]>
Commit | Line | Data |
---|---|---|
1 | OpenSSL CHANGES | |
2 | =============== | |
3 | ||
4 | This is a high-level summary of the most important changes. | |
5 | For a full list of changes, see the [git commit log][log] and | |
6 | pick the appropriate release branch. | |
7 | ||
8 | [log]: https://github.com/openssl/openssl/commits/ | |
9 | ||
10 | OpenSSL Releases | |
11 | ---------------- | |
12 | ||
13 | - [OpenSSL 3.0](#openssl-30) | |
14 | - [OpenSSL 1.1.1](#openssl-111) | |
15 | - [OpenSSL 1.1.0](#openssl-110) | |
16 | - [OpenSSL 1.0.2](#openssl-102) | |
17 | - [OpenSSL 1.0.1](#openssl-101) | |
18 | - [OpenSSL 1.0.0](#openssl-100) | |
19 | - [OpenSSL 0.9.x](#openssl-09x) | |
20 | ||
21 | OpenSSL 3.0 | |
22 | ----------- | |
23 | ||
24 | ### Changes between 1.1.1 and 3.0 [xx XXX xxxx] | |
25 | ||
26 | * A public key check is now performed during EVP_PKEY_derive_set_peer(). | |
27 | Previously DH was internally doing this during EVP_PKEY_derive(). | |
28 | To disable this check use EVP_PKEY_derive_set_peer_ex(dh, peer, 0). This | |
29 | may mean that an error can occur in EVP_PKEY_derive_set_peer() rather than | |
30 | during EVP_PKEY_derive(). | |
31 | ||
32 | *Shane Lontis* | |
33 | ||
34 | * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for | |
35 | more key types including RSA, DSA, ED25519, X25519, ED448 and X448. | |
36 | Previously (in 1.1.1) they would return -2. For key types that do not have | |
37 | parameters then EVP_PKEY_param_check() will always return 1. | |
38 | ||
39 | * The output from numerous "printing" functions such as X509_signature_print(), | |
40 | X509_print_ex(), X509_CRL_print_ex(), and other similar functions has been | |
41 | amended such that there may be cosmetic differences between the output | |
42 | observed in 1.1.1 and 3.0. This also applies to the "-text" output from the | |
43 | x509 and crl applications. | |
44 | ||
45 | *David von Oheimb* | |
46 | ||
47 | * Windows thread synchronization uses read/write primitives (SRWLock) when | |
48 | supported by the OS, otherwise CriticalSection continues to be used. | |
49 | ||
50 | *Vincent Drake* | |
51 | ||
52 | * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to | |
53 | work on read only BIO source/sinks that do not support these functions. | |
54 | This allows piping or redirection of a file BIO using stdin to be buffered | |
55 | into memory. This is used internally in OSSL_DECODER_from_bio(). | |
56 | ||
57 | *Shane Lontis* | |
58 | ||
59 | * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1 | |
60 | this function would return one of the values OSSL_STORE_INFO_NAME, | |
61 | OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or | |
62 | OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported | |
63 | as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now | |
64 | reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications | |
65 | using this function should be amended to handle the changed return value. | |
66 | ||
67 | *Richard Levitte* | |
68 | ||
69 | * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035) | |
70 | for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations. | |
71 | As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present. | |
72 | Correct the semantics of checking the validation chain in case ESSCertID{,v2} | |
73 | contains more than one certificate identifier: This means that all | |
74 | certificates referenced there MUST be part of the validation chain. | |
75 | ||
76 | *David von Oheimb* | |
77 | ||
78 | * The implementation of the EVP ciphers CAST5-ECB, CAST5-CBC, CAST5-OFB, | |
79 | CAST5-CFB, BF-ECB, BF-CBC, BF-OFB, BF-CFB, IDEA-ECB, IDEC-CBC, IDEA-OFB, | |
80 | IDEA-CFB, SEED-ECB, SEED-CBC, SEED-OFB, SEED-CFB, RC2-ECB, RC2-CBC, | |
81 | RC2-40-CBC, RC2-64-CBC, RC2-OFB, RC2-CFB, RC4, RC4-40, RC4-HMAC-MD5, RC5-ECB, | |
82 | RC5-CBC, RC5-OFB, RC5-CFB, DESX-CBC, DES-ECB, DES-CBC, DES-OFB, DES-CFB, | |
83 | DES-CFB1 and DES-CFB8 have been moved to the legacy provider. Applications | |
84 | using the EVP APIs to access these ciphers should instead use more modern | |
85 | ciphers. If that is not possible then these applications should ensure that | |
86 | the legacy provider has been loaded. This can be achieved either | |
87 | programmatically or via configuration. See the provider(7) man page for | |
88 | further details. | |
89 | ||
90 | *Matt Caswell* | |
91 | ||
92 | * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and | |
93 | RIPEMD-160 have been moved to the legacy provider. Applications using the | |
94 | EVP APIs to access these digests should instead use more modern digests. If | |
95 | that is not possible then these applications should ensure that the legacy | |
96 | provider has been loaded. This can be achieved either programmatically or via | |
97 | configuration. See the provider(7) man page for further details. | |
98 | ||
99 | *Matt Caswell* | |
100 | ||
101 | * The deprecated function EVP_PKEY_get0() now returns NULL being called for a | |
102 | provided key. | |
103 | ||
104 | *Dmitry Belyavskiy* | |
105 | ||
106 | * The deprecated functions EVP_PKEY_get0_RSA(), | |
107 | EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(), | |
108 | EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as | |
109 | well as the similarly named "get1" functions behave slightly differently in | |
110 | OpenSSL 3.0. Previously they returned a pointer to the low-level key used | |
111 | internally by libcrypto. From OpenSSL 3.0 this key may now be held in a | |
112 | provider. Calling these functions will only return a handle on the internal | |
113 | key where the EVP_PKEY was constructed using this key in the first place, for | |
114 | example using a function or macro such as EVP_PKEY_assign_RSA(), | |
115 | EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key, | |
116 | then these functions now return a cached copy of the key. Changes to | |
117 | the internal provider key that take place after the first time the cached key | |
118 | is accessed will not be reflected back in the cached copy. Similarly any | |
119 | changes made to the cached copy by application code will not be reflected | |
120 | back in the internal provider key. | |
121 | ||
122 | For the above reasons the keys returned from these functions should typically | |
123 | be treated as read-only. To emphasise this the value returned from | |
124 | EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), | |
125 | EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may | |
126 | break some existing code. Applications broken by this change should be | |
127 | modified. The preferred solution is to refactor the code to avoid the use of | |
128 | these deprecated functions. Failing this the code should be modified to use a | |
129 | const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(), | |
130 | EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a | |
131 | non-const pointer to enable them to be "freed". However they should also be | |
132 | treated as read-only. | |
133 | ||
134 | *Matt Caswell* | |
135 | ||
136 | * A number of functions handling low level keys or engines were deprecated | |
137 | including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(), | |
138 | EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and | |
139 | EVP_PKEY_get0_siphash(). Applications using engines should instead use | |
140 | providers. Applications getting or setting low level keys in an EVP_PKEY | |
141 | should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively | |
142 | use EVP_PKEY_fromdata() or EVP_PKEY_get_params(). | |
143 | ||
144 | *Matt Caswell* | |
145 | ||
146 | * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and | |
147 | EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed | |
148 | and require returning octet ptr parameters from providers that | |
149 | would like to support them which complicates provider implementations. | |
150 | ||
151 | *Tomáš Mráz* | |
152 | ||
153 | * The RAND_METHOD APIs have been deprecated. The functions deprecated are: | |
154 | RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and | |
155 | RAND_set_rand_method(). Provider based random number generators should | |
156 | be used instead via EVP_RAND(3). | |
157 | ||
158 | *Paul Dale* | |
159 | ||
160 | * The SRP APIs have been deprecated. The old APIs do not work via providers, | |
161 | and there is no EVP interface to them. Unfortunately there is no replacement | |
162 | for these APIs at this time. | |
163 | ||
164 | *Matt Caswell* | |
165 | ||
166 | * Add a compile time option to prevent the caching of provider fetched | |
167 | algorithms. This is enabled by including the no-cached-fetch option | |
168 | at configuration time. | |
169 | ||
170 | *Paul Dale* | |
171 | ||
172 | * The default algorithms for pkcs12 creation with the PKCS12_create() function | |
173 | were changed to more modern PBKDF2 and AES based algorithms. The default | |
174 | MAC iteration count was changed to PKCS12_DEFAULT_ITER to make it equal | |
175 | with the password-based encryption iteration count. The default digest | |
176 | algorithm for the MAC computation was changed to SHA-256. The pkcs12 | |
177 | application now supports -legacy option that restores the previous | |
178 | default algorithms to support interoperability with legacy systems. | |
179 | ||
180 | *Tomáš Mráz and Sahana Prasad* | |
181 | ||
182 | * The openssl speed command does not use low-level API calls anymore. This | |
183 | implies some of the performance numbers might not be fully comparable | |
184 | with the previous releases due to higher overhead. This applies | |
185 | particularly to measuring performance on smaller data chunks. | |
186 | ||
187 | *Tomáš Mráz* | |
188 | ||
189 | * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA | |
190 | capable processors. | |
191 | ||
192 | *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)* | |
193 | ||
194 | * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3. | |
195 | Typically if OpenSSL has no EC or DH algorithms then it cannot support | |
196 | connections with TLSv1.3. However OpenSSL now supports "pluggable" groups | |
197 | through providers. Therefore third party providers may supply group | |
198 | implementations even where there are no built-in ones. Attempting to create | |
199 | TLS connections in such a build without also disabling TLSv1.3 at run time or | |
200 | using third party provider groups may result in handshake failures. TLSv1.3 | |
201 | can be disabled at compile time using the "no-tls1_3" Configure option. | |
202 | ||
203 | *Matt Caswell* | |
204 | ||
205 | * The undocumented function X509_certificate_type() has been deprecated; | |
206 | applications can use X509_get0_pubkey() and X509_get0_signature() to | |
207 | get the same information. | |
208 | ||
209 | *Rich Salz* | |
210 | ||
211 | * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range() | |
212 | functions. They are identical to BN_rand() and BN_rand_range() | |
213 | respectively. | |
214 | ||
215 | *Tomáš Mráz* | |
216 | ||
217 | * Removed RSA padding mode for SSLv23 (which was only used for | |
218 | SSLv2). This includes the functions RSA_padding_check_SSLv23() and | |
219 | RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated | |
220 | `rsautl` command. | |
221 | ||
222 | *Rich Salz* | |
223 | ||
224 | * Deprecated the obsolete X9.31 RSA key generation related functions | |
225 | BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and | |
226 | BN_X931_generate_prime_ex(). | |
227 | ||
228 | *Tomáš Mráz* | |
229 | ||
230 | * The default key generation method for the regular 2-prime RSA keys was | |
231 | changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with | |
232 | Conditions Based on Auxiliary Probable Primes). This method is slower | |
233 | than the original method. | |
234 | ||
235 | *Shane Lontis* | |
236 | ||
237 | * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions. | |
238 | They are replaced with the BN_check_prime() function that avoids possible | |
239 | misuse and always uses at least 64 rounds of the Miller-Rabin | |
240 | primality test. At least 64 rounds of the Miller-Rabin test are now also | |
241 | used for all prime generation, including RSA key generation. | |
242 | This increases key generation time, especially for larger keys. | |
243 | ||
244 | *Kurt Roeckx* | |
245 | ||
246 | * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn() | |
247 | as they are not useful with non-deprecated functions. | |
248 | ||
249 | *Rich Salz* | |
250 | ||
251 | * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(), | |
252 | OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(), | |
253 | OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(), | |
254 | OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These | |
255 | were used to collect all necessary data to form a HTTP request, and to | |
256 | perform the HTTP transfer with that request. With OpenSSL 3.0, the | |
257 | type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced | |
258 | with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(), | |
259 | OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(), | |
260 | OSSL_HTTP_REQ_CTX_set1_req(), OSSL_HTTP_REQ_CTX_nbio(), | |
261 | OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and | |
262 | OSSL_HTTP_REQ_CTX_set_max_response_length(). | |
263 | ||
264 | *Rich Salz and Richard Levitte* | |
265 | ||
266 | * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`. | |
267 | ||
268 | *David von Oheimb* | |
269 | ||
270 | * Validation of SM2 keys has been separated from the validation of regular EC | |
271 | keys, allowing to improve the SM2 validation process to reject loaded private | |
272 | keys that are not conforming to the SM2 ISO standard. | |
273 | In particular, a private scalar `k` outside the range `1 <= k < n-1` is now | |
274 | correctly rejected. | |
275 | ||
276 | *Nicola Tuveri* | |
277 | ||
278 | * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck` | |
279 | switches: a validation failure triggers an early exit, returning a failure | |
280 | exit status to the parent process. | |
281 | ||
282 | *Nicola Tuveri* | |
283 | ||
284 | * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites() | |
285 | to ignore unknown ciphers. | |
286 | ||
287 | *Otto Hollmann* | |
288 | ||
289 | * The `-cipher-commands` and `-digest-commands` options | |
290 | of the command line utility `list` have been deprecated. | |
291 | Instead use the `-cipher-algorithms` and `-digest-algorithms` options. | |
292 | ||
293 | *Dmitry Belyavskiy* | |
294 | ||
295 | * All of the low level EC_KEY functions have been deprecated including: | |
296 | ||
297 | EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method, | |
298 | EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method | |
299 | EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init, | |
300 | EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key, | |
301 | EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify, | |
302 | EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen, | |
303 | EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign, | |
304 | EC_KEY_METHOD_get_verify, | |
305 | EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags, | |
306 | EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params, | |
307 | EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free, | |
308 | EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine, | |
309 | EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key, | |
310 | EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key, | |
311 | EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form, | |
312 | EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data, | |
313 | EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign, | |
314 | EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key, | |
315 | EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf. | |
316 | Applications that need to implement an EC_KEY_METHOD need to consider | |
317 | implementation of the functionality in a special provider. | |
318 | For replacement of the functions manipulating the EC_KEY objects | |
319 | see the EVP_PKEY-EC(7) manual page. | |
320 | ||
321 | Additionally functions that read and write EC_KEY objects such as | |
322 | o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp, | |
323 | d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio, | |
324 | d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp, | |
325 | i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio, | |
326 | i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp | |
327 | have also been deprecated. Applications should instead use the | |
328 | OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files. | |
329 | ||
330 | Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as | |
331 | EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and | |
332 | EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either | |
333 | read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER | |
334 | APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata(). | |
335 | ||
336 | *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz* | |
337 | ||
338 | * Deprecated all the libcrypto and libssl error string loading | |
339 | functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(), | |
340 | ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(), | |
341 | ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(), | |
342 | ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(), | |
343 | ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(), | |
344 | ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(), | |
345 | ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(), | |
346 | ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(), | |
347 | ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(), | |
348 | ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings(). | |
349 | ||
350 | Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL | |
351 | now loads error strings automatically. | |
352 | ||
353 | *Richard Levitte* | |
354 | ||
355 | * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as | |
356 | well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been | |
357 | deprecated. These are used to set the Diffie-Hellman (DH) parameters that | |
358 | are to be used by servers requiring ephemeral DH keys. Instead applications | |
359 | should consider using the built-in DH parameters that are available by | |
360 | calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are | |
361 | necessary then applications can use the alternative functions | |
362 | SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct | |
363 | replacement for the "callback" functions. The callback was originally useful | |
364 | in order to have different parameters for export and non-export ciphersuites. | |
365 | Export ciphersuites are no longer supported by OpenSSL. Use of the callback | |
366 | functions should be replaced by one of the other methods described above. | |
367 | ||
368 | *Matt Caswell* | |
369 | ||
370 | * The `-crypt` option to the `passwd` command line tool has been removed. | |
371 | ||
372 | *Paul Dale* | |
373 | ||
374 | * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands | |
375 | were removed. | |
376 | ||
377 | *Rich Salz* | |
378 | ||
379 | * Add support for AES Key Wrap inverse ciphers to the EVP layer. | |
380 | The algorithms are: | |
381 | "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV", | |
382 | "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV". | |
383 | The inverse ciphers use AES decryption for wrapping, and | |
384 | AES encryption for unwrapping. | |
385 | ||
386 | *Shane Lontis* | |
387 | ||
388 | * Deprecated EVP_PKEY_set1_tls_encodedpoint() and | |
389 | EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by | |
390 | libssl to set or get an encoded public key in/from an EVP_PKEY object. With | |
391 | OpenSSL 3.0 these are replaced by the more generic functions | |
392 | EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key(). | |
393 | The old versions have been converted to deprecated macros that just call the | |
394 | new functions. | |
395 | ||
396 | *Matt Caswell* | |
397 | ||
398 | * The security callback, which can be customised by application code, supports | |
399 | the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY | |
400 | in the "other" parameter. In most places this is what is passed. All these | |
401 | places occur server side. However there was one client side call of this | |
402 | security operation and it passed a DH object instead. This is incorrect | |
403 | according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all | |
404 | of the other locations. Therefore this client side call has been changed to | |
405 | pass an EVP_PKEY instead. | |
406 | ||
407 | *Matt Caswell* | |
408 | ||
409 | * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public | |
410 | interface. Their functionality remains unchanged. | |
411 | ||
412 | *Jordan Montgomery* | |
413 | ||
414 | * Added new option for 'openssl list', '-providers', which will display the | |
415 | list of loaded providers, their names, version and status. It optionally | |
416 | displays their gettable parameters. | |
417 | ||
418 | *Paul Dale* | |
419 | ||
420 | * Deprecated EVP_PKEY_set_alias_type(). This function was previously | |
421 | needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key | |
422 | type is internally recognised so the workaround is no longer needed. | |
423 | ||
424 | Functionality is still retained as it is, but will only work with | |
425 | EVP_PKEYs with a legacy internal key. | |
426 | ||
427 | *Richard Levitte* | |
428 | ||
429 | * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced | |
430 | `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred. | |
431 | ||
432 | *Jeremy Walch* | |
433 | ||
434 | * Changed all "STACK" functions to be macros instead of inline functions. Macro | |
435 | parameters are still checked for type safety at compile time via helper | |
436 | inline functions. | |
437 | ||
438 | *Matt Caswell* | |
439 | ||
440 | * Remove the RAND_DRBG API | |
441 | ||
442 | The RAND_DRBG API did not fit well into the new provider concept as | |
443 | implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the | |
444 | RAND_DRBG API is a mixture of 'front end' and 'back end' API calls | |
445 | and some of its API calls are rather low-level. This holds in particular | |
446 | for the callback mechanism (`RAND_DRBG_set_callbacks()`). | |
447 | ||
448 | Adding a compatibility layer to continue supporting the RAND_DRBG API as | |
449 | a legacy API for a regular deprecation period turned out to come at the | |
450 | price of complicating the new provider API unnecessarily. Since the | |
451 | RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC | |
452 | to drop it entirely. | |
453 | ||
454 | *Paul Dale and Matthias St. Pierre* | |
455 | ||
456 | * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses | |
457 | as well as actual hostnames. | |
458 | ||
459 | *David Woodhouse* | |
460 | ||
461 | * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently | |
462 | ignore TLS protocol version bounds when configuring DTLS-based contexts, and | |
463 | conversely, silently ignore DTLS protocol version bounds when configuring | |
464 | TLS-based contexts. The commands can be repeated to set bounds of both | |
465 | types. The same applies with the corresponding "min_protocol" and | |
466 | "max_protocol" command-line switches, in case some application uses both TLS | |
467 | and DTLS. | |
468 | ||
469 | SSL_CTX instances that are created for a fixed protocol version (e.g. | |
470 | `TLSv1_server_method()`) also silently ignore version bounds. Previously | |
471 | attempts to apply bounds to these protocol versions would result in an | |
472 | error. Now only the "version-flexible" SSL_CTX instances are subject to | |
473 | limits in configuration files in command-line options. | |
474 | ||
475 | *Viktor Dukhovni* | |
476 | ||
477 | * Deprecated the `ENGINE` API. Engines should be replaced with providers | |
478 | going forward. | |
479 | ||
480 | *Paul Dale* | |
481 | ||
482 | * Reworked the recorded ERR codes to make better space for system errors. | |
483 | To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the | |
484 | given code is a system error (true) or an OpenSSL error (false). | |
485 | ||
486 | *Richard Levitte* | |
487 | ||
488 | * Reworked the test perl framework to better allow parallel testing. | |
489 | ||
490 | *Nicola Tuveri and David von Oheimb* | |
491 | ||
492 | * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and | |
493 | AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported. | |
494 | ||
495 | *Shane Lontis* | |
496 | ||
497 | * 'Configure' has been changed to figure out the configuration target if | |
498 | none is given on the command line. Consequently, the 'config' script is | |
499 | now only a mere wrapper. All documentation is changed to only mention | |
500 | 'Configure'. | |
501 | ||
502 | *Rich Salz and Richard Levitte* | |
503 | ||
504 | * Added a library context `OSSL_LIB_CTX` that applications as well as | |
505 | other libraries can use to form a separate context within which | |
506 | libcrypto operations are performed. | |
507 | ||
508 | There are two ways this can be used: | |
509 | ||
510 | - Directly, by passing a library context to functions that take | |
511 | such an argument, such as `EVP_CIPHER_fetch` and similar algorithm | |
512 | fetching functions. | |
513 | - Indirectly, by creating a new library context and then assigning | |
514 | it as the new default, with `OSSL_LIB_CTX_set0_default`. | |
515 | ||
516 | All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer, | |
517 | apart from the functions directly related to `OSSL_LIB_CTX`, accept | |
518 | NULL to indicate that the default library context should be used. | |
519 | ||
520 | Library code that changes the default library context using | |
521 | `OSSL_LIB_CTX_set0_default` should take care to restore it with a | |
522 | second call before returning to the caller. | |
523 | ||
524 | _(Note: the library context was initially called `OPENSSL_CTX` and | |
525 | renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_ | |
526 | ||
527 | *Richard Levitte* | |
528 | ||
529 | * Handshake now fails if Extended Master Secret extension is dropped | |
530 | on renegotiation. | |
531 | ||
532 | *Tomáš Mráz* | |
533 | ||
534 | * Dropped interactive mode from the `openssl` program. From now on, | |
535 | running it without arguments is equivalent to `openssl help`. | |
536 | ||
537 | *Richard Levitte* | |
538 | ||
539 | * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and | |
540 | `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`. | |
541 | While the old function names have been retained for backward compatibility | |
542 | they should not be used in new developments | |
543 | because their return values are confusing: Unlike other `_cmp()` functions | |
544 | they do not return 0 in case their arguments are equal. | |
545 | ||
546 | *David von Oheimb* | |
547 | ||
548 | * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to | |
549 | `EC_GROUP_get_field_type()`. | |
550 | ||
551 | *Billy Bob Brumley* | |
552 | ||
553 | * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(), | |
554 | EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method() | |
555 | EC_GFp_nistp256_method(), and EC_GFp_nistp521_method(). | |
556 | Applications should rely on the library automatically assigning a suitable | |
557 | EC_METHOD internally upon EC_GROUP construction. | |
558 | ||
559 | *Billy Bob Brumley* | |
560 | ||
561 | * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of(). | |
562 | EC_METHOD is now an internal-only concept and a suitable EC_METHOD is | |
563 | assigned internally without application intervention. | |
564 | Users of EC_GROUP_new() should switch to a different suitable constructor. | |
565 | ||
566 | *Billy Bob Brumley* | |
567 | ||
568 | * Add CAdES-BES signature verification support, mostly derived | |
569 | from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein. | |
570 | ||
571 | *Filipe Raimundo da Silva* | |
572 | ||
573 | * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API. | |
574 | ||
575 | *Antonio Iacono* | |
576 | ||
577 | * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM | |
578 | parameter (RFC 5084) for the Cryptographic Message Syntax (CMS). Its purpose | |
579 | is to support encryption and decryption of a digital envelope that is both | |
580 | authenticated and encrypted using AES GCM mode. | |
581 | ||
582 | *Jakub Zelenka* | |
583 | ||
584 | * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These | |
585 | functions are not widely used and now OpenSSL automatically perform this | |
586 | conversion when needed. | |
587 | ||
588 | *Billy Bob Brumley* | |
589 | ||
590 | * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and | |
591 | EC_KEY_precompute_mult(). These functions are not widely used and | |
592 | applications should instead switch to named curves which OpenSSL has | |
593 | hardcoded lookup tables for. | |
594 | ||
595 | *Billy Bob Brumley* | |
596 | ||
597 | * Deprecated EC_POINTs_mul(). This function is not widely used and applications | |
598 | should instead use the L<EC_POINT_mul(3)> function. | |
599 | ||
600 | *Billy Bob Brumley* | |
601 | ||
602 | * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's | |
603 | that are not applicable to the new provider model. Applications should | |
604 | instead use EVP_default_properties_is_fips_enabled() and | |
605 | EVP_default_properties_enable_fips(). | |
606 | ||
607 | *Shane Lontis* | |
608 | ||
609 | * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option | |
610 | is set, an unexpected EOF is ignored, it pretends a close notify was received | |
611 | instead and so the returned error becomes SSL_ERROR_ZERO_RETURN. | |
612 | ||
613 | *Dmitry Belyavskiy* | |
614 | ||
615 | * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and | |
616 | EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely | |
617 | used and applications should instead use the | |
618 | L<EC_POINT_set_affine_coordinates(3)> and | |
619 | L<EC_POINT_get_affine_coordinates(3)> functions. | |
620 | ||
621 | *Billy Bob Brumley* | |
622 | ||
623 | * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM | |
624 | arrays to be more easily constructed via a series of utility functions. | |
625 | Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using | |
626 | the various push functions and finally convert to a passable OSSL_PARAM | |
627 | array using OSSL_PARAM_BLD_to_param(). | |
628 | ||
629 | *Paul Dale* | |
630 | ||
631 | * The security strength of SHA1 and MD5 based signatures in TLS has been | |
632 | reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer | |
633 | working at the default security level of 1 and instead requires security | |
634 | level 0. The security level can be changed either using the cipher string | |
635 | with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`. This also means | |
636 | that where the signature algorithms extension is missing from a ClientHello | |
637 | then the handshake will fail in TLS 1.2 at security level 1. This is because, | |
638 | although this extension is optional, failing to provide one means that | |
639 | OpenSSL will fallback to a default set of signature algorithms. This default | |
640 | set requires the availability of SHA1. | |
641 | ||
642 | *Kurt Roeckx* | |
643 | ||
644 | * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to | |
645 | contain a provider side internal key. | |
646 | ||
647 | *Richard Levitte* | |
648 | ||
649 | * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated. | |
650 | They are old functions that we don't use, and that you could disable with | |
651 | the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7. | |
652 | ||
653 | *Richard Levitte* | |
654 | ||
655 | * Project text documents not yet having a proper file name extension | |
656 | (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to | |
657 | `*.md` as far as reasonable, else `*.txt`, for better use with file managers. | |
658 | ||
659 | *David von Oheimb* | |
660 | ||
661 | * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT) | |
662 | have been converted to Markdown with the goal to produce documents | |
663 | which not only look pretty when viewed online in the browser, but | |
664 | remain well readable inside a plain text editor. | |
665 | ||
666 | To achieve this goal, a 'minimalistic' Markdown style has been applied | |
667 | which avoids formatting elements that interfere too much with the | |
668 | reading flow in the text file. For example, it | |
669 | ||
670 | * avoids [ATX headings][] and uses [setext headings][] instead | |
671 | (which works for `<h1>` and `<h2>` headings only). | |
672 | * avoids [inline links][] and uses [reference links][] instead. | |
673 | * avoids [fenced code blocks][] and uses [indented code blocks][] instead. | |
674 | ||
675 | [ATX headings]: https://github.github.com/gfm/#atx-headings | |
676 | [setext headings]: https://github.github.com/gfm/#setext-headings | |
677 | [inline links]: https://github.github.com/gfm/#inline-link | |
678 | [reference links]: https://github.github.com/gfm/#reference-link | |
679 | [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks | |
680 | [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks | |
681 | ||
682 | *Matthias St. Pierre* | |
683 | ||
684 | * The test suite is changed to preserve results of each test recipe. | |
685 | A new directory test-runs/ with subdirectories named like the | |
686 | test recipes are created in the build tree for this purpose. | |
687 | ||
688 | *Richard Levitte* | |
689 | ||
690 | * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712). | |
691 | This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`. | |
692 | See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points. | |
693 | ||
694 | *David von Oheimb, Martin Peylo* | |
695 | ||
696 | * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`. | |
697 | The legacy OCSP-focused and only partly documented API is retained for | |
698 | backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details. | |
699 | ||
700 | *David von Oheimb* | |
701 | ||
702 | * Added `util/check-format.pl`, a tool for checking adherence to the | |
703 | OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>. | |
704 | The checks performed are incomplete and yield some false positives. | |
705 | Still the tool should be useful for detecting most typical glitches. | |
706 | ||
707 | *David von Oheimb* | |
708 | ||
709 | * `BIO_do_connect()` and `BIO_do_handshake()` have been extended: | |
710 | If domain name resolution yields multiple IP addresses all of them are tried | |
711 | after `connect()` failures. | |
712 | ||
713 | *David von Oheimb* | |
714 | ||
715 | * All of the low level RSA functions have been deprecated including: | |
716 | ||
717 | RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params, | |
718 | RSA_get_version, RSA_get0_engine, RSA_generate_key_ex, | |
719 | RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex, | |
720 | RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt, | |
721 | RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method, | |
722 | RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method, | |
723 | RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify, | |
724 | RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on, | |
725 | RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1, | |
726 | RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2, | |
727 | RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP, | |
728 | RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1, | |
729 | RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23, | |
730 | RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none, | |
731 | RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id, | |
732 | RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1, | |
733 | RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data, | |
734 | RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name, | |
735 | RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags, | |
736 | RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc, | |
737 | RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec, | |
738 | RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec, | |
739 | RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp, | |
740 | RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init, | |
741 | RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish, | |
742 | RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify, | |
743 | RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen, | |
744 | RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen. | |
745 | ||
746 | Use of these low level functions has been informally discouraged for a long | |
747 | time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>, | |
748 | L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and | |
749 | L<EVP_PKEY_decrypt(3)>. | |
750 | ||
751 | *Paul Dale* | |
752 | ||
753 | * X509 certificates signed using SHA1 are no longer allowed at security | |
754 | level 1 and above. | |
755 | In TLS/SSL the default security level is 1. It can be set either | |
756 | using the cipher string with `@SECLEVEL`, or calling | |
757 | `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1, | |
758 | a call to `SSL_CTX_use_certificate()` will fail if the security level is not | |
759 | lowered first. | |
760 | Outside TLS/SSL, the default security level is -1 (effectively 0). It can | |
761 | be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level` | |
762 | options of the commands. | |
763 | ||
764 | *Kurt Roeckx* | |
765 | ||
766 | * The command line utilities dhparam, dsa, gendsa and dsaparam have been | |
767 | modified to use PKEY APIs. These commands are now in maintenance mode | |
768 | and no new features will be added to them. | |
769 | ||
770 | *Paul Dale* | |
771 | ||
772 | * The command line utility rsautl has been deprecated. | |
773 | Instead use the pkeyutl program. | |
774 | ||
775 | *Paul Dale* | |
776 | ||
777 | * The command line utilities genrsa and rsa have been modified to use PKEY | |
778 | APIs. They now write PKCS#8 keys by default. These commands are now in | |
779 | maintenance mode and no new features will be added to them. | |
780 | ||
781 | *Paul Dale* | |
782 | ||
783 | * All of the low level DH functions have been deprecated including: | |
784 | ||
785 | DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method, | |
786 | DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size, | |
787 | DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data, | |
788 | DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex, | |
789 | DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key, | |
790 | DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid, | |
791 | DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup, | |
792 | DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags, | |
793 | DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key, | |
794 | DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key, | |
795 | DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init, | |
796 | DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish, | |
797 | DH_meth_get_generate_params and DH_meth_set_generate_params. | |
798 | ||
799 | Use of these low level functions has been informally discouraged for a long | |
800 | time. Instead applications should use L<EVP_PKEY_derive_init(3)> | |
801 | and L<EVP_PKEY_derive(3)>. | |
802 | ||
803 | Additionally functions that read and write DH objects such as d2i_DHparams, | |
804 | i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar | |
805 | functions have also been deprecated. Applications should instead use the | |
806 | OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files. | |
807 | ||
808 | Finaly functions that assign or obtain DH objects from an EVP_PKEY such as | |
809 | `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and | |
810 | `EVP_PKEY_set1_DH()` are also deprecated. | |
811 | Applications should instead either read or write an | |
812 | EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs. | |
813 | Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`. | |
814 | ||
815 | *Paul Dale and Matt Caswell* | |
816 | ||
817 | * All of the low level DSA functions have been deprecated including: | |
818 | ||
819 | DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg, | |
820 | DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g, | |
821 | DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags, | |
822 | DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL, | |
823 | DSA_set_default_method, DSA_get_default_method, DSA_set_method, | |
824 | DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits, | |
825 | DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index, | |
826 | DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex, | |
827 | DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free, | |
828 | DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags, | |
829 | DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data, | |
830 | DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup, | |
831 | DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify, | |
832 | DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp, | |
833 | DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init, | |
834 | DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen, | |
835 | DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen. | |
836 | ||
837 | Use of these low level functions has been informally discouraged for a long | |
838 | time. Instead applications should use L<EVP_DigestSignInit_ex(3)>, | |
839 | L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>. | |
840 | ||
841 | Finaly functions that assign or obtain DSA objects from an EVP_PKEY such as | |
842 | `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and | |
843 | `EVP_PKEY_set1_DSA()` are also deprecated. | |
844 | Applications should instead either read or write an | |
845 | EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs, | |
846 | or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`. | |
847 | ||
848 | *Paul Dale* | |
849 | ||
850 | * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to | |
851 | automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC. | |
852 | This means that applications don't have to look at the curve NID and | |
853 | `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations. | |
854 | However, they still can, that `EVP_PKEY_set_alias_type()` call acts as | |
855 | a no-op when the EVP_PKEY is already of the given type. | |
856 | ||
857 | Parameter and key generation is also reworked to make it possible | |
858 | to generate EVP_PKEY_SM2 parameters and keys without having to go | |
859 | through EVP_PKEY_EC generation and then change the EVP_PKEY type. | |
860 | However, code that does the latter will still work as before. | |
861 | ||
862 | *Richard Levitte* | |
863 | ||
864 | * Deprecated low level ECDH and ECDSA functions. These include: | |
865 | ||
866 | ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify, | |
867 | ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and | |
868 | ECDSA_size. | |
869 | ||
870 | Use of these low level functions has been informally discouraged for a long | |
871 | time. Instead applications should use the EVP_PKEY_derive(3), | |
872 | EVP_DigestSign(3) and EVP_DigestVerify(3) functions. | |
873 | ||
874 | *Paul Dale* | |
875 | ||
876 | * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init() | |
877 | and EVP_PKEY_decrypt() instead. | |
878 | Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init() | |
879 | and EVP_PKEY_encrypt() instead. | |
880 | ||
881 | *Richard Levitte* | |
882 | ||
883 | * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits() | |
884 | and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed | |
885 | a new formulation to include all the things it can be used for, | |
886 | as well as words of caution. | |
887 | ||
888 | *Richard Levitte* | |
889 | ||
890 | * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated. | |
891 | Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function. | |
892 | ||
893 | *Paul Dale* | |
894 | ||
895 | * All of the low level HMAC functions have been deprecated including: | |
896 | ||
897 | HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free, | |
898 | HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags | |
899 | and HMAC_CTX_get_md. | |
900 | ||
901 | Use of these low level functions has been informally discouraged for a long | |
902 | time. Instead applications should use L<EVP_MAC_CTX_new(3)>, | |
903 | L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)> | |
904 | and L<EVP_MAC_final(3)>. | |
905 | ||
906 | *Paul Dale* | |
907 | ||
908 | * Over two thousand fixes were made to the documentation, including: | |
909 | - Common options (such as -rand/-writerand, TLS version control, etc) | |
910 | were refactored and point to newly-enhanced descriptions in openssl.pod. | |
911 | - Added style conformance for all options (with help from Richard Levitte), | |
912 | documented all reported missing options, added a CI build to check | |
913 | that all options are documented and that no unimplemented options | |
914 | are documented. | |
915 | - Documented some internals, such as all use of environment variables. | |
916 | - Addressed all internal broken L<> references. | |
917 | ||
918 | *Rich Salz* | |
919 | ||
920 | * All of the low level CMAC functions have been deprecated including: | |
921 | ||
922 | CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx, | |
923 | CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume. | |
924 | ||
925 | Use of these low level functions has been informally discouraged for a long | |
926 | time. Instead applications should use L<EVP_MAC_CTX_new(3)>, | |
927 | L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)> | |
928 | and L<EVP_MAC_final(3)>. | |
929 | ||
930 | *Paul Dale* | |
931 | ||
932 | * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256, | |
933 | SHA384, SHA512 and Whirlpool digest functions have been deprecated. | |
934 | These include: | |
935 | ||
936 | MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init, | |
937 | MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update, | |
938 | MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final, | |
939 | RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final, | |
940 | RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform, | |
941 | SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init, | |
942 | SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init, | |
943 | SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update, | |
944 | SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init, | |
945 | WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final. | |
946 | ||
947 | Use of these low level functions has been informally discouraged | |
948 | for a long time. Applications should use the EVP_DigestInit_ex(3), | |
949 | EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead. | |
950 | ||
951 | *Paul Dale* | |
952 | ||
953 | * Corrected the documentation of the return values from the `EVP_DigestSign*` | |
954 | set of functions. The documentation mentioned negative values for some | |
955 | errors, but this was never the case, so the mention of negative values | |
956 | was removed. | |
957 | ||
958 | Code that followed the documentation and thereby check with something | |
959 | like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed. | |
960 | ||
961 | *Richard Levitte* | |
962 | ||
963 | * All of the low level cipher functions have been deprecated including: | |
964 | ||
965 | AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt, | |
966 | AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt, | |
967 | AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt, | |
968 | AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt, | |
969 | BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt, | |
970 | BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt, | |
971 | Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt, | |
972 | Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt, | |
973 | Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt, | |
974 | CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt, | |
975 | CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2, | |
976 | DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt, | |
977 | DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt, | |
978 | DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key, | |
979 | DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key, | |
980 | DES_key_sched, DES_set_key_checked, DES_set_key_unchecked, | |
981 | DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity, | |
982 | DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt, | |
983 | DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt, | |
984 | DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt, | |
985 | DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options, | |
986 | IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key, | |
987 | IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key, | |
988 | RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt, | |
989 | RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key, | |
990 | RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt, | |
991 | RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt, | |
992 | SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt, | |
993 | SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt. | |
994 | ||
995 | Use of these low level functions has been informally discouraged for | |
996 | a long time. Applications should use the high level EVP APIs, e.g. | |
997 | EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the | |
998 | equivalently named decrypt functions instead. | |
999 | ||
1000 | *Matt Caswell and Paul Dale* | |
1001 | ||
1002 | * Removed include/openssl/opensslconf.h.in and replaced it with | |
1003 | include/openssl/configuration.h.in, which differs in not including | |
1004 | <openssl/macros.h>. A short header include/openssl/opensslconf.h | |
1005 | was added to include both. | |
1006 | ||
1007 | This allows internal hacks where one might need to modify the set | |
1008 | of configured macros, for example this if deprecated symbols are | |
1009 | still supposed to be available internally: | |
1010 | ||
1011 | #include <openssl/configuration.h> | |
1012 | ||
1013 | #undef OPENSSL_NO_DEPRECATED | |
1014 | #define OPENSSL_SUPPRESS_DEPRECATED | |
1015 | ||
1016 | #include <openssl/macros.h> | |
1017 | ||
1018 | This should not be used by applications that use the exported | |
1019 | symbols, as that will lead to linking errors. | |
1020 | ||
1021 | *Richard Levitte* | |
1022 | ||
1023 | * Fixed an overflow bug in the x64_64 Montgomery squaring procedure | |
1024 | used in exponentiation with 512-bit moduli. No EC algorithms are | |
1025 | affected. Analysis suggests that attacks against 2-prime RSA1024, | |
1026 | 3-prime RSA1536, and DSA1024 as a result of this defect would be very | |
1027 | difficult to perform and are not believed likely. Attacks against DH512 | |
1028 | are considered just feasible. However, for an attack the target would | |
1029 | have to re-use the DH512 private key, which is not recommended anyway. | |
1030 | Also applications directly using the low level API BN_mod_exp may be | |
1031 | affected if they use BN_FLG_CONSTTIME. | |
1032 | ([CVE-2019-1551]) | |
1033 | ||
1034 | *Andy Polyakov* | |
1035 | ||
1036 | * Most memory-debug features have been deprecated, and the functionality | |
1037 | replaced with no-ops. | |
1038 | ||
1039 | *Rich Salz* | |
1040 | ||
1041 | * Added documentation for the STACK API. | |
1042 | ||
1043 | *Rich Salz* | |
1044 | ||
1045 | * Introduced a new method type and API, OSSL_ENCODER, to | |
1046 | represent generic encoders. An implementation is expected to | |
1047 | be able to encode an object associated with a given name (such | |
1048 | as an algorithm name for an asymmetric key) into forms given by | |
1049 | implementation properties. | |
1050 | ||
1051 | Encoders are primarily used from inside libcrypto, through | |
1052 | calls to functions like EVP_PKEY_print_private(), | |
1053 | PEM_write_bio_PrivateKey() and similar. | |
1054 | ||
1055 | Encoders are specified in such a way that they can be made to | |
1056 | directly handle the provider side portion of an object, if this | |
1057 | provider side part comes from the same provider as the encoder | |
1058 | itself, but can also be made to handle objects in parametrized | |
1059 | form (as an OSSL_PARAM array of data). This allows a provider to | |
1060 | offer generic encoders as a service for any other provider. | |
1061 | ||
1062 | *Richard Levitte* | |
1063 | ||
1064 | * Added a .pragma directive to the syntax of configuration files, to | |
1065 | allow varying behavior in a supported and predictable manner. | |
1066 | Currently added pragma: | |
1067 | ||
1068 | .pragma dollarid:on | |
1069 | ||
1070 | This allows dollar signs to be a keyword character unless it's | |
1071 | followed by a opening brace or parenthesis. This is useful for | |
1072 | platforms where dollar signs are commonly used in names, such as | |
1073 | volume names and system directory names on VMS. | |
1074 | ||
1075 | *Richard Levitte* | |
1076 | ||
1077 | * Added functionality to create an EVP_PKEY from user data. This | |
1078 | is effectively the same as creating a RSA, DH or DSA object and | |
1079 | then assigning them to an EVP_PKEY, but directly using algorithm | |
1080 | agnostic EVP functions. A benefit is that this should be future | |
1081 | proof for public key algorithms to come. | |
1082 | ||
1083 | *Richard Levitte* | |
1084 | ||
1085 | * Change the interpretation of the '--api' configuration option to | |
1086 | mean that this is a desired API compatibility level with no | |
1087 | further meaning. The previous interpretation, that this would | |
1088 | also mean to remove all deprecated symbols up to and including | |
1089 | the given version, no requires that 'no-deprecated' is also used | |
1090 | in the configuration. | |
1091 | ||
1092 | When building applications, the desired API compatibility level | |
1093 | can be set with the OPENSSL_API_COMPAT macro like before. For | |
1094 | API compatibility version below 3.0, the old style numerical | |
1095 | value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L. | |
1096 | For version 3.0 and on, the value is expected to be the decimal | |
1097 | value calculated from the major and minor version like this: | |
1098 | ||
1099 | MAJOR * 10000 + MINOR * 100 | |
1100 | ||
1101 | Examples: | |
1102 | ||
1103 | -DOPENSSL_API_COMPAT=30000 For 3.0 | |
1104 | -DOPENSSL_API_COMPAT=30200 For 3.2 | |
1105 | ||
1106 | To hide declarations that are deprecated up to and including the | |
1107 | given API compatibility level, -DOPENSSL_NO_DEPRECATED must be | |
1108 | given when building the application as well. | |
1109 | ||
1110 | *Richard Levitte* | |
1111 | ||
1112 | * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow | |
1113 | access to certificate and CRL stores via URIs and OSSL_STORE | |
1114 | loaders. | |
1115 | ||
1116 | This adds the following functions: | |
1117 | ||
1118 | - X509_LOOKUP_store() | |
1119 | - X509_STORE_load_file() | |
1120 | - X509_STORE_load_path() | |
1121 | - X509_STORE_load_store() | |
1122 | - SSL_add_store_cert_subjects_to_stack() | |
1123 | - SSL_CTX_set_default_verify_store() | |
1124 | - SSL_CTX_load_verify_file() | |
1125 | - SSL_CTX_load_verify_dir() | |
1126 | - SSL_CTX_load_verify_store() | |
1127 | ||
1128 | *Richard Levitte* | |
1129 | ||
1130 | * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY. | |
1131 | The presence of this system service is determined at run-time. | |
1132 | ||
1133 | *Richard Levitte* | |
1134 | ||
1135 | * Added functionality to create an EVP_PKEY context based on data | |
1136 | for methods from providers. This takes an algorithm name and a | |
1137 | property query string and simply stores them, with the intent | |
1138 | that any operation that uses this context will use those strings | |
1139 | to fetch the needed methods implicitly, thereby making the port | |
1140 | of application written for pre-3.0 OpenSSL easier. | |
1141 | ||
1142 | *Richard Levitte* | |
1143 | ||
1144 | * The undocumented function NCONF_WIN32() has been deprecated; for | |
1145 | conversion details see the HISTORY section of doc/man5/config.pod | |
1146 | ||
1147 | *Rich Salz* | |
1148 | ||
1149 | * Introduced the new functions EVP_DigestSignInit_ex() and | |
1150 | EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and | |
1151 | EVP_DigestVerifyUpdate() have been converted to functions. See the man | |
1152 | pages for further details. | |
1153 | ||
1154 | *Matt Caswell* | |
1155 | ||
1156 | * Over two thousand fixes were made to the documentation, including: | |
1157 | adding missing command flags, better style conformance, documentation | |
1158 | of internals, etc. | |
1159 | ||
1160 | *Rich Salz, Richard Levitte* | |
1161 | ||
1162 | * s390x assembly pack: add hardware-support for P-256, P-384, P-521, | |
1163 | X25519, X448, Ed25519 and Ed448. | |
1164 | ||
1165 | *Patrick Steuer* | |
1166 | ||
1167 | * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just | |
1168 | the first value. | |
1169 | ||
1170 | *Jon Spillett* | |
1171 | ||
1172 | * Deprecated the public definition of `ERR_STATE` as well as the function | |
1173 | `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an | |
1174 | opaque type. | |
1175 | ||
1176 | *Richard Levitte* | |
1177 | ||
1178 | * Added ERR functionality to give callers access to the stored function | |
1179 | names that have replaced the older function code based functions. | |
1180 | ||
1181 | New functions are ERR_peek_error_func(), ERR_peek_last_error_func(), | |
1182 | ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(), | |
1183 | ERR_peek_error_all() and ERR_peek_last_error_all(). | |
1184 | ||
1185 | These functions have become deprecated: ERR_get_error_line(), | |
1186 | ERR_get_error_line_data(), ERR_peek_error_line_data(), | |
1187 | ERR_peek_last_error_line_data() and ERR_func_error_string(). | |
1188 | ||
1189 | Users are recommended to use ERR_get_error_all(), or to pick information | |
1190 | with ERR_peek functions and finish off with getting the error code by using | |
1191 | ERR_get_error(). | |
1192 | ||
1193 | *Richard Levitte* | |
1194 | ||
1195 | * Extended testing to be verbose for failing tests only. The make variables | |
1196 | VERBOSE_FAILURE or VF can be used to enable this: | |
1197 | ||
1198 | $ make VF=1 test # Unix | |
1199 | $ mms /macro=(VF=1) test ! OpenVMS | |
1200 | $ nmake VF=1 test # Windows | |
1201 | ||
1202 | *Richard Levitte* | |
1203 | ||
1204 | * Added the `-copy_extensions` option to the `x509` command for use with | |
1205 | `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument, | |
1206 | all extensions in the request are copied to the certificate or vice versa. | |
1207 | ||
1208 | *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>* | |
1209 | ||
1210 | * Added the `-copy_extensions` option to the `req` command for use with | |
1211 | `-x509`. When given with the `copy` or `copyall` argument, | |
1212 | all extensions in the certification request are copied to the certificate. | |
1213 | ||
1214 | *David von Oheimb* | |
1215 | ||
1216 | * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates | |
1217 | they generate are by default RFC 5280 compliant in the following sense: | |
1218 | There is a subjectKeyIdentifier extension with a hash value of the public key | |
1219 | and for not self-signed certs there is an authorityKeyIdentifier extension | |
1220 | with a keyIdentifier field or issuer information identifying the signing key. | |
1221 | This is done unless some configuration overrides the new default behavior, | |
1222 | such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`. | |
1223 | ||
1224 | *David von Oheimb* | |
1225 | ||
1226 | * Added several checks to `X509_verify_cert()` according to requirements in | |
1227 | RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set | |
1228 | (which may be done by using the CLI option `-x509_strict`): | |
1229 | * The basicConstraints of CA certificates must be marked critical. | |
1230 | * CA certificates must explicitly include the keyUsage extension. | |
1231 | * If a pathlenConstraint is given the key usage keyCertSign must be allowed. | |
1232 | * The issuer name of any certificate must not be empty. | |
1233 | * The subject name of CA certs, certs with keyUsage crlSign, | |
1234 | and certs without subjectAlternativeName must not be empty. | |
1235 | * If a subjectAlternativeName extension is given it must not be empty. | |
1236 | * The signatureAlgorithm field and the cert signature must be consistent. | |
1237 | * Any given authorityKeyIdentifier and any given subjectKeyIdentifier | |
1238 | must not be marked critical. | |
1239 | * The authorityKeyIdentifier must be given for X.509v3 certs | |
1240 | unless they are self-signed. | |
1241 | * The subjectKeyIdentifier must be given for all X.509v3 CA certs. | |
1242 | ||
1243 | *David von Oheimb* | |
1244 | ||
1245 | * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys | |
1246 | with explicit curve parameters (specifiedCurve) as required by RFC 5480. | |
1247 | ||
1248 | *Tomáš Mráz* | |
1249 | ||
1250 | * For built-in EC curves, ensure an EC_GROUP built from the curve name is | |
1251 | used even when parsing explicit parameters, when loading a encoded key | |
1252 | or calling `EC_GROUP_new_from_ecpkparameters()`/ | |
1253 | `EC_GROUP_new_from_ecparameters()`. | |
1254 | This prevents bypass of security hardening and performance gains, | |
1255 | especially for curves with specialized EC_METHODs. | |
1256 | By default, if a key encoded with explicit parameters is loaded and later | |
1257 | encoded, the output is still encoded with explicit parameters, even if | |
1258 | internally a "named" EC_GROUP is used for computation. | |
1259 | ||
1260 | *Nicola Tuveri* | |
1261 | ||
1262 | * Compute ECC cofactors if not provided during EC_GROUP construction. Before | |
1263 | this change, EC_GROUP_set_generator would accept order and/or cofactor as | |
1264 | NULL. After this change, only the cofactor parameter can be NULL. It also | |
1265 | does some minimal sanity checks on the passed order. | |
1266 | ([CVE-2019-1547]) | |
1267 | ||
1268 | *Billy Bob Brumley* | |
1269 | ||
1270 | * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey. | |
1271 | An attack is simple, if the first CMS_recipientInfo is valid but the | |
1272 | second CMS_recipientInfo is chosen ciphertext. If the second | |
1273 | recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct | |
1274 | encryption key will be replaced by garbage, and the message cannot be | |
1275 | decoded, but if the RSA decryption fails, the correct encryption key is | |
1276 | used and the recipient will not notice the attack. | |
1277 | As a work around for this potential attack the length of the decrypted | |
1278 | key must be equal to the cipher default key length, in case the | |
1279 | certifiate is not given and all recipientInfo are tried out. | |
1280 | The old behaviour can be re-enabled in the CMS code by setting the | |
1281 | CMS_DEBUG_DECRYPT flag. | |
1282 | ||
1283 | *Bernd Edlinger* | |
1284 | ||
1285 | * Early start up entropy quality from the DEVRANDOM seed source has been | |
1286 | improved for older Linux systems. The RAND subsystem will wait for | |
1287 | /dev/random to be producing output before seeding from /dev/urandom. | |
1288 | The seeded state is stored for future library initialisations using | |
1289 | a system global shared memory segment. The shared memory identifier | |
1290 | can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to | |
1291 | the desired value. The default identifier is 114. | |
1292 | ||
1293 | *Paul Dale* | |
1294 | ||
1295 | * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1 | |
1296 | when primes for RSA keys are computed. | |
1297 | Since we previously always generated primes == 2 (mod 3) for RSA keys, | |
1298 | the 2-prime and 3-prime RSA modules were easy to distinguish, since | |
1299 | `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting | |
1300 | 2-prime vs. 3-prime RSA keys was possible by computing N mod 3. | |
1301 | This avoids possible fingerprinting of newly generated RSA modules. | |
1302 | ||
1303 | *Bernd Edlinger* | |
1304 | ||
1305 | * Correct the extended master secret constant on EBCDIC systems. Without this | |
1306 | fix TLS connections between an EBCDIC system and a non-EBCDIC system that | |
1307 | negotiate EMS will fail. Unfortunately this also means that TLS connections | |
1308 | between EBCDIC systems with this fix, and EBCDIC systems without this | |
1309 | fix will fail if they negotiate EMS. | |
1310 | ||
1311 | *Matt Caswell* | |
1312 | ||
1313 | * Changed the library initialisation so that the config file is now loaded | |
1314 | by default. This was already the case for libssl. It now occurs for both | |
1315 | libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to | |
1316 | `OPENSSL_init_crypto()` to suppress automatic loading of a config file. | |
1317 | ||
1318 | *Matt Caswell* | |
1319 | ||
1320 | * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`, | |
1321 | where the former acts as a replacement for `ERR_put_error()`, and the | |
1322 | latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`. | |
1323 | `ERR_raise_data()` adds more flexibility by taking a format string and | |
1324 | an arbitrary number of arguments following it, to be processed with | |
1325 | `BIO_snprintf()`. | |
1326 | ||
1327 | *Richard Levitte* | |
1328 | ||
1329 | * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used | |
1330 | to check if a named provider is loaded and available. When called, it | |
1331 | will also activate all fallback providers if such are still present. | |
1332 | ||
1333 | *Richard Levitte* | |
1334 | ||
1335 | * Enforce a minimum DH modulus size of 512 bits. | |
1336 | ||
1337 | *Bernd Edlinger* | |
1338 | ||
1339 | * Changed DH parameters to generate the order q subgroup instead of 2q. | |
1340 | Previously generated DH parameters are still accepted by DH_check | |
1341 | but DH_generate_key works around that by clearing bit 0 of the | |
1342 | private key for those. This avoids leaking bit 0 of the private key. | |
1343 | ||
1344 | *Bernd Edlinger* | |
1345 | ||
1346 | * Significantly reduce secure memory usage by the randomness pools. | |
1347 | ||
1348 | *Paul Dale* | |
1349 | ||
1350 | * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been | |
1351 | deprecated. | |
1352 | ||
1353 | *Rich Salz* | |
1354 | ||
1355 | * A new type, EVP_KEYEXCH, has been introduced to represent key exchange | |
1356 | algorithms. An implementation of a key exchange algorithm can be obtained | |
1357 | by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be | |
1358 | used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to | |
1359 | the older EVP_PKEY_derive_init() function. See the man pages for the new | |
1360 | functions for further details. | |
1361 | ||
1362 | *Matt Caswell* | |
1363 | ||
1364 | * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function. | |
1365 | ||
1366 | *Matt Caswell* | |
1367 | ||
1368 | * Removed the function names from error messages and deprecated the | |
1369 | xxx_F_xxx define's. | |
1370 | ||
1371 | * Removed NextStep support and the macro OPENSSL_UNISTD | |
1372 | ||
1373 | *Rich Salz* | |
1374 | ||
1375 | * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL, | |
1376 | OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL. | |
1377 | Also removed "export var as function" capability; we do not export | |
1378 | variables, only functions. | |
1379 | ||
1380 | *Rich Salz* | |
1381 | ||
1382 | * RC5_32_set_key has been changed to return an int type, with 0 indicating | |
1383 | an error and 1 indicating success. In previous versions of OpenSSL this | |
1384 | was a void type. If a key was set longer than the maximum possible this | |
1385 | would crash. | |
1386 | ||
1387 | *Matt Caswell* | |
1388 | ||
1389 | * Support SM2 signing and verification schemes with X509 certificate. | |
1390 | ||
1391 | *Paul Yang* | |
1392 | ||
1393 | * Use SHA256 as the default digest for TS query in the `ts` app. | |
1394 | ||
1395 | *Tomáš Mráz* | |
1396 | ||
1397 | * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898. | |
1398 | This checks that the salt length is at least 128 bits, the derived key | |
1399 | length is at least 112 bits, and that the iteration count is at least 1000. | |
1400 | For backwards compatibility these checks are disabled by default in the | |
1401 | default provider, but are enabled by default in the fips provider. | |
1402 | To enable or disable these checks use the control | |
1403 | EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE. | |
1404 | ||
1405 | *Shane Lontis* | |
1406 | ||
1407 | * Default cipher lists/suites are now available via a function, the | |
1408 | #defines are deprecated. | |
1409 | ||
1410 | *Todd Short* | |
1411 | ||
1412 | * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and | |
1413 | VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries | |
1414 | for Windows Store apps easier. Also, the "no-uplink" option has been added. | |
1415 | ||
1416 | *Kenji Mouri* | |
1417 | ||
1418 | * Join the directories crypto/x509 and crypto/x509v3 | |
1419 | ||
1420 | *Richard Levitte* | |
1421 | ||
1422 | * Added command 'openssl kdf' that uses the EVP_KDF API. | |
1423 | ||
1424 | *Shane Lontis* | |
1425 | ||
1426 | * Added command 'openssl mac' that uses the EVP_MAC API. | |
1427 | ||
1428 | *Shane Lontis* | |
1429 | ||
1430 | * Added OPENSSL_info() to get diverse built-in OpenSSL data, such | |
1431 | as default directories. Also added the command 'openssl info' | |
1432 | for scripting purposes. | |
1433 | ||
1434 | *Richard Levitte* | |
1435 | ||
1436 | * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been | |
1437 | deprecated. These undocumented functions were never integrated into the EVP | |
1438 | layer and implement the AES Infinite Garble Extension (IGE) mode and AES | |
1439 | Bi-directional IGE mode. These modes were never formally standardised and | |
1440 | usage of these functions is believed to be very small. In particular | |
1441 | AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one | |
1442 | is ever used. The security implications are believed to be minimal, but | |
1443 | this issue was never fixed for backwards compatibility reasons. New code | |
1444 | should not use these modes. | |
1445 | ||
1446 | *Matt Caswell* | |
1447 | ||
1448 | * Add prediction resistance to the DRBG reseeding process. | |
1449 | ||
1450 | *Paul Dale* | |
1451 | ||
1452 | * Limit the number of blocks in a data unit for AES-XTS to 2^20 as | |
1453 | mandated by IEEE Std 1619-2018. | |
1454 | ||
1455 | *Paul Dale* | |
1456 | ||
1457 | * Added newline escaping functionality to a filename when using openssl dgst. | |
1458 | This output format is to replicate the output format found in the `*sum` | |
1459 | checksum programs. This aims to preserve backward compatibility. | |
1460 | ||
1461 | *Matt Eaton, Richard Levitte, and Paul Dale* | |
1462 | ||
1463 | * Removed the heartbeat message in DTLS feature, as it has very | |
1464 | little usage and doesn't seem to fulfill a valuable purpose. | |
1465 | The configuration option is now deprecated. | |
1466 | ||
1467 | *Richard Levitte* | |
1468 | ||
1469 | * Changed the output of 'openssl {digestname} < file' to display the | |
1470 | digest name in its output. | |
1471 | ||
1472 | *Richard Levitte* | |
1473 | ||
1474 | * Added a new generic trace API which provides support for enabling | |
1475 | instrumentation through trace output. This feature is mainly intended | |
1476 | as an aid for developers and is disabled by default. To utilize it, | |
1477 | OpenSSL needs to be configured with the `enable-trace` option. | |
1478 | ||
1479 | If the tracing API is enabled, the application can activate trace output | |
1480 | by registering BIOs as trace channels for a number of tracing and debugging | |
1481 | categories. | |
1482 | ||
1483 | The `openssl` program has been expanded to enable any of the types | |
1484 | available via environment variables defined by the user, and serves as | |
1485 | one possible example on how to use this functionality. | |
1486 | ||
1487 | *Richard Levitte & Matthias St. Pierre* | |
1488 | ||
1489 | * Added build tests for C++. These are generated files that only do one | |
1490 | thing, to include one public OpenSSL head file each. This tests that | |
1491 | the public header files can be usefully included in a C++ application. | |
1492 | ||
1493 | This test isn't enabled by default. It can be enabled with the option | |
1494 | 'enable-buildtest-c++'. | |
1495 | ||
1496 | *Richard Levitte* | |
1497 | ||
1498 | * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF. | |
1499 | ||
1500 | *Shane Lontis* | |
1501 | ||
1502 | * Add KMAC to EVP_MAC. | |
1503 | ||
1504 | *Shane Lontis* | |
1505 | ||
1506 | * Added property based algorithm implementation selection framework to | |
1507 | the core. | |
1508 | ||
1509 | *Paul Dale* | |
1510 | ||
1511 | * Added SCA hardening for modular field inversion in EC_GROUP through | |
1512 | a new dedicated field_inv() pointer in EC_METHOD. | |
1513 | This also addresses a leakage affecting conversions from projective | |
1514 | to affine coordinates. | |
1515 | ||
1516 | *Billy Bob Brumley, Nicola Tuveri* | |
1517 | ||
1518 | * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF | |
1519 | implementations. This includes an EVP_PKEY to EVP_KDF bridge for | |
1520 | those algorithms that were already supported through the EVP_PKEY API | |
1521 | (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2 | |
1522 | and scrypt are now wrappers that call EVP_KDF. | |
1523 | ||
1524 | *David Makepeace* | |
1525 | ||
1526 | * Build devcrypto engine as a dynamic engine. | |
1527 | ||
1528 | *Eneas U de Queiroz* | |
1529 | ||
1530 | * Add keyed BLAKE2 to EVP_MAC. | |
1531 | ||
1532 | *Antoine Salon* | |
1533 | ||
1534 | * Fix a bug in the computation of the endpoint-pair shared secret used | |
1535 | by DTLS over SCTP. This breaks interoperability with older versions | |
1536 | of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime | |
1537 | switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling | |
1538 | interoperability with such broken implementations. However, enabling | |
1539 | this switch breaks interoperability with correct implementations. | |
1540 | ||
1541 | * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a | |
1542 | re-used X509_PUBKEY object if the second PUBKEY is malformed. | |
1543 | ||
1544 | *Bernd Edlinger* | |
1545 | ||
1546 | * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0(). | |
1547 | ||
1548 | *Richard Levitte* | |
1549 | ||
1550 | * Changed the license to the Apache License v2.0. | |
1551 | ||
1552 | *Richard Levitte* | |
1553 | ||
1554 | * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH. | |
1555 | ||
1556 | - Major releases (indicated by incrementing the MAJOR release number) | |
1557 | may introduce incompatible API/ABI changes. | |
1558 | - Minor releases (indicated by incrementing the MINOR release number) | |
1559 | may introduce new features but retain API/ABI compatibility. | |
1560 | - Patch releases (indicated by incrementing the PATCH number) | |
1561 | are intended for bug fixes and other improvements of existing | |
1562 | features only (like improving performance or adding documentation) | |
1563 | and retain API/ABI compatibility. | |
1564 | ||
1565 | *Richard Levitte* | |
1566 | ||
1567 | * Add support for RFC5297 SIV mode (siv128), including AES-SIV. | |
1568 | ||
1569 | *Todd Short* | |
1570 | ||
1571 | * Remove the 'dist' target and add a tarball building script. The | |
1572 | 'dist' target has fallen out of use, and it shouldn't be | |
1573 | necessary to configure just to create a source distribution. | |
1574 | ||
1575 | *Richard Levitte* | |
1576 | ||
1577 | * Recreate the OS390-Unix config target. It no longer relies on a | |
1578 | special script like it did for OpenSSL pre-1.1.0. | |
1579 | ||
1580 | *Richard Levitte* | |
1581 | ||
1582 | * Instead of having the source directories listed in Configure, add | |
1583 | a 'build.info' keyword SUBDIRS to indicate what sub-directories to | |
1584 | look into. | |
1585 | ||
1586 | *Richard Levitte* | |
1587 | ||
1588 | * Add GMAC to EVP_MAC. | |
1589 | ||
1590 | *Paul Dale* | |
1591 | ||
1592 | * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC. | |
1593 | ||
1594 | *Richard Levitte* | |
1595 | ||
1596 | * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC | |
1597 | implementations. This includes a generic EVP_PKEY to EVP_MAC bridge, | |
1598 | to facilitate the continued use of MACs through raw private keys in | |
1599 | functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`. | |
1600 | ||
1601 | *Richard Levitte* | |
1602 | ||
1603 | * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users | |
1604 | should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type). | |
1605 | ||
1606 | *Antoine Salon* | |
1607 | ||
1608 | * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for | |
1609 | the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names | |
1610 | are retained for backwards compatibility. | |
1611 | ||
1612 | *Antoine Salon* | |
1613 | ||
1614 | * AES-XTS mode now enforces that its two keys are different to mitigate | |
1615 | the attacked described in "Efficient Instantiations of Tweakable | |
1616 | Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway. | |
1617 | Details of this attack can be obtained from: | |
1618 | <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf> | |
1619 | ||
1620 | *Paul Dale* | |
1621 | ||
1622 | * Rename the object files, i.e. give them other names than in previous | |
1623 | versions. Their names now include the name of the final product, as | |
1624 | well as its type mnemonic (bin, lib, shlib). | |
1625 | ||
1626 | *Richard Levitte* | |
1627 | ||
1628 | * Added new option for 'openssl list', '-objects', which will display the | |
1629 | list of built in objects, i.e. OIDs with names. | |
1630 | ||
1631 | *Richard Levitte* | |
1632 | ||
1633 | * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`, | |
1634 | allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to | |
1635 | be set explicitly. | |
1636 | ||
1637 | *Chris Novakovic* | |
1638 | ||
1639 | * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path | |
1640 | improves application performance by removing data copies and providing | |
1641 | applications with zero-copy system calls such as sendfile and splice. | |
1642 | ||
1643 | *Boris Pismenny* | |
1644 | ||
1645 | * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that | |
1646 | option is set, openssl cleanses (zeroize) plaintext bytes from | |
1647 | internal buffers after delivering them to the application. Note, | |
1648 | the application is still responsible for cleansing other copies | |
1649 | (e.g.: data received by SSL_read(3)). | |
1650 | ||
1651 | *Martin Elshuber* | |
1652 | ||
1653 | * `PKCS12_parse` now maintains the order of the parsed certificates | |
1654 | when outputting them via `*ca` (rather than reversing it). | |
1655 | ||
1656 | *David von Oheimb* | |
1657 | ||
1658 | * Deprecated pthread fork support methods. These were unused so no | |
1659 | replacement is required. | |
1660 | ||
1661 | - OPENSSL_fork_prepare() | |
1662 | - OPENSSL_fork_parent() | |
1663 | - OPENSSL_fork_child() | |
1664 | ||
1665 | *Randall S. Becker* | |
1666 | ||
1667 | OpenSSL 1.1.1 | |
1668 | ------------- | |
1669 | ||
1670 | ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx] | |
1671 | ||
1672 | * Fixed a problem with verifying a certificate chain when using the | |
1673 | X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of | |
1674 | the certificates present in a certificate chain. It is not set by default. | |
1675 | ||
1676 | Starting from OpenSSL version 1.1.1h a check to disallow certificates in | |
1677 | the chain that have explicitly encoded elliptic curve parameters was added | |
1678 | as an additional strict check. | |
1679 | ||
1680 | An error in the implementation of this check meant that the result of a | |
1681 | previous check to confirm that certificates in the chain are valid CA | |
1682 | certificates was overwritten. This effectively bypasses the check | |
1683 | that non-CA certificates must not be able to issue other certificates. | |
1684 | ||
1685 | If a "purpose" has been configured then there is a subsequent opportunity | |
1686 | for checks that the certificate is a valid CA. All of the named "purpose" | |
1687 | values implemented in libcrypto perform this check. Therefore, where | |
1688 | a purpose is set the certificate chain will still be rejected even when the | |
1689 | strict flag has been used. A purpose is set by default in libssl client and | |
1690 | server certificate verification routines, but it can be overridden or | |
1691 | removed by an application. | |
1692 | ||
1693 | In order to be affected, an application must explicitly set the | |
1694 | X509_V_FLAG_X509_STRICT verification flag and either not set a purpose | |
1695 | for the certificate verification or, in the case of TLS client or server | |
1696 | applications, override the default purpose. | |
1697 | ([CVE-2021-3450]) | |
1698 | ||
1699 | *Tomáš Mráz* | |
1700 | ||
1701 | * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously | |
1702 | crafted renegotiation ClientHello message from a client. If a TLSv1.2 | |
1703 | renegotiation ClientHello omits the signature_algorithms extension (where it | |
1704 | was present in the initial ClientHello), but includes a | |
1705 | signature_algorithms_cert extension then a NULL pointer dereference will | |
1706 | result, leading to a crash and a denial of service attack. | |
1707 | ||
1708 | A server is only vulnerable if it has TLSv1.2 and renegotiation enabled | |
1709 | (which is the default configuration). OpenSSL TLS clients are not impacted by | |
1710 | this issue. | |
1711 | ([CVE-2021-3449]) | |
1712 | ||
1713 | *Peter Kästle and Samuel Sapalski* | |
1714 | ||
1715 | ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021] | |
1716 | ||
1717 | * Fixed the X509_issuer_and_serial_hash() function. It attempts to | |
1718 | create a unique hash value based on the issuer and serial number data | |
1719 | contained within an X509 certificate. However it was failing to correctly | |
1720 | handle any errors that may occur while parsing the issuer field (which might | |
1721 | occur if the issuer field is maliciously constructed). This may subsequently | |
1722 | result in a NULL pointer deref and a crash leading to a potential denial of | |
1723 | service attack. | |
1724 | ([CVE-2021-23841]) | |
1725 | ||
1726 | *Matt Caswell* | |
1727 | ||
1728 | * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING | |
1729 | padding mode to correctly check for rollback attacks. This is considered a | |
1730 | bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is | |
1731 | CVE-2021-23839. | |
1732 | ||
1733 | *Matt Caswell* | |
1734 | ||
1735 | Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate | |
1736 | functions. Previously they could overflow the output length argument in some | |
1737 | cases where the input length is close to the maximum permissable length for | |
1738 | an integer on the platform. In such cases the return value from the function | |
1739 | call would be 1 (indicating success), but the output length value would be | |
1740 | negative. This could cause applications to behave incorrectly or crash. | |
1741 | ([CVE-2021-23840]) | |
1742 | ||
1743 | *Matt Caswell* | |
1744 | ||
1745 | * Fixed SRP_Calc_client_key so that it runs in constant time. The previous | |
1746 | implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This | |
1747 | could be exploited in a side channel attack to recover the password. Since | |
1748 | the attack is local host only this is outside of the current OpenSSL | |
1749 | threat model and therefore no CVE is assigned. | |
1750 | ||
1751 | Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this | |
1752 | issue. | |
1753 | ||
1754 | *Matt Caswell* | |
1755 | ||
1756 | ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020] | |
1757 | ||
1758 | * Fixed NULL pointer deref in the GENERAL_NAME_cmp function | |
1759 | This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME. | |
1760 | If an attacker can control both items being compared then this could lead | |
1761 | to a possible denial of service attack. OpenSSL itself uses the | |
1762 | GENERAL_NAME_cmp function for two purposes: | |
1763 | 1) Comparing CRL distribution point names between an available CRL and a | |
1764 | CRL distribution point embedded in an X509 certificate | |
1765 | 2) When verifying that a timestamp response token signer matches the | |
1766 | timestamp authority name (exposed via the API functions | |
1767 | TS_RESP_verify_response and TS_RESP_verify_token) | |
1768 | ([CVE-2020-1971]) | |
1769 | ||
1770 | *Matt Caswell* | |
1771 | ||
1772 | ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020] | |
1773 | ||
1774 | * Certificates with explicit curve parameters are now disallowed in | |
1775 | verification chains if the X509_V_FLAG_X509_STRICT flag is used. | |
1776 | ||
1777 | *Tomáš Mráz* | |
1778 | ||
1779 | * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently | |
1780 | ignore TLS protocol version bounds when configuring DTLS-based contexts, and | |
1781 | conversely, silently ignore DTLS protocol version bounds when configuring | |
1782 | TLS-based contexts. The commands can be repeated to set bounds of both | |
1783 | types. The same applies with the corresponding "min_protocol" and | |
1784 | "max_protocol" command-line switches, in case some application uses both TLS | |
1785 | and DTLS. | |
1786 | ||
1787 | SSL_CTX instances that are created for a fixed protocol version (e.g. | |
1788 | TLSv1_server_method()) also silently ignore version bounds. Previously | |
1789 | attempts to apply bounds to these protocol versions would result in an | |
1790 | error. Now only the "version-flexible" SSL_CTX instances are subject to | |
1791 | limits in configuration files in command-line options. | |
1792 | ||
1793 | *Viktor Dukhovni* | |
1794 | ||
1795 | * Handshake now fails if Extended Master Secret extension is dropped | |
1796 | on renegotiation. | |
1797 | ||
1798 | *Tomáš Mráz* | |
1799 | ||
1800 | * The Oracle Developer Studio compiler will start reporting deprecated APIs | |
1801 | ||
1802 | ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020] | |
1803 | ||
1804 | * Fixed segmentation fault in SSL_check_chain() | |
1805 | Server or client applications that call the SSL_check_chain() function | |
1806 | during or after a TLS 1.3 handshake may crash due to a NULL pointer | |
1807 | dereference as a result of incorrect handling of the | |
1808 | "signature_algorithms_cert" TLS extension. The crash occurs if an invalid | |
1809 | or unrecognised signature algorithm is received from the peer. This could | |
1810 | be exploited by a malicious peer in a Denial of Service attack. | |
1811 | ([CVE-2020-1967]) | |
1812 | ||
1813 | *Benjamin Kaduk* | |
1814 | ||
1815 | * Added AES consttime code for no-asm configurations | |
1816 | an optional constant time support for AES was added | |
1817 | when building openssl for no-asm. | |
1818 | Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME | |
1819 | Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME | |
1820 | At this time this feature is by default disabled. | |
1821 | It will be enabled by default in 3.0. | |
1822 | ||
1823 | *Bernd Edlinger* | |
1824 | ||
1825 | ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020] | |
1826 | ||
1827 | * Revert the change of EOF detection while reading in libssl to avoid | |
1828 | regressions in applications depending on the current way of reporting | |
1829 | the EOF. As the existing method is not fully accurate the change to | |
1830 | reporting the EOF via SSL_ERROR_SSL is kept on the current development | |
1831 | branch and will be present in the 3.0 release. | |
1832 | ||
1833 | *Tomáš Mráz* | |
1834 | ||
1835 | * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1 | |
1836 | when primes for RSA keys are computed. | |
1837 | Since we previously always generated primes == 2 (mod 3) for RSA keys, | |
1838 | the 2-prime and 3-prime RSA modules were easy to distinguish, since | |
1839 | N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting | |
1840 | 2-prime vs. 3-prime RSA keys was possible by computing N mod 3. | |
1841 | This avoids possible fingerprinting of newly generated RSA modules. | |
1842 | ||
1843 | *Bernd Edlinger* | |
1844 | ||
1845 | ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020] | |
1846 | ||
1847 | * Properly detect EOF while reading in libssl. Previously if we hit an EOF | |
1848 | while reading in libssl then we would report an error back to the | |
1849 | application (SSL_ERROR_SYSCALL) but errno would be 0. We now add | |
1850 | an error to the stack (which means we instead return SSL_ERROR_SSL) and | |
1851 | therefore give a hint as to what went wrong. | |
1852 | ||
1853 | *Matt Caswell* | |
1854 | ||
1855 | * Check that ed25519 and ed448 are allowed by the security level. Previously | |
1856 | signature algorithms not using an MD were not being checked that they were | |
1857 | allowed by the security level. | |
1858 | ||
1859 | *Kurt Roeckx* | |
1860 | ||
1861 | * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername() | |
1862 | was not quite right. The behaviour was not consistent between resumption | |
1863 | and normal handshakes, and also not quite consistent with historical | |
1864 | behaviour. The behaviour in various scenarios has been clarified and | |
1865 | it has been updated to make it match historical behaviour as closely as | |
1866 | possible. | |
1867 | ||
1868 | *Matt Caswell* | |
1869 | ||
1870 | * *[VMS only]* The header files that the VMS compilers include automatically, | |
1871 | `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas | |
1872 | that the C++ compiler doesn't understand. This is a shortcoming in the | |
1873 | compiler, but can be worked around with `__cplusplus` guards. | |
1874 | ||
1875 | C++ applications that use OpenSSL libraries must be compiled using the | |
1876 | qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL | |
1877 | functions. Otherwise, only functions with symbols of less than 31 | |
1878 | characters can be used, as the linker will not be able to successfully | |
1879 | resolve symbols with longer names. | |
1880 | ||
1881 | *Richard Levitte* | |
1882 | ||
1883 | * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY. | |
1884 | The presence of this system service is determined at run-time. | |
1885 | ||
1886 | *Richard Levitte* | |
1887 | ||
1888 | * Added newline escaping functionality to a filename when using openssl dgst. | |
1889 | This output format is to replicate the output format found in the `*sum` | |
1890 | checksum programs. This aims to preserve backward compatibility. | |
1891 | ||
1892 | *Matt Eaton, Richard Levitte, and Paul Dale* | |
1893 | ||
1894 | * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just | |
1895 | the first value. | |
1896 | ||
1897 | *Jon Spillett* | |
1898 | ||
1899 | ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019] | |
1900 | ||
1901 | * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random | |
1902 | number generator (RNG). This was intended to include protection in the | |
1903 | event of a fork() system call in order to ensure that the parent and child | |
1904 | processes did not share the same RNG state. However this protection was not | |
1905 | being used in the default case. | |
1906 | ||
1907 | A partial mitigation for this issue is that the output from a high | |
1908 | precision timer is mixed into the RNG state so the likelihood of a parent | |
1909 | and child process sharing state is significantly reduced. | |
1910 | ||
1911 | If an application already calls OPENSSL_init_crypto() explicitly using | |
1912 | OPENSSL_INIT_ATFORK then this problem does not occur at all. | |
1913 | ([CVE-2019-1549]) | |
1914 | ||
1915 | *Matthias St. Pierre* | |
1916 | ||
1917 | * For built-in EC curves, ensure an EC_GROUP built from the curve name is | |
1918 | used even when parsing explicit parameters, when loading a encoded key | |
1919 | or calling `EC_GROUP_new_from_ecpkparameters()`/ | |
1920 | `EC_GROUP_new_from_ecparameters()`. | |
1921 | This prevents bypass of security hardening and performance gains, | |
1922 | especially for curves with specialized EC_METHODs. | |
1923 | By default, if a key encoded with explicit parameters is loaded and later | |
1924 | encoded, the output is still encoded with explicit parameters, even if | |
1925 | internally a "named" EC_GROUP is used for computation. | |
1926 | ||
1927 | *Nicola Tuveri* | |
1928 | ||
1929 | * Compute ECC cofactors if not provided during EC_GROUP construction. Before | |
1930 | this change, EC_GROUP_set_generator would accept order and/or cofactor as | |
1931 | NULL. After this change, only the cofactor parameter can be NULL. It also | |
1932 | does some minimal sanity checks on the passed order. | |
1933 | ([CVE-2019-1547]) | |
1934 | ||
1935 | *Billy Bob Brumley* | |
1936 | ||
1937 | * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey. | |
1938 | An attack is simple, if the first CMS_recipientInfo is valid but the | |
1939 | second CMS_recipientInfo is chosen ciphertext. If the second | |
1940 | recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct | |
1941 | encryption key will be replaced by garbage, and the message cannot be | |
1942 | decoded, but if the RSA decryption fails, the correct encryption key is | |
1943 | used and the recipient will not notice the attack. | |
1944 | As a work around for this potential attack the length of the decrypted | |
1945 | key must be equal to the cipher default key length, in case the | |
1946 | certifiate is not given and all recipientInfo are tried out. | |
1947 | The old behaviour can be re-enabled in the CMS code by setting the | |
1948 | CMS_DEBUG_DECRYPT flag. | |
1949 | ([CVE-2019-1563]) | |
1950 | ||
1951 | *Bernd Edlinger* | |
1952 | ||
1953 | * Early start up entropy quality from the DEVRANDOM seed source has been | |
1954 | improved for older Linux systems. The RAND subsystem will wait for | |
1955 | /dev/random to be producing output before seeding from /dev/urandom. | |
1956 | The seeded state is stored for future library initialisations using | |
1957 | a system global shared memory segment. The shared memory identifier | |
1958 | can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to | |
1959 | the desired value. The default identifier is 114. | |
1960 | ||
1961 | *Paul Dale* | |
1962 | ||
1963 | * Correct the extended master secret constant on EBCDIC systems. Without this | |
1964 | fix TLS connections between an EBCDIC system and a non-EBCDIC system that | |
1965 | negotiate EMS will fail. Unfortunately this also means that TLS connections | |
1966 | between EBCDIC systems with this fix, and EBCDIC systems without this | |
1967 | fix will fail if they negotiate EMS. | |
1968 | ||
1969 | *Matt Caswell* | |
1970 | ||
1971 | * Use Windows installation paths in the mingw builds | |
1972 | ||
1973 | Mingw isn't a POSIX environment per se, which means that Windows | |
1974 | paths should be used for installation. | |
1975 | ([CVE-2019-1552]) | |
1976 | ||
1977 | *Richard Levitte* | |
1978 | ||
1979 | * Changed DH_check to accept parameters with order q and 2q subgroups. | |
1980 | With order 2q subgroups the bit 0 of the private key is not secret | |
1981 | but DH_generate_key works around that by clearing bit 0 of the | |
1982 | private key for those. This avoids leaking bit 0 of the private key. | |
1983 | ||
1984 | *Bernd Edlinger* | |
1985 | ||
1986 | * Significantly reduce secure memory usage by the randomness pools. | |
1987 | ||
1988 | *Paul Dale* | |
1989 | ||
1990 | * Revert the DEVRANDOM_WAIT feature for Linux systems | |
1991 | ||
1992 | The DEVRANDOM_WAIT feature added a select() call to wait for the | |
1993 | /dev/random device to become readable before reading from the | |
1994 | /dev/urandom device. | |
1995 | ||
1996 | It turned out that this change had negative side effects on | |
1997 | performance which were not acceptable. After some discussion it | |
1998 | was decided to revert this feature and leave it up to the OS | |
1999 | resp. the platform maintainer to ensure a proper initialization | |
2000 | during early boot time. | |
2001 | ||
2002 | *Matthias St. Pierre* | |
2003 | ||
2004 | ### Changes between 1.1.1b and 1.1.1c [28 May 2019] | |
2005 | ||
2006 | * Add build tests for C++. These are generated files that only do one | |
2007 | thing, to include one public OpenSSL head file each. This tests that | |
2008 | the public header files can be usefully included in a C++ application. | |
2009 | ||
2010 | This test isn't enabled by default. It can be enabled with the option | |
2011 | 'enable-buildtest-c++'. | |
2012 | ||
2013 | *Richard Levitte* | |
2014 | ||
2015 | * Enable SHA3 pre-hashing for ECDSA and DSA. | |
2016 | ||
2017 | *Patrick Steuer* | |
2018 | ||
2019 | * Change the default RSA, DSA and DH size to 2048 bit instead of 1024. | |
2020 | This changes the size when using the `genpkey` command when no size is given. | |
2021 | It fixes an omission in earlier changes that changed all RSA, DSA and DH | |
2022 | generation commands to use 2048 bits by default. | |
2023 | ||
2024 | *Kurt Roeckx* | |
2025 | ||
2026 | * Reorganize the manual pages to consistently have RETURN VALUES, | |
2027 | EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust | |
2028 | util/fix-doc-nits accordingly. | |
2029 | ||
2030 | *Paul Yang, Joshua Lock* | |
2031 | ||
2032 | * Add the missing accessor EVP_PKEY_get0_engine() | |
2033 | ||
2034 | *Matt Caswell* | |
2035 | ||
2036 | * Have commands like `s_client` and `s_server` output the signature scheme | |
2037 | along with other cipher suite parameters when debugging. | |
2038 | ||
2039 | *Lorinczy Zsigmond* | |
2040 | ||
2041 | * Make OPENSSL_config() error agnostic again. | |
2042 | ||
2043 | *Richard Levitte* | |
2044 | ||
2045 | * Do the error handling in RSA decryption constant time. | |
2046 | ||
2047 | *Bernd Edlinger* | |
2048 | ||
2049 | * Prevent over long nonces in ChaCha20-Poly1305. | |
2050 | ||
2051 | ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input | |
2052 | for every encryption operation. RFC 7539 specifies that the nonce value | |
2053 | (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length | |
2054 | and front pads the nonce with 0 bytes if it is less than 12 | |
2055 | bytes. However it also incorrectly allows a nonce to be set of up to 16 | |
2056 | bytes. In this case only the last 12 bytes are significant and any | |
2057 | additional leading bytes are ignored. | |
2058 | ||
2059 | It is a requirement of using this cipher that nonce values are | |
2060 | unique. Messages encrypted using a reused nonce value are susceptible to | |
2061 | serious confidentiality and integrity attacks. If an application changes | |
2062 | the default nonce length to be longer than 12 bytes and then makes a | |
2063 | change to the leading bytes of the nonce expecting the new value to be a | |
2064 | new unique nonce then such an application could inadvertently encrypt | |
2065 | messages with a reused nonce. | |
2066 | ||
2067 | Additionally the ignored bytes in a long nonce are not covered by the | |
2068 | integrity guarantee of this cipher. Any application that relies on the | |
2069 | integrity of these ignored leading bytes of a long nonce may be further | |
2070 | affected. Any OpenSSL internal use of this cipher, including in SSL/TLS, | |
2071 | is safe because no such use sets such a long nonce value. However user | |
2072 | applications that use this cipher directly and set a non-default nonce | |
2073 | length to be longer than 12 bytes may be vulnerable. | |
2074 | ||
2075 | This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk | |
2076 | Greef of Ronomon. | |
2077 | ([CVE-2019-1543]) | |
2078 | ||
2079 | *Matt Caswell* | |
2080 | ||
2081 | * Add DEVRANDOM_WAIT feature for Linux systems | |
2082 | ||
2083 | On older Linux systems where the getrandom() system call is not available, | |
2084 | OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG. | |
2085 | Contrary to getrandom(), the /dev/urandom device will not block during | |
2086 | early boot when the kernel CSPRNG has not been seeded yet. | |
2087 | ||
2088 | To mitigate this known weakness, use select() to wait for /dev/random to | |
2089 | become readable before reading from /dev/urandom. | |
2090 | ||
2091 | * Ensure that SM2 only uses SM3 as digest algorithm | |
2092 | ||
2093 | *Paul Yang* | |
2094 | ||
2095 | ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019] | |
2096 | ||
2097 | * Change the info callback signals for the start and end of a post-handshake | |
2098 | message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START | |
2099 | and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get | |
2100 | confused by this and assume that a TLSv1.2 renegotiation has started. This | |
2101 | can break KeyUpdate handling. Instead we no longer signal the start and end | |
2102 | of a post handshake message exchange (although the messages themselves are | |
2103 | still signalled). This could break some applications that were expecting | |
2104 | the old signals. However without this KeyUpdate is not usable for many | |
2105 | applications. | |
2106 | ||
2107 | *Matt Caswell* | |
2108 | ||
2109 | ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018] | |
2110 | ||
2111 | * Timing vulnerability in DSA signature generation | |
2112 | ||
2113 | The OpenSSL DSA signature algorithm has been shown to be vulnerable to a | |
2114 | timing side channel attack. An attacker could use variations in the signing | |
2115 | algorithm to recover the private key. | |
2116 | ||
2117 | This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser. | |
2118 | ([CVE-2018-0734]) | |
2119 | ||
2120 | *Paul Dale* | |
2121 | ||
2122 | * Timing vulnerability in ECDSA signature generation | |
2123 | ||
2124 | The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a | |
2125 | timing side channel attack. An attacker could use variations in the signing | |
2126 | algorithm to recover the private key. | |
2127 | ||
2128 | This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser. | |
2129 | ([CVE-2018-0735]) | |
2130 | ||
2131 | *Paul Dale* | |
2132 | ||
2133 | * Fixed the issue that RAND_add()/RAND_seed() silently discards random input | |
2134 | if its length exceeds 4096 bytes. The limit has been raised to a buffer size | |
2135 | of two gigabytes and the error handling improved. | |
2136 | ||
2137 | This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been | |
2138 | categorized as a normal bug, not a security issue, because the DRBG reseeds | |
2139 | automatically and is fully functional even without additional randomness | |
2140 | provided by the application. | |
2141 | ||
2142 | ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018] | |
2143 | ||
2144 | * Add a new ClientHello callback. Provides a callback interface that gives | |
2145 | the application the ability to adjust the nascent SSL object at the | |
2146 | earliest stage of ClientHello processing, immediately after extensions have | |
2147 | been collected but before they have been processed. In particular, this | |
2148 | callback can adjust the supported TLS versions in response to the contents | |
2149 | of the ClientHello | |
2150 | ||
2151 | *Benjamin Kaduk* | |
2152 | ||
2153 | * Add SM2 base algorithm support. | |
2154 | ||
2155 | *Jack Lloyd* | |
2156 | ||
2157 | * s390x assembly pack: add (improved) hardware-support for the following | |
2158 | cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb, | |
2159 | aes-cfb/cfb8, aes-ecb. | |
2160 | ||
2161 | *Patrick Steuer* | |
2162 | ||
2163 | * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str | |
2164 | parameter is no longer accepted, as it leads to a corrupt table. NULL | |
2165 | pem_str is reserved for alias entries only. | |
2166 | ||
2167 | *Richard Levitte* | |
2168 | ||
2169 | * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder | |
2170 | step for prime curves. The new implementation is based on formulae from | |
2171 | differential addition-and-doubling in homogeneous projective coordinates | |
2172 | from Izu-Takagi "A fast parallel elliptic curve multiplication resistant | |
2173 | against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves | |
2174 | and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified | |
2175 | to work in projective coordinates. | |
2176 | ||
2177 | *Billy Bob Brumley, Nicola Tuveri* | |
2178 | ||
2179 | * Change generating and checking of primes so that the error rate of not | |
2180 | being prime depends on the intended use based on the size of the input. | |
2181 | For larger primes this will result in more rounds of Miller-Rabin. | |
2182 | The maximal error rate for primes with more than 1080 bits is lowered | |
2183 | to 2^-128. | |
2184 | ||
2185 | *Kurt Roeckx, Annie Yousar* | |
2186 | ||
2187 | * Increase the number of Miller-Rabin rounds for DSA key generating to 64. | |
2188 | ||
2189 | *Kurt Roeckx* | |
2190 | ||
2191 | * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when | |
2192 | moving between systems, and to avoid confusion when a Windows build is | |
2193 | done with mingw vs with MSVC. For POSIX installs, there's still a | |
2194 | symlink or copy named 'tsget' to avoid that confusion as well. | |
2195 | ||
2196 | *Richard Levitte* | |
2197 | ||
2198 | * Revert blinding in ECDSA sign and instead make problematic addition | |
2199 | length-invariant. Switch even to fixed-length Montgomery multiplication. | |
2200 | ||
2201 | *Andy Polyakov* | |
2202 | ||
2203 | * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder | |
2204 | step for binary curves. The new implementation is based on formulae from | |
2205 | differential addition-and-doubling in mixed Lopez-Dahab projective | |
2206 | coordinates, modified to independently blind the operands. | |
2207 | ||
2208 | *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri* | |
2209 | ||
2210 | * Add a scaffold to optionally enhance the Montgomery ladder implementation | |
2211 | for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing | |
2212 | EC_METHODs to implement their own specialized "ladder step", to take | |
2213 | advantage of more favorable coordinate systems or more efficient | |
2214 | differential addition-and-doubling algorithms. | |
2215 | ||
2216 | *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri* | |
2217 | ||
2218 | * Modified the random device based seed sources to keep the relevant | |
2219 | file descriptors open rather than reopening them on each access. | |
2220 | This allows such sources to operate in a chroot() jail without | |
2221 | the associated device nodes being available. This behaviour can be | |
2222 | controlled using RAND_keep_random_devices_open(). | |
2223 | ||
2224 | *Paul Dale* | |
2225 | ||
2226 | * Numerous side-channel attack mitigations have been applied. This may have | |
2227 | performance impacts for some algorithms for the benefit of improved | |
2228 | security. Specific changes are noted in this change log by their respective | |
2229 | authors. | |
2230 | ||
2231 | *Matt Caswell* | |
2232 | ||
2233 | * AIX shared library support overhaul. Switch to AIX "natural" way of | |
2234 | handling shared libraries, which means collecting shared objects of | |
2235 | different versions and bitnesses in one common archive. This allows to | |
2236 | mitigate conflict between 1.0 and 1.1 side-by-side installations. It | |
2237 | doesn't affect the way 3rd party applications are linked, only how | |
2238 | multi-version installation is managed. | |
2239 | ||
2240 | *Andy Polyakov* | |
2241 | ||
2242 | * Make ec_group_do_inverse_ord() more robust and available to other | |
2243 | EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA | |
2244 | mitigations are applied to the fallback BN_mod_inverse(). | |
2245 | When using this function rather than BN_mod_inverse() directly, new | |
2246 | EC cryptosystem implementations are then safer-by-default. | |
2247 | ||
2248 | *Billy Bob Brumley* | |
2249 | ||
2250 | * Add coordinate blinding for EC_POINT and implement projective | |
2251 | coordinate blinding for generic prime curves as a countermeasure to | |
2252 | chosen point SCA attacks. | |
2253 | ||
2254 | *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley* | |
2255 | ||
2256 | * Add blinding to ECDSA and DSA signatures to protect against side channel | |
2257 | attacks discovered by Keegan Ryan (NCC Group). | |
2258 | ||
2259 | *Matt Caswell* | |
2260 | ||
2261 | * Enforce checking in the `pkeyutl` command to ensure that the input | |
