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 #include "file.h" 28 29 #ifndef lint 30 FILE_RCSID("@(#)$File: funcs.c,v 1.84 2015/09/10 13:32:19 christos Exp $") 31 #endif /* lint */ 32 33 #include "magic.h" 34 #include <assert.h> 35 #include <stdarg.h> 36 #include <stdlib.h> 37 #include <string.h> 38 #include <ctype.h> 39 #if defined(HAVE_WCHAR_H) 40 #include <wchar.h> 41 #endif 42 #if defined(HAVE_WCTYPE_H) 43 #include <wctype.h> 44 #endif 45 #if defined(HAVE_LIMITS_H) 46 #include <limits.h> 47 #endif 48 49 #ifndef SIZE_MAX 50 #define SIZE_MAX ((size_t)~0) 51 #endif 52 53 /* 54 * Like printf, only we append to a buffer. 55 */ 56 protected int 57 file_vprintf(struct magic_set *ms, const char *fmt, va_list ap) 58 { 59 int len; 60 char *buf, *newstr; 61 62 if (ms->event_flags & EVENT_HAD_ERR) 63 return 0; 64 len = vasprintf(&buf, fmt, ap); 65 if (len < 0) 66 goto out; 67 68 if (ms->o.buf != NULL) { 69 len = asprintf(&newstr, "%s%s", ms->o.buf, buf); 70 free(buf); 71 if (len < 0) 72 goto out; 73 free(ms->o.buf); 74 buf = newstr; 75 } 76 ms->o.buf = buf; 77 return 0; 78 out: 79 file_error(ms, errno, "vasprintf failed"); 80 return -1; 81 } 82 83 protected int 84 file_printf(struct magic_set *ms, const char *fmt, ...) 85 { 86 int rv; 87 va_list ap; 88 89 va_start(ap, fmt); 90 rv = file_vprintf(ms, fmt, ap); 91 va_end(ap); 92 return rv; 93 } 94 95 /* 96 * error - print best error message possible 97 */ 98 /*VARARGS*/ 99 __attribute__((__format__(__printf__, 3, 0))) 100 private void 101 file_error_core(struct magic_set *ms, int error, const char *f, va_list va, 102 size_t lineno) 103 { 104 /* Only the first error is ok */ 105 if (ms->event_flags & EVENT_HAD_ERR) 106 return; 107 if (lineno != 0) { 108 free(ms->o.buf); 109 ms->o.buf = NULL; 110 file_printf(ms, "line %" SIZE_T_FORMAT "u:", lineno); 111 } 112 if (ms->o.buf && *ms->o.buf) 113 file_printf(ms, " "); 114 file_vprintf(ms, f, va); 115 if (error > 0) 116 file_printf(ms, " (%s)", strerror(error)); 117 ms->event_flags |= EVENT_HAD_ERR; 118 ms->error = error; 119 } 120 121 /*VARARGS*/ 122 protected void 123 file_error(struct magic_set *ms, int error, const char *f, ...) 124 { 125 va_list va; 126 va_start(va, f); 127 file_error_core(ms, error, f, va, 0); 128 va_end(va); 129 } 130 131 /* 132 * Print an error with magic line number. 133 */ 134 /*VARARGS*/ 135 protected void 136 file_magerror(struct magic_set *ms, const char *f, ...) 137 { 138 va_list va; 139 va_start(va, f); 140 file_error_core(ms, 0, f, va, ms->line); 141 va_end(va); 142 } 143 144 protected void 145 file_oomem(struct magic_set *ms, size_t len) 146 { 147 file_error(ms, errno, "cannot allocate %" SIZE_T_FORMAT "u bytes", 148 len); 149 } 150 151 protected void 152 file_badseek(struct magic_set *ms) 153 { 154 file_error(ms, errno, "error seeking"); 155 } 156 157 protected void 158 file_badread(struct magic_set *ms) 159 { 160 file_error(ms, errno, "error reading"); 161 } 162 163 #ifndef COMPILE_ONLY 164 165 static int 166 checkdone(struct magic_set *ms, int *rv) 167 { 168 if ((ms->flags & MAGIC_CONTINUE) == 0) 169 return 1; 170 if (file_printf(ms, "\n- ") == -1) 171 *rv = -1; 172 return 0; 173 } 174 175 /*ARGSUSED*/ 176 protected int 177 file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((__unused__)), 178 