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 * The implementation of the EVP ciphers CAST5-ECB, CAST5-CBC, CAST5-OFB,
27 CAST5-CFB, BF-ECB, BF-CBC, BF-OFB, BF-CFB, IDEA-ECB, IDEC-CBC, IDEA-OFB,
28 IDEA-CFB, SEED-ECB, SEED-CBC, SEED-OFB, SEED-CFB, RC2-ECB, RC2-CBC,
29 RC2-40-CBC, RC2-64-CBC, RC2-OFB, RC2-CFB, RC4, RC4-40, RC4-HMAC-MD5, RC5-ECB,
30 RC5-CBC, RC5-OFB, RC5-CFB, DESX-CBC, DES-ECB, DES-CBC, DES-OFB, DES-CFB,
31 DES-CFB1 and DES-CFB8 have been moved to the legacy provider. Applications
32 using the EVP APIs to access these ciphers should instead use more modern
33 ciphers. If that is not possible then these applications should ensure that
34 the legacy provider has been loaded. This can be achieved either
35 programmatically or via configuration. See the provider(7) man page for
40 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
41 RIPEMD-160 have been moved to the legacy provider. Applications using the
42 EVP APIs to access these digests should instead use more modern digests. If
43 that is not possible then these applications should ensure that the legacy
44 provider has been loaded. This can be achieved either programmatically or via
45 configuration. See the provider(7) man page for further details.
49 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
54 * The deprecated functions EVP_PKEY_get0_RSA(),
55 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
56 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
57 well as the similarly named "get1" functions behave slightly differently in
58 OpenSSL 3.0. Previously they returned a pointer to the low-level key used
59 internally by libcrypto. From OpenSSL 3.0 this key may now be held in a
60 provider. Calling these functions will only return a handle on the internal
61 key where the EVP_PKEY was constructed using this key in the first place, for
62 example using a function or macro such as EVP_PKEY_assign_RSA(),
63 EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key,
64 then these functions now return a cached copy of the key. Changes to
65 the internal provider key that take place after the first time the cached key
66 is accessed will not be reflected back in the cached copy. Similarly any
67 changes made to the cached copy by application code will not be reflected
68 back in the internal provider key.
70 For the above reasons the keys returned from these functions should typically
71 be treated as read-only. To emphasise this the value returned from
72 EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(),
73 EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may
74 break some existing code. Applications broken by this change should be
75 modified. The preferred solution is to refactor the code to avoid the use of
76 these deprecated functions. Failing this the code should be modified to use a
77 const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(),
78 EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a
79 non-const pointer to enable them to be "freed". However they should also be
84 * A number of functions handling low level keys or engines were deprecated
85 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
86 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
87 EVP_PKEY_get0_siphash(). Applications using engines should instead use
88 providers. Applications getting or setting low level keys in an EVP_PKEY
89 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
90 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
94 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
95 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
96 and require returning octet ptr parameters from providers that
97 would like to support them which complicates provider implementations.
101 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
102 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
103 RAND_set_rand_method(). Provider based random number generators should
104 be used instead via EVP_RAND(3).
108 * The SRP APIs have been deprecated. The old APIs do not work via providers,
109 and there is no EVP interface to them. Unfortunately there is no replacement
110 for these APIs at this time.
114 * Add a compile time option to prevent the caching of provider fetched
115 algorithms. This is enabled by including the no-cached-fetch option
116 at configuration time.
120 * The openssl speed command does not use low-level API calls anymore. This
121 implies some of the performance numbers might not be fully comparable
122 with the previous releases due to higher overhead. This applies
123 particularly to measuring performance on smaller data chunks.
127 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
128 Typically if OpenSSL has no EC or DH algorithms then it cannot support
129 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
130 through providers. Therefore third party providers may supply group
131 implementations even where there are no built-in ones. Attempting to create
132 TLS connections in such a build without also disabling TLSv1.3 at run time or
133 using third party provider groups may result in handshake failures. TLSv1.3
134 can be disabled at compile time using the "no-tls1_3" Configure option.
138 * The undocumented function X509_certificate_type() has been deprecated;
139 applications can use X509_get0_pubkey() and X509_get0_signature() to
140 get the same information.
144 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
145 functions. They are identical to BN_rand() and BN_rand_range()
150 * Removed RSA padding mode for SSLv23 (which was only used for
151 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
152 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
157 * Deprecated the obsolete X9.31 RSA key generation related functions
158 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
159 BN_X931_generate_prime_ex().
163 * The default key generation method for the regular 2-prime RSA keys was
164 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
165 Conditions Based on Auxiliary Probable Primes). This method is slower
166 than the original method.
170 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
171 They are replaced with the BN_check_prime() function that avoids possible
172 misuse and always uses at least 64 rounds of the Miller-Rabin
173 primality test. At least 64 rounds of the Miller-Rabin test are now also
174 used for all prime generation, including RSA key generation.
175 This increases key generation time, especially for larger keys.
179 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
180 as they are not useful with non-deprecated functions.
184 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
185 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
186 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
187 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
188 were used to collect all necessary data to form a HTTP request, and to
189 perform the HTTP transfer with that request. With OpenSSL 3.0, the
190 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
191 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
192 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
193 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
194 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
195 OSSL_HTTP_REQ_CTX_set_max_response_length().
197 *Rich Salz and Richard Levitte*
199 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
203 * Validation of SM2 keys has been separated from the validation of regular EC
204 keys, allowing to improve the SM2 validation process to reject loaded private
205 keys that are not conforming to the SM2 ISO standard.
206 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
211 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
212 switches: a validation failure triggers an early exit, returning a failure
213 exit status to the parent process.
217 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
218 to ignore unknown ciphers.
222 * The `-cipher-commands` and `-digest-commands` options
223 of the command line utility `list` have been deprecated.
224 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
228 * All of the low level EC_KEY functions have been deprecated including:
230 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
231 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
232 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
233 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
234 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
235 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
236 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
237 EC_KEY_METHOD_get_verify,
238 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
239 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
240 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
241 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
242 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
243 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
244 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
245 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
246 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
247 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
248 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
249 Applications that need to implement an EC_KEY_METHOD need to consider
250 implementation of the functionality in a special provider.
251 For replacement of the functions manipulating the EC_KEY objects
252 see the EVP_PKEY-EC(7) manual page.
254 Additionally functions that read and write EC_KEY objects such as
255 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
256 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
257 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
258 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
259 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
260 have also been deprecated. Applications should instead use the
261 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
263 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
264 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
265 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
266 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
267 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
269 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
271 * Deprecated all the libcrypto and libssl error string loading
272 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
273 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
274 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
275 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
276 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
277 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
278 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
279 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
280 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
281 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
283 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
284 now loads error strings automatically.
288 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
289 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
290 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
291 are to be used by servers requiring ephemeral DH keys. Instead applications
292 should consider using the built-in DH parameters that are available by
293 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
294 necessary then applications can use the alternative functions
295 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
296 replacement for the "callback" functions. The callback was originally useful
297 in order to have different parameters for export and non-export ciphersuites.
298 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
299 functions should be replaced by one of the other methods described above.
303 * The `-crypt` option to the `passwd` command line tool has been removed.
307 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
312 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
314 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
315 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
316 The inverse ciphers use AES decryption for wrapping, and
317 AES encryption for unwrapping.
321 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
322 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
323 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
324 OpenSSL 3.0 these are replaced by the more generic functions
325 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
326 The old versions have been converted to deprecated macros that just call the
331 * The security callback, which can be customised by application code, supports
332 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
333 in the "other" parameter. In most places this is what is passed. All these
334 places occur server side. However there was one client side call of this
335 security operation and it passed a DH object instead. This is incorrect
336 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
337 of the other locations. Therefore this client side call has been changed to
338 pass an EVP_PKEY instead.
342 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
343 interface. Their functionality remains unchanged.
347 * Added new option for 'openssl list', '-providers', which will display the
348 list of loaded providers, their names, version and status. It optionally
349 displays their gettable parameters.
353 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
354 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
355 type is internally recognised so the workaround is no longer needed.
357 Functionality is still retained as it is, but will only work with
358 EVP_PKEYs with a legacy internal key.
362 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
363 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
367 * Changed all "STACK" functions to be macros instead of inline functions. Macro
368 parameters are still checked for type safety at compile time via helper
373 * Remove the RAND_DRBG API
375 The RAND_DRBG API did not fit well into the new provider concept as
376 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
377 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
378 and some of its API calls are rather low-level. This holds in particular
379 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
381 Adding a compatibility layer to continue supporting the RAND_DRBG API as
382 a legacy API for a regular deprecation period turned out to come at the
383 price of complicating the new provider API unnecessarily. Since the
384 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
387 *Paul Dale and Matthias St. Pierre*
389 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
390 as well as actual hostnames.
394 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
395 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
396 conversely, silently ignore DTLS protocol version bounds when configuring
397 TLS-based contexts. The commands can be repeated to set bounds of both
398 types. The same applies with the corresponding "min_protocol" and
399 "max_protocol" command-line switches, in case some application uses both TLS
402 SSL_CTX instances that are created for a fixed protocol version (e.g.
403 `TLSv1_server_method()`) also silently ignore version bounds. Previously
404 attempts to apply bounds to these protocol versions would result in an
405 error. Now only the "version-flexible" SSL_CTX instances are subject to
406 limits in configuration files in command-line options.
410 * Deprecated the `ENGINE` API. Engines should be replaced with providers
415 * Reworked the recorded ERR codes to make better space for system errors.
416 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
417 given code is a system error (true) or an OpenSSL error (false).
421 * Reworked the test perl framework to better allow parallel testing.
423 *Nicola Tuveri and David von Oheimb*
425 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
426 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
430 * 'Configure' has been changed to figure out the configuration target if
431 none is given on the command line. Consequently, the 'config' script is
432 now only a mere wrapper. All documentation is changed to only mention
435 *Rich Salz and Richard Levitte*
437 * Added a library context `OSSL_LIB_CTX` that applications as well as
438 other libraries can use to form a separate context within which
439 libcrypto operations are performed.
441 There are two ways this can be used:
443 - Directly, by passing a library context to functions that take
444 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
446 - Indirectly, by creating a new library context and then assigning
447 it as the new default, with `OSSL_LIB_CTX_set0_default`.
449 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
450 apart from the functions directly related to `OSSL_LIB_CTX`, accept
451 NULL to indicate that the default library context should be used.
453 Library code that changes the default library context using
454 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
455 second call before returning to the caller.
457 _(Note: the library context was initially called `OPENSSL_CTX` and
458 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
462 * Handshake now fails if Extended Master Secret extension is dropped
467 * Dropped interactive mode from the `openssl` program. From now on,
468 running it without arguments is equivalent to `openssl help`.
472 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
473 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
474 While the old function names have been retained for backward compatibility
475 they should not be used in new developments
476 because their return values are confusing: Unlike other `_cmp()` functions
477 they do not return 0 in case their arguments are equal.
481 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
482 `EC_GROUP_get_field_type()`.
486 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
487 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
488 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
489 Applications should rely on the library automatically assigning a suitable
490 EC_METHOD internally upon EC_GROUP construction.
494 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
495 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
496 assigned internally without application intervention.
497 Users of EC_GROUP_new() should switch to a different suitable constructor.
501 * Add CAdES-BES signature verification support, mostly derived
502 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
504 *Filipe Raimundo da Silva*
506 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
510 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
511 functions are not widely used and now OpenSSL automatically perform this
512 conversion when needed.
516 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
517 EC_KEY_precompute_mult(). These functions are not widely used and
518 applications should instead switch to named curves which OpenSSL has
519 hardcoded lookup tables for.
523 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
524 should instead use the L<EC_POINT_mul(3)> function.
528 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
529 that are not applicable to the new provider model. Applications should
530 instead use EVP_default_properties_is_fips_enabled() and
531 EVP_default_properties_enable_fips().
535 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
536 is set, an unexpected EOF is ignored, it pretends a close notify was received
537 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
541 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
542 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
543 used and applications should instead use the
544 L<EC_POINT_set_affine_coordinates(3)> and
545 L<EC_POINT_get_affine_coordinates(3)> functions.
549 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
550 arrays to be more easily constructed via a series of utility functions.
551 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
552 the various push functions and finally convert to a passable OSSL_PARAM
553 array using OSSL_PARAM_BLD_to_param().
557 * The security strength of SHA1 and MD5 based signatures in TLS has been
558 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
559 working at the default security level of 1 and instead requires security
560 level 0. The security level can be changed either using the cipher string
561 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
565 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
566 contain a provider side internal key.
570 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
571 They are old functions that we don't use, and that you could disable with
572 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
576 * Project text documents not yet having a proper file name extension
577 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
578 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
582 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
583 have been converted to Markdown with the goal to produce documents
584 which not only look pretty when viewed online in the browser, but
585 remain well readable inside a plain text editor.
587 To achieve this goal, a 'minimalistic' Markdown style has been applied
588 which avoids formatting elements that interfere too much with the
589 reading flow in the text file. For example, it
591 * avoids [ATX headings][] and uses [setext headings][] instead
592 (which works for `<h1>` and `<h2>` headings only).
593 * avoids [inline links][] and uses [reference links][] instead.
594 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
596 [ATX headings]: https://github.github.com/gfm/#atx-headings
597 [setext headings]: https://github.github.com/gfm/#setext-headings
598 [inline links]: https://github.github.com/gfm/#inline-link
599 [reference links]: https://github.github.com/gfm/#reference-link
600 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
601 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
603 *Matthias St. Pierre*
605 * The test suite is changed to preserve results of each test recipe.
606 A new directory test-runs/ with subdirectories named like the
607 test recipes are created in the build tree for this purpose.
611 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
612 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
613 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
615 *David von Oheimb, Martin Peylo*
617 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
618 The legacy OCSP-focused and only partly documented API is retained for
619 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
623 * Added `util/check-format.pl`, a tool for checking adherence to the
624 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
625 The checks performed are incomplete and yield some false positives.
626 Still the tool should be useful for detecting most typical glitches.
630 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
631 If domain name resolution yields multiple IP addresses all of them are tried
632 after `connect()` failures.
636 * All of the low level RSA functions have been deprecated including:
638 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
639 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
640 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
641 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
642 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
643 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
644 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
645 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
646 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
647 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
648 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
649 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
650 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
651 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
652 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
653 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
654 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
655 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
656 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
657 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
658 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
659 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
660 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
661 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
662 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
663 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
664 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
665 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
667 Use of these low level functions has been informally discouraged for a long
668 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
669 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
670 L<EVP_PKEY_decrypt(3)>.
674 * X509 certificates signed using SHA1 are no longer allowed at security
676 In TLS/SSL the default security level is 1. It can be set either
677 using the cipher string with `@SECLEVEL`, or calling
678 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
679 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
681 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
682 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
683 options of the commands.
687 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
688 modified to use PKEY APIs. These commands are now in maintenance mode
689 and no new features will be added to them.
693 * The command line utility rsautl has been deprecated.
694 Instead use the pkeyutl program.
698 * The command line utilities genrsa and rsa have been modified to use PKEY
699 APIs. They now write PKCS#8 keys by default. These commands are now in
700 maintenance mode and no new features will be added to them.
704 * All of the low level DH functions have been deprecated including:
706 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
707 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
708 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
709 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
710 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
711 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
712 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
713 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
714 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
715 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
716 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
717 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
718 DH_meth_get_generate_params and DH_meth_set_generate_params.
720 Use of these low level functions has been informally discouraged for a long
721 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
722 and L<EVP_PKEY_derive(3)>.
724 Additionally functions that read and write DH objects such as d2i_DHparams,
725 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
726 functions have also been deprecated. Applications should instead use the
727 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
729 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
730 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
731 `EVP_PKEY_set1_DH()` are also deprecated.
732 Applications should instead either read or write an
733 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
734 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
736 *Paul Dale and Matt Caswell*
738 * All of the low level DSA functions have been deprecated including:
740 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
741 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
742 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
743 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
744 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
745 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
746 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
747 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
748 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
749 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
750 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
751 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
752 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
753 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
754 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
755 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
756 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
758 Use of these low level functions has been informally discouraged for a long
759 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
760 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
762 Finaly functions that assign or obtain DSA objects from an EVP_PKEY such as
763 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
764 `EVP_PKEY_set1_DSA()` are also deprecated.
765 Applications should instead either read or write an
766 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs,
767 or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
771 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
772 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
773 This means that applications don't have to look at the curve NID and
774 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
775 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
776 a no-op when the EVP_PKEY is already of the given type.
778 Parameter and key generation is also reworked to make it possible
779 to generate EVP_PKEY_SM2 parameters and keys without having to go
780 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
781 However, code that does the latter will still work as before.
785 * Deprecated low level ECDH and ECDSA functions. These include:
787 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
788 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
791 Use of these low level functions has been informally discouraged for a long
792 time. Instead applications should use the EVP_PKEY_derive(3),
793 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
797 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
798 and EVP_PKEY_decrypt() instead.
799 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
800 and EVP_PKEY_encrypt() instead.
804 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
805 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
806 a new formulation to include all the things it can be used for,
807 as well as words of caution.
811 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
812 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
816 * All of the low level HMAC functions have been deprecated including:
818 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
819 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
822 Use of these low level functions has been informally discouraged for a long
823 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
824 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
825 and L<EVP_MAC_final(3)>.
829 * Over two thousand fixes were made to the documentation, including:
830 - Common options (such as -rand/-writerand, TLS version control, etc)
831 were refactored and point to newly-enhanced descriptions in openssl.pod.
832 - Added style conformance for all options (with help from Richard Levitte),
833 documented all reported missing options, added a CI build to check
834 that all options are documented and that no unimplemented options
836 - Documented some internals, such as all use of environment variables.
837 - Addressed all internal broken L<> references.
841 * All of the low level CMAC functions have been deprecated including:
843 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
844 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
846 Use of these low level functions has been informally discouraged for a long
847 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
848 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
849 and L<EVP_MAC_final(3)>.
853 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
854 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
857 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
858 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
859 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
860 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
861 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
862 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
863 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
864 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
865 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
866 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
868 Use of these low level functions has been informally discouraged
869 for a long time. Applications should use the EVP_DigestInit_ex(3),
870 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
874 * Corrected the documentation of the return values from the `EVP_DigestSign*`
875 set of functions. The documentation mentioned negative values for some
876 errors, but this was never the case, so the mention of negative values
879 Code that followed the documentation and thereby check with something
880 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
884 * All of the low level cipher functions have been deprecated including:
886 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
887 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
888 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
889 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
890 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
891 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
892 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
893 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
894 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
895 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
896 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
897 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
898 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
899 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
900 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
901 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
902 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
903 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
904 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
905 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
906 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
907 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
908 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
909 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
910 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
911 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
912 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
913 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
914 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
916 Use of these low level functions has been informally discouraged for
917 a long time. Applications should use the high level EVP APIs, e.g.
918 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
919 equivalently named decrypt functions instead.
921 *Matt Caswell and Paul Dale*
923 * Removed include/openssl/opensslconf.h.in and replaced it with
924 include/openssl/configuration.h.in, which differs in not including
925 <openssl/macros.h>. A short header include/openssl/opensslconf.h
926 was added to include both.
928 This allows internal hacks where one might need to modify the set
929 of configured macros, for example this if deprecated symbols are
930 still supposed to be available internally:
932 #include <openssl/configuration.h>
934 #undef OPENSSL_NO_DEPRECATED
935 #define OPENSSL_SUPPRESS_DEPRECATED
937 #include <openssl/macros.h>
939 This should not be used by applications that use the exported
940 symbols, as that will lead to linking errors.
944 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
945 used in exponentiation with 512-bit moduli. No EC algorithms are
946 affected. Analysis suggests that attacks against 2-prime RSA1024,
947 3-prime RSA1536, and DSA1024 as a result of this defect would be very
948 difficult to perform and are not believed likely. Attacks against DH512
949 are considered just feasible. However, for an attack the target would
950 have to re-use the DH512 private key, which is not recommended anyway.
951 Also applications directly using the low level API BN_mod_exp may be
952 affected if they use BN_FLG_CONSTTIME.
957 * Most memory-debug features have been deprecated, and the functionality
958 replaced with no-ops.
962 * Added documentation for the STACK API.
966 * Introduced a new method type and API, OSSL_ENCODER, to
967 represent generic encoders. An implementation is expected to
968 be able to encode an object associated with a given name (such
969 as an algorithm name for an asymmetric key) into forms given by
970 implementation properties.
972 Encoders are primarily used from inside libcrypto, through
973 calls to functions like EVP_PKEY_print_private(),
974 PEM_write_bio_PrivateKey() and similar.
976 Encoders are specified in such a way that they can be made to
977 directly handle the provider side portion of an object, if this
978 provider side part comes from the same provider as the encoder
979 itself, but can also be made to handle objects in parametrized
980 form (as an OSSL_PARAM array of data). This allows a provider to
981 offer generic encoders as a service for any other provider.
985 * Added a .pragma directive to the syntax of configuration files, to
986 allow varying behavior in a supported and predictable manner.
987 Currently added pragma:
991 This allows dollar signs to be a keyword character unless it's
992 followed by a opening brace or parenthesis. This is useful for
993 platforms where dollar signs are commonly used in names, such as
994 volume names and system directory names on VMS.
998 * Added functionality to create an EVP_PKEY from user data. This
999 is effectively the same as creating a RSA, DH or DSA object and
1000 then assigning them to an EVP_PKEY, but directly using algorithm
1001 agnostic EVP functions. A benefit is that this should be future
1002 proof for public key algorithms to come.
1006 * Change the interpretation of the '--api' configuration option to
1007 mean that this is a desired API compatibility level with no
1008 further meaning. The previous interpretation, that this would
1009 also mean to remove all deprecated symbols up to and including
1010 the given version, no requires that 'no-deprecated' is also used
1011 in the configuration.
1013 When building applications, the desired API compatibility level
1014 can be set with the OPENSSL_API_COMPAT macro like before. For
1015 API compatibility version below 3.0, the old style numerical
1016 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1017 For version 3.0 and on, the value is expected to be the decimal
1018 value calculated from the major and minor version like this:
1020 MAJOR * 10000 + MINOR * 100
1024 -DOPENSSL_API_COMPAT=30000 For 3.0
1025 -DOPENSSL_API_COMPAT=30200 For 3.2
1027 To hide declarations that are deprecated up to and including the
1028 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1029 given when building the application as well.
1033 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1034 access to certificate and CRL stores via URIs and OSSL_STORE
1037 This adds the following functions:
1039 - X509_LOOKUP_store()
1040 - X509_STORE_load_file()
1041 - X509_STORE_load_path()
1042 - X509_STORE_load_store()
1043 - SSL_add_store_cert_subjects_to_stack()
1044 - SSL_CTX_set_default_verify_store()
1045 - SSL_CTX_load_verify_file()
1046 - SSL_CTX_load_verify_dir()
1047 - SSL_CTX_load_verify_store()
1051 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1052 The presence of this system service is determined at run-time.
1056 * Added functionality to create an EVP_PKEY context based on data
1057 for methods from providers. This takes an algorithm name and a
1058 property query string and simply stores them, with the intent
1059 that any operation that uses this context will use those strings
1060 to fetch the needed methods implicitly, thereby making the port
1061 of application written for pre-3.0 OpenSSL easier.
1065 * The undocumented function NCONF_WIN32() has been deprecated; for
1066 conversion details see the HISTORY section of doc/man5/config.pod
1070 * Introduced the new functions EVP_DigestSignInit_ex() and
1071 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1072 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1073 pages for further details.
1077 * Over two thousand fixes were made to the documentation, including:
1078 adding missing command flags, better style conformance, documentation
1081 *Rich Salz, Richard Levitte*
1083 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1084 X25519, X448, Ed25519 and Ed448.
1088 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1093 * Deprecated the public definition of `ERR_STATE` as well as the function
1094 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1099 * Added ERR functionality to give callers access to the stored function
1100 names that have replaced the older function code based functions.
1102 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1103 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1104 ERR_peek_error_all() and ERR_peek_last_error_all().
1106 These functions have become deprecated: ERR_get_error_line(),
1107 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1108 ERR_peek_last_error_line_data() and ERR_func_error_string().
1110 Users are recommended to use ERR_get_error_all(), or to pick information
1111 with ERR_peek functions and finish off with getting the error code by using
1116 * Extended testing to be verbose for failing tests only. The make variables
1117 VERBOSE_FAILURE or VF can be used to enable this:
1119 $ make VF=1 test # Unix
1120 $ mms /macro=(VF=1) test ! OpenVMS
1121 $ nmake VF=1 test # Windows
1125 * Added the `-copy_extensions` option to the `x509` command for use with
1126 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1127 all extensions in the request are copied to the certificate or vice versa.
1129 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1131 * Added the `-copy_extensions` option to the `req` command for use with
1132 `-x509`. When given with the `copy` or `copyall` argument,
1133 all extensions in the certification request are copied to the certificate.
1137 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1138 they generate are by default RFC 5280 compliant in the following sense:
1139 There is a subjectKeyIdentifier extension with a hash value of the public key
1140 and for not self-signed certs there is an authorityKeyIdentifier extension
1141 with a keyIdentifier field or issuer information identifying the signing key.
1142 This is done unless some configuration overrides the new default behavior,
1143 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1147 * Added several checks to `X509_verify_cert()` according to requirements in
1148 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1149 (which may be done by using the CLI option `-x509_strict`):
1150 * The basicConstraints of CA certificates must be marked critical.
1151 * CA certificates must explicitly include the keyUsage extension.
1152 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1153 * The issuer name of any certificate must not be empty.
1154 * The subject name of CA certs, certs with keyUsage crlSign,
1155 and certs without subjectAlternativeName must not be empty.
1156 * If a subjectAlternativeName extension is given it must not be empty.
1157 * The signatureAlgorithm field and the cert signature must be consistent.
1158 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1159 must not be marked critical.
1160 * The authorityKeyIdentifier must be given for X.509v3 certs
1161 unless they are self-signed.
1162 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1166 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1167 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1171 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1172 used even when parsing explicit parameters, when loading a encoded key
1173 or calling `EC_GROUP_new_from_ecpkparameters()`/
1174 `EC_GROUP_new_from_ecparameters()`.
1175 This prevents bypass of security hardening and performance gains,
1176 especially for curves with specialized EC_METHODs.
1177 By default, if a key encoded with explicit parameters is loaded and later
1178 encoded, the output is still encoded with explicit parameters, even if
1179 internally a "named" EC_GROUP is used for computation.
1183 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1184 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1185 NULL. After this change, only the cofactor parameter can be NULL. It also
1186 does some minimal sanity checks on the passed order.
1191 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1192 An attack is simple, if the first CMS_recipientInfo is valid but the
1193 second CMS_recipientInfo is chosen ciphertext. If the second
1194 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1195 encryption key will be replaced by garbage, and the message cannot be
1196 decoded, but if the RSA decryption fails, the correct encryption key is
1197 used and the recipient will not notice the attack.
1198 As a work around for this potential attack the length of the decrypted
1199 key must be equal to the cipher default key length, in case the
1200 certifiate is not given and all recipientInfo are tried out.
1201 The old behaviour can be re-enabled in the CMS code by setting the
1202 CMS_DEBUG_DECRYPT flag.
1206 * Early start up entropy quality from the DEVRANDOM seed source has been
1207 improved for older Linux systems. The RAND subsystem will wait for
1208 /dev/random to be producing output before seeding from /dev/urandom.
1209 The seeded state is stored for future library initialisations using
1210 a system global shared memory segment. The shared memory identifier
1211 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1212 the desired value. The default identifier is 114.
1216 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1217 when primes for RSA keys are computed.
1218 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1219 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1220 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1221 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1222 This avoids possible fingerprinting of newly generated RSA modules.
1226 * Correct the extended master secret constant on EBCDIC systems. Without this
1227 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1228 negotiate EMS will fail. Unfortunately this also means that TLS connections
1229 between EBCDIC systems with this fix, and EBCDIC systems without this
1230 fix will fail if they negotiate EMS.
1234 * Changed the library initialisation so that the config file is now loaded
1235 by default. This was already the case for libssl. It now occurs for both
1236 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1237 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1241 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1242 where the former acts as a replacement for `ERR_put_error()`, and the
1243 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1244 `ERR_raise_data()` adds more flexibility by taking a format string and
1245 an arbitrary number of arguments following it, to be processed with
1250 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1251 to check if a named provider is loaded and available. When called, it
1252 will also activate all fallback providers if such are still present.
1256 * Enforce a minimum DH modulus size of 512 bits.
1260 * Changed DH parameters to generate the order q subgroup instead of 2q.
1261 Previously generated DH parameters are still accepted by DH_check
1262 but DH_generate_key works around that by clearing bit 0 of the
1263 private key for those. This avoids leaking bit 0 of the private key.
1267 * Significantly reduce secure memory usage by the randomness pools.
1271 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1276 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1277 algorithms. An implementation of a key exchange algorithm can be obtained
1278 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1279 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1280 the older EVP_PKEY_derive_init() function. See the man pages for the new
1281 functions for further details.
1285 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1289 * Removed the function names from error messages and deprecated the
1292 * Removed NextStep support and the macro OPENSSL_UNISTD
1296 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1297 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1298 Also removed "export var as function" capability; we do not export
1299 variables, only functions.
1303 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1304 an error and 1 indicating success. In previous versions of OpenSSL this
1305 was a void type. If a key was set longer than the maximum possible this
1310 * Support SM2 signing and verification schemes with X509 certificate.
1314 * Use SHA256 as the default digest for TS query in the `ts` app.
1318 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1319 This checks that the salt length is at least 128 bits, the derived key
1320 length is at least 112 bits, and that the iteration count is at least 1000.
1321 For backwards compatibility these checks are disabled by default in the
1322 default provider, but are enabled by default in the fips provider.
1323 To enable or disable these checks use the control
1324 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1328 * Default cipher lists/suites are now available via a function, the
1329 #defines are deprecated.
1333 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1334 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1335 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1339 * Join the directories crypto/x509 and crypto/x509v3
1343 * Added command 'openssl kdf' that uses the EVP_KDF API.
1347 * Added command 'openssl mac' that uses the EVP_MAC API.
1351 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1352 as default directories. Also added the command 'openssl info'
1353 for scripting purposes.
1357 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1358 deprecated. These undocumented functions were never integrated into the EVP
1359 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1360 Bi-directional IGE mode. These modes were never formally standardised and
1361 usage of these functions is believed to be very small. In particular
1362 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1363 is ever used. The security implications are believed to be minimal, but
1364 this issue was never fixed for backwards compatibility reasons. New code
1365 should not use these modes.
1369 * Add prediction resistance to the DRBG reseeding process.
1373 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1374 mandated by IEEE Std 1619-2018.
1378 * Added newline escaping functionality to a filename when using openssl dgst.
1379 This output format is to replicate the output format found in the `*sum`
1380 checksum programs. This aims to preserve backward compatibility.
1382 *Matt Eaton, Richard Levitte, and Paul Dale*
1384 * Removed the heartbeat message in DTLS feature, as it has very
1385 little usage and doesn't seem to fulfill a valuable purpose.
1386 The configuration option is now deprecated.
1390 * Changed the output of 'openssl {digestname} < file' to display the
1391 digest name in its output.
1395 * Added a new generic trace API which provides support for enabling
1396 instrumentation through trace output. This feature is mainly intended
1397 as an aid for developers and is disabled by default. To utilize it,
1398 OpenSSL needs to be configured with the `enable-trace` option.
1400 If the tracing API is enabled, the application can activate trace output
1401 by registering BIOs as trace channels for a number of tracing and debugging
1404 The `openssl` program has been expanded to enable any of the types
1405 available via environment variables defined by the user, and serves as
1406 one possible example on how to use this functionality.
1408 *Richard Levitte & Matthias St. Pierre*
1410 * Added build tests for C++. These are generated files that only do one
1411 thing, to include one public OpenSSL head file each. This tests that
1412 the public header files can be usefully included in a C++ application.
1414 This test isn't enabled by default. It can be enabled with the option
1415 'enable-buildtest-c++'.
1419 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1423 * Add KMAC to EVP_MAC.
1427 * Added property based algorithm implementation selection framework to
1432 * Added SCA hardening for modular field inversion in EC_GROUP through
1433 a new dedicated field_inv() pointer in EC_METHOD.
1434 This also addresses a leakage affecting conversions from projective
1435 to affine coordinates.
1437 *Billy Bob Brumley, Nicola Tuveri*
1439 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1440 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1441 those algorithms that were already supported through the EVP_PKEY API
1442 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1443 and scrypt are now wrappers that call EVP_KDF.
1447 * Build devcrypto engine as a dynamic engine.
1449 *Eneas U de Queiroz*
1451 * Add keyed BLAKE2 to EVP_MAC.
1455 * Fix a bug in the computation of the endpoint-pair shared secret used
1456 by DTLS over SCTP. This breaks interoperability with older versions
1457 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1458 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1459 interoperability with such broken implementations. However, enabling
1460 this switch breaks interoperability with correct implementations.
1462 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1463 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1467 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1471 * Change the license to the Apache License v2.0.
1475 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1477 - Major releases (indicated by incrementing the MAJOR release number)
1478 may introduce incompatible API/ABI changes.
1479 - Minor releases (indicated by incrementing the MINOR release number)
1480 may introduce new features but retain API/ABI compatibility.
1481 - Patch releases (indicated by incrementing the PATCH number)
1482 are intended for bug fixes and other improvements of existing
1483 features only (like improving performance or adding documentation)
1484 and retain API/ABI compatibility.
1488 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1492 * Remove the 'dist' target and add a tarball building script. The
1493 'dist' target has fallen out of use, and it shouldn't be
1494 necessary to configure just to create a source distribution.
1498 * Recreate the OS390-Unix config target. It no longer relies on a
1499 special script like it did for OpenSSL pre-1.1.0.
1503 * Instead of having the source directories listed in Configure, add
1504 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1509 * Add GMAC to EVP_MAC.
1513 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1517 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1518 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1519 to facilitate the continued use of MACs through raw private keys in
1520 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1524 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1525 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1529 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1530 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1531 are retained for backwards compatibility.
1535 * AES-XTS mode now enforces that its two keys are different to mitigate
1536 the attacked described in "Efficient Instantiations of Tweakable
1537 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1538 Details of this attack can be obtained from:
1539 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1543 * Rename the object files, i.e. give them other names than in previous
1544 versions. Their names now include the name of the final product, as
1545 well as its type mnemonic (bin, lib, shlib).
1549 * Added new option for 'openssl list', '-objects', which will display the
1550 list of built in objects, i.e. OIDs with names.
1554 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1555 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1560 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1561 improves application performance by removing data copies and providing
1562 applications with zero-copy system calls such as sendfile and splice.
1566 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1567 option is set, openssl cleanses (zeroize) plaintext bytes from
1568 internal buffers after delivering them to the application. Note,
1569 the application is still responsible for cleansing other copies
1570 (e.g.: data received by SSL_read(3)).
1574 * `PKCS12_parse` now maintains the order of the parsed certificates
1575 when outputting them via `*ca` (rather than reversing it).
1579 * Deprecated pthread fork support methods. These were unused so no
1580 replacement is required.
1582 - OPENSSL_fork_prepare()
1583 - OPENSSL_fork_parent()
1584 - OPENSSL_fork_child()
1591 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1593 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1595 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1596 create a unique hash value based on the issuer and serial number data
1597 contained within an X509 certificate. However it was failing to correctly
1598 handle any errors that may occur while parsing the issuer field (which might
1599 occur if the issuer field is maliciously constructed). This may subsequently
1600 result in a NULL pointer deref and a crash leading to a potential denial of
1606 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1607 padding mode to correctly check for rollback attacks. This is considered a
1608 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1613 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1614 functions. Previously they could overflow the output length argument in some
1615 cases where the input length is close to the maximum permissable length for
1616 an integer on the platform. In such cases the return value from the function
1617 call would be 1 (indicating success), but the output length value would be
1618 negative. This could cause applications to behave incorrectly or crash.
1623 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1624 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1625 could be exploited in a side channel attack to recover the password. Since
1626 the attack is local host only this is outside of the current OpenSSL
1627 threat model and therefore no CVE is assigned.
1629 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1634 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1636 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1637 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1638 If an attacker can control both items being compared then this could lead
1639 to a possible denial of service attack. OpenSSL itself uses the
1640 GENERAL_NAME_cmp function for two purposes:
1641 1) Comparing CRL distribution point names between an available CRL and a
1642 CRL distribution point embedded in an X509 certificate
1643 2) When verifying that a timestamp response token signer matches the
1644 timestamp authority name (exposed via the API functions
1645 TS_RESP_verify_response and TS_RESP_verify_token)
1650 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1652 * Certificates with explicit curve parameters are now disallowed in
1653 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1657 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1658 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1659 conversely, silently ignore DTLS protocol version bounds when configuring
1660 TLS-based contexts. The commands can be repeated to set bounds of both
1661 types. The same applies with the corresponding "min_protocol" and
1662 "max_protocol" command-line switches, in case some application uses both TLS
1665 SSL_CTX instances that are created for a fixed protocol version (e.g.
1666 TLSv1_server_method()) also silently ignore version bounds. Previously
1667 attempts to apply bounds to these protocol versions would result in an
1668 error. Now only the "version-flexible" SSL_CTX instances are subject to
1669 limits in configuration files in command-line options.
1673 * Handshake now fails if Extended Master Secret extension is dropped
1678 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1680 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1682 * Fixed segmentation fault in SSL_check_chain()
1683 Server or client applications that call the SSL_check_chain() function
1684 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1685 dereference as a result of incorrect handling of the
1686 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1687 or unrecognised signature algorithm is received from the peer. This could
1688 be exploited by a malicious peer in a Denial of Service attack.
1693 * Added AES consttime code for no-asm configurations
1694 an optional constant time support for AES was added
1695 when building openssl for no-asm.
1696 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1697 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1698 At this time this feature is by default disabled.
1699 It will be enabled by default in 3.0.
1703 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1705 * Revert the change of EOF detection while reading in libssl to avoid
1706 regressions in applications depending on the current way of reporting
1707 the EOF. As the existing method is not fully accurate the change to
1708 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1709 branch and will be present in the 3.0 release.
1713 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1714 when primes for RSA keys are computed.
1715 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1716 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1717 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1718 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1719 This avoids possible fingerprinting of newly generated RSA modules.
1723 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1725 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1726 while reading in libssl then we would report an error back to the
1727 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1728 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1729 therefore give a hint as to what went wrong.
1733 * Check that ed25519 and ed448 are allowed by the security level. Previously
1734 signature algorithms not using an MD were not being checked that they were
1735 allowed by the security level.
1739 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1740 was not quite right. The behaviour was not consistent between resumption
1741 and normal handshakes, and also not quite consistent with historical
1742 behaviour. The behaviour in various scenarios has been clarified and
1743 it has been updated to make it match historical behaviour as closely as
1748 * *[VMS only]* The header files that the VMS compilers include automatically,
1749 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1750 that the C++ compiler doesn't understand. This is a shortcoming in the
1751 compiler, but can be worked around with `__cplusplus` guards.
1753 C++ applications that use OpenSSL libraries must be compiled using the
1754 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1755 functions. Otherwise, only functions with symbols of less than 31
1756 characters can be used, as the linker will not be able to successfully
1757 resolve symbols with longer names.
1761 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1762 The presence of this system service is determined at run-time.
1766 * Added newline escaping functionality to a filename when using openssl dgst.
1767 This output format is to replicate the output format found in the `*sum`
1768 checksum programs. This aims to preserve backward compatibility.
1770 *Matt Eaton, Richard Levitte, and Paul Dale*
1772 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1777 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1779 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1780 number generator (RNG). This was intended to include protection in the
1781 event of a fork() system call in order to ensure that the parent and child
1782 processes did not share the same RNG state. However this protection was not
1783 being used in the default case.
1785 A partial mitigation for this issue is that the output from a high
1786 precision timer is mixed into the RNG state so the likelihood of a parent
1787 and child process sharing state is significantly reduced.
1789 If an application already calls OPENSSL_init_crypto() explicitly using
1790 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1793 *Matthias St. Pierre*
1795 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1796 used even when parsing explicit parameters, when loading a encoded key
1797 or calling `EC_GROUP_new_from_ecpkparameters()`/
1798 `EC_GROUP_new_from_ecparameters()`.
1799 This prevents bypass of security hardening and performance gains,
1800 especially for curves with specialized EC_METHODs.
1801 By default, if a key encoded with explicit parameters is loaded and later
1802 encoded, the output is still encoded with explicit parameters, even if
1803 internally a "named" EC_GROUP is used for computation.
1807 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1808 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1809 NULL. After this change, only the cofactor parameter can be NULL. It also
1810 does some minimal sanity checks on the passed order.
1815 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1816 An attack is simple, if the first CMS_recipientInfo is valid but the
1817 second CMS_recipientInfo is chosen ciphertext. If the second
1818 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1819 encryption key will be replaced by garbage, and the message cannot be
1820 decoded, but if the RSA decryption fails, the correct encryption key is
1821 used and the recipient will not notice the attack.
1822 As a work around for this potential attack the length of the decrypted
1823 key must be equal to the cipher default key length, in case the
1824 certifiate is not given and all recipientInfo are tried out.
1825 The old behaviour can be re-enabled in the CMS code by setting the
1826 CMS_DEBUG_DECRYPT flag.
1831 * Early start up entropy quality from the DEVRANDOM seed source has been
1832 improved for older Linux systems. The RAND subsystem will wait for
1833 /dev/random to be producing output before seeding from /dev/urandom.
1834 The seeded state is stored for future library initialisations using
1835 a system global shared memory segment. The shared memory identifier
1836 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1837 the desired value. The default identifier is 114.
1841 * Correct the extended master secret constant on EBCDIC systems. Without this
1842 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1843 negotiate EMS will fail. Unfortunately this also means that TLS connections
1844 between EBCDIC systems with this fix, and EBCDIC systems without this
1845 fix will fail if they negotiate EMS.
1849 * Use Windows installation paths in the mingw builds
1851 Mingw isn't a POSIX environment per se, which means that Windows
1852 paths should be used for installation.
1857 * Changed DH_check to accept parameters with order q and 2q subgroups.
1858 With order 2q subgroups the bit 0 of the private key is not secret
1859 but DH_generate_key works around that by clearing bit 0 of the
1860 private key for those. This avoids leaking bit 0 of the private key.
1864 * Significantly reduce secure memory usage by the randomness pools.
1868 * Revert the DEVRANDOM_WAIT feature for Linux systems
1870 The DEVRANDOM_WAIT feature added a select() call to wait for the
1871 /dev/random device to become readable before reading from the
1872 /dev/urandom device.
1874 It turned out that this change had negative side effects on
1875 performance which were not acceptable. After some discussion it
1876 was decided to revert this feature and leave it up to the OS
1877 resp. the platform maintainer to ensure a proper initialization
1878 during early boot time.
1880 *Matthias St. Pierre*
1882 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1884 * Add build tests for C++. These are generated files that only do one
1885 thing, to include one public OpenSSL head file each. This tests that
1886 the public header files can be usefully included in a C++ application.
1888 This test isn't enabled by default. It can be enabled with the option
1889 'enable-buildtest-c++'.
1893 * Enable SHA3 pre-hashing for ECDSA and DSA.
1897 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1898 This changes the size when using the `genpkey` command when no size is given.
1899 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1900 generation commands to use 2048 bits by default.
1904 * Reorganize the manual pages to consistently have RETURN VALUES,
1905 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1906 util/fix-doc-nits accordingly.
1908 *Paul Yang, Joshua Lock*
1910 * Add the missing accessor EVP_PKEY_get0_engine()
1914 * Have commands like `s_client` and `s_server` output the signature scheme
1915 along with other cipher suite parameters when debugging.
1919 * Make OPENSSL_config() error agnostic again.
1923 * Do the error handling in RSA decryption constant time.
1927 * Prevent over long nonces in ChaCha20-Poly1305.
1929 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1930 for every encryption operation. RFC 7539 specifies that the nonce value
1931 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1932 and front pads the nonce with 0 bytes if it is less than 12
1933 bytes. However it also incorrectly allows a nonce to be set of up to 16
1934 bytes. In this case only the last 12 bytes are significant and any
1935 additional leading bytes are ignored.
1937 It is a requirement of using this cipher that nonce values are
1938 unique. Messages encrypted using a reused nonce value are susceptible to
1939 serious confidentiality and integrity attacks. If an application changes
1940 the default nonce length to be longer than 12 bytes and then makes a
1941 change to the leading bytes of the nonce expecting the new value to be a
1942 new unique nonce then such an application could inadvertently encrypt
1943 messages with a reused nonce.
1945 Additionally the ignored bytes in a long nonce are not covered by the
1946 integrity guarantee of this cipher. Any application that relies on the
1947 integrity of these ignored leading bytes of a long nonce may be further
1948 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1949 is safe because no such use sets such a long nonce value. However user
1950 applications that use this cipher directly and set a non-default nonce
1951 length to be longer than 12 bytes may be vulnerable.
1953 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1959 * Add DEVRANDOM_WAIT feature for Linux systems
1961 On older Linux systems where the getrandom() system call is not available,
1962 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1963 Contrary to getrandom(), the /dev/urandom device will not block during
1964 early boot when the kernel CSPRNG has not been seeded yet.
1966 To mitigate this known weakness, use select() to wait for /dev/random to
1967 become readable before reading from /dev/urandom.
1969 * Ensure that SM2 only uses SM3 as digest algorithm
1973 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1975 * Change the info callback signals for the start and end of a post-handshake
1976 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1977 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1978 confused by this and assume that a TLSv1.2 renegotiation has started. This
1979 can break KeyUpdate handling. Instead we no longer signal the start and end
1980 of a post handshake message exchange (although the messages themselves are
1981 still signalled). This could break some applications that were expecting
1982 the old signals. However without this KeyUpdate is not usable for many
1987 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1989 * Timing vulnerability in DSA signature generation
1991 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1992 timing side channel attack. An attacker could use variations in the signing
1993 algorithm to recover the private key.
1995 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2000 * Timing vulnerability in ECDSA signature generation
2002 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2003 timing side channel attack. An attacker could use variations in the signing
2004 algorithm to recover the private key.
2006 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2011 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2012 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2013 of two gigabytes and the error handling improved.
2015 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2016 categorized as a normal bug, not a security issue, because the DRBG reseeds
2017 automatically and is fully functional even without additional randomness
2018 provided by the application.
2020 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2022 * Add a new ClientHello callback. Provides a callback interface that gives
2023 the application the ability to adjust the nascent SSL object at the
2024 earliest stage of ClientHello processing, immediately after extensions have
2025 been collected but before they have been processed. In particular, this
2026 callback can adjust the supported TLS versions in response to the contents
2031 * Add SM2 base algorithm support.
2035 * s390x assembly pack: add (improved) hardware-support for the following
2036 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2037 aes-cfb/cfb8, aes-ecb.
2041 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2042 parameter is no longer accepted, as it leads to a corrupt table. NULL
2043 pem_str is reserved for alias entries only.
2047 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2048 step for prime curves. The new implementation is based on formulae from
2049 differential addition-and-doubling in homogeneous projective coordinates
2050 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2051 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2052 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2053 to work in projective coordinates.
2055 *Billy Bob Brumley, Nicola Tuveri*
2057 * Change generating and checking of primes so that the error rate of not
2058 being prime depends on the intended use based on the size of the input.
2059 For larger primes this will result in more rounds of Miller-Rabin.
2060 The maximal error rate for primes with more than 1080 bits is lowered
2063 *Kurt Roeckx, Annie Yousar*
2065 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2069 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2070 moving between systems, and to avoid confusion when a Windows build is
2071 done with mingw vs with MSVC. For POSIX installs, there's still a
2072 symlink or copy named 'tsget' to avoid that confusion as well.
2076 * Revert blinding in ECDSA sign and instead make problematic addition
2077 length-invariant. Switch even to fixed-length Montgomery multiplication.
2081 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2082 step for binary curves. The new implementation is based on formulae from
2083 differential addition-and-doubling in mixed Lopez-Dahab projective
2084 coordinates, modified to independently blind the operands.
2086 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2088 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2089 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2090 EC_METHODs to implement their own specialized "ladder step", to take
2091 advantage of more favorable coordinate systems or more efficient
2092 differential addition-and-doubling algorithms.
2094 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2096 * Modified the random device based seed sources to keep the relevant
2097 file descriptors open rather than reopening them on each access.
2098 This allows such sources to operate in a chroot() jail without
2099 the associated device nodes being available. This behaviour can be
2100 controlled using RAND_keep_random_devices_open().
2104 * Numerous side-channel attack mitigations have been applied. This may have
2105 performance impacts for some algorithms for the benefit of improved
2106 security. Specific changes are noted in this change log by their respective
2111 * AIX shared library support overhaul. Switch to AIX "natural" way of
2112 handling shared libraries, which means collecting shared objects of
2113 different versions and bitnesses in one common archive. This allows to
2114 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2115 doesn't affect the way 3rd party applications are linked, only how
2116 multi-version installation is managed.
2120 * Make ec_group_do_inverse_ord() more robust and available to other
2121 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2122 mitigations are applied to the fallback BN_mod_inverse().
2123 When using this function rather than BN_mod_inverse() directly, new
2124 EC cryptosystem implementations are then safer-by-default.
2128 * Add coordinate blinding for EC_POINT and implement projective
2129 coordinate blinding for generic prime curves as a countermeasure to
2130 chosen point SCA attacks.
2132 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2134 * Add blinding to ECDSA and DSA signatures to protect against side channel
2135 attacks discovered by Keegan Ryan (NCC Group).
2139 * Enforce checking in the `pkeyutl` command to ensure that the input
2140 length does not exceed the maximum supported digest length when performing
2141 a sign, verify or verifyrecover operation.
2145 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2146 I/O in combination with something like select() or poll() will hang. This
2147 can be turned off again using SSL_CTX_clear_mode().
2148 Many applications do not properly handle non-application data records, and
2149 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2150 around the problems in those applications, but can also break some.
2151 It's recommended to read the manpages about SSL_read(), SSL_write(),
2152 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2153 SSL_CTX_set_read_ahead() again.
2157 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2158 now allow empty (zero character) pass phrases.
2162 * Apply blinding to binary field modular inversion and remove patent
2163 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2167 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2168 binary and prime elliptic curves.
2172 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2173 constant time fixed point multiplication.
2177 * Revise elliptic curve scalar multiplication with timing attack
2178 defenses: ec_wNAF_mul redirects to a constant time implementation
2179 when computing fixed point and variable point multiplication (which
2180 in OpenSSL are mostly used with secret scalars in keygen, sign,
2181 ECDH derive operations).
2182 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2185 * Updated CONTRIBUTING
2189 * Updated DRBG / RAND to request nonce and additional low entropy
2190 randomness from the system.
2192 *Matthias St. Pierre*
2194 * Updated 'openssl rehash' to use OpenSSL consistent default.
2198 * Moved the load of the ssl_conf module to libcrypto, which helps
2199 loading engines that libssl uses before libssl is initialised.
2203 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2207 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2209 *Ingo Schwarze, Rich Salz*
2211 * Added output of accepting IP address and port for 'openssl s_server'
2215 * Added a new API for TLSv1.3 ciphersuites:
2216 SSL_CTX_set_ciphersuites()
2217 SSL_set_ciphersuites()
2221 * Memory allocation failures consistently add an error to the error
2226 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2227 in libcrypto when run as setuid/setgid.
2231 * Load any config file by default when libssl is used.
2235 * Added new public header file <openssl/rand_drbg.h> and documentation
2236 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2238 *Matthias St. Pierre*
2240 * QNX support removed (cannot find contributors to get their approval
2241 for the license change).
2245 * TLSv1.3 replay protection for early data has been implemented. See the
2246 SSL_read_early_data() man page for further details.
2250 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2251 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2252 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2253 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2254 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2255 configuration has been separated out. See the ciphers man page or the
2256 SSL_CTX_set_ciphersuites() man page for more information.
2260 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2261 in responder mode now supports the new "-multi" option, which
2262 spawns the specified number of child processes to handle OCSP
2263 requests. The "-timeout" option now also limits the OCSP
2264 responder's patience to wait to receive the full client request
2265 on a newly accepted connection. Child processes are respawned
2266 as needed, and the CA index file is automatically reloaded
2267 when changed. This makes it possible to run the "ocsp" responder
2268 as a long-running service, making the OpenSSL CA somewhat more
2269 feature-complete. In this mode, most diagnostic messages logged
2270 after entering the event loop are logged via syslog(3) rather than
2275 * Added support for X448 and Ed448. Heavily based on original work by
2280 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2281 objects loaded. This adds the functions OSSL_STORE_expect() and
2282 OSSL_STORE_find() as well as needed tools to construct searches and
2283 get the search data out of them.
2287 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2288 version of OpenSSL should review their configuration settings to ensure
2289 that they are still appropriate for TLSv1.3. For further information see:
2290 <https://wiki.openssl.org/index.php/TLS1.3>
2294 * Grand redesign of the OpenSSL random generator
2296 The default RAND method now utilizes an AES-CTR DRBG according to
2297 NIST standard SP 800-90Ar1. The new random generator is essentially
2298 a port of the default random generator from the OpenSSL FIPS 2.0
2299 object module. It is a hybrid deterministic random bit generator
2300 using an AES-CTR bit stream and which seeds and reseeds itself
2301 automatically using trusted system entropy sources.
2303 Some of its new features are:
2304 - Support for multiple DRBG instances with seed chaining.
2305 - The default RAND method makes use of a DRBG.
2306 - There is a public and private DRBG instance.
2307 - The DRBG instances are fork-safe.
2308 - Keep all global DRBG instances on the secure heap if it is enabled.
2309 - The public and private DRBG instance are per thread for lock free
2312 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2314 * Changed Configure so it only says what it does and doesn't dump
2315 so much data. Instead, ./configdata.pm should be used as a script
2316 to display all sorts of configuration data.
2320 * Added processing of "make variables" to Configure.
2324 * Added SHA512/224 and SHA512/256 algorithm support.
2328 * The last traces of Netware support, first removed in 1.1.0, have
2333 * Get rid of Makefile.shared, and in the process, make the processing
2334 of certain files (rc.obj, or the .def/.map/.opt files produced from
2335 the ordinal files) more visible and hopefully easier to trace and
2336 debug (or make silent).
2340 * Make it possible to have environment variable assignments as
2341 arguments to config / Configure.
2345 * Add multi-prime RSA (RFC 8017) support.
2349 * Add SM3 implemented according to GB/T 32905-2016
2350 *Jack Lloyd <jack.lloyd@ribose.com>,*
2351 *Ronald Tse <ronald.tse@ribose.com>,*
2352 *Erick Borsboom <erick.borsboom@ribose.com>*
2354 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2355 as documented in RFC6066.
2356 Based on a patch from Tomasz Moń
2358 *Filipe Raimundo da Silva*
2360 * Add SM4 implemented according to GB/T 32907-2016.
2361 *Jack Lloyd <jack.lloyd@ribose.com>,*
2362 *Ronald Tse <ronald.tse@ribose.com>,*
2363 *Erick Borsboom <erick.borsboom@ribose.com>*
2365 * Reimplement -newreq-nodes and ERR_error_string_n; the
2366 original author does not agree with the license change.
2370 * Add ARIA AEAD TLS support.
2374 * Some macro definitions to support VS6 have been removed. Visual
2375 Studio 6 has not worked since 1.1.0
2379 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2380 without clearing the errors.
2384 * Add "atfork" functions. If building on a system that without
2385 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2386 requirements. The RAND facility now uses/requires this.
2394 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2395 not possible to disable entirely. However, it's still possible to
2396 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2399 To disable, configure with 'no-ui-console'. 'no-ui' is still
2400 possible to use as an alias. Check at compile time with the
2401 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2402 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2406 * Add a STORE module, which implements a uniform and URI based reader of
2407 stores that can contain keys, certificates, CRLs and numerous other
2408 objects. The main API is loosely based on a few stdio functions,
2409 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2410 OSSL_STORE_error and OSSL_STORE_close.
2411 The implementation uses backends called "loaders" to implement arbitrary
2412 URI schemes. There is one built in "loader" for the 'file' scheme.
2416 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2417 then adjusted to work on FreeBSD 8.4 as well.
2418 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2419 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2423 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2424 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2425 error code calls like this:
2427 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2429 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2430 that can be encoded in C. For the foreseeable future, this will only
2433 *Richard Levitte and Tim Hudson*
2435 * Removed BSD cryptodev engine.
2439 * Add a build target 'build_all_generated', to build all generated files
2440 and only that. This can be used to prepare everything that requires
2441 things like perl for a system that lacks perl and then move everything
2442 to that system and do the rest of the build there.
2446 * In the UI interface, make it possible to duplicate the user data. This
2447 can be used by engines that need to retain the data for a longer time
2448 than just the call where this user data is passed.
2452 * Ignore the '-named_curve auto' value for compatibility of applications
2455 *Tomáš Mráz <tmraz@fedoraproject.org>*
2457 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2458 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2459 alerts across multiple records (some of which could be empty). In practice
2460 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2461 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2462 support this at all. Supporting it adds significant complexity to the
2463 record layer, and its removal is unlikely to cause interoperability
2468 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2469 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2470 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2475 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2476 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2478 *Richard Levitte, Andy Polyakov*
2480 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2485 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2486 platform rather than 'mingw'.
2490 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2491 success if they are asked to add an object which already exists
2492 in the store. This change cascades to other functions which load
2493 certificates and CRLs.
2497 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2498 facilitate stack unwinding even from assembly subroutines.
2502 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2503 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2507 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2508 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2509 which is the minimum version we support.
2513 * Certificate time validation (X509_cmp_time) enforces stricter
2514 compliance with RFC 5280. Fractional seconds and timezone offsets
2515 are no longer allowed.
2519 * Add support for ARIA
2523 * s_client will now send the Server Name Indication (SNI) extension by
2524 default unless the new "-noservername" option is used. The server name is
2525 based on the host provided to the "-connect" option unless overridden by
2526 using "-servername".
