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 1.220 2020/06/08 17:38:27 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 */
93 #define ENABLE_CONDITIONALS
96 #define MAGIC "/etc/magic"
99 #if defined(__EMX__) || defined (WIN32)
105 #define private static
107 #if HAVE_VISIBILITY && !defined(WIN32)
108 #define public __attribute__ ((__visibility__("default")))
110 #define protected __attribute__ ((__visibility__("hidden")))
120 #define __arraycount(a) (sizeof(a) / sizeof(a[0]))
123 #ifndef __GNUC_PREREQ__
125 #define __GNUC_PREREQ__(x, y) \
126 ((__GNUC__ == (x) && __GNUC_MINOR__ >= (y)) || \
129 #define __GNUC_PREREQ__(x, y) 0
134 #ifndef __attribute__
135 #define __attribute__(a)
140 #define MIN(a,b) (((a) < (b)) ? (a) : (b))
144 #define MAX(a,b) (((a) > (b)) ? (a) : (b))
147 #define FILE_BADSIZE CAST(size_t, ~0ul)
148 #define MAXDESC 64 /* max len of text description/MIME type */
149 #define MAXMIME 80 /* max len of text MIME type */
150 #define MAXstring 128 /* max len of "string" types */
152 #define MAGICNO 0xF11E041C
154 #define FILE_MAGICSIZE 376
156 #define FILE_GUID_SIZE sizeof("XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX")
160 #define FILE_COMPILE 2
178 uint8_t hs[2]; /* 2 bytes of a fixed-endian "short" */
179 uint8_t hl[4]; /* 4 bytes of a fixed-endian "long" */
180 uint8_t hq[8]; /* 8 bytes of a fixed-endian "quad" */
181 char s[MAXstring]; /* the search string or regex pattern */
182 unsigned char us[MAXstring];
190 uint16_t cont_level; /* level of ">" */
192 #define INDIR 0x01 /* if '(...)' appears */
193 #define OFFADD 0x02 /* if '>&' or '>...(&' appears */
194 #define INDIROFFADD 0x04 /* if '>&(' appears */
195 #define UNSIGNED 0x08 /* comparison is unsigned */
196 #define NOSPACE 0x10 /* suppress space character before output */
197 #define BINTEST 0x20 /* test is for a binary type (set only
198 for top-level tests) */
199 #define TEXTTEST 0x40 /* for passing to file_softmagic */
200 #define OFFNEGATIVE 0x80 /* relative to the end of file */
205 uint8_t reln; /* relation (0=eq, '>'=gt, etc) */
206 uint8_t vallen; /* length of string value, if any */
207 uint8_t type; /* comparison type (FILE_*) */
208 uint8_t in_type; /* type of indirection */
209 #define FILE_INVALID 0
212 #define FILE_DEFAULT 3
214 #define FILE_STRING 5
216 #define FILE_BESHORT 7
217 #define FILE_BELONG 8
218 #define FILE_BEDATE 9
219 #define FILE_LESHORT 10
220 #define FILE_LELONG 11
221 #define FILE_LEDATE 12
222 #define FILE_PSTRING 13
223 #define FILE_LDATE 14
224 #define FILE_BELDATE 15
225 #define FILE_LELDATE 16
226 #define FILE_REGEX 17
227 #define FILE_BESTRING16 18
228 #define FILE_LESTRING16 19
229 #define FILE_SEARCH 20
230 #define FILE_MEDATE 21
231 #define FILE_MELDATE 22
232 #define FILE_MELONG 23
