4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * The deprecated functions EVP_PKEY_get0(), EVP_PKEY_get0_RSA(),
27 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
28 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
29 well as the similarly named "get1" functions behave slightly differently in
30 OpenSSL 3.0. Previously they returned a pointer to the low-level key used
31 internally by libcrypto. From OpenSSL 3.0 this key may now be held in a
32 provider. Calling these functions will only return a handle on the internal
33 key where the EVP_PKEY was constructed using this key in the first place, for
34 example using a function or macro such as EVP_PKEY_assign_RSA(),
35 EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key,
36 then these functions now return a cached copy of the key. Changes to
37 the internal provider key that take place after the first time the cached key
38 is accessed will not be reflected back in the cached copy. Similarly any
39 changes made to the cached copy by application code will not be reflected
40 back in the internal provider key.
42 For the above reasons the keys returned from these functions should typically
43 be treated as read-only. To emphasise this the value returned from
44 EVP_PKEY_get0(), EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(),
45 EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may
46 break some existing code. Applications broken by this change should be
47 modified. The preferred solution is to refactor the code to avoid the use of
48 these deprecated functions. Failing this the code should be modified to use a
49 const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(),
50 EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a
51 non-const pointer to enable them to be "freed". However they should also be
56 * A number of functions handling low level keys or engines were deprecated
57 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
58 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
59 EVP_PKEY_get0_siphash(). Applications using engines should instead use
60 providers. Applications getting or setting low level keys in an EVP_PKEY
61 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
62 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
66 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
67 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
68 and require returning octet ptr parameters from providers that
69 would like to support them which complicates provider implementations.
73 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
74 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
75 RAND_set_rand_method(). Provider based random number generators should
76 be used instead via EVP_RAND(3).
80 * The SRP APIs have been deprecated. The old APIs do not work via providers,
81 and there is no EVP interface to them. Unfortunately there is no replacement
82 for these APIs at this time.
86 * Add a compile time option to prevent the caching of provider fetched
87 algorithms. This is enabled by including the no-cached-fetch option
88 at configuration time.
92 * The openssl speed command does not use low-level API calls anymore. This
93 implies some of the performance numbers might not be fully comparable
94 with the previous releases due to higher overhead. This applies
95 particularly to measuring performance on smaller data chunks.
99 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
100 Typically if OpenSSL has no EC or DH algorithms then it cannot support
101 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
102 through providers. Therefore third party providers may supply group
103 implementations even where there are no built-in ones. Attempting to create
104 TLS connections in such a build without also disabling TLSv1.3 at run time or
105 using third party provider groups may result in handshake failures. TLSv1.3
106 can be disabled at compile time using the "no-tls1_3" Configure option.
110 * The undocumented function X509_certificate_type() has been deprecated;
111 applications can use X509_get0_pubkey() and X509_get0_signature() to
112 get the same information.
116 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
117 functions. They are identical to BN_rand() and BN_rand_range()
122 * Removed RSA padding mode for SSLv23 (which was only used for
123 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
124 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
129 * Deprecated the obsolete X9.31 RSA key generation related functions
130 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
131 BN_X931_generate_prime_ex().
135 * The default key generation method for the regular 2-prime RSA keys was
136 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
137 Conditions Based on Auxiliary Probable Primes). This method is slower
138 than the original method.
142 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
143 They are replaced with the BN_check_prime() function that avoids possible
144 misuse and always uses at least 64 rounds of the Miller-Rabin
145 primality test. At least 64 rounds of the Miller-Rabin test are now also
146 used for all prime generation, including RSA key generation.
147 This increases key generation time, especially for larger keys.
151 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
152 as they are not useful with non-deprecated functions.
156 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
157 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
158 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
159 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
160 were used to collect all necessary data to form a HTTP request, and to
161 perform the HTTP transfer with that request. With OpenSSL 3.0, the
162 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
163 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
164 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
165 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
166 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
167 OSSL_HTTP_REQ_CTX_set_max_response_length().
169 *Rich Salz and Richard Levitte*
171 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
175 * Validation of SM2 keys has been separated from the validation of regular EC
176 keys, allowing to improve the SM2 validation process to reject loaded private
177 keys that are not conforming to the SM2 ISO standard.
178 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
183 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
184 switches: a validation failure triggers an early exit, returning a failure
185 exit status to the parent process.
189 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
190 to ignore unknown ciphers.
194 * The `-cipher-commands` and `-digest-commands` options
195 of the command line utility `list` have been deprecated.
196 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
200 * All of the low level EC_KEY functions have been deprecated including:
202 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
203 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
204 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
205 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
206 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
207 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
208 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
209 EC_KEY_METHOD_get_verify,
210 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
211 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
212 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
213 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
214 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
215 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
216 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
217 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
218 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
219 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
220 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
221 Applications that need to implement an EC_KEY_METHOD need to consider
222 implementation of the functionality in a special provider.
223 For replacement of the functions manipulating the EC_KEY objects
224 see the EVP_PKEY-EC(7) manual page.
226 Additionally functions that read and write EC_KEY objects such as
227 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
228 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
229 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
230 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
231 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
232 have also been deprecated. Applications should instead use the
233 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
235 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
236 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
237 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
238 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
239 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
241 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
243 * Deprecated all the libcrypto and libssl error string loading
244 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
245 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
246 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
247 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
248 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
249 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
250 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
251 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
252 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
253 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
255 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
256 now loads error strings automatically.
260 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
261 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
262 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
263 are to be used by servers requiring ephemeral DH keys. Instead applications
264 should consider using the built-in DH parameters that are available by
265 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
266 necessary then applications can use the alternative functions
267 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
268 replacement for the "callback" functions. The callback was originally useful
269 in order to have different parameters for export and non-export ciphersuites.
270 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
271 functions should be replaced by one of the other methods described above.
275 * The `-crypt` option to the `passwd` command line tool has been removed.
279 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
284 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
286 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
287 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
288 The inverse ciphers use AES decryption for wrapping, and
289 AES encryption for unwrapping.
293 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
294 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
295 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
296 OpenSSL 3.0 these are replaced by the more generic functions
297 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
298 The old versions have been converted to deprecated macros that just call the
303 * The security callback, which can be customised by application code, supports
304 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
305 in the "other" parameter. In most places this is what is passed. All these
306 places occur server side. However there was one client side call of this
307 security operation and it passed a DH object instead. This is incorrect
308 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
309 of the other locations. Therefore this client side call has been changed to
310 pass an EVP_PKEY instead.
314 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
315 interface. Their functionality remains unchanged.
319 * Added new option for 'openssl list', '-providers', which will display the
320 list of loaded providers, their names, version and status. It optionally
321 displays their gettable parameters.
325 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
326 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
327 type is internally recognised so the workaround is no longer needed.
329 Functionality is still retained as it is, but will only work with
330 EVP_PKEYs with a legacy internal key.
334 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
335 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
339 * Changed all "STACK" functions to be macros instead of inline functions. Macro
340 parameters are still checked for type safety at compile time via helper
345 * Remove the RAND_DRBG API
347 The RAND_DRBG API did not fit well into the new provider concept as
348 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
349 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
350 and some of its API calls are rather low-level. This holds in particular
351 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
353 Adding a compatibility layer to continue supporting the RAND_DRBG API as
354 a legacy API for a regular deprecation period turned out to come at the
355 price of complicating the new provider API unnecessarily. Since the
356 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
359 *Paul Dale and Matthias St. Pierre*
361 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
362 as well as actual hostnames.
366 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
367 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
368 conversely, silently ignore DTLS protocol version bounds when configuring
369 TLS-based contexts. The commands can be repeated to set bounds of both
370 types. The same applies with the corresponding "min_protocol" and
371 "max_protocol" command-line switches, in case some application uses both TLS
374 SSL_CTX instances that are created for a fixed protocol version (e.g.
375 `TLSv1_server_method()`) also silently ignore version bounds. Previously
376 attempts to apply bounds to these protocol versions would result in an
377 error. Now only the "version-flexible" SSL_CTX instances are subject to
378 limits in configuration files in command-line options.
382 * Deprecated the `ENGINE` API. Engines should be replaced with providers
387 * Reworked the recorded ERR codes to make better space for system errors.
388 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
389 given code is a system error (true) or an OpenSSL error (false).
393 * Reworked the test perl framework to better allow parallel testing.
395 *Nicola Tuveri and David von Oheimb*
397 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
398 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
402 * 'Configure' has been changed to figure out the configuration target if
403 none is given on the command line. Consequently, the 'config' script is
404 now only a mere wrapper. All documentation is changed to only mention
407 *Rich Salz and Richard Levitte*
409 * Added a library context `OSSL_LIB_CTX` that applications as well as
410 other libraries can use to form a separate context within which
411 libcrypto operations are performed.
413 There are two ways this can be used:
415 - Directly, by passing a library context to functions that take
416 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
418 - Indirectly, by creating a new library context and then assigning
419 it as the new default, with `OSSL_LIB_CTX_set0_default`.
421 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
422 apart from the functions directly related to `OSSL_LIB_CTX`, accept
423 NULL to indicate that the default library context should be used.
425 Library code that changes the default library context using
426 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
427 second call before returning to the caller.
429 _(Note: the library context was initially called `OPENSSL_CTX` and
430 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
434 * Handshake now fails if Extended Master Secret extension is dropped
439 * Dropped interactive mode from the `openssl` program. From now on,
440 running it without arguments is equivalent to `openssl help`.
444 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
445 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
446 While the old function names have been retained for backward compatibility
447 they should not be used in new developments
448 because their return values are confusing: Unlike other `_cmp()` functions
449 they do not return 0 in case their arguments are equal.
453 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
454 `EC_GROUP_get_field_type()`.
458 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
459 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
460 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
461 Applications should rely on the library automatically assigning a suitable
462 EC_METHOD internally upon EC_GROUP construction.
466 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
467 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
468 assigned internally without application intervention.
469 Users of EC_GROUP_new() should switch to a different suitable constructor.
473 * Add CAdES-BES signature verification support, mostly derived
474 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
476 *Filipe Raimundo da Silva*
478 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
482 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
483 functions are not widely used and now OpenSSL automatically perform this
484 conversion when needed.
488 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
489 EC_KEY_precompute_mult(). These functions are not widely used and
490 applications should instead switch to named curves which OpenSSL has
491 hardcoded lookup tables for.
495 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
496 should instead use the L<EC_POINT_mul(3)> function.
500 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
501 that are not applicable to the new provider model. Applications should
502 instead use EVP_default_properties_is_fips_enabled() and
503 EVP_default_properties_enable_fips().
507 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
508 is set, an unexpected EOF is ignored, it pretends a close notify was received
509 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
513 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
514 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
515 used and applications should instead use the
516 L<EC_POINT_set_affine_coordinates(3)> and
517 L<EC_POINT_get_affine_coordinates(3)> functions.
521 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
522 arrays to be more easily constructed via a series of utility functions.
523 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
524 the various push functions and finally convert to a passable OSSL_PARAM
525 array using OSSL_PARAM_BLD_to_param().
529 * The security strength of SHA1 and MD5 based signatures in TLS has been
530 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
531 working at the default security level of 1 and instead requires security
532 level 0. The security level can be changed either using the cipher string
533 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
537 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
538 contain a provider side internal key.
542 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
543 They are old functions that we don't use, and that you could disable with
544 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
548 * Project text documents not yet having a proper file name extension
549 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
550 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
554 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
555 have been converted to Markdown with the goal to produce documents
556 which not only look pretty when viewed online in the browser, but
557 remain well readable inside a plain text editor.
559 To achieve this goal, a 'minimalistic' Markdown style has been applied
560 which avoids formatting elements that interfere too much with the
561 reading flow in the text file. For example, it
563 * avoids [ATX headings][] and uses [setext headings][] instead
564 (which works for `<h1>` and `<h2>` headings only).
565 * avoids [inline links][] and uses [reference links][] instead.
566 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
568 [ATX headings]: https://github.github.com/gfm/#atx-headings
569 [setext headings]: https://github.github.com/gfm/#setext-headings
570 [inline links]: https://github.github.com/gfm/#inline-link
571 [reference links]: https://github.github.com/gfm/#reference-link
572 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
573 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
575 *Matthias St. Pierre*
577 * The test suite is changed to preserve results of each test recipe.
578 A new directory test-runs/ with subdirectories named like the
579 test recipes are created in the build tree for this purpose.
583 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
584 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
585 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
587 *David von Oheimb, Martin Peylo*
589 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
590 The legacy OCSP-focused and only partly documented API is retained for
591 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
595 * Added `util/check-format.pl`, a tool for checking adherence to the
596 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
597 The checks performed are incomplete and yield some false positives.
598 Still the tool should be useful for detecting most typical glitches.
602 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
603 If domain name resolution yields multiple IP addresses all of them are tried
604 after `connect()` failures.
608 * All of the low level RSA functions have been deprecated including:
610 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
611 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
612 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
613 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
614 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
615 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
616 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
617 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
618 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
619 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
620 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
621 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
622 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
623 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
624 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
625 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
626 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
627 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
628 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
629 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
630 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
631 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
632 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
633 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
634 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
635 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
636 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
637 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
639 Use of these low level functions has been informally discouraged for a long
640 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
641 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
642 L<EVP_PKEY_decrypt(3)>.
646 * X509 certificates signed using SHA1 are no longer allowed at security
648 In TLS/SSL the default security level is 1. It can be set either
649 using the cipher string with `@SECLEVEL`, or calling
650 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
651 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
653 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
654 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
655 options of the commands.
659 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
660 modified to use PKEY APIs. These commands are now in maintenance mode
661 and no new features will be added to them.
665 * The command line utility rsautl has been deprecated.
666 Instead use the pkeyutl program.
670 * The command line utilities genrsa and rsa have been modified to use PKEY
671 APIs. They now write PKCS#8 keys by default. These commands are now in
672 maintenance mode and no new features will be added to them.
676 * All of the low level DH functions have been deprecated including:
678 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
679 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
680 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
681 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
682 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
683 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
684 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
685 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
686 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
687 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
688 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
689 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
690 DH_meth_get_generate_params and DH_meth_set_generate_params.
692 Use of these low level functions has been informally discouraged for a long
693 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
694 and L<EVP_PKEY_derive(3)>.
696 Additionally functions that read and write DH objects such as d2i_DHparams,
697 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
698 functions have also been deprecated. Applications should instead use the
699 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
701 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
702 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
703 `EVP_PKEY_set1_DH()` are also deprecated.
704 Applications should instead either read or write an
705 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
706 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
708 *Paul Dale and Matt Caswell*
710 * All of the low level DSA functions have been deprecated including:
712 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
713 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
714 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
715 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
716 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
717 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
718 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
719 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
720 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
721 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
722 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
723 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
724 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
725 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
726 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
727 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
728 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
730 Use of these low level functions has been informally discouraged for a long
731 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
732 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
734 Finaly functions that assign or obtain DSA objects from an EVP_PKEY such as
735 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
736 `EVP_PKEY_set1_DSA()` are also deprecated.
737 Applications should instead either read or write an
738 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs,
739 or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
743 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
744 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
745 This means that applications don't have to look at the curve NID and
746 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
747 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
748 a no-op when the EVP_PKEY is already of the given type.
750 Parameter and key generation is also reworked to make it possible
751 to generate EVP_PKEY_SM2 parameters and keys without having to go
752 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
753 However, code that does the latter will still work as before.
757 * Deprecated low level ECDH and ECDSA functions. These include:
759 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
760 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
763 Use of these low level functions has been informally discouraged for a long
764 time. Instead applications should use the EVP_PKEY_derive(3),
765 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
769 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
770 and EVP_PKEY_decrypt() instead.
771 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
772 and EVP_PKEY_encrypt() instead.
776 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
777 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
778 a new formulation to include all the things it can be used for,
779 as well as words of caution.
783 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
784 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
788 * All of the low level HMAC functions have been deprecated including:
790 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
791 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
794 Use of these low level functions has been informally discouraged for a long
795 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
796 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
797 and L<EVP_MAC_final(3)>.
801 * Over two thousand fixes were made to the documentation, including:
802 - Common options (such as -rand/-writerand, TLS version control, etc)
803 were refactored and point to newly-enhanced descriptions in openssl.pod.
804 - Added style conformance for all options (with help from Richard Levitte),
805 documented all reported missing options, added a CI build to check
806 that all options are documented and that no unimplemented options
808 - Documented some internals, such as all use of environment variables.
809 - Addressed all internal broken L<> references.
813 * All of the low level CMAC functions have been deprecated including:
815 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
816 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
818 Use of these low level functions has been informally discouraged for a long
819 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
820 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
821 and L<EVP_MAC_final(3)>.
825 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
826 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
829 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
830 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
831 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
832 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
833 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
834 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
835 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
836 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
837 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
838 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
840 Use of these low level functions has been informally discouraged
841 for a long time. Applications should use the EVP_DigestInit_ex(3),
842 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
846 * Corrected the documentation of the return values from the `EVP_DigestSign*`
847 set of functions. The documentation mentioned negative values for some
848 errors, but this was never the case, so the mention of negative values
851 Code that followed the documentation and thereby check with something
852 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
856 * All of the low level cipher functions have been deprecated including:
858 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
859 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
860 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
861 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
862 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
863 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
864 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
865 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
866 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
867 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
868 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
869 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
870 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
871 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
872 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
873 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
874 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
875 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
876 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
877 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
878 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
879 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
880 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
881 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
882 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
883 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
884 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
885 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
886 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
888 Use of these low level functions has been informally discouraged for
889 a long time. Applications should use the high level EVP APIs, e.g.
890 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
891 equivalently named decrypt functions instead.
893 *Matt Caswell and Paul Dale*
895 * Removed include/openssl/opensslconf.h.in and replaced it with
896 include/openssl/configuration.h.in, which differs in not including
897 <openssl/macros.h>. A short header include/openssl/opensslconf.h
898 was added to include both.
900 This allows internal hacks where one might need to modify the set
901 of configured macros, for example this if deprecated symbols are
902 still supposed to be available internally:
904 #include <openssl/configuration.h>
906 #undef OPENSSL_NO_DEPRECATED
907 #define OPENSSL_SUPPRESS_DEPRECATED
909 #include <openssl/macros.h>
911 This should not be used by applications that use the exported
912 symbols, as that will lead to linking errors.
916 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
917 used in exponentiation with 512-bit moduli. No EC algorithms are
918 affected. Analysis suggests that attacks against 2-prime RSA1024,
919 3-prime RSA1536, and DSA1024 as a result of this defect would be very
920 difficult to perform and are not believed likely. Attacks against DH512
921 are considered just feasible. However, for an attack the target would
922 have to re-use the DH512 private key, which is not recommended anyway.
923 Also applications directly using the low level API BN_mod_exp may be
924 affected if they use BN_FLG_CONSTTIME.
929 * Most memory-debug features have been deprecated, and the functionality
930 replaced with no-ops.
934 * Added documentation for the STACK API.
