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 1.114 2021/02/05 21:33:49 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 #include <limits.h> /* for PIPE_BUF */ 48 49 #if defined(HAVE_UTIMES) 50 # include <sys/time.h> 51 #elif defined(HAVE_UTIME) 52 # if defined(HAVE_SYS_UTIME_H) 53 # include <sys/utime.h> 54 # elif defined(HAVE_UTIME_H) 55 # include <utime.h> 56 # endif 57 #endif 58 59 #ifdef HAVE_UNISTD_H 60 #include <unistd.h> /* for read() */ 61 #endif 62 63 #ifndef PIPE_BUF 64 /* Get the PIPE_BUF from pathconf */ 65 #ifdef _PC_PIPE_BUF 66 #define PIPE_BUF pathconf(".", _PC_PIPE_BUF) 67 #else 68 #define PIPE_BUF 512 69 #endif 70 #endif 71 72 private void close_and_restore(const struct magic_set *, const char *, int, 73 const struct stat *); 74 private int unreadable_info(struct magic_set *, mode_t, const char *); 75 private const char* get_default_magic(void); 76 #ifndef COMPILE_ONLY 77 private const char *file_or_fd(struct magic_set *, const char *, int); 78 #endif 79 80 #ifndef STDIN_FILENO 81 #define STDIN_FILENO 0 82 #endif 83 84 #ifdef WIN32 85 /* HINSTANCE of this shared library. Needed for get_default_magic() */ 86 static HINSTANCE _w32_dll_instance = NULL; 87 88 static void 89 _w32_append_path(char **hmagicpath, const char *fmt, ...) 90 { 91 char *tmppath; 92 char *newpath; 93 va_list ap; 94 95 va_start(ap, fmt); 96 if (vasprintf(&tmppath, fmt, ap) < 0) { 97 va_end(ap); 98 return; 99 } 100 va_end(ap); 101 102 if (access(tmppath, R_OK) == -1) 103 goto out; 104 105 if (*hmagicpath == NULL) { 106 *hmagicpath = tmppath; 107 return; 108 } 109 110 if (asprintf(&newpath, "%s%c%s", *hmagicpath, PATHSEP, tmppath) < 0) 111 goto out; 112 113 free(*hmagicpath); 114 free(tmppath); 115 *hmagicpath = newpath; 116 return; 117 out: 118 free(tmppath); 119 } 120 121 static void 122 _w32_get_magic_relative_to(char **hmagicpath, HINSTANCE module) 123 { 124 static const char *trypaths[] = { 125 "%s/share/misc/magic.mgc", 126 "%s/magic.mgc", 127 }; 128 LPSTR dllpath; 129 size_t sp; 130 131 dllpath = calloc(MAX_PATH + 1, sizeof(*dllpath)); 132 133 if (!GetModuleFileNameA(module, dllpath, MAX_PATH)) 134 goto out; 135 136 PathRemoveFileSpecA(dllpath); 137 138 if (module) { 139 char exepath[MAX_PATH]; 140 GetModuleFileNameA(NULL, exepath, MAX_PATH); 141 PathRemoveFileSpecA(exepath); 142 if (stricmp(exepath, dllpath) == 0) 143 goto out; 144 } 145 146 sp = strlen(dllpath); 147 if (sp > 3 && stricmp(&dllpath[sp - 3], "bin") == 0) { 148 _w32_append_path(hmagicpath, 149 "%s/../share/misc/magic.mgc", dllpath); 150 goto out; 151 } 152 153 for (sp = 0; sp < __arraycount(trypaths); sp++) 154 _w32_append_path(hmagicpath, trypaths[sp], dllpath); 155 out: 156 free(dllpath); 157 } 158 159 /* Placate GCC by offering a sacrificial previous prototype */ 160 BOOL WINAPI DllMain(HINSTANCE, DWORD, LPVOID); 161 162 BOOL WINAPI 163 DllMain(HINSTANCE hinstDLL, DWORD fdwReason, 164 LPVOID lpvReserved __attribute__((__unused__))) 165 { 166 if (fdwReason == DLL_PROCESS_ATTACH) 167 _w32_dll_instance = hinstDLL; 168 return 1; 169 } 170 #endif 171 172 private const char * 173 get_default_magic(void) 174 { 175 static const char hmagic[] = "/.magic/magic.mgc"; 176 static char *default_magic; 177 char *home, *hmagicpath; 178 179 #ifndef WIN32 180 struct stat st; 181 182 if (default_magic) { 183 free(default_magic); 184 default_magic = NULL; 185 } 186 if ((home = getenv("HOME")) == NULL) 187 return MAGIC; 188 189 if (asprintf(&hmagicpath, "%s/.magic.mgc", home) < 0) 190 return MAGIC; 191 if (stat(hmagicpath, &st) == -1) { 192 free(hmagicpath); 193 if (asprintf(&hmagicpath, "%s/.magic", home) < 0) 194 return MAGIC; 195 if (stat(hmagicpath, &st) == -1) 196 goto out; 197 if (S_ISDIR(st.st_mode)) { 198 free(hmagicpath); 199 if (asprintf(&hmagicpath, "%s/%s", home, hmagic) < 0) 200 return MAGIC; 201 if (access(hmagicpath, R_OK) == -1) 202 goto out; 203 } 204 } 205 206 if (asprintf(&default_magic, "%s:%s", hmagicpath, MAGIC) < 0) 207 goto out; 208 free(hmagicpath); 209 return default_magic; 210 out: 211 default_magic = NULL; 212 free(hmagicpath); 213 return MAGIC; 214 #else 215 hmagicpath = NULL; 216 217 if (default_magic) { 218 free(default_magic); 219 default_magic = NULL; 220 } 221 222 /* First, try to get a magic file from user-application data */ 223 if ((home = getenv("LOCALAPPDATA")) != NULL) 224 _w32_append_path(&hmagicpath, "%s%s", home, hmagic); 225 226 /* Second, try to get a magic file from the user profile data */ 227 if ((home = getenv("USERPROFILE")) != NULL) 228 _w32_append_path(&hmagicpath, 229 "%s/Local Settings/Application Data%s", home, hmagic); 230 231 /* Third, try to get a magic file from Common Files */ 232 if ((home = getenv("COMMONPROGRAMFILES")) != NULL) 233 _w32_append_path(&hmagicpath, "%s%s", home, hmagic); 234 235 /* Fourth, try to get magic file relative to exe location */ 236 _w32_get_magic_relative_to(&hmagicpath, NULL); 237 238 /* Fifth, try to get magic file relative to dll location */ 239 _w32_get_magic_relative_to(&hmagicpath, _w32_dll_instance); 240 241 /* Avoid MAGIC constant - it likely points to a file within MSys tree */ 242 default_magic = hmagicpath; 243 return default_magic; 244 #endif 245 } 246 247 public const char * 248 magic_getpath(const char *magicfile, int action) 249 { 250 if (magicfile != NULL) 251 return magicfile; 252 253 magicfile = getenv("MAGIC"); 254 if (magicfile != NULL) 255 return magicfile; 256 257 return action == FILE_LOAD ? get_default_magic() : MAGIC; 258 } 259 260 public struct magic_set * 261 magic_open(int flags) 262 { 263 return file_ms_alloc(flags); 264 } 265 266 private int 267 unreadable_info(struct magic_set *ms, mode_t md, const char *file) 268 { 269 if (file) { 270 /* We cannot open it, but we were able to stat it. */ 271 if (access(file, W_OK) == 0) 272 if (file_printf(ms, "writable, ") == -1) 273 return -1; 274 if (access(file, X_OK) == 0) 275 if (file_printf(ms, "executable, ") == -1) 276 return -1; 277 } 278 if (S_ISREG(md)) 279 if (file_printf(ms, "regular file, ") == -1) 280 return -1; 281 if (file_printf(ms, "no read permission") == -1) 282 return -1; 283 return 0; 284 } 285 286 public void 287 magic_close(struct magic_set *ms) 288 { 289 if (ms == NULL) 290 return; 291 file_ms_free(ms); 292 } 293 294 /* 295 * load a magic file 296 */ 297 public int 298 magic_load(struct magic_set *ms, const char *magicfile) 299 { 300 if (ms == NULL) 301 return -1; 302 return file_apprentice(ms, magicfile, FILE_LOAD); 303 } 304 305 #ifndef COMPILE_ONLY 306 /* 307 * Install a set of compiled magic buffers. 