234 #define FILE_LEQUAD 25
235 #define FILE_BEQUAD 26
236 #define FILE_QDATE 27
237 #define FILE_LEQDATE 28
238 #define FILE_BEQDATE 29
239 #define FILE_QLDATE 30
240 #define FILE_LEQLDATE 31
241 #define FILE_BEQLDATE 32
242 #define FILE_FLOAT 33
243 #define FILE_BEFLOAT 34
244 #define FILE_LEFLOAT 35
245 #define FILE_DOUBLE 36
246 #define FILE_BEDOUBLE 37
247 #define FILE_LEDOUBLE 38
248 #define FILE_BEID3 39
249 #define FILE_LEID3 40
250 #define FILE_INDIRECT 41
251 #define FILE_QWDATE 42
252 #define FILE_LEQWDATE 43
253 #define FILE_BEQWDATE 44
256 #define FILE_CLEAR 47
259 #define FILE_OFFSET 50
260 #define FILE_NAMES_SIZE 51 /* size of array to contain all names */
262 #define IS_STRING(t) \
263 ((t) == FILE_STRING || \
264 (t) == FILE_PSTRING || \
265 (t) == FILE_BESTRING16 || \
266 (t) == FILE_LESTRING16 || \
267 (t) == FILE_REGEX || \
268 (t) == FILE_SEARCH || \
269 (t) == FILE_INDIRECT || \
270 (t) == FILE_NAME || \
273 #define FILE_FMT_NONE 0
274 #define FILE_FMT_NUM 1 /* "cduxXi" */
275 #define FILE_FMT_STR 2 /* "s" */
276 #define FILE_FMT_QUAD 3 /* "ll" */
277 #define FILE_FMT_FLOAT 4 /* "eEfFgG" */
278 #define FILE_FMT_DOUBLE 5 /* "eEfFgG" */
281 uint8_t in_op; /* operator for indirection */
282 uint8_t mask_op; /* operator for mask */
283 #ifdef ENABLE_CONDITIONALS
284 uint8_t cond; /* conditional type */
289 #define FILE_FACTOR_OP_PLUS '+'
290 #define FILE_FACTOR_OP_MINUS '-'
291 #define FILE_FACTOR_OP_TIMES '*'
292 #define FILE_FACTOR_OP_DIV '/'
293 #define FILE_FACTOR_OP_NONE '\0'
295 #define FILE_OPS "&|^+-*/%"
300 #define FILE_OPMINUS 4
301 #define FILE_OPMULTIPLY 5
302 #define FILE_OPDIVIDE 6
303 #define FILE_OPMODULO 7
304 #define FILE_OPS_MASK 0x07 /* mask for above ops */
305 #define FILE_UNUSED_1 0x08
306 #define FILE_UNUSED_2 0x10
307 #define FILE_OPSIGNED 0x20
308 #define FILE_OPINVERSE 0x40
309 #define FILE_OPINDIRECT 0x80
311 #ifdef ENABLE_CONDITIONALS
316 #endif /* ENABLE_CONDITIONALS */
319 int32_t offset; /* offset to magic number */
321 int32_t in_offset; /* offset from indirection */
323 uint32_t lineno; /* line number in magic file */
326 uint64_t _mask; /* for use with numeric and date types */
328 uint32_t _count; /* repeat/line count */
329 uint32_t _flags; /* modifier flags */
330 } _s; /* for use with string types */
332 #define num_mask _u._mask
333 #define str_range _u._s._count
334 #define str_flags _u._s._flags
336 union VALUETYPE value; /* either number or string */
338 char desc[MAXDESC]; /* description */
340 char mimetype[MAXMIME]; /* MIME type */
342 char apple[8]; /* APPLE CREATOR/TYPE */
344 char ext[64]; /* Popular extensions */
347 #define BIT(A) (1 << (A))
348 #define STRING_COMPACT_WHITESPACE BIT(0)
349 #define STRING_COMPACT_OPTIONAL_WHITESPACE BIT(1)
350 #define STRING_IGNORE_LOWERCASE BIT(2)
