xref: /freebsd/lib/libmd/mdX.3 (revision 8c784bb8cf36911b828652f0bf7e88f443abec50)
1.\"
2.\" ----------------------------------------------------------------------------
3.\" "THE BEER-WARE LICENSE" (Revision 42):
4.\" <phk@FreeBSD.org> wrote this file.  As long as you retain this notice you
5.\" can do whatever you want with this stuff. If we meet some day, and you think
6.\" this stuff is worth it, you can buy me a beer in return.   Poul-Henning Kamp
7.\" ----------------------------------------------------------------------------
8.\"
9.\" $FreeBSD$
10.\"
11.Dd May 21, 2019
12.Dt MDX 3
13.Os
14.Sh NAME
15.Nm MDXInit ,
16.Nm MDXUpdate ,
17.Nm MDXPad ,
18.Nm MDXFinal ,
19.Nm MDXEnd ,
20.Nm MDXFile ,
21.Nm MDXFileChunk ,
22.Nm MDXData
23.Nd calculate the RSA Data Security, Inc., ``MDX'' message digest
24.Sh LIBRARY
25.Lb libmd
26.Sh SYNOPSIS
27.In sys/types.h
28.In mdX.h
29.Ft void
30.Fn MDXInit "MDX_CTX *context"
31.Ft void
32.Fn MDXUpdate "MDX_CTX *context" "const void *data" "unsigned int len"
33.Ft void
34.Fn MDXPad "MDX_CTX *context"
35.Ft void
36.Fn MDXFinal "unsigned char digest[16]" "MDX_CTX *context"
37.Ft "char *"
38.Fn MDXEnd "MDX_CTX *context" "char *buf"
39.Ft "char *"
40.Fn MDXFile "const char *filename" "char *buf"
41.Ft "char *"
42.Fn MDXFileChunk "const char *filename" "char *buf" "off_t offset" "off_t length"
43.Ft "char *"
44.Fn MDXData "const void *data" "unsigned int len" "char *buf"
45.Sh DESCRIPTION
46The MDX functions calculate a 128-bit cryptographic checksum (digest)
47for any number of input bytes.
48A cryptographic checksum is a one-way
49hash-function, that is, you cannot find (except by exhaustive search)
50the input corresponding to a particular output.
51This net result is a
52.Dq fingerprint
53of the input-data, which does not disclose the actual input.
54.Pp
55MD4 is the fastest and MD5 is somewhat slower.
56MD4 has now been broken; it should only be used where necessary for
57backward compatibility.
58MD5 has not yet (1999-02-11) been broken, but sufficient attacks have been
59made that its security is in some doubt.
60The attacks on both MD4 and MD5
61are both in the nature of finding
62.Dq collisions
63\[en]
64that is, multiple
65inputs which hash to the same value; it is still unlikely for an attacker
66to be able to determine the exact original input given a hash value.
67.Pp
68The
69.Fn MDXInit ,
70.Fn MDXUpdate ,
71and
72.Fn MDXFinal
73functions are the core functions.
74Allocate an
75.Vt MDX_CTX ,
76initialize it with
77.Fn MDXInit ,
78run over the data with
79.Fn MDXUpdate ,
80and finally extract the result using
81.Fn MDXFinal ,
82which will also erase the
83.Vt MDX_CTX .
84.Pp
85The
86.Fn MDXPad
87function can be used to pad message data in same way
88as done by
89.Fn MDXFinal
90without terminating calculation.
91.Pp
92The
93.Fn MDXEnd
94function is a wrapper for
95.Fn MDXFinal
96which converts the return value to a 33-character
97(including the terminating '\e0')
98ASCII string which represents the 128 bits in hexadecimal.
99.Pp
100The
101.Fn MDXFile
102function calculates the digest of a file, and uses
103.Fn MDXEnd
104to return the result.
105If the file cannot be opened, a null pointer is returned.
106The
107.Fn MDXFileChunk
108function is similar to
109.Fn MDXFile ,
110but it only calculates the digest over a byte-range of the file specified,
111starting at
112.Fa offset
113and spanning
114.Fa length
115bytes.
116If the
117.Fa length
118parameter is specified as 0, or more than the length of the remaining part
119of the file,
120.Fn MDXFileChunk
121calculates the digest from
122.Fa offset
123to the end of file.
124The
125.Fn MDXData
126function calculates the digest of a chunk of data in memory, and uses
127.Fn MDXEnd
128to return the result.
129.Pp
130When using
131.Fn MDXEnd ,
132.Fn MDXFile ,
133or
134.Fn MDXData ,
135the
136.Fa buf
137argument can be a null pointer, in which case the returned string
138is allocated with
139.Xr malloc 3
140and subsequently must be explicitly deallocated using
141.Xr free 3
142after use.
143If the
144.Fa buf
145argument is non-null it must point to at least 33 characters of buffer space.
146.Sh ERRORS
147The
148.Fn MDXEnd
149function called with a null buf argument may fail and return NULL if:
150.Bl -tag -width Er
151.It Bq Er ENOMEM
152Insufficient storage space is available.
153.El
154.Pp
155The
156.Fn MDXFile
157and
158.Fn MDXFileChunk
159may return NULL when underlying
160.Xr open 2 ,
161.Xr fstat 2 ,
162.Xr lseek 2 ,
163or
164.Xr MDXEnd 3
165fail.
166.Sh SEE ALSO
167.Xr md4 3 ,
168.Xr md5 3 ,
169.Xr ripemd 3 ,
170.Xr sha 3 ,
171.Xr sha256 3 ,
172.Xr sha512 3 ,
173.Xr skein 3
174.Rs
175.%A R. Rivest
176.%T The MD4 Message-Digest Algorithm
177.%O RFC 1186
178.Re
179.Rs
180.%A R. Rivest
181.%T The MD5 Message-Digest Algorithm
182.%O RFC 1321
183.Re
184.Rs
185.%A H. Dobbertin
186.%T Alf Swindles Ann
187.%J CryptoBytes
188.%N 1(3):5
189.%D 1995
190.Re
191.Rs
192.%A MJ. B. Robshaw
193.%T On Recent Results for MD2, MD4 and MD5
194.%J RSA Laboratories Bulletin
195.%N 4
196.%D November 12, 1996
197.Re
198.Sh HISTORY
199These functions appeared in
200.Fx 2.0 .
201.Sh AUTHORS
202The original MDX routines were developed by
203RSA Data Security, Inc., and published in the above references.
204This code is derived directly from these implementations by
205.An Poul-Henning Kamp Aq Mt phk@FreeBSD.org .
206.Pp
207Phk ristede runen.
208.Sh BUGS
209The MD5 algorithm has been proven to be vulnerable to practical collision
210attacks and should not be relied upon to produce unique outputs,
211.Em nor should they be used as part of a cryptographic signature scheme.
212