308 */ 309 public int 310 magic_load_buffers(struct magic_set *ms, void **bufs, size_t *sizes, 311 size_t nbufs) 312 { 313 if (ms == NULL) 314 return -1; 315 return buffer_apprentice(ms, RCAST(struct magic **, bufs), 316 sizes, nbufs); 317 } 318 #endif 319 320 public int 321 magic_compile(struct magic_set *ms, const char *magicfile) 322 { 323 if (ms == NULL) 324 return -1; 325 return file_apprentice(ms, magicfile, FILE_COMPILE); 326 } 327 328 public int 329 magic_check(struct magic_set *ms, const char *magicfile) 330 { 331 if (ms == NULL) 332 return -1; 333 return file_apprentice(ms, magicfile, FILE_CHECK); 334 } 335 336 public int 337 magic_list(struct magic_set *ms, const char *magicfile) 338 { 339 if (ms == NULL) 340 return -1; 341 return file_apprentice(ms, magicfile, FILE_LIST); 342 } 343 344 private void 345 close_and_restore(const struct magic_set *ms, const char *name, int fd, 346 const struct stat *sb) 347 { 348 if (fd == STDIN_FILENO || name == NULL) 349 return; 350 (void) close(fd); 351 352 if ((ms->flags & MAGIC_PRESERVE_ATIME) != 0) { 353 /* 354 * Try to restore access, modification times if read it. 355 * This is really *bad* because it will modify the status 356 * time of the file... And of course this will affect 357 * backup programs 358 */ 359 #ifdef HAVE_UTIMES 360 struct timeval utsbuf[2]; 361 (void)memset(utsbuf, 0, sizeof(utsbuf)); 362 utsbuf[0].tv_sec = sb->st_atime; 363 utsbuf[1].tv_sec = sb->st_mtime; 364 365 (void) utimes(name, utsbuf); /* don't care if loses */ 366 #elif defined(HAVE_UTIME_H) || defined(HAVE_SYS_UTIME_H) 367 struct utimbuf utbuf; 368 369 (void)memset(&utbuf, 0, sizeof(utbuf)); 370 utbuf.actime = sb->st_atime; 371 utbuf.modtime = sb->st_mtime; 372 (void) utime(name, &utbuf); /* don't care if loses */ 373 #endif 374 } 375 } 376 377 #ifndef COMPILE_ONLY 378 379 /* 380 * find type of descriptor 381 */ 382 public const char * 383 magic_descriptor(struct magic_set *ms, int fd) 384 { 385 if (ms == NULL) 386 return NULL; 387 return file_or_fd(ms, NULL, fd); 388 } 389 390 /* 391 * find type of named file 392 */ 393 public const char * 394 magic_file(struct magic_set *ms, const char *inname) 395 { 396 if (ms == NULL) 397 return NULL; 398 return file_or_fd(ms, inname, STDIN_FILENO); 399 } 400 401 private const char * 402 file_or_fd(struct magic_set *ms, const char *inname, int fd) 403 { 404 int rv = -1; 405 unsigned char *buf; 406 struct stat sb; 407 ssize_t nbytes = 0; /* number of bytes read from a datafile */ 408 int ispipe = 0; 409 int okstat = 0; 410 off_t pos = CAST(off_t, -1); 411 412 if (file_reset(ms, 1) == -1) 413 goto out; 414 415 /* 416 * one extra for terminating '\0', and 417 * some overlapping space for matches near EOF 418 */ 419 #define SLOP (1 + sizeof(union VALUETYPE)) 420 if ((buf = CAST(unsigned char *, malloc(ms->bytes_max + SLOP))) == NULL) 421 return NULL; 422 423 switch (file_fsmagic(ms, inname, &sb)) { 424 case -1: /* error */ 425 goto done; 426 case 0: /* nothing found */ 427 break; 428 default: /* matched it and printed type */ 429 rv = 0; 430 goto done; 431 } 432 433 #ifdef WIN32 434 /* Place stdin in binary mode, so EOF (Ctrl+Z) doesn't stop early. */ 435 if (fd == STDIN_FILENO) 436 _setmode(STDIN_FILENO, O_BINARY); 437 #endif 438 if (inname != NULL) { 439 int flags = O_RDONLY|O_BINARY|O_NONBLOCK|O_CLOEXEC; 440 errno = 0; 441 if ((fd = open(inname, flags)) < 0) { 442 okstat = stat(inname, &sb) == 0; 443 if (okstat && S_ISFIFO(sb.st_mode)) 444 ispipe = 1; 445 #ifdef WIN32 446 /* 447 * Can't stat, can't open. It may have been opened in 448 * fsmagic, so if the user doesn't have read permission, 449 * allow it to say so; otherwise an error was probably 450 * displayed in fsmagic. 451 */ 452 if (!okstat && errno == EACCES) { 453 sb.st_mode = S_IFBLK; 454 okstat = 1; 455 } 456 #endif 457 if (okstat && 458 unreadable_info(ms, sb.st_mode, inname) == -1) 459 goto done; 460 rv = 0; 461 goto done; 462 } 463 #if O_CLOEXEC == 0 464 (void)fcntl(fd, F_SETFD, FD_CLOEXEC); 465 #endif 466 } 467 468 if (fd != -1) { 469 okstat = fstat(fd, &sb) == 0; 470 if (okstat && S_ISFIFO(sb.st_mode)) 471 ispipe = 1; 472 if (inname == NULL) 473 pos = lseek(fd, CAST(off_t, 0), SEEK_CUR); 474 } 475 476 /* 477 * try looking at the first ms->bytes_max bytes 478 */ 479 if (ispipe) { 480 if (fd != -1) { 481 ssize_t r = 0; 482 483 while ((r = sread(fd, RCAST(void *, &buf[nbytes]), 484 CAST(size_t, ms->bytes_max - nbytes), 1)) > 0) { 485 nbytes += r; 486 if (r < PIPE_BUF) break; 487 } 488 } 489 490 if (nbytes == 0 && inname) { 491 /* We can not read it, but we were able to stat it. */ 492 if (unreadable_info(ms, sb.st_mode, inname) == -1) 493 goto done; 494 rv = 0; 495 goto done; 496 } 497 498 } else if (fd != -1) { 499 /* Windows refuses to read from a big console buffer. */ 500 size_t howmany = 501 #if defined(WIN32) 502 _isatty(fd) ? 8 * 1024 : 503 #endif 504 ms->bytes_max; 505 if ((nbytes = read(fd, RCAST(void *, buf), howmany)) == -1) { 506 if (inname == NULL && fd != STDIN_FILENO) 507 file_error(ms, errno, "cannot read fd %d", fd); 508 else 509 file_error(ms, errno, "cannot read `%s'", 510 inname == NULL ? "/dev/stdin" : inname); 511 goto done; 512 } 513 } 514 515 (void)memset(buf + nbytes, 0, SLOP); /* NUL terminate */ 516 if (file_buffer(ms, fd, okstat ? &sb : NULL, inname, buf, CAST(size_t, nbytes)) == -1) 517 goto done; 518 rv = 0; 519 done: 520 free(buf); 521 if (fd != -1) { 522 if (pos != CAST(off_t, -1)) 523 (void)lseek(fd, pos, SEEK_SET); 524 close_and_restore(ms, inname, fd, &sb); 525 } 526 out: 527 return rv == 0 ? file_getbuffer(ms) : NULL; 528 } 529 530 531 public const char * 532 magic_buffer(struct magic_set *ms, const void *buf, size_t nb) 533 { 534 if (ms == NULL) 535 return NULL; 536 if (file_reset(ms, 1) == -1) 537 return NULL; 538 /* 539 * The main work is done here! 540 * We have the file name and/or the data buffer to be identified. 