2 * Copyright (c) Ian F. Darwin 1986-1995.
3 * Software written by Ian F. Darwin and others;
4 * maintained 1995-present by Christos Zoulas and others.
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
9 * 1. Redistributions of source code must retain the above copyright
10 * notice immediately at the beginning of the file, without modification,
11 * this list of conditions, and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
20 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * file.h - definitions for file(1) program
30 * @(#)$File: file.h,v 1.227 2021/06/30 10:08:48 christos Exp $
44 #ifdef HAVE_INTTYPES_H
48 #ifndef __STDC_LIMIT_MACROS
49 #define __STDC_LIMIT_MACROS
51 #ifndef __STDC_FORMAT_MACROS
52 #define __STDC_FORMAT_MACROS
58 # define SIZE_T_FORMAT PRIu64
60 # define SIZE_T_FORMAT PRIu32
62 # define INT64_T_FORMAT PRIi64
63 # define INTMAX_T_FORMAT PRIiMAX
66 # define SIZE_T_FORMAT "I64"
68 # define SIZE_T_FORMAT ""
70 # define INT64_T_FORMAT "I64"
71 # define INTMAX_T_FORMAT "I64"
74 # define SIZE_T_FORMAT "z"
75 # define INT64_T_FORMAT "ll"
76 # define INTMAX_T_FORMAT "j"
79 #include <stdio.h> /* Include that here, to make sure __P gets defined */
81 #include <fcntl.h> /* For open and flags */
84 #include <sys/types.h>
86 #include <sys/param.h>
88 /* Do this here and now, because struct stat gets re-defined on solaris */
92 #define ENABLE_CONDITIONALS
95 #define MAGIC "/etc/magic"
98 #if defined(__EMX__) || defined (WIN32)
104 #define private static
106 #if HAVE_VISIBILITY && !defined(WIN32)
107 #define public __attribute__ ((__visibility__("default")))
109 #define protected __attribute__ ((__visibility__("hidden")))
119 #define __arraycount(a) (sizeof(a) / sizeof(a[0]))
122 #ifndef __GNUC_PREREQ__
124 #define __GNUC_PREREQ__(x, y) \
125 ((__GNUC__ == (x) && __GNUC_MINOR__ >= (y)) || \
128 #define __GNUC_PREREQ__(x, y) 0
133 #ifndef __attribute__
134 #define __attribute__(a)
139 #define MIN(a,b) (((a) < (b)) ? (a) : (b))
143 #define MAX(a,b) (((a) > (b)) ? (a) : (b))
151 # define FD_CLOEXEC 1
154 #define FILE_BADSIZE CAST(size_t, ~0ul)
155 #define MAXDESC 64 /* max len of text description/MIME type */
156 #define MAXMIME 80 /* max len of text MIME type */
157 #define MAXstring 128 /* max len of "string" types */
159 #define MAGICNO 0xF11E041C
161 #define FILE_MAGICSIZE 376
163 #define FILE_GUID_SIZE sizeof("XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX")
167 #define FILE_COMPILE 2
185 uint8_t hs[2]; /* 2 bytes of a fixed-endian "short" */
186 uint8_t hl[4]; /* 4 bytes of a fixed-endian "long" */
187 uint8_t hq[8]; /* 8 bytes of a fixed-endian "quad" */
188 char s[MAXstring]; /* the search string or regex pattern */
189 unsigned char us[MAXstring];
197 uint16_t cont_level; /* level of ">" */
199 #define INDIR 0x01 /* if '(...)' appears */
200 #define OFFADD 0x02 /* if '>&' or '>...(&' appears */
201 #define INDIROFFADD 0x04 /* if '>&(' appears */
202 #define UNSIGNED 0x08 /* comparison is unsigned */
203 #define NOSPACE 0x10 /* suppress space character before output */
204 #define BINTEST 0x20 /* test is for a binary type (set only
205 for top-level tests) */
206 #define TEXTTEST 0x40 /* for passing to file_softmagic */