938 * Introduced a new method type and API, OSSL_ENCODER, to
939 represent generic encoders. An implementation is expected to
940 be able to encode an object associated with a given name (such
941 as an algorithm name for an asymmetric key) into forms given by
942 implementation properties.
944 Encoders are primarily used from inside libcrypto, through
945 calls to functions like EVP_PKEY_print_private(),
946 PEM_write_bio_PrivateKey() and similar.
948 Encoders are specified in such a way that they can be made to
949 directly handle the provider side portion of an object, if this
950 provider side part comes from the same provider as the encoder
951 itself, but can also be made to handle objects in parametrized
952 form (as an OSSL_PARAM array of data). This allows a provider to
953 offer generic encoders as a service for any other provider.
957 * Added a .pragma directive to the syntax of configuration files, to
958 allow varying behavior in a supported and predictable manner.
959 Currently added pragma:
963 This allows dollar signs to be a keyword character unless it's
964 followed by a opening brace or parenthesis. This is useful for
965 platforms where dollar signs are commonly used in names, such as
966 volume names and system directory names on VMS.
970 * Added functionality to create an EVP_PKEY from user data. This
971 is effectively the same as creating a RSA, DH or DSA object and
972 then assigning them to an EVP_PKEY, but directly using algorithm
973 agnostic EVP functions. A benefit is that this should be future
974 proof for public key algorithms to come.
978 * Change the interpretation of the '--api' configuration option to
979 mean that this is a desired API compatibility level with no
980 further meaning. The previous interpretation, that this would
981 also mean to remove all deprecated symbols up to and including
982 the given version, no requires that 'no-deprecated' is also used
983 in the configuration.
985 When building applications, the desired API compatibility level
986 can be set with the OPENSSL_API_COMPAT macro like before. For
987 API compatibility version below 3.0, the old style numerical
988 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
989 For version 3.0 and on, the value is expected to be the decimal
990 value calculated from the major and minor version like this:
992 MAJOR * 10000 + MINOR * 100
996 -DOPENSSL_API_COMPAT=30000 For 3.0
997 -DOPENSSL_API_COMPAT=30200 For 3.2
999 To hide declarations that are deprecated up to and including the
1000 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1001 given when building the application as well.
1005 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1006 access to certificate and CRL stores via URIs and OSSL_STORE
1009 This adds the following functions:
1011 - X509_LOOKUP_store()
1012 - X509_STORE_load_file()
1013 - X509_STORE_load_path()
1014 - X509_STORE_load_store()
1015 - SSL_add_store_cert_subjects_to_stack()
1016 - SSL_CTX_set_default_verify_store()
1017 - SSL_CTX_load_verify_file()
1018 - SSL_CTX_load_verify_dir()
1019 - SSL_CTX_load_verify_store()
1023 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1024 The presence of this system service is determined at run-time.
1028 * Added functionality to create an EVP_PKEY context based on data
1029 for methods from providers. This takes an algorithm name and a
1030 property query string and simply stores them, with the intent
1031 that any operation that uses this context will use those strings
1032 to fetch the needed methods implicitly, thereby making the port
1033 of application written for pre-3.0 OpenSSL easier.
1037 * The undocumented function NCONF_WIN32() has been deprecated; for
1038 conversion details see the HISTORY section of doc/man5/config.pod
1042 * Introduced the new functions EVP_DigestSignInit_ex() and
1043 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1044 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1045 pages for further details.
1049 * Over two thousand fixes were made to the documentation, including:
1050 adding missing command flags, better style conformance, documentation
1053 *Rich Salz, Richard Levitte*
1055 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1056 X25519, X448, Ed25519 and Ed448.
1060 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1065 * Deprecated the public definition of `ERR_STATE` as well as the function
1066 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1071 * Added ERR functionality to give callers access to the stored function
1072 names that have replaced the older function code based functions.
1074 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1075 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1076 ERR_peek_error_all() and ERR_peek_last_error_all().
1078 These functions have become deprecated: ERR_get_error_line(),
1079 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1080 ERR_peek_last_error_line_data() and ERR_func_error_string().
1082 Users are recommended to use ERR_get_error_all(), or to pick information
1083 with ERR_peek functions and finish off with getting the error code by using
1088 * Extended testing to be verbose for failing tests only. The make variables
1089 VERBOSE_FAILURE or VF can be used to enable this:
1091 $ make VF=1 test # Unix
1092 $ mms /macro=(VF=1) test ! OpenVMS
1093 $ nmake VF=1 test # Windows
1097 * Added the `-copy_extensions` option to the `x509` command for use with
1098 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1099 all extensions in the request are copied to the certificate or vice versa.
1101 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1103 * Added the `-copy_extensions` option to the `req` command for use with
1104 `-x509`. When given with the `copy` or `copyall` argument,
1105 all extensions in the certification request are copied to the certificate.
1109 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1110 they generate are by default RFC 5280 compliant in the following sense:
1111 There is a subjectKeyIdentifier extension with a hash value of the public key
1112 and for not self-signed certs there is an authorityKeyIdentifier extension
1113 with a keyIdentifier field or issuer information identifying the signing key.
1114 This is done unless some configuration overrides the new default behavior,
1115 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1119 * Added several checks to `X509_verify_cert()` according to requirements in
1120 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1121 (which may be done by using the CLI option `-x509_strict`):
1122 * The basicConstraints of CA certificates must be marked critical.
1123 * CA certificates must explicitly include the keyUsage extension.
1124 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1125 * The issuer name of any certificate must not be empty.
1126 * The subject name of CA certs, certs with keyUsage crlSign,
1127 and certs without subjectAlternativeName must not be empty.
1128 * If a subjectAlternativeName extension is given it must not be empty.
1129 * The signatureAlgorithm field and the cert signature must be consistent.
1130 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1131 must not be marked critical.
1132 * The authorityKeyIdentifier must be given for X.509v3 certs
1133 unless they are self-signed.
1134 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1138 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1139 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1143 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1144 used even when parsing explicit parameters, when loading a encoded key
1145 or calling `EC_GROUP_new_from_ecpkparameters()`/
1146 `EC_GROUP_new_from_ecparameters()`.
1147 This prevents bypass of security hardening and performance gains,
1148 especially for curves with specialized EC_METHODs.
1149 By default, if a key encoded with explicit parameters is loaded and later
1150 encoded, the output is still encoded with explicit parameters, even if
1151 internally a "named" EC_GROUP is used for computation.
1155 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1156 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1157 NULL. After this change, only the cofactor parameter can be NULL. It also
1158 does some minimal sanity checks on the passed order.
1163 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1164 An attack is simple, if the first CMS_recipientInfo is valid but the
1165 second CMS_recipientInfo is chosen ciphertext. If the second
1166 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1167 encryption key will be replaced by garbage, and the message cannot be
1168 decoded, but if the RSA decryption fails, the correct encryption key is
1169 used and the recipient will not notice the attack.
1170 As a work around for this potential attack the length of the decrypted
1171 key must be equal to the cipher default key length, in case the
1172 certifiate is not given and all recipientInfo are tried out.
1173 The old behaviour can be re-enabled in the CMS code by setting the
1174 CMS_DEBUG_DECRYPT flag.
1178 * Early start up entropy quality from the DEVRANDOM seed source has been
1179 improved for older Linux systems. The RAND subsystem will wait for
1180 /dev/random to be producing output before seeding from /dev/urandom.
1181 The seeded state is stored for future library initialisations using
1182 a system global shared memory segment. The shared memory identifier
1183 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1184 the desired value. The default identifier is 114.
1188 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1189 when primes for RSA keys are computed.
1190 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1191 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1192 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1193 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1194 This avoids possible fingerprinting of newly generated RSA modules.
1198 * Correct the extended master secret constant on EBCDIC systems. Without this
1199 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1200 negotiate EMS will fail. Unfortunately this also means that TLS connections
1201 between EBCDIC systems with this fix, and EBCDIC systems without this
1202 fix will fail if they negotiate EMS.
1206 * Changed the library initialisation so that the config file is now loaded
1207 by default. This was already the case for libssl. It now occurs for both
1208 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1209 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1213 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1214 where the former acts as a replacement for `ERR_put_error()`, and the
1215 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1216 `ERR_raise_data()` adds more flexibility by taking a format string and
1217 an arbitrary number of arguments following it, to be processed with
1222 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1223 to check if a named provider is loaded and available. When called, it
1224 will also activate all fallback providers if such are still present.
1228 * Enforce a minimum DH modulus size of 512 bits.
1232 * Changed DH parameters to generate the order q subgroup instead of 2q.
1233 Previously generated DH parameters are still accepted by DH_check
1234 but DH_generate_key works around that by clearing bit 0 of the
1235 private key for those. This avoids leaking bit 0 of the private key.
1239 * Significantly reduce secure memory usage by the randomness pools.
1243 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1248 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1249 algorithms. An implementation of a key exchange algorithm can be obtained
1250 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1251 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1252 the older EVP_PKEY_derive_init() function. See the man pages for the new
1253 functions for further details.
1257 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1261 * Removed the function names from error messages and deprecated the
1264 * Removed NextStep support and the macro OPENSSL_UNISTD
1268 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1269 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1270 Also removed "export var as function" capability; we do not export
1271 variables, only functions.
1275 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1276 an error and 1 indicating success. In previous versions of OpenSSL this
1277 was a void type. If a key was set longer than the maximum possible this
1282 * Support SM2 signing and verification schemes with X509 certificate.
1286 * Use SHA256 as the default digest for TS query in the `ts` app.
1290 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1291 This checks that the salt length is at least 128 bits, the derived key
1292 length is at least 112 bits, and that the iteration count is at least 1000.
1293 For backwards compatibility these checks are disabled by default in the
1294 default provider, but are enabled by default in the fips provider.
1295 To enable or disable these checks use the control
1296 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1300 * Default cipher lists/suites are now available via a function, the
1301 #defines are deprecated.
1305 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1306 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1307 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1311 * Join the directories crypto/x509 and crypto/x509v3
1315 * Added command 'openssl kdf' that uses the EVP_KDF API.
1319 * Added command 'openssl mac' that uses the EVP_MAC API.
1323 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1324 as default directories. Also added the command 'openssl info'
1325 for scripting purposes.
1329 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1330 deprecated. These undocumented functions were never integrated into the EVP
1331 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1332 Bi-directional IGE mode. These modes were never formally standardised and
1333 usage of these functions is believed to be very small. In particular
1334 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1335 is ever used. The security implications are believed to be minimal, but
1336 this issue was never fixed for backwards compatibility reasons. New code
1337 should not use these modes.
1341 * Add prediction resistance to the DRBG reseeding process.
1345 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1346 mandated by IEEE Std 1619-2018.
1350 * Added newline escaping functionality to a filename when using openssl dgst.
1351 This output format is to replicate the output format found in the `*sum`
1352 checksum programs. This aims to preserve backward compatibility.
1354 *Matt Eaton, Richard Levitte, and Paul Dale*
1356 * Removed the heartbeat message in DTLS feature, as it has very
1357 little usage and doesn't seem to fulfill a valuable purpose.
1358 The configuration option is now deprecated.
1362 * Changed the output of 'openssl {digestname} < file' to display the
1363 digest name in its output.
1367 * Added a new generic trace API which provides support for enabling
1368 instrumentation through trace output. This feature is mainly intended
1369 as an aid for developers and is disabled by default. To utilize it,
1370 OpenSSL needs to be configured with the `enable-trace` option.
1372 If the tracing API is enabled, the application can activate trace output
1373 by registering BIOs as trace channels for a number of tracing and debugging
1376 The `openssl` program has been expanded to enable any of the types
1377 available via environment variables defined by the user, and serves as
1378 one possible example on how to use this functionality.
1380 *Richard Levitte & Matthias St. Pierre*
1382 * Added build tests for C++. These are generated files that only do one
1383 thing, to include one public OpenSSL head file each. This tests that
1384 the public header files can be usefully included in a C++ application.
1386 This test isn't enabled by default. It can be enabled with the option
1387 'enable-buildtest-c++'.
1391 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1395 * Add KMAC to EVP_MAC.
1399 * Added property based algorithm implementation selection framework to
1404 * Added SCA hardening for modular field inversion in EC_GROUP through
1405 a new dedicated field_inv() pointer in EC_METHOD.
1406 This also addresses a leakage affecting conversions from projective
1407 to affine coordinates.
1409 *Billy Bob Brumley, Nicola Tuveri*
1411 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1412 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1413 those algorithms that were already supported through the EVP_PKEY API
1414 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1415 and scrypt are now wrappers that call EVP_KDF.
1419 * Build devcrypto engine as a dynamic engine.
1421 *Eneas U de Queiroz*
1423 * Add keyed BLAKE2 to EVP_MAC.
1427 * Fix a bug in the computation of the endpoint-pair shared secret used
1428 by DTLS over SCTP. This breaks interoperability with older versions
1429 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1430 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1431 interoperability with such broken implementations. However, enabling
1432 this switch breaks interoperability with correct implementations.
1434 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1435 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1439 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1443 * Change the license to the Apache License v2.0.
1447 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1449 - Major releases (indicated by incrementing the MAJOR release number)
1450 may introduce incompatible API/ABI changes.
1451 - Minor releases (indicated by incrementing the MINOR release number)
1452 may introduce new features but retain API/ABI compatibility.
1453 - Patch releases (indicated by incrementing the PATCH number)
1454 are intended for bug fixes and other improvements of existing
1455 features only (like improving performance or adding documentation)
1456 and retain API/ABI compatibility.
1460 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1464 * Remove the 'dist' target and add a tarball building script. The
1465 'dist' target has fallen out of use, and it shouldn't be
1466 necessary to configure just to create a source distribution.
1470 * Recreate the OS390-Unix config target. It no longer relies on a
1471 special script like it did for OpenSSL pre-1.1.0.
1475 * Instead of having the source directories listed in Configure, add
1476 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1481 * Add GMAC to EVP_MAC.
1485 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1489 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1490 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1491 to facilitate the continued use of MACs through raw private keys in
1492 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1496 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1497 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1501 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1502 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1503 are retained for backwards compatibility.
1507 * AES-XTS mode now enforces that its two keys are different to mitigate
1508 the attacked described in "Efficient Instantiations of Tweakable
1509 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1510 Details of this attack can be obtained from:
1511 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1515 * Rename the object files, i.e. give them other names than in previous
1516 versions. Their names now include the name of the final product, as
1517 well as its type mnemonic (bin, lib, shlib).
1521 * Added new option for 'openssl list', '-objects', which will display the
1522 list of built in objects, i.e. OIDs with names.
1526 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1527 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1532 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1533 improves application performance by removing data copies and providing
1534 applications with zero-copy system calls such as sendfile and splice.
1538 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1539 option is set, openssl cleanses (zeroize) plaintext bytes from
1540 internal buffers after delivering them to the application. Note,
1541 the application is still responsible for cleansing other copies
1542 (e.g.: data received by SSL_read(3)).
1546 * `PKCS12_parse` now maintains the order of the parsed certificates
1547 when outputting them via `*ca` (rather than reversing it).
1551 * Deprecated pthread fork support methods. These were unused so no
1552 replacement is required.
1554 - OPENSSL_fork_prepare()
1555 - OPENSSL_fork_parent()
1556 - OPENSSL_fork_child()
1563 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1565 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1567 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1568 create a unique hash value based on the issuer and serial number data
1569 contained within an X509 certificate. However it was failing to correctly
1570 handle any errors that may occur while parsing the issuer field (which might
1571 occur if the issuer field is maliciously constructed). This may subsequently
1572 result in a NULL pointer deref and a crash leading to a potential denial of
1578 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1579 padding mode to correctly check for rollback attacks. This is considered a
1580 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1585 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1586 functions. Previously they could overflow the output length argument in some
1587 cases where the input length is close to the maximum permissable length for
1588 an integer on the platform. In such cases the return value from the function
1589 call would be 1 (indicating success), but the output length value would be
1590 negative. This could cause applications to behave incorrectly or crash.
1595 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1596 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1597 could be exploited in a side channel attack to recover the password. Since
1598 the attack is local host only this is outside of the current OpenSSL
1599 threat model and therefore no CVE is assigned.
1601 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1606 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1608 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1609 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1610 If an attacker can control both items being compared then this could lead
1611 to a possible denial of service attack. OpenSSL itself uses the
1612 GENERAL_NAME_cmp function for two purposes:
1613 1) Comparing CRL distribution point names between an available CRL and a
1614 CRL distribution point embedded in an X509 certificate
1615 2) When verifying that a timestamp response token signer matches the
1616 timestamp authority name (exposed via the API functions
1617 TS_RESP_verify_response and TS_RESP_verify_token)
1622 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1624 * Certificates with explicit curve parameters are now disallowed in
1625 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1629 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1630 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1631 conversely, silently ignore DTLS protocol version bounds when configuring
1632 TLS-based contexts. The commands can be repeated to set bounds of both
1633 types. The same applies with the corresponding "min_protocol" and
1634 "max_protocol" command-line switches, in case some application uses both TLS
1637 SSL_CTX instances that are created for a fixed protocol version (e.g.
1638 TLSv1_server_method()) also silently ignore version bounds. Previously
1639 attempts to apply bounds to these protocol versions would result in an
1640 error. Now only the "version-flexible" SSL_CTX instances are subject to
1641 limits in configuration files in command-line options.
1645 * Handshake now fails if Extended Master Secret extension is dropped
1650 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1652 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1654 * Fixed segmentation fault in SSL_check_chain()
1655 Server or client applications that call the SSL_check_chain() function
1656 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1657 dereference as a result of incorrect handling of the
1658 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1659 or unrecognised signature algorithm is received from the peer. This could
1660 be exploited by a malicious peer in a Denial of Service attack.
1665 * Added AES consttime code for no-asm configurations
1666 an optional constant time support for AES was added
1667 when building openssl for no-asm.
1668 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1669 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1670 At this time this feature is by default disabled.
1671 It will be enabled by default in 3.0.
1675 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1677 * Revert the change of EOF detection while reading in libssl to avoid
1678 regressions in applications depending on the current way of reporting
1679 the EOF. As the existing method is not fully accurate the change to
1680 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1681 branch and will be present in the 3.0 release.
1685 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1686 when primes for RSA keys are computed.
1687 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1688 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1689 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1690 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1691 This avoids possible fingerprinting of newly generated RSA modules.
1695 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1697 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1698 while reading in libssl then we would report an error back to the
1699 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1700 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1701 therefore give a hint as to what went wrong.
1705 * Check that ed25519 and ed448 are allowed by the security level. Previously
1706 signature algorithms not using an MD were not being checked that they were
1707 allowed by the security level.
1711 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1712 was not quite right. The behaviour was not consistent between resumption
1713 and normal handshakes, and also not quite consistent with historical
1714 behaviour. The behaviour in various scenarios has been clarified and
1715 it has been updated to make it match historical behaviour as closely as
1720 * *[VMS only]* The header files that the VMS compilers include automatically,
1721 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1722 that the C++ compiler doesn't understand. This is a shortcoming in the
1723 compiler, but can be worked around with `__cplusplus` guards.