2262 | length does not exceed the maximum supported digest length when performing | |
2263 | a sign, verify or verifyrecover operation. | |
2264 | ||
2265 | *Matt Caswell* | |
2266 | ||
2267 | * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking | |
2268 | I/O in combination with something like select() or poll() will hang. This | |
2269 | can be turned off again using SSL_CTX_clear_mode(). | |
2270 | Many applications do not properly handle non-application data records, and | |
2271 | TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works | |
2272 | around the problems in those applications, but can also break some. | |
2273 | It's recommended to read the manpages about SSL_read(), SSL_write(), | |
2274 | SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and | |
2275 | SSL_CTX_set_read_ahead() again. | |
2276 | ||
2277 | *Kurt Roeckx* | |
2278 | ||
2279 | * When unlocking a pass phrase protected PEM file or PKCS#8 container, we | |
2280 | now allow empty (zero character) pass phrases. | |
2281 | ||
2282 | *Richard Levitte* | |
2283 | ||
2284 | * Apply blinding to binary field modular inversion and remove patent | |
2285 | pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation. | |
2286 | ||
2287 | *Billy Bob Brumley* | |
2288 | ||
2289 | * Deprecate ec2_mult.c and unify scalar multiplication code paths for | |
2290 | binary and prime elliptic curves. | |
2291 | ||
2292 | *Billy Bob Brumley* | |
2293 | ||
2294 | * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for | |
2295 | constant time fixed point multiplication. | |
2296 | ||
2297 | *Billy Bob Brumley* | |
2298 | ||
2299 | * Revise elliptic curve scalar multiplication with timing attack | |
2300 | defenses: ec_wNAF_mul redirects to a constant time implementation | |
2301 | when computing fixed point and variable point multiplication (which | |
2302 | in OpenSSL are mostly used with secret scalars in keygen, sign, | |
2303 | ECDH derive operations). | |
2304 | *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García, | |
2305 | Sohaib ul Hassan* | |
2306 | ||
2307 | * Updated CONTRIBUTING | |
2308 | ||
2309 | *Rich Salz* | |
2310 | ||
2311 | * Updated DRBG / RAND to request nonce and additional low entropy | |
2312 | randomness from the system. | |
2313 | ||
2314 | *Matthias St. Pierre* | |
2315 | ||
2316 | * Updated 'openssl rehash' to use OpenSSL consistent default. | |
2317 | ||
2318 | *Richard Levitte* | |
2319 | ||
2320 | * Moved the load of the ssl_conf module to libcrypto, which helps | |
2321 | loading engines that libssl uses before libssl is initialised. | |
2322 | ||
2323 | *Matt Caswell* | |
2324 | ||
2325 | * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA | |
2326 | ||
2327 | *Matt Caswell* | |
2328 | ||
2329 | * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases. | |
2330 | ||
2331 | *Ingo Schwarze, Rich Salz* | |
2332 | ||
2333 | * Added output of accepting IP address and port for 'openssl s_server' | |
2334 | ||
2335 | *Richard Levitte* | |
2336 | ||
2337 | * Added a new API for TLSv1.3 ciphersuites: | |
2338 | SSL_CTX_set_ciphersuites() | |
2339 | SSL_set_ciphersuites() | |
2340 | ||
2341 | *Matt Caswell* | |
2342 | ||
2343 | * Memory allocation failures consistently add an error to the error | |
2344 | stack. | |
2345 | ||
2346 | *Rich Salz* | |
2347 | ||
2348 | * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values | |
2349 | in libcrypto when run as setuid/setgid. | |
2350 | ||
2351 | *Bernd Edlinger* | |
2352 | ||
2353 | * Load any config file by default when libssl is used. | |
2354 | ||
2355 | *Matt Caswell* | |
2356 | ||
2357 | * Added new public header file <openssl/rand_drbg.h> and documentation | |
2358 | for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview. | |
2359 | ||
2360 | *Matthias St. Pierre* | |
2361 | ||
2362 | * QNX support removed (cannot find contributors to get their approval | |
2363 | for the license change). | |
2364 | ||
2365 | *Rich Salz* | |
2366 | ||
2367 | * TLSv1.3 replay protection for early data has been implemented. See the | |
2368 | SSL_read_early_data() man page for further details. | |
2369 | ||
2370 | *Matt Caswell* | |
2371 | ||
2372 | * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite | |
2373 | configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and | |
2374 | below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3. | |
2375 | In order to avoid issues where legacy TLSv1.2 ciphersuite configuration | |
2376 | would otherwise inadvertently disable all TLSv1.3 ciphersuites the | |
2377 | configuration has been separated out. See the ciphers man page or the | |
2378 | SSL_CTX_set_ciphersuites() man page for more information. | |
2379 | ||
2380 | *Matt Caswell* | |
2381 | ||
2382 | * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running | |
2383 | in responder mode now supports the new "-multi" option, which | |
2384 | spawns the specified number of child processes to handle OCSP | |
2385 | requests. The "-timeout" option now also limits the OCSP | |
2386 | responder's patience to wait to receive the full client request | |
2387 | on a newly accepted connection. Child processes are respawned | |
2388 | as needed, and the CA index file is automatically reloaded | |
2389 | when changed. This makes it possible to run the "ocsp" responder | |
2390 | as a long-running service, making the OpenSSL CA somewhat more | |
2391 | feature-complete. In this mode, most diagnostic messages logged | |
2392 | after entering the event loop are logged via syslog(3) rather than | |
2393 | written to stderr. | |
2394 | ||
2395 | *Viktor Dukhovni* | |
2396 | ||
2397 | * Added support for X448 and Ed448. Heavily based on original work by | |
2398 | Mike Hamburg. | |
2399 | ||
2400 | *Matt Caswell* | |
2401 | ||
2402 | * Extend OSSL_STORE with capabilities to search and to narrow the set of | |
2403 | objects loaded. This adds the functions OSSL_STORE_expect() and | |
2404 | OSSL_STORE_find() as well as needed tools to construct searches and | |
2405 | get the search data out of them. | |
2406 | ||
2407 | *Richard Levitte* | |
2408 | ||
2409 | * Support for TLSv1.3 added. Note that users upgrading from an earlier | |
2410 | version of OpenSSL should review their configuration settings to ensure | |
2411 | that they are still appropriate for TLSv1.3. For further information see: | |
2412 | <https://wiki.openssl.org/index.php/TLS1.3> | |
2413 | ||
2414 | *Matt Caswell* | |
2415 | ||
2416 | * Grand redesign of the OpenSSL random generator | |
2417 | ||
2418 | The default RAND method now utilizes an AES-CTR DRBG according to | |
2419 | NIST standard SP 800-90Ar1. The new random generator is essentially | |
2420 | a port of the default random generator from the OpenSSL FIPS 2.0 | |
2421 | object module. It is a hybrid deterministic random bit generator | |
2422 | using an AES-CTR bit stream and which seeds and reseeds itself | |
2423 | automatically using trusted system entropy sources. | |
2424 | ||
2425 | Some of its new features are: | |
2426 | - Support for multiple DRBG instances with seed chaining. | |
2427 | - The default RAND method makes use of a DRBG. | |
2428 | - There is a public and private DRBG instance. | |
2429 | - The DRBG instances are fork-safe. | |
2430 | - Keep all global DRBG instances on the secure heap if it is enabled. | |
2431 | - The public and private DRBG instance are per thread for lock free | |
2432 | operation | |
2433 | ||
2434 | *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre* | |
2435 | ||
2436 | * Changed Configure so it only says what it does and doesn't dump | |
2437 | so much data. Instead, ./configdata.pm should be used as a script | |
2438 | to display all sorts of configuration data. | |
2439 | ||
2440 | *Richard Levitte* | |
2441 | ||
2442 | * Added processing of "make variables" to Configure. | |
2443 | ||
2444 | *Richard Levitte* | |
2445 | ||
2446 | * Added SHA512/224 and SHA512/256 algorithm support. | |
2447 | ||
2448 | *Paul Dale* | |
2449 | ||
2450 | * The last traces of Netware support, first removed in 1.1.0, have | |
2451 | now been removed. | |
2452 | ||
2453 | *Rich Salz* | |
2454 | ||
2455 | * Get rid of Makefile.shared, and in the process, make the processing | |
2456 | of certain files (rc.obj, or the .def/.map/.opt files produced from | |
2457 | the ordinal files) more visible and hopefully easier to trace and | |
2458 | debug (or make silent). | |
2459 | ||
2460 | *Richard Levitte* | |
2461 | ||
2462 | * Make it possible to have environment variable assignments as | |
2463 | arguments to config / Configure. | |
2464 | ||
2465 | *Richard Levitte* | |
2466 | ||
2467 | * Add multi-prime RSA (RFC 8017) support. | |
2468 | ||
2469 | *Paul Yang* | |
2470 | ||
2471 | * Add SM3 implemented according to GB/T 32905-2016 | |
2472 | *Jack Lloyd <jack.lloyd@ribose.com>,* | |
2473 | *Ronald Tse <ronald.tse@ribose.com>,* | |
2474 | *Erick Borsboom <erick.borsboom@ribose.com>* | |
2475 | ||
2476 | * Add 'Maximum Fragment Length' TLS extension negotiation and support | |
2477 | as documented in RFC6066. | |
2478 | Based on a patch from Tomasz Moń | |
2479 | ||
2480 | *Filipe Raimundo da Silva* | |
2481 | ||
2482 | * Add SM4 implemented according to GB/T 32907-2016. | |
2483 | *Jack Lloyd <jack.lloyd@ribose.com>,* | |
2484 | *Ronald Tse <ronald.tse@ribose.com>,* | |
2485 | *Erick Borsboom <erick.borsboom@ribose.com>* | |
2486 | ||
2487 | * Reimplement -newreq-nodes and ERR_error_string_n; the | |
2488 | original author does not agree with the license change. | |
2489 | ||
2490 | *Rich Salz* | |
2491 | ||
2492 | * Add ARIA AEAD TLS support. | |
2493 | ||
2494 | *Jon Spillett* | |
2495 | ||
2496 | * Some macro definitions to support VS6 have been removed. Visual | |
2497 | Studio 6 has not worked since 1.1.0 | |
2498 | ||
2499 | *Rich Salz* | |
2500 | ||
2501 | * Add ERR_clear_last_mark(), to allow callers to clear the last mark | |
2502 | without clearing the errors. | |
2503 | ||
2504 | *Richard Levitte* | |
2505 | ||
2506 | * Add "atfork" functions. If building on a system that without | |
2507 | pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application | |
2508 | requirements. The RAND facility now uses/requires this. | |
2509 | ||
2510 | *Rich Salz* | |
2511 | ||
2512 | * Add SHA3. | |
2513 | ||
2514 | *Andy Polyakov* | |
2515 | ||
2516 | * The UI API becomes a permanent and integral part of libcrypto, i.e. | |
2517 | not possible to disable entirely. However, it's still possible to | |
2518 | disable the console reading UI method, UI_OpenSSL() (use UI_null() | |
2519 | as a fallback). | |
2520 | ||
2521 | To disable, configure with 'no-ui-console'. 'no-ui' is still | |
2522 | possible to use as an alias. Check at compile time with the | |
2523 | macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still | |
2524 | possible to check and is an alias for OPENSSL_NO_UI_CONSOLE. | |
2525 | ||
2526 | *Richard Levitte* | |
2527 | ||
2528 | * Add a STORE module, which implements a uniform and URI based reader of | |
2529 | stores that can contain keys, certificates, CRLs and numerous other | |
2530 | objects. The main API is loosely based on a few stdio functions, | |
2531 | and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof, | |
2532 | OSSL_STORE_error and OSSL_STORE_close. | |
2533 | The implementation uses backends called "loaders" to implement arbitrary | |
2534 | URI schemes. There is one built in "loader" for the 'file' scheme. | |
2535 | ||
2536 | *Richard Levitte* | |
2537 | ||
2538 | * Add devcrypto engine. This has been implemented against cryptodev-linux, | |
2539 | then adjusted to work on FreeBSD 8.4 as well. | |
2540 | Enable by configuring with 'enable-devcryptoeng'. This is done by default | |
2541 | on BSD implementations, as cryptodev.h is assumed to exist on all of them. | |
2542 | ||
2543 | *Richard Levitte* | |
2544 | ||
2545 | * Module names can prefixed with OSSL_ or OPENSSL_. This affects | |
2546 | util/mkerr.pl, which is adapted to allow those prefixes, leading to | |
2547 | error code calls like this: | |
2548 | ||
2549 | OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER); | |
2550 | ||
2551 | With this change, we claim the namespaces OSSL and OPENSSL in a manner | |
2552 | that can be encoded in C. For the foreseeable future, this will only | |
2553 | affect new modules. | |
2554 | ||
2555 | *Richard Levitte and Tim Hudson* | |
2556 | ||
2557 | * Removed BSD cryptodev engine. | |
2558 | ||
2559 | *Rich Salz* | |
2560 | ||
2561 | * Add a build target 'build_all_generated', to build all generated files | |
2562 | and only that. This can be used to prepare everything that requires | |
2563 | things like perl for a system that lacks perl and then move everything | |
2564 | to that system and do the rest of the build there. | |
2565 | ||
2566 | *Richard Levitte* | |
2567 | ||
2568 | * In the UI interface, make it possible to duplicate the user data. This | |
2569 | can be used by engines that need to retain the data for a longer time | |
2570 | than just the call where this user data is passed. | |
2571 | ||
2572 | *Richard Levitte* | |
2573 | ||
2574 | * Ignore the '-named_curve auto' value for compatibility of applications | |
2575 | with OpenSSL 1.0.2. | |
2576 | ||
2577 | *Tomáš Mráz <tmraz@fedoraproject.org>* | |
2578 | ||
2579 | * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2 | |
2580 | bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such | |
2581 | alerts across multiple records (some of which could be empty). In practice | |
2582 | it make no sense to send an empty alert record, or to fragment one. TLSv1.3 | |
2583 | prohibits this altogether and other libraries (BoringSSL, NSS) do not | |
2584 | support this at all. Supporting it adds significant complexity to the | |
2585 | record layer, and its removal is unlikely to cause interoperability | |
2586 | issues. | |
2587 | ||
2588 | *Matt Caswell* | |
2589 | ||
2590 | * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed | |
2591 | with Z. These are meant to replace LONG and ZLONG and to be size safe. | |
2592 | The use of LONG and ZLONG is discouraged and scheduled for deprecation | |
2593 | in OpenSSL 1.2.0. | |
2594 | ||
2595 | *Richard Levitte* | |
2596 | ||
2597 | * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string, | |
2598 | 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t. | |
2599 | ||
2600 | *Richard Levitte, Andy Polyakov* | |
2601 | ||
2602 | * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine() | |
2603 | does for RSA, etc. | |
2604 | ||
2605 | *Richard Levitte* | |
2606 | ||
2607 | * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target | |
2608 | platform rather than 'mingw'. | |
2609 | ||
2610 | *Richard Levitte* | |
2611 | ||
2612 | * The functions X509_STORE_add_cert and X509_STORE_add_crl return | |
2613 | success if they are asked to add an object which already exists | |
2614 | in the store. This change cascades to other functions which load | |
2615 | certificates and CRLs. | |
2616 | ||
2617 | *Paul Dale* | |
2618 | ||
2619 | * x86_64 assembly pack: annotate code with DWARF CFI directives to | |
2620 | facilitate stack unwinding even from assembly subroutines. | |
2621 | ||
2622 | *Andy Polyakov* | |
2623 | ||
2624 | * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN. | |
2625 | Also remove OPENSSL_GLOBAL entirely, as it became a no-op. | |
2626 | ||
2627 | *Richard Levitte* | |
2628 | ||
2629 | * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c. | |
2630 | VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1, | |
2631 | which is the minimum version we support. | |
2632 | ||
2633 | *Richard Levitte* | |
2634 | ||
2635 | * Certificate time validation (X509_cmp_time) enforces stricter | |
2636 | compliance with RFC 5280. Fractional seconds and timezone offsets | |
2637 | are no longer allowed. | |
2638 | ||
2639 | *Emilia Käsper* | |
2640 | ||
2641 | * Add support for ARIA | |
2642 | ||
2643 | *Paul Dale* | |
2644 | ||
2645 | * s_client will now send the Server Name Indication (SNI) extension by | |
2646 | default unless the new "-noservername" option is used. The server name is | |
2647 | based on the host provided to the "-connect" option unless overridden by | |
2648 | using "-servername". | |
2649 | ||
2650 | *Matt Caswell* | |
2651 | ||
2652 | * Add support for SipHash | |
2653 | ||
2654 | *Todd Short* | |
2655 | ||
2656 | * OpenSSL now fails if it receives an unrecognised record type in TLS1.0 | |
2657 | or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to | |
2658 | prevent issues where no progress is being made and the peer continually | |
2659 | sends unrecognised record types, using up resources processing them. | |
2660 | ||
2661 | *Matt Caswell* | |
2662 | ||
2663 | * 'openssl passwd' can now produce SHA256 and SHA512 based output, | |
2664 | using the algorithm defined in | |
2665 | <https://www.akkadia.org/drepper/SHA-crypt.txt> | |
2666 | ||
2667 | *Richard Levitte* | |
2668 | ||
2669 | * Heartbeat support has been removed; the ABI is changed for now. | |
2670 | ||
2671 | *Richard Levitte, Rich Salz* | |
2672 | ||
2673 | * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd. | |
2674 | ||
2675 | *Emilia Käsper* | |
2676 | ||
2677 | * The RSA "null" method, which was partially supported to avoid patent | |
2678 | issues, has been replaced to always returns NULL. | |
2679 | ||
2680 | *Rich Salz* | |
2681 | ||
2682 | OpenSSL 1.1.0 | |
2683 | ------------- | |
2684 | ||
2685 | ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019] | |
2686 | ||
2687 | * For built-in EC curves, ensure an EC_GROUP built from the curve name is | |
2688 | used even when parsing explicit parameters, when loading a encoded key | |
2689 | or calling `EC_GROUP_new_from_ecpkparameters()`/ | |
2690 | `EC_GROUP_new_from_ecparameters()`. | |
2691 | This prevents bypass of security hardening and performance gains, | |
2692 | especially for curves with specialized EC_METHODs. | |
2693 | By default, if a key encoded with explicit parameters is loaded and later | |
2694 | encoded, the output is still encoded with explicit parameters, even if | |
2695 | internally a "named" EC_GROUP is used for computation. | |
2696 | ||
2697 | *Nicola Tuveri* | |
2698 | ||
2699 | * Compute ECC cofactors if not provided during EC_GROUP construction. Before | |
2700 | this change, EC_GROUP_set_generator would accept order and/or cofactor as | |
2701 | NULL. After this change, only the cofactor parameter can be NULL. It also | |
2702 | does some minimal sanity checks on the passed order. | |
2703 | ([CVE-2019-1547]) | |
2704 | ||
2705 | *Billy Bob Brumley* | |
2706 | ||
2707 | * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey. | |
2708 | An attack is simple, if the first CMS_recipientInfo is valid but the | |
2709 | second CMS_recipientInfo is chosen ciphertext. If the second | |
2710 | recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct | |
2711 | encryption key will be replaced by garbage, and the message cannot be | |
2712 | decoded, but if the RSA decryption fails, the correct encryption key is | |
2713 | used and the recipient will not notice the attack. | |
2714 | As a work around for this potential attack the length of the decrypted | |
2715 | key must be equal to the cipher default key length, in case the | |
2716 | certifiate is not given and all recipientInfo are tried out. | |
2717 | The old behaviour can be re-enabled in the CMS code by setting the | |
2718 | CMS_DEBUG_DECRYPT flag. | |
2719 | ([CVE-2019-1563]) | |
2720 | ||
2721 | *Bernd Edlinger* | |
2722 | ||
2723 | * Use Windows installation paths in the mingw builds | |
2724 | ||
2725 | Mingw isn't a POSIX environment per se, which means that Windows | |
2726 | paths should be used for installation. | |
2727 | ([CVE-2019-1552]) | |
2728 | ||
2729 | *Richard Levitte* | |
2730 | ||
2731 | ### Changes between 1.1.0j and 1.1.0k [28 May 2019] | |
2732 | ||
2733 | * Change the default RSA, DSA and DH size to 2048 bit instead of 1024. | |
2734 | This changes the size when using the `genpkey` command when no size is given. | |
2735 | It fixes an omission in earlier changes that changed all RSA, DSA and DH | |
2736 | generation commands to use 2048 bits by default. | |
2737 | ||
2738 | *Kurt Roeckx* | |
2739 | ||
2740 | * Prevent over long nonces in ChaCha20-Poly1305. | |
2741 | ||
2742 | ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input | |
2743 | for every encryption operation. RFC 7539 specifies that the nonce value | |
2744 | (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length | |
2745 | and front pads the nonce with 0 bytes if it is less than 12 | |
2746 | bytes. However it also incorrectly allows a nonce to be set of up to 16 | |
2747 | bytes. In this case only the last 12 bytes are significant and any | |
2748 | additional leading bytes are ignored. | |
2749 | ||
2750 | It is a requirement of using this cipher that nonce values are | |
2751 | unique. Messages encrypted using a reused nonce value are susceptible to | |
2752 | serious confidentiality and integrity attacks. If an application changes | |
2753 | the default nonce length to be longer than 12 bytes and then makes a | |
2754 | change to the leading bytes of the nonce expecting the new value to be a | |
2755 | new unique nonce then such an application could inadvertently encrypt | |
2756 | messages with a reused nonce. | |
2757 | ||
2758 | Additionally the ignored bytes in a long nonce are not covered by the | |
2759 | integrity guarantee of this cipher. Any application that relies on the | |
2760 | integrity of these ignored leading bytes of a long nonce may be further | |
2761 | affected. Any OpenSSL internal use of this cipher, including in SSL/TLS, | |
2762 | is safe because no such use sets such a long nonce value. However user | |
2763 | applications that use this cipher directly and set a non-default nonce | |
2764 | length to be longer than 12 bytes may be vulnerable. | |
2765 | ||
2766 | This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk | |
2767 | Greef of Ronomon. | |
2768 | ([CVE-2019-1543]) | |
2769 | ||
2770 | *Matt Caswell* | |
2771 | ||
2772 | * Added SCA hardening for modular field inversion in EC_GROUP through | |
2773 | a new dedicated field_inv() pointer in EC_METHOD. | |
2774 | This also addresses a leakage affecting conversions from projective | |
2775 | to affine coordinates. | |
2776 | ||
2777 | *Billy Bob Brumley, Nicola Tuveri* | |
2778 | ||
2779 | * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a | |
2780 | re-used X509_PUBKEY object if the second PUBKEY is malformed. | |
2781 | ||
2782 | *Bernd Edlinger* | |
2783 | ||
2784 | * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0(). | |
2785 | ||
2786 | *Richard Levitte* | |
2787 | ||
2788 | * Remove the 'dist' target and add a tarball building script. The | |
2789 | 'dist' target has fallen out of use, and it shouldn't be | |
2790 | necessary to configure just to create a source distribution. | |
2791 | ||
2792 | *Richard Levitte* | |
2793 | ||
2794 | ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018] | |
2795 | ||
2796 | * Timing vulnerability in DSA signature generation | |
2797 | ||
2798 | The OpenSSL DSA signature algorithm has been shown to be vulnerable to a | |
2799 | timing side channel attack. An attacker could use variations in the signing | |
2800 | algorithm to recover the private key. | |
2801 | ||
2802 | This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser. | |
2803 | ([CVE-2018-0734]) | |
2804 | ||
2805 | *Paul Dale* | |
2806 | ||
2807 | * Timing vulnerability in ECDSA signature generation | |
2808 | ||
2809 | The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a | |
2810 | timing side channel attack. An attacker could use variations in the signing | |
2811 | algorithm to recover the private key. | |
2812 | ||
2813 | This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser. | |
2814 | ([CVE-2018-0735]) | |
2815 | ||
2816 | *Paul Dale* | |
2817 | ||
2818 | * Add coordinate blinding for EC_POINT and implement projective | |
2819 | coordinate blinding for generic prime curves as a countermeasure to | |
2820 | chosen point SCA attacks. | |
2821 | ||
2822 | *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley* | |
2823 | ||
2824 | ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018] | |
2825 | ||
2826 | * Client DoS due to large DH parameter | |
2827 | ||
2828 | During key agreement in a TLS handshake using a DH(E) based ciphersuite a | |
2829 | malicious server can send a very large prime value to the client. This will | |
2830 | cause the client to spend an unreasonably long period of time generating a | |
2831 | key for this prime resulting in a hang until the client has finished. This | |
2832 | could be exploited in a Denial Of Service attack. | |
2833 | ||
2834 | This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken | |
2835 | ([CVE-2018-0732]) | |
2836 | ||
2837 | *Guido Vranken* | |
2838 | ||
2839 | * Cache timing vulnerability in RSA Key Generation | |
2840 | ||
2841 | The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to | |
2842 | a cache timing side channel attack. An attacker with sufficient access to | |
2843 | mount cache timing attacks during the RSA key generation process could | |
2844 | recover the private key. | |
2845 | ||
2846 | This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera | |
2847 | Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia. | |
2848 | ([CVE-2018-0737]) | |
2849 | ||
2850 | *Billy Brumley* | |
2851 | ||
2852 | * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str | |
2853 | parameter is no longer accepted, as it leads to a corrupt table. NULL | |
2854 | pem_str is reserved for alias entries only. | |
2855 | ||
2856 | *Richard Levitte* | |
2857 | ||
2858 | * Revert blinding in ECDSA sign and instead make problematic addition | |
2859 | length-invariant. Switch even to fixed-length Montgomery multiplication. | |
2860 | ||
2861 | *Andy Polyakov* | |
2862 | ||
2863 | * Change generating and checking of primes so that the error rate of not | |
2864 | being prime depends on the intended use based on the size of the input. | |
2865 | For larger primes this will result in more rounds of Miller-Rabin. | |
2866 | The maximal error rate for primes with more than 1080 bits is lowered | |
2867 | to 2^-128. | |
2868 | ||
2869 | *Kurt Roeckx, Annie Yousar* | |
2870 | ||
2871 | * Increase the number of Miller-Rabin rounds for DSA key generating to 64. | |
2872 | ||
2873 | *Kurt Roeckx* | |
2874 | ||
2875 | * Add blinding to ECDSA and DSA signatures to protect against side channel | |
2876 | attacks discovered by Keegan Ryan (NCC Group). | |
2877 | ||
2878 | *Matt Caswell* | |
2879 | ||
2880 | * When unlocking a pass phrase protected PEM file or PKCS#8 container, we | |
2881 | now allow empty (zero character) pass phrases. | |
2882 | ||
2883 | *Richard Levitte* | |
2884 | ||
2885 | * Certificate time validation (X509_cmp_time) enforces stricter | |
2886 | compliance with RFC 5280. Fractional seconds and timezone offsets | |
2887 | are no longer allowed. | |
2888 | ||
2889 | *Emilia Käsper* | |
2890 | ||
2891 | * Fixed a text canonicalisation bug in CMS | |
2892 | ||
2893 | Where a CMS detached signature is used with text content the text goes | |
2894 | through a canonicalisation process first prior to signing or verifying a | |
2895 | signature. This process strips trailing space at the end of lines, converts | |
2896 | line terminators to CRLF and removes additional trailing line terminators | |
2897 | at the end of a file. A bug in the canonicalisation process meant that | |
2898 | some characters, such as form-feed, were incorrectly treated as whitespace | |
2899 | and removed. This is contrary to the specification (RFC5485). This fix | |
2900 | could mean that detached text data signed with an earlier version of | |
2901 | OpenSSL 1.1.0 may fail to verify using the fixed version, or text data | |
2902 | signed with a fixed OpenSSL may fail to verify with an earlier version of | |
2903 | OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data | |
2904 | and use the "-binary" flag (for the "cms" command line application) or set | |
2905 | the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()). | |
2906 | ||
2907 | *Matt Caswell* | |
2908 | ||
2909 | ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018] | |
2910 | ||
2911 | * Constructed ASN.1 types with a recursive definition could exceed the stack | |
2912 | ||
2913 | Constructed ASN.1 types with a recursive definition (such as can be found | |
2914 | in PKCS7) could eventually exceed the stack given malicious input with | |
2915 | excessive recursion. This could result in a Denial Of Service attack. There | |
2916 | are no such structures used within SSL/TLS that come from untrusted sources | |
2917 | so this is considered safe. | |
2918 | ||
2919 | This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz | |
2920 | project. | |
2921 | ([CVE-2018-0739]) | |
2922 | ||
2923 | *Matt Caswell* | |
2924 | ||
2925 | * Incorrect CRYPTO_memcmp on HP-UX PA-RISC | |
2926 | ||
2927 | Because of an implementation bug the PA-RISC CRYPTO_memcmp function is | |
2928 | effectively reduced to only comparing the least significant bit of each | |
2929 | byte. This allows an attacker to forge messages that would be considered as | |
2930 | authenticated in an amount of tries lower than that guaranteed by the | |
2931 | security claims of the scheme. The module can only be compiled by the | |
2932 | HP-UX assembler, so that only HP-UX PA-RISC targets are affected. | |
2933 | ||
2934 | This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg | |
2935 | (IBM). | |
2936 | ([CVE-2018-0733]) | |
2937 | ||
2938 | *Andy Polyakov* | |
2939 | ||
2940 | * Add a build target 'build_all_generated', to build all generated files | |
2941 | and only that. This can be used to prepare everything that requires | |
2942 | things like perl for a system that lacks perl and then move everything | |
2943 | to that system and do the rest of the build there. | |
2944 | ||
2945 | *Richard Levitte* | |
2946 | ||
2947 | * Backport SSL_OP_NO_RENGOTIATION | |
2948 | ||
2949 | OpenSSL 1.0.2 and below had the ability to disable renegotiation using the | |
2950 | (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity | |
2951 | changes this is no longer possible in 1.1.0. Therefore the new | |
2952 | SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to | |
2953 | 1.1.0 to provide equivalent functionality. | |
2954 | ||
2955 | Note that if an application built against 1.1.0h headers (or above) is run | |
2956 | using an older version of 1.1.0 (prior to 1.1.0h) then the option will be | |
2957 | accepted but nothing will happen, i.e. renegotiation will not be prevented. | |
2958 | ||
2959 | *Matt Caswell* | |
2960 | ||
2961 | * Removed the OS390-Unix config target. It relied on a script that doesn't | |
2962 | exist. | |
2963 | ||
2964 | *Rich Salz* | |
2965 | ||
2966 | * rsaz_1024_mul_avx2 overflow bug on x86_64 | |
2967 | ||
2968 | There is an overflow bug in the AVX2 Montgomery multiplication procedure | |
2969 | used in exponentiation with 1024-bit moduli. No EC algorithms are affected. | |
2970 | Analysis suggests that attacks against RSA and DSA as a result of this | |
2971 | defect would be very difficult to perform and are not believed likely. | |
2972 | Attacks against DH1024 are considered just feasible, because most of the | |
2973 | work necessary to deduce information about a private key may be performed | |
2974 | offline. The amount of resources required for such an attack would be | |
2975 | significant. However, for an attack on TLS to be meaningful, the server | |
2976 | would have to share the DH1024 private key among multiple clients, which is | |
2977 | no longer an option since CVE-2016-0701. | |
2978 | ||
2979 | This only affects processors that support the AVX2 but not ADX extensions | |
2980 | like Intel Haswell (4th generation). | |
2981 | ||
2982 | This issue was reported to OpenSSL by David Benjamin (Google). The issue | |
2983 | was originally found via the OSS-Fuzz project. | |
2984 | ([CVE-2017-3738]) | |
2985 | ||
2986 | *Andy Polyakov* | |
2987 | ||
2988 | ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017] | |
2989 | ||
2990 | * bn_sqrx8x_internal carry bug on x86_64 | |
2991 | ||
2992 | There is a carry propagating bug in the x86_64 Montgomery squaring | |
2993 | procedure. No EC algorithms are affected. Analysis suggests that attacks | |
2994 | against RSA and DSA as a result of this defect would be very difficult to | |
2995 | perform and are not believed likely. Attacks against DH are considered just | |
2996 | feasible (although very difficult) because most of the work necessary to | |
2997 | deduce information about a private key may be performed offline. The amount | |
2998 | of resources required for such an attack would be very significant and | |
2999 | likely only accessible to a limited number of attackers. An attacker would | |
3000 | additionally need online access to an unpatched system using the target | |
3001 | private key in a scenario with persistent DH parameters and a private | |
3002 | key that is shared between multiple clients. | |
3003 | ||
3004 | This only affects processors that support the BMI1, BMI2 and ADX extensions | |
3005 | like Intel Broadwell (5th generation) and later or AMD Ryzen. | |
3006 | ||
3007 | This issue was reported to OpenSSL by the OSS-Fuzz project. | |
3008 | ([CVE-2017-3736]) | |
3009 | ||
3010 | *Andy Polyakov* | |
3011 | ||
3012 | * Malformed X.509 IPAddressFamily could cause OOB read | |
3013 | ||
3014 | If an X.509 certificate has a malformed IPAddressFamily extension, | |
3015 | OpenSSL could do a one-byte buffer overread. The most likely result | |
3016 | would be an erroneous display of the certificate in text format. | |
3017 | ||
3018 | This issue was reported to OpenSSL by the OSS-Fuzz project. | |
3019 | ([CVE-2017-3735]) | |
3020 | ||
3021 | *Rich Salz* | |
3022 | ||
3023 | ### Changes between 1.1.0e and 1.1.0f [25 May 2017] | |
3024 | ||
3025 | * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target | |
3026 | platform rather than 'mingw'. | |
3027 | ||
3028 | *Richard Levitte* | |
3029 | ||
3030 | * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c. | |
3031 | VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1, | |
3032 | which is the minimum version we support. | |
3033 | ||
3034 | *Richard Levitte* | |
3035 | ||
3036 | ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017] | |
3037 | ||
3038 | * Encrypt-Then-Mac renegotiation crash | |
3039 | ||
3040 | During a renegotiation handshake if the Encrypt-Then-Mac extension is | |
3041 | negotiated where it was not in the original handshake (or vice-versa) then | |
3042 | this can cause OpenSSL to crash (dependant on ciphersuite). Both clients | |
3043 | and servers are affected. | |
3044 | ||
3045 | This issue was reported to OpenSSL by Joe Orton (Red Hat). | |
3046 | ([CVE-2017-3733]) | |
3047 | ||
3048 | *Matt Caswell* | |
3049 | ||
3050 | ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017] | |
3051 | ||
3052 | * Truncated packet could crash via OOB read | |
3053 | ||
3054 | If one side of an SSL/TLS path is running on a 32-bit host and a specific | |
3055 | cipher is being used, then a truncated packet can cause that host to | |
3056 | perform an out-of-bounds read, usually resulting in a crash. | |
3057 | ||
3058 | This issue was reported to OpenSSL by Robert Święcki of Google. | |
3059 | ([CVE-2017-3731]) | |
3060 | ||
3061 | *Andy Polyakov* | |
3062 | ||
3063 | * Bad (EC)DHE parameters cause a client crash | |
3064 | ||
3065 | If a malicious server supplies bad parameters for a DHE or ECDHE key | |
3066 | exchange then this can result in the client attempting to dereference a | |
3067 | NULL pointer leading to a client crash. This could be exploited in a Denial | |
3068 | of Service attack. | |
3069 | ||
3070 | This issue was reported to OpenSSL by Guido Vranken. | |
3071 | ([CVE-2017-3730]) | |
3072 | ||
3073 | *Matt Caswell* | |
3074 | ||
3075 | * BN_mod_exp may produce incorrect results on x86_64 | |
3076 | ||
3077 | There is a carry propagating bug in the x86_64 Montgomery squaring | |
3078 | procedure. No EC algorithms are affected. Analysis suggests that attacks | |
3079 | against RSA and DSA as a result of this defect would be very difficult to | |
3080 | perform and are not believed likely. Attacks against DH are considered just | |
3081 | feasible (although very difficult) because most of the work necessary to | |
3082 | deduce information about a private key may be performed offline. The amount | |
3083 | of resources required for such an attack would be very significant and | |
3084 | likely only accessible to a limited number of attackers. An attacker would | |
3085 | additionally need online access to an unpatched system using the target | |
3086 | private key in a scenario with persistent DH parameters and a private | |
3087 | key that is shared between multiple clients. For example this can occur by | |
3088 | default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very | |
3089 | similar to CVE-2015-3193 but must be treated as a separate problem. | |
3090 | ||
3091 | This issue was reported to OpenSSL by the OSS-Fuzz project. | |
3092 | ([CVE-2017-3732]) | |
3093 | ||
3094 | *Andy Polyakov* | |
3095 | ||
3096 | ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016] | |
3097 | ||
3098 | * ChaCha20/Poly1305 heap-buffer-overflow | |
3099 | ||
3100 | TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to | |
3101 | a DoS attack by corrupting larger payloads. This can result in an OpenSSL | |
3102 | crash. This issue is not considered to be exploitable beyond a DoS. | |
3103 | ||
3104 | This issue was reported to OpenSSL by Robert Święcki (Google Security Team) | |
3105 | ([CVE-2016-7054]) | |
3106 | ||
3107 | *Richard Levitte* | |
3108 | ||
3109 | * CMS Null dereference | |
3110 | ||
3111 | Applications parsing invalid CMS structures can crash with a NULL pointer | |
3112 | dereference. This is caused by a bug in the handling of the ASN.1 CHOICE | |
3113 | type in OpenSSL 1.1.0 which can result in a NULL value being passed to the | |
3114 | structure callback if an attempt is made to free certain invalid encodings. | |
3115 | Only CHOICE structures using a callback which do not handle NULL value are | |
3116 | affected. | |
3117 | ||
3118 | This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure. | |
3119 | ([CVE-2016-7053]) | |
3120 | ||
3121 | *Stephen Henson* | |
3122 | ||
3123 | * Montgomery multiplication may produce incorrect results | |
3124 | ||
3125 | There is a carry propagating bug in the Broadwell-specific Montgomery | |
3126 | multiplication procedure that handles input lengths divisible by, but | |
3127 | longer than 256 bits. Analysis suggests that attacks against RSA, DSA | |
3128 | and DH private keys are impossible. This is because the subroutine in | |
3129 | question is not used in operations with the private key itself and an input | |
3130 | of the attacker's direct choice. Otherwise the bug can manifest itself as | |
3131 | transient authentication and key negotiation failures or reproducible | |
3132 | erroneous outcome of public-key operations with specially crafted input. | |
3133 | Among EC algorithms only Brainpool P-512 curves are affected and one | |
3134 | presumably can attack ECDH key negotiation. Impact was not analyzed in | |
3135 | detail, because pre-requisites for attack are considered unlikely. Namely | |
3136 | multiple clients have to choose the curve in question and the server has to | |
3137 | share the private key among them, neither of which is default behaviour. | |
3138 | Even then only clients that chose the curve will be affected. | |
3139 | ||
3140 | This issue was publicly reported as transient failures and was not | |
3141 | initially recognized as a security issue. Thanks to Richard Morgan for | |
3142 | providing reproducible case. | |
3143 | ([CVE-2016-7055]) | |
3144 | ||
3145 | *Andy Polyakov* | |
3146 | ||
3147 | * Removed automatic addition of RPATH in shared libraries and executables, | |
3148 | as this was a remainder from OpenSSL 1.0.x and isn't needed any more. | |
3149 | ||
3150 | *Richard Levitte* | |
3151 | ||
3152 | ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016] | |
3153 | ||
3154 | * Fix Use After Free for large message sizes | |
3155 | ||
3156 | The patch applied to address CVE-2016-6307 resulted in an issue where if a | |
3157 | message larger than approx 16k is received then the underlying buffer to | |
3158 | store the incoming message is reallocated and moved. Unfortunately a | |
3159 | dangling pointer to the old location is left which results in an attempt to | |
3160 | write to the previously freed location. This is likely to result in a | |
3161 | crash, however it could potentially lead to execution of arbitrary code. | |
3162 | ||
3163 | This issue only affects OpenSSL 1.1.0a. | |
3164 | ||
3165 | This issue was reported to OpenSSL by Robert Święcki. | |
3166 | ([CVE-2016-6309]) | |
3167 | ||
3168 | *Matt Caswell* | |
3169 | ||
3170 | ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016] | |
3171 | ||
3172 | * OCSP Status Request extension unbounded memory growth | |
3173 | ||
3174 | A malicious client can send an excessively large OCSP Status Request | |
3175 | extension. If that client continually requests renegotiation, sending a | |
3176 | large OCSP Status Request extension each time, then there will be unbounded | |
3177 | memory growth on the server. This will eventually lead to a Denial Of | |
3178 | Service attack through memory exhaustion. Servers with a default | |
3179 | configuration are vulnerable even if they do not support OCSP. Builds using | |
3180 | the "no-ocsp" build time option are not affected. | |
3181 | ||
3182 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
3183 | ([CVE-2016-6304]) | |
3184 | ||
3185 | *Matt Caswell* | |
3186 | ||
3187 | * SSL_peek() hang on empty record | |
3188 | ||
3189 | OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer | |
3190 | sends an empty record. This could be exploited by a malicious peer in a | |
3191 | Denial Of Service attack. | |
3192 | ||
3193 | This issue was reported to OpenSSL by Alex Gaynor. | |
3194 | ([CVE-2016-6305]) | |
3195 | ||
3196 | *Matt Caswell* | |
3197 | ||
3198 | * Excessive allocation of memory in tls_get_message_header() and | |
3199 | dtls1_preprocess_fragment() | |
3200 | ||
3201 | A (D)TLS message includes 3 bytes for its length in the header for the | |
3202 | message. This would allow for messages up to 16Mb in length. Messages of | |
3203 | this length are excessive and OpenSSL includes a check to ensure that a | |
3204 | peer is sending reasonably sized messages in order to avoid too much memory | |
3205 | being consumed to service a connection. A flaw in the logic of version | |
3206 | 1.1.0 means that memory for the message is allocated too early, prior to | |
3207 | the excessive message length check. Due to way memory is allocated in | |
3208 | OpenSSL this could mean an attacker could force up to 21Mb to be allocated | |
3209 | to service a connection. This could lead to a Denial of Service through | |
3210 | memory exhaustion. However, the excessive message length check still takes | |
3211 | place, and this would cause the connection to immediately fail. Assuming | |
3212 | that the application calls SSL_free() on the failed connection in a timely | |
3213 | manner then the 21Mb of allocated memory will then be immediately freed | |
3214 | again. Therefore the excessive memory allocation will be transitory in | |
3215 | nature. This then means that there is only a security impact if: | |
3216 | ||
3217 | 1) The application does not call SSL_free() in a timely manner in the event | |
3218 | that the connection fails | |
3219 | or | |
3220 | 2) The application is working in a constrained environment where there is | |
3221 | very little free memory | |
3222 | or | |
3223 | 3) The attacker initiates multiple connection attempts such that there are | |
3224 | multiple connections in a state where memory has been allocated for the | |
3225 | connection; SSL_free() has not yet been called; and there is insufficient | |
3226 | memory to service the multiple requests. | |
3227 | ||
3228 | Except in the instance of (1) above any Denial Of Service is likely to be | |
3229 | transitory because as soon as the connection fails the memory is | |
3230 | subsequently freed again in the SSL_free() call. However there is an | |
3231 | increased risk during this period of application crashes due to the lack of | |
3232 | memory - which would then mean a more serious Denial of Service. | |
3233 | ||
3234 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
3235 | (CVE-2016-6307 and CVE-2016-6308) | |
3236 | ||
3237 | *Matt Caswell* | |
3238 | ||
3239 | * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler, | |
3240 | had to be removed. Primary reason is that vendor assembler can't | |
3241 | assemble our modules with -KPIC flag. As result it, assembly | |
3242 | support, was not even available as option. But its lack means | |
3243 | lack of side-channel resistant code, which is incompatible with | |
3244 | security by todays standards. Fortunately gcc is readily available | |
3245 | prepackaged option, which we firmly point at... | |
3246 | ||
3247 | *Andy Polyakov* | |
3248 | ||
3249 | ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016] | |
3250 | ||
3251 | * Windows command-line tool supports UTF-8 opt-in option for arguments | |
3252 | and console input. Setting OPENSSL_WIN32_UTF8 environment variable | |
3253 | (to any value) allows Windows user to access PKCS#12 file generated | |
3254 | with Windows CryptoAPI and protected with non-ASCII password, as well | |
3255 | as files generated under UTF-8 locale on Linux also protected with | |
3256 | non-ASCII password. | |
3257 | ||
3258 | *Andy Polyakov* | |
3259 | ||
3260 | * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites | |
3261 | have been disabled by default and removed from DEFAULT, just like RC4. | |
3262 | See the RC4 item below to re-enable both. | |
3263 | ||
3264 | *Rich Salz* | |
3265 | ||
3266 | * The method for finding the storage location for the Windows RAND seed file | |
3267 | has changed. First we check %RANDFILE%. If that is not set then we check | |
3268 | the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If | |
3269 | all else fails we fall back to C:\. | |
3270 | ||
3271 | *Matt Caswell* | |
3272 | ||
3273 | * The EVP_EncryptUpdate() function has had its return type changed from void | |
3274 | to int. A return of 0 indicates and error while a return of 1 indicates | |
3275 | success. | |
3276 | ||
3277 | *Matt Caswell* | |
3278 | ||
3279 | * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and | |
3280 | DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch | |
3281 | off the constant time implementation for RSA, DSA and DH have been made | |
3282 | no-ops and deprecated. | |
3283 | ||
3284 | *Matt Caswell* | |
3285 | ||
3286 | * Windows RAND implementation was simplified to only get entropy by | |
3287 | calling CryptGenRandom(). Various other RAND-related tickets | |
3288 | were also closed. | |
3289 | ||
3290 | *Joseph Wylie Yandle, Rich Salz* | |
3291 | ||
3292 | * The stack and lhash API's were renamed to start with `OPENSSL_SK_` | |
3293 | and `OPENSSL_LH_`, respectively. The old names are available | |
3294 | with API compatibility. They new names are now completely documented. | |
3295 | ||
3296 | *Rich Salz* | |
3297 | ||
3298 | * Unify TYPE_up_ref(obj) methods signature. | |
3299 | SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(), | |
3300 | X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an | |
3301 | int (instead of void) like all others TYPE_up_ref() methods. | |
3302 | So now these methods also check the return value of CRYPTO_atomic_add(), | |
3303 | and the validity of object reference counter. | |
3304 | ||
3305 | *fdasilvayy@gmail.com* | |
3306 | ||
3307 | * With Windows Visual Studio builds, the .pdb files are installed | |
3308 | alongside the installed libraries and executables. For a static | |
3309 | library installation, ossl_static.pdb is the associate compiler | |
3310 | generated .pdb file to be used when linking programs. | |
3311 | ||
3312 | *Richard Levitte* | |
3313 | ||
3314 | * Remove openssl.spec. Packaging files belong with the packagers. | |
3315 | ||
3316 | *Richard Levitte* | |
3317 | ||
3318 | * Automatic Darwin/OSX configuration has had a refresh, it will now | |
3319 | recognise x86_64 architectures automatically. You can still decide | |
3320 | to build for a different bitness with the environment variable | |
3321 | KERNEL_BITS (can be 32 or 64), for example: | |
3322 | ||
3323 | KERNEL_BITS=32 ./config | |
3324 | ||
3325 | *Richard Levitte* | |
3326 | ||
3327 | * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0, | |
3328 | 256 bit AES and HMAC with SHA256. | |
3329 | ||
3330 | *Steve Henson* | |
3331 | ||
3332 | * Remove support for MIPS o32 ABI on IRIX (and IRIX only). | |
3333 | ||
3334 | *Andy Polyakov* | |
3335 | ||
3336 | * Triple-DES ciphers have been moved from HIGH to MEDIUM. | |
3337 | ||
3338 | *Rich Salz* | |
3339 | ||
3340 | * To enable users to have their own config files and build file templates, | |
3341 | Configure looks in the directory indicated by the environment variable | |
3342 | OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/ | |
3343 | directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical | |
3344 | name and is used as is. | |
3345 | ||
3346 | *Richard Levitte* | |
3347 | ||
3348 | * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX, | |
3349 | X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type | |
3350 | X509_CERT_FILE_CTX was removed. | |
3351 | ||
3352 | *Rich Salz* | |
3353 | ||
3354 | * "shared" builds are now the default. To create only static libraries use | |
3355 | the "no-shared" Configure option. | |
3356 | ||
3357 | *Matt Caswell* | |
3358 | ||
3359 | * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options. | |
3360 | All of these option have not worked for some while and are fundamental | |
3361 | algorithms. | |
3362 | ||
3363 | *Matt Caswell* | |
3364 | ||
3365 | * Make various cleanup routines no-ops and mark them as deprecated. Most | |
3366 | global cleanup functions are no longer required because they are handled | |
3367 | via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages). | |
3368 | Explicitly de-initing can cause problems (e.g. where a library that uses | |
3369 | OpenSSL de-inits, but an application is still using it). The affected | |
3370 | functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(), | |
3371 | EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(), | |
3372 | RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and | |
3373 | COMP_zlib_cleanup(). | |
3374 | ||
3375 | *Matt Caswell* | |
3376 | ||
3377 | * --strict-warnings no longer enables runtime debugging options | |
3378 | such as REF_DEBUG. Instead, debug options are automatically | |
3379 | enabled with '--debug' builds. | |
3380 | ||
3381 | *Andy Polyakov, Emilia Käsper* | |
3382 | ||
3383 | * Made DH and DH_METHOD opaque. The structures for managing DH objects | |
3384 | have been moved out of the public header files. New functions for managing | |
3385 | these have been added. | |
3386 | ||
3387 | *Matt Caswell* | |
3388 | ||
3389 | * Made RSA and RSA_METHOD opaque. The structures for managing RSA | |
3390 | objects have been moved out of the public header files. New | |
3391 | functions for managing these have been added. | |
3392 | ||
3393 | *Richard Levitte* | |
3394 | ||
3395 | * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects | |
3396 | have been moved out of the public header files. New functions for managing | |
3397 | these have been added. | |
3398 | ||
3399 | *Matt Caswell* | |
3400 | ||
3401 | * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been | |
3402 | moved out of the public header files. New functions for managing these | |
3403 | have been added. | |
3404 | ||
3405 | *Matt Caswell* | |
3406 | ||
3407 | * Removed no-rijndael as a config option. Rijndael is an old name for AES. | |
3408 | ||
3409 | *Matt Caswell* | |
3410 | ||
3411 | * Removed the mk1mf build scripts. | |
3412 | ||
3413 | *Richard Levitte* | |
3414 | ||
3415 | * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so | |
3416 | it is always safe to #include a header now. | |
3417 | ||
3418 | *Rich Salz* | |
3419 | ||
3420 | * Removed the aged BC-32 config and all its supporting scripts | |
3421 | ||
3422 | *Richard Levitte* | |
3423 | ||
3424 | * Removed support for Ultrix, Netware, and OS/2. | |
3425 | ||
3426 | *Rich Salz* | |
3427 | ||
3428 | * Add support for HKDF. | |
3429 | ||
3430 | *Alessandro Ghedini* | |
3431 | ||
3432 | * Add support for blake2b and blake2s | |
3433 | ||
3434 | *Bill Cox* | |
3435 | ||
3436 | * Added support for "pipelining". Ciphers that have the | |
3437 | EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple | |
3438 | encryptions/decryptions simultaneously. There are currently no built-in | |
3439 | ciphers with this property but the expectation is that engines will be able | |
3440 | to offer it to significantly improve throughput. Support has been extended | |
3441 | into libssl so that multiple records for a single connection can be | |
3442 | processed in one go (for >=TLS 1.1). | |
3443 | ||
3444 | *Matt Caswell* | |
3445 | ||
3446 | * Added the AFALG engine. This is an async capable engine which is able to | |
3447 | offload work to the Linux kernel. In this initial version it only supports | |
3448 | AES128-CBC. The kernel must be version 4.1.0 or greater. | |
3449 | ||
3450 | *Catriona Lucey* | |
3451 | ||
3452 | * OpenSSL now uses a new threading API. It is no longer necessary to | |
3453 | set locking callbacks to use OpenSSL in a multi-threaded environment. There | |
3454 | are two supported threading models: pthreads and windows threads. It is | |
3455 | also possible to configure OpenSSL at compile time for "no-threads". The | |
3456 | old threading API should no longer be used. The functions have been | |
3457 | replaced with "no-op" compatibility macros. | |
3458 | ||
3459 | *Alessandro Ghedini, Matt Caswell* | |
3460 | ||
3461 | * Modify behavior of ALPN to invoke callback after SNI/servername | |
3462 | callback, such that updates to the SSL_CTX affect ALPN. | |
3463 | ||
3464 | *Todd Short* | |
3465 | ||
3466 | * Add SSL_CIPHER queries for authentication and key-exchange. | |
3467 | ||
3468 | *Todd Short* | |
3469 | ||
3470 | * Changes to the DEFAULT cipherlist: | |
3471 | - Prefer (EC)DHE handshakes over plain RSA. | |
3472 | - Prefer AEAD ciphers over legacy ciphers. | |
3473 | - Prefer ECDSA over RSA when both certificates are available. | |
3474 | - Prefer TLSv1.2 ciphers/PRF. | |
3475 | - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the | |
3476 | default cipherlist. | |
3477 | ||
3478 | *Emilia Käsper* | |
3479 | ||
3480 | * Change the ECC default curve list to be this, in order: x25519, | |
3481 | secp256r1, secp521r1, secp384r1. | |
3482 | ||
3483 | *Rich Salz* | |
3484 | ||
3485 | * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are | |
3486 | disabled by default. They can be re-enabled using the | |
3487 | enable-weak-ssl-ciphers option to Configure. | |
3488 | ||
3489 | *Matt Caswell* | |
3490 | ||
3491 | * If the server has ALPN configured, but supports no protocols that the | |
3492 | client advertises, send a fatal "no_application_protocol" alert. | |
3493 | This behaviour is SHALL in RFC 7301, though it isn't universally | |
3494 | implemented by other servers. | |
3495 | ||
3496 | *Emilia Käsper* | |
3497 | ||
3498 | * Add X25519 support. | |
3499 | Add ASN.1 and EVP_PKEY methods for X25519. This includes support | |
3500 | for public and private key encoding using the format documented in | |
3501 | draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports | |
3502 | key generation and key derivation. | |
3503 | ||
3504 | TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses | |
3505 | X25519(29). | |
3506 | ||
3507 | *Steve Henson* | |
3508 | ||
3509 | * Deprecate SRP_VBASE_get_by_user. | |
3510 | SRP_VBASE_get_by_user had inconsistent memory management behaviour. | |
3511 | In order to fix an unavoidable memory leak ([CVE-2016-0798]), | |
3512 | SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP | |
3513 | seed, even if the seed is configured. | |
3514 | ||
3515 | Users should use SRP_VBASE_get1_by_user instead. Note that in | |
3516 | SRP_VBASE_get1_by_user, caller must free the returned value. Note | |
3517 | also that even though configuring the SRP seed attempts to hide | |
3518 | invalid usernames by continuing the handshake with fake | |
3519 | credentials, this behaviour is not constant time and no strong | |
3520 | guarantees are made that the handshake is indistinguishable from | |
3521 | that of a valid user. | |
3522 | ||
3523 | *Emilia Käsper* | |
3524 | ||
3525 | * Configuration change; it's now possible to build dynamic engines | |
3526 | without having to build shared libraries and vice versa. This | |
3527 | only applies to the engines in `engines/`, those in `crypto/engine/` | |
3528 | will always be built into libcrypto (i.e. "static"). | |
3529 | ||
3530 | Building dynamic engines is enabled by default; to disable, use | |
3531 | the configuration option "disable-dynamic-engine". | |
3532 | ||
3533 | The only requirements for building dynamic engines are the | |
3534 | presence of the DSO module and building with position independent | |
3535 | code, so they will also automatically be disabled if configuring | |
3536 | with "disable-dso" or "disable-pic". | |
3537 | ||
3538 | The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE | |
3539 | are also taken away from openssl/opensslconf.h, as they are | |
3540 | irrelevant. | |
3541 | ||
3542 | *Richard Levitte* | |
3543 | ||
3544 | * Configuration change; if there is a known flag to compile | |
3545 | position independent code, it will always be applied on the | |
3546 | libcrypto and libssl object files, and never on the application | |
3547 | object files. This means other libraries that use routines from | |
3548 | libcrypto / libssl can be made into shared libraries regardless | |
3549 | of how OpenSSL was configured. | |
3550 | ||
3551 | If this isn't desirable, the configuration options "disable-pic" | |
3552 | or "no-pic" can be used to disable the use of PIC. This will | |
3553 | also disable building shared libraries and dynamic engines. | |
3554 | ||
3555 | *Richard Levitte* | |
3556 | ||
3557 | * Removed JPAKE code. It was experimental and has no wide use. | |
3558 | ||
3559 | *Rich Salz* | |
3560 | ||
3561 | * The INSTALL_PREFIX Makefile variable has been renamed to | |
3562 | DESTDIR. That makes for less confusion on what this variable | |
3563 | is for. Also, the configuration option --install_prefix is | |
3564 | removed. | |
3565 | ||
3566 | *Richard Levitte* | |
3567 | ||
3568 | * Heartbeat for TLS has been removed and is disabled by default | |
3569 | for DTLS; configure with enable-heartbeats. Code that uses the | |
3570 | old #define's might need to be updated. | |
3571 | ||
3572 | *Emilia Käsper, Rich Salz* | |
3573 | ||
3574 | * Rename REF_CHECK to REF_DEBUG. | |
3575 | ||
3576 | *Rich Salz* | |
3577 | ||
3578 | * New "unified" build system | |
3579 | ||
3580 | The "unified" build system is aimed to be a common system for all | |
3581 | platforms we support. With it comes new support for VMS. | |
3582 | ||
3583 | This system builds supports building in a different directory tree | |
3584 | than the source tree. It produces one Makefile (for unix family | |
3585 | or lookalikes), or one descrip.mms (for VMS). | |
3586 | ||
3587 | The source of information to make the Makefile / descrip.mms is | |
3588 | small files called 'build.info', holding the necessary | |
3589 | information for each directory with source to compile, and a | |
3590 | template in Configurations, like unix-Makefile.tmpl or | |
3591 | descrip.mms.tmpl. | |
3592 | ||
3593 | With this change, the library names were also renamed on Windows | |
3594 | and on VMS. They now have names that are closer to the standard | |
3595 | on Unix, and include the major version number, and in certain | |
3596 | cases, the architecture they are built for. See "Notes on shared | |
3597 | libraries" in INSTALL. | |
3598 | ||
3599 | We rely heavily on the perl module Text::Template. | |
3600 | ||
3601 | *Richard Levitte* | |
3602 | ||
3603 | * Added support for auto-initialisation and de-initialisation of the library. | |
3604 | OpenSSL no longer requires explicit init or deinit routines to be called, | |
3605 | except in certain circumstances. See the OPENSSL_init_crypto() and | |
3606 | OPENSSL_init_ssl() man pages for further information. | |
3607 | ||
3608 | *Matt Caswell* | |
3609 | ||
3610 | * The arguments to the DTLSv1_listen function have changed. Specifically the | |
3611 | "peer" argument is now expected to be a BIO_ADDR object. | |
3612 | ||
3613 | * Rewrite of BIO networking library. The BIO library lacked consistent | |
3614 | support of IPv6, and adding it required some more extensive | |
3615 | modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types, | |
3616 | which hold all types of addresses and chains of address information. | |
3617 | It also introduces a new API, with functions like BIO_socket, | |
3618 | BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept. | |
3619 | The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram | |
3620 | have been adapted accordingly. | |
3621 | ||
3622 | *Richard Levitte* | |
3623 | ||
3624 | * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without | |
3625 | the leading 0-byte. | |
3626 | ||
3627 | *Emilia Käsper* | |
3628 | ||
3629 | * CRIME protection: disable compression by default, even if OpenSSL is | |
3630 | compiled with zlib enabled. Applications can still enable compression | |
3631 | by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by | |
3632 | using the SSL_CONF library to configure compression. | |
3633 | ||
3634 | *Emilia Käsper* | |
3635 | ||
3636 | * The signature of the session callback configured with | |
3637 | SSL_CTX_sess_set_get_cb was changed. The read-only input buffer | |
3638 | was explicitly marked as `const unsigned char*` instead of | |
3639 | `unsigned char*`. | |
3640 | ||
3641 | *Emilia Käsper* | |
3642 | ||
3643 | * Always DPURIFY. Remove the use of uninitialized memory in the | |
3644 | RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op. | |
3645 | ||
3646 | *Emilia Käsper* | |
3647 | ||
3648 | * Removed many obsolete configuration items, including | |
3649 | DES_PTR, DES_RISC1, DES_RISC2, DES_INT | |
3650 | MD2_CHAR, MD2_INT, MD2_LONG | |
3651 | BF_PTR, BF_PTR2 | |
3652 | IDEA_SHORT, IDEA_LONG | |
3653 | RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX | |
3654 | ||
3655 | *Rich Salz, with advice from Andy Polyakov* | |
3656 | ||
3657 | * Many BN internals have been moved to an internal header file. | |
3658 | ||
3659 | *Rich Salz with help from Andy Polyakov* | |
3660 | ||
3661 | * Configuration and writing out the results from it has changed. | |
3662 | Files such as Makefile include/openssl/opensslconf.h and are now | |
3663 | produced through general templates, such as Makefile.in and | |
3664 | crypto/opensslconf.h.in and some help from the perl module | |
3665 | Text::Template. | |
3666 | ||
3667 | Also, the center of configuration information is no longer | |
3668 | Makefile. Instead, Configure produces a perl module in | |
3669 | configdata.pm which holds most of the config data (in the hash | |
3670 | table %config), the target data that comes from the target | |
3671 | configuration in one of the `Configurations/*.conf` files (in | |
3672 | %target). | |
3673 | ||
3674 | *Richard Levitte* | |
3675 | ||
3676 | * To clarify their intended purposes, the Configure options | |
3677 | --prefix and --openssldir change their semantics, and become more | |
3678 | straightforward and less interdependent. | |
3679 | ||
3680 | --prefix shall be used exclusively to give the location INSTALLTOP | |
3681 | where programs, scripts, libraries, include files and manuals are | |
3682 | going to be installed. The default is now /usr/local. | |
3683 | ||
3684 | --openssldir shall be used exclusively to give the default | |
3685 | location OPENSSLDIR where certificates, private keys, CRLs are | |
3686 | managed. This is also where the default openssl.cnf gets | |
3687 | installed. | |
3688 | If the directory given with this option is a relative path, the | |
3689 | values of both the --prefix value and the --openssldir value will | |
3690 | be combined to become OPENSSLDIR. | |
3691 | The default for --openssldir is INSTALLTOP/ssl. | |
3692 | ||
3693 | Anyone who uses --openssldir to specify where OpenSSL is to be | |
3694 | installed MUST change to use --prefix instead. | |
3695 | ||
3696 | *Richard Levitte* | |
3697 | ||
3698 | * The GOST engine was out of date and therefore it has been removed. An up | |
3699 | to date GOST engine is now being maintained in an external repository. | |
3700 | See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains | |
3701 | support for GOST ciphersuites (these are only activated if a GOST engine | |
3702 | is present). | |
3703 | ||
3704 | *Matt Caswell* | |
3705 | ||
3706 | * EGD is no longer supported by default; use enable-egd when | |
3707 | configuring. | |
3708 | ||
3709 | *Ben Kaduk and Rich Salz* | |
3710 | ||
3711 | * The distribution now has Makefile.in files, which are used to | |
3712 | create Makefile's when Configure is run. *Configure must be run | |
3713 | before trying to build now.* | |
3714 | ||
3715 | *Rich Salz* | |
3716 | ||
3717 | * The return value for SSL_CIPHER_description() for error conditions | |
3718 | has changed. | |
3719 | ||
3720 | *Rich Salz* | |
3721 | ||
3722 | * Support for RFC6698/RFC7671 DANE TLSA peer authentication. | |
3723 | ||
3724 | Obtaining and performing DNSSEC validation of TLSA records is | |
3725 | the application's responsibility. The application provides | |
3726 | the TLSA records of its choice to OpenSSL, and these are then | |
3727 | used to authenticate the peer. | |
3728 | ||
3729 | The TLSA records need not even come from DNS. They can, for | |
3730 | example, be used to implement local end-entity certificate or | |
3731 | trust-anchor "pinning", where the "pin" data takes the form | |
3732 | of TLSA records, which can augment or replace verification | |
3733 | based on the usual WebPKI public certification authorities. | |
3734 | ||
3735 | *Viktor Dukhovni* | |
3736 | ||
3737 | * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL | |
3738 | continues to support deprecated interfaces in default builds. | |
3739 | However, applications are strongly advised to compile their | |
3740 | source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides | |
3741 | the declarations of all interfaces deprecated in 0.9.8, 1.0.0 | |
3742 | or the 1.1.0 releases. | |
3743 | ||
3744 | In environments in which all applications have been ported to | |
3745 | not use any deprecated interfaces OpenSSL's Configure script | |
3746 | should be used with the --api=1.1.0 option to entirely remove | |
3747 | support for the deprecated features from the library and | |
3748 | unconditionally disable them in the installed headers. | |
3749 | Essentially the same effect can be achieved with the "no-deprecated" | |
3750 | argument to Configure, except that this will always restrict | |
3751 | the build to just the latest API, rather than a fixed API | |
3752 | version. | |
3753 | ||
3754 | As applications are ported to future revisions of the API, | |
3755 | they should update their compile-time OPENSSL_API_COMPAT define | |
3756 | accordingly, but in most cases should be able to continue to | |
3757 | compile with later releases. | |
3758 | ||
3759 | The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are | |
3760 | 0x10000000L and 0x00908000L, respectively. However those | |
3761 | versions did not support the OPENSSL_API_COMPAT feature, and | |
3762 | so applications are not typically tested for explicit support | |
3763 | of just the undeprecated features of either release. | |
3764 | ||
3765 | *Viktor Dukhovni* | |
3766 | ||
3767 | * Add support for setting the minimum and maximum supported protocol. | |
3768 | It can bet set via the SSL_set_min_proto_version() and | |
3769 | SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and | |
3770 | MaxProtocol. It's recommended to use the new APIs to disable | |
3771 | protocols instead of disabling individual protocols using | |
3772 | SSL_set_options() or SSL_CONF's Protocol. This change also | |
3773 | removes support for disabling TLS 1.2 in the OpenSSL TLS | |
3774 | client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT. | |
3775 | ||
3776 | *Kurt Roeckx* | |
3777 | ||
3778 | * Support for ChaCha20 and Poly1305 added to libcrypto and libssl. | |
3779 | ||
3780 | *Andy Polyakov* | |
3781 | ||
3782 | * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD | |
3783 | and integrates ECDSA and ECDH functionality into EC. Implementations can | |
3784 | now redirect key generation and no longer need to convert to or from | |
3785 | ECDSA_SIG format. | |
3786 | ||
3787 | Note: the ecdsa.h and ecdh.h headers are now no longer needed and just | |
3788 | include the ec.h header file instead. | |
3789 | ||
3790 | *Steve Henson* | |
3791 | ||
3792 | * Remove support for all 40 and 56 bit ciphers. This includes all the export | |
3793 | ciphers who are no longer supported and drops support the ephemeral RSA key | |
3794 | exchange. The LOW ciphers currently doesn't have any ciphers in it. | |
3795 | ||
3796 | *Kurt Roeckx* | |
3797 | ||
3798 | * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX | |
3799 | opaque. For HMAC_CTX, the following constructors and destructors | |
3800 | were added: | |
3801 | ||
3802 | HMAC_CTX *HMAC_CTX_new(void); | |
3803 | void HMAC_CTX_free(HMAC_CTX *ctx); | |
3804 | ||
3805 | For EVP_MD and EVP_CIPHER, complete APIs to create, fill and | |
3806 | destroy such methods has been added. See EVP_MD_meth_new(3) and | |
3807 | EVP_CIPHER_meth_new(3) for documentation. | |
3808 | ||
3809 | Additional changes: | |
3810 | 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and | |
3811 | `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and | |
3812 | `EVP_MD_CTX_reset()` should be called instead to reinitialise | |
3813 | an already created structure. | |
3814 | 2) For consistency with the majority of our object creators and | |
3815 | destructors, `EVP_MD_CTX_(create|destroy)` were renamed to | |
3816 | `EVP_MD_CTX_(new|free)`. The old names are retained as macros | |
3817 | for deprecated builds. | |
3818 | ||
3819 | *Richard Levitte* | |
3820 | ||
3821 | * Added ASYNC support. Libcrypto now includes the async sub-library to enable | |
3822 | cryptographic operations to be performed asynchronously as long as an | |
3823 | asynchronous capable engine is used. See the ASYNC_start_job() man page for | |
3824 | further details. Libssl has also had this capability integrated with the | |
3825 | introduction of the new mode SSL_MODE_ASYNC and associated error | |
3826 | SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man | |
3827 | pages. This work was developed in partnership with Intel Corp. | |
3828 | ||
3829 | *Matt Caswell* | |
3830 | ||
3831 | * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is | |
3832 | always enabled now. If you want to disable the support you should | |
3833 | exclude it using the list of supported ciphers. This also means that the | |
3834 | "-no_ecdhe" option has been removed from s_server. | |
3835 | ||
3836 | *Kurt Roeckx* | |
3837 | ||
3838 | * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls | |
3839 | SSL_{CTX_}set1_curves() which can set a list. | |
3840 | ||
3841 | *Kurt Roeckx* | |
3842 | ||
3843 | * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the | |
3844 | curve you want to support using SSL_{CTX_}set1_curves(). | |
3845 | ||
3846 | *Kurt Roeckx* | |
3847 | ||
3848 | * State machine rewrite. The state machine code has been significantly | |
3849 | refactored in order to remove much duplication of code and solve issues | |
3850 | with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for | |
3851 | further details). This change does have some associated API changes. | |
3852 | Notably the SSL_state() function has been removed and replaced by | |
3853 | SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int. | |
3854 | SSL_set_state() has been removed altogether. The previous handshake states | |
3855 | defined in ssl.h and ssl3.h have also been removed. | |
3856 | ||
3857 | *Matt Caswell* | |
3858 | ||
3859 | * All instances of the string "ssleay" in the public API were replaced | |
3860 | with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's) | |
3861 | Some error codes related to internal RSA_eay API's were renamed. | |
3862 | ||
3863 | *Rich Salz* | |
3864 | ||
3865 | * The demo files in crypto/threads were moved to demo/threads. | |
3866 | ||
3867 | *Rich Salz* | |
3868 | ||
3869 | * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp, | |
3870 | sureware and ubsec. | |
3871 | ||
3872 | *Matt Caswell, Rich Salz* | |
3873 | ||
3874 | * New ASN.1 embed macro. | |
3875 | ||
3876 | New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the | |
3877 | structure is not allocated: it is part of the parent. That is instead of | |
3878 | ||
3879 | FOO *x; | |
3880 | ||
3881 | it must be: | |
3882 | ||
3883 | FOO x; | |
3884 | ||
3885 | This reduces memory fragmentation and make it impossible to accidentally | |
3886 | set a mandatory field to NULL. | |
3887 | ||
3888 | This currently only works for some fields specifically a SEQUENCE, CHOICE, | |
3889 | or ASN1_STRING type which is part of a parent SEQUENCE. Since it is | |
3890 | equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or | |
3891 | SEQUENCE OF. | |
3892 | ||
3893 | *Steve Henson* | |
3894 | ||
3895 | * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled. | |
3896 | ||
3897 | *Emilia Käsper* | |
3898 | ||
3899 | * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although | |
3900 | in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also | |
3901 | an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add | |
3902 | DES and RC4 ciphersuites. | |
3903 | ||
3904 | *Matt Caswell* | |
3905 | ||
3906 | * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. | |
3907 | This changes the decoding behaviour for some invalid messages, | |
3908 | though the change is mostly in the more lenient direction, and | |
3909 | legacy behaviour is preserved as much as possible. | |
3910 | ||
3911 | *Emilia Käsper* | |
3912 | ||
3913 | * Fix no-stdio build. | |
3914 | *David Woodhouse <David.Woodhouse@intel.com> and also* | |
3915 | *Ivan Nestlerode <ivan.nestlerode@sonos.com>* | |
3916 | ||
3917 | * New testing framework | |
3918 | The testing framework has been largely rewritten and is now using | |
3919 | perl and the perl modules Test::Harness and an extended variant of | |
3920 | Test::More called OpenSSL::Test to do its work. All test scripts in | |
3921 | test/ have been rewritten into test recipes, and all direct calls to | |
3922 | executables in test/Makefile have become individual recipes using the | |
3923 | simplified testing OpenSSL::Test::Simple. | |
3924 | ||
3925 | For documentation on our testing modules, do: | |
3926 | ||
3927 | perldoc test/testlib/OpenSSL/Test/Simple.pm | |
3928 | perldoc test/testlib/OpenSSL/Test.pm | |
3929 | ||
3930 | *Richard Levitte* | |
3931 | ||
3932 | * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT | |
3933 | are used; the latter aborts on memory leaks (usually checked on exit). | |
3934 | Some undocumented "set malloc, etc., hooks" functions were removed | |
3935 | and others were changed. All are now documented. | |
3936 | ||
3937 | *Rich Salz* | |
3938 | ||
3939 | * In DSA_generate_parameters_ex, if the provided seed is too short, | |
3940 | return an error | |
3941 | ||
3942 | *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>* | |
3943 | ||
3944 | * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites | |
3945 | from RFC4279, RFC4785, RFC5487, RFC5489. | |
3946 | ||
3947 | Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the | |
3948 | original RSA_PSK patch. | |
3949 | ||
3950 | *Steve Henson* | |
3951 | ||
3952 | * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay | |
3953 | era flag was never set throughout the codebase (only read). Also removed | |
3954 | SSL3_FLAGS_POP_BUFFER which was only used if | |
3955 | SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set. | |
3956 | ||
3957 | *Matt Caswell* | |
3958 | ||
3959 | * Changed the default name options in the "ca", "crl", "req" and "x509" | |
3960 | to be "oneline" instead of "compat". | |
3961 | ||
3962 | *Richard Levitte* | |
3963 | ||
3964 | * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're | |
3965 | not aware of clients that still exhibit this bug, and the workaround | |
3966 | hasn't been working properly for a while. | |
3967 | ||
3968 | *Emilia Käsper* | |
3969 | ||
3970 | * The return type of BIO_number_read() and BIO_number_written() as well as | |
3971 | the corresponding num_read and num_write members in the BIO structure has | |
3972 | changed from unsigned long to uint64_t. On platforms where an unsigned | |
3973 | long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is | |
3974 | transferred. | |
3975 | ||
3976 | *Matt Caswell* | |
3977 | ||
3978 | * Given the pervasive nature of TLS extensions it is inadvisable to run | |
3979 | OpenSSL without support for them. It also means that maintaining | |
3980 | the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably | |
3981 | not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed. | |
3982 | ||
3983 | *Matt Caswell* | |
3984 | ||
3985 | * Removed support for the two export grade static DH ciphersuites | |
3986 | EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites | |
3987 | were newly added (along with a number of other static DH ciphersuites) to | |
3988 | 1.0.2. However the two export ones have *never* worked since they were | |
3989 | introduced. It seems strange in any case to be adding new export | |
3990 | ciphersuites, and given "logjam" it also does not seem correct to fix them. | |
3991 | ||
3992 | *Matt Caswell* | |
3993 | ||
3994 | * Version negotiation has been rewritten. In particular SSLv23_method(), | |
3995 | SSLv23_client_method() and SSLv23_server_method() have been deprecated, | |
3996 | and turned into macros which simply call the new preferred function names | |
3997 | TLS_method(), TLS_client_method() and TLS_server_method(). All new code | |
3998 | should use the new names instead. Also as part of this change the ssl23.h | |
3999 | header file has been removed. | |
4000 | ||
4001 | *Matt Caswell* | |
4002 | ||
4003 | * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This | |
4004 | code and the associated standard is no longer considered fit-for-purpose. | |
4005 | ||
4006 | *Matt Caswell* | |
4007 | ||
4008 | * RT2547 was closed. When generating a private key, try to make the | |
4009 | output file readable only by the owner. This behavior change might | |
4010 | be noticeable when interacting with other software. | |
4011 | ||
4012 | * Documented all exdata functions. Added CRYPTO_free_ex_index. | |
4013 | Added a test. | |
4014 | ||
4015 | *Rich Salz* | |
4016 | ||
4017 | * Added HTTP GET support to the ocsp command. | |
4018 | ||
4019 | *Rich Salz* | |
4020 | ||
4021 | * Changed default digest for the dgst and enc commands from MD5 to | |
4022 | sha256 | |
4023 | ||
4024 | *Rich Salz* | |
4025 | ||
4026 | * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead. | |
4027 | ||
4028 | *Matt Caswell* | |
4029 | ||
4030 | * Added support for TLS extended master secret from | |
4031 | draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an | |
4032 | initial patch which was a great help during development. | |
4033 | ||
4034 | *Steve Henson* | |
4035 | ||
4036 | * All libssl internal structures have been removed from the public header | |
4037 | files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is | |
4038 | now redundant). Users should not attempt to access internal structures | |
4039 | directly. Instead they should use the provided API functions. | |
4040 | ||
4041 | *Matt Caswell* | |
4042 | ||
4043 | * config has been changed so that by default OPENSSL_NO_DEPRECATED is used. | |
4044 | Access to deprecated functions can be re-enabled by running config with | |
4045 | "enable-deprecated". In addition applications wishing to use deprecated | |
4046 | functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour | |
4047 | will, by default, disable some transitive includes that previously existed | |
4048 | in the header files (e.g. ec.h will no longer, by default, include bn.h) | |
4049 | ||
4050 | *Matt Caswell* | |
4051 | ||
4052 | * Added support for OCB mode. OpenSSL has been granted a patent license | |
4053 | compatible with the OpenSSL license for use of OCB. Details are available | |
4054 | at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support | |
4055 | for OCB can be removed by calling config with no-ocb. | |
4056 | ||
4057 | *Matt Caswell* | |
4058 | ||
4059 | * SSLv2 support has been removed. It still supports receiving a SSLv2 | |
4060 | compatible client hello. | |
4061 | ||
4062 | *Kurt Roeckx* | |
4063 | ||
4064 | * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz], | |
4065 | done while fixing the error code for the key-too-small case. | |
4066 | ||
4067 | *Annie Yousar <a.yousar@informatik.hu-berlin.de>* | |
4068 | ||
4069 | * CA.sh has been removed; use CA.pl instead. | |
4070 | ||
4071 | *Rich Salz* | |
4072 | ||
4073 | * Removed old DES API. | |
4074 | ||
4075 | *Rich Salz* | |
4076 | ||
4077 | * Remove various unsupported platforms: | |
4078 | Sony NEWS4 | |
4079 | BEOS and BEOS_R5 | |
4080 | NeXT | |
4081 | SUNOS | |
4082 | MPE/iX | |
4083 | Sinix/ReliantUNIX RM400 | |
4084 | DGUX | |
4085 | NCR | |
4086 | Tandem | |
4087 | Cray | |
4088 | 16-bit platforms such as WIN16 | |
4089 | ||
4090 | *Rich Salz* | |
4091 | ||
4092 | * Clean up OPENSSL_NO_xxx #define's | |
4093 | - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF | |
4094 | - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx | |
4095 | - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC | |
4096 | - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160 | |
4097 | - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO | |
4098 | - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY | |
4099 | OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP | |
4100 | OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK | |
4101 | OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY | |
4102 | - Remove MS_STATIC; it's a relic from platforms <32 bits. | |
4103 | ||
4104 | *Rich Salz* | |
4105 | ||
4106 | * Cleaned up dead code | |
4107 | Remove all but one '#ifdef undef' which is to be looked at. | |
4108 | ||
4109 | *Rich Salz* | |
4110 | ||
4111 | * Clean up calling of xxx_free routines. | |
4112 | Just like free(), fix most of the xxx_free routines to accept | |
4113 | NULL. Remove the non-null checks from callers. Save much code. | |
4114 | ||
4115 | *Rich Salz* | |
4116 | ||
4117 | * Add secure heap for storage of private keys (when possible). | |
4118 | Add BIO_s_secmem(), CBIGNUM, etc. | |
4119 | Contributed by Akamai Technologies under our Corporate CLA. | |
4120 | ||
4121 | *Rich Salz* | |
4122 | ||
4123 | * Experimental support for a new, fast, unbiased prime candidate generator, | |
4124 | bn_probable_prime_dh_coprime(). Not currently used by any prime generator. | |
4125 | ||
4126 | *Felix Laurie von Massenbach <felix@erbridge.co.uk>* | |
4127 | ||
4128 | * New output format NSS in the sess_id command line tool. This allows | |
4129 | exporting the session id and the master key in NSS keylog format. | |
4130 | ||
4131 | *Martin Kaiser <martin@kaiser.cx>* | |
4132 | ||
4133 | * Harmonize version and its documentation. -f flag is used to display | |
4134 | compilation flags. | |
4135 | ||
4136 | *mancha <mancha1@zoho.com>* | |
4137 | ||
4138 | * Fix eckey_priv_encode so it immediately returns an error upon a failure | |
4139 | in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue. | |
4140 | ||
4141 | *mancha <mancha1@zoho.com>* | |
4142 | ||
4143 | * Fix some double frees. These are not thought to be exploitable. | |
4144 | ||
4145 | *mancha <mancha1@zoho.com>* | |
4146 | ||
4147 | * A missing bounds check in the handling of the TLS heartbeat extension | |
4148 | can be used to reveal up to 64k of memory to a connected client or | |
4149 | server. | |
4150 | ||
4151 | Thanks for Neel Mehta of Google Security for discovering this bug and to | |
4152 | Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for | |
4153 | preparing the fix ([CVE-2014-0160]) | |
4154 | ||
4155 | *Adam Langley, Bodo Moeller* | |
4156 | ||
4157 | * Fix for the attack described in the paper "Recovering OpenSSL | |
4158 | ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" | |
4159 | by Yuval Yarom and Naomi Benger. Details can be obtained from: | |
4160 | <http://eprint.iacr.org/2014/140> | |
4161 | ||
4162 | Thanks to Yuval Yarom and Naomi Benger for discovering this | |
4163 | flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076]) | |
4164 | ||
4165 | *Yuval Yarom and Naomi Benger* | |
4166 | ||
4167 | * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file(): | |
4168 | this fixes a limitation in previous versions of OpenSSL. | |
4169 | ||
4170 | *Steve Henson* | |
4171 | ||
4172 | * Experimental encrypt-then-mac support. | |
4173 | ||
4174 | Experimental support for encrypt then mac from | |
4175 | draft-gutmann-tls-encrypt-then-mac-02.txt | |
4176 | ||
4177 | To enable it set the appropriate extension number (0x42 for the test | |
4178 | server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42 | |
4179 | ||
4180 | For non-compliant peers (i.e. just about everything) this should have no | |
4181 | effect. | |
4182 | ||
4183 | WARNING: EXPERIMENTAL, SUBJECT TO CHANGE. | |
4184 | ||
4185 | *Steve Henson* | |
4186 | ||
4187 | * Add EVP support for key wrapping algorithms, to avoid problems with | |
4188 | existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in | |
4189 | the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap | |
4190 | algorithms and include tests cases. | |
4191 | ||
4192 | *Steve Henson* | |
4193 | ||
4194 | * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for | |
4195 | enveloped data. | |
4196 | ||
4197 | *Steve Henson* | |
4198 | ||
4199 | * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest, | |
4200 | MGF1 digest and OAEP label. | |
4201 | ||
4202 | *Steve Henson* | |
4203 | ||
4204 | * Make openssl verify return errors. | |
4205 | ||
4206 | *Chris Palmer <palmer@google.com> and Ben Laurie* | |
4207 | ||
4208 | * New function ASN1_TIME_diff to calculate the difference between two | |
4209 | ASN1_TIME structures or one structure and the current time. | |
4210 | ||
4211 | *Steve Henson* | |
4212 | ||
4213 | * Update fips_test_suite to support multiple command line options. New | |
4214 | test to induce all self test errors in sequence and check expected | |
4215 | failures. | |
4216 | ||
4217 | *Steve Henson* | |
4218 | ||
4219 | * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and | |
4220 | sign or verify all in one operation. | |
4221 | ||
4222 | *Steve Henson* | |
4223 | ||
4224 | * Add fips_algvs: a multicall fips utility incorporating all the algorithm | |
4225 | test programs and fips_test_suite. Includes functionality to parse | |
4226 | the minimal script output of fipsalgest.pl directly. | |
4227 | ||
4228 | *Steve Henson* | |
4229 | ||
4230 | * Add authorisation parameter to FIPS_module_mode_set(). | |
4231 | ||
4232 | *Steve Henson* | |
4233 | ||
4234 | * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves. | |
4235 | ||
4236 | *Steve Henson* | |
4237 | ||
4238 | * Use separate DRBG fields for internal and external flags. New function | |
4239 | FIPS_drbg_health_check() to perform on demand health checking. Add | |
4240 | generation tests to fips_test_suite with reduced health check interval to | |
4241 | demonstrate periodic health checking. Add "nodh" option to | |
4242 | fips_test_suite to skip very slow DH test. | |
4243 | ||
4244 | *Steve Henson* | |
4245 | ||
4246 | * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers | |
4247 | based on NID. | |
4248 | ||
4249 | *Steve Henson* | |
4250 | ||
4251 | * More extensive health check for DRBG checking many more failure modes. | |
4252 | New function FIPS_selftest_drbg_all() to handle every possible DRBG | |
4253 | combination: call this in fips_test_suite. | |
4254 | ||
4255 | *Steve Henson* | |
4256 | ||
4257 | * Add support for canonical generation of DSA parameter 'g'. See | |
4258 | FIPS 186-3 A.2.3. | |
4259 | ||
4260 | * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and | |
4261 | POST to handle HMAC cases. | |
4262 | ||
4263 | *Steve Henson* | |
4264 | ||
4265 | * Add functions FIPS_module_version() and FIPS_module_version_text() | |
4266 | to return numerical and string versions of the FIPS module number. | |
4267 | ||
4268 | *Steve Henson* | |
4269 | ||
4270 | * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and | |
4271 | FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented | |
4272 | outside the validated module in the FIPS capable OpenSSL. | |
4273 | ||
4274 | *Steve Henson* | |
4275 | ||
4276 | * Minor change to DRBG entropy callback semantics. In some cases | |
4277 | there is no multiple of the block length between min_len and | |
4278 | max_len. Allow the callback to return more than max_len bytes | |
4279 | of entropy but discard any extra: it is the callback's responsibility | |
4280 | to ensure that the extra data discarded does not impact the | |
4281 | requested amount of entropy. | |
4282 | ||
4283 | *Steve Henson* | |
4284 | ||
4285 | * Add PRNG security strength checks to RSA, DSA and ECDSA using | |
4286 | information in FIPS186-3, SP800-57 and SP800-131A. | |
4287 | ||
4288 | *Steve Henson* | |
4289 | ||
4290 | * CCM support via EVP. Interface is very similar to GCM case except we | |
4291 | must supply all data in one chunk (i.e. no update, final) and the | |
4292 | message length must be supplied if AAD is used. Add algorithm test | |
4293 | support. | |
4294 | ||
4295 | *Steve Henson* | |
4296 | ||
4297 | * Initial version of POST overhaul. Add POST callback to allow the status | |
4298 | of POST to be monitored and/or failures induced. Modify fips_test_suite | |
4299 | to use callback. Always run all selftests even if one fails. | |
4300 | ||
4301 | *Steve Henson* | |
4302 | ||
4303 | * XTS support including algorithm test driver in the fips_gcmtest program. | |
4304 | Note: this does increase the maximum key length from 32 to 64 bytes but | |
4305 | there should be no binary compatibility issues as existing applications | |
4306 | will never use XTS mode. | |
4307 | ||
4308 | *Steve Henson* | |
4309 | ||
4310 | * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies | |
4311 | to OpenSSL RAND code and replace with a tiny FIPS RAND API which also | |
4312 | performs algorithm blocking for unapproved PRNG types. Also do not | |
4313 | set PRNG type in FIPS_mode_set(): leave this to the application. | |
4314 | Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with | |
4315 | the standard OpenSSL PRNG: set additional data to a date time vector. | |
4316 | ||
4317 | *Steve Henson* | |
4318 | ||
4319 | * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`. | |
4320 | This shouldn't present any incompatibility problems because applications | |
4321 | shouldn't be using these directly and any that are will need to rethink | |
4322 | anyway as the X9.31 PRNG is now deprecated by FIPS 140-2 | |
4323 | ||
4324 | *Steve Henson* | |
4325 | ||
4326 | * Extensive self tests and health checking required by SP800-90 DRBG. | |
4327 | Remove strength parameter from FIPS_drbg_instantiate and always | |
4328 | instantiate at maximum supported strength. | |
4329 | ||
4330 | *Steve Henson* | |
4331 | ||
4332 | * Add ECDH code to fips module and fips_ecdhvs for primitives only testing. | |
4333 | ||
4334 | *Steve Henson* | |
4335 | ||
4336 | * New algorithm test program fips_dhvs to handle DH primitives only testing. | |
4337 | ||
4338 | *Steve Henson* | |
4339 | ||
4340 | * New function DH_compute_key_padded() to compute a DH key and pad with | |
4341 | leading zeroes if needed: this complies with SP800-56A et al. | |
4342 | ||
4343 | *Steve Henson* | |
4344 | ||
4345 | * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by | |
4346 | anything, incomplete, subject to change and largely untested at present. | |
4347 | ||
4348 | *Steve Henson* | |
4349 | ||
4350 | * Modify fipscanisteronly build option to only build the necessary object | |
4351 | files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile. | |
4352 | ||
4353 | *Steve Henson* | |
4354 | ||
4355 | * Add experimental option FIPSSYMS to give all symbols in | |
4356 | fipscanister.o and FIPS or fips prefix. This will avoid | |
4357 | conflicts with future versions of OpenSSL. Add perl script | |
4358 | util/fipsas.pl to preprocess assembly language source files | |
4359 | and rename any affected symbols. | |
4360 | ||
4361 | *Steve Henson* | |
4362 | ||
4363 | * Add selftest checks and algorithm block of non-fips algorithms in | |
4364 | FIPS mode. Remove DES2 from selftests. | |
4365 | ||
4366 | *Steve Henson* | |
4367 | ||
4368 | * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just | |
4369 | return internal method without any ENGINE dependencies. Add new | |
4370 | tiny fips sign and verify functions. | |
4371 | ||
4372 | *Steve Henson* | |
4373 | ||
4374 | * New build option no-ec2m to disable characteristic 2 code. | |
4375 | ||
4376 | *Steve Henson* | |
4377 | ||
4378 | * New build option "fipscanisteronly". This only builds fipscanister.o | |
4379 | and (currently) associated fips utilities. Uses the file Makefile.fips | |
4380 | instead of Makefile.org as the prototype. | |
4381 | ||
4382 | *Steve Henson* | |
4383 | ||
4384 | * Add some FIPS mode restrictions to GCM. Add internal IV generator. | |
4385 | Update fips_gcmtest to use IV generator. | |
4386 | ||
4387 | *Steve Henson* | |
4388 | ||
4389 | * Initial, experimental EVP support for AES-GCM. AAD can be input by | |
4390 | setting output buffer to NULL. The `*Final` function must be | |
4391 | called although it will not retrieve any additional data. The tag | |
4392 | can be set or retrieved with a ctrl. The IV length is by default 12 | |
4393 | bytes (96 bits) but can be set to an alternative value. If the IV | |
4394 | length exceeds the maximum IV length (currently 16 bytes) it cannot be | |
4395 | set before the key. | |
4396 | ||
4397 | *Steve Henson* | |
4398 | ||
4399 | * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the | |
4400 | underlying do_cipher function handles all cipher semantics itself | |
4401 | including padding and finalisation. This is useful if (for example) | |
4402 | an ENGINE cipher handles block padding itself. The behaviour of | |
4403 | do_cipher is subtly changed if this flag is set: the return value | |
4404 | is the number of characters written to the output buffer (zero is | |
4405 | no longer an error code) or a negative error code. Also if the | |
4406 | input buffer is NULL and length 0 finalisation should be performed. | |
4407 | ||
4408 | *Steve Henson* | |
4409 | ||
4410 | * If a candidate issuer certificate is already part of the constructed | |
4411 | path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case. | |
4412 | ||
4413 | *Steve Henson* | |
4414 | ||
4415 | * Improve forward-security support: add functions | |
4416 | ||
4417 | void SSL_CTX_set_not_resumable_session_callback( | |
4418 | SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure)) | |
4419 | void SSL_set_not_resumable_session_callback( | |
4420 | SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure)) | |
4421 | ||
4422 | for use by SSL/TLS servers; the callback function will be called whenever a | |
4423 | new session is created, and gets to decide whether the session may be | |
4424 | cached to make it resumable (return 0) or not (return 1). (As by the | |
4425 | SSL/TLS protocol specifications, the session_id sent by the server will be | |
4426 | empty to indicate that the session is not resumable; also, the server will | |
4427 | not generate RFC 4507 (RFC 5077) session tickets.) | |
4428 | ||
4429 | A simple reasonable callback implementation is to return is_forward_secure. | |
4430 | This parameter will be set to 1 or 0 depending on the ciphersuite selected | |
4431 | by the SSL/TLS server library, indicating whether it can provide forward | |
4432 | security. | |
4433 | ||
4434 | *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)* | |
4435 | ||
4436 | * New -verify_name option in command line utilities to set verification | |
4437 | parameters by name. | |
4438 | ||
4439 | *Steve Henson* | |
4440 | ||
4441 | * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE. | |
4442 | Add CMAC pkey methods. | |
4443 | ||
4444 | *Steve Henson* | |
4445 | ||
4446 | * Experimental renegotiation in s_server -www mode. If the client | |
4447 | browses /reneg connection is renegotiated. If /renegcert it is | |
4448 | renegotiated requesting a certificate. | |
4449 | ||
4450 | *Steve Henson* | |
4451 | ||
4452 | * Add an "external" session cache for debugging purposes to s_server. This | |
4453 | should help trace issues which normally are only apparent in deployed | |
4454 | multi-process servers. | |
4455 | ||
4456 | *Steve Henson* | |
4457 | ||
4458 | * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where | |
4459 | return value is ignored. NB. The functions RAND_add(), RAND_seed(), | |
4460 | BIO_set_cipher() and some obscure PEM functions were changed so they | |
4461 | can now return an error. The RAND changes required a change to the | |
4462 | RAND_METHOD structure. | |
4463 | ||
4464 | *Steve Henson* | |
4465 | ||
4466 | * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of | |
4467 | a gcc attribute to warn if the result of a function is ignored. This | |
4468 | is enable if DEBUG_UNUSED is set. Add to several functions in evp.h | |
4469 | whose return value is often ignored. | |
4470 | ||
4471 | *Steve Henson* | |
4472 | ||
4473 | * New -noct, -requestct, -requirect and -ctlogfile options for s_client. | |
4474 | These allow SCTs (signed certificate timestamps) to be requested and | |
4475 | validated when establishing a connection. | |
4476 | ||
4477 | *Rob Percival <robpercival@google.com>* | |
4478 | ||
4479 | OpenSSL 1.0.2 | |
4480 | ------------- | |
4481 | ||
4482 | ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019] | |
4483 | ||
4484 | * For built-in EC curves, ensure an EC_GROUP built from the curve name is | |
4485 | used even when parsing explicit parameters, when loading a encoded key | |
4486 | or calling `EC_GROUP_new_from_ecpkparameters()`/ | |
4487 | `EC_GROUP_new_from_ecparameters()`. | |
4488 | This prevents bypass of security hardening and performance gains, | |
4489 | especially for curves with specialized EC_METHODs. | |
4490 | By default, if a key encoded with explicit parameters is loaded and later | |
4491 | encoded, the output is still encoded with explicit parameters, even if | |
4492 | internally a "named" EC_GROUP is used for computation. | |
4493 | ||
4494 | *Nicola Tuveri* | |
4495 | ||
4496 | * Compute ECC cofactors if not provided during EC_GROUP construction. Before | |
4497 | this change, EC_GROUP_set_generator would accept order and/or cofactor as | |
4498 | NULL. After this change, only the cofactor parameter can be NULL. It also | |
4499 | does some minimal sanity checks on the passed order. | |
4500 | ([CVE-2019-1547]) | |
4501 | ||
4502 | *Billy Bob Brumley* | |
4503 | ||
4504 | * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey. | |
4505 | An attack is simple, if the first CMS_recipientInfo is valid but the | |
4506 | second CMS_recipientInfo is chosen ciphertext. If the second | |
4507 | recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct | |
4508 | encryption key will be replaced by garbage, and the message cannot be | |
4509 | decoded, but if the RSA decryption fails, the correct encryption key is | |
4510 | used and the recipient will not notice the attack. | |
4511 | As a work around for this potential attack the length of the decrypted | |
4512 | key must be equal to the cipher default key length, in case the | |
4513 | certifiate is not given and all recipientInfo are tried out. | |
4514 | The old behaviour can be re-enabled in the CMS code by setting the | |
4515 | CMS_DEBUG_DECRYPT flag. | |
4516 | ([CVE-2019-1563]) | |
4517 | ||
4518 | *Bernd Edlinger* | |
4519 | ||
4520 | * Document issue with installation paths in diverse Windows builds | |
4521 | ||
4522 | '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL | |
4523 | binaries and run-time config file. | |
4524 | ([CVE-2019-1552]) | |
4525 | ||
4526 | *Richard Levitte* | |
4527 | ||
4528 | ### Changes between 1.0.2r and 1.0.2s [28 May 2019] | |
4529 | ||
4530 | * Change the default RSA, DSA and DH size to 2048 bit instead of 1024. | |
4531 | This changes the size when using the `genpkey` command when no size is given. | |
4532 | It fixes an omission in earlier changes that changed all RSA, DSA and DH | |
4533 | generation commands to use 2048 bits by default. | |
4534 | ||
4535 | *Kurt Roeckx* | |
4536 | ||
4537 | * Add FIPS support for Android Arm 64-bit | |
4538 | ||
4539 | Support for Android Arm 64-bit was added to the OpenSSL FIPS Object | |
4540 | Module in Version 2.0.10. For some reason, the corresponding target | |
4541 | 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be | |
4542 | built with FIPS support on Android Arm 64-bit. This omission has been | |
4543 | fixed. | |
4544 | ||
4545 | *Matthias St. Pierre* | |
4546 | ||
4547 | ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019] | |
4548 | ||
4549 | * 0-byte record padding oracle | |
4550 | ||
4551 | If an application encounters a fatal protocol error and then calls | |
4552 | SSL_shutdown() twice (once to send a close_notify, and once to receive one) | |
4553 | then OpenSSL can respond differently to the calling application if a 0 byte | |
4554 | record is received with invalid padding compared to if a 0 byte record is | |
4555 | received with an invalid MAC. If the application then behaves differently | |
4556 | based on that in a way that is detectable to the remote peer, then this | |
4557 | amounts to a padding oracle that could be used to decrypt data. | |
4558 | ||
4559 | In order for this to be exploitable "non-stitched" ciphersuites must be in | |
4560 | use. Stitched ciphersuites are optimised implementations of certain | |
4561 | commonly used ciphersuites. Also the application must call SSL_shutdown() | |
4562 | twice even if a protocol error has occurred (applications should not do | |
4563 | this but some do anyway). | |
4564 | ||
4565 | This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod | |
4566 | Aviram, with additional investigation by Steven Collison and Andrew | |
4567 | Hourselt. It was reported to OpenSSL on 10th December 2018. | |
4568 | ([CVE-2019-1559]) | |
4569 | ||
4570 | *Matt Caswell* | |
4571 | ||
4572 | * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0(). | |
4573 | ||
4574 | *Richard Levitte* | |
4575 | ||
4576 | ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018] | |
4577 | ||
4578 | * Microarchitecture timing vulnerability in ECC scalar multiplication | |
4579 | ||
4580 | OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been | |
4581 | shown to be vulnerable to a microarchitecture timing side channel attack. | |
4582 | An attacker with sufficient access to mount local timing attacks during | |
4583 | ECDSA signature generation could recover the private key. | |
4584 | ||
4585 | This issue was reported to OpenSSL on 26th October 2018 by Alejandro | |
4586 | Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and | |
4587 | Nicola Tuveri. | |
4588 | ([CVE-2018-5407]) | |
4589 | ||
4590 | *Billy Brumley* | |
4591 | ||
4592 | * Timing vulnerability in DSA signature generation | |
4593 | ||
4594 | The OpenSSL DSA signature algorithm has been shown to be vulnerable to a | |
4595 | timing side channel attack. An attacker could use variations in the signing | |
4596 | algorithm to recover the private key. | |
4597 | ||
4598 | This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser. | |
4599 | ([CVE-2018-0734]) | |
4600 | ||
4601 | *Paul Dale* | |
4602 | ||
4603 | * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object | |
4604 | Module, accidentally introduced while backporting security fixes from the | |
4605 | development branch and hindering the use of ECC in FIPS mode. | |
4606 | ||
4607 | *Nicola Tuveri* | |
4608 | ||
4609 | ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018] | |
4610 | ||
4611 | * Client DoS due to large DH parameter | |
4612 | ||
4613 | During key agreement in a TLS handshake using a DH(E) based ciphersuite a | |
4614 | malicious server can send a very large prime value to the client. This will | |
4615 | cause the client to spend an unreasonably long period of time generating a | |
4616 | key for this prime resulting in a hang until the client has finished. This | |
4617 | could be exploited in a Denial Of Service attack. | |
4618 | ||
4619 | This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken | |
4620 | ([CVE-2018-0732]) | |
4621 | ||
4622 | *Guido Vranken* | |
4623 | ||
4624 | * Cache timing vulnerability in RSA Key Generation | |
4625 | ||
4626 | The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to | |
4627 | a cache timing side channel attack. An attacker with sufficient access to | |
4628 | mount cache timing attacks during the RSA key generation process could | |
4629 | recover the private key. | |
4630 | ||
4631 | This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera | |
4632 | Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia. | |
4633 | ([CVE-2018-0737]) | |
4634 | ||
4635 | *Billy Brumley* | |
4636 | ||
4637 | * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str | |
4638 | parameter is no longer accepted, as it leads to a corrupt table. NULL | |
4639 | pem_str is reserved for alias entries only. | |
4640 | ||
4641 | *Richard Levitte* | |
4642 | ||
4643 | * Revert blinding in ECDSA sign and instead make problematic addition | |
4644 | length-invariant. Switch even to fixed-length Montgomery multiplication. | |
4645 | ||
4646 | *Andy Polyakov* | |
4647 | ||
4648 | * Change generating and checking of primes so that the error rate of not | |
4649 | being prime depends on the intended use based on the size of the input. | |
4650 | For larger primes this will result in more rounds of Miller-Rabin. | |
4651 | The maximal error rate for primes with more than 1080 bits is lowered | |
4652 | to 2^-128. | |
4653 | ||
4654 | *Kurt Roeckx, Annie Yousar* | |
4655 | ||
4656 | * Increase the number of Miller-Rabin rounds for DSA key generating to 64. | |
4657 | ||
4658 | *Kurt Roeckx* | |
4659 | ||
4660 | * Add blinding to ECDSA and DSA signatures to protect against side channel | |
4661 | attacks discovered by Keegan Ryan (NCC Group). | |
4662 | ||
4663 | *Matt Caswell* | |
4664 | ||
4665 | * When unlocking a pass phrase protected PEM file or PKCS#8 container, we | |
4666 | now allow empty (zero character) pass phrases. | |
4667 | ||
4668 | *Richard Levitte* | |
4669 | ||
4670 | * Certificate time validation (X509_cmp_time) enforces stricter | |
4671 | compliance with RFC 5280. Fractional seconds and timezone offsets | |
4672 | are no longer allowed. | |
4673 | ||
4674 | *Emilia Käsper* | |
4675 | ||
4676 | ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018] | |
4677 | ||
4678 | * Constructed ASN.1 types with a recursive definition could exceed the stack | |
4679 | ||
4680 | Constructed ASN.1 types with a recursive definition (such as can be found | |
4681 | in PKCS7) could eventually exceed the stack given malicious input with | |
4682 | excessive recursion. This could result in a Denial Of Service attack. There | |
4683 | are no such structures used within SSL/TLS that come from untrusted sources | |
4684 | so this is considered safe. | |
4685 | ||
4686 | This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz | |
4687 | project. | |
4688 | ([CVE-2018-0739]) | |
4689 | ||
4690 | *Matt Caswell* | |
4691 | ||
4692 | ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017] | |
4693 | ||
4694 | * Read/write after SSL object in error state | |
4695 | ||
4696 | OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state" | |
4697 | mechanism. The intent was that if a fatal error occurred during a handshake | |
4698 | then OpenSSL would move into the error state and would immediately fail if | |
4699 | you attempted to continue the handshake. This works as designed for the | |
4700 | explicit handshake functions (SSL_do_handshake(), SSL_accept() and | |
4701 | SSL_connect()), however due to a bug it does not work correctly if | |
4702 | SSL_read() or SSL_write() is called directly. In that scenario, if the | |
4703 | handshake fails then a fatal error will be returned in the initial function | |
4704 | call. If SSL_read()/SSL_write() is subsequently called by the application | |
4705 | for the same SSL object then it will succeed and the data is passed without | |
4706 | being decrypted/encrypted directly from the SSL/TLS record layer. | |
4707 | ||
