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.
8 [log]: https://github.com/openssl/openssl/commits/
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)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * Windows thread synchronization uses read/write primitives (SRWLock) when
27 supported by the OS, otherwise CriticalSection continues to be used.
31 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
32 work on read only BIO source/sinks that do not support these functions.
33 This allows piping or redirection of a file BIO using stdin to be buffered
34 into memory. This is used internally in OSSL_DECODER_from_bio().
38 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
39 this function would return one of the values OSSL_STORE_INFO_NAME,
40 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
41 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
42 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
43 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
44 using this function should be amended to handle the changed return value.
48 * The implementation of the EVP ciphers CAST5-ECB, CAST5-CBC, CAST5-OFB,
49 CAST5-CFB, BF-ECB, BF-CBC, BF-OFB, BF-CFB, IDEA-ECB, IDEC-CBC, IDEA-OFB,
50 IDEA-CFB, SEED-ECB, SEED-CBC, SEED-OFB, SEED-CFB, RC2-ECB, RC2-CBC,
51 RC2-40-CBC, RC2-64-CBC, RC2-OFB, RC2-CFB, RC4, RC4-40, RC4-HMAC-MD5, RC5-ECB,
52 RC5-CBC, RC5-OFB, RC5-CFB, DESX-CBC, DES-ECB, DES-CBC, DES-OFB, DES-CFB,
53 DES-CFB1 and DES-CFB8 have been moved to the legacy provider. Applications
54 using the EVP APIs to access these ciphers should instead use more modern
55 ciphers. If that is not possible then these applications should ensure that
56 the legacy provider has been loaded. This can be achieved either
57 programmatically or via configuration. See the provider(7) man page for
62 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
63 RIPEMD-160 have been moved to the legacy provider. Applications using the
64 EVP APIs to access these digests should instead use more modern digests. If
65 that is not possible then these applications should ensure that the legacy
66 provider has been loaded. This can be achieved either programmatically or via
67 configuration. See the provider(7) man page for further details.
71 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
76 * The deprecated functions EVP_PKEY_get0_RSA(),
77 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
78 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
79 well as the similarly named "get1" functions behave slightly differently in
80 OpenSSL 3.0. Previously they returned a pointer to the low-level key used
81 internally by libcrypto. From OpenSSL 3.0 this key may now be held in a
82 provider. Calling these functions will only return a handle on the internal
83 key where the EVP_PKEY was constructed using this key in the first place, for
84 example using a function or macro such as EVP_PKEY_assign_RSA(),
85 EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key,
86 then these functions now return a cached copy of the key. Changes to
87 the internal provider key that take place after the first time the cached key
88 is accessed will not be reflected back in the cached copy. Similarly any
89 changes made to the cached copy by application code will not be reflected
90 back in the internal provider key.
92 For the above reasons the keys returned from these functions should typically
93 be treated as read-only. To emphasise this the value returned from
94 EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(),
95 EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may
96 break some existing code. Applications broken by this change should be
97 modified. The preferred solution is to refactor the code to avoid the use of
98 these deprecated functions. Failing this the code should be modified to use a
99 const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(),
100 EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a
101 non-const pointer to enable them to be "freed". However they should also be
102 treated as read-only.
106 * A number of functions handling low level keys or engines were deprecated
107 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
108 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
109 EVP_PKEY_get0_siphash(). Applications using engines should instead use
110 providers. Applications getting or setting low level keys in an EVP_PKEY
111 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
112 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
116 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
117 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
118 and require returning octet ptr parameters from providers that
119 would like to support them which complicates provider implementations.
123 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
124 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
125 RAND_set_rand_method(). Provider based random number generators should
126 be used instead via EVP_RAND(3).
130 * The SRP APIs have been deprecated. The old APIs do not work via providers,
131 and there is no EVP interface to them. Unfortunately there is no replacement
132 for these APIs at this time.
136 * Add a compile time option to prevent the caching of provider fetched
137 algorithms. This is enabled by including the no-cached-fetch option
138 at configuration time.
142 * The default algorithms for pkcs12 creation with the PKCS12_create() function
143 were changed to more modern PBKDF2 and AES based algorithms. The default
144 MAC iteration count was changed to PKCS12_DEFAULT_ITER to make it equal
145 with the password-based encryption iteration count. The default digest
146 algorithm for the MAC computation was changed to SHA-256. The pkcs12
147 application now supports -legacy option that restores the previous
148 default algorithms to support interoperability with legacy systems.
150 *Tomáš Mráz and Sahana Prasad*
152 * The openssl speed command does not use low-level API calls anymore. This
153 implies some of the performance numbers might not be fully comparable
154 with the previous releases due to higher overhead. This applies
155 particularly to measuring performance on smaller data chunks.
159 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
160 Typically if OpenSSL has no EC or DH algorithms then it cannot support
161 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
162 through providers. Therefore third party providers may supply group
163 implementations even where there are no built-in ones. Attempting to create
164 TLS connections in such a build without also disabling TLSv1.3 at run time or
165 using third party provider groups may result in handshake failures. TLSv1.3
166 can be disabled at compile time using the "no-tls1_3" Configure option.
170 * The undocumented function X509_certificate_type() has been deprecated;
171 applications can use X509_get0_pubkey() and X509_get0_signature() to
172 get the same information.
176 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
177 functions. They are identical to BN_rand() and BN_rand_range()
182 * Removed RSA padding mode for SSLv23 (which was only used for
183 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
184 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
189 * Deprecated the obsolete X9.31 RSA key generation related functions
190 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
191 BN_X931_generate_prime_ex().
195 * The default key generation method for the regular 2-prime RSA keys was
196 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
197 Conditions Based on Auxiliary Probable Primes). This method is slower
198 than the original method.
202 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
203 They are replaced with the BN_check_prime() function that avoids possible
204 misuse and always uses at least 64 rounds of the Miller-Rabin
205 primality test. At least 64 rounds of the Miller-Rabin test are now also
206 used for all prime generation, including RSA key generation.
207 This increases key generation time, especially for larger keys.
211 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
212 as they are not useful with non-deprecated functions.
216 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
217 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
218 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
219 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
220 were used to collect all necessary data to form a HTTP request, and to
221 perform the HTTP transfer with that request. With OpenSSL 3.0, the
222 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
223 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
224 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
225 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
226 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
227 OSSL_HTTP_REQ_CTX_set_max_response_length().
229 *Rich Salz and Richard Levitte*
231 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
235 * Validation of SM2 keys has been separated from the validation of regular EC
236 keys, allowing to improve the SM2 validation process to reject loaded private
237 keys that are not conforming to the SM2 ISO standard.
238 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
243 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
244 switches: a validation failure triggers an early exit, returning a failure
245 exit status to the parent process.
249 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
250 to ignore unknown ciphers.
254 * The `-cipher-commands` and `-digest-commands` options
255 of the command line utility `list` have been deprecated.
256 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
260 * All of the low level EC_KEY functions have been deprecated including:
262 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
263 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
264 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
265 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
266 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
267 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
268 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
269 EC_KEY_METHOD_get_verify,
270 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
271 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
272 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
273 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
274 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
275 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
276 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
277 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
278 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
279 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
280 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
281 Applications that need to implement an EC_KEY_METHOD need to consider
282 implementation of the functionality in a special provider.
283 For replacement of the functions manipulating the EC_KEY objects
284 see the EVP_PKEY-EC(7) manual page.
286 Additionally functions that read and write EC_KEY objects such as
287 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
288 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
289 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
290 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
291 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
292 have also been deprecated. Applications should instead use the
293 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
295 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
296 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
297 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
298 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
299 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
301 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
303 * Deprecated all the libcrypto and libssl error string loading
304 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
305 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
306 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
307 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
308 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
309 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
310 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
311 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
312 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
313 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
315 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
316 now loads error strings automatically.
320 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
321 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
322 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
323 are to be used by servers requiring ephemeral DH keys. Instead applications
324 should consider using the built-in DH parameters that are available by
325 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
326 necessary then applications can use the alternative functions
327 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
328 replacement for the "callback" functions. The callback was originally useful
329 in order to have different parameters for export and non-export ciphersuites.
330 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
331 functions should be replaced by one of the other methods described above.
335 * The `-crypt` option to the `passwd` command line tool has been removed.
339 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
344 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
346 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
347 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
348 The inverse ciphers use AES decryption for wrapping, and
349 AES encryption for unwrapping.
353 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
354 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
355 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
356 OpenSSL 3.0 these are replaced by the more generic functions
357 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
358 The old versions have been converted to deprecated macros that just call the
363 * The security callback, which can be customised by application code, supports
364 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
365 in the "other" parameter. In most places this is what is passed. All these
366 places occur server side. However there was one client side call of this
367 security operation and it passed a DH object instead. This is incorrect
368 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
369 of the other locations. Therefore this client side call has been changed to
370 pass an EVP_PKEY instead.
374 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
375 interface. Their functionality remains unchanged.
379 * Added new option for 'openssl list', '-providers', which will display the
380 list of loaded providers, their names, version and status. It optionally
381 displays their gettable parameters.
385 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
386 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
387 type is internally recognised so the workaround is no longer needed.
389 Functionality is still retained as it is, but will only work with
390 EVP_PKEYs with a legacy internal key.
394 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
395 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
399 * Changed all "STACK" functions to be macros instead of inline functions. Macro
400 parameters are still checked for type safety at compile time via helper
405 * Remove the RAND_DRBG API
407 The RAND_DRBG API did not fit well into the new provider concept as
408 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
409 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
410 and some of its API calls are rather low-level. This holds in particular
411 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
413 Adding a compatibility layer to continue supporting the RAND_DRBG API as
414 a legacy API for a regular deprecation period turned out to come at the
415 price of complicating the new provider API unnecessarily. Since the
416 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
419 *Paul Dale and Matthias St. Pierre*
421 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
422 as well as actual hostnames.
426 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
427 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
428 conversely, silently ignore DTLS protocol version bounds when configuring
429 TLS-based contexts. The commands can be repeated to set bounds of both
430 types. The same applies with the corresponding "min_protocol" and
431 "max_protocol" command-line switches, in case some application uses both TLS
434 SSL_CTX instances that are created for a fixed protocol version (e.g.
435 `TLSv1_server_method()`) also silently ignore version bounds. Previously
436 attempts to apply bounds to these protocol versions would result in an
437 error. Now only the "version-flexible" SSL_CTX instances are subject to
438 limits in configuration files in command-line options.
442 * Deprecated the `ENGINE` API. Engines should be replaced with providers
447 * Reworked the recorded ERR codes to make better space for system errors.
448 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
449 given code is a system error (true) or an OpenSSL error (false).
453 * Reworked the test perl framework to better allow parallel testing.
455 *Nicola Tuveri and David von Oheimb*
457 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
458 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
462 * 'Configure' has been changed to figure out the configuration target if
463 none is given on the command line. Consequently, the 'config' script is
464 now only a mere wrapper. All documentation is changed to only mention
467 *Rich Salz and Richard Levitte*
469 * Added a library context `OSSL_LIB_CTX` that applications as well as
470 other libraries can use to form a separate context within which
471 libcrypto operations are performed.
473 There are two ways this can be used:
475 - Directly, by passing a library context to functions that take
476 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
478 - Indirectly, by creating a new library context and then assigning
479 it as the new default, with `OSSL_LIB_CTX_set0_default`.
481 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
482 apart from the functions directly related to `OSSL_LIB_CTX`, accept
483 NULL to indicate that the default library context should be used.
485 Library code that changes the default library context using
486 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
487 second call before returning to the caller.
489 _(Note: the library context was initially called `OPENSSL_CTX` and
490 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
494 * Handshake now fails if Extended Master Secret extension is dropped
499 * Dropped interactive mode from the `openssl` program. From now on,
500 running it without arguments is equivalent to `openssl help`.
504 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
505 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
506 While the old function names have been retained for backward compatibility
507 they should not be used in new developments
508 because their return values are confusing: Unlike other `_cmp()` functions
509 they do not return 0 in case their arguments are equal.
513 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
514 `EC_GROUP_get_field_type()`.
518 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
519 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
520 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
521 Applications should rely on the library automatically assigning a suitable
522 EC_METHOD internally upon EC_GROUP construction.
526 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
527 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
528 assigned internally without application intervention.
529 Users of EC_GROUP_new() should switch to a different suitable constructor.
533 * Add CAdES-BES signature verification support, mostly derived
534 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
536 *Filipe Raimundo da Silva*
538 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
542 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
543 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS). Its purpose
544 is to support encryption and decryption of a digital envelope that is both
545 authenticated and encrypted using AES GCM mode.
549 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
550 functions are not widely used and now OpenSSL automatically perform this
551 conversion when needed.
555 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
556 EC_KEY_precompute_mult(). These functions are not widely used and
557 applications should instead switch to named curves which OpenSSL has
558 hardcoded lookup tables for.
562 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
563 should instead use the L<EC_POINT_mul(3)> function.
567 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
568 that are not applicable to the new provider model. Applications should
569 instead use EVP_default_properties_is_fips_enabled() and
570 EVP_default_properties_enable_fips().
574 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
575 is set, an unexpected EOF is ignored, it pretends a close notify was received
576 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
580 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
581 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
582 used and applications should instead use the
583 L<EC_POINT_set_affine_coordinates(3)> and
584 L<EC_POINT_get_affine_coordinates(3)> functions.
588 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
589 arrays to be more easily constructed via a series of utility functions.
590 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
591 the various push functions and finally convert to a passable OSSL_PARAM
592 array using OSSL_PARAM_BLD_to_param().
596 * The security strength of SHA1 and MD5 based signatures in TLS has been
597 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
598 working at the default security level of 1 and instead requires security
599 level 0. The security level can be changed either using the cipher string
600 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`. This also means
601 that where the signature algorithms extension is missing from a ClientHello
602 then the handshake will fail in TLS 1.2 at security level 1. This is because,
603 although this extension is optional, failing to provide one means that
604 OpenSSL will fallback to a default set of signature algorithms. This default
605 set requires the availability of SHA1.
609 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
610 contain a provider side internal key.
614 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
615 They are old functions that we don't use, and that you could disable with
616 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
620 * Project text documents not yet having a proper file name extension
621 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
622 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
626 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
627 have been converted to Markdown with the goal to produce documents
628 which not only look pretty when viewed online in the browser, but
629 remain well readable inside a plain text editor.
631 To achieve this goal, a 'minimalistic' Markdown style has been applied
632 which avoids formatting elements that interfere too much with the
633 reading flow in the text file. For example, it
635 * avoids [ATX headings][] and uses [setext headings][] instead
636 (which works for `<h1>` and `<h2>` headings only).
637 * avoids [inline links][] and uses [reference links][] instead.
638 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
640 [ATX headings]: https://github.github.com/gfm/#atx-headings
641 [setext headings]: https://github.github.com/gfm/#setext-headings
642 [inline links]: https://github.github.com/gfm/#inline-link
643 [reference links]: https://github.github.com/gfm/#reference-link
644 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
645 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
647 *Matthias St. Pierre*
649 * The test suite is changed to preserve results of each test recipe.
650 A new directory test-runs/ with subdirectories named like the
651 test recipes are created in the build tree for this purpose.
655 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
656 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
657 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
659 *David von Oheimb, Martin Peylo*
661 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
662 The legacy OCSP-focused and only partly documented API is retained for
663 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
667 * Added `util/check-format.pl`, a tool for checking adherence to the
668 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
669 The checks performed are incomplete and yield some false positives.
670 Still the tool should be useful for detecting most typical glitches.
674 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
675 If domain name resolution yields multiple IP addresses all of them are tried
676 after `connect()` failures.
680 * All of the low level RSA functions have been deprecated including:
682 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
683 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
684 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
685 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
686 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
687 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
688 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
689 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
690 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
691 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
692 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
693 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
694 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
695 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
696 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
697 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
698 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
699 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
700 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
701 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
702 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
703 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
704 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
705 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
706 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
707 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
708 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
709 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
711 Use of these low level functions has been informally discouraged for a long
712 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
713 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
714 L<EVP_PKEY_decrypt(3)>.
718 * X509 certificates signed using SHA1 are no longer allowed at security
720 In TLS/SSL the default security level is 1. It can be set either
721 using the cipher string with `@SECLEVEL`, or calling
722 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
723 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
725 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
726 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
727 options of the commands.
731 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
732 modified to use PKEY APIs. These commands are now in maintenance mode
733 and no new features will be added to them.
737 * The command line utility rsautl has been deprecated.
738 Instead use the pkeyutl program.
742 * The command line utilities genrsa and rsa have been modified to use PKEY
743 APIs. They now write PKCS#8 keys by default. These commands are now in
744 maintenance mode and no new features will be added to them.
748 * All of the low level DH functions have been deprecated including:
750 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
751 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
752 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
753 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
754 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
755 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
756 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
757 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
758 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
759 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
760 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
761 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
762 DH_meth_get_generate_params and DH_meth_set_generate_params.
764 Use of these low level functions has been informally discouraged for a long
765 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
766 and L<EVP_PKEY_derive(3)>.
768 Additionally functions that read and write DH objects such as d2i_DHparams,
769 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
770 functions have also been deprecated. Applications should instead use the
771 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
773 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
774 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
775 `EVP_PKEY_set1_DH()` are also deprecated.
776 Applications should instead either read or write an
777 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
778 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
780 *Paul Dale and Matt Caswell*
782 * All of the low level DSA functions have been deprecated including:
784 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
785 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
786 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
787 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
788 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
789 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
790 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
791 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
792 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
793 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
794 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
795 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
796 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
797 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
798 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
799 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
800 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
802 Use of these low level functions has been informally discouraged for a long
803 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
804 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
806 Finaly functions that assign or obtain DSA objects from an EVP_PKEY such as
807 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
808 `EVP_PKEY_set1_DSA()` are also deprecated.
809 Applications should instead either read or write an
810 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs,
811 or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
815 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
816 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
817 This means that applications don't have to look at the curve NID and
818 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
819 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
820 a no-op when the EVP_PKEY is already of the given type.
822 Parameter and key generation is also reworked to make it possible
823 to generate EVP_PKEY_SM2 parameters and keys without having to go
824 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
825 However, code that does the latter will still work as before.
829 * Deprecated low level ECDH and ECDSA functions. These include:
831 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
832 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
835 Use of these low level functions has been informally discouraged for a long
836 time. Instead applications should use the EVP_PKEY_derive(3),
837 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
841 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
842 and EVP_PKEY_decrypt() instead.
843 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
844 and EVP_PKEY_encrypt() instead.
848 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
849 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
850 a new formulation to include all the things it can be used for,
851 as well as words of caution.
855 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
856 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
860 * All of the low level HMAC functions have been deprecated including:
862 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
863 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
866 Use of these low level functions has been informally discouraged for a long
867 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
868 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
869 and L<EVP_MAC_final(3)>.
873 * Over two thousand fixes were made to the documentation, including:
874 - Common options (such as -rand/-writerand, TLS version control, etc)
875 were refactored and point to newly-enhanced descriptions in openssl.pod.
876 - Added style conformance for all options (with help from Richard Levitte),
877 documented all reported missing options, added a CI build to check
878 that all options are documented and that no unimplemented options
880 - Documented some internals, such as all use of environment variables.
881 - Addressed all internal broken L<> references.
885 * All of the low level CMAC functions have been deprecated including:
887 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
888 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
890 Use of these low level functions has been informally discouraged for a long
891 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
892 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
893 and L<EVP_MAC_final(3)>.
897 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
898 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
901 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
902 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
903 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
904 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
905 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
906 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
907 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
908 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
909 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
910 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
912 Use of these low level functions has been informally discouraged
913 for a long time. Applications should use the EVP_DigestInit_ex(3),
914 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
918 * Corrected the documentation of the return values from the `EVP_DigestSign*`
919 set of functions. The documentation mentioned negative values for some
920 errors, but this was never the case, so the mention of negative values
923 Code that followed the documentation and thereby check with something
924 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
928 * All of the low level cipher functions have been deprecated including:
930 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
931 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
932 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
933 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
934 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
935 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
936 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
937 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
938 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
939 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
940 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
941 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
942 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
943 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
944 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
945 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
946 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
947 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
948 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
949 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
950 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
951 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
952 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
953 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
954 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
955 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
956 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
957 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
958 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
960 Use of these low level functions has been informally discouraged for
961 a long time. Applications should use the high level EVP APIs, e.g.
962 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
963 equivalently named decrypt functions instead.
965 *Matt Caswell and Paul Dale*
967 * Removed include/openssl/opensslconf.h.in and replaced it with
968 include/openssl/configuration.h.in, which differs in not including
969 <openssl/macros.h>. A short header include/openssl/opensslconf.h
970 was added to include both.
972 This allows internal hacks where one might need to modify the set
973 of configured macros, for example this if deprecated symbols are
974 still supposed to be available internally:
976 #include <openssl/configuration.h>
978 #undef OPENSSL_NO_DEPRECATED
979 #define OPENSSL_SUPPRESS_DEPRECATED
981 #include <openssl/macros.h>
983 This should not be used by applications that use the exported
984 symbols, as that will lead to linking errors.
988 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
989 used in exponentiation with 512-bit moduli. No EC algorithms are
990 affected. Analysis suggests that attacks against 2-prime RSA1024,
991 3-prime RSA1536, and DSA1024 as a result of this defect would be very
992 difficult to perform and are not believed likely. Attacks against DH512
993 are considered just feasible. However, for an attack the target would
994 have to re-use the DH512 private key, which is not recommended anyway.
995 Also applications directly using the low level API BN_mod_exp may be
996 affected if they use BN_FLG_CONSTTIME.
1001 * Most memory-debug features have been deprecated, and the functionality
1002 replaced with no-ops.
1006 * Added documentation for the STACK API.
1010 * Introduced a new method type and API, OSSL_ENCODER, to
1011 represent generic encoders. An implementation is expected to
1012 be able to encode an object associated with a given name (such
1013 as an algorithm name for an asymmetric key) into forms given by
1014 implementation properties.
1016 Encoders are primarily used from inside libcrypto, through
1017 calls to functions like EVP_PKEY_print_private(),
1018 PEM_write_bio_PrivateKey() and similar.
1020 Encoders are specified in such a way that they can be made to
1021 directly handle the provider side portion of an object, if this
1022 provider side part comes from the same provider as the encoder
1023 itself, but can also be made to handle objects in parametrized
1024 form (as an OSSL_PARAM array of data). This allows a provider to
1025 offer generic encoders as a service for any other provider.
1029 * Added a .pragma directive to the syntax of configuration files, to
1030 allow varying behavior in a supported and predictable manner.
1031 Currently added pragma:
1035 This allows dollar signs to be a keyword character unless it's
1036 followed by a opening brace or parenthesis. This is useful for
1037 platforms where dollar signs are commonly used in names, such as
1038 volume names and system directory names on VMS.
1042 * Added functionality to create an EVP_PKEY from user data. This
1043 is effectively the same as creating a RSA, DH or DSA object and
1044 then assigning them to an EVP_PKEY, but directly using algorithm
1045 agnostic EVP functions. A benefit is that this should be future
1046 proof for public key algorithms to come.
1050 * Change the interpretation of the '--api' configuration option to
1051 mean that this is a desired API compatibility level with no
1052 further meaning. The previous interpretation, that this would
1053 also mean to remove all deprecated symbols up to and including
1054 the given version, no requires that 'no-deprecated' is also used
1055 in the configuration.
1057 When building applications, the desired API compatibility level
1058 can be set with the OPENSSL_API_COMPAT macro like before. For
1059 API compatibility version below 3.0, the old style numerical
1060 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1061 For version 3.0 and on, the value is expected to be the decimal
1062 value calculated from the major and minor version like this:
1064 MAJOR * 10000 + MINOR * 100
1068 -DOPENSSL_API_COMPAT=30000 For 3.0
1069 -DOPENSSL_API_COMPAT=30200 For 3.2
1071 To hide declarations that are deprecated up to and including the
1072 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1073 given when building the application as well.
1077 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1078 access to certificate and CRL stores via URIs and OSSL_STORE
1081 This adds the following functions:
1083 - X509_LOOKUP_store()
1084 - X509_STORE_load_file()
1085 - X509_STORE_load_path()
1086 - X509_STORE_load_store()
1087 - SSL_add_store_cert_subjects_to_stack()
1088 - SSL_CTX_set_default_verify_store()
1089 - SSL_CTX_load_verify_file()
1090 - SSL_CTX_load_verify_dir()
1091 - SSL_CTX_load_verify_store()
1095 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1096 The presence of this system service is determined at run-time.
1100 * Added functionality to create an EVP_PKEY context based on data
1101 for methods from providers. This takes an algorithm name and a
1102 property query string and simply stores them, with the intent
1103 that any operation that uses this context will use those strings
1104 to fetch the needed methods implicitly, thereby making the port
1105 of application written for pre-3.0 OpenSSL easier.
1109 * The undocumented function NCONF_WIN32() has been deprecated; for
1110 conversion details see the HISTORY section of doc/man5/config.pod
1114 * Introduced the new functions EVP_DigestSignInit_ex() and
1115 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1116 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1117 pages for further details.
1121 * Over two thousand fixes were made to the documentation, including:
1122 adding missing command flags, better style conformance, documentation
1125 *Rich Salz, Richard Levitte*
1127 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1128 X25519, X448, Ed25519 and Ed448.
1132 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1137 * Deprecated the public definition of `ERR_STATE` as well as the function
1138 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1143 * Added ERR functionality to give callers access to the stored function
1144 names that have replaced the older function code based functions.
1146 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1147 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1148 ERR_peek_error_all() and ERR_peek_last_error_all().
1150 These functions have become deprecated: ERR_get_error_line(),
1151 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1152 ERR_peek_last_error_line_data() and ERR_func_error_string().
1154 Users are recommended to use ERR_get_error_all(), or to pick information
1155 with ERR_peek functions and finish off with getting the error code by using
1160 * Extended testing to be verbose for failing tests only. The make variables
1161 VERBOSE_FAILURE or VF can be used to enable this:
1163 $ make VF=1 test # Unix
1164 $ mms /macro=(VF=1) test ! OpenVMS
1165 $ nmake VF=1 test # Windows
1169 * Added the `-copy_extensions` option to the `x509` command for use with
1170 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1171 all extensions in the request are copied to the certificate or vice versa.
1173 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1175 * Added the `-copy_extensions` option to the `req` command for use with
1176 `-x509`. When given with the `copy` or `copyall` argument,
1177 all extensions in the certification request are copied to the certificate.
1181 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1182 they generate are by default RFC 5280 compliant in the following sense:
1183 There is a subjectKeyIdentifier extension with a hash value of the public key
1184 and for not self-signed certs there is an authorityKeyIdentifier extension
1185 with a keyIdentifier field or issuer information identifying the signing key.
1186 This is done unless some configuration overrides the new default behavior,
1187 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1191 * Added several checks to `X509_verify_cert()` according to requirements in
1192 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1193 (which may be done by using the CLI option `-x509_strict`):
1194 * The basicConstraints of CA certificates must be marked critical.
1195 * CA certificates must explicitly include the keyUsage extension.
1196 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1197 * The issuer name of any certificate must not be empty.
1198 * The subject name of CA certs, certs with keyUsage crlSign,
1199 and certs without subjectAlternativeName must not be empty.
1200 * If a subjectAlternativeName extension is given it must not be empty.
1201 * The signatureAlgorithm field and the cert signature must be consistent.
1202 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1203 must not be marked critical.
1204 * The authorityKeyIdentifier must be given for X.509v3 certs
1205 unless they are self-signed.
1206 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1210 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1211 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1215 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1216 used even when parsing explicit parameters, when loading a encoded key
1217 or calling `EC_GROUP_new_from_ecpkparameters()`/
1218 `EC_GROUP_new_from_ecparameters()`.
1219 This prevents bypass of security hardening and performance gains,
1220 especially for curves with specialized EC_METHODs.
1221 By default, if a key encoded with explicit parameters is loaded and later
1222 encoded, the output is still encoded with explicit parameters, even if
1223 internally a "named" EC_GROUP is used for computation.
1227 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1228 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1229 NULL. After this change, only the cofactor parameter can be NULL. It also
1230 does some minimal sanity checks on the passed order.
1235 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1236 An attack is simple, if the first CMS_recipientInfo is valid but the
1237 second CMS_recipientInfo is chosen ciphertext. If the second
1238 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1239 encryption key will be replaced by garbage, and the message cannot be
1240 decoded, but if the RSA decryption fails, the correct encryption key is
1241 used and the recipient will not notice the attack.
1242 As a work around for this potential attack the length of the decrypted
1243 key must be equal to the cipher default key length, in case the
1244 certifiate is not given and all recipientInfo are tried out.
1245 The old behaviour can be re-enabled in the CMS code by setting the
1246 CMS_DEBUG_DECRYPT flag.
1250 * Early start up entropy quality from the DEVRANDOM seed source has been
1251 improved for older Linux systems. The RAND subsystem will wait for
1252 /dev/random to be producing output before seeding from /dev/urandom.
1253 The seeded state is stored for future library initialisations using
1254 a system global shared memory segment. The shared memory identifier
1255 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1256 the desired value. The default identifier is 114.
1260 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1261 when primes for RSA keys are computed.
1262 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1263 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1264 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1265 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1266 This avoids possible fingerprinting of newly generated RSA modules.
1270 * Correct the extended master secret constant on EBCDIC systems. Without this
1271 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1272 negotiate EMS will fail. Unfortunately this also means that TLS connections
1273 between EBCDIC systems with this fix, and EBCDIC systems without this
1274 fix will fail if they negotiate EMS.
1278 * Changed the library initialisation so that the config file is now loaded
1279 by default. This was already the case for libssl. It now occurs for both
1280 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1281 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1285 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1286 where the former acts as a replacement for `ERR_put_error()`, and the
1287 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1288 `ERR_raise_data()` adds more flexibility by taking a format string and
1289 an arbitrary number of arguments following it, to be processed with
1294 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1295 to check if a named provider is loaded and available. When called, it
1296 will also activate all fallback providers if such are still present.
1300 * Enforce a minimum DH modulus size of 512 bits.
1304 * Changed DH parameters to generate the order q subgroup instead of 2q.
1305 Previously generated DH parameters are still accepted by DH_check
1306 but DH_generate_key works around that by clearing bit 0 of the
1307 private key for those. This avoids leaking bit 0 of the private key.
1311 * Significantly reduce secure memory usage by the randomness pools.
1315 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1320 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1321 algorithms. An implementation of a key exchange algorithm can be obtained
1322 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1323 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1324 the older EVP_PKEY_derive_init() function. See the man pages for the new
1325 functions for further details.
1329 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1333 * Removed the function names from error messages and deprecated the
1336 * Removed NextStep support and the macro OPENSSL_UNISTD
1340 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1341 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1342 Also removed "export var as function" capability; we do not export
1343 variables, only functions.
1347 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1348 an error and 1 indicating success. In previous versions of OpenSSL this
1349 was a void type. If a key was set longer than the maximum possible this
1354 * Support SM2 signing and verification schemes with X509 certificate.
1358 * Use SHA256 as the default digest for TS query in the `ts` app.
1362 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1363 This checks that the salt length is at least 128 bits, the derived key
1364 length is at least 112 bits, and that the iteration count is at least 1000.
1365 For backwards compatibility these checks are disabled by default in the
1366 default provider, but are enabled by default in the fips provider.
1367 To enable or disable these checks use the control
1368 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1372 * Default cipher lists/suites are now available via a function, the
1373 #defines are deprecated.
1377 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1378 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1379 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1383 * Join the directories crypto/x509 and crypto/x509v3
1387 * Added command 'openssl kdf' that uses the EVP_KDF API.
1391 * Added command 'openssl mac' that uses the EVP_MAC API.
1395 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1396 as default directories. Also added the command 'openssl info'
1397 for scripting purposes.
1401 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1402 deprecated. These undocumented functions were never integrated into the EVP
1403 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1404 Bi-directional IGE mode. These modes were never formally standardised and
1405 usage of these functions is believed to be very small. In particular
1406 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1407 is ever used. The security implications are believed to be minimal, but
1408 this issue was never fixed for backwards compatibility reasons. New code
1409 should not use these modes.
1413 * Add prediction resistance to the DRBG reseeding process.
1417 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1418 mandated by IEEE Std 1619-2018.
1422 * Added newline escaping functionality to a filename when using openssl dgst.
1423 This output format is to replicate the output format found in the `*sum`
1424 checksum programs. This aims to preserve backward compatibility.
1426 *Matt Eaton, Richard Levitte, and Paul Dale*
1428 * Removed the heartbeat message in DTLS feature, as it has very
1429 little usage and doesn't seem to fulfill a valuable purpose.
1430 The configuration option is now deprecated.
1434 * Changed the output of 'openssl {digestname} < file' to display the
1435 digest name in its output.
1439 * Added a new generic trace API which provides support for enabling
1440 instrumentation through trace output. This feature is mainly intended
1441 as an aid for developers and is disabled by default. To utilize it,
1442 OpenSSL needs to be configured with the `enable-trace` option.
1444 If the tracing API is enabled, the application can activate trace output
1445 by registering BIOs as trace channels for a number of tracing and debugging
1448 The `openssl` program has been expanded to enable any of the types
1449 available via environment variables defined by the user, and serves as
1450 one possible example on how to use this functionality.
1452 *Richard Levitte & Matthias St. Pierre*
1454 * Added build tests for C++. These are generated files that only do one
1455 thing, to include one public OpenSSL head file each. This tests that
1456 the public header files can be usefully included in a C++ application.
1458 This test isn't enabled by default. It can be enabled with the option
1459 'enable-buildtest-c++'.
1463 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1467 * Add KMAC to EVP_MAC.
1471 * Added property based algorithm implementation selection framework to
1476 * Added SCA hardening for modular field inversion in EC_GROUP through
1477 a new dedicated field_inv() pointer in EC_METHOD.
1478 This also addresses a leakage affecting conversions from projective
1479 to affine coordinates.
1481 *Billy Bob Brumley, Nicola Tuveri*
1483 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1484 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1485 those algorithms that were already supported through the EVP_PKEY API
1486 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1487 and scrypt are now wrappers that call EVP_KDF.
1491 * Build devcrypto engine as a dynamic engine.
1493 *Eneas U de Queiroz*
1495 * Add keyed BLAKE2 to EVP_MAC.
1499 * Fix a bug in the computation of the endpoint-pair shared secret used
1500 by DTLS over SCTP. This breaks interoperability with older versions
1501 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1502 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1503 interoperability with such broken implementations. However, enabling
1504 this switch breaks interoperability with correct implementations.
1506 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1507 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1511 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1515 * Changed the license to the Apache License v2.0.
1519 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1521 - Major releases (indicated by incrementing the MAJOR release number)
1522 may introduce incompatible API/ABI changes.
1523 - Minor releases (indicated by incrementing the MINOR release number)
1524 may introduce new features but retain API/ABI compatibility.
1525 - Patch releases (indicated by incrementing the PATCH number)
1526 are intended for bug fixes and other improvements of existing
1527 features only (like improving performance or adding documentation)
1528 and retain API/ABI compatibility.
1532 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1536 * Remove the 'dist' target and add a tarball building script. The
1537 'dist' target has fallen out of use, and it shouldn't be
1538 necessary to configure just to create a source distribution.
1542 * Recreate the OS390-Unix config target. It no longer relies on a
1543 special script like it did for OpenSSL pre-1.1.0.
1547 * Instead of having the source directories listed in Configure, add
1548 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1553 * Add GMAC to EVP_MAC.
1557 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1561 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1562 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1563 to facilitate the continued use of MACs through raw private keys in
1564 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1568 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1569 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1573 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1574 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1575 are retained for backwards compatibility.
1579 * AES-XTS mode now enforces that its two keys are different to mitigate
1580 the attacked described in "Efficient Instantiations of Tweakable
1581 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1582 Details of this attack can be obtained from:
1583 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1587 * Rename the object files, i.e. give them other names than in previous
1588 versions. Their names now include the name of the final product, as
1589 well as its type mnemonic (bin, lib, shlib).
1593 * Added new option for 'openssl list', '-objects', which will display the
1594 list of built in objects, i.e. OIDs with names.
1598 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1599 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1604 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1605 improves application performance by removing data copies and providing
1606 applications with zero-copy system calls such as sendfile and splice.
1610 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1611 option is set, openssl cleanses (zeroize) plaintext bytes from
1612 internal buffers after delivering them to the application. Note,
1613 the application is still responsible for cleansing other copies
1614 (e.g.: data received by SSL_read(3)).
1618 * `PKCS12_parse` now maintains the order of the parsed certificates
1619 when outputting them via `*ca` (rather than reversing it).
1623 * Deprecated pthread fork support methods. These were unused so no
1624 replacement is required.
1626 - OPENSSL_fork_prepare()
1627 - OPENSSL_fork_parent()
1628 - OPENSSL_fork_child()
1635 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1637 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1639 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1640 create a unique hash value based on the issuer and serial number data
1641 contained within an X509 certificate. However it was failing to correctly
1642 handle any errors that may occur while parsing the issuer field (which might
1643 occur if the issuer field is maliciously constructed). This may subsequently
1644 result in a NULL pointer deref and a crash leading to a potential denial of
1650 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1651 padding mode to correctly check for rollback attacks. This is considered a
1652 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1657 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1658 functions. Previously they could overflow the output length argument in some
1659 cases where the input length is close to the maximum permissable length for
1660 an integer on the platform. In such cases the return value from the function
1661 call would be 1 (indicating success), but the output length value would be
1662 negative. This could cause applications to behave incorrectly or crash.
1667 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1668 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1669 could be exploited in a side channel attack to recover the password. Since
1670 the attack is local host only this is outside of the current OpenSSL
1671 threat model and therefore no CVE is assigned.
1673 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1678 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1680 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1681 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1682 If an attacker can control both items being compared then this could lead
1683 to a possible denial of service attack. OpenSSL itself uses the
1684 GENERAL_NAME_cmp function for two purposes:
1685 1) Comparing CRL distribution point names between an available CRL and a
1686 CRL distribution point embedded in an X509 certificate
1687 2) When verifying that a timestamp response token signer matches the
1688 timestamp authority name (exposed via the API functions
1689 TS_RESP_verify_response and TS_RESP_verify_token)
1694 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1696 * Certificates with explicit curve parameters are now disallowed in
1697 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1701 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1702 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1703 conversely, silently ignore DTLS protocol version bounds when configuring
1704 TLS-based contexts. The commands can be repeated to set bounds of both
1705 types. The same applies with the corresponding "min_protocol" and
1706 "max_protocol" command-line switches, in case some application uses both TLS
1709 SSL_CTX instances that are created for a fixed protocol version (e.g.
1710 TLSv1_server_method()) also silently ignore version bounds. Previously
1711 attempts to apply bounds to these protocol versions would result in an
1712 error. Now only the "version-flexible" SSL_CTX instances are subject to
1713 limits in configuration files in command-line options.
1717 * Handshake now fails if Extended Master Secret extension is dropped
1722 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1724 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1726 * Fixed segmentation fault in SSL_check_chain()
1727 Server or client applications that call the SSL_check_chain() function
1728 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1729 dereference as a result of incorrect handling of the
1730 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1731 or unrecognised signature algorithm is received from the peer. This could
1732 be exploited by a malicious peer in a Denial of Service attack.
1737 * Added AES consttime code for no-asm configurations
1738 an optional constant time support for AES was added
1739 when building openssl for no-asm.
1740 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1741 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1742 At this time this feature is by default disabled.
1743 It will be enabled by default in 3.0.
1747 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1749 * Revert the change of EOF detection while reading in libssl to avoid
1750 regressions in applications depending on the current way of reporting
1751 the EOF. As the existing method is not fully accurate the change to
1752 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1753 branch and will be present in the 3.0 release.
1757 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1758 when primes for RSA keys are computed.
1759 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1760 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1761 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1762 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1763 This avoids possible fingerprinting of newly generated RSA modules.
1767 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1769 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1770 while reading in libssl then we would report an error back to the
1771 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1772 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1773 therefore give a hint as to what went wrong.
1777 * Check that ed25519 and ed448 are allowed by the security level. Previously
1778 signature algorithms not using an MD were not being checked that they were
1779 allowed by the security level.
1783 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1784 was not quite right. The behaviour was not consistent between resumption
1785 and normal handshakes, and also not quite consistent with historical
1786 behaviour. The behaviour in various scenarios has been clarified and
1787 it has been updated to make it match historical behaviour as closely as
1792 * *[VMS only]* The header files that the VMS compilers include automatically,
1793 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1794 that the C++ compiler doesn't understand. This is a shortcoming in the
1795 compiler, but can be worked around with `__cplusplus` guards.
1797 C++ applications that use OpenSSL libraries must be compiled using the
1798 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1799 functions. Otherwise, only functions with symbols of less than 31
1800 characters can be used, as the linker will not be able to successfully
1801 resolve symbols with longer names.
1805 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1806 The presence of this system service is determined at run-time.
1810 * Added newline escaping functionality to a filename when using openssl dgst.
1811 This output format is to replicate the output format found in the `*sum`
1812 checksum programs. This aims to preserve backward compatibility.
1814 *Matt Eaton, Richard Levitte, and Paul Dale*
1816 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1821 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1823 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1824 number generator (RNG). This was intended to include protection in the
1825 event of a fork() system call in order to ensure that the parent and child
1826 processes did not share the same RNG state. However this protection was not
1827 being used in the default case.
1829 A partial mitigation for this issue is that the output from a high
1830 precision timer is mixed into the RNG state so the likelihood of a parent
1831 and child process sharing state is significantly reduced.
1833 If an application already calls OPENSSL_init_crypto() explicitly using
1834 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1837 *Matthias St. Pierre*
1839 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1840 used even when parsing explicit parameters, when loading a encoded key
1841 or calling `EC_GROUP_new_from_ecpkparameters()`/
1842 `EC_GROUP_new_from_ecparameters()`.
1843 This prevents bypass of security hardening and performance gains,
1844 especially for curves with specialized EC_METHODs.
1845 By default, if a key encoded with explicit parameters is loaded and later
1846 encoded, the output is still encoded with explicit parameters, even if
1847 internally a "named" EC_GROUP is used for computation.
1851 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1852 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1853 NULL. After this change, only the cofactor parameter can be NULL. It also
1854 does some minimal sanity checks on the passed order.
1859 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1860 An attack is simple, if the first CMS_recipientInfo is valid but the
1861 second CMS_recipientInfo is chosen ciphertext. If the second
1862 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1863 encryption key will be replaced by garbage, and the message cannot be
1864 decoded, but if the RSA decryption fails, the correct encryption key is
1865 used and the recipient will not notice the attack.
1866 As a work around for this potential attack the length of the decrypted
1867 key must be equal to the cipher default key length, in case the
1868 certifiate is not given and all recipientInfo are tried out.
1869 The old behaviour can be re-enabled in the CMS code by setting the
1870 CMS_DEBUG_DECRYPT flag.
1875 * Early start up entropy quality from the DEVRANDOM seed source has been
1876 improved for older Linux systems. The RAND subsystem will wait for
1877 /dev/random to be producing output before seeding from /dev/urandom.
1878 The seeded state is stored for future library initialisations using
1879 a system global shared memory segment. The shared memory identifier
1880 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1881 the desired value. The default identifier is 114.
1885 * Correct the extended master secret constant on EBCDIC systems. Without this
1886 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1887 negotiate EMS will fail. Unfortunately this also means that TLS connections
1888 between EBCDIC systems with this fix, and EBCDIC systems without this
1889 fix will fail if they negotiate EMS.
1893 * Use Windows installation paths in the mingw builds
1895 Mingw isn't a POSIX environment per se, which means that Windows
1896 paths should be used for installation.
1901 * Changed DH_check to accept parameters with order q and 2q subgroups.
1902 With order 2q subgroups the bit 0 of the private key is not secret
1903 but DH_generate_key works around that by clearing bit 0 of the
1904 private key for those. This avoids leaking bit 0 of the private key.
1908 * Significantly reduce secure memory usage by the randomness pools.
1912 * Revert the DEVRANDOM_WAIT feature for Linux systems
1914 The DEVRANDOM_WAIT feature added a select() call to wait for the
1915 /dev/random device to become readable before reading from the
1916 /dev/urandom device.
1918 It turned out that this change had negative side effects on
1919 performance which were not acceptable. After some discussion it
1920 was decided to revert this feature and leave it up to the OS
1921 resp. the platform maintainer to ensure a proper initialization
1922 during early boot time.
1924 *Matthias St. Pierre*
1926 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1928 * Add build tests for C++. These are generated files that only do one
1929 thing, to include one public OpenSSL head file each. This tests that
1930 the public header files can be usefully included in a C++ application.
1932 This test isn't enabled by default. It can be enabled with the option
1933 'enable-buildtest-c++'.
1937 * Enable SHA3 pre-hashing for ECDSA and DSA.
1941 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1942 This changes the size when using the `genpkey` command when no size is given.
1943 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1944 generation commands to use 2048 bits by default.
1948 * Reorganize the manual pages to consistently have RETURN VALUES,
1949 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1950 util/fix-doc-nits accordingly.
1952 *Paul Yang, Joshua Lock*
1954 * Add the missing accessor EVP_PKEY_get0_engine()
1958 * Have commands like `s_client` and `s_server` output the signature scheme
1959 along with other cipher suite parameters when debugging.
1963 * Make OPENSSL_config() error agnostic again.
1967 * Do the error handling in RSA decryption constant time.
1971 * Prevent over long nonces in ChaCha20-Poly1305.
1973 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1974 for every encryption operation. RFC 7539 specifies that the nonce value
1975 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1976 and front pads the nonce with 0 bytes if it is less than 12
1977 bytes. However it also incorrectly allows a nonce to be set of up to 16
1978 bytes. In this case only the last 12 bytes are significant and any
1979 additional leading bytes are ignored.
1981 It is a requirement of using this cipher that nonce values are
1982 unique. Messages encrypted using a reused nonce value are susceptible to
1983 serious confidentiality and integrity attacks. If an application changes
1984 the default nonce length to be longer than 12 bytes and then makes a
1985 change to the leading bytes of the nonce expecting the new value to be a
1986 new unique nonce then such an application could inadvertently encrypt
1987 messages with a reused nonce.
1989 Additionally the ignored bytes in a long nonce are not covered by the
1990 integrity guarantee of this cipher. Any application that relies on the
1991 integrity of these ignored leading bytes of a long nonce may be further
1992 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1993 is safe because no such use sets such a long nonce value. However user
1994 applications that use this cipher directly and set a non-default nonce
1995 length to be longer than 12 bytes may be vulnerable.
1997 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2003 * Add DEVRANDOM_WAIT feature for Linux systems
2005 On older Linux systems where the getrandom() system call is not available,
2006 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2007 Contrary to getrandom(), the /dev/urandom device will not block during
2008 early boot when the kernel CSPRNG has not been seeded yet.
2010 To mitigate this known weakness, use select() to wait for /dev/random to
2011 become readable before reading from /dev/urandom.
2013 * Ensure that SM2 only uses SM3 as digest algorithm
2017 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2019 * Change the info callback signals for the start and end of a post-handshake
2020 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2021 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2022 confused by this and assume that a TLSv1.2 renegotiation has started. This
2023 can break KeyUpdate handling. Instead we no longer signal the start and end
2024 of a post handshake message exchange (although the messages themselves are
2025 still signalled). This could break some applications that were expecting
2026 the old signals. However without this KeyUpdate is not usable for many
2031 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2033 * Timing vulnerability in DSA signature generation
2035 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2036 timing side channel attack. An attacker could use variations in the signing
2037 algorithm to recover the private key.
2039 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2044 * Timing vulnerability in ECDSA signature generation
2046 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2047 timing side channel attack. An attacker could use variations in the signing
2048 algorithm to recover the private key.
2050 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2055 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2056 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2057 of two gigabytes and the error handling improved.
2059 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2060 categorized as a normal bug, not a security issue, because the DRBG reseeds
2061 automatically and is fully functional even without additional randomness
2062 provided by the application.
2064 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2066 * Add a new ClientHello callback. Provides a callback interface that gives
2067 the application the ability to adjust the nascent SSL object at the
2068 earliest stage of ClientHello processing, immediately after extensions have
2069 been collected but before they have been processed. In particular, this
2070 callback can adjust the supported TLS versions in response to the contents
2075 * Add SM2 base algorithm support.
2079 * s390x assembly pack: add (improved) hardware-support for the following
2080 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2081 aes-cfb/cfb8, aes-ecb.
2085 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2086 parameter is no longer accepted, as it leads to a corrupt table. NULL
2087 pem_str is reserved for alias entries only.
2091 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2092 step for prime curves. The new implementation is based on formulae from
2093 differential addition-and-doubling in homogeneous projective coordinates
2094 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2095 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2096 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2097 to work in projective coordinates.
2099 *Billy Bob Brumley, Nicola Tuveri*
2101 * Change generating and checking of primes so that the error rate of not
2102 being prime depends on the intended use based on the size of the input.
2103 For larger primes this will result in more rounds of Miller-Rabin.
2104 The maximal error rate for primes with more than 1080 bits is lowered
2107 *Kurt Roeckx, Annie Yousar*
2109 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2113 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2114 moving between systems, and to avoid confusion when a Windows build is
2115 done with mingw vs with MSVC. For POSIX installs, there's still a
2116 symlink or copy named 'tsget' to avoid that confusion as well.
2120 * Revert blinding in ECDSA sign and instead make problematic addition
2121 length-invariant. Switch even to fixed-length Montgomery multiplication.
2125 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2126 step for binary curves. The new implementation is based on formulae from
2127 differential addition-and-doubling in mixed Lopez-Dahab projective
2128 coordinates, modified to independently blind the operands.
2130 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2132 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2133 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2134 EC_METHODs to implement their own specialized "ladder step", to take
2135 advantage of more favorable coordinate systems or more efficient
2136 differential addition-and-doubling algorithms.
2138 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2140 * Modified the random device based seed sources to keep the relevant
2141 file descriptors open rather than reopening them on each access.
2142 This allows such sources to operate in a chroot() jail without
2143 the associated device nodes being available. This behaviour can be
2144 controlled using RAND_keep_random_devices_open().
2148 * Numerous side-channel attack mitigations have been applied. This may have
2149 performance impacts for some algorithms for the benefit of improved
2150 security. Specific changes are noted in this change log by their respective
2155 * AIX shared library support overhaul. Switch to AIX "natural" way of
2156 handling shared libraries, which means collecting shared objects of
2157 different versions and bitnesses in one common archive. This allows to
2158 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2159 doesn't affect the way 3rd party applications are linked, only how
2160 multi-version installation is managed.
2164 * Make ec_group_do_inverse_ord() more robust and available to other
2165 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2166 mitigations are applied to the fallback BN_mod_inverse().
2167 When using this function rather than BN_mod_inverse() directly, new
2168 EC cryptosystem implementations are then safer-by-default.
2172 * Add coordinate blinding for EC_POINT and implement projective
2173 coordinate blinding for generic prime curves as a countermeasure to
2174 chosen point SCA attacks.
2176 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2178 * Add blinding to ECDSA and DSA signatures to protect against side channel
2179 attacks discovered by Keegan Ryan (NCC Group).
2183 * Enforce checking in the `pkeyutl` command to ensure that the input
2184 length does not exceed the maximum supported digest length when performing
2185 a sign, verify or verifyrecover operation.
2189 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2190 I/O in combination with something like select() or poll() will hang. This
2191 can be turned off again using SSL_CTX_clear_mode().
2192 Many applications do not properly handle non-application data records, and
2193 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2194 around the problems in those applications, but can also break some.
2195 It's recommended to read the manpages about SSL_read(), SSL_write(),
2196 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2197 SSL_CTX_set_read_ahead() again.
2201 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2202 now allow empty (zero character) pass phrases.
2206 * Apply blinding to binary field modular inversion and remove patent
2207 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2211 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2212 binary and prime elliptic curves.
2216 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2217 constant time fixed point multiplication.
2221 * Revise elliptic curve scalar multiplication with timing attack
2222 defenses: ec_wNAF_mul redirects to a constant time implementation
2223 when computing fixed point and variable point multiplication (which
2224 in OpenSSL are mostly used with secret scalars in keygen, sign,
2225 ECDH derive operations).
2226 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2229 * Updated CONTRIBUTING
2233 * Updated DRBG / RAND to request nonce and additional low entropy
2234 randomness from the system.
2236 *Matthias St. Pierre*
2238 * Updated 'openssl rehash' to use OpenSSL consistent default.
2242 * Moved the load of the ssl_conf module to libcrypto, which helps
2243 loading engines that libssl uses before libssl is initialised.
2247 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2251 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2253 *Ingo Schwarze, Rich Salz*
2255 * Added output of accepting IP address and port for 'openssl s_server'
2259 * Added a new API for TLSv1.3 ciphersuites:
2260 SSL_CTX_set_ciphersuites()
2261 SSL_set_ciphersuites()
2265 * Memory allocation failures consistently add an error to the error
2270 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2271 in libcrypto when run as setuid/setgid.
2275 * Load any config file by default when libssl is used.
2279 * Added new public header file <openssl/rand_drbg.h> and documentation
2280 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2282 *Matthias St. Pierre*
2284 * QNX support removed (cannot find contributors to get their approval
2285 for the license change).
2289 * TLSv1.3 replay protection for early data has been implemented. See the
2290 SSL_read_early_data() man page for further details.
2294 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2295 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2296 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2297 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2298 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2299 configuration has been separated out. See the ciphers man page or the
2300 SSL_CTX_set_ciphersuites() man page for more information.
2304 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2305 in responder mode now supports the new "-multi" option, which
2306 spawns the specified number of child processes to handle OCSP
2307 requests. The "-timeout" option now also limits the OCSP
2308 responder's patience to wait to receive the full client request
2309 on a newly accepted connection. Child processes are respawned
2310 as needed, and the CA index file is automatically reloaded
2311 when changed. This makes it possible to run the "ocsp" responder
2312 as a long-running service, making the OpenSSL CA somewhat more
2313 feature-complete. In this mode, most diagnostic messages logged
2314 after entering the event loop are logged via syslog(3) rather than
2319 * Added support for X448 and Ed448. Heavily based on original work by
2324 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2325 objects loaded. This adds the functions OSSL_STORE_expect() and
2326 OSSL_STORE_find() as well as needed tools to construct searches and
2327 get the search data out of them.
2331 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2332 version of OpenSSL should review their configuration settings to ensure
2333 that they are still appropriate for TLSv1.3. For further information see:
2334 <https://wiki.openssl.org/index.php/TLS1.3>
2338 * Grand redesign of the OpenSSL random generator
2340 The default RAND method now utilizes an AES-CTR DRBG according to
2341 NIST standard SP 800-90Ar1. The new random generator is essentially
2342 a port of the default random generator from the OpenSSL FIPS 2.0
2343 object module. It is a hybrid deterministic random bit generator
2344 using an AES-CTR bit stream and which seeds and reseeds itself
2345 automatically using trusted system entropy sources.
2347 Some of its new features are:
2348 - Support for multiple DRBG instances with seed chaining.
2349 - The default RAND method makes use of a DRBG.
2350 - There is a public and private DRBG instance.
2351 - The DRBG instances are fork-safe.
2352 - Keep all global DRBG instances on the secure heap if it is enabled.
2353 - The public and private DRBG instance are per thread for lock free
2356 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2358 * Changed Configure so it only says what it does and doesn't dump
2359 so much data. Instead, ./configdata.pm should be used as a script
2360 to display all sorts of configuration data.
2364 * Added processing of "make variables" to Configure.
2368 * Added SHA512/224 and SHA512/256 algorithm support.
2372 * The last traces of Netware support, first removed in 1.1.0, have
2377 * Get rid of Makefile.shared, and in the process, make the processing
2378 of certain files (rc.obj, or the .def/.map/.opt files produced from
2379 the ordinal files) more visible and hopefully easier to trace and
2380 debug (or make silent).
2384 * Make it possible to have environment variable assignments as
2385 arguments to config / Configure.
2389 * Add multi-prime RSA (RFC 8017) support.
2393 * Add SM3 implemented according to GB/T 32905-2016
2394 *Jack Lloyd <jack.lloyd@ribose.com>,*
2395 *Ronald Tse <ronald.tse@ribose.com>,*
2396 *Erick Borsboom <erick.borsboom@ribose.com>*
2398 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2399 as documented in RFC6066.
2400 Based on a patch from Tomasz Moń
2402 *Filipe Raimundo da Silva*
2404 * Add SM4 implemented according to GB/T 32907-2016.
2405 *Jack Lloyd <jack.lloyd@ribose.com>,*
2406 *Ronald Tse <ronald.tse@ribose.com>,*
2407 *Erick Borsboom <erick.borsboom@ribose.com>*
2409 * Reimplement -newreq-nodes and ERR_error_string_n; the
2410 original author does not agree with the license change.
2414 * Add ARIA AEAD TLS support.
2418 * Some macro definitions to support VS6 have been removed. Visual
2419 Studio 6 has not worked since 1.1.0
2423 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2424 without clearing the errors.
2428 * Add "atfork" functions. If building on a system that without
2429 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2430 requirements. The RAND facility now uses/requires this.
2438 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2439 not possible to disable entirely. However, it's still possible to
2440 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2443 To disable, configure with 'no-ui-console'. 'no-ui' is still
2444 possible to use as an alias. Check at compile time with the
2445 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2446 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2450 * Add a STORE module, which implements a uniform and URI based reader of
2451 stores that can contain keys, certificates, CRLs and numerous other
2452 objects. The main API is loosely based on a few stdio functions,
2453 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2454 OSSL_STORE_error and OSSL_STORE_close.
2455 The implementation uses backends called "loaders" to implement arbitrary
2456 URI schemes. There is one built in "loader" for the 'file' scheme.
2460 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2461 then adjusted to work on FreeBSD 8.4 as well.
2462 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2463 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2467 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2468 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2469 error code calls like this:
2471 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2473 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2474 that can be encoded in C. For the foreseeable future, this will only
2477 *Richard Levitte and Tim Hudson*
2479 * Removed BSD cryptodev engine.
2483 * Add a build target 'build_all_generated', to build all generated files
2484 and only that. This can be used to prepare everything that requires
2485 things like perl for a system that lacks perl and then move everything
2486 to that system and do the rest of the build there.