207 #define OFFNEGATIVE 0x80 /* relative to the end of file */
212 uint8_t reln; /* relation (0=eq, '>'=gt, etc) */
213 uint8_t vallen; /* length of string value, if any */
214 uint8_t type; /* comparison type (FILE_*) */
215 uint8_t in_type; /* type of indirection */
216 #define FILE_INVALID 0
219 #define FILE_DEFAULT 3
221 #define FILE_STRING 5
223 #define FILE_BESHORT 7
224 #define FILE_BELONG 8
225 #define FILE_BEDATE 9
226 #define FILE_LESHORT 10
227 #define FILE_LELONG 11
228 #define FILE_LEDATE 12
229 #define FILE_PSTRING 13
230 #define FILE_LDATE 14
231 #define FILE_BELDATE 15
232 #define FILE_LELDATE 16
233 #define FILE_REGEX 17
234 #define FILE_BESTRING16 18
235 #define FILE_LESTRING16 19
236 #define FILE_SEARCH 20
237 #define FILE_MEDATE 21
238 #define FILE_MELDATE 22
239 #define FILE_MELONG 23
241 #define FILE_LEQUAD 25
242 #define FILE_BEQUAD 26
243 #define FILE_QDATE 27
244 #define FILE_LEQDATE 28
245 #define FILE_BEQDATE 29
246 #define FILE_QLDATE 30
247 #define FILE_LEQLDATE 31
248 #define FILE_BEQLDATE 32
249 #define FILE_FLOAT 33
250 #define FILE_BEFLOAT 34
251 #define FILE_LEFLOAT 35
252 #define FILE_DOUBLE 36
253 #define FILE_BEDOUBLE 37
254 #define FILE_LEDOUBLE 38
255 #define FILE_BEID3 39
256 #define FILE_LEID3 40
257 #define FILE_INDIRECT 41
258 #define FILE_QWDATE 42
259 #define FILE_LEQWDATE 43
260 #define FILE_BEQWDATE 44
263 #define FILE_CLEAR 47
266 #define FILE_OFFSET 50
267 #define FILE_BEVARINT 51
268 #define FILE_LEVARINT 52
269 #define FILE_NAMES_SIZE 53 /* size of array to contain all names */
271 #define IS_STRING(t) \
272 ((t) == FILE_STRING || \
273 (t) == FILE_PSTRING || \
274 (t) == FILE_BESTRING16 || \
275 (t) == FILE_LESTRING16 || \
276 (t) == FILE_REGEX || \
277 (t) == FILE_SEARCH || \
278 (t) == FILE_INDIRECT || \
279 (t) == FILE_NAME || \
282 #define FILE_FMT_NONE 0
283 #define FILE_FMT_NUM 1 /* "cduxXi" */
284 #define FILE_FMT_STR 2 /* "s" */
285 #define FILE_FMT_QUAD 3 /* "ll" */
286 #define FILE_FMT_FLOAT 4 /* "eEfFgG" */
287 #define FILE_FMT_DOUBLE 5 /* "eEfFgG" */
290 uint8_t in_op; /* operator for indirection */
291 uint8_t mask_op; /* operator for mask */
292 #ifdef ENABLE_CONDITIONALS
293 uint8_t cond; /* conditional type */
298 #define FILE_FACTOR_OP_PLUS '+'
299 #define FILE_FACTOR_OP_MINUS '-'
300 #define FILE_FACTOR_OP_TIMES '*'
301 #define FILE_FACTOR_OP_DIV '/'
302 #define FILE_FACTOR_OP_NONE '\0'
304 #define FILE_OPS "&|^+-*/%"
309 #define FILE_OPMINUS 4
310 #define FILE_OPMULTIPLY 5
311 #define FILE_OPDIVIDE 6
312 #define FILE_OPMODULO 7
313 #define FILE_OPS_MASK 0x07 /* mask for above ops */
314 #define FILE_UNUSED_1 0x08
315 #define FILE_UNUSED_2 0x10
316 #define FILE_OPSIGNED 0x20
317 #define FILE_OPINVERSE 0x40
318 #define FILE_OPINDIRECT 0x80
320 #ifdef ENABLE_CONDITIONALS
325 #endif /* ENABLE_CONDITIONALS */
328 int32_t offset; /* offset to magic number */
330 int32_t in_offset; /* offset from indirection */
332 uint32_t lineno; /* line number in magic file */
335 uint64_t _mask; /* for use with numeric and date types */
337 uint32_t _count; /* repeat/line count */
338 uint32_t _flags; /* modifier flags */
339 } _s; /* for use with string types */
341 #define num_mask _u._mask
342 #define str_range _u._s._count
343 #define str_flags _u._s._flags