4708 | In order to exploit this issue an application bug would have to be present | |
4709 | that resulted in a call to SSL_read()/SSL_write() being issued after having | |
4710 | already received a fatal error. | |
4711 | ||
4712 | This issue was reported to OpenSSL by David Benjamin (Google). | |
4713 | ([CVE-2017-3737]) | |
4714 | ||
4715 | *Matt Caswell* | |
4716 | ||
4717 | * rsaz_1024_mul_avx2 overflow bug on x86_64 | |
4718 | ||
4719 | There is an overflow bug in the AVX2 Montgomery multiplication procedure | |
4720 | used in exponentiation with 1024-bit moduli. No EC algorithms are affected. | |
4721 | Analysis suggests that attacks against RSA and DSA as a result of this | |
4722 | defect would be very difficult to perform and are not believed likely. | |
4723 | Attacks against DH1024 are considered just feasible, because most of the | |
4724 | work necessary to deduce information about a private key may be performed | |
4725 | offline. The amount of resources required for such an attack would be | |
4726 | significant. However, for an attack on TLS to be meaningful, the server | |
4727 | would have to share the DH1024 private key among multiple clients, which is | |
4728 | no longer an option since CVE-2016-0701. | |
4729 | ||
4730 | This only affects processors that support the AVX2 but not ADX extensions | |
4731 | like Intel Haswell (4th generation). | |
4732 | ||
4733 | This issue was reported to OpenSSL by David Benjamin (Google). The issue | |
4734 | was originally found via the OSS-Fuzz project. | |
4735 | ([CVE-2017-3738]) | |
4736 | ||
4737 | *Andy Polyakov* | |
4738 | ||
4739 | ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017] | |
4740 | ||
4741 | * bn_sqrx8x_internal carry bug on x86_64 | |
4742 | ||
4743 | There is a carry propagating bug in the x86_64 Montgomery squaring | |
4744 | procedure. No EC algorithms are affected. Analysis suggests that attacks | |
4745 | against RSA and DSA as a result of this defect would be very difficult to | |
4746 | perform and are not believed likely. Attacks against DH are considered just | |
4747 | feasible (although very difficult) because most of the work necessary to | |
4748 | deduce information about a private key may be performed offline. The amount | |
4749 | of resources required for such an attack would be very significant and | |
4750 | likely only accessible to a limited number of attackers. An attacker would | |
4751 | additionally need online access to an unpatched system using the target | |
4752 | private key in a scenario with persistent DH parameters and a private | |
4753 | key that is shared between multiple clients. | |
4754 | ||
4755 | This only affects processors that support the BMI1, BMI2 and ADX extensions | |
4756 | like Intel Broadwell (5th generation) and later or AMD Ryzen. | |
4757 | ||
4758 | This issue was reported to OpenSSL by the OSS-Fuzz project. | |
4759 | ([CVE-2017-3736]) | |
4760 | ||
4761 | *Andy Polyakov* | |
4762 | ||
4763 | * Malformed X.509 IPAddressFamily could cause OOB read | |
4764 | ||
4765 | If an X.509 certificate has a malformed IPAddressFamily extension, | |
4766 | OpenSSL could do a one-byte buffer overread. The most likely result | |
4767 | would be an erroneous display of the certificate in text format. | |
4768 | ||
4769 | This issue was reported to OpenSSL by the OSS-Fuzz project. | |
4770 | ([CVE-2017-3735]) | |
4771 | ||
4772 | *Rich Salz* | |
4773 | ||
4774 | ### Changes between 1.0.2k and 1.0.2l [25 May 2017] | |
4775 | ||
4776 | * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target | |
4777 | platform rather than 'mingw'. | |
4778 | ||
4779 | *Richard Levitte* | |
4780 | ||
4781 | ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017] | |
4782 | ||
4783 | * Truncated packet could crash via OOB read | |
4784 | ||
4785 | If one side of an SSL/TLS path is running on a 32-bit host and a specific | |
4786 | cipher is being used, then a truncated packet can cause that host to | |
4787 | perform an out-of-bounds read, usually resulting in a crash. | |
4788 | ||
4789 | This issue was reported to OpenSSL by Robert Święcki of Google. | |
4790 | ([CVE-2017-3731]) | |
4791 | ||
4792 | *Andy Polyakov* | |
4793 | ||
4794 | * BN_mod_exp may produce incorrect results on x86_64 | |
4795 | ||
4796 | There is a carry propagating bug in the x86_64 Montgomery squaring | |
4797 | procedure. No EC algorithms are affected. Analysis suggests that attacks | |
4798 | against RSA and DSA as a result of this defect would be very difficult to | |
4799 | perform and are not believed likely. Attacks against DH are considered just | |
4800 | feasible (although very difficult) because most of the work necessary to | |
4801 | deduce information about a private key may be performed offline. The amount | |
4802 | of resources required for such an attack would be very significant and | |
4803 | likely only accessible to a limited number of attackers. An attacker would | |
4804 | additionally need online access to an unpatched system using the target | |
4805 | private key in a scenario with persistent DH parameters and a private | |
4806 | key that is shared between multiple clients. For example this can occur by | |
4807 | default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very | |
4808 | similar to CVE-2015-3193 but must be treated as a separate problem. | |
4809 | ||
4810 | This issue was reported to OpenSSL by the OSS-Fuzz project. | |
4811 | ([CVE-2017-3732]) | |
4812 | ||
4813 | *Andy Polyakov* | |
4814 | ||
4815 | * Montgomery multiplication may produce incorrect results | |
4816 | ||
4817 | There is a carry propagating bug in the Broadwell-specific Montgomery | |
4818 | multiplication procedure that handles input lengths divisible by, but | |
4819 | longer than 256 bits. Analysis suggests that attacks against RSA, DSA | |
4820 | and DH private keys are impossible. This is because the subroutine in | |
4821 | question is not used in operations with the private key itself and an input | |
4822 | of the attacker's direct choice. Otherwise the bug can manifest itself as | |
4823 | transient authentication and key negotiation failures or reproducible | |
4824 | erroneous outcome of public-key operations with specially crafted input. | |
4825 | Among EC algorithms only Brainpool P-512 curves are affected and one | |
4826 | presumably can attack ECDH key negotiation. Impact was not analyzed in | |
4827 | detail, because pre-requisites for attack are considered unlikely. Namely | |
4828 | multiple clients have to choose the curve in question and the server has to | |
4829 | share the private key among them, neither of which is default behaviour. | |
4830 | Even then only clients that chose the curve will be affected. | |
4831 | ||
4832 | This issue was publicly reported as transient failures and was not | |
4833 | initially recognized as a security issue. Thanks to Richard Morgan for | |
4834 | providing reproducible case. | |
4835 | ([CVE-2016-7055]) | |
4836 | ||
4837 | *Andy Polyakov* | |
4838 | ||
4839 | * OpenSSL now fails if it receives an unrecognised record type in TLS1.0 | |
4840 | or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to | |
4841 | prevent issues where no progress is being made and the peer continually | |
4842 | sends unrecognised record types, using up resources processing them. | |
4843 | ||
4844 | *Matt Caswell* | |
4845 | ||
4846 | ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016] | |
4847 | ||
4848 | * Missing CRL sanity check | |
4849 | ||
4850 | A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0 | |
4851 | but was omitted from OpenSSL 1.0.2i. As a result any attempt to use | |
4852 | CRLs in OpenSSL 1.0.2i will crash with a null pointer exception. | |
4853 | ||
4854 | This issue only affects the OpenSSL 1.0.2i | |
4855 | ([CVE-2016-7052]) | |
4856 | ||
4857 | *Matt Caswell* | |
4858 | ||
4859 | ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016] | |
4860 | ||
4861 | * OCSP Status Request extension unbounded memory growth | |
4862 | ||
4863 | A malicious client can send an excessively large OCSP Status Request | |
4864 | extension. If that client continually requests renegotiation, sending a | |
4865 | large OCSP Status Request extension each time, then there will be unbounded | |
4866 | memory growth on the server. This will eventually lead to a Denial Of | |
4867 | Service attack through memory exhaustion. Servers with a default | |
4868 | configuration are vulnerable even if they do not support OCSP. Builds using | |
4869 | the "no-ocsp" build time option are not affected. | |
4870 | ||
4871 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
4872 | ([CVE-2016-6304]) | |
4873 | ||
4874 | *Matt Caswell* | |
4875 | ||
4876 | * In order to mitigate the SWEET32 attack, the DES ciphers were moved from | |
4877 | HIGH to MEDIUM. | |
4878 | ||
4879 | This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan | |
4880 | Leurent (INRIA) | |
4881 | ([CVE-2016-2183]) | |
4882 | ||
4883 | *Rich Salz* | |
4884 | ||
4885 | * OOB write in MDC2_Update() | |
4886 | ||
4887 | An overflow can occur in MDC2_Update() either if called directly or | |
4888 | through the EVP_DigestUpdate() function using MDC2. If an attacker | |
4889 | is able to supply very large amounts of input data after a previous | |
4890 | call to EVP_EncryptUpdate() with a partial block then a length check | |
4891 | can overflow resulting in a heap corruption. | |
4892 | ||
4893 | The amount of data needed is comparable to SIZE_MAX which is impractical | |
4894 | on most platforms. | |
4895 | ||
4896 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
4897 | ([CVE-2016-6303]) | |
4898 | ||
4899 | *Stephen Henson* | |
4900 | ||
4901 | * Malformed SHA512 ticket DoS | |
4902 | ||
4903 | If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a | |
4904 | DoS attack where a malformed ticket will result in an OOB read which will | |
4905 | ultimately crash. | |
4906 | ||
4907 | The use of SHA512 in TLS session tickets is comparatively rare as it requires | |
4908 | a custom server callback and ticket lookup mechanism. | |
4909 | ||
4910 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
4911 | ([CVE-2016-6302]) | |
4912 | ||
4913 | *Stephen Henson* | |
4914 | ||
4915 | * OOB write in BN_bn2dec() | |
4916 | ||
4917 | The function BN_bn2dec() does not check the return value of BN_div_word(). | |
4918 | This can cause an OOB write if an application uses this function with an | |
4919 | overly large BIGNUM. This could be a problem if an overly large certificate | |
4920 | or CRL is printed out from an untrusted source. TLS is not affected because | |
4921 | record limits will reject an oversized certificate before it is parsed. | |
4922 | ||
4923 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
4924 | ([CVE-2016-2182]) | |
4925 | ||
4926 | *Stephen Henson* | |
4927 | ||
4928 | * OOB read in TS_OBJ_print_bio() | |
4929 | ||
4930 | The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is | |
4931 | the total length the OID text representation would use and not the amount | |
4932 | of data written. This will result in OOB reads when large OIDs are | |
4933 | presented. | |
4934 | ||
4935 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
4936 | ([CVE-2016-2180]) | |
4937 | ||
4938 | *Stephen Henson* | |
4939 | ||
4940 | * Pointer arithmetic undefined behaviour | |
4941 | ||
4942 | Avoid some undefined pointer arithmetic | |
4943 | ||
4944 | A common idiom in the codebase is to check limits in the following manner: | |
4945 | "p + len > limit" | |
4946 | ||
4947 | Where "p" points to some malloc'd data of SIZE bytes and | |
4948 | limit == p + SIZE | |
4949 | ||
4950 | "len" here could be from some externally supplied data (e.g. from a TLS | |
4951 | message). | |
4952 | ||
4953 | The rules of C pointer arithmetic are such that "p + len" is only well | |
4954 | defined where len <= SIZE. Therefore the above idiom is actually | |
4955 | undefined behaviour. | |
4956 | ||
4957 | For example this could cause problems if some malloc implementation | |
4958 | provides an address for "p" such that "p + len" actually overflows for | |
4959 | values of len that are too big and therefore p + len < limit. | |
4960 | ||
4961 | This issue was reported to OpenSSL by Guido Vranken | |
4962 | ([CVE-2016-2177]) | |
4963 | ||
4964 | *Matt Caswell* | |
4965 | ||
4966 | * Constant time flag not preserved in DSA signing | |
4967 | ||
4968 | Operations in the DSA signing algorithm should run in constant time in | |
4969 | order to avoid side channel attacks. A flaw in the OpenSSL DSA | |
4970 | implementation means that a non-constant time codepath is followed for | |
4971 | certain operations. This has been demonstrated through a cache-timing | |
4972 | attack to be sufficient for an attacker to recover the private DSA key. | |
4973 | ||
4974 | This issue was reported by César Pereida (Aalto University), Billy Brumley | |
4975 | (Tampere University of Technology), and Yuval Yarom (The University of | |
4976 | Adelaide and NICTA). | |
4977 | ([CVE-2016-2178]) | |
4978 | ||
4979 | *César Pereida* | |
4980 | ||
4981 | * DTLS buffered message DoS | |
4982 | ||
4983 | In a DTLS connection where handshake messages are delivered out-of-order | |
4984 | those messages that OpenSSL is not yet ready to process will be buffered | |
4985 | for later use. Under certain circumstances, a flaw in the logic means that | |
4986 | those messages do not get removed from the buffer even though the handshake | |
4987 | has been completed. An attacker could force up to approx. 15 messages to | |
4988 | remain in the buffer when they are no longer required. These messages will | |
4989 | be cleared when the DTLS connection is closed. The default maximum size for | |
4990 | a message is 100k. Therefore the attacker could force an additional 1500k | |
4991 | to be consumed per connection. By opening many simulataneous connections an | |
4992 | attacker could cause a DoS attack through memory exhaustion. | |
4993 | ||
4994 | This issue was reported to OpenSSL by Quan Luo. | |
4995 | ([CVE-2016-2179]) | |
4996 | ||
4997 | *Matt Caswell* | |
4998 | ||
4999 | * DTLS replay protection DoS | |
5000 | ||
5001 | A flaw in the DTLS replay attack protection mechanism means that records | |
5002 | that arrive for future epochs update the replay protection "window" before | |
5003 | the MAC for the record has been validated. This could be exploited by an | |
5004 | attacker by sending a record for the next epoch (which does not have to | |
5005 | decrypt or have a valid MAC), with a very large sequence number. This means | |
5006 | that all subsequent legitimate packets are dropped causing a denial of | |
5007 | service for a specific DTLS connection. | |
5008 | ||
5009 | This issue was reported to OpenSSL by the OCAP audit team. | |
5010 | ([CVE-2016-2181]) | |
5011 | ||
5012 | *Matt Caswell* | |
5013 | ||
5014 | * Certificate message OOB reads | |
5015 | ||
5016 | In OpenSSL 1.0.2 and earlier some missing message length checks can result | |
5017 | in OOB reads of up to 2 bytes beyond an allocated buffer. There is a | |
5018 | theoretical DoS risk but this has not been observed in practice on common | |
5019 | platforms. | |
5020 | ||
5021 | The messages affected are client certificate, client certificate request | |
5022 | and server certificate. As a result the attack can only be performed | |
5023 | against a client or a server which enables client authentication. | |
5024 | ||
5025 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
5026 | ([CVE-2016-6306]) | |
5027 | ||
5028 | *Stephen Henson* | |
5029 | ||
5030 | ### Changes between 1.0.2g and 1.0.2h [3 May 2016] | |
5031 | ||
5032 | * Prevent padding oracle in AES-NI CBC MAC check | |
5033 | ||
5034 | A MITM attacker can use a padding oracle attack to decrypt traffic | |
5035 | when the connection uses an AES CBC cipher and the server support | |
5036 | AES-NI. | |
5037 | ||
5038 | This issue was introduced as part of the fix for Lucky 13 padding | |
5039 | attack ([CVE-2013-0169]). The padding check was rewritten to be in | |
5040 | constant time by making sure that always the same bytes are read and | |
5041 | compared against either the MAC or padding bytes. But it no longer | |
5042 | checked that there was enough data to have both the MAC and padding | |
5043 | bytes. | |
5044 | ||
5045 | This issue was reported by Juraj Somorovsky using TLS-Attacker. | |
5046 | ([CVE-2016-2107]) | |
5047 | ||
5048 | *Kurt Roeckx* | |
5049 | ||
5050 | * Fix EVP_EncodeUpdate overflow | |
5051 | ||
5052 | An overflow can occur in the EVP_EncodeUpdate() function which is used for | |
5053 | Base64 encoding of binary data. If an attacker is able to supply very large | |
5054 | amounts of input data then a length check can overflow resulting in a heap | |
5055 | corruption. | |
5056 | ||
5057 | Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by | |
5058 | the `PEM_write_bio*` family of functions. These are mainly used within the | |
5059 | OpenSSL command line applications, so any application which processes data | |
5060 | from an untrusted source and outputs it as a PEM file should be considered | |
5061 | vulnerable to this issue. User applications that call these APIs directly | |
5062 | with large amounts of untrusted data may also be vulnerable. | |
5063 | ||
5064 | This issue was reported by Guido Vranken. | |
5065 | ([CVE-2016-2105]) | |
5066 | ||
5067 | *Matt Caswell* | |
5068 | ||
5069 | * Fix EVP_EncryptUpdate overflow | |
5070 | ||
5071 | An overflow can occur in the EVP_EncryptUpdate() function. If an attacker | |
5072 | is able to supply very large amounts of input data after a previous call to | |
5073 | EVP_EncryptUpdate() with a partial block then a length check can overflow | |
5074 | resulting in a heap corruption. Following an analysis of all OpenSSL | |
5075 | internal usage of the EVP_EncryptUpdate() function all usage is one of two | |
5076 | forms. The first form is where the EVP_EncryptUpdate() call is known to be | |
5077 | the first called function after an EVP_EncryptInit(), and therefore that | |
5078 | specific call must be safe. The second form is where the length passed to | |
5079 | EVP_EncryptUpdate() can be seen from the code to be some small value and | |
5080 | therefore there is no possibility of an overflow. Since all instances are | |
5081 | one of these two forms, it is believed that there can be no overflows in | |
5082 | internal code due to this problem. It should be noted that | |
5083 | EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths. | |
5084 | Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances | |
5085 | of these calls have also been analysed too and it is believed there are no | |
5086 | instances in internal usage where an overflow could occur. | |
5087 | ||
5088 | This issue was reported by Guido Vranken. | |
5089 | ([CVE-2016-2106]) | |
5090 | ||
5091 | *Matt Caswell* | |
5092 | ||
5093 | * Prevent ASN.1 BIO excessive memory allocation | |
5094 | ||
5095 | When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() | |
5096 | a short invalid encoding can cause allocation of large amounts of memory | |
5097 | potentially consuming excessive resources or exhausting memory. | |
5098 | ||
5099 | Any application parsing untrusted data through d2i BIO functions is | |
5100 | affected. The memory based functions such as d2i_X509() are *not* affected. | |
5101 | Since the memory based functions are used by the TLS library, TLS | |
5102 | applications are not affected. | |
5103 | ||
5104 | This issue was reported by Brian Carpenter. | |
5105 | ([CVE-2016-2109]) | |
5106 | ||
5107 | *Stephen Henson* | |
5108 | ||
5109 | * EBCDIC overread | |
5110 | ||
5111 | ASN1 Strings that are over 1024 bytes can cause an overread in applications | |
5112 | using the X509_NAME_oneline() function on EBCDIC systems. This could result | |
5113 | in arbitrary stack data being returned in the buffer. | |
5114 | ||
5115 | This issue was reported by Guido Vranken. | |
5116 | ([CVE-2016-2176]) | |
5117 | ||
5118 | *Matt Caswell* | |
5119 | ||
5120 | * Modify behavior of ALPN to invoke callback after SNI/servername | |
5121 | callback, such that updates to the SSL_CTX affect ALPN. | |
5122 | ||
5123 | *Todd Short* | |
5124 | ||
5125 | * Remove LOW from the DEFAULT cipher list. This removes singles DES from the | |
5126 | default. | |
5127 | ||
5128 | *Kurt Roeckx* | |
5129 | ||
5130 | * Only remove the SSLv2 methods with the no-ssl2-method option. When the | |
5131 | methods are enabled and ssl2 is disabled the methods return NULL. | |
5132 | ||
5133 | *Kurt Roeckx* | |
5134 | ||
5135 | ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016] | |
5136 | ||
5137 | * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL. | |
5138 | Builds that are not configured with "enable-weak-ssl-ciphers" will not | |
5139 | provide any "EXPORT" or "LOW" strength ciphers. | |
5140 | ||
5141 | *Viktor Dukhovni* | |
5142 | ||
5143 | * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2 | |
5144 | is by default disabled at build-time. Builds that are not configured with | |
5145 | "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used, | |
5146 | users who want to negotiate SSLv2 via the version-flexible SSLv23_method() | |
5147 | will need to explicitly call either of: | |
5148 | ||
5149 | SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2); | |
5150 | or | |
5151 | SSL_clear_options(ssl, SSL_OP_NO_SSLv2); | |
5152 | ||
5153 | as appropriate. Even if either of those is used, or the application | |
5154 | explicitly uses the version-specific SSLv2_method() or its client and | |
5155 | server variants, SSLv2 ciphers vulnerable to exhaustive search key | |
5156 | recovery have been removed. Specifically, the SSLv2 40-bit EXPORT | |
5157 | ciphers, and SSLv2 56-bit DES are no longer available. | |
5158 | ([CVE-2016-0800]) | |
5159 | ||
5160 | *Viktor Dukhovni* | |
5161 | ||
5162 | * Fix a double-free in DSA code | |
5163 | ||
5164 | A double free bug was discovered when OpenSSL parses malformed DSA private | |
5165 | keys and could lead to a DoS attack or memory corruption for applications | |
5166 | that receive DSA private keys from untrusted sources. This scenario is | |
5167 | considered rare. | |
5168 | ||
5169 | This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using | |
5170 | libFuzzer. | |
5171 | ([CVE-2016-0705]) | |
5172 | ||
5173 | *Stephen Henson* | |
5174 | ||
5175 | * Disable SRP fake user seed to address a server memory leak. | |
5176 | ||
5177 | Add a new method SRP_VBASE_get1_by_user that handles the seed properly. | |
5178 | ||
5179 | SRP_VBASE_get_by_user had inconsistent memory management behaviour. | |
5180 | In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user | |
5181 | was changed to ignore the "fake user" SRP seed, even if the seed | |
5182 | is configured. | |
5183 | ||
5184 | Users should use SRP_VBASE_get1_by_user instead. Note that in | |
5185 | SRP_VBASE_get1_by_user, caller must free the returned value. Note | |
5186 | also that even though configuring the SRP seed attempts to hide | |
5187 | invalid usernames by continuing the handshake with fake | |
5188 | credentials, this behaviour is not constant time and no strong | |
5189 | guarantees are made that the handshake is indistinguishable from | |
5190 | that of a valid user. | |
5191 | ([CVE-2016-0798]) | |
5192 | ||
5193 | *Emilia Käsper* | |
5194 | ||
5195 | * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption | |
5196 | ||
5197 | In the BN_hex2bn function the number of hex digits is calculated using an | |
5198 | int value `i`. Later `bn_expand` is called with a value of `i * 4`. For | |
5199 | large values of `i` this can result in `bn_expand` not allocating any | |
5200 | memory because `i * 4` is negative. This can leave the internal BIGNUM data | |
5201 | field as NULL leading to a subsequent NULL ptr deref. For very large values | |
5202 | of `i`, the calculation `i * 4` could be a positive value smaller than `i`. | |
5203 | In this case memory is allocated to the internal BIGNUM data field, but it | |
5204 | is insufficiently sized leading to heap corruption. A similar issue exists | |
5205 | in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn | |
5206 | is ever called by user applications with very large untrusted hex/dec data. | |
5207 | This is anticipated to be a rare occurrence. | |
5208 | ||
5209 | All OpenSSL internal usage of these functions use data that is not expected | |
5210 | to be untrusted, e.g. config file data or application command line | |
5211 | arguments. If user developed applications generate config file data based | |
5212 | on untrusted data then it is possible that this could also lead to security | |
5213 | consequences. This is also anticipated to be rare. | |
5214 | ||
5215 | This issue was reported to OpenSSL by Guido Vranken. | |
5216 | ([CVE-2016-0797]) | |
5217 | ||
5218 | *Matt Caswell* | |
5219 | ||
5220 | * Fix memory issues in `BIO_*printf` functions | |
5221 | ||
5222 | The internal `fmtstr` function used in processing a "%s" format string in | |
5223 | the `BIO_*printf` functions could overflow while calculating the length of a | |
5224 | string and cause an OOB read when printing very long strings. | |
5225 | ||
5226 | Additionally the internal `doapr_outch` function can attempt to write to an | |
5227 | OOB memory location (at an offset from the NULL pointer) in the event of a | |
5228 | memory allocation failure. In 1.0.2 and below this could be caused where | |
5229 | the size of a buffer to be allocated is greater than INT_MAX. E.g. this | |
5230 | could be in processing a very long "%s" format string. Memory leaks can | |
5231 | also occur. | |
5232 | ||
5233 | The first issue may mask the second issue dependent on compiler behaviour. | |
5234 | These problems could enable attacks where large amounts of untrusted data | |
5235 | is passed to the `BIO_*printf` functions. If applications use these functions | |
5236 | in this way then they could be vulnerable. OpenSSL itself uses these | |
5237 | functions when printing out human-readable dumps of ASN.1 data. Therefore | |
5238 | applications that print this data could be vulnerable if the data is from | |
5239 | untrusted sources. OpenSSL command line applications could also be | |
5240 | vulnerable where they print out ASN.1 data, or if untrusted data is passed | |
5241 | as command line arguments. | |
5242 | ||
5243 | Libssl is not considered directly vulnerable. Additionally certificates etc | |
5244 | received via remote connections via libssl are also unlikely to be able to | |
5245 | trigger these issues because of message size limits enforced within libssl. | |
5246 | ||
5247 | This issue was reported to OpenSSL Guido Vranken. | |
5248 | ([CVE-2016-0799]) | |
5249 | ||
5250 | *Matt Caswell* | |
5251 | ||
5252 | * Side channel attack on modular exponentiation | |
5253 | ||
5254 | A side-channel attack was found which makes use of cache-bank conflicts on | |
5255 | the Intel Sandy-Bridge microarchitecture which could lead to the recovery | |
5256 | of RSA keys. The ability to exploit this issue is limited as it relies on | |
5257 | an attacker who has control of code in a thread running on the same | |
5258 | hyper-threaded core as the victim thread which is performing decryptions. | |
5259 | ||
5260 | This issue was reported to OpenSSL by Yuval Yarom, The University of | |
5261 | Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and | |
5262 | Nadia Heninger, University of Pennsylvania with more information at | |
5263 | <http://cachebleed.info>. | |
5264 | ([CVE-2016-0702]) | |
5265 | ||
5266 | *Andy Polyakov* | |
5267 | ||
5268 | * Change the `req` command to generate a 2048-bit RSA/DSA key by default, | |
5269 | if no keysize is specified with default_bits. This fixes an | |
5270 | omission in an earlier change that changed all RSA/DSA key generation | |
5271 | commands to use 2048 bits by default. | |
5272 | ||
5273 | *Emilia Käsper* | |
5274 | ||
5275 | ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016] | |
5276 | ||
5277 | * DH small subgroups | |
5278 | ||
5279 | Historically OpenSSL only ever generated DH parameters based on "safe" | |
5280 | primes. More recently (in version 1.0.2) support was provided for | |
5281 | generating X9.42 style parameter files such as those required for RFC 5114 | |
5282 | support. The primes used in such files may not be "safe". Where an | |
5283 | application is using DH configured with parameters based on primes that are | |
5284 | not "safe" then an attacker could use this fact to find a peer's private | |
5285 | DH exponent. This attack requires that the attacker complete multiple | |
5286 | handshakes in which the peer uses the same private DH exponent. For example | |
5287 | this could be used to discover a TLS server's private DH exponent if it's | |
5288 | reusing the private DH exponent or it's using a static DH ciphersuite. | |
5289 | ||
5290 | OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in | |
5291 | TLS. It is not on by default. If the option is not set then the server | |
5292 | reuses the same private DH exponent for the life of the server process and | |
5293 | would be vulnerable to this attack. It is believed that many popular | |
5294 | applications do set this option and would therefore not be at risk. | |
5295 | ||
5296 | The fix for this issue adds an additional check where a "q" parameter is | |
5297 | available (as is the case in X9.42 based parameters). This detects the | |
5298 | only known attack, and is the only possible defense for static DH | |
5299 | ciphersuites. This could have some performance impact. | |
5300 | ||
5301 | Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by | |
5302 | default and cannot be disabled. This could have some performance impact. | |
5303 | ||
5304 | This issue was reported to OpenSSL by Antonio Sanso (Adobe). | |
5305 | ([CVE-2016-0701]) | |
5306 | ||
5307 | *Matt Caswell* | |
5308 | ||
5309 | * SSLv2 doesn't block disabled ciphers | |
5310 | ||
5311 | A malicious client can negotiate SSLv2 ciphers that have been disabled on | |
5312 | the server and complete SSLv2 handshakes even if all SSLv2 ciphers have | |
5313 | been disabled, provided that the SSLv2 protocol was not also disabled via | |
5314 | SSL_OP_NO_SSLv2. | |
5315 | ||
5316 | This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram | |
5317 | and Sebastian Schinzel. | |
5318 | ([CVE-2015-3197]) | |
5319 | ||
5320 | *Viktor Dukhovni* | |
5321 | ||
5322 | ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015] | |
5323 | ||
5324 | * BN_mod_exp may produce incorrect results on x86_64 | |
5325 | ||
5326 | There is a carry propagating bug in the x86_64 Montgomery squaring | |
5327 | procedure. No EC algorithms are affected. Analysis suggests that attacks | |
5328 | against RSA and DSA as a result of this defect would be very difficult to | |
5329 | perform and are not believed likely. Attacks against DH are considered just | |
5330 | feasible (although very difficult) because most of the work necessary to | |
5331 | deduce information about a private key may be performed offline. The amount | |
5332 | of resources required for such an attack would be very significant and | |
5333 | likely only accessible to a limited number of attackers. An attacker would | |
5334 | additionally need online access to an unpatched system using the target | |
5335 | private key in a scenario with persistent DH parameters and a private | |
5336 | key that is shared between multiple clients. For example this can occur by | |
5337 | default in OpenSSL DHE based SSL/TLS ciphersuites. | |
5338 | ||
5339 | This issue was reported to OpenSSL by Hanno Böck. | |
5340 | ([CVE-2015-3193]) | |
5341 | ||
5342 | *Andy Polyakov* | |
5343 | ||
5344 | * Certificate verify crash with missing PSS parameter | |
5345 | ||
5346 | The signature verification routines will crash with a NULL pointer | |
5347 | dereference if presented with an ASN.1 signature using the RSA PSS | |
5348 | algorithm and absent mask generation function parameter. Since these | |
5349 | routines are used to verify certificate signature algorithms this can be | |
5350 | used to crash any certificate verification operation and exploited in a | |
5351 | DoS attack. Any application which performs certificate verification is | |
5352 | vulnerable including OpenSSL clients and servers which enable client | |
5353 | authentication. | |
5354 | ||
5355 | This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG). | |
5356 | ([CVE-2015-3194]) | |
5357 | ||
5358 | *Stephen Henson* | |
5359 | ||
5360 | * X509_ATTRIBUTE memory leak | |
5361 | ||
5362 | When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak | |
5363 | memory. This structure is used by the PKCS#7 and CMS routines so any | |
5364 | application which reads PKCS#7 or CMS data from untrusted sources is | |
5365 | affected. SSL/TLS is not affected. | |
5366 | ||
5367 | This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using | |
5368 | libFuzzer. | |
5369 | ([CVE-2015-3195]) | |
5370 | ||
5371 | *Stephen Henson* | |
5372 | ||
5373 | * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. | |
5374 | This changes the decoding behaviour for some invalid messages, | |
5375 | though the change is mostly in the more lenient direction, and | |
5376 | legacy behaviour is preserved as much as possible. | |
5377 | ||
5378 | *Emilia Käsper* | |
5379 | ||
5380 | * In DSA_generate_parameters_ex, if the provided seed is too short, | |
5381 | return an error | |
5382 | ||
5383 | *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>* | |
5384 | ||
5385 | ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015] | |
5386 | ||
5387 | * Alternate chains certificate forgery | |
5388 | ||
5389 | During certificate verification, OpenSSL will attempt to find an | |
5390 | alternative certificate chain if the first attempt to build such a chain | |
5391 | fails. An error in the implementation of this logic can mean that an | |
5392 | attacker could cause certain checks on untrusted certificates to be | |
5393 | bypassed, such as the CA flag, enabling them to use a valid leaf | |
5394 | certificate to act as a CA and "issue" an invalid certificate. | |
5395 | ||
5396 | This issue was reported to OpenSSL by Adam Langley/David Benjamin | |
5397 | (Google/BoringSSL). | |
5398 | ||
5399 | *Matt Caswell* | |
5400 | ||
5401 | ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015] | |
5402 | ||
5403 | * Fix HMAC ABI incompatibility. The previous version introduced an ABI | |
5404 | incompatibility in the handling of HMAC. The previous ABI has now been | |
5405 | restored. | |
5406 | ||
5407 | *Matt Caswell* | |
5408 | ||
5409 | ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015] | |
5410 | ||
5411 | * Malformed ECParameters causes infinite loop | |
5412 | ||
5413 | When processing an ECParameters structure OpenSSL enters an infinite loop | |
5414 | if the curve specified is over a specially malformed binary polynomial | |
5415 | field. | |
5416 | ||
5417 | This can be used to perform denial of service against any | |
5418 | system which processes public keys, certificate requests or | |
5419 | certificates. This includes TLS clients and TLS servers with | |
5420 | client authentication enabled. | |
5421 | ||
5422 | This issue was reported to OpenSSL by Joseph Barr-Pixton. | |
5423 | ([CVE-2015-1788]) | |
5424 | ||
5425 | *Andy Polyakov* | |
5426 | ||
5427 | * Exploitable out-of-bounds read in X509_cmp_time | |
5428 | ||
5429 | X509_cmp_time does not properly check the length of the ASN1_TIME | |
5430 | string and can read a few bytes out of bounds. In addition, | |
5431 | X509_cmp_time accepts an arbitrary number of fractional seconds in the | |
5432 | time string. | |
5433 | ||
5434 | An attacker can use this to craft malformed certificates and CRLs of | |
5435 | various sizes and potentially cause a segmentation fault, resulting in | |
5436 | a DoS on applications that verify certificates or CRLs. TLS clients | |
5437 | that verify CRLs are affected. TLS clients and servers with client | |
5438 | authentication enabled may be affected if they use custom verification | |
5439 | callbacks. | |
5440 | ||
5441 | This issue was reported to OpenSSL by Robert Swiecki (Google), and | |
5442 | independently by Hanno Böck. | |
5443 | ([CVE-2015-1789]) | |
5444 | ||
5445 | *Emilia Käsper* | |
5446 | ||
5447 | * PKCS7 crash with missing EnvelopedContent | |
5448 | ||
5449 | The PKCS#7 parsing code does not handle missing inner EncryptedContent | |
5450 | correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs | |
5451 | with missing content and trigger a NULL pointer dereference on parsing. | |
5452 | ||
5453 | Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 | |
5454 | structures from untrusted sources are affected. OpenSSL clients and | |
5455 | servers are not affected. | |
5456 | ||
5457 | This issue was reported to OpenSSL by Michal Zalewski (Google). | |
5458 | ([CVE-2015-1790]) | |
5459 | ||
5460 | *Emilia Käsper* | |
5461 | ||
5462 | * CMS verify infinite loop with unknown hash function | |
5463 | ||
5464 | When verifying a signedData message the CMS code can enter an infinite loop | |
5465 | if presented with an unknown hash function OID. This can be used to perform | |
5466 | denial of service against any system which verifies signedData messages using | |
5467 | the CMS code. | |
5468 | This issue was reported to OpenSSL by Johannes Bauer. | |
5469 | ([CVE-2015-1792]) | |
5470 | ||
5471 | *Stephen Henson* | |
5472 | ||
5473 | * Race condition handling NewSessionTicket | |
5474 | ||
5475 | If a NewSessionTicket is received by a multi-threaded client when attempting to | |
5476 | reuse a previous ticket then a race condition can occur potentially leading to | |
5477 | a double free of the ticket data. | |
5478 | ([CVE-2015-1791]) | |
5479 | ||
5480 | *Matt Caswell* | |
5481 | ||
5482 | * Only support 256-bit or stronger elliptic curves with the | |
5483 | 'ecdh_auto' setting (server) or by default (client). Of supported | |
5484 | curves, prefer P-256 (both). | |
5485 | ||
5486 | *Emilia Kasper* | |
5487 | ||
5488 | ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015] | |
5489 | ||
5490 | * ClientHello sigalgs DoS fix | |
5491 | ||
5492 | If a client connects to an OpenSSL 1.0.2 server and renegotiates with an | |
5493 | invalid signature algorithms extension a NULL pointer dereference will | |
5494 | occur. This can be exploited in a DoS attack against the server. | |
5495 | ||
5496 | This issue was was reported to OpenSSL by David Ramos of Stanford | |
5497 | University. | |
5498 | ([CVE-2015-0291]) | |
5499 | ||
5500 | *Stephen Henson and Matt Caswell* | |
5501 | ||
5502 | * Multiblock corrupted pointer fix | |
5503 | ||
5504 | OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This | |
5505 | feature only applies on 64 bit x86 architecture platforms that support AES | |
5506 | NI instructions. A defect in the implementation of "multiblock" can cause | |
5507 | OpenSSL's internal write buffer to become incorrectly set to NULL when | |
5508 | using non-blocking IO. Typically, when the user application is using a | |
5509 | socket BIO for writing, this will only result in a failed connection. | |
5510 | However if some other BIO is used then it is likely that a segmentation | |
5511 | fault will be triggered, thus enabling a potential DoS attack. | |
5512 | ||
5513 | This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller. | |
5514 | ([CVE-2015-0290]) | |
5515 | ||
5516 | *Matt Caswell* | |
5517 | ||
5518 | * Segmentation fault in DTLSv1_listen fix | |
5519 | ||
5520 | The DTLSv1_listen function is intended to be stateless and processes the | |
5521 | initial ClientHello from many peers. It is common for user code to loop | |
5522 | over the call to DTLSv1_listen until a valid ClientHello is received with | |
5523 | an associated cookie. A defect in the implementation of DTLSv1_listen means | |
5524 | that state is preserved in the SSL object from one invocation to the next | |
5525 | that can lead to a segmentation fault. Errors processing the initial | |
5526 | ClientHello can trigger this scenario. An example of such an error could be | |
5527 | that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only | |
5528 | server. | |
5529 | ||
5530 | This issue was reported to OpenSSL by Per Allansson. | |
5531 | ([CVE-2015-0207]) | |
5532 | ||
5533 | *Matt Caswell* | |
5534 | ||
5535 | * Segmentation fault in ASN1_TYPE_cmp fix | |
5536 | ||
5537 | The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is | |
5538 | made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check | |
5539 | certificate signature algorithm consistency this can be used to crash any | |
5540 | certificate verification operation and exploited in a DoS attack. Any | |
5541 | application which performs certificate verification is vulnerable including | |
5542 | OpenSSL clients and servers which enable client authentication. | |
5543 | ([CVE-2015-0286]) | |
5544 | ||
5545 | *Stephen Henson* | |
5546 | ||
5547 | * Segmentation fault for invalid PSS parameters fix | |
5548 | ||
5549 | The signature verification routines will crash with a NULL pointer | |
5550 | dereference if presented with an ASN.1 signature using the RSA PSS | |
5551 | algorithm and invalid parameters. Since these routines are used to verify | |
5552 | certificate signature algorithms this can be used to crash any | |
5553 | certificate verification operation and exploited in a DoS attack. Any | |
5554 | application which performs certificate verification is vulnerable including | |
5555 | OpenSSL clients and servers which enable client authentication. | |
5556 | ||
5557 | This issue was was reported to OpenSSL by Brian Carpenter. | |
5558 | ([CVE-2015-0208]) | |
5559 | ||
5560 | *Stephen Henson* | |
5561 | ||
5562 | * ASN.1 structure reuse memory corruption fix | |
5563 | ||
5564 | Reusing a structure in ASN.1 parsing may allow an attacker to cause | |
5565 | memory corruption via an invalid write. Such reuse is and has been | |
5566 | strongly discouraged and is believed to be rare. | |
5567 | ||
5568 | Applications that parse structures containing CHOICE or ANY DEFINED BY | |
5569 | components may be affected. Certificate parsing (d2i_X509 and related | |
5570 | functions) are however not affected. OpenSSL clients and servers are | |
5571 | not affected. | |
5572 | ([CVE-2015-0287]) | |
5573 | ||
5574 | *Stephen Henson* | |
5575 | ||
5576 | * PKCS7 NULL pointer dereferences fix | |
5577 | ||
5578 | The PKCS#7 parsing code does not handle missing outer ContentInfo | |
5579 | correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with | |
5580 | missing content and trigger a NULL pointer dereference on parsing. | |
5581 | ||
5582 | Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or | |
5583 | otherwise parse PKCS#7 structures from untrusted sources are | |
5584 | affected. OpenSSL clients and servers are not affected. | |
5585 | ||
5586 | This issue was reported to OpenSSL by Michal Zalewski (Google). | |
5587 | ([CVE-2015-0289]) | |
5588 | ||
5589 | *Emilia Käsper* | |
5590 | ||
5591 | * DoS via reachable assert in SSLv2 servers fix | |
5592 | ||
5593 | A malicious client can trigger an OPENSSL_assert (i.e., an abort) in | |
5594 | servers that both support SSLv2 and enable export cipher suites by sending | |
5595 | a specially crafted SSLv2 CLIENT-MASTER-KEY message. | |
5596 | ||
5597 | This issue was discovered by Sean Burford (Google) and Emilia Käsper | |
5598 | (OpenSSL development team). | |
5599 | ([CVE-2015-0293]) | |
5600 | ||
5601 | *Emilia Käsper* | |
5602 | ||
5603 | * Empty CKE with client auth and DHE fix | |
5604 | ||
5605 | If client auth is used then a server can seg fault in the event of a DHE | |
5606 | ciphersuite being selected and a zero length ClientKeyExchange message | |
5607 | being sent by the client. This could be exploited in a DoS attack. | |
5608 | ([CVE-2015-1787]) | |
5609 | ||
5610 | *Matt Caswell* | |
5611 | ||
5612 | * Handshake with unseeded PRNG fix | |
5613 | ||
5614 | Under certain conditions an OpenSSL 1.0.2 client can complete a handshake | |
5615 | with an unseeded PRNG. The conditions are: | |
5616 | - The client is on a platform where the PRNG has not been seeded | |
5617 | automatically, and the user has not seeded manually | |
5618 | - A protocol specific client method version has been used (i.e. not | |
5619 | SSL_client_methodv23) | |
5620 | - A ciphersuite is used that does not require additional random data from | |
5621 | the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA). | |
5622 | ||
5623 | If the handshake succeeds then the client random that has been used will | |
5624 | have been generated from a PRNG with insufficient entropy and therefore the | |
5625 | output may be predictable. | |
5626 | ||
5627 | For example using the following command with an unseeded openssl will | |
5628 | succeed on an unpatched platform: | |
5629 | ||
5630 | openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA | |
5631 | ([CVE-2015-0285]) | |
5632 | ||
5633 | *Matt Caswell* | |
5634 | ||
5635 | * Use After Free following d2i_ECPrivatekey error fix | |
5636 | ||
5637 | A malformed EC private key file consumed via the d2i_ECPrivateKey function | |
5638 | could cause a use after free condition. This, in turn, could cause a double | |
5639 | free in several private key parsing functions (such as d2i_PrivateKey | |
5640 | or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption | |
5641 | for applications that receive EC private keys from untrusted | |
5642 | sources. This scenario is considered rare. | |
5643 | ||
5644 | This issue was discovered by the BoringSSL project and fixed in their | |
5645 | commit 517073cd4b. | |
5646 | ([CVE-2015-0209]) | |
5647 | ||
5648 | *Matt Caswell* | |
5649 | ||
5650 | * X509_to_X509_REQ NULL pointer deref fix | |
5651 | ||
5652 | The function X509_to_X509_REQ will crash with a NULL pointer dereference if | |
5653 | the certificate key is invalid. This function is rarely used in practice. | |
5654 | ||
5655 | This issue was discovered by Brian Carpenter. | |
5656 | ([CVE-2015-0288]) | |
5657 | ||
5658 | *Stephen Henson* | |
5659 | ||
5660 | * Removed the export ciphers from the DEFAULT ciphers | |
5661 | ||
5662 | *Kurt Roeckx* | |
5663 | ||
5664 | ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015] | |
5665 | ||
5666 | * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g. | |
5667 | ARMv5 through ARMv8, as opposite to "locking" it to single one. | |
5668 | So far those who have to target multiple platforms would compromise | |
5669 | and argue that binary targeting say ARMv5 would still execute on | |
5670 | ARMv8. "Universal" build resolves this compromise by providing | |
5671 | near-optimal performance even on newer platforms. | |
5672 | ||
5673 | *Andy Polyakov* | |
5674 | ||
5675 | * Accelerated NIST P-256 elliptic curve implementation for x86_64 | |
5676 | (other platforms pending). | |
5677 | ||
5678 | *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov* | |
5679 | ||
5680 | * Add support for the SignedCertificateTimestampList certificate and | |
5681 | OCSP response extensions from RFC6962. | |
5682 | ||
5683 | *Rob Stradling* | |
5684 | ||
5685 | * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) | |
5686 | for corner cases. (Certain input points at infinity could lead to | |
5687 | bogus results, with non-infinity inputs mapped to infinity too.) | |
5688 | ||
5689 | *Bodo Moeller* | |
5690 | ||
5691 | * Initial support for PowerISA 2.0.7, first implemented in POWER8. | |
5692 | This covers AES, SHA256/512 and GHASH. "Initial" means that most | |
5693 | common cases are optimized and there still is room for further | |
5694 | improvements. Vector Permutation AES for Altivec is also added. | |
5695 | ||
5696 | *Andy Polyakov* | |
5697 | ||
5698 | * Add support for little-endian ppc64 Linux target. | |
5699 | ||
5700 | *Marcelo Cerri (IBM)* | |
5701 | ||
5702 | * Initial support for AMRv8 ISA crypto extensions. This covers AES, | |
5703 | SHA1, SHA256 and GHASH. "Initial" means that most common cases | |
5704 | are optimized and there still is room for further improvements. | |
5705 | Both 32- and 64-bit modes are supported. | |
5706 | ||
5707 | *Andy Polyakov, Ard Biesheuvel (Linaro)* | |
5708 | ||
5709 | * Improved ARMv7 NEON support. | |
5710 | ||
5711 | *Andy Polyakov* | |
5712 | ||
5713 | * Support for SPARC Architecture 2011 crypto extensions, first | |
5714 | implemented in SPARC T4. This covers AES, DES, Camellia, SHA1, | |
5715 | SHA256/512, MD5, GHASH and modular exponentiation. | |
5716 | ||
5717 | *Andy Polyakov, David Miller* | |
5718 | ||
5719 | * Accelerated modular exponentiation for Intel processors, a.k.a. | |
5720 | RSAZ. | |
5721 | ||
5722 | *Shay Gueron & Vlad Krasnov (Intel Corp)* | |
5723 | ||
5724 | * Support for new and upcoming Intel processors, including AVX2, | |
5725 | BMI and SHA ISA extensions. This includes additional "stitched" | |
5726 | implementations, AESNI-SHA256 and GCM, and multi-buffer support | |
5727 | for TLS encrypt. | |
5728 | ||
5729 | This work was sponsored by Intel Corp. | |
5730 | ||
5731 | *Andy Polyakov* | |
5732 | ||
5733 | * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method() | |
5734 | supports both DTLS 1.2 and 1.0 and should use whatever version the peer | |
5735 | supports and DTLSv1_2_*_method() which supports DTLS 1.2 only. | |
5736 | ||
5737 | *Steve Henson* | |
5738 | ||
5739 | * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file(): | |
5740 | this fixes a limitation in previous versions of OpenSSL. | |
5741 | ||
5742 | *Steve Henson* | |
5743 | ||
5744 | * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest, | |
5745 | MGF1 digest and OAEP label. | |
5746 | ||
5747 | *Steve Henson* | |
5748 | ||
5749 | * Add EVP support for key wrapping algorithms, to avoid problems with | |
5750 | existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in | |
5751 | the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap | |
5752 | algorithms and include tests cases. | |
5753 | ||
5754 | *Steve Henson* | |
5755 | ||
5756 | * Add functions to allocate and set the fields of an ECDSA_METHOD | |
5757 | structure. | |
5758 | ||
5759 | *Douglas E. Engert, Steve Henson* | |
5760 | ||
5761 | * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the | |
5762 | difference in days and seconds between two tm or ASN1_TIME structures. | |
5763 | ||
5764 | *Steve Henson* | |
5765 | ||
5766 | * Add -rev test option to s_server to just reverse order of characters | |
5767 | received by client and send back to server. Also prints an abbreviated | |
5768 | summary of the connection parameters. | |
5769 | ||
5770 | *Steve Henson* | |
5771 | ||
5772 | * New option -brief for s_client and s_server to print out a brief summary | |
5773 | of connection parameters. | |
5774 | ||
5775 | *Steve Henson* | |
5776 | ||
5777 | * Add callbacks for arbitrary TLS extensions. | |
5778 | ||
5779 | *Trevor Perrin <trevp@trevp.net> and Ben Laurie* | |
5780 | ||
5781 | * New option -crl_download in several openssl utilities to download CRLs | |
5782 | from CRLDP extension in certificates. | |
5783 | ||
5784 | *Steve Henson* | |
5785 | ||
5786 | * New options -CRL and -CRLform for s_client and s_server for CRLs. | |
5787 | ||
5788 | *Steve Henson* | |
5789 | ||
5790 | * New function X509_CRL_diff to generate a delta CRL from the difference | |
5791 | of two full CRLs. Add support to "crl" utility. | |
5792 | ||
5793 | *Steve Henson* | |
5794 | ||
5795 | * New functions to set lookup_crls function and to retrieve | |
5796 | X509_STORE from X509_STORE_CTX. | |
5797 | ||
5798 | *Steve Henson* | |
5799 | ||
5800 | * Print out deprecated issuer and subject unique ID fields in | |
5801 | certificates. | |
5802 | ||
5803 | *Steve Henson* | |
5804 | ||
5805 | * Extend OCSP I/O functions so they can be used for simple general purpose | |
5806 | HTTP as well as OCSP. New wrapper function which can be used to download | |
5807 | CRLs using the OCSP API. | |
5808 | ||
5809 | *Steve Henson* | |
5810 | ||
5811 | * Delegate command line handling in s_client/s_server to SSL_CONF APIs. | |
5812 | ||
5813 | *Steve Henson* | |
5814 | ||
5815 | * `SSL_CONF*` functions. These provide a common framework for application | |
5816 | configuration using configuration files or command lines. | |
5817 | ||
5818 | *Steve Henson* | |
5819 | ||
5820 | * SSL/TLS tracing code. This parses out SSL/TLS records using the | |
5821 | message callback and prints the results. Needs compile time option | |
5822 | "enable-ssl-trace". New options to s_client and s_server to enable | |
5823 | tracing. | |
5824 | ||
5825 | *Steve Henson* | |
5826 | ||
5827 | * New ctrl and macro to retrieve supported points extensions. | |
5828 | Print out extension in s_server and s_client. | |
5829 | ||
5830 | *Steve Henson* | |
5831 | ||
5832 | * New functions to retrieve certificate signature and signature | |
5833 | OID NID. | |
5834 | ||
5835 | *Steve Henson* | |
5836 | ||
5837 | * Add functions to retrieve and manipulate the raw cipherlist sent by a | |
5838 | client to OpenSSL. | |
5839 | ||
5840 | *Steve Henson* | |
5841 | ||
5842 | * New Suite B modes for TLS code. These use and enforce the requirements | |
5843 | of RFC6460: restrict ciphersuites, only permit Suite B algorithms and | |
5844 | only use Suite B curves. The Suite B modes can be set by using the | |
5845 | strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring. | |
5846 | ||
5847 | *Steve Henson* | |
5848 | ||
5849 | * New chain verification flags for Suite B levels of security. Check | |
5850 | algorithms are acceptable when flags are set in X509_verify_cert. | |
5851 | ||
5852 | *Steve Henson* | |
5853 | ||
5854 | * Make tls1_check_chain return a set of flags indicating checks passed | |
5855 | by a certificate chain. Add additional tests to handle client | |
5856 | certificates: checks for matching certificate type and issuer name | |
5857 | comparison. | |
5858 | ||
5859 | *Steve Henson* | |
5860 | ||
5861 | * If an attempt is made to use a signature algorithm not in the peer | |
5862 | preference list abort the handshake. If client has no suitable | |
5863 | signature algorithms in response to a certificate request do not | |
5864 | use the certificate. | |
5865 | ||
5866 | *Steve Henson* | |
5867 | ||
5868 | * If server EC tmp key is not in client preference list abort handshake. | |
5869 | ||
5870 | *Steve Henson* | |
5871 | ||
5872 | * Add support for certificate stores in CERT structure. This makes it | |
5873 | possible to have different stores per SSL structure or one store in | |
5874 | the parent SSL_CTX. Include distinct stores for certificate chain | |
5875 | verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN | |
5876 | to build and store a certificate chain in CERT structure: returning | |
5877 | an error if the chain cannot be built: this will allow applications | |
5878 | to test if a chain is correctly configured. | |
5879 | ||
5880 | Note: if the CERT based stores are not set then the parent SSL_CTX | |
5881 | store is used to retain compatibility with existing behaviour. | |
5882 | ||
5883 | *Steve Henson* | |
5884 | ||
5885 | * New function ssl_set_client_disabled to set a ciphersuite disabled | |
5886 | mask based on the current session, check mask when sending client | |
5887 | hello and checking the requested ciphersuite. | |
5888 | ||
5889 | *Steve Henson* | |
5890 | ||
5891 | * New ctrls to retrieve and set certificate types in a certificate | |
5892 | request message. Print out received values in s_client. If certificate | |
5893 | types is not set with custom values set sensible values based on | |
5894 | supported signature algorithms. | |
5895 | ||
5896 | *Steve Henson* | |
5897 | ||
5898 | * Support for distinct client and server supported signature algorithms. | |
5899 | ||
5900 | *Steve Henson* | |
5901 | ||
5902 | * Add certificate callback. If set this is called whenever a certificate | |
5903 | is required by client or server. An application can decide which | |
5904 | certificate chain to present based on arbitrary criteria: for example | |
5905 | supported signature algorithms. Add very simple example to s_server. | |
5906 | This fixes many of the problems and restrictions of the existing client | |
5907 | certificate callback: for example you can now clear an existing | |
5908 | certificate and specify the whole chain. | |
5909 | ||
5910 | *Steve Henson* | |
5911 | ||
5912 | * Add new "valid_flags" field to CERT_PKEY structure which determines what | |
5913 | the certificate can be used for (if anything). Set valid_flags field | |
5914 | in new tls1_check_chain function. Simplify ssl_set_cert_masks which used | |
5915 | to have similar checks in it. | |
5916 | ||
5917 | Add new "cert_flags" field to CERT structure and include a "strict mode". | |
5918 | This enforces some TLS certificate requirements (such as only permitting | |
5919 | certificate signature algorithms contained in the supported algorithms | |
5920 | extension) which some implementations ignore: this option should be used | |
5921 | with caution as it could cause interoperability issues. | |
5922 | ||
5923 | *Steve Henson* | |
5924 | ||
5925 | * Update and tidy signature algorithm extension processing. Work out | |
5926 | shared signature algorithms based on preferences and peer algorithms | |
5927 | and print them out in s_client and s_server. Abort handshake if no | |
5928 | shared signature algorithms. | |
5929 | ||
5930 | *Steve Henson* | |
5931 | ||
5932 | * Add new functions to allow customised supported signature algorithms | |
5933 | for SSL and SSL_CTX structures. Add options to s_client and s_server | |
5934 | to support them. | |
5935 | ||
5936 | *Steve Henson* | |
5937 | ||
5938 | * New function SSL_certs_clear() to delete all references to certificates | |
5939 | from an SSL structure. Before this once a certificate had been added | |
5940 | it couldn't be removed. | |
5941 | ||
5942 | *Steve Henson* | |
5943 | ||
5944 | * Integrate hostname, email address and IP address checking with certificate | |
5945 | verification. New verify options supporting checking in openssl utility. | |
5946 | ||
5947 | *Steve Henson* | |
5948 | ||
5949 | * Fixes and wildcard matching support to hostname and email checking | |
5950 | functions. Add manual page. | |
5951 | ||
5952 | *Florian Weimer (Red Hat Product Security Team)* | |
5953 | ||
5954 | * New functions to check a hostname email or IP address against a | |
5955 | certificate. Add options x509 utility to print results of checks against | |
5956 | a certificate. | |
5957 | ||
5958 | *Steve Henson* | |
5959 | ||
5960 | * Fix OCSP checking. | |
5961 | ||
5962 | *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie* | |
5963 | ||
5964 | * Initial experimental support for explicitly trusted non-root CAs. | |
5965 | OpenSSL still tries to build a complete chain to a root but if an | |
5966 | intermediate CA has a trust setting included that is used. The first | |
5967 | setting is used: whether to trust (e.g., -addtrust option to the x509 | |
5968 | utility) or reject. | |
5969 | ||
5970 | *Steve Henson* | |
5971 | ||
5972 | * Add -trusted_first option which attempts to find certificates in the | |
5973 | trusted store even if an untrusted chain is also supplied. | |
5974 | ||
5975 | *Steve Henson* | |
5976 | ||
5977 | * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE, | |
5978 | platform support for Linux and Android. | |
5979 | ||
5980 | *Andy Polyakov* | |
5981 | ||
5982 | * Support for linux-x32, ILP32 environment in x86_64 framework. | |
5983 | ||
5984 | *Andy Polyakov* | |
5985 | ||
5986 | * Experimental multi-implementation support for FIPS capable OpenSSL. | |
5987 | When in FIPS mode the approved implementations are used as normal, | |
5988 | when not in FIPS mode the internal unapproved versions are used instead. | |
5989 | This means that the FIPS capable OpenSSL isn't forced to use the | |
5990 | (often lower performance) FIPS implementations outside FIPS mode. | |
5991 | ||
5992 | *Steve Henson* | |
5993 | ||
5994 | * Transparently support X9.42 DH parameters when calling | |
5995 | PEM_read_bio_DHparameters. This means existing applications can handle | |
5996 | the new parameter format automatically. | |
5997 | ||
5998 | *Steve Henson* | |
5999 | ||
6000 | * Initial experimental support for X9.42 DH parameter format: mainly | |
6001 | to support use of 'q' parameter for RFC5114 parameters. | |
6002 | ||
6003 | *Steve Henson* | |
6004 | ||
6005 | * Add DH parameters from RFC5114 including test data to dhtest. | |
6006 | ||
6007 | *Steve Henson* | |
6008 | ||
6009 | * Support for automatic EC temporary key parameter selection. If enabled | |
6010 | the most preferred EC parameters are automatically used instead of | |
6011 | hardcoded fixed parameters. Now a server just has to call: | |
6012 | SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically | |
6013 | support ECDH and use the most appropriate parameters. | |
6014 | ||
6015 | *Steve Henson* | |
6016 | ||
6017 | * Enhance and tidy EC curve and point format TLS extension code. Use | |
6018 | static structures instead of allocation if default values are used. | |
6019 | New ctrls to set curves we wish to support and to retrieve shared curves. | |
6020 | Print out shared curves in s_server. New options to s_server and s_client | |
6021 | to set list of supported curves. | |
6022 | ||
6023 | *Steve Henson* | |
6024 | ||
6025 | * New ctrls to retrieve supported signature algorithms and | |
6026 | supported curve values as an array of NIDs. Extend openssl utility | |
6027 | to print out received values. | |
6028 | ||
6029 | *Steve Henson* | |
6030 | ||
6031 | * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert | |
6032 | between NIDs and the more common NIST names such as "P-256". Enhance | |
6033 | ecparam utility and ECC method to recognise the NIST names for curves. | |
6034 | ||
6035 | *Steve Henson* | |
6036 | ||
6037 | * Enhance SSL/TLS certificate chain handling to support different | |
6038 | chains for each certificate instead of one chain in the parent SSL_CTX. | |
6039 | ||
6040 | *Steve Henson* | |
6041 | ||
6042 | * Support for fixed DH ciphersuite client authentication: where both | |
6043 | server and client use DH certificates with common parameters. | |
6044 | ||
6045 | *Steve Henson* | |
6046 | ||
6047 | * Support for fixed DH ciphersuites: those requiring DH server | |
6048 | certificates. | |
6049 | ||
6050 | *Steve Henson* | |
6051 | ||
6052 | * New function i2d_re_X509_tbs for re-encoding the TBS portion of | |
6053 | the certificate. | |
6054 | Note: Related 1.0.2-beta specific macros X509_get_cert_info, | |
6055 | X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and | |
6056 | X509_CINF_get_signature were reverted post internal team review. | |
6057 | ||
6058 | OpenSSL 1.0.1 | |
6059 | ------------- | |
6060 | ||
6061 | ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016] | |
6062 | ||
6063 | * OCSP Status Request extension unbounded memory growth | |
6064 | ||
6065 | A malicious client can send an excessively large OCSP Status Request | |
6066 | extension. If that client continually requests renegotiation, sending a | |
6067 | large OCSP Status Request extension each time, then there will be unbounded | |
6068 | memory growth on the server. This will eventually lead to a Denial Of | |
6069 | Service attack through memory exhaustion. Servers with a default | |
6070 | configuration are vulnerable even if they do not support OCSP. Builds using | |
6071 | the "no-ocsp" build time option are not affected. | |
6072 | ||
6073 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
6074 | ([CVE-2016-6304]) | |
6075 | ||
6076 | *Matt Caswell* | |
6077 | ||
6078 | * In order to mitigate the SWEET32 attack, the DES ciphers were moved from | |
6079 | HIGH to MEDIUM. | |
6080 | ||
6081 | This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan | |
6082 | Leurent (INRIA) | |
6083 | ([CVE-2016-2183]) | |
6084 | ||
6085 | *Rich Salz* | |
6086 | ||
6087 | * OOB write in MDC2_Update() | |
6088 | ||
6089 | An overflow can occur in MDC2_Update() either if called directly or | |
6090 | through the EVP_DigestUpdate() function using MDC2. If an attacker | |
6091 | is able to supply very large amounts of input data after a previous | |
6092 | call to EVP_EncryptUpdate() with a partial block then a length check | |
6093 | can overflow resulting in a heap corruption. | |
6094 | ||
6095 | The amount of data needed is comparable to SIZE_MAX which is impractical | |
6096 | on most platforms. | |
6097 | ||
6098 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
6099 | ([CVE-2016-6303]) | |
6100 | ||
6101 | *Stephen Henson* | |
6102 | ||
6103 | * Malformed SHA512 ticket DoS | |
6104 | ||
6105 | If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a | |
6106 | DoS attack where a malformed ticket will result in an OOB read which will | |
6107 | ultimately crash. | |
6108 | ||
6109 | The use of SHA512 in TLS session tickets is comparatively rare as it requires | |
6110 | a custom server callback and ticket lookup mechanism. | |
6111 | ||
6112 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
6113 | ([CVE-2016-6302]) | |
6114 | ||
6115 | *Stephen Henson* | |
6116 | ||
6117 | * OOB write in BN_bn2dec() | |
6118 | ||
6119 | The function BN_bn2dec() does not check the return value of BN_div_word(). | |
6120 | This can cause an OOB write if an application uses this function with an | |
6121 | overly large BIGNUM. This could be a problem if an overly large certificate | |
6122 | or CRL is printed out from an untrusted source. TLS is not affected because | |
6123 | record limits will reject an oversized certificate before it is parsed. | |
6124 | ||
6125 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
6126 | ([CVE-2016-2182]) | |
6127 | ||
6128 | *Stephen Henson* | |
6129 | ||
6130 | * OOB read in TS_OBJ_print_bio() | |
6131 | ||
6132 | The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is | |
6133 | the total length the OID text representation would use and not the amount | |
6134 | of data written. This will result in OOB reads when large OIDs are | |
6135 | presented. | |
6136 | ||
6137 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
6138 | ([CVE-2016-2180]) | |
6139 | ||
6140 | *Stephen Henson* | |
6141 | ||
6142 | * Pointer arithmetic undefined behaviour | |
6143 | ||
6144 | Avoid some undefined pointer arithmetic | |
6145 | ||
6146 | A common idiom in the codebase is to check limits in the following manner: | |
6147 | "p + len > limit" | |
6148 | ||
6149 | Where "p" points to some malloc'd data of SIZE bytes and | |
6150 | limit == p + SIZE | |
6151 | ||
6152 | "len" here could be from some externally supplied data (e.g. from a TLS | |
6153 | message). | |
6154 | ||
6155 | The rules of C pointer arithmetic are such that "p + len" is only well | |
6156 | defined where len <= SIZE. Therefore the above idiom is actually | |
6157 | undefined behaviour. | |
6158 | ||
6159 | For example this could cause problems if some malloc implementation | |
6160 | provides an address for "p" such that "p + len" actually overflows for | |
6161 | values of len that are too big and therefore p + len < limit. | |
6162 | ||
6163 | This issue was reported to OpenSSL by Guido Vranken | |
6164 | ([CVE-2016-2177]) | |
6165 | ||
6166 | *Matt Caswell* | |
6167 | ||
6168 | * Constant time flag not preserved in DSA signing | |
6169 | ||
6170 | Operations in the DSA signing algorithm should run in constant time in | |
6171 | order to avoid side channel attacks. A flaw in the OpenSSL DSA | |
6172 | implementation means that a non-constant time codepath is followed for | |
6173 | certain operations. This has been demonstrated through a cache-timing | |
6174 | attack to be sufficient for an attacker to recover the private DSA key. | |
6175 | ||
6176 | This issue was reported by César Pereida (Aalto University), Billy Brumley | |
6177 | (Tampere University of Technology), and Yuval Yarom (The University of | |
6178 | Adelaide and NICTA). | |
6179 | ([CVE-2016-2178]) | |
6180 | ||
6181 | *César Pereida* | |
6182 | ||
6183 | * DTLS buffered message DoS | |
6184 | ||
6185 | In a DTLS connection where handshake messages are delivered out-of-order | |
6186 | those messages that OpenSSL is not yet ready to process will be buffered | |
6187 | for later use. Under certain circumstances, a flaw in the logic means that | |
6188 | those messages do not get removed from the buffer even though the handshake | |
6189 | has been completed. An attacker could force up to approx. 15 messages to | |
6190 | remain in the buffer when they are no longer required. These messages will | |
6191 | be cleared when the DTLS connection is closed. The default maximum size for | |
6192 | a message is 100k. Therefore the attacker could force an additional 1500k | |
6193 | to be consumed per connection. By opening many simulataneous connections an | |
6194 | attacker could cause a DoS attack through memory exhaustion. | |
6195 | ||
6196 | This issue was reported to OpenSSL by Quan Luo. | |
6197 | ([CVE-2016-2179]) | |
6198 | ||
6199 | *Matt Caswell* | |
6200 | ||
6201 | * DTLS replay protection DoS | |
6202 | ||
6203 | A flaw in the DTLS replay attack protection mechanism means that records | |
6204 | that arrive for future epochs update the replay protection "window" before | |
6205 | the MAC for the record has been validated. This could be exploited by an | |
6206 | attacker by sending a record for the next epoch (which does not have to | |
6207 | decrypt or have a valid MAC), with a very large sequence number. This means | |
6208 | that all subsequent legitimate packets are dropped causing a denial of | |
6209 | service for a specific DTLS connection. | |
6210 | ||
6211 | This issue was reported to OpenSSL by the OCAP audit team. | |
6212 | ([CVE-2016-2181]) | |
6213 | ||
6214 | *Matt Caswell* | |
6215 | ||
6216 | * Certificate message OOB reads | |
6217 | ||
6218 | In OpenSSL 1.0.2 and earlier some missing message length checks can result | |
6219 | in OOB reads of up to 2 bytes beyond an allocated buffer. There is a | |
6220 | theoretical DoS risk but this has not been observed in practice on common | |
6221 | platforms. | |
6222 | ||
6223 | The messages affected are client certificate, client certificate request | |
6224 | and server certificate. As a result the attack can only be performed | |
6225 | against a client or a server which enables client authentication. | |
6226 | ||
6227 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
6228 | ([CVE-2016-6306]) | |
6229 | ||
6230 | *Stephen Henson* | |
6231 | ||
6232 | ### Changes between 1.0.1s and 1.0.1t [3 May 2016] | |
6233 | ||
6234 | * Prevent padding oracle in AES-NI CBC MAC check | |
6235 | ||
6236 | A MITM attacker can use a padding oracle attack to decrypt traffic | |
6237 | when the connection uses an AES CBC cipher and the server support | |
6238 | AES-NI. | |
6239 | ||
6240 | This issue was introduced as part of the fix for Lucky 13 padding | |
6241 | attack ([CVE-2013-0169]). The padding check was rewritten to be in | |
6242 | constant time by making sure that always the same bytes are read and | |
6243 | compared against either the MAC or padding bytes. But it no longer | |
6244 | checked that there was enough data to have both the MAC and padding | |
6245 | bytes. | |
6246 | ||
6247 | This issue was reported by Juraj Somorovsky using TLS-Attacker. | |
6248 | ([CVE-2016-2107]) | |
6249 | ||
6250 | *Kurt Roeckx* | |
6251 | ||
6252 | * Fix EVP_EncodeUpdate overflow | |
6253 | ||
6254 | An overflow can occur in the EVP_EncodeUpdate() function which is used for | |
6255 | Base64 encoding of binary data. If an attacker is able to supply very large | |
6256 | amounts of input data then a length check can overflow resulting in a heap | |
6257 | corruption. | |
6258 | ||
6259 | Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by | |
6260 | the `PEM_write_bio*` family of functions. These are mainly used within the | |
6261 | OpenSSL command line applications, so any application which processes data | |
6262 | from an untrusted source and outputs it as a PEM file should be considered | |
6263 | vulnerable to this issue. User applications that call these APIs directly | |
6264 | with large amounts of untrusted data may also be vulnerable. | |
6265 | ||
6266 | This issue was reported by Guido Vranken. | |
6267 | ([CVE-2016-2105]) | |
6268 | ||
6269 | *Matt Caswell* | |
6270 | ||
6271 | * Fix EVP_EncryptUpdate overflow | |
6272 | ||
6273 | An overflow can occur in the EVP_EncryptUpdate() function. If an attacker | |
6274 | is able to supply very large amounts of input data after a previous call to | |
6275 | EVP_EncryptUpdate() with a partial block then a length check can overflow | |
6276 | resulting in a heap corruption. Following an analysis of all OpenSSL | |
6277 | internal usage of the EVP_EncryptUpdate() function all usage is one of two | |
6278 | forms. The first form is where the EVP_EncryptUpdate() call is known to be | |
6279 | the first called function after an EVP_EncryptInit(), and therefore that | |
6280 | specific call must be safe. The second form is where the length passed to | |
6281 | EVP_EncryptUpdate() can be seen from the code to be some small value and | |
6282 | therefore there is no possibility of an overflow. Since all instances are | |
6283 | one of these two forms, it is believed that there can be no overflows in | |
6284 | internal code due to this problem. It should be noted that | |
6285 | EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths. | |
6286 | Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances | |
6287 | of these calls have also been analysed too and it is believed there are no | |
6288 | instances in internal usage where an overflow could occur. | |
6289 | ||
6290 | This issue was reported by Guido Vranken. | |
6291 | ([CVE-2016-2106]) | |
6292 | ||
6293 | *Matt Caswell* | |
6294 | ||
6295 | * Prevent ASN.1 BIO excessive memory allocation | |
6296 | ||
6297 | When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() | |
6298 | a short invalid encoding can casuse allocation of large amounts of memory | |
6299 | potentially consuming excessive resources or exhausting memory. | |
6300 | ||
6301 | Any application parsing untrusted data through d2i BIO functions is | |
6302 | affected. The memory based functions such as d2i_X509() are *not* affected. | |
6303 | Since the memory based functions are used by the TLS library, TLS | |
6304 | applications are not affected. | |
6305 | ||
6306 | This issue was reported by Brian Carpenter. | |
6307 | ([CVE-2016-2109]) | |
6308 | ||
6309 | *Stephen Henson* | |
6310 | ||
6311 | * EBCDIC overread | |
6312 | ||
6313 | ASN1 Strings that are over 1024 bytes can cause an overread in applications | |
6314 | using the X509_NAME_oneline() function on EBCDIC systems. This could result | |
6315 | in arbitrary stack data being returned in the buffer. | |
6316 | ||
6317 | This issue was reported by Guido Vranken. | |
6318 | ([CVE-2016-2176]) | |
6319 | ||
6320 | *Matt Caswell* | |
6321 | ||
6322 | * Modify behavior of ALPN to invoke callback after SNI/servername | |
6323 | callback, such that updates to the SSL_CTX affect ALPN. | |
6324 | ||
6325 | *Todd Short* | |
6326 | ||
6327 | * Remove LOW from the DEFAULT cipher list. This removes singles DES from the | |
6328 | default. | |
6329 | ||
6330 | *Kurt Roeckx* | |
6331 | ||
6332 | * Only remove the SSLv2 methods with the no-ssl2-method option. When the | |
6333 | methods are enabled and ssl2 is disabled the methods return NULL. | |
6334 | ||
6335 | *Kurt Roeckx* | |
6336 | ||
6337 | ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016] | |
6338 | ||
6339 | * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL. | |
6340 | Builds that are not configured with "enable-weak-ssl-ciphers" will not | |
6341 | provide any "EXPORT" or "LOW" strength ciphers. | |
6342 | ||
6343 | *Viktor Dukhovni* | |
6344 | ||
6345 | * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2 | |
6346 | is by default disabled at build-time. Builds that are not configured with | |
6347 | "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used, | |
6348 | users who want to negotiate SSLv2 via the version-flexible SSLv23_method() | |
6349 | will need to explicitly call either of: | |
6350 | ||
6351 | SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2); | |
6352 | or | |
6353 | SSL_clear_options(ssl, SSL_OP_NO_SSLv2); | |
6354 | ||
6355 | as appropriate. Even if either of those is used, or the application | |
6356 | explicitly uses the version-specific SSLv2_method() or its client and | |
6357 | server variants, SSLv2 ciphers vulnerable to exhaustive search key | |
6358 | recovery have been removed. Specifically, the SSLv2 40-bit EXPORT | |
6359 | ciphers, and SSLv2 56-bit DES are no longer available. | |
6360 | ([CVE-2016-0800]) | |
6361 | ||
6362 | *Viktor Dukhovni* | |
6363 | ||
6364 | * Fix a double-free in DSA code | |
6365 | ||
6366 | A double free bug was discovered when OpenSSL parses malformed DSA private | |
6367 | keys and could lead to a DoS attack or memory corruption for applications | |
6368 | that receive DSA private keys from untrusted sources. This scenario is | |
6369 | considered rare. | |
6370 | ||
6371 | This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using | |
6372 | libFuzzer. | |
6373 | ([CVE-2016-0705]) | |
6374 | ||
6375 | *Stephen Henson* | |
6376 | ||
6377 | * Disable SRP fake user seed to address a server memory leak. | |
6378 | ||
6379 | Add a new method SRP_VBASE_get1_by_user that handles the seed properly. | |
6380 | ||
6381 | SRP_VBASE_get_by_user had inconsistent memory management behaviour. | |
6382 | In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user | |
6383 | was changed to ignore the "fake user" SRP seed, even if the seed | |
6384 | is configured. | |
6385 | ||
6386 | Users should use SRP_VBASE_get1_by_user instead. Note that in | |
6387 | SRP_VBASE_get1_by_user, caller must free the returned value. Note | |
6388 | also that even though configuring the SRP seed attempts to hide | |
6389 | invalid usernames by continuing the handshake with fake | |
6390 | credentials, this behaviour is not constant time and no strong | |
6391 | guarantees are made that the handshake is indistinguishable from | |
6392 | that of a valid user. | |
6393 | ([CVE-2016-0798]) | |
6394 | ||
6395 | *Emilia Käsper* | |
6396 | ||
6397 | * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption | |
6398 | ||
6399 | In the BN_hex2bn function the number of hex digits is calculated using an | |
6400 | int value `i`. Later `bn_expand` is called with a value of `i * 4`. For | |
6401 | large values of `i` this can result in `bn_expand` not allocating any | |
6402 | memory because `i * 4` is negative. This can leave the internal BIGNUM data | |
6403 | field as NULL leading to a subsequent NULL ptr deref. For very large values | |
6404 | of `i`, the calculation `i * 4` could be a positive value smaller than `i`. | |
6405 | In this case memory is allocated to the internal BIGNUM data field, but it | |
6406 | is insufficiently sized leading to heap corruption. A similar issue exists | |
6407 | in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn | |
6408 | is ever called by user applications with very large untrusted hex/dec data. | |
6409 | This is anticipated to be a rare occurrence. | |
6410 | ||
6411 | All OpenSSL internal usage of these functions use data that is not expected | |
6412 | to be untrusted, e.g. config file data or application command line | |
6413 | arguments. If user developed applications generate config file data based | |
6414 | on untrusted data then it is possible that this could also lead to security | |
6415 | consequences. This is also anticipated to be rare. | |
6416 | ||
6417 | This issue was reported to OpenSSL by Guido Vranken. | |
6418 | ([CVE-2016-0797]) | |
6419 | ||
6420 | *Matt Caswell* | |
6421 | ||
6422 | * Fix memory issues in `BIO_*printf` functions | |
6423 | ||
6424 | The internal `fmtstr` function used in processing a "%s" format string in | |
6425 | the `BIO_*printf` functions could overflow while calculating the length of a | |
6426 | string and cause an OOB read when printing very long strings. | |
6427 | ||
6428 | Additionally the internal `doapr_outch` function can attempt to write to an | |
6429 | OOB memory location (at an offset from the NULL pointer) in the event of a | |
6430 | memory allocation failure. In 1.0.2 and below this could be caused where | |
6431 | the size of a buffer to be allocated is greater than INT_MAX. E.g. this | |
6432 | could be in processing a very long "%s" format string. Memory leaks can | |
6433 | also occur. | |
6434 | ||
6435 | The first issue may mask the second issue dependent on compiler behaviour. | |
6436 | These problems could enable attacks where large amounts of untrusted data | |
6437 | is passed to the `BIO_*printf` functions. If applications use these functions | |
6438 | in this way then they could be vulnerable. OpenSSL itself uses these | |
6439 | functions when printing out human-readable dumps of ASN.1 data. Therefore | |
6440 | applications that print this data could be vulnerable if the data is from | |
6441 | untrusted sources. OpenSSL command line applications could also be | |
6442 | vulnerable where they print out ASN.1 data, or if untrusted data is passed | |
6443 | as command line arguments. | |
6444 | ||
6445 | Libssl is not considered directly vulnerable. Additionally certificates etc | |
6446 | received via remote connections via libssl are also unlikely to be able to | |
6447 | trigger these issues because of message size limits enforced within libssl. | |
6448 | ||
6449 | This issue was reported to OpenSSL Guido Vranken. | |
6450 | ([CVE-2016-0799]) | |
6451 | ||
6452 | *Matt Caswell* | |
6453 | ||
6454 | * Side channel attack on modular exponentiation | |
6455 | ||
6456 | A side-channel attack was found which makes use of cache-bank conflicts on | |
6457 | the Intel Sandy-Bridge microarchitecture which could lead to the recovery | |
6458 | of RSA keys. The ability to exploit this issue is limited as it relies on | |
6459 | an attacker who has control of code in a thread running on the same | |
6460 | hyper-threaded core as the victim thread which is performing decryptions. | |
6461 | ||
6462 | This issue was reported to OpenSSL by Yuval Yarom, The University of | |
6463 | Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and | |
6464 | Nadia Heninger, University of Pennsylvania with more information at | |
6465 | <http://cachebleed.info>. | |
6466 | ([CVE-2016-0702]) | |
6467 | ||
6468 | *Andy Polyakov* | |
6469 | ||
6470 | * Change the req command to generate a 2048-bit RSA/DSA key by default, | |
6471 | if no keysize is specified with default_bits. This fixes an | |
6472 | omission in an earlier change that changed all RSA/DSA key generation | |
6473 | commands to use 2048 bits by default. | |
6474 | ||
6475 | *Emilia Käsper* | |
6476 | ||
6477 | ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016] | |
6478 | ||
6479 | * Protection for DH small subgroup attacks | |
6480 | ||
6481 | As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been | |
6482 | switched on by default and cannot be disabled. This could have some | |
6483 | performance impact. | |
6484 | ||
6485 | *Matt Caswell* | |
6486 | ||
6487 | * SSLv2 doesn't block disabled ciphers | |
6488 | ||
6489 | A malicious client can negotiate SSLv2 ciphers that have been disabled on | |
6490 | the server and complete SSLv2 handshakes even if all SSLv2 ciphers have | |
6491 | been disabled, provided that the SSLv2 protocol was not also disabled via | |
6492 | SSL_OP_NO_SSLv2. | |
6493 | ||
6494 | This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram | |
6495 | and Sebastian Schinzel. | |
6496 | ([CVE-2015-3197]) | |
6497 | ||
6498 | *Viktor Dukhovni* | |
6499 | ||
6500 | * Reject DH handshakes with parameters shorter than 1024 bits. | |
6501 | ||
6502 | *Kurt Roeckx* | |
6503 | ||
6504 | ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015] | |
6505 | ||
6506 | * Certificate verify crash with missing PSS parameter | |
6507 | ||
6508 | The signature verification routines will crash with a NULL pointer | |
6509 | dereference if presented with an ASN.1 signature using the RSA PSS | |
6510 | algorithm and absent mask generation function parameter. Since these | |
6511 | routines are used to verify certificate signature algorithms this can be | |
6512 | used to crash any certificate verification operation and exploited in a | |
6513 | DoS attack. Any application which performs certificate verification is | |
6514 | vulnerable including OpenSSL clients and servers which enable client | |
6515 | authentication. | |
6516 | ||
6517 | This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG). | |
6518 | ([CVE-2015-3194]) | |
6519 | ||
6520 | *Stephen Henson* | |
6521 | ||
6522 | * X509_ATTRIBUTE memory leak | |
6523 | ||
6524 | When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak | |
6525 | memory. This structure is used by the PKCS#7 and CMS routines so any | |
6526 | application which reads PKCS#7 or CMS data from untrusted sources is | |
6527 | affected. SSL/TLS is not affected. | |
6528 | ||
6529 | This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using | |
6530 | libFuzzer. | |
6531 | ([CVE-2015-3195]) | |
6532 | ||
6533 | *Stephen Henson* | |
6534 | ||
6535 | * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. | |
6536 | This changes the decoding behaviour for some invalid messages, | |
6537 | though the change is mostly in the more lenient direction, and | |
6538 | legacy behaviour is preserved as much as possible. | |
6539 | ||
6540 | *Emilia Käsper* | |
6541 | ||
6542 | * In DSA_generate_parameters_ex, if the provided seed is too short, | |
6543 | use a random seed, as already documented. | |
6544 | ||
6545 | *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>* | |
6546 | ||
6547 | ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015] | |
6548 | ||
6549 | * Alternate chains certificate forgery | |
6550 | ||
6551 | During certificate verfification, OpenSSL will attempt to find an | |
6552 | alternative certificate chain if the first attempt to build such a chain | |
6553 | fails. An error in the implementation of this logic can mean that an | |
6554 | attacker could cause certain checks on untrusted certificates to be | |
6555 | bypassed, such as the CA flag, enabling them to use a valid leaf | |
6556 | certificate to act as a CA and "issue" an invalid certificate. | |
6557 | ||
6558 | This issue was reported to OpenSSL by Adam Langley/David Benjamin | |
6559 | (Google/BoringSSL). | |
6560 | ([CVE-2015-1793]) | |
6561 | ||
6562 | *Matt Caswell* | |
6563 | ||
6564 | * Race condition handling PSK identify hint | |
6565 | ||
6566 | If PSK identity hints are received by a multi-threaded client then | |
6567 | the values are wrongly updated in the parent SSL_CTX structure. This can | |
6568 | result in a race condition potentially leading to a double free of the | |
6569 | identify hint data. | |
6570 | ([CVE-2015-3196]) | |
6571 | ||
6572 | *Stephen Henson* | |
6573 | ||
6574 | ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015] | |
6575 | ||
6576 | * Fix HMAC ABI incompatibility. The previous version introduced an ABI | |
6577 | incompatibility in the handling of HMAC. The previous ABI has now been | |
6578 | restored. | |
6579 | ||
6580 | ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015] | |
6581 | ||
6582 | * Malformed ECParameters causes infinite loop | |
6583 | ||
6584 | When processing an ECParameters structure OpenSSL enters an infinite loop | |
6585 | if the curve specified is over a specially malformed binary polynomial | |
6586 | field. | |
6587 | ||
6588 | This can be used to perform denial of service against any | |
6589 | system which processes public keys, certificate requests or | |
6590 | certificates. This includes TLS clients and TLS servers with | |
6591 | client authentication enabled. | |
6592 | ||
6593 | This issue was reported to OpenSSL by Joseph Barr-Pixton. | |
6594 | ([CVE-2015-1788]) | |
6595 | ||
6596 | *Andy Polyakov* | |
6597 | ||
6598 | * Exploitable out-of-bounds read in X509_cmp_time | |
6599 | ||
6600 | X509_cmp_time does not properly check the length of the ASN1_TIME | |
6601 | string and can read a few bytes out of bounds. In addition, | |
6602 | X509_cmp_time accepts an arbitrary number of fractional seconds in the | |
6603 | time string. | |
6604 | ||
6605 | An attacker can use this to craft malformed certificates and CRLs of | |
6606 | various sizes and potentially cause a segmentation fault, resulting in | |
6607 | a DoS on applications that verify certificates or CRLs. TLS clients | |
6608 | that verify CRLs are affected. TLS clients and servers with client | |
6609 | authentication enabled may be affected if they use custom verification | |
6610 | callbacks. | |
6611 | ||
6612 | This issue was reported to OpenSSL by Robert Swiecki (Google), and | |
6613 | independently by Hanno Böck. | |
6614 | ([CVE-2015-1789]) | |
6615 | ||
6616 | *Emilia Käsper* | |
6617 | ||
6618 | * PKCS7 crash with missing EnvelopedContent | |
6619 | ||
6620 | The PKCS#7 parsing code does not handle missing inner EncryptedContent | |
6621 | correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs | |
6622 | with missing content and trigger a NULL pointer dereference on parsing. | |
6623 | ||
6624 | Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 | |
6625 | structures from untrusted sources are affected. OpenSSL clients and | |
6626 | servers are not affected. | |
6627 | ||
6628 | This issue was reported to OpenSSL by Michal Zalewski (Google). | |
6629 | ([CVE-2015-1790]) | |
6630 | ||
6631 | *Emilia Käsper* | |
6632 | ||
6633 | * CMS verify infinite loop with unknown hash function | |
6634 | ||
6635 | When verifying a signedData message the CMS code can enter an infinite loop | |
6636 | if presented with an unknown hash function OID. This can be used to perform | |
6637 | denial of service against any system which verifies signedData messages using | |
6638 | the CMS code. | |
6639 | This issue was reported to OpenSSL by Johannes Bauer. | |
6640 | ([CVE-2015-1792]) | |
6641 | ||
6642 | *Stephen Henson* | |
6643 | ||
6644 | * Race condition handling NewSessionTicket | |
6645 | ||
6646 | If a NewSessionTicket is received by a multi-threaded client when attempting to | |
6647 | reuse a previous ticket then a race condition can occur potentially leading to | |
6648 | a double free of the ticket data. | |
6649 | ([CVE-2015-1791]) | |
6650 | ||
6651 | *Matt Caswell* | |
6652 | ||
6653 | * Reject DH handshakes with parameters shorter than 768 bits. | |
6654 | ||
6655 | *Kurt Roeckx and Emilia Kasper* | |
6656 | ||
6657 | * dhparam: generate 2048-bit parameters by default. | |
6658 | ||
6659 | *Kurt Roeckx and Emilia Kasper* | |
6660 | ||
6661 | ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015] | |
6662 | ||
6663 | * Segmentation fault in ASN1_TYPE_cmp fix | |
6664 | ||
6665 | The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is | |
6666 | made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check | |
6667 | certificate signature algorithm consistency this can be used to crash any | |
6668 | certificate verification operation and exploited in a DoS attack. Any | |
6669 | application which performs certificate verification is vulnerable including | |
6670 | OpenSSL clients and servers which enable client authentication. | |
6671 | ([CVE-2015-0286]) | |
6672 | ||
6673 | *Stephen Henson* | |
6674 | ||
6675 | * ASN.1 structure reuse memory corruption fix | |
6676 | ||
6677 | Reusing a structure in ASN.1 parsing may allow an attacker to cause | |
6678 | memory corruption via an invalid write. Such reuse is and has been | |
6679 | strongly discouraged and is believed to be rare. | |
6680 | ||
6681 | Applications that parse structures containing CHOICE or ANY DEFINED BY | |
6682 | components may be affected. Certificate parsing (d2i_X509 and related | |
6683 | functions) are however not affected. OpenSSL clients and servers are | |
6684 | not affected. | |
6685 | ([CVE-2015-0287]) | |
6686 | ||
6687 | *Stephen Henson* | |
6688 | ||
6689 | * PKCS7 NULL pointer dereferences fix | |
6690 | ||
6691 | The PKCS#7 parsing code does not handle missing outer ContentInfo | |
6692 | correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with | |
6693 | missing content and trigger a NULL pointer dereference on parsing. | |
6694 | ||
6695 | Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or | |
6696 | otherwise parse PKCS#7 structures from untrusted sources are | |
6697 | affected. OpenSSL clients and servers are not affected. | |
6698 | ||
6699 | This issue was reported to OpenSSL by Michal Zalewski (Google). | |
6700 | ([CVE-2015-0289]) | |
6701 | ||
6702 | *Emilia Käsper* | |
6703 | ||
6704 | * DoS via reachable assert in SSLv2 servers fix | |
6705 | ||
6706 | A malicious client can trigger an OPENSSL_assert (i.e., an abort) in | |
6707 | servers that both support SSLv2 and enable export cipher suites by sending | |
6708 | a specially crafted SSLv2 CLIENT-MASTER-KEY message. | |
6709 | ||
6710 | This issue was discovered by Sean Burford (Google) and Emilia Käsper | |
6711 | (OpenSSL development team). | |
6712 | ([CVE-2015-0293]) | |
6713 | ||
6714 | *Emilia Käsper* | |
6715 | ||
6716 | * Use After Free following d2i_ECPrivatekey error fix | |
6717 | ||
6718 | A malformed EC private key file consumed via the d2i_ECPrivateKey function | |
6719 | could cause a use after free condition. This, in turn, could cause a double | |
6720 | free in several private key parsing functions (such as d2i_PrivateKey | |
6721 | or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption | |
6722 | for applications that receive EC private keys from untrusted | |
6723 | sources. This scenario is considered rare. | |
6724 | ||
6725 | This issue was discovered by the BoringSSL project and fixed in their | |
6726 | commit 517073cd4b. | |
6727 | ([CVE-2015-0209]) | |
6728 | ||
6729 | *Matt Caswell* | |
6730 | ||
6731 | * X509_to_X509_REQ NULL pointer deref fix | |
6732 | ||
6733 | The function X509_to_X509_REQ will crash with a NULL pointer dereference if | |
6734 | the certificate key is invalid. This function is rarely used in practice. | |
6735 | ||
6736 | This issue was discovered by Brian Carpenter. | |
6737 | ([CVE-2015-0288]) | |
6738 | ||
6739 | *Stephen Henson* | |
6740 | ||
6741 | * Removed the export ciphers from the DEFAULT ciphers | |
6742 | ||
6743 | *Kurt Roeckx* | |
6744 | ||
6745 | ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015] | |
6746 | ||
6747 | * Build fixes for the Windows and OpenVMS platforms | |
6748 | ||
6749 | *Matt Caswell and Richard Levitte* | |
6750 | ||
6751 | ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015] | |
6752 | ||
6753 | * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS | |
6754 | message can cause a segmentation fault in OpenSSL due to a NULL pointer | |
6755 | dereference. This could lead to a Denial Of Service attack. Thanks to | |
6756 | Markus Stenberg of Cisco Systems, Inc. for reporting this issue. | |
6757 | ([CVE-2014-3571]) | |
6758 | ||
6759 | *Steve Henson* | |
6760 | ||
6761 | * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the | |
6762 | dtls1_buffer_record function under certain conditions. In particular this | |
6763 | could occur if an attacker sent repeated DTLS records with the same | |
6764 | sequence number but for the next epoch. The memory leak could be exploited | |
6765 | by an attacker in a Denial of Service attack through memory exhaustion. | |
6766 | Thanks to Chris Mueller for reporting this issue. | |
6767 | ([CVE-2015-0206]) | |
6768 | ||
6769 | *Matt Caswell* | |
6770 | ||
6771 | * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is | |
6772 | built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl | |
6773 | method would be set to NULL which could later result in a NULL pointer | |
6774 | dereference. Thanks to Frank Schmirler for reporting this issue. | |
6775 | ([CVE-2014-3569]) | |
6776 | ||
6777 | *Kurt Roeckx* | |
6778 | ||
6779 | * Abort handshake if server key exchange message is omitted for ephemeral | |
6780 | ECDH ciphersuites. | |
6781 | ||
6782 | Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for | |
6783 | reporting this issue. | |
6784 | ([CVE-2014-3572]) | |
6785 | ||
6786 | *Steve Henson* | |
6787 | ||
6788 | * Remove non-export ephemeral RSA code on client and server. This code | |
6789 | violated the TLS standard by allowing the use of temporary RSA keys in | |
6790 | non-export ciphersuites and could be used by a server to effectively | |
6791 | downgrade the RSA key length used to a value smaller than the server | |
6792 | certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at | |
6793 | INRIA or reporting this issue. | |
6794 | ([CVE-2015-0204]) | |
6795 | ||
6796 | *Steve Henson* | |
6797 | ||
6798 | * Fixed issue where DH client certificates are accepted without verification. | |
6799 | An OpenSSL server will accept a DH certificate for client authentication | |
6800 | without the certificate verify message. This effectively allows a client to | |
6801 | authenticate without the use of a private key. This only affects servers | |
6802 | which trust a client certificate authority which issues certificates | |
6803 | containing DH keys: these are extremely rare and hardly ever encountered. | |
6804 | Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting | |
6805 | this issue. | |
6806 | ([CVE-2015-0205]) | |
6807 | ||
6808 | *Steve Henson* | |
6809 | ||
6810 | * Ensure that the session ID context of an SSL is updated when its | |
6811 | SSL_CTX is updated via SSL_set_SSL_CTX. | |
6812 | ||
6813 | The session ID context is typically set from the parent SSL_CTX, | |
6814 | and can vary with the CTX. | |
6815 | ||
6816 | *Adam Langley* | |
6817 | ||
6818 | * Fix various certificate fingerprint issues. | |
6819 | ||
6820 | By using non-DER or invalid encodings outside the signed portion of a | |
6821 | certificate the fingerprint can be changed without breaking the signature. | |
6822 | Although no details of the signed portion of the certificate can be changed | |
6823 | this can cause problems with some applications: e.g. those using the | |
6824 | certificate fingerprint for blacklists. | |
6825 | ||
6826 | 1. Reject signatures with non zero unused bits. | |
6827 | ||
6828 | If the BIT STRING containing the signature has non zero unused bits reject | |
6829 | the signature. All current signature algorithms require zero unused bits. | |
6830 | ||
6831 | 2. Check certificate algorithm consistency. | |
6832 | ||
6833 | Check the AlgorithmIdentifier inside TBS matches the one in the | |
6834 | certificate signature. NB: this will result in signature failure | |
6835 | errors for some broken certificates. | |
6836 | ||
6837 | Thanks to Konrad Kraszewski from Google for reporting this issue. | |
6838 | ||
6839 | 3. Check DSA/ECDSA signatures use DER. | |
6840 | ||
6841 | Re-encode DSA/ECDSA signatures and compare with the original received | |
6842 | signature. Return an error if there is a mismatch. | |
6843 | ||
6844 | This will reject various cases including garbage after signature | |
6845 | (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS | |
6846 | program for discovering this case) and use of BER or invalid ASN.1 INTEGERs | |
6847 | (negative or with leading zeroes). | |
6848 | ||
6849 | Further analysis was conducted and fixes were developed by Stephen Henson | |
6850 | of the OpenSSL core team. | |
6851 | ||
6852 | ([CVE-2014-8275]) | |
6853 | ||
6854 | *Steve Henson* | |
6855 | ||
6856 | * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect | |
6857 | results on some platforms, including x86_64. This bug occurs at random | |
6858 | with a very low probability, and is not known to be exploitable in any | |
6859 | way, though its exact impact is difficult to determine. Thanks to Pieter | |
6860 | Wuille (Blockstream) who reported this issue and also suggested an initial | |
6861 | fix. Further analysis was conducted by the OpenSSL development team and | |
6862 | Adam Langley of Google. The final fix was developed by Andy Polyakov of | |
6863 | the OpenSSL core team. | |
6864 | ([CVE-2014-3570]) | |
6865 | ||
6866 | *Andy Polyakov* | |
6867 | ||
6868 | * Do not resume sessions on the server if the negotiated protocol | |
6869 | version does not match the session's version. Resuming with a different | |
6870 | version, while not strictly forbidden by the RFC, is of questionable | |
6871 | sanity and breaks all known clients. | |
6872 | ||
6873 | *David Benjamin, Emilia Käsper* | |
6874 | ||
6875 | * Tighten handling of the ChangeCipherSpec (CCS) message: reject | |
6876 | early CCS messages during renegotiation. (Note that because | |
6877 | renegotiation is encrypted, this early CCS was not exploitable.) | |
6878 | ||
6879 | *Emilia Käsper* | |
6880 | ||
6881 | * Tighten client-side session ticket handling during renegotiation: | |
6882 | ensure that the client only accepts a session ticket if the server sends | |
6883 | the extension anew in the ServerHello. Previously, a TLS client would | |
6884 | reuse the old extension state and thus accept a session ticket if one was | |
6885 | announced in the initial ServerHello. | |
6886 | ||
6887 | Similarly, ensure that the client requires a session ticket if one | |
6888 | was advertised in the ServerHello. Previously, a TLS client would | |
6889 | ignore a missing NewSessionTicket message. | |
6890 | ||
6891 | *Emilia Käsper* | |
6892 | ||
6893 | ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014] | |
6894 | ||
6895 | * SRTP Memory Leak. | |
6896 | ||
6897 | A flaw in the DTLS SRTP extension parsing code allows an attacker, who | |
6898 | sends a carefully crafted handshake message, to cause OpenSSL to fail | |
6899 | to free up to 64k of memory causing a memory leak. This could be | |
6900 | exploited in a Denial Of Service attack. This issue affects OpenSSL | |
6901 | 1.0.1 server implementations for both SSL/TLS and DTLS regardless of | |
6902 | whether SRTP is used or configured. Implementations of OpenSSL that | |
6903 | have been compiled with OPENSSL_NO_SRTP defined are not affected. | |
6904 | ||
6905 | The fix was developed by the OpenSSL team. | |
6906 | ([CVE-2014-3513]) | |
6907 | ||
6908 | *OpenSSL team* | |
6909 | ||
6910 | * Session Ticket Memory Leak. | |
6911 | ||
6912 | When an OpenSSL SSL/TLS/DTLS server receives a session ticket the | |
6913 | integrity of that ticket is first verified. In the event of a session | |
6914 | ticket integrity check failing, OpenSSL will fail to free memory | |
6915 | causing a memory leak. By sending a large number of invalid session | |
6916 | tickets an attacker could exploit this issue in a Denial Of Service | |
6917 | attack. | |
6918 | ([CVE-2014-3567]) | |
6919 | ||
6920 | *Steve Henson* | |
6921 | ||
6922 | * Build option no-ssl3 is incomplete. | |
6923 | ||
6924 | When OpenSSL is configured with "no-ssl3" as a build option, servers | |
6925 | could accept and complete a SSL 3.0 handshake, and clients could be | |
6926 | configured to send them. | |
6927 | ([CVE-2014-3568]) | |
6928 | ||
6929 | *Akamai and the OpenSSL team* | |
6930 | ||
6931 | * Add support for TLS_FALLBACK_SCSV. | |
6932 | Client applications doing fallback retries should call | |
6933 | SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). | |
6934 | ([CVE-2014-3566]) | |
6935 | ||
6936 | *Adam Langley, Bodo Moeller* | |
6937 | ||
6938 | * Add additional DigestInfo checks. | |
6939 | ||
6940 | Re-encode DigestInto in DER and check against the original when | |
6941 | verifying RSA signature: this will reject any improperly encoded | |
6942 | DigestInfo structures. | |
6943 | ||
6944 | Note: this is a precautionary measure and no attacks are currently known. | |
6945 | ||
6946 | *Steve Henson* | |
6947 | ||
6948 | ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014] | |
6949 | ||
6950 | * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the | |
6951 | SRP code can be overrun an internal buffer. Add sanity check that | |
6952 | g, A, B < N to SRP code. | |
6953 | ||
6954 | Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC | |
6955 | Group for discovering this issue. | |
6956 | ([CVE-2014-3512]) | |
6957 | ||
6958 | *Steve Henson* | |
6959 | ||
6960 | * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate | |
6961 | TLS 1.0 instead of higher protocol versions when the ClientHello message | |
6962 | is badly fragmented. This allows a man-in-the-middle attacker to force a | |
6963 | downgrade to TLS 1.0 even if both the server and the client support a | |
6964 | higher protocol version, by modifying the client's TLS records. | |
6965 | ||
6966 | Thanks to David Benjamin and Adam Langley (Google) for discovering and | |
6967 | researching this issue. | |
6968 | ([CVE-2014-3511]) | |
6969 | ||
6970 | *David Benjamin* | |
6971 | ||
6972 | * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject | |
6973 | to a denial of service attack. A malicious server can crash the client | |
6974 | with a null pointer dereference (read) by specifying an anonymous (EC)DH | |
6975 | ciphersuite and sending carefully crafted handshake messages. | |
6976 | ||
6977 | Thanks to Felix Gröbert (Google) for discovering and researching this | |
6978 | issue. | |
6979 | ([CVE-2014-3510]) | |
6980 | ||
6981 | *Emilia Käsper* | |
6982 | ||
6983 | * By sending carefully crafted DTLS packets an attacker could cause openssl | |
6984 | to leak memory. This can be exploited through a Denial of Service attack. | |
6985 | Thanks to Adam Langley for discovering and researching this issue. | |
6986 | ([CVE-2014-3507]) | |
6987 | ||
6988 | *Adam Langley* | |
6989 | ||
6990 | * An attacker can force openssl to consume large amounts of memory whilst | |
6991 | processing DTLS handshake messages. This can be exploited through a | |
6992 | Denial of Service attack. | |
6993 | Thanks to Adam Langley for discovering and researching this issue. | |
6994 | ([CVE-2014-3506]) | |
6995 | ||
6996 | *Adam Langley* | |
6997 | ||
6998 | * An attacker can force an error condition which causes openssl to crash | |
6999 | whilst processing DTLS packets due to memory being freed twice. This | |
7000 | can be exploited through a Denial of Service attack. | |
7001 | Thanks to Adam Langley and Wan-Teh Chang for discovering and researching | |
7002 | this issue. | |
7003 | ([CVE-2014-3505]) | |
7004 | ||
7005 | *Adam Langley* | |
7006 | ||
7007 | * If a multithreaded client connects to a malicious server using a resumed | |
7008 | session and the server sends an ec point format extension it could write | |
7009 | up to 255 bytes to freed memory. | |
7010 | ||
7011 | Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this | |
7012 | issue. | |
7013 | ([CVE-2014-3509]) | |
7014 | ||
7015 | *Gabor Tyukasz* | |
7016 | ||
7017 | * A malicious server can crash an OpenSSL client with a null pointer | |
7018 | dereference (read) by specifying an SRP ciphersuite even though it was not | |
7019 | properly negotiated with the client. This can be exploited through a | |
7020 | Denial of Service attack. | |
7021 | ||
7022 | Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for | |
7023 | discovering and researching this issue. | |
7024 | ([CVE-2014-5139]) | |
7025 | ||
7026 | *Steve Henson* | |
7027 | ||
7028 | * A flaw in OBJ_obj2txt may cause pretty printing functions such as | |
7029 | X509_name_oneline, X509_name_print_ex et al. to leak some information | |
7030 | from the stack. Applications may be affected if they echo pretty printing | |
7031 | output to the attacker. | |
7032 | ||
7033 | Thanks to Ivan Fratric (Google) for discovering this issue. | |
7034 | ([CVE-2014-3508]) | |
7035 | ||
7036 | *Emilia Käsper, and Steve Henson* | |
7037 | ||
7038 | * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) | |
7039 | for corner cases. (Certain input points at infinity could lead to | |
7040 | bogus results, with non-infinity inputs mapped to infinity too.) | |
7041 | ||
7042 | *Bodo Moeller* | |
7043 | ||
7044 | ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014] | |
7045 | ||
7046 | * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted | |
7047 | handshake can force the use of weak keying material in OpenSSL | |
7048 | SSL/TLS clients and servers. | |
7049 | ||
7050 | Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and | |
7051 | researching this issue. ([CVE-2014-0224]) | |
7052 | ||
7053 | *KIKUCHI Masashi, Steve Henson* | |
7054 | ||
7055 | * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an | |
7056 | OpenSSL DTLS client the code can be made to recurse eventually crashing | |
7057 | in a DoS attack. | |
7058 | ||
7059 | Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. | |
7060 | ([CVE-2014-0221]) | |
7061 | ||
7062 | *Imre Rad, Steve Henson* | |
7063 | ||
7064 | * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can | |
7065 | be triggered by sending invalid DTLS fragments to an OpenSSL DTLS | |
7066 | client or server. This is potentially exploitable to run arbitrary | |
7067 | code on a vulnerable client or server. | |
7068 | ||
7069 | Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195]) | |
7070 | ||
7071 | *Jüri Aedla, Steve Henson* | |
7072 | ||
7073 | * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites | |
7074 | are subject to a denial of service attack. | |
7075 | ||
7076 | Thanks to Felix Gröbert and Ivan Fratric at Google for discovering | |
7077 | this issue. ([CVE-2014-3470]) | |
7078 | ||
7079 | *Felix Gröbert, Ivan Fratric, Steve Henson* | |
7080 | ||
7081 | * Harmonize version and its documentation. -f flag is used to display | |
