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]
25 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
30 * The deprecated functions EVP_PKEY_get0_RSA(),
31 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
32 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
33 well as the similarly named "get1" functions behave slightly differently in
34 OpenSSL 3.0. Previously they returned a pointer to the low-level key used
35 internally by libcrypto. From OpenSSL 3.0 this key may now be held in a
36 provider. Calling these functions will only return a handle on the internal
37 key where the EVP_PKEY was constructed using this key in the first place, for
38 example using a function or macro such as EVP_PKEY_assign_RSA(),
39 EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key,
40 then these functions now return a cached copy of the key. Changes to
41 the internal provider key that take place after the first time the cached key
42 is accessed will not be reflected back in the cached copy. Similarly any
43 changes made to the cached copy by application code will not be reflected
44 back in the internal provider key.
46 For the above reasons the keys returned from these functions should typically
47 be treated as read-only. To emphasise this the value returned from
48 EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(),
49 EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may
50 break some existing code. Applications broken by this change should be
51 modified. The preferred solution is to refactor the code to avoid the use of
52 these deprecated functions. Failing this the code should be modified to use a
53 const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(),
54 EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a
55 non-const pointer to enable them to be "freed". However they should also be
60 * A number of functions handling low level keys or engines were deprecated
61 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
62 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
63 EVP_PKEY_get0_siphash(). Applications using engines should instead use
64 providers. Applications getting or setting low level keys in an EVP_PKEY
65 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
66 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
70 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
71 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
72 and require returning octet ptr parameters from providers that
73 would like to support them which complicates provider implementations.
77 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
78 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
79 RAND_set_rand_method(). Provider based random number generators should
80 be used instead via EVP_RAND(3).
84 * The SRP APIs have been deprecated. The old APIs do not work via providers,
85 and there is no EVP interface to them. Unfortunately there is no replacement
86 for these APIs at this time.
90 * Add a compile time option to prevent the caching of provider fetched
91 algorithms. This is enabled by including the no-cached-fetch option
92 at configuration time.
96 * The openssl speed command does not use low-level API calls anymore. This
97 implies some of the performance numbers might not be fully comparable
98 with the previous releases due to higher overhead. This applies
99 particularly to measuring performance on smaller data chunks.
103 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
104 Typically if OpenSSL has no EC or DH algorithms then it cannot support
105 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
106 through providers. Therefore third party providers may supply group
107 implementations even where there are no built-in ones. Attempting to create
108 TLS connections in such a build without also disabling TLSv1.3 at run time or
109 using third party provider groups may result in handshake failures. TLSv1.3
110 can be disabled at compile time using the "no-tls1_3" Configure option.
114 * The undocumented function X509_certificate_type() has been deprecated;
115 applications can use X509_get0_pubkey() and X509_get0_signature() to
116 get the same information.
120 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
121 functions. They are identical to BN_rand() and BN_rand_range()
126 * Removed RSA padding mode for SSLv23 (which was only used for
127 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
128 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
133 * Deprecated the obsolete X9.31 RSA key generation related functions
134 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
135 BN_X931_generate_prime_ex().
139 * The default key generation method for the regular 2-prime RSA keys was
140 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
141 Conditions Based on Auxiliary Probable Primes). This method is slower
142 than the original method.
146 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
147 They are replaced with the BN_check_prime() function that avoids possible
148 misuse and always uses at least 64 rounds of the Miller-Rabin
149 primality test. At least 64 rounds of the Miller-Rabin test are now also
150 used for all prime generation, including RSA key generation.
151 This increases key generation time, especially for larger keys.
155 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
156 as they are not useful with non-deprecated functions.
160 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
161 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
162 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
163 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
164 were used to collect all necessary data to form a HTTP request, and to
165 perform the HTTP transfer with that request. With OpenSSL 3.0, the
166 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
167 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
168 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
169 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
170 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
171 OSSL_HTTP_REQ_CTX_set_max_response_length().
173 *Rich Salz and Richard Levitte*
175 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
179 * Validation of SM2 keys has been separated from the validation of regular EC
180 keys, allowing to improve the SM2 validation process to reject loaded private
181 keys that are not conforming to the SM2 ISO standard.
182 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
187 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
188 switches: a validation failure triggers an early exit, returning a failure
189 exit status to the parent process.
193 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
194 to ignore unknown ciphers.
198 * The `-cipher-commands` and `-digest-commands` options
199 of the command line utility `list` have been deprecated.
200 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
204 * All of the low level EC_KEY functions have been deprecated including:
206 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
207 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
208 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
209 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
210 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
211 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
212 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
213 EC_KEY_METHOD_get_verify,
214 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
215 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
216 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
217 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
218 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
219 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
220 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
221 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
222 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
223 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
224 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
225 Applications that need to implement an EC_KEY_METHOD need to consider
226 implementation of the functionality in a special provider.
227 For replacement of the functions manipulating the EC_KEY objects
228 see the EVP_PKEY-EC(7) manual page.
230 Additionally functions that read and write EC_KEY objects such as
231 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
232 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
233 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
234 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
235 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
236 have also been deprecated. Applications should instead use the
237 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
239 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
240 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
241 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
242 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
243 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
245 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
247 * Deprecated all the libcrypto and libssl error string loading
248 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
249 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
250 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
251 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
252 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
253 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
254 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
255 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
256 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
257 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
259 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
260 now loads error strings automatically.
264 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
265 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
266 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
267 are to be used by servers requiring ephemeral DH keys. Instead applications
268 should consider using the built-in DH parameters that are available by
269 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
270 necessary then applications can use the alternative functions
271 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
272 replacement for the "callback" functions. The callback was originally useful
273 in order to have different parameters for export and non-export ciphersuites.
274 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
275 functions should be replaced by one of the other methods described above.
279 * The `-crypt` option to the `passwd` command line tool has been removed.
283 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
288 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
290 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
291 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
292 The inverse ciphers use AES decryption for wrapping, and
293 AES encryption for unwrapping.
297 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
298 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
299 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
300 OpenSSL 3.0 these are replaced by the more generic functions
301 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
302 The old versions have been converted to deprecated macros that just call the
307 * The security callback, which can be customised by application code, supports
308 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
309 in the "other" parameter. In most places this is what is passed. All these
310 places occur server side. However there was one client side call of this
311 security operation and it passed a DH object instead. This is incorrect
312 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
313 of the other locations. Therefore this client side call has been changed to
314 pass an EVP_PKEY instead.
318 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
319 interface. Their functionality remains unchanged.
323 * Added new option for 'openssl list', '-providers', which will display the
324 list of loaded providers, their names, version and status. It optionally
325 displays their gettable parameters.
329 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
330 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
331 type is internally recognised so the workaround is no longer needed.
333 Functionality is still retained as it is, but will only work with
334 EVP_PKEYs with a legacy internal key.
338 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
339 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
343 * Changed all "STACK" functions to be macros instead of inline functions. Macro
344 parameters are still checked for type safety at compile time via helper
349 * Remove the RAND_DRBG API
351 The RAND_DRBG API did not fit well into the new provider concept as
352 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
353 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
354 and some of its API calls are rather low-level. This holds in particular
355 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
357 Adding a compatibility layer to continue supporting the RAND_DRBG API as
358 a legacy API for a regular deprecation period turned out to come at the
359 price of complicating the new provider API unnecessarily. Since the
360 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
363 *Paul Dale and Matthias St. Pierre*
365 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
366 as well as actual hostnames.
370 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
371 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
372 conversely, silently ignore DTLS protocol version bounds when configuring
373 TLS-based contexts. The commands can be repeated to set bounds of both
374 types. The same applies with the corresponding "min_protocol" and
375 "max_protocol" command-line switches, in case some application uses both TLS
378 SSL_CTX instances that are created for a fixed protocol version (e.g.
379 `TLSv1_server_method()`) also silently ignore version bounds. Previously
380 attempts to apply bounds to these protocol versions would result in an
381 error. Now only the "version-flexible" SSL_CTX instances are subject to
382 limits in configuration files in command-line options.
386 * Deprecated the `ENGINE` API. Engines should be replaced with providers
391 * Reworked the recorded ERR codes to make better space for system errors.
392 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
393 given code is a system error (true) or an OpenSSL error (false).
397 * Reworked the test perl framework to better allow parallel testing.
399 *Nicola Tuveri and David von Oheimb*
401 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
402 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
406 * 'Configure' has been changed to figure out the configuration target if
407 none is given on the command line. Consequently, the 'config' script is
408 now only a mere wrapper. All documentation is changed to only mention
411 *Rich Salz and Richard Levitte*
413 * Added a library context `OSSL_LIB_CTX` that applications as well as
414 other libraries can use to form a separate context within which
415 libcrypto operations are performed.
417 There are two ways this can be used:
419 - Directly, by passing a library context to functions that take
420 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
422 - Indirectly, by creating a new library context and then assigning
423 it as the new default, with `OSSL_LIB_CTX_set0_default`.
425 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
426 apart from the functions directly related to `OSSL_LIB_CTX`, accept
427 NULL to indicate that the default library context should be used.
429 Library code that changes the default library context using
430 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
431 second call before returning to the caller.
433 _(Note: the library context was initially called `OPENSSL_CTX` and
434 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
438 * Handshake now fails if Extended Master Secret extension is dropped
443 * Dropped interactive mode from the `openssl` program. From now on,
444 running it without arguments is equivalent to `openssl help`.
448 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
449 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
450 While the old function names have been retained for backward compatibility
451 they should not be used in new developments
452 because their return values are confusing: Unlike other `_cmp()` functions
453 they do not return 0 in case their arguments are equal.
457 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
458 `EC_GROUP_get_field_type()`.
462 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
463 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
464 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
465 Applications should rely on the library automatically assigning a suitable
466 EC_METHOD internally upon EC_GROUP construction.
470 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
471 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
472 assigned internally without application intervention.
473 Users of EC_GROUP_new() should switch to a different suitable constructor.
477 * Add CAdES-BES signature verification support, mostly derived
478 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
480 *Filipe Raimundo da Silva*
482 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
486 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
487 functions are not widely used and now OpenSSL automatically perform this
488 conversion when needed.
492 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
493 EC_KEY_precompute_mult(). These functions are not widely used and
494 applications should instead switch to named curves which OpenSSL has
495 hardcoded lookup tables for.
499 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
500 should instead use the L<EC_POINT_mul(3)> function.
504 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
505 that are not applicable to the new provider model. Applications should
506 instead use EVP_default_properties_is_fips_enabled() and
507 EVP_default_properties_enable_fips().
511 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
512 is set, an unexpected EOF is ignored, it pretends a close notify was received
513 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
517 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
518 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
519 used and applications should instead use the
520 L<EC_POINT_set_affine_coordinates(3)> and
521 L<EC_POINT_get_affine_coordinates(3)> functions.
525 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
526 arrays to be more easily constructed via a series of utility functions.
527 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
528 the various push functions and finally convert to a passable OSSL_PARAM
529 array using OSSL_PARAM_BLD_to_param().
533 * The security strength of SHA1 and MD5 based signatures in TLS has been
534 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
535 working at the default security level of 1 and instead requires security
536 level 0. The security level can be changed either using the cipher string
537 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
541 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
542 contain a provider side internal key.
546 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
547 They are old functions that we don't use, and that you could disable with
548 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
552 * Project text documents not yet having a proper file name extension
553 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
554 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
558 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
559 have been converted to Markdown with the goal to produce documents
560 which not only look pretty when viewed online in the browser, but
561 remain well readable inside a plain text editor.
563 To achieve this goal, a 'minimalistic' Markdown style has been applied
564 which avoids formatting elements that interfere too much with the
565 reading flow in the text file. For example, it
567 * avoids [ATX headings][] and uses [setext headings][] instead
568 (which works for `<h1>` and `<h2>` headings only).
569 * avoids [inline links][] and uses [reference links][] instead.
570 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
572 [ATX headings]: https://github.github.com/gfm/#atx-headings
573 [setext headings]: https://github.github.com/gfm/#setext-headings
574 [inline links]: https://github.github.com/gfm/#inline-link
575 [reference links]: https://github.github.com/gfm/#reference-link
576 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
577 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
579 *Matthias St. Pierre*
581 * The test suite is changed to preserve results of each test recipe.
582 A new directory test-runs/ with subdirectories named like the
583 test recipes are created in the build tree for this purpose.
587 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
588 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
589 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
591 *David von Oheimb, Martin Peylo*
593 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
594 The legacy OCSP-focused and only partly documented API is retained for
595 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
599 * Added `util/check-format.pl`, a tool for checking adherence to the
600 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
601 The checks performed are incomplete and yield some false positives.
602 Still the tool should be useful for detecting most typical glitches.
606 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
607 If domain name resolution yields multiple IP addresses all of them are tried
608 after `connect()` failures.
612 * All of the low level RSA functions have been deprecated including:
614 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
615 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
616 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
617 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
618 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
619 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
620 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
621 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
622 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
623 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
624 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
625 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
626 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
627 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
628 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
629 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
630 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
631 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
632 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
633 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
634 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
635 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
636 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
637 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
638 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
639 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
640 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
641 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
643 Use of these low level functions has been informally discouraged for a long
644 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
645 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
646 L<EVP_PKEY_decrypt(3)>.
650 * X509 certificates signed using SHA1 are no longer allowed at security
652 In TLS/SSL the default security level is 1. It can be set either
653 using the cipher string with `@SECLEVEL`, or calling
654 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
655 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
657 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
658 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
659 options of the commands.
663 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
664 modified to use PKEY APIs. These commands are now in maintenance mode
665 and no new features will be added to them.
669 * The command line utility rsautl has been deprecated.
670 Instead use the pkeyutl program.
674 * The command line utilities genrsa and rsa have been modified to use PKEY
675 APIs. They now write PKCS#8 keys by default. These commands are now in
676 maintenance mode and no new features will be added to them.
680 * All of the low level DH functions have been deprecated including:
682 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
683 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
684 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
685 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
686 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
687 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
688 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
689 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
690 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
691 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
692 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
693 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
694 DH_meth_get_generate_params and DH_meth_set_generate_params.
696 Use of these low level functions has been informally discouraged for a long
697 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
698 and L<EVP_PKEY_derive(3)>.
700 Additionally functions that read and write DH objects such as d2i_DHparams,
701 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
702 functions have also been deprecated. Applications should instead use the
703 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
705 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
706 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
707 `EVP_PKEY_set1_DH()` are also deprecated.
708 Applications should instead either read or write an
709 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
710 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
712 *Paul Dale and Matt Caswell*
714 * All of the low level DSA functions have been deprecated including:
716 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
717 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
718 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
719 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
720 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
721 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
722 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
723 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
724 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
725 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
726 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
727 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
728 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
729 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
730 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
731 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
732 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
734 Use of these low level functions has been informally discouraged for a long
735 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
736 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
738 Finaly functions that assign or obtain DSA objects from an EVP_PKEY such as
739 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
740 `EVP_PKEY_set1_DSA()` are also deprecated.
741 Applications should instead either read or write an
742 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs,
743 or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
747 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
748 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
749 This means that applications don't have to look at the curve NID and
750 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
751 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
752 a no-op when the EVP_PKEY is already of the given type.
754 Parameter and key generation is also reworked to make it possible
755 to generate EVP_PKEY_SM2 parameters and keys without having to go
756 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
757 However, code that does the latter will still work as before.
761 * Deprecated low level ECDH and ECDSA functions. These include:
763 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
764 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
767 Use of these low level functions has been informally discouraged for a long
768 time. Instead applications should use the EVP_PKEY_derive(3),
769 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
773 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
774 and EVP_PKEY_decrypt() instead.
775 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
776 and EVP_PKEY_encrypt() instead.
780 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
781 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
782 a new formulation to include all the things it can be used for,
783 as well as words of caution.
787 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
788 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
792 * All of the low level HMAC functions have been deprecated including:
794 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
795 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
798 Use of these low level functions has been informally discouraged for a long
799 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
800 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
801 and L<EVP_MAC_final(3)>.
805 * Over two thousand fixes were made to the documentation, including:
806 - Common options (such as -rand/-writerand, TLS version control, etc)
807 were refactored and point to newly-enhanced descriptions in openssl.pod.
808 - Added style conformance for all options (with help from Richard Levitte),
809 documented all reported missing options, added a CI build to check
810 that all options are documented and that no unimplemented options
812 - Documented some internals, such as all use of environment variables.
813 - Addressed all internal broken L<> references.
817 * All of the low level CMAC functions have been deprecated including:
819 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
820 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
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 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
830 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
833 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
834 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
835 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
836 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
837 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
838 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
839 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
840 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
841 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
842 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
844 Use of these low level functions has been informally discouraged
845 for a long time. Applications should use the EVP_DigestInit_ex(3),
846 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
850 * Corrected the documentation of the return values from the `EVP_DigestSign*`
851 set of functions. The documentation mentioned negative values for some
852 errors, but this was never the case, so the mention of negative values
855 Code that followed the documentation and thereby check with something
856 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
860 * All of the low level cipher functions have been deprecated including:
862 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
863 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
864 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
865 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
866 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
867 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
868 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
869 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
870 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
871 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
872 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
873 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
874 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
875 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
876 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
877 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
878 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
879 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
880 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
881 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
882 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
883 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
884 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
885 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
886 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
887 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
888 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
889 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
890 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
892 Use of these low level functions has been informally discouraged for
893 a long time. Applications should use the high level EVP APIs, e.g.
894 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
895 equivalently named decrypt functions instead.
897 *Matt Caswell and Paul Dale*
899 * Removed include/openssl/opensslconf.h.in and replaced it with
900 include/openssl/configuration.h.in, which differs in not including
901 <openssl/macros.h>. A short header include/openssl/opensslconf.h
902 was added to include both.
904 This allows internal hacks where one might need to modify the set
905 of configured macros, for example this if deprecated symbols are
906 still supposed to be available internally:
908 #include <openssl/configuration.h>
910 #undef OPENSSL_NO_DEPRECATED
911 #define OPENSSL_SUPPRESS_DEPRECATED
913 #include <openssl/macros.h>
915 This should not be used by applications that use the exported
916 symbols, as that will lead to linking errors.
920 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
921 used in exponentiation with 512-bit moduli. No EC algorithms are
922 affected. Analysis suggests that attacks against 2-prime RSA1024,
923 3-prime RSA1536, and DSA1024 as a result of this defect would be very
924 difficult to perform and are not believed likely. Attacks against DH512
925 are considered just feasible. However, for an attack the target would
926 have to re-use the DH512 private key, which is not recommended anyway.
927 Also applications directly using the low level API BN_mod_exp may be
928 affected if they use BN_FLG_CONSTTIME.
933 * Most memory-debug features have been deprecated, and the functionality
934 replaced with no-ops.
938 * Added documentation for the STACK API.
942 * Introduced a new method type and API, OSSL_ENCODER, to
943 represent generic encoders. An implementation is expected to
944 be able to encode an object associated with a given name (such
945 as an algorithm name for an asymmetric key) into forms given by
946 implementation properties.
948 Encoders are primarily used from inside libcrypto, through
949 calls to functions like EVP_PKEY_print_private(),
950 PEM_write_bio_PrivateKey() and similar.
952 Encoders are specified in such a way that they can be made to
953 directly handle the provider side portion of an object, if this
954 provider side part comes from the same provider as the encoder
955 itself, but can also be made to handle objects in parametrized
956 form (as an OSSL_PARAM array of data). This allows a provider to
957 offer generic encoders as a service for any other provider.
961 * Added a .pragma directive to the syntax of configuration files, to
962 allow varying behavior in a supported and predictable manner.
963 Currently added pragma:
967 This allows dollar signs to be a keyword character unless it's
968 followed by a opening brace or parenthesis. This is useful for
969 platforms where dollar signs are commonly used in names, such as
970 volume names and system directory names on VMS.
974 * Added functionality to create an EVP_PKEY from user data. This
975 is effectively the same as creating a RSA, DH or DSA object and
976 then assigning them to an EVP_PKEY, but directly using algorithm
977 agnostic EVP functions. A benefit is that this should be future
978 proof for public key algorithms to come.
982 * Change the interpretation of the '--api' configuration option to
983 mean that this is a desired API compatibility level with no
984 further meaning. The previous interpretation, that this would
985 also mean to remove all deprecated symbols up to and including
986 the given version, no requires that 'no-deprecated' is also used
987 in the configuration.
989 When building applications, the desired API compatibility level
990 can be set with the OPENSSL_API_COMPAT macro like before. For
991 API compatibility version below 3.0, the old style numerical
992 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
993 For version 3.0 and on, the value is expected to be the decimal
994 value calculated from the major and minor version like this:
996 MAJOR * 10000 + MINOR * 100
1000 -DOPENSSL_API_COMPAT=30000 For 3.0
1001 -DOPENSSL_API_COMPAT=30200 For 3.2
1003 To hide declarations that are deprecated up to and including the
1004 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1005 given when building the application as well.
1009 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1010 access to certificate and CRL stores via URIs and OSSL_STORE
1013 This adds the following functions:
1015 - X509_LOOKUP_store()
1016 - X509_STORE_load_file()
1017 - X509_STORE_load_path()
1018 - X509_STORE_load_store()
1019 - SSL_add_store_cert_subjects_to_stack()
1020 - SSL_CTX_set_default_verify_store()
1021 - SSL_CTX_load_verify_file()
1022 - SSL_CTX_load_verify_dir()
1023 - SSL_CTX_load_verify_store()
1027 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1028 The presence of this system service is determined at run-time.
1032 * Added functionality to create an EVP_PKEY context based on data
1033 for methods from providers. This takes an algorithm name and a
1034 property query string and simply stores them, with the intent
1035 that any operation that uses this context will use those strings
1036 to fetch the needed methods implicitly, thereby making the port
1037 of application written for pre-3.0 OpenSSL easier.
1041 * The undocumented function NCONF_WIN32() has been deprecated; for
1042 conversion details see the HISTORY section of doc/man5/config.pod
1046 * Introduced the new functions EVP_DigestSignInit_ex() and
1047 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1048 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1049 pages for further details.
1053 * Over two thousand fixes were made to the documentation, including:
1054 adding missing command flags, better style conformance, documentation
1057 *Rich Salz, Richard Levitte*
1059 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1060 X25519, X448, Ed25519 and Ed448.
1064 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1069 * Deprecated the public definition of `ERR_STATE` as well as the function
1070 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1075 * Added ERR functionality to give callers access to the stored function
1076 names that have replaced the older function code based functions.
1078 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1079 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1080 ERR_peek_error_all() and ERR_peek_last_error_all().
1082 These functions have become deprecated: ERR_get_error_line(),
1083 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1084 ERR_peek_last_error_line_data() and ERR_func_error_string().
1086 Users are recommended to use ERR_get_error_all(), or to pick information
1087 with ERR_peek functions and finish off with getting the error code by using
1092 * Extended testing to be verbose for failing tests only. The make variables
1093 VERBOSE_FAILURE or VF can be used to enable this:
1095 $ make VF=1 test # Unix
1096 $ mms /macro=(VF=1) test ! OpenVMS
1097 $ nmake VF=1 test # Windows
1101 * Added the `-copy_extensions` option to the `x509` command for use with
1102 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1103 all extensions in the request are copied to the certificate or vice versa.
1105 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1107 * Added the `-copy_extensions` option to the `req` command for use with
1108 `-x509`. When given with the `copy` or `copyall` argument,
1109 all extensions in the certification request are copied to the certificate.
1113 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1114 they generate are by default RFC 5280 compliant in the following sense:
1115 There is a subjectKeyIdentifier extension with a hash value of the public key
1116 and for not self-signed certs there is an authorityKeyIdentifier extension
1117 with a keyIdentifier field or issuer information identifying the signing key.
1118 This is done unless some configuration overrides the new default behavior,
1119 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1123 * Added several checks to `X509_verify_cert()` according to requirements in
1124 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1125 (which may be done by using the CLI option `-x509_strict`):
1126 * The basicConstraints of CA certificates must be marked critical.
1127 * CA certificates must explicitly include the keyUsage extension.
1128 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1129 * The issuer name of any certificate must not be empty.
1130 * The subject name of CA certs, certs with keyUsage crlSign,
1131 and certs without subjectAlternativeName must not be empty.
1132 * If a subjectAlternativeName extension is given it must not be empty.
1133 * The signatureAlgorithm field and the cert signature must be consistent.
1134 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1135 must not be marked critical.
1136 * The authorityKeyIdentifier must be given for X.509v3 certs
1137 unless they are self-signed.
1138 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1142 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1143 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1147 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1148 used even when parsing explicit parameters, when loading a encoded key
1149 or calling `EC_GROUP_new_from_ecpkparameters()`/
1150 `EC_GROUP_new_from_ecparameters()`.
1151 This prevents bypass of security hardening and performance gains,
1152 especially for curves with specialized EC_METHODs.
1153 By default, if a key encoded with explicit parameters is loaded and later
1154 encoded, the output is still encoded with explicit parameters, even if
1155 internally a "named" EC_GROUP is used for computation.
1159 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1160 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1161 NULL. After this change, only the cofactor parameter can be NULL. It also
1162 does some minimal sanity checks on the passed order.
1167 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1168 An attack is simple, if the first CMS_recipientInfo is valid but the
1169 second CMS_recipientInfo is chosen ciphertext. If the second
1170 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1171 encryption key will be replaced by garbage, and the message cannot be
1172 decoded, but if the RSA decryption fails, the correct encryption key is
1173 used and the recipient will not notice the attack.
1174 As a work around for this potential attack the length of the decrypted
1175 key must be equal to the cipher default key length, in case the
1176 certifiate is not given and all recipientInfo are tried out.
1177 The old behaviour can be re-enabled in the CMS code by setting the
1178 CMS_DEBUG_DECRYPT flag.
1182 * Early start up entropy quality from the DEVRANDOM seed source has been
1183 improved for older Linux systems. The RAND subsystem will wait for
1184 /dev/random to be producing output before seeding from /dev/urandom.
1185 The seeded state is stored for future library initialisations using
1186 a system global shared memory segment. The shared memory identifier
1187 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1188 the desired value. The default identifier is 114.
1192 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1193 when primes for RSA keys are computed.
1194 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1195 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1196 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1197 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1198 This avoids possible fingerprinting of newly generated RSA modules.
1202 * Correct the extended master secret constant on EBCDIC systems. Without this
1203 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1204 negotiate EMS will fail. Unfortunately this also means that TLS connections
1205 between EBCDIC systems with this fix, and EBCDIC systems without this
1206 fix will fail if they negotiate EMS.
1210 * Changed the library initialisation so that the config file is now loaded
1211 by default. This was already the case for libssl. It now occurs for both
1212 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1213 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1217 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1218 where the former acts as a replacement for `ERR_put_error()`, and the
1219 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1220 `ERR_raise_data()` adds more flexibility by taking a format string and
1221 an arbitrary number of arguments following it, to be processed with
1226 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1227 to check if a named provider is loaded and available. When called, it
1228 will also activate all fallback providers if such are still present.
1232 * Enforce a minimum DH modulus size of 512 bits.
1236 * Changed DH parameters to generate the order q subgroup instead of 2q.
1237 Previously generated DH parameters are still accepted by DH_check
1238 but DH_generate_key works around that by clearing bit 0 of the
1239 private key for those. This avoids leaking bit 0 of the private key.
1243 * Significantly reduce secure memory usage by the randomness pools.
1247 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1252 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1253 algorithms. An implementation of a key exchange algorithm can be obtained
1254 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1255 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1256 the older EVP_PKEY_derive_init() function. See the man pages for the new
1257 functions for further details.
1261 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1265 * Removed the function names from error messages and deprecated the
1268 * Removed NextStep support and the macro OPENSSL_UNISTD
1272 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1273 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1274 Also removed "export var as function" capability; we do not export
1275 variables, only functions.
1279 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1280 an error and 1 indicating success. In previous versions of OpenSSL this
1281 was a void type. If a key was set longer than the maximum possible this
1286 * Support SM2 signing and verification schemes with X509 certificate.
1290 * Use SHA256 as the default digest for TS query in the `ts` app.
1294 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1295 This checks that the salt length is at least 128 bits, the derived key
1296 length is at least 112 bits, and that the iteration count is at least 1000.
1297 For backwards compatibility these checks are disabled by default in the
1298 default provider, but are enabled by default in the fips provider.
1299 To enable or disable these checks use the control
1300 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1304 * Default cipher lists/suites are now available via a function, the
1305 #defines are deprecated.
1309 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1310 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1311 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1315 * Join the directories crypto/x509 and crypto/x509v3
1319 * Added command 'openssl kdf' that uses the EVP_KDF API.
1323 * Added command 'openssl mac' that uses the EVP_MAC API.
1327 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1328 as default directories. Also added the command 'openssl info'
1329 for scripting purposes.
1333 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1334 deprecated. These undocumented functions were never integrated into the EVP
1335 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1336 Bi-directional IGE mode. These modes were never formally standardised and
1337 usage of these functions is believed to be very small. In particular
1338 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1339 is ever used. The security implications are believed to be minimal, but
1340 this issue was never fixed for backwards compatibility reasons. New code
1341 should not use these modes.
1345 * Add prediction resistance to the DRBG reseeding process.
1349 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1350 mandated by IEEE Std 1619-2018.
1354 * Added newline escaping functionality to a filename when using openssl dgst.
1355 This output format is to replicate the output format found in the `*sum`
1356 checksum programs. This aims to preserve backward compatibility.
1358 *Matt Eaton, Richard Levitte, and Paul Dale*
1360 * Removed the heartbeat message in DTLS feature, as it has very
1361 little usage and doesn't seem to fulfill a valuable purpose.
1362 The configuration option is now deprecated.
1366 * Changed the output of 'openssl {digestname} < file' to display the
1367 digest name in its output.
1371 * Added a new generic trace API which provides support for enabling
1372 instrumentation through trace output. This feature is mainly intended
1373 as an aid for developers and is disabled by default. To utilize it,
1374 OpenSSL needs to be configured with the `enable-trace` option.
1376 If the tracing API is enabled, the application can activate trace output
1377 by registering BIOs as trace channels for a number of tracing and debugging
1380 The `openssl` program has been expanded to enable any of the types
1381 available via environment variables defined by the user, and serves as
1382 one possible example on how to use this functionality.
1384 *Richard Levitte & Matthias St. Pierre*
1386 * Added build tests for C++. These are generated files that only do one
1387 thing, to include one public OpenSSL head file each. This tests that
1388 the public header files can be usefully included in a C++ application.
1390 This test isn't enabled by default. It can be enabled with the option
1391 'enable-buildtest-c++'.
1395 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1399 * Add KMAC to EVP_MAC.
1403 * Added property based algorithm implementation selection framework to
1408 * Added SCA hardening for modular field inversion in EC_GROUP through
1409 a new dedicated field_inv() pointer in EC_METHOD.
1410 This also addresses a leakage affecting conversions from projective
1411 to affine coordinates.
1413 *Billy Bob Brumley, Nicola Tuveri*
1415 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1416 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1417 those algorithms that were already supported through the EVP_PKEY API
1418 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1419 and scrypt are now wrappers that call EVP_KDF.
1423 * Build devcrypto engine as a dynamic engine.
1425 *Eneas U de Queiroz*
1427 * Add keyed BLAKE2 to EVP_MAC.
1431 * Fix a bug in the computation of the endpoint-pair shared secret used
1432 by DTLS over SCTP. This breaks interoperability with older versions
1433 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1434 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1435 interoperability with such broken implementations. However, enabling
1436 this switch breaks interoperability with correct implementations.
1438 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1439 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1443 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1447 * Change the license to the Apache License v2.0.
1451 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1453 - Major releases (indicated by incrementing the MAJOR release number)
1454 may introduce incompatible API/ABI changes.
1455 - Minor releases (indicated by incrementing the MINOR release number)
1456 may introduce new features but retain API/ABI compatibility.
1457 - Patch releases (indicated by incrementing the PATCH number)
1458 are intended for bug fixes and other improvements of existing
1459 features only (like improving performance or adding documentation)
1460 and retain API/ABI compatibility.
1464 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1468 * Remove the 'dist' target and add a tarball building script. The
1469 'dist' target has fallen out of use, and it shouldn't be
1470 necessary to configure just to create a source distribution.
1474 * Recreate the OS390-Unix config target. It no longer relies on a
1475 special script like it did for OpenSSL pre-1.1.0.
1479 * Instead of having the source directories listed in Configure, add
1480 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1485 * Add GMAC to EVP_MAC.
1489 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1493 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1494 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1495 to facilitate the continued use of MACs through raw private keys in
1496 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1500 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1501 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1505 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1506 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1507 are retained for backwards compatibility.
1511 * AES-XTS mode now enforces that its two keys are different to mitigate
1512 the attacked described in "Efficient Instantiations of Tweakable
1513 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1514 Details of this attack can be obtained from:
1515 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1519 * Rename the object files, i.e. give them other names than in previous
1520 versions. Their names now include the name of the final product, as
1521 well as its type mnemonic (bin, lib, shlib).
1525 * Added new option for 'openssl list', '-objects', which will display the
1526 list of built in objects, i.e. OIDs with names.
1530 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1531 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1536 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1537 improves application performance by removing data copies and providing
1538 applications with zero-copy system calls such as sendfile and splice.
1542 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1543 option is set, openssl cleanses (zeroize) plaintext bytes from
1544 internal buffers after delivering them to the application. Note,
1545 the application is still responsible for cleansing other copies
1546 (e.g.: data received by SSL_read(3)).
1550 * `PKCS12_parse` now maintains the order of the parsed certificates
1551 when outputting them via `*ca` (rather than reversing it).
1555 * Deprecated pthread fork support methods. These were unused so no
1556 replacement is required.
1558 - OPENSSL_fork_prepare()
1559 - OPENSSL_fork_parent()
1560 - OPENSSL_fork_child()
1567 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1569 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1571 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1572 create a unique hash value based on the issuer and serial number data
1573 contained within an X509 certificate. However it was failing to correctly
1574 handle any errors that may occur while parsing the issuer field (which might
1575 occur if the issuer field is maliciously constructed). This may subsequently
1576 result in a NULL pointer deref and a crash leading to a potential denial of
1582 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1583 padding mode to correctly check for rollback attacks. This is considered a
1584 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1589 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1590 functions. Previously they could overflow the output length argument in some
1591 cases where the input length is close to the maximum permissable length for
1592 an integer on the platform. In such cases the return value from the function
1593 call would be 1 (indicating success), but the output length value would be
1594 negative. This could cause applications to behave incorrectly or crash.
1599 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1600 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1601 could be exploited in a side channel attack to recover the password. Since
1602 the attack is local host only this is outside of the current OpenSSL
1603 threat model and therefore no CVE is assigned.
1605 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1610 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1612 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1613 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1614 If an attacker can control both items being compared then this could lead
1615 to a possible denial of service attack. OpenSSL itself uses the
1616 GENERAL_NAME_cmp function for two purposes:
1617 1) Comparing CRL distribution point names between an available CRL and a
1618 CRL distribution point embedded in an X509 certificate
1619 2) When verifying that a timestamp response token signer matches the
1620 timestamp authority name (exposed via the API functions
1621 TS_RESP_verify_response and TS_RESP_verify_token)
1626 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1628 * Certificates with explicit curve parameters are now disallowed in
1629 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1633 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1634 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1635 conversely, silently ignore DTLS protocol version bounds when configuring
1636 TLS-based contexts. The commands can be repeated to set bounds of both
1637 types. The same applies with the corresponding "min_protocol" and
1638 "max_protocol" command-line switches, in case some application uses both TLS
1641 SSL_CTX instances that are created for a fixed protocol version (e.g.
1642 TLSv1_server_method()) also silently ignore version bounds. Previously
1643 attempts to apply bounds to these protocol versions would result in an
1644 error. Now only the "version-flexible" SSL_CTX instances are subject to
1645 limits in configuration files in command-line options.
1649 * Handshake now fails if Extended Master Secret extension is dropped
1654 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1656 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1658 * Fixed segmentation fault in SSL_check_chain()
1659 Server or client applications that call the SSL_check_chain() function
1660 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1661 dereference as a result of incorrect handling of the
1662 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1663 or unrecognised signature algorithm is received from the peer. This could
1664 be exploited by a malicious peer in a Denial of Service attack.
1669 * Added AES consttime code for no-asm configurations
1670 an optional constant time support for AES was added
1671 when building openssl for no-asm.
1672 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1673 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1674 At this time this feature is by default disabled.
1675 It will be enabled by default in 3.0.
1679 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1681 * Revert the change of EOF detection while reading in libssl to avoid
1682 regressions in applications depending on the current way of reporting
1683 the EOF. As the existing method is not fully accurate the change to
1684 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1685 branch and will be present in the 3.0 release.
1689 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1690 when primes for RSA keys are computed.
1691 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1692 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1693 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1694 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1695 This avoids possible fingerprinting of newly generated RSA modules.
1699 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1701 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1702 while reading in libssl then we would report an error back to the
1703 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1704 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1705 therefore give a hint as to what went wrong.
1709 * Check that ed25519 and ed448 are allowed by the security level. Previously
1710 signature algorithms not using an MD were not being checked that they were
1711 allowed by the security level.
1715 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1716 was not quite right. The behaviour was not consistent between resumption
1717 and normal handshakes, and also not quite consistent with historical
1718 behaviour. The behaviour in various scenarios has been clarified and
1719 it has been updated to make it match historical behaviour as closely as
1724 * *[VMS only]* The header files that the VMS compilers include automatically,
1725 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1726 that the C++ compiler doesn't understand. This is a shortcoming in the
1727 compiler, but can be worked around with `__cplusplus` guards.
1729 C++ applications that use OpenSSL libraries must be compiled using the
1730 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1731 functions. Otherwise, only functions with symbols of less than 31
1732 characters can be used, as the linker will not be able to successfully
1733 resolve symbols with longer names.
1737 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1738 The presence of this system service is determined at run-time.
1742 * Added newline escaping functionality to a filename when using openssl dgst.
1743 This output format is to replicate the output format found in the `*sum`
1744 checksum programs. This aims to preserve backward compatibility.
1746 *Matt Eaton, Richard Levitte, and Paul Dale*
1748 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1753 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1755 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1756 number generator (RNG). This was intended to include protection in the
1757 event of a fork() system call in order to ensure that the parent and child
1758 processes did not share the same RNG state. However this protection was not
1759 being used in the default case.
1761 A partial mitigation for this issue is that the output from a high
1762 precision timer is mixed into the RNG state so the likelihood of a parent
1763 and child process sharing state is significantly reduced.
1765 If an application already calls OPENSSL_init_crypto() explicitly using
1766 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1769 *Matthias St. Pierre*
1771 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1772 used even when parsing explicit parameters, when loading a encoded key
1773 or calling `EC_GROUP_new_from_ecpkparameters()`/
1774 `EC_GROUP_new_from_ecparameters()`.
1775 This prevents bypass of security hardening and performance gains,
1776 especially for curves with specialized EC_METHODs.
1777 By default, if a key encoded with explicit parameters is loaded and later
1778 encoded, the output is still encoded with explicit parameters, even if
1779 internally a "named" EC_GROUP is used for computation.
1783 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1784 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1785 NULL. After this change, only the cofactor parameter can be NULL. It also
1786 does some minimal sanity checks on the passed order.
1791 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1792 An attack is simple, if the first CMS_recipientInfo is valid but the
1793 second CMS_recipientInfo is chosen ciphertext. If the second
1794 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1795 encryption key will be replaced by garbage, and the message cannot be
1796 decoded, but if the RSA decryption fails, the correct encryption key is
1797 used and the recipient will not notice the attack.
1798 As a work around for this potential attack the length of the decrypted
1799 key must be equal to the cipher default key length, in case the
1800 certifiate is not given and all recipientInfo are tried out.
1801 The old behaviour can be re-enabled in the CMS code by setting the
1802 CMS_DEBUG_DECRYPT flag.
1807 * Early start up entropy quality from the DEVRANDOM seed source has been
1808 improved for older Linux systems. The RAND subsystem will wait for
1809 /dev/random to be producing output before seeding from /dev/urandom.
1810 The seeded state is stored for future library initialisations using
1811 a system global shared memory segment. The shared memory identifier
1812 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1813 the desired value. The default identifier is 114.
1817 * Correct the extended master secret constant on EBCDIC systems. Without this
1818 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1819 negotiate EMS will fail. Unfortunately this also means that TLS connections
1820 between EBCDIC systems with this fix, and EBCDIC systems without this
1821 fix will fail if they negotiate EMS.
1825 * Use Windows installation paths in the mingw builds
1827 Mingw isn't a POSIX environment per se, which means that Windows
1828 paths should be used for installation.
1833 * Changed DH_check to accept parameters with order q and 2q subgroups.
1834 With order 2q subgroups the bit 0 of the private key is not secret
1835 but DH_generate_key works around that by clearing bit 0 of the
1836 private key for those. This avoids leaking bit 0 of the private key.
1840 * Significantly reduce secure memory usage by the randomness pools.
1844 * Revert the DEVRANDOM_WAIT feature for Linux systems
1846 The DEVRANDOM_WAIT feature added a select() call to wait for the
1847 /dev/random device to become readable before reading from the
1848 /dev/urandom device.
1850 It turned out that this change had negative side effects on
1851 performance which were not acceptable. After some discussion it
1852 was decided to revert this feature and leave it up to the OS
1853 resp. the platform maintainer to ensure a proper initialization
1854 during early boot time.
1856 *Matthias St. Pierre*
1858 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1860 * Add build tests for C++. These are generated files that only do one
1861 thing, to include one public OpenSSL head file each. This tests that
1862 the public header files can be usefully included in a C++ application.
1864 This test isn't enabled by default. It can be enabled with the option
1865 'enable-buildtest-c++'.
1869 * Enable SHA3 pre-hashing for ECDSA and DSA.
1873 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1874 This changes the size when using the `genpkey` command when no size is given.
1875 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1876 generation commands to use 2048 bits by default.
1880 * Reorganize the manual pages to consistently have RETURN VALUES,
1881 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1882 util/fix-doc-nits accordingly.
1884 *Paul Yang, Joshua Lock*
1886 * Add the missing accessor EVP_PKEY_get0_engine()
1890 * Have commands like `s_client` and `s_server` output the signature scheme
1891 along with other cipher suite parameters when debugging.
1895 * Make OPENSSL_config() error agnostic again.
1899 * Do the error handling in RSA decryption constant time.
1903 * Prevent over long nonces in ChaCha20-Poly1305.
1905 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1906 for every encryption operation. RFC 7539 specifies that the nonce value
1907 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1908 and front pads the nonce with 0 bytes if it is less than 12
1909 bytes. However it also incorrectly allows a nonce to be set of up to 16
1910 bytes. In this case only the last 12 bytes are significant and any
1911 additional leading bytes are ignored.
1913 It is a requirement of using this cipher that nonce values are
1914 unique. Messages encrypted using a reused nonce value are susceptible to
1915 serious confidentiality and integrity attacks. If an application changes
1916 the default nonce length to be longer than 12 bytes and then makes a
1917 change to the leading bytes of the nonce expecting the new value to be a
1918 new unique nonce then such an application could inadvertently encrypt
1919 messages with a reused nonce.
1921 Additionally the ignored bytes in a long nonce are not covered by the
1922 integrity guarantee of this cipher. Any application that relies on the
1923 integrity of these ignored leading bytes of a long nonce may be further
1924 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1925 is safe because no such use sets such a long nonce value. However user
1926 applications that use this cipher directly and set a non-default nonce
1927 length to be longer than 12 bytes may be vulnerable.
1929 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1935 * Add DEVRANDOM_WAIT feature for Linux systems
1937 On older Linux systems where the getrandom() system call is not available,
1938 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1939 Contrary to getrandom(), the /dev/urandom device will not block during
1940 early boot when the kernel CSPRNG has not been seeded yet.
1942 To mitigate this known weakness, use select() to wait for /dev/random to
1943 become readable before reading from /dev/urandom.
1945 * Ensure that SM2 only uses SM3 as digest algorithm
1949 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1951 * Change the info callback signals for the start and end of a post-handshake
1952 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1953 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1954 confused by this and assume that a TLSv1.2 renegotiation has started. This
1955 can break KeyUpdate handling. Instead we no longer signal the start and end
1956 of a post handshake message exchange (although the messages themselves are
1957 still signalled). This could break some applications that were expecting
1958 the old signals. However without this KeyUpdate is not usable for many
1963 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1965 * Timing vulnerability in DSA signature generation
1967 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1968 timing side channel attack. An attacker could use variations in the signing
1969 algorithm to recover the private key.
1971 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1976 * Timing vulnerability in ECDSA signature generation
1978 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1979 timing side channel attack. An attacker could use variations in the signing
1980 algorithm to recover the private key.
1982 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1987 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1988 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1989 of two gigabytes and the error handling improved.
1991 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1992 categorized as a normal bug, not a security issue, because the DRBG reseeds
1993 automatically and is fully functional even without additional randomness
1994 provided by the application.
1996 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1998 * Add a new ClientHello callback. Provides a callback interface that gives
1999 the application the ability to adjust the nascent SSL object at the
2000 earliest stage of ClientHello processing, immediately after extensions have
2001 been collected but before they have been processed. In particular, this
2002 callback can adjust the supported TLS versions in response to the contents
2007 * Add SM2 base algorithm support.
2011 * s390x assembly pack: add (improved) hardware-support for the following
2012 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2013 aes-cfb/cfb8, aes-ecb.
2017 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2018 parameter is no longer accepted, as it leads to a corrupt table. NULL
2019 pem_str is reserved for alias entries only.
2023 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2024 step for prime curves. The new implementation is based on formulae from
2025 differential addition-and-doubling in homogeneous projective coordinates
2026 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2027 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2028 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2029 to work in projective coordinates.
2031 *Billy Bob Brumley, Nicola Tuveri*
2033 * Change generating and checking of primes so that the error rate of not
2034 being prime depends on the intended use based on the size of the input.
2035 For larger primes this will result in more rounds of Miller-Rabin.
2036 The maximal error rate for primes with more than 1080 bits is lowered
2039 *Kurt Roeckx, Annie Yousar*
2041 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2045 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2046 moving between systems, and to avoid confusion when a Windows build is
2047 done with mingw vs with MSVC. For POSIX installs, there's still a
2048 symlink or copy named 'tsget' to avoid that confusion as well.
2052 * Revert blinding in ECDSA sign and instead make problematic addition
2053 length-invariant. Switch even to fixed-length Montgomery multiplication.
2057 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2058 step for binary curves. The new implementation is based on formulae from
2059 differential addition-and-doubling in mixed Lopez-Dahab projective
2060 coordinates, modified to independently blind the operands.
2062 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2064 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2065 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2066 EC_METHODs to implement their own specialized "ladder step", to take
2067 advantage of more favorable coordinate systems or more efficient
2068 differential addition-and-doubling algorithms.
2070 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2072 * Modified the random device based seed sources to keep the relevant
2073 file descriptors open rather than reopening them on each access.
2074 This allows such sources to operate in a chroot() jail without
2075 the associated device nodes being available. This behaviour can be
2076 controlled using RAND_keep_random_devices_open().
2080 * Numerous side-channel attack mitigations have been applied. This may have
2081 performance impacts for some algorithms for the benefit of improved
2082 security. Specific changes are noted in this change log by their respective
2087 * AIX shared library support overhaul. Switch to AIX "natural" way of
2088 handling shared libraries, which means collecting shared objects of
2089 different versions and bitnesses in one common archive. This allows to
2090 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2091 doesn't affect the way 3rd party applications are linked, only how
2092 multi-version installation is managed.
2096 * Make ec_group_do_inverse_ord() more robust and available to other
2097 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2098 mitigations are applied to the fallback BN_mod_inverse().
2099 When using this function rather than BN_mod_inverse() directly, new
2100 EC cryptosystem implementations are then safer-by-default.
2104 * Add coordinate blinding for EC_POINT and implement projective
2105 coordinate blinding for generic prime curves as a countermeasure to
2106 chosen point SCA attacks.
2108 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2110 * Add blinding to ECDSA and DSA signatures to protect against side channel
2111 attacks discovered by Keegan Ryan (NCC Group).
2115 * Enforce checking in the `pkeyutl` command to ensure that the input
2116 length does not exceed the maximum supported digest length when performing
2117 a sign, verify or verifyrecover operation.
2121 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2122 I/O in combination with something like select() or poll() will hang. This
2123 can be turned off again using SSL_CTX_clear_mode().
2124 Many applications do not properly handle non-application data records, and
2125 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2126 around the problems in those applications, but can also break some.
2127 It's recommended to read the manpages about SSL_read(), SSL_write(),
2128 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2129 SSL_CTX_set_read_ahead() again.
2133 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2134 now allow empty (zero character) pass phrases.
2138 * Apply blinding to binary field modular inversion and remove patent
2139 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2143 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2144 binary and prime elliptic curves.
2148 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2149 constant time fixed point multiplication.
2153 * Revise elliptic curve scalar multiplication with timing attack
2154 defenses: ec_wNAF_mul redirects to a constant time implementation
2155 when computing fixed point and variable point multiplication (which
2156 in OpenSSL are mostly used with secret scalars in keygen, sign,
2157 ECDH derive operations).
2158 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2161 * Updated CONTRIBUTING
2165 * Updated DRBG / RAND to request nonce and additional low entropy
2166 randomness from the system.
2168 *Matthias St. Pierre*
2170 * Updated 'openssl rehash' to use OpenSSL consistent default.
2174 * Moved the load of the ssl_conf module to libcrypto, which helps
2175 loading engines that libssl uses before libssl is initialised.
2179 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2183 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2185 *Ingo Schwarze, Rich Salz*
2187 * Added output of accepting IP address and port for 'openssl s_server'
2191 * Added a new API for TLSv1.3 ciphersuites:
2192 SSL_CTX_set_ciphersuites()
2193 SSL_set_ciphersuites()
2197 * Memory allocation failures consistently add an error to the error
2202 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2203 in libcrypto when run as setuid/setgid.
2207 * Load any config file by default when libssl is used.
2211 * Added new public header file <openssl/rand_drbg.h> and documentation
2212 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2214 *Matthias St. Pierre*
2216 * QNX support removed (cannot find contributors to get their approval
2217 for the license change).
2221 * TLSv1.3 replay protection for early data has been implemented. See the
2222 SSL_read_early_data() man page for further details.
2226 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2227 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2228 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2229 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2230 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2231 configuration has been separated out. See the ciphers man page or the
2232 SSL_CTX_set_ciphersuites() man page for more information.
2236 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2237 in responder mode now supports the new "-multi" option, which
2238 spawns the specified number of child processes to handle OCSP
2239 requests. The "-timeout" option now also limits the OCSP
2240 responder's patience to wait to receive the full client request
2241 on a newly accepted connection. Child processes are respawned
2242 as needed, and the CA index file is automatically reloaded
2243 when changed. This makes it possible to run the "ocsp" responder
2244 as a long-running service, making the OpenSSL CA somewhat more
2245 feature-complete. In this mode, most diagnostic messages logged
2246 after entering the event loop are logged via syslog(3) rather than
2251 * Added support for X448 and Ed448. Heavily based on original work by
2256 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2257 objects loaded. This adds the functions OSSL_STORE_expect() and
2258 OSSL_STORE_find() as well as needed tools to construct searches and
2259 get the search data out of them.
2263 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2264 version of OpenSSL should review their configuration settings to ensure
2265 that they are still appropriate for TLSv1.3. For further information see:
2266 <https://wiki.openssl.org/index.php/TLS1.3>
2270 * Grand redesign of the OpenSSL random generator
2272 The default RAND method now utilizes an AES-CTR DRBG according to
2273 NIST standard SP 800-90Ar1. The new random generator is essentially
2274 a port of the default random generator from the OpenSSL FIPS 2.0
2275 object module. It is a hybrid deterministic random bit generator
2276 using an AES-CTR bit stream and which seeds and reseeds itself
2277 automatically using trusted system entropy sources.
2279 Some of its new features are:
2280 - Support for multiple DRBG instances with seed chaining.
2281 - The default RAND method makes use of a DRBG.
2282 - There is a public and private DRBG instance.
2283 - The DRBG instances are fork-safe.
2284 - Keep all global DRBG instances on the secure heap if it is enabled.
2285 - The public and private DRBG instance are per thread for lock free
2288 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2290 * Changed Configure so it only says what it does and doesn't dump
2291 so much data. Instead, ./configdata.pm should be used as a script
2292 to display all sorts of configuration data.
2296 * Added processing of "make variables" to Configure.
2300 * Added SHA512/224 and SHA512/256 algorithm support.
2304 * The last traces of Netware support, first removed in 1.1.0, have
2309 * Get rid of Makefile.shared, and in the process, make the processing
2310 of certain files (rc.obj, or the .def/.map/.opt files produced from
2311 the ordinal files) more visible and hopefully easier to trace and
2312 debug (or make silent).
2316 * Make it possible to have environment variable assignments as
2317 arguments to config / Configure.
2321 * Add multi-prime RSA (RFC 8017) support.
2325 * Add SM3 implemented according to GB/T 32905-2016
2326 *Jack Lloyd <jack.lloyd@ribose.com>,*
2327 *Ronald Tse <ronald.tse@ribose.com>,*
2328 *Erick Borsboom <erick.borsboom@ribose.com>*
2330 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2331 as documented in RFC6066.
2332 Based on a patch from Tomasz Moń
2334 *Filipe Raimundo da Silva*
2336 * Add SM4 implemented according to GB/T 32907-2016.
2337 *Jack Lloyd <jack.lloyd@ribose.com>,*
2338 *Ronald Tse <ronald.tse@ribose.com>,*
2339 *Erick Borsboom <erick.borsboom@ribose.com>*
2341 * Reimplement -newreq-nodes and ERR_error_string_n; the
2342 original author does not agree with the license change.
2346 * Add ARIA AEAD TLS support.
2350 * Some macro definitions to support VS6 have been removed. Visual
2351 Studio 6 has not worked since 1.1.0
2355 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2356 without clearing the errors.
2360 * Add "atfork" functions. If building on a system that without
2361 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2362 requirements. The RAND facility now uses/requires this.
2370 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2371 not possible to disable entirely. However, it's still possible to
2372 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2375 To disable, configure with 'no-ui-console'. 'no-ui' is still
2376 possible to use as an alias. Check at compile time with the
2377 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2378 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2382 * Add a STORE module, which implements a uniform and URI based reader of
2383 stores that can contain keys, certificates, CRLs and numerous other
2384 objects. The main API is loosely based on a few stdio functions,
2385 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2386 OSSL_STORE_error and OSSL_STORE_close.
2387 The implementation uses backends called "loaders" to implement arbitrary
2388 URI schemes. There is one built in "loader" for the 'file' scheme.
2392 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2393 then adjusted to work on FreeBSD 8.4 as well.
2394 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2395 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2399 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2400 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2401 error code calls like this:
2403 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2405 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2406 that can be encoded in C. For the foreseeable future, this will only
2409 *Richard Levitte and Tim Hudson*
2411 * Removed BSD cryptodev engine.
2415 * Add a build target 'build_all_generated', to build all generated files
2416 and only that. This can be used to prepare everything that requires
2417 things like perl for a system that lacks perl and then move everything
2418 to that system and do the rest of the build there.
2422 * In the UI interface, make it possible to duplicate the user data. This
2423 can be used by engines that need to retain the data for a longer time
2424 than just the call where this user data is passed.
2428 * Ignore the '-named_curve auto' value for compatibility of applications
2431 *Tomáš Mráz <tmraz@fedoraproject.org>*
2433 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2434 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2435 alerts across multiple records (some of which could be empty). In practice
2436 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2437 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2438 support this at all. Supporting it adds significant complexity to the
2439 record layer, and its removal is unlikely to cause interoperability
2444 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2445 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2446 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2451 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2452 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2454 *Richard Levitte, Andy Polyakov*
2456 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2461 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2462 platform rather than 'mingw'.
2466 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2467 success if they are asked to add an object which already exists
2468 in the store. This change cascades to other functions which load
2469 certificates and CRLs.
2473 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2474 facilitate stack unwinding even from assembly subroutines.
2478 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2479 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2483 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2484 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2485 which is the minimum version we support.
2489 * Certificate time validation (X509_cmp_time) enforces stricter
2490 compliance with RFC 5280. Fractional seconds and timezone offsets
2491 are no longer allowed.
2495 * Add support for ARIA
2499 * s_client will now send the Server Name Indication (SNI) extension by
2500 default unless the new "-noservername" option is used. The server name is
2501 based on the host provided to the "-connect" option unless overridden by
2502 using "-servername".
2506 * Add support for SipHash
2510 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2511 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2512 prevent issues where no progress is being made and the peer continually
2513 sends unrecognised record types, using up resources processing them.
2517 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2518 using the algorithm defined in
2519 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2523 * Heartbeat support has been removed; the ABI is changed for now.
2525 *Richard Levitte, Rich Salz*
2527 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2531 * The RSA "null" method, which was partially supported to avoid patent
2532 issues, has been replaced to always returns NULL.
2539 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2541 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2542 used even when parsing explicit parameters, when loading a encoded key
2543 or calling `EC_GROUP_new_from_ecpkparameters()`/
2544 `EC_GROUP_new_from_ecparameters()`.
2545 This prevents bypass of security hardening and performance gains,
2546 especially for curves with specialized EC_METHODs.
2547 By default, if a key encoded with explicit parameters is loaded and later
2548 encoded, the output is still encoded with explicit parameters, even if
2549 internally a "named" EC_GROUP is used for computation.
2553 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2554 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2555 NULL. After this change, only the cofactor parameter can be NULL. It also
2556 does some minimal sanity checks on the passed order.
2561 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2562 An attack is simple, if the first CMS_recipientInfo is valid but the
2563 second CMS_recipientInfo is chosen ciphertext. If the second
2564 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2565 encryption key will be replaced by garbage, and the message cannot be
2566 decoded, but if the RSA decryption fails, the correct encryption key is
2567 used and the recipient will not notice the attack.