351 #define STRING_IGNORE_UPPERCASE BIT(3)
352 #define REGEX_OFFSET_START BIT(4)
353 #define STRING_TEXTTEST BIT(5)
354 #define STRING_BINTEST BIT(6)
355 #define PSTRING_1_BE BIT(7)
356 #define PSTRING_1_LE BIT(7)
357 #define PSTRING_2_BE BIT(8)
358 #define PSTRING_2_LE BIT(9)
359 #define PSTRING_4_BE BIT(10)
360 #define PSTRING_4_LE BIT(11)
361 #define REGEX_LINE_COUNT BIT(11)
362 #define PSTRING_LEN \
363 (PSTRING_1_BE|PSTRING_2_LE|PSTRING_2_BE|PSTRING_4_LE|PSTRING_4_BE)
364 #define PSTRING_LENGTH_INCLUDES_ITSELF BIT(12)
365 #define STRING_TRIM BIT(13)
366 #define CHAR_COMPACT_WHITESPACE 'W'
367 #define CHAR_COMPACT_OPTIONAL_WHITESPACE 'w'
368 #define CHAR_IGNORE_LOWERCASE 'c'
369 #define CHAR_IGNORE_UPPERCASE 'C'
370 #define CHAR_REGEX_OFFSET_START 's'
371 #define CHAR_TEXTTEST 't'
372 #define CHAR_TRIM 'T'
373 #define CHAR_BINTEST 'b'
374 #define CHAR_PSTRING_1_BE 'B'
375 #define CHAR_PSTRING_1_LE 'B'
376 #define CHAR_PSTRING_2_BE 'H'
377 #define CHAR_PSTRING_2_LE 'h'
378 #define CHAR_PSTRING_4_BE 'L'
379 #define CHAR_PSTRING_4_LE 'l'
380 #define CHAR_PSTRING_LENGTH_INCLUDES_ITSELF 'J'
381 #define STRING_IGNORE_CASE (STRING_IGNORE_LOWERCASE|STRING_IGNORE_UPPERCASE)
382 #define STRING_DEFAULT_RANGE 100
384 #define INDIRECT_RELATIVE BIT(0)
385 #define CHAR_INDIRECT_RELATIVE 'r'
387 /* list of magic entries */
389 struct magic *magic; /* array of magic entries */
390 uint32_t nmagic; /* number of entries in array */
391 void *map; /* internal resources used by entry */
392 struct mlist *next, *prev;
396 #define CAST(T, b) static_cast<T>(b)
397 #define RCAST(T, b) reinterpret_cast<T>(b)
398 #define CCAST(T, b) const_cast<T>(b)
400 #define CAST(T, b) ((T)(b))
401 #define RCAST(T, b) ((T)(uintptr_t)(b))
402 #define CCAST(T, b) ((T)(uintptr_t)(b))
408 #ifdef ENABLE_CONDITIONALS
410 int last_cond; /* used for error checking by parse() */
417 struct mlist *mlist[MAGIC_SETS]; /* list of regular entries */
420 struct level_info *li;
423 char *buf; /* Accumulation buffer */
424 size_t blen; /* Length of buffer */
425 char *pbuf; /* Printable buffer */
427 uint32_t offset; /* a copy of m->offset while we */
428 /* are working on the magic entry */
429 uint32_t eoffset; /* offset from end of file */
431 int flags; /* Control magic tests. */
432 int event_flags; /* Note things that happened. */
433 #define EVENT_HAD_ERR 0x01
435 size_t line; /* current magic line number */
436 mode_t mode; /* copy of current stat mode */
438 /* data for searches */
440 const char *s; /* start of search in original source */
441 size_t s_len; /* length of search region */
442 size_t offset; /* starting offset in source: XXX - should this be off_t? */
443 size_t rm_len; /* match length */
446 /* FIXME: Make the string dynamically allocated so that e.g.