2490 * In the UI interface, make it possible to duplicate the user data. This
2491 can be used by engines that need to retain the data for a longer time
2492 than just the call where this user data is passed.
2496 * Ignore the '-named_curve auto' value for compatibility of applications
2499 *Tomáš Mráz <tmraz@fedoraproject.org>*
2501 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2502 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2503 alerts across multiple records (some of which could be empty). In practice
2504 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2505 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2506 support this at all. Supporting it adds significant complexity to the
2507 record layer, and its removal is unlikely to cause interoperability
2512 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2513 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2514 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2519 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2520 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2522 *Richard Levitte, Andy Polyakov*
2524 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2529 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2530 platform rather than 'mingw'.
2534 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2535 success if they are asked to add an object which already exists
2536 in the store. This change cascades to other functions which load
2537 certificates and CRLs.
2541 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2542 facilitate stack unwinding even from assembly subroutines.
2546 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2547 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2551 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2552 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2553 which is the minimum version we support.
2557 * Certificate time validation (X509_cmp_time) enforces stricter
2558 compliance with RFC 5280. Fractional seconds and timezone offsets
2559 are no longer allowed.
2563 * Add support for ARIA
2567 * s_client will now send the Server Name Indication (SNI) extension by
2568 default unless the new "-noservername" option is used. The server name is
2569 based on the host provided to the "-connect" option unless overridden by
2570 using "-servername".
2574 * Add support for SipHash
2578 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2579 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2580 prevent issues where no progress is being made and the peer continually
2581 sends unrecognised record types, using up resources processing them.
2585 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2586 using the algorithm defined in
2587 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2591 * Heartbeat support has been removed; the ABI is changed for now.
2593 *Richard Levitte, Rich Salz*
2595 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2599 * The RSA "null" method, which was partially supported to avoid patent
2600 issues, has been replaced to always returns NULL.
2607 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2609 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2610 used even when parsing explicit parameters, when loading a encoded key
2611 or calling `EC_GROUP_new_from_ecpkparameters()`/
2612 `EC_GROUP_new_from_ecparameters()`.
2613 This prevents bypass of security hardening and performance gains,
2614 especially for curves with specialized EC_METHODs.
2615 By default, if a key encoded with explicit parameters is loaded and later
2616 encoded, the output is still encoded with explicit parameters, even if
2617 internally a "named" EC_GROUP is used for computation.
2621 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2622 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2623 NULL. After this change, only the cofactor parameter can be NULL. It also
2624 does some minimal sanity checks on the passed order.
2629 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2630 An attack is simple, if the first CMS_recipientInfo is valid but the
2631 second CMS_recipientInfo is chosen ciphertext. If the second
2632 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2633 encryption key will be replaced by garbage, and the message cannot be
2634 decoded, but if the RSA decryption fails, the correct encryption key is
2635 used and the recipient will not notice the attack.
2636 As a work around for this potential attack the length of the decrypted
2637 key must be equal to the cipher default key length, in case the
2638 certifiate is not given and all recipientInfo are tried out.
2639 The old behaviour can be re-enabled in the CMS code by setting the
2640 CMS_DEBUG_DECRYPT flag.
2645 * Use Windows installation paths in the mingw builds
2647 Mingw isn't a POSIX environment per se, which means that Windows
2648 paths should be used for installation.
2653 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2655 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2656 This changes the size when using the `genpkey` command when no size is given.
2657 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2658 generation commands to use 2048 bits by default.
2662 * Prevent over long nonces in ChaCha20-Poly1305.
2664 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2665 for every encryption operation. RFC 7539 specifies that the nonce value
2666 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2667 and front pads the nonce with 0 bytes if it is less than 12
2668 bytes. However it also incorrectly allows a nonce to be set of up to 16
2669 bytes. In this case only the last 12 bytes are significant and any
2670 additional leading bytes are ignored.
2672 It is a requirement of using this cipher that nonce values are
2673 unique. Messages encrypted using a reused nonce value are susceptible to
2674 serious confidentiality and integrity attacks. If an application changes
2675 the default nonce length to be longer than 12 bytes and then makes a
2676 change to the leading bytes of the nonce expecting the new value to be a
2677 new unique nonce then such an application could inadvertently encrypt
2678 messages with a reused nonce.
2680 Additionally the ignored bytes in a long nonce are not covered by the
2681 integrity guarantee of this cipher. Any application that relies on the
2682 integrity of these ignored leading bytes of a long nonce may be further
2683 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2684 is safe because no such use sets such a long nonce value. However user
2685 applications that use this cipher directly and set a non-default nonce
2686 length to be longer than 12 bytes may be vulnerable.
2688 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2694 * Added SCA hardening for modular field inversion in EC_GROUP through
2695 a new dedicated field_inv() pointer in EC_METHOD.
2696 This also addresses a leakage affecting conversions from projective
2697 to affine coordinates.
2699 *Billy Bob Brumley, Nicola Tuveri*
2701 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2702 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2706 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2710 * Remove the 'dist' target and add a tarball building script. The
2711 'dist' target has fallen out of use, and it shouldn't be
2712 necessary to configure just to create a source distribution.
2716 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2718 * Timing vulnerability in DSA signature generation
2720 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2721 timing side channel attack. An attacker could use variations in the signing
2722 algorithm to recover the private key.
2724 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2729 * Timing vulnerability in ECDSA signature generation
2731 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2732 timing side channel attack. An attacker could use variations in the signing
2733 algorithm to recover the private key.
2735 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2740 * Add coordinate blinding for EC_POINT and implement projective
2741 coordinate blinding for generic prime curves as a countermeasure to
2742 chosen point SCA attacks.
2744 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2746 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2748 * Client DoS due to large DH parameter
2750 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2751 malicious server can send a very large prime value to the client. This will
2752 cause the client to spend an unreasonably long period of time generating a
2753 key for this prime resulting in a hang until the client has finished. This
2754 could be exploited in a Denial Of Service attack.
2756 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2761 * Cache timing vulnerability in RSA Key Generation
2763 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2764 a cache timing side channel attack. An attacker with sufficient access to
2765 mount cache timing attacks during the RSA key generation process could
2766 recover the private key.
2768 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2769 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2774 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2775 parameter is no longer accepted, as it leads to a corrupt table. NULL
2776 pem_str is reserved for alias entries only.
2780 * Revert blinding in ECDSA sign and instead make problematic addition
2781 length-invariant. Switch even to fixed-length Montgomery multiplication.
2785 * Change generating and checking of primes so that the error rate of not
2786 being prime depends on the intended use based on the size of the input.
2787 For larger primes this will result in more rounds of Miller-Rabin.
2788 The maximal error rate for primes with more than 1080 bits is lowered
2791 *Kurt Roeckx, Annie Yousar*
2793 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2797 * Add blinding to ECDSA and DSA signatures to protect against side channel
2798 attacks discovered by Keegan Ryan (NCC Group).
2802 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2803 now allow empty (zero character) pass phrases.
2807 * Certificate time validation (X509_cmp_time) enforces stricter
2808 compliance with RFC 5280. Fractional seconds and timezone offsets
2809 are no longer allowed.
2813 * Fixed a text canonicalisation bug in CMS
2815 Where a CMS detached signature is used with text content the text goes
2816 through a canonicalisation process first prior to signing or verifying a
2817 signature. This process strips trailing space at the end of lines, converts
2818 line terminators to CRLF and removes additional trailing line terminators
2819 at the end of a file. A bug in the canonicalisation process meant that
2820 some characters, such as form-feed, were incorrectly treated as whitespace
2821 and removed. This is contrary to the specification (RFC5485). This fix
2822 could mean that detached text data signed with an earlier version of
2823 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2824 signed with a fixed OpenSSL may fail to verify with an earlier version of
2825 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2826 and use the "-binary" flag (for the "cms" command line application) or set
2827 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2831 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2833 * Constructed ASN.1 types with a recursive definition could exceed the stack
2835 Constructed ASN.1 types with a recursive definition (such as can be found
2836 in PKCS7) could eventually exceed the stack given malicious input with
2837 excessive recursion. This could result in a Denial Of Service attack. There
2838 are no such structures used within SSL/TLS that come from untrusted sources
2839 so this is considered safe.
2841 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2847 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2849 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2850 effectively reduced to only comparing the least significant bit of each
2851 byte. This allows an attacker to forge messages that would be considered as
2852 authenticated in an amount of tries lower than that guaranteed by the
2853 security claims of the scheme. The module can only be compiled by the
2854 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2856 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2862 * Add a build target 'build_all_generated', to build all generated files
2863 and only that. This can be used to prepare everything that requires
2864 things like perl for a system that lacks perl and then move everything
2865 to that system and do the rest of the build there.
2869 * Backport SSL_OP_NO_RENGOTIATION
2871 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2872 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2873 changes this is no longer possible in 1.1.0. Therefore the new
2874 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2875 1.1.0 to provide equivalent functionality.
2877 Note that if an application built against 1.1.0h headers (or above) is run
2878 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2879 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2883 * Removed the OS390-Unix config target. It relied on a script that doesn't
2888 * rsaz_1024_mul_avx2 overflow bug on x86_64
2890 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2891 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2892 Analysis suggests that attacks against RSA and DSA as a result of this
2893 defect would be very difficult to perform and are not believed likely.
2894 Attacks against DH1024 are considered just feasible, because most of the
2895 work necessary to deduce information about a private key may be performed
2896 offline. The amount of resources required for such an attack would be
2897 significant. However, for an attack on TLS to be meaningful, the server
2898 would have to share the DH1024 private key among multiple clients, which is
2899 no longer an option since CVE-2016-0701.
2901 This only affects processors that support the AVX2 but not ADX extensions
2902 like Intel Haswell (4th generation).
2904 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2905 was originally found via the OSS-Fuzz project.
2910 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2912 * bn_sqrx8x_internal carry bug on x86_64
2914 There is a carry propagating bug in the x86_64 Montgomery squaring
2915 procedure. No EC algorithms are affected. Analysis suggests that attacks
2916 against RSA and DSA as a result of this defect would be very difficult to
2917 perform and are not believed likely. Attacks against DH are considered just
2918 feasible (although very difficult) because most of the work necessary to
2919 deduce information about a private key may be performed offline. The amount
2920 of resources required for such an attack would be very significant and
2921 likely only accessible to a limited number of attackers. An attacker would
2922 additionally need online access to an unpatched system using the target
2923 private key in a scenario with persistent DH parameters and a private
2924 key that is shared between multiple clients.
2926 This only affects processors that support the BMI1, BMI2 and ADX extensions
2927 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2929 This issue was reported to OpenSSL by the OSS-Fuzz project.
2934 * Malformed X.509 IPAddressFamily could cause OOB read
2936 If an X.509 certificate has a malformed IPAddressFamily extension,
2937 OpenSSL could do a one-byte buffer overread. The most likely result
2938 would be an erroneous display of the certificate in text format.
2940 This issue was reported to OpenSSL by the OSS-Fuzz project.
2945 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2947 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2948 platform rather than 'mingw'.
2952 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2953 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2954 which is the minimum version we support.
2958 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2960 * Encrypt-Then-Mac renegotiation crash
2962 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2963 negotiated where it was not in the original handshake (or vice-versa) then
2964 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2965 and servers are affected.
2967 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2972 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2974 * Truncated packet could crash via OOB read
2976 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2977 cipher is being used, then a truncated packet can cause that host to
2978 perform an out-of-bounds read, usually resulting in a crash.
2980 This issue was reported to OpenSSL by Robert Święcki of Google.
2985 * Bad (EC)DHE parameters cause a client crash
2987 If a malicious server supplies bad parameters for a DHE or ECDHE key
2988 exchange then this can result in the client attempting to dereference a
2989 NULL pointer leading to a client crash. This could be exploited in a Denial
2992 This issue was reported to OpenSSL by Guido Vranken.
2997 * BN_mod_exp may produce incorrect results on x86_64
2999 There is a carry propagating bug in the x86_64 Montgomery squaring
3000 procedure. No EC algorithms are affected. Analysis suggests that attacks
3001 against RSA and DSA as a result of this defect would be very difficult to
3002 perform and are not believed likely. Attacks against DH are considered just
3003 feasible (although very difficult) because most of the work necessary to
3004 deduce information about a private key may be performed offline. The amount
3005 of resources required for such an attack would be very significant and
3006 likely only accessible to a limited number of attackers. An attacker would
3007 additionally need online access to an unpatched system using the target
3008 private key in a scenario with persistent DH parameters and a private
3009 key that is shared between multiple clients. For example this can occur by
3010 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3011 similar to CVE-2015-3193 but must be treated as a separate problem.
3013 This issue was reported to OpenSSL by the OSS-Fuzz project.
3018 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3020 * ChaCha20/Poly1305 heap-buffer-overflow
3022 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3023 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3024 crash. This issue is not considered to be exploitable beyond a DoS.
3026 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3031 * CMS Null dereference
3033 Applications parsing invalid CMS structures can crash with a NULL pointer
3034 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3035 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3036 structure callback if an attempt is made to free certain invalid encodings.
3037 Only CHOICE structures using a callback which do not handle NULL value are
3040 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3045 * Montgomery multiplication may produce incorrect results
3047 There is a carry propagating bug in the Broadwell-specific Montgomery
3048 multiplication procedure that handles input lengths divisible by, but
3049 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3050 and DH private keys are impossible. This is because the subroutine in
3051 question is not used in operations with the private key itself and an input
3052 of the attacker's direct choice. Otherwise the bug can manifest itself as
3053 transient authentication and key negotiation failures or reproducible
3054 erroneous outcome of public-key operations with specially crafted input.
3055 Among EC algorithms only Brainpool P-512 curves are affected and one
3056 presumably can attack ECDH key negotiation. Impact was not analyzed in
3057 detail, because pre-requisites for attack are considered unlikely. Namely
3058 multiple clients have to choose the curve in question and the server has to
3059 share the private key among them, neither of which is default behaviour.
3060 Even then only clients that chose the curve will be affected.
3062 This issue was publicly reported as transient failures and was not
3063 initially recognized as a security issue. Thanks to Richard Morgan for
3064 providing reproducible case.
3069 * Removed automatic addition of RPATH in shared libraries and executables,
3070 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3074 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3076 * Fix Use After Free for large message sizes
3078 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3079 message larger than approx 16k is received then the underlying buffer to
3080 store the incoming message is reallocated and moved. Unfortunately a
3081 dangling pointer to the old location is left which results in an attempt to
3082 write to the previously freed location. This is likely to result in a
3083 crash, however it could potentially lead to execution of arbitrary code.
3085 This issue only affects OpenSSL 1.1.0a.
3087 This issue was reported to OpenSSL by Robert Święcki.
3092 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3094 * OCSP Status Request extension unbounded memory growth
3096 A malicious client can send an excessively large OCSP Status Request
3097 extension. If that client continually requests renegotiation, sending a
3098 large OCSP Status Request extension each time, then there will be unbounded
3099 memory growth on the server. This will eventually lead to a Denial Of
3100 Service attack through memory exhaustion. Servers with a default
3101 configuration are vulnerable even if they do not support OCSP. Builds using
3102 the "no-ocsp" build time option are not affected.
3104 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3109 * SSL_peek() hang on empty record
3111 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3112 sends an empty record. This could be exploited by a malicious peer in a
3113 Denial Of Service attack.
3115 This issue was reported to OpenSSL by Alex Gaynor.
3120 * Excessive allocation of memory in tls_get_message_header() and
3121 dtls1_preprocess_fragment()
3123 A (D)TLS message includes 3 bytes for its length in the header for the
3124 message. This would allow for messages up to 16Mb in length. Messages of
3125 this length are excessive and OpenSSL includes a check to ensure that a
3126 peer is sending reasonably sized messages in order to avoid too much memory
3127 being consumed to service a connection. A flaw in the logic of version
3128 1.1.0 means that memory for the message is allocated too early, prior to
3129 the excessive message length check. Due to way memory is allocated in
3130 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3131 to service a connection. This could lead to a Denial of Service through
3132 memory exhaustion. However, the excessive message length check still takes
3133 place, and this would cause the connection to immediately fail. Assuming
3134 that the application calls SSL_free() on the failed connection in a timely
3135 manner then the 21Mb of allocated memory will then be immediately freed
3136 again. Therefore the excessive memory allocation will be transitory in
3137 nature. This then means that there is only a security impact if:
3139 1) The application does not call SSL_free() in a timely manner in the event
3140 that the connection fails
3142 2) The application is working in a constrained environment where there is
3143 very little free memory
3145 3) The attacker initiates multiple connection attempts such that there are
3146 multiple connections in a state where memory has been allocated for the
3147 connection; SSL_free() has not yet been called; and there is insufficient
3148 memory to service the multiple requests.
3150 Except in the instance of (1) above any Denial Of Service is likely to be
3151 transitory because as soon as the connection fails the memory is
3152 subsequently freed again in the SSL_free() call. However there is an
3153 increased risk during this period of application crashes due to the lack of
3154 memory - which would then mean a more serious Denial of Service.
3156 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3157 (CVE-2016-6307 and CVE-2016-6308)
3161 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3162 had to be removed. Primary reason is that vendor assembler can't
3163 assemble our modules with -KPIC flag. As result it, assembly
3164 support, was not even available as option. But its lack means
3165 lack of side-channel resistant code, which is incompatible with
3166 security by todays standards. Fortunately gcc is readily available
3167 prepackaged option, which we firmly point at...
3171 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3173 * Windows command-line tool supports UTF-8 opt-in option for arguments
3174 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3175 (to any value) allows Windows user to access PKCS#12 file generated
3176 with Windows CryptoAPI and protected with non-ASCII password, as well
3177 as files generated under UTF-8 locale on Linux also protected with
3182 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3183 have been disabled by default and removed from DEFAULT, just like RC4.
3184 See the RC4 item below to re-enable both.
3188 * The method for finding the storage location for the Windows RAND seed file
3189 has changed. First we check %RANDFILE%. If that is not set then we check
3190 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3191 all else fails we fall back to C:\.
3195 * The EVP_EncryptUpdate() function has had its return type changed from void
3196 to int. A return of 0 indicates and error while a return of 1 indicates
3201 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3202 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3203 off the constant time implementation for RSA, DSA and DH have been made
3204 no-ops and deprecated.
3208 * Windows RAND implementation was simplified to only get entropy by
3209 calling CryptGenRandom(). Various other RAND-related tickets
3212 *Joseph Wylie Yandle, Rich Salz*
3214 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3215 and `OPENSSL_LH_`, respectively. The old names are available
3216 with API compatibility. They new names are now completely documented.
3220 * Unify TYPE_up_ref(obj) methods signature.
3221 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3222 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3223 int (instead of void) like all others TYPE_up_ref() methods.
3224 So now these methods also check the return value of CRYPTO_atomic_add(),
3225 and the validity of object reference counter.
3227 *fdasilvayy@gmail.com*
3229 * With Windows Visual Studio builds, the .pdb files are installed
3230 alongside the installed libraries and executables. For a static
3231 library installation, ossl_static.pdb is the associate compiler
3232 generated .pdb file to be used when linking programs.
3236 * Remove openssl.spec. Packaging files belong with the packagers.
3240 * Automatic Darwin/OSX configuration has had a refresh, it will now
3241 recognise x86_64 architectures automatically. You can still decide
3242 to build for a different bitness with the environment variable
3243 KERNEL_BITS (can be 32 or 64), for example:
3245 KERNEL_BITS=32 ./config
3249 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3250 256 bit AES and HMAC with SHA256.
3254 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3258 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3262 * To enable users to have their own config files and build file templates,
3263 Configure looks in the directory indicated by the environment variable
3264 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3265 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3266 name and is used as is.
3270 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3271 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3272 X509_CERT_FILE_CTX was removed.
3276 * "shared" builds are now the default. To create only static libraries use
3277 the "no-shared" Configure option.
3281 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3282 All of these option have not worked for some while and are fundamental
3287 * Make various cleanup routines no-ops and mark them as deprecated. Most
3288 global cleanup functions are no longer required because they are handled
3289 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3290 Explicitly de-initing can cause problems (e.g. where a library that uses
3291 OpenSSL de-inits, but an application is still using it). The affected
3292 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3293 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3294 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3295 COMP_zlib_cleanup().
3299 * --strict-warnings no longer enables runtime debugging options
3300 such as REF_DEBUG. Instead, debug options are automatically
3301 enabled with '--debug' builds.
3303 *Andy Polyakov, Emilia Käsper*
3305 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3306 have been moved out of the public header files. New functions for managing
3307 these have been added.
3311 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3312 objects have been moved out of the public header files. New
3313 functions for managing these have been added.
3317 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3318 have been moved out of the public header files. New functions for managing
3319 these have been added.
3323 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3324 moved out of the public header files. New functions for managing these
3329 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3333 * Removed the mk1mf build scripts.
3337 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3338 it is always safe to #include a header now.
3342 * Removed the aged BC-32 config and all its supporting scripts
3346 * Removed support for Ultrix, Netware, and OS/2.
3350 * Add support for HKDF.
3352 *Alessandro Ghedini*
3354 * Add support for blake2b and blake2s
3358 * Added support for "pipelining". Ciphers that have the
3359 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3360 encryptions/decryptions simultaneously. There are currently no built-in
3361 ciphers with this property but the expectation is that engines will be able
3362 to offer it to significantly improve throughput. Support has been extended
3363 into libssl so that multiple records for a single connection can be
3364 processed in one go (for >=TLS 1.1).
3368 * Added the AFALG engine. This is an async capable engine which is able to
3369 offload work to the Linux kernel. In this initial version it only supports
3370 AES128-CBC. The kernel must be version 4.1.0 or greater.
3374 * OpenSSL now uses a new threading API. It is no longer necessary to
3375 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3376 are two supported threading models: pthreads and windows threads. It is
3377 also possible to configure OpenSSL at compile time for "no-threads". The
3378 old threading API should no longer be used. The functions have been
3379 replaced with "no-op" compatibility macros.
3381 *Alessandro Ghedini, Matt Caswell*
3383 * Modify behavior of ALPN to invoke callback after SNI/servername
3384 callback, such that updates to the SSL_CTX affect ALPN.
3388 * Add SSL_CIPHER queries for authentication and key-exchange.
3392 * Changes to the DEFAULT cipherlist:
3393 - Prefer (EC)DHE handshakes over plain RSA.
3394 - Prefer AEAD ciphers over legacy ciphers.
3395 - Prefer ECDSA over RSA when both certificates are available.
3396 - Prefer TLSv1.2 ciphers/PRF.
3397 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3402 * Change the ECC default curve list to be this, in order: x25519,
3403 secp256r1, secp521r1, secp384r1.
3407 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3408 disabled by default. They can be re-enabled using the
3409 enable-weak-ssl-ciphers option to Configure.
3413 * If the server has ALPN configured, but supports no protocols that the
3414 client advertises, send a fatal "no_application_protocol" alert.
3415 This behaviour is SHALL in RFC 7301, though it isn't universally
3416 implemented by other servers.
3420 * Add X25519 support.
3421 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3422 for public and private key encoding using the format documented in
3423 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3424 key generation and key derivation.
3426 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3431 * Deprecate SRP_VBASE_get_by_user.
3432 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3433 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3434 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3435 seed, even if the seed is configured.
3437 Users should use SRP_VBASE_get1_by_user instead. Note that in
3438 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3439 also that even though configuring the SRP seed attempts to hide
3440 invalid usernames by continuing the handshake with fake
3441 credentials, this behaviour is not constant time and no strong
3442 guarantees are made that the handshake is indistinguishable from
3443 that of a valid user.
3447 * Configuration change; it's now possible to build dynamic engines
3448 without having to build shared libraries and vice versa. This
3449 only applies to the engines in `engines/`, those in `crypto/engine/`
3450 will always be built into libcrypto (i.e. "static").
3452 Building dynamic engines is enabled by default; to disable, use
3453 the configuration option "disable-dynamic-engine".
3455 The only requirements for building dynamic engines are the
3456 presence of the DSO module and building with position independent
3457 code, so they will also automatically be disabled if configuring
3458 with "disable-dso" or "disable-pic".
3460 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3461 are also taken away from openssl/opensslconf.h, as they are
3466 * Configuration change; if there is a known flag to compile
3467 position independent code, it will always be applied on the
3468 libcrypto and libssl object files, and never on the application
3469 object files. This means other libraries that use routines from
3470 libcrypto / libssl can be made into shared libraries regardless
3471 of how OpenSSL was configured.
3473 If this isn't desirable, the configuration options "disable-pic"
3474 or "no-pic" can be used to disable the use of PIC. This will
3475 also disable building shared libraries and dynamic engines.
3479 * Removed JPAKE code. It was experimental and has no wide use.
3483 * The INSTALL_PREFIX Makefile variable has been renamed to
3484 DESTDIR. That makes for less confusion on what this variable
3485 is for. Also, the configuration option --install_prefix is
3490 * Heartbeat for TLS has been removed and is disabled by default
3491 for DTLS; configure with enable-heartbeats. Code that uses the
3492 old #define's might need to be updated.
3494 *Emilia Käsper, Rich Salz*
3496 * Rename REF_CHECK to REF_DEBUG.
3500 * New "unified" build system
3502 The "unified" build system is aimed to be a common system for all
3503 platforms we support. With it comes new support for VMS.
3505 This system builds supports building in a different directory tree
3506 than the source tree. It produces one Makefile (for unix family
3507 or lookalikes), or one descrip.mms (for VMS).
3509 The source of information to make the Makefile / descrip.mms is
3510 small files called 'build.info', holding the necessary
3511 information for each directory with source to compile, and a
3512 template in Configurations, like unix-Makefile.tmpl or
3515 With this change, the library names were also renamed on Windows
3516 and on VMS. They now have names that are closer to the standard
3517 on Unix, and include the major version number, and in certain
3518 cases, the architecture they are built for. See "Notes on shared
3519 libraries" in INSTALL.
3521 We rely heavily on the perl module Text::Template.
3525 * Added support for auto-initialisation and de-initialisation of the library.
3526 OpenSSL no longer requires explicit init or deinit routines to be called,
3527 except in certain circumstances. See the OPENSSL_init_crypto() and
3528 OPENSSL_init_ssl() man pages for further information.
3532 * The arguments to the DTLSv1_listen function have changed. Specifically the
3533 "peer" argument is now expected to be a BIO_ADDR object.
3535 * Rewrite of BIO networking library. The BIO library lacked consistent
3536 support of IPv6, and adding it required some more extensive
3537 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3538 which hold all types of addresses and chains of address information.
3539 It also introduces a new API, with functions like BIO_socket,
3540 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3541 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3542 have been adapted accordingly.
3546 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3551 * CRIME protection: disable compression by default, even if OpenSSL is
3552 compiled with zlib enabled. Applications can still enable compression
3553 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3554 using the SSL_CONF library to configure compression.
3558 * The signature of the session callback configured with
3559 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3560 was explicitly marked as `const unsigned char*` instead of
3565 * Always DPURIFY. Remove the use of uninitialized memory in the
3566 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3570 * Removed many obsolete configuration items, including
3571 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3572 MD2_CHAR, MD2_INT, MD2_LONG
3574 IDEA_SHORT, IDEA_LONG
3575 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3577 *Rich Salz, with advice from Andy Polyakov*
3579 * Many BN internals have been moved to an internal header file.
3581 *Rich Salz with help from Andy Polyakov*
3583 * Configuration and writing out the results from it has changed.
3584 Files such as Makefile include/openssl/opensslconf.h and are now
3585 produced through general templates, such as Makefile.in and
3586 crypto/opensslconf.h.in and some help from the perl module
3589 Also, the center of configuration information is no longer
3590 Makefile. Instead, Configure produces a perl module in
3591 configdata.pm which holds most of the config data (in the hash
3592 table %config), the target data that comes from the target
3593 configuration in one of the `Configurations/*.conf` files (in
3598 * To clarify their intended purposes, the Configure options
3599 --prefix and --openssldir change their semantics, and become more
3600 straightforward and less interdependent.
3602 --prefix shall be used exclusively to give the location INSTALLTOP
3603 where programs, scripts, libraries, include files and manuals are
3604 going to be installed. The default is now /usr/local.
3606 --openssldir shall be used exclusively to give the default
3607 location OPENSSLDIR where certificates, private keys, CRLs are
3608 managed. This is also where the default openssl.cnf gets
3610 If the directory given with this option is a relative path, the
3611 values of both the --prefix value and the --openssldir value will
3612 be combined to become OPENSSLDIR.
3613 The default for --openssldir is INSTALLTOP/ssl.
3615 Anyone who uses --openssldir to specify where OpenSSL is to be
3616 installed MUST change to use --prefix instead.
3620 * The GOST engine was out of date and therefore it has been removed. An up
3621 to date GOST engine is now being maintained in an external repository.
3622 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3623 support for GOST ciphersuites (these are only activated if a GOST engine
3628 * EGD is no longer supported by default; use enable-egd when
3631 *Ben Kaduk and Rich Salz*
3633 * The distribution now has Makefile.in files, which are used to
3634 create Makefile's when Configure is run. *Configure must be run
3635 before trying to build now.*
3639 * The return value for SSL_CIPHER_description() for error conditions
3644 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3646 Obtaining and performing DNSSEC validation of TLSA records is
3647 the application's responsibility. The application provides
3648 the TLSA records of its choice to OpenSSL, and these are then
3649 used to authenticate the peer.
3651 The TLSA records need not even come from DNS. They can, for
3652 example, be used to implement local end-entity certificate or
3653 trust-anchor "pinning", where the "pin" data takes the form
3654 of TLSA records, which can augment or replace verification
3655 based on the usual WebPKI public certification authorities.
3659 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3660 continues to support deprecated interfaces in default builds.
3661 However, applications are strongly advised to compile their
3662 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3663 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3664 or the 1.1.0 releases.
3666 In environments in which all applications have been ported to
3667 not use any deprecated interfaces OpenSSL's Configure script
3668 should be used with the --api=1.1.0 option to entirely remove
3669 support for the deprecated features from the library and
3670 unconditionally disable them in the installed headers.
3671 Essentially the same effect can be achieved with the "no-deprecated"
3672 argument to Configure, except that this will always restrict
3673 the build to just the latest API, rather than a fixed API
3676 As applications are ported to future revisions of the API,
3677 they should update their compile-time OPENSSL_API_COMPAT define
3678 accordingly, but in most cases should be able to continue to
3679 compile with later releases.
3681 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3682 0x10000000L and 0x00908000L, respectively. However those
3683 versions did not support the OPENSSL_API_COMPAT feature, and
3684 so applications are not typically tested for explicit support
3685 of just the undeprecated features of either release.
3689 * Add support for setting the minimum and maximum supported protocol.
3690 It can bet set via the SSL_set_min_proto_version() and
3691 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3692 MaxProtocol. It's recommended to use the new APIs to disable
3693 protocols instead of disabling individual protocols using
3694 SSL_set_options() or SSL_CONF's Protocol. This change also
3695 removes support for disabling TLS 1.2 in the OpenSSL TLS
3696 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3700 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3704 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3705 and integrates ECDSA and ECDH functionality into EC. Implementations can
3706 now redirect key generation and no longer need to convert to or from
3709 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3710 include the ec.h header file instead.
3714 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3715 ciphers who are no longer supported and drops support the ephemeral RSA key
3716 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3720 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3721 opaque. For HMAC_CTX, the following constructors and destructors
3724 HMAC_CTX *HMAC_CTX_new(void);
3725 void HMAC_CTX_free(HMAC_CTX *ctx);
3727 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3728 destroy such methods has been added. See EVP_MD_meth_new(3) and
3729 EVP_CIPHER_meth_new(3) for documentation.
3732 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3733 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3734 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3735 an already created structure.
3736 2) For consistency with the majority of our object creators and
3737 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3738 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3739 for deprecated builds.
3743 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3744 cryptographic operations to be performed asynchronously as long as an
3745 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3746 further details. Libssl has also had this capability integrated with the
3747 introduction of the new mode SSL_MODE_ASYNC and associated error
3748 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3749 pages. This work was developed in partnership with Intel Corp.
3753 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3754 always enabled now. If you want to disable the support you should
3755 exclude it using the list of supported ciphers. This also means that the
3756 "-no_ecdhe" option has been removed from s_server.
3760 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3761 SSL_{CTX_}set1_curves() which can set a list.
3765 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3766 curve you want to support using SSL_{CTX_}set1_curves().
3770 * State machine rewrite. The state machine code has been significantly
3771 refactored in order to remove much duplication of code and solve issues
3772 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3773 further details). This change does have some associated API changes.
3774 Notably the SSL_state() function has been removed and replaced by
3775 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3776 SSL_set_state() has been removed altogether. The previous handshake states
3777 defined in ssl.h and ssl3.h have also been removed.
3781 * All instances of the string "ssleay" in the public API were replaced
3782 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3783 Some error codes related to internal RSA_eay API's were renamed.
3787 * The demo files in crypto/threads were moved to demo/threads.
3791 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3794 *Matt Caswell, Rich Salz*
3796 * New ASN.1 embed macro.
3798 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3799 structure is not allocated: it is part of the parent. That is instead of
3807 This reduces memory fragmentation and make it impossible to accidentally
3808 set a mandatory field to NULL.
3810 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3811 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3812 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3817 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3821 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3822 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3823 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3824 DES and RC4 ciphersuites.
3828 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3829 This changes the decoding behaviour for some invalid messages,
3830 though the change is mostly in the more lenient direction, and
3831 legacy behaviour is preserved as much as possible.
3835 * Fix no-stdio build.
3836 *David Woodhouse <David.Woodhouse@intel.com> and also*
3837 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3839 * New testing framework
3840 The testing framework has been largely rewritten and is now using
3841 perl and the perl modules Test::Harness and an extended variant of
3842 Test::More called OpenSSL::Test to do its work. All test scripts in
3843 test/ have been rewritten into test recipes, and all direct calls to
3844 executables in test/Makefile have become individual recipes using the
3845 simplified testing OpenSSL::Test::Simple.
3847 For documentation on our testing modules, do:
3849 perldoc test/testlib/OpenSSL/Test/Simple.pm
3850 perldoc test/testlib/OpenSSL/Test.pm
3854 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3855 are used; the latter aborts on memory leaks (usually checked on exit).
3856 Some undocumented "set malloc, etc., hooks" functions were removed
3857 and others were changed. All are now documented.
3861 * In DSA_generate_parameters_ex, if the provided seed is too short,
3864 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3866 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3867 from RFC4279, RFC4785, RFC5487, RFC5489.
3869 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3870 original RSA_PSK patch.
3874 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3875 era flag was never set throughout the codebase (only read). Also removed
3876 SSL3_FLAGS_POP_BUFFER which was only used if
3877 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3881 * Changed the default name options in the "ca", "crl", "req" and "x509"
3882 to be "oneline" instead of "compat".
3886 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3887 not aware of clients that still exhibit this bug, and the workaround
3888 hasn't been working properly for a while.
3892 * The return type of BIO_number_read() and BIO_number_written() as well as
3893 the corresponding num_read and num_write members in the BIO structure has
3894 changed from unsigned long to uint64_t. On platforms where an unsigned
3895 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3900 * Given the pervasive nature of TLS extensions it is inadvisable to run
3901 OpenSSL without support for them. It also means that maintaining
3902 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3903 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3907 * Removed support for the two export grade static DH ciphersuites
3908 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3909 were newly added (along with a number of other static DH ciphersuites) to
3910 1.0.2. However the two export ones have *never* worked since they were
3911 introduced. It seems strange in any case to be adding new export
3912 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3916 * Version negotiation has been rewritten. In particular SSLv23_method(),
3917 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3918 and turned into macros which simply call the new preferred function names
3919 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3920 should use the new names instead. Also as part of this change the ssl23.h
3921 header file has been removed.
3925 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3926 code and the associated standard is no longer considered fit-for-purpose.
3930 * RT2547 was closed. When generating a private key, try to make the
3931 output file readable only by the owner. This behavior change might
3932 be noticeable when interacting with other software.
3934 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3939 * Added HTTP GET support to the ocsp command.
3943 * Changed default digest for the dgst and enc commands from MD5 to
3948 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3952 * Added support for TLS extended master secret from
3953 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3954 initial patch which was a great help during development.
3958 * All libssl internal structures have been removed from the public header
3959 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3960 now redundant). Users should not attempt to access internal structures
3961 directly. Instead they should use the provided API functions.
3965 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3966 Access to deprecated functions can be re-enabled by running config with
3967 "enable-deprecated". In addition applications wishing to use deprecated
3968 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3969 will, by default, disable some transitive includes that previously existed
3970 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3974 * Added support for OCB mode. OpenSSL has been granted a patent license
3975 compatible with the OpenSSL license for use of OCB. Details are available
3976 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3977 for OCB can be removed by calling config with no-ocb.
3981 * SSLv2 support has been removed. It still supports receiving a SSLv2
3982 compatible client hello.
3986 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3987 done while fixing the error code for the key-too-small case.
3989 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3991 * CA.sh has been removed; use CA.pl instead.
3995 * Removed old DES API.
3999 * Remove various unsupported platforms:
4005 Sinix/ReliantUNIX RM400
4010 16-bit platforms such as WIN16
4014 * Clean up OPENSSL_NO_xxx #define's
4015 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4016 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4017 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4018 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4019 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4020 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4021 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4022 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4023 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4024 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4028 * Cleaned up dead code
4029 Remove all but one '#ifdef undef' which is to be looked at.
4033 * Clean up calling of xxx_free routines.
4034 Just like free(), fix most of the xxx_free routines to accept
4035 NULL. Remove the non-null checks from callers. Save much code.
4039 * Add secure heap for storage of private keys (when possible).
4040 Add BIO_s_secmem(), CBIGNUM, etc.
4041 Contributed by Akamai Technologies under our Corporate CLA.
4045 * Experimental support for a new, fast, unbiased prime candidate generator,
4046 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4048 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4050 * New output format NSS in the sess_id command line tool. This allows
4051 exporting the session id and the master key in NSS keylog format.
4053 *Martin Kaiser <martin@kaiser.cx>*
4055 * Harmonize version and its documentation. -f flag is used to display
4058 *mancha <mancha1@zoho.com>*
4060 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4061 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4063 *mancha <mancha1@zoho.com>*
4065 * Fix some double frees. These are not thought to be exploitable.
4067 *mancha <mancha1@zoho.com>*
4069 * A missing bounds check in the handling of the TLS heartbeat extension
4070 can be used to reveal up to 64k of memory to a connected client or
4073 Thanks for Neel Mehta of Google Security for discovering this bug and to
4074 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4075 preparing the fix ([CVE-2014-0160])
4077 *Adam Langley, Bodo Moeller*
4079 * Fix for the attack described in the paper "Recovering OpenSSL
4080 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4081 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4082 <http://eprint.iacr.org/2014/140>
4084 Thanks to Yuval Yarom and Naomi Benger for discovering this
4085 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4087 *Yuval Yarom and Naomi Benger*
4089 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4090 this fixes a limitation in previous versions of OpenSSL.
4094 * Experimental encrypt-then-mac support.
4096 Experimental support for encrypt then mac from
4097 draft-gutmann-tls-encrypt-then-mac-02.txt
4099 To enable it set the appropriate extension number (0x42 for the test
4100 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4102 For non-compliant peers (i.e. just about everything) this should have no
4105 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4109 * Add EVP support for key wrapping algorithms, to avoid problems with
4110 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4111 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4112 algorithms and include tests cases.
4116 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4121 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4122 MGF1 digest and OAEP label.
4126 * Make openssl verify return errors.
4128 *Chris Palmer <palmer@google.com> and Ben Laurie*
4130 * New function ASN1_TIME_diff to calculate the difference between two
4131 ASN1_TIME structures or one structure and the current time.
4135 * Update fips_test_suite to support multiple command line options. New
4136 test to induce all self test errors in sequence and check expected
4141 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4142 sign or verify all in one operation.
4146 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4147 test programs and fips_test_suite. Includes functionality to parse
4148 the minimal script output of fipsalgest.pl directly.
4152 * Add authorisation parameter to FIPS_module_mode_set().
4156 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4160 * Use separate DRBG fields for internal and external flags. New function
4161 FIPS_drbg_health_check() to perform on demand health checking. Add
4162 generation tests to fips_test_suite with reduced health check interval to
4163 demonstrate periodic health checking. Add "nodh" option to
4164 fips_test_suite to skip very slow DH test.
4168 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4173 * More extensive health check for DRBG checking many more failure modes.
4174 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4175 combination: call this in fips_test_suite.
4179 * Add support for canonical generation of DSA parameter 'g'. See
4182 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4183 POST to handle HMAC cases.
4187 * Add functions FIPS_module_version() and FIPS_module_version_text()
4188 to return numerical and string versions of the FIPS module number.
4192 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4193 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4194 outside the validated module in the FIPS capable OpenSSL.
4198 * Minor change to DRBG entropy callback semantics. In some cases
4199 there is no multiple of the block length between min_len and
4200 max_len. Allow the callback to return more than max_len bytes
4201 of entropy but discard any extra: it is the callback's responsibility
4202 to ensure that the extra data discarded does not impact the
4203 requested amount of entropy.
4207 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4208 information in FIPS186-3, SP800-57 and SP800-131A.
4212 * CCM support via EVP. Interface is very similar to GCM case except we
4213 must supply all data in one chunk (i.e. no update, final) and the
4214 message length must be supplied if AAD is used. Add algorithm test
4219 * Initial version of POST overhaul. Add POST callback to allow the status
4220 of POST to be monitored and/or failures induced. Modify fips_test_suite
4221 to use callback. Always run all selftests even if one fails.
4225 * XTS support including algorithm test driver in the fips_gcmtest program.
4226 Note: this does increase the maximum key length from 32 to 64 bytes but
4227 there should be no binary compatibility issues as existing applications
4228 will never use XTS mode.
4232 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4233 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4234 performs algorithm blocking for unapproved PRNG types. Also do not
4235 set PRNG type in FIPS_mode_set(): leave this to the application.
4236 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4237 the standard OpenSSL PRNG: set additional data to a date time vector.
4241 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4242 This shouldn't present any incompatibility problems because applications
4243 shouldn't be using these directly and any that are will need to rethink
4244 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4248 * Extensive self tests and health checking required by SP800-90 DRBG.
4249 Remove strength parameter from FIPS_drbg_instantiate and always
4250 instantiate at maximum supported strength.
4254 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4258 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4262 * New function DH_compute_key_padded() to compute a DH key and pad with
4263 leading zeroes if needed: this complies with SP800-56A et al.
4267 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4268 anything, incomplete, subject to change and largely untested at present.
4272 * Modify fipscanisteronly build option to only build the necessary object
4273 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4277 * Add experimental option FIPSSYMS to give all symbols in
4278 fipscanister.o and FIPS or fips prefix. This will avoid
4279 conflicts with future versions of OpenSSL. Add perl script
4280 util/fipsas.pl to preprocess assembly language source files
4281 and rename any affected symbols.
4285 * Add selftest checks and algorithm block of non-fips algorithms in
4286 FIPS mode. Remove DES2 from selftests.
4290 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4291 return internal method without any ENGINE dependencies. Add new
4292 tiny fips sign and verify functions.
4296 * New build option no-ec2m to disable characteristic 2 code.
4300 * New build option "fipscanisteronly". This only builds fipscanister.o
4301 and (currently) associated fips utilities. Uses the file Makefile.fips
4302 instead of Makefile.org as the prototype.
4306 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4307 Update fips_gcmtest to use IV generator.
4311 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4312 setting output buffer to NULL. The `*Final` function must be
4313 called although it will not retrieve any additional data. The tag
4314 can be set or retrieved with a ctrl. The IV length is by default 12
4315 bytes (96 bits) but can be set to an alternative value. If the IV
4316 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4321 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4322 underlying do_cipher function handles all cipher semantics itself
4323 including padding and finalisation. This is useful if (for example)
4324 an ENGINE cipher handles block padding itself. The behaviour of
4325 do_cipher is subtly changed if this flag is set: the return value
4326 is the number of characters written to the output buffer (zero is
4327 no longer an error code) or a negative error code. Also if the
4328 input buffer is NULL and length 0 finalisation should be performed.
4332 * If a candidate issuer certificate is already part of the constructed
4333 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4337 * Improve forward-security support: add functions
4339 void SSL_CTX_set_not_resumable_session_callback(
4340 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4341 void SSL_set_not_resumable_session_callback(
4342 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4344 for use by SSL/TLS servers; the callback function will be called whenever a
4345 new session is created, and gets to decide whether the session may be
4346 cached to make it resumable (return 0) or not (return 1). (As by the
4347 SSL/TLS protocol specifications, the session_id sent by the server will be
4348 empty to indicate that the session is not resumable; also, the server will
4349 not generate RFC 4507 (RFC 5077) session tickets.)
4351 A simple reasonable callback implementation is to return is_forward_secure.
4352 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4353 by the SSL/TLS server library, indicating whether it can provide forward
4356 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4358 * New -verify_name option in command line utilities to set verification
4363 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4364 Add CMAC pkey methods.
4368 * Experimental renegotiation in s_server -www mode. If the client
4369 browses /reneg connection is renegotiated. If /renegcert it is
4370 renegotiated requesting a certificate.
4374 * Add an "external" session cache for debugging purposes to s_server. This
4375 should help trace issues which normally are only apparent in deployed
4376 multi-process servers.
4380 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4381 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4382 BIO_set_cipher() and some obscure PEM functions were changed so they
4383 can now return an error. The RAND changes required a change to the
4384 RAND_METHOD structure.
4388 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4389 a gcc attribute to warn if the result of a function is ignored. This
4390 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4391 whose return value is often ignored.
4395 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4396 These allow SCTs (signed certificate timestamps) to be requested and
4397 validated when establishing a connection.
4399 *Rob Percival <robpercival@google.com>*
4404 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4406 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4407 used even when parsing explicit parameters, when loading a encoded key
4408 or calling `EC_GROUP_new_from_ecpkparameters()`/
4409 `EC_GROUP_new_from_ecparameters()`.
4410 This prevents bypass of security hardening and performance gains,
4411 especially for curves with specialized EC_METHODs.
4412 By default, if a key encoded with explicit parameters is loaded and later
4413 encoded, the output is still encoded with explicit parameters, even if
4414 internally a "named" EC_GROUP is used for computation.
4418 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4419 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4420 NULL. After this change, only the cofactor parameter can be NULL. It also
4421 does some minimal sanity checks on the passed order.
4426 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4427 An attack is simple, if the first CMS_recipientInfo is valid but the
4428 second CMS_recipientInfo is chosen ciphertext. If the second
4429 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4430 encryption key will be replaced by garbage, and the message cannot be
4431 decoded, but if the RSA decryption fails, the correct encryption key is
4432 used and the recipient will not notice the attack.
4433 As a work around for this potential attack the length of the decrypted
4434 key must be equal to the cipher default key length, in case the
4435 certifiate is not given and all recipientInfo are tried out.
4436 The old behaviour can be re-enabled in the CMS code by setting the
4437 CMS_DEBUG_DECRYPT flag.
4442 * Document issue with installation paths in diverse Windows builds
4444 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4445 binaries and run-time config file.
4450 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4452 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4453 This changes the size when using the `genpkey` command when no size is given.
4454 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4455 generation commands to use 2048 bits by default.
4459 * Add FIPS support for Android Arm 64-bit
4461 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4462 Module in Version 2.0.10. For some reason, the corresponding target
4463 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4464 built with FIPS support on Android Arm 64-bit. This omission has been
4467 *Matthias St. Pierre*
4469 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4471 * 0-byte record padding oracle
4473 If an application encounters a fatal protocol error and then calls
4474 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4475 then OpenSSL can respond differently to the calling application if a 0 byte
4476 record is received with invalid padding compared to if a 0 byte record is
4477 received with an invalid MAC. If the application then behaves differently
4478 based on that in a way that is detectable to the remote peer, then this
4479 amounts to a padding oracle that could be used to decrypt data.
4481 In order for this to be exploitable "non-stitched" ciphersuites must be in
4482 use. Stitched ciphersuites are optimised implementations of certain
4483 commonly used ciphersuites. Also the application must call SSL_shutdown()
4484 twice even if a protocol error has occurred (applications should not do
4485 this but some do anyway).
4487 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4488 Aviram, with additional investigation by Steven Collison and Andrew
4489 Hourselt. It was reported to OpenSSL on 10th December 2018.
4494 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4498 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4500 * Microarchitecture timing vulnerability in ECC scalar multiplication
4502 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4503 shown to be vulnerable to a microarchitecture timing side channel attack.
4504 An attacker with sufficient access to mount local timing attacks during
4505 ECDSA signature generation could recover the private key.
4507 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4508 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4514 * Timing vulnerability in DSA signature generation
4516 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4517 timing side channel attack. An attacker could use variations in the signing
4518 algorithm to recover the private key.
4520 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4525 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4526 Module, accidentally introduced while backporting security fixes from the
4527 development branch and hindering the use of ECC in FIPS mode.
4531 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4533 * Client DoS due to large DH parameter
4535 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4536 malicious server can send a very large prime value to the client. This will
4537 cause the client to spend an unreasonably long period of time generating a
4538 key for this prime resulting in a hang until the client has finished. This
4539 could be exploited in a Denial Of Service attack.
4541 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4546 * Cache timing vulnerability in RSA Key Generation
4548 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4549 a cache timing side channel attack. An attacker with sufficient access to
4550 mount cache timing attacks during the RSA key generation process could
4551 recover the private key.
4553 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4554 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4559 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4560 parameter is no longer accepted, as it leads to a corrupt table. NULL
4561 pem_str is reserved for alias entries only.
4565 * Revert blinding in ECDSA sign and instead make problematic addition
4566 length-invariant. Switch even to fixed-length Montgomery multiplication.
4570 * Change generating and checking of primes so that the error rate of not
4571 being prime depends on the intended use based on the size of the input.
4572 For larger primes this will result in more rounds of Miller-Rabin.
4573 The maximal error rate for primes with more than 1080 bits is lowered
4576 *Kurt Roeckx, Annie Yousar*
4578 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4582 * Add blinding to ECDSA and DSA signatures to protect against side channel
4583 attacks discovered by Keegan Ryan (NCC Group).
4587 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4588 now allow empty (zero character) pass phrases.
4592 * Certificate time validation (X509_cmp_time) enforces stricter
4593 compliance with RFC 5280. Fractional seconds and timezone offsets
4594 are no longer allowed.
4598 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4600 * Constructed ASN.1 types with a recursive definition could exceed the stack
4602 Constructed ASN.1 types with a recursive definition (such as can be found
4603 in PKCS7) could eventually exceed the stack given malicious input with
4604 excessive recursion. This could result in a Denial Of Service attack. There
4605 are no such structures used within SSL/TLS that come from untrusted sources
4606 so this is considered safe.
4608 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4614 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4616 * Read/write after SSL object in error state
4618 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4619 mechanism. The intent was that if a fatal error occurred during a handshake
4620 then OpenSSL would move into the error state and would immediately fail if
4621 you attempted to continue the handshake. This works as designed for the
4622 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4623 SSL_connect()), however due to a bug it does not work correctly if
4624 SSL_read() or SSL_write() is called directly. In that scenario, if the
4625 handshake fails then a fatal error will be returned in the initial function
4626 call. If SSL_read()/SSL_write() is subsequently called by the application
4627 for the same SSL object then it will succeed and the data is passed without
4628 being decrypted/encrypted directly from the SSL/TLS record layer.
4630 In order to exploit this issue an application bug would have to be present
4631 that resulted in a call to SSL_read()/SSL_write() being issued after having
4632 already received a fatal error.
4634 This issue was reported to OpenSSL by David Benjamin (Google).
4639 * rsaz_1024_mul_avx2 overflow bug on x86_64
4641 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4642 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4643 Analysis suggests that attacks against RSA and DSA as a result of this
4644 defect would be very difficult to perform and are not believed likely.
4645 Attacks against DH1024 are considered just feasible, because most of the
4646 work necessary to deduce information about a private key may be performed
4647 offline. The amount of resources required for such an attack would be
4648 significant. However, for an attack on TLS to be meaningful, the server
4649 would have to share the DH1024 private key among multiple clients, which is
4650 no longer an option since CVE-2016-0701.
4652 This only affects processors that support the AVX2 but not ADX extensions
4653 like Intel Haswell (4th generation).
4655 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4656 was originally found via the OSS-Fuzz project.
4661 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4663 * bn_sqrx8x_internal carry bug on x86_64
4665 There is a carry propagating bug in the x86_64 Montgomery squaring
4666 procedure. No EC algorithms are affected. Analysis suggests that attacks
4667 against RSA and DSA as a result of this defect would be very difficult to
4668 perform and are not believed likely. Attacks against DH are considered just
4669 feasible (although very difficult) because most of the work necessary to
4670 deduce information about a private key may be performed offline. The amount
4671 of resources required for such an attack would be very significant and
4672 likely only accessible to a limited number of attackers. An attacker would
4673 additionally need online access to an unpatched system using the target
4674 private key in a scenario with persistent DH parameters and a private
4675 key that is shared between multiple clients.
4677 This only affects processors that support the BMI1, BMI2 and ADX extensions
4678 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4680 This issue was reported to OpenSSL by the OSS-Fuzz project.
4685 * Malformed X.509 IPAddressFamily could cause OOB read
4687 If an X.509 certificate has a malformed IPAddressFamily extension,
4688 OpenSSL could do a one-byte buffer overread. The most likely result
4689 would be an erroneous display of the certificate in text format.
4691 This issue was reported to OpenSSL by the OSS-Fuzz project.
4696 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4698 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4699 platform rather than 'mingw'.
4703 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4705 * Truncated packet could crash via OOB read
4707 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4708 cipher is being used, then a truncated packet can cause that host to
4709 perform an out-of-bounds read, usually resulting in a crash.
4711 This issue was reported to OpenSSL by Robert Święcki of Google.
4716 * BN_mod_exp may produce incorrect results on x86_64
4718 There is a carry propagating bug in the x86_64 Montgomery squaring
4719 procedure. No EC algorithms are affected. Analysis suggests that attacks
4720 against RSA and DSA as a result of this defect would be very difficult to
4721 perform and are not believed likely. Attacks against DH are considered just
4722 feasible (although very difficult) because most of the work necessary to
4723 deduce information about a private key may be performed offline. The amount
4724 of resources required for such an attack would be very significant and
4725 likely only accessible to a limited number of attackers. An attacker would
4726 additionally need online access to an unpatched system using the target
4727 private key in a scenario with persistent DH parameters and a private
4728 key that is shared between multiple clients. For example this can occur by
4729 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4730 similar to CVE-2015-3193 but must be treated as a separate problem.
4732 This issue was reported to OpenSSL by the OSS-Fuzz project.
4737 * Montgomery multiplication may produce incorrect results
4739 There is a carry propagating bug in the Broadwell-specific Montgomery
4740 multiplication procedure that handles input lengths divisible by, but
4741 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4742 and DH private keys are impossible. This is because the subroutine in
4743 question is not used in operations with the private key itself and an input
4744 of the attacker's direct choice. Otherwise the bug can manifest itself as
4745 transient authentication and key negotiation failures or reproducible
4746 erroneous outcome of public-key operations with specially crafted input.
4747 Among EC algorithms only Brainpool P-512 curves are affected and one
4748 presumably can attack ECDH key negotiation. Impact was not analyzed in
4749 detail, because pre-requisites for attack are considered unlikely. Namely
4750 multiple clients have to choose the curve in question and the server has to
4751 share the private key among them, neither of which is default behaviour.
4752 Even then only clients that chose the curve will be affected.
4754 This issue was publicly reported as transient failures and was not
4755 initially recognized as a security issue. Thanks to Richard Morgan for
4756 providing reproducible case.
4761 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4762 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4763 prevent issues where no progress is being made and the peer continually
4764 sends unrecognised record types, using up resources processing them.
4768 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4770 * Missing CRL sanity check
4772 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4773 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4774 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4776 This issue only affects the OpenSSL 1.0.2i
4781 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4783 * OCSP Status Request extension unbounded memory growth
4785 A malicious client can send an excessively large OCSP Status Request
4786 extension. If that client continually requests renegotiation, sending a
4787 large OCSP Status Request extension each time, then there will be unbounded
4788 memory growth on the server. This will eventually lead to a Denial Of
4789 Service attack through memory exhaustion. Servers with a default
4790 configuration are vulnerable even if they do not support OCSP. Builds using
4791 the "no-ocsp" build time option are not affected.
4793 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4798 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4801 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4807 * OOB write in MDC2_Update()
4809 An overflow can occur in MDC2_Update() either if called directly or
4810 through the EVP_DigestUpdate() function using MDC2. If an attacker
4811 is able to supply very large amounts of input data after a previous
4812 call to EVP_EncryptUpdate() with a partial block then a length check
4813 can overflow resulting in a heap corruption.
4815 The amount of data needed is comparable to SIZE_MAX which is impractical
4818 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4823 * Malformed SHA512 ticket DoS
4825 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4826 DoS attack where a malformed ticket will result in an OOB read which will
4829 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4830 a custom server callback and ticket lookup mechanism.
4832 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4837 * OOB write in BN_bn2dec()
4839 The function BN_bn2dec() does not check the return value of BN_div_word().
4840 This can cause an OOB write if an application uses this function with an
4841 overly large BIGNUM. This could be a problem if an overly large certificate
4842 or CRL is printed out from an untrusted source. TLS is not affected because
4843 record limits will reject an oversized certificate before it is parsed.
4845 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4850 * OOB read in TS_OBJ_print_bio()
4852 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4853 the total length the OID text representation would use and not the amount
4854 of data written. This will result in OOB reads when large OIDs are
4857 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4862 * Pointer arithmetic undefined behaviour
4864 Avoid some undefined pointer arithmetic
4866 A common idiom in the codebase is to check limits in the following manner:
4869 Where "p" points to some malloc'd data of SIZE bytes and
4872 "len" here could be from some externally supplied data (e.g. from a TLS
4875 The rules of C pointer arithmetic are such that "p + len" is only well
4876 defined where len <= SIZE. Therefore the above idiom is actually
4877 undefined behaviour.