1725 C++ applications that use OpenSSL libraries must be compiled using the
1726 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1727 functions. Otherwise, only functions with symbols of less than 31
1728 characters can be used, as the linker will not be able to successfully
1729 resolve symbols with longer names.
1733 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1734 The presence of this system service is determined at run-time.
1738 * Added newline escaping functionality to a filename when using openssl dgst.
1739 This output format is to replicate the output format found in the `*sum`
1740 checksum programs. This aims to preserve backward compatibility.
1742 *Matt Eaton, Richard Levitte, and Paul Dale*
1744 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1749 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1751 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1752 number generator (RNG). This was intended to include protection in the
1753 event of a fork() system call in order to ensure that the parent and child
1754 processes did not share the same RNG state. However this protection was not
1755 being used in the default case.
1757 A partial mitigation for this issue is that the output from a high
1758 precision timer is mixed into the RNG state so the likelihood of a parent
1759 and child process sharing state is significantly reduced.
1761 If an application already calls OPENSSL_init_crypto() explicitly using
1762 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1765 *Matthias St. Pierre*
1767 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1768 used even when parsing explicit parameters, when loading a encoded key
1769 or calling `EC_GROUP_new_from_ecpkparameters()`/
1770 `EC_GROUP_new_from_ecparameters()`.
1771 This prevents bypass of security hardening and performance gains,
1772 especially for curves with specialized EC_METHODs.
1773 By default, if a key encoded with explicit parameters is loaded and later
1774 encoded, the output is still encoded with explicit parameters, even if
1775 internally a "named" EC_GROUP is used for computation.
1779 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1780 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1781 NULL. After this change, only the cofactor parameter can be NULL. It also
1782 does some minimal sanity checks on the passed order.
1787 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1788 An attack is simple, if the first CMS_recipientInfo is valid but the
1789 second CMS_recipientInfo is chosen ciphertext. If the second
1790 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1791 encryption key will be replaced by garbage, and the message cannot be
1792 decoded, but if the RSA decryption fails, the correct encryption key is
1793 used and the recipient will not notice the attack.
1794 As a work around for this potential attack the length of the decrypted
1795 key must be equal to the cipher default key length, in case the
1796 certifiate is not given and all recipientInfo are tried out.
1797 The old behaviour can be re-enabled in the CMS code by setting the
1798 CMS_DEBUG_DECRYPT flag.
1803 * Early start up entropy quality from the DEVRANDOM seed source has been
1804 improved for older Linux systems. The RAND subsystem will wait for
1805 /dev/random to be producing output before seeding from /dev/urandom.
1806 The seeded state is stored for future library initialisations using
1807 a system global shared memory segment. The shared memory identifier
1808 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1809 the desired value. The default identifier is 114.
1813 * Correct the extended master secret constant on EBCDIC systems. Without this
1814 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1815 negotiate EMS will fail. Unfortunately this also means that TLS connections
1816 between EBCDIC systems with this fix, and EBCDIC systems without this
1817 fix will fail if they negotiate EMS.
1821 * Use Windows installation paths in the mingw builds
1823 Mingw isn't a POSIX environment per se, which means that Windows
1824 paths should be used for installation.
1829 * Changed DH_check to accept parameters with order q and 2q subgroups.
1830 With order 2q subgroups the bit 0 of the private key is not secret
1831 but DH_generate_key works around that by clearing bit 0 of the
1832 private key for those. This avoids leaking bit 0 of the private key.
1836 * Significantly reduce secure memory usage by the randomness pools.
1840 * Revert the DEVRANDOM_WAIT feature for Linux systems
1842 The DEVRANDOM_WAIT feature added a select() call to wait for the
1843 /dev/random device to become readable before reading from the
1844 /dev/urandom device.
1846 It turned out that this change had negative side effects on
1847 performance which were not acceptable. After some discussion it
1848 was decided to revert this feature and leave it up to the OS
1849 resp. the platform maintainer to ensure a proper initialization
1850 during early boot time.
1852 *Matthias St. Pierre*
1854 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1856 * Add build tests for C++. These are generated files that only do one
1857 thing, to include one public OpenSSL head file each. This tests that
1858 the public header files can be usefully included in a C++ application.
1860 This test isn't enabled by default. It can be enabled with the option
1861 'enable-buildtest-c++'.
1865 * Enable SHA3 pre-hashing for ECDSA and DSA.
1869 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1870 This changes the size when using the `genpkey` command when no size is given.
1871 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1872 generation commands to use 2048 bits by default.
1876 * Reorganize the manual pages to consistently have RETURN VALUES,
1877 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1878 util/fix-doc-nits accordingly.
1880 *Paul Yang, Joshua Lock*
1882 * Add the missing accessor EVP_PKEY_get0_engine()
1886 * Have commands like `s_client` and `s_server` output the signature scheme
1887 along with other cipher suite parameters when debugging.
1891 * Make OPENSSL_config() error agnostic again.
1895 * Do the error handling in RSA decryption constant time.
1899 * Prevent over long nonces in ChaCha20-Poly1305.
1901 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1902 for every encryption operation. RFC 7539 specifies that the nonce value
1903 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1904 and front pads the nonce with 0 bytes if it is less than 12
1905 bytes. However it also incorrectly allows a nonce to be set of up to 16
1906 bytes. In this case only the last 12 bytes are significant and any
1907 additional leading bytes are ignored.
1909 It is a requirement of using this cipher that nonce values are
1910 unique. Messages encrypted using a reused nonce value are susceptible to
1911 serious confidentiality and integrity attacks. If an application changes
1912 the default nonce length to be longer than 12 bytes and then makes a
1913 change to the leading bytes of the nonce expecting the new value to be a
1914 new unique nonce then such an application could inadvertently encrypt
1915 messages with a reused nonce.
1917 Additionally the ignored bytes in a long nonce are not covered by the
1918 integrity guarantee of this cipher. Any application that relies on the
1919 integrity of these ignored leading bytes of a long nonce may be further
1920 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1921 is safe because no such use sets such a long nonce value. However user
1922 applications that use this cipher directly and set a non-default nonce
1923 length to be longer than 12 bytes may be vulnerable.
1925 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1931 * Add DEVRANDOM_WAIT feature for Linux systems
1933 On older Linux systems where the getrandom() system call is not available,
1934 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1935 Contrary to getrandom(), the /dev/urandom device will not block during
1936 early boot when the kernel CSPRNG has not been seeded yet.
1938 To mitigate this known weakness, use select() to wait for /dev/random to
1939 become readable before reading from /dev/urandom.
1941 * Ensure that SM2 only uses SM3 as digest algorithm
1945 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1947 * Change the info callback signals for the start and end of a post-handshake
1948 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1949 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1950 confused by this and assume that a TLSv1.2 renegotiation has started. This
1951 can break KeyUpdate handling. Instead we no longer signal the start and end
1952 of a post handshake message exchange (although the messages themselves are
1953 still signalled). This could break some applications that were expecting
1954 the old signals. However without this KeyUpdate is not usable for many
1959 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1961 * Timing vulnerability in DSA signature generation
1963 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1964 timing side channel attack. An attacker could use variations in the signing
1965 algorithm to recover the private key.
1967 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1972 * Timing vulnerability in ECDSA signature generation
1974 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1975 timing side channel attack. An attacker could use variations in the signing
1976 algorithm to recover the private key.
1978 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1983 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1984 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1985 of two gigabytes and the error handling improved.
1987 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1988 categorized as a normal bug, not a security issue, because the DRBG reseeds
1989 automatically and is fully functional even without additional randomness
1990 provided by the application.
1992 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1994 * Add a new ClientHello callback. Provides a callback interface that gives
1995 the application the ability to adjust the nascent SSL object at the
1996 earliest stage of ClientHello processing, immediately after extensions have
1997 been collected but before they have been processed. In particular, this
1998 callback can adjust the supported TLS versions in response to the contents
2003 * Add SM2 base algorithm support.
2007 * s390x assembly pack: add (improved) hardware-support for the following
2008 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2009 aes-cfb/cfb8, aes-ecb.
2013 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2014 parameter is no longer accepted, as it leads to a corrupt table. NULL
2015 pem_str is reserved for alias entries only.
2019 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2020 step for prime curves. The new implementation is based on formulae from
2021 differential addition-and-doubling in homogeneous projective coordinates
2022 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2023 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2024 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2025 to work in projective coordinates.
2027 *Billy Bob Brumley, Nicola Tuveri*
2029 * Change generating and checking of primes so that the error rate of not
2030 being prime depends on the intended use based on the size of the input.
2031 For larger primes this will result in more rounds of Miller-Rabin.
2032 The maximal error rate for primes with more than 1080 bits is lowered
2035 *Kurt Roeckx, Annie Yousar*
2037 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2041 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2042 moving between systems, and to avoid confusion when a Windows build is
2043 done with mingw vs with MSVC. For POSIX installs, there's still a
2044 symlink or copy named 'tsget' to avoid that confusion as well.
2048 * Revert blinding in ECDSA sign and instead make problematic addition
2049 length-invariant. Switch even to fixed-length Montgomery multiplication.
2053 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2054 step for binary curves. The new implementation is based on formulae from
2055 differential addition-and-doubling in mixed Lopez-Dahab projective
2056 coordinates, modified to independently blind the operands.
2058 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2060 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2061 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2062 EC_METHODs to implement their own specialized "ladder step", to take
2063 advantage of more favorable coordinate systems or more efficient
2064 differential addition-and-doubling algorithms.
2066 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2068 * Modified the random device based seed sources to keep the relevant
2069 file descriptors open rather than reopening them on each access.
2070 This allows such sources to operate in a chroot() jail without
2071 the associated device nodes being available. This behaviour can be
2072 controlled using RAND_keep_random_devices_open().
2076 * Numerous side-channel attack mitigations have been applied. This may have
2077 performance impacts for some algorithms for the benefit of improved
2078 security. Specific changes are noted in this change log by their respective
2083 * AIX shared library support overhaul. Switch to AIX "natural" way of
2084 handling shared libraries, which means collecting shared objects of
2085 different versions and bitnesses in one common archive. This allows to
2086 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2087 doesn't affect the way 3rd party applications are linked, only how
2088 multi-version installation is managed.
2092 * Make ec_group_do_inverse_ord() more robust and available to other
2093 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2094 mitigations are applied to the fallback BN_mod_inverse().
2095 When using this function rather than BN_mod_inverse() directly, new
2096 EC cryptosystem implementations are then safer-by-default.
2100 * Add coordinate blinding for EC_POINT and implement projective
2101 coordinate blinding for generic prime curves as a countermeasure to
2102 chosen point SCA attacks.
2104 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2106 * Add blinding to ECDSA and DSA signatures to protect against side channel
2107 attacks discovered by Keegan Ryan (NCC Group).
2111 * Enforce checking in the `pkeyutl` command to ensure that the input
2112 length does not exceed the maximum supported digest length when performing
2113 a sign, verify or verifyrecover operation.
2117 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2118 I/O in combination with something like select() or poll() will hang. This
2119 can be turned off again using SSL_CTX_clear_mode().
2120 Many applications do not properly handle non-application data records, and
2121 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2122 around the problems in those applications, but can also break some.
2123 It's recommended to read the manpages about SSL_read(), SSL_write(),
2124 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2125 SSL_CTX_set_read_ahead() again.
2129 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2130 now allow empty (zero character) pass phrases.
2134 * Apply blinding to binary field modular inversion and remove patent
2135 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2139 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2140 binary and prime elliptic curves.
2144 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2145 constant time fixed point multiplication.
2149 * Revise elliptic curve scalar multiplication with timing attack
2150 defenses: ec_wNAF_mul redirects to a constant time implementation
2151 when computing fixed point and variable point multiplication (which
2152 in OpenSSL are mostly used with secret scalars in keygen, sign,
2153 ECDH derive operations).
2154 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2157 * Updated CONTRIBUTING
2161 * Updated DRBG / RAND to request nonce and additional low entropy
2162 randomness from the system.
2164 *Matthias St. Pierre*
2166 * Updated 'openssl rehash' to use OpenSSL consistent default.
2170 * Moved the load of the ssl_conf module to libcrypto, which helps
2171 loading engines that libssl uses before libssl is initialised.
2175 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2179 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2181 *Ingo Schwarze, Rich Salz*
2183 * Added output of accepting IP address and port for 'openssl s_server'
2187 * Added a new API for TLSv1.3 ciphersuites:
2188 SSL_CTX_set_ciphersuites()
2189 SSL_set_ciphersuites()
2193 * Memory allocation failures consistently add an error to the error
2198 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2199 in libcrypto when run as setuid/setgid.
2203 * Load any config file by default when libssl is used.
2207 * Added new public header file <openssl/rand_drbg.h> and documentation
2208 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2210 *Matthias St. Pierre*
2212 * QNX support removed (cannot find contributors to get their approval
2213 for the license change).
2217 * TLSv1.3 replay protection for early data has been implemented. See the
2218 SSL_read_early_data() man page for further details.
2222 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2223 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2224 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2225 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2226 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2227 configuration has been separated out. See the ciphers man page or the
2228 SSL_CTX_set_ciphersuites() man page for more information.
2232 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2233 in responder mode now supports the new "-multi" option, which
2234 spawns the specified number of child processes to handle OCSP
2235 requests. The "-timeout" option now also limits the OCSP
2236 responder's patience to wait to receive the full client request
2237 on a newly accepted connection. Child processes are respawned
2238 as needed, and the CA index file is automatically reloaded
2239 when changed. This makes it possible to run the "ocsp" responder
2240 as a long-running service, making the OpenSSL CA somewhat more
2241 feature-complete. In this mode, most diagnostic messages logged
2242 after entering the event loop are logged via syslog(3) rather than
2247 * Added support for X448 and Ed448. Heavily based on original work by
2252 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2253 objects loaded. This adds the functions OSSL_STORE_expect() and
2254 OSSL_STORE_find() as well as needed tools to construct searches and
2255 get the search data out of them.
2259 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2260 version of OpenSSL should review their configuration settings to ensure
2261 that they are still appropriate for TLSv1.3. For further information see:
2262 <https://wiki.openssl.org/index.php/TLS1.3>
2266 * Grand redesign of the OpenSSL random generator
2268 The default RAND method now utilizes an AES-CTR DRBG according to
2269 NIST standard SP 800-90Ar1. The new random generator is essentially
2270 a port of the default random generator from the OpenSSL FIPS 2.0
2271 object module. It is a hybrid deterministic random bit generator
2272 using an AES-CTR bit stream and which seeds and reseeds itself
2273 automatically using trusted system entropy sources.
2275 Some of its new features are:
2276 - Support for multiple DRBG instances with seed chaining.
2277 - The default RAND method makes use of a DRBG.
2278 - There is a public and private DRBG instance.
2279 - The DRBG instances are fork-safe.
2280 - Keep all global DRBG instances on the secure heap if it is enabled.
2281 - The public and private DRBG instance are per thread for lock free
2284 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2286 * Changed Configure so it only says what it does and doesn't dump
2287 so much data. Instead, ./configdata.pm should be used as a script
2288 to display all sorts of configuration data.
2292 * Added processing of "make variables" to Configure.
2296 * Added SHA512/224 and SHA512/256 algorithm support.
2300 * The last traces of Netware support, first removed in 1.1.0, have
2305 * Get rid of Makefile.shared, and in the process, make the processing
2306 of certain files (rc.obj, or the .def/.map/.opt files produced from
2307 the ordinal files) more visible and hopefully easier to trace and
2308 debug (or make silent).
2312 * Make it possible to have environment variable assignments as
2313 arguments to config / Configure.
2317 * Add multi-prime RSA (RFC 8017) support.
2321 * Add SM3 implemented according to GB/T 32905-2016
2322 *Jack Lloyd <jack.lloyd@ribose.com>,*
2323 *Ronald Tse <ronald.tse@ribose.com>,*
2324 *Erick Borsboom <erick.borsboom@ribose.com>*
2326 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2327 as documented in RFC6066.
2328 Based on a patch from Tomasz Moń
2330 *Filipe Raimundo da Silva*
2332 * Add SM4 implemented according to GB/T 32907-2016.
2333 *Jack Lloyd <jack.lloyd@ribose.com>,*
2334 *Ronald Tse <ronald.tse@ribose.com>,*
2335 *Erick Borsboom <erick.borsboom@ribose.com>*
2337 * Reimplement -newreq-nodes and ERR_error_string_n; the
2338 original author does not agree with the license change.
2342 * Add ARIA AEAD TLS support.
2346 * Some macro definitions to support VS6 have been removed. Visual
2347 Studio 6 has not worked since 1.1.0
2351 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2352 without clearing the errors.
2356 * Add "atfork" functions. If building on a system that without
2357 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2358 requirements. The RAND facility now uses/requires this.
2366 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2367 not possible to disable entirely. However, it's still possible to
2368 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2371 To disable, configure with 'no-ui-console'. 'no-ui' is still
2372 possible to use as an alias. Check at compile time with the
2373 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2374 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2378 * Add a STORE module, which implements a uniform and URI based reader of
2379 stores that can contain keys, certificates, CRLs and numerous other
2380 objects. The main API is loosely based on a few stdio functions,
2381 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2382 OSSL_STORE_error and OSSL_STORE_close.
2383 The implementation uses backends called "loaders" to implement arbitrary
2384 URI schemes. There is one built in "loader" for the 'file' scheme.
2388 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2389 then adjusted to work on FreeBSD 8.4 as well.
2390 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2391 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2395 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2396 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2397 error code calls like this:
2399 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2401 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2402 that can be encoded in C. For the foreseeable future, this will only
2405 *Richard Levitte and Tim Hudson*
2407 * Removed BSD cryptodev engine.
2411 * Add a build target 'build_all_generated', to build all generated files
2412 and only that. This can be used to prepare everything that requires
2413 things like perl for a system that lacks perl and then move everything
2414 to that system and do the rest of the build there.
2418 * In the UI interface, make it possible to duplicate the user data. This
2419 can be used by engines that need to retain the data for a longer time
2420 than just the call where this user data is passed.
2424 * Ignore the '-named_curve auto' value for compatibility of applications
2427 *Tomáš Mráz <tmraz@fedoraproject.org>*
2429 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2430 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2431 alerts across multiple records (some of which could be empty). In practice
2432 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2433 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2434 support this at all. Supporting it adds significant complexity to the
2435 record layer, and its removal is unlikely to cause interoperability
2440 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2441 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2442 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2447 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2448 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2450 *Richard Levitte, Andy Polyakov*
2452 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2457 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2458 platform rather than 'mingw'.
2462 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2463 success if they are asked to add an object which already exists
2464 in the store. This change cascades to other functions which load
2465 certificates and CRLs.
2469 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2470 facilitate stack unwinding even from assembly subroutines.
2474 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2475 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2479 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2480 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2481 which is the minimum version we support.
2485 * Certificate time validation (X509_cmp_time) enforces stricter
2486 compliance with RFC 5280. Fractional seconds and timezone offsets
2487 are no longer allowed.
