1MAGIC(5) BSD File Formats Manual MAGIC(5)
2
4 magic — file command's magic pattern file
5
7 This manual page documents the format of the magic file as used by the
8 file(1) command, version 5.11. The file(1) command identifies the type
9 of a file using, among other tests, a test for whether the file contains
10 certain “magic patterns”. The file /usr/share/misc/magic specifies what
11 patterns are to be tested for, what message or MIME type to print if a
12 particular pattern is found, and additional information to extract from
13 the file.
14 Each line of the file specifies a test to be performed. A test compares
16 the data starting at a particular offset in the file with a byte value, a
17 string or a numeric value. If the test succeeds, a message is printed.
18 The line consists of the following fields:
19 offset A number specifying the offset, in bytes, into the file of the
21 data which is to be tested.
22 type The type of the data to be tested. The possible values are:
24 byte A one-byte value.
26 short A two-byte value in this machine's native byte
28 order.
29 long A four-byte value in this machine's native byte
31 order.
32 quad An eight-byte value in this machine's native byte
34 order.
35 float A 32-bit single precision IEEE floating point number
37 in this machine's native byte order.
38 double A 64-bit double precision IEEE floating point number
40 in this machine's native byte order.
41 string A string of bytes. The string type specification
43 can be optionally followed by /[WwcCtb]*. The “W”
44 flag compacts whitespace in the target, which must
45 contain at least one whitespace character. If the
46 magic has n consecutive blanks, the target needs at
47 least n consecutive blanks to match. The “w” flag
48 treats every blank in the magic as an optional
49 blank. The “c” flag specifies case insensitive
50 matching: lower case characters in the magic match
51 both lower and upper case characters in the target,
52 whereas upper case characters in the magic only
53 match upper case characters in the target. The “C”
54 flag specifies case insensitive matching: upper case
55 characters in the magic match both lower and upper
56 case characters in the target, whereas lower case
57 characters in the magic only match upper case char‐
58 acters in the target. To do a complete case insen‐
59 sitive match, specify both “c” and “C”. The “t”
60 flag forces the test to be done for text files,
61 while the “b” flag forces the test to be done for
62 binary files.
63 pstring A Pascal-style string where the first byte/short/int
65 is interpreted as the an unsigned length. The
66 length defaults to byte and can be specified as a
67 modifier. The following modifiers are supported:
68 B A byte length (default).
69 H A 2 byte big endian length.
70 h A 2 byte big little length.
71 L A 4 byte big endian length.
72 l A 4 byte big little length.
73 J The length includes itself in its count.
74 The string is not NUL terminated. “J” is used
75 rather than the more valuable “I” because this type
76 of length is a feature of the JPEG format.
77 date A four-byte value interpreted as a UNIX date.
79 qdate A eight-byte value interpreted as a UNIX date.
81 ldate A four-byte value interpreted as a UNIX-style date,
83 but interpreted as local time rather than UTC.
84 qldate An eight-byte value interpreted as a UNIX-style
86 date, but interpreted as local time rather than UTC.
87 beid3 A 32-bit ID3 length in big-endian byte order.
89 beshort A two-byte value in big-endian byte order.
91 belong A four-byte value in big-endian byte order.
93 bequad An eight-byte value in big-endian byte order.
95 befloat A 32-bit single precision IEEE floating point number
97 in big-endian byte order.
98 bedouble A 64-bit double precision IEEE floating point number
100 in big-endian byte order.
101 bedate A four-byte value in big-endian byte order, inter‐
103 preted as a Unix date.
104 beqdate An eight-byte value in big-endian byte order, inter‐
106 preted as a Unix date.
107 beldate A four-byte value in big-endian byte order, inter‐
109 preted as a UNIX-style date, but interpreted as
110 local time rather than UTC.
111 beqldate An eight-byte value in big-endian byte order, inter‐
113 preted as a UNIX-style date, but interpreted as
114 local time rather than UTC.
115 bestring16 A two-byte unicode (UCS16) string in big-endian byte
117 order.
118 leid3 A 32-bit ID3 length in little-endian byte order.
120 leshort A two-byte value in little-endian byte order.
122 lelong A four-byte value in little-endian byte order.
124 lequad An eight-byte value in little-endian byte order.
126 lefloat A 32-bit single precision IEEE floating point number
128 in little-endian byte order.
129 ledouble A 64-bit double precision IEEE floating point number
131 in little-endian byte order.
132 ledate A four-byte value in little-endian byte order,
134 interpreted as a UNIX date.
135 leqdate An eight-byte value in little-endian byte order,
137 interpreted as a UNIX date.
138 leldate A four-byte value in little-endian byte order,
140 interpreted as a UNIX-style date, but interpreted as
141 local time rather than UTC.
142 leqldate An eight-byte value in little-endian byte order,
144 interpreted as a UNIX-style date, but interpreted as
145 local time rather than UTC.
146 lestring16 A two-byte unicode (UCS16) string in little-endian
148 byte order.
149 melong A four-byte value in middle-endian (PDP-11) byte
151 order.
152 medate A four-byte value in middle-endian (PDP-11) byte
154 order, interpreted as a UNIX date.
