contrib/file/src/magic.c

   1 /*
   2  * Copyright (c) Christos Zoulas 2003.
   3  * All Rights Reserved.
   4  *
   5  * Redistribution and use in source and binary forms, with or without
   6  * modification, are permitted provided that the following conditions
   7  * are met:
   8  * 1. Redistributions of source code must retain the above copyright
   9  *    notice immediately at the beginning of the file, without modification,
  10  *    this list of conditions, and the following disclaimer.
  11  * 2. Redistributions in binary form must reproduce the above copyright
  12  *    notice, this list of conditions and the following disclaimer in the
  13  *    documentation and/or other materials provided with the distribution.
  14  *
  15  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
  16  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  17  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  18  * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
  19  * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  20  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  21  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  22  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  23  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  24  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  25  * SUCH DAMAGE.
  26  */
  27
  28 #ifdef WIN32
  29 #include <windows.h>
  30 #include <shlwapi.h>
  31 #endif
  32
  33 #include "file.h"
  34
  35 #ifndef lint
  36 FILE_RCSID("@(#)$File: magic.c,v 1.91 2014/12/16 23:18:40 christos Exp $")
  37 #endif  /* lint */
  38
  39 #include "magic.h"
  40
  41 #include <stdlib.h>
  42 #include <unistd.h>
  43 #include <string.h>
  44 #ifdef QUICK
  45 #include <sys/mman.h>
  46 #endif
  47 #ifdef HAVE_LIMITS_H
  48 #include <limits.h>     /* for PIPE_BUF */
  49 #endif
  50
  51 #if defined(HAVE_UTIMES)
  52 # include <sys/time.h>
  53 #elif defined(HAVE_UTIME)
  54 # if defined(HAVE_SYS_UTIME_H)
  55 #  include <sys/utime.h>
  56 # elif defined(HAVE_UTIME_H)
  57 #  include <utime.h>
  58 # endif
  59 #endif
  60
  61 #ifdef HAVE_UNISTD_H
  62 #include <unistd.h>     /* for read() */
  63 #endif
  64
  65 #ifndef PIPE_BUF
  66 /* Get the PIPE_BUF from pathconf */
  67 #ifdef _PC_PIPE_BUF
  68 #define PIPE_BUF pathconf(".", _PC_PIPE_BUF)
  69 #else
  70 #define PIPE_BUF 512
  71 #endif
  72 #endif
  73
  74 private void close_and_restore(const struct magic_set *, const char *, int,
  75     const struct stat *);
  76 private int unreadable_info(struct magic_set *, mode_t, const char *);
  77 private const char* get_default_magic(void);
  78 #ifndef COMPILE_ONLY
  79 private const char *file_or_fd(struct magic_set *, const char *, int);
  80 #endif
  81
  82 #ifndef STDIN_FILENO
  83 #define STDIN_FILENO    0
  84 #endif
  85
  86 private const char *
  87 get_default_magic(void)
  88 {
  89         static const char hmagic[] = "/.magic/magic.mgc";
  90         static char *default_magic;
  91         char *home, *hmagicpath;
  92
  93 #ifndef WIN32
  94         struct stat st;
  95
  96         if (default_magic) {
  97                 free(default_magic);
  98                 default_magic = NULL;
  99         }
 100         if ((home = getenv("HOME")) == NULL)
 101                 return MAGIC;
 102
 103         if (asprintf(&hmagicpath, "%s/.magic.mgc", home) < 0)
 104                 return MAGIC;
 105         if (stat(hmagicpath, &st) == -1) {
 106                 free(hmagicpath);
 107                 if (asprintf(&hmagicpath, "%s/.magic", home) < 0)
 108                         return MAGIC;
 109                 if (stat(hmagicpath, &st) == -1)
 110                         goto out;
 111                 if (S_ISDIR(st.st_mode)) {
 112                         free(hmagicpath);
 113                         if (asprintf(&hmagicpath, "%s/%s", home, hmagic) < 0)
 114                                 return MAGIC;
 115                         if (access(hmagicpath, R_OK) == -1)
 116                                 goto out;
 117                 }
 118         }
 119
 120         if (asprintf(&default_magic, "%s:%s", hmagicpath, MAGIC) < 0)
 121                 goto out;
 122         free(hmagicpath);
 123         return default_magic;
 124 out:
 125         default_magic = NULL;
 126         free(hmagicpath);
 127         return MAGIC;
 128 #else
 129         char *hmagicp;
 130         char *tmppath = NULL;
 131         LPTSTR dllpath;
 132         hmagicpath = NULL;
 133
 134 #define APPENDPATH() \
 135         do { \
 136                 if (tmppath && access(tmppath, R_OK) != -1) { \
 137                         if (hmagicpath == NULL) \
