1 /* $NetBSD: magic.c,v 1.4 2011/09/16 21:06:26 christos Exp $ */ 2 3 /* 4 * Copyright (c) Christos Zoulas 2003. 5 * All Rights Reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice immediately at the beginning of the file, without modification, 12 * this list of conditions, and the following disclaimer. 13 * 2. Redistributions in binary form must reproduce the above copyright 14 * notice, this list of conditions and the following disclaimer in the 15 * documentation and/or other materials provided with the distribution. 16 * 17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 20 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR 21 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 27 * SUCH DAMAGE. 28 */ 29 30 #ifdef WIN32 31 #include <windows.h> 32 #include <shlwapi.h> 33 #endif 34 35 #include "file.h" 36 37 #ifndef lint 38 #if 0 39 FILE_RCSID("@(#)$File: magic.c,v 1.74 2011/05/26 01:27:59 christos Exp $") 40 #else 41 __RCSID("$NetBSD: magic.c,v 1.4 2011/09/16 21:06:26 christos Exp $"); 42 #endif 43 #endif /* lint */ 44 45 #include "magic.h" 46 47 #include <stdlib.h> 48 #include <unistd.h> 49 #include <string.h> 50 #ifdef QUICK 51 #include <sys/mman.h> 52 #endif 53 #ifdef HAVE_LIMITS_H 54 #include <limits.h> /* for PIPE_BUF */ 55 #endif 56 57 #if defined(HAVE_UTIMES) 58 # include <sys/time.h> 59 #elif defined(HAVE_UTIME) 60 # if defined(HAVE_SYS_UTIME_H) 61 # include <sys/utime.h> 62 # elif defined(HAVE_UTIME_H) 63 # include <utime.h> 64 # endif 65 #endif 66 67 #ifdef HAVE_UNISTD_H 68 #include <unistd.h> /* for read() */ 69 #endif 70 71 #ifndef PIPE_BUF 72 /* Get the PIPE_BUF from pathconf */ 73 #ifdef _PC_PIPE_BUF 74 #define PIPE_BUF pathconf(".", _PC_PIPE_BUF) 75 #else 76 #define PIPE_BUF 512 77 #endif 78 #endif 79 80 private void free_mlist(struct mlist *); 81 private void close_and_restore(const struct magic_set *, const char *, int, 82 const struct stat *); 83 private int unreadable_info(struct magic_set *, mode_t, const char *); 84 private const char* get_default_magic(void); 85 #ifndef COMPILE_ONLY 86 private const char *file_or_fd(struct magic_set *, const char *, int); 87 #endif 88 89 #ifndef STDIN_FILENO 90 #define STDIN_FILENO 0 91 #endif 92 93 private const char * 94 get_default_magic(void) 95 { 96 static const char hmagic[] = "/.magic/magic.mgc"; 97 static char *default_magic; 98 char *home, *hmagicpath; 99 100 #ifndef WIN32 101 struct stat st; 102 103 if (default_magic) { 104 free(default_magic); 105 default_magic = NULL; 106 } 107 if ((home = getenv("HOME")) == NULL) 108 return MAGIC; 109 110 if (asprintf(&hmagicpath, "%s/.magic", home) < 0) 111 return MAGIC; 112 if (stat(hmagicpath, &st) == -1) 113 goto out; 114 if (S_ISDIR(st.st_mode)) { 115 free(hmagicpath); 116 if (asprintf(&hmagicpath, "%s/%s", home, hmagic) < 0) 117 return MAGIC; 118 if (access(hmagicpath, R_OK) == -1) 119 goto out; 120 } 121 122 if (asprintf(&default_magic, "%s:%s", hmagicpath, MAGIC) < 0) 123 goto out; 124 free(hmagicpath); 125 return default_magic; 126 out: 127 default_magic = NULL; 128 free(hmagicpath); 129 return MAGIC; 130 #else 131 char *hmagicp = hmagicpath; 132 char *tmppath = 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 LPTSTR 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 struct magic_set *ms; 220 size_t len; 221 222 if ((ms = CAST(struct magic_set *, calloc((size_t)1, 223 sizeof(struct magic_set)))) == NULL) 