1 /* gc.h --- Header file for implementation agnostic crypto wrapper API.
2 * Copyright (C) 2002, 2003, 2004, 2005 Simon Josefsson
4 * This file is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License as published
6 * by the Free Software Foundation; either version 2, or (at your
7 * option) any later version.
9 * This file is distributed in the hope that it will be useful, but
10 * WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * General Public License for more details.
14 * You should have received a copy of the GNU General Public License
15 * along with this file; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
35 GC_PKCS5_INVALID_ITERATION_COUNT,
36 GC_PKCS5_INVALID_DERIVED_KEY_LENGTH,
37 GC_PKCS5_DERIVED_KEY_TOO_LONG
39 typedef enum Gc_rc Gc_rc;
46 typedef enum Gc_hash Gc_hash;
48 #define GC_MD5_DIGEST_SIZE 16
50 /* Call before respectively after any other functions. */
51 extern int gc_init (void);
52 extern void gc_done (void);
54 /* Memory allocation (avoid). */
55 typedef void *(*gc_malloc_t) (size_t n);
56 typedef int (*gc_secure_check_t) (const void *);
57 typedef void *(*gc_realloc_t) (void *p, size_t n);
58 typedef void (*gc_free_t) (void *);
59 extern void gc_set_allocators (gc_malloc_t func_malloc,
60 gc_malloc_t secure_malloc,
61 gc_secure_check_t secure_check,
62 gc_realloc_t func_realloc,
67 gc_hash_buffer (int hash, const void *in, size_t inlen, char *out);
69 /* One-call interface. */
70 extern int gc_md5 (const void *in, size_t inlen, void *resbuf);
71 extern int gc_hmac_md5 (const void *key, size_t keylen,
72 const void *in, size_t inlen,
78 From: Simon Josefsson <jas@extundo.com>
79 Subject: Re: generic crypto
80 Newsgroups: gmane.comp.lib.gnulib.bugs
81 Cc: bug-gnulib@gnu.org
82 Date: Fri, 07 Oct 2005 12:50:57 +0200
83 Mail-Copies-To: nobody
85 Paul Eggert <eggert@CS.UCLA.EDU> writes:
87 > Simon Josefsson <jas@extundo.com> writes:
89 >> * Perhaps the /dev/*random reading should be separated into a separate
90 >> module? It might be useful outside of the gc layer too.
92 > Absolutely. I've been meaning to do that for months (for a "shuffle"
93 > program I want to add to coreutils), but hadn't gotten around to it.
94 > It would have to be generalized a bit. I'd like to have the file
95 > descriptor cached, for example.
97 I'll write a separate module for that part.
99 I think we should even add a good PRNG that is re-seeded from
100 /dev/*random frequently. GnuTLS can need a lot of random data on a
101 big server, more than /dev/random can supply. And /dev/urandom might
102 not be strong enough. Further, the security of /dev/*random can also
105 >> I'm also not sure about the names of those functions, they suggest
106 >> a more higher-level API than what is really offered (i.e., the
107 >> names "nonce" and "pseudo_random" and "random" imply certain
108 >> cryptographic properties).
110 > Could you expand a bit more on that? What is the relationship between
111 > nonce/pseudorandom/random and the /dev/ values you are using?
113 There is none, that is the problem.
115 Applications generally need different kind of "random" numbers.
116 Sometimes they just need some random data and doesn't care whether it
117 is possible for an attacker to compute the string (aka a "nonce").
118 Sometimes they need data that is very difficult to compute (i.e.,
119 computing it require inverting SHA1 or similar). Sometimes they need
120 data that is not possible to compute, i.e., it wants real entropy
121 collected over time on the system. Collecting the last kind of random
122 data is very expensive, so it must not be used too often. The second
123 kind of random data ("pseudo random") is typically generated by
124 seeding a good PRNG with a couple of hundred bytes of real entropy
125 from the "real random" data pool. The "nonce" is usually computed
126 using the PRNG as well, because PRNGs are usually fast.
128 Pseudo-random data is typically used for session keys. Strong random
129 data is often used to generate long-term keys (e.g., private RSA
132 Of course, there are many subtleties. There are several different
133 kind of nonce:s. Sometimes a nonce is just an ever-increasing
134 integer, starting from 0. Sometimes it is assumed to be unlikely to
135 be the same as previous nonces, but without a requirement that the
136 nonce is possible to guess. MD5(system clock) would thus suffice, if
137 it isn't called too often. You can guess what the next value will be,
138 but it will always be different.
140 The problem is that /dev/*random doesn't offer any kind of semantic
141 guarantees. But applications need an API that make that promise.
143 I think we should do this in several steps:
145 1) Write a module that can read from /dev/*random.
147 2) Add a module for a known-good PRNG suitable for random number
148 generation, that can be continuously re-seeded.
150 3) Add a high-level module that provide various different randomness
151 functions. One for nonces, perhaps even different kind of nonces,
152 one for pseudo random data, and one for strong random data. It is
153 not clear whether we can hope to achieve the last one in a portable
156 Further, it would be useful to allow users to provide their own
157 entropy source as a file, used to seed the PRNG or initialize the
158 strong randomness pool. This is used on embedded platforms that
159 doesn't have enough interrupts to hope to generate good random data.
161 > For example, why not use OpenBSD's /dev/arandom?
163 I don't trust ARC4. For example, recent cryptographic efforts
164 indicate that you must throw away the first 512 bytes generated from
165 the PRNG for it to be secure. I don't know whether OpenBSD do this.
166 Further, I recall some eprint paper on RC4 security that didn't
169 While I trust the random devices in OpenBSD more than
170 Solaris/AIX/HPUX/etc, I think that since we need something better on
171 Solaris/AIX/HPUX we'd might as well use it on OpenBSD or even Linux
174 > Here is one thought. The user could specify a desired quality level
175 > range, and the implementation then would supply random data that is at
176 > least as good as the lower bound of the range. I.e., ihe
177 > implementation refuses to produce any random data if it can't generate
178 > data that is at least as good as the lower end of the range. The
179 > upper bound of the range is advice from the user not to be any more
180 > expensive than that, but the implementation can ignore the advice if
181 > it doesn't have anything cheaper.
183 I'm not sure this is a good idea. Users can't really be expected to
184 understand this. Further, applications need many different kind of
185 random data. Selecting the randomness level for each by the user will
188 I think it is better if the application decide, from its cryptographic
189 requirement, what entropy quality it require, and call the proper API.
190 Meeting the implied semantic properties should be the job for gnulib.
192 >> Perhaps gc_dev_random and gc_dev_urandom?
194 > To some extent. I'd rather insulate the user from the details of
195 > where the random numbers come from. On the other hand we need to
196 > provide a way for applications to specify a file that contains
197 > random bits, so that people can override the defaults.
201 This may require some thinking before it is finalized. Is it ok to
202 install the GC module as-is meanwhile? Then I can continue to add the
203 stuff that GnuTLS need, and then come back to re-working the
204 randomness module. That way, we have two different projects that use
205 the code. GnuTLS includes the same randomness code that was in GNU
206 SASL and that is in the current gc module. I feel much more
207 comfortable working in small steps at a time, rather then working on
208 this for a long time in gnulib and only later integrate the stuff in