2568 As a work around for this potential attack the length of the decrypted
2569 key must be equal to the cipher default key length, in case the
2570 certifiate is not given and all recipientInfo are tried out.
2571 The old behaviour can be re-enabled in the CMS code by setting the
2572 CMS_DEBUG_DECRYPT flag.
2577 * Use Windows installation paths in the mingw builds
2579 Mingw isn't a POSIX environment per se, which means that Windows
2580 paths should be used for installation.
2585 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2587 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2588 This changes the size when using the `genpkey` command when no size is given.
2589 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2590 generation commands to use 2048 bits by default.
2594 * Prevent over long nonces in ChaCha20-Poly1305.
2596 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2597 for every encryption operation. RFC 7539 specifies that the nonce value
2598 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2599 and front pads the nonce with 0 bytes if it is less than 12
2600 bytes. However it also incorrectly allows a nonce to be set of up to 16
2601 bytes. In this case only the last 12 bytes are significant and any
2602 additional leading bytes are ignored.
2604 It is a requirement of using this cipher that nonce values are
2605 unique. Messages encrypted using a reused nonce value are susceptible to
2606 serious confidentiality and integrity attacks. If an application changes
2607 the default nonce length to be longer than 12 bytes and then makes a
2608 change to the leading bytes of the nonce expecting the new value to be a
2609 new unique nonce then such an application could inadvertently encrypt
2610 messages with a reused nonce.
2612 Additionally the ignored bytes in a long nonce are not covered by the
2613 integrity guarantee of this cipher. Any application that relies on the
2614 integrity of these ignored leading bytes of a long nonce may be further
2615 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2616 is safe because no such use sets such a long nonce value. However user
2617 applications that use this cipher directly and set a non-default nonce
2618 length to be longer than 12 bytes may be vulnerable.
2620 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2626 * Added SCA hardening for modular field inversion in EC_GROUP through
2627 a new dedicated field_inv() pointer in EC_METHOD.
2628 This also addresses a leakage affecting conversions from projective
2629 to affine coordinates.
2631 *Billy Bob Brumley, Nicola Tuveri*
2633 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2634 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2638 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2642 * Remove the 'dist' target and add a tarball building script. The
2643 'dist' target has fallen out of use, and it shouldn't be
2644 necessary to configure just to create a source distribution.
2648 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2650 * Timing vulnerability in DSA signature generation
2652 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2653 timing side channel attack. An attacker could use variations in the signing
2654 algorithm to recover the private key.
2656 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2661 * Timing vulnerability in ECDSA signature generation
2663 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2664 timing side channel attack. An attacker could use variations in the signing
2665 algorithm to recover the private key.
2667 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2672 * Add coordinate blinding for EC_POINT and implement projective
2673 coordinate blinding for generic prime curves as a countermeasure to
2674 chosen point SCA attacks.
2676 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2678 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2680 * Client DoS due to large DH parameter
2682 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2683 malicious server can send a very large prime value to the client. This will
2684 cause the client to spend an unreasonably long period of time generating a
2685 key for this prime resulting in a hang until the client has finished. This
2686 could be exploited in a Denial Of Service attack.
2688 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2693 * Cache timing vulnerability in RSA Key Generation
2695 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2696 a cache timing side channel attack. An attacker with sufficient access to
2697 mount cache timing attacks during the RSA key generation process could
2698 recover the private key.
2700 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2701 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2706 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2707 parameter is no longer accepted, as it leads to a corrupt table. NULL
2708 pem_str is reserved for alias entries only.
2712 * Revert blinding in ECDSA sign and instead make problematic addition
2713 length-invariant. Switch even to fixed-length Montgomery multiplication.
2717 * Change generating and checking of primes so that the error rate of not
2718 being prime depends on the intended use based on the size of the input.
2719 For larger primes this will result in more rounds of Miller-Rabin.
2720 The maximal error rate for primes with more than 1080 bits is lowered
2723 *Kurt Roeckx, Annie Yousar*
2725 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2729 * Add blinding to ECDSA and DSA signatures to protect against side channel
2730 attacks discovered by Keegan Ryan (NCC Group).
2734 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2735 now allow empty (zero character) pass phrases.
2739 * Certificate time validation (X509_cmp_time) enforces stricter
2740 compliance with RFC 5280. Fractional seconds and timezone offsets
2741 are no longer allowed.
2745 * Fixed a text canonicalisation bug in CMS
2747 Where a CMS detached signature is used with text content the text goes
2748 through a canonicalisation process first prior to signing or verifying a
2749 signature. This process strips trailing space at the end of lines, converts
2750 line terminators to CRLF and removes additional trailing line terminators
2751 at the end of a file. A bug in the canonicalisation process meant that
2752 some characters, such as form-feed, were incorrectly treated as whitespace
2753 and removed. This is contrary to the specification (RFC5485). This fix
2754 could mean that detached text data signed with an earlier version of
2755 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2756 signed with a fixed OpenSSL may fail to verify with an earlier version of
2757 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2758 and use the "-binary" flag (for the "cms" command line application) or set
2759 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2763 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2765 * Constructed ASN.1 types with a recursive definition could exceed the stack
2767 Constructed ASN.1 types with a recursive definition (such as can be found
2768 in PKCS7) could eventually exceed the stack given malicious input with
2769 excessive recursion. This could result in a Denial Of Service attack. There
2770 are no such structures used within SSL/TLS that come from untrusted sources
2771 so this is considered safe.
2773 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2779 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2781 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2782 effectively reduced to only comparing the least significant bit of each
2783 byte. This allows an attacker to forge messages that would be considered as
2784 authenticated in an amount of tries lower than that guaranteed by the
2785 security claims of the scheme. The module can only be compiled by the
2786 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2788 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2794 * Add a build target 'build_all_generated', to build all generated files
2795 and only that. This can be used to prepare everything that requires
2796 things like perl for a system that lacks perl and then move everything
2797 to that system and do the rest of the build there.
2801 * Backport SSL_OP_NO_RENGOTIATION
2803 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2804 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2805 changes this is no longer possible in 1.1.0. Therefore the new
2806 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2807 1.1.0 to provide equivalent functionality.
2809 Note that if an application built against 1.1.0h headers (or above) is run
2810 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2811 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2815 * Removed the OS390-Unix config target. It relied on a script that doesn't
2820 * rsaz_1024_mul_avx2 overflow bug on x86_64
2822 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2823 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2824 Analysis suggests that attacks against RSA and DSA as a result of this
2825 defect would be very difficult to perform and are not believed likely.
2826 Attacks against DH1024 are considered just feasible, because most of the
2827 work necessary to deduce information about a private key may be performed
2828 offline. The amount of resources required for such an attack would be
2829 significant. However, for an attack on TLS to be meaningful, the server
2830 would have to share the DH1024 private key among multiple clients, which is
2831 no longer an option since CVE-2016-0701.
2833 This only affects processors that support the AVX2 but not ADX extensions
2834 like Intel Haswell (4th generation).
2836 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2837 was originally found via the OSS-Fuzz project.
2842 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2844 * bn_sqrx8x_internal carry bug on x86_64
2846 There is a carry propagating bug in the x86_64 Montgomery squaring
2847 procedure. No EC algorithms are affected. Analysis suggests that attacks
2848 against RSA and DSA as a result of this defect would be very difficult to
2849 perform and are not believed likely. Attacks against DH are considered just
2850 feasible (although very difficult) because most of the work necessary to
2851 deduce information about a private key may be performed offline. The amount
2852 of resources required for such an attack would be very significant and
2853 likely only accessible to a limited number of attackers. An attacker would
2854 additionally need online access to an unpatched system using the target
2855 private key in a scenario with persistent DH parameters and a private
2856 key that is shared between multiple clients.
2858 This only affects processors that support the BMI1, BMI2 and ADX extensions
2859 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2861 This issue was reported to OpenSSL by the OSS-Fuzz project.
2866 * Malformed X.509 IPAddressFamily could cause OOB read
2868 If an X.509 certificate has a malformed IPAddressFamily extension,
2869 OpenSSL could do a one-byte buffer overread. The most likely result
2870 would be an erroneous display of the certificate in text format.
2872 This issue was reported to OpenSSL by the OSS-Fuzz project.
2877 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2879 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2880 platform rather than 'mingw'.
2884 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2885 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2886 which is the minimum version we support.
2890 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2892 * Encrypt-Then-Mac renegotiation crash
2894 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2895 negotiated where it was not in the original handshake (or vice-versa) then
2896 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2897 and servers are affected.
2899 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2904 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2906 * Truncated packet could crash via OOB read
2908 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2909 cipher is being used, then a truncated packet can cause that host to
2910 perform an out-of-bounds read, usually resulting in a crash.
2912 This issue was reported to OpenSSL by Robert Święcki of Google.
2917 * Bad (EC)DHE parameters cause a client crash
2919 If a malicious server supplies bad parameters for a DHE or ECDHE key
2920 exchange then this can result in the client attempting to dereference a
2921 NULL pointer leading to a client crash. This could be exploited in a Denial
2924 This issue was reported to OpenSSL by Guido Vranken.
2929 * BN_mod_exp may produce incorrect results on x86_64
2931 There is a carry propagating bug in the x86_64 Montgomery squaring
2932 procedure. No EC algorithms are affected. Analysis suggests that attacks
2933 against RSA and DSA as a result of this defect would be very difficult to
2934 perform and are not believed likely. Attacks against DH are considered just
2935 feasible (although very difficult) because most of the work necessary to
2936 deduce information about a private key may be performed offline. The amount
2937 of resources required for such an attack would be very significant and
2938 likely only accessible to a limited number of attackers. An attacker would
2939 additionally need online access to an unpatched system using the target
2940 private key in a scenario with persistent DH parameters and a private
2941 key that is shared between multiple clients. For example this can occur by
2942 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2943 similar to CVE-2015-3193 but must be treated as a separate problem.
2945 This issue was reported to OpenSSL by the OSS-Fuzz project.
2950 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2952 * ChaCha20/Poly1305 heap-buffer-overflow
2954 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2955 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2956 crash. This issue is not considered to be exploitable beyond a DoS.
2958 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2963 * CMS Null dereference
2965 Applications parsing invalid CMS structures can crash with a NULL pointer
2966 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2967 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2968 structure callback if an attempt is made to free certain invalid encodings.
2969 Only CHOICE structures using a callback which do not handle NULL value are
2972 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2977 * Montgomery multiplication may produce incorrect results
2979 There is a carry propagating bug in the Broadwell-specific Montgomery
2980 multiplication procedure that handles input lengths divisible by, but
2981 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2982 and DH private keys are impossible. This is because the subroutine in
2983 question is not used in operations with the private key itself and an input
2984 of the attacker's direct choice. Otherwise the bug can manifest itself as
2985 transient authentication and key negotiation failures or reproducible
2986 erroneous outcome of public-key operations with specially crafted input.
2987 Among EC algorithms only Brainpool P-512 curves are affected and one
2988 presumably can attack ECDH key negotiation. Impact was not analyzed in
2989 detail, because pre-requisites for attack are considered unlikely. Namely
2990 multiple clients have to choose the curve in question and the server has to
2991 share the private key among them, neither of which is default behaviour.
2992 Even then only clients that chose the curve will be affected.
2994 This issue was publicly reported as transient failures and was not
2995 initially recognized as a security issue. Thanks to Richard Morgan for
2996 providing reproducible case.
3001 * Removed automatic addition of RPATH in shared libraries and executables,
3002 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3006 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3008 * Fix Use After Free for large message sizes
3010 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3011 message larger than approx 16k is received then the underlying buffer to
3012 store the incoming message is reallocated and moved. Unfortunately a
3013 dangling pointer to the old location is left which results in an attempt to
3014 write to the previously freed location. This is likely to result in a
3015 crash, however it could potentially lead to execution of arbitrary code.
3017 This issue only affects OpenSSL 1.1.0a.
3019 This issue was reported to OpenSSL by Robert Święcki.
3024 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3026 * OCSP Status Request extension unbounded memory growth
3028 A malicious client can send an excessively large OCSP Status Request
3029 extension. If that client continually requests renegotiation, sending a
3030 large OCSP Status Request extension each time, then there will be unbounded
3031 memory growth on the server. This will eventually lead to a Denial Of
3032 Service attack through memory exhaustion. Servers with a default
3033 configuration are vulnerable even if they do not support OCSP. Builds using
3034 the "no-ocsp" build time option are not affected.
3036 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3041 * SSL_peek() hang on empty record
3043 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3044 sends an empty record. This could be exploited by a malicious peer in a
3045 Denial Of Service attack.
3047 This issue was reported to OpenSSL by Alex Gaynor.
3052 * Excessive allocation of memory in tls_get_message_header() and
3053 dtls1_preprocess_fragment()
3055 A (D)TLS message includes 3 bytes for its length in the header for the
3056 message. This would allow for messages up to 16Mb in length. Messages of
3057 this length are excessive and OpenSSL includes a check to ensure that a
3058 peer is sending reasonably sized messages in order to avoid too much memory
3059 being consumed to service a connection. A flaw in the logic of version
3060 1.1.0 means that memory for the message is allocated too early, prior to
3061 the excessive message length check. Due to way memory is allocated in
3062 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3063 to service a connection. This could lead to a Denial of Service through
3064 memory exhaustion. However, the excessive message length check still takes
3065 place, and this would cause the connection to immediately fail. Assuming
3066 that the application calls SSL_free() on the failed connection in a timely
3067 manner then the 21Mb of allocated memory will then be immediately freed
3068 again. Therefore the excessive memory allocation will be transitory in
3069 nature. This then means that there is only a security impact if:
3071 1) The application does not call SSL_free() in a timely manner in the event
3072 that the connection fails
3074 2) The application is working in a constrained environment where there is
3075 very little free memory
3077 3) The attacker initiates multiple connection attempts such that there are
3078 multiple connections in a state where memory has been allocated for the
3079 connection; SSL_free() has not yet been called; and there is insufficient
3080 memory to service the multiple requests.
3082 Except in the instance of (1) above any Denial Of Service is likely to be
3083 transitory because as soon as the connection fails the memory is
3084 subsequently freed again in the SSL_free() call. However there is an
3085 increased risk during this period of application crashes due to the lack of
3086 memory - which would then mean a more serious Denial of Service.
3088 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3089 (CVE-2016-6307 and CVE-2016-6308)
3093 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3094 had to be removed. Primary reason is that vendor assembler can't
3095 assemble our modules with -KPIC flag. As result it, assembly
3096 support, was not even available as option. But its lack means
3097 lack of side-channel resistant code, which is incompatible with
3098 security by todays standards. Fortunately gcc is readily available
3099 prepackaged option, which we firmly point at...
3103 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3105 * Windows command-line tool supports UTF-8 opt-in option for arguments
3106 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3107 (to any value) allows Windows user to access PKCS#12 file generated
3108 with Windows CryptoAPI and protected with non-ASCII password, as well
3109 as files generated under UTF-8 locale on Linux also protected with
3114 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3115 have been disabled by default and removed from DEFAULT, just like RC4.
3116 See the RC4 item below to re-enable both.
3120 * The method for finding the storage location for the Windows RAND seed file
3121 has changed. First we check %RANDFILE%. If that is not set then we check
3122 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3123 all else fails we fall back to C:\.
3127 * The EVP_EncryptUpdate() function has had its return type changed from void
3128 to int. A return of 0 indicates and error while a return of 1 indicates
3133 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3134 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3135 off the constant time implementation for RSA, DSA and DH have been made
3136 no-ops and deprecated.
3140 * Windows RAND implementation was simplified to only get entropy by
3141 calling CryptGenRandom(). Various other RAND-related tickets
3144 *Joseph Wylie Yandle, Rich Salz*
3146 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3147 and `OPENSSL_LH_`, respectively. The old names are available
3148 with API compatibility. They new names are now completely documented.
3152 * Unify TYPE_up_ref(obj) methods signature.
3153 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3154 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3155 int (instead of void) like all others TYPE_up_ref() methods.
3156 So now these methods also check the return value of CRYPTO_atomic_add(),
3157 and the validity of object reference counter.
3159 *fdasilvayy@gmail.com*
3161 * With Windows Visual Studio builds, the .pdb files are installed
3162 alongside the installed libraries and executables. For a static
3163 library installation, ossl_static.pdb is the associate compiler
3164 generated .pdb file to be used when linking programs.
3168 * Remove openssl.spec. Packaging files belong with the packagers.
3172 * Automatic Darwin/OSX configuration has had a refresh, it will now
3173 recognise x86_64 architectures automatically. You can still decide
3174 to build for a different bitness with the environment variable
3175 KERNEL_BITS (can be 32 or 64), for example:
3177 KERNEL_BITS=32 ./config
3181 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3182 256 bit AES and HMAC with SHA256.
3186 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3190 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3194 * To enable users to have their own config files and build file templates,
3195 Configure looks in the directory indicated by the environment variable
3196 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3197 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3198 name and is used as is.
3202 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3203 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3204 X509_CERT_FILE_CTX was removed.
3208 * "shared" builds are now the default. To create only static libraries use
3209 the "no-shared" Configure option.
3213 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3214 All of these option have not worked for some while and are fundamental
3219 * Make various cleanup routines no-ops and mark them as deprecated. Most
3220 global cleanup functions are no longer required because they are handled
3221 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3222 Explicitly de-initing can cause problems (e.g. where a library that uses
3223 OpenSSL de-inits, but an application is still using it). The affected
3224 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3225 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3226 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3227 COMP_zlib_cleanup().
3231 * --strict-warnings no longer enables runtime debugging options
3232 such as REF_DEBUG. Instead, debug options are automatically
3233 enabled with '--debug' builds.
3235 *Andy Polyakov, Emilia Käsper*
3237 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3238 have been moved out of the public header files. New functions for managing
3239 these have been added.
3243 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3244 objects have been moved out of the public header files. New
3245 functions for managing these have been added.
3249 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3250 have been moved out of the public header files. New functions for managing
3251 these have been added.
3255 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3256 moved out of the public header files. New functions for managing these
3261 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3265 * Removed the mk1mf build scripts.
3269 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3270 it is always safe to #include a header now.
3274 * Removed the aged BC-32 config and all its supporting scripts
3278 * Removed support for Ultrix, Netware, and OS/2.
3282 * Add support for HKDF.
3284 *Alessandro Ghedini*
3286 * Add support for blake2b and blake2s
3290 * Added support for "pipelining". Ciphers that have the
3291 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3292 encryptions/decryptions simultaneously. There are currently no built-in
3293 ciphers with this property but the expectation is that engines will be able
3294 to offer it to significantly improve throughput. Support has been extended
3295 into libssl so that multiple records for a single connection can be
3296 processed in one go (for >=TLS 1.1).
3300 * Added the AFALG engine. This is an async capable engine which is able to
3301 offload work to the Linux kernel. In this initial version it only supports
3302 AES128-CBC. The kernel must be version 4.1.0 or greater.
3306 * OpenSSL now uses a new threading API. It is no longer necessary to
3307 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3308 are two supported threading models: pthreads and windows threads. It is
3309 also possible to configure OpenSSL at compile time for "no-threads". The
3310 old threading API should no longer be used. The functions have been
3311 replaced with "no-op" compatibility macros.
3313 *Alessandro Ghedini, Matt Caswell*
3315 * Modify behavior of ALPN to invoke callback after SNI/servername
3316 callback, such that updates to the SSL_CTX affect ALPN.
3320 * Add SSL_CIPHER queries for authentication and key-exchange.
3324 * Changes to the DEFAULT cipherlist:
3325 - Prefer (EC)DHE handshakes over plain RSA.
3326 - Prefer AEAD ciphers over legacy ciphers.
3327 - Prefer ECDSA over RSA when both certificates are available.
3328 - Prefer TLSv1.2 ciphers/PRF.
3329 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3334 * Change the ECC default curve list to be this, in order: x25519,
3335 secp256r1, secp521r1, secp384r1.
3339 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3340 disabled by default. They can be re-enabled using the
3341 enable-weak-ssl-ciphers option to Configure.
3345 * If the server has ALPN configured, but supports no protocols that the
3346 client advertises, send a fatal "no_application_protocol" alert.
3347 This behaviour is SHALL in RFC 7301, though it isn't universally
3348 implemented by other servers.
3352 * Add X25519 support.
3353 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3354 for public and private key encoding using the format documented in
3355 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3356 key generation and key derivation.
3358 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3363 * Deprecate SRP_VBASE_get_by_user.
3364 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3365 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3366 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3367 seed, even if the seed is configured.
3369 Users should use SRP_VBASE_get1_by_user instead. Note that in
3370 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3371 also that even though configuring the SRP seed attempts to hide
3372 invalid usernames by continuing the handshake with fake
3373 credentials, this behaviour is not constant time and no strong
3374 guarantees are made that the handshake is indistinguishable from
3375 that of a valid user.
3379 * Configuration change; it's now possible to build dynamic engines
3380 without having to build shared libraries and vice versa. This
3381 only applies to the engines in `engines/`, those in `crypto/engine/`
3382 will always be built into libcrypto (i.e. "static").
3384 Building dynamic engines is enabled by default; to disable, use
3385 the configuration option "disable-dynamic-engine".
3387 The only requirements for building dynamic engines are the
3388 presence of the DSO module and building with position independent
3389 code, so they will also automatically be disabled if configuring
3390 with "disable-dso" or "disable-pic".
3392 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3393 are also taken away from openssl/opensslconf.h, as they are
3398 * Configuration change; if there is a known flag to compile
3399 position independent code, it will always be applied on the
3400 libcrypto and libssl object files, and never on the application
3401 object files. This means other libraries that use routines from
3402 libcrypto / libssl can be made into shared libraries regardless
3403 of how OpenSSL was configured.
3405 If this isn't desirable, the configuration options "disable-pic"
3406 or "no-pic" can be used to disable the use of PIC. This will
3407 also disable building shared libraries and dynamic engines.
3411 * Removed JPAKE code. It was experimental and has no wide use.
3415 * The INSTALL_PREFIX Makefile variable has been renamed to
3416 DESTDIR. That makes for less confusion on what this variable
3417 is for. Also, the configuration option --install_prefix is
3422 * Heartbeat for TLS has been removed and is disabled by default
3423 for DTLS; configure with enable-heartbeats. Code that uses the
3424 old #define's might need to be updated.
3426 *Emilia Käsper, Rich Salz*
3428 * Rename REF_CHECK to REF_DEBUG.
3432 * New "unified" build system
3434 The "unified" build system is aimed to be a common system for all
3435 platforms we support. With it comes new support for VMS.
3437 This system builds supports building in a different directory tree
3438 than the source tree. It produces one Makefile (for unix family
3439 or lookalikes), or one descrip.mms (for VMS).
3441 The source of information to make the Makefile / descrip.mms is
3442 small files called 'build.info', holding the necessary
3443 information for each directory with source to compile, and a
3444 template in Configurations, like unix-Makefile.tmpl or
3447 With this change, the library names were also renamed on Windows
3448 and on VMS. They now have names that are closer to the standard
3449 on Unix, and include the major version number, and in certain
3450 cases, the architecture they are built for. See "Notes on shared
3451 libraries" in INSTALL.
3453 We rely heavily on the perl module Text::Template.
3457 * Added support for auto-initialisation and de-initialisation of the library.
3458 OpenSSL no longer requires explicit init or deinit routines to be called,
3459 except in certain circumstances. See the OPENSSL_init_crypto() and
3460 OPENSSL_init_ssl() man pages for further information.
3464 * The arguments to the DTLSv1_listen function have changed. Specifically the
3465 "peer" argument is now expected to be a BIO_ADDR object.
3467 * Rewrite of BIO networking library. The BIO library lacked consistent
3468 support of IPv6, and adding it required some more extensive
3469 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3470 which hold all types of addresses and chains of address information.
3471 It also introduces a new API, with functions like BIO_socket,
3472 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3473 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3474 have been adapted accordingly.
3478 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3483 * CRIME protection: disable compression by default, even if OpenSSL is
3484 compiled with zlib enabled. Applications can still enable compression
3485 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3486 using the SSL_CONF library to configure compression.
3490 * The signature of the session callback configured with
3491 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3492 was explicitly marked as `const unsigned char*` instead of
3497 * Always DPURIFY. Remove the use of uninitialized memory in the
3498 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3502 * Removed many obsolete configuration items, including
3503 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3504 MD2_CHAR, MD2_INT, MD2_LONG
3506 IDEA_SHORT, IDEA_LONG
3507 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3509 *Rich Salz, with advice from Andy Polyakov*
3511 * Many BN internals have been moved to an internal header file.
3513 *Rich Salz with help from Andy Polyakov*
3515 * Configuration and writing out the results from it has changed.
3516 Files such as Makefile include/openssl/opensslconf.h and are now
3517 produced through general templates, such as Makefile.in and
3518 crypto/opensslconf.h.in and some help from the perl module
3521 Also, the center of configuration information is no longer
3522 Makefile. Instead, Configure produces a perl module in
3523 configdata.pm which holds most of the config data (in the hash
3524 table %config), the target data that comes from the target
3525 configuration in one of the `Configurations/*.conf` files (in
3530 * To clarify their intended purposes, the Configure options
3531 --prefix and --openssldir change their semantics, and become more
3532 straightforward and less interdependent.
3534 --prefix shall be used exclusively to give the location INSTALLTOP
3535 where programs, scripts, libraries, include files and manuals are
3536 going to be installed. The default is now /usr/local.
3538 --openssldir shall be used exclusively to give the default
3539 location OPENSSLDIR where certificates, private keys, CRLs are
3540 managed. This is also where the default openssl.cnf gets
3542 If the directory given with this option is a relative path, the
3543 values of both the --prefix value and the --openssldir value will
3544 be combined to become OPENSSLDIR.
3545 The default for --openssldir is INSTALLTOP/ssl.
3547 Anyone who uses --openssldir to specify where OpenSSL is to be
3548 installed MUST change to use --prefix instead.
3552 * The GOST engine was out of date and therefore it has been removed. An up
3553 to date GOST engine is now being maintained in an external repository.
3554 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3555 support for GOST ciphersuites (these are only activated if a GOST engine
3560 * EGD is no longer supported by default; use enable-egd when
3563 *Ben Kaduk and Rich Salz*
3565 * The distribution now has Makefile.in files, which are used to
3566 create Makefile's when Configure is run. *Configure must be run
3567 before trying to build now.*
3571 * The return value for SSL_CIPHER_description() for error conditions
3576 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3578 Obtaining and performing DNSSEC validation of TLSA records is
3579 the application's responsibility. The application provides
3580 the TLSA records of its choice to OpenSSL, and these are then
3581 used to authenticate the peer.
3583 The TLSA records need not even come from DNS. They can, for
3584 example, be used to implement local end-entity certificate or
3585 trust-anchor "pinning", where the "pin" data takes the form
3586 of TLSA records, which can augment or replace verification
3587 based on the usual WebPKI public certification authorities.
3591 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3592 continues to support deprecated interfaces in default builds.
3593 However, applications are strongly advised to compile their
3594 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3595 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3596 or the 1.1.0 releases.
3598 In environments in which all applications have been ported to
3599 not use any deprecated interfaces OpenSSL's Configure script
3600 should be used with the --api=1.1.0 option to entirely remove
3601 support for the deprecated features from the library and
3602 unconditionally disable them in the installed headers.
3603 Essentially the same effect can be achieved with the "no-deprecated"
3604 argument to Configure, except that this will always restrict
3605 the build to just the latest API, rather than a fixed API
3608 As applications are ported to future revisions of the API,
3609 they should update their compile-time OPENSSL_API_COMPAT define
3610 accordingly, but in most cases should be able to continue to
3611 compile with later releases.
3613 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3614 0x10000000L and 0x00908000L, respectively. However those
3615 versions did not support the OPENSSL_API_COMPAT feature, and
3616 so applications are not typically tested for explicit support
3617 of just the undeprecated features of either release.
3621 * Add support for setting the minimum and maximum supported protocol.
3622 It can bet set via the SSL_set_min_proto_version() and
3623 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3624 MaxProtocol. It's recommended to use the new APIs to disable
3625 protocols instead of disabling individual protocols using
3626 SSL_set_options() or SSL_CONF's Protocol. This change also
3627 removes support for disabling TLS 1.2 in the OpenSSL TLS
3628 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3632 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3636 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3637 and integrates ECDSA and ECDH functionality into EC. Implementations can
3638 now redirect key generation and no longer need to convert to or from
3641 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3642 include the ec.h header file instead.
3646 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3647 ciphers who are no longer supported and drops support the ephemeral RSA key
3648 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3652 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3653 opaque. For HMAC_CTX, the following constructors and destructors
3656 HMAC_CTX *HMAC_CTX_new(void);
3657 void HMAC_CTX_free(HMAC_CTX *ctx);
3659 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3660 destroy such methods has been added. See EVP_MD_meth_new(3) and
3661 EVP_CIPHER_meth_new(3) for documentation.
3664 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3665 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3666 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3667 an already created structure.
3668 2) For consistency with the majority of our object creators and
3669 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3670 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3671 for deprecated builds.
3675 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3676 cryptographic operations to be performed asynchronously as long as an
3677 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3678 further details. Libssl has also had this capability integrated with the
3679 introduction of the new mode SSL_MODE_ASYNC and associated error
3680 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3681 pages. This work was developed in partnership with Intel Corp.
3685 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3686 always enabled now. If you want to disable the support you should
3687 exclude it using the list of supported ciphers. This also means that the
3688 "-no_ecdhe" option has been removed from s_server.
3692 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3693 SSL_{CTX_}set1_curves() which can set a list.
3697 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3698 curve you want to support using SSL_{CTX_}set1_curves().
3702 * State machine rewrite. The state machine code has been significantly
3703 refactored in order to remove much duplication of code and solve issues
3704 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3705 further details). This change does have some associated API changes.
3706 Notably the SSL_state() function has been removed and replaced by
3707 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3708 SSL_set_state() has been removed altogether. The previous handshake states
3709 defined in ssl.h and ssl3.h have also been removed.
3713 * All instances of the string "ssleay" in the public API were replaced
3714 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3715 Some error codes related to internal RSA_eay API's were renamed.
3719 * The demo files in crypto/threads were moved to demo/threads.
3723 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3726 *Matt Caswell, Rich Salz*
3728 * New ASN.1 embed macro.
3730 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3731 structure is not allocated: it is part of the parent. That is instead of
3739 This reduces memory fragmentation and make it impossible to accidentally
3740 set a mandatory field to NULL.
3742 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3743 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3744 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3749 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3753 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3754 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3755 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3756 DES and RC4 ciphersuites.
3760 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3761 This changes the decoding behaviour for some invalid messages,
3762 though the change is mostly in the more lenient direction, and
3763 legacy behaviour is preserved as much as possible.
3767 * Fix no-stdio build.
3768 *David Woodhouse <David.Woodhouse@intel.com> and also*
3769 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3771 * New testing framework
3772 The testing framework has been largely rewritten and is now using
3773 perl and the perl modules Test::Harness and an extended variant of
3774 Test::More called OpenSSL::Test to do its work. All test scripts in
3775 test/ have been rewritten into test recipes, and all direct calls to
3776 executables in test/Makefile have become individual recipes using the
3777 simplified testing OpenSSL::Test::Simple.
3779 For documentation on our testing modules, do:
3781 perldoc test/testlib/OpenSSL/Test/Simple.pm
3782 perldoc test/testlib/OpenSSL/Test.pm
3786 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3787 are used; the latter aborts on memory leaks (usually checked on exit).
3788 Some undocumented "set malloc, etc., hooks" functions were removed
3789 and others were changed. All are now documented.
3793 * In DSA_generate_parameters_ex, if the provided seed is too short,
3796 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3798 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3799 from RFC4279, RFC4785, RFC5487, RFC5489.
3801 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3802 original RSA_PSK patch.
3806 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3807 era flag was never set throughout the codebase (only read). Also removed
3808 SSL3_FLAGS_POP_BUFFER which was only used if
3809 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3813 * Changed the default name options in the "ca", "crl", "req" and "x509"
3814 to be "oneline" instead of "compat".
3818 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3819 not aware of clients that still exhibit this bug, and the workaround
3820 hasn't been working properly for a while.
3824 * The return type of BIO_number_read() and BIO_number_written() as well as
3825 the corresponding num_read and num_write members in the BIO structure has
3826 changed from unsigned long to uint64_t. On platforms where an unsigned
3827 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3832 * Given the pervasive nature of TLS extensions it is inadvisable to run
3833 OpenSSL without support for them. It also means that maintaining
3834 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3835 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3839 * Removed support for the two export grade static DH ciphersuites
3840 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3841 were newly added (along with a number of other static DH ciphersuites) to
3842 1.0.2. However the two export ones have *never* worked since they were
3843 introduced. It seems strange in any case to be adding new export
3844 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3848 * Version negotiation has been rewritten. In particular SSLv23_method(),
3849 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3850 and turned into macros which simply call the new preferred function names
3851 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3852 should use the new names instead. Also as part of this change the ssl23.h
3853 header file has been removed.
3857 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3858 code and the associated standard is no longer considered fit-for-purpose.
3862 * RT2547 was closed. When generating a private key, try to make the
3863 output file readable only by the owner. This behavior change might
3864 be noticeable when interacting with other software.
3866 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3871 * Added HTTP GET support to the ocsp command.
3875 * Changed default digest for the dgst and enc commands from MD5 to
3880 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3884 * Added support for TLS extended master secret from
3885 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3886 initial patch which was a great help during development.
3890 * All libssl internal structures have been removed from the public header
3891 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3892 now redundant). Users should not attempt to access internal structures
3893 directly. Instead they should use the provided API functions.
3897 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3898 Access to deprecated functions can be re-enabled by running config with
3899 "enable-deprecated". In addition applications wishing to use deprecated
3900 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3901 will, by default, disable some transitive includes that previously existed
3902 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3906 * Added support for OCB mode. OpenSSL has been granted a patent license
3907 compatible with the OpenSSL license for use of OCB. Details are available
3908 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3909 for OCB can be removed by calling config with no-ocb.
3913 * SSLv2 support has been removed. It still supports receiving a SSLv2
3914 compatible client hello.
3918 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3919 done while fixing the error code for the key-too-small case.
3921 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3923 * CA.sh has been removed; use CA.pl instead.
3927 * Removed old DES API.
3931 * Remove various unsupported platforms:
3937 Sinix/ReliantUNIX RM400
3942 16-bit platforms such as WIN16
3946 * Clean up OPENSSL_NO_xxx #define's
3947 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3948 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3949 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3950 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3951 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3952 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3953 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3954 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3955 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3956 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3960 * Cleaned up dead code
3961 Remove all but one '#ifdef undef' which is to be looked at.
3965 * Clean up calling of xxx_free routines.
3966 Just like free(), fix most of the xxx_free routines to accept
3967 NULL. Remove the non-null checks from callers. Save much code.
3971 * Add secure heap for storage of private keys (when possible).
3972 Add BIO_s_secmem(), CBIGNUM, etc.
3973 Contributed by Akamai Technologies under our Corporate CLA.
3977 * Experimental support for a new, fast, unbiased prime candidate generator,
3978 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3980 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3982 * New output format NSS in the sess_id command line tool. This allows
3983 exporting the session id and the master key in NSS keylog format.
3985 *Martin Kaiser <martin@kaiser.cx>*
3987 * Harmonize version and its documentation. -f flag is used to display
3990 *mancha <mancha1@zoho.com>*
3992 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3993 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3995 *mancha <mancha1@zoho.com>*
3997 * Fix some double frees. These are not thought to be exploitable.
3999 *mancha <mancha1@zoho.com>*
4001 * A missing bounds check in the handling of the TLS heartbeat extension
4002 can be used to reveal up to 64k of memory to a connected client or
4005 Thanks for Neel Mehta of Google Security for discovering this bug and to
4006 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4007 preparing the fix ([CVE-2014-0160])
4009 *Adam Langley, Bodo Moeller*
4011 * Fix for the attack described in the paper "Recovering OpenSSL
4012 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4013 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4014 <http://eprint.iacr.org/2014/140>
4016 Thanks to Yuval Yarom and Naomi Benger for discovering this
4017 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4019 *Yuval Yarom and Naomi Benger*
4021 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4022 this fixes a limitation in previous versions of OpenSSL.
4026 * Experimental encrypt-then-mac support.
4028 Experimental support for encrypt then mac from
4029 draft-gutmann-tls-encrypt-then-mac-02.txt
4031 To enable it set the appropriate extension number (0x42 for the test
4032 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4034 For non-compliant peers (i.e. just about everything) this should have no
4037 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4041 * Add EVP support for key wrapping algorithms, to avoid problems with
4042 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4043 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4044 algorithms and include tests cases.
4048 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4053 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4054 MGF1 digest and OAEP label.
4058 * Make openssl verify return errors.
4060 *Chris Palmer <palmer@google.com> and Ben Laurie*
4062 * New function ASN1_TIME_diff to calculate the difference between two
4063 ASN1_TIME structures or one structure and the current time.
4067 * Update fips_test_suite to support multiple command line options. New
4068 test to induce all self test errors in sequence and check expected
4073 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4074 sign or verify all in one operation.
4078 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4079 test programs and fips_test_suite. Includes functionality to parse
4080 the minimal script output of fipsalgest.pl directly.
4084 * Add authorisation parameter to FIPS_module_mode_set().
4088 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4092 * Use separate DRBG fields for internal and external flags. New function
4093 FIPS_drbg_health_check() to perform on demand health checking. Add
4094 generation tests to fips_test_suite with reduced health check interval to
4095 demonstrate periodic health checking. Add "nodh" option to
4096 fips_test_suite to skip very slow DH test.
4100 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4105 * More extensive health check for DRBG checking many more failure modes.
4106 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4107 combination: call this in fips_test_suite.
4111 * Add support for canonical generation of DSA parameter 'g'. See
4114 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4115 POST to handle HMAC cases.
4119 * Add functions FIPS_module_version() and FIPS_module_version_text()
4120 to return numerical and string versions of the FIPS module number.
4124 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4125 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4126 outside the validated module in the FIPS capable OpenSSL.
4130 * Minor change to DRBG entropy callback semantics. In some cases
4131 there is no multiple of the block length between min_len and
4132 max_len. Allow the callback to return more than max_len bytes
4133 of entropy but discard any extra: it is the callback's responsibility
4134 to ensure that the extra data discarded does not impact the
4135 requested amount of entropy.
4139 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4140 information in FIPS186-3, SP800-57 and SP800-131A.
4144 * CCM support via EVP. Interface is very similar to GCM case except we
4145 must supply all data in one chunk (i.e. no update, final) and the
4146 message length must be supplied if AAD is used. Add algorithm test
4151 * Initial version of POST overhaul. Add POST callback to allow the status
4152 of POST to be monitored and/or failures induced. Modify fips_test_suite
4153 to use callback. Always run all selftests even if one fails.
4157 * XTS support including algorithm test driver in the fips_gcmtest program.
4158 Note: this does increase the maximum key length from 32 to 64 bytes but
4159 there should be no binary compatibility issues as existing applications
4160 will never use XTS mode.
4164 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4165 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4166 performs algorithm blocking for unapproved PRNG types. Also do not
4167 set PRNG type in FIPS_mode_set(): leave this to the application.
4168 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4169 the standard OpenSSL PRNG: set additional data to a date time vector.
4173 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4174 This shouldn't present any incompatibility problems because applications
4175 shouldn't be using these directly and any that are will need to rethink
4176 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4180 * Extensive self tests and health checking required by SP800-90 DRBG.
4181 Remove strength parameter from FIPS_drbg_instantiate and always
4182 instantiate at maximum supported strength.
4186 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4190 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4194 * New function DH_compute_key_padded() to compute a DH key and pad with
4195 leading zeroes if needed: this complies with SP800-56A et al.
4199 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4200 anything, incomplete, subject to change and largely untested at present.
4204 * Modify fipscanisteronly build option to only build the necessary object
4205 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4209 * Add experimental option FIPSSYMS to give all symbols in
4210 fipscanister.o and FIPS or fips prefix. This will avoid
4211 conflicts with future versions of OpenSSL. Add perl script
4212 util/fipsas.pl to preprocess assembly language source files
4213 and rename any affected symbols.
4217 * Add selftest checks and algorithm block of non-fips algorithms in
4218 FIPS mode. Remove DES2 from selftests.
4222 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4223 return internal method without any ENGINE dependencies. Add new
4224 tiny fips sign and verify functions.
4228 * New build option no-ec2m to disable characteristic 2 code.
4232 * New build option "fipscanisteronly". This only builds fipscanister.o
4233 and (currently) associated fips utilities. Uses the file Makefile.fips
4234 instead of Makefile.org as the prototype.
4238 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4239 Update fips_gcmtest to use IV generator.
4243 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4244 setting output buffer to NULL. The `*Final` function must be
4245 called although it will not retrieve any additional data. The tag
4246 can be set or retrieved with a ctrl. The IV length is by default 12
4247 bytes (96 bits) but can be set to an alternative value. If the IV
4248 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4253 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4254 underlying do_cipher function handles all cipher semantics itself
4255 including padding and finalisation. This is useful if (for example)
4256 an ENGINE cipher handles block padding itself. The behaviour of
4257 do_cipher is subtly changed if this flag is set: the return value
4258 is the number of characters written to the output buffer (zero is
4259 no longer an error code) or a negative error code. Also if the
4260 input buffer is NULL and length 0 finalisation should be performed.
4264 * If a candidate issuer certificate is already part of the constructed
4265 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4269 * Improve forward-security support: add functions
4271 void SSL_CTX_set_not_resumable_session_callback(
4272 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4273 void SSL_set_not_resumable_session_callback(
4274 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4276 for use by SSL/TLS servers; the callback function will be called whenever a
4277 new session is created, and gets to decide whether the session may be
4278 cached to make it resumable (return 0) or not (return 1). (As by the
4279 SSL/TLS protocol specifications, the session_id sent by the server will be
4280 empty to indicate that the session is not resumable; also, the server will
4281 not generate RFC 4507 (RFC 5077) session tickets.)
4283 A simple reasonable callback implementation is to return is_forward_secure.
4284 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4285 by the SSL/TLS server library, indicating whether it can provide forward
4288 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4290 * New -verify_name option in command line utilities to set verification
4295 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4296 Add CMAC pkey methods.
4300 * Experimental renegotiation in s_server -www mode. If the client
4301 browses /reneg connection is renegotiated. If /renegcert it is
4302 renegotiated requesting a certificate.
4306 * Add an "external" session cache for debugging purposes to s_server. This
4307 should help trace issues which normally are only apparent in deployed
4308 multi-process servers.
4312 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4313 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4314 BIO_set_cipher() and some obscure PEM functions were changed so they
4315 can now return an error. The RAND changes required a change to the
4316 RAND_METHOD structure.
4320 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4321 a gcc attribute to warn if the result of a function is ignored. This
4322 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4323 whose return value is often ignored.
4327 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4328 These allow SCTs (signed certificate timestamps) to be requested and
4329 validated when establishing a connection.
4331 *Rob Percival <robpercival@google.com>*
4336 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4338 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4339 used even when parsing explicit parameters, when loading a encoded key
4340 or calling `EC_GROUP_new_from_ecpkparameters()`/
4341 `EC_GROUP_new_from_ecparameters()`.
4342 This prevents bypass of security hardening and performance gains,
4343 especially for curves with specialized EC_METHODs.
4344 By default, if a key encoded with explicit parameters is loaded and later
4345 encoded, the output is still encoded with explicit parameters, even if
4346 internally a "named" EC_GROUP is used for computation.
4350 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4351 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4352 NULL. After this change, only the cofactor parameter can be NULL. It also
4353 does some minimal sanity checks on the passed order.
4358 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4359 An attack is simple, if the first CMS_recipientInfo is valid but the
4360 second CMS_recipientInfo is chosen ciphertext. If the second
4361 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4362 encryption key will be replaced by garbage, and the message cannot be
4363 decoded, but if the RSA decryption fails, the correct encryption key is
4364 used and the recipient will not notice the attack.
4365 As a work around for this potential attack the length of the decrypted
4366 key must be equal to the cipher default key length, in case the
4367 certifiate is not given and all recipientInfo are tried out.
4368 The old behaviour can be re-enabled in the CMS code by setting the
4369 CMS_DEBUG_DECRYPT flag.
4374 * Document issue with installation paths in diverse Windows builds
4376 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4377 binaries and run-time config file.
4382 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4384 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4385 This changes the size when using the `genpkey` command when no size is given.
4386 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4387 generation commands to use 2048 bits by default.
4391 * Add FIPS support for Android Arm 64-bit
4393 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4394 Module in Version 2.0.10. For some reason, the corresponding target
4395 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4396 built with FIPS support on Android Arm 64-bit. This omission has been
4399 *Matthias St. Pierre*
4401 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4403 * 0-byte record padding oracle
4405 If an application encounters a fatal protocol error and then calls
4406 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4407 then OpenSSL can respond differently to the calling application if a 0 byte
4408 record is received with invalid padding compared to if a 0 byte record is
4409 received with an invalid MAC. If the application then behaves differently
4410 based on that in a way that is detectable to the remote peer, then this
4411 amounts to a padding oracle that could be used to decrypt data.
4413 In order for this to be exploitable "non-stitched" ciphersuites must be in
4414 use. Stitched ciphersuites are optimised implementations of certain
4415 commonly used ciphersuites. Also the application must call SSL_shutdown()
4416 twice even if a protocol error has occurred (applications should not do
4417 this but some do anyway).
4419 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4420 Aviram, with additional investigation by Steven Collison and Andrew
4421 Hourselt. It was reported to OpenSSL on 10th December 2018.
4426 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4430 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4432 * Microarchitecture timing vulnerability in ECC scalar multiplication
4434 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4435 shown to be vulnerable to a microarchitecture timing side channel attack.
4436 An attacker with sufficient access to mount local timing attacks during
4437 ECDSA signature generation could recover the private key.
4439 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4440 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4446 * Timing vulnerability in DSA signature generation
4448 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4449 timing side channel attack. An attacker could use variations in the signing
4450 algorithm to recover the private key.
4452 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4457 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4458 Module, accidentally introduced while backporting security fixes from the
4459 development branch and hindering the use of ECC in FIPS mode.
4463 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4465 * Client DoS due to large DH parameter
4467 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4468 malicious server can send a very large prime value to the client. This will
4469 cause the client to spend an unreasonably long period of time generating a
4470 key for this prime resulting in a hang until the client has finished. This
4471 could be exploited in a Denial Of Service attack.
4473 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4478 * Cache timing vulnerability in RSA Key Generation
4480 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4481 a cache timing side channel attack. An attacker with sufficient access to
4482 mount cache timing attacks during the RSA key generation process could
4483 recover the private key.
4485 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4486 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4491 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4492 parameter is no longer accepted, as it leads to a corrupt table. NULL
4493 pem_str is reserved for alias entries only.
4497 * Revert blinding in ECDSA sign and instead make problematic addition
4498 length-invariant. Switch even to fixed-length Montgomery multiplication.
4502 * Change generating and checking of primes so that the error rate of not
4503 being prime depends on the intended use based on the size of the input.
4504 For larger primes this will result in more rounds of Miller-Rabin.
4505 The maximal error rate for primes with more than 1080 bits is lowered
4508 *Kurt Roeckx, Annie Yousar*
4510 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4514 * Add blinding to ECDSA and DSA signatures to protect against side channel
4515 attacks discovered by Keegan Ryan (NCC Group).
4519 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4520 now allow empty (zero character) pass phrases.
4524 * Certificate time validation (X509_cmp_time) enforces stricter
4525 compliance with RFC 5280. Fractional seconds and timezone offsets
4526 are no longer allowed.
4530 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4532 * Constructed ASN.1 types with a recursive definition could exceed the stack
4534 Constructed ASN.1 types with a recursive definition (such as can be found
4535 in PKCS7) could eventually exceed the stack given malicious input with
4536 excessive recursion. This could result in a Denial Of Service attack. There
4537 are no such structures used within SSL/TLS that come from untrusted sources
4538 so this is considered safe.
4540 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4546 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4548 * Read/write after SSL object in error state
4550 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4551 mechanism. The intent was that if a fatal error occurred during a handshake
4552 then OpenSSL would move into the error state and would immediately fail if
4553 you attempted to continue the handshake. This works as designed for the
4554 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4555 SSL_connect()), however due to a bug it does not work correctly if
4556 SSL_read() or SSL_write() is called directly. In that scenario, if the
4557 handshake fails then a fatal error will be returned in the initial function
4558 call. If SSL_read()/SSL_write() is subsequently called by the application
4559 for the same SSL object then it will succeed and the data is passed without
4560 being decrypted/encrypted directly from the SSL/TLS record layer.
4562 In order to exploit this issue an application bug would have to be present
4563 that resulted in a call to SSL_read()/SSL_write() being issued after having
4564 already received a fatal error.
4566 This issue was reported to OpenSSL by David Benjamin (Google).
4571 * rsaz_1024_mul_avx2 overflow bug on x86_64
4573 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4574 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4575 Analysis suggests that attacks against RSA and DSA as a result of this
4576 defect would be very difficult to perform and are not believed likely.
4577 Attacks against DH1024 are considered just feasible, because most of the
4578 work necessary to deduce information about a private key may be performed
4579 offline. The amount of resources required for such an attack would be
4580 significant. However, for an attack on TLS to be meaningful, the server
4581 would have to share the DH1024 private key among multiple clients, which is
4582 no longer an option since CVE-2016-0701.
4584 This only affects processors that support the AVX2 but not ADX extensions
4585 like Intel Haswell (4th generation).
4587 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4588 was originally found via the OSS-Fuzz project.
4593 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4595 * bn_sqrx8x_internal carry bug on x86_64
4597 There is a carry propagating bug in the x86_64 Montgomery squaring
4598 procedure. No EC algorithms are affected. Analysis suggests that attacks
4599 against RSA and DSA as a result of this defect would be very difficult to
4600 perform and are not believed likely. Attacks against DH are considered just
4601 feasible (although very difficult) because most of the work necessary to
4602 deduce information about a private key may be performed offline. The amount
4603 of resources required for such an attack would be very significant and
4604 likely only accessible to a limited number of attackers. An attacker would
4605 additionally need online access to an unpatched system using the target
4606 private key in a scenario with persistent DH parameters and a private
4607 key that is shared between multiple clients.
4609 This only affects processors that support the BMI1, BMI2 and ADX extensions
4610 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4612 This issue was reported to OpenSSL by the OSS-Fuzz project.
4617 * Malformed X.509 IPAddressFamily could cause OOB read
4619 If an X.509 certificate has a malformed IPAddressFamily extension,
4620 OpenSSL could do a one-byte buffer overread. The most likely result
4621 would be an erroneous display of the certificate in text format.
4623 This issue was reported to OpenSSL by the OSS-Fuzz project.
4628 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4630 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4631 platform rather than 'mingw'.
4635 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4637 * Truncated packet could crash via OOB read
4639 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4640 cipher is being used, then a truncated packet can cause that host to
4641 perform an out-of-bounds read, usually resulting in a crash.
4643 This issue was reported to OpenSSL by Robert Święcki of Google.
4648 * BN_mod_exp may produce incorrect results on x86_64
4650 There is a carry propagating bug in the x86_64 Montgomery squaring
4651 procedure. No EC algorithms are affected. Analysis suggests that attacks
4652 against RSA and DSA as a result of this defect would be very difficult to
4653 perform and are not believed likely. Attacks against DH are considered just
4654 feasible (although very difficult) because most of the work necessary to
4655 deduce information about a private key may be performed offline. The amount
4656 of resources required for such an attack would be very significant and
4657 likely only accessible to a limited number of attackers. An attacker would
4658 additionally need online access to an unpatched system using the target
4659 private key in a scenario with persistent DH parameters and a private
4660 key that is shared between multiple clients. For example this can occur by
4661 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4662 similar to CVE-2015-3193 but must be treated as a separate problem.
4664 This issue was reported to OpenSSL by the OSS-Fuzz project.
4669 * Montgomery multiplication may produce incorrect results
4671 There is a carry propagating bug in the Broadwell-specific Montgomery
4672 multiplication procedure that handles input lengths divisible by, but
4673 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4674 and DH private keys are impossible. This is because the subroutine in
4675 question is not used in operations with the private key itself and an input
4676 of the attacker's direct choice. Otherwise the bug can manifest itself as
4677 transient authentication and key negotiation failures or reproducible
4678 erroneous outcome of public-key operations with specially crafted input.
4679 Among EC algorithms only Brainpool P-512 curves are affected and one
4680 presumably can attack ECDH key negotiation. Impact was not analyzed in
4681 detail, because pre-requisites for attack are considered unlikely. Namely
4682 multiple clients have to choose the curve in question and the server has to
4683 share the private key among them, neither of which is default behaviour.
4684 Even then only clients that chose the curve will be affected.
4686 This issue was publicly reported as transient failures and was not
4687 initially recognized as a security issue. Thanks to Richard Morgan for
4688 providing reproducible case.
4693 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4694 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4695 prevent issues where no progress is being made and the peer continually
4696 sends unrecognised record types, using up resources processing them.
4700 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4702 * Missing CRL sanity check
4704 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4705 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4706 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4708 This issue only affects the OpenSSL 1.0.2i
4713 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4715 * OCSP Status Request extension unbounded memory growth
4717 A malicious client can send an excessively large OCSP Status Request
4718 extension. If that client continually requests renegotiation, sending a
4719 large OCSP Status Request extension each time, then there will be unbounded
4720 memory growth on the server. This will eventually lead to a Denial Of
4721 Service attack through memory exhaustion. Servers with a default
4722 configuration are vulnerable even if they do not support OCSP. Builds using
4723 the "no-ocsp" build time option are not affected.
4725 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4730 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4733 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4739 * OOB write in MDC2_Update()
4741 An overflow can occur in MDC2_Update() either if called directly or
4742 through the EVP_DigestUpdate() function using MDC2. If an attacker
4743 is able to supply very large amounts of input data after a previous
4744 call to EVP_EncryptUpdate() with a partial block then a length check
4745 can overflow resulting in a heap corruption.
4747 The amount of data needed is comparable to SIZE_MAX which is impractical
4750 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4755 * Malformed SHA512 ticket DoS
4757 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4758 DoS attack where a malformed ticket will result in an OOB read which will
4761 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4762 a custom server callback and ticket lookup mechanism.
4764 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4769 * OOB write in BN_bn2dec()
4771 The function BN_bn2dec() does not check the return value of BN_div_word().
4772 This can cause an OOB write if an application uses this function with an
4773 overly large BIGNUM. This could be a problem if an overly large certificate
4774 or CRL is printed out from an untrusted source. TLS is not affected because
4775 record limits will reject an oversized certificate before it is parsed.
4777 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4782 * OOB read in TS_OBJ_print_bio()
4784 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4785 the total length the OID text representation would use and not the amount
4786 of data written. This will result in OOB reads when large OIDs are
4789 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4794 * Pointer arithmetic undefined behaviour
4796 Avoid some undefined pointer arithmetic
4798 A common idiom in the codebase is to check limits in the following manner:
4801 Where "p" points to some malloc'd data of SIZE bytes and
4804 "len" here could be from some externally supplied data (e.g. from a TLS
4807 The rules of C pointer arithmetic are such that "p + len" is only well
4808 defined where len <= SIZE. Therefore the above idiom is actually
4809 undefined behaviour.
4811 For example this could cause problems if some malloc implementation
4812 provides an address for "p" such that "p + len" actually overflows for
4813 values of len that are too big and therefore p + len < limit.
4815 This issue was reported to OpenSSL by Guido Vranken
4820 * Constant time flag not preserved in DSA signing
4822 Operations in the DSA signing algorithm should run in constant time in
4823 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4824 implementation means that a non-constant time codepath is followed for
4825 certain operations. This has been demonstrated through a cache-timing
4826 attack to be sufficient for an attacker to recover the private DSA key.
4828 This issue was reported by César Pereida (Aalto University), Billy Brumley
4829 (Tampere University of Technology), and Yuval Yarom (The University of
4830 Adelaide and NICTA).
4835 * DTLS buffered message DoS
4837 In a DTLS connection where handshake messages are delivered out-of-order
4838 those messages that OpenSSL is not yet ready to process will be buffered
4839 for later use. Under certain circumstances, a flaw in the logic means that
4840 those messages do not get removed from the buffer even though the handshake
4841 has been completed. An attacker could force up to approx. 15 messages to
4842 remain in the buffer when they are no longer required. These messages will
4843 be cleared when the DTLS connection is closed. The default maximum size for
4844 a message is 100k. Therefore the attacker could force an additional 1500k
4845 to be consumed per connection. By opening many simulataneous connections an
4846 attacker could cause a DoS attack through memory exhaustion.
4848 This issue was reported to OpenSSL by Quan Luo.
4853 * DTLS replay protection DoS
4855 A flaw in the DTLS replay attack protection mechanism means that records
4856 that arrive for future epochs update the replay protection "window" before
4857 the MAC for the record has been validated. This could be exploited by an
4858 attacker by sending a record for the next epoch (which does not have to
4859 decrypt or have a valid MAC), with a very large sequence number. This means
4860 that all subsequent legitimate packets are dropped causing a denial of
4861 service for a specific DTLS connection.
4863 This issue was reported to OpenSSL by the OCAP audit team.
4868 * Certificate message OOB reads
4870 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4871 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4872 theoretical DoS risk but this has not been observed in practice on common
4875 The messages affected are client certificate, client certificate request
4876 and server certificate. As a result the attack can only be performed
4877 against a client or a server which enables client authentication.