2491 * Add support for ARIA
2495 * s_client will now send the Server Name Indication (SNI) extension by
2496 default unless the new "-noservername" option is used. The server name is
2497 based on the host provided to the "-connect" option unless overridden by
2498 using "-servername".
2502 * Add support for SipHash
2506 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2507 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2508 prevent issues where no progress is being made and the peer continually
2509 sends unrecognised record types, using up resources processing them.
2513 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2514 using the algorithm defined in
2515 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2519 * Heartbeat support has been removed; the ABI is changed for now.
2521 *Richard Levitte, Rich Salz*
2523 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2527 * The RSA "null" method, which was partially supported to avoid patent
2528 issues, has been replaced to always returns NULL.
2535 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2537 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2538 used even when parsing explicit parameters, when loading a encoded key
2539 or calling `EC_GROUP_new_from_ecpkparameters()`/
2540 `EC_GROUP_new_from_ecparameters()`.
2541 This prevents bypass of security hardening and performance gains,
2542 especially for curves with specialized EC_METHODs.
2543 By default, if a key encoded with explicit parameters is loaded and later
2544 encoded, the output is still encoded with explicit parameters, even if
2545 internally a "named" EC_GROUP is used for computation.
2549 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2550 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2551 NULL. After this change, only the cofactor parameter can be NULL. It also
2552 does some minimal sanity checks on the passed order.
2557 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2558 An attack is simple, if the first CMS_recipientInfo is valid but the
2559 second CMS_recipientInfo is chosen ciphertext. If the second
2560 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2561 encryption key will be replaced by garbage, and the message cannot be
2562 decoded, but if the RSA decryption fails, the correct encryption key is
2563 used and the recipient will not notice the attack.
2564 As a work around for this potential attack the length of the decrypted
2565 key must be equal to the cipher default key length, in case the
2566 certifiate is not given and all recipientInfo are tried out.
2567 The old behaviour can be re-enabled in the CMS code by setting the
2568 CMS_DEBUG_DECRYPT flag.
2573 * Use Windows installation paths in the mingw builds
2575 Mingw isn't a POSIX environment per se, which means that Windows
2576 paths should be used for installation.
2581 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2583 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2584 This changes the size when using the `genpkey` command when no size is given.
2585 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2586 generation commands to use 2048 bits by default.
2590 * Prevent over long nonces in ChaCha20-Poly1305.
2592 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2593 for every encryption operation. RFC 7539 specifies that the nonce value
2594 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2595 and front pads the nonce with 0 bytes if it is less than 12
2596 bytes. However it also incorrectly allows a nonce to be set of up to 16
2597 bytes. In this case only the last 12 bytes are significant and any
2598 additional leading bytes are ignored.
2600 It is a requirement of using this cipher that nonce values are
2601 unique. Messages encrypted using a reused nonce value are susceptible to
2602 serious confidentiality and integrity attacks. If an application changes
2603 the default nonce length to be longer than 12 bytes and then makes a
2604 change to the leading bytes of the nonce expecting the new value to be a
2605 new unique nonce then such an application could inadvertently encrypt
2606 messages with a reused nonce.
2608 Additionally the ignored bytes in a long nonce are not covered by the
2609 integrity guarantee of this cipher. Any application that relies on the
2610 integrity of these ignored leading bytes of a long nonce may be further
2611 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2612 is safe because no such use sets such a long nonce value. However user
2613 applications that use this cipher directly and set a non-default nonce
2614 length to be longer than 12 bytes may be vulnerable.
2616 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2622 * Added SCA hardening for modular field inversion in EC_GROUP through
2623 a new dedicated field_inv() pointer in EC_METHOD.
2624 This also addresses a leakage affecting conversions from projective
2625 to affine coordinates.
2627 *Billy Bob Brumley, Nicola Tuveri*
2629 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2630 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2634 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2638 * Remove the 'dist' target and add a tarball building script. The
2639 'dist' target has fallen out of use, and it shouldn't be
2640 necessary to configure just to create a source distribution.
2644 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2646 * Timing vulnerability in DSA signature generation
2648 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2649 timing side channel attack. An attacker could use variations in the signing
2650 algorithm to recover the private key.
2652 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2657 * Timing vulnerability in ECDSA signature generation
2659 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2660 timing side channel attack. An attacker could use variations in the signing
2661 algorithm to recover the private key.
2663 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2668 * Add coordinate blinding for EC_POINT and implement projective
2669 coordinate blinding for generic prime curves as a countermeasure to
2670 chosen point SCA attacks.
2672 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2674 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2676 * Client DoS due to large DH parameter
2678 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2679 malicious server can send a very large prime value to the client. This will
2680 cause the client to spend an unreasonably long period of time generating a
2681 key for this prime resulting in a hang until the client has finished. This
2682 could be exploited in a Denial Of Service attack.
2684 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2689 * Cache timing vulnerability in RSA Key Generation
2691 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2692 a cache timing side channel attack. An attacker with sufficient access to
2693 mount cache timing attacks during the RSA key generation process could
2694 recover the private key.
2696 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2697 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2702 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2703 parameter is no longer accepted, as it leads to a corrupt table. NULL
2704 pem_str is reserved for alias entries only.
2708 * Revert blinding in ECDSA sign and instead make problematic addition
2709 length-invariant. Switch even to fixed-length Montgomery multiplication.
2713 * Change generating and checking of primes so that the error rate of not
2714 being prime depends on the intended use based on the size of the input.
2715 For larger primes this will result in more rounds of Miller-Rabin.
2716 The maximal error rate for primes with more than 1080 bits is lowered
2719 *Kurt Roeckx, Annie Yousar*
2721 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2725 * Add blinding to ECDSA and DSA signatures to protect against side channel
2726 attacks discovered by Keegan Ryan (NCC Group).
2730 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2731 now allow empty (zero character) pass phrases.
2735 * Certificate time validation (X509_cmp_time) enforces stricter
2736 compliance with RFC 5280. Fractional seconds and timezone offsets
2737 are no longer allowed.
2741 * Fixed a text canonicalisation bug in CMS
2743 Where a CMS detached signature is used with text content the text goes
2744 through a canonicalisation process first prior to signing or verifying a
2745 signature. This process strips trailing space at the end of lines, converts
2746 line terminators to CRLF and removes additional trailing line terminators
2747 at the end of a file. A bug in the canonicalisation process meant that
2748 some characters, such as form-feed, were incorrectly treated as whitespace
2749 and removed. This is contrary to the specification (RFC5485). This fix
2750 could mean that detached text data signed with an earlier version of
2751 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2752 signed with a fixed OpenSSL may fail to verify with an earlier version of
2753 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2754 and use the "-binary" flag (for the "cms" command line application) or set
2755 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2759 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2761 * Constructed ASN.1 types with a recursive definition could exceed the stack
2763 Constructed ASN.1 types with a recursive definition (such as can be found
2764 in PKCS7) could eventually exceed the stack given malicious input with
2765 excessive recursion. This could result in a Denial Of Service attack. There
2766 are no such structures used within SSL/TLS that come from untrusted sources
2767 so this is considered safe.
2769 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2775 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2777 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2778 effectively reduced to only comparing the least significant bit of each
2779 byte. This allows an attacker to forge messages that would be considered as
2780 authenticated in an amount of tries lower than that guaranteed by the
2781 security claims of the scheme. The module can only be compiled by the
2782 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2784 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2790 * Add a build target 'build_all_generated', to build all generated files
2791 and only that. This can be used to prepare everything that requires
2792 things like perl for a system that lacks perl and then move everything
2793 to that system and do the rest of the build there.
2797 * Backport SSL_OP_NO_RENGOTIATION
2799 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2800 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2801 changes this is no longer possible in 1.1.0. Therefore the new
2802 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2803 1.1.0 to provide equivalent functionality.
2805 Note that if an application built against 1.1.0h headers (or above) is run
2806 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2807 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2811 * Removed the OS390-Unix config target. It relied on a script that doesn't
2816 * rsaz_1024_mul_avx2 overflow bug on x86_64
2818 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2819 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2820 Analysis suggests that attacks against RSA and DSA as a result of this
2821 defect would be very difficult to perform and are not believed likely.
2822 Attacks against DH1024 are considered just feasible, because most of the
2823 work necessary to deduce information about a private key may be performed
2824 offline. The amount of resources required for such an attack would be
2825 significant. However, for an attack on TLS to be meaningful, the server
2826 would have to share the DH1024 private key among multiple clients, which is
2827 no longer an option since CVE-2016-0701.
2829 This only affects processors that support the AVX2 but not ADX extensions
2830 like Intel Haswell (4th generation).
2832 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2833 was originally found via the OSS-Fuzz project.
2838 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2840 * bn_sqrx8x_internal carry bug on x86_64
2842 There is a carry propagating bug in the x86_64 Montgomery squaring
2843 procedure. No EC algorithms are affected. Analysis suggests that attacks
2844 against RSA and DSA as a result of this defect would be very difficult to
2845 perform and are not believed likely. Attacks against DH are considered just
2846 feasible (although very difficult) because most of the work necessary to
2847 deduce information about a private key may be performed offline. The amount
2848 of resources required for such an attack would be very significant and
2849 likely only accessible to a limited number of attackers. An attacker would
2850 additionally need online access to an unpatched system using the target
2851 private key in a scenario with persistent DH parameters and a private
2852 key that is shared between multiple clients.
2854 This only affects processors that support the BMI1, BMI2 and ADX extensions
2855 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2857 This issue was reported to OpenSSL by the OSS-Fuzz project.
2862 * Malformed X.509 IPAddressFamily could cause OOB read
2864 If an X.509 certificate has a malformed IPAddressFamily extension,
2865 OpenSSL could do a one-byte buffer overread. The most likely result
2866 would be an erroneous display of the certificate in text format.
2868 This issue was reported to OpenSSL by the OSS-Fuzz project.
2873 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2875 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2876 platform rather than 'mingw'.
2880 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2881 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2882 which is the minimum version we support.
2886 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2888 * Encrypt-Then-Mac renegotiation crash
2890 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2891 negotiated where it was not in the original handshake (or vice-versa) then
2892 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2893 and servers are affected.
2895 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2900 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2902 * Truncated packet could crash via OOB read
2904 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2905 cipher is being used, then a truncated packet can cause that host to
2906 perform an out-of-bounds read, usually resulting in a crash.
2908 This issue was reported to OpenSSL by Robert Święcki of Google.
2913 * Bad (EC)DHE parameters cause a client crash
2915 If a malicious server supplies bad parameters for a DHE or ECDHE key
2916 exchange then this can result in the client attempting to dereference a
2917 NULL pointer leading to a client crash. This could be exploited in a Denial
2920 This issue was reported to OpenSSL by Guido Vranken.
2925 * BN_mod_exp may produce incorrect results on x86_64
2927 There is a carry propagating bug in the x86_64 Montgomery squaring
2928 procedure. No EC algorithms are affected. Analysis suggests that attacks
2929 against RSA and DSA as a result of this defect would be very difficult to
2930 perform and are not believed likely. Attacks against DH are considered just
2931 feasible (although very difficult) because most of the work necessary to
2932 deduce information about a private key may be performed offline. The amount
2933 of resources required for such an attack would be very significant and
2934 likely only accessible to a limited number of attackers. An attacker would
2935 additionally need online access to an unpatched system using the target
2936 private key in a scenario with persistent DH parameters and a private
2937 key that is shared between multiple clients. For example this can occur by
2938 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2939 similar to CVE-2015-3193 but must be treated as a separate problem.
2941 This issue was reported to OpenSSL by the OSS-Fuzz project.
2946 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2948 * ChaCha20/Poly1305 heap-buffer-overflow
2950 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2951 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2952 crash. This issue is not considered to be exploitable beyond a DoS.
2954 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2959 * CMS Null dereference
2961 Applications parsing invalid CMS structures can crash with a NULL pointer
2962 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2963 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2964 structure callback if an attempt is made to free certain invalid encodings.
2965 Only CHOICE structures using a callback which do not handle NULL value are
2968 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2973 * Montgomery multiplication may produce incorrect results
2975 There is a carry propagating bug in the Broadwell-specific Montgomery
2976 multiplication procedure that handles input lengths divisible by, but
2977 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2978 and DH private keys are impossible. This is because the subroutine in
2979 question is not used in operations with the private key itself and an input
2980 of the attacker's direct choice. Otherwise the bug can manifest itself as
2981 transient authentication and key negotiation failures or reproducible
2982 erroneous outcome of public-key operations with specially crafted input.
2983 Among EC algorithms only Brainpool P-512 curves are affected and one
2984 presumably can attack ECDH key negotiation. Impact was not analyzed in
2985 detail, because pre-requisites for attack are considered unlikely. Namely
2986 multiple clients have to choose the curve in question and the server has to
2987 share the private key among them, neither of which is default behaviour.
2988 Even then only clients that chose the curve will be affected.
2990 This issue was publicly reported as transient failures and was not
2991 initially recognized as a security issue. Thanks to Richard Morgan for
2992 providing reproducible case.
2997 * Removed automatic addition of RPATH in shared libraries and executables,
2998 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3002 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3004 * Fix Use After Free for large message sizes
3006 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3007 message larger than approx 16k is received then the underlying buffer to
3008 store the incoming message is reallocated and moved. Unfortunately a
3009 dangling pointer to the old location is left which results in an attempt to
3010 write to the previously freed location. This is likely to result in a
3011 crash, however it could potentially lead to execution of arbitrary code.
3013 This issue only affects OpenSSL 1.1.0a.
3015 This issue was reported to OpenSSL by Robert Święcki.
3020 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3022 * OCSP Status Request extension unbounded memory growth
3024 A malicious client can send an excessively large OCSP Status Request
3025 extension. If that client continually requests renegotiation, sending a
3026 large OCSP Status Request extension each time, then there will be unbounded
3027 memory growth on the server. This will eventually lead to a Denial Of
3028 Service attack through memory exhaustion. Servers with a default
3029 configuration are vulnerable even if they do not support OCSP. Builds using
3030 the "no-ocsp" build time option are not affected.
3032 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3037 * SSL_peek() hang on empty record
3039 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3040 sends an empty record. This could be exploited by a malicious peer in a
3041 Denial Of Service attack.
3043 This issue was reported to OpenSSL by Alex Gaynor.
3048 * Excessive allocation of memory in tls_get_message_header() and
3049 dtls1_preprocess_fragment()
3051 A (D)TLS message includes 3 bytes for its length in the header for the
3052 message. This would allow for messages up to 16Mb in length. Messages of
3053 this length are excessive and OpenSSL includes a check to ensure that a
3054 peer is sending reasonably sized messages in order to avoid too much memory
3055 being consumed to service a connection. A flaw in the logic of version
3056 1.1.0 means that memory for the message is allocated too early, prior to
3057 the excessive message length check. Due to way memory is allocated in
3058 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3059 to service a connection. This could lead to a Denial of Service through
3060 memory exhaustion. However, the excessive message length check still takes
3061 place, and this would cause the connection to immediately fail. Assuming
3062 that the application calls SSL_free() on the failed connection in a timely
3063 manner then the 21Mb of allocated memory will then be immediately freed
3064 again. Therefore the excessive memory allocation will be transitory in
3065 nature. This then means that there is only a security impact if:
3067 1) The application does not call SSL_free() in a timely manner in the event
3068 that the connection fails
3070 2) The application is working in a constrained environment where there is
3071 very little free memory
3073 3) The attacker initiates multiple connection attempts such that there are
3074 multiple connections in a state where memory has been allocated for the
3075 connection; SSL_free() has not yet been called; and there is insufficient
3076 memory to service the multiple requests.
3078 Except in the instance of (1) above any Denial Of Service is likely to be
3079 transitory because as soon as the connection fails the memory is
3080 subsequently freed again in the SSL_free() call. However there is an
3081 increased risk during this period of application crashes due to the lack of
3082 memory - which would then mean a more serious Denial of Service.
3084 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3085 (CVE-2016-6307 and CVE-2016-6308)
3089 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3090 had to be removed. Primary reason is that vendor assembler can't
3091 assemble our modules with -KPIC flag. As result it, assembly
3092 support, was not even available as option. But its lack means
3093 lack of side-channel resistant code, which is incompatible with
3094 security by todays standards. Fortunately gcc is readily available
3095 prepackaged option, which we firmly point at...
3099 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3101 * Windows command-line tool supports UTF-8 opt-in option for arguments
3102 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3103 (to any value) allows Windows user to access PKCS#12 file generated
3104 with Windows CryptoAPI and protected with non-ASCII password, as well
3105 as files generated under UTF-8 locale on Linux also protected with
3110 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3111 have been disabled by default and removed from DEFAULT, just like RC4.
3112 See the RC4 item below to re-enable both.
3116 * The method for finding the storage location for the Windows RAND seed file
3117 has changed. First we check %RANDFILE%. If that is not set then we check
3118 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3119 all else fails we fall back to C:\.
3123 * The EVP_EncryptUpdate() function has had its return type changed from void
3124 to int. A return of 0 indicates and error while a return of 1 indicates
3129 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3130 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3131 off the constant time implementation for RSA, DSA and DH have been made
3132 no-ops and deprecated.
3136 * Windows RAND implementation was simplified to only get entropy by
3137 calling CryptGenRandom(). Various other RAND-related tickets
3140 *Joseph Wylie Yandle, Rich Salz*
3142 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3143 and `OPENSSL_LH_`, respectively. The old names are available
3144 with API compatibility. They new names are now completely documented.
3148 * Unify TYPE_up_ref(obj) methods signature.
3149 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3150 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3151 int (instead of void) like all others TYPE_up_ref() methods.
3152 So now these methods also check the return value of CRYPTO_atomic_add(),
3153 and the validity of object reference counter.
3155 *fdasilvayy@gmail.com*
3157 * With Windows Visual Studio builds, the .pdb files are installed
3158 alongside the installed libraries and executables. For a static
3159 library installation, ossl_static.pdb is the associate compiler
3160 generated .pdb file to be used when linking programs.
3164 * Remove openssl.spec. Packaging files belong with the packagers.
3168 * Automatic Darwin/OSX configuration has had a refresh, it will now
3169 recognise x86_64 architectures automatically. You can still decide
3170 to build for a different bitness with the environment variable
3171 KERNEL_BITS (can be 32 or 64), for example:
3173 KERNEL_BITS=32 ./config
3177 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3178 256 bit AES and HMAC with SHA256.
3182 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3186 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3190 * To enable users to have their own config files and build file templates,
3191 Configure looks in the directory indicated by the environment variable
3192 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3193 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3194 name and is used as is.
3198 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3199 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3200 X509_CERT_FILE_CTX was removed.
3204 * "shared" builds are now the default. To create only static libraries use
3205 the "no-shared" Configure option.
3209 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3210 All of these option have not worked for some while and are fundamental
3215 * Make various cleanup routines no-ops and mark them as deprecated. Most
3216 global cleanup functions are no longer required because they are handled
3217 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3218 Explicitly de-initing can cause problems (e.g. where a library that uses
3219 OpenSSL de-inits, but an application is still using it). The affected
3220 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3221 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3222 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3223 COMP_zlib_cleanup().
