xref: /openbsd-src/usr.bin/file/file.h (revision 2b0358df1d88d06ef4139321dd05bd5e05d91eaf)
1 /*	$OpenBSD: file.h,v 1.19 2008/05/08 01:40:56 chl 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  * @(#)$Id: file.h,v 1.19 2008/05/08 01:40:56 chl 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 #include <stdio.h>	/* Include that here, to make sure __P gets defined */
42 #include <errno.h>
43 #include <fcntl.h>	/* For open and flags */
44 #ifdef HAVE_STDINT_H
45 #include <stdint.h>
46 #endif
47 #ifdef HAVE_INTTYPES_H
48 #include <inttypes.h>
49 #endif
50 #include <regex.h>
51 #include <sys/types.h>
52 /* Do this here and now, because struct stat gets re-defined on solaris */
53 #include <sys/stat.h>
54 
55 #define ENABLE_CONDITIONALS
56 
57 #ifndef MAGIC
58 #define MAGIC "/etc/magic"
59 #endif
60 
61 #ifdef __EMX__
62 #define PATHSEP	';'
63 #else
64 #define PATHSEP	':'
65 #endif
66 
67 #define private static
68 #ifndef protected
69 #define protected
70 #endif
71 #define public
72 
73 #ifndef __GNUC_PREREQ__
74 #ifdef __GNUC__
75 #define	__GNUC_PREREQ__(x, y)						\
76 	((__GNUC__ == (x) && __GNUC_MINOR__ >= (y)) ||			\
77 	 (__GNUC__ > (x)))
78 #else
79 #define	__GNUC_PREREQ__(x, y)	0
80 #endif
81 #endif
82 
83 #ifndef MIN
84 #define	MIN(a,b)	(((a) < (b)) ? (a) : (b))
85 #endif
86 
87 #ifndef HOWMANY
88 # define HOWMANY (256 * 1024)	/* how much of the file to look at */
89 #endif
90 #define MAXMAGIS 8192		/* max entries in /etc/magic */
91 #define MAXDESC	64		/* max leng of text description */
92 #define MAXstring 32		/* max leng of "string" types */
93 
94 #define MAGICNO		0xF11E041C
95 #define VERSIONNO	4
96 #define FILE_MAGICSIZE	(32 * 4)
97 
98 #define	FILE_LOAD	0
99 #define FILE_CHECK	1
100 #define FILE_COMPILE	2
101 
102 struct magic {
103 	/* Word 1 */
104 	uint16_t cont_level;	/* level of ">" */
105 	uint8_t nospflag;	/* suppress space character */
106 	uint8_t flag;
107 #define INDIR		1	/* if '(...)' appears */
108 #define OFFADD		2	/* if '>&' or '>...(&' appears */
109 #define INDIROFFADD	4	/* if '>&(' appears */
110 #define	UNSIGNED	8	/* comparison is unsigned */
111 
112 	/* Word 2 */
113 	uint8_t reln;		/* relation (0=eq, '>'=gt, etc) */
114 	uint8_t vallen;		/* length of string value, if any */
115 	uint8_t type;		/* int, short, long or string. */
116 	uint8_t in_type;	/* type of indirrection */
117 #define 			FILE_INVALID	0
118 #define 			FILE_BYTE	1
119 #define				FILE_SHORT	2
120 #define				FILE_DEFAULT	3
121 #define				FILE_LONG	4
122 #define				FILE_STRING	5
123 #define				FILE_DATE	6
124 #define				FILE_BESHORT	7
125 #define				FILE_BELONG	8
126 #define				FILE_BEDATE	9
127 #define				FILE_LESHORT	10
128 #define				FILE_LELONG	11
129 #define				FILE_LEDATE	12
130 #define				FILE_PSTRING	13
131 #define				FILE_LDATE	14
132 #define				FILE_BELDATE	15
133 #define				FILE_LELDATE	16
134 #define				FILE_REGEX	17
135 #define				FILE_BESTRING16	18
136 #define				FILE_LESTRING16	19
137 #define				FILE_SEARCH	20
138 #define				FILE_MEDATE	21
139 #define				FILE_MELDATE	22
140 #define				FILE_MELONG	23
141 #define				FILE_QUAD	24
142 #define				FILE_LEQUAD	25
143 #define				FILE_BEQUAD	26
144 #define				FILE_QDATE	27
145 #define				FILE_LEQDATE	28
146 #define				FILE_BEQDATE	29
147 #define				FILE_QLDATE	30
148 #define				FILE_LEQLDATE	31
149 #define				FILE_BEQLDATE	32
150 #define				FILE_NAMES_SIZE	33/* size of array to contain all names */
151 
152 #define IS_STRING(t) \
153 	((t) == FILE_STRING || \
154 	 (t) == FILE_PSTRING || \