4879 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4884 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4886 * Prevent padding oracle in AES-NI CBC MAC check
4888 A MITM attacker can use a padding oracle attack to decrypt traffic
4889 when the connection uses an AES CBC cipher and the server support
4892 This issue was introduced as part of the fix for Lucky 13 padding
4893 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4894 constant time by making sure that always the same bytes are read and
4895 compared against either the MAC or padding bytes. But it no longer
4896 checked that there was enough data to have both the MAC and padding
4899 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4904 * Fix EVP_EncodeUpdate overflow
4906 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4907 Base64 encoding of binary data. If an attacker is able to supply very large
4908 amounts of input data then a length check can overflow resulting in a heap
4911 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4912 the `PEM_write_bio*` family of functions. These are mainly used within the
4913 OpenSSL command line applications, so any application which processes data
4914 from an untrusted source and outputs it as a PEM file should be considered
4915 vulnerable to this issue. User applications that call these APIs directly
4916 with large amounts of untrusted data may also be vulnerable.
4918 This issue was reported by Guido Vranken.
4923 * Fix EVP_EncryptUpdate overflow
4925 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4926 is able to supply very large amounts of input data after a previous call to
4927 EVP_EncryptUpdate() with a partial block then a length check can overflow
4928 resulting in a heap corruption. Following an analysis of all OpenSSL
4929 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4930 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4931 the first called function after an EVP_EncryptInit(), and therefore that
4932 specific call must be safe. The second form is where the length passed to
4933 EVP_EncryptUpdate() can be seen from the code to be some small value and
4934 therefore there is no possibility of an overflow. Since all instances are
4935 one of these two forms, it is believed that there can be no overflows in
4936 internal code due to this problem. It should be noted that
4937 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4938 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4939 of these calls have also been analysed too and it is believed there are no
4940 instances in internal usage where an overflow could occur.
4942 This issue was reported by Guido Vranken.
4947 * Prevent ASN.1 BIO excessive memory allocation
4949 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4950 a short invalid encoding can cause allocation of large amounts of memory
4951 potentially consuming excessive resources or exhausting memory.
4953 Any application parsing untrusted data through d2i BIO functions is
4954 affected. The memory based functions such as d2i_X509() are *not* affected.
4955 Since the memory based functions are used by the TLS library, TLS
4956 applications are not affected.
4958 This issue was reported by Brian Carpenter.
4965 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4966 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4967 in arbitrary stack data being returned in the buffer.
4969 This issue was reported by Guido Vranken.
4974 * Modify behavior of ALPN to invoke callback after SNI/servername
4975 callback, such that updates to the SSL_CTX affect ALPN.
4979 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4984 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4985 methods are enabled and ssl2 is disabled the methods return NULL.
4989 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4991 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4992 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4993 provide any "EXPORT" or "LOW" strength ciphers.
4997 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4998 is by default disabled at build-time. Builds that are not configured with
4999 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5000 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5001 will need to explicitly call either of:
5003 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5005 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5007 as appropriate. Even if either of those is used, or the application
5008 explicitly uses the version-specific SSLv2_method() or its client and
5009 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5010 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5011 ciphers, and SSLv2 56-bit DES are no longer available.
5016 * Fix a double-free in DSA code
5018 A double free bug was discovered when OpenSSL parses malformed DSA private
5019 keys and could lead to a DoS attack or memory corruption for applications
5020 that receive DSA private keys from untrusted sources. This scenario is
5023 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5029 * Disable SRP fake user seed to address a server memory leak.
5031 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5033 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5034 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5035 was changed to ignore the "fake user" SRP seed, even if the seed
5038 Users should use SRP_VBASE_get1_by_user instead. Note that in
5039 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5040 also that even though configuring the SRP seed attempts to hide
5041 invalid usernames by continuing the handshake with fake
5042 credentials, this behaviour is not constant time and no strong
5043 guarantees are made that the handshake is indistinguishable from
5044 that of a valid user.
5049 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5051 In the BN_hex2bn function the number of hex digits is calculated using an
5052 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5053 large values of `i` this can result in `bn_expand` not allocating any
5054 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5055 field as NULL leading to a subsequent NULL ptr deref. For very large values
5056 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5057 In this case memory is allocated to the internal BIGNUM data field, but it
5058 is insufficiently sized leading to heap corruption. A similar issue exists
5059 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5060 is ever called by user applications with very large untrusted hex/dec data.
5061 This is anticipated to be a rare occurrence.
5063 All OpenSSL internal usage of these functions use data that is not expected
5064 to be untrusted, e.g. config file data or application command line
5065 arguments. If user developed applications generate config file data based
5066 on untrusted data then it is possible that this could also lead to security
5067 consequences. This is also anticipated to be rare.
5069 This issue was reported to OpenSSL by Guido Vranken.
5074 * Fix memory issues in `BIO_*printf` functions
5076 The internal `fmtstr` function used in processing a "%s" format string in
5077 the `BIO_*printf` functions could overflow while calculating the length of a
5078 string and cause an OOB read when printing very long strings.
5080 Additionally the internal `doapr_outch` function can attempt to write to an
5081 OOB memory location (at an offset from the NULL pointer) in the event of a
5082 memory allocation failure. In 1.0.2 and below this could be caused where
5083 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5084 could be in processing a very long "%s" format string. Memory leaks can
5087 The first issue may mask the second issue dependent on compiler behaviour.
5088 These problems could enable attacks where large amounts of untrusted data
5089 is passed to the `BIO_*printf` functions. If applications use these functions
5090 in this way then they could be vulnerable. OpenSSL itself uses these
5091 functions when printing out human-readable dumps of ASN.1 data. Therefore
5092 applications that print this data could be vulnerable if the data is from
5093 untrusted sources. OpenSSL command line applications could also be
5094 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5095 as command line arguments.
5097 Libssl is not considered directly vulnerable. Additionally certificates etc
5098 received via remote connections via libssl are also unlikely to be able to
5099 trigger these issues because of message size limits enforced within libssl.
5101 This issue was reported to OpenSSL Guido Vranken.
5106 * Side channel attack on modular exponentiation
5108 A side-channel attack was found which makes use of cache-bank conflicts on
5109 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5110 of RSA keys. The ability to exploit this issue is limited as it relies on
5111 an attacker who has control of code in a thread running on the same
5112 hyper-threaded core as the victim thread which is performing decryptions.
5114 This issue was reported to OpenSSL by Yuval Yarom, The University of
5115 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5116 Nadia Heninger, University of Pennsylvania with more information at
5117 <http://cachebleed.info>.
5122 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5123 if no keysize is specified with default_bits. This fixes an
5124 omission in an earlier change that changed all RSA/DSA key generation
5125 commands to use 2048 bits by default.
5129 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5131 * DH small subgroups
5133 Historically OpenSSL only ever generated DH parameters based on "safe"
5134 primes. More recently (in version 1.0.2) support was provided for
5135 generating X9.42 style parameter files such as those required for RFC 5114
5136 support. The primes used in such files may not be "safe". Where an
5137 application is using DH configured with parameters based on primes that are
5138 not "safe" then an attacker could use this fact to find a peer's private
5139 DH exponent. This attack requires that the attacker complete multiple
5140 handshakes in which the peer uses the same private DH exponent. For example
5141 this could be used to discover a TLS server's private DH exponent if it's
5142 reusing the private DH exponent or it's using a static DH ciphersuite.
5144 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5145 TLS. It is not on by default. If the option is not set then the server
5146 reuses the same private DH exponent for the life of the server process and
5147 would be vulnerable to this attack. It is believed that many popular
5148 applications do set this option and would therefore not be at risk.
5150 The fix for this issue adds an additional check where a "q" parameter is
5151 available (as is the case in X9.42 based parameters). This detects the
5152 only known attack, and is the only possible defense for static DH
5153 ciphersuites. This could have some performance impact.
5155 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5156 default and cannot be disabled. This could have some performance impact.
5158 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5163 * SSLv2 doesn't block disabled ciphers
5165 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5166 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5167 been disabled, provided that the SSLv2 protocol was not also disabled via
5170 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5171 and Sebastian Schinzel.
5176 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5178 * BN_mod_exp may produce incorrect results on x86_64
5180 There is a carry propagating bug in the x86_64 Montgomery squaring
5181 procedure. No EC algorithms are affected. Analysis suggests that attacks
5182 against RSA and DSA as a result of this defect would be very difficult to
5183 perform and are not believed likely. Attacks against DH are considered just
5184 feasible (although very difficult) because most of the work necessary to
5185 deduce information about a private key may be performed offline. The amount
5186 of resources required for such an attack would be very significant and
5187 likely only accessible to a limited number of attackers. An attacker would
5188 additionally need online access to an unpatched system using the target
5189 private key in a scenario with persistent DH parameters and a private
5190 key that is shared between multiple clients. For example this can occur by
5191 default in OpenSSL DHE based SSL/TLS ciphersuites.
5193 This issue was reported to OpenSSL by Hanno Böck.
5198 * Certificate verify crash with missing PSS parameter
5200 The signature verification routines will crash with a NULL pointer
5201 dereference if presented with an ASN.1 signature using the RSA PSS
5202 algorithm and absent mask generation function parameter. Since these
5203 routines are used to verify certificate signature algorithms this can be
5204 used to crash any certificate verification operation and exploited in a
5205 DoS attack. Any application which performs certificate verification is
5206 vulnerable including OpenSSL clients and servers which enable client
5209 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5214 * X509_ATTRIBUTE memory leak
5216 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5217 memory. This structure is used by the PKCS#7 and CMS routines so any
5218 application which reads PKCS#7 or CMS data from untrusted sources is
5219 affected. SSL/TLS is not affected.
5221 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5227 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5228 This changes the decoding behaviour for some invalid messages,
5229 though the change is mostly in the more lenient direction, and
5230 legacy behaviour is preserved as much as possible.
5234 * In DSA_generate_parameters_ex, if the provided seed is too short,
5237 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5239 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5241 * Alternate chains certificate forgery
5243 During certificate verification, OpenSSL will attempt to find an
5244 alternative certificate chain if the first attempt to build such a chain
5245 fails. An error in the implementation of this logic can mean that an
5246 attacker could cause certain checks on untrusted certificates to be
5247 bypassed, such as the CA flag, enabling them to use a valid leaf
5248 certificate to act as a CA and "issue" an invalid certificate.
5250 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5255 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5257 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5258 incompatibility in the handling of HMAC. The previous ABI has now been
5263 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5265 * Malformed ECParameters causes infinite loop
5267 When processing an ECParameters structure OpenSSL enters an infinite loop
5268 if the curve specified is over a specially malformed binary polynomial
5271 This can be used to perform denial of service against any
5272 system which processes public keys, certificate requests or
5273 certificates. This includes TLS clients and TLS servers with
5274 client authentication enabled.
5276 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5281 * Exploitable out-of-bounds read in X509_cmp_time
5283 X509_cmp_time does not properly check the length of the ASN1_TIME
5284 string and can read a few bytes out of bounds. In addition,
5285 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5288 An attacker can use this to craft malformed certificates and CRLs of
5289 various sizes and potentially cause a segmentation fault, resulting in
5290 a DoS on applications that verify certificates or CRLs. TLS clients
5291 that verify CRLs are affected. TLS clients and servers with client
5292 authentication enabled may be affected if they use custom verification
5295 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5296 independently by Hanno Böck.
5301 * PKCS7 crash with missing EnvelopedContent
5303 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5304 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5305 with missing content and trigger a NULL pointer dereference on parsing.
5307 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5308 structures from untrusted sources are affected. OpenSSL clients and
5309 servers are not affected.
5311 This issue was reported to OpenSSL by Michal Zalewski (Google).
5316 * CMS verify infinite loop with unknown hash function
5318 When verifying a signedData message the CMS code can enter an infinite loop
5319 if presented with an unknown hash function OID. This can be used to perform
5320 denial of service against any system which verifies signedData messages using
5322 This issue was reported to OpenSSL by Johannes Bauer.
5327 * Race condition handling NewSessionTicket
5329 If a NewSessionTicket is received by a multi-threaded client when attempting to
5330 reuse a previous ticket then a race condition can occur potentially leading to
5331 a double free of the ticket data.
5336 * Only support 256-bit or stronger elliptic curves with the
5337 'ecdh_auto' setting (server) or by default (client). Of supported
5338 curves, prefer P-256 (both).
5342 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5344 * ClientHello sigalgs DoS fix
5346 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5347 invalid signature algorithms extension a NULL pointer dereference will
5348 occur. This can be exploited in a DoS attack against the server.
5350 This issue was was reported to OpenSSL by David Ramos of Stanford
5354 *Stephen Henson and Matt Caswell*
5356 * Multiblock corrupted pointer fix
5358 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5359 feature only applies on 64 bit x86 architecture platforms that support AES
5360 NI instructions. A defect in the implementation of "multiblock" can cause
5361 OpenSSL's internal write buffer to become incorrectly set to NULL when
5362 using non-blocking IO. Typically, when the user application is using a
5363 socket BIO for writing, this will only result in a failed connection.
5364 However if some other BIO is used then it is likely that a segmentation
5365 fault will be triggered, thus enabling a potential DoS attack.
5367 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5372 * Segmentation fault in DTLSv1_listen fix
5374 The DTLSv1_listen function is intended to be stateless and processes the
5375 initial ClientHello from many peers. It is common for user code to loop
5376 over the call to DTLSv1_listen until a valid ClientHello is received with
5377 an associated cookie. A defect in the implementation of DTLSv1_listen means
5378 that state is preserved in the SSL object from one invocation to the next
5379 that can lead to a segmentation fault. Errors processing the initial
5380 ClientHello can trigger this scenario. An example of such an error could be
5381 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5384 This issue was reported to OpenSSL by Per Allansson.
5389 * Segmentation fault in ASN1_TYPE_cmp fix
5391 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5392 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5393 certificate signature algorithm consistency this can be used to crash any
5394 certificate verification operation and exploited in a DoS attack. Any
5395 application which performs certificate verification is vulnerable including
5396 OpenSSL clients and servers which enable client authentication.
5401 * Segmentation fault for invalid PSS parameters fix
5403 The signature verification routines will crash with a NULL pointer
5404 dereference if presented with an ASN.1 signature using the RSA PSS
5405 algorithm and invalid parameters. Since these routines are used to verify
5406 certificate signature algorithms this can be used to crash any
5407 certificate verification operation and exploited in a DoS attack. Any
5408 application which performs certificate verification is vulnerable including
5409 OpenSSL clients and servers which enable client authentication.
5411 This issue was was reported to OpenSSL by Brian Carpenter.
5416 * ASN.1 structure reuse memory corruption fix
5418 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5419 memory corruption via an invalid write. Such reuse is and has been
5420 strongly discouraged and is believed to be rare.
5422 Applications that parse structures containing CHOICE or ANY DEFINED BY
5423 components may be affected. Certificate parsing (d2i_X509 and related
5424 functions) are however not affected. OpenSSL clients and servers are
5430 * PKCS7 NULL pointer dereferences fix
5432 The PKCS#7 parsing code does not handle missing outer ContentInfo
5433 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5434 missing content and trigger a NULL pointer dereference on parsing.
5436 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5437 otherwise parse PKCS#7 structures from untrusted sources are
5438 affected. OpenSSL clients and servers are not affected.
5440 This issue was reported to OpenSSL by Michal Zalewski (Google).
5445 * DoS via reachable assert in SSLv2 servers fix
5447 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5448 servers that both support SSLv2 and enable export cipher suites by sending
5449 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5451 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5452 (OpenSSL development team).
5457 * Empty CKE with client auth and DHE fix
5459 If client auth is used then a server can seg fault in the event of a DHE
5460 ciphersuite being selected and a zero length ClientKeyExchange message
5461 being sent by the client. This could be exploited in a DoS attack.
5466 * Handshake with unseeded PRNG fix
5468 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5469 with an unseeded PRNG. The conditions are:
5470 - The client is on a platform where the PRNG has not been seeded
5471 automatically, and the user has not seeded manually
5472 - A protocol specific client method version has been used (i.e. not
5473 SSL_client_methodv23)
5474 - A ciphersuite is used that does not require additional random data from
5475 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5477 If the handshake succeeds then the client random that has been used will
5478 have been generated from a PRNG with insufficient entropy and therefore the
5479 output may be predictable.
5481 For example using the following command with an unseeded openssl will
5482 succeed on an unpatched platform:
5484 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5489 * Use After Free following d2i_ECPrivatekey error fix
5491 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5492 could cause a use after free condition. This, in turn, could cause a double
5493 free in several private key parsing functions (such as d2i_PrivateKey
5494 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5495 for applications that receive EC private keys from untrusted
5496 sources. This scenario is considered rare.
5498 This issue was discovered by the BoringSSL project and fixed in their
5504 * X509_to_X509_REQ NULL pointer deref fix
5506 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5507 the certificate key is invalid. This function is rarely used in practice.
5509 This issue was discovered by Brian Carpenter.
5514 * Removed the export ciphers from the DEFAULT ciphers
5518 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5520 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5521 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5522 So far those who have to target multiple platforms would compromise
5523 and argue that binary targeting say ARMv5 would still execute on
5524 ARMv8. "Universal" build resolves this compromise by providing
5525 near-optimal performance even on newer platforms.
5529 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5530 (other platforms pending).
5532 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5534 * Add support for the SignedCertificateTimestampList certificate and
5535 OCSP response extensions from RFC6962.
5539 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5540 for corner cases. (Certain input points at infinity could lead to
5541 bogus results, with non-infinity inputs mapped to infinity too.)
5545 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5546 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5547 common cases are optimized and there still is room for further
5548 improvements. Vector Permutation AES for Altivec is also added.
5552 * Add support for little-endian ppc64 Linux target.
5554 *Marcelo Cerri (IBM)*
5556 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5557 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5558 are optimized and there still is room for further improvements.
5559 Both 32- and 64-bit modes are supported.
5561 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5563 * Improved ARMv7 NEON support.
5567 * Support for SPARC Architecture 2011 crypto extensions, first
5568 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5569 SHA256/512, MD5, GHASH and modular exponentiation.
5571 *Andy Polyakov, David Miller*
5573 * Accelerated modular exponentiation for Intel processors, a.k.a.
5576 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5578 * Support for new and upcoming Intel processors, including AVX2,
5579 BMI and SHA ISA extensions. This includes additional "stitched"
5580 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5583 This work was sponsored by Intel Corp.
5587 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5588 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5589 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5593 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5594 this fixes a limitation in previous versions of OpenSSL.
5598 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5599 MGF1 digest and OAEP label.
5603 * Add EVP support for key wrapping algorithms, to avoid problems with
5604 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5605 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5606 algorithms and include tests cases.
5610 * Add functions to allocate and set the fields of an ECDSA_METHOD
5613 *Douglas E. Engert, Steve Henson*
5615 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5616 difference in days and seconds between two tm or ASN1_TIME structures.
5620 * Add -rev test option to s_server to just reverse order of characters
5621 received by client and send back to server. Also prints an abbreviated
5622 summary of the connection parameters.
5626 * New option -brief for s_client and s_server to print out a brief summary
5627 of connection parameters.
5631 * Add callbacks for arbitrary TLS extensions.
5633 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5635 * New option -crl_download in several openssl utilities to download CRLs
5636 from CRLDP extension in certificates.
5640 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5644 * New function X509_CRL_diff to generate a delta CRL from the difference
5645 of two full CRLs. Add support to "crl" utility.
5649 * New functions to set lookup_crls function and to retrieve
5650 X509_STORE from X509_STORE_CTX.
5654 * Print out deprecated issuer and subject unique ID fields in
5659 * Extend OCSP I/O functions so they can be used for simple general purpose
5660 HTTP as well as OCSP. New wrapper function which can be used to download
5661 CRLs using the OCSP API.
5665 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5669 * `SSL_CONF*` functions. These provide a common framework for application
5670 configuration using configuration files or command lines.
5674 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5675 message callback and prints the results. Needs compile time option
5676 "enable-ssl-trace". New options to s_client and s_server to enable
5681 * New ctrl and macro to retrieve supported points extensions.
5682 Print out extension in s_server and s_client.
5686 * New functions to retrieve certificate signature and signature
5691 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5696 * New Suite B modes for TLS code. These use and enforce the requirements
5697 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5698 only use Suite B curves. The Suite B modes can be set by using the
5699 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5703 * New chain verification flags for Suite B levels of security. Check
5704 algorithms are acceptable when flags are set in X509_verify_cert.
5708 * Make tls1_check_chain return a set of flags indicating checks passed
5709 by a certificate chain. Add additional tests to handle client
5710 certificates: checks for matching certificate type and issuer name
5715 * If an attempt is made to use a signature algorithm not in the peer
5716 preference list abort the handshake. If client has no suitable
5717 signature algorithms in response to a certificate request do not
5718 use the certificate.
5722 * If server EC tmp key is not in client preference list abort handshake.
5726 * Add support for certificate stores in CERT structure. This makes it
5727 possible to have different stores per SSL structure or one store in
5728 the parent SSL_CTX. Include distinct stores for certificate chain
5729 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5730 to build and store a certificate chain in CERT structure: returning
5731 an error if the chain cannot be built: this will allow applications
5732 to test if a chain is correctly configured.
5734 Note: if the CERT based stores are not set then the parent SSL_CTX
5735 store is used to retain compatibility with existing behaviour.
5739 * New function ssl_set_client_disabled to set a ciphersuite disabled
5740 mask based on the current session, check mask when sending client
5741 hello and checking the requested ciphersuite.
5745 * New ctrls to retrieve and set certificate types in a certificate
5746 request message. Print out received values in s_client. If certificate
5747 types is not set with custom values set sensible values based on
5748 supported signature algorithms.
5752 * Support for distinct client and server supported signature algorithms.
5756 * Add certificate callback. If set this is called whenever a certificate
5757 is required by client or server. An application can decide which
5758 certificate chain to present based on arbitrary criteria: for example
5759 supported signature algorithms. Add very simple example to s_server.
5760 This fixes many of the problems and restrictions of the existing client
5761 certificate callback: for example you can now clear an existing
5762 certificate and specify the whole chain.
5766 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5767 the certificate can be used for (if anything). Set valid_flags field
5768 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5769 to have similar checks in it.
5771 Add new "cert_flags" field to CERT structure and include a "strict mode".
5772 This enforces some TLS certificate requirements (such as only permitting
5773 certificate signature algorithms contained in the supported algorithms
5774 extension) which some implementations ignore: this option should be used
5775 with caution as it could cause interoperability issues.
5779 * Update and tidy signature algorithm extension processing. Work out
5780 shared signature algorithms based on preferences and peer algorithms
5781 and print them out in s_client and s_server. Abort handshake if no
5782 shared signature algorithms.
5786 * Add new functions to allow customised supported signature algorithms
5787 for SSL and SSL_CTX structures. Add options to s_client and s_server
5792 * New function SSL_certs_clear() to delete all references to certificates
5793 from an SSL structure. Before this once a certificate had been added
5794 it couldn't be removed.
5798 * Integrate hostname, email address and IP address checking with certificate
5799 verification. New verify options supporting checking in openssl utility.
5803 * Fixes and wildcard matching support to hostname and email checking
5804 functions. Add manual page.
5806 *Florian Weimer (Red Hat Product Security Team)*
5808 * New functions to check a hostname email or IP address against a
5809 certificate. Add options x509 utility to print results of checks against
5814 * Fix OCSP checking.
5816 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5818 * Initial experimental support for explicitly trusted non-root CAs.
5819 OpenSSL still tries to build a complete chain to a root but if an
5820 intermediate CA has a trust setting included that is used. The first
5821 setting is used: whether to trust (e.g., -addtrust option to the x509
5826 * Add -trusted_first option which attempts to find certificates in the
5827 trusted store even if an untrusted chain is also supplied.
5831 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5832 platform support for Linux and Android.
5836 * Support for linux-x32, ILP32 environment in x86_64 framework.
5840 * Experimental multi-implementation support for FIPS capable OpenSSL.
5841 When in FIPS mode the approved implementations are used as normal,
5842 when not in FIPS mode the internal unapproved versions are used instead.
5843 This means that the FIPS capable OpenSSL isn't forced to use the
5844 (often lower performance) FIPS implementations outside FIPS mode.
5848 * Transparently support X9.42 DH parameters when calling
5849 PEM_read_bio_DHparameters. This means existing applications can handle
5850 the new parameter format automatically.
5854 * Initial experimental support for X9.42 DH parameter format: mainly
5855 to support use of 'q' parameter for RFC5114 parameters.
5859 * Add DH parameters from RFC5114 including test data to dhtest.
5863 * Support for automatic EC temporary key parameter selection. If enabled
5864 the most preferred EC parameters are automatically used instead of
5865 hardcoded fixed parameters. Now a server just has to call:
5866 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5867 support ECDH and use the most appropriate parameters.
5871 * Enhance and tidy EC curve and point format TLS extension code. Use
5872 static structures instead of allocation if default values are used.
5873 New ctrls to set curves we wish to support and to retrieve shared curves.
5874 Print out shared curves in s_server. New options to s_server and s_client
5875 to set list of supported curves.
5879 * New ctrls to retrieve supported signature algorithms and
5880 supported curve values as an array of NIDs. Extend openssl utility
5881 to print out received values.
5885 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5886 between NIDs and the more common NIST names such as "P-256". Enhance
5887 ecparam utility and ECC method to recognise the NIST names for curves.
5891 * Enhance SSL/TLS certificate chain handling to support different
5892 chains for each certificate instead of one chain in the parent SSL_CTX.
5896 * Support for fixed DH ciphersuite client authentication: where both
5897 server and client use DH certificates with common parameters.
5901 * Support for fixed DH ciphersuites: those requiring DH server
5906 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5908 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5909 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5910 X509_CINF_get_signature were reverted post internal team review.
5915 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5917 * OCSP Status Request extension unbounded memory growth
5919 A malicious client can send an excessively large OCSP Status Request
5920 extension. If that client continually requests renegotiation, sending a
5921 large OCSP Status Request extension each time, then there will be unbounded
5922 memory growth on the server. This will eventually lead to a Denial Of
5923 Service attack through memory exhaustion. Servers with a default
5924 configuration are vulnerable even if they do not support OCSP. Builds using
5925 the "no-ocsp" build time option are not affected.
5927 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5932 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5935 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5941 * OOB write in MDC2_Update()
5943 An overflow can occur in MDC2_Update() either if called directly or
5944 through the EVP_DigestUpdate() function using MDC2. If an attacker
5945 is able to supply very large amounts of input data after a previous
5946 call to EVP_EncryptUpdate() with a partial block then a length check
5947 can overflow resulting in a heap corruption.
5949 The amount of data needed is comparable to SIZE_MAX which is impractical
5952 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5957 * Malformed SHA512 ticket DoS
5959 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5960 DoS attack where a malformed ticket will result in an OOB read which will
5963 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5964 a custom server callback and ticket lookup mechanism.
5966 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5971 * OOB write in BN_bn2dec()
5973 The function BN_bn2dec() does not check the return value of BN_div_word().
5974 This can cause an OOB write if an application uses this function with an
5975 overly large BIGNUM. This could be a problem if an overly large certificate
5976 or CRL is printed out from an untrusted source. TLS is not affected because
5977 record limits will reject an oversized certificate before it is parsed.
5979 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5984 * OOB read in TS_OBJ_print_bio()
5986 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5987 the total length the OID text representation would use and not the amount
5988 of data written. This will result in OOB reads when large OIDs are
5991 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5996 * Pointer arithmetic undefined behaviour
5998 Avoid some undefined pointer arithmetic
6000 A common idiom in the codebase is to check limits in the following manner:
6003 Where "p" points to some malloc'd data of SIZE bytes and
6006 "len" here could be from some externally supplied data (e.g. from a TLS
6009 The rules of C pointer arithmetic are such that "p + len" is only well
6010 defined where len <= SIZE. Therefore the above idiom is actually
6011 undefined behaviour.
6013 For example this could cause problems if some malloc implementation
6014 provides an address for "p" such that "p + len" actually overflows for
6015 values of len that are too big and therefore p + len < limit.
6017 This issue was reported to OpenSSL by Guido Vranken
6022 * Constant time flag not preserved in DSA signing
6024 Operations in the DSA signing algorithm should run in constant time in
6025 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6026 implementation means that a non-constant time codepath is followed for
6027 certain operations. This has been demonstrated through a cache-timing
6028 attack to be sufficient for an attacker to recover the private DSA key.
6030 This issue was reported by César Pereida (Aalto University), Billy Brumley
6031 (Tampere University of Technology), and Yuval Yarom (The University of
6032 Adelaide and NICTA).
6037 * DTLS buffered message DoS
6039 In a DTLS connection where handshake messages are delivered out-of-order
6040 those messages that OpenSSL is not yet ready to process will be buffered
6041 for later use. Under certain circumstances, a flaw in the logic means that
6042 those messages do not get removed from the buffer even though the handshake
6043 has been completed. An attacker could force up to approx. 15 messages to
6044 remain in the buffer when they are no longer required. These messages will
6045 be cleared when the DTLS connection is closed. The default maximum size for
6046 a message is 100k. Therefore the attacker could force an additional 1500k
6047 to be consumed per connection. By opening many simulataneous connections an
6048 attacker could cause a DoS attack through memory exhaustion.
6050 This issue was reported to OpenSSL by Quan Luo.
6055 * DTLS replay protection DoS
6057 A flaw in the DTLS replay attack protection mechanism means that records
6058 that arrive for future epochs update the replay protection "window" before
6059 the MAC for the record has been validated. This could be exploited by an
6060 attacker by sending a record for the next epoch (which does not have to
6061 decrypt or have a valid MAC), with a very large sequence number. This means
6062 that all subsequent legitimate packets are dropped causing a denial of
6063 service for a specific DTLS connection.
6065 This issue was reported to OpenSSL by the OCAP audit team.
6070 * Certificate message OOB reads
6072 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6073 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6074 theoretical DoS risk but this has not been observed in practice on common
6077 The messages affected are client certificate, client certificate request
6078 and server certificate. As a result the attack can only be performed
6079 against a client or a server which enables client authentication.
6081 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6086 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6088 * Prevent padding oracle in AES-NI CBC MAC check
6090 A MITM attacker can use a padding oracle attack to decrypt traffic
6091 when the connection uses an AES CBC cipher and the server support
6094 This issue was introduced as part of the fix for Lucky 13 padding
6095 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6096 constant time by making sure that always the same bytes are read and
6097 compared against either the MAC or padding bytes. But it no longer
6098 checked that there was enough data to have both the MAC and padding
6101 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6106 * Fix EVP_EncodeUpdate overflow
6108 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6109 Base64 encoding of binary data. If an attacker is able to supply very large
6110 amounts of input data then a length check can overflow resulting in a heap
6113 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6114 the `PEM_write_bio*` family of functions. These are mainly used within the
6115 OpenSSL command line applications, so any application which processes data
6116 from an untrusted source and outputs it as a PEM file should be considered
6117 vulnerable to this issue. User applications that call these APIs directly
6118 with large amounts of untrusted data may also be vulnerable.
6120 This issue was reported by Guido Vranken.
6125 * Fix EVP_EncryptUpdate overflow
6127 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6128 is able to supply very large amounts of input data after a previous call to
6129 EVP_EncryptUpdate() with a partial block then a length check can overflow
6130 resulting in a heap corruption. Following an analysis of all OpenSSL
6131 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6132 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6133 the first called function after an EVP_EncryptInit(), and therefore that
6134 specific call must be safe. The second form is where the length passed to
6135 EVP_EncryptUpdate() can be seen from the code to be some small value and
6136 therefore there is no possibility of an overflow. Since all instances are
6137 one of these two forms, it is believed that there can be no overflows in
6138 internal code due to this problem. It should be noted that
6139 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6140 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6141 of these calls have also been analysed too and it is believed there are no
6142 instances in internal usage where an overflow could occur.
6144 This issue was reported by Guido Vranken.
6149 * Prevent ASN.1 BIO excessive memory allocation
6151 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6152 a short invalid encoding can casuse allocation of large amounts of memory
6153 potentially consuming excessive resources or exhausting memory.
6155 Any application parsing untrusted data through d2i BIO functions is
6156 affected. The memory based functions such as d2i_X509() are *not* affected.
6157 Since the memory based functions are used by the TLS library, TLS
6158 applications are not affected.
6160 This issue was reported by Brian Carpenter.
6167 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6168 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6169 in arbitrary stack data being returned in the buffer.
6171 This issue was reported by Guido Vranken.
6176 * Modify behavior of ALPN to invoke callback after SNI/servername
6177 callback, such that updates to the SSL_CTX affect ALPN.
6181 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6186 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6187 methods are enabled and ssl2 is disabled the methods return NULL.
6191 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6193 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6194 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6195 provide any "EXPORT" or "LOW" strength ciphers.
6199 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6200 is by default disabled at build-time. Builds that are not configured with
6201 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6202 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6203 will need to explicitly call either of:
6205 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6207 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6209 as appropriate. Even if either of those is used, or the application
6210 explicitly uses the version-specific SSLv2_method() or its client and
6211 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6212 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6213 ciphers, and SSLv2 56-bit DES are no longer available.
6218 * Fix a double-free in DSA code
6220 A double free bug was discovered when OpenSSL parses malformed DSA private
6221 keys and could lead to a DoS attack or memory corruption for applications
6222 that receive DSA private keys from untrusted sources. This scenario is
6225 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6231 * Disable SRP fake user seed to address a server memory leak.
6233 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6235 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6236 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6237 was changed to ignore the "fake user" SRP seed, even if the seed
6240 Users should use SRP_VBASE_get1_by_user instead. Note that in
6241 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6242 also that even though configuring the SRP seed attempts to hide
6243 invalid usernames by continuing the handshake with fake
6244 credentials, this behaviour is not constant time and no strong
6245 guarantees are made that the handshake is indistinguishable from
6246 that of a valid user.
6251 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6253 In the BN_hex2bn function the number of hex digits is calculated using an
6254 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6255 large values of `i` this can result in `bn_expand` not allocating any
6256 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6257 field as NULL leading to a subsequent NULL ptr deref. For very large values
6258 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6259 In this case memory is allocated to the internal BIGNUM data field, but it
6260 is insufficiently sized leading to heap corruption. A similar issue exists
6261 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6262 is ever called by user applications with very large untrusted hex/dec data.
6263 This is anticipated to be a rare occurrence.
6265 All OpenSSL internal usage of these functions use data that is not expected
6266 to be untrusted, e.g. config file data or application command line
6267 arguments. If user developed applications generate config file data based
6268 on untrusted data then it is possible that this could also lead to security
6269 consequences. This is also anticipated to be rare.
6271 This issue was reported to OpenSSL by Guido Vranken.
6276 * Fix memory issues in `BIO_*printf` functions
6278 The internal `fmtstr` function used in processing a "%s" format string in
6279 the `BIO_*printf` functions could overflow while calculating the length of a
6280 string and cause an OOB read when printing very long strings.
6282 Additionally the internal `doapr_outch` function can attempt to write to an
6283 OOB memory location (at an offset from the NULL pointer) in the event of a
6284 memory allocation failure. In 1.0.2 and below this could be caused where
6285 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6286 could be in processing a very long "%s" format string. Memory leaks can
6289 The first issue may mask the second issue dependent on compiler behaviour.
6290 These problems could enable attacks where large amounts of untrusted data
6291 is passed to the `BIO_*printf` functions. If applications use these functions
6292 in this way then they could be vulnerable. OpenSSL itself uses these
6293 functions when printing out human-readable dumps of ASN.1 data. Therefore
6294 applications that print this data could be vulnerable if the data is from
6295 untrusted sources. OpenSSL command line applications could also be
6296 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6297 as command line arguments.
6299 Libssl is not considered directly vulnerable. Additionally certificates etc
6300 received via remote connections via libssl are also unlikely to be able to
6301 trigger these issues because of message size limits enforced within libssl.
6303 This issue was reported to OpenSSL Guido Vranken.
6308 * Side channel attack on modular exponentiation
6310 A side-channel attack was found which makes use of cache-bank conflicts on
6311 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6312 of RSA keys. The ability to exploit this issue is limited as it relies on
6313 an attacker who has control of code in a thread running on the same
6314 hyper-threaded core as the victim thread which is performing decryptions.
6316 This issue was reported to OpenSSL by Yuval Yarom, The University of
6317 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6318 Nadia Heninger, University of Pennsylvania with more information at
6319 <http://cachebleed.info>.
6324 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6325 if no keysize is specified with default_bits. This fixes an
6326 omission in an earlier change that changed all RSA/DSA key generation
6327 commands to use 2048 bits by default.
6331 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6333 * Protection for DH small subgroup attacks
6335 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6336 switched on by default and cannot be disabled. This could have some
6341 * SSLv2 doesn't block disabled ciphers
6343 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6344 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6345 been disabled, provided that the SSLv2 protocol was not also disabled via
6348 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6349 and Sebastian Schinzel.
6354 * Reject DH handshakes with parameters shorter than 1024 bits.
6358 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6360 * Certificate verify crash with missing PSS parameter
6362 The signature verification routines will crash with a NULL pointer
6363 dereference if presented with an ASN.1 signature using the RSA PSS
6364 algorithm and absent mask generation function parameter. Since these
6365 routines are used to verify certificate signature algorithms this can be
6366 used to crash any certificate verification operation and exploited in a
6367 DoS attack. Any application which performs certificate verification is
6368 vulnerable including OpenSSL clients and servers which enable client
6371 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6376 * X509_ATTRIBUTE memory leak
6378 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6379 memory. This structure is used by the PKCS#7 and CMS routines so any
6380 application which reads PKCS#7 or CMS data from untrusted sources is
6381 affected. SSL/TLS is not affected.
6383 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6389 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6390 This changes the decoding behaviour for some invalid messages,
6391 though the change is mostly in the more lenient direction, and
6392 legacy behaviour is preserved as much as possible.
6396 * In DSA_generate_parameters_ex, if the provided seed is too short,
6397 use a random seed, as already documented.
6399 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6401 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6403 * Alternate chains certificate forgery
6405 During certificate verfification, OpenSSL will attempt to find an
6406 alternative certificate chain if the first attempt to build such a chain
6407 fails. An error in the implementation of this logic can mean that an
6408 attacker could cause certain checks on untrusted certificates to be
6409 bypassed, such as the CA flag, enabling them to use a valid leaf
6410 certificate to act as a CA and "issue" an invalid certificate.
6412 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6418 * Race condition handling PSK identify hint
6420 If PSK identity hints are received by a multi-threaded client then
6421 the values are wrongly updated in the parent SSL_CTX structure. This can
6422 result in a race condition potentially leading to a double free of the
6428 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6430 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6431 incompatibility in the handling of HMAC. The previous ABI has now been
6434 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6436 * Malformed ECParameters causes infinite loop
6438 When processing an ECParameters structure OpenSSL enters an infinite loop
6439 if the curve specified is over a specially malformed binary polynomial
6442 This can be used to perform denial of service against any
6443 system which processes public keys, certificate requests or
6444 certificates. This includes TLS clients and TLS servers with
6445 client authentication enabled.
6447 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6452 * Exploitable out-of-bounds read in X509_cmp_time
6454 X509_cmp_time does not properly check the length of the ASN1_TIME
6455 string and can read a few bytes out of bounds. In addition,
6456 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6459 An attacker can use this to craft malformed certificates and CRLs of
6460 various sizes and potentially cause a segmentation fault, resulting in
6461 a DoS on applications that verify certificates or CRLs. TLS clients
6462 that verify CRLs are affected. TLS clients and servers with client
6463 authentication enabled may be affected if they use custom verification
6466 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6467 independently by Hanno Böck.
6472 * PKCS7 crash with missing EnvelopedContent
6474 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6475 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6476 with missing content and trigger a NULL pointer dereference on parsing.
6478 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6479 structures from untrusted sources are affected. OpenSSL clients and
6480 servers are not affected.
6482 This issue was reported to OpenSSL by Michal Zalewski (Google).
6487 * CMS verify infinite loop with unknown hash function
6489 When verifying a signedData message the CMS code can enter an infinite loop
6490 if presented with an unknown hash function OID. This can be used to perform
6491 denial of service against any system which verifies signedData messages using
6493 This issue was reported to OpenSSL by Johannes Bauer.
6498 * Race condition handling NewSessionTicket
6500 If a NewSessionTicket is received by a multi-threaded client when attempting to
6501 reuse a previous ticket then a race condition can occur potentially leading to
6502 a double free of the ticket data.
6507 * Reject DH handshakes with parameters shorter than 768 bits.
6509 *Kurt Roeckx and Emilia Kasper*
6511 * dhparam: generate 2048-bit parameters by default.
6513 *Kurt Roeckx and Emilia Kasper*
6515 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6517 * Segmentation fault in ASN1_TYPE_cmp fix
6519 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6520 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6521 certificate signature algorithm consistency this can be used to crash any
6522 certificate verification operation and exploited in a DoS attack. Any
6523 application which performs certificate verification is vulnerable including
6524 OpenSSL clients and servers which enable client authentication.
6529 * ASN.1 structure reuse memory corruption fix
6531 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6532 memory corruption via an invalid write. Such reuse is and has been
6533 strongly discouraged and is believed to be rare.
6535 Applications that parse structures containing CHOICE or ANY DEFINED BY
6536 components may be affected. Certificate parsing (d2i_X509 and related
6537 functions) are however not affected. OpenSSL clients and servers are
6543 * PKCS7 NULL pointer dereferences fix
6545 The PKCS#7 parsing code does not handle missing outer ContentInfo
6546 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6547 missing content and trigger a NULL pointer dereference on parsing.
6549 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6550 otherwise parse PKCS#7 structures from untrusted sources are
6551 affected. OpenSSL clients and servers are not affected.
6553 This issue was reported to OpenSSL by Michal Zalewski (Google).
6558 * DoS via reachable assert in SSLv2 servers fix
6560 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6561 servers that both support SSLv2 and enable export cipher suites by sending
6562 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6564 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6565 (OpenSSL development team).
6570 * Use After Free following d2i_ECPrivatekey error fix
6572 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6573 could cause a use after free condition. This, in turn, could cause a double
6574 free in several private key parsing functions (such as d2i_PrivateKey
6575 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6576 for applications that receive EC private keys from untrusted
6577 sources. This scenario is considered rare.
6579 This issue was discovered by the BoringSSL project and fixed in their
6585 * X509_to_X509_REQ NULL pointer deref fix
6587 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6588 the certificate key is invalid. This function is rarely used in practice.
6590 This issue was discovered by Brian Carpenter.
6595 * Removed the export ciphers from the DEFAULT ciphers
6599 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6601 * Build fixes for the Windows and OpenVMS platforms
6603 *Matt Caswell and Richard Levitte*
6605 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6607 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6608 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6609 dereference. This could lead to a Denial Of Service attack. Thanks to
6610 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6615 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6616 dtls1_buffer_record function under certain conditions. In particular this
6617 could occur if an attacker sent repeated DTLS records with the same
6618 sequence number but for the next epoch. The memory leak could be exploited
6619 by an attacker in a Denial of Service attack through memory exhaustion.
6620 Thanks to Chris Mueller for reporting this issue.
6625 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6626 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6627 method would be set to NULL which could later result in a NULL pointer
6628 dereference. Thanks to Frank Schmirler for reporting this issue.
6633 * Abort handshake if server key exchange message is omitted for ephemeral
6636 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6637 reporting this issue.
6642 * Remove non-export ephemeral RSA code on client and server. This code
6643 violated the TLS standard by allowing the use of temporary RSA keys in
6644 non-export ciphersuites and could be used by a server to effectively
6645 downgrade the RSA key length used to a value smaller than the server
6646 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6647 INRIA or reporting this issue.
6652 * Fixed issue where DH client certificates are accepted without verification.
6653 An OpenSSL server will accept a DH certificate for client authentication
6654 without the certificate verify message. This effectively allows a client to
6655 authenticate without the use of a private key. This only affects servers
6656 which trust a client certificate authority which issues certificates
6657 containing DH keys: these are extremely rare and hardly ever encountered.
6658 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6664 * Ensure that the session ID context of an SSL is updated when its
6665 SSL_CTX is updated via SSL_set_SSL_CTX.
6667 The session ID context is typically set from the parent SSL_CTX,
6668 and can vary with the CTX.
6672 * Fix various certificate fingerprint issues.
6674 By using non-DER or invalid encodings outside the signed portion of a
6675 certificate the fingerprint can be changed without breaking the signature.
6676 Although no details of the signed portion of the certificate can be changed
6677 this can cause problems with some applications: e.g. those using the
6678 certificate fingerprint for blacklists.
6680 1. Reject signatures with non zero unused bits.
6682 If the BIT STRING containing the signature has non zero unused bits reject
6683 the signature. All current signature algorithms require zero unused bits.
6685 2. Check certificate algorithm consistency.
6687 Check the AlgorithmIdentifier inside TBS matches the one in the
6688 certificate signature. NB: this will result in signature failure
6689 errors for some broken certificates.
6691 Thanks to Konrad Kraszewski from Google for reporting this issue.
6693 3. Check DSA/ECDSA signatures use DER.
6695 Re-encode DSA/ECDSA signatures and compare with the original received
6696 signature. Return an error if there is a mismatch.
6698 This will reject various cases including garbage after signature
6699 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6700 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6701 (negative or with leading zeroes).
6703 Further analysis was conducted and fixes were developed by Stephen Henson
6704 of the OpenSSL core team.
6710 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6711 results on some platforms, including x86_64. This bug occurs at random
6712 with a very low probability, and is not known to be exploitable in any
6713 way, though its exact impact is difficult to determine. Thanks to Pieter
6714 Wuille (Blockstream) who reported this issue and also suggested an initial
6715 fix. Further analysis was conducted by the OpenSSL development team and
6716 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6717 the OpenSSL core team.
6722 * Do not resume sessions on the server if the negotiated protocol
6723 version does not match the session's version. Resuming with a different
6724 version, while not strictly forbidden by the RFC, is of questionable
6725 sanity and breaks all known clients.
6727 *David Benjamin, Emilia Käsper*
6729 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6730 early CCS messages during renegotiation. (Note that because
6731 renegotiation is encrypted, this early CCS was not exploitable.)
6735 * Tighten client-side session ticket handling during renegotiation:
6736 ensure that the client only accepts a session ticket if the server sends
6737 the extension anew in the ServerHello. Previously, a TLS client would
6738 reuse the old extension state and thus accept a session ticket if one was
6739 announced in the initial ServerHello.
6741 Similarly, ensure that the client requires a session ticket if one
6742 was advertised in the ServerHello. Previously, a TLS client would
6743 ignore a missing NewSessionTicket message.
6747 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6751 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6752 sends a carefully crafted handshake message, to cause OpenSSL to fail
6753 to free up to 64k of memory causing a memory leak. This could be
6754 exploited in a Denial Of Service attack. This issue affects OpenSSL
6755 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6756 whether SRTP is used or configured. Implementations of OpenSSL that
6757 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6759 The fix was developed by the OpenSSL team.
6764 * Session Ticket Memory Leak.
6766 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6767 integrity of that ticket is first verified. In the event of a session
6768 ticket integrity check failing, OpenSSL will fail to free memory
6769 causing a memory leak. By sending a large number of invalid session
6770 tickets an attacker could exploit this issue in a Denial Of Service
6776 * Build option no-ssl3 is incomplete.
6778 When OpenSSL is configured with "no-ssl3" as a build option, servers
6779 could accept and complete a SSL 3.0 handshake, and clients could be
6780 configured to send them.
6783 *Akamai and the OpenSSL team*
6785 * Add support for TLS_FALLBACK_SCSV.
6786 Client applications doing fallback retries should call
6787 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6790 *Adam Langley, Bodo Moeller*
6792 * Add additional DigestInfo checks.
6794 Re-encode DigestInto in DER and check against the original when
6795 verifying RSA signature: this will reject any improperly encoded
6796 DigestInfo structures.
6798 Note: this is a precautionary measure and no attacks are currently known.
6802 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6804 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6805 SRP code can be overrun an internal buffer. Add sanity check that
6806 g, A, B < N to SRP code.
6808 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6809 Group for discovering this issue.
6814 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6815 TLS 1.0 instead of higher protocol versions when the ClientHello message
6816 is badly fragmented. This allows a man-in-the-middle attacker to force a
6817 downgrade to TLS 1.0 even if both the server and the client support a
6818 higher protocol version, by modifying the client's TLS records.
6820 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6821 researching this issue.
6826 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6827 to a denial of service attack. A malicious server can crash the client
6828 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6829 ciphersuite and sending carefully crafted handshake messages.
6831 Thanks to Felix Gröbert (Google) for discovering and researching this
6837 * By sending carefully crafted DTLS packets an attacker could cause openssl
6838 to leak memory. This can be exploited through a Denial of Service attack.
6839 Thanks to Adam Langley for discovering and researching this issue.
6844 * An attacker can force openssl to consume large amounts of memory whilst
6845 processing DTLS handshake messages. This can be exploited through a
6846 Denial of Service attack.
6847 Thanks to Adam Langley for discovering and researching this issue.
6852 * An attacker can force an error condition which causes openssl to crash
6853 whilst processing DTLS packets due to memory being freed twice. This
6854 can be exploited through a Denial of Service attack.
6855 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6861 * If a multithreaded client connects to a malicious server using a resumed
6862 session and the server sends an ec point format extension it could write
6863 up to 255 bytes to freed memory.
6865 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6871 * A malicious server can crash an OpenSSL client with a null pointer
6872 dereference (read) by specifying an SRP ciphersuite even though it was not
6873 properly negotiated with the client. This can be exploited through a
6874 Denial of Service attack.
6876 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6877 discovering and researching this issue.
6882 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6883 X509_name_oneline, X509_name_print_ex et al. to leak some information
6884 from the stack. Applications may be affected if they echo pretty printing
6885 output to the attacker.
6887 Thanks to Ivan Fratric (Google) for discovering this issue.
6890 *Emilia Käsper, and Steve Henson*
6892 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6893 for corner cases. (Certain input points at infinity could lead to
6894 bogus results, with non-infinity inputs mapped to infinity too.)
6898 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6900 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6901 handshake can force the use of weak keying material in OpenSSL
6902 SSL/TLS clients and servers.
6904 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6905 researching this issue. ([CVE-2014-0224])
6907 *KIKUCHI Masashi, Steve Henson*
6909 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6910 OpenSSL DTLS client the code can be made to recurse eventually crashing
6913 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6916 *Imre Rad, Steve Henson*
6918 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6919 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6920 client or server. This is potentially exploitable to run arbitrary
6921 code on a vulnerable client or server.
6923 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6925 *Jüri Aedla, Steve Henson*
6927 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6928 are subject to a denial of service attack.
6930 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6931 this issue. ([CVE-2014-3470])
6933 *Felix Gröbert, Ivan Fratric, Steve Henson*
6935 * Harmonize version and its documentation. -f flag is used to display
6938 *mancha <mancha1@zoho.com>*
6940 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6941 in i2d_ECPrivateKey.
6943 *mancha <mancha1@zoho.com>*
6945 * Fix some double frees. These are not thought to be exploitable.
6947 *mancha <mancha1@zoho.com>*
6949 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6951 * A missing bounds check in the handling of the TLS heartbeat extension
6952 can be used to reveal up to 64k of memory to a connected client or
6955 Thanks for Neel Mehta of Google Security for discovering this bug and to
6956 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6957 preparing the fix ([CVE-2014-0160])
6959 *Adam Langley, Bodo Moeller*
6961 * Fix for the attack described in the paper "Recovering OpenSSL
6962 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6963 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6964 <http://eprint.iacr.org/2014/140>
6966 Thanks to Yuval Yarom and Naomi Benger for discovering this
6967 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6969 *Yuval Yarom and Naomi Benger*
6971 * TLS pad extension: draft-agl-tls-padding-03
6973 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6974 TLS client Hello record length value would otherwise be > 255 and
6975 less that 512 pad with a dummy extension containing zeroes so it
6976 is at least 512 bytes long.
6978 *Adam Langley, Steve Henson*
6980 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6982 * Fix for TLS record tampering bug. A carefully crafted invalid
6983 handshake could crash OpenSSL with a NULL pointer exception.
6984 Thanks to Anton Johansson for reporting this issues.
6987 * Keep original DTLS digest and encryption contexts in retransmission
6988 structures so we can use the previous session parameters if they need
6989 to be resent. ([CVE-2013-6450])
6993 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6994 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6995 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6996 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6997 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6998 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7000 *Rob Stradling, Adam Langley*
7002 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7004 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7005 supporting platforms or when small records were transferred.
7007 *Andy Polyakov, Steve Henson*
7009 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7011 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7013 This addresses the flaw in CBC record processing discovered by
7014 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7015 at: <http://www.isg.rhul.ac.uk/tls/>
7017 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7018 Security Group at Royal Holloway, University of London
7019 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7020 Emilia Käsper for the initial patch.
7023 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7025 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7026 ciphersuites which can be exploited in a denial of service attack.
7027 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7028 and detecting this bug and to Wolfgang Ettlinger
7029 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7034 * Return an error when checking OCSP signatures when key is NULL.
7035 This fixes a DoS attack. ([CVE-2013-0166])
7039 * Make openssl verify return errors.
7041 *Chris Palmer <palmer@google.com> and Ben Laurie*
7043 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7044 the right response is stapled. Also change SSL_get_certificate()
7045 so it returns the certificate actually sent.
7046 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7048 *Rob Stradling <rob.stradling@comodo.com>*
7050 * Fix possible deadlock when decoding public keys.
7054 * Don't use TLS 1.0 record version number in initial client hello
7059 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7061 * Sanity check record length before skipping explicit IV in TLS
7062 1.2, 1.1 and DTLS to fix DoS attack.
7064 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7065 fuzzing as a service testing platform.
7070 * Initialise tkeylen properly when encrypting CMS messages.
7071 Thanks to Solar Designer of Openwall for reporting this issue.
7075 * In FIPS mode don't try to use composite ciphers as they are not
7080 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7082 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7083 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7084 mean any application compiled against OpenSSL 1.0.0 headers setting
7085 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7086 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7087 0x10000000L Any application which was previously compiled against
7088 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7089 will need to be recompiled as a result. Letting be results in
7090 inability to disable specifically TLS 1.1 and in client context,
7091 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7095 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7096 disable just protocol X, but all protocols above X *if* there are
7097 protocols *below* X still enabled. In more practical terms it means
7098 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7099 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7100 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7105 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7107 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7108 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7109 in CRYPTO_realloc_clean.
7111 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7112 issue and to Adam Langley <agl@chromium.org> for fixing it.
7115 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7117 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7121 * Workarounds for some broken servers that "hang" if a client hello
7122 record length exceeds 255 bytes.
7124 1. Do not use record version number > TLS 1.0 in initial client
7125 hello: some (but not all) hanging servers will now work.
7126 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7127 the number of ciphers sent in the client hello. This should be
7128 set to an even number, such as 50, for example by passing:
7129 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7130 Most broken servers should now work.
7131 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7132 TLS 1.2 client support entirely.
7136 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7140 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7142 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7143 STRING form instead of a DigestInfo.
7147 * The format used for MDC2 RSA signatures is inconsistent between EVP
7148 and the RSA_sign/RSA_verify functions. This was made more apparent when
7149 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7150 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7151 the correct format in RSA_verify so both forms transparently work.
7155 * Some servers which support TLS 1.0 can choke if we initially indicate
7156 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7157 encrypted premaster secret. As a workaround use the maximum permitted
7158 client version in client hello, this should keep such servers happy
7159 and still work with previous versions of OpenSSL.
7163 * Add support for TLS/DTLS heartbeats.
7165 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7167 * Add support for SCTP.
7169 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7171 * Improved PRNG seeding for VOS.
7173 *Paul Green <Paul.Green@stratus.com>*
7175 * Extensive assembler packs updates, most notably:
7177 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7178 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7179 - x86_64: bit-sliced AES implementation;
7180 - ARM: NEON support, contemporary platforms optimizations;
7181 - s390x: z196 support;
7182 - `*`: GHASH and GF(2^m) multiplication implementations;
7186 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7187 (removal of unnecessary code)
7189 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7191 * Add TLS key material exporter from RFC 5705.
7195 * Add DTLS-SRTP negotiation from RFC 5764.
7199 * Add Next Protocol Negotiation,
7200 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7201 disabled with a no-npn flag to config or Configure. Code donated
7204 *Adam Langley <agl@google.com> and Ben Laurie*
7206 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7207 NIST-P256, NIST-P521, with constant-time single point multiplication on
7208 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7209 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7210 Code made available under Apache License version 2.0.
7212 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7213 line to include this in your build of OpenSSL, and run "make depend" (or
7214 "make update"). This enables the following EC_METHODs:
7216 EC_GFp_nistp224_method()
7217 EC_GFp_nistp256_method()
7218 EC_GFp_nistp521_method()
7220 EC_GROUP_new_by_curve_name() will automatically use these (while
7221 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7224 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7226 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7227 all platforms. Move ssize_t definition from e_os.h to the public
7228 header file e_os2.h as it now appears in public header file cms.h
7232 * New -sigopt option to the ca, req and x509 utilities. Additional
7233 signature parameters can be passed using this option and in
7238 * Add RSA PSS signing function. This will generate and set the
7239 appropriate AlgorithmIdentifiers for PSS based on those in the
7240 corresponding EVP_MD_CTX structure. No application support yet.
7244 * Support for companion algorithm specific ASN1 signing routines.
7245 New function ASN1_item_sign_ctx() signs a pre-initialised
7246 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7247 the appropriate parameters.
7251 * Add new algorithm specific ASN1 verification initialisation function
7252 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7253 handling will be the same no matter what EVP_PKEY_METHOD is used.
7254 Add a PSS handler to support verification of PSS signatures: checked
7255 against a number of sample certificates.
7259 * Add signature printing for PSS. Add PSS OIDs.
7261 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7263 * Add algorithm specific signature printing. An individual ASN1 method
7264 can now print out signatures instead of the standard hex dump.