 138                                 hmagicpath = tmppath; \
 139                         else { \
 140                                 if (asprintf(&hmagicp, "%s%c%s", hmagicpath, \
 141                                     PATHSEP, tmppath) >= 0) { \
 142                                         free(hmagicpath); \
 143                                         hmagicpath = hmagicp; \
 144                                 } \
 145                                 free(tmppath); \
 146                         } \
 147                         tmppath = NULL; \
 148                 } \
 149         } while (/*CONSTCOND*/0)
 150
 151         if (default_magic) {
 152                 free(default_magic);
 153                 default_magic = NULL;
 154         }
 155
 156         /* First, try to get user-specific magic file */
 157         if ((home = getenv("LOCALAPPDATA")) == NULL) {
 158                 if ((home = getenv("USERPROFILE")) != NULL)
 159                         if (asprintf(&tmppath,
 160                             "%s/Local Settings/Application Data%s", home,
 161                             hmagic) < 0)
 162                                 tmppath = NULL;
 163         } else {
 164                 if (asprintf(&tmppath, "%s%s", home, hmagic) < 0)
 165                         tmppath = NULL;
 166         }
 167
 168         APPENDPATH();
 169
 170         /* Second, try to get a magic file from Common Files */
 171         if ((home = getenv("COMMONPROGRAMFILES")) != NULL) {
 172                 if (asprintf(&tmppath, "%s%s", home, hmagic) >= 0)
 173                         APPENDPATH();
 174         }
 175
 176         /* Third, try to get magic file relative to dll location */
 177         dllpath = malloc(sizeof(*dllpath) * (MAX_PATH + 1));
 178         dllpath[MAX_PATH] = 0;  /* just in case long path gets truncated and not null terminated */
 179         if (GetModuleFileNameA(NULL, dllpath, MAX_PATH)){
 180                 PathRemoveFileSpecA(dllpath);
 181                 if (strlen(dllpath) > 3 &&
 182                     stricmp(&dllpath[strlen(dllpath) - 3], "bin") == 0) {
 183                         if (asprintf(&tmppath,
 184                             "%s/../share/misc/magic.mgc", dllpath) >= 0)
 185                                 APPENDPATH();
 186                 } else {
 187                         if (asprintf(&tmppath,
 188                             "%s/share/misc/magic.mgc", dllpath) >= 0)
 189                                 APPENDPATH();
 190                         else if (asprintf(&tmppath,
 191                             "%s/magic.mgc", dllpath) >= 0)
 192                                 APPENDPATH();
 193                 }
 194         }
 195
 196         /* Don't put MAGIC constant - it likely points to a file within MSys
 197         tree */
 198         default_magic = hmagicpath;
 199         return default_magic;
 200 #endif
 201 }
 202
 203 public const char *
 204 magic_getpath(const char *magicfile, int action)
 205 {
 206         if (magicfile != NULL)
 207                 return magicfile;
 208
 209         magicfile = getenv("MAGIC");
 210         if (magicfile != NULL)
 211                 return magicfile;
 212
 213         return action == FILE_LOAD ? get_default_magic() : MAGIC;
 214 }
 215
 216 public struct magic_set *
 217 magic_open(int flags)
 218 {
 219         return file_ms_alloc(flags);
 220 }
 221
 222 private int
 223 unreadable_info(struct magic_set *ms, mode_t md, const char *file)
 224 {
 225         if (file) {
 226                 /* We cannot open it, but we were able to stat it. */
 227                 if (access(file, W_OK) == 0)
 228                         if (file_printf(ms, "writable, ") == -1)
 229                                 return -1;
 230                 if (access(file, X_OK) == 0)
 231                         if (file_printf(ms, "executable, ") == -1)
 232                                 return -1;
 233         }
 234         if (S_ISREG(md))
 235                 if (file_printf(ms, "regular file, ") == -1)
 236                         return -1;
 237         if (file_printf(ms, "no read permission") == -1)
 238                 return -1;
 239         return 0;
 240 }
 241
 242 public void
 243 magic_close(struct magic_set *ms)
 244 {
 245         if (ms == NULL)
 246                 return;
 247         file_ms_free(ms);
 248 }
 249
 250 /*
 251  * load a magic file
 252  */
 253 public int
 254 magic_load(struct magic_set *ms, const char *magicfile)
 255 {
 256         if (ms == NULL)
 257                 return -1;
 258         return file_apprentice(ms, magicfile, FILE_LOAD);
 259 }
 260
 261 #ifndef COMPILE_ONLY
 262 /*
 263  * Install a set of compiled magic buffers.
