1 /* $NetBSD: file.h,v 1.11 2014/06/13 02:08:06 christos Exp $ */ 2 /* 3 * Copyright (c) Ian F. Darwin 1986-1995. 4 * Software written by Ian F. Darwin and others; 5 * maintained 1995-present by Christos Zoulas and others. 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 * file.h - definitions for file(1) program 31 * @(#)$File: file.h,v 1.152 2014/06/03 19:01:34 christos Exp $ 32 */ 33 34 #ifndef __file_h__ 35 #define __file_h__ 36 37 #ifdef HAVE_CONFIG_H 38 #include <config.h> 39 #endif 40 41 #ifdef WIN32 42 #ifdef _WIN64 43 #define SIZE_T_FORMAT "I64" 44 #else 45 #define SIZE_T_FORMAT "" 46 #endif 47 #define INT64_T_FORMAT "I64" 48 #else 49 #define SIZE_T_FORMAT "z" 50 #define INT64_T_FORMAT "ll" 51 #endif 52 53 #include <stdio.h> /* Include that here, to make sure __P gets defined */ 54 #include <errno.h> 55 #include <fcntl.h> /* For open and flags */ 56 #ifdef HAVE_STDINT_H 57 #ifndef __STDC_LIMIT_MACROS 58 #define __STDC_LIMIT_MACROS 59 #endif 60 #include <stdint.h> 61 #endif 62 #ifdef HAVE_INTTYPES_H 63 #include <inttypes.h> 64 #endif 65 #include <regex.h> 66 #include <time.h> 67 #include <sys/types.h> 68 #include <sys/param.h> 69 /* Do this here and now, because struct stat gets re-defined on solaris */ 70 #include <sys/stat.h> 71 #include <stdarg.h> 72 73 #define ENABLE_CONDITIONALS 74 75 #ifndef MAGIC 76 #define MAGIC "/etc/magic" 77 #endif 78 79 #if defined(__EMX__) || defined (WIN32) 80 #define PATHSEP ';' 81 #else 82 #define PATHSEP ':' 83 #endif 84 85 #define private static 86 87 #if HAVE_VISIBILITY && !defined(WIN32) 88 #define public __attribute__ ((__visibility__("default"))) 89 #ifndef protected 90 #define protected __attribute__ ((__visibility__("hidden"))) 91 #endif 92 #else 93 #define public 94 #ifndef protected 95 #define protected 96 #endif 97 #endif 98 99 #ifndef __arraycount 100 #define __arraycount(a) (sizeof(a) / sizeof(a[0])) 101 #endif 102 103 #ifndef __GNUC_PREREQ__ 104 #ifdef __GNUC__ 105 #define __GNUC_PREREQ__(x, y) \ 106 ((__GNUC__ == (x) && __GNUC_MINOR__ >= (y)) || \ 107 (__GNUC__ > (x))) 108 #else 109 #define __GNUC_PREREQ__(x, y) 0 110 #endif 111 #endif 112 113 #ifndef __GNUC__ 114 #ifndef __attribute__ 115 #define __attribute__(a) 116 #endif 117 #endif 118 119 #ifndef MIN 120 #define MIN(a,b) (((a) < (b)) ? (a) : (b)) 121 #endif 122 123 #ifndef MAX 124 #define MAX(a,b) (((a) > (b)) ? (a) : (b)) 125 #endif 126 127 #ifndef HOWMANY 128 # define HOWMANY (256 * 1024) /* how much of the file to look at */ 129 #endif 130 #define MAXMAGIS 8192 /* max entries in any one magic file 131 or directory */ 132 #define MAXDESC 64 /* max len of text description/MIME type */ 133 #define MAXMIME 80 /* max len of text MIME type */ 134 #define MAXstring 64 /* max len of "string" types */ 135 136 #define MAGICNO 0xF11E041C 137 #define VERSIONNO 12 138 #define FILE_MAGICSIZE 248 139 140 #define FILE_LOAD 0 141 #define FILE_CHECK 1 142 #define FILE_COMPILE 2 143 #define FILE_LIST 3 144 145 union VALUETYPE { 146 uint8_t b; 147 uint16_t h; 148 uint32_t l; 149 uint64_t q; 150 uint8_t hs[2]; /* 2 bytes of a fixed-endian "short" */ 151 uint8_t hl[4]; /* 4 bytes of a fixed-endian "long" */ 152 uint8_t hq[8]; /* 8 bytes of a fixed-endian "quad" */ 153 char s[MAXstring]; /* the search string or regex pattern */ 154 unsigned char us[MAXstring]; 155 float f; 156 double d; 157 }; 158 159 struct magic { 160 /* Word 1 */ 161 uint16_t cont_level; /* level of ">" */ 162 uint8_t flag; 163 #define INDIR 0x01 /* if '(...)' appears */ 164 #define OFFADD 0x02 /* if '>&' or '>...