7266 More complex signatures (e.g. PSS) can print out more meaningful
7267 information. Include DSA version that prints out the signature
7272 * Password based recipient info support for CMS library: implementing
7277 * Split password based encryption into PBES2 and PBKDF2 functions. This
7278 neatly separates the code into cipher and PBE sections and is required
7279 for some algorithms that split PBES2 into separate pieces (such as
7280 password based CMS).
7284 * Session-handling fixes:
7285 - Fix handling of connections that are resuming with a session ID,
7286 but also support Session Tickets.
7287 - Fix a bug that suppressed issuing of a new ticket if the client
7288 presented a ticket with an expired session.
7289 - Try to set the ticket lifetime hint to something reasonable.
7290 - Make tickets shorter by excluding irrelevant information.
7291 - On the client side, don't ignore renewed tickets.
7293 *Adam Langley, Bodo Moeller (Google)*
7295 * Fix PSK session representation.
7299 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7301 This work was sponsored by Intel.
7305 * Add GCM support to TLS library. Some custom code is needed to split
7306 the IV between the fixed (from PRF) and explicit (from TLS record)
7307 portions. This adds all GCM ciphersuites supported by RFC5288 and
7308 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7309 add a special AESGCM string for GCM only.
7313 * Expand range of ctrls for AES GCM. Permit setting invocation
7314 field on decrypt and retrieval of invocation field only on encrypt.
7318 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7319 As required by RFC5289 these ciphersuites cannot be used if for
7320 versions of TLS earlier than 1.2.
7324 * For FIPS capable OpenSSL interpret a NULL default public key method
7325 as unset and return the appropriate default but do *not* set the default.
7326 This means we can return the appropriate method in applications that
7327 switch between FIPS and non-FIPS modes.
7331 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7332 ENGINE is used then we cannot handle that in the FIPS module so we
7333 keep original code iff non-FIPS operations are allowed.
7337 * Add -attime option to openssl utilities.
7339 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7341 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7345 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7346 FIPS EC methods unconditionally for now.
7350 * New build option no-ec2m to disable characteristic 2 code.
7354 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7355 all cases can be covered as some introduce binary incompatibilities.
7359 * Redirect RSA operations to FIPS module including keygen,
7360 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7364 * Add similar low level API blocking to ciphers.
7368 * Low level digest APIs are not approved in FIPS mode: any attempt
7369 to use these will cause a fatal error. Applications that *really* want
7370 to use them can use the `private_*` version instead.
7374 * Redirect cipher operations to FIPS module for FIPS builds.
7378 * Redirect digest operations to FIPS module for FIPS builds.
7382 * Update build system to add "fips" flag which will link in fipscanister.o
7383 for static and shared library builds embedding a signature if needed.
7387 * Output TLS supported curves in preference order instead of numerical
7388 order. This is currently hardcoded for the highest order curves first.
7389 This should be configurable so applications can judge speed vs strength.
7393 * Add TLS v1.2 server support for client authentication.
7397 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7402 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7403 FIPS modules versions.
7407 * Add TLS v1.2 client side support for client authentication. Keep cache
7408 of handshake records longer as we don't know the hash algorithm to use
7409 until after the certificate request message is received.
7413 * Initial TLS v1.2 client support. Add a default signature algorithms
7414 extension including all the algorithms we support. Parse new signature
7415 format in client key exchange. Relax some ECC signing restrictions for
7416 TLS v1.2 as indicated in RFC5246.
7420 * Add server support for TLS v1.2 signature algorithms extension. Switch
7421 to new signature format when needed using client digest preference.
7422 All server ciphersuites should now work correctly in TLS v1.2. No client
7423 support yet and no support for client certificates.
7427 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7428 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7429 ciphersuites. At present only RSA key exchange ciphersuites work with
7430 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7431 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7432 and version checking.
7436 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7437 with this defined it will not be affected by any changes to ssl internal
7438 structures. Add several utility functions to allow openssl application
7439 to work with OPENSSL_NO_SSL_INTERN defined.
7443 * A long standing patch to add support for SRP from EdelWeb (Peter
7444 Sylvester and Christophe Renou) was integrated.
7445 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7446 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7449 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7453 * Permit abbreviated handshakes when renegotiating using the function
7454 SSL_renegotiate_abbreviated().
7456 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7458 * Add call to ENGINE_register_all_complete() to
7459 ENGINE_load_builtin_engines(), so some implementations get used
7460 automatically instead of needing explicit application support.
7464 * Add support for TLS key exporter as described in RFC5705.
7466 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7468 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7469 a few changes are required:
7471 Add SSL_OP_NO_TLSv1_1 flag.
7472 Add TLSv1_1 methods.
7473 Update version checking logic to handle version 1.1.
7474 Add explicit IV handling (ported from DTLS code).
7475 Add command line options to s_client/s_server.
7482 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7484 * X509_ATTRIBUTE memory leak
7486 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7487 memory. This structure is used by the PKCS#7 and CMS routines so any
7488 application which reads PKCS#7 or CMS data from untrusted sources is
7489 affected. SSL/TLS is not affected.
7491 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7497 * Race condition handling PSK identify hint
7499 If PSK identity hints are received by a multi-threaded client then
7500 the values are wrongly updated in the parent SSL_CTX structure. This can
7501 result in a race condition potentially leading to a double free of the
7507 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7509 * Malformed ECParameters causes infinite loop
7511 When processing an ECParameters structure OpenSSL enters an infinite loop
7512 if the curve specified is over a specially malformed binary polynomial
7515 This can be used to perform denial of service against any
7516 system which processes public keys, certificate requests or
7517 certificates. This includes TLS clients and TLS servers with
7518 client authentication enabled.
7520 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7525 * Exploitable out-of-bounds read in X509_cmp_time
7527 X509_cmp_time does not properly check the length of the ASN1_TIME
7528 string and can read a few bytes out of bounds. In addition,
7529 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7532 An attacker can use this to craft malformed certificates and CRLs of
7533 various sizes and potentially cause a segmentation fault, resulting in
7534 a DoS on applications that verify certificates or CRLs. TLS clients
7535 that verify CRLs are affected. TLS clients and servers with client
7536 authentication enabled may be affected if they use custom verification
7539 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7540 independently by Hanno Böck.
7545 * PKCS7 crash with missing EnvelopedContent
7547 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7548 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7549 with missing content and trigger a NULL pointer dereference on parsing.
7551 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7552 structures from untrusted sources are affected. OpenSSL clients and
7553 servers are not affected.
7555 This issue was reported to OpenSSL by Michal Zalewski (Google).
7560 * CMS verify infinite loop with unknown hash function
7562 When verifying a signedData message the CMS code can enter an infinite loop
7563 if presented with an unknown hash function OID. This can be used to perform
7564 denial of service against any system which verifies signedData messages using
7566 This issue was reported to OpenSSL by Johannes Bauer.
7571 * Race condition handling NewSessionTicket
7573 If a NewSessionTicket is received by a multi-threaded client when attempting to
7574 reuse a previous ticket then a race condition can occur potentially leading to
7575 a double free of the ticket data.
7580 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7582 * Segmentation fault in ASN1_TYPE_cmp fix
7584 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7585 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7586 certificate signature algorithm consistency this can be used to crash any
7587 certificate verification operation and exploited in a DoS attack. Any
7588 application which performs certificate verification is vulnerable including
7589 OpenSSL clients and servers which enable client authentication.
7594 * ASN.1 structure reuse memory corruption fix
7596 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7597 memory corruption via an invalid write. Such reuse is and has been
7598 strongly discouraged and is believed to be rare.
7600 Applications that parse structures containing CHOICE or ANY DEFINED BY
7601 components may be affected. Certificate parsing (d2i_X509 and related
7602 functions) are however not affected. OpenSSL clients and servers are
7608 * PKCS7 NULL pointer dereferences fix
7610 The PKCS#7 parsing code does not handle missing outer ContentInfo
7611 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7612 missing content and trigger a NULL pointer dereference on parsing.
7614 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7615 otherwise parse PKCS#7 structures from untrusted sources are
7616 affected. OpenSSL clients and servers are not affected.
7618 This issue was reported to OpenSSL by Michal Zalewski (Google).
7623 * DoS via reachable assert in SSLv2 servers fix
7625 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7626 servers that both support SSLv2 and enable export cipher suites by sending
7627 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7629 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7630 (OpenSSL development team).
7635 * Use After Free following d2i_ECPrivatekey error fix
7637 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7638 could cause a use after free condition. This, in turn, could cause a double
7639 free in several private key parsing functions (such as d2i_PrivateKey
7640 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7641 for applications that receive EC private keys from untrusted
7642 sources. This scenario is considered rare.
7644 This issue was discovered by the BoringSSL project and fixed in their
7650 * X509_to_X509_REQ NULL pointer deref fix
7652 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7653 the certificate key is invalid. This function is rarely used in practice.
7655 This issue was discovered by Brian Carpenter.
7660 * Removed the export ciphers from the DEFAULT ciphers
7664 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7666 * Build fixes for the Windows and OpenVMS platforms
7668 *Matt Caswell and Richard Levitte*
7670 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7672 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7673 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7674 dereference. This could lead to a Denial Of Service attack. Thanks to
7675 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7680 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7681 dtls1_buffer_record function under certain conditions. In particular this
7682 could occur if an attacker sent repeated DTLS records with the same
7683 sequence number but for the next epoch. The memory leak could be exploited
7684 by an attacker in a Denial of Service attack through memory exhaustion.
7685 Thanks to Chris Mueller for reporting this issue.
7690 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7691 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7692 method would be set to NULL which could later result in a NULL pointer
7693 dereference. Thanks to Frank Schmirler for reporting this issue.
7698 * Abort handshake if server key exchange message is omitted for ephemeral
7701 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7702 reporting this issue.
7707 * Remove non-export ephemeral RSA code on client and server. This code
7708 violated the TLS standard by allowing the use of temporary RSA keys in
7709 non-export ciphersuites and could be used by a server to effectively
7710 downgrade the RSA key length used to a value smaller than the server
7711 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7712 INRIA or reporting this issue.
7717 * Fixed issue where DH client certificates are accepted without verification.
7718 An OpenSSL server will accept a DH certificate for client authentication
7719 without the certificate verify message. This effectively allows a client to
7720 authenticate without the use of a private key. This only affects servers
7721 which trust a client certificate authority which issues certificates
7722 containing DH keys: these are extremely rare and hardly ever encountered.
7723 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7729 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7730 results on some platforms, including x86_64. This bug occurs at random
7731 with a very low probability, and is not known to be exploitable in any
7732 way, though its exact impact is difficult to determine. Thanks to Pieter
7733 Wuille (Blockstream) who reported this issue and also suggested an initial
7734 fix. Further analysis was conducted by the OpenSSL development team and
7735 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7736 the OpenSSL core team.
7741 * Fix various certificate fingerprint issues.
7743 By using non-DER or invalid encodings outside the signed portion of a
7744 certificate the fingerprint can be changed without breaking the signature.
7745 Although no details of the signed portion of the certificate can be changed
7746 this can cause problems with some applications: e.g. those using the
7747 certificate fingerprint for blacklists.
7749 1. Reject signatures with non zero unused bits.
7751 If the BIT STRING containing the signature has non zero unused bits reject
7752 the signature. All current signature algorithms require zero unused bits.
7754 2. Check certificate algorithm consistency.
7756 Check the AlgorithmIdentifier inside TBS matches the one in the
7757 certificate signature. NB: this will result in signature failure
7758 errors for some broken certificates.
7760 Thanks to Konrad Kraszewski from Google for reporting this issue.
7762 3. Check DSA/ECDSA signatures use DER.
7764 Reencode DSA/ECDSA signatures and compare with the original received
7765 signature. Return an error if there is a mismatch.
7767 This will reject various cases including garbage after signature
7768 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7769 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7770 (negative or with leading zeroes).
7772 Further analysis was conducted and fixes were developed by Stephen Henson
7773 of the OpenSSL core team.
7779 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7781 * Session Ticket Memory Leak.
7783 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7784 integrity of that ticket is first verified. In the event of a session
7785 ticket integrity check failing, OpenSSL will fail to free memory
7786 causing a memory leak. By sending a large number of invalid session
7787 tickets an attacker could exploit this issue in a Denial Of Service
7793 * Build option no-ssl3 is incomplete.
7795 When OpenSSL is configured with "no-ssl3" as a build option, servers
7796 could accept and complete a SSL 3.0 handshake, and clients could be
7797 configured to send them.
7800 *Akamai and the OpenSSL team*
7802 * Add support for TLS_FALLBACK_SCSV.
7803 Client applications doing fallback retries should call
7804 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7807 *Adam Langley, Bodo Moeller*
7809 * Add additional DigestInfo checks.
7811 Reencode DigestInto in DER and check against the original when
7812 verifying RSA signature: this will reject any improperly encoded
7813 DigestInfo structures.
7815 Note: this is a precautionary measure and no attacks are currently known.
7819 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7821 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7822 to a denial of service attack. A malicious server can crash the client
7823 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7824 ciphersuite and sending carefully crafted handshake messages.
7826 Thanks to Felix Gröbert (Google) for discovering and researching this
7832 * By sending carefully crafted DTLS packets an attacker could cause openssl
7833 to leak memory. This can be exploited through a Denial of Service attack.
7834 Thanks to Adam Langley for discovering and researching this issue.
7839 * An attacker can force openssl to consume large amounts of memory whilst
7840 processing DTLS handshake messages. This can be exploited through a
7841 Denial of Service attack.
7842 Thanks to Adam Langley for discovering and researching this issue.
7847 * An attacker can force an error condition which causes openssl to crash
7848 whilst processing DTLS packets due to memory being freed twice. This
7849 can be exploited through a Denial of Service attack.
7850 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7856 * If a multithreaded client connects to a malicious server using a resumed
7857 session and the server sends an ec point format extension it could write
7858 up to 255 bytes to freed memory.
7860 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7866 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7867 X509_name_oneline, X509_name_print_ex et al. to leak some information
7868 from the stack. Applications may be affected if they echo pretty printing
7869 output to the attacker.
7871 Thanks to Ivan Fratric (Google) for discovering this issue.
7874 *Emilia Käsper, and Steve Henson*
7876 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7877 for corner cases. (Certain input points at infinity could lead to
7878 bogus results, with non-infinity inputs mapped to infinity too.)
7882 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7884 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7885 handshake can force the use of weak keying material in OpenSSL
7886 SSL/TLS clients and servers.
7888 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7889 researching this issue. ([CVE-2014-0224])
7891 *KIKUCHI Masashi, Steve Henson*
7893 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7894 OpenSSL DTLS client the code can be made to recurse eventually crashing
7897 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7900 *Imre Rad, Steve Henson*
7902 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7903 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7904 client or server. This is potentially exploitable to run arbitrary
7905 code on a vulnerable client or server.
7907 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7909 *Jüri Aedla, Steve Henson*
7911 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7912 are subject to a denial of service attack.
7914 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7915 this issue. ([CVE-2014-3470])
7917 *Felix Gröbert, Ivan Fratric, Steve Henson*
7919 * Harmonize version and its documentation. -f flag is used to display
7922 *mancha <mancha1@zoho.com>*
7924 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7925 in i2d_ECPrivateKey.
7927 *mancha <mancha1@zoho.com>*
7929 * Fix some double frees. These are not thought to be exploitable.
7931 *mancha <mancha1@zoho.com>*
7933 * Fix for the attack described in the paper "Recovering OpenSSL
7934 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7935 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7936 <http://eprint.iacr.org/2014/140>
7938 Thanks to Yuval Yarom and Naomi Benger for discovering this
7939 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7941 *Yuval Yarom and Naomi Benger*
7943 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7945 * Keep original DTLS digest and encryption contexts in retransmission
7946 structures so we can use the previous session parameters if they need
7947 to be resent. ([CVE-2013-6450])
7951 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7952 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7953 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7954 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7955 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7956 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7958 *Rob Stradling, Adam Langley*
7960 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7962 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7964 This addresses the flaw in CBC record processing discovered by
7965 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7966 at: <http://www.isg.rhul.ac.uk/tls/>
7968 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7969 Security Group at Royal Holloway, University of London
7970 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7971 Emilia Käsper for the initial patch.
7974 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7976 * Return an error when checking OCSP signatures when key is NULL.
7977 This fixes a DoS attack. ([CVE-2013-0166])
7981 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7982 the right response is stapled. Also change SSL_get_certificate()
7983 so it returns the certificate actually sent.
7984 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7985 (This is a backport)
7987 *Rob Stradling <rob.stradling@comodo.com>*
7989 * Fix possible deadlock when decoding public keys.
7993 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7995 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7998 * Sanity check record length before skipping explicit IV in DTLS
8001 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8002 fuzzing as a service testing platform.
8007 * Initialise tkeylen properly when encrypting CMS messages.
8008 Thanks to Solar Designer of Openwall for reporting this issue.
8012 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8014 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8015 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8016 in CRYPTO_realloc_clean.
8018 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8019 issue and to Adam Langley <agl@chromium.org> for fixing it.
8022 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8024 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8026 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8027 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8028 content decryption and always return the same error. Note: this attack
8029 needs on average 2^20 messages so it only affects automated senders. The
8030 old behaviour can be re-enabled in the CMS code by setting the
8031 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8032 an MMA defence is not necessary.
8033 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8034 this issue. ([CVE-2012-0884])
8038 * Fix CVE-2011-4619: make sure we really are receiving a
8039 client hello before rejecting multiple SGC restarts. Thanks to
8040 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8044 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8046 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8047 Thanks to Antonio Martin, Enterprise Secure Access Research and
8048 Development, Cisco Systems, Inc. for discovering this bug and
8049 preparing a fix. ([CVE-2012-0050])
8053 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8055 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8056 of the Vaudenay padding oracle attack on CBC mode encryption
8057 which enables an efficient plaintext recovery attack against
8058 the OpenSSL implementation of DTLS. Their attack exploits timing
8059 differences arising during decryption processing. A research
8060 paper describing this attack can be found at:
8061 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8062 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8063 Security Group at Royal Holloway, University of London
8064 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8065 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8066 for preparing the fix. ([CVE-2011-4108])
8068 *Robin Seggelmann, Michael Tuexen*
8070 * Clear bytes used for block padding of SSL 3.0 records.
8073 *Adam Langley (Google)*
8075 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8076 Kadianakis <desnacked@gmail.com> for discovering this issue and
8077 Adam Langley for preparing the fix. ([CVE-2011-4619])
8079 *Adam Langley (Google)*
8081 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8083 *Andrey Kulikov <amdeich@gmail.com>*
8085 * Prevent malformed RFC3779 data triggering an assertion failure.
8086 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8087 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8089 *Rob Austein <sra@hactrn.net>*
8091 * Improved PRNG seeding for VOS.
8093 *Paul Green <Paul.Green@stratus.com>*
8095 * Fix ssl_ciph.c set-up race.
8097 *Adam Langley (Google)*
8099 * Fix spurious failures in ecdsatest.c.
8101 *Emilia Käsper (Google)*
8103 * Fix the BIO_f_buffer() implementation (which was mixing different
8104 interpretations of the `..._len` fields).
8106 *Adam Langley (Google)*
8108 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8109 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8110 threads won't reuse the same blinding coefficients.
8112 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8113 lock to call BN_BLINDING_invert_ex, and avoids one use of
8114 BN_BLINDING_update for each BN_BLINDING structure (previously,
8115 the last update always remained unused).
8117 *Emilia Käsper (Google)*
8119 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8121 *Bob Buckholz (Google)*
8123 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8125 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8126 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8128 *Kaspar Brand <ossl@velox.ch>*
8130 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8131 for multi-threaded use of ECDH. ([CVE-2011-3210])
8133 *Adam Langley (Google)*
8135 * Fix x509_name_ex_d2i memory leak on bad inputs.
8139 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8140 signature public key algorithm by using OID xref utilities instead.
8141 Before this you could only use some ECC ciphersuites with SHA1 only.
8145 * Add protection against ECDSA timing attacks as mentioned in the paper
8146 by Billy Bob Brumley and Nicola Tuveri, see:
8147 <http://eprint.iacr.org/2011/232.pdf>
8149 *Billy Bob Brumley and Nicola Tuveri*
8151 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8153 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8155 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8157 * Fix bug in string printing code: if *any* escaping is enabled we must
8158 escape the escape character (backslash) or the resulting string is
8163 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8165 * Disable code workaround for ancient and obsolete Netscape browsers
8166 and servers: an attacker can use it in a ciphersuite downgrade attack.
8167 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8171 * Fixed J-PAKE implementation error, originally discovered by
8172 Sebastien Martini, further info and confirmation from Stefan
8173 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8177 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8179 * Fix extension code to avoid race conditions which can result in a buffer
8180 overrun vulnerability: resumed sessions must not be modified as they can
8181 be shared by multiple threads. CVE-2010-3864
8185 * Fix WIN32 build system to correctly link an ENGINE directory into
8190 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8192 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8195 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8197 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8199 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8200 context. The operation can be customised via the ctrl mechanism in
8201 case ENGINEs want to include additional functionality.
8205 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8209 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8210 output hashes compatible with older versions of OpenSSL.
8212 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8214 * Fix compression algorithm handling: if resuming a session use the
8215 compression algorithm of the resumed session instead of determining
8216 it from client hello again. Don't allow server to change algorithm.
8220 * Add load_crls() function to commands tidying load_certs() too. Add option
8221 to verify utility to allow additional CRLs to be included.
8225 * Update OCSP request code to permit adding custom headers to the request:
8226 some responders need this.
8230 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8233 *Julia Lawall <julia@diku.dk>*
8235 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8236 needlessly dereferenced structures, used obsolete functions and
8237 didn't handle all updated verify codes correctly.
8241 * Disable MD2 in the default configuration.
8245 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8246 indicate the initial BIO being pushed or popped. This makes it possible
8247 to determine whether the BIO is the one explicitly called or as a result
8248 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8249 it handles reference counts correctly and doesn't zero out the I/O bio
8250 when it is not being explicitly popped. WARNING: applications which
8251 included workarounds for the old buggy behaviour will need to be modified
8252 or they could free up already freed BIOs.
8256 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8257 renaming to all platforms (within the 0.9.8 branch, this was
8258 done conditionally on Netware platforms to avoid a name clash).
8260 *Guenter <lists@gknw.net>*
8262 * Add ECDHE and PSK support to DTLS.
8264 *Michael Tuexen <tuexen@fh-muenster.de>*
8266 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8271 * Add "missing" function EVP_MD_flags() (without this the only way to
8272 retrieve a digest flags is by accessing the structure directly. Update
8273 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8274 or cipher is registered as in the "from" argument. Print out all
8275 registered digests in the dgst usage message instead of manually
8276 attempting to work them out.
8280 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8281 this allows the use of compression and extensions. Change default cipher
8282 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8283 by default unless an application cipher string requests it.
8287 * Alter match criteria in PKCS12_parse(). It used to try to use local
8288 key ids to find matching certificates and keys but some PKCS#12 files
8289 don't follow the (somewhat unwritten) rules and this strategy fails.
8290 Now just gather all certificates together and the first private key
8291 then look for the first certificate that matches the key.
8295 * Support use of registered digest and cipher names for dgst and cipher
8296 commands instead of having to add each one as a special case. So now
8303 openssl dgst -sha256 foo
8305 and this works for ENGINE based algorithms too.
8309 * Update Gost ENGINE to support parameter files.
8311 *Victor B. Wagner <vitus@cryptocom.ru>*
8313 * Support GeneralizedTime in ca utility.
8315 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8317 * Enhance the hash format used for certificate directory links. The new
8318 form uses the canonical encoding (meaning equivalent names will work
8319 even if they aren't identical) and uses SHA1 instead of MD5. This form
8320 is incompatible with the older format and as a result c_rehash should
8321 be used to rebuild symbolic links.
8325 * Make PKCS#8 the default write format for private keys, replacing the
8326 traditional format. This form is standardised, more secure and doesn't
8327 include an implicit MD5 dependency.
8331 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8332 committed to OpenSSL should pass this lot as a minimum.
8336 * Add session ticket override functionality for use by EAP-FAST.
8338 *Jouni Malinen <j@w1.fi>*
8340 * Modify HMAC functions to return a value. Since these can be implemented
8341 in an ENGINE errors can occur.
8345 * Type-checked OBJ_bsearch_ex.
8349 * Type-checked OBJ_bsearch. Also some constification necessitated
8350 by type-checking. Still to come: TXT_DB, bsearch(?),
8351 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8356 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8357 seconds to a tm structure directly, instead of going through OS
8358 specific date routines. This avoids any issues with OS routines such
8359 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8360 and X509_time_adj_ex() to cover the extended range. The existing
8361 X509_time_adj() is still usable and will no longer have any date issues.
8365 * Delta CRL support. New use deltas option which will attempt to locate
8366 and search any appropriate delta CRLs available.
8368 This work was sponsored by Google.
8372 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8373 code and add additional score elements. Validate alternate CRL paths
8374 as part of the CRL checking and indicate a new error "CRL path validation
8375 error" in this case. Applications wanting additional details can use
8376 the verify callback and check the new "parent" field. If this is not
8377 NULL CRL path validation is taking place. Existing applications won't
8378 see this because it requires extended CRL support which is off by
8381 This work was sponsored by Google.
8385 * Support for freshest CRL extension.
8387 This work was sponsored by Google.
8391 * Initial indirect CRL support. Currently only supported in the CRLs
8392 passed directly and not via lookup. Process certificate issuer
8393 CRL entry extension and lookup CRL entries by bother issuer name
8394 and serial number. Check and process CRL issuer entry in IDP extension.
8396 This work was sponsored by Google.
8400 * Add support for distinct certificate and CRL paths. The CRL issuer
8401 certificate is validated separately in this case. Only enabled if
8402 an extended CRL support flag is set: this flag will enable additional
8403 CRL functionality in future.
8405 This work was sponsored by Google.
8409 * Add support for policy mappings extension.
8411 This work was sponsored by Google.
8415 * Fixes to pathlength constraint, self issued certificate handling,
8416 policy processing to align with RFC3280 and PKITS tests.
8418 This work was sponsored by Google.
8422 * Support for name constraints certificate extension. DN, email, DNS
8423 and URI types are currently supported.
8425 This work was sponsored by Google.
8429 * To cater for systems that provide a pointer-based thread ID rather
8430 than numeric, deprecate the current numeric thread ID mechanism and
8431 replace it with a structure and associated callback type. This
8432 mechanism allows a numeric "hash" to be extracted from a thread ID in
8433 either case, and on platforms where pointers are larger than 'long',
8434 mixing is done to help ensure the numeric 'hash' is usable even if it
8435 can't be guaranteed unique. The default mechanism is to use "&errno"
8436 as a pointer-based thread ID to distinguish between threads.
8438 Applications that want to provide their own thread IDs should now use
8439 CRYPTO_THREADID_set_callback() to register a callback that will call
8440 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8442 Note that ERR_remove_state() is now deprecated, because it is tied
8443 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8444 to free the current thread's error state should be replaced by
8445 ERR_remove_thread_state(NULL).
8447 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8448 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8449 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8450 application was previously providing a numeric thread callback that
8451 was inappropriate for distinguishing threads, then uniqueness might
8452 have been obtained with &errno that happened immediately in the
8453 intermediate development versions of OpenSSL; this is no longer the
8454 case, the numeric thread callback will now override the automatic use
8457 *Geoff Thorpe, with help from Bodo Moeller*
8459 * Initial support for different CRL issuing certificates. This covers a
8460 simple case where the self issued certificates in the chain exist and
8461 the real CRL issuer is higher in the existing chain.
8463 This work was sponsored by Google.
8467 * Removed effectively defunct crypto/store from the build.
8471 * Revamp of STACK to provide stronger type-checking. Still to come:
8472 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8473 ASN1_STRING, CONF_VALUE.
8477 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8478 RAM on SSL connections. This option can save about 34k per idle SSL.
8482 * Revamp of LHASH to provide stronger type-checking. Still to come:
8483 STACK, TXT_DB, bsearch, qsort.
8487 * Initial support for Cryptographic Message Syntax (aka CMS) based
8488 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8489 support for data, signedData, compressedData, digestedData and
8490 encryptedData, envelopedData types included. Scripts to check against
8491 RFC4134 examples draft and interop and consistency checks of many
8492 content types and variants.
8496 * Add options to enc utility to support use of zlib compression BIO.
8500 * Extend mk1mf to support importing of options and assembly language
8501 files from Configure script, currently only included in VC-WIN32.
8502 The assembly language rules can now optionally generate the source
8503 files from the associated perl scripts.
8507 * Implement remaining functionality needed to support GOST ciphersuites.
8508 Interop testing has been performed using CryptoPro implementations.
8510 *Victor B. Wagner <vitus@cryptocom.ru>*
8512 * s390x assembler pack.
8516 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8521 * Implement Opaque PRF Input TLS extension as specified in
8522 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8523 official specification yet and no extension type assignment by
8524 IANA exists, this extension (for now) will have to be explicitly
8525 enabled when building OpenSSL by providing the extension number
8526 to use. For example, specify an option
8528 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8530 to the "config" or "Configure" script to enable the extension,
8531 assuming extension number 0x9527 (which is a completely arbitrary
8532 and unofficial assignment based on the MD5 hash of the Internet
8533 Draft). Note that by doing so, you potentially lose
8534 interoperability with other TLS implementations since these might
8535 be using the same extension number for other purposes.
8537 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8538 opaque PRF input value to use in the handshake. This will create
8539 an internal copy of the length-'len' string at 'src', and will
8540 return non-zero for success.
8542 To get more control and flexibility, provide a callback function
8545 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8546 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8550 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8553 Callback function 'cb' will be called in handshakes, and is
8554 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8555 Argument 'arg' is for application purposes (the value as given to
8556 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8557 be provided to the callback function). The callback function
8558 has to return non-zero to report success: usually 1 to use opaque
8559 PRF input just if possible, or 2 to enforce use of the opaque PRF
8560 input. In the latter case, the library will abort the handshake
8561 if opaque PRF input is not successfully negotiated.
8563 Arguments 'peerinput' and 'len' given to the callback function
8564 will always be NULL and 0 in the case of a client. A server will
8565 see the client's opaque PRF input through these variables if
8566 available (NULL and 0 otherwise). Note that if the server
8567 provides an opaque PRF input, the length must be the same as the
8568 length of the client's opaque PRF input.
8570 Note that the callback function will only be called when creating
8571 a new session (session resumption can resume whatever was
8572 previously negotiated), and will not be called in SSL 2.0
8573 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8574 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8575 for applications that need to enforce opaque PRF input.
8579 * Update ssl code to support digests other than SHA1+MD5 for handshake
8582 *Victor B. Wagner <vitus@cryptocom.ru>*
8584 * Add RFC4507 support to OpenSSL. This includes the corrections in
8585 RFC4507bis. The encrypted ticket format is an encrypted encoded
8586 SSL_SESSION structure, that way new session features are automatically
8589 If a client application caches session in an SSL_SESSION structure
8590 support is transparent because tickets are now stored in the encoded
8593 The SSL_CTX structure automatically generates keys for ticket
8594 protection in servers so again support should be possible
8595 with no application modification.
8597 If a client or server wishes to disable RFC4507 support then the option
8598 SSL_OP_NO_TICKET can be set.
8600 Add a TLS extension debugging callback to allow the contents of any client
8601 or server extensions to be examined.
8603 This work was sponsored by Google.
8607 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8608 OpenSSL should now compile cleanly on gcc 4.2
8610 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8612 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8613 support including streaming MAC support: this is required for GOST
8614 ciphersuite support.
8616 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8618 * Add option -stream to use PKCS#7 streaming in smime utility. New
8619 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8620 to output in BER and PEM format.
8624 * Experimental support for use of HMAC via EVP_PKEY interface. This
8625 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8626 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8627 ENGINE support for HMAC keys which are unextractable. New -mac and
8628 -macopt options to dgst utility.
8632 * New option -sigopt to dgst utility. Update dgst to use
8633 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8634 alternative signing parameters such as X9.31 or PSS in the dgst
8639 * Change ssl_cipher_apply_rule(), the internal function that does
8640 the work each time a ciphersuite string requests enabling
8641 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8642 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8643 the order of disabled ciphersuites such that those ciphersuites
8644 that most recently went from enabled to disabled not only stay
8645 in order with respect to each other, but also have higher priority
8646 than other disabled ciphersuites the next time ciphersuites are
8649 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8650 the same ciphersuites as with "HIGH" alone, but in a specific
8651 order where the PSK ciphersuites come first (since they are the
8652 most recently disabled ciphersuites when "HIGH" is parsed).
8654 Also, change ssl_create_cipher_list() (using this new
8655 functionality) such that between otherwise identical
8656 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8661 * Change ssl_create_cipher_list() so that it automatically
8662 arranges the ciphersuites in reasonable order before starting
8663 to process the rule string. Thus, the definition for "DEFAULT"
8664 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8665 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8666 This makes it much easier to arrive at a reasonable default order
8667 in applications for which anonymous ciphers are OK (meaning
8668 that you can't actually use DEFAULT).
8670 *Bodo Moeller; suggested by Victor Duchovni*
8672 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8673 processing) into multiple integers instead of setting
8674 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8675 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8676 (These masks as well as the individual bit definitions are hidden
8677 away into the non-exported interface ssl/ssl_locl.h, so this
8678 change to the definition of the SSL_CIPHER structure shouldn't
8679 affect applications.) This give us more bits for each of these
8680 categories, so there is no longer a need to coagulate AES128 and
8681 AES256 into a single algorithm bit, and to coagulate Camellia128
8682 and Camellia256 into a single algorithm bit, which has led to all
8685 Thus, among other things, the kludge introduced in 0.9.7m and
8686 0.9.8e for masking out AES256 independently of AES128 or masking
8687 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8689 With the change, we also introduce new ciphersuite aliases that
8690 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8695 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8696 Use the leftmost N bytes of the signature input if the input is
8697 larger than the prime q (with N being the size in bytes of q).
8701 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8702 it yet and it is largely untested.
8706 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8710 * Initial incomplete changes to avoid need for function casts in OpenSSL
8711 some compilers (gcc 4.2 and later) reject their use. Safestack is
8712 reimplemented. Update ASN1 to avoid use of legacy functions.
8716 * Win32/64 targets are linked with Winsock2.
8720 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8721 to external functions. This can be used to increase CRL handling
8722 efficiency especially when CRLs are very large by (for example) storing
8723 the CRL revoked certificates in a database.
8727 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8728 new CRLs added to a directory can be used. New command line option
8729 -verify_return_error to s_client and s_server. This causes real errors
8730 to be returned by the verify callback instead of carrying on no matter
8731 what. This reflects the way a "real world" verify callback would behave.
8735 * GOST engine, supporting several GOST algorithms and public key formats.
8736 Kindly donated by Cryptocom.
8740 * Partial support for Issuing Distribution Point CRL extension. CRLs
8741 partitioned by DP are handled but no indirect CRL or reason partitioning
8742 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8743 selected via a scoring technique which handles IDP and AKID in CRLs.
8747 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8748 will ultimately be used for all verify operations: this will remove the
8749 X509_STORE dependency on certificate verification and allow alternative
8750 lookup methods. X509_STORE based implementations of these two callbacks.
8754 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8755 Modify get_crl() to find a valid (unexpired) CRL if possible.
8759 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8760 this would be called X509_CRL_cmp() but that name is already used by
8761 a function that just compares CRL issuer names. Cache several CRL
8762 extensions in X509_CRL structure and cache CRLDP in X509.
8766 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8767 this maps equivalent X509_NAME structures into a consistent structure.
8768 Name comparison can then be performed rapidly using memcmp().
8772 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8777 * Allow digests to supply their own micalg string for S/MIME type using
8778 the ctrl EVP_MD_CTRL_MICALG.
8782 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8783 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8784 ctrl. It can then customise the structure before and/or after signing
8789 * New function OBJ_add_sigid() to allow application defined signature OIDs
8790 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8791 to free up any added signature OIDs.
8795 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8796 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8797 digest and cipher tables. New options added to openssl utility:
8798 list-message-digest-algorithms and list-cipher-algorithms.
8802 * Change the array representation of binary polynomials: the list
8803 of degrees of non-zero coefficients is now terminated with -1.
8804 Previously it was terminated with 0, which was also part of the
8805 value; thus, the array representation was not applicable to
8806 polynomials where t^0 has coefficient zero. This change makes
8807 the array representation useful in a more general context.
8811 * Various modifications and fixes to SSL/TLS cipher string
8812 handling. For ECC, the code now distinguishes between fixed ECDH
8813 with RSA certificates on the one hand and with ECDSA certificates
8814 on the other hand, since these are separate ciphersuites. The
8815 unused code for Fortezza ciphersuites has been removed.
8817 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8818 (not "ECDHE"). For consistency with the code for DH
8819 certificates, use of ECDH certificates is now considered ECDH
8820 authentication, not RSA or ECDSA authentication (the latter is
8821 merely the CA's signing algorithm and not actively used in the
8824 The temporary ciphersuite alias "ECCdraft" is no longer
8825 available, and ECC ciphersuites are no longer excluded from "ALL"
8826 and "DEFAULT". The following aliases now exist for RFC 4492
8827 ciphersuites, most of these by analogy with the DH case:
8829 kECDHr - ECDH cert, signed with RSA
8830 kECDHe - ECDH cert, signed with ECDSA
8831 kECDH - ECDH cert (signed with either RSA or ECDSA)
8832 kEECDH - ephemeral ECDH
8833 ECDH - ECDH cert or ephemeral ECDH
8839 AECDH - anonymous ECDH
8840 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8844 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8845 Use correct micalg parameters depending on digest(s) in signed message.
8849 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8850 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8854 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8855 an engine to register a method. Add ENGINE lookups for methods and
8856 functional reference processing.
8860 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8861 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8866 * New -resign option to smime utility. This adds one or more signers
8867 to an existing PKCS#7 signedData structure. Also -md option to use an
8868 alternative message digest algorithm for signing.
8872 * Tidy up PKCS#7 routines and add new functions to make it easier to
8873 create PKCS7 structures containing multiple signers. Update smime
8874 application to support multiple signers.
8878 * New -macalg option to pkcs12 utility to allow setting of an alternative
8883 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8884 Reorganize PBE internals to lookup from a static table using NIDs,
8885 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8886 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8887 PRF which will be automatically used with PBES2.
8891 * Replace the algorithm specific calls to generate keys in "req" with the
8896 * Update PKCS#7 enveloped data routines to use new API. This is now
8897 supported by any public key method supporting the encrypt operation. A
8898 ctrl is added to allow the public key algorithm to examine or modify
8899 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8904 * Add a ctrl to asn1 method to allow a public key algorithm to express
8905 a default digest type to use. In most cases this will be SHA1 but some
8906 algorithms (such as GOST) need to specify an alternative digest. The
8907 return value indicates how strong the preference is 1 means optional and
8908 2 is mandatory (that is it is the only supported type). Modify
8909 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8910 use the default md. Update openssl utilities to use the default digest
8911 type for signing if it is not explicitly indicated.
8915 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8916 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8917 signing method from the key type. This effectively removes the link
8918 between digests and public key types.
8922 * Add an OID cross reference table and utility functions. Its purpose is to
8923 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8924 rsaEncryption. This will allow some of the algorithm specific hackery
8925 needed to use the correct OID to be removed.
8929 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8930 structures for PKCS7_sign(). They are now set up by the relevant public
8935 * Add provisional EC pkey method with support for ECDSA and ECDH.
8939 * Add support for key derivation (agreement) in the API, DH method and
8944 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8945 public and private key formats. As a side effect these add additional
8946 command line functionality not previously available: DSA signatures can be
8947 generated and verified using pkeyutl and DH key support and generation in
8954 *Oliver Tappe <zooey@hirschkaefer.de>*
8956 * New make target "install_html_docs" installs HTML renditions of the
8959 *Oliver Tappe <zooey@hirschkaefer.de>*
8961 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8962 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8963 support key and parameter generation and add initial key generation
8964 functionality for RSA.
8968 * Add functions for main EVP_PKEY_method operations. The undocumented
8969 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8970 `EVP_PKEY_{encrypt,decrypt}_old`.
8974 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8975 key API, doesn't do much yet.
8979 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8980 public key algorithms. New option to openssl utility:
8981 "list-public-key-algorithms" to print out info.
8985 * Implement the Supported Elliptic Curves Extension for
8986 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8990 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8991 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8995 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8996 utilities such as rsa, dsa, dsaparam etc except they process any key
9001 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9002 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9003 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9008 * Initial support for pluggable public key ASN1.
9009 De-spaghettify the public key ASN1 handling. Move public and private
9010 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9011 algorithm specific handling to a single module within the relevant
9012 algorithm directory. Add functions to allow (near) opaque processing
9013 of public and private key structures.
9017 * Implement the Supported Point Formats Extension for
9018 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9022 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9023 for the psk identity [hint] and the psk callback functions to the
9024 SSL_SESSION, SSL and SSL_CTX structure.
9027 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9031 SSL_CTX_use_psk_identity_hint
9032 SSL_get_psk_identity_hint
9033 SSL_get_psk_identity
9034 SSL_use_psk_identity_hint
9036 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9038 * Add RFC 3161 compliant time stamp request creation, response generation
9039 and response verification functionality.
9041 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9043 * Add initial support for TLS extensions, specifically for the server_name
9044 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9045 have new members for a host name. The SSL data structure has an
9046 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9047 stored in that context to allow for session resumption, even after the
9048 SSL has been switched to a new SSL_CTX in reaction to a client's
9049 server_name extension.
9051 New functions (subject to change):
9053 SSL_get_servername()
9054 SSL_get_servername_type()
9057 New CTRL codes and macros (subject to change):
9059 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9060 - SSL_CTX_set_tlsext_servername_callback()
9061 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9062 - SSL_CTX_set_tlsext_servername_arg()
9063 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9065 openssl s_client has a new '-servername ...' option.
9067 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9068 '-key2 ...', '-servername_fatal' (subject to change). This allows
9069 testing the HostName extension for a specific single host name ('-cert'
9070 and '-key' remain fallbacks for handshakes without HostName
9071 negotiation). If the unrecognized_name alert has to be sent, this by
9072 default is a warning; it becomes fatal with the '-servername_fatal'
9075 *Peter Sylvester, Remy Allais, Christophe Renou*
9077 * Whirlpool hash implementation is added.
9081 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9082 bn(64,32). Because of instruction set limitations it doesn't have
9083 any negative impact on performance. This was done mostly in order
9084 to make it possible to share assembler modules, such as bn_mul_mont
9085 implementations, between 32- and 64-bit builds without hassle.
9089 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9090 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9095 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9096 dedicated Montgomery multiplication procedure, is introduced.
9097 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9098 "64-bit" performance on certain 32-bit targets.
9102 * New option SSL_OP_NO_COMP to disable use of compression selectively
9103 in SSL structures. New SSL ctrl to set maximum send fragment size.
9104 Save memory by setting the I/O buffer sizes dynamically instead of
9105 using the maximum available value.
9109 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9110 in addition to the text details.
9114 * Very, very preliminary EXPERIMENTAL support for printing of general
9115 ASN1 structures. This currently produces rather ugly output and doesn't
9116 handle several customised structures at all.
9120 * Integrated support for PVK file format and some related formats such
9121 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9122 these in the 'rsa' and 'dsa' utilities.
9126 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9130 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9131 place for the (very old) "NETSCAPE" format certificates which are now
9132 handled using new ASN1 code equivalents.
9136 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9137 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9138 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9142 * Modify CRL distribution points extension code to print out previously
9143 unsupported fields. Enhance extension setting code to allow setting of
9148 * Add print and set support for Issuing Distribution Point CRL extension.
9152 * Change 'Configure' script to enable Camellia by default.
9159 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9161 * When rejecting SSL/TLS records due to an incorrect version number, never
9162 update s->server with a new major version number. As of
9163 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9164 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9165 the previous behavior could result in a read attempt at NULL when
9166 receiving specific incorrect SSL/TLS records once record payload
9167 protection is active. ([CVE-2010-0740])
9169 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9171 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9172 could be crashed if the relevant tables were not present (e.g. chrooted).
9174 *Tomas Hoger <thoger@redhat.com>*
9176 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9178 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9180 *Martin Olsson, Neel Mehta*
9182 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9183 accommodate for stack sorting, always a write lock!).
9187 * On some versions of WIN32 Heap32Next is very slow. This can cause
9188 excessive delays in the RAND_poll(): over a minute. As a workaround
9189 include a time check in the inner Heap32Next loop too.
9193 * The code that handled flushing of data in SSL/TLS originally used the
9194 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9195 the problem outlined in PR#1949. The fix suggested there however can
9196 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9197 of Apache). So instead simplify the code to flush unconditionally.
9198 This should be fine since flushing with no data to flush is a no op.
9202 * Handle TLS versions 2.0 and later properly and correctly use the
9203 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9204 off ancient servers have a habit of sticking around for a while...
9208 * Modify compression code so it frees up structures without using the
9209 ex_data callbacks. This works around a problem where some applications
9210 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9211 restarting) then use compression (e.g. SSL with compression) later.
9212 This results in significant per-connection memory leaks and
9213 has caused some security issues including CVE-2008-1678 and
9218 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9219 change when encrypting or decrypting.
9223 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9224 connect and renegotiate with servers which do not support RI.
9225 Until RI is more widely deployed this option is enabled by default.
9229 * Add "missing" ssl ctrls to clear options and mode.
9233 * If client attempts to renegotiate and doesn't support RI respond with
9234 a no_renegotiation alert as required by RFC5746. Some renegotiating
9235 TLS clients will continue a connection gracefully when they receive
9236 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9237 waiting for a server hello which it will never receive. Now we treat a
9238 received no_renegotiation alert as a fatal error. This is because
9239 applications requesting a renegotiation might well expect it to succeed
9240 and would have no code in place to handle the server denying it so the
9241 only safe thing to do is to terminate the connection.
9245 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9246 peer supports secure renegotiation and 0 otherwise. Print out peer
9247 renegotiation support in s_client/s_server.
9251 * Replace the highly broken and deprecated SPKAC certification method with
9252 the updated NID creation version. This should correctly handle UTF8.
9256 * Implement RFC5746. Re-enable renegotiation but require the extension
9257 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9258 turns out to be a bad idea. It has been replaced by
9259 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9260 SSL_CTX_set_options(). This is really not recommended unless you
9261 know what you are doing.
9263 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9265 * Fixes to stateless session resumption handling. Use initial_ctx when
9266 issuing and attempting to decrypt tickets in case it has changed during
9267 servername handling. Use a non-zero length session ID when attempting
9268 stateless session resumption: this makes it possible to determine if
9269 a resumption has occurred immediately after receiving server hello
9270 (several places in OpenSSL subtly assume this) instead of later in
9275 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9276 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9277 fixes for a few places where the return code is not checked
9280 *Julia Lawall <julia@diku.dk>*
9282 * Add --strict-warnings option to Configure script to include devteam
9283 warnings in other configurations.
9287 * Add support for --libdir option and LIBDIR variable in makefiles. This
9288 makes it possible to install openssl libraries in locations which
9289 have names other than "lib", for example "/usr/lib64" which some
9292 *Steve Henson, based on patch from Jeremy Utley*
9294 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9295 X690 8.9.12 and can produce some misleading textual output of OIDs.
9297 *Steve Henson, reported by Dan Kaminsky*
9299 * Delete MD2 from algorithm tables. This follows the recommendation in
9300 several standards that it is not used in new applications due to
9301 several cryptographic weaknesses. For binary compatibility reasons
9302 the MD2 API is still compiled in by default.
9306 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9311 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9312 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9315 *Guenter <lists@gknw.net>*
9317 * Fix the server certificate chain building code to use X509_verify_cert(),
9318 it used to have an ad-hoc builder which was unable to cope with anything
9319 other than a simple chain.
9321 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9323 * Don't check self signed certificate signatures in X509_verify_cert()
9324 by default (a flag can override this): it just wastes time without
9325 adding any security. As a useful side effect self signed root CAs
9326 with non-FIPS digests are now usable in FIPS mode.
9330 * In dtls1_process_out_of_seq_message() the check if the current message
9331 is already buffered was missing. For every new message was memory
9332 allocated, allowing an attacker to perform an denial of service attack
9333 with sending out of seq handshake messages until there is no memory
9334 left. Additionally every future message was buffered, even if the
9335 sequence number made no sense and would be part of another handshake.
9336 So only messages with sequence numbers less than 10 in advance will be
9337 buffered. ([CVE-2009-1378])
9339 *Robin Seggelmann, discovered by Daniel Mentz*
9341 * Records are buffered if they arrive with a future epoch to be
9342 processed after finishing the corresponding handshake. There is
9343 currently no limitation to this buffer allowing an attacker to perform
9344 a DOS attack with sending records with future epochs until there is no
9345 memory left. This patch adds the pqueue_size() function to determine
9346 the size of a buffer and limits the record buffer to 100 entries.
9349 *Robin Seggelmann, discovered by Daniel Mentz*
9351 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9352 parent structure is freed. ([CVE-2009-1379])
9356 * Handle non-blocking I/O properly in SSL_shutdown() call.
9358 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9360 * Add `2.5.4.*` OIDs
9362 *Ilya O. <vrghost@gmail.com>*
9364 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9366 * Disable renegotiation completely - this fixes a severe security
9367 problem ([CVE-2009-3555]) at the cost of breaking all
9368 renegotiation. Renegotiation can be re-enabled by setting
9369 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9370 run-time. This is really not recommended unless you know what
9375 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9377 * Don't set val to NULL when freeing up structures, it is freed up by
9378 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9379 zeroing past the valid field. ([CVE-2009-0789])
9381 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9383 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9384 checked correctly. This would allow some invalid signed attributes to
9385 appear to verify correctly. ([CVE-2009-0591])
9387 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9389 * Reject UniversalString and BMPString types with invalid lengths. This
9390 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9391 a legal length. ([CVE-2009-0590])
9395 * Set S/MIME signing as the default purpose rather than setting it
9396 unconditionally. This allows applications to override it at the store
9401 * Permit restricted recursion of ASN1 strings. This is needed in practice
9402 to handle some structures.
9406 * Improve efficiency of mem_gets: don't search whole buffer each time
9409 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9411 * New -hex option for openssl rand.
9415 * Print out UTF8String and NumericString when parsing ASN1.
9419 * Support NumericString type for name components.
9423 * Allow CC in the environment to override the automatically chosen
9424 compiler. Note that nothing is done to ensure flags work with the
9429 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9431 * Properly check EVP_VerifyFinal() and similar return values
9434 *Ben Laurie, Bodo Moeller, Google Security Team*
9436 * Enable TLS extensions by default.
9440 * Allow the CHIL engine to be loaded, whether the application is
9441 multithreaded or not. (This does not release the developer from the
9442 obligation to set up the dynamic locking callbacks.)
9444 *Sander Temme <sander@temme.net>*
9446 * Use correct exit code if there is an error in dgst command.
9448 *Steve Henson; problem pointed out by Roland Dirlewanger*
9450 * Tweak Configure so that you need to say "experimental-jpake" to enable
9451 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9455 * Add experimental JPAKE support, including demo authentication in
9456 s_client and s_server.
9460 * Set the comparison function in v3_addr_canonize().
9462 *Rob Austein <sra@hactrn.net>*
9464 * Add support for XMPP STARTTLS in s_client.
9466 *Philip Paeps <philip@freebsd.org>*
9468 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9469 to ensure that even with this option, only ciphersuites in the
9470 server's preference list will be accepted. (Note that the option
9471 applies only when resuming a session, so the earlier behavior was
9472 just about the algorithm choice for symmetric cryptography.)
9476 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9478 * Fix NULL pointer dereference if a DTLS server received
9479 ChangeCipherSpec as first record ([CVE-2009-1386]).
9483 * Fix a state transition in s3_srvr.c and d1_srvr.c
9484 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9488 * The fix in 0.9.8c that supposedly got rid of unsafe
9489 double-checked locking was incomplete for RSA blinding,
9490 addressing just one layer of what turns out to have been
9491 doubly unsafe triple-checked locking.
9493 So now fix this for real by retiring the MONT_HELPER macro
9494 in crypto/rsa/rsa_eay.c.
9496 *Bodo Moeller; problem pointed out by Marius Schilder*
9498 * Various precautionary measures:
9500 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9502 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9503 (NB: This would require knowledge of the secret session ticket key
9504 to exploit, in which case you'd be SOL either way.)
9506 - Change bn_nist.c so that it will properly handle input BIGNUMs
9507 outside the expected range.
9509 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9512 *Neel Mehta, Bodo Moeller*
9514 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9515 the load fails. Useful for distros.
9517 *Ben Laurie and the FreeBSD team*
9519 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9523 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9527 * Expand ENGINE to support engine supplied SSL client certificate functions.
9529 This work was sponsored by Logica.
9533 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9534 keystores. Support for SSL/TLS client authentication too.
9535 Not compiled unless enable-capieng specified to Configure.
9537 This work was sponsored by Logica.
9541 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9542 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9543 attribute creation routines such as certificate requests and PKCS#12
9548 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9550 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9551 handshake which could lead to a client crash as found using the
9552 Codenomicon TLS test suite ([CVE-2008-1672])
9554 *Steve Henson, Mark Cox*
9556 * Fix double free in TLS server name extensions which could lead to
9557 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9561 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9563 Clear the error queue to ensure that error entries left from
9564 older function calls do not interfere with the correct operation.
9566 *Lutz Jaenicke, Erik de Castro Lopo*
9568 * Remove root CA certificates of commercial CAs:
9570 The OpenSSL project does not recommend any specific CA and does not
9571 have any policy with respect to including or excluding any CA.
9572 Therefore it does not make any sense to ship an arbitrary selection
9573 of root CA certificates with the OpenSSL software.
9577 * RSA OAEP patches to fix two separate invalid memory reads.
9578 The first one involves inputs when 'lzero' is greater than
9579 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9580 before the beginning of from). The second one involves inputs where
9581 the 'db' section contains nothing but zeroes (there is a one-byte
9582 invalid read after the end of 'db').
9584 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9586 * Partial backport from 0.9.9-dev:
9588 Introduce bn_mul_mont (dedicated Montgomery multiplication
9589 procedure) as a candidate for BIGNUM assembler implementation.
9590 While 0.9.9-dev uses assembler for various architectures, only
9591 x86_64 is available by default here in the 0.9.8 branch, and
9592 32-bit x86 is available through a compile-time setting.
9594 To try the 32-bit x86 assembler implementation, use Configure
9595 option "enable-montasm" (which exists only for this backport).
9597 As "enable-montasm" for 32-bit x86 disclaims code stability
9598 anyway, in this constellation we activate additional code
9599 backported from 0.9.9-dev for further performance improvements,
9600 namely BN_from_montgomery_word. (To enable this otherwise,
9601 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9603 *Andy Polyakov (backport partially by Bodo Moeller)*
9605 * Add TLS session ticket callback. This allows an application to set
9606 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9607 values. This is useful for key rollover for example where several key
9608 sets may exist with different names.
9612 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9613 This was broken until now in 0.9.8 releases, such that the only way
9614 a registered ENGINE could be used (assuming it initialises
9615 successfully on the host) was to explicitly set it as the default
9616 for the relevant algorithms. This is in contradiction with 0.9.7
9617 behaviour and the documentation. With this fix, when an ENGINE is
9618 registered into a given algorithm's table of implementations, the
9619 'uptodate' flag is reset so that auto-discovery will be used next
9620 time a new context for that algorithm attempts to select an
9623 *Ian Lister (tweaked by Geoff Thorpe)*
9625 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9626 implementation in the following ways:
9628 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9631 Lack of BER streaming support means one pass streaming processing is
9632 only supported if data is detached: setting the streaming flag is
9633 ignored for embedded content.
9635 CMS support is disabled by default and must be explicitly enabled
9636 with the enable-cms configuration option.
9640 * Update the GMP engine glue to do direct copies between BIGNUM and
9641 mpz_t when openssl and GMP use the same limb size. Otherwise the
9642 existing "conversion via a text string export" trick is still used.
9644 *Paul Sheer <paulsheer@gmail.com>*
9646 * Zlib compression BIO. This is a filter BIO which compressed and
9647 uncompresses any data passed through it.
9651 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9652 RFC3394 compatible AES key wrapping.
9656 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9657 sets string data without copying. X509_ALGOR_set0() and
9658 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9659 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9660 from an X509_ATTRIBUTE structure optionally checking it occurs only
9661 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9666 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9667 to get the expected BN_FLG_CONSTTIME behavior.
9669 *Bodo Moeller (Google)*
9673 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9674 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9675 - added some more tests to do_tests.pl
9676 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9677 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9678 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9679 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9680 - various changes to netware.pl to enable gcc-cross builds on Win32
9682 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9683 - various changes to fix missing prototype warnings
9684 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9685 - added AES, WHIRLPOOL and CPUID assembler code to build files
9686 - added missing AES assembler make rules to mk1mf.pl
9687 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9689 *Guenter Knauf <eflash@gmx.net>*
9691 * Implement certificate status request TLS extension defined in RFC3546.
9692 A client can set the appropriate parameters and receive the encoded
9693 OCSP response via a callback. A server can query the supplied parameters
9694 and set the encoded OCSP response in the callback. Add simplified examples
9695 to s_client and s_server.
9699 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9702 + Binary incompatibility of ssl_ctx_st structure
9703 + DTLS interoperation with non-compliant servers
9704 + Don't call get_session_cb() without proposed session
9705 + Fix ia64 assembler code
9707 *Andy Polyakov, Steve Henson*
9709 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9711 * DTLS Handshake overhaul. There were longstanding issues with
9712 OpenSSL DTLS implementation, which were making it impossible for
9713 RFC 4347 compliant client to communicate with OpenSSL server.
9714 Unfortunately just fixing these incompatibilities would "cut off"
9715 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9716 server keeps tolerating non RFC compliant syntax. The opposite is
9717 not true, 0.9.8f client can not communicate with earlier server.