 264  */
 265 public int
 266 magic_load_buffers(struct magic_set *ms, void **bufs, size_t *sizes,
 267     size_t nbufs)
 268 {
 269         if (ms == NULL)
 270                 return -1;
 271         return buffer_apprentice(ms, (struct magic **)bufs, sizes, nbufs);
 272 }
 273 #endif
 274
 275 public int
 276 magic_compile(struct magic_set *ms, const char *magicfile)
 277 {
 278         if (ms == NULL)
 279                 return -1;
 280         return file_apprentice(ms, magicfile, FILE_COMPILE);
 281 }
 282
 283 public int
 284 magic_check(struct magic_set *ms, const char *magicfile)
 285 {
 286         if (ms == NULL)
 287                 return -1;
 288         return file_apprentice(ms, magicfile, FILE_CHECK);
 289 }
 290
 291 public int
 292 magic_list(struct magic_set *ms, const char *magicfile)
 293 {
 294         if (ms == NULL)
 295                 return -1;
 296         return file_apprentice(ms, magicfile, FILE_LIST);
 297 }
 298
 299 private void
 300 close_and_restore(const struct magic_set *ms, const char *name, int fd,
 301     const struct stat *sb)
 302 {
 303         if (fd == STDIN_FILENO || name == NULL)
 304                 return;
 305         (void) close(fd);
 306
 307         if ((ms->flags & MAGIC_PRESERVE_ATIME) != 0) {
 308                 /*
 309                  * Try to restore access, modification times if read it.
 310                  * This is really *bad* because it will modify the status
 311                  * time of the file... And of course this will affect
 312                  * backup programs
 313                  */
 314 #ifdef HAVE_UTIMES
 315                 struct timeval  utsbuf[2];
 316                 (void)memset(utsbuf, 0, sizeof(utsbuf));
 317                 utsbuf[0].tv_sec = sb->st_atime;
 318                 utsbuf[1].tv_sec = sb->st_mtime;
 319
 320                 (void) utimes(name, utsbuf); /* don't care if loses */
 321 #elif defined(HAVE_UTIME_H) || defined(HAVE_SYS_UTIME_H)
 322                 struct utimbuf  utbuf;
 323
 324                 (void)memset(&utbuf, 0, sizeof(utbuf));
 325                 utbuf.actime = sb->st_atime;
 326                 utbuf.modtime = sb->st_mtime;
 327                 (void) utime(name, &utbuf); /* don't care if loses */
 328 #endif
 329         }
 330 }
 331
 332 #ifndef COMPILE_ONLY
 333
 334 /*
 335  * find type of descriptor
 336  */
 337 public const char *
 338 magic_descriptor(struct magic_set *ms, int fd)
 339 {
 340         if (ms == NULL)
 341                 return NULL;
 342         return file_or_fd(ms, NULL, fd);
 343 }
 344
 345 /*
 346  * find type of named file
 347  */
 348 public const char *
 349 magic_file(struct magic_set *ms, const char *inname)
 350 {
 351         if (ms == NULL)
 352                 return NULL;
 353         return file_or_fd(ms, inname, STDIN_FILENO);
 354 }
 355
 356 private const char *
 357 file_or_fd(struct magic_set *ms, const char *inname, int fd)
 358 {
 359         int     rv = -1;
 360         unsigned char *buf;
 361         struct stat     sb;
 362         ssize_t nbytes = 0;     /* number of bytes read from a datafile */
 363         int     ispipe = 0;
 364         off_t   pos = (off_t)-1;
 365
 366         if (file_reset(ms) == -1)
 367                 goto out;
 368
 369         /*
 370          * one extra for terminating '\0', and
 371          * some overlapping space for matches near EOF
 372          */
 373 #define SLOP (1 + sizeof(union VALUETYPE))
 374         if ((buf = CAST(unsigned char *, malloc(HOWMANY + SLOP))) == NULL)
 375                 return NULL;
 376
 377         switch (file_fsmagic(ms, inname, &sb)) {
 378         case -1:                /* error */
 379                 goto done;
 380         case 0:                 /* nothing found */
 381                 break;
 382         default:                /* matched it and printed type */
 383                 rv = 0;
 384                 goto done;
 385         }
 386
 387 #ifdef WIN32
 388         /* Place stdin in binary mode, so EOF (Ctrl+Z) doesn't stop early. */
 389         if (fd == STDIN_FILENO)
 390                 _setmode(STDIN_FILENO, O_BINARY);
