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    -            Do not scan only the initialized data section of object files.
  25
  26    --print-file-name
  27    -f           Print the name of the file before each string.
  28
  29    --bytes=min-len
  30    -n min-len
  31    -min-len     Print graphic char sequences, MIN-LEN or more bytes long,
  32                 that are followed by a NUL or a newline.  Default is 4.
  33
  34    --radix={o,x,d}
  35    -t {o,x,d}   Print the offset within the file before each string,
  36                 in octal/hex/decimal.
  37
  38    -o           Like -to.  (Some other implementations have -o like -to,
  39                 others like -td.  We chose one arbitrarily.)
  40
  41    --encoding={s,S,b,l,B,L}
  42    -e {s,S,b,l,B,L}
  43                 Select character encoding: 7-bit-character, 8-bit-character,
  44                 bigendian 16-bit, littleendian 16-bit, bigendian 32-bit,
  45                 littleendian 32-bit.
  46
  47    --target=BFDNAME
  48    -T {bfdname}
  49                 Specify a non-default object file format.
  50
  51    --help
  52    -h           Print the usage message on the standard output.
  53
  54    --version
  55    -V
  56    -v           Print the program version number.
  57
  58    Written by Richard Stallman <rms@gnu.ai.mit.edu>
  59    and David MacKenzie <djm@gnu.ai.mit.edu>.  */
  60
  61 #include "sysdep.h"
  62 #include "bfd.h"
  63 #include "getopt.h"
  64 #include "libiberty.h"
  65 #include "safe-ctype.h"
  66 #include "bucomm.h"
  67
  68 #define STRING_ISGRAPHIC(c) \
  69       (   (c) >= 0 \
  70        && (c) <= 255 \
  71        && ((c) == '\t' || ISPRINT (c) || (encoding == 'S' && (c) > 127)))
  72
  73 #ifndef errno
  74 extern int errno;
  75 #endif
  76
  77 /* The BFD section flags that identify an initialized data section.  */
  78 #define DATA_FLAGS (SEC_ALLOC | SEC_LOAD | SEC_HAS_CONTENTS)
  79
  80 /* Radix for printing addresses (must be 8, 10 or 16).  */
  81 static int address_radix;
  82
  83 /* Minimum length of sequence of graphic chars to trigger output.  */
  84 static int string_min;
  85
  86 /* TRUE means print address within file for each string.  */
  87 static bfd_boolean print_addresses;
  88
  89 /* TRUE means print filename for each string.  */
  90 static bfd_boolean print_filenames;
  91
  92 /* TRUE means for object files scan only the data section.  */
  93 static bfd_boolean datasection_only;
  94
  95 /* TRUE if we found an initialized data section in the current file.  */
  96 static bfd_boolean got_a_section;
  97
  98 /* The BFD object file format.  */
  99 static char *target;
 100
 101 /* The character encoding format.  */
 102 static char encoding;
 103 static int encoding_bytes;
 104
 105 static struct option long_options[] =
 106 {
 107   {"all", no_argument, NULL, 'a'},
 108   {"print-file-name", no_argument, NULL, 'f'},
 109   {"bytes", required_argument, NULL, 'n'},
 110   {"radix", required_argument, NULL, 't'},
 111   {"encoding", required_argument, NULL, 'e'},
 112   {"target", required_argument, NULL, 'T'},
 113   {"help", no_argument, NULL, 'h'},
 114   {"version", no_argument, NULL, 'v'},
 115   {NULL, 0, NULL, 0}
 116 };
 117
 118 /* Records the size of a named file so that we
 119    do not repeatedly run bfd_stat() on it.  */
 120
 121 typedef struct
 122 {
 123   const char *  filename;
 124   bfd_size_type filesize;
 125 } filename_and_size_t;
 126
 127 static void strings_a_section (bfd *, asection *, void *);
 128 static bfd_boolean strings_object_file (const char *);
 129 static bfd_boolean strings_file (char *file);
 130 static void print_strings (const char *, FILE *, file_ptr, int, int, char *);
 131 static void usage (FILE *, int);
 132 static long get_char (FILE *, file_ptr *, int *, char **);
 133 \f
 134 int main (int, char **);
 135
 136 int
 137 main (int argc, char **argv)
 138 {
 139   int optc;
 140   int exit_status = 0;
 141   bfd_boolean files_given = FALSE;
 142   char *s;
 143   int numeric_opt = 0;
 144