(&' appears */ 165 #define INDIROFFADD 0x04 /* if '>&(' appears */ 166 #define UNSIGNED 0x08 /* comparison is unsigned */ 167 #define NOSPACE 0x10 /* suppress space character before output */ 168 #define BINTEST 0x20 /* test is for a binary type (set only 169 for top-level tests) */ 170 #define TEXTTEST 0x40 /* for passing to file_softmagic */ 171 172 uint8_t factor; 173 174 /* Word 2 */ 175 uint8_t reln; /* relation (0=eq, '>'=gt, etc) */ 176 uint8_t vallen; /* length of string value, if any */ 177 uint8_t type; /* comparison type (FILE_*) */ 178 uint8_t in_type; /* type of indirection */ 179 #define FILE_INVALID 0 180 #define FILE_BYTE 1 181 #define FILE_SHORT 2 182 #define FILE_DEFAULT 3 183 #define FILE_LONG 4 184 #define FILE_STRING 5 185 #define FILE_DATE 6 186 #define FILE_BESHORT 7 187 #define FILE_BELONG 8 188 #define FILE_BEDATE 9 189 #define FILE_LESHORT 10 190 #define FILE_LELONG 11 191 #define FILE_LEDATE 12 192 #define FILE_PSTRING 13 193 #define FILE_LDATE 14 194 #define FILE_BELDATE 15 195 #define FILE_LELDATE 16 196 #define FILE_REGEX 17 197 #define FILE_BESTRING16 18 198 #define FILE_LESTRING16 19 199 #define FILE_SEARCH 20 200 #define FILE_MEDATE 21 201 #define FILE_MELDATE 22 202 #define FILE_MELONG 23 203 #define FILE_QUAD 24 204 #define FILE_LEQUAD 25 205 #define FILE_BEQUAD 26 206 #define FILE_QDATE 27 207 #define FILE_LEQDATE 28 208 #define FILE_BEQDATE 29 209 #define FILE_QLDATE 30 210 #define FILE_LEQLDATE 31 211 #define FILE_BEQLDATE 32 212 #define FILE_FLOAT 33 213 #define FILE_BEFLOAT 34 214 #define FILE_LEFLOAT 35 215 #define FILE_DOUBLE 36 216 #define FILE_BEDOUBLE 37 217 #define FILE_LEDOUBLE 38 218 #define FILE_BEID3 39 219 #define FILE_LEID3 40 220 #define FILE_INDIRECT 41 221 #define FILE_QWDATE 42 222 #define FILE_LEQWDATE 43 223 #define FILE_BEQWDATE 44 224 #define FILE_NAME 45 225 #define FILE_USE 46 226 #define FILE_CLEAR 47 227 #define FILE_NAMES_SIZE 48 /* size of array to contain all names */ 228 229 #define IS_STRING(t) \ 230 ((t) == FILE_STRING || \ 231 (t) == FILE_PSTRING || \ 232 (t) == FILE_BESTRING16 || \ 233 (t) == FILE_LESTRING16 || \ 234 (t) == FILE_REGEX || \ 235 (t) == FILE_SEARCH || \ 236 (t) == FILE_NAME || \ 237 (t) == FILE_USE) 238 239 #define FILE_FMT_NONE 0 240 #define FILE_FMT_NUM 1 /* "cduxXi" */ 241 #define FILE_FMT_STR 2 /* "s" */ 242 #define FILE_FMT_QUAD 3 /* "ll" */ 243 #define FILE_FMT_FLOAT 4 /* "eEfFgG" */ 244 #define FILE_FMT_DOUBLE 5 /* "eEfFgG" */ 245 246 /* Word 3 */ 247 uint8_t in_op; /* operator for indirection */ 248 uint8_t mask_op; /* operator for mask */ 249 #ifdef ENABLE_CONDITIONALS 250 uint8_t cond; /* conditional type */ 251 #else 252 uint8_t dummy; 253 #endif 254 uint8_t factor_op; 255 #define FILE_FACTOR_OP_PLUS '+' 256 #define FILE_FACTOR_OP_MINUS '-' 257 #define FILE_FACTOR_OP_TIMES '*' 258 #define FILE_FACTOR_OP_DIV '/' 259 #define FILE_FACTOR_OP_NONE '\0' 260 261 #define FILE_OPS "&|^+-*/%" 262 #define FILE_OPAND 0 263 #define FILE_OPOR 1 264 #define FILE_OPXOR 2 265 #define FILE_OPADD 3 266 #define FILE_OPMINUS 4 267 #define FILE_OPMULTIPLY 5 268 #define FILE_OPDIVIDE 6 269 #define FILE_OPMODULO 7 270 #define