2530 * Add support for SipHash
2534 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2535 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2536 prevent issues where no progress is being made and the peer continually
2537 sends unrecognised record types, using up resources processing them.
2541 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2542 using the algorithm defined in
2543 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2547 * Heartbeat support has been removed; the ABI is changed for now.
2549 *Richard Levitte, Rich Salz*
2551 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2555 * The RSA "null" method, which was partially supported to avoid patent
2556 issues, has been replaced to always returns NULL.
2563 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2565 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2566 used even when parsing explicit parameters, when loading a encoded key
2567 or calling `EC_GROUP_new_from_ecpkparameters()`/
2568 `EC_GROUP_new_from_ecparameters()`.
2569 This prevents bypass of security hardening and performance gains,
2570 especially for curves with specialized EC_METHODs.
2571 By default, if a key encoded with explicit parameters is loaded and later
2572 encoded, the output is still encoded with explicit parameters, even if
2573 internally a "named" EC_GROUP is used for computation.
2577 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2578 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2579 NULL. After this change, only the cofactor parameter can be NULL. It also
2580 does some minimal sanity checks on the passed order.
2585 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2586 An attack is simple, if the first CMS_recipientInfo is valid but the
2587 second CMS_recipientInfo is chosen ciphertext. If the second
2588 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2589 encryption key will be replaced by garbage, and the message cannot be
2590 decoded, but if the RSA decryption fails, the correct encryption key is
2591 used and the recipient will not notice the attack.
2592 As a work around for this potential attack the length of the decrypted
2593 key must be equal to the cipher default key length, in case the
2594 certifiate is not given and all recipientInfo are tried out.
2595 The old behaviour can be re-enabled in the CMS code by setting the
2596 CMS_DEBUG_DECRYPT flag.
2601 * Use Windows installation paths in the mingw builds
2603 Mingw isn't a POSIX environment per se, which means that Windows
2604 paths should be used for installation.
2609 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2611 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2612 This changes the size when using the `genpkey` command when no size is given.
2613 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2614 generation commands to use 2048 bits by default.
2618 * Prevent over long nonces in ChaCha20-Poly1305.
2620 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2621 for every encryption operation. RFC 7539 specifies that the nonce value
2622 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2623 and front pads the nonce with 0 bytes if it is less than 12
2624 bytes. However it also incorrectly allows a nonce to be set of up to 16
2625 bytes. In this case only the last 12 bytes are significant and any
2626 additional leading bytes are ignored.
2628 It is a requirement of using this cipher that nonce values are
2629 unique. Messages encrypted using a reused nonce value are susceptible to
2630 serious confidentiality and integrity attacks. If an application changes
2631 the default nonce length to be longer than 12 bytes and then makes a
2632 change to the leading bytes of the nonce expecting the new value to be a
2633 new unique nonce then such an application could inadvertently encrypt
2634 messages with a reused nonce.
2636 Additionally the ignored bytes in a long nonce are not covered by the
2637 integrity guarantee of this cipher. Any application that relies on the
2638 integrity of these ignored leading bytes of a long nonce may be further
2639 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2640 is safe because no such use sets such a long nonce value. However user
2641 applications that use this cipher directly and set a non-default nonce
2642 length to be longer than 12 bytes may be vulnerable.
2644 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2650 * Added SCA hardening for modular field inversion in EC_GROUP through
2651 a new dedicated field_inv() pointer in EC_METHOD.
2652 This also addresses a leakage affecting conversions from projective
2653 to affine coordinates.
2655 *Billy Bob Brumley, Nicola Tuveri*
2657 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2658 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2662 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2666 * Remove the 'dist' target and add a tarball building script. The
2667 'dist' target has fallen out of use, and it shouldn't be
2668 necessary to configure just to create a source distribution.
2672 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2674 * Timing vulnerability in DSA signature generation
2676 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2677 timing side channel attack. An attacker could use variations in the signing
2678 algorithm to recover the private key.
2680 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2685 * Timing vulnerability in ECDSA signature generation
2687 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2688 timing side channel attack. An attacker could use variations in the signing
2689 algorithm to recover the private key.
2691 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2696 * Add coordinate blinding for EC_POINT and implement projective
2697 coordinate blinding for generic prime curves as a countermeasure to
2698 chosen point SCA attacks.
2700 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2702 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2704 * Client DoS due to large DH parameter
2706 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2707 malicious server can send a very large prime value to the client. This will
2708 cause the client to spend an unreasonably long period of time generating a
2709 key for this prime resulting in a hang until the client has finished. This
2710 could be exploited in a Denial Of Service attack.
2712 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2717 * Cache timing vulnerability in RSA Key Generation
2719 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2720 a cache timing side channel attack. An attacker with sufficient access to
2721 mount cache timing attacks during the RSA key generation process could
2722 recover the private key.
2724 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2725 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2730 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2731 parameter is no longer accepted, as it leads to a corrupt table. NULL
2732 pem_str is reserved for alias entries only.
2736 * Revert blinding in ECDSA sign and instead make problematic addition
2737 length-invariant. Switch even to fixed-length Montgomery multiplication.
2741 * Change generating and checking of primes so that the error rate of not
2742 being prime depends on the intended use based on the size of the input.
2743 For larger primes this will result in more rounds of Miller-Rabin.
2744 The maximal error rate for primes with more than 1080 bits is lowered
2747 *Kurt Roeckx, Annie Yousar*
2749 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2753 * Add blinding to ECDSA and DSA signatures to protect against side channel
2754 attacks discovered by Keegan Ryan (NCC Group).
2758 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2759 now allow empty (zero character) pass phrases.
2763 * Certificate time validation (X509_cmp_time) enforces stricter
2764 compliance with RFC 5280. Fractional seconds and timezone offsets
2765 are no longer allowed.
2769 * Fixed a text canonicalisation bug in CMS
2771 Where a CMS detached signature is used with text content the text goes
2772 through a canonicalisation process first prior to signing or verifying a
2773 signature. This process strips trailing space at the end of lines, converts
2774 line terminators to CRLF and removes additional trailing line terminators
2775 at the end of a file. A bug in the canonicalisation process meant that
2776 some characters, such as form-feed, were incorrectly treated as whitespace
2777 and removed. This is contrary to the specification (RFC5485). This fix
2778 could mean that detached text data signed with an earlier version of
2779 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2780 signed with a fixed OpenSSL may fail to verify with an earlier version of
2781 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2782 and use the "-binary" flag (for the "cms" command line application) or set
2783 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2787 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2789 * Constructed ASN.1 types with a recursive definition could exceed the stack
2791 Constructed ASN.1 types with a recursive definition (such as can be found
2792 in PKCS7) could eventually exceed the stack given malicious input with
2793 excessive recursion. This could result in a Denial Of Service attack. There
2794 are no such structures used within SSL/TLS that come from untrusted sources
2795 so this is considered safe.
2797 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2803 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2805 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2806 effectively reduced to only comparing the least significant bit of each
2807 byte. This allows an attacker to forge messages that would be considered as
2808 authenticated in an amount of tries lower than that guaranteed by the
2809 security claims of the scheme. The module can only be compiled by the
2810 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2812 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2818 * Add a build target 'build_all_generated', to build all generated files
2819 and only that. This can be used to prepare everything that requires
2820 things like perl for a system that lacks perl and then move everything
2821 to that system and do the rest of the build there.
2825 * Backport SSL_OP_NO_RENGOTIATION
2827 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2828 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2829 changes this is no longer possible in 1.1.0. Therefore the new
2830 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2831 1.1.0 to provide equivalent functionality.
2833 Note that if an application built against 1.1.0h headers (or above) is run
2834 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2835 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2839 * Removed the OS390-Unix config target. It relied on a script that doesn't
2844 * rsaz_1024_mul_avx2 overflow bug on x86_64
2846 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2847 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2848 Analysis suggests that attacks against RSA and DSA as a result of this
2849 defect would be very difficult to perform and are not believed likely.
2850 Attacks against DH1024 are considered just feasible, because most of the
2851 work necessary to deduce information about a private key may be performed
2852 offline. The amount of resources required for such an attack would be
2853 significant. However, for an attack on TLS to be meaningful, the server
2854 would have to share the DH1024 private key among multiple clients, which is
2855 no longer an option since CVE-2016-0701.
2857 This only affects processors that support the AVX2 but not ADX extensions
2858 like Intel Haswell (4th generation).
2860 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2861 was originally found via the OSS-Fuzz project.
2866 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2868 * bn_sqrx8x_internal carry bug on x86_64
2870 There is a carry propagating bug in the x86_64 Montgomery squaring
2871 procedure. No EC algorithms are affected. Analysis suggests that attacks
2872 against RSA and DSA as a result of this defect would be very difficult to
2873 perform and are not believed likely. Attacks against DH are considered just
2874 feasible (although very difficult) because most of the work necessary to
2875 deduce information about a private key may be performed offline. The amount
2876 of resources required for such an attack would be very significant and
2877 likely only accessible to a limited number of attackers. An attacker would
2878 additionally need online access to an unpatched system using the target
2879 private key in a scenario with persistent DH parameters and a private
2880 key that is shared between multiple clients.
2882 This only affects processors that support the BMI1, BMI2 and ADX extensions
2883 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2885 This issue was reported to OpenSSL by the OSS-Fuzz project.
2890 * Malformed X.509 IPAddressFamily could cause OOB read
2892 If an X.509 certificate has a malformed IPAddressFamily extension,
2893 OpenSSL could do a one-byte buffer overread. The most likely result
2894 would be an erroneous display of the certificate in text format.
2896 This issue was reported to OpenSSL by the OSS-Fuzz project.
2901 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2903 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2904 platform rather than 'mingw'.
2908 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2909 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2910 which is the minimum version we support.
2914 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2916 * Encrypt-Then-Mac renegotiation crash
2918 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2919 negotiated where it was not in the original handshake (or vice-versa) then
2920 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2921 and servers are affected.
2923 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2928 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2930 * Truncated packet could crash via OOB read
2932 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2933 cipher is being used, then a truncated packet can cause that host to
2934 perform an out-of-bounds read, usually resulting in a crash.
2936 This issue was reported to OpenSSL by Robert Święcki of Google.
2941 * Bad (EC)DHE parameters cause a client crash
2943 If a malicious server supplies bad parameters for a DHE or ECDHE key
2944 exchange then this can result in the client attempting to dereference a
2945 NULL pointer leading to a client crash. This could be exploited in a Denial
2948 This issue was reported to OpenSSL by Guido Vranken.
2953 * BN_mod_exp may produce incorrect results on x86_64
2955 There is a carry propagating bug in the x86_64 Montgomery squaring
2956 procedure. No EC algorithms are affected. Analysis suggests that attacks
2957 against RSA and DSA as a result of this defect would be very difficult to
2958 perform and are not believed likely. Attacks against DH are considered just
2959 feasible (although very difficult) because most of the work necessary to
2960 deduce information about a private key may be performed offline. The amount
2961 of resources required for such an attack would be very significant and
2962 likely only accessible to a limited number of attackers. An attacker would
2963 additionally need online access to an unpatched system using the target
2964 private key in a scenario with persistent DH parameters and a private
2965 key that is shared between multiple clients. For example this can occur by
2966 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2967 similar to CVE-2015-3193 but must be treated as a separate problem.
2969 This issue was reported to OpenSSL by the OSS-Fuzz project.
2974 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2976 * ChaCha20/Poly1305 heap-buffer-overflow
2978 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2979 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2980 crash. This issue is not considered to be exploitable beyond a DoS.
2982 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2987 * CMS Null dereference
2989 Applications parsing invalid CMS structures can crash with a NULL pointer
2990 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2991 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2992 structure callback if an attempt is made to free certain invalid encodings.
2993 Only CHOICE structures using a callback which do not handle NULL value are
2996 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3001 * Montgomery multiplication may produce incorrect results
3003 There is a carry propagating bug in the Broadwell-specific Montgomery
3004 multiplication procedure that handles input lengths divisible by, but
3005 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3006 and DH private keys are impossible. This is because the subroutine in
3007 question is not used in operations with the private key itself and an input
3008 of the attacker's direct choice. Otherwise the bug can manifest itself as
3009 transient authentication and key negotiation failures or reproducible
3010 erroneous outcome of public-key operations with specially crafted input.
3011 Among EC algorithms only Brainpool P-512 curves are affected and one
3012 presumably can attack ECDH key negotiation. Impact was not analyzed in
3013 detail, because pre-requisites for attack are considered unlikely. Namely
3014 multiple clients have to choose the curve in question and the server has to
3015 share the private key among them, neither of which is default behaviour.
3016 Even then only clients that chose the curve will be affected.
3018 This issue was publicly reported as transient failures and was not
3019 initially recognized as a security issue. Thanks to Richard Morgan for
3020 providing reproducible case.
3025 * Removed automatic addition of RPATH in shared libraries and executables,
3026 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3030 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3032 * Fix Use After Free for large message sizes
3034 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3035 message larger than approx 16k is received then the underlying buffer to
3036 store the incoming message is reallocated and moved. Unfortunately a
3037 dangling pointer to the old location is left which results in an attempt to
3038 write to the previously freed location. This is likely to result in a
3039 crash, however it could potentially lead to execution of arbitrary code.
3041 This issue only affects OpenSSL 1.1.0a.
3043 This issue was reported to OpenSSL by Robert Święcki.
3048 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3050 * OCSP Status Request extension unbounded memory growth
3052 A malicious client can send an excessively large OCSP Status Request
3053 extension. If that client continually requests renegotiation, sending a
3054 large OCSP Status Request extension each time, then there will be unbounded
3055 memory growth on the server. This will eventually lead to a Denial Of
3056 Service attack through memory exhaustion. Servers with a default
3057 configuration are vulnerable even if they do not support OCSP. Builds using
3058 the "no-ocsp" build time option are not affected.
3060 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3065 * SSL_peek() hang on empty record
3067 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3068 sends an empty record. This could be exploited by a malicious peer in a
3069 Denial Of Service attack.
3071 This issue was reported to OpenSSL by Alex Gaynor.
3076 * Excessive allocation of memory in tls_get_message_header() and
3077 dtls1_preprocess_fragment()
3079 A (D)TLS message includes 3 bytes for its length in the header for the
3080 message. This would allow for messages up to 16Mb in length. Messages of
3081 this length are excessive and OpenSSL includes a check to ensure that a
3082 peer is sending reasonably sized messages in order to avoid too much memory
3083 being consumed to service a connection. A flaw in the logic of version
3084 1.1.0 means that memory for the message is allocated too early, prior to
3085 the excessive message length check. Due to way memory is allocated in
3086 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3087 to service a connection. This could lead to a Denial of Service through
3088 memory exhaustion. However, the excessive message length check still takes
3089 place, and this would cause the connection to immediately fail. Assuming
3090 that the application calls SSL_free() on the failed connection in a timely
3091 manner then the 21Mb of allocated memory will then be immediately freed
3092 again. Therefore the excessive memory allocation will be transitory in
3093 nature. This then means that there is only a security impact if:
3095 1) The application does not call SSL_free() in a timely manner in the event
3096 that the connection fails
3098 2) The application is working in a constrained environment where there is
3099 very little free memory
3101 3) The attacker initiates multiple connection attempts such that there are
3102 multiple connections in a state where memory has been allocated for the
3103 connection; SSL_free() has not yet been called; and there is insufficient
3104 memory to service the multiple requests.
3106 Except in the instance of (1) above any Denial Of Service is likely to be
3107 transitory because as soon as the connection fails the memory is
3108 subsequently freed again in the SSL_free() call. However there is an
3109 increased risk during this period of application crashes due to the lack of
3110 memory - which would then mean a more serious Denial of Service.
3112 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3113 (CVE-2016-6307 and CVE-2016-6308)
3117 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3118 had to be removed. Primary reason is that vendor assembler can't
3119 assemble our modules with -KPIC flag. As result it, assembly
3120 support, was not even available as option. But its lack means
3121 lack of side-channel resistant code, which is incompatible with
3122 security by todays standards. Fortunately gcc is readily available
3123 prepackaged option, which we firmly point at...
3127 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3129 * Windows command-line tool supports UTF-8 opt-in option for arguments
3130 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3131 (to any value) allows Windows user to access PKCS#12 file generated
3132 with Windows CryptoAPI and protected with non-ASCII password, as well
3133 as files generated under UTF-8 locale on Linux also protected with
3138 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3139 have been disabled by default and removed from DEFAULT, just like RC4.
3140 See the RC4 item below to re-enable both.
3144 * The method for finding the storage location for the Windows RAND seed file
3145 has changed. First we check %RANDFILE%. If that is not set then we check
3146 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3147 all else fails we fall back to C:\.
3151 * The EVP_EncryptUpdate() function has had its return type changed from void
3152 to int. A return of 0 indicates and error while a return of 1 indicates
3157 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3158 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3159 off the constant time implementation for RSA, DSA and DH have been made
3160 no-ops and deprecated.
3164 * Windows RAND implementation was simplified to only get entropy by
3165 calling CryptGenRandom(). Various other RAND-related tickets
3168 *Joseph Wylie Yandle, Rich Salz*
3170 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3171 and `OPENSSL_LH_`, respectively. The old names are available
3172 with API compatibility. They new names are now completely documented.
3176 * Unify TYPE_up_ref(obj) methods signature.
3177 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3178 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3179 int (instead of void) like all others TYPE_up_ref() methods.
3180 So now these methods also check the return value of CRYPTO_atomic_add(),
3181 and the validity of object reference counter.
3183 *fdasilvayy@gmail.com*
3185 * With Windows Visual Studio builds, the .pdb files are installed
3186 alongside the installed libraries and executables. For a static
3187 library installation, ossl_static.pdb is the associate compiler
3188 generated .pdb file to be used when linking programs.
3192 * Remove openssl.spec. Packaging files belong with the packagers.
3196 * Automatic Darwin/OSX configuration has had a refresh, it will now
3197 recognise x86_64 architectures automatically. You can still decide
3198 to build for a different bitness with the environment variable
3199 KERNEL_BITS (can be 32 or 64), for example:
3201 KERNEL_BITS=32 ./config
3205 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3206 256 bit AES and HMAC with SHA256.
3210 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3214 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3218 * To enable users to have their own config files and build file templates,
3219 Configure looks in the directory indicated by the environment variable
3220 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3221 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3222 name and is used as is.
3226 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3227 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3228 X509_CERT_FILE_CTX was removed.
3232 * "shared" builds are now the default. To create only static libraries use
3233 the "no-shared" Configure option.
3237 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3238 All of these option have not worked for some while and are fundamental
3243 * Make various cleanup routines no-ops and mark them as deprecated. Most
3244 global cleanup functions are no longer required because they are handled
3245 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3246 Explicitly de-initing can cause problems (e.g. where a library that uses
3247 OpenSSL de-inits, but an application is still using it). The affected
3248 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3249 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3250 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3251 COMP_zlib_cleanup().
3255 * --strict-warnings no longer enables runtime debugging options
3256 such as REF_DEBUG. Instead, debug options are automatically
3257 enabled with '--debug' builds.
3259 *Andy Polyakov, Emilia Käsper*
3261 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3262 have been moved out of the public header files. New functions for managing
3263 these have been added.
3267 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3268 objects have been moved out of the public header files. New
3269 functions for managing these have been added.
3273 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3274 have been moved out of the public header files. New functions for managing
3275 these have been added.
3279 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3280 moved out of the public header files. New functions for managing these
3285 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3289 * Removed the mk1mf build scripts.
3293 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3294 it is always safe to #include a header now.
3298 * Removed the aged BC-32 config and all its supporting scripts
3302 * Removed support for Ultrix, Netware, and OS/2.
3306 * Add support for HKDF.
3308 *Alessandro Ghedini*
3310 * Add support for blake2b and blake2s
3314 * Added support for "pipelining". Ciphers that have the
3315 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3316 encryptions/decryptions simultaneously. There are currently no built-in
3317 ciphers with this property but the expectation is that engines will be able
3318 to offer it to significantly improve throughput. Support has been extended
3319 into libssl so that multiple records for a single connection can be
3320 processed in one go (for >=TLS 1.1).
3324 * Added the AFALG engine. This is an async capable engine which is able to
3325 offload work to the Linux kernel. In this initial version it only supports
3326 AES128-CBC. The kernel must be version 4.1.0 or greater.
3330 * OpenSSL now uses a new threading API. It is no longer necessary to
3331 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3332 are two supported threading models: pthreads and windows threads. It is
3333 also possible to configure OpenSSL at compile time for "no-threads". The
3334 old threading API should no longer be used. The functions have been
3335 replaced with "no-op" compatibility macros.
3337 *Alessandro Ghedini, Matt Caswell*
3339 * Modify behavior of ALPN to invoke callback after SNI/servername
3340 callback, such that updates to the SSL_CTX affect ALPN.
3344 * Add SSL_CIPHER queries for authentication and key-exchange.
3348 * Changes to the DEFAULT cipherlist:
3349 - Prefer (EC)DHE handshakes over plain RSA.
3350 - Prefer AEAD ciphers over legacy ciphers.
3351 - Prefer ECDSA over RSA when both certificates are available.
3352 - Prefer TLSv1.2 ciphers/PRF.
3353 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3358 * Change the ECC default curve list to be this, in order: x25519,
3359 secp256r1, secp521r1, secp384r1.
3363 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3364 disabled by default. They can be re-enabled using the
3365 enable-weak-ssl-ciphers option to Configure.
3369 * If the server has ALPN configured, but supports no protocols that the
3370 client advertises, send a fatal "no_application_protocol" alert.
3371 This behaviour is SHALL in RFC 7301, though it isn't universally
3372 implemented by other servers.
3376 * Add X25519 support.
3377 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3378 for public and private key encoding using the format documented in
3379 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3380 key generation and key derivation.
3382 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3387 * Deprecate SRP_VBASE_get_by_user.
3388 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3389 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3390 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3391 seed, even if the seed is configured.
3393 Users should use SRP_VBASE_get1_by_user instead. Note that in
3394 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3395 also that even though configuring the SRP seed attempts to hide
3396 invalid usernames by continuing the handshake with fake
3397 credentials, this behaviour is not constant time and no strong
3398 guarantees are made that the handshake is indistinguishable from
3399 that of a valid user.
3403 * Configuration change; it's now possible to build dynamic engines
3404 without having to build shared libraries and vice versa. This
3405 only applies to the engines in `engines/`, those in `crypto/engine/`
3406 will always be built into libcrypto (i.e. "static").
3408 Building dynamic engines is enabled by default; to disable, use
3409 the configuration option "disable-dynamic-engine".
3411 The only requirements for building dynamic engines are the
3412 presence of the DSO module and building with position independent
3413 code, so they will also automatically be disabled if configuring
3414 with "disable-dso" or "disable-pic".
3416 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3417 are also taken away from openssl/opensslconf.h, as they are
3422 * Configuration change; if there is a known flag to compile
3423 position independent code, it will always be applied on the
3424 libcrypto and libssl object files, and never on the application
3425 object files. This means other libraries that use routines from
3426 libcrypto / libssl can be made into shared libraries regardless
3427 of how OpenSSL was configured.
3429 If this isn't desirable, the configuration options "disable-pic"
3430 or "no-pic" can be used to disable the use of PIC. This will
3431 also disable building shared libraries and dynamic engines.
3435 * Removed JPAKE code. It was experimental and has no wide use.
3439 * The INSTALL_PREFIX Makefile variable has been renamed to
3440 DESTDIR. That makes for less confusion on what this variable
3441 is for. Also, the configuration option --install_prefix is
3446 * Heartbeat for TLS has been removed and is disabled by default
3447 for DTLS; configure with enable-heartbeats. Code that uses the
3448 old #define's might need to be updated.
3450 *Emilia Käsper, Rich Salz*
3452 * Rename REF_CHECK to REF_DEBUG.
3456 * New "unified" build system
3458 The "unified" build system is aimed to be a common system for all
3459 platforms we support. With it comes new support for VMS.
3461 This system builds supports building in a different directory tree
3462 than the source tree. It produces one Makefile (for unix family
3463 or lookalikes), or one descrip.mms (for VMS).
3465 The source of information to make the Makefile / descrip.mms is
3466 small files called 'build.info', holding the necessary
3467 information for each directory with source to compile, and a
3468 template in Configurations, like unix-Makefile.tmpl or
3471 With this change, the library names were also renamed on Windows
3472 and on VMS. They now have names that are closer to the standard
3473 on Unix, and include the major version number, and in certain
3474 cases, the architecture they are built for. See "Notes on shared
3475 libraries" in INSTALL.
3477 We rely heavily on the perl module Text::Template.
3481 * Added support for auto-initialisation and de-initialisation of the library.
3482 OpenSSL no longer requires explicit init or deinit routines to be called,
3483 except in certain circumstances. See the OPENSSL_init_crypto() and
3484 OPENSSL_init_ssl() man pages for further information.
3488 * The arguments to the DTLSv1_listen function have changed. Specifically the
3489 "peer" argument is now expected to be a BIO_ADDR object.
3491 * Rewrite of BIO networking library. The BIO library lacked consistent
3492 support of IPv6, and adding it required some more extensive
3493 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3494 which hold all types of addresses and chains of address information.
3495 It also introduces a new API, with functions like BIO_socket,
3496 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3497 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3498 have been adapted accordingly.
3502 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3507 * CRIME protection: disable compression by default, even if OpenSSL is
3508 compiled with zlib enabled. Applications can still enable compression
3509 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3510 using the SSL_CONF library to configure compression.
3514 * The signature of the session callback configured with
3515 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3516 was explicitly marked as `const unsigned char*` instead of
3521 * Always DPURIFY. Remove the use of uninitialized memory in the
3522 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3526 * Removed many obsolete configuration items, including
3527 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3528 MD2_CHAR, MD2_INT, MD2_LONG
3530 IDEA_SHORT, IDEA_LONG
3531 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3533 *Rich Salz, with advice from Andy Polyakov*
3535 * Many BN internals have been moved to an internal header file.
3537 *Rich Salz with help from Andy Polyakov*
3539 * Configuration and writing out the results from it has changed.
3540 Files such as Makefile include/openssl/opensslconf.h and are now
3541 produced through general templates, such as Makefile.in and
3542 crypto/opensslconf.h.in and some help from the perl module
3545 Also, the center of configuration information is no longer
3546 Makefile. Instead, Configure produces a perl module in
3547 configdata.pm which holds most of the config data (in the hash
3548 table %config), the target data that comes from the target
3549 configuration in one of the `Configurations/*.conf` files (in
3554 * To clarify their intended purposes, the Configure options
3555 --prefix and --openssldir change their semantics, and become more
3556 straightforward and less interdependent.
3558 --prefix shall be used exclusively to give the location INSTALLTOP
3559 where programs, scripts, libraries, include files and manuals are
3560 going to be installed. The default is now /usr/local.
3562 --openssldir shall be used exclusively to give the default
3563 location OPENSSLDIR where certificates, private keys, CRLs are
3564 managed. This is also where the default openssl.cnf gets
3566 If the directory given with this option is a relative path, the
3567 values of both the --prefix value and the --openssldir value will
3568 be combined to become OPENSSLDIR.
3569 The default for --openssldir is INSTALLTOP/ssl.
3571 Anyone who uses --openssldir to specify where OpenSSL is to be
3572 installed MUST change to use --prefix instead.
3576 * The GOST engine was out of date and therefore it has been removed. An up
3577 to date GOST engine is now being maintained in an external repository.
3578 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3579 support for GOST ciphersuites (these are only activated if a GOST engine
3584 * EGD is no longer supported by default; use enable-egd when
3587 *Ben Kaduk and Rich Salz*
3589 * The distribution now has Makefile.in files, which are used to
3590 create Makefile's when Configure is run. *Configure must be run
3591 before trying to build now.*
3595 * The return value for SSL_CIPHER_description() for error conditions
3600 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3602 Obtaining and performing DNSSEC validation of TLSA records is
3603 the application's responsibility. The application provides
3604 the TLSA records of its choice to OpenSSL, and these are then
3605 used to authenticate the peer.
3607 The TLSA records need not even come from DNS. They can, for
3608 example, be used to implement local end-entity certificate or
3609 trust-anchor "pinning", where the "pin" data takes the form
3610 of TLSA records, which can augment or replace verification
3611 based on the usual WebPKI public certification authorities.
3615 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3616 continues to support deprecated interfaces in default builds.
3617 However, applications are strongly advised to compile their
3618 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3619 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3620 or the 1.1.0 releases.
3622 In environments in which all applications have been ported to
3623 not use any deprecated interfaces OpenSSL's Configure script
3624 should be used with the --api=1.1.0 option to entirely remove
3625 support for the deprecated features from the library and
3626 unconditionally disable them in the installed headers.
3627 Essentially the same effect can be achieved with the "no-deprecated"
3628 argument to Configure, except that this will always restrict
3629 the build to just the latest API, rather than a fixed API
3632 As applications are ported to future revisions of the API,
3633 they should update their compile-time OPENSSL_API_COMPAT define
3634 accordingly, but in most cases should be able to continue to
3635 compile with later releases.
3637 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3638 0x10000000L and 0x00908000L, respectively. However those
3639 versions did not support the OPENSSL_API_COMPAT feature, and
3640 so applications are not typically tested for explicit support
3641 of just the undeprecated features of either release.
3645 * Add support for setting the minimum and maximum supported protocol.
3646 It can bet set via the SSL_set_min_proto_version() and
3647 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3648 MaxProtocol. It's recommended to use the new APIs to disable
3649 protocols instead of disabling individual protocols using
3650 SSL_set_options() or SSL_CONF's Protocol. This change also
3651 removes support for disabling TLS 1.2 in the OpenSSL TLS
3652 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3656 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3660 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3661 and integrates ECDSA and ECDH functionality into EC. Implementations can
3662 now redirect key generation and no longer need to convert to or from
3665 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3666 include the ec.h header file instead.
3670 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3671 ciphers who are no longer supported and drops support the ephemeral RSA key
3672 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3676 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3677 opaque. For HMAC_CTX, the following constructors and destructors
3680 HMAC_CTX *HMAC_CTX_new(void);
3681 void HMAC_CTX_free(HMAC_CTX *ctx);
3683 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3684 destroy such methods has been added. See EVP_MD_meth_new(3) and
3685 EVP_CIPHER_meth_new(3) for documentation.
3688 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3689 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3690 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3691 an already created structure.
3692 2) For consistency with the majority of our object creators and
3693 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3694 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3695 for deprecated builds.
3699 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3700 cryptographic operations to be performed asynchronously as long as an
3701 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3702 further details. Libssl has also had this capability integrated with the
3703 introduction of the new mode SSL_MODE_ASYNC and associated error
3704 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3705 pages. This work was developed in partnership with Intel Corp.
3709 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3710 always enabled now. If you want to disable the support you should
3711 exclude it using the list of supported ciphers. This also means that the
3712 "-no_ecdhe" option has been removed from s_server.
3716 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3717 SSL_{CTX_}set1_curves() which can set a list.
3721 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3722 curve you want to support using SSL_{CTX_}set1_curves().
3726 * State machine rewrite. The state machine code has been significantly
3727 refactored in order to remove much duplication of code and solve issues
3728 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3729 further details). This change does have some associated API changes.
3730 Notably the SSL_state() function has been removed and replaced by
3731 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3732 SSL_set_state() has been removed altogether. The previous handshake states
3733 defined in ssl.h and ssl3.h have also been removed.
3737 * All instances of the string "ssleay" in the public API were replaced
3738 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3739 Some error codes related to internal RSA_eay API's were renamed.
3743 * The demo files in crypto/threads were moved to demo/threads.
3747 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3750 *Matt Caswell, Rich Salz*
3752 * New ASN.1 embed macro.
3754 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3755 structure is not allocated: it is part of the parent. That is instead of
3763 This reduces memory fragmentation and make it impossible to accidentally
3764 set a mandatory field to NULL.
3766 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3767 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3768 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3773 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3777 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3778 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3779 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3780 DES and RC4 ciphersuites.
3784 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3785 This changes the decoding behaviour for some invalid messages,
3786 though the change is mostly in the more lenient direction, and
3787 legacy behaviour is preserved as much as possible.
3791 * Fix no-stdio build.
3792 *David Woodhouse <David.Woodhouse@intel.com> and also*
3793 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3795 * New testing framework
3796 The testing framework has been largely rewritten and is now using
3797 perl and the perl modules Test::Harness and an extended variant of
3798 Test::More called OpenSSL::Test to do its work. All test scripts in
3799 test/ have been rewritten into test recipes, and all direct calls to
3800 executables in test/Makefile have become individual recipes using the
3801 simplified testing OpenSSL::Test::Simple.
3803 For documentation on our testing modules, do:
3805 perldoc test/testlib/OpenSSL/Test/Simple.pm
3806 perldoc test/testlib/OpenSSL/Test.pm
3810 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3811 are used; the latter aborts on memory leaks (usually checked on exit).
3812 Some undocumented "set malloc, etc., hooks" functions were removed
3813 and others were changed. All are now documented.
3817 * In DSA_generate_parameters_ex, if the provided seed is too short,
3820 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3822 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3823 from RFC4279, RFC4785, RFC5487, RFC5489.
3825 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3826 original RSA_PSK patch.
3830 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3831 era flag was never set throughout the codebase (only read). Also removed
3832 SSL3_FLAGS_POP_BUFFER which was only used if
3833 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3837 * Changed the default name options in the "ca", "crl", "req" and "x509"
3838 to be "oneline" instead of "compat".
3842 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3843 not aware of clients that still exhibit this bug, and the workaround
3844 hasn't been working properly for a while.
3848 * The return type of BIO_number_read() and BIO_number_written() as well as
3849 the corresponding num_read and num_write members in the BIO structure has
3850 changed from unsigned long to uint64_t. On platforms where an unsigned
3851 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3856 * Given the pervasive nature of TLS extensions it is inadvisable to run
3857 OpenSSL without support for them. It also means that maintaining
3858 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3859 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3863 * Removed support for the two export grade static DH ciphersuites
3864 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3865 were newly added (along with a number of other static DH ciphersuites) to
3866 1.0.2. However the two export ones have *never* worked since they were
3867 introduced. It seems strange in any case to be adding new export
3868 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3872 * Version negotiation has been rewritten. In particular SSLv23_method(),
3873 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3874 and turned into macros which simply call the new preferred function names
3875 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3876 should use the new names instead. Also as part of this change the ssl23.h
3877 header file has been removed.
3881 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3882 code and the associated standard is no longer considered fit-for-purpose.
3886 * RT2547 was closed. When generating a private key, try to make the
3887 output file readable only by the owner. This behavior change might
3888 be noticeable when interacting with other software.
3890 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3895 * Added HTTP GET support to the ocsp command.
3899 * Changed default digest for the dgst and enc commands from MD5 to
3904 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3908 * Added support for TLS extended master secret from
3909 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3910 initial patch which was a great help during development.
3914 * All libssl internal structures have been removed from the public header
3915 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3916 now redundant). Users should not attempt to access internal structures
3917 directly. Instead they should use the provided API functions.
3921 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3922 Access to deprecated functions can be re-enabled by running config with
3923 "enable-deprecated". In addition applications wishing to use deprecated
3924 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3925 will, by default, disable some transitive includes that previously existed
3926 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3930 * Added support for OCB mode. OpenSSL has been granted a patent license
3931 compatible with the OpenSSL license for use of OCB. Details are available
3932 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3933 for OCB can be removed by calling config with no-ocb.
3937 * SSLv2 support has been removed. It still supports receiving a SSLv2
3938 compatible client hello.
3942 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3943 done while fixing the error code for the key-too-small case.
3945 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3947 * CA.sh has been removed; use CA.pl instead.
3951 * Removed old DES API.
3955 * Remove various unsupported platforms:
3961 Sinix/ReliantUNIX RM400
3966 16-bit platforms such as WIN16
3970 * Clean up OPENSSL_NO_xxx #define's
3971 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3972 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3973 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3974 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3975 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3976 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3977 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3978 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3979 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3980 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3984 * Cleaned up dead code
3985 Remove all but one '#ifdef undef' which is to be looked at.
3989 * Clean up calling of xxx_free routines.
3990 Just like free(), fix most of the xxx_free routines to accept
3991 NULL. Remove the non-null checks from callers. Save much code.
3995 * Add secure heap for storage of private keys (when possible).
3996 Add BIO_s_secmem(), CBIGNUM, etc.
3997 Contributed by Akamai Technologies under our Corporate CLA.
4001 * Experimental support for a new, fast, unbiased prime candidate generator,
4002 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4004 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4006 * New output format NSS in the sess_id command line tool. This allows
4007 exporting the session id and the master key in NSS keylog format.
4009 *Martin Kaiser <martin@kaiser.cx>*
4011 * Harmonize version and its documentation. -f flag is used to display
4014 *mancha <mancha1@zoho.com>*
4016 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4017 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4019 *mancha <mancha1@zoho.com>*
4021 * Fix some double frees. These are not thought to be exploitable.
4023 *mancha <mancha1@zoho.com>*
4025 * A missing bounds check in the handling of the TLS heartbeat extension
4026 can be used to reveal up to 64k of memory to a connected client or
4029 Thanks for Neel Mehta of Google Security for discovering this bug and to
4030 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4031 preparing the fix ([CVE-2014-0160])
4033 *Adam Langley, Bodo Moeller*
4035 * Fix for the attack described in the paper "Recovering OpenSSL
4036 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4037 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4038 <http://eprint.iacr.org/2014/140>
4040 Thanks to Yuval Yarom and Naomi Benger for discovering this
4041 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4043 *Yuval Yarom and Naomi Benger*
4045 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4046 this fixes a limitation in previous versions of OpenSSL.
4050 * Experimental encrypt-then-mac support.
4052 Experimental support for encrypt then mac from
4053 draft-gutmann-tls-encrypt-then-mac-02.txt
4055 To enable it set the appropriate extension number (0x42 for the test
4056 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4058 For non-compliant peers (i.e. just about everything) this should have no
4061 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4065 * Add EVP support for key wrapping algorithms, to avoid problems with
4066 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4067 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4068 algorithms and include tests cases.
4072 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4077 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4078 MGF1 digest and OAEP label.
4082 * Make openssl verify return errors.
4084 *Chris Palmer <palmer@google.com> and Ben Laurie*
4086 * New function ASN1_TIME_diff to calculate the difference between two
4087 ASN1_TIME structures or one structure and the current time.
4091 * Update fips_test_suite to support multiple command line options. New
4092 test to induce all self test errors in sequence and check expected
4097 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4098 sign or verify all in one operation.
4102 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4103 test programs and fips_test_suite. Includes functionality to parse
4104 the minimal script output of fipsalgest.pl directly.
4108 * Add authorisation parameter to FIPS_module_mode_set().
4112 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4116 * Use separate DRBG fields for internal and external flags. New function
4117 FIPS_drbg_health_check() to perform on demand health checking. Add
4118 generation tests to fips_test_suite with reduced health check interval to
4119 demonstrate periodic health checking. Add "nodh" option to
4120 fips_test_suite to skip very slow DH test.
4124 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4129 * More extensive health check for DRBG checking many more failure modes.
4130 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4131 combination: call this in fips_test_suite.
4135 * Add support for canonical generation of DSA parameter 'g'. See
4138 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4139 POST to handle HMAC cases.
4143 * Add functions FIPS_module_version() and FIPS_module_version_text()
4144 to return numerical and string versions of the FIPS module number.
4148 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4149 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4150 outside the validated module in the FIPS capable OpenSSL.
4154 * Minor change to DRBG entropy callback semantics. In some cases
4155 there is no multiple of the block length between min_len and
4156 max_len. Allow the callback to return more than max_len bytes
4157 of entropy but discard any extra: it is the callback's responsibility
4158 to ensure that the extra data discarded does not impact the
4159 requested amount of entropy.
4163 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4164 information in FIPS186-3, SP800-57 and SP800-131A.
4168 * CCM support via EVP. Interface is very similar to GCM case except we
4169 must supply all data in one chunk (i.e. no update, final) and the
4170 message length must be supplied if AAD is used. Add algorithm test
4175 * Initial version of POST overhaul. Add POST callback to allow the status
4176 of POST to be monitored and/or failures induced. Modify fips_test_suite
4177 to use callback. Always run all selftests even if one fails.
4181 * XTS support including algorithm test driver in the fips_gcmtest program.
4182 Note: this does increase the maximum key length from 32 to 64 bytes but
4183 there should be no binary compatibility issues as existing applications
4184 will never use XTS mode.
4188 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4189 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4190 performs algorithm blocking for unapproved PRNG types. Also do not
4191 set PRNG type in FIPS_mode_set(): leave this to the application.
4192 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4193 the standard OpenSSL PRNG: set additional data to a date time vector.
4197 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4198 This shouldn't present any incompatibility problems because applications
4199 shouldn't be using these directly and any that are will need to rethink
4200 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4204 * Extensive self tests and health checking required by SP800-90 DRBG.
4205 Remove strength parameter from FIPS_drbg_instantiate and always
4206 instantiate at maximum supported strength.
4210 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4214 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4218 * New function DH_compute_key_padded() to compute a DH key and pad with
4219 leading zeroes if needed: this complies with SP800-56A et al.
4223 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4224 anything, incomplete, subject to change and largely untested at present.
4228 * Modify fipscanisteronly build option to only build the necessary object
4229 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4233 * Add experimental option FIPSSYMS to give all symbols in
4234 fipscanister.o and FIPS or fips prefix. This will avoid
4235 conflicts with future versions of OpenSSL. Add perl script
4236 util/fipsas.pl to preprocess assembly language source files
4237 and rename any affected symbols.
4241 * Add selftest checks and algorithm block of non-fips algorithms in
4242 FIPS mode. Remove DES2 from selftests.
4246 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4247 return internal method without any ENGINE dependencies. Add new
4248 tiny fips sign and verify functions.
4252 * New build option no-ec2m to disable characteristic 2 code.
4256 * New build option "fipscanisteronly". This only builds fipscanister.o
4257 and (currently) associated fips utilities. Uses the file Makefile.fips
4258 instead of Makefile.org as the prototype.
4262 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4263 Update fips_gcmtest to use IV generator.
4267 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4268 setting output buffer to NULL. The `*Final` function must be
4269 called although it will not retrieve any additional data. The tag
4270 can be set or retrieved with a ctrl. The IV length is by default 12
4271 bytes (96 bits) but can be set to an alternative value. If the IV
4272 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4277 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4278 underlying do_cipher function handles all cipher semantics itself
4279 including padding and finalisation. This is useful if (for example)
4280 an ENGINE cipher handles block padding itself. The behaviour of
4281 do_cipher is subtly changed if this flag is set: the return value
4282 is the number of characters written to the output buffer (zero is
4283 no longer an error code) or a negative error code. Also if the
4284 input buffer is NULL and length 0 finalisation should be performed.
4288 * If a candidate issuer certificate is already part of the constructed
4289 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4293 * Improve forward-security support: add functions
4295 void SSL_CTX_set_not_resumable_session_callback(
4296 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4297 void SSL_set_not_resumable_session_callback(
4298 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4300 for use by SSL/TLS servers; the callback function will be called whenever a
4301 new session is created, and gets to decide whether the session may be
4302 cached to make it resumable (return 0) or not (return 1). (As by the
4303 SSL/TLS protocol specifications, the session_id sent by the server will be
4304 empty to indicate that the session is not resumable; also, the server will
4305 not generate RFC 4507 (RFC 5077) session tickets.)
4307 A simple reasonable callback implementation is to return is_forward_secure.
4308 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4309 by the SSL/TLS server library, indicating whether it can provide forward
4312 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4314 * New -verify_name option in command line utilities to set verification
4319 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4320 Add CMAC pkey methods.
4324 * Experimental renegotiation in s_server -www mode. If the client
4325 browses /reneg connection is renegotiated. If /renegcert it is
4326 renegotiated requesting a certificate.
4330 * Add an "external" session cache for debugging purposes to s_server. This
4331 should help trace issues which normally are only apparent in deployed
4332 multi-process servers.
4336 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4337 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4338 BIO_set_cipher() and some obscure PEM functions were changed so they
4339 can now return an error. The RAND changes required a change to the
4340 RAND_METHOD structure.
4344 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4345 a gcc attribute to warn if the result of a function is ignored. This
4346 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4347 whose return value is often ignored.
4351 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4352 These allow SCTs (signed certificate timestamps) to be requested and
4353 validated when establishing a connection.
4355 *Rob Percival <robpercival@google.com>*
4360 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4362 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4363 used even when parsing explicit parameters, when loading a encoded key
4364 or calling `EC_GROUP_new_from_ecpkparameters()`/
4365 `EC_GROUP_new_from_ecparameters()`.
4366 This prevents bypass of security hardening and performance gains,
4367 especially for curves with specialized EC_METHODs.
4368 By default, if a key encoded with explicit parameters is loaded and later
4369 encoded, the output is still encoded with explicit parameters, even if
4370 internally a "named" EC_GROUP is used for computation.
4374 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4375 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4376 NULL. After this change, only the cofactor parameter can be NULL. It also
4377 does some minimal sanity checks on the passed order.
4382 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4383 An attack is simple, if the first CMS_recipientInfo is valid but the
4384 second CMS_recipientInfo is chosen ciphertext. If the second
4385 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4386 encryption key will be replaced by garbage, and the message cannot be
4387 decoded, but if the RSA decryption fails, the correct encryption key is
4388 used and the recipient will not notice the attack.
4389 As a work around for this potential attack the length of the decrypted
4390 key must be equal to the cipher default key length, in case the
4391 certifiate is not given and all recipientInfo are tried out.
4392 The old behaviour can be re-enabled in the CMS code by setting the
4393 CMS_DEBUG_DECRYPT flag.
4398 * Document issue with installation paths in diverse Windows builds
4400 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4401 binaries and run-time config file.
4406 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4408 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4409 This changes the size when using the `genpkey` command when no size is given.
4410 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4411 generation commands to use 2048 bits by default.
4415 * Add FIPS support for Android Arm 64-bit
4417 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4418 Module in Version 2.0.10. For some reason, the corresponding target
4419 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4420 built with FIPS support on Android Arm 64-bit. This omission has been
4423 *Matthias St. Pierre*
4425 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4427 * 0-byte record padding oracle
4429 If an application encounters a fatal protocol error and then calls
4430 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4431 then OpenSSL can respond differently to the calling application if a 0 byte
4432 record is received with invalid padding compared to if a 0 byte record is
4433 received with an invalid MAC. If the application then behaves differently
4434 based on that in a way that is detectable to the remote peer, then this
4435 amounts to a padding oracle that could be used to decrypt data.
4437 In order for this to be exploitable "non-stitched" ciphersuites must be in
4438 use. Stitched ciphersuites are optimised implementations of certain
4439 commonly used ciphersuites. Also the application must call SSL_shutdown()
4440 twice even if a protocol error has occurred (applications should not do
4441 this but some do anyway).
4443 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4444 Aviram, with additional investigation by Steven Collison and Andrew
4445 Hourselt. It was reported to OpenSSL on 10th December 2018.
4450 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4454 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4456 * Microarchitecture timing vulnerability in ECC scalar multiplication
4458 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4459 shown to be vulnerable to a microarchitecture timing side channel attack.
4460 An attacker with sufficient access to mount local timing attacks during
4461 ECDSA signature generation could recover the private key.
4463 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4464 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4470 * Timing vulnerability in DSA signature generation
4472 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4473 timing side channel attack. An attacker could use variations in the signing
4474 algorithm to recover the private key.
4476 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4481 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4482 Module, accidentally introduced while backporting security fixes from the
4483 development branch and hindering the use of ECC in FIPS mode.
4487 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4489 * Client DoS due to large DH parameter
4491 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4492 malicious server can send a very large prime value to the client. This will
4493 cause the client to spend an unreasonably long period of time generating a
4494 key for this prime resulting in a hang until the client has finished. This
4495 could be exploited in a Denial Of Service attack.
4497 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4502 * Cache timing vulnerability in RSA Key Generation
4504 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4505 a cache timing side channel attack. An attacker with sufficient access to
4506 mount cache timing attacks during the RSA key generation process could
4507 recover the private key.
4509 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4510 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4515 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4516 parameter is no longer accepted, as it leads to a corrupt table. NULL
4517 pem_str is reserved for alias entries only.
4521 * Revert blinding in ECDSA sign and instead make problematic addition
4522 length-invariant. Switch even to fixed-length Montgomery multiplication.
4526 * Change generating and checking of primes so that the error rate of not
4527 being prime depends on the intended use based on the size of the input.
4528 For larger primes this will result in more rounds of Miller-Rabin.
4529 The maximal error rate for primes with more than 1080 bits is lowered
4532 *Kurt Roeckx, Annie Yousar*
4534 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4538 * Add blinding to ECDSA and DSA signatures to protect against side channel
4539 attacks discovered by Keegan Ryan (NCC Group).
4543 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4544 now allow empty (zero character) pass phrases.
4548 * Certificate time validation (X509_cmp_time) enforces stricter
4549 compliance with RFC 5280. Fractional seconds and timezone offsets
4550 are no longer allowed.
4554 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4556 * Constructed ASN.1 types with a recursive definition could exceed the stack
4558 Constructed ASN.1 types with a recursive definition (such as can be found
4559 in PKCS7) could eventually exceed the stack given malicious input with
4560 excessive recursion. This could result in a Denial Of Service attack. There
4561 are no such structures used within SSL/TLS that come from untrusted sources
4562 so this is considered safe.
4564 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4570 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4572 * Read/write after SSL object in error state
4574 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4575 mechanism. The intent was that if a fatal error occurred during a handshake
4576 then OpenSSL would move into the error state and would immediately fail if
4577 you attempted to continue the handshake. This works as designed for the
4578 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4579 SSL_connect()), however due to a bug it does not work correctly if
4580 SSL_read() or SSL_write() is called directly. In that scenario, if the
4581 handshake fails then a fatal error will be returned in the initial function
4582 call. If SSL_read()/SSL_write() is subsequently called by the application
4583 for the same SSL object then it will succeed and the data is passed without
4584 being decrypted/encrypted directly from the SSL/TLS record layer.
4586 In order to exploit this issue an application bug would have to be present
4587 that resulted in a call to SSL_read()/SSL_write() being issued after having
4588 already received a fatal error.
4590 This issue was reported to OpenSSL by David Benjamin (Google).
4595 * rsaz_1024_mul_avx2 overflow bug on x86_64
4597 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4598 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4599 Analysis suggests that attacks against RSA and DSA as a result of this
4600 defect would be very difficult to perform and are not believed likely.
4601 Attacks against DH1024 are considered just feasible, because most of the
4602 work necessary to deduce information about a private key may be performed
4603 offline. The amount of resources required for such an attack would be
4604 significant. However, for an attack on TLS to be meaningful, the server
4605 would have to share the DH1024 private key among multiple clients, which is
4606 no longer an option since CVE-2016-0701.
4608 This only affects processors that support the AVX2 but not ADX extensions
4609 like Intel Haswell (4th generation).
4611 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4612 was originally found via the OSS-Fuzz project.
4617 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4619 * bn_sqrx8x_internal carry bug on x86_64
4621 There is a carry propagating bug in the x86_64 Montgomery squaring
4622 procedure. No EC algorithms are affected. Analysis suggests that attacks
4623 against RSA and DSA as a result of this defect would be very difficult to
4624 perform and are not believed likely. Attacks against DH are considered just
4625 feasible (although very difficult) because most of the work necessary to
4626 deduce information about a private key may be performed offline. The amount
4627 of resources required for such an attack would be very significant and
4628 likely only accessible to a limited number of attackers. An attacker would
4629 additionally need online access to an unpatched system using the target
4630 private key in a scenario with persistent DH parameters and a private
4631 key that is shared between multiple clients.
4633 This only affects processors that support the BMI1, BMI2 and ADX extensions
4634 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4636 This issue was reported to OpenSSL by the OSS-Fuzz project.
4641 * Malformed X.509 IPAddressFamily could cause OOB read
4643 If an X.509 certificate has a malformed IPAddressFamily extension,
4644 OpenSSL could do a one-byte buffer overread. The most likely result
4645 would be an erroneous display of the certificate in text format.
4647 This issue was reported to OpenSSL by the OSS-Fuzz project.
4652 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4654 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4655 platform rather than 'mingw'.
4659 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4661 * Truncated packet could crash via OOB read
4663 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4664 cipher is being used, then a truncated packet can cause that host to
4665 perform an out-of-bounds read, usually resulting in a crash.
4667 This issue was reported to OpenSSL by Robert Święcki of Google.
4672 * BN_mod_exp may produce incorrect results on x86_64
4674 There is a carry propagating bug in the x86_64 Montgomery squaring
4675 procedure. No EC algorithms are affected. Analysis suggests that attacks
4676 against RSA and DSA as a result of this defect would be very difficult to
4677 perform and are not believed likely. Attacks against DH are considered just
4678 feasible (although very difficult) because most of the work necessary to
4679 deduce information about a private key may be performed offline. The amount
4680 of resources required for such an attack would be very significant and
4681 likely only accessible to a limited number of attackers. An attacker would
4682 additionally need online access to an unpatched system using the target
4683 private key in a scenario with persistent DH parameters and a private
4684 key that is shared between multiple clients. For example this can occur by
4685 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4686 similar to CVE-2015-3193 but must be treated as a separate problem.
4688 This issue was reported to OpenSSL by the OSS-Fuzz project.
4693 * Montgomery multiplication may produce incorrect results
4695 There is a carry propagating bug in the Broadwell-specific Montgomery
4696 multiplication procedure that handles input lengths divisible by, but
4697 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4698 and DH private keys are impossible. This is because the subroutine in
4699 question is not used in operations with the private key itself and an input
4700 of the attacker's direct choice. Otherwise the bug can manifest itself as
4701 transient authentication and key negotiation failures or reproducible
4702 erroneous outcome of public-key operations with specially crafted input.
4703 Among EC algorithms only Brainpool P-512 curves are affected and one
4704 presumably can attack ECDH key negotiation. Impact was not analyzed in
4705 detail, because pre-requisites for attack are considered unlikely. Namely
4706 multiple clients have to choose the curve in question and the server has to
4707 share the private key among them, neither of which is default behaviour.
4708 Even then only clients that chose the curve will be affected.
4710 This issue was publicly reported as transient failures and was not
4711 initially recognized as a security issue. Thanks to Richard Morgan for
4712 providing reproducible case.
4717 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4718 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4719 prevent issues where no progress is being made and the peer continually
4720 sends unrecognised record types, using up resources processing them.
4724 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4726 * Missing CRL sanity check
4728 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4729 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4730 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4732 This issue only affects the OpenSSL 1.0.2i
4737 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4739 * OCSP Status Request extension unbounded memory growth
4741 A malicious client can send an excessively large OCSP Status Request
4742 extension. If that client continually requests renegotiation, sending a
4743 large OCSP Status Request extension each time, then there will be unbounded
4744 memory growth on the server. This will eventually lead to a Denial Of
4745 Service attack through memory exhaustion. Servers with a default
4746 configuration are vulnerable even if they do not support OCSP. Builds using
4747 the "no-ocsp" build time option are not affected.
4749 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4754 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4757 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4763 * OOB write in MDC2_Update()
4765 An overflow can occur in MDC2_Update() either if called directly or
4766 through the EVP_DigestUpdate() function using MDC2. If an attacker
4767 is able to supply very large amounts of input data after a previous
4768 call to EVP_EncryptUpdate() with a partial block then a length check
4769 can overflow resulting in a heap corruption.
4771 The amount of data needed is comparable to SIZE_MAX which is impractical
4774 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4779 * Malformed SHA512 ticket DoS
4781 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4782 DoS attack where a malformed ticket will result in an OOB read which will
4785 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4786 a custom server callback and ticket lookup mechanism.
4788 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4793 * OOB write in BN_bn2dec()
4795 The function BN_bn2dec() does not check the return value of BN_div_word().
4796 This can cause an OOB write if an application uses this function with an
4797 overly large BIGNUM. This could be a problem if an overly large certificate
4798 or CRL is printed out from an untrusted source. TLS is not affected because
4799 record limits will reject an oversized certificate before it is parsed.
4801 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4806 * OOB read in TS_OBJ_print_bio()
4808 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4809 the total length the OID text representation would use and not the amount
4810 of data written. This will result in OOB reads when large OIDs are
4813 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4818 * Pointer arithmetic undefined behaviour
4820 Avoid some undefined pointer arithmetic
4822 A common idiom in the codebase is to check limits in the following manner:
4825 Where "p" points to some malloc'd data of SIZE bytes and
4828 "len" here could be from some externally supplied data (e.g. from a TLS
4831 The rules of C pointer arithmetic are such that "p + len" is only well
4832 defined where len <= SIZE. Therefore the above idiom is actually
4833 undefined behaviour.
4835 For example this could cause problems if some malloc implementation
4836 provides an address for "p" such that "p + len" actually overflows for
4837 values of len that are too big and therefore p + len < limit.
4839 This issue was reported to OpenSSL by Guido Vranken
4844 * Constant time flag not preserved in DSA signing
4846 Operations in the DSA signing algorithm should run in constant time in
4847 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4848 implementation means that a non-constant time codepath is followed for
4849 certain operations. This has been demonstrated through a cache-timing
4850 attack to be sufficient for an attacker to recover the private DSA key.
4852 This issue was reported by César Pereida (Aalto University), Billy Brumley
4853 (Tampere University of Technology), and Yuval Yarom (The University of
4854 Adelaide and NICTA).
4859 * DTLS buffered message DoS
4861 In a DTLS connection where handshake messages are delivered out-of-order
4862 those messages that OpenSSL is not yet ready to process will be buffered
4863 for later use. Under certain circumstances, a flaw in the logic means that
4864 those messages do not get removed from the buffer even though the handshake
4865 has been completed. An attacker could force up to approx. 15 messages to
4866 remain in the buffer when they are no longer required. These messages will
4867 be cleared when the DTLS connection is closed. The default maximum size for
4868 a message is 100k. Therefore the attacker could force an additional 1500k
4869 to be consumed per connection. By opening many simulataneous connections an
4870 attacker could cause a DoS attack through memory exhaustion.