 145 #if defined (HAVE_SETLOCALE)
 146   setlocale (LC_ALL, "");
 147 #endif
 148   bindtextdomain (PACKAGE, LOCALEDIR);
 149   textdomain (PACKAGE);
 150
 151   program_name = argv[0];
 152   xmalloc_set_program_name (program_name);
 153
 154   expandargv (&argc, &argv);
 155
 156   string_min = 4;
 157   print_addresses = FALSE;
 158   print_filenames = FALSE;
 159   datasection_only = TRUE;
 160   target = NULL;
 161   encoding = 's';
 162
 163   while ((optc = getopt_long (argc, argv, "afhHn:ot:e:T:Vv0123456789",
 164                               long_options, (int *) 0)) != EOF)
 165     {
 166       switch (optc)
 167         {
 168         case 'a':
 169           datasection_only = FALSE;
 170           break;
 171
 172         case 'f':
 173           print_filenames = TRUE;
 174           break;
 175
 176         case 'H':
 177         case 'h':
 178           usage (stdout, 0);
 179
 180         case 'n':
 181           string_min = (int) strtoul (optarg, &s, 0);
 182           if (s != NULL && *s != 0)
 183             fatal (_("invalid integer argument %s"), optarg);
 184           break;
 185
 186         case 'o':
 187           print_addresses = TRUE;
 188           address_radix = 8;
 189           break;
 190
 191         case 't':
 192           print_addresses = TRUE;
 193           if (optarg[1] != '\0')
 194             usage (stderr, 1);
 195           switch (optarg[0])
 196             {
 197             case 'o':
 198               address_radix = 8;
 199               break;
 200
 201             case 'd':
 202               address_radix = 10;
 203               break;
 204
 205             case 'x':
 206               address_radix = 16;
 207               break;
 208
 209             default:
 210               usage (stderr, 1);
 211             }
 212           break;
 213
 214         case 'T':
 215           target = optarg;
 216           break;
 217
 218         case 'e':
 219           if (optarg[1] != '\0')
 220             usage (stderr, 1);
 221           encoding = optarg[0];
 222           break;
 223
 224         case 'V':
 225         case 'v':
 226           print_version ("strings");
 227           break;
 228
 229         case '?':
 230           usage (stderr, 1);
 231
 232         default:
 233           numeric_opt = optind;
 234           break;
 235         }
 236     }
 237
 238   if (numeric_opt != 0)
 239     {
 240       string_min = (int) strtoul (argv[numeric_opt - 1] + 1, &s, 0);
 241       if (s != NULL && *s != 0)
 242         fatal (_("invalid integer argument %s"), argv[numeric_opt - 1] + 1);
 243     }
 244   if (string_min < 1)
 245     fatal (_("invalid minimum string length %d"), string_min);
 246
 247   switch (encoding)
 248     {
 249     case 'S':
 250     case 's':
 251       encoding_bytes = 1;
 252       break;
 253     case 'b':
 254     case 'l':
 255       encoding_bytes = 2;
 256       break;
 257     case 'B':
 258     case 'L':
 259       encoding_bytes = 4;
 260       break;
 261     default:
 262       usage (stderr, 1);
 263     }
 264
 265   bfd_init ();
 266   set_default_bfd_target ();
 267
 268   if (optind >= argc)
 269     {
 270       datasection_only = FALSE;
 271       SET_BINARY (fileno (stdin));
 272       print_strings ("{standard input}", stdin, 0, 0, 0, (char *) NULL);
 273       files_given = TRUE;
 274     }
 275   else
 276     {
 277       for (; optind < argc; ++optind)
 278         {
 279           if (strcmp (argv[optind], "-") == 0)
 280             datasection_only = FALSE;
 281           else
 282             {
 283               files_given = TRUE;
 284               exit_status |= strings_file (argv[optind]) == FALSE;
 285             }
 286         }
 287     }
 288
 289   if (!files_given)
 290     usage (stderr, 1);
 291
 292   return (exit_status);
 293 }
 294 \f
 295 /* Scan section SECT of the file ABFD, whose printable name is in
 296    ARG->filename and whose size might be in ARG->filesize.  If it
 297    contains initialized data set `got_a_section' and print the
 298    strings in it.