155 	 (t) == FILE_BESTRING16 || \
156 	 (t) == FILE_LESTRING16 || \
157 	 (t) == FILE_REGEX || \
158 	 (t) == FILE_SEARCH || \
159 	 (t) == FILE_DEFAULT)
160 
161 #define FILE_FMT_NONE 0
162 #define FILE_FMT_NUM  1 /* "cduxXi" */
163 #define FILE_FMT_STR  2 /* "s" */
164 #define FILE_FMT_QUAD 3 /* "ll" */
165 
166 	/* Word 3 */
167 	uint8_t in_op;		/* operator for indirection */
168 	uint8_t mask_op;	/* operator for mask */
169 #ifdef ENABLE_CONDITIONALS
170 	uint8_t cond;		/* conditional type */
171 	uint8_t dummy1;
172 #else
173 	uint8_t dummy1;
174 	uint8_t dummy2;
175 #endif
176 
177 #define				FILE_OPS	"&|^+-*/%"
178 #define				FILE_OPAND	0
179 #define				FILE_OPOR	1
180 #define				FILE_OPXOR	2
181 #define				FILE_OPADD	3
182 #define				FILE_OPMINUS	4
183 #define				FILE_OPMULTIPLY	5
184 #define				FILE_OPDIVIDE	6
185 #define				FILE_OPMODULO	7
186 #define				FILE_OPS_MASK	0x07 /* mask for above ops */
187 #define				FILE_UNUSED_1	0x08
188 #define				FILE_UNUSED_2	0x10
189 #define				FILE_UNUSED_3	0x20
190 #define				FILE_OPINVERSE	0x40
191 #define				FILE_OPINDIRECT	0x80
192 
193 #ifdef ENABLE_CONDITIONALS
194 #define				COND_NONE	0
195 #define				COND_IF		1
196 #define				COND_ELIF	2
197 #define				COND_ELSE	3
198 #endif /* ENABLE_CONDITIONALS */
199 
200 	/* Word 4 */
201 	uint32_t offset;	/* offset to magic number */
202 	/* Word 5 */
203 	int32_t in_offset;	/* offset from indirection */
204 	/* Word 6 */
205 	uint32_t lineno;	/* line number in magic file */
206 	/* Word 7,8 */
207 	union {
208 		uint64_t _mask;	/* for use with numeric and date types */
209 		struct {
210 			uint32_t _count;	/* repeat/line count */
211 			uint32_t _flags;	/* modifier flags */
212 		} _s;		/* for use with string types */
213 	} _u;
214 #define num_mask _u._mask
215 #define str_count _u._s._count
216 #define str_flags _u._s._flags
217 
218 	/* Words 9-16 */
219 	union VALUETYPE {
220 		uint8_t b;
221 		uint16_t h;
222 		uint32_t l;
223 		uint64_t q;
224 		uint8_t hs[2];	/* 2 bytes of a fixed-endian "short" */
225 		uint8_t hl[4];	/* 4 bytes of a fixed-endian "long" */
226 		uint8_t hq[8];	/* 8 bytes of a fixed-endian "quad" */
227 		char s[MAXstring];	/* the search string or regex pattern */
228 	} value;		/* either number or string */
229 	/* Words 17..31 */
230 	char desc[MAXDESC];	/* description */
231 };
232 
233 #define BIT(A)   (1 << (A))
234 #define STRING_COMPACT_BLANK		BIT(0)
235 #define STRING_COMPACT_OPTIONAL_BLANK	BIT(1)
236 #define STRING_IGNORE_LOWERCASE		BIT(2)
237 #define STRING_IGNORE_UPPERCASE		BIT(3)
238 #define REGEX_OFFSET_START		BIT(4)
239 #define CHAR_COMPACT_BLANK		'B'
240 #define CHAR_COMPACT_OPTIONAL_BLANK	'b'
241 #define CHAR_IGNORE_LOWERCASE		'c'
242 #define CHAR_IGNORE_UPPERCASE		'C'
243 #define CHAR_REGEX_OFFSET_START		's'
244 #define STRING_IGNORE_CASE		(STRING_IGNORE_LOWERCASE|STRING_IGNORE_UPPERCASE)
245 
246 
247 /* list of magic entries */
248 struct mlist {
249 	struct magic *magic;		/* array of magic entries */
250 	uint32_t nmagic;			/* number of entries in array */
251 	int mapped;  /* allocation type: 0 => apprentice_file
252 		      *                  1 => apprentice_map + malloc
253 		      *                  2 => apprentice_map + mmap */
254 	struct mlist *next, *prev;
255 };
256 
257 struct magic_set {
258 	struct mlist *mlist;
259 	struct cont {
260 		size_t len;
261 		struct level_info {
262 			int32_t off;
263 			int got_match;
264 #ifdef ENABLE_CONDITIONALS
265 			int last_match;