4879 For example this could cause problems if some malloc implementation
4880 provides an address for "p" such that "p + len" actually overflows for
4881 values of len that are too big and therefore p + len < limit.
4883 This issue was reported to OpenSSL by Guido Vranken
4888 * Constant time flag not preserved in DSA signing
4890 Operations in the DSA signing algorithm should run in constant time in
4891 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4892 implementation means that a non-constant time codepath is followed for
4893 certain operations. This has been demonstrated through a cache-timing
4894 attack to be sufficient for an attacker to recover the private DSA key.
4896 This issue was reported by César Pereida (Aalto University), Billy Brumley
4897 (Tampere University of Technology), and Yuval Yarom (The University of
4898 Adelaide and NICTA).
4903 * DTLS buffered message DoS
4905 In a DTLS connection where handshake messages are delivered out-of-order
4906 those messages that OpenSSL is not yet ready to process will be buffered
4907 for later use. Under certain circumstances, a flaw in the logic means that
4908 those messages do not get removed from the buffer even though the handshake
4909 has been completed. An attacker could force up to approx. 15 messages to
4910 remain in the buffer when they are no longer required. These messages will
4911 be cleared when the DTLS connection is closed. The default maximum size for
4912 a message is 100k. Therefore the attacker could force an additional 1500k
4913 to be consumed per connection. By opening many simulataneous connections an
4914 attacker could cause a DoS attack through memory exhaustion.
4916 This issue was reported to OpenSSL by Quan Luo.
4921 * DTLS replay protection DoS
4923 A flaw in the DTLS replay attack protection mechanism means that records
4924 that arrive for future epochs update the replay protection "window" before
4925 the MAC for the record has been validated. This could be exploited by an
4926 attacker by sending a record for the next epoch (which does not have to
4927 decrypt or have a valid MAC), with a very large sequence number. This means
4928 that all subsequent legitimate packets are dropped causing a denial of
4929 service for a specific DTLS connection.
4931 This issue was reported to OpenSSL by the OCAP audit team.
4936 * Certificate message OOB reads
4938 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4939 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4940 theoretical DoS risk but this has not been observed in practice on common
4943 The messages affected are client certificate, client certificate request
4944 and server certificate. As a result the attack can only be performed
4945 against a client or a server which enables client authentication.
4947 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4952 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4954 * Prevent padding oracle in AES-NI CBC MAC check
4956 A MITM attacker can use a padding oracle attack to decrypt traffic
4957 when the connection uses an AES CBC cipher and the server support
4960 This issue was introduced as part of the fix for Lucky 13 padding
4961 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4962 constant time by making sure that always the same bytes are read and
4963 compared against either the MAC or padding bytes. But it no longer
4964 checked that there was enough data to have both the MAC and padding
4967 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4972 * Fix EVP_EncodeUpdate overflow
4974 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4975 Base64 encoding of binary data. If an attacker is able to supply very large
4976 amounts of input data then a length check can overflow resulting in a heap
4979 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4980 the `PEM_write_bio*` family of functions. These are mainly used within the
4981 OpenSSL command line applications, so any application which processes data
4982 from an untrusted source and outputs it as a PEM file should be considered
4983 vulnerable to this issue. User applications that call these APIs directly
4984 with large amounts of untrusted data may also be vulnerable.
4986 This issue was reported by Guido Vranken.
4991 * Fix EVP_EncryptUpdate overflow
4993 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4994 is able to supply very large amounts of input data after a previous call to
4995 EVP_EncryptUpdate() with a partial block then a length check can overflow
4996 resulting in a heap corruption. Following an analysis of all OpenSSL
4997 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4998 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4999 the first called function after an EVP_EncryptInit(), and therefore that
5000 specific call must be safe. The second form is where the length passed to
5001 EVP_EncryptUpdate() can be seen from the code to be some small value and
5002 therefore there is no possibility of an overflow. Since all instances are
5003 one of these two forms, it is believed that there can be no overflows in
5004 internal code due to this problem. It should be noted that
5005 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5006 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5007 of these calls have also been analysed too and it is believed there are no
5008 instances in internal usage where an overflow could occur.
5010 This issue was reported by Guido Vranken.
5015 * Prevent ASN.1 BIO excessive memory allocation
5017 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5018 a short invalid encoding can cause allocation of large amounts of memory
5019 potentially consuming excessive resources or exhausting memory.
5021 Any application parsing untrusted data through d2i BIO functions is
5022 affected. The memory based functions such as d2i_X509() are *not* affected.
5023 Since the memory based functions are used by the TLS library, TLS
5024 applications are not affected.
5026 This issue was reported by Brian Carpenter.
5033 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5034 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5035 in arbitrary stack data being returned in the buffer.
5037 This issue was reported by Guido Vranken.
5042 * Modify behavior of ALPN to invoke callback after SNI/servername
5043 callback, such that updates to the SSL_CTX affect ALPN.
5047 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5052 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5053 methods are enabled and ssl2 is disabled the methods return NULL.
5057 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5059 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5060 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5061 provide any "EXPORT" or "LOW" strength ciphers.
5065 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5066 is by default disabled at build-time. Builds that are not configured with
5067 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5068 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5069 will need to explicitly call either of:
5071 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5073 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5075 as appropriate. Even if either of those is used, or the application
5076 explicitly uses the version-specific SSLv2_method() or its client and
5077 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5078 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5079 ciphers, and SSLv2 56-bit DES are no longer available.
5084 * Fix a double-free in DSA code
5086 A double free bug was discovered when OpenSSL parses malformed DSA private
5087 keys and could lead to a DoS attack or memory corruption for applications
5088 that receive DSA private keys from untrusted sources. This scenario is
5091 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5097 * Disable SRP fake user seed to address a server memory leak.
5099 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5101 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5102 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5103 was changed to ignore the "fake user" SRP seed, even if the seed
5106 Users should use SRP_VBASE_get1_by_user instead. Note that in
5107 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5108 also that even though configuring the SRP seed attempts to hide
5109 invalid usernames by continuing the handshake with fake
5110 credentials, this behaviour is not constant time and no strong
5111 guarantees are made that the handshake is indistinguishable from
5112 that of a valid user.
5117 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5119 In the BN_hex2bn function the number of hex digits is calculated using an
5120 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5121 large values of `i` this can result in `bn_expand` not allocating any
5122 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5123 field as NULL leading to a subsequent NULL ptr deref. For very large values
5124 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5125 In this case memory is allocated to the internal BIGNUM data field, but it
5126 is insufficiently sized leading to heap corruption. A similar issue exists
5127 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5128 is ever called by user applications with very large untrusted hex/dec data.
5129 This is anticipated to be a rare occurrence.
5131 All OpenSSL internal usage of these functions use data that is not expected
5132 to be untrusted, e.g. config file data or application command line
5133 arguments. If user developed applications generate config file data based
5134 on untrusted data then it is possible that this could also lead to security
5135 consequences. This is also anticipated to be rare.
5137 This issue was reported to OpenSSL by Guido Vranken.
5142 * Fix memory issues in `BIO_*printf` functions
5144 The internal `fmtstr` function used in processing a "%s" format string in
5145 the `BIO_*printf` functions could overflow while calculating the length of a
5146 string and cause an OOB read when printing very long strings.
5148 Additionally the internal `doapr_outch` function can attempt to write to an
5149 OOB memory location (at an offset from the NULL pointer) in the event of a
5150 memory allocation failure. In 1.0.2 and below this could be caused where
5151 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5152 could be in processing a very long "%s" format string. Memory leaks can
5155 The first issue may mask the second issue dependent on compiler behaviour.
5156 These problems could enable attacks where large amounts of untrusted data
5157 is passed to the `BIO_*printf` functions. If applications use these functions
5158 in this way then they could be vulnerable. OpenSSL itself uses these
5159 functions when printing out human-readable dumps of ASN.1 data. Therefore
5160 applications that print this data could be vulnerable if the data is from
5161 untrusted sources. OpenSSL command line applications could also be
5162 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5163 as command line arguments.
5165 Libssl is not considered directly vulnerable. Additionally certificates etc
5166 received via remote connections via libssl are also unlikely to be able to
5167 trigger these issues because of message size limits enforced within libssl.
5169 This issue was reported to OpenSSL Guido Vranken.
5174 * Side channel attack on modular exponentiation
5176 A side-channel attack was found which makes use of cache-bank conflicts on
5177 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5178 of RSA keys. The ability to exploit this issue is limited as it relies on
5179 an attacker who has control of code in a thread running on the same
5180 hyper-threaded core as the victim thread which is performing decryptions.
5182 This issue was reported to OpenSSL by Yuval Yarom, The University of
5183 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5184 Nadia Heninger, University of Pennsylvania with more information at
5185 <http://cachebleed.info>.
5190 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5191 if no keysize is specified with default_bits. This fixes an
5192 omission in an earlier change that changed all RSA/DSA key generation
5193 commands to use 2048 bits by default.
5197 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5199 * DH small subgroups
5201 Historically OpenSSL only ever generated DH parameters based on "safe"
5202 primes. More recently (in version 1.0.2) support was provided for
5203 generating X9.42 style parameter files such as those required for RFC 5114
5204 support. The primes used in such files may not be "safe". Where an
5205 application is using DH configured with parameters based on primes that are
5206 not "safe" then an attacker could use this fact to find a peer's private
5207 DH exponent. This attack requires that the attacker complete multiple
5208 handshakes in which the peer uses the same private DH exponent. For example
5209 this could be used to discover a TLS server's private DH exponent if it's
5210 reusing the private DH exponent or it's using a static DH ciphersuite.
5212 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5213 TLS. It is not on by default. If the option is not set then the server
5214 reuses the same private DH exponent for the life of the server process and
5215 would be vulnerable to this attack. It is believed that many popular
5216 applications do set this option and would therefore not be at risk.
5218 The fix for this issue adds an additional check where a "q" parameter is
5219 available (as is the case in X9.42 based parameters). This detects the
5220 only known attack, and is the only possible defense for static DH
5221 ciphersuites. This could have some performance impact.
5223 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5224 default and cannot be disabled. This could have some performance impact.
5226 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5231 * SSLv2 doesn't block disabled ciphers
5233 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5234 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5235 been disabled, provided that the SSLv2 protocol was not also disabled via
5238 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5239 and Sebastian Schinzel.
5244 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5246 * BN_mod_exp may produce incorrect results on x86_64
5248 There is a carry propagating bug in the x86_64 Montgomery squaring
5249 procedure. No EC algorithms are affected. Analysis suggests that attacks
5250 against RSA and DSA as a result of this defect would be very difficult to
5251 perform and are not believed likely. Attacks against DH are considered just
5252 feasible (although very difficult) because most of the work necessary to
5253 deduce information about a private key may be performed offline. The amount
5254 of resources required for such an attack would be very significant and
5255 likely only accessible to a limited number of attackers. An attacker would
5256 additionally need online access to an unpatched system using the target
5257 private key in a scenario with persistent DH parameters and a private
5258 key that is shared between multiple clients. For example this can occur by
5259 default in OpenSSL DHE based SSL/TLS ciphersuites.
5261 This issue was reported to OpenSSL by Hanno Böck.
5266 * Certificate verify crash with missing PSS parameter
5268 The signature verification routines will crash with a NULL pointer
5269 dereference if presented with an ASN.1 signature using the RSA PSS
5270 algorithm and absent mask generation function parameter. Since these
5271 routines are used to verify certificate signature algorithms this can be
5272 used to crash any certificate verification operation and exploited in a
5273 DoS attack. Any application which performs certificate verification is
5274 vulnerable including OpenSSL clients and servers which enable client
5277 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5282 * X509_ATTRIBUTE memory leak
5284 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5285 memory. This structure is used by the PKCS#7 and CMS routines so any
5286 application which reads PKCS#7 or CMS data from untrusted sources is
5287 affected. SSL/TLS is not affected.
5289 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5295 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5296 This changes the decoding behaviour for some invalid messages,
5297 though the change is mostly in the more lenient direction, and
5298 legacy behaviour is preserved as much as possible.
5302 * In DSA_generate_parameters_ex, if the provided seed is too short,
5305 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5307 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5309 * Alternate chains certificate forgery
5311 During certificate verification, OpenSSL will attempt to find an
5312 alternative certificate chain if the first attempt to build such a chain
5313 fails. An error in the implementation of this logic can mean that an
5314 attacker could cause certain checks on untrusted certificates to be
5315 bypassed, such as the CA flag, enabling them to use a valid leaf
5316 certificate to act as a CA and "issue" an invalid certificate.
5318 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5323 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5325 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5326 incompatibility in the handling of HMAC. The previous ABI has now been
5331 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5333 * Malformed ECParameters causes infinite loop
5335 When processing an ECParameters structure OpenSSL enters an infinite loop
5336 if the curve specified is over a specially malformed binary polynomial
5339 This can be used to perform denial of service against any
5340 system which processes public keys, certificate requests or
5341 certificates. This includes TLS clients and TLS servers with
5342 client authentication enabled.
5344 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5349 * Exploitable out-of-bounds read in X509_cmp_time
5351 X509_cmp_time does not properly check the length of the ASN1_TIME
5352 string and can read a few bytes out of bounds. In addition,
5353 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5356 An attacker can use this to craft malformed certificates and CRLs of
5357 various sizes and potentially cause a segmentation fault, resulting in
5358 a DoS on applications that verify certificates or CRLs. TLS clients
5359 that verify CRLs are affected. TLS clients and servers with client
5360 authentication enabled may be affected if they use custom verification
5363 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5364 independently by Hanno Böck.
5369 * PKCS7 crash with missing EnvelopedContent
5371 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5372 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5373 with missing content and trigger a NULL pointer dereference on parsing.
5375 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5376 structures from untrusted sources are affected. OpenSSL clients and
5377 servers are not affected.
5379 This issue was reported to OpenSSL by Michal Zalewski (Google).
5384 * CMS verify infinite loop with unknown hash function
5386 When verifying a signedData message the CMS code can enter an infinite loop
5387 if presented with an unknown hash function OID. This can be used to perform
5388 denial of service against any system which verifies signedData messages using
5390 This issue was reported to OpenSSL by Johannes Bauer.
5395 * Race condition handling NewSessionTicket
5397 If a NewSessionTicket is received by a multi-threaded client when attempting to
5398 reuse a previous ticket then a race condition can occur potentially leading to
5399 a double free of the ticket data.
5404 * Only support 256-bit or stronger elliptic curves with the
5405 'ecdh_auto' setting (server) or by default (client). Of supported
5406 curves, prefer P-256 (both).
5410 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5412 * ClientHello sigalgs DoS fix
5414 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5415 invalid signature algorithms extension a NULL pointer dereference will
5416 occur. This can be exploited in a DoS attack against the server.
5418 This issue was was reported to OpenSSL by David Ramos of Stanford
5422 *Stephen Henson and Matt Caswell*
5424 * Multiblock corrupted pointer fix
5426 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5427 feature only applies on 64 bit x86 architecture platforms that support AES
5428 NI instructions. A defect in the implementation of "multiblock" can cause
5429 OpenSSL's internal write buffer to become incorrectly set to NULL when
5430 using non-blocking IO. Typically, when the user application is using a
5431 socket BIO for writing, this will only result in a failed connection.
5432 However if some other BIO is used then it is likely that a segmentation
5433 fault will be triggered, thus enabling a potential DoS attack.
5435 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5440 * Segmentation fault in DTLSv1_listen fix
5442 The DTLSv1_listen function is intended to be stateless and processes the
5443 initial ClientHello from many peers. It is common for user code to loop
5444 over the call to DTLSv1_listen until a valid ClientHello is received with
5445 an associated cookie. A defect in the implementation of DTLSv1_listen means
5446 that state is preserved in the SSL object from one invocation to the next
5447 that can lead to a segmentation fault. Errors processing the initial
5448 ClientHello can trigger this scenario. An example of such an error could be
5449 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5452 This issue was reported to OpenSSL by Per Allansson.
5457 * Segmentation fault in ASN1_TYPE_cmp fix
5459 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5460 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5461 certificate signature algorithm consistency this can be used to crash any
5462 certificate verification operation and exploited in a DoS attack. Any
5463 application which performs certificate verification is vulnerable including
5464 OpenSSL clients and servers which enable client authentication.
5469 * Segmentation fault for invalid PSS parameters fix
5471 The signature verification routines will crash with a NULL pointer
5472 dereference if presented with an ASN.1 signature using the RSA PSS
5473 algorithm and invalid parameters. Since these routines are used to verify
5474 certificate signature algorithms this can be used to crash any
5475 certificate verification operation and exploited in a DoS attack. Any
5476 application which performs certificate verification is vulnerable including
5477 OpenSSL clients and servers which enable client authentication.
5479 This issue was was reported to OpenSSL by Brian Carpenter.
5484 * ASN.1 structure reuse memory corruption fix
5486 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5487 memory corruption via an invalid write. Such reuse is and has been
5488 strongly discouraged and is believed to be rare.
5490 Applications that parse structures containing CHOICE or ANY DEFINED BY
5491 components may be affected. Certificate parsing (d2i_X509 and related
5492 functions) are however not affected. OpenSSL clients and servers are
5498 * PKCS7 NULL pointer dereferences fix
5500 The PKCS#7 parsing code does not handle missing outer ContentInfo
5501 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5502 missing content and trigger a NULL pointer dereference on parsing.
5504 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5505 otherwise parse PKCS#7 structures from untrusted sources are
5506 affected. OpenSSL clients and servers are not affected.
5508 This issue was reported to OpenSSL by Michal Zalewski (Google).
5513 * DoS via reachable assert in SSLv2 servers fix
5515 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5516 servers that both support SSLv2 and enable export cipher suites by sending
5517 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5519 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5520 (OpenSSL development team).
5525 * Empty CKE with client auth and DHE fix
5527 If client auth is used then a server can seg fault in the event of a DHE
5528 ciphersuite being selected and a zero length ClientKeyExchange message
5529 being sent by the client. This could be exploited in a DoS attack.
5534 * Handshake with unseeded PRNG fix
5536 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5537 with an unseeded PRNG. The conditions are:
5538 - The client is on a platform where the PRNG has not been seeded
5539 automatically, and the user has not seeded manually
5540 - A protocol specific client method version has been used (i.e. not
5541 SSL_client_methodv23)
5542 - A ciphersuite is used that does not require additional random data from
5543 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5545 If the handshake succeeds then the client random that has been used will
5546 have been generated from a PRNG with insufficient entropy and therefore the
5547 output may be predictable.
5549 For example using the following command with an unseeded openssl will
5550 succeed on an unpatched platform:
5552 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5557 * Use After Free following d2i_ECPrivatekey error fix
5559 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5560 could cause a use after free condition. This, in turn, could cause a double
5561 free in several private key parsing functions (such as d2i_PrivateKey
5562 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5563 for applications that receive EC private keys from untrusted
5564 sources. This scenario is considered rare.
5566 This issue was discovered by the BoringSSL project and fixed in their
5572 * X509_to_X509_REQ NULL pointer deref fix
5574 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5575 the certificate key is invalid. This function is rarely used in practice.
5577 This issue was discovered by Brian Carpenter.
5582 * Removed the export ciphers from the DEFAULT ciphers
5586 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5588 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5589 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5590 So far those who have to target multiple platforms would compromise
5591 and argue that binary targeting say ARMv5 would still execute on
5592 ARMv8. "Universal" build resolves this compromise by providing
5593 near-optimal performance even on newer platforms.
5597 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5598 (other platforms pending).
5600 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5602 * Add support for the SignedCertificateTimestampList certificate and
5603 OCSP response extensions from RFC6962.
5607 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5608 for corner cases. (Certain input points at infinity could lead to
5609 bogus results, with non-infinity inputs mapped to infinity too.)
5613 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5614 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5615 common cases are optimized and there still is room for further
5616 improvements. Vector Permutation AES for Altivec is also added.
5620 * Add support for little-endian ppc64 Linux target.
5622 *Marcelo Cerri (IBM)*
5624 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5625 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5626 are optimized and there still is room for further improvements.
5627 Both 32- and 64-bit modes are supported.
5629 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5631 * Improved ARMv7 NEON support.
5635 * Support for SPARC Architecture 2011 crypto extensions, first
5636 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5637 SHA256/512, MD5, GHASH and modular exponentiation.
5639 *Andy Polyakov, David Miller*
5641 * Accelerated modular exponentiation for Intel processors, a.k.a.
5644 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5646 * Support for new and upcoming Intel processors, including AVX2,
5647 BMI and SHA ISA extensions. This includes additional "stitched"
5648 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5651 This work was sponsored by Intel Corp.
5655 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5656 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5657 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5661 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5662 this fixes a limitation in previous versions of OpenSSL.
5666 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5667 MGF1 digest and OAEP label.
5671 * Add EVP support for key wrapping algorithms, to avoid problems with
5672 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5673 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5674 algorithms and include tests cases.
5678 * Add functions to allocate and set the fields of an ECDSA_METHOD
5681 *Douglas E. Engert, Steve Henson*
5683 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5684 difference in days and seconds between two tm or ASN1_TIME structures.
5688 * Add -rev test option to s_server to just reverse order of characters
5689 received by client and send back to server. Also prints an abbreviated
5690 summary of the connection parameters.
5694 * New option -brief for s_client and s_server to print out a brief summary
5695 of connection parameters.
5699 * Add callbacks for arbitrary TLS extensions.
5701 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5703 * New option -crl_download in several openssl utilities to download CRLs
5704 from CRLDP extension in certificates.
5708 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5712 * New function X509_CRL_diff to generate a delta CRL from the difference
5713 of two full CRLs. Add support to "crl" utility.
5717 * New functions to set lookup_crls function and to retrieve
5718 X509_STORE from X509_STORE_CTX.
5722 * Print out deprecated issuer and subject unique ID fields in
5727 * Extend OCSP I/O functions so they can be used for simple general purpose
5728 HTTP as well as OCSP. New wrapper function which can be used to download
5729 CRLs using the OCSP API.
5733 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5737 * `SSL_CONF*` functions. These provide a common framework for application
5738 configuration using configuration files or command lines.
5742 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5743 message callback and prints the results. Needs compile time option
5744 "enable-ssl-trace". New options to s_client and s_server to enable
5749 * New ctrl and macro to retrieve supported points extensions.
5750 Print out extension in s_server and s_client.
5754 * New functions to retrieve certificate signature and signature
5759 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5764 * New Suite B modes for TLS code. These use and enforce the requirements
5765 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5766 only use Suite B curves. The Suite B modes can be set by using the
5767 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5771 * New chain verification flags for Suite B levels of security. Check
5772 algorithms are acceptable when flags are set in X509_verify_cert.
5776 * Make tls1_check_chain return a set of flags indicating checks passed
5777 by a certificate chain. Add additional tests to handle client
5778 certificates: checks for matching certificate type and issuer name
5783 * If an attempt is made to use a signature algorithm not in the peer
5784 preference list abort the handshake. If client has no suitable
5785 signature algorithms in response to a certificate request do not
5786 use the certificate.
5790 * If server EC tmp key is not in client preference list abort handshake.
5794 * Add support for certificate stores in CERT structure. This makes it
5795 possible to have different stores per SSL structure or one store in
5796 the parent SSL_CTX. Include distinct stores for certificate chain
5797 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5798 to build and store a certificate chain in CERT structure: returning
5799 an error if the chain cannot be built: this will allow applications
5800 to test if a chain is correctly configured.
5802 Note: if the CERT based stores are not set then the parent SSL_CTX
5803 store is used to retain compatibility with existing behaviour.
5807 * New function ssl_set_client_disabled to set a ciphersuite disabled
5808 mask based on the current session, check mask when sending client
5809 hello and checking the requested ciphersuite.
5813 * New ctrls to retrieve and set certificate types in a certificate
5814 request message. Print out received values in s_client. If certificate
5815 types is not set with custom values set sensible values based on
5816 supported signature algorithms.
5820 * Support for distinct client and server supported signature algorithms.
5824 * Add certificate callback. If set this is called whenever a certificate
5825 is required by client or server. An application can decide which
5826 certificate chain to present based on arbitrary criteria: for example
5827 supported signature algorithms. Add very simple example to s_server.
5828 This fixes many of the problems and restrictions of the existing client
5829 certificate callback: for example you can now clear an existing
5830 certificate and specify the whole chain.
5834 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5835 the certificate can be used for (if anything). Set valid_flags field
5836 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5837 to have similar checks in it.
5839 Add new "cert_flags" field to CERT structure and include a "strict mode".
5840 This enforces some TLS certificate requirements (such as only permitting
5841 certificate signature algorithms contained in the supported algorithms
5842 extension) which some implementations ignore: this option should be used
5843 with caution as it could cause interoperability issues.
5847 * Update and tidy signature algorithm extension processing. Work out
5848 shared signature algorithms based on preferences and peer algorithms
5849 and print them out in s_client and s_server. Abort handshake if no
5850 shared signature algorithms.
5854 * Add new functions to allow customised supported signature algorithms
5855 for SSL and SSL_CTX structures. Add options to s_client and s_server
5860 * New function SSL_certs_clear() to delete all references to certificates
5861 from an SSL structure. Before this once a certificate had been added
5862 it couldn't be removed.
5866 * Integrate hostname, email address and IP address checking with certificate
5867 verification. New verify options supporting checking in openssl utility.
5871 * Fixes and wildcard matching support to hostname and email checking
5872 functions. Add manual page.
5874 *Florian Weimer (Red Hat Product Security Team)*
5876 * New functions to check a hostname email or IP address against a
5877 certificate. Add options x509 utility to print results of checks against
5882 * Fix OCSP checking.
5884 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5886 * Initial experimental support for explicitly trusted non-root CAs.
5887 OpenSSL still tries to build a complete chain to a root but if an
5888 intermediate CA has a trust setting included that is used. The first
5889 setting is used: whether to trust (e.g., -addtrust option to the x509
5894 * Add -trusted_first option which attempts to find certificates in the
5895 trusted store even if an untrusted chain is also supplied.
5899 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5900 platform support for Linux and Android.
5904 * Support for linux-x32, ILP32 environment in x86_64 framework.
5908 * Experimental multi-implementation support for FIPS capable OpenSSL.
5909 When in FIPS mode the approved implementations are used as normal,
5910 when not in FIPS mode the internal unapproved versions are used instead.
5911 This means that the FIPS capable OpenSSL isn't forced to use the
5912 (often lower performance) FIPS implementations outside FIPS mode.
5916 * Transparently support X9.42 DH parameters when calling
5917 PEM_read_bio_DHparameters. This means existing applications can handle
5918 the new parameter format automatically.
5922 * Initial experimental support for X9.42 DH parameter format: mainly
5923 to support use of 'q' parameter for RFC5114 parameters.
5927 * Add DH parameters from RFC5114 including test data to dhtest.
5931 * Support for automatic EC temporary key parameter selection. If enabled
5932 the most preferred EC parameters are automatically used instead of
5933 hardcoded fixed parameters. Now a server just has to call:
5934 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5935 support ECDH and use the most appropriate parameters.
5939 * Enhance and tidy EC curve and point format TLS extension code. Use
5940 static structures instead of allocation if default values are used.
5941 New ctrls to set curves we wish to support and to retrieve shared curves.
5942 Print out shared curves in s_server. New options to s_server and s_client
5943 to set list of supported curves.
5947 * New ctrls to retrieve supported signature algorithms and
5948 supported curve values as an array of NIDs. Extend openssl utility
5949 to print out received values.
5953 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5954 between NIDs and the more common NIST names such as "P-256". Enhance
5955 ecparam utility and ECC method to recognise the NIST names for curves.
5959 * Enhance SSL/TLS certificate chain handling to support different
5960 chains for each certificate instead of one chain in the parent SSL_CTX.
5964 * Support for fixed DH ciphersuite client authentication: where both
5965 server and client use DH certificates with common parameters.
5969 * Support for fixed DH ciphersuites: those requiring DH server
5974 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5976 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5977 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5978 X509_CINF_get_signature were reverted post internal team review.
5983 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5985 * OCSP Status Request extension unbounded memory growth
5987 A malicious client can send an excessively large OCSP Status Request
5988 extension. If that client continually requests renegotiation, sending a
5989 large OCSP Status Request extension each time, then there will be unbounded
5990 memory growth on the server. This will eventually lead to a Denial Of
5991 Service attack through memory exhaustion. Servers with a default
5992 configuration are vulnerable even if they do not support OCSP. Builds using
5993 the "no-ocsp" build time option are not affected.
5995 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6000 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6003 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6009 * OOB write in MDC2_Update()
6011 An overflow can occur in MDC2_Update() either if called directly or
6012 through the EVP_DigestUpdate() function using MDC2. If an attacker
6013 is able to supply very large amounts of input data after a previous
6014 call to EVP_EncryptUpdate() with a partial block then a length check
6015 can overflow resulting in a heap corruption.
6017 The amount of data needed is comparable to SIZE_MAX which is impractical
6020 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6025 * Malformed SHA512 ticket DoS
6027 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6028 DoS attack where a malformed ticket will result in an OOB read which will
6031 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6032 a custom server callback and ticket lookup mechanism.
6034 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6039 * OOB write in BN_bn2dec()
6041 The function BN_bn2dec() does not check the return value of BN_div_word().
6042 This can cause an OOB write if an application uses this function with an
6043 overly large BIGNUM. This could be a problem if an overly large certificate
6044 or CRL is printed out from an untrusted source. TLS is not affected because
6045 record limits will reject an oversized certificate before it is parsed.
6047 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6052 * OOB read in TS_OBJ_print_bio()
6054 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6055 the total length the OID text representation would use and not the amount
6056 of data written. This will result in OOB reads when large OIDs are
6059 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6064 * Pointer arithmetic undefined behaviour
6066 Avoid some undefined pointer arithmetic
6068 A common idiom in the codebase is to check limits in the following manner:
6071 Where "p" points to some malloc'd data of SIZE bytes and
6074 "len" here could be from some externally supplied data (e.g. from a TLS
6077 The rules of C pointer arithmetic are such that "p + len" is only well
6078 defined where len <= SIZE. Therefore the above idiom is actually
6079 undefined behaviour.
6081 For example this could cause problems if some malloc implementation
6082 provides an address for "p" such that "p + len" actually overflows for
6083 values of len that are too big and therefore p + len < limit.
6085 This issue was reported to OpenSSL by Guido Vranken
6090 * Constant time flag not preserved in DSA signing
6092 Operations in the DSA signing algorithm should run in constant time in
6093 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6094 implementation means that a non-constant time codepath is followed for
6095 certain operations. This has been demonstrated through a cache-timing
6096 attack to be sufficient for an attacker to recover the private DSA key.
6098 This issue was reported by César Pereida (Aalto University), Billy Brumley
6099 (Tampere University of Technology), and Yuval Yarom (The University of
6100 Adelaide and NICTA).
6105 * DTLS buffered message DoS
6107 In a DTLS connection where handshake messages are delivered out-of-order
6108 those messages that OpenSSL is not yet ready to process will be buffered
6109 for later use. Under certain circumstances, a flaw in the logic means that
6110 those messages do not get removed from the buffer even though the handshake
6111 has been completed. An attacker could force up to approx. 15 messages to
6112 remain in the buffer when they are no longer required. These messages will
6113 be cleared when the DTLS connection is closed. The default maximum size for
6114 a message is 100k. Therefore the attacker could force an additional 1500k
6115 to be consumed per connection. By opening many simulataneous connections an
6116 attacker could cause a DoS attack through memory exhaustion.
6118 This issue was reported to OpenSSL by Quan Luo.
6123 * DTLS replay protection DoS
6125 A flaw in the DTLS replay attack protection mechanism means that records
6126 that arrive for future epochs update the replay protection "window" before
6127 the MAC for the record has been validated. This could be exploited by an
6128 attacker by sending a record for the next epoch (which does not have to
6129 decrypt or have a valid MAC), with a very large sequence number. This means
6130 that all subsequent legitimate packets are dropped causing a denial of
6131 service for a specific DTLS connection.
6133 This issue was reported to OpenSSL by the OCAP audit team.
6138 * Certificate message OOB reads
6140 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6141 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6142 theoretical DoS risk but this has not been observed in practice on common
6145 The messages affected are client certificate, client certificate request
6146 and server certificate. As a result the attack can only be performed
6147 against a client or a server which enables client authentication.
6149 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6154 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6156 * Prevent padding oracle in AES-NI CBC MAC check
6158 A MITM attacker can use a padding oracle attack to decrypt traffic
6159 when the connection uses an AES CBC cipher and the server support
6162 This issue was introduced as part of the fix for Lucky 13 padding
6163 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6164 constant time by making sure that always the same bytes are read and
6165 compared against either the MAC or padding bytes. But it no longer
6166 checked that there was enough data to have both the MAC and padding
6169 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6174 * Fix EVP_EncodeUpdate overflow
6176 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6177 Base64 encoding of binary data. If an attacker is able to supply very large
6178 amounts of input data then a length check can overflow resulting in a heap
6181 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6182 the `PEM_write_bio*` family of functions. These are mainly used within the
6183 OpenSSL command line applications, so any application which processes data
6184 from an untrusted source and outputs it as a PEM file should be considered
6185 vulnerable to this issue. User applications that call these APIs directly
6186 with large amounts of untrusted data may also be vulnerable.
6188 This issue was reported by Guido Vranken.
6193 * Fix EVP_EncryptUpdate overflow
6195 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6196 is able to supply very large amounts of input data after a previous call to
6197 EVP_EncryptUpdate() with a partial block then a length check can overflow
6198 resulting in a heap corruption. Following an analysis of all OpenSSL
6199 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6200 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6201 the first called function after an EVP_EncryptInit(), and therefore that
6202 specific call must be safe. The second form is where the length passed to
6203 EVP_EncryptUpdate() can be seen from the code to be some small value and
6204 therefore there is no possibility of an overflow. Since all instances are
6205 one of these two forms, it is believed that there can be no overflows in
6206 internal code due to this problem. It should be noted that
6207 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6208 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6209 of these calls have also been analysed too and it is believed there are no
6210 instances in internal usage where an overflow could occur.
6212 This issue was reported by Guido Vranken.
6217 * Prevent ASN.1 BIO excessive memory allocation
6219 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6220 a short invalid encoding can casuse allocation of large amounts of memory
6221 potentially consuming excessive resources or exhausting memory.
6223 Any application parsing untrusted data through d2i BIO functions is
6224 affected. The memory based functions such as d2i_X509() are *not* affected.
6225 Since the memory based functions are used by the TLS library, TLS
6226 applications are not affected.
6228 This issue was reported by Brian Carpenter.
6235 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6236 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6237 in arbitrary stack data being returned in the buffer.
6239 This issue was reported by Guido Vranken.
6244 * Modify behavior of ALPN to invoke callback after SNI/servername
6245 callback, such that updates to the SSL_CTX affect ALPN.
6249 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6254 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6255 methods are enabled and ssl2 is disabled the methods return NULL.
6259 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6261 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6262 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6263 provide any "EXPORT" or "LOW" strength ciphers.
6267 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6268 is by default disabled at build-time. Builds that are not configured with
6269 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6270 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6271 will need to explicitly call either of:
6273 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6275 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6277 as appropriate. Even if either of those is used, or the application
6278 explicitly uses the version-specific SSLv2_method() or its client and
6279 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6280 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6281 ciphers, and SSLv2 56-bit DES are no longer available.
6286 * Fix a double-free in DSA code
6288 A double free bug was discovered when OpenSSL parses malformed DSA private
6289 keys and could lead to a DoS attack or memory corruption for applications
6290 that receive DSA private keys from untrusted sources. This scenario is
6293 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6299 * Disable SRP fake user seed to address a server memory leak.
6301 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6303 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6304 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6305 was changed to ignore the "fake user" SRP seed, even if the seed
6308 Users should use SRP_VBASE_get1_by_user instead. Note that in
6309 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6310 also that even though configuring the SRP seed attempts to hide
6311 invalid usernames by continuing the handshake with fake
6312 credentials, this behaviour is not constant time and no strong
6313 guarantees are made that the handshake is indistinguishable from
6314 that of a valid user.
6319 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6321 In the BN_hex2bn function the number of hex digits is calculated using an
6322 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6323 large values of `i` this can result in `bn_expand` not allocating any
6324 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6325 field as NULL leading to a subsequent NULL ptr deref. For very large values
6326 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6327 In this case memory is allocated to the internal BIGNUM data field, but it
6328 is insufficiently sized leading to heap corruption. A similar issue exists
6329 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6330 is ever called by user applications with very large untrusted hex/dec data.
6331 This is anticipated to be a rare occurrence.
6333 All OpenSSL internal usage of these functions use data that is not expected
6334 to be untrusted, e.g. config file data or application command line
6335 arguments. If user developed applications generate config file data based
6336 on untrusted data then it is possible that this could also lead to security
6337 consequences. This is also anticipated to be rare.
6339 This issue was reported to OpenSSL by Guido Vranken.
6344 * Fix memory issues in `BIO_*printf` functions
6346 The internal `fmtstr` function used in processing a "%s" format string in
6347 the `BIO_*printf` functions could overflow while calculating the length of a
6348 string and cause an OOB read when printing very long strings.
6350 Additionally the internal `doapr_outch` function can attempt to write to an
6351 OOB memory location (at an offset from the NULL pointer) in the event of a
6352 memory allocation failure. In 1.0.2 and below this could be caused where
6353 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6354 could be in processing a very long "%s" format string. Memory leaks can
6357 The first issue may mask the second issue dependent on compiler behaviour.
6358 These problems could enable attacks where large amounts of untrusted data
6359 is passed to the `BIO_*printf` functions. If applications use these functions
6360 in this way then they could be vulnerable. OpenSSL itself uses these
6361 functions when printing out human-readable dumps of ASN.1 data. Therefore
6362 applications that print this data could be vulnerable if the data is from
6363 untrusted sources. OpenSSL command line applications could also be
6364 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6365 as command line arguments.
6367 Libssl is not considered directly vulnerable. Additionally certificates etc
6368 received via remote connections via libssl are also unlikely to be able to
6369 trigger these issues because of message size limits enforced within libssl.
6371 This issue was reported to OpenSSL Guido Vranken.
6376 * Side channel attack on modular exponentiation
6378 A side-channel attack was found which makes use of cache-bank conflicts on
6379 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6380 of RSA keys. The ability to exploit this issue is limited as it relies on
6381 an attacker who has control of code in a thread running on the same
6382 hyper-threaded core as the victim thread which is performing decryptions.
6384 This issue was reported to OpenSSL by Yuval Yarom, The University of
6385 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6386 Nadia Heninger, University of Pennsylvania with more information at
6387 <http://cachebleed.info>.
6392 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6393 if no keysize is specified with default_bits. This fixes an
6394 omission in an earlier change that changed all RSA/DSA key generation
6395 commands to use 2048 bits by default.
6399 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6401 * Protection for DH small subgroup attacks
6403 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6404 switched on by default and cannot be disabled. This could have some
6409 * SSLv2 doesn't block disabled ciphers
6411 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6412 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6413 been disabled, provided that the SSLv2 protocol was not also disabled via
6416 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6417 and Sebastian Schinzel.
6422 * Reject DH handshakes with parameters shorter than 1024 bits.
6426 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6428 * Certificate verify crash with missing PSS parameter
6430 The signature verification routines will crash with a NULL pointer
6431 dereference if presented with an ASN.1 signature using the RSA PSS
6432 algorithm and absent mask generation function parameter. Since these
6433 routines are used to verify certificate signature algorithms this can be
6434 used to crash any certificate verification operation and exploited in a
6435 DoS attack. Any application which performs certificate verification is
6436 vulnerable including OpenSSL clients and servers which enable client
6439 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6444 * X509_ATTRIBUTE memory leak
6446 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6447 memory. This structure is used by the PKCS#7 and CMS routines so any
6448 application which reads PKCS#7 or CMS data from untrusted sources is
6449 affected. SSL/TLS is not affected.
6451 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6457 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6458 This changes the decoding behaviour for some invalid messages,
6459 though the change is mostly in the more lenient direction, and
6460 legacy behaviour is preserved as much as possible.
6464 * In DSA_generate_parameters_ex, if the provided seed is too short,
6465 use a random seed, as already documented.
6467 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6469 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6471 * Alternate chains certificate forgery
6473 During certificate verfification, OpenSSL will attempt to find an
6474 alternative certificate chain if the first attempt to build such a chain
6475 fails. An error in the implementation of this logic can mean that an
6476 attacker could cause certain checks on untrusted certificates to be
6477 bypassed, such as the CA flag, enabling them to use a valid leaf
6478 certificate to act as a CA and "issue" an invalid certificate.
6480 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6486 * Race condition handling PSK identify hint
6488 If PSK identity hints are received by a multi-threaded client then
6489 the values are wrongly updated in the parent SSL_CTX structure. This can
6490 result in a race condition potentially leading to a double free of the
6496 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6498 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6499 incompatibility in the handling of HMAC. The previous ABI has now been
6502 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6504 * Malformed ECParameters causes infinite loop
6506 When processing an ECParameters structure OpenSSL enters an infinite loop
6507 if the curve specified is over a specially malformed binary polynomial
6510 This can be used to perform denial of service against any
6511 system which processes public keys, certificate requests or
6512 certificates. This includes TLS clients and TLS servers with
6513 client authentication enabled.
6515 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6520 * Exploitable out-of-bounds read in X509_cmp_time
6522 X509_cmp_time does not properly check the length of the ASN1_TIME
6523 string and can read a few bytes out of bounds. In addition,
6524 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6527 An attacker can use this to craft malformed certificates and CRLs of
6528 various sizes and potentially cause a segmentation fault, resulting in
6529 a DoS on applications that verify certificates or CRLs. TLS clients
6530 that verify CRLs are affected. TLS clients and servers with client
6531 authentication enabled may be affected if they use custom verification
6534 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6535 independently by Hanno Böck.
6540 * PKCS7 crash with missing EnvelopedContent
6542 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6543 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6544 with missing content and trigger a NULL pointer dereference on parsing.
6546 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6547 structures from untrusted sources are affected. OpenSSL clients and
6548 servers are not affected.
6550 This issue was reported to OpenSSL by Michal Zalewski (Google).
6555 * CMS verify infinite loop with unknown hash function
6557 When verifying a signedData message the CMS code can enter an infinite loop
6558 if presented with an unknown hash function OID. This can be used to perform
6559 denial of service against any system which verifies signedData messages using
6561 This issue was reported to OpenSSL by Johannes Bauer.
6566 * Race condition handling NewSessionTicket
6568 If a NewSessionTicket is received by a multi-threaded client when attempting to
6569 reuse a previous ticket then a race condition can occur potentially leading to
6570 a double free of the ticket data.
6575 * Reject DH handshakes with parameters shorter than 768 bits.
6577 *Kurt Roeckx and Emilia Kasper*
6579 * dhparam: generate 2048-bit parameters by default.
6581 *Kurt Roeckx and Emilia Kasper*
6583 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6585 * Segmentation fault in ASN1_TYPE_cmp fix
6587 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6588 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6589 certificate signature algorithm consistency this can be used to crash any
6590 certificate verification operation and exploited in a DoS attack. Any
6591 application which performs certificate verification is vulnerable including
6592 OpenSSL clients and servers which enable client authentication.
6597 * ASN.1 structure reuse memory corruption fix
6599 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6600 memory corruption via an invalid write. Such reuse is and has been
6601 strongly discouraged and is believed to be rare.
6603 Applications that parse structures containing CHOICE or ANY DEFINED BY
6604 components may be affected. Certificate parsing (d2i_X509 and related
6605 functions) are however not affected. OpenSSL clients and servers are
6611 * PKCS7 NULL pointer dereferences fix
6613 The PKCS#7 parsing code does not handle missing outer ContentInfo
6614 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6615 missing content and trigger a NULL pointer dereference on parsing.
6617 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6618 otherwise parse PKCS#7 structures from untrusted sources are
6619 affected. OpenSSL clients and servers are not affected.
6621 This issue was reported to OpenSSL by Michal Zalewski (Google).
6626 * DoS via reachable assert in SSLv2 servers fix
6628 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6629 servers that both support SSLv2 and enable export cipher suites by sending
6630 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6632 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6633 (OpenSSL development team).
6638 * Use After Free following d2i_ECPrivatekey error fix
6640 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6641 could cause a use after free condition. This, in turn, could cause a double
6642 free in several private key parsing functions (such as d2i_PrivateKey
6643 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6644 for applications that receive EC private keys from untrusted
6645 sources. This scenario is considered rare.
6647 This issue was discovered by the BoringSSL project and fixed in their
6653 * X509_to_X509_REQ NULL pointer deref fix
6655 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6656 the certificate key is invalid. This function is rarely used in practice.
6658 This issue was discovered by Brian Carpenter.
6663 * Removed the export ciphers from the DEFAULT ciphers
6667 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6669 * Build fixes for the Windows and OpenVMS platforms
6671 *Matt Caswell and Richard Levitte*
6673 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6675 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6676 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6677 dereference. This could lead to a Denial Of Service attack. Thanks to
6678 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6683 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6684 dtls1_buffer_record function under certain conditions. In particular this
6685 could occur if an attacker sent repeated DTLS records with the same
6686 sequence number but for the next epoch. The memory leak could be exploited
6687 by an attacker in a Denial of Service attack through memory exhaustion.
6688 Thanks to Chris Mueller for reporting this issue.
6693 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6694 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6695 method would be set to NULL which could later result in a NULL pointer
6696 dereference. Thanks to Frank Schmirler for reporting this issue.
6701 * Abort handshake if server key exchange message is omitted for ephemeral
6704 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6705 reporting this issue.
6710 * Remove non-export ephemeral RSA code on client and server. This code
6711 violated the TLS standard by allowing the use of temporary RSA keys in
6712 non-export ciphersuites and could be used by a server to effectively
6713 downgrade the RSA key length used to a value smaller than the server
6714 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6715 INRIA or reporting this issue.
6720 * Fixed issue where DH client certificates are accepted without verification.
6721 An OpenSSL server will accept a DH certificate for client authentication
6722 without the certificate verify message. This effectively allows a client to
6723 authenticate without the use of a private key. This only affects servers
6724 which trust a client certificate authority which issues certificates
6725 containing DH keys: these are extremely rare and hardly ever encountered.
6726 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6732 * Ensure that the session ID context of an SSL is updated when its
6733 SSL_CTX is updated via SSL_set_SSL_CTX.
6735 The session ID context is typically set from the parent SSL_CTX,
6736 and can vary with the CTX.
6740 * Fix various certificate fingerprint issues.
6742 By using non-DER or invalid encodings outside the signed portion of a
6743 certificate the fingerprint can be changed without breaking the signature.
6744 Although no details of the signed portion of the certificate can be changed
6745 this can cause problems with some applications: e.g. those using the
6746 certificate fingerprint for blacklists.
6748 1. Reject signatures with non zero unused bits.
6750 If the BIT STRING containing the signature has non zero unused bits reject
6751 the signature. All current signature algorithms require zero unused bits.
6753 2. Check certificate algorithm consistency.
6755 Check the AlgorithmIdentifier inside TBS matches the one in the
6756 certificate signature. NB: this will result in signature failure
6757 errors for some broken certificates.
6759 Thanks to Konrad Kraszewski from Google for reporting this issue.
6761 3. Check DSA/ECDSA signatures use DER.
6763 Re-encode DSA/ECDSA signatures and compare with the original received
6764 signature. Return an error if there is a mismatch.
6766 This will reject various cases including garbage after signature
6767 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6768 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6769 (negative or with leading zeroes).
6771 Further analysis was conducted and fixes were developed by Stephen Henson
6772 of the OpenSSL core team.
6778 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6779 results on some platforms, including x86_64. This bug occurs at random
6780 with a very low probability, and is not known to be exploitable in any
6781 way, though its exact impact is difficult to determine. Thanks to Pieter
6782 Wuille (Blockstream) who reported this issue and also suggested an initial
6783 fix. Further analysis was conducted by the OpenSSL development team and
6784 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6785 the OpenSSL core team.
6790 * Do not resume sessions on the server if the negotiated protocol
6791 version does not match the session's version. Resuming with a different
6792 version, while not strictly forbidden by the RFC, is of questionable
6793 sanity and breaks all known clients.
6795 *David Benjamin, Emilia Käsper*
6797 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6798 early CCS messages during renegotiation. (Note that because
6799 renegotiation is encrypted, this early CCS was not exploitable.)
6803 * Tighten client-side session ticket handling during renegotiation:
6804 ensure that the client only accepts a session ticket if the server sends
6805 the extension anew in the ServerHello. Previously, a TLS client would
6806 reuse the old extension state and thus accept a session ticket if one was
6807 announced in the initial ServerHello.
6809 Similarly, ensure that the client requires a session ticket if one
6810 was advertised in the ServerHello. Previously, a TLS client would
6811 ignore a missing NewSessionTicket message.
6815 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6819 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6820 sends a carefully crafted handshake message, to cause OpenSSL to fail
6821 to free up to 64k of memory causing a memory leak. This could be
6822 exploited in a Denial Of Service attack. This issue affects OpenSSL
6823 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6824 whether SRTP is used or configured. Implementations of OpenSSL that
6825 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6827 The fix was developed by the OpenSSL team.
6832 * Session Ticket Memory Leak.
6834 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6835 integrity of that ticket is first verified. In the event of a session
6836 ticket integrity check failing, OpenSSL will fail to free memory
6837 causing a memory leak. By sending a large number of invalid session
6838 tickets an attacker could exploit this issue in a Denial Of Service
6844 * Build option no-ssl3 is incomplete.
6846 When OpenSSL is configured with "no-ssl3" as a build option, servers
6847 could accept and complete a SSL 3.0 handshake, and clients could be
6848 configured to send them.
6851 *Akamai and the OpenSSL team*
6853 * Add support for TLS_FALLBACK_SCSV.
6854 Client applications doing fallback retries should call
6855 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6858 *Adam Langley, Bodo Moeller*
6860 * Add additional DigestInfo checks.
6862 Re-encode DigestInto in DER and check against the original when
6863 verifying RSA signature: this will reject any improperly encoded
6864 DigestInfo structures.
6866 Note: this is a precautionary measure and no attacks are currently known.
6870 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6872 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6873 SRP code can be overrun an internal buffer. Add sanity check that
6874 g, A, B < N to SRP code.
6876 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6877 Group for discovering this issue.
6882 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6883 TLS 1.0 instead of higher protocol versions when the ClientHello message
6884 is badly fragmented. This allows a man-in-the-middle attacker to force a
6885 downgrade to TLS 1.0 even if both the server and the client support a
6886 higher protocol version, by modifying the client's TLS records.
6888 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6889 researching this issue.
6894 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6895 to a denial of service attack. A malicious server can crash the client
6896 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6897 ciphersuite and sending carefully crafted handshake messages.
6899 Thanks to Felix Gröbert (Google) for discovering and researching this
6905 * By sending carefully crafted DTLS packets an attacker could cause openssl
6906 to leak memory. This can be exploited through a Denial of Service attack.
6907 Thanks to Adam Langley for discovering and researching this issue.
6912 * An attacker can force openssl to consume large amounts of memory whilst
6913 processing DTLS handshake messages. This can be exploited through a
6914 Denial of Service attack.
6915 Thanks to Adam Langley for discovering and researching this issue.
6920 * An attacker can force an error condition which causes openssl to crash
6921 whilst processing DTLS packets due to memory being freed twice. This
6922 can be exploited through a Denial of Service attack.
6923 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6929 * If a multithreaded client connects to a malicious server using a resumed
6930 session and the server sends an ec point format extension it could write
6931 up to 255 bytes to freed memory.
6933 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6939 * A malicious server can crash an OpenSSL client with a null pointer
6940 dereference (read) by specifying an SRP ciphersuite even though it was not
6941 properly negotiated with the client. This can be exploited through a
6942 Denial of Service attack.
6944 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6945 discovering and researching this issue.
6950 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6951 X509_name_oneline, X509_name_print_ex et al. to leak some information
6952 from the stack. Applications may be affected if they echo pretty printing
6953 output to the attacker.
6955 Thanks to Ivan Fratric (Google) for discovering this issue.
6958 *Emilia Käsper, and Steve Henson*
6960 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6961 for corner cases. (Certain input points at infinity could lead to
6962 bogus results, with non-infinity inputs mapped to infinity too.)
6966 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6968 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6969 handshake can force the use of weak keying material in OpenSSL
6970 SSL/TLS clients and servers.
6972 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6973 researching this issue. ([CVE-2014-0224])
6975 *KIKUCHI Masashi, Steve Henson*
6977 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6978 OpenSSL DTLS client the code can be made to recurse eventually crashing
6981 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6984 *Imre Rad, Steve Henson*
6986 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6987 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6988 client or server. This is potentially exploitable to run arbitrary
6989 code on a vulnerable client or server.
6991 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6993 *Jüri Aedla, Steve Henson*
6995 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6996 are subject to a denial of service attack.
6998 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6999 this issue. ([CVE-2014-3470])
7001 *Felix Gröbert, Ivan Fratric, Steve Henson*
7003 * Harmonize version and its documentation. -f flag is used to display
7006 *mancha <mancha1@zoho.com>*
7008 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7009 in i2d_ECPrivateKey.
7011 *mancha <mancha1@zoho.com>*
7013 * Fix some double frees. These are not thought to be exploitable.
7015 *mancha <mancha1@zoho.com>*
7017 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7019 * A missing bounds check in the handling of the TLS heartbeat extension
7020 can be used to reveal up to 64k of memory to a connected client or
7023 Thanks for Neel Mehta of Google Security for discovering this bug and to
7024 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7025 preparing the fix ([CVE-2014-0160])
7027 *Adam Langley, Bodo Moeller*
7029 * Fix for the attack described in the paper "Recovering OpenSSL
7030 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7031 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7032 <http://eprint.iacr.org/2014/140>
7034 Thanks to Yuval Yarom and Naomi Benger for discovering this
7035 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7037 *Yuval Yarom and Naomi Benger*
7039 * TLS pad extension: draft-agl-tls-padding-03
7041 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7042 TLS client Hello record length value would otherwise be > 255 and
7043 less that 512 pad with a dummy extension containing zeroes so it
7044 is at least 512 bytes long.
7046 *Adam Langley, Steve Henson*
7048 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7050 * Fix for TLS record tampering bug. A carefully crafted invalid
7051 handshake could crash OpenSSL with a NULL pointer exception.
7052 Thanks to Anton Johansson for reporting this issues.
7055 * Keep original DTLS digest and encryption contexts in retransmission
7056 structures so we can use the previous session parameters if they need
7057 to be resent. ([CVE-2013-6450])
7061 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7062 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7063 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7064 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7065 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7066 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7068 *Rob Stradling, Adam Langley*
7070 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7072 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7073 supporting platforms or when small records were transferred.
7075 *Andy Polyakov, Steve Henson*
7077 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7079 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7081 This addresses the flaw in CBC record processing discovered by
7082 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7083 at: <http://www.isg.rhul.ac.uk/tls/>
7085 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7086 Security Group at Royal Holloway, University of London
7087 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7088 Emilia Käsper for the initial patch.
7091 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7093 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7094 ciphersuites which can be exploited in a denial of service attack.
7095 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7096 and detecting this bug and to Wolfgang Ettlinger
7097 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7102 * Return an error when checking OCSP signatures when key is NULL.
7103 This fixes a DoS attack. ([CVE-2013-0166])
7107 * Make openssl verify return errors.
7109 *Chris Palmer <palmer@google.com> and Ben Laurie*
7111 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7112 the right response is stapled. Also change SSL_get_certificate()
7113 so it returns the certificate actually sent.
7114 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7116 *Rob Stradling <rob.stradling@comodo.com>*
7118 * Fix possible deadlock when decoding public keys.
7122 * Don't use TLS 1.0 record version number in initial client hello
7127 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7129 * Sanity check record length before skipping explicit IV in TLS
7130 1.2, 1.1 and DTLS to fix DoS attack.
7132 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7133 fuzzing as a service testing platform.
7138 * Initialise tkeylen properly when encrypting CMS messages.
7139 Thanks to Solar Designer of Openwall for reporting this issue.
7143 * In FIPS mode don't try to use composite ciphers as they are not
7148 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7150 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7151 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7152 mean any application compiled against OpenSSL 1.0.0 headers setting
7153 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7154 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7155 0x10000000L Any application which was previously compiled against
7156 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7157 will need to be recompiled as a result. Letting be results in
7158 inability to disable specifically TLS 1.1 and in client context,
7159 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7163 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7164 disable just protocol X, but all protocols above X *if* there are
7165 protocols *below* X still enabled. In more practical terms it means
7166 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7167 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7168 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7173 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7175 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7176 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7177 in CRYPTO_realloc_clean.
7179 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7180 issue and to Adam Langley <agl@chromium.org> for fixing it.
7183 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7185 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7189 * Workarounds for some broken servers that "hang" if a client hello
7190 record length exceeds 255 bytes.
7192 1. Do not use record version number > TLS 1.0 in initial client
7193 hello: some (but not all) hanging servers will now work.
7194 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7195 the number of ciphers sent in the client hello. This should be
7196 set to an even number, such as 50, for example by passing:
7197 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7198 Most broken servers should now work.
7199 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7200 TLS 1.2 client support entirely.
7204 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7208 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7210 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7211 STRING form instead of a DigestInfo.
7215 * The format used for MDC2 RSA signatures is inconsistent between EVP
7216 and the RSA_sign/RSA_verify functions. This was made more apparent when
7217 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7218 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7219 the correct format in RSA_verify so both forms transparently work.
7223 * Some servers which support TLS 1.0 can choke if we initially indicate
7224 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7225 encrypted premaster secret. As a workaround use the maximum permitted
7226 client version in client hello, this should keep such servers happy
7227 and still work with previous versions of OpenSSL.
7231 * Add support for TLS/DTLS heartbeats.
7233 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7235 * Add support for SCTP.
7237 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7239 * Improved PRNG seeding for VOS.
7241 *Paul Green <Paul.Green@stratus.com>*
7243 * Extensive assembler packs updates, most notably:
7245 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7246 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7247 - x86_64: bit-sliced AES implementation;
7248 - ARM: NEON support, contemporary platforms optimizations;
7249 - s390x: z196 support;
7250 - `*`: GHASH and GF(2^m) multiplication implementations;
7254 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7255 (removal of unnecessary code)
7257 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7259 * Add TLS key material exporter from RFC 5705.
