]> git.ipfire.org Git - thirdparty/openssl.git/blame - CHANGES
Code to avoid the use of non-standard strptime(). By
[thirdparty/openssl.git] / CHANGES
CommitLineData
81a6c781 1
f1c236f8 2 OpenSSL CHANGES
651d0aff
RE
3 _______________
4
a9d2bc49 5 Changes between 0.9.6 and 0.9.7 [xx XXX 2001]
a43cf9fa 6
e9ad0d2c
BM
7 OpenSSL 0.9.6a/0.9.6b (bugfix releases, 5 Apr 2001 and 9 July 2001)
8 and OpenSSL 0.9.7 were developped in parallel, based on OpenSSL 0.9.6.
9
a9d2bc49 10 Change log entries are tagged as follows:
e9ad0d2c
BM
11 -) applies to 0.9.6a/0.9.6b only
12 *) applies to 0.9.6a/0.9.6b and 0.9.7
a9d2bc49
BM
13 +) applies to 0.9.7 only
14
e9ad0d2c
BM
15 -) OpenSSL 0.9.6b released [9 July 2001]
16
17 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
18 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
19 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
20 PRNG state recovery was possible based on the output of
21 one PRNG request appropriately sized to gain knowledge on
22 'md' followed by enough consecutive 1-byte PRNG requests
23 to traverse all of 'state'.
24
25 1. When updating 'md_local' (the current thread's copy of 'md')
26 during PRNG output generation, hash all of the previous
27 'md_local' value, not just the half used for PRNG output.
28
29 2. Make the number of bytes from 'state' included into the hash
30 independent from the number of PRNG bytes requested.
31
32 The first measure alone would be sufficient to avoid
33 Markku-Juhani's attack. (Actually it had never occurred
34 to me that the half of 'md_local' used for chaining was the
35 half from which PRNG output bytes were taken -- I had always
36 assumed that the secret half would be used.) The second
37 measure makes sure that additional data from 'state' is never
38 mixed into 'md_local' in small portions; this heuristically
39 further strengthens the PRNG.
40 [Bodo Moeller]
41
f31b1250
BL
42 +) Speed up EVP routines.
43 Before:
44encrypt
45type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
46des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
47des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
48des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
49decrypt
50des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
51des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
52des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
53 After:
54encrypt
c148d709 55des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
f31b1250 56decrypt
c148d709 57des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
f31b1250
BL
58 [Ben Laurie]
59
93dbd835
BM
60 *) Fix crypto/bn/asm/mips3.s.
61 [Andy Polyakov]
62
43f9391b
LJ
63 *) When only the key is given to "enc", the IV is undefined. Print out
64 an error message in this case.
65 [Lutz Jaenicke]
66
c80410c5
RL
67 +) Added the OS2-EMX target.
68 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
69
b7a26e6d
DSH
70 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
71 to support NCONF routines in extension code. New function CONF_set_nconf()
72 to allow functions which take an NCONF to also handle the old LHASH
73 structure: this means that the old CONF compatible routines can be
74 retained (in particular wrt extensions) without having to duplicate the
75 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
76 [Steve Henson]
77
1e325f61
DSH
78 *) Handle special case when X509_NAME is empty in X509 printing routines.
79 [Steve Henson]
80
c458a331
BM
81 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
82 positive and less than q.
83 [Bodo Moeller]
84
fd3e027f 85 +) Enhance the general user interface with mechanisms for inner control
235dd0a2
RL
86 and with pssibilities to have yes/no kind of prompts.
87 [Richard Levitte]
88
d63c6bd3 89 +) Change all calls to low level digest routines in the library and
323f289c
DSH
90 applications to use EVP. Add missing calls to HMAC_cleanup() and
91 don't assume HMAC_CTX can be copied using memcpy().
92 [Verdon Walker <VWalker@novell.com>, Steve Henson]
93
839590f5
RL
94 +) Add the possibility to control engines through control names but with
95 arbitrary arguments instead of just a string.
96 Change the key loaders to take a UI_METHOD instead of a callback
97 function pointer. NOTE: this breaks binary compatibility with earlier
98 versions of OpenSSL [engine].
99 Addapt the nCipher code for these new conditions and add a card insertion
100 callback.
101 [Richard Levitte]
102
9ad0f681
RL
103 +) Enhance the general user interface with mechanisms to better support
104 dialog box interfaces, application-defined prompts, the possibility
105 to use defaults (for example default passwords from somewhere else)
106 and interrupts/cancelations.
107 [Richard Levitte]
108
3cc1f498
DSH
109 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
110 used: it isn't thread safe and the add_lock_callback should handle
111 that itself.
112 [Paul Rose <Paul.Rose@bridge.com>]
113
285b4275
BM
114 *) Verify that incoming data obeys the block size in
115 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
116 [Bodo Moeller]
117
f2a253e0
DSH
118 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
119 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
120 [Steve Henson]
121
ecf18606
BM
122 *) Fix OAEP check.
123