9718 This update even addresses CVE-2007-4995.
9722 * Changes to avoid need for function casts in OpenSSL: some compilers
9723 (gcc 4.2 and later) reject their use.
9724 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9727 * Add RFC4507 support to OpenSSL. This includes the corrections in
9728 RFC4507bis. The encrypted ticket format is an encrypted encoded
9729 SSL_SESSION structure, that way new session features are automatically
9732 If a client application caches session in an SSL_SESSION structure
9733 support is transparent because tickets are now stored in the encoded
9736 The SSL_CTX structure automatically generates keys for ticket
9737 protection in servers so again support should be possible
9738 with no application modification.
9740 If a client or server wishes to disable RFC4507 support then the option
9741 SSL_OP_NO_TICKET can be set.
9743 Add a TLS extension debugging callback to allow the contents of any client
9744 or server extensions to be examined.
9746 This work was sponsored by Google.
9750 * Add initial support for TLS extensions, specifically for the server_name
9751 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9752 have new members for a host name. The SSL data structure has an
9753 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9754 stored in that context to allow for session resumption, even after the
9755 SSL has been switched to a new SSL_CTX in reaction to a client's
9756 server_name extension.
9758 New functions (subject to change):
9760 SSL_get_servername()
9761 SSL_get_servername_type()
9764 New CTRL codes and macros (subject to change):
9766 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9767 - SSL_CTX_set_tlsext_servername_callback()
9768 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9769 - SSL_CTX_set_tlsext_servername_arg()
9770 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9772 openssl s_client has a new '-servername ...' option.
9774 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9775 '-key2 ...', '-servername_fatal' (subject to change). This allows
9776 testing the HostName extension for a specific single host name ('-cert'
9777 and '-key' remain fallbacks for handshakes without HostName
9778 negotiation). If the unrecognized_name alert has to be sent, this by
9779 default is a warning; it becomes fatal with the '-servername_fatal'
9782 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9784 * Add AES and SSE2 assembly language support to VC++ build.
9788 * Mitigate attack on final subtraction in Montgomery reduction.
9792 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9793 (which previously caused an internal error).
9797 * Squeeze another 10% out of IGE mode when in != out.
9801 * AES IGE mode speedup.
9803 *Dean Gaudet (Google)*
9805 * Add the Korean symmetric 128-bit cipher SEED (see
9806 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9807 add SEED ciphersuites from RFC 4162:
9809 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9810 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9811 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9812 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9814 To minimize changes between patchlevels in the OpenSSL 0.9.8
9815 series, SEED remains excluded from compilation unless OpenSSL
9816 is configured with 'enable-seed'.
9818 *KISA, Bodo Moeller*
9820 * Mitigate branch prediction attacks, which can be practical if a
9821 single processor is shared, allowing a spy process to extract
9822 information. For detailed background information, see
9823 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9824 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9825 and Necessary Software Countermeasures"). The core of the change
9826 are new versions BN_div_no_branch() and
9827 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9828 respectively, which are slower, but avoid the security-relevant
9829 conditional branches. These are automatically called by BN_div()
9830 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9831 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9832 remove a conditional branch.
9834 BN_FLG_CONSTTIME is the new name for the previous
9835 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9836 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9837 in the exponent causes BN_mod_exp_mont() to use the alternative
9838 implementation in BN_mod_exp_mont_consttime().) The old name
9839 remains as a deprecated alias.
9841 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9842 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9843 constant-time implementations for more than just exponentiation.
9844 Here too the old name is kept as a deprecated alias.
9846 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9847 the BN_BLINDING structure gets an independent copy of the
9848 modulus. This means that the previous `BIGNUM *m` argument to
9849 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9850 essentially becomes `const BIGNUM *m`, although we can't actually
9851 change this in the header file before 0.9.9. It allows
9852 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9853 enable BN_FLG_CONSTTIME.
9855 *Matthew D Wood (Intel Corp)*
9857 * In the SSL/TLS server implementation, be strict about session ID
9858 context matching (which matters if an application uses a single
9859 external cache for different purposes). Previously,
9860 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9861 set. This did ensure strict client verification, but meant that,
9862 with applications using a single external cache for quite
9863 different requirements, clients could circumvent ciphersuite
9864 restrictions for a given session ID context by starting a session
9865 in a different context.
9869 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9870 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9871 authentication-only ciphersuites.
9875 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9876 not complete and could lead to a possible single byte overflow
9877 ([CVE-2007-5135]) [Ben Laurie]
9879 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9881 * Since AES128 and AES256 (and similarly Camellia128 and
9882 Camellia256) share a single mask bit in the logic of
9883 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9884 kludge to work properly if AES128 is available and AES256 isn't
9885 (or if Camellia128 is available and Camellia256 isn't).
9889 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9890 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9891 When a point or a seed is encoded in a BIT STRING, we need to
9892 prevent the removal of trailing zero bits to get the proper DER
9893 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9894 of a NamedBitList, for which trailing 0 bits need to be removed.)
9898 * Have SSL/TLS server implementation tolerate "mismatched" record
9899 protocol version while receiving ClientHello even if the
9900 ClientHello is fragmented. (The server can't insist on the
9901 particular protocol version it has chosen before the ServerHello
9902 message has informed the client about his choice.)
9906 * Add RFC 3779 support.
9908 *Rob Austein for ARIN, Ben Laurie*
9910 * Load error codes if they are not already present instead of using a
9911 static variable. This allows them to be cleanly unloaded and reloaded.
9912 Improve header file function name parsing.
9916 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9917 or CAPABILITY handshake as required by RFCs.
9921 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9923 * Introduce limits to prevent malicious keys being able to
9924 cause a denial of service. ([CVE-2006-2940])
9926 *Steve Henson, Bodo Moeller*
9928 * Fix ASN.1 parsing of certain invalid structures that can result
9929 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9931 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9932 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9934 * Fix SSL client code which could crash if connecting to a
9935 malicious SSLv2 server. ([CVE-2006-4343])
9937 *Tavis Ormandy and Will Drewry, Google Security Team*
9939 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9940 match only those. Before that, "AES256-SHA" would be interpreted
9941 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9942 the same strength classification in 0.9.7h) as we currently only
9943 have a single AES bit in the ciphersuite description bitmap.
9944 That change, however, also applied to ciphersuite strings such as
9945 "RC4-MD5" that intentionally matched multiple ciphersuites --
9946 namely, SSL 2.0 ciphersuites in addition to the more common ones
9947 from SSL 3.0/TLS 1.0.
9949 So we change the selection algorithm again: Naming an explicit
9950 ciphersuite selects this one ciphersuite, and any other similar
9951 ciphersuite (same bitmap) from *other* protocol versions.
9952 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9953 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9955 Since SSL 2.0 does not have any ciphersuites for which the
9956 128/256 bit distinction would be relevant, this works for now.
9957 The proper fix will be to use different bits for AES128 and
9958 AES256, which would have avoided the problems from the beginning;
9959 however, bits are scarce, so we can only do this in a new release
9960 (not just a patchlevel) when we can change the SSL_CIPHER
9961 definition to split the single 'unsigned long mask' bitmap into
9962 multiple values to extend the available space.
9966 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9968 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9969 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9971 * Add AES IGE and biIGE modes.
9975 * Change the Unix randomness entropy gathering to use poll() when
9976 possible instead of select(), since the latter has some
9977 undesirable limitations.
9979 *Darryl Miles via Richard Levitte and Bodo Moeller*
9981 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9982 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9983 cannot be implicitly activated as part of, e.g., the "AES" alias.
9984 However, please upgrade to OpenSSL 0.9.9[-dev] for
9985 non-experimental use of the ECC ciphersuites to get TLS extension
9986 support, which is required for curve and point format negotiation
9987 to avoid potential handshake problems.
9991 * Disable rogue ciphersuites:
9993 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9994 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9995 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9997 The latter two were purportedly from
9998 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10001 Also deactivate the remaining ciphersuites from
10002 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10003 unofficial, and the ID has long expired.
10007 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10008 dual-core machines) and other potential thread-safety issues.
10012 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10013 versions), which is now available for royalty-free use
10014 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10015 Also, add Camellia TLS ciphersuites from RFC 4132.
10017 To minimize changes between patchlevels in the OpenSSL 0.9.8
10018 series, Camellia remains excluded from compilation unless OpenSSL
10019 is configured with 'enable-camellia'.
10023 * Disable the padding bug check when compression is in use. The padding
10024 bug check assumes the first packet is of even length, this is not
10025 necessarily true if compression is enabled and can result in false
10026 positives causing handshake failure. The actual bug test is ancient
10027 code so it is hoped that implementations will either have fixed it by
10028 now or any which still have the bug do not support compression.
10032 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10034 * When applying a cipher rule check to see if string match is an explicit
10035 cipher suite and only match that one cipher suite if it is.
10039 * Link in manifests for VC++ if needed.
10041 *Austin Ziegler <halostatue@gmail.com>*
10043 * Update support for ECC-based TLS ciphersuites according to
10044 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10045 TLS extensions, which are supported starting with the 0.9.9
10046 branch, not in the OpenSSL 0.9.8 branch).
10050 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10051 opaque EVP_CIPHER_CTX handling.
10055 * Fixes and enhancements to zlib compression code. We now only use
10056 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10057 to conform with the standards mentioned here:
10058 <http://www.zlib.net/DLL_FAQ.txt>
10059 Static zlib linking now works on Windows and the new --with-zlib-include
10060 --with-zlib-lib options to Configure can be used to supply the location
10061 of the headers and library. Gracefully handle case where zlib library
10066 * Several fixes and enhancements to the OID generation code. The old code
10067 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10068 handle numbers larger than ULONG_MAX, truncated printing and had a
10069 non standard OBJ_obj2txt() behaviour.
10073 * Add support for building of engines under engine/ as shared libraries
10074 under VC++ build system.
10078 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10079 Hopefully, we will not see any false combination of paths any more.
10083 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10085 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10086 (part of SSL_OP_ALL). This option used to disable the
10087 countermeasure against man-in-the-middle protocol-version
10088 rollback in the SSL 2.0 server implementation, which is a bad
10089 idea. ([CVE-2005-2969])
10091 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10092 for Information Security, National Institute of Advanced Industrial
10093 Science and Technology [AIST], Japan)*
10095 * Add two function to clear and return the verify parameter flags.
10099 * Keep cipherlists sorted in the source instead of sorting them at
10100 runtime, thus removing the need for a lock.
10104 * Avoid some small subgroup attacks in Diffie-Hellman.
10106 *Nick Mathewson and Ben Laurie*
10108 * Add functions for well-known primes.
10112 * Extended Windows CE support.
10114 *Satoshi Nakamura and Andy Polyakov*
10116 * Initialize SSL_METHOD structures at compile time instead of during
10117 runtime, thus removing the need for a lock.
10121 * Make PKCS7_decrypt() work even if no certificate is supplied by
10122 attempting to decrypt each encrypted key in turn. Add support to
10127 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10129 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10132 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10136 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10137 key into the same file any more.
10141 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10145 * Add -utf8 command line and config file option to 'ca'.
10147 *Stefan <stf@udoma.org*
10149 * Removed the macro des_crypt(), as it seems to conflict with some
10150 libraries. Use DES_crypt().
10154 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10155 involves renaming the source and generated shared-libs for
10156 both. The engines will accept the corrected or legacy ids
10157 ('ncipher' and '4758_cca' respectively) when binding. NB,
10158 this only applies when building 'shared'.
10160 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10162 * Add attribute functions to EVP_PKEY structure. Modify
10163 PKCS12_create() to recognize a CSP name attribute and
10164 use it. Make -CSP option work again in pkcs12 utility.
10168 * Add new functionality to the bn blinding code:
10169 - automatic re-creation of the BN_BLINDING parameters after
10170 a fixed number of uses (currently 32)
10171 - add new function for parameter creation
10172 - introduce flags to control the update behaviour of the
10173 BN_BLINDING parameters
10174 - hide BN_BLINDING structure
10175 Add a second BN_BLINDING slot to the RSA structure to improve
10176 performance when a single RSA object is shared among several
10181 * Add support for DTLS.
10183 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10185 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10186 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10190 * Remove buggy and incomplete DH cert support from
10191 ssl/ssl_rsa.c and ssl/s3_both.c
10195 * Use SHA-1 instead of MD5 as the default digest algorithm for
10196 the `apps/openssl` commands.
10200 * Compile clean with "-Wall -Wmissing-prototypes
10201 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10202 DEBUG_SAFESTACK must also be set.
10206 * Change ./Configure so that certain algorithms can be disabled by default.
10207 The new counterpiece to "no-xxx" is "enable-xxx".
10209 The patented RC5 and MDC2 algorithms will now be disabled unless
10210 "enable-rc5" and "enable-mdc2", respectively, are specified.
10212 (IDEA remains enabled despite being patented. This is because IDEA
10213 is frequently required for interoperability, and there is no license
10214 fee for non-commercial use. As before, "no-idea" can be used to
10215 avoid this algorithm.)
10219 * Add processing of proxy certificates (see RFC 3820). This work was
10220 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10221 EGEE (Enabling Grids for E-science in Europe).
10225 * RC4 performance overhaul on modern architectures/implementations, such
10226 as Intel P4, IA-64 and AMD64.
10230 * New utility extract-section.pl. This can be used specify an alternative
10231 section number in a pod file instead of having to treat each file as
10232 a separate case in Makefile. This can be done by adding two lines to the
10235 =for comment openssl_section:XXX
10237 The blank line is mandatory.
10241 * New arguments -certform, -keyform and -pass for s_client and s_server
10242 to allow alternative format key and certificate files and passphrase
10247 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10248 update associated structures and add various utility functions.
10250 Add new policy related verify parameters, include policy checking in
10251 standard verify code. Enhance 'smime' application with extra parameters
10252 to support policy checking and print out.
10256 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10257 Nehemiah processors. These extensions support AES encryption in hardware
10258 as well as RNG (though RNG support is currently disabled).
10260 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10262 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10266 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10268 *Andy Polyakov and a number of other people*
10270 * Improved PowerPC platform support. Most notably BIGNUM assembler
10271 implementation contributed by IBM.
10273 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10275 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10276 exponent rather than 'unsigned long'. There is a corresponding change to
10277 the new 'rsa_keygen' element of the RSA_METHOD structure.
10279 *Jelte Jansen, Geoff Thorpe*
10281 * Functionality for creating the initial serial number file is now
10282 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10284 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10285 number file to 1, which is bound to cause problems. To avoid
10286 the problems while respecting compatibility between different 0.9.7
10287 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10288 CA.pl for serial number initialization. With the new release 0.9.8,
10289 we can fix the problem directly in the 'ca' utility.)
10293 * Reduced header interdependencies by declaring more opaque objects in
10294 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10295 give fewer recursive includes, which could break lazy source code - so
10296 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10297 developers should define this symbol when building and using openssl to
10298 ensure they track the recommended behaviour, interfaces, [etc], but
10299 backwards-compatible behaviour prevails when this isn't defined.
10303 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10307 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10308 This will generate a random key of the appropriate length based on the
10309 cipher context. The EVP_CIPHER can provide its own random key generation
10310 routine to support keys of a specific form. This is used in the des and
10311 3des routines to generate a key of the correct parity. Update S/MIME
10312 code to use new functions and hence generate correct parity DES keys.
10313 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10314 valid (weak or incorrect parity).
10318 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10319 as looking them up. This is useful when the verified structure may contain
10320 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10321 present unless the new PKCS7_NO_CRL flag is asserted.
10325 * Extend ASN1 oid configuration module. It now additionally accepts the
10328 shortName = some long name, 1.2.3.4
10332 * Reimplemented the BN_CTX implementation. There is now no more static
10333 limitation on the number of variables it can handle nor the depth of the
10334 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10335 information can now expand as required, and rather than having a single
10336 static array of bignums, BN_CTX now uses a linked-list of such arrays
10337 allowing it to expand on demand whilst maintaining the usefulness of
10338 BN_CTX's "bundling".
10342 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10343 to allow all RSA operations to function using a single BN_CTX.
10347 * Preliminary support for certificate policy evaluation and checking. This
10348 is initially intended to pass the tests outlined in "Conformance Testing
10349 of Relying Party Client Certificate Path Processing Logic" v1.07.
10353 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10354 remained unused and not that useful. A variety of other little bignum
10355 tweaks and fixes have also been made continuing on from the audit (see
10360 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10361 associated ASN1, EVP and SSL functions and old ASN1 macros.
10365 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10366 and this should never fail. So the return value from the use of
10367 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10368 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10372 * BN_CTX_get() should return zero-valued bignums, providing the same
10373 initialised value as BN_new().
10375 *Geoff Thorpe, suggested by Ulf Möller*
10377 * Support for inhibitAnyPolicy certificate extension.
10381 * An audit of the BIGNUM code is underway, for which debugging code is
10382 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10383 is considered valid when processing BIGNUMs, and causes execution to
10384 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10385 further steps are taken to deliberately pollute unused data in BIGNUM
10386 structures to try and expose faulty code further on. For now, openssl will
10387 (in its default mode of operation) continue to tolerate the inconsistent
10388 forms that it has tolerated in the past, but authors and packagers should
10389 consider trying openssl and their own applications when compiled with
10390 these debugging symbols defined. It will help highlight potential bugs in
10391 their own code, and will improve the test coverage for OpenSSL itself. At
10392 some point, these tighter rules will become openssl's default to improve
10393 maintainability, though the assert()s and other overheads will remain only
10394 in debugging configurations. See bn.h for more details.
10396 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10398 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10399 that can only be obtained through BN_CTX_new() (which implicitly
10400 initialises it). The presence of this function only made it possible
10401 to overwrite an existing structure (and cause memory leaks).
10405 * Because of the callback-based approach for implementing LHASH as a
10406 template type, lh_insert() adds opaque objects to hash-tables and
10407 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10408 to clean up those corresponding objects before destroying the hash table
10409 (and losing the object pointers). So some over-zealous constifications in
10410 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10411 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10412 prototyped to have "const" restrictions on the object pointers they are
10413 given (and so aren't required to cast them away any more).
10417 * The tmdiff.h API was so ugly and minimal that our own timing utility
10418 (speed) prefers to use its own implementation. The two implementations
10419 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10420 its object type properly exposed (MS_TM) instead of casting to/from
10421 `char *`. This may still change yet if someone realises MS_TM and
10423 aren't necessarily the greatest nomenclatures - but this is what was used
10424 internally to the implementation so I've used that for now.
10428 * Ensure that deprecated functions do not get compiled when
10429 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10430 the self-tests were still using deprecated key-generation functions so
10431 these have been updated also.
10435 * Reorganise PKCS#7 code to separate the digest location functionality
10436 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10437 New function PKCS7_set_digest() to set the digest type for PKCS#7
10438 digestedData type. Add additional code to correctly generate the
10439 digestedData type and add support for this type in PKCS7 initialization
10444 * New function PKCS7_set0_type_other() this initializes a PKCS7
10445 structure of type "other".
10449 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10450 sure the loop does correctly stop and breaking ("division by zero")
10451 modulus operations are not performed. The (pre-generated) prime
10452 table crypto/bn/bn_prime.h was already correct, but it could not be
10453 re-generated on some platforms because of the "division by zero"
10454 situation in the script.
10456 *Ralf S. Engelschall*
10458 * Update support for ECC-based TLS ciphersuites according to
10459 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10460 SHA-1 now is only used for "small" curves (where the
10461 representation of a field element takes up to 24 bytes); for
10462 larger curves, the field element resulting from ECDH is directly
10463 used as premaster secret.
10465 *Douglas Stebila (Sun Microsystems Laboratories)*
10467 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10468 curve secp160r1 to the tests.
10470 *Douglas Stebila (Sun Microsystems Laboratories)*
10472 * Add the possibility to load symbols globally with DSO.
10474 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10476 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10477 control of the error stack.
10481 * Add support for STORE in ENGINE.
10485 * Add the STORE type. The intention is to provide a common interface
10486 to certificate and key stores, be they simple file-based stores, or
10487 HSM-type store, or LDAP stores, or...
10488 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10492 * Add a generic structure called OPENSSL_ITEM. This can be used to
10493 pass a list of arguments to any function as well as provide a way
10494 for a function to pass data back to the caller.
10498 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10499 works like BUF_strdup() but can be used to duplicate a portion of
10500 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10505 * Add the function sk_find_ex() which works like sk_find(), but will
10506 return an index to an element even if an exact match couldn't be
10507 found. The index is guaranteed to point at the element where the
10508 searched-for key would be inserted to preserve sorting order.
10512 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10513 takes an extra flags argument for optional functionality. Currently,
10514 the following flags are defined:
10516 OBJ_BSEARCH_VALUE_ON_NOMATCH
10517 This one gets OBJ_bsearch_ex() to return a pointer to the first
10518 element where the comparing function returns a negative or zero
10521 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10522 This one gets OBJ_bsearch_ex() to return a pointer to the first
10523 element where the comparing function returns zero. This is useful
10524 if there are more than one element where the comparing function
10529 * Make it possible to create self-signed certificates with 'openssl ca'
10530 in such a way that the self-signed certificate becomes part of the
10531 CA database and uses the same mechanisms for serial number generation
10532 as all other certificate signing. The new flag '-selfsign' enables
10533 this functionality. Adapt CA.sh and CA.pl.in.
10537 * Add functionality to check the public key of a certificate request
10538 against a given private. This is useful to check that a certificate
10539 request can be signed by that key (self-signing).
10543 * Make it possible to have multiple active certificates with the same
10544 subject in the CA index file. This is done only if the keyword
10545 'unique_subject' is set to 'no' in the main CA section (default
10546 if 'CA_default') of the configuration file. The value is saved
10547 with the database itself in a separate index attribute file,
10548 named like the index file with '.attr' appended to the name.
10552 * Generate multi-valued AVAs using '+' notation in config files for
10557 * Support for nameConstraints certificate extension.
10561 * Support for policyConstraints certificate extension.
10565 * Support for policyMappings certificate extension.
10569 * Make sure the default DSA_METHOD implementation only uses its
10570 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10571 and change its own handlers to be NULL so as to remove unnecessary
10572 indirection. This lets alternative implementations fallback to the
10573 default implementation more easily.
10577 * Support for directoryName in GeneralName related extensions
10582 * Make it possible to link applications using Makefile.shared.
10583 Make that possible even when linking against static libraries!
10587 * Support for single pass processing for S/MIME signing. This now
10588 means that S/MIME signing can be done from a pipe, in addition
10589 cleartext signing (multipart/signed type) is effectively streaming
10590 and the signed data does not need to be all held in memory.
10592 This is done with a new flag PKCS7_STREAM. When this flag is set
10593 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10594 is done after the data is output (and digests calculated) in
10595 SMIME_write_PKCS7().
10599 * Add full support for -rpath/-R, both in shared libraries and
10600 applications, at least on the platforms where it's known how
10605 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10606 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10607 will now compute a table of multiples of the generator that
10608 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10609 faster (notably in the case of a single point multiplication,
10610 scalar * generator).
10612 *Nils Larsch, Bodo Moeller*
10614 * IPv6 support for certificate extensions. The various extensions
10615 which use the IP:a.b.c.d can now take IPv6 addresses using the
10616 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10621 * Added an ENGINE that implements RSA by performing private key
10622 exponentiations with the GMP library. The conversions to and from
10623 GMP's mpz_t format aren't optimised nor are any montgomery forms
10624 cached, and on x86 it appears OpenSSL's own performance has caught up.
10625 However there are likely to be other architectures where GMP could
10626 provide a boost. This ENGINE is not built in by default, but it can be
10627 specified at Configure time and should be accompanied by the necessary
10628 linker additions, eg;
10629 ./config -DOPENSSL_USE_GMP -lgmp
10633 * "openssl engine" will not display ENGINE/DSO load failure errors when
10634 testing availability of engines with "-t" - the old behaviour is
10635 produced by increasing the feature's verbosity with "-tt".
10639 * ECDSA routines: under certain error conditions uninitialized BN objects
10640 could be freed. Solution: make sure initialization is performed early
10641 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10646 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10647 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10648 software implementations. For DSA and DH, parameter generation can
10649 also be overridden by providing the appropriate method callbacks.
10653 * Change the "progress" mechanism used in key-generation and
10654 primality testing to functions that take a new BN_GENCB pointer in
10655 place of callback/argument pairs. The new API functions have `_ex`
10656 postfixes and the older functions are reimplemented as wrappers for
10657 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10658 declarations of the old functions to help (graceful) attempts to
10659 migrate to the new functions. Also, the new key-generation API
10660 functions operate on a caller-supplied key-structure and return
10661 success/failure rather than returning a key or NULL - this is to
10662 help make "keygen" another member function of RSA_METHOD etc.
10664 Example for using the new callback interface:
10666 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10667 void *my_arg = ...;
10670 BN_GENCB_set(&my_cb, my_callback, my_arg);
10672 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10673 /* For the meaning of a, b in calls to my_callback(), see the
10674 * documentation of the function that calls the callback.
10675 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10676 * my_callback should return 1 if it wants BN_is_prime_ex()
10677 * to continue, or 0 to stop.
10682 * Change the ZLIB compression method to be stateful, and make it
10683 available to TLS with the number defined in
10684 draft-ietf-tls-compression-04.txt.
10688 * Add the ASN.1 structures and functions for CertificatePair, which
10689 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10691 CertificatePair ::= SEQUENCE {
10692 forward [0] Certificate OPTIONAL,
10693 reverse [1] Certificate OPTIONAL,
10694 -- at least one of the pair shall be present -- }
10696 Also implement the PEM functions to read and write certificate
10697 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10699 This needed to be defined, mostly for the sake of the LDAP
10700 attribute crossCertificatePair, but may prove useful elsewhere as
10705 * Make it possible to inhibit symlinking of shared libraries in
10706 Makefile.shared, for Cygwin's sake.
10710 * Extend the BIGNUM API by creating a function
10711 void BN_set_negative(BIGNUM *a, int neg);
10712 and a macro that behave like
10713 int BN_is_negative(const BIGNUM *a);
10715 to avoid the need to access 'a->neg' directly in applications.
10719 * Implement fast modular reduction for pseudo-Mersenne primes
10720 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10721 EC_GROUP_new_curve_GFp() will now automatically use this
10724 *Nils Larsch <nla@trustcenter.de>*
10726 * Add new lock type (CRYPTO_LOCK_BN).
10730 * Change the ENGINE framework to automatically load engines
10731 dynamically from specific directories unless they could be
10732 found to already be built in or loaded. Move all the
10733 current engines except for the cryptodev one to a new
10734 directory engines/.
10735 The engines in engines/ are built as shared libraries if
10736 the "shared" options was given to ./Configure or ./config.
10737 Otherwise, they are inserted in libcrypto.a.
10738 /usr/local/ssl/engines is the default directory for dynamic
10739 engines, but that can be overridden at configure time through
10740 the usual use of --prefix and/or --openssldir, and at run
10741 time with the environment variable OPENSSL_ENGINES.
10743 *Geoff Thorpe and Richard Levitte*
10745 * Add Makefile.shared, a helper makefile to build shared
10746 libraries. Adapt Makefile.org.
10750 * Add version info to Win32 DLLs.
10752 *Peter 'Luna' Runestig" <peter@runestig.com>*
10754 * Add new 'medium level' PKCS#12 API. Certificates and keys
10755 can be added using this API to created arbitrary PKCS#12
10756 files while avoiding the low level API.
10758 New options to PKCS12_create(), key or cert can be NULL and
10759 will then be omitted from the output file. The encryption
10760 algorithm NIDs can be set to -1 for no encryption, the mac
10761 iteration count can be set to 0 to omit the mac.
10763 Enhance pkcs12 utility by making the -nokeys and -nocerts
10764 options work when creating a PKCS#12 file. New option -nomac
10765 to omit the mac, NONE can be set for an encryption algorithm.
10766 New code is modified to use the enhanced PKCS12_create()
10767 instead of the low level API.
10771 * Extend ASN1 encoder to support indefinite length constructed
10772 encoding. This can output sequences tags and octet strings in
10773 this form. Modify pk7_asn1.c to support indefinite length
10774 encoding. This is experimental and needs additional code to
10775 be useful, such as an ASN1 bio and some enhanced streaming
10778 Extend template encode functionality so that tagging is passed
10779 down to the template encoder.
10783 * Let 'openssl req' fail if an argument to '-newkey' is not
10784 recognized instead of using RSA as a default.
10788 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10789 As these are not official, they are not included in "ALL";
10790 the "ECCdraft" ciphersuite group alias can be used to select them.
10792 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10794 * Add ECDH engine support.
10796 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10798 * Add ECDH in new directory crypto/ecdh/.
10800 *Douglas Stebila (Sun Microsystems Laboratories)*
10802 * Let BN_rand_range() abort with an error after 100 iterations
10803 without success (which indicates a broken PRNG).
10807 * Change BN_mod_sqrt() so that it verifies that the input value
10808 is really the square of the return value. (Previously,
10809 BN_mod_sqrt would show GIGO behaviour.)
10813 * Add named elliptic curves over binary fields from X9.62, SECG,
10814 and WAP/WTLS; add OIDs that were still missing.
10816 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10818 * Extend the EC library for elliptic curves over binary fields
10819 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10822 EC_GF2m_simple_method
10826 EC_GROUP_new_curve_GF2m
10827 EC_GROUP_set_curve_GF2m
10828 EC_GROUP_get_curve_GF2m
10829 EC_POINT_set_affine_coordinates_GF2m
10830 EC_POINT_get_affine_coordinates_GF2m
10831 EC_POINT_set_compressed_coordinates_GF2m
10833 Point compression for binary fields is disabled by default for
10834 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10837 As binary polynomials are represented as BIGNUMs, various members
10838 of the EC_GROUP and EC_POINT data structures can be shared
10839 between the implementations for prime fields and binary fields;
10840 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10841 are essentially identical to their `..._GFp` counterparts.
10842 (For simplicity, the `..._GFp` prefix has been dropped from
10843 various internal method names.)
10845 An internal 'field_div' method (similar to 'field_mul' and
10846 'field_sqr') has been added; this is used only for binary fields.
10848 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10850 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10851 through methods ('mul', 'precompute_mult').
10853 The generic implementations (now internally called 'ec_wNAF_mul'
10854 and 'ec_wNAF_precomputed_mult') remain the default if these
10855 methods are undefined.
10857 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10859 * New function EC_GROUP_get_degree, which is defined through
10860 EC_METHOD. For curves over prime fields, this returns the bit
10861 length of the modulus.
10863 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10865 * New functions EC_GROUP_dup, EC_POINT_dup.
10866 (These simply call ..._new and ..._copy).
10868 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10870 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10871 Polynomials are represented as BIGNUMs (where the sign bit is not
10872 used) in the following functions [macros]:
10875 BN_GF2m_sub [= BN_GF2m_add]
10876 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10877 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10878 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10880 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10881 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10882 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10883 BN_GF2m_cmp [= BN_ucmp]
10885 (Note that only the 'mod' functions are actually for fields GF(2^m).
10886 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10888 For some functions, an the irreducible polynomial defining a
10889 field can be given as an 'unsigned int[]' with strictly
10890 decreasing elements giving the indices of those bits that are set;
10891 i.e., p[] represents the polynomial
10892 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10894 p[0] > p[1] > ... > p[k] = 0.
10895 This applies to the following functions:
10898 BN_GF2m_mod_mul_arr
10899 BN_GF2m_mod_sqr_arr
10900 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10901 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10902 BN_GF2m_mod_exp_arr
10903 BN_GF2m_mod_sqrt_arr
10904 BN_GF2m_mod_solve_quad_arr
10908 Conversion can be performed by the following functions:
10913 bntest.c has additional tests for binary polynomial arithmetic.
10915 Two implementations for BN_GF2m_mod_div() are available.
10916 The default algorithm simply uses BN_GF2m_mod_inv() and
10917 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10918 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10919 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10921 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10923 * Add new error code 'ERR_R_DISABLED' that can be used when some
10924 functionality is disabled at compile-time.
10926 *Douglas Stebila <douglas.stebila@sun.com>*
10928 * Change default behaviour of 'openssl asn1parse' so that more
10929 information is visible when viewing, e.g., a certificate:
10931 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10932 mode the content of non-printable OCTET STRINGs is output in a
10933 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10934 avoid the appearance of a printable string.
10936 *Nils Larsch <nla@trustcenter.de>*
10938 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10940 EC_GROUP_set_asn1_flag()
10941 EC_GROUP_get_asn1_flag()
10942 EC_GROUP_set_point_conversion_form()
10943 EC_GROUP_get_point_conversion_form()
10944 These control ASN1 encoding details:
10945 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10946 has been set to OPENSSL_EC_NAMED_CURVE.
10947 - Points are encoded in uncompressed form by default; options for
10948 asn1_for are as for point2oct, namely
10949 POINT_CONVERSION_COMPRESSED
10950 POINT_CONVERSION_UNCOMPRESSED
10951 POINT_CONVERSION_HYBRID
10953 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10955 EC_GROUP_set_seed()
10956 EC_GROUP_get0_seed()
10957 EC_GROUP_get_seed_len()
10958 This is used only for ASN1 purposes (so far).
10960 *Nils Larsch <nla@trustcenter.de>*
10962 * Add 'field_type' member to EC_METHOD, which holds the NID
10963 of the appropriate field type OID. The new function
10964 EC_METHOD_get_field_type() returns this value.
10966 *Nils Larsch <nla@trustcenter.de>*
10969 EC_POINT_point2bn()
10970 EC_POINT_bn2point()
10971 EC_POINT_point2hex()
10972 EC_POINT_hex2point()
10973 providing useful interfaces to EC_POINT_point2oct() and
10974 EC_POINT_oct2point().
10976 *Nils Larsch <nla@trustcenter.de>*
10978 * Change internals of the EC library so that the functions
10979 EC_GROUP_set_generator()
10980 EC_GROUP_get_generator()
10981 EC_GROUP_get_order()
10982 EC_GROUP_get_cofactor()
10983 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10984 to methods, which would lead to unnecessary code duplication when
10985 adding different types of curves.
10987 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10989 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10990 arithmetic, and such that modified wNAFs are generated
10991 (which avoid length expansion in many cases).
10995 * Add a function EC_GROUP_check_discriminant() (defined via
10996 EC_METHOD) that verifies that the curve discriminant is non-zero.
10998 Add a function EC_GROUP_check() that makes some sanity tests
10999 on a EC_GROUP, its generator and order. This includes
11000 EC_GROUP_check_discriminant().
11002 *Nils Larsch <nla@trustcenter.de>*
11004 * Add ECDSA in new directory crypto/ecdsa/.
11006 Add applications 'openssl ecparam' and 'openssl ecdsa'
11007 (these are based on 'openssl dsaparam' and 'openssl dsa').
11009 ECDSA support is also included in various other files across the
11010 library. Most notably,
11011 - 'openssl req' now has a '-newkey ecdsa:file' option;
11012 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11013 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11014 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11015 them suitable for ECDSA where domain parameters must be
11016 extracted before the specific public key;
11017 - ECDSA engine support has been added.
11019 *Nils Larsch <nla@trustcenter.de>*
11021 * Include some named elliptic curves, and add OIDs from X9.62,
11022 SECG, and WAP/WTLS. Each curve can be obtained from the new
11024 EC_GROUP_new_by_curve_name(),
11025 and the list of available named curves can be obtained with
11026 EC_get_builtin_curves().
11027 Also add a 'curve_name' member to EC_GROUP objects, which can be
11029 EC_GROUP_set_curve_name()
11030 EC_GROUP_get_curve_name()
11032 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11034 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11035 was actually never needed) and in BN_mul(). The removal in BN_mul()
11036 required a small change in bn_mul_part_recursive() and the addition
11037 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11038 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11039 bn_sub_words() and bn_add_words() except they take arrays with
11044 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11046 * Cleanse PEM buffers before freeing them since they may contain
11049 *Benjamin Bennett <ben@psc.edu>*
11051 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11052 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11053 authentication-only ciphersuites.
11057 * Since AES128 and AES256 share a single mask bit in the logic of
11058 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11059 kludge to work properly if AES128 is available and AES256 isn't.
11063 * Expand security boundary to match 1.1.1 module.
11067 * Remove redundant features: hash file source, editing of test vectors
11068 modify fipsld to use external fips_premain.c signature.
11072 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11073 run algorithm test programs.
11077 * Make algorithm test programs more tolerant of whitespace.
11081 * Have SSL/TLS server implementation tolerate "mismatched" record
11082 protocol version while receiving ClientHello even if the
11083 ClientHello is fragmented. (The server can't insist on the
11084 particular protocol version it has chosen before the ServerHello
11085 message has informed the client about his choice.)
11089 * Load error codes if they are not already present instead of using a
11090 static variable. This allows them to be cleanly unloaded and reloaded.
11094 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11096 * Introduce limits to prevent malicious keys being able to
11097 cause a denial of service. ([CVE-2006-2940])
11099 *Steve Henson, Bodo Moeller*
11101 * Fix ASN.1 parsing of certain invalid structures that can result
11102 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11104 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11105 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11107 * Fix SSL client code which could crash if connecting to a
11108 malicious SSLv2 server. ([CVE-2006-4343])
11110 *Tavis Ormandy and Will Drewry, Google Security Team*
11112 * Change ciphersuite string processing so that an explicit
11113 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11114 will no longer include "AES128-SHA"), and any other similar
11115 ciphersuite (same bitmap) from *other* protocol versions (so that
11116 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11117 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11118 changes from 0.9.8b and 0.9.8d.
11122 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11124 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11125 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11127 * Change the Unix randomness entropy gathering to use poll() when
11128 possible instead of select(), since the latter has some
11129 undesirable limitations.
11131 *Darryl Miles via Richard Levitte and Bodo Moeller*
11133 * Disable rogue ciphersuites:
11135 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11136 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11137 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11139 The latter two were purportedly from
11140 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11143 Also deactivate the remaining ciphersuites from
11144 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11145 unofficial, and the ID has long expired.
11149 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11150 dual-core machines) and other potential thread-safety issues.
11154 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11156 * Adapt fipsld and the build system to link against the validated FIPS
11157 module in FIPS mode.
11161 * Fixes for VC++ 2005 build under Windows.
11165 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11166 from a Windows bash shell such as MSYS. It is autodetected from the
11167 "config" script when run from a VC++ environment. Modify standard VC++
11168 build to use fipscanister.o from the GNU make build.
11172 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11174 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11175 The value now differs depending on if you build for FIPS or not.
11176 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11177 safely run with a non-FIPSed libcrypto, as it may crash because of
11178 the difference induced by this change.
11182 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11184 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11185 (part of SSL_OP_ALL). This option used to disable the
11186 countermeasure against man-in-the-middle protocol-version
11187 rollback in the SSL 2.0 server implementation, which is a bad
11188 idea. ([CVE-2005-2969])
11190 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11191 for Information Security, National Institute of Advanced Industrial
11192 Science and Technology [AIST, Japan)]*
11194 * Minimal support for X9.31 signatures and PSS padding modes. This is
11195 mainly for FIPS compliance and not fully integrated at this stage.
11199 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11200 the exponentiation using a fixed-length exponent. (Otherwise,
11201 the information leaked through timing could expose the secret key
11202 after many signatures; cf. Bleichenbacher's attack on DSA with
11207 * Make a new fixed-window mod_exp implementation the default for
11208 RSA, DSA, and DH private-key operations so that the sequence of
11209 squares and multiplies and the memory access pattern are
11210 independent of the particular secret key. This will mitigate
11211 cache-timing and potential related attacks.
11213 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11214 and this is automatically used by BN_mod_exp_mont() if the new flag
11215 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11216 will use this BN flag for private exponents unless the flag
11217 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11218 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11220 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11222 * Change the client implementation for SSLv23_method() and
11223 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11224 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11225 (Previously, the SSL 2.0 backwards compatible Client Hello
11226 message format would be used even with SSL_OP_NO_SSLv2.)
11230 * Add support for smime-type MIME parameter in S/MIME messages which some
11235 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11236 a threadsafe manner. Modify rsa code to use new function and add calls
11237 to dsa and dh code (which had race conditions before).
11241 * Include the fixed error library code in the C error file definitions
11242 instead of fixing them up at runtime. This keeps the error code
11243 structures constant.
11247 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11249 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11252 * Fixes for newer kerberos headers. NB: the casts are needed because
11253 the 'length' field is signed on one version and unsigned on another
11254 with no (?) obvious way to tell the difference, without these VC++
11255 complains. Also the "definition" of FAR (blank) is no longer included
11256 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11257 some needed definitions.
11261 * Undo Cygwin change.
11265 * Added support for proxy certificates according to RFC 3820.
11266 Because they may be a security thread to unaware applications,
11267 they must be explicitly allowed in run-time. See
11268 docs/HOWTO/proxy_certificates.txt for further information.
11272 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11274 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11275 server and client random values. Previously
11276 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11277 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11279 This change has negligible security impact because:
11281 1. Server and client random values still have 24 bytes of pseudo random
11284 2. Server and client random values are sent in the clear in the initial
11287 3. The master secret is derived using the premaster secret (48 bytes in
11288 size for static RSA ciphersuites) as well as client server and random
11291 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11294 *Stephen Henson, reported by UK NISCC*
11296 * Use Windows randomness collection on Cygwin.
11300 * Fix hang in EGD/PRNGD query when communication socket is closed
11301 prematurely by EGD/PRNGD.
11303 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11305 * Prompt for pass phrases when appropriate for PKCS12 input format.
11309 * Back-port of selected performance improvements from development
11310 branch, as well as improved support for PowerPC platforms.
11314 * Add lots of checks for memory allocation failure, error codes to indicate
11315 failure and freeing up memory if a failure occurs.
11317 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11319 * Add new -passin argument to dgst.
11323 * Perform some character comparisons of different types in X509_NAME_cmp:
11324 this is needed for some certificates that re-encode DNs into UTF8Strings
11325 (in violation of RFC3280) and can't or won't issue name rollover
11330 * Make an explicit check during certificate validation to see that
11331 the CA setting in each certificate on the chain is correct. As a
11332 side effect always do the following basic checks on extensions,
11333 not just when there's an associated purpose to the check:
11335 - if there is an unhandled critical extension (unless the user
11336 has chosen to ignore this fault)
11337 - if the path length has been exceeded (if one is set at all)
11338 - that certain extensions fit the associated purpose (if one has
11343 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11345 * Avoid a race condition when CRLs are checked in a multi threaded
11346 environment. This would happen due to the reordering of the revoked
11347 entries during signature checking and serial number lookup. Now the
11348 encoding is cached and the serial number sort performed under a lock.
11349 Add new STACK function sk_is_sorted().
11353 * Add Delta CRL to the extension code.
11357 * Various fixes to s3_pkt.c so alerts are sent properly.
11359 *David Holmes <d.holmes@f5.com>*
11361 * Reduce the chances of duplicate issuer name and serial numbers (in
11362 violation of RFC3280) using the OpenSSL certificate creation utilities.
11363 This is done by creating a random 64 bit value for the initial serial
11364 number when a serial number file is created or when a self signed
11365 certificate is created using 'openssl req -x509'. The initial serial
11366 number file is created using 'openssl x509 -next_serial' in CA.pl
11367 rather than being initialized to 1.
11371 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11373 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11374 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11376 *Joe Orton, Steve Henson*
11378 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11381 *Joe Orton, Steve Henson*
11383 * Make it possible to have multiple active certificates with the same
11384 subject in the CA index file. This is done only if the keyword
11385 'unique_subject' is set to 'no' in the main CA section (default
11386 if 'CA_default') of the configuration file. The value is saved
11387 with the database itself in a separate index attribute file,
11388 named like the index file with '.attr' appended to the name.
11392 * X509 verify fixes. Disable broken certificate workarounds when
11393 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11394 keyUsage extension present. Don't accept CRLs with unhandled critical
11395 extensions: since verify currently doesn't process CRL extensions this
11396 rejects a CRL with *any* critical extensions. Add new verify error codes
11401 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11402 A clarification of RFC2560 will require the use of OCTET STRINGs and
11403 some implementations cannot handle the current raw format. Since OpenSSL
11404 copies and compares OCSP nonces as opaque blobs without any attempt at
11405 parsing them this should not create any compatibility issues.
11409 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11410 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11411 this HMAC (and other) operations are several times slower than OpenSSL
11416 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11418 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11420 * Use the correct content when signing type "other".
11424 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11426 * Fix various bugs revealed by running the NISCC test suite:
11428 Stop out of bounds reads in the ASN1 code when presented with
11429 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11431 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11433 If verify callback ignores invalid public key errors don't try to check
11434 certificate signature with the NULL public key.
11438 * New -ignore_err option in ocsp application to stop the server
11439 exiting on the first error in a request.
11443 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11444 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11449 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11450 extra data after the compression methods not only for TLS 1.0
11451 but also for SSL 3.0 (as required by the specification).
11453 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11455 * Change X509_certificate_type() to mark the key as exported/exportable
11456 when it's 512 *bits* long, not 512 bytes.
11460 * Change AES_cbc_encrypt() so it outputs exact multiple of
11461 blocks during encryption.
11465 * Various fixes to base64 BIO and non blocking I/O. On write
11466 flushes were not handled properly if the BIO retried. On read
11467 data was not being buffered properly and had various logic bugs.
11468 This also affects blocking I/O when the data being decoded is a
11473 * Various S/MIME bugfixes and compatibility changes:
11474 output correct application/pkcs7 MIME type if
11475 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11476 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11477 of files as .eml work). Correctly handle very long lines in MIME
11482 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11484 * Countermeasure against the Klima-Pokorny-Rosa extension of
11485 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11486 a protocol version number mismatch like a decryption error
11487 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11491 * Turn on RSA blinding by default in the default implementation
11492 to avoid a timing attack. Applications that don't want it can call
11493 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11494 They would be ill-advised to do so in most cases.
11496 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11498 * Change RSA blinding code so that it works when the PRNG is not
11499 seeded (in this case, the secret RSA exponent is abused as
11500 an unpredictable seed -- if it is not unpredictable, there
11501 is no point in blinding anyway). Make RSA blinding thread-safe
11502 by remembering the creator's thread ID in rsa->blinding and
11503 having all other threads use local one-time blinding factors
11504 (this requires more computation than sharing rsa->blinding, but
11505 avoids excessive locking; and if an RSA object is not shared
11506 between threads, blinding will still be very fast).
11510 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11511 ENGINE as defaults for all supported algorithms irrespective of
11512 the 'flags' parameter. 'flags' is now honoured, so applications
11513 should make sure they are passing it correctly.
11517 * Target "mingw" now allows native Windows code to be generated in
11518 the Cygwin environment as well as with the MinGW compiler.
11522 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11524 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11525 via timing by performing a MAC computation even if incorrect
11526 block cipher padding has been found. This is a countermeasure
11527 against active attacks where the attacker has to distinguish
11528 between bad padding and a MAC verification error. ([CVE-2003-0078])
11530 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11531 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11532 Martin Vuagnoux (EPFL, Ilion)*
11534 * Make the no-err option work as intended. The intention with no-err
11535 is not to have the whole error stack handling routines removed from
11536 libcrypto, it's only intended to remove all the function name and
11537 reason texts, thereby removing some of the footprint that may not
11538 be interesting if those errors aren't displayed anyway.
11540 NOTE: it's still possible for any application or module to have its
11541 own set of error texts inserted. The routines are there, just not
11542 used by default when no-err is given.
11546 * Add support for FreeBSD on IA64.
11548 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11550 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11551 Kerberos function mit_des_cbc_cksum(). Before this change,
11552 the value returned by DES_cbc_cksum() was like the one from
11553 mit_des_cbc_cksum(), except the bytes were swapped.
11555 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11557 * Allow an application to disable the automatic SSL chain building.
11558 Before this a rather primitive chain build was always performed in
11559 ssl3_output_cert_chain(): an application had no way to send the
11560 correct chain if the automatic operation produced an incorrect result.
11562 Now the chain builder is disabled if either:
11564 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11566 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11568 The reasoning behind this is that an application would not want the
11569 auto chain building to take place if extra chain certificates are
11570 present and it might also want a means of sending no additional
11571 certificates (for example the chain has two certificates and the
11576 * Add the possibility to build without the ENGINE framework.
11578 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11580 * Under Win32 gmtime() can return NULL: check return value in
11581 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11585 * DSA routines: under certain error conditions uninitialized BN objects
11586 could be freed. Solution: make sure initialization is performed early
11587 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11588 Nils Larsch <nla@trustcenter.de> via PR#459)
11592 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11593 checked on reconnect on the client side, therefore session resumption
11594 could still fail with a "ssl session id is different" error. This
11595 behaviour is masked when SSL_OP_ALL is used due to
11596 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11597 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11598 followup to PR #377.
11602 * IA-32 assembler support enhancements: unified ELF targets, support
11603 for SCO/Caldera platforms, fix for Cygwin shared build.
11607 * Add support for FreeBSD on sparc64. As a consequence, support for
11608 FreeBSD on non-x86 processors is separate from x86 processors on
11609 the config script, much like the NetBSD support.
11611 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11613 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11615 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11618 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11619 code (06) was taken as the first octet of the session ID and the last
11620 octet was ignored consequently. As a result SSLv2 client side session
11621 caching could not have worked due to the session ID mismatch between
11623 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11628 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11629 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11634 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11635 seems that in spite of existing for more than a year, many application
11636 author have done nothing to provide the necessary callbacks, which
11637 means that this particular engine will not work properly anywhere.
11638 This is a very unfortunate situation which forces us, in the name
11639 of usability, to give the hw_ncipher.c a static lock, which is part
11641 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11642 appear in 0.9.8 or later. We EXPECT application authors to have
11643 dealt properly with this when 0.9.8 is released (unless we actually
11644 make such changes in the libcrypto locking code that changes will
11645 have to be made anyway).
11649 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11650 octets have been read, EOF or an error occurs. Without this change
11651 some truncated ASN1 structures will not produce an error.
11655 * Disable Heimdal support, since it hasn't been fully implemented.
11656 Still give the possibility to force the use of Heimdal, but with
11657 warnings and a request that patches get sent to openssl-dev.
11661 * Add the VC-CE target, introduce the WINCE sysname, and add
11662 INSTALL.WCE and appropriate conditionals to make it build.
11664 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11666 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11667 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11668 edit numbers of the version.
11670 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11672 * Introduce safe string copy and catenation functions
11673 (BUF_strlcpy() and BUF_strlcat()).
11675 *Ben Laurie (CHATS) and Richard Levitte*
11677 * Avoid using fixed-size buffers for one-line DNs.
11679 *Ben Laurie (CHATS)*
11681 * Add BUF_MEM_grow_clean() to avoid information leakage when
11682 resizing buffers containing secrets, and use where appropriate.
11684 *Ben Laurie (CHATS)*
11686 * Avoid using fixed size buffers for configuration file location.
11688 *Ben Laurie (CHATS)*
11690 * Avoid filename truncation for various CA files.
11692 *Ben Laurie (CHATS)*
11694 * Use sizeof in preference to magic numbers.
11696 *Ben Laurie (CHATS)*
11698 * Avoid filename truncation in cert requests.
11700 *Ben Laurie (CHATS)*
11702 * Add assertions to check for (supposedly impossible) buffer
11705 *Ben Laurie (CHATS)*
11707 * Don't cache truncated DNS entries in the local cache (this could
11708 potentially lead to a spoofing attack).
11710 *Ben Laurie (CHATS)*
11712 * Fix various buffers to be large enough for hex/decimal
11713 representations in a platform independent manner.
11715 *Ben Laurie (CHATS)*
11717 * Add CRYPTO_realloc_clean() to avoid information leakage when
11718 resizing buffers containing secrets, and use where appropriate.
11720 *Ben Laurie (CHATS)*
11722 * Add BIO_indent() to avoid much slightly worrying code to do
11725 *Ben Laurie (CHATS)*
11727 * Convert sprintf()/BIO_puts() to BIO_printf().
11729 *Ben Laurie (CHATS)*
11731 * buffer_gets() could terminate with the buffer only half
11734 *Ben Laurie (CHATS)*
11736 * Add assertions to prevent user-supplied crypto functions from
11737 overflowing internal buffers by having large block sizes, etc.
11739 *Ben Laurie (CHATS)*
11741 * New OPENSSL_assert() macro (similar to assert(), but enabled
11744 *Ben Laurie (CHATS)*
11746 * Eliminate unused copy of key in RC4.
11748 *Ben Laurie (CHATS)*
11750 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11752 *Ben Laurie (CHATS)*
11754 * Fix off-by-one error in EGD path.
11756 *Ben Laurie (CHATS)*
11758 * If RANDFILE path is too long, ignore instead of truncating.
11760 *Ben Laurie (CHATS)*
11762 * Eliminate unused and incorrectly sized X.509 structure
11765 *Ben Laurie (CHATS)*
11767 * Eliminate unused and dangerous function knumber().
11769 *Ben Laurie (CHATS)*
11771 * Eliminate unused and dangerous structure, KSSL_ERR.
11773 *Ben Laurie (CHATS)*
11775 * Protect against overlong session ID context length in an encoded
11776 session object. Since these are local, this does not appear to be
11779 *Ben Laurie (CHATS)*
11781 * Change from security patch (see 0.9.6e below) that did not affect
11782 the 0.9.6 release series:
11784 Remote buffer overflow in SSL3 protocol - an attacker could
11785 supply an oversized master key in Kerberos-enabled versions.
11788 *Ben Laurie (CHATS)*
11790 * Change the SSL kerb5 codes to match RFC 2712.
11794 * Make -nameopt work fully for req and add -reqopt switch.
11796 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11798 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11800 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11802 * Make sure tests can be performed even if the corresponding algorithms
11803 have been removed entirely. This was also the last step to make
11804 OpenSSL compilable with DJGPP under all reasonable conditions.
11806 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11808 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11809 to allow version independent disabling of normally unselected ciphers,
11810 which may be activated as a side-effect of selecting a single cipher.
11812 (E.g., cipher list string "RSA" enables ciphersuites that are left
11813 out of "ALL" because they do not provide symmetric encryption.
11814 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11816 *Lutz Jaenicke, Bodo Moeller*
11818 * Add appropriate support for separate platform-dependent build
11819 directories. The recommended way to make a platform-dependent
11820 build directory is the following (tested on Linux), maybe with
11823 # Place yourself outside of the OpenSSL source tree. In
11824 # this example, the environment variable OPENSSL_SOURCE
11825 # is assumed to contain the absolute OpenSSL source directory.
11826 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11827 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11828 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11829 mkdir -p `dirname $F`
11830 ln -s $OPENSSL_SOURCE/$F $F
11833 To be absolutely sure not to disturb the source tree, a "make clean"
11834 is a good thing. If it isn't successful, don't worry about it,
11835 it probably means the source directory is very clean.
11839 * Make sure any ENGINE control commands make local copies of string
11840 pointers passed to them whenever necessary. Otherwise it is possible
11841 the caller may have overwritten (or deallocated) the original string
11842 data when a later ENGINE operation tries to use the stored values.
11844 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11846 * Improve diagnostics in file reading and command-line digests.
11848 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11850 * Add AES modes CFB and OFB to the object database. Correct an
11851 error in AES-CFB decryption.
11855 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11856 allows existing EVP_CIPHER_CTX structures to be reused after
11857 calling `EVP_*Final()`. This behaviour is used by encryption
11858 BIOs and some applications. This has the side effect that
11859 applications must explicitly clean up cipher contexts with
11860 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11864 * Check the values of dna and dnb in bn_mul_recursive before calling
11865 bn_mul_comba (a non zero value means the a or b arrays do not contain
11866 n2 elements) and fallback to bn_mul_normal if either is not zero.
11870 * Fix escaping of non-ASCII characters when using the -subj option
11871 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11875 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11876 form for "surname", serialNumber has no short form.
11877 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11878 therefore remove "mail" short name for "internet 7".
11879 The OID for unique identifiers in X509 certificates is
11880 x500UniqueIdentifier, not uniqueIdentifier.
11881 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11885 * Add an "init" command to the ENGINE config module and auto initialize
11886 ENGINEs. Without any "init" command the ENGINE will be initialized
11887 after all ctrl commands have been executed on it. If init=1 the
11888 ENGINE is initialized at that point (ctrls before that point are run
11889 on the uninitialized ENGINE and after on the initialized one). If
11890 init=0 then the ENGINE will not be initialized at all.
11894 * Fix the 'app_verify_callback' interface so that the user-defined
11895 argument is actually passed to the callback: In the
11896 SSL_CTX_set_cert_verify_callback() prototype, the callback
11897 declaration has been changed from
11900 int (*cb)(X509_STORE_CTX *,void *);
11901 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11902 i=s->ctx->app_verify_callback(&ctx)
11903 has been changed into
11904 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11906 To update applications using SSL_CTX_set_cert_verify_callback(),
11907 a dummy argument can be added to their callback functions.
11909 *D. K. Smetters <smetters@parc.xerox.com>*
11911 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11913 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11915 * Add and OPENSSL_LOAD_CONF define which will cause
11916 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11917 This allows older applications to transparently support certain
11918 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11919 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11920 load the config file and OPENSSL_add_all_algorithms_conf() which will
11921 always load it have also been added.
11925 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11926 Adjust NIDs and EVP layer.
11928 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11930 * Config modules support in openssl utility.
11932 Most commands now load modules from the config file,
11933 though in a few (such as version) this isn't done
11934 because it couldn't be used for anything.
11936 In the case of ca and req the config file used is
11937 the same as the utility itself: that is the -config
11938 command line option can be used to specify an
11943 * Move default behaviour from OPENSSL_config(). If appname is NULL
11944 use "openssl_conf" if filename is NULL use default openssl config file.