447 strings matched in files can be longer than MAXstring */
448 union VALUETYPE ms_value; /* either number or string */
451 uint16_t elf_shnum_max;
452 uint16_t elf_phnum_max;
453 uint16_t elf_notes_max;
455 size_t bytes_max; /* number of bytes to read from file */
456 #ifndef FILE_BYTES_MAX
457 # define FILE_BYTES_MAX (1024 * 1024) /* how much of the file to look at */
459 #define FILE_ELF_NOTES_MAX 256
460 #define FILE_ELF_PHNUM_MAX 2048
461 #define FILE_ELF_SHNUM_MAX 32768
462 #define FILE_INDIR_MAX 50
463 #define FILE_NAME_MAX 50
464 #define FILE_REGEX_MAX 8192
467 /* Type for Unicode characters */
468 typedef unsigned long unichar;
471 #define FILE_T_LOCAL 1
472 #define FILE_T_WINDOWS 2
473 protected const char *file_fmttime(char *, size_t, uint64_t, int);
474 protected struct magic_set *file_ms_alloc(int);
475 protected void file_ms_free(struct magic_set *);
476 protected int file_default(struct magic_set *, size_t);
477 protected int file_buffer(struct magic_set *, int, struct stat *, const char *,
478 const void *, size_t);
479 protected int file_fsmagic(struct magic_set *, const char *, struct stat *);
480 protected int file_pipe2file(struct magic_set *, int, const void *, size_t);
481 protected int file_vprintf(struct magic_set *, const char *, va_list)
482 __attribute__((__format__(__printf__, 2, 0)));
483 protected int file_separator(struct magic_set *);
484 protected char *file_copystr(char *, size_t, size_t, const char *);
485 protected int file_checkfmt(char *, size_t, const char *);
486 protected size_t file_printedlen(const struct magic_set *);
487 protected int file_print_guid(char *, size_t, const uint64_t *);
488 protected int file_parse_guid(const char *, uint64_t *);
489 protected int file_replace(struct magic_set *, const char *, const char *);
490 protected int file_printf(struct magic_set *, const char *, ...)
491 __attribute__((__format__(__printf__, 2, 3)));
492 protected int file_reset(struct magic_set *, int);
493 protected int file_tryelf(struct magic_set *, const struct buffer *);
494 protected int file_trycdf(struct magic_set *, const struct buffer *);
496 protected int file_zmagic(struct magic_set *, const struct buffer *,
499 protected int file_ascmagic(struct magic_set *, const struct buffer *,
501 protected int file_ascmagic_with_encoding(struct magic_set *,
502 const struct buffer *, unichar *, size_t, const char *, const char *, int);
503 protected int file_encoding(struct magic_set *, const struct buffer *,
504 unichar **, size_t *, const char **, const char **, const char **);
505 protected int file_is_json(struct magic_set *, const struct buffer *);
506 protected int file_is_csv(struct magic_set *, const struct buffer *, int);
507 protected int file_is_tar(struct magic_set *, const struct buffer *);
508 protected int file_softmagic(struct magic_set *, const struct buffer *,
509 uint16_t *, uint16_t *, int, int);
510 protected int file_apprentice(struct magic_set *, const char *, int);
511 protected int buffer_apprentice(struct magic_set *, struct magic **,
513 protected int file_magicfind(struct magic_set *, const char *, struct mlist *);
514 protected uint64_t file_signextend(struct magic_set *, struct magic *,
516 protected void file_badread(struct magic_set *);
517 protected void file_badseek(struct magic_set *);
518 protected void file_oomem(struct magic_set *, size_t);
519 protected void file_error(struct magic_set *, int, const char *, ...)
520 __attribute__((__format__(__printf__, 3, 4)));
521 protected void file_magerror(struct magic_set *, const char *, ...)
522 __attribute__((__format__(__printf__, 2, 3)));
523 protected void file_magwarn(struct magic_set *, const char *, ...)