3227 * --strict-warnings no longer enables runtime debugging options
3228 such as REF_DEBUG. Instead, debug options are automatically
3229 enabled with '--debug' builds.
3231 *Andy Polyakov, Emilia Käsper*
3233 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3234 have been moved out of the public header files. New functions for managing
3235 these have been added.
3239 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3240 objects have been moved out of the public header files. New
3241 functions for managing these have been added.
3245 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3246 have been moved out of the public header files. New functions for managing
3247 these have been added.
3251 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3252 moved out of the public header files. New functions for managing these
3257 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3261 * Removed the mk1mf build scripts.
3265 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3266 it is always safe to #include a header now.
3270 * Removed the aged BC-32 config and all its supporting scripts
3274 * Removed support for Ultrix, Netware, and OS/2.
3278 * Add support for HKDF.
3280 *Alessandro Ghedini*
3282 * Add support for blake2b and blake2s
3286 * Added support for "pipelining". Ciphers that have the
3287 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3288 encryptions/decryptions simultaneously. There are currently no built-in
3289 ciphers with this property but the expectation is that engines will be able
3290 to offer it to significantly improve throughput. Support has been extended
3291 into libssl so that multiple records for a single connection can be
3292 processed in one go (for >=TLS 1.1).
3296 * Added the AFALG engine. This is an async capable engine which is able to
3297 offload work to the Linux kernel. In this initial version it only supports
3298 AES128-CBC. The kernel must be version 4.1.0 or greater.
3302 * OpenSSL now uses a new threading API. It is no longer necessary to
3303 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3304 are two supported threading models: pthreads and windows threads. It is
3305 also possible to configure OpenSSL at compile time for "no-threads". The
3306 old threading API should no longer be used. The functions have been
3307 replaced with "no-op" compatibility macros.
3309 *Alessandro Ghedini, Matt Caswell*
3311 * Modify behavior of ALPN to invoke callback after SNI/servername
3312 callback, such that updates to the SSL_CTX affect ALPN.
3316 * Add SSL_CIPHER queries for authentication and key-exchange.
3320 * Changes to the DEFAULT cipherlist:
3321 - Prefer (EC)DHE handshakes over plain RSA.
3322 - Prefer AEAD ciphers over legacy ciphers.
3323 - Prefer ECDSA over RSA when both certificates are available.
3324 - Prefer TLSv1.2 ciphers/PRF.
3325 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3330 * Change the ECC default curve list to be this, in order: x25519,
3331 secp256r1, secp521r1, secp384r1.
3335 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3336 disabled by default. They can be re-enabled using the
3337 enable-weak-ssl-ciphers option to Configure.
3341 * If the server has ALPN configured, but supports no protocols that the
3342 client advertises, send a fatal "no_application_protocol" alert.
3343 This behaviour is SHALL in RFC 7301, though it isn't universally
3344 implemented by other servers.
3348 * Add X25519 support.
3349 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3350 for public and private key encoding using the format documented in
3351 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3352 key generation and key derivation.
3354 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3359 * Deprecate SRP_VBASE_get_by_user.
3360 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3361 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3362 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3363 seed, even if the seed is configured.
3365 Users should use SRP_VBASE_get1_by_user instead. Note that in
3366 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3367 also that even though configuring the SRP seed attempts to hide
3368 invalid usernames by continuing the handshake with fake
3369 credentials, this behaviour is not constant time and no strong
3370 guarantees are made that the handshake is indistinguishable from
3371 that of a valid user.
3375 * Configuration change; it's now possible to build dynamic engines
3376 without having to build shared libraries and vice versa. This
3377 only applies to the engines in `engines/`, those in `crypto/engine/`
3378 will always be built into libcrypto (i.e. "static").
3380 Building dynamic engines is enabled by default; to disable, use
3381 the configuration option "disable-dynamic-engine".
3383 The only requirements for building dynamic engines are the
3384 presence of the DSO module and building with position independent
3385 code, so they will also automatically be disabled if configuring
3386 with "disable-dso" or "disable-pic".
3388 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3389 are also taken away from openssl/opensslconf.h, as they are
3394 * Configuration change; if there is a known flag to compile
3395 position independent code, it will always be applied on the
3396 libcrypto and libssl object files, and never on the application
3397 object files. This means other libraries that use routines from
3398 libcrypto / libssl can be made into shared libraries regardless
3399 of how OpenSSL was configured.
3401 If this isn't desirable, the configuration options "disable-pic"
3402 or "no-pic" can be used to disable the use of PIC. This will
3403 also disable building shared libraries and dynamic engines.
3407 * Removed JPAKE code. It was experimental and has no wide use.
3411 * The INSTALL_PREFIX Makefile variable has been renamed to
3412 DESTDIR. That makes for less confusion on what this variable
3413 is for. Also, the configuration option --install_prefix is
3418 * Heartbeat for TLS has been removed and is disabled by default
3419 for DTLS; configure with enable-heartbeats. Code that uses the
3420 old #define's might need to be updated.
3422 *Emilia Käsper, Rich Salz*
3424 * Rename REF_CHECK to REF_DEBUG.
3428 * New "unified" build system
3430 The "unified" build system is aimed to be a common system for all
3431 platforms we support. With it comes new support for VMS.
3433 This system builds supports building in a different directory tree
3434 than the source tree. It produces one Makefile (for unix family
3435 or lookalikes), or one descrip.mms (for VMS).
3437 The source of information to make the Makefile / descrip.mms is
3438 small files called 'build.info', holding the necessary
3439 information for each directory with source to compile, and a
3440 template in Configurations, like unix-Makefile.tmpl or
3443 With this change, the library names were also renamed on Windows
3444 and on VMS. They now have names that are closer to the standard
3445 on Unix, and include the major version number, and in certain
3446 cases, the architecture they are built for. See "Notes on shared
3447 libraries" in INSTALL.
3449 We rely heavily on the perl module Text::Template.
3453 * Added support for auto-initialisation and de-initialisation of the library.
3454 OpenSSL no longer requires explicit init or deinit routines to be called,
3455 except in certain circumstances. See the OPENSSL_init_crypto() and
3456 OPENSSL_init_ssl() man pages for further information.
3460 * The arguments to the DTLSv1_listen function have changed. Specifically the
3461 "peer" argument is now expected to be a BIO_ADDR object.
3463 * Rewrite of BIO networking library. The BIO library lacked consistent
3464 support of IPv6, and adding it required some more extensive
3465 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3466 which hold all types of addresses and chains of address information.
3467 It also introduces a new API, with functions like BIO_socket,
3468 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3469 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3470 have been adapted accordingly.
3474 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3479 * CRIME protection: disable compression by default, even if OpenSSL is
3480 compiled with zlib enabled. Applications can still enable compression
3481 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3482 using the SSL_CONF library to configure compression.
3486 * The signature of the session callback configured with
3487 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3488 was explicitly marked as `const unsigned char*` instead of
3493 * Always DPURIFY. Remove the use of uninitialized memory in the
3494 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3498 * Removed many obsolete configuration items, including
3499 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3500 MD2_CHAR, MD2_INT, MD2_LONG
3502 IDEA_SHORT, IDEA_LONG
3503 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3505 *Rich Salz, with advice from Andy Polyakov*
3507 * Many BN internals have been moved to an internal header file.
3509 *Rich Salz with help from Andy Polyakov*
3511 * Configuration and writing out the results from it has changed.
3512 Files such as Makefile include/openssl/opensslconf.h and are now
3513 produced through general templates, such as Makefile.in and
3514 crypto/opensslconf.h.in and some help from the perl module
3517 Also, the center of configuration information is no longer
3518 Makefile. Instead, Configure produces a perl module in
3519 configdata.pm which holds most of the config data (in the hash
3520 table %config), the target data that comes from the target
3521 configuration in one of the `Configurations/*.conf` files (in
3526 * To clarify their intended purposes, the Configure options
3527 --prefix and --openssldir change their semantics, and become more
3528 straightforward and less interdependent.
3530 --prefix shall be used exclusively to give the location INSTALLTOP
3531 where programs, scripts, libraries, include files and manuals are
3532 going to be installed. The default is now /usr/local.
3534 --openssldir shall be used exclusively to give the default
3535 location OPENSSLDIR where certificates, private keys, CRLs are
3536 managed. This is also where the default openssl.cnf gets
3538 If the directory given with this option is a relative path, the
3539 values of both the --prefix value and the --openssldir value will
3540 be combined to become OPENSSLDIR.
3541 The default for --openssldir is INSTALLTOP/ssl.
3543 Anyone who uses --openssldir to specify where OpenSSL is to be
3544 installed MUST change to use --prefix instead.
3548 * The GOST engine was out of date and therefore it has been removed. An up
3549 to date GOST engine is now being maintained in an external repository.
3550 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3551 support for GOST ciphersuites (these are only activated if a GOST engine
3556 * EGD is no longer supported by default; use enable-egd when
3559 *Ben Kaduk and Rich Salz*
3561 * The distribution now has Makefile.in files, which are used to
3562 create Makefile's when Configure is run. *Configure must be run
3563 before trying to build now.*
3567 * The return value for SSL_CIPHER_description() for error conditions
3572 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3574 Obtaining and performing DNSSEC validation of TLSA records is
3575 the application's responsibility. The application provides
3576 the TLSA records of its choice to OpenSSL, and these are then
3577 used to authenticate the peer.
3579 The TLSA records need not even come from DNS. They can, for
3580 example, be used to implement local end-entity certificate or
3581 trust-anchor "pinning", where the "pin" data takes the form
3582 of TLSA records, which can augment or replace verification
3583 based on the usual WebPKI public certification authorities.
3587 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3588 continues to support deprecated interfaces in default builds.
3589 However, applications are strongly advised to compile their
3590 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3591 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3592 or the 1.1.0 releases.
3594 In environments in which all applications have been ported to
3595 not use any deprecated interfaces OpenSSL's Configure script
3596 should be used with the --api=1.1.0 option to entirely remove
3597 support for the deprecated features from the library and
3598 unconditionally disable them in the installed headers.
3599 Essentially the same effect can be achieved with the "no-deprecated"
3600 argument to Configure, except that this will always restrict
3601 the build to just the latest API, rather than a fixed API
3604 As applications are ported to future revisions of the API,
3605 they should update their compile-time OPENSSL_API_COMPAT define
3606 accordingly, but in most cases should be able to continue to
3607 compile with later releases.
3609 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3610 0x10000000L and 0x00908000L, respectively. However those
3611 versions did not support the OPENSSL_API_COMPAT feature, and
3612 so applications are not typically tested for explicit support
3613 of just the undeprecated features of either release.
3617 * Add support for setting the minimum and maximum supported protocol.
3618 It can bet set via the SSL_set_min_proto_version() and
3619 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3620 MaxProtocol. It's recommended to use the new APIs to disable
3621 protocols instead of disabling individual protocols using
3622 SSL_set_options() or SSL_CONF's Protocol. This change also
3623 removes support for disabling TLS 1.2 in the OpenSSL TLS
3624 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3628 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3632 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3633 and integrates ECDSA and ECDH functionality into EC. Implementations can
3634 now redirect key generation and no longer need to convert to or from
3637 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3638 include the ec.h header file instead.
3642 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3643 ciphers who are no longer supported and drops support the ephemeral RSA key
3644 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3648 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3649 opaque. For HMAC_CTX, the following constructors and destructors
3652 HMAC_CTX *HMAC_CTX_new(void);
3653 void HMAC_CTX_free(HMAC_CTX *ctx);
3655 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3656 destroy such methods has been added. See EVP_MD_meth_new(3) and
3657 EVP_CIPHER_meth_new(3) for documentation.
3660 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3661 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3662 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3663 an already created structure.
3664 2) For consistency with the majority of our object creators and
3665 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3666 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3667 for deprecated builds.
3671 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3672 cryptographic operations to be performed asynchronously as long as an
3673 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3674 further details. Libssl has also had this capability integrated with the
3675 introduction of the new mode SSL_MODE_ASYNC and associated error
3676 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3677 pages. This work was developed in partnership with Intel Corp.
3681 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3682 always enabled now. If you want to disable the support you should
3683 exclude it using the list of supported ciphers. This also means that the
3684 "-no_ecdhe" option has been removed from s_server.
3688 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3689 SSL_{CTX_}set1_curves() which can set a list.
3693 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3694 curve you want to support using SSL_{CTX_}set1_curves().
3698 * State machine rewrite. The state machine code has been significantly
3699 refactored in order to remove much duplication of code and solve issues
3700 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3701 further details). This change does have some associated API changes.
3702 Notably the SSL_state() function has been removed and replaced by
3703 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3704 SSL_set_state() has been removed altogether. The previous handshake states
3705 defined in ssl.h and ssl3.h have also been removed.
3709 * All instances of the string "ssleay" in the public API were replaced
3710 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3711 Some error codes related to internal RSA_eay API's were renamed.
3715 * The demo files in crypto/threads were moved to demo/threads.
3719 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3722 *Matt Caswell, Rich Salz*
3724 * New ASN.1 embed macro.
3726 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3727 structure is not allocated: it is part of the parent. That is instead of
3735 This reduces memory fragmentation and make it impossible to accidentally
3736 set a mandatory field to NULL.
3738 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3739 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3740 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3745 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3749 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3750 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3751 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3752 DES and RC4 ciphersuites.
3756 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3757 This changes the decoding behaviour for some invalid messages,
3758 though the change is mostly in the more lenient direction, and
3759 legacy behaviour is preserved as much as possible.
3763 * Fix no-stdio build.
3764 *David Woodhouse <David.Woodhouse@intel.com> and also*
3765 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3767 * New testing framework
3768 The testing framework has been largely rewritten and is now using
3769 perl and the perl modules Test::Harness and an extended variant of
3770 Test::More called OpenSSL::Test to do its work. All test scripts in
3771 test/ have been rewritten into test recipes, and all direct calls to
3772 executables in test/Makefile have become individual recipes using the
3773 simplified testing OpenSSL::Test::Simple.
3775 For documentation on our testing modules, do:
3777 perldoc test/testlib/OpenSSL/Test/Simple.pm
3778 perldoc test/testlib/OpenSSL/Test.pm
3782 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3783 are used; the latter aborts on memory leaks (usually checked on exit).
3784 Some undocumented "set malloc, etc., hooks" functions were removed
3785 and others were changed. All are now documented.
3789 * In DSA_generate_parameters_ex, if the provided seed is too short,
3792 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3794 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3795 from RFC4279, RFC4785, RFC5487, RFC5489.
3797 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3798 original RSA_PSK patch.
3802 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3803 era flag was never set throughout the codebase (only read). Also removed
3804 SSL3_FLAGS_POP_BUFFER which was only used if
3805 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3809 * Changed the default name options in the "ca", "crl", "req" and "x509"
3810 to be "oneline" instead of "compat".
3814 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3815 not aware of clients that still exhibit this bug, and the workaround
3816 hasn't been working properly for a while.
3820 * The return type of BIO_number_read() and BIO_number_written() as well as
3821 the corresponding num_read and num_write members in the BIO structure has
3822 changed from unsigned long to uint64_t. On platforms where an unsigned
3823 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3828 * Given the pervasive nature of TLS extensions it is inadvisable to run
3829 OpenSSL without support for them. It also means that maintaining
3830 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3831 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3835 * Removed support for the two export grade static DH ciphersuites
3836 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3837 were newly added (along with a number of other static DH ciphersuites) to
3838 1.0.2. However the two export ones have *never* worked since they were
3839 introduced. It seems strange in any case to be adding new export
3840 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3844 * Version negotiation has been rewritten. In particular SSLv23_method(),
3845 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3846 and turned into macros which simply call the new preferred function names
3847 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3848 should use the new names instead. Also as part of this change the ssl23.h
3849 header file has been removed.
3853 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3854 code and the associated standard is no longer considered fit-for-purpose.
3858 * RT2547 was closed. When generating a private key, try to make the
3859 output file readable only by the owner. This behavior change might
3860 be noticeable when interacting with other software.
3862 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3867 * Added HTTP GET support to the ocsp command.
3871 * Changed default digest for the dgst and enc commands from MD5 to
3876 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3880 * Added support for TLS extended master secret from
3881 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3882 initial patch which was a great help during development.
3886 * All libssl internal structures have been removed from the public header
3887 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3888 now redundant). Users should not attempt to access internal structures
3889 directly. Instead they should use the provided API functions.
3893 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3894 Access to deprecated functions can be re-enabled by running config with
3895 "enable-deprecated". In addition applications wishing to use deprecated
3896 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3897 will, by default, disable some transitive includes that previously existed
3898 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3902 * Added support for OCB mode. OpenSSL has been granted a patent license
3903 compatible with the OpenSSL license for use of OCB. Details are available
3904 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3905 for OCB can be removed by calling config with no-ocb.
3909 * SSLv2 support has been removed. It still supports receiving a SSLv2
3910 compatible client hello.
3914 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3915 done while fixing the error code for the key-too-small case.
3917 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3919 * CA.sh has been removed; use CA.pl instead.
3923 * Removed old DES API.
3927 * Remove various unsupported platforms:
3933 Sinix/ReliantUNIX RM400
3938 16-bit platforms such as WIN16
3942 * Clean up OPENSSL_NO_xxx #define's
3943 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3944 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3945 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3946 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3947 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3948 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3949 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3950 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3951 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3952 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3956 * Cleaned up dead code
3957 Remove all but one '#ifdef undef' which is to be looked at.
3961 * Clean up calling of xxx_free routines.
3962 Just like free(), fix most of the xxx_free routines to accept
3963 NULL. Remove the non-null checks from callers. Save much code.
3967 * Add secure heap for storage of private keys (when possible).
3968 Add BIO_s_secmem(), CBIGNUM, etc.
3969 Contributed by Akamai Technologies under our Corporate CLA.
3973 * Experimental support for a new, fast, unbiased prime candidate generator,
3974 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3976 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3978 * New output format NSS in the sess_id command line tool. This allows
3979 exporting the session id and the master key in NSS keylog format.
3981 *Martin Kaiser <martin@kaiser.cx>*
3983 * Harmonize version and its documentation. -f flag is used to display
3986 *mancha <mancha1@zoho.com>*
3988 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3989 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3991 *mancha <mancha1@zoho.com>*
3993 * Fix some double frees. These are not thought to be exploitable.
3995 *mancha <mancha1@zoho.com>*
3997 * A missing bounds check in the handling of the TLS heartbeat extension
3998 can be used to reveal up to 64k of memory to a connected client or
4001 Thanks for Neel Mehta of Google Security for discovering this bug and to
4002 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4003 preparing the fix ([CVE-2014-0160])
4005 *Adam Langley, Bodo Moeller*
4007 * Fix for the attack described in the paper "Recovering OpenSSL
4008 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4009 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4010 <http://eprint.iacr.org/2014/140>
4012 Thanks to Yuval Yarom and Naomi Benger for discovering this
4013 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4015 *Yuval Yarom and Naomi Benger*
4017 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4018 this fixes a limitation in previous versions of OpenSSL.