224 return NULL; 225 226 if (magic_setflags(ms, flags) == -1) { 227 errno = EINVAL; 228 goto free; 229 } 230 231 ms->o.buf = ms->o.pbuf = NULL; 232 len = (ms->c.len = 10) * sizeof(*ms->c.li); 233 234 if ((ms->c.li = CAST(struct level_info *, malloc(len))) == NULL) 235 goto free; 236 237 ms->event_flags = 0; 238 ms->error = -1; 239 ms->mlist = NULL; 240 ms->file = "unknown"; 241 ms->line = 0; 242 return ms; 243 free: 244 free(ms); 245 return NULL; 246 } 247 248 private void 249 free_mlist(struct mlist *mlist) 250 { 251 struct mlist *ml; 252 253 if (mlist == NULL) 254 return; 255 256 for (ml = mlist->next; ml != mlist;) { 257 struct mlist *next = ml->next; 258 struct magic *mg = ml->magic; 259 file_delmagic(mg, ml->mapped, ml->nmagic); 260 free(ml); 261 ml = next; 262 } 263 free(ml); 264 } 265 266 private int 267 unreadable_info(struct magic_set *ms, mode_t md, const char *file) 268 { 269 /* We cannot open it, but we were able to stat it. */ 270 if (access(file, W_OK) == 0) 271 if (file_printf(ms, "writable, ") == -1) 272 return -1; 273 if (access(file, X_OK) == 0) 274 if (file_printf(ms, "executable, ") == -1) 275 return -1; 276 if (S_ISREG(md)) 277 if (file_printf(ms, "regular file, ") == -1) 278 return -1; 279 if (file_printf(ms, "no read permission") == -1) 280 return -1; 281 return 0; 282 } 283 284 public void 285 magic_close(struct magic_set *ms) 286 { 287 free_mlist(ms->mlist); 288 free(ms->o.pbuf); 289 free(ms->o.buf); 290 free(ms->c.li); 291 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 struct mlist *ml = file_apprentice(ms, magicfile, FILE_LOAD); 301 if (ml) { 302 free_mlist(ms->mlist); 303 ms->mlist = ml; 304 return 0; 305 } 306 return -1; 307 } 308 309 public int 310 magic_compile(struct magic_set *ms, const char *magicfile) 311 { 312 struct mlist *ml = file_apprentice(ms, magicfile, FILE_COMPILE); 313 free_mlist(ml); 314 return ml ? 0 : -1; 315 } 316 317 public int 318 magic_check(struct magic_set *ms, const char *magicfile) 319 { 320 struct mlist *ml = file_apprentice(ms, magicfile, FILE_CHECK); 321 free_mlist(ml); 322 return ml ? 0 : -1; 323 } 324 325 public int 326 magic_list(struct magic_set *ms, const char *magicfile) 327 { 328 struct mlist *ml = file_apprentice(ms, magicfile, FILE_LIST); 329 free_mlist(ml); 330 return ml ? 0 : -1; 331 } 332 333 private void 334 close_and_restore(const struct magic_set *ms, const char *name, int fd, 335 const struct stat *sb) 336 { 337 if (fd == STDIN_FILENO) 338 return; 339 (void) close(fd); 340 341 if ((ms->flags & MAGIC_PRESERVE_ATIME) != 0) { 342 /* 343 * Try to restore access, modification times if read it. 344 * This is really *bad* because it will modify the status 345 * time of the file... And of course this will affect 346 * backup programs 347 */ 348 #ifdef HAVE_UTIMES 349 struct timeval utsbuf[2]; 350 (void)memset(utsbuf, 0, sizeof(utsbuf)); 351 utsbuf[0].tv_sec = sb->st_atime; 352 utsbuf[1].tv_sec = sb->st_mtime; 353 354 (void) utimes(name, utsbuf); /* don't care if loses */ 355 #elif defined(HAVE_UTIME_H) || defined(HAVE_SYS_UTIME_H) 356 struct utimbuf utbuf; 357 358 (void)memset(&utbuf, 0, sizeof(utbuf)); 359 utbuf.actime = sb->st_atime; 360 utbuf.modtime = sb->st_mtime; 361 (void) utime(name, &utbuf); /* don't care if loses */ 362 #endif 363 } 364 } 365 366 #ifndef COMPILE_ONLY 367 368 /* 369 * find type of descriptor 370 */ 371 public const char * 372 magic_descriptor(struct