7082 | compilation flags. | |
7083 | ||
7084 | *mancha <mancha1@zoho.com>* | |
7085 | ||
7086 | * Fix eckey_priv_encode so it immediately returns an error upon a failure | |
7087 | in i2d_ECPrivateKey. | |
7088 | ||
7089 | *mancha <mancha1@zoho.com>* | |
7090 | ||
7091 | * Fix some double frees. These are not thought to be exploitable. | |
7092 | ||
7093 | *mancha <mancha1@zoho.com>* | |
7094 | ||
7095 | ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014] | |
7096 | ||
7097 | * A missing bounds check in the handling of the TLS heartbeat extension | |
7098 | can be used to reveal up to 64k of memory to a connected client or | |
7099 | server. | |
7100 | ||
7101 | Thanks for Neel Mehta of Google Security for discovering this bug and to | |
7102 | Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for | |
7103 | preparing the fix ([CVE-2014-0160]) | |
7104 | ||
7105 | *Adam Langley, Bodo Moeller* | |
7106 | ||
7107 | * Fix for the attack described in the paper "Recovering OpenSSL | |
7108 | ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" | |
7109 | by Yuval Yarom and Naomi Benger. Details can be obtained from: | |
7110 | <http://eprint.iacr.org/2014/140> | |
7111 | ||
7112 | Thanks to Yuval Yarom and Naomi Benger for discovering this | |
7113 | flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076]) | |
7114 | ||
7115 | *Yuval Yarom and Naomi Benger* | |
7116 | ||
7117 | * TLS pad extension: draft-agl-tls-padding-03 | |
7118 | ||
7119 | Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the | |
7120 | TLS client Hello record length value would otherwise be > 255 and | |
7121 | less that 512 pad with a dummy extension containing zeroes so it | |
7122 | is at least 512 bytes long. | |
7123 | ||
7124 | *Adam Langley, Steve Henson* | |
7125 | ||
7126 | ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014] | |
7127 | ||
7128 | * Fix for TLS record tampering bug. A carefully crafted invalid | |
7129 | handshake could crash OpenSSL with a NULL pointer exception. | |
7130 | Thanks to Anton Johansson for reporting this issues. | |
7131 | ([CVE-2013-4353]) | |
7132 | ||
7133 | * Keep original DTLS digest and encryption contexts in retransmission | |
7134 | structures so we can use the previous session parameters if they need | |
7135 | to be resent. ([CVE-2013-6450]) | |
7136 | ||
7137 | *Steve Henson* | |
7138 | ||
7139 | * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which | |
7140 | avoids preferring ECDHE-ECDSA ciphers when the client appears to be | |
7141 | Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for | |
7142 | several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug | |
7143 | is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing | |
7144 | 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. | |
7145 | ||
7146 | *Rob Stradling, Adam Langley* | |
7147 | ||
7148 | ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013] | |
7149 | ||
7150 | * Correct fix for CVE-2013-0169. The original didn't work on AES-NI | |
7151 | supporting platforms or when small records were transferred. | |
7152 | ||
7153 | *Andy Polyakov, Steve Henson* | |
7154 | ||
7155 | ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013] | |
7156 | ||
7157 | * Make the decoding of SSLv3, TLS and DTLS CBC records constant time. | |
7158 | ||
7159 | This addresses the flaw in CBC record processing discovered by | |
7160 | Nadhem Alfardan and Kenny Paterson. Details of this attack can be found | |
7161 | at: <http://www.isg.rhul.ac.uk/tls/> | |
7162 | ||
7163 | Thanks go to Nadhem Alfardan and Kenny Paterson of the Information | |
7164 | Security Group at Royal Holloway, University of London | |
7165 | (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and | |
7166 | Emilia Käsper for the initial patch. | |
7167 | ([CVE-2013-0169]) | |
7168 | ||
7169 | *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson* | |
7170 | ||
7171 | * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode | |
7172 | ciphersuites which can be exploited in a denial of service attack. | |
7173 | Thanks go to and to Adam Langley <agl@chromium.org> for discovering | |
7174 | and detecting this bug and to Wolfgang Ettlinger | |
7175 | <wolfgang.ettlinger@gmail.com> for independently discovering this issue. | |
7176 | ([CVE-2012-2686]) | |
7177 | ||
7178 | *Adam Langley* | |
7179 | ||
7180 | * Return an error when checking OCSP signatures when key is NULL. | |
7181 | This fixes a DoS attack. ([CVE-2013-0166]) | |
7182 | ||
7183 | *Steve Henson* | |
7184 | ||
7185 | * Make openssl verify return errors. | |
7186 | ||
7187 | *Chris Palmer <palmer@google.com> and Ben Laurie* | |
7188 | ||
7189 | * Call OCSP Stapling callback after ciphersuite has been chosen, so | |
7190 | the right response is stapled. Also change SSL_get_certificate() | |
7191 | so it returns the certificate actually sent. | |
7192 | See <http://rt.openssl.org/Ticket/Display.html?id=2836>. | |
7193 | ||
7194 | *Rob Stradling <rob.stradling@comodo.com>* | |
7195 | ||
7196 | * Fix possible deadlock when decoding public keys. | |
7197 | ||
7198 | *Steve Henson* | |
7199 | ||
7200 | * Don't use TLS 1.0 record version number in initial client hello | |
7201 | if renegotiating. | |
7202 | ||
7203 | *Steve Henson* | |
7204 | ||
7205 | ### Changes between 1.0.1b and 1.0.1c [10 May 2012] | |
7206 | ||
7207 | * Sanity check record length before skipping explicit IV in TLS | |
7208 | 1.2, 1.1 and DTLS to fix DoS attack. | |
7209 | ||
7210 | Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic | |
7211 | fuzzing as a service testing platform. | |
7212 | ([CVE-2012-2333]) | |
7213 | ||
7214 | *Steve Henson* | |
7215 | ||
7216 | * Initialise tkeylen properly when encrypting CMS messages. | |
7217 | Thanks to Solar Designer of Openwall for reporting this issue. | |
7218 | ||
7219 | *Steve Henson* | |
7220 | ||
7221 | * In FIPS mode don't try to use composite ciphers as they are not | |
7222 | approved. | |
7223 | ||
7224 | *Steve Henson* | |
7225 | ||
7226 | ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012] | |
7227 | ||
7228 | * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and | |
7229 | 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately | |
7230 | mean any application compiled against OpenSSL 1.0.0 headers setting | |
7231 | SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling | |
7232 | TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to | |
7233 | 0x10000000L Any application which was previously compiled against | |
7234 | OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 | |
7235 | will need to be recompiled as a result. Letting be results in | |
7236 | inability to disable specifically TLS 1.1 and in client context, | |
7237 | in unlike event, limit maximum offered version to TLS 1.0 [see below]. | |
7238 | ||
7239 | *Steve Henson* | |
7240 | ||
7241 | * In order to ensure interoperability SSL_OP_NO_protocolX does not | |
7242 | disable just protocol X, but all protocols above X *if* there are | |
7243 | protocols *below* X still enabled. In more practical terms it means | |
7244 | that if application wants to disable TLS1.0 in favor of TLS1.1 and | |
7245 | above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass | |
7246 | `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to | |
7247 | client side. | |
7248 | ||
7249 | *Andy Polyakov* | |
7250 | ||
7251 | ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012] | |
7252 | ||
7253 | * Check for potentially exploitable overflows in asn1_d2i_read_bio | |
7254 | BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer | |
7255 | in CRYPTO_realloc_clean. | |
7256 | ||
7257 | Thanks to Tavis Ormandy, Google Security Team, for discovering this | |
7258 | issue and to Adam Langley <agl@chromium.org> for fixing it. | |
7259 | ([CVE-2012-2110]) | |
7260 | ||
7261 | *Adam Langley (Google), Tavis Ormandy, Google Security Team* | |
7262 | ||
7263 | * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. | |
7264 | ||
7265 | *Adam Langley* | |
7266 | ||
7267 | * Workarounds for some broken servers that "hang" if a client hello | |
7268 | record length exceeds 255 bytes. | |
7269 | ||
7270 | 1. Do not use record version number > TLS 1.0 in initial client | |
7271 | hello: some (but not all) hanging servers will now work. | |
7272 | 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate | |
7273 | the number of ciphers sent in the client hello. This should be | |
7274 | set to an even number, such as 50, for example by passing: | |
7275 | -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. | |
7276 | Most broken servers should now work. | |
7277 | 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable | |
7278 | TLS 1.2 client support entirely. | |
7279 | ||
7280 | *Steve Henson* | |
7281 | ||
7282 | * Fix SEGV in Vector Permutation AES module observed in OpenSSH. | |
7283 | ||
7284 | *Andy Polyakov* | |
7285 | ||
7286 | ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012] | |
7287 | ||
7288 | * Add compatibility with old MDC2 signatures which use an ASN1 OCTET | |
7289 | STRING form instead of a DigestInfo. | |
7290 | ||
7291 | *Steve Henson* | |
7292 | ||
7293 | * The format used for MDC2 RSA signatures is inconsistent between EVP | |
7294 | and the RSA_sign/RSA_verify functions. This was made more apparent when | |
7295 | OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular | |
7296 | those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect | |
7297 | the correct format in RSA_verify so both forms transparently work. | |
7298 | ||
7299 | *Steve Henson* | |
7300 | ||
7301 | * Some servers which support TLS 1.0 can choke if we initially indicate | |
7302 | support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA | |
7303 | encrypted premaster secret. As a workaround use the maximum permitted | |
7304 | client version in client hello, this should keep such servers happy | |
7305 | and still work with previous versions of OpenSSL. | |
7306 | ||
7307 | *Steve Henson* | |
7308 | ||
7309 | * Add support for TLS/DTLS heartbeats. | |
7310 | ||
7311 | *Robin Seggelmann <seggelmann@fh-muenster.de>* | |
7312 | ||
7313 | * Add support for SCTP. | |
7314 | ||
7315 | *Robin Seggelmann <seggelmann@fh-muenster.de>* | |
7316 | ||
7317 | * Improved PRNG seeding for VOS. | |
7318 | ||
7319 | *Paul Green <Paul.Green@stratus.com>* | |
7320 | ||
7321 | * Extensive assembler packs updates, most notably: | |
7322 | ||
7323 | - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; | |
7324 | - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); | |
7325 | - x86_64: bit-sliced AES implementation; | |
7326 | - ARM: NEON support, contemporary platforms optimizations; | |
7327 | - s390x: z196 support; | |
7328 | - `*`: GHASH and GF(2^m) multiplication implementations; | |
7329 | ||
7330 | *Andy Polyakov* | |
7331 | ||
7332 | * Make TLS-SRP code conformant with RFC 5054 API cleanup | |
7333 | (removal of unnecessary code) | |
7334 | ||
7335 | *Peter Sylvester <peter.sylvester@edelweb.fr>* | |
7336 | ||
7337 | * Add TLS key material exporter from RFC 5705. | |
7338 | ||
7339 | *Eric Rescorla* | |
7340 | ||
7341 | * Add DTLS-SRTP negotiation from RFC 5764. | |
7342 | ||
7343 | *Eric Rescorla* | |
7344 | ||
7345 | * Add Next Protocol Negotiation, | |
7346 | <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be | |
7347 | disabled with a no-npn flag to config or Configure. Code donated | |
7348 | by Google. | |
7349 | ||
7350 | *Adam Langley <agl@google.com> and Ben Laurie* | |
7351 | ||
7352 | * Add optional 64-bit optimized implementations of elliptic curves NIST-P224, | |
7353 | NIST-P256, NIST-P521, with constant-time single point multiplication on | |
7354 | typical inputs. Compiler support for the nonstandard type `__uint128_t` is | |
7355 | required to use this (present in gcc 4.4 and later, for 64-bit builds). | |
7356 | Code made available under Apache License version 2.0. | |
7357 | ||
7358 | Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command | |
7359 | line to include this in your build of OpenSSL, and run "make depend" (or | |
7360 | "make update"). This enables the following EC_METHODs: | |
7361 | ||
7362 | EC_GFp_nistp224_method() | |
7363 | EC_GFp_nistp256_method() | |
7364 | EC_GFp_nistp521_method() | |
7365 | ||
7366 | EC_GROUP_new_by_curve_name() will automatically use these (while | |
7367 | EC_GROUP_new_curve_GFp() currently prefers the more flexible | |
7368 | implementations). | |
7369 | ||
7370 | *Emilia Käsper, Adam Langley, Bodo Moeller (Google)* | |
7371 | ||
7372 | * Use type ossl_ssize_t instead of ssize_t which isn't available on | |
7373 | all platforms. Move ssize_t definition from e_os.h to the public | |
7374 | header file e_os2.h as it now appears in public header file cms.h | |
7375 | ||
7376 | *Steve Henson* | |
7377 | ||
7378 | * New -sigopt option to the ca, req and x509 utilities. Additional | |
7379 | signature parameters can be passed using this option and in | |
7380 | particular PSS. | |
7381 | ||
7382 | *Steve Henson* | |
7383 | ||
7384 | * Add RSA PSS signing function. This will generate and set the | |
7385 | appropriate AlgorithmIdentifiers for PSS based on those in the | |
7386 | corresponding EVP_MD_CTX structure. No application support yet. | |
7387 | ||
7388 | *Steve Henson* | |
7389 | ||
7390 | * Support for companion algorithm specific ASN1 signing routines. | |
7391 | New function ASN1_item_sign_ctx() signs a pre-initialised | |
7392 | EVP_MD_CTX structure and sets AlgorithmIdentifiers based on | |
7393 | the appropriate parameters. | |
7394 | ||
7395 | *Steve Henson* | |
7396 | ||
7397 | * Add new algorithm specific ASN1 verification initialisation function | |
7398 | to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 | |
7399 | handling will be the same no matter what EVP_PKEY_METHOD is used. | |
7400 | Add a PSS handler to support verification of PSS signatures: checked | |
7401 | against a number of sample certificates. | |
7402 | ||
7403 | *Steve Henson* | |
7404 | ||
7405 | * Add signature printing for PSS. Add PSS OIDs. | |
7406 | ||
7407 | *Steve Henson, Martin Kaiser <lists@kaiser.cx>* | |
7408 | ||
7409 | * Add algorithm specific signature printing. An individual ASN1 method | |
7410 | can now print out signatures instead of the standard hex dump. | |
7411 | ||
7412 | More complex signatures (e.g. PSS) can print out more meaningful | |
7413 | information. Include DSA version that prints out the signature | |
7414 | parameters r, s. | |
7415 | ||
7416 | *Steve Henson* | |
7417 | ||
7418 | * Password based recipient info support for CMS library: implementing | |
7419 | RFC3211. | |
7420 | ||
7421 | *Steve Henson* | |
7422 | ||
7423 | * Split password based encryption into PBES2 and PBKDF2 functions. This | |
7424 | neatly separates the code into cipher and PBE sections and is required | |
7425 | for some algorithms that split PBES2 into separate pieces (such as | |
7426 | password based CMS). | |
7427 | ||
7428 | *Steve Henson* | |
7429 | ||
7430 | * Session-handling fixes: | |
7431 | - Fix handling of connections that are resuming with a session ID, | |
7432 | but also support Session Tickets. | |
7433 | - Fix a bug that suppressed issuing of a new ticket if the client | |
7434 | presented a ticket with an expired session. | |
7435 | - Try to set the ticket lifetime hint to something reasonable. | |
7436 | - Make tickets shorter by excluding irrelevant information. | |
7437 | - On the client side, don't ignore renewed tickets. | |
7438 | ||
7439 | *Adam Langley, Bodo Moeller (Google)* | |
7440 | ||
7441 | * Fix PSK session representation. | |
7442 | ||
7443 | *Bodo Moeller* | |
7444 | ||
7445 | * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. | |
7446 | ||
7447 | This work was sponsored by Intel. | |
7448 | ||
7449 | *Andy Polyakov* | |
7450 | ||
7451 | * Add GCM support to TLS library. Some custom code is needed to split | |
7452 | the IV between the fixed (from PRF) and explicit (from TLS record) | |
7453 | portions. This adds all GCM ciphersuites supported by RFC5288 and | |
7454 | RFC5289. Generalise some `AES*` cipherstrings to include GCM and | |
7455 | add a special AESGCM string for GCM only. | |
7456 | ||
7457 | *Steve Henson* | |
7458 | ||
7459 | * Expand range of ctrls for AES GCM. Permit setting invocation | |
7460 | field on decrypt and retrieval of invocation field only on encrypt. | |
7461 | ||
7462 | *Steve Henson* | |
7463 | ||
7464 | * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. | |
7465 | As required by RFC5289 these ciphersuites cannot be used if for | |
7466 | versions of TLS earlier than 1.2. | |
7467 | ||
7468 | *Steve Henson* | |
7469 | ||
7470 | * For FIPS capable OpenSSL interpret a NULL default public key method | |
7471 | as unset and return the appropriate default but do *not* set the default. | |
7472 | This means we can return the appropriate method in applications that | |
7473 | switch between FIPS and non-FIPS modes. | |
7474 | ||
7475 | *Steve Henson* | |
7476 | ||
7477 | * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an | |
7478 | ENGINE is used then we cannot handle that in the FIPS module so we | |
7479 | keep original code iff non-FIPS operations are allowed. | |
7480 | ||
7481 | *Steve Henson* | |
7482 | ||
7483 | * Add -attime option to openssl utilities. | |
7484 | ||
7485 | *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson* | |
7486 | ||
7487 | * Redirect DSA and DH operations to FIPS module in FIPS mode. | |
7488 | ||
7489 | *Steve Henson* | |
7490 | ||
7491 | * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use | |
7492 | FIPS EC methods unconditionally for now. | |
7493 | ||
7494 | *Steve Henson* | |
7495 | ||
7496 | * New build option no-ec2m to disable characteristic 2 code. | |
7497 | ||
7498 | *Steve Henson* | |
7499 | ||
7500 | * Backport libcrypto audit of return value checking from 1.1.0-dev; not | |
7501 | all cases can be covered as some introduce binary incompatibilities. | |
7502 | ||
7503 | *Steve Henson* | |
7504 | ||
7505 | * Redirect RSA operations to FIPS module including keygen, | |
7506 | encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. | |
7507 | ||
7508 | *Steve Henson* | |
7509 | ||
7510 | * Add similar low level API blocking to ciphers. | |
7511 | ||
7512 | *Steve Henson* | |
7513 | ||
7514 | * Low level digest APIs are not approved in FIPS mode: any attempt | |
7515 | to use these will cause a fatal error. Applications that *really* want | |
7516 | to use them can use the `private_*` version instead. | |
7517 | ||
7518 | *Steve Henson* | |
7519 | ||
7520 | * Redirect cipher operations to FIPS module for FIPS builds. | |
7521 | ||
7522 | *Steve Henson* | |
7523 | ||
7524 | * Redirect digest operations to FIPS module for FIPS builds. | |
7525 | ||
7526 | *Steve Henson* | |
7527 | ||
7528 | * Update build system to add "fips" flag which will link in fipscanister.o | |
7529 | for static and shared library builds embedding a signature if needed. | |
7530 | ||
7531 | *Steve Henson* | |
7532 | ||
7533 | * Output TLS supported curves in preference order instead of numerical | |
7534 | order. This is currently hardcoded for the highest order curves first. | |
7535 | This should be configurable so applications can judge speed vs strength. | |
7536 | ||
7537 | *Steve Henson* | |
7538 | ||
7539 | * Add TLS v1.2 server support for client authentication. | |
7540 | ||
7541 | *Steve Henson* | |
7542 | ||
7543 | * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers | |
7544 | and enable MD5. | |
7545 | ||
7546 | *Steve Henson* | |
7547 | ||
7548 | * Functions FIPS_mode_set() and FIPS_mode() which call the underlying | |
7549 | FIPS modules versions. | |
7550 | ||
7551 | *Steve Henson* | |
7552 | ||
7553 | * Add TLS v1.2 client side support for client authentication. Keep cache | |
7554 | of handshake records longer as we don't know the hash algorithm to use | |
7555 | until after the certificate request message is received. | |
7556 | ||
7557 | *Steve Henson* | |
7558 | ||
7559 | * Initial TLS v1.2 client support. Add a default signature algorithms | |
7560 | extension including all the algorithms we support. Parse new signature | |
7561 | format in client key exchange. Relax some ECC signing restrictions for | |
7562 | TLS v1.2 as indicated in RFC5246. | |
7563 | ||
7564 | *Steve Henson* | |
7565 | ||
7566 | * Add server support for TLS v1.2 signature algorithms extension. Switch | |
7567 | to new signature format when needed using client digest preference. | |
7568 | All server ciphersuites should now work correctly in TLS v1.2. No client | |
7569 | support yet and no support for client certificates. | |
7570 | ||
7571 | *Steve Henson* | |
7572 | ||
7573 | * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch | |
7574 | to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based | |
7575 | ciphersuites. At present only RSA key exchange ciphersuites work with | |
7576 | TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete | |
7577 | SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods | |
7578 | and version checking. | |
7579 | ||
7580 | *Steve Henson* | |
7581 | ||
7582 | * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled | |
7583 | with this defined it will not be affected by any changes to ssl internal | |
7584 | structures. Add several utility functions to allow openssl application | |
7585 | to work with OPENSSL_NO_SSL_INTERN defined. | |
7586 | ||
7587 | *Steve Henson* | |
7588 | ||
7589 | * A long standing patch to add support for SRP from EdelWeb (Peter | |
7590 | Sylvester and Christophe Renou) was integrated. | |
7591 | *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester | |
7592 | <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and | |
7593 | Ben Laurie* | |
7594 | ||
7595 | * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. | |
7596 | ||
7597 | *Steve Henson* | |
7598 | ||
7599 | * Permit abbreviated handshakes when renegotiating using the function | |
7600 | SSL_renegotiate_abbreviated(). | |
7601 | ||
7602 | *Robin Seggelmann <seggelmann@fh-muenster.de>* | |
7603 | ||
7604 | * Add call to ENGINE_register_all_complete() to | |
7605 | ENGINE_load_builtin_engines(), so some implementations get used | |
7606 | automatically instead of needing explicit application support. | |
7607 | ||
7608 | *Steve Henson* | |
7609 | ||
7610 | * Add support for TLS key exporter as described in RFC5705. | |
7611 | ||
7612 | *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson* | |
7613 | ||
7614 | * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only | |
7615 | a few changes are required: | |
7616 | ||
7617 | Add SSL_OP_NO_TLSv1_1 flag. | |
7618 | Add TLSv1_1 methods. | |
7619 | Update version checking logic to handle version 1.1. | |
7620 | Add explicit IV handling (ported from DTLS code). | |
7621 | Add command line options to s_client/s_server. | |
7622 | ||
7623 | *Steve Henson* | |
7624 | ||
7625 | OpenSSL 1.0.0 | |
7626 | ------------- | |
7627 | ||
7628 | ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015] | |
7629 | ||
7630 | * X509_ATTRIBUTE memory leak | |
7631 | ||
7632 | When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak | |
7633 | memory. This structure is used by the PKCS#7 and CMS routines so any | |
7634 | application which reads PKCS#7 or CMS data from untrusted sources is | |
7635 | affected. SSL/TLS is not affected. | |
7636 | ||
7637 | This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using | |
7638 | libFuzzer. | |
7639 | ([CVE-2015-3195]) | |
7640 | ||
7641 | *Stephen Henson* | |
7642 | ||
7643 | * Race condition handling PSK identify hint | |
7644 | ||
7645 | If PSK identity hints are received by a multi-threaded client then | |
7646 | the values are wrongly updated in the parent SSL_CTX structure. This can | |
7647 | result in a race condition potentially leading to a double free of the | |
7648 | identify hint data. | |
7649 | ([CVE-2015-3196]) | |
7650 | ||
7651 | *Stephen Henson* | |
7652 | ||
7653 | ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015] | |
7654 | ||
7655 | * Malformed ECParameters causes infinite loop | |
7656 | ||
7657 | When processing an ECParameters structure OpenSSL enters an infinite loop | |
7658 | if the curve specified is over a specially malformed binary polynomial | |
7659 | field. | |
7660 | ||
7661 | This can be used to perform denial of service against any | |
7662 | system which processes public keys, certificate requests or | |
7663 | certificates. This includes TLS clients and TLS servers with | |
7664 | client authentication enabled. | |
7665 | ||
7666 | This issue was reported to OpenSSL by Joseph Barr-Pixton. | |
7667 | ([CVE-2015-1788]) | |
7668 | ||
7669 | *Andy Polyakov* | |
7670 | ||
7671 | * Exploitable out-of-bounds read in X509_cmp_time | |
7672 | ||
7673 | X509_cmp_time does not properly check the length of the ASN1_TIME | |
7674 | string and can read a few bytes out of bounds. In addition, | |
7675 | X509_cmp_time accepts an arbitrary number of fractional seconds in the | |
7676 | time string. | |
7677 | ||
7678 | An attacker can use this to craft malformed certificates and CRLs of | |
7679 | various sizes and potentially cause a segmentation fault, resulting in | |
7680 | a DoS on applications that verify certificates or CRLs. TLS clients | |
7681 | that verify CRLs are affected. TLS clients and servers with client | |
7682 | authentication enabled may be affected if they use custom verification | |
7683 | callbacks. | |
7684 | ||
7685 | This issue was reported to OpenSSL by Robert Swiecki (Google), and | |
7686 | independently by Hanno Böck. | |
7687 | ([CVE-2015-1789]) | |
7688 | ||
7689 | *Emilia Käsper* | |
7690 | ||
7691 | * PKCS7 crash with missing EnvelopedContent | |
7692 | ||
7693 | The PKCS#7 parsing code does not handle missing inner EncryptedContent | |
7694 | correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs | |
7695 | with missing content and trigger a NULL pointer dereference on parsing. | |
7696 | ||
7697 | Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 | |
7698 | structures from untrusted sources are affected. OpenSSL clients and | |
7699 | servers are not affected. | |
7700 | ||
7701 | This issue was reported to OpenSSL by Michal Zalewski (Google). | |
7702 | ([CVE-2015-1790]) | |
7703 | ||
7704 | *Emilia Käsper* | |
7705 | ||
7706 | * CMS verify infinite loop with unknown hash function | |
7707 | ||
7708 | When verifying a signedData message the CMS code can enter an infinite loop | |
7709 | if presented with an unknown hash function OID. This can be used to perform | |
7710 | denial of service against any system which verifies signedData messages using | |
7711 | the CMS code. | |
7712 | This issue was reported to OpenSSL by Johannes Bauer. | |
7713 | ([CVE-2015-1792]) | |
7714 | ||
7715 | *Stephen Henson* | |
7716 | ||
7717 | * Race condition handling NewSessionTicket | |
7718 | ||
7719 | If a NewSessionTicket is received by a multi-threaded client when attempting to | |
7720 | reuse a previous ticket then a race condition can occur potentially leading to | |
7721 | a double free of the ticket data. | |
7722 | ([CVE-2015-1791]) | |
7723 | ||
7724 | *Matt Caswell* | |
7725 | ||
7726 | ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015] | |
7727 | ||
7728 | * Segmentation fault in ASN1_TYPE_cmp fix | |
7729 | ||
7730 | The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is | |
7731 | made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check | |
7732 | certificate signature algorithm consistency this can be used to crash any | |
7733 | certificate verification operation and exploited in a DoS attack. Any | |
7734 | application which performs certificate verification is vulnerable including | |
7735 | OpenSSL clients and servers which enable client authentication. | |
7736 | ([CVE-2015-0286]) | |
7737 | ||
7738 | *Stephen Henson* | |
7739 | ||
7740 | * ASN.1 structure reuse memory corruption fix | |
7741 | ||
7742 | Reusing a structure in ASN.1 parsing may allow an attacker to cause | |
7743 | memory corruption via an invalid write. Such reuse is and has been | |
7744 | strongly discouraged and is believed to be rare. | |
7745 | ||
7746 | Applications that parse structures containing CHOICE or ANY DEFINED BY | |
7747 | components may be affected. Certificate parsing (d2i_X509 and related | |
7748 | functions) are however not affected. OpenSSL clients and servers are | |
7749 | not affected. | |
7750 | ([CVE-2015-0287]) | |
7751 | ||
7752 | *Stephen Henson* | |
7753 | ||
7754 | * PKCS7 NULL pointer dereferences fix | |
7755 | ||
7756 | The PKCS#7 parsing code does not handle missing outer ContentInfo | |
7757 | correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with | |
7758 | missing content and trigger a NULL pointer dereference on parsing. | |
7759 | ||
7760 | Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or | |
7761 | otherwise parse PKCS#7 structures from untrusted sources are | |
7762 | affected. OpenSSL clients and servers are not affected. | |
7763 | ||
7764 | This issue was reported to OpenSSL by Michal Zalewski (Google). | |
7765 | ([CVE-2015-0289]) | |
7766 | ||
7767 | *Emilia Käsper* | |
7768 | ||
7769 | * DoS via reachable assert in SSLv2 servers fix | |
7770 | ||
7771 | A malicious client can trigger an OPENSSL_assert (i.e., an abort) in | |
7772 | servers that both support SSLv2 and enable export cipher suites by sending | |
7773 | a specially crafted SSLv2 CLIENT-MASTER-KEY message. | |
7774 | ||
7775 | This issue was discovered by Sean Burford (Google) and Emilia Käsper | |
7776 | (OpenSSL development team). | |
7777 | ([CVE-2015-0293]) | |
7778 | ||
7779 | *Emilia Käsper* | |
7780 | ||
7781 | * Use After Free following d2i_ECPrivatekey error fix | |
7782 | ||
7783 | A malformed EC private key file consumed via the d2i_ECPrivateKey function | |
7784 | could cause a use after free condition. This, in turn, could cause a double | |
7785 | free in several private key parsing functions (such as d2i_PrivateKey | |
7786 | or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption | |
7787 | for applications that receive EC private keys from untrusted | |
7788 | sources. This scenario is considered rare. | |
7789 | ||
7790 | This issue was discovered by the BoringSSL project and fixed in their | |
7791 | commit 517073cd4b. | |
7792 | ([CVE-2015-0209]) | |
7793 | ||
7794 | *Matt Caswell* | |
7795 | ||
7796 | * X509_to_X509_REQ NULL pointer deref fix | |
7797 | ||
7798 | The function X509_to_X509_REQ will crash with a NULL pointer dereference if | |
7799 | the certificate key is invalid. This function is rarely used in practice. | |
7800 | ||
7801 | This issue was discovered by Brian Carpenter. | |
7802 | ([CVE-2015-0288]) | |
7803 | ||
7804 | *Stephen Henson* | |
7805 | ||
7806 | * Removed the export ciphers from the DEFAULT ciphers | |
7807 | ||
7808 | *Kurt Roeckx* | |
7809 | ||
7810 | ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015] | |
7811 | ||
7812 | * Build fixes for the Windows and OpenVMS platforms | |
7813 | ||
7814 | *Matt Caswell and Richard Levitte* | |
7815 | ||
7816 | ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015] | |
7817 | ||
7818 | * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS | |
7819 | message can cause a segmentation fault in OpenSSL due to a NULL pointer | |
7820 | dereference. This could lead to a Denial Of Service attack. Thanks to | |
7821 | Markus Stenberg of Cisco Systems, Inc. for reporting this issue. | |
7822 | ([CVE-2014-3571]) | |
7823 | ||
7824 | *Steve Henson* | |
7825 | ||
7826 | * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the | |
7827 | dtls1_buffer_record function under certain conditions. In particular this | |
7828 | could occur if an attacker sent repeated DTLS records with the same | |
7829 | sequence number but for the next epoch. The memory leak could be exploited | |
7830 | by an attacker in a Denial of Service attack through memory exhaustion. | |
7831 | Thanks to Chris Mueller for reporting this issue. | |
7832 | ([CVE-2015-0206]) | |
7833 | ||
7834 | *Matt Caswell* | |
7835 | ||
7836 | * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is | |
7837 | built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl | |
7838 | method would be set to NULL which could later result in a NULL pointer | |
7839 | dereference. Thanks to Frank Schmirler for reporting this issue. | |
7840 | ([CVE-2014-3569]) | |
7841 | ||
7842 | *Kurt Roeckx* | |
7843 | ||
7844 | * Abort handshake if server key exchange message is omitted for ephemeral | |
7845 | ECDH ciphersuites. | |
7846 | ||
7847 | Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for | |
7848 | reporting this issue. | |
7849 | ([CVE-2014-3572]) | |
7850 | ||
7851 | *Steve Henson* | |
7852 | ||
7853 | * Remove non-export ephemeral RSA code on client and server. This code | |
7854 | violated the TLS standard by allowing the use of temporary RSA keys in | |
7855 | non-export ciphersuites and could be used by a server to effectively | |
7856 | downgrade the RSA key length used to a value smaller than the server | |
7857 | certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at | |
7858 | INRIA or reporting this issue. | |
7859 | ([CVE-2015-0204]) | |
7860 | ||
7861 | *Steve Henson* | |
7862 | ||
7863 | * Fixed issue where DH client certificates are accepted without verification. | |
7864 | An OpenSSL server will accept a DH certificate for client authentication | |
7865 | without the certificate verify message. This effectively allows a client to | |
7866 | authenticate without the use of a private key. This only affects servers | |
7867 | which trust a client certificate authority which issues certificates | |
7868 | containing DH keys: these are extremely rare and hardly ever encountered. | |
7869 | Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting | |
7870 | this issue. | |
7871 | ([CVE-2015-0205]) | |
7872 | ||
7873 | *Steve Henson* | |
7874 | ||
7875 | * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect | |
7876 | results on some platforms, including x86_64. This bug occurs at random | |
7877 | with a very low probability, and is not known to be exploitable in any | |
7878 | way, though its exact impact is difficult to determine. Thanks to Pieter | |
7879 | Wuille (Blockstream) who reported this issue and also suggested an initial | |
7880 | fix. Further analysis was conducted by the OpenSSL development team and | |
7881 | Adam Langley of Google. The final fix was developed by Andy Polyakov of | |
7882 | the OpenSSL core team. | |
7883 | ([CVE-2014-3570]) | |
7884 | ||
7885 | *Andy Polyakov* | |
7886 | ||
7887 | * Fix various certificate fingerprint issues. | |
7888 | ||
7889 | By using non-DER or invalid encodings outside the signed portion of a | |
7890 | certificate the fingerprint can be changed without breaking the signature. | |
7891 | Although no details of the signed portion of the certificate can be changed | |
7892 | this can cause problems with some applications: e.g. those using the | |
7893 | certificate fingerprint for blacklists. | |
7894 | ||
7895 | 1. Reject signatures with non zero unused bits. | |
7896 | ||
7897 | If the BIT STRING containing the signature has non zero unused bits reject | |
7898 | the signature. All current signature algorithms require zero unused bits. | |
7899 | ||
7900 | 2. Check certificate algorithm consistency. | |
7901 | ||
7902 | Check the AlgorithmIdentifier inside TBS matches the one in the | |
7903 | certificate signature. NB: this will result in signature failure | |
7904 | errors for some broken certificates. | |
7905 | ||
7906 | Thanks to Konrad Kraszewski from Google for reporting this issue. | |
7907 | ||
7908 | 3. Check DSA/ECDSA signatures use DER. | |
7909 | ||
7910 | Reencode DSA/ECDSA signatures and compare with the original received | |
7911 | signature. Return an error if there is a mismatch. | |
7912 | ||
7913 | This will reject various cases including garbage after signature | |
7914 | (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS | |
7915 | program for discovering this case) and use of BER or invalid ASN.1 INTEGERs | |
7916 | (negative or with leading zeroes). | |
7917 | ||
7918 | Further analysis was conducted and fixes were developed by Stephen Henson | |
7919 | of the OpenSSL core team. | |
7920 | ||
7921 | ([CVE-2014-8275]) | |
7922 | ||
7923 | *Steve Henson* | |
7924 | ||
7925 | ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014] | |
7926 | ||
7927 | * Session Ticket Memory Leak. | |
7928 | ||
7929 | When an OpenSSL SSL/TLS/DTLS server receives a session ticket the | |
7930 | integrity of that ticket is first verified. In the event of a session | |
7931 | ticket integrity check failing, OpenSSL will fail to free memory | |
7932 | causing a memory leak. By sending a large number of invalid session | |
7933 | tickets an attacker could exploit this issue in a Denial Of Service | |
7934 | attack. | |
7935 | ([CVE-2014-3567]) | |
7936 | ||
7937 | *Steve Henson* | |
7938 | ||
7939 | * Build option no-ssl3 is incomplete. | |
7940 | ||
7941 | When OpenSSL is configured with "no-ssl3" as a build option, servers | |
7942 | could accept and complete a SSL 3.0 handshake, and clients could be | |
7943 | configured to send them. | |
7944 | ([CVE-2014-3568]) | |
7945 | ||
7946 | *Akamai and the OpenSSL team* | |
7947 | ||
7948 | * Add support for TLS_FALLBACK_SCSV. | |
7949 | Client applications doing fallback retries should call | |
7950 | SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). | |
7951 | ([CVE-2014-3566]) | |
7952 | ||
7953 | *Adam Langley, Bodo Moeller* | |
7954 | ||
7955 | * Add additional DigestInfo checks. | |
7956 | ||
7957 | Reencode DigestInto in DER and check against the original when | |
7958 | verifying RSA signature: this will reject any improperly encoded | |
7959 | DigestInfo structures. | |
7960 | ||
7961 | Note: this is a precautionary measure and no attacks are currently known. | |
7962 | ||
7963 | *Steve Henson* | |
7964 | ||
7965 | ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014] | |
7966 | ||
7967 | * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject | |
7968 | to a denial of service attack. A malicious server can crash the client | |
7969 | with a null pointer dereference (read) by specifying an anonymous (EC)DH | |
7970 | ciphersuite and sending carefully crafted handshake messages. | |
7971 | ||
7972 | Thanks to Felix Gröbert (Google) for discovering and researching this | |
7973 | issue. | |
7974 | ([CVE-2014-3510]) | |
7975 | ||
7976 | *Emilia Käsper* | |
7977 | ||
7978 | * By sending carefully crafted DTLS packets an attacker could cause openssl | |
7979 | to leak memory. This can be exploited through a Denial of Service attack. | |
7980 | Thanks to Adam Langley for discovering and researching this issue. | |
7981 | ([CVE-2014-3507]) | |
7982 | ||
7983 | *Adam Langley* | |
7984 | ||
7985 | * An attacker can force openssl to consume large amounts of memory whilst | |
7986 | processing DTLS handshake messages. This can be exploited through a | |
7987 | Denial of Service attack. | |
7988 | Thanks to Adam Langley for discovering and researching this issue. | |
7989 | ([CVE-2014-3506]) | |
7990 | ||
7991 | *Adam Langley* | |
7992 | ||
7993 | * An attacker can force an error condition which causes openssl to crash | |
7994 | whilst processing DTLS packets due to memory being freed twice. This | |
7995 | can be exploited through a Denial of Service attack. | |
7996 | Thanks to Adam Langley and Wan-Teh Chang for discovering and researching | |
7997 | this issue. | |
7998 | ([CVE-2014-3505]) | |
7999 | ||
8000 | *Adam Langley* | |
8001 | ||
8002 | * If a multithreaded client connects to a malicious server using a resumed | |
8003 | session and the server sends an ec point format extension it could write | |
8004 | up to 255 bytes to freed memory. | |
8005 | ||
8006 | Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this | |
8007 | issue. | |
8008 | ([CVE-2014-3509]) | |
8009 | ||
8010 | *Gabor Tyukasz* | |
8011 | ||
8012 | * A flaw in OBJ_obj2txt may cause pretty printing functions such as | |
8013 | X509_name_oneline, X509_name_print_ex et al. to leak some information | |
8014 | from the stack. Applications may be affected if they echo pretty printing | |
8015 | output to the attacker. | |
8016 | ||
8017 | Thanks to Ivan Fratric (Google) for discovering this issue. | |
8018 | ([CVE-2014-3508]) | |
8019 | ||
8020 | *Emilia Käsper, and Steve Henson* | |
8021 | ||
8022 | * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) | |
8023 | for corner cases. (Certain input points at infinity could lead to | |
8024 | bogus results, with non-infinity inputs mapped to infinity too.) | |
8025 | ||
8026 | *Bodo Moeller* | |
8027 | ||
8028 | ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014] | |
8029 | ||
8030 | * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted | |
8031 | handshake can force the use of weak keying material in OpenSSL | |
8032 | SSL/TLS clients and servers. | |
8033 | ||
8034 | Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and | |
8035 | researching this issue. ([CVE-2014-0224]) | |
8036 | ||
8037 | *KIKUCHI Masashi, Steve Henson* | |
8038 | ||
8039 | * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an | |
8040 | OpenSSL DTLS client the code can be made to recurse eventually crashing | |
8041 | in a DoS attack. | |
8042 | ||
8043 | Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. | |
8044 | ([CVE-2014-0221]) | |
8045 | ||
8046 | *Imre Rad, Steve Henson* | |
8047 | ||
8048 | * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can | |
8049 | be triggered by sending invalid DTLS fragments to an OpenSSL DTLS | |
8050 | client or server. This is potentially exploitable to run arbitrary | |
8051 | code on a vulnerable client or server. | |
8052 | ||
8053 | Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195]) | |
8054 | ||
8055 | *Jüri Aedla, Steve Henson* | |
8056 | ||
8057 | * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites | |
8058 | are subject to a denial of service attack. | |
8059 | ||
8060 | Thanks to Felix Gröbert and Ivan Fratric at Google for discovering | |
8061 | this issue. ([CVE-2014-3470]) | |
8062 | ||
8063 | *Felix Gröbert, Ivan Fratric, Steve Henson* | |
8064 | ||
8065 | * Harmonize version and its documentation. -f flag is used to display | |
8066 | compilation flags. | |
8067 | ||
8068 | *mancha <mancha1@zoho.com>* | |
8069 | ||
8070 | * Fix eckey_priv_encode so it immediately returns an error upon a failure | |
8071 | in i2d_ECPrivateKey. | |
8072 | ||
8073 | *mancha <mancha1@zoho.com>* | |
8074 | ||
8075 | * Fix some double frees. These are not thought to be exploitable. | |
8076 | ||
8077 | *mancha <mancha1@zoho.com>* | |
8078 | ||
8079 | * Fix for the attack described in the paper "Recovering OpenSSL | |
8080 | ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" | |
8081 | by Yuval Yarom and Naomi Benger. Details can be obtained from: | |
8082 | <http://eprint.iacr.org/2014/140> | |
8083 | ||
8084 | Thanks to Yuval Yarom and Naomi Benger for discovering this | |
8085 | flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076]) | |
8086 | ||
8087 | *Yuval Yarom and Naomi Benger* | |
8088 | ||
8089 | ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014] | |
8090 | ||
8091 | * Keep original DTLS digest and encryption contexts in retransmission | |
8092 | structures so we can use the previous session parameters if they need | |
8093 | to be resent. ([CVE-2013-6450]) | |
8094 | ||
8095 | *Steve Henson* | |
8096 | ||
8097 | * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which | |
8098 | avoids preferring ECDHE-ECDSA ciphers when the client appears to be | |
8099 | Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for | |
8100 | several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug | |
8101 | is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing | |
8102 | 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. | |
8103 | ||
8104 | *Rob Stradling, Adam Langley* | |
8105 | ||
8106 | ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013] | |
8107 | ||
8108 | * Make the decoding of SSLv3, TLS and DTLS CBC records constant time. | |
8109 | ||
8110 | This addresses the flaw in CBC record processing discovered by | |
8111 | Nadhem Alfardan and Kenny Paterson. Details of this attack can be found | |
8112 | at: <http://www.isg.rhul.ac.uk/tls/> | |
8113 | ||
8114 | Thanks go to Nadhem Alfardan and Kenny Paterson of the Information | |
8115 | Security Group at Royal Holloway, University of London | |
8116 | (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and | |
8117 | Emilia Käsper for the initial patch. | |
8118 | ([CVE-2013-0169]) | |
8119 | ||
8120 | *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson* | |
8121 | ||
8122 | * Return an error when checking OCSP signatures when key is NULL. | |
8123 | This fixes a DoS attack. ([CVE-2013-0166]) | |
8124 | ||
8125 | *Steve Henson* | |
8126 | ||
8127 | * Call OCSP Stapling callback after ciphersuite has been chosen, so | |
8128 | the right response is stapled. Also change SSL_get_certificate() | |
8129 | so it returns the certificate actually sent. | |
8130 | See <http://rt.openssl.org/Ticket/Display.html?id=2836>. | |
8131 | (This is a backport) | |
8132 | ||
8133 | *Rob Stradling <rob.stradling@comodo.com>* | |
8134 | ||
8135 | * Fix possible deadlock when decoding public keys. | |
8136 | ||
8137 | *Steve Henson* | |
8138 | ||
8139 | ### Changes between 1.0.0i and 1.0.0j [10 May 2012] | |
8140 | ||
8141 | [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after | |
8142 | OpenSSL 1.0.1.] | |
8143 | ||
8144 | * Sanity check record length before skipping explicit IV in DTLS | |
8145 | to fix DoS attack. | |
8146 | ||
8147 | Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic | |
8148 | fuzzing as a service testing platform. | |
8149 | ([CVE-2012-2333]) | |
8150 | ||
8151 | *Steve Henson* | |
8152 | ||
8153 | * Initialise tkeylen properly when encrypting CMS messages. | |
8154 | Thanks to Solar Designer of Openwall for reporting this issue. | |
8155 | ||
8156 | *Steve Henson* | |
8157 | ||
8158 | ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012] | |
8159 | ||
8160 | * Check for potentially exploitable overflows in asn1_d2i_read_bio | |
8161 | BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer | |
8162 | in CRYPTO_realloc_clean. | |
8163 | ||
8164 | Thanks to Tavis Ormandy, Google Security Team, for discovering this | |
8165 | issue and to Adam Langley <agl@chromium.org> for fixing it. | |
8166 | ([CVE-2012-2110]) | |
8167 | ||
8168 | *Adam Langley (Google), Tavis Ormandy, Google Security Team* | |
8169 | ||
8170 | ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012] | |
8171 | ||
8172 | * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness | |
8173 | in CMS and PKCS7 code. When RSA decryption fails use a random key for | |
8174 | content decryption and always return the same error. Note: this attack | |
8175 | needs on average 2^20 messages so it only affects automated senders. The | |
8176 | old behaviour can be re-enabled in the CMS code by setting the | |
8177 | CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where | |
8178 | an MMA defence is not necessary. | |
8179 | Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering | |
8180 | this issue. ([CVE-2012-0884]) | |
8181 | ||
8182 | *Steve Henson* | |
8183 | ||
8184 | * Fix CVE-2011-4619: make sure we really are receiving a | |
8185 | client hello before rejecting multiple SGC restarts. Thanks to | |
8186 | Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. | |
8187 | ||
8188 | *Steve Henson* | |
8189 | ||
8190 | ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012] | |
8191 | ||
8192 | * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. | |
8193 | Thanks to Antonio Martin, Enterprise Secure Access Research and | |
8194 | Development, Cisco Systems, Inc. for discovering this bug and | |
8195 | preparing a fix. ([CVE-2012-0050]) | |
8196 | ||
8197 | *Antonio Martin* | |
8198 | ||
8199 | ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012] | |
8200 | ||
8201 | * Nadhem Alfardan and Kenny Paterson have discovered an extension | |
8202 | of the Vaudenay padding oracle attack on CBC mode encryption | |
8203 | which enables an efficient plaintext recovery attack against | |
8204 | the OpenSSL implementation of DTLS. Their attack exploits timing | |
8205 | differences arising during decryption processing. A research | |
8206 | paper describing this attack can be found at: | |
8207 | <http://www.isg.rhul.ac.uk/~kp/dtls.pdf> | |
8208 | Thanks go to Nadhem Alfardan and Kenny Paterson of the Information | |
8209 | Security Group at Royal Holloway, University of London | |
8210 | (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann | |
8211 | <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> | |
8212 | for preparing the fix. ([CVE-2011-4108]) | |
8213 | ||
8214 | *Robin Seggelmann, Michael Tuexen* | |
8215 | ||
8216 | * Clear bytes used for block padding of SSL 3.0 records. | |
8217 | ([CVE-2011-4576]) | |
8218 | ||
8219 | *Adam Langley (Google)* | |
8220 | ||
8221 | * Only allow one SGC handshake restart for SSL/TLS. Thanks to George | |
8222 | Kadianakis <desnacked@gmail.com> for discovering this issue and | |
8223 | Adam Langley for preparing the fix. ([CVE-2011-4619]) | |
8224 | ||
8225 | *Adam Langley (Google)* | |
8226 | ||
8227 | * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027]) | |
8228 | ||
8229 | *Andrey Kulikov <amdeich@gmail.com>* | |
8230 | ||
8231 | * Prevent malformed RFC3779 data triggering an assertion failure. | |
8232 | Thanks to Andrew Chi, BBN Technologies, for discovering the flaw | |
8233 | and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577]) | |
8234 | ||
8235 | *Rob Austein <sra@hactrn.net>* | |
8236 | ||
8237 | * Improved PRNG seeding for VOS. | |
8238 | ||
8239 | *Paul Green <Paul.Green@stratus.com>* | |
8240 | ||
8241 | * Fix ssl_ciph.c set-up race. | |
8242 | ||
8243 | *Adam Langley (Google)* | |
8244 | ||
8245 | * Fix spurious failures in ecdsatest.c. | |
8246 | ||
8247 | *Emilia Käsper (Google)* | |
8248 | ||
8249 | * Fix the BIO_f_buffer() implementation (which was mixing different | |
8250 | interpretations of the `..._len` fields). | |
8251 | ||
8252 | *Adam Langley (Google)* | |
8253 | ||
8254 | * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than | |
8255 | BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent | |
8256 | threads won't reuse the same blinding coefficients. | |
8257 | ||
8258 | This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING | |
8259 | lock to call BN_BLINDING_invert_ex, and avoids one use of | |
8260 | BN_BLINDING_update for each BN_BLINDING structure (previously, | |
8261 | the last update always remained unused). | |
8262 | ||
8263 | *Emilia Käsper (Google)* | |
8264 | ||
8265 | * In ssl3_clear, preserve s3->init_extra along with s3->rbuf. | |
8266 | ||
8267 | *Bob Buckholz (Google)* | |
8268 | ||
8269 | ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011] | |
8270 | ||
8271 | * Fix bug where CRLs with nextUpdate in the past are sometimes accepted | |
8272 | by initialising X509_STORE_CTX properly. ([CVE-2011-3207]) | |
8273 | ||
8274 | *Kaspar Brand <ossl@velox.ch>* | |
8275 | ||
8276 | * Fix SSL memory handling for (EC)DH ciphersuites, in particular | |
8277 | for multi-threaded use of ECDH. ([CVE-2011-3210]) | |
8278 | ||
8279 | *Adam Langley (Google)* | |
8280 | ||
8281 | * Fix x509_name_ex_d2i memory leak on bad inputs. | |
8282 | ||
8283 | *Bodo Moeller* | |
8284 | ||
8285 | * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check | |
8286 | signature public key algorithm by using OID xref utilities instead. | |
8287 | Before this you could only use some ECC ciphersuites with SHA1 only. | |
8288 | ||
8289 | *Steve Henson* | |
8290 | ||
8291 | * Add protection against ECDSA timing attacks as mentioned in the paper | |
8292 | by Billy Bob Brumley and Nicola Tuveri, see: | |
8293 | <http://eprint.iacr.org/2011/232.pdf> | |
8294 | ||
8295 | *Billy Bob Brumley and Nicola Tuveri* | |
8296 | ||
8297 | ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011] | |
8298 | ||
8299 | * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 | |
8300 | ||
8301 | *Neel Mehta, Adam Langley, Bodo Moeller (Google)* | |
8302 | ||
8303 | * Fix bug in string printing code: if *any* escaping is enabled we must | |
8304 | escape the escape character (backslash) or the resulting string is | |
8305 | ambiguous. | |
8306 | ||
8307 | *Steve Henson* | |
8308 | ||
8309 | ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010] | |
8310 | ||
8311 | * Disable code workaround for ancient and obsolete Netscape browsers | |
8312 | and servers: an attacker can use it in a ciphersuite downgrade attack. | |
8313 | Thanks to Martin Rex for discovering this bug. CVE-2010-4180 | |
8314 | ||
8315 | *Steve Henson* | |
8316 | ||
8317 | * Fixed J-PAKE implementation error, originally discovered by | |
8318 | Sebastien Martini, further info and confirmation from Stefan | |
8319 | Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 | |
8320 | ||
8321 | *Ben Laurie* | |
8322 | ||
8323 | ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010] | |
8324 | ||
8325 | * Fix extension code to avoid race conditions which can result in a buffer | |
8326 | overrun vulnerability: resumed sessions must not be modified as they can | |
8327 | be shared by multiple threads. CVE-2010-3864 | |
8328 | ||
8329 | *Steve Henson* | |
8330 | ||
8331 | * Fix WIN32 build system to correctly link an ENGINE directory into | |
8332 | a DLL. | |
8333 | ||
8334 | *Steve Henson* | |
8335 | ||
8336 | ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010] | |
8337 | ||
8338 | * Check return value of int_rsa_verify in pkey_rsa_verifyrecover | |
8339 | ([CVE-2010-1633]) | |
8340 | ||
8341 | *Steve Henson, Peter-Michael Hager <hager@dortmund.net>* | |
8342 | ||
8343 | ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010] | |
8344 | ||
8345 | * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher | |
8346 | context. The operation can be customised via the ctrl mechanism in | |
8347 | case ENGINEs want to include additional functionality. | |
8348 | ||
8349 | *Steve Henson* | |
8350 | ||
8351 | * Tolerate yet another broken PKCS#8 key format: private key value negative. | |
8352 | ||
8353 | *Steve Henson* | |
8354 | ||
8355 | * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to | |
8356 | output hashes compatible with older versions of OpenSSL. | |
8357 | ||
8358 | *Willy Weisz <weisz@vcpc.univie.ac.at>* | |
8359 | ||
8360 | * Fix compression algorithm handling: if resuming a session use the | |
8361 | compression algorithm of the resumed session instead of determining | |
8362 | it from client hello again. Don't allow server to change algorithm. | |
8363 | ||
8364 | *Steve Henson* | |
8365 | ||
8366 | * Add load_crls() function to commands tidying load_certs() too. Add option | |
8367 | to verify utility to allow additional CRLs to be included. | |
8368 | ||
8369 | *Steve Henson* | |
8370 | ||
8371 | * Update OCSP request code to permit adding custom headers to the request: | |
8372 | some responders need this. | |
8373 | ||
8374 | *Steve Henson* | |
8375 | ||
8376 | * The function EVP_PKEY_sign() returns <=0 on error: check return code | |
8377 | correctly. | |
8378 | ||
8379 | *Julia Lawall <julia@diku.dk>* | |
8380 | ||
8381 | * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it | |
8382 | needlessly dereferenced structures, used obsolete functions and | |
8383 | didn't handle all updated verify codes correctly. | |
8384 | ||
8385 | *Steve Henson* | |
8386 | ||
8387 | * Disable MD2 in the default configuration. | |
8388 | ||
8389 | *Steve Henson* | |
8390 | ||
8391 | * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to | |
8392 | indicate the initial BIO being pushed or popped. This makes it possible | |
8393 | to determine whether the BIO is the one explicitly called or as a result | |
8394 | of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so | |
8395 | it handles reference counts correctly and doesn't zero out the I/O bio | |
8396 | when it is not being explicitly popped. WARNING: applications which | |
8397 | included workarounds for the old buggy behaviour will need to be modified | |
8398 | or they could free up already freed BIOs. | |
8399 | ||
8400 | *Steve Henson* | |
8401 | ||
8402 | * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni | |
8403 | renaming to all platforms (within the 0.9.8 branch, this was | |
8404 | done conditionally on Netware platforms to avoid a name clash). | |
8405 | ||
8406 | *Guenter <lists@gknw.net>* | |
8407 | ||
8408 | * Add ECDHE and PSK support to DTLS. | |
8409 | ||
8410 | *Michael Tuexen <tuexen@fh-muenster.de>* | |
8411 | ||
8412 | * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't | |
8413 | be used on C++. | |
8414 | ||
8415 | *Steve Henson* | |
8416 | ||
8417 | * Add "missing" function EVP_MD_flags() (without this the only way to | |
8418 | retrieve a digest flags is by accessing the structure directly. Update | |
8419 | `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest | |
8420 | or cipher is registered as in the "from" argument. Print out all | |
8421 | registered digests in the dgst usage message instead of manually | |
8422 | attempting to work them out. | |
8423 | ||
8424 | *Steve Henson* | |
8425 | ||
8426 | * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: | |
8427 | this allows the use of compression and extensions. Change default cipher | |
8428 | string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 | |
8429 | by default unless an application cipher string requests it. | |
8430 | ||
8431 | *Steve Henson* | |
8432 | ||
8433 | * Alter match criteria in PKCS12_parse(). It used to try to use local | |
8434 | key ids to find matching certificates and keys but some PKCS#12 files | |
8435 | don't follow the (somewhat unwritten) rules and this strategy fails. | |
8436 | Now just gather all certificates together and the first private key | |
8437 | then look for the first certificate that matches the key. | |
8438 | ||
8439 | *Steve Henson* | |
8440 | ||
8441 | * Support use of registered digest and cipher names for dgst and cipher | |
8442 | commands instead of having to add each one as a special case. So now | |
8443 | you can do: | |
8444 | ||
8445 | openssl sha256 foo | |
8446 | ||
8447 | as well as: | |
8448 | ||
8449 | openssl dgst -sha256 foo | |
8450 | ||
8451 | and this works for ENGINE based algorithms too. | |
8452 | ||
8453 | *Steve Henson* | |
8454 | ||
8455 | * Update Gost ENGINE to support parameter files. | |
8456 | ||
8457 | *Victor B. Wagner <vitus@cryptocom.ru>* | |
8458 | ||
8459 | * Support GeneralizedTime in ca utility. | |
8460 | ||
8461 | *Oliver Martin <oliver@volatilevoid.net>, Steve Henson* | |
8462 | ||
8463 | * Enhance the hash format used for certificate directory links. The new | |
8464 | form uses the canonical encoding (meaning equivalent names will work | |
8465 | even if they aren't identical) and uses SHA1 instead of MD5. This form | |
8466 | is incompatible with the older format and as a result c_rehash should | |
8467 | be used to rebuild symbolic links. | |
8468 | ||
8469 | *Steve Henson* | |
8470 | ||
8471 | * Make PKCS#8 the default write format for private keys, replacing the | |
8472 | traditional format. This form is standardised, more secure and doesn't | |
8473 | include an implicit MD5 dependency. | |
8474 | ||
8475 | *Steve Henson* | |
8476 | ||
8477 | * Add a $gcc_devteam_warn option to Configure. The idea is that any code | |
8478 | committed to OpenSSL should pass this lot as a minimum. | |
8479 | ||
8480 | *Steve Henson* | |
8481 | ||
8482 | * Add session ticket override functionality for use by EAP-FAST. | |
8483 | ||
8484 | *Jouni Malinen <j@w1.fi>* | |
8485 | ||
8486 | * Modify HMAC functions to return a value. Since these can be implemented | |
8487 | in an ENGINE errors can occur. | |
8488 | ||
8489 | *Steve Henson* | |
8490 | ||
8491 | * Type-checked OBJ_bsearch_ex. | |
8492 | ||
8493 | *Ben Laurie* | |
8494 | ||
8495 | * Type-checked OBJ_bsearch. Also some constification necessitated | |
8496 | by type-checking. Still to come: TXT_DB, bsearch(?), | |
8497 | OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, | |
8498 | CONF_VALUE. | |
8499 | ||
8500 | *Ben Laurie* | |
8501 | ||
8502 | * New function OPENSSL_gmtime_adj() to add a specific number of days and | |
8503 | seconds to a tm structure directly, instead of going through OS | |
8504 | specific date routines. This avoids any issues with OS routines such | |
8505 | as the year 2038 bug. New `*_adj()` functions for ASN1 time structures | |
8506 | and X509_time_adj_ex() to cover the extended range. The existing | |
8507 | X509_time_adj() is still usable and will no longer have any date issues. | |
8508 | ||
8509 | *Steve Henson* | |
8510 | ||
8511 | * Delta CRL support. New use deltas option which will attempt to locate | |
8512 | and search any appropriate delta CRLs available. | |
8513 | ||
8514 | This work was sponsored by Google. | |
8515 | ||
8516 | *Steve Henson* | |
8517 | ||
8518 | * Support for CRLs partitioned by reason code. Reorganise CRL processing | |
8519 | code and add additional score elements. Validate alternate CRL paths | |
8520 | as part of the CRL checking and indicate a new error "CRL path validation | |
8521 | error" in this case. Applications wanting additional details can use | |
8522 | the verify callback and check the new "parent" field. If this is not | |
8523 | NULL CRL path validation is taking place. Existing applications won't | |
8524 | see this because it requires extended CRL support which is off by | |
8525 | default. | |
8526 | ||
8527 | This work was sponsored by Google. | |
8528 | ||
8529 | *Steve Henson* | |
8530 | ||
8531 | * Support for freshest CRL extension. | |
8532 | ||
8533 | This work was sponsored by Google. | |
8534 | ||
8535 | *Steve Henson* | |
8536 | ||
8537 | * Initial indirect CRL support. Currently only supported in the CRLs | |
8538 | passed directly and not via lookup. Process certificate issuer | |
8539 | CRL entry extension and lookup CRL entries by bother issuer name | |
8540 | and serial number. Check and process CRL issuer entry in IDP extension. | |
8541 | ||
8542 | This work was sponsored by Google. | |
8543 | ||
8544 | *Steve Henson* | |
8545 | ||
8546 | * Add support for distinct certificate and CRL paths. The CRL issuer | |
8547 | certificate is validated separately in this case. Only enabled if | |
8548 | an extended CRL support flag is set: this flag will enable additional | |
8549 | CRL functionality in future. | |
8550 | ||
8551 | This work was sponsored by Google. | |
8552 | ||
8553 | *Steve Henson* | |
8554 | ||
8555 | * Add support for policy mappings extension. | |
8556 | ||
8557 | This work was sponsored by Google. | |
8558 | ||
8559 | *Steve Henson* | |
8560 | ||
8561 | * Fixes to pathlength constraint, self issued certificate handling, | |
8562 | policy processing to align with RFC3280 and PKITS tests. | |
8563 | ||
8564 | This work was sponsored by Google. | |
8565 | ||
8566 | *Steve Henson* | |
8567 | ||
8568 | * Support for name constraints certificate extension. DN, email, DNS | |
8569 | and URI types are currently supported. | |
8570 | ||
8571 | This work was sponsored by Google. | |
8572 | ||
8573 | *Steve Henson* | |
8574 | ||
8575 | * To cater for systems that provide a pointer-based thread ID rather | |
8576 | than numeric, deprecate the current numeric thread ID mechanism and | |
8577 | replace it with a structure and associated callback type. This | |
8578 | mechanism allows a numeric "hash" to be extracted from a thread ID in | |
8579 | either case, and on platforms where pointers are larger than 'long', | |
8580 | mixing is done to help ensure the numeric 'hash' is usable even if it | |
8581 | can't be guaranteed unique. The default mechanism is to use "&errno" | |
8582 | as a pointer-based thread ID to distinguish between threads. | |
8583 | ||
8584 | Applications that want to provide their own thread IDs should now use | |
8585 | CRYPTO_THREADID_set_callback() to register a callback that will call | |
8586 | either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). | |
8587 | ||
8588 | Note that ERR_remove_state() is now deprecated, because it is tied | |
8589 | to the assumption that thread IDs are numeric. ERR_remove_state(0) | |
8590 | to free the current thread's error state should be replaced by | |
8591 | ERR_remove_thread_state(NULL). | |
8592 | ||
8593 | (This new approach replaces the functions CRYPTO_set_idptr_callback(), | |
8594 | CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in | |
8595 | OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an | |
8596 | application was previously providing a numeric thread callback that | |
8597 | was inappropriate for distinguishing threads, then uniqueness might | |
8598 | have been obtained with &errno that happened immediately in the | |
8599 | intermediate development versions of OpenSSL; this is no longer the | |
8600 | case, the numeric thread callback will now override the automatic use | |
8601 | of &errno.) | |
8602 | ||
8603 | *Geoff Thorpe, with help from Bodo Moeller* | |
8604 | ||
8605 | * Initial support for different CRL issuing certificates. This covers a | |
8606 | simple case where the self issued certificates in the chain exist and | |
8607 | the real CRL issuer is higher in the existing chain. | |
8608 | ||
8609 | This work was sponsored by Google. | |
8610 | ||
8611 | *Steve Henson* | |
8612 | ||
8613 | * Removed effectively defunct crypto/store from the build. | |
8614 | ||
8615 | *Ben Laurie* | |
8616 | ||
8617 | * Revamp of STACK to provide stronger type-checking. Still to come: | |
8618 | TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, | |
8619 | ASN1_STRING, CONF_VALUE. | |
8620 | ||
8621 | *Ben Laurie* | |
8622 | ||
8623 | * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer | |
8624 | RAM on SSL connections. This option can save about 34k per idle SSL. | |
8625 | ||
8626 | *Nick Mathewson* | |
8627 | ||
8628 | * Revamp of LHASH to provide stronger type-checking. Still to come: | |
8629 | STACK, TXT_DB, bsearch, qsort. | |
8630 | ||
8631 | *Ben Laurie* | |
8632 | ||
8633 | * Initial support for Cryptographic Message Syntax (aka CMS) based | |
8634 | on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, | |
8635 | support for data, signedData, compressedData, digestedData and | |
8636 | encryptedData, envelopedData types included. Scripts to check against | |
8637 | RFC4134 examples draft and interop and consistency checks of many | |
8638 | content types and variants. | |
8639 | ||
8640 | *Steve Henson* | |
8641 | ||
8642 | * Add options to enc utility to support use of zlib compression BIO. | |
8643 | ||
8644 | *Steve Henson* | |
8645 | ||
8646 | * Extend mk1mf to support importing of options and assembly language | |
8647 | files from Configure script, currently only included in VC-WIN32. | |
8648 | The assembly language rules can now optionally generate the source | |
8649 | files from the associated perl scripts. | |
8650 | ||
8651 | *Steve Henson* | |
8652 | ||
8653 | * Implement remaining functionality needed to support GOST ciphersuites. | |
8654 | Interop testing has been performed using CryptoPro implementations. | |
8655 | ||
8656 | *Victor B. Wagner <vitus@cryptocom.ru>* | |
8657 | ||
8658 | * s390x assembler pack. | |
8659 | ||
8660 | *Andy Polyakov* | |
8661 | ||
8662 | * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU | |
8663 | "family." | |
8664 | ||
8665 | *Andy Polyakov* | |
8666 | ||
8667 | * Implement Opaque PRF Input TLS extension as specified in | |
8668 | draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an | |
8669 | official specification yet and no extension type assignment by | |
8670 | IANA exists, this extension (for now) will have to be explicitly | |
8671 | enabled when building OpenSSL by providing the extension number | |
8672 | to use. For example, specify an option | |
8673 | ||
8674 | -DTLSEXT_TYPE_opaque_prf_input=0x9527 | |
8675 | ||
8676 | to the "config" or "Configure" script to enable the extension, | |
8677 | assuming extension number 0x9527 (which is a completely arbitrary | |
8678 | and unofficial assignment based on the MD5 hash of the Internet | |
8679 | Draft). Note that by doing so, you potentially lose | |
8680 | interoperability with other TLS implementations since these might | |
8681 | be using the same extension number for other purposes. | |
8682 | ||
8683 | SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the | |
8684 | opaque PRF input value to use in the handshake. This will create | |
8685 | an internal copy of the length-'len' string at 'src', and will | |
8686 | return non-zero for success. | |
8687 | ||
8688 | To get more control and flexibility, provide a callback function | |
8689 | by using | |
8690 | ||
8691 | SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) | |
8692 | SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) | |
8693 | ||
8694 | where | |
8695 | ||
8696 | int (*cb)(SSL *, void *peerinput, size_t len, void *arg); | |
8697 | void *arg; | |
8698 | ||
8699 | Callback function 'cb' will be called in handshakes, and is | |
8700 | expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. | |
8701 | Argument 'arg' is for application purposes (the value as given to | |
8702 | SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly | |
8703 | be provided to the callback function). The callback function | |
8704 | has to return non-zero to report success: usually 1 to use opaque | |
8705 | PRF input just if possible, or 2 to enforce use of the opaque PRF | |
8706 | input. In the latter case, the library will abort the handshake | |
8707 | if opaque PRF input is not successfully negotiated. | |
8708 | ||
8709 | Arguments 'peerinput' and 'len' given to the callback function | |
8710 | will always be NULL and 0 in the case of a client. A server will | |
8711 | see the client's opaque PRF input through these variables if | |
8712 | available (NULL and 0 otherwise). Note that if the server | |
8713 | provides an opaque PRF input, the length must be the same as the | |
8714 | length of the client's opaque PRF input. | |
8715 | ||
8716 | Note that the callback function will only be called when creating | |
8717 | a new session (session resumption can resume whatever was | |
8718 | previously negotiated), and will not be called in SSL 2.0 | |
8719 | handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or | |
8720 | SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended | |
8721 | for applications that need to enforce opaque PRF input. | |
8722 | ||
8723 | *Bodo Moeller* | |
8724 | ||
8725 | * Update ssl code to support digests other than SHA1+MD5 for handshake | |
8726 | MAC. | |
8727 | ||
8728 | *Victor B. Wagner <vitus@cryptocom.ru>* | |
8729 | ||
8730 | * Add RFC4507 support to OpenSSL. This includes the corrections in | |
8731 | RFC4507bis. The encrypted ticket format is an encrypted encoded | |
8732 | SSL_SESSION structure, that way new session features are automatically | |
8733 | supported. | |
8734 | ||
8735 | If a client application caches session in an SSL_SESSION structure | |
8736 | support is transparent because tickets are now stored in the encoded | |
8737 | SSL_SESSION. | |
8738 | ||
8739 | The SSL_CTX structure automatically generates keys for ticket | |
8740 | protection in servers so again support should be possible | |
8741 | with no application modification. | |
8742 | ||
8743 | If a client or server wishes to disable RFC4507 support then the option | |
8744 | SSL_OP_NO_TICKET can be set. | |
8745 | ||
8746 | Add a TLS extension debugging callback to allow the contents of any client | |
8747 | or server extensions to be examined. | |
8748 | ||
8749 | This work was sponsored by Google. | |
8750 | ||
8751 | *Steve Henson* | |
8752 | ||
8753 | * Final changes to avoid use of pointer pointer casts in OpenSSL. | |
8754 | OpenSSL should now compile cleanly on gcc 4.2 | |
8755 | ||
8756 | *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson* | |
8757 | ||
8758 | * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC | |
8759 | support including streaming MAC support: this is required for GOST | |
8760 | ciphersuite support. | |
8761 | ||
8762 | *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson* | |
8763 | ||
8764 | * Add option -stream to use PKCS#7 streaming in smime utility. New | |
8765 | function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() | |
8766 | to output in BER and PEM format. | |
8767 | ||
8768 | *Steve Henson* | |
8769 | ||
8770 | * Experimental support for use of HMAC via EVP_PKEY interface. This | |
8771 | allows HMAC to be handled via the `EVP_DigestSign*()` interface. The | |
8772 | EVP_PKEY "key" in this case is the HMAC key, potentially allowing | |
8773 | ENGINE support for HMAC keys which are unextractable. New -mac and | |
8774 | -macopt options to dgst utility. | |
8775 | ||
8776 | *Steve Henson* | |
8777 | ||
8778 | * New option -sigopt to dgst utility. Update dgst to use | |
8779 | `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use | |
8780 | alternative signing parameters such as X9.31 or PSS in the dgst | |
8781 | utility. | |
8782 | ||
8783 | *Steve Henson* | |
8784 | ||
8785 | * Change ssl_cipher_apply_rule(), the internal function that does | |
8786 | the work each time a ciphersuite string requests enabling | |
8787 | ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or | |
8788 | removing ("!foo+bar") a class of ciphersuites: Now it maintains | |
8789 | the order of disabled ciphersuites such that those ciphersuites | |
8790 | that most recently went from enabled to disabled not only stay | |
8791 | in order with respect to each other, but also have higher priority | |
8792 | than other disabled ciphersuites the next time ciphersuites are | |
8793 | enabled again. | |
8794 | ||
8795 | This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable | |
8796 | the same ciphersuites as with "HIGH" alone, but in a specific | |
8797 | order where the PSK ciphersuites come first (since they are the | |
8798 | most recently disabled ciphersuites when "HIGH" is parsed). | |
8799 | ||
8800 | Also, change ssl_create_cipher_list() (using this new | |
8801 | functionality) such that between otherwise identical | |
8802 | ciphersuites, ephemeral ECDH is preferred over ephemeral DH in | |
8803 | the default order. | |
8804 | ||
8805 | *Bodo Moeller* | |
8806 | ||
8807 | * Change ssl_create_cipher_list() so that it automatically | |
8808 | arranges the ciphersuites in reasonable order before starting | |
8809 | to process the rule string. Thus, the definition for "DEFAULT" | |
8810 | (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but | |
8811 | remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`. | |
8812 | This makes it much easier to arrive at a reasonable default order | |
8813 | in applications for which anonymous ciphers are OK (meaning | |
8814 | that you can't actually use DEFAULT). | |
8815 | ||
8816 | *Bodo Moeller; suggested by Victor Duchovni* | |
8817 | ||
8818 | * Split the SSL/TLS algorithm mask (as used for ciphersuite string | |
8819 | processing) into multiple integers instead of setting | |
8820 | "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", | |
8821 | "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. | |
8822 | (These masks as well as the individual bit definitions are hidden | |
8823 | away into the non-exported interface ssl/ssl_locl.h, so this | |
8824 | change to the definition of the SSL_CIPHER structure shouldn't | |
8825 | affect applications.) This give us more bits for each of these | |
8826 | categories, so there is no longer a need to coagulate AES128 and | |
8827 | AES256 into a single algorithm bit, and to coagulate Camellia128 | |
8828 | and Camellia256 into a single algorithm bit, which has led to all | |
8829 | kinds of kludges. | |
8830 | ||
8831 | Thus, among other things, the kludge introduced in 0.9.7m and | |
8832 | 0.9.8e for masking out AES256 independently of AES128 or masking | |
8833 | out Camellia256 independently of AES256 is not needed here in 0.9.9. | |
8834 | ||
8835 | With the change, we also introduce new ciphersuite aliases that | |
8836 | so far were missing: "AES128", "AES256", "CAMELLIA128", and | |
8837 | "CAMELLIA256". | |
8838 | ||
8839 | *Bodo Moeller* | |
8840 | ||
8841 | * Add support for dsa-with-SHA224 and dsa-with-SHA256. | |
8842 | Use the leftmost N bytes of the signature input if the input is | |
8843 | larger than the prime q (with N being the size in bytes of q). | |
8844 | ||
8845 | *Nils Larsch* | |
8846 | ||
8847 | * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses | |
8848 | it yet and it is largely untested. | |
8849 | ||
8850 | *Steve Henson* | |
8851 | ||
8852 | * Add support for the ecdsa-with-SHA224/256/384/512 signature types. | |
8853 | ||
8854 | *Nils Larsch* | |
8855 | ||
8856 | * Initial incomplete changes to avoid need for function casts in OpenSSL | |
8857 | some compilers (gcc 4.2 and later) reject their use. Safestack is | |
8858 | reimplemented. Update ASN1 to avoid use of legacy functions. | |
8859 | ||
8860 | *Steve Henson* | |
8861 | ||
8862 | * Win32/64 targets are linked with Winsock2. | |
8863 | ||
8864 | *Andy Polyakov* | |
8865 | ||
8866 | * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected | |
8867 | to external functions. This can be used to increase CRL handling | |
8868 | efficiency especially when CRLs are very large by (for example) storing | |
8869 | the CRL revoked certificates in a database. | |
8870 | ||
8871 | *Steve Henson* | |
8872 | ||
8873 | * Overhaul of by_dir code. Add support for dynamic loading of CRLs so | |
8874 | new CRLs added to a directory can be used. New command line option | |
8875 | -verify_return_error to s_client and s_server. This causes real errors | |
8876 | to be returned by the verify callback instead of carrying on no matter | |
8877 | what. This reflects the way a "real world" verify callback would behave. | |
8878 | ||
8879 | *Steve Henson* | |
8880 | ||
8881 | * GOST engine, supporting several GOST algorithms and public key formats. | |
8882 | Kindly donated by Cryptocom. | |
8883 | ||
8884 | *Cryptocom* | |
8885 | ||
8886 | * Partial support for Issuing Distribution Point CRL extension. CRLs | |
8887 | partitioned by DP are handled but no indirect CRL or reason partitioning | |
8888 | (yet). Complete overhaul of CRL handling: now the most suitable CRL is | |
8889 | selected via a scoring technique which handles IDP and AKID in CRLs. | |
8890 | ||
8891 | *Steve Henson* | |
8892 | ||
8893 | * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which | |
8894 | will ultimately be used for all verify operations: this will remove the | |
8895 | X509_STORE dependency on certificate verification and allow alternative | |
8896 | lookup methods. X509_STORE based implementations of these two callbacks. | |
8897 | ||
8898 | *Steve Henson* | |
8899 | ||
8900 | * Allow multiple CRLs to exist in an X509_STORE with matching issuer names. | |
8901 | Modify get_crl() to find a valid (unexpired) CRL if possible. | |
8902 | ||
8903 | *Steve Henson* | |
8904 | ||
8905 | * New function X509_CRL_match() to check if two CRLs are identical. Normally | |
8906 | this would be called X509_CRL_cmp() but that name is already used by | |
8907 | a function that just compares CRL issuer names. Cache several CRL | |
8908 | extensions in X509_CRL structure and cache CRLDP in X509. | |
8909 | ||
8910 | *Steve Henson* | |
8911 | ||
8912 | * Store a "canonical" representation of X509_NAME structure (ASN1 Name) | |
8913 | this maps equivalent X509_NAME structures into a consistent structure. | |
8914 | Name comparison can then be performed rapidly using memcmp(). | |
8915 | ||
8916 | *Steve Henson* | |
8917 | ||
8918 | * Non-blocking OCSP request processing. Add -timeout option to ocsp | |
8919 | utility. | |
8920 | ||
8921 | *Steve Henson* | |
8922 | ||
8923 | * Allow digests to supply their own micalg string for S/MIME type using | |
8924 | the ctrl EVP_MD_CTRL_MICALG. | |
8925 | ||
8926 | *Steve Henson* | |
8927 | ||
8928 | * During PKCS7 signing pass the PKCS7 SignerInfo structure to the | |
8929 | EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN | |
8930 | ctrl. It can then customise the structure before and/or after signing | |
8931 | if necessary. | |
8932 | ||
8933 | *Steve Henson* | |
8934 | ||
8935 | * New function OBJ_add_sigid() to allow application defined signature OIDs | |
8936 | to be added to OpenSSLs internal tables. New function OBJ_sigid_free() | |
8937 | to free up any added signature OIDs. | |
8938 | ||
8939 | *Steve Henson* | |
8940 | ||
8941 | * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), | |
8942 | EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal | |
8943 | digest and cipher tables. New options added to openssl utility: | |
8944 | list-message-digest-algorithms and list-cipher-algorithms. | |
8945 | ||
8946 | *Steve Henson* | |
8947 | ||
8948 | * Change the array representation of binary polynomials: the list | |
8949 | of degrees of non-zero coefficients is now terminated with -1. | |
8950 | Previously it was terminated with 0, which was also part of the | |
8951 | value; thus, the array representation was not applicable to | |
8952 | polynomials where t^0 has coefficient zero. This change makes | |
8953 | the array representation useful in a more general context. | |
8954 | ||
8955 | *Douglas Stebila* | |
8956 | ||
8957 | * Various modifications and fixes to SSL/TLS cipher string | |
8958 | handling. For ECC, the code now distinguishes between fixed ECDH | |
8959 | with RSA certificates on the one hand and with ECDSA certificates | |
8960 | on the other hand, since these are separate ciphersuites. The | |
8961 | unused code for Fortezza ciphersuites has been removed. | |
8962 | ||
8963 | For consistency with EDH, ephemeral ECDH is now called "EECDH" | |
8964 | (not "ECDHE"). For consistency with the code for DH | |
8965 | certificates, use of ECDH certificates is now considered ECDH | |
8966 | authentication, not RSA or ECDSA authentication (the latter is | |
8967 | merely the CA's signing algorithm and not actively used in the | |
8968 | protocol). | |
8969 | ||
8970 | The temporary ciphersuite alias "ECCdraft" is no longer | |
8971 | available, and ECC ciphersuites are no longer excluded from "ALL" | |
8972 | and "DEFAULT". The following aliases now exist for RFC 4492 | |
8973 | ciphersuites, most of these by analogy with the DH case: | |
8974 | ||
8975 | kECDHr - ECDH cert, signed with RSA | |
8976 | kECDHe - ECDH cert, signed with ECDSA | |
8977 | kECDH - ECDH cert (signed with either RSA or ECDSA) | |
8978 | kEECDH - ephemeral ECDH | |
8979 | ECDH - ECDH cert or ephemeral ECDH | |
8980 | ||
8981 | aECDH - ECDH cert | |
8982 | aECDSA - ECDSA cert | |
8983 | ECDSA - ECDSA cert | |
8984 | ||
8985 | AECDH - anonymous ECDH | |
8986 | EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") | |
8987 | ||
8988 | *Bodo Moeller* | |
8989 | ||
8990 | * Add additional S/MIME capabilities for AES and GOST ciphers if supported. | |
8991 | Use correct micalg parameters depending on digest(s) in signed message. | |
8992 | ||
8993 | *Steve Henson* | |
8994 | ||
8995 | * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process | |
8996 | an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. | |
8997 | ||
8998 | *Steve Henson* | |
8999 | ||
9000 | * Initial engine support for EVP_PKEY_METHOD. New functions to permit | |
9001 | an engine to register a method. Add ENGINE lookups for methods and | |
9002 | functional reference processing. | |
9003 | ||
9004 | *Steve Henson* | |
9005 | ||
9006 | * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of | |
9007 | `EVP_{Sign,Verify}*` which allow an application to customise the signature | |
9008 | process. | |
9009 | ||
9010 | *Steve Henson* | |
9011 | ||
9012 | * New -resign option to smime utility. This adds one or more signers | |
9013 | to an existing PKCS#7 signedData structure. Also -md option to use an | |
9014 | alternative message digest algorithm for signing. | |
9015 | ||
9016 | *Steve Henson* | |
9017 | ||
9018 | * Tidy up PKCS#7 routines and add new functions to make it easier to | |
9019 | create PKCS7 structures containing multiple signers. Update smime | |
9020 | application to support multiple signers. | |
9021 | ||
9022 | *Steve Henson* | |
9023 | ||
9024 | * New -macalg option to pkcs12 utility to allow setting of an alternative | |
9025 | digest MAC. | |
9026 | ||
9027 | *Steve Henson* | |
9028 | ||
9029 | * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. | |
9030 | Reorganize PBE internals to lookup from a static table using NIDs, | |
9031 | add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: | |
9032 | EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative | |
9033 | PRF which will be automatically used with PBES2. | |
9034 | ||
9035 | *Steve Henson* | |
9036 | ||
9037 | * Replace the algorithm specific calls to generate keys in "req" with the | |
9038 | new API. | |
9039 | ||
9040 | *Steve Henson* | |
9041 | ||
9042 | * Update PKCS#7 enveloped data routines to use new API. This is now | |
9043 | supported by any public key method supporting the encrypt operation. A | |
9044 | ctrl is added to allow the public key algorithm to examine or modify | |
9045 | the PKCS#7 RecipientInfo structure if it needs to: for RSA this is | |
9046 | a no op. | |
9047 | ||
9048 | *Steve Henson* | |
9049 | ||
9050 | * Add a ctrl to asn1 method to allow a public key algorithm to express | |
9051 | a default digest type to use. In most cases this will be SHA1 but some | |
9052 | algorithms (such as GOST) need to specify an alternative digest. The | |
9053 | return value indicates how strong the preference is 1 means optional and | |
9054 | 2 is mandatory (that is it is the only supported type). Modify | |
9055 | ASN1_item_sign() to accept a NULL digest argument to indicate it should | |
9056 | use the default md. Update openssl utilities to use the default digest | |
9057 | type for signing if it is not explicitly indicated. | |
9058 | ||
9059 | *Steve Henson* | |
9060 | ||
9061 | * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New | |
9062 | EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant | |
9063 | signing method from the key type. This effectively removes the link | |
9064 | between digests and public key types. | |
9065 | ||
9066 | *Steve Henson* | |
9067 | ||
9068 | * Add an OID cross reference table and utility functions. Its purpose is to | |
9069 | translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, | |
9070 | rsaEncryption. This will allow some of the algorithm specific hackery | |
9071 | needed to use the correct OID to be removed. | |
9072 | ||
9073 | *Steve Henson* | |
9074 | ||
9075 | * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO | |
9076 | structures for PKCS7_sign(). They are now set up by the relevant public | |
9077 | key ASN1 method. | |
9078 | ||
9079 | *Steve Henson* | |
9080 | ||
9081 | * Add provisional EC pkey method with support for ECDSA and ECDH. | |
9082 | ||
9083 | *Steve Henson* | |
9084 | ||
9085 | * Add support for key derivation (agreement) in the API, DH method and | |
9086 | pkeyutl. | |
9087 | ||
9088 | *Steve Henson* | |
9089 | ||
9090 | * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support | |
9091 | public and private key formats. As a side effect these add additional | |
9092 | command line functionality not previously available: DSA signatures can be | |
9093 | generated and verified using pkeyutl and DH key support and generation in | |
9094 | pkey, genpkey. | |
9095 | ||
9096 | *Steve Henson* | |
9097 | ||
9098 | * BeOS support. | |
9099 | ||
9100 | *Oliver Tappe <zooey@hirschkaefer.de>* | |
9101 | ||
9102 | * New make target "install_html_docs" installs HTML renditions of the | |
9103 | manual pages. | |
9104 | ||
9105 | *Oliver Tappe <zooey@hirschkaefer.de>* | |
9106 | ||
9107 | * New utility "genpkey" this is analogous to "genrsa" etc except it can | |
9108 | generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to | |
9109 | support key and parameter generation and add initial key generation | |
9110 | functionality for RSA. | |
9111 | ||
9112 | *Steve Henson* | |
9113 | ||
9114 | * Add functions for main EVP_PKEY_method operations. The undocumented | |
9115 | functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to | |
9116 | `EVP_PKEY_{encrypt,decrypt}_old`. | |
9117 | ||
9118 | *Steve Henson* | |
9119 | ||
9120 | * Initial definitions for EVP_PKEY_METHOD. This will be a high level public | |
9121 | key API, doesn't do much yet. | |
9122 | ||
9123 | *Steve Henson* | |
9124 | ||
9125 | * New function EVP_PKEY_asn1_get0_info() to retrieve information about | |
9126 | public key algorithms. New option to openssl utility: | |
9127 | "list-public-key-algorithms" to print out info. | |
9128 | ||
9129 | *Steve Henson* | |
9130 | ||
9131 | * Implement the Supported Elliptic Curves Extension for | |
9132 | ECC ciphersuites from draft-ietf-tls-ecc-12.txt. | |
9133 | ||
9134 | *Douglas Stebila* | |
9135 | ||
9136 | * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or | |
9137 | EVP_CIPHER structures to avoid later problems in EVP_cleanup(). | |
9138 | ||
9139 | *Steve Henson* | |
9140 | ||
9141 | * New utilities pkey and pkeyparam. These are similar to algorithm specific | |
9142 | utilities such as rsa, dsa, dsaparam etc except they process any key | |
9143 | type. | |
9144 | ||
9145 | *Steve Henson* | |
9146 | ||
9147 | * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New | |
9148 | functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), | |
9149 | EVP_PKEY_print_param() to print public key data from an EVP_PKEY | |
9150 | structure. | |
9151 | ||
9152 | *Steve Henson* | |
9153 | ||
9154 | * Initial support for pluggable public key ASN1. | |
9155 | De-spaghettify the public key ASN1 handling. Move public and private | |
9156 | key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate | |
9157 | algorithm specific handling to a single module within the relevant | |
9158 | algorithm directory. Add functions to allow (near) opaque processing | |
9159 | of public and private key structures. | |
9160 | ||
9161 | *Steve Henson* | |
9162 | ||
9163 | * Implement the Supported Point Formats Extension for | |
9164 | ECC ciphersuites from draft-ietf-tls-ecc-12.txt. | |
9165 | ||
9166 | *Douglas Stebila* | |
9167 | ||
9168 | * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members | |
9169 | for the psk identity [hint] and the psk callback functions to the | |
9170 | SSL_SESSION, SSL and SSL_CTX structure. | |
9171 | ||
9172 | New ciphersuites: | |
9173 | PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, | |
9174 | PSK-AES256-CBC-SHA | |
9175 | ||
9176 | New functions: | |
9177 | SSL_CTX_use_psk_identity_hint | |
9178 | SSL_get_psk_identity_hint | |
9179 | SSL_get_psk_identity | |
9180 | SSL_use_psk_identity_hint | |
9181 | ||
9182 | *Mika Kousa and Pasi Eronen of Nokia Corporation* | |
9183 | ||
9184 | * Add RFC 3161 compliant time stamp request creation, response generation | |
9185 | and response verification functionality. | |
9186 | ||
9187 | *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project* | |
9188 | ||
9189 | * Add initial support for TLS extensions, specifically for the server_name | |
9190 | extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now | |
9191 | have new members for a host name. The SSL data structure has an | |
9192 | additional member `SSL_CTX *initial_ctx` so that new sessions can be | |
9193 | stored in that context to allow for session resumption, even after the | |
9194 | SSL has been switched to a new SSL_CTX in reaction to a client's | |
9195 | server_name extension. | |
9196 | ||
9197 | New functions (subject to change): | |
9198 | ||
9199 | SSL_get_servername() | |
9200 | SSL_get_servername_type() | |
9201 | SSL_set_SSL_CTX() | |
9202 | ||
9203 | New CTRL codes and macros (subject to change): | |
9204 | ||
9205 | SSL_CTRL_SET_TLSEXT_SERVERNAME_CB | |
9206 | - SSL_CTX_set_tlsext_servername_callback() | |
9207 | SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG | |
9208 | - SSL_CTX_set_tlsext_servername_arg() | |
9209 | SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() | |
9210 | ||
9211 | openssl s_client has a new '-servername ...' option. | |
9212 | ||
9213 | openssl s_server has new options '-servername_host ...', '-cert2 ...', | |
9214 | '-key2 ...', '-servername_fatal' (subject to change). This allows | |
9215 | testing the HostName extension for a specific single host name ('-cert' | |
9216 | and '-key' remain fallbacks for handshakes without HostName | |
9217 | negotiation). If the unrecognized_name alert has to be sent, this by | |
9218 | default is a warning; it becomes fatal with the '-servername_fatal' | |
9219 | option. | |
9220 | ||
9221 | *Peter Sylvester, Remy Allais, Christophe Renou* | |
9222 | ||
9223 | * Whirlpool hash implementation is added. | |
9224 | ||
9225 | *Andy Polyakov* | |
9226 | ||
9227 | * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to | |
9228 | bn(64,32). Because of instruction set limitations it doesn't have | |
9229 | any negative impact on performance. This was done mostly in order | |
9230 | to make it possible to share assembler modules, such as bn_mul_mont | |
9231 | implementations, between 32- and 64-bit builds without hassle. | |
9232 | ||
9233 | *Andy Polyakov* | |
9234 | ||
9235 | * Move code previously exiled into file crypto/ec/ec2_smpt.c | |
9236 | to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP | |
9237 | macro. | |
9238 | ||
9239 | *Bodo Moeller* | |
9240 | ||
9241 | * New candidate for BIGNUM assembler implementation, bn_mul_mont, | |
9242 | dedicated Montgomery multiplication procedure, is introduced. | |
9243 | BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher | |
9244 | "64-bit" performance on certain 32-bit targets. | |
9245 | ||
9246 | *Andy Polyakov* | |
9247 | ||
9248 | * New option SSL_OP_NO_COMP to disable use of compression selectively | |
9249 | in SSL structures. New SSL ctrl to set maximum send fragment size. | |
9250 | Save memory by setting the I/O buffer sizes dynamically instead of | |
9251 | using the maximum available value. | |
9252 | ||
9253 | *Steve Henson* | |
9254 | ||
9255 | * New option -V for 'openssl ciphers'. This prints the ciphersuite code | |
9256 | in addition to the text details. | |
9257 | ||
9258 | *Bodo Moeller* | |
9259 | ||
9260 | * Very, very preliminary EXPERIMENTAL support for printing of general | |
9261 | ASN1 structures. This currently produces rather ugly output and doesn't | |
9262 | handle several customised structures at all. | |
9263 | ||
9264 | *Steve Henson* | |
9265 | ||
9266 | * Integrated support for PVK file format and some related formats such | |
9267 | as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support | |
9268 | these in the 'rsa' and 'dsa' utilities. | |
9269 | ||
9270 | *Steve Henson* | |
9271 | ||
9272 | * Support for PKCS#1 RSAPublicKey format on rsa utility command line. | |
9273 | ||
9274 | *Steve Henson* | |
9275 | ||
9276 | * Remove the ancient ASN1_METHOD code. This was only ever used in one | |
9277 | place for the (very old) "NETSCAPE" format certificates which are now | |
9278 | handled using new ASN1 code equivalents. | |
9279 | ||
9280 | *Steve Henson* | |
9281 | ||
9282 | * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD | |
9283 | pointer and make the SSL_METHOD parameter in SSL_CTX_new, | |
9284 | SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. | |
9285 | ||
9286 | *Nils Larsch* | |
9287 | ||
9288 | * Modify CRL distribution points extension code to print out previously | |
9289 | unsupported fields. Enhance extension setting code to allow setting of | |
9290 | all fields. | |
9291 | ||
9292 | *Steve Henson* | |
9293 | ||
9294 | * Add print and set support for Issuing Distribution Point CRL extension. | |
9295 | ||
9296 | *Steve Henson* | |
9297 | ||
9298 | * Change 'Configure' script to enable Camellia by default. | |
9299 | ||
9300 | *NTT* | |
9301 | ||
9302 | OpenSSL 0.9.x | |
9303 | ------------- | |
9304 | ||
9305 | ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010] | |
9306 | ||
9307 | * When rejecting SSL/TLS records due to an incorrect version number, never | |
9308 | update s->server with a new major version number. As of | |
9309 | - OpenSSL 0.9.8m if 'short' is a 16-bit type, | |
9310 | - OpenSSL 0.9.8f if 'short' is longer than 16 bits, | |
9311 | the previous behavior could result in a read attempt at NULL when | |
9312 | receiving specific incorrect SSL/TLS records once record payload | |
9313 | protection is active. ([CVE-2010-0740]) | |
9314 | ||
9315 | *Bodo Moeller, Adam Langley <agl@chromium.org>* | |
9316 | ||
9317 | * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL | |
9318 | could be crashed if the relevant tables were not present (e.g. chrooted). | |
9319 | ||
9320 | *Tomas Hoger <thoger@redhat.com>* | |
9321 | ||
9322 | ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010] | |
9323 | ||
9324 | * Always check bn_wexpand() return values for failure. ([CVE-2009-3245]) | |
9325 | ||
9326 | *Martin Olsson, Neel Mehta* | |
9327 | ||
9328 | * Fix X509_STORE locking: Every 'objs' access requires a lock (to | |
9329 | accommodate for stack sorting, always a write lock!). | |
9330 | ||
9331 | *Bodo Moeller* | |
9332 | ||
9333 | * On some versions of WIN32 Heap32Next is very slow. This can cause | |
9334 | excessive delays in the RAND_poll(): over a minute. As a workaround | |
9335 | include a time check in the inner Heap32Next loop too. | |
9336 | ||
9337 | *Steve Henson* | |
9338 | ||
9339 | * The code that handled flushing of data in SSL/TLS originally used the | |
9340 | BIO_CTRL_INFO ctrl to see if any data was pending first. This caused | |
9341 | the problem outlined in PR#1949. The fix suggested there however can | |
9342 | trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions | |
9343 | of Apache). So instead simplify the code to flush unconditionally. | |
9344 | This should be fine since flushing with no data to flush is a no op. | |
9345 | ||
9346 | *Steve Henson* | |
9347 | ||
9348 | * Handle TLS versions 2.0 and later properly and correctly use the | |
9349 | highest version of TLS/SSL supported. Although TLS >= 2.0 is some way | |
9350 | off ancient servers have a habit of sticking around for a while... | |
9351 | ||
9352 | *Steve Henson* | |
9353 | ||
9354 | * Modify compression code so it frees up structures without using the | |
9355 | ex_data callbacks. This works around a problem where some applications | |
9356 | call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when | |
9357 | restarting) then use compression (e.g. SSL with compression) later. | |
9358 | This results in significant per-connection memory leaks and | |
9359 | has caused some security issues including CVE-2008-1678 and | |
9360 | CVE-2009-4355. | |
9361 | ||
9362 | *Steve Henson* | |
9363 | ||
9364 | * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't | |
9365 | change when encrypting or decrypting. | |
9366 | ||
9367 | *Bodo Moeller* | |
9368 | ||
9369 | * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to | |
9370 | connect and renegotiate with servers which do not support RI. | |
9371 | Until RI is more widely deployed this option is enabled by default. | |
9372 | ||
9373 | *Steve Henson* | |
9374 | ||
9375 | * Add "missing" ssl ctrls to clear options and mode. | |
9376 | ||
9377 | *Steve Henson* | |
9378 | ||
9379 | * If client attempts to renegotiate and doesn't support RI respond with | |
9380 | a no_renegotiation alert as required by RFC5746. Some renegotiating | |
9381 | TLS clients will continue a connection gracefully when they receive | |
9382 | the alert. Unfortunately OpenSSL mishandled this alert and would hang | |
9383 | waiting for a server hello which it will never receive. Now we treat a | |
9384 | received no_renegotiation alert as a fatal error. This is because | |
9385 | applications requesting a renegotiation might well expect it to succeed | |
9386 | and would have no code in place to handle the server denying it so the | |
9387 | only safe thing to do is to terminate the connection. | |
9388 | ||
9389 | *Steve Henson* | |
9390 | ||
9391 | * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if | |
9392 | peer supports secure renegotiation and 0 otherwise. Print out peer | |
9393 | renegotiation support in s_client/s_server. | |
9394 | ||
9395 | *Steve Henson* | |
9396 | ||
9397 | * Replace the highly broken and deprecated SPKAC certification method with | |
9398 | the updated NID creation version. This should correctly handle UTF8. | |
9399 | ||
9400 | *Steve Henson* | |
9401 | ||
9402 | * Implement RFC5746. Re-enable renegotiation but require the extension | |
9403 | as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION | |
9404 | turns out to be a bad idea. It has been replaced by | |
9405 | SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with | |
9406 | SSL_CTX_set_options(). This is really not recommended unless you | |
9407 | know what you are doing. | |
9408 | ||
9409 | *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson* | |
9410 | ||
9411 | * Fixes to stateless session resumption handling. Use initial_ctx when | |
9412 | issuing and attempting to decrypt tickets in case it has changed during | |
9413 | servername handling. Use a non-zero length session ID when attempting | |
9414 | stateless session resumption: this makes it possible to determine if | |
9415 | a resumption has occurred immediately after receiving server hello | |
9416 | (several places in OpenSSL subtly assume this) instead of later in | |
9417 | the handshake. | |
9418 | ||
9419 | *Steve Henson* | |
9420 | ||
9421 | * The functions ENGINE_ctrl(), OPENSSL_isservice(), | |
9422 | CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error | |
9423 | fixes for a few places where the return code is not checked | |
9424 | correctly. | |
9425 | ||
9426 | *Julia Lawall <julia@diku.dk>* | |
9427 | ||
9428 | * Add --strict-warnings option to Configure script to include devteam | |
9429 | warnings in other configurations. | |
9430 | ||
9431 | *Steve Henson* | |
9432 | ||
9433 | * Add support for --libdir option and LIBDIR variable in makefiles. This | |
9434 | makes it possible to install openssl libraries in locations which | |
9435 | have names other than "lib", for example "/usr/lib64" which some | |
9436 | systems need. | |
9437 | ||
9438 | *Steve Henson, based on patch from Jeremy Utley* | |
9439 | ||
9440 | * Don't allow the use of leading 0x80 in OIDs. This is a violation of | |
9441 | X690 8.9.12 and can produce some misleading textual output of OIDs. | |
9442 | ||
9443 | *Steve Henson, reported by Dan Kaminsky* | |
9444 | ||
9445 | * Delete MD2 from algorithm tables. This follows the recommendation in | |
9446 | several standards that it is not used in new applications due to | |
9447 | several cryptographic weaknesses. For binary compatibility reasons | |
9448 | the MD2 API is still compiled in by default. | |
9449 | ||
9450 | *Steve Henson* | |
9451 | ||
9452 | * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved | |
9453 | and restored. | |
9454 | ||
9455 | *Steve Henson* | |
9456 | ||
9457 | * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and | |
9458 | OPENSSL_asc2uni conditionally on Netware platforms to avoid a name | |
9459 | clash. | |
9460 | ||
9461 | *Guenter <lists@gknw.net>* | |
9462 | ||
9463 | * Fix the server certificate chain building code to use X509_verify_cert(), | |
9464 | it used to have an ad-hoc builder which was unable to cope with anything | |
9465 | other than a simple chain. | |
9466 | ||
9467 | *David Woodhouse <dwmw2@infradead.org>, Steve Henson* | |
9468 | ||
9469 | * Don't check self signed certificate signatures in X509_verify_cert() | |
9470 | by default (a flag can override this): it just wastes time without | |
9471 | adding any security. As a useful side effect self signed root CAs | |
9472 | with non-FIPS digests are now usable in FIPS mode. | |
9473 | ||
9474 | *Steve Henson* | |
9475 | ||
9476 | * In dtls1_process_out_of_seq_message() the check if the current message | |
9477 | is already buffered was missing. For every new message was memory | |
9478 | allocated, allowing an attacker to perform an denial of service attack | |
9479 | with sending out of seq handshake messages until there is no memory | |
9480 | left. Additionally every future message was buffered, even if the | |
9481 | sequence number made no sense and would be part of another handshake. | |
9482 | So only messages with sequence numbers less than 10 in advance will be | |
9483 | buffered. ([CVE-2009-1378]) | |
9484 | ||
9485 | *Robin Seggelmann, discovered by Daniel Mentz* | |
9486 | ||
9487 | * Records are buffered if they arrive with a future epoch to be | |
9488 | processed after finishing the corresponding handshake. There is | |
9489 | currently no limitation to this buffer allowing an attacker to perform | |
9490 | a DOS attack with sending records with future epochs until there is no | |
9491 | memory left. This patch adds the pqueue_size() function to determine | |
9492 | the size of a buffer and limits the record buffer to 100 entries. | |
9493 | ([CVE-2009-1377]) | |
9494 | ||
9495 | *Robin Seggelmann, discovered by Daniel Mentz* | |
9496 | ||
9497 | * Keep a copy of frag->msg_header.frag_len so it can be used after the | |
9498 | parent structure is freed. ([CVE-2009-1379]) | |
9499 | ||
9500 | *Daniel Mentz* | |
9501 | ||
9502 | * Handle non-blocking I/O properly in SSL_shutdown() call. | |
9503 | ||
9504 | *Darryl Miles <darryl-mailinglists@netbauds.net>* | |
9505 | ||
9506 | * Add `2.5.4.*` OIDs | |
9507 | ||
9508 | *Ilya O. <vrghost@gmail.com>* | |
9509 | ||
9510 | ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009] | |
9511 | ||
9512 | * Disable renegotiation completely - this fixes a severe security | |
9513 | problem ([CVE-2009-3555]) at the cost of breaking all | |
9514 | renegotiation. Renegotiation can be re-enabled by setting | |
9515 | SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at | |
9516 | run-time. This is really not recommended unless you know what | |
9517 | you're doing. | |
9518 | ||
9519 | *Ben Laurie* | |
9520 | ||
9521 | ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009] | |
9522 | ||
9523 | * Don't set val to NULL when freeing up structures, it is freed up by | |
9524 | underlying code. If `sizeof(void *) > sizeof(long)` this can result in | |
9525 | zeroing past the valid field. ([CVE-2009-0789]) | |
9526 | ||
9527 | *Paolo Ganci <Paolo.Ganci@AdNovum.CH>* | |
9528 | ||
9529 | * Fix bug where return value of CMS_SignerInfo_verify_content() was not | |
9530 | checked correctly. This would allow some invalid signed attributes to | |
9531 | appear to verify correctly. ([CVE-2009-0591]) | |
9532 | ||
9533 | *Ivan Nestlerode <inestlerode@us.ibm.com>* | |
9534 | ||
9535 | * Reject UniversalString and BMPString types with invalid lengths. This | |
9536 | prevents a crash in ASN1_STRING_print_ex() which assumes the strings have | |
9537 | a legal length. ([CVE-2009-0590]) | |
9538 | ||
9539 | *Steve Henson* | |
9540 | ||
9541 | * Set S/MIME signing as the default purpose rather than setting it | |
9542 | unconditionally. This allows applications to override it at the store | |
9543 | level. | |
9544 | ||
9545 | *Steve Henson* | |
9546 | ||
9547 | * Permit restricted recursion of ASN1 strings. This is needed in practice | |
9548 | to handle some structures. | |
9549 | ||
9550 | *Steve Henson* | |
9551 | ||
9552 | * Improve efficiency of mem_gets: don't search whole buffer each time | |
9553 | for a '\n' | |
9554 | ||
9555 | *Jeremy Shapiro <jnshapir@us.ibm.com>* | |
9556 | ||
9557 | * New -hex option for openssl rand. | |
9558 | ||
9559 | *Matthieu Herrb* | |
9560 | ||
9561 | * Print out UTF8String and NumericString when parsing ASN1. | |
9562 | ||
9563 | *Steve Henson* | |
9564 | ||
9565 | * Support NumericString type for name components. | |
9566 | ||
9567 | *Steve Henson* | |
9568 | ||
9569 | * Allow CC in the environment to override the automatically chosen | |
9570 | compiler. Note that nothing is done to ensure flags work with the | |
9571 | chosen compiler. | |
9572 | ||
9573 | *Ben Laurie* | |
9574 | ||
9575 | ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009] | |
9576 | ||
9577 | * Properly check EVP_VerifyFinal() and similar return values | |
9578 | ([CVE-2008-5077]). | |
9579 | ||
9580 | *Ben Laurie, Bodo Moeller, Google Security Team* | |
9581 | ||
9582 | * Enable TLS extensions by default. | |
9583 | ||
9584 | *Ben Laurie* | |
9585 | ||
9586 | * Allow the CHIL engine to be loaded, whether the application is | |
9587 | multithreaded or not. (This does not release the developer from the | |
9588 | obligation to set up the dynamic locking callbacks.) | |
9589 | ||
9590 | *Sander Temme <sander@temme.net>* | |
9591 | ||
9592 | * Use correct exit code if there is an error in dgst command. | |
9593 | ||
9594 | *Steve Henson; problem pointed out by Roland Dirlewanger* | |
9595 | ||
9596 | * Tweak Configure so that you need to say "experimental-jpake" to enable | |
9597 | JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. | |
9598 | ||
9599 | *Bodo Moeller* | |
9600 | ||
9601 | * Add experimental JPAKE support, including demo authentication in | |
9602 | s_client and s_server. | |
9603 | ||
9604 | *Ben Laurie* | |
9605 | ||
9606 | * Set the comparison function in v3_addr_canonize(). | |
9607 | ||
9608 | *Rob Austein <sra@hactrn.net>* | |
9609 | ||
9610 | * Add support for XMPP STARTTLS in s_client. | |
9611 | ||
9612 | *Philip Paeps <philip@freebsd.org>* | |
9613 | ||
9614 | * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior | |
9615 | to ensure that even with this option, only ciphersuites in the | |
9616 | server's preference list will be accepted. (Note that the option | |
9617 | applies only when resuming a session, so the earlier behavior was | |
9618 | just about the algorithm choice for symmetric cryptography.) | |
9619 | ||
9620 | *Bodo Moeller* | |
9621 | ||
9622 | ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008] | |
9623 | ||
9624 | * Fix NULL pointer dereference if a DTLS server received | |
9625 | ChangeCipherSpec as first record ([CVE-2009-1386]). | |
9626 | ||
9627 | *PR #1679* | |
9628 | ||
9629 | * Fix a state transition in s3_srvr.c and d1_srvr.c | |
9630 | (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`). | |
9631 | ||
9632 | *Nagendra Modadugu* | |
9633 | ||
9634 | * The fix in 0.9.8c that supposedly got rid of unsafe | |
9635 | double-checked locking was incomplete for RSA blinding, | |
9636 | addressing just one layer of what turns out to have been | |
9637 | doubly unsafe triple-checked locking. | |
9638 | ||
9639 | So now fix this for real by retiring the MONT_HELPER macro | |
9640 | in crypto/rsa/rsa_eay.c. | |
9641 | ||
9642 | *Bodo Moeller; problem pointed out by Marius Schilder* | |
9643 | ||
9644 | * Various precautionary measures: | |
9645 | ||
9646 | - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). | |
9647 | ||
9648 | - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). | |
9649 | (NB: This would require knowledge of the secret session ticket key | |
9650 | to exploit, in which case you'd be SOL either way.) | |
9651 | ||
9652 | - Change bn_nist.c so that it will properly handle input BIGNUMs | |
9653 | outside the expected range. | |
9654 | ||
9655 | - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG | |
9656 | builds. | |
9657 | ||
9658 | *Neel Mehta, Bodo Moeller* | |
9659 | ||
9660 | * Allow engines to be "soft loaded" - i.e. optionally don't die if | |
9661 | the load fails. Useful for distros. | |
9662 | ||
9663 | *Ben Laurie and the FreeBSD team* | |
9664 | ||
9665 | * Add support for Local Machine Keyset attribute in PKCS#12 files. | |
9666 | ||
9667 | *Steve Henson* | |
9668 | ||
9669 | * Fix BN_GF2m_mod_arr() top-bit cleanup code. | |
9670 | ||
9671 | *Huang Ying* | |
9672 | ||
9673 | * Expand ENGINE to support engine supplied SSL client certificate functions. | |
9674 | ||
9675 | This work was sponsored by Logica. | |
9676 | ||
9677 | *Steve Henson* | |
9678 | ||
9679 | * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows | |
9680 | keystores. Support for SSL/TLS client authentication too. | |
9681 | Not compiled unless enable-capieng specified to Configure. | |
9682 | ||
9683 | This work was sponsored by Logica. | |
9684 | ||
9685 | *Steve Henson* | |
9686 | ||
9687 | * Fix bug in X509_ATTRIBUTE creation: don't set attribute using | |
9688 | ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain | |
9689 | attribute creation routines such as certificate requests and PKCS#12 | |
9690 | files. | |
9691 | ||
9692 | *Steve Henson* | |
9693 | ||
9694 | ### Changes between 0.9.8g and 0.9.8h [28 May 2008] | |
9695 | ||
9696 | * Fix flaw if 'Server Key exchange message' is omitted from a TLS | |
9697 | handshake which could lead to a client crash as found using the | |
9698 | Codenomicon TLS test suite ([CVE-2008-1672]) | |
9699 | ||
9700 | *Steve Henson, Mark Cox* | |
9701 | ||
9702 | * Fix double free in TLS server name extensions which could lead to | |
9703 | a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891]) | |
9704 | ||
9705 | *Joe Orton* | |
9706 | ||
9707 | * Clear error queue in SSL_CTX_use_certificate_chain_file() | |
9708 | ||
9709 | Clear the error queue to ensure that error entries left from | |
9710 | older function calls do not interfere with the correct operation. | |
9711 | ||
9712 | *Lutz Jaenicke, Erik de Castro Lopo* | |
9713 | ||
9714 | * Remove root CA certificates of commercial CAs: | |
9715 | ||
9716 | The OpenSSL project does not recommend any specific CA and does not | |
9717 | have any policy with respect to including or excluding any CA. | |
9718 | Therefore it does not make any sense to ship an arbitrary selection | |
9719 | of root CA certificates with the OpenSSL software. | |
9720 | ||
9721 | *Lutz Jaenicke* | |
9722 | ||
9723 | * RSA OAEP patches to fix two separate invalid memory reads. | |
9724 | The first one involves inputs when 'lzero' is greater than | |
9725 | 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes | |
9726 | before the beginning of from). The second one involves inputs where | |
9727 | the 'db' section contains nothing but zeroes (there is a one-byte | |
9728 | invalid read after the end of 'db'). | |
9729 | ||
9730 | *Ivan Nestlerode <inestlerode@us.ibm.com>* | |
9731 | ||
9732 | * Partial backport from 0.9.9-dev: | |
9733 | ||
9734 | Introduce bn_mul_mont (dedicated Montgomery multiplication | |
9735 | procedure) as a candidate for BIGNUM assembler implementation. | |
9736 | While 0.9.9-dev uses assembler for various architectures, only | |
9737 | x86_64 is available by default here in the 0.9.8 branch, and | |
9738 | 32-bit x86 is available through a compile-time setting. | |
9739 | ||
9740 | To try the 32-bit x86 assembler implementation, use Configure | |
9741 | option "enable-montasm" (which exists only for this backport). | |
9742 | ||
9743 | As "enable-montasm" for 32-bit x86 disclaims code stability | |
9744 | anyway, in this constellation we activate additional code | |
9745 | backported from 0.9.9-dev for further performance improvements, | |
9746 | namely BN_from_montgomery_word. (To enable this otherwise, | |
9747 | e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.) | |
9748 | ||
9749 | *Andy Polyakov (backport partially by Bodo Moeller)* | |
9750 | ||
9751 | * Add TLS session ticket callback. This allows an application to set | |
9752 | TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed | |
9753 | values. This is useful for key rollover for example where several key | |
9754 | sets may exist with different names. | |
9755 | ||
9756 | *Steve Henson* | |
9757 | ||
9758 | * Reverse ENGINE-internal logic for caching default ENGINE handles. | |
9759 | This was broken until now in 0.9.8 releases, such that the only way | |
9760 | a registered ENGINE could be used (assuming it initialises | |
9761 | successfully on the host) was to explicitly set it as the default | |
9762 | for the relevant algorithms. This is in contradiction with 0.9.7 | |
9763 | behaviour and the documentation. With this fix, when an ENGINE is | |
9764 | registered into a given algorithm's table of implementations, the | |
9765 | 'uptodate' flag is reset so that auto-discovery will be used next | |
9766 | time a new context for that algorithm attempts to select an | |
9767 | implementation. | |
9768 | ||
9769 | *Ian Lister (tweaked by Geoff Thorpe)* | |
9770 | ||
9771 | * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 | |
9772 | implementation in the following ways: | |
9773 | ||
9774 | Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be | |
9775 | hard coded. | |
9776 | ||
9777 | Lack of BER streaming support means one pass streaming processing is | |
9778 | only supported if data is detached: setting the streaming flag is | |
9779 | ignored for embedded content. | |
9780 | ||
9781 | CMS support is disabled by default and must be explicitly enabled | |
9782 | with the enable-cms configuration option. | |
9783 | ||
9784 | *Steve Henson* | |
9785 | ||
9786 | * Update the GMP engine glue to do direct copies between BIGNUM and | |
9787 | mpz_t when openssl and GMP use the same limb size. Otherwise the | |
9788 | existing "conversion via a text string export" trick is still used. | |
9789 | ||
9790 | *Paul Sheer <paulsheer@gmail.com>* | |
9791 | ||
9792 | * Zlib compression BIO. This is a filter BIO which compressed and | |
9793 | uncompresses any data passed through it. | |
9794 | ||
9795 | *Steve Henson* | |
9796 | ||
9797 | * Add AES_wrap_key() and AES_unwrap_key() functions to implement | |
9798 | RFC3394 compatible AES key wrapping. | |
9799 | ||
9800 | *Steve Henson* | |
9801 | ||
9802 | * Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): | |
9803 | sets string data without copying. X509_ALGOR_set0() and | |
9804 | X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) | |
9805 | data. Attribute function X509at_get0_data_by_OBJ(): retrieves data | |
9806 | from an X509_ATTRIBUTE structure optionally checking it occurs only | |
9807 | once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied | |
9808 | data. | |
9809 | ||
9810 | *Steve Henson* | |
9811 | ||
9812 | * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() | |
9813 | to get the expected BN_FLG_CONSTTIME behavior. | |
9814 | ||
9815 | *Bodo Moeller (Google)* | |
9816 | ||
9817 | * Netware support: | |
9818 | ||
9819 | - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets | |
9820 | - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) | |
9821 | - added some more tests to do_tests.pl | |
9822 | - fixed RunningProcess usage so that it works with newer LIBC NDKs too | |
9823 | - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency | |
9824 | - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, | |
9825 | netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc | |
9826 | - various changes to netware.pl to enable gcc-cross builds on Win32 | |
9827 | platform | |
9828 | - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) | |
9829 | - various changes to fix missing prototype warnings | |
9830 | - fixed x86nasm.pl to create correct asm files for NASM COFF output | |
9831 | - added AES, WHIRLPOOL and CPUID assembler code to build files | |
9832 | - added missing AES assembler make rules to mk1mf.pl | |
9833 | - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply | |
9834 | ||
9835 | *Guenter Knauf <eflash@gmx.net>* | |
9836 | ||
9837 | * Implement certificate status request TLS extension defined in RFC3546. | |
9838 | A client can set the appropriate parameters and receive the encoded | |
9839 | OCSP response via a callback. A server can query the supplied parameters | |
9840 | and set the encoded OCSP response in the callback. Add simplified examples | |
9841 | to s_client and s_server. | |
9842 | ||
9843 | *Steve Henson* | |
9844 | ||
9845 | ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007] | |
9846 | ||
9847 | * Fix various bugs: | |
9848 | + Binary incompatibility of ssl_ctx_st structure | |
9849 | + DTLS interoperation with non-compliant servers | |
9850 | + Don't call get_session_cb() without proposed session | |
9851 | + Fix ia64 assembler code | |
9852 | ||
9853 | *Andy Polyakov, Steve Henson* | |
9854 | ||
9855 | ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007] | |
9856 | ||
9857 | * DTLS Handshake overhaul. There were longstanding issues with | |
9858 | OpenSSL DTLS implementation, which were making it impossible for | |
9859 | RFC 4347 compliant client to communicate with OpenSSL server. | |
9860 | Unfortunately just fixing these incompatibilities would "cut off" | |
9861 | pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e | |
9862 | server keeps tolerating non RFC compliant syntax. The opposite is | |
9863 | not true, 0.9.8f client can not communicate with earlier server. | |
9864 | This update even addresses CVE-2007-4995. | |
9865 | ||
9866 | *Andy Polyakov* | |
9867 | ||
9868 | * Changes to avoid need for function casts in OpenSSL: some compilers | |
9869 | (gcc 4.2 and later) reject their use. | |
9870 | *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, | |
9871 | Steve Henson* | |
9872 | ||
9873 | * Add RFC4507 support to OpenSSL. This includes the corrections in | |
9874 | RFC4507bis. The encrypted ticket format is an encrypted encoded | |
9875 | SSL_SESSION structure, that way new session features are automatically | |
9876 | supported. | |
9877 | ||
9878 | If a client application caches session in an SSL_SESSION structure | |
9879 | support is transparent because tickets are now stored in the encoded | |
9880 | SSL_SESSION. | |
9881 | ||
9882 | The SSL_CTX structure automatically generates keys for ticket | |
9883 | protection in servers so again support should be possible | |
9884 | with no application modification. | |
9885 | ||
9886 | If a client or server wishes to disable RFC4507 support then the option | |
9887 | SSL_OP_NO_TICKET can be set. | |
9888 | ||
9889 | Add a TLS extension debugging callback to allow the contents of any client | |
9890 | or server extensions to be examined. | |
9891 | ||
9892 | This work was sponsored by Google. | |
9893 | ||
9894 | *Steve Henson* | |
9895 | ||
9896 | * Add initial support for TLS extensions, specifically for the server_name | |
9897 | extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now | |
9898 | have new members for a host name. The SSL data structure has an | |
9899 | additional member `SSL_CTX *initial_ctx` so that new sessions can be | |
9900 | stored in that context to allow for session resumption, even after the | |
9901 | SSL has been switched to a new SSL_CTX in reaction to a client's | |
9902 | server_name extension. | |
9903 | ||
9904 | New functions (subject to change): | |
9905 | ||
9906 | SSL_get_servername() | |
9907 | SSL_get_servername_type() | |
9908 | SSL_set_SSL_CTX() | |
9909 | ||
9910 | New CTRL codes and macros (subject to change): | |
9911 | ||
9912 | SSL_CTRL_SET_TLSEXT_SERVERNAME_CB | |
9913 | - SSL_CTX_set_tlsext_servername_callback() | |
9914 | SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG | |
9915 | - SSL_CTX_set_tlsext_servername_arg() | |
9916 | SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() | |
9917 | ||
9918 | openssl s_client has a new '-servername ...' option. | |
9919 | ||
9920 | openssl s_server has new options '-servername_host ...', '-cert2 ...', | |
9921 | '-key2 ...', '-servername_fatal' (subject to change). This allows | |
9922 | testing the HostName extension for a specific single host name ('-cert' | |
9923 | and '-key' remain fallbacks for handshakes without HostName | |
9924 | negotiation). If the unrecognized_name alert has to be sent, this by | |
9925 | default is a warning; it becomes fatal with the '-servername_fatal' | |
9926 | option. | |
9927 | ||
9928 | *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson* | |
9929 | ||
9930 | * Add AES and SSE2 assembly language support to VC++ build. | |
9931 | ||
9932 | *Steve Henson* | |
9933 | ||
9934 | * Mitigate attack on final subtraction in Montgomery reduction. | |
9935 | ||
9936 | *Andy Polyakov* | |
9937 | ||
9938 | * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 | |
9939 | (which previously caused an internal error). | |
9940 | ||
9941 | *Bodo Moeller* | |
9942 | ||
9943 | * Squeeze another 10% out of IGE mode when in != out. | |
9944 | ||
9945 | *Ben Laurie* | |
9946 | ||
9947 | * AES IGE mode speedup. | |
9948 | ||
9949 | *Dean Gaudet (Google)* | |
9950 | ||
9951 | * Add the Korean symmetric 128-bit cipher SEED (see | |
9952 | <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and | |
9953 | add SEED ciphersuites from RFC 4162: | |
9954 | ||
9955 | TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" | |
9956 | TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" | |
9957 | TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" | |
9958 | TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" | |
9959 | ||
9960 | To minimize changes between patchlevels in the OpenSSL 0.9.8 | |
9961 | series, SEED remains excluded from compilation unless OpenSSL | |
9962 | is configured with 'enable-seed'. | |
9963 | ||
9964 | *KISA, Bodo Moeller* | |
9965 | ||
9966 | * Mitigate branch prediction attacks, which can be practical if a | |
9967 | single processor is shared, allowing a spy process to extract | |
9968 | information. For detailed background information, see | |
9969 | <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron, | |
9970 | J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL | |
9971 | and Necessary Software Countermeasures"). The core of the change | |
9972 | are new versions BN_div_no_branch() and | |
9973 | BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), | |
9974 | respectively, which are slower, but avoid the security-relevant | |
9975 | conditional branches. These are automatically called by BN_div() | |
9976 | and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one | |
9977 | of the input BIGNUMs. Also, BN_is_bit_set() has been changed to | |
9978 | remove a conditional branch. | |
9979 | ||
9980 | BN_FLG_CONSTTIME is the new name for the previous | |
9981 | BN_FLG_EXP_CONSTTIME flag, since it now affects more than just | |
9982 | modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag | |
9983 | in the exponent causes BN_mod_exp_mont() to use the alternative | |
9984 | implementation in BN_mod_exp_mont_consttime().) The old name | |
9985 | remains as a deprecated alias. | |
9986 | ||
9987 | Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general | |
9988 | RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses | |
9989 | constant-time implementations for more than just exponentiation. | |
9990 | Here too the old name is kept as a deprecated alias. | |
9991 | ||
9992 | BN_BLINDING_new() will now use BN_dup() for the modulus so that | |
9993 | the BN_BLINDING structure gets an independent copy of the | |
9994 | modulus. This means that the previous `BIGNUM *m` argument to | |
9995 | BN_BLINDING_new() and to BN_BLINDING_create_param() now | |
9996 | essentially becomes `const BIGNUM *m`, although we can't actually | |
9997 | change this in the header file before 0.9.9. It allows | |
9998 | RSA_setup_blinding() to use BN_with_flags() on the modulus to | |
9999 | enable BN_FLG_CONSTTIME. | |
10000 | ||
10001 | *Matthew D Wood (Intel Corp)* | |
10002 | ||
10003 | * In the SSL/TLS server implementation, be strict about session ID | |
10004 | context matching (which matters if an application uses a single | |
10005 | external cache for different purposes). Previously, | |
10006 | out-of-context reuse was forbidden only if SSL_VERIFY_PEER was | |
10007 | set. This did ensure strict client verification, but meant that, | |
10008 | with applications using a single external cache for quite | |
10009 | different requirements, clients could circumvent ciphersuite | |
10010 | restrictions for a given session ID context by starting a session | |
10011 | in a different context. | |
10012 | ||
10013 | *Bodo Moeller* | |
10014 | ||
10015 | * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that | |
10016 | a ciphersuite string such as "DEFAULT:RSA" cannot enable | |
10017 | authentication-only ciphersuites. | |
10018 | ||
10019 | *Bodo Moeller* | |
10020 | ||
10021 | * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was | |
10022 | not complete and could lead to a possible single byte overflow | |
10023 | ([CVE-2007-5135]) [Ben Laurie] | |
10024 | ||
10025 | ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007] | |
10026 | ||
10027 | * Since AES128 and AES256 (and similarly Camellia128 and | |
10028 | Camellia256) share a single mask bit in the logic of | |
10029 | ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a | |
10030 | kludge to work properly if AES128 is available and AES256 isn't | |
10031 | (or if Camellia128 is available and Camellia256 isn't). | |
10032 | ||
10033 | *Victor Duchovni* | |
10034 | ||
10035 | * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c | |
10036 | (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): | |
10037 | When a point or a seed is encoded in a BIT STRING, we need to | |
10038 | prevent the removal of trailing zero bits to get the proper DER | |
10039 | encoding. (By default, crypto/asn1/a_bitstr.c assumes the case | |
10040 | of a NamedBitList, for which trailing 0 bits need to be removed.) | |
10041 | ||
10042 | *Bodo Moeller* | |
10043 | ||
10044 | * Have SSL/TLS server implementation tolerate "mismatched" record | |
10045 | protocol version while receiving ClientHello even if the | |
10046 | ClientHello is fragmented. (The server can't insist on the | |
10047 | particular protocol version it has chosen before the ServerHello | |
10048 | message has informed the client about his choice.) | |
10049 | ||
10050 | *Bodo Moeller* | |
10051 | ||
10052 | * Add RFC 3779 support. | |
10053 | ||
10054 | *Rob Austein for ARIN, Ben Laurie* | |
10055 | ||
10056 | * Load error codes if they are not already present instead of using a | |
10057 | static variable. This allows them to be cleanly unloaded and reloaded. | |
10058 | Improve header file function name parsing. | |
10059 | ||
10060 | *Steve Henson* | |
10061 | ||
10062 | * extend SMTP and IMAP protocol emulation in s_client to use EHLO | |
10063 | or CAPABILITY handshake as required by RFCs. | |
10064 | ||
10065 | *Goetz Babin-Ebell* | |
10066 | ||
10067 | ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006] | |
10068 | ||
10069 | * Introduce limits to prevent malicious keys being able to | |
10070 | cause a denial of service. ([CVE-2006-2940]) | |
10071 | ||
10072 | *Steve Henson, Bodo Moeller* | |
10073 | ||
10074 | * Fix ASN.1 parsing of certain invalid structures that can result | |
10075 | in a denial of service. ([CVE-2006-2937]) [Steve Henson] | |
10076 | ||
10077 | * Fix buffer overflow in SSL_get_shared_ciphers() function. | |
10078 | ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team] | |
10079 | ||
10080 | * Fix SSL client code which could crash if connecting to a | |
10081 | malicious SSLv2 server. ([CVE-2006-4343]) | |
10082 | ||
10083 | *Tavis Ormandy and Will Drewry, Google Security Team* | |
10084 | ||
10085 | * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites | |
10086 | match only those. Before that, "AES256-SHA" would be interpreted | |
10087 | as a pattern and match "AES128-SHA" too (since AES128-SHA got | |
10088 | the same strength classification in 0.9.7h) as we currently only | |
10089 | have a single AES bit in the ciphersuite description bitmap. | |
10090 | That change, however, also applied to ciphersuite strings such as | |
10091 | "RC4-MD5" that intentionally matched multiple ciphersuites -- | |
10092 | namely, SSL 2.0 ciphersuites in addition to the more common ones | |
10093 | from SSL 3.0/TLS 1.0. | |
10094 | ||
10095 | So we change the selection algorithm again: Naming an explicit | |
10096 | ciphersuite selects this one ciphersuite, and any other similar | |
10097 | ciphersuite (same bitmap) from *other* protocol versions. | |
10098 | Thus, "RC4-MD5" again will properly select both the SSL 2.0 | |
10099 | ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. | |
10100 | ||
10101 | Since SSL 2.0 does not have any ciphersuites for which the | |
10102 | 128/256 bit distinction would be relevant, this works for now. | |
10103 | The proper fix will be to use different bits for AES128 and | |
10104 | AES256, which would have avoided the problems from the beginning; | |
10105 | however, bits are scarce, so we can only do this in a new release | |
10106 | (not just a patchlevel) when we can change the SSL_CIPHER | |
10107 | definition to split the single 'unsigned long mask' bitmap into | |
10108 | multiple values to extend the available space. | |
10109 | ||
10110 | *Bodo Moeller* | |
10111 | ||
10112 | ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006] | |
10113 | ||
10114 | * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher | |
10115 | ([CVE-2006-4339]) [Ben Laurie and Google Security Team] | |
10116 | ||
10117 | * Add AES IGE and biIGE modes. | |
10118 | ||
10119 | *Ben Laurie* | |
10120 | ||
10121 | * Change the Unix randomness entropy gathering to use poll() when | |
10122 | possible instead of select(), since the latter has some | |
10123 | undesirable limitations. | |
10124 | ||
10125 | *Darryl Miles via Richard Levitte and Bodo Moeller* | |
10126 | ||
10127 | * Disable "ECCdraft" ciphersuites more thoroughly. Now special | |
10128 | treatment in ssl/ssl_ciph.s makes sure that these ciphersuites | |
10129 | cannot be implicitly activated as part of, e.g., the "AES" alias. | |
10130 | However, please upgrade to OpenSSL 0.9.9[-dev] for | |
10131 | non-experimental use of the ECC ciphersuites to get TLS extension | |
10132 | support, which is required for curve and point format negotiation | |
10133 | to avoid potential handshake problems. | |
10134 | ||
10135 | *Bodo Moeller* | |
10136 | ||
10137 | * Disable rogue ciphersuites: | |
10138 | ||
10139 | - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") | |
10140 | - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") | |
10141 | - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") | |
10142 | ||
10143 | The latter two were purportedly from | |
10144 | draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really | |
10145 | appear there. | |
10146 | ||
10147 | Also deactivate the remaining ciphersuites from | |
10148 | draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as | |
10149 | unofficial, and the ID has long expired. | |
10150 | ||
10151 | *Bodo Moeller* | |
10152 | ||
10153 | * Fix RSA blinding Heisenbug (problems sometimes occurred on | |
10154 | dual-core machines) and other potential thread-safety issues. | |
10155 | ||
10156 | *Bodo Moeller* | |
10157 | ||
10158 | * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key | |
10159 | versions), which is now available for royalty-free use | |
10160 | (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>). | |
10161 | Also, add Camellia TLS ciphersuites from RFC 4132. | |
10162 | ||
10163 | To minimize changes between patchlevels in the OpenSSL 0.9.8 | |
10164 | series, Camellia remains excluded from compilation unless OpenSSL | |
10165 | is configured with 'enable-camellia'. | |
10166 | ||
10167 | *NTT* | |
10168 | ||
10169 | * Disable the padding bug check when compression is in use. The padding | |
10170 | bug check assumes the first packet is of even length, this is not | |
10171 | necessarily true if compression is enabled and can result in false | |
10172 | positives causing handshake failure. The actual bug test is ancient | |
10173 | code so it is hoped that implementations will either have fixed it by | |
10174 | now or any which still have the bug do not support compression. | |
10175 | ||
10176 | *Steve Henson* | |
10177 | ||
10178 | ### Changes between 0.9.8a and 0.9.8b [04 May 2006] | |
10179 | ||
10180 | * When applying a cipher rule check to see if string match is an explicit | |
10181 | cipher suite and only match that one cipher suite if it is. | |
10182 | ||
10183 | *Steve Henson* | |
10184 | ||
10185 | * Link in manifests for VC++ if needed. | |
10186 | ||
10187 | *Austin Ziegler <halostatue@gmail.com>* | |
10188 | ||
10189 | * Update support for ECC-based TLS ciphersuites according to | |
10190 | draft-ietf-tls-ecc-12.txt with proposed changes (but without | |
10191 | TLS extensions, which are supported starting with the 0.9.9 | |
10192 | branch, not in the OpenSSL 0.9.8 branch). | |
10193 | ||
10194 | *Douglas Stebila* | |
10195 | ||
10196 | * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support | |
10197 | opaque EVP_CIPHER_CTX handling. | |
10198 | ||
10199 | *Steve Henson* | |
10200 | ||
10201 | * Fixes and enhancements to zlib compression code. We now only use | |
10202 | "zlib1.dll" and use the default `__cdecl` calling convention on Win32 | |
10203 | to conform with the standards mentioned here: | |
10204 | <http://www.zlib.net/DLL_FAQ.txt> | |
10205 | Static zlib linking now works on Windows and the new --with-zlib-include | |
10206 | --with-zlib-lib options to Configure can be used to supply the location | |
10207 | of the headers and library. Gracefully handle case where zlib library | |
10208 | can't be loaded. | |
10209 | ||
10210 | *Steve Henson* | |
10211 | ||
10212 | * Several fixes and enhancements to the OID generation code. The old code | |
10213 | sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't | |
10214 | handle numbers larger than ULONG_MAX, truncated printing and had a | |
10215 | non standard OBJ_obj2txt() behaviour. | |
10216 | ||
10217 | *Steve Henson* | |
10218 | ||
10219 | * Add support for building of engines under engine/ as shared libraries | |
10220 | under VC++ build system. | |
10221 | ||
10222 | *Steve Henson* | |
10223 | ||
10224 | * Corrected the numerous bugs in the Win32 path splitter in DSO. | |
10225 | Hopefully, we will not see any false combination of paths any more. | |
10226 | ||
10227 | *Richard Levitte* | |
10228 | ||
10229 | ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005] | |
10230 | ||
10231 | * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING | |
10232 | (part of SSL_OP_ALL). This option used to disable the | |
10233 | countermeasure against man-in-the-middle protocol-version | |
10234 | rollback in the SSL 2.0 server implementation, which is a bad | |
10235 | idea. ([CVE-2005-2969]) | |
10236 | ||
10237 | *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center | |
10238 | for Information Security, National Institute of Advanced Industrial | |
10239 | Science and Technology [AIST], Japan)* | |
10240 | ||
10241 | * Add two function to clear and return the verify parameter flags. | |
10242 | ||
10243 | *Steve Henson* | |
10244 | ||
10245 | * Keep cipherlists sorted in the source instead of sorting them at | |
10246 | runtime, thus removing the need for a lock. | |
10247 | ||
10248 | *Nils Larsch* | |
10249 | ||
10250 | * Avoid some small subgroup attacks in Diffie-Hellman. | |
10251 | ||
10252 | *Nick Mathewson and Ben Laurie* | |
10253 | ||
10254 | * Add functions for well-known primes. | |
10255 | ||
10256 | *Nick Mathewson* | |
10257 | ||
10258 | * Extended Windows CE support. | |
10259 | ||
10260 | *Satoshi Nakamura and Andy Polyakov* | |
10261 | ||
10262 | * Initialize SSL_METHOD structures at compile time instead of during | |
10263 | runtime, thus removing the need for a lock. | |
10264 | ||
10265 | *Steve Henson* | |
10266 | ||
10267 | * Make PKCS7_decrypt() work even if no certificate is supplied by | |
10268 | attempting to decrypt each encrypted key in turn. Add support to | |
10269 | smime utility. | |
10270 | ||
10271 | *Steve Henson* | |
10272 | ||
10273 | ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005] | |
10274 | ||
10275 | [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after | |
10276 | OpenSSL 0.9.8.] | |
10277 | ||
10278 | * Add libcrypto.pc and libssl.pc for those who feel they need them. | |
10279 | ||
10280 | *Richard Levitte* | |
10281 | ||
10282 | * Change CA.sh and CA.pl so they don't bundle the CSR and the private | |
10283 | key into the same file any more. | |
10284 | ||
10285 | *Richard Levitte* | |
10286 | ||
10287 | * Add initial support for Win64, both IA64 and AMD64/x64 flavors. | |
10288 | ||
10289 | *Andy Polyakov* | |
10290 | ||
10291 | * Add -utf8 command line and config file option to 'ca'. | |
10292 | ||
10293 | *Stefan <stf@udoma.org* | |
10294 | ||
10295 | * Removed the macro des_crypt(), as it seems to conflict with some | |
10296 | libraries. Use DES_crypt(). | |
10297 | ||
10298 | *Richard Levitte* | |
10299 | ||
10300 | * Correct naming of the 'chil' and '4758cca' ENGINEs. This | |
10301 | involves renaming the source and generated shared-libs for | |
10302 | both. The engines will accept the corrected or legacy ids | |
10303 | ('ncipher' and '4758_cca' respectively) when binding. NB, | |
10304 | this only applies when building 'shared'. | |
10305 | ||
10306 | *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe* | |
10307 | ||
10308 | * Add attribute functions to EVP_PKEY structure. Modify | |
10309 | PKCS12_create() to recognize a CSP name attribute and | |
10310 | use it. Make -CSP option work again in pkcs12 utility. | |
10311 | ||
10312 | *Steve Henson* | |
10313 | ||
10314 | * Add new functionality to the bn blinding code: | |
10315 | - automatic re-creation of the BN_BLINDING parameters after | |
10316 | a fixed number of uses (currently 32) | |
10317 | - add new function for parameter creation | |
10318 | - introduce flags to control the update behaviour of the | |
10319 | BN_BLINDING parameters | |
10320 | - hide BN_BLINDING structure | |
10321 | Add a second BN_BLINDING slot to the RSA structure to improve | |
10322 | performance when a single RSA object is shared among several | |
10323 | threads. | |
10324 | ||
10325 | *Nils Larsch* | |
10326 | ||
10327 | * Add support for DTLS. | |
10328 | ||
10329 | *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie* | |
10330 | ||
10331 | * Add support for DER encoded private keys (SSL_FILETYPE_ASN1) | |
10332 | to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() | |
10333 | ||
10334 | *Walter Goulet* | |
10335 | ||
10336 | * Remove buggy and incomplete DH cert support from | |
10337 | ssl/ssl_rsa.c and ssl/s3_both.c | |
10338 | ||
10339 | *Nils Larsch* | |
10340 | ||
10341 | * Use SHA-1 instead of MD5 as the default digest algorithm for | |
10342 | the `apps/openssl` commands. | |
10343 | ||
10344 | *Nils Larsch* | |
10345 | ||
10346 | * Compile clean with "-Wall -Wmissing-prototypes | |
10347 | -Wstrict-prototypes -Wmissing-declarations -Werror". Currently | |
10348 | DEBUG_SAFESTACK must also be set. | |
10349 | ||
10350 | *Ben Laurie* | |
10351 | ||
10352 | * Change ./Configure so that certain algorithms can be disabled by default. | |
10353 | The new counterpiece to "no-xxx" is "enable-xxx". | |
10354 | ||
10355 | The patented RC5 and MDC2 algorithms will now be disabled unless | |
10356 | "enable-rc5" and "enable-mdc2", respectively, are specified. | |
10357 | ||
10358 | (IDEA remains enabled despite being patented. This is because IDEA | |
10359 | is frequently required for interoperability, and there is no license | |
10360 | fee for non-commercial use. As before, "no-idea" can be used to | |
10361 | avoid this algorithm.) | |
10362 | ||
10363 | *Bodo Moeller* | |
10364 | ||
10365 | * Add processing of proxy certificates (see RFC 3820). This work was | |
10366 | sponsored by KTH (The Royal Institute of Technology in Stockholm) and | |
10367 | EGEE (Enabling Grids for E-science in Europe). | |
10368 | ||
10369 | *Richard Levitte* | |
10370 | ||
10371 | * RC4 performance overhaul on modern architectures/implementations, such | |
10372 | as Intel P4, IA-64 and AMD64. | |
10373 | ||
10374 | *Andy Polyakov* | |
10375 | ||
10376 | * New utility extract-section.pl. This can be used specify an alternative | |
10377 | section number in a pod file instead of having to treat each file as | |
10378 | a separate case in Makefile. This can be done by adding two lines to the | |
10379 | pod file: | |
10380 | ||
10381 | =for comment openssl_section:XXX | |
10382 | ||
10383 | The blank line is mandatory. | |
10384 | ||
10385 | *Steve Henson* | |
10386 | ||
10387 | * New arguments -certform, -keyform and -pass for s_client and s_server | |
10388 | to allow alternative format key and certificate files and passphrase | |
10389 | sources. | |
10390 | ||
10391 | *Steve Henson* | |
10392 | ||
10393 | * New structure X509_VERIFY_PARAM which combines current verify parameters, | |
10394 | update associated structures and add various utility functions. | |
10395 | ||
10396 | Add new policy related verify parameters, include policy checking in | |
10397 | standard verify code. Enhance 'smime' application with extra parameters | |
10398 | to support policy checking and print out. | |
10399 | ||
10400 | *Steve Henson* | |
10401 | ||
10402 | * Add a new engine to support VIA PadLock ACE extensions in the VIA C3 | |
10403 | Nehemiah processors. These extensions support AES encryption in hardware | |
10404 | as well as RNG (though RNG support is currently disabled). | |
10405 | ||
10406 | *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov* | |
10407 | ||
10408 | * Deprecate `BN_[get|set]_params()` functions (they were ignored internally). | |
10409 | ||
10410 | *Geoff Thorpe* | |
10411 | ||
10412 | * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. | |
10413 | ||
10414 | *Andy Polyakov and a number of other people* | |
10415 | ||
10416 | * Improved PowerPC platform support. Most notably BIGNUM assembler | |
10417 | implementation contributed by IBM. | |
10418 | ||
10419 | *Suresh Chari, Peter Waltenberg, Andy Polyakov* | |
10420 | ||
10421 | * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public | |
10422 | exponent rather than 'unsigned long'. There is a corresponding change to | |
10423 | the new 'rsa_keygen' element of the RSA_METHOD structure. | |
10424 | ||
10425 | *Jelte Jansen, Geoff Thorpe* | |
10426 | ||
10427 | * Functionality for creating the initial serial number file is now | |
10428 | moved from CA.pl to the 'ca' utility with a new option -create_serial. | |
10429 | ||
10430 | (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial | |
10431 | number file to 1, which is bound to cause problems. To avoid | |
10432 | the problems while respecting compatibility between different 0.9.7 | |
10433 | patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in | |
10434 | CA.pl for serial number initialization. With the new release 0.9.8, | |
10435 | we can fix the problem directly in the 'ca' utility.) | |
10436 | ||
10437 | *Steve Henson* | |
10438 | ||
10439 | * Reduced header interdependencies by declaring more opaque objects in | |
10440 | ossl_typ.h. As a consequence, including some headers (eg. engine.h) will | |
10441 | give fewer recursive includes, which could break lazy source code - so | |
10442 | this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, | |
10443 | developers should define this symbol when building and using openssl to | |
10444 | ensure they track the recommended behaviour, interfaces, [etc], but | |
10445 | backwards-compatible behaviour prevails when this isn't defined. | |
10446 | ||
10447 | *Geoff Thorpe* | |
10448 | ||
10449 | * New function X509_POLICY_NODE_print() which prints out policy nodes. | |
10450 | ||
10451 | *Steve Henson* | |
10452 | ||
10453 | * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. | |
10454 | This will generate a random key of the appropriate length based on the | |
10455 | cipher context. The EVP_CIPHER can provide its own random key generation | |
10456 | routine to support keys of a specific form. This is used in the des and | |
10457 | 3des routines to generate a key of the correct parity. Update S/MIME | |
10458 | code to use new functions and hence generate correct parity DES keys. | |
10459 | Add EVP_CHECK_DES_KEY #define to return an error if the key is not | |
10460 | valid (weak or incorrect parity). | |
10461 | ||
10462 | *Steve Henson* | |
10463 | ||
10464 | * Add a local set of CRLs that can be used by X509_verify_cert() as well | |
10465 | as looking them up. This is useful when the verified structure may contain | |
10466 | CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs | |
10467 | present unless the new PKCS7_NO_CRL flag is asserted. | |
10468 | ||
10469 | *Steve Henson* | |
10470 | ||
10471 | * Extend ASN1 oid configuration module. It now additionally accepts the | |
10472 | syntax: | |
10473 | ||
10474 | shortName = some long name, 1.2.3.4 | |
10475 | ||
10476 | *Steve Henson* | |
10477 | ||
10478 | * Reimplemented the BN_CTX implementation. There is now no more static | |
10479 | limitation on the number of variables it can handle nor the depth of the | |
10480 | "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack | |
10481 | information can now expand as required, and rather than having a single | |
10482 | static array of bignums, BN_CTX now uses a linked-list of such arrays | |
10483 | allowing it to expand on demand whilst maintaining the usefulness of | |
10484 | BN_CTX's "bundling". | |
10485 | ||
10486 | *Geoff Thorpe* | |
10487 | ||
10488 | * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD | |
10489 | to allow all RSA operations to function using a single BN_CTX. | |
10490 | ||
10491 | *Geoff Thorpe* | |
10492 | ||
10493 | * Preliminary support for certificate policy evaluation and checking. This | |
10494 | is initially intended to pass the tests outlined in "Conformance Testing | |
10495 | of Relying Party Client Certificate Path Processing Logic" v1.07. | |
10496 | ||
10497 | *Steve Henson* | |
10498 | ||
10499 | * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and | |
10500 | remained unused and not that useful. A variety of other little bignum | |
10501 | tweaks and fixes have also been made continuing on from the audit (see | |
10502 | below). | |
10503 | ||
10504 | *Geoff Thorpe* | |
10505 | ||
10506 | * Constify all or almost all d2i, c2i, s2i and r2i functions, along with | |
10507 | associated ASN1, EVP and SSL functions and old ASN1 macros. | |
10508 | ||
10509 | *Richard Levitte* | |
10510 | ||
10511 | * BN_zero() only needs to set 'top' and 'neg' to zero for correct results, | |
10512 | and this should never fail. So the return value from the use of | |
10513 | BN_set_word() (which can fail due to needless expansion) is now deprecated; | |
10514 | if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. | |
10515 | ||
10516 | *Geoff Thorpe* | |
10517 | ||
10518 | * BN_CTX_get() should return zero-valued bignums, providing the same | |
10519 | initialised value as BN_new(). | |
10520 | ||
10521 | *Geoff Thorpe, suggested by Ulf Möller* | |
10522 | ||
10523 | * Support for inhibitAnyPolicy certificate extension. | |
10524 | ||
10525 | *Steve Henson* | |
10526 | ||
10527 | * An audit of the BIGNUM code is underway, for which debugging code is | |
10528 | enabled when BN_DEBUG is defined. This makes stricter enforcements on what | |
10529 | is considered valid when processing BIGNUMs, and causes execution to | |
10530 | assert() when a problem is discovered. If BN_DEBUG_RAND is defined, | |
10531 | further steps are taken to deliberately pollute unused data in BIGNUM | |
10532 | structures to try and expose faulty code further on. For now, openssl will | |
10533 | (in its default mode of operation) continue to tolerate the inconsistent | |
10534 | forms that it has tolerated in the past, but authors and packagers should | |
10535 | consider trying openssl and their own applications when compiled with | |
10536 | these debugging symbols defined. It will help highlight potential bugs in | |
10537 | their own code, and will improve the test coverage for OpenSSL itself. At | |
10538 | some point, these tighter rules will become openssl's default to improve | |
10539 | maintainability, though the assert()s and other overheads will remain only | |
10540 | in debugging configurations. See bn.h for more details. | |
10541 | ||
10542 | *Geoff Thorpe, Nils Larsch, Ulf Möller* | |
10543 | ||
10544 | * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure | |
10545 | that can only be obtained through BN_CTX_new() (which implicitly | |
10546 | initialises it). The presence of this function only made it possible | |
10547 | to overwrite an existing structure (and cause memory leaks). | |
10548 | ||
10549 | *Geoff Thorpe* | |
10550 | ||
10551 | * Because of the callback-based approach for implementing LHASH as a | |
10552 | template type, lh_insert() adds opaque objects to hash-tables and | |
10553 | lh_doall() or lh_doall_arg() are typically used with a destructor callback | |
10554 | to clean up those corresponding objects before destroying the hash table | |
10555 | (and losing the object pointers). So some over-zealous constifications in | |
10556 | LHASH have been relaxed so that lh_insert() does not take (nor store) the | |
10557 | objects as "const" and the `lh_doall[_arg]` callback wrappers are not | |
10558 | prototyped to have "const" restrictions on the object pointers they are | |
10559 | given (and so aren't required to cast them away any more). | |
10560 | ||
10561 | *Geoff Thorpe* | |
10562 | ||
10563 | * The tmdiff.h API was so ugly and minimal that our own timing utility | |
10564 | (speed) prefers to use its own implementation. The two implementations | |
10565 | haven't been consolidated as yet (volunteers?) but the tmdiff API has had | |
10566 | its object type properly exposed (MS_TM) instead of casting to/from | |
10567 | `char *`. This may still change yet if someone realises MS_TM and | |
10568 | `ms_time_***` | |
10569 | aren't necessarily the greatest nomenclatures - but this is what was used | |
10570 | internally to the implementation so I've used that for now. | |
10571 | ||
10572 | *Geoff Thorpe* | |
10573 | ||
10574 | * Ensure that deprecated functions do not get compiled when | |
10575 | OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of | |
10576 | the self-tests were still using deprecated key-generation functions so | |
10577 | these have been updated also. | |
10578 | ||
10579 | *Geoff Thorpe* | |
10580 | ||
10581 | * Reorganise PKCS#7 code to separate the digest location functionality | |
10582 | into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest(). | |
10583 | New function PKCS7_set_digest() to set the digest type for PKCS#7 | |
10584 | digestedData type. Add additional code to correctly generate the | |
10585 | digestedData type and add support for this type in PKCS7 initialization | |
10586 | functions. | |
10587 | ||
10588 | *Steve Henson* | |
10589 | ||
10590 | * New function PKCS7_set0_type_other() this initializes a PKCS7 | |
10591 | structure of type "other". | |
10592 | ||
10593 | *Steve Henson* | |
10594 | ||
10595 | * Fix prime generation loop in crypto/bn/bn_prime.pl by making | |
10596 | sure the loop does correctly stop and breaking ("division by zero") | |
10597 | modulus operations are not performed. The (pre-generated) prime | |
10598 | table crypto/bn/bn_prime.h was already correct, but it could not be | |
10599 | re-generated on some platforms because of the "division by zero" | |
10600 | situation in the script. | |
10601 | ||
10602 | *Ralf S. Engelschall* | |
10603 | ||
10604 | * Update support for ECC-based TLS ciphersuites according to | |
10605 | draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with | |
10606 | SHA-1 now is only used for "small" curves (where the | |
10607 | representation of a field element takes up to 24 bytes); for | |
10608 | larger curves, the field element resulting from ECDH is directly | |
10609 | used as premaster secret. | |
10610 | ||
10611 | *Douglas Stebila (Sun Microsystems Laboratories)* | |
10612 | ||
10613 | * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 | |
10614 | curve secp160r1 to the tests. | |
10615 | ||
10616 | *Douglas Stebila (Sun Microsystems Laboratories)* | |
10617 | ||
10618 | * Add the possibility to load symbols globally with DSO. | |
10619 | ||
10620 | *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte* | |
10621 | ||
10622 | * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better | |
10623 | control of the error stack. | |
10624 | ||
10625 | *Richard Levitte* | |
10626 | ||
10627 | * Add support for STORE in ENGINE. | |
10628 | ||
10629 | *Richard Levitte* | |
10630 | ||
10631 | * Add the STORE type. The intention is to provide a common interface | |
10632 | to certificate and key stores, be they simple file-based stores, or | |
10633 | HSM-type store, or LDAP stores, or... | |
10634 | NOTE: The code is currently UNTESTED and isn't really used anywhere. | |
10635 | ||
10636 | *Richard Levitte* | |
10637 | ||
10638 | * Add a generic structure called OPENSSL_ITEM. This can be used to | |
10639 | pass a list of arguments to any function as well as provide a way | |
10640 | for a function to pass data back to the caller. | |
10641 | ||
10642 | *Richard Levitte* | |
10643 | ||
10644 | * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() | |
10645 | works like BUF_strdup() but can be used to duplicate a portion of | |
10646 | a string. The copy gets NUL-terminated. BUF_memdup() duplicates | |
10647 | a memory area. | |
10648 | ||
10649 | *Richard Levitte* | |
10650 | ||
10651 | * Add the function sk_find_ex() which works like sk_find(), but will | |
10652 | return an index to an element even if an exact match couldn't be | |
10653 | found. The index is guaranteed to point at the element where the | |
10654 | searched-for key would be inserted to preserve sorting order. | |
10655 | ||
10656 | *Richard Levitte* | |
10657 | ||
10658 | * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but | |
10659 | takes an extra flags argument for optional functionality. Currently, | |
10660 | the following flags are defined: | |
10661 | ||
10662 | OBJ_BSEARCH_VALUE_ON_NOMATCH | |
10663 | This one gets OBJ_bsearch_ex() to return a pointer to the first | |
10664 | element where the comparing function returns a negative or zero | |
10665 | number. | |
10666 | ||
10667 | OBJ_BSEARCH_FIRST_VALUE_ON_MATCH | |
10668 | This one gets OBJ_bsearch_ex() to return a pointer to the first | |
10669 | element where the comparing function returns zero. This is useful | |
10670 | if there are more than one element where the comparing function | |
10671 | returns zero. | |
10672 | ||
10673 | *Richard Levitte* | |
10674 | ||
10675 | * Make it possible to create self-signed certificates with 'openssl ca' | |
10676 | in such a way that the self-signed certificate becomes part of the | |
10677 | CA database and uses the same mechanisms for serial number generation | |
10678 | as all other certificate signing. The new flag '-selfsign' enables | |
10679 | this functionality. Adapt CA.sh and CA.pl.in. | |
10680 | ||
10681 | *Richard Levitte* | |
10682 | ||
10683 | * Add functionality to check the public key of a certificate request | |
10684 | against a given private. This is useful to check that a certificate | |
10685 | request can be signed by that key (self-signing). | |
10686 | ||
10687 | *Richard Levitte* | |
10688 | ||
10689 | * Make it possible to have multiple active certificates with the same | |
10690 | subject in the CA index file. This is done only if the keyword | |
10691 | 'unique_subject' is set to 'no' in the main CA section (default | |
10692 | if 'CA_default') of the configuration file. The value is saved | |
10693 | with the database itself in a separate index attribute file, | |
10694 | named like the index file with '.attr' appended to the name. | |
10695 | ||
10696 | *Richard Levitte* | |
10697 | ||
10698 | * Generate multi-valued AVAs using '+' notation in config files for | |
10699 | req and dirName. | |
10700 | ||
10701 | *Steve Henson* | |
10702 | ||
10703 | * Support for nameConstraints certificate extension. | |
10704 | ||
10705 | *Steve Henson* | |
10706 | ||
10707 | * Support for policyConstraints certificate extension. | |
10708 | ||
10709 | *Steve Henson* | |
10710 | ||
10711 | * Support for policyMappings certificate extension. | |
10712 | ||
10713 | *Steve Henson* | |
10714 | ||
10715 | * Make sure the default DSA_METHOD implementation only uses its | |
10716 | dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, | |
10717 | and change its own handlers to be NULL so as to remove unnecessary | |
10718 | indirection. This lets alternative implementations fallback to the | |
10719 | default implementation more easily. | |
10720 | ||
10721 | *Geoff Thorpe* | |
10722 | ||
10723 | * Support for directoryName in GeneralName related extensions | |
10724 | in config files. | |
10725 | ||
10726 | *Steve Henson* | |
10727 | ||
10728 | * Make it possible to link applications using Makefile.shared. | |
10729 | Make that possible even when linking against static libraries! | |
10730 | ||
10731 | *Richard Levitte* | |
10732 | ||
10733 | * Support for single pass processing for S/MIME signing. This now | |
10734 | means that S/MIME signing can be done from a pipe, in addition | |
10735 | cleartext signing (multipart/signed type) is effectively streaming | |
10736 | and the signed data does not need to be all held in memory. | |
10737 | ||
10738 | This is done with a new flag PKCS7_STREAM. When this flag is set | |
10739 | PKCS7_sign() only initializes the PKCS7 structure and the actual signing | |
10740 | is done after the data is output (and digests calculated) in | |
10741 | SMIME_write_PKCS7(). | |
10742 | ||
10743 | *Steve Henson* | |
10744 | ||
10745 | * Add full support for -rpath/-R, both in shared libraries and | |
10746 | applications, at least on the platforms where it's known how | |
10747 | to do it. | |
10748 | ||
10749 | *Richard Levitte* | |
10750 | ||
10751 | * In crypto/ec/ec_mult.c, implement fast point multiplication with | |
10752 | precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() | |
10753 | will now compute a table of multiples of the generator that | |
10754 | makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() | |
10755 | faster (notably in the case of a single point multiplication, | |
10756 | scalar * generator). | |
10757 | ||
10758 | *Nils Larsch, Bodo Moeller* | |
10759 | ||
10760 | * IPv6 support for certificate extensions. The various extensions | |
10761 | which use the IP:a.b.c.d can now take IPv6 addresses using the | |
10762 | formats of RFC1884 2.2 . IPv6 addresses are now also displayed | |
10763 | correctly. | |
10764 | ||
10765 | *Steve Henson* | |
10766 | ||
10767 | * Added an ENGINE that implements RSA by performing private key | |
10768 | exponentiations with the GMP library. The conversions to and from | |
10769 | GMP's mpz_t format aren't optimised nor are any montgomery forms | |
10770 | cached, and on x86 it appears OpenSSL's own performance has caught up. | |
10771 | However there are likely to be other architectures where GMP could | |
10772 | provide a boost. This ENGINE is not built in by default, but it can be | |
10773 | specified at Configure time and should be accompanied by the necessary | |
10774 | linker additions, eg; | |
10775 | ./config -DOPENSSL_USE_GMP -lgmp | |
10776 | ||
10777 | *Geoff Thorpe* | |
10778 | ||
10779 | * "openssl engine" will not display ENGINE/DSO load failure errors when | |
10780 | testing availability of engines with "-t" - the old behaviour is | |
10781 | produced by increasing the feature's verbosity with "-tt". | |
10782 | ||
10783 | *Geoff Thorpe* | |
10784 | ||
10785 | * ECDSA routines: under certain error conditions uninitialized BN objects | |
10786 | could be freed. Solution: make sure initialization is performed early | |
10787 | enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> | |
10788 | via PR#459) | |
10789 | ||
10790 | *Lutz Jaenicke* | |
10791 | ||
10792 | * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD | |
10793 | and DH_METHOD (eg. by ENGINE implementations) to override the normal | |
10794 | software implementations. For DSA and DH, parameter generation can | |
10795 | also be overridden by providing the appropriate method callbacks. | |
10796 | ||
10797 | *Geoff Thorpe* | |
10798 | ||
10799 | * Change the "progress" mechanism used in key-generation and | |
10800 | primality testing to functions that take a new BN_GENCB pointer in | |
10801 | place of callback/argument pairs. The new API functions have `_ex` | |
10802 | postfixes and the older functions are reimplemented as wrappers for | |
10803 | the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide | |
10804 | declarations of the old functions to help (graceful) attempts to | |
10805 | migrate to the new functions. Also, the new key-generation API | |
10806 | functions operate on a caller-supplied key-structure and return | |
10807 | success/failure rather than returning a key or NULL - this is to | |
10808 | help make "keygen" another member function of RSA_METHOD etc. | |
10809 | ||
10810 | Example for using the new callback interface: | |
10811 | ||
10812 | int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; | |
10813 | void *my_arg = ...; | |
10814 | BN_GENCB my_cb; | |
10815 | ||
10816 | BN_GENCB_set(&my_cb, my_callback, my_arg); | |
10817 | ||
10818 | return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); | |
10819 | /* For the meaning of a, b in calls to my_callback(), see the | |
10820 | * documentation of the function that calls the callback. | |
10821 | * cb will point to my_cb; my_arg can be retrieved as cb->arg. | |
10822 | * my_callback should return 1 if it wants BN_is_prime_ex() | |
10823 | * to continue, or 0 to stop. | |
10824 | */ | |
10825 | ||
10826 | *Geoff Thorpe* | |
10827 | ||
10828 | * Change the ZLIB compression method to be stateful, and make it | |
10829 | available to TLS with the number defined in | |
10830 | draft-ietf-tls-compression-04.txt. | |
10831 | ||
10832 | *Richard Levitte* | |
10833 | ||
10834 | * Add the ASN.1 structures and functions for CertificatePair, which | |
10835 | is defined as follows (according to X.509_4thEditionDraftV6.pdf): | |
10836 | ||
10837 | CertificatePair ::= SEQUENCE { | |
10838 | forward [0] Certificate OPTIONAL, | |
10839 | reverse [1] Certificate OPTIONAL, | |
10840 | -- at least one of the pair shall be present -- } | |
10841 | ||
10842 | Also implement the PEM functions to read and write certificate | |
10843 | pairs, and defined the PEM tag as "CERTIFICATE PAIR". | |
10844 | ||
10845 | This needed to be defined, mostly for the sake of the LDAP | |
10846 | attribute crossCertificatePair, but may prove useful elsewhere as | |
10847 | well. | |
10848 | ||
10849 | *Richard Levitte* | |
10850 | ||
10851 | * Make it possible to inhibit symlinking of shared libraries in | |
10852 | Makefile.shared, for Cygwin's sake. | |
10853 | ||
10854 | *Richard Levitte* | |
10855 | ||
10856 | * Extend the BIGNUM API by creating a function | |
10857 | void BN_set_negative(BIGNUM *a, int neg); | |
10858 | and a macro that behave like | |
10859 | int BN_is_negative(const BIGNUM *a); | |
10860 | ||
10861 | to avoid the need to access 'a->neg' directly in applications. | |
10862 | ||
10863 | *Nils Larsch* | |
10864 | ||
10865 | * Implement fast modular reduction for pseudo-Mersenne primes | |
10866 | used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). | |
10867 | EC_GROUP_new_curve_GFp() will now automatically use this | |
10868 | if applicable. | |
10869 | ||
10870 | *Nils Larsch <nla@trustcenter.de>* | |
10871 | ||
10872 | * Add new lock type (CRYPTO_LOCK_BN). | |
10873 | ||
10874 | *Bodo Moeller* | |
10875 | ||
10876 | * Change the ENGINE framework to automatically load engines | |
10877 | dynamically from specific directories unless they could be | |
10878 | found to already be built in or loaded. Move all the | |
10879 | current engines except for the cryptodev one to a new | |
10880 | directory engines/. | |
10881 | The engines in engines/ are built as shared libraries if | |
10882 | the "shared" options was given to ./Configure or ./config. | |
10883 | Otherwise, they are inserted in libcrypto.a. | |
10884 | /usr/local/ssl/engines is the default directory for dynamic | |
10885 | engines, but that can be overridden at configure time through | |
10886 | the usual use of --prefix and/or --openssldir, and at run | |
10887 | time with the environment variable OPENSSL_ENGINES. | |
10888 | ||
10889 | *Geoff Thorpe and Richard Levitte* | |
10890 | ||
10891 | * Add Makefile.shared, a helper makefile to build shared | |
10892 | libraries. Adapt Makefile.org. | |
10893 | ||
10894 | *Richard Levitte* | |
10895 | ||
10896 | * Add version info to Win32 DLLs. | |
10897 | ||
10898 | *Peter 'Luna' Runestig" <peter@runestig.com>* | |
10899 | ||
10900 | * Add new 'medium level' PKCS#12 API. Certificates and keys | |
10901 | can be added using this API to created arbitrary PKCS#12 | |
10902 | files while avoiding the low level API. | |
10903 | ||
10904 | New options to PKCS12_create(), key or cert can be NULL and | |
10905 | will then be omitted from the output file. The encryption | |
10906 | algorithm NIDs can be set to -1 for no encryption, the mac | |
10907 | iteration count can be set to 0 to omit the mac. | |
10908 | ||
10909 | Enhance pkcs12 utility by making the -nokeys and -nocerts | |
10910 | options work when creating a PKCS#12 file. New option -nomac | |
10911 | to omit the mac, NONE can be set for an encryption algorithm. | |
10912 | New code is modified to use the enhanced PKCS12_create() | |
10913 | instead of the low level API. | |
10914 | ||
10915 | *Steve Henson* | |
10916 | ||
10917 | * Extend ASN1 encoder to support indefinite length constructed | |
10918 | encoding. This can output sequences tags and octet strings in | |
10919 | this form. Modify pk7_asn1.c to support indefinite length | |
10920 | encoding. This is experimental and needs additional code to | |
10921 | be useful, such as an ASN1 bio and some enhanced streaming | |
10922 | PKCS#7 code. | |
10923 | ||
10924 | Extend template encode functionality so that tagging is passed | |
10925 | down to the template encoder. | |
10926 | ||
10927 | *Steve Henson* | |
10928 | ||
10929 | * Let 'openssl req' fail if an argument to '-newkey' is not | |
10930 | recognized instead of using RSA as a default. | |
10931 | ||
10932 | *Bodo Moeller* | |
10933 | ||
10934 | * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. | |
10935 | As these are not official, they are not included in "ALL"; | |
10936 | the "ECCdraft" ciphersuite group alias can be used to select them. | |
10937 | ||
10938 | *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)* | |
10939 | ||
10940 | * Add ECDH engine support. | |
10941 | ||
10942 | *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)* | |
10943 | ||
10944 | * Add ECDH in new directory crypto/ecdh/. | |
10945 | ||
10946 | *Douglas Stebila (Sun Microsystems Laboratories)* | |
10947 | ||
10948 | * Let BN_rand_range() abort with an error after 100 iterations | |
10949 | without success (which indicates a broken PRNG). | |
10950 | ||
10951 | *Bodo Moeller* | |
10952 | ||
10953 | * Change BN_mod_sqrt() so that it verifies that the input value | |
10954 | is really the square of the return value. (Previously, | |
10955 | BN_mod_sqrt would show GIGO behaviour.) | |
10956 | ||
10957 | *Bodo Moeller* | |
10958 | ||
10959 | * Add named elliptic curves over binary fields from X9.62, SECG, | |
10960 | and WAP/WTLS; add OIDs that were still missing. | |
10961 | ||
10962 | *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* | |
10963 | ||
10964 | * Extend the EC library for elliptic curves over binary fields | |
10965 | (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). | |
10966 | New EC_METHOD: | |
10967 | ||
10968 | EC_GF2m_simple_method | |
10969 | ||
10970 | New API functions: | |
10971 | ||
10972 | EC_GROUP_new_curve_GF2m | |
10973 | EC_GROUP_set_curve_GF2m | |
10974 | EC_GROUP_get_curve_GF2m | |
10975 | EC_POINT_set_affine_coordinates_GF2m | |
10976 | EC_POINT_get_affine_coordinates_GF2m | |
10977 | EC_POINT_set_compressed_coordinates_GF2m | |
10978 | ||
10979 | Point compression for binary fields is disabled by default for | |
10980 | patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to | |
10981 | enable it). | |
10982 | ||
10983 | As binary polynomials are represented as BIGNUMs, various members | |
10984 | of the EC_GROUP and EC_POINT data structures can be shared | |
10985 | between the implementations for prime fields and binary fields; | |
10986 | the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m) | |
10987 | are essentially identical to their `..._GFp` counterparts. | |
10988 | (For simplicity, the `..._GFp` prefix has been dropped from | |
10989 | various internal method names.) | |
10990 | ||
10991 | An internal 'field_div' method (similar to 'field_mul' and | |
10992 | 'field_sqr') has been added; this is used only for binary fields. | |
10993 | ||
10994 | *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* | |
10995 | ||
10996 | * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() | |
10997 | through methods ('mul', 'precompute_mult'). | |
10998 | ||
10999 | The generic implementations (now internally called 'ec_wNAF_mul' | |
11000 | and 'ec_wNAF_precomputed_mult') remain the default if these | |
11001 | methods are undefined. | |
11002 | ||
11003 | *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* | |
11004 | ||
11005 | * New function EC_GROUP_get_degree, which is defined through | |
11006 | EC_METHOD. For curves over prime fields, this returns the bit | |
11007 | length of the modulus. | |
11008 | ||
11009 | *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* | |
11010 | ||
11011 | * New functions EC_GROUP_dup, EC_POINT_dup. | |
11012 | (These simply call ..._new and ..._copy). | |
11013 | ||
11014 | *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* | |
11015 | ||
11016 | * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. | |
11017 | Polynomials are represented as BIGNUMs (where the sign bit is not | |
11018 | used) in the following functions [macros]: | |
11019 | ||
11020 | BN_GF2m_add | |
11021 | BN_GF2m_sub [= BN_GF2m_add] | |
11022 | BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] | |
11023 | BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] | |
11024 | BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] | |
11025 | BN_GF2m_mod_inv | |
11026 | BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] | |
11027 | BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] | |
11028 | BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] | |
11029 | BN_GF2m_cmp [= BN_ucmp] | |
11030 | ||
11031 | (Note that only the 'mod' functions are actually for fields GF(2^m). | |
11032 | BN_GF2m_add() is misnomer, but this is for the sake of consistency.) | |
11033 | ||
11034 | For some functions, an the irreducible polynomial defining a | |
11035 | field can be given as an 'unsigned int[]' with strictly | |
11036 | decreasing elements giving the indices of those bits that are set; | |
11037 | i.e., p[] represents the polynomial | |
11038 | f(t) = t^p[0] + t^p[1] + ... + t^p[k] | |
11039 | where | |
11040 | p[0] > p[1] > ... > p[k] = 0. | |
11041 | This applies to the following functions: | |
11042 | ||
11043 | BN_GF2m_mod_arr | |
11044 | BN_GF2m_mod_mul_arr | |
11045 | BN_GF2m_mod_sqr_arr | |
11046 | BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] | |
11047 | BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] | |
11048 | BN_GF2m_mod_exp_arr | |
11049 | BN_GF2m_mod_sqrt_arr | |
11050 | BN_GF2m_mod_solve_quad_arr | |
11051 | BN_GF2m_poly2arr | |
11052 | BN_GF2m_arr2poly | |
11053 | ||
11054 | Conversion can be performed by the following functions: | |
11055 | ||
11056 | BN_GF2m_poly2arr | |
11057 | BN_GF2m_arr2poly | |
11058 | ||
11059 | bntest.c has additional tests for binary polynomial arithmetic. | |
11060 | ||
11061 | Two implementations for BN_GF2m_mod_div() are available. | |
11062 | The default algorithm simply uses BN_GF2m_mod_inv() and | |
11063 | BN_GF2m_mod_mul(). The alternative algorithm is compiled in only | |
11064 | if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the | |
11065 | copyright notice in crypto/bn/bn_gf2m.c before enabling it). | |
11066 | ||
11067 | *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* | |
11068 | ||
11069 | * Add new error code 'ERR_R_DISABLED' that can be used when some | |
11070 | functionality is disabled at compile-time. | |
11071 | ||
11072 | *Douglas Stebila <douglas.stebila@sun.com>* | |
11073 | ||
11074 | * Change default behaviour of 'openssl asn1parse' so that more | |
11075 | information is visible when viewing, e.g., a certificate: | |
11076 | ||
11077 | Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' | |
11078 | mode the content of non-printable OCTET STRINGs is output in a | |
11079 | style similar to INTEGERs, but with '[HEX DUMP]' prepended to | |
11080 | avoid the appearance of a printable string. | |
11081 | ||
11082 | *Nils Larsch <nla@trustcenter.de>* | |
11083 | ||
11084 | * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access | |
11085 | functions | |
11086 | EC_GROUP_set_asn1_flag() | |
11087 | EC_GROUP_get_asn1_flag() | |
11088 | EC_GROUP_set_point_conversion_form() | |
11089 | EC_GROUP_get_point_conversion_form() | |
11090 | These control ASN1 encoding details: | |
11091 | - Curves (i.e., groups) are encoded explicitly unless asn1_flag | |
11092 | has been set to OPENSSL_EC_NAMED_CURVE. | |
11093 | - Points are encoded in uncompressed form by default; options for | |
11094 | asn1_for are as for point2oct, namely | |
11095 | POINT_CONVERSION_COMPRESSED | |
11096 | POINT_CONVERSION_UNCOMPRESSED | |
11097 | POINT_CONVERSION_HYBRID | |
11098 | ||
11099 | Also add 'seed' and 'seed_len' members to EC_GROUP with access | |
11100 | functions | |
11101 | EC_GROUP_set_seed() | |
11102 | EC_GROUP_get0_seed() | |
11103 | EC_GROUP_get_seed_len() | |
11104 | This is used only for ASN1 purposes (so far). | |
11105 | ||
11106 | *Nils Larsch <nla@trustcenter.de>* | |
11107 | ||
11108 | * Add 'field_type' member to EC_METHOD, which holds the NID | |
11109 | of the appropriate field type OID. The new function | |
11110 | EC_METHOD_get_field_type() returns this value. | |
11111 | ||
11112 | *Nils Larsch <nla@trustcenter.de>* | |
11113 | ||
11114 | * Add functions | |
11115 | EC_POINT_point2bn() | |
11116 | EC_POINT_bn2point() | |
11117 | EC_POINT_point2hex() | |
11118 | EC_POINT_hex2point() | |
11119 | providing useful interfaces to EC_POINT_point2oct() and | |
11120 | EC_POINT_oct2point(). | |
11121 | ||
11122 | *Nils Larsch <nla@trustcenter.de>* | |
11123 | ||
11124 | * Change internals of the EC library so that the functions | |
11125 | EC_GROUP_set_generator() | |
11126 | EC_GROUP_get_generator() | |
11127 | EC_GROUP_get_order() | |
11128 | EC_GROUP_get_cofactor() | |
11129 | are implemented directly in crypto/ec/ec_lib.c and not dispatched | |
11130 | to methods, which would lead to unnecessary code duplication when | |
11131 | adding different types of curves. | |
11132 | ||
11133 | *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller* | |
11134 | ||
11135 | * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM | |
11136 | arithmetic, and such that modified wNAFs are generated | |
11137 | (which avoid length expansion in many cases). | |
11138 | ||
11139 | *Bodo Moeller* | |
11140 | ||
11141 | * Add a function EC_GROUP_check_discriminant() (defined via | |
11142 | EC_METHOD) that verifies that the curve discriminant is non-zero. | |
11143 | ||
11144 | Add a function EC_GROUP_check() that makes some sanity tests | |
11145 | on a EC_GROUP, its generator and order. This includes | |
11146 | EC_GROUP_check_discriminant(). | |
11147 | ||
11148 | *Nils Larsch <nla@trustcenter.de>* | |
11149 | ||
11150 | * Add ECDSA in new directory crypto/ecdsa/. | |
11151 | ||
11152 | Add applications 'openssl ecparam' and 'openssl ecdsa' | |
11153 | (these are based on 'openssl dsaparam' and 'openssl dsa'). | |
11154 | ||
11155 | ECDSA support is also included in various other files across the | |
11156 | library. Most notably, | |
11157 | - 'openssl req' now has a '-newkey ecdsa:file' option; | |
11158 | - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; | |
11159 | - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and | |
11160 | d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make | |
11161 | them suitable for ECDSA where domain parameters must be | |
11162 | extracted before the specific public key; | |
11163 | - ECDSA engine support has been added. | |
11164 | ||
11165 | *Nils Larsch <nla@trustcenter.de>* | |
11166 | ||
11167 | * Include some named elliptic curves, and add OIDs from X9.62, | |
11168 | SECG, and WAP/WTLS. Each curve can be obtained from the new | |
11169 | function | |
11170 | EC_GROUP_new_by_curve_name(), | |
11171 | and the list of available named curves can be obtained with | |
11172 | EC_get_builtin_curves(). | |
11173 | Also add a 'curve_name' member to EC_GROUP objects, which can be | |
11174 | accessed via | |
11175 | EC_GROUP_set_curve_name() | |
11176 | EC_GROUP_get_curve_name() | |
11177 | ||
11178 | *Nils Larsch <larsch@trustcenter.de, Bodo Moeller* | |
11179 | ||
11180 | * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there | |
11181 | was actually never needed) and in BN_mul(). The removal in BN_mul() | |
11182 | required a small change in bn_mul_part_recursive() and the addition | |
11183 | of the functions bn_cmp_part_words(), bn_sub_part_words() and | |
11184 | bn_add_part_words(), which do the same thing as bn_cmp_words(), | |
11185 | bn_sub_words() and bn_add_words() except they take arrays with | |
11186 | differing sizes. | |
11187 | ||
11188 | *Richard Levitte* | |
11189 | ||
11190 | ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007] | |
11191 | ||
11192 | * Cleanse PEM buffers before freeing them since they may contain | |
11193 | sensitive data. | |
11194 | ||
11195 | *Benjamin Bennett <ben@psc.edu>* | |
11196 | ||
11197 | * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that | |
11198 | a ciphersuite string such as "DEFAULT:RSA" cannot enable | |
11199 | authentication-only ciphersuites. | |
11200 | ||
11201 | *Bodo Moeller* | |
11202 | ||
11203 | * Since AES128 and AES256 share a single mask bit in the logic of | |
11204 | ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a | |
11205 | kludge to work properly if AES128 is available and AES256 isn't. | |
11206 | ||
11207 | *Victor Duchovni* | |
11208 | ||
11209 | * Expand security boundary to match 1.1.1 module. | |
11210 | ||
11211 | *Steve Henson* | |
11212 | ||
11213 | * Remove redundant features: hash file source, editing of test vectors | |
11214 | modify fipsld to use external fips_premain.c signature. | |
11215 | ||
11216 | *Steve Henson* | |
11217 | ||
11218 | * New perl script mkfipsscr.pl to create shell scripts or batch files to | |
11219 | run algorithm test programs. | |
11220 | ||
11221 | *Steve Henson* | |
11222 | ||
11223 | * Make algorithm test programs more tolerant of whitespace. | |
11224 | ||
11225 | *Steve Henson* | |
11226 | ||
11227 | * Have SSL/TLS server implementation tolerate "mismatched" record | |
11228 | protocol version while receiving ClientHello even if the | |
11229 | ClientHello is fragmented. (The server can't insist on the | |
11230 | particular protocol version it has chosen before the ServerHello | |
11231 | message has informed the client about his choice.) | |
11232 | ||
11233 | *Bodo Moeller* | |
11234 | ||
11235 | * Load error codes if they are not already present instead of using a | |
11236 | static variable. This allows them to be cleanly unloaded and reloaded. | |
11237 | ||
11238 | *Steve Henson* | |
11239 | ||
11240 | ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006] | |
11241 | ||
11242 | * Introduce limits to prevent malicious keys being able to | |
11243 | cause a denial of service. ([CVE-2006-2940]) | |
11244 | ||
11245 | *Steve Henson, Bodo Moeller* | |
11246 | ||
11247 | * Fix ASN.1 parsing of certain invalid structures that can result | |
11248 | in a denial of service. ([CVE-2006-2937]) [Steve Henson] | |
11249 | ||
11250 | * Fix buffer overflow in SSL_get_shared_ciphers() function. | |
11251 | ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team] | |
11252 | ||
11253 | * Fix SSL client code which could crash if connecting to a | |
11254 | malicious SSLv2 server. ([CVE-2006-4343]) | |
11255 | ||
11256 | *Tavis Ormandy and Will Drewry, Google Security Team* | |
11257 | ||
11258 | * Change ciphersuite string processing so that an explicit | |
11259 | ciphersuite selects this one ciphersuite (so that "AES256-SHA" | |
11260 | will no longer include "AES128-SHA"), and any other similar | |
11261 | ciphersuite (same bitmap) from *other* protocol versions (so that | |
11262 | "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the | |
11263 | SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining | |
11264 | changes from 0.9.8b and 0.9.8d. | |
11265 | ||
11266 | *Bodo Moeller* | |
11267 | ||
11268 | ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006] | |
11269 | ||
11270 | * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher | |
11271 | ([CVE-2006-4339]) [Ben Laurie and Google Security Team] | |
11272 | ||
11273 | * Change the Unix randomness entropy gathering to use poll() when | |
11274 | possible instead of select(), since the latter has some | |
11275 | undesirable limitations. | |
11276 | ||
11277 | *Darryl Miles via Richard Levitte and Bodo Moeller* | |
11278 | ||
11279 | * Disable rogue ciphersuites: | |
11280 | ||
11281 | - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") | |
11282 | - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") | |
11283 | - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") | |
11284 | ||
11285 | The latter two were purportedly from | |
11286 | draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really | |
11287 | appear there. | |
11288 | ||
11289 | Also deactivate the remaining ciphersuites from | |
11290 | draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as | |
11291 | unofficial, and the ID has long expired. | |
11292 | ||
11293 | *Bodo Moeller* | |
11294 | ||
11295 | * Fix RSA blinding Heisenbug (problems sometimes occurred on | |
11296 | dual-core machines) and other potential thread-safety issues. | |
11297 | ||
11298 | *Bodo Moeller* | |
11299 | ||
11300 | ### Changes between 0.9.7i and 0.9.7j [04 May 2006] | |
11301 | ||
11302 | * Adapt fipsld and the build system to link against the validated FIPS | |
11303 | module in FIPS mode. | |
11304 | ||
11305 | *Steve Henson* | |
11306 | ||
11307 | * Fixes for VC++ 2005 build under Windows. | |
11308 | ||
11309 | *Steve Henson* | |
11310 | ||
11311 | * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make | |
11312 | from a Windows bash shell such as MSYS. It is autodetected from the | |
11313 | "config" script when run from a VC++ environment. Modify standard VC++ | |
11314 | build to use fipscanister.o from the GNU make build. | |
11315 | ||
11316 | *Steve Henson* | |
11317 | ||
11318 | ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005] | |
11319 | ||
11320 | * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. | |
11321 | The value now differs depending on if you build for FIPS or not. | |
11322 | BEWARE! A program linked with a shared FIPSed libcrypto can't be | |
11323 | safely run with a non-FIPSed libcrypto, as it may crash because of | |
11324 | the difference induced by this change. | |
11325 | ||
11326 | *Andy Polyakov* | |
11327 | ||
11328 | ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005] | |
11329 | ||
11330 | * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING | |
11331 | (part of SSL_OP_ALL). This option used to disable the | |
11332 | countermeasure against man-in-the-middle protocol-version | |
11333 | rollback in the SSL 2.0 server implementation, which is a bad | |
11334 | idea. ([CVE-2005-2969]) | |
11335 | ||
11336 | *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center | |
11337 | for Information Security, National Institute of Advanced Industrial | |
11338 | Science and Technology [AIST, Japan)]* | |
11339 | ||
11340 | * Minimal support for X9.31 signatures and PSS padding modes. This is | |
11341 | mainly for FIPS compliance and not fully integrated at this stage. | |
11342 | ||
11343 | *Steve Henson* | |
11344 | ||
11345 | * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform | |
11346 | the exponentiation using a fixed-length exponent. (Otherwise, | |
11347 | the information leaked through timing could expose the secret key | |
11348 | after many signatures; cf. Bleichenbacher's attack on DSA with | |
11349 | biased k.) | |
11350 | ||
11351 | *Bodo Moeller* | |
11352 | ||
11353 | * Make a new fixed-window mod_exp implementation the default for | |
11354 | RSA, DSA, and DH private-key operations so that the sequence of | |
11355 | squares and multiplies and the memory access pattern are | |
11356 | independent of the particular secret key. This will mitigate | |
11357 | cache-timing and potential related attacks. | |
11358 | ||
11359 | BN_mod_exp_mont_consttime() is the new exponentiation implementation, | |
11360 | and this is automatically used by BN_mod_exp_mont() if the new flag | |
11361 | BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH | |
11362 | will use this BN flag for private exponents unless the flag | |
11363 | RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or | |
11364 | DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. | |
11365 | ||
11366 | *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller* | |
11367 | ||
11368 | * Change the client implementation for SSLv23_method() and | |
11369 | SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 | |
11370 | Client Hello message format if the SSL_OP_NO_SSLv2 option is set. | |
11371 | (Previously, the SSL 2.0 backwards compatible Client Hello | |
11372 | message format would be used even with SSL_OP_NO_SSLv2.) | |
11373 | ||
11374 | *Bodo Moeller* | |
11375 | ||
11376 | * Add support for smime-type MIME parameter in S/MIME messages which some | |
11377 | clients need. | |
11378 | ||
11379 | *Steve Henson* | |
11380 | ||
11381 | * New function BN_MONT_CTX_set_locked() to set montgomery parameters in | |
11382 | a threadsafe manner. Modify rsa code to use new function and add calls | |
11383 | to dsa and dh code (which had race conditions before). | |
11384 | ||
11385 | *Steve Henson* | |
11386 | ||
11387 | * Include the fixed error library code in the C error file definitions | |
11388 | instead of fixing them up at runtime. This keeps the error code | |
11389 | structures constant. | |
11390 | ||
11391 | *Steve Henson* | |
11392 | ||
11393 | ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005] | |
11394 | ||
11395 | [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after | |
11396 | OpenSSL 0.9.8.] | |
11397 | ||
11398 | * Fixes for newer kerberos headers. NB: the casts are needed because | |
11399 | the 'length' field is signed on one version and unsigned on another | |
11400 | with no (?) obvious way to tell the difference, without these VC++ | |
11401 | complains. Also the "definition" of FAR (blank) is no longer included | |
11402 | nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up | |
11403 | some needed definitions. | |
11404 | ||
11405 | *Steve Henson* | |
11406 | ||
11407 | * Undo Cygwin change. | |
11408 | ||
11409 | *Ulf Möller* | |
11410 | ||
11411 | * Added support for proxy certificates according to RFC 3820. | |
11412 | Because they may be a security thread to unaware applications, | |
11413 | they must be explicitly allowed in run-time. See | |
11414 | docs/HOWTO/proxy_certificates.txt for further information. | |
11415 | ||
11416 | *Richard Levitte* | |
11417 | ||
11418 | ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005] | |
11419 | ||
11420 | * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating | |
11421 | server and client random values. Previously | |
11422 | (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in | |
11423 | less random data when sizeof(time_t) > 4 (some 64 bit platforms). | |
11424 | ||
11425 | This change has negligible security impact because: | |
11426 | ||
11427 | 1. Server and client random values still have 24 bytes of pseudo random | |
11428 | data. | |
11429 | ||
11430 | 2. Server and client random values are sent in the clear in the initial | |
11431 | handshake. | |
11432 | ||
11433 | 3. The master secret is derived using the premaster secret (48 bytes in | |
11434 | size for static RSA ciphersuites) as well as client server and random | |
11435 | values. | |
11436 | ||
11437 | The OpenSSL team would like to thank the UK NISCC for bringing this issue | |
11438 | to our attention. | |
11439 | ||
11440 | *Stephen Henson, reported by UK NISCC* | |
11441 | ||
11442 | * Use Windows randomness collection on Cygwin. | |
11443 | ||
11444 | *Ulf Möller* | |
11445 | ||
11446 | * Fix hang in EGD/PRNGD query when communication socket is closed | |
11447 | prematurely by EGD/PRNGD. | |
11448 | ||
11449 | *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014* | |
11450 | ||
11451 | * Prompt for pass phrases when appropriate for PKCS12 input format. | |
11452 | ||
11453 | *Steve Henson* | |
11454 | ||
11455 | * Back-port of selected performance improvements from development | |
11456 | branch, as well as improved support for PowerPC platforms. | |
11457 | ||
11458 | *Andy Polyakov* | |
11459 | ||
11460 | * Add lots of checks for memory allocation failure, error codes to indicate | |
11461 | failure and freeing up memory if a failure occurs. | |
11462 | ||
11463 | *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson* | |
11464 | ||
11465 | * Add new -passin argument to dgst. | |
11466 | ||
11467 | *Steve Henson* | |
11468 | ||
11469 | * Perform some character comparisons of different types in X509_NAME_cmp: | |
11470 | this is needed for some certificates that re-encode DNs into UTF8Strings | |
11471 | (in violation of RFC3280) and can't or won't issue name rollover | |
11472 | certificates. | |
11473 | ||
11474 | *Steve Henson* | |
11475 | ||
11476 | * Make an explicit check during certificate validation to see that | |
11477 | the CA setting in each certificate on the chain is correct. As a | |
11478 | side effect always do the following basic checks on extensions, | |
11479 | not just when there's an associated purpose to the check: | |
11480 | ||
11481 | - if there is an unhandled critical extension (unless the user | |
11482 | has chosen to ignore this fault) | |
11483 | - if the path length has been exceeded (if one is set at all) | |
11484 | - that certain extensions fit the associated purpose (if one has | |
11485 | been given) | |
11486 | ||
11487 | *Richard Levitte* | |
11488 | ||
11489 | ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004] | |
11490 | ||
11491 | * Avoid a race condition when CRLs are checked in a multi threaded | |
11492 | environment. This would happen due to the reordering of the revoked | |
11493 | entries during signature checking and serial number lookup. Now the | |
11494 | encoding is cached and the serial number sort performed under a lock. | |
11495 | Add new STACK function sk_is_sorted(). | |
11496 | ||
11497 | *Steve Henson* | |
11498 | ||
11499 | * Add Delta CRL to the extension code. | |
11500 | ||
11501 | *Steve Henson* | |
11502 | ||
11503 | * Various fixes to s3_pkt.c so alerts are sent properly. | |
11504 | ||
11505 | *David Holmes <d.holmes@f5.com>* | |
11506 | ||
11507 | * Reduce the chances of duplicate issuer name and serial numbers (in | |
11508 | violation of RFC3280) using the OpenSSL certificate creation utilities. | |
11509 | This is done by creating a random 64 bit value for the initial serial | |
11510 | number when a serial number file is created or when a self signed | |
11511 | certificate is created using 'openssl req -x509'. The initial serial | |
11512 | number file is created using 'openssl x509 -next_serial' in CA.pl | |
11513 | rather than being initialized to 1. | |
11514 | ||
11515 | *Steve Henson* | |
11516 | ||
11517 | ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004] | |
11518 | ||
11519 | * Fix null-pointer assignment in do_change_cipher_spec() revealed | |
11520 | by using the Codenomicon TLS Test Tool ([CVE-2004-0079]) | |
11521 | ||
11522 | *Joe Orton, Steve Henson* | |
11523 | ||
11524 | * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites | |
11525 | ([CVE-2004-0112]) | |
11526 | ||
11527 | *Joe Orton, Steve Henson* | |
11528 | ||
11529 | * Make it possible to have multiple active certificates with the same | |
11530 | subject in the CA index file. This is done only if the keyword | |
11531 | 'unique_subject' is set to 'no' in the main CA section (default | |
11532 | if 'CA_default') of the configuration file. The value is saved | |
11533 | with the database itself in a separate index attribute file, | |
11534 | named like the index file with '.attr' appended to the name. | |
11535 | ||
11536 | *Richard Levitte* | |
11537 | ||
11538 | * X509 verify fixes. Disable broken certificate workarounds when | |
11539 | X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if | |
11540 | keyUsage extension present. Don't accept CRLs with unhandled critical | |
11541 | extensions: since verify currently doesn't process CRL extensions this | |
11542 | rejects a CRL with *any* critical extensions. Add new verify error codes | |
11543 | for these cases. | |
11544 | ||
11545 | *Steve Henson* | |
11546 | ||
11547 | * When creating an OCSP nonce use an OCTET STRING inside the extnValue. | |
11548 | A clarification of RFC2560 will require the use of OCTET STRINGs and | |
11549 | some implementations cannot handle the current raw format. Since OpenSSL | |
11550 | copies and compares OCSP nonces as opaque blobs without any attempt at | |
11551 | parsing them this should not create any compatibility issues. | |
11552 | ||
11553 | *Steve Henson* | |
11554 | ||
11555 | * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when | |
11556 | calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without | |
11557 | this HMAC (and other) operations are several times slower than OpenSSL | |
11558 | < 0.9.7. | |
11559 | ||
11560 | *Steve Henson* | |
11561 | ||
11562 | * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). | |
11563 | ||
11564 | *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>* | |
11565 | ||
11566 | * Use the correct content when signing type "other". | |
11567 | ||
11568 | *Steve Henson* | |
11569 | ||
11570 | ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003] | |
11571 | ||
11572 | * Fix various bugs revealed by running the NISCC test suite: | |
11573 | ||
11574 | Stop out of bounds reads in the ASN1 code when presented with | |
11575 | invalid tags (CVE-2003-0543 and CVE-2003-0544). | |
11576 | ||
11577 | Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]). | |
11578 | ||
11579 | If verify callback ignores invalid public key errors don't try to check | |
11580 | certificate signature with the NULL public key. | |
11581 | ||
11582 | *Steve Henson* | |
11583 | ||
11584 | * New -ignore_err option in ocsp application to stop the server | |
11585 | exiting on the first error in a request. | |
11586 | ||
11587 | *Steve Henson* | |
11588 | ||
11589 | * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate | |
11590 | if the server requested one: as stated in TLS 1.0 and SSL 3.0 | |
11591 | specifications. | |
11592 | ||
11593 | *Steve Henson* | |
11594 | ||
11595 | * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional | |
11596 | extra data after the compression methods not only for TLS 1.0 | |
11597 | but also for SSL 3.0 (as required by the specification). | |
11598 | ||
11599 | *Bodo Moeller; problem pointed out by Matthias Loepfe* | |
11600 | ||
11601 | * Change X509_certificate_type() to mark the key as exported/exportable | |
11602 | when it's 512 *bits* long, not 512 bytes. | |
11603 | ||
11604 | *Richard Levitte* | |
11605 | ||
11606 | * Change AES_cbc_encrypt() so it outputs exact multiple of | |
11607 | blocks during encryption. | |
11608 | ||
11609 | *Richard Levitte* | |
11610 | ||
11611 | * Various fixes to base64 BIO and non blocking I/O. On write | |
11612 | flushes were not handled properly if the BIO retried. On read | |
11613 | data was not being buffered properly and had various logic bugs. | |
11614 | This also affects blocking I/O when the data being decoded is a | |
11615 | certain size. | |
11616 | ||
11617 | *Steve Henson* | |
11618 | ||
11619 | * Various S/MIME bugfixes and compatibility changes: | |
11620 | output correct application/pkcs7 MIME type if | |
11621 | PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. | |
11622 | Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening | |
11623 | of files as .eml work). Correctly handle very long lines in MIME | |
11624 | parser. | |
11625 | ||
11626 | *Steve Henson* | |
11627 | ||
11628 | ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003] | |
11629 | ||
11630 | * Countermeasure against the Klima-Pokorny-Rosa extension of | |
11631 | Bleichbacher's attack on PKCS #1 v1.5 padding: treat | |
11632 | a protocol version number mismatch like a decryption error | |
11633 | in ssl3_get_client_key_exchange (ssl/s3_srvr.c). | |
11634 | ||
11635 | *Bodo Moeller* | |
11636 | ||
11637 | * Turn on RSA blinding by default in the default implementation | |
11638 | to avoid a timing attack. Applications that don't want it can call | |
11639 | RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. | |
11640 | They would be ill-advised to do so in most cases. | |
11641 | ||
11642 | *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller* | |
11643 | ||
11644 | * Change RSA blinding code so that it works when the PRNG is not | |
11645 | seeded (in this case, the secret RSA exponent is abused as | |
11646 | an unpredictable seed -- if it is not unpredictable, there | |
11647 | is no point in blinding anyway). Make RSA blinding thread-safe | |
11648 | by remembering the creator's thread ID in rsa->blinding and | |
11649 | having all other threads use local one-time blinding factors | |
11650 | (this requires more computation than sharing rsa->blinding, but | |
11651 | avoids excessive locking; and if an RSA object is not shared | |
11652 | between threads, blinding will still be very fast). | |
11653 | ||
11654 | *Bodo Moeller* | |
11655 | ||
11656 | * Fixed a typo bug that would cause ENGINE_set_default() to set an | |
11657 | ENGINE as defaults for all supported algorithms irrespective of | |
11658 | the 'flags' parameter. 'flags' is now honoured, so applications | |
11659 | should make sure they are passing it correctly. | |
11660 | ||
11661 | *Geoff Thorpe* | |
11662 | ||
11663 | * Target "mingw" now allows native Windows code to be generated in | |
11664 | the Cygwin environment as well as with the MinGW compiler. | |
11665 | ||
11666 | *Ulf Moeller* | |
11667 | ||
11668 | ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003] | |
11669 | ||
11670 | * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked | |
11671 | via timing by performing a MAC computation even if incorrect | |
11672 | block cipher padding has been found. This is a countermeasure | |
11673 | against active attacks where the attacker has to distinguish | |
11674 | between bad padding and a MAC verification error. ([CVE-2003-0078]) | |
11675 | ||
11676 | *Bodo Moeller; problem pointed out by Brice Canvel (EPFL), | |
11677 | Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and | |
11678 | Martin Vuagnoux (EPFL, Ilion)* | |
11679 | ||
11680 | * Make the no-err option work as intended. The intention with no-err | |
11681 | is not to have the whole error stack handling routines removed from | |
11682 | libcrypto, it's only intended to remove all the function name and | |
11683 | reason texts, thereby removing some of the footprint that may not | |
11684 | be interesting if those errors aren't displayed anyway. | |
11685 | ||
11686 | NOTE: it's still possible for any application or module to have its | |
11687 | own set of error texts inserted. The routines are there, just not | |
11688 | used by default when no-err is given. | |
11689 | ||
11690 | *Richard Levitte* | |
11691 | ||
11692 | * Add support for FreeBSD on IA64. | |
11693 | ||
11694 | *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454* | |
11695 | ||
11696 | * Adjust DES_cbc_cksum() so it returns the same value as the MIT | |
11697 | Kerberos function mit_des_cbc_cksum(). Before this change, | |
11698 | the value returned by DES_cbc_cksum() was like the one from | |
11699 | mit_des_cbc_cksum(), except the bytes were swapped. | |
11700 | ||
11701 | *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte* | |
11702 | ||
11703 | * Allow an application to disable the automatic SSL chain building. | |
11704 | Before this a rather primitive chain build was always performed in | |
11705 | ssl3_output_cert_chain(): an application had no way to send the | |
11706 | correct chain if the automatic operation produced an incorrect result. | |
11707 | ||
11708 | Now the chain builder is disabled if either: | |
11709 | ||
11710 | 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). | |
11711 | ||
11712 | 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. | |
11713 | ||
11714 | The reasoning behind this is that an application would not want the | |
11715 | auto chain building to take place if extra chain certificates are | |
11716 | present and it might also want a means of sending no additional | |
11717 | certificates (for example the chain has two certificates and the | |
11718 | root is omitted). | |
11719 | ||
11720 | *Steve Henson* | |
11721 | ||
11722 | * Add the possibility to build without the ENGINE framework. | |
11723 | ||
11724 | *Steven Reddie <smr@essemer.com.au> via Richard Levitte* | |
11725 | ||
11726 | * Under Win32 gmtime() can return NULL: check return value in | |
11727 | OPENSSL_gmtime(). Add error code for case where gmtime() fails. | |
11728 | ||
11729 | *Steve Henson* | |
11730 | ||
11731 | * DSA routines: under certain error conditions uninitialized BN objects | |
11732 | could be freed. Solution: make sure initialization is performed early | |
11733 | enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, | |
11734 | Nils Larsch <nla@trustcenter.de> via PR#459) | |
11735 | ||
11736 | *Lutz Jaenicke* | |
11737 | ||
11738 | * Another fix for SSLv2 session ID handling: the session ID was incorrectly | |
11739 | checked on reconnect on the client side, therefore session resumption | |
11740 | could still fail with a "ssl session id is different" error. This | |
11741 | behaviour is masked when SSL_OP_ALL is used due to | |
11742 | SSL_OP_MICROSOFT_SESS_ID_BUG being set. | |
11743 | Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as | |
11744 | followup to PR #377. | |
11745 | ||
11746 | *Lutz Jaenicke* | |
11747 | ||
11748 | * IA-32 assembler support enhancements: unified ELF targets, support | |
11749 | for SCO/Caldera platforms, fix for Cygwin shared build. | |
11750 | ||
11751 | *Andy Polyakov* | |
11752 | ||
11753 | * Add support for FreeBSD on sparc64. As a consequence, support for | |
11754 | FreeBSD on non-x86 processors is separate from x86 processors on | |
11755 | the config script, much like the NetBSD support. | |
11756 | ||
11757 | *Richard Levitte & Kris Kennaway <kris@obsecurity.org>* | |
11758 | ||
11759 | ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002] | |
11760 | ||
11761 | [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after | |
11762 | OpenSSL 0.9.7.] | |
11763 | ||
11764 | * Fix session ID handling in SSLv2 client code: the SERVER FINISHED | |
11765 | code (06) was taken as the first octet of the session ID and the last | |
11766 | octet was ignored consequently. As a result SSLv2 client side session | |
11767 | caching could not have worked due to the session ID mismatch between | |
11768 | client and server. | |
11769 | Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as | |
11770 | PR #377. | |
11771 | ||
11772 | *Lutz Jaenicke* | |
11773 | ||
11774 | * Change the declaration of needed Kerberos libraries to use EX_LIBS | |
11775 | instead of the special (and badly supported) LIBKRB5. LIBKRB5 is | |
11776 | removed entirely. | |
11777 | ||
11778 | *Richard Levitte* | |
11779 | ||
11780 | * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it | |
11781 | seems that in spite of existing for more than a year, many application | |
11782 | author have done nothing to provide the necessary callbacks, which | |
11783 | means that this particular engine will not work properly anywhere. | |
11784 | This is a very unfortunate situation which forces us, in the name | |
11785 | of usability, to give the hw_ncipher.c a static lock, which is part | |
11786 | of libcrypto. | |
11787 | NOTE: This is for the 0.9.7 series ONLY. This hack will never | |
11788 | appear in 0.9.8 or later. We EXPECT application authors to have | |
11789 | dealt properly with this when 0.9.8 is released (unless we actually | |
11790 | make such changes in the libcrypto locking code that changes will | |
11791 | have to be made anyway). | |
11792 | ||
11793 | *Richard Levitte* | |
11794 | ||
11795 | * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content | |
11796 | octets have been read, EOF or an error occurs. Without this change | |
11797 | some truncated ASN1 structures will not produce an error. | |
11798 | ||
11799 | *Steve Henson* | |
11800 | ||
11801 | * Disable Heimdal support, since it hasn't been fully implemented. | |
11802 | Still give the possibility to force the use of Heimdal, but with | |
11803 | warnings and a request that patches get sent to openssl-dev. | |
11804 | ||
11805 | *Richard Levitte* | |
11806 | ||
11807 | * Add the VC-CE target, introduce the WINCE sysname, and add | |
11808 | INSTALL.WCE and appropriate conditionals to make it build. | |
11809 | ||
11810 | *Steven Reddie <smr@essemer.com.au> via Richard Levitte* | |
11811 | ||
11812 | * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and | |
11813 | cygssl-x.y.z.dll, where x, y and z are the major, minor and | |
11814 | edit numbers of the version. | |
11815 | ||
11816 | *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte* | |
11817 | ||
11818 | * Introduce safe string copy and catenation functions | |
11819 | (BUF_strlcpy() and BUF_strlcat()). | |
11820 | ||
11821 | *Ben Laurie (CHATS) and Richard Levitte* | |
11822 | ||
11823 | * Avoid using fixed-size buffers for one-line DNs. | |
11824 | ||
11825 | *Ben Laurie (CHATS)* | |
11826 | ||
11827 | * Add BUF_MEM_grow_clean() to avoid information leakage when | |
11828 | resizing buffers containing secrets, and use where appropriate. | |
11829 | ||
11830 | *Ben Laurie (CHATS)* | |
11831 | ||
11832 | * Avoid using fixed size buffers for configuration file location. | |
11833 | ||
11834 | *Ben Laurie (CHATS)* | |
11835 | ||
11836 | * Avoid filename truncation for various CA files. | |
11837 | ||
11838 | *Ben Laurie (CHATS)* | |
11839 | ||
11840 | * Use sizeof in preference to magic numbers. | |
11841 | ||
11842 | *Ben Laurie (CHATS)* | |
11843 | ||
11844 | * Avoid filename truncation in cert requests. | |
11845 | ||
11846 | *Ben Laurie (CHATS)* | |
11847 | ||
11848 | * Add assertions to check for (supposedly impossible) buffer | |
11849 | overflows. | |
11850 | ||
11851 | *Ben Laurie (CHATS)* | |
11852 | ||
11853 | * Don't cache truncated DNS entries in the local cache (this could | |
11854 | potentially lead to a spoofing attack). | |
11855 | ||
11856 | *Ben Laurie (CHATS)* | |
11857 | ||
11858 | * Fix various buffers to be large enough for hex/decimal | |
11859 | representations in a platform independent manner. | |
11860 | ||
11861 | *Ben Laurie (CHATS)* | |
11862 | ||
11863 | * Add CRYPTO_realloc_clean() to avoid information leakage when | |
11864 | resizing buffers containing secrets, and use where appropriate. | |
11865 | ||
11866 | *Ben Laurie (CHATS)* | |
11867 | ||
11868 | * Add BIO_indent() to avoid much slightly worrying code to do | |
11869 | indents. | |
11870 | ||
11871 | *Ben Laurie (CHATS)* | |
11872 | ||
11873 | * Convert sprintf()/BIO_puts() to BIO_printf(). | |
11874 | ||
11875 | *Ben Laurie (CHATS)* | |
11876 | ||
11877 | * buffer_gets() could terminate with the buffer only half | |
11878 | full. Fixed. | |
11879 | ||
11880 | *Ben Laurie (CHATS)* | |
11881 | ||
11882 | * Add assertions to prevent user-supplied crypto functions from | |
11883 | overflowing internal buffers by having large block sizes, etc. | |
11884 | ||
11885 | *Ben Laurie (CHATS)* | |
11886 | ||
11887 | * New OPENSSL_assert() macro (similar to assert(), but enabled | |
11888 | unconditionally). | |
11889 | ||
11890 | *Ben Laurie (CHATS)* | |
11891 | ||
11892 | * Eliminate unused copy of key in RC4. | |
11893 | ||
11894 | *Ben Laurie (CHATS)* | |
11895 | ||
11896 | * Eliminate unused and incorrectly sized buffers for IV in pem.h. | |
11897 | ||
11898 | *Ben Laurie (CHATS)* | |
11899 | ||
11900 | * Fix off-by-one error in EGD path. | |
11901 | ||
11902 | *Ben Laurie (CHATS)* | |
11903 | ||
11904 | * If RANDFILE path is too long, ignore instead of truncating. | |
11905 | ||
11906 | *Ben Laurie (CHATS)* | |
11907 | ||
11908 | * Eliminate unused and incorrectly sized X.509 structure | |
11909 | CBCParameter. | |
11910 | ||
11911 | *Ben Laurie (CHATS)* | |
11912 | ||
11913 | * Eliminate unused and dangerous function knumber(). | |
11914 | ||
11915 | *Ben Laurie (CHATS)* | |
11916 | ||
11917 | * Eliminate unused and dangerous structure, KSSL_ERR. | |
11918 | ||
11919 | *Ben Laurie (CHATS)* | |
11920 | ||
11921 | * Protect against overlong session ID context length in an encoded | |
11922 | session object. Since these are local, this does not appear to be | |
11923 | exploitable. | |
11924 | ||
11925 | *Ben Laurie (CHATS)* | |
11926 | ||
11927 | * Change from security patch (see 0.9.6e below) that did not affect | |
11928 | the 0.9.6 release series: | |
11929 | ||
11930 | Remote buffer overflow in SSL3 protocol - an attacker could | |
11931 | supply an oversized master key in Kerberos-enabled versions. | |
11932 | ([CVE-2002-0657]) | |
11933 | ||
11934 | *Ben Laurie (CHATS)* | |
11935 | ||
11936 | * Change the SSL kerb5 codes to match RFC 2712. | |
11937 | ||
11938 | *Richard Levitte* | |
11939 | ||
11940 | * Make -nameopt work fully for req and add -reqopt switch. | |
11941 | ||
11942 | *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson* | |
11943 | ||
11944 | * The "block size" for block ciphers in CFB and OFB mode should be 1. | |
11945 | ||
11946 | *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>* | |
11947 | ||
11948 | * Make sure tests can be performed even if the corresponding algorithms | |
11949 | have been removed entirely. This was also the last step to make | |
11950 | OpenSSL compilable with DJGPP under all reasonable conditions. | |
11951 | ||
11952 | *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>* | |
11953 | ||
11954 | * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT | |
11955 | to allow version independent disabling of normally unselected ciphers, | |
11956 | which may be activated as a side-effect of selecting a single cipher. | |
11957 | ||
11958 | (E.g., cipher list string "RSA" enables ciphersuites that are left | |
11959 | out of "ALL" because they do not provide symmetric encryption. | |
11960 | "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) | |
11961 | ||
11962 | *Lutz Jaenicke, Bodo Moeller* | |
11963 | ||
11964 | * Add appropriate support for separate platform-dependent build | |
11965 | directories. The recommended way to make a platform-dependent | |
11966 | build directory is the following (tested on Linux), maybe with | |
11967 | some local tweaks: | |
11968 | ||
11969 | # Place yourself outside of the OpenSSL source tree. In | |
11970 | # this example, the environment variable OPENSSL_SOURCE | |
11971 | # is assumed to contain the absolute OpenSSL source directory. | |
11972 | mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" | |
11973 | cd objtree/"`uname -s`-`uname -r`-`uname -m`" | |
11974 | (cd $OPENSSL_SOURCE; find . -type f) | while read F; do | |
11975 | mkdir -p `dirname $F` | |
11976 | ln -s $OPENSSL_SOURCE/$F $F | |
11977 | done | |
11978 | ||
11979 | To be absolutely sure not to disturb the source tree, a "make clean" | |
11980 | is a good thing. If it isn't successful, don't worry about it, | |
11981 | it probably means the source directory is very clean. | |
11982 | ||
11983 | *Richard Levitte* | |
11984 | ||
11985 | * Make sure any ENGINE control commands make local copies of string | |
11986 | pointers passed to them whenever necessary. Otherwise it is possible | |
11987 | the caller may have overwritten (or deallocated) the original string | |
11988 | data when a later ENGINE operation tries to use the stored values. | |
11989 | ||
11990 | *Götz Babin-Ebell <babinebell@trustcenter.de>* | |
11991 | ||
11992 | * Improve diagnostics in file reading and command-line digests. | |
11993 | ||
11994 | *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>* | |
11995 | ||
11996 | * Add AES modes CFB and OFB to the object database. Correct an | |
11997 | error in AES-CFB decryption. | |
11998 | ||
11999 | *Richard Levitte* | |
12000 | ||
12001 | * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this | |
12002 | allows existing EVP_CIPHER_CTX structures to be reused after | |
12003 | calling `EVP_*Final()`. This behaviour is used by encryption | |
12004 | BIOs and some applications. This has the side effect that | |
12005 | applications must explicitly clean up cipher contexts with | |
12006 | EVP_CIPHER_CTX_cleanup() or they will leak memory. | |
12007 | ||
12008 | *Steve Henson* | |
12009 | ||
12010 | * Check the values of dna and dnb in bn_mul_recursive before calling | |
12011 | bn_mul_comba (a non zero value means the a or b arrays do not contain | |
12012 | n2 elements) and fallback to bn_mul_normal if either is not zero. | |
12013 | ||
12014 | *Steve Henson* | |
12015 | ||
12016 | * Fix escaping of non-ASCII characters when using the -subj option | |
12017 | of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) | |
12018 | ||
12019 | *Lutz Jaenicke* | |
12020 | ||
12021 | * Make object definitions compliant to LDAP (RFC2256): SN is the short | |
12022 | form for "surname", serialNumber has no short form. | |
12023 | Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; | |
12024 | therefore remove "mail" short name for "internet 7". | |
12025 | The OID for unique identifiers in X509 certificates is | |
12026 | x500UniqueIdentifier, not uniqueIdentifier. | |
12027 | Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) | |
12028 | ||
12029 | *Lutz Jaenicke* | |
12030 | ||
12031 | * Add an "init" command to the ENGINE config module and auto initialize | |
12032 | ENGINEs. Without any "init" command the ENGINE will be initialized | |
12033 | after all ctrl commands have been executed on it. If init=1 the | |
12034 | ENGINE is initialized at that point (ctrls before that point are run | |
12035 | on the uninitialized ENGINE and after on the initialized one). If | |
12036 | init=0 then the ENGINE will not be initialized at all. | |
12037 | ||
12038 | *Steve Henson* | |
12039 | ||
12040 | * Fix the 'app_verify_callback' interface so that the user-defined | |
12041 | argument is actually passed to the callback: In the | |