7263 * Add DTLS-SRTP negotiation from RFC 5764.
7267 * Add Next Protocol Negotiation,
7268 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7269 disabled with a no-npn flag to config or Configure. Code donated
7272 *Adam Langley <agl@google.com> and Ben Laurie*
7274 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7275 NIST-P256, NIST-P521, with constant-time single point multiplication on
7276 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7277 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7278 Code made available under Apache License version 2.0.
7280 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7281 line to include this in your build of OpenSSL, and run "make depend" (or
7282 "make update"). This enables the following EC_METHODs:
7284 EC_GFp_nistp224_method()
7285 EC_GFp_nistp256_method()
7286 EC_GFp_nistp521_method()
7288 EC_GROUP_new_by_curve_name() will automatically use these (while
7289 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7292 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7294 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7295 all platforms. Move ssize_t definition from e_os.h to the public
7296 header file e_os2.h as it now appears in public header file cms.h
7300 * New -sigopt option to the ca, req and x509 utilities. Additional
7301 signature parameters can be passed using this option and in
7306 * Add RSA PSS signing function. This will generate and set the
7307 appropriate AlgorithmIdentifiers for PSS based on those in the
7308 corresponding EVP_MD_CTX structure. No application support yet.
7312 * Support for companion algorithm specific ASN1 signing routines.
7313 New function ASN1_item_sign_ctx() signs a pre-initialised
7314 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7315 the appropriate parameters.
7319 * Add new algorithm specific ASN1 verification initialisation function
7320 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7321 handling will be the same no matter what EVP_PKEY_METHOD is used.
7322 Add a PSS handler to support verification of PSS signatures: checked
7323 against a number of sample certificates.
7327 * Add signature printing for PSS. Add PSS OIDs.
7329 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7331 * Add algorithm specific signature printing. An individual ASN1 method
7332 can now print out signatures instead of the standard hex dump.
7334 More complex signatures (e.g. PSS) can print out more meaningful
7335 information. Include DSA version that prints out the signature
7340 * Password based recipient info support for CMS library: implementing
7345 * Split password based encryption into PBES2 and PBKDF2 functions. This
7346 neatly separates the code into cipher and PBE sections and is required
7347 for some algorithms that split PBES2 into separate pieces (such as
7348 password based CMS).
7352 * Session-handling fixes:
7353 - Fix handling of connections that are resuming with a session ID,
7354 but also support Session Tickets.
7355 - Fix a bug that suppressed issuing of a new ticket if the client
7356 presented a ticket with an expired session.
7357 - Try to set the ticket lifetime hint to something reasonable.
7358 - Make tickets shorter by excluding irrelevant information.
7359 - On the client side, don't ignore renewed tickets.
7361 *Adam Langley, Bodo Moeller (Google)*
7363 * Fix PSK session representation.
7367 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7369 This work was sponsored by Intel.
7373 * Add GCM support to TLS library. Some custom code is needed to split
7374 the IV between the fixed (from PRF) and explicit (from TLS record)
7375 portions. This adds all GCM ciphersuites supported by RFC5288 and
7376 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7377 add a special AESGCM string for GCM only.
7381 * Expand range of ctrls for AES GCM. Permit setting invocation
7382 field on decrypt and retrieval of invocation field only on encrypt.
7386 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7387 As required by RFC5289 these ciphersuites cannot be used if for
7388 versions of TLS earlier than 1.2.
7392 * For FIPS capable OpenSSL interpret a NULL default public key method
7393 as unset and return the appropriate default but do *not* set the default.
7394 This means we can return the appropriate method in applications that
7395 switch between FIPS and non-FIPS modes.
7399 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7400 ENGINE is used then we cannot handle that in the FIPS module so we
7401 keep original code iff non-FIPS operations are allowed.
7405 * Add -attime option to openssl utilities.
7407 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7409 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7413 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7414 FIPS EC methods unconditionally for now.
7418 * New build option no-ec2m to disable characteristic 2 code.
7422 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7423 all cases can be covered as some introduce binary incompatibilities.
7427 * Redirect RSA operations to FIPS module including keygen,
7428 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7432 * Add similar low level API blocking to ciphers.
7436 * Low level digest APIs are not approved in FIPS mode: any attempt
7437 to use these will cause a fatal error. Applications that *really* want
7438 to use them can use the `private_*` version instead.
7442 * Redirect cipher operations to FIPS module for FIPS builds.
7446 * Redirect digest operations to FIPS module for FIPS builds.
7450 * Update build system to add "fips" flag which will link in fipscanister.o
7451 for static and shared library builds embedding a signature if needed.
7455 * Output TLS supported curves in preference order instead of numerical
7456 order. This is currently hardcoded for the highest order curves first.
7457 This should be configurable so applications can judge speed vs strength.
7461 * Add TLS v1.2 server support for client authentication.
7465 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7470 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7471 FIPS modules versions.
7475 * Add TLS v1.2 client side support for client authentication. Keep cache
7476 of handshake records longer as we don't know the hash algorithm to use
7477 until after the certificate request message is received.
7481 * Initial TLS v1.2 client support. Add a default signature algorithms
7482 extension including all the algorithms we support. Parse new signature
7483 format in client key exchange. Relax some ECC signing restrictions for
7484 TLS v1.2 as indicated in RFC5246.
7488 * Add server support for TLS v1.2 signature algorithms extension. Switch
7489 to new signature format when needed using client digest preference.
7490 All server ciphersuites should now work correctly in TLS v1.2. No client
7491 support yet and no support for client certificates.
7495 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7496 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7497 ciphersuites. At present only RSA key exchange ciphersuites work with
7498 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7499 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7500 and version checking.
7504 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7505 with this defined it will not be affected by any changes to ssl internal
7506 structures. Add several utility functions to allow openssl application
7507 to work with OPENSSL_NO_SSL_INTERN defined.
7511 * A long standing patch to add support for SRP from EdelWeb (Peter
7512 Sylvester and Christophe Renou) was integrated.
7513 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7514 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7517 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7521 * Permit abbreviated handshakes when renegotiating using the function
7522 SSL_renegotiate_abbreviated().
7524 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7526 * Add call to ENGINE_register_all_complete() to
7527 ENGINE_load_builtin_engines(), so some implementations get used
7528 automatically instead of needing explicit application support.
7532 * Add support for TLS key exporter as described in RFC5705.
7534 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7536 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7537 a few changes are required:
7539 Add SSL_OP_NO_TLSv1_1 flag.
7540 Add TLSv1_1 methods.
7541 Update version checking logic to handle version 1.1.
7542 Add explicit IV handling (ported from DTLS code).
7543 Add command line options to s_client/s_server.
7550 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7552 * X509_ATTRIBUTE memory leak
7554 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7555 memory. This structure is used by the PKCS#7 and CMS routines so any
7556 application which reads PKCS#7 or CMS data from untrusted sources is
7557 affected. SSL/TLS is not affected.
7559 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7565 * Race condition handling PSK identify hint
7567 If PSK identity hints are received by a multi-threaded client then
7568 the values are wrongly updated in the parent SSL_CTX structure. This can
7569 result in a race condition potentially leading to a double free of the
7575 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7577 * Malformed ECParameters causes infinite loop
7579 When processing an ECParameters structure OpenSSL enters an infinite loop
7580 if the curve specified is over a specially malformed binary polynomial
7583 This can be used to perform denial of service against any
7584 system which processes public keys, certificate requests or
7585 certificates. This includes TLS clients and TLS servers with
7586 client authentication enabled.
7588 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7593 * Exploitable out-of-bounds read in X509_cmp_time
7595 X509_cmp_time does not properly check the length of the ASN1_TIME
7596 string and can read a few bytes out of bounds. In addition,
7597 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7600 An attacker can use this to craft malformed certificates and CRLs of
7601 various sizes and potentially cause a segmentation fault, resulting in
7602 a DoS on applications that verify certificates or CRLs. TLS clients
7603 that verify CRLs are affected. TLS clients and servers with client
7604 authentication enabled may be affected if they use custom verification
7607 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7608 independently by Hanno Böck.
7613 * PKCS7 crash with missing EnvelopedContent
7615 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7616 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7617 with missing content and trigger a NULL pointer dereference on parsing.
7619 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7620 structures from untrusted sources are affected. OpenSSL clients and
7621 servers are not affected.
7623 This issue was reported to OpenSSL by Michal Zalewski (Google).
7628 * CMS verify infinite loop with unknown hash function
7630 When verifying a signedData message the CMS code can enter an infinite loop
7631 if presented with an unknown hash function OID. This can be used to perform
7632 denial of service against any system which verifies signedData messages using
7634 This issue was reported to OpenSSL by Johannes Bauer.
7639 * Race condition handling NewSessionTicket
7641 If a NewSessionTicket is received by a multi-threaded client when attempting to
7642 reuse a previous ticket then a race condition can occur potentially leading to
7643 a double free of the ticket data.
7648 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7650 * Segmentation fault in ASN1_TYPE_cmp fix
7652 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7653 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7654 certificate signature algorithm consistency this can be used to crash any
7655 certificate verification operation and exploited in a DoS attack. Any
7656 application which performs certificate verification is vulnerable including
7657 OpenSSL clients and servers which enable client authentication.
7662 * ASN.1 structure reuse memory corruption fix
7664 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7665 memory corruption via an invalid write. Such reuse is and has been
7666 strongly discouraged and is believed to be rare.
7668 Applications that parse structures containing CHOICE or ANY DEFINED BY
7669 components may be affected. Certificate parsing (d2i_X509 and related
7670 functions) are however not affected. OpenSSL clients and servers are
7676 * PKCS7 NULL pointer dereferences fix
7678 The PKCS#7 parsing code does not handle missing outer ContentInfo
7679 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7680 missing content and trigger a NULL pointer dereference on parsing.
7682 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7683 otherwise parse PKCS#7 structures from untrusted sources are
7684 affected. OpenSSL clients and servers are not affected.
7686 This issue was reported to OpenSSL by Michal Zalewski (Google).
7691 * DoS via reachable assert in SSLv2 servers fix
7693 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7694 servers that both support SSLv2 and enable export cipher suites by sending
7695 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7697 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7698 (OpenSSL development team).
7703 * Use After Free following d2i_ECPrivatekey error fix
7705 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7706 could cause a use after free condition. This, in turn, could cause a double
7707 free in several private key parsing functions (such as d2i_PrivateKey
7708 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7709 for applications that receive EC private keys from untrusted
7710 sources. This scenario is considered rare.
7712 This issue was discovered by the BoringSSL project and fixed in their
7718 * X509_to_X509_REQ NULL pointer deref fix
7720 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7721 the certificate key is invalid. This function is rarely used in practice.
7723 This issue was discovered by Brian Carpenter.
7728 * Removed the export ciphers from the DEFAULT ciphers
7732 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7734 * Build fixes for the Windows and OpenVMS platforms
7736 *Matt Caswell and Richard Levitte*
7738 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7740 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7741 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7742 dereference. This could lead to a Denial Of Service attack. Thanks to
7743 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7748 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7749 dtls1_buffer_record function under certain conditions. In particular this
7750 could occur if an attacker sent repeated DTLS records with the same
7751 sequence number but for the next epoch. The memory leak could be exploited
7752 by an attacker in a Denial of Service attack through memory exhaustion.
7753 Thanks to Chris Mueller for reporting this issue.
7758 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7759 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7760 method would be set to NULL which could later result in a NULL pointer
7761 dereference. Thanks to Frank Schmirler for reporting this issue.
7766 * Abort handshake if server key exchange message is omitted for ephemeral
7769 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7770 reporting this issue.
7775 * Remove non-export ephemeral RSA code on client and server. This code
7776 violated the TLS standard by allowing the use of temporary RSA keys in
7777 non-export ciphersuites and could be used by a server to effectively
7778 downgrade the RSA key length used to a value smaller than the server
7779 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7780 INRIA or reporting this issue.
7785 * Fixed issue where DH client certificates are accepted without verification.
7786 An OpenSSL server will accept a DH certificate for client authentication
7787 without the certificate verify message. This effectively allows a client to
7788 authenticate without the use of a private key. This only affects servers
7789 which trust a client certificate authority which issues certificates
7790 containing DH keys: these are extremely rare and hardly ever encountered.
7791 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7797 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7798 results on some platforms, including x86_64. This bug occurs at random
7799 with a very low probability, and is not known to be exploitable in any
7800 way, though its exact impact is difficult to determine. Thanks to Pieter
7801 Wuille (Blockstream) who reported this issue and also suggested an initial
7802 fix. Further analysis was conducted by the OpenSSL development team and
7803 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7804 the OpenSSL core team.
7809 * Fix various certificate fingerprint issues.
7811 By using non-DER or invalid encodings outside the signed portion of a
7812 certificate the fingerprint can be changed without breaking the signature.
7813 Although no details of the signed portion of the certificate can be changed
7814 this can cause problems with some applications: e.g. those using the
7815 certificate fingerprint for blacklists.
7817 1. Reject signatures with non zero unused bits.
7819 If the BIT STRING containing the signature has non zero unused bits reject
7820 the signature. All current signature algorithms require zero unused bits.
7822 2. Check certificate algorithm consistency.
7824 Check the AlgorithmIdentifier inside TBS matches the one in the
7825 certificate signature. NB: this will result in signature failure
7826 errors for some broken certificates.
7828 Thanks to Konrad Kraszewski from Google for reporting this issue.
7830 3. Check DSA/ECDSA signatures use DER.
7832 Reencode DSA/ECDSA signatures and compare with the original received
7833 signature. Return an error if there is a mismatch.
7835 This will reject various cases including garbage after signature
7836 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7837 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7838 (negative or with leading zeroes).
7840 Further analysis was conducted and fixes were developed by Stephen Henson
7841 of the OpenSSL core team.
7847 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7849 * Session Ticket Memory Leak.
7851 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7852 integrity of that ticket is first verified. In the event of a session
7853 ticket integrity check failing, OpenSSL will fail to free memory
7854 causing a memory leak. By sending a large number of invalid session
7855 tickets an attacker could exploit this issue in a Denial Of Service
7861 * Build option no-ssl3 is incomplete.
7863 When OpenSSL is configured with "no-ssl3" as a build option, servers
7864 could accept and complete a SSL 3.0 handshake, and clients could be
7865 configured to send them.
7868 *Akamai and the OpenSSL team*
7870 * Add support for TLS_FALLBACK_SCSV.
7871 Client applications doing fallback retries should call
7872 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7875 *Adam Langley, Bodo Moeller*
7877 * Add additional DigestInfo checks.
7879 Reencode DigestInto in DER and check against the original when
7880 verifying RSA signature: this will reject any improperly encoded
7881 DigestInfo structures.
7883 Note: this is a precautionary measure and no attacks are currently known.
7887 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7889 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7890 to a denial of service attack. A malicious server can crash the client
7891 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7892 ciphersuite and sending carefully crafted handshake messages.
7894 Thanks to Felix Gröbert (Google) for discovering and researching this
7900 * By sending carefully crafted DTLS packets an attacker could cause openssl
7901 to leak memory. This can be exploited through a Denial of Service attack.
7902 Thanks to Adam Langley for discovering and researching this issue.
7907 * An attacker can force openssl to consume large amounts of memory whilst
7908 processing DTLS handshake messages. This can be exploited through a
7909 Denial of Service attack.
7910 Thanks to Adam Langley for discovering and researching this issue.
7915 * An attacker can force an error condition which causes openssl to crash
7916 whilst processing DTLS packets due to memory being freed twice. This
7917 can be exploited through a Denial of Service attack.
7918 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7924 * If a multithreaded client connects to a malicious server using a resumed
7925 session and the server sends an ec point format extension it could write
7926 up to 255 bytes to freed memory.
7928 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7934 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7935 X509_name_oneline, X509_name_print_ex et al. to leak some information
7936 from the stack. Applications may be affected if they echo pretty printing
7937 output to the attacker.
7939 Thanks to Ivan Fratric (Google) for discovering this issue.
7942 *Emilia Käsper, and Steve Henson*
7944 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7945 for corner cases. (Certain input points at infinity could lead to
7946 bogus results, with non-infinity inputs mapped to infinity too.)
7950 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7952 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7953 handshake can force the use of weak keying material in OpenSSL
7954 SSL/TLS clients and servers.
7956 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7957 researching this issue. ([CVE-2014-0224])
7959 *KIKUCHI Masashi, Steve Henson*
7961 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7962 OpenSSL DTLS client the code can be made to recurse eventually crashing
7965 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7968 *Imre Rad, Steve Henson*
7970 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7971 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7972 client or server. This is potentially exploitable to run arbitrary
7973 code on a vulnerable client or server.
7975 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7977 *Jüri Aedla, Steve Henson*
7979 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7980 are subject to a denial of service attack.
7982 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7983 this issue. ([CVE-2014-3470])
7985 *Felix Gröbert, Ivan Fratric, Steve Henson*
7987 * Harmonize version and its documentation. -f flag is used to display
7990 *mancha <mancha1@zoho.com>*
7992 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7993 in i2d_ECPrivateKey.
7995 *mancha <mancha1@zoho.com>*
7997 * Fix some double frees. These are not thought to be exploitable.
7999 *mancha <mancha1@zoho.com>*
8001 * Fix for the attack described in the paper "Recovering OpenSSL
8002 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8003 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8004 <http://eprint.iacr.org/2014/140>
8006 Thanks to Yuval Yarom and Naomi Benger for discovering this
8007 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8009 *Yuval Yarom and Naomi Benger*
8011 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8013 * Keep original DTLS digest and encryption contexts in retransmission
8014 structures so we can use the previous session parameters if they need
8015 to be resent. ([CVE-2013-6450])
8019 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8020 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8021 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8022 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8023 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8024 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8026 *Rob Stradling, Adam Langley*
8028 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8030 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8032 This addresses the flaw in CBC record processing discovered by
8033 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8034 at: <http://www.isg.rhul.ac.uk/tls/>
8036 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8037 Security Group at Royal Holloway, University of London
8038 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8039 Emilia Käsper for the initial patch.
8042 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8044 * Return an error when checking OCSP signatures when key is NULL.
8045 This fixes a DoS attack. ([CVE-2013-0166])
8049 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8050 the right response is stapled. Also change SSL_get_certificate()
8051 so it returns the certificate actually sent.
8052 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8053 (This is a backport)
8055 *Rob Stradling <rob.stradling@comodo.com>*
8057 * Fix possible deadlock when decoding public keys.
8061 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8063 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8066 * Sanity check record length before skipping explicit IV in DTLS
8069 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8070 fuzzing as a service testing platform.
8075 * Initialise tkeylen properly when encrypting CMS messages.
8076 Thanks to Solar Designer of Openwall for reporting this issue.
8080 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8082 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8083 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8084 in CRYPTO_realloc_clean.
8086 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8087 issue and to Adam Langley <agl@chromium.org> for fixing it.
8090 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8092 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8094 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8095 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8096 content decryption and always return the same error. Note: this attack
8097 needs on average 2^20 messages so it only affects automated senders. The
8098 old behaviour can be re-enabled in the CMS code by setting the
8099 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8100 an MMA defence is not necessary.
8101 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8102 this issue. ([CVE-2012-0884])
8106 * Fix CVE-2011-4619: make sure we really are receiving a
8107 client hello before rejecting multiple SGC restarts. Thanks to
8108 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8112 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8114 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8115 Thanks to Antonio Martin, Enterprise Secure Access Research and
8116 Development, Cisco Systems, Inc. for discovering this bug and
8117 preparing a fix. ([CVE-2012-0050])
8121 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8123 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8124 of the Vaudenay padding oracle attack on CBC mode encryption
8125 which enables an efficient plaintext recovery attack against
8126 the OpenSSL implementation of DTLS. Their attack exploits timing
8127 differences arising during decryption processing. A research
8128 paper describing this attack can be found at:
8129 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8130 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8131 Security Group at Royal Holloway, University of London
8132 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8133 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8134 for preparing the fix. ([CVE-2011-4108])
8136 *Robin Seggelmann, Michael Tuexen*
8138 * Clear bytes used for block padding of SSL 3.0 records.
8141 *Adam Langley (Google)*
8143 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8144 Kadianakis <desnacked@gmail.com> for discovering this issue and
8145 Adam Langley for preparing the fix. ([CVE-2011-4619])
8147 *Adam Langley (Google)*
8149 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8151 *Andrey Kulikov <amdeich@gmail.com>*
8153 * Prevent malformed RFC3779 data triggering an assertion failure.
8154 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8155 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8157 *Rob Austein <sra@hactrn.net>*
8159 * Improved PRNG seeding for VOS.
8161 *Paul Green <Paul.Green@stratus.com>*
8163 * Fix ssl_ciph.c set-up race.
8165 *Adam Langley (Google)*
8167 * Fix spurious failures in ecdsatest.c.
8169 *Emilia Käsper (Google)*
8171 * Fix the BIO_f_buffer() implementation (which was mixing different
8172 interpretations of the `..._len` fields).
8174 *Adam Langley (Google)*
8176 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8177 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8178 threads won't reuse the same blinding coefficients.
8180 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8181 lock to call BN_BLINDING_invert_ex, and avoids one use of
8182 BN_BLINDING_update for each BN_BLINDING structure (previously,
8183 the last update always remained unused).
8185 *Emilia Käsper (Google)*
8187 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8189 *Bob Buckholz (Google)*
8191 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8193 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8194 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8196 *Kaspar Brand <ossl@velox.ch>*
8198 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8199 for multi-threaded use of ECDH. ([CVE-2011-3210])
8201 *Adam Langley (Google)*
8203 * Fix x509_name_ex_d2i memory leak on bad inputs.
8207 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8208 signature public key algorithm by using OID xref utilities instead.
8209 Before this you could only use some ECC ciphersuites with SHA1 only.
8213 * Add protection against ECDSA timing attacks as mentioned in the paper
8214 by Billy Bob Brumley and Nicola Tuveri, see:
8215 <http://eprint.iacr.org/2011/232.pdf>
8217 *Billy Bob Brumley and Nicola Tuveri*
8219 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8221 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8223 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8225 * Fix bug in string printing code: if *any* escaping is enabled we must
8226 escape the escape character (backslash) or the resulting string is
8231 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8233 * Disable code workaround for ancient and obsolete Netscape browsers
8234 and servers: an attacker can use it in a ciphersuite downgrade attack.
8235 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8239 * Fixed J-PAKE implementation error, originally discovered by
8240 Sebastien Martini, further info and confirmation from Stefan
8241 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8245 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8247 * Fix extension code to avoid race conditions which can result in a buffer
8248 overrun vulnerability: resumed sessions must not be modified as they can
8249 be shared by multiple threads. CVE-2010-3864
8253 * Fix WIN32 build system to correctly link an ENGINE directory into
8258 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8260 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8263 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8265 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8267 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8268 context. The operation can be customised via the ctrl mechanism in
8269 case ENGINEs want to include additional functionality.
8273 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8277 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8278 output hashes compatible with older versions of OpenSSL.
8280 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8282 * Fix compression algorithm handling: if resuming a session use the
8283 compression algorithm of the resumed session instead of determining
8284 it from client hello again. Don't allow server to change algorithm.
8288 * Add load_crls() function to commands tidying load_certs() too. Add option
8289 to verify utility to allow additional CRLs to be included.
8293 * Update OCSP request code to permit adding custom headers to the request:
8294 some responders need this.
8298 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8301 *Julia Lawall <julia@diku.dk>*
8303 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8304 needlessly dereferenced structures, used obsolete functions and
8305 didn't handle all updated verify codes correctly.
8309 * Disable MD2 in the default configuration.
8313 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8314 indicate the initial BIO being pushed or popped. This makes it possible
8315 to determine whether the BIO is the one explicitly called or as a result
8316 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8317 it handles reference counts correctly and doesn't zero out the I/O bio
8318 when it is not being explicitly popped. WARNING: applications which
8319 included workarounds for the old buggy behaviour will need to be modified
8320 or they could free up already freed BIOs.
8324 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8325 renaming to all platforms (within the 0.9.8 branch, this was
8326 done conditionally on Netware platforms to avoid a name clash).
8328 *Guenter <lists@gknw.net>*
8330 * Add ECDHE and PSK support to DTLS.
8332 *Michael Tuexen <tuexen@fh-muenster.de>*
8334 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8339 * Add "missing" function EVP_MD_flags() (without this the only way to
8340 retrieve a digest flags is by accessing the structure directly. Update
8341 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8342 or cipher is registered as in the "from" argument. Print out all
8343 registered digests in the dgst usage message instead of manually
8344 attempting to work them out.
8348 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8349 this allows the use of compression and extensions. Change default cipher
8350 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8351 by default unless an application cipher string requests it.
8355 * Alter match criteria in PKCS12_parse(). It used to try to use local
8356 key ids to find matching certificates and keys but some PKCS#12 files
8357 don't follow the (somewhat unwritten) rules and this strategy fails.
8358 Now just gather all certificates together and the first private key
8359 then look for the first certificate that matches the key.
8363 * Support use of registered digest and cipher names for dgst and cipher
8364 commands instead of having to add each one as a special case. So now
8371 openssl dgst -sha256 foo
8373 and this works for ENGINE based algorithms too.
8377 * Update Gost ENGINE to support parameter files.
8379 *Victor B. Wagner <vitus@cryptocom.ru>*
8381 * Support GeneralizedTime in ca utility.
8383 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8385 * Enhance the hash format used for certificate directory links. The new
8386 form uses the canonical encoding (meaning equivalent names will work
8387 even if they aren't identical) and uses SHA1 instead of MD5. This form
8388 is incompatible with the older format and as a result c_rehash should
8389 be used to rebuild symbolic links.
8393 * Make PKCS#8 the default write format for private keys, replacing the
8394 traditional format. This form is standardised, more secure and doesn't
8395 include an implicit MD5 dependency.
8399 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8400 committed to OpenSSL should pass this lot as a minimum.
8404 * Add session ticket override functionality for use by EAP-FAST.
8406 *Jouni Malinen <j@w1.fi>*
8408 * Modify HMAC functions to return a value. Since these can be implemented
8409 in an ENGINE errors can occur.
8413 * Type-checked OBJ_bsearch_ex.
8417 * Type-checked OBJ_bsearch. Also some constification necessitated
8418 by type-checking. Still to come: TXT_DB, bsearch(?),
8419 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8424 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8425 seconds to a tm structure directly, instead of going through OS
8426 specific date routines. This avoids any issues with OS routines such
8427 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8428 and X509_time_adj_ex() to cover the extended range. The existing
8429 X509_time_adj() is still usable and will no longer have any date issues.
8433 * Delta CRL support. New use deltas option which will attempt to locate
8434 and search any appropriate delta CRLs available.
8436 This work was sponsored by Google.
8440 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8441 code and add additional score elements. Validate alternate CRL paths
8442 as part of the CRL checking and indicate a new error "CRL path validation
8443 error" in this case. Applications wanting additional details can use
8444 the verify callback and check the new "parent" field. If this is not
8445 NULL CRL path validation is taking place. Existing applications won't
8446 see this because it requires extended CRL support which is off by
8449 This work was sponsored by Google.
8453 * Support for freshest CRL extension.
8455 This work was sponsored by Google.
8459 * Initial indirect CRL support. Currently only supported in the CRLs
8460 passed directly and not via lookup. Process certificate issuer
8461 CRL entry extension and lookup CRL entries by bother issuer name
8462 and serial number. Check and process CRL issuer entry in IDP extension.
8464 This work was sponsored by Google.
8468 * Add support for distinct certificate and CRL paths. The CRL issuer
8469 certificate is validated separately in this case. Only enabled if
8470 an extended CRL support flag is set: this flag will enable additional
8471 CRL functionality in future.
8473 This work was sponsored by Google.
8477 * Add support for policy mappings extension.
8479 This work was sponsored by Google.
8483 * Fixes to pathlength constraint, self issued certificate handling,
8484 policy processing to align with RFC3280 and PKITS tests.
8486 This work was sponsored by Google.
8490 * Support for name constraints certificate extension. DN, email, DNS
8491 and URI types are currently supported.
8493 This work was sponsored by Google.
8497 * To cater for systems that provide a pointer-based thread ID rather
8498 than numeric, deprecate the current numeric thread ID mechanism and
8499 replace it with a structure and associated callback type. This
8500 mechanism allows a numeric "hash" to be extracted from a thread ID in
8501 either case, and on platforms where pointers are larger than 'long',
8502 mixing is done to help ensure the numeric 'hash' is usable even if it
8503 can't be guaranteed unique. The default mechanism is to use "&errno"
8504 as a pointer-based thread ID to distinguish between threads.
8506 Applications that want to provide their own thread IDs should now use
8507 CRYPTO_THREADID_set_callback() to register a callback that will call
8508 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8510 Note that ERR_remove_state() is now deprecated, because it is tied
8511 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8512 to free the current thread's error state should be replaced by
8513 ERR_remove_thread_state(NULL).
8515 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8516 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8517 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8518 application was previously providing a numeric thread callback that
8519 was inappropriate for distinguishing threads, then uniqueness might
8520 have been obtained with &errno that happened immediately in the
8521 intermediate development versions of OpenSSL; this is no longer the
8522 case, the numeric thread callback will now override the automatic use
8525 *Geoff Thorpe, with help from Bodo Moeller*
8527 * Initial support for different CRL issuing certificates. This covers a
8528 simple case where the self issued certificates in the chain exist and
8529 the real CRL issuer is higher in the existing chain.
8531 This work was sponsored by Google.
8535 * Removed effectively defunct crypto/store from the build.
8539 * Revamp of STACK to provide stronger type-checking. Still to come:
8540 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8541 ASN1_STRING, CONF_VALUE.
8545 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8546 RAM on SSL connections. This option can save about 34k per idle SSL.
8550 * Revamp of LHASH to provide stronger type-checking. Still to come:
8551 STACK, TXT_DB, bsearch, qsort.
8555 * Initial support for Cryptographic Message Syntax (aka CMS) based
8556 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8557 support for data, signedData, compressedData, digestedData and
8558 encryptedData, envelopedData types included. Scripts to check against
8559 RFC4134 examples draft and interop and consistency checks of many
8560 content types and variants.
8564 * Add options to enc utility to support use of zlib compression BIO.
8568 * Extend mk1mf to support importing of options and assembly language
8569 files from Configure script, currently only included in VC-WIN32.
8570 The assembly language rules can now optionally generate the source
8571 files from the associated perl scripts.
8575 * Implement remaining functionality needed to support GOST ciphersuites.
8576 Interop testing has been performed using CryptoPro implementations.
8578 *Victor B. Wagner <vitus@cryptocom.ru>*
8580 * s390x assembler pack.
8584 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8589 * Implement Opaque PRF Input TLS extension as specified in
8590 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8591 official specification yet and no extension type assignment by
8592 IANA exists, this extension (for now) will have to be explicitly
8593 enabled when building OpenSSL by providing the extension number
8594 to use. For example, specify an option
8596 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8598 to the "config" or "Configure" script to enable the extension,
8599 assuming extension number 0x9527 (which is a completely arbitrary
8600 and unofficial assignment based on the MD5 hash of the Internet
8601 Draft). Note that by doing so, you potentially lose
8602 interoperability with other TLS implementations since these might
8603 be using the same extension number for other purposes.
8605 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8606 opaque PRF input value to use in the handshake. This will create
8607 an internal copy of the length-'len' string at 'src', and will
8608 return non-zero for success.
8610 To get more control and flexibility, provide a callback function
8613 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8614 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8618 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8621 Callback function 'cb' will be called in handshakes, and is
8622 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8623 Argument 'arg' is for application purposes (the value as given to
8624 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8625 be provided to the callback function). The callback function
8626 has to return non-zero to report success: usually 1 to use opaque
8627 PRF input just if possible, or 2 to enforce use of the opaque PRF
8628 input. In the latter case, the library will abort the handshake
8629 if opaque PRF input is not successfully negotiated.
8631 Arguments 'peerinput' and 'len' given to the callback function
8632 will always be NULL and 0 in the case of a client. A server will
8633 see the client's opaque PRF input through these variables if
8634 available (NULL and 0 otherwise). Note that if the server
8635 provides an opaque PRF input, the length must be the same as the
8636 length of the client's opaque PRF input.
8638 Note that the callback function will only be called when creating
8639 a new session (session resumption can resume whatever was
8640 previously negotiated), and will not be called in SSL 2.0
8641 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8642 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8643 for applications that need to enforce opaque PRF input.
8647 * Update ssl code to support digests other than SHA1+MD5 for handshake
8650 *Victor B. Wagner <vitus@cryptocom.ru>*
8652 * Add RFC4507 support to OpenSSL. This includes the corrections in
8653 RFC4507bis. The encrypted ticket format is an encrypted encoded
8654 SSL_SESSION structure, that way new session features are automatically
8657 If a client application caches session in an SSL_SESSION structure
8658 support is transparent because tickets are now stored in the encoded
8661 The SSL_CTX structure automatically generates keys for ticket
8662 protection in servers so again support should be possible
8663 with no application modification.
8665 If a client or server wishes to disable RFC4507 support then the option
8666 SSL_OP_NO_TICKET can be set.
8668 Add a TLS extension debugging callback to allow the contents of any client
8669 or server extensions to be examined.
8671 This work was sponsored by Google.
8675 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8676 OpenSSL should now compile cleanly on gcc 4.2
8678 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8680 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8681 support including streaming MAC support: this is required for GOST
8682 ciphersuite support.
8684 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8686 * Add option -stream to use PKCS#7 streaming in smime utility. New
8687 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8688 to output in BER and PEM format.
8692 * Experimental support for use of HMAC via EVP_PKEY interface. This
8693 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8694 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8695 ENGINE support for HMAC keys which are unextractable. New -mac and
8696 -macopt options to dgst utility.
8700 * New option -sigopt to dgst utility. Update dgst to use
8701 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8702 alternative signing parameters such as X9.31 or PSS in the dgst
8707 * Change ssl_cipher_apply_rule(), the internal function that does
8708 the work each time a ciphersuite string requests enabling
8709 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8710 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8711 the order of disabled ciphersuites such that those ciphersuites
8712 that most recently went from enabled to disabled not only stay
8713 in order with respect to each other, but also have higher priority
8714 than other disabled ciphersuites the next time ciphersuites are
8717 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8718 the same ciphersuites as with "HIGH" alone, but in a specific
8719 order where the PSK ciphersuites come first (since they are the
8720 most recently disabled ciphersuites when "HIGH" is parsed).
8722 Also, change ssl_create_cipher_list() (using this new
8723 functionality) such that between otherwise identical
8724 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8729 * Change ssl_create_cipher_list() so that it automatically
8730 arranges the ciphersuites in reasonable order before starting
8731 to process the rule string. Thus, the definition for "DEFAULT"
8732 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8733 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8734 This makes it much easier to arrive at a reasonable default order
8735 in applications for which anonymous ciphers are OK (meaning
8736 that you can't actually use DEFAULT).
8738 *Bodo Moeller; suggested by Victor Duchovni*
8740 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8741 processing) into multiple integers instead of setting
8742 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8743 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8744 (These masks as well as the individual bit definitions are hidden
8745 away into the non-exported interface ssl/ssl_locl.h, so this
8746 change to the definition of the SSL_CIPHER structure shouldn't
8747 affect applications.) This give us more bits for each of these
8748 categories, so there is no longer a need to coagulate AES128 and
8749 AES256 into a single algorithm bit, and to coagulate Camellia128
8750 and Camellia256 into a single algorithm bit, which has led to all
8753 Thus, among other things, the kludge introduced in 0.9.7m and
8754 0.9.8e for masking out AES256 independently of AES128 or masking
8755 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8757 With the change, we also introduce new ciphersuite aliases that
8758 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8763 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8764 Use the leftmost N bytes of the signature input if the input is
8765 larger than the prime q (with N being the size in bytes of q).
8769 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8770 it yet and it is largely untested.
8774 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8778 * Initial incomplete changes to avoid need for function casts in OpenSSL
8779 some compilers (gcc 4.2 and later) reject their use. Safestack is
8780 reimplemented. Update ASN1 to avoid use of legacy functions.
8784 * Win32/64 targets are linked with Winsock2.
8788 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8789 to external functions. This can be used to increase CRL handling
8790 efficiency especially when CRLs are very large by (for example) storing
8791 the CRL revoked certificates in a database.
8795 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8796 new CRLs added to a directory can be used. New command line option
8797 -verify_return_error to s_client and s_server. This causes real errors
8798 to be returned by the verify callback instead of carrying on no matter
8799 what. This reflects the way a "real world" verify callback would behave.
8803 * GOST engine, supporting several GOST algorithms and public key formats.
8804 Kindly donated by Cryptocom.
8808 * Partial support for Issuing Distribution Point CRL extension. CRLs
8809 partitioned by DP are handled but no indirect CRL or reason partitioning
8810 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8811 selected via a scoring technique which handles IDP and AKID in CRLs.
8815 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8816 will ultimately be used for all verify operations: this will remove the
8817 X509_STORE dependency on certificate verification and allow alternative
8818 lookup methods. X509_STORE based implementations of these two callbacks.
8822 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8823 Modify get_crl() to find a valid (unexpired) CRL if possible.
8827 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8828 this would be called X509_CRL_cmp() but that name is already used by
8829 a function that just compares CRL issuer names. Cache several CRL
8830 extensions in X509_CRL structure and cache CRLDP in X509.
8834 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8835 this maps equivalent X509_NAME structures into a consistent structure.
8836 Name comparison can then be performed rapidly using memcmp().
8840 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8845 * Allow digests to supply their own micalg string for S/MIME type using
8846 the ctrl EVP_MD_CTRL_MICALG.
8850 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8851 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8852 ctrl. It can then customise the structure before and/or after signing
8857 * New function OBJ_add_sigid() to allow application defined signature OIDs
8858 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8859 to free up any added signature OIDs.
8863 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8864 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8865 digest and cipher tables. New options added to openssl utility:
8866 list-message-digest-algorithms and list-cipher-algorithms.
8870 * Change the array representation of binary polynomials: the list
8871 of degrees of non-zero coefficients is now terminated with -1.
8872 Previously it was terminated with 0, which was also part of the
8873 value; thus, the array representation was not applicable to
8874 polynomials where t^0 has coefficient zero. This change makes
8875 the array representation useful in a more general context.
8879 * Various modifications and fixes to SSL/TLS cipher string
8880 handling. For ECC, the code now distinguishes between fixed ECDH
8881 with RSA certificates on the one hand and with ECDSA certificates
8882 on the other hand, since these are separate ciphersuites. The
8883 unused code for Fortezza ciphersuites has been removed.
8885 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8886 (not "ECDHE"). For consistency with the code for DH
8887 certificates, use of ECDH certificates is now considered ECDH
8888 authentication, not RSA or ECDSA authentication (the latter is
8889 merely the CA's signing algorithm and not actively used in the
8892 The temporary ciphersuite alias "ECCdraft" is no longer
8893 available, and ECC ciphersuites are no longer excluded from "ALL"
8894 and "DEFAULT". The following aliases now exist for RFC 4492
8895 ciphersuites, most of these by analogy with the DH case:
8897 kECDHr - ECDH cert, signed with RSA
8898 kECDHe - ECDH cert, signed with ECDSA
8899 kECDH - ECDH cert (signed with either RSA or ECDSA)
8900 kEECDH - ephemeral ECDH
8901 ECDH - ECDH cert or ephemeral ECDH
8907 AECDH - anonymous ECDH
8908 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8912 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8913 Use correct micalg parameters depending on digest(s) in signed message.
8917 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8918 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8922 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8923 an engine to register a method. Add ENGINE lookups for methods and
8924 functional reference processing.
8928 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8929 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8934 * New -resign option to smime utility. This adds one or more signers
8935 to an existing PKCS#7 signedData structure. Also -md option to use an
8936 alternative message digest algorithm for signing.
8940 * Tidy up PKCS#7 routines and add new functions to make it easier to
8941 create PKCS7 structures containing multiple signers. Update smime
8942 application to support multiple signers.
8946 * New -macalg option to pkcs12 utility to allow setting of an alternative
8951 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8952 Reorganize PBE internals to lookup from a static table using NIDs,
8953 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8954 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8955 PRF which will be automatically used with PBES2.
8959 * Replace the algorithm specific calls to generate keys in "req" with the
8964 * Update PKCS#7 enveloped data routines to use new API. This is now
8965 supported by any public key method supporting the encrypt operation. A
8966 ctrl is added to allow the public key algorithm to examine or modify
8967 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8972 * Add a ctrl to asn1 method to allow a public key algorithm to express
8973 a default digest type to use. In most cases this will be SHA1 but some
8974 algorithms (such as GOST) need to specify an alternative digest. The
8975 return value indicates how strong the preference is 1 means optional and
8976 2 is mandatory (that is it is the only supported type). Modify
8977 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8978 use the default md. Update openssl utilities to use the default digest
8979 type for signing if it is not explicitly indicated.
8983 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8984 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8985 signing method from the key type. This effectively removes the link
8986 between digests and public key types.
8990 * Add an OID cross reference table and utility functions. Its purpose is to
8991 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8992 rsaEncryption. This will allow some of the algorithm specific hackery
8993 needed to use the correct OID to be removed.
8997 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8998 structures for PKCS7_sign(). They are now set up by the relevant public
9003 * Add provisional EC pkey method with support for ECDSA and ECDH.
9007 * Add support for key derivation (agreement) in the API, DH method and
9012 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9013 public and private key formats. As a side effect these add additional
9014 command line functionality not previously available: DSA signatures can be
9015 generated and verified using pkeyutl and DH key support and generation in
9022 *Oliver Tappe <zooey@hirschkaefer.de>*
9024 * New make target "install_html_docs" installs HTML renditions of the
9027 *Oliver Tappe <zooey@hirschkaefer.de>*
9029 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9030 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9031 support key and parameter generation and add initial key generation
9032 functionality for RSA.
9036 * Add functions for main EVP_PKEY_method operations. The undocumented
9037 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9038 `EVP_PKEY_{encrypt,decrypt}_old`.
9042 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9043 key API, doesn't do much yet.
9047 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9048 public key algorithms. New option to openssl utility:
9049 "list-public-key-algorithms" to print out info.
9053 * Implement the Supported Elliptic Curves Extension for
9054 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9058 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9059 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9063 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9064 utilities such as rsa, dsa, dsaparam etc except they process any key
9069 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9070 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9071 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9076 * Initial support for pluggable public key ASN1.
9077 De-spaghettify the public key ASN1 handling. Move public and private
9078 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9079 algorithm specific handling to a single module within the relevant
9080 algorithm directory. Add functions to allow (near) opaque processing
9081 of public and private key structures.
9085 * Implement the Supported Point Formats Extension for
9086 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9090 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9091 for the psk identity [hint] and the psk callback functions to the
9092 SSL_SESSION, SSL and SSL_CTX structure.
9095 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9099 SSL_CTX_use_psk_identity_hint
9100 SSL_get_psk_identity_hint
9101 SSL_get_psk_identity
9102 SSL_use_psk_identity_hint
9104 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9106 * Add RFC 3161 compliant time stamp request creation, response generation
9107 and response verification functionality.
9109 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9111 * Add initial support for TLS extensions, specifically for the server_name
9112 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9113 have new members for a host name. The SSL data structure has an
9114 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9115 stored in that context to allow for session resumption, even after the
9116 SSL has been switched to a new SSL_CTX in reaction to a client's
9117 server_name extension.
9119 New functions (subject to change):
9121 SSL_get_servername()
9122 SSL_get_servername_type()
9125 New CTRL codes and macros (subject to change):
9127 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9128 - SSL_CTX_set_tlsext_servername_callback()
9129 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9130 - SSL_CTX_set_tlsext_servername_arg()
9131 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9133 openssl s_client has a new '-servername ...' option.
9135 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9136 '-key2 ...', '-servername_fatal' (subject to change). This allows
9137 testing the HostName extension for a specific single host name ('-cert'
9138 and '-key' remain fallbacks for handshakes without HostName
9139 negotiation). If the unrecognized_name alert has to be sent, this by
9140 default is a warning; it becomes fatal with the '-servername_fatal'
9143 *Peter Sylvester, Remy Allais, Christophe Renou*
9145 * Whirlpool hash implementation is added.
9149 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9150 bn(64,32). Because of instruction set limitations it doesn't have
9151 any negative impact on performance. This was done mostly in order
9152 to make it possible to share assembler modules, such as bn_mul_mont
9153 implementations, between 32- and 64-bit builds without hassle.
9157 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9158 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9163 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9164 dedicated Montgomery multiplication procedure, is introduced.
9165 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9166 "64-bit" performance on certain 32-bit targets.
9170 * New option SSL_OP_NO_COMP to disable use of compression selectively
9171 in SSL structures. New SSL ctrl to set maximum send fragment size.
9172 Save memory by setting the I/O buffer sizes dynamically instead of
9173 using the maximum available value.
9177 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9178 in addition to the text details.
9182 * Very, very preliminary EXPERIMENTAL support for printing of general
9183 ASN1 structures. This currently produces rather ugly output and doesn't
9184 handle several customised structures at all.
9188 * Integrated support for PVK file format and some related formats such
9189 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9190 these in the 'rsa' and 'dsa' utilities.
9194 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9198 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9199 place for the (very old) "NETSCAPE" format certificates which are now
9200 handled using new ASN1 code equivalents.
9204 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9205 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9206 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9210 * Modify CRL distribution points extension code to print out previously
9211 unsupported fields. Enhance extension setting code to allow setting of
9216 * Add print and set support for Issuing Distribution Point CRL extension.
9220 * Change 'Configure' script to enable Camellia by default.
9227 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9229 * When rejecting SSL/TLS records due to an incorrect version number, never
9230 update s->server with a new major version number. As of
9231 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9232 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9233 the previous behavior could result in a read attempt at NULL when
9234 receiving specific incorrect SSL/TLS records once record payload
9235 protection is active. ([CVE-2010-0740])
9237 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9239 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9240 could be crashed if the relevant tables were not present (e.g. chrooted).
9242 *Tomas Hoger <thoger@redhat.com>*
9244 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9246 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9248 *Martin Olsson, Neel Mehta*
9250 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9251 accommodate for stack sorting, always a write lock!).
9255 * On some versions of WIN32 Heap32Next is very slow. This can cause
9256 excessive delays in the RAND_poll(): over a minute. As a workaround
9257 include a time check in the inner Heap32Next loop too.
9261 * The code that handled flushing of data in SSL/TLS originally used the
9262 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9263 the problem outlined in PR#1949. The fix suggested there however can
9264 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9265 of Apache). So instead simplify the code to flush unconditionally.
9266 This should be fine since flushing with no data to flush is a no op.
9270 * Handle TLS versions 2.0 and later properly and correctly use the
9271 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9272 off ancient servers have a habit of sticking around for a while...
9276 * Modify compression code so it frees up structures without using the
9277 ex_data callbacks. This works around a problem where some applications
9278 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9279 restarting) then use compression (e.g. SSL with compression) later.
9280 This results in significant per-connection memory leaks and
9281 has caused some security issues including CVE-2008-1678 and
9286 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9287 change when encrypting or decrypting.
9291 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9292 connect and renegotiate with servers which do not support RI.
9293 Until RI is more widely deployed this option is enabled by default.
9297 * Add "missing" ssl ctrls to clear options and mode.
9301 * If client attempts to renegotiate and doesn't support RI respond with
9302 a no_renegotiation alert as required by RFC5746. Some renegotiating
9303 TLS clients will continue a connection gracefully when they receive
9304 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9305 waiting for a server hello which it will never receive. Now we treat a
9306 received no_renegotiation alert as a fatal error. This is because
9307 applications requesting a renegotiation might well expect it to succeed
9308 and would have no code in place to handle the server denying it so the
9309 only safe thing to do is to terminate the connection.
9313 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9314 peer supports secure renegotiation and 0 otherwise. Print out peer
9315 renegotiation support in s_client/s_server.
9319 * Replace the highly broken and deprecated SPKAC certification method with
9320 the updated NID creation version. This should correctly handle UTF8.
9324 * Implement RFC5746. Re-enable renegotiation but require the extension
9325 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9326 turns out to be a bad idea. It has been replaced by
9327 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9328 SSL_CTX_set_options(). This is really not recommended unless you
9329 know what you are doing.
9331 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9333 * Fixes to stateless session resumption handling. Use initial_ctx when
9334 issuing and attempting to decrypt tickets in case it has changed during
9335 servername handling. Use a non-zero length session ID when attempting
9336 stateless session resumption: this makes it possible to determine if
9337 a resumption has occurred immediately after receiving server hello
9338 (several places in OpenSSL subtly assume this) instead of later in
9343 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9344 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9345 fixes for a few places where the return code is not checked
9348 *Julia Lawall <julia@diku.dk>*
9350 * Add --strict-warnings option to Configure script to include devteam
9351 warnings in other configurations.
9355 * Add support for --libdir option and LIBDIR variable in makefiles. This
9356 makes it possible to install openssl libraries in locations which
9357 have names other than "lib", for example "/usr/lib64" which some
9360 *Steve Henson, based on patch from Jeremy Utley*
9362 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9363 X690 8.9.12 and can produce some misleading textual output of OIDs.
9365 *Steve Henson, reported by Dan Kaminsky*
9367 * Delete MD2 from algorithm tables. This follows the recommendation in
9368 several standards that it is not used in new applications due to
9369 several cryptographic weaknesses. For binary compatibility reasons
9370 the MD2 API is still compiled in by default.
9374 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9379 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9380 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9383 *Guenter <lists@gknw.net>*
9385 * Fix the server certificate chain building code to use X509_verify_cert(),
9386 it used to have an ad-hoc builder which was unable to cope with anything
9387 other than a simple chain.
9389 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9391 * Don't check self signed certificate signatures in X509_verify_cert()
9392 by default (a flag can override this): it just wastes time without
9393 adding any security. As a useful side effect self signed root CAs
9394 with non-FIPS digests are now usable in FIPS mode.
9398 * In dtls1_process_out_of_seq_message() the check if the current message
9399 is already buffered was missing. For every new message was memory
9400 allocated, allowing an attacker to perform an denial of service attack
9401 with sending out of seq handshake messages until there is no memory
9402 left. Additionally every future message was buffered, even if the
9403 sequence number made no sense and would be part of another handshake.
9404 So only messages with sequence numbers less than 10 in advance will be
9405 buffered. ([CVE-2009-1378])
9407 *Robin Seggelmann, discovered by Daniel Mentz*
9409 * Records are buffered if they arrive with a future epoch to be
9410 processed after finishing the corresponding handshake. There is
9411 currently no limitation to this buffer allowing an attacker to perform
9412 a DOS attack with sending records with future epochs until there is no
9413 memory left. This patch adds the pqueue_size() function to determine
9414 the size of a buffer and limits the record buffer to 100 entries.
9417 *Robin Seggelmann, discovered by Daniel Mentz*
9419 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9420 parent structure is freed. ([CVE-2009-1379])
9424 * Handle non-blocking I/O properly in SSL_shutdown() call.
9426 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9428 * Add `2.5.4.*` OIDs
9430 *Ilya O. <vrghost@gmail.com>*
9432 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9434 * Disable renegotiation completely - this fixes a severe security
9435 problem ([CVE-2009-3555]) at the cost of breaking all
9436 renegotiation. Renegotiation can be re-enabled by setting
9437 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9438 run-time. This is really not recommended unless you know what
9443 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9445 * Don't set val to NULL when freeing up structures, it is freed up by
9446 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9447 zeroing past the valid field. ([CVE-2009-0789])
9449 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9451 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9452 checked correctly. This would allow some invalid signed attributes to
9453 appear to verify correctly. ([CVE-2009-0591])
9455 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9457 * Reject UniversalString and BMPString types with invalid lengths. This
9458 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9459 a legal length. ([CVE-2009-0590])
9463 * Set S/MIME signing as the default purpose rather than setting it
9464 unconditionally. This allows applications to override it at the store
9469 * Permit restricted recursion of ASN1 strings. This is needed in practice
9470 to handle some structures.
9474 * Improve efficiency of mem_gets: don't search whole buffer each time
9477 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9479 * New -hex option for openssl rand.
9483 * Print out UTF8String and NumericString when parsing ASN1.
9487 * Support NumericString type for name components.
9491 * Allow CC in the environment to override the automatically chosen
9492 compiler. Note that nothing is done to ensure flags work with the
9497 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9499 * Properly check EVP_VerifyFinal() and similar return values
9502 *Ben Laurie, Bodo Moeller, Google Security Team*
9504 * Enable TLS extensions by default.
9508 * Allow the CHIL engine to be loaded, whether the application is
9509 multithreaded or not. (This does not release the developer from the
9510 obligation to set up the dynamic locking callbacks.)
9512 *Sander Temme <sander@temme.net>*
9514 * Use correct exit code if there is an error in dgst command.
9516 *Steve Henson; problem pointed out by Roland Dirlewanger*
9518 * Tweak Configure so that you need to say "experimental-jpake" to enable
9519 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9523 * Add experimental JPAKE support, including demo authentication in
9524 s_client and s_server.
9528 * Set the comparison function in v3_addr_canonize().
9530 *Rob Austein <sra@hactrn.net>*
9532 * Add support for XMPP STARTTLS in s_client.
9534 *Philip Paeps <philip@freebsd.org>*
9536 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9537 to ensure that even with this option, only ciphersuites in the
9538 server's preference list will be accepted. (Note that the option
9539 applies only when resuming a session, so the earlier behavior was
9540 just about the algorithm choice for symmetric cryptography.)
9544 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9546 * Fix NULL pointer dereference if a DTLS server received
9547 ChangeCipherSpec as first record ([CVE-2009-1386]).
9551 * Fix a state transition in s3_srvr.c and d1_srvr.c
9552 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9556 * The fix in 0.9.8c that supposedly got rid of unsafe
9557 double-checked locking was incomplete for RSA blinding,
9558 addressing just one layer of what turns out to have been
9559 doubly unsafe triple-checked locking.
9561 So now fix this for real by retiring the MONT_HELPER macro
9562 in crypto/rsa/rsa_eay.c.
9564 *Bodo Moeller; problem pointed out by Marius Schilder*
9566 * Various precautionary measures:
9568 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9570 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9571 (NB: This would require knowledge of the secret session ticket key
9572 to exploit, in which case you'd be SOL either way.)
9574 - Change bn_nist.c so that it will properly handle input BIGNUMs
9575 outside the expected range.
9577 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9580 *Neel Mehta, Bodo Moeller*
9582 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9583 the load fails. Useful for distros.
9585 *Ben Laurie and the FreeBSD team*
9587 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9591 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9595 * Expand ENGINE to support engine supplied SSL client certificate functions.
9597 This work was sponsored by Logica.
9601 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9602 keystores. Support for SSL/TLS client authentication too.
9603 Not compiled unless enable-capieng specified to Configure.
9605 This work was sponsored by Logica.
9609 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9610 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9611 attribute creation routines such as certificate requests and PKCS#12
9616 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9618 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9619 handshake which could lead to a client crash as found using the
9620 Codenomicon TLS test suite ([CVE-2008-1672])
9622 *Steve Henson, Mark Cox*
9624 * Fix double free in TLS server name extensions which could lead to
9625 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9629 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9631 Clear the error queue to ensure that error entries left from
9632 older function calls do not interfere with the correct operation.
9634 *Lutz Jaenicke, Erik de Castro Lopo*
9636 * Remove root CA certificates of commercial CAs:
9638 The OpenSSL project does not recommend any specific CA and does not
9639 have any policy with respect to including or excluding any CA.
9640 Therefore it does not make any sense to ship an arbitrary selection
9641 of root CA certificates with the OpenSSL software.
9645 * RSA OAEP patches to fix two separate invalid memory reads.
9646 The first one involves inputs when 'lzero' is greater than
9647 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9648 before the beginning of from). The second one involves inputs where
9649 the 'db' section contains nothing but zeroes (there is a one-byte
9650 invalid read after the end of 'db').
9652 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9654 * Partial backport from 0.9.9-dev:
9656 Introduce bn_mul_mont (dedicated Montgomery multiplication
9657 procedure) as a candidate for BIGNUM assembler implementation.
9658 While 0.9.9-dev uses assembler for various architectures, only
9659 x86_64 is available by default here in the 0.9.8 branch, and
9660 32-bit x86 is available through a compile-time setting.
9662 To try the 32-bit x86 assembler implementation, use Configure
9663 option "enable-montasm" (which exists only for this backport).
9665 As "enable-montasm" for 32-bit x86 disclaims code stability
9666 anyway, in this constellation we activate additional code
9667 backported from 0.9.9-dev for further performance improvements,
9668 namely BN_from_montgomery_word. (To enable this otherwise,
9669 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9671 *Andy Polyakov (backport partially by Bodo Moeller)*
9673 * Add TLS session ticket callback. This allows an application to set
9674 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9675 values. This is useful for key rollover for example where several key
9676 sets may exist with different names.
9680 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9681 This was broken until now in 0.9.8 releases, such that the only way
9682 a registered ENGINE could be used (assuming it initialises
9683 successfully on the host) was to explicitly set it as the default
9684 for the relevant algorithms. This is in contradiction with 0.9.7
9685 behaviour and the documentation. With this fix, when an ENGINE is
9686 registered into a given algorithm's table of implementations, the
9687 'uptodate' flag is reset so that auto-discovery will be used next
9688 time a new context for that algorithm attempts to select an
9691 *Ian Lister (tweaked by Geoff Thorpe)*
9693 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9694 implementation in the following ways:
9696 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9699 Lack of BER streaming support means one pass streaming processing is
9700 only supported if data is detached: setting the streaming flag is
9701 ignored for embedded content.
9703 CMS support is disabled by default and must be explicitly enabled
9704 with the enable-cms configuration option.
9708 * Update the GMP engine glue to do direct copies between BIGNUM and
9709 mpz_t when openssl and GMP use the same limb size. Otherwise the
9710 existing "conversion via a text string export" trick is still used.
9712 *Paul Sheer <paulsheer@gmail.com>*
9714 * Zlib compression BIO. This is a filter BIO which compressed and
9715 uncompresses any data passed through it.
