xref: /freebsd/contrib/file/doc/magic.man (revision 49b49cda41feabe3439f7318e8bf40e3896c7bf4)
1.\" $File: magic.man,v 1.86 2015/09/08 13:48:44 christos Exp $
2.Dd January 1, 2015
3.Dt MAGIC __FSECTION__
4.Os
5.\" install as magic.4 on USG, magic.5 on V7, Berkeley and Linux systems.
6.Sh NAME
7.Nm magic
8.Nd file command's magic pattern file
9.Sh DESCRIPTION
10This manual page documents the format of the magic file as
11used by the
12.Xr file __CSECTION__
13command, version __VERSION__.
14The
15.Xr file __CSECTION__
16command identifies the type of a file using,
17among other tests,
18a test for whether the file contains certain
19.Dq "magic patterns" .
20The file
21.Pa __MAGIC__
22specifies what patterns are to be tested for, what message or
23MIME type to print if a particular pattern is found,
24and additional information to extract from the file.
25.Pp
26Each line of the file specifies a test to be performed.
27A test compares the data starting at a particular offset
28in the file with a byte value, a string or a numeric value.
29If the test succeeds, a message is printed.
30The line consists of the following fields:
31.Bl -tag -width ".Dv message"
32.It Dv offset
33A number specifying the offset, in bytes, into the file of the data
34which is to be tested.
35.It Dv type
36The type of the data to be tested.
37The possible values are:
38.Bl -tag -width ".Dv lestring16"
39.It Dv byte
40A one-byte value.
41.It Dv short
42A two-byte value in this machine's native byte order.
43.It Dv long
44A four-byte value in this machine's native byte order.
45.It Dv quad
46An eight-byte value in this machine's native byte order.
47.It Dv float
48A 32-bit single precision IEEE floating point number in this machine's native byte order.
49.It Dv double
50A 64-bit double precision IEEE floating point number in this machine's native byte order.
51.It Dv string
52A string of bytes.
53The string type specification can be optionally followed
54by /[WwcCtbT]*.
55The
56.Dq W
57flag compacts whitespace in the target, which must
58contain at least one whitespace character.
59If the magic has
60.Dv n
61consecutive blanks, the target needs at least
62.Dv n
63consecutive blanks to match.
64The
65.Dq w
66flag treats every blank in the magic as an optional blank.
67The
68.Dq c
69flag specifies case insensitive matching: lower case
70characters in the magic match both lower and upper case characters in the
71target, whereas upper case characters in the magic only match upper case
72characters in the target.
73The
74.Dq C
75flag specifies case insensitive matching: upper case
76characters in the magic match both lower and upper case characters in the
77target, whereas lower case characters in the magic only match upper case
78characters in the target.
79To do a complete case insensitive match, specify both
80.Dq c
81and
82.Dq C .
83The
84.Dq t
85flag forces the test to be done for text files, while the
86.Dq b
87flag forces the test to be done for binary files.
88The
89.Dq T
90flag causes the string to be trimmed, i.e. leading and trailing whitespace
91is deleted before the string is printed.
92.It Dv pstring
93A Pascal-style string where the first byte/short/int is interpreted as the
94unsigned length.
95The length defaults to byte and can be specified as a modifier.
96The following modifiers are supported:
97.Bl -tag -compact -width B
98.It B
99A byte length (default).
100.It H
101A 2 byte big endian length.
102.It h
103A 2 byte big little length.
104.It L
105A 4 byte big endian length.
106.It l
107A 4 byte big little length.
108.It J
109The length includes itself in its count.
110.El
111The string is not NUL terminated.
112.Dq J
113is used rather than the more
114valuable
115.Dq I
116because this type of length is a feature of the JPEG
117format.
118.It Dv date
119A four-byte value interpreted as a UNIX date.
120.It Dv qdate
121A eight-byte value interpreted as a UNIX date.
122.It Dv ldate
123A four-byte value interpreted as a UNIX-style date, but interpreted as
124local time rather than UTC.
125.It Dv qldate
126An eight-byte value interpreted as a UNIX-style date, but interpreted as
127local time rather than UTC.