4022 * Experimental encrypt-then-mac support.
4024 Experimental support for encrypt then mac from
4025 draft-gutmann-tls-encrypt-then-mac-02.txt
4027 To enable it set the appropriate extension number (0x42 for the test
4028 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4030 For non-compliant peers (i.e. just about everything) this should have no
4033 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4037 * Add EVP support for key wrapping algorithms, to avoid problems with
4038 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4039 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4040 algorithms and include tests cases.
4044 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4049 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4050 MGF1 digest and OAEP label.
4054 * Make openssl verify return errors.
4056 *Chris Palmer <palmer@google.com> and Ben Laurie*
4058 * New function ASN1_TIME_diff to calculate the difference between two
4059 ASN1_TIME structures or one structure and the current time.
4063 * Update fips_test_suite to support multiple command line options. New
4064 test to induce all self test errors in sequence and check expected
4069 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4070 sign or verify all in one operation.
4074 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4075 test programs and fips_test_suite. Includes functionality to parse
4076 the minimal script output of fipsalgest.pl directly.
4080 * Add authorisation parameter to FIPS_module_mode_set().
4084 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4088 * Use separate DRBG fields for internal and external flags. New function
4089 FIPS_drbg_health_check() to perform on demand health checking. Add
4090 generation tests to fips_test_suite with reduced health check interval to
4091 demonstrate periodic health checking. Add "nodh" option to
4092 fips_test_suite to skip very slow DH test.
4096 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4101 * More extensive health check for DRBG checking many more failure modes.
4102 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4103 combination: call this in fips_test_suite.
4107 * Add support for canonical generation of DSA parameter 'g'. See
4110 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4111 POST to handle HMAC cases.
4115 * Add functions FIPS_module_version() and FIPS_module_version_text()
4116 to return numerical and string versions of the FIPS module number.
4120 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4121 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4122 outside the validated module in the FIPS capable OpenSSL.
4126 * Minor change to DRBG entropy callback semantics. In some cases
4127 there is no multiple of the block length between min_len and
4128 max_len. Allow the callback to return more than max_len bytes
4129 of entropy but discard any extra: it is the callback's responsibility
4130 to ensure that the extra data discarded does not impact the
4131 requested amount of entropy.
4135 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4136 information in FIPS186-3, SP800-57 and SP800-131A.
4140 * CCM support via EVP. Interface is very similar to GCM case except we
4141 must supply all data in one chunk (i.e. no update, final) and the
4142 message length must be supplied if AAD is used. Add algorithm test
4147 * Initial version of POST overhaul. Add POST callback to allow the status
4148 of POST to be monitored and/or failures induced. Modify fips_test_suite
4149 to use callback. Always run all selftests even if one fails.
4153 * XTS support including algorithm test driver in the fips_gcmtest program.
4154 Note: this does increase the maximum key length from 32 to 64 bytes but
4155 there should be no binary compatibility issues as existing applications
4156 will never use XTS mode.
4160 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4161 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4162 performs algorithm blocking for unapproved PRNG types. Also do not
4163 set PRNG type in FIPS_mode_set(): leave this to the application.
4164 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4165 the standard OpenSSL PRNG: set additional data to a date time vector.
4169 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4170 This shouldn't present any incompatibility problems because applications
4171 shouldn't be using these directly and any that are will need to rethink
4172 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4176 * Extensive self tests and health checking required by SP800-90 DRBG.
4177 Remove strength parameter from FIPS_drbg_instantiate and always
4178 instantiate at maximum supported strength.
4182 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4186 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4190 * New function DH_compute_key_padded() to compute a DH key and pad with
4191 leading zeroes if needed: this complies with SP800-56A et al.
4195 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4196 anything, incomplete, subject to change and largely untested at present.
4200 * Modify fipscanisteronly build option to only build the necessary object
4201 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4205 * Add experimental option FIPSSYMS to give all symbols in
4206 fipscanister.o and FIPS or fips prefix. This will avoid
4207 conflicts with future versions of OpenSSL. Add perl script
4208 util/fipsas.pl to preprocess assembly language source files
4209 and rename any affected symbols.
4213 * Add selftest checks and algorithm block of non-fips algorithms in
4214 FIPS mode. Remove DES2 from selftests.
4218 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4219 return internal method without any ENGINE dependencies. Add new
4220 tiny fips sign and verify functions.
4224 * New build option no-ec2m to disable characteristic 2 code.
4228 * New build option "fipscanisteronly". This only builds fipscanister.o
4229 and (currently) associated fips utilities. Uses the file Makefile.fips
4230 instead of Makefile.org as the prototype.
4234 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4235 Update fips_gcmtest to use IV generator.
4239 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4240 setting output buffer to NULL. The `*Final` function must be
4241 called although it will not retrieve any additional data. The tag
4242 can be set or retrieved with a ctrl. The IV length is by default 12
4243 bytes (96 bits) but can be set to an alternative value. If the IV
4244 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4249 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4250 underlying do_cipher function handles all cipher semantics itself
4251 including padding and finalisation. This is useful if (for example)
4252 an ENGINE cipher handles block padding itself. The behaviour of
4253 do_cipher is subtly changed if this flag is set: the return value
4254 is the number of characters written to the output buffer (zero is
4255 no longer an error code) or a negative error code. Also if the
4256 input buffer is NULL and length 0 finalisation should be performed.
4260 * If a candidate issuer certificate is already part of the constructed
4261 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4265 * Improve forward-security support: add functions
4267 void SSL_CTX_set_not_resumable_session_callback(
4268 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4269 void SSL_set_not_resumable_session_callback(
4270 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4272 for use by SSL/TLS servers; the callback function will be called whenever a
4273 new session is created, and gets to decide whether the session may be
4274 cached to make it resumable (return 0) or not (return 1). (As by the
4275 SSL/TLS protocol specifications, the session_id sent by the server will be
4276 empty to indicate that the session is not resumable; also, the server will
4277 not generate RFC 4507 (RFC 5077) session tickets.)
4279 A simple reasonable callback implementation is to return is_forward_secure.
4280 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4281 by the SSL/TLS server library, indicating whether it can provide forward
4284 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4286 * New -verify_name option in command line utilities to set verification
4291 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4292 Add CMAC pkey methods.
4296 * Experimental renegotiation in s_server -www mode. If the client
4297 browses /reneg connection is renegotiated. If /renegcert it is
4298 renegotiated requesting a certificate.
4302 * Add an "external" session cache for debugging purposes to s_server. This
4303 should help trace issues which normally are only apparent in deployed
4304 multi-process servers.
4308 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4309 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4310 BIO_set_cipher() and some obscure PEM functions were changed so they
4311 can now return an error. The RAND changes required a change to the
4312 RAND_METHOD structure.
4316 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4317 a gcc attribute to warn if the result of a function is ignored. This
4318 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4319 whose return value is often ignored.
4323 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4324 These allow SCTs (signed certificate timestamps) to be requested and
4325 validated when establishing a connection.
4327 *Rob Percival <robpercival@google.com>*
4332 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4334 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4335 used even when parsing explicit parameters, when loading a encoded key
4336 or calling `EC_GROUP_new_from_ecpkparameters()`/
4337 `EC_GROUP_new_from_ecparameters()`.
4338 This prevents bypass of security hardening and performance gains,
4339 especially for curves with specialized EC_METHODs.
4340 By default, if a key encoded with explicit parameters is loaded and later
4341 encoded, the output is still encoded with explicit parameters, even if
4342 internally a "named" EC_GROUP is used for computation.
4346 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4347 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4348 NULL. After this change, only the cofactor parameter can be NULL. It also
4349 does some minimal sanity checks on the passed order.
4354 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4355 An attack is simple, if the first CMS_recipientInfo is valid but the
4356 second CMS_recipientInfo is chosen ciphertext. If the second
4357 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4358 encryption key will be replaced by garbage, and the message cannot be
4359 decoded, but if the RSA decryption fails, the correct encryption key is
4360 used and the recipient will not notice the attack.
4361 As a work around for this potential attack the length of the decrypted
4362 key must be equal to the cipher default key length, in case the
4363 certifiate is not given and all recipientInfo are tried out.
4364 The old behaviour can be re-enabled in the CMS code by setting the
4365 CMS_DEBUG_DECRYPT flag.
4370 * Document issue with installation paths in diverse Windows builds
4372 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4373 binaries and run-time config file.
4378 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4380 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4381 This changes the size when using the `genpkey` command when no size is given.
4382 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4383 generation commands to use 2048 bits by default.
4387 * Add FIPS support for Android Arm 64-bit
4389 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4390 Module in Version 2.0.10. For some reason, the corresponding target
4391 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4392 built with FIPS support on Android Arm 64-bit. This omission has been
4395 *Matthias St. Pierre*
4397 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4399 * 0-byte record padding oracle
4401 If an application encounters a fatal protocol error and then calls
4402 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4403 then OpenSSL can respond differently to the calling application if a 0 byte
4404 record is received with invalid padding compared to if a 0 byte record is
4405 received with an invalid MAC. If the application then behaves differently
4406 based on that in a way that is detectable to the remote peer, then this
4407 amounts to a padding oracle that could be used to decrypt data.
4409 In order for this to be exploitable "non-stitched" ciphersuites must be in
4410 use. Stitched ciphersuites are optimised implementations of certain
4411 commonly used ciphersuites. Also the application must call SSL_shutdown()
4412 twice even if a protocol error has occurred (applications should not do
4413 this but some do anyway).
4415 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4416 Aviram, with additional investigation by Steven Collison and Andrew
4417 Hourselt. It was reported to OpenSSL on 10th December 2018.
4422 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4426 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4428 * Microarchitecture timing vulnerability in ECC scalar multiplication
4430 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4431 shown to be vulnerable to a microarchitecture timing side channel attack.
4432 An attacker with sufficient access to mount local timing attacks during
4433 ECDSA signature generation could recover the private key.
4435 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4436 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4442 * Timing vulnerability in DSA signature generation
4444 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4445 timing side channel attack. An attacker could use variations in the signing
4446 algorithm to recover the private key.
4448 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4453 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4454 Module, accidentally introduced while backporting security fixes from the
4455 development branch and hindering the use of ECC in FIPS mode.
4459 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4461 * Client DoS due to large DH parameter
4463 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4464 malicious server can send a very large prime value to the client. This will
4465 cause the client to spend an unreasonably long period of time generating a
4466 key for this prime resulting in a hang until the client has finished. This
4467 could be exploited in a Denial Of Service attack.
4469 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4474 * Cache timing vulnerability in RSA Key Generation
4476 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4477 a cache timing side channel attack. An attacker with sufficient access to
4478 mount cache timing attacks during the RSA key generation process could
4479 recover the private key.
4481 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4482 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4487 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4488 parameter is no longer accepted, as it leads to a corrupt table. NULL
4489 pem_str is reserved for alias entries only.
4493 * Revert blinding in ECDSA sign and instead make problematic addition
4494 length-invariant. Switch even to fixed-length Montgomery multiplication.
4498 * Change generating and checking of primes so that the error rate of not
4499 being prime depends on the intended use based on the size of the input.
4500 For larger primes this will result in more rounds of Miller-Rabin.
4501 The maximal error rate for primes with more than 1080 bits is lowered
4504 *Kurt Roeckx, Annie Yousar*
4506 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4510 * Add blinding to ECDSA and DSA signatures to protect against side channel
4511 attacks discovered by Keegan Ryan (NCC Group).
4515 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4516 now allow empty (zero character) pass phrases.
4520 * Certificate time validation (X509_cmp_time) enforces stricter
4521 compliance with RFC 5280. Fractional seconds and timezone offsets
4522 are no longer allowed.
4526 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4528 * Constructed ASN.1 types with a recursive definition could exceed the stack
4530 Constructed ASN.1 types with a recursive definition (such as can be found
4531 in PKCS7) could eventually exceed the stack given malicious input with
4532 excessive recursion. This could result in a Denial Of Service attack. There
4533 are no such structures used within SSL/TLS that come from untrusted sources
4534 so this is considered safe.
4536 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4542 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4544 * Read/write after SSL object in error state
4546 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4547 mechanism. The intent was that if a fatal error occurred during a handshake
4548 then OpenSSL would move into the error state and would immediately fail if
4549 you attempted to continue the handshake. This works as designed for the
4550 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4551 SSL_connect()), however due to a bug it does not work correctly if
4552 SSL_read() or SSL_write() is called directly. In that scenario, if the
4553 handshake fails then a fatal error will be returned in the initial function
4554 call. If SSL_read()/SSL_write() is subsequently called by the application
4555 for the same SSL object then it will succeed and the data is passed without
4556 being decrypted/encrypted directly from the SSL/TLS record layer.
4558 In order to exploit this issue an application bug would have to be present
4559 that resulted in a call to SSL_read()/SSL_write() being issued after having
4560 already received a fatal error.
4562 This issue was reported to OpenSSL by David Benjamin (Google).
4567 * rsaz_1024_mul_avx2 overflow bug on x86_64
4569 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4570 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4571 Analysis suggests that attacks against RSA and DSA as a result of this
4572 defect would be very difficult to perform and are not believed likely.
4573 Attacks against DH1024 are considered just feasible, because most of the
4574 work necessary to deduce information about a private key may be performed
4575 offline. The amount of resources required for such an attack would be
4576 significant. However, for an attack on TLS to be meaningful, the server
4577 would have to share the DH1024 private key among multiple clients, which is
4578 no longer an option since CVE-2016-0701.
4580 This only affects processors that support the AVX2 but not ADX extensions
4581 like Intel Haswell (4th generation).
4583 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4584 was originally found via the OSS-Fuzz project.
4589 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4591 * bn_sqrx8x_internal carry bug on x86_64
4593 There is a carry propagating bug in the x86_64 Montgomery squaring
4594 procedure. No EC algorithms are affected. Analysis suggests that attacks
4595 against RSA and DSA as a result of this defect would be very difficult to
4596 perform and are not believed likely. Attacks against DH are considered just
4597 feasible (although very difficult) because most of the work necessary to
4598 deduce information about a private key may be performed offline. The amount
4599 of resources required for such an attack would be very significant and
4600 likely only accessible to a limited number of attackers. An attacker would
4601 additionally need online access to an unpatched system using the target
4602 private key in a scenario with persistent DH parameters and a private
4603 key that is shared between multiple clients.
4605 This only affects processors that support the BMI1, BMI2 and ADX extensions
4606 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4608 This issue was reported to OpenSSL by the OSS-Fuzz project.
4613 * Malformed X.509 IPAddressFamily could cause OOB read
4615 If an X.509 certificate has a malformed IPAddressFamily extension,
4616 OpenSSL could do a one-byte buffer overread. The most likely result
4617 would be an erroneous display of the certificate in text format.
4619 This issue was reported to OpenSSL by the OSS-Fuzz project.
4624 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4626 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4627 platform rather than 'mingw'.
4631 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4633 * Truncated packet could crash via OOB read
4635 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4636 cipher is being used, then a truncated packet can cause that host to
4637 perform an out-of-bounds read, usually resulting in a crash.
4639 This issue was reported to OpenSSL by Robert Święcki of Google.
4644 * BN_mod_exp may produce incorrect results on x86_64
4646 There is a carry propagating bug in the x86_64 Montgomery squaring
4647 procedure. No EC algorithms are affected. Analysis suggests that attacks
4648 against RSA and DSA as a result of this defect would be very difficult to
4649 perform and are not believed likely. Attacks against DH are considered just
4650 feasible (although very difficult) because most of the work necessary to
4651 deduce information about a private key may be performed offline. The amount
4652 of resources required for such an attack would be very significant and
4653 likely only accessible to a limited number of attackers. An attacker would
4654 additionally need online access to an unpatched system using the target
4655 private key in a scenario with persistent DH parameters and a private
4656 key that is shared between multiple clients. For example this can occur by
4657 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4658 similar to CVE-2015-3193 but must be treated as a separate problem.
4660 This issue was reported to OpenSSL by the OSS-Fuzz project.
4665 * Montgomery multiplication may produce incorrect results
4667 There is a carry propagating bug in the Broadwell-specific Montgomery
4668 multiplication procedure that handles input lengths divisible by, but
4669 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4670 and DH private keys are impossible. This is because the subroutine in
4671 question is not used in operations with the private key itself and an input
4672 of the attacker's direct choice. Otherwise the bug can manifest itself as
4673 transient authentication and key negotiation failures or reproducible
4674 erroneous outcome of public-key operations with specially crafted input.
4675 Among EC algorithms only Brainpool P-512 curves are affected and one
4676 presumably can attack ECDH key negotiation. Impact was not analyzed in
4677 detail, because pre-requisites for attack are considered unlikely. Namely
4678 multiple clients have to choose the curve in question and the server has to
4679 share the private key among them, neither of which is default behaviour.
4680 Even then only clients that chose the curve will be affected.
4682 This issue was publicly reported as transient failures and was not
4683 initially recognized as a security issue. Thanks to Richard Morgan for
4684 providing reproducible case.
4689 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4690 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4691 prevent issues where no progress is being made and the peer continually
4692 sends unrecognised record types, using up resources processing them.
4696 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4698 * Missing CRL sanity check
4700 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4701 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4702 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4704 This issue only affects the OpenSSL 1.0.2i
4709 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4711 * OCSP Status Request extension unbounded memory growth
4713 A malicious client can send an excessively large OCSP Status Request
4714 extension. If that client continually requests renegotiation, sending a
4715 large OCSP Status Request extension each time, then there will be unbounded
4716 memory growth on the server. This will eventually lead to a Denial Of
4717 Service attack through memory exhaustion. Servers with a default
4718 configuration are vulnerable even if they do not support OCSP. Builds using
4719 the "no-ocsp" build time option are not affected.
4721 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4726 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4729 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4735 * OOB write in MDC2_Update()
4737 An overflow can occur in MDC2_Update() either if called directly or
4738 through the EVP_DigestUpdate() function using MDC2. If an attacker
4739 is able to supply very large amounts of input data after a previous
4740 call to EVP_EncryptUpdate() with a partial block then a length check
4741 can overflow resulting in a heap corruption.
4743 The amount of data needed is comparable to SIZE_MAX which is impractical
4746 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4751 * Malformed SHA512 ticket DoS
4753 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4754 DoS attack where a malformed ticket will result in an OOB read which will
4757 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4758 a custom server callback and ticket lookup mechanism.
4760 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4765 * OOB write in BN_bn2dec()
4767 The function BN_bn2dec() does not check the return value of BN_div_word().
4768 This can cause an OOB write if an application uses this function with an
4769 overly large BIGNUM. This could be a problem if an overly large certificate
4770 or CRL is printed out from an untrusted source. TLS is not affected because
4771 record limits will reject an oversized certificate before it is parsed.