9719 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9720 RFC3394 compatible AES key wrapping.
9724 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9725 sets string data without copying. X509_ALGOR_set0() and
9726 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9727 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9728 from an X509_ATTRIBUTE structure optionally checking it occurs only
9729 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9734 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9735 to get the expected BN_FLG_CONSTTIME behavior.
9737 *Bodo Moeller (Google)*
9741 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9742 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9743 - added some more tests to do_tests.pl
9744 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9745 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9746 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9747 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9748 - various changes to netware.pl to enable gcc-cross builds on Win32
9750 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9751 - various changes to fix missing prototype warnings
9752 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9753 - added AES, WHIRLPOOL and CPUID assembler code to build files
9754 - added missing AES assembler make rules to mk1mf.pl
9755 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9757 *Guenter Knauf <eflash@gmx.net>*
9759 * Implement certificate status request TLS extension defined in RFC3546.
9760 A client can set the appropriate parameters and receive the encoded
9761 OCSP response via a callback. A server can query the supplied parameters
9762 and set the encoded OCSP response in the callback. Add simplified examples
9763 to s_client and s_server.
9767 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9770 + Binary incompatibility of ssl_ctx_st structure
9771 + DTLS interoperation with non-compliant servers
9772 + Don't call get_session_cb() without proposed session
9773 + Fix ia64 assembler code
9775 *Andy Polyakov, Steve Henson*
9777 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9779 * DTLS Handshake overhaul. There were longstanding issues with
9780 OpenSSL DTLS implementation, which were making it impossible for
9781 RFC 4347 compliant client to communicate with OpenSSL server.
9782 Unfortunately just fixing these incompatibilities would "cut off"
9783 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9784 server keeps tolerating non RFC compliant syntax. The opposite is
9785 not true, 0.9.8f client can not communicate with earlier server.
9786 This update even addresses CVE-2007-4995.
9790 * Changes to avoid need for function casts in OpenSSL: some compilers
9791 (gcc 4.2 and later) reject their use.
9792 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9795 * Add RFC4507 support to OpenSSL. This includes the corrections in
9796 RFC4507bis. The encrypted ticket format is an encrypted encoded
9797 SSL_SESSION structure, that way new session features are automatically
9800 If a client application caches session in an SSL_SESSION structure
9801 support is transparent because tickets are now stored in the encoded
9804 The SSL_CTX structure automatically generates keys for ticket
9805 protection in servers so again support should be possible
9806 with no application modification.
9808 If a client or server wishes to disable RFC4507 support then the option
9809 SSL_OP_NO_TICKET can be set.
9811 Add a TLS extension debugging callback to allow the contents of any client
9812 or server extensions to be examined.
9814 This work was sponsored by Google.
9818 * Add initial support for TLS extensions, specifically for the server_name
9819 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9820 have new members for a host name. The SSL data structure has an
9821 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9822 stored in that context to allow for session resumption, even after the
9823 SSL has been switched to a new SSL_CTX in reaction to a client's
9824 server_name extension.
9826 New functions (subject to change):
9828 SSL_get_servername()
9829 SSL_get_servername_type()
9832 New CTRL codes and macros (subject to change):
9834 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9835 - SSL_CTX_set_tlsext_servername_callback()
9836 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9837 - SSL_CTX_set_tlsext_servername_arg()
9838 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9840 openssl s_client has a new '-servername ...' option.
9842 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9843 '-key2 ...', '-servername_fatal' (subject to change). This allows
9844 testing the HostName extension for a specific single host name ('-cert'
9845 and '-key' remain fallbacks for handshakes without HostName
9846 negotiation). If the unrecognized_name alert has to be sent, this by
9847 default is a warning; it becomes fatal with the '-servername_fatal'
9850 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9852 * Add AES and SSE2 assembly language support to VC++ build.
9856 * Mitigate attack on final subtraction in Montgomery reduction.
9860 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9861 (which previously caused an internal error).
9865 * Squeeze another 10% out of IGE mode when in != out.
9869 * AES IGE mode speedup.
9871 *Dean Gaudet (Google)*
9873 * Add the Korean symmetric 128-bit cipher SEED (see
9874 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9875 add SEED ciphersuites from RFC 4162:
9877 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9878 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9879 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9880 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9882 To minimize changes between patchlevels in the OpenSSL 0.9.8
9883 series, SEED remains excluded from compilation unless OpenSSL
9884 is configured with 'enable-seed'.
9886 *KISA, Bodo Moeller*
9888 * Mitigate branch prediction attacks, which can be practical if a
9889 single processor is shared, allowing a spy process to extract
9890 information. For detailed background information, see
9891 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9892 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9893 and Necessary Software Countermeasures"). The core of the change
9894 are new versions BN_div_no_branch() and
9895 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9896 respectively, which are slower, but avoid the security-relevant
9897 conditional branches. These are automatically called by BN_div()
9898 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9899 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9900 remove a conditional branch.
9902 BN_FLG_CONSTTIME is the new name for the previous
9903 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9904 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9905 in the exponent causes BN_mod_exp_mont() to use the alternative
9906 implementation in BN_mod_exp_mont_consttime().) The old name
9907 remains as a deprecated alias.
9909 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9910 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9911 constant-time implementations for more than just exponentiation.
9912 Here too the old name is kept as a deprecated alias.
9914 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9915 the BN_BLINDING structure gets an independent copy of the
9916 modulus. This means that the previous `BIGNUM *m` argument to
9917 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9918 essentially becomes `const BIGNUM *m`, although we can't actually
9919 change this in the header file before 0.9.9. It allows
9920 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9921 enable BN_FLG_CONSTTIME.
9923 *Matthew D Wood (Intel Corp)*
9925 * In the SSL/TLS server implementation, be strict about session ID
9926 context matching (which matters if an application uses a single
9927 external cache for different purposes). Previously,
9928 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9929 set. This did ensure strict client verification, but meant that,
9930 with applications using a single external cache for quite
9931 different requirements, clients could circumvent ciphersuite
9932 restrictions for a given session ID context by starting a session
9933 in a different context.
9937 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9938 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9939 authentication-only ciphersuites.
9943 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9944 not complete and could lead to a possible single byte overflow
9945 ([CVE-2007-5135]) [Ben Laurie]
9947 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9949 * Since AES128 and AES256 (and similarly Camellia128 and
9950 Camellia256) share a single mask bit in the logic of
9951 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9952 kludge to work properly if AES128 is available and AES256 isn't
9953 (or if Camellia128 is available and Camellia256 isn't).
9957 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9958 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9959 When a point or a seed is encoded in a BIT STRING, we need to
9960 prevent the removal of trailing zero bits to get the proper DER
9961 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9962 of a NamedBitList, for which trailing 0 bits need to be removed.)
9966 * Have SSL/TLS server implementation tolerate "mismatched" record
9967 protocol version while receiving ClientHello even if the
9968 ClientHello is fragmented. (The server can't insist on the
9969 particular protocol version it has chosen before the ServerHello
9970 message has informed the client about his choice.)
9974 * Add RFC 3779 support.
9976 *Rob Austein for ARIN, Ben Laurie*
9978 * Load error codes if they are not already present instead of using a
9979 static variable. This allows them to be cleanly unloaded and reloaded.
9980 Improve header file function name parsing.
9984 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9985 or CAPABILITY handshake as required by RFCs.
9989 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9991 * Introduce limits to prevent malicious keys being able to
9992 cause a denial of service. ([CVE-2006-2940])
9994 *Steve Henson, Bodo Moeller*
9996 * Fix ASN.1 parsing of certain invalid structures that can result
9997 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9999 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10000 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10002 * Fix SSL client code which could crash if connecting to a
10003 malicious SSLv2 server. ([CVE-2006-4343])
10005 *Tavis Ormandy and Will Drewry, Google Security Team*
10007 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10008 match only those. Before that, "AES256-SHA" would be interpreted
10009 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10010 the same strength classification in 0.9.7h) as we currently only
10011 have a single AES bit in the ciphersuite description bitmap.
10012 That change, however, also applied to ciphersuite strings such as
10013 "RC4-MD5" that intentionally matched multiple ciphersuites --
10014 namely, SSL 2.0 ciphersuites in addition to the more common ones
10015 from SSL 3.0/TLS 1.0.
10017 So we change the selection algorithm again: Naming an explicit
10018 ciphersuite selects this one ciphersuite, and any other similar
10019 ciphersuite (same bitmap) from *other* protocol versions.
10020 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10021 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10023 Since SSL 2.0 does not have any ciphersuites for which the
10024 128/256 bit distinction would be relevant, this works for now.
10025 The proper fix will be to use different bits for AES128 and
10026 AES256, which would have avoided the problems from the beginning;
10027 however, bits are scarce, so we can only do this in a new release
10028 (not just a patchlevel) when we can change the SSL_CIPHER
10029 definition to split the single 'unsigned long mask' bitmap into
10030 multiple values to extend the available space.
10034 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10036 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10037 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10039 * Add AES IGE and biIGE modes.
10043 * Change the Unix randomness entropy gathering to use poll() when
10044 possible instead of select(), since the latter has some
10045 undesirable limitations.
10047 *Darryl Miles via Richard Levitte and Bodo Moeller*
10049 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10050 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10051 cannot be implicitly activated as part of, e.g., the "AES" alias.
10052 However, please upgrade to OpenSSL 0.9.9[-dev] for
10053 non-experimental use of the ECC ciphersuites to get TLS extension
10054 support, which is required for curve and point format negotiation
10055 to avoid potential handshake problems.
10059 * Disable rogue ciphersuites:
10061 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10062 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10063 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10065 The latter two were purportedly from
10066 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10069 Also deactivate the remaining ciphersuites from
10070 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10071 unofficial, and the ID has long expired.
10075 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10076 dual-core machines) and other potential thread-safety issues.
10080 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10081 versions), which is now available for royalty-free use
10082 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10083 Also, add Camellia TLS ciphersuites from RFC 4132.
10085 To minimize changes between patchlevels in the OpenSSL 0.9.8
10086 series, Camellia remains excluded from compilation unless OpenSSL
10087 is configured with 'enable-camellia'.
10091 * Disable the padding bug check when compression is in use. The padding
10092 bug check assumes the first packet is of even length, this is not
10093 necessarily true if compression is enabled and can result in false
10094 positives causing handshake failure. The actual bug test is ancient
10095 code so it is hoped that implementations will either have fixed it by
10096 now or any which still have the bug do not support compression.
10100 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10102 * When applying a cipher rule check to see if string match is an explicit
10103 cipher suite and only match that one cipher suite if it is.
10107 * Link in manifests for VC++ if needed.
10109 *Austin Ziegler <halostatue@gmail.com>*
10111 * Update support for ECC-based TLS ciphersuites according to
10112 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10113 TLS extensions, which are supported starting with the 0.9.9
10114 branch, not in the OpenSSL 0.9.8 branch).
10118 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10119 opaque EVP_CIPHER_CTX handling.
10123 * Fixes and enhancements to zlib compression code. We now only use
10124 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10125 to conform with the standards mentioned here:
10126 <http://www.zlib.net/DLL_FAQ.txt>
10127 Static zlib linking now works on Windows and the new --with-zlib-include
10128 --with-zlib-lib options to Configure can be used to supply the location
10129 of the headers and library. Gracefully handle case where zlib library
10134 * Several fixes and enhancements to the OID generation code. The old code
10135 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10136 handle numbers larger than ULONG_MAX, truncated printing and had a
10137 non standard OBJ_obj2txt() behaviour.
10141 * Add support for building of engines under engine/ as shared libraries
10142 under VC++ build system.
10146 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10147 Hopefully, we will not see any false combination of paths any more.
10151 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10153 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10154 (part of SSL_OP_ALL). This option used to disable the
10155 countermeasure against man-in-the-middle protocol-version
10156 rollback in the SSL 2.0 server implementation, which is a bad
10157 idea. ([CVE-2005-2969])
10159 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10160 for Information Security, National Institute of Advanced Industrial
10161 Science and Technology [AIST], Japan)*
10163 * Add two function to clear and return the verify parameter flags.
10167 * Keep cipherlists sorted in the source instead of sorting them at
10168 runtime, thus removing the need for a lock.
10172 * Avoid some small subgroup attacks in Diffie-Hellman.
10174 *Nick Mathewson and Ben Laurie*
10176 * Add functions for well-known primes.
10180 * Extended Windows CE support.
10182 *Satoshi Nakamura and Andy Polyakov*
10184 * Initialize SSL_METHOD structures at compile time instead of during
10185 runtime, thus removing the need for a lock.
10189 * Make PKCS7_decrypt() work even if no certificate is supplied by
10190 attempting to decrypt each encrypted key in turn. Add support to
10195 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10197 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10200 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10204 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10205 key into the same file any more.
10209 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10213 * Add -utf8 command line and config file option to 'ca'.
10215 *Stefan <stf@udoma.org*
10217 * Removed the macro des_crypt(), as it seems to conflict with some
10218 libraries. Use DES_crypt().
10222 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10223 involves renaming the source and generated shared-libs for
10224 both. The engines will accept the corrected or legacy ids
10225 ('ncipher' and '4758_cca' respectively) when binding. NB,
10226 this only applies when building 'shared'.
10228 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10230 * Add attribute functions to EVP_PKEY structure. Modify
10231 PKCS12_create() to recognize a CSP name attribute and
10232 use it. Make -CSP option work again in pkcs12 utility.
10236 * Add new functionality to the bn blinding code:
10237 - automatic re-creation of the BN_BLINDING parameters after
10238 a fixed number of uses (currently 32)
10239 - add new function for parameter creation
10240 - introduce flags to control the update behaviour of the
10241 BN_BLINDING parameters
10242 - hide BN_BLINDING structure
10243 Add a second BN_BLINDING slot to the RSA structure to improve
10244 performance when a single RSA object is shared among several
10249 * Add support for DTLS.
10251 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10253 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10254 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10258 * Remove buggy and incomplete DH cert support from
10259 ssl/ssl_rsa.c and ssl/s3_both.c
10263 * Use SHA-1 instead of MD5 as the default digest algorithm for
10264 the `apps/openssl` commands.
10268 * Compile clean with "-Wall -Wmissing-prototypes
10269 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10270 DEBUG_SAFESTACK must also be set.
10274 * Change ./Configure so that certain algorithms can be disabled by default.
10275 The new counterpiece to "no-xxx" is "enable-xxx".
10277 The patented RC5 and MDC2 algorithms will now be disabled unless
10278 "enable-rc5" and "enable-mdc2", respectively, are specified.
10280 (IDEA remains enabled despite being patented. This is because IDEA
10281 is frequently required for interoperability, and there is no license
10282 fee for non-commercial use. As before, "no-idea" can be used to
10283 avoid this algorithm.)
10287 * Add processing of proxy certificates (see RFC 3820). This work was
10288 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10289 EGEE (Enabling Grids for E-science in Europe).
10293 * RC4 performance overhaul on modern architectures/implementations, such
10294 as Intel P4, IA-64 and AMD64.
10298 * New utility extract-section.pl. This can be used specify an alternative
10299 section number in a pod file instead of having to treat each file as
10300 a separate case in Makefile. This can be done by adding two lines to the
10303 =for comment openssl_section:XXX
10305 The blank line is mandatory.
10309 * New arguments -certform, -keyform and -pass for s_client and s_server
10310 to allow alternative format key and certificate files and passphrase
10315 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10316 update associated structures and add various utility functions.
10318 Add new policy related verify parameters, include policy checking in
10319 standard verify code. Enhance 'smime' application with extra parameters
10320 to support policy checking and print out.
10324 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10325 Nehemiah processors. These extensions support AES encryption in hardware
10326 as well as RNG (though RNG support is currently disabled).
10328 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10330 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10334 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10336 *Andy Polyakov and a number of other people*
10338 * Improved PowerPC platform support. Most notably BIGNUM assembler
10339 implementation contributed by IBM.
10341 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10343 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10344 exponent rather than 'unsigned long'. There is a corresponding change to
10345 the new 'rsa_keygen' element of the RSA_METHOD structure.
10347 *Jelte Jansen, Geoff Thorpe*
10349 * Functionality for creating the initial serial number file is now
10350 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10352 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10353 number file to 1, which is bound to cause problems. To avoid
10354 the problems while respecting compatibility between different 0.9.7
10355 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10356 CA.pl for serial number initialization. With the new release 0.9.8,
10357 we can fix the problem directly in the 'ca' utility.)
10361 * Reduced header interdependencies by declaring more opaque objects in
10362 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10363 give fewer recursive includes, which could break lazy source code - so
10364 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10365 developers should define this symbol when building and using openssl to
10366 ensure they track the recommended behaviour, interfaces, [etc], but
10367 backwards-compatible behaviour prevails when this isn't defined.
10371 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10375 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10376 This will generate a random key of the appropriate length based on the
10377 cipher context. The EVP_CIPHER can provide its own random key generation
10378 routine to support keys of a specific form. This is used in the des and
10379 3des routines to generate a key of the correct parity. Update S/MIME
10380 code to use new functions and hence generate correct parity DES keys.
10381 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10382 valid (weak or incorrect parity).
10386 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10387 as looking them up. This is useful when the verified structure may contain
10388 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10389 present unless the new PKCS7_NO_CRL flag is asserted.
10393 * Extend ASN1 oid configuration module. It now additionally accepts the
10396 shortName = some long name, 1.2.3.4
10400 * Reimplemented the BN_CTX implementation. There is now no more static
10401 limitation on the number of variables it can handle nor the depth of the
10402 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10403 information can now expand as required, and rather than having a single
10404 static array of bignums, BN_CTX now uses a linked-list of such arrays
10405 allowing it to expand on demand whilst maintaining the usefulness of
10406 BN_CTX's "bundling".
10410 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10411 to allow all RSA operations to function using a single BN_CTX.
10415 * Preliminary support for certificate policy evaluation and checking. This
10416 is initially intended to pass the tests outlined in "Conformance Testing
10417 of Relying Party Client Certificate Path Processing Logic" v1.07.
10421 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10422 remained unused and not that useful. A variety of other little bignum
10423 tweaks and fixes have also been made continuing on from the audit (see
10428 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10429 associated ASN1, EVP and SSL functions and old ASN1 macros.
10433 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10434 and this should never fail. So the return value from the use of
10435 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10436 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10440 * BN_CTX_get() should return zero-valued bignums, providing the same
10441 initialised value as BN_new().
10443 *Geoff Thorpe, suggested by Ulf Möller*
10445 * Support for inhibitAnyPolicy certificate extension.
10449 * An audit of the BIGNUM code is underway, for which debugging code is
10450 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10451 is considered valid when processing BIGNUMs, and causes execution to
10452 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10453 further steps are taken to deliberately pollute unused data in BIGNUM
10454 structures to try and expose faulty code further on. For now, openssl will
10455 (in its default mode of operation) continue to tolerate the inconsistent
10456 forms that it has tolerated in the past, but authors and packagers should
10457 consider trying openssl and their own applications when compiled with
10458 these debugging symbols defined. It will help highlight potential bugs in
10459 their own code, and will improve the test coverage for OpenSSL itself. At
10460 some point, these tighter rules will become openssl's default to improve
10461 maintainability, though the assert()s and other overheads will remain only
10462 in debugging configurations. See bn.h for more details.
10464 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10466 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10467 that can only be obtained through BN_CTX_new() (which implicitly
10468 initialises it). The presence of this function only made it possible
10469 to overwrite an existing structure (and cause memory leaks).
10473 * Because of the callback-based approach for implementing LHASH as a
10474 template type, lh_insert() adds opaque objects to hash-tables and
10475 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10476 to clean up those corresponding objects before destroying the hash table
10477 (and losing the object pointers). So some over-zealous constifications in
10478 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10479 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10480 prototyped to have "const" restrictions on the object pointers they are
10481 given (and so aren't required to cast them away any more).
10485 * The tmdiff.h API was so ugly and minimal that our own timing utility
10486 (speed) prefers to use its own implementation. The two implementations
10487 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10488 its object type properly exposed (MS_TM) instead of casting to/from
10489 `char *`. This may still change yet if someone realises MS_TM and
10491 aren't necessarily the greatest nomenclatures - but this is what was used
10492 internally to the implementation so I've used that for now.
10496 * Ensure that deprecated functions do not get compiled when
10497 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10498 the self-tests were still using deprecated key-generation functions so
10499 these have been updated also.
10503 * Reorganise PKCS#7 code to separate the digest location functionality
10504 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10505 New function PKCS7_set_digest() to set the digest type for PKCS#7
10506 digestedData type. Add additional code to correctly generate the
10507 digestedData type and add support for this type in PKCS7 initialization
10512 * New function PKCS7_set0_type_other() this initializes a PKCS7
10513 structure of type "other".
10517 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10518 sure the loop does correctly stop and breaking ("division by zero")
10519 modulus operations are not performed. The (pre-generated) prime
10520 table crypto/bn/bn_prime.h was already correct, but it could not be
10521 re-generated on some platforms because of the "division by zero"
10522 situation in the script.
10524 *Ralf S. Engelschall*
10526 * Update support for ECC-based TLS ciphersuites according to
10527 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10528 SHA-1 now is only used for "small" curves (where the
10529 representation of a field element takes up to 24 bytes); for
10530 larger curves, the field element resulting from ECDH is directly
10531 used as premaster secret.
10533 *Douglas Stebila (Sun Microsystems Laboratories)*
10535 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10536 curve secp160r1 to the tests.
10538 *Douglas Stebila (Sun Microsystems Laboratories)*
10540 * Add the possibility to load symbols globally with DSO.
10542 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10544 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10545 control of the error stack.
10549 * Add support for STORE in ENGINE.
10553 * Add the STORE type. The intention is to provide a common interface
10554 to certificate and key stores, be they simple file-based stores, or
10555 HSM-type store, or LDAP stores, or...
10556 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10560 * Add a generic structure called OPENSSL_ITEM. This can be used to
10561 pass a list of arguments to any function as well as provide a way
10562 for a function to pass data back to the caller.
10566 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10567 works like BUF_strdup() but can be used to duplicate a portion of
10568 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10573 * Add the function sk_find_ex() which works like sk_find(), but will
10574 return an index to an element even if an exact match couldn't be
10575 found. The index is guaranteed to point at the element where the
10576 searched-for key would be inserted to preserve sorting order.
10580 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10581 takes an extra flags argument for optional functionality. Currently,
10582 the following flags are defined:
10584 OBJ_BSEARCH_VALUE_ON_NOMATCH
10585 This one gets OBJ_bsearch_ex() to return a pointer to the first
10586 element where the comparing function returns a negative or zero
10589 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10590 This one gets OBJ_bsearch_ex() to return a pointer to the first
10591 element where the comparing function returns zero. This is useful
10592 if there are more than one element where the comparing function
10597 * Make it possible to create self-signed certificates with 'openssl ca'
10598 in such a way that the self-signed certificate becomes part of the
10599 CA database and uses the same mechanisms for serial number generation
10600 as all other certificate signing. The new flag '-selfsign' enables
10601 this functionality. Adapt CA.sh and CA.pl.in.
10605 * Add functionality to check the public key of a certificate request
10606 against a given private. This is useful to check that a certificate
10607 request can be signed by that key (self-signing).
10611 * Make it possible to have multiple active certificates with the same
10612 subject in the CA index file. This is done only if the keyword
10613 'unique_subject' is set to 'no' in the main CA section (default
10614 if 'CA_default') of the configuration file. The value is saved
10615 with the database itself in a separate index attribute file,
10616 named like the index file with '.attr' appended to the name.
10620 * Generate multi-valued AVAs using '+' notation in config files for
10625 * Support for nameConstraints certificate extension.
10629 * Support for policyConstraints certificate extension.
10633 * Support for policyMappings certificate extension.
10637 * Make sure the default DSA_METHOD implementation only uses its
10638 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10639 and change its own handlers to be NULL so as to remove unnecessary
10640 indirection. This lets alternative implementations fallback to the
10641 default implementation more easily.
10645 * Support for directoryName in GeneralName related extensions
10650 * Make it possible to link applications using Makefile.shared.
10651 Make that possible even when linking against static libraries!
10655 * Support for single pass processing for S/MIME signing. This now
10656 means that S/MIME signing can be done from a pipe, in addition
10657 cleartext signing (multipart/signed type) is effectively streaming
10658 and the signed data does not need to be all held in memory.
10660 This is done with a new flag PKCS7_STREAM. When this flag is set
10661 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10662 is done after the data is output (and digests calculated) in
10663 SMIME_write_PKCS7().
10667 * Add full support for -rpath/-R, both in shared libraries and
10668 applications, at least on the platforms where it's known how
10673 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10674 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10675 will now compute a table of multiples of the generator that
10676 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10677 faster (notably in the case of a single point multiplication,
10678 scalar * generator).
10680 *Nils Larsch, Bodo Moeller*
10682 * IPv6 support for certificate extensions. The various extensions
10683 which use the IP:a.b.c.d can now take IPv6 addresses using the
10684 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10689 * Added an ENGINE that implements RSA by performing private key
10690 exponentiations with the GMP library. The conversions to and from
10691 GMP's mpz_t format aren't optimised nor are any montgomery forms
10692 cached, and on x86 it appears OpenSSL's own performance has caught up.
10693 However there are likely to be other architectures where GMP could
10694 provide a boost. This ENGINE is not built in by default, but it can be
10695 specified at Configure time and should be accompanied by the necessary
10696 linker additions, eg;
10697 ./config -DOPENSSL_USE_GMP -lgmp
10701 * "openssl engine" will not display ENGINE/DSO load failure errors when
10702 testing availability of engines with "-t" - the old behaviour is
10703 produced by increasing the feature's verbosity with "-tt".
10707 * ECDSA routines: under certain error conditions uninitialized BN objects
10708 could be freed. Solution: make sure initialization is performed early
10709 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10714 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10715 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10716 software implementations. For DSA and DH, parameter generation can
10717 also be overridden by providing the appropriate method callbacks.
10721 * Change the "progress" mechanism used in key-generation and
10722 primality testing to functions that take a new BN_GENCB pointer in
10723 place of callback/argument pairs. The new API functions have `_ex`
10724 postfixes and the older functions are reimplemented as wrappers for
10725 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10726 declarations of the old functions to help (graceful) attempts to
10727 migrate to the new functions. Also, the new key-generation API
10728 functions operate on a caller-supplied key-structure and return
10729 success/failure rather than returning a key or NULL - this is to
10730 help make "keygen" another member function of RSA_METHOD etc.
10732 Example for using the new callback interface:
10734 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10735 void *my_arg = ...;
10738 BN_GENCB_set(&my_cb, my_callback, my_arg);
10740 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10741 /* For the meaning of a, b in calls to my_callback(), see the
10742 * documentation of the function that calls the callback.
10743 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10744 * my_callback should return 1 if it wants BN_is_prime_ex()
10745 * to continue, or 0 to stop.
10750 * Change the ZLIB compression method to be stateful, and make it
10751 available to TLS with the number defined in
10752 draft-ietf-tls-compression-04.txt.
10756 * Add the ASN.1 structures and functions for CertificatePair, which
10757 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10759 CertificatePair ::= SEQUENCE {
10760 forward [0] Certificate OPTIONAL,
10761 reverse [1] Certificate OPTIONAL,
10762 -- at least one of the pair shall be present -- }
10764 Also implement the PEM functions to read and write certificate
10765 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10767 This needed to be defined, mostly for the sake of the LDAP
10768 attribute crossCertificatePair, but may prove useful elsewhere as
10773 * Make it possible to inhibit symlinking of shared libraries in
10774 Makefile.shared, for Cygwin's sake.
10778 * Extend the BIGNUM API by creating a function
10779 void BN_set_negative(BIGNUM *a, int neg);
10780 and a macro that behave like
10781 int BN_is_negative(const BIGNUM *a);
10783 to avoid the need to access 'a->neg' directly in applications.
10787 * Implement fast modular reduction for pseudo-Mersenne primes
10788 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10789 EC_GROUP_new_curve_GFp() will now automatically use this
10792 *Nils Larsch <nla@trustcenter.de>*
10794 * Add new lock type (CRYPTO_LOCK_BN).
10798 * Change the ENGINE framework to automatically load engines
10799 dynamically from specific directories unless they could be
10800 found to already be built in or loaded. Move all the
10801 current engines except for the cryptodev one to a new
10802 directory engines/.
10803 The engines in engines/ are built as shared libraries if
10804 the "shared" options was given to ./Configure or ./config.
10805 Otherwise, they are inserted in libcrypto.a.
10806 /usr/local/ssl/engines is the default directory for dynamic
10807 engines, but that can be overridden at configure time through
10808 the usual use of --prefix and/or --openssldir, and at run
10809 time with the environment variable OPENSSL_ENGINES.
10811 *Geoff Thorpe and Richard Levitte*
10813 * Add Makefile.shared, a helper makefile to build shared
10814 libraries. Adapt Makefile.org.
10818 * Add version info to Win32 DLLs.
10820 *Peter 'Luna' Runestig" <peter@runestig.com>*
10822 * Add new 'medium level' PKCS#12 API. Certificates and keys
10823 can be added using this API to created arbitrary PKCS#12
10824 files while avoiding the low level API.
10826 New options to PKCS12_create(), key or cert can be NULL and
10827 will then be omitted from the output file. The encryption
10828 algorithm NIDs can be set to -1 for no encryption, the mac
10829 iteration count can be set to 0 to omit the mac.
10831 Enhance pkcs12 utility by making the -nokeys and -nocerts
10832 options work when creating a PKCS#12 file. New option -nomac
10833 to omit the mac, NONE can be set for an encryption algorithm.
10834 New code is modified to use the enhanced PKCS12_create()
10835 instead of the low level API.
10839 * Extend ASN1 encoder to support indefinite length constructed
10840 encoding. This can output sequences tags and octet strings in
10841 this form. Modify pk7_asn1.c to support indefinite length
10842 encoding. This is experimental and needs additional code to
10843 be useful, such as an ASN1 bio and some enhanced streaming
10846 Extend template encode functionality so that tagging is passed
10847 down to the template encoder.
10851 * Let 'openssl req' fail if an argument to '-newkey' is not
10852 recognized instead of using RSA as a default.
10856 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10857 As these are not official, they are not included in "ALL";
10858 the "ECCdraft" ciphersuite group alias can be used to select them.
10860 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10862 * Add ECDH engine support.
10864 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10866 * Add ECDH in new directory crypto/ecdh/.
10868 *Douglas Stebila (Sun Microsystems Laboratories)*
10870 * Let BN_rand_range() abort with an error after 100 iterations
10871 without success (which indicates a broken PRNG).
10875 * Change BN_mod_sqrt() so that it verifies that the input value
10876 is really the square of the return value. (Previously,
10877 BN_mod_sqrt would show GIGO behaviour.)
10881 * Add named elliptic curves over binary fields from X9.62, SECG,
10882 and WAP/WTLS; add OIDs that were still missing.
10884 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10886 * Extend the EC library for elliptic curves over binary fields
10887 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10890 EC_GF2m_simple_method
10894 EC_GROUP_new_curve_GF2m
10895 EC_GROUP_set_curve_GF2m
10896 EC_GROUP_get_curve_GF2m
10897 EC_POINT_set_affine_coordinates_GF2m
10898 EC_POINT_get_affine_coordinates_GF2m
10899 EC_POINT_set_compressed_coordinates_GF2m
10901 Point compression for binary fields is disabled by default for
10902 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10905 As binary polynomials are represented as BIGNUMs, various members
10906 of the EC_GROUP and EC_POINT data structures can be shared
10907 between the implementations for prime fields and binary fields;
10908 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10909 are essentially identical to their `..._GFp` counterparts.
10910 (For simplicity, the `..._GFp` prefix has been dropped from
10911 various internal method names.)
10913 An internal 'field_div' method (similar to 'field_mul' and
10914 'field_sqr') has been added; this is used only for binary fields.
10916 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10918 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10919 through methods ('mul', 'precompute_mult').
10921 The generic implementations (now internally called 'ec_wNAF_mul'
10922 and 'ec_wNAF_precomputed_mult') remain the default if these
10923 methods are undefined.
10925 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10927 * New function EC_GROUP_get_degree, which is defined through
10928 EC_METHOD. For curves over prime fields, this returns the bit
10929 length of the modulus.
10931 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10933 * New functions EC_GROUP_dup, EC_POINT_dup.
10934 (These simply call ..._new and ..._copy).
10936 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10938 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10939 Polynomials are represented as BIGNUMs (where the sign bit is not
10940 used) in the following functions [macros]:
10943 BN_GF2m_sub [= BN_GF2m_add]
10944 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10945 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10946 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10948 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10949 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10950 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10951 BN_GF2m_cmp [= BN_ucmp]
10953 (Note that only the 'mod' functions are actually for fields GF(2^m).
10954 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10956 For some functions, an the irreducible polynomial defining a
10957 field can be given as an 'unsigned int[]' with strictly
10958 decreasing elements giving the indices of those bits that are set;
10959 i.e., p[] represents the polynomial
10960 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10962 p[0] > p[1] > ... > p[k] = 0.
10963 This applies to the following functions:
10966 BN_GF2m_mod_mul_arr
10967 BN_GF2m_mod_sqr_arr
10968 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10969 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10970 BN_GF2m_mod_exp_arr
10971 BN_GF2m_mod_sqrt_arr
10972 BN_GF2m_mod_solve_quad_arr
10976 Conversion can be performed by the following functions:
10981 bntest.c has additional tests for binary polynomial arithmetic.
10983 Two implementations for BN_GF2m_mod_div() are available.
10984 The default algorithm simply uses BN_GF2m_mod_inv() and
10985 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10986 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10987 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10989 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10991 * Add new error code 'ERR_R_DISABLED' that can be used when some
10992 functionality is disabled at compile-time.
10994 *Douglas Stebila <douglas.stebila@sun.com>*
10996 * Change default behaviour of 'openssl asn1parse' so that more
10997 information is visible when viewing, e.g., a certificate:
10999 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11000 mode the content of non-printable OCTET STRINGs is output in a
11001 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11002 avoid the appearance of a printable string.
11004 *Nils Larsch <nla@trustcenter.de>*
11006 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11008 EC_GROUP_set_asn1_flag()
11009 EC_GROUP_get_asn1_flag()
11010 EC_GROUP_set_point_conversion_form()
11011 EC_GROUP_get_point_conversion_form()
11012 These control ASN1 encoding details:
11013 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11014 has been set to OPENSSL_EC_NAMED_CURVE.
11015 - Points are encoded in uncompressed form by default; options for
11016 asn1_for are as for point2oct, namely
11017 POINT_CONVERSION_COMPRESSED
11018 POINT_CONVERSION_UNCOMPRESSED
11019 POINT_CONVERSION_HYBRID
11021 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11023 EC_GROUP_set_seed()
11024 EC_GROUP_get0_seed()
11025 EC_GROUP_get_seed_len()
11026 This is used only for ASN1 purposes (so far).
11028 *Nils Larsch <nla@trustcenter.de>*
11030 * Add 'field_type' member to EC_METHOD, which holds the NID
11031 of the appropriate field type OID. The new function
11032 EC_METHOD_get_field_type() returns this value.
11034 *Nils Larsch <nla@trustcenter.de>*
11037 EC_POINT_point2bn()
11038 EC_POINT_bn2point()
11039 EC_POINT_point2hex()
11040 EC_POINT_hex2point()
11041 providing useful interfaces to EC_POINT_point2oct() and
11042 EC_POINT_oct2point().
11044 *Nils Larsch <nla@trustcenter.de>*
11046 * Change internals of the EC library so that the functions
11047 EC_GROUP_set_generator()
11048 EC_GROUP_get_generator()
11049 EC_GROUP_get_order()
11050 EC_GROUP_get_cofactor()
11051 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11052 to methods, which would lead to unnecessary code duplication when
11053 adding different types of curves.
11055 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11057 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11058 arithmetic, and such that modified wNAFs are generated
11059 (which avoid length expansion in many cases).
11063 * Add a function EC_GROUP_check_discriminant() (defined via
11064 EC_METHOD) that verifies that the curve discriminant is non-zero.
11066 Add a function EC_GROUP_check() that makes some sanity tests
11067 on a EC_GROUP, its generator and order. This includes
11068 EC_GROUP_check_discriminant().
11070 *Nils Larsch <nla@trustcenter.de>*
11072 * Add ECDSA in new directory crypto/ecdsa/.
11074 Add applications 'openssl ecparam' and 'openssl ecdsa'
11075 (these are based on 'openssl dsaparam' and 'openssl dsa').
11077 ECDSA support is also included in various other files across the
11078 library. Most notably,
11079 - 'openssl req' now has a '-newkey ecdsa:file' option;
11080 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11081 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11082 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11083 them suitable for ECDSA where domain parameters must be
11084 extracted before the specific public key;
11085 - ECDSA engine support has been added.
11087 *Nils Larsch <nla@trustcenter.de>*
11089 * Include some named elliptic curves, and add OIDs from X9.62,
11090 SECG, and WAP/WTLS. Each curve can be obtained from the new
11092 EC_GROUP_new_by_curve_name(),
11093 and the list of available named curves can be obtained with
11094 EC_get_builtin_curves().
11095 Also add a 'curve_name' member to EC_GROUP objects, which can be
11097 EC_GROUP_set_curve_name()
11098 EC_GROUP_get_curve_name()
11100 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11102 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11103 was actually never needed) and in BN_mul(). The removal in BN_mul()
11104 required a small change in bn_mul_part_recursive() and the addition
11105 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11106 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11107 bn_sub_words() and bn_add_words() except they take arrays with
11112 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11114 * Cleanse PEM buffers before freeing them since they may contain
11117 *Benjamin Bennett <ben@psc.edu>*
11119 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11120 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11121 authentication-only ciphersuites.
11125 * Since AES128 and AES256 share a single mask bit in the logic of
11126 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11127 kludge to work properly if AES128 is available and AES256 isn't.
11131 * Expand security boundary to match 1.1.1 module.
11135 * Remove redundant features: hash file source, editing of test vectors
11136 modify fipsld to use external fips_premain.c signature.
11140 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11141 run algorithm test programs.
11145 * Make algorithm test programs more tolerant of whitespace.
11149 * Have SSL/TLS server implementation tolerate "mismatched" record
11150 protocol version while receiving ClientHello even if the
11151 ClientHello is fragmented. (The server can't insist on the
11152 particular protocol version it has chosen before the ServerHello
11153 message has informed the client about his choice.)
11157 * Load error codes if they are not already present instead of using a
11158 static variable. This allows them to be cleanly unloaded and reloaded.
11162 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11164 * Introduce limits to prevent malicious keys being able to
11165 cause a denial of service. ([CVE-2006-2940])
11167 *Steve Henson, Bodo Moeller*
11169 * Fix ASN.1 parsing of certain invalid structures that can result
11170 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11172 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11173 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11175 * Fix SSL client code which could crash if connecting to a
11176 malicious SSLv2 server. ([CVE-2006-4343])
11178 *Tavis Ormandy and Will Drewry, Google Security Team*
11180 * Change ciphersuite string processing so that an explicit
11181 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11182 will no longer include "AES128-SHA"), and any other similar
11183 ciphersuite (same bitmap) from *other* protocol versions (so that
11184 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11185 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11186 changes from 0.9.8b and 0.9.8d.
11190 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11192 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11193 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11195 * Change the Unix randomness entropy gathering to use poll() when
11196 possible instead of select(), since the latter has some
11197 undesirable limitations.
11199 *Darryl Miles via Richard Levitte and Bodo Moeller*
11201 * Disable rogue ciphersuites:
11203 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11204 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11205 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11207 The latter two were purportedly from
11208 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11211 Also deactivate the remaining ciphersuites from
11212 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11213 unofficial, and the ID has long expired.
11217 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11218 dual-core machines) and other potential thread-safety issues.
11222 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11224 * Adapt fipsld and the build system to link against the validated FIPS
11225 module in FIPS mode.
11229 * Fixes for VC++ 2005 build under Windows.
11233 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11234 from a Windows bash shell such as MSYS. It is autodetected from the
11235 "config" script when run from a VC++ environment. Modify standard VC++
11236 build to use fipscanister.o from the GNU make build.
11240 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11242 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11243 The value now differs depending on if you build for FIPS or not.
11244 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11245 safely run with a non-FIPSed libcrypto, as it may crash because of
11246 the difference induced by this change.
11250 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11252 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11253 (part of SSL_OP_ALL). This option used to disable the
11254 countermeasure against man-in-the-middle protocol-version
11255 rollback in the SSL 2.0 server implementation, which is a bad
11256 idea. ([CVE-2005-2969])
11258 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11259 for Information Security, National Institute of Advanced Industrial
11260 Science and Technology [AIST, Japan)]*
11262 * Minimal support for X9.31 signatures and PSS padding modes. This is
11263 mainly for FIPS compliance and not fully integrated at this stage.
11267 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11268 the exponentiation using a fixed-length exponent. (Otherwise,
11269 the information leaked through timing could expose the secret key
11270 after many signatures; cf. Bleichenbacher's attack on DSA with
11275 * Make a new fixed-window mod_exp implementation the default for
11276 RSA, DSA, and DH private-key operations so that the sequence of
11277 squares and multiplies and the memory access pattern are
11278 independent of the particular secret key. This will mitigate
11279 cache-timing and potential related attacks.
11281 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11282 and this is automatically used by BN_mod_exp_mont() if the new flag
11283 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11284 will use this BN flag for private exponents unless the flag
11285 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11286 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11288 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11290 * Change the client implementation for SSLv23_method() and
11291 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11292 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11293 (Previously, the SSL 2.0 backwards compatible Client Hello
11294 message format would be used even with SSL_OP_NO_SSLv2.)
11298 * Add support for smime-type MIME parameter in S/MIME messages which some
11303 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11304 a threadsafe manner. Modify rsa code to use new function and add calls
11305 to dsa and dh code (which had race conditions before).
11309 * Include the fixed error library code in the C error file definitions
11310 instead of fixing them up at runtime. This keeps the error code
11311 structures constant.
11315 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11317 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11320 * Fixes for newer kerberos headers. NB: the casts are needed because
11321 the 'length' field is signed on one version and unsigned on another
11322 with no (?) obvious way to tell the difference, without these VC++
11323 complains. Also the "definition" of FAR (blank) is no longer included
11324 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11325 some needed definitions.
11329 * Undo Cygwin change.
11333 * Added support for proxy certificates according to RFC 3820.
11334 Because they may be a security thread to unaware applications,
11335 they must be explicitly allowed in run-time. See
11336 docs/HOWTO/proxy_certificates.txt for further information.
11340 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11342 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11343 server and client random values. Previously
11344 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11345 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11347 This change has negligible security impact because:
11349 1. Server and client random values still have 24 bytes of pseudo random
11352 2. Server and client random values are sent in the clear in the initial
11355 3. The master secret is derived using the premaster secret (48 bytes in
11356 size for static RSA ciphersuites) as well as client server and random
11359 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11362 *Stephen Henson, reported by UK NISCC*
11364 * Use Windows randomness collection on Cygwin.
11368 * Fix hang in EGD/PRNGD query when communication socket is closed
11369 prematurely by EGD/PRNGD.
11371 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11373 * Prompt for pass phrases when appropriate for PKCS12 input format.
11377 * Back-port of selected performance improvements from development
11378 branch, as well as improved support for PowerPC platforms.
11382 * Add lots of checks for memory allocation failure, error codes to indicate
11383 failure and freeing up memory if a failure occurs.
11385 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11387 * Add new -passin argument to dgst.
11391 * Perform some character comparisons of different types in X509_NAME_cmp:
11392 this is needed for some certificates that re-encode DNs into UTF8Strings
11393 (in violation of RFC3280) and can't or won't issue name rollover
11398 * Make an explicit check during certificate validation to see that
11399 the CA setting in each certificate on the chain is correct. As a
11400 side effect always do the following basic checks on extensions,
11401 not just when there's an associated purpose to the check:
11403 - if there is an unhandled critical extension (unless the user
11404 has chosen to ignore this fault)
11405 - if the path length has been exceeded (if one is set at all)
11406 - that certain extensions fit the associated purpose (if one has
11411 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11413 * Avoid a race condition when CRLs are checked in a multi threaded
11414 environment. This would happen due to the reordering of the revoked
11415 entries during signature checking and serial number lookup. Now the
11416 encoding is cached and the serial number sort performed under a lock.
11417 Add new STACK function sk_is_sorted().
11421 * Add Delta CRL to the extension code.
11425 * Various fixes to s3_pkt.c so alerts are sent properly.
11427 *David Holmes <d.holmes@f5.com>*
11429 * Reduce the chances of duplicate issuer name and serial numbers (in
11430 violation of RFC3280) using the OpenSSL certificate creation utilities.
11431 This is done by creating a random 64 bit value for the initial serial
11432 number when a serial number file is created or when a self signed
11433 certificate is created using 'openssl req -x509'. The initial serial
11434 number file is created using 'openssl x509 -next_serial' in CA.pl
11435 rather than being initialized to 1.
11439 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11441 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11442 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11444 *Joe Orton, Steve Henson*
11446 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11449 *Joe Orton, Steve Henson*
11451 * Make it possible to have multiple active certificates with the same
11452 subject in the CA index file. This is done only if the keyword
11453 'unique_subject' is set to 'no' in the main CA section (default
11454 if 'CA_default') of the configuration file. The value is saved
11455 with the database itself in a separate index attribute file,
11456 named like the index file with '.attr' appended to the name.
11460 * X509 verify fixes. Disable broken certificate workarounds when
11461 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11462 keyUsage extension present. Don't accept CRLs with unhandled critical
11463 extensions: since verify currently doesn't process CRL extensions this
11464 rejects a CRL with *any* critical extensions. Add new verify error codes
11469 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11470 A clarification of RFC2560 will require the use of OCTET STRINGs and
11471 some implementations cannot handle the current raw format. Since OpenSSL
11472 copies and compares OCSP nonces as opaque blobs without any attempt at
11473 parsing them this should not create any compatibility issues.
11477 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11478 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11479 this HMAC (and other) operations are several times slower than OpenSSL
11484 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11486 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11488 * Use the correct content when signing type "other".
11492 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11494 * Fix various bugs revealed by running the NISCC test suite:
11496 Stop out of bounds reads in the ASN1 code when presented with
11497 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11499 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11501 If verify callback ignores invalid public key errors don't try to check
11502 certificate signature with the NULL public key.
11506 * New -ignore_err option in ocsp application to stop the server
11507 exiting on the first error in a request.
11511 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11512 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11517 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11518 extra data after the compression methods not only for TLS 1.0
11519 but also for SSL 3.0 (as required by the specification).
11521 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11523 * Change X509_certificate_type() to mark the key as exported/exportable
11524 when it's 512 *bits* long, not 512 bytes.
11528 * Change AES_cbc_encrypt() so it outputs exact multiple of
11529 blocks during encryption.
11533 * Various fixes to base64 BIO and non blocking I/O. On write
11534 flushes were not handled properly if the BIO retried. On read
11535 data was not being buffered properly and had various logic bugs.
11536 This also affects blocking I/O when the data being decoded is a
11541 * Various S/MIME bugfixes and compatibility changes:
11542 output correct application/pkcs7 MIME type if
11543 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11544 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11545 of files as .eml work). Correctly handle very long lines in MIME
11550 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11552 * Countermeasure against the Klima-Pokorny-Rosa extension of
11553 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11554 a protocol version number mismatch like a decryption error
11555 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11559 * Turn on RSA blinding by default in the default implementation
11560 to avoid a timing attack. Applications that don't want it can call
11561 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11562 They would be ill-advised to do so in most cases.
11564 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11566 * Change RSA blinding code so that it works when the PRNG is not
11567 seeded (in this case, the secret RSA exponent is abused as
11568 an unpredictable seed -- if it is not unpredictable, there
11569 is no point in blinding anyway). Make RSA blinding thread-safe
11570 by remembering the creator's thread ID in rsa->blinding and
11571 having all other threads use local one-time blinding factors
11572 (this requires more computation than sharing rsa->blinding, but
11573 avoids excessive locking; and if an RSA object is not shared
11574 between threads, blinding will still be very fast).
11578 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11579 ENGINE as defaults for all supported algorithms irrespective of
11580 the 'flags' parameter. 'flags' is now honoured, so applications
11581 should make sure they are passing it correctly.
11585 * Target "mingw" now allows native Windows code to be generated in
11586 the Cygwin environment as well as with the MinGW compiler.
11590 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11592 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11593 via timing by performing a MAC computation even if incorrect
11594 block cipher padding has been found. This is a countermeasure
11595 against active attacks where the attacker has to distinguish
11596 between bad padding and a MAC verification error. ([CVE-2003-0078])
11598 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11599 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11600 Martin Vuagnoux (EPFL, Ilion)*
11602 * Make the no-err option work as intended. The intention with no-err
11603 is not to have the whole error stack handling routines removed from
11604 libcrypto, it's only intended to remove all the function name and
11605 reason texts, thereby removing some of the footprint that may not
11606 be interesting if those errors aren't displayed anyway.
11608 NOTE: it's still possible for any application or module to have its
11609 own set of error texts inserted. The routines are there, just not
11610 used by default when no-err is given.
11614 * Add support for FreeBSD on IA64.
11616 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11618 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11619 Kerberos function mit_des_cbc_cksum(). Before this change,
11620 the value returned by DES_cbc_cksum() was like the one from
11621 mit_des_cbc_cksum(), except the bytes were swapped.
11623 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11625 * Allow an application to disable the automatic SSL chain building.
11626 Before this a rather primitive chain build was always performed in
11627 ssl3_output_cert_chain(): an application had no way to send the
11628 correct chain if the automatic operation produced an incorrect result.
11630 Now the chain builder is disabled if either:
11632 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11634 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11636 The reasoning behind this is that an application would not want the
11637 auto chain building to take place if extra chain certificates are
11638 present and it might also want a means of sending no additional
11639 certificates (for example the chain has two certificates and the
11644 * Add the possibility to build without the ENGINE framework.
11646 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11648 * Under Win32 gmtime() can return NULL: check return value in
11649 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11653 * DSA routines: under certain error conditions uninitialized BN objects
11654 could be freed. Solution: make sure initialization is performed early
11655 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11656 Nils Larsch <nla@trustcenter.de> via PR#459)
11660 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11661 checked on reconnect on the client side, therefore session resumption
11662 could still fail with a "ssl session id is different" error. This
11663 behaviour is masked when SSL_OP_ALL is used due to
11664 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11665 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11666 followup to PR #377.
11670 * IA-32 assembler support enhancements: unified ELF targets, support
11671 for SCO/Caldera platforms, fix for Cygwin shared build.
11675 * Add support for FreeBSD on sparc64. As a consequence, support for
11676 FreeBSD on non-x86 processors is separate from x86 processors on
11677 the config script, much like the NetBSD support.
11679 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11681 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11683 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11686 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11687 code (06) was taken as the first octet of the session ID and the last
11688 octet was ignored consequently. As a result SSLv2 client side session
11689 caching could not have worked due to the session ID mismatch between
11691 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11696 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11697 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11702 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11703 seems that in spite of existing for more than a year, many application
11704 author have done nothing to provide the necessary callbacks, which
11705 means that this particular engine will not work properly anywhere.
11706 This is a very unfortunate situation which forces us, in the name
11707 of usability, to give the hw_ncipher.c a static lock, which is part
11709 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11710 appear in 0.9.8 or later. We EXPECT application authors to have
11711 dealt properly with this when 0.9.8 is released (unless we actually
11712 make such changes in the libcrypto locking code that changes will
11713 have to be made anyway).
11717 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11718 octets have been read, EOF or an error occurs. Without this change
11719 some truncated ASN1 structures will not produce an error.
11723 * Disable Heimdal support, since it hasn't been fully implemented.
11724 Still give the possibility to force the use of Heimdal, but with
11725 warnings and a request that patches get sent to openssl-dev.
11729 * Add the VC-CE target, introduce the WINCE sysname, and add
11730 INSTALL.WCE and appropriate conditionals to make it build.
11732 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11734 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11735 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11736 edit numbers of the version.
11738 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11740 * Introduce safe string copy and catenation functions
11741 (BUF_strlcpy() and BUF_strlcat()).
11743 *Ben Laurie (CHATS) and Richard Levitte*
11745 * Avoid using fixed-size buffers for one-line DNs.
11747 *Ben Laurie (CHATS)*
11749 * Add BUF_MEM_grow_clean() to avoid information leakage when
11750 resizing buffers containing secrets, and use where appropriate.
11752 *Ben Laurie (CHATS)*
11754 * Avoid using fixed size buffers for configuration file location.
11756 *Ben Laurie (CHATS)*
11758 * Avoid filename truncation for various CA files.
11760 *Ben Laurie (CHATS)*
11762 * Use sizeof in preference to magic numbers.
11764 *Ben Laurie (CHATS)*
11766 * Avoid filename truncation in cert requests.
11768 *Ben Laurie (CHATS)*
11770 * Add assertions to check for (supposedly impossible) buffer
11773 *Ben Laurie (CHATS)*
11775 * Don't cache truncated DNS entries in the local cache (this could
11776 potentially lead to a spoofing attack).
11778 *Ben Laurie (CHATS)*
11780 * Fix various buffers to be large enough for hex/decimal
11781 representations in a platform independent manner.
11783 *Ben Laurie (CHATS)*
11785 * Add CRYPTO_realloc_clean() to avoid information leakage when
11786 resizing buffers containing secrets, and use where appropriate.
11788 *Ben Laurie (CHATS)*
11790 * Add BIO_indent() to avoid much slightly worrying code to do
11793 *Ben Laurie (CHATS)*
11795 * Convert sprintf()/BIO_puts() to BIO_printf().
11797 *Ben Laurie (CHATS)*
11799 * buffer_gets() could terminate with the buffer only half
11802 *Ben Laurie (CHATS)*
11804 * Add assertions to prevent user-supplied crypto functions from
11805 overflowing internal buffers by having large block sizes, etc.
11807 *Ben Laurie (CHATS)*
11809 * New OPENSSL_assert() macro (similar to assert(), but enabled
11812 *Ben Laurie (CHATS)*
11814 * Eliminate unused copy of key in RC4.
11816 *Ben Laurie (CHATS)*
11818 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11820 *Ben Laurie (CHATS)*
11822 * Fix off-by-one error in EGD path.
11824 *Ben Laurie (CHATS)*
11826 * If RANDFILE path is too long, ignore instead of truncating.
11828 *Ben Laurie (CHATS)*
11830 * Eliminate unused and incorrectly sized X.509 structure
11833 *Ben Laurie (CHATS)*
11835 * Eliminate unused and dangerous function knumber().
11837 *Ben Laurie (CHATS)*
11839 * Eliminate unused and dangerous structure, KSSL_ERR.
11841 *Ben Laurie (CHATS)*
11843 * Protect against overlong session ID context length in an encoded
11844 session object. Since these are local, this does not appear to be
11847 *Ben Laurie (CHATS)*
11849 * Change from security patch (see 0.9.6e below) that did not affect
11850 the 0.9.6 release series:
11852 Remote buffer overflow in SSL3 protocol - an attacker could
11853 supply an oversized master key in Kerberos-enabled versions.
11856 *Ben Laurie (CHATS)*
11858 * Change the SSL kerb5 codes to match RFC 2712.
11862 * Make -nameopt work fully for req and add -reqopt switch.
11864 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11866 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11868 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11870 * Make sure tests can be performed even if the corresponding algorithms
11871 have been removed entirely. This was also the last step to make
11872 OpenSSL compilable with DJGPP under all reasonable conditions.
11874 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11876 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11877 to allow version independent disabling of normally unselected ciphers,
11878 which may be activated as a side-effect of selecting a single cipher.
11880 (E.g., cipher list string "RSA" enables ciphersuites that are left
11881 out of "ALL" because they do not provide symmetric encryption.
11882 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11884 *Lutz Jaenicke, Bodo Moeller*
11886 * Add appropriate support for separate platform-dependent build
11887 directories. The recommended way to make a platform-dependent
11888 build directory is the following (tested on Linux), maybe with
11891 # Place yourself outside of the OpenSSL source tree. In
11892 # this example, the environment variable OPENSSL_SOURCE
11893 # is assumed to contain the absolute OpenSSL source directory.
11894 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11895 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11896 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11897 mkdir -p `dirname $F`
11898 ln -s $OPENSSL_SOURCE/$F $F
11901 To be absolutely sure not to disturb the source tree, a "make clean"
11902 is a good thing. If it isn't successful, don't worry about it,
11903 it probably means the source directory is very clean.
11907 * Make sure any ENGINE control commands make local copies of string
11908 pointers passed to them whenever necessary. Otherwise it is possible
11909 the caller may have overwritten (or deallocated) the original string
11910 data when a later ENGINE operation tries to use the stored values.
11912 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11914 * Improve diagnostics in file reading and command-line digests.
11916 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11918 * Add AES modes CFB and OFB to the object database. Correct an
11919 error in AES-CFB decryption.
11923 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11924 allows existing EVP_CIPHER_CTX structures to be reused after
11925 calling `EVP_*Final()`. This behaviour is used by encryption
11926 BIOs and some applications. This has the side effect that
11927 applications must explicitly clean up cipher contexts with
11928 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11932 * Check the values of dna and dnb in bn_mul_recursive before calling
11933 bn_mul_comba (a non zero value means the a or b arrays do not contain
11934 n2 elements) and fallback to bn_mul_normal if either is not zero.
11938 * Fix escaping of non-ASCII characters when using the -subj option
11939 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11943 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11944 form for "surname", serialNumber has no short form.
11945 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11946 therefore remove "mail" short name for "internet 7".
11947 The OID for unique identifiers in X509 certificates is
11948 x500UniqueIdentifier, not uniqueIdentifier.
11949 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11953 * Add an "init" command to the ENGINE config module and auto initialize
11954 ENGINEs. Without any "init" command the ENGINE will be initialized
11955 after all ctrl commands have been executed on it. If init=1 the
11956 ENGINE is initialized at that point (ctrls before that point are run
11957 on the uninitialized ENGINE and after on the initialized one). If
11958 init=0 then the ENGINE will not be initialized at all.
