");
 161 #endif
 162   bindtextdomain (PACKAGE, LOCALEDIR);
 163   textdomain (PACKAGE);
 164
 165   program_name = argv[0];
 166   xmalloc_set_program_name (program_name);
 167
 168   expandargv (&argc, &argv);
 169
 170   string_min = 4;
 171   include_all_whitespace = FALSE;
 172   print_addresses = FALSE;
 173   print_filenames = FALSE;
 174   if (DEFAULT_STRINGS_ALL)
 175     datasection_only = FALSE;
 176   else
 177     datasection_only = TRUE;
 178   target = NULL;
 179   encoding = 's';
 180
 181   while ((optc = getopt_long (argc, argv, "adfhHn:wot:e:T:Vv0123456789",
 182                               long_options, (int *) 0)) != EOF)
 183     {
 184       switch (optc)
 185         {
 186         case 'a':
 187           datasection_only = FALSE;
 188           break;
 189
 190         case 'd':
 191           datasection_only = TRUE;
 192           break;
 193
 194         case 'f':
 195           print_filenames = TRUE;
 196           break;
 197
 198         case 'H':
 199         case 'h':
 200           usage (stdout, 0);
 201
 202         case 'n':
 203           string_min = (int) strtoul (optarg, &s, 0);
 204           if (s != NULL && *s != 0)
 205             fatal (_("invalid integer argument %s"), optarg);
 206           break;
 207
 208         case 'w':
 209           include_all_whitespace = TRUE;
 210           break;
 211
 212         case 'o':
 213           print_addresses = TRUE;
 214           address_radix = 8;
 215           break;
 216
 217         case 't':
 218           print_addresses = TRUE;
 219           if (optarg[1] != '\0')
 220             usage (stderr, 1);
 221           switch (optarg[0])
 222             {
 223             case 'o':
 224               address_radix = 8;
 225               break;
 226
 227             case 'd':
 228               address_radix = 10;
 229               break;
 230
 231             case 'x':
 232               address_radix = 16;
 233               break;
 234
 235             default:
 236               usage (stderr, 1);
 237             }
 238           break;
 239
 240         case 'T':
 241           target = optarg;
 242           break;
 243
 244         case 'e':
 245           if (optarg[1] != '\0')
 246             usage (stderr, 1);
 247           encoding = optarg[0];
 248           break;
 249
 250         case 'V':
 251         case 'v':
 252           print_version ("strings");
 253           break;
 254
 255         case '?':
 256           usage (stderr, 1);
 257
 258         default:
 259           numeric_opt = optind;
 260           break;
 261         }
 262     }
 263
 264   if (numeric_opt != 0)
 265     {
 266       string_min = (int) strtoul (argv[numeric_opt - 1] + 1, &s, 0);
 267       if (s != NULL && *s != 0)
 268         fatal (_("invalid integer argument %s"), argv[numeric_opt - 1] + 1);
 269     }
 270   if (string_min < 1)
 271     fatal (_("invalid minimum string length %d"), string_min);
 272
 273   switch (encoding)
 274     {
 275     case 'S':
 276     case 's':
 277       encoding_bytes = 1;
 278       break;
 279     case 'b':
 280     case 'l':
 281       encoding_bytes = 2;
 282       break;
 283     case 'B':
 284     case 'L':
 285       encoding_bytes = 4;
 286       break;
 287     default:
 288       usage (stderr, 1);
 289     }
 290
 291   bfd_init ();
 292   set_default_bfd_target ();
 293
 294   if (optind >= argc)
 295     {
 296       datasection_only = FALSE;
 297       SET_BINARY (fileno (stdin));
 298       print_strings ("{standard input}", stdin, 0, 0, 0, (char *) NULL);
 299       files_given = TRUE;
 300     }
 301   else
 302     {
 303       for (; optind < argc; ++optind)
 304         {
 305           if (strcmp (argv[optind], "-") == 0)
 306             datasection_only = FALSE;
 307           else
 308             {
 309               files_given = TRUE;
 310               exit_status |= strings_file (argv[optind]) == FALSE;
 311             }
 312         }
 313     }
 314
 315   if (!files_given)
 316     usage (stderr, 1);
 317
 318   return (exit_status);
 319 }
 320 \f
 321 /* Scan section SECT of the file ABFD, whose printable name is in
 322    ARG->filename and whose size might be in ARG->filesize.  If it
 323    contains initialized data set `got_a_section' and print the
 324    strings in it.