12042 | SSL_CTX_set_cert_verify_callback() prototype, the callback | |
12043 | declaration has been changed from | |
12044 | int (*cb)() | |
12045 | into | |
12046 | int (*cb)(X509_STORE_CTX *,void *); | |
12047 | in ssl_verify_cert_chain (ssl/ssl_cert.c), the call | |
12048 | i=s->ctx->app_verify_callback(&ctx) | |
12049 | has been changed into | |
12050 | i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). | |
12051 | ||
12052 | To update applications using SSL_CTX_set_cert_verify_callback(), | |
12053 | a dummy argument can be added to their callback functions. | |
12054 | ||
12055 | *D. K. Smetters <smetters@parc.xerox.com>* | |
12056 | ||
12057 | * Added the '4758cca' ENGINE to support IBM 4758 cards. | |
12058 | ||
12059 | *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe* | |
12060 | ||
12061 | * Add and OPENSSL_LOAD_CONF define which will cause | |
12062 | OpenSSL_add_all_algorithms() to load the openssl.cnf config file. | |
12063 | This allows older applications to transparently support certain | |
12064 | OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. | |
12065 | Two new functions OPENSSL_add_all_algorithms_noconf() which will never | |
12066 | load the config file and OPENSSL_add_all_algorithms_conf() which will | |
12067 | always load it have also been added. | |
12068 | ||
12069 | *Steve Henson* | |
12070 | ||
12071 | * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. | |
12072 | Adjust NIDs and EVP layer. | |
12073 | ||
12074 | *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte* | |
12075 | ||
12076 | * Config modules support in openssl utility. | |
12077 | ||
12078 | Most commands now load modules from the config file, | |
12079 | though in a few (such as version) this isn't done | |
12080 | because it couldn't be used for anything. | |
12081 | ||
12082 | In the case of ca and req the config file used is | |
12083 | the same as the utility itself: that is the -config | |
12084 | command line option can be used to specify an | |
12085 | alternative file. | |
12086 | ||
12087 | *Steve Henson* | |
12088 | ||
12089 | * Move default behaviour from OPENSSL_config(). If appname is NULL | |
12090 | use "openssl_conf" if filename is NULL use default openssl config file. | |
12091 | ||
12092 | *Steve Henson* | |
12093 | ||
12094 | * Add an argument to OPENSSL_config() to allow the use of an alternative | |
12095 | config section name. Add a new flag to tolerate a missing config file | |
12096 | and move code to CONF_modules_load_file(). | |
12097 | ||
12098 | *Steve Henson* | |
12099 | ||
12100 | * Support for crypto accelerator cards from Accelerated Encryption | |
12101 | Processing, www.aep.ie. (Use engine 'aep') | |
12102 | The support was copied from 0.9.6c [engine] and adapted/corrected | |
12103 | to work with the new engine framework. | |
12104 | ||
12105 | *AEP Inc. and Richard Levitte* | |
12106 | ||
12107 | * Support for SureWare crypto accelerator cards from Baltimore | |
12108 | Technologies. (Use engine 'sureware') | |
12109 | The support was copied from 0.9.6c [engine] and adapted | |
12110 | to work with the new engine framework. | |
12111 | ||
12112 | *Richard Levitte* | |
12113 | ||
12114 | * Have the CHIL engine fork-safe (as defined by nCipher) and actually | |
12115 | make the newer ENGINE framework commands for the CHIL engine work. | |
12116 | ||
12117 | *Toomas Kiisk <vix@cyber.ee> and Richard Levitte* | |
12118 | ||
12119 | * Make it possible to produce shared libraries on ReliantUNIX. | |
12120 | ||
12121 | *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte* | |
12122 | ||
12123 | * Add the configuration target debug-linux-ppro. | |
12124 | Make 'openssl rsa' use the general key loading routines | |
12125 | implemented in `apps.c`, and make those routines able to | |
12126 | handle the key format FORMAT_NETSCAPE and the variant | |
12127 | FORMAT_IISSGC. | |
12128 | ||
12129 | *Toomas Kiisk <vix@cyber.ee> via Richard Levitte* | |
12130 | ||
12131 | * Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). | |
12132 | ||
12133 | *Toomas Kiisk <vix@cyber.ee> via Richard Levitte* | |
12134 | ||
12135 | * Add -keyform to rsautl, and document -engine. | |
12136 | ||
12137 | *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>* | |
12138 | ||
12139 | * Change BIO_new_file (crypto/bio/bss_file.c) to use new | |
12140 | BIO_R_NO_SUCH_FILE error code rather than the generic | |
12141 | ERR_R_SYS_LIB error code if fopen() fails with ENOENT. | |
12142 | ||
12143 | *Ben Laurie* | |
12144 | ||
12145 | * Add new functions | |
12146 | ERR_peek_last_error | |
12147 | ERR_peek_last_error_line | |
12148 | ERR_peek_last_error_line_data. | |
12149 | These are similar to | |
12150 | ERR_peek_error | |
12151 | ERR_peek_error_line | |
12152 | ERR_peek_error_line_data, | |
12153 | but report on the latest error recorded rather than the first one | |
12154 | still in the error queue. | |
12155 | ||
12156 | *Ben Laurie, Bodo Moeller* | |
12157 | ||
12158 | * default_algorithms option in ENGINE config module. This allows things | |
12159 | like: | |
12160 | default_algorithms = ALL | |
12161 | default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS | |
12162 | ||
12163 | *Steve Henson* | |
12164 | ||
12165 | * Preliminary ENGINE config module. | |
12166 | ||
12167 | *Steve Henson* | |
12168 | ||
12169 | * New experimental application configuration code. | |
12170 | ||
12171 | *Steve Henson* | |
12172 | ||
12173 | * Change the AES code to follow the same name structure as all other | |
12174 | symmetric ciphers, and behave the same way. Move everything to | |
12175 | the directory crypto/aes, thereby obsoleting crypto/rijndael. | |
12176 | ||
12177 | *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte* | |
12178 | ||
12179 | * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. | |
12180 | ||
12181 | *Ben Laurie and Theo de Raadt* | |
12182 | ||
12183 | * Add option to output public keys in req command. | |
12184 | ||
12185 | *Massimiliano Pala madwolf@openca.org* | |
12186 | ||
12187 | * Use wNAFs in EC_POINTs_mul() for improved efficiency | |
12188 | (up to about 10% better than before for P-192 and P-224). | |
12189 | ||
12190 | *Bodo Moeller* | |
12191 | ||
12192 | * New functions/macros | |
12193 | ||
12194 | SSL_CTX_set_msg_callback(ctx, cb) | |
12195 | SSL_CTX_set_msg_callback_arg(ctx, arg) | |
12196 | SSL_set_msg_callback(ssl, cb) | |
12197 | SSL_set_msg_callback_arg(ssl, arg) | |
12198 | ||
12199 | to request calling a callback function | |
12200 | ||
12201 | void cb(int write_p, int version, int content_type, | |
12202 | const void *buf, size_t len, SSL *ssl, void *arg) | |
12203 | ||
12204 | whenever a protocol message has been completely received | |
12205 | (write_p == 0) or sent (write_p == 1). Here 'version' is the | |
12206 | protocol version according to which the SSL library interprets | |
12207 | the current protocol message (SSL2_VERSION, SSL3_VERSION, or | |
12208 | TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or | |
12209 | the content type as defined in the SSL 3.0/TLS 1.0 protocol | |
12210 | specification (change_cipher_spec(20), alert(21), handshake(22)). | |
12211 | 'buf' and 'len' point to the actual message, 'ssl' to the | |
12212 | SSL object, and 'arg' is the application-defined value set by | |
12213 | SSL[_CTX]_set_msg_callback_arg(). | |
12214 | ||
12215 | 'openssl s_client' and 'openssl s_server' have new '-msg' options | |
12216 | to enable a callback that displays all protocol messages. | |
12217 | ||
12218 | *Bodo Moeller* | |
12219 | ||
12220 | * Change the shared library support so shared libraries are built as | |
12221 | soon as the corresponding static library is finished, and thereby get | |
12222 | openssl and the test programs linked against the shared library. | |
12223 | This still only happens when the keyword "shard" has been given to | |
12224 | the configuration scripts. | |
12225 | ||
12226 | NOTE: shared library support is still an experimental thing, and | |
12227 | backward binary compatibility is still not guaranteed. | |
12228 | ||
12229 | *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte* | |
12230 | ||
12231 | * Add support for Subject Information Access extension. | |
12232 | ||
12233 | *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>* | |
12234 | ||
12235 | * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero | |
12236 | additional bytes when new memory had to be allocated, not just | |
12237 | when reusing an existing buffer. | |
12238 | ||
12239 | *Bodo Moeller* | |
12240 | ||
12241 | * New command line and configuration option 'utf8' for the req command. | |
12242 | This allows field values to be specified as UTF8 strings. | |
12243 | ||
12244 | *Steve Henson* | |
12245 | ||
12246 | * Add -multi and -mr options to "openssl speed" - giving multiple parallel | |
12247 | runs for the former and machine-readable output for the latter. | |
12248 | ||
12249 | *Ben Laurie* | |
12250 | ||
12251 | * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion | |
12252 | of the e-mail address in the DN (i.e., it will go into a certificate | |
12253 | extension only). The new configuration file option 'email_in_dn = no' | |
12254 | has the same effect. | |
12255 | ||
12256 | *Massimiliano Pala madwolf@openca.org* | |
12257 | ||
12258 | * Change all functions with names starting with `des_` to be starting | |
12259 | with `DES_` instead. Add wrappers that are compatible with libdes, | |
12260 | but are named `_ossl_old_des_*`. Finally, add macros that map the | |
12261 | `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes | |
12262 | compatibility is desired. If OpenSSL 0.9.6c compatibility is | |
12263 | desired, the `des_*` symbols will be mapped to `DES_*`, with one | |
12264 | exception. | |
12265 | ||
12266 | Since we provide two compatibility mappings, the user needs to | |
12267 | define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes | |
12268 | compatibility is desired. The default (i.e., when that macro | |
12269 | isn't defined) is OpenSSL 0.9.6c compatibility. | |
12270 | ||
12271 | There are also macros that enable and disable the support of old | |
12272 | des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT | |
12273 | and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those | |
12274 | are defined, the default will apply: to support the old des routines. | |
12275 | ||
12276 | In either case, one must include openssl/des.h to get the correct | |
12277 | definitions. Do not try to just include openssl/des_old.h, that | |
12278 | won't work. | |
12279 | ||
12280 | NOTE: This is a major break of an old API into a new one. Software | |
12281 | authors are encouraged to switch to the `DES_` style functions. Some | |
12282 | time in the future, des_old.h and the libdes compatibility functions | |
12283 | will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the | |
12284 | default), and then completely removed. | |
12285 | ||
12286 | *Richard Levitte* | |
12287 | ||
12288 | * Test for certificates which contain unsupported critical extensions. | |
12289 | If such a certificate is found during a verify operation it is | |
12290 | rejected by default: this behaviour can be overridden by either | |
12291 | handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or | |
12292 | by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function | |
12293 | X509_supported_extension() has also been added which returns 1 if a | |
12294 | particular extension is supported. | |
12295 | ||
12296 | *Steve Henson* | |
12297 | ||
12298 | * Modify the behaviour of EVP cipher functions in similar way to digests | |
12299 | to retain compatibility with existing code. | |
12300 | ||
12301 | *Steve Henson* | |
12302 | ||
12303 | * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain | |
12304 | compatibility with existing code. In particular the 'ctx' parameter does | |
12305 | not have to be to be initialized before the call to EVP_DigestInit() and | |
12306 | it is tidied up after a call to EVP_DigestFinal(). New function | |
12307 | EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function | |
12308 | EVP_MD_CTX_copy() changed to not require the destination to be | |
12309 | initialized valid and new function EVP_MD_CTX_copy_ex() added which | |
12310 | requires the destination to be valid. | |
12311 | ||
12312 | Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), | |
12313 | EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). | |
12314 | ||
12315 | *Steve Henson* | |
12316 | ||
12317 | * Change ssl3_get_message (ssl/s3_both.c) and the functions using it | |
12318 | so that complete 'Handshake' protocol structures are kept in memory | |
12319 | instead of overwriting 'msg_type' and 'length' with 'body' data. | |
12320 | ||
12321 | *Bodo Moeller* | |
12322 | ||
12323 | * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. | |
12324 | ||
12325 | *Massimo Santin via Richard Levitte* | |
12326 | ||
12327 | * Major restructuring to the underlying ENGINE code. This includes | |
12328 | reduction of linker bloat, separation of pure "ENGINE" manipulation | |
12329 | (initialisation, etc) from functionality dealing with implementations | |
12330 | of specific crypto interfaces. This change also introduces integrated | |
12331 | support for symmetric ciphers and digest implementations - so ENGINEs | |
12332 | can now accelerate these by providing EVP_CIPHER and EVP_MD | |
12333 | implementations of their own. This is detailed in | |
12334 | [crypto/engine/README.md](crypto/engine/README.md) | |
12335 | as it couldn't be adequately described here. However, there are a few | |
12336 | API changes worth noting - some RSA, DSA, DH, and RAND functions that | |
12337 | were changed in the original introduction of ENGINE code have now | |
12338 | reverted back - the hooking from this code to ENGINE is now a good | |
12339 | deal more passive and at run-time, operations deal directly with | |
12340 | RSA_METHODs, DSA_METHODs (etc) as they did before, rather than | |
12341 | dereferencing through an ENGINE pointer any more. Also, the ENGINE | |
12342 | functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed - | |
12343 | they were not being used by the framework as there is no concept of a | |
12344 | BIGNUM_METHOD and they could not be generalised to the new | |
12345 | 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, | |
12346 | ENGINE_cpy() has been removed as it cannot be consistently defined in | |
12347 | the new code. | |
12348 | ||
12349 | *Geoff Thorpe* | |
12350 | ||
12351 | * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. | |
12352 | ||
12353 | *Steve Henson* | |
12354 | ||
12355 | * Change mkdef.pl to sort symbols that get the same entry number, | |
12356 | and make sure the automatically generated functions `ERR_load_*` | |
12357 | become part of libeay.num as well. | |
12358 | ||
12359 | *Richard Levitte* | |
12360 | ||
12361 | * New function SSL_renegotiate_pending(). This returns true once | |
12362 | renegotiation has been requested (either SSL_renegotiate() call | |
12363 | or HelloRequest/ClientHello received from the peer) and becomes | |
12364 | false once a handshake has been completed. | |
12365 | (For servers, SSL_renegotiate() followed by SSL_do_handshake() | |
12366 | sends a HelloRequest, but does not ensure that a handshake takes | |
12367 | place. SSL_renegotiate_pending() is useful for checking if the | |
12368 | client has followed the request.) | |
12369 | ||
12370 | *Bodo Moeller* | |
12371 | ||
12372 | * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. | |
12373 | By default, clients may request session resumption even during | |
12374 | renegotiation (if session ID contexts permit); with this option, | |
12375 | session resumption is possible only in the first handshake. | |
12376 | ||
12377 | SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes | |
12378 | more bits available for options that should not be part of | |
12379 | SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). | |
12380 | ||
12381 | *Bodo Moeller* | |
12382 | ||
12383 | * Add some demos for certificate and certificate request creation. | |
12384 | ||
12385 | *Steve Henson* | |
12386 | ||
12387 | * Make maximum certificate chain size accepted from the peer application | |
12388 | settable (`SSL*_get/set_max_cert_list()`), as proposed by | |
12389 | "Douglas E. Engert" <deengert@anl.gov>. | |
12390 | ||
12391 | *Lutz Jaenicke* | |
12392 | ||
12393 | * Add support for shared libraries for Unixware-7 | |
12394 | (Boyd Lynn Gerber <gerberb@zenez.com>). | |
12395 | ||
12396 | *Lutz Jaenicke* | |
12397 | ||
12398 | * Add a "destroy" handler to ENGINEs that allows structural cleanup to | |
12399 | be done prior to destruction. Use this to unload error strings from | |
12400 | ENGINEs that load their own error strings. NB: This adds two new API | |
12401 | functions to "get" and "set" this destroy handler in an ENGINE. | |
12402 | ||
12403 | *Geoff Thorpe* | |
12404 | ||
12405 | * Alter all existing ENGINE implementations (except "openssl" and | |
12406 | "openbsd") to dynamically instantiate their own error strings. This | |
12407 | makes them more flexible to be built both as statically-linked ENGINEs | |
12408 | and self-contained shared-libraries loadable via the "dynamic" ENGINE. | |
12409 | Also, add stub code to each that makes building them as self-contained | |
12410 | shared-libraries easier (see [README-Engine.md](README-Engine.md)). | |
12411 | ||
12412 | *Geoff Thorpe* | |
12413 | ||
12414 | * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE | |
12415 | implementations into applications that are completely implemented in | |
12416 | self-contained shared-libraries. The "dynamic" ENGINE exposes control | |
12417 | commands that can be used to configure what shared-library to load and | |
12418 | to control aspects of the way it is handled. Also, made an update to | |
12419 | the [README-Engine.md](README-Engine.md) file | |
12420 | that brings its information up-to-date and | |
12421 | provides some information and instructions on the "dynamic" ENGINE | |
12422 | (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). | |
12423 | ||
12424 | *Geoff Thorpe* | |
12425 | ||
12426 | * Make it possible to unload ranges of ERR strings with a new | |
12427 | "ERR_unload_strings" function. | |
12428 | ||
12429 | *Geoff Thorpe* | |
12430 | ||
12431 | * Add a copy() function to EVP_MD. | |
12432 | ||
12433 | *Ben Laurie* | |
12434 | ||
12435 | * Make EVP_MD routines take a context pointer instead of just the | |
12436 | md_data void pointer. | |
12437 | ||
12438 | *Ben Laurie* | |
12439 | ||
12440 | * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates | |
12441 | that the digest can only process a single chunk of data | |
12442 | (typically because it is provided by a piece of | |
12443 | hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application | |
12444 | is only going to provide a single chunk of data, and hence the | |
12445 | framework needn't accumulate the data for oneshot drivers. | |
12446 | ||
12447 | *Ben Laurie* | |
12448 | ||
12449 | * As with "ERR", make it possible to replace the underlying "ex_data" | |
12450 | functions. This change also alters the storage and management of global | |
12451 | ex_data state - it's now all inside ex_data.c and all "class" code (eg. | |
12452 | RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class | |
12453 | index counters. The API functions that use this state have been changed | |
12454 | to take a "class_index" rather than pointers to the class's local STACK | |
12455 | and counter, and there is now an API function to dynamically create new | |
12456 | classes. This centralisation allows us to (a) plug a lot of the | |
12457 | thread-safety problems that existed, and (b) makes it possible to clean | |
12458 | up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) | |
12459 | such data would previously have always leaked in application code and | |
12460 | workarounds were in place to make the memory debugging turn a blind eye | |
12461 | to it. Application code that doesn't use this new function will still | |
12462 | leak as before, but their memory debugging output will announce it now | |
12463 | rather than letting it slide. | |
12464 | ||
12465 | Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change | |
12466 | induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now | |
12467 | has a return value to indicate success or failure. | |
12468 | ||
12469 | *Geoff Thorpe* | |
12470 | ||
12471 | * Make it possible to replace the underlying "ERR" functions such that the | |
12472 | global state (2 LHASH tables and 2 locks) is only used by the "default" | |
12473 | implementation. This change also adds two functions to "get" and "set" | |
12474 | the implementation prior to it being automatically set the first time | |
12475 | any other ERR function takes place. Ie. an application can call "get", | |
12476 | pass the return value to a module it has just loaded, and that module | |
12477 | can call its own "set" function using that value. This means the | |
12478 | module's "ERR" operations will use (and modify) the error state in the | |
12479 | application and not in its own statically linked copy of OpenSSL code. | |
12480 | ||
12481 | *Geoff Thorpe* | |
12482 | ||
12483 | * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment | |
12484 | reference counts. This performs normal REF_PRINT/REF_CHECK macros on | |
12485 | the operation, and provides a more encapsulated way for external code | |
12486 | (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code | |
12487 | to use these functions rather than manually incrementing the counts. | |
12488 | ||
12489 | Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". | |
12490 | ||
12491 | *Geoff Thorpe* | |
12492 | ||
12493 | * Add EVP test program. | |
12494 | ||
12495 | *Ben Laurie* | |
12496 | ||
12497 | * Add symmetric cipher support to ENGINE. Expect the API to change! | |
12498 | ||
12499 | *Ben Laurie* | |
12500 | ||
12501 | * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() | |
12502 | X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), | |
12503 | X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). | |
12504 | These allow a CRL to be built without having to access X509_CRL fields | |
12505 | directly. Modify 'ca' application to use new functions. | |
12506 | ||
12507 | *Steve Henson* | |
12508 | ||
12509 | * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended | |
12510 | bug workarounds. Rollback attack detection is a security feature. | |
12511 | The problem will only arise on OpenSSL servers when TLSv1 is not | |
12512 | available (sslv3_server_method() or SSL_OP_NO_TLSv1). | |
12513 | Software authors not wanting to support TLSv1 will have special reasons | |
12514 | for their choice and can explicitly enable this option. | |
12515 | ||
12516 | *Bodo Moeller, Lutz Jaenicke* | |
12517 | ||
12518 | * Rationalise EVP so it can be extended: don't include a union of | |
12519 | cipher/digest structures, add init/cleanup functions for EVP_MD_CTX | |
12520 | (similar to those existing for EVP_CIPHER_CTX). | |
12521 | Usage example: | |
12522 | ||
12523 | EVP_MD_CTX md; | |
12524 | ||
12525 | EVP_MD_CTX_init(&md); /* new function call */ | |
12526 | EVP_DigestInit(&md, EVP_sha1()); | |
12527 | EVP_DigestUpdate(&md, in, len); | |
12528 | EVP_DigestFinal(&md, out, NULL); | |
12529 | EVP_MD_CTX_cleanup(&md); /* new function call */ | |
12530 | ||
12531 | *Ben Laurie* | |
12532 | ||
12533 | * Make DES key schedule conform to the usual scheme, as well as | |
12534 | correcting its structure. This means that calls to DES functions | |
12535 | now have to pass a pointer to a des_key_schedule instead of a | |
12536 | plain des_key_schedule (which was actually always a pointer | |
12537 | anyway): E.g., | |
12538 | ||
12539 | des_key_schedule ks; | |
12540 | ||
12541 | des_set_key_checked(..., &ks); | |
12542 | des_ncbc_encrypt(..., &ks, ...); | |
12543 | ||
12544 | (Note that a later change renames 'des_...' into 'DES_...'.) | |
12545 | ||
12546 | *Ben Laurie* | |
12547 | ||
12548 | * Initial reduction of linker bloat: the use of some functions, such as | |
12549 | PEM causes large amounts of unused functions to be linked in due to | |
12550 | poor organisation. For example pem_all.c contains every PEM function | |
12551 | which has a knock on effect of linking in large amounts of (unused) | |
12552 | ASN1 code. Grouping together similar functions and splitting unrelated | |
12553 | functions prevents this. | |
12554 | ||
12555 | *Steve Henson* | |
12556 | ||
12557 | * Cleanup of EVP macros. | |
12558 | ||
12559 | *Ben Laurie* | |
12560 | ||
12561 | * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the | |
12562 | correct `_ecb suffix`. | |
12563 | ||
12564 | *Ben Laurie* | |
12565 | ||
12566 | * Add initial OCSP responder support to ocsp application. The | |
12567 | revocation information is handled using the text based index | |
12568 | use by the ca application. The responder can either handle | |
12569 | requests generated internally, supplied in files (for example | |
12570 | via a CGI script) or using an internal minimal server. | |
12571 | ||
12572 | *Steve Henson* | |
12573 | ||
12574 | * Add configuration choices to get zlib compression for TLS. | |
12575 | ||
12576 | *Richard Levitte* | |
12577 | ||
12578 | * Changes to Kerberos SSL for RFC 2712 compliance: | |
12579 | 1. Implemented real KerberosWrapper, instead of just using | |
12580 | KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] | |
12581 | 2. Implemented optional authenticator field of KerberosWrapper. | |
12582 | ||
12583 | Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, | |
12584 | and authenticator structs; see crypto/krb5/. | |
12585 | ||
12586 | Generalized Kerberos calls to support multiple Kerberos libraries. | |
12587 | *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu> | |
12588 | via Richard Levitte* | |
12589 | ||
12590 | * Cause 'openssl speed' to use fully hard-coded DSA keys as it | |
12591 | already does with RSA. testdsa.h now has 'priv_key/pub_key' | |
12592 | values for each of the key sizes rather than having just | |
12593 | parameters (and 'speed' generating keys each time). | |
12594 | ||
12595 | *Geoff Thorpe* | |
12596 | ||
12597 | * Speed up EVP routines. | |
12598 | Before: | |
12599 | crypt | |
12600 | pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes | |
12601 | s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k | |
12602 | s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k | |
12603 | s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k | |
12604 | crypt | |
12605 | s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k | |
12606 | s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k | |
12607 | s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k | |
12608 | After: | |
12609 | crypt | |
12610 | s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k | |
12611 | crypt | |
12612 | s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k | |
12613 | ||
12614 | *Ben Laurie* | |
12615 | ||
12616 | * Added the OS2-EMX target. | |
12617 | ||
12618 | *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte* | |
12619 | ||
12620 | * Rewrite commands to use `NCONF` routines instead of the old `CONF`. | |
12621 | New functions to support `NCONF `routines in extension code. | |
12622 | New function `CONF_set_nconf()` | |
12623 | to allow functions which take an `NCONF` to also handle the old `LHASH` | |
12624 | structure: this means that the old `CONF` compatible routines can be | |
12625 | retained (in particular w.rt. extensions) without having to duplicate the | |
12626 | code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack. | |
12627 | ||
12628 | *Steve Henson* | |
12629 | ||
12630 | * Enhance the general user interface with mechanisms for inner control | |
12631 | and with possibilities to have yes/no kind of prompts. | |
12632 | ||
12633 | *Richard Levitte* | |
12634 | ||
12635 | * Change all calls to low level digest routines in the library and | |
12636 | applications to use EVP. Add missing calls to HMAC_cleanup() and | |
12637 | don't assume HMAC_CTX can be copied using memcpy(). | |
12638 | ||
12639 | *Verdon Walker <VWalker@novell.com>, Steve Henson* | |
12640 | ||
12641 | * Add the possibility to control engines through control names but with | |
12642 | arbitrary arguments instead of just a string. | |
12643 | Change the key loaders to take a UI_METHOD instead of a callback | |
12644 | function pointer. NOTE: this breaks binary compatibility with earlier | |
12645 | versions of OpenSSL [engine]. | |
12646 | Adapt the nCipher code for these new conditions and add a card insertion | |
12647 | callback. | |
12648 | ||
12649 | *Richard Levitte* | |
12650 | ||
12651 | * Enhance the general user interface with mechanisms to better support | |
12652 | dialog box interfaces, application-defined prompts, the possibility | |
12653 | to use defaults (for example default passwords from somewhere else) | |
12654 | and interrupts/cancellations. | |
12655 | ||
12656 | *Richard Levitte* | |
12657 | ||
12658 | * Tidy up PKCS#12 attribute handling. Add support for the CSP name | |
12659 | attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. | |
12660 | ||
12661 | *Steve Henson* | |
12662 | ||
12663 | * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also | |
12664 | tidy up some unnecessarily weird code in 'sk_new()'). | |
12665 | ||
12666 | *Geoff, reported by Diego Tartara <dtartara@novamens.com>* | |
12667 | ||
12668 | * Change the key loading routines for ENGINEs to use the same kind | |
12669 | callback (pem_password_cb) as all other routines that need this | |
12670 | kind of callback. | |
12671 | ||
12672 | *Richard Levitte* | |
12673 | ||
12674 | * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with | |
12675 | 256 bit (=32 byte) keys. Of course seeding with more entropy bytes | |
12676 | than this minimum value is recommended. | |
12677 | ||
12678 | *Lutz Jaenicke* | |
12679 | ||
12680 | * New random seeder for OpenVMS, using the system process statistics | |
12681 | that are easily reachable. | |
12682 | ||
12683 | *Richard Levitte* | |
12684 | ||
12685 | * Windows apparently can't transparently handle global | |
12686 | variables defined in DLLs. Initialisations such as: | |
12687 | ||
12688 | const ASN1_ITEM *it = &ASN1_INTEGER_it; | |
12689 | ||
12690 | won't compile. This is used by the any applications that need to | |
12691 | declare their own ASN1 modules. This was fixed by adding the option | |
12692 | EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly | |
12693 | needed for static libraries under Win32. | |
12694 | ||
12695 | *Steve Henson* | |
12696 | ||
12697 | * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle | |
12698 | setting of purpose and trust fields. New X509_STORE trust and | |
12699 | purpose functions and tidy up setting in other SSL functions. | |
12700 | ||
12701 | *Steve Henson* | |
12702 | ||
12703 | * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE | |
12704 | structure. These are inherited by X509_STORE_CTX when it is | |
12705 | initialised. This allows various defaults to be set in the | |
12706 | X509_STORE structure (such as flags for CRL checking and custom | |
12707 | purpose or trust settings) for functions which only use X509_STORE_CTX | |
12708 | internally such as S/MIME. | |
12709 | ||
12710 | Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and | |
12711 | trust settings if they are not set in X509_STORE. This allows X509_STORE | |
12712 | purposes and trust (in S/MIME for example) to override any set by default. | |
12713 | ||
12714 | Add command line options for CRL checking to smime, s_client and s_server | |
12715 | applications. | |
12716 | ||
12717 | *Steve Henson* | |
12718 | ||
12719 | * Initial CRL based revocation checking. If the CRL checking flag(s) | |
12720 | are set then the CRL is looked up in the X509_STORE structure and | |
12721 | its validity and signature checked, then if the certificate is found | |
12722 | in the CRL the verify fails with a revoked error. | |
12723 | ||
12724 | Various new CRL related callbacks added to X509_STORE_CTX structure. | |
12725 | ||
12726 | Command line options added to 'verify' application to support this. | |
12727 | ||
12728 | This needs some additional work, such as being able to handle multiple | |
12729 | CRLs with different times, extension based lookup (rather than just | |
12730 | by subject name) and ultimately more complete V2 CRL extension | |
12731 | handling. | |
12732 | ||
12733 | *Steve Henson* | |
12734 | ||
12735 | * Add a general user interface API (crypto/ui/). This is designed | |
12736 | to replace things like des_read_password and friends (backward | |
12737 | compatibility functions using this new API are provided). | |
12738 | The purpose is to remove prompting functions from the DES code | |
12739 | section as well as provide for prompting through dialog boxes in | |
12740 | a window system and the like. | |
12741 | ||
12742 | *Richard Levitte* | |
12743 | ||
12744 | * Add "ex_data" support to ENGINE so implementations can add state at a | |
12745 | per-structure level rather than having to store it globally. | |
12746 | ||
12747 | *Geoff* | |
12748 | ||
12749 | * Make it possible for ENGINE structures to be copied when retrieved by | |
12750 | ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. | |
12751 | This causes the "original" ENGINE structure to act like a template, | |
12752 | analogous to the RSA vs. RSA_METHOD type of separation. Because of this | |
12753 | operational state can be localised to each ENGINE structure, despite the | |
12754 | fact they all share the same "methods". New ENGINE structures returned in | |
12755 | this case have no functional references and the return value is the single | |
12756 | structural reference. This matches the single structural reference returned | |
12757 | by ENGINE_by_id() normally, when it is incremented on the pre-existing | |
12758 | ENGINE structure. | |
12759 | ||
12760 | *Geoff* | |
12761 | ||
12762 | * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this | |
12763 | needs to match any other type at all we need to manually clear the | |
12764 | tag cache. | |
12765 | ||
12766 | *Steve Henson* | |
12767 | ||
12768 | * Changes to the "openssl engine" utility to include; | |
12769 | - verbosity levels ('-v', '-vv', and '-vvv') that provide information | |
12770 | about an ENGINE's available control commands. | |
12771 | - executing control commands from command line arguments using the | |
12772 | '-pre' and '-post' switches. '-post' is only used if '-t' is | |
12773 | specified and the ENGINE is successfully initialised. The syntax for | |
12774 | the individual commands are colon-separated, for example; | |
12775 | openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so | |
12776 | ||
12777 | *Geoff* | |
12778 | ||
12779 | * New dynamic control command support for ENGINEs. ENGINEs can now | |
12780 | declare their own commands (numbers), names (strings), descriptions, | |
12781 | and input types for run-time discovery by calling applications. A | |
12782 | subset of these commands are implicitly classed as "executable" | |
12783 | depending on their input type, and only these can be invoked through | |
12784 | the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this | |
12785 | can be based on user input, config files, etc). The distinction is | |
12786 | that "executable" commands cannot return anything other than a boolean | |
12787 | result and can only support numeric or string input, whereas some | |
12788 | discoverable commands may only be for direct use through | |
12789 | ENGINE_ctrl(), eg. supporting the exchange of binary data, function | |
12790 | pointers, or other custom uses. The "executable" commands are to | |
12791 | support parameterisations of ENGINE behaviour that can be | |
12792 | unambiguously defined by ENGINEs and used consistently across any | |
12793 | OpenSSL-based application. Commands have been added to all the | |
12794 | existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow | |
12795 | control over shared-library paths without source code alterations. | |
12796 | ||
12797 | *Geoff* | |
12798 | ||
12799 | * Changed all ENGINE implementations to dynamically allocate their | |
12800 | ENGINEs rather than declaring them statically. Apart from this being | |
12801 | necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, | |
12802 | this also allows the implementations to compile without using the | |
12803 | internal engine_int.h header. | |
12804 | ||
12805 | *Geoff* | |
12806 | ||
12807 | * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a | |
12808 | 'const' value. Any code that should be able to modify a RAND_METHOD | |
12809 | should already have non-const pointers to it (ie. they should only | |
12810 | modify their own ones). | |
12811 | ||
12812 | *Geoff* | |
12813 | ||
12814 | * Made a variety of little tweaks to the ENGINE code. | |
12815 | - "atalla" and "ubsec" string definitions were moved from header files | |
12816 | to C code. "nuron" string definitions were placed in variables | |
12817 | rather than hard-coded - allowing parameterisation of these values | |
12818 | later on via ctrl() commands. | |
12819 | - Removed unused "#if 0"'d code. | |
12820 | - Fixed engine list iteration code so it uses ENGINE_free() to release | |
12821 | structural references. | |
12822 | - Constified the RAND_METHOD element of ENGINE structures. | |
12823 | - Constified various get/set functions as appropriate and added | |
12824 | missing functions (including a catch-all ENGINE_cpy that duplicates | |
12825 | all ENGINE values onto a new ENGINE except reference counts/state). | |
12826 | - Removed NULL parameter checks in get/set functions. Setting a method | |
12827 | or function to NULL is a way of cancelling out a previously set | |
12828 | value. Passing a NULL ENGINE parameter is just plain stupid anyway | |
12829 | and doesn't justify the extra error symbols and code. | |
12830 | - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for | |
12831 | flags from engine_int.h to engine.h. | |
12832 | - Changed prototypes for ENGINE handler functions (init(), finish(), | |
12833 | ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. | |
12834 | ||
12835 | *Geoff* | |
12836 | ||
12837 | * Implement binary inversion algorithm for BN_mod_inverse in addition | |
12838 | to the algorithm using long division. The binary algorithm can be | |
12839 | used only if the modulus is odd. On 32-bit systems, it is faster | |
12840 | only for relatively small moduli (roughly 20-30% for 128-bit moduli, | |
12841 | roughly 5-15% for 256-bit moduli), so we use it only for moduli | |
12842 | up to 450 bits. In 64-bit environments, the binary algorithm | |
12843 | appears to be advantageous for much longer moduli; here we use it | |
12844 | for moduli up to 2048 bits. | |
12845 | ||
12846 | *Bodo Moeller* | |
12847 | ||
12848 | * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code | |
12849 | could not support the combine flag in choice fields. | |
12850 | ||
12851 | *Steve Henson* | |
12852 | ||
12853 | * Add a 'copy_extensions' option to the 'ca' utility. This copies | |
12854 | extensions from a certificate request to the certificate. | |
12855 | ||
12856 | *Steve Henson* | |
12857 | ||
12858 | * Allow multiple 'certopt' and 'nameopt' options to be separated | |
12859 | by commas. Add 'namopt' and 'certopt' options to the 'ca' config | |
12860 | file: this allows the display of the certificate about to be | |
12861 | signed to be customised, to allow certain fields to be included | |
12862 | or excluded and extension details. The old system didn't display | |
12863 | multicharacter strings properly, omitted fields not in the policy | |
12864 | and couldn't display additional details such as extensions. | |
12865 | ||
12866 | *Steve Henson* | |
12867 | ||
12868 | * Function EC_POINTs_mul for multiple scalar multiplication | |
12869 | of an arbitrary number of elliptic curve points | |
12870 | \sum scalars[i]*points[i], | |
12871 | optionally including the generator defined for the EC_GROUP: | |
12872 | scalar*generator + \sum scalars[i]*points[i]. | |
12873 | ||
12874 | EC_POINT_mul is a simple wrapper function for the typical case | |
12875 | that the point list has just one item (besides the optional | |
12876 | generator). | |
12877 | ||
12878 | *Bodo Moeller* | |
12879 | ||
12880 | * First EC_METHODs for curves over GF(p): | |
12881 | ||
12882 | EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr | |
12883 | operations and provides various method functions that can also | |
12884 | operate with faster implementations of modular arithmetic. | |
12885 | ||
12886 | EC_GFp_mont_method() reuses most functions that are part of | |
12887 | EC_GFp_simple_method, but uses Montgomery arithmetic. | |
12888 | ||
12889 | *Bodo Moeller; point addition and point doubling | |
12890 | implementation directly derived from source code provided by | |
12891 | Lenka Fibikova <fibikova@exp-math.uni-essen.de>* | |
12892 | ||
12893 | * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, | |
12894 | crypto/ec/ec_lib.c): | |
12895 | ||
12896 | Curves are EC_GROUP objects (with an optional group generator) | |
12897 | based on EC_METHODs that are built into the library. | |
12898 | ||
12899 | Points are EC_POINT objects based on EC_GROUP objects. | |
12900 | ||
12901 | Most of the framework would be able to handle curves over arbitrary | |
12902 | finite fields, but as there are no obvious types for fields other | |
12903 | than GF(p), some functions are limited to that for now. | |
12904 | ||
12905 | *Bodo Moeller* | |
12906 | ||
12907 | * Add the -HTTP option to s_server. It is similar to -WWW, but requires | |
12908 | that the file contains a complete HTTP response. | |
12909 | ||
12910 | *Richard Levitte* | |
12911 | ||
12912 | * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl | |
12913 | change the def and num file printf format specifier from "%-40sXXX" | |
12914 | to "%-39s XXX". The latter will always guarantee a space after the | |
12915 | field while the former will cause them to run together if the field | |
12916 | is 40 of more characters long. | |
12917 | ||
12918 | *Steve Henson* | |
12919 | ||
12920 | * Constify the cipher and digest 'method' functions and structures | |
12921 | and modify related functions to take constant EVP_MD and EVP_CIPHER | |
12922 | pointers. | |
12923 | ||
12924 | *Steve Henson* | |
12925 | ||
12926 | * Hide BN_CTX structure details in bn_lcl.h instead of publishing them | |
12927 | in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. | |
12928 | ||
12929 | *Bodo Moeller* | |
12930 | ||
12931 | * Modify `EVP_Digest*()` routines so they now return values. Although the | |
12932 | internal software routines can never fail additional hardware versions | |
12933 | might. | |
12934 | ||
12935 | *Steve Henson* | |
12936 | ||
12937 | * Clean up crypto/err/err.h and change some error codes to avoid conflicts: | |
12938 | ||
12939 | Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 | |
12940 | (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. | |
12941 | ||
12942 | ASN1 error codes | |
12943 | ERR_R_NESTED_ASN1_ERROR | |
12944 | ... | |
12945 | ERR_R_MISSING_ASN1_EOS | |
12946 | were 4 .. 9, conflicting with | |
12947 | ERR_LIB_RSA (= ERR_R_RSA_LIB) | |
12948 | ... | |
12949 | ERR_LIB_PEM (= ERR_R_PEM_LIB). | |
12950 | They are now 58 .. 63 (i.e., just below ERR_R_FATAL). | |
12951 | ||
12952 | Add new error code 'ERR_R_INTERNAL_ERROR'. | |
12953 | ||
12954 | *Bodo Moeller* | |
12955 | ||
12956 | * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock | |
12957 | suffices. | |
12958 | ||
12959 | *Bodo Moeller* | |
12960 | ||
12961 | * New option '-subj arg' for 'openssl req' and 'openssl ca'. This | |
12962 | sets the subject name for a new request or supersedes the | |
12963 | subject name in a given request. Formats that can be parsed are | |
12964 | 'CN=Some Name, OU=myOU, C=IT' | |
12965 | and | |
12966 | 'CN=Some Name/OU=myOU/C=IT'. | |
12967 | ||
12968 | Add options '-batch' and '-verbose' to 'openssl req'. | |
12969 | ||
12970 | *Massimiliano Pala <madwolf@hackmasters.net>* | |
12971 | ||
12972 | * Introduce the possibility to access global variables through | |
12973 | functions on platform were that's the best way to handle exporting | |
12974 | global variables in shared libraries. To enable this functionality, | |
12975 | one must configure with "EXPORT_VAR_AS_FN" or defined the C macro | |
12976 | "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter | |
12977 | is normally done by Configure or something similar). | |
12978 | ||
12979 | To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL | |
12980 | in the source file (foo.c) like this: | |
12981 | ||
12982 | OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; | |
12983 | OPENSSL_IMPLEMENT_GLOBAL(double,bar); | |
12984 | ||
12985 | To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL | |
12986 | and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: | |
12987 | ||
12988 | OPENSSL_DECLARE_GLOBAL(int,foo); | |
12989 | #define foo OPENSSL_GLOBAL_REF(foo) | |
12990 | OPENSSL_DECLARE_GLOBAL(double,bar); | |
12991 | #define bar OPENSSL_GLOBAL_REF(bar) | |
12992 | ||
12993 | The #defines are very important, and therefore so is including the | |
12994 | header file everywhere where the defined globals are used. | |
12995 | ||
12996 | The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition | |
12997 | of ASN.1 items, but that structure is a bit different. | |
12998 | ||
12999 | The largest change is in util/mkdef.pl which has been enhanced with | |
13000 | better and easier to understand logic to choose which symbols should | |
13001 | go into the Windows .def files as well as a number of fixes and code | |
13002 | cleanup (among others, algorithm keywords are now sorted | |
13003 | lexicographically to avoid constant rewrites). | |
13004 | ||
13005 | *Richard Levitte* | |
13006 | ||
13007 | * In BN_div() keep a copy of the sign of 'num' before writing the | |
13008 | result to 'rm' because if rm==num the value will be overwritten | |
13009 | and produce the wrong result if 'num' is negative: this caused | |
13010 | problems with BN_mod() and BN_nnmod(). | |
13011 | ||
13012 | *Steve Henson* | |
13013 | ||
13014 | * Function OCSP_request_verify(). This checks the signature on an | |
13015 | OCSP request and verifies the signer certificate. The signer | |
13016 | certificate is just checked for a generic purpose and OCSP request | |
13017 | trust settings. | |
13018 | ||
13019 | *Steve Henson* | |
13020 | ||
13021 | * Add OCSP_check_validity() function to check the validity of OCSP | |
13022 | responses. OCSP responses are prepared in real time and may only | |
13023 | be a few seconds old. Simply checking that the current time lies | |
13024 | between thisUpdate and nextUpdate max reject otherwise valid responses | |
13025 | caused by either OCSP responder or client clock inaccuracy. Instead | |
13026 | we allow thisUpdate and nextUpdate to fall within a certain period of | |
13027 | the current time. The age of the response can also optionally be | |
13028 | checked. Two new options -validity_period and -status_age added to | |
13029 | ocsp utility. | |
13030 | ||
13031 | *Steve Henson* | |
13032 | ||
13033 | * If signature or public key algorithm is unrecognized print out its | |
13034 | OID rather that just UNKNOWN. | |
13035 | ||
13036 | *Steve Henson* | |
13037 | ||
13038 | * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and | |
13039 | OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate | |
13040 | ID to be generated from the issuer certificate alone which can then be | |
13041 | passed to OCSP_id_issuer_cmp(). | |
13042 | ||
13043 | *Steve Henson* | |
13044 | ||
13045 | * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new | |
13046 | ASN1 modules to export functions returning ASN1_ITEM pointers | |
13047 | instead of the ASN1_ITEM structures themselves. This adds several | |
13048 | new macros which allow the underlying ASN1 function/structure to | |
13049 | be accessed transparently. As a result code should not use ASN1_ITEM | |
13050 | references directly (such as &X509_it) but instead use the relevant | |
13051 | macros (such as ASN1_ITEM_rptr(X509)). This option is to allow | |
13052 | use of the new ASN1 code on platforms where exporting structures | |
13053 | is problematical (for example in shared libraries) but exporting | |
13054 | functions returning pointers to structures is not. | |
13055 | ||
13056 | *Steve Henson* | |
13057 | ||
13058 | * Add support for overriding the generation of SSL/TLS session IDs. | |
13059 | These callbacks can be registered either in an SSL_CTX or per SSL. | |
13060 | The purpose of this is to allow applications to control, if they wish, | |
13061 | the arbitrary values chosen for use as session IDs, particularly as it | |
13062 | can be useful for session caching in multiple-server environments. A | |
13063 | command-line switch for testing this (and any client code that wishes | |
13064 | to use such a feature) has been added to "s_server". | |
13065 | ||
13066 | *Geoff Thorpe, Lutz Jaenicke* | |
13067 | ||
13068 | * Modify mkdef.pl to recognise and parse preprocessor conditionals | |
13069 | of the form `#if defined(...) || defined(...) || ...` and | |
13070 | `#if !defined(...) && !defined(...) && ...`. This also avoids | |
13071 | the growing number of special cases it was previously handling. | |
13072 | ||
13073 | *Richard Levitte* | |
13074 | ||
13075 | * Make all configuration macros available for application by making | |
13076 | sure they are available in opensslconf.h, by giving them names starting | |
13077 | with `OPENSSL_` to avoid conflicts with other packages and by making | |
13078 | sure e_os2.h will cover all platform-specific cases together with | |
13079 | opensslconf.h. | |
13080 | Additionally, it is now possible to define configuration/platform- | |
13081 | specific names (called "system identities"). In the C code, these | |
13082 | are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another | |
13083 | macro with the name beginning with `OPENSSL_SYS_`, which is determined | |
13084 | from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on | |
13085 | what is available. | |
13086 | ||
13087 | *Richard Levitte* | |
13088 | ||
13089 | * New option -set_serial to 'req' and 'x509' this allows the serial | |
13090 | number to use to be specified on the command line. Previously self | |
13091 | signed certificates were hard coded with serial number 0 and the | |
13092 | CA options of 'x509' had to use a serial number in a file which was | |
13093 | auto incremented. | |
13094 | ||
13095 | *Steve Henson* | |
13096 | ||
13097 | * New options to 'ca' utility to support V2 CRL entry extensions. | |
13098 | Currently CRL reason, invalidity date and hold instruction are | |
13099 | supported. Add new CRL extensions to V3 code and some new objects. | |
13100 | ||
13101 | *Steve Henson* | |
13102 | ||
13103 | * New function EVP_CIPHER_CTX_set_padding() this is used to | |
13104 | disable standard block padding (aka PKCS#5 padding) in the EVP | |
13105 | API, which was previously mandatory. This means that the data is | |
13106 | not padded in any way and so the total length much be a multiple | |
13107 | of the block size, otherwise an error occurs. | |
13108 | ||
13109 | *Steve Henson* | |
13110 | ||
13111 | * Initial (incomplete) OCSP SSL support. | |
13112 | ||
13113 | *Steve Henson* | |
13114 | ||
13115 | * New function OCSP_parse_url(). This splits up a URL into its host, | |
13116 | port and path components: primarily to parse OCSP URLs. New -url | |
13117 | option to ocsp utility. | |
13118 | ||
13119 | *Steve Henson* | |
13120 | ||
13121 | * New nonce behavior. The return value of OCSP_check_nonce() now | |
13122 | reflects the various checks performed. Applications can decide | |
13123 | whether to tolerate certain situations such as an absent nonce | |
13124 | in a response when one was present in a request: the ocsp application | |
13125 | just prints out a warning. New function OCSP_add1_basic_nonce() | |
13126 | this is to allow responders to include a nonce in a response even if | |
13127 | the request is nonce-less. | |
13128 | ||
13129 | *Steve Henson* | |
13130 | ||
13131 | * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are | |
13132 | skipped when using openssl x509 multiple times on a single input file, | |
13133 | e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`. | |
13134 | ||
13135 | *Bodo Moeller* | |
13136 | ||
13137 | * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() | |
13138 | set string type: to handle setting ASN1_TIME structures. Fix ca | |
13139 | utility to correctly initialize revocation date of CRLs. | |
13140 | ||
13141 | *Steve Henson* | |
13142 | ||
13143 | * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override | |
13144 | the clients preferred ciphersuites and rather use its own preferences. | |
13145 | Should help to work around M$ SGC (Server Gated Cryptography) bug in | |
13146 | Internet Explorer by ensuring unchanged hash method during stepup. | |
13147 | (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) | |
13148 | ||
13149 | *Lutz Jaenicke* | |
13150 | ||
13151 | * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael | |
13152 | to aes and add a new 'exist' option to print out symbols that don't | |
13153 | appear to exist. | |
13154 | ||
13155 | *Steve Henson* | |
13156 | ||
13157 | * Additional options to ocsp utility to allow flags to be set and | |
13158 | additional certificates supplied. | |
13159 | ||
13160 | *Steve Henson* | |
13161 | ||
13162 | * Add the option -VAfile to 'openssl ocsp', so the user can give the | |
13163 | OCSP client a number of certificate to only verify the response | |
13164 | signature against. | |
13165 | ||
13166 | *Richard Levitte* | |
13167 | ||
13168 | * Update Rijndael code to version 3.0 and change EVP AES ciphers to | |
13169 | handle the new API. Currently only ECB, CBC modes supported. Add new | |
13170 | AES OIDs. | |
13171 | ||
13172 | Add TLS AES ciphersuites as described in RFC3268, "Advanced | |
13173 | Encryption Standard (AES) Ciphersuites for Transport Layer | |
13174 | Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were | |
13175 | not enabled by default and were not part of the "ALL" ciphersuite | |
13176 | alias because they were not yet official; they could be | |
13177 | explicitly requested by specifying the "AESdraft" ciphersuite | |
13178 | group alias. In the final release of OpenSSL 0.9.7, the group | |
13179 | alias is called "AES" and is part of "ALL".) | |
13180 | ||
13181 | *Ben Laurie, Steve Henson, Bodo Moeller* | |
13182 | ||
13183 | * New function OCSP_copy_nonce() to copy nonce value (if present) from | |
13184 | request to response. | |
13185 | ||
13186 | *Steve Henson* | |
13187 | ||
13188 | * Functions for OCSP responders. OCSP_request_onereq_count(), | |
13189 | OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() | |
13190 | extract information from a certificate request. OCSP_response_create() | |
13191 | creates a response and optionally adds a basic response structure. | |
13192 | OCSP_basic_add1_status() adds a complete single response to a basic | |
13193 | response and returns the OCSP_SINGLERESP structure just added (to allow | |
13194 | extensions to be included for example). OCSP_basic_add1_cert() adds a | |
13195 | certificate to a basic response and OCSP_basic_sign() signs a basic | |
13196 | response with various flags. New helper functions ASN1_TIME_check() | |
13197 | (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() | |
13198 | (converts ASN1_TIME to GeneralizedTime). | |
13199 | ||
13200 | *Steve Henson* | |
13201 | ||
13202 | * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() | |
13203 | in a single operation. X509_get0_pubkey_bitstr() extracts the public_key | |
13204 | structure from a certificate. X509_pubkey_digest() digests the public_key | |
13205 | contents: this is used in various key identifiers. | |
13206 | ||
13207 | *Steve Henson* | |
13208 | ||
13209 | * Make sk_sort() tolerate a NULL argument. | |
13210 | ||
13211 | *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>* | |
13212 | ||
13213 | * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates | |
13214 | passed by the function are trusted implicitly. If any of them signed the | |
13215 | response then it is assumed to be valid and is not verified. | |
13216 | ||
13217 | *Steve Henson* | |
13218 | ||
13219 | * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT | |
13220 | to data. This was previously part of the PKCS7 ASN1 code. This | |
13221 | was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. | |
13222 | *Steve Henson, reported by Kenneth R. Robinette | |
13223 | <support@securenetterm.com>* | |
13224 | ||
13225 | * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 | |
13226 | routines: without these tracing memory leaks is very painful. | |
13227 | Fix leaks in PKCS12 and PKCS7 routines. | |
13228 | ||
13229 | *Steve Henson* | |
13230 | ||
13231 | * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). | |
13232 | Previously it initialised the 'type' argument to V_ASN1_UTCTIME which | |
13233 | effectively meant GeneralizedTime would never be used. Now it | |
13234 | is initialised to -1 but X509_time_adj() now has to check the value | |
13235 | and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or | |
13236 | V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. | |
13237 | *Steve Henson, reported by Kenneth R. Robinette | |
13238 | <support@securenetterm.com>* | |
13239 | ||
13240 | * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously | |
13241 | result in a zero length in the ASN1_INTEGER structure which was | |
13242 | not consistent with the structure when d2i_ASN1_INTEGER() was used | |
13243 | and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() | |
13244 | to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() | |
13245 | where it did not print out a minus for negative ASN1_INTEGER. | |
13246 | ||
13247 | *Steve Henson* | |
13248 | ||
13249 | * Add summary printout to ocsp utility. The various functions which | |
13250 | convert status values to strings have been renamed to: | |
13251 | OCSP_response_status_str(), OCSP_cert_status_str() and | |
13252 | OCSP_crl_reason_str() and are no longer static. New options | |
13253 | to verify nonce values and to disable verification. OCSP response | |
13254 | printout format cleaned up. | |
13255 | ||
13256 | *Steve Henson* | |
13257 | ||
13258 | * Add additional OCSP certificate checks. These are those specified | |
13259 | in RFC2560. This consists of two separate checks: the CA of the | |
13260 | certificate being checked must either be the OCSP signer certificate | |
13261 | or the issuer of the OCSP signer certificate. In the latter case the | |
13262 | OCSP signer certificate must contain the OCSP signing extended key | |
13263 | usage. This check is performed by attempting to match the OCSP | |
13264 | signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash | |
13265 | in the OCSP_CERTID structures of the response. | |
13266 | ||
13267 | *Steve Henson* | |
13268 | ||
13269 | * Initial OCSP certificate verification added to OCSP_basic_verify() | |
13270 | and related routines. This uses the standard OpenSSL certificate | |
13271 | verify routines to perform initial checks (just CA validity) and | |
13272 | to obtain the certificate chain. Then additional checks will be | |
13273 | performed on the chain. Currently the root CA is checked to see | |
13274 | if it is explicitly trusted for OCSP signing. This is used to set | |
13275 | a root CA as a global signing root: that is any certificate that | |
13276 | chains to that CA is an acceptable OCSP signing certificate. | |
13277 | ||
13278 | *Steve Henson* | |
13279 | ||
13280 | * New '-extfile ...' option to 'openssl ca' for reading X.509v3 | |
13281 | extensions from a separate configuration file. | |
13282 | As when reading extensions from the main configuration file, | |
13283 | the '-extensions ...' option may be used for specifying the | |
13284 | section to use. | |
13285 | ||
13286 | *Massimiliano Pala <madwolf@comune.modena.it>* | |
13287 | ||
13288 | * New OCSP utility. Allows OCSP requests to be generated or | |
13289 | read. The request can be sent to a responder and the output | |
13290 | parsed, outputted or printed in text form. Not complete yet: | |
13291 | still needs to check the OCSP response validity. | |
13292 | ||
13293 | *Steve Henson* | |
13294 | ||
13295 | * New subcommands for 'openssl ca': | |
13296 | `openssl ca -status <serial>` prints the status of the cert with | |
13297 | the given serial number (according to the index file). | |
13298 | `openssl ca -updatedb` updates the expiry status of certificates | |
13299 | in the index file. | |
13300 | ||
13301 | *Massimiliano Pala <madwolf@comune.modena.it>* | |
13302 | ||
13303 | * New '-newreq-nodes' command option to CA.pl. This is like | |
13304 | '-newreq', but calls 'openssl req' with the '-nodes' option | |
13305 | so that the resulting key is not encrypted. | |
13306 | ||
13307 | *Damien Miller <djm@mindrot.org>* | |
13308 | ||
13309 | * New configuration for the GNU Hurd. | |
13310 | ||
13311 | *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte* | |
13312 | ||
13313 | * Initial code to implement OCSP basic response verify. This | |
13314 | is currently incomplete. Currently just finds the signer's | |
13315 | certificate and verifies the signature on the response. | |
13316 | ||
13317 | *Steve Henson* | |
13318 | ||
13319 | * New SSLeay_version code SSLEAY_DIR to determine the compiled-in | |
13320 | value of OPENSSLDIR. This is available via the new '-d' option | |
13321 | to 'openssl version', and is also included in 'openssl version -a'. | |
13322 | ||
13323 | *Bodo Moeller* | |
13324 | ||
13325 | * Allowing defining memory allocation callbacks that will be given | |
13326 | file name and line number information in additional arguments | |
13327 | (a `const char*` and an int). The basic functionality remains, as | |
13328 | well as the original possibility to just replace malloc(), | |
13329 | realloc() and free() by functions that do not know about these | |
13330 | additional arguments. To register and find out the current | |
13331 | settings for extended allocation functions, the following | |
13332 | functions are provided: | |
13333 | ||
13334 | CRYPTO_set_mem_ex_functions | |
13335 | CRYPTO_set_locked_mem_ex_functions | |
13336 | CRYPTO_get_mem_ex_functions | |
13337 | CRYPTO_get_locked_mem_ex_functions | |
13338 | ||
13339 | These work the same way as CRYPTO_set_mem_functions and friends. | |
13340 | `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an | |
13341 | extended allocation function is enabled. | |
13342 | Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where | |
13343 | a conventional allocation function is enabled. | |
13344 | ||
13345 | *Richard Levitte, Bodo Moeller* | |
13346 | ||
13347 | * Finish off removing the remaining LHASH function pointer casts. | |
13348 | There should no longer be any prototype-casting required when using | |
13349 | the LHASH abstraction, and any casts that remain are "bugs". See | |
13350 | the callback types and macros at the head of lhash.h for details | |
13351 | (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). | |
13352 | ||
13353 | *Geoff Thorpe* | |
13354 | ||
13355 | * Add automatic query of EGD sockets in RAND_poll() for the unix variant. | |
13356 | If /dev/[u]random devices are not available or do not return enough | |
13357 | entropy, EGD style sockets (served by EGD or PRNGD) will automatically | |
13358 | be queried. | |
13359 | The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and | |
13360 | /etc/entropy will be queried once each in this sequence, querying stops | |
13361 | when enough entropy was collected without querying more sockets. | |
13362 | ||
13363 | *Lutz Jaenicke* | |
13364 | ||
13365 | * Change the Unix RAND_poll() variant to be able to poll several | |
13366 | random devices, as specified by DEVRANDOM, until a sufficient amount | |
13367 | of data has been collected. We spend at most 10 ms on each file | |
13368 | (select timeout) and read in non-blocking mode. DEVRANDOM now | |
13369 | defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" | |
13370 | (previously it was just the string "/dev/urandom"), so on typical | |
13371 | platforms the 10 ms delay will never occur. | |
13372 | Also separate out the Unix variant to its own file, rand_unix.c. | |
13373 | For VMS, there's a currently-empty rand_vms.c. | |
13374 | ||
13375 | *Richard Levitte* | |
13376 | ||
13377 | * Move OCSP client related routines to ocsp_cl.c. These | |
13378 | provide utility functions which an application needing | |
13379 | to issue a request to an OCSP responder and analyse the | |
13380 | response will typically need: as opposed to those which an | |
13381 | OCSP responder itself would need which will be added later. | |
13382 | ||
13383 | OCSP_request_sign() signs an OCSP request with an API similar | |
13384 | to PKCS7_sign(). OCSP_response_status() returns status of OCSP | |
13385 | response. OCSP_response_get1_basic() extracts basic response | |
13386 | from response. OCSP_resp_find_status(): finds and extracts status | |
13387 | information from an OCSP_CERTID structure (which will be created | |
13388 | when the request structure is built). These are built from lower | |
13389 | level functions which work on OCSP_SINGLERESP structures but | |
13390 | won't normally be used unless the application wishes to examine | |
13391 | extensions in the OCSP response for example. | |
13392 | ||
13393 | Replace nonce routines with a pair of functions. | |
13394 | OCSP_request_add1_nonce() adds a nonce value and optionally | |
13395 | generates a random value. OCSP_check_nonce() checks the | |
13396 | validity of the nonce in an OCSP response. | |
13397 | ||
13398 | *Steve Henson* | |
13399 | ||
13400 | * Change function OCSP_request_add() to OCSP_request_add0_id(). | |
13401 | This doesn't copy the supplied OCSP_CERTID and avoids the | |
13402 | need to free up the newly created id. Change return type | |
13403 | to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. | |
13404 | This can then be used to add extensions to the request. | |
13405 | Deleted OCSP_request_new(), since most of its functionality | |
13406 | is now in OCSP_REQUEST_new() (and the case insensitive name | |
13407 | clash) apart from the ability to set the request name which | |
13408 | will be added elsewhere. | |
13409 | ||
13410 | *Steve Henson* | |
13411 | ||
13412 | * Update OCSP API. Remove obsolete extensions argument from | |
13413 | various functions. Extensions are now handled using the new | |
13414 | OCSP extension code. New simple OCSP HTTP function which | |
13415 | can be used to send requests and parse the response. | |
13416 | ||
13417 | *Steve Henson* | |
13418 | ||
13419 | * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new | |
13420 | ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN | |
13421 | uses the special reorder version of SET OF to sort the attributes | |
13422 | and reorder them to match the encoded order. This resolves a long | |
13423 | standing problem: a verify on a PKCS7 structure just after signing | |
13424 | it used to fail because the attribute order did not match the | |
13425 | encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: | |
13426 | it uses the received order. This is necessary to tolerate some broken | |
13427 | software that does not order SET OF. This is handled by encoding | |
13428 | as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) | |
13429 | to produce the required SET OF. | |
13430 | ||
13431 | *Steve Henson* | |
13432 | ||
13433 | * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and | |
13434 | OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header | |
13435 | files to get correct declarations of the ASN.1 item variables. | |
13436 | ||
13437 | *Richard Levitte* | |
13438 | ||
13439 | * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many | |
13440 | PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: | |
13441 | asn1_check_tlen() would sometimes attempt to use 'ctx' when it was | |
13442 | NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). | |
13443 | New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant | |
13444 | ASN1_ITEM and no wrapper functions. | |
13445 | ||
13446 | *Steve Henson* | |
13447 | ||
13448 | * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These | |
13449 | replace the old function pointer based I/O routines. Change most of | |
13450 | the `*_d2i_bio()` and `*_d2i_fp()` functions to use these. | |
13451 | ||
13452 | *Steve Henson* | |
13453 | ||
13454 | * Enhance mkdef.pl to be more accepting about spacing in C preprocessor | |
13455 | lines, recognize more "algorithms" that can be deselected, and make | |
13456 | it complain about algorithm deselection that isn't recognised. | |
13457 | ||
13458 | *Richard Levitte* | |
13459 | ||
13460 | * New ASN1 functions to handle dup, sign, verify, digest, pack and | |
13461 | unpack operations in terms of ASN1_ITEM. Modify existing wrappers | |
13462 | to use new functions. Add NO_ASN1_OLD which can be set to remove | |
13463 | some old style ASN1 functions: this can be used to determine if old | |
13464 | code will still work when these eventually go away. | |
13465 | ||
13466 | *Steve Henson* | |
13467 | ||
13468 | * New extension functions for OCSP structures, these follow the | |
13469 | same conventions as certificates and CRLs. | |
13470 | ||
13471 | *Steve Henson* | |
13472 | ||
13473 | * New function X509V3_add1_i2d(). This automatically encodes and | |
13474 | adds an extension. Its behaviour can be customised with various | |
13475 | flags to append, replace or delete. Various wrappers added for | |
13476 | certificates and CRLs. | |
13477 | ||
13478 | *Steve Henson* | |
13479 | ||
13480 | * Fix to avoid calling the underlying ASN1 print routine when | |
13481 | an extension cannot be parsed. Correct a typo in the | |
13482 | OCSP_SERVICELOC extension. Tidy up print OCSP format. | |
13483 | ||
13484 | *Steve Henson* | |
13485 | ||
13486 | * Make mkdef.pl parse some of the ASN1 macros and add appropriate | |
13487 | entries for variables. | |
13488 | ||
13489 | *Steve Henson* | |
13490 | ||
13491 | * Add functionality to `apps/openssl.c` for detecting locking | |
13492 | problems: As the program is single-threaded, all we have | |
13493 | to do is register a locking callback using an array for | |
13494 | storing which locks are currently held by the program. | |
13495 | ||
13496 | *Bodo Moeller* | |
13497 | ||
13498 | * Use a lock around the call to CRYPTO_get_ex_new_index() in | |
13499 | SSL_get_ex_data_X509_STORE_idx(), which is used in | |
13500 | ssl_verify_cert_chain() and thus can be called at any time | |
13501 | during TLS/SSL handshakes so that thread-safety is essential. | |
13502 | Unfortunately, the ex_data design is not at all suited | |
13503 | for multi-threaded use, so it probably should be abolished. | |
13504 | ||
13505 | *Bodo Moeller* | |
13506 | ||
13507 | * Added Broadcom "ubsec" ENGINE to OpenSSL. | |
13508 | ||
13509 | *Broadcom, tweaked and integrated by Geoff Thorpe* | |
13510 | ||
13511 | * Move common extension printing code to new function | |
13512 | X509V3_print_extensions(). Reorganise OCSP print routines and | |
13513 | implement some needed OCSP ASN1 functions. Add OCSP extensions. | |
13514 | ||
13515 | *Steve Henson* | |
13516 | ||
13517 | * New function X509_signature_print() to remove duplication in some | |
13518 | print routines. | |
13519 | ||
13520 | *Steve Henson* | |
13521 | ||
13522 | * Add a special meaning when SET OF and SEQUENCE OF flags are both | |
13523 | set (this was treated exactly the same as SET OF previously). This | |
13524 | is used to reorder the STACK representing the structure to match the | |
13525 | encoding. This will be used to get round a problem where a PKCS7 | |
13526 | structure which was signed could not be verified because the STACK | |
13527 | order did not reflect the encoded order. | |
13528 | ||
13529 | *Steve Henson* | |
13530 | ||
13531 | * Reimplement the OCSP ASN1 module using the new code. | |
13532 | ||
13533 | *Steve Henson* | |
13534 | ||
13535 | * Update the X509V3 code to permit the use of an ASN1_ITEM structure | |
13536 | for its ASN1 operations. The old style function pointers still exist | |
13537 | for now but they will eventually go away. | |
13538 | ||
13539 | *Steve Henson* | |
13540 | ||
13541 | * Merge in replacement ASN1 code from the ASN1 branch. This almost | |
13542 | completely replaces the old ASN1 functionality with a table driven | |
13543 | encoder and decoder which interprets an ASN1_ITEM structure describing | |
13544 | the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is | |
13545 | largely maintained. Almost all of the old asn1_mac.h macro based ASN1 | |
13546 | has also been converted to the new form. | |
13547 | ||
13548 | *Steve Henson* | |
13549 | ||
13550 | * Change BN_mod_exp_recp so that negative moduli are tolerated | |
13551 | (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set | |
13552 | so that BN_mod_exp_mont and BN_mod_exp_mont_word work | |
13553 | for negative moduli. | |
13554 | ||
13555 | *Bodo Moeller* | |
13556 | ||
13557 | * Fix BN_uadd and BN_usub: Always return non-negative results instead | |
13558 | of not touching the result's sign bit. | |
13559 | ||
13560 | *Bodo Moeller* | |
13561 | ||
13562 | * BN_div bugfix: If the result is 0, the sign (res->neg) must not be | |
13563 | set. | |
13564 | ||
13565 | *Bodo Moeller* | |
13566 | ||
13567 | * Changed the LHASH code to use prototypes for callbacks, and created | |
13568 | macros to declare and implement thin (optionally static) functions | |
13569 | that provide type-safety and avoid function pointer casting for the | |
13570 | type-specific callbacks. | |
13571 | ||
13572 | *Geoff Thorpe* | |
13573 | ||
13574 | * Added Kerberos Cipher Suites to be used with TLS, as written in | |
13575 | RFC 2712. | |
13576 | *Veers Staats <staatsvr@asc.hpc.mil>, | |
13577 | Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte* | |
13578 | ||
13579 | * Reformat the FAQ so the different questions and answers can be divided | |
13580 | in sections depending on the subject. | |
13581 | ||
13582 | *Richard Levitte* | |
13583 | ||
13584 | * Have the zlib compression code load ZLIB.DLL dynamically under | |
13585 | Windows. | |
13586 | ||
13587 | *Richard Levitte* | |
13588 | ||
13589 | * New function BN_mod_sqrt for computing square roots modulo a prime | |
13590 | (using the probabilistic Tonelli-Shanks algorithm unless | |
13591 | p == 3 (mod 4) or p == 5 (mod 8), which are cases that can | |
13592 | be handled deterministically). | |
13593 | ||
13594 | *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller* | |
13595 | ||
13596 | * Make BN_mod_inverse faster by explicitly handling small quotients | |
13597 | in the Euclid loop. (Speed gain about 20% for small moduli [256 or | |
13598 | 512 bits], about 30% for larger ones [1024 or 2048 bits].) | |
13599 | ||
13600 | *Bodo Moeller* | |
13601 | ||
13602 | * New function BN_kronecker. | |
13603 | ||
13604 | *Bodo Moeller* | |
13605 | ||
13606 | * Fix BN_gcd so that it works on negative inputs; the result is | |
13607 | positive unless both parameters are zero. | |
13608 | Previously something reasonably close to an infinite loop was | |
13609 | possible because numbers could be growing instead of shrinking | |
13610 | in the implementation of Euclid's algorithm. | |
13611 | ||
13612 | *Bodo Moeller* | |
13613 | ||
13614 | * Fix BN_is_word() and BN_is_one() macros to take into account the | |
13615 | sign of the number in question. | |
13616 | ||
13617 | Fix BN_is_word(a,w) to work correctly for w == 0. | |
13618 | ||
13619 | The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) | |
13620 | because its test if the absolute value of 'a' equals 'w'. | |
13621 | Note that BN_abs_is_word does *not* handle w == 0 reliably; | |
13622 | it exists mostly for use in the implementations of BN_is_zero(), | |
13623 | BN_is_one(), and BN_is_word(). | |
13624 | ||
13625 | *Bodo Moeller* | |
13626 | ||
13627 | * New function BN_swap. | |
13628 | ||
13629 | *Bodo Moeller* | |
13630 | ||
13631 | * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that | |
13632 | the exponentiation functions are more likely to produce reasonable | |
13633 | results on negative inputs. | |
13634 | ||
13635 | *Bodo Moeller* | |
13636 | ||
13637 | * Change BN_mod_mul so that the result is always non-negative. | |
13638 | Previously, it could be negative if one of the factors was negative; | |
13639 | I don't think anyone really wanted that behaviour. | |
13640 | ||
13641 | *Bodo Moeller* | |
13642 | ||
13643 | * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c` | |
13644 | (except for exponentiation, which stays in `crypto/bn/bn_exp.c`, | |
13645 | and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`) | |
13646 | and add new functions: | |
13647 | ||
13648 | BN_nnmod | |
13649 | BN_mod_sqr | |
13650 | BN_mod_add | |
13651 | BN_mod_add_quick | |
13652 | BN_mod_sub | |
13653 | BN_mod_sub_quick | |
13654 | BN_mod_lshift1 | |
13655 | BN_mod_lshift1_quick | |
13656 | BN_mod_lshift | |
13657 | BN_mod_lshift_quick | |
13658 | ||
13659 | These functions always generate non-negative results. | |
13660 | ||
13661 | `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r` | |
13662 | such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead). | |
13663 | ||
13664 | `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as | |
13665 | `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`] | |
13666 | be reduced modulo `m`. | |
13667 | ||
13668 | *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller* | |
13669 | ||
13670 | <!-- | |
13671 | The following entry accidentally appeared in the CHANGES file | |
13672 | distributed with OpenSSL 0.9.7. The modifications described in | |
13673 | it do *not* apply to OpenSSL 0.9.7. | |
13674 | ||
13675 | * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there | |
13676 | was actually never needed) and in BN_mul(). The removal in BN_mul() | |
13677 | required a small change in bn_mul_part_recursive() and the addition | |
13678 | of the functions bn_cmp_part_words(), bn_sub_part_words() and | |
13679 | bn_add_part_words(), which do the same thing as bn_cmp_words(), | |
13680 | bn_sub_words() and bn_add_words() except they take arrays with | |
13681 | differing sizes. | |
13682 | ||
13683 | *Richard Levitte* | |
13684 | --> | |
13685 | ||
13686 | * In 'openssl passwd', verify passwords read from the terminal | |
13687 | unless the '-salt' option is used (which usually means that | |
13688 | verification would just waste user's time since the resulting | |
13689 | hash is going to be compared with some given password hash) | |
13690 | or the new '-noverify' option is used. | |
13691 | ||
13692 | This is an incompatible change, but it does not affect | |
13693 | non-interactive use of 'openssl passwd' (passwords on the command | |
13694 | line, '-stdin' option, '-in ...' option) and thus should not | |
13695 | cause any problems. | |
13696 | ||
13697 | *Bodo Moeller* | |
13698 | ||
13699 | * Remove all references to RSAref, since there's no more need for it. | |
13700 | ||
13701 | *Richard Levitte* | |
13702 | ||
13703 | * Make DSO load along a path given through an environment variable | |
13704 | (SHLIB_PATH) with shl_load(). | |
13705 | ||
13706 | *Richard Levitte* | |
13707 | ||
13708 | * Constify the ENGINE code as a result of BIGNUM constification. | |
13709 | Also constify the RSA code and most things related to it. In a | |
13710 | few places, most notable in the depth of the ASN.1 code, ugly | |
13711 | casts back to non-const were required (to be solved at a later | |
13712 | time) | |
13713 | ||
13714 | *Richard Levitte* | |
13715 | ||
13716 | * Make it so the openssl application has all engines loaded by default. | |
13717 | ||
13718 | *Richard Levitte* | |
13719 | ||
13720 | * Constify the BIGNUM routines a little more. | |
13721 | ||
13722 | *Richard Levitte* | |
13723 | ||
13724 | * Add the following functions: | |
13725 | ||
13726 | ENGINE_load_cswift() | |
13727 | ENGINE_load_chil() | |
13728 | ENGINE_load_atalla() | |
13729 | ENGINE_load_nuron() | |
13730 | ENGINE_load_builtin_engines() | |
13731 | ||
13732 | That way, an application can itself choose if external engines that | |
13733 | are built-in in OpenSSL shall ever be used or not. The benefit is | |
13734 | that applications won't have to be linked with libdl or other dso | |
13735 | libraries unless it's really needed. | |
13736 | ||
13737 | Changed 'openssl engine' to load all engines on demand. | |
13738 | Changed the engine header files to avoid the duplication of some | |
13739 | declarations (they differed!). | |
13740 | ||
13741 | *Richard Levitte* | |
13742 | ||
13743 | * 'openssl engine' can now list capabilities. | |
13744 | ||
13745 | *Richard Levitte* | |
13746 | ||
13747 | * Better error reporting in 'openssl engine'. | |
13748 | ||
13749 | *Richard Levitte* | |
13750 | ||
13751 | * Never call load_dh_param(NULL) in s_server. | |
13752 | ||
13753 | *Bodo Moeller* | |
13754 | ||
13755 | * Add engine application. It can currently list engines by name and | |
13756 | identity, and test if they are actually available. | |
13757 | ||
13758 | *Richard Levitte* | |
13759 | ||
13760 | * Improve RPM specification file by forcing symbolic linking and making | |
13761 | sure the installed documentation is also owned by root.root. | |
13762 | ||
13763 | *Damien Miller <djm@mindrot.org>* | |
13764 | ||
13765 | * Give the OpenSSL applications more possibilities to make use of | |
13766 | keys (public as well as private) handled by engines. | |
13767 | ||
13768 | *Richard Levitte* | |
13769 | ||
13770 | * Add OCSP code that comes from CertCo. | |
13771 | ||
13772 | *Richard Levitte* | |
13773 | ||
13774 | * Add VMS support for the Rijndael code. | |
13775 | ||
13776 | *Richard Levitte* | |
13777 | ||
13778 | * Added untested support for Nuron crypto accelerator. | |
13779 | ||
13780 | *Ben Laurie* | |
13781 | ||
13782 | * Add support for external cryptographic devices. This code was | |
13783 | previously distributed separately as the "engine" branch. | |
13784 | ||
13785 | *Geoff Thorpe, Richard Levitte* | |
13786 | ||
13787 | * Rework the filename-translation in the DSO code. It is now possible to | |
13788 | have far greater control over how a "name" is turned into a filename | |
13789 | depending on the operating environment and any oddities about the | |
13790 | different shared library filenames on each system. | |
13791 | ||
13792 | *Geoff Thorpe* | |
13793 | ||
13794 | * Support threads on FreeBSD-elf in Configure. | |
13795 | ||
13796 | *Richard Levitte* | |
13797 | ||
13798 | * Fix for SHA1 assembly problem with MASM: it produces | |
13799 | warnings about corrupt line number information when assembling | |
13800 | with debugging information. This is caused by the overlapping | |
13801 | of two sections. | |
13802 | ||
13803 | *Bernd Matthes <mainbug@celocom.de>, Steve Henson* | |
13804 | ||
13805 | * NCONF changes. | |
13806 | NCONF_get_number() has no error checking at all. As a replacement, | |
13807 | NCONF_get_number_e() is defined (`_e` for "error checking") and is | |
13808 | promoted strongly. The old NCONF_get_number is kept around for | |
13809 | binary backward compatibility. | |
13810 | Make it possible for methods to load from something other than a BIO, | |
13811 | by providing a function pointer that is given a name instead of a BIO. | |
13812 | For example, this could be used to load configuration data from an | |
13813 | LDAP server. | |
13814 | ||
13815 | *Richard Levitte* | |
13816 | ||
13817 | * Fix for non blocking accept BIOs. Added new I/O special reason | |
13818 | BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs | |
13819 | with non blocking I/O was not possible because no retry code was | |
13820 | implemented. Also added new SSL code SSL_WANT_ACCEPT to cover | |
13821 | this case. | |
13822 | ||
13823 | *Steve Henson* | |
13824 | ||
13825 | * Added the beginnings of Rijndael support. | |
13826 | ||
13827 | *Ben Laurie* | |
13828 | ||
13829 | * Fix for bug in DirectoryString mask setting. Add support for | |
13830 | X509_NAME_print_ex() in 'req' and X509_print_ex() function | |
13831 | to allow certificate printing to more controllable, additional | |
13832 | 'certopt' option to 'x509' to allow new printing options to be | |
13833 | set. | |
13834 | ||
13835 | *Steve Henson* | |
13836 | ||
13837 | * Clean old EAY MD5 hack from e_os.h. | |
13838 | ||
13839 | *Richard Levitte* | |
13840 | ||
13841 | ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004] | |
13842 | ||
13843 | * Fix null-pointer assignment in do_change_cipher_spec() revealed | |
13844 | by using the Codenomicon TLS Test Tool ([CVE-2004-0079]) | |
13845 | ||
13846 | *Joe Orton, Steve Henson* | |
13847 | ||
13848 | ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003] | |
13849 | ||
13850 | * Fix additional bug revealed by the NISCC test suite: | |
13851 | ||
13852 | Stop bug triggering large recursion when presented with | |
13853 | certain ASN.1 tags ([CVE-2003-0851]) | |
13854 | ||
13855 | *Steve Henson* | |
13856 | ||
13857 | ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003] | |
13858 | ||
13859 | * Fix various bugs revealed by running the NISCC test suite: | |
13860 | ||
13861 | Stop out of bounds reads in the ASN1 code when presented with | |
13862 | invalid tags (CVE-2003-0543 and CVE-2003-0544). | |
13863 | ||
13864 | If verify callback ignores invalid public key errors don't try to check | |
13865 | certificate signature with the NULL public key. | |
13866 | ||
13867 | *Steve Henson* | |
13868 | ||
13869 | * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate | |
13870 | if the server requested one: as stated in TLS 1.0 and SSL 3.0 | |
13871 | specifications. | |
13872 | ||
13873 | *Steve Henson* | |
13874 | ||
13875 | * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional | |
13876 | extra data after the compression methods not only for TLS 1.0 | |
13877 | but also for SSL 3.0 (as required by the specification). | |
13878 | ||
13879 | *Bodo Moeller; problem pointed out by Matthias Loepfe* | |
13880 | ||
13881 | * Change X509_certificate_type() to mark the key as exported/exportable | |
13882 | when it's 512 *bits* long, not 512 bytes. | |
13883 | ||
13884 | *Richard Levitte* | |
13885 | ||
13886 | ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003] | |
13887 | ||
13888 | * Countermeasure against the Klima-Pokorny-Rosa extension of | |
13889 | Bleichbacher's attack on PKCS #1 v1.5 padding: treat | |
13890 | a protocol version number mismatch like a decryption error | |
13891 | in ssl3_get_client_key_exchange (ssl/s3_srvr.c). | |
13892 | ||
13893 | *Bodo Moeller* | |
13894 | ||
13895 | * Turn on RSA blinding by default in the default implementation | |
13896 | to avoid a timing attack. Applications that don't want it can call | |
13897 | RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. | |
13898 | They would be ill-advised to do so in most cases. | |
13899 | ||
13900 | *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller* | |
13901 | ||
13902 | * Change RSA blinding code so that it works when the PRNG is not | |
13903 | seeded (in this case, the secret RSA exponent is abused as | |
13904 | an unpredictable seed -- if it is not unpredictable, there | |
13905 | is no point in blinding anyway). Make RSA blinding thread-safe | |
13906 | by remembering the creator's thread ID in rsa->blinding and | |
13907 | having all other threads use local one-time blinding factors | |
13908 | (this requires more computation than sharing rsa->blinding, but | |
13909 | avoids excessive locking; and if an RSA object is not shared | |
13910 | between threads, blinding will still be very fast). | |
13911 | ||
13912 | *Bodo Moeller* | |
13913 | ||
13914 | ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003] | |
13915 | ||
13916 | * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked | |
13917 | via timing by performing a MAC computation even if incorrect | |
13918 | block cipher padding has been found. This is a countermeasure | |
13919 | against active attacks where the attacker has to distinguish | |
13920 | between bad padding and a MAC verification error. ([CVE-2003-0078]) | |
13921 | ||
13922 | *Bodo Moeller; problem pointed out by Brice Canvel (EPFL), | |
13923 | Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and | |
13924 | Martin Vuagnoux (EPFL, Ilion)* | |
13925 | ||
13926 | ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002] | |
13927 | ||
13928 | * New function OPENSSL_cleanse(), which is used to cleanse a section of | |
13929 | memory from its contents. This is done with a counter that will | |
13930 | place alternating values in each byte. This can be used to solve | |
13931 | two issues: 1) the removal of calls to memset() by highly optimizing | |
13932 | compilers, and 2) cleansing with other values than 0, since those can | |
13933 | be read through on certain media, for example a swap space on disk. | |
13934 | ||
13935 | *Geoff Thorpe* | |
13936 | ||
13937 | * Bugfix: client side session caching did not work with external caching, | |
13938 | because the session->cipher setting was not restored when reloading | |
13939 | from the external cache. This problem was masked, when | |
13940 | SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. | |
13941 | (Found by Steve Haslam <steve@araqnid.ddts.net>.) | |
13942 | ||
13943 | *Lutz Jaenicke* | |
13944 | ||
13945 | * Fix client_certificate (ssl/s2_clnt.c): The permissible total | |
13946 | length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. | |
13947 | ||
13948 | *Zeev Lieber <zeev-l@yahoo.com>* | |
13949 | ||
13950 | * Undo an undocumented change introduced in 0.9.6e which caused | |
13951 | repeated calls to OpenSSL_add_all_ciphers() and | |
13952 | OpenSSL_add_all_digests() to be ignored, even after calling | |
13953 | EVP_cleanup(). | |
13954 | ||
13955 | *Richard Levitte* | |
13956 | ||
13957 | * Change the default configuration reader to deal with last line not | |
13958 | being properly terminated. | |
13959 | ||
13960 | *Richard Levitte* | |
13961 | ||
13962 | * Change X509_NAME_cmp() so it applies the special rules on handling | |
13963 | DN values that are of type PrintableString, as well as RDNs of type | |
13964 | emailAddress where the value has the type ia5String. | |
13965 | ||
13966 | *stefank@valicert.com via Richard Levitte* | |
13967 | ||
13968 | * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half | |
13969 | the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently | |
13970 | doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be | |
13971 | the bitwise-OR of the two for use by the majority of applications | |
13972 | wanting this behaviour, and update the docs. The documented | |
13973 | behaviour and actual behaviour were inconsistent and had been | |
13974 | changing anyway, so this is more a bug-fix than a behavioural | |
13975 | change. | |
13976 | ||
13977 | *Geoff Thorpe, diagnosed by Nadav Har'El* | |
13978 | ||
13979 | * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c | |
13980 | (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). | |
13981 | ||
13982 | *Bodo Moeller* | |
13983 | ||
13984 | * Fix initialization code race conditions in | |
13985 | SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), | |
13986 | SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), | |
13987 | SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), | |
13988 | TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), | |
13989 | ssl2_get_cipher_by_char(), | |
13990 | ssl3_get_cipher_by_char(). | |
13991 | ||
13992 | *Patrick McCormick <patrick@tellme.com>, Bodo Moeller* | |
13993 | ||
13994 | * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after | |
13995 | the cached sessions are flushed, as the remove_cb() might use ex_data | |
13996 | contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> | |
13997 | (see [openssl.org #212]). | |
13998 | ||
13999 | *Geoff Thorpe, Lutz Jaenicke* | |
14000 | ||
14001 | * Fix typo in OBJ_txt2obj which incorrectly passed the content | |
14002 | length, instead of the encoding length to d2i_ASN1_OBJECT. | |
14003 | ||
14004 | *Steve Henson* | |
14005 | ||
14006 | ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002] | |
14007 | ||
14008 | * [In 0.9.6g-engine release:] | |
14009 | Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`). | |
14010 | ||
14011 | *Lynn Gazis <lgazis@rainbow.com>* | |
14012 | ||
14013 | ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002] | |
14014 | ||
14015 | * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX | |
14016 | and get fix the header length calculation. | |
14017 | *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, | |
14018 | Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson* | |
14019 | ||
14020 | * Use proper error handling instead of 'assertions' in buffer | |
14021 | overflow checks added in 0.9.6e. This prevents DoS (the | |
14022 | assertions could call abort()). | |
14023 | ||
14024 | *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller* | |
14025 | ||
14026 | ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002] | |
14027 | ||
14028 | * Add various sanity checks to asn1_get_length() to reject | |
14029 | the ASN1 length bytes if they exceed sizeof(long), will appear | |
14030 | negative or the content length exceeds the length of the | |
14031 | supplied buffer. | |
14032 | ||
14033 | *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>* | |
14034 | ||
14035 | * Fix cipher selection routines: ciphers without encryption had no flags | |
14036 | for the cipher strength set and where therefore not handled correctly | |
14037 | by the selection routines (PR #130). | |
14038 | ||
14039 | *Lutz Jaenicke* | |
14040 | ||
14041 | * Fix EVP_dsa_sha macro. | |
14042 | ||
14043 | *Nils Larsch* | |
14044 | ||
14045 | * New option | |
14046 | SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS | |
14047 | for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure | |
14048 | that was added in OpenSSL 0.9.6d. | |
14049 | ||
14050 | As the countermeasure turned out to be incompatible with some | |
14051 | broken SSL implementations, the new option is part of SSL_OP_ALL. | |
14052 | SSL_OP_ALL is usually employed when compatibility with weird SSL | |
14053 | implementations is desired (e.g. '-bugs' option to 's_client' and | |
14054 | 's_server'), so the new option is automatically set in many | |
14055 | applications. | |
14056 | ||
14057 | *Bodo Moeller* | |
14058 | ||
14059 | * Changes in security patch: | |
14060 | ||
14061 | Changes marked "(CHATS)" were sponsored by the Defense Advanced | |
14062 | Research Projects Agency (DARPA) and Air Force Research Laboratory, | |
14063 | Air Force Materiel Command, USAF, under agreement number | |
14064 | F30602-01-2-0537. | |
14065 | ||
14066 | * Add various sanity checks to asn1_get_length() to reject | |
14067 | the ASN1 length bytes if they exceed sizeof(long), will appear | |
14068 | negative or the content length exceeds the length of the | |
14069 | supplied buffer. ([CVE-2002-0659]) | |
14070 | ||
14071 | *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>* | |
14072 | ||
14073 | * Assertions for various potential buffer overflows, not known to | |
14074 | happen in practice. | |
14075 | ||
14076 | *Ben Laurie (CHATS)* | |
14077 | ||
14078 | * Various temporary buffers to hold ASCII versions of integers were | |
14079 | too small for 64 bit platforms. ([CVE-2002-0655]) | |
14080 | *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>* | |
14081 | ||
14082 | * Remote buffer overflow in SSL3 protocol - an attacker could | |
14083 | supply an oversized session ID to a client. ([CVE-2002-0656]) | |
14084 | ||
14085 | *Ben Laurie (CHATS)* | |
14086 | ||
14087 | * Remote buffer overflow in SSL2 protocol - an attacker could | |
14088 | supply an oversized client master key. ([CVE-2002-0656]) | |
14089 | ||
14090 | *Ben Laurie (CHATS)* | |
14091 | ||
14092 | ### Changes between 0.9.6c and 0.9.6d [9 May 2002] | |
14093 | ||
14094 | * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not | |
14095 | encoded as NULL) with id-dsa-with-sha1. | |
14096 | ||
14097 | *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller* | |
14098 | ||
14099 | * Check various `X509_...()` return values in `apps/req.c`. | |
14100 | ||
14101 | *Nils Larsch <nla@trustcenter.de>* | |
14102 | ||
14103 | * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: | |
14104 | an end-of-file condition would erroneously be flagged, when the CRLF | |
14105 | was just at the end of a processed block. The bug was discovered when | |
14106 | processing data through a buffering memory BIO handing the data to a | |
14107 | BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov | |
14108 | <ptsekov@syntrex.com> and Nedelcho Stanev. | |
14109 | ||
14110 | *Lutz Jaenicke* | |
14111 | ||
14112 | * Implement a countermeasure against a vulnerability recently found | |
14113 | in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment | |
14114 | before application data chunks to avoid the use of known IVs | |
14115 | with data potentially chosen by the attacker. | |
14116 | ||
14117 | *Bodo Moeller* | |
14118 | ||
14119 | * Fix length checks in ssl3_get_client_hello(). | |
14120 | ||
14121 | *Bodo Moeller* | |
14122 | ||
14123 | * TLS/SSL library bugfix: use s->s3->in_read_app_data differently | |
14124 | to prevent ssl3_read_internal() from incorrectly assuming that | |
14125 | ssl3_read_bytes() found application data while handshake | |
14126 | processing was enabled when in fact s->s3->in_read_app_data was | |
14127 | merely automatically cleared during the initial handshake. | |
14128 | ||
14129 | *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>* | |
14130 | ||
14131 | * Fix object definitions for Private and Enterprise: they were not | |
14132 | recognized in their shortname (=lowercase) representation. Extend | |
14133 | obj_dat.pl to issue an error when using undefined keywords instead | |
14134 | of silently ignoring the problem (Svenning Sorensen | |
14135 | <sss@sss.dnsalias.net>). | |
14136 | ||
14137 | *Lutz Jaenicke* | |
14138 | ||
14139 | * Fix DH_generate_parameters() so that it works for 'non-standard' | |
14140 | generators, i.e. generators other than 2 and 5. (Previously, the | |
14141 | code did not properly initialise the 'add' and 'rem' values to | |
14142 | BN_generate_prime().) | |
14143 | ||
14144 | In the new general case, we do not insist that 'generator' is | |
14145 | actually a primitive root: This requirement is rather pointless; | |
14146 | a generator of the order-q subgroup is just as good, if not | |
14147 | better. | |
14148 | ||
14149 | *Bodo Moeller* | |
14150 | ||
14151 | * Map new X509 verification errors to alerts. Discovered and submitted by | |
14152 | Tom Wu <tom@arcot.com>. | |
14153 | ||
14154 | *Lutz Jaenicke* | |
14155 | ||
14156 | * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from | |
14157 | returning non-zero before the data has been completely received | |
14158 | when using non-blocking I/O. | |
14159 | ||
14160 | *Bodo Moeller; problem pointed out by John Hughes* | |
14161 | ||
14162 | * Some of the ciphers missed the strength entry (SSL_LOW etc). | |
14163 | ||
14164 | *Ben Laurie, Lutz Jaenicke* | |
14165 | ||
14166 | * Fix bug in SSL_clear(): bad sessions were not removed (found by | |
14167 | Yoram Zahavi <YoramZ@gilian.com>). | |
14168 | ||
14169 | *Lutz Jaenicke* | |
14170 | ||
14171 | * Add information about CygWin 1.3 and on, and preserve proper | |
14172 | configuration for the versions before that. | |
14173 | ||
14174 | *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte* | |
14175 | ||
14176 | * Make removal from session cache (SSL_CTX_remove_session()) more robust: | |
14177 | check whether we deal with a copy of a session and do not delete from | |
14178 | the cache in this case. Problem reported by "Izhar Shoshani Levi" | |
14179 | <izhar@checkpoint.com>. | |
14180 | ||
14181 | *Lutz Jaenicke* | |
14182 | ||
14183 | * Do not store session data into the internal session cache, if it | |
14184 | is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP | |
14185 | flag is set). Proposed by Aslam <aslam@funk.com>. | |
14186 | ||
14187 | *Lutz Jaenicke* | |
14188 | ||
14189 | * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested | |
14190 | value is 0. | |
14191 | ||
14192 | *Richard Levitte* | |
14193 | ||
14194 | * [In 0.9.6d-engine release:] | |
14195 | Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). | |
14196 | ||
14197 | *Toomas Kiisk <vix@cyber.ee> via Richard Levitte* | |
14198 | ||
14199 | * Add the configuration target linux-s390x. | |
14200 | ||
14201 | *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte* | |
14202 | ||
14203 | * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of | |
14204 | ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag | |
14205 | variable as an indication that a ClientHello message has been | |
14206 | received. As the flag value will be lost between multiple | |
14207 | invocations of ssl3_accept when using non-blocking I/O, the | |
14208 | function may not be aware that a handshake has actually taken | |
14209 | place, thus preventing a new session from being added to the | |
14210 | session cache. | |
14211 | ||
14212 | To avoid this problem, we now set s->new_session to 2 instead of | |
14213 | using a local variable. | |
14214 | ||
14215 | *Lutz Jaenicke, Bodo Moeller* | |
14216 | ||
14217 | * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) | |
14218 | if the SSL_R_LENGTH_MISMATCH error is detected. | |
14219 | ||
14220 | *Geoff Thorpe, Bodo Moeller* | |
14221 | ||
14222 | * New 'shared_ldflag' column in Configure platform table. | |
14223 | ||
14224 | *Richard Levitte* | |
14225 | ||
14226 | * Fix EVP_CIPHER_mode macro. | |
14227 | ||
14228 | *"Dan S. Camper" <dan@bti.net>* | |
14229 | ||
14230 | * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown | |
14231 | type, we must throw them away by setting rr->length to 0. | |
14232 | ||
14233 | *D P Chang <dpc@qualys.com>* | |
14234 | ||
14235 | ### Changes between 0.9.6b and 0.9.6c [21 dec 2001] | |
14236 | ||
14237 | * Fix BN_rand_range bug pointed out by Dominikus Scherkl | |
14238 | <Dominikus.Scherkl@biodata.com>. (The previous implementation | |
14239 | worked incorrectly for those cases where range = `10..._2` and | |
14240 | `3*range` is two bits longer than range.) | |
14241 | ||
14242 | *Bodo Moeller* | |
14243 | ||
14244 | * Only add signing time to PKCS7 structures if it is not already | |
14245 | present. | |
14246 | ||
14247 | *Steve Henson* | |
14248 | ||
14249 | * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", | |
14250 | OBJ_ld_ce should be OBJ_id_ce. | |
14251 | Also some ip-pda OIDs in crypto/objects/objects.txt were | |
14252 | incorrect (cf. RFC 3039). | |
14253 | ||
14254 | *Matt Cooper, Frederic Giudicelli, Bodo Moeller* | |
14255 | ||
14256 | * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() | |
14257 | returns early because it has nothing to do. | |
14258 | ||
14259 | *Andy Schneider <andy.schneider@bjss.co.uk>* | |
14260 | ||
14261 | * [In 0.9.6c-engine release:] | |
14262 | Fix mutex callback return values in crypto/engine/hw_ncipher.c. | |
14263 | ||
14264 | *Andy Schneider <andy.schneider@bjss.co.uk>* | |
14265 | ||
14266 | * [In 0.9.6c-engine release:] | |
14267 | Add support for Cryptographic Appliance's keyserver technology. | |
14268 | (Use engine 'keyclient') | |
14269 | ||
14270 | *Cryptographic Appliances and Geoff Thorpe* | |
14271 | ||
14272 | * Add a configuration entry for OS/390 Unix. The C compiler 'c89' | |
14273 | is called via tools/c89.sh because arguments have to be | |
14274 | rearranged (all '-L' options must appear before the first object | |
14275 | modules). | |
14276 | ||
14277 | *Richard Shapiro <rshapiro@abinitio.com>* | |
14278 | ||
14279 | * [In 0.9.6c-engine release:] | |
14280 | Add support for Broadcom crypto accelerator cards, backported | |
14281 | from 0.9.7. | |
14282 | ||
14283 | *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox* | |
14284 | ||
14285 | * [In 0.9.6c-engine release:] | |
14286 | Add support for SureWare crypto accelerator cards from | |
14287 | Baltimore Technologies. (Use engine 'sureware') | |
14288 | ||
14289 | *Baltimore Technologies and Mark Cox* | |
14290 | ||
14291 | * [In 0.9.6c-engine release:] | |
14292 | Add support for crypto accelerator cards from Accelerated | |
14293 | Encryption Processing, www.aep.ie. (Use engine 'aep') | |
14294 | ||
14295 | *AEP Inc. and Mark Cox* | |
14296 | ||
14297 | * Add a configuration entry for gcc on UnixWare. | |
14298 | ||
14299 | *Gary Benson <gbenson@redhat.com>* | |
14300 | ||
14301 | * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake | |
14302 | messages are stored in a single piece (fixed-length part and | |
14303 | variable-length part combined) and fix various bugs found on the way. | |
14304 | ||
14305 | *Bodo Moeller* | |
14306 | ||
14307 | * Disable caching in BIO_gethostbyname(), directly use gethostbyname() | |
14308 | instead. BIO_gethostbyname() does not know what timeouts are | |
14309 | appropriate, so entries would stay in cache even when they have | |
14310 | become invalid. | |
14311 | *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>* | |
14312 | ||
14313 | * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when | |
14314 | faced with a pathologically small ClientHello fragment that does | |
14315 | not contain client_version: Instead of aborting with an error, | |
14316 | simply choose the highest available protocol version (i.e., | |
14317 | TLS 1.0 unless it is disabled). In practice, ClientHello | |
14318 | messages are never sent like this, but this change gives us | |
14319 | strictly correct behaviour at least for TLS. | |
14320 | ||
14321 | *Bodo Moeller* | |
14322 | ||
14323 | * Fix SSL handshake functions and SSL_clear() such that SSL_clear() | |
14324 | never resets s->method to s->ctx->method when called from within | |
14325 | one of the SSL handshake functions. | |
14326 | ||
14327 | *Bodo Moeller; problem pointed out by Niko Baric* | |
14328 | ||
14329 | * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert | |
14330 | (sent using the client's version number) if client_version is | |
14331 | smaller than the protocol version in use. Also change | |
14332 | ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if | |
14333 | the client demanded SSL 3.0 but only TLS 1.0 is enabled; then | |
14334 | the client will at least see that alert. | |
14335 | ||
14336 | *Bodo Moeller* | |
14337 | ||
14338 | * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation | |
14339 | correctly. | |
14340 | ||
14341 | *Bodo Moeller* | |
14342 | ||
14343 | * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a | |
14344 | client receives HelloRequest while in a handshake. | |
14345 | ||
14346 | *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>* | |
14347 | ||
14348 | * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C | |
14349 | should end in 'break', not 'goto end' which circumvents various | |
14350 | cleanups done in state SSL_ST_OK. But session related stuff | |
14351 | must be disabled for SSL_ST_OK in the case that we just sent a | |
14352 | HelloRequest. | |
14353 | ||
14354 | Also avoid some overhead by not calling ssl_init_wbio_buffer() | |
14355 | before just sending a HelloRequest. | |
14356 | ||
14357 | *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>* | |
14358 | ||
14359 | * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't | |
14360 | reveal whether illegal block cipher padding was found or a MAC | |
14361 | verification error occurred. (Neither SSLerr() codes nor alerts | |
14362 | are directly visible to potential attackers, but the information | |
14363 | may leak via logfiles.) | |
14364 | ||
14365 | Similar changes are not required for the SSL 2.0 implementation | |
14366 | because the number of padding bytes is sent in clear for SSL 2.0, | |
14367 | and the extra bytes are just ignored. However ssl/s2_pkt.c | |
14368 | failed to verify that the purported number of padding bytes is in | |
14369 | the legal range. | |
14370 | ||
14371 | *Bodo Moeller* | |
14372 | ||
14373 | * Add OpenUNIX-8 support including shared libraries | |
14374 | (Boyd Lynn Gerber <gerberb@zenez.com>). | |
14375 | ||
14376 | *Lutz Jaenicke* | |
14377 | ||
14378 | * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid | |
14379 | 'wristwatch attack' using huge encoding parameters (cf. | |
14380 | James H. Manger's CRYPTO 2001 paper). Note that the | |
14381 | RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use | |
14382 | encoding parameters and hence was not vulnerable. | |
14383 | ||
14384 | *Bodo Moeller* | |
14385 | ||
14386 | * BN_sqr() bug fix. | |
14387 | ||
14388 | *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>* | |
14389 | ||
14390 | * Rabin-Miller test analyses assume uniformly distributed witnesses, | |
14391 | so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() | |
14392 | followed by modular reduction. | |
14393 | ||
14394 | *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>* | |
14395 | ||
14396 | * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() | |
