]>
Commit | Line | Data |
---|---|---|
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 | 7 | OpenSSL 0.9.6a/0.9.6b (bugfix releases, 5 Apr 2001 and 9 July 2001) |
e3fefbfd | 8 | and OpenSSL 0.9.7 were developed in parallel, based on OpenSSL 0.9.6. |
e9ad0d2c | 9 | |
a9d2bc49 | 10 | Change log entries are tagged as follows: |
daba492c BM |
11 | -) applies to 0.9.6a/0.9.6b/0.9.6c only |
12 | *) applies to 0.9.6a/0.9.6b/0.9.6c and 0.9.7 | |
a9d2bc49 BM |
13 | +) applies to 0.9.7 only |
14 | ||
c602e7f4 BM |
15 | *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() |
16 | instead. BIO_gethostbyname() does not know what timeouts are | |
17 | appropriate, so entries would stay in cache even when they hade | |
18 | become invalid. | |
19 | [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> | |
20 | ||
1fc6d41b DSH |
21 | +) New command line and configuration option 'utf8' for the req command. |
22 | This allows field values to be specified as UTF8 strings. | |
23 | [Steve Henson] | |
24 | ||
0e211563 BL |
25 | +) Add -multi and -mr options to "openssl speed" - giving multiple parallel |
26 | runs for the former and machine-readable output for the latter. | |
27 | [Ben Laurie] | |
28 | ||
89da653f BM |
29 | +) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion |
30 | of the e-mail address in the DN (i.e., it will go into a certificate | |
31 | extension only). The new configuration file option 'email_in_dn = no' | |
32 | has the same effect. | |
33 | [Massimiliano Pala madwolf@openca.org] | |
34 | ||
ba1c6022 BM |
35 | *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when |
36 | faced with a pathologically small ClientHello fragment that does | |
37 | not contain client_version: Instead of aborting with an error, | |
38 | simply choose the highest available protocol version (i.e., | |
39 | TLS 1.0 unless it is disabled). In practice, ClientHello | |
40 | messages are never sent like this, but this change gives us | |
41 | strictly correct behaviour at least for TLS. | |
42 | [Bodo Moeller] | |
43 | ||
c2e4f17c RL |
44 | +) Change all functions with names starting with des_ to be starting |
45 | with DES_ instead. This because there are increasing clashes with | |
46 | libdes and other des libraries that are currently used by other | |
47 | projects. The old libdes interface is provided, as well as crypt(), | |
48 | if openssl/des_old.h is included. Note that crypt() is no longer | |
49 | declared in openssl/des.h. | |
50 | ||
51 | NOTE: This is a major break of an old API into a new one. Software | |
52 | authors are encouraged to switch to the DES_ style functions. Some | |
53 | time in the future, des_old.h and the libdes compatibility functions | |
54 | will be completely removed. | |
55 | [Richard Levitte] | |
56 | ||
979689aa BM |
57 | *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() |
58 | never resets s->method to s->ctx->method when called from within | |
59 | one of the SSL handshake functions. | |
60 | [Bodo Moeller; problem pointed out by Niko Baric] | |
61 | ||
f1558bb4 DSH |
62 | +) Test for certificates which contain unsupported critical extensions. |
63 | If such a certificate is found during a verify operation it is | |
64 | rejected by default: this behaviour can be overridden by either | |
65 | handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or | |
66 | by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function | |
67 | X509_supported_extension() has also been added which returns 1 if a | |
68 | particular extension is supported. | |
69 | [Steve Henson] | |
70 | ||
a661b653 BM |
71 | +) New functions/macros |
72 | ||
73 | SSL_CTX_set_msg_callback(ctx, cb) | |
74 | SSL_CTX_set_msg_callback_arg(ctx, arg) | |
75 | SSL_set_msg_callback(ssl, cb) | |
76 | SSL_set_msg_callback_arg(ssl, arg) | |
77 | ||
78 | to request calling a callback function | |
79 | ||
80 | void cb(int write_p, int version, int content_type, | |
81 | const void *buf, size_t len, SSL *ssl, void *arg) | |
82 | ||
83 | whenever a protocol message has been completely received | |
84 | (write_p == 0) or sent (write_p == 1). Here 'version' is the | |
85 | protocol version according to which the SSL library interprets | |
86 | the current protocol message (SSL2_VERSION, SSL3_VERSION, or | |
87 | TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or | |
88 | the content type as defined in the SSL 3.0/TLS 1.0 protocol | |
89 | specification (change_cipher_spec(20), alert(21), handshake(22)). | |
90 | 'buf' and 'len' point to the actual message, 'ssl' to the | |