11962 * Fix the 'app_verify_callback' interface so that the user-defined
11963 argument is actually passed to the callback: In the
11964 SSL_CTX_set_cert_verify_callback() prototype, the callback
11965 declaration has been changed from
11968 int (*cb)(X509_STORE_CTX *,void *);
11969 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11970 i=s->ctx->app_verify_callback(&ctx)
11971 has been changed into
11972 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11974 To update applications using SSL_CTX_set_cert_verify_callback(),
11975 a dummy argument can be added to their callback functions.
11977 *D. K. Smetters <smetters@parc.xerox.com>*
11979 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11981 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11983 * Add and OPENSSL_LOAD_CONF define which will cause
11984 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11985 This allows older applications to transparently support certain
11986 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11987 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11988 load the config file and OPENSSL_add_all_algorithms_conf() which will
11989 always load it have also been added.
11993 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11994 Adjust NIDs and EVP layer.
11996 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11998 * Config modules support in openssl utility.
12000 Most commands now load modules from the config file,
12001 though in a few (such as version) this isn't done
12002 because it couldn't be used for anything.
12004 In the case of ca and req the config file used is
12005 the same as the utility itself: that is the -config
12006 command line option can be used to specify an
12011 * Move default behaviour from OPENSSL_config(). If appname is NULL
12012 use "openssl_conf" if filename is NULL use default openssl config file.
12016 * Add an argument to OPENSSL_config() to allow the use of an alternative
12017 config section name. Add a new flag to tolerate a missing config file
12018 and move code to CONF_modules_load_file().
12022 * Support for crypto accelerator cards from Accelerated Encryption
12023 Processing, www.aep.ie. (Use engine 'aep')
12024 The support was copied from 0.9.6c [engine] and adapted/corrected
12025 to work with the new engine framework.
12027 *AEP Inc. and Richard Levitte*
12029 * Support for SureWare crypto accelerator cards from Baltimore
12030 Technologies. (Use engine 'sureware')
12031 The support was copied from 0.9.6c [engine] and adapted
12032 to work with the new engine framework.
12036 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12037 make the newer ENGINE framework commands for the CHIL engine work.
12039 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12041 * Make it possible to produce shared libraries on ReliantUNIX.
12043 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12045 * Add the configuration target debug-linux-ppro.
12046 Make 'openssl rsa' use the general key loading routines
12047 implemented in `apps.c`, and make those routines able to
12048 handle the key format FORMAT_NETSCAPE and the variant
12051 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12053 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12055 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12057 * Add -keyform to rsautl, and document -engine.
12059 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12061 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12062 BIO_R_NO_SUCH_FILE error code rather than the generic
12063 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12067 * Add new functions
12068 ERR_peek_last_error
12069 ERR_peek_last_error_line
12070 ERR_peek_last_error_line_data.
12071 These are similar to
12073 ERR_peek_error_line
12074 ERR_peek_error_line_data,
12075 but report on the latest error recorded rather than the first one
12076 still in the error queue.
12078 *Ben Laurie, Bodo Moeller*
12080 * default_algorithms option in ENGINE config module. This allows things
12082 default_algorithms = ALL
12083 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12087 * Preliminary ENGINE config module.
12091 * New experimental application configuration code.
12095 * Change the AES code to follow the same name structure as all other
12096 symmetric ciphers, and behave the same way. Move everything to
12097 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12099 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12101 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12103 *Ben Laurie and Theo de Raadt*
12105 * Add option to output public keys in req command.
12107 *Massimiliano Pala madwolf@openca.org*
12109 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12110 (up to about 10% better than before for P-192 and P-224).
12114 * New functions/macros
12116 SSL_CTX_set_msg_callback(ctx, cb)
12117 SSL_CTX_set_msg_callback_arg(ctx, arg)
12118 SSL_set_msg_callback(ssl, cb)
12119 SSL_set_msg_callback_arg(ssl, arg)
12121 to request calling a callback function
12123 void cb(int write_p, int version, int content_type,
12124 const void *buf, size_t len, SSL *ssl, void *arg)
12126 whenever a protocol message has been completely received
12127 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12128 protocol version according to which the SSL library interprets
12129 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12130 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12131 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12132 specification (change_cipher_spec(20), alert(21), handshake(22)).
12133 'buf' and 'len' point to the actual message, 'ssl' to the
12134 SSL object, and 'arg' is the application-defined value set by
12135 SSL[_CTX]_set_msg_callback_arg().
12137 'openssl s_client' and 'openssl s_server' have new '-msg' options
12138 to enable a callback that displays all protocol messages.
12142 * Change the shared library support so shared libraries are built as
12143 soon as the corresponding static library is finished, and thereby get
12144 openssl and the test programs linked against the shared library.
12145 This still only happens when the keyword "shard" has been given to
12146 the configuration scripts.
12148 NOTE: shared library support is still an experimental thing, and
12149 backward binary compatibility is still not guaranteed.
12151 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12153 * Add support for Subject Information Access extension.
12155 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12157 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12158 additional bytes when new memory had to be allocated, not just
12159 when reusing an existing buffer.
12163 * New command line and configuration option 'utf8' for the req command.
12164 This allows field values to be specified as UTF8 strings.
12168 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12169 runs for the former and machine-readable output for the latter.
12173 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12174 of the e-mail address in the DN (i.e., it will go into a certificate
12175 extension only). The new configuration file option 'email_in_dn = no'
12176 has the same effect.
12178 *Massimiliano Pala madwolf@openca.org*
12180 * Change all functions with names starting with `des_` to be starting
12181 with `DES_` instead. Add wrappers that are compatible with libdes,
12182 but are named `_ossl_old_des_*`. Finally, add macros that map the
12183 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12184 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12185 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12188 Since we provide two compatibility mappings, the user needs to
12189 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12190 compatibility is desired. The default (i.e., when that macro
12191 isn't defined) is OpenSSL 0.9.6c compatibility.
12193 There are also macros that enable and disable the support of old
12194 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12195 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12196 are defined, the default will apply: to support the old des routines.
12198 In either case, one must include openssl/des.h to get the correct
12199 definitions. Do not try to just include openssl/des_old.h, that
12202 NOTE: This is a major break of an old API into a new one. Software
12203 authors are encouraged to switch to the `DES_` style functions. Some
12204 time in the future, des_old.h and the libdes compatibility functions
12205 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12206 default), and then completely removed.
12210 * Test for certificates which contain unsupported critical extensions.
12211 If such a certificate is found during a verify operation it is
12212 rejected by default: this behaviour can be overridden by either
12213 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12214 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12215 X509_supported_extension() has also been added which returns 1 if a
12216 particular extension is supported.
12220 * Modify the behaviour of EVP cipher functions in similar way to digests
12221 to retain compatibility with existing code.
12225 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12226 compatibility with existing code. In particular the 'ctx' parameter does
12227 not have to be to be initialized before the call to EVP_DigestInit() and
12228 it is tidied up after a call to EVP_DigestFinal(). New function
12229 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12230 EVP_MD_CTX_copy() changed to not require the destination to be
12231 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12232 requires the destination to be valid.
12234 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12235 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12239 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12240 so that complete 'Handshake' protocol structures are kept in memory
12241 instead of overwriting 'msg_type' and 'length' with 'body' data.
12245 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12247 *Massimo Santin via Richard Levitte*
12249 * Major restructuring to the underlying ENGINE code. This includes
12250 reduction of linker bloat, separation of pure "ENGINE" manipulation
12251 (initialisation, etc) from functionality dealing with implementations
12252 of specific crypto interfaces. This change also introduces integrated
12253 support for symmetric ciphers and digest implementations - so ENGINEs
12254 can now accelerate these by providing EVP_CIPHER and EVP_MD
12255 implementations of their own. This is detailed in
12256 [crypto/engine/README.md](crypto/engine/README.md)
12257 as it couldn't be adequately described here. However, there are a few
12258 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12259 were changed in the original introduction of ENGINE code have now
12260 reverted back - the hooking from this code to ENGINE is now a good
12261 deal more passive and at run-time, operations deal directly with
12262 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12263 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12264 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12265 they were not being used by the framework as there is no concept of a
12266 BIGNUM_METHOD and they could not be generalised to the new
12267 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12268 ENGINE_cpy() has been removed as it cannot be consistently defined in
12273 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12277 * Change mkdef.pl to sort symbols that get the same entry number,
12278 and make sure the automatically generated functions `ERR_load_*`
12279 become part of libeay.num as well.
12283 * New function SSL_renegotiate_pending(). This returns true once
12284 renegotiation has been requested (either SSL_renegotiate() call
12285 or HelloRequest/ClientHello received from the peer) and becomes
12286 false once a handshake has been completed.
12287 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12288 sends a HelloRequest, but does not ensure that a handshake takes
12289 place. SSL_renegotiate_pending() is useful for checking if the
12290 client has followed the request.)
12294 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12295 By default, clients may request session resumption even during
12296 renegotiation (if session ID contexts permit); with this option,
12297 session resumption is possible only in the first handshake.
12299 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12300 more bits available for options that should not be part of
12301 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12305 * Add some demos for certificate and certificate request creation.
12309 * Make maximum certificate chain size accepted from the peer application
12310 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12311 "Douglas E. Engert" <deengert@anl.gov>.
12315 * Add support for shared libraries for Unixware-7
12316 (Boyd Lynn Gerber <gerberb@zenez.com>).
12320 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12321 be done prior to destruction. Use this to unload error strings from
12322 ENGINEs that load their own error strings. NB: This adds two new API
12323 functions to "get" and "set" this destroy handler in an ENGINE.
12327 * Alter all existing ENGINE implementations (except "openssl" and
12328 "openbsd") to dynamically instantiate their own error strings. This
12329 makes them more flexible to be built both as statically-linked ENGINEs
12330 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12331 Also, add stub code to each that makes building them as self-contained
12332 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12336 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12337 implementations into applications that are completely implemented in
12338 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12339 commands that can be used to configure what shared-library to load and
12340 to control aspects of the way it is handled. Also, made an update to
12341 the [README-Engine.md](README-Engine.md) file
12342 that brings its information up-to-date and
12343 provides some information and instructions on the "dynamic" ENGINE
12344 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12348 * Make it possible to unload ranges of ERR strings with a new
12349 "ERR_unload_strings" function.
12353 * Add a copy() function to EVP_MD.
12357 * Make EVP_MD routines take a context pointer instead of just the
12358 md_data void pointer.
12362 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12363 that the digest can only process a single chunk of data
12364 (typically because it is provided by a piece of
12365 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12366 is only going to provide a single chunk of data, and hence the
12367 framework needn't accumulate the data for oneshot drivers.
12371 * As with "ERR", make it possible to replace the underlying "ex_data"
12372 functions. This change also alters the storage and management of global
12373 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12374 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12375 index counters. The API functions that use this state have been changed
12376 to take a "class_index" rather than pointers to the class's local STACK
12377 and counter, and there is now an API function to dynamically create new
12378 classes. This centralisation allows us to (a) plug a lot of the
12379 thread-safety problems that existed, and (b) makes it possible to clean
12380 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12381 such data would previously have always leaked in application code and
12382 workarounds were in place to make the memory debugging turn a blind eye
12383 to it. Application code that doesn't use this new function will still
12384 leak as before, but their memory debugging output will announce it now
12385 rather than letting it slide.
12387 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12388 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12389 has a return value to indicate success or failure.
12393 * Make it possible to replace the underlying "ERR" functions such that the
12394 global state (2 LHASH tables and 2 locks) is only used by the "default"
12395 implementation. This change also adds two functions to "get" and "set"
12396 the implementation prior to it being automatically set the first time
12397 any other ERR function takes place. Ie. an application can call "get",
12398 pass the return value to a module it has just loaded, and that module
12399 can call its own "set" function using that value. This means the
12400 module's "ERR" operations will use (and modify) the error state in the
12401 application and not in its own statically linked copy of OpenSSL code.
12405 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12406 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12407 the operation, and provides a more encapsulated way for external code
12408 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12409 to use these functions rather than manually incrementing the counts.
12411 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12415 * Add EVP test program.
12419 * Add symmetric cipher support to ENGINE. Expect the API to change!
12423 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12424 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12425 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12426 These allow a CRL to be built without having to access X509_CRL fields
12427 directly. Modify 'ca' application to use new functions.
12431 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12432 bug workarounds. Rollback attack detection is a security feature.
12433 The problem will only arise on OpenSSL servers when TLSv1 is not
12434 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12435 Software authors not wanting to support TLSv1 will have special reasons
12436 for their choice and can explicitly enable this option.
12438 *Bodo Moeller, Lutz Jaenicke*
12440 * Rationalise EVP so it can be extended: don't include a union of
12441 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12442 (similar to those existing for EVP_CIPHER_CTX).
12447 EVP_MD_CTX_init(&md); /* new function call */
12448 EVP_DigestInit(&md, EVP_sha1());
12449 EVP_DigestUpdate(&md, in, len);
12450 EVP_DigestFinal(&md, out, NULL);
12451 EVP_MD_CTX_cleanup(&md); /* new function call */
12455 * Make DES key schedule conform to the usual scheme, as well as
12456 correcting its structure. This means that calls to DES functions
12457 now have to pass a pointer to a des_key_schedule instead of a
12458 plain des_key_schedule (which was actually always a pointer
12461 des_key_schedule ks;
12463 des_set_key_checked(..., &ks);
12464 des_ncbc_encrypt(..., &ks, ...);
12466 (Note that a later change renames 'des_...' into 'DES_...'.)
12470 * Initial reduction of linker bloat: the use of some functions, such as
12471 PEM causes large amounts of unused functions to be linked in due to
12472 poor organisation. For example pem_all.c contains every PEM function
12473 which has a knock on effect of linking in large amounts of (unused)
12474 ASN1 code. Grouping together similar functions and splitting unrelated
12475 functions prevents this.
12479 * Cleanup of EVP macros.
12483 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12484 correct `_ecb suffix`.
12488 * Add initial OCSP responder support to ocsp application. The
12489 revocation information is handled using the text based index
12490 use by the ca application. The responder can either handle
12491 requests generated internally, supplied in files (for example
12492 via a CGI script) or using an internal minimal server.
12496 * Add configuration choices to get zlib compression for TLS.
12500 * Changes to Kerberos SSL for RFC 2712 compliance:
12501 1. Implemented real KerberosWrapper, instead of just using
12502 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12503 2. Implemented optional authenticator field of KerberosWrapper.
12505 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12506 and authenticator structs; see crypto/krb5/.
12508 Generalized Kerberos calls to support multiple Kerberos libraries.
12509 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12510 via Richard Levitte*
12512 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12513 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12514 values for each of the key sizes rather than having just
12515 parameters (and 'speed' generating keys each time).
12519 * Speed up EVP routines.
12522 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12523 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12524 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12525 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12527 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12528 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12529 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12532 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12534 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12538 * Added the OS2-EMX target.
12540 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12542 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12543 New functions to support `NCONF `routines in extension code.
12544 New function `CONF_set_nconf()`
12545 to allow functions which take an `NCONF` to also handle the old `LHASH`
12546 structure: this means that the old `CONF` compatible routines can be
12547 retained (in particular w.rt. extensions) without having to duplicate the
12548 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12552 * Enhance the general user interface with mechanisms for inner control
12553 and with possibilities to have yes/no kind of prompts.
12557 * Change all calls to low level digest routines in the library and
12558 applications to use EVP. Add missing calls to HMAC_cleanup() and
12559 don't assume HMAC_CTX can be copied using memcpy().
12561 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12563 * Add the possibility to control engines through control names but with
12564 arbitrary arguments instead of just a string.
12565 Change the key loaders to take a UI_METHOD instead of a callback
12566 function pointer. NOTE: this breaks binary compatibility with earlier
12567 versions of OpenSSL [engine].
12568 Adapt the nCipher code for these new conditions and add a card insertion
12573 * Enhance the general user interface with mechanisms to better support
12574 dialog box interfaces, application-defined prompts, the possibility
12575 to use defaults (for example default passwords from somewhere else)
12576 and interrupts/cancellations.
12580 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12581 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12585 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12586 tidy up some unnecessarily weird code in 'sk_new()').
12588 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12590 * Change the key loading routines for ENGINEs to use the same kind
12591 callback (pem_password_cb) as all other routines that need this
12596 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12597 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12598 than this minimum value is recommended.
12602 * New random seeder for OpenVMS, using the system process statistics
12603 that are easily reachable.
12607 * Windows apparently can't transparently handle global
12608 variables defined in DLLs. Initialisations such as:
12610 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12612 won't compile. This is used by the any applications that need to
12613 declare their own ASN1 modules. This was fixed by adding the option
12614 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12615 needed for static libraries under Win32.
12619 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12620 setting of purpose and trust fields. New X509_STORE trust and
12621 purpose functions and tidy up setting in other SSL functions.
12625 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12626 structure. These are inherited by X509_STORE_CTX when it is
12627 initialised. This allows various defaults to be set in the
12628 X509_STORE structure (such as flags for CRL checking and custom
12629 purpose or trust settings) for functions which only use X509_STORE_CTX
12630 internally such as S/MIME.
12632 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12633 trust settings if they are not set in X509_STORE. This allows X509_STORE
12634 purposes and trust (in S/MIME for example) to override any set by default.
12636 Add command line options for CRL checking to smime, s_client and s_server
12641 * Initial CRL based revocation checking. If the CRL checking flag(s)
12642 are set then the CRL is looked up in the X509_STORE structure and
12643 its validity and signature checked, then if the certificate is found
12644 in the CRL the verify fails with a revoked error.
12646 Various new CRL related callbacks added to X509_STORE_CTX structure.
12648 Command line options added to 'verify' application to support this.
12650 This needs some additional work, such as being able to handle multiple
12651 CRLs with different times, extension based lookup (rather than just
12652 by subject name) and ultimately more complete V2 CRL extension
12657 * Add a general user interface API (crypto/ui/). This is designed
12658 to replace things like des_read_password and friends (backward
12659 compatibility functions using this new API are provided).
12660 The purpose is to remove prompting functions from the DES code
12661 section as well as provide for prompting through dialog boxes in
12662 a window system and the like.
12666 * Add "ex_data" support to ENGINE so implementations can add state at a
12667 per-structure level rather than having to store it globally.
12671 * Make it possible for ENGINE structures to be copied when retrieved by
12672 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12673 This causes the "original" ENGINE structure to act like a template,
12674 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12675 operational state can be localised to each ENGINE structure, despite the
12676 fact they all share the same "methods". New ENGINE structures returned in
12677 this case have no functional references and the return value is the single
12678 structural reference. This matches the single structural reference returned
12679 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12684 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12685 needs to match any other type at all we need to manually clear the
12690 * Changes to the "openssl engine" utility to include;
12691 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12692 about an ENGINE's available control commands.
12693 - executing control commands from command line arguments using the
12694 '-pre' and '-post' switches. '-post' is only used if '-t' is
12695 specified and the ENGINE is successfully initialised. The syntax for
12696 the individual commands are colon-separated, for example;
12697 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12701 * New dynamic control command support for ENGINEs. ENGINEs can now
12702 declare their own commands (numbers), names (strings), descriptions,
12703 and input types for run-time discovery by calling applications. A
12704 subset of these commands are implicitly classed as "executable"
12705 depending on their input type, and only these can be invoked through
12706 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12707 can be based on user input, config files, etc). The distinction is
12708 that "executable" commands cannot return anything other than a boolean
12709 result and can only support numeric or string input, whereas some
12710 discoverable commands may only be for direct use through
12711 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12712 pointers, or other custom uses. The "executable" commands are to
12713 support parameterisations of ENGINE behaviour that can be
12714 unambiguously defined by ENGINEs and used consistently across any
12715 OpenSSL-based application. Commands have been added to all the
12716 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12717 control over shared-library paths without source code alterations.
12721 * Changed all ENGINE implementations to dynamically allocate their
12722 ENGINEs rather than declaring them statically. Apart from this being
12723 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12724 this also allows the implementations to compile without using the
12725 internal engine_int.h header.
12729 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12730 'const' value. Any code that should be able to modify a RAND_METHOD
12731 should already have non-const pointers to it (ie. they should only
12732 modify their own ones).
12736 * Made a variety of little tweaks to the ENGINE code.
12737 - "atalla" and "ubsec" string definitions were moved from header files
12738 to C code. "nuron" string definitions were placed in variables
12739 rather than hard-coded - allowing parameterisation of these values
12740 later on via ctrl() commands.
12741 - Removed unused "#if 0"'d code.
12742 - Fixed engine list iteration code so it uses ENGINE_free() to release
12743 structural references.
12744 - Constified the RAND_METHOD element of ENGINE structures.
12745 - Constified various get/set functions as appropriate and added
12746 missing functions (including a catch-all ENGINE_cpy that duplicates
12747 all ENGINE values onto a new ENGINE except reference counts/state).
12748 - Removed NULL parameter checks in get/set functions. Setting a method
12749 or function to NULL is a way of cancelling out a previously set
12750 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12751 and doesn't justify the extra error symbols and code.
12752 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12753 flags from engine_int.h to engine.h.
12754 - Changed prototypes for ENGINE handler functions (init(), finish(),
12755 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12759 * Implement binary inversion algorithm for BN_mod_inverse in addition
12760 to the algorithm using long division. The binary algorithm can be
12761 used only if the modulus is odd. On 32-bit systems, it is faster
12762 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12763 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12764 up to 450 bits. In 64-bit environments, the binary algorithm
12765 appears to be advantageous for much longer moduli; here we use it
12766 for moduli up to 2048 bits.
12770 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12771 could not support the combine flag in choice fields.
12775 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12776 extensions from a certificate request to the certificate.
12780 * Allow multiple 'certopt' and 'nameopt' options to be separated
12781 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12782 file: this allows the display of the certificate about to be
12783 signed to be customised, to allow certain fields to be included
12784 or excluded and extension details. The old system didn't display
12785 multicharacter strings properly, omitted fields not in the policy
12786 and couldn't display additional details such as extensions.
12790 * Function EC_POINTs_mul for multiple scalar multiplication
12791 of an arbitrary number of elliptic curve points
12792 \sum scalars[i]*points[i],
12793 optionally including the generator defined for the EC_GROUP:
12794 scalar*generator + \sum scalars[i]*points[i].
12796 EC_POINT_mul is a simple wrapper function for the typical case
12797 that the point list has just one item (besides the optional
12802 * First EC_METHODs for curves over GF(p):
12804 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12805 operations and provides various method functions that can also
12806 operate with faster implementations of modular arithmetic.
12808 EC_GFp_mont_method() reuses most functions that are part of
12809 EC_GFp_simple_method, but uses Montgomery arithmetic.
12811 *Bodo Moeller; point addition and point doubling
12812 implementation directly derived from source code provided by
12813 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12815 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12816 crypto/ec/ec_lib.c):
12818 Curves are EC_GROUP objects (with an optional group generator)
12819 based on EC_METHODs that are built into the library.
12821 Points are EC_POINT objects based on EC_GROUP objects.
12823 Most of the framework would be able to handle curves over arbitrary
12824 finite fields, but as there are no obvious types for fields other
12825 than GF(p), some functions are limited to that for now.
12829 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12830 that the file contains a complete HTTP response.
12834 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12835 change the def and num file printf format specifier from "%-40sXXX"
12836 to "%-39s XXX". The latter will always guarantee a space after the
12837 field while the former will cause them to run together if the field
12838 is 40 of more characters long.
12842 * Constify the cipher and digest 'method' functions and structures
12843 and modify related functions to take constant EVP_MD and EVP_CIPHER
12848 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12849 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12853 * Modify `EVP_Digest*()` routines so they now return values. Although the
12854 internal software routines can never fail additional hardware versions
12859 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12861 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12862 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12865 ERR_R_NESTED_ASN1_ERROR
12867 ERR_R_MISSING_ASN1_EOS
12868 were 4 .. 9, conflicting with
12869 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12871 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12872 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12874 Add new error code 'ERR_R_INTERNAL_ERROR'.
12878 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12883 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12884 sets the subject name for a new request or supersedes the
12885 subject name in a given request. Formats that can be parsed are
12886 'CN=Some Name, OU=myOU, C=IT'
12888 'CN=Some Name/OU=myOU/C=IT'.
12890 Add options '-batch' and '-verbose' to 'openssl req'.
12892 *Massimiliano Pala <madwolf@hackmasters.net>*
12894 * Introduce the possibility to access global variables through
12895 functions on platform were that's the best way to handle exporting
12896 global variables in shared libraries. To enable this functionality,
12897 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12898 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12899 is normally done by Configure or something similar).
12901 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12902 in the source file (foo.c) like this:
12904 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12905 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12907 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12908 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12910 OPENSSL_DECLARE_GLOBAL(int,foo);
12911 #define foo OPENSSL_GLOBAL_REF(foo)
12912 OPENSSL_DECLARE_GLOBAL(double,bar);
12913 #define bar OPENSSL_GLOBAL_REF(bar)
12915 The #defines are very important, and therefore so is including the
12916 header file everywhere where the defined globals are used.
12918 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12919 of ASN.1 items, but that structure is a bit different.
12921 The largest change is in util/mkdef.pl which has been enhanced with
12922 better and easier to understand logic to choose which symbols should
12923 go into the Windows .def files as well as a number of fixes and code
12924 cleanup (among others, algorithm keywords are now sorted
12925 lexicographically to avoid constant rewrites).
12929 * In BN_div() keep a copy of the sign of 'num' before writing the
12930 result to 'rm' because if rm==num the value will be overwritten
12931 and produce the wrong result if 'num' is negative: this caused
12932 problems with BN_mod() and BN_nnmod().
12936 * Function OCSP_request_verify(). This checks the signature on an
12937 OCSP request and verifies the signer certificate. The signer
12938 certificate is just checked for a generic purpose and OCSP request
12943 * Add OCSP_check_validity() function to check the validity of OCSP
12944 responses. OCSP responses are prepared in real time and may only
12945 be a few seconds old. Simply checking that the current time lies
12946 between thisUpdate and nextUpdate max reject otherwise valid responses
12947 caused by either OCSP responder or client clock inaccuracy. Instead
12948 we allow thisUpdate and nextUpdate to fall within a certain period of
12949 the current time. The age of the response can also optionally be
12950 checked. Two new options -validity_period and -status_age added to
12955 * If signature or public key algorithm is unrecognized print out its
12956 OID rather that just UNKNOWN.
12960 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12961 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12962 ID to be generated from the issuer certificate alone which can then be
12963 passed to OCSP_id_issuer_cmp().
12967 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12968 ASN1 modules to export functions returning ASN1_ITEM pointers
12969 instead of the ASN1_ITEM structures themselves. This adds several
12970 new macros which allow the underlying ASN1 function/structure to
12971 be accessed transparently. As a result code should not use ASN1_ITEM
12972 references directly (such as &X509_it) but instead use the relevant
12973 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12974 use of the new ASN1 code on platforms where exporting structures
12975 is problematical (for example in shared libraries) but exporting
12976 functions returning pointers to structures is not.
12980 * Add support for overriding the generation of SSL/TLS session IDs.
12981 These callbacks can be registered either in an SSL_CTX or per SSL.
12982 The purpose of this is to allow applications to control, if they wish,
12983 the arbitrary values chosen for use as session IDs, particularly as it
12984 can be useful for session caching in multiple-server environments. A
12985 command-line switch for testing this (and any client code that wishes
12986 to use such a feature) has been added to "s_server".
12988 *Geoff Thorpe, Lutz Jaenicke*
12990 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12991 of the form `#if defined(...) || defined(...) || ...` and
12992 `#if !defined(...) && !defined(...) && ...`. This also avoids
12993 the growing number of special cases it was previously handling.
12997 * Make all configuration macros available for application by making
12998 sure they are available in opensslconf.h, by giving them names starting
12999 with `OPENSSL_` to avoid conflicts with other packages and by making
13000 sure e_os2.h will cover all platform-specific cases together with
13002 Additionally, it is now possible to define configuration/platform-
13003 specific names (called "system identities"). In the C code, these
13004 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13005 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13006 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13011 * New option -set_serial to 'req' and 'x509' this allows the serial
13012 number to use to be specified on the command line. Previously self
13013 signed certificates were hard coded with serial number 0 and the
13014 CA options of 'x509' had to use a serial number in a file which was
13019 * New options to 'ca' utility to support V2 CRL entry extensions.
13020 Currently CRL reason, invalidity date and hold instruction are
13021 supported. Add new CRL extensions to V3 code and some new objects.
13025 * New function EVP_CIPHER_CTX_set_padding() this is used to
13026 disable standard block padding (aka PKCS#5 padding) in the EVP
13027 API, which was previously mandatory. This means that the data is
13028 not padded in any way and so the total length much be a multiple
13029 of the block size, otherwise an error occurs.
13033 * Initial (incomplete) OCSP SSL support.
13037 * New function OCSP_parse_url(). This splits up a URL into its host,
13038 port and path components: primarily to parse OCSP URLs. New -url
13039 option to ocsp utility.
13043 * New nonce behavior. The return value of OCSP_check_nonce() now
13044 reflects the various checks performed. Applications can decide
13045 whether to tolerate certain situations such as an absent nonce
13046 in a response when one was present in a request: the ocsp application
13047 just prints out a warning. New function OCSP_add1_basic_nonce()
13048 this is to allow responders to include a nonce in a response even if
13049 the request is nonce-less.
13053 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13054 skipped when using openssl x509 multiple times on a single input file,
13055 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13059 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13060 set string type: to handle setting ASN1_TIME structures. Fix ca
13061 utility to correctly initialize revocation date of CRLs.
13065 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13066 the clients preferred ciphersuites and rather use its own preferences.
13067 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13068 Internet Explorer by ensuring unchanged hash method during stepup.
13069 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13073 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13074 to aes and add a new 'exist' option to print out symbols that don't
13079 * Additional options to ocsp utility to allow flags to be set and
13080 additional certificates supplied.
13084 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13085 OCSP client a number of certificate to only verify the response
13090 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13091 handle the new API. Currently only ECB, CBC modes supported. Add new
13094 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13095 Encryption Standard (AES) Ciphersuites for Transport Layer
13096 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13097 not enabled by default and were not part of the "ALL" ciphersuite
13098 alias because they were not yet official; they could be
13099 explicitly requested by specifying the "AESdraft" ciphersuite
13100 group alias. In the final release of OpenSSL 0.9.7, the group
13101 alias is called "AES" and is part of "ALL".)
13103 *Ben Laurie, Steve Henson, Bodo Moeller*
13105 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13106 request to response.
13110 * Functions for OCSP responders. OCSP_request_onereq_count(),
13111 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13112 extract information from a certificate request. OCSP_response_create()
13113 creates a response and optionally adds a basic response structure.
13114 OCSP_basic_add1_status() adds a complete single response to a basic
13115 response and returns the OCSP_SINGLERESP structure just added (to allow
13116 extensions to be included for example). OCSP_basic_add1_cert() adds a
13117 certificate to a basic response and OCSP_basic_sign() signs a basic
13118 response with various flags. New helper functions ASN1_TIME_check()
13119 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13120 (converts ASN1_TIME to GeneralizedTime).
13124 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13125 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13126 structure from a certificate. X509_pubkey_digest() digests the public_key
13127 contents: this is used in various key identifiers.
13131 * Make sk_sort() tolerate a NULL argument.
13133 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13135 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13136 passed by the function are trusted implicitly. If any of them signed the
13137 response then it is assumed to be valid and is not verified.
13141 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13142 to data. This was previously part of the PKCS7 ASN1 code. This
13143 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13144 *Steve Henson, reported by Kenneth R. Robinette
13145 <support@securenetterm.com>*
13147 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13148 routines: without these tracing memory leaks is very painful.
13149 Fix leaks in PKCS12 and PKCS7 routines.
13153 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13154 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13155 effectively meant GeneralizedTime would never be used. Now it
13156 is initialised to -1 but X509_time_adj() now has to check the value
13157 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13158 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13159 *Steve Henson, reported by Kenneth R. Robinette
13160 <support@securenetterm.com>*
13162 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13163 result in a zero length in the ASN1_INTEGER structure which was
13164 not consistent with the structure when d2i_ASN1_INTEGER() was used
13165 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13166 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13167 where it did not print out a minus for negative ASN1_INTEGER.
13171 * Add summary printout to ocsp utility. The various functions which
13172 convert status values to strings have been renamed to:
13173 OCSP_response_status_str(), OCSP_cert_status_str() and
13174 OCSP_crl_reason_str() and are no longer static. New options
13175 to verify nonce values and to disable verification. OCSP response
13176 printout format cleaned up.
13180 * Add additional OCSP certificate checks. These are those specified
13181 in RFC2560. This consists of two separate checks: the CA of the
13182 certificate being checked must either be the OCSP signer certificate
13183 or the issuer of the OCSP signer certificate. In the latter case the
13184 OCSP signer certificate must contain the OCSP signing extended key
13185 usage. This check is performed by attempting to match the OCSP
13186 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13187 in the OCSP_CERTID structures of the response.
13191 * Initial OCSP certificate verification added to OCSP_basic_verify()
13192 and related routines. This uses the standard OpenSSL certificate
13193 verify routines to perform initial checks (just CA validity) and
13194 to obtain the certificate chain. Then additional checks will be
13195 performed on the chain. Currently the root CA is checked to see
13196 if it is explicitly trusted for OCSP signing. This is used to set
13197 a root CA as a global signing root: that is any certificate that
13198 chains to that CA is an acceptable OCSP signing certificate.
13202 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13203 extensions from a separate configuration file.
13204 As when reading extensions from the main configuration file,
13205 the '-extensions ...' option may be used for specifying the
13208 *Massimiliano Pala <madwolf@comune.modena.it>*
13210 * New OCSP utility. Allows OCSP requests to be generated or
13211 read. The request can be sent to a responder and the output
13212 parsed, outputted or printed in text form. Not complete yet:
13213 still needs to check the OCSP response validity.
13217 * New subcommands for 'openssl ca':
13218 `openssl ca -status <serial>` prints the status of the cert with
13219 the given serial number (according to the index file).
13220 `openssl ca -updatedb` updates the expiry status of certificates
13223 *Massimiliano Pala <madwolf@comune.modena.it>*
13225 * New '-newreq-nodes' command option to CA.pl. This is like
13226 '-newreq', but calls 'openssl req' with the '-nodes' option
13227 so that the resulting key is not encrypted.
13229 *Damien Miller <djm@mindrot.org>*
13231 * New configuration for the GNU Hurd.
13233 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13235 * Initial code to implement OCSP basic response verify. This
13236 is currently incomplete. Currently just finds the signer's
13237 certificate and verifies the signature on the response.
13241 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13242 value of OPENSSLDIR. This is available via the new '-d' option
13243 to 'openssl version', and is also included in 'openssl version -a'.
13247 * Allowing defining memory allocation callbacks that will be given
13248 file name and line number information in additional arguments
13249 (a `const char*` and an int). The basic functionality remains, as
13250 well as the original possibility to just replace malloc(),
13251 realloc() and free() by functions that do not know about these
13252 additional arguments. To register and find out the current
13253 settings for extended allocation functions, the following
13254 functions are provided:
13256 CRYPTO_set_mem_ex_functions
13257 CRYPTO_set_locked_mem_ex_functions
13258 CRYPTO_get_mem_ex_functions
13259 CRYPTO_get_locked_mem_ex_functions
13261 These work the same way as CRYPTO_set_mem_functions and friends.
13262 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13263 extended allocation function is enabled.
13264 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13265 a conventional allocation function is enabled.
13267 *Richard Levitte, Bodo Moeller*
13269 * Finish off removing the remaining LHASH function pointer casts.
13270 There should no longer be any prototype-casting required when using
13271 the LHASH abstraction, and any casts that remain are "bugs". See
13272 the callback types and macros at the head of lhash.h for details
13273 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13277 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13278 If /dev/[u]random devices are not available or do not return enough
13279 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13281 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13282 /etc/entropy will be queried once each in this sequence, querying stops
13283 when enough entropy was collected without querying more sockets.
13287 * Change the Unix RAND_poll() variant to be able to poll several
13288 random devices, as specified by DEVRANDOM, until a sufficient amount
13289 of data has been collected. We spend at most 10 ms on each file
13290 (select timeout) and read in non-blocking mode. DEVRANDOM now
13291 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13292 (previously it was just the string "/dev/urandom"), so on typical
13293 platforms the 10 ms delay will never occur.
13294 Also separate out the Unix variant to its own file, rand_unix.c.
13295 For VMS, there's a currently-empty rand_vms.c.
13299 * Move OCSP client related routines to ocsp_cl.c. These
13300 provide utility functions which an application needing
13301 to issue a request to an OCSP responder and analyse the
13302 response will typically need: as opposed to those which an
13303 OCSP responder itself would need which will be added later.
13305 OCSP_request_sign() signs an OCSP request with an API similar
13306 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13307 response. OCSP_response_get1_basic() extracts basic response
13308 from response. OCSP_resp_find_status(): finds and extracts status
13309 information from an OCSP_CERTID structure (which will be created
13310 when the request structure is built). These are built from lower
13311 level functions which work on OCSP_SINGLERESP structures but
13312 won't normally be used unless the application wishes to examine
13313 extensions in the OCSP response for example.
13315 Replace nonce routines with a pair of functions.
13316 OCSP_request_add1_nonce() adds a nonce value and optionally
13317 generates a random value. OCSP_check_nonce() checks the
13318 validity of the nonce in an OCSP response.
13322 * Change function OCSP_request_add() to OCSP_request_add0_id().
13323 This doesn't copy the supplied OCSP_CERTID and avoids the
13324 need to free up the newly created id. Change return type
13325 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13326 This can then be used to add extensions to the request.
13327 Deleted OCSP_request_new(), since most of its functionality
13328 is now in OCSP_REQUEST_new() (and the case insensitive name
13329 clash) apart from the ability to set the request name which
13330 will be added elsewhere.
13334 * Update OCSP API. Remove obsolete extensions argument from
13335 various functions. Extensions are now handled using the new
13336 OCSP extension code. New simple OCSP HTTP function which
13337 can be used to send requests and parse the response.
13341 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13342 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13343 uses the special reorder version of SET OF to sort the attributes
13344 and reorder them to match the encoded order. This resolves a long
13345 standing problem: a verify on a PKCS7 structure just after signing
13346 it used to fail because the attribute order did not match the
13347 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13348 it uses the received order. This is necessary to tolerate some broken
13349 software that does not order SET OF. This is handled by encoding
13350 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13351 to produce the required SET OF.
13355 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13356 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13357 files to get correct declarations of the ASN.1 item variables.
13361 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13362 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13363 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13364 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13365 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13366 ASN1_ITEM and no wrapper functions.
13370 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13371 replace the old function pointer based I/O routines. Change most of
13372 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13376 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13377 lines, recognize more "algorithms" that can be deselected, and make
13378 it complain about algorithm deselection that isn't recognised.
13382 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13383 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13384 to use new functions. Add NO_ASN1_OLD which can be set to remove
13385 some old style ASN1 functions: this can be used to determine if old
13386 code will still work when these eventually go away.
13390 * New extension functions for OCSP structures, these follow the
13391 same conventions as certificates and CRLs.
13395 * New function X509V3_add1_i2d(). This automatically encodes and
13396 adds an extension. Its behaviour can be customised with various
13397 flags to append, replace or delete. Various wrappers added for
13398 certificates and CRLs.
13402 * Fix to avoid calling the underlying ASN1 print routine when
13403 an extension cannot be parsed. Correct a typo in the
13404 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13408 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13409 entries for variables.
13413 * Add functionality to `apps/openssl.c` for detecting locking
13414 problems: As the program is single-threaded, all we have
13415 to do is register a locking callback using an array for
13416 storing which locks are currently held by the program.
13420 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13421 SSL_get_ex_data_X509_STORE_idx(), which is used in
13422 ssl_verify_cert_chain() and thus can be called at any time
13423 during TLS/SSL handshakes so that thread-safety is essential.
13424 Unfortunately, the ex_data design is not at all suited
13425 for multi-threaded use, so it probably should be abolished.
13429 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13431 *Broadcom, tweaked and integrated by Geoff Thorpe*
13433 * Move common extension printing code to new function
13434 X509V3_print_extensions(). Reorganise OCSP print routines and
13435 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13439 * New function X509_signature_print() to remove duplication in some
13444 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13445 set (this was treated exactly the same as SET OF previously). This
13446 is used to reorder the STACK representing the structure to match the
13447 encoding. This will be used to get round a problem where a PKCS7
13448 structure which was signed could not be verified because the STACK
13449 order did not reflect the encoded order.
13453 * Reimplement the OCSP ASN1 module using the new code.
13457 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13458 for its ASN1 operations. The old style function pointers still exist
13459 for now but they will eventually go away.
13463 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13464 completely replaces the old ASN1 functionality with a table driven
13465 encoder and decoder which interprets an ASN1_ITEM structure describing
13466 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13467 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13468 has also been converted to the new form.
13472 * Change BN_mod_exp_recp so that negative moduli are tolerated
13473 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13474 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13475 for negative moduli.
13479 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13480 of not touching the result's sign bit.
13484 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13489 * Changed the LHASH code to use prototypes for callbacks, and created
13490 macros to declare and implement thin (optionally static) functions
13491 that provide type-safety and avoid function pointer casting for the
13492 type-specific callbacks.
13496 * Added Kerberos Cipher Suites to be used with TLS, as written in
13498 *Veers Staats <staatsvr@asc.hpc.mil>,
13499 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13501 * Reformat the FAQ so the different questions and answers can be divided
13502 in sections depending on the subject.
13506 * Have the zlib compression code load ZLIB.DLL dynamically under
13511 * New function BN_mod_sqrt for computing square roots modulo a prime
13512 (using the probabilistic Tonelli-Shanks algorithm unless
13513 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13514 be handled deterministically).
13516 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13518 * Make BN_mod_inverse faster by explicitly handling small quotients
13519 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13520 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13524 * New function BN_kronecker.
13528 * Fix BN_gcd so that it works on negative inputs; the result is
13529 positive unless both parameters are zero.
13530 Previously something reasonably close to an infinite loop was
13531 possible because numbers could be growing instead of shrinking
13532 in the implementation of Euclid's algorithm.
13536 * Fix BN_is_word() and BN_is_one() macros to take into account the
13537 sign of the number in question.
13539 Fix BN_is_word(a,w) to work correctly for w == 0.
13541 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13542 because its test if the absolute value of 'a' equals 'w'.
13543 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13544 it exists mostly for use in the implementations of BN_is_zero(),
13545 BN_is_one(), and BN_is_word().
13549 * New function BN_swap.
13553 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13554 the exponentiation functions are more likely to produce reasonable
13555 results on negative inputs.
13559 * Change BN_mod_mul so that the result is always non-negative.
13560 Previously, it could be negative if one of the factors was negative;
13561 I don't think anyone really wanted that behaviour.
13565 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13566 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13567 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13568 and add new functions:
13577 BN_mod_lshift1_quick
13579 BN_mod_lshift_quick
13581 These functions always generate non-negative results.
13583 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13584 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13586 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13587 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13588 be reduced modulo `m`.
13590 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13593 The following entry accidentally appeared in the CHANGES file
13594 distributed with OpenSSL 0.9.7. The modifications described in
13595 it do *not* apply to OpenSSL 0.9.7.
13597 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13598 was actually never needed) and in BN_mul(). The removal in BN_mul()
13599 required a small change in bn_mul_part_recursive() and the addition
13600 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13601 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13602 bn_sub_words() and bn_add_words() except they take arrays with
13608 * In 'openssl passwd', verify passwords read from the terminal
13609 unless the '-salt' option is used (which usually means that
13610 verification would just waste user's time since the resulting
13611 hash is going to be compared with some given password hash)
13612 or the new '-noverify' option is used.
13614 This is an incompatible change, but it does not affect
13615 non-interactive use of 'openssl passwd' (passwords on the command
13616 line, '-stdin' option, '-in ...' option) and thus should not
13617 cause any problems.
13621 * Remove all references to RSAref, since there's no more need for it.
13625 * Make DSO load along a path given through an environment variable
13626 (SHLIB_PATH) with shl_load().
13630 * Constify the ENGINE code as a result of BIGNUM constification.
13631 Also constify the RSA code and most things related to it. In a
13632 few places, most notable in the depth of the ASN.1 code, ugly
13633 casts back to non-const were required (to be solved at a later
13638 * Make it so the openssl application has all engines loaded by default.
13642 * Constify the BIGNUM routines a little more.
13646 * Add the following functions:
13648 ENGINE_load_cswift()
13650 ENGINE_load_atalla()
13651 ENGINE_load_nuron()
13652 ENGINE_load_builtin_engines()
13654 That way, an application can itself choose if external engines that
13655 are built-in in OpenSSL shall ever be used or not. The benefit is
13656 that applications won't have to be linked with libdl or other dso
13657 libraries unless it's really needed.
13659 Changed 'openssl engine' to load all engines on demand.
13660 Changed the engine header files to avoid the duplication of some
13661 declarations (they differed!).
13665 * 'openssl engine' can now list capabilities.
13669 * Better error reporting in 'openssl engine'.
13673 * Never call load_dh_param(NULL) in s_server.
13677 * Add engine application. It can currently list engines by name and
13678 identity, and test if they are actually available.
13682 * Improve RPM specification file by forcing symbolic linking and making
13683 sure the installed documentation is also owned by root.root.
13685 *Damien Miller <djm@mindrot.org>*
13687 * Give the OpenSSL applications more possibilities to make use of
13688 keys (public as well as private) handled by engines.
13692 * Add OCSP code that comes from CertCo.
13696 * Add VMS support for the Rijndael code.
13700 * Added untested support for Nuron crypto accelerator.
13704 * Add support for external cryptographic devices. This code was
13705 previously distributed separately as the "engine" branch.
13707 *Geoff Thorpe, Richard Levitte*
13709 * Rework the filename-translation in the DSO code. It is now possible to
13710 have far greater control over how a "name" is turned into a filename
13711 depending on the operating environment and any oddities about the
13712 different shared library filenames on each system.
13716 * Support threads on FreeBSD-elf in Configure.
13720 * Fix for SHA1 assembly problem with MASM: it produces
13721 warnings about corrupt line number information when assembling
13722 with debugging information. This is caused by the overlapping
13725 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13728 NCONF_get_number() has no error checking at all. As a replacement,
13729 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13730 promoted strongly. The old NCONF_get_number is kept around for
13731 binary backward compatibility.
13732 Make it possible for methods to load from something other than a BIO,
13733 by providing a function pointer that is given a name instead of a BIO.
13734 For example, this could be used to load configuration data from an
13739 * Fix for non blocking accept BIOs. Added new I/O special reason
13740 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13741 with non blocking I/O was not possible because no retry code was
13742 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13747 * Added the beginnings of Rijndael support.
13751 * Fix for bug in DirectoryString mask setting. Add support for
13752 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13753 to allow certificate printing to more controllable, additional
13754 'certopt' option to 'x509' to allow new printing options to be
13759 * Clean old EAY MD5 hack from e_os.h.
13763 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13765 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13766 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13768 *Joe Orton, Steve Henson*
13770 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13772 * Fix additional bug revealed by the NISCC test suite:
13774 Stop bug triggering large recursion when presented with
13775 certain ASN.1 tags ([CVE-2003-0851])
13779 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13781 * Fix various bugs revealed by running the NISCC test suite:
13783 Stop out of bounds reads in the ASN1 code when presented with
13784 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13786 If verify callback ignores invalid public key errors don't try to check
13787 certificate signature with the NULL public key.
13791 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13792 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13797 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13798 extra data after the compression methods not only for TLS 1.0
13799 but also for SSL 3.0 (as required by the specification).
13801 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13803 * Change X509_certificate_type() to mark the key as exported/exportable
13804 when it's 512 *bits* long, not 512 bytes.
13808 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13810 * Countermeasure against the Klima-Pokorny-Rosa extension of
13811 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13812 a protocol version number mismatch like a decryption error
13813 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13817 * Turn on RSA blinding by default in the default implementation
13818 to avoid a timing attack. Applications that don't want it can call
13819 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13820 They would be ill-advised to do so in most cases.
13822 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13824 * Change RSA blinding code so that it works when the PRNG is not
13825 seeded (in this case, the secret RSA exponent is abused as
13826 an unpredictable seed -- if it is not unpredictable, there
13827 is no point in blinding anyway). Make RSA blinding thread-safe
13828 by remembering the creator's thread ID in rsa->blinding and
13829 having all other threads use local one-time blinding factors
13830 (this requires more computation than sharing rsa->blinding, but
13831 avoids excessive locking; and if an RSA object is not shared
13832 between threads, blinding will still be very fast).
13836 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13838 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13839 via timing by performing a MAC computation even if incorrect
13840 block cipher padding has been found. This is a countermeasure
13841 against active attacks where the attacker has to distinguish
13842 between bad padding and a MAC verification error. ([CVE-2003-0078])
13844 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13845 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13846 Martin Vuagnoux (EPFL, Ilion)*
13848 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13850 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13851 memory from its contents. This is done with a counter that will
13852 place alternating values in each byte. This can be used to solve
13853 two issues: 1) the removal of calls to memset() by highly optimizing
13854 compilers, and 2) cleansing with other values than 0, since those can
13855 be read through on certain media, for example a swap space on disk.
13859 * Bugfix: client side session caching did not work with external caching,
13860 because the session->cipher setting was not restored when reloading
13861 from the external cache. This problem was masked, when
13862 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13863 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13867 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13868 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13870 *Zeev Lieber <zeev-l@yahoo.com>*
13872 * Undo an undocumented change introduced in 0.9.6e which caused
13873 repeated calls to OpenSSL_add_all_ciphers() and
13874 OpenSSL_add_all_digests() to be ignored, even after calling
13879 * Change the default configuration reader to deal with last line not
13880 being properly terminated.
13884 * Change X509_NAME_cmp() so it applies the special rules on handling
13885 DN values that are of type PrintableString, as well as RDNs of type
13886 emailAddress where the value has the type ia5String.
13888 *stefank@valicert.com via Richard Levitte*
13890 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13891 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13892 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13893 the bitwise-OR of the two for use by the majority of applications
13894 wanting this behaviour, and update the docs. The documented
13895 behaviour and actual behaviour were inconsistent and had been
13896 changing anyway, so this is more a bug-fix than a behavioural
13899 *Geoff Thorpe, diagnosed by Nadav Har'El*
13901 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13902 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13906 * Fix initialization code race conditions in
13907 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13908 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13909 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13910 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13911 ssl2_get_cipher_by_char(),
13912 ssl3_get_cipher_by_char().
13914 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13916 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13917 the cached sessions are flushed, as the remove_cb() might use ex_data
13918 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13919 (see [openssl.org #212]).
13921 *Geoff Thorpe, Lutz Jaenicke*
13923 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13924 length, instead of the encoding length to d2i_ASN1_OBJECT.
13928 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13930 * [In 0.9.6g-engine release:]
13931 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13933 *Lynn Gazis <lgazis@rainbow.com>*
13935 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13937 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13938 and get fix the header length calculation.
13939 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13940 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13942 * Use proper error handling instead of 'assertions' in buffer
13943 overflow checks added in 0.9.6e. This prevents DoS (the
13944 assertions could call abort()).
13946 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13948 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13950 * Add various sanity checks to asn1_get_length() to reject
13951 the ASN1 length bytes if they exceed sizeof(long), will appear
13952 negative or the content length exceeds the length of the
13955 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13957 * Fix cipher selection routines: ciphers without encryption had no flags
13958 for the cipher strength set and where therefore not handled correctly
13959 by the selection routines (PR #130).
13963 * Fix EVP_dsa_sha macro.
13968 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13969 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13970 that was added in OpenSSL 0.9.6d.
13972 As the countermeasure turned out to be incompatible with some
13973 broken SSL implementations, the new option is part of SSL_OP_ALL.
13974 SSL_OP_ALL is usually employed when compatibility with weird SSL
13975 implementations is desired (e.g. '-bugs' option to 's_client' and
13976 's_server'), so the new option is automatically set in many
13981 * Changes in security patch:
13983 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13984 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13985 Air Force Materiel Command, USAF, under agreement number
13988 * Add various sanity checks to asn1_get_length() to reject
13989 the ASN1 length bytes if they exceed sizeof(long), will appear
13990 negative or the content length exceeds the length of the
13991 supplied buffer. ([CVE-2002-0659])
13993 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13995 * Assertions for various potential buffer overflows, not known to
13996 happen in practice.
13998 *Ben Laurie (CHATS)*
14000 * Various temporary buffers to hold ASCII versions of integers were
14001 too small for 64 bit platforms. ([CVE-2002-0655])
14002 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14004 * Remote buffer overflow in SSL3 protocol - an attacker could
14005 supply an oversized session ID to a client. ([CVE-2002-0656])
14007 *Ben Laurie (CHATS)*
14009 * Remote buffer overflow in SSL2 protocol - an attacker could
14010 supply an oversized client master key. ([CVE-2002-0656])
14012 *Ben Laurie (CHATS)*
14014 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14016 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14017 encoded as NULL) with id-dsa-with-sha1.
14019 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14021 * Check various `X509_...()` return values in `apps/req.c`.
14023 *Nils Larsch <nla@trustcenter.de>*
14025 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14026 an end-of-file condition would erroneously be flagged, when the CRLF
14027 was just at the end of a processed block. The bug was discovered when
14028 processing data through a buffering memory BIO handing the data to a
14029 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14030 <ptsekov@syntrex.com> and Nedelcho Stanev.
14034 * Implement a countermeasure against a vulnerability recently found
14035 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14036 before application data chunks to avoid the use of known IVs
14037 with data potentially chosen by the attacker.
14041 * Fix length checks in ssl3_get_client_hello().
14045 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14046 to prevent ssl3_read_internal() from incorrectly assuming that
14047 ssl3_read_bytes() found application data while handshake
14048 processing was enabled when in fact s->s3->in_read_app_data was
14049 merely automatically cleared during the initial handshake.
14051 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14053 * Fix object definitions for Private and Enterprise: they were not
14054 recognized in their shortname (=lowercase) representation. Extend
14055 obj_dat.pl to issue an error when using undefined keywords instead
14056 of silently ignoring the problem (Svenning Sorensen
14057 <sss@sss.dnsalias.net>).
14061 * Fix DH_generate_parameters() so that it works for 'non-standard'
14062 generators, i.e. generators other than 2 and 5. (Previously, the
14063 code did not properly initialise the 'add' and 'rem' values to
14064 BN_generate_prime().)
14066 In the new general case, we do not insist that 'generator' is
14067 actually a primitive root: This requirement is rather pointless;
14068 a generator of the order-q subgroup is just as good, if not
14073 * Map new X509 verification errors to alerts. Discovered and submitted by
14074 Tom Wu <tom@arcot.com>.
14078 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14079 returning non-zero before the data has been completely received
14080 when using non-blocking I/O.
14082 *Bodo Moeller; problem pointed out by John Hughes*
14084 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14086 *Ben Laurie, Lutz Jaenicke*
14088 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14089 Yoram Zahavi <YoramZ@gilian.com>).
14093 * Add information about CygWin 1.3 and on, and preserve proper
14094 configuration for the versions before that.
14096 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14098 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14099 check whether we deal with a copy of a session and do not delete from
14100 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14101 <izhar@checkpoint.com>.
14105 * Do not store session data into the internal session cache, if it
14106 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14107 flag is set). Proposed by Aslam <aslam@funk.com>.
14111 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14116 * [In 0.9.6d-engine release:]
14117 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14119 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14121 * Add the configuration target linux-s390x.
14123 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14125 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14126 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14127 variable as an indication that a ClientHello message has been
14128 received. As the flag value will be lost between multiple
14129 invocations of ssl3_accept when using non-blocking I/O, the
14130 function may not be aware that a handshake has actually taken
14131 place, thus preventing a new session from being added to the
14134 To avoid this problem, we now set s->new_session to 2 instead of
14135 using a local variable.
14137 *Lutz Jaenicke, Bodo Moeller*
14139 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14140 if the SSL_R_LENGTH_MISMATCH error is detected.
14142 *Geoff Thorpe, Bodo Moeller*
14144 * New 'shared_ldflag' column in Configure platform table.
14148 * Fix EVP_CIPHER_mode macro.
14150 *"Dan S. Camper" <dan@bti.net>*
14152 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14153 type, we must throw them away by setting rr->length to 0.
14155 *D P Chang <dpc@qualys.com>*
14157 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14159 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14160 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14161 worked incorrectly for those cases where range = `10..._2` and
14162 `3*range` is two bits longer than range.)
14166 * Only add signing time to PKCS7 structures if it is not already
14171 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14172 OBJ_ld_ce should be OBJ_id_ce.
14173 Also some ip-pda OIDs in crypto/objects/objects.txt were
14174 incorrect (cf. RFC 3039).
14176 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14178 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14179 returns early because it has nothing to do.
14181 *Andy Schneider <andy.schneider@bjss.co.uk>*
14183 * [In 0.9.6c-engine release:]
14184 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14186 *Andy Schneider <andy.schneider@bjss.co.uk>*
14188 * [In 0.9.6c-engine release:]
14189 Add support for Cryptographic Appliance's keyserver technology.
14190 (Use engine 'keyclient')
14192 *Cryptographic Appliances and Geoff Thorpe*
14194 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14195 is called via tools/c89.sh because arguments have to be
14196 rearranged (all '-L' options must appear before the first object
14199 *Richard Shapiro <rshapiro@abinitio.com>*
14201 * [In 0.9.6c-engine release:]
14202 Add support for Broadcom crypto accelerator cards, backported
14205 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14207 * [In 0.9.6c-engine release:]
14208 Add support for SureWare crypto accelerator cards from
14209 Baltimore Technologies. (Use engine 'sureware')
14211 *Baltimore Technologies and Mark Cox*
14213 * [In 0.9.6c-engine release:]
14214 Add support for crypto accelerator cards from Accelerated
14215 Encryption Processing, www.aep.ie. (Use engine 'aep')
14217 *AEP Inc. and Mark Cox*
14219 * Add a configuration entry for gcc on UnixWare.
14221 *Gary Benson <gbenson@redhat.com>*
14223 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14224 messages are stored in a single piece (fixed-length part and
14225 variable-length part combined) and fix various bugs found on the way.