11948 * Add an argument to OPENSSL_config() to allow the use of an alternative
11949 config section name. Add a new flag to tolerate a missing config file
11950 and move code to CONF_modules_load_file().
11954 * Support for crypto accelerator cards from Accelerated Encryption
11955 Processing, www.aep.ie. (Use engine 'aep')
11956 The support was copied from 0.9.6c [engine] and adapted/corrected
11957 to work with the new engine framework.
11959 *AEP Inc. and Richard Levitte*
11961 * Support for SureWare crypto accelerator cards from Baltimore
11962 Technologies. (Use engine 'sureware')
11963 The support was copied from 0.9.6c [engine] and adapted
11964 to work with the new engine framework.
11968 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11969 make the newer ENGINE framework commands for the CHIL engine work.
11971 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11973 * Make it possible to produce shared libraries on ReliantUNIX.
11975 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11977 * Add the configuration target debug-linux-ppro.
11978 Make 'openssl rsa' use the general key loading routines
11979 implemented in `apps.c`, and make those routines able to
11980 handle the key format FORMAT_NETSCAPE and the variant
11983 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11985 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11987 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11989 * Add -keyform to rsautl, and document -engine.
11991 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11993 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11994 BIO_R_NO_SUCH_FILE error code rather than the generic
11995 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11999 * Add new functions
12000 ERR_peek_last_error
12001 ERR_peek_last_error_line
12002 ERR_peek_last_error_line_data.
12003 These are similar to
12005 ERR_peek_error_line
12006 ERR_peek_error_line_data,
12007 but report on the latest error recorded rather than the first one
12008 still in the error queue.
12010 *Ben Laurie, Bodo Moeller*
12012 * default_algorithms option in ENGINE config module. This allows things
12014 default_algorithms = ALL
12015 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12019 * Preliminary ENGINE config module.
12023 * New experimental application configuration code.
12027 * Change the AES code to follow the same name structure as all other
12028 symmetric ciphers, and behave the same way. Move everything to
12029 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12031 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12033 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12035 *Ben Laurie and Theo de Raadt*
12037 * Add option to output public keys in req command.
12039 *Massimiliano Pala madwolf@openca.org*
12041 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12042 (up to about 10% better than before for P-192 and P-224).
12046 * New functions/macros
12048 SSL_CTX_set_msg_callback(ctx, cb)
12049 SSL_CTX_set_msg_callback_arg(ctx, arg)
12050 SSL_set_msg_callback(ssl, cb)
12051 SSL_set_msg_callback_arg(ssl, arg)
12053 to request calling a callback function
12055 void cb(int write_p, int version, int content_type,
12056 const void *buf, size_t len, SSL *ssl, void *arg)
12058 whenever a protocol message has been completely received
12059 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12060 protocol version according to which the SSL library interprets
12061 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12062 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12063 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12064 specification (change_cipher_spec(20), alert(21), handshake(22)).
12065 'buf' and 'len' point to the actual message, 'ssl' to the
12066 SSL object, and 'arg' is the application-defined value set by
12067 SSL[_CTX]_set_msg_callback_arg().
12069 'openssl s_client' and 'openssl s_server' have new '-msg' options
12070 to enable a callback that displays all protocol messages.
12074 * Change the shared library support so shared libraries are built as
12075 soon as the corresponding static library is finished, and thereby get
12076 openssl and the test programs linked against the shared library.
12077 This still only happens when the keyword "shard" has been given to
12078 the configuration scripts.
12080 NOTE: shared library support is still an experimental thing, and
12081 backward binary compatibility is still not guaranteed.
12083 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12085 * Add support for Subject Information Access extension.
12087 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12089 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12090 additional bytes when new memory had to be allocated, not just
12091 when reusing an existing buffer.
12095 * New command line and configuration option 'utf8' for the req command.
12096 This allows field values to be specified as UTF8 strings.
12100 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12101 runs for the former and machine-readable output for the latter.
12105 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12106 of the e-mail address in the DN (i.e., it will go into a certificate
12107 extension only). The new configuration file option 'email_in_dn = no'
12108 has the same effect.
12110 *Massimiliano Pala madwolf@openca.org*
12112 * Change all functions with names starting with `des_` to be starting
12113 with `DES_` instead. Add wrappers that are compatible with libdes,
12114 but are named `_ossl_old_des_*`. Finally, add macros that map the
12115 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12116 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12117 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12120 Since we provide two compatibility mappings, the user needs to
12121 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12122 compatibility is desired. The default (i.e., when that macro
12123 isn't defined) is OpenSSL 0.9.6c compatibility.
12125 There are also macros that enable and disable the support of old
12126 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12127 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12128 are defined, the default will apply: to support the old des routines.
12130 In either case, one must include openssl/des.h to get the correct
12131 definitions. Do not try to just include openssl/des_old.h, that
12134 NOTE: This is a major break of an old API into a new one. Software
12135 authors are encouraged to switch to the `DES_` style functions. Some
12136 time in the future, des_old.h and the libdes compatibility functions
12137 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12138 default), and then completely removed.
12142 * Test for certificates which contain unsupported critical extensions.
12143 If such a certificate is found during a verify operation it is
12144 rejected by default: this behaviour can be overridden by either
12145 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12146 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12147 X509_supported_extension() has also been added which returns 1 if a
12148 particular extension is supported.
12152 * Modify the behaviour of EVP cipher functions in similar way to digests
12153 to retain compatibility with existing code.
12157 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12158 compatibility with existing code. In particular the 'ctx' parameter does
12159 not have to be to be initialized before the call to EVP_DigestInit() and
12160 it is tidied up after a call to EVP_DigestFinal(). New function
12161 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12162 EVP_MD_CTX_copy() changed to not require the destination to be
12163 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12164 requires the destination to be valid.
12166 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12167 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12171 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12172 so that complete 'Handshake' protocol structures are kept in memory
12173 instead of overwriting 'msg_type' and 'length' with 'body' data.
12177 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12179 *Massimo Santin via Richard Levitte*
12181 * Major restructuring to the underlying ENGINE code. This includes
12182 reduction of linker bloat, separation of pure "ENGINE" manipulation
12183 (initialisation, etc) from functionality dealing with implementations
12184 of specific crypto interfaces. This change also introduces integrated
12185 support for symmetric ciphers and digest implementations - so ENGINEs
12186 can now accelerate these by providing EVP_CIPHER and EVP_MD
12187 implementations of their own. This is detailed in
12188 [crypto/engine/README.md](crypto/engine/README.md)
12189 as it couldn't be adequately described here. However, there are a few
12190 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12191 were changed in the original introduction of ENGINE code have now
12192 reverted back - the hooking from this code to ENGINE is now a good
12193 deal more passive and at run-time, operations deal directly with
12194 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12195 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12196 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12197 they were not being used by the framework as there is no concept of a
12198 BIGNUM_METHOD and they could not be generalised to the new
12199 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12200 ENGINE_cpy() has been removed as it cannot be consistently defined in
12205 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12209 * Change mkdef.pl to sort symbols that get the same entry number,
12210 and make sure the automatically generated functions `ERR_load_*`
12211 become part of libeay.num as well.
12215 * New function SSL_renegotiate_pending(). This returns true once
12216 renegotiation has been requested (either SSL_renegotiate() call
12217 or HelloRequest/ClientHello received from the peer) and becomes
12218 false once a handshake has been completed.
12219 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12220 sends a HelloRequest, but does not ensure that a handshake takes
12221 place. SSL_renegotiate_pending() is useful for checking if the
12222 client has followed the request.)
12226 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12227 By default, clients may request session resumption even during
12228 renegotiation (if session ID contexts permit); with this option,
12229 session resumption is possible only in the first handshake.
12231 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12232 more bits available for options that should not be part of
12233 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12237 * Add some demos for certificate and certificate request creation.
12241 * Make maximum certificate chain size accepted from the peer application
12242 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12243 "Douglas E. Engert" <deengert@anl.gov>.
12247 * Add support for shared libraries for Unixware-7
12248 (Boyd Lynn Gerber <gerberb@zenez.com>).
12252 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12253 be done prior to destruction. Use this to unload error strings from
12254 ENGINEs that load their own error strings. NB: This adds two new API
12255 functions to "get" and "set" this destroy handler in an ENGINE.
12259 * Alter all existing ENGINE implementations (except "openssl" and
12260 "openbsd") to dynamically instantiate their own error strings. This
12261 makes them more flexible to be built both as statically-linked ENGINEs
12262 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12263 Also, add stub code to each that makes building them as self-contained
12264 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12268 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12269 implementations into applications that are completely implemented in
12270 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12271 commands that can be used to configure what shared-library to load and
12272 to control aspects of the way it is handled. Also, made an update to
12273 the [README-Engine.md](README-Engine.md) file
12274 that brings its information up-to-date and
12275 provides some information and instructions on the "dynamic" ENGINE
12276 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12280 * Make it possible to unload ranges of ERR strings with a new
12281 "ERR_unload_strings" function.
12285 * Add a copy() function to EVP_MD.
12289 * Make EVP_MD routines take a context pointer instead of just the
12290 md_data void pointer.
12294 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12295 that the digest can only process a single chunk of data
12296 (typically because it is provided by a piece of
12297 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12298 is only going to provide a single chunk of data, and hence the
12299 framework needn't accumulate the data for oneshot drivers.
12303 * As with "ERR", make it possible to replace the underlying "ex_data"
12304 functions. This change also alters the storage and management of global
12305 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12306 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12307 index counters. The API functions that use this state have been changed
12308 to take a "class_index" rather than pointers to the class's local STACK
12309 and counter, and there is now an API function to dynamically create new
12310 classes. This centralisation allows us to (a) plug a lot of the
12311 thread-safety problems that existed, and (b) makes it possible to clean
12312 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12313 such data would previously have always leaked in application code and
12314 workarounds were in place to make the memory debugging turn a blind eye
12315 to it. Application code that doesn't use this new function will still
12316 leak as before, but their memory debugging output will announce it now
12317 rather than letting it slide.
12319 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12320 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12321 has a return value to indicate success or failure.
12325 * Make it possible to replace the underlying "ERR" functions such that the
12326 global state (2 LHASH tables and 2 locks) is only used by the "default"
12327 implementation. This change also adds two functions to "get" and "set"
12328 the implementation prior to it being automatically set the first time
12329 any other ERR function takes place. Ie. an application can call "get",
12330 pass the return value to a module it has just loaded, and that module
12331 can call its own "set" function using that value. This means the
12332 module's "ERR" operations will use (and modify) the error state in the
12333 application and not in its own statically linked copy of OpenSSL code.
12337 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12338 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12339 the operation, and provides a more encapsulated way for external code
12340 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12341 to use these functions rather than manually incrementing the counts.
12343 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12347 * Add EVP test program.
12351 * Add symmetric cipher support to ENGINE. Expect the API to change!
12355 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12356 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12357 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12358 These allow a CRL to be built without having to access X509_CRL fields
12359 directly. Modify 'ca' application to use new functions.
12363 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12364 bug workarounds. Rollback attack detection is a security feature.
12365 The problem will only arise on OpenSSL servers when TLSv1 is not
12366 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12367 Software authors not wanting to support TLSv1 will have special reasons
12368 for their choice and can explicitly enable this option.
12370 *Bodo Moeller, Lutz Jaenicke*
12372 * Rationalise EVP so it can be extended: don't include a union of
12373 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12374 (similar to those existing for EVP_CIPHER_CTX).
12379 EVP_MD_CTX_init(&md); /* new function call */
12380 EVP_DigestInit(&md, EVP_sha1());
12381 EVP_DigestUpdate(&md, in, len);
12382 EVP_DigestFinal(&md, out, NULL);
12383 EVP_MD_CTX_cleanup(&md); /* new function call */
12387 * Make DES key schedule conform to the usual scheme, as well as
12388 correcting its structure. This means that calls to DES functions
12389 now have to pass a pointer to a des_key_schedule instead of a
12390 plain des_key_schedule (which was actually always a pointer
12393 des_key_schedule ks;
12395 des_set_key_checked(..., &ks);
12396 des_ncbc_encrypt(..., &ks, ...);
12398 (Note that a later change renames 'des_...' into 'DES_...'.)
12402 * Initial reduction of linker bloat: the use of some functions, such as
12403 PEM causes large amounts of unused functions to be linked in due to
12404 poor organisation. For example pem_all.c contains every PEM function
12405 which has a knock on effect of linking in large amounts of (unused)
12406 ASN1 code. Grouping together similar functions and splitting unrelated
12407 functions prevents this.
12411 * Cleanup of EVP macros.
12415 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12416 correct `_ecb suffix`.
12420 * Add initial OCSP responder support to ocsp application. The
12421 revocation information is handled using the text based index
12422 use by the ca application. The responder can either handle
12423 requests generated internally, supplied in files (for example
12424 via a CGI script) or using an internal minimal server.
12428 * Add configuration choices to get zlib compression for TLS.
12432 * Changes to Kerberos SSL for RFC 2712 compliance:
12433 1. Implemented real KerberosWrapper, instead of just using
12434 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12435 2. Implemented optional authenticator field of KerberosWrapper.
12437 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12438 and authenticator structs; see crypto/krb5/.
12440 Generalized Kerberos calls to support multiple Kerberos libraries.
12441 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12442 via Richard Levitte*
12444 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12445 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12446 values for each of the key sizes rather than having just
12447 parameters (and 'speed' generating keys each time).
12451 * Speed up EVP routines.
12454 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12455 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12456 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12457 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12459 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12460 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12461 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12464 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12466 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12470 * Added the OS2-EMX target.
12472 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12474 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12475 New functions to support `NCONF `routines in extension code.
12476 New function `CONF_set_nconf()`
12477 to allow functions which take an `NCONF` to also handle the old `LHASH`
12478 structure: this means that the old `CONF` compatible routines can be
12479 retained (in particular w.rt. extensions) without having to duplicate the
12480 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12484 * Enhance the general user interface with mechanisms for inner control
12485 and with possibilities to have yes/no kind of prompts.
12489 * Change all calls to low level digest routines in the library and
12490 applications to use EVP. Add missing calls to HMAC_cleanup() and
12491 don't assume HMAC_CTX can be copied using memcpy().
12493 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12495 * Add the possibility to control engines through control names but with
12496 arbitrary arguments instead of just a string.
12497 Change the key loaders to take a UI_METHOD instead of a callback
12498 function pointer. NOTE: this breaks binary compatibility with earlier
12499 versions of OpenSSL [engine].
12500 Adapt the nCipher code for these new conditions and add a card insertion
12505 * Enhance the general user interface with mechanisms to better support
12506 dialog box interfaces, application-defined prompts, the possibility
12507 to use defaults (for example default passwords from somewhere else)
12508 and interrupts/cancellations.
12512 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12513 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12517 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12518 tidy up some unnecessarily weird code in 'sk_new()').
12520 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12522 * Change the key loading routines for ENGINEs to use the same kind
12523 callback (pem_password_cb) as all other routines that need this
12528 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12529 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12530 than this minimum value is recommended.
12534 * New random seeder for OpenVMS, using the system process statistics
12535 that are easily reachable.
12539 * Windows apparently can't transparently handle global
12540 variables defined in DLLs. Initialisations such as:
12542 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12544 won't compile. This is used by the any applications that need to
12545 declare their own ASN1 modules. This was fixed by adding the option
12546 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12547 needed for static libraries under Win32.
12551 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12552 setting of purpose and trust fields. New X509_STORE trust and
12553 purpose functions and tidy up setting in other SSL functions.
12557 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12558 structure. These are inherited by X509_STORE_CTX when it is
12559 initialised. This allows various defaults to be set in the
12560 X509_STORE structure (such as flags for CRL checking and custom
12561 purpose or trust settings) for functions which only use X509_STORE_CTX
12562 internally such as S/MIME.
12564 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12565 trust settings if they are not set in X509_STORE. This allows X509_STORE
12566 purposes and trust (in S/MIME for example) to override any set by default.
12568 Add command line options for CRL checking to smime, s_client and s_server
12573 * Initial CRL based revocation checking. If the CRL checking flag(s)
12574 are set then the CRL is looked up in the X509_STORE structure and
12575 its validity and signature checked, then if the certificate is found
12576 in the CRL the verify fails with a revoked error.
12578 Various new CRL related callbacks added to X509_STORE_CTX structure.
12580 Command line options added to 'verify' application to support this.
12582 This needs some additional work, such as being able to handle multiple
12583 CRLs with different times, extension based lookup (rather than just
12584 by subject name) and ultimately more complete V2 CRL extension
12589 * Add a general user interface API (crypto/ui/). This is designed
12590 to replace things like des_read_password and friends (backward
12591 compatibility functions using this new API are provided).
12592 The purpose is to remove prompting functions from the DES code
12593 section as well as provide for prompting through dialog boxes in
12594 a window system and the like.
12598 * Add "ex_data" support to ENGINE so implementations can add state at a
12599 per-structure level rather than having to store it globally.
12603 * Make it possible for ENGINE structures to be copied when retrieved by
12604 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12605 This causes the "original" ENGINE structure to act like a template,
12606 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12607 operational state can be localised to each ENGINE structure, despite the
12608 fact they all share the same "methods". New ENGINE structures returned in
12609 this case have no functional references and the return value is the single
12610 structural reference. This matches the single structural reference returned
12611 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12616 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12617 needs to match any other type at all we need to manually clear the
12622 * Changes to the "openssl engine" utility to include;
12623 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12624 about an ENGINE's available control commands.
12625 - executing control commands from command line arguments using the
12626 '-pre' and '-post' switches. '-post' is only used if '-t' is
12627 specified and the ENGINE is successfully initialised. The syntax for
12628 the individual commands are colon-separated, for example;
12629 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12633 * New dynamic control command support for ENGINEs. ENGINEs can now
12634 declare their own commands (numbers), names (strings), descriptions,
12635 and input types for run-time discovery by calling applications. A
12636 subset of these commands are implicitly classed as "executable"
12637 depending on their input type, and only these can be invoked through
12638 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12639 can be based on user input, config files, etc). The distinction is
12640 that "executable" commands cannot return anything other than a boolean
12641 result and can only support numeric or string input, whereas some
12642 discoverable commands may only be for direct use through
12643 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12644 pointers, or other custom uses. The "executable" commands are to
12645 support parameterisations of ENGINE behaviour that can be
12646 unambiguously defined by ENGINEs and used consistently across any
12647 OpenSSL-based application. Commands have been added to all the
12648 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12649 control over shared-library paths without source code alterations.
12653 * Changed all ENGINE implementations to dynamically allocate their
12654 ENGINEs rather than declaring them statically. Apart from this being
12655 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12656 this also allows the implementations to compile without using the
12657 internal engine_int.h header.
12661 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12662 'const' value. Any code that should be able to modify a RAND_METHOD
12663 should already have non-const pointers to it (ie. they should only
12664 modify their own ones).
12668 * Made a variety of little tweaks to the ENGINE code.
12669 - "atalla" and "ubsec" string definitions were moved from header files
12670 to C code. "nuron" string definitions were placed in variables
12671 rather than hard-coded - allowing parameterisation of these values
12672 later on via ctrl() commands.
12673 - Removed unused "#if 0"'d code.
12674 - Fixed engine list iteration code so it uses ENGINE_free() to release
12675 structural references.
12676 - Constified the RAND_METHOD element of ENGINE structures.
12677 - Constified various get/set functions as appropriate and added
12678 missing functions (including a catch-all ENGINE_cpy that duplicates
12679 all ENGINE values onto a new ENGINE except reference counts/state).
12680 - Removed NULL parameter checks in get/set functions. Setting a method
12681 or function to NULL is a way of cancelling out a previously set
12682 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12683 and doesn't justify the extra error symbols and code.
12684 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12685 flags from engine_int.h to engine.h.
12686 - Changed prototypes for ENGINE handler functions (init(), finish(),
12687 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12691 * Implement binary inversion algorithm for BN_mod_inverse in addition
12692 to the algorithm using long division. The binary algorithm can be
12693 used only if the modulus is odd. On 32-bit systems, it is faster
12694 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12695 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12696 up to 450 bits. In 64-bit environments, the binary algorithm
12697 appears to be advantageous for much longer moduli; here we use it
12698 for moduli up to 2048 bits.
12702 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12703 could not support the combine flag in choice fields.
12707 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12708 extensions from a certificate request to the certificate.
12712 * Allow multiple 'certopt' and 'nameopt' options to be separated
12713 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12714 file: this allows the display of the certificate about to be
12715 signed to be customised, to allow certain fields to be included
12716 or excluded and extension details. The old system didn't display
12717 multicharacter strings properly, omitted fields not in the policy
12718 and couldn't display additional details such as extensions.
12722 * Function EC_POINTs_mul for multiple scalar multiplication
12723 of an arbitrary number of elliptic curve points
12724 \sum scalars[i]*points[i],
12725 optionally including the generator defined for the EC_GROUP:
12726 scalar*generator + \sum scalars[i]*points[i].
12728 EC_POINT_mul is a simple wrapper function for the typical case
12729 that the point list has just one item (besides the optional
12734 * First EC_METHODs for curves over GF(p):
12736 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12737 operations and provides various method functions that can also
12738 operate with faster implementations of modular arithmetic.
12740 EC_GFp_mont_method() reuses most functions that are part of
12741 EC_GFp_simple_method, but uses Montgomery arithmetic.
12743 *Bodo Moeller; point addition and point doubling
12744 implementation directly derived from source code provided by
12745 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12747 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12748 crypto/ec/ec_lib.c):
12750 Curves are EC_GROUP objects (with an optional group generator)
12751 based on EC_METHODs that are built into the library.
12753 Points are EC_POINT objects based on EC_GROUP objects.
12755 Most of the framework would be able to handle curves over arbitrary
12756 finite fields, but as there are no obvious types for fields other
12757 than GF(p), some functions are limited to that for now.
12761 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12762 that the file contains a complete HTTP response.
12766 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12767 change the def and num file printf format specifier from "%-40sXXX"
12768 to "%-39s XXX". The latter will always guarantee a space after the
12769 field while the former will cause them to run together if the field
12770 is 40 of more characters long.
12774 * Constify the cipher and digest 'method' functions and structures
12775 and modify related functions to take constant EVP_MD and EVP_CIPHER
12780 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12781 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12785 * Modify `EVP_Digest*()` routines so they now return values. Although the
12786 internal software routines can never fail additional hardware versions
12791 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12793 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12794 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12797 ERR_R_NESTED_ASN1_ERROR
12799 ERR_R_MISSING_ASN1_EOS
12800 were 4 .. 9, conflicting with
12801 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12803 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12804 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12806 Add new error code 'ERR_R_INTERNAL_ERROR'.
12810 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12815 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12816 sets the subject name for a new request or supersedes the
12817 subject name in a given request. Formats that can be parsed are
12818 'CN=Some Name, OU=myOU, C=IT'
12820 'CN=Some Name/OU=myOU/C=IT'.
12822 Add options '-batch' and '-verbose' to 'openssl req'.
12824 *Massimiliano Pala <madwolf@hackmasters.net>*
12826 * Introduce the possibility to access global variables through
12827 functions on platform were that's the best way to handle exporting
12828 global variables in shared libraries. To enable this functionality,
12829 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12830 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12831 is normally done by Configure or something similar).
12833 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12834 in the source file (foo.c) like this:
12836 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12837 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12839 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12840 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12842 OPENSSL_DECLARE_GLOBAL(int,foo);
12843 #define foo OPENSSL_GLOBAL_REF(foo)
12844 OPENSSL_DECLARE_GLOBAL(double,bar);
12845 #define bar OPENSSL_GLOBAL_REF(bar)
12847 The #defines are very important, and therefore so is including the
12848 header file everywhere where the defined globals are used.
12850 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12851 of ASN.1 items, but that structure is a bit different.
12853 The largest change is in util/mkdef.pl which has been enhanced with
12854 better and easier to understand logic to choose which symbols should
12855 go into the Windows .def files as well as a number of fixes and code
12856 cleanup (among others, algorithm keywords are now sorted
12857 lexicographically to avoid constant rewrites).
12861 * In BN_div() keep a copy of the sign of 'num' before writing the
12862 result to 'rm' because if rm==num the value will be overwritten
12863 and produce the wrong result if 'num' is negative: this caused
12864 problems with BN_mod() and BN_nnmod().
12868 * Function OCSP_request_verify(). This checks the signature on an
12869 OCSP request and verifies the signer certificate. The signer
12870 certificate is just checked for a generic purpose and OCSP request
12875 * Add OCSP_check_validity() function to check the validity of OCSP
12876 responses. OCSP responses are prepared in real time and may only
12877 be a few seconds old. Simply checking that the current time lies
12878 between thisUpdate and nextUpdate max reject otherwise valid responses
12879 caused by either OCSP responder or client clock inaccuracy. Instead
12880 we allow thisUpdate and nextUpdate to fall within a certain period of
12881 the current time. The age of the response can also optionally be
12882 checked. Two new options -validity_period and -status_age added to
12887 * If signature or public key algorithm is unrecognized print out its
12888 OID rather that just UNKNOWN.
12892 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12893 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12894 ID to be generated from the issuer certificate alone which can then be
12895 passed to OCSP_id_issuer_cmp().
12899 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12900 ASN1 modules to export functions returning ASN1_ITEM pointers
12901 instead of the ASN1_ITEM structures themselves. This adds several
12902 new macros which allow the underlying ASN1 function/structure to
12903 be accessed transparently. As a result code should not use ASN1_ITEM
12904 references directly (such as &X509_it) but instead use the relevant
12905 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12906 use of the new ASN1 code on platforms where exporting structures
12907 is problematical (for example in shared libraries) but exporting
12908 functions returning pointers to structures is not.
12912 * Add support for overriding the generation of SSL/TLS session IDs.
12913 These callbacks can be registered either in an SSL_CTX or per SSL.
12914 The purpose of this is to allow applications to control, if they wish,
12915 the arbitrary values chosen for use as session IDs, particularly as it
12916 can be useful for session caching in multiple-server environments. A
12917 command-line switch for testing this (and any client code that wishes
12918 to use such a feature) has been added to "s_server".
12920 *Geoff Thorpe, Lutz Jaenicke*
12922 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12923 of the form `#if defined(...) || defined(...) || ...` and
12924 `#if !defined(...) && !defined(...) && ...`. This also avoids
12925 the growing number of special cases it was previously handling.
12929 * Make all configuration macros available for application by making
12930 sure they are available in opensslconf.h, by giving them names starting
12931 with `OPENSSL_` to avoid conflicts with other packages and by making
12932 sure e_os2.h will cover all platform-specific cases together with
12934 Additionally, it is now possible to define configuration/platform-
12935 specific names (called "system identities"). In the C code, these
12936 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12937 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12938 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12943 * New option -set_serial to 'req' and 'x509' this allows the serial
12944 number to use to be specified on the command line. Previously self
12945 signed certificates were hard coded with serial number 0 and the
12946 CA options of 'x509' had to use a serial number in a file which was
12951 * New options to 'ca' utility to support V2 CRL entry extensions.
12952 Currently CRL reason, invalidity date and hold instruction are
12953 supported. Add new CRL extensions to V3 code and some new objects.
12957 * New function EVP_CIPHER_CTX_set_padding() this is used to
12958 disable standard block padding (aka PKCS#5 padding) in the EVP
12959 API, which was previously mandatory. This means that the data is
12960 not padded in any way and so the total length much be a multiple
12961 of the block size, otherwise an error occurs.
12965 * Initial (incomplete) OCSP SSL support.
12969 * New function OCSP_parse_url(). This splits up a URL into its host,
12970 port and path components: primarily to parse OCSP URLs. New -url
12971 option to ocsp utility.
12975 * New nonce behavior. The return value of OCSP_check_nonce() now
12976 reflects the various checks performed. Applications can decide
12977 whether to tolerate certain situations such as an absent nonce
12978 in a response when one was present in a request: the ocsp application
12979 just prints out a warning. New function OCSP_add1_basic_nonce()
12980 this is to allow responders to include a nonce in a response even if
12981 the request is nonce-less.
12985 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12986 skipped when using openssl x509 multiple times on a single input file,
12987 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12991 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12992 set string type: to handle setting ASN1_TIME structures. Fix ca
12993 utility to correctly initialize revocation date of CRLs.
12997 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12998 the clients preferred ciphersuites and rather use its own preferences.
12999 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13000 Internet Explorer by ensuring unchanged hash method during stepup.
13001 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13005 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13006 to aes and add a new 'exist' option to print out symbols that don't
13011 * Additional options to ocsp utility to allow flags to be set and
13012 additional certificates supplied.
13016 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13017 OCSP client a number of certificate to only verify the response
13022 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13023 handle the new API. Currently only ECB, CBC modes supported. Add new
13026 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13027 Encryption Standard (AES) Ciphersuites for Transport Layer
13028 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13029 not enabled by default and were not part of the "ALL" ciphersuite
13030 alias because they were not yet official; they could be
13031 explicitly requested by specifying the "AESdraft" ciphersuite
13032 group alias. In the final release of OpenSSL 0.9.7, the group
13033 alias is called "AES" and is part of "ALL".)
13035 *Ben Laurie, Steve Henson, Bodo Moeller*
13037 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13038 request to response.
13042 * Functions for OCSP responders. OCSP_request_onereq_count(),
13043 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13044 extract information from a certificate request. OCSP_response_create()
13045 creates a response and optionally adds a basic response structure.
13046 OCSP_basic_add1_status() adds a complete single response to a basic
13047 response and returns the OCSP_SINGLERESP structure just added (to allow
13048 extensions to be included for example). OCSP_basic_add1_cert() adds a
13049 certificate to a basic response and OCSP_basic_sign() signs a basic
13050 response with various flags. New helper functions ASN1_TIME_check()
13051 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13052 (converts ASN1_TIME to GeneralizedTime).
13056 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13057 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13058 structure from a certificate. X509_pubkey_digest() digests the public_key
13059 contents: this is used in various key identifiers.
13063 * Make sk_sort() tolerate a NULL argument.
13065 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13067 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13068 passed by the function are trusted implicitly. If any of them signed the
13069 response then it is assumed to be valid and is not verified.
13073 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13074 to data. This was previously part of the PKCS7 ASN1 code. This
13075 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13076 *Steve Henson, reported by Kenneth R. Robinette
13077 <support@securenetterm.com>*
13079 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13080 routines: without these tracing memory leaks is very painful.
13081 Fix leaks in PKCS12 and PKCS7 routines.
13085 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13086 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13087 effectively meant GeneralizedTime would never be used. Now it
13088 is initialised to -1 but X509_time_adj() now has to check the value
13089 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13090 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13091 *Steve Henson, reported by Kenneth R. Robinette
13092 <support@securenetterm.com>*
13094 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13095 result in a zero length in the ASN1_INTEGER structure which was
13096 not consistent with the structure when d2i_ASN1_INTEGER() was used
13097 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13098 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13099 where it did not print out a minus for negative ASN1_INTEGER.
13103 * Add summary printout to ocsp utility. The various functions which
13104 convert status values to strings have been renamed to:
13105 OCSP_response_status_str(), OCSP_cert_status_str() and
13106 OCSP_crl_reason_str() and are no longer static. New options
13107 to verify nonce values and to disable verification. OCSP response
13108 printout format cleaned up.
13112 * Add additional OCSP certificate checks. These are those specified
13113 in RFC2560. This consists of two separate checks: the CA of the
13114 certificate being checked must either be the OCSP signer certificate
13115 or the issuer of the OCSP signer certificate. In the latter case the
13116 OCSP signer certificate must contain the OCSP signing extended key
13117 usage. This check is performed by attempting to match the OCSP
13118 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13119 in the OCSP_CERTID structures of the response.
13123 * Initial OCSP certificate verification added to OCSP_basic_verify()
13124 and related routines. This uses the standard OpenSSL certificate
13125 verify routines to perform initial checks (just CA validity) and
13126 to obtain the certificate chain. Then additional checks will be
13127 performed on the chain. Currently the root CA is checked to see
13128 if it is explicitly trusted for OCSP signing. This is used to set
13129 a root CA as a global signing root: that is any certificate that
13130 chains to that CA is an acceptable OCSP signing certificate.
13134 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13135 extensions from a separate configuration file.
13136 As when reading extensions from the main configuration file,
13137 the '-extensions ...' option may be used for specifying the
13140 *Massimiliano Pala <madwolf@comune.modena.it>*
13142 * New OCSP utility. Allows OCSP requests to be generated or
13143 read. The request can be sent to a responder and the output
13144 parsed, outputted or printed in text form. Not complete yet:
13145 still needs to check the OCSP response validity.
13149 * New subcommands for 'openssl ca':
13150 `openssl ca -status <serial>` prints the status of the cert with
13151 the given serial number (according to the index file).
13152 `openssl ca -updatedb` updates the expiry status of certificates
13155 *Massimiliano Pala <madwolf@comune.modena.it>*
13157 * New '-newreq-nodes' command option to CA.pl. This is like
13158 '-newreq', but calls 'openssl req' with the '-nodes' option
13159 so that the resulting key is not encrypted.
13161 *Damien Miller <djm@mindrot.org>*
13163 * New configuration for the GNU Hurd.
13165 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13167 * Initial code to implement OCSP basic response verify. This
13168 is currently incomplete. Currently just finds the signer's
13169 certificate and verifies the signature on the response.
13173 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13174 value of OPENSSLDIR. This is available via the new '-d' option
13175 to 'openssl version', and is also included in 'openssl version -a'.
13179 * Allowing defining memory allocation callbacks that will be given
13180 file name and line number information in additional arguments
13181 (a `const char*` and an int). The basic functionality remains, as
13182 well as the original possibility to just replace malloc(),
13183 realloc() and free() by functions that do not know about these
13184 additional arguments. To register and find out the current
13185 settings for extended allocation functions, the following
13186 functions are provided:
13188 CRYPTO_set_mem_ex_functions
13189 CRYPTO_set_locked_mem_ex_functions
13190 CRYPTO_get_mem_ex_functions
13191 CRYPTO_get_locked_mem_ex_functions
13193 These work the same way as CRYPTO_set_mem_functions and friends.
13194 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13195 extended allocation function is enabled.
13196 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13197 a conventional allocation function is enabled.
13199 *Richard Levitte, Bodo Moeller*
13201 * Finish off removing the remaining LHASH function pointer casts.
13202 There should no longer be any prototype-casting required when using
13203 the LHASH abstraction, and any casts that remain are "bugs". See
13204 the callback types and macros at the head of lhash.h for details
13205 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13209 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13210 If /dev/[u]random devices are not available or do not return enough
13211 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13213 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13214 /etc/entropy will be queried once each in this sequence, querying stops
13215 when enough entropy was collected without querying more sockets.
13219 * Change the Unix RAND_poll() variant to be able to poll several
13220 random devices, as specified by DEVRANDOM, until a sufficient amount
13221 of data has been collected. We spend at most 10 ms on each file
13222 (select timeout) and read in non-blocking mode. DEVRANDOM now
13223 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13224 (previously it was just the string "/dev/urandom"), so on typical
13225 platforms the 10 ms delay will never occur.
13226 Also separate out the Unix variant to its own file, rand_unix.c.
13227 For VMS, there's a currently-empty rand_vms.c.
13231 * Move OCSP client related routines to ocsp_cl.c. These
13232 provide utility functions which an application needing
13233 to issue a request to an OCSP responder and analyse the
13234 response will typically need: as opposed to those which an
13235 OCSP responder itself would need which will be added later.
13237 OCSP_request_sign() signs an OCSP request with an API similar
13238 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13239 response. OCSP_response_get1_basic() extracts basic response
13240 from response. OCSP_resp_find_status(): finds and extracts status
13241 information from an OCSP_CERTID structure (which will be created
13242 when the request structure is built). These are built from lower
13243 level functions which work on OCSP_SINGLERESP structures but
13244 won't normally be used unless the application wishes to examine
13245 extensions in the OCSP response for example.
13247 Replace nonce routines with a pair of functions.
13248 OCSP_request_add1_nonce() adds a nonce value and optionally
13249 generates a random value. OCSP_check_nonce() checks the
13250 validity of the nonce in an OCSP response.
13254 * Change function OCSP_request_add() to OCSP_request_add0_id().
13255 This doesn't copy the supplied OCSP_CERTID and avoids the
13256 need to free up the newly created id. Change return type
13257 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13258 This can then be used to add extensions to the request.
13259 Deleted OCSP_request_new(), since most of its functionality
13260 is now in OCSP_REQUEST_new() (and the case insensitive name
13261 clash) apart from the ability to set the request name which
13262 will be added elsewhere.
13266 * Update OCSP API. Remove obsolete extensions argument from
13267 various functions. Extensions are now handled using the new
13268 OCSP extension code. New simple OCSP HTTP function which
13269 can be used to send requests and parse the response.
13273 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13274 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13275 uses the special reorder version of SET OF to sort the attributes
13276 and reorder them to match the encoded order. This resolves a long
13277 standing problem: a verify on a PKCS7 structure just after signing
13278 it used to fail because the attribute order did not match the
13279 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13280 it uses the received order. This is necessary to tolerate some broken
13281 software that does not order SET OF. This is handled by encoding
13282 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13283 to produce the required SET OF.
13287 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13288 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13289 files to get correct declarations of the ASN.1 item variables.
13293 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13294 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13295 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13296 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13297 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13298 ASN1_ITEM and no wrapper functions.
13302 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13303 replace the old function pointer based I/O routines. Change most of
13304 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13308 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13309 lines, recognize more "algorithms" that can be deselected, and make
13310 it complain about algorithm deselection that isn't recognised.
13314 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13315 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13316 to use new functions. Add NO_ASN1_OLD which can be set to remove
13317 some old style ASN1 functions: this can be used to determine if old
13318 code will still work when these eventually go away.
13322 * New extension functions for OCSP structures, these follow the
13323 same conventions as certificates and CRLs.
13327 * New function X509V3_add1_i2d(). This automatically encodes and
13328 adds an extension. Its behaviour can be customised with various
13329 flags to append, replace or delete. Various wrappers added for
13330 certificates and CRLs.
13334 * Fix to avoid calling the underlying ASN1 print routine when
13335 an extension cannot be parsed. Correct a typo in the
13336 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13340 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13341 entries for variables.
13345 * Add functionality to `apps/openssl.c` for detecting locking
13346 problems: As the program is single-threaded, all we have
13347 to do is register a locking callback using an array for
13348 storing which locks are currently held by the program.
13352 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13353 SSL_get_ex_data_X509_STORE_idx(), which is used in
13354 ssl_verify_cert_chain() and thus can be called at any time
13355 during TLS/SSL handshakes so that thread-safety is essential.
13356 Unfortunately, the ex_data design is not at all suited
13357 for multi-threaded use, so it probably should be abolished.
13361 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13363 *Broadcom, tweaked and integrated by Geoff Thorpe*
13365 * Move common extension printing code to new function
13366 X509V3_print_extensions(). Reorganise OCSP print routines and
13367 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13371 * New function X509_signature_print() to remove duplication in some
13376 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13377 set (this was treated exactly the same as SET OF previously). This
13378 is used to reorder the STACK representing the structure to match the
13379 encoding. This will be used to get round a problem where a PKCS7
13380 structure which was signed could not be verified because the STACK
13381 order did not reflect the encoded order.
13385 * Reimplement the OCSP ASN1 module using the new code.
13389 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13390 for its ASN1 operations. The old style function pointers still exist
13391 for now but they will eventually go away.
13395 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13396 completely replaces the old ASN1 functionality with a table driven
13397 encoder and decoder which interprets an ASN1_ITEM structure describing
13398 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13399 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13400 has also been converted to the new form.
13404 * Change BN_mod_exp_recp so that negative moduli are tolerated
13405 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13406 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13407 for negative moduli.
13411 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13412 of not touching the result's sign bit.
13416 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13421 * Changed the LHASH code to use prototypes for callbacks, and created
13422 macros to declare and implement thin (optionally static) functions
13423 that provide type-safety and avoid function pointer casting for the
13424 type-specific callbacks.
13428 * Added Kerberos Cipher Suites to be used with TLS, as written in
13430 *Veers Staats <staatsvr@asc.hpc.mil>,
13431 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13433 * Reformat the FAQ so the different questions and answers can be divided
13434 in sections depending on the subject.
13438 * Have the zlib compression code load ZLIB.DLL dynamically under
13443 * New function BN_mod_sqrt for computing square roots modulo a prime
13444 (using the probabilistic Tonelli-Shanks algorithm unless
13445 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13446 be handled deterministically).
13448 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13450 * Make BN_mod_inverse faster by explicitly handling small quotients
13451 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13452 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13456 * New function BN_kronecker.
13460 * Fix BN_gcd so that it works on negative inputs; the result is
13461 positive unless both parameters are zero.
13462 Previously something reasonably close to an infinite loop was
13463 possible because numbers could be growing instead of shrinking
13464 in the implementation of Euclid's algorithm.
13468 * Fix BN_is_word() and BN_is_one() macros to take into account the
13469 sign of the number in question.
13471 Fix BN_is_word(a,w) to work correctly for w == 0.
13473 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13474 because its test if the absolute value of 'a' equals 'w'.
13475 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13476 it exists mostly for use in the implementations of BN_is_zero(),
13477 BN_is_one(), and BN_is_word().
13481 * New function BN_swap.
13485 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13486 the exponentiation functions are more likely to produce reasonable
13487 results on negative inputs.
13491 * Change BN_mod_mul so that the result is always non-negative.
13492 Previously, it could be negative if one of the factors was negative;
13493 I don't think anyone really wanted that behaviour.
13497 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13498 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13499 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13500 and add new functions:
13509 BN_mod_lshift1_quick
13511 BN_mod_lshift_quick
13513 These functions always generate non-negative results.
13515 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13516 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13518 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13519 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13520 be reduced modulo `m`.
13522 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13525 The following entry accidentally appeared in the CHANGES file
13526 distributed with OpenSSL 0.9.7. The modifications described in
13527 it do *not* apply to OpenSSL 0.9.7.
13529 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13530 was actually never needed) and in BN_mul(). The removal in BN_mul()
13531 required a small change in bn_mul_part_recursive() and the addition
13532 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13533 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13534 bn_sub_words() and bn_add_words() except they take arrays with
13540 * In 'openssl passwd', verify passwords read from the terminal
13541 unless the '-salt' option is used (which usually means that
13542 verification would just waste user's time since the resulting
13543 hash is going to be compared with some given password hash)
13544 or the new '-noverify' option is used.
13546 This is an incompatible change, but it does not affect
13547 non-interactive use of 'openssl passwd' (passwords on the command
13548 line, '-stdin' option, '-in ...' option) and thus should not
13549 cause any problems.
13553 * Remove all references to RSAref, since there's no more need for it.
13557 * Make DSO load along a path given through an environment variable
13558 (SHLIB_PATH) with shl_load().
13562 * Constify the ENGINE code as a result of BIGNUM constification.
13563 Also constify the RSA code and most things related to it. In a
13564 few places, most notable in the depth of the ASN.1 code, ugly
13565 casts back to non-const were required (to be solved at a later
13570 * Make it so the openssl application has all engines loaded by default.
13574 * Constify the BIGNUM routines a little more.
13578 * Add the following functions:
13580 ENGINE_load_cswift()
13582 ENGINE_load_atalla()
13583 ENGINE_load_nuron()
13584 ENGINE_load_builtin_engines()
13586 That way, an application can itself choose if external engines that
13587 are built-in in OpenSSL shall ever be used or not. The benefit is
13588 that applications won't have to be linked with libdl or other dso
13589 libraries unless it's really needed.
13591 Changed 'openssl engine' to load all engines on demand.
13592 Changed the engine header files to avoid the duplication of some
13593 declarations (they differed!).
13597 * 'openssl engine' can now list capabilities.
13601 * Better error reporting in 'openssl engine'.
13605 * Never call load_dh_param(NULL) in s_server.
13609 * Add engine application. It can currently list engines by name and
13610 identity, and test if they are actually available.
13614 * Improve RPM specification file by forcing symbolic linking and making
13615 sure the installed documentation is also owned by root.root.
13617 *Damien Miller <djm@mindrot.org>*
13619 * Give the OpenSSL applications more possibilities to make use of
13620 keys (public as well as private) handled by engines.
13624 * Add OCSP code that comes from CertCo.
13628 * Add VMS support for the Rijndael code.
13632 * Added untested support for Nuron crypto accelerator.
13636 * Add support for external cryptographic devices. This code was
13637 previously distributed separately as the "engine" branch.
13639 *Geoff Thorpe, Richard Levitte*
13641 * Rework the filename-translation in the DSO code. It is now possible to
13642 have far greater control over how a "name" is turned into a filename
13643 depending on the operating environment and any oddities about the
13644 different shared library filenames on each system.
13648 * Support threads on FreeBSD-elf in Configure.
13652 * Fix for SHA1 assembly problem with MASM: it produces
13653 warnings about corrupt line number information when assembling
13654 with debugging information. This is caused by the overlapping
13657 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13660 NCONF_get_number() has no error checking at all. As a replacement,
13661 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13662 promoted strongly. The old NCONF_get_number is kept around for
13663 binary backward compatibility.
13664 Make it possible for methods to load from something other than a BIO,
13665 by providing a function pointer that is given a name instead of a BIO.
13666 For example, this could be used to load configuration data from an
13671 * Fix for non blocking accept BIOs. Added new I/O special reason
13672 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13673 with non blocking I/O was not possible because no retry code was
13674 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13679 * Added the beginnings of Rijndael support.
13683 * Fix for bug in DirectoryString mask setting. Add support for
13684 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13685 to allow certificate printing to more controllable, additional
13686 'certopt' option to 'x509' to allow new printing options to be
13691 * Clean old EAY MD5 hack from e_os.h.
13695 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13697 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13698 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13700 *Joe Orton, Steve Henson*
13702 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13704 * Fix additional bug revealed by the NISCC test suite:
13706 Stop bug triggering large recursion when presented with
13707 certain ASN.1 tags ([CVE-2003-0851])
13711 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13713 * Fix various bugs revealed by running the NISCC test suite:
13715 Stop out of bounds reads in the ASN1 code when presented with
13716 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13718 If verify callback ignores invalid public key errors don't try to check
13719 certificate signature with the NULL public key.
13723 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13724 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13729 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13730 extra data after the compression methods not only for TLS 1.0
13731 but also for SSL 3.0 (as required by the specification).
13733 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13735 * Change X509_certificate_type() to mark the key as exported/exportable
13736 when it's 512 *bits* long, not 512 bytes.
13740 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13742 * Countermeasure against the Klima-Pokorny-Rosa extension of
13743 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13744 a protocol version number mismatch like a decryption error
13745 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13749 * Turn on RSA blinding by default in the default implementation
13750 to avoid a timing attack. Applications that don't want it can call
13751 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13752 They would be ill-advised to do so in most cases.
13754 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13756 * Change RSA blinding code so that it works when the PRNG is not
13757 seeded (in this case, the secret RSA exponent is abused as
13758 an unpredictable seed -- if it is not unpredictable, there
13759 is no point in blinding anyway). Make RSA blinding thread-safe
13760 by remembering the creator's thread ID in rsa->blinding and
13761 having all other threads use local one-time blinding factors
13762 (this requires more computation than sharing rsa->blinding, but
13763 avoids excessive locking; and if an RSA object is not shared
13764 between threads, blinding will still be very fast).
13768 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13770 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13771 via timing by performing a MAC computation even if incorrect
13772 block cipher padding has been found. This is a countermeasure
13773 against active attacks where the attacker has to distinguish
13774 between bad padding and a MAC verification error. ([CVE-2003-0078])
13776 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13777 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13778 Martin Vuagnoux (EPFL, Ilion)*
13780 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13782 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13783 memory from its contents. This is done with a counter that will
13784 place alternating values in each byte. This can be used to solve
13785 two issues: 1) the removal of calls to memset() by highly optimizing
13786 compilers, and 2) cleansing with other values than 0, since those can
13787 be read through on certain media, for example a swap space on disk.
13791 * Bugfix: client side session caching did not work with external caching,
13792 because the session->cipher setting was not restored when reloading
13793 from the external cache. This problem was masked, when
13794 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13795 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13799 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13800 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13802 *Zeev Lieber <zeev-l@yahoo.com>*
13804 * Undo an undocumented change introduced in 0.9.6e which caused
13805 repeated calls to OpenSSL_add_all_ciphers() and
13806 OpenSSL_add_all_digests() to be ignored, even after calling
13811 * Change the default configuration reader to deal with last line not
13812 being properly terminated.
13816 * Change X509_NAME_cmp() so it applies the special rules on handling
13817 DN values that are of type PrintableString, as well as RDNs of type
13818 emailAddress where the value has the type ia5String.
13820 *stefank@valicert.com via Richard Levitte*
13822 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13823 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13824 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13825 the bitwise-OR of the two for use by the majority of applications
13826 wanting this behaviour, and update the docs. The documented
13827 behaviour and actual behaviour were inconsistent and had been
13828 changing anyway, so this is more a bug-fix than a behavioural
13831 *Geoff Thorpe, diagnosed by Nadav Har'El*
13833 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13834 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13838 * Fix initialization code race conditions in
13839 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13840 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13841 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13842 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13843 ssl2_get_cipher_by_char(),
13844 ssl3_get_cipher_by_char().
13846 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13848 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13849 the cached sessions are flushed, as the remove_cb() might use ex_data
13850 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13851 (see [openssl.org #212]).
13853 *Geoff Thorpe, Lutz Jaenicke*
13855 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13856 length, instead of the encoding length to d2i_ASN1_OBJECT.
13860 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13862 * [In 0.9.6g-engine release:]
13863 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13865 *Lynn Gazis <lgazis@rainbow.com>*
13867 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13869 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13870 and get fix the header length calculation.
13871 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13872 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13874 * Use proper error handling instead of 'assertions' in buffer
13875 overflow checks added in 0.9.6e. This prevents DoS (the
13876 assertions could call abort()).
13878 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13880 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13882 * Add various sanity checks to asn1_get_length() to reject
13883 the ASN1 length bytes if they exceed sizeof(long), will appear
13884 negative or the content length exceeds the length of the
13887 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13889 * Fix cipher selection routines: ciphers without encryption had no flags
13890 for the cipher strength set and where therefore not handled correctly
13891 by the selection routines (PR #130).
13895 * Fix EVP_dsa_sha macro.
13900 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13901 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13902 that was added in OpenSSL 0.9.6d.
13904 As the countermeasure turned out to be incompatible with some
13905 broken SSL implementations, the new option is part of SSL_OP_ALL.
13906 SSL_OP_ALL is usually employed when compatibility with weird SSL
13907 implementations is desired (e.g. '-bugs' option to 's_client' and
13908 's_server'), so the new option is automatically set in many
13913 * Changes in security patch:
13915 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13916 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13917 Air Force Materiel Command, USAF, under agreement number
13920 * Add various sanity checks to asn1_get_length() to reject
13921 the ASN1 length bytes if they exceed sizeof(long), will appear
13922 negative or the content length exceeds the length of the
13923 supplied buffer. ([CVE-2002-0659])
13925 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13927 * Assertions for various potential buffer overflows, not known to
13928 happen in practice.
13930 *Ben Laurie (CHATS)*
13932 * Various temporary buffers to hold ASCII versions of integers were
13933 too small for 64 bit platforms. ([CVE-2002-0655])
13934 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13936 * Remote buffer overflow in SSL3 protocol - an attacker could
13937 supply an oversized session ID to a client. ([CVE-2002-0656])
13939 *Ben Laurie (CHATS)*
13941 * Remote buffer overflow in SSL2 protocol - an attacker could
13942 supply an oversized client master key. ([CVE-2002-0656])
13944 *Ben Laurie (CHATS)*
13946 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13948 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13949 encoded as NULL) with id-dsa-with-sha1.
13951 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13953 * Check various `X509_...()` return values in `apps/req.c`.
13955 *Nils Larsch <nla@trustcenter.de>*
13957 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13958 an end-of-file condition would erroneously be flagged, when the CRLF
13959 was just at the end of a processed block. The bug was discovered when
13960 processing data through a buffering memory BIO handing the data to a
13961 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13962 <ptsekov@syntrex.com> and Nedelcho Stanev.
13966 * Implement a countermeasure against a vulnerability recently found
13967 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13968 before application data chunks to avoid the use of known IVs
13969 with data potentially chosen by the attacker.
13973 * Fix length checks in ssl3_get_client_hello().
13977 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13978 to prevent ssl3_read_internal() from incorrectly assuming that
13979 ssl3_read_bytes() found application data while handshake
13980 processing was enabled when in fact s->s3->in_read_app_data was
13981 merely automatically cleared during the initial handshake.
13983 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13985 * Fix object definitions for Private and Enterprise: they were not
13986 recognized in their shortname (=lowercase) representation. Extend
13987 obj_dat.pl to issue an error when using undefined keywords instead
13988 of silently ignoring the problem (Svenning Sorensen
13989 <sss@sss.dnsalias.net>).
13993 * Fix DH_generate_parameters() so that it works for 'non-standard'
13994 generators, i.e. generators other than 2 and 5. (Previously, the
13995 code did not properly initialise the 'add' and 'rem' values to
13996 BN_generate_prime().)
13998 In the new general case, we do not insist that 'generator' is
13999 actually a primitive root: This requirement is rather pointless;
14000 a generator of the order-q subgroup is just as good, if not
14005 * Map new X509 verification errors to alerts. Discovered and submitted by
14006 Tom Wu <tom@arcot.com>.
14010 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14011 returning non-zero before the data has been completely received
14012 when using non-blocking I/O.
14014 *Bodo Moeller; problem pointed out by John Hughes*
14016 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14018 *Ben Laurie, Lutz Jaenicke*
14020 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14021 Yoram Zahavi <YoramZ@gilian.com>).
14025 * Add information about CygWin 1.3 and on, and preserve proper
14026 configuration for the versions before that.
14028 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14030 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14031 check whether we deal with a copy of a session and do not delete from
14032 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14033 <izhar@checkpoint.com>.
14037 * Do not store session data into the internal session cache, if it
14038 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14039 flag is set). Proposed by Aslam <aslam@funk.com>.
14043 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14048 * [In 0.9.6d-engine release:]
14049 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14051 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14053 * Add the configuration target linux-s390x.
14055 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14057 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14058 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14059 variable as an indication that a ClientHello message has been
14060 received. As the flag value will be lost between multiple
14061 invocations of ssl3_accept when using non-blocking I/O, the
14062 function may not be aware that a handshake has actually taken
14063 place, thus preventing a new session from being added to the
14066 To avoid this problem, we now set s->new_session to 2 instead of
14067 using a local variable.
14069 *Lutz Jaenicke, Bodo Moeller*
14071 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14072 if the SSL_R_LENGTH_MISMATCH error is detected.
14074 *Geoff Thorpe, Bodo Moeller*
14076 * New 'shared_ldflag' column in Configure platform table.
14080 * Fix EVP_CIPHER_mode macro.
14082 *"Dan S. Camper" <dan@bti.net>*
14084 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14085 type, we must throw them away by setting rr->length to 0.
14087 *D P Chang <dpc@qualys.com>*
14089 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14091 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14092 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14093 worked incorrectly for those cases where range = `10..._2` and
14094 `3*range` is two bits longer than range.)
14098 * Only add signing time to PKCS7 structures if it is not already
14103 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14104 OBJ_ld_ce should be OBJ_id_ce.
14105 Also some ip-pda OIDs in crypto/objects/objects.txt were
14106 incorrect (cf. RFC 3039).
14108 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14110 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14111 returns early because it has nothing to do.
14113 *Andy Schneider <andy.schneider@bjss.co.uk>*
14115 * [In 0.9.6c-engine release:]
14116 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14118 *Andy Schneider <andy.schneider@bjss.co.uk>*
14120 * [In 0.9.6c-engine release:]
14121 Add support for Cryptographic Appliance's keyserver technology.
14122 (Use engine 'keyclient')
14124 *Cryptographic Appliances and Geoff Thorpe*
14126 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14127 is called via tools/c89.sh because arguments have to be
14128 rearranged (all '-L' options must appear before the first object
14131 *Richard Shapiro <rshapiro@abinitio.com>*
14133 * [In 0.9.6c-engine release:]
14134 Add support for Broadcom crypto accelerator cards, backported
14137 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14139 * [In 0.9.6c-engine release:]
14140 Add support for SureWare crypto accelerator cards from
14141 Baltimore Technologies. (Use engine 'sureware')
14143 *Baltimore Technologies and Mark Cox*
14145 * [In 0.9.6c-engine release:]
14146 Add support for crypto accelerator cards from Accelerated
14147 Encryption Processing, www.aep.ie. (Use engine 'aep')
14149 *AEP Inc. and Mark Cox*
14151 * Add a configuration entry for gcc on UnixWare.
14153 *Gary Benson <gbenson@redhat.com>*
14155 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14156 messages are stored in a single piece (fixed-length part and
14157 variable-length part combined) and fix various bugs found on the way.
14161 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14162 instead. BIO_gethostbyname() does not know what timeouts are
14163 appropriate, so entries would stay in cache even when they have
14165 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14167 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14168 faced with a pathologically small ClientHello fragment that does
14169 not contain client_version: Instead of aborting with an error,
14170 simply choose the highest available protocol version (i.e.,
14171 TLS 1.0 unless it is disabled). In practice, ClientHello
14172 messages are never sent like this, but this change gives us
14173 strictly correct behaviour at least for TLS.
14177 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14178 never resets s->method to s->ctx->method when called from within
14179 one of the SSL handshake functions.
14181 *Bodo Moeller; problem pointed out by Niko Baric*
14183 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14184 (sent using the client's version number) if client_version is
14185 smaller than the protocol version in use. Also change
14186 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14187 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14188 the client will at least see that alert.
14192 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14197 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14198 client receives HelloRequest while in a handshake.