128.It Dv qwdate
129An eight-byte value interpreted as a Windows-style date.
130.It Dv beid3
131A 32-bit ID3 length in big-endian byte order.
132.It Dv beshort
133A two-byte value in big-endian byte order.
134.It Dv belong
135A four-byte value in big-endian byte order.
136.It Dv bequad
137An eight-byte value in big-endian byte order.
138.It Dv befloat
139A 32-bit single precision IEEE floating point number in big-endian byte order.
140.It Dv bedouble
141A 64-bit double precision IEEE floating point number in big-endian byte order.
142.It Dv bedate
143A four-byte value in big-endian byte order,
144interpreted as a Unix date.
145.It Dv beqdate
146An eight-byte value in big-endian byte order,
147interpreted as a Unix date.
148.It Dv beldate
149A four-byte value in big-endian byte order,
150interpreted as a UNIX-style date, but interpreted as local time rather
151than UTC.
152.It Dv beqldate
153An eight-byte value in big-endian byte order,
154interpreted as a UNIX-style date, but interpreted as local time rather
155than UTC.
156.It Dv beqwdate
157An eight-byte value in big-endian byte order,
158interpreted as a Windows-style date.
159.It Dv bestring16
160A two-byte unicode (UCS16) string in big-endian byte order.
161.It Dv leid3
162A 32-bit ID3 length in little-endian byte order.
163.It Dv leshort
164A two-byte value in little-endian byte order.
165.It Dv lelong
166A four-byte value in little-endian byte order.
167.It Dv lequad
168An eight-byte value in little-endian byte order.
169.It Dv lefloat
170A 32-bit single precision IEEE floating point number in little-endian byte order.
171.It Dv ledouble
172A 64-bit double precision IEEE floating point number in little-endian byte order.
173.It Dv ledate
174A four-byte value in little-endian byte order,
175interpreted as a UNIX date.
176.It Dv leqdate
177An eight-byte value in little-endian byte order,
178interpreted as a UNIX date.
179.It Dv leldate
180A four-byte value in little-endian byte order,
181interpreted as a UNIX-style date, but interpreted as local time rather
182than UTC.
183.It Dv leqldate
184An eight-byte value in little-endian byte order,
185interpreted as a UNIX-style date, but interpreted as local time rather
186than UTC.
187.It Dv leqwdate
188An eight-byte value in little-endian byte order,
189interpreted as a Windows-style date.
190.It Dv lestring16
191A two-byte unicode (UCS16) string in little-endian byte order.
192.It Dv melong
193A four-byte value in middle-endian (PDP-11) byte order.
194.It Dv medate
195A four-byte value in middle-endian (PDP-11) byte order,
196interpreted as a UNIX date.
197.It Dv meldate
198A four-byte value in middle-endian (PDP-11) byte order,
199interpreted as a UNIX-style date, but interpreted as local time rather
200than UTC.
201.It Dv indirect
202Starting at the given offset, consult the magic database again.
203The offset of the
204.Dv indirect
205magic is by default absolute in the file, but one can specify
206.Dv /r
207to indicate that the offset is relative from the beginning of the entry.
208.It Dv name
209Define a
210.Dq named
211magic instance that can be called from another
212.Dv use
213magic entry, like a subroutine call.
214Named instance direct magic offsets are relative to the offset of the
215previous matched entry, but indirect offsets are relative to the beginning
216of the file as usual.
217Named magic entries always match.
218.It Dv use
219Recursively call the named magic starting from the current offset.
220If the name of the referenced begins with a
221.Dv ^
222then the endianness of the magic is switched; if the magic mentioned
223.Dv leshort
224for example,
225it is treated as
226.Dv beshort
227and vice versa.
228This is useful to avoid duplicating the rules for different endianness.
229.It Dv regex
230A regular expression match in extended POSIX regular expression syntax
231(like egrep).
232Regular expressions can take exponential time to process, and their
233performance is hard to predict, so their use is discouraged.
234When used in production environments, their performance
235should be carefully checked.