91 | SSL object, and 'arg' is the application-defined value set by | |
92 | SSL[_CTX]_set_msg_callback_arg(). | |
93 | ||
94 | 'openssl s_client' and 'openssl s_server' have new '-msg' options | |
95 | to enable a callback that displays all protocol messages. | |
96 | ||
97 | TODO: SSL 2.0, doc/ssl/, doc/apps/ | |
98 | [Bodo Moeller] | |
99 | ||
100 | *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert | |
101 | (sent using the client's version number) if client_version is | |
102 | smaller than the protocol version in use. Also change | |
103 | ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if | |
104 | the client demanded SSL 3.0 but only TLS 1.0 is enabled; then | |
105 | the client will at least see that alert. | |
106 | [Bodo Moeller] | |
107 | ||
581f1c84 DSH |
108 | +) Modify the behaviour of EVP cipher functions in similar way to digests |
109 | to retain compatibility with existing code. | |
110 | [Steve Henson] | |
111 | ||
20d2186c | 112 | +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain |
50d194af DSH |
113 | compatibility with existing code. In particular the 'ctx' parameter does |
114 | not have to be to be initialized before the call to EVP_DigestInit() and | |
115 | it is tidied up after a call to EVP_DigestFinal(). New function | |
116 | EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function | |
117 | EVP_MD_CTX_copy() changed to not require the destination to be | |
118 | initialized valid and new function EVP_MD_CTX_copy_ex() added which | |
119 | requires the destination to be valid. | |
120 | ||
121 | Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), | |
122 | EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). | |
20d2186c DSH |
123 | [Steve Henson] |
124 | ||
48948d53 BM |
125 | +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it |
126 | so that complete 'Handshake' protocol structures are kept in memory | |
127 | instead of overwriting 'msg_type' and 'length' with 'body' data. | |
128 | [Bodo Moeller] | |
129 | ||
130 | *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation | |
131 | correctly. | |
132 | [Bodo Moeller] | |
133 | ||
285046ec RL |
134 | +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. |
135 | [Massimo Santin via Richard Levitte] | |
136 | ||
07cee702 GT |
137 | +) Major restructuring to the underlying ENGINE code. This includes |
138 | reduction of linker bloat, separation of pure "ENGINE" manipulation | |
139 | (initialisation, etc) from functionality dealing with implementations | |
140 | of specific crypto iterfaces. This change also introduces integrated | |
141 | support for symmetric ciphers and digest implementations - so ENGINEs | |
142 | can now accelerate these by providing EVP_CIPHER and EVP_MD | |
143 | implementations of their own. This is detailed in crypto/engine/README | |
144 | as it couldn't be adequately described here. However, there are a few | |
145 | API changes worth noting - some RSA, DSA, DH, and RAND functions that | |
146 | were changed in the original introduction of ENGINE code have now | |
147 | reverted back - the hooking from this code to ENGINE is now a good | |
148 | deal more passive and at run-time, operations deal directly with | |
149 | RSA_METHODs, DSA_METHODs (etc) as they did before, rather than | |
150 | dereferencing through an ENGINE pointer any more. Also, the ENGINE | |
151 | functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - | |
152 | they were not being used by the framework as there is no concept of a | |
153 | BIGNUM_METHOD and they could not be generalised to the new | |
154 | 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, | |
155 | ENGINE_cpy() has been removed as it cannot be consistently defined in | |
156 | the new code. | |
157 | [Geoff Thorpe] | |
158 | ||
d46c1a81 DSH |
159 | +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. |
160 | [Steve Henson] | |
161 | ||
89eeccac RL |
162 | +) Change mkdef.pl to sort symbols that get the same entry number, |
163 | and make sure the automatically generated functions ERR_load_* | |
164 | become part of libeay.num as well. | |
165 | [Richard Levitte] | |
166 | ||
3b0b5aba BM |
167 | *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a |
168 | client receives HelloRequest while in a handshake. | |
169 | [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] | |
170 | ||
6b0e9fac BM |
171 | +) New function SSL_renegotiate_pending(). This returns true once |
172 | renegotiation has been requested (either SSL_renegotiate() call | |
