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 1.82 2015/06/03 18:01:20 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 file_vprintf(ms, f, va); 113 if (error > 0) 114 file_printf(ms, " (%s)", strerror(error)); 115 ms->event_flags |= EVENT_HAD_ERR; 116 ms->error = error; 117 } 118 119 /*VARARGS*/ 120 protected void 121 file_error(struct magic_set *ms, int error, const char *f, ...) 122 { 123 va_list va; 124 va_start(va, f); 125 file_error_core(ms, error, f, va, 0); 126 va_end(va); 127 } 128 129 /* 130 * Print an error with magic line number. 131 */ 132 /*VARARGS*/ 133 protected void 134 file_magerror(struct magic_set *ms, const char *f, ...) 135 { 136 va_list va; 137 va_start(va, f); 138 file_error_core(ms, 0, f, va, ms->line); 139 va_end(va); 140 } 141 142 protected void 143 file_oomem(struct magic_set *ms, size_t len) 144 { 145 file_error(ms, errno, "cannot allocate %" SIZE_T_FORMAT "u bytes", 146 len); 147 } 148 149 protected void 150 file_badseek(struct magic_set *ms) 151 { 152 file_error(ms, errno, "error seeking"); 153 } 154 155 protected void 156 file_badread(struct magic_set *ms) 157 { 158 file_error(ms, errno, "error reading"); 159 } 160 161 #ifndef COMPILE_ONLY 162 163 static int 164 checkdone(struct magic_set *ms, int *rv) 165 { 166 if ((ms->flags & MAGIC_CONTINUE) == 0) 167 return 1; 168 if (file_printf(ms, "\n- ") == -1) 169 *rv = -1; 170 return 0; 171 } 172 173 /*ARGSUSED*/ 174 protected int 175 file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((__unused__)), 176 const void *buf, size_t nb) 177 { 178 int m = 0, rv = 0, looks_text = 0; 179 int mime = ms->flags & MAGIC_MIME; 180 const unsigned char *ubuf = CAST(const unsigned char *, buf); 181 unichar *u8buf = NULL; 182 size_t ulen; 183 const char *code = NULL; 184 const char *code_mime = "binary"; 185 const char *type = "application/octet-stream"; 186 const char *def = "data"; 187 const char *ftype = NULL; 188 189 if (nb == 0) { 190 def = "empty"; 191 type = "application/x-empty"; 192 goto simple; 193 } else if (nb == 1) { 194 def = "very short file (no magic)"; 195 goto simple; 196 } 197 198 if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) { 199 looks_text = file_encoding(ms, ubuf, nb, &u8buf, &ulen, 200 &code, &code_mime, &ftype); 201 } 202 203 #ifdef __EMX__ 204 if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) { 205 switch (file_os2_apptype(ms, inname, buf, nb)) { 206 case -1: 207 return -1; 208 case 0: 209 break; 210 default: 211 return 1; 212 } 213 } 214 #endif 215 #if HAVE_FORK 216 /* try compression stuff */ 217 if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0) 218 if ((m = file_zmagic(ms, fd, inname, ubuf, nb)) != 0) { 219 if ((ms->flags & MAGIC_DEBUG) != 0) 220 (void)fprintf(stderr, "zmagic %d\n", m); 221 goto done_encoding; 222 } 223 #endif 224 /* Check if we have a tar file */ 225 if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0) 226 if ((m = file_is_tar(ms, ubuf, nb)) != 0) { 227 if ((ms->flags & MAGIC_DEBUG) != 0) 228 (void)fprintf(stderr, "tar %d\n", m); 229 if (checkdone(ms, &rv)) 230 goto done; 231 } 232 233 /* Check if we have a CDF file */ 234 if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0) 235 if ((m = file_trycdf(ms, fd, ubuf, nb)) != 0) { 236 if ((ms->flags & MAGIC_DEBUG) != 0) 237 (void)fprintf(stderr, "cdf %d\n", m); 238 if (checkdone(ms, &rv)) 239 goto done; 240 } 241 242 /* try soft magic tests */ 243 if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0) 244 if ((m = file_softmagic(ms, ubuf, nb, 0, NULL, BINTEST, 245 looks_text)) != 0) { 246 if ((ms->flags & MAGIC_DEBUG) != 0) 247 (void)fprintf(stderr, "softmagic %d\n", m); 248 #ifdef BUILTIN_ELF 249 if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 && 250 nb > 5 && fd != -1) { 251 /* 252 * We matched something in the file, so this 253 * *might* be an ELF file, and the file is at 254 * least 5 bytes long, so if it's an ELF file 255 * it has at least one byte past the ELF magic 256 * number - try extracting information from the 257 * ELF headers that cannot easily * be 258 * extracted with rules in the magic file. 259 */ 260 if ((m = file_tryelf(ms, fd, ubuf, nb)) != 0) 261 if ((ms->flags & MAGIC_DEBUG) != 0) 262 (void)fprintf(stderr, 263 "elf %d\n", m); 264 } 265 #endif 266 if (checkdone(ms, &rv)) 267 goto done; 268 } 269 270 /* try text properties */ 271 if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) { 272 273 if ((m = file_ascmagic(ms, ubuf, nb, looks_text)) != 0) { 274 if ((ms->flags & MAGIC_DEBUG) != 0) 275 (void)fprintf(stderr, "ascmagic %d\n", m); 276 if (checkdone(ms, &rv)) 277 goto done; 278 } 279 } 280 281 simple: 282 /* give up */ 283 m = 1; 284 if ((!mime || (mime & MAGIC_MIME_TYPE)) && 285 file_printf(ms, "%s", mime ? type : def) == -1) { 286 rv = -1; 287 } 288 done: 289 if ((ms->flags & MAGIC_MIME_ENCODING) != 0) { 290 if (ms->flags & MAGIC_MIME_TYPE) 291 if (file_printf(ms, "; charset=") == -1) 292 rv = -1; 293 if (file_printf(ms, "%s", code_mime) == -1) 294 rv = -1; 295 } 296 #if HAVE_FORK 297 done_encoding: 298 #endif 299 free(u8buf); 300 if (rv) 301 return rv; 302 303 return m; 304 } 305 #endif 306 307 protected int 308 file_reset(struct magic_set *ms) 309 { 310 if (ms->mlist[0] == NULL) { 311 file_error(ms, 0, "no magic files loaded"); 312 return -1; 313 } 314 if (ms->o.buf) { 315 free(ms->o.buf); 316 ms->o.buf = NULL; 317 } 318 if (ms->o.pbuf) { 319 free(ms->o.pbuf); 320 ms->o.pbuf = NULL; 321 } 322 ms->event_flags &= ~EVENT_HAD_ERR; 323 ms->error = -1; 324 return 0; 325 } 326 327 #define OCTALIFY(n, o) \ 328 /*LINTED*/ \ 329 (void)(*(n)++ = '\\', \ 330 *(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \ 331 *(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \ 332 *(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \ 333 (o)++) 334 335 protected const char * 336 file_getbuffer(struct magic_set *ms) 337 { 338 char *pbuf, *op, *np; 339 size_t psize, len; 340 341 if (ms->event_flags & EVENT_HAD_ERR) 342 return NULL; 343 344 if (ms->flags & MAGIC_RAW) 345 return ms->o.buf; 346 347 if (ms->o.buf == NULL) 348 return NULL; 349 350 /* * 4 is for octal representation, + 1 is for NUL */ 351 len = strlen(ms->o.buf); 352 if (len > (SIZE_MAX - 1) / 4) { 353 file_oomem(ms, len); 354 return NULL; 355 } 356 psize = len * 4 + 1; 357 if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) { 358 file_oomem(ms, psize); 359 return NULL; 360 } 361 ms->o.pbuf = pbuf; 362 363 #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH) 364 { 365 mbstate_t state; 366 wchar_t nextchar; 367 int mb_conv = 1; 368 size_t bytesconsumed; 369 char *eop; 370 (void)memset(&state, 0, sizeof(mbstate_t)); 371 372 np = ms->o.pbuf; 373 op = ms->o.buf; 374 eop = op + len; 375 376 while (op < eop) { 377 bytesconsumed = mbrtowc(&nextchar, op, 378 (size_t)(eop - op), &state); 379 if (bytesconsumed == (size_t)(-1) || 380 bytesconsumed == (size_t)(-2)) { 381 mb_conv = 0; 382 break; 383 } 384 385 if (iswprint(nextchar)) { 386 (void)memcpy(np, op, bytesconsumed); 387 op += bytesconsumed; 388 np += bytesconsumed; 389 } else { 390 while (bytesconsumed-- > 0) 391 OCTALIFY(np, op); 392 } 393 } 394 *np = '\0'; 395 396 /* Parsing succeeded as a multi-byte sequence */ 397 if (mb_conv != 0) 398 return ms->o.pbuf; 399 } 400 #endif 401 402 for (np = ms->o.pbuf, op = ms->o.buf; *op;) { 403 if (isprint((unsigned char)*op)) { 404 *np++ = *op++; 405 } else { 406 OCTALIFY(np, op); 407 } 408 } 409 *np = '\0'; 410 return ms->o.pbuf; 411 } 412 413 protected int 414 file_check_mem(struct magic_set *ms, unsigned int level) 415 { 416 size_t len; 417 418 if (level >= ms->c.len) { 419 len = (ms->c.len = 20 + level) * sizeof(*ms->c.li); 420 ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ? 