236The size of the string to search should also be limited by specifying
237.Dv /<length> ,
238to avoid performance issues scanning long files.
239The type specification can also be optionally followed by
240.Dv /[c][s][l] .
241The
242.Dq c
243flag makes the match case insensitive, while the
244.Dq s
245flag update the offset to the start offset of the match, rather than the end.
246The
247.Dq l
248modifier, changes the limit of length to mean number of lines instead of a
249byte count.
250Lines are delimited by the platforms native line delimiter.
251When a line count is specified, an implicit byte count also computed assuming
252each line is 80 characters long.
253If neither a byte or line count is specified, the search is limited automatically
254to 8KiB.
255.Dv ^
256and
257.Dv $
258match the beginning and end of individual lines, respectively,
259not beginning and end of file.
260.It Dv search
261A literal string search starting at the given offset.
262The same modifier flags can be used as for string patterns.
263The search expression must contain the range in the form
264.Dv /number,
265that is the number of positions at which the match will be
266attempted, starting from the start offset.
267This is suitable for
268searching larger binary expressions with variable offsets, using
269.Dv \e
270escapes for special characters.
271The order of modifier and number is not relevant.
272.It Dv default
273This is intended to be used with the test
274.Em x
275(which is always true) and it has no type.
276It matches when no other test at that continuation level has matched before.
277Clearing that matched tests for a continuation level, can be done using the
278.Dv clear
279test.
280.It Dv clear
281This test is always true and clears the match flag for that continuation level.
282It is intended to be used with the
283.Dv default
284test.
285.El
286.Pp
287For compatibility with the Single
288.Ux
289Standard, the type specifiers
290.Dv dC
291and
292.Dv d1
293are equivalent to
294.Dv byte ,
295the type specifiers
296.Dv uC
297and
298.Dv u1
299are equivalent to
300.Dv ubyte ,
301the type specifiers
302.Dv dS
303and
304.Dv d2
305are equivalent to
306.Dv short ,
307the type specifiers
308.Dv uS
309and
310.Dv u2
311are equivalent to
312.Dv ushort ,
313the type specifiers
314.Dv dI ,
315.Dv dL ,
316and
317.Dv d4
318are equivalent to
319.Dv long ,
320the type specifiers
321.Dv uI ,
322.Dv uL ,
323and
324.Dv u4
325are equivalent to
326.Dv ulong ,
327the type specifier
328.Dv d8
329is equivalent to
330.Dv quad ,
331the type specifier
332.Dv u8
333is equivalent to
334.Dv uquad ,
335and the type specifier
336.Dv s
337is equivalent to
338.Dv string .
339In addition, the type specifier
340.Dv dQ
341is equivalent to
342.Dv quad
343and the type specifier
344.Dv uQ
345is equivalent to
346.Dv uquad .
347.Pp
348Each top-level magic pattern (see below for an explanation of levels)
349is classified as text or binary according to the types used.
350Types
351.Dq regex
352and
353.Dq search
354are classified as text tests, unless non-printable characters are used
355in the pattern.
356All other tests are classified as binary.
357A top-level
358pattern is considered to be a test text when all its patterns are text
359patterns; otherwise, it is considered to be a binary pattern.
360When
361matching a file, binary patterns are tried first; if no match is
362found, and the file looks like text, then its encoding is determined
363and the text patterns are tried.
364.Pp
365The numeric types may optionally be followed by
366.Dv \*[Am]
367and a numeric value,
368to specify that the value is to be AND'ed with the
369numeric value before any comparisons are done.
370Prepending a
371.Dv u
372to the type indicates that ordered comparisons should be unsigned.
373.It Dv test
374The value to be compared with the value from the file.
375If the type is
376numeric, this value
377is specified in C form; if it is a string, it is specified as a C string
378with the usual escapes permitted (e.g. \en for new-line).
379.Pp
380Numeric values
381may be preceded by a character indicating the operation to be performed.