14229 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14230 instead. BIO_gethostbyname() does not know what timeouts are
14231 appropriate, so entries would stay in cache even when they have
14233 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14235 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14236 faced with a pathologically small ClientHello fragment that does
14237 not contain client_version: Instead of aborting with an error,
14238 simply choose the highest available protocol version (i.e.,
14239 TLS 1.0 unless it is disabled). In practice, ClientHello
14240 messages are never sent like this, but this change gives us
14241 strictly correct behaviour at least for TLS.
14245 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14246 never resets s->method to s->ctx->method when called from within
14247 one of the SSL handshake functions.
14249 *Bodo Moeller; problem pointed out by Niko Baric*
14251 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14252 (sent using the client's version number) if client_version is
14253 smaller than the protocol version in use. Also change
14254 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14255 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14256 the client will at least see that alert.
14260 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14265 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14266 client receives HelloRequest while in a handshake.
14268 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14270 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14271 should end in 'break', not 'goto end' which circumvents various
14272 cleanups done in state SSL_ST_OK. But session related stuff
14273 must be disabled for SSL_ST_OK in the case that we just sent a
14276 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14277 before just sending a HelloRequest.
14279 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14281 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14282 reveal whether illegal block cipher padding was found or a MAC
14283 verification error occurred. (Neither SSLerr() codes nor alerts
14284 are directly visible to potential attackers, but the information
14285 may leak via logfiles.)
14287 Similar changes are not required for the SSL 2.0 implementation
14288 because the number of padding bytes is sent in clear for SSL 2.0,
14289 and the extra bytes are just ignored. However ssl/s2_pkt.c
14290 failed to verify that the purported number of padding bytes is in
14295 * Add OpenUNIX-8 support including shared libraries
14296 (Boyd Lynn Gerber <gerberb@zenez.com>).
14300 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14301 'wristwatch attack' using huge encoding parameters (cf.
14302 James H. Manger's CRYPTO 2001 paper). Note that the
14303 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14304 encoding parameters and hence was not vulnerable.
14308 * BN_sqr() bug fix.
14310 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14312 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14313 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14314 followed by modular reduction.
14316 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14318 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14319 equivalent based on BN_pseudo_rand() instead of BN_rand().
14323 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14324 This function was broken, as the check for a new client hello message
14325 to handle SGC did not allow these large messages.
14326 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14330 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14334 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14335 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14339 * Rework the configuration and shared library support for Tru64 Unix.
14340 The configuration part makes use of modern compiler features and
14341 still retains old compiler behavior for those that run older versions
14342 of the OS. The shared library support part includes a variant that
14343 uses the RPATH feature, and is available through the special
14344 configuration target "alpha-cc-rpath", which will never be selected
14347 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14349 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14350 with the same message size as in ssl3_get_certificate_request().
14351 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14352 messages might inadvertently be reject as too long.
14354 *Petr Lampa <lampa@fee.vutbr.cz>*
14356 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14360 * Modified SSL library such that the verify_callback that has been set
14361 specifically for an SSL object with SSL_set_verify() is actually being
14362 used. Before the change, a verify_callback set with this function was
14363 ignored and the verify_callback() set in the SSL_CTX at the time of
14364 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14365 to allow the necessary settings.
14369 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14370 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14371 done automatically (in contradiction to the requirements of the C
14372 standard). This made problems when used from OpenSSH.
14376 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14377 dh->length and always used
14379 BN_rand_range(priv_key, dh->p).
14381 BN_rand_range() is not necessary for Diffie-Hellman, and this
14382 specific range makes Diffie-Hellman unnecessarily inefficient if
14383 dh->length (recommended exponent length) is much smaller than the
14384 length of dh->p. We could use BN_rand_range() if the order of
14385 the subgroup was stored in the DH structure, but we only have
14390 BN_rand(priv_key, l, ...)
14392 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14399 RSA_eay_public_encrypt
14400 RSA_eay_private_decrypt
14401 RSA_eay_private_encrypt (signing)
14402 RSA_eay_public_decrypt (signature verification)
14404 (default implementations for RSA_public_encrypt,
14405 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14406 always reject numbers >= n.
14410 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14411 to synchronize access to 'locking_thread'. This is necessary on
14412 systems where access to 'locking_thread' (an 'unsigned long'
14413 variable) is not atomic.
14417 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14418 *before* setting the 'crypto_lock_rand' flag. The previous code had
14419 a race condition if 0 is a valid thread ID.
14421 *Travis Vitek <vitek@roguewave.com>*
14423 * Add support for shared libraries under Irix.
14425 *Albert Chin-A-Young <china@thewrittenword.com>*
14427 * Add configuration option to build on Linux on both big-endian and
14428 little-endian MIPS.
14430 *Ralf Baechle <ralf@uni-koblenz.de>*
14432 * Add the possibility to create shared libraries on HP-UX.
14436 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14438 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14439 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14440 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14441 PRNG state recovery was possible based on the output of
14442 one PRNG request appropriately sized to gain knowledge on
14443 'md' followed by enough consecutive 1-byte PRNG requests
14444 to traverse all of 'state'.
14446 1. When updating 'md_local' (the current thread's copy of 'md')
14447 during PRNG output generation, hash all of the previous
14448 'md_local' value, not just the half used for PRNG output.
14450 2. Make the number of bytes from 'state' included into the hash
14451 independent from the number of PRNG bytes requested.
14453 The first measure alone would be sufficient to avoid
14454 Markku-Juhani's attack. (Actually it had never occurred
14455 to me that the half of 'md_local' used for chaining was the
14456 half from which PRNG output bytes were taken -- I had always
14457 assumed that the secret half would be used.) The second
14458 measure makes sure that additional data from 'state' is never
14459 mixed into 'md_local' in small portions; this heuristically
14460 further strengthens the PRNG.
14464 * Fix crypto/bn/asm/mips3.s.
14468 * When only the key is given to "enc", the IV is undefined. Print out
14469 an error message in this case.
14473 * Handle special case when X509_NAME is empty in X509 printing routines.
14477 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14478 positive and less than q.
14482 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14483 used: it isn't thread safe and the add_lock_callback should handle
14486 *Paul Rose <Paul.Rose@bridge.com>*
14488 * Verify that incoming data obeys the block size in
14489 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14495 *Ulf Möller, Bodo Möller*
14497 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14498 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14499 when fixing the server behaviour for backwards-compatible 'client
14500 hello' messages. (Note that the attack is impractical against
14501 SSL 3.0 and TLS 1.0 anyway because length and version checking
14502 means that the probability of guessing a valid ciphertext is
14503 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14506 Before 0.9.5, the countermeasure (hide the error by generating a
14507 random 'decryption result') did not work properly because
14508 ERR_clear_error() was missing, meaning that SSL_get_error() would
14509 detect the supposedly ignored error.
14511 Both problems are now fixed.
14515 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14516 (previously it was 1024).
14520 * Fix for compatibility mode trust settings: ignore trust settings
14521 unless some valid trust or reject settings are present.
14525 * Fix for blowfish EVP: its a variable length cipher.
14529 * Fix various bugs related to DSA S/MIME verification. Handle missing
14530 parameters in DSA public key structures and return an error in the
14531 DSA routines if parameters are absent.
14535 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14536 in the current directory if neither $RANDFILE nor $HOME was set.
14537 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14538 caused some confusion to Windows users who haven't defined $HOME.
14539 Thus RAND_file_name() is changed again: e_os.h can define a
14540 DEFAULT_HOME, which will be used if $HOME is not set.
14541 For Windows, we use "C:"; on other platforms, we still require
14542 environment variables.
14544 * Move 'if (!initialized) RAND_poll()' into regions protected by
14545 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14546 having multiple threads call RAND_poll() concurrently.
14550 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14551 combination of a flag and a thread ID variable.
14552 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14553 flag), *other* threads can enter ssleay_add_bytes without obeying
14554 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14555 that they do not hold after the first thread unsets add_do_not_lock).
14559 * Change bctest again: '-x' expressions are not available in all
14560 versions of 'test'.
14564 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14566 * Fix a couple of memory leaks in PKCS7_dataDecode()
14568 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14570 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14571 the default extension for executables, if any. Also, make the perl
14572 scripts that use symlink() to test if it really exists and use "cp"
14573 if it doesn't. All this made OpenSSL compilable and installable in
14578 * Fix for asn1_GetSequence() for indefinite length constructed data.
14579 If SEQUENCE is length is indefinite just set c->slen to the total
14580 amount of data available.
14582 *Steve Henson, reported by shige@FreeBSD.org*
14584 *This change does not apply to 0.9.7.*
14586 * Change bctest to avoid here-documents inside command substitution
14587 (workaround for FreeBSD /bin/sh bug).
14588 For compatibility with Ultrix, avoid shell functions (introduced
14589 in the bctest version that searches along $PATH).
14593 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14594 with des_encrypt() defined on some operating systems, like Solaris
14599 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14600 On the Importance of Eliminating Errors in Cryptographic
14601 Computations, J. Cryptology 14 (2001) 2, 101-119,
14602 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14606 * MIPS assembler BIGNUM division bug fix.
14610 * Disabled incorrect Alpha assembler code.
14614 * Fix PKCS#7 decode routines so they correctly update the length
14615 after reading an EOC for the EXPLICIT tag.
14619 *This change does not apply to 0.9.7.*
14621 * Fix bug in PKCS#12 key generation routines. This was triggered
14622 if a 3DES key was generated with a 0 initial byte. Include
14623 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14624 (but broken) behaviour.
14628 * Enhance bctest to search for a working bc along $PATH and print
14631 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14633 * Fix memory leaks in err.c: free err_data string if necessary;
14634 don't write to the wrong index in ERR_set_error_data.
14638 * Implement ssl23_peek (analogous to ssl23_read), which previously
14643 * Replace rdtsc with `_emit` statements for VC++ version 5.
14645 *Jeremy Cooper <jeremy@baymoo.org>*
14647 * Make it possible to reuse SSLv2 sessions.
14651 * In copy_email() check for >= 0 as a return value for
14652 X509_NAME_get_index_by_NID() since 0 is a valid index.
14654 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14656 * Avoid coredump with unsupported or invalid public keys by checking if
14657 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14658 PKCS7_verify() fails with non detached data.
14662 * Don't use getenv in library functions when run as setuid/setgid.
14663 New function OPENSSL_issetugid().
14667 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14668 due to incorrect handling of multi-threading:
14670 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14672 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14674 3. Count how many times MemCheck_off() has been called so that
14675 nested use can be treated correctly. This also avoids
14676 inband-signalling in the previous code (which relied on the
14677 assumption that thread ID 0 is impossible).
14681 * Add "-rand" option also to s_client and s_server.
14685 * Fix CPU detection on Irix 6.x.
14686 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14687 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14689 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14694 *This change does not apply to 0.9.7.*
14696 * Use the cached encoding of an X509_NAME structure rather than
14697 copying it. This is apparently the reason for the libsafe "errors"
14698 but the code is actually correct.
14702 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14703 Bleichenbacher's DSA attack.
14704 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14705 to be set and top=0 forces the highest bit to be set; top=-1 is new
14706 and leaves the highest bit random.
14708 *Ulf Moeller, Bodo Moeller*
14710 * In the `NCONF_...`-based implementations for `CONF_...` queries
14711 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14712 a temporary CONF structure with the data component set to NULL
14713 (which gives segmentation faults in lh_retrieve).
14714 Instead, use NULL for the CONF pointer in CONF_get_string and
14715 CONF_get_number (which may use environment variables) and directly
14716 return NULL from CONF_get_section.
14720 * Fix potential buffer overrun for EBCDIC.
14724 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14725 keyUsage if basicConstraints absent for a CA.
14729 * Make SMIME_write_PKCS7() write mail header values with a format that
14730 is more generally accepted (no spaces before the semicolon), since
14731 some programs can't parse those values properly otherwise. Also make
14732 sure BIO's that break lines after each write do not create invalid
14737 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14738 macros previously used would not encode an empty SEQUENCE OF
14739 and break the signature.
14743 *This change does not apply to 0.9.7.*
14745 * Zero the premaster secret after deriving the master secret in
14750 * Add some EVP_add_digest_alias registrations (as found in
14751 OpenSSL_add_all_digests()) to SSL_library_init()
14752 aka OpenSSL_add_ssl_algorithms(). This provides improved
14753 compatibility with peers using X.509 certificates
14754 with unconventional AlgorithmIdentifier OIDs.
14758 * Fix for Irix with NO_ASM.
14760 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14762 * ./config script fixes.
14764 *Ulf Moeller, Richard Levitte*
14766 * Fix 'openssl passwd -1'.
14770 * Change PKCS12_key_gen_asc() so it can cope with non null
14771 terminated strings whose length is passed in the passlen
14772 parameter, for example from PEM callbacks. This was done
14773 by adding an extra length parameter to asc2uni().
14775 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14777 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14778 call failed, free the DSA structure.
14782 * Fix to uni2asc() to cope with zero length Unicode strings.
14783 These are present in some PKCS#12 files.
14787 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14788 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14789 when writing a 32767 byte record.
14791 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14793 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14794 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14796 (RSA objects have a reference count access to which is protected
14797 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14798 so they are meant to be shared between threads.)
14799 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14800 "Reddie, Steven" <Steven.Reddie@ca.com>*
14802 * Fix a deadlock in CRYPTO_mem_leaks().
14806 * Use better test patterns in bntest.
14810 * rand_win.c fix for Borland C.
14814 * BN_rshift bugfix for n == 0.
14818 * Add a 'bctest' script that checks for some known 'bc' bugs
14819 so that 'make test' does not abort just because 'bc' is broken.
14823 * Store verify_result within SSL_SESSION also for client side to
14824 avoid potential security hole. (Re-used sessions on the client side
14825 always resulted in verify_result==X509_V_OK, not using the original
14826 result of the server certificate verification.)
14830 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14831 SSL3_RT_APPLICATION_DATA, return 0.
14832 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14837 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14838 releases, have been re-implemented by renaming the previous
14839 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14840 and ssl3_read_internal, respectively, and adding 'peek' parameters
14841 to them. The new ssl[23]_{read,peek} functions are calls to
14842 ssl[23]_read_internal with the 'peek' flag set appropriately.
14843 A 'peek' parameter has also been added to ssl3_read_bytes, which
14844 does the actual work for ssl3_read_internal.
14848 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14849 the method-specific "init()" handler. Also clean up ex_data after
14850 calling the method-specific "finish()" handler. Previously, this was
14851 happening the other way round.
14855 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14856 The previous value, 12, was not always sufficient for BN_mod_exp().
14860 * Make sure that shared libraries get the internal name engine with
14861 the full version number and not just 0. This should mark the
14862 shared libraries as not backward compatible. Of course, this should
14863 be changed again when we can guarantee backward binary compatibility.
14867 * Fix typo in get_cert_by_subject() in by_dir.c
14869 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14871 * Rework the system to generate shared libraries:
14873 - Make note of the expected extension for the shared libraries and
14874 if there is a need for symbolic links from for example libcrypto.so.0
14875 to libcrypto.so.0.9.7. There is extended info in Configure for
14878 - Make as few rebuilds of the shared libraries as possible.
14880 - Still avoid linking the OpenSSL programs with the shared libraries.
14882 - When installing, install the shared libraries separately from the
14887 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14889 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14890 and not in SSL_clear because the latter is also used by the
14891 accept/connect functions; previously, the settings made by
14892 SSL_set_read_ahead would be lost during the handshake.
14894 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14896 * Correct util/mkdef.pl to be selective about disabled algorithms.
14897 Previously, it would create entries for disabled algorithms no
14902 * Added several new manual pages for SSL_* function.
14906 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14908 * In ssl23_get_client_hello, generate an error message when faced
14909 with an initial SSL 3.0/TLS record that is too small to contain the
14910 first two bytes of the ClientHello message, i.e. client_version.
14911 (Note that this is a pathologic case that probably has never happened
14912 in real life.) The previous approach was to use the version number
14913 from the record header as a substitute; but our protocol choice
14914 should not depend on that one because it is not authenticated
14915 by the Finished messages.
14919 * More robust randomness gathering functions for Windows.
14921 *Jeffrey Altman <jaltman@columbia.edu>*
14923 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14924 not set then we don't setup the error code for issuer check errors
14925 to avoid possibly overwriting other errors which the callback does
14926 handle. If an application does set the flag then we assume it knows
14927 what it is doing and can handle the new informational codes
14932 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14933 a general "ANY" type, as such it should be able to decode anything
14934 including tagged types. However it didn't check the class so it would
14935 wrongly interpret tagged types in the same way as their universal
14936 counterpart and unknown types were just rejected. Changed so that the
14937 tagged and unknown types are handled in the same way as a SEQUENCE:
14938 that is the encoding is stored intact. There is also a new type
14939 "V_ASN1_OTHER" which is used when the class is not universal, in this
14940 case we have no idea what the actual type is so we just lump them all
14945 * On VMS, stdout may very well lead to a file that is written to
14946 in a record-oriented fashion. That means that every write() will
14947 write a separate record, which will be read separately by the
14948 programs trying to read from it. This can be very confusing.
14950 The solution is to put a BIO filter in the way that will buffer
14951 text until a linefeed is reached, and then write everything a
14952 line at a time, so every record written will be an actual line,
14953 not chunks of lines and not (usually doesn't happen, but I've
14954 seen it once) several lines in one record. BIO_f_linebuffer() is
14957 Currently, it's a VMS-only method, because that's where it has
14958 been tested well enough.
14962 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14963 it can return incorrect results.
14964 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14965 but it was in 0.9.6-beta[12].)
14969 * Disable the check for content being present when verifying detached
14970 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14971 include zero length content when signing messages.
14975 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14976 BIO_ctrl (for BIO pairs).
14980 * Add DSO method for VMS.
14984 * Bug fix: Montgomery multiplication could produce results with the
14989 * Add RPM specification openssl.spec and modify it to build three
14990 packages. The default package contains applications, application
14991 documentation and run-time libraries. The devel package contains
14992 include files, static libraries and function documentation. The
14993 doc package contains the contents of the doc directory. The original
14994 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14998 * Add a large number of documentation files for many SSL routines.
15000 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15002 * Add a configuration entry for Sony News 4.
15004 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15006 * Don't set the two most significant bits to one when generating a
15007 random number < q in the DSA library.
15011 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15012 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15013 the underlying transport is blocking) if a handshake took place.
15014 (The default behaviour is needed by applications such as s_client
15015 and s_server that use select() to determine when to use SSL_read;
15016 but for applications that know in advance when to expect data, it
15017 just makes things more complicated.)
15021 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15026 * Add a few more EBCDIC conditionals that make `req` and `x509`
15027 work better on such systems.
15029 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15031 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15032 Update PKCS12_parse() so it copies the friendlyName and the
15033 keyid to the certificates aux info.
15037 * Fix bug in PKCS7_verify() which caused an infinite loop
15038 if there was more than one signature.
15040 *Sven Uszpelkat <su@celocom.de>*
15042 * Major change in util/mkdef.pl to include extra information
15043 about each symbol, as well as presenting variables as well
15044 as functions. This change means that there's n more need
15045 to rebuild the .num files when some algorithms are excluded.
15049 * Allow the verify time to be set by an application,
15050 rather than always using the current time.
15054 * Phase 2 verify code reorganisation. The certificate
15055 verify code now looks up an issuer certificate by a
15056 number of criteria: subject name, authority key id
15057 and key usage. It also verifies self signed certificates
15058 by the same criteria. The main comparison function is
15059 X509_check_issued() which performs these checks.
15061 Lot of changes were necessary in order to support this
15062 without completely rewriting the lookup code.
15064 Authority and subject key identifier are now cached.
15066 The LHASH 'certs' is X509_STORE has now been replaced
15067 by a STACK_OF(X509_OBJECT). This is mainly because an
15068 LHASH can't store or retrieve multiple objects with
15069 the same hash value.
15071 As a result various functions (which were all internal
15072 use only) have changed to handle the new X509_STORE
15073 structure. This will break anything that messed round
15074 with X509_STORE internally.
15076 The functions X509_STORE_add_cert() now checks for an
15077 exact match, rather than just subject name.
15079 The X509_STORE API doesn't directly support the retrieval
15080 of multiple certificates matching a given criteria, however
15081 this can be worked round by performing a lookup first
15082 (which will fill the cache with candidate certificates)
15083 and then examining the cache for matches. This is probably
15084 the best we can do without throwing out X509_LOOKUP
15085 entirely (maybe later...).
15087 The X509_VERIFY_CTX structure has been enhanced considerably.
15089 All certificate lookup operations now go via a get_issuer()
15090 callback. Although this currently uses an X509_STORE it
15091 can be replaced by custom lookups. This is a simple way
15092 to bypass the X509_STORE hackery necessary to make this
15093 work and makes it possible to use more efficient techniques
15094 in future. A very simple version which uses a simple
15095 STACK for its trusted certificate store is also provided
15096 using X509_STORE_CTX_trusted_stack().
15098 The verify_cb() and verify() callbacks now have equivalents
15099 in the X509_STORE_CTX structure.
15101 X509_STORE_CTX also has a 'flags' field which can be used
15102 to customise the verify behaviour.
15106 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15107 excludes S/MIME capabilities.
15111 * When a certificate request is read in keep a copy of the
15112 original encoding of the signed data and use it when outputting
15113 again. Signatures then use the original encoding rather than
15114 a decoded, encoded version which may cause problems if the
15115 request is improperly encoded.
15119 * For consistency with other BIO_puts implementations, call
15120 buffer_write(b, ...) directly in buffer_puts instead of calling
15123 In BIO_puts, increment b->num_write as in BIO_write.
15125 *Peter.Sylvester@EdelWeb.fr*
15127 * Fix BN_mul_word for the case where the word is 0. (We have to use
15128 BN_zero, we may not return a BIGNUM with an array consisting of
15129 words set to zero.)
15133 * Avoid calling abort() from within the library when problems are
15134 detected, except if preprocessor symbols have been defined
15135 (such as REF_CHECK, BN_DEBUG etc.).
15139 * New openssl application 'rsautl'. This utility can be
15140 used for low level RSA operations. DER public key
15141 BIO/fp routines also added.
15145 * New Configure entry and patches for compiling on QNX 4.
15147 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15149 * A demo state-machine implementation was sponsored by
15150 Nuron (<http://www.nuron.com/>) and is now available in
15151 demos/state_machine.
15155 * New options added to the 'dgst' utility for signature
15156 generation and verification.
15160 * Unrecognized PKCS#7 content types are now handled via a
15161 catch all ASN1_TYPE structure. This allows unsupported
15162 types to be stored as a "blob" and an application can
15163 encode and decode it manually.
15167 * Fix various signed/unsigned issues to make a_strex.c
15168 compile under VC++.
15170 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15172 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15173 length if passed a buffer. ASN1_INTEGER_to_BN failed
15174 if passed a NULL BN and its argument was negative.
15176 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15178 * Modification to PKCS#7 encoding routines to output definite
15179 length encoding. Since currently the whole structures are in
15180 memory there's not real point in using indefinite length
15181 constructed encoding. However if OpenSSL is compiled with
15182 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15186 * Added BIO_vprintf() and BIO_vsnprintf().
15190 * Added more prefixes to parse for in the strings written
15191 through a logging bio, to cover all the levels that are available
15192 through syslog. The prefixes are now:
15194 PANIC, EMERG, EMR => LOG_EMERG
15195 ALERT, ALR => LOG_ALERT
15196 CRIT, CRI => LOG_CRIT
15197 ERROR, ERR => LOG_ERR
15198 WARNING, WARN, WAR => LOG_WARNING
15199 NOTICE, NOTE, NOT => LOG_NOTICE
15200 INFO, INF => LOG_INFO
15201 DEBUG, DBG => LOG_DEBUG
15203 and as before, if none of those prefixes are present at the
15204 beginning of the string, LOG_ERR is chosen.
15206 On Win32, the `LOG_*` levels are mapped according to this:
15208 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15209 LOG_WARNING => EVENTLOG_WARNING_TYPE
15210 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15214 * Made it possible to reconfigure with just the configuration
15215 argument "reconf" or "reconfigure". The command line arguments
15216 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15217 and are retrieved from there when reconfiguring.
15223 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15225 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15229 * The obj_dat.pl script was messing up the sorting of object
15230 names. The reason was that it compared the quoted version
15231 of strings as a result "OCSP" > "OCSP Signing" because
15232 " > SPACE. Changed script to store unquoted versions of
15233 names and add quotes on output. It was also omitting some
15234 names from the lookup table if they were given a default
15235 value (that is if SN is missing it is given the same
15236 value as LN and vice versa), these are now added on the
15237 grounds that if an object has a name we should be able to
15238 look it up. Finally added warning output when duplicate
15239 short or long names are found.
15243 * Changes needed for Tandem NSK.
15245 *Scott Uroff <scott@xypro.com>*
15247 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15248 RSA_padding_check_SSLv23(), special padding was never detected
15249 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15250 version rollback attacks was not effective.
15252 In s23_clnt.c, don't use special rollback-attack detection padding
15253 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15254 client; similarly, in s23_srvr.c, don't do the rollback check if
15255 SSL 2.0 is the only protocol enabled in the server.
15259 * Make it possible to get hexdumps of unprintable data with 'openssl
15260 asn1parse'. By implication, the functions ASN1_parse_dump() and
15261 BIO_dump_indent() are added.
15265 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15266 these print out strings and name structures based on various
15267 flags including RFC2253 support and proper handling of
15268 multibyte characters. Added options to the 'x509' utility
15269 to allow the various flags to be set.
15273 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15274 Also change the functions X509_cmp_current_time() and
15275 X509_gmtime_adj() work with an ASN1_TIME structure,
15276 this will enable certificates using GeneralizedTime in validity
15277 dates to be checked.
15281 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15282 negative public key encodings) on by default,
15283 NO_NEG_PUBKEY_BUG can be set to disable it.
15287 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15288 content octets. An i2c_ASN1_OBJECT is unnecessary because
15289 the encoding can be trivially obtained from the structure.
15293 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15294 not read locks (`CRYPTO_r_[un]lock`).
15298 * A first attempt at creating official support for shared
15299 libraries through configuration. I've kept it so the
15300 default is static libraries only, and the OpenSSL programs
15301 are always statically linked for now, but there are
15302 preparations for dynamic linking in place.
15303 This has been tested on Linux and Tru64.
15307 * Randomness polling function for Win9x, as described in:
15308 Peter Gutmann, Software Generation of Practically Strong
15313 * Fix so PRNG is seeded in req if using an already existing
15318 * New options to smime application. -inform and -outform
15319 allow alternative formats for the S/MIME message including
15320 PEM and DER. The -content option allows the content to be
15321 specified separately. This should allow things like Netscape
15322 form signing output easier to verify.
15326 * Fix the ASN1 encoding of tags using the 'long form'.
15330 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15331 STRING types. These convert content octets to and from the
15332 underlying type. The actual tag and length octets are
15333 already assumed to have been read in and checked. These
15334 are needed because all other string types have virtually
15335 identical handling apart from the tag. By having versions
15336 of the ASN1 functions that just operate on content octets
15337 IMPLICIT tagging can be handled properly. It also allows
15338 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15339 and ASN1_INTEGER are identical apart from the tag.
15343 * Change the handling of OID objects as follows:
15345 - New object identifiers are inserted in objects.txt, following
15346 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15347 - objects.pl is used to process obj_mac.num and create a new
15349 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15352 This is currently kind of a hack, and the perl code in objects.pl
15353 isn't very elegant, but it works as I intended. The simplest way
15354 to check that it worked correctly is to look in obj_dat.h and
15355 check the array nid_objs and make sure the objects haven't moved
15356 around (this is important!). Additions are OK, as well as
15357 consistent name changes.
15361 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15365 * Addition of the command line parameter '-rand file' to 'openssl req'.
15366 The given file adds to whatever has already been seeded into the
15367 random pool through the RANDFILE configuration file option or
15368 environment variable, or the default random state file.
15372 * mkstack.pl now sorts each macro group into lexical order.
15373 Previously the output order depended on the order the files
15374 appeared in the directory, resulting in needless rewriting
15379 * Patches to make OpenSSL compile under Win32 again. Mostly
15380 work arounds for the VC++ problem that it treats func() as
15381 func(void). Also stripped out the parts of mkdef.pl that
15382 added extra typesafe functions: these no longer exist.
15386 * Reorganisation of the stack code. The macros are now all
15387 collected in safestack.h . Each macro is defined in terms of
15388 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15389 DEBUG_SAFESTACK is now handled in terms of function casts,
15390 this has the advantage of retaining type safety without the
15391 use of additional functions. If DEBUG_SAFESTACK is not defined
15392 then the non typesafe macros are used instead. Also modified the
15393 mkstack.pl script to handle the new form. Needs testing to see
15394 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15395 the default if no major problems. Similar behaviour for ASN1_SET_OF
15396 and PKCS12_STACK_OF.
15400 * When some versions of IIS use the 'NET' form of private key the
15401 key derivation algorithm is different. Normally MD5(password) is
15402 used as a 128 bit RC4 key. In the modified case
15403 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15404 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15405 as the old Netscape_RSA functions except they have an additional
15406 'sgckey' parameter which uses the modified algorithm. Also added
15407 an -sgckey command line option to the rsa utility. Thanks to
15408 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15409 algorithm to openssl-dev.
15413 * The evp_local.h macros were using 'c.##kname' which resulted in
15414 invalid expansion on some systems (SCO 5.0.5 for example).
15415 Corrected to 'c.kname'.
15417 *Phillip Porch <root@theporch.com>*
15419 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15420 a STACK of email addresses from a certificate or request, these look
15421 in the subject name and the subject alternative name extensions and
15422 omit any duplicate addresses.
15426 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15427 This makes DSA verification about 2 % faster.
15431 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15432 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15433 plus overhead for 1024 bit moduli).
15434 This makes exponentiations about 0.5 % faster for 1024 bit
15435 exponents (as measured by "openssl speed rsa2048").
15439 * Rename memory handling macros to avoid conflicts with other
15441 Malloc => OPENSSL_malloc
15442 Malloc_locked => OPENSSL_malloc_locked
15443 Realloc => OPENSSL_realloc
15444 Free => OPENSSL_free
15448 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15449 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15453 * CygWin32 support.
15455 *John Jarvie <jjarvie@newsguy.com>*
15457 * The type-safe stack code has been rejigged. It is now only compiled
15458 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15459 by default all type-specific stack functions are "#define"d back to
15460 standard stack functions. This results in more streamlined output
15461 but retains the type-safety checking possibilities of the original
15466 * The STACK code has been cleaned up, and certain type declarations
15467 that didn't make a lot of sense have been brought in line. This has
15468 also involved a cleanup of sorts in safestack.h to more correctly
15469 map type-safe stack functions onto their plain stack counterparts.
15470 This work has also resulted in a variety of "const"ifications of
15471 lots of the code, especially `_cmp` operations which should normally
15472 be prototyped with "const" parameters anyway.
15476 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15477 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15478 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15479 where all of 'md' is used each time the PRNG is used, but 'state'
15480 is used only indexed by a cyclic counter. As entropy may not be
15481 well distributed from the beginning, 'md' is important as a
15482 chaining variable. However, the output function chains only half
15483 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15484 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15485 in all of 'state' being rewritten, with the new values depending
15486 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15490 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15491 the handshake is continued after ssl_verify_cert_chain();
15492 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15493 can lead to 'unexplainable' connection aborts later.
15495 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15497 * Major EVP API cipher revision.
15498 Add hooks for extra EVP features. This allows various cipher
15499 parameters to be set in the EVP interface. Support added for variable
15500 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15501 setting of RC2 and RC5 parameters.
15503 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15506 Remove lots of duplicated code from the EVP library. For example *every*
15507 cipher init() function handles the 'iv' in the same way according to the
15508 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15509 for CFB and OFB modes they zero ctx->num.
15511 New functionality allows removal of S/MIME code RC2 hack.
15513 Most of the routines have the same form and so can be declared in terms
15516 By shifting this to the top level EVP_CipherInit() it can be removed from
15517 all individual ciphers. If the cipher wants to handle IVs or keys
15518 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15521 Change lots of functions like EVP_EncryptUpdate() to now return a
15522 value: although software versions of the algorithms cannot fail
15523 any installed hardware versions can.
15527 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15528 this option is set, tolerate broken clients that send the negotiated
15529 protocol version number instead of the requested protocol version
15534 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15535 i.e. non-zero for export ciphersuites, zero otherwise.
15536 Previous versions had this flag inverted, inconsistent with
15537 rsa_tmp_cb (..._TMP_RSA_CB).
15539 *Bodo Moeller; problem reported by Amit Chopra*
15541 * Add missing DSA library text string. Work around for some IIS
15542 key files with invalid SEQUENCE encoding.
15546 * Add a document (doc/standards.txt) that list all kinds of standards
15547 and so on that are implemented in OpenSSL.
15551 * Enhance c_rehash script. Old version would mishandle certificates
15552 with the same subject name hash and wouldn't handle CRLs at all.
15553 Added -fingerprint option to crl utility, to support new c_rehash
15558 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15562 * Fix for SSL server purpose checking. Server checking was
15563 rejecting certificates which had extended key usage present
15564 but no ssl client purpose.
15566 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15568 * Make PKCS#12 code work with no password. The PKCS#12 spec
15569 is a little unclear about how a blank password is handled.
15570 Since the password in encoded as a BMPString with terminating
15571 double NULL a zero length password would end up as just the
15572 double NULL. However no password at all is different and is
15573 handled differently in the PKCS#12 key generation code. NS
15574 treats a blank password as zero length. MSIE treats it as no
15575 password on export: but it will try both on import. We now do
15576 the same: PKCS12_parse() tries zero length and no password if
15577 the password is set to "" or NULL (NULL is now a valid password:
15578 it wasn't before) as does the pkcs12 application.
15582 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15583 perror when PEM_read_bio_X509_REQ fails, the error message must
15584 be obtained from the error queue.
15588 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15589 it in ERR_remove_state if appropriate, and change ERR_get_state
15590 accordingly to avoid race conditions (this is necessary because
15591 thread_hash is no longer constant once set).
15595 * Bugfix for linux-elf makefile.one.
15599 * RSA_get_default_method() will now cause a default
15600 RSA_METHOD to be chosen if one doesn't exist already.
15601 Previously this was only set during a call to RSA_new()
15602 or RSA_new_method(NULL) meaning it was possible for
15603 RSA_get_default_method() to return NULL.
15607 * Added native name translation to the existing DSO code
15608 that will convert (if the flag to do so is set) filenames
15609 that are sufficiently small and have no path information
15610 into a canonical native form. Eg. "blah" converted to
15611 "libblah.so" or "blah.dll" etc.
15615 * New function ERR_error_string_n(e, buf, len) which is like
15616 ERR_error_string(e, buf), but writes at most 'len' bytes
15617 including the 0 terminator. For ERR_error_string_n, 'buf'
15620 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15622 * CONF library reworked to become more general. A new CONF
15623 configuration file reader "class" is implemented as well as a
15624 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15625 old `CONF_*` functions are still there, but are reimplemented to
15626 work in terms of the new functions. Also, a set of functions
15627 to handle the internal storage of the configuration data is
15628 provided to make it easier to write new configuration file
15629 reader "classes" (I can definitely see something reading a
15630 configuration file in XML format, for example), called `_CONF_*`,
15631 or "the configuration storage API"...
15633 The new configuration file reading functions are:
15635 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15636 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15638 NCONF_default, NCONF_WIN32
15640 NCONF_dump_fp, NCONF_dump_bio
15642 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15643 NCONF_new creates a new CONF object. This works in the same way
15644 as other interfaces in OpenSSL, like the BIO interface.
15645 `NCONF_dump_*` dump the internal storage of the configuration file,
15646 which is useful for debugging. All other functions take the same
15647 arguments as the old `CONF_*` functions with the exception of the
15648 first that must be a `CONF *` instead of a `LHASH *`.
15650 To make it easier to use the new classes with the old `CONF_*` functions,
15651 the function CONF_set_default_method is provided.
15655 * Add '-tls1' option to 'openssl ciphers', which was already
15656 mentioned in the documentation but had not been implemented.
15657 (This option is not yet really useful because even the additional
15658 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15662 * Initial DSO code added into libcrypto for letting OpenSSL (and
15663 OpenSSL-based applications) load shared libraries and bind to
15664 them in a portable way.
15666 *Geoff Thorpe, with contributions from Richard Levitte*
15668 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15670 * Make sure _lrotl and _lrotr are only used with MSVC.
15672 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15673 (the default implementation of RAND_status).
15675 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15676 to '-clrext' (= clear extensions), as intended and documented.
15677 *Bodo Moeller; inconsistency pointed out by Michael Attili
15678 <attili@amaxo.com>*
15680 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15681 was larger than the MD block size.
15683 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15685 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15686 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15687 using the passed key: if the passed key was a private key the result
15688 of X509_print(), for example, would be to print out all the private key
15693 * des_quad_cksum() byte order bug fix.
15694 *Ulf Möller, using the problem description in krb4-0.9.7, where
15695 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15697 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15700 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15702 * For easily testing in shell scripts whether some command
15703 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15704 returns with exit code 0 iff no command of the given name is available.
15705 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15706 the output goes to stdout and nothing is printed to stderr.
15707 Additional arguments are always ignored.
15709 Since for each cipher there is a command of the same name,
15710 the 'no-cipher' compilation switches can be tested this way.
15712 ('openssl no-XXX' is not able to detect pseudo-commands such
15713 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15717 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15721 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15722 is set; it will be thrown away anyway because each handshake creates
15724 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15725 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15726 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15727 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15731 * New s_client option -ign_eof: EOF at stdin is ignored, and
15732 'Q' and 'R' lose their special meanings (quit/renegotiate).
15733 This is part of what -quiet does; unlike -quiet, -ign_eof
15734 does not suppress any output.
15738 * Add compatibility options to the purpose and trust code. The
15739 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15740 accepts a certificate or CA, this was the previous behaviour,
15741 with all the associated security issues.
15743 X509_TRUST_COMPAT is the old trust behaviour: only and
15744 automatically trust self signed roots in certificate store. A
15745 new trust setting X509_TRUST_DEFAULT is used to specify that
15746 a purpose has no associated trust setting and it should instead
15747 use the value in the default purpose.
15751 * Fix the PKCS#8 DSA private key code so it decodes keys again
15752 and fix a memory leak.
15756 * In util/mkerr.pl (which implements 'make errors'), preserve
15757 reason strings from the previous version of the .c file, as
15758 the default to have only downcase letters (and digits) in
15759 automatically generated reasons codes is not always appropriate.
15763 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15764 using strerror. Previously, ERR_reason_error_string() returned
15765 library names as reason strings for SYSerr; but SYSerr is a special
15766 case where small numbers are errno values, not library numbers.
15770 * Add '-dsaparam' option to 'openssl dhparam' application. This
15771 converts DSA parameters into DH parameters. (When creating parameters,
15772 DSA_generate_parameters is used.)
15776 * Include 'length' (recommended exponent length) in C code generated
15777 by 'openssl dhparam -C'.
15781 * The second argument to set_label in perlasm was already being used
15782 so couldn't be used as a "file scope" flag. Moved to third argument
15787 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15788 instead of RAND_bytes for encryption IVs and salts.
15792 * Include RAND_status() into RAND_METHOD instead of implementing
15793 it only for md_rand.c Otherwise replacing the PRNG by calling
15794 RAND_set_rand_method would be impossible.
15798 * Don't let DSA_generate_key() enter an infinite loop if the random
15799 number generation fails.
15803 * New 'rand' application for creating pseudo-random output.
15807 * Added configuration support for Linux/IA64
15809 *Rolf Haberrecker <rolf@suse.de>*
15811 * Assembler module support for Mingw32.
15815 * Shared library support for HPUX (in shlib/).
15817 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15819 * Shared library support for Solaris gcc.
15821 *Lutz Behnke <behnke@trustcenter.de>*
15823 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15825 * PKCS7_encrypt() was adding text MIME headers twice because they
15826 were added manually and by SMIME_crlf_copy().
15830 * In bntest.c don't call BN_rand with zero bits argument.
15832 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15834 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15835 case was implemented. This caused BN_div_recp() to fail occasionally.
15839 * Add an optional second argument to the set_label() in the perl
15840 assembly language builder. If this argument exists and is set
15841 to 1 it signals that the assembler should use a symbol whose
15842 scope is the entire file, not just the current function. This
15843 is needed with MASM which uses the format label:: for this scope.
15845 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15847 * Change the ASN1 types so they are typedefs by default. Before
15848 almost all types were #define'd to ASN1_STRING which was causing
15849 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15854 * Change names of new functions to the new get1/get0 naming
15855 convention: After 'get1', the caller owns a reference count
15856 and has to call `..._free`; 'get0' returns a pointer to some
15857 data structure without incrementing reference counters.
15858 (Some of the existing 'get' functions increment a reference
15859 counter, some don't.)
15860 Similarly, 'set1' and 'add1' functions increase reference
15861 counters or duplicate objects.
15865 * Allow for the possibility of temp RSA key generation failure:
15866 the code used to assume it always worked and crashed on failure.
15870 * Fix potential buffer overrun problem in BIO_printf().
15871 *Ulf Möller, using public domain code by Patrick Powell; problem
15872 pointed out by David Sacerdote <das33@cornell.edu>*
15874 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15875 RAND_egd() and RAND_status(). In the command line application,
15876 the EGD socket can be specified like a seed file using RANDFILE
15881 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15882 Some CAs (e.g. Verisign) distribute certificates in this form.
15886 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15887 list to exclude them. This means that no special compilation option
15888 is needed to use anonymous DH: it just needs to be included in the
15893 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15894 EVP_MD_type. The old functionality is available in a new macro called
15895 EVP_MD_md(). Change code that uses it and update docs.
15899 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15900 where the `void *` argument is replaced by a function pointer argument.
15901 Previously `void *` was abused to point to functions, which works on
15902 many platforms, but is not correct. As these functions are usually
15903 called by macros defined in OpenSSL header files, most source code
15904 should work without changes.
15908 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15909 sections with information on -D... compiler switches used for
15910 compiling the library so that applications can see them. To enable
15911 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15912 must be defined. E.g.,
15913 #define OPENSSL_ALGORITHM_DEFINES
15914 #include <openssl/opensslconf.h>
15915 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15917 *Richard Levitte, Ulf and Bodo Möller*
15919 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15924 * Change the 'other' type in certificate aux info to a STACK_OF
15925 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15926 the required ASN1 format: arbitrary types determined by an OID.
15930 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15931 argument to 'req'. This is not because the function is newer or
15932 better than others it just uses the work 'NEW' in the certificate
15933 request header lines. Some software needs this.
15937 * Reorganise password command line arguments: now passwords can be
15938 obtained from various sources. Delete the PEM_cb function and make
15939 it the default behaviour: i.e. if the callback is NULL and the
15940 usrdata argument is not NULL interpret it as a null terminated pass
15941 phrase. If usrdata and the callback are NULL then the pass phrase
15942 is prompted for as usual.
15946 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15947 the support is automatically enabled. The resulting binaries will
15948 autodetect the card and use it if present.
15950 *Ben Laurie and Compaq Inc.*
15952 * Work around for Netscape hang bug. This sends certificate request
15953 and server done in one record. Since this is perfectly legal in the
15954 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15955 the bugs/SSLv3 entry for more info.
15959 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15963 * Add -rand argument to smime and pkcs12 applications and read/write
15968 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15972 * Add command line password options to the remaining applications.
15976 * Bug fix for BN_div_recp() for numerators with an even number of
15981 * More tests in bntest.c, and changed test_bn output.
15985 * ./config recognizes MacOS X now.
15989 * Bug fix for BN_div() when the first words of num and divisor are
15990 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15994 * Add support for various broken PKCS#8 formats, and command line
15995 options to produce them.
15999 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16000 get temporary BIGNUMs from a BN_CTX.
16004 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16009 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16010 include a #define from the old name to the new. The original intent
16011 was that statically linked binaries could for example just call
16012 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16013 link with digests. This never worked because SSLeay_add_all_digests()
16014 and SSLeay_add_all_ciphers() were in the same source file so calling
16015 one would link with the other. They are now in separate source files.
16019 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16023 * Use a less unusual form of the Miller-Rabin primality test (it used
16024 a binary algorithm for exponentiation integrated into the Miller-Rabin
16025 loop, our standard modexp algorithms are faster).
16029 * Support for the EBCDIC character set completed.
16031 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16033 * Source code cleanups: use const where appropriate, eliminate casts,
16034 use `void *` instead of `char *` in lhash.
16038 * Bugfix: ssl3_send_server_key_exchange was not restartable
16039 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16040 this the server could overwrite ephemeral keys that the client
16045 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16046 using 50 iterations of the Rabin-Miller test.
16048 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16049 iterations of the Rabin-Miller test as required by the appendix
16050 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16051 As BN_is_prime_fasttest includes trial division, DSA parameter
16052 generation becomes much faster.
16054 This implies a change for the callback functions in DSA_is_prime
16055 and DSA_generate_parameters: The callback function is called once
16056 for each positive witness in the Rabin-Miller test, not just
16057 occasionally in the inner loop; and the parameters to the
16058 callback function now provide an iteration count for the outer
16059 loop rather than for the current invocation of the inner loop.
16060 DSA_generate_parameters additionally can call the callback
16061 function with an 'iteration count' of -1, meaning that a
16062 candidate has passed the trial division test (when q is generated
16063 from an application-provided seed, trial division is skipped).
16067 * New function BN_is_prime_fasttest that optionally does trial
16068 division before starting the Rabin-Miller test and has
16069 an additional BN_CTX * argument (whereas BN_is_prime always
16070 has to allocate at least one BN_CTX).
16071 'callback(1, -1, cb_arg)' is called when a number has passed the
16072 trial division stage.
16076 * Fix for bug in CRL encoding. The validity dates weren't being handled
16081 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16085 * New function BN_pseudo_rand().
16089 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16090 bignum version of BN_from_montgomery() with the working code from
16091 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16096 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16097 made it impossible to use the same SSL_SESSION data structure in
16098 SSL2 clients in multiple threads.
16102 * The return value of RAND_load_file() no longer counts bytes obtained
16103 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16104 to seed the PRNG (previously an explicit byte count was required).
16106 *Ulf Möller, Bodo Möller*
16108 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16109 used `char *` instead of `void *` and had casts all over the place.
16113 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16117 * Retain source code compatibility for BN_prime_checks macro:
16118 BN_is_prime(..., BN_prime_checks, ...) now uses
16119 BN_prime_checks_for_size to determine the appropriate number of
16120 Rabin-Miller iterations.
16124 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16125 DH_CHECK_P_NOT_SAFE_PRIME.
16126 (Check if this is true? OpenPGP calls them "strong".)
16130 * Merge the functionality of "dh" and "gendh" programs into a new program
16131 "dhparam". The old programs are retained for now but will handle DH keys
16132 (instead of parameters) in future.
16136 * Make the ciphers, s_server and s_client programs check the return values
16137 when a new cipher list is set.
16141 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16142 ciphers. Before when the 56bit ciphers were enabled the sorting was
16145 The syntax for the cipher sorting has been extended to support sorting by
16146 cipher-strength (using the strength_bits hard coded in the tables).
16147 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16149 Fix a bug in the cipher-command parser: when supplying a cipher command
16150 string with an "undefined" symbol (neither command nor alphanumeric
16151 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16152 an error is flagged.
16154 Due to the strength-sorting extension, the code of the
16155 ssl_create_cipher_list() function was completely rearranged. I hope that
16156 the readability was also increased :-)
16158 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16160 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16161 for the first serial number and places 2 in the serial number file. This
16162 avoids problems when the root CA is created with serial number zero and
16163 the first user certificate has the same issuer name and serial number
16168 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16169 the new code. Add documentation for this stuff.
16173 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16174 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16175 structures and behave in an analogous way to the X509v3 functions:
16176 they shouldn't be called directly but wrapper functions should be used
16179 So we also now have some wrapper functions that call the X509at functions
16180 when passed certificate requests. (TO DO: similar things can be done with
16181 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16182 things. Some of these need some d2i or i2d and print functionality
16183 because they handle more complex structures.)
16187 * Add missing #ifndefs that caused missing symbols when building libssl
16188 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16189 NO_RSA in `ssl/s2*.c`.
16191 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16193 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16194 has a return value which indicates the quality of the random data
16195 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16196 error queue. New function RAND_pseudo_bytes() generates output that is
16197 guaranteed to be unique but not unpredictable. RAND_add is like
16198 RAND_seed, but takes an extra argument for an entropy estimate
16199 (RAND_seed always assumes full entropy).
16203 * Do more iterations of Rabin-Miller probable prime test (specifically,
16204 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16205 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16206 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16207 false-positive rate of at most 2^-80 for random input.
16211 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16215 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16216 in the 0.9.5 release), this returns the chain
16217 from an X509_CTX structure with a dup of the stack and all
16218 the X509 reference counts upped: so the stack will exist
16219 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16222 Also make SSL_SESSION_print() print out the verify return
16227 * Add manpage for the pkcs12 command. Also change the default
16228 behaviour so MAC iteration counts are used unless the new
16229 -nomaciter option is used. This improves file security and
16230 only older versions of MSIE (4.0 for example) need it.
16234 * Honor the no-xxx Configure options when creating .DEF files.
16238 * Add PKCS#10 attributes to field table: challengePassword,
16239 unstructuredName and unstructuredAddress. These are taken from
16240 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16241 international characters are used.
16243 More changes to X509_ATTRIBUTE code: allow the setting of types
16244 based on strings. Remove the 'loc' parameter when adding
16245 attributes because these will be a SET OF encoding which is sorted
16250 * Initial changes to the 'req' utility to allow request generation
16251 automation. This will allow an application to just generate a template
16252 file containing all the field values and have req construct the
16255 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16256 used all over the place including certificate requests and PKCS#7
16257 structures. They are currently handled manually where necessary with
16258 some primitive wrappers for PKCS#7. The new functions behave in a
16259 manner analogous to the X509 extension functions: they allow
16260 attributes to be looked up by NID and added.
16262 Later something similar to the X509V3 code would be desirable to
16263 automatically handle the encoding, decoding and printing of the
16264 more complex types. The string types like challengePassword can
16265 be handled by the string table functions.
16267 Also modified the multi byte string table handling. Now there is
16268 a 'global mask' which masks out certain types. The table itself
16269 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16270 is useful when for example there is only one permissible type
16271 (as in countryName) and using the mask might result in no valid
16276 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16277 SSL_get_peer_finished to allow applications to obtain the latest
16278 Finished messages sent to the peer or expected from the peer,
16279 respectively. (SSL_get_peer_finished is usually the Finished message
16280 actually received from the peer, otherwise the protocol will be aborted.)
16282 As the Finished message are message digests of the complete handshake
16283 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16284 be used for external authentication procedures when the authentication
16285 provided by SSL/TLS is not desired or is not enough.
16289 * Enhanced support for Alpha Linux is added. Now ./config checks if
16290 the host supports BWX extension and if Compaq C is present on the
16291 $PATH. Just exploiting of the BWX extension results in 20-30%
16292 performance kick for some algorithms, e.g. DES and RC4 to mention
16293 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16298 * Add support for MS "fast SGC". This is arguably a violation of the
16299 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16300 weak crypto and after checking the certificate is SGC a second one
16301 with strong crypto. MS SGC stops the first handshake after receiving
16302 the server certificate message and sends a second client hello. Since
16303 a server will typically do all the time consuming operations before
16304 expecting any further messages from the client (server key exchange
16305 is the most expensive) there is little difference between the two.
16307 To get OpenSSL to support MS SGC we have to permit a second client
16308 hello message after we have sent server done. In addition we have to
16309 reset the MAC if we do get this second client hello.
16313 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16314 if a DER encoded private key is RSA or DSA traditional format. Changed
16315 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16316 format DER encoded private key. Newer code should use PKCS#8 format which
16317 has the key type encoded in the ASN1 structure. Added DER private key
16318 support to pkcs8 application.
16322 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16323 ciphersuites has been selected (as required by the SSL 3/TLS 1
16324 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16325 is set, we interpret this as a request to violate the specification
16326 (the worst that can happen is a handshake failure, and 'correct'
16327 behaviour would result in a handshake failure anyway).
16331 * In SSL_CTX_add_session, take into account that there might be multiple
16332 SSL_SESSION structures with the same session ID (e.g. when two threads
16333 concurrently obtain them from an external cache).
16334 The internal cache can handle only one SSL_SESSION with a given ID,
16335 so if there's a conflict, we now throw out the old one to achieve
16340 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16341 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16342 some routines that use cipher OIDs: some ciphers do not have OIDs
16343 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16348 * Simplify the trust setting structure and code. Now we just have
16349 two sequences of OIDs for trusted and rejected settings. These will
16350 typically have values the same as the extended key usage extension
16351 and any application specific purposes.
16353 The trust checking code now has a default behaviour: it will just
16354 check for an object with the same NID as the passed id. Functions can
16355 be provided to override either the default behaviour or the behaviour
16356 for a given id. SSL client, server and email already have functions
16357 in place for compatibility: they check the NID and also return "trusted"
16358 if the certificate is self signed.
16362 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16363 traditional format into an EVP_PKEY structure.
16367 * Add a password callback function PEM_cb() which either prompts for
16368 a password if usr_data is NULL or otherwise assumes it is a null
16369 terminated password. Allow passwords to be passed on command line
16370 environment or config files in a few more utilities.
16374 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16375 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16376 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16377 Update documentation.
16381 * Support for ASN1 "NULL" type. This could be handled before by using
16382 ASN1_TYPE but there wasn't any function that would try to read a NULL
16383 and produce an error if it couldn't. For compatibility we also have
16384 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16385 don't allocate anything because they don't need to.
16389 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16392 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16394 * Rebuild of the memory allocation routines used by OpenSSL code and
16395 possibly others as well. The purpose is to make an interface that
16396 provide hooks so anyone can build a separate set of allocation and
16397 deallocation routines to be used by OpenSSL, for example memory
16398 pool implementations, or something else, which was previously hard
16399 since Malloc(), Realloc() and Free() were defined as macros having
16400 the values malloc, realloc and free, respectively (except for Win32
16401 compilations). The same is provided for memory debugging code.
16402 OpenSSL already comes with functionality to find memory leaks, but
16403 this gives people a chance to debug other memory problems.
16405 With these changes, a new set of functions and macros have appeared:
16407 CRYPTO_set_mem_debug_functions() [F]
16408 CRYPTO_get_mem_debug_functions() [F]
16409 CRYPTO_dbg_set_options() [F]
16410 CRYPTO_dbg_get_options() [F]
16411 CRYPTO_malloc_debug_init() [M]
16413 The memory debug functions are NULL by default, unless the library
16414 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16415 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16416 gives the standard debugging functions that come with OpenSSL) or
16417 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16418 provided by the library user) must be used. When the standard
16419 debugging functions are used, CRYPTO_dbg_set_options can be used to
16420 request additional information:
16421 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16422 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16424 Also, things like CRYPTO_set_mem_functions will always give the
16425 expected result (the new set of functions is used for allocation
16426 and deallocation) at all times, regardless of platform and compiler
16429 To finish it up, some functions that were never use in any other
16430 way than through macros have a new API and new semantic:
16432 CRYPTO_dbg_malloc()
16433 CRYPTO_dbg_realloc()
16436 All macros of value have retained their old syntax.
16438 *Richard Levitte and Bodo Moeller*
16440 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16441 ordering of SMIMECapabilities wasn't in "strength order" and there
16442 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16447 * Some ASN1 types with illegal zero length encoding (INTEGER,
16448 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16450 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16452 * Merge in my S/MIME library for OpenSSL. This provides a simple
16453 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16454 functionality to handle multipart/signed properly) and a utility
16455 called 'smime' to call all this stuff. This is based on code I
16456 originally wrote for Celo who have kindly allowed it to be
16457 included in OpenSSL.
16461 * Add variants des_set_key_checked and des_set_key_unchecked of
16462 des_set_key (aka des_key_sched). Global variable des_check_key
16463 decides which of these is called by des_set_key; this way
16464 des_check_key behaves as it always did, but applications and
16465 the library itself, which was buggy for des_check_key == 1,
16466 have a cleaner way to pick the version they need.
16470 * New function PKCS12_newpass() which changes the password of a
16475 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16476 dynamic mix. In both cases the ids can be used as an index into the
16477 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16478 functions so they accept a list of the field values and the
16479 application doesn't need to directly manipulate the X509_TRUST
16484 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16489 * Modify the way the V3 extension code looks up extensions. This now
16490 works in a similar way to the object code: we have some "standard"
16491 extensions in a static table which is searched with OBJ_bsearch()
16492 and the application can add dynamic ones if needed. The file
16493 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16494 updated whenever a new extension is added to the core code and kept
16495 in ext_nid order. There is a simple program 'tabtest.c' which checks
16496 this. New extensions are not added too often so this file can readily
16497 be maintained manually.
16499 There are two big advantages in doing things this way. The extensions
16500 can be looked up immediately and no longer need to be "added" using
16501 X509V3_add_standard_extensions(): this function now does nothing.
16502 Side note: I get *lots* of email saying the extension code doesn't
16503 work because people forget to call this function.
16504 Also no dynamic allocation is done unless new extensions are added:
16505 so if we don't add custom extensions there is no need to call
16506 X509V3_EXT_cleanup().
16510 * Modify enc utility's salting as follows: make salting the default. Add a
16511 magic header, so unsalted files fail gracefully instead of just decrypting
16512 to garbage. This is because not salting is a big security hole, so people
16513 should be discouraged from doing it.
16517 * Fixes and enhancements to the 'x509' utility. It allowed a message
16518 digest to be passed on the command line but it only used this
16519 parameter when signing a certificate. Modified so all relevant
16520 operations are affected by the digest parameter including the
16521 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16522 DSA key was used because it didn't fix the digest.
16526 * Initial certificate chain verify code. Currently tests the untrusted
16527 certificates for consistency with the verify purpose (which is set
16528 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16530 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16531 this is because it will reject chains with invalid extensions whereas
16532 every previous version of OpenSSL and SSLeay made no checks at all.
16534 Trust code: checks the root CA for the relevant trust settings. Trust
16535 settings have an initial value consistent with the verify purpose: e.g.