 299
 300    FIXME: We ought to be able to return error codes/messages for
 301    certain conditions.  */
 302
 303 static void
 304 strings_a_section (bfd *abfd, asection *sect, void *arg)
 305 {
 306   filename_and_size_t * filename_and_sizep;
 307   bfd_size_type *filesizep;
 308   bfd_size_type sectsize;
 309   void *mem;
 310
 311   if ((sect->flags & DATA_FLAGS) != DATA_FLAGS)
 312     return;
 313
 314   sectsize = bfd_get_section_size (sect);
 315
 316   if (sectsize <= 0)
 317     return;
 318
 319   /* Get the size of the file.  This might have been cached for us.  */
 320   filename_and_sizep = (filename_and_size_t *) arg;
 321   filesizep = & filename_and_sizep->filesize;
 322
 323   if (*filesizep == 0)
 324     {
 325       struct stat st;
 326
 327       if (bfd_stat (abfd, &st))
 328         return;
 329
 330       /* Cache the result so that we do not repeatedly stat this file.  */
 331       *filesizep = st.st_size;
 332     }
 333
 334   /* Compare the size of the section against the size of the file.
 335      If the section is bigger then the file must be corrupt and
 336      we should not try dumping it.  */
 337   if (sectsize >= *filesizep)
 338     return;
 339
 340   mem = xmalloc (sectsize);
 341
 342   if (bfd_get_section_contents (abfd, sect, mem, (file_ptr) 0, sectsize))
 343     {
 344       got_a_section = TRUE;
 345
 346       print_strings (filename_and_sizep->filename, NULL, sect->filepos,
 347                      0, sectsize, (char *) mem);
 348     }
 349
 350   free (mem);
 351 }
 352
 353 /* Scan all of the sections in FILE, and print the strings
 354    in the initialized data section(s).
 355
 356    Return TRUE if successful,
 357    FALSE if not (such as if FILE is not an object file).  */
 358
 359 static bfd_boolean
 360 strings_object_file (const char *file)
 361 {
 362   filename_and_size_t filename_and_size;
 363   bfd *abfd;
 364
 365   abfd = bfd_openr (file, target);
 366
 367   if (abfd == NULL)
 368     /* Treat the file as a non-object file.  */
 369     return FALSE;
 370
 371   /* This call is mainly for its side effect of reading in the sections.
 372      We follow the traditional behavior of `strings' in that we don't
 373      complain if we don't recognize a file to be an object file.  */
 374   if (!bfd_check_format (abfd, bfd_object))
 375     {
 376       bfd_close (abfd);
 377       return FALSE;
 378     }
 379
 380   got_a_section = FALSE;
 381   filename_and_size.filename = file;
 382   filename_and_size.filesize = 0;
 383   bfd_map_over_sections (abfd, strings_a_section, & filename_and_size);
 384
 385   if (!bfd_close (abfd))
 386     {
 387       bfd_nonfatal (file);
 388       return FALSE;
 389     }
 390
 391   return got_a_section;
 392 }
 393
 394 /* Print the strings in FILE.  Return TRUE if ok, FALSE if an error occurs.  */
 395
 396 static bfd_boolean
 397 strings_file (char *file)
 398 {
 399   struct stat st;
 400
 401   /* get_file_size does not support non-S_ISREG files.  */
 402
 403   if (stat (file, &st) < 0)
 404     {
 405       if (errno == ENOENT)
 406         non_fatal (_("'%s': No such file"), file);
 407       else
 408         non_fatal (_("Warning: could not locate '%s'.  reason: %s"),
 409                    file, strerror (errno));
 410       return FALSE;
 411     }
 412
 413   /* If we weren't told to scan the whole file,
 414      try to open it as an object file and only look at
 415      initialized data sections.  If that fails, fall back to the
 416      whole file.  */
 417   if (!datasection_only || !strings_object_file (file))
 418     {
 419       FILE *stream;
 420
 421       stream = fopen (file, FOPEN_RB);
 422       if (stream == NULL)
 423         {
 424           fprintf (stderr, "%s: ", program_name);
 425           perror (file);
 426           return FALSE;
 427         }
 428
 429       print_strings (file, stream, (file_ptr) 0, 0, 0, (char *) 0);
 430
 431       if (fclose (stream) == EOF)
 432         {
 433           fprintf (stderr, "%s: ", program_name);
 434           perror (file);
 435           return FALSE;
 436         }
 437     }
 438
 439   return TRUE;
 440 }
 441 \f
 442 /* Read the next character, return EOF if none available.