const void *buf, size_t nb) 179 { 180 int m = 0, rv = 0, looks_text = 0; 181 int mime = ms->flags & MAGIC_MIME; 182 const unsigned char *ubuf = CAST(const unsigned char *, buf); 183 unichar *u8buf = NULL; 184 size_t ulen; 185 const char *code = NULL; 186 const char *code_mime = "binary"; 187 const char *type = "application/octet-stream"; 188 const char *def = "data"; 189 const char *ftype = NULL; 190 191 if (nb == 0) { 192 def = "empty"; 193 type = "application/x-empty"; 194 goto simple; 195 } else if (nb == 1) { 196 def = "very short file (no magic)"; 197 goto simple; 198 } 199 200 if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) { 201 looks_text = file_encoding(ms, ubuf, nb, &u8buf, &ulen, 202 &code, &code_mime, &ftype); 203 } 204 205 #ifdef __EMX__ 206 if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) { 207 m = file_os2_apptype(ms, inname, buf, nb); 208 if ((ms->flags & MAGIC_DEBUG) != 0) 209 (void)fprintf(stderr, "[try os2_apptype %d]\n", m); 210 switch (m) { 211 case -1: 212 return -1; 213 case 0: 214 break; 215 default: 216 return 1; 217 } 218 } 219 #endif 220 #if HAVE_FORK 221 /* try compression stuff */ 222 if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0) { 223 m = file_zmagic(ms, fd, inname, ubuf, nb); 224 if ((ms->flags & MAGIC_DEBUG) != 0) 225 (void)fprintf(stderr, "[try zmagic %d]\n", m); 226 if (m) { 227 goto done_encoding; 228 } 229 } 230 #endif 231 /* Check if we have a tar file */ 232 if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0) { 233 m = file_is_tar(ms, ubuf, nb); 234 if ((ms->flags & MAGIC_DEBUG) != 0) 235 (void)fprintf(stderr, "[try tar %d]\n", m); 236 if (m) { 237 if (checkdone(ms, &rv)) 238 goto done; 239 } 240 } 241 242 /* Check if we have a CDF file */ 243 if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0) { 244 m = file_trycdf(ms, fd, ubuf, nb); 245 if ((ms->flags & MAGIC_DEBUG) != 0) 246 (void)fprintf(stderr, "[try cdf %d]\n", m); 247 if (m) { 248 if (checkdone(ms, &rv)) 249 goto done; 250 } 251 } 252 253 /* try soft magic tests */ 254 if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0) 255 m = file_softmagic(ms, ubuf, nb, 0, NULL, BINTEST, looks_text); 256 if ((ms->flags & MAGIC_DEBUG) != 0) 257 (void)fprintf(stderr, "[try softmagic %d]\n", m); 258 if (m) { 259 #ifdef BUILTIN_ELF 260 if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 && 261 nb > 5 && fd != -1) { 262 /* 263 * We matched something in the file, so this 264 * *might* be an ELF file, and the file is at 265 * least 5 bytes long, so if it's an ELF file 266 * it has at least one byte past the ELF magic 267 * number - try extracting information from the 268 * ELF headers that cannot easily * be 269 * extracted with rules in the magic file. 270 */ 271 m = file_tryelf(ms, fd, ubuf, nb); 272 if ((ms->flags & MAGIC_DEBUG) != 0) 273 (void)fprintf(stderr, "[try elf %d]\n", 274 m); 275 } 276 #endif 277 if (checkdone(ms, &rv)) 278 goto done; 279 } 280 281 /* try text properties */ 282 if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) { 283 284 m = file_ascmagic(ms, ubuf, nb, looks_text); 285 if ((ms->flags & MAGIC_DEBUG) != 0) 286 (void)fprintf(stderr, "[try ascmagic %d]\n", m); 287 if (m) { 288 if (checkdone(ms, &rv)) 289 goto done; 290 } 291 } 292 293 simple: 294 /* give up */ 295 m = 1; 296 