541 */ 542 if (file_buffer(ms, -1, NULL, NULL, buf, nb) == -1) { 543 return NULL; 544 } 545 return file_getbuffer(ms); 546 } 547 #endif 548 549 public const char * 550 magic_error(struct magic_set *ms) 551 { 552 if (ms == NULL) 553 return "Magic database is not open"; 554 return (ms->event_flags & EVENT_HAD_ERR) ? ms->o.buf : NULL; 555 } 556 557 public int 558 magic_errno(struct magic_set *ms) 559 { 560 if (ms == NULL) 561 return EINVAL; 562 return (ms->event_flags & EVENT_HAD_ERR) ? ms->error : 0; 563 } 564 565 public int 566 magic_getflags(struct magic_set *ms) 567 { 568 if (ms == NULL) 569 return -1; 570 571 return ms->flags; 572 } 573 574 public int 575 magic_setflags(struct magic_set *ms, int flags) 576 { 577 if (ms == NULL) 578 return -1; 579 #if !defined(HAVE_UTIME) && !defined(HAVE_UTIMES) 580 if (flags & MAGIC_PRESERVE_ATIME) 581 return -1; 582 #endif 583 ms->flags = flags; 584 return 0; 585 } 586 587 public int 588 magic_version(void) 589 { 590 return MAGIC_VERSION; 591 } 592 593 public int 594 magic_setparam(struct magic_set *ms, int param, const void *val) 595 { 596 if (ms == NULL) 597 return -1; 598 switch (param) { 599 case MAGIC_PARAM_INDIR_MAX: 600 ms->indir_max = CAST(uint16_t, *CAST(const size_t *, val)); 601 return 0; 602 case MAGIC_PARAM_NAME_MAX: 603 ms->name_max = CAST(uint16_t, *CAST(const size_t *, val)); 604 return 0; 605 case MAGIC_PARAM_ELF_PHNUM_MAX: 606 ms->elf_phnum_max = CAST(uint16_t, *CAST(const size_t *, val)); 607 return 0; 608 case MAGIC_PARAM_ELF_SHNUM_MAX: 609 ms->elf_shnum_max = CAST(uint16_t, *CAST(const size_t *, val)); 610 return 0; 611 case MAGIC_PARAM_ELF_NOTES_MAX: 612 ms->elf_notes_max = CAST(uint16_t, *CAST(const size_t *, val)); 613 return 0; 614 case MAGIC_PARAM_REGEX_MAX: 615 ms->regex_max = CAST(uint16_t, *CAST(const size_t *, val)); 616 return 0; 617 case MAGIC_PARAM_BYTES_MAX: 618 ms->bytes_max = *CAST(const size_t *, val); 619 return 0; 620 case MAGIC_PARAM_ENCODING_MAX: 621 ms->encoding_max = *CAST(const size_t *, val); 622 return 0; 623 default: 624 errno = EINVAL; 625 return -1; 626 } 627 } 628 629 public int 630 magic_getparam(struct magic_set *ms, int param, void *val) 631 { 632 if (ms == NULL) 633 return -1; 634 switch (param) { 635 case MAGIC_PARAM_INDIR_MAX: 636 *CAST(size_t *, val) = ms->indir_max; 637 return 0; 638 case MAGIC_PARAM_NAME_MAX: 639 *CAST(size_t *, val) = ms->name_max; 640 return 0; 641 case MAGIC_PARAM_ELF_PHNUM_MAX: 642 *CAST(size_t *, val) = ms->elf_phnum_max; 643 return 0; 644 case MAGIC_PARAM_ELF_SHNUM_MAX: 645 *CAST(size_t *, val) = ms->elf_shnum_max; 646 return 0; 647 case MAGIC_PARAM_ELF_NOTES_MAX: 648 *CAST(size_t *, val) = ms->elf_notes_max; 649 return 0; 650 case MAGIC_PARAM_REGEX_MAX: 651 *CAST(size_t *, val) = ms->regex_max; 652 return 0; 653 case MAGIC_PARAM_BYTES_MAX: 654 *CAST(size_t *, val) = ms->bytes_max; 655 return 0; 656 case MAGIC_PARAM_ENCODING_MAX: 657 *CAST(size_t *, val) = ms->encoding_max; 658 return 0; 659 default: 660 errno = EINVAL; 661 return -1; 662 } 663 } 664