155 meldate A four-byte value in middle-endian (PDP-11) byte
157 order, interpreted as a UNIX-style date, but inter‐
158 preted as local time rather than UTC.
159 indirect Starting at the given offset, consult the magic
161 database again.
162 regex A regular expression match in extended POSIX regular
164 expression syntax (like egrep). Regular expressions
165 can take exponential time to process, and their per‐
166 formance is hard to predict, so their use is dis‐
167 couraged. When used in production environments,
168 their performance should be carefully checked. The
169 type specification can be optionally followed by
170 /[c][s]. The “c” flag makes the match case insensi‐
171 tive, while the “s” flag update the offset to the
172 start offset of the match, rather than the end. The
173 regular expression is tested against line N + 1
174 onwards, where N is the given offset. Line endings
175 are assumed to be in the machine's native format. ^
176 and $ match the beginning and end of individual
177 lines, respectively, not beginning and end of file.
178 search A literal string search starting at the given off‐
180 set. The same modifier flags can be used as for
181 string patterns. The modifier flags (if any) must
182 be followed by /number the range, that is, the num‐
183 ber of positions at which the match will be
184 attempted, starting from the start offset. This is
185 suitable for searching larger binary expressions
186 with variable offsets, using \ escapes for special
187 characters. The offset works as for regex.
188 default This is intended to be used with the test x (which
190 is always true) and a message that is to be used if
191 there are no other matches.
192 Each top-level magic pattern (see below for an explanation of
194 levels) is classified as text or binary according to the types
195 used. Types “regex” and “search” are classified as text tests,
196 unless non-printable characters are used in the pattern. All
197 other tests are classified as binary. A top-level pattern is
198 considered to be a test text when all its patterns are text pat‐
199 terns; otherwise, it is considered to be a binary pattern. When
200 matching a file, binary patterns are tried first; if no match is
201 found, and the file looks like text, then its encoding is deter‐
202 mined and the text patterns are tried.
203 The numeric types may optionally be followed by & and a numeric
205 value, to specify that the value is to be AND'ed with the
206 numeric value before any comparisons are done. Prepending a u
207 to the type indicates that ordered comparisons should be
208 unsigned.
209 test The value to be compared with the value from the file. If the
211 type is numeric, this value is specified in C form; if it is a
212 string, it is specified as a C string with the usual escapes
213 permitted (e.g. \n for new-line).
214 Numeric values may be preceded by a character indicating the
216 operation to be performed. It may be =, to specify that the
217 value from the file must equal the specified value, <, to spec‐
218 ify that the value from the file must be less than the specified
219 value, >, to specify that the value from the file must be
220 greater than the specified value, &, to specify that the value
221 from the file must have set all of the bits that are set in the
222 specified value, ^, to specify that the value from the file must
223 have clear any of the bits that are set in the specified value,
224 or ~, the value specified after is negated before tested. x, to
225 specify that any value will match. If the character is omitted,
226 it is assumed to be =. Operators &, ^, and ~ don't work with
227 floats and doubles. The operator ! specifies that the line
228 matches if the test does not succeed.
229 Numeric values are specified in C form; e.g. 13 is decimal, 013
231 is octal, and 0x13 is hexadecimal.
232 For string values, the string from the file must match the spec‐
234 ified string. The operators =, < and > (but not &) can be
235 applied to strings. The length used for matching is that of the
236 string argument in the magic file. This means that a line can
237 match any non-empty string (usually used to then print the
238 string), with >\0 (because all non-empty strings are greater
239 than the empty string).
240 The special test x always evaluates to true.
242 message The message to be printed if the comparison succeeds. If the
244 string contains a printf(3) format specification, the value from
245 the file (with any specified masking performed) is printed using
246 the message as the format string. If the string begins with
247 “\b”, the message printed is the remainder of the string with no
248 whitespace added before it: multiple matches are normally sepa‐
249 rated by a single space.
250 An APPLE 4+4 character APPLE creator and type can be specified as:
252 !:apple CREATYPE
254 A MIME type is given on a separate line, which must be the next non-blank
256 or comment line after the magic line that identifies the file type, and
257 has the following format:
258 !:mime MIMETYPE
260 i.e. the literal string “!:mime” followed by the MIME type.
262 An optional strength can be supplied on a separate line which refers to
264 the current magic description using the following format:
265 !:strength OP VALUE
267 The operand OP can be: +, -, *, or / and VALUE is a constant between 0
269 and 255. This constant is applied using the specified operand to the
270 currently computed default magic strength.
271 Some file formats contain additional information which is to be printed
273 along with the file type or need additional tests to determine the true
274 file type. These additional tests are introduced by one or more > char‐
275 acters preceding the offset. The number of > on the line indicates the
276 level of the test; a line with no > at the beginning is considered to be
277 at level 0. Tests are arranged in a tree-like hierarchy: if the test on
278 a line at level n succeeds, all following tests at level n+1 are per‐
279 formed, and the messages printed if the tests succeed, until a line with
280 level n (or less) appears. For more complex files, one can use empty
281 messages to get just the "if/then" effect, in the following way:
282 0 string MZ
284 >0x18 leshort <0x40 MS-DOS executable
285 >0x18 leshort >0x3f extended PC executable (e.g., MS Windows)
286 Offsets do not need to be constant, but can also be read from the file
288 being examined. If the first character following the last > is a ( then
289 the string after the parenthesis is interpreted as an indirect offset.