FILE_OPS_MASK 0x07 /* mask for above ops */ 271 #define FILE_UNUSED_1 0x08 272 #define FILE_UNUSED_2 0x10 273 #define FILE_UNUSED_3 0x20 274 #define FILE_OPINVERSE 0x40 275 #define FILE_OPINDIRECT 0x80 276 277 #ifdef ENABLE_CONDITIONALS 278 #define COND_NONE 0 279 #define COND_IF 1 280 #define COND_ELIF 2 281 #define COND_ELSE 3 282 #endif /* ENABLE_CONDITIONALS */ 283 284 /* Word 4 */ 285 uint32_t offset; /* offset to magic number */ 286 /* Word 5 */ 287 int32_t in_offset; /* offset from indirection */ 288 /* Word 6 */ 289 uint32_t lineno; /* line number in magic file */ 290 /* Word 7,8 */ 291 union { 292 uint64_t _mask; /* for use with numeric and date types */ 293 struct { 294 uint32_t _count; /* repeat/line count */ 295 uint32_t _flags; /* modifier flags */ 296 } _s; /* for use with string types */ 297 } _u; 298 #define num_mask _u._mask 299 #define str_range _u._s._count 300 #define str_flags _u._s._flags 301 /* Words 9-16 */ 302 union VALUETYPE value; /* either number or string */ 303 /* Words 17-32 */ 304 char desc[MAXDESC]; /* description */ 305 /* Words 33-52 */ 306 char mimetype[MAXMIME]; /* MIME type */ 307 /* Words 53-54 */ 308 char apple[8]; 309 }; 310 311 #define BIT(A) (1 << (A)) 312 #define STRING_COMPACT_WHITESPACE BIT(0) 313 #define STRING_COMPACT_OPTIONAL_WHITESPACE BIT(1) 314 #define STRING_IGNORE_LOWERCASE BIT(2) 315 #define STRING_IGNORE_UPPERCASE BIT(3) 316 #define REGEX_OFFSET_START BIT(4) 317 #define STRING_TEXTTEST BIT(5) 318 #define STRING_BINTEST BIT(6) 319 #define PSTRING_1_BE BIT(7) 320 #define PSTRING_1_LE BIT(7) 321 #define PSTRING_2_BE BIT(8) 322 #define PSTRING_2_LE BIT(9) 323 #define PSTRING_4_BE BIT(10) 324 #define PSTRING_4_LE BIT(11) 325 #define REGEX_LINE_COUNT BIT(11) 326 #define PSTRING_LEN \ 327 (PSTRING_1_BE|PSTRING_2_LE|PSTRING_2_BE|PSTRING_4_LE|PSTRING_4_BE) 328 #define PSTRING_LENGTH_INCLUDES_ITSELF BIT(12) 329 #define STRING_TRIM BIT(13) 330 #define CHAR_COMPACT_WHITESPACE 'W' 331 #define CHAR_COMPACT_OPTIONAL_WHITESPACE 'w' 332 #define CHAR_IGNORE_LOWERCASE 'c' 333 #define CHAR_IGNORE_UPPERCASE 'C' 334 #define CHAR_REGEX_OFFSET_START 's' 335 #define CHAR_TEXTTEST 't' 336 #define CHAR_TRIM 'T' 337 #define CHAR_BINTEST 'b' 338 #define CHAR_PSTRING_1_BE 'B' 339 #define CHAR_PSTRING_1_LE 'B' 340 #define CHAR_PSTRING_2_BE 'H' 341 #define CHAR_PSTRING_2_LE 'h' 342 #define CHAR_PSTRING_4_BE 'L' 343 #define CHAR_PSTRING_4_LE 'l' 344 #define CHAR_PSTRING_LENGTH_INCLUDES_ITSELF 'J' 345 #define STRING_IGNORE_CASE (STRING_IGNORE_LOWERCASE|STRING_IGNORE_UPPERCASE) 346 #define STRING_DEFAULT_RANGE 100 347 348 349 /* list of magic entries */ 350 struct mlist { 351 struct magic *magic; /* array of magic entries */ 352 uint32_t nmagic; /* number of entries in array */ 353 void *map; /* internal resources used by entry */ 354 struct mlist *next, *prev; 355 }; 356 357 #ifdef __cplusplus 358 #define CAST(T, b) static_cast<T>(b) 359 #define RCAST(T, b) reinterpret_cast<T>(b) 360 #else 361 #define CAST(T, b) (T)(b) 362 #define RCAST(T, b) (T)(b) 363 #endif 364 365 struct level_info { 366 int32_t off; 367 int got_match; 368 #ifdef ENABLE_CONDITIONALS 369 int last_match; 370 int last_cond; /* used for error checking by parse() */ 371 #endif 372 }; 373 374 #define MAGIC_SETS 2 375 376 struct magic_set { 377 struct mlist *mlist[MAGIC_SETS]; /* list of regular