1=pod
2
3=head1 NAME
4
5rand - pseudo-random number generator
6
7=head1 SYNOPSIS
8
9 #include <openssl/rand.h>
10
11 int  RAND_set_rand_engine(ENGINE *engine);
12
13 int  RAND_bytes(unsigned char *buf, int num);
14 int  RAND_pseudo_bytes(unsigned char *buf, int num);
15
16 void RAND_seed(const void *buf, int num);
17 void RAND_add(const void *buf, int num, int entropy);
18 int  RAND_status(void);
19
20 int  RAND_load_file(const char *file, long max_bytes);
21 int  RAND_write_file(const char *file);
22 const char *RAND_file_name(char *file, size_t num);
23
24 int  RAND_egd(const char *path);
25
26 void RAND_set_rand_method(const RAND_METHOD *meth);
27 const RAND_METHOD *RAND_get_rand_method(void);
28 RAND_METHOD *RAND_SSLeay(void);
29
30 void RAND_cleanup(void);
31
32 /* For Win32 only */
33 void RAND_screen(void);
34 int RAND_event(UINT, WPARAM, LPARAM);
35
36=head1 DESCRIPTION
37
38Since the introduction of the ENGINE API, the recommended way of controlling
39default implementations is by using the ENGINE API functions. The default
40B<RAND_METHOD>, as set by RAND_set_rand_method() and returned by
41RAND_get_rand_method(), is only used if no ENGINE has been set as the default
42"rand" implementation. Hence, these two functions are no longer the recommened
43way to control defaults.
44
45If an alternative B<RAND_METHOD> implementation is being used (either set
46directly or as provided by an ENGINE module), then it is entirely responsible
47for the generation and management of a cryptographically secure PRNG stream. The
48mechanisms described below relate solely to the software PRNG implementation
49built in to OpenSSL and used by default.
50
51These functions implement a cryptographically secure pseudo-random
52number generator (PRNG). It is used by other library functions for
53example to generate random keys, and applications can use it when they
54need randomness.
55
56A cryptographic PRNG must be seeded with unpredictable data such as
57mouse movements or keys pressed at random by the user. This is
58described in L<RAND_add(3)|RAND_add(3)>. Its state can be saved in a seed file
59(see L<RAND_load_file(3)|RAND_load_file(3)>) to avoid having to go through the
60seeding process whenever the application is started.
61
62L<RAND_bytes(3)|RAND_bytes(3)> describes how to obtain random data from the
63PRNG.
64
65=head1 INTERNALS
66
67The RAND_SSLeay() method implements a PRNG based on a cryptographic
68hash function.
69
70The following description of its design is based on the SSLeay
71documentation:
72
73First up I will state the things I believe I need for a good RNG.
74
75=over 4
76
77=item 1
78
79A good hashing algorithm to mix things up and to convert the RNG 'state'
80to random numbers.
81
82=item 2
83
84An initial source of random 'state'.
85
86=item 3
87
88The state should be very large.  If the RNG is being used to generate
894096 bit RSA keys, 2 2048 bit random strings are required (at a minimum).
90If your RNG state only has 128 bits, you are obviously limiting the
91search space to 128 bits, not 2048.  I'm probably getting a little
92carried away on this last point but it does indicate that it may not be
93a bad idea to keep quite a lot of RNG state.  It should be easier to
94break a cipher than guess the RNG seed data.
95
96=item 4
97
98Any RNG seed data should influence all subsequent random numbers
99generated.  This implies that any random seed data entered will have
100an influence on all subsequent random numbers generated.
101
102=item 5
103
104When using data to seed the RNG state, the data used should not be
105extractable from the RNG state.  I believe this should be a
106requirement because one possible source of 'secret' semi random
107data would be a private key or a password.  This data must
108not be disclosed by either subsequent random numbers or a
109'core' dump left by a program crash.
110
111=item 6
112
113Given the same initial 'state', 2 systems should deviate in their RNG state
114(and hence the random numbers generated) over time if at all possible.
115
116=item 7
117
118Given the random number output stream, it should not be possible to determine
119the RNG state or the next random number.
120
121=back
122
123The algorithm is as follows.
124
125There is global state made up of a 1023 byte buffer (the 'state'), a
126working hash value ('md'), and a counter ('count').
127
128Whenever seed data is added, it is inserted into the 'state' as
129follows.
130
131The input is chopped up into units of 20 bytes (or less for
132the last block).  Each of these blocks is run through the hash
133function as follows:  The data passed to the hash function
134is the current 'md', the same number of bytes from the 'state'
135(the location determined by in incremented looping index) as
136the current 'block', the new key data 'block', and 'count'
137(which is incremented after each use).
138The result of this is kept in 'md' and also xored into the
139'state' at the same locations that were used as input into the
140hash function. I
141believe this system addresses points 1 (hash function; currently
142SHA-1), 3 (the 'state'), 4 (via the 'md'), 5 (by the use of a hash
143function and xor).
144
145When bytes are extracted from the RNG, the following process is used.
146For each group of 10 bytes (or less), we do the following:
147
148Input into the hash function the local 'md' (which is initialized from
149the global 'md' before any bytes are generated), the bytes that are to
150be overwritten by the random bytes, and bytes from the 'state'
151(incrementing looping index). From this digest output (which is kept
152in 'md'), the top (up to) 10 bytes are returned to the caller and the
153bottom 10 bytes are xored into the 'state'.
154
155Finally, after we have finished 'num' random bytes for the caller,
156'count' (which is incremented) and the local and global 'md' are fed
157into the hash function and the results are kept in the global 'md'.
158
159I believe the above addressed points 1 (use of SHA-1), 6 (by hashing
160into the 'state' the 'old' data from the caller that is about to be
161overwritten) and 7 (by not using the 10 bytes given to the caller to
162update the 'state', but they are used to update 'md').
163
164So of the points raised, only 2 is not addressed (but see
165L<RAND_add(3)|RAND_add(3)>).
166
167=head1 SEE ALSO
168
169L<BN_rand(3)|BN_rand(3)>, L<RAND_add(3)|RAND_add(3)>,
170L<RAND_load_file(3)|RAND_load_file(3)>, L<RAND_egd(3)|RAND_egd(3)>,
171L<RAND_bytes(3)|RAND_bytes(3)>,
172L<RAND_set_rand_method(3)|RAND_set_rand_method(3)>,
173L<RAND_cleanup(3)|RAND_cleanup(3)>
174
175=cut
176