4773 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4778 * OOB read in TS_OBJ_print_bio()
4780 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4781 the total length the OID text representation would use and not the amount
4782 of data written. This will result in OOB reads when large OIDs are
4785 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4790 * Pointer arithmetic undefined behaviour
4792 Avoid some undefined pointer arithmetic
4794 A common idiom in the codebase is to check limits in the following manner:
4797 Where "p" points to some malloc'd data of SIZE bytes and
4800 "len" here could be from some externally supplied data (e.g. from a TLS
4803 The rules of C pointer arithmetic are such that "p + len" is only well
4804 defined where len <= SIZE. Therefore the above idiom is actually
4805 undefined behaviour.
4807 For example this could cause problems if some malloc implementation
4808 provides an address for "p" such that "p + len" actually overflows for
4809 values of len that are too big and therefore p + len < limit.
4811 This issue was reported to OpenSSL by Guido Vranken
4816 * Constant time flag not preserved in DSA signing
4818 Operations in the DSA signing algorithm should run in constant time in
4819 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4820 implementation means that a non-constant time codepath is followed for
4821 certain operations. This has been demonstrated through a cache-timing
4822 attack to be sufficient for an attacker to recover the private DSA key.
4824 This issue was reported by César Pereida (Aalto University), Billy Brumley
4825 (Tampere University of Technology), and Yuval Yarom (The University of
4826 Adelaide and NICTA).
4831 * DTLS buffered message DoS
4833 In a DTLS connection where handshake messages are delivered out-of-order
4834 those messages that OpenSSL is not yet ready to process will be buffered
4835 for later use. Under certain circumstances, a flaw in the logic means that
4836 those messages do not get removed from the buffer even though the handshake
4837 has been completed. An attacker could force up to approx. 15 messages to
4838 remain in the buffer when they are no longer required. These messages will
4839 be cleared when the DTLS connection is closed. The default maximum size for
4840 a message is 100k. Therefore the attacker could force an additional 1500k
4841 to be consumed per connection. By opening many simulataneous connections an
4842 attacker could cause a DoS attack through memory exhaustion.
4844 This issue was reported to OpenSSL by Quan Luo.
4849 * DTLS replay protection DoS
4851 A flaw in the DTLS replay attack protection mechanism means that records
4852 that arrive for future epochs update the replay protection "window" before
4853 the MAC for the record has been validated. This could be exploited by an
4854 attacker by sending a record for the next epoch (which does not have to
4855 decrypt or have a valid MAC), with a very large sequence number. This means
4856 that all subsequent legitimate packets are dropped causing a denial of
4857 service for a specific DTLS connection.
4859 This issue was reported to OpenSSL by the OCAP audit team.
4864 * Certificate message OOB reads
4866 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4867 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4868 theoretical DoS risk but this has not been observed in practice on common
4871 The messages affected are client certificate, client certificate request
4872 and server certificate. As a result the attack can only be performed
4873 against a client or a server which enables client authentication.
4875 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4880 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4882 * Prevent padding oracle in AES-NI CBC MAC check
4884 A MITM attacker can use a padding oracle attack to decrypt traffic
4885 when the connection uses an AES CBC cipher and the server support
4888 This issue was introduced as part of the fix for Lucky 13 padding
4889 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4890 constant time by making sure that always the same bytes are read and
4891 compared against either the MAC or padding bytes. But it no longer
4892 checked that there was enough data to have both the MAC and padding
4895 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4900 * Fix EVP_EncodeUpdate overflow
4902 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4903 Base64 encoding of binary data. If an attacker is able to supply very large
4904 amounts of input data then a length check can overflow resulting in a heap
4907 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4908 the `PEM_write_bio*` family of functions. These are mainly used within the
4909 OpenSSL command line applications, so any application which processes data
4910 from an untrusted source and outputs it as a PEM file should be considered
4911 vulnerable to this issue. User applications that call these APIs directly
4912 with large amounts of untrusted data may also be vulnerable.
4914 This issue was reported by Guido Vranken.
4919 * Fix EVP_EncryptUpdate overflow
4921 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4922 is able to supply very large amounts of input data after a previous call to
4923 EVP_EncryptUpdate() with a partial block then a length check can overflow
4924 resulting in a heap corruption. Following an analysis of all OpenSSL
4925 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4926 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4927 the first called function after an EVP_EncryptInit(), and therefore that
4928 specific call must be safe. The second form is where the length passed to
4929 EVP_EncryptUpdate() can be seen from the code to be some small value and
4930 therefore there is no possibility of an overflow. Since all instances are
4931 one of these two forms, it is believed that there can be no overflows in
4932 internal code due to this problem. It should be noted that
4933 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4934 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4935 of these calls have also been analysed too and it is believed there are no
4936 instances in internal usage where an overflow could occur.
4938 This issue was reported by Guido Vranken.
4943 * Prevent ASN.1 BIO excessive memory allocation
4945 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4946 a short invalid encoding can cause allocation of large amounts of memory
4947 potentially consuming excessive resources or exhausting memory.
4949 Any application parsing untrusted data through d2i BIO functions is
4950 affected. The memory based functions such as d2i_X509() are *not* affected.
4951 Since the memory based functions are used by the TLS library, TLS
4952 applications are not affected.
4954 This issue was reported by Brian Carpenter.
4961 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4962 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4963 in arbitrary stack data being returned in the buffer.
4965 This issue was reported by Guido Vranken.
4970 * Modify behavior of ALPN to invoke callback after SNI/servername
4971 callback, such that updates to the SSL_CTX affect ALPN.
4975 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4980 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4981 methods are enabled and ssl2 is disabled the methods return NULL.
4985 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4987 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4988 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4989 provide any "EXPORT" or "LOW" strength ciphers.
4993 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4994 is by default disabled at build-time. Builds that are not configured with
4995 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4996 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4997 will need to explicitly call either of:
4999 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5001 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5003 as appropriate. Even if either of those is used, or the application
5004 explicitly uses the version-specific SSLv2_method() or its client and
5005 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5006 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5007 ciphers, and SSLv2 56-bit DES are no longer available.
5012 * Fix a double-free in DSA code
5014 A double free bug was discovered when OpenSSL parses malformed DSA private
5015 keys and could lead to a DoS attack or memory corruption for applications
5016 that receive DSA private keys from untrusted sources. This scenario is
5019 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5025 * Disable SRP fake user seed to address a server memory leak.
5027 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5029 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5030 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5031 was changed to ignore the "fake user" SRP seed, even if the seed
5034 Users should use SRP_VBASE_get1_by_user instead. Note that in
5035 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5036 also that even though configuring the SRP seed attempts to hide
5037 invalid usernames by continuing the handshake with fake
5038 credentials, this behaviour is not constant time and no strong
5039 guarantees are made that the handshake is indistinguishable from
5040 that of a valid user.
5045 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5047 In the BN_hex2bn function the number of hex digits is calculated using an
5048 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5049 large values of `i` this can result in `bn_expand` not allocating any
5050 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5051 field as NULL leading to a subsequent NULL ptr deref. For very large values
5052 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5053 In this case memory is allocated to the internal BIGNUM data field, but it
5054 is insufficiently sized leading to heap corruption. A similar issue exists
5055 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5056 is ever called by user applications with very large untrusted hex/dec data.
5057 This is anticipated to be a rare occurrence.
5059 All OpenSSL internal usage of these functions use data that is not expected
5060 to be untrusted, e.g. config file data or application command line
5061 arguments. If user developed applications generate config file data based
5062 on untrusted data then it is possible that this could also lead to security
5063 consequences. This is also anticipated to be rare.
5065 This issue was reported to OpenSSL by Guido Vranken.
5070 * Fix memory issues in `BIO_*printf` functions
5072 The internal `fmtstr` function used in processing a "%s" format string in
5073 the `BIO_*printf` functions could overflow while calculating the length of a
5074 string and cause an OOB read when printing very long strings.
5076 Additionally the internal `doapr_outch` function can attempt to write to an
5077 OOB memory location (at an offset from the NULL pointer) in the event of a
5078 memory allocation failure. In 1.0.2 and below this could be caused where
5079 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5080 could be in processing a very long "%s" format string. Memory leaks can
5083 The first issue may mask the second issue dependent on compiler behaviour.
5084 These problems could enable attacks where large amounts of untrusted data
5085 is passed to the `BIO_*printf` functions. If applications use these functions
5086 in this way then they could be vulnerable. OpenSSL itself uses these
5087 functions when printing out human-readable dumps of ASN.1 data. Therefore
5088 applications that print this data could be vulnerable if the data is from
5089 untrusted sources. OpenSSL command line applications could also be
5090 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5091 as command line arguments.
5093 Libssl is not considered directly vulnerable. Additionally certificates etc
5094 received via remote connections via libssl are also unlikely to be able to
5095 trigger these issues because of message size limits enforced within libssl.
5097 This issue was reported to OpenSSL Guido Vranken.
5102 * Side channel attack on modular exponentiation
5104 A side-channel attack was found which makes use of cache-bank conflicts on
5105 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5106 of RSA keys. The ability to exploit this issue is limited as it relies on
5107 an attacker who has control of code in a thread running on the same
5108 hyper-threaded core as the victim thread which is performing decryptions.
5110 This issue was reported to OpenSSL by Yuval Yarom, The University of
5111 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5112 Nadia Heninger, University of Pennsylvania with more information at
5113 <http://cachebleed.info>.
5118 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5119 if no keysize is specified with default_bits. This fixes an
5120 omission in an earlier change that changed all RSA/DSA key generation
5121 commands to use 2048 bits by default.
5125 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5127 * DH small subgroups
5129 Historically OpenSSL only ever generated DH parameters based on "safe"
5130 primes. More recently (in version 1.0.2) support was provided for
5131 generating X9.42 style parameter files such as those required for RFC 5114
5132 support. The primes used in such files may not be "safe". Where an
5133 application is using DH configured with parameters based on primes that are
5134 not "safe" then an attacker could use this fact to find a peer's private
5135 DH exponent. This attack requires that the attacker complete multiple
5136 handshakes in which the peer uses the same private DH exponent. For example
5137 this could be used to discover a TLS server's private DH exponent if it's
5138 reusing the private DH exponent or it's using a static DH ciphersuite.
5140 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5141 TLS. It is not on by default. If the option is not set then the server
5142 reuses the same private DH exponent for the life of the server process and
5143 would be vulnerable to this attack. It is believed that many popular
5144 applications do set this option and would therefore not be at risk.
5146 The fix for this issue adds an additional check where a "q" parameter is
5147 available (as is the case in X9.42 based parameters). This detects the
5148 only known attack, and is the only possible defense for static DH
5149 ciphersuites. This could have some performance impact.
5151 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5152 default and cannot be disabled. This could have some performance impact.
5154 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5159 * SSLv2 doesn't block disabled ciphers
5161 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5162 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5163 been disabled, provided that the SSLv2 protocol was not also disabled via
5166 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5167 and Sebastian Schinzel.
5172 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5174 * BN_mod_exp may produce incorrect results on x86_64
5176 There is a carry propagating bug in the x86_64 Montgomery squaring
5177 procedure. No EC algorithms are affected. Analysis suggests that attacks
5178 against RSA and DSA as a result of this defect would be very difficult to
5179 perform and are not believed likely. Attacks against DH are considered just
5180 feasible (although very difficult) because most of the work necessary to
5181 deduce information about a private key may be performed offline. The amount
5182 of resources required for such an attack would be very significant and
5183 likely only accessible to a limited number of attackers. An attacker would
5184 additionally need online access to an unpatched system using the target
5185 private key in a scenario with persistent DH parameters and a private
5186 key that is shared between multiple clients. For example this can occur by
5187 default in OpenSSL DHE based SSL/TLS ciphersuites.
5189 This issue was reported to OpenSSL by Hanno Böck.
5194 * Certificate verify crash with missing PSS parameter
5196 The signature verification routines will crash with a NULL pointer
5197 dereference if presented with an ASN.1 signature using the RSA PSS
5198 algorithm and absent mask generation function parameter. Since these
5199 routines are used to verify certificate signature algorithms this can be
5200 used to crash any certificate verification operation and exploited in a
5201 DoS attack. Any application which performs certificate verification is
5202 vulnerable including OpenSSL clients and servers which enable client
5205 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5210 * X509_ATTRIBUTE memory leak
5212 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5213 memory. This structure is used by the PKCS#7 and CMS routines so any
5214 application which reads PKCS#7 or CMS data from untrusted sources is
5215 affected. SSL/TLS is not affected.
5217 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5223 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5224 This changes the decoding behaviour for some invalid messages,
5225 though the change is mostly in the more lenient direction, and
5226 legacy behaviour is preserved as much as possible.
5230 * In DSA_generate_parameters_ex, if the provided seed is too short,
5233 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5235 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5237 * Alternate chains certificate forgery
5239 During certificate verification, OpenSSL will attempt to find an
5240 alternative certificate chain if the first attempt to build such a chain
5241 fails. An error in the implementation of this logic can mean that an
5242 attacker could cause certain checks on untrusted certificates to be
5243 bypassed, such as the CA flag, enabling them to use a valid leaf
5244 certificate to act as a CA and "issue" an invalid certificate.
5246 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5251 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5253 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5254 incompatibility in the handling of HMAC. The previous ABI has now been
5259 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5261 * Malformed ECParameters causes infinite loop
5263 When processing an ECParameters structure OpenSSL enters an infinite loop
5264 if the curve specified is over a specially malformed binary polynomial
5267 This can be used to perform denial of service against any
5268 system which processes public keys, certificate requests or
5269 certificates. This includes TLS clients and TLS servers with
5270 client authentication enabled.
5272 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5277 * Exploitable out-of-bounds read in X509_cmp_time
5279 X509_cmp_time does not properly check the length of the ASN1_TIME
5280 string and can read a few bytes out of bounds. In addition,
5281 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5284 An attacker can use this to craft malformed certificates and CRLs of
5285 various sizes and potentially cause a segmentation fault, resulting in
5286 a DoS on applications that verify certificates or CRLs. TLS clients
5287 that verify CRLs are affected. TLS clients and servers with client
5288 authentication enabled may be affected if they use custom verification
5291 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5292 independently by Hanno Böck.
5297 * PKCS7 crash with missing EnvelopedContent
5299 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5300 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5301 with missing content and trigger a NULL pointer dereference on parsing.
5303 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5304 structures from untrusted sources are affected. OpenSSL clients and
5305 servers are not affected.
5307 This issue was reported to OpenSSL by Michal Zalewski (Google).
5312 * CMS verify infinite loop with unknown hash function
5314 When verifying a signedData message the CMS code can enter an infinite loop
5315 if presented with an unknown hash function OID. This can be used to perform
5316 denial of service against any system which verifies signedData messages using
5318 This issue was reported to OpenSSL by Johannes Bauer.
5323 * Race condition handling NewSessionTicket
5325 If a NewSessionTicket is received by a multi-threaded client when attempting to
5326 reuse a previous ticket then a race condition can occur potentially leading to
5327 a double free of the ticket data.
5332 * Only support 256-bit or stronger elliptic curves with the
5333 'ecdh_auto' setting (server) or by default (client). Of supported
5334 curves, prefer P-256 (both).
5338 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5340 * ClientHello sigalgs DoS fix
5342 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5343 invalid signature algorithms extension a NULL pointer dereference will
5344 occur. This can be exploited in a DoS attack against the server.
5346 This issue was was reported to OpenSSL by David Ramos of Stanford
5350 *Stephen Henson and Matt Caswell*
5352 * Multiblock corrupted pointer fix
5354 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5355 feature only applies on 64 bit x86 architecture platforms that support AES
5356 NI instructions. A defect in the implementation of "multiblock" can cause
5357 OpenSSL's internal write buffer to become incorrectly set to NULL when
5358 using non-blocking IO. Typically, when the user application is using a
5359 socket BIO for writing, this will only result in a failed connection.
5360 However if some other BIO is used then it is likely that a segmentation
5361 fault will be triggered, thus enabling a potential DoS attack.
5363 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5368 * Segmentation fault in DTLSv1_listen fix
5370 The DTLSv1_listen function is intended to be stateless and processes the
5371 initial ClientHello from many peers. It is common for user code to loop
5372 over the call to DTLSv1_listen until a valid ClientHello is received with
5373 an associated cookie. A defect in the implementation of DTLSv1_listen means
5374 that state is preserved in the SSL object from one invocation to the next
5375 that can lead to a segmentation fault. Errors processing the initial
5376 ClientHello can trigger this scenario. An example of such an error could be
5377 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5380 This issue was reported to OpenSSL by Per Allansson.
5385 * Segmentation fault in ASN1_TYPE_cmp fix
5387 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5388 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5389 certificate signature algorithm consistency this can be used to crash any
5390 certificate verification operation and exploited in a DoS attack. Any
5391 application which performs certificate verification is vulnerable including
5392 OpenSSL clients and servers which enable client authentication.
5397 * Segmentation fault for invalid PSS parameters fix
5399 The signature verification routines will crash with a NULL pointer
5400 dereference if presented with an ASN.1 signature using the RSA PSS
5401 algorithm and invalid parameters. Since these routines are used to verify
5402 certificate signature algorithms this can be used to crash any
5403 certificate verification operation and exploited in a DoS attack. Any
5404 application which performs certificate verification is vulnerable including
5405 OpenSSL clients and servers which enable client authentication.
5407 This issue was was reported to OpenSSL by Brian Carpenter.
5412 * ASN.1 structure reuse memory corruption fix
5414 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5415 memory corruption via an invalid write. Such reuse is and has been
5416 strongly discouraged and is believed to be rare.
5418 Applications that parse structures containing CHOICE or ANY DEFINED BY
5419 components may be affected. Certificate parsing (d2i_X509 and related
5420 functions) are however not affected. OpenSSL clients and servers are
5426 * PKCS7 NULL pointer dereferences fix
5428 The PKCS#7 parsing code does not handle missing outer ContentInfo
5429 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5430 missing content and trigger a NULL pointer dereference on parsing.
5432 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5433 otherwise parse PKCS#7 structures from untrusted sources are
5434 affected. OpenSSL clients and servers are not affected.
5436 This issue was reported to OpenSSL by Michal Zalewski (Google).
5441 * DoS via reachable assert in SSLv2 servers fix
5443 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5444 servers that both support SSLv2 and enable export cipher suites by sending
5445 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5447 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5448 (OpenSSL development team).
5453 * Empty CKE with client auth and DHE fix
5455 If client auth is used then a server can seg fault in the event of a DHE
5456 ciphersuite being selected and a zero length ClientKeyExchange message
5457 being sent by the client. This could be exploited in a DoS attack.
5462 * Handshake with unseeded PRNG fix
5464 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5465 with an unseeded PRNG. The conditions are:
5466 - The client is on a platform where the PRNG has not been seeded
5467 automatically, and the user has not seeded manually
5468 - A protocol specific client method version has been used (i.e. not
5469 SSL_client_methodv23)
5470 - A ciphersuite is used that does not require additional random data from
5471 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5473 If the handshake succeeds then the client random that has been used will
5474 have been generated from a PRNG with insufficient entropy and therefore the
5475 output may be predictable.