4872 This issue was reported to OpenSSL by Quan Luo.
4877 * DTLS replay protection DoS
4879 A flaw in the DTLS replay attack protection mechanism means that records
4880 that arrive for future epochs update the replay protection "window" before
4881 the MAC for the record has been validated. This could be exploited by an
4882 attacker by sending a record for the next epoch (which does not have to
4883 decrypt or have a valid MAC), with a very large sequence number. This means
4884 that all subsequent legitimate packets are dropped causing a denial of
4885 service for a specific DTLS connection.
4887 This issue was reported to OpenSSL by the OCAP audit team.
4892 * Certificate message OOB reads
4894 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4895 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4896 theoretical DoS risk but this has not been observed in practice on common
4899 The messages affected are client certificate, client certificate request
4900 and server certificate. As a result the attack can only be performed
4901 against a client or a server which enables client authentication.
4903 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4908 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4910 * Prevent padding oracle in AES-NI CBC MAC check
4912 A MITM attacker can use a padding oracle attack to decrypt traffic
4913 when the connection uses an AES CBC cipher and the server support
4916 This issue was introduced as part of the fix for Lucky 13 padding
4917 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4918 constant time by making sure that always the same bytes are read and
4919 compared against either the MAC or padding bytes. But it no longer
4920 checked that there was enough data to have both the MAC and padding
4923 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4928 * Fix EVP_EncodeUpdate overflow
4930 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4931 Base64 encoding of binary data. If an attacker is able to supply very large
4932 amounts of input data then a length check can overflow resulting in a heap
4935 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4936 the `PEM_write_bio*` family of functions. These are mainly used within the
4937 OpenSSL command line applications, so any application which processes data
4938 from an untrusted source and outputs it as a PEM file should be considered
4939 vulnerable to this issue. User applications that call these APIs directly
4940 with large amounts of untrusted data may also be vulnerable.
4942 This issue was reported by Guido Vranken.
4947 * Fix EVP_EncryptUpdate overflow
4949 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4950 is able to supply very large amounts of input data after a previous call to
4951 EVP_EncryptUpdate() with a partial block then a length check can overflow
4952 resulting in a heap corruption. Following an analysis of all OpenSSL
4953 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4954 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4955 the first called function after an EVP_EncryptInit(), and therefore that
4956 specific call must be safe. The second form is where the length passed to
4957 EVP_EncryptUpdate() can be seen from the code to be some small value and
4958 therefore there is no possibility of an overflow. Since all instances are
4959 one of these two forms, it is believed that there can be no overflows in
4960 internal code due to this problem. It should be noted that
4961 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4962 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4963 of these calls have also been analysed too and it is believed there are no
4964 instances in internal usage where an overflow could occur.
4966 This issue was reported by Guido Vranken.
4971 * Prevent ASN.1 BIO excessive memory allocation
4973 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4974 a short invalid encoding can cause allocation of large amounts of memory
4975 potentially consuming excessive resources or exhausting memory.
4977 Any application parsing untrusted data through d2i BIO functions is
4978 affected. The memory based functions such as d2i_X509() are *not* affected.
4979 Since the memory based functions are used by the TLS library, TLS
4980 applications are not affected.
4982 This issue was reported by Brian Carpenter.
4989 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4990 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4991 in arbitrary stack data being returned in the buffer.
4993 This issue was reported by Guido Vranken.
4998 * Modify behavior of ALPN to invoke callback after SNI/servername
4999 callback, such that updates to the SSL_CTX affect ALPN.
5003 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5008 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5009 methods are enabled and ssl2 is disabled the methods return NULL.
5013 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5015 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5016 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5017 provide any "EXPORT" or "LOW" strength ciphers.
5021 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5022 is by default disabled at build-time. Builds that are not configured with
5023 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5024 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5025 will need to explicitly call either of:
5027 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5029 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5031 as appropriate. Even if either of those is used, or the application
5032 explicitly uses the version-specific SSLv2_method() or its client and
5033 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5034 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5035 ciphers, and SSLv2 56-bit DES are no longer available.
5040 * Fix a double-free in DSA code
5042 A double free bug was discovered when OpenSSL parses malformed DSA private
5043 keys and could lead to a DoS attack or memory corruption for applications
5044 that receive DSA private keys from untrusted sources. This scenario is
5047 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5053 * Disable SRP fake user seed to address a server memory leak.
5055 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5057 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5058 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5059 was changed to ignore the "fake user" SRP seed, even if the seed
5062 Users should use SRP_VBASE_get1_by_user instead. Note that in
5063 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5064 also that even though configuring the SRP seed attempts to hide
5065 invalid usernames by continuing the handshake with fake
5066 credentials, this behaviour is not constant time and no strong
5067 guarantees are made that the handshake is indistinguishable from
5068 that of a valid user.
5073 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5075 In the BN_hex2bn function the number of hex digits is calculated using an
5076 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5077 large values of `i` this can result in `bn_expand` not allocating any
5078 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5079 field as NULL leading to a subsequent NULL ptr deref. For very large values
5080 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5081 In this case memory is allocated to the internal BIGNUM data field, but it
5082 is insufficiently sized leading to heap corruption. A similar issue exists
5083 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5084 is ever called by user applications with very large untrusted hex/dec data.
5085 This is anticipated to be a rare occurrence.
5087 All OpenSSL internal usage of these functions use data that is not expected
5088 to be untrusted, e.g. config file data or application command line
5089 arguments. If user developed applications generate config file data based
5090 on untrusted data then it is possible that this could also lead to security
5091 consequences. This is also anticipated to be rare.
5093 This issue was reported to OpenSSL by Guido Vranken.
5098 * Fix memory issues in `BIO_*printf` functions
5100 The internal `fmtstr` function used in processing a "%s" format string in
5101 the `BIO_*printf` functions could overflow while calculating the length of a
5102 string and cause an OOB read when printing very long strings.
5104 Additionally the internal `doapr_outch` function can attempt to write to an
5105 OOB memory location (at an offset from the NULL pointer) in the event of a
5106 memory allocation failure. In 1.0.2 and below this could be caused where
5107 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5108 could be in processing a very long "%s" format string. Memory leaks can
5111 The first issue may mask the second issue dependent on compiler behaviour.
5112 These problems could enable attacks where large amounts of untrusted data
5113 is passed to the `BIO_*printf` functions. If applications use these functions
5114 in this way then they could be vulnerable. OpenSSL itself uses these
5115 functions when printing out human-readable dumps of ASN.1 data. Therefore
5116 applications that print this data could be vulnerable if the data is from
5117 untrusted sources. OpenSSL command line applications could also be
5118 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5119 as command line arguments.
5121 Libssl is not considered directly vulnerable. Additionally certificates etc
5122 received via remote connections via libssl are also unlikely to be able to
5123 trigger these issues because of message size limits enforced within libssl.
5125 This issue was reported to OpenSSL Guido Vranken.
5130 * Side channel attack on modular exponentiation
5132 A side-channel attack was found which makes use of cache-bank conflicts on
5133 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5134 of RSA keys. The ability to exploit this issue is limited as it relies on
5135 an attacker who has control of code in a thread running on the same
5136 hyper-threaded core as the victim thread which is performing decryptions.
5138 This issue was reported to OpenSSL by Yuval Yarom, The University of
5139 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5140 Nadia Heninger, University of Pennsylvania with more information at
5141 <http://cachebleed.info>.
5146 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5147 if no keysize is specified with default_bits. This fixes an
5148 omission in an earlier change that changed all RSA/DSA key generation
5149 commands to use 2048 bits by default.
5153 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5155 * DH small subgroups
5157 Historically OpenSSL only ever generated DH parameters based on "safe"
5158 primes. More recently (in version 1.0.2) support was provided for
5159 generating X9.42 style parameter files such as those required for RFC 5114
5160 support. The primes used in such files may not be "safe". Where an
5161 application is using DH configured with parameters based on primes that are
5162 not "safe" then an attacker could use this fact to find a peer's private
5163 DH exponent. This attack requires that the attacker complete multiple
5164 handshakes in which the peer uses the same private DH exponent. For example
5165 this could be used to discover a TLS server's private DH exponent if it's
5166 reusing the private DH exponent or it's using a static DH ciphersuite.
5168 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5169 TLS. It is not on by default. If the option is not set then the server
5170 reuses the same private DH exponent for the life of the server process and
5171 would be vulnerable to this attack. It is believed that many popular
5172 applications do set this option and would therefore not be at risk.
5174 The fix for this issue adds an additional check where a "q" parameter is
5175 available (as is the case in X9.42 based parameters). This detects the
5176 only known attack, and is the only possible defense for static DH
5177 ciphersuites. This could have some performance impact.
5179 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5180 default and cannot be disabled. This could have some performance impact.
5182 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5187 * SSLv2 doesn't block disabled ciphers
5189 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5190 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5191 been disabled, provided that the SSLv2 protocol was not also disabled via
5194 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5195 and Sebastian Schinzel.
5200 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5202 * BN_mod_exp may produce incorrect results on x86_64
5204 There is a carry propagating bug in the x86_64 Montgomery squaring
5205 procedure. No EC algorithms are affected. Analysis suggests that attacks
5206 against RSA and DSA as a result of this defect would be very difficult to
5207 perform and are not believed likely. Attacks against DH are considered just
5208 feasible (although very difficult) because most of the work necessary to
5209 deduce information about a private key may be performed offline. The amount
5210 of resources required for such an attack would be very significant and
5211 likely only accessible to a limited number of attackers. An attacker would
5212 additionally need online access to an unpatched system using the target
5213 private key in a scenario with persistent DH parameters and a private
5214 key that is shared between multiple clients. For example this can occur by
5215 default in OpenSSL DHE based SSL/TLS ciphersuites.
5217 This issue was reported to OpenSSL by Hanno Böck.
5222 * Certificate verify crash with missing PSS parameter
5224 The signature verification routines will crash with a NULL pointer
5225 dereference if presented with an ASN.1 signature using the RSA PSS
5226 algorithm and absent mask generation function parameter. Since these
5227 routines are used to verify certificate signature algorithms this can be
5228 used to crash any certificate verification operation and exploited in a
5229 DoS attack. Any application which performs certificate verification is
5230 vulnerable including OpenSSL clients and servers which enable client
5233 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5238 * X509_ATTRIBUTE memory leak
5240 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5241 memory. This structure is used by the PKCS#7 and CMS routines so any
5242 application which reads PKCS#7 or CMS data from untrusted sources is
5243 affected. SSL/TLS is not affected.
5245 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5251 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5252 This changes the decoding behaviour for some invalid messages,
5253 though the change is mostly in the more lenient direction, and
5254 legacy behaviour is preserved as much as possible.
5258 * In DSA_generate_parameters_ex, if the provided seed is too short,
5261 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5263 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5265 * Alternate chains certificate forgery
5267 During certificate verification, OpenSSL will attempt to find an
5268 alternative certificate chain if the first attempt to build such a chain
5269 fails. An error in the implementation of this logic can mean that an
5270 attacker could cause certain checks on untrusted certificates to be
5271 bypassed, such as the CA flag, enabling them to use a valid leaf
5272 certificate to act as a CA and "issue" an invalid certificate.
5274 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5279 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5281 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5282 incompatibility in the handling of HMAC. The previous ABI has now been
5287 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5289 * Malformed ECParameters causes infinite loop
5291 When processing an ECParameters structure OpenSSL enters an infinite loop
5292 if the curve specified is over a specially malformed binary polynomial
5295 This can be used to perform denial of service against any
5296 system which processes public keys, certificate requests or
5297 certificates. This includes TLS clients and TLS servers with
5298 client authentication enabled.
5300 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5305 * Exploitable out-of-bounds read in X509_cmp_time
5307 X509_cmp_time does not properly check the length of the ASN1_TIME
5308 string and can read a few bytes out of bounds. In addition,
5309 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5312 An attacker can use this to craft malformed certificates and CRLs of
5313 various sizes and potentially cause a segmentation fault, resulting in
5314 a DoS on applications that verify certificates or CRLs. TLS clients
5315 that verify CRLs are affected. TLS clients and servers with client
5316 authentication enabled may be affected if they use custom verification
5319 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5320 independently by Hanno Böck.
5325 * PKCS7 crash with missing EnvelopedContent
5327 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5328 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5329 with missing content and trigger a NULL pointer dereference on parsing.
5331 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5332 structures from untrusted sources are affected. OpenSSL clients and
5333 servers are not affected.
5335 This issue was reported to OpenSSL by Michal Zalewski (Google).
5340 * CMS verify infinite loop with unknown hash function
5342 When verifying a signedData message the CMS code can enter an infinite loop
5343 if presented with an unknown hash function OID. This can be used to perform
5344 denial of service against any system which verifies signedData messages using
5346 This issue was reported to OpenSSL by Johannes Bauer.
5351 * Race condition handling NewSessionTicket
5353 If a NewSessionTicket is received by a multi-threaded client when attempting to
5354 reuse a previous ticket then a race condition can occur potentially leading to
5355 a double free of the ticket data.
5360 * Only support 256-bit or stronger elliptic curves with the
5361 'ecdh_auto' setting (server) or by default (client). Of supported
5362 curves, prefer P-256 (both).
5366 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5368 * ClientHello sigalgs DoS fix
5370 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5371 invalid signature algorithms extension a NULL pointer dereference will
5372 occur. This can be exploited in a DoS attack against the server.
5374 This issue was was reported to OpenSSL by David Ramos of Stanford
5378 *Stephen Henson and Matt Caswell*
5380 * Multiblock corrupted pointer fix
5382 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5383 feature only applies on 64 bit x86 architecture platforms that support AES
5384 NI instructions. A defect in the implementation of "multiblock" can cause
5385 OpenSSL's internal write buffer to become incorrectly set to NULL when
5386 using non-blocking IO. Typically, when the user application is using a
5387 socket BIO for writing, this will only result in a failed connection.
5388 However if some other BIO is used then it is likely that a segmentation
5389 fault will be triggered, thus enabling a potential DoS attack.
5391 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5396 * Segmentation fault in DTLSv1_listen fix
5398 The DTLSv1_listen function is intended to be stateless and processes the
5399 initial ClientHello from many peers. It is common for user code to loop
5400 over the call to DTLSv1_listen until a valid ClientHello is received with
5401 an associated cookie. A defect in the implementation of DTLSv1_listen means
5402 that state is preserved in the SSL object from one invocation to the next
5403 that can lead to a segmentation fault. Errors processing the initial
5404 ClientHello can trigger this scenario. An example of such an error could be
5405 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5408 This issue was reported to OpenSSL by Per Allansson.
5413 * Segmentation fault in ASN1_TYPE_cmp fix
5415 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5416 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5417 certificate signature algorithm consistency this can be used to crash any
5418 certificate verification operation and exploited in a DoS attack. Any
5419 application which performs certificate verification is vulnerable including
5420 OpenSSL clients and servers which enable client authentication.
5425 * Segmentation fault for invalid PSS parameters fix
5427 The signature verification routines will crash with a NULL pointer
5428 dereference if presented with an ASN.1 signature using the RSA PSS
5429 algorithm and invalid parameters. Since these routines are used to verify
5430 certificate signature algorithms this can be used to crash any
5431 certificate verification operation and exploited in a DoS attack. Any
5432 application which performs certificate verification is vulnerable including
5433 OpenSSL clients and servers which enable client authentication.
5435 This issue was was reported to OpenSSL by Brian Carpenter.
5440 * ASN.1 structure reuse memory corruption fix
5442 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5443 memory corruption via an invalid write. Such reuse is and has been
5444 strongly discouraged and is believed to be rare.
5446 Applications that parse structures containing CHOICE or ANY DEFINED BY
5447 components may be affected. Certificate parsing (d2i_X509 and related
5448 functions) are however not affected. OpenSSL clients and servers are
5454 * PKCS7 NULL pointer dereferences fix
5456 The PKCS#7 parsing code does not handle missing outer ContentInfo
5457 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5458 missing content and trigger a NULL pointer dereference on parsing.
5460 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5461 otherwise parse PKCS#7 structures from untrusted sources are
5462 affected. OpenSSL clients and servers are not affected.
5464 This issue was reported to OpenSSL by Michal Zalewski (Google).
5469 * DoS via reachable assert in SSLv2 servers fix
5471 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5472 servers that both support SSLv2 and enable export cipher suites by sending
5473 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5475 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5476 (OpenSSL development team).
5481 * Empty CKE with client auth and DHE fix
5483 If client auth is used then a server can seg fault in the event of a DHE
5484 ciphersuite being selected and a zero length ClientKeyExchange message
5485 being sent by the client. This could be exploited in a DoS attack.
5490 * Handshake with unseeded PRNG fix
5492 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5493 with an unseeded PRNG. The conditions are:
5494 - The client is on a platform where the PRNG has not been seeded
5495 automatically, and the user has not seeded manually
5496 - A protocol specific client method version has been used (i.e. not
5497 SSL_client_methodv23)
5498 - A ciphersuite is used that does not require additional random data from
5499 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5501 If the handshake succeeds then the client random that has been used will
5502 have been generated from a PRNG with insufficient entropy and therefore the
5503 output may be predictable.
5505 For example using the following command with an unseeded openssl will
5506 succeed on an unpatched platform:
5508 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5513 * Use After Free following d2i_ECPrivatekey error fix
5515 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5516 could cause a use after free condition. This, in turn, could cause a double
5517 free in several private key parsing functions (such as d2i_PrivateKey
5518 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5519 for applications that receive EC private keys from untrusted
5520 sources. This scenario is considered rare.
5522 This issue was discovered by the BoringSSL project and fixed in their
5528 * X509_to_X509_REQ NULL pointer deref fix
5530 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5531 the certificate key is invalid. This function is rarely used in practice.
5533 This issue was discovered by Brian Carpenter.
5538 * Removed the export ciphers from the DEFAULT ciphers
5542 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5544 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5545 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5546 So far those who have to target multiple platforms would compromise
5547 and argue that binary targeting say ARMv5 would still execute on
5548 ARMv8. "Universal" build resolves this compromise by providing
5549 near-optimal performance even on newer platforms.
5553 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5554 (other platforms pending).
5556 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5558 * Add support for the SignedCertificateTimestampList certificate and
5559 OCSP response extensions from RFC6962.
5563 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5564 for corner cases. (Certain input points at infinity could lead to
5565 bogus results, with non-infinity inputs mapped to infinity too.)
5569 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5570 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5571 common cases are optimized and there still is room for further
5572 improvements. Vector Permutation AES for Altivec is also added.
5576 * Add support for little-endian ppc64 Linux target.
5578 *Marcelo Cerri (IBM)*
5580 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5581 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5582 are optimized and there still is room for further improvements.
5583 Both 32- and 64-bit modes are supported.
5585 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5587 * Improved ARMv7 NEON support.
5591 * Support for SPARC Architecture 2011 crypto extensions, first
5592 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5593 SHA256/512, MD5, GHASH and modular exponentiation.
5595 *Andy Polyakov, David Miller*
5597 * Accelerated modular exponentiation for Intel processors, a.k.a.
5600 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5602 * Support for new and upcoming Intel processors, including AVX2,
5603 BMI and SHA ISA extensions. This includes additional "stitched"
5604 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5607 This work was sponsored by Intel Corp.
5611 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5612 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5613 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5617 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5618 this fixes a limitation in previous versions of OpenSSL.
5622 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5623 MGF1 digest and OAEP label.
5627 * Add EVP support for key wrapping algorithms, to avoid problems with
5628 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5629 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5630 algorithms and include tests cases.
5634 * Add functions to allocate and set the fields of an ECDSA_METHOD
5637 *Douglas E. Engert, Steve Henson*
5639 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5640 difference in days and seconds between two tm or ASN1_TIME structures.
5644 * Add -rev test option to s_server to just reverse order of characters
5645 received by client and send back to server. Also prints an abbreviated
5646 summary of the connection parameters.
5650 * New option -brief for s_client and s_server to print out a brief summary
5651 of connection parameters.
5655 * Add callbacks for arbitrary TLS extensions.
5657 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5659 * New option -crl_download in several openssl utilities to download CRLs
5660 from CRLDP extension in certificates.
5664 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5668 * New function X509_CRL_diff to generate a delta CRL from the difference
5669 of two full CRLs. Add support to "crl" utility.
5673 * New functions to set lookup_crls function and to retrieve
5674 X509_STORE from X509_STORE_CTX.
5678 * Print out deprecated issuer and subject unique ID fields in
5683 * Extend OCSP I/O functions so they can be used for simple general purpose
5684 HTTP as well as OCSP. New wrapper function which can be used to download
5685 CRLs using the OCSP API.
5689 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5693 * `SSL_CONF*` functions. These provide a common framework for application
5694 configuration using configuration files or command lines.
5698 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5699 message callback and prints the results. Needs compile time option
5700 "enable-ssl-trace". New options to s_client and s_server to enable
5705 * New ctrl and macro to retrieve supported points extensions.
5706 Print out extension in s_server and s_client.
5710 * New functions to retrieve certificate signature and signature
5715 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5720 * New Suite B modes for TLS code. These use and enforce the requirements
5721 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5722 only use Suite B curves. The Suite B modes can be set by using the
5723 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5727 * New chain verification flags for Suite B levels of security. Check
5728 algorithms are acceptable when flags are set in X509_verify_cert.
5732 * Make tls1_check_chain return a set of flags indicating checks passed
5733 by a certificate chain. Add additional tests to handle client
5734 certificates: checks for matching certificate type and issuer name
5739 * If an attempt is made to use a signature algorithm not in the peer
5740 preference list abort the handshake. If client has no suitable
5741 signature algorithms in response to a certificate request do not
5742 use the certificate.
5746 * If server EC tmp key is not in client preference list abort handshake.
5750 * Add support for certificate stores in CERT structure. This makes it
5751 possible to have different stores per SSL structure or one store in
5752 the parent SSL_CTX. Include distinct stores for certificate chain
5753 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5754 to build and store a certificate chain in CERT structure: returning
5755 an error if the chain cannot be built: this will allow applications
5756 to test if a chain is correctly configured.
5758 Note: if the CERT based stores are not set then the parent SSL_CTX
5759 store is used to retain compatibility with existing behaviour.
5763 * New function ssl_set_client_disabled to set a ciphersuite disabled
5764 mask based on the current session, check mask when sending client
5765 hello and checking the requested ciphersuite.
5769 * New ctrls to retrieve and set certificate types in a certificate
5770 request message. Print out received values in s_client. If certificate
5771 types is not set with custom values set sensible values based on
5772 supported signature algorithms.
5776 * Support for distinct client and server supported signature algorithms.
5780 * Add certificate callback. If set this is called whenever a certificate
5781 is required by client or server. An application can decide which
5782 certificate chain to present based on arbitrary criteria: for example
5783 supported signature algorithms. Add very simple example to s_server.
5784 This fixes many of the problems and restrictions of the existing client
5785 certificate callback: for example you can now clear an existing
5786 certificate and specify the whole chain.
5790 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5791 the certificate can be used for (if anything). Set valid_flags field
5792 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5793 to have similar checks in it.
5795 Add new "cert_flags" field to CERT structure and include a "strict mode".
5796 This enforces some TLS certificate requirements (such as only permitting
5797 certificate signature algorithms contained in the supported algorithms
5798 extension) which some implementations ignore: this option should be used
5799 with caution as it could cause interoperability issues.
5803 * Update and tidy signature algorithm extension processing. Work out
5804 shared signature algorithms based on preferences and peer algorithms
5805 and print them out in s_client and s_server. Abort handshake if no
5806 shared signature algorithms.
5810 * Add new functions to allow customised supported signature algorithms
5811 for SSL and SSL_CTX structures. Add options to s_client and s_server
5816 * New function SSL_certs_clear() to delete all references to certificates
5817 from an SSL structure. Before this once a certificate had been added
5818 it couldn't be removed.
5822 * Integrate hostname, email address and IP address checking with certificate
5823 verification. New verify options supporting checking in openssl utility.
5827 * Fixes and wildcard matching support to hostname and email checking
5828 functions. Add manual page.
5830 *Florian Weimer (Red Hat Product Security Team)*
5832 * New functions to check a hostname email or IP address against a
5833 certificate. Add options x509 utility to print results of checks against
5838 * Fix OCSP checking.
5840 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5842 * Initial experimental support for explicitly trusted non-root CAs.
5843 OpenSSL still tries to build a complete chain to a root but if an
5844 intermediate CA has a trust setting included that is used. The first
5845 setting is used: whether to trust (e.g., -addtrust option to the x509
5850 * Add -trusted_first option which attempts to find certificates in the
5851 trusted store even if an untrusted chain is also supplied.
5855 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5856 platform support for Linux and Android.
5860 * Support for linux-x32, ILP32 environment in x86_64 framework.
5864 * Experimental multi-implementation support for FIPS capable OpenSSL.
5865 When in FIPS mode the approved implementations are used as normal,
5866 when not in FIPS mode the internal unapproved versions are used instead.
5867 This means that the FIPS capable OpenSSL isn't forced to use the
5868 (often lower performance) FIPS implementations outside FIPS mode.
5872 * Transparently support X9.42 DH parameters when calling
5873 PEM_read_bio_DHparameters. This means existing applications can handle
5874 the new parameter format automatically.
5878 * Initial experimental support for X9.42 DH parameter format: mainly
5879 to support use of 'q' parameter for RFC5114 parameters.
5883 * Add DH parameters from RFC5114 including test data to dhtest.
5887 * Support for automatic EC temporary key parameter selection. If enabled
5888 the most preferred EC parameters are automatically used instead of
5889 hardcoded fixed parameters. Now a server just has to call:
5890 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5891 support ECDH and use the most appropriate parameters.
5895 * Enhance and tidy EC curve and point format TLS extension code. Use
5896 static structures instead of allocation if default values are used.
5897 New ctrls to set curves we wish to support and to retrieve shared curves.
5898 Print out shared curves in s_server. New options to s_server and s_client
5899 to set list of supported curves.
5903 * New ctrls to retrieve supported signature algorithms and
5904 supported curve values as an array of NIDs. Extend openssl utility
5905 to print out received values.
5909 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5910 between NIDs and the more common NIST names such as "P-256". Enhance
5911 ecparam utility and ECC method to recognise the NIST names for curves.
5915 * Enhance SSL/TLS certificate chain handling to support different
5916 chains for each certificate instead of one chain in the parent SSL_CTX.
5920 * Support for fixed DH ciphersuite client authentication: where both
5921 server and client use DH certificates with common parameters.
5925 * Support for fixed DH ciphersuites: those requiring DH server
5930 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5932 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5933 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5934 X509_CINF_get_signature were reverted post internal team review.
5939 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5941 * OCSP Status Request extension unbounded memory growth
5943 A malicious client can send an excessively large OCSP Status Request
5944 extension. If that client continually requests renegotiation, sending a
5945 large OCSP Status Request extension each time, then there will be unbounded
5946 memory growth on the server. This will eventually lead to a Denial Of
5947 Service attack through memory exhaustion. Servers with a default
5948 configuration are vulnerable even if they do not support OCSP. Builds using
5949 the "no-ocsp" build time option are not affected.
5951 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5956 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5959 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5965 * OOB write in MDC2_Update()
5967 An overflow can occur in MDC2_Update() either if called directly or
5968 through the EVP_DigestUpdate() function using MDC2. If an attacker
5969 is able to supply very large amounts of input data after a previous
5970 call to EVP_EncryptUpdate() with a partial block then a length check
5971 can overflow resulting in a heap corruption.
5973 The amount of data needed is comparable to SIZE_MAX which is impractical
5976 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5981 * Malformed SHA512 ticket DoS
5983 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5984 DoS attack where a malformed ticket will result in an OOB read which will
5987 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5988 a custom server callback and ticket lookup mechanism.
5990 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5995 * OOB write in BN_bn2dec()
5997 The function BN_bn2dec() does not check the return value of BN_div_word().
5998 This can cause an OOB write if an application uses this function with an
5999 overly large BIGNUM. This could be a problem if an overly large certificate
6000 or CRL is printed out from an untrusted source. TLS is not affected because
6001 record limits will reject an oversized certificate before it is parsed.
6003 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6008 * OOB read in TS_OBJ_print_bio()
6010 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6011 the total length the OID text representation would use and not the amount
6012 of data written. This will result in OOB reads when large OIDs are
6015 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6020 * Pointer arithmetic undefined behaviour
6022 Avoid some undefined pointer arithmetic
6024 A common idiom in the codebase is to check limits in the following manner:
6027 Where "p" points to some malloc'd data of SIZE bytes and
6030 "len" here could be from some externally supplied data (e.g. from a TLS
6033 The rules of C pointer arithmetic are such that "p + len" is only well
6034 defined where len <= SIZE. Therefore the above idiom is actually
6035 undefined behaviour.
6037 For example this could cause problems if some malloc implementation
6038 provides an address for "p" such that "p + len" actually overflows for
6039 values of len that are too big and therefore p + len < limit.
6041 This issue was reported to OpenSSL by Guido Vranken
6046 * Constant time flag not preserved in DSA signing
6048 Operations in the DSA signing algorithm should run in constant time in
6049 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6050 implementation means that a non-constant time codepath is followed for
6051 certain operations. This has been demonstrated through a cache-timing
6052 attack to be sufficient for an attacker to recover the private DSA key.
6054 This issue was reported by César Pereida (Aalto University), Billy Brumley
6055 (Tampere University of Technology), and Yuval Yarom (The University of
6056 Adelaide and NICTA).
6061 * DTLS buffered message DoS
6063 In a DTLS connection where handshake messages are delivered out-of-order
6064 those messages that OpenSSL is not yet ready to process will be buffered
6065 for later use. Under certain circumstances, a flaw in the logic means that
6066 those messages do not get removed from the buffer even though the handshake
6067 has been completed. An attacker could force up to approx. 15 messages to
6068 remain in the buffer when they are no longer required. These messages will
6069 be cleared when the DTLS connection is closed. The default maximum size for
6070 a message is 100k. Therefore the attacker could force an additional 1500k
6071 to be consumed per connection. By opening many simulataneous connections an
6072 attacker could cause a DoS attack through memory exhaustion.
6074 This issue was reported to OpenSSL by Quan Luo.
6079 * DTLS replay protection DoS
6081 A flaw in the DTLS replay attack protection mechanism means that records
6082 that arrive for future epochs update the replay protection "window" before
6083 the MAC for the record has been validated. This could be exploited by an
6084 attacker by sending a record for the next epoch (which does not have to
6085 decrypt or have a valid MAC), with a very large sequence number. This means
6086 that all subsequent legitimate packets are dropped causing a denial of
6087 service for a specific DTLS connection.
6089 This issue was reported to OpenSSL by the OCAP audit team.
6094 * Certificate message OOB reads
6096 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6097 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6098 theoretical DoS risk but this has not been observed in practice on common
6101 The messages affected are client certificate, client certificate request
6102 and server certificate. As a result the attack can only be performed
6103 against a client or a server which enables client authentication.
6105 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6110 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6112 * Prevent padding oracle in AES-NI CBC MAC check
6114 A MITM attacker can use a padding oracle attack to decrypt traffic
6115 when the connection uses an AES CBC cipher and the server support
6118 This issue was introduced as part of the fix for Lucky 13 padding
6119 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6120 constant time by making sure that always the same bytes are read and
6121 compared against either the MAC or padding bytes. But it no longer
6122 checked that there was enough data to have both the MAC and padding
6125 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6130 * Fix EVP_EncodeUpdate overflow
6132 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6133 Base64 encoding of binary data. If an attacker is able to supply very large
6134 amounts of input data then a length check can overflow resulting in a heap
6137 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6138 the `PEM_write_bio*` family of functions. These are mainly used within the
6139 OpenSSL command line applications, so any application which processes data
6140 from an untrusted source and outputs it as a PEM file should be considered
6141 vulnerable to this issue. User applications that call these APIs directly
6142 with large amounts of untrusted data may also be vulnerable.
6144 This issue was reported by Guido Vranken.
6149 * Fix EVP_EncryptUpdate overflow
6151 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6152 is able to supply very large amounts of input data after a previous call to
6153 EVP_EncryptUpdate() with a partial block then a length check can overflow
6154 resulting in a heap corruption. Following an analysis of all OpenSSL
6155 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6156 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6157 the first called function after an EVP_EncryptInit(), and therefore that
6158 specific call must be safe. The second form is where the length passed to
6159 EVP_EncryptUpdate() can be seen from the code to be some small value and
6160 therefore there is no possibility of an overflow. Since all instances are
6161 one of these two forms, it is believed that there can be no overflows in
6162 internal code due to this problem. It should be noted that
6163 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6164 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6165 of these calls have also been analysed too and it is believed there are no
6166 instances in internal usage where an overflow could occur.
6168 This issue was reported by Guido Vranken.
6173 * Prevent ASN.1 BIO excessive memory allocation
6175 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6176 a short invalid encoding can casuse allocation of large amounts of memory
6177 potentially consuming excessive resources or exhausting memory.
6179 Any application parsing untrusted data through d2i BIO functions is
6180 affected. The memory based functions such as d2i_X509() are *not* affected.
6181 Since the memory based functions are used by the TLS library, TLS
6182 applications are not affected.
6184 This issue was reported by Brian Carpenter.
6191 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6192 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6193 in arbitrary stack data being returned in the buffer.
6195 This issue was reported by Guido Vranken.
6200 * Modify behavior of ALPN to invoke callback after SNI/servername
6201 callback, such that updates to the SSL_CTX affect ALPN.
6205 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6210 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6211 methods are enabled and ssl2 is disabled the methods return NULL.
6215 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6217 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6218 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6219 provide any "EXPORT" or "LOW" strength ciphers.
6223 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6224 is by default disabled at build-time. Builds that are not configured with
6225 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6226 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6227 will need to explicitly call either of:
6229 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6231 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6233 as appropriate. Even if either of those is used, or the application
6234 explicitly uses the version-specific SSLv2_method() or its client and
6235 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6236 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6237 ciphers, and SSLv2 56-bit DES are no longer available.
6242 * Fix a double-free in DSA code
6244 A double free bug was discovered when OpenSSL parses malformed DSA private
6245 keys and could lead to a DoS attack or memory corruption for applications
6246 that receive DSA private keys from untrusted sources. This scenario is
6249 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6255 * Disable SRP fake user seed to address a server memory leak.
6257 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6259 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6260 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6261 was changed to ignore the "fake user" SRP seed, even if the seed
6264 Users should use SRP_VBASE_get1_by_user instead. Note that in
6265 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6266 also that even though configuring the SRP seed attempts to hide
6267 invalid usernames by continuing the handshake with fake
6268 credentials, this behaviour is not constant time and no strong
6269 guarantees are made that the handshake is indistinguishable from
6270 that of a valid user.
6275 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6277 In the BN_hex2bn function the number of hex digits is calculated using an
6278 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6279 large values of `i` this can result in `bn_expand` not allocating any
6280 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6281 field as NULL leading to a subsequent NULL ptr deref. For very large values
6282 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6283 In this case memory is allocated to the internal BIGNUM data field, but it
6284 is insufficiently sized leading to heap corruption. A similar issue exists
6285 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6286 is ever called by user applications with very large untrusted hex/dec data.
6287 This is anticipated to be a rare occurrence.
6289 All OpenSSL internal usage of these functions use data that is not expected
6290 to be untrusted, e.g. config file data or application command line
6291 arguments. If user developed applications generate config file data based
6292 on untrusted data then it is possible that this could also lead to security
6293 consequences. This is also anticipated to be rare.
6295 This issue was reported to OpenSSL by Guido Vranken.
6300 * Fix memory issues in `BIO_*printf` functions
6302 The internal `fmtstr` function used in processing a "%s" format string in
6303 the `BIO_*printf` functions could overflow while calculating the length of a
6304 string and cause an OOB read when printing very long strings.
6306 Additionally the internal `doapr_outch` function can attempt to write to an
6307 OOB memory location (at an offset from the NULL pointer) in the event of a
6308 memory allocation failure. In 1.0.2 and below this could be caused where
6309 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6310 could be in processing a very long "%s" format string. Memory leaks can
6313 The first issue may mask the second issue dependent on compiler behaviour.
6314 These problems could enable attacks where large amounts of untrusted data
6315 is passed to the `BIO_*printf` functions. If applications use these functions
6316 in this way then they could be vulnerable. OpenSSL itself uses these
6317 functions when printing out human-readable dumps of ASN.1 data. Therefore
6318 applications that print this data could be vulnerable if the data is from
6319 untrusted sources. OpenSSL command line applications could also be
6320 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6321 as command line arguments.
6323 Libssl is not considered directly vulnerable. Additionally certificates etc
6324 received via remote connections via libssl are also unlikely to be able to
6325 trigger these issues because of message size limits enforced within libssl.
6327 This issue was reported to OpenSSL Guido Vranken.
6332 * Side channel attack on modular exponentiation
6334 A side-channel attack was found which makes use of cache-bank conflicts on
6335 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6336 of RSA keys. The ability to exploit this issue is limited as it relies on
6337 an attacker who has control of code in a thread running on the same
6338 hyper-threaded core as the victim thread which is performing decryptions.
6340 This issue was reported to OpenSSL by Yuval Yarom, The University of
6341 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6342 Nadia Heninger, University of Pennsylvania with more information at
6343 <http://cachebleed.info>.
6348 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6349 if no keysize is specified with default_bits. This fixes an
6350 omission in an earlier change that changed all RSA/DSA key generation
6351 commands to use 2048 bits by default.
6355 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6357 * Protection for DH small subgroup attacks
6359 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6360 switched on by default and cannot be disabled. This could have some
6365 * SSLv2 doesn't block disabled ciphers
6367 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6368 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6369 been disabled, provided that the SSLv2 protocol was not also disabled via
6372 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6373 and Sebastian Schinzel.
6378 * Reject DH handshakes with parameters shorter than 1024 bits.
6382 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6384 * Certificate verify crash with missing PSS parameter
6386 The signature verification routines will crash with a NULL pointer
6387 dereference if presented with an ASN.1 signature using the RSA PSS
6388 algorithm and absent mask generation function parameter. Since these
6389 routines are used to verify certificate signature algorithms this can be
6390 used to crash any certificate verification operation and exploited in a
6391 DoS attack. Any application which performs certificate verification is
6392 vulnerable including OpenSSL clients and servers which enable client
6395 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6400 * X509_ATTRIBUTE memory leak
6402 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6403 memory. This structure is used by the PKCS#7 and CMS routines so any
6404 application which reads PKCS#7 or CMS data from untrusted sources is
6405 affected. SSL/TLS is not affected.
6407 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6413 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6414 This changes the decoding behaviour for some invalid messages,
6415 though the change is mostly in the more lenient direction, and
6416 legacy behaviour is preserved as much as possible.
6420 * In DSA_generate_parameters_ex, if the provided seed is too short,
6421 use a random seed, as already documented.
6423 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6425 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6427 * Alternate chains certificate forgery
6429 During certificate verfification, OpenSSL will attempt to find an
6430 alternative certificate chain if the first attempt to build such a chain
6431 fails. An error in the implementation of this logic can mean that an
6432 attacker could cause certain checks on untrusted certificates to be
6433 bypassed, such as the CA flag, enabling them to use a valid leaf
6434 certificate to act as a CA and "issue" an invalid certificate.
6436 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6442 * Race condition handling PSK identify hint
6444 If PSK identity hints are received by a multi-threaded client then
6445 the values are wrongly updated in the parent SSL_CTX structure. This can
6446 result in a race condition potentially leading to a double free of the
6452 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6454 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6455 incompatibility in the handling of HMAC. The previous ABI has now been
6458 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6460 * Malformed ECParameters causes infinite loop
6462 When processing an ECParameters structure OpenSSL enters an infinite loop
6463 if the curve specified is over a specially malformed binary polynomial
6466 This can be used to perform denial of service against any
6467 system which processes public keys, certificate requests or
6468 certificates. This includes TLS clients and TLS servers with
6469 client authentication enabled.
6471 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6476 * Exploitable out-of-bounds read in X509_cmp_time
6478 X509_cmp_time does not properly check the length of the ASN1_TIME
6479 string and can read a few bytes out of bounds. In addition,
6480 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6483 An attacker can use this to craft malformed certificates and CRLs of
6484 various sizes and potentially cause a segmentation fault, resulting in
6485 a DoS on applications that verify certificates or CRLs. TLS clients
6486 that verify CRLs are affected. TLS clients and servers with client
6487 authentication enabled may be affected if they use custom verification
6490 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6491 independently by Hanno Böck.
6496 * PKCS7 crash with missing EnvelopedContent
6498 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6499 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6500 with missing content and trigger a NULL pointer dereference on parsing.
6502 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6503 structures from untrusted sources are affected. OpenSSL clients and
6504 servers are not affected.
6506 This issue was reported to OpenSSL by Michal Zalewski (Google).
6511 * CMS verify infinite loop with unknown hash function
6513 When verifying a signedData message the CMS code can enter an infinite loop
6514 if presented with an unknown hash function OID. This can be used to perform
6515 denial of service against any system which verifies signedData messages using
6517 This issue was reported to OpenSSL by Johannes Bauer.
6522 * Race condition handling NewSessionTicket
6524 If a NewSessionTicket is received by a multi-threaded client when attempting to
6525 reuse a previous ticket then a race condition can occur potentially leading to
6526 a double free of the ticket data.
6531 * Reject DH handshakes with parameters shorter than 768 bits.
6533 *Kurt Roeckx and Emilia Kasper*
6535 * dhparam: generate 2048-bit parameters by default.
6537 *Kurt Roeckx and Emilia Kasper*
6539 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6541 * Segmentation fault in ASN1_TYPE_cmp fix
6543 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6544 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6545 certificate signature algorithm consistency this can be used to crash any
6546 certificate verification operation and exploited in a DoS attack. Any
6547 application which performs certificate verification is vulnerable including
6548 OpenSSL clients and servers which enable client authentication.
6553 * ASN.1 structure reuse memory corruption fix
6555 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6556 memory corruption via an invalid write. Such reuse is and has been
6557 strongly discouraged and is believed to be rare.
6559 Applications that parse structures containing CHOICE or ANY DEFINED BY
6560 components may be affected. Certificate parsing (d2i_X509 and related
6561 functions) are however not affected. OpenSSL clients and servers are
6567 * PKCS7 NULL pointer dereferences fix
6569 The PKCS#7 parsing code does not handle missing outer ContentInfo
6570 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6571 missing content and trigger a NULL pointer dereference on parsing.
6573 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6574 otherwise parse PKCS#7 structures from untrusted sources are
6575 affected. OpenSSL clients and servers are not affected.
6577 This issue was reported to OpenSSL by Michal Zalewski (Google).
6582 * DoS via reachable assert in SSLv2 servers fix
6584 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6585 servers that both support SSLv2 and enable export cipher suites by sending
6586 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6588 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6589 (OpenSSL development team).
6594 * Use After Free following d2i_ECPrivatekey error fix
6596 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6597 could cause a use after free condition. This, in turn, could cause a double
6598 free in several private key parsing functions (such as d2i_PrivateKey
6599 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6600 for applications that receive EC private keys from untrusted
6601 sources. This scenario is considered rare.
6603 This issue was discovered by the BoringSSL project and fixed in their
6609 * X509_to_X509_REQ NULL pointer deref fix
6611 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6612 the certificate key is invalid. This function is rarely used in practice.
6614 This issue was discovered by Brian Carpenter.
6619 * Removed the export ciphers from the DEFAULT ciphers
6623 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6625 * Build fixes for the Windows and OpenVMS platforms
6627 *Matt Caswell and Richard Levitte*
6629 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6631 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6632 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6633 dereference. This could lead to a Denial Of Service attack. Thanks to
6634 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6639 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6640 dtls1_buffer_record function under certain conditions. In particular this
6641 could occur if an attacker sent repeated DTLS records with the same
6642 sequence number but for the next epoch. The memory leak could be exploited
6643 by an attacker in a Denial of Service attack through memory exhaustion.
6644 Thanks to Chris Mueller for reporting this issue.
6649 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6650 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6651 method would be set to NULL which could later result in a NULL pointer
6652 dereference. Thanks to Frank Schmirler for reporting this issue.
6657 * Abort handshake if server key exchange message is omitted for ephemeral
6660 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6661 reporting this issue.
6666 * Remove non-export ephemeral RSA code on client and server. This code
6667 violated the TLS standard by allowing the use of temporary RSA keys in
6668 non-export ciphersuites and could be used by a server to effectively
6669 downgrade the RSA key length used to a value smaller than the server
6670 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6671 INRIA or reporting this issue.
6676 * Fixed issue where DH client certificates are accepted without verification.
6677 An OpenSSL server will accept a DH certificate for client authentication
6678 without the certificate verify message. This effectively allows a client to
6679 authenticate without the use of a private key. This only affects servers
6680 which trust a client certificate authority which issues certificates
6681 containing DH keys: these are extremely rare and hardly ever encountered.
6682 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6688 * Ensure that the session ID context of an SSL is updated when its
6689 SSL_CTX is updated via SSL_set_SSL_CTX.
6691 The session ID context is typically set from the parent SSL_CTX,
6692 and can vary with the CTX.
6696 * Fix various certificate fingerprint issues.
6698 By using non-DER or invalid encodings outside the signed portion of a
6699 certificate the fingerprint can be changed without breaking the signature.
6700 Although no details of the signed portion of the certificate can be changed
6701 this can cause problems with some applications: e.g. those using the
6702 certificate fingerprint for blacklists.
6704 1. Reject signatures with non zero unused bits.
6706 If the BIT STRING containing the signature has non zero unused bits reject
6707 the signature. All current signature algorithms require zero unused bits.
6709 2. Check certificate algorithm consistency.
6711 Check the AlgorithmIdentifier inside TBS matches the one in the
6712 certificate signature. NB: this will result in signature failure
6713 errors for some broken certificates.
6715 Thanks to Konrad Kraszewski from Google for reporting this issue.
6717 3. Check DSA/ECDSA signatures use DER.
6719 Re-encode DSA/ECDSA signatures and compare with the original received
6720 signature. Return an error if there is a mismatch.
6722 This will reject various cases including garbage after signature
6723 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6724 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6725 (negative or with leading zeroes).
6727 Further analysis was conducted and fixes were developed by Stephen Henson
6728 of the OpenSSL core team.
6734 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6735 results on some platforms, including x86_64. This bug occurs at random
6736 with a very low probability, and is not known to be exploitable in any
6737 way, though its exact impact is difficult to determine. Thanks to Pieter
6738 Wuille (Blockstream) who reported this issue and also suggested an initial
6739 fix. Further analysis was conducted by the OpenSSL development team and
6740 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6741 the OpenSSL core team.
6746 * Do not resume sessions on the server if the negotiated protocol
6747 version does not match the session's version. Resuming with a different
6748 version, while not strictly forbidden by the RFC, is of questionable
6749 sanity and breaks all known clients.
6751 *David Benjamin, Emilia Käsper*
6753 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6754 early CCS messages during renegotiation. (Note that because
6755 renegotiation is encrypted, this early CCS was not exploitable.)
6759 * Tighten client-side session ticket handling during renegotiation:
6760 ensure that the client only accepts a session ticket if the server sends
6761 the extension anew in the ServerHello. Previously, a TLS client would
6762 reuse the old extension state and thus accept a session ticket if one was
6763 announced in the initial ServerHello.
6765 Similarly, ensure that the client requires a session ticket if one
6766 was advertised in the ServerHello. Previously, a TLS client would
6767 ignore a missing NewSessionTicket message.
6771 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6775 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6776 sends a carefully crafted handshake message, to cause OpenSSL to fail
6777 to free up to 64k of memory causing a memory leak. This could be
6778 exploited in a Denial Of Service attack. This issue affects OpenSSL
6779 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6780 whether SRTP is used or configured. Implementations of OpenSSL that
6781 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6783 The fix was developed by the OpenSSL team.
6788 * Session Ticket Memory Leak.
6790 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6791 integrity of that ticket is first verified. In the event of a session
6792 ticket integrity check failing, OpenSSL will fail to free memory
6793 causing a memory leak. By sending a large number of invalid session
6794 tickets an attacker could exploit this issue in a Denial Of Service
6800 * Build option no-ssl3 is incomplete.
6802 When OpenSSL is configured with "no-ssl3" as a build option, servers
6803 could accept and complete a SSL 3.0 handshake, and clients could be
6804 configured to send them.
6807 *Akamai and the OpenSSL team*
6809 * Add support for TLS_FALLBACK_SCSV.
6810 Client applications doing fallback retries should call
6811 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6814 *Adam Langley, Bodo Moeller*
6816 * Add additional DigestInfo checks.
6818 Re-encode DigestInto in DER and check against the original when
6819 verifying RSA signature: this will reject any improperly encoded
6820 DigestInfo structures.
6822 Note: this is a precautionary measure and no attacks are currently known.
6826 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6828 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6829 SRP code can be overrun an internal buffer. Add sanity check that
6830 g, A, B < N to SRP code.
6832 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6833 Group for discovering this issue.
6838 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6839 TLS 1.0 instead of higher protocol versions when the ClientHello message
6840 is badly fragmented. This allows a man-in-the-middle attacker to force a
6841 downgrade to TLS 1.0 even if both the server and the client support a
6842 higher protocol version, by modifying the client's TLS records.
6844 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6845 researching this issue.
6850 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6851 to a denial of service attack. A malicious server can crash the client
6852 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6853 ciphersuite and sending carefully crafted handshake messages.
6855 Thanks to Felix Gröbert (Google) for discovering and researching this
6861 * By sending carefully crafted DTLS packets an attacker could cause openssl
6862 to leak memory. This can be exploited through a Denial of Service attack.
6863 Thanks to Adam Langley for discovering and researching this issue.
6868 * An attacker can force openssl to consume large amounts of memory whilst
6869 processing DTLS handshake messages. This can be exploited through a
6870 Denial of Service attack.
6871 Thanks to Adam Langley for discovering and researching this issue.
6876 * An attacker can force an error condition which causes openssl to crash
6877 whilst processing DTLS packets due to memory being freed twice. This
6878 can be exploited through a Denial of Service attack.
6879 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6885 * If a multithreaded client connects to a malicious server using a resumed
6886 session and the server sends an ec point format extension it could write
6887 up to 255 bytes to freed memory.
6889 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6895 * A malicious server can crash an OpenSSL client with a null pointer
6896 dereference (read) by specifying an SRP ciphersuite even though it was not
6897 properly negotiated with the client. This can be exploited through a
6898 Denial of Service attack.
6900 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6901 discovering and researching this issue.
6906 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6907 X509_name_oneline, X509_name_print_ex et al. to leak some information
6908 from the stack. Applications may be affected if they echo pretty printing
6909 output to the attacker.
6911 Thanks to Ivan Fratric (Google) for discovering this issue.
6914 *Emilia Käsper, and Steve Henson*
6916 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6917 for corner cases. (Certain input points at infinity could lead to
6918 bogus results, with non-infinity inputs mapped to infinity too.)
6922 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6924 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6925 handshake can force the use of weak keying material in OpenSSL
6926 SSL/TLS clients and servers.
6928 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6929 researching this issue. ([CVE-2014-0224])
6931 *KIKUCHI Masashi, Steve Henson*
6933 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6934 OpenSSL DTLS client the code can be made to recurse eventually crashing
6937 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6940 *Imre Rad, Steve Henson*
6942 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6943 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6944 client or server. This is potentially exploitable to run arbitrary
6945 code on a vulnerable client or server.
6947 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6949 *Jüri Aedla, Steve Henson*
6951 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6952 are subject to a denial of service attack.
6954 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6955 this issue. ([CVE-2014-3470])
6957 *Felix Gröbert, Ivan Fratric, Steve Henson*
6959 * Harmonize version and its documentation. -f flag is used to display
6962 *mancha <mancha1@zoho.com>*
6964 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6965 in i2d_ECPrivateKey.
6967 *mancha <mancha1@zoho.com>*
6969 * Fix some double frees. These are not thought to be exploitable.
6971 *mancha <mancha1@zoho.com>*
6973 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6975 * A missing bounds check in the handling of the TLS heartbeat extension
6976 can be used to reveal up to 64k of memory to a connected client or
6979 Thanks for Neel Mehta of Google Security for discovering this bug and to
6980 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6981 preparing the fix ([CVE-2014-0160])
6983 *Adam Langley, Bodo Moeller*
6985 * Fix for the attack described in the paper "Recovering OpenSSL
6986 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6987 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6988 <http://eprint.iacr.org/2014/140>
6990 Thanks to Yuval Yarom and Naomi Benger for discovering this
6991 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6993 *Yuval Yarom and Naomi Benger*
6995 * TLS pad extension: draft-agl-tls-padding-03
6997 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6998 TLS client Hello record length value would otherwise be > 255 and
6999 less that 512 pad with a dummy extension containing zeroes so it
7000 is at least 512 bytes long.
7002 *Adam Langley, Steve Henson*
7004 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7006 * Fix for TLS record tampering bug. A carefully crafted invalid
7007 handshake could crash OpenSSL with a NULL pointer exception.
7008 Thanks to Anton Johansson for reporting this issues.
7011 * Keep original DTLS digest and encryption contexts in retransmission
7012 structures so we can use the previous session parameters if they need
7013 to be resent. ([CVE-2013-6450])
7017 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7018 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7019 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7020 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7021 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7022 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7024 *Rob Stradling, Adam Langley*
7026 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7028 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7029 supporting platforms or when small records were transferred.
7031 *Andy Polyakov, Steve Henson*
7033 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7035 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7037 This addresses the flaw in CBC record processing discovered by
7038 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7039 at: <http://www.isg.rhul.ac.uk/tls/>
7041 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7042 Security Group at Royal Holloway, University of London
7043 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7044 Emilia Käsper for the initial patch.
7047 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7049 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7050 ciphersuites which can be exploited in a denial of service attack.
7051 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7052 and detecting this bug and to Wolfgang Ettlinger
7053 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7058 * Return an error when checking OCSP signatures when key is NULL.
7059 This fixes a DoS attack. ([CVE-2013-0166])
7063 * Make openssl verify return errors.
7065 *Chris Palmer <palmer@google.com> and Ben Laurie*
7067 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7068 the right response is stapled. Also change SSL_get_certificate()
7069 so it returns the certificate actually sent.
7070 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7072 *Rob Stradling <rob.stradling@comodo.com>*
7074 * Fix possible deadlock when decoding public keys.
7078 * Don't use TLS 1.0 record version number in initial client hello
7083 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7085 * Sanity check record length before skipping explicit IV in TLS
7086 1.2, 1.1 and DTLS to fix DoS attack.
7088 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7089 fuzzing as a service testing platform.
7094 * Initialise tkeylen properly when encrypting CMS messages.
7095 Thanks to Solar Designer of Openwall for reporting this issue.
7099 * In FIPS mode don't try to use composite ciphers as they are not
7104 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7106 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7107 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7108 mean any application compiled against OpenSSL 1.0.0 headers setting
7109 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7110 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7111 0x10000000L Any application which was previously compiled against
7112 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7113 will need to be recompiled as a result. Letting be results in
7114 inability to disable specifically TLS 1.1 and in client context,
7115 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7119 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7120 disable just protocol X, but all protocols above X *if* there are
7121 protocols *below* X still enabled. In more practical terms it means
7122 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7123 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7124 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7129 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7131 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7132 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7133 in CRYPTO_realloc_clean.
7135 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7136 issue and to Adam Langley <agl@chromium.org> for fixing it.
7139 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7141 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7145 * Workarounds for some broken servers that "hang" if a client hello
7146 record length exceeds 255 bytes.
7148 1. Do not use record version number > TLS 1.0 in initial client
7149 hello: some (but not all) hanging servers will now work.
7150 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7151 the number of ciphers sent in the client hello. This should be
7152 set to an even number, such as 50, for example by passing:
7153 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7154 Most broken servers should now work.
7155 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7156 TLS 1.2 client support entirely.
7160 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7164 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7166 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7167 STRING form instead of a DigestInfo.
7171 * The format used for MDC2 RSA signatures is inconsistent between EVP
7172 and the RSA_sign/RSA_verify functions. This was made more apparent when
7173 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7174 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7175 the correct format in RSA_verify so both forms transparently work.
7179 * Some servers which support TLS 1.0 can choke if we initially indicate
7180 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7181 encrypted premaster secret. As a workaround use the maximum permitted
7182 client version in client hello, this should keep such servers happy
7183 and still work with previous versions of OpenSSL.
7187 * Add support for TLS/DTLS heartbeats.
7189 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7191 * Add support for SCTP.
7193 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7195 * Improved PRNG seeding for VOS.
7197 *Paul Green <Paul.Green@stratus.com>*
7199 * Extensive assembler packs updates, most notably:
7201 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7202 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7203 - x86_64: bit-sliced AES implementation;
7204 - ARM: NEON support, contemporary platforms optimizations;
7205 - s390x: z196 support;
7206 - `*`: GHASH and GF(2^m) multiplication implementations;
7210 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7211 (removal of unnecessary code)
7213 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7215 * Add TLS key material exporter from RFC 5705.
7219 * Add DTLS-SRTP negotiation from RFC 5764.
7223 * Add Next Protocol Negotiation,
7224 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7225 disabled with a no-npn flag to config or Configure. Code donated
7228 *Adam Langley <agl@google.com> and Ben Laurie*
7230 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7231 NIST-P256, NIST-P521, with constant-time single point multiplication on
7232 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7233 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7234 Code made available under Apache License version 2.0.
7236 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7237 line to include this in your build of OpenSSL, and run "make depend" (or
7238 "make update"). This enables the following EC_METHODs:
7240 EC_GFp_nistp224_method()
7241 EC_GFp_nistp256_method()
7242 EC_GFp_nistp521_method()
7244 EC_GROUP_new_by_curve_name() will automatically use these (while
7245 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7248 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7250 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7251 all platforms. Move ssize_t definition from e_os.h to the public
7252 header file e_os2.h as it now appears in public header file cms.h
7256 * New -sigopt option to the ca, req and x509 utilities. Additional
7257 signature parameters can be passed using this option and in
7262 * Add RSA PSS signing function. This will generate and set the
7263 appropriate AlgorithmIdentifiers for PSS based on those in the
7264 corresponding EVP_MD_CTX structure. No application support yet.