14200 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14202 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14203 should end in 'break', not 'goto end' which circumvents various
14204 cleanups done in state SSL_ST_OK. But session related stuff
14205 must be disabled for SSL_ST_OK in the case that we just sent a
14208 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14209 before just sending a HelloRequest.
14211 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14213 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14214 reveal whether illegal block cipher padding was found or a MAC
14215 verification error occurred. (Neither SSLerr() codes nor alerts
14216 are directly visible to potential attackers, but the information
14217 may leak via logfiles.)
14219 Similar changes are not required for the SSL 2.0 implementation
14220 because the number of padding bytes is sent in clear for SSL 2.0,
14221 and the extra bytes are just ignored. However ssl/s2_pkt.c
14222 failed to verify that the purported number of padding bytes is in
14227 * Add OpenUNIX-8 support including shared libraries
14228 (Boyd Lynn Gerber <gerberb@zenez.com>).
14232 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14233 'wristwatch attack' using huge encoding parameters (cf.
14234 James H. Manger's CRYPTO 2001 paper). Note that the
14235 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14236 encoding parameters and hence was not vulnerable.
14240 * BN_sqr() bug fix.
14242 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14244 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14245 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14246 followed by modular reduction.
14248 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14250 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14251 equivalent based on BN_pseudo_rand() instead of BN_rand().
14255 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14256 This function was broken, as the check for a new client hello message
14257 to handle SGC did not allow these large messages.
14258 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14262 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14266 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14267 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14271 * Rework the configuration and shared library support for Tru64 Unix.
14272 The configuration part makes use of modern compiler features and
14273 still retains old compiler behavior for those that run older versions
14274 of the OS. The shared library support part includes a variant that
14275 uses the RPATH feature, and is available through the special
14276 configuration target "alpha-cc-rpath", which will never be selected
14279 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14281 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14282 with the same message size as in ssl3_get_certificate_request().
14283 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14284 messages might inadvertently be reject as too long.
14286 *Petr Lampa <lampa@fee.vutbr.cz>*
14288 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14292 * Modified SSL library such that the verify_callback that has been set
14293 specifically for an SSL object with SSL_set_verify() is actually being
14294 used. Before the change, a verify_callback set with this function was
14295 ignored and the verify_callback() set in the SSL_CTX at the time of
14296 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14297 to allow the necessary settings.
14301 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14302 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14303 done automatically (in contradiction to the requirements of the C
14304 standard). This made problems when used from OpenSSH.
14308 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14309 dh->length and always used
14311 BN_rand_range(priv_key, dh->p).
14313 BN_rand_range() is not necessary for Diffie-Hellman, and this
14314 specific range makes Diffie-Hellman unnecessarily inefficient if
14315 dh->length (recommended exponent length) is much smaller than the
14316 length of dh->p. We could use BN_rand_range() if the order of
14317 the subgroup was stored in the DH structure, but we only have
14322 BN_rand(priv_key, l, ...)
14324 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14331 RSA_eay_public_encrypt
14332 RSA_eay_private_decrypt
14333 RSA_eay_private_encrypt (signing)
14334 RSA_eay_public_decrypt (signature verification)
14336 (default implementations for RSA_public_encrypt,
14337 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14338 always reject numbers >= n.
14342 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14343 to synchronize access to 'locking_thread'. This is necessary on
14344 systems where access to 'locking_thread' (an 'unsigned long'
14345 variable) is not atomic.
14349 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14350 *before* setting the 'crypto_lock_rand' flag. The previous code had
14351 a race condition if 0 is a valid thread ID.
14353 *Travis Vitek <vitek@roguewave.com>*
14355 * Add support for shared libraries under Irix.
14357 *Albert Chin-A-Young <china@thewrittenword.com>*
14359 * Add configuration option to build on Linux on both big-endian and
14360 little-endian MIPS.
14362 *Ralf Baechle <ralf@uni-koblenz.de>*
14364 * Add the possibility to create shared libraries on HP-UX.
14368 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14370 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14371 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14372 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14373 PRNG state recovery was possible based on the output of
14374 one PRNG request appropriately sized to gain knowledge on
14375 'md' followed by enough consecutive 1-byte PRNG requests
14376 to traverse all of 'state'.
14378 1. When updating 'md_local' (the current thread's copy of 'md')
14379 during PRNG output generation, hash all of the previous
14380 'md_local' value, not just the half used for PRNG output.
14382 2. Make the number of bytes from 'state' included into the hash
14383 independent from the number of PRNG bytes requested.
14385 The first measure alone would be sufficient to avoid
14386 Markku-Juhani's attack. (Actually it had never occurred
14387 to me that the half of 'md_local' used for chaining was the
14388 half from which PRNG output bytes were taken -- I had always
14389 assumed that the secret half would be used.) The second
14390 measure makes sure that additional data from 'state' is never
14391 mixed into 'md_local' in small portions; this heuristically
14392 further strengthens the PRNG.
14396 * Fix crypto/bn/asm/mips3.s.
14400 * When only the key is given to "enc", the IV is undefined. Print out
14401 an error message in this case.
14405 * Handle special case when X509_NAME is empty in X509 printing routines.
14409 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14410 positive and less than q.
14414 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14415 used: it isn't thread safe and the add_lock_callback should handle
14418 *Paul Rose <Paul.Rose@bridge.com>*
14420 * Verify that incoming data obeys the block size in
14421 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14427 *Ulf Möller, Bodo Möller*
14429 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14430 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14431 when fixing the server behaviour for backwards-compatible 'client
14432 hello' messages. (Note that the attack is impractical against
14433 SSL 3.0 and TLS 1.0 anyway because length and version checking
14434 means that the probability of guessing a valid ciphertext is
14435 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14438 Before 0.9.5, the countermeasure (hide the error by generating a
14439 random 'decryption result') did not work properly because
14440 ERR_clear_error() was missing, meaning that SSL_get_error() would
14441 detect the supposedly ignored error.
14443 Both problems are now fixed.
14447 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14448 (previously it was 1024).
14452 * Fix for compatibility mode trust settings: ignore trust settings
14453 unless some valid trust or reject settings are present.
14457 * Fix for blowfish EVP: its a variable length cipher.
14461 * Fix various bugs related to DSA S/MIME verification. Handle missing
14462 parameters in DSA public key structures and return an error in the
14463 DSA routines if parameters are absent.
14467 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14468 in the current directory if neither $RANDFILE nor $HOME was set.
14469 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14470 caused some confusion to Windows users who haven't defined $HOME.
14471 Thus RAND_file_name() is changed again: e_os.h can define a
14472 DEFAULT_HOME, which will be used if $HOME is not set.
14473 For Windows, we use "C:"; on other platforms, we still require
14474 environment variables.
14476 * Move 'if (!initialized) RAND_poll()' into regions protected by
14477 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14478 having multiple threads call RAND_poll() concurrently.
14482 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14483 combination of a flag and a thread ID variable.
14484 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14485 flag), *other* threads can enter ssleay_add_bytes without obeying
14486 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14487 that they do not hold after the first thread unsets add_do_not_lock).
14491 * Change bctest again: '-x' expressions are not available in all
14492 versions of 'test'.
14496 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14498 * Fix a couple of memory leaks in PKCS7_dataDecode()
14500 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14502 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14503 the default extension for executables, if any. Also, make the perl
14504 scripts that use symlink() to test if it really exists and use "cp"
14505 if it doesn't. All this made OpenSSL compilable and installable in
14510 * Fix for asn1_GetSequence() for indefinite length constructed data.
14511 If SEQUENCE is length is indefinite just set c->slen to the total
14512 amount of data available.
14514 *Steve Henson, reported by shige@FreeBSD.org*
14516 *This change does not apply to 0.9.7.*
14518 * Change bctest to avoid here-documents inside command substitution
14519 (workaround for FreeBSD /bin/sh bug).
14520 For compatibility with Ultrix, avoid shell functions (introduced
14521 in the bctest version that searches along $PATH).
14525 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14526 with des_encrypt() defined on some operating systems, like Solaris
14531 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14532 On the Importance of Eliminating Errors in Cryptographic
14533 Computations, J. Cryptology 14 (2001) 2, 101-119,
14534 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14538 * MIPS assembler BIGNUM division bug fix.
14542 * Disabled incorrect Alpha assembler code.
14546 * Fix PKCS#7 decode routines so they correctly update the length
14547 after reading an EOC for the EXPLICIT tag.
14551 *This change does not apply to 0.9.7.*
14553 * Fix bug in PKCS#12 key generation routines. This was triggered
14554 if a 3DES key was generated with a 0 initial byte. Include
14555 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14556 (but broken) behaviour.
14560 * Enhance bctest to search for a working bc along $PATH and print
14563 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14565 * Fix memory leaks in err.c: free err_data string if necessary;
14566 don't write to the wrong index in ERR_set_error_data.
14570 * Implement ssl23_peek (analogous to ssl23_read), which previously
14575 * Replace rdtsc with `_emit` statements for VC++ version 5.
14577 *Jeremy Cooper <jeremy@baymoo.org>*
14579 * Make it possible to reuse SSLv2 sessions.
14583 * In copy_email() check for >= 0 as a return value for
14584 X509_NAME_get_index_by_NID() since 0 is a valid index.
14586 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14588 * Avoid coredump with unsupported or invalid public keys by checking if
14589 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14590 PKCS7_verify() fails with non detached data.
14594 * Don't use getenv in library functions when run as setuid/setgid.
14595 New function OPENSSL_issetugid().
14599 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14600 due to incorrect handling of multi-threading:
14602 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14604 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14606 3. Count how many times MemCheck_off() has been called so that
14607 nested use can be treated correctly. This also avoids
14608 inband-signalling in the previous code (which relied on the
14609 assumption that thread ID 0 is impossible).
14613 * Add "-rand" option also to s_client and s_server.
14617 * Fix CPU detection on Irix 6.x.
14618 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14619 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14621 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14626 *This change does not apply to 0.9.7.*
14628 * Use the cached encoding of an X509_NAME structure rather than
14629 copying it. This is apparently the reason for the libsafe "errors"
14630 but the code is actually correct.
14634 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14635 Bleichenbacher's DSA attack.
14636 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14637 to be set and top=0 forces the highest bit to be set; top=-1 is new
14638 and leaves the highest bit random.
14640 *Ulf Moeller, Bodo Moeller*
14642 * In the `NCONF_...`-based implementations for `CONF_...` queries
14643 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14644 a temporary CONF structure with the data component set to NULL
14645 (which gives segmentation faults in lh_retrieve).
14646 Instead, use NULL for the CONF pointer in CONF_get_string and
14647 CONF_get_number (which may use environment variables) and directly
14648 return NULL from CONF_get_section.
14652 * Fix potential buffer overrun for EBCDIC.
14656 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14657 keyUsage if basicConstraints absent for a CA.
14661 * Make SMIME_write_PKCS7() write mail header values with a format that
14662 is more generally accepted (no spaces before the semicolon), since
14663 some programs can't parse those values properly otherwise. Also make
14664 sure BIO's that break lines after each write do not create invalid
14669 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14670 macros previously used would not encode an empty SEQUENCE OF
14671 and break the signature.
14675 *This change does not apply to 0.9.7.*
14677 * Zero the premaster secret after deriving the master secret in
14682 * Add some EVP_add_digest_alias registrations (as found in
14683 OpenSSL_add_all_digests()) to SSL_library_init()
14684 aka OpenSSL_add_ssl_algorithms(). This provides improved
14685 compatibility with peers using X.509 certificates
14686 with unconventional AlgorithmIdentifier OIDs.
14690 * Fix for Irix with NO_ASM.
14692 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14694 * ./config script fixes.
14696 *Ulf Moeller, Richard Levitte*
14698 * Fix 'openssl passwd -1'.
14702 * Change PKCS12_key_gen_asc() so it can cope with non null
14703 terminated strings whose length is passed in the passlen
14704 parameter, for example from PEM callbacks. This was done
14705 by adding an extra length parameter to asc2uni().
14707 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14709 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14710 call failed, free the DSA structure.
14714 * Fix to uni2asc() to cope with zero length Unicode strings.
14715 These are present in some PKCS#12 files.
14719 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14720 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14721 when writing a 32767 byte record.
14723 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14725 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14726 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14728 (RSA objects have a reference count access to which is protected
14729 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14730 so they are meant to be shared between threads.)
14731 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14732 "Reddie, Steven" <Steven.Reddie@ca.com>*
14734 * Fix a deadlock in CRYPTO_mem_leaks().
14738 * Use better test patterns in bntest.
14742 * rand_win.c fix for Borland C.
14746 * BN_rshift bugfix for n == 0.
14750 * Add a 'bctest' script that checks for some known 'bc' bugs
14751 so that 'make test' does not abort just because 'bc' is broken.
14755 * Store verify_result within SSL_SESSION also for client side to
14756 avoid potential security hole. (Re-used sessions on the client side
14757 always resulted in verify_result==X509_V_OK, not using the original
14758 result of the server certificate verification.)
14762 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14763 SSL3_RT_APPLICATION_DATA, return 0.
14764 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14769 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14770 releases, have been re-implemented by renaming the previous
14771 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14772 and ssl3_read_internal, respectively, and adding 'peek' parameters
14773 to them. The new ssl[23]_{read,peek} functions are calls to
14774 ssl[23]_read_internal with the 'peek' flag set appropriately.
14775 A 'peek' parameter has also been added to ssl3_read_bytes, which
14776 does the actual work for ssl3_read_internal.
14780 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14781 the method-specific "init()" handler. Also clean up ex_data after
14782 calling the method-specific "finish()" handler. Previously, this was
14783 happening the other way round.
14787 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14788 The previous value, 12, was not always sufficient for BN_mod_exp().
14792 * Make sure that shared libraries get the internal name engine with
14793 the full version number and not just 0. This should mark the
14794 shared libraries as not backward compatible. Of course, this should
14795 be changed again when we can guarantee backward binary compatibility.
14799 * Fix typo in get_cert_by_subject() in by_dir.c
14801 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14803 * Rework the system to generate shared libraries:
14805 - Make note of the expected extension for the shared libraries and
14806 if there is a need for symbolic links from for example libcrypto.so.0
14807 to libcrypto.so.0.9.7. There is extended info in Configure for
14810 - Make as few rebuilds of the shared libraries as possible.
14812 - Still avoid linking the OpenSSL programs with the shared libraries.
14814 - When installing, install the shared libraries separately from the
14819 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14821 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14822 and not in SSL_clear because the latter is also used by the
14823 accept/connect functions; previously, the settings made by
14824 SSL_set_read_ahead would be lost during the handshake.
14826 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14828 * Correct util/mkdef.pl to be selective about disabled algorithms.
14829 Previously, it would create entries for disabled algorithms no
14834 * Added several new manual pages for SSL_* function.
14838 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14840 * In ssl23_get_client_hello, generate an error message when faced
14841 with an initial SSL 3.0/TLS record that is too small to contain the
14842 first two bytes of the ClientHello message, i.e. client_version.
14843 (Note that this is a pathologic case that probably has never happened
14844 in real life.) The previous approach was to use the version number
14845 from the record header as a substitute; but our protocol choice
14846 should not depend on that one because it is not authenticated
14847 by the Finished messages.
14851 * More robust randomness gathering functions for Windows.
14853 *Jeffrey Altman <jaltman@columbia.edu>*
14855 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14856 not set then we don't setup the error code for issuer check errors
14857 to avoid possibly overwriting other errors which the callback does
14858 handle. If an application does set the flag then we assume it knows
14859 what it is doing and can handle the new informational codes
14864 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14865 a general "ANY" type, as such it should be able to decode anything
14866 including tagged types. However it didn't check the class so it would
14867 wrongly interpret tagged types in the same way as their universal
14868 counterpart and unknown types were just rejected. Changed so that the
14869 tagged and unknown types are handled in the same way as a SEQUENCE:
14870 that is the encoding is stored intact. There is also a new type
14871 "V_ASN1_OTHER" which is used when the class is not universal, in this
14872 case we have no idea what the actual type is so we just lump them all
14877 * On VMS, stdout may very well lead to a file that is written to
14878 in a record-oriented fashion. That means that every write() will
14879 write a separate record, which will be read separately by the
14880 programs trying to read from it. This can be very confusing.
14882 The solution is to put a BIO filter in the way that will buffer
14883 text until a linefeed is reached, and then write everything a
14884 line at a time, so every record written will be an actual line,
14885 not chunks of lines and not (usually doesn't happen, but I've
14886 seen it once) several lines in one record. BIO_f_linebuffer() is
14889 Currently, it's a VMS-only method, because that's where it has
14890 been tested well enough.
14894 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14895 it can return incorrect results.
14896 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14897 but it was in 0.9.6-beta[12].)
14901 * Disable the check for content being present when verifying detached
14902 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14903 include zero length content when signing messages.
14907 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14908 BIO_ctrl (for BIO pairs).
14912 * Add DSO method for VMS.
14916 * Bug fix: Montgomery multiplication could produce results with the
14921 * Add RPM specification openssl.spec and modify it to build three
14922 packages. The default package contains applications, application
14923 documentation and run-time libraries. The devel package contains
14924 include files, static libraries and function documentation. The
14925 doc package contains the contents of the doc directory. The original
14926 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14930 * Add a large number of documentation files for many SSL routines.
14932 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14934 * Add a configuration entry for Sony News 4.
14936 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14938 * Don't set the two most significant bits to one when generating a
14939 random number < q in the DSA library.
14943 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14944 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14945 the underlying transport is blocking) if a handshake took place.
14946 (The default behaviour is needed by applications such as s_client
14947 and s_server that use select() to determine when to use SSL_read;
14948 but for applications that know in advance when to expect data, it
14949 just makes things more complicated.)
14953 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14958 * Add a few more EBCDIC conditionals that make `req` and `x509`
14959 work better on such systems.
14961 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14963 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14964 Update PKCS12_parse() so it copies the friendlyName and the
14965 keyid to the certificates aux info.
14969 * Fix bug in PKCS7_verify() which caused an infinite loop
14970 if there was more than one signature.
14972 *Sven Uszpelkat <su@celocom.de>*
14974 * Major change in util/mkdef.pl to include extra information
14975 about each symbol, as well as presenting variables as well
14976 as functions. This change means that there's n more need
14977 to rebuild the .num files when some algorithms are excluded.
14981 * Allow the verify time to be set by an application,
14982 rather than always using the current time.
14986 * Phase 2 verify code reorganisation. The certificate
14987 verify code now looks up an issuer certificate by a
14988 number of criteria: subject name, authority key id
14989 and key usage. It also verifies self signed certificates
14990 by the same criteria. The main comparison function is
14991 X509_check_issued() which performs these checks.
14993 Lot of changes were necessary in order to support this
14994 without completely rewriting the lookup code.
14996 Authority and subject key identifier are now cached.
14998 The LHASH 'certs' is X509_STORE has now been replaced
14999 by a STACK_OF(X509_OBJECT). This is mainly because an
15000 LHASH can't store or retrieve multiple objects with
15001 the same hash value.
15003 As a result various functions (which were all internal
15004 use only) have changed to handle the new X509_STORE
15005 structure. This will break anything that messed round
15006 with X509_STORE internally.
15008 The functions X509_STORE_add_cert() now checks for an
15009 exact match, rather than just subject name.
15011 The X509_STORE API doesn't directly support the retrieval
15012 of multiple certificates matching a given criteria, however
15013 this can be worked round by performing a lookup first
15014 (which will fill the cache with candidate certificates)
15015 and then examining the cache for matches. This is probably
15016 the best we can do without throwing out X509_LOOKUP
15017 entirely (maybe later...).
15019 The X509_VERIFY_CTX structure has been enhanced considerably.
15021 All certificate lookup operations now go via a get_issuer()
15022 callback. Although this currently uses an X509_STORE it
15023 can be replaced by custom lookups. This is a simple way
15024 to bypass the X509_STORE hackery necessary to make this
15025 work and makes it possible to use more efficient techniques
15026 in future. A very simple version which uses a simple
15027 STACK for its trusted certificate store is also provided
15028 using X509_STORE_CTX_trusted_stack().
15030 The verify_cb() and verify() callbacks now have equivalents
15031 in the X509_STORE_CTX structure.
15033 X509_STORE_CTX also has a 'flags' field which can be used
15034 to customise the verify behaviour.
15038 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15039 excludes S/MIME capabilities.
15043 * When a certificate request is read in keep a copy of the
15044 original encoding of the signed data and use it when outputting
15045 again. Signatures then use the original encoding rather than
15046 a decoded, encoded version which may cause problems if the
15047 request is improperly encoded.
15051 * For consistency with other BIO_puts implementations, call
15052 buffer_write(b, ...) directly in buffer_puts instead of calling
15055 In BIO_puts, increment b->num_write as in BIO_write.
15057 *Peter.Sylvester@EdelWeb.fr*
15059 * Fix BN_mul_word for the case where the word is 0. (We have to use
15060 BN_zero, we may not return a BIGNUM with an array consisting of
15061 words set to zero.)
15065 * Avoid calling abort() from within the library when problems are
15066 detected, except if preprocessor symbols have been defined
15067 (such as REF_CHECK, BN_DEBUG etc.).
15071 * New openssl application 'rsautl'. This utility can be
15072 used for low level RSA operations. DER public key
15073 BIO/fp routines also added.
15077 * New Configure entry and patches for compiling on QNX 4.
15079 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15081 * A demo state-machine implementation was sponsored by
15082 Nuron (<http://www.nuron.com/>) and is now available in
15083 demos/state_machine.
15087 * New options added to the 'dgst' utility for signature
15088 generation and verification.
15092 * Unrecognized PKCS#7 content types are now handled via a
15093 catch all ASN1_TYPE structure. This allows unsupported
15094 types to be stored as a "blob" and an application can
15095 encode and decode it manually.
15099 * Fix various signed/unsigned issues to make a_strex.c
15100 compile under VC++.
15102 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15104 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15105 length if passed a buffer. ASN1_INTEGER_to_BN failed
15106 if passed a NULL BN and its argument was negative.
15108 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15110 * Modification to PKCS#7 encoding routines to output definite
15111 length encoding. Since currently the whole structures are in
15112 memory there's not real point in using indefinite length
15113 constructed encoding. However if OpenSSL is compiled with
15114 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15118 * Added BIO_vprintf() and BIO_vsnprintf().
15122 * Added more prefixes to parse for in the strings written
15123 through a logging bio, to cover all the levels that are available
15124 through syslog. The prefixes are now:
15126 PANIC, EMERG, EMR => LOG_EMERG
15127 ALERT, ALR => LOG_ALERT
15128 CRIT, CRI => LOG_CRIT
15129 ERROR, ERR => LOG_ERR
15130 WARNING, WARN, WAR => LOG_WARNING
15131 NOTICE, NOTE, NOT => LOG_NOTICE
15132 INFO, INF => LOG_INFO
15133 DEBUG, DBG => LOG_DEBUG
15135 and as before, if none of those prefixes are present at the
15136 beginning of the string, LOG_ERR is chosen.
15138 On Win32, the `LOG_*` levels are mapped according to this:
15140 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15141 LOG_WARNING => EVENTLOG_WARNING_TYPE
15142 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15146 * Made it possible to reconfigure with just the configuration
15147 argument "reconf" or "reconfigure". The command line arguments
15148 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15149 and are retrieved from there when reconfiguring.
15155 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15157 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15161 * The obj_dat.pl script was messing up the sorting of object
15162 names. The reason was that it compared the quoted version
15163 of strings as a result "OCSP" > "OCSP Signing" because
15164 " > SPACE. Changed script to store unquoted versions of
15165 names and add quotes on output. It was also omitting some
15166 names from the lookup table if they were given a default
15167 value (that is if SN is missing it is given the same
15168 value as LN and vice versa), these are now added on the
15169 grounds that if an object has a name we should be able to
15170 look it up. Finally added warning output when duplicate
15171 short or long names are found.
15175 * Changes needed for Tandem NSK.
15177 *Scott Uroff <scott@xypro.com>*
15179 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15180 RSA_padding_check_SSLv23(), special padding was never detected
15181 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15182 version rollback attacks was not effective.
15184 In s23_clnt.c, don't use special rollback-attack detection padding
15185 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15186 client; similarly, in s23_srvr.c, don't do the rollback check if
15187 SSL 2.0 is the only protocol enabled in the server.
15191 * Make it possible to get hexdumps of unprintable data with 'openssl
15192 asn1parse'. By implication, the functions ASN1_parse_dump() and
15193 BIO_dump_indent() are added.
15197 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15198 these print out strings and name structures based on various
15199 flags including RFC2253 support and proper handling of
15200 multibyte characters. Added options to the 'x509' utility
15201 to allow the various flags to be set.
15205 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15206 Also change the functions X509_cmp_current_time() and
15207 X509_gmtime_adj() work with an ASN1_TIME structure,
15208 this will enable certificates using GeneralizedTime in validity
15209 dates to be checked.
15213 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15214 negative public key encodings) on by default,
15215 NO_NEG_PUBKEY_BUG can be set to disable it.
15219 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15220 content octets. An i2c_ASN1_OBJECT is unnecessary because
15221 the encoding can be trivially obtained from the structure.
15225 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15226 not read locks (`CRYPTO_r_[un]lock`).
15230 * A first attempt at creating official support for shared
15231 libraries through configuration. I've kept it so the
15232 default is static libraries only, and the OpenSSL programs
15233 are always statically linked for now, but there are
15234 preparations for dynamic linking in place.
15235 This has been tested on Linux and Tru64.
15239 * Randomness polling function for Win9x, as described in:
15240 Peter Gutmann, Software Generation of Practically Strong
15245 * Fix so PRNG is seeded in req if using an already existing
15250 * New options to smime application. -inform and -outform
15251 allow alternative formats for the S/MIME message including
15252 PEM and DER. The -content option allows the content to be
15253 specified separately. This should allow things like Netscape
15254 form signing output easier to verify.
15258 * Fix the ASN1 encoding of tags using the 'long form'.
15262 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15263 STRING types. These convert content octets to and from the
15264 underlying type. The actual tag and length octets are
15265 already assumed to have been read in and checked. These
15266 are needed because all other string types have virtually
15267 identical handling apart from the tag. By having versions
15268 of the ASN1 functions that just operate on content octets
15269 IMPLICIT tagging can be handled properly. It also allows
15270 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15271 and ASN1_INTEGER are identical apart from the tag.
15275 * Change the handling of OID objects as follows:
15277 - New object identifiers are inserted in objects.txt, following
15278 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15279 - objects.pl is used to process obj_mac.num and create a new
15281 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15284 This is currently kind of a hack, and the perl code in objects.pl
15285 isn't very elegant, but it works as I intended. The simplest way
15286 to check that it worked correctly is to look in obj_dat.h and
15287 check the array nid_objs and make sure the objects haven't moved
15288 around (this is important!). Additions are OK, as well as
15289 consistent name changes.
15293 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15297 * Addition of the command line parameter '-rand file' to 'openssl req'.
15298 The given file adds to whatever has already been seeded into the
15299 random pool through the RANDFILE configuration file option or
15300 environment variable, or the default random state file.
15304 * mkstack.pl now sorts each macro group into lexical order.
15305 Previously the output order depended on the order the files
15306 appeared in the directory, resulting in needless rewriting
15311 * Patches to make OpenSSL compile under Win32 again. Mostly
15312 work arounds for the VC++ problem that it treats func() as
15313 func(void). Also stripped out the parts of mkdef.pl that
15314 added extra typesafe functions: these no longer exist.
15318 * Reorganisation of the stack code. The macros are now all
15319 collected in safestack.h . Each macro is defined in terms of
15320 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15321 DEBUG_SAFESTACK is now handled in terms of function casts,
15322 this has the advantage of retaining type safety without the
15323 use of additional functions. If DEBUG_SAFESTACK is not defined
15324 then the non typesafe macros are used instead. Also modified the
15325 mkstack.pl script to handle the new form. Needs testing to see
15326 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15327 the default if no major problems. Similar behaviour for ASN1_SET_OF
15328 and PKCS12_STACK_OF.
15332 * When some versions of IIS use the 'NET' form of private key the
15333 key derivation algorithm is different. Normally MD5(password) is
15334 used as a 128 bit RC4 key. In the modified case
15335 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15336 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15337 as the old Netscape_RSA functions except they have an additional
15338 'sgckey' parameter which uses the modified algorithm. Also added
15339 an -sgckey command line option to the rsa utility. Thanks to
15340 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15341 algorithm to openssl-dev.
15345 * The evp_local.h macros were using 'c.##kname' which resulted in
15346 invalid expansion on some systems (SCO 5.0.5 for example).
15347 Corrected to 'c.kname'.
15349 *Phillip Porch <root@theporch.com>*
15351 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15352 a STACK of email addresses from a certificate or request, these look
15353 in the subject name and the subject alternative name extensions and
15354 omit any duplicate addresses.
15358 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15359 This makes DSA verification about 2 % faster.
15363 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15364 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15365 plus overhead for 1024 bit moduli).
15366 This makes exponentiations about 0.5 % faster for 1024 bit
15367 exponents (as measured by "openssl speed rsa2048").
15371 * Rename memory handling macros to avoid conflicts with other
15373 Malloc => OPENSSL_malloc
15374 Malloc_locked => OPENSSL_malloc_locked
15375 Realloc => OPENSSL_realloc
15376 Free => OPENSSL_free
15380 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15381 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15385 * CygWin32 support.
15387 *John Jarvie <jjarvie@newsguy.com>*
15389 * The type-safe stack code has been rejigged. It is now only compiled
15390 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15391 by default all type-specific stack functions are "#define"d back to
15392 standard stack functions. This results in more streamlined output
15393 but retains the type-safety checking possibilities of the original
15398 * The STACK code has been cleaned up, and certain type declarations
15399 that didn't make a lot of sense have been brought in line. This has
15400 also involved a cleanup of sorts in safestack.h to more correctly
15401 map type-safe stack functions onto their plain stack counterparts.
15402 This work has also resulted in a variety of "const"ifications of
15403 lots of the code, especially `_cmp` operations which should normally
15404 be prototyped with "const" parameters anyway.
15408 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15409 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15410 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15411 where all of 'md' is used each time the PRNG is used, but 'state'
15412 is used only indexed by a cyclic counter. As entropy may not be
15413 well distributed from the beginning, 'md' is important as a
15414 chaining variable. However, the output function chains only half
15415 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15416 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15417 in all of 'state' being rewritten, with the new values depending
15418 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15422 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15423 the handshake is continued after ssl_verify_cert_chain();
15424 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15425 can lead to 'unexplainable' connection aborts later.
15427 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15429 * Major EVP API cipher revision.
15430 Add hooks for extra EVP features. This allows various cipher
15431 parameters to be set in the EVP interface. Support added for variable
15432 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15433 setting of RC2 and RC5 parameters.
15435 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15438 Remove lots of duplicated code from the EVP library. For example *every*
15439 cipher init() function handles the 'iv' in the same way according to the
15440 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15441 for CFB and OFB modes they zero ctx->num.
15443 New functionality allows removal of S/MIME code RC2 hack.
15445 Most of the routines have the same form and so can be declared in terms
15448 By shifting this to the top level EVP_CipherInit() it can be removed from
15449 all individual ciphers. If the cipher wants to handle IVs or keys
15450 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15453 Change lots of functions like EVP_EncryptUpdate() to now return a
15454 value: although software versions of the algorithms cannot fail
15455 any installed hardware versions can.
15459 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15460 this option is set, tolerate broken clients that send the negotiated
15461 protocol version number instead of the requested protocol version
15466 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15467 i.e. non-zero for export ciphersuites, zero otherwise.
15468 Previous versions had this flag inverted, inconsistent with
15469 rsa_tmp_cb (..._TMP_RSA_CB).
15471 *Bodo Moeller; problem reported by Amit Chopra*
15473 * Add missing DSA library text string. Work around for some IIS
15474 key files with invalid SEQUENCE encoding.
15478 * Add a document (doc/standards.txt) that list all kinds of standards
15479 and so on that are implemented in OpenSSL.
15483 * Enhance c_rehash script. Old version would mishandle certificates
15484 with the same subject name hash and wouldn't handle CRLs at all.
15485 Added -fingerprint option to crl utility, to support new c_rehash
15490 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15494 * Fix for SSL server purpose checking. Server checking was
15495 rejecting certificates which had extended key usage present
15496 but no ssl client purpose.
15498 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15500 * Make PKCS#12 code work with no password. The PKCS#12 spec
15501 is a little unclear about how a blank password is handled.
15502 Since the password in encoded as a BMPString with terminating
15503 double NULL a zero length password would end up as just the
15504 double NULL. However no password at all is different and is
15505 handled differently in the PKCS#12 key generation code. NS
15506 treats a blank password as zero length. MSIE treats it as no
15507 password on export: but it will try both on import. We now do
15508 the same: PKCS12_parse() tries zero length and no password if
15509 the password is set to "" or NULL (NULL is now a valid password:
15510 it wasn't before) as does the pkcs12 application.
15514 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15515 perror when PEM_read_bio_X509_REQ fails, the error message must
15516 be obtained from the error queue.
15520 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15521 it in ERR_remove_state if appropriate, and change ERR_get_state
15522 accordingly to avoid race conditions (this is necessary because
15523 thread_hash is no longer constant once set).
15527 * Bugfix for linux-elf makefile.one.
15531 * RSA_get_default_method() will now cause a default
15532 RSA_METHOD to be chosen if one doesn't exist already.
15533 Previously this was only set during a call to RSA_new()
15534 or RSA_new_method(NULL) meaning it was possible for
15535 RSA_get_default_method() to return NULL.
15539 * Added native name translation to the existing DSO code
15540 that will convert (if the flag to do so is set) filenames
15541 that are sufficiently small and have no path information
15542 into a canonical native form. Eg. "blah" converted to
15543 "libblah.so" or "blah.dll" etc.
15547 * New function ERR_error_string_n(e, buf, len) which is like
15548 ERR_error_string(e, buf), but writes at most 'len' bytes
15549 including the 0 terminator. For ERR_error_string_n, 'buf'
15552 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15554 * CONF library reworked to become more general. A new CONF
15555 configuration file reader "class" is implemented as well as a
15556 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15557 old `CONF_*` functions are still there, but are reimplemented to
15558 work in terms of the new functions. Also, a set of functions
15559 to handle the internal storage of the configuration data is
15560 provided to make it easier to write new configuration file
15561 reader "classes" (I can definitely see something reading a
15562 configuration file in XML format, for example), called `_CONF_*`,
15563 or "the configuration storage API"...
15565 The new configuration file reading functions are:
15567 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15568 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15570 NCONF_default, NCONF_WIN32
15572 NCONF_dump_fp, NCONF_dump_bio
15574 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15575 NCONF_new creates a new CONF object. This works in the same way
15576 as other interfaces in OpenSSL, like the BIO interface.
15577 `NCONF_dump_*` dump the internal storage of the configuration file,
15578 which is useful for debugging. All other functions take the same
15579 arguments as the old `CONF_*` functions with the exception of the
15580 first that must be a `CONF *` instead of a `LHASH *`.
15582 To make it easier to use the new classes with the old `CONF_*` functions,
15583 the function CONF_set_default_method is provided.
15587 * Add '-tls1' option to 'openssl ciphers', which was already
15588 mentioned in the documentation but had not been implemented.
15589 (This option is not yet really useful because even the additional
15590 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15594 * Initial DSO code added into libcrypto for letting OpenSSL (and
15595 OpenSSL-based applications) load shared libraries and bind to
15596 them in a portable way.
15598 *Geoff Thorpe, with contributions from Richard Levitte*
15600 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15602 * Make sure _lrotl and _lrotr are only used with MSVC.
15604 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15605 (the default implementation of RAND_status).
15607 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15608 to '-clrext' (= clear extensions), as intended and documented.
15609 *Bodo Moeller; inconsistency pointed out by Michael Attili
15610 <attili@amaxo.com>*
15612 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15613 was larger than the MD block size.
15615 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15617 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15618 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15619 using the passed key: if the passed key was a private key the result
15620 of X509_print(), for example, would be to print out all the private key
15625 * des_quad_cksum() byte order bug fix.
15626 *Ulf Möller, using the problem description in krb4-0.9.7, where
15627 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15629 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15632 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15634 * For easily testing in shell scripts whether some command
15635 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15636 returns with exit code 0 iff no command of the given name is available.
15637 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15638 the output goes to stdout and nothing is printed to stderr.
15639 Additional arguments are always ignored.
15641 Since for each cipher there is a command of the same name,
15642 the 'no-cipher' compilation switches can be tested this way.
15644 ('openssl no-XXX' is not able to detect pseudo-commands such
15645 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15649 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15653 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15654 is set; it will be thrown away anyway because each handshake creates
15656 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15657 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15658 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15659 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15663 * New s_client option -ign_eof: EOF at stdin is ignored, and
15664 'Q' and 'R' lose their special meanings (quit/renegotiate).
15665 This is part of what -quiet does; unlike -quiet, -ign_eof
15666 does not suppress any output.
15670 * Add compatibility options to the purpose and trust code. The
15671 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15672 accepts a certificate or CA, this was the previous behaviour,
15673 with all the associated security issues.
15675 X509_TRUST_COMPAT is the old trust behaviour: only and
15676 automatically trust self signed roots in certificate store. A
15677 new trust setting X509_TRUST_DEFAULT is used to specify that
15678 a purpose has no associated trust setting and it should instead
15679 use the value in the default purpose.
15683 * Fix the PKCS#8 DSA private key code so it decodes keys again
15684 and fix a memory leak.
15688 * In util/mkerr.pl (which implements 'make errors'), preserve
15689 reason strings from the previous version of the .c file, as
15690 the default to have only downcase letters (and digits) in
15691 automatically generated reasons codes is not always appropriate.
15695 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15696 using strerror. Previously, ERR_reason_error_string() returned
15697 library names as reason strings for SYSerr; but SYSerr is a special
15698 case where small numbers are errno values, not library numbers.
15702 * Add '-dsaparam' option to 'openssl dhparam' application. This
15703 converts DSA parameters into DH parameters. (When creating parameters,
15704 DSA_generate_parameters is used.)
15708 * Include 'length' (recommended exponent length) in C code generated
15709 by 'openssl dhparam -C'.
15713 * The second argument to set_label in perlasm was already being used
15714 so couldn't be used as a "file scope" flag. Moved to third argument
15719 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15720 instead of RAND_bytes for encryption IVs and salts.
15724 * Include RAND_status() into RAND_METHOD instead of implementing
15725 it only for md_rand.c Otherwise replacing the PRNG by calling
15726 RAND_set_rand_method would be impossible.
15730 * Don't let DSA_generate_key() enter an infinite loop if the random
15731 number generation fails.
15735 * New 'rand' application for creating pseudo-random output.
15739 * Added configuration support for Linux/IA64
15741 *Rolf Haberrecker <rolf@suse.de>*
15743 * Assembler module support for Mingw32.
15747 * Shared library support for HPUX (in shlib/).
15749 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15751 * Shared library support for Solaris gcc.
15753 *Lutz Behnke <behnke@trustcenter.de>*
15755 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15757 * PKCS7_encrypt() was adding text MIME headers twice because they
15758 were added manually and by SMIME_crlf_copy().
15762 * In bntest.c don't call BN_rand with zero bits argument.
15764 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15766 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15767 case was implemented. This caused BN_div_recp() to fail occasionally.
15771 * Add an optional second argument to the set_label() in the perl
15772 assembly language builder. If this argument exists and is set
15773 to 1 it signals that the assembler should use a symbol whose
15774 scope is the entire file, not just the current function. This
15775 is needed with MASM which uses the format label:: for this scope.
15777 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15779 * Change the ASN1 types so they are typedefs by default. Before
15780 almost all types were #define'd to ASN1_STRING which was causing
15781 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15786 * Change names of new functions to the new get1/get0 naming
15787 convention: After 'get1', the caller owns a reference count
15788 and has to call `..._free`; 'get0' returns a pointer to some
15789 data structure without incrementing reference counters.
15790 (Some of the existing 'get' functions increment a reference
15791 counter, some don't.)
15792 Similarly, 'set1' and 'add1' functions increase reference
15793 counters or duplicate objects.
15797 * Allow for the possibility of temp RSA key generation failure:
15798 the code used to assume it always worked and crashed on failure.
15802 * Fix potential buffer overrun problem in BIO_printf().
15803 *Ulf Möller, using public domain code by Patrick Powell; problem
15804 pointed out by David Sacerdote <das33@cornell.edu>*
15806 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15807 RAND_egd() and RAND_status(). In the command line application,
15808 the EGD socket can be specified like a seed file using RANDFILE
15813 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15814 Some CAs (e.g. Verisign) distribute certificates in this form.
15818 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15819 list to exclude them. This means that no special compilation option
15820 is needed to use anonymous DH: it just needs to be included in the
15825 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15826 EVP_MD_type. The old functionality is available in a new macro called
15827 EVP_MD_md(). Change code that uses it and update docs.
15831 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15832 where the `void *` argument is replaced by a function pointer argument.
15833 Previously `void *` was abused to point to functions, which works on
15834 many platforms, but is not correct. As these functions are usually
15835 called by macros defined in OpenSSL header files, most source code
15836 should work without changes.
15840 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15841 sections with information on -D... compiler switches used for
15842 compiling the library so that applications can see them. To enable
15843 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15844 must be defined. E.g.,
15845 #define OPENSSL_ALGORITHM_DEFINES
15846 #include <openssl/opensslconf.h>
15847 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15849 *Richard Levitte, Ulf and Bodo Möller*
15851 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15856 * Change the 'other' type in certificate aux info to a STACK_OF
15857 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15858 the required ASN1 format: arbitrary types determined by an OID.
15862 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15863 argument to 'req'. This is not because the function is newer or
15864 better than others it just uses the work 'NEW' in the certificate
15865 request header lines. Some software needs this.
15869 * Reorganise password command line arguments: now passwords can be
15870 obtained from various sources. Delete the PEM_cb function and make
15871 it the default behaviour: i.e. if the callback is NULL and the
15872 usrdata argument is not NULL interpret it as a null terminated pass
15873 phrase. If usrdata and the callback are NULL then the pass phrase
15874 is prompted for as usual.
15878 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15879 the support is automatically enabled. The resulting binaries will
15880 autodetect the card and use it if present.
15882 *Ben Laurie and Compaq Inc.*
15884 * Work around for Netscape hang bug. This sends certificate request
15885 and server done in one record. Since this is perfectly legal in the
15886 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15887 the bugs/SSLv3 entry for more info.
15891 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15895 * Add -rand argument to smime and pkcs12 applications and read/write
15900 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15904 * Add command line password options to the remaining applications.
15908 * Bug fix for BN_div_recp() for numerators with an even number of
15913 * More tests in bntest.c, and changed test_bn output.
15917 * ./config recognizes MacOS X now.
15921 * Bug fix for BN_div() when the first words of num and divisor are
15922 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15926 * Add support for various broken PKCS#8 formats, and command line
15927 options to produce them.
15931 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15932 get temporary BIGNUMs from a BN_CTX.
15936 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15941 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15942 include a #define from the old name to the new. The original intent
15943 was that statically linked binaries could for example just call
15944 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15945 link with digests. This never worked because SSLeay_add_all_digests()
15946 and SSLeay_add_all_ciphers() were in the same source file so calling
15947 one would link with the other. They are now in separate source files.
15951 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15955 * Use a less unusual form of the Miller-Rabin primality test (it used
15956 a binary algorithm for exponentiation integrated into the Miller-Rabin
15957 loop, our standard modexp algorithms are faster).
15961 * Support for the EBCDIC character set completed.
15963 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15965 * Source code cleanups: use const where appropriate, eliminate casts,
15966 use `void *` instead of `char *` in lhash.
15970 * Bugfix: ssl3_send_server_key_exchange was not restartable
15971 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15972 this the server could overwrite ephemeral keys that the client
15977 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15978 using 50 iterations of the Rabin-Miller test.
15980 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15981 iterations of the Rabin-Miller test as required by the appendix
15982 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15983 As BN_is_prime_fasttest includes trial division, DSA parameter
15984 generation becomes much faster.
15986 This implies a change for the callback functions in DSA_is_prime
15987 and DSA_generate_parameters: The callback function is called once
15988 for each positive witness in the Rabin-Miller test, not just
15989 occasionally in the inner loop; and the parameters to the
15990 callback function now provide an iteration count for the outer
15991 loop rather than for the current invocation of the inner loop.
15992 DSA_generate_parameters additionally can call the callback
15993 function with an 'iteration count' of -1, meaning that a
15994 candidate has passed the trial division test (when q is generated
15995 from an application-provided seed, trial division is skipped).
15999 * New function BN_is_prime_fasttest that optionally does trial
16000 division before starting the Rabin-Miller test and has
16001 an additional BN_CTX * argument (whereas BN_is_prime always
16002 has to allocate at least one BN_CTX).
16003 'callback(1, -1, cb_arg)' is called when a number has passed the
16004 trial division stage.
16008 * Fix for bug in CRL encoding. The validity dates weren't being handled
16013 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16017 * New function BN_pseudo_rand().
16021 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16022 bignum version of BN_from_montgomery() with the working code from
16023 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16028 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16029 made it impossible to use the same SSL_SESSION data structure in
16030 SSL2 clients in multiple threads.
16034 * The return value of RAND_load_file() no longer counts bytes obtained
16035 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16036 to seed the PRNG (previously an explicit byte count was required).
16038 *Ulf Möller, Bodo Möller*
16040 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16041 used `char *` instead of `void *` and had casts all over the place.
16045 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16049 * Retain source code compatibility for BN_prime_checks macro:
16050 BN_is_prime(..., BN_prime_checks, ...) now uses
16051 BN_prime_checks_for_size to determine the appropriate number of
16052 Rabin-Miller iterations.
16056 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16057 DH_CHECK_P_NOT_SAFE_PRIME.
16058 (Check if this is true? OpenPGP calls them "strong".)
16062 * Merge the functionality of "dh" and "gendh" programs into a new program
16063 "dhparam". The old programs are retained for now but will handle DH keys
16064 (instead of parameters) in future.
16068 * Make the ciphers, s_server and s_client programs check the return values
16069 when a new cipher list is set.
16073 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16074 ciphers. Before when the 56bit ciphers were enabled the sorting was
16077 The syntax for the cipher sorting has been extended to support sorting by
16078 cipher-strength (using the strength_bits hard coded in the tables).
16079 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16081 Fix a bug in the cipher-command parser: when supplying a cipher command
16082 string with an "undefined" symbol (neither command nor alphanumeric
16083 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16084 an error is flagged.
16086 Due to the strength-sorting extension, the code of the
16087 ssl_create_cipher_list() function was completely rearranged. I hope that
16088 the readability was also increased :-)
16090 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16092 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16093 for the first serial number and places 2 in the serial number file. This
16094 avoids problems when the root CA is created with serial number zero and
16095 the first user certificate has the same issuer name and serial number
16100 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16101 the new code. Add documentation for this stuff.
16105 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16106 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16107 structures and behave in an analogous way to the X509v3 functions:
16108 they shouldn't be called directly but wrapper functions should be used
16111 So we also now have some wrapper functions that call the X509at functions
16112 when passed certificate requests. (TO DO: similar things can be done with
16113 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16114 things. Some of these need some d2i or i2d and print functionality
16115 because they handle more complex structures.)
16119 * Add missing #ifndefs that caused missing symbols when building libssl
16120 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16121 NO_RSA in `ssl/s2*.c`.
16123 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16125 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16126 has a return value which indicates the quality of the random data
16127 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16128 error queue. New function RAND_pseudo_bytes() generates output that is
16129 guaranteed to be unique but not unpredictable. RAND_add is like
16130 RAND_seed, but takes an extra argument for an entropy estimate
16131 (RAND_seed always assumes full entropy).
16135 * Do more iterations of Rabin-Miller probable prime test (specifically,
16136 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16137 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16138 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16139 false-positive rate of at most 2^-80 for random input.
16143 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16147 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16148 in the 0.9.5 release), this returns the chain
16149 from an X509_CTX structure with a dup of the stack and all
16150 the X509 reference counts upped: so the stack will exist
16151 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16154 Also make SSL_SESSION_print() print out the verify return
16159 * Add manpage for the pkcs12 command. Also change the default
16160 behaviour so MAC iteration counts are used unless the new
16161 -nomaciter option is used. This improves file security and
16162 only older versions of MSIE (4.0 for example) need it.
16166 * Honor the no-xxx Configure options when creating .DEF files.
16170 * Add PKCS#10 attributes to field table: challengePassword,
16171 unstructuredName and unstructuredAddress. These are taken from
16172 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16173 international characters are used.
16175 More changes to X509_ATTRIBUTE code: allow the setting of types
16176 based on strings. Remove the 'loc' parameter when adding
16177 attributes because these will be a SET OF encoding which is sorted
16182 * Initial changes to the 'req' utility to allow request generation
16183 automation. This will allow an application to just generate a template
16184 file containing all the field values and have req construct the
16187 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16188 used all over the place including certificate requests and PKCS#7
16189 structures. They are currently handled manually where necessary with
16190 some primitive wrappers for PKCS#7. The new functions behave in a
16191 manner analogous to the X509 extension functions: they allow
16192 attributes to be looked up by NID and added.
16194 Later something similar to the X509V3 code would be desirable to
16195 automatically handle the encoding, decoding and printing of the
16196 more complex types. The string types like challengePassword can
16197 be handled by the string table functions.
16199 Also modified the multi byte string table handling. Now there is
16200 a 'global mask' which masks out certain types. The table itself
16201 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16202 is useful when for example there is only one permissible type
16203 (as in countryName) and using the mask might result in no valid
16208 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16209 SSL_get_peer_finished to allow applications to obtain the latest
16210 Finished messages sent to the peer or expected from the peer,
16211 respectively. (SSL_get_peer_finished is usually the Finished message
16212 actually received from the peer, otherwise the protocol will be aborted.)
16214 As the Finished message are message digests of the complete handshake
16215 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16216 be used for external authentication procedures when the authentication
16217 provided by SSL/TLS is not desired or is not enough.
16221 * Enhanced support for Alpha Linux is added. Now ./config checks if
16222 the host supports BWX extension and if Compaq C is present on the
16223 $PATH. Just exploiting of the BWX extension results in 20-30%
16224 performance kick for some algorithms, e.g. DES and RC4 to mention
16225 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16230 * Add support for MS "fast SGC". This is arguably a violation of the
16231 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16232 weak crypto and after checking the certificate is SGC a second one
16233 with strong crypto. MS SGC stops the first handshake after receiving
16234 the server certificate message and sends a second client hello. Since
16235 a server will typically do all the time consuming operations before
16236 expecting any further messages from the client (server key exchange
16237 is the most expensive) there is little difference between the two.
16239 To get OpenSSL to support MS SGC we have to permit a second client
16240 hello message after we have sent server done. In addition we have to
16241 reset the MAC if we do get this second client hello.
16245 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16246 if a DER encoded private key is RSA or DSA traditional format. Changed
16247 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16248 format DER encoded private key. Newer code should use PKCS#8 format which
16249 has the key type encoded in the ASN1 structure. Added DER private key
16250 support to pkcs8 application.
16254 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16255 ciphersuites has been selected (as required by the SSL 3/TLS 1
16256 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16257 is set, we interpret this as a request to violate the specification
16258 (the worst that can happen is a handshake failure, and 'correct'
16259 behaviour would result in a handshake failure anyway).
16263 * In SSL_CTX_add_session, take into account that there might be multiple
16264 SSL_SESSION structures with the same session ID (e.g. when two threads
16265 concurrently obtain them from an external cache).
16266 The internal cache can handle only one SSL_SESSION with a given ID,
16267 so if there's a conflict, we now throw out the old one to achieve
16272 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16273 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16274 some routines that use cipher OIDs: some ciphers do not have OIDs
16275 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16280 * Simplify the trust setting structure and code. Now we just have
16281 two sequences of OIDs for trusted and rejected settings. These will
16282 typically have values the same as the extended key usage extension
16283 and any application specific purposes.
16285 The trust checking code now has a default behaviour: it will just
16286 check for an object with the same NID as the passed id. Functions can
16287 be provided to override either the default behaviour or the behaviour
16288 for a given id. SSL client, server and email already have functions
16289 in place for compatibility: they check the NID and also return "trusted"
16290 if the certificate is self signed.
16294 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16295 traditional format into an EVP_PKEY structure.
16299 * Add a password callback function PEM_cb() which either prompts for
16300 a password if usr_data is NULL or otherwise assumes it is a null
16301 terminated password. Allow passwords to be passed on command line
16302 environment or config files in a few more utilities.
16306 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16307 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16308 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16309 Update documentation.
16313 * Support for ASN1 "NULL" type. This could be handled before by using
16314 ASN1_TYPE but there wasn't any function that would try to read a NULL
16315 and produce an error if it couldn't. For compatibility we also have
16316 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16317 don't allocate anything because they don't need to.
16321 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16324 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16326 * Rebuild of the memory allocation routines used by OpenSSL code and
16327 possibly others as well. The purpose is to make an interface that
16328 provide hooks so anyone can build a separate set of allocation and
16329 deallocation routines to be used by OpenSSL, for example memory
16330 pool implementations, or something else, which was previously hard
16331 since Malloc(), Realloc() and Free() were defined as macros having
16332 the values malloc, realloc and free, respectively (except for Win32
16333 compilations). The same is provided for memory debugging code.
16334 OpenSSL already comes with functionality to find memory leaks, but
16335 this gives people a chance to debug other memory problems.
16337 With these changes, a new set of functions and macros have appeared:
16339 CRYPTO_set_mem_debug_functions() [F]
16340 CRYPTO_get_mem_debug_functions() [F]
16341 CRYPTO_dbg_set_options() [F]
16342 CRYPTO_dbg_get_options() [F]
16343 CRYPTO_malloc_debug_init() [M]
16345 The memory debug functions are NULL by default, unless the library
16346 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16347 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16348 gives the standard debugging functions that come with OpenSSL) or
16349 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16350 provided by the library user) must be used. When the standard
16351 debugging functions are used, CRYPTO_dbg_set_options can be used to
16352 request additional information:
16353 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16354 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16356 Also, things like CRYPTO_set_mem_functions will always give the
16357 expected result (the new set of functions is used for allocation
16358 and deallocation) at all times, regardless of platform and compiler
16361 To finish it up, some functions that were never use in any other
16362 way than through macros have a new API and new semantic:
16364 CRYPTO_dbg_malloc()
16365 CRYPTO_dbg_realloc()
16368 All macros of value have retained their old syntax.
16370 *Richard Levitte and Bodo Moeller*
16372 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16373 ordering of SMIMECapabilities wasn't in "strength order" and there
16374 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16379 * Some ASN1 types with illegal zero length encoding (INTEGER,
16380 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16382 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16384 * Merge in my S/MIME library for OpenSSL. This provides a simple
16385 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16386 functionality to handle multipart/signed properly) and a utility
16387 called 'smime' to call all this stuff. This is based on code I
16388 originally wrote for Celo who have kindly allowed it to be
16389 included in OpenSSL.
16393 * Add variants des_set_key_checked and des_set_key_unchecked of
16394 des_set_key (aka des_key_sched). Global variable des_check_key
16395 decides which of these is called by des_set_key; this way
16396 des_check_key behaves as it always did, but applications and
16397 the library itself, which was buggy for des_check_key == 1,
16398 have a cleaner way to pick the version they need.
16402 * New function PKCS12_newpass() which changes the password of a
16407 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16408 dynamic mix. In both cases the ids can be used as an index into the
16409 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16410 functions so they accept a list of the field values and the
16411 application doesn't need to directly manipulate the X509_TRUST
16416 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16421 * Modify the way the V3 extension code looks up extensions. This now
16422 works in a similar way to the object code: we have some "standard"
16423 extensions in a static table which is searched with OBJ_bsearch()
16424 and the application can add dynamic ones if needed. The file
16425 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16426 updated whenever a new extension is added to the core code and kept
16427 in ext_nid order. There is a simple program 'tabtest.c' which checks
16428 this. New extensions are not added too often so this file can readily
16429 be maintained manually.
16431 There are two big advantages in doing things this way. The extensions
16432 can be looked up immediately and no longer need to be "added" using
16433 X509V3_add_standard_extensions(): this function now does nothing.
16434 Side note: I get *lots* of email saying the extension code doesn't
16435 work because people forget to call this function.
16436 Also no dynamic allocation is done unless new extensions are added:
16437 so if we don't add custom extensions there is no need to call
16438 X509V3_EXT_cleanup().
16442 * Modify enc utility's salting as follows: make salting the default. Add a
16443 magic header, so unsalted files fail gracefully instead of just decrypting
16444 to garbage. This is because not salting is a big security hole, so people
16445 should be discouraged from doing it.
16449 * Fixes and enhancements to the 'x509' utility. It allowed a message
16450 digest to be passed on the command line but it only used this
16451 parameter when signing a certificate. Modified so all relevant
16452 operations are affected by the digest parameter including the
16453 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16454 DSA key was used because it didn't fix the digest.
16458 * Initial certificate chain verify code. Currently tests the untrusted
16459 certificates for consistency with the verify purpose (which is set
16460 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16462 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16463 this is because it will reject chains with invalid extensions whereas
16464 every previous version of OpenSSL and SSLeay made no checks at all.
16466 Trust code: checks the root CA for the relevant trust settings. Trust
16467 settings have an initial value consistent with the verify purpose: e.g.
16468 if the verify purpose is for SSL client use it expects the CA to be
16469 trusted for SSL client use. However the default value can be changed to
16470 permit custom trust settings: one example of this would be to only trust
16471 certificates from a specific "secure" set of CAs.
16473 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16474 which should be used for version portability: especially since the
16475 verify structure is likely to change more often now.