if ((!mime || (mime & MAGIC_MIME_TYPE)) && 297 file_printf(ms, "%s", mime ? type : def) == -1) { 298 rv = -1; 299 } 300 done: 301 if ((ms->flags & MAGIC_MIME_ENCODING) != 0) { 302 if (ms->flags & MAGIC_MIME_TYPE) 303 if (file_printf(ms, "; charset=") == -1) 304 rv = -1; 305 if (file_printf(ms, "%s", code_mime) == -1) 306 rv = -1; 307 } 308 #if HAVE_FORK 309 done_encoding: 310 #endif 311 free(u8buf); 312 if (rv) 313 return rv; 314 315 return m; 316 } 317 #endif 318 319 protected int 320 file_reset(struct magic_set *ms) 321 { 322 if (ms->mlist[0] == NULL) { 323 file_error(ms, 0, "no magic files loaded"); 324 return -1; 325 } 326 if (ms->o.buf) { 327 free(ms->o.buf); 328 ms->o.buf = NULL; 329 } 330 if (ms->o.pbuf) { 331 free(ms->o.pbuf); 332 ms->o.pbuf = NULL; 333 } 334 ms->event_flags &= ~EVENT_HAD_ERR; 335 ms->error = -1; 336 return 0; 337 } 338 339 #define OCTALIFY(n, o) \ 340 /*LINTED*/ \ 341 (void)(*(n)++ = '\\', \ 342 *(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \ 343 *(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \ 344 *(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \ 345 (o)++) 346 347 protected const char * 348 file_getbuffer(struct magic_set *ms) 349 { 350 char *pbuf, *op, *np; 351 size_t psize, len; 352 353 if (ms->event_flags & EVENT_HAD_ERR) 354 return NULL; 355 356 if (ms->flags & MAGIC_RAW) 357 return ms->o.buf; 358 359 if (ms->o.buf == NULL) 360 return NULL; 361 362 /* * 4 is for octal representation, + 1 is for NUL */ 363 len = strlen(ms->o.buf); 364 if (len > (SIZE_MAX - 1) / 4) { 365 file_oomem(ms, len); 366 return NULL; 367 } 368 psize = len * 4 + 1; 369 if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) { 370 file_oomem(ms, psize); 371 return NULL; 372 } 373 ms->o.pbuf = pbuf; 374 375 #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH) 376 { 377 mbstate_t state; 378 wchar_t nextchar; 379 int mb_conv = 1; 380 size_t bytesconsumed; 381 char *eop; 382 (void)memset(&state, 0, sizeof(mbstate_t)); 383 384 np = ms->o.pbuf; 385 op = ms->o.buf; 386 eop = op + len; 387 388 while (op < eop) { 389 bytesconsumed = mbrtowc(&nextchar, op, 390 (size_t)(eop - op), &state); 391 if (bytesconsumed == (size_t)(-1) || 392 bytesconsumed == (size_t)(-2)) { 393 mb_conv = 0; 394 break; 395 } 396 397 if (iswprint(nextchar)) { 398 (void)memcpy(np, op, bytesconsumed); 399 op += bytesconsumed; 400 np += bytesconsumed; 401 } else { 402 while (bytesconsumed-- > 0) 403 OCTALIFY(np, op); 404 } 405 } 406 *np = '\0'; 407 408 /* Parsing succeeded as a multi-byte sequence */ 409 if (mb_conv != 0) 410 return ms->o.pbuf; 411 } 412 #endif 413 414 for (np = ms->o.pbuf, op = ms->o.buf; *op;) { 415 if (isprint((unsigned char)*op)) { 416 *np++ = *op++; 417 } else { 418 OCTALIFY(np, op); 419 } 420 } 421 *np = '\0'; 422 return ms->o.pbuf; 423 } 424 425 protected int 426 file_check_mem(struct magic_set *ms, unsigned int level) 427 { 428 size_t len; 429 430 if (level >= ms->c.len) { 431 len = (ms->c.len = 20 + level) * sizeof(*ms->c.li); 432 ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ? 433 malloc(len) : 434 realloc(ms->c.li, len)); 435 if (ms->c.li == NULL) { 436 file_oomem(ms, len); 437 return -1; 438 } 439 } 440 ms->c.li[level].got_match = 0; 441 #ifdef ENABLE_CONDITIONALS 442 ms->c.li[level].last_match = 0; 443 ms->c.li[level].last_cond = COND_NONE; 444 #endif /* ENABLE_CONDITIONALS */ 445 return 0; 446 } 447 448 protected size_t 449 file_printedlen(const struct magic_set *ms) 450 { 451 return ms->o.buf == NULL ? 