14397 | equivalent based on BN_pseudo_rand() instead of BN_rand(). | |
14398 | ||
14399 | *Bodo Moeller* | |
14400 | ||
14401 | * s3_srvr.c: allow sending of large client certificate lists (> 16 kB). | |
14402 | This function was broken, as the check for a new client hello message | |
14403 | to handle SGC did not allow these large messages. | |
14404 | (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) | |
14405 | ||
14406 | *Lutz Jaenicke* | |
14407 | ||
14408 | * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`. | |
14409 | ||
14410 | *Lutz Jaenicke* | |
14411 | ||
14412 | * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() | |
14413 | for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). | |
14414 | ||
14415 | *Lutz Jaenicke* | |
14416 | ||
14417 | * Rework the configuration and shared library support for Tru64 Unix. | |
14418 | The configuration part makes use of modern compiler features and | |
14419 | still retains old compiler behavior for those that run older versions | |
14420 | of the OS. The shared library support part includes a variant that | |
14421 | uses the RPATH feature, and is available through the special | |
14422 | configuration target "alpha-cc-rpath", which will never be selected | |
14423 | automatically. | |
14424 | ||
14425 | *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte* | |
14426 | ||
14427 | * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() | |
14428 | with the same message size as in ssl3_get_certificate_request(). | |
14429 | Otherwise, if no ServerKeyExchange message occurs, CertificateRequest | |
14430 | messages might inadvertently be reject as too long. | |
14431 | ||
14432 | *Petr Lampa <lampa@fee.vutbr.cz>* | |
14433 | ||
14434 | * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). | |
14435 | ||
14436 | *Andy Polyakov* | |
14437 | ||
14438 | * Modified SSL library such that the verify_callback that has been set | |
14439 | specifically for an SSL object with SSL_set_verify() is actually being | |
14440 | used. Before the change, a verify_callback set with this function was | |
14441 | ignored and the verify_callback() set in the SSL_CTX at the time of | |
14442 | the call was used. New function X509_STORE_CTX_set_verify_cb() introduced | |
14443 | to allow the necessary settings. | |
14444 | ||
14445 | *Lutz Jaenicke* | |
14446 | ||
14447 | * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c | |
14448 | explicitly to NULL, as at least on Solaris 8 this seems not always to be | |
14449 | done automatically (in contradiction to the requirements of the C | |
14450 | standard). This made problems when used from OpenSSH. | |
14451 | ||
14452 | *Lutz Jaenicke* | |
14453 | ||
14454 | * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored | |
14455 | dh->length and always used | |
14456 | ||
14457 | BN_rand_range(priv_key, dh->p). | |
14458 | ||
14459 | BN_rand_range() is not necessary for Diffie-Hellman, and this | |
14460 | specific range makes Diffie-Hellman unnecessarily inefficient if | |
14461 | dh->length (recommended exponent length) is much smaller than the | |
14462 | length of dh->p. We could use BN_rand_range() if the order of | |
14463 | the subgroup was stored in the DH structure, but we only have | |
14464 | dh->length. | |
14465 | ||
14466 | So switch back to | |
14467 | ||
14468 | BN_rand(priv_key, l, ...) | |
14469 | ||
14470 | where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 | |
14471 | otherwise. | |
14472 | ||
14473 | *Bodo Moeller* | |
14474 | ||
14475 | * In | |
14476 | ||
14477 | RSA_eay_public_encrypt | |
14478 | RSA_eay_private_decrypt | |
14479 | RSA_eay_private_encrypt (signing) | |
14480 | RSA_eay_public_decrypt (signature verification) | |
14481 | ||
14482 | (default implementations for RSA_public_encrypt, | |
14483 | RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), | |
14484 | always reject numbers >= n. | |
14485 | ||
14486 | *Bodo Moeller* | |
14487 | ||
14488 | * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 | |
14489 | to synchronize access to 'locking_thread'. This is necessary on | |
14490 | systems where access to 'locking_thread' (an 'unsigned long' | |
14491 | variable) is not atomic. | |
14492 | ||
14493 | *Bodo Moeller* | |
14494 | ||
14495 | * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID | |
14496 | *before* setting the 'crypto_lock_rand' flag. The previous code had | |
14497 | a race condition if 0 is a valid thread ID. | |
14498 | ||
14499 | *Travis Vitek <vitek@roguewave.com>* | |
14500 | ||
14501 | * Add support for shared libraries under Irix. | |
14502 | ||
14503 | *Albert Chin-A-Young <china@thewrittenword.com>* | |
14504 | ||
14505 | * Add configuration option to build on Linux on both big-endian and | |
14506 | little-endian MIPS. | |
14507 | ||
14508 | *Ralf Baechle <ralf@uni-koblenz.de>* | |
14509 | ||
14510 | * Add the possibility to create shared libraries on HP-UX. | |
14511 | ||
14512 | *Richard Levitte* | |
14513 | ||
14514 | ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001] | |
14515 | ||
14516 | * Change ssleay_rand_bytes (crypto/rand/md_rand.c) | |
14517 | to avoid a SSLeay/OpenSSL PRNG weakness pointed out by | |
14518 | Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: | |
14519 | PRNG state recovery was possible based on the output of | |
14520 | one PRNG request appropriately sized to gain knowledge on | |
14521 | 'md' followed by enough consecutive 1-byte PRNG requests | |
14522 | to traverse all of 'state'. | |
14523 | ||
14524 | 1. When updating 'md_local' (the current thread's copy of 'md') | |
14525 | during PRNG output generation, hash all of the previous | |
14526 | 'md_local' value, not just the half used for PRNG output. | |
14527 | ||
14528 | 2. Make the number of bytes from 'state' included into the hash | |
14529 | independent from the number of PRNG bytes requested. | |
14530 | ||
14531 | The first measure alone would be sufficient to avoid | |
14532 | Markku-Juhani's attack. (Actually it had never occurred | |
14533 | to me that the half of 'md_local' used for chaining was the | |
14534 | half from which PRNG output bytes were taken -- I had always | |
14535 | assumed that the secret half would be used.) The second | |
14536 | measure makes sure that additional data from 'state' is never | |
14537 | mixed into 'md_local' in small portions; this heuristically | |
14538 | further strengthens the PRNG. | |
14539 | ||
14540 | *Bodo Moeller* | |
14541 | ||
14542 | * Fix crypto/bn/asm/mips3.s. | |
14543 | ||
14544 | *Andy Polyakov* | |
14545 | ||
14546 | * When only the key is given to "enc", the IV is undefined. Print out | |
14547 | an error message in this case. | |
14548 | ||
14549 | *Lutz Jaenicke* | |
14550 | ||
14551 | * Handle special case when X509_NAME is empty in X509 printing routines. | |
14552 | ||
14553 | *Steve Henson* | |
14554 | ||
14555 | * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are | |
14556 | positive and less than q. | |
14557 | ||
14558 | *Bodo Moeller* | |
14559 | ||
14560 | * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is | |
14561 | used: it isn't thread safe and the add_lock_callback should handle | |
14562 | that itself. | |
14563 | ||
14564 | *Paul Rose <Paul.Rose@bridge.com>* | |
14565 | ||
14566 | * Verify that incoming data obeys the block size in | |
14567 | ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). | |
14568 | ||
14569 | *Bodo Moeller* | |
14570 | ||
14571 | * Fix OAEP check. | |
14572 | ||
14573 | *Ulf Möller, Bodo Möller* | |
14574 | ||
14575 | * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 | |
14576 | RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 | |
14577 | when fixing the server behaviour for backwards-compatible 'client | |
14578 | hello' messages. (Note that the attack is impractical against | |
14579 | SSL 3.0 and TLS 1.0 anyway because length and version checking | |
14580 | means that the probability of guessing a valid ciphertext is | |
14581 | around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 | |
14582 | paper.) | |
14583 | ||
14584 | Before 0.9.5, the countermeasure (hide the error by generating a | |
14585 | random 'decryption result') did not work properly because | |
14586 | ERR_clear_error() was missing, meaning that SSL_get_error() would | |
14587 | detect the supposedly ignored error. | |
14588 | ||
14589 | Both problems are now fixed. | |
14590 | ||
14591 | *Bodo Moeller* | |
14592 | ||
14593 | * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 | |
14594 | (previously it was 1024). | |
14595 | ||
14596 | *Bodo Moeller* | |
14597 | ||
14598 | * Fix for compatibility mode trust settings: ignore trust settings | |
14599 | unless some valid trust or reject settings are present. | |
14600 | ||
14601 | *Steve Henson* | |
14602 | ||
14603 | * Fix for blowfish EVP: its a variable length cipher. | |
14604 | ||
14605 | *Steve Henson* | |
14606 | ||
14607 | * Fix various bugs related to DSA S/MIME verification. Handle missing | |
14608 | parameters in DSA public key structures and return an error in the | |
14609 | DSA routines if parameters are absent. | |
14610 | ||
14611 | *Steve Henson* | |
14612 | ||
14613 | * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" | |
14614 | in the current directory if neither $RANDFILE nor $HOME was set. | |
14615 | RAND_file_name() in 0.9.6a returned NULL in this case. This has | |
14616 | caused some confusion to Windows users who haven't defined $HOME. | |
14617 | Thus RAND_file_name() is changed again: e_os.h can define a | |
14618 | DEFAULT_HOME, which will be used if $HOME is not set. | |
14619 | For Windows, we use "C:"; on other platforms, we still require | |
14620 | environment variables. | |
14621 | ||
14622 | * Move 'if (!initialized) RAND_poll()' into regions protected by | |
14623 | CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids | |
14624 | having multiple threads call RAND_poll() concurrently. | |
14625 | ||
14626 | *Bodo Moeller* | |
14627 | ||
14628 | * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a | |
14629 | combination of a flag and a thread ID variable. | |
14630 | Otherwise while one thread is in ssleay_rand_bytes (which sets the | |
14631 | flag), *other* threads can enter ssleay_add_bytes without obeying | |
14632 | the CRYPTO_LOCK_RAND lock (and may even illegally release the lock | |
14633 | that they do not hold after the first thread unsets add_do_not_lock). | |
14634 | ||
14635 | *Bodo Moeller* | |
14636 | ||
14637 | * Change bctest again: '-x' expressions are not available in all | |
14638 | versions of 'test'. | |
14639 | ||
14640 | *Bodo Moeller* | |
14641 | ||
14642 | ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001] | |
14643 | ||
14644 | * Fix a couple of memory leaks in PKCS7_dataDecode() | |
14645 | ||
14646 | *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>* | |
14647 | ||
14648 | * Change Configure and Makefiles to provide EXE_EXT, which will contain | |
14649 | the default extension for executables, if any. Also, make the perl | |
14650 | scripts that use symlink() to test if it really exists and use "cp" | |
14651 | if it doesn't. All this made OpenSSL compilable and installable in | |
14652 | CygWin. | |
14653 | ||
14654 | *Richard Levitte* | |
14655 | ||
14656 | * Fix for asn1_GetSequence() for indefinite length constructed data. | |
14657 | If SEQUENCE is length is indefinite just set c->slen to the total | |
14658 | amount of data available. | |
14659 | ||
14660 | *Steve Henson, reported by shige@FreeBSD.org* | |
14661 | ||
14662 | *This change does not apply to 0.9.7.* | |
14663 | ||
14664 | * Change bctest to avoid here-documents inside command substitution | |
14665 | (workaround for FreeBSD /bin/sh bug). | |
14666 | For compatibility with Ultrix, avoid shell functions (introduced | |
14667 | in the bctest version that searches along $PATH). | |
14668 | ||
14669 | *Bodo Moeller* | |
14670 | ||
14671 | * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes | |
14672 | with des_encrypt() defined on some operating systems, like Solaris | |
14673 | and UnixWare. | |
14674 | ||
14675 | *Richard Levitte* | |
14676 | ||
14677 | * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: | |
14678 | On the Importance of Eliminating Errors in Cryptographic | |
14679 | Computations, J. Cryptology 14 (2001) 2, 101-119, | |
14680 | <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>). | |
14681 | ||
14682 | *Ulf Moeller* | |
14683 | ||
14684 | * MIPS assembler BIGNUM division bug fix. | |
14685 | ||
14686 | *Andy Polyakov* | |
14687 | ||
14688 | * Disabled incorrect Alpha assembler code. | |
14689 | ||
14690 | *Richard Levitte* | |
14691 | ||
14692 | * Fix PKCS#7 decode routines so they correctly update the length | |
14693 | after reading an EOC for the EXPLICIT tag. | |
14694 | ||
14695 | *Steve Henson* | |
14696 | ||
14697 | *This change does not apply to 0.9.7.* | |
14698 | ||
14699 | * Fix bug in PKCS#12 key generation routines. This was triggered | |
14700 | if a 3DES key was generated with a 0 initial byte. Include | |
14701 | PKCS12_BROKEN_KEYGEN compilation option to retain the old | |
14702 | (but broken) behaviour. | |
14703 | ||
14704 | *Steve Henson* | |
14705 | ||
14706 | * Enhance bctest to search for a working bc along $PATH and print | |
14707 | it when found. | |
14708 | ||
14709 | *Tim Rice <tim@multitalents.net> via Richard Levitte* | |
14710 | ||
14711 | * Fix memory leaks in err.c: free err_data string if necessary; | |
14712 | don't write to the wrong index in ERR_set_error_data. | |
14713 | ||
14714 | *Bodo Moeller* | |
14715 | ||
14716 | * Implement ssl23_peek (analogous to ssl23_read), which previously | |
14717 | did not exist. | |
14718 | ||
14719 | *Bodo Moeller* | |
14720 | ||
14721 | * Replace rdtsc with `_emit` statements for VC++ version 5. | |
14722 | ||
14723 | *Jeremy Cooper <jeremy@baymoo.org>* | |
14724 | ||
14725 | * Make it possible to reuse SSLv2 sessions. | |
14726 | ||
14727 | *Richard Levitte* | |
14728 | ||
14729 | * In copy_email() check for >= 0 as a return value for | |
14730 | X509_NAME_get_index_by_NID() since 0 is a valid index. | |
14731 | ||
14732 | *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>* | |
14733 | ||
14734 | * Avoid coredump with unsupported or invalid public keys by checking if | |
14735 | X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when | |
14736 | PKCS7_verify() fails with non detached data. | |
14737 | ||
14738 | *Steve Henson* | |
14739 | ||
14740 | * Don't use getenv in library functions when run as setuid/setgid. | |
14741 | New function OPENSSL_issetugid(). | |
14742 | ||
14743 | *Ulf Moeller* | |
14744 | ||
14745 | * Avoid false positives in memory leak detection code (crypto/mem_dbg.c) | |
14746 | due to incorrect handling of multi-threading: | |
14747 | ||
14748 | 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). | |
14749 | ||
14750 | 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). | |
14751 | ||
14752 | 3. Count how many times MemCheck_off() has been called so that | |
14753 | nested use can be treated correctly. This also avoids | |
14754 | inband-signalling in the previous code (which relied on the | |
14755 | assumption that thread ID 0 is impossible). | |
14756 | ||
14757 | *Bodo Moeller* | |
14758 | ||
14759 | * Add "-rand" option also to s_client and s_server. | |
14760 | ||
14761 | *Lutz Jaenicke* | |
14762 | ||
14763 | * Fix CPU detection on Irix 6.x. | |
14764 | *Kurt Hockenbury <khockenb@stevens-tech.edu> and | |
14765 | "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>* | |
14766 | ||
14767 | * Fix X509_NAME bug which produced incorrect encoding if X509_NAME | |
14768 | was empty. | |
14769 | ||
14770 | *Steve Henson* | |
14771 | ||
14772 | *This change does not apply to 0.9.7.* | |
14773 | ||
14774 | * Use the cached encoding of an X509_NAME structure rather than | |
14775 | copying it. This is apparently the reason for the libsafe "errors" | |
14776 | but the code is actually correct. | |
14777 | ||
14778 | *Steve Henson* | |
14779 | ||
14780 | * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent | |
14781 | Bleichenbacher's DSA attack. | |
14782 | Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits | |
14783 | to be set and top=0 forces the highest bit to be set; top=-1 is new | |
14784 | and leaves the highest bit random. | |
14785 | ||
14786 | *Ulf Moeller, Bodo Moeller* | |
14787 | ||
14788 | * In the `NCONF_...`-based implementations for `CONF_...` queries | |
14789 | (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using | |
14790 | a temporary CONF structure with the data component set to NULL | |
14791 | (which gives segmentation faults in lh_retrieve). | |
14792 | Instead, use NULL for the CONF pointer in CONF_get_string and | |
14793 | CONF_get_number (which may use environment variables) and directly | |
14794 | return NULL from CONF_get_section. | |
14795 | ||
14796 | *Bodo Moeller* | |
14797 | ||
14798 | * Fix potential buffer overrun for EBCDIC. | |
14799 | ||
14800 | *Ulf Moeller* | |
14801 | ||
14802 | * Tolerate nonRepudiation as being valid for S/MIME signing and certSign | |
14803 | keyUsage if basicConstraints absent for a CA. | |
14804 | ||
14805 | *Steve Henson* | |
14806 | ||
14807 | * Make SMIME_write_PKCS7() write mail header values with a format that | |
14808 | is more generally accepted (no spaces before the semicolon), since | |
14809 | some programs can't parse those values properly otherwise. Also make | |
14810 | sure BIO's that break lines after each write do not create invalid | |
14811 | headers. | |
14812 | ||
14813 | *Richard Levitte* | |
14814 | ||
14815 | * Make the CRL encoding routines work with empty SEQUENCE OF. The | |
14816 | macros previously used would not encode an empty SEQUENCE OF | |
14817 | and break the signature. | |
14818 | ||
14819 | *Steve Henson* | |
14820 | ||
14821 | *This change does not apply to 0.9.7.* | |
14822 | ||
14823 | * Zero the premaster secret after deriving the master secret in | |
14824 | DH ciphersuites. | |
14825 | ||
14826 | *Steve Henson* | |
14827 | ||
14828 | * Add some EVP_add_digest_alias registrations (as found in | |
14829 | OpenSSL_add_all_digests()) to SSL_library_init() | |
14830 | aka OpenSSL_add_ssl_algorithms(). This provides improved | |
14831 | compatibility with peers using X.509 certificates | |
14832 | with unconventional AlgorithmIdentifier OIDs. | |
14833 | ||
14834 | *Bodo Moeller* | |
14835 | ||
14836 | * Fix for Irix with NO_ASM. | |
14837 | ||
14838 | *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>* | |
14839 | ||
14840 | * ./config script fixes. | |
14841 | ||
14842 | *Ulf Moeller, Richard Levitte* | |
14843 | ||
14844 | * Fix 'openssl passwd -1'. | |
14845 | ||
14846 | *Bodo Moeller* | |
14847 | ||
14848 | * Change PKCS12_key_gen_asc() so it can cope with non null | |
14849 | terminated strings whose length is passed in the passlen | |
14850 | parameter, for example from PEM callbacks. This was done | |
14851 | by adding an extra length parameter to asc2uni(). | |
14852 | ||
14853 | *Steve Henson, reported by <oddissey@samsung.co.kr>* | |
14854 | ||
14855 | * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn | |
14856 | call failed, free the DSA structure. | |
14857 | ||
14858 | *Bodo Moeller* | |
14859 | ||
14860 | * Fix to uni2asc() to cope with zero length Unicode strings. | |
14861 | These are present in some PKCS#12 files. | |
14862 | ||
14863 | *Steve Henson* | |
14864 | ||
14865 | * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). | |
14866 | Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits | |
14867 | when writing a 32767 byte record. | |
14868 | ||
14869 | *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>* | |
14870 | ||
14871 | * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c), | |
14872 | obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`. | |
14873 | ||
14874 | (RSA objects have a reference count access to which is protected | |
14875 | by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], | |
14876 | so they are meant to be shared between threads.) | |
14877 | *Bodo Moeller, Geoff Thorpe; original patch submitted by | |
14878 | "Reddie, Steven" <Steven.Reddie@ca.com>* | |
14879 | ||
14880 | * Fix a deadlock in CRYPTO_mem_leaks(). | |
14881 | ||
14882 | *Bodo Moeller* | |
14883 | ||
14884 | * Use better test patterns in bntest. | |
14885 | ||
14886 | *Ulf Möller* | |
14887 | ||
14888 | * rand_win.c fix for Borland C. | |
14889 | ||
14890 | *Ulf Möller* | |
14891 | ||
14892 | * BN_rshift bugfix for n == 0. | |
14893 | ||
14894 | *Bodo Moeller* | |
14895 | ||
14896 | * Add a 'bctest' script that checks for some known 'bc' bugs | |
14897 | so that 'make test' does not abort just because 'bc' is broken. | |
14898 | ||
14899 | *Bodo Moeller* | |
14900 | ||
14901 | * Store verify_result within SSL_SESSION also for client side to | |
14902 | avoid potential security hole. (Re-used sessions on the client side | |
14903 | always resulted in verify_result==X509_V_OK, not using the original | |
14904 | result of the server certificate verification.) | |
14905 | ||
14906 | *Lutz Jaenicke* | |
14907 | ||
14908 | * Fix ssl3_pending: If the record in s->s3->rrec is not of type | |
14909 | SSL3_RT_APPLICATION_DATA, return 0. | |
14910 | Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. | |
14911 | ||
14912 | *Bodo Moeller* | |
14913 | ||
14914 | * Fix SSL_peek: | |
14915 | Both ssl2_peek and ssl3_peek, which were totally broken in earlier | |
14916 | releases, have been re-implemented by renaming the previous | |
14917 | implementations of ssl2_read and ssl3_read to ssl2_read_internal | |
14918 | and ssl3_read_internal, respectively, and adding 'peek' parameters | |
14919 | to them. The new ssl[23]_{read,peek} functions are calls to | |
14920 | ssl[23]_read_internal with the 'peek' flag set appropriately. | |
14921 | A 'peek' parameter has also been added to ssl3_read_bytes, which | |
14922 | does the actual work for ssl3_read_internal. | |
14923 | ||
14924 | *Bodo Moeller* | |
14925 | ||
14926 | * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling | |
14927 | the method-specific "init()" handler. Also clean up ex_data after | |
14928 | calling the method-specific "finish()" handler. Previously, this was | |
14929 | happening the other way round. | |
14930 | ||
14931 | *Geoff Thorpe* | |
14932 | ||
14933 | * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. | |
14934 | The previous value, 12, was not always sufficient for BN_mod_exp(). | |
14935 | ||
14936 | *Bodo Moeller* | |
14937 | ||
14938 | * Make sure that shared libraries get the internal name engine with | |
14939 | the full version number and not just 0. This should mark the | |
14940 | shared libraries as not backward compatible. Of course, this should | |
14941 | be changed again when we can guarantee backward binary compatibility. | |
14942 | ||
14943 | *Richard Levitte* | |
14944 | ||
14945 | * Fix typo in get_cert_by_subject() in by_dir.c | |
14946 | ||
14947 | *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>* | |
14948 | ||
14949 | * Rework the system to generate shared libraries: | |
14950 | ||
14951 | - Make note of the expected extension for the shared libraries and | |
14952 | if there is a need for symbolic links from for example libcrypto.so.0 | |
14953 | to libcrypto.so.0.9.7. There is extended info in Configure for | |
14954 | that. | |
14955 | ||
14956 | - Make as few rebuilds of the shared libraries as possible. | |
14957 | ||
14958 | - Still avoid linking the OpenSSL programs with the shared libraries. | |
14959 | ||
14960 | - When installing, install the shared libraries separately from the | |
14961 | static ones. | |
14962 | ||
14963 | *Richard Levitte* | |
14964 | ||
14965 | * Fix SSL_CTX_set_read_ahead macro to actually use its argument. | |
14966 | ||
14967 | Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new | |
14968 | and not in SSL_clear because the latter is also used by the | |
14969 | accept/connect functions; previously, the settings made by | |
14970 | SSL_set_read_ahead would be lost during the handshake. | |
14971 | ||
14972 | *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>* | |
14973 | ||
14974 | * Correct util/mkdef.pl to be selective about disabled algorithms. | |
14975 | Previously, it would create entries for disabled algorithms no | |
14976 | matter what. | |
14977 | ||
14978 | *Richard Levitte* | |
14979 | ||
14980 | * Added several new manual pages for SSL_* function. | |
14981 | ||
14982 | *Lutz Jaenicke* | |
14983 | ||
14984 | ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000] | |
14985 | ||
14986 | * In ssl23_get_client_hello, generate an error message when faced | |
14987 | with an initial SSL 3.0/TLS record that is too small to contain the | |
14988 | first two bytes of the ClientHello message, i.e. client_version. | |
14989 | (Note that this is a pathologic case that probably has never happened | |
14990 | in real life.) The previous approach was to use the version number | |
14991 | from the record header as a substitute; but our protocol choice | |
14992 | should not depend on that one because it is not authenticated | |
14993 | by the Finished messages. | |
14994 | ||
14995 | *Bodo Moeller* | |
14996 | ||
14997 | * More robust randomness gathering functions for Windows. | |
14998 | ||
14999 | *Jeffrey Altman <jaltman@columbia.edu>* | |
15000 | ||
15001 | * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is | |
15002 | not set then we don't setup the error code for issuer check errors | |
15003 | to avoid possibly overwriting other errors which the callback does | |
15004 | handle. If an application does set the flag then we assume it knows | |
15005 | what it is doing and can handle the new informational codes | |
15006 | appropriately. | |
15007 | ||
15008 | *Steve Henson* | |
15009 | ||
15010 | * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for | |
15011 | a general "ANY" type, as such it should be able to decode anything | |
15012 | including tagged types. However it didn't check the class so it would | |
15013 | wrongly interpret tagged types in the same way as their universal | |
15014 | counterpart and unknown types were just rejected. Changed so that the | |
15015 | tagged and unknown types are handled in the same way as a SEQUENCE: | |
15016 | that is the encoding is stored intact. There is also a new type | |
15017 | "V_ASN1_OTHER" which is used when the class is not universal, in this | |
15018 | case we have no idea what the actual type is so we just lump them all | |
15019 | together. | |
15020 | ||
15021 | *Steve Henson* | |
15022 | ||
15023 | * On VMS, stdout may very well lead to a file that is written to | |
15024 | in a record-oriented fashion. That means that every write() will | |
15025 | write a separate record, which will be read separately by the | |
15026 | programs trying to read from it. This can be very confusing. | |
15027 | ||
15028 | The solution is to put a BIO filter in the way that will buffer | |
15029 | text until a linefeed is reached, and then write everything a | |
15030 | line at a time, so every record written will be an actual line, | |
15031 | not chunks of lines and not (usually doesn't happen, but I've | |
15032 | seen it once) several lines in one record. BIO_f_linebuffer() is | |
15033 | the answer. | |
15034 | ||
15035 | Currently, it's a VMS-only method, because that's where it has | |
15036 | been tested well enough. | |
15037 | ||
15038 | *Richard Levitte* | |
15039 | ||
15040 | * Remove 'optimized' squaring variant in BN_mod_mul_montgomery, | |
15041 | it can return incorrect results. | |
15042 | (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, | |
15043 | but it was in 0.9.6-beta[12].) | |
15044 | ||
15045 | *Bodo Moeller* | |
15046 | ||
15047 | * Disable the check for content being present when verifying detached | |
15048 | signatures in pk7_smime.c. Some versions of Netscape (wrongly) | |
15049 | include zero length content when signing messages. | |
15050 | ||
15051 | *Steve Henson* | |
15052 | ||
15053 | * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR | |
15054 | BIO_ctrl (for BIO pairs). | |
15055 | ||
15056 | *Bodo Möller* | |
15057 | ||
15058 | * Add DSO method for VMS. | |
15059 | ||
15060 | *Richard Levitte* | |
15061 | ||
15062 | * Bug fix: Montgomery multiplication could produce results with the | |
15063 | wrong sign. | |
15064 | ||
15065 | *Ulf Möller* | |
15066 | ||
15067 | * Add RPM specification openssl.spec and modify it to build three | |
15068 | packages. The default package contains applications, application | |
15069 | documentation and run-time libraries. The devel package contains | |
15070 | include files, static libraries and function documentation. The | |
15071 | doc package contains the contents of the doc directory. The original | |
15072 | openssl.spec was provided by Damien Miller <djm@mindrot.org>. | |
15073 | ||
15074 | *Richard Levitte* | |
15075 | ||
15076 | * Add a large number of documentation files for many SSL routines. | |
15077 | ||
15078 | *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>* | |
15079 | ||
15080 | * Add a configuration entry for Sony News 4. | |
15081 | ||
15082 | *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>* | |
15083 | ||
15084 | * Don't set the two most significant bits to one when generating a | |
15085 | random number < q in the DSA library. | |
15086 | ||
15087 | *Ulf Möller* | |
15088 | ||
15089 | * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default | |
15090 | behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if | |
15091 | the underlying transport is blocking) if a handshake took place. | |
15092 | (The default behaviour is needed by applications such as s_client | |
15093 | and s_server that use select() to determine when to use SSL_read; | |
15094 | but for applications that know in advance when to expect data, it | |
15095 | just makes things more complicated.) | |
15096 | ||
15097 | *Bodo Moeller* | |
15098 | ||
15099 | * Add RAND_egd_bytes(), which gives control over the number of bytes read | |
15100 | from EGD. | |
15101 | ||
15102 | *Ben Laurie* | |
15103 | ||
15104 | * Add a few more EBCDIC conditionals that make `req` and `x509` | |
15105 | work better on such systems. | |
15106 | ||
15107 | *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>* | |
15108 | ||
15109 | * Add two demo programs for PKCS12_parse() and PKCS12_create(). | |
15110 | Update PKCS12_parse() so it copies the friendlyName and the | |
15111 | keyid to the certificates aux info. | |
15112 | ||
15113 | *Steve Henson* | |
15114 | ||
15115 | * Fix bug in PKCS7_verify() which caused an infinite loop | |
15116 | if there was more than one signature. | |
15117 | ||
15118 | *Sven Uszpelkat <su@celocom.de>* | |
15119 | ||
15120 | * Major change in util/mkdef.pl to include extra information | |
15121 | about each symbol, as well as presenting variables as well | |
15122 | as functions. This change means that there's n more need | |
15123 | to rebuild the .num files when some algorithms are excluded. | |
15124 | ||
15125 | *Richard Levitte* | |
15126 | ||
15127 | * Allow the verify time to be set by an application, | |
15128 | rather than always using the current time. | |
15129 | ||
15130 | *Steve Henson* | |
15131 | ||
15132 | * Phase 2 verify code reorganisation. The certificate | |
15133 | verify code now looks up an issuer certificate by a | |
15134 | number of criteria: subject name, authority key id | |
15135 | and key usage. It also verifies self signed certificates | |
15136 | by the same criteria. The main comparison function is | |
15137 | X509_check_issued() which performs these checks. | |
15138 | ||
15139 | Lot of changes were necessary in order to support this | |
15140 | without completely rewriting the lookup code. | |
15141 | ||
15142 | Authority and subject key identifier are now cached. | |
15143 | ||
15144 | The LHASH 'certs' is X509_STORE has now been replaced | |
15145 | by a STACK_OF(X509_OBJECT). This is mainly because an | |
15146 | LHASH can't store or retrieve multiple objects with | |
15147 | the same hash value. | |
15148 | ||
15149 | As a result various functions (which were all internal | |
15150 | use only) have changed to handle the new X509_STORE | |
15151 | structure. This will break anything that messed round | |
15152 | with X509_STORE internally. | |
15153 | ||
15154 | The functions X509_STORE_add_cert() now checks for an | |
15155 | exact match, rather than just subject name. | |
15156 | ||
15157 | The X509_STORE API doesn't directly support the retrieval | |
15158 | of multiple certificates matching a given criteria, however | |
15159 | this can be worked round by performing a lookup first | |
15160 | (which will fill the cache with candidate certificates) | |
15161 | and then examining the cache for matches. This is probably | |
15162 | the best we can do without throwing out X509_LOOKUP | |
15163 | entirely (maybe later...). | |
15164 | ||
15165 | The X509_VERIFY_CTX structure has been enhanced considerably. | |
15166 | ||
15167 | All certificate lookup operations now go via a get_issuer() | |
15168 | callback. Although this currently uses an X509_STORE it | |
15169 | can be replaced by custom lookups. This is a simple way | |
15170 | to bypass the X509_STORE hackery necessary to make this | |
15171 | work and makes it possible to use more efficient techniques | |
15172 | in future. A very simple version which uses a simple | |
15173 | STACK for its trusted certificate store is also provided | |
15174 | using X509_STORE_CTX_trusted_stack(). | |
15175 | ||
15176 | The verify_cb() and verify() callbacks now have equivalents | |
15177 | in the X509_STORE_CTX structure. | |
15178 | ||
15179 | X509_STORE_CTX also has a 'flags' field which can be used | |
15180 | to customise the verify behaviour. | |
15181 | ||
15182 | *Steve Henson* | |
15183 | ||
15184 | * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which | |
15185 | excludes S/MIME capabilities. | |
15186 | ||
15187 | *Steve Henson* | |
15188 | ||
15189 | * When a certificate request is read in keep a copy of the | |
15190 | original encoding of the signed data and use it when outputting | |
15191 | again. Signatures then use the original encoding rather than | |
15192 | a decoded, encoded version which may cause problems if the | |
15193 | request is improperly encoded. | |
15194 | ||
15195 | *Steve Henson* | |
15196 | ||
15197 | * For consistency with other BIO_puts implementations, call | |
15198 | buffer_write(b, ...) directly in buffer_puts instead of calling | |
15199 | BIO_write(b, ...). | |
15200 | ||
15201 | In BIO_puts, increment b->num_write as in BIO_write. | |
15202 | ||
15203 | *Peter.Sylvester@EdelWeb.fr* | |
15204 | ||
15205 | * Fix BN_mul_word for the case where the word is 0. (We have to use | |
15206 | BN_zero, we may not return a BIGNUM with an array consisting of | |
15207 | words set to zero.) | |
15208 | ||
15209 | *Bodo Moeller* | |
15210 | ||
15211 | * Avoid calling abort() from within the library when problems are | |
15212 | detected, except if preprocessor symbols have been defined | |
15213 | (such as REF_CHECK, BN_DEBUG etc.). | |
15214 | ||
15215 | *Bodo Moeller* | |
15216 | ||
15217 | * New openssl application 'rsautl'. This utility can be | |
15218 | used for low level RSA operations. DER public key | |
15219 | BIO/fp routines also added. | |
15220 | ||
15221 | *Steve Henson* | |
15222 | ||
15223 | * New Configure entry and patches for compiling on QNX 4. | |
15224 | ||
15225 | *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>* | |
15226 | ||
15227 | * A demo state-machine implementation was sponsored by | |
15228 | Nuron (<http://www.nuron.com/>) and is now available in | |
15229 | demos/state_machine. | |
15230 | ||
15231 | *Ben Laurie* | |
15232 | ||
15233 | * New options added to the 'dgst' utility for signature | |
15234 | generation and verification. | |
15235 | ||
15236 | *Steve Henson* | |
15237 | ||
15238 | * Unrecognized PKCS#7 content types are now handled via a | |
15239 | catch all ASN1_TYPE structure. This allows unsupported | |
15240 | types to be stored as a "blob" and an application can | |
15241 | encode and decode it manually. | |
15242 | ||
15243 | *Steve Henson* | |
15244 | ||
15245 | * Fix various signed/unsigned issues to make a_strex.c | |
15246 | compile under VC++. | |
15247 | ||
15248 | *Oscar Jacobsson <oscar.jacobsson@celocom.com>* | |
15249 | ||
15250 | * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct | |
15251 | length if passed a buffer. ASN1_INTEGER_to_BN failed | |
15252 | if passed a NULL BN and its argument was negative. | |
15253 | ||
15254 | *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>* | |
15255 | ||
15256 | * Modification to PKCS#7 encoding routines to output definite | |
15257 | length encoding. Since currently the whole structures are in | |
15258 | memory there's not real point in using indefinite length | |
15259 | constructed encoding. However if OpenSSL is compiled with | |
15260 | the flag PKCS7_INDEFINITE_ENCODING the old form is used. | |
15261 | ||
15262 | *Steve Henson* | |
15263 | ||
15264 | * Added BIO_vprintf() and BIO_vsnprintf(). | |
15265 | ||
15266 | *Richard Levitte* | |
15267 | ||
15268 | * Added more prefixes to parse for in the strings written | |
15269 | through a logging bio, to cover all the levels that are available | |
15270 | through syslog. The prefixes are now: | |
15271 | ||
15272 | PANIC, EMERG, EMR => LOG_EMERG | |
15273 | ALERT, ALR => LOG_ALERT | |
15274 | CRIT, CRI => LOG_CRIT | |
15275 | ERROR, ERR => LOG_ERR | |
15276 | WARNING, WARN, WAR => LOG_WARNING | |
15277 | NOTICE, NOTE, NOT => LOG_NOTICE | |
15278 | INFO, INF => LOG_INFO | |
15279 | DEBUG, DBG => LOG_DEBUG | |
15280 | ||
15281 | and as before, if none of those prefixes are present at the | |
15282 | beginning of the string, LOG_ERR is chosen. | |
15283 | ||
15284 | On Win32, the `LOG_*` levels are mapped according to this: | |
15285 | ||
15286 | LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE | |
15287 | LOG_WARNING => EVENTLOG_WARNING_TYPE | |
15288 | LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE | |
15289 | ||
15290 | *Richard Levitte* | |
15291 | ||
15292 | * Made it possible to reconfigure with just the configuration | |
15293 | argument "reconf" or "reconfigure". The command line arguments | |
15294 | are stored in Makefile.ssl in the variable CONFIGURE_ARGS, | |
15295 | and are retrieved from there when reconfiguring. | |
15296 | ||
15297 | *Richard Levitte* | |
15298 | ||
15299 | * MD4 implemented. | |
15300 | ||
15301 | *Assar Westerlund <assar@sics.se>, Richard Levitte* | |
15302 | ||
15303 | * Add the arguments -CAfile and -CApath to the pkcs12 utility. | |
15304 | ||
15305 | *Richard Levitte* | |
15306 | ||
15307 | * The obj_dat.pl script was messing up the sorting of object | |
15308 | names. The reason was that it compared the quoted version | |
15309 | of strings as a result "OCSP" > "OCSP Signing" because | |
15310 | " > SPACE. Changed script to store unquoted versions of | |
15311 | names and add quotes on output. It was also omitting some | |
15312 | names from the lookup table if they were given a default | |
15313 | value (that is if SN is missing it is given the same | |
15314 | value as LN and vice versa), these are now added on the | |
15315 | grounds that if an object has a name we should be able to | |
15316 | look it up. Finally added warning output when duplicate | |
15317 | short or long names are found. | |
15318 | ||
15319 | *Steve Henson* | |
15320 | ||
15321 | * Changes needed for Tandem NSK. | |
15322 | ||
15323 | *Scott Uroff <scott@xypro.com>* | |
15324 | ||
15325 | * Fix SSL 2.0 rollback checking: Due to an off-by-one error in | |
15326 | RSA_padding_check_SSLv23(), special padding was never detected | |
15327 | and thus the SSL 3.0/TLS 1.0 countermeasure against protocol | |
15328 | version rollback attacks was not effective. | |
15329 | ||
15330 | In s23_clnt.c, don't use special rollback-attack detection padding | |
15331 | (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the | |
15332 | client; similarly, in s23_srvr.c, don't do the rollback check if | |
15333 | SSL 2.0 is the only protocol enabled in the server. | |
15334 | ||
15335 | *Bodo Moeller* | |
15336 | ||
15337 | * Make it possible to get hexdumps of unprintable data with 'openssl | |
15338 | asn1parse'. By implication, the functions ASN1_parse_dump() and | |
15339 | BIO_dump_indent() are added. | |
15340 | ||
15341 | *Richard Levitte* | |
15342 | ||
15343 | * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() | |
15344 | these print out strings and name structures based on various | |
15345 | flags including RFC2253 support and proper handling of | |
15346 | multibyte characters. Added options to the 'x509' utility | |
15347 | to allow the various flags to be set. | |
15348 | ||
15349 | *Steve Henson* | |
15350 | ||
15351 | * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. | |
15352 | Also change the functions X509_cmp_current_time() and | |
15353 | X509_gmtime_adj() work with an ASN1_TIME structure, | |
15354 | this will enable certificates using GeneralizedTime in validity | |
15355 | dates to be checked. | |
15356 | ||
15357 | *Steve Henson* | |
15358 | ||
15359 | * Make the NEG_PUBKEY_BUG code (which tolerates invalid | |
15360 | negative public key encodings) on by default, | |
15361 | NO_NEG_PUBKEY_BUG can be set to disable it. | |
15362 | ||
15363 | *Steve Henson* | |
15364 | ||
15365 | * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT | |
15366 | content octets. An i2c_ASN1_OBJECT is unnecessary because | |
15367 | the encoding can be trivially obtained from the structure. | |
15368 | ||
15369 | *Steve Henson* | |
15370 | ||
15371 | * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`), | |
15372 | not read locks (`CRYPTO_r_[un]lock`). | |
15373 | ||
15374 | *Bodo Moeller* | |
15375 | ||
15376 | * A first attempt at creating official support for shared | |
15377 | libraries through configuration. I've kept it so the | |
15378 | default is static libraries only, and the OpenSSL programs | |
15379 | are always statically linked for now, but there are | |
15380 | preparations for dynamic linking in place. | |
15381 | This has been tested on Linux and Tru64. | |
15382 | ||
15383 | *Richard Levitte* | |
15384 | ||
15385 | * Randomness polling function for Win9x, as described in: | |
15386 | Peter Gutmann, Software Generation of Practically Strong | |
15387 | Random Numbers. | |
15388 | ||
15389 | *Ulf Möller* | |
15390 | ||
15391 | * Fix so PRNG is seeded in req if using an already existing | |
15392 | DSA key. | |
15393 | ||
15394 | *Steve Henson* | |
15395 | ||
15396 | * New options to smime application. -inform and -outform | |
15397 | allow alternative formats for the S/MIME message including | |
15398 | PEM and DER. The -content option allows the content to be | |
15399 | specified separately. This should allow things like Netscape | |
15400 | form signing output easier to verify. | |
15401 | ||
15402 | *Steve Henson* | |
15403 | ||
15404 | * Fix the ASN1 encoding of tags using the 'long form'. | |
15405 | ||
15406 | *Steve Henson* | |
15407 | ||
15408 | * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT | |
15409 | STRING types. These convert content octets to and from the | |
15410 | underlying type. The actual tag and length octets are | |
15411 | already assumed to have been read in and checked. These | |
15412 | are needed because all other string types have virtually | |
15413 | identical handling apart from the tag. By having versions | |
15414 | of the ASN1 functions that just operate on content octets | |
15415 | IMPLICIT tagging can be handled properly. It also allows | |
15416 | the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED | |
15417 | and ASN1_INTEGER are identical apart from the tag. | |
15418 | ||
15419 | *Steve Henson* | |
15420 | ||
15421 | * Change the handling of OID objects as follows: | |
15422 | ||
15423 | - New object identifiers are inserted in objects.txt, following | |
15424 | the syntax given in [crypto/objects/README.md](crypto/objects/README.md). | |
15425 | - objects.pl is used to process obj_mac.num and create a new | |
15426 | obj_mac.h. | |
15427 | - obj_dat.pl is used to create a new obj_dat.h, using the data in | |
15428 | obj_mac.h. | |
15429 | ||
15430 | This is currently kind of a hack, and the perl code in objects.pl | |
15431 | isn't very elegant, but it works as I intended. The simplest way | |
15432 | to check that it worked correctly is to look in obj_dat.h and | |
15433 | check the array nid_objs and make sure the objects haven't moved | |
15434 | around (this is important!). Additions are OK, as well as | |
15435 | consistent name changes. | |
15436 | ||
15437 | *Richard Levitte* | |
15438 | ||
15439 | * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). | |
15440 | ||
15441 | *Bodo Moeller* | |
15442 | ||
15443 | * Addition of the command line parameter '-rand file' to 'openssl req'. | |
15444 | The given file adds to whatever has already been seeded into the | |
15445 | random pool through the RANDFILE configuration file option or | |
15446 | environment variable, or the default random state file. | |
15447 | ||
15448 | *Richard Levitte* | |
15449 | ||
15450 | * mkstack.pl now sorts each macro group into lexical order. | |
15451 | Previously the output order depended on the order the files | |
15452 | appeared in the directory, resulting in needless rewriting | |
15453 | of safestack.h . | |
15454 | ||
15455 | *Steve Henson* | |
15456 | ||
15457 | * Patches to make OpenSSL compile under Win32 again. Mostly | |
15458 | work arounds for the VC++ problem that it treats func() as | |
15459 | func(void). Also stripped out the parts of mkdef.pl that | |
15460 | added extra typesafe functions: these no longer exist. | |
15461 | ||
15462 | *Steve Henson* | |
15463 | ||
15464 | * Reorganisation of the stack code. The macros are now all | |
15465 | collected in safestack.h . Each macro is defined in terms of | |
15466 | a "stack macro" of the form `SKM_<name>(type, a, b)`. The | |
15467 | DEBUG_SAFESTACK is now handled in terms of function casts, | |
15468 | this has the advantage of retaining type safety without the | |
15469 | use of additional functions. If DEBUG_SAFESTACK is not defined | |
15470 | then the non typesafe macros are used instead. Also modified the | |
15471 | mkstack.pl script to handle the new form. Needs testing to see | |
15472 | if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK | |
15473 | the default if no major problems. Similar behaviour for ASN1_SET_OF | |
15474 | and PKCS12_STACK_OF. | |
15475 | ||
15476 | *Steve Henson* | |
15477 | ||
15478 | * When some versions of IIS use the 'NET' form of private key the | |
15479 | key derivation algorithm is different. Normally MD5(password) is | |
15480 | used as a 128 bit RC4 key. In the modified case | |
15481 | MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some | |
15482 | new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same | |
15483 | as the old Netscape_RSA functions except they have an additional | |
15484 | 'sgckey' parameter which uses the modified algorithm. Also added | |
15485 | an -sgckey command line option to the rsa utility. Thanks to | |
15486 | Adrian Peck <bertie@ncipher.com> for posting details of the modified | |
15487 | algorithm to openssl-dev. | |
15488 | ||
15489 | *Steve Henson* | |
15490 | ||
15491 | * The evp_local.h macros were using 'c.##kname' which resulted in | |
15492 | invalid expansion on some systems (SCO 5.0.5 for example). | |
15493 | Corrected to 'c.kname'. | |
15494 | ||
15495 | *Phillip Porch <root@theporch.com>* | |
15496 | ||
15497 | * New X509_get1_email() and X509_REQ_get1_email() functions that return | |
15498 | a STACK of email addresses from a certificate or request, these look | |
15499 | in the subject name and the subject alternative name extensions and | |
15500 | omit any duplicate addresses. | |
15501 | ||
15502 | *Steve Henson* | |
15503 | ||
15504 | * Re-implement BN_mod_exp2_mont using independent (and larger) windows. | |
15505 | This makes DSA verification about 2 % faster. | |
15506 | ||
15507 | *Bodo Moeller* | |
15508 | ||
15509 | * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5 | |
15510 | (meaning that now 2^5 values will be precomputed, which is only 4 KB | |
15511 | plus overhead for 1024 bit moduli). | |
15512 | This makes exponentiations about 0.5 % faster for 1024 bit | |
15513 | exponents (as measured by "openssl speed rsa2048"). | |
15514 | ||
15515 | *Bodo Moeller* | |
15516 | ||
15517 | * Rename memory handling macros to avoid conflicts with other | |
15518 | software: | |
15519 | Malloc => OPENSSL_malloc | |
15520 | Malloc_locked => OPENSSL_malloc_locked | |
15521 | Realloc => OPENSSL_realloc | |
15522 | Free => OPENSSL_free | |
15523 | ||
15524 | *Richard Levitte* | |
15525 | ||
15526 | * New function BN_mod_exp_mont_word for small bases (roughly 15% | |
15527 | faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). | |
15528 | ||
15529 | *Bodo Moeller* | |
15530 | ||
15531 | * CygWin32 support. | |
15532 | ||
15533 | *John Jarvie <jjarvie@newsguy.com>* | |
15534 | ||
15535 | * The type-safe stack code has been rejigged. It is now only compiled | |
15536 | in when OpenSSL is configured with the DEBUG_SAFESTACK option and | |
15537 | by default all type-specific stack functions are "#define"d back to | |
15538 | standard stack functions. This results in more streamlined output | |
15539 | but retains the type-safety checking possibilities of the original | |
15540 | approach. | |
15541 | ||
15542 | *Geoff Thorpe* | |
15543 | ||
15544 | * The STACK code has been cleaned up, and certain type declarations | |
15545 | that didn't make a lot of sense have been brought in line. This has | |
15546 | also involved a cleanup of sorts in safestack.h to more correctly | |
15547 | map type-safe stack functions onto their plain stack counterparts. | |
15548 | This work has also resulted in a variety of "const"ifications of | |
15549 | lots of the code, especially `_cmp` operations which should normally | |
15550 | be prototyped with "const" parameters anyway. | |
15551 | ||
15552 | *Geoff Thorpe* | |
15553 | ||
15554 | * When generating bytes for the first time in md_rand.c, 'stir the pool' | |
15555 | by seeding with STATE_SIZE dummy bytes (with zero entropy count). | |
15556 | (The PRNG state consists of two parts, the large pool 'state' and 'md', | |
15557 | where all of 'md' is used each time the PRNG is used, but 'state' | |
15558 | is used only indexed by a cyclic counter. As entropy may not be | |
15559 | well distributed from the beginning, 'md' is important as a | |
15560 | chaining variable. However, the output function chains only half | |
15561 | of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains | |
15562 | all of 'md', and seeding with STATE_SIZE dummy bytes will result | |
15563 | in all of 'state' being rewritten, with the new values depending | |
15564 | on virtually all of 'md'. This overcomes the 80 bit limitation.) | |
15565 | ||
15566 | *Bodo Moeller* | |
15567 | ||
15568 | * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when | |
15569 | the handshake is continued after ssl_verify_cert_chain(); | |
15570 | otherwise, if SSL_VERIFY_NONE is set, remaining error codes | |
15571 | can lead to 'unexplainable' connection aborts later. | |
15572 | ||
15573 | *Bodo Moeller; problem tracked down by Lutz Jaenicke* | |
15574 | ||
15575 | * Major EVP API cipher revision. | |
15576 | Add hooks for extra EVP features. This allows various cipher | |
15577 | parameters to be set in the EVP interface. Support added for variable | |
15578 | key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and | |
15579 | setting of RC2 and RC5 parameters. | |
15580 | ||
15581 | Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length | |
15582 | ciphers. | |
15583 | ||
15584 | Remove lots of duplicated code from the EVP library. For example *every* | |
15585 | cipher init() function handles the 'iv' in the same way according to the | |
15586 | cipher mode. They also all do nothing if the 'key' parameter is NULL and | |
15587 | for CFB and OFB modes they zero ctx->num. | |
15588 | ||
15589 | New functionality allows removal of S/MIME code RC2 hack. | |
15590 | ||
15591 | Most of the routines have the same form and so can be declared in terms | |
15592 | of macros. | |
15593 | ||
15594 | By shifting this to the top level EVP_CipherInit() it can be removed from | |
15595 | all individual ciphers. If the cipher wants to handle IVs or keys | |
15596 | differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT | |
15597 | flags. | |
15598 | ||
15599 | Change lots of functions like EVP_EncryptUpdate() to now return a | |
15600 | value: although software versions of the algorithms cannot fail | |
15601 | any installed hardware versions can. | |
15602 | ||
15603 | *Steve Henson* | |
15604 | ||
15605 | * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if | |
15606 | this option is set, tolerate broken clients that send the negotiated | |
15607 | protocol version number instead of the requested protocol version | |
15608 | number. | |
15609 | ||
15610 | *Bodo Moeller* | |
15611 | ||
15612 | * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag; | |
15613 | i.e. non-zero for export ciphersuites, zero otherwise. | |
15614 | Previous versions had this flag inverted, inconsistent with | |
15615 | rsa_tmp_cb (..._TMP_RSA_CB). | |
15616 | ||
15617 | *Bodo Moeller; problem reported by Amit Chopra* | |
15618 | ||
15619 | * Add missing DSA library text string. Work around for some IIS | |
15620 | key files with invalid SEQUENCE encoding. | |
15621 | ||
15622 | *Steve Henson* | |
15623 | ||
15624 | * Add a document (doc/standards.txt) that list all kinds of standards | |
15625 | and so on that are implemented in OpenSSL. | |
15626 | ||
15627 | *Richard Levitte* | |
15628 | ||
15629 | * Enhance c_rehash script. Old version would mishandle certificates | |
15630 | with the same subject name hash and wouldn't handle CRLs at all. | |
15631 | Added -fingerprint option to crl utility, to support new c_rehash | |
15632 | features. | |
15633 | ||
15634 | *Steve Henson* | |
15635 | ||
15636 | * Eliminate non-ANSI declarations in crypto.h and stack.h. | |
15637 | ||
15638 | *Ulf Möller* | |
15639 | ||
15640 | * Fix for SSL server purpose checking. Server checking was | |
15641 | rejecting certificates which had extended key usage present | |
15642 | but no ssl client purpose. | |
15643 | ||
15644 | *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>* | |
15645 | ||
15646 | * Make PKCS#12 code work with no password. The PKCS#12 spec | |
15647 | is a little unclear about how a blank password is handled. | |
15648 | Since the password in encoded as a BMPString with terminating | |
15649 | double NULL a zero length password would end up as just the | |
15650 | double NULL. However no password at all is different and is | |
15651 | handled differently in the PKCS#12 key generation code. NS | |
15652 | treats a blank password as zero length. MSIE treats it as no | |
15653 | password on export: but it will try both on import. We now do | |
15654 | the same: PKCS12_parse() tries zero length and no password if | |
15655 | the password is set to "" or NULL (NULL is now a valid password: | |
15656 | it wasn't before) as does the pkcs12 application. | |
15657 | ||
15658 | *Steve Henson* | |
15659 | ||
15660 | * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use | |
15661 | perror when PEM_read_bio_X509_REQ fails, the error message must | |
15662 | be obtained from the error queue. | |
15663 | ||
15664 | *Bodo Moeller* | |
15665 | ||
15666 | * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing | |
15667 | it in ERR_remove_state if appropriate, and change ERR_get_state | |
15668 | accordingly to avoid race conditions (this is necessary because | |
15669 | thread_hash is no longer constant once set). | |
15670 | ||
15671 | *Bodo Moeller* | |
15672 | ||
15673 | * Bugfix for linux-elf makefile.one. | |
15674 | ||
15675 | *Ulf Möller* | |
15676 | ||
15677 | * RSA_get_default_method() will now cause a default | |
15678 | RSA_METHOD to be chosen if one doesn't exist already. | |
15679 | Previously this was only set during a call to RSA_new() | |
15680 | or RSA_new_method(NULL) meaning it was possible for | |
15681 | RSA_get_default_method() to return NULL. | |
15682 | ||
15683 | *Geoff Thorpe* | |
15684 | ||
15685 | * Added native name translation to the existing DSO code | |
15686 | that will convert (if the flag to do so is set) filenames | |
15687 | that are sufficiently small and have no path information | |
15688 | into a canonical native form. Eg. "blah" converted to | |
15689 | "libblah.so" or "blah.dll" etc. | |
15690 | ||
15691 | *Geoff Thorpe* | |
15692 | ||
15693 | * New function ERR_error_string_n(e, buf, len) which is like | |
15694 | ERR_error_string(e, buf), but writes at most 'len' bytes | |
15695 | including the 0 terminator. For ERR_error_string_n, 'buf' | |
15696 | may not be NULL. | |
15697 | ||
15698 | *Damien Miller <djm@mindrot.org>, Bodo Moeller* | |
15699 | ||
15700 | * CONF library reworked to become more general. A new CONF | |
15701 | configuration file reader "class" is implemented as well as a | |
15702 | new functions (`NCONF_*`, for "New CONF") to handle it. The now | |
15703 | old `CONF_*` functions are still there, but are reimplemented to | |
15704 | work in terms of the new functions. Also, a set of functions | |
15705 | to handle the internal storage of the configuration data is | |
15706 | provided to make it easier to write new configuration file | |
15707 | reader "classes" (I can definitely see something reading a | |
15708 | configuration file in XML format, for example), called `_CONF_*`, | |
15709 | or "the configuration storage API"... | |
15710 | ||
15711 | The new configuration file reading functions are: | |
15712 | ||
15713 | NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, | |
15714 | NCONF_get_section, NCONF_get_string, NCONF_get_numbre | |
15715 | ||
15716 | NCONF_default, NCONF_WIN32 | |
15717 | ||
15718 | NCONF_dump_fp, NCONF_dump_bio | |
15719 | ||
15720 | NCONF_default and NCONF_WIN32 are method (or "class") choosers, | |
15721 | NCONF_new creates a new CONF object. This works in the same way | |
15722 | as other interfaces in OpenSSL, like the BIO interface. | |
15723 | `NCONF_dump_*` dump the internal storage of the configuration file, | |
15724 | which is useful for debugging. All other functions take the same | |
15725 | arguments as the old `CONF_*` functions with the exception of the | |
15726 | first that must be a `CONF *` instead of a `LHASH *`. | |
15727 | ||
15728 | To make it easier to use the new classes with the old `CONF_*` functions, | |
15729 | the function CONF_set_default_method is provided. | |
15730 | ||
15731 | *Richard Levitte* | |
15732 | ||
15733 | * Add '-tls1' option to 'openssl ciphers', which was already | |
15734 | mentioned in the documentation but had not been implemented. | |
15735 | (This option is not yet really useful because even the additional | |
15736 | experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) | |
15737 | ||
15738 | *Bodo Moeller* | |
15739 | ||
15740 | * Initial DSO code added into libcrypto for letting OpenSSL (and | |
15741 | OpenSSL-based applications) load shared libraries and bind to | |
15742 | them in a portable way. | |
15743 | ||
15744 | *Geoff Thorpe, with contributions from Richard Levitte* | |
15745 | ||
15746 | ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000] | |
15747 | ||
15748 | * Make sure _lrotl and _lrotr are only used with MSVC. | |
15749 | ||
15750 | * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status | |
15751 | (the default implementation of RAND_status). | |
15752 | ||
15753 | * Rename openssl x509 option '-crlext', which was added in 0.9.5, | |
15754 | to '-clrext' (= clear extensions), as intended and documented. | |
15755 | *Bodo Moeller; inconsistency pointed out by Michael Attili | |
15756 | <attili@amaxo.com>* | |
15757 | ||
15758 | * Fix for HMAC. It wasn't zeroing the rest of the block if the key length | |
15759 | was larger than the MD block size. | |
15760 | ||
15761 | *Steve Henson, pointed out by Yost William <YostW@tce.com>* | |
15762 | ||
15763 | * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument | |
15764 | fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() | |
15765 | using the passed key: if the passed key was a private key the result | |
15766 | of X509_print(), for example, would be to print out all the private key | |
15767 | components. | |
15768 | ||
15769 | *Steve Henson* | |
15770 | ||
15771 | * des_quad_cksum() byte order bug fix. | |
15772 | *Ulf Möller, using the problem description in krb4-0.9.7, where | |
15773 | the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>* | |
15774 | ||
15775 | * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly | |
15776 | discouraged. | |
15777 | ||
15778 | *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>* | |
15779 | ||
15780 | * For easily testing in shell scripts whether some command | |
15781 | 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' | |
15782 | returns with exit code 0 iff no command of the given name is available. | |
15783 | 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, | |
15784 | the output goes to stdout and nothing is printed to stderr. | |
15785 | Additional arguments are always ignored. | |
15786 | ||
15787 | Since for each cipher there is a command of the same name, | |
15788 | the 'no-cipher' compilation switches can be tested this way. | |
15789 | ||
15790 | ('openssl no-XXX' is not able to detect pseudo-commands such | |
15791 | as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) | |
15792 | ||
15793 | *Bodo Moeller* | |
15794 | ||
15795 | * Update test suite so that 'make test' succeeds in 'no-rsa' configuration. | |
15796 | ||
15797 | *Bodo Moeller* | |
15798 | ||
15799 | * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE | |
15800 | is set; it will be thrown away anyway because each handshake creates | |
15801 | its own key. | |
15802 | ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition | |
15803 | to parameters -- in previous versions (since OpenSSL 0.9.3) the | |
15804 | 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning | |
15805 | you effectively got SSL_OP_SINGLE_DH_USE when using this macro. | |
15806 | ||
15807 | *Bodo Moeller* | |
15808 | ||
15809 | * New s_client option -ign_eof: EOF at stdin is ignored, and | |
15810 | 'Q' and 'R' lose their special meanings (quit/renegotiate). | |
15811 | This is part of what -quiet does; unlike -quiet, -ign_eof | |
15812 | does not suppress any output. | |
15813 | ||
15814 | *Richard Levitte* | |
15815 | ||
15816 | * Add compatibility options to the purpose and trust code. The | |
15817 | purpose X509_PURPOSE_ANY is "any purpose" which automatically | |
15818 | accepts a certificate or CA, this was the previous behaviour, | |
15819 | with all the associated security issues. | |
15820 | ||
15821 | X509_TRUST_COMPAT is the old trust behaviour: only and | |
15822 | automatically trust self signed roots in certificate store. A | |
15823 | new trust setting X509_TRUST_DEFAULT is used to specify that | |
15824 | a purpose has no associated trust setting and it should instead | |
15825 | use the value in the default purpose. | |
15826 | ||
15827 | *Steve Henson* | |
15828 | ||
15829 | * Fix the PKCS#8 DSA private key code so it decodes keys again | |
15830 | and fix a memory leak. | |
15831 | ||
15832 | *Steve Henson* | |
15833 | ||
15834 | * In util/mkerr.pl (which implements 'make errors'), preserve | |
15835 | reason strings from the previous version of the .c file, as | |
15836 | the default to have only downcase letters (and digits) in | |
15837 | automatically generated reasons codes is not always appropriate. | |
15838 | ||
15839 | *Bodo Moeller* | |
15840 | ||
15841 | * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table | |
15842 | using strerror. Previously, ERR_reason_error_string() returned | |
15843 | library names as reason strings for SYSerr; but SYSerr is a special | |
15844 | case where small numbers are errno values, not library numbers. | |
15845 | ||
15846 | *Bodo Moeller* | |
15847 | ||
15848 | * Add '-dsaparam' option to 'openssl dhparam' application. This | |
15849 | converts DSA parameters into DH parameters. (When creating parameters, | |
15850 | DSA_generate_parameters is used.) | |
15851 | ||
15852 | *Bodo Moeller* | |
15853 | ||
15854 | * Include 'length' (recommended exponent length) in C code generated | |
15855 | by 'openssl dhparam -C'. | |
15856 | ||
15857 | *Bodo Moeller* | |
15858 | ||
15859 | * The second argument to set_label in perlasm was already being used | |
15860 | so couldn't be used as a "file scope" flag. Moved to third argument | |
15861 | which was free. | |
15862 | ||
15863 | *Steve Henson* | |
15864 | ||
15865 | * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes | |
15866 | instead of RAND_bytes for encryption IVs and salts. | |
15867 | ||
15868 | *Bodo Moeller* | |
15869 | ||
15870 | * Include RAND_status() into RAND_METHOD instead of implementing | |
15871 | it only for md_rand.c Otherwise replacing the PRNG by calling | |
15872 | RAND_set_rand_method would be impossible. | |
15873 | ||
15874 | *Bodo Moeller* | |
15875 | ||
15876 | * Don't let DSA_generate_key() enter an infinite loop if the random | |
15877 | number generation fails. | |
15878 | ||
15879 | *Bodo Moeller* | |
15880 | ||
15881 | * New 'rand' application for creating pseudo-random output. | |
15882 | ||
15883 | *Bodo Moeller* | |
15884 | ||
15885 | * Added configuration support for Linux/IA64 | |
15886 | ||
15887 | *Rolf Haberrecker <rolf@suse.de>* | |
15888 | ||
15889 | * Assembler module support for Mingw32. | |
15890 | ||
15891 | *Ulf Möller* | |
15892 | ||
15893 | * Shared library support for HPUX (in shlib/). | |
15894 | ||
15895 | *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous* | |
15896 | ||
15897 | * Shared library support for Solaris gcc. | |
15898 | ||
15899 | *Lutz Behnke <behnke@trustcenter.de>* | |
15900 | ||
15901 | ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000] | |
15902 | ||
15903 | * PKCS7_encrypt() was adding text MIME headers twice because they | |
15904 | were added manually and by SMIME_crlf_copy(). | |
15905 | ||
15906 | *Steve Henson* | |
15907 | ||
15908 | * In bntest.c don't call BN_rand with zero bits argument. | |
15909 | ||
15910 | *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>* | |
15911 | ||
15912 | * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] | |
15913 | case was implemented. This caused BN_div_recp() to fail occasionally. | |
15914 | ||
15915 | *Ulf Möller* | |
15916 | ||
15917 | * Add an optional second argument to the set_label() in the perl | |
15918 | assembly language builder. If this argument exists and is set | |
15919 | to 1 it signals that the assembler should use a symbol whose | |
15920 | scope is the entire file, not just the current function. This | |
15921 | is needed with MASM which uses the format label:: for this scope. | |
15922 | ||
15923 | *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>* | |
15924 | ||
15925 | * Change the ASN1 types so they are typedefs by default. Before | |
15926 | almost all types were #define'd to ASN1_STRING which was causing | |
15927 | STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) | |
15928 | for example. | |
15929 | ||
15930 | *Steve Henson* | |
15931 | ||
15932 | * Change names of new functions to the new get1/get0 naming | |
15933 | convention: After 'get1', the caller owns a reference count | |
15934 | and has to call `..._free`; 'get0' returns a pointer to some | |
15935 | data structure without incrementing reference counters. | |
15936 | (Some of the existing 'get' functions increment a reference | |
15937 | counter, some don't.) | |
15938 | Similarly, 'set1' and 'add1' functions increase reference | |
15939 | counters or duplicate objects. | |
15940 | ||
15941 | *Steve Henson* | |
15942 | ||
15943 | * Allow for the possibility of temp RSA key generation failure: | |
15944 | the code used to assume it always worked and crashed on failure. | |
15945 | ||
15946 | *Steve Henson* | |
15947 | ||
15948 | * Fix potential buffer overrun problem in BIO_printf(). | |
15949 | *Ulf Möller, using public domain code by Patrick Powell; problem | |
15950 | pointed out by David Sacerdote <das33@cornell.edu>* | |
15951 | ||
15952 | * Support EGD <http://www.lothar.com/tech/crypto/>. New functions | |
15953 | RAND_egd() and RAND_status(). In the command line application, | |
15954 | the EGD socket can be specified like a seed file using RANDFILE | |
15955 | or -rand. | |
15956 | ||
15957 | *Ulf Möller* | |
15958 | ||
15959 | * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. | |
15960 | Some CAs (e.g. Verisign) distribute certificates in this form. | |
15961 | ||
15962 | *Steve Henson* | |
15963 | ||
15964 | * Remove the SSL_ALLOW_ADH compile option and set the default cipher | |
15965 | list to exclude them. This means that no special compilation option | |
15966 | is needed to use anonymous DH: it just needs to be included in the | |
15967 | cipher list. | |
15968 | ||
15969 | *Steve Henson* | |
15970 | ||
15971 | * Change the EVP_MD_CTX_type macro so its meaning consistent with | |
15972 | EVP_MD_type. The old functionality is available in a new macro called | |
15973 | EVP_MD_md(). Change code that uses it and update docs. | |
15974 | ||
15975 | *Steve Henson* | |
15976 | ||
15977 | * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions | |
15978 | where the `void *` argument is replaced by a function pointer argument. | |
15979 | Previously `void *` was abused to point to functions, which works on | |
15980 | many platforms, but is not correct. As these functions are usually | |
15981 | called by macros defined in OpenSSL header files, most source code | |
15982 | should work without changes. | |
15983 | ||
15984 | *Richard Levitte* | |
15985 | ||
15986 | * `<openssl/opensslconf.h>` (which is created by Configure) now contains | |
15987 | sections with information on -D... compiler switches used for | |
15988 | compiling the library so that applications can see them. To enable | |
15989 | one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES` | |
15990 | must be defined. E.g., | |
15991 | #define OPENSSL_ALGORITHM_DEFINES | |
15992 | #include <openssl/opensslconf.h> | |
15993 | defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc. | |
15994 | ||
15995 | *Richard Levitte, Ulf and Bodo Möller* | |
15996 | ||
15997 | * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS | |
15998 | record layer. | |
15999 | ||
16000 | *Bodo Moeller* | |
16001 | ||
16002 | * Change the 'other' type in certificate aux info to a STACK_OF | |
16003 | X509_ALGOR. Although not an AlgorithmIdentifier as such it has | |
16004 | the required ASN1 format: arbitrary types determined by an OID. | |
16005 | ||
16006 | *Steve Henson* | |
16007 | ||
16008 | * Add some PEM_write_X509_REQ_NEW() functions and a command line | |
16009 | argument to 'req'. This is not because the function is newer or | |
16010 | better than others it just uses the work 'NEW' in the certificate | |
16011 | request header lines. Some software needs this. | |
16012 | ||
16013 | *Steve Henson* | |
16014 | ||
16015 | * Reorganise password command line arguments: now passwords can be | |
16016 | obtained from various sources. Delete the PEM_cb function and make | |
16017 | it the default behaviour: i.e. if the callback is NULL and the | |
16018 | usrdata argument is not NULL interpret it as a null terminated pass | |
16019 | phrase. If usrdata and the callback are NULL then the pass phrase | |
16020 | is prompted for as usual. | |
16021 | ||
16022 | *Steve Henson* | |
16023 | ||
16024 | * Add support for the Compaq Atalla crypto accelerator. If it is installed, | |
16025 | the support is automatically enabled. The resulting binaries will | |
16026 | autodetect the card and use it if present. | |
16027 | ||
16028 | *Ben Laurie and Compaq Inc.* | |
16029 | ||
16030 | * Work around for Netscape hang bug. This sends certificate request | |
16031 | and server done in one record. Since this is perfectly legal in the | |
16032 | SSL/TLS protocol it isn't a "bug" option and is on by default. See | |
16033 | the bugs/SSLv3 entry for more info. | |
16034 | ||
16035 | *Steve Henson* | |
16036 | ||
16037 | * HP-UX tune-up: new unified configs, HP C compiler bug workaround. | |
16038 | ||
16039 | *Andy Polyakov* | |
16040 | ||
16041 | * Add -rand argument to smime and pkcs12 applications and read/write | |
16042 | of seed file. | |
16043 | ||
16044 | *Steve Henson* | |
16045 | ||
16046 | * New 'passwd' tool for crypt(3) and apr1 password hashes. | |
16047 | ||
16048 | *Bodo Moeller* | |
16049 | ||
16050 | * Add command line password options to the remaining applications. | |
16051 | ||
16052 | *Steve Henson* | |
16053 | ||
16054 | * Bug fix for BN_div_recp() for numerators with an even number of | |
16055 | bits. | |
16056 | ||
16057 | *Ulf Möller* | |
16058 | ||
16059 | * More tests in bntest.c, and changed test_bn output. | |
16060 | ||
16061 | *Ulf Möller* | |
16062 | ||
16063 | * ./config recognizes MacOS X now. | |
16064 | ||
16065 | *Andy Polyakov* | |
16066 | ||
16067 | * Bug fix for BN_div() when the first words of num and divisor are | |
16068 | equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`. | |
16069 | ||
16070 | *Ulf Möller* | |
16071 | ||
16072 | * Add support for various broken PKCS#8 formats, and command line | |
16073 | options to produce them. | |
16074 | ||
16075 | *Steve Henson* | |
16076 | ||
16077 | * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to | |
16078 | get temporary BIGNUMs from a BN_CTX. | |
16079 | ||
16080 | *Ulf Möller* | |
16081 | ||
16082 | * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() | |
16083 | for p == 0. | |
16084 | ||
16085 | *Ulf Möller* | |
16086 | ||
16087 | * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and | |
16088 | include a #define from the old name to the new. The original intent | |
16089 | was that statically linked binaries could for example just call | |
16090 | SSLeay_add_all_ciphers() to just add ciphers to the table and not | |
16091 | link with digests. This never worked because SSLeay_add_all_digests() | |
16092 | and SSLeay_add_all_ciphers() were in the same source file so calling | |
16093 | one would link with the other. They are now in separate source files. | |
16094 | ||
16095 | *Steve Henson* | |
16096 | ||
16097 | * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. | |
16098 | ||
16099 | *Steve Henson* | |
16100 | ||
16101 | * Use a less unusual form of the Miller-Rabin primality test (it used | |
16102 | a binary algorithm for exponentiation integrated into the Miller-Rabin | |
16103 | loop, our standard modexp algorithms are faster). | |
16104 | ||
16105 | *Bodo Moeller* | |
16106 | ||
16107 | * Support for the EBCDIC character set completed. | |
16108 | ||
16109 | *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>* | |
16110 | ||
16111 | * Source code cleanups: use const where appropriate, eliminate casts, | |
16112 | use `void *` instead of `char *` in lhash. | |
16113 | ||
16114 | *Ulf Möller* | |
16115 | ||
16116 | * Bugfix: ssl3_send_server_key_exchange was not restartable | |
16117 | (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of | |
16118 | this the server could overwrite ephemeral keys that the client | |
16119 | has already seen). | |
16120 | ||
16121 | *Bodo Moeller* | |
16122 | ||
16123 | * Turn DSA_is_prime into a macro that calls BN_is_prime, | |
16124 | using 50 iterations of the Rabin-Miller test. | |
16125 | ||
16126 | DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 | |
16127 | iterations of the Rabin-Miller test as required by the appendix | |
16128 | to FIPS PUB 186[-1]) instead of DSA_is_prime. | |
16129 | As BN_is_prime_fasttest includes trial division, DSA parameter | |
16130 | generation becomes much faster. | |
16131 | ||
16132 | This implies a change for the callback functions in DSA_is_prime | |
16133 | and DSA_generate_parameters: The callback function is called once | |
16134 | for each positive witness in the Rabin-Miller test, not just | |
16135 | occasionally in the inner loop; and the parameters to the | |
16136 | callback function now provide an iteration count for the outer | |
16137 | loop rather than for the current invocation of the inner loop. | |
16138 | DSA_generate_parameters additionally can call the callback | |
16139 | function with an 'iteration count' of -1, meaning that a | |
16140 | candidate has passed the trial division test (when q is generated | |
16141 | from an application-provided seed, trial division is skipped). | |
16142 | ||
16143 | *Bodo Moeller* | |
16144 | ||
16145 | * New function BN_is_prime_fasttest that optionally does trial | |
16146 | division before starting the Rabin-Miller test and has | |
16147 | an additional BN_CTX * argument (whereas BN_is_prime always | |
16148 | has to allocate at least one BN_CTX). | |
16149 | 'callback(1, -1, cb_arg)' is called when a number has passed the | |
16150 | trial division stage. | |
16151 | ||
16152 | *Bodo Moeller* | |
16153 | ||
16154 | * Fix for bug in CRL encoding. The validity dates weren't being handled | |
16155 | as ASN1_TIME. | |
16156 | ||
16157 | *Steve Henson* | |
16158 | ||
16159 | * New -pkcs12 option to CA.pl script to write out a PKCS#12 file. | |
16160 | ||
16161 | *Steve Henson* | |
16162 | ||
16163 | * New function BN_pseudo_rand(). | |
16164 | ||
16165 | *Ulf Möller* | |
16166 | ||
16167 | * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) | |
16168 | bignum version of BN_from_montgomery() with the working code from | |
16169 | SSLeay 0.9.0 (the word based version is faster anyway), and clean up | |
16170 | the comments. | |
16171 | ||
16172 | *Ulf Möller* | |
16173 | ||
16174 | * Avoid a race condition in s2_clnt.c (function get_server_hello) that | |
16175 | made it impossible to use the same SSL_SESSION data structure in | |
16176 | SSL2 clients in multiple threads. | |
16177 | ||
16178 | *Bodo Moeller* | |
16179 | ||
16180 | * The return value of RAND_load_file() no longer counts bytes obtained | |
16181 | by stat(). RAND_load_file(..., -1) is new and uses the complete file | |
16182 | to seed the PRNG (previously an explicit byte count was required). | |
16183 | ||
16184 | *Ulf Möller, Bodo Möller* | |
16185 | ||
16186 | * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes | |
16187 | used `char *` instead of `void *` and had casts all over the place. | |
16188 | ||
16189 | *Steve Henson* | |
16190 | ||
16191 | * Make BN_generate_prime() return NULL on error if ret!=NULL. | |
16192 | ||
16193 | *Ulf Möller* | |
16194 | ||
16195 | * Retain source code compatibility for BN_prime_checks macro: | |
16196 | BN_is_prime(..., BN_prime_checks, ...) now uses | |
16197 | BN_prime_checks_for_size to determine the appropriate number of | |
16198 | Rabin-Miller iterations. | |
16199 | ||
16200 | *Ulf Möller* | |
16201 | ||
16202 | * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to | |
16203 | DH_CHECK_P_NOT_SAFE_PRIME. | |
16204 | (Check if this is true? OpenPGP calls them "strong".) | |
16205 | ||
16206 | *Ulf Möller* | |
16207 | ||
16208 | * Merge the functionality of "dh" and "gendh" programs into a new program | |
16209 | "dhparam". The old programs are retained for now but will handle DH keys | |
16210 | (instead of parameters) in future. | |
16211 | ||
16212 | *Steve Henson* | |
16213 | ||
16214 | * Make the ciphers, s_server and s_client programs check the return values | |
16215 | when a new cipher list is set. | |
16216 | ||
16217 | *Steve Henson* | |
16218 | ||
16219 | * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit | |
16220 | ciphers. Before when the 56bit ciphers were enabled the sorting was | |
16221 | wrong. | |
16222 | ||
16223 | The syntax for the cipher sorting has been extended to support sorting by | |
16224 | cipher-strength (using the strength_bits hard coded in the tables). | |
16225 | The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`). | |
16226 | ||
16227 | Fix a bug in the cipher-command parser: when supplying a cipher command | |
16228 | string with an "undefined" symbol (neither command nor alphanumeric | |
16229 | *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now | |
16230 | an error is flagged. | |
16231 | ||
16232 | Due to the strength-sorting extension, the code of the | |
16233 | ssl_create_cipher_list() function was completely rearranged. I hope that | |
16234 | the readability was also increased :-) | |
16235 | ||
16236 | *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>* | |
16237 | ||
16238 | * Minor change to 'x509' utility. The -CAcreateserial option now uses 1 | |
16239 | for the first serial number and places 2 in the serial number file. This | |
16240 | avoids problems when the root CA is created with serial number zero and | |
16241 | the first user certificate has the same issuer name and serial number | |
16242 | as the root CA. | |
16243 | ||
16244 | *Steve Henson* | |
16245 | ||
16246 | * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses | |
16247 | the new code. Add documentation for this stuff. | |
16248 | ||
16249 | *Steve Henson* | |
16250 | ||
16251 | * Changes to X509_ATTRIBUTE utilities. These have been renamed from | |
16252 | `X509_*()` to `X509at_*()` on the grounds that they don't handle X509 | |
16253 | structures and behave in an analogous way to the X509v3 functions: | |
16254 | they shouldn't be called directly but wrapper functions should be used | |
16255 | instead. | |
16256 | ||
16257 | So we also now have some wrapper functions that call the X509at functions | |
16258 | when passed certificate requests. (TO DO: similar things can be done with | |
16259 | PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other | |
16260 | things. Some of these need some d2i or i2d and print functionality | |
16261 | because they handle more complex structures.) | |
16262 | ||
16263 | *Steve Henson* | |
16264 | ||
16265 | * Add missing #ifndefs that caused missing symbols when building libssl | |
16266 | as a shared library without RSA. Use #ifndef NO_SSL2 instead of | |
16267 | NO_RSA in `ssl/s2*.c`. | |
16268 | ||
16269 | *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller* | |
16270 | ||
16271 | * Precautions against using the PRNG uninitialized: RAND_bytes() now | |
16272 | has a return value which indicates the quality of the random data | |
16273 | (1 = ok, 0 = not seeded). Also an error is recorded on the thread's | |
16274 | error queue. New function RAND_pseudo_bytes() generates output that is | |
16275 | guaranteed to be unique but not unpredictable. RAND_add is like | |
16276 | RAND_seed, but takes an extra argument for an entropy estimate | |
16277 | (RAND_seed always assumes full entropy). | |
16278 | ||
16279 | *Ulf Möller* | |
16280 | ||
16281 | * Do more iterations of Rabin-Miller probable prime test (specifically, | |
16282 | 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes | |
16283 | instead of only 2 for all lengths; see BN_prime_checks_for_size definition | |
16284 | in crypto/bn/bn_prime.c for the complete table). This guarantees a | |
16285 | false-positive rate of at most 2^-80 for random input. | |
16286 | ||
16287 | *Bodo Moeller* | |
16288 | ||
16289 | * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. | |
16290 | ||
16291 | *Bodo Moeller* | |
16292 | ||
16293 | * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain | |
16294 | in the 0.9.5 release), this returns the chain | |
16295 | from an X509_CTX structure with a dup of the stack and all | |
16296 | the X509 reference counts upped: so the stack will exist | |
16297 | after X509_CTX_cleanup() has been called. Modify pkcs12.c | |
16298 | to use this. | |
16299 | ||
16300 | Also make SSL_SESSION_print() print out the verify return | |
16301 | code. | |
16302 | ||
16303 | *Steve Henson* | |
16304 | ||
16305 | * Add manpage for the pkcs12 command. Also change the default | |
16306 | behaviour so MAC iteration counts are used unless the new | |
16307 | -nomaciter option is used. This improves file security and | |
16308 | only older versions of MSIE (4.0 for example) need it. | |
16309 | ||
16310 | *Steve Henson* | |
16311 | ||
16312 | * Honor the no-xxx Configure options when creating .DEF files. | |
16313 | ||
16314 | *Ulf Möller* | |
16315 | ||
16316 | * Add PKCS#10 attributes to field table: challengePassword, | |
16317 | unstructuredName and unstructuredAddress. These are taken from | |
16318 | draft PKCS#9 v2.0 but are compatible with v1.2 provided no | |
16319 | international characters are used. | |
16320 | ||
16321 | More changes to X509_ATTRIBUTE code: allow the setting of types | |
16322 | based on strings. Remove the 'loc' parameter when adding | |
16323 | attributes because these will be a SET OF encoding which is sorted | |
16324 | in ASN1 order. | |
16325 | ||
16326 | *Steve Henson* | |
16327 | ||
16328 | * Initial changes to the 'req' utility to allow request generation | |
16329 | automation. This will allow an application to just generate a template | |
16330 | file containing all the field values and have req construct the | |
16331 | request. | |
16332 | ||
16333 | Initial support for X509_ATTRIBUTE handling. Stacks of these are | |
16334 | used all over the place including certificate requests and PKCS#7 | |
16335 | structures. They are currently handled manually where necessary with | |
16336 | some primitive wrappers for PKCS#7. The new functions behave in a | |
16337 | manner analogous to the X509 extension functions: they allow | |
16338 | attributes to be looked up by NID and added. | |
16339 | ||
16340 | Later something similar to the X509V3 code would be desirable to | |
16341 | automatically handle the encoding, decoding and printing of the | |
16342 | more complex types. The string types like challengePassword can | |
16343 | be handled by the string table functions. | |
16344 | ||
16345 | Also modified the multi byte string table handling. Now there is | |
16346 | a 'global mask' which masks out certain types. The table itself | |
16347 | can use the flag STABLE_NO_MASK to ignore the mask setting: this | |
16348 | is useful when for example there is only one permissible type | |
16349 | (as in countryName) and using the mask might result in no valid | |
16350 | types at all. | |
16351 | ||
16352 | *Steve Henson* | |
16353 | ||
16354 | * Clean up 'Finished' handling, and add functions SSL_get_finished and | |
16355 | SSL_get_peer_finished to allow applications to obtain the latest | |
16356 | Finished messages sent to the peer or expected from the peer, | |
16357 | respectively. (SSL_get_peer_finished is usually the Finished message | |
16358 | actually received from the peer, otherwise the protocol will be aborted.) | |
16359 | ||
16360 | As the Finished message are message digests of the complete handshake | |
16361 | (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can | |
16362 | be used for external authentication procedures when the authentication | |
16363 | provided by SSL/TLS is not desired or is not enough. | |
16364 | ||
16365 | *Bodo Moeller* | |
16366 | ||
16367 | * Enhanced support for Alpha Linux is added. Now ./config checks if | |
16368 | the host supports BWX extension and if Compaq C is present on the | |
16369 | $PATH. Just exploiting of the BWX extension results in 20-30% | |
16370 | performance kick for some algorithms, e.g. DES and RC4 to mention | |
16371 | a couple. Compaq C in turn generates ~20% faster code for MD5 and | |
16372 | SHA1. | |
16373 | ||
16374 | *Andy Polyakov* | |
16375 | ||
16376 | * Add support for MS "fast SGC". This is arguably a violation of the | |
16377 | SSL3/TLS protocol. Netscape SGC does two handshakes: the first with | |
16378 | weak crypto and after checking the certificate is SGC a second one | |
16379 | with strong crypto. MS SGC stops the first handshake after receiving | |
16380 | the server certificate message and sends a second client hello. Since | |
16381 | a server will typically do all the time consuming operations before | |
16382 | expecting any further messages from the client (server key exchange | |
16383 | is the most expensive) there is little difference between the two. | |
16384 | ||
16385 | To get OpenSSL to support MS SGC we have to permit a second client | |
16386 | hello message after we have sent server done. In addition we have to | |
16387 | reset the MAC if we do get this second client hello. | |
16388 | ||
16389 | *Steve Henson* | |
16390 | ||
16391 | * Add a function 'd2i_AutoPrivateKey()' this will automatically decide | |
16392 | if a DER encoded private key is RSA or DSA traditional format. Changed | |
16393 | d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" | |
16394 | format DER encoded private key. Newer code should use PKCS#8 format which | |
16395 | has the key type encoded in the ASN1 structure. Added DER private key | |
16396 | support to pkcs8 application. | |
16397 | ||
16398 | *Steve Henson* | |
16399 | ||
16400 | * SSL 3/TLS 1 servers now don't request certificates when an anonymous | |
16401 | ciphersuites has been selected (as required by the SSL 3/TLS 1 | |
16402 | specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT | |
16403 | is set, we interpret this as a request to violate the specification | |
16404 | (the worst that can happen is a handshake failure, and 'correct' | |
16405 | behaviour would result in a handshake failure anyway). | |
16406 | ||
16407 | *Bodo Moeller* | |
16408 | ||
16409 | * In SSL_CTX_add_session, take into account that there might be multiple | |
16410 | SSL_SESSION structures with the same session ID (e.g. when two threads | |
16411 | concurrently obtain them from an external cache). | |
16412 | The internal cache can handle only one SSL_SESSION with a given ID, | |
16413 | so if there's a conflict, we now throw out the old one to achieve | |
16414 | consistency. | |
16415 | ||
16416 | *Bodo Moeller* | |
16417 | ||
16418 | * Add OIDs for idea and blowfish in CBC mode. This will allow both | |
16419 | to be used in PKCS#5 v2.0 and S/MIME. Also add checking to | |
16420 | some routines that use cipher OIDs: some ciphers do not have OIDs | |
16421 | defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for | |
16422 | example. | |
16423 | ||
16424 | *Steve Henson* | |
16425 | ||
16426 | * Simplify the trust setting structure and code. Now we just have | |
16427 | two sequences of OIDs for trusted and rejected settings. These will | |
16428 | typically have values the same as the extended key usage extension | |
16429 | and any application specific purposes. | |
16430 | ||
16431 | The trust checking code now has a default behaviour: it will just | |
16432 | check for an object with the same NID as the passed id. Functions can | |
16433 | be provided to override either the default behaviour or the behaviour | |
16434 | for a given id. SSL client, server and email already have functions | |
16435 | in place for compatibility: they check the NID and also return "trusted" | |
16436 | if the certificate is self signed. | |
16437 | ||
16438 | *Steve Henson* | |
16439 | ||
16440 | * Add d2i,i2d bio/fp functions for PrivateKey: these convert the | |
16441 | traditional format into an EVP_PKEY structure. | |
16442 | ||
16443 | *Steve Henson* | |
16444 | ||
16445 | * Add a password callback function PEM_cb() which either prompts for | |
16446 | a password if usr_data is NULL or otherwise assumes it is a null | |
16447 | terminated password. Allow passwords to be passed on command line | |
16448 | environment or config files in a few more utilities. | |
16449 | ||
16450 | *Steve Henson* | |
16451 | ||
16452 | * Add a bunch of DER and PEM functions to handle PKCS#8 format private | |
16453 | keys. Add some short names for PKCS#8 PBE algorithms and allow them | |
16454 | to be specified on the command line for the pkcs8 and pkcs12 utilities. | |
16455 | Update documentation. | |
16456 | ||
16457 | *Steve Henson* | |
16458 | ||
16459 | * Support for ASN1 "NULL" type. This could be handled before by using | |
16460 | ASN1_TYPE but there wasn't any function that would try to read a NULL | |
16461 | and produce an error if it couldn't. For compatibility we also have | |
16462 | ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and | |
16463 | don't allocate anything because they don't need to. | |
16464 | ||
16465 | *Steve Henson* | |
16466 | ||
16467 | * Initial support for MacOS is now provided. Examine INSTALL.MacOS | |
16468 | for details. | |
16469 | ||
16470 | *Andy Polyakov, Roy Woods <roy@centicsystems.ca>* | |
16471 | ||
16472 | * Rebuild of the memory allocation routines used by OpenSSL code and | |
16473 | possibly others as well. The purpose is to make an interface that | |
16474 | provide hooks so anyone can build a separate set of allocation and | |
16475 | deallocation routines to be used by OpenSSL, for example memory | |
16476 | pool implementations, or something else, which was previously hard | |
16477 | since Malloc(), Realloc() and Free() were defined as macros having | |
16478 | the values malloc, realloc and free, respectively (except for Win32 | |
16479 | compilations). The same is provided for memory debugging code. | |
16480 | OpenSSL already comes with functionality to find memory leaks, but | |
16481 | this gives people a chance to debug other memory problems. | |
16482 | ||
16483 | With these changes, a new set of functions and macros have appeared: | |
16484 | ||
16485 | CRYPTO_set_mem_debug_functions() [F] | |
16486 | CRYPTO_get_mem_debug_functions() [F] | |
16487 | CRYPTO_dbg_set_options() [F] | |
16488 | CRYPTO_dbg_get_options() [F] | |
16489 | CRYPTO_malloc_debug_init() [M] | |
16490 | ||
16491 | The memory debug functions are NULL by default, unless the library | |
16492 | is compiled with CRYPTO_MDEBUG or friends is defined. If someone | |
16493 | wants to debug memory anyway, CRYPTO_malloc_debug_init() (which | |
16494 | gives the standard debugging functions that come with OpenSSL) or | |
16495 | CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions | |
16496 | provided by the library user) must be used. When the standard | |
16497 | debugging functions are used, CRYPTO_dbg_set_options can be used to | |
16498 | request additional information: | |
16499 | CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting | |
16500 | the CRYPTO_MDEBUG_xxx macro when compiling the library. | |
16501 | ||
16502 | Also, things like CRYPTO_set_mem_functions will always give the | |
16503 | expected result (the new set of functions is used for allocation | |
16504 | and deallocation) at all times, regardless of platform and compiler | |
16505 | options. | |
16506 | ||
16507 | To finish it up, some functions that were never use in any other | |
16508 | way than through macros have a new API and new semantic: | |
16509 | ||
16510 | CRYPTO_dbg_malloc() | |
16511 | CRYPTO_dbg_realloc() | |
16512 | CRYPTO_dbg_free() | |
16513 | ||
16514 | All macros of value have retained their old syntax. | |
16515 | ||
16516 | *Richard Levitte and Bodo Moeller* | |
16517 | ||
16518 | * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the | |
16519 | ordering of SMIMECapabilities wasn't in "strength order" and there | |
16520 | was a missing NULL in the AlgorithmIdentifier for the SHA1 signature | |
16521 | algorithm. | |
16522 | ||
16523 | *Steve Henson* | |
16524 | ||
16525 | * Some ASN1 types with illegal zero length encoding (INTEGER, | |
16526 | ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. | |
16527 | ||
16528 | *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson* | |
16529 | ||
16530 | * Merge in my S/MIME library for OpenSSL. This provides a simple | |
16531 | S/MIME API on top of the PKCS#7 code, a MIME parser (with enough | |
16532 | functionality to handle multipart/signed properly) and a utility | |
16533 | called 'smime' to call all this stuff. This is based on code I | |
16534 | originally wrote for Celo who have kindly allowed it to be | |
16535 | included in OpenSSL. | |
16536 | ||
16537 | *Steve Henson* | |
16538 | ||
16539 | * Add variants des_set_key_checked and des_set_key_unchecked of | |
16540 | des_set_key (aka des_key_sched). Global variable des_check_key | |
16541 | decides which of these is called by des_set_key; this way | |
16542 | des_check_key behaves as it always did, but applications and | |
16543 | the library itself, which was buggy for des_check_key == 1, | |
16544 | have a cleaner way to pick the version they need. | |
16545 | ||
16546 | *Bodo Moeller* | |
16547 | ||
16548 | * New function PKCS12_newpass() which changes the password of a | |
16549 | PKCS12 structure. | |
16550 | ||
16551 | *Steve Henson* | |
16552 | ||
16553 | * Modify X509_TRUST and X509_PURPOSE so it also uses a static and | |
16554 | dynamic mix. In both cases the ids can be used as an index into the | |
16555 | table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() | |
16556 | functions so they accept a list of the field values and the | |
16557 | application doesn't need to directly manipulate the X509_TRUST | |
16558 | structure. | |
16559 | ||
16560 | *Steve Henson* | |
16561 | ||
16562 | * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't | |
16563 | need initialising. | |
16564 | ||
16565 | *Steve Henson* | |
16566 | ||
16567 | * Modify the way the V3 extension code looks up extensions. This now | |
16568 | works in a similar way to the object code: we have some "standard" | |
16569 | extensions in a static table which is searched with OBJ_bsearch() | |
16570 | and the application can add dynamic ones if needed. The file | |
16571 | crypto/x509v3/ext_dat.h now has the info: this file needs to be | |
16572 | updated whenever a new extension is added to the core code and kept | |
16573 | in ext_nid order. There is a simple program 'tabtest.c' which checks | |
16574 | this. New extensions are not added too often so this file can readily | |
16575 | be maintained manually. | |
16576 | ||
16577 | There are two big advantages in doing things this way. The extensions | |
16578 | can be looked up immediately and no longer need to be "added" using | |
16579 | X509V3_add_standard_extensions(): this function now does nothing. | |
16580 | Side note: I get *lots* of email saying the extension code doesn't | |
16581 | work because people forget to call this function. | |
16582 | Also no dynamic allocation is done unless new extensions are added: | |
16583 | so if we don't add custom extensions there is no need to call | |
16584 | X509V3_EXT_cleanup(). | |
16585 | ||
16586 | *Steve Henson* | |
16587 | ||
16588 | * Modify enc utility's salting as follows: make salting the default. Add a | |
16589 | magic header, so unsalted files fail gracefully instead of just decrypting | |
16590 | to garbage. This is because not salting is a big security hole, so people | |
16591 | should be discouraged from doing it. | |
16592 | ||
16593 | *Ben Laurie* | |
16594 | ||
16595 | * Fixes and enhancements to the 'x509' utility. It allowed a message | |
16596 | digest to be passed on the command line but it only used this | |
16597 | parameter when signing a certificate. Modified so all relevant | |
16598 | operations are affected by the digest parameter including the | |
16599 | -fingerprint and -x509toreq options. Also -x509toreq choked if a | |
16600 | DSA key was used because it didn't fix the digest. | |
16601 | ||
16602 | *Steve Henson* | |
16603 | ||
16604 | * Initial certificate chain verify code. Currently tests the untrusted | |
16605 | certificates for consistency with the verify purpose (which is set | |
16606 | when the X509_STORE_CTX structure is set up) and checks the pathlength. | |
16607 | ||
16608 | There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: | |
16609 | this is because it will reject chains with invalid extensions whereas | |
16610 | every previous version of OpenSSL and SSLeay made no checks at all. | |
16611 | ||
16612 | Trust code: checks the root CA for the relevant trust settings. Trust | |
16613 | settings have an initial value consistent with the verify purpose: e.g. | |
16614 | if the verify purpose is for SSL client use it expects the CA to be | |
16615 | trusted for SSL client use. However the default value can be changed to | |
16616 | permit custom trust settings: one example of this would be to only trust | |
16617 | certificates from a specific "secure" set of CAs. | |
16618 | ||
16619 | Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions | |
16620 | which should be used for version portability: especially since the | |
16621 | verify structure is likely to change more often now. | |
16622 | ||
16623 | SSL integration. Add purpose and trust to SSL_CTX and SSL and functions | |
16624 | to set them. If not set then assume SSL clients will verify SSL servers | |
16625 | and vice versa. | |
16626 | ||
16627 | Two new options to the verify program: -untrusted allows a set of | |
16628 | untrusted certificates to be passed in and -purpose which sets the | |
16629 | intended purpose of the certificate. If a purpose is set then the | |
16630 | new chain verify code is used to check extension consistency. | |
16631 | ||
16632 | *Steve Henson* | |
16633 | ||
16634 | * Support for the authority information access extension. | |
16635 | ||
16636 | *Steve Henson* | |
16637 | ||
16638 | * Modify RSA and DSA PEM read routines to transparently handle | |
16639 | PKCS#8 format private keys. New *_PUBKEY_* functions that handle | |
16640 | public keys in a format compatible with certificate | |
16641 | SubjectPublicKeyInfo structures. Unfortunately there were already | |
16642 | functions called *_PublicKey_* which used various odd formats so | |
16643 | these are retained for compatibility: however the DSA variants were | |
16644 | never in a public release so they have been deleted. Changed dsa/rsa | |
16645 | utilities to handle the new format: note no releases ever handled public | |
16646 | keys so we should be OK. | |
16647 | ||
16648 | The primary motivation for this change is to avoid the same fiasco | |
16649 | that dogs private keys: there are several incompatible private key | |
16650 | formats some of which are standard and some OpenSSL specific and | |
16651 | require various evil hacks to allow partial transparent handling and | |
16652 | even then it doesn't work with DER formats. Given the option anything | |
16653 | other than PKCS#8 should be dumped: but the other formats have to | |
16654 | stay in the name of compatibility. | |
16655 | ||
16656 | With public keys and the benefit of hindsight one standard format | |
16657 | is used which works with EVP_PKEY, RSA or DSA structures: though | |
16658 | it clearly returns an error if you try to read the wrong kind of key. | |
16659 | ||
16660 | Added a -pubkey option to the 'x509' utility to output the public key. | |
16661 | Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()` | |
16662 | (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add | |
16663 | `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`) | |
16664 | that do the same as the `EVP_PKEY_assign_*()` except they up the | |
16665 | reference count of the added key (they don't "swallow" the | |
16666 | supplied key). | |
16667 | ||
16668 | *Steve Henson* | |
16669 | ||
16670 | * Fixes to crypto/x509/by_file.c the code to read in certificates and | |
16671 | CRLs would fail if the file contained no certificates or no CRLs: | |
16672 | added a new function to read in both types and return the number | |
16673 | read: this means that if none are read it will be an error. The | |
16674 | DER versions of the certificate and CRL reader would always fail | |
16675 | because it isn't possible to mix certificates and CRLs in DER format | |
16676 | without choking one or the other routine. Changed this to just read | |
16677 | a certificate: this is the best we can do. Also modified the code | |
16678 | in `apps/verify.c` to take notice of return codes: it was previously | |
16679 | attempting to read in certificates from NULL pointers and ignoring | |
16680 | any errors: this is one reason why the cert and CRL reader seemed | |
16681 | to work. It doesn't check return codes from the default certificate | |
16682 | routines: these may well fail if the certificates aren't installed. | |
16683 | ||
16684 | *Steve Henson* | |
16685 | ||
16686 | * Code to support otherName option in GeneralName. | |
16687 | ||
16688 | *Steve Henson* | |
16689 | ||
16690 | * First update to verify code. Change the verify utility | |
16691 | so it warns if it is passed a self signed certificate: | |
16692 | for consistency with the normal behaviour. X509_verify | |
16693 | has been modified to it will now verify a self signed | |
16694 | certificate if *exactly* the same certificate appears | |
16695 | in the store: it was previously impossible to trust a | |
16696 | single self signed certificate. This means that: | |
16697 | openssl verify ss.pem | |
16698 | now gives a warning about a self signed certificate but | |
16699 | openssl verify -CAfile ss.pem ss.pem | |
16700 | is OK. | |
16701 | ||
16702 | *Steve Henson* | |
16703 | ||
16704 | * For servers, store verify_result in SSL_SESSION data structure | |
16705 | (and add it to external session representation). | |
16706 | This is needed when client certificate verifications fails, | |
16707 | but an application-provided verification callback (set by | |
16708 | SSL_CTX_set_cert_verify_callback) allows accepting the session | |
16709 | anyway (i.e. leaves x509_store_ctx->error != X509_V_OK | |
16710 | but returns 1): When the session is reused, we have to set | |
16711 | ssl->verify_result to the appropriate error code to avoid | |
16712 | security holes. | |
16713 | ||
16714 | *Bodo Moeller, problem pointed out by Lutz Jaenicke* | |
16715 | ||
16716 | * Fix a bug in the new PKCS#7 code: it didn't consider the | |
16717 | case in PKCS7_dataInit() where the signed PKCS7 structure | |
16718 | didn't contain any existing data because it was being created. | |
16719 | ||
16720 | *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson* | |
16721 | ||
16722 | * Add a salt to the key derivation routines in enc.c. This | |
16723 | forms the first 8 bytes of the encrypted file. Also add a | |
16724 | -S option to allow a salt to be input on the command line. | |
16725 | ||
16726 | *Steve Henson* | |
16727 | ||
16728 | * New function X509_cmp(). Oddly enough there wasn't a function | |
16729 | to compare two certificates. We do this by working out the SHA1 | |
16730 | hash and comparing that. X509_cmp() will be needed by the trust | |
16731 | code. | |
16732 | ||
16733 | *Steve Henson* | |
16734 | ||
16735 | * SSL_get1_session() is like SSL_get_session(), but increments | |
16736 | the reference count in the SSL_SESSION returned. | |
16737 | ||
16738 | *Geoff Thorpe <geoff@eu.c2.net>* | |
16739 | ||
16740 | * Fix for 'req': it was adding a null to request attributes. | |
16741 | Also change the X509_LOOKUP and X509_INFO code to handle | |
16742 | certificate auxiliary information. | |
16743 | ||
16744 | *Steve Henson* | |
16745 | ||
16746 | * Add support for 40 and 64 bit RC2 and RC4 algorithms: document | |
16747 | the 'enc' command. | |
16748 | ||
16749 | *Steve Henson* | |
16750 | ||
16751 | * Add the possibility to add extra information to the memory leak | |
16752 | detecting output, to form tracebacks, showing from where each | |
16753 | allocation was originated: CRYPTO_push_info("constant string") adds | |
16754 | the string plus current file name and line number to a per-thread | |
16755 | stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() | |
16756 | is like calling CYRPTO_pop_info() until the stack is empty. | |
16757 | Also updated memory leak detection code to be multi-thread-safe. | |
16758 | ||
16759 | *Richard Levitte* | |
16760 | ||
16761 | * Add options -text and -noout to pkcs7 utility and delete the | |
16762 | encryption options which never did anything. Update docs. | |
16763 | ||
16764 | *Steve Henson* | |
16765 | ||
16766 | * Add options to some of the utilities to allow the pass phrase | |
16767 | to be included on either the command line (not recommended on | |
16768 | OSes like Unix) or read from the environment. Update the | |
16769 | manpages and fix a few bugs. | |
16770 | ||
16771 | *Steve Henson* | |
16772 | ||
16773 | * Add a few manpages for some of the openssl commands. | |
16774 | ||
16775 | *Steve Henson* | |
16776 | ||
16777 | * Fix the -revoke option in ca. It was freeing up memory twice, | |
16778 | leaking and not finding already revoked certificates. | |
16779 | ||
16780 | *Steve Henson* | |
16781 | ||
16782 | * Extensive changes to support certificate auxiliary information. | |
16783 | This involves the use of X509_CERT_AUX structure and X509_AUX | |
16784 | functions. An X509_AUX function such as PEM_read_X509_AUX() | |
16785 | can still read in a certificate file in the usual way but it | |
16786 | will also read in any additional "auxiliary information". By | |
16787 | doing things this way a fair degree of compatibility can be | |
16788 | retained: existing certificates can have this information added | |
16789 | using the new 'x509' options. | |
16790 | ||
16791 | Current auxiliary information includes an "alias" and some trust | |
16792 | settings. The trust settings will ultimately be used in enhanced | |
16793 | certificate chain verification routines: currently a certificate | |
16794 | can only be trusted if it is self signed and then it is trusted | |
16795 | for all purposes. | |
16796 | ||
16797 | *Steve Henson* | |