 391 #endif
 392
 393         if (inname == NULL) {
 394                 if (fstat(fd, &sb) == 0 && S_ISFIFO(sb.st_mode))
 395                         ispipe = 1;
 396                 else
 397                         pos = lseek(fd, (off_t)0, SEEK_CUR);
 398         } else {
 399                 int flags = O_RDONLY|O_BINARY;
 400                 int okstat = stat(inname, &sb) == 0;
 401
 402                 if (okstat && S_ISFIFO(sb.st_mode)) {
 403 #ifdef O_NONBLOCK
 404                         flags |= O_NONBLOCK;
 405 #endif
 406                         ispipe = 1;
 407                 }
 408
 409                 errno = 0;
 410                 if ((fd = open(inname, flags)) < 0) {
 411 #ifdef WIN32
 412                         /*
 413                          * Can't stat, can't open.  It may have been opened in
 414                          * fsmagic, so if the user doesn't have read permission,
 415                          * allow it to say so; otherwise an error was probably
 416                          * displayed in fsmagic.
 417                          */
 418                         if (!okstat && errno == EACCES) {
 419                                 sb.st_mode = S_IFBLK;
 420                                 okstat = 1;
 421                         }
 422 #endif
 423                         if (okstat &&
 424                             unreadable_info(ms, sb.st_mode, inname) == -1)
 425                                 goto done;
 426                         rv = 0;
 427                         goto done;
 428                 }
 429 #ifdef O_NONBLOCK
 430                 if ((flags = fcntl(fd, F_GETFL)) != -1) {
 431                         flags &= ~O_NONBLOCK;
 432                         (void)fcntl(fd, F_SETFL, flags);
 433                 }
 434 #endif
 435         }
 436
 437         /*
 438          * try looking at the first HOWMANY bytes
 439          */
 440         if (ispipe) {
 441                 ssize_t r = 0;
 442
 443                 while ((r = sread(fd, (void *)&buf[nbytes],
 444                     (size_t)(HOWMANY - nbytes), 1)) > 0) {
 445                         nbytes += r;
 446                         if (r < PIPE_BUF) break;
 447                 }
 448
 449                 if (nbytes == 0) {
 450                         /* We can not read it, but we were able to stat it. */
 451                         if (unreadable_info(ms, sb.st_mode, inname) == -1)
 452                                 goto done;
 453                         rv = 0;
 454                         goto done;
 455                 }
 456
 457         } else {
 458                 /* Windows refuses to read from a big console buffer. */
 459                 size_t howmany =
 460 #if defined(WIN32) && HOWMANY > 8 * 1024
 461                                 _isatty(fd) ? 8 * 1024 :
 462 #endif
 463                                 HOWMANY;
 464                 if ((nbytes = read(fd, (char *)buf, howmany)) == -1) {
 465                         if (inname == NULL && fd != STDIN_FILENO)
 466                                 file_error(ms, errno, "cannot read fd %d", fd);
 467                         else
 468                                 file_error(ms, errno, "cannot read `%s'",
 469                                     inname == NULL ? "/dev/stdin" : inname);
 470                         goto done;
 471                 }
 472         }
 473
 474         (void)memset(buf + nbytes, 0, SLOP); /* NUL terminate */
 475         if (file_buffer(ms, fd, inname, buf, (size_t)nbytes) == -1)
 476                 goto done;
 477         rv = 0;
 478 done:
 479         free(buf);
 480         if (pos != (off_t)-1)
 481                 (void)lseek(fd, pos, SEEK_SET);
 482         close_and_restore(ms, inname, fd, &sb);
 483 out:
 484         return rv == 0 ? file_getbuffer(ms) : NULL;
 485 }
 486
 487
 488 public const char *
 489 magic_buffer(struct magic_set *ms, const void *buf, size_t nb)
 490 {
 491         if (ms == NULL)
 492                 return NULL;
 493         if (file_reset(ms) == -1)
 494                 return NULL;
 495         /*
 496          * The main work is done here!