 325
 326    FIXME: We ought to be able to return error codes/messages for
 327    certain conditions.  */
 328
 329 static void
 330 strings_a_section (bfd *abfd, asection *sect, void *arg)
 331 {
 332   filename_and_size_t * filename_and_sizep;
 333   bfd_size_type *filesizep;
 334   bfd_size_type sectsize;
 335   void *mem;
 336
 337   if ((sect->flags & DATA_FLAGS) != DATA_FLAGS)
 338     return;
 339
 340   sectsize = bfd_get_section_size (sect);
 341
 342   if (sectsize <= 0)
 343     return;
 344
 345   /* Get the size of the file.  This might have been cached for us.  */
 346   filename_and_sizep = (filename_and_size_t *) arg;
 347   filesizep = & filename_and_sizep->filesize;
 348
 349   if (*filesizep == 0)
 350     {
 351       struct stat st;
 352
 353       if (bfd_stat (abfd, &st))
 354         return;
 355
 356       /* Cache the result so that we do not repeatedly stat this file.  */
 357       *filesizep = st.st_size;
 358     }
 359
 360   /* Compare the size of the section against the size of the file.
 361      If the section is bigger then the file must be corrupt and
 362      we should not try dumping it.  */
 363   if (sectsize >= *filesizep)
 364     return;
 365
 366   mem = xmalloc (sectsize);
 367
 368   if (bfd_get_section_contents (abfd, sect, mem, (file_ptr) 0, sectsize))
 369     {
 370       got_a_section = TRUE;
 371
 372       print_strings (filename_and_sizep->filename, NULL, sect->filepos,
 373                      0, sectsize, (char *) mem);
 374     }
 375
 376   free (mem);
 377 }
 378
 379 /* Scan all of the sections in FILE, and print the strings
 380    in the initialized data section(s).
 381
 382    Return TRUE if successful,
 383    FALSE if not (such as if FILE is not an object file).  */
 384
 385 static bfd_boolean
 386 strings_object_file (const char *file)
 387 {
 388   filename_and_size_t filename_and_size;
 389   bfd *abfd;
 390
 391   abfd = bfd_openr (file, target);
 392
 393   if (abfd == NULL)
 394     /* Treat the file as a non-object file.  */
 395     return FALSE;
 396
 397   /* This call is mainly for its side effect of reading in the sections.
 398      We follow the traditional behavior of `strings' in that we don't
 399      complain if we don't recognize a file to be an object file.  */
 400   if (!bfd_check_format (abfd, bfd_object))
 401     {
 402       bfd_close (abfd);
 403       return FALSE;
 404     }
 405
 406   got_a_section = FALSE;
 407   filename_and_size.filename = file;
 408   filename_and_size.filesize = 0;
 409   bfd_map_over_sections (abfd, strings_a_section, & filename_and_size);
 410
 411   if (!bfd_close (abfd))
 412     {
 413       bfd_nonfatal (file);
 414       return FALSE;
 415     }
 416
 417   return got_a_section;
 418 }
 419
 420 /* Print the strings in FILE.  Return TRUE if ok, FALSE if an error occurs.  */
 421
 422 static bfd_boolean
 423 strings_file (char *file)
 424 {
 425   struct stat st;
 426
 427   /* get_file_size does not support non-S_ISREG files.  */
 428
 429   if (stat (file, &st) < 0)
 430     {
 431       if (errno == ENOENT)
 432         non_fatal (_("'%s': No such file"), file);
 433       else
 434         non_fatal (_("Warning: could not locate '%s'.  reason: %s"),
 435                    file, strerror (errno));
 436       return FALSE;
 437     }
 438
 439   /* If we weren't told to scan the whole file,
 440      try to open it as an object file and only look at
 441      initialized data sections.  If that fails, fall back to the
 442      whole file.  */
 443   if (!datasection_only || !strings_object_file (file))
 444     {
 445       FILE *stream;
 446
 447       stream = fopen (file, FOPEN_RB);
 448       if (stream == NULL)
 449         {
 450           fprintf (stderr, "%s: ", program_name);
 451           perror (file);
 452           return FALSE;
 453         }
 454
 455       print_strings (file, stream, (file_ptr) 0, 0, 0, (char *) 0);
 456
 457       if (fclose (stream) == EOF)
 458         {
 459           fprintf (stderr, "%s: ", program_name);
 460           perror (file);
 461           return FALSE;
 462         }
 463     }
 464
 465   return TRUE;
 466 }
 467 \f
 468 /* Read the next character, return EOF if none available.