magic_set *ms, int fd) 373 { 374 return file_or_fd(ms, NULL, fd); 375 } 376 377 /* 378 * find type of named file 379 */ 380 public const char * 381 magic_file(struct magic_set *ms, const char *inname) 382 { 383 return file_or_fd(ms, inname, STDIN_FILENO); 384 } 385 386 private const char * 387 file_or_fd(struct magic_set *ms, const char *inname, int fd) 388 { 389 int rv = -1; 390 unsigned char *buf; 391 struct stat sb; 392 ssize_t nbytes = 0; /* number of bytes read from a datafile */ 393 int ispipe = 0; 394 395 /* 396 * one extra for terminating '\0', and 397 * some overlapping space for matches near EOF 398 */ 399 #define SLOP (1 + sizeof(union VALUETYPE)) 400 if ((buf = CAST(unsigned char *, malloc(HOWMANY + SLOP))) == NULL) 401 return NULL; 402 403 if (file_reset(ms) == -1) 404 goto done; 405 406 switch (file_fsmagic(ms, inname, &sb)) { 407 case -1: /* error */ 408 goto done; 409 case 0: /* nothing found */ 410 break; 411 default: /* matched it and printed type */ 412 rv = 0; 413 goto done; 414 } 415 416 if (inname == NULL) { 417 if (fstat(fd, &sb) == 0 && S_ISFIFO(sb.st_mode)) 418 ispipe = 1; 419 } else { 420 int flags = O_RDONLY|O_BINARY; 421 422 if (stat(inname, &sb) == 0 && S_ISFIFO(sb.st_mode)) { 423 #ifdef O_NONBLOCK 424 flags |= O_NONBLOCK; 425 #endif 426 ispipe = 1; 427 } 428 429 errno = 0; 430 if ((fd = open(inname, flags)) < 0) { 431 if (unreadable_info(ms, sb.st_mode, inname) == -1) 432 goto done; 433 rv = 0; 434 goto done; 435 } 436 #ifdef O_NONBLOCK 437 if ((flags = fcntl(fd, F_GETFL)) != -1) { 438 flags &= ~O_NONBLOCK; 439 (void)fcntl(fd, F_SETFL, flags); 440 } 441 #endif 442 } 443 444 /* 445 * try looking at the first HOWMANY bytes 446 */ 447 if (ispipe) { 448 ssize_t r = 0; 449 450 while ((r = sread(fd, (void *)&buf[nbytes], 451 (size_t)(HOWMANY - nbytes), 1)) > 0) { 452 nbytes += r; 453 if (r < PIPE_BUF) break; 454 } 455 456 if (nbytes == 0) { 457 /* We can not read it, but we were able to stat it. */ 458 if (unreadable_info(ms, sb.st_mode, inname) == -1) 459 goto done; 460 rv = 0; 461 goto done; 462 } 463 464 } else { 465 if ((nbytes = read(fd, (char *)buf, HOWMANY)) == -1) { 466 file_error(ms, errno, "cannot read `%s'", inname); 467 goto done; 468 } 469 } 470 471 (void)memset(buf + nbytes, 0, SLOP); /* NUL terminate */ 472 if (file_buffer(ms, fd, inname, buf, (size_t)nbytes) == -1) 473 goto done; 474 rv = 0; 475 done: 476 free(buf); 477 close_and_restore(ms, inname, fd, &sb); 478 return rv == 0 ? file_getbuffer(ms) : NULL; 479 } 480 481 482 public const char * 483 magic_buffer(struct magic_set *ms, const void *buf, size_t nb) 484 { 485 if (file_reset(ms) == -1) 486 return NULL; 487 /* 488 * The main work is done here! 489 * We have the file name and/or the data buffer to be identified. 490 */ 491 if (file_buffer(ms, -1, NULL, buf, nb) == -1) { 492 return NULL; 493 } 494 return file_getbuffer(ms); 495 } 496 #endif 497 498 public const char * 499 magic_error(struct magic_set *ms) 500 { 501 return (ms->event_flags & EVENT_HAD_ERR) ? ms->o.buf : NULL; 502 } 503 504 public int 505 magic_errno(struct magic_set *ms) 506 { 507 return (ms->event_flags & EVENT_HAD_ERR) ? ms->error : 0; 508 } 509 510 public int 511 magic_setflags(struct magic_set *ms, int flags) 512 { 513 #if !defined(HAVE_UTIME) && !defined(HAVE_UTIMES) 514 if (flags & MAGIC_PRESERVE_ATIME) 515 return -1; 516 #endif 517 ms->flags = flags; 518 return 0; 519 } 520