16477 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16478 to set them. If not set then assume SSL clients will verify SSL servers
16481 Two new options to the verify program: -untrusted allows a set of
16482 untrusted certificates to be passed in and -purpose which sets the
16483 intended purpose of the certificate. If a purpose is set then the
16484 new chain verify code is used to check extension consistency.
16488 * Support for the authority information access extension.
16492 * Modify RSA and DSA PEM read routines to transparently handle
16493 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16494 public keys in a format compatible with certificate
16495 SubjectPublicKeyInfo structures. Unfortunately there were already
16496 functions called *_PublicKey_* which used various odd formats so
16497 these are retained for compatibility: however the DSA variants were
16498 never in a public release so they have been deleted. Changed dsa/rsa
16499 utilities to handle the new format: note no releases ever handled public
16500 keys so we should be OK.
16502 The primary motivation for this change is to avoid the same fiasco
16503 that dogs private keys: there are several incompatible private key
16504 formats some of which are standard and some OpenSSL specific and
16505 require various evil hacks to allow partial transparent handling and
16506 even then it doesn't work with DER formats. Given the option anything
16507 other than PKCS#8 should be dumped: but the other formats have to
16508 stay in the name of compatibility.
16510 With public keys and the benefit of hindsight one standard format
16511 is used which works with EVP_PKEY, RSA or DSA structures: though
16512 it clearly returns an error if you try to read the wrong kind of key.
16514 Added a -pubkey option to the 'x509' utility to output the public key.
16515 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16516 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16517 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16518 that do the same as the `EVP_PKEY_assign_*()` except they up the
16519 reference count of the added key (they don't "swallow" the
16524 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16525 CRLs would fail if the file contained no certificates or no CRLs:
16526 added a new function to read in both types and return the number
16527 read: this means that if none are read it will be an error. The
16528 DER versions of the certificate and CRL reader would always fail
16529 because it isn't possible to mix certificates and CRLs in DER format
16530 without choking one or the other routine. Changed this to just read
16531 a certificate: this is the best we can do. Also modified the code
16532 in `apps/verify.c` to take notice of return codes: it was previously
16533 attempting to read in certificates from NULL pointers and ignoring
16534 any errors: this is one reason why the cert and CRL reader seemed
16535 to work. It doesn't check return codes from the default certificate
16536 routines: these may well fail if the certificates aren't installed.
16540 * Code to support otherName option in GeneralName.
16544 * First update to verify code. Change the verify utility
16545 so it warns if it is passed a self signed certificate:
16546 for consistency with the normal behaviour. X509_verify
16547 has been modified to it will now verify a self signed
16548 certificate if *exactly* the same certificate appears
16549 in the store: it was previously impossible to trust a
16550 single self signed certificate. This means that:
16551 openssl verify ss.pem
16552 now gives a warning about a self signed certificate but
16553 openssl verify -CAfile ss.pem ss.pem
16558 * For servers, store verify_result in SSL_SESSION data structure
16559 (and add it to external session representation).
16560 This is needed when client certificate verifications fails,
16561 but an application-provided verification callback (set by
16562 SSL_CTX_set_cert_verify_callback) allows accepting the session
16563 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16564 but returns 1): When the session is reused, we have to set
16565 ssl->verify_result to the appropriate error code to avoid
16568 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16570 * Fix a bug in the new PKCS#7 code: it didn't consider the
16571 case in PKCS7_dataInit() where the signed PKCS7 structure
16572 didn't contain any existing data because it was being created.
16574 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16576 * Add a salt to the key derivation routines in enc.c. This
16577 forms the first 8 bytes of the encrypted file. Also add a
16578 -S option to allow a salt to be input on the command line.
16582 * New function X509_cmp(). Oddly enough there wasn't a function
16583 to compare two certificates. We do this by working out the SHA1
16584 hash and comparing that. X509_cmp() will be needed by the trust
16589 * SSL_get1_session() is like SSL_get_session(), but increments
16590 the reference count in the SSL_SESSION returned.
16592 *Geoff Thorpe <geoff@eu.c2.net>*
16594 * Fix for 'req': it was adding a null to request attributes.
16595 Also change the X509_LOOKUP and X509_INFO code to handle
16596 certificate auxiliary information.
16600 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16605 * Add the possibility to add extra information to the memory leak
16606 detecting output, to form tracebacks, showing from where each
16607 allocation was originated: CRYPTO_push_info("constant string") adds
16608 the string plus current file name and line number to a per-thread
16609 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16610 is like calling CYRPTO_pop_info() until the stack is empty.
16611 Also updated memory leak detection code to be multi-thread-safe.
16615 * Add options -text and -noout to pkcs7 utility and delete the
16616 encryption options which never did anything. Update docs.
16620 * Add options to some of the utilities to allow the pass phrase
16621 to be included on either the command line (not recommended on
16622 OSes like Unix) or read from the environment. Update the
16623 manpages and fix a few bugs.
16627 * Add a few manpages for some of the openssl commands.
16631 * Fix the -revoke option in ca. It was freeing up memory twice,
16632 leaking and not finding already revoked certificates.
16636 * Extensive changes to support certificate auxiliary information.
16637 This involves the use of X509_CERT_AUX structure and X509_AUX
16638 functions. An X509_AUX function such as PEM_read_X509_AUX()
16639 can still read in a certificate file in the usual way but it
16640 will also read in any additional "auxiliary information". By
16641 doing things this way a fair degree of compatibility can be
16642 retained: existing certificates can have this information added
16643 using the new 'x509' options.
16645 Current auxiliary information includes an "alias" and some trust
16646 settings. The trust settings will ultimately be used in enhanced
16647 certificate chain verification routines: currently a certificate
16648 can only be trusted if it is self signed and then it is trusted
16653 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16654 The problem was that one of the replacement routines had not been working
16655 since SSLeay releases. For now the offending routine has been replaced
16656 with non-optimised assembler. Even so, this now gives around 95%
16657 performance improvement for 1024 bit RSA signs.
16661 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16662 handling. Most clients have the effective key size in bits equal to
16663 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16664 A few however don't do this and instead use the size of the decrypted key
16665 to determine the RC2 key length and the AlgorithmIdentifier to determine
16666 the effective key length. In this case the effective key length can still
16667 be 40 bits but the key length can be 168 bits for example. This is fixed
16668 by manually forcing an RC2 key into the EVP_PKEY structure because the
16669 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16670 the key length and effective key length are equal.
16674 * Add a bunch of functions that should simplify the creation of
16675 X509_NAME structures. Now you should be able to do:
16676 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16677 and have it automatically work out the correct field type and fill in
16678 the structures. The more adventurous can try:
16679 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16680 and it will (hopefully) work out the correct multibyte encoding.
16684 * Change the 'req' utility to use the new field handling and multibyte
16685 copy routines. Before the DN field creation was handled in an ad hoc
16686 way in req, ca, and x509 which was rather broken and didn't support
16687 BMPStrings or UTF8Strings. Since some software doesn't implement
16688 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16689 using the dirstring_type option. See the new comment in the default
16690 openssl.cnf for more info.
16694 * Make crypto/rand/md_rand.c more robust:
16695 - Assure unique random numbers after fork().
16696 - Make sure that concurrent threads access the global counter and
16697 md serializably so that we never lose entropy in them
16698 or use exactly the same state in multiple threads.
16699 Access to the large state is not always serializable because
16700 the additional locking could be a performance killer, and
16701 md should be large enough anyway.
16705 * New file `apps/app_rand.c` with commonly needed functionality
16706 for handling the random seed file.
16708 Use the random seed file in some applications that previously did not:
16710 dsaparam -genkey (which also ignored its '-rand' option),
16713 x509 (when signing).
16714 Except on systems with /dev/urandom, it is crucial to have a random
16715 seed file at least for key creation, DSA signing, and for DH exchanges;
16716 for RSA signatures we could do without one.
16718 gendh and gendsa (unlike genrsa) used to read only the first byte
16719 of each file listed in the '-rand' option. The function as previously
16720 found in genrsa is now in app_rand.c and is used by all programs
16721 that support '-rand'.
16725 * In RAND_write_file, use mode 0600 for creating files;
16726 don't just chmod when it may be too late.
16730 * Report an error from X509_STORE_load_locations
16731 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16735 * New function ASN1_mbstring_copy() this copies a string in either
16736 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16737 into an ASN1_STRING type. A mask of permissible types is passed
16738 and it chooses the "minimal" type to use or an error if not type
16743 * Add function equivalents to the various macros in asn1.h. The old
16744 macros are retained with an `M_` prefix. Code inside the library can
16745 use the `M_` macros. External code (including the openssl utility)
16746 should *NOT* in order to be "shared library friendly".
16750 * Add various functions that can check a certificate's extensions
16751 to see if it usable for various purposes such as SSL client,
16752 server or S/MIME and CAs of these types. This is currently
16753 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16754 verification. Also added a -purpose flag to x509 utility to
16755 print out all the purposes.
16759 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16764 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16765 for, obtain and decode and extension and obtain its critical flag.
16766 This allows all the necessary extension code to be handled in a
16767 single function call.
16771 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16772 platforms. See crypto/rc4/rc4_enc.c for further details.
16776 * New -noout option to asn1parse. This causes no output to be produced
16777 its main use is when combined with -strparse and -out to extract data
16778 from a file (which may not be in ASN.1 format).
16782 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16783 when producing the local key id.
16785 *Richard Levitte <levitte@stacken.kth.se>*
16787 * New option -dhparam in s_server. This allows a DH parameter file to be
16788 stated explicitly. If it is not stated then it tries the first server
16789 certificate file. The previous behaviour hard coded the filename
16794 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16795 a public key to be input or output. For example:
16796 openssl rsa -in key.pem -pubout -out pubkey.pem
16797 Also added necessary DSA public key functions to handle this.
16801 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16802 in the message. This was handled by allowing
16803 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16805 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16807 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16808 to the end of the strings whereas this didn't. This would cause problems
16809 if strings read with d2i_ASN1_bytes() were later modified.
16811 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16813 * Fix for base64 decode bug. When a base64 bio reads only one line of
16814 data and it contains EOF it will end up returning an error. This is
16815 caused by input 46 bytes long. The cause is due to the way base64
16816 BIOs find the start of base64 encoded data. They do this by trying a
16817 trial decode on each line until they find one that works. When they
16818 do a flag is set and it starts again knowing it can pass all the
16819 data directly through the decoder. Unfortunately it doesn't reset
16820 the context it uses. This means that if EOF is reached an attempt
16821 is made to pass two EOFs through the context and this causes the
16822 resulting error. This can also cause other problems as well. As is
16823 usual with these problems it takes *ages* to find and the fix is
16824 trivial: move one line.
16826 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16828 * Ugly workaround to get s_client and s_server working under Windows. The
16829 old code wouldn't work because it needed to select() on sockets and the
16830 tty (for keypresses and to see if data could be written). Win32 only
16831 supports select() on sockets so we select() with a 1s timeout on the
16832 sockets and then see if any characters are waiting to be read, if none
16833 are present then we retry, we also assume we can always write data to
16834 the tty. This isn't nice because the code then blocks until we've
16835 received a complete line of data and it is effectively polling the
16836 keyboard at 1s intervals: however it's quite a bit better than not
16837 working at all :-) A dedicated Windows application might handle this
16838 with an event loop for example.
16842 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16843 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16844 will be called when RSA_sign() and RSA_verify() are used. This is useful
16845 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16846 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16847 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16848 This necessitated the support of an extra signature type NID_md5_sha1
16849 for SSL signatures and modifications to the SSL library to use it instead
16850 of calling RSA_public_decrypt() and RSA_private_encrypt().
16854 * Add new -verify -CAfile and -CApath options to the crl program, these
16855 will lookup a CRL issuers certificate and verify the signature in a
16856 similar way to the verify program. Tidy up the crl program so it
16857 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16858 less strict. It will now permit CRL extensions even if it is not
16859 a V2 CRL: this will allow it to tolerate some broken CRLs.
16863 * Initialize all non-automatic variables each time one of the openssl
16864 sub-programs is started (this is necessary as they may be started
16865 multiple times from the "OpenSSL>" prompt).
16867 *Lennart Bang, Bodo Moeller*
16869 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16870 removing all other RSA functionality (this is what NO_RSA does). This
16871 is so (for example) those in the US can disable those operations covered
16872 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16877 * Non-copying interface to BIO pairs.
16878 (still largely untested)
16882 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16883 ASCII string. This was handled independently in various places before.
16887 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16888 UTF8 strings a character at a time.
16892 * Use client_version from client hello to select the protocol
16893 (s23_srvr.c) and for RSA client key exchange verification
16894 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16898 * Add various utility functions to handle SPKACs, these were previously
16899 handled by poking round in the structure internals. Added new function
16900 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16901 print, verify and generate SPKACs. Based on an original idea from
16902 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16906 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16910 * Allow the config file extension section to be overwritten on the
16911 command line. Based on an original idea from Massimiliano Pala
16912 <madwolf@comune.modena.it>. The new option is called -extensions
16913 and can be applied to ca, req and x509. Also -reqexts to override
16914 the request extensions in req and -crlexts to override the crl extensions
16919 * Add new feature to the SPKAC handling in ca. Now you can include
16920 the same field multiple times by preceding it by "XXXX." for example:
16923 this is the same syntax as used in the req config file.
16927 * Allow certificate extensions to be added to certificate requests. These
16928 are specified in a 'req_extensions' option of the req section of the
16929 config file. They can be printed out with the -text option to req but
16930 are otherwise ignored at present.
16934 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16935 data read consists of only the final block it would not decrypted because
16936 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16937 A misplaced 'break' also meant the decrypted final block might not be
16938 copied until the next read.
16942 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16943 a few extra parameters to the DH structure: these will be useful if
16944 for example we want the value of 'q' or implement X9.42 DH.
16948 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16949 provides hooks that allow the default DSA functions or functions on a
16950 "per key" basis to be replaced. This allows hardware acceleration and
16951 hardware key storage to be handled without major modification to the
16952 library. Also added low level modexp hooks and CRYPTO_EX structure and
16953 associated functions.
16957 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16958 as "read only": it can't be written to and the buffer it points to will
16959 not be freed. Reading from a read only BIO is much more efficient than
16960 a normal memory BIO. This was added because there are several times when
16961 an area of memory needs to be read from a BIO. The previous method was
16962 to create a memory BIO and write the data to it, this results in two
16963 copies of the data and an O(n^2) reading algorithm. There is a new
16964 function BIO_new_mem_buf() which creates a read only memory BIO from
16965 an area of memory. Also modified the PKCS#7 routines to use read only
16970 * Bugfix: ssl23_get_client_hello did not work properly when called in
16971 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16972 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16973 but a retry condition occurred while trying to read the rest.
16977 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16978 NID_pkcs7_encrypted by default: this was wrong since this should almost
16979 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16980 the encrypted data type: this is a more sensible place to put it and it
16981 allows the PKCS#12 code to be tidied up that duplicated this
16986 * Changed obj_dat.pl script so it takes its input and output files on
16987 the command line. This should avoid shell escape redirection problems
16992 * Initial support for certificate extension requests, these are included
16993 in things like Xenroll certificate requests. Included functions to allow
16994 extensions to be obtained and added.
16998 * -crlf option to s_client and s_server for sending newlines as
16999 CRLF (as required by many protocols).
17003 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17005 * Install libRSAglue.a when OpenSSL is built with RSAref.
17007 *Ralf S. Engelschall*
17009 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17011 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17013 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17018 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17019 DH parameters/keys (q is lost during that conversion, but the resulting
17020 DH parameters contain its length).
17022 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17023 much faster than DH_generate_parameters (which creates parameters
17024 where `p = 2*q + 1`), and also the smaller q makes DH computations
17025 much more efficient (160-bit exponentiation instead of 1024-bit
17026 exponentiation); so this provides a convenient way to support DHE
17027 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17028 utter importance to use
17029 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17031 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17032 when such DH parameters are used, because otherwise small subgroup
17033 attacks may become possible!
17037 * Avoid memory leak in i2d_DHparams.
17041 * Allow the -k option to be used more than once in the enc program:
17042 this allows the same encrypted message to be read by multiple recipients.
17046 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17047 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17048 it will always use the numerical form of the OID, even if it has a short
17053 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17054 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17055 otherwise bn_mod_exp was called. In the case of hardware keys for example
17056 no private key components need be present and it might store extra data
17057 in the RSA structure, which cannot be accessed from bn_mod_exp.
17058 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17059 private key operations.
17063 * Added support for SPARC Linux.
17067 * pem_password_cb function type incompatibly changed from
17068 typedef int pem_password_cb(char *buf, int size, int rwflag);
17070 ....(char *buf, int size, int rwflag, void *userdata);
17071 so that applications can pass data to their callbacks:
17072 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17073 additional void * argument, which is just handed through whenever
17074 the password callback is called.
17076 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17078 New function SSL_CTX_set_default_passwd_cb_userdata.
17080 Compatibility note: As many C implementations push function arguments
17081 onto the stack in reverse order, the new library version is likely to
17082 interoperate with programs that have been compiled with the old
17083 pem_password_cb definition (PEM_whatever takes some data that
17084 happens to be on the stack as its last argument, and the callback
17085 just ignores this garbage); but there is no guarantee whatsoever that
17088 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17089 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17090 problems not only on Windows, but also on some Unix platforms.
17091 To avoid problematic command lines, these definitions are now in an
17092 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17093 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17097 * MIPS III/IV assembler module is reimplemented.
17101 * More DES library cleanups: remove references to srand/rand and
17102 delete an unused file.
17106 * Add support for the free Netwide assembler (NASM) under Win32,
17107 since not many people have MASM (ml) and it can be hard to obtain.
17108 This is currently experimental but it seems to work OK and pass all
17109 the tests. Check out INSTALL.W32 for info.
17113 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17114 without temporary keys kept an extra copy of the server key,
17115 and connections with temporary keys did not free everything in case
17120 * New function RSA_check_key and new openssl rsa option -check
17121 for verifying the consistency of RSA keys.
17123 *Ulf Moeller, Bodo Moeller*
17125 * Various changes to make Win32 compile work:
17126 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17127 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17128 comparison" warnings.
17129 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17133 * Add a debugging option to PKCS#5 v2 key generation function: when
17134 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17135 derived keys are printed to stderr.
17139 * Copy the flags in ASN1_STRING_dup().
17141 *Roman E. Pavlov <pre@mo.msk.ru>*
17143 * The x509 application mishandled signing requests containing DSA
17144 keys when the signing key was also DSA and the parameters didn't match.
17146 It was supposed to omit the parameters when they matched the signing key:
17147 the verifying software was then supposed to automatically use the CA's
17148 parameters if they were absent from the end user certificate.
17150 Omitting parameters is no longer recommended. The test was also
17151 the wrong way round! This was probably due to unusual behaviour in
17152 EVP_cmp_parameters() which returns 1 if the parameters match.
17153 This meant that parameters were omitted when they *didn't* match and
17154 the certificate was useless. Certificates signed with 'ca' didn't have
17157 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17159 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17160 The interface is as follows:
17161 Applications can use
17162 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17163 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17164 "off" is now the default.
17165 The library internally uses
17166 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17167 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17168 to disable memory-checking temporarily.
17170 Some inconsistent states that previously were possible (and were
17171 even the default) are now avoided.
17173 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17174 with each memory chunk allocated; this is occasionally more helpful
17175 than just having a counter.
17177 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17179 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17184 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17185 which largely parallels "options", but is for changing API behaviour,
17186 whereas "options" are about protocol behaviour.
17187 Initial "mode" flags are:
17189 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17190 a single record has been written.
17191 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17192 retries use the same buffer location.
17193 (But all of the contents must be
17198 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17201 * Fix problems with no-hmac etc.
17203 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17205 * New functions RSA_get_default_method(), RSA_set_method() and
17206 RSA_get_method(). These allows replacement of RSA_METHODs without having
17207 to mess around with the internals of an RSA structure.
17211 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17212 Also really enable memory leak checks in openssl.c and in some
17215 *Chad C. Mulligan, Bodo Moeller*
17217 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17218 up the length of negative integers. This has now been simplified to just
17219 store the length when it is first determined and use it later, rather
17220 than trying to keep track of where data is copied and updating it to
17222 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17224 * Add a new function PKCS7_signatureVerify. This allows the verification
17225 of a PKCS#7 signature but with the signing certificate passed to the
17226 function itself. This contrasts with PKCS7_dataVerify which assumes the
17227 certificate is present in the PKCS#7 structure. This isn't always the
17228 case: certificates can be omitted from a PKCS#7 structure and be
17229 distributed by "out of band" means (such as a certificate database).
17233 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17234 function prototypes in pem.h, also change util/mkdef.pl to add the
17235 necessary function names.
17239 * mk1mf.pl (used by Windows builds) did not properly read the
17240 options set by Configure in the top level Makefile, and Configure
17241 was not even able to write more than one option correctly.
17242 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17246 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17247 file to be loaded from a BIO or FILE pointer. The BIO version will
17248 for example allow memory BIOs to contain config info.
17252 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17253 Whoever hopes to achieve shared-library compatibility across versions
17254 must use this, not the compile-time macro.
17255 (Exercise 0.9.4: Which is the minimum library version required by
17257 Note: All this applies only to multi-threaded programs, others don't
17262 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17263 through a BIO pair triggered the default case, i.e.
17264 SSLerr(...,SSL_R_UNKNOWN_STATE).
17268 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17269 can use the SSL library even if none of the specific BIOs is
17274 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17275 for the encoded length.
17277 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17279 * Add initial documentation of the X509V3 functions.
17283 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17284 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17285 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17286 secure PKCS#8 private key format with a high iteration count.
17290 * Fix determination of Perl interpreter: A perl or perl5
17291 *directory* in $PATH was also accepted as the interpreter.
17293 *Ralf S. Engelschall*
17295 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17296 wrong with it but it was very old and did things like calling
17297 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17298 unusual formatting.
17302 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17303 to use the new extension code.
17307 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17308 with macros. This should make it easier to change their form, add extra
17309 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17314 * Add to configuration table a new entry that can specify an alternative
17315 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17316 according to Mark Crispin <MRC@Panda.COM>.
17320 * DES CBC did not update the IV. Weird.
17324 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17325 Changing the behaviour of the former might break existing programs --
17326 where IV updating is needed, des_ncbc_encrypt can be used.
17329 * When bntest is run from "make test" it drives bc to check its
17330 calculations, as well as internally checking them. If an internal check
17331 fails, it needs to cause bc to give a non-zero result or make test carries
17332 on without noticing the failure. Fixed.
17336 * DES library cleanups.
17340 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17341 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17342 ciphers. NOTE: although the key derivation function has been verified
17343 against some published test vectors it has not been extensively tested
17344 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17349 * Instead of "mkdir -p", which is not fully portable, use new
17350 Perl script "util/mkdir-p.pl".
17354 * Rewrite the way password based encryption (PBE) is handled. It used to
17355 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17356 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17357 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17358 the 'parameter' field of the AlgorithmIdentifier is passed to the
17359 underlying key generation function so it must do its own ASN1 parsing.
17360 This has also changed the EVP_PBE_CipherInit() function which now has a
17361 'parameter' argument instead of literal salt and iteration count values
17362 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17366 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17367 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17368 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17369 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17370 value was just used as a "magic string" and not used directly its
17371 value doesn't matter.
17375 * Introduce some semblance of const correctness to BN. Shame C doesn't
17380 * "linux-sparc64" configuration (ultrapenguin).
17382 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17383 "linux-sparc" configuration.
17385 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17387 * config now generates no-xxx options for missing ciphers.
17391 * Support the EBCDIC character set (work in progress).
17392 File ebcdic.c not yet included because it has a different license.
17394 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17396 * Support BS2000/OSD-POSIX.
17398 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17400 * Make callbacks for key generation use `void *` instead of `char *`.
17404 * Make S/MIME samples compile (not yet tested).
17408 * Additional typesafe stacks.
17412 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17416 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17418 * New configuration variant "sco5-gcc".
17420 * Updated some demos.
17422 *Sean O Riordain, Wade Scholine*
17424 * Add missing BIO_free at exit of pkcs12 application.
17428 * Fix memory leak in conf.c.
17432 * Updates for Win32 to assembler version of MD5.
17436 * Set #! path to perl in `apps/der_chop` to where we found it
17437 instead of using a fixed path.
17441 * SHA library changes for irix64-mips4-cc.
17445 * Improvements for VMS support.
17449 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17451 * Bignum library bug fix. IRIX 6 passes "make test" now!
17452 This also avoids the problems with SC4.2 and unpatched SC5.
17454 *Andy Polyakov <appro@fy.chalmers.se>*
17456 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17457 These are required because of the typesafe stack would otherwise break
17458 existing code. If old code used a structure member which used to be STACK
17459 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17460 sk_num or sk_value it would produce an error because the num, data members
17461 are not present in STACK_OF. Now it just produces a warning. sk_set
17462 replaces the old method of assigning a value to sk_value
17463 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17464 that does this will no longer work (and should use sk_set instead) but
17465 this could be regarded as a "questionable" behaviour anyway.
17469 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17470 correctly handle encrypted S/MIME data.
17474 * Change type of various DES function arguments from des_cblock
17475 (which means, in function argument declarations, pointer to char)
17476 to des_cblock * (meaning pointer to array with 8 char elements),
17477 which allows the compiler to do more typechecking; it was like
17478 that back in SSLeay, but with lots of ugly casts.
17480 Introduce new type const_des_cblock.
17484 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17485 problems: find RecipientInfo structure that matches recipient certificate
17486 and initialise the ASN1 structures properly based on passed cipher.
17490 * Belatedly make the BN tests actually check the results.
17494 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17495 to and from BNs: it was completely broken. New compilation option
17496 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17497 key elements as negative integers.
17501 * Reorganize and speed up MD5.
17503 *Andy Polyakov <appro@fy.chalmers.se>*
17507 *Richard Levitte <richard@levitte.org>*
17509 * New option -out to asn1parse to allow the parsed structure to be
17510 output to a file. This is most useful when combined with the -strparse
17511 option to examine the output of things like OCTET STRINGS.
17515 * Make SSL library a little more fool-proof by not requiring any longer
17516 that `SSL_set_{accept,connect}_state` be called before
17517 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17518 in many applications because usually everything *appeared* to work as
17519 intended anyway -- now it really works as intended).
17523 * Move openssl.cnf out of lib/.
17527 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17528 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17529 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17531 *Ralf S. Engelschall*
17533 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17534 handle PKCS#7 enveloped data properly.
17536 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17538 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17539 copying pointers. The cert_st handling is changed by this in
17540 various ways (and thus what used to be known as ctx->default_cert
17541 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17542 any longer when s->cert does not give us what we need).
17543 ssl_cert_instantiate becomes obsolete by this change.
17544 As soon as we've got the new code right (possibly it already is?),
17545 we have solved a couple of bugs of the earlier code where s->cert
17546 was used as if it could not have been shared with other SSL structures.
17548 Note that using the SSL API in certain dirty ways now will result
17549 in different behaviour than observed with earlier library versions:
17550 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17551 does not influence s as it used to.
17553 In order to clean up things more thoroughly, inside SSL_SESSION
17554 we don't use CERT any longer, but a new structure SESS_CERT
17555 that holds per-session data (if available); currently, this is
17556 the peer's certificate chain and, for clients, the server's certificate
17557 and temporary key. CERT holds only those values that can have
17558 meaningful defaults in an SSL_CTX.
17562 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17563 from the internal representation. Various PKCS#7 fixes: remove some
17564 evil casts and set the enc_dig_alg field properly based on the signing
17569 * Allow PKCS#12 password to be set from the command line or the
17570 environment. Let 'ca' get its config file name from the environment
17571 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17576 * Allow certificate policies extension to use an IA5STRING for the
17577 organization field. This is contrary to the PKIX definition but
17578 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17583 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17584 without disallowing inline assembler and the like for non-pedantic builds.
17588 * Support Borland C++ builder.
17590 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17596 * SHA-1 cleanups and performance enhancements.
17598 *Andy Polyakov <appro@fy.chalmers.se>*
17600 * Sparc v8plus assembler for the bignum library.
17602 *Andy Polyakov <appro@fy.chalmers.se>*
17604 * Accept any -xxx and +xxx compiler options in Configure.
17608 * Update HPUX configuration.
17612 * Add missing `sk_<type>_unshift()` function to safestack.h
17614 *Ralf S. Engelschall*
17616 * New function SSL_CTX_use_certificate_chain_file that sets the
17617 "extra_cert"s in addition to the certificate. (This makes sense
17618 only for "PEM" format files, as chains as a whole are not
17623 * Support verify_depth from the SSL API.
17624 x509_vfy.c had what can be considered an off-by-one-error:
17625 Its depth (which was not part of the external interface)
17626 was actually counting the number of certificates in a chain;
17627 now it really counts the depth.
17631 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17632 instead of X509err, which often resulted in confusing error
17633 messages since the error codes are not globally unique
17634 (e.g. an alleged error in ssl3_accept when a certificate
17635 didn't match the private key).
17637 * New function SSL_CTX_set_session_id_context that allows to set a default
17638 value (so that you don't need SSL_set_session_id_context for each
17639 connection using the SSL_CTX).
17643 * OAEP decoding bug fix.
17647 * Support INSTALL_PREFIX for package builders, as proposed by
17652 * New Configure options "threads" and "no-threads". For systems
17653 where the proper compiler options are known (currently Solaris
17654 and Linux), "threads" is the default.
17658 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17662 * Install various scripts to $(OPENSSLDIR)/misc, not to
17663 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17664 such as /usr/local/bin.
17668 * "make linux-shared" to build shared libraries.
17670 *Niels Poppe <niels@netbox.org>*
17672 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17676 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17677 extension adding in x509 utility.
17681 * Remove NOPROTO sections and error code comments.
17685 * Partial rewrite of the DEF file generator to now parse the ANSI
17690 * New Configure options --prefix=DIR and --openssldir=DIR.
17694 * Complete rewrite of the error code script(s). It is all now handled
17695 by one script at the top level which handles error code gathering,
17696 header rewriting and C source file generation. It should be much better
17697 than the old method: it now uses a modified version of Ulf's parser to
17698 read the ANSI prototypes in all header files (thus the old K&R definitions
17699 aren't needed for error creation any more) and do a better job of
17700 translating function codes into names. The old 'ASN1 error code embedded
17701 in a comment' is no longer necessary and it doesn't use .err files which
17702 have now been deleted. Also the error code call doesn't have to appear all
17703 on one line (which resulted in some large lines...).
17707 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17711 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17712 0 (which usually indicates a closed connection), but continue reading.
17716 * Fix some race conditions.
17720 * Add support for CRL distribution points extension. Add Certificate
17721 Policies and CRL distribution points documentation.
17725 * Move the autogenerated header file parts to crypto/opensslconf.h.
17729 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17730 8 of keying material. Merlin has also confirmed interop with this fix
17731 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17733 *Merlin Hughes <merlin@baltimore.ie>*
17735 * Fix lots of warnings.
17737 *Richard Levitte <levitte@stacken.kth.se>*
17739 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17740 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17742 *Richard Levitte <levitte@stacken.kth.se>*
17744 * Fix problems with sizeof(long) == 8.
17746 *Andy Polyakov <appro@fy.chalmers.se>*
17748 * Change functions to ANSI C.
17752 * Fix typos in error codes.
17754 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17756 * Remove defunct assembler files from Configure.
17760 * SPARC v8 assembler BIGNUM implementation.
17762 *Andy Polyakov <appro@fy.chalmers.se>*
17764 * Support for Certificate Policies extension: both print and set.
17765 Various additions to support the r2i method this uses.
17769 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17770 return a const string when you are expecting an allocated buffer.
17774 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17775 types DirectoryString and DisplayText.
17779 * Add code to allow r2i extensions to access the configuration database,
17780 add an LHASH database driver and add several ctx helper functions.
17784 * Fix an evil bug in bn_expand2() which caused various BN functions to
17785 fail when they extended the size of a BIGNUM.
17789 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17790 support typesafe stack.
17794 * Fix typo in SSL_[gs]et_options().
17796 *Nils Frostberg <nils@medcom.se>*
17798 * Delete various functions and files that belonged to the (now obsolete)
17799 old X509V3 handling code.
17803 * New Configure option "rsaref".
17807 * Don't auto-generate pem.h.
17811 * Introduce type-safe ASN.1 SETs.
17815 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17817 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17819 * Introduce type-safe STACKs. This will almost certainly break lots of code
17820 that links with OpenSSL (well at least cause lots of warnings), but fear
17821 not: the conversion is trivial, and it eliminates loads of evil casts. A
17822 few STACKed things have been converted already. Feel free to convert more.
17823 In the fullness of time, I'll do away with the STACK type altogether.
17827 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17828 specified in `<certfile>` by updating the entry in the index.txt file.
17829 This way one no longer has to edit the index.txt file manually for
17830 revoking a certificate. The -revoke option does the gory details now.
17832 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17834 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17835 `-text` option at all and this way the `-noout -text` combination was
17836 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17838 *Ralf S. Engelschall*
17840 * Make sure a corresponding plain text error message exists for the
17841 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17842 verify callback function determined that a certificate was revoked.
17844 *Ralf S. Engelschall*
17846 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17847 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17848 all available ciphers including rc5, which was forgotten until now.
17849 In order to let the testing shell script know which algorithms
17850 are available, a new (up to now undocumented) command
17851 `openssl list-cipher-commands` is used.
17855 * Bugfix: s_client occasionally would sleep in select() when
17856 it should have checked SSL_pending() first.
17860 * New functions DSA_do_sign and DSA_do_verify to provide access to
17861 the raw DSA values prior to ASN.1 encoding.
17865 * Tweaks to Configure
17867 *Niels Poppe <niels@netbox.org>*
17869 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17874 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17878 * New config option to avoid instructions that are illegal on the 80386.
17879 The default code is faster, but requires at least a 486.
17883 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17884 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17885 same as SSL2_VERSION anyway.
17889 * New "-showcerts" option for s_client.
17893 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17894 application. Various cleanups and fixes.
17898 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17899 modify error routines to work internally. Add error codes and PBE init
17900 to library startup routines.
17904 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17905 packing functions to asn1 and evp. Changed function names and error
17906 codes along the way.
17910 * PKCS12 integration: and so it begins... First of several patches to
17911 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17912 objects to objects.h
17916 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17917 and display support for Thawte strong extranet extension.
17921 * Add LinuxPPC support.
17923 *Jeff Dubrule <igor@pobox.org>*
17925 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17926 bn_div_words in alpha.s.
17928 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17930 * Make sure the RSA OAEP test is skipped under -DRSAref because
17931 OAEP isn't supported when OpenSSL is built with RSAref.
17933 *Ulf Moeller <ulf@fitug.de>*
17935 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17936 so they no longer are missing under -DNOPROTO.
17938 *Soren S. Jorvang <soren@t.dk>*
17940 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17942 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17943 doesn't work when the session is reused. Coming soon!
17947 * Fix a security hole, that allows sessions to be reused in the wrong
17948 context thus bypassing client cert protection! All software that uses
17949 client certs and session caches in multiple contexts NEEDS PATCHING to
17950 allow session reuse! A fuller solution is in the works.
17952 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17954 * Some more source tree cleanups (removed obsolete files
17955 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17956 permission on "config" script to be executable) and a fix for the INSTALL
17959 *Ulf Moeller <ulf@fitug.de>*
17961 * Remove some legacy and erroneous uses of malloc, free instead of
17964 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17966 * Make rsa_oaep_test return non-zero on error.
17968 *Ulf Moeller <ulf@fitug.de>*
17970 * Add support for native Solaris shared libraries. Configure
17971 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17972 if someone would make that last step automatic.
17974 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17976 * ctx_size was not built with the right compiler during "make links". Fixed.
17980 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17981 except NULL ciphers". This means the default cipher list will no longer
17982 enable NULL ciphers. They need to be specifically enabled e.g. with
17983 the string "DEFAULT:eNULL".
17987 * Fix to RSA private encryption routines: if p < q then it would
17988 occasionally produce an invalid result. This will only happen with
17989 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17993 * Be less restrictive and allow also `perl util/perlpath.pl
17994 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17995 because this way one can also use an interpreter named `perl5` (which is
17996 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17997 installed as `perl`).
17999 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18001 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18003 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18005 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18006 advapi32.lib to Win32 build and change the pem test comparison
18007 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18008 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18009 and crypto/des/ede_cbcm_enc.c.
18013 * DES quad checksum was broken on big-endian architectures. Fixed.
18017 * Comment out two functions in bio.h that aren't implemented. Fix up the
18018 Win32 test batch file so it (might) work again. The Win32 test batch file
18019 is horrible: I feel ill....
18023 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18024 in e_os.h. Audit of header files to check ANSI and non ANSI
18025 sections: 10 functions were absent from non ANSI section and not exported
18026 from Windows DLLs. Fixed up libeay.num for new functions.
18030 * Make `openssl version` output lines consistent.
18032 *Ralf S. Engelschall*
18034 * Fix Win32 symbol export lists for BIO functions: Added
18035 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18036 to ms/libeay{16,32}.def.
18038 *Ralf S. Engelschall*
18040 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18041 fine under Unix and passes some trivial tests I've now added. But the
18042 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18043 added to make sure no one expects that this stuff really works in the
18044 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18045 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18048 *Ralf S. Engelschall*
18050 * Fix the generation of two part addresses in perl.
18052 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18054 * Add config entry for Linux on MIPS.
18056 *John Tobey <jtobey@channel1.com>*
18058 * Make links whenever Configure is run, unless we are on Windoze.
18062 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18063 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18068 * Add a useful kludge to allow package maintainers to specify compiler and
18069 other platforms details on the command line without having to patch the
18070 Configure script every time: One now can use
18071 `perl Configure <id>:<details>`,
18072 i.e. platform ids are allowed to have details appended
18073 to them (separated by colons). This is treated as there would be a static
18074 pre-configured entry in Configure's %table under key `<id>` with value
18075 `<details>` and `perl Configure <id>` is called. So, when you want to
18076 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18077 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18078 now, which overrides the FreeBSD-elf entry on-the-fly.
18080 *Ralf S. Engelschall*
18082 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18086 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18087 on the `perl Configure ...` command line. This way one can compile
18088 OpenSSL libraries with Position Independent Code (PIC) which is needed
18089 for linking it into DSOs.
18091 *Ralf S. Engelschall*
18093 * Remarkably, export ciphers were totally broken and no-one had noticed!
18098 * Cleaned up the LICENSE document: The official contact for any license
18099 questions now is the OpenSSL core team under openssl-core@openssl.org.
18100 And add a paragraph about the dual-license situation to make sure people
18101 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18102 to the OpenSSL toolkit.
18104 *Ralf S. Engelschall*
18106 * General source tree makefile cleanups: Made `making xxx in yyy...`
18107 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18108 Additionally cleaned up the `make links` target: Remove unnecessary
18109 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18110 to speed processing and no longer clutter the display with confusing
18111 stuff. Instead only the actually done links are displayed.
18113 *Ralf S. Engelschall*
18115 * Permit null encryption ciphersuites, used for authentication only. It used
18116 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18117 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18122 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18123 signed attributes when verifying signatures (this would break them),
18124 the detached data encoding was wrong and public keys obtained using
18125 X509_get_pubkey() weren't freed.
18129 * Add text documentation for the BUFFER functions. Also added a work around
18130 to a Win95 console bug. This was triggered by the password read stuff: the
18131 last character typed gets carried over to the next fread(). If you were
18132 generating a new cert request using 'req' for example then the last
18133 character of the passphrase would be CR which would then enter the first
18138 * Added the new 'Includes OpenSSL Cryptography Software' button as
18139 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18140 button and can be used by applications based on OpenSSL to show the
18141 relationship to the OpenSSL project.
18143 *Ralf S. Engelschall*
18145 * Remove confusing variables in function signatures in files
18146 ssl/ssl_lib.c and ssl/ssl.h.
18148 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18150 * Don't install bss_file.c under PREFIX/include/
18152 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18154 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18155 functions that return function pointers and has support for NT specific
18156 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18157 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18158 unsigned to signed types: this was killing the Win32 compile.
18162 * Add new certificate file to stack functions,
18163 SSL_add_dir_cert_subjects_to_stack() and
18164 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18165 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18166 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18167 This means that Apache-SSL and similar packages don't have to mess around
18168 to add as many CAs as they want to the preferred list.
18172 * Experiment with doxygen documentation. Currently only partially applied to
18174 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18175 openssl.doxy as the configuration file.
18179 * Get rid of remaining C++-style comments which strict C compilers hate.
18181 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18183 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18184 compiled in by default: it has problems with large keys.
18188 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18189 DH private keys and/or callback functions which directly correspond to
18190 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18191 is needed for applications which have to configure certificates on a
18192 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18194 For the RSA certificate situation is makes no difference, but
18195 for the DSA certificate situation this fixes the "no shared cipher"
18196 problem where the OpenSSL cipher selection procedure failed because the
18197 temporary keys were not overtaken from the context and the API provided
18198 no way to reconfigure them.
18199 The new functions now let applications reconfigure the stuff and they
18200 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18201 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18202 non-public-API function ssl_cert_instantiate() is used as a helper
18203 function and also to reduce code redundancy inside ssl_rsa.c.
18205 *Ralf S. Engelschall*
18207 * Move s_server -dcert and -dkey options out of the undocumented feature
18208 area because they are useful for the DSA situation and should be
18209 recognized by the users.
18211 *Ralf S. Engelschall*
18213 * Fix the cipher decision scheme for export ciphers: the export bits are
18214 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18215 SSL_EXP_MASK. So, the original variable has to be used instead of the
18216 already masked variable.
18218 *Richard Levitte <levitte@stacken.kth.se>*
18220 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18222 *Richard Levitte <levitte@stacken.kth.se>*
18224 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18225 from `int` to `unsigned int` because it is a length and initialized by
18226 EVP_DigestFinal() which expects an `unsigned int *`.
18228 *Richard Levitte <levitte@stacken.kth.se>*
18230 * Don't hard-code path to Perl interpreter on shebang line of Configure
18231 script. Instead use the usual Shell->Perl transition trick.
18233 *Ralf S. Engelschall*
18235 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18236 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18237 -noout -modulus` as it's already the case for `openssl rsa -noout
18238 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18239 currently the public key is printed (a decision which was already done by
18240 `openssl dsa -modulus` in the past) which serves a similar purpose.
18241 Additionally the NO_RSA no longer completely removes the whole -modulus
18242 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18245 *Ralf S. Engelschall*
18247 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18248 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18250 *Arne Ansper <arne@ats.cyber.ee>*
18252 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18253 to be added. Now both 'req' and 'ca' can use new objects defined in the
18258 * Add cool BIO that does syslog (or event log on NT).
18260 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18262 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18263 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18264 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18265 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18269 * Add preliminary config info for new extension code.
18273 * Make RSA_NO_PADDING really use no padding.
18275 *Ulf Moeller <ulf@fitug.de>*
18277 * Generate errors when private/public key check is done.
18281 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18282 for some CRL extensions and new objects added.
18286 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18287 key usage extension and fuller support for authority key id.
18291 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18292 padding method for RSA, which is recommended for new applications in PKCS
18293 #1 v2.0 (RFC 2437, October 1998).
18294 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18295 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18296 against Bleichbacher's attack on RSA.
18297 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18300 * Updates to the new SSL compression code
18302 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18304 * Fix so that the version number in the master secret, when passed
18305 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18306 (because the server will not accept higher), that the version number
18307 is 0x03,0x01, not 0x03,0x00
18309 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18311 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18312 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18313 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18317 * Support for RAW extensions where an arbitrary extension can be
18318 created by including its DER encoding. See `apps/openssl.cnf` for
18323 * Make sure latest Perl versions don't interpret some generated C array
18324 code as Perl array code in the crypto/err/err_genc.pl script.
18326 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18328 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18329 not many people have the assembler. Various Win32 compilation fixes and
18330 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18331 build instructions.
18335 * Modify configure script 'Configure' to automatically create crypto/date.h
18336 file under Win32 and also build pem.h from pem.org. New script
18337 util/mkfiles.pl to create the MINFO file on environments that can't do a
18338 'make files': perl util/mkfiles.pl >MINFO should work.
18342 * Major rework of DES function declarations, in the pursuit of correctness
18343 and purity. As a result, many evil casts evaporated, and some weirdness,
18344 too. You may find this causes warnings in your code. Zapping your evil
18345 casts will probably fix them. Mostly.
18349 * Fix for a typo in asn1.h. Bug fix to object creation script
18350 obj_dat.pl. It considered a zero in an object definition to mean
18351 "end of object": none of the objects in objects.h have any zeros
18352 so it wasn't spotted.
18354 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18356 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18357 Masking (CBCM). In the absence of test vectors, the best I have been able
18358 to do is check that the decrypt undoes the encrypt, so far. Send me test
18359 vectors if you have them.
18363 * Correct calculation of key length for export ciphers (too much space was
18364 allocated for null ciphers). This has not been tested!
18368 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18369 message is now correct (it understands "crypto" and "ssl" on its
18370 command line). There is also now an "update" option. This will update
18371 the util/ssleay.num and util/libeay.num files with any new functions.
18373 perl util/mkdef.pl crypto ssl update
18374 it will update them.
18378 * Overhauled the Perl interface:
18379 - ported BN stuff to OpenSSL's different BN library
18380 - made the perl/ source tree CVS-aware
18381 - renamed the package from SSLeay to OpenSSL (the files still contain
18382 their history because I've copied them in the repository)
18383 - removed obsolete files (the test scripts will be replaced
18384 by better Test::Harness variants in the future)
18386 *Ralf S. Engelschall*
18388 * First cut for a very conservative source tree cleanup:
18389 1. merge various obsolete readme texts into doc/ssleay.txt
18390 where we collect the old documents and readme texts.
18391 2. remove the first part of files where I'm already sure that we no
18392 longer need them because of three reasons: either they are just temporary
18393 files which were left by Eric or they are preserved original files where
18394 I've verified that the diff is also available in the CVS via "cvs diff
18395 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18396 the crypto/md/ stuff).
18398 *Ralf S. Engelschall*
18400 * More extension code. Incomplete support for subject and issuer alt
18401 name, issuer and authority key id. Change the i2v function parameters
18402 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18403 what that's for :-) Fix to ASN1 macro which messed up
18404 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18408 * Preliminary support for ENUMERATED type. This is largely copied from the
18413 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18415 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18417 * Make sure `make rehash` target really finds the `openssl` program.
18419 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18421 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18422 like to hear about it if this slows down other processors.
18426 * Add CygWin32 platform information to Configure script.
18428 *Alan Batie <batie@aahz.jf.intel.com>*
18430 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18432 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18434 * New program nseq to manipulate netscape certificate sequences
18438 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18443 * Fixes to BN code. Previously the default was to define BN_RECURSION
18444 but the BN code had some problems that would cause failures when
18445 doing certificate verification and some other functions.
18447 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18449 * Add ASN1 and PEM code to support netscape certificate sequences.
18453 * Add ASN1 and PEM code to support netscape certificate sequences.
18457 * Add several PKIX and private extended key usage OIDs.
18461 * Modify the 'ca' program to handle the new extension code. Modify
18462 openssl.cnf for new extension format, add comments.
18466 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18467 and add a sample to openssl.cnf so req -x509 now adds appropriate
18472 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18473 error code, add initial support to X509_print() and x509 application.
18477 * Takes a deep breath and start adding X509 V3 extension support code. Add
18478 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18479 stuff is currently isolated and isn't even compiled yet.
18483 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18484 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18485 Removed the versions check from X509 routines when loading extensions:
18486 this allows certain broken certificates that don't set the version
18487 properly to be processed.
18491 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18492 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18493 can still be regenerated with "make depend".
18497 * Spelling mistake in C version of CAST-128.
18499 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18501 * Changes to the error generation code. The perl script err-code.pl
18502 now reads in the old error codes and retains the old numbers, only
18503 adding new ones if necessary. It also only changes the .err files if new
18504 codes are added. The makefiles have been modified to only insert errors
18505 when needed (to avoid needlessly modifying header files). This is done
18506 by only inserting errors if the .err file is newer than the auto generated
18507 C file. To rebuild all the error codes from scratch (the old behaviour)
18508 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18509 or delete all the .err files.
18513 * CAST-128 was incorrectly implemented for short keys. The C version has
18514 been fixed, but is untested. The assembler versions are also fixed, but
18515 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18516 to regenerate it if needed.
18517 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18518 Hagino <itojun@kame.net>*
18520 * File was opened incorrectly in randfile.c.
18522 *Ulf Möller <ulf@fitug.de>*
18524 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18525 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18526 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18527 al: it's just almost always a UTCTime. Note this patch adds new error
18528 codes so do a "make errors" if there are problems.
18532 * Correct Linux 1 recognition in config.
18534 *Ulf Möller <ulf@fitug.de>*
18536 * Remove pointless MD5 hash when using DSA keys in ca.
18538 *Anonymous <nobody@replay.com>*
18540 * Generate an error if given an empty string as a cert directory. Also
18541 generate an error if handed NULL (previously returned 0 to indicate an
18542 error, but didn't set one).
18544 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18546 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18550 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18551 parameters. This was causing a warning which killed off the Win32 compile.
18555 * Remove C++ style comments from crypto/bn/bn_local.h.
18557 *Neil Costigan <neil.costigan@celocom.com>*
18559 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18560 based on a text string, looking up short and long names and finally
18561 "dot" format. The "dot" format stuff didn't work. Added new function
18562 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18563 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18564 OID is not part of the table.
18568 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18569 X509_LOOKUP_by_alias().
18573 * Sort openssl functions by name.
18577 * Get the `gendsa` command working and add it to the `list` command. Remove
18578 encryption from sample DSA keys (in case anyone is interested the password
18583 * Make *all* `*_free` functions accept a NULL pointer.
18585 *Frans Heymans <fheymans@isaserver.be>*
18587 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18590 *Anonymous <nobody@replay.com>*
18592 * s_server should send the CAfile as acceptable CAs, not its own cert.
18594 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18596 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18598 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18600 * Temp key "for export" tests were wrong in s3_srvr.c.
18602 *Anonymous <nobody@replay.com>*
18604 * Add prototype for temp key callback functions
18605 SSL_CTX_set_tmp_{rsa,dh}_callback().
18609 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18610 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18614 * X509_name_add_entry() freed the wrong thing after an error.
18616 *Arne Ansper <arne@ats.cyber.ee>*
18618 * rsa_eay.c would attempt to free a NULL context.
18620 *Arne Ansper <arne@ats.cyber.ee>*
18622 * BIO_s_socket() had a broken should_retry() on Windoze.
18624 *Arne Ansper <arne@ats.cyber.ee>*
18626 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18628 *Arne Ansper <arne@ats.cyber.ee>*
18630 * Make sure the already existing X509_STORE->depth variable is initialized
18631 in X509_STORE_new(), but document the fact that this variable is still
18632 unused in the certificate verification process.
18634 *Ralf S. Engelschall*
18636 * Fix the various library and `apps/` files to free up pkeys obtained from
18637 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18641 * Fix reference counting in X509_PUBKEY_get(). This makes
18642 demos/maurice/example2.c work, amongst others, probably.
18644 *Steve Henson and Ben Laurie*
18646 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18647 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18648 are no longer created. This way we have a single and consistent command
18649 line interface `openssl <command>`, similar to `cvs <command>`.
18651 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18653 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18654 BIT STRING wrapper always have zero unused bits.
18658 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18662 * Make the top-level INSTALL documentation easier to understand.
18666 * Makefiles updated to exit if an error occurs in a sub-directory
18667 make (including if user presses ^C) [Paul Sutton]
18669 * Make Montgomery context stuff explicit in RSA data structure.
18673 * Fix build order of pem and err to allow for generated pem.h.
18677 * Fix renumbering bug in X509_NAME_delete_entry().
18681 * Enhanced the err-ins.pl script so it makes the error library number
18682 global and can add a library name. This is needed for external ASN1 and
18683 other error libraries.
18687 * Fixed sk_insert which never worked properly.
18691 * Fix ASN1 macros so they can handle indefinite length constructed
18692 EXPLICIT tags. Some non standard certificates use these: they can now
18697 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18698 into a single doc/ssleay.txt bundle. This way the information is still
18699 preserved but no longer messes up this directory. Now it's new room for
18700 the new set of documentation files.
18702 *Ralf S. Engelschall*
18704 * SETs were incorrectly DER encoded. This was a major pain, because they
18705 shared code with SEQUENCEs, which aren't coded the same. This means that
18706 almost everything to do with SETs or SEQUENCEs has either changed name or
18707 number of arguments.
18709 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18711 * Fix test data to work with the above.
18715 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18716 was already fixed by Eric for 0.9.1 it seems.
18718 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18720 * Autodetect FreeBSD3.
18724 * Fix various bugs in Configure. This affects the following platforms:
18728 unixware-2.0-pentium
18733 * Eliminate generated files from CVS. Reorder tests to regenerate files
18734 before they are needed.
18738 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18742 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18744 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18745 changed SSLeay to OpenSSL in version strings.
18747 *Ralf S. Engelschall*
18749 * Some fixups to the top-level documents.
18753 * Fixed the nasty bug where rsaref.h was not found under compile-time
18754 because the symlink to include/ was missing.
18756 *Ralf S. Engelschall*
18758 * Incorporated the popular no-RSA/DSA-only patches
18759 which allow to compile a RSA-free SSLeay.
18761 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18763 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18764 when "ssleay" is still not found.
18766 *Ralf S. Engelschall*
18768 * Added more platforms to Configure: Cray T3E, HPUX 11,
18770 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18772 * Updated the README file.
18774 *Ralf S. Engelschall*
18776 * Added various .cvsignore files in the CVS repository subdirs
18777 to make a "cvs update" really silent.
18779 *Ralf S. Engelschall*
18781 * Recompiled the error-definition header files and added
18782 missing symbols to the Win32 linker tables.
18784 *Ralf S. Engelschall*
18786 * Cleaned up the top-level documents;
18787 o new files: CHANGES and LICENSE
18788 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18789 o merged COPYRIGHT into LICENSE
18790 o removed obsolete TODO file
18791 o renamed MICROSOFT to INSTALL.W32
18793 *Ralf S. Engelschall*
18795 * Removed dummy files from the 0.9.1b source tree:
18797 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18798 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18799 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18800 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18801 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18804 *Ralf S. Engelschall*
18806 * Added various platform portability fixes.
18810 * The Genesis of the OpenSSL rpject:
18811 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18812 Young and Tim J. Hudson created while they were working for C2Net until
18815 *The OpenSSL Project*
18817 ### Changes between 0.9.0b and 0.9.1b [not released]
18819 * Updated a few CA certificates under certs/
18823 * Changed some BIGNUM api stuff.
18827 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18828 DGUX x86, Linux Alpha, etc.
18832 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18833 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18838 * Add -strparse option to asn1pars program which parses nested
18841 *Dr Stephen Henson <shenson@bigfoot.com>*
18843 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18847 * DSA fix for "ca" program.
18851 * Added "-genkey" option to "dsaparam" program.
18855 * Added RIPE MD160 (rmd160) message digest.
18859 * Added -a (all) option to "ssleay version" command.
18863 * Added PLATFORM define which is the id given to Configure.
18867 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18871 * Extended the ASN.1 parser routines.
18875 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18879 * Added a BN_CTX to the BN library.
18883 * Fixed the weak key values in DES library
18887 * Changed API in EVP library for cipher aliases.
18891 * Added support for RC2/64bit cipher.
18895 * Converted the lhash library to the crypto/mem.c functions.
18899 * Added more recognized ASN.1 object ids.
18903 * Added more RSA padding checks for SSL/TLS.
18907 * Added BIO proxy/filter functionality.
18911 * Added extra_certs to SSL_CTX which can be used
18912 send extra CA certificates to the client in the CA cert chain sending
18913 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18917 * Now Fortezza is denied in the authentication phase because
18918 this is key exchange mechanism is not supported by SSLeay at all.
18922 * Additional PKCS1 checks.
18926 * Support the string "TLSv1" for all TLS v1 ciphers.
18930 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18931 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18935 * Fixed a few memory leaks.
18939 * Fixed various code and comment typos.
18943 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18944 bytes sent in the client random.
18946 *Edward Bishop <ebishop@spyglass.com>*
18950 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18951 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18952 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18953 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18954 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18955 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18956 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18957 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18958 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18959 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18960 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18961 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18962 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18963 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18964 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18965 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18966 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18967 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18968 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18969 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18970 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18971 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18972 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18973 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18974 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18975 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18976 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18977 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18978 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18979 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18980 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18981 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18982 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18983 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18984 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18985 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18986 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18987 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18988 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18989 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18990 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18991 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18992 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18993 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18994 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18995 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18996 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18997 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18998 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18999 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19000 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19001 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19002 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19003 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19004 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19005 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19006 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19007 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19008 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19009 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19010 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19011 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19012 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19013 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19014 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19015 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19016 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19017 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19018 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19019 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19020 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19021 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19022 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19023 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19024 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19025 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19026 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19027 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19028 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19029 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19030 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19031 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19032 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19033 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19034 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19035 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19036 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19037 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19038 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19039 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19040 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19041 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19042 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19043 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19044 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19045 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19046 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19047 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19048 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19049 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19050 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19051 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19052 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19053 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19054 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19055 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19056 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19057 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19058 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19059 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19060 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19061 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19062 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19063 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19064 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19065 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19066 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19067 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19068 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19069 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19070 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19071 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19072 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19073 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19074 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19075 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19076 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19077 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19078 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19079 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19080 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19081 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19082 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19083 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19084 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19085 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19086 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19087 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19088 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19089 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19090 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19091 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19092 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19093 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19094 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19095 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19096 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19097 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19098 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19099 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19100 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19101 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19102 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19103 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19104 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19105 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19106 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19107 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19108 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19109 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19110 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19111 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655