entries */ 378 struct cont { 379 size_t len; 380 struct level_info *li; 381 } c; 382 struct out { 383 char *buf; /* Accumulation buffer */ 384 char *pbuf; /* Printable buffer */ 385 } o; 386 uint32_t offset; 387 int error; 388 int flags; /* Control magic tests. */ 389 int event_flags; /* Note things that happened. */ 390 #define EVENT_HAD_ERR 0x01 391 const char *file; 392 size_t line; /* current magic line number */ 393 394 /* data for searches */ 395 struct { 396 const char *s; /* start of search in original source */ 397 size_t s_len; /* length of search region */ 398 size_t offset; /* starting offset in source: XXX - should this be off_t? */ 399 size_t rm_len; /* match length */ 400 } search; 401 402 /* FIXME: Make the string dynamically allocated so that e.g. 403 strings matched in files can be longer than MAXstring */ 404 union VALUETYPE ms_value; /* either number or string */ 405 }; 406 407 /* Type for Unicode characters */ 408 typedef unsigned long unichar; 409 410 struct stat; 411 #define FILE_T_LOCAL 1 412 #define FILE_T_WINDOWS 2 413 protected const char *file_fmttime(uint64_t, int, char *); 414 protected struct magic_set *file_ms_alloc(int); 415 protected void file_ms_free(struct magic_set *); 416 protected int file_buffer(struct magic_set *, int, const char *, const void *, 417 size_t); 418 protected int file_fsmagic(struct magic_set *, const char *, struct stat *); 419 protected int file_pipe2file(struct magic_set *, int, const void *, size_t); 420 protected int file_vprintf(struct magic_set *, const char *, va_list) 421 __attribute__((__format__(__printf__, 2, 0))); 422 protected size_t file_printedlen(const struct magic_set *); 423 protected int file_replace(struct magic_set *, const char *, const char *); 424 protected int file_printf(struct magic_set *, const char *, ...) 425 __attribute__((__format__(__printf__, 2, 3))); 426 protected int file_reset(struct magic_set *); 427 protected int file_tryelf(struct magic_set *, int, const unsigned char *, 428 size_t); 429 protected int file_trycdf(struct magic_set *, int, const unsigned char *, 430 size_t); 431 #if HAVE_FORK 432 protected int file_zmagic(struct magic_set *, int, const char *, 433 const unsigned char *, size_t); 434 #endif 435 protected int file_ascmagic(struct magic_set *, const unsigned char *, size_t, 436 int); 437 protected int file_ascmagic_with_encoding(struct magic_set *, 438 const unsigned char *, size_t, unichar *, size_t, const char *, 439 const char *, int); 440 protected int file_encoding(struct magic_set *, const unsigned char *, size_t, 441 unichar **, size_t *, const char **, const char **, const char **); 442 protected int file_is_tar(struct magic_set *, const unsigned char *, size_t); 443 protected int file_softmagic(struct magic_set *, const unsigned char *, size_t, 444 size_t, int, int); 445 protected int file_apprentice(struct magic_set *, const char *, int); 446 protected int file_magicfind(struct magic_set *, const char *, struct mlist *); 447 protected uint64_t file_signextend(struct magic_set *, struct magic *, 448 uint64_t); 449 protected void file_badread(struct magic_set *); 450 protected void file_badseek(struct magic_set *); 451 protected void file_oomem(struct magic_set *, size_t); 452 protected void file_error(struct magic_set *, int, const char *, ...) 