16536 if the verify purpose is for SSL client use it expects the CA to be
16537 trusted for SSL client use. However the default value can be changed to
16538 permit custom trust settings: one example of this would be to only trust
16539 certificates from a specific "secure" set of CAs.
16541 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16542 which should be used for version portability: especially since the
16543 verify structure is likely to change more often now.
16545 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16546 to set them. If not set then assume SSL clients will verify SSL servers
16549 Two new options to the verify program: -untrusted allows a set of
16550 untrusted certificates to be passed in and -purpose which sets the
16551 intended purpose of the certificate. If a purpose is set then the
16552 new chain verify code is used to check extension consistency.
16556 * Support for the authority information access extension.
16560 * Modify RSA and DSA PEM read routines to transparently handle
16561 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16562 public keys in a format compatible with certificate
16563 SubjectPublicKeyInfo structures. Unfortunately there were already
16564 functions called *_PublicKey_* which used various odd formats so
16565 these are retained for compatibility: however the DSA variants were
16566 never in a public release so they have been deleted. Changed dsa/rsa
16567 utilities to handle the new format: note no releases ever handled public
16568 keys so we should be OK.
16570 The primary motivation for this change is to avoid the same fiasco
16571 that dogs private keys: there are several incompatible private key
16572 formats some of which are standard and some OpenSSL specific and
16573 require various evil hacks to allow partial transparent handling and
16574 even then it doesn't work with DER formats. Given the option anything
16575 other than PKCS#8 should be dumped: but the other formats have to
16576 stay in the name of compatibility.
16578 With public keys and the benefit of hindsight one standard format
16579 is used which works with EVP_PKEY, RSA or DSA structures: though
16580 it clearly returns an error if you try to read the wrong kind of key.
16582 Added a -pubkey option to the 'x509' utility to output the public key.
16583 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16584 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16585 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16586 that do the same as the `EVP_PKEY_assign_*()` except they up the
16587 reference count of the added key (they don't "swallow" the
16592 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16593 CRLs would fail if the file contained no certificates or no CRLs:
16594 added a new function to read in both types and return the number
16595 read: this means that if none are read it will be an error. The
16596 DER versions of the certificate and CRL reader would always fail
16597 because it isn't possible to mix certificates and CRLs in DER format
16598 without choking one or the other routine. Changed this to just read
16599 a certificate: this is the best we can do. Also modified the code
16600 in `apps/verify.c` to take notice of return codes: it was previously
16601 attempting to read in certificates from NULL pointers and ignoring
16602 any errors: this is one reason why the cert and CRL reader seemed
16603 to work. It doesn't check return codes from the default certificate
16604 routines: these may well fail if the certificates aren't installed.
16608 * Code to support otherName option in GeneralName.
16612 * First update to verify code. Change the verify utility
16613 so it warns if it is passed a self signed certificate:
16614 for consistency with the normal behaviour. X509_verify
16615 has been modified to it will now verify a self signed
16616 certificate if *exactly* the same certificate appears
16617 in the store: it was previously impossible to trust a
16618 single self signed certificate. This means that:
16619 openssl verify ss.pem
16620 now gives a warning about a self signed certificate but
16621 openssl verify -CAfile ss.pem ss.pem
16626 * For servers, store verify_result in SSL_SESSION data structure
16627 (and add it to external session representation).
16628 This is needed when client certificate verifications fails,
16629 but an application-provided verification callback (set by
16630 SSL_CTX_set_cert_verify_callback) allows accepting the session
16631 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16632 but returns 1): When the session is reused, we have to set
16633 ssl->verify_result to the appropriate error code to avoid
16636 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16638 * Fix a bug in the new PKCS#7 code: it didn't consider the
16639 case in PKCS7_dataInit() where the signed PKCS7 structure
16640 didn't contain any existing data because it was being created.
16642 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16644 * Add a salt to the key derivation routines in enc.c. This
16645 forms the first 8 bytes of the encrypted file. Also add a
16646 -S option to allow a salt to be input on the command line.
16650 * New function X509_cmp(). Oddly enough there wasn't a function
16651 to compare two certificates. We do this by working out the SHA1
16652 hash and comparing that. X509_cmp() will be needed by the trust
16657 * SSL_get1_session() is like SSL_get_session(), but increments
16658 the reference count in the SSL_SESSION returned.
16660 *Geoff Thorpe <geoff@eu.c2.net>*
16662 * Fix for 'req': it was adding a null to request attributes.
16663 Also change the X509_LOOKUP and X509_INFO code to handle
16664 certificate auxiliary information.
16668 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16673 * Add the possibility to add extra information to the memory leak
16674 detecting output, to form tracebacks, showing from where each
16675 allocation was originated: CRYPTO_push_info("constant string") adds
16676 the string plus current file name and line number to a per-thread
16677 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16678 is like calling CYRPTO_pop_info() until the stack is empty.
16679 Also updated memory leak detection code to be multi-thread-safe.
16683 * Add options -text and -noout to pkcs7 utility and delete the
16684 encryption options which never did anything. Update docs.
16688 * Add options to some of the utilities to allow the pass phrase
16689 to be included on either the command line (not recommended on
16690 OSes like Unix) or read from the environment. Update the
16691 manpages and fix a few bugs.
16695 * Add a few manpages for some of the openssl commands.
16699 * Fix the -revoke option in ca. It was freeing up memory twice,
16700 leaking and not finding already revoked certificates.
16704 * Extensive changes to support certificate auxiliary information.
16705 This involves the use of X509_CERT_AUX structure and X509_AUX
16706 functions. An X509_AUX function such as PEM_read_X509_AUX()
16707 can still read in a certificate file in the usual way but it
16708 will also read in any additional "auxiliary information". By
16709 doing things this way a fair degree of compatibility can be
16710 retained: existing certificates can have this information added
16711 using the new 'x509' options.
16713 Current auxiliary information includes an "alias" and some trust
16714 settings. The trust settings will ultimately be used in enhanced
16715 certificate chain verification routines: currently a certificate
16716 can only be trusted if it is self signed and then it is trusted
16721 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16722 The problem was that one of the replacement routines had not been working
16723 since SSLeay releases. For now the offending routine has been replaced
16724 with non-optimised assembler. Even so, this now gives around 95%
16725 performance improvement for 1024 bit RSA signs.
16729 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16730 handling. Most clients have the effective key size in bits equal to
16731 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16732 A few however don't do this and instead use the size of the decrypted key
16733 to determine the RC2 key length and the AlgorithmIdentifier to determine
16734 the effective key length. In this case the effective key length can still
16735 be 40 bits but the key length can be 168 bits for example. This is fixed
16736 by manually forcing an RC2 key into the EVP_PKEY structure because the
16737 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16738 the key length and effective key length are equal.
16742 * Add a bunch of functions that should simplify the creation of
16743 X509_NAME structures. Now you should be able to do:
16744 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16745 and have it automatically work out the correct field type and fill in
16746 the structures. The more adventurous can try:
16747 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16748 and it will (hopefully) work out the correct multibyte encoding.
16752 * Change the 'req' utility to use the new field handling and multibyte
16753 copy routines. Before the DN field creation was handled in an ad hoc
16754 way in req, ca, and x509 which was rather broken and didn't support
16755 BMPStrings or UTF8Strings. Since some software doesn't implement
16756 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16757 using the dirstring_type option. See the new comment in the default
16758 openssl.cnf for more info.
16762 * Make crypto/rand/md_rand.c more robust:
16763 - Assure unique random numbers after fork().
16764 - Make sure that concurrent threads access the global counter and
16765 md serializably so that we never lose entropy in them
16766 or use exactly the same state in multiple threads.
16767 Access to the large state is not always serializable because
16768 the additional locking could be a performance killer, and
16769 md should be large enough anyway.
16773 * New file `apps/app_rand.c` with commonly needed functionality
16774 for handling the random seed file.
16776 Use the random seed file in some applications that previously did not:
16778 dsaparam -genkey (which also ignored its '-rand' option),
16781 x509 (when signing).
16782 Except on systems with /dev/urandom, it is crucial to have a random
16783 seed file at least for key creation, DSA signing, and for DH exchanges;
16784 for RSA signatures we could do without one.
16786 gendh and gendsa (unlike genrsa) used to read only the first byte
16787 of each file listed in the '-rand' option. The function as previously
16788 found in genrsa is now in app_rand.c and is used by all programs
16789 that support '-rand'.
16793 * In RAND_write_file, use mode 0600 for creating files;
16794 don't just chmod when it may be too late.
16798 * Report an error from X509_STORE_load_locations
16799 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16803 * New function ASN1_mbstring_copy() this copies a string in either
16804 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16805 into an ASN1_STRING type. A mask of permissible types is passed
16806 and it chooses the "minimal" type to use or an error if not type
16811 * Add function equivalents to the various macros in asn1.h. The old
16812 macros are retained with an `M_` prefix. Code inside the library can
16813 use the `M_` macros. External code (including the openssl utility)
16814 should *NOT* in order to be "shared library friendly".
16818 * Add various functions that can check a certificate's extensions
16819 to see if it usable for various purposes such as SSL client,
16820 server or S/MIME and CAs of these types. This is currently
16821 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16822 verification. Also added a -purpose flag to x509 utility to
16823 print out all the purposes.
16827 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16832 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16833 for, obtain and decode and extension and obtain its critical flag.
16834 This allows all the necessary extension code to be handled in a
16835 single function call.
16839 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16840 platforms. See crypto/rc4/rc4_enc.c for further details.
16844 * New -noout option to asn1parse. This causes no output to be produced
16845 its main use is when combined with -strparse and -out to extract data
16846 from a file (which may not be in ASN.1 format).
16850 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16851 when producing the local key id.
16853 *Richard Levitte <levitte@stacken.kth.se>*
16855 * New option -dhparam in s_server. This allows a DH parameter file to be
16856 stated explicitly. If it is not stated then it tries the first server
16857 certificate file. The previous behaviour hard coded the filename
16862 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16863 a public key to be input or output. For example:
16864 openssl rsa -in key.pem -pubout -out pubkey.pem
16865 Also added necessary DSA public key functions to handle this.
16869 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16870 in the message. This was handled by allowing
16871 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16873 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16875 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16876 to the end of the strings whereas this didn't. This would cause problems
16877 if strings read with d2i_ASN1_bytes() were later modified.
16879 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16881 * Fix for base64 decode bug. When a base64 bio reads only one line of
16882 data and it contains EOF it will end up returning an error. This is
16883 caused by input 46 bytes long. The cause is due to the way base64
16884 BIOs find the start of base64 encoded data. They do this by trying a
16885 trial decode on each line until they find one that works. When they
16886 do a flag is set and it starts again knowing it can pass all the
16887 data directly through the decoder. Unfortunately it doesn't reset
16888 the context it uses. This means that if EOF is reached an attempt
16889 is made to pass two EOFs through the context and this causes the
16890 resulting error. This can also cause other problems as well. As is
16891 usual with these problems it takes *ages* to find and the fix is
16892 trivial: move one line.
16894 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16896 * Ugly workaround to get s_client and s_server working under Windows. The
16897 old code wouldn't work because it needed to select() on sockets and the
16898 tty (for keypresses and to see if data could be written). Win32 only
16899 supports select() on sockets so we select() with a 1s timeout on the
16900 sockets and then see if any characters are waiting to be read, if none
16901 are present then we retry, we also assume we can always write data to
16902 the tty. This isn't nice because the code then blocks until we've
16903 received a complete line of data and it is effectively polling the
16904 keyboard at 1s intervals: however it's quite a bit better than not
16905 working at all :-) A dedicated Windows application might handle this
16906 with an event loop for example.
16910 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16911 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16912 will be called when RSA_sign() and RSA_verify() are used. This is useful
16913 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16914 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16915 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16916 This necessitated the support of an extra signature type NID_md5_sha1
16917 for SSL signatures and modifications to the SSL library to use it instead
16918 of calling RSA_public_decrypt() and RSA_private_encrypt().
16922 * Add new -verify -CAfile and -CApath options to the crl program, these
16923 will lookup a CRL issuers certificate and verify the signature in a
16924 similar way to the verify program. Tidy up the crl program so it
16925 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16926 less strict. It will now permit CRL extensions even if it is not
16927 a V2 CRL: this will allow it to tolerate some broken CRLs.
16931 * Initialize all non-automatic variables each time one of the openssl
16932 sub-programs is started (this is necessary as they may be started
16933 multiple times from the "OpenSSL>" prompt).
16935 *Lennart Bang, Bodo Moeller*
16937 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16938 removing all other RSA functionality (this is what NO_RSA does). This
16939 is so (for example) those in the US can disable those operations covered
16940 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16945 * Non-copying interface to BIO pairs.
16946 (still largely untested)
16950 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16951 ASCII string. This was handled independently in various places before.
16955 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16956 UTF8 strings a character at a time.
16960 * Use client_version from client hello to select the protocol
16961 (s23_srvr.c) and for RSA client key exchange verification
16962 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16966 * Add various utility functions to handle SPKACs, these were previously
16967 handled by poking round in the structure internals. Added new function
16968 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16969 print, verify and generate SPKACs. Based on an original idea from
16970 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16974 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16978 * Allow the config file extension section to be overwritten on the
16979 command line. Based on an original idea from Massimiliano Pala
16980 <madwolf@comune.modena.it>. The new option is called -extensions
16981 and can be applied to ca, req and x509. Also -reqexts to override
16982 the request extensions in req and -crlexts to override the crl extensions
16987 * Add new feature to the SPKAC handling in ca. Now you can include
16988 the same field multiple times by preceding it by "XXXX." for example:
16991 this is the same syntax as used in the req config file.
16995 * Allow certificate extensions to be added to certificate requests. These
16996 are specified in a 'req_extensions' option of the req section of the
16997 config file. They can be printed out with the -text option to req but
16998 are otherwise ignored at present.
17002 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17003 data read consists of only the final block it would not decrypted because
17004 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17005 A misplaced 'break' also meant the decrypted final block might not be
17006 copied until the next read.
17010 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17011 a few extra parameters to the DH structure: these will be useful if
17012 for example we want the value of 'q' or implement X9.42 DH.
17016 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17017 provides hooks that allow the default DSA functions or functions on a
17018 "per key" basis to be replaced. This allows hardware acceleration and
17019 hardware key storage to be handled without major modification to the
17020 library. Also added low level modexp hooks and CRYPTO_EX structure and
17021 associated functions.
17025 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17026 as "read only": it can't be written to and the buffer it points to will
17027 not be freed. Reading from a read only BIO is much more efficient than
17028 a normal memory BIO. This was added because there are several times when
17029 an area of memory needs to be read from a BIO. The previous method was
17030 to create a memory BIO and write the data to it, this results in two
17031 copies of the data and an O(n^2) reading algorithm. There is a new
17032 function BIO_new_mem_buf() which creates a read only memory BIO from
17033 an area of memory. Also modified the PKCS#7 routines to use read only
17038 * Bugfix: ssl23_get_client_hello did not work properly when called in
17039 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17040 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17041 but a retry condition occurred while trying to read the rest.
17045 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17046 NID_pkcs7_encrypted by default: this was wrong since this should almost
17047 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17048 the encrypted data type: this is a more sensible place to put it and it
17049 allows the PKCS#12 code to be tidied up that duplicated this
17054 * Changed obj_dat.pl script so it takes its input and output files on
17055 the command line. This should avoid shell escape redirection problems
17060 * Initial support for certificate extension requests, these are included
17061 in things like Xenroll certificate requests. Included functions to allow
17062 extensions to be obtained and added.
17066 * -crlf option to s_client and s_server for sending newlines as
17067 CRLF (as required by many protocols).
17071 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17073 * Install libRSAglue.a when OpenSSL is built with RSAref.
17075 *Ralf S. Engelschall*
17077 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17079 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17081 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17086 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17087 DH parameters/keys (q is lost during that conversion, but the resulting
17088 DH parameters contain its length).
17090 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17091 much faster than DH_generate_parameters (which creates parameters
17092 where `p = 2*q + 1`), and also the smaller q makes DH computations
17093 much more efficient (160-bit exponentiation instead of 1024-bit
17094 exponentiation); so this provides a convenient way to support DHE
17095 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17096 utter importance to use
17097 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17099 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17100 when such DH parameters are used, because otherwise small subgroup
17101 attacks may become possible!
17105 * Avoid memory leak in i2d_DHparams.
17109 * Allow the -k option to be used more than once in the enc program:
17110 this allows the same encrypted message to be read by multiple recipients.
17114 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17115 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17116 it will always use the numerical form of the OID, even if it has a short
17121 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17122 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17123 otherwise bn_mod_exp was called. In the case of hardware keys for example
17124 no private key components need be present and it might store extra data
17125 in the RSA structure, which cannot be accessed from bn_mod_exp.
17126 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17127 private key operations.
17131 * Added support for SPARC Linux.
17135 * pem_password_cb function type incompatibly changed from
17136 typedef int pem_password_cb(char *buf, int size, int rwflag);
17138 ....(char *buf, int size, int rwflag, void *userdata);
17139 so that applications can pass data to their callbacks:
17140 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17141 additional void * argument, which is just handed through whenever
17142 the password callback is called.
17144 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17146 New function SSL_CTX_set_default_passwd_cb_userdata.
17148 Compatibility note: As many C implementations push function arguments
17149 onto the stack in reverse order, the new library version is likely to
17150 interoperate with programs that have been compiled with the old
17151 pem_password_cb definition (PEM_whatever takes some data that
17152 happens to be on the stack as its last argument, and the callback
17153 just ignores this garbage); but there is no guarantee whatsoever that
17156 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17157 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17158 problems not only on Windows, but also on some Unix platforms.
17159 To avoid problematic command lines, these definitions are now in an
17160 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17161 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17165 * MIPS III/IV assembler module is reimplemented.
17169 * More DES library cleanups: remove references to srand/rand and
17170 delete an unused file.
17174 * Add support for the free Netwide assembler (NASM) under Win32,
17175 since not many people have MASM (ml) and it can be hard to obtain.
17176 This is currently experimental but it seems to work OK and pass all
17177 the tests. Check out INSTALL.W32 for info.
17181 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17182 without temporary keys kept an extra copy of the server key,
17183 and connections with temporary keys did not free everything in case
17188 * New function RSA_check_key and new openssl rsa option -check
17189 for verifying the consistency of RSA keys.
17191 *Ulf Moeller, Bodo Moeller*
17193 * Various changes to make Win32 compile work:
17194 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17195 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17196 comparison" warnings.
17197 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17201 * Add a debugging option to PKCS#5 v2 key generation function: when
17202 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17203 derived keys are printed to stderr.
17207 * Copy the flags in ASN1_STRING_dup().
17209 *Roman E. Pavlov <pre@mo.msk.ru>*
17211 * The x509 application mishandled signing requests containing DSA
17212 keys when the signing key was also DSA and the parameters didn't match.
17214 It was supposed to omit the parameters when they matched the signing key:
17215 the verifying software was then supposed to automatically use the CA's
17216 parameters if they were absent from the end user certificate.
17218 Omitting parameters is no longer recommended. The test was also
17219 the wrong way round! This was probably due to unusual behaviour in
17220 EVP_cmp_parameters() which returns 1 if the parameters match.
17221 This meant that parameters were omitted when they *didn't* match and
17222 the certificate was useless. Certificates signed with 'ca' didn't have
17225 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17227 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17228 The interface is as follows:
17229 Applications can use
17230 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17231 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17232 "off" is now the default.
17233 The library internally uses
17234 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17235 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17236 to disable memory-checking temporarily.
17238 Some inconsistent states that previously were possible (and were
17239 even the default) are now avoided.
17241 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17242 with each memory chunk allocated; this is occasionally more helpful
17243 than just having a counter.
17245 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17247 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17252 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17253 which largely parallels "options", but is for changing API behaviour,
17254 whereas "options" are about protocol behaviour.
17255 Initial "mode" flags are:
17257 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17258 a single record has been written.
17259 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17260 retries use the same buffer location.
17261 (But all of the contents must be
17266 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17269 * Fix problems with no-hmac etc.
17271 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17273 * New functions RSA_get_default_method(), RSA_set_method() and
17274 RSA_get_method(). These allows replacement of RSA_METHODs without having
17275 to mess around with the internals of an RSA structure.
17279 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17280 Also really enable memory leak checks in openssl.c and in some
17283 *Chad C. Mulligan, Bodo Moeller*
17285 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17286 up the length of negative integers. This has now been simplified to just
17287 store the length when it is first determined and use it later, rather
17288 than trying to keep track of where data is copied and updating it to
17290 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17292 * Add a new function PKCS7_signatureVerify. This allows the verification
17293 of a PKCS#7 signature but with the signing certificate passed to the
17294 function itself. This contrasts with PKCS7_dataVerify which assumes the
17295 certificate is present in the PKCS#7 structure. This isn't always the
17296 case: certificates can be omitted from a PKCS#7 structure and be
17297 distributed by "out of band" means (such as a certificate database).
17301 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17302 function prototypes in pem.h, also change util/mkdef.pl to add the
17303 necessary function names.
17307 * mk1mf.pl (used by Windows builds) did not properly read the
17308 options set by Configure in the top level Makefile, and Configure
17309 was not even able to write more than one option correctly.
17310 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17314 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17315 file to be loaded from a BIO or FILE pointer. The BIO version will
17316 for example allow memory BIOs to contain config info.
17320 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17321 Whoever hopes to achieve shared-library compatibility across versions
17322 must use this, not the compile-time macro.
17323 (Exercise 0.9.4: Which is the minimum library version required by
17325 Note: All this applies only to multi-threaded programs, others don't
17330 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17331 through a BIO pair triggered the default case, i.e.
17332 SSLerr(...,SSL_R_UNKNOWN_STATE).
17336 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17337 can use the SSL library even if none of the specific BIOs is
17342 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17343 for the encoded length.
17345 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17347 * Add initial documentation of the X509V3 functions.
17351 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17352 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17353 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17354 secure PKCS#8 private key format with a high iteration count.
17358 * Fix determination of Perl interpreter: A perl or perl5
17359 *directory* in $PATH was also accepted as the interpreter.
17361 *Ralf S. Engelschall*
17363 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17364 wrong with it but it was very old and did things like calling
17365 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17366 unusual formatting.
17370 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17371 to use the new extension code.
17375 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17376 with macros. This should make it easier to change their form, add extra
17377 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17382 * Add to configuration table a new entry that can specify an alternative
17383 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17384 according to Mark Crispin <MRC@Panda.COM>.
17388 * DES CBC did not update the IV. Weird.
17392 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17393 Changing the behaviour of the former might break existing programs --
17394 where IV updating is needed, des_ncbc_encrypt can be used.
17397 * When bntest is run from "make test" it drives bc to check its
17398 calculations, as well as internally checking them. If an internal check
17399 fails, it needs to cause bc to give a non-zero result or make test carries
17400 on without noticing the failure. Fixed.
17404 * DES library cleanups.
17408 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17409 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17410 ciphers. NOTE: although the key derivation function has been verified
17411 against some published test vectors it has not been extensively tested
17412 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17417 * Instead of "mkdir -p", which is not fully portable, use new
17418 Perl script "util/mkdir-p.pl".
17422 * Rewrite the way password based encryption (PBE) is handled. It used to
17423 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17424 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17425 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17426 the 'parameter' field of the AlgorithmIdentifier is passed to the
17427 underlying key generation function so it must do its own ASN1 parsing.
17428 This has also changed the EVP_PBE_CipherInit() function which now has a
17429 'parameter' argument instead of literal salt and iteration count values
17430 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17434 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17435 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17436 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17437 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17438 value was just used as a "magic string" and not used directly its
17439 value doesn't matter.
17443 * Introduce some semblance of const correctness to BN. Shame C doesn't
17448 * "linux-sparc64" configuration (ultrapenguin).
17450 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17451 "linux-sparc" configuration.
17453 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17455 * config now generates no-xxx options for missing ciphers.
17459 * Support the EBCDIC character set (work in progress).
17460 File ebcdic.c not yet included because it has a different license.
17462 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17464 * Support BS2000/OSD-POSIX.
17466 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17468 * Make callbacks for key generation use `void *` instead of `char *`.
17472 * Make S/MIME samples compile (not yet tested).
17476 * Additional typesafe stacks.
17480 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17484 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17486 * New configuration variant "sco5-gcc".
17488 * Updated some demos.
17490 *Sean O Riordain, Wade Scholine*
17492 * Add missing BIO_free at exit of pkcs12 application.
17496 * Fix memory leak in conf.c.
17500 * Updates for Win32 to assembler version of MD5.
17504 * Set #! path to perl in `apps/der_chop` to where we found it
17505 instead of using a fixed path.
17509 * SHA library changes for irix64-mips4-cc.
17513 * Improvements for VMS support.
17517 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17519 * Bignum library bug fix. IRIX 6 passes "make test" now!
17520 This also avoids the problems with SC4.2 and unpatched SC5.
17522 *Andy Polyakov <appro@fy.chalmers.se>*
17524 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17525 These are required because of the typesafe stack would otherwise break
17526 existing code. If old code used a structure member which used to be STACK
17527 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17528 sk_num or sk_value it would produce an error because the num, data members
17529 are not present in STACK_OF. Now it just produces a warning. sk_set
17530 replaces the old method of assigning a value to sk_value
17531 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17532 that does this will no longer work (and should use sk_set instead) but
17533 this could be regarded as a "questionable" behaviour anyway.
17537 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17538 correctly handle encrypted S/MIME data.
17542 * Change type of various DES function arguments from des_cblock
17543 (which means, in function argument declarations, pointer to char)
17544 to des_cblock * (meaning pointer to array with 8 char elements),
17545 which allows the compiler to do more typechecking; it was like
17546 that back in SSLeay, but with lots of ugly casts.
17548 Introduce new type const_des_cblock.
17552 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17553 problems: find RecipientInfo structure that matches recipient certificate
17554 and initialise the ASN1 structures properly based on passed cipher.
17558 * Belatedly make the BN tests actually check the results.
17562 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17563 to and from BNs: it was completely broken. New compilation option
17564 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17565 key elements as negative integers.
17569 * Reorganize and speed up MD5.
17571 *Andy Polyakov <appro@fy.chalmers.se>*
17575 *Richard Levitte <richard@levitte.org>*
17577 * New option -out to asn1parse to allow the parsed structure to be
17578 output to a file. This is most useful when combined with the -strparse
17579 option to examine the output of things like OCTET STRINGS.
17583 * Make SSL library a little more fool-proof by not requiring any longer
17584 that `SSL_set_{accept,connect}_state` be called before
17585 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17586 in many applications because usually everything *appeared* to work as
17587 intended anyway -- now it really works as intended).
17591 * Move openssl.cnf out of lib/.
17595 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17596 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17597 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17599 *Ralf S. Engelschall*
17601 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17602 handle PKCS#7 enveloped data properly.
17604 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17606 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17607 copying pointers. The cert_st handling is changed by this in
17608 various ways (and thus what used to be known as ctx->default_cert
17609 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17610 any longer when s->cert does not give us what we need).
17611 ssl_cert_instantiate becomes obsolete by this change.
17612 As soon as we've got the new code right (possibly it already is?),
17613 we have solved a couple of bugs of the earlier code where s->cert
17614 was used as if it could not have been shared with other SSL structures.
17616 Note that using the SSL API in certain dirty ways now will result
17617 in different behaviour than observed with earlier library versions:
17618 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17619 does not influence s as it used to.
17621 In order to clean up things more thoroughly, inside SSL_SESSION
17622 we don't use CERT any longer, but a new structure SESS_CERT
17623 that holds per-session data (if available); currently, this is
17624 the peer's certificate chain and, for clients, the server's certificate
17625 and temporary key. CERT holds only those values that can have
17626 meaningful defaults in an SSL_CTX.
17630 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17631 from the internal representation. Various PKCS#7 fixes: remove some
17632 evil casts and set the enc_dig_alg field properly based on the signing
17637 * Allow PKCS#12 password to be set from the command line or the
17638 environment. Let 'ca' get its config file name from the environment
17639 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17644 * Allow certificate policies extension to use an IA5STRING for the
17645 organization field. This is contrary to the PKIX definition but
17646 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17651 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17652 without disallowing inline assembler and the like for non-pedantic builds.
17656 * Support Borland C++ builder.
17658 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17664 * SHA-1 cleanups and performance enhancements.
17666 *Andy Polyakov <appro@fy.chalmers.se>*
17668 * Sparc v8plus assembler for the bignum library.
17670 *Andy Polyakov <appro@fy.chalmers.se>*
17672 * Accept any -xxx and +xxx compiler options in Configure.
17676 * Update HPUX configuration.
17680 * Add missing `sk_<type>_unshift()` function to safestack.h
17682 *Ralf S. Engelschall*
17684 * New function SSL_CTX_use_certificate_chain_file that sets the
17685 "extra_cert"s in addition to the certificate. (This makes sense
17686 only for "PEM" format files, as chains as a whole are not
17691 * Support verify_depth from the SSL API.
17692 x509_vfy.c had what can be considered an off-by-one-error:
17693 Its depth (which was not part of the external interface)
17694 was actually counting the number of certificates in a chain;
17695 now it really counts the depth.
17699 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17700 instead of X509err, which often resulted in confusing error
17701 messages since the error codes are not globally unique
17702 (e.g. an alleged error in ssl3_accept when a certificate
17703 didn't match the private key).
17705 * New function SSL_CTX_set_session_id_context that allows to set a default
17706 value (so that you don't need SSL_set_session_id_context for each
17707 connection using the SSL_CTX).
17711 * OAEP decoding bug fix.
17715 * Support INSTALL_PREFIX for package builders, as proposed by
17720 * New Configure options "threads" and "no-threads". For systems
17721 where the proper compiler options are known (currently Solaris
17722 and Linux), "threads" is the default.
17726 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17730 * Install various scripts to $(OPENSSLDIR)/misc, not to
17731 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17732 such as /usr/local/bin.
17736 * "make linux-shared" to build shared libraries.
17738 *Niels Poppe <niels@netbox.org>*
17740 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17744 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17745 extension adding in x509 utility.
17749 * Remove NOPROTO sections and error code comments.
17753 * Partial rewrite of the DEF file generator to now parse the ANSI
17758 * New Configure options --prefix=DIR and --openssldir=DIR.
17762 * Complete rewrite of the error code script(s). It is all now handled
17763 by one script at the top level which handles error code gathering,
17764 header rewriting and C source file generation. It should be much better
17765 than the old method: it now uses a modified version of Ulf's parser to
17766 read the ANSI prototypes in all header files (thus the old K&R definitions
17767 aren't needed for error creation any more) and do a better job of
17768 translating function codes into names. The old 'ASN1 error code embedded
17769 in a comment' is no longer necessary and it doesn't use .err files which
17770 have now been deleted. Also the error code call doesn't have to appear all
17771 on one line (which resulted in some large lines...).
17775 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17779 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17780 0 (which usually indicates a closed connection), but continue reading.
17784 * Fix some race conditions.
17788 * Add support for CRL distribution points extension. Add Certificate
17789 Policies and CRL distribution points documentation.
17793 * Move the autogenerated header file parts to crypto/opensslconf.h.
17797 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17798 8 of keying material. Merlin has also confirmed interop with this fix
17799 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17801 *Merlin Hughes <merlin@baltimore.ie>*
17803 * Fix lots of warnings.
17805 *Richard Levitte <levitte@stacken.kth.se>*
17807 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17808 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17810 *Richard Levitte <levitte@stacken.kth.se>*
17812 * Fix problems with sizeof(long) == 8.
17814 *Andy Polyakov <appro@fy.chalmers.se>*
17816 * Change functions to ANSI C.
17820 * Fix typos in error codes.
17822 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17824 * Remove defunct assembler files from Configure.
17828 * SPARC v8 assembler BIGNUM implementation.
17830 *Andy Polyakov <appro@fy.chalmers.se>*
17832 * Support for Certificate Policies extension: both print and set.
17833 Various additions to support the r2i method this uses.
17837 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17838 return a const string when you are expecting an allocated buffer.
17842 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17843 types DirectoryString and DisplayText.
17847 * Add code to allow r2i extensions to access the configuration database,
17848 add an LHASH database driver and add several ctx helper functions.
17852 * Fix an evil bug in bn_expand2() which caused various BN functions to
17853 fail when they extended the size of a BIGNUM.
17857 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17858 support typesafe stack.
17862 * Fix typo in SSL_[gs]et_options().
17864 *Nils Frostberg <nils@medcom.se>*
17866 * Delete various functions and files that belonged to the (now obsolete)
17867 old X509V3 handling code.
17871 * New Configure option "rsaref".
17875 * Don't auto-generate pem.h.
17879 * Introduce type-safe ASN.1 SETs.
17883 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17885 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17887 * Introduce type-safe STACKs. This will almost certainly break lots of code
17888 that links with OpenSSL (well at least cause lots of warnings), but fear
17889 not: the conversion is trivial, and it eliminates loads of evil casts. A
17890 few STACKed things have been converted already. Feel free to convert more.
17891 In the fullness of time, I'll do away with the STACK type altogether.
17895 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17896 specified in `<certfile>` by updating the entry in the index.txt file.
17897 This way one no longer has to edit the index.txt file manually for
17898 revoking a certificate. The -revoke option does the gory details now.
17900 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17902 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17903 `-text` option at all and this way the `-noout -text` combination was
17904 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17906 *Ralf S. Engelschall*
17908 * Make sure a corresponding plain text error message exists for the
17909 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17910 verify callback function determined that a certificate was revoked.
17912 *Ralf S. Engelschall*
17914 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17915 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17916 all available ciphers including rc5, which was forgotten until now.
17917 In order to let the testing shell script know which algorithms
17918 are available, a new (up to now undocumented) command
17919 `openssl list-cipher-commands` is used.
17923 * Bugfix: s_client occasionally would sleep in select() when
17924 it should have checked SSL_pending() first.
17928 * New functions DSA_do_sign and DSA_do_verify to provide access to
17929 the raw DSA values prior to ASN.1 encoding.
17933 * Tweaks to Configure
17935 *Niels Poppe <niels@netbox.org>*
17937 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17942 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17946 * New config option to avoid instructions that are illegal on the 80386.
17947 The default code is faster, but requires at least a 486.
17951 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17952 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17953 same as SSL2_VERSION anyway.
17957 * New "-showcerts" option for s_client.
17961 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17962 application. Various cleanups and fixes.
17966 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17967 modify error routines to work internally. Add error codes and PBE init
17968 to library startup routines.
17972 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17973 packing functions to asn1 and evp. Changed function names and error
17974 codes along the way.
17978 * PKCS12 integration: and so it begins... First of several patches to
17979 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17980 objects to objects.h
17984 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17985 and display support for Thawte strong extranet extension.
17989 * Add LinuxPPC support.
17991 *Jeff Dubrule <igor@pobox.org>*
17993 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17994 bn_div_words in alpha.s.
17996 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17998 * Make sure the RSA OAEP test is skipped under -DRSAref because
17999 OAEP isn't supported when OpenSSL is built with RSAref.
18001 *Ulf Moeller <ulf@fitug.de>*
18003 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18004 so they no longer are missing under -DNOPROTO.
18006 *Soren S. Jorvang <soren@t.dk>*
18008 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18010 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18011 doesn't work when the session is reused. Coming soon!
18015 * Fix a security hole, that allows sessions to be reused in the wrong
18016 context thus bypassing client cert protection! All software that uses
18017 client certs and session caches in multiple contexts NEEDS PATCHING to
18018 allow session reuse! A fuller solution is in the works.
18020 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18022 * Some more source tree cleanups (removed obsolete files
18023 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18024 permission on "config" script to be executable) and a fix for the INSTALL
18027 *Ulf Moeller <ulf@fitug.de>*
18029 * Remove some legacy and erroneous uses of malloc, free instead of
18032 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18034 * Make rsa_oaep_test return non-zero on error.
18036 *Ulf Moeller <ulf@fitug.de>*
18038 * Add support for native Solaris shared libraries. Configure
18039 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18040 if someone would make that last step automatic.
18042 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18044 * ctx_size was not built with the right compiler during "make links". Fixed.
18048 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18049 except NULL ciphers". This means the default cipher list will no longer
18050 enable NULL ciphers. They need to be specifically enabled e.g. with
18051 the string "DEFAULT:eNULL".
18055 * Fix to RSA private encryption routines: if p < q then it would
18056 occasionally produce an invalid result. This will only happen with
18057 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18061 * Be less restrictive and allow also `perl util/perlpath.pl
18062 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18063 because this way one can also use an interpreter named `perl5` (which is
18064 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18065 installed as `perl`).
18067 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18069 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18071 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18073 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18074 advapi32.lib to Win32 build and change the pem test comparison
18075 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18076 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18077 and crypto/des/ede_cbcm_enc.c.
18081 * DES quad checksum was broken on big-endian architectures. Fixed.
18085 * Comment out two functions in bio.h that aren't implemented. Fix up the
18086 Win32 test batch file so it (might) work again. The Win32 test batch file
18087 is horrible: I feel ill....
18091 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18092 in e_os.h. Audit of header files to check ANSI and non ANSI
18093 sections: 10 functions were absent from non ANSI section and not exported
18094 from Windows DLLs. Fixed up libeay.num for new functions.
18098 * Make `openssl version` output lines consistent.
18100 *Ralf S. Engelschall*
18102 * Fix Win32 symbol export lists for BIO functions: Added
18103 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18104 to ms/libeay{16,32}.def.
18106 *Ralf S. Engelschall*
18108 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18109 fine under Unix and passes some trivial tests I've now added. But the
18110 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18111 added to make sure no one expects that this stuff really works in the
18112 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18113 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18116 *Ralf S. Engelschall*
18118 * Fix the generation of two part addresses in perl.
18120 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18122 * Add config entry for Linux on MIPS.
18124 *John Tobey <jtobey@channel1.com>*
18126 * Make links whenever Configure is run, unless we are on Windoze.
18130 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18131 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18136 * Add a useful kludge to allow package maintainers to specify compiler and
18137 other platforms details on the command line without having to patch the
18138 Configure script every time: One now can use
18139 `perl Configure <id>:<details>`,
18140 i.e. platform ids are allowed to have details appended
18141 to them (separated by colons). This is treated as there would be a static
18142 pre-configured entry in Configure's %table under key `<id>` with value
18143 `<details>` and `perl Configure <id>` is called. So, when you want to
18144 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18145 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18146 now, which overrides the FreeBSD-elf entry on-the-fly.
18148 *Ralf S. Engelschall*
18150 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18154 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18155 on the `perl Configure ...` command line. This way one can compile
18156 OpenSSL libraries with Position Independent Code (PIC) which is needed
18157 for linking it into DSOs.
18159 *Ralf S. Engelschall*
18161 * Remarkably, export ciphers were totally broken and no-one had noticed!
18166 * Cleaned up the LICENSE document: The official contact for any license
18167 questions now is the OpenSSL core team under openssl-core@openssl.org.
18168 And add a paragraph about the dual-license situation to make sure people
18169 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18170 to the OpenSSL toolkit.
18172 *Ralf S. Engelschall*
18174 * General source tree makefile cleanups: Made `making xxx in yyy...`
18175 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18176 Additionally cleaned up the `make links` target: Remove unnecessary
18177 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18178 to speed processing and no longer clutter the display with confusing
18179 stuff. Instead only the actually done links are displayed.
18181 *Ralf S. Engelschall*
18183 * Permit null encryption ciphersuites, used for authentication only. It used
18184 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18185 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18190 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18191 signed attributes when verifying signatures (this would break them),
18192 the detached data encoding was wrong and public keys obtained using
18193 X509_get_pubkey() weren't freed.
18197 * Add text documentation for the BUFFER functions. Also added a work around
18198 to a Win95 console bug. This was triggered by the password read stuff: the
18199 last character typed gets carried over to the next fread(). If you were
18200 generating a new cert request using 'req' for example then the last
18201 character of the passphrase would be CR which would then enter the first
18206 * Added the new 'Includes OpenSSL Cryptography Software' button as
18207 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18208 button and can be used by applications based on OpenSSL to show the
18209 relationship to the OpenSSL project.
18211 *Ralf S. Engelschall*
18213 * Remove confusing variables in function signatures in files
18214 ssl/ssl_lib.c and ssl/ssl.h.
18216 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18218 * Don't install bss_file.c under PREFIX/include/
18220 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18222 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18223 functions that return function pointers and has support for NT specific
18224 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18225 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18226 unsigned to signed types: this was killing the Win32 compile.
18230 * Add new certificate file to stack functions,
18231 SSL_add_dir_cert_subjects_to_stack() and
18232 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18233 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18234 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18235 This means that Apache-SSL and similar packages don't have to mess around
18236 to add as many CAs as they want to the preferred list.
18240 * Experiment with doxygen documentation. Currently only partially applied to
18242 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18243 openssl.doxy as the configuration file.
18247 * Get rid of remaining C++-style comments which strict C compilers hate.
18249 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18251 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18252 compiled in by default: it has problems with large keys.
18256 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18257 DH private keys and/or callback functions which directly correspond to
18258 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18259 is needed for applications which have to configure certificates on a
18260 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18262 For the RSA certificate situation is makes no difference, but
18263 for the DSA certificate situation this fixes the "no shared cipher"
18264 problem where the OpenSSL cipher selection procedure failed because the
18265 temporary keys were not overtaken from the context and the API provided
18266 no way to reconfigure them.
18267 The new functions now let applications reconfigure the stuff and they
18268 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18269 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18270 non-public-API function ssl_cert_instantiate() is used as a helper
18271 function and also to reduce code redundancy inside ssl_rsa.c.
18273 *Ralf S. Engelschall*
18275 * Move s_server -dcert and -dkey options out of the undocumented feature
18276 area because they are useful for the DSA situation and should be
18277 recognized by the users.
18279 *Ralf S. Engelschall*
18281 * Fix the cipher decision scheme for export ciphers: the export bits are
18282 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18283 SSL_EXP_MASK. So, the original variable has to be used instead of the
18284 already masked variable.
18286 *Richard Levitte <levitte@stacken.kth.se>*
18288 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18290 *Richard Levitte <levitte@stacken.kth.se>*
18292 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18293 from `int` to `unsigned int` because it is a length and initialized by
18294 EVP_DigestFinal() which expects an `unsigned int *`.
18296 *Richard Levitte <levitte@stacken.kth.se>*
18298 * Don't hard-code path to Perl interpreter on shebang line of Configure
18299 script. Instead use the usual Shell->Perl transition trick.
18301 *Ralf S. Engelschall*
18303 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18304 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18305 -noout -modulus` as it's already the case for `openssl rsa -noout
18306 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18307 currently the public key is printed (a decision which was already done by
18308 `openssl dsa -modulus` in the past) which serves a similar purpose.
18309 Additionally the NO_RSA no longer completely removes the whole -modulus
18310 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18313 *Ralf S. Engelschall*
18315 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18316 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18318 *Arne Ansper <arne@ats.cyber.ee>*
18320 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18321 to be added. Now both 'req' and 'ca' can use new objects defined in the
18326 * Add cool BIO that does syslog (or event log on NT).
18328 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18330 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18331 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18332 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18333 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18337 * Add preliminary config info for new extension code.
18341 * Make RSA_NO_PADDING really use no padding.
18343 *Ulf Moeller <ulf@fitug.de>*
18345 * Generate errors when private/public key check is done.
18349 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18350 for some CRL extensions and new objects added.
18354 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18355 key usage extension and fuller support for authority key id.
18359 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18360 padding method for RSA, which is recommended for new applications in PKCS
18361 #1 v2.0 (RFC 2437, October 1998).
18362 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18363 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18364 against Bleichbacher's attack on RSA.
18365 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18368 * Updates to the new SSL compression code
18370 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18372 * Fix so that the version number in the master secret, when passed
18373 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18374 (because the server will not accept higher), that the version number
18375 is 0x03,0x01, not 0x03,0x00
18377 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18379 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18380 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18381 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18385 * Support for RAW extensions where an arbitrary extension can be
18386 created by including its DER encoding. See `apps/openssl.cnf` for
18391 * Make sure latest Perl versions don't interpret some generated C array
18392 code as Perl array code in the crypto/err/err_genc.pl script.
18394 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18396 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18397 not many people have the assembler. Various Win32 compilation fixes and
18398 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18399 build instructions.
18403 * Modify configure script 'Configure' to automatically create crypto/date.h
18404 file under Win32 and also build pem.h from pem.org. New script
18405 util/mkfiles.pl to create the MINFO file on environments that can't do a
18406 'make files': perl util/mkfiles.pl >MINFO should work.
18410 * Major rework of DES function declarations, in the pursuit of correctness
18411 and purity. As a result, many evil casts evaporated, and some weirdness,
18412 too. You may find this causes warnings in your code. Zapping your evil
18413 casts will probably fix them. Mostly.
18417 * Fix for a typo in asn1.h. Bug fix to object creation script
18418 obj_dat.pl. It considered a zero in an object definition to mean
18419 "end of object": none of the objects in objects.h have any zeros
18420 so it wasn't spotted.
18422 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18424 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18425 Masking (CBCM). In the absence of test vectors, the best I have been able
18426 to do is check that the decrypt undoes the encrypt, so far. Send me test
18427 vectors if you have them.
18431 * Correct calculation of key length for export ciphers (too much space was
18432 allocated for null ciphers). This has not been tested!
18436 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18437 message is now correct (it understands "crypto" and "ssl" on its
18438 command line). There is also now an "update" option. This will update
18439 the util/ssleay.num and util/libeay.num files with any new functions.
18441 perl util/mkdef.pl crypto ssl update
18442 it will update them.
18446 * Overhauled the Perl interface:
18447 - ported BN stuff to OpenSSL's different BN library
18448 - made the perl/ source tree CVS-aware
18449 - renamed the package from SSLeay to OpenSSL (the files still contain
18450 their history because I've copied them in the repository)
18451 - removed obsolete files (the test scripts will be replaced
18452 by better Test::Harness variants in the future)
18454 *Ralf S. Engelschall*
18456 * First cut for a very conservative source tree cleanup:
18457 1. merge various obsolete readme texts into doc/ssleay.txt
18458 where we collect the old documents and readme texts.
18459 2. remove the first part of files where I'm already sure that we no
18460 longer need them because of three reasons: either they are just temporary
18461 files which were left by Eric or they are preserved original files where
18462 I've verified that the diff is also available in the CVS via "cvs diff
18463 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18464 the crypto/md/ stuff).
18466 *Ralf S. Engelschall*
18468 * More extension code. Incomplete support for subject and issuer alt
18469 name, issuer and authority key id. Change the i2v function parameters
18470 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18471 what that's for :-) Fix to ASN1 macro which messed up
18472 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18476 * Preliminary support for ENUMERATED type. This is largely copied from the
18481 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18483 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18485 * Make sure `make rehash` target really finds the `openssl` program.
18487 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18489 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18490 like to hear about it if this slows down other processors.
18494 * Add CygWin32 platform information to Configure script.
18496 *Alan Batie <batie@aahz.jf.intel.com>*
18498 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18500 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18502 * New program nseq to manipulate netscape certificate sequences
18506 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18511 * Fixes to BN code. Previously the default was to define BN_RECURSION
18512 but the BN code had some problems that would cause failures when
18513 doing certificate verification and some other functions.
18515 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18517 * Add ASN1 and PEM code to support netscape certificate sequences.
18521 * Add ASN1 and PEM code to support netscape certificate sequences.
18525 * Add several PKIX and private extended key usage OIDs.
18529 * Modify the 'ca' program to handle the new extension code. Modify
18530 openssl.cnf for new extension format, add comments.
18534 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18535 and add a sample to openssl.cnf so req -x509 now adds appropriate
18540 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18541 error code, add initial support to X509_print() and x509 application.
18545 * Takes a deep breath and start adding X509 V3 extension support code. Add
18546 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18547 stuff is currently isolated and isn't even compiled yet.
18551 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18552 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18553 Removed the versions check from X509 routines when loading extensions:
18554 this allows certain broken certificates that don't set the version
18555 properly to be processed.
18559 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18560 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18561 can still be regenerated with "make depend".
18565 * Spelling mistake in C version of CAST-128.
18567 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18569 * Changes to the error generation code. The perl script err-code.pl
18570 now reads in the old error codes and retains the old numbers, only
18571 adding new ones if necessary. It also only changes the .err files if new
18572 codes are added. The makefiles have been modified to only insert errors
18573 when needed (to avoid needlessly modifying header files). This is done
18574 by only inserting errors if the .err file is newer than the auto generated
18575 C file. To rebuild all the error codes from scratch (the old behaviour)
18576 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18577 or delete all the .err files.
18581 * CAST-128 was incorrectly implemented for short keys. The C version has
18582 been fixed, but is untested. The assembler versions are also fixed, but
18583 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18584 to regenerate it if needed.
18585 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18586 Hagino <itojun@kame.net>*
18588 * File was opened incorrectly in randfile.c.
18590 *Ulf Möller <ulf@fitug.de>*
18592 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18593 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18594 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18595 al: it's just almost always a UTCTime. Note this patch adds new error
18596 codes so do a "make errors" if there are problems.
18600 * Correct Linux 1 recognition in config.
18602 *Ulf Möller <ulf@fitug.de>*
18604 * Remove pointless MD5 hash when using DSA keys in ca.
18606 *Anonymous <nobody@replay.com>*
18608 * Generate an error if given an empty string as a cert directory. Also
18609 generate an error if handed NULL (previously returned 0 to indicate an
18610 error, but didn't set one).
18612 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18614 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18618 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18619 parameters. This was causing a warning which killed off the Win32 compile.
18623 * Remove C++ style comments from crypto/bn/bn_local.h.
18625 *Neil Costigan <neil.costigan@celocom.com>*
18627 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18628 based on a text string, looking up short and long names and finally
18629 "dot" format. The "dot" format stuff didn't work. Added new function
18630 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18631 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18632 OID is not part of the table.
18636 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18637 X509_LOOKUP_by_alias().
18641 * Sort openssl functions by name.
18645 * Get the `gendsa` command working and add it to the `list` command. Remove
18646 encryption from sample DSA keys (in case anyone is interested the password
18651 * Make *all* `*_free` functions accept a NULL pointer.
18653 *Frans Heymans <fheymans@isaserver.be>*
18655 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18658 *Anonymous <nobody@replay.com>*
18660 * s_server should send the CAfile as acceptable CAs, not its own cert.
18662 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18664 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18666 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18668 * Temp key "for export" tests were wrong in s3_srvr.c.
18670 *Anonymous <nobody@replay.com>*
18672 * Add prototype for temp key callback functions
18673 SSL_CTX_set_tmp_{rsa,dh}_callback().
18677 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18678 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18682 * X509_name_add_entry() freed the wrong thing after an error.
18684 *Arne Ansper <arne@ats.cyber.ee>*
18686 * rsa_eay.c would attempt to free a NULL context.
18688 *Arne Ansper <arne@ats.cyber.ee>*
18690 * BIO_s_socket() had a broken should_retry() on Windoze.
18692 *Arne Ansper <arne@ats.cyber.ee>*
18694 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18696 *Arne Ansper <arne@ats.cyber.ee>*
18698 * Make sure the already existing X509_STORE->depth variable is initialized
18699 in X509_STORE_new(), but document the fact that this variable is still
18700 unused in the certificate verification process.
18702 *Ralf S. Engelschall*
18704 * Fix the various library and `apps/` files to free up pkeys obtained from
18705 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18709 * Fix reference counting in X509_PUBKEY_get(). This makes
18710 demos/maurice/example2.c work, amongst others, probably.
18712 *Steve Henson and Ben Laurie*
18714 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18715 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18716 are no longer created. This way we have a single and consistent command
18717 line interface `openssl <command>`, similar to `cvs <command>`.
18719 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18721 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18722 BIT STRING wrapper always have zero unused bits.
18726 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18730 * Make the top-level INSTALL documentation easier to understand.
18734 * Makefiles updated to exit if an error occurs in a sub-directory
18735 make (including if user presses ^C) [Paul Sutton]
18737 * Make Montgomery context stuff explicit in RSA data structure.
18741 * Fix build order of pem and err to allow for generated pem.h.
18745 * Fix renumbering bug in X509_NAME_delete_entry().
18749 * Enhanced the err-ins.pl script so it makes the error library number
18750 global and can add a library name. This is needed for external ASN1 and
18751 other error libraries.
18755 * Fixed sk_insert which never worked properly.
18759 * Fix ASN1 macros so they can handle indefinite length constructed
18760 EXPLICIT tags. Some non standard certificates use these: they can now
18765 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18766 into a single doc/ssleay.txt bundle. This way the information is still
18767 preserved but no longer messes up this directory. Now it's new room for
18768 the new set of documentation files.
18770 *Ralf S. Engelschall*
18772 * SETs were incorrectly DER encoded. This was a major pain, because they
18773 shared code with SEQUENCEs, which aren't coded the same. This means that
18774 almost everything to do with SETs or SEQUENCEs has either changed name or
18775 number of arguments.
18777 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18779 * Fix test data to work with the above.
18783 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18784 was already fixed by Eric for 0.9.1 it seems.
18786 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18788 * Autodetect FreeBSD3.
18792 * Fix various bugs in Configure. This affects the following platforms:
18796 unixware-2.0-pentium
18801 * Eliminate generated files from CVS. Reorder tests to regenerate files
18802 before they are needed.
18806 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18810 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18812 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18813 changed SSLeay to OpenSSL in version strings.
18815 *Ralf S. Engelschall*
18817 * Some fixups to the top-level documents.
18821 * Fixed the nasty bug where rsaref.h was not found under compile-time
18822 because the symlink to include/ was missing.
18824 *Ralf S. Engelschall*
18826 * Incorporated the popular no-RSA/DSA-only patches
18827 which allow to compile a RSA-free SSLeay.
18829 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18831 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18832 when "ssleay" is still not found.
18834 *Ralf S. Engelschall*
18836 * Added more platforms to Configure: Cray T3E, HPUX 11,
18838 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18840 * Updated the README file.
18842 *Ralf S. Engelschall*
18844 * Added various .cvsignore files in the CVS repository subdirs
18845 to make a "cvs update" really silent.
18847 *Ralf S. Engelschall*
18849 * Recompiled the error-definition header files and added
18850 missing symbols to the Win32 linker tables.
18852 *Ralf S. Engelschall*
18854 * Cleaned up the top-level documents;
18855 o new files: CHANGES and LICENSE
18856 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18857 o merged COPYRIGHT into LICENSE
18858 o removed obsolete TODO file
18859 o renamed MICROSOFT to INSTALL.W32
18861 *Ralf S. Engelschall*
18863 * Removed dummy files from the 0.9.1b source tree:
18865 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18866 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18867 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18868 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18869 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18872 *Ralf S. Engelschall*
18874 * Added various platform portability fixes.
18878 * The Genesis of the OpenSSL rpject:
18879 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18880 Young and Tim J. Hudson created while they were working for C2Net until
18883 *The OpenSSL Project*
18885 ### Changes between 0.9.0b and 0.9.1b [not released]
18887 * Updated a few CA certificates under certs/
18891 * Changed some BIGNUM api stuff.
18895 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18896 DGUX x86, Linux Alpha, etc.
18900 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18901 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18906 * Add -strparse option to asn1pars program which parses nested
18909 *Dr Stephen Henson <shenson@bigfoot.com>*
18911 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18915 * DSA fix for "ca" program.
18919 * Added "-genkey" option to "dsaparam" program.
18923 * Added RIPE MD160 (rmd160) message digest.
18927 * Added -a (all) option to "ssleay version" command.
18931 * Added PLATFORM define which is the id given to Configure.
18935 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18939 * Extended the ASN.1 parser routines.
18943 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18947 * Added a BN_CTX to the BN library.
18951 * Fixed the weak key values in DES library
18955 * Changed API in EVP library for cipher aliases.
18959 * Added support for RC2/64bit cipher.
18963 * Converted the lhash library to the crypto/mem.c functions.
18967 * Added more recognized ASN.1 object ids.
18971 * Added more RSA padding checks for SSL/TLS.
18975 * Added BIO proxy/filter functionality.
18979 * Added extra_certs to SSL_CTX which can be used
18980 send extra CA certificates to the client in the CA cert chain sending
18981 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18985 * Now Fortezza is denied in the authentication phase because
18986 this is key exchange mechanism is not supported by SSLeay at all.
18990 * Additional PKCS1 checks.
18994 * Support the string "TLSv1" for all TLS v1 ciphers.
18998 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18999 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19003 * Fixed a few memory leaks.
19007 * Fixed various code and comment typos.
19011 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19012 bytes sent in the client random.
19014 *Edward Bishop <ebishop@spyglass.com>*
19018 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19019 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19020 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19021 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19022 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19023 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19024 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19025 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19026 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19027 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19028 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19029 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19030 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19031 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19032 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19033 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19034 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19035 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19036 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19037 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19038 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19039 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19040 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19041 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19042 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19043 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19044 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19045 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19046 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19047 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19048 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19049 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19050 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19051 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19052 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19053 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19054 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19055 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19056 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19057 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19058 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19059 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19060 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19061 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19062 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19063 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19064 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19065 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19066 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19067 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19068 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19069 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19070 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19071 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19072 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19073 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19074 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19075 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19076 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19077 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19078 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19079 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19080 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19081 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19082 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19083 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19084 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19085 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19086 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19087 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19088 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19089 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19090 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19091 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19092 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19093 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19094 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19095 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19096 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19097 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19098 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19099 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19100 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19101 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19102 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19103 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19104 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19105 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19106 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19107 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19108 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19109 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19110 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19111 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19112 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19113 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19114 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19115 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19116 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19117 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19118 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19119 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19120 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19121 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19122 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19123 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19124 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19125 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19126 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19127 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19128 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19129 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19130 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19131 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19132 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19133 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19134 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19135 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19136 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19137 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19138 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19139 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19140 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19141 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19142 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19143 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19144 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19145 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19146 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19147 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19148 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19149 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19150 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19151 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19152 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19153 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19154 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19155 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19156 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19157 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19158 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19159 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19160 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19161 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19162 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19163 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19164 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19165 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19166 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19167 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19168 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19169 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19170 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19171 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19172 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19173 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19174 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19175 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19176 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19177 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19178 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19179 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655