7268 * Support for companion algorithm specific ASN1 signing routines.
7269 New function ASN1_item_sign_ctx() signs a pre-initialised
7270 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7271 the appropriate parameters.
7275 * Add new algorithm specific ASN1 verification initialisation function
7276 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7277 handling will be the same no matter what EVP_PKEY_METHOD is used.
7278 Add a PSS handler to support verification of PSS signatures: checked
7279 against a number of sample certificates.
7283 * Add signature printing for PSS. Add PSS OIDs.
7285 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7287 * Add algorithm specific signature printing. An individual ASN1 method
7288 can now print out signatures instead of the standard hex dump.
7290 More complex signatures (e.g. PSS) can print out more meaningful
7291 information. Include DSA version that prints out the signature
7296 * Password based recipient info support for CMS library: implementing
7301 * Split password based encryption into PBES2 and PBKDF2 functions. This
7302 neatly separates the code into cipher and PBE sections and is required
7303 for some algorithms that split PBES2 into separate pieces (such as
7304 password based CMS).
7308 * Session-handling fixes:
7309 - Fix handling of connections that are resuming with a session ID,
7310 but also support Session Tickets.
7311 - Fix a bug that suppressed issuing of a new ticket if the client
7312 presented a ticket with an expired session.
7313 - Try to set the ticket lifetime hint to something reasonable.
7314 - Make tickets shorter by excluding irrelevant information.
7315 - On the client side, don't ignore renewed tickets.
7317 *Adam Langley, Bodo Moeller (Google)*
7319 * Fix PSK session representation.
7323 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7325 This work was sponsored by Intel.
7329 * Add GCM support to TLS library. Some custom code is needed to split
7330 the IV between the fixed (from PRF) and explicit (from TLS record)
7331 portions. This adds all GCM ciphersuites supported by RFC5288 and
7332 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7333 add a special AESGCM string for GCM only.
7337 * Expand range of ctrls for AES GCM. Permit setting invocation
7338 field on decrypt and retrieval of invocation field only on encrypt.
7342 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7343 As required by RFC5289 these ciphersuites cannot be used if for
7344 versions of TLS earlier than 1.2.
7348 * For FIPS capable OpenSSL interpret a NULL default public key method
7349 as unset and return the appropriate default but do *not* set the default.
7350 This means we can return the appropriate method in applications that
7351 switch between FIPS and non-FIPS modes.
7355 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7356 ENGINE is used then we cannot handle that in the FIPS module so we
7357 keep original code iff non-FIPS operations are allowed.
7361 * Add -attime option to openssl utilities.
7363 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7365 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7369 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7370 FIPS EC methods unconditionally for now.
7374 * New build option no-ec2m to disable characteristic 2 code.
7378 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7379 all cases can be covered as some introduce binary incompatibilities.
7383 * Redirect RSA operations to FIPS module including keygen,
7384 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7388 * Add similar low level API blocking to ciphers.
7392 * Low level digest APIs are not approved in FIPS mode: any attempt
7393 to use these will cause a fatal error. Applications that *really* want
7394 to use them can use the `private_*` version instead.
7398 * Redirect cipher operations to FIPS module for FIPS builds.
7402 * Redirect digest operations to FIPS module for FIPS builds.
7406 * Update build system to add "fips" flag which will link in fipscanister.o
7407 for static and shared library builds embedding a signature if needed.
7411 * Output TLS supported curves in preference order instead of numerical
7412 order. This is currently hardcoded for the highest order curves first.
7413 This should be configurable so applications can judge speed vs strength.
7417 * Add TLS v1.2 server support for client authentication.
7421 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7426 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7427 FIPS modules versions.
7431 * Add TLS v1.2 client side support for client authentication. Keep cache
7432 of handshake records longer as we don't know the hash algorithm to use
7433 until after the certificate request message is received.
7437 * Initial TLS v1.2 client support. Add a default signature algorithms
7438 extension including all the algorithms we support. Parse new signature
7439 format in client key exchange. Relax some ECC signing restrictions for
7440 TLS v1.2 as indicated in RFC5246.
7444 * Add server support for TLS v1.2 signature algorithms extension. Switch
7445 to new signature format when needed using client digest preference.
7446 All server ciphersuites should now work correctly in TLS v1.2. No client
7447 support yet and no support for client certificates.
7451 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7452 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7453 ciphersuites. At present only RSA key exchange ciphersuites work with
7454 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7455 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7456 and version checking.
7460 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7461 with this defined it will not be affected by any changes to ssl internal
7462 structures. Add several utility functions to allow openssl application
7463 to work with OPENSSL_NO_SSL_INTERN defined.
7467 * A long standing patch to add support for SRP from EdelWeb (Peter
7468 Sylvester and Christophe Renou) was integrated.
7469 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7470 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7473 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7477 * Permit abbreviated handshakes when renegotiating using the function
7478 SSL_renegotiate_abbreviated().
7480 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7482 * Add call to ENGINE_register_all_complete() to
7483 ENGINE_load_builtin_engines(), so some implementations get used
7484 automatically instead of needing explicit application support.
7488 * Add support for TLS key exporter as described in RFC5705.
7490 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7492 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7493 a few changes are required:
7495 Add SSL_OP_NO_TLSv1_1 flag.
7496 Add TLSv1_1 methods.
7497 Update version checking logic to handle version 1.1.
7498 Add explicit IV handling (ported from DTLS code).
7499 Add command line options to s_client/s_server.
7506 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7508 * X509_ATTRIBUTE memory leak
7510 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7511 memory. This structure is used by the PKCS#7 and CMS routines so any
7512 application which reads PKCS#7 or CMS data from untrusted sources is
7513 affected. SSL/TLS is not affected.
7515 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7521 * Race condition handling PSK identify hint
7523 If PSK identity hints are received by a multi-threaded client then
7524 the values are wrongly updated in the parent SSL_CTX structure. This can
7525 result in a race condition potentially leading to a double free of the
7531 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7533 * Malformed ECParameters causes infinite loop
7535 When processing an ECParameters structure OpenSSL enters an infinite loop
7536 if the curve specified is over a specially malformed binary polynomial
7539 This can be used to perform denial of service against any
7540 system which processes public keys, certificate requests or
7541 certificates. This includes TLS clients and TLS servers with
7542 client authentication enabled.
7544 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7549 * Exploitable out-of-bounds read in X509_cmp_time
7551 X509_cmp_time does not properly check the length of the ASN1_TIME
7552 string and can read a few bytes out of bounds. In addition,
7553 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7556 An attacker can use this to craft malformed certificates and CRLs of
7557 various sizes and potentially cause a segmentation fault, resulting in
7558 a DoS on applications that verify certificates or CRLs. TLS clients
7559 that verify CRLs are affected. TLS clients and servers with client
7560 authentication enabled may be affected if they use custom verification
7563 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7564 independently by Hanno Böck.
7569 * PKCS7 crash with missing EnvelopedContent
7571 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7572 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7573 with missing content and trigger a NULL pointer dereference on parsing.
7575 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7576 structures from untrusted sources are affected. OpenSSL clients and
7577 servers are not affected.
7579 This issue was reported to OpenSSL by Michal Zalewski (Google).
7584 * CMS verify infinite loop with unknown hash function
7586 When verifying a signedData message the CMS code can enter an infinite loop
7587 if presented with an unknown hash function OID. This can be used to perform
7588 denial of service against any system which verifies signedData messages using
7590 This issue was reported to OpenSSL by Johannes Bauer.
7595 * Race condition handling NewSessionTicket
7597 If a NewSessionTicket is received by a multi-threaded client when attempting to
7598 reuse a previous ticket then a race condition can occur potentially leading to
7599 a double free of the ticket data.
7604 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7606 * Segmentation fault in ASN1_TYPE_cmp fix
7608 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7609 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7610 certificate signature algorithm consistency this can be used to crash any
7611 certificate verification operation and exploited in a DoS attack. Any
7612 application which performs certificate verification is vulnerable including
7613 OpenSSL clients and servers which enable client authentication.
7618 * ASN.1 structure reuse memory corruption fix
7620 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7621 memory corruption via an invalid write. Such reuse is and has been
7622 strongly discouraged and is believed to be rare.
7624 Applications that parse structures containing CHOICE or ANY DEFINED BY
7625 components may be affected. Certificate parsing (d2i_X509 and related
7626 functions) are however not affected. OpenSSL clients and servers are
7632 * PKCS7 NULL pointer dereferences fix
7634 The PKCS#7 parsing code does not handle missing outer ContentInfo
7635 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7636 missing content and trigger a NULL pointer dereference on parsing.
7638 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7639 otherwise parse PKCS#7 structures from untrusted sources are
7640 affected. OpenSSL clients and servers are not affected.
7642 This issue was reported to OpenSSL by Michal Zalewski (Google).
7647 * DoS via reachable assert in SSLv2 servers fix
7649 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7650 servers that both support SSLv2 and enable export cipher suites by sending
7651 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7653 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7654 (OpenSSL development team).
7659 * Use After Free following d2i_ECPrivatekey error fix
7661 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7662 could cause a use after free condition. This, in turn, could cause a double
7663 free in several private key parsing functions (such as d2i_PrivateKey
7664 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7665 for applications that receive EC private keys from untrusted
7666 sources. This scenario is considered rare.
7668 This issue was discovered by the BoringSSL project and fixed in their
7674 * X509_to_X509_REQ NULL pointer deref fix
7676 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7677 the certificate key is invalid. This function is rarely used in practice.
7679 This issue was discovered by Brian Carpenter.
7684 * Removed the export ciphers from the DEFAULT ciphers
7688 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7690 * Build fixes for the Windows and OpenVMS platforms
7692 *Matt Caswell and Richard Levitte*
7694 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7696 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7697 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7698 dereference. This could lead to a Denial Of Service attack. Thanks to
7699 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7704 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7705 dtls1_buffer_record function under certain conditions. In particular this
7706 could occur if an attacker sent repeated DTLS records with the same
7707 sequence number but for the next epoch. The memory leak could be exploited
7708 by an attacker in a Denial of Service attack through memory exhaustion.
7709 Thanks to Chris Mueller for reporting this issue.
7714 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7715 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7716 method would be set to NULL which could later result in a NULL pointer
7717 dereference. Thanks to Frank Schmirler for reporting this issue.
7722 * Abort handshake if server key exchange message is omitted for ephemeral
7725 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7726 reporting this issue.
7731 * Remove non-export ephemeral RSA code on client and server. This code
7732 violated the TLS standard by allowing the use of temporary RSA keys in
7733 non-export ciphersuites and could be used by a server to effectively
7734 downgrade the RSA key length used to a value smaller than the server
7735 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7736 INRIA or reporting this issue.
7741 * Fixed issue where DH client certificates are accepted without verification.
7742 An OpenSSL server will accept a DH certificate for client authentication
7743 without the certificate verify message. This effectively allows a client to
7744 authenticate without the use of a private key. This only affects servers
7745 which trust a client certificate authority which issues certificates
7746 containing DH keys: these are extremely rare and hardly ever encountered.
7747 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7753 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7754 results on some platforms, including x86_64. This bug occurs at random
7755 with a very low probability, and is not known to be exploitable in any
7756 way, though its exact impact is difficult to determine. Thanks to Pieter
7757 Wuille (Blockstream) who reported this issue and also suggested an initial
7758 fix. Further analysis was conducted by the OpenSSL development team and
7759 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7760 the OpenSSL core team.
7765 * Fix various certificate fingerprint issues.
7767 By using non-DER or invalid encodings outside the signed portion of a
7768 certificate the fingerprint can be changed without breaking the signature.
7769 Although no details of the signed portion of the certificate can be changed
7770 this can cause problems with some applications: e.g. those using the
7771 certificate fingerprint for blacklists.
7773 1. Reject signatures with non zero unused bits.
7775 If the BIT STRING containing the signature has non zero unused bits reject
7776 the signature. All current signature algorithms require zero unused bits.
7778 2. Check certificate algorithm consistency.
7780 Check the AlgorithmIdentifier inside TBS matches the one in the
7781 certificate signature. NB: this will result in signature failure
7782 errors for some broken certificates.
7784 Thanks to Konrad Kraszewski from Google for reporting this issue.
7786 3. Check DSA/ECDSA signatures use DER.
7788 Reencode DSA/ECDSA signatures and compare with the original received
7789 signature. Return an error if there is a mismatch.
7791 This will reject various cases including garbage after signature
7792 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7793 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7794 (negative or with leading zeroes).
7796 Further analysis was conducted and fixes were developed by Stephen Henson
7797 of the OpenSSL core team.
7803 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7805 * Session Ticket Memory Leak.
7807 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7808 integrity of that ticket is first verified. In the event of a session
7809 ticket integrity check failing, OpenSSL will fail to free memory
7810 causing a memory leak. By sending a large number of invalid session
7811 tickets an attacker could exploit this issue in a Denial Of Service
7817 * Build option no-ssl3 is incomplete.
7819 When OpenSSL is configured with "no-ssl3" as a build option, servers
7820 could accept and complete a SSL 3.0 handshake, and clients could be
7821 configured to send them.
7824 *Akamai and the OpenSSL team*
7826 * Add support for TLS_FALLBACK_SCSV.
7827 Client applications doing fallback retries should call
7828 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7831 *Adam Langley, Bodo Moeller*
7833 * Add additional DigestInfo checks.
7835 Reencode DigestInto in DER and check against the original when
7836 verifying RSA signature: this will reject any improperly encoded
7837 DigestInfo structures.
7839 Note: this is a precautionary measure and no attacks are currently known.
7843 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7845 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7846 to a denial of service attack. A malicious server can crash the client
7847 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7848 ciphersuite and sending carefully crafted handshake messages.
7850 Thanks to Felix Gröbert (Google) for discovering and researching this
7856 * By sending carefully crafted DTLS packets an attacker could cause openssl
7857 to leak memory. This can be exploited through a Denial of Service attack.
7858 Thanks to Adam Langley for discovering and researching this issue.
7863 * An attacker can force openssl to consume large amounts of memory whilst
7864 processing DTLS handshake messages. This can be exploited through a
7865 Denial of Service attack.
7866 Thanks to Adam Langley for discovering and researching this issue.
7871 * An attacker can force an error condition which causes openssl to crash
7872 whilst processing DTLS packets due to memory being freed twice. This
7873 can be exploited through a Denial of Service attack.
7874 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7880 * If a multithreaded client connects to a malicious server using a resumed
7881 session and the server sends an ec point format extension it could write
7882 up to 255 bytes to freed memory.
7884 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7890 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7891 X509_name_oneline, X509_name_print_ex et al. to leak some information
7892 from the stack. Applications may be affected if they echo pretty printing
7893 output to the attacker.
7895 Thanks to Ivan Fratric (Google) for discovering this issue.
7898 *Emilia Käsper, and Steve Henson*
7900 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7901 for corner cases. (Certain input points at infinity could lead to
7902 bogus results, with non-infinity inputs mapped to infinity too.)
7906 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7908 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7909 handshake can force the use of weak keying material in OpenSSL
7910 SSL/TLS clients and servers.
7912 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7913 researching this issue. ([CVE-2014-0224])
7915 *KIKUCHI Masashi, Steve Henson*
7917 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7918 OpenSSL DTLS client the code can be made to recurse eventually crashing
7921 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7924 *Imre Rad, Steve Henson*
7926 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7927 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7928 client or server. This is potentially exploitable to run arbitrary
7929 code on a vulnerable client or server.
7931 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7933 *Jüri Aedla, Steve Henson*
7935 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7936 are subject to a denial of service attack.
7938 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7939 this issue. ([CVE-2014-3470])
7941 *Felix Gröbert, Ivan Fratric, Steve Henson*
7943 * Harmonize version and its documentation. -f flag is used to display
7946 *mancha <mancha1@zoho.com>*
7948 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7949 in i2d_ECPrivateKey.
7951 *mancha <mancha1@zoho.com>*
7953 * Fix some double frees. These are not thought to be exploitable.
7955 *mancha <mancha1@zoho.com>*
7957 * Fix for the attack described in the paper "Recovering OpenSSL
7958 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7959 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7960 <http://eprint.iacr.org/2014/140>
7962 Thanks to Yuval Yarom and Naomi Benger for discovering this
7963 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7965 *Yuval Yarom and Naomi Benger*
7967 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7969 * Keep original DTLS digest and encryption contexts in retransmission
7970 structures so we can use the previous session parameters if they need
7971 to be resent. ([CVE-2013-6450])
7975 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7976 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7977 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7978 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7979 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7980 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7982 *Rob Stradling, Adam Langley*
7984 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7986 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7988 This addresses the flaw in CBC record processing discovered by
7989 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7990 at: <http://www.isg.rhul.ac.uk/tls/>
7992 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7993 Security Group at Royal Holloway, University of London
7994 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7995 Emilia Käsper for the initial patch.
7998 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8000 * Return an error when checking OCSP signatures when key is NULL.
8001 This fixes a DoS attack. ([CVE-2013-0166])
8005 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8006 the right response is stapled. Also change SSL_get_certificate()
8007 so it returns the certificate actually sent.
8008 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8009 (This is a backport)
8011 *Rob Stradling <rob.stradling@comodo.com>*
8013 * Fix possible deadlock when decoding public keys.
8017 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8019 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8022 * Sanity check record length before skipping explicit IV in DTLS
8025 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8026 fuzzing as a service testing platform.
8031 * Initialise tkeylen properly when encrypting CMS messages.
8032 Thanks to Solar Designer of Openwall for reporting this issue.
8036 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8038 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8039 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8040 in CRYPTO_realloc_clean.
8042 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8043 issue and to Adam Langley <agl@chromium.org> for fixing it.
8046 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8048 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8050 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8051 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8052 content decryption and always return the same error. Note: this attack
8053 needs on average 2^20 messages so it only affects automated senders. The
8054 old behaviour can be re-enabled in the CMS code by setting the
8055 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8056 an MMA defence is not necessary.
8057 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8058 this issue. ([CVE-2012-0884])
8062 * Fix CVE-2011-4619: make sure we really are receiving a
8063 client hello before rejecting multiple SGC restarts. Thanks to
8064 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8068 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8070 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8071 Thanks to Antonio Martin, Enterprise Secure Access Research and
8072 Development, Cisco Systems, Inc. for discovering this bug and
8073 preparing a fix. ([CVE-2012-0050])
8077 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8079 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8080 of the Vaudenay padding oracle attack on CBC mode encryption
8081 which enables an efficient plaintext recovery attack against
8082 the OpenSSL implementation of DTLS. Their attack exploits timing
8083 differences arising during decryption processing. A research
8084 paper describing this attack can be found at:
8085 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8086 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8087 Security Group at Royal Holloway, University of London
8088 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8089 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8090 for preparing the fix. ([CVE-2011-4108])
8092 *Robin Seggelmann, Michael Tuexen*
8094 * Clear bytes used for block padding of SSL 3.0 records.
8097 *Adam Langley (Google)*
8099 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8100 Kadianakis <desnacked@gmail.com> for discovering this issue and
8101 Adam Langley for preparing the fix. ([CVE-2011-4619])
8103 *Adam Langley (Google)*
8105 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8107 *Andrey Kulikov <amdeich@gmail.com>*
8109 * Prevent malformed RFC3779 data triggering an assertion failure.
8110 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8111 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8113 *Rob Austein <sra@hactrn.net>*
8115 * Improved PRNG seeding for VOS.
8117 *Paul Green <Paul.Green@stratus.com>*
8119 * Fix ssl_ciph.c set-up race.
8121 *Adam Langley (Google)*
8123 * Fix spurious failures in ecdsatest.c.
8125 *Emilia Käsper (Google)*
8127 * Fix the BIO_f_buffer() implementation (which was mixing different
8128 interpretations of the `..._len` fields).
8130 *Adam Langley (Google)*
8132 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8133 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8134 threads won't reuse the same blinding coefficients.
8136 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8137 lock to call BN_BLINDING_invert_ex, and avoids one use of
8138 BN_BLINDING_update for each BN_BLINDING structure (previously,
8139 the last update always remained unused).
8141 *Emilia Käsper (Google)*
8143 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8145 *Bob Buckholz (Google)*
8147 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8149 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8150 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8152 *Kaspar Brand <ossl@velox.ch>*
8154 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8155 for multi-threaded use of ECDH. ([CVE-2011-3210])
8157 *Adam Langley (Google)*
8159 * Fix x509_name_ex_d2i memory leak on bad inputs.
8163 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8164 signature public key algorithm by using OID xref utilities instead.
8165 Before this you could only use some ECC ciphersuites with SHA1 only.
8169 * Add protection against ECDSA timing attacks as mentioned in the paper
8170 by Billy Bob Brumley and Nicola Tuveri, see:
8171 <http://eprint.iacr.org/2011/232.pdf>
8173 *Billy Bob Brumley and Nicola Tuveri*
8175 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8177 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8179 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8181 * Fix bug in string printing code: if *any* escaping is enabled we must
8182 escape the escape character (backslash) or the resulting string is
8187 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8189 * Disable code workaround for ancient and obsolete Netscape browsers
8190 and servers: an attacker can use it in a ciphersuite downgrade attack.
8191 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8195 * Fixed J-PAKE implementation error, originally discovered by
8196 Sebastien Martini, further info and confirmation from Stefan
8197 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8201 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8203 * Fix extension code to avoid race conditions which can result in a buffer
8204 overrun vulnerability: resumed sessions must not be modified as they can
8205 be shared by multiple threads. CVE-2010-3864
8209 * Fix WIN32 build system to correctly link an ENGINE directory into
8214 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8216 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8219 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8221 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8223 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8224 context. The operation can be customised via the ctrl mechanism in
8225 case ENGINEs want to include additional functionality.
8229 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8233 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8234 output hashes compatible with older versions of OpenSSL.
8236 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8238 * Fix compression algorithm handling: if resuming a session use the
8239 compression algorithm of the resumed session instead of determining
8240 it from client hello again. Don't allow server to change algorithm.
8244 * Add load_crls() function to commands tidying load_certs() too. Add option
8245 to verify utility to allow additional CRLs to be included.
8249 * Update OCSP request code to permit adding custom headers to the request:
8250 some responders need this.
8254 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8257 *Julia Lawall <julia@diku.dk>*
8259 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8260 needlessly dereferenced structures, used obsolete functions and
8261 didn't handle all updated verify codes correctly.
8265 * Disable MD2 in the default configuration.
8269 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8270 indicate the initial BIO being pushed or popped. This makes it possible
8271 to determine whether the BIO is the one explicitly called or as a result
8272 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8273 it handles reference counts correctly and doesn't zero out the I/O bio
8274 when it is not being explicitly popped. WARNING: applications which
8275 included workarounds for the old buggy behaviour will need to be modified
8276 or they could free up already freed BIOs.
8280 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8281 renaming to all platforms (within the 0.9.8 branch, this was
8282 done conditionally on Netware platforms to avoid a name clash).
8284 *Guenter <lists@gknw.net>*
8286 * Add ECDHE and PSK support to DTLS.
8288 *Michael Tuexen <tuexen@fh-muenster.de>*
8290 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8295 * Add "missing" function EVP_MD_flags() (without this the only way to
8296 retrieve a digest flags is by accessing the structure directly. Update
8297 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8298 or cipher is registered as in the "from" argument. Print out all
8299 registered digests in the dgst usage message instead of manually
8300 attempting to work them out.
8304 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8305 this allows the use of compression and extensions. Change default cipher
8306 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8307 by default unless an application cipher string requests it.
8311 * Alter match criteria in PKCS12_parse(). It used to try to use local
8312 key ids to find matching certificates and keys but some PKCS#12 files
8313 don't follow the (somewhat unwritten) rules and this strategy fails.
8314 Now just gather all certificates together and the first private key
8315 then look for the first certificate that matches the key.
8319 * Support use of registered digest and cipher names for dgst and cipher
8320 commands instead of having to add each one as a special case. So now
8327 openssl dgst -sha256 foo
8329 and this works for ENGINE based algorithms too.
8333 * Update Gost ENGINE to support parameter files.
8335 *Victor B. Wagner <vitus@cryptocom.ru>*
8337 * Support GeneralizedTime in ca utility.
8339 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8341 * Enhance the hash format used for certificate directory links. The new
8342 form uses the canonical encoding (meaning equivalent names will work
8343 even if they aren't identical) and uses SHA1 instead of MD5. This form
8344 is incompatible with the older format and as a result c_rehash should
8345 be used to rebuild symbolic links.
8349 * Make PKCS#8 the default write format for private keys, replacing the
8350 traditional format. This form is standardised, more secure and doesn't
8351 include an implicit MD5 dependency.
8355 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8356 committed to OpenSSL should pass this lot as a minimum.
8360 * Add session ticket override functionality for use by EAP-FAST.
8362 *Jouni Malinen <j@w1.fi>*
8364 * Modify HMAC functions to return a value. Since these can be implemented
8365 in an ENGINE errors can occur.
8369 * Type-checked OBJ_bsearch_ex.
8373 * Type-checked OBJ_bsearch. Also some constification necessitated
8374 by type-checking. Still to come: TXT_DB, bsearch(?),
8375 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8380 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8381 seconds to a tm structure directly, instead of going through OS
8382 specific date routines. This avoids any issues with OS routines such
8383 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8384 and X509_time_adj_ex() to cover the extended range. The existing
8385 X509_time_adj() is still usable and will no longer have any date issues.
8389 * Delta CRL support. New use deltas option which will attempt to locate
8390 and search any appropriate delta CRLs available.
8392 This work was sponsored by Google.
8396 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8397 code and add additional score elements. Validate alternate CRL paths
8398 as part of the CRL checking and indicate a new error "CRL path validation
8399 error" in this case. Applications wanting additional details can use
8400 the verify callback and check the new "parent" field. If this is not
8401 NULL CRL path validation is taking place. Existing applications won't
8402 see this because it requires extended CRL support which is off by
8405 This work was sponsored by Google.
8409 * Support for freshest CRL extension.
8411 This work was sponsored by Google.
8415 * Initial indirect CRL support. Currently only supported in the CRLs
8416 passed directly and not via lookup. Process certificate issuer
8417 CRL entry extension and lookup CRL entries by bother issuer name
8418 and serial number. Check and process CRL issuer entry in IDP extension.
8420 This work was sponsored by Google.
8424 * Add support for distinct certificate and CRL paths. The CRL issuer
8425 certificate is validated separately in this case. Only enabled if
8426 an extended CRL support flag is set: this flag will enable additional
8427 CRL functionality in future.
8429 This work was sponsored by Google.
8433 * Add support for policy mappings extension.
8435 This work was sponsored by Google.
8439 * Fixes to pathlength constraint, self issued certificate handling,
8440 policy processing to align with RFC3280 and PKITS tests.
8442 This work was sponsored by Google.
8446 * Support for name constraints certificate extension. DN, email, DNS
8447 and URI types are currently supported.
8449 This work was sponsored by Google.
8453 * To cater for systems that provide a pointer-based thread ID rather
8454 than numeric, deprecate the current numeric thread ID mechanism and
8455 replace it with a structure and associated callback type. This
8456 mechanism allows a numeric "hash" to be extracted from a thread ID in
8457 either case, and on platforms where pointers are larger than 'long',
8458 mixing is done to help ensure the numeric 'hash' is usable even if it
8459 can't be guaranteed unique. The default mechanism is to use "&errno"
8460 as a pointer-based thread ID to distinguish between threads.
8462 Applications that want to provide their own thread IDs should now use
8463 CRYPTO_THREADID_set_callback() to register a callback that will call
8464 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8466 Note that ERR_remove_state() is now deprecated, because it is tied
8467 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8468 to free the current thread's error state should be replaced by
8469 ERR_remove_thread_state(NULL).
8471 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8472 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8473 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8474 application was previously providing a numeric thread callback that
8475 was inappropriate for distinguishing threads, then uniqueness might
8476 have been obtained with &errno that happened immediately in the
8477 intermediate development versions of OpenSSL; this is no longer the
8478 case, the numeric thread callback will now override the automatic use
8481 *Geoff Thorpe, with help from Bodo Moeller*
8483 * Initial support for different CRL issuing certificates. This covers a
8484 simple case where the self issued certificates in the chain exist and
8485 the real CRL issuer is higher in the existing chain.
8487 This work was sponsored by Google.
8491 * Removed effectively defunct crypto/store from the build.
8495 * Revamp of STACK to provide stronger type-checking. Still to come:
8496 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8497 ASN1_STRING, CONF_VALUE.
8501 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8502 RAM on SSL connections. This option can save about 34k per idle SSL.
8506 * Revamp of LHASH to provide stronger type-checking. Still to come:
8507 STACK, TXT_DB, bsearch, qsort.
8511 * Initial support for Cryptographic Message Syntax (aka CMS) based
8512 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8513 support for data, signedData, compressedData, digestedData and
8514 encryptedData, envelopedData types included. Scripts to check against
8515 RFC4134 examples draft and interop and consistency checks of many
8516 content types and variants.
8520 * Add options to enc utility to support use of zlib compression BIO.
8524 * Extend mk1mf to support importing of options and assembly language
8525 files from Configure script, currently only included in VC-WIN32.
8526 The assembly language rules can now optionally generate the source
8527 files from the associated perl scripts.
8531 * Implement remaining functionality needed to support GOST ciphersuites.
8532 Interop testing has been performed using CryptoPro implementations.
8534 *Victor B. Wagner <vitus@cryptocom.ru>*
8536 * s390x assembler pack.
8540 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8545 * Implement Opaque PRF Input TLS extension as specified in
8546 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8547 official specification yet and no extension type assignment by
8548 IANA exists, this extension (for now) will have to be explicitly
8549 enabled when building OpenSSL by providing the extension number
8550 to use. For example, specify an option
8552 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8554 to the "config" or "Configure" script to enable the extension,
8555 assuming extension number 0x9527 (which is a completely arbitrary
8556 and unofficial assignment based on the MD5 hash of the Internet
8557 Draft). Note that by doing so, you potentially lose
8558 interoperability with other TLS implementations since these might
8559 be using the same extension number for other purposes.
8561 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8562 opaque PRF input value to use in the handshake. This will create
8563 an internal copy of the length-'len' string at 'src', and will
8564 return non-zero for success.
8566 To get more control and flexibility, provide a callback function
8569 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8570 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8574 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8577 Callback function 'cb' will be called in handshakes, and is
8578 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8579 Argument 'arg' is for application purposes (the value as given to
8580 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8581 be provided to the callback function). The callback function
8582 has to return non-zero to report success: usually 1 to use opaque
8583 PRF input just if possible, or 2 to enforce use of the opaque PRF
8584 input. In the latter case, the library will abort the handshake
8585 if opaque PRF input is not successfully negotiated.
8587 Arguments 'peerinput' and 'len' given to the callback function
8588 will always be NULL and 0 in the case of a client. A server will
8589 see the client's opaque PRF input through these variables if
8590 available (NULL and 0 otherwise). Note that if the server
8591 provides an opaque PRF input, the length must be the same as the
8592 length of the client's opaque PRF input.
8594 Note that the callback function will only be called when creating
8595 a new session (session resumption can resume whatever was
8596 previously negotiated), and will not be called in SSL 2.0
8597 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8598 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8599 for applications that need to enforce opaque PRF input.
8603 * Update ssl code to support digests other than SHA1+MD5 for handshake
8606 *Victor B. Wagner <vitus@cryptocom.ru>*
8608 * Add RFC4507 support to OpenSSL. This includes the corrections in
8609 RFC4507bis. The encrypted ticket format is an encrypted encoded
8610 SSL_SESSION structure, that way new session features are automatically
8613 If a client application caches session in an SSL_SESSION structure
8614 support is transparent because tickets are now stored in the encoded
8617 The SSL_CTX structure automatically generates keys for ticket
8618 protection in servers so again support should be possible
8619 with no application modification.
8621 If a client or server wishes to disable RFC4507 support then the option
8622 SSL_OP_NO_TICKET can be set.
8624 Add a TLS extension debugging callback to allow the contents of any client
8625 or server extensions to be examined.
8627 This work was sponsored by Google.
8631 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8632 OpenSSL should now compile cleanly on gcc 4.2
8634 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8636 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8637 support including streaming MAC support: this is required for GOST
8638 ciphersuite support.
8640 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8642 * Add option -stream to use PKCS#7 streaming in smime utility. New
8643 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8644 to output in BER and PEM format.
8648 * Experimental support for use of HMAC via EVP_PKEY interface. This
8649 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8650 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8651 ENGINE support for HMAC keys which are unextractable. New -mac and
8652 -macopt options to dgst utility.
8656 * New option -sigopt to dgst utility. Update dgst to use
8657 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8658 alternative signing parameters such as X9.31 or PSS in the dgst
8663 * Change ssl_cipher_apply_rule(), the internal function that does
8664 the work each time a ciphersuite string requests enabling
8665 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8666 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8667 the order of disabled ciphersuites such that those ciphersuites
8668 that most recently went from enabled to disabled not only stay
8669 in order with respect to each other, but also have higher priority
8670 than other disabled ciphersuites the next time ciphersuites are
8673 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8674 the same ciphersuites as with "HIGH" alone, but in a specific
8675 order where the PSK ciphersuites come first (since they are the
8676 most recently disabled ciphersuites when "HIGH" is parsed).
8678 Also, change ssl_create_cipher_list() (using this new
8679 functionality) such that between otherwise identical
8680 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8685 * Change ssl_create_cipher_list() so that it automatically
8686 arranges the ciphersuites in reasonable order before starting
8687 to process the rule string. Thus, the definition for "DEFAULT"
8688 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8689 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8690 This makes it much easier to arrive at a reasonable default order
8691 in applications for which anonymous ciphers are OK (meaning
8692 that you can't actually use DEFAULT).
8694 *Bodo Moeller; suggested by Victor Duchovni*
8696 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8697 processing) into multiple integers instead of setting
8698 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8699 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8700 (These masks as well as the individual bit definitions are hidden
8701 away into the non-exported interface ssl/ssl_locl.h, so this
8702 change to the definition of the SSL_CIPHER structure shouldn't
8703 affect applications.) This give us more bits for each of these
8704 categories, so there is no longer a need to coagulate AES128 and
8705 AES256 into a single algorithm bit, and to coagulate Camellia128
8706 and Camellia256 into a single algorithm bit, which has led to all
8709 Thus, among other things, the kludge introduced in 0.9.7m and
8710 0.9.8e for masking out AES256 independently of AES128 or masking
8711 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8713 With the change, we also introduce new ciphersuite aliases that
8714 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8719 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8720 Use the leftmost N bytes of the signature input if the input is
8721 larger than the prime q (with N being the size in bytes of q).
8725 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8726 it yet and it is largely untested.
8730 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8734 * Initial incomplete changes to avoid need for function casts in OpenSSL
8735 some compilers (gcc 4.2 and later) reject their use. Safestack is
8736 reimplemented. Update ASN1 to avoid use of legacy functions.
8740 * Win32/64 targets are linked with Winsock2.
8744 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8745 to external functions. This can be used to increase CRL handling
8746 efficiency especially when CRLs are very large by (for example) storing
8747 the CRL revoked certificates in a database.
8751 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8752 new CRLs added to a directory can be used. New command line option
8753 -verify_return_error to s_client and s_server. This causes real errors
8754 to be returned by the verify callback instead of carrying on no matter
8755 what. This reflects the way a "real world" verify callback would behave.
8759 * GOST engine, supporting several GOST algorithms and public key formats.
8760 Kindly donated by Cryptocom.
8764 * Partial support for Issuing Distribution Point CRL extension. CRLs
8765 partitioned by DP are handled but no indirect CRL or reason partitioning
8766 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8767 selected via a scoring technique which handles IDP and AKID in CRLs.
8771 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8772 will ultimately be used for all verify operations: this will remove the
8773 X509_STORE dependency on certificate verification and allow alternative
8774 lookup methods. X509_STORE based implementations of these two callbacks.
8778 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8779 Modify get_crl() to find a valid (unexpired) CRL if possible.
8783 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8784 this would be called X509_CRL_cmp() but that name is already used by
8785 a function that just compares CRL issuer names. Cache several CRL
8786 extensions in X509_CRL structure and cache CRLDP in X509.
8790 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8791 this maps equivalent X509_NAME structures into a consistent structure.
8792 Name comparison can then be performed rapidly using memcmp().
8796 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8801 * Allow digests to supply their own micalg string for S/MIME type using
8802 the ctrl EVP_MD_CTRL_MICALG.
8806 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8807 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8808 ctrl. It can then customise the structure before and/or after signing
8813 * New function OBJ_add_sigid() to allow application defined signature OIDs
8814 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8815 to free up any added signature OIDs.
8819 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8820 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8821 digest and cipher tables. New options added to openssl utility:
8822 list-message-digest-algorithms and list-cipher-algorithms.
8826 * Change the array representation of binary polynomials: the list
8827 of degrees of non-zero coefficients is now terminated with -1.
8828 Previously it was terminated with 0, which was also part of the
8829 value; thus, the array representation was not applicable to
8830 polynomials where t^0 has coefficient zero. This change makes
8831 the array representation useful in a more general context.
8835 * Various modifications and fixes to SSL/TLS cipher string
8836 handling. For ECC, the code now distinguishes between fixed ECDH
8837 with RSA certificates on the one hand and with ECDSA certificates
8838 on the other hand, since these are separate ciphersuites. The
8839 unused code for Fortezza ciphersuites has been removed.
8841 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8842 (not "ECDHE"). For consistency with the code for DH
8843 certificates, use of ECDH certificates is now considered ECDH
8844 authentication, not RSA or ECDSA authentication (the latter is
8845 merely the CA's signing algorithm and not actively used in the
8848 The temporary ciphersuite alias "ECCdraft" is no longer
8849 available, and ECC ciphersuites are no longer excluded from "ALL"
8850 and "DEFAULT". The following aliases now exist for RFC 4492
8851 ciphersuites, most of these by analogy with the DH case:
8853 kECDHr - ECDH cert, signed with RSA
8854 kECDHe - ECDH cert, signed with ECDSA
8855 kECDH - ECDH cert (signed with either RSA or ECDSA)
8856 kEECDH - ephemeral ECDH
8857 ECDH - ECDH cert or ephemeral ECDH
8863 AECDH - anonymous ECDH
8864 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8868 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8869 Use correct micalg parameters depending on digest(s) in signed message.
8873 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8874 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8878 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8879 an engine to register a method. Add ENGINE lookups for methods and
8880 functional reference processing.
8884 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8885 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8890 * New -resign option to smime utility. This adds one or more signers
8891 to an existing PKCS#7 signedData structure. Also -md option to use an
8892 alternative message digest algorithm for signing.
8896 * Tidy up PKCS#7 routines and add new functions to make it easier to
8897 create PKCS7 structures containing multiple signers. Update smime
8898 application to support multiple signers.
8902 * New -macalg option to pkcs12 utility to allow setting of an alternative
8907 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8908 Reorganize PBE internals to lookup from a static table using NIDs,
8909 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8910 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8911 PRF which will be automatically used with PBES2.
8915 * Replace the algorithm specific calls to generate keys in "req" with the
8920 * Update PKCS#7 enveloped data routines to use new API. This is now
8921 supported by any public key method supporting the encrypt operation. A
8922 ctrl is added to allow the public key algorithm to examine or modify
8923 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8928 * Add a ctrl to asn1 method to allow a public key algorithm to express
8929 a default digest type to use. In most cases this will be SHA1 but some
8930 algorithms (such as GOST) need to specify an alternative digest. The
8931 return value indicates how strong the preference is 1 means optional and
8932 2 is mandatory (that is it is the only supported type). Modify
8933 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8934 use the default md. Update openssl utilities to use the default digest
8935 type for signing if it is not explicitly indicated.
8939 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8940 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8941 signing method from the key type. This effectively removes the link
8942 between digests and public key types.
8946 * Add an OID cross reference table and utility functions. Its purpose is to
8947 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8948 rsaEncryption. This will allow some of the algorithm specific hackery
8949 needed to use the correct OID to be removed.
8953 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8954 structures for PKCS7_sign(). They are now set up by the relevant public
8959 * Add provisional EC pkey method with support for ECDSA and ECDH.
8963 * Add support for key derivation (agreement) in the API, DH method and
8968 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8969 public and private key formats. As a side effect these add additional
8970 command line functionality not previously available: DSA signatures can be
8971 generated and verified using pkeyutl and DH key support and generation in
8978 *Oliver Tappe <zooey@hirschkaefer.de>*
8980 * New make target "install_html_docs" installs HTML renditions of the
8983 *Oliver Tappe <zooey@hirschkaefer.de>*
8985 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8986 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8987 support key and parameter generation and add initial key generation
8988 functionality for RSA.
8992 * Add functions for main EVP_PKEY_method operations. The undocumented
8993 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8994 `EVP_PKEY_{encrypt,decrypt}_old`.
8998 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8999 key API, doesn't do much yet.
9003 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9004 public key algorithms. New option to openssl utility:
9005 "list-public-key-algorithms" to print out info.
9009 * Implement the Supported Elliptic Curves Extension for
9010 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9014 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9015 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9019 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9020 utilities such as rsa, dsa, dsaparam etc except they process any key
9025 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9026 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9027 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9032 * Initial support for pluggable public key ASN1.
9033 De-spaghettify the public key ASN1 handling. Move public and private
9034 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9035 algorithm specific handling to a single module within the relevant
9036 algorithm directory. Add functions to allow (near) opaque processing
9037 of public and private key structures.
9041 * Implement the Supported Point Formats Extension for
9042 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9046 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9047 for the psk identity [hint] and the psk callback functions to the
9048 SSL_SESSION, SSL and SSL_CTX structure.
9051 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9055 SSL_CTX_use_psk_identity_hint
9056 SSL_get_psk_identity_hint
9057 SSL_get_psk_identity
9058 SSL_use_psk_identity_hint
9060 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9062 * Add RFC 3161 compliant time stamp request creation, response generation
9063 and response verification functionality.
9065 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9067 * Add initial support for TLS extensions, specifically for the server_name
9068 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9069 have new members for a host name. The SSL data structure has an
9070 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9071 stored in that context to allow for session resumption, even after the
9072 SSL has been switched to a new SSL_CTX in reaction to a client's
9073 server_name extension.
9075 New functions (subject to change):
9077 SSL_get_servername()
9078 SSL_get_servername_type()
9081 New CTRL codes and macros (subject to change):
9083 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9084 - SSL_CTX_set_tlsext_servername_callback()
9085 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9086 - SSL_CTX_set_tlsext_servername_arg()
9087 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9089 openssl s_client has a new '-servername ...' option.
9091 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9092 '-key2 ...', '-servername_fatal' (subject to change). This allows
9093 testing the HostName extension for a specific single host name ('-cert'
9094 and '-key' remain fallbacks for handshakes without HostName
9095 negotiation). If the unrecognized_name alert has to be sent, this by
9096 default is a warning; it becomes fatal with the '-servername_fatal'
9099 *Peter Sylvester, Remy Allais, Christophe Renou*
9101 * Whirlpool hash implementation is added.
9105 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9106 bn(64,32). Because of instruction set limitations it doesn't have
9107 any negative impact on performance. This was done mostly in order
9108 to make it possible to share assembler modules, such as bn_mul_mont
9109 implementations, between 32- and 64-bit builds without hassle.
9113 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9114 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9119 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9120 dedicated Montgomery multiplication procedure, is introduced.
9121 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9122 "64-bit" performance on certain 32-bit targets.
9126 * New option SSL_OP_NO_COMP to disable use of compression selectively
9127 in SSL structures. New SSL ctrl to set maximum send fragment size.
9128 Save memory by setting the I/O buffer sizes dynamically instead of
9129 using the maximum available value.
9133 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9134 in addition to the text details.
9138 * Very, very preliminary EXPERIMENTAL support for printing of general
9139 ASN1 structures. This currently produces rather ugly output and doesn't
9140 handle several customised structures at all.
9144 * Integrated support for PVK file format and some related formats such
9145 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9146 these in the 'rsa' and 'dsa' utilities.
9150 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9154 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9155 place for the (very old) "NETSCAPE" format certificates which are now
9156 handled using new ASN1 code equivalents.
9160 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9161 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9162 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9166 * Modify CRL distribution points extension code to print out previously
9167 unsupported fields. Enhance extension setting code to allow setting of
9172 * Add print and set support for Issuing Distribution Point CRL extension.
9176 * Change 'Configure' script to enable Camellia by default.
9183 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9185 * When rejecting SSL/TLS records due to an incorrect version number, never
9186 update s->server with a new major version number. As of
9187 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9188 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9189 the previous behavior could result in a read attempt at NULL when
9190 receiving specific incorrect SSL/TLS records once record payload
9191 protection is active. ([CVE-2010-0740])
9193 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9195 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9196 could be crashed if the relevant tables were not present (e.g. chrooted).
9198 *Tomas Hoger <thoger@redhat.com>*
9200 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9202 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9204 *Martin Olsson, Neel Mehta*
9206 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9207 accommodate for stack sorting, always a write lock!).
9211 * On some versions of WIN32 Heap32Next is very slow. This can cause
9212 excessive delays in the RAND_poll(): over a minute. As a workaround
9213 include a time check in the inner Heap32Next loop too.
9217 * The code that handled flushing of data in SSL/TLS originally used the
9218 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9219 the problem outlined in PR#1949. The fix suggested there however can
9220 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9221 of Apache). So instead simplify the code to flush unconditionally.
9222 This should be fine since flushing with no data to flush is a no op.
9226 * Handle TLS versions 2.0 and later properly and correctly use the
9227 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9228 off ancient servers have a habit of sticking around for a while...
9232 * Modify compression code so it frees up structures without using the
9233 ex_data callbacks. This works around a problem where some applications
9234 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9235 restarting) then use compression (e.g. SSL with compression) later.
9236 This results in significant per-connection memory leaks and
9237 has caused some security issues including CVE-2008-1678 and
9242 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9243 change when encrypting or decrypting.
9247 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9248 connect and renegotiate with servers which do not support RI.
9249 Until RI is more widely deployed this option is enabled by default.
9253 * Add "missing" ssl ctrls to clear options and mode.
9257 * If client attempts to renegotiate and doesn't support RI respond with
9258 a no_renegotiation alert as required by RFC5746. Some renegotiating
9259 TLS clients will continue a connection gracefully when they receive
9260 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9261 waiting for a server hello which it will never receive. Now we treat a
9262 received no_renegotiation alert as a fatal error. This is because
9263 applications requesting a renegotiation might well expect it to succeed
9264 and would have no code in place to handle the server denying it so the
9265 only safe thing to do is to terminate the connection.
9269 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9270 peer supports secure renegotiation and 0 otherwise. Print out peer
9271 renegotiation support in s_client/s_server.
9275 * Replace the highly broken and deprecated SPKAC certification method with
9276 the updated NID creation version. This should correctly handle UTF8.
9280 * Implement RFC5746. Re-enable renegotiation but require the extension
9281 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9282 turns out to be a bad idea. It has been replaced by
9283 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9284 SSL_CTX_set_options(). This is really not recommended unless you
9285 know what you are doing.
9287 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9289 * Fixes to stateless session resumption handling. Use initial_ctx when
9290 issuing and attempting to decrypt tickets in case it has changed during
9291 servername handling. Use a non-zero length session ID when attempting
9292 stateless session resumption: this makes it possible to determine if
9293 a resumption has occurred immediately after receiving server hello
9294 (several places in OpenSSL subtly assume this) instead of later in
9299 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9300 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9301 fixes for a few places where the return code is not checked
9304 *Julia Lawall <julia@diku.dk>*
9306 * Add --strict-warnings option to Configure script to include devteam
9307 warnings in other configurations.
9311 * Add support for --libdir option and LIBDIR variable in makefiles. This
9312 makes it possible to install openssl libraries in locations which
9313 have names other than "lib", for example "/usr/lib64" which some
9316 *Steve Henson, based on patch from Jeremy Utley*
9318 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9319 X690 8.9.12 and can produce some misleading textual output of OIDs.
9321 *Steve Henson, reported by Dan Kaminsky*
9323 * Delete MD2 from algorithm tables. This follows the recommendation in
9324 several standards that it is not used in new applications due to
9325 several cryptographic weaknesses. For binary compatibility reasons
9326 the MD2 API is still compiled in by default.
9330 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9335 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9336 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9339 *Guenter <lists@gknw.net>*
9341 * Fix the server certificate chain building code to use X509_verify_cert(),
9342 it used to have an ad-hoc builder which was unable to cope with anything
9343 other than a simple chain.
9345 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9347 * Don't check self signed certificate signatures in X509_verify_cert()
9348 by default (a flag can override this): it just wastes time without
9349 adding any security. As a useful side effect self signed root CAs
9350 with non-FIPS digests are now usable in FIPS mode.
9354 * In dtls1_process_out_of_seq_message() the check if the current message
9355 is already buffered was missing. For every new message was memory
9356 allocated, allowing an attacker to perform an denial of service attack
9357 with sending out of seq handshake messages until there is no memory
9358 left. Additionally every future message was buffered, even if the
9359 sequence number made no sense and would be part of another handshake.
9360 So only messages with sequence numbers less than 10 in advance will be
9361 buffered. ([CVE-2009-1378])
9363 *Robin Seggelmann, discovered by Daniel Mentz*
9365 * Records are buffered if they arrive with a future epoch to be
9366 processed after finishing the corresponding handshake. There is
9367 currently no limitation to this buffer allowing an attacker to perform
9368 a DOS attack with sending records with future epochs until there is no
9369 memory left. This patch adds the pqueue_size() function to determine
9370 the size of a buffer and limits the record buffer to 100 entries.
9373 *Robin Seggelmann, discovered by Daniel Mentz*
9375 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9376 parent structure is freed. ([CVE-2009-1379])
9380 * Handle non-blocking I/O properly in SSL_shutdown() call.
9382 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9384 * Add `2.5.4.*` OIDs
9386 *Ilya O. <vrghost@gmail.com>*
9388 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9390 * Disable renegotiation completely - this fixes a severe security
9391 problem ([CVE-2009-3555]) at the cost of breaking all
9392 renegotiation. Renegotiation can be re-enabled by setting
9393 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9394 run-time. This is really not recommended unless you know what
9399 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9401 * Don't set val to NULL when freeing up structures, it is freed up by
9402 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9403 zeroing past the valid field. ([CVE-2009-0789])
9405 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9407 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9408 checked correctly. This would allow some invalid signed attributes to
9409 appear to verify correctly. ([CVE-2009-0591])
9411 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9413 * Reject UniversalString and BMPString types with invalid lengths. This
9414 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9415 a legal length. ([CVE-2009-0590])
9419 * Set S/MIME signing as the default purpose rather than setting it
9420 unconditionally. This allows applications to override it at the store
9425 * Permit restricted recursion of ASN1 strings. This is needed in practice
9426 to handle some structures.
9430 * Improve efficiency of mem_gets: don't search whole buffer each time
9433 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9435 * New -hex option for openssl rand.
9439 * Print out UTF8String and NumericString when parsing ASN1.
9443 * Support NumericString type for name components.
9447 * Allow CC in the environment to override the automatically chosen
9448 compiler. Note that nothing is done to ensure flags work with the
9453 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9455 * Properly check EVP_VerifyFinal() and similar return values
9458 *Ben Laurie, Bodo Moeller, Google Security Team*
9460 * Enable TLS extensions by default.
9464 * Allow the CHIL engine to be loaded, whether the application is
9465 multithreaded or not. (This does not release the developer from the
9466 obligation to set up the dynamic locking callbacks.)
9468 *Sander Temme <sander@temme.net>*
9470 * Use correct exit code if there is an error in dgst command.
9472 *Steve Henson; problem pointed out by Roland Dirlewanger*
9474 * Tweak Configure so that you need to say "experimental-jpake" to enable
9475 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9479 * Add experimental JPAKE support, including demo authentication in
9480 s_client and s_server.
9484 * Set the comparison function in v3_addr_canonize().
9486 *Rob Austein <sra@hactrn.net>*
9488 * Add support for XMPP STARTTLS in s_client.
9490 *Philip Paeps <philip@freebsd.org>*
9492 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9493 to ensure that even with this option, only ciphersuites in the
9494 server's preference list will be accepted. (Note that the option
9495 applies only when resuming a session, so the earlier behavior was
9496 just about the algorithm choice for symmetric cryptography.)
9500 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9502 * Fix NULL pointer dereference if a DTLS server received
9503 ChangeCipherSpec as first record ([CVE-2009-1386]).
9507 * Fix a state transition in s3_srvr.c and d1_srvr.c
9508 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9512 * The fix in 0.9.8c that supposedly got rid of unsafe
9513 double-checked locking was incomplete for RSA blinding,
9514 addressing just one layer of what turns out to have been
9515 doubly unsafe triple-checked locking.
9517 So now fix this for real by retiring the MONT_HELPER macro
9518 in crypto/rsa/rsa_eay.c.
9520 *Bodo Moeller; problem pointed out by Marius Schilder*
9522 * Various precautionary measures:
9524 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9526 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9527 (NB: This would require knowledge of the secret session ticket key
9528 to exploit, in which case you'd be SOL either way.)
9530 - Change bn_nist.c so that it will properly handle input BIGNUMs
9531 outside the expected range.
9533 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9536 *Neel Mehta, Bodo Moeller*
9538 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9539 the load fails. Useful for distros.
9541 *Ben Laurie and the FreeBSD team*
9543 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9547 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9551 * Expand ENGINE to support engine supplied SSL client certificate functions.
9553 This work was sponsored by Logica.
9557 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9558 keystores. Support for SSL/TLS client authentication too.
9559 Not compiled unless enable-capieng specified to Configure.
9561 This work was sponsored by Logica.
9565 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9566 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9567 attribute creation routines such as certificate requests and PKCS#12
9572 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9574 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9575 handshake which could lead to a client crash as found using the
9576 Codenomicon TLS test suite ([CVE-2008-1672])
9578 *Steve Henson, Mark Cox*
9580 * Fix double free in TLS server name extensions which could lead to
9581 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9585 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9587 Clear the error queue to ensure that error entries left from
9588 older function calls do not interfere with the correct operation.
9590 *Lutz Jaenicke, Erik de Castro Lopo*
9592 * Remove root CA certificates of commercial CAs:
9594 The OpenSSL project does not recommend any specific CA and does not
9595 have any policy with respect to including or excluding any CA.
9596 Therefore it does not make any sense to ship an arbitrary selection
9597 of root CA certificates with the OpenSSL software.
9601 * RSA OAEP patches to fix two separate invalid memory reads.
9602 The first one involves inputs when 'lzero' is greater than
9603 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9604 before the beginning of from). The second one involves inputs where
9605 the 'db' section contains nothing but zeroes (there is a one-byte
9606 invalid read after the end of 'db').
9608 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9610 * Partial backport from 0.9.9-dev:
9612 Introduce bn_mul_mont (dedicated Montgomery multiplication
9613 procedure) as a candidate for BIGNUM assembler implementation.
9614 While 0.9.9-dev uses assembler for various architectures, only
9615 x86_64 is available by default here in the 0.9.8 branch, and
9616 32-bit x86 is available through a compile-time setting.
9618 To try the 32-bit x86 assembler implementation, use Configure
9619 option "enable-montasm" (which exists only for this backport).
9621 As "enable-montasm" for 32-bit x86 disclaims code stability
9622 anyway, in this constellation we activate additional code
9623 backported from 0.9.9-dev for further performance improvements,
9624 namely BN_from_montgomery_word. (To enable this otherwise,
9625 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9627 *Andy Polyakov (backport partially by Bodo Moeller)*
9629 * Add TLS session ticket callback. This allows an application to set
9630 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9631 values. This is useful for key rollover for example where several key
9632 sets may exist with different names.
9636 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9637 This was broken until now in 0.9.8 releases, such that the only way
9638 a registered ENGINE could be used (assuming it initialises
9639 successfully on the host) was to explicitly set it as the default
9640 for the relevant algorithms. This is in contradiction with 0.9.7
9641 behaviour and the documentation. With this fix, when an ENGINE is
9642 registered into a given algorithm's table of implementations, the
9643 'uptodate' flag is reset so that auto-discovery will be used next
9644 time a new context for that algorithm attempts to select an
9647 *Ian Lister (tweaked by Geoff Thorpe)*
9649 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9650 implementation in the following ways:
9652 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9655 Lack of BER streaming support means one pass streaming processing is
9656 only supported if data is detached: setting the streaming flag is
9657 ignored for embedded content.
9659 CMS support is disabled by default and must be explicitly enabled
9660 with the enable-cms configuration option.
9664 * Update the GMP engine glue to do direct copies between BIGNUM and
9665 mpz_t when openssl and GMP use the same limb size. Otherwise the
9666 existing "conversion via a text string export" trick is still used.
9668 *Paul Sheer <paulsheer@gmail.com>*
9670 * Zlib compression BIO. This is a filter BIO which compressed and
9671 uncompresses any data passed through it.
9675 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9676 RFC3394 compatible AES key wrapping.
9680 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9681 sets string data without copying. X509_ALGOR_set0() and
9682 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9683 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9684 from an X509_ATTRIBUTE structure optionally checking it occurs only
9685 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9690 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9691 to get the expected BN_FLG_CONSTTIME behavior.
9693 *Bodo Moeller (Google)*
9697 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9698 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9699 - added some more tests to do_tests.pl
9700 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9701 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9702 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9703 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9704 - various changes to netware.pl to enable gcc-cross builds on Win32
9706 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9707 - various changes to fix missing prototype warnings
9708 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9709 - added AES, WHIRLPOOL and CPUID assembler code to build files
9710 - added missing AES assembler make rules to mk1mf.pl
9711 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9713 *Guenter Knauf <eflash@gmx.net>*
9715 * Implement certificate status request TLS extension defined in RFC3546.