382It may be
383.Dv = ,
384to specify that the value from the file must equal the specified value,
385.Dv \*[Lt] ,
386to specify that the value from the file must be less than the specified
387value,
388.Dv \*[Gt] ,
389to specify that the value from the file must be greater than the specified
390value,
391.Dv \*[Am] ,
392to specify that the value from the file must have set all of the bits
393that are set in the specified value,
394.Dv ^ ,
395to specify that the value from the file must have clear any of the bits
396that are set in the specified value, or
397.Dv ~ ,
398the value specified after is negated before tested.
399.Dv x ,
400to specify that any value will match.
401If the character is omitted, it is assumed to be
402.Dv = .
403Operators
404.Dv \*[Am] ,
405.Dv ^ ,
406and
407.Dv ~
408don't work with floats and doubles.
409The operator
410.Dv !\&
411specifies that the line matches if the test does
412.Em not
413succeed.
414.Pp
415Numeric values are specified in C form; e.g.
416.Dv 13
417is decimal,
418.Dv 013
419is octal, and
420.Dv 0x13
421is hexadecimal.
422.Pp
423Numeric operations are not performed on date types, instead the numeric
424value is interpreted as an offset.
425.Pp
426For string values, the string from the
427file must match the specified string.
428The operators
429.Dv = ,
430.Dv \*[Lt]
431and
432.Dv \*[Gt]
433(but not
434.Dv \*[Am] )
435can be applied to strings.
436The length used for matching is that of the string argument
437in the magic file.
438This means that a line can match any non-empty string (usually used to
439then print the string), with
440.Em \*[Gt]\e0
441(because all non-empty strings are greater than the empty string).
442.Pp
443Dates are treated as numerical values in the respective internal
444representation.
445.Pp
446The special test
447.Em x
448always evaluates to true.
449.It Dv message
450The message to be printed if the comparison succeeds.
451If the string contains a
452.Xr printf 3
453format specification, the value from the file (with any specified masking
454performed) is printed using the message as the format string.
455If the string begins with
456.Dq \eb ,
457the message printed is the remainder of the string with no whitespace
458added before it: multiple matches are normally separated by a single
459space.
460.El
461.Pp
462An APPLE 4+4 character APPLE creator and type can be specified as:
463.Bd -literal -offset indent
464!:apple	CREATYPE
465.Ed
466.Pp
467A MIME type is given on a separate line, which must be the next
468non-blank or comment line after the magic line that identifies the
469file type, and has the following format:
470.Bd -literal -offset indent
471!:mime	MIMETYPE
472.Ed
473.Pp
474i.e. the literal string
475.Dq !:mime
476followed by the MIME type.
477.Pp
478An optional strength can be supplied on a separate line which refers to
479the current magic description using the following format:
480.Bd -literal -offset indent
481!:strength OP VALUE
482.Ed
483.Pp
484The operand
485.Dv OP
486can be:
487.Dv + ,
488.Dv - ,
489.Dv * ,
490or
491.Dv /
492and
493.Dv VALUE
494is a constant between 0 and 255.
495This constant is applied using the specified operand
496to the currently computed default magic strength.
497.Pp
498Some file formats contain additional information which is to be printed
499along with the file type or need additional tests to determine the true
500file type.
501These additional tests are introduced by one or more
502.Em \*[Gt]
503characters preceding the offset.
504The number of
505.Em \*[Gt]
506on the line indicates the level of the test; a line with no
507.Em \*[Gt]
508at the beginning is considered to be at level 0.
509Tests are arranged in a tree-like hierarchy:
510if the test on a line at level
511.Em n
512succeeds, all following tests at level
513.Em n+1
514are performed, and the messages printed if the tests succeed, until a line
515with level
516.Em n
517(or less) appears.
518For more complex files, one can use empty messages to get just the
519"if/then" effect, in the following way:
520.Bd -literal -offset indent
5210      string   MZ
522\*[Gt]0x18  leshort  \*[Lt]0x40   MS-DOS executable
523\*[Gt]0x18  leshort  \*[Gt]0x3f   extended PC executable (e.g., MS Windows)
524.Ed
525.Pp
526Offsets do not need to be constant, but can also be read from the file
527being examined.