 469    Assume that STREAM is positioned so that the next byte read
 470    is at address ADDRESS in the file.
 471
 472    If STREAM is NULL, do not read from it.
 473    The caller can supply a buffer of characters
 474    to be processed before the data in STREAM.
 475    MAGIC is the address of the buffer and
 476    MAGICCOUNT is how many characters are in it.  */
 477
 478 static long
 479 get_char (FILE *stream, file_ptr *address, int *magiccount, char **magic)
 480 {
 481   int c, i;
 482   long r = 0;
 483
 484   for (i = 0; i < encoding_bytes; i++)
 485     {
 486       if (*magiccount)
 487         {
 488           (*magiccount)--;
 489           c = *(*magic)++;
 490         }
 491       else
 492         {
 493           if (stream == NULL)
 494             return EOF;
 495
 496           /* Only use getc_unlocked if we found a declaration for it.
 497              Otherwise, libc is not thread safe by default, and we
 498              should not use it.  */
 499
 500 #if defined(HAVE_GETC_UNLOCKED) && HAVE_DECL_GETC_UNLOCKED
 501           c = getc_unlocked (stream);
 502 #else
 503           c = getc (stream);
 504 #endif
 505           if (c == EOF)
 506             return EOF;
 507         }
 508
 509       (*address)++;
 510       r = (r << 8) | (c & 0xff);
 511     }
 512
 513   switch (encoding)
 514     {
 515     default:
 516       break;
 517     case 'l':
 518       r = ((r & 0xff) << 8) | ((r & 0xff00) >> 8);
 519       break;
 520     case 'L':
 521       r = (((r & 0xff) << 24) | ((r & 0xff00) << 8)
 522            | ((r & 0xff0000) >> 8) | ((r & 0xff000000) >> 24));
 523       break;
 524     }
 525
 526   return r;
 527 }
 528 \f
 529 /* Find the strings in file FILENAME, read from STREAM.
 530    Assume that STREAM is positioned so that the next byte read
 531    is at address ADDRESS in the file.
 532    Stop reading at address STOP_POINT in the file, if nonzero.
 533
 534    If STREAM is NULL, do not read from it.
 535    The caller can supply a buffer of characters
 536    to be processed before the data in STREAM.
 537    MAGIC is the address of the buffer and
 538    MAGICCOUNT is how many characters are in it.
 539    Those characters come at address ADDRESS and the data in STREAM follow.  */
 540
 541 static void
 542 print_strings (const char *filename, FILE *stream, file_ptr address,
 543                int stop_point, int magiccount, char *magic)
 544 {
 545   char *buf = (char *) xmalloc (sizeof (char) * (string_min + 1));
 546
 547   while (1)
 548     {
 549       file_ptr start;
 550       int i;
 551       long c;
 552
 553       /* See if the next `string_min' chars are all graphic chars.  */
 554     tryline:
 555       if (stop_point && address >= stop_point)
 556         break;
 557       start = address;
 558       for (i = 0; i < string_min; i++)
 559         {
 560           c = get_char (stream, &address, &magiccount, &magic);
 561           if (c == EOF)
 562             {
 563               free (buf);
 564               return;
 565             }
 566           if (! STRING_ISGRAPHIC (c))
 567             /* Found a non-graphic.  Try again starting with next char.  */
 568             goto tryline;
 569           buf[i] = c;
 570         }
 571
 572       /* We found a run of `string_min' graphic characters.  Print up
 573          to the next non-graphic character.  */
 574
 575       if (print_filenames)
 576         printf ("%s: ", filename);
 577       if (print_addresses)
 578         switch (address_radix)
 579           {
 580           case 8:
 581 #ifdef HAVE_LONG_LONG
 582             if (sizeof (start) > sizeof (long))
 583               {
 584 # ifndef __MSVCRT__
 585                 printf ("%7llo ", (unsigned long long) start);
 586 # else
 587                 printf ("%7I64o ", (unsigned long long) start);
 588 # endif
 589               }
 590             else