266 			int last_cond;	/* used for error checking by parse() */
267 #endif
268 		} *li;
269 	} c;
270 	struct out {
271 		/* Accumulation buffer */
272 		char *buf;
273 		char *ptr;
274 		size_t left;
275 		size_t size;
276 		/* Printable buffer */
277 		char *pbuf;
278 		size_t psize;
279 	} o;
280 	uint32_t offset;
281 	int error;
282 	int flags;
283 	int haderr;
284 	const char *file;
285 	size_t line;			/* current magic line number */
286 
287 	/* data for searches */
288 	struct {
289 		const char *s;		/* start of search in original source */
290 		size_t s_len;		/* length of search region */
291 		size_t offset;		/* starting offset in source: XXX - should this be off_t? */
292 		size_t rm_len;		/* match length */
293 	} search;
294 
295 	union VALUETYPE ms_value;	/* either number or string */
296 };
297 
298 struct stat;
299 protected const char *file_fmttime(uint32_t, int);
300 protected int file_buffer(struct magic_set *, int, const char *, const void *,
301     size_t);
302 protected int file_fsmagic(struct magic_set *, const char *, struct stat *);
303 protected int file_pipe2file(struct magic_set *, int, const void *, size_t);
304 protected int file_printf(struct magic_set *, const char *, ...);
305 protected int file_reset(struct magic_set *);
306 protected int file_tryelf(struct magic_set *, int, const unsigned char *,
307     size_t);
308 protected int file_zmagic(struct magic_set *, int, const char *,
309     const unsigned char *, size_t);
310 protected int file_ascmagic(struct magic_set *, const unsigned char *, size_t);
311 protected int file_is_tar(struct magic_set *, const unsigned char *, size_t);
312 protected int file_softmagic(struct magic_set *, const unsigned char *, size_t);
313 protected struct mlist *file_apprentice(struct magic_set *, const char *, int);
314 protected uint64_t file_signextend(struct magic_set *, struct magic *,
315     uint64_t);
316 protected void file_delmagic(struct magic *, int type, size_t entries);
317 protected void file_badread(struct magic_set *);
318 protected void file_badseek(struct magic_set *);
319 protected void file_oomem(struct magic_set *, size_t);
320 protected void file_error(struct magic_set *, int, const char *, ...);
321 protected void file_magerror(struct magic_set *, const char *, ...);
322 protected void file_magwarn(struct magic_set *, const char *, ...);
323 protected void file_mdump(struct magic *);
324 protected void file_showstr(FILE *, const char *, size_t);
325 protected size_t file_mbswidth(const char *);
326 protected const char *file_getbuffer(struct magic_set *);
327 protected ssize_t sread(int, void *, size_t, int);
328 protected int file_check_mem(struct magic_set *, unsigned int);
329 
330 #ifndef COMPILE_ONLY
331 extern const char *file_names[];
332 extern const size_t file_nnames;
333 #endif
334 
335 #ifndef HAVE_STRERROR
336 extern int sys_nerr;
337 extern char *sys_errlist[];
338 #define strerror(e) \
339 	(((e) >= 0 && (e) < sys_nerr) ? sys_errlist[(e)] : "Unknown error")
340 #endif
341 
342 #ifndef HAVE_STRTOUL
343 #define strtoul(a, b, c)	strtol(a, b, c)
344 #endif
345 
346 #ifndef HAVE_SNPRINTF
347 int snprintf(char *, size_t, const char *, ...);
348 #endif
349 
350 #if defined(HAVE_MMAP) && defined(HAVE_SYS_MMAN_H) && !defined(QUICK)
351 #define QUICK
352 #endif
353 
354 #ifndef O_BINARY
355 #define O_BINARY	0
356 #endif
357 
358 #define FILE_RCSID(id) \
359 static const char *rcsid(const char *p) { \
360 	return rcsid(p = id); \
361 }
362 
363 #endif /* __file_h__ */
364