0 : strlen(ms->o.buf); 452 } 453 454 protected int 455 file_replace(struct magic_set *ms, const char *pat, const char *rep) 456 { 457 file_regex_t rx; 458 int rc, rv = -1; 459 460 rc = file_regcomp(&rx, pat, REG_EXTENDED); 461 if (rc) { 462 file_regerror(&rx, rc, ms); 463 } else { 464 regmatch_t rm; 465 int nm = 0; 466 while (file_regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) { 467 ms->o.buf[rm.rm_so] = '\0'; 468 if (file_printf(ms, "%s%s", rep, 469 rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1) 470 goto out; 471 nm++; 472 } 473 rv = nm; 474 } 475 out: 476 file_regfree(&rx); 477 return rv; 478 } 479 480 protected int 481 file_regcomp(file_regex_t *rx, const char *pat, int flags) 482 { 483 #ifdef USE_C_LOCALE 484 rx->c_lc_ctype = newlocale(LC_CTYPE_MASK, "C", 0); 485 assert(rx->c_lc_ctype != NULL); 486 rx->old_lc_ctype = uselocale(rx->c_lc_ctype); 487 assert(rx->old_lc_ctype != NULL); 488 #endif 489 rx->pat = pat; 490 491 return rx->rc = regcomp(&rx->rx, pat, flags); 492 } 493 494 protected int 495 file_regexec(file_regex_t *rx, const char *str, size_t nmatch, 496 regmatch_t* pmatch, int eflags) 497 { 498 assert(rx->rc == 0); 499 return regexec(&rx->rx, str, nmatch, pmatch, eflags); 500 } 501 502 protected void 503 file_regfree(file_regex_t *rx) 504 { 505 if (rx->rc == 0) 506 regfree(&rx->rx); 507 #ifdef USE_C_LOCALE 508 (void)uselocale(rx->old_lc_ctype); 509 freelocale(rx->c_lc_ctype); 510 #endif 511 } 512 513 protected void 514 file_regerror(file_regex_t *rx, int rc, struct magic_set *ms) 515 { 516 char errmsg[512]; 517 518 (void)regerror(rc, &rx->rx, errmsg, sizeof(errmsg)); 519 file_magerror(ms, "regex error %d for `%s', (%s)", rc, rx->pat, 520 errmsg); 521 } 522 523 protected file_pushbuf_t * 524 file_push_buffer(struct magic_set *ms) 525 { 526 file_pushbuf_t *pb; 527 528 if (ms->event_flags & EVENT_HAD_ERR) 529 return NULL; 530 531 if ((pb = (CAST(file_pushbuf_t *, malloc(sizeof(*pb))))) == NULL) 532 return NULL; 533 534 pb->buf = ms->o.buf; 535 pb->offset = ms->offset; 536 537 ms->o.buf = NULL; 538 ms->offset = 0; 539 540 return pb; 541 } 542 543 protected char * 544 file_pop_buffer(struct magic_set *ms, file_pushbuf_t *pb) 545 { 546 char *rbuf; 547 548 if (ms->event_flags & EVENT_HAD_ERR) { 549 free(pb->buf); 550 free(pb); 551 return NULL; 552 } 553 554 rbuf = ms->o.buf; 555 556 ms->o.buf = pb->buf; 557 ms->offset = pb->offset; 558 559 free(pb); 560 return rbuf; 561 } 562 563 /* 564 * convert string to ascii printable format. 565 */ 566 protected char * 567 file_printable(char *buf, size_t bufsiz, const char *str) 568 { 569 char *ptr, *eptr; 570 const unsigned char *s = (const unsigned char *)str; 571 572 for (ptr = buf, eptr = ptr + bufsiz - 1; ptr < eptr && *s; s++) { 573 if (isprint(*s)) { 574 *ptr++ = *s; 575 continue; 576 } 577 if (ptr >= eptr - 3) 578 break; 579 *ptr++ = '\\'; 580 *ptr++ = ((CAST(unsigned int, *s) >> 6) & 7) + '0'; 581 *ptr++ = ((CAST(unsigned int, *s) >> 3) & 7) + '0'; 582 *ptr++ = ((CAST(unsigned int, *s) >> 0) & 7) + '0'; 583 } 584 *ptr = '\0'; 585 return buf; 586 } 587