345 union VALUETYPE value; /* either number or string */
347 char desc[MAXDESC]; /* description */
349 char mimetype[MAXMIME]; /* MIME type */
351 char apple[8]; /* APPLE CREATOR/TYPE */
353 char ext[64]; /* Popular extensions */
356 #define BIT(A) (1 << (A))
357 #define STRING_COMPACT_WHITESPACE BIT(0)
358 #define STRING_COMPACT_OPTIONAL_WHITESPACE BIT(1)
359 #define STRING_IGNORE_LOWERCASE BIT(2)
360 #define STRING_IGNORE_UPPERCASE BIT(3)
361 #define REGEX_OFFSET_START BIT(4)
362 #define STRING_TEXTTEST BIT(5)
363 #define STRING_BINTEST BIT(6)
364 #define PSTRING_1_BE BIT(7)
365 #define PSTRING_1_LE BIT(7)
366 #define PSTRING_2_BE BIT(8)
367 #define PSTRING_2_LE BIT(9)
368 #define PSTRING_4_BE BIT(10)
369 #define PSTRING_4_LE BIT(11)
370 #define REGEX_LINE_COUNT BIT(11)
371 #define PSTRING_LEN \
372 (PSTRING_1_BE|PSTRING_2_LE|PSTRING_2_BE|PSTRING_4_LE|PSTRING_4_BE)
373 #define PSTRING_LENGTH_INCLUDES_ITSELF BIT(12)
374 #define STRING_TRIM BIT(13)
375 #define STRING_FULL_WORD BIT(14)
376 #define CHAR_COMPACT_WHITESPACE 'W'
377 #define CHAR_COMPACT_OPTIONAL_WHITESPACE 'w'
378 #define CHAR_IGNORE_LOWERCASE 'c'
379 #define CHAR_IGNORE_UPPERCASE 'C'
380 #define CHAR_REGEX_OFFSET_START 's'
381 #define CHAR_TEXTTEST 't'
382 #define CHAR_TRIM 'T'
383 #define CHAR_FULL_WORD 'f'
384 #define CHAR_BINTEST 'b'
385 #define CHAR_PSTRING_1_BE 'B'
386 #define CHAR_PSTRING_1_LE 'B'
387 #define CHAR_PSTRING_2_BE 'H'
388 #define CHAR_PSTRING_2_LE 'h'
389 #define CHAR_PSTRING_4_BE 'L'
390 #define CHAR_PSTRING_4_LE 'l'
391 #define CHAR_PSTRING_LENGTH_INCLUDES_ITSELF 'J'
392 #define STRING_IGNORE_CASE (STRING_IGNORE_LOWERCASE|STRING_IGNORE_UPPERCASE)
393 #define STRING_DEFAULT_RANGE 100
395 #define INDIRECT_RELATIVE BIT(0)
396 #define CHAR_INDIRECT_RELATIVE 'r'
398 /* list of magic entries */
400 struct magic *magic; /* array of magic entries */
401 uint32_t nmagic; /* number of entries in array */
402 void *map; /* internal resources used by entry */
403 struct mlist *next, *prev;
407 #define CAST(T, b) static_cast<T>(b)
408 #define RCAST(T, b) reinterpret_cast<T>(b)
409 #define CCAST(T, b) const_cast<T>(b)
411 #define CAST(T, b) ((T)(b))
412 #define RCAST(T, b) ((T)(uintptr_t)(b))
413 #define CCAST(T, b) ((T)(uintptr_t)(b))
419 #ifdef ENABLE_CONDITIONALS
421 int last_cond; /* used for error checking by parse() */
427 struct level_info *li;
433 struct mlist *mlist[MAGIC_SETS]; /* list of regular entries */
436 char *buf; /* Accumulation buffer */
437 size_t blen; /* Length of buffer */
438 char *pbuf; /* Printable buffer */
440 uint32_t offset; /* a copy of m->offset while we */
441 /* are working on the magic entry */
442 uint32_t eoffset; /* offset from end of file */
444 int flags; /* Control magic tests. */
445 int event_flags; /* Note things that happened. */
446 #define EVENT_HAD_ERR 0x01
448 size_t line; /* current magic line number */
449 mode_t mode; /* copy of current stat mode */
451 /* data for searches */
453 const char *s; /* start of search in original source */
454 size_t s_len; /* length of search region */
455 size_t offset; /* starting offset in source: XXX - should this be off_t? */
456 size_t rm_len; /* match length */
459 /* FIXME: Make the string dynamically allocated so that e.g.