 443    Assume that STREAM is positioned so that the next byte read
 444    is at address ADDRESS in the file.
 445
 446    If STREAM is NULL, do not read from it.
 447    The caller can supply a buffer of characters
 448    to be processed before the data in STREAM.
 449    MAGIC is the address of the buffer and
 450    MAGICCOUNT is how many characters are in it.  */
 451
 452 static long
 453 get_char (FILE *stream, file_ptr *address, int *magiccount, char **magic)
 454 {
 455   int c, i;
 456   long r = 0;
 457
 458   for (i = 0; i < encoding_bytes; i++)
 459     {
 460       if (*magiccount)
 461         {
 462           (*magiccount)--;
 463           c = *(*magic)++;
 464         }
 465       else
 466         {
 467           if (stream == NULL)
 468             return EOF;
 469
 470           /* Only use getc_unlocked if we found a declaration for it.
 471              Otherwise, libc is not thread safe by default, and we
 472              should not use it.  */
 473
 474 #if defined(HAVE_GETC_UNLOCKED) && HAVE_DECL_GETC_UNLOCKED
 475           c = getc_unlocked (stream);
 476 #else
 477           c = getc (stream);
 478 #endif
 479           if (c == EOF)
 480             return EOF;
 481         }
 482
 483       (*address)++;
 484       r = (r << 8) | (c & 0xff);
 485     }
 486
 487   switch (encoding)
 488     {
 489     default:
 490       break;
 491     case 'l':
 492       r = ((r & 0xff) << 8) | ((r & 0xff00) >> 8);
 493       break;
 494     case 'L':
 495       r = (((r & 0xff) << 24) | ((r & 0xff00) << 8)
 496            | ((r & 0xff0000) >> 8) | ((r & 0xff000000) >> 24));
 497       break;
 498     }
 499
 500   return r;
 501 }
 502 \f
 503 /* Find the strings in file FILENAME, read from STREAM.
 504    Assume that STREAM is positioned so that the next byte read
 505    is at address ADDRESS in the file.
 506    Stop reading at address STOP_POINT in the file, if nonzero.
 507
 508    If STREAM is NULL, do not read from it.
 509    The caller can supply a buffer of characters
 510    to be processed before the data in STREAM.