528If the first character following the last
529.Em \*[Gt]
530is a
531.Em \&(
532then the string after the parenthesis is interpreted as an indirect offset.
533That means that the number after the parenthesis is used as an offset in
534the file.
535The value at that offset is read, and is used again as an offset
536in the file.
537Indirect offsets are of the form:
538.Em (( x [.[bislBISL]][+\-][ y ]) .
539The value of
540.Em x
541is used as an offset in the file.
542A byte, id3 length, short or long is read at that offset depending on the
543.Em [bislBISLm]
544type specifier.
545The capitalized types interpret the number as a big endian
546value, whereas the small letter versions interpret the number as a little
547endian value;
548the
549.Em m
550type interprets the number as a middle endian (PDP-11) value.
551To that number the value of
552.Em y
553is added and the result is used as an offset in the file.
554The default type if one is not specified is long.
555.Pp
556That way variable length structures can be examined:
557.Bd -literal -offset indent
558# MS Windows executables are also valid MS-DOS executables
5590           string  MZ
560\*[Gt]0x18       leshort \*[Lt]0x40   MZ executable (MS-DOS)
561# skip the whole block below if it is not an extended executable
562\*[Gt]0x18       leshort \*[Gt]0x3f
563\*[Gt]\*[Gt](0x3c.l)  string  PE\e0\e0  PE executable (MS-Windows)
564\*[Gt]\*[Gt](0x3c.l)  string  LX\e0\e0  LX executable (OS/2)
565.Ed
566.Pp
567This strategy of examining has a drawback: You must make sure that
568you eventually print something, or users may get empty output (like, when
569there is neither PE\e0\e0 nor LE\e0\e0 in the above example)
570.Pp
571If this indirect offset cannot be used directly, simple calculations are
572possible: appending
573.Em [+-*/%\*[Am]|^]number
574inside parentheses allows one to modify
575the value read from the file before it is used as an offset:
576.Bd -literal -offset indent
577# MS Windows executables are also valid MS-DOS executables
5780           string  MZ
579# sometimes, the value at 0x18 is less that 0x40 but there's still an
580# extended executable, simply appended to the file
581\*[Gt]0x18       leshort \*[Lt]0x40
582\*[Gt]\*[Gt](4.s*512) leshort 0x014c  COFF executable (MS-DOS, DJGPP)
583\*[Gt]\*[Gt](4.s*512) leshort !0x014c MZ executable (MS-DOS)
584.Ed
585.Pp
586Sometimes you do not know the exact offset as this depends on the length or
587position (when indirection was used before) of preceding fields.
588You can specify an offset relative to the end of the last up-level
589field using
590.Sq \*[Am]
591as a prefix to the offset:
592.Bd -literal -offset indent
5930           string  MZ
594\*[Gt]0x18       leshort \*[Gt]0x3f
595\*[Gt]\*[Gt](0x3c.l)  string  PE\e0\e0    PE executable (MS-Windows)
596# immediately following the PE signature is the CPU type
597\*[Gt]\*[Gt]\*[Gt]\*[Am]0       leshort 0x14c     for Intel 80386
598\*[Gt]\*[Gt]\*[Gt]\*[Am]0       leshort 0x184     for DEC Alpha
599.Ed
600.Pp
601Indirect and relative offsets can be combined:
602.Bd -literal -offset indent
6030             string  MZ
604\*[Gt]0x18         leshort \*[Lt]0x40
605\*[Gt]\*[Gt](4.s*512)   leshort !0x014c MZ executable (MS-DOS)
606# if it's not COFF, go back 512 bytes and add the offset taken
607# from byte 2/3, which is yet another way of finding the start
608# of the extended executable
609\*[Gt]\*[Gt]\*[Gt]\*[Am](2.s-514) string  LE      LE executable (MS Windows VxD driver)
610.Ed
611.Pp
612Or the other way around:
613.Bd -literal -offset indent
6140                 string  MZ
615\*[Gt]0x18             leshort \*[Gt]0x3f
616\*[Gt]\*[Gt](0x3c.l)        string  LE\e0\e0  LE executable (MS-Windows)
617# at offset 0x80 (-4, since relative offsets start at the end
618# of the up-level match) inside the LE header, we find the absolute
619# offset to the code area, where we look for a specific signature
620\*[Gt]\*[Gt]\*[Gt](\*[Am]0x7c.l+0x26) string  UPX     \eb, UPX compressed
621.Ed
622.Pp
623Or even both!