9716 A client can set the appropriate parameters and receive the encoded
9717 OCSP response via a callback. A server can query the supplied parameters
9718 and set the encoded OCSP response in the callback. Add simplified examples
9719 to s_client and s_server.
9723 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9726 + Binary incompatibility of ssl_ctx_st structure
9727 + DTLS interoperation with non-compliant servers
9728 + Don't call get_session_cb() without proposed session
9729 + Fix ia64 assembler code
9731 *Andy Polyakov, Steve Henson*
9733 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9735 * DTLS Handshake overhaul. There were longstanding issues with
9736 OpenSSL DTLS implementation, which were making it impossible for
9737 RFC 4347 compliant client to communicate with OpenSSL server.
9738 Unfortunately just fixing these incompatibilities would "cut off"
9739 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9740 server keeps tolerating non RFC compliant syntax. The opposite is
9741 not true, 0.9.8f client can not communicate with earlier server.
9742 This update even addresses CVE-2007-4995.
9746 * Changes to avoid need for function casts in OpenSSL: some compilers
9747 (gcc 4.2 and later) reject their use.
9748 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9751 * Add RFC4507 support to OpenSSL. This includes the corrections in
9752 RFC4507bis. The encrypted ticket format is an encrypted encoded
9753 SSL_SESSION structure, that way new session features are automatically
9756 If a client application caches session in an SSL_SESSION structure
9757 support is transparent because tickets are now stored in the encoded
9760 The SSL_CTX structure automatically generates keys for ticket
9761 protection in servers so again support should be possible
9762 with no application modification.
9764 If a client or server wishes to disable RFC4507 support then the option
9765 SSL_OP_NO_TICKET can be set.
9767 Add a TLS extension debugging callback to allow the contents of any client
9768 or server extensions to be examined.
9770 This work was sponsored by Google.
9774 * Add initial support for TLS extensions, specifically for the server_name
9775 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9776 have new members for a host name. The SSL data structure has an
9777 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9778 stored in that context to allow for session resumption, even after the
9779 SSL has been switched to a new SSL_CTX in reaction to a client's
9780 server_name extension.
9782 New functions (subject to change):
9784 SSL_get_servername()
9785 SSL_get_servername_type()
9788 New CTRL codes and macros (subject to change):
9790 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9791 - SSL_CTX_set_tlsext_servername_callback()
9792 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9793 - SSL_CTX_set_tlsext_servername_arg()
9794 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9796 openssl s_client has a new '-servername ...' option.
9798 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9799 '-key2 ...', '-servername_fatal' (subject to change). This allows
9800 testing the HostName extension for a specific single host name ('-cert'
9801 and '-key' remain fallbacks for handshakes without HostName
9802 negotiation). If the unrecognized_name alert has to be sent, this by
9803 default is a warning; it becomes fatal with the '-servername_fatal'
9806 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9808 * Add AES and SSE2 assembly language support to VC++ build.
9812 * Mitigate attack on final subtraction in Montgomery reduction.
9816 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9817 (which previously caused an internal error).
9821 * Squeeze another 10% out of IGE mode when in != out.
9825 * AES IGE mode speedup.
9827 *Dean Gaudet (Google)*
9829 * Add the Korean symmetric 128-bit cipher SEED (see
9830 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9831 add SEED ciphersuites from RFC 4162:
9833 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9834 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9835 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9836 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9838 To minimize changes between patchlevels in the OpenSSL 0.9.8
9839 series, SEED remains excluded from compilation unless OpenSSL
9840 is configured with 'enable-seed'.
9842 *KISA, Bodo Moeller*
9844 * Mitigate branch prediction attacks, which can be practical if a
9845 single processor is shared, allowing a spy process to extract
9846 information. For detailed background information, see
9847 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9848 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9849 and Necessary Software Countermeasures"). The core of the change
9850 are new versions BN_div_no_branch() and
9851 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9852 respectively, which are slower, but avoid the security-relevant
9853 conditional branches. These are automatically called by BN_div()
9854 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9855 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9856 remove a conditional branch.
9858 BN_FLG_CONSTTIME is the new name for the previous
9859 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9860 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9861 in the exponent causes BN_mod_exp_mont() to use the alternative
9862 implementation in BN_mod_exp_mont_consttime().) The old name
9863 remains as a deprecated alias.
9865 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9866 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9867 constant-time implementations for more than just exponentiation.
9868 Here too the old name is kept as a deprecated alias.
9870 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9871 the BN_BLINDING structure gets an independent copy of the
9872 modulus. This means that the previous `BIGNUM *m` argument to
9873 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9874 essentially becomes `const BIGNUM *m`, although we can't actually
9875 change this in the header file before 0.9.9. It allows
9876 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9877 enable BN_FLG_CONSTTIME.
9879 *Matthew D Wood (Intel Corp)*
9881 * In the SSL/TLS server implementation, be strict about session ID
9882 context matching (which matters if an application uses a single
9883 external cache for different purposes). Previously,
9884 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9885 set. This did ensure strict client verification, but meant that,
9886 with applications using a single external cache for quite
9887 different requirements, clients could circumvent ciphersuite
9888 restrictions for a given session ID context by starting a session
9889 in a different context.
9893 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9894 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9895 authentication-only ciphersuites.
9899 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9900 not complete and could lead to a possible single byte overflow
9901 ([CVE-2007-5135]) [Ben Laurie]
9903 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9905 * Since AES128 and AES256 (and similarly Camellia128 and
9906 Camellia256) share a single mask bit in the logic of
9907 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9908 kludge to work properly if AES128 is available and AES256 isn't
9909 (or if Camellia128 is available and Camellia256 isn't).
9913 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9914 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9915 When a point or a seed is encoded in a BIT STRING, we need to
9916 prevent the removal of trailing zero bits to get the proper DER
9917 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9918 of a NamedBitList, for which trailing 0 bits need to be removed.)
9922 * Have SSL/TLS server implementation tolerate "mismatched" record
9923 protocol version while receiving ClientHello even if the
9924 ClientHello is fragmented. (The server can't insist on the
9925 particular protocol version it has chosen before the ServerHello
9926 message has informed the client about his choice.)
9930 * Add RFC 3779 support.
9932 *Rob Austein for ARIN, Ben Laurie*
9934 * Load error codes if they are not already present instead of using a
9935 static variable. This allows them to be cleanly unloaded and reloaded.
9936 Improve header file function name parsing.
9940 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9941 or CAPABILITY handshake as required by RFCs.
9945 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9947 * Introduce limits to prevent malicious keys being able to
9948 cause a denial of service. ([CVE-2006-2940])
9950 *Steve Henson, Bodo Moeller*
9952 * Fix ASN.1 parsing of certain invalid structures that can result
9953 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9955 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9956 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9958 * Fix SSL client code which could crash if connecting to a
9959 malicious SSLv2 server. ([CVE-2006-4343])
9961 *Tavis Ormandy and Will Drewry, Google Security Team*
9963 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9964 match only those. Before that, "AES256-SHA" would be interpreted
9965 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9966 the same strength classification in 0.9.7h) as we currently only
9967 have a single AES bit in the ciphersuite description bitmap.
9968 That change, however, also applied to ciphersuite strings such as
9969 "RC4-MD5" that intentionally matched multiple ciphersuites --
9970 namely, SSL 2.0 ciphersuites in addition to the more common ones
9971 from SSL 3.0/TLS 1.0.
9973 So we change the selection algorithm again: Naming an explicit
9974 ciphersuite selects this one ciphersuite, and any other similar
9975 ciphersuite (same bitmap) from *other* protocol versions.
9976 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9977 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9979 Since SSL 2.0 does not have any ciphersuites for which the
9980 128/256 bit distinction would be relevant, this works for now.
9981 The proper fix will be to use different bits for AES128 and
9982 AES256, which would have avoided the problems from the beginning;
9983 however, bits are scarce, so we can only do this in a new release
9984 (not just a patchlevel) when we can change the SSL_CIPHER
9985 definition to split the single 'unsigned long mask' bitmap into
9986 multiple values to extend the available space.
9990 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9992 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9993 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9995 * Add AES IGE and biIGE modes.
9999 * Change the Unix randomness entropy gathering to use poll() when
10000 possible instead of select(), since the latter has some
10001 undesirable limitations.
10003 *Darryl Miles via Richard Levitte and Bodo Moeller*
10005 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10006 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10007 cannot be implicitly activated as part of, e.g., the "AES" alias.
10008 However, please upgrade to OpenSSL 0.9.9[-dev] for
10009 non-experimental use of the ECC ciphersuites to get TLS extension
10010 support, which is required for curve and point format negotiation
10011 to avoid potential handshake problems.
10015 * Disable rogue ciphersuites:
10017 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10018 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10019 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10021 The latter two were purportedly from
10022 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10025 Also deactivate the remaining ciphersuites from
10026 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10027 unofficial, and the ID has long expired.
10031 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10032 dual-core machines) and other potential thread-safety issues.
10036 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10037 versions), which is now available for royalty-free use
10038 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10039 Also, add Camellia TLS ciphersuites from RFC 4132.
10041 To minimize changes between patchlevels in the OpenSSL 0.9.8
10042 series, Camellia remains excluded from compilation unless OpenSSL
10043 is configured with 'enable-camellia'.
10047 * Disable the padding bug check when compression is in use. The padding
10048 bug check assumes the first packet is of even length, this is not
10049 necessarily true if compression is enabled and can result in false
10050 positives causing handshake failure. The actual bug test is ancient
10051 code so it is hoped that implementations will either have fixed it by
10052 now or any which still have the bug do not support compression.
10056 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10058 * When applying a cipher rule check to see if string match is an explicit
10059 cipher suite and only match that one cipher suite if it is.
10063 * Link in manifests for VC++ if needed.
10065 *Austin Ziegler <halostatue@gmail.com>*
10067 * Update support for ECC-based TLS ciphersuites according to
10068 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10069 TLS extensions, which are supported starting with the 0.9.9
10070 branch, not in the OpenSSL 0.9.8 branch).
10074 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10075 opaque EVP_CIPHER_CTX handling.
10079 * Fixes and enhancements to zlib compression code. We now only use
10080 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10081 to conform with the standards mentioned here:
10082 <http://www.zlib.net/DLL_FAQ.txt>
10083 Static zlib linking now works on Windows and the new --with-zlib-include
10084 --with-zlib-lib options to Configure can be used to supply the location
10085 of the headers and library. Gracefully handle case where zlib library
10090 * Several fixes and enhancements to the OID generation code. The old code
10091 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10092 handle numbers larger than ULONG_MAX, truncated printing and had a
10093 non standard OBJ_obj2txt() behaviour.
10097 * Add support for building of engines under engine/ as shared libraries
10098 under VC++ build system.
10102 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10103 Hopefully, we will not see any false combination of paths any more.
10107 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10109 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10110 (part of SSL_OP_ALL). This option used to disable the
10111 countermeasure against man-in-the-middle protocol-version
10112 rollback in the SSL 2.0 server implementation, which is a bad
10113 idea. ([CVE-2005-2969])
10115 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10116 for Information Security, National Institute of Advanced Industrial
10117 Science and Technology [AIST], Japan)*
10119 * Add two function to clear and return the verify parameter flags.
10123 * Keep cipherlists sorted in the source instead of sorting them at
10124 runtime, thus removing the need for a lock.
10128 * Avoid some small subgroup attacks in Diffie-Hellman.
10130 *Nick Mathewson and Ben Laurie*
10132 * Add functions for well-known primes.
10136 * Extended Windows CE support.
10138 *Satoshi Nakamura and Andy Polyakov*
10140 * Initialize SSL_METHOD structures at compile time instead of during
10141 runtime, thus removing the need for a lock.
10145 * Make PKCS7_decrypt() work even if no certificate is supplied by
10146 attempting to decrypt each encrypted key in turn. Add support to
10151 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10153 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10156 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10160 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10161 key into the same file any more.
10165 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10169 * Add -utf8 command line and config file option to 'ca'.
10171 *Stefan <stf@udoma.org*
10173 * Removed the macro des_crypt(), as it seems to conflict with some
10174 libraries. Use DES_crypt().
10178 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10179 involves renaming the source and generated shared-libs for
10180 both. The engines will accept the corrected or legacy ids
10181 ('ncipher' and '4758_cca' respectively) when binding. NB,
10182 this only applies when building 'shared'.
10184 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10186 * Add attribute functions to EVP_PKEY structure. Modify
10187 PKCS12_create() to recognize a CSP name attribute and
10188 use it. Make -CSP option work again in pkcs12 utility.
10192 * Add new functionality to the bn blinding code:
10193 - automatic re-creation of the BN_BLINDING parameters after
10194 a fixed number of uses (currently 32)
10195 - add new function for parameter creation
10196 - introduce flags to control the update behaviour of the
10197 BN_BLINDING parameters
10198 - hide BN_BLINDING structure
10199 Add a second BN_BLINDING slot to the RSA structure to improve
10200 performance when a single RSA object is shared among several
10205 * Add support for DTLS.
10207 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10209 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10210 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10214 * Remove buggy and incomplete DH cert support from
10215 ssl/ssl_rsa.c and ssl/s3_both.c
10219 * Use SHA-1 instead of MD5 as the default digest algorithm for
10220 the `apps/openssl` commands.
10224 * Compile clean with "-Wall -Wmissing-prototypes
10225 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10226 DEBUG_SAFESTACK must also be set.
10230 * Change ./Configure so that certain algorithms can be disabled by default.
10231 The new counterpiece to "no-xxx" is "enable-xxx".
10233 The patented RC5 and MDC2 algorithms will now be disabled unless
10234 "enable-rc5" and "enable-mdc2", respectively, are specified.
10236 (IDEA remains enabled despite being patented. This is because IDEA
10237 is frequently required for interoperability, and there is no license
10238 fee for non-commercial use. As before, "no-idea" can be used to
10239 avoid this algorithm.)
10243 * Add processing of proxy certificates (see RFC 3820). This work was
10244 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10245 EGEE (Enabling Grids for E-science in Europe).
10249 * RC4 performance overhaul on modern architectures/implementations, such
10250 as Intel P4, IA-64 and AMD64.
10254 * New utility extract-section.pl. This can be used specify an alternative
10255 section number in a pod file instead of having to treat each file as
10256 a separate case in Makefile. This can be done by adding two lines to the
10259 =for comment openssl_section:XXX
10261 The blank line is mandatory.
10265 * New arguments -certform, -keyform and -pass for s_client and s_server
10266 to allow alternative format key and certificate files and passphrase
10271 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10272 update associated structures and add various utility functions.
10274 Add new policy related verify parameters, include policy checking in
10275 standard verify code. Enhance 'smime' application with extra parameters
10276 to support policy checking and print out.
10280 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10281 Nehemiah processors. These extensions support AES encryption in hardware
10282 as well as RNG (though RNG support is currently disabled).
10284 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10286 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10290 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10292 *Andy Polyakov and a number of other people*
10294 * Improved PowerPC platform support. Most notably BIGNUM assembler
10295 implementation contributed by IBM.
10297 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10299 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10300 exponent rather than 'unsigned long'. There is a corresponding change to
10301 the new 'rsa_keygen' element of the RSA_METHOD structure.
10303 *Jelte Jansen, Geoff Thorpe*
10305 * Functionality for creating the initial serial number file is now
10306 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10308 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10309 number file to 1, which is bound to cause problems. To avoid
10310 the problems while respecting compatibility between different 0.9.7
10311 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10312 CA.pl for serial number initialization. With the new release 0.9.8,
10313 we can fix the problem directly in the 'ca' utility.)
10317 * Reduced header interdependencies by declaring more opaque objects in
10318 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10319 give fewer recursive includes, which could break lazy source code - so
10320 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10321 developers should define this symbol when building and using openssl to
10322 ensure they track the recommended behaviour, interfaces, [etc], but
10323 backwards-compatible behaviour prevails when this isn't defined.
10327 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10331 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10332 This will generate a random key of the appropriate length based on the
10333 cipher context. The EVP_CIPHER can provide its own random key generation
10334 routine to support keys of a specific form. This is used in the des and
10335 3des routines to generate a key of the correct parity. Update S/MIME
10336 code to use new functions and hence generate correct parity DES keys.
10337 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10338 valid (weak or incorrect parity).
10342 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10343 as looking them up. This is useful when the verified structure may contain
10344 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10345 present unless the new PKCS7_NO_CRL flag is asserted.
10349 * Extend ASN1 oid configuration module. It now additionally accepts the
10352 shortName = some long name, 1.2.3.4
10356 * Reimplemented the BN_CTX implementation. There is now no more static
10357 limitation on the number of variables it can handle nor the depth of the
10358 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10359 information can now expand as required, and rather than having a single
10360 static array of bignums, BN_CTX now uses a linked-list of such arrays
10361 allowing it to expand on demand whilst maintaining the usefulness of
10362 BN_CTX's "bundling".
10366 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10367 to allow all RSA operations to function using a single BN_CTX.
10371 * Preliminary support for certificate policy evaluation and checking. This
10372 is initially intended to pass the tests outlined in "Conformance Testing
10373 of Relying Party Client Certificate Path Processing Logic" v1.07.
10377 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10378 remained unused and not that useful. A variety of other little bignum
10379 tweaks and fixes have also been made continuing on from the audit (see
10384 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10385 associated ASN1, EVP and SSL functions and old ASN1 macros.
10389 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10390 and this should never fail. So the return value from the use of
10391 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10392 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10396 * BN_CTX_get() should return zero-valued bignums, providing the same
10397 initialised value as BN_new().
10399 *Geoff Thorpe, suggested by Ulf Möller*
10401 * Support for inhibitAnyPolicy certificate extension.
10405 * An audit of the BIGNUM code is underway, for which debugging code is
10406 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10407 is considered valid when processing BIGNUMs, and causes execution to
10408 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10409 further steps are taken to deliberately pollute unused data in BIGNUM
10410 structures to try and expose faulty code further on. For now, openssl will
10411 (in its default mode of operation) continue to tolerate the inconsistent
10412 forms that it has tolerated in the past, but authors and packagers should
10413 consider trying openssl and their own applications when compiled with
10414 these debugging symbols defined. It will help highlight potential bugs in
10415 their own code, and will improve the test coverage for OpenSSL itself. At
10416 some point, these tighter rules will become openssl's default to improve
10417 maintainability, though the assert()s and other overheads will remain only
10418 in debugging configurations. See bn.h for more details.
10420 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10422 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10423 that can only be obtained through BN_CTX_new() (which implicitly
10424 initialises it). The presence of this function only made it possible
10425 to overwrite an existing structure (and cause memory leaks).
10429 * Because of the callback-based approach for implementing LHASH as a
10430 template type, lh_insert() adds opaque objects to hash-tables and
10431 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10432 to clean up those corresponding objects before destroying the hash table
10433 (and losing the object pointers). So some over-zealous constifications in
10434 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10435 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10436 prototyped to have "const" restrictions on the object pointers they are
10437 given (and so aren't required to cast them away any more).
10441 * The tmdiff.h API was so ugly and minimal that our own timing utility
10442 (speed) prefers to use its own implementation. The two implementations
10443 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10444 its object type properly exposed (MS_TM) instead of casting to/from
10445 `char *`. This may still change yet if someone realises MS_TM and
10447 aren't necessarily the greatest nomenclatures - but this is what was used
10448 internally to the implementation so I've used that for now.
10452 * Ensure that deprecated functions do not get compiled when
10453 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10454 the self-tests were still using deprecated key-generation functions so
10455 these have been updated also.
10459 * Reorganise PKCS#7 code to separate the digest location functionality
10460 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10461 New function PKCS7_set_digest() to set the digest type for PKCS#7
10462 digestedData type. Add additional code to correctly generate the
10463 digestedData type and add support for this type in PKCS7 initialization
10468 * New function PKCS7_set0_type_other() this initializes a PKCS7
10469 structure of type "other".
10473 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10474 sure the loop does correctly stop and breaking ("division by zero")
10475 modulus operations are not performed. The (pre-generated) prime
10476 table crypto/bn/bn_prime.h was already correct, but it could not be
10477 re-generated on some platforms because of the "division by zero"
10478 situation in the script.
10480 *Ralf S. Engelschall*
10482 * Update support for ECC-based TLS ciphersuites according to
10483 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10484 SHA-1 now is only used for "small" curves (where the
10485 representation of a field element takes up to 24 bytes); for
10486 larger curves, the field element resulting from ECDH is directly
10487 used as premaster secret.
10489 *Douglas Stebila (Sun Microsystems Laboratories)*
10491 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10492 curve secp160r1 to the tests.
10494 *Douglas Stebila (Sun Microsystems Laboratories)*
10496 * Add the possibility to load symbols globally with DSO.
10498 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10500 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10501 control of the error stack.
10505 * Add support for STORE in ENGINE.
10509 * Add the STORE type. The intention is to provide a common interface
10510 to certificate and key stores, be they simple file-based stores, or
10511 HSM-type store, or LDAP stores, or...
10512 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10516 * Add a generic structure called OPENSSL_ITEM. This can be used to
10517 pass a list of arguments to any function as well as provide a way
10518 for a function to pass data back to the caller.
10522 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10523 works like BUF_strdup() but can be used to duplicate a portion of
10524 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10529 * Add the function sk_find_ex() which works like sk_find(), but will
10530 return an index to an element even if an exact match couldn't be
10531 found. The index is guaranteed to point at the element where the
10532 searched-for key would be inserted to preserve sorting order.
10536 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10537 takes an extra flags argument for optional functionality. Currently,
10538 the following flags are defined:
10540 OBJ_BSEARCH_VALUE_ON_NOMATCH
10541 This one gets OBJ_bsearch_ex() to return a pointer to the first
10542 element where the comparing function returns a negative or zero
10545 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10546 This one gets OBJ_bsearch_ex() to return a pointer to the first
10547 element where the comparing function returns zero. This is useful
10548 if there are more than one element where the comparing function
10553 * Make it possible to create self-signed certificates with 'openssl ca'
10554 in such a way that the self-signed certificate becomes part of the
10555 CA database and uses the same mechanisms for serial number generation
10556 as all other certificate signing. The new flag '-selfsign' enables
10557 this functionality. Adapt CA.sh and CA.pl.in.
10561 * Add functionality to check the public key of a certificate request
10562 against a given private. This is useful to check that a certificate
10563 request can be signed by that key (self-signing).
10567 * Make it possible to have multiple active certificates with the same
10568 subject in the CA index file. This is done only if the keyword
10569 'unique_subject' is set to 'no' in the main CA section (default
10570 if 'CA_default') of the configuration file. The value is saved
10571 with the database itself in a separate index attribute file,
10572 named like the index file with '.attr' appended to the name.
10576 * Generate multi-valued AVAs using '+' notation in config files for
10581 * Support for nameConstraints certificate extension.
10585 * Support for policyConstraints certificate extension.
10589 * Support for policyMappings certificate extension.
10593 * Make sure the default DSA_METHOD implementation only uses its
10594 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10595 and change its own handlers to be NULL so as to remove unnecessary
10596 indirection. This lets alternative implementations fallback to the
10597 default implementation more easily.
10601 * Support for directoryName in GeneralName related extensions
10606 * Make it possible to link applications using Makefile.shared.
10607 Make that possible even when linking against static libraries!
10611 * Support for single pass processing for S/MIME signing. This now
10612 means that S/MIME signing can be done from a pipe, in addition
10613 cleartext signing (multipart/signed type) is effectively streaming
10614 and the signed data does not need to be all held in memory.
10616 This is done with a new flag PKCS7_STREAM. When this flag is set
10617 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10618 is done after the data is output (and digests calculated) in
10619 SMIME_write_PKCS7().
10623 * Add full support for -rpath/-R, both in shared libraries and
10624 applications, at least on the platforms where it's known how
10629 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10630 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10631 will now compute a table of multiples of the generator that
10632 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10633 faster (notably in the case of a single point multiplication,
10634 scalar * generator).
10636 *Nils Larsch, Bodo Moeller*
10638 * IPv6 support for certificate extensions. The various extensions
10639 which use the IP:a.b.c.d can now take IPv6 addresses using the
10640 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10645 * Added an ENGINE that implements RSA by performing private key
10646 exponentiations with the GMP library. The conversions to and from
10647 GMP's mpz_t format aren't optimised nor are any montgomery forms
10648 cached, and on x86 it appears OpenSSL's own performance has caught up.
10649 However there are likely to be other architectures where GMP could
10650 provide a boost. This ENGINE is not built in by default, but it can be
10651 specified at Configure time and should be accompanied by the necessary
10652 linker additions, eg;
10653 ./config -DOPENSSL_USE_GMP -lgmp
10657 * "openssl engine" will not display ENGINE/DSO load failure errors when
10658 testing availability of engines with "-t" - the old behaviour is
10659 produced by increasing the feature's verbosity with "-tt".
10663 * ECDSA routines: under certain error conditions uninitialized BN objects
10664 could be freed. Solution: make sure initialization is performed early
10665 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10670 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10671 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10672 software implementations. For DSA and DH, parameter generation can
10673 also be overridden by providing the appropriate method callbacks.
10677 * Change the "progress" mechanism used in key-generation and
10678 primality testing to functions that take a new BN_GENCB pointer in
10679 place of callback/argument pairs. The new API functions have `_ex`
10680 postfixes and the older functions are reimplemented as wrappers for
10681 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10682 declarations of the old functions to help (graceful) attempts to
10683 migrate to the new functions. Also, the new key-generation API
10684 functions operate on a caller-supplied key-structure and return
10685 success/failure rather than returning a key or NULL - this is to
10686 help make "keygen" another member function of RSA_METHOD etc.
10688 Example for using the new callback interface:
10690 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10691 void *my_arg = ...;
10694 BN_GENCB_set(&my_cb, my_callback, my_arg);
10696 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10697 /* For the meaning of a, b in calls to my_callback(), see the
10698 * documentation of the function that calls the callback.
10699 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10700 * my_callback should return 1 if it wants BN_is_prime_ex()
10701 * to continue, or 0 to stop.
10706 * Change the ZLIB compression method to be stateful, and make it
10707 available to TLS with the number defined in
10708 draft-ietf-tls-compression-04.txt.
10712 * Add the ASN.1 structures and functions for CertificatePair, which
10713 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10715 CertificatePair ::= SEQUENCE {
10716 forward [0] Certificate OPTIONAL,
10717 reverse [1] Certificate OPTIONAL,
10718 -- at least one of the pair shall be present -- }
10720 Also implement the PEM functions to read and write certificate
10721 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10723 This needed to be defined, mostly for the sake of the LDAP
10724 attribute crossCertificatePair, but may prove useful elsewhere as
10729 * Make it possible to inhibit symlinking of shared libraries in
10730 Makefile.shared, for Cygwin's sake.
10734 * Extend the BIGNUM API by creating a function
10735 void BN_set_negative(BIGNUM *a, int neg);
10736 and a macro that behave like
10737 int BN_is_negative(const BIGNUM *a);
10739 to avoid the need to access 'a->neg' directly in applications.
10743 * Implement fast modular reduction for pseudo-Mersenne primes
10744 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10745 EC_GROUP_new_curve_GFp() will now automatically use this
10748 *Nils Larsch <nla@trustcenter.de>*
10750 * Add new lock type (CRYPTO_LOCK_BN).
10754 * Change the ENGINE framework to automatically load engines
10755 dynamically from specific directories unless they could be
10756 found to already be built in or loaded. Move all the
10757 current engines except for the cryptodev one to a new
10758 directory engines/.
10759 The engines in engines/ are built as shared libraries if
10760 the "shared" options was given to ./Configure or ./config.
10761 Otherwise, they are inserted in libcrypto.a.
10762 /usr/local/ssl/engines is the default directory for dynamic
10763 engines, but that can be overridden at configure time through
10764 the usual use of --prefix and/or --openssldir, and at run
10765 time with the environment variable OPENSSL_ENGINES.
10767 *Geoff Thorpe and Richard Levitte*
10769 * Add Makefile.shared, a helper makefile to build shared
10770 libraries. Adapt Makefile.org.
10774 * Add version info to Win32 DLLs.
10776 *Peter 'Luna' Runestig" <peter@runestig.com>*
10778 * Add new 'medium level' PKCS#12 API. Certificates and keys
10779 can be added using this API to created arbitrary PKCS#12
10780 files while avoiding the low level API.
10782 New options to PKCS12_create(), key or cert can be NULL and
10783 will then be omitted from the output file. The encryption
10784 algorithm NIDs can be set to -1 for no encryption, the mac
10785 iteration count can be set to 0 to omit the mac.
10787 Enhance pkcs12 utility by making the -nokeys and -nocerts
10788 options work when creating a PKCS#12 file. New option -nomac
10789 to omit the mac, NONE can be set for an encryption algorithm.
10790 New code is modified to use the enhanced PKCS12_create()
10791 instead of the low level API.
10795 * Extend ASN1 encoder to support indefinite length constructed
10796 encoding. This can output sequences tags and octet strings in
10797 this form. Modify pk7_asn1.c to support indefinite length
10798 encoding. This is experimental and needs additional code to
10799 be useful, such as an ASN1 bio and some enhanced streaming
10802 Extend template encode functionality so that tagging is passed
10803 down to the template encoder.
10807 * Let 'openssl req' fail if an argument to '-newkey' is not
10808 recognized instead of using RSA as a default.
10812 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10813 As these are not official, they are not included in "ALL";
10814 the "ECCdraft" ciphersuite group alias can be used to select them.
10816 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10818 * Add ECDH engine support.
10820 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10822 * Add ECDH in new directory crypto/ecdh/.
10824 *Douglas Stebila (Sun Microsystems Laboratories)*
10826 * Let BN_rand_range() abort with an error after 100 iterations
10827 without success (which indicates a broken PRNG).
10831 * Change BN_mod_sqrt() so that it verifies that the input value
10832 is really the square of the return value. (Previously,
10833 BN_mod_sqrt would show GIGO behaviour.)
10837 * Add named elliptic curves over binary fields from X9.62, SECG,
10838 and WAP/WTLS; add OIDs that were still missing.
10840 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10842 * Extend the EC library for elliptic curves over binary fields
10843 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10846 EC_GF2m_simple_method
10850 EC_GROUP_new_curve_GF2m
10851 EC_GROUP_set_curve_GF2m
10852 EC_GROUP_get_curve_GF2m
10853 EC_POINT_set_affine_coordinates_GF2m
10854 EC_POINT_get_affine_coordinates_GF2m
10855 EC_POINT_set_compressed_coordinates_GF2m
10857 Point compression for binary fields is disabled by default for
10858 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10861 As binary polynomials are represented as BIGNUMs, various members
10862 of the EC_GROUP and EC_POINT data structures can be shared
10863 between the implementations for prime fields and binary fields;
10864 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10865 are essentially identical to their `..._GFp` counterparts.
10866 (For simplicity, the `..._GFp` prefix has been dropped from
10867 various internal method names.)
10869 An internal 'field_div' method (similar to 'field_mul' and
10870 'field_sqr') has been added; this is used only for binary fields.
10872 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10874 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10875 through methods ('mul', 'precompute_mult').
10877 The generic implementations (now internally called 'ec_wNAF_mul'
10878 and 'ec_wNAF_precomputed_mult') remain the default if these
10879 methods are undefined.
10881 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10883 * New function EC_GROUP_get_degree, which is defined through
10884 EC_METHOD. For curves over prime fields, this returns the bit
10885 length of the modulus.
10887 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10889 * New functions EC_GROUP_dup, EC_POINT_dup.
10890 (These simply call ..._new and ..._copy).
10892 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10894 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10895 Polynomials are represented as BIGNUMs (where the sign bit is not
10896 used) in the following functions [macros]:
10899 BN_GF2m_sub [= BN_GF2m_add]
10900 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10901 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10902 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10904 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10905 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10906 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10907 BN_GF2m_cmp [= BN_ucmp]
10909 (Note that only the 'mod' functions are actually for fields GF(2^m).
10910 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10912 For some functions, an the irreducible polynomial defining a
10913 field can be given as an 'unsigned int[]' with strictly
10914 decreasing elements giving the indices of those bits that are set;
10915 i.e., p[] represents the polynomial
10916 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10918 p[0] > p[1] > ... > p[k] = 0.
10919 This applies to the following functions:
10922 BN_GF2m_mod_mul_arr
10923 BN_GF2m_mod_sqr_arr
10924 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10925 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10926 BN_GF2m_mod_exp_arr
10927 BN_GF2m_mod_sqrt_arr
10928 BN_GF2m_mod_solve_quad_arr
10932 Conversion can be performed by the following functions:
10937 bntest.c has additional tests for binary polynomial arithmetic.
10939 Two implementations for BN_GF2m_mod_div() are available.
10940 The default algorithm simply uses BN_GF2m_mod_inv() and
10941 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10942 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10943 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10945 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10947 * Add new error code 'ERR_R_DISABLED' that can be used when some
10948 functionality is disabled at compile-time.
10950 *Douglas Stebila <douglas.stebila@sun.com>*
10952 * Change default behaviour of 'openssl asn1parse' so that more
10953 information is visible when viewing, e.g., a certificate:
10955 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10956 mode the content of non-printable OCTET STRINGs is output in a
10957 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10958 avoid the appearance of a printable string.
10960 *Nils Larsch <nla@trustcenter.de>*
10962 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10964 EC_GROUP_set_asn1_flag()
10965 EC_GROUP_get_asn1_flag()
10966 EC_GROUP_set_point_conversion_form()
10967 EC_GROUP_get_point_conversion_form()
10968 These control ASN1 encoding details:
10969 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10970 has been set to OPENSSL_EC_NAMED_CURVE.
10971 - Points are encoded in uncompressed form by default; options for
10972 asn1_for are as for point2oct, namely
10973 POINT_CONVERSION_COMPRESSED
10974 POINT_CONVERSION_UNCOMPRESSED
10975 POINT_CONVERSION_HYBRID
10977 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10979 EC_GROUP_set_seed()
10980 EC_GROUP_get0_seed()
10981 EC_GROUP_get_seed_len()
10982 This is used only for ASN1 purposes (so far).
10984 *Nils Larsch <nla@trustcenter.de>*
10986 * Add 'field_type' member to EC_METHOD, which holds the NID
10987 of the appropriate field type OID. The new function
10988 EC_METHOD_get_field_type() returns this value.
10990 *Nils Larsch <nla@trustcenter.de>*
10993 EC_POINT_point2bn()
10994 EC_POINT_bn2point()
10995 EC_POINT_point2hex()
10996 EC_POINT_hex2point()
10997 providing useful interfaces to EC_POINT_point2oct() and
10998 EC_POINT_oct2point().
11000 *Nils Larsch <nla@trustcenter.de>*
11002 * Change internals of the EC library so that the functions
11003 EC_GROUP_set_generator()
11004 EC_GROUP_get_generator()
11005 EC_GROUP_get_order()
11006 EC_GROUP_get_cofactor()
11007 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11008 to methods, which would lead to unnecessary code duplication when
11009 adding different types of curves.
11011 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11013 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11014 arithmetic, and such that modified wNAFs are generated
11015 (which avoid length expansion in many cases).
11019 * Add a function EC_GROUP_check_discriminant() (defined via
11020 EC_METHOD) that verifies that the curve discriminant is non-zero.
11022 Add a function EC_GROUP_check() that makes some sanity tests
11023 on a EC_GROUP, its generator and order. This includes
11024 EC_GROUP_check_discriminant().
11026 *Nils Larsch <nla@trustcenter.de>*
11028 * Add ECDSA in new directory crypto/ecdsa/.
11030 Add applications 'openssl ecparam' and 'openssl ecdsa'
11031 (these are based on 'openssl dsaparam' and 'openssl dsa').
11033 ECDSA support is also included in various other files across the
11034 library. Most notably,
11035 - 'openssl req' now has a '-newkey ecdsa:file' option;
11036 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11037 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11038 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11039 them suitable for ECDSA where domain parameters must be
11040 extracted before the specific public key;
11041 - ECDSA engine support has been added.
11043 *Nils Larsch <nla@trustcenter.de>*
11045 * Include some named elliptic curves, and add OIDs from X9.62,
11046 SECG, and WAP/WTLS. Each curve can be obtained from the new
11048 EC_GROUP_new_by_curve_name(),
11049 and the list of available named curves can be obtained with
11050 EC_get_builtin_curves().
11051 Also add a 'curve_name' member to EC_GROUP objects, which can be
11053 EC_GROUP_set_curve_name()
11054 EC_GROUP_get_curve_name()
11056 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11058 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11059 was actually never needed) and in BN_mul(). The removal in BN_mul()
11060 required a small change in bn_mul_part_recursive() and the addition
11061 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11062 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11063 bn_sub_words() and bn_add_words() except they take arrays with
11068 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11070 * Cleanse PEM buffers before freeing them since they may contain
11073 *Benjamin Bennett <ben@psc.edu>*
11075 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11076 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11077 authentication-only ciphersuites.
11081 * Since AES128 and AES256 share a single mask bit in the logic of
11082 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11083 kludge to work properly if AES128 is available and AES256 isn't.
11087 * Expand security boundary to match 1.1.1 module.
11091 * Remove redundant features: hash file source, editing of test vectors
11092 modify fipsld to use external fips_premain.c signature.
11096 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11097 run algorithm test programs.
11101 * Make algorithm test programs more tolerant of whitespace.
11105 * Have SSL/TLS server implementation tolerate "mismatched" record
11106 protocol version while receiving ClientHello even if the
11107 ClientHello is fragmented. (The server can't insist on the
11108 particular protocol version it has chosen before the ServerHello
11109 message has informed the client about his choice.)
11113 * Load error codes if they are not already present instead of using a
11114 static variable. This allows them to be cleanly unloaded and reloaded.
11118 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11120 * Introduce limits to prevent malicious keys being able to
11121 cause a denial of service. ([CVE-2006-2940])
11123 *Steve Henson, Bodo Moeller*
11125 * Fix ASN.1 parsing of certain invalid structures that can result
11126 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11128 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11129 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11131 * Fix SSL client code which could crash if connecting to a
11132 malicious SSLv2 server. ([CVE-2006-4343])
11134 *Tavis Ormandy and Will Drewry, Google Security Team*
11136 * Change ciphersuite string processing so that an explicit
11137 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11138 will no longer include "AES128-SHA"), and any other similar
11139 ciphersuite (same bitmap) from *other* protocol versions (so that
11140 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11141 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11142 changes from 0.9.8b and 0.9.8d.
11146 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11148 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11149 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11151 * Change the Unix randomness entropy gathering to use poll() when
11152 possible instead of select(), since the latter has some
11153 undesirable limitations.
11155 *Darryl Miles via Richard Levitte and Bodo Moeller*
11157 * Disable rogue ciphersuites:
11159 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11160 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11161 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11163 The latter two were purportedly from
11164 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11167 Also deactivate the remaining ciphersuites from
11168 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11169 unofficial, and the ID has long expired.
11173 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11174 dual-core machines) and other potential thread-safety issues.
11178 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11180 * Adapt fipsld and the build system to link against the validated FIPS
11181 module in FIPS mode.
11185 * Fixes for VC++ 2005 build under Windows.
11189 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11190 from a Windows bash shell such as MSYS. It is autodetected from the
11191 "config" script when run from a VC++ environment. Modify standard VC++
11192 build to use fipscanister.o from the GNU make build.
11196 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11198 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11199 The value now differs depending on if you build for FIPS or not.
11200 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11201 safely run with a non-FIPSed libcrypto, as it may crash because of
11202 the difference induced by this change.
11206 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11208 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11209 (part of SSL_OP_ALL). This option used to disable the
11210 countermeasure against man-in-the-middle protocol-version
11211 rollback in the SSL 2.0 server implementation, which is a bad
11212 idea. ([CVE-2005-2969])
11214 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11215 for Information Security, National Institute of Advanced Industrial
11216 Science and Technology [AIST, Japan)]*
11218 * Minimal support for X9.31 signatures and PSS padding modes. This is
11219 mainly for FIPS compliance and not fully integrated at this stage.
11223 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11224 the exponentiation using a fixed-length exponent. (Otherwise,
11225 the information leaked through timing could expose the secret key
11226 after many signatures; cf. Bleichenbacher's attack on DSA with
11231 * Make a new fixed-window mod_exp implementation the default for
11232 RSA, DSA, and DH private-key operations so that the sequence of
11233 squares and multiplies and the memory access pattern are
11234 independent of the particular secret key. This will mitigate
11235 cache-timing and potential related attacks.
11237 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11238 and this is automatically used by BN_mod_exp_mont() if the new flag
11239 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11240 will use this BN flag for private exponents unless the flag
11241 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11242 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11244 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11246 * Change the client implementation for SSLv23_method() and
11247 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11248 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11249 (Previously, the SSL 2.0 backwards compatible Client Hello
11250 message format would be used even with SSL_OP_NO_SSLv2.)
11254 * Add support for smime-type MIME parameter in S/MIME messages which some
11259 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11260 a threadsafe manner. Modify rsa code to use new function and add calls
11261 to dsa and dh code (which had race conditions before).
11265 * Include the fixed error library code in the C error file definitions
11266 instead of fixing them up at runtime. This keeps the error code
11267 structures constant.
11271 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11273 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11276 * Fixes for newer kerberos headers. NB: the casts are needed because
11277 the 'length' field is signed on one version and unsigned on another
11278 with no (?) obvious way to tell the difference, without these VC++
11279 complains. Also the "definition" of FAR (blank) is no longer included
11280 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11281 some needed definitions.
11285 * Undo Cygwin change.
11289 * Added support for proxy certificates according to RFC 3820.
11290 Because they may be a security thread to unaware applications,
11291 they must be explicitly allowed in run-time. See
11292 docs/HOWTO/proxy_certificates.txt for further information.
11296 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11298 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11299 server and client random values. Previously
11300 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11301 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11303 This change has negligible security impact because:
11305 1. Server and client random values still have 24 bytes of pseudo random
11308 2. Server and client random values are sent in the clear in the initial
11311 3. The master secret is derived using the premaster secret (48 bytes in
11312 size for static RSA ciphersuites) as well as client server and random
11315 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11318 *Stephen Henson, reported by UK NISCC*
11320 * Use Windows randomness collection on Cygwin.
11324 * Fix hang in EGD/PRNGD query when communication socket is closed
11325 prematurely by EGD/PRNGD.
11327 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11329 * Prompt for pass phrases when appropriate for PKCS12 input format.
11333 * Back-port of selected performance improvements from development
11334 branch, as well as improved support for PowerPC platforms.
11338 * Add lots of checks for memory allocation failure, error codes to indicate
11339 failure and freeing up memory if a failure occurs.
11341 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11343 * Add new -passin argument to dgst.
11347 * Perform some character comparisons of different types in X509_NAME_cmp:
11348 this is needed for some certificates that re-encode DNs into UTF8Strings
11349 (in violation of RFC3280) and can't or won't issue name rollover
11354 * Make an explicit check during certificate validation to see that
11355 the CA setting in each certificate on the chain is correct. As a
11356 side effect always do the following basic checks on extensions,
11357 not just when there's an associated purpose to the check:
11359 - if there is an unhandled critical extension (unless the user
11360 has chosen to ignore this fault)
11361 - if the path length has been exceeded (if one is set at all)
11362 - that certain extensions fit the associated purpose (if one has
11367 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11369 * Avoid a race condition when CRLs are checked in a multi threaded
11370 environment. This would happen due to the reordering of the revoked
11371 entries during signature checking and serial number lookup. Now the
11372 encoding is cached and the serial number sort performed under a lock.
11373 Add new STACK function sk_is_sorted().
11377 * Add Delta CRL to the extension code.
11381 * Various fixes to s3_pkt.c so alerts are sent properly.
11383 *David Holmes <d.holmes@f5.com>*
11385 * Reduce the chances of duplicate issuer name and serial numbers (in
11386 violation of RFC3280) using the OpenSSL certificate creation utilities.
11387 This is done by creating a random 64 bit value for the initial serial
11388 number when a serial number file is created or when a self signed
11389 certificate is created using 'openssl req -x509'. The initial serial
11390 number file is created using 'openssl x509 -next_serial' in CA.pl
11391 rather than being initialized to 1.
11395 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11397 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11398 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11400 *Joe Orton, Steve Henson*
11402 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11405 *Joe Orton, Steve Henson*
11407 * Make it possible to have multiple active certificates with the same
11408 subject in the CA index file. This is done only if the keyword
11409 'unique_subject' is set to 'no' in the main CA section (default
11410 if 'CA_default') of the configuration file. The value is saved
11411 with the database itself in a separate index attribute file,
11412 named like the index file with '.attr' appended to the name.
11416 * X509 verify fixes. Disable broken certificate workarounds when
11417 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11418 keyUsage extension present. Don't accept CRLs with unhandled critical
11419 extensions: since verify currently doesn't process CRL extensions this
11420 rejects a CRL with *any* critical extensions. Add new verify error codes
11425 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11426 A clarification of RFC2560 will require the use of OCTET STRINGs and
11427 some implementations cannot handle the current raw format. Since OpenSSL
11428 copies and compares OCSP nonces as opaque blobs without any attempt at
11429 parsing them this should not create any compatibility issues.
11433 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11434 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11435 this HMAC (and other) operations are several times slower than OpenSSL
11440 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11442 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11444 * Use the correct content when signing type "other".
11448 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11450 * Fix various bugs revealed by running the NISCC test suite:
11452 Stop out of bounds reads in the ASN1 code when presented with
11453 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11455 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11457 If verify callback ignores invalid public key errors don't try to check
11458 certificate signature with the NULL public key.
11462 * New -ignore_err option in ocsp application to stop the server
11463 exiting on the first error in a request.
11467 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11468 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11473 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11474 extra data after the compression methods not only for TLS 1.0
11475 but also for SSL 3.0 (as required by the specification).
11477 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11479 * Change X509_certificate_type() to mark the key as exported/exportable
11480 when it's 512 *bits* long, not 512 bytes.
11484 * Change AES_cbc_encrypt() so it outputs exact multiple of
11485 blocks during encryption.
11489 * Various fixes to base64 BIO and non blocking I/O. On write
11490 flushes were not handled properly if the BIO retried. On read
11491 data was not being buffered properly and had various logic bugs.
11492 This also affects blocking I/O when the data being decoded is a
11497 * Various S/MIME bugfixes and compatibility changes:
11498 output correct application/pkcs7 MIME type if
11499 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11500 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11501 of files as .eml work). Correctly handle very long lines in MIME
11506 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11508 * Countermeasure against the Klima-Pokorny-Rosa extension of
11509 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11510 a protocol version number mismatch like a decryption error
11511 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11515 * Turn on RSA blinding by default in the default implementation
11516 to avoid a timing attack. Applications that don't want it can call
11517 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11518 They would be ill-advised to do so in most cases.
11520 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11522 * Change RSA blinding code so that it works when the PRNG is not
11523 seeded (in this case, the secret RSA exponent is abused as
11524 an unpredictable seed -- if it is not unpredictable, there
11525 is no point in blinding anyway). Make RSA blinding thread-safe
11526 by remembering the creator's thread ID in rsa->blinding and
11527 having all other threads use local one-time blinding factors
11528 (this requires more computation than sharing rsa->blinding, but
11529 avoids excessive locking; and if an RSA object is not shared
11530 between threads, blinding will still be very fast).
11534 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11535 ENGINE as defaults for all supported algorithms irrespective of
11536 the 'flags' parameter. 'flags' is now honoured, so applications
11537 should make sure they are passing it correctly.
11541 * Target "mingw" now allows native Windows code to be generated in
11542 the Cygwin environment as well as with the MinGW compiler.
11546 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11548 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11549 via timing by performing a MAC computation even if incorrect
11550 block cipher padding has been found. This is a countermeasure
11551 against active attacks where the attacker has to distinguish
11552 between bad padding and a MAC verification error. ([CVE-2003-0078])
11554 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11555 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11556 Martin Vuagnoux (EPFL, Ilion)*
11558 * Make the no-err option work as intended. The intention with no-err
11559 is not to have the whole error stack handling routines removed from
11560 libcrypto, it's only intended to remove all the function name and
11561 reason texts, thereby removing some of the footprint that may not
11562 be interesting if those errors aren't displayed anyway.
11564 NOTE: it's still possible for any application or module to have its
11565 own set of error texts inserted. The routines are there, just not
11566 used by default when no-err is given.
11570 * Add support for FreeBSD on IA64.
11572 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11574 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11575 Kerberos function mit_des_cbc_cksum(). Before this change,
11576 the value returned by DES_cbc_cksum() was like the one from
11577 mit_des_cbc_cksum(), except the bytes were swapped.
11579 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11581 * Allow an application to disable the automatic SSL chain building.
11582 Before this a rather primitive chain build was always performed in
11583 ssl3_output_cert_chain(): an application had no way to send the
11584 correct chain if the automatic operation produced an incorrect result.
11586 Now the chain builder is disabled if either:
11588 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11590 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11592 The reasoning behind this is that an application would not want the
11593 auto chain building to take place if extra chain certificates are
11594 present and it might also want a means of sending no additional
11595 certificates (for example the chain has two certificates and the
11600 * Add the possibility to build without the ENGINE framework.
11602 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11604 * Under Win32 gmtime() can return NULL: check return value in
11605 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11609 * DSA routines: under certain error conditions uninitialized BN objects
11610 could be freed. Solution: make sure initialization is performed early
11611 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11612 Nils Larsch <nla@trustcenter.de> via PR#459)
11616 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11617 checked on reconnect on the client side, therefore session resumption
11618 could still fail with a "ssl session id is different" error. This
11619 behaviour is masked when SSL_OP_ALL is used due to
11620 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11621 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11622 followup to PR #377.
11626 * IA-32 assembler support enhancements: unified ELF targets, support
11627 for SCO/Caldera platforms, fix for Cygwin shared build.
11631 * Add support for FreeBSD on sparc64. As a consequence, support for
11632 FreeBSD on non-x86 processors is separate from x86 processors on
11633 the config script, much like the NetBSD support.
11635 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11637 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11639 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11642 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11643 code (06) was taken as the first octet of the session ID and the last
11644 octet was ignored consequently. As a result SSLv2 client side session
11645 caching could not have worked due to the session ID mismatch between
11647 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11652 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11653 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11658 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11659 seems that in spite of existing for more than a year, many application
11660 author have done nothing to provide the necessary callbacks, which
11661 means that this particular engine will not work properly anywhere.
11662 This is a very unfortunate situation which forces us, in the name
11663 of usability, to give the hw_ncipher.c a static lock, which is part
11665 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11666 appear in 0.9.8 or later. We EXPECT application authors to have
11667 dealt properly with this when 0.9.8 is released (unless we actually
11668 make such changes in the libcrypto locking code that changes will
11669 have to be made anyway).
11673 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11674 octets have been read, EOF or an error occurs. Without this change
11675 some truncated ASN1 structures will not produce an error.
11679 * Disable Heimdal support, since it hasn't been fully implemented.
11680 Still give the possibility to force the use of Heimdal, but with
11681 warnings and a request that patches get sent to openssl-dev.
11685 * Add the VC-CE target, introduce the WINCE sysname, and add
11686 INSTALL.WCE and appropriate conditionals to make it build.
11688 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11690 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11691 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11692 edit numbers of the version.
11694 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11696 * Introduce safe string copy and catenation functions
11697 (BUF_strlcpy() and BUF_strlcat()).
11699 *Ben Laurie (CHATS) and Richard Levitte*
11701 * Avoid using fixed-size buffers for one-line DNs.
11703 *Ben Laurie (CHATS)*
11705 * Add BUF_MEM_grow_clean() to avoid information leakage when
11706 resizing buffers containing secrets, and use where appropriate.
11708 *Ben Laurie (CHATS)*
11710 * Avoid using fixed size buffers for configuration file location.
11712 *Ben Laurie (CHATS)*
11714 * Avoid filename truncation for various CA files.
11716 *Ben Laurie (CHATS)*
11718 * Use sizeof in preference to magic numbers.
11720 *Ben Laurie (CHATS)*
11722 * Avoid filename truncation in cert requests.
11724 *Ben Laurie (CHATS)*
11726 * Add assertions to check for (supposedly impossible) buffer
11729 *Ben Laurie (CHATS)*
11731 * Don't cache truncated DNS entries in the local cache (this could
11732 potentially lead to a spoofing attack).
11734 *Ben Laurie (CHATS)*
11736 * Fix various buffers to be large enough for hex/decimal
11737 representations in a platform independent manner.
11739 *Ben Laurie (CHATS)*
11741 * Add CRYPTO_realloc_clean() to avoid information leakage when
11742 resizing buffers containing secrets, and use where appropriate.
11744 *Ben Laurie (CHATS)*
11746 * Add BIO_indent() to avoid much slightly worrying code to do
11749 *Ben Laurie (CHATS)*
11751 * Convert sprintf()/BIO_puts() to BIO_printf().
11753 *Ben Laurie (CHATS)*
11755 * buffer_gets() could terminate with the buffer only half
11758 *Ben Laurie (CHATS)*
11760 * Add assertions to prevent user-supplied crypto functions from
11761 overflowing internal buffers by having large block sizes, etc.
11763 *Ben Laurie (CHATS)*
11765 * New OPENSSL_assert() macro (similar to assert(), but enabled
11768 *Ben Laurie (CHATS)*
11770 * Eliminate unused copy of key in RC4.
11772 *Ben Laurie (CHATS)*
11774 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11776 *Ben Laurie (CHATS)*
11778 * Fix off-by-one error in EGD path.
11780 *Ben Laurie (CHATS)*
11782 * If RANDFILE path is too long, ignore instead of truncating.
11784 *Ben Laurie (CHATS)*
11786 * Eliminate unused and incorrectly sized X.509 structure
11789 *Ben Laurie (CHATS)*
11791 * Eliminate unused and dangerous function knumber().
11793 *Ben Laurie (CHATS)*
11795 * Eliminate unused and dangerous structure, KSSL_ERR.
11797 *Ben Laurie (CHATS)*
11799 * Protect against overlong session ID context length in an encoded
11800 session object. Since these are local, this does not appear to be
11803 *Ben Laurie (CHATS)*
11805 * Change from security patch (see 0.9.6e below) that did not affect
11806 the 0.9.6 release series:
11808 Remote buffer overflow in SSL3 protocol - an attacker could
11809 supply an oversized master key in Kerberos-enabled versions.
11812 *Ben Laurie (CHATS)*
11814 * Change the SSL kerb5 codes to match RFC 2712.
11818 * Make -nameopt work fully for req and add -reqopt switch.
11820 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11822 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11824 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11826 * Make sure tests can be performed even if the corresponding algorithms
11827 have been removed entirely. This was also the last step to make
11828 OpenSSL compilable with DJGPP under all reasonable conditions.
11830 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11832 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11833 to allow version independent disabling of normally unselected ciphers,
11834 which may be activated as a side-effect of selecting a single cipher.
11836 (E.g., cipher list string "RSA" enables ciphersuites that are left
11837 out of "ALL" because they do not provide symmetric encryption.
11838 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11840 *Lutz Jaenicke, Bodo Moeller*
11842 * Add appropriate support for separate platform-dependent build
11843 directories. The recommended way to make a platform-dependent
11844 build directory is the following (tested on Linux), maybe with
11847 # Place yourself outside of the OpenSSL source tree. In
11848 # this example, the environment variable OPENSSL_SOURCE
11849 # is assumed to contain the absolute OpenSSL source directory.
11850 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11851 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11852 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11853 mkdir -p `dirname $F`
11854 ln -s $OPENSSL_SOURCE/$F $F
11857 To be absolutely sure not to disturb the source tree, a "make clean"
11858 is a good thing. If it isn't successful, don't worry about it,
11859 it probably means the source directory is very clean.
11863 * Make sure any ENGINE control commands make local copies of string
11864 pointers passed to them whenever necessary. Otherwise it is possible
11865 the caller may have overwritten (or deallocated) the original string
11866 data when a later ENGINE operation tries to use the stored values.
11868 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11870 * Improve diagnostics in file reading and command-line digests.
11872 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11874 * Add AES modes CFB and OFB to the object database. Correct an
11875 error in AES-CFB decryption.
11879 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11880 allows existing EVP_CIPHER_CTX structures to be reused after
11881 calling `EVP_*Final()`. This behaviour is used by encryption
11882 BIOs and some applications. This has the side effect that
11883 applications must explicitly clean up cipher contexts with
11884 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11888 * Check the values of dna and dnb in bn_mul_recursive before calling
11889 bn_mul_comba (a non zero value means the a or b arrays do not contain
11890 n2 elements) and fallback to bn_mul_normal if either is not zero.
11894 * Fix escaping of non-ASCII characters when using the -subj option
11895 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11899 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11900 form for "surname", serialNumber has no short form.
11901 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11902 therefore remove "mail" short name for "internet 7".
11903 The OID for unique identifiers in X509 certificates is
11904 x500UniqueIdentifier, not uniqueIdentifier.
11905 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11909 * Add an "init" command to the ENGINE config module and auto initialize
11910 ENGINEs. Without any "init" command the ENGINE will be initialized
11911 after all ctrl commands have been executed on it. If init=1 the
11912 ENGINE is initialized at that point (ctrls before that point are run
11913 on the uninitialized ENGINE and after on the initialized one). If
11914 init=0 then the ENGINE will not be initialized at all.
11918 * Fix the 'app_verify_callback' interface so that the user-defined
11919 argument is actually passed to the callback: In the
11920 SSL_CTX_set_cert_verify_callback() prototype, the callback
11921 declaration has been changed from
11924 int (*cb)(X509_STORE_CTX *,void *);
11925 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11926 i=s->ctx->app_verify_callback(&ctx)
11927 has been changed into
11928 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11930 To update applications using SSL_CTX_set_cert_verify_callback(),
11931 a dummy argument can be added to their callback functions.
11933 *D. K. Smetters <smetters@parc.xerox.com>*
11935 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11937 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11939 * Add and OPENSSL_LOAD_CONF define which will cause
11940 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11941 This allows older applications to transparently support certain
11942 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11943 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11944 load the config file and OPENSSL_add_all_algorithms_conf() which will
11945 always load it have also been added.