 497          * We have the file name and/or the data buffer to be identified.
 498          */
 499         if (file_buffer(ms, -1, NULL, buf, nb) == -1) {
 500                 return NULL;
 501         }
 502         return file_getbuffer(ms);
 503 }
 504 #endif
 505
 506 public const char *
 507 magic_error(struct magic_set *ms)
 508 {
 509         if (ms == NULL)
 510                 return "Magic database is not open";
 511         return (ms->event_flags & EVENT_HAD_ERR) ? ms->o.buf : NULL;
 512 }
 513
 514 public int
 515 magic_errno(struct magic_set *ms)
 516 {
 517         if (ms == NULL)
 518                 return EINVAL;
 519         return (ms->event_flags & EVENT_HAD_ERR) ? ms->error : 0;
 520 }
 521
 522 public int
 523 magic_setflags(struct magic_set *ms, int flags)
 524 {
 525         if (ms == NULL)
 526                 return -1;
 527 #if !defined(HAVE_UTIME) && !defined(HAVE_UTIMES)
 528         if (flags & MAGIC_PRESERVE_ATIME)
 529                 return -1;
 530 #endif
 531         ms->flags = flags;
 532         return 0;
 533 }
 534
 535 public int
 536 magic_version(void)
 537 {
 538         return MAGIC_VERSION;
 539 }
 540
 541 public int
 542 magic_setparam(struct magic_set *ms, int param, const void *val)
 543 {
 544         switch (param) {
 545         case MAGIC_PARAM_INDIR_MAX:
 546                 ms->indir_max = *(const size_t *)val;
 547                 return 0;
 548         case MAGIC_PARAM_NAME_MAX:
 549                 ms->name_max = *(const size_t *)val;
 550                 return 0;
 551         case MAGIC_PARAM_ELF_PHNUM_MAX:
 552                 ms->elf_phnum_max = *(const size_t *)val;
 553                 return 0;
 554         case MAGIC_PARAM_ELF_SHNUM_MAX:
 555                 ms->elf_shnum_max = *(const size_t *)val;
 556                 return 0;
 557         case MAGIC_PARAM_ELF_NOTES_MAX:
 558                 ms->elf_notes_max = *(const size_t *)val;
 559                 return 0;
 560         default:
 561                 errno = EINVAL;
 562                 return -1;
 563         }
 564 }
 565
 566 public int
 567 magic_getparam(struct magic_set *ms, int param, void *val)
 568 {
 569         switch (param) {
 570         case MAGIC_PARAM_INDIR_MAX:
 571                 *(size_t *)val = ms->indir_max;
 572                 return 0;
 573         case MAGIC_PARAM_NAME_MAX:
 574                 *(size_t *)val = ms->name_max;
 575                 return 0;
 576         case MAGIC_PARAM_ELF_PHNUM_MAX:
 577                 *(size_t *)val = ms->elf_phnum_max;
 578                 return 0;
 579         case MAGIC_PARAM_ELF_SHNUM_MAX:
 580                 *(size_t *)val = ms->elf_shnum_max;
 581                 return 0;
 582         case MAGIC_PARAM_ELF_NOTES_MAX:
 583                 *(size_t *)val = ms->elf_notes_max;
 584                 return 0;
 585         default:
 586                 errno = EINVAL;
 587                 return -1;
 588         }
 589 }