524 __attribute__((__format__(__printf__, 2, 3)));
525 protected void file_mdump(struct magic *);
526 protected void file_showstr(FILE *, const char *, size_t);
527 protected size_t file_mbswidth(const char *);
528 protected const char *file_getbuffer(struct magic_set *);
529 protected ssize_t sread(int, void *, size_t, int);
530 protected int file_check_mem(struct magic_set *, unsigned int);
531 protected int file_looks_utf8(const unsigned char *, size_t, unichar *,
533 protected size_t file_pstring_length_size(struct magic_set *,
534 const struct magic *);
535 protected size_t file_pstring_get_length(struct magic_set *,
536 const struct magic *, const char *);
537 protected char * file_printable(char *, size_t, const char *, size_t);
539 protected int file_os2_apptype(struct magic_set *, const char *, const void *,
543 protected void buffer_init(struct buffer *, int, const struct stat *,
544 const void *, size_t);
545 protected void buffer_fini(struct buffer *);
546 protected int buffer_fill(const struct buffer *);
549 #if defined(HAVE_XLOCALE_H)
555 #if defined(HAVE_NEWLOCALE) && defined(HAVE_USELOCALE) && defined(HAVE_FREELOCALE)
557 locale_t old_lc_ctype;
566 protected int file_regcomp(file_regex_t *, const char *, int);
567 protected int file_regexec(file_regex_t *, const char *, size_t, regmatch_t *,
569 protected void file_regfree(file_regex_t *);
570 protected void file_regerror(file_regex_t *, int, struct magic_set *);
578 protected file_pushbuf_t *file_push_buffer(struct magic_set *);
579 protected char *file_pop_buffer(struct magic_set *, file_pushbuf_t *);
582 extern const char *file_names[];
583 extern const size_t file_nnames;
587 ssize_t pread(int, void *, size_t, off_t);
589 #ifndef HAVE_VASPRINTF
590 int vasprintf(char **, const char *, va_list);
592 #ifndef HAVE_ASPRINTF
593 int asprintf(char **, const char *, ...);
596 int dprintf(int, const char *, ...);
600 size_t strlcpy(char *, const char *, size_t);
603 size_t strlcat(char *, const char *, size_t);
605 #ifndef HAVE_STRCASESTR
606 char *strcasestr(const char *, const char *);
609 ssize_t getline(char **, size_t *, FILE *);
610 ssize_t getdelim(char **, size_t *, int, FILE *);
613 char *ctime_r(const time_t *, char *);
615 #ifndef HAVE_ASCTIME_R
616 char *asctime_r(const struct tm *, char *);
618 #ifndef HAVE_GMTIME_R
619 struct tm *gmtime_r(const time_t *, struct tm *);
621 #ifndef HAVE_LOCALTIME_R
622 struct tm *localtime_r(const time_t *, struct tm *);
624 #ifndef HAVE_FMTCHECK
625 const char *fmtcheck(const char *, const char *)
626 __attribute__((__format_arg__(2)));
629 #ifdef HAVE_LIBSECCOMP
631 // this mode should not interfere with normal operations
632 // only some dangerous syscalls are blacklisted
633 int enable_sandbox_basic(void);
636 // this mode allows only the necessary syscalls used during normal operation
637 // extensive testing required !!!
638 int enable_sandbox_full(void);
641 protected const char *file_getprogname(void);
642 protected void file_setprogname(const char *);
643 protected void file_err(int, const char *, ...)
644 __attribute__((__format__(__printf__, 2, 3), __noreturn__));
645 protected void file_errx(int, const char *, ...)
646 __attribute__((__format__(__printf__, 2, 3), __noreturn__));
647 protected void file_warn(const char *, ...)
648 __attribute__((__format__(__printf__, 1, 2)));
649 protected void file_warnx(const char *, ...)
650 __attribute__((__format__(__printf__, 1, 2)));
652 #if defined(HAVE_MMAP) && defined(HAVE_SYS_MMAN_H) && !defined(QUICK)
664 #if defined(__GNUC__) && (__GNUC__ >= 3)
665 #define FILE_RCSID(id) \
666 static const char rcsid[] __attribute__((__used__)) = id;
668 #define FILE_RCSID(id) \
669 static const char *rcsid(const char *p) { \
670 return rcsid(p = id); \
674 #define FILE_RCSID(id)
680 #endif /* __file_h__ */