16798 | ||
16799 | * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`). | |
16800 | The problem was that one of the replacement routines had not been working | |
16801 | since SSLeay releases. For now the offending routine has been replaced | |
16802 | with non-optimised assembler. Even so, this now gives around 95% | |
16803 | performance improvement for 1024 bit RSA signs. | |
16804 | ||
16805 | *Mark Cox* | |
16806 | ||
16807 | * Hack to fix PKCS#7 decryption when used with some unorthodox RC2 | |
16808 | handling. Most clients have the effective key size in bits equal to | |
16809 | the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. | |
16810 | A few however don't do this and instead use the size of the decrypted key | |
16811 | to determine the RC2 key length and the AlgorithmIdentifier to determine | |
16812 | the effective key length. In this case the effective key length can still | |
16813 | be 40 bits but the key length can be 168 bits for example. This is fixed | |
16814 | by manually forcing an RC2 key into the EVP_PKEY structure because the | |
16815 | EVP code can't currently handle unusual RC2 key sizes: it always assumes | |
16816 | the key length and effective key length are equal. | |
16817 | ||
16818 | *Steve Henson* | |
16819 | ||
16820 | * Add a bunch of functions that should simplify the creation of | |
16821 | X509_NAME structures. Now you should be able to do: | |
16822 | X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); | |
16823 | and have it automatically work out the correct field type and fill in | |
16824 | the structures. The more adventurous can try: | |
16825 | X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); | |
16826 | and it will (hopefully) work out the correct multibyte encoding. | |
16827 | ||
16828 | *Steve Henson* | |
16829 | ||
16830 | * Change the 'req' utility to use the new field handling and multibyte | |
16831 | copy routines. Before the DN field creation was handled in an ad hoc | |
16832 | way in req, ca, and x509 which was rather broken and didn't support | |
16833 | BMPStrings or UTF8Strings. Since some software doesn't implement | |
16834 | BMPStrings or UTF8Strings yet, they can be enabled using the config file | |
16835 | using the dirstring_type option. See the new comment in the default | |
16836 | openssl.cnf for more info. | |
16837 | ||
16838 | *Steve Henson* | |
16839 | ||
16840 | * Make crypto/rand/md_rand.c more robust: | |
16841 | - Assure unique random numbers after fork(). | |
16842 | - Make sure that concurrent threads access the global counter and | |
16843 | md serializably so that we never lose entropy in them | |
16844 | or use exactly the same state in multiple threads. | |
16845 | Access to the large state is not always serializable because | |
16846 | the additional locking could be a performance killer, and | |
16847 | md should be large enough anyway. | |
16848 | ||
16849 | *Bodo Moeller* | |
16850 | ||
16851 | * New file `apps/app_rand.c` with commonly needed functionality | |
16852 | for handling the random seed file. | |
16853 | ||
16854 | Use the random seed file in some applications that previously did not: | |
16855 | ca, | |
16856 | dsaparam -genkey (which also ignored its '-rand' option), | |
16857 | s_client, | |
16858 | s_server, | |
16859 | x509 (when signing). | |
16860 | Except on systems with /dev/urandom, it is crucial to have a random | |
16861 | seed file at least for key creation, DSA signing, and for DH exchanges; | |
16862 | for RSA signatures we could do without one. | |
16863 | ||
16864 | gendh and gendsa (unlike genrsa) used to read only the first byte | |
16865 | of each file listed in the '-rand' option. The function as previously | |
16866 | found in genrsa is now in app_rand.c and is used by all programs | |
16867 | that support '-rand'. | |
16868 | ||
16869 | *Bodo Moeller* | |
16870 | ||
16871 | * In RAND_write_file, use mode 0600 for creating files; | |
16872 | don't just chmod when it may be too late. | |
16873 | ||
16874 | *Bodo Moeller* | |
16875 | ||
16876 | * Report an error from X509_STORE_load_locations | |
16877 | when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. | |
16878 | ||
16879 | *Bill Perry* | |
16880 | ||
16881 | * New function ASN1_mbstring_copy() this copies a string in either | |
16882 | ASCII, Unicode, Universal (4 bytes per character) or UTF8 format | |
16883 | into an ASN1_STRING type. A mask of permissible types is passed | |
16884 | and it chooses the "minimal" type to use or an error if not type | |
16885 | is suitable. | |
16886 | ||
16887 | *Steve Henson* | |
16888 | ||
16889 | * Add function equivalents to the various macros in asn1.h. The old | |
16890 | macros are retained with an `M_` prefix. Code inside the library can | |
16891 | use the `M_` macros. External code (including the openssl utility) | |
16892 | should *NOT* in order to be "shared library friendly". | |
16893 | ||
16894 | *Steve Henson* | |
16895 | ||
16896 | * Add various functions that can check a certificate's extensions | |
16897 | to see if it usable for various purposes such as SSL client, | |
16898 | server or S/MIME and CAs of these types. This is currently | |
16899 | VERY EXPERIMENTAL but will ultimately be used for certificate chain | |
16900 | verification. Also added a -purpose flag to x509 utility to | |
16901 | print out all the purposes. | |
16902 | ||
16903 | *Steve Henson* | |
16904 | ||
16905 | * Add a CRYPTO_EX_DATA to X509 certificate structure and associated | |
16906 | functions. | |
16907 | ||
16908 | *Steve Henson* | |
16909 | ||
16910 | * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search | |
16911 | for, obtain and decode and extension and obtain its critical flag. | |
16912 | This allows all the necessary extension code to be handled in a | |
16913 | single function call. | |
16914 | ||
16915 | *Steve Henson* | |
16916 | ||
16917 | * RC4 tune-up featuring 30-40% performance improvement on most RISC | |
16918 | platforms. See crypto/rc4/rc4_enc.c for further details. | |
16919 | ||
16920 | *Andy Polyakov* | |
16921 | ||
16922 | * New -noout option to asn1parse. This causes no output to be produced | |
16923 | its main use is when combined with -strparse and -out to extract data | |
16924 | from a file (which may not be in ASN.1 format). | |
16925 | ||
16926 | *Steve Henson* | |
16927 | ||
16928 | * Fix for pkcs12 program. It was hashing an invalid certificate pointer | |
16929 | when producing the local key id. | |
16930 | ||
16931 | *Richard Levitte <levitte@stacken.kth.se>* | |
16932 | ||
16933 | * New option -dhparam in s_server. This allows a DH parameter file to be | |
16934 | stated explicitly. If it is not stated then it tries the first server | |
16935 | certificate file. The previous behaviour hard coded the filename | |
16936 | "server.pem". | |
16937 | ||
16938 | *Steve Henson* | |
16939 | ||
16940 | * Add -pubin and -pubout options to the rsa and dsa commands. These allow | |
16941 | a public key to be input or output. For example: | |
16942 | openssl rsa -in key.pem -pubout -out pubkey.pem | |
16943 | Also added necessary DSA public key functions to handle this. | |
16944 | ||
16945 | *Steve Henson* | |
16946 | ||
16947 | * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained | |
16948 | in the message. This was handled by allowing | |
16949 | X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. | |
16950 | ||
16951 | *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>* | |
16952 | ||
16953 | * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null | |
16954 | to the end of the strings whereas this didn't. This would cause problems | |
16955 | if strings read with d2i_ASN1_bytes() were later modified. | |
16956 | ||
16957 | *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>* | |
16958 | ||
16959 | * Fix for base64 decode bug. When a base64 bio reads only one line of | |
16960 | data and it contains EOF it will end up returning an error. This is | |
16961 | caused by input 46 bytes long. The cause is due to the way base64 | |
16962 | BIOs find the start of base64 encoded data. They do this by trying a | |
16963 | trial decode on each line until they find one that works. When they | |
16964 | do a flag is set and it starts again knowing it can pass all the | |
16965 | data directly through the decoder. Unfortunately it doesn't reset | |
16966 | the context it uses. This means that if EOF is reached an attempt | |
16967 | is made to pass two EOFs through the context and this causes the | |
16968 | resulting error. This can also cause other problems as well. As is | |
16969 | usual with these problems it takes *ages* to find and the fix is | |
16970 | trivial: move one line. | |
16971 | ||
16972 | *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)* | |
16973 | ||
16974 | * Ugly workaround to get s_client and s_server working under Windows. The | |
16975 | old code wouldn't work because it needed to select() on sockets and the | |
16976 | tty (for keypresses and to see if data could be written). Win32 only | |
16977 | supports select() on sockets so we select() with a 1s timeout on the | |
16978 | sockets and then see if any characters are waiting to be read, if none | |
16979 | are present then we retry, we also assume we can always write data to | |
16980 | the tty. This isn't nice because the code then blocks until we've | |
16981 | received a complete line of data and it is effectively polling the | |
16982 | keyboard at 1s intervals: however it's quite a bit better than not | |
16983 | working at all :-) A dedicated Windows application might handle this | |
16984 | with an event loop for example. | |
16985 | ||
16986 | *Steve Henson* | |
16987 | ||
16988 | * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign | |
16989 | and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions | |
16990 | will be called when RSA_sign() and RSA_verify() are used. This is useful | |
16991 | if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. | |
16992 | For this to work properly RSA_public_decrypt() and RSA_private_encrypt() | |
16993 | should *not* be used: RSA_sign() and RSA_verify() must be used instead. | |
16994 | This necessitated the support of an extra signature type NID_md5_sha1 | |
16995 | for SSL signatures and modifications to the SSL library to use it instead | |
16996 | of calling RSA_public_decrypt() and RSA_private_encrypt(). | |
16997 | ||
16998 | *Steve Henson* | |
16999 | ||
17000 | * Add new -verify -CAfile and -CApath options to the crl program, these | |
17001 | will lookup a CRL issuers certificate and verify the signature in a | |
17002 | similar way to the verify program. Tidy up the crl program so it | |
17003 | no longer accesses structures directly. Make the ASN1 CRL parsing a bit | |
17004 | less strict. It will now permit CRL extensions even if it is not | |
17005 | a V2 CRL: this will allow it to tolerate some broken CRLs. | |
17006 | ||
17007 | *Steve Henson* | |
17008 | ||
17009 | * Initialize all non-automatic variables each time one of the openssl | |
17010 | sub-programs is started (this is necessary as they may be started | |
17011 | multiple times from the "OpenSSL>" prompt). | |
17012 | ||
17013 | *Lennart Bang, Bodo Moeller* | |
17014 | ||
17015 | * Preliminary compilation option RSA_NULL which disables RSA crypto without | |
17016 | removing all other RSA functionality (this is what NO_RSA does). This | |
17017 | is so (for example) those in the US can disable those operations covered | |
17018 | by the RSA patent while allowing storage and parsing of RSA keys and RSA | |
17019 | key generation. | |
17020 | ||
17021 | *Steve Henson* | |
17022 | ||
17023 | * Non-copying interface to BIO pairs. | |
17024 | (still largely untested) | |
17025 | ||
17026 | *Bodo Moeller* | |
17027 | ||
17028 | * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive | |
17029 | ASCII string. This was handled independently in various places before. | |
17030 | ||
17031 | *Steve Henson* | |
17032 | ||
17033 | * New functions UTF8_getc() and UTF8_putc() that parse and generate | |
17034 | UTF8 strings a character at a time. | |
17035 | ||
17036 | *Steve Henson* | |
17037 | ||
17038 | * Use client_version from client hello to select the protocol | |
17039 | (s23_srvr.c) and for RSA client key exchange verification | |
17040 | (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. | |
17041 | ||
17042 | *Bodo Moeller* | |
17043 | ||
17044 | * Add various utility functions to handle SPKACs, these were previously | |
17045 | handled by poking round in the structure internals. Added new function | |
17046 | NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to | |
17047 | print, verify and generate SPKACs. Based on an original idea from | |
17048 | Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. | |
17049 | ||
17050 | *Steve Henson* | |
17051 | ||
17052 | * RIPEMD160 is operational on all platforms and is back in 'make test'. | |
17053 | ||
17054 | *Andy Polyakov* | |
17055 | ||
17056 | * Allow the config file extension section to be overwritten on the | |
17057 | command line. Based on an original idea from Massimiliano Pala | |
17058 | <madwolf@comune.modena.it>. The new option is called -extensions | |
17059 | and can be applied to ca, req and x509. Also -reqexts to override | |
17060 | the request extensions in req and -crlexts to override the crl extensions | |
17061 | in ca. | |
17062 | ||
17063 | *Steve Henson* | |
17064 | ||
17065 | * Add new feature to the SPKAC handling in ca. Now you can include | |
17066 | the same field multiple times by preceding it by "XXXX." for example: | |
17067 | 1.OU="Unit name 1" | |
17068 | 2.OU="Unit name 2" | |
17069 | this is the same syntax as used in the req config file. | |
17070 | ||
17071 | *Steve Henson* | |
17072 | ||
17073 | * Allow certificate extensions to be added to certificate requests. These | |
17074 | are specified in a 'req_extensions' option of the req section of the | |
17075 | config file. They can be printed out with the -text option to req but | |
17076 | are otherwise ignored at present. | |
17077 | ||
17078 | *Steve Henson* | |
17079 | ||
17080 | * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first | |
17081 | data read consists of only the final block it would not decrypted because | |
17082 | EVP_CipherUpdate() would correctly report zero bytes had been decrypted. | |
17083 | A misplaced 'break' also meant the decrypted final block might not be | |
17084 | copied until the next read. | |
17085 | ||
17086 | *Steve Henson* | |
17087 | ||
17088 | * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added | |
17089 | a few extra parameters to the DH structure: these will be useful if | |
17090 | for example we want the value of 'q' or implement X9.42 DH. | |
17091 | ||
17092 | *Steve Henson* | |
17093 | ||
17094 | * Initial support for DSA_METHOD. This is based on the RSA_METHOD and | |
17095 | provides hooks that allow the default DSA functions or functions on a | |
17096 | "per key" basis to be replaced. This allows hardware acceleration and | |
17097 | hardware key storage to be handled without major modification to the | |
17098 | library. Also added low level modexp hooks and CRYPTO_EX structure and | |
17099 | associated functions. | |
17100 | ||
17101 | *Steve Henson* | |
17102 | ||
17103 | * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO | |
17104 | as "read only": it can't be written to and the buffer it points to will | |
17105 | not be freed. Reading from a read only BIO is much more efficient than | |
17106 | a normal memory BIO. This was added because there are several times when | |
17107 | an area of memory needs to be read from a BIO. The previous method was | |
17108 | to create a memory BIO and write the data to it, this results in two | |
17109 | copies of the data and an O(n^2) reading algorithm. There is a new | |
17110 | function BIO_new_mem_buf() which creates a read only memory BIO from | |
17111 | an area of memory. Also modified the PKCS#7 routines to use read only | |
17112 | memory BIOs. | |
17113 | ||
17114 | *Steve Henson* | |
17115 | ||
17116 | * Bugfix: ssl23_get_client_hello did not work properly when called in | |
17117 | state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of | |
17118 | a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, | |
17119 | but a retry condition occurred while trying to read the rest. | |
17120 | ||
17121 | *Bodo Moeller* | |
17122 | ||
17123 | * The PKCS7_ENC_CONTENT_new() function was setting the content type as | |
17124 | NID_pkcs7_encrypted by default: this was wrong since this should almost | |
17125 | always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle | |
17126 | the encrypted data type: this is a more sensible place to put it and it | |
17127 | allows the PKCS#12 code to be tidied up that duplicated this | |
17128 | functionality. | |
17129 | ||
17130 | *Steve Henson* | |
17131 | ||
17132 | * Changed obj_dat.pl script so it takes its input and output files on | |
17133 | the command line. This should avoid shell escape redirection problems | |
17134 | under Win32. | |
17135 | ||
17136 | *Steve Henson* | |
17137 | ||
17138 | * Initial support for certificate extension requests, these are included | |
17139 | in things like Xenroll certificate requests. Included functions to allow | |
17140 | extensions to be obtained and added. | |
17141 | ||
17142 | *Steve Henson* | |
17143 | ||
17144 | * -crlf option to s_client and s_server for sending newlines as | |
17145 | CRLF (as required by many protocols). | |
17146 | ||
17147 | *Bodo Moeller* | |
17148 | ||
17149 | ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999] | |
17150 | ||
17151 | * Install libRSAglue.a when OpenSSL is built with RSAref. | |
17152 | ||
17153 | *Ralf S. Engelschall* | |
17154 | ||
17155 | * A few more `#ifndef NO_FP_API / #endif` pairs for consistency. | |
17156 | ||
17157 | *Andrija Antonijevic <TheAntony2@bigfoot.com>* | |
17158 | ||
17159 | * Fix -startdate and -enddate (which was missing) arguments to 'ca' | |
17160 | program. | |
17161 | ||
17162 | *Steve Henson* | |
17163 | ||
17164 | * New function DSA_dup_DH, which duplicates DSA parameters/keys as | |
17165 | DH parameters/keys (q is lost during that conversion, but the resulting | |
17166 | DH parameters contain its length). | |
17167 | ||
17168 | For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is | |
17169 | much faster than DH_generate_parameters (which creates parameters | |
17170 | where `p = 2*q + 1`), and also the smaller q makes DH computations | |
17171 | much more efficient (160-bit exponentiation instead of 1024-bit | |
17172 | exponentiation); so this provides a convenient way to support DHE | |
17173 | ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of | |
17174 | utter importance to use | |
17175 | SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); | |
17176 | or | |
17177 | SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); | |
17178 | when such DH parameters are used, because otherwise small subgroup | |
17179 | attacks may become possible! | |
17180 | ||
17181 | *Bodo Moeller* | |
17182 | ||
17183 | * Avoid memory leak in i2d_DHparams. | |
17184 | ||
17185 | *Bodo Moeller* | |
17186 | ||
17187 | * Allow the -k option to be used more than once in the enc program: | |
17188 | this allows the same encrypted message to be read by multiple recipients. | |
17189 | ||
17190 | *Steve Henson* | |
17191 | ||
17192 | * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts | |
17193 | an ASN1_OBJECT to a text string. If the "no_name" parameter is set then | |
17194 | it will always use the numerical form of the OID, even if it has a short | |
17195 | or long name. | |
17196 | ||
17197 | *Steve Henson* | |
17198 | ||
17199 | * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp | |
17200 | method only got called if p,q,dmp1,dmq1,iqmp components were present, | |
17201 | otherwise bn_mod_exp was called. In the case of hardware keys for example | |
17202 | no private key components need be present and it might store extra data | |
17203 | in the RSA structure, which cannot be accessed from bn_mod_exp. | |
17204 | By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for | |
17205 | private key operations. | |
17206 | ||
17207 | *Steve Henson* | |
17208 | ||
17209 | * Added support for SPARC Linux. | |
17210 | ||
17211 | *Andy Polyakov* | |
17212 | ||
17213 | * pem_password_cb function type incompatibly changed from | |
17214 | typedef int pem_password_cb(char *buf, int size, int rwflag); | |
17215 | to | |
17216 | ....(char *buf, int size, int rwflag, void *userdata); | |
17217 | so that applications can pass data to their callbacks: | |
17218 | The `PEM[_ASN1]_{read,write}...` functions and macros now take an | |
17219 | additional void * argument, which is just handed through whenever | |
17220 | the password callback is called. | |
17221 | ||
17222 | *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller* | |
17223 | ||
17224 | New function SSL_CTX_set_default_passwd_cb_userdata. | |
17225 | ||
17226 | Compatibility note: As many C implementations push function arguments | |
17227 | onto the stack in reverse order, the new library version is likely to | |
17228 | interoperate with programs that have been compiled with the old | |
17229 | pem_password_cb definition (PEM_whatever takes some data that | |
17230 | happens to be on the stack as its last argument, and the callback | |
17231 | just ignores this garbage); but there is no guarantee whatsoever that | |
17232 | this will work. | |
17233 | ||
17234 | * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... | |
17235 | (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused | |
17236 | problems not only on Windows, but also on some Unix platforms. | |
17237 | To avoid problematic command lines, these definitions are now in an | |
17238 | auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl | |
17239 | for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). | |
17240 | ||
17241 | *Bodo Moeller* | |
17242 | ||
17243 | * MIPS III/IV assembler module is reimplemented. | |
17244 | ||
17245 | *Andy Polyakov* | |
17246 | ||
17247 | * More DES library cleanups: remove references to srand/rand and | |
17248 | delete an unused file. | |
17249 | ||
17250 | *Ulf Möller* | |
17251 | ||
17252 | * Add support for the free Netwide assembler (NASM) under Win32, | |
17253 | since not many people have MASM (ml) and it can be hard to obtain. | |
17254 | This is currently experimental but it seems to work OK and pass all | |
17255 | the tests. Check out INSTALL.W32 for info. | |
17256 | ||
17257 | *Steve Henson* | |
17258 | ||
17259 | * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections | |
17260 | without temporary keys kept an extra copy of the server key, | |
17261 | and connections with temporary keys did not free everything in case | |
17262 | of an error. | |
17263 | ||
17264 | *Bodo Moeller* | |
17265 | ||
17266 | * New function RSA_check_key and new openssl rsa option -check | |
17267 | for verifying the consistency of RSA keys. | |
17268 | ||
17269 | *Ulf Moeller, Bodo Moeller* | |
17270 | ||
17271 | * Various changes to make Win32 compile work: | |
17272 | 1. Casts to avoid "loss of data" warnings in p5_crpt2.c | |
17273 | 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned | |
17274 | comparison" warnings. | |
17275 | 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update. | |
17276 | ||
17277 | *Steve Henson* | |
17278 | ||
17279 | * Add a debugging option to PKCS#5 v2 key generation function: when | |
17280 | you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and | |
17281 | derived keys are printed to stderr. | |
17282 | ||
17283 | *Steve Henson* | |
17284 | ||
17285 | * Copy the flags in ASN1_STRING_dup(). | |
17286 | ||
17287 | *Roman E. Pavlov <pre@mo.msk.ru>* | |
17288 | ||
17289 | * The x509 application mishandled signing requests containing DSA | |
17290 | keys when the signing key was also DSA and the parameters didn't match. | |
17291 | ||
17292 | It was supposed to omit the parameters when they matched the signing key: | |
17293 | the verifying software was then supposed to automatically use the CA's | |
17294 | parameters if they were absent from the end user certificate. | |
17295 | ||
17296 | Omitting parameters is no longer recommended. The test was also | |
17297 | the wrong way round! This was probably due to unusual behaviour in | |
17298 | EVP_cmp_parameters() which returns 1 if the parameters match. | |
17299 | This meant that parameters were omitted when they *didn't* match and | |
17300 | the certificate was useless. Certificates signed with 'ca' didn't have | |
17301 | this bug. | |
17302 | ||
17303 | *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>* | |
17304 | ||
17305 | * Memory leak checking (-DCRYPTO_MDEBUG) had some problems. | |
17306 | The interface is as follows: | |
17307 | Applications can use | |
17308 | CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), | |
17309 | CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); | |
17310 | "off" is now the default. | |
17311 | The library internally uses | |
17312 | CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), | |
17313 | CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() | |
17314 | to disable memory-checking temporarily. | |
17315 | ||
17316 | Some inconsistent states that previously were possible (and were | |
17317 | even the default) are now avoided. | |
17318 | ||
17319 | -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time | |
17320 | with each memory chunk allocated; this is occasionally more helpful | |
17321 | than just having a counter. | |
17322 | ||
17323 | -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. | |
17324 | ||
17325 | -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future | |
17326 | extensions. | |
17327 | ||
17328 | *Bodo Moeller* | |
17329 | ||
17330 | * Introduce "mode" for SSL structures (with defaults in SSL_CTX), | |
17331 | which largely parallels "options", but is for changing API behaviour, | |
17332 | whereas "options" are about protocol behaviour. | |
17333 | Initial "mode" flags are: | |
17334 | ||
17335 | SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when | |
17336 | a single record has been written. | |
17337 | SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write | |
17338 | retries use the same buffer location. | |
17339 | (But all of the contents must be | |
17340 | copied!) | |
17341 | ||
17342 | *Bodo Moeller* | |
17343 | ||
17344 | * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options | |
17345 | worked. | |
17346 | ||
17347 | * Fix problems with no-hmac etc. | |
17348 | ||
17349 | *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>* | |
17350 | ||
17351 | * New functions RSA_get_default_method(), RSA_set_method() and | |
17352 | RSA_get_method(). These allows replacement of RSA_METHODs without having | |
17353 | to mess around with the internals of an RSA structure. | |
17354 | ||
17355 | *Steve Henson* | |
17356 | ||
17357 | * Fix memory leaks in DSA_do_sign and DSA_is_prime. | |
17358 | Also really enable memory leak checks in openssl.c and in some | |
17359 | test programs. | |
17360 | ||
17361 | *Chad C. Mulligan, Bodo Moeller* | |
17362 | ||
17363 | * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess | |
17364 | up the length of negative integers. This has now been simplified to just | |
17365 | store the length when it is first determined and use it later, rather | |
17366 | than trying to keep track of where data is copied and updating it to | |
17367 | point to the end. | |
17368 | *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>* | |
17369 | ||
17370 | * Add a new function PKCS7_signatureVerify. This allows the verification | |
17371 | of a PKCS#7 signature but with the signing certificate passed to the | |
17372 | function itself. This contrasts with PKCS7_dataVerify which assumes the | |
17373 | certificate is present in the PKCS#7 structure. This isn't always the | |
17374 | case: certificates can be omitted from a PKCS#7 structure and be | |
17375 | distributed by "out of band" means (such as a certificate database). | |
17376 | ||
17377 | *Steve Henson* | |
17378 | ||
17379 | * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the | |
17380 | function prototypes in pem.h, also change util/mkdef.pl to add the | |
17381 | necessary function names. | |
17382 | ||
17383 | *Steve Henson* | |
17384 | ||
17385 | * mk1mf.pl (used by Windows builds) did not properly read the | |
17386 | options set by Configure in the top level Makefile, and Configure | |
17387 | was not even able to write more than one option correctly. | |
17388 | Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. | |
17389 | ||
17390 | *Bodo Moeller* | |
17391 | ||
17392 | * New functions CONF_load_bio() and CONF_load_fp() to allow a config | |
17393 | file to be loaded from a BIO or FILE pointer. The BIO version will | |
17394 | for example allow memory BIOs to contain config info. | |
17395 | ||
17396 | *Steve Henson* | |
17397 | ||
17398 | * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. | |
17399 | Whoever hopes to achieve shared-library compatibility across versions | |
17400 | must use this, not the compile-time macro. | |
17401 | (Exercise 0.9.4: Which is the minimum library version required by | |
17402 | such programs?) | |
17403 | Note: All this applies only to multi-threaded programs, others don't | |
17404 | need locks. | |
17405 | ||
17406 | *Bodo Moeller* | |
17407 | ||
17408 | * Add missing case to s3_clnt.c state machine -- one of the new SSL tests | |
17409 | through a BIO pair triggered the default case, i.e. | |
17410 | SSLerr(...,SSL_R_UNKNOWN_STATE). | |
17411 | ||
17412 | *Bodo Moeller* | |
17413 | ||
17414 | * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications | |
17415 | can use the SSL library even if none of the specific BIOs is | |
17416 | appropriate. | |
17417 | ||
17418 | *Bodo Moeller* | |
17419 | ||
17420 | * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value | |
17421 | for the encoded length. | |
17422 | ||
17423 | *Jeon KyoungHo <khjeon@sds.samsung.co.kr>* | |
17424 | ||
17425 | * Add initial documentation of the X509V3 functions. | |
17426 | ||
17427 | *Steve Henson* | |
17428 | ||
17429 | * Add a new pair of functions PEM_write_PKCS8PrivateKey() and | |
17430 | PEM_write_bio_PKCS8PrivateKey() that are equivalent to | |
17431 | PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more | |
17432 | secure PKCS#8 private key format with a high iteration count. | |
17433 | ||
17434 | *Steve Henson* | |
17435 | ||
17436 | * Fix determination of Perl interpreter: A perl or perl5 | |
17437 | *directory* in $PATH was also accepted as the interpreter. | |
17438 | ||
17439 | *Ralf S. Engelschall* | |
17440 | ||
17441 | * Fix demos/sign/sign.c: well there wasn't anything strictly speaking | |
17442 | wrong with it but it was very old and did things like calling | |
17443 | PEM_ASN1_read() directly and used MD5 for the hash not to mention some | |
17444 | unusual formatting. | |
17445 | ||
17446 | *Steve Henson* | |
17447 | ||
17448 | * Fix demos/selfsign.c: it used obsolete and deleted functions, changed | |
17449 | to use the new extension code. | |
17450 | ||
17451 | *Steve Henson* | |
17452 | ||
17453 | * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c | |
17454 | with macros. This should make it easier to change their form, add extra | |
17455 | arguments etc. Fix a few PEM prototypes which didn't have cipher as a | |
17456 | constant. | |
17457 | ||
17458 | *Steve Henson* | |
17459 | ||
17460 | * Add to configuration table a new entry that can specify an alternative | |
17461 | name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, | |
17462 | according to Mark Crispin <MRC@Panda.COM>. | |
17463 | ||
17464 | *Bodo Moeller* | |
17465 | ||
17466 | * DES CBC did not update the IV. Weird. | |
17467 | ||
17468 | *Ben Laurie* | |
17469 | lse | |
17470 | des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. | |
17471 | Changing the behaviour of the former might break existing programs -- | |
17472 | where IV updating is needed, des_ncbc_encrypt can be used. | |
17473 | ndif | |
17474 | ||
17475 | * When bntest is run from "make test" it drives bc to check its | |
17476 | calculations, as well as internally checking them. If an internal check | |
17477 | fails, it needs to cause bc to give a non-zero result or make test carries | |
17478 | on without noticing the failure. Fixed. | |
17479 | ||
17480 | *Ben Laurie* | |
17481 | ||
17482 | * DES library cleanups. | |
17483 | ||
17484 | *Ulf Möller* | |
17485 | ||
17486 | * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be | |
17487 | used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit | |
17488 | ciphers. NOTE: although the key derivation function has been verified | |
17489 | against some published test vectors it has not been extensively tested | |
17490 | yet. Added a -v2 "cipher" option to pkcs8 application to allow the use | |
17491 | of v2.0. | |
17492 | ||
17493 | *Steve Henson* | |
17494 | ||
17495 | * Instead of "mkdir -p", which is not fully portable, use new | |
17496 | Perl script "util/mkdir-p.pl". | |
17497 | ||
17498 | *Bodo Moeller* | |
17499 | ||
17500 | * Rewrite the way password based encryption (PBE) is handled. It used to | |
17501 | assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter | |
17502 | structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms | |
17503 | but doesn't apply to PKCS#5 v2.0 where it can be something else. Now | |
17504 | the 'parameter' field of the AlgorithmIdentifier is passed to the | |
17505 | underlying key generation function so it must do its own ASN1 parsing. | |
17506 | This has also changed the EVP_PBE_CipherInit() function which now has a | |
17507 | 'parameter' argument instead of literal salt and iteration count values | |
17508 | and the function EVP_PBE_ALGOR_CipherInit() has been deleted. | |
17509 | ||
17510 | *Steve Henson* | |
17511 | ||
17512 | * Support for PKCS#5 v1.5 compatible password based encryption algorithms | |
17513 | and PKCS#8 functionality. New 'pkcs8' application linked to openssl. | |
17514 | Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE | |
17515 | KEY" because this clashed with PKCS#8 unencrypted string. Since this | |
17516 | value was just used as a "magic string" and not used directly its | |
17517 | value doesn't matter. | |
17518 | ||
17519 | *Steve Henson* | |
17520 | ||
17521 | * Introduce some semblance of const correctness to BN. Shame C doesn't | |
17522 | support mutable. | |
17523 | ||
17524 | *Ben Laurie* | |
17525 | ||
17526 | * "linux-sparc64" configuration (ultrapenguin). | |
17527 | ||
17528 | *Ray Miller <ray.miller@oucs.ox.ac.uk>* | |
17529 | "linux-sparc" configuration. | |
17530 | ||
17531 | *Christian Forster <fo@hawo.stw.uni-erlangen.de>* | |
17532 | ||
17533 | * config now generates no-xxx options for missing ciphers. | |
17534 | ||
17535 | *Ulf Möller* | |
17536 | ||
17537 | * Support the EBCDIC character set (work in progress). | |
17538 | File ebcdic.c not yet included because it has a different license. | |
17539 | ||
17540 | *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>* | |
17541 | ||
17542 | * Support BS2000/OSD-POSIX. | |
17543 | ||
17544 | *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>* | |
17545 | ||
17546 | * Make callbacks for key generation use `void *` instead of `char *`. | |
17547 | ||
17548 | *Ben Laurie* | |
17549 | ||
17550 | * Make S/MIME samples compile (not yet tested). | |
17551 | ||
17552 | *Ben Laurie* | |
17553 | ||
17554 | * Additional typesafe stacks. | |
17555 | ||
17556 | *Ben Laurie* | |
17557 | ||
17558 | * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). | |
17559 | ||
17560 | *Bodo Moeller* | |
17561 | ||
17562 | ### Changes between 0.9.3 and 0.9.3a [29 May 1999] | |
17563 | ||
17564 | * New configuration variant "sco5-gcc". | |
17565 | ||
17566 | * Updated some demos. | |
17567 | ||
17568 | *Sean O Riordain, Wade Scholine* | |
17569 | ||
17570 | * Add missing BIO_free at exit of pkcs12 application. | |
17571 | ||
17572 | *Wu Zhigang* | |
17573 | ||
17574 | * Fix memory leak in conf.c. | |
17575 | ||
17576 | *Steve Henson* | |
17577 | ||
17578 | * Updates for Win32 to assembler version of MD5. | |
17579 | ||
17580 | *Steve Henson* | |
17581 | ||
17582 | * Set #! path to perl in `apps/der_chop` to where we found it | |
17583 | instead of using a fixed path. | |
17584 | ||
17585 | *Bodo Moeller* | |
17586 | ||
17587 | * SHA library changes for irix64-mips4-cc. | |
17588 | ||
17589 | *Andy Polyakov* | |
17590 | ||
17591 | * Improvements for VMS support. | |
17592 | ||
17593 | *Richard Levitte* | |
17594 | ||
17595 | ### Changes between 0.9.2b and 0.9.3 [24 May 1999] | |
17596 | ||
17597 | * Bignum library bug fix. IRIX 6 passes "make test" now! | |
17598 | This also avoids the problems with SC4.2 and unpatched SC5. | |
17599 | ||
17600 | *Andy Polyakov <appro@fy.chalmers.se>* | |
17601 | ||
17602 | * New functions sk_num, sk_value and sk_set to replace the previous macros. | |
17603 | These are required because of the typesafe stack would otherwise break | |
17604 | existing code. If old code used a structure member which used to be STACK | |
17605 | and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with | |
17606 | sk_num or sk_value it would produce an error because the num, data members | |
17607 | are not present in STACK_OF. Now it just produces a warning. sk_set | |
17608 | replaces the old method of assigning a value to sk_value | |
17609 | (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code | |
17610 | that does this will no longer work (and should use sk_set instead) but | |
17611 | this could be regarded as a "questionable" behaviour anyway. | |
17612 | ||
17613 | *Steve Henson* | |
17614 | ||
17615 | * Fix most of the other PKCS#7 bugs. The "experimental" code can now | |
17616 | correctly handle encrypted S/MIME data. | |
17617 | ||
17618 | *Steve Henson* | |
17619 | ||
17620 | * Change type of various DES function arguments from des_cblock | |
17621 | (which means, in function argument declarations, pointer to char) | |
17622 | to des_cblock * (meaning pointer to array with 8 char elements), | |
17623 | which allows the compiler to do more typechecking; it was like | |
17624 | that back in SSLeay, but with lots of ugly casts. | |
17625 | ||
17626 | Introduce new type const_des_cblock. | |
17627 | ||
17628 | *Bodo Moeller* | |
17629 | ||
17630 | * Reorganise the PKCS#7 library and get rid of some of the more obvious | |
17631 | problems: find RecipientInfo structure that matches recipient certificate | |
17632 | and initialise the ASN1 structures properly based on passed cipher. | |
17633 | ||
17634 | *Steve Henson* | |
17635 | ||
17636 | * Belatedly make the BN tests actually check the results. | |
17637 | ||
17638 | *Ben Laurie* | |
17639 | ||
17640 | * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion | |
17641 | to and from BNs: it was completely broken. New compilation option | |
17642 | NEG_PUBKEY_BUG to allow for some broken certificates that encode public | |
17643 | key elements as negative integers. | |
17644 | ||
17645 | *Steve Henson* | |
17646 | ||
17647 | * Reorganize and speed up MD5. | |
17648 | ||
17649 | *Andy Polyakov <appro@fy.chalmers.se>* | |
17650 | ||
17651 | * VMS support. | |
17652 | ||
17653 | *Richard Levitte <richard@levitte.org>* | |
17654 | ||
17655 | * New option -out to asn1parse to allow the parsed structure to be | |
17656 | output to a file. This is most useful when combined with the -strparse | |
17657 | option to examine the output of things like OCTET STRINGS. | |
17658 | ||
17659 | *Steve Henson* | |
17660 | ||
17661 | * Make SSL library a little more fool-proof by not requiring any longer | |
17662 | that `SSL_set_{accept,connect}_state` be called before | |
17663 | `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted | |
17664 | in many applications because usually everything *appeared* to work as | |
17665 | intended anyway -- now it really works as intended). | |
17666 | ||
17667 | *Bodo Moeller* | |
17668 | ||
17669 | * Move openssl.cnf out of lib/. | |
17670 | ||
17671 | *Ulf Möller* | |
17672 | ||
17673 | * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall | |
17674 | -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes | |
17675 | -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+ | |
17676 | ||
17677 | *Ralf S. Engelschall* | |
17678 | ||
17679 | * Various fixes to the EVP and PKCS#7 code. It may now be able to | |
17680 | handle PKCS#7 enveloped data properly. | |
17681 | ||
17682 | *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve* | |
17683 | ||
17684 | * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of | |
17685 | copying pointers. The cert_st handling is changed by this in | |
17686 | various ways (and thus what used to be known as ctx->default_cert | |
17687 | is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert` | |
17688 | any longer when s->cert does not give us what we need). | |
17689 | ssl_cert_instantiate becomes obsolete by this change. | |
17690 | As soon as we've got the new code right (possibly it already is?), | |
17691 | we have solved a couple of bugs of the earlier code where s->cert | |
17692 | was used as if it could not have been shared with other SSL structures. | |
17693 | ||
17694 | Note that using the SSL API in certain dirty ways now will result | |
17695 | in different behaviour than observed with earlier library versions: | |
17696 | Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx) | |
17697 | does not influence s as it used to. | |
17698 | ||
17699 | In order to clean up things more thoroughly, inside SSL_SESSION | |
17700 | we don't use CERT any longer, but a new structure SESS_CERT | |
17701 | that holds per-session data (if available); currently, this is | |
17702 | the peer's certificate chain and, for clients, the server's certificate | |
17703 | and temporary key. CERT holds only those values that can have | |
17704 | meaningful defaults in an SSL_CTX. | |
17705 | ||
17706 | *Bodo Moeller* | |
17707 | ||
17708 | * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure | |
17709 | from the internal representation. Various PKCS#7 fixes: remove some | |
17710 | evil casts and set the enc_dig_alg field properly based on the signing | |
17711 | key type. | |
17712 | ||
17713 | *Steve Henson* | |
17714 | ||
17715 | * Allow PKCS#12 password to be set from the command line or the | |
17716 | environment. Let 'ca' get its config file name from the environment | |
17717 | variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' | |
17718 | and 'x509'). | |
17719 | ||
17720 | *Steve Henson* | |
17721 | ||
17722 | * Allow certificate policies extension to use an IA5STRING for the | |
17723 | organization field. This is contrary to the PKIX definition but | |
17724 | VeriSign uses it and IE5 only recognises this form. Document 'x509' | |
17725 | extension option. | |
17726 | ||
17727 | *Steve Henson* | |
17728 | ||
17729 | * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, | |
17730 | without disallowing inline assembler and the like for non-pedantic builds. | |
17731 | ||
17732 | *Ben Laurie* | |
17733 | ||
17734 | * Support Borland C++ builder. | |
17735 | ||
17736 | *Janez Jere <jj@void.si>, modified by Ulf Möller* | |
17737 | ||
17738 | * Support Mingw32. | |
17739 | ||
17740 | *Ulf Möller* | |
17741 | ||
17742 | * SHA-1 cleanups and performance enhancements. | |
17743 | ||
17744 | *Andy Polyakov <appro@fy.chalmers.se>* | |
17745 | ||
17746 | * Sparc v8plus assembler for the bignum library. | |
17747 | ||
17748 | *Andy Polyakov <appro@fy.chalmers.se>* | |
17749 | ||
17750 | * Accept any -xxx and +xxx compiler options in Configure. | |
17751 | ||
17752 | *Ulf Möller* | |
17753 | ||
17754 | * Update HPUX configuration. | |
17755 | ||
17756 | *Anonymous* | |
17757 | ||
17758 | * Add missing `sk_<type>_unshift()` function to safestack.h | |
17759 | ||
17760 | *Ralf S. Engelschall* | |
17761 | ||
17762 | * New function SSL_CTX_use_certificate_chain_file that sets the | |
17763 | "extra_cert"s in addition to the certificate. (This makes sense | |
17764 | only for "PEM" format files, as chains as a whole are not | |
17765 | DER-encoded.) | |
17766 | ||
17767 | *Bodo Moeller* | |
17768 | ||
17769 | * Support verify_depth from the SSL API. | |
17770 | x509_vfy.c had what can be considered an off-by-one-error: | |
17771 | Its depth (which was not part of the external interface) | |
17772 | was actually counting the number of certificates in a chain; | |
17773 | now it really counts the depth. | |
17774 | ||
17775 | *Bodo Moeller* | |
17776 | ||
17777 | * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used | |
17778 | instead of X509err, which often resulted in confusing error | |
17779 | messages since the error codes are not globally unique | |
17780 | (e.g. an alleged error in ssl3_accept when a certificate | |
17781 | didn't match the private key). | |
17782 | ||
17783 | * New function SSL_CTX_set_session_id_context that allows to set a default | |
17784 | value (so that you don't need SSL_set_session_id_context for each | |
17785 | connection using the SSL_CTX). | |
17786 | ||
17787 | *Bodo Moeller* | |
17788 | ||
17789 | * OAEP decoding bug fix. | |
17790 | ||
17791 | *Ulf Möller* | |
17792 | ||
17793 | * Support INSTALL_PREFIX for package builders, as proposed by | |
17794 | David Harris. | |
17795 | ||
17796 | *Bodo Moeller* | |
17797 | ||
17798 | * New Configure options "threads" and "no-threads". For systems | |
17799 | where the proper compiler options are known (currently Solaris | |
17800 | and Linux), "threads" is the default. | |
17801 | ||
17802 | *Bodo Moeller* | |
17803 | ||
17804 | * New script util/mklink.pl as a faster substitute for util/mklink.sh. | |
17805 | ||
17806 | *Bodo Moeller* | |
17807 | ||
17808 | * Install various scripts to $(OPENSSLDIR)/misc, not to | |
17809 | $(INSTALLTOP)/bin -- they shouldn't clutter directories | |
17810 | such as /usr/local/bin. | |
17811 | ||
17812 | *Bodo Moeller* | |
17813 | ||
17814 | * "make linux-shared" to build shared libraries. | |
17815 | ||
17816 | *Niels Poppe <niels@netbox.org>* | |
17817 | ||
17818 | * New Configure option `no-<cipher>` (rsa, idea, rc5, ...). | |
17819 | ||
17820 | *Ulf Möller* | |
17821 | ||
17822 | * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for | |
17823 | extension adding in x509 utility. | |
17824 | ||
17825 | *Steve Henson* | |
17826 | ||
17827 | * Remove NOPROTO sections and error code comments. | |
17828 | ||
17829 | *Ulf Möller* | |
17830 | ||
17831 | * Partial rewrite of the DEF file generator to now parse the ANSI | |
17832 | prototypes. | |
17833 | ||
17834 | *Steve Henson* | |
17835 | ||
17836 | * New Configure options --prefix=DIR and --openssldir=DIR. | |
17837 | ||
17838 | *Ulf Möller* | |
17839 | ||
17840 | * Complete rewrite of the error code script(s). It is all now handled | |
17841 | by one script at the top level which handles error code gathering, | |
17842 | header rewriting and C source file generation. It should be much better | |
17843 | than the old method: it now uses a modified version of Ulf's parser to | |
17844 | read the ANSI prototypes in all header files (thus the old K&R definitions | |
17845 | aren't needed for error creation any more) and do a better job of | |
17846 | translating function codes into names. The old 'ASN1 error code embedded | |
17847 | in a comment' is no longer necessary and it doesn't use .err files which | |
17848 | have now been deleted. Also the error code call doesn't have to appear all | |
17849 | on one line (which resulted in some large lines...). | |
17850 | ||
17851 | *Steve Henson* | |
17852 | ||
17853 | * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`. | |
17854 | ||
17855 | *Bodo Moeller* | |
17856 | ||
17857 | * Change behaviour of ssl2_read when facing length-0 packets: Don't return | |
17858 | 0 (which usually indicates a closed connection), but continue reading. | |
17859 | ||
17860 | *Bodo Moeller* | |
17861 | ||
17862 | * Fix some race conditions. | |
17863 | ||
17864 | *Bodo Moeller* | |
17865 | ||
17866 | * Add support for CRL distribution points extension. Add Certificate | |
17867 | Policies and CRL distribution points documentation. | |
17868 | ||
17869 | *Steve Henson* | |
17870 | ||
17871 | * Move the autogenerated header file parts to crypto/opensslconf.h. | |
17872 | ||
17873 | *Ulf Möller* | |
17874 | ||
17875 | * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of | |
17876 | 8 of keying material. Merlin has also confirmed interop with this fix | |
17877 | between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. | |
17878 | ||
17879 | *Merlin Hughes <merlin@baltimore.ie>* | |
17880 | ||
17881 | * Fix lots of warnings. | |
17882 | ||
17883 | *Richard Levitte <levitte@stacken.kth.se>* | |
17884 | ||
17885 | * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if | |
17886 | the directory spec didn't end with a LIST_SEPARATOR_CHAR. | |
17887 | ||
17888 | *Richard Levitte <levitte@stacken.kth.se>* | |
17889 | ||
17890 | * Fix problems with sizeof(long) == 8. | |
17891 | ||
17892 | *Andy Polyakov <appro@fy.chalmers.se>* | |
17893 | ||
17894 | * Change functions to ANSI C. | |
17895 | ||
17896 | *Ulf Möller* | |
17897 | ||
17898 | * Fix typos in error codes. | |
17899 | ||
17900 | *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller* | |
17901 | ||
17902 | * Remove defunct assembler files from Configure. | |
17903 | ||
17904 | *Ulf Möller* | |
17905 | ||
17906 | * SPARC v8 assembler BIGNUM implementation. | |
17907 | ||
17908 | *Andy Polyakov <appro@fy.chalmers.se>* | |
17909 | ||
17910 | * Support for Certificate Policies extension: both print and set. | |
17911 | Various additions to support the r2i method this uses. | |
17912 | ||
17913 | *Steve Henson* | |
17914 | ||
17915 | * A lot of constification, and fix a bug in X509_NAME_oneline() that could | |
17916 | return a const string when you are expecting an allocated buffer. | |
17917 | ||
17918 | *Ben Laurie* | |
17919 | ||
17920 | * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE | |
17921 | types DirectoryString and DisplayText. | |
17922 | ||
17923 | *Steve Henson* | |
17924 | ||
17925 | * Add code to allow r2i extensions to access the configuration database, | |
17926 | add an LHASH database driver and add several ctx helper functions. | |
17927 | ||
17928 | *Steve Henson* | |
17929 | ||
17930 | * Fix an evil bug in bn_expand2() which caused various BN functions to | |
17931 | fail when they extended the size of a BIGNUM. | |
17932 | ||
17933 | *Steve Henson* | |
17934 | ||
17935 | * Various utility functions to handle SXNet extension. Modify mkdef.pl to | |
17936 | support typesafe stack. | |
17937 | ||
17938 | *Steve Henson* | |
17939 | ||
17940 | * Fix typo in SSL_[gs]et_options(). | |
17941 | ||
17942 | *Nils Frostberg <nils@medcom.se>* | |
17943 | ||
17944 | * Delete various functions and files that belonged to the (now obsolete) | |
17945 | old X509V3 handling code. | |
17946 | ||
17947 | *Steve Henson* | |
17948 | ||
17949 | * New Configure option "rsaref". | |
17950 | ||
17951 | *Ulf Möller* | |
17952 | ||
17953 | * Don't auto-generate pem.h. | |
17954 | ||
17955 | *Bodo Moeller* | |
17956 | ||
17957 | * Introduce type-safe ASN.1 SETs. | |
17958 | ||
17959 | *Ben Laurie* | |
17960 | ||
17961 | * Convert various additional casted stacks to type-safe STACK_OF() variants. | |
17962 | ||
17963 | *Ben Laurie, Ralf S. Engelschall, Steve Henson* | |
17964 | ||
17965 | * Introduce type-safe STACKs. This will almost certainly break lots of code | |
17966 | that links with OpenSSL (well at least cause lots of warnings), but fear | |
17967 | not: the conversion is trivial, and it eliminates loads of evil casts. A | |
17968 | few STACKed things have been converted already. Feel free to convert more. | |
17969 | In the fullness of time, I'll do away with the STACK type altogether. | |
17970 | ||
17971 | *Ben Laurie* | |
17972 | ||
17973 | * Add `openssl ca -revoke <certfile>` facility which revokes a certificate | |
17974 | specified in `<certfile>` by updating the entry in the index.txt file. | |
17975 | This way one no longer has to edit the index.txt file manually for | |
17976 | revoking a certificate. The -revoke option does the gory details now. | |
17977 | ||
17978 | *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall* | |
17979 | ||
17980 | * Fix `openssl crl -noout -text` combination where `-noout` killed the | |
17981 | `-text` option at all and this way the `-noout -text` combination was | |
17982 | inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`. | |
17983 | ||
17984 | *Ralf S. Engelschall* | |
17985 | ||
17986 | * Make sure a corresponding plain text error message exists for the | |
17987 | X509_V_ERR_CERT_REVOKED/23 error number which can occur when a | |
17988 | verify callback function determined that a certificate was revoked. | |
17989 | ||
17990 | *Ralf S. Engelschall* | |
17991 | ||
17992 | * Bugfix: In test/testenc, don't test `openssl <cipher>` for | |
17993 | ciphers that were excluded, e.g. by -DNO_IDEA. Also, test | |
17994 | all available ciphers including rc5, which was forgotten until now. | |
17995 | In order to let the testing shell script know which algorithms | |
17996 | are available, a new (up to now undocumented) command | |
17997 | `openssl list-cipher-commands` is used. | |
17998 | ||
17999 | *Bodo Moeller* | |
18000 | ||
18001 | * Bugfix: s_client occasionally would sleep in select() when | |
18002 | it should have checked SSL_pending() first. | |
18003 | ||
18004 | *Bodo Moeller* | |
18005 | ||
18006 | * New functions DSA_do_sign and DSA_do_verify to provide access to | |
18007 | the raw DSA values prior to ASN.1 encoding. | |
18008 | ||
18009 | *Ulf Möller* | |
18010 | ||
18011 | * Tweaks to Configure | |
18012 | ||
18013 | *Niels Poppe <niels@netbox.org>* | |
18014 | ||
18015 | * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, | |
18016 | yet... | |
18017 | ||
18018 | *Steve Henson* | |
18019 | ||
18020 | * New variables $(RANLIB) and $(PERL) in the Makefiles. | |
18021 | ||
18022 | *Ulf Möller* | |
18023 | ||
18024 | * New config option to avoid instructions that are illegal on the 80386. | |
18025 | The default code is faster, but requires at least a 486. | |
18026 | ||
18027 | *Ulf Möller* | |
18028 | ||
18029 | * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and | |
18030 | SSL2_SERVER_VERSION (not used at all) macros, which are now the | |
18031 | same as SSL2_VERSION anyway. | |
18032 | ||
18033 | *Bodo Moeller* | |
18034 | ||
18035 | * New "-showcerts" option for s_client. | |
18036 | ||
18037 | *Bodo Moeller* | |
18038 | ||
18039 | * Still more PKCS#12 integration. Add pkcs12 application to openssl | |
18040 | application. Various cleanups and fixes. | |
18041 | ||
18042 | *Steve Henson* | |
18043 | ||
18044 | * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and | |
18045 | modify error routines to work internally. Add error codes and PBE init | |
18046 | to library startup routines. | |
18047 | ||
18048 | *Steve Henson* | |
18049 | ||
18050 | * Further PKCS#12 integration. Added password based encryption, PKCS#8 and | |
18051 | packing functions to asn1 and evp. Changed function names and error | |
18052 | codes along the way. | |
18053 | ||
18054 | *Steve Henson* | |
18055 | ||
18056 | * PKCS12 integration: and so it begins... First of several patches to | |
18057 | slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 | |
18058 | objects to objects.h | |
18059 | ||
18060 | *Steve Henson* | |
18061 | ||
18062 | * Add a new 'indent' option to some X509V3 extension code. Initial ASN1 | |
18063 | and display support for Thawte strong extranet extension. | |
18064 | ||
18065 | *Steve Henson* | |
18066 | ||
18067 | * Add LinuxPPC support. | |
18068 | ||
18069 | *Jeff Dubrule <igor@pobox.org>* | |
18070 | ||
18071 | * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to | |
18072 | bn_div_words in alpha.s. | |
18073 | ||
18074 | *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie* | |
18075 | ||
18076 | * Make sure the RSA OAEP test is skipped under -DRSAref because | |
18077 | OAEP isn't supported when OpenSSL is built with RSAref. | |
18078 | ||
18079 | *Ulf Moeller <ulf@fitug.de>* | |
18080 | ||
18081 | * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h | |
18082 | so they no longer are missing under -DNOPROTO. | |
18083 | ||
18084 | *Soren S. Jorvang <soren@t.dk>* | |
18085 | ||
18086 | ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999] | |
18087 | ||
18088 | * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still | |
18089 | doesn't work when the session is reused. Coming soon! | |
18090 | ||
18091 | *Ben Laurie* | |
18092 | ||
18093 | * Fix a security hole, that allows sessions to be reused in the wrong | |
18094 | context thus bypassing client cert protection! All software that uses | |
18095 | client certs and session caches in multiple contexts NEEDS PATCHING to | |
18096 | allow session reuse! A fuller solution is in the works. | |
18097 | ||
18098 | *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)* | |
18099 | ||
18100 | * Some more source tree cleanups (removed obsolete files | |
18101 | crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed | |
18102 | permission on "config" script to be executable) and a fix for the INSTALL | |
18103 | document. | |
18104 | ||
18105 | *Ulf Moeller <ulf@fitug.de>* | |
18106 | ||
18107 | * Remove some legacy and erroneous uses of malloc, free instead of | |
18108 | Malloc, Free. | |
18109 | ||
18110 | *Lennart Bang <lob@netstream.se>, with minor changes by Steve* | |
18111 | ||
18112 | * Make rsa_oaep_test return non-zero on error. | |
18113 | ||
18114 | *Ulf Moeller <ulf@fitug.de>* | |
18115 | ||
18116 | * Add support for native Solaris shared libraries. Configure | |
18117 | solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice | |
18118 | if someone would make that last step automatic. | |
18119 | ||
18120 | *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>* | |
18121 | ||
18122 | * ctx_size was not built with the right compiler during "make links". Fixed. | |
18123 | ||
18124 | *Ben Laurie* | |
18125 | ||
18126 | * Change the meaning of 'ALL' in the cipher list. It now means "everything | |
18127 | except NULL ciphers". This means the default cipher list will no longer | |
18128 | enable NULL ciphers. They need to be specifically enabled e.g. with | |
18129 | the string "DEFAULT:eNULL". | |
18130 | ||
18131 | *Steve Henson* | |
18132 | ||
18133 | * Fix to RSA private encryption routines: if p < q then it would | |
18134 | occasionally produce an invalid result. This will only happen with | |
18135 | externally generated keys because OpenSSL (and SSLeay) ensure p > q. | |
18136 | ||
18137 | *Steve Henson* | |
18138 | ||
18139 | * Be less restrictive and allow also `perl util/perlpath.pl | |
18140 | /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`, | |
18141 | because this way one can also use an interpreter named `perl5` (which is | |
18142 | usually the name of Perl 5.xxx on platforms where an Perl 4.x is still | |
18143 | installed as `perl`). | |
18144 | ||
18145 | *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>* | |
18146 | ||
18147 | * Let util/clean-depend.pl work also with older Perl 5.00x versions. | |
18148 | ||
18149 | *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>* | |
18150 | ||
18151 | * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add | |
18152 | advapi32.lib to Win32 build and change the pem test comparison | |
18153 | to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the | |
18154 | suggestion). Fix misplaced ASNI prototypes and declarations in evp.h | |
18155 | and crypto/des/ede_cbcm_enc.c. | |
18156 | ||
18157 | *Steve Henson* | |
18158 | ||
18159 | * DES quad checksum was broken on big-endian architectures. Fixed. | |
18160 | ||
18161 | *Ben Laurie* | |
18162 | ||
18163 | * Comment out two functions in bio.h that aren't implemented. Fix up the | |
18164 | Win32 test batch file so it (might) work again. The Win32 test batch file | |
18165 | is horrible: I feel ill.... | |
18166 | ||
18167 | *Steve Henson* | |
18168 | ||
18169 | * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected | |
18170 | in e_os.h. Audit of header files to check ANSI and non ANSI | |
18171 | sections: 10 functions were absent from non ANSI section and not exported | |
18172 | from Windows DLLs. Fixed up libeay.num for new functions. | |
18173 | ||
18174 | *Steve Henson* | |
18175 | ||
18176 | * Make `openssl version` output lines consistent. | |
18177 | ||
18178 | *Ralf S. Engelschall* | |
18179 | ||
18180 | * Fix Win32 symbol export lists for BIO functions: Added | |
18181 | BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data | |
18182 | to ms/libeay{16,32}.def. | |
18183 | ||
18184 | *Ralf S. Engelschall* | |
18185 | ||
18186 | * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled | |
18187 | fine under Unix and passes some trivial tests I've now added. But the | |
18188 | whole stuff is horribly incomplete, so a README.1ST with a disclaimer was | |
18189 | added to make sure no one expects that this stuff really works in the | |
18190 | OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources | |
18191 | up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and | |
18192 | openssl_bio.xs. | |
18193 | ||
18194 | *Ralf S. Engelschall* | |
18195 | ||
18196 | * Fix the generation of two part addresses in perl. | |
18197 | ||
18198 | *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie* | |
18199 | ||
18200 | * Add config entry for Linux on MIPS. | |
18201 | ||
18202 | *John Tobey <jtobey@channel1.com>* | |
18203 | ||
18204 | * Make links whenever Configure is run, unless we are on Windoze. | |
18205 | ||
18206 | *Ben Laurie* | |
18207 | ||
18208 | * Permit extensions to be added to CRLs using crl_section in openssl.cnf. | |
18209 | Currently only issuerAltName and AuthorityKeyIdentifier make any sense | |
18210 | in CRLs. | |
18211 | ||
18212 | *Steve Henson* | |
18213 | ||
18214 | * Add a useful kludge to allow package maintainers to specify compiler and | |
18215 | other platforms details on the command line without having to patch the | |
18216 | Configure script every time: One now can use | |
18217 | `perl Configure <id>:<details>`, | |
18218 | i.e. platform ids are allowed to have details appended | |
18219 | to them (separated by colons). This is treated as there would be a static | |
18220 | pre-configured entry in Configure's %table under key `<id>` with value | |
18221 | `<details>` and `perl Configure <id>` is called. So, when you want to | |
18222 | perform a quick test-compile under FreeBSD 3.1 with pgcc and without | |
18223 | assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"` | |
18224 | now, which overrides the FreeBSD-elf entry on-the-fly. | |
18225 | ||
18226 | *Ralf S. Engelschall* | |
18227 | ||
18228 | * Disable new TLS1 ciphersuites by default: they aren't official yet. | |
18229 | ||
18230 | *Ben Laurie* | |
18231 | ||
18232 | * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified | |
18233 | on the `perl Configure ...` command line. This way one can compile | |
18234 | OpenSSL libraries with Position Independent Code (PIC) which is needed | |
18235 | for linking it into DSOs. | |
18236 | ||
18237 | *Ralf S. Engelschall* | |
18238 | ||
18239 | * Remarkably, export ciphers were totally broken and no-one had noticed! | |
18240 | Fixed. | |
18241 | ||
18242 | *Ben Laurie* | |
18243 | ||
18244 | * Cleaned up the LICENSE document: The official contact for any license | |
18245 | questions now is the OpenSSL core team under openssl-core@openssl.org. | |
18246 | And add a paragraph about the dual-license situation to make sure people | |
18247 | recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply | |
18248 | to the OpenSSL toolkit. | |
18249 | ||
18250 | *Ralf S. Engelschall* | |
18251 | ||
18252 | * General source tree makefile cleanups: Made `making xxx in yyy...` | |
18253 | display consistent in the source tree and replaced `/bin/rm` by `rm`. | |
18254 | Additionally cleaned up the `make links` target: Remove unnecessary | |
18255 | semicolons, subsequent redundant removes, inline point.sh into mklink.sh | |
18256 | to speed processing and no longer clutter the display with confusing | |
18257 | stuff. Instead only the actually done links are displayed. | |
18258 | ||
18259 | *Ralf S. Engelschall* | |
18260 | ||
18261 | * Permit null encryption ciphersuites, used for authentication only. It used | |
18262 | to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. | |
18263 | It is now necessary to set SSL_FORBID_ENULL to prevent the use of null | |
18264 | encryption. | |
18265 | ||
18266 | *Ben Laurie* | |
18267 | ||
18268 | * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder | |
18269 | signed attributes when verifying signatures (this would break them), | |
18270 | the detached data encoding was wrong and public keys obtained using | |
18271 | X509_get_pubkey() weren't freed. | |
18272 | ||
18273 | *Steve Henson* | |
18274 | ||
18275 | * Add text documentation for the BUFFER functions. Also added a work around | |
18276 | to a Win95 console bug. This was triggered by the password read stuff: the | |
18277 | last character typed gets carried over to the next fread(). If you were | |
18278 | generating a new cert request using 'req' for example then the last | |
18279 | character of the passphrase would be CR which would then enter the first | |
18280 | field as blank. | |
18281 | ||
18282 | *Steve Henson* | |
18283 | ||
18284 | * Added the new 'Includes OpenSSL Cryptography Software' button as | |
18285 | doc/openssl_button.{gif,html} which is similar in style to the old SSLeay | |
18286 | button and can be used by applications based on OpenSSL to show the | |
18287 | relationship to the OpenSSL project. | |
18288 | ||
18289 | *Ralf S. Engelschall* | |
18290 | ||
18291 | * Remove confusing variables in function signatures in files | |
18292 | ssl/ssl_lib.c and ssl/ssl.h. | |
18293 | ||
18294 | *Lennart Bong <lob@kulthea.stacken.kth.se>* | |
18295 | ||
18296 | * Don't install bss_file.c under PREFIX/include/ | |
18297 | ||
18298 | *Lennart Bong <lob@kulthea.stacken.kth.se>* | |
18299 | ||
18300 | * Get the Win32 compile working again. Modify mkdef.pl so it can handle | |
18301 | functions that return function pointers and has support for NT specific | |
18302 | stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various | |
18303 | #ifdef WIN32 and WINNTs sprinkled about the place and some changes from | |
18304 | unsigned to signed types: this was killing the Win32 compile. | |
18305 | ||
18306 | *Steve Henson* | |
18307 | ||
18308 | * Add new certificate file to stack functions, | |
18309 | SSL_add_dir_cert_subjects_to_stack() and | |
18310 | SSL_add_file_cert_subjects_to_stack(). These largely supplant | |
18311 | SSL_load_client_CA_file(), and can be used to add multiple certs easily | |
18312 | to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). | |
18313 | This means that Apache-SSL and similar packages don't have to mess around | |
18314 | to add as many CAs as they want to the preferred list. | |
18315 | ||
18316 | *Ben Laurie* | |
18317 | ||
18318 | * Experiment with doxygen documentation. Currently only partially applied to | |
18319 | ssl/ssl_lib.c. | |
18320 | See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with | |
18321 | openssl.doxy as the configuration file. | |
18322 | ||
18323 | *Ben Laurie* | |
18324 | ||
18325 | * Get rid of remaining C++-style comments which strict C compilers hate. | |
18326 | ||
18327 | *Ralf S. Engelschall, pointed out by Carlos Amengual* | |
18328 | ||
18329 | * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not | |
18330 | compiled in by default: it has problems with large keys. | |
18331 | ||
18332 | *Steve Henson* | |
18333 | ||
18334 | * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and | |
18335 | DH private keys and/or callback functions which directly correspond to | |
18336 | their SSL_CTX_xxx() counterparts but work on a per-connection basis. This | |
18337 | is needed for applications which have to configure certificates on a | |
18338 | per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis | |
18339 | (e.g. s_server). | |
18340 | For the RSA certificate situation is makes no difference, but | |
18341 | for the DSA certificate situation this fixes the "no shared cipher" | |
18342 | problem where the OpenSSL cipher selection procedure failed because the | |
18343 | temporary keys were not overtaken from the context and the API provided | |
18344 | no way to reconfigure them. | |
18345 | The new functions now let applications reconfigure the stuff and they | |
18346 | are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, | |
18347 | SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new | |
18348 | non-public-API function ssl_cert_instantiate() is used as a helper | |
18349 | function and also to reduce code redundancy inside ssl_rsa.c. | |
18350 | ||
18351 | *Ralf S. Engelschall* | |
18352 | ||
18353 | * Move s_server -dcert and -dkey options out of the undocumented feature | |
18354 | area because they are useful for the DSA situation and should be | |
18355 | recognized by the users. | |
18356 | ||
18357 | *Ralf S. Engelschall* | |
18358 | ||
18359 | * Fix the cipher decision scheme for export ciphers: the export bits are | |
18360 | *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within | |
18361 | SSL_EXP_MASK. So, the original variable has to be used instead of the | |
18362 | already masked variable. | |
18363 | ||
18364 | *Richard Levitte <levitte@stacken.kth.se>* | |
18365 | ||
18366 | * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c | |
18367 | ||
18368 | *Richard Levitte <levitte@stacken.kth.se>* | |
18369 | ||
18370 | * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() | |
18371 | from `int` to `unsigned int` because it is a length and initialized by | |
18372 | EVP_DigestFinal() which expects an `unsigned int *`. | |
18373 | ||
18374 | *Richard Levitte <levitte@stacken.kth.se>* | |
18375 | ||
18376 | * Don't hard-code path to Perl interpreter on shebang line of Configure | |
18377 | script. Instead use the usual Shell->Perl transition trick. | |
18378 | ||
18379 | *Ralf S. Engelschall* | |
18380 | ||
18381 | * Make `openssl x509 -noout -modulus`' functional also for DSA certificates | |
18382 | (in addition to RSA certificates) to match the behaviour of `openssl dsa | |
18383 | -noout -modulus` as it's already the case for `openssl rsa -noout | |
18384 | -modulus`. For RSA the -modulus is the real "modulus" while for DSA | |
18385 | currently the public key is printed (a decision which was already done by | |
18386 | `openssl dsa -modulus` in the past) which serves a similar purpose. | |
18387 | Additionally the NO_RSA no longer completely removes the whole -modulus | |
18388 | option; it now only avoids using the RSA stuff. Same applies to NO_DSA | |
18389 | now, too. | |
18390 | ||
18391 | *Ralf S. Engelschall* | |
18392 | ||
18393 | * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested | |
18394 | BIO. See the source (crypto/evp/bio_ok.c) for more info. | |
18395 | ||
18396 | *Arne Ansper <arne@ats.cyber.ee>* | |
18397 | ||
18398 | * Dump the old yucky req code that tried (and failed) to allow raw OIDs | |
18399 | to be added. Now both 'req' and 'ca' can use new objects defined in the | |
18400 | config file. | |
18401 | ||
18402 | *Steve Henson* | |
18403 | ||
18404 | * Add cool BIO that does syslog (or event log on NT). | |
18405 | ||
18406 | *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie* | |
18407 | ||
18408 | * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, | |
18409 | TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and | |
18410 | TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher | |
18411 | Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. | |
18412 | ||
18413 | *Ben Laurie* | |
18414 | ||
18415 | * Add preliminary config info for new extension code. | |
18416 | ||
18417 | *Steve Henson* | |
18418 | ||
18419 | * Make RSA_NO_PADDING really use no padding. | |
18420 | ||
18421 | *Ulf Moeller <ulf@fitug.de>* | |
18422 | ||
18423 | * Generate errors when private/public key check is done. | |
18424 | ||
18425 | *Ben Laurie* | |
18426 | ||
18427 | * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support | |
18428 | for some CRL extensions and new objects added. | |
18429 | ||
18430 | *Steve Henson* | |
18431 | ||
18432 | * Really fix the ASN1 IMPLICIT bug this time... Partial support for private | |
18433 | key usage extension and fuller support for authority key id. | |
18434 | ||
18435 | *Steve Henson* | |
18436 | ||
18437 | * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved | |
18438 | padding method for RSA, which is recommended for new applications in PKCS | |
18439 | #1 v2.0 (RFC 2437, October 1998). | |
18440 | OAEP (Optimal Asymmetric Encryption Padding) has better theoretical | |
18441 | foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure | |
18442 | against Bleichbacher's attack on RSA. | |
18443 | *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by | |
18444 | Ben Laurie* | |
18445 | ||
18446 | * Updates to the new SSL compression code | |
18447 | ||
18448 | *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)* | |
18449 | ||
18450 | * Fix so that the version number in the master secret, when passed | |
18451 | via RSA, checks that if TLS was proposed, but we roll back to SSLv3 | |
18452 | (because the server will not accept higher), that the version number | |
18453 | is 0x03,0x01, not 0x03,0x00 | |
18454 | ||
18455 | *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)* | |
18456 | ||
18457 | * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory | |
18458 | leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes | |
18459 | in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`. | |
18460 | ||
18461 | *Steve Henson* | |
18462 | ||
18463 | * Support for RAW extensions where an arbitrary extension can be | |
18464 | created by including its DER encoding. See `apps/openssl.cnf` for | |
18465 | an example. | |
18466 | ||
18467 | *Steve Henson* | |
18468 | ||
18469 | * Make sure latest Perl versions don't interpret some generated C array | |
18470 | code as Perl array code in the crypto/err/err_genc.pl script. | |
18471 | ||
18472 | *Lars Weber <3weber@informatik.uni-hamburg.de>* | |
18473 | ||
18474 | * Modify ms/do_ms.bat to not generate assembly language makefiles since | |
18475 | not many people have the assembler. Various Win32 compilation fixes and | |
18476 | update to the INSTALL.W32 file with (hopefully) more accurate Win32 | |
18477 | build instructions. | |
18478 | ||
18479 | *Steve Henson* | |
18480 | ||
18481 | * Modify configure script 'Configure' to automatically create crypto/date.h | |
18482 | file under Win32 and also build pem.h from pem.org. New script | |
18483 | util/mkfiles.pl to create the MINFO file on environments that can't do a | |
18484 | 'make files': perl util/mkfiles.pl >MINFO should work. | |
18485 | ||
18486 | *Steve Henson* | |
18487 | ||
18488 | * Major rework of DES function declarations, in the pursuit of correctness | |
18489 | and purity. As a result, many evil casts evaporated, and some weirdness, | |
18490 | too. You may find this causes warnings in your code. Zapping your evil | |
18491 | casts will probably fix them. Mostly. | |
18492 | ||
18493 | *Ben Laurie* | |
18494 | ||
18495 | * Fix for a typo in asn1.h. Bug fix to object creation script | |
18496 | obj_dat.pl. It considered a zero in an object definition to mean | |
18497 | "end of object": none of the objects in objects.h have any zeros | |
18498 | so it wasn't spotted. | |
18499 | ||
18500 | *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>* | |
18501 | ||
18502 | * Add support for Triple DES Cipher Block Chaining with Output Feedback | |
18503 | Masking (CBCM). In the absence of test vectors, the best I have been able | |
18504 | to do is check that the decrypt undoes the encrypt, so far. Send me test | |
18505 | vectors if you have them. | |
18506 | ||
18507 | *Ben Laurie* | |
18508 | ||
18509 | * Correct calculation of key length for export ciphers (too much space was | |
18510 | allocated for null ciphers). This has not been tested! | |
18511 | ||
18512 | *Ben Laurie* | |
18513 | ||
18514 | * Modifications to the mkdef.pl for Win32 DEF file creation. The usage | |
18515 | message is now correct (it understands "crypto" and "ssl" on its | |
18516 | command line). There is also now an "update" option. This will update | |
18517 | the util/ssleay.num and util/libeay.num files with any new functions. | |
18518 | If you do a: | |
18519 | perl util/mkdef.pl crypto ssl update | |
18520 | it will update them. | |
18521 | ||
18522 | *Steve Henson* | |
18523 | ||
18524 | * Overhauled the Perl interface: | |
18525 | - ported BN stuff to OpenSSL's different BN library | |
18526 | - made the perl/ source tree CVS-aware | |
18527 | - renamed the package from SSLeay to OpenSSL (the files still contain | |
18528 | their history because I've copied them in the repository) | |
18529 | - removed obsolete files (the test scripts will be replaced | |
18530 | by better Test::Harness variants in the future) | |
18531 | ||
18532 | *Ralf S. Engelschall* | |
18533 | ||
18534 | * First cut for a very conservative source tree cleanup: | |
18535 | 1. merge various obsolete readme texts into doc/ssleay.txt | |
18536 | where we collect the old documents and readme texts. | |
18537 | 2. remove the first part of files where I'm already sure that we no | |
18538 | longer need them because of three reasons: either they are just temporary | |
18539 | files which were left by Eric or they are preserved original files where | |
18540 | I've verified that the diff is also available in the CVS via "cvs diff | |
18541 | -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for | |
18542 | the crypto/md/ stuff). | |
18543 | ||
18544 | *Ralf S. Engelschall* | |
18545 | ||
18546 | * More extension code. Incomplete support for subject and issuer alt | |
18547 | name, issuer and authority key id. Change the i2v function parameters | |
18548 | and add an extra 'crl' parameter in the X509V3_CTX structure: guess | |
18549 | what that's for :-) Fix to ASN1 macro which messed up | |
18550 | IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. | |
18551 | ||
18552 | *Steve Henson* | |
18553 | ||
18554 | * Preliminary support for ENUMERATED type. This is largely copied from the | |
18555 | INTEGER code. | |
18556 | ||
18557 | *Steve Henson* | |
18558 | ||
18559 | * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. | |
18560 | ||
18561 | *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)* | |
18562 | ||
18563 | * Make sure `make rehash` target really finds the `openssl` program. | |
18564 | ||
18565 | *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>* | |
18566 | ||
18567 | * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd | |
18568 | like to hear about it if this slows down other processors. | |
18569 | ||
18570 | *Ben Laurie* | |
18571 | ||
18572 | * Add CygWin32 platform information to Configure script. | |
18573 | ||
18574 | *Alan Batie <batie@aahz.jf.intel.com>* | |
18575 | ||
18576 | * Fixed ms/32all.bat script: `no_asm` -> `no-asm` | |
18577 | ||
18578 | *Rainer W. Gerling <gerling@mpg-gv.mpg.de>* | |
18579 | ||
18580 | * New program nseq to manipulate netscape certificate sequences | |
18581 | ||
18582 | *Steve Henson* | |
18583 | ||
18584 | * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a | |
18585 | few typos. | |
18586 | ||
18587 | *Steve Henson* | |
18588 | ||
18589 | * Fixes to BN code. Previously the default was to define BN_RECURSION | |
18590 | but the BN code had some problems that would cause failures when | |
18591 | doing certificate verification and some other functions. | |
18592 | ||
18593 | *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)* | |
18594 | ||
18595 | * Add ASN1 and PEM code to support netscape certificate sequences. | |
18596 | ||
18597 | *Steve Henson* | |
18598 | ||
18599 | * Add ASN1 and PEM code to support netscape certificate sequences. | |
18600 | ||
18601 | *Steve Henson* | |
18602 | ||
18603 | * Add several PKIX and private extended key usage OIDs. | |
18604 | ||
18605 | *Steve Henson* | |
18606 | ||
18607 | * Modify the 'ca' program to handle the new extension code. Modify | |
18608 | openssl.cnf for new extension format, add comments. | |
18609 | ||
18610 | *Steve Henson* | |
18611 | ||
18612 | * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' | |
18613 | and add a sample to openssl.cnf so req -x509 now adds appropriate | |
18614 | CA extensions. | |
18615 | ||
18616 | *Steve Henson* | |
18617 | ||
18618 | * Continued X509 V3 changes. Add to other makefiles, integrate with the | |
18619 | error code, add initial support to X509_print() and x509 application. | |
18620 | ||
18621 | *Steve Henson* | |
18622 | ||
18623 | * Takes a deep breath and start adding X509 V3 extension support code. Add | |
18624 | files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this | |
18625 | stuff is currently isolated and isn't even compiled yet. | |
18626 | ||
18627 | *Steve Henson* | |
18628 | ||
18629 | * Continuing patches for GeneralizedTime. Fix up certificate and CRL | |
18630 | ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. | |
18631 | Removed the versions check from X509 routines when loading extensions: | |
18632 | this allows certain broken certificates that don't set the version | |
18633 | properly to be processed. | |
18634 | ||
18635 | *Steve Henson* | |
18636 | ||
18637 | * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another | |
18638 | Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which | |
18639 | can still be regenerated with "make depend". | |
18640 | ||
18641 | *Ben Laurie* | |
18642 | ||
18643 | * Spelling mistake in C version of CAST-128. | |
18644 | ||
18645 | *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>* | |
18646 | ||
18647 | * Changes to the error generation code. The perl script err-code.pl | |
18648 | now reads in the old error codes and retains the old numbers, only | |
18649 | adding new ones if necessary. It also only changes the .err files if new | |
18650 | codes are added. The makefiles have been modified to only insert errors | |
18651 | when needed (to avoid needlessly modifying header files). This is done | |
18652 | by only inserting errors if the .err file is newer than the auto generated | |
18653 | C file. To rebuild all the error codes from scratch (the old behaviour) | |
18654 | either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl | |
18655 | or delete all the .err files. | |
18656 | ||
18657 | *Steve Henson* | |
18658 | ||
18659 | * CAST-128 was incorrectly implemented for short keys. The C version has | |
18660 | been fixed, but is untested. The assembler versions are also fixed, but | |
18661 | new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing | |
18662 | to regenerate it if needed. | |
18663 | *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun | |
18664 | Hagino <itojun@kame.net>* | |
18665 | ||
18666 | * File was opened incorrectly in randfile.c. | |
18667 | ||
18668 | *Ulf Möller <ulf@fitug.de>* | |
18669 | ||
18670 | * Beginning of support for GeneralizedTime. d2i, i2d, check and print | |
18671 | functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or | |
18672 | GeneralizedTime. ASN1_TIME is the proper type used in certificates et | |
18673 | al: it's just almost always a UTCTime. Note this patch adds new error | |
18674 | codes so do a "make errors" if there are problems. | |
18675 | ||
18676 | *Steve Henson* | |
18677 | ||
18678 | * Correct Linux 1 recognition in config. | |
18679 | ||
18680 | *Ulf Möller <ulf@fitug.de>* | |
18681 | ||
18682 | * Remove pointless MD5 hash when using DSA keys in ca. | |
18683 | ||
18684 | *Anonymous <nobody@replay.com>* | |
18685 | ||
18686 | * Generate an error if given an empty string as a cert directory. Also | |
18687 | generate an error if handed NULL (previously returned 0 to indicate an | |
18688 | error, but didn't set one). | |
18689 | ||
18690 | *Ben Laurie, reported by Anonymous <nobody@replay.com>* | |
18691 | ||
18692 | * Add prototypes to SSL methods. Make SSL_write's buffer const, at last. | |
18693 | ||
18694 | *Ben Laurie* | |
18695 | ||
18696 | * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct | |
18697 | parameters. This was causing a warning which killed off the Win32 compile. | |
18698 | ||
18699 | *Steve Henson* | |
18700 | ||
18701 | * Remove C++ style comments from crypto/bn/bn_local.h. | |
18702 | ||
18703 | *Neil Costigan <neil.costigan@celocom.com>* | |
18704 | ||
18705 | * The function OBJ_txt2nid was broken. It was supposed to return a nid | |
18706 | based on a text string, looking up short and long names and finally | |
18707 | "dot" format. The "dot" format stuff didn't work. Added new function | |
18708 | OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote | |
18709 | OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the | |
18710 | OID is not part of the table. | |
18711 | ||
18712 | *Steve Henson* | |
18713 | ||
18714 | * Add prototypes to X509 lookup/verify methods, fixing a bug in | |
18715 | X509_LOOKUP_by_alias(). | |
18716 | ||
18717 | *Ben Laurie* | |
18718 | ||
18719 | * Sort openssl functions by name. | |
18720 | ||
18721 | *Ben Laurie* | |
18722 | ||
18723 | * Get the `gendsa` command working and add it to the `list` command. Remove | |
18724 | encryption from sample DSA keys (in case anyone is interested the password | |
18725 | was "1234"). | |
18726 | ||
18727 | *Steve Henson* | |
18728 | ||
18729 | * Make *all* `*_free` functions accept a NULL pointer. | |
18730 | ||
18731 | *Frans Heymans <fheymans@isaserver.be>* | |
18732 | ||
18733 | * If a DH key is generated in s3_srvr.c, don't blow it by trying to use | |
18734 | NULL pointers. | |
18735 | ||
18736 | *Anonymous <nobody@replay.com>* | |
18737 | ||
18738 | * s_server should send the CAfile as acceptable CAs, not its own cert. | |
18739 | ||
18740 | *Bodo Moeller <3moeller@informatik.uni-hamburg.de>* | |
18741 | ||
18742 | * Don't blow it for numeric `-newkey` arguments to `apps/req`. | |
18743 | ||
18744 | *Bodo Moeller <3moeller@informatik.uni-hamburg.de>* | |
18745 | ||
18746 | * Temp key "for export" tests were wrong in s3_srvr.c. | |
18747 | ||
18748 | *Anonymous <nobody@replay.com>* | |
18749 | ||
18750 | * Add prototype for temp key callback functions | |
18751 | SSL_CTX_set_tmp_{rsa,dh}_callback(). | |
18752 | ||
18753 | *Ben Laurie* | |
18754 | ||
18755 | * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and | |
18756 | DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). | |
18757 | ||
18758 | *Steve Henson* | |
18759 | ||
18760 | * X509_name_add_entry() freed the wrong thing after an error. | |
18761 | ||
18762 | *Arne Ansper <arne@ats.cyber.ee>* | |
18763 | ||
18764 | * rsa_eay.c would attempt to free a NULL context. | |
18765 | ||
18766 | *Arne Ansper <arne@ats.cyber.ee>* | |
18767 | ||
18768 | * BIO_s_socket() had a broken should_retry() on Windoze. | |
18769 | ||
18770 | *Arne Ansper <arne@ats.cyber.ee>* | |
18771 | ||
18772 | * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. | |
18773 | ||
18774 | *Arne Ansper <arne@ats.cyber.ee>* | |
18775 | ||
18776 | * Make sure the already existing X509_STORE->depth variable is initialized | |
18777 | in X509_STORE_new(), but document the fact that this variable is still | |
18778 | unused in the certificate verification process. | |
18779 | ||
18780 | *Ralf S. Engelschall* | |
18781 | ||
18782 | * Fix the various library and `apps/` files to free up pkeys obtained from | |
18783 | X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. | |
18784 | ||
18785 | *Steve Henson* | |
18786 | ||
18787 | * Fix reference counting in X509_PUBKEY_get(). This makes | |
18788 | demos/maurice/example2.c work, amongst others, probably. | |
18789 | ||
18790 | *Steve Henson and Ben Laurie* | |
18791 | ||
18792 | * First cut of a cleanup for `apps/`. First the `ssleay` program is now named | |
18793 | `openssl` and second, the shortcut symlinks for the `openssl <command>` | |
18794 | are no longer created. This way we have a single and consistent command | |
18795 | line interface `openssl <command>`, similar to `cvs <command>`. | |
18796 | ||
18797 | *Ralf S. Engelschall, Paul Sutton and Ben Laurie* | |
18798 | ||
18799 | * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey | |
18800 | BIT STRING wrapper always have zero unused bits. | |
18801 | ||
18802 | *Steve Henson* | |
18803 | ||
18804 | * Add CA.pl, perl version of CA.sh, add extended key usage OID. | |
18805 | ||
18806 | *Steve Henson* | |
18807 | ||
18808 | * Make the top-level INSTALL documentation easier to understand. | |
18809 | ||
18810 | *Paul Sutton* | |
18811 | ||
18812 | * Makefiles updated to exit if an error occurs in a sub-directory | |
18813 | make (including if user presses ^C) [Paul Sutton] | |
18814 | ||
18815 | * Make Montgomery context stuff explicit in RSA data structure. | |
18816 | ||
18817 | *Ben Laurie* | |
18818 | ||
18819 | * Fix build order of pem and err to allow for generated pem.h. | |
18820 | ||
18821 | *Ben Laurie* | |
18822 | ||
18823 | * Fix renumbering bug in X509_NAME_delete_entry(). | |
18824 | ||
18825 | *Ben Laurie* | |
18826 | ||
18827 | * Enhanced the err-ins.pl script so it makes the error library number | |
18828 | global and can add a library name. This is needed for external ASN1 and | |
18829 | other error libraries. | |
18830 | ||
18831 | *Steve Henson* | |
18832 | ||
18833 | * Fixed sk_insert which never worked properly. | |
18834 | ||
18835 | *Steve Henson* | |
18836 | ||
18837 | * Fix ASN1 macros so they can handle indefinite length constructed | |
18838 | EXPLICIT tags. Some non standard certificates use these: they can now | |
18839 | be read in. | |
18840 | ||
18841 | *Steve Henson* | |
18842 | ||
18843 | * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) | |
18844 | into a single doc/ssleay.txt bundle. This way the information is still | |
18845 | preserved but no longer messes up this directory. Now it's new room for | |
18846 | the new set of documentation files. | |
18847 | ||
18848 | *Ralf S. Engelschall* | |
18849 | ||
18850 | * SETs were incorrectly DER encoded. This was a major pain, because they | |
18851 | shared code with SEQUENCEs, which aren't coded the same. This means that | |
18852 | almost everything to do with SETs or SEQUENCEs has either changed name or | |
18853 | number of arguments. | |
18854 | ||
18855 | *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>* | |
18856 | ||
18857 | * Fix test data to work with the above. | |
18858 | ||
18859 | *Ben Laurie* | |
18860 | ||
18861 | * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but | |
18862 | was already fixed by Eric for 0.9.1 it seems. | |
18863 | ||
18864 | *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>* | |
18865 | ||
18866 | * Autodetect FreeBSD3. | |
18867 | ||
18868 | *Ben Laurie* | |
18869 | ||
18870 | * Fix various bugs in Configure. This affects the following platforms: | |
18871 | nextstep | |
18872 | ncr-scde | |
18873 | unixware-2.0 | |
18874 | unixware-2.0-pentium | |
18875 | sco5-cc. | |
18876 | ||
18877 | *Ben Laurie* | |
18878 | ||
18879 | * Eliminate generated files from CVS. Reorder tests to regenerate files | |
18880 | before they are needed. | |
18881 | ||
18882 | *Ben Laurie* | |
18883 | ||
18884 | * Generate Makefile.ssl from Makefile.org (to keep CVS happy). | |
18885 | ||
18886 | *Ben Laurie* | |
18887 | ||
18888 | ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998] | |
18889 | ||
18890 | * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and | |
18891 | changed SSLeay to OpenSSL in version strings. | |
18892 | ||
18893 | *Ralf S. Engelschall* | |
18894 | ||
18895 | * Some fixups to the top-level documents. | |
18896 | ||
18897 | *Paul Sutton* | |
18898 | ||
18899 | * Fixed the nasty bug where rsaref.h was not found under compile-time | |
18900 | because the symlink to include/ was missing. | |
18901 | ||
18902 | *Ralf S. Engelschall* | |
18903 | ||
18904 | * Incorporated the popular no-RSA/DSA-only patches | |
18905 | which allow to compile a RSA-free SSLeay. | |
18906 | ||
18907 | *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall* | |
18908 | ||
18909 | * Fixed nasty rehash problem under `make -f Makefile.ssl links` | |
18910 | when "ssleay" is still not found. | |
18911 | ||
18912 | *Ralf S. Engelschall* | |
18913 | ||
18914 | * Added more platforms to Configure: Cray T3E, HPUX 11, | |
18915 | ||
18916 | *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>* | |
18917 | ||
18918 | * Updated the README file. | |
18919 | ||
18920 | *Ralf S. Engelschall* | |
18921 | ||
18922 | * Added various .cvsignore files in the CVS repository subdirs | |
18923 | to make a "cvs update" really silent. | |
18924 | ||
18925 | *Ralf S. Engelschall* | |
18926 | ||
18927 | * Recompiled the error-definition header files and added | |
18928 | missing symbols to the Win32 linker tables. | |
18929 | ||
18930 | *Ralf S. Engelschall* | |
18931 | ||
18932 | * Cleaned up the top-level documents; | |
18933 | o new files: CHANGES and LICENSE | |
18934 | o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay | |
18935 | o merged COPYRIGHT into LICENSE | |
18936 | o removed obsolete TODO file | |
18937 | o renamed MICROSOFT to INSTALL.W32 | |
18938 | ||
18939 | *Ralf S. Engelschall* | |
18940 | ||
18941 | * Removed dummy files from the 0.9.1b source tree: | |
18942 | ``` | |
18943 | crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi | |
18944 | crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f | |
18945 | crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f | |
18946 | crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f | |
18947 | util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f | |
18948 | ``` | |
18949 | ||
18950 | *Ralf S. Engelschall* | |
18951 | ||
18952 | * Added various platform portability fixes. | |
18953 | ||
18954 | *Mark J. Cox* | |
18955 | ||
18956 | * The Genesis of the OpenSSL rpject: | |
18957 | We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. | |
18958 | Young and Tim J. Hudson created while they were working for C2Net until | |
18959 | summer 1998. | |
18960 | ||
18961 | *The OpenSSL Project* | |
18962 | ||
18963 | ### Changes between 0.9.0b and 0.9.1b [not released] | |
18964 | ||
18965 | * Updated a few CA certificates under certs/ | |
18966 | ||
18967 | *Eric A. Young* | |
18968 | ||
18969 | * Changed some BIGNUM api stuff. | |
18970 | ||
18971 | *Eric A. Young* | |
18972 | ||
18973 | * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, | |
18974 | DGUX x86, Linux Alpha, etc. | |
18975 | ||
18976 | *Eric A. Young* | |
18977 | ||
18978 | * New COMP library [crypto/comp/] for SSL Record Layer Compression: | |
18979 | RLE (dummy implemented) and ZLIB (really implemented when ZLIB is | |
18980 | available). | |
18981 | ||
18982 | *Eric A. Young* | |
18983 | ||
18984 | * Add -strparse option to asn1pars program which parses nested | |
18985 | binary structures | |
18986 | ||
18987 | *Dr Stephen Henson <shenson@bigfoot.com>* | |
18988 | ||
18989 | * Added "oid_file" to ssleay.cnf for "ca" and "req" programs. | |
18990 | ||
18991 | *Eric A. Young* | |
18992 | ||
18993 | * DSA fix for "ca" program. | |
18994 | ||
18995 | *Eric A. Young* | |
18996 | ||
18997 | * Added "-genkey" option to "dsaparam" program. | |
18998 | ||
18999 | *Eric A. Young* | |
19000 | ||
19001 | * Added RIPE MD160 (rmd160) message digest. | |
19002 | ||
19003 | *Eric A. Young* | |
19004 | ||
19005 | * Added -a (all) option to "ssleay version" command. | |
19006 | ||
19007 | *Eric A. Young* | |
19008 | ||
19009 | * Added PLATFORM define which is the id given to Configure. | |
19010 | ||
19011 | *Eric A. Young* | |
19012 | ||
19013 | * Added MemCheck_XXXX functions to crypto/mem.c for memory checking. | |
19014 | ||
19015 | *Eric A. Young* | |
19016 | ||
19017 | * Extended the ASN.1 parser routines. | |
19018 | ||
19019 | *Eric A. Young* | |
19020 | ||
19021 | * Extended BIO routines to support REUSEADDR, seek, tell, etc. | |
19022 | ||
19023 | *Eric A. Young* | |
19024 | ||
19025 | * Added a BN_CTX to the BN library. | |
19026 | ||
19027 | *Eric A. Young* | |
19028 | ||
19029 | * Fixed the weak key values in DES library | |
19030 | ||
19031 | *Eric A. Young* | |
19032 | ||
19033 | * Changed API in EVP library for cipher aliases. | |
19034 | ||
19035 | *Eric A. Young* | |
19036 | ||
19037 | * Added support for RC2/64bit cipher. | |
19038 | ||
19039 | *Eric A. Young* | |
19040 | ||
19041 | * Converted the lhash library to the crypto/mem.c functions. | |
19042 | ||
19043 | *Eric A. Young* | |
19044 | ||
19045 | * Added more recognized ASN.1 object ids. | |
19046 | ||
19047 | *Eric A. Young* | |
19048 | ||
19049 | * Added more RSA padding checks for SSL/TLS. | |
19050 | ||
19051 | *Eric A. Young* | |
19052 | ||
19053 | * Added BIO proxy/filter functionality. | |
19054 | ||
19055 | *Eric A. Young* | |
19056 | ||
19057 | * Added extra_certs to SSL_CTX which can be used | |
19058 | send extra CA certificates to the client in the CA cert chain sending | |
19059 | process. It can be configured with SSL_CTX_add_extra_chain_cert(). | |
19060 | ||
19061 | *Eric A. Young* | |
19062 | ||
19063 | * Now Fortezza is denied in the authentication phase because | |
19064 | this is key exchange mechanism is not supported by SSLeay at all. | |
19065 | ||
19066 | *Eric A. Young* | |
19067 | ||
19068 | * Additional PKCS1 checks. | |
19069 | ||
19070 | *Eric A. Young* | |
19071 | ||
19072 | * Support the string "TLSv1" for all TLS v1 ciphers. | |
19073 | ||
19074 | *Eric A. Young* | |
19075 | ||
19076 | * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the | |
19077 | ex_data index of the SSL context in the X509_STORE_CTX ex_data. | |
19078 | ||
19079 | *Eric A. Young* | |
19080 | ||
19081 | * Fixed a few memory leaks. | |
19082 | ||
19083 | *Eric A. Young* | |
19084 | ||
19085 | * Fixed various code and comment typos. | |
19086 | ||
19087 | *Eric A. Young* | |
19088 | ||
19089 | * A minor bug in ssl/s3_clnt.c where there would always be 4 0 | |
19090 | bytes sent in the client random. | |
19091 | ||
19092 | *Edward Bishop <ebishop@spyglass.com>* | |
19093 | ||
19094 | <!-- Links --> | |
19095 | ||
19096 | [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971 | |
19097 | [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967 | |
19098 | [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563 | |
19099 | [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559 | |
19100 | [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552 | |
19101 | [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551 | |
19102 | [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549 | |
19103 | [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547 | |
19104 | [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543 | |
19105 | [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407 | |
19106 | [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739 | |
19107 | [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737 | |
19108 | [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735 | |
19109 | [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734 | |
19110 | [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733 | |
19111 | [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732 | |
19112 | [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738 | |
19113 | [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737 | |
19114 | [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736 | |
19115 | [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735 | |
19116 | [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733 | |
19117 | [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732 | |
19118 | [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731 | |
19119 | [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730 | |
19120 | [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055 | |
19121 | [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054 | |
19122 | [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053 | |
19123 | [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052 | |
19124 | [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309 | |
19125 | [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308 | |
19126 | [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307 | |
19127 | [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306 | |
19128 | [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305 | |
19129 | [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304 | |
19130 | [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303 | |
19131 | [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302 | |
19132 | [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183 | |
19133 | [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182 | |
19134 | [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181 | |
19135 | [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180 | |
19136 | [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179 | |
19137 | [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178 | |
19138 | [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177 | |
19139 | [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176 | |
19140 | [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109 | |
19141 | [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107 | |
19142 | [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106 | |
19143 | [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105 | |
19144 | [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800 | |
19145 | [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799 | |
19146 | [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798 | |
19147 | [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797 | |
19148 | [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705 | |
19149 | [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702 | |
19150 | [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701 | |
19151 | [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197 | |
19152 | [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196 | |
19153 | [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195 | |
19154 | [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194 | |
19155 | [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193 | |
19156 | [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793 | |
19157 | [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792 | |
19158 | [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791 | |
19159 | [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790 | |
19160 | [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789 | |
19161 | [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788 | |
19162 | [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787 | |
19163 | [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293 | |
19164 | [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291 | |
19165 | [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290 | |
19166 | [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289 | |
19167 | [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288 | |
19168 | [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287 | |
19169 | [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286 | |
19170 | [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285 | |
19171 | [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209 | |
19172 | [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208 | |
19173 | [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207 | |
19174 | [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206 | |
19175 | [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205 | |
19176 | [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204 | |
19177 | [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275 | |
19178 | [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139 | |
19179 | [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572 | |
19180 | [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571 | |
19181 | [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570 | |
19182 | [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569 | |
19183 | [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568 | |
19184 | [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567 | |
19185 | [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566 | |
19186 | [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513 | |
19187 | [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512 | |
19188 | [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511 | |
19189 | [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510 | |
19190 | [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509 | |
19191 | [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508 | |
19192 | [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507 | |
19193 | [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506 | |
19194 | [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505 | |
19195 | [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470 | |
19196 | [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224 | |
19197 | [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221 | |
19198 | [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195 | |
19199 | [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160 | |
19200 | [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076 | |
19201 | [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450 | |
19202 | [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353 | |
19203 | [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169 | |
19204 | [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166 | |
19205 | [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686 | |
19206 | [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333 | |
19207 | [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110 | |
19208 | [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884 | |
19209 | [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050 | |
19210 | [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027 | |
19211 | [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619 | |
19212 | [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577 | |
19213 | [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576 | |
19214 | [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109 | |
19215 | [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108 | |
19216 | [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210 | |
19217 | [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207 | |
19218 | [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014 | |
19219 | [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252 | |
19220 | [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180 | |
19221 | [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864 | |
19222 | [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633 | |
19223 | [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740 | |
19224 | [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433 | |
19225 | [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355 | |
19226 | [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555 | |
19227 | [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245 | |
19228 | [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386 | |
19229 | [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379 | |
19230 | [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378 | |
19231 | [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377 | |
19232 | [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789 | |
19233 | [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591 | |
19234 | [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590 | |
19235 | [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077 | |
19236 | [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678 | |
19237 | [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672 | |
19238 | [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891 | |
19239 | [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135 | |
19240 | [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995 | |
19241 | [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343 | |
19242 | [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339 | |
19243 | [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738 | |
19244 | [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940 | |
19245 | [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937 | |
19246 | [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969 | |
19247 | [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112 | |
19248 | [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079 | |
19249 | [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851 | |
19250 | [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545 | |
19251 | [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544 | |
19252 | [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543 | |
19253 | [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078 | |
19254 | [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659 | |
19255 | [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657 | |
19256 | [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656 | |
19257 | [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655 |