460 strings matched in files can be longer than MAXstring */
461 union VALUETYPE ms_value; /* either number or string */
464 uint16_t elf_shnum_max;
465 uint16_t elf_phnum_max;
466 uint16_t elf_notes_max;
468 size_t bytes_max; /* number of bytes to read from file */
469 size_t encoding_max; /* bytes to look for encoding */
470 #ifndef FILE_BYTES_MAX
471 # define FILE_BYTES_MAX (1024 * 1024) /* how much of the file to look at */
473 #define FILE_ELF_NOTES_MAX 256
474 #define FILE_ELF_PHNUM_MAX 2048
475 #define FILE_ELF_SHNUM_MAX 32768
476 #define FILE_INDIR_MAX 50
477 #define FILE_NAME_MAX 50
478 #define FILE_REGEX_MAX 8192
479 #define FILE_ENCODING_MAX (64 * 1024)
482 /* Type for Unicode characters */
483 typedef unsigned long file_unichar_t;
486 #define FILE_T_LOCAL 1
487 #define FILE_T_WINDOWS 2
488 protected const char *file_fmttime(char *, size_t, uint64_t, int);
489 protected const char *file_fmtvarint(const unsigned char *, int, char *,
491 protected struct magic_set *file_ms_alloc(int);
492 protected void file_ms_free(struct magic_set *);
493 protected int file_default(struct magic_set *, size_t);
494 protected int file_buffer(struct magic_set *, int, struct stat *, const char *,
495 const void *, size_t);
496 protected int file_fsmagic(struct magic_set *, const char *, struct stat *);
497 protected int file_pipe2file(struct magic_set *, int, const void *, size_t);
498 protected int file_vprintf(struct magic_set *, const char *, va_list)
499 __attribute__((__format__(__printf__, 2, 0)));
500 protected int file_separator(struct magic_set *);
501 protected char *file_copystr(char *, size_t, size_t, const char *);
502 protected int file_checkfmt(char *, size_t, const char *);
503 protected size_t file_printedlen(const struct magic_set *);
504 protected int file_print_guid(char *, size_t, const uint64_t *);
505 protected int file_parse_guid(const char *, uint64_t *);
506 protected int file_replace(struct magic_set *, const char *, const char *);
507 protected int file_printf(struct magic_set *, const char *, ...)
508 __attribute__((__format__(__printf__, 2, 3)));
509 protected int file_reset(struct magic_set *, int);
510 protected int file_tryelf(struct magic_set *, const struct buffer *);
511 protected int file_trycdf(struct magic_set *, const struct buffer *);
513 protected int file_zmagic(struct magic_set *, const struct buffer *,
516 protected int file_ascmagic(struct magic_set *, const struct buffer *,
518 protected int file_ascmagic_with_encoding(struct magic_set *,
519 const struct buffer *, file_unichar_t *, size_t, const char *, const char *, int);
520 protected int file_encoding(struct magic_set *, const struct buffer *,
521 file_unichar_t **, size_t *, const char **, const char **, const char **);
522 protected int file_is_json(struct magic_set *, const struct buffer *);
523 protected int file_is_csv(struct magic_set *, const struct buffer *, int);
524 protected int file_is_tar(struct magic_set *, const struct buffer *);
525 protected int file_softmagic(struct magic_set *, const struct buffer *,
526 uint16_t *, uint16_t *, int, int);
527 protected int file_apprentice(struct magic_set *, const char *, int);
528 protected int buffer_apprentice(struct magic_set *, struct magic **,
530 protected int file_magicfind(struct magic_set *, const char *, struct mlist *);
531 protected uint64_t file_signextend(struct magic_set *, struct magic *,
533 protected uintmax_t file_varint2uintmax_t(const unsigned char *, int, size_t *);
535 protected void file_badread(struct magic_set *);
536 protected void file_badseek(struct magic_set *);
537 protected void file_oomem(struct magic_set *, size_t);
538 protected void file_error(struct magic_set *, int, const char *, ...)
539 __attribute__((__format__(__printf__, 3, 4)));
540 protected void file_magerror(struct magic_set *, const char *, ...)
541 __attribute__((__format__(__printf__, 2, 3)));
542 protected void file_magwarn(struct magic_set *, const char *, ...)