11949 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11950 Adjust NIDs and EVP layer.
11952 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11954 * Config modules support in openssl utility.
11956 Most commands now load modules from the config file,
11957 though in a few (such as version) this isn't done
11958 because it couldn't be used for anything.
11960 In the case of ca and req the config file used is
11961 the same as the utility itself: that is the -config
11962 command line option can be used to specify an
11967 * Move default behaviour from OPENSSL_config(). If appname is NULL
11968 use "openssl_conf" if filename is NULL use default openssl config file.
11972 * Add an argument to OPENSSL_config() to allow the use of an alternative
11973 config section name. Add a new flag to tolerate a missing config file
11974 and move code to CONF_modules_load_file().
11978 * Support for crypto accelerator cards from Accelerated Encryption
11979 Processing, www.aep.ie. (Use engine 'aep')
11980 The support was copied from 0.9.6c [engine] and adapted/corrected
11981 to work with the new engine framework.
11983 *AEP Inc. and Richard Levitte*
11985 * Support for SureWare crypto accelerator cards from Baltimore
11986 Technologies. (Use engine 'sureware')
11987 The support was copied from 0.9.6c [engine] and adapted
11988 to work with the new engine framework.
11992 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11993 make the newer ENGINE framework commands for the CHIL engine work.
11995 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11997 * Make it possible to produce shared libraries on ReliantUNIX.
11999 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12001 * Add the configuration target debug-linux-ppro.
12002 Make 'openssl rsa' use the general key loading routines
12003 implemented in `apps.c`, and make those routines able to
12004 handle the key format FORMAT_NETSCAPE and the variant
12007 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12009 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12011 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12013 * Add -keyform to rsautl, and document -engine.
12015 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12017 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12018 BIO_R_NO_SUCH_FILE error code rather than the generic
12019 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12023 * Add new functions
12024 ERR_peek_last_error
12025 ERR_peek_last_error_line
12026 ERR_peek_last_error_line_data.
12027 These are similar to
12029 ERR_peek_error_line
12030 ERR_peek_error_line_data,
12031 but report on the latest error recorded rather than the first one
12032 still in the error queue.
12034 *Ben Laurie, Bodo Moeller*
12036 * default_algorithms option in ENGINE config module. This allows things
12038 default_algorithms = ALL
12039 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12043 * Preliminary ENGINE config module.
12047 * New experimental application configuration code.
12051 * Change the AES code to follow the same name structure as all other
12052 symmetric ciphers, and behave the same way. Move everything to
12053 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12055 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12057 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12059 *Ben Laurie and Theo de Raadt*
12061 * Add option to output public keys in req command.
12063 *Massimiliano Pala madwolf@openca.org*
12065 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12066 (up to about 10% better than before for P-192 and P-224).
12070 * New functions/macros
12072 SSL_CTX_set_msg_callback(ctx, cb)
12073 SSL_CTX_set_msg_callback_arg(ctx, arg)
12074 SSL_set_msg_callback(ssl, cb)
12075 SSL_set_msg_callback_arg(ssl, arg)
12077 to request calling a callback function
12079 void cb(int write_p, int version, int content_type,
12080 const void *buf, size_t len, SSL *ssl, void *arg)
12082 whenever a protocol message has been completely received
12083 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12084 protocol version according to which the SSL library interprets
12085 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12086 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12087 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12088 specification (change_cipher_spec(20), alert(21), handshake(22)).
12089 'buf' and 'len' point to the actual message, 'ssl' to the
12090 SSL object, and 'arg' is the application-defined value set by
12091 SSL[_CTX]_set_msg_callback_arg().
12093 'openssl s_client' and 'openssl s_server' have new '-msg' options
12094 to enable a callback that displays all protocol messages.
12098 * Change the shared library support so shared libraries are built as
12099 soon as the corresponding static library is finished, and thereby get
12100 openssl and the test programs linked against the shared library.
12101 This still only happens when the keyword "shard" has been given to
12102 the configuration scripts.
12104 NOTE: shared library support is still an experimental thing, and
12105 backward binary compatibility is still not guaranteed.
12107 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12109 * Add support for Subject Information Access extension.
12111 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12113 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12114 additional bytes when new memory had to be allocated, not just
12115 when reusing an existing buffer.
12119 * New command line and configuration option 'utf8' for the req command.
12120 This allows field values to be specified as UTF8 strings.
12124 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12125 runs for the former and machine-readable output for the latter.
12129 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12130 of the e-mail address in the DN (i.e., it will go into a certificate
12131 extension only). The new configuration file option 'email_in_dn = no'
12132 has the same effect.
12134 *Massimiliano Pala madwolf@openca.org*
12136 * Change all functions with names starting with `des_` to be starting
12137 with `DES_` instead. Add wrappers that are compatible with libdes,
12138 but are named `_ossl_old_des_*`. Finally, add macros that map the
12139 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12140 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12141 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12144 Since we provide two compatibility mappings, the user needs to
12145 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12146 compatibility is desired. The default (i.e., when that macro
12147 isn't defined) is OpenSSL 0.9.6c compatibility.
12149 There are also macros that enable and disable the support of old
12150 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12151 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12152 are defined, the default will apply: to support the old des routines.
12154 In either case, one must include openssl/des.h to get the correct
12155 definitions. Do not try to just include openssl/des_old.h, that
12158 NOTE: This is a major break of an old API into a new one. Software
12159 authors are encouraged to switch to the `DES_` style functions. Some
12160 time in the future, des_old.h and the libdes compatibility functions
12161 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12162 default), and then completely removed.
12166 * Test for certificates which contain unsupported critical extensions.
12167 If such a certificate is found during a verify operation it is
12168 rejected by default: this behaviour can be overridden by either
12169 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12170 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12171 X509_supported_extension() has also been added which returns 1 if a
12172 particular extension is supported.
12176 * Modify the behaviour of EVP cipher functions in similar way to digests
12177 to retain compatibility with existing code.
12181 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12182 compatibility with existing code. In particular the 'ctx' parameter does
12183 not have to be to be initialized before the call to EVP_DigestInit() and
12184 it is tidied up after a call to EVP_DigestFinal(). New function
12185 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12186 EVP_MD_CTX_copy() changed to not require the destination to be
12187 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12188 requires the destination to be valid.
12190 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12191 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12195 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12196 so that complete 'Handshake' protocol structures are kept in memory
12197 instead of overwriting 'msg_type' and 'length' with 'body' data.
12201 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12203 *Massimo Santin via Richard Levitte*
12205 * Major restructuring to the underlying ENGINE code. This includes
12206 reduction of linker bloat, separation of pure "ENGINE" manipulation
12207 (initialisation, etc) from functionality dealing with implementations
12208 of specific crypto interfaces. This change also introduces integrated
12209 support for symmetric ciphers and digest implementations - so ENGINEs
12210 can now accelerate these by providing EVP_CIPHER and EVP_MD
12211 implementations of their own. This is detailed in
12212 [crypto/engine/README.md](crypto/engine/README.md)
12213 as it couldn't be adequately described here. However, there are a few
12214 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12215 were changed in the original introduction of ENGINE code have now
12216 reverted back - the hooking from this code to ENGINE is now a good
12217 deal more passive and at run-time, operations deal directly with
12218 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12219 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12220 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12221 they were not being used by the framework as there is no concept of a
12222 BIGNUM_METHOD and they could not be generalised to the new
12223 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12224 ENGINE_cpy() has been removed as it cannot be consistently defined in
12229 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12233 * Change mkdef.pl to sort symbols that get the same entry number,
12234 and make sure the automatically generated functions `ERR_load_*`
12235 become part of libeay.num as well.
12239 * New function SSL_renegotiate_pending(). This returns true once
12240 renegotiation has been requested (either SSL_renegotiate() call
12241 or HelloRequest/ClientHello received from the peer) and becomes
12242 false once a handshake has been completed.
12243 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12244 sends a HelloRequest, but does not ensure that a handshake takes
12245 place. SSL_renegotiate_pending() is useful for checking if the
12246 client has followed the request.)
12250 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12251 By default, clients may request session resumption even during
12252 renegotiation (if session ID contexts permit); with this option,
12253 session resumption is possible only in the first handshake.
12255 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12256 more bits available for options that should not be part of
12257 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12261 * Add some demos for certificate and certificate request creation.
12265 * Make maximum certificate chain size accepted from the peer application
12266 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12267 "Douglas E. Engert" <deengert@anl.gov>.
12271 * Add support for shared libraries for Unixware-7
12272 (Boyd Lynn Gerber <gerberb@zenez.com>).
12276 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12277 be done prior to destruction. Use this to unload error strings from
12278 ENGINEs that load their own error strings. NB: This adds two new API
12279 functions to "get" and "set" this destroy handler in an ENGINE.
12283 * Alter all existing ENGINE implementations (except "openssl" and
12284 "openbsd") to dynamically instantiate their own error strings. This
12285 makes them more flexible to be built both as statically-linked ENGINEs
12286 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12287 Also, add stub code to each that makes building them as self-contained
12288 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12292 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12293 implementations into applications that are completely implemented in
12294 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12295 commands that can be used to configure what shared-library to load and
12296 to control aspects of the way it is handled. Also, made an update to
12297 the [README-Engine.md](README-Engine.md) file
12298 that brings its information up-to-date and
12299 provides some information and instructions on the "dynamic" ENGINE
12300 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12304 * Make it possible to unload ranges of ERR strings with a new
12305 "ERR_unload_strings" function.
12309 * Add a copy() function to EVP_MD.
12313 * Make EVP_MD routines take a context pointer instead of just the
12314 md_data void pointer.
12318 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12319 that the digest can only process a single chunk of data
12320 (typically because it is provided by a piece of
12321 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12322 is only going to provide a single chunk of data, and hence the
12323 framework needn't accumulate the data for oneshot drivers.
12327 * As with "ERR", make it possible to replace the underlying "ex_data"
12328 functions. This change also alters the storage and management of global
12329 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12330 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12331 index counters. The API functions that use this state have been changed
12332 to take a "class_index" rather than pointers to the class's local STACK
12333 and counter, and there is now an API function to dynamically create new
12334 classes. This centralisation allows us to (a) plug a lot of the
12335 thread-safety problems that existed, and (b) makes it possible to clean
12336 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12337 such data would previously have always leaked in application code and
12338 workarounds were in place to make the memory debugging turn a blind eye
12339 to it. Application code that doesn't use this new function will still
12340 leak as before, but their memory debugging output will announce it now
12341 rather than letting it slide.
12343 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12344 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12345 has a return value to indicate success or failure.
12349 * Make it possible to replace the underlying "ERR" functions such that the
12350 global state (2 LHASH tables and 2 locks) is only used by the "default"
12351 implementation. This change also adds two functions to "get" and "set"
12352 the implementation prior to it being automatically set the first time
12353 any other ERR function takes place. Ie. an application can call "get",
12354 pass the return value to a module it has just loaded, and that module
12355 can call its own "set" function using that value. This means the
12356 module's "ERR" operations will use (and modify) the error state in the
12357 application and not in its own statically linked copy of OpenSSL code.
12361 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12362 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12363 the operation, and provides a more encapsulated way for external code
12364 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12365 to use these functions rather than manually incrementing the counts.
12367 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12371 * Add EVP test program.
12375 * Add symmetric cipher support to ENGINE. Expect the API to change!
12379 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12380 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12381 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12382 These allow a CRL to be built without having to access X509_CRL fields
12383 directly. Modify 'ca' application to use new functions.
12387 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12388 bug workarounds. Rollback attack detection is a security feature.
12389 The problem will only arise on OpenSSL servers when TLSv1 is not
12390 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12391 Software authors not wanting to support TLSv1 will have special reasons
12392 for their choice and can explicitly enable this option.
12394 *Bodo Moeller, Lutz Jaenicke*
12396 * Rationalise EVP so it can be extended: don't include a union of
12397 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12398 (similar to those existing for EVP_CIPHER_CTX).
12403 EVP_MD_CTX_init(&md); /* new function call */
12404 EVP_DigestInit(&md, EVP_sha1());
12405 EVP_DigestUpdate(&md, in, len);
12406 EVP_DigestFinal(&md, out, NULL);
12407 EVP_MD_CTX_cleanup(&md); /* new function call */
12411 * Make DES key schedule conform to the usual scheme, as well as
12412 correcting its structure. This means that calls to DES functions
12413 now have to pass a pointer to a des_key_schedule instead of a
12414 plain des_key_schedule (which was actually always a pointer
12417 des_key_schedule ks;
12419 des_set_key_checked(..., &ks);
12420 des_ncbc_encrypt(..., &ks, ...);
12422 (Note that a later change renames 'des_...' into 'DES_...'.)
12426 * Initial reduction of linker bloat: the use of some functions, such as
12427 PEM causes large amounts of unused functions to be linked in due to
12428 poor organisation. For example pem_all.c contains every PEM function
12429 which has a knock on effect of linking in large amounts of (unused)
12430 ASN1 code. Grouping together similar functions and splitting unrelated
12431 functions prevents this.
12435 * Cleanup of EVP macros.
12439 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12440 correct `_ecb suffix`.
12444 * Add initial OCSP responder support to ocsp application. The
12445 revocation information is handled using the text based index
12446 use by the ca application. The responder can either handle
12447 requests generated internally, supplied in files (for example
12448 via a CGI script) or using an internal minimal server.
12452 * Add configuration choices to get zlib compression for TLS.
12456 * Changes to Kerberos SSL for RFC 2712 compliance:
12457 1. Implemented real KerberosWrapper, instead of just using
12458 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12459 2. Implemented optional authenticator field of KerberosWrapper.
12461 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12462 and authenticator structs; see crypto/krb5/.
12464 Generalized Kerberos calls to support multiple Kerberos libraries.
12465 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12466 via Richard Levitte*
12468 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12469 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12470 values for each of the key sizes rather than having just
12471 parameters (and 'speed' generating keys each time).
12475 * Speed up EVP routines.
12478 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12479 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12480 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12481 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12483 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12484 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12485 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12488 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12490 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12494 * Added the OS2-EMX target.
12496 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12498 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12499 New functions to support `NCONF `routines in extension code.
12500 New function `CONF_set_nconf()`
12501 to allow functions which take an `NCONF` to also handle the old `LHASH`
12502 structure: this means that the old `CONF` compatible routines can be
12503 retained (in particular w.rt. extensions) without having to duplicate the
12504 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12508 * Enhance the general user interface with mechanisms for inner control
12509 and with possibilities to have yes/no kind of prompts.
12513 * Change all calls to low level digest routines in the library and
12514 applications to use EVP. Add missing calls to HMAC_cleanup() and
12515 don't assume HMAC_CTX can be copied using memcpy().
12517 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12519 * Add the possibility to control engines through control names but with
12520 arbitrary arguments instead of just a string.
12521 Change the key loaders to take a UI_METHOD instead of a callback
12522 function pointer. NOTE: this breaks binary compatibility with earlier
12523 versions of OpenSSL [engine].
12524 Adapt the nCipher code for these new conditions and add a card insertion
12529 * Enhance the general user interface with mechanisms to better support
12530 dialog box interfaces, application-defined prompts, the possibility
12531 to use defaults (for example default passwords from somewhere else)
12532 and interrupts/cancellations.
12536 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12537 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12541 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12542 tidy up some unnecessarily weird code in 'sk_new()').
12544 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12546 * Change the key loading routines for ENGINEs to use the same kind
12547 callback (pem_password_cb) as all other routines that need this
12552 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12553 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12554 than this minimum value is recommended.
12558 * New random seeder for OpenVMS, using the system process statistics
12559 that are easily reachable.
12563 * Windows apparently can't transparently handle global
12564 variables defined in DLLs. Initialisations such as:
12566 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12568 won't compile. This is used by the any applications that need to
12569 declare their own ASN1 modules. This was fixed by adding the option
12570 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12571 needed for static libraries under Win32.
12575 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12576 setting of purpose and trust fields. New X509_STORE trust and
12577 purpose functions and tidy up setting in other SSL functions.
12581 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12582 structure. These are inherited by X509_STORE_CTX when it is
12583 initialised. This allows various defaults to be set in the
12584 X509_STORE structure (such as flags for CRL checking and custom
12585 purpose or trust settings) for functions which only use X509_STORE_CTX
12586 internally such as S/MIME.
12588 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12589 trust settings if they are not set in X509_STORE. This allows X509_STORE
12590 purposes and trust (in S/MIME for example) to override any set by default.
12592 Add command line options for CRL checking to smime, s_client and s_server
12597 * Initial CRL based revocation checking. If the CRL checking flag(s)
12598 are set then the CRL is looked up in the X509_STORE structure and
12599 its validity and signature checked, then if the certificate is found
12600 in the CRL the verify fails with a revoked error.
12602 Various new CRL related callbacks added to X509_STORE_CTX structure.
12604 Command line options added to 'verify' application to support this.
12606 This needs some additional work, such as being able to handle multiple
12607 CRLs with different times, extension based lookup (rather than just
12608 by subject name) and ultimately more complete V2 CRL extension
12613 * Add a general user interface API (crypto/ui/). This is designed
12614 to replace things like des_read_password and friends (backward
12615 compatibility functions using this new API are provided).
12616 The purpose is to remove prompting functions from the DES code
12617 section as well as provide for prompting through dialog boxes in
12618 a window system and the like.
12622 * Add "ex_data" support to ENGINE so implementations can add state at a
12623 per-structure level rather than having to store it globally.
12627 * Make it possible for ENGINE structures to be copied when retrieved by
12628 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12629 This causes the "original" ENGINE structure to act like a template,
12630 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12631 operational state can be localised to each ENGINE structure, despite the
12632 fact they all share the same "methods". New ENGINE structures returned in
12633 this case have no functional references and the return value is the single
12634 structural reference. This matches the single structural reference returned
12635 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12640 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12641 needs to match any other type at all we need to manually clear the
12646 * Changes to the "openssl engine" utility to include;
12647 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12648 about an ENGINE's available control commands.
12649 - executing control commands from command line arguments using the
12650 '-pre' and '-post' switches. '-post' is only used if '-t' is
12651 specified and the ENGINE is successfully initialised. The syntax for
12652 the individual commands are colon-separated, for example;
12653 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12657 * New dynamic control command support for ENGINEs. ENGINEs can now
12658 declare their own commands (numbers), names (strings), descriptions,
12659 and input types for run-time discovery by calling applications. A
12660 subset of these commands are implicitly classed as "executable"
12661 depending on their input type, and only these can be invoked through
12662 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12663 can be based on user input, config files, etc). The distinction is
12664 that "executable" commands cannot return anything other than a boolean
12665 result and can only support numeric or string input, whereas some
12666 discoverable commands may only be for direct use through
12667 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12668 pointers, or other custom uses. The "executable" commands are to
12669 support parameterisations of ENGINE behaviour that can be
12670 unambiguously defined by ENGINEs and used consistently across any
12671 OpenSSL-based application. Commands have been added to all the
12672 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12673 control over shared-library paths without source code alterations.
12677 * Changed all ENGINE implementations to dynamically allocate their
12678 ENGINEs rather than declaring them statically. Apart from this being
12679 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12680 this also allows the implementations to compile without using the
12681 internal engine_int.h header.
12685 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12686 'const' value. Any code that should be able to modify a RAND_METHOD
12687 should already have non-const pointers to it (ie. they should only
12688 modify their own ones).
12692 * Made a variety of little tweaks to the ENGINE code.
12693 - "atalla" and "ubsec" string definitions were moved from header files
12694 to C code. "nuron" string definitions were placed in variables
12695 rather than hard-coded - allowing parameterisation of these values
12696 later on via ctrl() commands.
12697 - Removed unused "#if 0"'d code.
12698 - Fixed engine list iteration code so it uses ENGINE_free() to release
12699 structural references.
12700 - Constified the RAND_METHOD element of ENGINE structures.
12701 - Constified various get/set functions as appropriate and added
12702 missing functions (including a catch-all ENGINE_cpy that duplicates
12703 all ENGINE values onto a new ENGINE except reference counts/state).
12704 - Removed NULL parameter checks in get/set functions. Setting a method
12705 or function to NULL is a way of cancelling out a previously set
12706 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12707 and doesn't justify the extra error symbols and code.
12708 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12709 flags from engine_int.h to engine.h.
12710 - Changed prototypes for ENGINE handler functions (init(), finish(),
12711 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12715 * Implement binary inversion algorithm for BN_mod_inverse in addition
12716 to the algorithm using long division. The binary algorithm can be
12717 used only if the modulus is odd. On 32-bit systems, it is faster
12718 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12719 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12720 up to 450 bits. In 64-bit environments, the binary algorithm
12721 appears to be advantageous for much longer moduli; here we use it
12722 for moduli up to 2048 bits.
12726 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12727 could not support the combine flag in choice fields.
12731 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12732 extensions from a certificate request to the certificate.
12736 * Allow multiple 'certopt' and 'nameopt' options to be separated
12737 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12738 file: this allows the display of the certificate about to be
12739 signed to be customised, to allow certain fields to be included
12740 or excluded and extension details. The old system didn't display
12741 multicharacter strings properly, omitted fields not in the policy
12742 and couldn't display additional details such as extensions.
12746 * Function EC_POINTs_mul for multiple scalar multiplication
12747 of an arbitrary number of elliptic curve points
12748 \sum scalars[i]*points[i],
12749 optionally including the generator defined for the EC_GROUP:
12750 scalar*generator + \sum scalars[i]*points[i].
12752 EC_POINT_mul is a simple wrapper function for the typical case
12753 that the point list has just one item (besides the optional
12758 * First EC_METHODs for curves over GF(p):
12760 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12761 operations and provides various method functions that can also
12762 operate with faster implementations of modular arithmetic.
12764 EC_GFp_mont_method() reuses most functions that are part of
12765 EC_GFp_simple_method, but uses Montgomery arithmetic.
12767 *Bodo Moeller; point addition and point doubling
12768 implementation directly derived from source code provided by
12769 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12771 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12772 crypto/ec/ec_lib.c):
12774 Curves are EC_GROUP objects (with an optional group generator)
12775 based on EC_METHODs that are built into the library.
12777 Points are EC_POINT objects based on EC_GROUP objects.
12779 Most of the framework would be able to handle curves over arbitrary
12780 finite fields, but as there are no obvious types for fields other
12781 than GF(p), some functions are limited to that for now.
12785 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12786 that the file contains a complete HTTP response.
12790 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12791 change the def and num file printf format specifier from "%-40sXXX"
12792 to "%-39s XXX". The latter will always guarantee a space after the
12793 field while the former will cause them to run together if the field
12794 is 40 of more characters long.
12798 * Constify the cipher and digest 'method' functions and structures
12799 and modify related functions to take constant EVP_MD and EVP_CIPHER
12804 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12805 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12809 * Modify `EVP_Digest*()` routines so they now return values. Although the
12810 internal software routines can never fail additional hardware versions
12815 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12817 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12818 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12821 ERR_R_NESTED_ASN1_ERROR
12823 ERR_R_MISSING_ASN1_EOS
12824 were 4 .. 9, conflicting with
12825 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12827 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12828 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12830 Add new error code 'ERR_R_INTERNAL_ERROR'.
12834 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12839 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12840 sets the subject name for a new request or supersedes the
12841 subject name in a given request. Formats that can be parsed are
12842 'CN=Some Name, OU=myOU, C=IT'
12844 'CN=Some Name/OU=myOU/C=IT'.
12846 Add options '-batch' and '-verbose' to 'openssl req'.
12848 *Massimiliano Pala <madwolf@hackmasters.net>*
12850 * Introduce the possibility to access global variables through
12851 functions on platform were that's the best way to handle exporting
12852 global variables in shared libraries. To enable this functionality,
12853 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12854 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12855 is normally done by Configure or something similar).
12857 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12858 in the source file (foo.c) like this:
12860 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12861 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12863 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12864 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12866 OPENSSL_DECLARE_GLOBAL(int,foo);
12867 #define foo OPENSSL_GLOBAL_REF(foo)
12868 OPENSSL_DECLARE_GLOBAL(double,bar);
12869 #define bar OPENSSL_GLOBAL_REF(bar)
12871 The #defines are very important, and therefore so is including the
12872 header file everywhere where the defined globals are used.
12874 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12875 of ASN.1 items, but that structure is a bit different.
12877 The largest change is in util/mkdef.pl which has been enhanced with
12878 better and easier to understand logic to choose which symbols should
12879 go into the Windows .def files as well as a number of fixes and code
12880 cleanup (among others, algorithm keywords are now sorted
12881 lexicographically to avoid constant rewrites).
12885 * In BN_div() keep a copy of the sign of 'num' before writing the
12886 result to 'rm' because if rm==num the value will be overwritten
12887 and produce the wrong result if 'num' is negative: this caused
12888 problems with BN_mod() and BN_nnmod().
12892 * Function OCSP_request_verify(). This checks the signature on an
12893 OCSP request and verifies the signer certificate. The signer
12894 certificate is just checked for a generic purpose and OCSP request
12899 * Add OCSP_check_validity() function to check the validity of OCSP
12900 responses. OCSP responses are prepared in real time and may only
12901 be a few seconds old. Simply checking that the current time lies
12902 between thisUpdate and nextUpdate max reject otherwise valid responses
12903 caused by either OCSP responder or client clock inaccuracy. Instead
12904 we allow thisUpdate and nextUpdate to fall within a certain period of
12905 the current time. The age of the response can also optionally be
12906 checked. Two new options -validity_period and -status_age added to
12911 * If signature or public key algorithm is unrecognized print out its
12912 OID rather that just UNKNOWN.
12916 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12917 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12918 ID to be generated from the issuer certificate alone which can then be
12919 passed to OCSP_id_issuer_cmp().
12923 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12924 ASN1 modules to export functions returning ASN1_ITEM pointers
12925 instead of the ASN1_ITEM structures themselves. This adds several
12926 new macros which allow the underlying ASN1 function/structure to
12927 be accessed transparently. As a result code should not use ASN1_ITEM
12928 references directly (such as &X509_it) but instead use the relevant
12929 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12930 use of the new ASN1 code on platforms where exporting structures
12931 is problematical (for example in shared libraries) but exporting
12932 functions returning pointers to structures is not.
12936 * Add support for overriding the generation of SSL/TLS session IDs.
12937 These callbacks can be registered either in an SSL_CTX or per SSL.
12938 The purpose of this is to allow applications to control, if they wish,
12939 the arbitrary values chosen for use as session IDs, particularly as it
12940 can be useful for session caching in multiple-server environments. A
12941 command-line switch for testing this (and any client code that wishes
12942 to use such a feature) has been added to "s_server".
12944 *Geoff Thorpe, Lutz Jaenicke*
12946 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12947 of the form `#if defined(...) || defined(...) || ...` and
12948 `#if !defined(...) && !defined(...) && ...`. This also avoids
12949 the growing number of special cases it was previously handling.
12953 * Make all configuration macros available for application by making
12954 sure they are available in opensslconf.h, by giving them names starting
12955 with `OPENSSL_` to avoid conflicts with other packages and by making
12956 sure e_os2.h will cover all platform-specific cases together with
12958 Additionally, it is now possible to define configuration/platform-
12959 specific names (called "system identities"). In the C code, these
12960 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12961 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12962 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12967 * New option -set_serial to 'req' and 'x509' this allows the serial
12968 number to use to be specified on the command line. Previously self
12969 signed certificates were hard coded with serial number 0 and the
12970 CA options of 'x509' had to use a serial number in a file which was
12975 * New options to 'ca' utility to support V2 CRL entry extensions.
12976 Currently CRL reason, invalidity date and hold instruction are
12977 supported. Add new CRL extensions to V3 code and some new objects.
12981 * New function EVP_CIPHER_CTX_set_padding() this is used to
12982 disable standard block padding (aka PKCS#5 padding) in the EVP
12983 API, which was previously mandatory. This means that the data is
12984 not padded in any way and so the total length much be a multiple
12985 of the block size, otherwise an error occurs.
12989 * Initial (incomplete) OCSP SSL support.
12993 * New function OCSP_parse_url(). This splits up a URL into its host,
12994 port and path components: primarily to parse OCSP URLs. New -url
12995 option to ocsp utility.
12999 * New nonce behavior. The return value of OCSP_check_nonce() now
13000 reflects the various checks performed. Applications can decide
13001 whether to tolerate certain situations such as an absent nonce
13002 in a response when one was present in a request: the ocsp application
13003 just prints out a warning. New function OCSP_add1_basic_nonce()
13004 this is to allow responders to include a nonce in a response even if
13005 the request is nonce-less.
13009 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13010 skipped when using openssl x509 multiple times on a single input file,
13011 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13015 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13016 set string type: to handle setting ASN1_TIME structures. Fix ca
13017 utility to correctly initialize revocation date of CRLs.
13021 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13022 the clients preferred ciphersuites and rather use its own preferences.
13023 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13024 Internet Explorer by ensuring unchanged hash method during stepup.
13025 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13029 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13030 to aes and add a new 'exist' option to print out symbols that don't
13035 * Additional options to ocsp utility to allow flags to be set and
13036 additional certificates supplied.
13040 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13041 OCSP client a number of certificate to only verify the response
13046 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13047 handle the new API. Currently only ECB, CBC modes supported. Add new
13050 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13051 Encryption Standard (AES) Ciphersuites for Transport Layer
13052 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13053 not enabled by default and were not part of the "ALL" ciphersuite
13054 alias because they were not yet official; they could be
13055 explicitly requested by specifying the "AESdraft" ciphersuite
13056 group alias. In the final release of OpenSSL 0.9.7, the group
13057 alias is called "AES" and is part of "ALL".)
13059 *Ben Laurie, Steve Henson, Bodo Moeller*
13061 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13062 request to response.
13066 * Functions for OCSP responders. OCSP_request_onereq_count(),
13067 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13068 extract information from a certificate request. OCSP_response_create()
13069 creates a response and optionally adds a basic response structure.
13070 OCSP_basic_add1_status() adds a complete single response to a basic
13071 response and returns the OCSP_SINGLERESP structure just added (to allow
13072 extensions to be included for example). OCSP_basic_add1_cert() adds a
13073 certificate to a basic response and OCSP_basic_sign() signs a basic
13074 response with various flags. New helper functions ASN1_TIME_check()
13075 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13076 (converts ASN1_TIME to GeneralizedTime).
13080 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13081 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13082 structure from a certificate. X509_pubkey_digest() digests the public_key
13083 contents: this is used in various key identifiers.
13087 * Make sk_sort() tolerate a NULL argument.
13089 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13091 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13092 passed by the function are trusted implicitly. If any of them signed the
13093 response then it is assumed to be valid and is not verified.
13097 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13098 to data. This was previously part of the PKCS7 ASN1 code. This
13099 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13100 *Steve Henson, reported by Kenneth R. Robinette
13101 <support@securenetterm.com>*
13103 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13104 routines: without these tracing memory leaks is very painful.
13105 Fix leaks in PKCS12 and PKCS7 routines.
13109 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13110 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13111 effectively meant GeneralizedTime would never be used. Now it
13112 is initialised to -1 but X509_time_adj() now has to check the value
13113 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13114 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13115 *Steve Henson, reported by Kenneth R. Robinette
13116 <support@securenetterm.com>*
13118 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13119 result in a zero length in the ASN1_INTEGER structure which was
13120 not consistent with the structure when d2i_ASN1_INTEGER() was used
13121 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13122 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13123 where it did not print out a minus for negative ASN1_INTEGER.
13127 * Add summary printout to ocsp utility. The various functions which
13128 convert status values to strings have been renamed to:
13129 OCSP_response_status_str(), OCSP_cert_status_str() and
13130 OCSP_crl_reason_str() and are no longer static. New options
13131 to verify nonce values and to disable verification. OCSP response
13132 printout format cleaned up.
13136 * Add additional OCSP certificate checks. These are those specified
13137 in RFC2560. This consists of two separate checks: the CA of the
13138 certificate being checked must either be the OCSP signer certificate
13139 or the issuer of the OCSP signer certificate. In the latter case the
13140 OCSP signer certificate must contain the OCSP signing extended key
13141 usage. This check is performed by attempting to match the OCSP
13142 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13143 in the OCSP_CERTID structures of the response.
13147 * Initial OCSP certificate verification added to OCSP_basic_verify()
13148 and related routines. This uses the standard OpenSSL certificate
13149 verify routines to perform initial checks (just CA validity) and
13150 to obtain the certificate chain. Then additional checks will be
13151 performed on the chain. Currently the root CA is checked to see
13152 if it is explicitly trusted for OCSP signing. This is used to set
13153 a root CA as a global signing root: that is any certificate that
13154 chains to that CA is an acceptable OCSP signing certificate.
13158 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13159 extensions from a separate configuration file.
13160 As when reading extensions from the main configuration file,
13161 the '-extensions ...' option may be used for specifying the
13164 *Massimiliano Pala <madwolf@comune.modena.it>*
13166 * New OCSP utility. Allows OCSP requests to be generated or
13167 read. The request can be sent to a responder and the output
13168 parsed, outputted or printed in text form. Not complete yet:
13169 still needs to check the OCSP response validity.
13173 * New subcommands for 'openssl ca':
13174 `openssl ca -status <serial>` prints the status of the cert with
13175 the given serial number (according to the index file).
13176 `openssl ca -updatedb` updates the expiry status of certificates
13179 *Massimiliano Pala <madwolf@comune.modena.it>*
13181 * New '-newreq-nodes' command option to CA.pl. This is like
13182 '-newreq', but calls 'openssl req' with the '-nodes' option
13183 so that the resulting key is not encrypted.
13185 *Damien Miller <djm@mindrot.org>*
13187 * New configuration for the GNU Hurd.
13189 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13191 * Initial code to implement OCSP basic response verify. This
13192 is currently incomplete. Currently just finds the signer's
13193 certificate and verifies the signature on the response.
13197 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13198 value of OPENSSLDIR. This is available via the new '-d' option
13199 to 'openssl version', and is also included in 'openssl version -a'.
13203 * Allowing defining memory allocation callbacks that will be given
13204 file name and line number information in additional arguments
13205 (a `const char*` and an int). The basic functionality remains, as
13206 well as the original possibility to just replace malloc(),
13207 realloc() and free() by functions that do not know about these
13208 additional arguments. To register and find out the current
13209 settings for extended allocation functions, the following
13210 functions are provided:
13212 CRYPTO_set_mem_ex_functions
13213 CRYPTO_set_locked_mem_ex_functions
13214 CRYPTO_get_mem_ex_functions
13215 CRYPTO_get_locked_mem_ex_functions
13217 These work the same way as CRYPTO_set_mem_functions and friends.
13218 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13219 extended allocation function is enabled.
13220 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13221 a conventional allocation function is enabled.
13223 *Richard Levitte, Bodo Moeller*
13225 * Finish off removing the remaining LHASH function pointer casts.
13226 There should no longer be any prototype-casting required when using
13227 the LHASH abstraction, and any casts that remain are "bugs". See
13228 the callback types and macros at the head of lhash.h for details
13229 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13233 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13234 If /dev/[u]random devices are not available or do not return enough
13235 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13237 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13238 /etc/entropy will be queried once each in this sequence, querying stops
13239 when enough entropy was collected without querying more sockets.
13243 * Change the Unix RAND_poll() variant to be able to poll several
13244 random devices, as specified by DEVRANDOM, until a sufficient amount
13245 of data has been collected. We spend at most 10 ms on each file
13246 (select timeout) and read in non-blocking mode. DEVRANDOM now
13247 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13248 (previously it was just the string "/dev/urandom"), so on typical
13249 platforms the 10 ms delay will never occur.
13250 Also separate out the Unix variant to its own file, rand_unix.c.
13251 For VMS, there's a currently-empty rand_vms.c.
13255 * Move OCSP client related routines to ocsp_cl.c. These
13256 provide utility functions which an application needing
13257 to issue a request to an OCSP responder and analyse the
13258 response will typically need: as opposed to those which an
13259 OCSP responder itself would need which will be added later.
13261 OCSP_request_sign() signs an OCSP request with an API similar
13262 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13263 response. OCSP_response_get1_basic() extracts basic response
13264 from response. OCSP_resp_find_status(): finds and extracts status
13265 information from an OCSP_CERTID structure (which will be created
13266 when the request structure is built). These are built from lower
13267 level functions which work on OCSP_SINGLERESP structures but
13268 won't normally be used unless the application wishes to examine
13269 extensions in the OCSP response for example.
13271 Replace nonce routines with a pair of functions.
13272 OCSP_request_add1_nonce() adds a nonce value and optionally
13273 generates a random value. OCSP_check_nonce() checks the
13274 validity of the nonce in an OCSP response.
13278 * Change function OCSP_request_add() to OCSP_request_add0_id().
13279 This doesn't copy the supplied OCSP_CERTID and avoids the
13280 need to free up the newly created id. Change return type
13281 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13282 This can then be used to add extensions to the request.
13283 Deleted OCSP_request_new(), since most of its functionality
13284 is now in OCSP_REQUEST_new() (and the case insensitive name
13285 clash) apart from the ability to set the request name which
13286 will be added elsewhere.
13290 * Update OCSP API. Remove obsolete extensions argument from
13291 various functions. Extensions are now handled using the new
13292 OCSP extension code. New simple OCSP HTTP function which
13293 can be used to send requests and parse the response.
13297 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13298 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13299 uses the special reorder version of SET OF to sort the attributes
13300 and reorder them to match the encoded order. This resolves a long
13301 standing problem: a verify on a PKCS7 structure just after signing
13302 it used to fail because the attribute order did not match the
13303 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13304 it uses the received order. This is necessary to tolerate some broken
13305 software that does not order SET OF. This is handled by encoding
13306 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13307 to produce the required SET OF.
13311 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13312 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13313 files to get correct declarations of the ASN.1 item variables.
13317 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13318 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13319 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13320 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13321 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13322 ASN1_ITEM and no wrapper functions.
13326 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13327 replace the old function pointer based I/O routines. Change most of
13328 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13332 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13333 lines, recognize more "algorithms" that can be deselected, and make
13334 it complain about algorithm deselection that isn't recognised.
13338 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13339 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13340 to use new functions. Add NO_ASN1_OLD which can be set to remove
13341 some old style ASN1 functions: this can be used to determine if old
13342 code will still work when these eventually go away.
13346 * New extension functions for OCSP structures, these follow the
13347 same conventions as certificates and CRLs.
13351 * New function X509V3_add1_i2d(). This automatically encodes and
13352 adds an extension. Its behaviour can be customised with various
13353 flags to append, replace or delete. Various wrappers added for
13354 certificates and CRLs.
13358 * Fix to avoid calling the underlying ASN1 print routine when
13359 an extension cannot be parsed. Correct a typo in the
13360 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13364 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13365 entries for variables.
13369 * Add functionality to `apps/openssl.c` for detecting locking
13370 problems: As the program is single-threaded, all we have
13371 to do is register a locking callback using an array for
13372 storing which locks are currently held by the program.
13376 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13377 SSL_get_ex_data_X509_STORE_idx(), which is used in
13378 ssl_verify_cert_chain() and thus can be called at any time
13379 during TLS/SSL handshakes so that thread-safety is essential.
13380 Unfortunately, the ex_data design is not at all suited
13381 for multi-threaded use, so it probably should be abolished.
13385 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13387 *Broadcom, tweaked and integrated by Geoff Thorpe*
13389 * Move common extension printing code to new function
13390 X509V3_print_extensions(). Reorganise OCSP print routines and
13391 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13395 * New function X509_signature_print() to remove duplication in some
13400 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13401 set (this was treated exactly the same as SET OF previously). This
13402 is used to reorder the STACK representing the structure to match the
13403 encoding. This will be used to get round a problem where a PKCS7
13404 structure which was signed could not be verified because the STACK
13405 order did not reflect the encoded order.
13409 * Reimplement the OCSP ASN1 module using the new code.
13413 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13414 for its ASN1 operations. The old style function pointers still exist
13415 for now but they will eventually go away.
13419 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13420 completely replaces the old ASN1 functionality with a table driven
13421 encoder and decoder which interprets an ASN1_ITEM structure describing
13422 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13423 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13424 has also been converted to the new form.
13428 * Change BN_mod_exp_recp so that negative moduli are tolerated
13429 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13430 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13431 for negative moduli.
13435 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13436 of not touching the result's sign bit.
13440 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13445 * Changed the LHASH code to use prototypes for callbacks, and created
13446 macros to declare and implement thin (optionally static) functions
13447 that provide type-safety and avoid function pointer casting for the
13448 type-specific callbacks.
13452 * Added Kerberos Cipher Suites to be used with TLS, as written in
13454 *Veers Staats <staatsvr@asc.hpc.mil>,
13455 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13457 * Reformat the FAQ so the different questions and answers can be divided
13458 in sections depending on the subject.
13462 * Have the zlib compression code load ZLIB.DLL dynamically under
13467 * New function BN_mod_sqrt for computing square roots modulo a prime
13468 (using the probabilistic Tonelli-Shanks algorithm unless
13469 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13470 be handled deterministically).
13472 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13474 * Make BN_mod_inverse faster by explicitly handling small quotients
13475 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13476 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13480 * New function BN_kronecker.
13484 * Fix BN_gcd so that it works on negative inputs; the result is
13485 positive unless both parameters are zero.
13486 Previously something reasonably close to an infinite loop was
13487 possible because numbers could be growing instead of shrinking
13488 in the implementation of Euclid's algorithm.
13492 * Fix BN_is_word() and BN_is_one() macros to take into account the
13493 sign of the number in question.
13495 Fix BN_is_word(a,w) to work correctly for w == 0.
13497 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13498 because its test if the absolute value of 'a' equals 'w'.
13499 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13500 it exists mostly for use in the implementations of BN_is_zero(),
13501 BN_is_one(), and BN_is_word().
13505 * New function BN_swap.
13509 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13510 the exponentiation functions are more likely to produce reasonable
13511 results on negative inputs.
13515 * Change BN_mod_mul so that the result is always non-negative.
13516 Previously, it could be negative if one of the factors was negative;
13517 I don't think anyone really wanted that behaviour.
13521 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13522 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13523 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13524 and add new functions:
13533 BN_mod_lshift1_quick
13535 BN_mod_lshift_quick
13537 These functions always generate non-negative results.
13539 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13540 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13542 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13543 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13544 be reduced modulo `m`.
13546 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13549 The following entry accidentally appeared in the CHANGES file
13550 distributed with OpenSSL 0.9.7. The modifications described in
13551 it do *not* apply to OpenSSL 0.9.7.
13553 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13554 was actually never needed) and in BN_mul(). The removal in BN_mul()
13555 required a small change in bn_mul_part_recursive() and the addition
13556 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13557 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13558 bn_sub_words() and bn_add_words() except they take arrays with
13564 * In 'openssl passwd', verify passwords read from the terminal
13565 unless the '-salt' option is used (which usually means that
13566 verification would just waste user's time since the resulting
13567 hash is going to be compared with some given password hash)
13568 or the new '-noverify' option is used.
13570 This is an incompatible change, but it does not affect
13571 non-interactive use of 'openssl passwd' (passwords on the command
13572 line, '-stdin' option, '-in ...' option) and thus should not
13573 cause any problems.
13577 * Remove all references to RSAref, since there's no more need for it.
13581 * Make DSO load along a path given through an environment variable
13582 (SHLIB_PATH) with shl_load().
13586 * Constify the ENGINE code as a result of BIGNUM constification.
13587 Also constify the RSA code and most things related to it. In a
13588 few places, most notable in the depth of the ASN.1 code, ugly
13589 casts back to non-const were required (to be solved at a later
13594 * Make it so the openssl application has all engines loaded by default.
13598 * Constify the BIGNUM routines a little more.
13602 * Add the following functions:
13604 ENGINE_load_cswift()
13606 ENGINE_load_atalla()
13607 ENGINE_load_nuron()
13608 ENGINE_load_builtin_engines()
13610 That way, an application can itself choose if external engines that
13611 are built-in in OpenSSL shall ever be used or not. The benefit is
13612 that applications won't have to be linked with libdl or other dso
13613 libraries unless it's really needed.
13615 Changed 'openssl engine' to load all engines on demand.
13616 Changed the engine header files to avoid the duplication of some
13617 declarations (they differed!).
13621 * 'openssl engine' can now list capabilities.
13625 * Better error reporting in 'openssl engine'.
13629 * Never call load_dh_param(NULL) in s_server.
13633 * Add engine application. It can currently list engines by name and
13634 identity, and test if they are actually available.
13638 * Improve RPM specification file by forcing symbolic linking and making
13639 sure the installed documentation is also owned by root.root.
13641 *Damien Miller <djm@mindrot.org>*
13643 * Give the OpenSSL applications more possibilities to make use of
13644 keys (public as well as private) handled by engines.
13648 * Add OCSP code that comes from CertCo.
13652 * Add VMS support for the Rijndael code.
13656 * Added untested support for Nuron crypto accelerator.
13660 * Add support for external cryptographic devices. This code was
13661 previously distributed separately as the "engine" branch.
13663 *Geoff Thorpe, Richard Levitte*
13665 * Rework the filename-translation in the DSO code. It is now possible to
13666 have far greater control over how a "name" is turned into a filename
13667 depending on the operating environment and any oddities about the
13668 different shared library filenames on each system.
13672 * Support threads on FreeBSD-elf in Configure.
13676 * Fix for SHA1 assembly problem with MASM: it produces
13677 warnings about corrupt line number information when assembling
13678 with debugging information. This is caused by the overlapping
13681 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13684 NCONF_get_number() has no error checking at all. As a replacement,
13685 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13686 promoted strongly. The old NCONF_get_number is kept around for
13687 binary backward compatibility.
13688 Make it possible for methods to load from something other than a BIO,
13689 by providing a function pointer that is given a name instead of a BIO.
13690 For example, this could be used to load configuration data from an
13695 * Fix for non blocking accept BIOs. Added new I/O special reason
13696 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13697 with non blocking I/O was not possible because no retry code was
13698 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13703 * Added the beginnings of Rijndael support.
13707 * Fix for bug in DirectoryString mask setting. Add support for
13708 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13709 to allow certificate printing to more controllable, additional
13710 'certopt' option to 'x509' to allow new printing options to be
13715 * Clean old EAY MD5 hack from e_os.h.
13719 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13721 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13722 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13724 *Joe Orton, Steve Henson*
13726 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13728 * Fix additional bug revealed by the NISCC test suite:
13730 Stop bug triggering large recursion when presented with
13731 certain ASN.1 tags ([CVE-2003-0851])
13735 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13737 * Fix various bugs revealed by running the NISCC test suite:
13739 Stop out of bounds reads in the ASN1 code when presented with
13740 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13742 If verify callback ignores invalid public key errors don't try to check
13743 certificate signature with the NULL public key.
13747 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13748 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13753 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13754 extra data after the compression methods not only for TLS 1.0
13755 but also for SSL 3.0 (as required by the specification).
13757 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13759 * Change X509_certificate_type() to mark the key as exported/exportable
13760 when it's 512 *bits* long, not 512 bytes.
13764 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13766 * Countermeasure against the Klima-Pokorny-Rosa extension of
13767 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13768 a protocol version number mismatch like a decryption error
13769 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13773 * Turn on RSA blinding by default in the default implementation
13774 to avoid a timing attack. Applications that don't want it can call
13775 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13776 They would be ill-advised to do so in most cases.
13778 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13780 * Change RSA blinding code so that it works when the PRNG is not
13781 seeded (in this case, the secret RSA exponent is abused as
13782 an unpredictable seed -- if it is not unpredictable, there
13783 is no point in blinding anyway). Make RSA blinding thread-safe
13784 by remembering the creator's thread ID in rsa->blinding and
13785 having all other threads use local one-time blinding factors
13786 (this requires more computation than sharing rsa->blinding, but
13787 avoids excessive locking; and if an RSA object is not shared
13788 between threads, blinding will still be very fast).
13792 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13794 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13795 via timing by performing a MAC computation even if incorrect
13796 block cipher padding has been found. This is a countermeasure
13797 against active attacks where the attacker has to distinguish
13798 between bad padding and a MAC verification error. ([CVE-2003-0078])
13800 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13801 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13802 Martin Vuagnoux (EPFL, Ilion)*
13804 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13806 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13807 memory from its contents. This is done with a counter that will
13808 place alternating values in each byte. This can be used to solve
13809 two issues: 1) the removal of calls to memset() by highly optimizing
13810 compilers, and 2) cleansing with other values than 0, since those can
13811 be read through on certain media, for example a swap space on disk.
13815 * Bugfix: client side session caching did not work with external caching,
13816 because the session->cipher setting was not restored when reloading
13817 from the external cache. This problem was masked, when
13818 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13819 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13823 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13824 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13826 *Zeev Lieber <zeev-l@yahoo.com>*
13828 * Undo an undocumented change introduced in 0.9.6e which caused
13829 repeated calls to OpenSSL_add_all_ciphers() and
13830 OpenSSL_add_all_digests() to be ignored, even after calling
13835 * Change the default configuration reader to deal with last line not
13836 being properly terminated.
13840 * Change X509_NAME_cmp() so it applies the special rules on handling
13841 DN values that are of type PrintableString, as well as RDNs of type
13842 emailAddress where the value has the type ia5String.
13844 *stefank@valicert.com via Richard Levitte*
13846 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13847 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13848 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13849 the bitwise-OR of the two for use by the majority of applications
13850 wanting this behaviour, and update the docs. The documented
13851 behaviour and actual behaviour were inconsistent and had been
13852 changing anyway, so this is more a bug-fix than a behavioural
13855 *Geoff Thorpe, diagnosed by Nadav Har'El*
13857 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13858 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13862 * Fix initialization code race conditions in
13863 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13864 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13865 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13866 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13867 ssl2_get_cipher_by_char(),
13868 ssl3_get_cipher_by_char().
13870 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13872 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13873 the cached sessions are flushed, as the remove_cb() might use ex_data
13874 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13875 (see [openssl.org #212]).
13877 *Geoff Thorpe, Lutz Jaenicke*
13879 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13880 length, instead of the encoding length to d2i_ASN1_OBJECT.
13884 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13886 * [In 0.9.6g-engine release:]
13887 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13889 *Lynn Gazis <lgazis@rainbow.com>*
13891 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13893 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13894 and get fix the header length calculation.
13895 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13896 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13898 * Use proper error handling instead of 'assertions' in buffer
13899 overflow checks added in 0.9.6e. This prevents DoS (the
13900 assertions could call abort()).
13902 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13904 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13906 * Add various sanity checks to asn1_get_length() to reject
13907 the ASN1 length bytes if they exceed sizeof(long), will appear
13908 negative or the content length exceeds the length of the
13911 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13913 * Fix cipher selection routines: ciphers without encryption had no flags
13914 for the cipher strength set and where therefore not handled correctly
13915 by the selection routines (PR #130).
13919 * Fix EVP_dsa_sha macro.
13924 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13925 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13926 that was added in OpenSSL 0.9.6d.
13928 As the countermeasure turned out to be incompatible with some
13929 broken SSL implementations, the new option is part of SSL_OP_ALL.
13930 SSL_OP_ALL is usually employed when compatibility with weird SSL
13931 implementations is desired (e.g. '-bugs' option to 's_client' and
13932 's_server'), so the new option is automatically set in many
13937 * Changes in security patch:
13939 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13940 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13941 Air Force Materiel Command, USAF, under agreement number
13944 * Add various sanity checks to asn1_get_length() to reject
13945 the ASN1 length bytes if they exceed sizeof(long), will appear
13946 negative or the content length exceeds the length of the
13947 supplied buffer. ([CVE-2002-0659])
13949 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13951 * Assertions for various potential buffer overflows, not known to
13952 happen in practice.
13954 *Ben Laurie (CHATS)*
13956 * Various temporary buffers to hold ASCII versions of integers were
13957 too small for 64 bit platforms. ([CVE-2002-0655])
13958 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13960 * Remote buffer overflow in SSL3 protocol - an attacker could
13961 supply an oversized session ID to a client. ([CVE-2002-0656])
13963 *Ben Laurie (CHATS)*
13965 * Remote buffer overflow in SSL2 protocol - an attacker could
13966 supply an oversized client master key. ([CVE-2002-0656])
13968 *Ben Laurie (CHATS)*
13970 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13972 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13973 encoded as NULL) with id-dsa-with-sha1.
13975 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13977 * Check various `X509_...()` return values in `apps/req.c`.
13979 *Nils Larsch <nla@trustcenter.de>*
13981 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13982 an end-of-file condition would erroneously be flagged, when the CRLF
13983 was just at the end of a processed block. The bug was discovered when
13984 processing data through a buffering memory BIO handing the data to a
13985 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13986 <ptsekov@syntrex.com> and Nedelcho Stanev.
13990 * Implement a countermeasure against a vulnerability recently found
13991 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13992 before application data chunks to avoid the use of known IVs
13993 with data potentially chosen by the attacker.
13997 * Fix length checks in ssl3_get_client_hello().
14001 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14002 to prevent ssl3_read_internal() from incorrectly assuming that
14003 ssl3_read_bytes() found application data while handshake
14004 processing was enabled when in fact s->s3->in_read_app_data was
14005 merely automatically cleared during the initial handshake.
14007 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14009 * Fix object definitions for Private and Enterprise: they were not
14010 recognized in their shortname (=lowercase) representation. Extend
14011 obj_dat.pl to issue an error when using undefined keywords instead
14012 of silently ignoring the problem (Svenning Sorensen
14013 <sss@sss.dnsalias.net>).
14017 * Fix DH_generate_parameters() so that it works for 'non-standard'
14018 generators, i.e. generators other than 2 and 5. (Previously, the
14019 code did not properly initialise the 'add' and 'rem' values to
14020 BN_generate_prime().)
14022 In the new general case, we do not insist that 'generator' is
14023 actually a primitive root: This requirement is rather pointless;
14024 a generator of the order-q subgroup is just as good, if not
14029 * Map new X509 verification errors to alerts. Discovered and submitted by
14030 Tom Wu <tom@arcot.com>.
14034 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14035 returning non-zero before the data has been completely received
14036 when using non-blocking I/O.
14038 *Bodo Moeller; problem pointed out by John Hughes*
14040 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14042 *Ben Laurie, Lutz Jaenicke*
14044 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14045 Yoram Zahavi <YoramZ@gilian.com>).
14049 * Add information about CygWin 1.3 and on, and preserve proper
14050 configuration for the versions before that.
14052 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14054 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14055 check whether we deal with a copy of a session and do not delete from
14056 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14057 <izhar@checkpoint.com>.
14061 * Do not store session data into the internal session cache, if it
14062 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14063 flag is set). Proposed by Aslam <aslam@funk.com>.
14067 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14072 * [In 0.9.6d-engine release:]
14073 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14075 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14077 * Add the configuration target linux-s390x.
14079 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14081 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14082 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14083 variable as an indication that a ClientHello message has been
14084 received. As the flag value will be lost between multiple
14085 invocations of ssl3_accept when using non-blocking I/O, the
14086 function may not be aware that a handshake has actually taken
14087 place, thus preventing a new session from being added to the
14090 To avoid this problem, we now set s->new_session to 2 instead of
14091 using a local variable.
14093 *Lutz Jaenicke, Bodo Moeller*
14095 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14096 if the SSL_R_LENGTH_MISMATCH error is detected.
14098 *Geoff Thorpe, Bodo Moeller*
14100 * New 'shared_ldflag' column in Configure platform table.
14104 * Fix EVP_CIPHER_mode macro.
14106 *"Dan S. Camper" <dan@bti.net>*
14108 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14109 type, we must throw them away by setting rr->length to 0.
14111 *D P Chang <dpc@qualys.com>*
14113 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14115 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14116 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14117 worked incorrectly for those cases where range = `10..._2` and
14118 `3*range` is two bits longer than range.)
14122 * Only add signing time to PKCS7 structures if it is not already
14127 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14128 OBJ_ld_ce should be OBJ_id_ce.
14129 Also some ip-pda OIDs in crypto/objects/objects.txt were
14130 incorrect (cf. RFC 3039).
14132 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14134 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14135 returns early because it has nothing to do.
14137 *Andy Schneider <andy.schneider@bjss.co.uk>*
14139 * [In 0.9.6c-engine release:]
14140 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14142 *Andy Schneider <andy.schneider@bjss.co.uk>*
14144 * [In 0.9.6c-engine release:]
14145 Add support for Cryptographic Appliance's keyserver technology.
14146 (Use engine 'keyclient')
14148 *Cryptographic Appliances and Geoff Thorpe*
14150 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14151 is called via tools/c89.sh because arguments have to be
14152 rearranged (all '-L' options must appear before the first object
14155 *Richard Shapiro <rshapiro@abinitio.com>*
14157 * [In 0.9.6c-engine release:]
14158 Add support for Broadcom crypto accelerator cards, backported
14161 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14163 * [In 0.9.6c-engine release:]
14164 Add support for SureWare crypto accelerator cards from
14165 Baltimore Technologies. (Use engine 'sureware')
14167 *Baltimore Technologies and Mark Cox*
14169 * [In 0.9.6c-engine release:]
14170 Add support for crypto accelerator cards from Accelerated
14171 Encryption Processing, www.aep.ie. (Use engine 'aep')
14173 *AEP Inc. and Mark Cox*
14175 * Add a configuration entry for gcc on UnixWare.
14177 *Gary Benson <gbenson@redhat.com>*
14179 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14180 messages are stored in a single piece (fixed-length part and
14181 variable-length part combined) and fix various bugs found on the way.
14185 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14186 instead. BIO_gethostbyname() does not know what timeouts are
14187 appropriate, so entries would stay in cache even when they have
14189 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14191 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14192 faced with a pathologically small ClientHello fragment that does
14193 not contain client_version: Instead of aborting with an error,
14194 simply choose the highest available protocol version (i.e.,
14195 TLS 1.0 unless it is disabled). In practice, ClientHello
14196 messages are never sent like this, but this change gives us
14197 strictly correct behaviour at least for TLS.