624.Bd -literal -offset indent
6250                string  MZ
626\*[Gt]0x18            leshort \*[Gt]0x3f
627\*[Gt]\*[Gt](0x3c.l)       string  LE\e0\e0 LE executable (MS-Windows)
628# at offset 0x58 inside the LE header, we find the relative offset
629# to a data area where we look for a specific signature
630\*[Gt]\*[Gt]\*[Gt]\*[Am](\*[Am]0x54.l-3)  string  UNACE  \eb, ACE self-extracting archive
631.Ed
632.Pp
633If you have to deal with offset/length pairs in your file, even the
634second value in a parenthesized expression can be taken from the file itself,
635using another set of parentheses.
636Note that this additional indirect offset is always relative to the
637start of the main indirect offset.
638.Bd -literal -offset indent
6390                 string       MZ
640\*[Gt]0x18             leshort      \*[Gt]0x3f
641\*[Gt]\*[Gt](0x3c.l)        string       PE\e0\e0 PE executable (MS-Windows)
642# search for the PE section called ".idata"...
643\*[Gt]\*[Gt]\*[Gt]\*[Am]0xf4          search/0x140 .idata
644# ...and go to the end of it, calculated from start+length;
645# these are located 14 and 10 bytes after the section name
646\*[Gt]\*[Gt]\*[Gt]\*[Gt](\*[Am]0xe.l+(-4)) string       PK\e3\e4 \eb, ZIP self-extracting archive
647.Ed
648.Pp
649If you have a list of known avalues at a particular continuation level,
650and you want to provide a switch-like default case:
651.Bd -literal -offset indent
652# clear that continuation level match
653\*[Gt]18	clear
654\*[Gt]18	lelong	1	one
655\*[Gt]18	lelong	2	two
656\*[Gt]18	default	x
657# print default match
658\*[Gt]\*[Gt]18	lelong	x	unmatched 0x%x
659.Ed
660.Sh SEE ALSO
661.Xr file __CSECTION__
662\- the command that reads this file.
663.Sh BUGS
664The formats
665.Dv long ,
666.Dv belong ,
667.Dv lelong ,
668.Dv melong ,
669.Dv short ,
670.Dv beshort ,
671and
672.Dv leshort
673do not depend on the length of the C data types
674.Dv short
675and
676.Dv long
677on the platform, even though the Single
678.Ux
679Specification implies that they do.  However, as OS X Mountain Lion has
680passed the Single
681.Ux
682Specification validation suite, and supplies a version of
683.Xr file __CSECTION__
684in which they do not depend on the sizes of the C data types and that is
685built for a 64-bit environment in which
686.Dv long
687is 8 bytes rather than 4 bytes, presumably the validation suite does not
688test whether, for example
689.Dv long
690refers to an item with the same size as the C data type
691.Dv long .
692There should probably be
693.Dv type
694names
695.Dv int8 ,
696.Dv uint8 ,
697.Dv int16 ,
698.Dv uint16 ,
699.Dv int32 ,
700.Dv uint32 ,
701.Dv int64 ,
702and
703.Dv uint64 ,
704and specified-byte-order variants of them,
705to make it clearer that those types have specified widths.
706.\"
707.\" From: guy@sun.uucp (Guy Harris)
708.\" Newsgroups: net.bugs.usg
709.\" Subject: /etc/magic's format isn't well documented
710.\" Message-ID: <2752@sun.uucp>
711.\" Date: 3 Sep 85 08:19:07 GMT
712.\" Organization: Sun Microsystems, Inc.
713.\" Lines: 136
714.\"
715.\" Here's a manual page for the format accepted by the "file" made by adding
716.\" the changes I posted to the S5R2 version.
717.\"
718.\" Modified for Ian Darwin's version of the file command.
719