421 malloc(len) : 422 realloc(ms->c.li, len)); 423 if (ms->c.li == NULL) { 424 file_oomem(ms, len); 425 return -1; 426 } 427 } 428 ms->c.li[level].got_match = 0; 429 #ifdef ENABLE_CONDITIONALS 430 ms->c.li[level].last_match = 0; 431 ms->c.li[level].last_cond = COND_NONE; 432 #endif /* ENABLE_CONDITIONALS */ 433 return 0; 434 } 435 436 protected size_t 437 file_printedlen(const struct magic_set *ms) 438 { 439 return ms->o.buf == NULL ? 0 : strlen(ms->o.buf); 440 } 441 442 protected int 443 file_replace(struct magic_set *ms, const char *pat, const char *rep) 444 { 445 file_regex_t rx; 446 int rc, rv = -1; 447 448 rc = file_regcomp(&rx, pat, REG_EXTENDED); 449 if (rc) { 450 file_regerror(&rx, rc, ms); 451 } else { 452 regmatch_t rm; 453 int nm = 0; 454 while (file_regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) { 455 ms->o.buf[rm.rm_so] = '\0'; 456 if (file_printf(ms, "%s%s", rep, 457 rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1) 458 goto out; 459 nm++; 460 } 461 rv = nm; 462 } 463 out: 464 file_regfree(&rx); 465 return rv; 466 } 467 468 protected int 469 file_regcomp(file_regex_t *rx, const char *pat, int flags) 470 { 471 #ifdef USE_C_LOCALE 472 rx->c_lc_ctype = newlocale(LC_CTYPE_MASK, "C", 0); 473 assert(rx->c_lc_ctype != NULL); 474 rx->old_lc_ctype = uselocale(rx->c_lc_ctype); 475 assert(rx->old_lc_ctype != NULL); 476 #endif 477 rx->pat = pat; 478 479 return rx->rc = regcomp(&rx->rx, pat, flags); 480 } 481 482 protected int 483 file_regexec(file_regex_t *rx, const char *str, size_t nmatch, 484 regmatch_t* pmatch, int eflags) 485 { 486 assert(rx->rc == 0); 487 return regexec(&rx->rx, str, nmatch, pmatch, eflags); 488 } 489 490 protected void 491 file_regfree(file_regex_t *rx) 492 { 493 if (rx->rc == 0) 494 regfree(&rx->rx); 495 #ifdef USE_C_LOCALE 496 (void)uselocale(rx->old_lc_ctype); 497 freelocale(rx->c_lc_ctype); 498 #endif 499 } 500 501 protected void 502 file_regerror(file_regex_t *rx, int rc, struct magic_set *ms) 503 { 504 char errmsg[512]; 505 506 (void)regerror(rc, &rx->rx, errmsg, sizeof(errmsg)); 507 file_magerror(ms, "regex error %d for `%s', (%s)", rc, rx->pat, 508 errmsg); 509 } 510 511 protected file_pushbuf_t * 512 file_push_buffer(struct magic_set *ms) 513 { 514 file_pushbuf_t *pb; 515 516 if (ms->event_flags & EVENT_HAD_ERR) 517 return NULL; 518 519 if ((pb = (CAST(file_pushbuf_t *, malloc(sizeof(*pb))))) == NULL) 520 return NULL; 521 522 pb->buf = ms->o.buf; 523 pb->offset = ms->offset; 524 525 ms->o.buf = NULL; 526 ms->offset = 0; 527 528 return pb; 529 } 530 531 protected char * 532 file_pop_buffer(struct magic_set *ms, file_pushbuf_t *pb) 533 { 534 char *rbuf; 535 536 if (ms->event_flags & EVENT_HAD_ERR) { 537 free(pb->buf); 538 free(pb); 539 return NULL; 540 } 541 542 rbuf = ms->o.buf; 543 544 ms->o.buf = pb->buf; 545 ms->offset = pb->offset; 546 547 free(pb); 548 return rbuf; 549 } 550 551 /* 552 * convert string to ascii printable format. 553 */ 554 protected char * 555 file_printable(char *buf, size_t bufsiz, const char *str) 556 { 557 char *ptr, *eptr; 558 const unsigned char *s = (const unsigned char *)str; 559 560 for (ptr = buf, eptr = ptr + bufsiz - 1; ptr < eptr && *s; s++) { 561 if (isprint(*s)) { 562 *ptr++ = *s; 563 continue; 564 } 565 if (ptr >= eptr - 3) 566 break; 567 *ptr++ = '\\'; 568 *ptr++ = ((CAST(unsigned int, *s) >> 6) & 7) + '0'; 569 *ptr++ = ((CAST(unsigned int, *s) >> 3) & 7) + '0'; 570 *ptr++ = ((CAST(unsigned int, *s) >> 0) & 7) + '0'; 571 } 572 *ptr = '\0'; 573 return buf; 574 } 575