14201 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14202 never resets s->method to s->ctx->method when called from within
14203 one of the SSL handshake functions.
14205 *Bodo Moeller; problem pointed out by Niko Baric*
14207 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14208 (sent using the client's version number) if client_version is
14209 smaller than the protocol version in use. Also change
14210 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14211 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14212 the client will at least see that alert.
14216 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14221 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14222 client receives HelloRequest while in a handshake.
14224 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14226 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14227 should end in 'break', not 'goto end' which circumvents various
14228 cleanups done in state SSL_ST_OK. But session related stuff
14229 must be disabled for SSL_ST_OK in the case that we just sent a
14232 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14233 before just sending a HelloRequest.
14235 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14237 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14238 reveal whether illegal block cipher padding was found or a MAC
14239 verification error occurred. (Neither SSLerr() codes nor alerts
14240 are directly visible to potential attackers, but the information
14241 may leak via logfiles.)
14243 Similar changes are not required for the SSL 2.0 implementation
14244 because the number of padding bytes is sent in clear for SSL 2.0,
14245 and the extra bytes are just ignored. However ssl/s2_pkt.c
14246 failed to verify that the purported number of padding bytes is in
14251 * Add OpenUNIX-8 support including shared libraries
14252 (Boyd Lynn Gerber <gerberb@zenez.com>).
14256 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14257 'wristwatch attack' using huge encoding parameters (cf.
14258 James H. Manger's CRYPTO 2001 paper). Note that the
14259 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14260 encoding parameters and hence was not vulnerable.
14264 * BN_sqr() bug fix.
14266 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14268 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14269 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14270 followed by modular reduction.
14272 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14274 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14275 equivalent based on BN_pseudo_rand() instead of BN_rand().
14279 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14280 This function was broken, as the check for a new client hello message
14281 to handle SGC did not allow these large messages.
14282 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14286 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14290 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14291 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14295 * Rework the configuration and shared library support for Tru64 Unix.
14296 The configuration part makes use of modern compiler features and
14297 still retains old compiler behavior for those that run older versions
14298 of the OS. The shared library support part includes a variant that
14299 uses the RPATH feature, and is available through the special
14300 configuration target "alpha-cc-rpath", which will never be selected
14303 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14305 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14306 with the same message size as in ssl3_get_certificate_request().
14307 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14308 messages might inadvertently be reject as too long.
14310 *Petr Lampa <lampa@fee.vutbr.cz>*
14312 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14316 * Modified SSL library such that the verify_callback that has been set
14317 specifically for an SSL object with SSL_set_verify() is actually being
14318 used. Before the change, a verify_callback set with this function was
14319 ignored and the verify_callback() set in the SSL_CTX at the time of
14320 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14321 to allow the necessary settings.
14325 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14326 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14327 done automatically (in contradiction to the requirements of the C
14328 standard). This made problems when used from OpenSSH.
14332 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14333 dh->length and always used
14335 BN_rand_range(priv_key, dh->p).
14337 BN_rand_range() is not necessary for Diffie-Hellman, and this
14338 specific range makes Diffie-Hellman unnecessarily inefficient if
14339 dh->length (recommended exponent length) is much smaller than the
14340 length of dh->p. We could use BN_rand_range() if the order of
14341 the subgroup was stored in the DH structure, but we only have
14346 BN_rand(priv_key, l, ...)
14348 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14355 RSA_eay_public_encrypt
14356 RSA_eay_private_decrypt
14357 RSA_eay_private_encrypt (signing)
14358 RSA_eay_public_decrypt (signature verification)
14360 (default implementations for RSA_public_encrypt,
14361 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14362 always reject numbers >= n.
14366 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14367 to synchronize access to 'locking_thread'. This is necessary on
14368 systems where access to 'locking_thread' (an 'unsigned long'
14369 variable) is not atomic.
14373 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14374 *before* setting the 'crypto_lock_rand' flag. The previous code had
14375 a race condition if 0 is a valid thread ID.
14377 *Travis Vitek <vitek@roguewave.com>*
14379 * Add support for shared libraries under Irix.
14381 *Albert Chin-A-Young <china@thewrittenword.com>*
14383 * Add configuration option to build on Linux on both big-endian and
14384 little-endian MIPS.
14386 *Ralf Baechle <ralf@uni-koblenz.de>*
14388 * Add the possibility to create shared libraries on HP-UX.
14392 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14394 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14395 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14396 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14397 PRNG state recovery was possible based on the output of
14398 one PRNG request appropriately sized to gain knowledge on
14399 'md' followed by enough consecutive 1-byte PRNG requests
14400 to traverse all of 'state'.
14402 1. When updating 'md_local' (the current thread's copy of 'md')
14403 during PRNG output generation, hash all of the previous
14404 'md_local' value, not just the half used for PRNG output.
14406 2. Make the number of bytes from 'state' included into the hash
14407 independent from the number of PRNG bytes requested.
14409 The first measure alone would be sufficient to avoid
14410 Markku-Juhani's attack. (Actually it had never occurred
14411 to me that the half of 'md_local' used for chaining was the
14412 half from which PRNG output bytes were taken -- I had always
14413 assumed that the secret half would be used.) The second
14414 measure makes sure that additional data from 'state' is never
14415 mixed into 'md_local' in small portions; this heuristically
14416 further strengthens the PRNG.
14420 * Fix crypto/bn/asm/mips3.s.
14424 * When only the key is given to "enc", the IV is undefined. Print out
14425 an error message in this case.
14429 * Handle special case when X509_NAME is empty in X509 printing routines.
14433 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14434 positive and less than q.
14438 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14439 used: it isn't thread safe and the add_lock_callback should handle
14442 *Paul Rose <Paul.Rose@bridge.com>*
14444 * Verify that incoming data obeys the block size in
14445 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14451 *Ulf Möller, Bodo Möller*
14453 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14454 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14455 when fixing the server behaviour for backwards-compatible 'client
14456 hello' messages. (Note that the attack is impractical against
14457 SSL 3.0 and TLS 1.0 anyway because length and version checking
14458 means that the probability of guessing a valid ciphertext is
14459 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14462 Before 0.9.5, the countermeasure (hide the error by generating a
14463 random 'decryption result') did not work properly because
14464 ERR_clear_error() was missing, meaning that SSL_get_error() would
14465 detect the supposedly ignored error.
14467 Both problems are now fixed.
14471 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14472 (previously it was 1024).
14476 * Fix for compatibility mode trust settings: ignore trust settings
14477 unless some valid trust or reject settings are present.
14481 * Fix for blowfish EVP: its a variable length cipher.
14485 * Fix various bugs related to DSA S/MIME verification. Handle missing
14486 parameters in DSA public key structures and return an error in the
14487 DSA routines if parameters are absent.
14491 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14492 in the current directory if neither $RANDFILE nor $HOME was set.
14493 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14494 caused some confusion to Windows users who haven't defined $HOME.
14495 Thus RAND_file_name() is changed again: e_os.h can define a
14496 DEFAULT_HOME, which will be used if $HOME is not set.
14497 For Windows, we use "C:"; on other platforms, we still require
14498 environment variables.
14500 * Move 'if (!initialized) RAND_poll()' into regions protected by
14501 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14502 having multiple threads call RAND_poll() concurrently.
14506 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14507 combination of a flag and a thread ID variable.
14508 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14509 flag), *other* threads can enter ssleay_add_bytes without obeying
14510 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14511 that they do not hold after the first thread unsets add_do_not_lock).
14515 * Change bctest again: '-x' expressions are not available in all
14516 versions of 'test'.
14520 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14522 * Fix a couple of memory leaks in PKCS7_dataDecode()
14524 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14526 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14527 the default extension for executables, if any. Also, make the perl
14528 scripts that use symlink() to test if it really exists and use "cp"
14529 if it doesn't. All this made OpenSSL compilable and installable in
14534 * Fix for asn1_GetSequence() for indefinite length constructed data.
14535 If SEQUENCE is length is indefinite just set c->slen to the total
14536 amount of data available.
14538 *Steve Henson, reported by shige@FreeBSD.org*
14540 *This change does not apply to 0.9.7.*
14542 * Change bctest to avoid here-documents inside command substitution
14543 (workaround for FreeBSD /bin/sh bug).
14544 For compatibility with Ultrix, avoid shell functions (introduced
14545 in the bctest version that searches along $PATH).
14549 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14550 with des_encrypt() defined on some operating systems, like Solaris
14555 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14556 On the Importance of Eliminating Errors in Cryptographic
14557 Computations, J. Cryptology 14 (2001) 2, 101-119,
14558 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14562 * MIPS assembler BIGNUM division bug fix.
14566 * Disabled incorrect Alpha assembler code.
14570 * Fix PKCS#7 decode routines so they correctly update the length
14571 after reading an EOC for the EXPLICIT tag.
14575 *This change does not apply to 0.9.7.*
14577 * Fix bug in PKCS#12 key generation routines. This was triggered
14578 if a 3DES key was generated with a 0 initial byte. Include
14579 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14580 (but broken) behaviour.
14584 * Enhance bctest to search for a working bc along $PATH and print
14587 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14589 * Fix memory leaks in err.c: free err_data string if necessary;
14590 don't write to the wrong index in ERR_set_error_data.
14594 * Implement ssl23_peek (analogous to ssl23_read), which previously
14599 * Replace rdtsc with `_emit` statements for VC++ version 5.
14601 *Jeremy Cooper <jeremy@baymoo.org>*
14603 * Make it possible to reuse SSLv2 sessions.
14607 * In copy_email() check for >= 0 as a return value for
14608 X509_NAME_get_index_by_NID() since 0 is a valid index.
14610 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14612 * Avoid coredump with unsupported or invalid public keys by checking if
14613 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14614 PKCS7_verify() fails with non detached data.
14618 * Don't use getenv in library functions when run as setuid/setgid.
14619 New function OPENSSL_issetugid().
14623 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14624 due to incorrect handling of multi-threading:
14626 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14628 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14630 3. Count how many times MemCheck_off() has been called so that
14631 nested use can be treated correctly. This also avoids
14632 inband-signalling in the previous code (which relied on the
14633 assumption that thread ID 0 is impossible).
14637 * Add "-rand" option also to s_client and s_server.
14641 * Fix CPU detection on Irix 6.x.
14642 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14643 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14645 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14650 *This change does not apply to 0.9.7.*
14652 * Use the cached encoding of an X509_NAME structure rather than
14653 copying it. This is apparently the reason for the libsafe "errors"
14654 but the code is actually correct.
14658 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14659 Bleichenbacher's DSA attack.
14660 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14661 to be set and top=0 forces the highest bit to be set; top=-1 is new
14662 and leaves the highest bit random.
14664 *Ulf Moeller, Bodo Moeller*
14666 * In the `NCONF_...`-based implementations for `CONF_...` queries
14667 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14668 a temporary CONF structure with the data component set to NULL
14669 (which gives segmentation faults in lh_retrieve).
14670 Instead, use NULL for the CONF pointer in CONF_get_string and
14671 CONF_get_number (which may use environment variables) and directly
14672 return NULL from CONF_get_section.
14676 * Fix potential buffer overrun for EBCDIC.
14680 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14681 keyUsage if basicConstraints absent for a CA.
14685 * Make SMIME_write_PKCS7() write mail header values with a format that
14686 is more generally accepted (no spaces before the semicolon), since
14687 some programs can't parse those values properly otherwise. Also make
14688 sure BIO's that break lines after each write do not create invalid
14693 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14694 macros previously used would not encode an empty SEQUENCE OF
14695 and break the signature.
14699 *This change does not apply to 0.9.7.*
14701 * Zero the premaster secret after deriving the master secret in
14706 * Add some EVP_add_digest_alias registrations (as found in
14707 OpenSSL_add_all_digests()) to SSL_library_init()
14708 aka OpenSSL_add_ssl_algorithms(). This provides improved
14709 compatibility with peers using X.509 certificates
14710 with unconventional AlgorithmIdentifier OIDs.
14714 * Fix for Irix with NO_ASM.
14716 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14718 * ./config script fixes.
14720 *Ulf Moeller, Richard Levitte*
14722 * Fix 'openssl passwd -1'.
14726 * Change PKCS12_key_gen_asc() so it can cope with non null
14727 terminated strings whose length is passed in the passlen
14728 parameter, for example from PEM callbacks. This was done
14729 by adding an extra length parameter to asc2uni().
14731 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14733 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14734 call failed, free the DSA structure.
14738 * Fix to uni2asc() to cope with zero length Unicode strings.
14739 These are present in some PKCS#12 files.
14743 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14744 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14745 when writing a 32767 byte record.
14747 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14749 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14750 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14752 (RSA objects have a reference count access to which is protected
14753 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14754 so they are meant to be shared between threads.)
14755 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14756 "Reddie, Steven" <Steven.Reddie@ca.com>*
14758 * Fix a deadlock in CRYPTO_mem_leaks().
14762 * Use better test patterns in bntest.
14766 * rand_win.c fix for Borland C.
14770 * BN_rshift bugfix for n == 0.
14774 * Add a 'bctest' script that checks for some known 'bc' bugs
14775 so that 'make test' does not abort just because 'bc' is broken.
14779 * Store verify_result within SSL_SESSION also for client side to
14780 avoid potential security hole. (Re-used sessions on the client side
14781 always resulted in verify_result==X509_V_OK, not using the original
14782 result of the server certificate verification.)
14786 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14787 SSL3_RT_APPLICATION_DATA, return 0.
14788 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14793 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14794 releases, have been re-implemented by renaming the previous
14795 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14796 and ssl3_read_internal, respectively, and adding 'peek' parameters
14797 to them. The new ssl[23]_{read,peek} functions are calls to
14798 ssl[23]_read_internal with the 'peek' flag set appropriately.
14799 A 'peek' parameter has also been added to ssl3_read_bytes, which
14800 does the actual work for ssl3_read_internal.
14804 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14805 the method-specific "init()" handler. Also clean up ex_data after
14806 calling the method-specific "finish()" handler. Previously, this was
14807 happening the other way round.
14811 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14812 The previous value, 12, was not always sufficient for BN_mod_exp().
14816 * Make sure that shared libraries get the internal name engine with
14817 the full version number and not just 0. This should mark the
14818 shared libraries as not backward compatible. Of course, this should
14819 be changed again when we can guarantee backward binary compatibility.
14823 * Fix typo in get_cert_by_subject() in by_dir.c
14825 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14827 * Rework the system to generate shared libraries:
14829 - Make note of the expected extension for the shared libraries and
14830 if there is a need for symbolic links from for example libcrypto.so.0
14831 to libcrypto.so.0.9.7. There is extended info in Configure for
14834 - Make as few rebuilds of the shared libraries as possible.
14836 - Still avoid linking the OpenSSL programs with the shared libraries.
14838 - When installing, install the shared libraries separately from the
14843 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14845 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14846 and not in SSL_clear because the latter is also used by the
14847 accept/connect functions; previously, the settings made by
14848 SSL_set_read_ahead would be lost during the handshake.
14850 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14852 * Correct util/mkdef.pl to be selective about disabled algorithms.
14853 Previously, it would create entries for disabled algorithms no
14858 * Added several new manual pages for SSL_* function.
14862 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14864 * In ssl23_get_client_hello, generate an error message when faced
14865 with an initial SSL 3.0/TLS record that is too small to contain the
14866 first two bytes of the ClientHello message, i.e. client_version.
14867 (Note that this is a pathologic case that probably has never happened
14868 in real life.) The previous approach was to use the version number
14869 from the record header as a substitute; but our protocol choice
14870 should not depend on that one because it is not authenticated
14871 by the Finished messages.
14875 * More robust randomness gathering functions for Windows.
14877 *Jeffrey Altman <jaltman@columbia.edu>*
14879 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14880 not set then we don't setup the error code for issuer check errors
14881 to avoid possibly overwriting other errors which the callback does
14882 handle. If an application does set the flag then we assume it knows
14883 what it is doing and can handle the new informational codes
14888 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14889 a general "ANY" type, as such it should be able to decode anything
14890 including tagged types. However it didn't check the class so it would
14891 wrongly interpret tagged types in the same way as their universal
14892 counterpart and unknown types were just rejected. Changed so that the
14893 tagged and unknown types are handled in the same way as a SEQUENCE:
14894 that is the encoding is stored intact. There is also a new type
14895 "V_ASN1_OTHER" which is used when the class is not universal, in this
14896 case we have no idea what the actual type is so we just lump them all
14901 * On VMS, stdout may very well lead to a file that is written to
14902 in a record-oriented fashion. That means that every write() will
14903 write a separate record, which will be read separately by the
14904 programs trying to read from it. This can be very confusing.
14906 The solution is to put a BIO filter in the way that will buffer
14907 text until a linefeed is reached, and then write everything a
14908 line at a time, so every record written will be an actual line,
14909 not chunks of lines and not (usually doesn't happen, but I've
14910 seen it once) several lines in one record. BIO_f_linebuffer() is
14913 Currently, it's a VMS-only method, because that's where it has
14914 been tested well enough.
14918 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14919 it can return incorrect results.
14920 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14921 but it was in 0.9.6-beta[12].)
14925 * Disable the check for content being present when verifying detached
14926 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14927 include zero length content when signing messages.
14931 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14932 BIO_ctrl (for BIO pairs).
14936 * Add DSO method for VMS.
14940 * Bug fix: Montgomery multiplication could produce results with the
14945 * Add RPM specification openssl.spec and modify it to build three
14946 packages. The default package contains applications, application
14947 documentation and run-time libraries. The devel package contains
14948 include files, static libraries and function documentation. The
14949 doc package contains the contents of the doc directory. The original
14950 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14954 * Add a large number of documentation files for many SSL routines.
14956 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14958 * Add a configuration entry for Sony News 4.
14960 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14962 * Don't set the two most significant bits to one when generating a
14963 random number < q in the DSA library.
14967 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14968 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14969 the underlying transport is blocking) if a handshake took place.
14970 (The default behaviour is needed by applications such as s_client
14971 and s_server that use select() to determine when to use SSL_read;
14972 but for applications that know in advance when to expect data, it
14973 just makes things more complicated.)
14977 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14982 * Add a few more EBCDIC conditionals that make `req` and `x509`
14983 work better on such systems.
14985 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14987 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14988 Update PKCS12_parse() so it copies the friendlyName and the
14989 keyid to the certificates aux info.
14993 * Fix bug in PKCS7_verify() which caused an infinite loop
14994 if there was more than one signature.
14996 *Sven Uszpelkat <su@celocom.de>*
14998 * Major change in util/mkdef.pl to include extra information
14999 about each symbol, as well as presenting variables as well
15000 as functions. This change means that there's n more need
15001 to rebuild the .num files when some algorithms are excluded.
15005 * Allow the verify time to be set by an application,
15006 rather than always using the current time.
15010 * Phase 2 verify code reorganisation. The certificate
15011 verify code now looks up an issuer certificate by a
15012 number of criteria: subject name, authority key id
15013 and key usage. It also verifies self signed certificates
15014 by the same criteria. The main comparison function is
15015 X509_check_issued() which performs these checks.
15017 Lot of changes were necessary in order to support this
15018 without completely rewriting the lookup code.
15020 Authority and subject key identifier are now cached.
15022 The LHASH 'certs' is X509_STORE has now been replaced
15023 by a STACK_OF(X509_OBJECT). This is mainly because an
15024 LHASH can't store or retrieve multiple objects with
15025 the same hash value.
15027 As a result various functions (which were all internal
15028 use only) have changed to handle the new X509_STORE
15029 structure. This will break anything that messed round
15030 with X509_STORE internally.
15032 The functions X509_STORE_add_cert() now checks for an
15033 exact match, rather than just subject name.
15035 The X509_STORE API doesn't directly support the retrieval
15036 of multiple certificates matching a given criteria, however
15037 this can be worked round by performing a lookup first
15038 (which will fill the cache with candidate certificates)
15039 and then examining the cache for matches. This is probably
15040 the best we can do without throwing out X509_LOOKUP
15041 entirely (maybe later...).
15043 The X509_VERIFY_CTX structure has been enhanced considerably.
15045 All certificate lookup operations now go via a get_issuer()
15046 callback. Although this currently uses an X509_STORE it
15047 can be replaced by custom lookups. This is a simple way
15048 to bypass the X509_STORE hackery necessary to make this
15049 work and makes it possible to use more efficient techniques
15050 in future. A very simple version which uses a simple
15051 STACK for its trusted certificate store is also provided
15052 using X509_STORE_CTX_trusted_stack().
15054 The verify_cb() and verify() callbacks now have equivalents
15055 in the X509_STORE_CTX structure.
15057 X509_STORE_CTX also has a 'flags' field which can be used
15058 to customise the verify behaviour.
15062 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15063 excludes S/MIME capabilities.
15067 * When a certificate request is read in keep a copy of the
15068 original encoding of the signed data and use it when outputting
15069 again. Signatures then use the original encoding rather than
15070 a decoded, encoded version which may cause problems if the
15071 request is improperly encoded.
15075 * For consistency with other BIO_puts implementations, call
15076 buffer_write(b, ...) directly in buffer_puts instead of calling
15079 In BIO_puts, increment b->num_write as in BIO_write.
15081 *Peter.Sylvester@EdelWeb.fr*
15083 * Fix BN_mul_word for the case where the word is 0. (We have to use
15084 BN_zero, we may not return a BIGNUM with an array consisting of
15085 words set to zero.)
15089 * Avoid calling abort() from within the library when problems are
15090 detected, except if preprocessor symbols have been defined
15091 (such as REF_CHECK, BN_DEBUG etc.).
15095 * New openssl application 'rsautl'. This utility can be
15096 used for low level RSA operations. DER public key
15097 BIO/fp routines also added.
15101 * New Configure entry and patches for compiling on QNX 4.
15103 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15105 * A demo state-machine implementation was sponsored by
15106 Nuron (<http://www.nuron.com/>) and is now available in
15107 demos/state_machine.
15111 * New options added to the 'dgst' utility for signature
15112 generation and verification.
15116 * Unrecognized PKCS#7 content types are now handled via a
15117 catch all ASN1_TYPE structure. This allows unsupported
15118 types to be stored as a "blob" and an application can
15119 encode and decode it manually.
15123 * Fix various signed/unsigned issues to make a_strex.c
15124 compile under VC++.
15126 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15128 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15129 length if passed a buffer. ASN1_INTEGER_to_BN failed
15130 if passed a NULL BN and its argument was negative.
15132 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15134 * Modification to PKCS#7 encoding routines to output definite
15135 length encoding. Since currently the whole structures are in
15136 memory there's not real point in using indefinite length
15137 constructed encoding. However if OpenSSL is compiled with
15138 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15142 * Added BIO_vprintf() and BIO_vsnprintf().
15146 * Added more prefixes to parse for in the strings written
15147 through a logging bio, to cover all the levels that are available
15148 through syslog. The prefixes are now:
15150 PANIC, EMERG, EMR => LOG_EMERG
15151 ALERT, ALR => LOG_ALERT
15152 CRIT, CRI => LOG_CRIT
15153 ERROR, ERR => LOG_ERR
15154 WARNING, WARN, WAR => LOG_WARNING
15155 NOTICE, NOTE, NOT => LOG_NOTICE
15156 INFO, INF => LOG_INFO
15157 DEBUG, DBG => LOG_DEBUG
15159 and as before, if none of those prefixes are present at the
15160 beginning of the string, LOG_ERR is chosen.
15162 On Win32, the `LOG_*` levels are mapped according to this:
15164 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15165 LOG_WARNING => EVENTLOG_WARNING_TYPE
15166 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15170 * Made it possible to reconfigure with just the configuration
15171 argument "reconf" or "reconfigure". The command line arguments
15172 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15173 and are retrieved from there when reconfiguring.
15179 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15181 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15185 * The obj_dat.pl script was messing up the sorting of object
15186 names. The reason was that it compared the quoted version
15187 of strings as a result "OCSP" > "OCSP Signing" because
15188 " > SPACE. Changed script to store unquoted versions of
15189 names and add quotes on output. It was also omitting some
15190 names from the lookup table if they were given a default
15191 value (that is if SN is missing it is given the same
15192 value as LN and vice versa), these are now added on the
15193 grounds that if an object has a name we should be able to
15194 look it up. Finally added warning output when duplicate
15195 short or long names are found.
15199 * Changes needed for Tandem NSK.
15201 *Scott Uroff <scott@xypro.com>*
15203 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15204 RSA_padding_check_SSLv23(), special padding was never detected
15205 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15206 version rollback attacks was not effective.
15208 In s23_clnt.c, don't use special rollback-attack detection padding
15209 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15210 client; similarly, in s23_srvr.c, don't do the rollback check if
15211 SSL 2.0 is the only protocol enabled in the server.
15215 * Make it possible to get hexdumps of unprintable data with 'openssl
15216 asn1parse'. By implication, the functions ASN1_parse_dump() and
15217 BIO_dump_indent() are added.
15221 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15222 these print out strings and name structures based on various
15223 flags including RFC2253 support and proper handling of
15224 multibyte characters. Added options to the 'x509' utility
15225 to allow the various flags to be set.
15229 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15230 Also change the functions X509_cmp_current_time() and
15231 X509_gmtime_adj() work with an ASN1_TIME structure,
15232 this will enable certificates using GeneralizedTime in validity
15233 dates to be checked.
15237 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15238 negative public key encodings) on by default,
15239 NO_NEG_PUBKEY_BUG can be set to disable it.
15243 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15244 content octets. An i2c_ASN1_OBJECT is unnecessary because
15245 the encoding can be trivially obtained from the structure.
15249 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15250 not read locks (`CRYPTO_r_[un]lock`).
15254 * A first attempt at creating official support for shared
15255 libraries through configuration. I've kept it so the
15256 default is static libraries only, and the OpenSSL programs
15257 are always statically linked for now, but there are
15258 preparations for dynamic linking in place.
15259 This has been tested on Linux and Tru64.
15263 * Randomness polling function for Win9x, as described in:
15264 Peter Gutmann, Software Generation of Practically Strong
15269 * Fix so PRNG is seeded in req if using an already existing
15274 * New options to smime application. -inform and -outform
15275 allow alternative formats for the S/MIME message including
15276 PEM and DER. The -content option allows the content to be
15277 specified separately. This should allow things like Netscape
15278 form signing output easier to verify.
15282 * Fix the ASN1 encoding of tags using the 'long form'.
15286 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15287 STRING types. These convert content octets to and from the
15288 underlying type. The actual tag and length octets are
15289 already assumed to have been read in and checked. These
15290 are needed because all other string types have virtually
15291 identical handling apart from the tag. By having versions
15292 of the ASN1 functions that just operate on content octets
15293 IMPLICIT tagging can be handled properly. It also allows
15294 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15295 and ASN1_INTEGER are identical apart from the tag.
15299 * Change the handling of OID objects as follows:
15301 - New object identifiers are inserted in objects.txt, following
15302 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15303 - objects.pl is used to process obj_mac.num and create a new
15305 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15308 This is currently kind of a hack, and the perl code in objects.pl
15309 isn't very elegant, but it works as I intended. The simplest way
15310 to check that it worked correctly is to look in obj_dat.h and
15311 check the array nid_objs and make sure the objects haven't moved
15312 around (this is important!). Additions are OK, as well as
15313 consistent name changes.
15317 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15321 * Addition of the command line parameter '-rand file' to 'openssl req'.
15322 The given file adds to whatever has already been seeded into the
15323 random pool through the RANDFILE configuration file option or
15324 environment variable, or the default random state file.
15328 * mkstack.pl now sorts each macro group into lexical order.
15329 Previously the output order depended on the order the files
15330 appeared in the directory, resulting in needless rewriting
15335 * Patches to make OpenSSL compile under Win32 again. Mostly
15336 work arounds for the VC++ problem that it treats func() as
15337 func(void). Also stripped out the parts of mkdef.pl that
15338 added extra typesafe functions: these no longer exist.
15342 * Reorganisation of the stack code. The macros are now all
15343 collected in safestack.h . Each macro is defined in terms of
15344 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15345 DEBUG_SAFESTACK is now handled in terms of function casts,
15346 this has the advantage of retaining type safety without the
15347 use of additional functions. If DEBUG_SAFESTACK is not defined
15348 then the non typesafe macros are used instead. Also modified the
15349 mkstack.pl script to handle the new form. Needs testing to see
15350 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15351 the default if no major problems. Similar behaviour for ASN1_SET_OF
15352 and PKCS12_STACK_OF.
15356 * When some versions of IIS use the 'NET' form of private key the
15357 key derivation algorithm is different. Normally MD5(password) is
15358 used as a 128 bit RC4 key. In the modified case
15359 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15360 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15361 as the old Netscape_RSA functions except they have an additional
15362 'sgckey' parameter which uses the modified algorithm. Also added
15363 an -sgckey command line option to the rsa utility. Thanks to
15364 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15365 algorithm to openssl-dev.
15369 * The evp_local.h macros were using 'c.##kname' which resulted in
15370 invalid expansion on some systems (SCO 5.0.5 for example).
15371 Corrected to 'c.kname'.
15373 *Phillip Porch <root@theporch.com>*
15375 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15376 a STACK of email addresses from a certificate or request, these look
15377 in the subject name and the subject alternative name extensions and
15378 omit any duplicate addresses.
15382 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15383 This makes DSA verification about 2 % faster.
15387 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15388 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15389 plus overhead for 1024 bit moduli).
15390 This makes exponentiations about 0.5 % faster for 1024 bit
15391 exponents (as measured by "openssl speed rsa2048").
15395 * Rename memory handling macros to avoid conflicts with other
15397 Malloc => OPENSSL_malloc
15398 Malloc_locked => OPENSSL_malloc_locked
15399 Realloc => OPENSSL_realloc
15400 Free => OPENSSL_free
15404 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15405 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15409 * CygWin32 support.
15411 *John Jarvie <jjarvie@newsguy.com>*
15413 * The type-safe stack code has been rejigged. It is now only compiled
15414 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15415 by default all type-specific stack functions are "#define"d back to
15416 standard stack functions. This results in more streamlined output
15417 but retains the type-safety checking possibilities of the original
15422 * The STACK code has been cleaned up, and certain type declarations
15423 that didn't make a lot of sense have been brought in line. This has
15424 also involved a cleanup of sorts in safestack.h to more correctly
15425 map type-safe stack functions onto their plain stack counterparts.
15426 This work has also resulted in a variety of "const"ifications of
15427 lots of the code, especially `_cmp` operations which should normally
15428 be prototyped with "const" parameters anyway.
15432 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15433 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15434 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15435 where all of 'md' is used each time the PRNG is used, but 'state'
15436 is used only indexed by a cyclic counter. As entropy may not be
15437 well distributed from the beginning, 'md' is important as a
15438 chaining variable. However, the output function chains only half
15439 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15440 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15441 in all of 'state' being rewritten, with the new values depending
15442 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15446 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15447 the handshake is continued after ssl_verify_cert_chain();
15448 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15449 can lead to 'unexplainable' connection aborts later.
15451 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15453 * Major EVP API cipher revision.
15454 Add hooks for extra EVP features. This allows various cipher
15455 parameters to be set in the EVP interface. Support added for variable
15456 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15457 setting of RC2 and RC5 parameters.
15459 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15462 Remove lots of duplicated code from the EVP library. For example *every*
15463 cipher init() function handles the 'iv' in the same way according to the
15464 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15465 for CFB and OFB modes they zero ctx->num.
15467 New functionality allows removal of S/MIME code RC2 hack.
15469 Most of the routines have the same form and so can be declared in terms
15472 By shifting this to the top level EVP_CipherInit() it can be removed from
15473 all individual ciphers. If the cipher wants to handle IVs or keys
15474 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15477 Change lots of functions like EVP_EncryptUpdate() to now return a
15478 value: although software versions of the algorithms cannot fail
15479 any installed hardware versions can.
15483 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15484 this option is set, tolerate broken clients that send the negotiated
15485 protocol version number instead of the requested protocol version
15490 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15491 i.e. non-zero for export ciphersuites, zero otherwise.
15492 Previous versions had this flag inverted, inconsistent with
15493 rsa_tmp_cb (..._TMP_RSA_CB).
15495 *Bodo Moeller; problem reported by Amit Chopra*
15497 * Add missing DSA library text string. Work around for some IIS
15498 key files with invalid SEQUENCE encoding.
15502 * Add a document (doc/standards.txt) that list all kinds of standards
15503 and so on that are implemented in OpenSSL.
15507 * Enhance c_rehash script. Old version would mishandle certificates
15508 with the same subject name hash and wouldn't handle CRLs at all.
15509 Added -fingerprint option to crl utility, to support new c_rehash
15514 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15518 * Fix for SSL server purpose checking. Server checking was
15519 rejecting certificates which had extended key usage present
15520 but no ssl client purpose.
15522 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15524 * Make PKCS#12 code work with no password. The PKCS#12 spec
15525 is a little unclear about how a blank password is handled.
15526 Since the password in encoded as a BMPString with terminating
15527 double NULL a zero length password would end up as just the
15528 double NULL. However no password at all is different and is
15529 handled differently in the PKCS#12 key generation code. NS
15530 treats a blank password as zero length. MSIE treats it as no
15531 password on export: but it will try both on import. We now do
15532 the same: PKCS12_parse() tries zero length and no password if
15533 the password is set to "" or NULL (NULL is now a valid password:
15534 it wasn't before) as does the pkcs12 application.
15538 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15539 perror when PEM_read_bio_X509_REQ fails, the error message must
15540 be obtained from the error queue.
15544 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15545 it in ERR_remove_state if appropriate, and change ERR_get_state
15546 accordingly to avoid race conditions (this is necessary because
15547 thread_hash is no longer constant once set).
15551 * Bugfix for linux-elf makefile.one.
15555 * RSA_get_default_method() will now cause a default
15556 RSA_METHOD to be chosen if one doesn't exist already.
15557 Previously this was only set during a call to RSA_new()
15558 or RSA_new_method(NULL) meaning it was possible for
15559 RSA_get_default_method() to return NULL.
15563 * Added native name translation to the existing DSO code
15564 that will convert (if the flag to do so is set) filenames
15565 that are sufficiently small and have no path information
15566 into a canonical native form. Eg. "blah" converted to
15567 "libblah.so" or "blah.dll" etc.
15571 * New function ERR_error_string_n(e, buf, len) which is like
15572 ERR_error_string(e, buf), but writes at most 'len' bytes
15573 including the 0 terminator. For ERR_error_string_n, 'buf'
15576 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15578 * CONF library reworked to become more general. A new CONF
15579 configuration file reader "class" is implemented as well as a
15580 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15581 old `CONF_*` functions are still there, but are reimplemented to
15582 work in terms of the new functions. Also, a set of functions
15583 to handle the internal storage of the configuration data is
15584 provided to make it easier to write new configuration file
15585 reader "classes" (I can definitely see something reading a
15586 configuration file in XML format, for example), called `_CONF_*`,
15587 or "the configuration storage API"...
15589 The new configuration file reading functions are:
15591 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15592 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15594 NCONF_default, NCONF_WIN32
15596 NCONF_dump_fp, NCONF_dump_bio
15598 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15599 NCONF_new creates a new CONF object. This works in the same way
15600 as other interfaces in OpenSSL, like the BIO interface.
15601 `NCONF_dump_*` dump the internal storage of the configuration file,
15602 which is useful for debugging. All other functions take the same
15603 arguments as the old `CONF_*` functions with the exception of the
15604 first that must be a `CONF *` instead of a `LHASH *`.
15606 To make it easier to use the new classes with the old `CONF_*` functions,
15607 the function CONF_set_default_method is provided.
15611 * Add '-tls1' option to 'openssl ciphers', which was already
15612 mentioned in the documentation but had not been implemented.
15613 (This option is not yet really useful because even the additional
15614 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15618 * Initial DSO code added into libcrypto for letting OpenSSL (and
15619 OpenSSL-based applications) load shared libraries and bind to
15620 them in a portable way.
15622 *Geoff Thorpe, with contributions from Richard Levitte*
15624 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15626 * Make sure _lrotl and _lrotr are only used with MSVC.
15628 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15629 (the default implementation of RAND_status).
15631 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15632 to '-clrext' (= clear extensions), as intended and documented.
15633 *Bodo Moeller; inconsistency pointed out by Michael Attili
15634 <attili@amaxo.com>*
15636 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15637 was larger than the MD block size.
15639 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15641 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15642 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15643 using the passed key: if the passed key was a private key the result
15644 of X509_print(), for example, would be to print out all the private key
15649 * des_quad_cksum() byte order bug fix.
15650 *Ulf Möller, using the problem description in krb4-0.9.7, where
15651 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15653 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15656 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15658 * For easily testing in shell scripts whether some command
15659 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15660 returns with exit code 0 iff no command of the given name is available.
15661 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15662 the output goes to stdout and nothing is printed to stderr.
15663 Additional arguments are always ignored.
15665 Since for each cipher there is a command of the same name,
15666 the 'no-cipher' compilation switches can be tested this way.
15668 ('openssl no-XXX' is not able to detect pseudo-commands such
15669 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15673 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15677 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15678 is set; it will be thrown away anyway because each handshake creates
15680 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15681 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15682 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15683 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15687 * New s_client option -ign_eof: EOF at stdin is ignored, and
15688 'Q' and 'R' lose their special meanings (quit/renegotiate).
15689 This is part of what -quiet does; unlike -quiet, -ign_eof
15690 does not suppress any output.
15694 * Add compatibility options to the purpose and trust code. The
15695 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15696 accepts a certificate or CA, this was the previous behaviour,
15697 with all the associated security issues.
15699 X509_TRUST_COMPAT is the old trust behaviour: only and
15700 automatically trust self signed roots in certificate store. A
15701 new trust setting X509_TRUST_DEFAULT is used to specify that
15702 a purpose has no associated trust setting and it should instead
15703 use the value in the default purpose.
15707 * Fix the PKCS#8 DSA private key code so it decodes keys again
15708 and fix a memory leak.
15712 * In util/mkerr.pl (which implements 'make errors'), preserve
15713 reason strings from the previous version of the .c file, as
15714 the default to have only downcase letters (and digits) in
15715 automatically generated reasons codes is not always appropriate.
15719 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15720 using strerror. Previously, ERR_reason_error_string() returned
15721 library names as reason strings for SYSerr; but SYSerr is a special
15722 case where small numbers are errno values, not library numbers.
15726 * Add '-dsaparam' option to 'openssl dhparam' application. This
15727 converts DSA parameters into DH parameters. (When creating parameters,
15728 DSA_generate_parameters is used.)
15732 * Include 'length' (recommended exponent length) in C code generated
15733 by 'openssl dhparam -C'.
15737 * The second argument to set_label in perlasm was already being used
15738 so couldn't be used as a "file scope" flag. Moved to third argument
15743 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15744 instead of RAND_bytes for encryption IVs and salts.
15748 * Include RAND_status() into RAND_METHOD instead of implementing
15749 it only for md_rand.c Otherwise replacing the PRNG by calling
15750 RAND_set_rand_method would be impossible.
15754 * Don't let DSA_generate_key() enter an infinite loop if the random
15755 number generation fails.
15759 * New 'rand' application for creating pseudo-random output.
15763 * Added configuration support for Linux/IA64
15765 *Rolf Haberrecker <rolf@suse.de>*
15767 * Assembler module support for Mingw32.
15771 * Shared library support for HPUX (in shlib/).
15773 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15775 * Shared library support for Solaris gcc.
15777 *Lutz Behnke <behnke@trustcenter.de>*
15779 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15781 * PKCS7_encrypt() was adding text MIME headers twice because they
15782 were added manually and by SMIME_crlf_copy().
15786 * In bntest.c don't call BN_rand with zero bits argument.
15788 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15790 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15791 case was implemented. This caused BN_div_recp() to fail occasionally.
15795 * Add an optional second argument to the set_label() in the perl
15796 assembly language builder. If this argument exists and is set
15797 to 1 it signals that the assembler should use a symbol whose
15798 scope is the entire file, not just the current function. This
15799 is needed with MASM which uses the format label:: for this scope.
15801 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15803 * Change the ASN1 types so they are typedefs by default. Before
15804 almost all types were #define'd to ASN1_STRING which was causing
15805 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15810 * Change names of new functions to the new get1/get0 naming
15811 convention: After 'get1', the caller owns a reference count
15812 and has to call `..._free`; 'get0' returns a pointer to some
15813 data structure without incrementing reference counters.
15814 (Some of the existing 'get' functions increment a reference
15815 counter, some don't.)
15816 Similarly, 'set1' and 'add1' functions increase reference
15817 counters or duplicate objects.
15821 * Allow for the possibility of temp RSA key generation failure:
15822 the code used to assume it always worked and crashed on failure.
15826 * Fix potential buffer overrun problem in BIO_printf().
15827 *Ulf Möller, using public domain code by Patrick Powell; problem
15828 pointed out by David Sacerdote <das33@cornell.edu>*
15830 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15831 RAND_egd() and RAND_status(). In the command line application,
15832 the EGD socket can be specified like a seed file using RANDFILE
15837 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15838 Some CAs (e.g. Verisign) distribute certificates in this form.
15842 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15843 list to exclude them. This means that no special compilation option
15844 is needed to use anonymous DH: it just needs to be included in the
15849 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15850 EVP_MD_type. The old functionality is available in a new macro called
15851 EVP_MD_md(). Change code that uses it and update docs.
15855 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15856 where the `void *` argument is replaced by a function pointer argument.
15857 Previously `void *` was abused to point to functions, which works on
15858 many platforms, but is not correct. As these functions are usually
15859 called by macros defined in OpenSSL header files, most source code
15860 should work without changes.
15864 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15865 sections with information on -D... compiler switches used for
15866 compiling the library so that applications can see them. To enable
15867 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15868 must be defined. E.g.,
15869 #define OPENSSL_ALGORITHM_DEFINES
15870 #include <openssl/opensslconf.h>
15871 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15873 *Richard Levitte, Ulf and Bodo Möller*
15875 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15880 * Change the 'other' type in certificate aux info to a STACK_OF
15881 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15882 the required ASN1 format: arbitrary types determined by an OID.
15886 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15887 argument to 'req'. This is not because the function is newer or
15888 better than others it just uses the work 'NEW' in the certificate
15889 request header lines. Some software needs this.
15893 * Reorganise password command line arguments: now passwords can be
15894 obtained from various sources. Delete the PEM_cb function and make
15895 it the default behaviour: i.e. if the callback is NULL and the
15896 usrdata argument is not NULL interpret it as a null terminated pass
15897 phrase. If usrdata and the callback are NULL then the pass phrase
15898 is prompted for as usual.
15902 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15903 the support is automatically enabled. The resulting binaries will
15904 autodetect the card and use it if present.
15906 *Ben Laurie and Compaq Inc.*
15908 * Work around for Netscape hang bug. This sends certificate request
15909 and server done in one record. Since this is perfectly legal in the
15910 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15911 the bugs/SSLv3 entry for more info.
15915 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15919 * Add -rand argument to smime and pkcs12 applications and read/write
15924 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15928 * Add command line password options to the remaining applications.
15932 * Bug fix for BN_div_recp() for numerators with an even number of
15937 * More tests in bntest.c, and changed test_bn output.
15941 * ./config recognizes MacOS X now.
15945 * Bug fix for BN_div() when the first words of num and divisor are
15946 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15950 * Add support for various broken PKCS#8 formats, and command line
15951 options to produce them.
15955 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15956 get temporary BIGNUMs from a BN_CTX.
15960 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15965 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15966 include a #define from the old name to the new. The original intent
15967 was that statically linked binaries could for example just call
15968 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15969 link with digests. This never worked because SSLeay_add_all_digests()
15970 and SSLeay_add_all_ciphers() were in the same source file so calling
15971 one would link with the other. They are now in separate source files.
15975 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15979 * Use a less unusual form of the Miller-Rabin primality test (it used
15980 a binary algorithm for exponentiation integrated into the Miller-Rabin
15981 loop, our standard modexp algorithms are faster).
15985 * Support for the EBCDIC character set completed.
15987 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15989 * Source code cleanups: use const where appropriate, eliminate casts,
15990 use `void *` instead of `char *` in lhash.
15994 * Bugfix: ssl3_send_server_key_exchange was not restartable
15995 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15996 this the server could overwrite ephemeral keys that the client
16001 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16002 using 50 iterations of the Rabin-Miller test.
16004 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16005 iterations of the Rabin-Miller test as required by the appendix
16006 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16007 As BN_is_prime_fasttest includes trial division, DSA parameter
16008 generation becomes much faster.
16010 This implies a change for the callback functions in DSA_is_prime
16011 and DSA_generate_parameters: The callback function is called once
16012 for each positive witness in the Rabin-Miller test, not just
16013 occasionally in the inner loop; and the parameters to the
16014 callback function now provide an iteration count for the outer
16015 loop rather than for the current invocation of the inner loop.
16016 DSA_generate_parameters additionally can call the callback
16017 function with an 'iteration count' of -1, meaning that a
16018 candidate has passed the trial division test (when q is generated
16019 from an application-provided seed, trial division is skipped).
16023 * New function BN_is_prime_fasttest that optionally does trial
16024 division before starting the Rabin-Miller test and has
16025 an additional BN_CTX * argument (whereas BN_is_prime always
16026 has to allocate at least one BN_CTX).
16027 'callback(1, -1, cb_arg)' is called when a number has passed the
16028 trial division stage.
16032 * Fix for bug in CRL encoding. The validity dates weren't being handled
16037 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16041 * New function BN_pseudo_rand().
16045 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16046 bignum version of BN_from_montgomery() with the working code from
16047 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16052 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16053 made it impossible to use the same SSL_SESSION data structure in
16054 SSL2 clients in multiple threads.
16058 * The return value of RAND_load_file() no longer counts bytes obtained
16059 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16060 to seed the PRNG (previously an explicit byte count was required).
16062 *Ulf Möller, Bodo Möller*
16064 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16065 used `char *` instead of `void *` and had casts all over the place.
16069 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16073 * Retain source code compatibility for BN_prime_checks macro:
16074 BN_is_prime(..., BN_prime_checks, ...) now uses
16075 BN_prime_checks_for_size to determine the appropriate number of
16076 Rabin-Miller iterations.
16080 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16081 DH_CHECK_P_NOT_SAFE_PRIME.
16082 (Check if this is true? OpenPGP calls them "strong".)
16086 * Merge the functionality of "dh" and "gendh" programs into a new program
16087 "dhparam". The old programs are retained for now but will handle DH keys
16088 (instead of parameters) in future.
16092 * Make the ciphers, s_server and s_client programs check the return values
16093 when a new cipher list is set.
16097 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16098 ciphers. Before when the 56bit ciphers were enabled the sorting was
16101 The syntax for the cipher sorting has been extended to support sorting by
16102 cipher-strength (using the strength_bits hard coded in the tables).
16103 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16105 Fix a bug in the cipher-command parser: when supplying a cipher command
16106 string with an "undefined" symbol (neither command nor alphanumeric
16107 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16108 an error is flagged.
16110 Due to the strength-sorting extension, the code of the
16111 ssl_create_cipher_list() function was completely rearranged. I hope that
16112 the readability was also increased :-)
16114 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16116 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16117 for the first serial number and places 2 in the serial number file. This
16118 avoids problems when the root CA is created with serial number zero and
16119 the first user certificate has the same issuer name and serial number
16124 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16125 the new code. Add documentation for this stuff.
16129 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16130 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16131 structures and behave in an analogous way to the X509v3 functions:
16132 they shouldn't be called directly but wrapper functions should be used
16135 So we also now have some wrapper functions that call the X509at functions
16136 when passed certificate requests. (TO DO: similar things can be done with
16137 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16138 things. Some of these need some d2i or i2d and print functionality
16139 because they handle more complex structures.)
16143 * Add missing #ifndefs that caused missing symbols when building libssl
16144 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16145 NO_RSA in `ssl/s2*.c`.
16147 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16149 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16150 has a return value which indicates the quality of the random data
16151 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16152 error queue. New function RAND_pseudo_bytes() generates output that is
16153 guaranteed to be unique but not unpredictable. RAND_add is like
16154 RAND_seed, but takes an extra argument for an entropy estimate
16155 (RAND_seed always assumes full entropy).
16159 * Do more iterations of Rabin-Miller probable prime test (specifically,
16160 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16161 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16162 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16163 false-positive rate of at most 2^-80 for random input.
16167 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16171 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16172 in the 0.9.5 release), this returns the chain
16173 from an X509_CTX structure with a dup of the stack and all
16174 the X509 reference counts upped: so the stack will exist
16175 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16178 Also make SSL_SESSION_print() print out the verify return
16183 * Add manpage for the pkcs12 command. Also change the default
16184 behaviour so MAC iteration counts are used unless the new
16185 -nomaciter option is used. This improves file security and
16186 only older versions of MSIE (4.0 for example) need it.
16190 * Honor the no-xxx Configure options when creating .DEF files.
16194 * Add PKCS#10 attributes to field table: challengePassword,
16195 unstructuredName and unstructuredAddress. These are taken from
16196 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16197 international characters are used.
16199 More changes to X509_ATTRIBUTE code: allow the setting of types
16200 based on strings. Remove the 'loc' parameter when adding
16201 attributes because these will be a SET OF encoding which is sorted
16206 * Initial changes to the 'req' utility to allow request generation
16207 automation. This will allow an application to just generate a template
16208 file containing all the field values and have req construct the
16211 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16212 used all over the place including certificate requests and PKCS#7
16213 structures. They are currently handled manually where necessary with
16214 some primitive wrappers for PKCS#7. The new functions behave in a
16215 manner analogous to the X509 extension functions: they allow
16216 attributes to be looked up by NID and added.
16218 Later something similar to the X509V3 code would be desirable to
16219 automatically handle the encoding, decoding and printing of the
16220 more complex types. The string types like challengePassword can
16221 be handled by the string table functions.
16223 Also modified the multi byte string table handling. Now there is
16224 a 'global mask' which masks out certain types. The table itself
16225 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16226 is useful when for example there is only one permissible type
16227 (as in countryName) and using the mask might result in no valid
16232 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16233 SSL_get_peer_finished to allow applications to obtain the latest
16234 Finished messages sent to the peer or expected from the peer,
16235 respectively. (SSL_get_peer_finished is usually the Finished message
16236 actually received from the peer, otherwise the protocol will be aborted.)
16238 As the Finished message are message digests of the complete handshake
16239 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16240 be used for external authentication procedures when the authentication
16241 provided by SSL/TLS is not desired or is not enough.
16245 * Enhanced support for Alpha Linux is added. Now ./config checks if
16246 the host supports BWX extension and if Compaq C is present on the
16247 $PATH. Just exploiting of the BWX extension results in 20-30%
16248 performance kick for some algorithms, e.g. DES and RC4 to mention
16249 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16254 * Add support for MS "fast SGC". This is arguably a violation of the
16255 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16256 weak crypto and after checking the certificate is SGC a second one
16257 with strong crypto. MS SGC stops the first handshake after receiving
16258 the server certificate message and sends a second client hello. Since
16259 a server will typically do all the time consuming operations before
16260 expecting any further messages from the client (server key exchange
16261 is the most expensive) there is little difference between the two.
16263 To get OpenSSL to support MS SGC we have to permit a second client
16264 hello message after we have sent server done. In addition we have to
16265 reset the MAC if we do get this second client hello.
16269 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16270 if a DER encoded private key is RSA or DSA traditional format. Changed
16271 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16272 format DER encoded private key. Newer code should use PKCS#8 format which
16273 has the key type encoded in the ASN1 structure. Added DER private key
16274 support to pkcs8 application.
16278 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16279 ciphersuites has been selected (as required by the SSL 3/TLS 1
16280 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16281 is set, we interpret this as a request to violate the specification
16282 (the worst that can happen is a handshake failure, and 'correct'
16283 behaviour would result in a handshake failure anyway).
16287 * In SSL_CTX_add_session, take into account that there might be multiple
16288 SSL_SESSION structures with the same session ID (e.g. when two threads
16289 concurrently obtain them from an external cache).
16290 The internal cache can handle only one SSL_SESSION with a given ID,
16291 so if there's a conflict, we now throw out the old one to achieve
16296 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16297 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16298 some routines that use cipher OIDs: some ciphers do not have OIDs
16299 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16304 * Simplify the trust setting structure and code. Now we just have
16305 two sequences of OIDs for trusted and rejected settings. These will
16306 typically have values the same as the extended key usage extension
16307 and any application specific purposes.
16309 The trust checking code now has a default behaviour: it will just
16310 check for an object with the same NID as the passed id. Functions can
16311 be provided to override either the default behaviour or the behaviour
16312 for a given id. SSL client, server and email already have functions
16313 in place for compatibility: they check the NID and also return "trusted"
16314 if the certificate is self signed.
16318 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16319 traditional format into an EVP_PKEY structure.
16323 * Add a password callback function PEM_cb() which either prompts for
16324 a password if usr_data is NULL or otherwise assumes it is a null
16325 terminated password. Allow passwords to be passed on command line
16326 environment or config files in a few more utilities.
16330 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16331 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16332 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16333 Update documentation.
16337 * Support for ASN1 "NULL" type. This could be handled before by using
16338 ASN1_TYPE but there wasn't any function that would try to read a NULL
16339 and produce an error if it couldn't. For compatibility we also have
16340 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16341 don't allocate anything because they don't need to.
16345 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16348 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16350 * Rebuild of the memory allocation routines used by OpenSSL code and
16351 possibly others as well. The purpose is to make an interface that
16352 provide hooks so anyone can build a separate set of allocation and
16353 deallocation routines to be used by OpenSSL, for example memory
16354 pool implementations, or something else, which was previously hard
16355 since Malloc(), Realloc() and Free() were defined as macros having
16356 the values malloc, realloc and free, respectively (except for Win32
16357 compilations). The same is provided for memory debugging code.
16358 OpenSSL already comes with functionality to find memory leaks, but
16359 this gives people a chance to debug other memory problems.
16361 With these changes, a new set of functions and macros have appeared:
16363 CRYPTO_set_mem_debug_functions() [F]
16364 CRYPTO_get_mem_debug_functions() [F]
16365 CRYPTO_dbg_set_options() [F]
16366 CRYPTO_dbg_get_options() [F]
16367 CRYPTO_malloc_debug_init() [M]
16369 The memory debug functions are NULL by default, unless the library
16370 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16371 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16372 gives the standard debugging functions that come with OpenSSL) or
16373 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16374 provided by the library user) must be used. When the standard
16375 debugging functions are used, CRYPTO_dbg_set_options can be used to
16376 request additional information:
16377 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16378 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16380 Also, things like CRYPTO_set_mem_functions will always give the
16381 expected result (the new set of functions is used for allocation
16382 and deallocation) at all times, regardless of platform and compiler
16385 To finish it up, some functions that were never use in any other
16386 way than through macros have a new API and new semantic:
16388 CRYPTO_dbg_malloc()
16389 CRYPTO_dbg_realloc()
16392 All macros of value have retained their old syntax.
16394 *Richard Levitte and Bodo Moeller*
16396 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16397 ordering of SMIMECapabilities wasn't in "strength order" and there
16398 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16403 * Some ASN1 types with illegal zero length encoding (INTEGER,
16404 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16406 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16408 * Merge in my S/MIME library for OpenSSL. This provides a simple
16409 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16410 functionality to handle multipart/signed properly) and a utility
16411 called 'smime' to call all this stuff. This is based on code I
16412 originally wrote for Celo who have kindly allowed it to be
16413 included in OpenSSL.
16417 * Add variants des_set_key_checked and des_set_key_unchecked of
16418 des_set_key (aka des_key_sched). Global variable des_check_key
16419 decides which of these is called by des_set_key; this way
16420 des_check_key behaves as it always did, but applications and
16421 the library itself, which was buggy for des_check_key == 1,
16422 have a cleaner way to pick the version they need.
16426 * New function PKCS12_newpass() which changes the password of a
16431 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16432 dynamic mix. In both cases the ids can be used as an index into the
16433 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16434 functions so they accept a list of the field values and the
16435 application doesn't need to directly manipulate the X509_TRUST
16440 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16445 * Modify the way the V3 extension code looks up extensions. This now
16446 works in a similar way to the object code: we have some "standard"
16447 extensions in a static table which is searched with OBJ_bsearch()
16448 and the application can add dynamic ones if needed. The file
16449 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16450 updated whenever a new extension is added to the core code and kept
16451 in ext_nid order. There is a simple program 'tabtest.c' which checks
16452 this. New extensions are not added too often so this file can readily
16453 be maintained manually.
16455 There are two big advantages in doing things this way. The extensions
16456 can be looked up immediately and no longer need to be "added" using
16457 X509V3_add_standard_extensions(): this function now does nothing.
16458 Side note: I get *lots* of email saying the extension code doesn't
16459 work because people forget to call this function.
16460 Also no dynamic allocation is done unless new extensions are added:
16461 so if we don't add custom extensions there is no need to call
16462 X509V3_EXT_cleanup().
16466 * Modify enc utility's salting as follows: make salting the default. Add a
16467 magic header, so unsalted files fail gracefully instead of just decrypting
16468 to garbage. This is because not salting is a big security hole, so people
16469 should be discouraged from doing it.
16473 * Fixes and enhancements to the 'x509' utility. It allowed a message
16474 digest to be passed on the command line but it only used this
16475 parameter when signing a certificate. Modified so all relevant
16476 operations are affected by the digest parameter including the
16477 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16478 DSA key was used because it didn't fix the digest.
16482 * Initial certificate chain verify code. Currently tests the untrusted
16483 certificates for consistency with the verify purpose (which is set
16484 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16486 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16487 this is because it will reject chains with invalid extensions whereas
16488 every previous version of OpenSSL and SSLeay made no checks at all.
16490 Trust code: checks the root CA for the relevant trust settings. Trust
16491 settings have an initial value consistent with the verify purpose: e.g.
16492 if the verify purpose is for SSL client use it expects the CA to be
16493 trusted for SSL client use. However the default value can be changed to
16494 permit custom trust settings: one example of this would be to only trust
16495 certificates from a specific "secure" set of CAs.