 511    MAGIC is the address of the buffer and
 512    MAGICCOUNT is how many characters are in it.
 513    Those characters come at address ADDRESS and the data in STREAM follow.  */
 514
 515 static void
 516 print_strings (const char *filename, FILE *stream, file_ptr address,
 517                int stop_point, int magiccount, char *magic)
 518 {
 519   char *buf = (char *) xmalloc (sizeof (char) * (string_min + 1));
 520
 521   while (1)
 522     {
 523       file_ptr start;
 524       int i;
 525       long c;
 526
 527       /* See if the next `string_min' chars are all graphic chars.  */
 528     tryline:
 529       if (stop_point && address >= stop_point)
 530         break;
 531       start = address;
 532       for (i = 0; i < string_min; i++)
 533         {
 534           c = get_char (stream, &address, &magiccount, &magic);
 535           if (c == EOF)
 536             {
 537               free (buf);
 538               return;
 539             }
 540           if (! STRING_ISGRAPHIC (c))
 541             /* Found a non-graphic.  Try again starting with next char.  */
 542             goto tryline;
 543           buf[i] = c;
 544         }
 545
 546       /* We found a run of `string_min' graphic characters.  Print up
 547          to the next non-graphic character.  */
 548
 549       if (print_filenames)
 550         printf ("%s: ", filename);
 551       if (print_addresses)
 552         switch (address_radix)
 553           {
 554           case 8:
 555 #if __STDC_VERSION__ >= 199901L || (defined(__GNUC__) && __GNUC__ >= 2)
 556             if (sizeof (start) > sizeof (long))
 557               {
 558 #ifndef __MSVCRT__
 559                 printf ("%7llo ", (unsigned long long) start);
 560 #else
 561                 printf ("%7I64o ", (unsigned long long) start);
 562 #endif
 563               }
 564             else
 565 #elif !BFD_HOST_64BIT_LONG
 566             if (start != (unsigned long) start)
 567               printf ("++%7lo ", (unsigned long) start);
 568             else
 569 #endif
 570               printf ("%7lo ", (unsigned long) start);
 571             break;
 572
 573           case 10:
 574 #if __STDC_VERSION__ >= 199901L || (defined(__GNUC__) && __GNUC__ >= 2)
 575             if (sizeof (start) > sizeof (long))
 576               {
 577 #ifndef __MSVCRT__
 578                 printf ("%7lld ", (unsigned long long) start);
 579 #else
 580                 printf ("%7I64d ", (unsigned long long) start);
 581 #endif
 582               }
 583             else
 584 #elif !BFD_HOST_64BIT_LONG
 585             if (start != (unsigned long) start)
 586               printf ("++%7ld ", (unsigned long) start);
 587             else
 588 #endif
 589               printf ("%7ld ", (long) start);
 590             break;
 591
 592           case 16:
 593 #if __STDC_VERSION__ >= 199901L || (defined(__GNUC__) && __GNUC__ >= 2)
 594             if (sizeof (start) > sizeof (long))
 595               {
 596 #ifndef __MSVCRT__
 597                 printf ("%7llx ", (unsigned long long) start);
 598 #else
 599                 printf ("%7I64x ", (unsigned long long) start);
 600 #endif
 601               }
 602             else
 603 #elif !BFD_HOST_64BIT_LONG
 604             if (start != (unsigned long) start)
 605               printf ("%lx%8.8lx ", (unsigned long) (start >> 32),
 606                       (unsigned long) (start & 0xffffffff));
 607             else
 608 #endif
 609               printf ("%7lx ", (unsigned long) start);
 610             break;
 611           }
 612
 613       buf[i] = '\0';
 614       fputs (buf, stdout);
 615
 616       while (1)
 617         {
 618           c = get_char (stream, &address, &magiccount, &magic);
 619           if (c == EOF)
 620             break;
 621           if (! STRING_ISGRAPHIC (c))
 622             break;
 623           putchar (c);
 624         }
 625
 626       putchar ('\n');
 627     }
 628   free (buf);
 629 }
 630 \f
 631 static void
 632 usage (FILE *stream, int status)
 633 {
 634   fprintf (stream, _("Usage: %s [option(s)] [file(s)]\n"), program_name);
 635   fprintf (stream, _(" Display printable strings in [file(s)] (stdin by default)\n"));
 636   fprintf (stream, _(" The options are:\n\
 637   -a - --all                Scan the entire file, not just the data section\n\
 638   -f --print-file-name      Print the name of the file before each string\n\
 639   -n --bytes=[number]       Locate & print any NUL-terminated sequence of at\n\
 640   -<number>                   least [number] characters (default 4).\n\
 641   -t --radix={o,d,x}        Print the location of the string in base 8, 10 or 16\n\
 642   -o                        An alias for --radix=o\n\
 643   -T --target=<BFDNAME>     Specify the binary file format\n\
 644   -e --encoding={s,S,b,l,B,L} Select character size and endianness:\n\
 645                             s = 7-bit, S = 8-bit, {b,l} = 16-bit, {B,L} = 32-bit\n\
 646   @<file>                   Read options from <file>\n\
 647   -h --help                 Display this information\n\
 648   -v -V --version           Print the program's version number\n"));
 649   list_supported_targets (program_name, stream);
 650   if (REPORT_BUGS_TO[0] && status == 0)
 651     fprintf (stream, _("Report bugs to %s\n"), REPORT_BUGS_TO);
 652   exit (status);
 653 }