173 | or HelloRequest/ClientHello receveived from the peer) and becomes | |
174 | false once a handshake has been completed. | |
175 | (For servers, SSL_renegotiate() followed by SSL_do_handshake() | |
176 | sends a HelloRequest, but does not ensure that a handshake takes | |
177 | place. SSL_renegotiate_pending() is useful for checking if the | |
178 | client has followed the request.) | |
179 | [Bodo Moeller] | |
180 | ||
181 | +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. | |
182 | By default, clients may request session resumption even during | |
183 | renegotiation (if session ID contexts permit); with this option, | |
184 | session resumption is possible only in the first handshake. | |
185 | [Bodo Moeller] | |
186 | ||
2260ad21 BM |
187 | *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C |
188 | should end in 'break', not 'goto end' which circuments various | |
b49124f6 BM |
189 | cleanups done in state SSL_ST_OK. But session related stuff |
190 | must be disabled for SSL_ST_OK in the case that we just sent a | |
191 | HelloRequest. | |
2260ad21 BM |
192 | |
193 | Also avoid some overhead by not calling ssl_init_wbio_buffer() | |
194 | before just sending a HelloRequest. | |
b49124f6 | 195 | [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] |
8e2f6b79 | 196 | |
ee60d9fb BM |
197 | *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't |
198 | reveal whether illegal block cipher padding was found or a MAC | |
199 | verification error occured. (Neither SSLerr() codes nor alerts | |
200 | are directly visible to potential attackers, but the information | |
201 | may leak via logfiles.) | |
202 | ||
203 | Similar changes are not required for the SSL 2.0 implementation | |
204 | because the number of padding bytes is sent in clear for SSL 2.0, | |
205 | and the extra bytes are just ignored. However ssl/s2_pkt.c | |
206 | failed to verify that the purported number of padding bytes is in | |
207 | the legal range. | |
208 | [Bodo Moeller] | |
209 | ||
96bd6f73 DSH |
210 | +) Add some demos for certificate and certificate request creation. |
211 | [Steve Henson] | |
212 | ||
c0f5dd07 LJ |
213 | +) Make maximum certificate chain size accepted from the peer application |
214 | settable (SSL*_get/set_max_cert_list()), as proposed by | |
215 | "Douglas E. Engert" <deengert@anl.gov>. | |
216 | [Lutz Jaenicke] | |
217 | ||
6c36f7a9 LJ |
218 | +) Add support for shared libraries for Unixware-7 and support including |
219 | shared libraries for OpenUNIX-8 (Boyd Lynn Gerber <gerberb@zenez.com>). | |
220 | [Lutz Jaenicke] | |
221 | ||
a9ed4da8 BM |
222 | *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid |
223 | 'wristwatch attack' using huge encoding parameters (cf. | |
224 | James H. Manger's CRYPTO 2001 paper). Note that the | |
225 | RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use | |
e3fefbfd | 226 | encoding parameters and hence was not vulnerable. |
a9ed4da8 BM |
227 | [Bodo Moeller] |
228 | ||
4450107a | 229 | +) Add a "destroy" handler to ENGINEs that allows structural cleanup to |
908efd3b GT |
230 | be done prior to destruction. Use this to unload error strings from |
231 | ENGINEs that load their own error strings. NB: This adds two new API | |
232 | functions to "get" and "set" this destroy handler in an ENGINE. | |
a9ed4da8 | 233 | [Geoff Thorpe] |
908efd3b | 234 | |
4450107a | 235 | +) Alter all existing ENGINE implementations (except "openssl" and |
541814c4 GT |
236 | "openbsd") to dynamically instantiate their own error strings. This |
237 | makes them more flexible to be built both as statically-linked ENGINEs | |
238 | and self-contained shared-libraries loadable via the "dynamic" ENGINE. | |
239 | Also, add stub code to each that makes building them as self-contained | |
240 | shared-libraries easier (see README.ENGINE). | |
241 | [Geoff Thorpe] | |
242 | ||
4450107a | 243 | +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE |
541814c4 GT |
244 | implementations into applications that are completely implemented in |
245 | self-contained shared-libraries. The "dynamic" ENGINE exposes control | |
246 | commands that can be used to configure what shared-library to load and | |
247 | to control aspects of the way it is handled. Also, made an update to | |
248 | the README.ENGINE file that brings its information up-to-date and | |
249 | provides some information and instructions on the "dynamic" ENGINE | |
250 | (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). | |
251 | [Geoff Thorpe] | |
252 | ||
d98a4b73 UM |
253 | *) BN_sqr() bug fix. |
254 |