543 __attribute__((__format__(__printf__, 2, 3)));
544 protected void file_mdump(struct magic *);
545 protected void file_showstr(FILE *, const char *, size_t);
546 protected size_t file_mbswidth(const char *);
547 protected const char *file_getbuffer(struct magic_set *);
548 protected ssize_t sread(int, void *, size_t, int);
549 protected int file_check_mem(struct magic_set *, unsigned int);
550 protected int file_looks_utf8(const unsigned char *, size_t, file_unichar_t *,
552 protected size_t file_pstring_length_size(struct magic_set *,
553 const struct magic *);
554 protected size_t file_pstring_get_length(struct magic_set *,
555 const struct magic *, const char *);
556 protected char * file_printable(struct magic_set *, char *, size_t,
557 const char *, size_t);
559 protected int file_os2_apptype(struct magic_set *, const char *, const void *,
562 protected int file_pipe_closexec(int *);
563 protected int file_clear_closexec(int);
564 protected char *file_strtrim(char *);
566 protected void buffer_init(struct buffer *, int, const struct stat *,
567 const void *, size_t);
568 protected void buffer_fini(struct buffer *);
569 protected int buffer_fill(const struct buffer *);
572 #if defined(HAVE_XLOCALE_H)
578 #if defined(HAVE_NEWLOCALE) && defined(HAVE_USELOCALE) && defined(HAVE_FREELOCALE)
580 locale_t old_lc_ctype;
589 protected int file_regcomp(file_regex_t *, const char *, int);
590 protected int file_regexec(file_regex_t *, const char *, size_t, regmatch_t *,
592 protected void file_regfree(file_regex_t *);
593 protected void file_regerror(file_regex_t *, int, struct magic_set *);
601 protected file_pushbuf_t *file_push_buffer(struct magic_set *);
602 protected char *file_pop_buffer(struct magic_set *, file_pushbuf_t *);
605 extern const char *file_names[];
606 extern const size_t file_nnames;
610 ssize_t pread(int, void *, size_t, off_t);
612 #ifndef HAVE_VASPRINTF
613 int vasprintf(char **, const char *, va_list);
615 #ifndef HAVE_ASPRINTF
616 int asprintf(char **, const char *, ...);
619 int dprintf(int, const char *, ...);
623 size_t strlcpy(char *, const char *, size_t);
626 size_t strlcat(char *, const char *, size_t);
628 #ifndef HAVE_STRCASESTR
629 char *strcasestr(const char *, const char *);
632 ssize_t getline(char **, size_t *, FILE *);
633 ssize_t getdelim(char **, size_t *, int, FILE *);
636 char *ctime_r(const time_t *, char *);
638 #ifndef HAVE_ASCTIME_R
639 char *asctime_r(const struct tm *, char *);
641 #ifndef HAVE_GMTIME_R
642 struct tm *gmtime_r(const time_t *, struct tm *);
644 #ifndef HAVE_LOCALTIME_R
645 struct tm *localtime_r(const time_t *, struct tm *);
647 #ifndef HAVE_FMTCHECK
648 const char *fmtcheck(const char *, const char *)
649 __attribute__((__format_arg__(2)));
652 #ifdef HAVE_LIBSECCOMP
654 // this mode should not interfere with normal operations
655 // only some dangerous syscalls are blacklisted
656 int enable_sandbox_basic(void);
659 // this mode allows only the necessary syscalls used during normal operation
660 // extensive testing required !!!
661 int enable_sandbox_full(void);
664 protected const char *file_getprogname(void);
665 protected void file_setprogname(const char *);
666 protected void file_err(int, const char *, ...)
667 __attribute__((__format__(__printf__, 2, 3), __noreturn__));
668 protected void file_errx(int, const char *, ...)
669 __attribute__((__format__(__printf__, 2, 3), __noreturn__));
670 protected void file_warn(const char *, ...)
671 __attribute__((__format__(__printf__, 1, 2)));
672 protected void file_warnx(const char *, ...)
673 __attribute__((__format__(__printf__, 1, 2)));
675 #if defined(HAVE_MMAP) && defined(HAVE_SYS_MMAN_H) && !defined(QUICK)
687 #if defined(__GNUC__) && (__GNUC__ >= 3)
688 #define FILE_RCSID(id) \
689 static const char rcsid[] __attribute__((__used__)) = id;
691 #define FILE_RCSID(id) \
692 static const char *rcsid(const char *p) { \
693 return rcsid(p = id); \
697 #define FILE_RCSID(id)
703 #endif /* __file_h__ */