 591 #elif !BFD_HOST_64BIT_LONG
 592             if (start != (unsigned long) start)
 593               printf ("++%7lo ", (unsigned long) start);
 594             else
 595 #endif
 596               printf ("%7lo ", (unsigned long) start);
 597             break;
 598
 599           case 10:
 600 #ifdef HAVE_LONG_LONG
 601             if (sizeof (start) > sizeof (long))
 602               {
 603 # ifndef __MSVCRT__
 604                 printf ("%7lld ", (unsigned long long) start);
 605 # else
 606                 printf ("%7I64d ", (unsigned long long) start);
 607 # endif
 608               }
 609             else
 610 #elif !BFD_HOST_64BIT_LONG
 611             if (start != (unsigned long) start)
 612               printf ("++%7ld ", (unsigned long) start);
 613             else
 614 #endif
 615               printf ("%7ld ", (long) start);
 616             break;
 617
 618           case 16:
 619 #ifdef HAVE_LONG_LONG
 620             if (sizeof (start) > sizeof (long))
 621               {
 622 # ifndef __MSVCRT__
 623                 printf ("%7llx ", (unsigned long long) start);
 624 # else
 625                 printf ("%7I64x ", (unsigned long long) start);
 626 # endif
 627               }
 628             else
 629 #elif !BFD_HOST_64BIT_LONG
 630             if (start != (unsigned long) start)
 631               printf ("%lx%8.8lx ", (unsigned long) (start >> 32),
 632                       (unsigned long) (start & 0xffffffff));
 633             else
 634 #endif
 635               printf ("%7lx ", (unsigned long) start);
 636             break;
 637           }
 638
 639       buf[i] = '\0';
 640       fputs (buf, stdout);
 641
 642       while (1)
 643         {
 644           c = get_char (stream, &address, &magiccount, &magic);
 645           if (c == EOF)
 646             break;
 647           if (! STRING_ISGRAPHIC (c))
 648             break;
 649           putchar (c);
 650         }
 651
 652       putchar ('\n');
 653     }
 654   free (buf);
 655 }
 656 \f
 657 static void
 658 usage (FILE *stream, int status)
 659 {
 660   fprintf (stream, _("Usage: %s [option(s)] [file(s)]\n"), program_name);
 661   fprintf (stream, _(" Display printable strings in [file(s)] (stdin by default)\n"));
 662   fprintf (stream, _(" The options are:\n"));
 663
 664   if (DEFAULT_STRINGS_ALL)
 665     fprintf (stream, _("\
 666   -a - --all                Scan the entire file, not just the data section [default]\n\
 667   -d --data                 Only scan the data sections in the file\n"));
 668   else
 669     fprintf (stream, _("\
 670   -a - --all                Scan the entire file, not just the data section\n\
 671   -d --data                 Only scan the data sections in the file [default]\n"));
 672
 673   fprintf (stream, _("\
 674   -f --print-file-name      Print the name of the file before each string\n\
 675   -n --bytes=[number]       Locate & print any NUL-terminated sequence of at\n\
 676   -<number>                   least [number] characters (default 4).\n\
 677   -t --radix={o,d,x}        Print the location of the string in base 8, 10 or 16\n\
 678   -w --include-all-whitespace Include all whitespace as valid string characters\n\
 679   -o                        An alias for --radix=o\n\
 680   -T --target=<BFDNAME>     Specify the binary file format\n\
 681   -e --encoding={s,S,b,l,B,L} Select character size and endianness:\n\
 682                             s = 7-bit, S = 8-bit, {b,l} = 16-bit, {B,L} = 32-bit\n\
 683   @<file>                   Read options from <file>\n\
 684   -h --help                 Display this information\n\
 685   -v -V --version           Print the program's version number\n"));
 686   list_supported_targets (program_name, stream);
 687   if (REPORT_BUGS_TO[0] && status == 0)
 688     fprintf (stream, _("Report bugs to %s\n"), REPORT_BUGS_TO);
 689   exit (status);
 690 }