453 __attribute__((__format__(__printf__, 3, 4))); 454 protected void file_magerror(struct magic_set *, const char *, ...) 455 __attribute__((__format__(__printf__, 2, 3))); 456 protected void file_magwarn(struct magic_set *, const char *, ...) 457 __attribute__((__format__(__printf__, 2, 3))); 458 protected void file_mdump(struct magic *); 459 protected void file_showstr(FILE *, const char *, size_t); 460 protected size_t file_mbswidth(const char *); 461 protected const char *file_getbuffer(struct magic_set *); 462 protected ssize_t sread(int, void *, size_t, int); 463 protected int file_check_mem(struct magic_set *, unsigned int); 464 protected int file_looks_utf8(const unsigned char *, size_t, unichar *, 465 size_t *); 466 protected size_t file_pstring_length_size(const struct magic *); 467 protected size_t file_pstring_get_length(const struct magic *, const char *); 468 #ifdef __EMX__ 469 protected int file_os2_apptype(struct magic_set *, const char *, const void *, 470 size_t); 471 #endif /* __EMX__ */ 472 473 typedef struct { 474 const char *pat; 475 char *old_lc_ctype; 476 int rc; 477 regex_t rx; 478 } file_regex_t; 479 480 protected int file_regcomp(file_regex_t *, const char *, int); 481 protected int file_regexec(file_regex_t *, const char *, size_t, regmatch_t *, 482 int); 483 protected void file_regfree(file_regex_t *); 484 protected void file_regerror(file_regex_t *, int, struct magic_set *); 485 486 #ifndef COMPILE_ONLY 487 extern const char *file_names[]; 488 extern const size_t file_nnames; 489 #endif 490 491 #ifndef HAVE_STRERROR 492 extern int sys_nerr; 493 extern char *sys_errlist[]; 494 #define strerror(e) \ 495 (((e) >= 0 && (e) < sys_nerr) ? sys_errlist[(e)] : "Unknown error") 496 #endif 497 498 #ifndef HAVE_STRTOUL 499 #define strtoul(a, b, c) strtol(a, b, c) 500 #endif 501 502 #ifndef HAVE_PREAD 503 ssize_t pread(int, void *, size_t, off_t); 504 #endif 505 #ifndef HAVE_VASPRINTF 506 int vasprintf(char **, const char *, va_list); 507 #endif 508 #ifndef HAVE_ASPRINTF 509 int asprintf(char **, const char *, ...); 510 #endif 511 512 #ifndef HAVE_STRLCPY 513 size_t strlcpy(char *, const char *, size_t); 514 #endif 515 #ifndef HAVE_STRLCAT 516 size_t strlcat(char *, const char *, size_t); 517 #endif 518 #ifndef HAVE_STRCASESTR 519 char *strcasestr(const char *, const char *); 520 #endif 521 #ifndef HAVE_GETLINE 522 ssize_t getline(char **, size_t *, FILE *); 523 ssize_t getdelim(char **, size_t *, int, FILE *); 524 #endif 525 #ifndef HAVE_CTIME_R 526 char *ctime_r(const time_t *, char *); 527 #endif 528 #ifndef HAVE_ASCTIME_R 529 char *asctime_r(const struct tm *, char *); 530 #endif 531 #ifndef HAVE_FMTCHECK 532 const char *fmtcheck(const char *, const char *) 533 __attribute__((__format_arg__(2))); 534 #endif 535 536 #if defined(HAVE_MMAP) && defined(HAVE_SYS_MMAN_H) && !defined(QUICK) 537 #define QUICK 538 #endif 539 540 #ifndef O_BINARY 541 #define O_BINARY 0 542 #endif 543 544 #ifndef __cplusplus 545 #if defined(__GNUC__) && (__GNUC__ >= 3) 546 #define FILE_RCSID(id) \ 547 static const char rcsid[] __attribute__((__used__)) = id; 548 #else 549 #define FILE_RCSID(id) \ 550 static const char *rcsid(const char *p) { \ 551 return rcsid(p = id); \ 552 } 553 #endif 554 #else 555 #define FILE_RCSID(id) 556 #endif 557 #ifndef __RCSID 558 #define __RCSID(a) 559 #endif 560 561 #endif /* __file_h__ */ 562