16497 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16498 which should be used for version portability: especially since the
16499 verify structure is likely to change more often now.
16501 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16502 to set them. If not set then assume SSL clients will verify SSL servers
16505 Two new options to the verify program: -untrusted allows a set of
16506 untrusted certificates to be passed in and -purpose which sets the
16507 intended purpose of the certificate. If a purpose is set then the
16508 new chain verify code is used to check extension consistency.
16512 * Support for the authority information access extension.
16516 * Modify RSA and DSA PEM read routines to transparently handle
16517 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16518 public keys in a format compatible with certificate
16519 SubjectPublicKeyInfo structures. Unfortunately there were already
16520 functions called *_PublicKey_* which used various odd formats so
16521 these are retained for compatibility: however the DSA variants were
16522 never in a public release so they have been deleted. Changed dsa/rsa
16523 utilities to handle the new format: note no releases ever handled public
16524 keys so we should be OK.
16526 The primary motivation for this change is to avoid the same fiasco
16527 that dogs private keys: there are several incompatible private key
16528 formats some of which are standard and some OpenSSL specific and
16529 require various evil hacks to allow partial transparent handling and
16530 even then it doesn't work with DER formats. Given the option anything
16531 other than PKCS#8 should be dumped: but the other formats have to
16532 stay in the name of compatibility.
16534 With public keys and the benefit of hindsight one standard format
16535 is used which works with EVP_PKEY, RSA or DSA structures: though
16536 it clearly returns an error if you try to read the wrong kind of key.
16538 Added a -pubkey option to the 'x509' utility to output the public key.
16539 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16540 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16541 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16542 that do the same as the `EVP_PKEY_assign_*()` except they up the
16543 reference count of the added key (they don't "swallow" the
16548 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16549 CRLs would fail if the file contained no certificates or no CRLs:
16550 added a new function to read in both types and return the number
16551 read: this means that if none are read it will be an error. The
16552 DER versions of the certificate and CRL reader would always fail
16553 because it isn't possible to mix certificates and CRLs in DER format
16554 without choking one or the other routine. Changed this to just read
16555 a certificate: this is the best we can do. Also modified the code
16556 in `apps/verify.c` to take notice of return codes: it was previously
16557 attempting to read in certificates from NULL pointers and ignoring
16558 any errors: this is one reason why the cert and CRL reader seemed
16559 to work. It doesn't check return codes from the default certificate
16560 routines: these may well fail if the certificates aren't installed.
16564 * Code to support otherName option in GeneralName.
16568 * First update to verify code. Change the verify utility
16569 so it warns if it is passed a self signed certificate:
16570 for consistency with the normal behaviour. X509_verify
16571 has been modified to it will now verify a self signed
16572 certificate if *exactly* the same certificate appears
16573 in the store: it was previously impossible to trust a
16574 single self signed certificate. This means that:
16575 openssl verify ss.pem
16576 now gives a warning about a self signed certificate but
16577 openssl verify -CAfile ss.pem ss.pem
16582 * For servers, store verify_result in SSL_SESSION data structure
16583 (and add it to external session representation).
16584 This is needed when client certificate verifications fails,
16585 but an application-provided verification callback (set by
16586 SSL_CTX_set_cert_verify_callback) allows accepting the session
16587 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16588 but returns 1): When the session is reused, we have to set
16589 ssl->verify_result to the appropriate error code to avoid
16592 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16594 * Fix a bug in the new PKCS#7 code: it didn't consider the
16595 case in PKCS7_dataInit() where the signed PKCS7 structure
16596 didn't contain any existing data because it was being created.
16598 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16600 * Add a salt to the key derivation routines in enc.c. This
16601 forms the first 8 bytes of the encrypted file. Also add a
16602 -S option to allow a salt to be input on the command line.
16606 * New function X509_cmp(). Oddly enough there wasn't a function
16607 to compare two certificates. We do this by working out the SHA1
16608 hash and comparing that. X509_cmp() will be needed by the trust
16613 * SSL_get1_session() is like SSL_get_session(), but increments
16614 the reference count in the SSL_SESSION returned.
16616 *Geoff Thorpe <geoff@eu.c2.net>*
16618 * Fix for 'req': it was adding a null to request attributes.
16619 Also change the X509_LOOKUP and X509_INFO code to handle
16620 certificate auxiliary information.
16624 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16629 * Add the possibility to add extra information to the memory leak
16630 detecting output, to form tracebacks, showing from where each
16631 allocation was originated: CRYPTO_push_info("constant string") adds
16632 the string plus current file name and line number to a per-thread
16633 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16634 is like calling CYRPTO_pop_info() until the stack is empty.
16635 Also updated memory leak detection code to be multi-thread-safe.
16639 * Add options -text and -noout to pkcs7 utility and delete the
16640 encryption options which never did anything. Update docs.
16644 * Add options to some of the utilities to allow the pass phrase
16645 to be included on either the command line (not recommended on
16646 OSes like Unix) or read from the environment. Update the
16647 manpages and fix a few bugs.
16651 * Add a few manpages for some of the openssl commands.
16655 * Fix the -revoke option in ca. It was freeing up memory twice,
16656 leaking and not finding already revoked certificates.
16660 * Extensive changes to support certificate auxiliary information.
16661 This involves the use of X509_CERT_AUX structure and X509_AUX
16662 functions. An X509_AUX function such as PEM_read_X509_AUX()
16663 can still read in a certificate file in the usual way but it
16664 will also read in any additional "auxiliary information". By
16665 doing things this way a fair degree of compatibility can be
16666 retained: existing certificates can have this information added
16667 using the new 'x509' options.
16669 Current auxiliary information includes an "alias" and some trust
16670 settings. The trust settings will ultimately be used in enhanced
16671 certificate chain verification routines: currently a certificate
16672 can only be trusted if it is self signed and then it is trusted
16677 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16678 The problem was that one of the replacement routines had not been working
16679 since SSLeay releases. For now the offending routine has been replaced
16680 with non-optimised assembler. Even so, this now gives around 95%
16681 performance improvement for 1024 bit RSA signs.
16685 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16686 handling. Most clients have the effective key size in bits equal to
16687 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16688 A few however don't do this and instead use the size of the decrypted key
16689 to determine the RC2 key length and the AlgorithmIdentifier to determine
16690 the effective key length. In this case the effective key length can still
16691 be 40 bits but the key length can be 168 bits for example. This is fixed
16692 by manually forcing an RC2 key into the EVP_PKEY structure because the
16693 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16694 the key length and effective key length are equal.
16698 * Add a bunch of functions that should simplify the creation of
16699 X509_NAME structures. Now you should be able to do:
16700 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16701 and have it automatically work out the correct field type and fill in
16702 the structures. The more adventurous can try:
16703 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16704 and it will (hopefully) work out the correct multibyte encoding.
16708 * Change the 'req' utility to use the new field handling and multibyte
16709 copy routines. Before the DN field creation was handled in an ad hoc
16710 way in req, ca, and x509 which was rather broken and didn't support
16711 BMPStrings or UTF8Strings. Since some software doesn't implement
16712 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16713 using the dirstring_type option. See the new comment in the default
16714 openssl.cnf for more info.
16718 * Make crypto/rand/md_rand.c more robust:
16719 - Assure unique random numbers after fork().
16720 - Make sure that concurrent threads access the global counter and
16721 md serializably so that we never lose entropy in them
16722 or use exactly the same state in multiple threads.
16723 Access to the large state is not always serializable because
16724 the additional locking could be a performance killer, and
16725 md should be large enough anyway.
16729 * New file `apps/app_rand.c` with commonly needed functionality
16730 for handling the random seed file.
16732 Use the random seed file in some applications that previously did not:
16734 dsaparam -genkey (which also ignored its '-rand' option),
16737 x509 (when signing).
16738 Except on systems with /dev/urandom, it is crucial to have a random
16739 seed file at least for key creation, DSA signing, and for DH exchanges;
16740 for RSA signatures we could do without one.
16742 gendh and gendsa (unlike genrsa) used to read only the first byte
16743 of each file listed in the '-rand' option. The function as previously
16744 found in genrsa is now in app_rand.c and is used by all programs
16745 that support '-rand'.
16749 * In RAND_write_file, use mode 0600 for creating files;
16750 don't just chmod when it may be too late.
16754 * Report an error from X509_STORE_load_locations
16755 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16759 * New function ASN1_mbstring_copy() this copies a string in either
16760 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16761 into an ASN1_STRING type. A mask of permissible types is passed
16762 and it chooses the "minimal" type to use or an error if not type
16767 * Add function equivalents to the various macros in asn1.h. The old
16768 macros are retained with an `M_` prefix. Code inside the library can
16769 use the `M_` macros. External code (including the openssl utility)
16770 should *NOT* in order to be "shared library friendly".
16774 * Add various functions that can check a certificate's extensions
16775 to see if it usable for various purposes such as SSL client,
16776 server or S/MIME and CAs of these types. This is currently
16777 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16778 verification. Also added a -purpose flag to x509 utility to
16779 print out all the purposes.
16783 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16788 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16789 for, obtain and decode and extension and obtain its critical flag.
16790 This allows all the necessary extension code to be handled in a
16791 single function call.
16795 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16796 platforms. See crypto/rc4/rc4_enc.c for further details.
16800 * New -noout option to asn1parse. This causes no output to be produced
16801 its main use is when combined with -strparse and -out to extract data
16802 from a file (which may not be in ASN.1 format).
16806 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16807 when producing the local key id.
16809 *Richard Levitte <levitte@stacken.kth.se>*
16811 * New option -dhparam in s_server. This allows a DH parameter file to be
16812 stated explicitly. If it is not stated then it tries the first server
16813 certificate file. The previous behaviour hard coded the filename
16818 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16819 a public key to be input or output. For example:
16820 openssl rsa -in key.pem -pubout -out pubkey.pem
16821 Also added necessary DSA public key functions to handle this.
16825 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16826 in the message. This was handled by allowing
16827 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16829 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16831 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16832 to the end of the strings whereas this didn't. This would cause problems
16833 if strings read with d2i_ASN1_bytes() were later modified.
16835 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16837 * Fix for base64 decode bug. When a base64 bio reads only one line of
16838 data and it contains EOF it will end up returning an error. This is
16839 caused by input 46 bytes long. The cause is due to the way base64
16840 BIOs find the start of base64 encoded data. They do this by trying a
16841 trial decode on each line until they find one that works. When they
16842 do a flag is set and it starts again knowing it can pass all the
16843 data directly through the decoder. Unfortunately it doesn't reset
16844 the context it uses. This means that if EOF is reached an attempt
16845 is made to pass two EOFs through the context and this causes the
16846 resulting error. This can also cause other problems as well. As is
16847 usual with these problems it takes *ages* to find and the fix is
16848 trivial: move one line.
16850 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16852 * Ugly workaround to get s_client and s_server working under Windows. The
16853 old code wouldn't work because it needed to select() on sockets and the
16854 tty (for keypresses and to see if data could be written). Win32 only
16855 supports select() on sockets so we select() with a 1s timeout on the
16856 sockets and then see if any characters are waiting to be read, if none
16857 are present then we retry, we also assume we can always write data to
16858 the tty. This isn't nice because the code then blocks until we've
16859 received a complete line of data and it is effectively polling the
16860 keyboard at 1s intervals: however it's quite a bit better than not
16861 working at all :-) A dedicated Windows application might handle this
16862 with an event loop for example.
16866 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16867 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16868 will be called when RSA_sign() and RSA_verify() are used. This is useful
16869 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16870 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16871 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16872 This necessitated the support of an extra signature type NID_md5_sha1
16873 for SSL signatures and modifications to the SSL library to use it instead
16874 of calling RSA_public_decrypt() and RSA_private_encrypt().
16878 * Add new -verify -CAfile and -CApath options to the crl program, these
16879 will lookup a CRL issuers certificate and verify the signature in a
16880 similar way to the verify program. Tidy up the crl program so it
16881 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16882 less strict. It will now permit CRL extensions even if it is not
16883 a V2 CRL: this will allow it to tolerate some broken CRLs.
16887 * Initialize all non-automatic variables each time one of the openssl
16888 sub-programs is started (this is necessary as they may be started
16889 multiple times from the "OpenSSL>" prompt).
16891 *Lennart Bang, Bodo Moeller*
16893 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16894 removing all other RSA functionality (this is what NO_RSA does). This
16895 is so (for example) those in the US can disable those operations covered
16896 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16901 * Non-copying interface to BIO pairs.
16902 (still largely untested)
16906 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16907 ASCII string. This was handled independently in various places before.
16911 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16912 UTF8 strings a character at a time.
16916 * Use client_version from client hello to select the protocol
16917 (s23_srvr.c) and for RSA client key exchange verification
16918 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16922 * Add various utility functions to handle SPKACs, these were previously
16923 handled by poking round in the structure internals. Added new function
16924 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16925 print, verify and generate SPKACs. Based on an original idea from
16926 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16930 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16934 * Allow the config file extension section to be overwritten on the
16935 command line. Based on an original idea from Massimiliano Pala
16936 <madwolf@comune.modena.it>. The new option is called -extensions
16937 and can be applied to ca, req and x509. Also -reqexts to override
16938 the request extensions in req and -crlexts to override the crl extensions
16943 * Add new feature to the SPKAC handling in ca. Now you can include
16944 the same field multiple times by preceding it by "XXXX." for example:
16947 this is the same syntax as used in the req config file.
16951 * Allow certificate extensions to be added to certificate requests. These
16952 are specified in a 'req_extensions' option of the req section of the
16953 config file. They can be printed out with the -text option to req but
16954 are otherwise ignored at present.
16958 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16959 data read consists of only the final block it would not decrypted because
16960 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16961 A misplaced 'break' also meant the decrypted final block might not be
16962 copied until the next read.
16966 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16967 a few extra parameters to the DH structure: these will be useful if
16968 for example we want the value of 'q' or implement X9.42 DH.
16972 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16973 provides hooks that allow the default DSA functions or functions on a
16974 "per key" basis to be replaced. This allows hardware acceleration and
16975 hardware key storage to be handled without major modification to the
16976 library. Also added low level modexp hooks and CRYPTO_EX structure and
16977 associated functions.
16981 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16982 as "read only": it can't be written to and the buffer it points to will
16983 not be freed. Reading from a read only BIO is much more efficient than
16984 a normal memory BIO. This was added because there are several times when
16985 an area of memory needs to be read from a BIO. The previous method was
16986 to create a memory BIO and write the data to it, this results in two
16987 copies of the data and an O(n^2) reading algorithm. There is a new
16988 function BIO_new_mem_buf() which creates a read only memory BIO from
16989 an area of memory. Also modified the PKCS#7 routines to use read only
16994 * Bugfix: ssl23_get_client_hello did not work properly when called in
16995 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16996 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16997 but a retry condition occurred while trying to read the rest.
17001 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17002 NID_pkcs7_encrypted by default: this was wrong since this should almost
17003 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17004 the encrypted data type: this is a more sensible place to put it and it
17005 allows the PKCS#12 code to be tidied up that duplicated this
17010 * Changed obj_dat.pl script so it takes its input and output files on
17011 the command line. This should avoid shell escape redirection problems
17016 * Initial support for certificate extension requests, these are included
17017 in things like Xenroll certificate requests. Included functions to allow
17018 extensions to be obtained and added.
17022 * -crlf option to s_client and s_server for sending newlines as
17023 CRLF (as required by many protocols).
17027 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17029 * Install libRSAglue.a when OpenSSL is built with RSAref.
17031 *Ralf S. Engelschall*
17033 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17035 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17037 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17042 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17043 DH parameters/keys (q is lost during that conversion, but the resulting
17044 DH parameters contain its length).
17046 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17047 much faster than DH_generate_parameters (which creates parameters
17048 where `p = 2*q + 1`), and also the smaller q makes DH computations
17049 much more efficient (160-bit exponentiation instead of 1024-bit
17050 exponentiation); so this provides a convenient way to support DHE
17051 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17052 utter importance to use
17053 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17055 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17056 when such DH parameters are used, because otherwise small subgroup
17057 attacks may become possible!
17061 * Avoid memory leak in i2d_DHparams.
17065 * Allow the -k option to be used more than once in the enc program:
17066 this allows the same encrypted message to be read by multiple recipients.
17070 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17071 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17072 it will always use the numerical form of the OID, even if it has a short
17077 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17078 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17079 otherwise bn_mod_exp was called. In the case of hardware keys for example
17080 no private key components need be present and it might store extra data
17081 in the RSA structure, which cannot be accessed from bn_mod_exp.
17082 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17083 private key operations.
17087 * Added support for SPARC Linux.
17091 * pem_password_cb function type incompatibly changed from
17092 typedef int pem_password_cb(char *buf, int size, int rwflag);
17094 ....(char *buf, int size, int rwflag, void *userdata);
17095 so that applications can pass data to their callbacks:
17096 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17097 additional void * argument, which is just handed through whenever
17098 the password callback is called.
17100 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17102 New function SSL_CTX_set_default_passwd_cb_userdata.
17104 Compatibility note: As many C implementations push function arguments
17105 onto the stack in reverse order, the new library version is likely to
17106 interoperate with programs that have been compiled with the old
17107 pem_password_cb definition (PEM_whatever takes some data that
17108 happens to be on the stack as its last argument, and the callback
17109 just ignores this garbage); but there is no guarantee whatsoever that
17112 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17113 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17114 problems not only on Windows, but also on some Unix platforms.
17115 To avoid problematic command lines, these definitions are now in an
17116 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17117 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17121 * MIPS III/IV assembler module is reimplemented.
17125 * More DES library cleanups: remove references to srand/rand and
17126 delete an unused file.
17130 * Add support for the free Netwide assembler (NASM) under Win32,
17131 since not many people have MASM (ml) and it can be hard to obtain.
17132 This is currently experimental but it seems to work OK and pass all
17133 the tests. Check out INSTALL.W32 for info.
17137 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17138 without temporary keys kept an extra copy of the server key,
17139 and connections with temporary keys did not free everything in case
17144 * New function RSA_check_key and new openssl rsa option -check
17145 for verifying the consistency of RSA keys.
17147 *Ulf Moeller, Bodo Moeller*
17149 * Various changes to make Win32 compile work:
17150 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17151 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17152 comparison" warnings.
17153 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17157 * Add a debugging option to PKCS#5 v2 key generation function: when
17158 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17159 derived keys are printed to stderr.
17163 * Copy the flags in ASN1_STRING_dup().
17165 *Roman E. Pavlov <pre@mo.msk.ru>*
17167 * The x509 application mishandled signing requests containing DSA
17168 keys when the signing key was also DSA and the parameters didn't match.
17170 It was supposed to omit the parameters when they matched the signing key:
17171 the verifying software was then supposed to automatically use the CA's
17172 parameters if they were absent from the end user certificate.
17174 Omitting parameters is no longer recommended. The test was also
17175 the wrong way round! This was probably due to unusual behaviour in
17176 EVP_cmp_parameters() which returns 1 if the parameters match.
17177 This meant that parameters were omitted when they *didn't* match and
17178 the certificate was useless. Certificates signed with 'ca' didn't have
17181 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17183 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17184 The interface is as follows:
17185 Applications can use
17186 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17187 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17188 "off" is now the default.
17189 The library internally uses
17190 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17191 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17192 to disable memory-checking temporarily.
17194 Some inconsistent states that previously were possible (and were
17195 even the default) are now avoided.
17197 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17198 with each memory chunk allocated; this is occasionally more helpful
17199 than just having a counter.
17201 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17203 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17208 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17209 which largely parallels "options", but is for changing API behaviour,
17210 whereas "options" are about protocol behaviour.
17211 Initial "mode" flags are:
17213 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17214 a single record has been written.
17215 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17216 retries use the same buffer location.
17217 (But all of the contents must be
17222 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17225 * Fix problems with no-hmac etc.
17227 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17229 * New functions RSA_get_default_method(), RSA_set_method() and
17230 RSA_get_method(). These allows replacement of RSA_METHODs without having
17231 to mess around with the internals of an RSA structure.
17235 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17236 Also really enable memory leak checks in openssl.c and in some
17239 *Chad C. Mulligan, Bodo Moeller*
17241 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17242 up the length of negative integers. This has now been simplified to just
17243 store the length when it is first determined and use it later, rather
17244 than trying to keep track of where data is copied and updating it to
17246 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17248 * Add a new function PKCS7_signatureVerify. This allows the verification
17249 of a PKCS#7 signature but with the signing certificate passed to the
17250 function itself. This contrasts with PKCS7_dataVerify which assumes the
17251 certificate is present in the PKCS#7 structure. This isn't always the
17252 case: certificates can be omitted from a PKCS#7 structure and be
17253 distributed by "out of band" means (such as a certificate database).
17257 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17258 function prototypes in pem.h, also change util/mkdef.pl to add the
17259 necessary function names.
17263 * mk1mf.pl (used by Windows builds) did not properly read the
17264 options set by Configure in the top level Makefile, and Configure
17265 was not even able to write more than one option correctly.
17266 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17270 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17271 file to be loaded from a BIO or FILE pointer. The BIO version will
17272 for example allow memory BIOs to contain config info.
17276 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17277 Whoever hopes to achieve shared-library compatibility across versions
17278 must use this, not the compile-time macro.
17279 (Exercise 0.9.4: Which is the minimum library version required by
17281 Note: All this applies only to multi-threaded programs, others don't
17286 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17287 through a BIO pair triggered the default case, i.e.
17288 SSLerr(...,SSL_R_UNKNOWN_STATE).
17292 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17293 can use the SSL library even if none of the specific BIOs is
17298 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17299 for the encoded length.
17301 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17303 * Add initial documentation of the X509V3 functions.
17307 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17308 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17309 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17310 secure PKCS#8 private key format with a high iteration count.
17314 * Fix determination of Perl interpreter: A perl or perl5
17315 *directory* in $PATH was also accepted as the interpreter.
17317 *Ralf S. Engelschall*
17319 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17320 wrong with it but it was very old and did things like calling
17321 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17322 unusual formatting.
17326 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17327 to use the new extension code.
17331 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17332 with macros. This should make it easier to change their form, add extra
17333 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17338 * Add to configuration table a new entry that can specify an alternative
17339 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17340 according to Mark Crispin <MRC@Panda.COM>.
17344 * DES CBC did not update the IV. Weird.
17348 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17349 Changing the behaviour of the former might break existing programs --
17350 where IV updating is needed, des_ncbc_encrypt can be used.
17353 * When bntest is run from "make test" it drives bc to check its
17354 calculations, as well as internally checking them. If an internal check
17355 fails, it needs to cause bc to give a non-zero result or make test carries
17356 on without noticing the failure. Fixed.
17360 * DES library cleanups.
17364 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17365 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17366 ciphers. NOTE: although the key derivation function has been verified
17367 against some published test vectors it has not been extensively tested
17368 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17373 * Instead of "mkdir -p", which is not fully portable, use new
17374 Perl script "util/mkdir-p.pl".
17378 * Rewrite the way password based encryption (PBE) is handled. It used to
17379 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17380 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17381 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17382 the 'parameter' field of the AlgorithmIdentifier is passed to the
17383 underlying key generation function so it must do its own ASN1 parsing.
17384 This has also changed the EVP_PBE_CipherInit() function which now has a
17385 'parameter' argument instead of literal salt and iteration count values
17386 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17390 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17391 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17392 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17393 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17394 value was just used as a "magic string" and not used directly its
17395 value doesn't matter.
17399 * Introduce some semblance of const correctness to BN. Shame C doesn't
17404 * "linux-sparc64" configuration (ultrapenguin).
17406 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17407 "linux-sparc" configuration.
17409 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17411 * config now generates no-xxx options for missing ciphers.
17415 * Support the EBCDIC character set (work in progress).
17416 File ebcdic.c not yet included because it has a different license.
17418 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17420 * Support BS2000/OSD-POSIX.
17422 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17424 * Make callbacks for key generation use `void *` instead of `char *`.
17428 * Make S/MIME samples compile (not yet tested).
17432 * Additional typesafe stacks.
17436 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17440 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17442 * New configuration variant "sco5-gcc".
17444 * Updated some demos.
17446 *Sean O Riordain, Wade Scholine*
17448 * Add missing BIO_free at exit of pkcs12 application.
17452 * Fix memory leak in conf.c.
17456 * Updates for Win32 to assembler version of MD5.
17460 * Set #! path to perl in `apps/der_chop` to where we found it
17461 instead of using a fixed path.
17465 * SHA library changes for irix64-mips4-cc.
17469 * Improvements for VMS support.
17473 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17475 * Bignum library bug fix. IRIX 6 passes "make test" now!
17476 This also avoids the problems with SC4.2 and unpatched SC5.
17478 *Andy Polyakov <appro@fy.chalmers.se>*
17480 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17481 These are required because of the typesafe stack would otherwise break
17482 existing code. If old code used a structure member which used to be STACK
17483 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17484 sk_num or sk_value it would produce an error because the num, data members
17485 are not present in STACK_OF. Now it just produces a warning. sk_set
17486 replaces the old method of assigning a value to sk_value
17487 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17488 that does this will no longer work (and should use sk_set instead) but
17489 this could be regarded as a "questionable" behaviour anyway.
17493 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17494 correctly handle encrypted S/MIME data.
17498 * Change type of various DES function arguments from des_cblock
17499 (which means, in function argument declarations, pointer to char)
17500 to des_cblock * (meaning pointer to array with 8 char elements),
17501 which allows the compiler to do more typechecking; it was like
17502 that back in SSLeay, but with lots of ugly casts.
17504 Introduce new type const_des_cblock.
17508 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17509 problems: find RecipientInfo structure that matches recipient certificate
17510 and initialise the ASN1 structures properly based on passed cipher.
17514 * Belatedly make the BN tests actually check the results.
17518 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17519 to and from BNs: it was completely broken. New compilation option
17520 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17521 key elements as negative integers.
17525 * Reorganize and speed up MD5.
17527 *Andy Polyakov <appro@fy.chalmers.se>*
17531 *Richard Levitte <richard@levitte.org>*
17533 * New option -out to asn1parse to allow the parsed structure to be
17534 output to a file. This is most useful when combined with the -strparse
17535 option to examine the output of things like OCTET STRINGS.
17539 * Make SSL library a little more fool-proof by not requiring any longer
17540 that `SSL_set_{accept,connect}_state` be called before
17541 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17542 in many applications because usually everything *appeared* to work as
17543 intended anyway -- now it really works as intended).
17547 * Move openssl.cnf out of lib/.
17551 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17552 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17553 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17555 *Ralf S. Engelschall*
17557 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17558 handle PKCS#7 enveloped data properly.
17560 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17562 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17563 copying pointers. The cert_st handling is changed by this in
17564 various ways (and thus what used to be known as ctx->default_cert
17565 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17566 any longer when s->cert does not give us what we need).
17567 ssl_cert_instantiate becomes obsolete by this change.
17568 As soon as we've got the new code right (possibly it already is?),
17569 we have solved a couple of bugs of the earlier code where s->cert
17570 was used as if it could not have been shared with other SSL structures.
17572 Note that using the SSL API in certain dirty ways now will result
17573 in different behaviour than observed with earlier library versions:
17574 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17575 does not influence s as it used to.
17577 In order to clean up things more thoroughly, inside SSL_SESSION
17578 we don't use CERT any longer, but a new structure SESS_CERT
17579 that holds per-session data (if available); currently, this is
17580 the peer's certificate chain and, for clients, the server's certificate
17581 and temporary key. CERT holds only those values that can have
17582 meaningful defaults in an SSL_CTX.
17586 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17587 from the internal representation. Various PKCS#7 fixes: remove some
17588 evil casts and set the enc_dig_alg field properly based on the signing
17593 * Allow PKCS#12 password to be set from the command line or the
17594 environment. Let 'ca' get its config file name from the environment
17595 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17600 * Allow certificate policies extension to use an IA5STRING for the
17601 organization field. This is contrary to the PKIX definition but
17602 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17607 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17608 without disallowing inline assembler and the like for non-pedantic builds.
17612 * Support Borland C++ builder.
17614 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17620 * SHA-1 cleanups and performance enhancements.
17622 *Andy Polyakov <appro@fy.chalmers.se>*
17624 * Sparc v8plus assembler for the bignum library.
17626 *Andy Polyakov <appro@fy.chalmers.se>*
17628 * Accept any -xxx and +xxx compiler options in Configure.
17632 * Update HPUX configuration.
17636 * Add missing `sk_<type>_unshift()` function to safestack.h
17638 *Ralf S. Engelschall*
17640 * New function SSL_CTX_use_certificate_chain_file that sets the
17641 "extra_cert"s in addition to the certificate. (This makes sense
17642 only for "PEM" format files, as chains as a whole are not
17647 * Support verify_depth from the SSL API.
17648 x509_vfy.c had what can be considered an off-by-one-error:
17649 Its depth (which was not part of the external interface)
17650 was actually counting the number of certificates in a chain;
17651 now it really counts the depth.
17655 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17656 instead of X509err, which often resulted in confusing error
17657 messages since the error codes are not globally unique
17658 (e.g. an alleged error in ssl3_accept when a certificate
17659 didn't match the private key).
17661 * New function SSL_CTX_set_session_id_context that allows to set a default
17662 value (so that you don't need SSL_set_session_id_context for each
17663 connection using the SSL_CTX).
17667 * OAEP decoding bug fix.
17671 * Support INSTALL_PREFIX for package builders, as proposed by
17676 * New Configure options "threads" and "no-threads". For systems
17677 where the proper compiler options are known (currently Solaris
17678 and Linux), "threads" is the default.
17682 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17686 * Install various scripts to $(OPENSSLDIR)/misc, not to
17687 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17688 such as /usr/local/bin.
17692 * "make linux-shared" to build shared libraries.
17694 *Niels Poppe <niels@netbox.org>*
17696 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17700 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17701 extension adding in x509 utility.
17705 * Remove NOPROTO sections and error code comments.
17709 * Partial rewrite of the DEF file generator to now parse the ANSI
17714 * New Configure options --prefix=DIR and --openssldir=DIR.
17718 * Complete rewrite of the error code script(s). It is all now handled
17719 by one script at the top level which handles error code gathering,
17720 header rewriting and C source file generation. It should be much better
17721 than the old method: it now uses a modified version of Ulf's parser to
17722 read the ANSI prototypes in all header files (thus the old K&R definitions
17723 aren't needed for error creation any more) and do a better job of
17724 translating function codes into names. The old 'ASN1 error code embedded
17725 in a comment' is no longer necessary and it doesn't use .err files which
17726 have now been deleted. Also the error code call doesn't have to appear all
17727 on one line (which resulted in some large lines...).
17731 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17735 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17736 0 (which usually indicates a closed connection), but continue reading.
17740 * Fix some race conditions.
17744 * Add support for CRL distribution points extension. Add Certificate
17745 Policies and CRL distribution points documentation.
17749 * Move the autogenerated header file parts to crypto/opensslconf.h.
17753 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17754 8 of keying material. Merlin has also confirmed interop with this fix
17755 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17757 *Merlin Hughes <merlin@baltimore.ie>*
17759 * Fix lots of warnings.
17761 *Richard Levitte <levitte@stacken.kth.se>*
17763 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17764 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17766 *Richard Levitte <levitte@stacken.kth.se>*
17768 * Fix problems with sizeof(long) == 8.
17770 *Andy Polyakov <appro@fy.chalmers.se>*
17772 * Change functions to ANSI C.
17776 * Fix typos in error codes.
17778 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17780 * Remove defunct assembler files from Configure.
17784 * SPARC v8 assembler BIGNUM implementation.
17786 *Andy Polyakov <appro@fy.chalmers.se>*
17788 * Support for Certificate Policies extension: both print and set.
17789 Various additions to support the r2i method this uses.
17793 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17794 return a const string when you are expecting an allocated buffer.
17798 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17799 types DirectoryString and DisplayText.
17803 * Add code to allow r2i extensions to access the configuration database,
17804 add an LHASH database driver and add several ctx helper functions.
17808 * Fix an evil bug in bn_expand2() which caused various BN functions to
17809 fail when they extended the size of a BIGNUM.
17813 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17814 support typesafe stack.
17818 * Fix typo in SSL_[gs]et_options().
17820 *Nils Frostberg <nils@medcom.se>*
17822 * Delete various functions and files that belonged to the (now obsolete)
17823 old X509V3 handling code.
17827 * New Configure option "rsaref".
17831 * Don't auto-generate pem.h.
17835 * Introduce type-safe ASN.1 SETs.
17839 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17841 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17843 * Introduce type-safe STACKs. This will almost certainly break lots of code
17844 that links with OpenSSL (well at least cause lots of warnings), but fear
17845 not: the conversion is trivial, and it eliminates loads of evil casts. A
17846 few STACKed things have been converted already. Feel free to convert more.
17847 In the fullness of time, I'll do away with the STACK type altogether.
17851 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17852 specified in `<certfile>` by updating the entry in the index.txt file.
17853 This way one no longer has to edit the index.txt file manually for
17854 revoking a certificate. The -revoke option does the gory details now.
17856 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17858 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17859 `-text` option at all and this way the `-noout -text` combination was
17860 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17862 *Ralf S. Engelschall*
17864 * Make sure a corresponding plain text error message exists for the
17865 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17866 verify callback function determined that a certificate was revoked.
17868 *Ralf S. Engelschall*
17870 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17871 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17872 all available ciphers including rc5, which was forgotten until now.
17873 In order to let the testing shell script know which algorithms
17874 are available, a new (up to now undocumented) command
17875 `openssl list-cipher-commands` is used.
17879 * Bugfix: s_client occasionally would sleep in select() when
17880 it should have checked SSL_pending() first.
17884 * New functions DSA_do_sign and DSA_do_verify to provide access to
17885 the raw DSA values prior to ASN.1 encoding.
17889 * Tweaks to Configure
17891 *Niels Poppe <niels@netbox.org>*
17893 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17898 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17902 * New config option to avoid instructions that are illegal on the 80386.
17903 The default code is faster, but requires at least a 486.
17907 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17908 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17909 same as SSL2_VERSION anyway.
17913 * New "-showcerts" option for s_client.
17917 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17918 application. Various cleanups and fixes.
17922 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17923 modify error routines to work internally. Add error codes and PBE init
17924 to library startup routines.
17928 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17929 packing functions to asn1 and evp. Changed function names and error
17930 codes along the way.
17934 * PKCS12 integration: and so it begins... First of several patches to
17935 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17936 objects to objects.h
17940 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17941 and display support for Thawte strong extranet extension.
17945 * Add LinuxPPC support.
17947 *Jeff Dubrule <igor@pobox.org>*
17949 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17950 bn_div_words in alpha.s.
17952 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17954 * Make sure the RSA OAEP test is skipped under -DRSAref because
17955 OAEP isn't supported when OpenSSL is built with RSAref.
17957 *Ulf Moeller <ulf@fitug.de>*
17959 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17960 so they no longer are missing under -DNOPROTO.
17962 *Soren S. Jorvang <soren@t.dk>*
17964 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17966 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17967 doesn't work when the session is reused. Coming soon!
17971 * Fix a security hole, that allows sessions to be reused in the wrong
17972 context thus bypassing client cert protection! All software that uses
17973 client certs and session caches in multiple contexts NEEDS PATCHING to
17974 allow session reuse! A fuller solution is in the works.
17976 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17978 * Some more source tree cleanups (removed obsolete files
17979 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17980 permission on "config" script to be executable) and a fix for the INSTALL
17983 *Ulf Moeller <ulf@fitug.de>*
17985 * Remove some legacy and erroneous uses of malloc, free instead of
17988 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17990 * Make rsa_oaep_test return non-zero on error.
17992 *Ulf Moeller <ulf@fitug.de>*
17994 * Add support for native Solaris shared libraries. Configure
17995 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17996 if someone would make that last step automatic.
17998 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18000 * ctx_size was not built with the right compiler during "make links". Fixed.
18004 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18005 except NULL ciphers". This means the default cipher list will no longer
18006 enable NULL ciphers. They need to be specifically enabled e.g. with
18007 the string "DEFAULT:eNULL".
18011 * Fix to RSA private encryption routines: if p < q then it would
18012 occasionally produce an invalid result. This will only happen with
18013 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18017 * Be less restrictive and allow also `perl util/perlpath.pl
18018 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18019 because this way one can also use an interpreter named `perl5` (which is
18020 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18021 installed as `perl`).
18023 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18025 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18027 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18029 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18030 advapi32.lib to Win32 build and change the pem test comparison
18031 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18032 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18033 and crypto/des/ede_cbcm_enc.c.
18037 * DES quad checksum was broken on big-endian architectures. Fixed.
18041 * Comment out two functions in bio.h that aren't implemented. Fix up the
18042 Win32 test batch file so it (might) work again. The Win32 test batch file
18043 is horrible: I feel ill....
18047 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18048 in e_os.h. Audit of header files to check ANSI and non ANSI
18049 sections: 10 functions were absent from non ANSI section and not exported
18050 from Windows DLLs. Fixed up libeay.num for new functions.
18054 * Make `openssl version` output lines consistent.
18056 *Ralf S. Engelschall*
18058 * Fix Win32 symbol export lists for BIO functions: Added
18059 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18060 to ms/libeay{16,32}.def.
18062 *Ralf S. Engelschall*
18064 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18065 fine under Unix and passes some trivial tests I've now added. But the
18066 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18067 added to make sure no one expects that this stuff really works in the
18068 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18069 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18072 *Ralf S. Engelschall*
18074 * Fix the generation of two part addresses in perl.
18076 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18078 * Add config entry for Linux on MIPS.
18080 *John Tobey <jtobey@channel1.com>*
18082 * Make links whenever Configure is run, unless we are on Windoze.
18086 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18087 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18092 * Add a useful kludge to allow package maintainers to specify compiler and
18093 other platforms details on the command line without having to patch the
18094 Configure script every time: One now can use
18095 `perl Configure <id>:<details>`,
18096 i.e. platform ids are allowed to have details appended
18097 to them (separated by colons). This is treated as there would be a static
18098 pre-configured entry in Configure's %table under key `<id>` with value
18099 `<details>` and `perl Configure <id>` is called. So, when you want to
18100 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18101 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18102 now, which overrides the FreeBSD-elf entry on-the-fly.
18104 *Ralf S. Engelschall*
18106 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18110 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18111 on the `perl Configure ...` command line. This way one can compile
18112 OpenSSL libraries with Position Independent Code (PIC) which is needed
18113 for linking it into DSOs.
18115 *Ralf S. Engelschall*
18117 * Remarkably, export ciphers were totally broken and no-one had noticed!
18122 * Cleaned up the LICENSE document: The official contact for any license
18123 questions now is the OpenSSL core team under openssl-core@openssl.org.
18124 And add a paragraph about the dual-license situation to make sure people
18125 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18126 to the OpenSSL toolkit.
18128 *Ralf S. Engelschall*
18130 * General source tree makefile cleanups: Made `making xxx in yyy...`
18131 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18132 Additionally cleaned up the `make links` target: Remove unnecessary
18133 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18134 to speed processing and no longer clutter the display with confusing
18135 stuff. Instead only the actually done links are displayed.
18137 *Ralf S. Engelschall*
18139 * Permit null encryption ciphersuites, used for authentication only. It used
18140 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18141 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18146 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18147 signed attributes when verifying signatures (this would break them),
18148 the detached data encoding was wrong and public keys obtained using
18149 X509_get_pubkey() weren't freed.
18153 * Add text documentation for the BUFFER functions. Also added a work around
18154 to a Win95 console bug. This was triggered by the password read stuff: the
18155 last character typed gets carried over to the next fread(). If you were
18156 generating a new cert request using 'req' for example then the last
18157 character of the passphrase would be CR which would then enter the first
18162 * Added the new 'Includes OpenSSL Cryptography Software' button as
18163 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18164 button and can be used by applications based on OpenSSL to show the
18165 relationship to the OpenSSL project.
18167 *Ralf S. Engelschall*
18169 * Remove confusing variables in function signatures in files
18170 ssl/ssl_lib.c and ssl/ssl.h.
18172 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18174 * Don't install bss_file.c under PREFIX/include/
18176 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18178 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18179 functions that return function pointers and has support for NT specific
18180 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18181 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18182 unsigned to signed types: this was killing the Win32 compile.
18186 * Add new certificate file to stack functions,
18187 SSL_add_dir_cert_subjects_to_stack() and
18188 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18189 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18190 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18191 This means that Apache-SSL and similar packages don't have to mess around
18192 to add as many CAs as they want to the preferred list.
18196 * Experiment with doxygen documentation. Currently only partially applied to
18198 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18199 openssl.doxy as the configuration file.
18203 * Get rid of remaining C++-style comments which strict C compilers hate.
18205 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18207 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18208 compiled in by default: it has problems with large keys.
18212 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18213 DH private keys and/or callback functions which directly correspond to
18214 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18215 is needed for applications which have to configure certificates on a
18216 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18218 For the RSA certificate situation is makes no difference, but
18219 for the DSA certificate situation this fixes the "no shared cipher"
18220 problem where the OpenSSL cipher selection procedure failed because the
18221 temporary keys were not overtaken from the context and the API provided
18222 no way to reconfigure them.
18223 The new functions now let applications reconfigure the stuff and they
18224 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18225 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18226 non-public-API function ssl_cert_instantiate() is used as a helper
18227 function and also to reduce code redundancy inside ssl_rsa.c.
18229 *Ralf S. Engelschall*
18231 * Move s_server -dcert and -dkey options out of the undocumented feature
18232 area because they are useful for the DSA situation and should be
18233 recognized by the users.
18235 *Ralf S. Engelschall*
18237 * Fix the cipher decision scheme for export ciphers: the export bits are
18238 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18239 SSL_EXP_MASK. So, the original variable has to be used instead of the
18240 already masked variable.
18242 *Richard Levitte <levitte@stacken.kth.se>*
18244 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18246 *Richard Levitte <levitte@stacken.kth.se>*
18248 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18249 from `int` to `unsigned int` because it is a length and initialized by
18250 EVP_DigestFinal() which expects an `unsigned int *`.
18252 *Richard Levitte <levitte@stacken.kth.se>*
18254 * Don't hard-code path to Perl interpreter on shebang line of Configure
18255 script. Instead use the usual Shell->Perl transition trick.
18257 *Ralf S. Engelschall*
18259 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18260 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18261 -noout -modulus` as it's already the case for `openssl rsa -noout
18262 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18263 currently the public key is printed (a decision which was already done by
18264 `openssl dsa -modulus` in the past) which serves a similar purpose.
18265 Additionally the NO_RSA no longer completely removes the whole -modulus
18266 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18269 *Ralf S. Engelschall*
18271 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18272 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18274 *Arne Ansper <arne@ats.cyber.ee>*
18276 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18277 to be added. Now both 'req' and 'ca' can use new objects defined in the
18282 * Add cool BIO that does syslog (or event log on NT).
18284 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18286 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18287 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18288 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18289 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18293 * Add preliminary config info for new extension code.
18297 * Make RSA_NO_PADDING really use no padding.
18299 *Ulf Moeller <ulf@fitug.de>*
18301 * Generate errors when private/public key check is done.
18305 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18306 for some CRL extensions and new objects added.
18310 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18311 key usage extension and fuller support for authority key id.
18315 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18316 padding method for RSA, which is recommended for new applications in PKCS
18317 #1 v2.0 (RFC 2437, October 1998).
18318 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18319 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18320 against Bleichbacher's attack on RSA.
18321 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18324 * Updates to the new SSL compression code
18326 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18328 * Fix so that the version number in the master secret, when passed
18329 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18330 (because the server will not accept higher), that the version number
18331 is 0x03,0x01, not 0x03,0x00
18333 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18335 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18336 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18337 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18341 * Support for RAW extensions where an arbitrary extension can be
18342 created by including its DER encoding. See `apps/openssl.cnf` for
18347 * Make sure latest Perl versions don't interpret some generated C array
18348 code as Perl array code in the crypto/err/err_genc.pl script.
18350 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18352 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18353 not many people have the assembler. Various Win32 compilation fixes and
18354 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18355 build instructions.
18359 * Modify configure script 'Configure' to automatically create crypto/date.h
18360 file under Win32 and also build pem.h from pem.org. New script
18361 util/mkfiles.pl to create the MINFO file on environments that can't do a
18362 'make files': perl util/mkfiles.pl >MINFO should work.
18366 * Major rework of DES function declarations, in the pursuit of correctness
18367 and purity. As a result, many evil casts evaporated, and some weirdness,
18368 too. You may find this causes warnings in your code. Zapping your evil
18369 casts will probably fix them. Mostly.
18373 * Fix for a typo in asn1.h. Bug fix to object creation script
18374 obj_dat.pl. It considered a zero in an object definition to mean
18375 "end of object": none of the objects in objects.h have any zeros
18376 so it wasn't spotted.
18378 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18380 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18381 Masking (CBCM). In the absence of test vectors, the best I have been able
18382 to do is check that the decrypt undoes the encrypt, so far. Send me test
18383 vectors if you have them.
18387 * Correct calculation of key length for export ciphers (too much space was
18388 allocated for null ciphers). This has not been tested!
18392 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18393 message is now correct (it understands "crypto" and "ssl" on its
18394 command line). There is also now an "update" option. This will update
18395 the util/ssleay.num and util/libeay.num files with any new functions.
18397 perl util/mkdef.pl crypto ssl update
18398 it will update them.
18402 * Overhauled the Perl interface:
18403 - ported BN stuff to OpenSSL's different BN library
18404 - made the perl/ source tree CVS-aware
18405 - renamed the package from SSLeay to OpenSSL (the files still contain
18406 their history because I've copied them in the repository)
18407 - removed obsolete files (the test scripts will be replaced
18408 by better Test::Harness variants in the future)
18410 *Ralf S. Engelschall*
18412 * First cut for a very conservative source tree cleanup:
18413 1. merge various obsolete readme texts into doc/ssleay.txt
18414 where we collect the old documents and readme texts.
18415 2. remove the first part of files where I'm already sure that we no
18416 longer need them because of three reasons: either they are just temporary
18417 files which were left by Eric or they are preserved original files where
18418 I've verified that the diff is also available in the CVS via "cvs diff
18419 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18420 the crypto/md/ stuff).
18422 *Ralf S. Engelschall*
18424 * More extension code. Incomplete support for subject and issuer alt
18425 name, issuer and authority key id. Change the i2v function parameters
18426 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18427 what that's for :-) Fix to ASN1 macro which messed up
18428 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18432 * Preliminary support for ENUMERATED type. This is largely copied from the
18437 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18439 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18441 * Make sure `make rehash` target really finds the `openssl` program.
18443 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18445 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18446 like to hear about it if this slows down other processors.
18450 * Add CygWin32 platform information to Configure script.
18452 *Alan Batie <batie@aahz.jf.intel.com>*
18454 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18456 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18458 * New program nseq to manipulate netscape certificate sequences
18462 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18467 * Fixes to BN code. Previously the default was to define BN_RECURSION
18468 but the BN code had some problems that would cause failures when
18469 doing certificate verification and some other functions.
18471 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18473 * Add ASN1 and PEM code to support netscape certificate sequences.
18477 * Add ASN1 and PEM code to support netscape certificate sequences.
18481 * Add several PKIX and private extended key usage OIDs.
18485 * Modify the 'ca' program to handle the new extension code. Modify
18486 openssl.cnf for new extension format, add comments.
18490 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18491 and add a sample to openssl.cnf so req -x509 now adds appropriate
18496 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18497 error code, add initial support to X509_print() and x509 application.
18501 * Takes a deep breath and start adding X509 V3 extension support code. Add
18502 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18503 stuff is currently isolated and isn't even compiled yet.
18507 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18508 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18509 Removed the versions check from X509 routines when loading extensions:
18510 this allows certain broken certificates that don't set the version
18511 properly to be processed.
18515 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18516 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18517 can still be regenerated with "make depend".
18521 * Spelling mistake in C version of CAST-128.
18523 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18525 * Changes to the error generation code. The perl script err-code.pl
18526 now reads in the old error codes and retains the old numbers, only
18527 adding new ones if necessary. It also only changes the .err files if new
18528 codes are added. The makefiles have been modified to only insert errors
18529 when needed (to avoid needlessly modifying header files). This is done
18530 by only inserting errors if the .err file is newer than the auto generated
18531 C file. To rebuild all the error codes from scratch (the old behaviour)
18532 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18533 or delete all the .err files.
18537 * CAST-128 was incorrectly implemented for short keys. The C version has
18538 been fixed, but is untested. The assembler versions are also fixed, but
18539 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18540 to regenerate it if needed.
18541 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18542 Hagino <itojun@kame.net>*
18544 * File was opened incorrectly in randfile.c.
18546 *Ulf Möller <ulf@fitug.de>*
18548 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18549 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18550 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18551 al: it's just almost always a UTCTime. Note this patch adds new error
18552 codes so do a "make errors" if there are problems.
18556 * Correct Linux 1 recognition in config.
18558 *Ulf Möller <ulf@fitug.de>*
18560 * Remove pointless MD5 hash when using DSA keys in ca.
18562 *Anonymous <nobody@replay.com>*
18564 * Generate an error if given an empty string as a cert directory. Also
18565 generate an error if handed NULL (previously returned 0 to indicate an
18566 error, but didn't set one).
18568 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18570 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18574 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18575 parameters. This was causing a warning which killed off the Win32 compile.
18579 * Remove C++ style comments from crypto/bn/bn_local.h.
18581 *Neil Costigan <neil.costigan@celocom.com>*
18583 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18584 based on a text string, looking up short and long names and finally
18585 "dot" format. The "dot" format stuff didn't work. Added new function
18586 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18587 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18588 OID is not part of the table.
18592 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18593 X509_LOOKUP_by_alias().
18597 * Sort openssl functions by name.
18601 * Get the `gendsa` command working and add it to the `list` command. Remove
18602 encryption from sample DSA keys (in case anyone is interested the password
18607 * Make *all* `*_free` functions accept a NULL pointer.
18609 *Frans Heymans <fheymans@isaserver.be>*
18611 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18614 *Anonymous <nobody@replay.com>*
18616 * s_server should send the CAfile as acceptable CAs, not its own cert.
18618 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18620 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18622 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18624 * Temp key "for export" tests were wrong in s3_srvr.c.
18626 *Anonymous <nobody@replay.com>*
18628 * Add prototype for temp key callback functions
18629 SSL_CTX_set_tmp_{rsa,dh}_callback().
18633 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18634 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18638 * X509_name_add_entry() freed the wrong thing after an error.
18640 *Arne Ansper <arne@ats.cyber.ee>*
18642 * rsa_eay.c would attempt to free a NULL context.
18644 *Arne Ansper <arne@ats.cyber.ee>*
18646 * BIO_s_socket() had a broken should_retry() on Windoze.
18648 *Arne Ansper <arne@ats.cyber.ee>*
18650 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18652 *Arne Ansper <arne@ats.cyber.ee>*
18654 * Make sure the already existing X509_STORE->depth variable is initialized
18655 in X509_STORE_new(), but document the fact that this variable is still
18656 unused in the certificate verification process.
18658 *Ralf S. Engelschall*
18660 * Fix the various library and `apps/` files to free up pkeys obtained from
18661 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18665 * Fix reference counting in X509_PUBKEY_get(). This makes
18666 demos/maurice/example2.c work, amongst others, probably.
18668 *Steve Henson and Ben Laurie*
18670 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18671 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18672 are no longer created. This way we have a single and consistent command
18673 line interface `openssl <command>`, similar to `cvs <command>`.
18675 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18677 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18678 BIT STRING wrapper always have zero unused bits.
18682 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18686 * Make the top-level INSTALL documentation easier to understand.
18690 * Makefiles updated to exit if an error occurs in a sub-directory
18691 make (including if user presses ^C) [Paul Sutton]
18693 * Make Montgomery context stuff explicit in RSA data structure.
18697 * Fix build order of pem and err to allow for generated pem.h.
18701 * Fix renumbering bug in X509_NAME_delete_entry().
18705 * Enhanced the err-ins.pl script so it makes the error library number
18706 global and can add a library name. This is needed for external ASN1 and
18707 other error libraries.
18711 * Fixed sk_insert which never worked properly.
18715 * Fix ASN1 macros so they can handle indefinite length constructed
18716 EXPLICIT tags. Some non standard certificates use these: they can now
18721 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18722 into a single doc/ssleay.txt bundle. This way the information is still
18723 preserved but no longer messes up this directory. Now it's new room for
18724 the new set of documentation files.
18726 *Ralf S. Engelschall*
18728 * SETs were incorrectly DER encoded. This was a major pain, because they
18729 shared code with SEQUENCEs, which aren't coded the same. This means that
18730 almost everything to do with SETs or SEQUENCEs has either changed name or
18731 number of arguments.
18733 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18735 * Fix test data to work with the above.
18739 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18740 was already fixed by Eric for 0.9.1 it seems.
18742 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18744 * Autodetect FreeBSD3.
18748 * Fix various bugs in Configure. This affects the following platforms:
18752 unixware-2.0-pentium
18757 * Eliminate generated files from CVS. Reorder tests to regenerate files
18758 before they are needed.
18762 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18766 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18768 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18769 changed SSLeay to OpenSSL in version strings.
18771 *Ralf S. Engelschall*
18773 * Some fixups to the top-level documents.
18777 * Fixed the nasty bug where rsaref.h was not found under compile-time
18778 because the symlink to include/ was missing.
18780 *Ralf S. Engelschall*
18782 * Incorporated the popular no-RSA/DSA-only patches
18783 which allow to compile a RSA-free SSLeay.
18785 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18787 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18788 when "ssleay" is still not found.
18790 *Ralf S. Engelschall*
18792 * Added more platforms to Configure: Cray T3E, HPUX 11,
18794 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18796 * Updated the README file.
18798 *Ralf S. Engelschall*
18800 * Added various .cvsignore files in the CVS repository subdirs
18801 to make a "cvs update" really silent.
18803 *Ralf S. Engelschall*
18805 * Recompiled the error-definition header files and added
18806 missing symbols to the Win32 linker tables.
18808 *Ralf S. Engelschall*
18810 * Cleaned up the top-level documents;
18811 o new files: CHANGES and LICENSE
18812 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18813 o merged COPYRIGHT into LICENSE
18814 o removed obsolete TODO file
18815 o renamed MICROSOFT to INSTALL.W32
18817 *Ralf S. Engelschall*
18819 * Removed dummy files from the 0.9.1b source tree:
18821 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18822 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18823 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18824 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18825 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18828 *Ralf S. Engelschall*
18830 * Added various platform portability fixes.
18834 * The Genesis of the OpenSSL rpject:
18835 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18836 Young and Tim J. Hudson created while they were working for C2Net until
18839 *The OpenSSL Project*
18841 ### Changes between 0.9.0b and 0.9.1b [not released]
18843 * Updated a few CA certificates under certs/
18847 * Changed some BIGNUM api stuff.
18851 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18852 DGUX x86, Linux Alpha, etc.
18856 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18857 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18862 * Add -strparse option to asn1pars program which parses nested
18865 *Dr Stephen Henson <shenson@bigfoot.com>*
18867 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18871 * DSA fix for "ca" program.
18875 * Added "-genkey" option to "dsaparam" program.
18879 * Added RIPE MD160 (rmd160) message digest.
18883 * Added -a (all) option to "ssleay version" command.
18887 * Added PLATFORM define which is the id given to Configure.
18891 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18895 * Extended the ASN.1 parser routines.
18899 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18903 * Added a BN_CTX to the BN library.
18907 * Fixed the weak key values in DES library
18911 * Changed API in EVP library for cipher aliases.
18915 * Added support for RC2/64bit cipher.
18919 * Converted the lhash library to the crypto/mem.c functions.
18923 * Added more recognized ASN.1 object ids.
18927 * Added more RSA padding checks for SSL/TLS.
18931 * Added BIO proxy/filter functionality.
18935 * Added extra_certs to SSL_CTX which can be used
18936 send extra CA certificates to the client in the CA cert chain sending
18937 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18941 * Now Fortezza is denied in the authentication phase because
18942 this is key exchange mechanism is not supported by SSLeay at all.
18946 * Additional PKCS1 checks.
18950 * Support the string "TLSv1" for all TLS v1 ciphers.
18954 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18955 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18959 * Fixed a few memory leaks.
18963 * Fixed various code and comment typos.
18967 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18968 bytes sent in the client random.
18970 *Edward Bishop <ebishop@spyglass.com>*
18974 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18975 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18976 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18977 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18978 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18979 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18980 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18981 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18982 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18983 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18984 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18985 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18986 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18987 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18988 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18989 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18990 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18991 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18992 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18993 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18994 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18995 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18996 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18997 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18998 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18999 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19000 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19001 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19002 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19003 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19004 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19005 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19006 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19007 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19008 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19009 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19010 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19011 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19012 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19013 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19014 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19015 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19016 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19017 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19018 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19019 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19020 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19021 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19022 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19023 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19024 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19025 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19026 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19027 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19028 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19029 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19030 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19031 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19032 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19033 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19034 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19035 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19036 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19037 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19038 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19039 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19040 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19041 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19042 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19043 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19044 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19045 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19046 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19047 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19048 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19049 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19050 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19051 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19052 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19053 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19054 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19055 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19056 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19057 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19058 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19059 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19060 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19061 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19062 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19063 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19064 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19065 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19066 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19067 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19068 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19069 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19070 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19071 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19072 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19073 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19074 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19075 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19076 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19077 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19078 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19079 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19080 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19081 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19082 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19083 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19084 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19085 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19086 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19087 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19088 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19089 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19090 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19091 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19092 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19093 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19094 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19095 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19096 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19097 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19098 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19099 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19100 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19101 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19102 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19103 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19104 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19105 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19106 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19107 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19108 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19109 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19110 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19111 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19112 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19113 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19114 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19115 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19116 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19117 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19118 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19119 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19120 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19121 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19122 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19123 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19124 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19125 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19126 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19127 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19128 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19129 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19130 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19131 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19132 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19133 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19134 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19135 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655