290 That means that the number after the parenthesis is used as an offset in
291 the file. The value at that offset is read, and is used again as an off‐
292 set in the file. Indirect offsets are of the form: (( x
293 [.[bislBISL]][+-][ y ]). The value of x is used as an offset in the
294 file. A byte, id3 length, short or long is read at that offset depending
295 on the [bislBISLm] type specifier. The capitalized types interpret the
296 number as a big endian value, whereas the small letter versions interpret
297 the number as a little endian value; the m type interprets the number as
298 a middle endian (PDP-11) value. To that number the value of y is added
299 and the result is used as an offset in the file. The default type if one
300 is not specified is long.
301 That way variable length structures can be examined:
303 # MS Windows executables are also valid MS-DOS executables
305 0 string MZ
306 >0x18 leshort <0x40 MZ executable (MS-DOS)
307 # skip the whole block below if it is not an extended executable
308 >0x18 leshort >0x3f
309 >>(0x3c.l) string PE\0\0 PE executable (MS-Windows)
310 >>(0x3c.l) string LX\0\0 LX executable (OS/2)
311 This strategy of examining has a drawback: You must make sure that you
313 eventually print something, or users may get empty output (like, when
314 there is neither PE\0\0 nor LE\0\0 in the above example)
315 If this indirect offset cannot be used directly, simple calculations are
317 possible: appending [+-*/%&|^]number inside parentheses allows one to
318 modify the value read from the file before it is used as an offset:
319 # MS Windows executables are also valid MS-DOS executables
321 0 string MZ
322 # sometimes, the value at 0x18 is less that 0x40 but there's still an
323 # extended executable, simply appended to the file
324 >0x18 leshort <0x40
325 >>(4.s*512) leshort 0x014c COFF executable (MS-DOS, DJGPP)
326 >>(4.s*512) leshort !0x014c MZ executable (MS-DOS)
327 Sometimes you do not know the exact offset as this depends on the length
329 or position (when indirection was used before) of preceding fields. You
330 can specify an offset relative to the end of the last up-level field
331 using ‘&’ as a prefix to the offset:
332 0 string MZ
334 >0x18 leshort >0x3f
335 >>(0x3c.l) string PE\0\0 PE executable (MS-Windows)
336 # immediately following the PE signature is the CPU type
337 >>>&0 leshort 0x14c for Intel 80386
338 >>>&0 leshort 0x184 for DEC Alpha
339 Indirect and relative offsets can be combined:
341 0 string MZ
343 >0x18 leshort <0x40
344 >>(4.s*512) leshort !0x014c MZ executable (MS-DOS)
345 # if it's not COFF, go back 512 bytes and add the offset taken
346 # from byte 2/3, which is yet another way of finding the start
347 # of the extended executable
348 >>>&(2.s-514) string LE LE executable (MS Windows VxD driver)
349 Or the other way around:
351 0 string MZ
353 >0x18 leshort >0x3f
354 >>(0x3c.l) string LE\0\0 LE executable (MS-Windows)
355 # at offset 0x80 (-4, since relative offsets start at the end
356 # of the up-level match) inside the LE header, we find the absolute
357 # offset to the code area, where we look for a specific signature
358 >>>(&0x7c.l+0x26) string UPX \b, UPX compressed
359 Or even both!
361 0 string MZ
363 >0x18 leshort >0x3f
364 >>(0x3c.l) string LE\0\0 LE executable (MS-Windows)
365 # at offset 0x58 inside the LE header, we find the relative offset
366 # to a data area where we look for a specific signature
367 >>>&(&0x54.l-3) string UNACE \b, ACE self-extracting archive
368 Finally, if you have to deal with offset/length pairs in your file, even
370 the second value in a parenthesized expression can be taken from the file
371 itself, using another set of parentheses. Note that this additional
372 indirect offset is always relative to the start of the main indirect off‐
373 set.
374 0 string MZ
376 >0x18 leshort >0x3f
377 >>(0x3c.l) string PE\0\0 PE executable (MS-Windows)
378 # search for the PE section called ".idata"...
379 >>>&0xf4 search/0x140 .idata
380 # ...and go to the end of it, calculated from start+length;
381 # these are located 14 and 10 bytes after the section name
382 >>>>(&0xe.l+(-4)) string PK\3\4 \b, ZIP self-extracting archive
383
385 file(1) - the command that reads this file.
386
388 The formats long, belong, lelong, melong, short, beshort, leshort, date,
389 bedate, medate, ledate, beldate, leldate, and meldate are system-depen‐
390 dent; perhaps they should be specified as a number of bytes (2B, 4B,
391 etc), since the files being recognized typically come from a system on
392 which the lengths are invariant.
393BSD April 20, 2011 BSD