5477 For example using the following command with an unseeded openssl will
5478 succeed on an unpatched platform:
5480 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5485 * Use After Free following d2i_ECPrivatekey error fix
5487 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5488 could cause a use after free condition. This, in turn, could cause a double
5489 free in several private key parsing functions (such as d2i_PrivateKey
5490 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5491 for applications that receive EC private keys from untrusted
5492 sources. This scenario is considered rare.
5494 This issue was discovered by the BoringSSL project and fixed in their
5500 * X509_to_X509_REQ NULL pointer deref fix
5502 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5503 the certificate key is invalid. This function is rarely used in practice.
5505 This issue was discovered by Brian Carpenter.
5510 * Removed the export ciphers from the DEFAULT ciphers
5514 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5516 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5517 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5518 So far those who have to target multiple platforms would compromise
5519 and argue that binary targeting say ARMv5 would still execute on
5520 ARMv8. "Universal" build resolves this compromise by providing
5521 near-optimal performance even on newer platforms.
5525 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5526 (other platforms pending).
5528 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5530 * Add support for the SignedCertificateTimestampList certificate and
5531 OCSP response extensions from RFC6962.
5535 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5536 for corner cases. (Certain input points at infinity could lead to
5537 bogus results, with non-infinity inputs mapped to infinity too.)
5541 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5542 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5543 common cases are optimized and there still is room for further
5544 improvements. Vector Permutation AES for Altivec is also added.
5548 * Add support for little-endian ppc64 Linux target.
5550 *Marcelo Cerri (IBM)*
5552 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5553 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5554 are optimized and there still is room for further improvements.
5555 Both 32- and 64-bit modes are supported.
5557 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5559 * Improved ARMv7 NEON support.
5563 * Support for SPARC Architecture 2011 crypto extensions, first
5564 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5565 SHA256/512, MD5, GHASH and modular exponentiation.
5567 *Andy Polyakov, David Miller*
5569 * Accelerated modular exponentiation for Intel processors, a.k.a.
5572 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5574 * Support for new and upcoming Intel processors, including AVX2,
5575 BMI and SHA ISA extensions. This includes additional "stitched"
5576 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5579 This work was sponsored by Intel Corp.
5583 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5584 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5585 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5589 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5590 this fixes a limitation in previous versions of OpenSSL.
5594 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5595 MGF1 digest and OAEP label.
5599 * Add EVP support for key wrapping algorithms, to avoid problems with
5600 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5601 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5602 algorithms and include tests cases.
5606 * Add functions to allocate and set the fields of an ECDSA_METHOD
5609 *Douglas E. Engert, Steve Henson*
5611 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5612 difference in days and seconds between two tm or ASN1_TIME structures.
5616 * Add -rev test option to s_server to just reverse order of characters
5617 received by client and send back to server. Also prints an abbreviated
5618 summary of the connection parameters.
5622 * New option -brief for s_client and s_server to print out a brief summary
5623 of connection parameters.
5627 * Add callbacks for arbitrary TLS extensions.
5629 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5631 * New option -crl_download in several openssl utilities to download CRLs
5632 from CRLDP extension in certificates.
5636 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5640 * New function X509_CRL_diff to generate a delta CRL from the difference
5641 of two full CRLs. Add support to "crl" utility.
5645 * New functions to set lookup_crls function and to retrieve
5646 X509_STORE from X509_STORE_CTX.
5650 * Print out deprecated issuer and subject unique ID fields in
5655 * Extend OCSP I/O functions so they can be used for simple general purpose
5656 HTTP as well as OCSP. New wrapper function which can be used to download
5657 CRLs using the OCSP API.
5661 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5665 * `SSL_CONF*` functions. These provide a common framework for application
5666 configuration using configuration files or command lines.
5670 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5671 message callback and prints the results. Needs compile time option
5672 "enable-ssl-trace". New options to s_client and s_server to enable
5677 * New ctrl and macro to retrieve supported points extensions.
5678 Print out extension in s_server and s_client.
5682 * New functions to retrieve certificate signature and signature
5687 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5692 * New Suite B modes for TLS code. These use and enforce the requirements
5693 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5694 only use Suite B curves. The Suite B modes can be set by using the
5695 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5699 * New chain verification flags for Suite B levels of security. Check
5700 algorithms are acceptable when flags are set in X509_verify_cert.
5704 * Make tls1_check_chain return a set of flags indicating checks passed
5705 by a certificate chain. Add additional tests to handle client
5706 certificates: checks for matching certificate type and issuer name
5711 * If an attempt is made to use a signature algorithm not in the peer
5712 preference list abort the handshake. If client has no suitable
5713 signature algorithms in response to a certificate request do not
5714 use the certificate.
5718 * If server EC tmp key is not in client preference list abort handshake.
5722 * Add support for certificate stores in CERT structure. This makes it
5723 possible to have different stores per SSL structure or one store in
5724 the parent SSL_CTX. Include distinct stores for certificate chain
5725 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5726 to build and store a certificate chain in CERT structure: returning
5727 an error if the chain cannot be built: this will allow applications
5728 to test if a chain is correctly configured.
5730 Note: if the CERT based stores are not set then the parent SSL_CTX
5731 store is used to retain compatibility with existing behaviour.
5735 * New function ssl_set_client_disabled to set a ciphersuite disabled
5736 mask based on the current session, check mask when sending client
5737 hello and checking the requested ciphersuite.
5741 * New ctrls to retrieve and set certificate types in a certificate
5742 request message. Print out received values in s_client. If certificate
5743 types is not set with custom values set sensible values based on
5744 supported signature algorithms.
5748 * Support for distinct client and server supported signature algorithms.
5752 * Add certificate callback. If set this is called whenever a certificate
5753 is required by client or server. An application can decide which
5754 certificate chain to present based on arbitrary criteria: for example
5755 supported signature algorithms. Add very simple example to s_server.
5756 This fixes many of the problems and restrictions of the existing client
5757 certificate callback: for example you can now clear an existing
5758 certificate and specify the whole chain.
5762 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5763 the certificate can be used for (if anything). Set valid_flags field
5764 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5765 to have similar checks in it.
5767 Add new "cert_flags" field to CERT structure and include a "strict mode".
5768 This enforces some TLS certificate requirements (such as only permitting
5769 certificate signature algorithms contained in the supported algorithms
5770 extension) which some implementations ignore: this option should be used
5771 with caution as it could cause interoperability issues.
5775 * Update and tidy signature algorithm extension processing. Work out
5776 shared signature algorithms based on preferences and peer algorithms
5777 and print them out in s_client and s_server. Abort handshake if no
5778 shared signature algorithms.
5782 * Add new functions to allow customised supported signature algorithms
5783 for SSL and SSL_CTX structures. Add options to s_client and s_server
5788 * New function SSL_certs_clear() to delete all references to certificates
5789 from an SSL structure. Before this once a certificate had been added
5790 it couldn't be removed.
5794 * Integrate hostname, email address and IP address checking with certificate
5795 verification. New verify options supporting checking in openssl utility.
5799 * Fixes and wildcard matching support to hostname and email checking
5800 functions. Add manual page.
5802 *Florian Weimer (Red Hat Product Security Team)*
5804 * New functions to check a hostname email or IP address against a
5805 certificate. Add options x509 utility to print results of checks against
5810 * Fix OCSP checking.
5812 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5814 * Initial experimental support for explicitly trusted non-root CAs.
5815 OpenSSL still tries to build a complete chain to a root but if an
5816 intermediate CA has a trust setting included that is used. The first
5817 setting is used: whether to trust (e.g., -addtrust option to the x509
5822 * Add -trusted_first option which attempts to find certificates in the
5823 trusted store even if an untrusted chain is also supplied.
5827 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5828 platform support for Linux and Android.
5832 * Support for linux-x32, ILP32 environment in x86_64 framework.
5836 * Experimental multi-implementation support for FIPS capable OpenSSL.
5837 When in FIPS mode the approved implementations are used as normal,
5838 when not in FIPS mode the internal unapproved versions are used instead.
5839 This means that the FIPS capable OpenSSL isn't forced to use the
5840 (often lower performance) FIPS implementations outside FIPS mode.
5844 * Transparently support X9.42 DH parameters when calling
5845 PEM_read_bio_DHparameters. This means existing applications can handle
5846 the new parameter format automatically.
5850 * Initial experimental support for X9.42 DH parameter format: mainly
5851 to support use of 'q' parameter for RFC5114 parameters.
5855 * Add DH parameters from RFC5114 including test data to dhtest.
5859 * Support for automatic EC temporary key parameter selection. If enabled
5860 the most preferred EC parameters are automatically used instead of
5861 hardcoded fixed parameters. Now a server just has to call:
5862 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5863 support ECDH and use the most appropriate parameters.
5867 * Enhance and tidy EC curve and point format TLS extension code. Use
5868 static structures instead of allocation if default values are used.
5869 New ctrls to set curves we wish to support and to retrieve shared curves.
5870 Print out shared curves in s_server. New options to s_server and s_client
5871 to set list of supported curves.
5875 * New ctrls to retrieve supported signature algorithms and
5876 supported curve values as an array of NIDs. Extend openssl utility
5877 to print out received values.
5881 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5882 between NIDs and the more common NIST names such as "P-256". Enhance
5883 ecparam utility and ECC method to recognise the NIST names for curves.
5887 * Enhance SSL/TLS certificate chain handling to support different
5888 chains for each certificate instead of one chain in the parent SSL_CTX.
5892 * Support for fixed DH ciphersuite client authentication: where both
5893 server and client use DH certificates with common parameters.
5897 * Support for fixed DH ciphersuites: those requiring DH server
5902 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5904 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5905 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5906 X509_CINF_get_signature were reverted post internal team review.
5911 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5913 * OCSP Status Request extension unbounded memory growth
5915 A malicious client can send an excessively large OCSP Status Request
5916 extension. If that client continually requests renegotiation, sending a
5917 large OCSP Status Request extension each time, then there will be unbounded
5918 memory growth on the server. This will eventually lead to a Denial Of
5919 Service attack through memory exhaustion. Servers with a default
5920 configuration are vulnerable even if they do not support OCSP. Builds using
5921 the "no-ocsp" build time option are not affected.
5923 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5928 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5931 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5937 * OOB write in MDC2_Update()
5939 An overflow can occur in MDC2_Update() either if called directly or
5940 through the EVP_DigestUpdate() function using MDC2. If an attacker
5941 is able to supply very large amounts of input data after a previous
5942 call to EVP_EncryptUpdate() with a partial block then a length check
5943 can overflow resulting in a heap corruption.
5945 The amount of data needed is comparable to SIZE_MAX which is impractical
5948 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5953 * Malformed SHA512 ticket DoS
5955 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5956 DoS attack where a malformed ticket will result in an OOB read which will
5959 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5960 a custom server callback and ticket lookup mechanism.
5962 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5967 * OOB write in BN_bn2dec()
5969 The function BN_bn2dec() does not check the return value of BN_div_word().
5970 This can cause an OOB write if an application uses this function with an
5971 overly large BIGNUM. This could be a problem if an overly large certificate
5972 or CRL is printed out from an untrusted source. TLS is not affected because
5973 record limits will reject an oversized certificate before it is parsed.
5975 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5980 * OOB read in TS_OBJ_print_bio()
5982 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5983 the total length the OID text representation would use and not the amount
5984 of data written. This will result in OOB reads when large OIDs are
5987 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5992 * Pointer arithmetic undefined behaviour
5994 Avoid some undefined pointer arithmetic
5996 A common idiom in the codebase is to check limits in the following manner:
5999 Where "p" points to some malloc'd data of SIZE bytes and
6002 "len" here could be from some externally supplied data (e.g. from a TLS
6005 The rules of C pointer arithmetic are such that "p + len" is only well
6006 defined where len <= SIZE. Therefore the above idiom is actually
6007 undefined behaviour.
6009 For example this could cause problems if some malloc implementation
6010 provides an address for "p" such that "p + len" actually overflows for
6011 values of len that are too big and therefore p + len < limit.
6013 This issue was reported to OpenSSL by Guido Vranken
6018 * Constant time flag not preserved in DSA signing
6020 Operations in the DSA signing algorithm should run in constant time in
6021 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6022 implementation means that a non-constant time codepath is followed for
6023 certain operations. This has been demonstrated through a cache-timing
6024 attack to be sufficient for an attacker to recover the private DSA key.
6026 This issue was reported by César Pereida (Aalto University), Billy Brumley
6027 (Tampere University of Technology), and Yuval Yarom (The University of
6028 Adelaide and NICTA).
6033 * DTLS buffered message DoS
6035 In a DTLS connection where handshake messages are delivered out-of-order
6036 those messages that OpenSSL is not yet ready to process will be buffered
6037 for later use. Under certain circumstances, a flaw in the logic means that
6038 those messages do not get removed from the buffer even though the handshake
6039 has been completed. An attacker could force up to approx. 15 messages to
6040 remain in the buffer when they are no longer required. These messages will
6041 be cleared when the DTLS connection is closed. The default maximum size for
6042 a message is 100k. Therefore the attacker could force an additional 1500k
6043 to be consumed per connection. By opening many simulataneous connections an
6044 attacker could cause a DoS attack through memory exhaustion.
6046 This issue was reported to OpenSSL by Quan Luo.
6051 * DTLS replay protection DoS
6053 A flaw in the DTLS replay attack protection mechanism means that records
6054 that arrive for future epochs update the replay protection "window" before
6055 the MAC for the record has been validated. This could be exploited by an
6056 attacker by sending a record for the next epoch (which does not have to
6057 decrypt or have a valid MAC), with a very large sequence number. This means
6058 that all subsequent legitimate packets are dropped causing a denial of
6059 service for a specific DTLS connection.
6061 This issue was reported to OpenSSL by the OCAP audit team.
6066 * Certificate message OOB reads
6068 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6069 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6070 theoretical DoS risk but this has not been observed in practice on common
6073 The messages affected are client certificate, client certificate request
6074 and server certificate. As a result the attack can only be performed
6075 against a client or a server which enables client authentication.
6077 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6082 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6084 * Prevent padding oracle in AES-NI CBC MAC check
6086 A MITM attacker can use a padding oracle attack to decrypt traffic
6087 when the connection uses an AES CBC cipher and the server support
6090 This issue was introduced as part of the fix for Lucky 13 padding
6091 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6092 constant time by making sure that always the same bytes are read and
6093 compared against either the MAC or padding bytes. But it no longer
6094 checked that there was enough data to have both the MAC and padding
6097 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6102 * Fix EVP_EncodeUpdate overflow
6104 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6105 Base64 encoding of binary data. If an attacker is able to supply very large
6106 amounts of input data then a length check can overflow resulting in a heap
6109 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6110 the `PEM_write_bio*` family of functions. These are mainly used within the
6111 OpenSSL command line applications, so any application which processes data
6112 from an untrusted source and outputs it as a PEM file should be considered
6113 vulnerable to this issue. User applications that call these APIs directly
6114 with large amounts of untrusted data may also be vulnerable.
6116 This issue was reported by Guido Vranken.
6121 * Fix EVP_EncryptUpdate overflow
6123 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6124 is able to supply very large amounts of input data after a previous call to
6125 EVP_EncryptUpdate() with a partial block then a length check can overflow
6126 resulting in a heap corruption. Following an analysis of all OpenSSL
6127 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6128 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6129 the first called function after an EVP_EncryptInit(), and therefore that
6130 specific call must be safe. The second form is where the length passed to
6131 EVP_EncryptUpdate() can be seen from the code to be some small value and
6132 therefore there is no possibility of an overflow. Since all instances are
6133 one of these two forms, it is believed that there can be no overflows in
6134 internal code due to this problem. It should be noted that
6135 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6136 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6137 of these calls have also been analysed too and it is believed there are no
6138 instances in internal usage where an overflow could occur.
6140 This issue was reported by Guido Vranken.
6145 * Prevent ASN.1 BIO excessive memory allocation
6147 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6148 a short invalid encoding can casuse allocation of large amounts of memory
6149 potentially consuming excessive resources or exhausting memory.
6151 Any application parsing untrusted data through d2i BIO functions is
6152 affected. The memory based functions such as d2i_X509() are *not* affected.
6153 Since the memory based functions are used by the TLS library, TLS
6154 applications are not affected.
6156 This issue was reported by Brian Carpenter.
6163 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6164 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6165 in arbitrary stack data being returned in the buffer.
6167 This issue was reported by Guido Vranken.
6172 * Modify behavior of ALPN to invoke callback after SNI/servername
6173 callback, such that updates to the SSL_CTX affect ALPN.
6177 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6182 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6183 methods are enabled and ssl2 is disabled the methods return NULL.
6187 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6189 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6190 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6191 provide any "EXPORT" or "LOW" strength ciphers.
6195 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6196 is by default disabled at build-time. Builds that are not configured with
6197 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6198 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6199 will need to explicitly call either of:
6201 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6203 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6205 as appropriate. Even if either of those is used, or the application
6206 explicitly uses the version-specific SSLv2_method() or its client and
6207 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6208 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6209 ciphers, and SSLv2 56-bit DES are no longer available.
6214 * Fix a double-free in DSA code
6216 A double free bug was discovered when OpenSSL parses malformed DSA private
6217 keys and could lead to a DoS attack or memory corruption for applications
6218 that receive DSA private keys from untrusted sources. This scenario is
6221 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6227 * Disable SRP fake user seed to address a server memory leak.
6229 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6231 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6232 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6233 was changed to ignore the "fake user" SRP seed, even if the seed
6236 Users should use SRP_VBASE_get1_by_user instead. Note that in
6237 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6238 also that even though configuring the SRP seed attempts to hide
6239 invalid usernames by continuing the handshake with fake
6240 credentials, this behaviour is not constant time and no strong
6241 guarantees are made that the handshake is indistinguishable from
6242 that of a valid user.
6247 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6249 In the BN_hex2bn function the number of hex digits is calculated using an
6250 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6251 large values of `i` this can result in `bn_expand` not allocating any
6252 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6253 field as NULL leading to a subsequent NULL ptr deref. For very large values
6254 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6255 In this case memory is allocated to the internal BIGNUM data field, but it
6256 is insufficiently sized leading to heap corruption. A similar issue exists
6257 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6258 is ever called by user applications with very large untrusted hex/dec data.
6259 This is anticipated to be a rare occurrence.
6261 All OpenSSL internal usage of these functions use data that is not expected
6262 to be untrusted, e.g. config file data or application command line
6263 arguments. If user developed applications generate config file data based
6264 on untrusted data then it is possible that this could also lead to security
6265 consequences. This is also anticipated to be rare.
6267 This issue was reported to OpenSSL by Guido Vranken.
6272 * Fix memory issues in `BIO_*printf` functions
6274 The internal `fmtstr` function used in processing a "%s" format string in
6275 the `BIO_*printf` functions could overflow while calculating the length of a
6276 string and cause an OOB read when printing very long strings.
6278 Additionally the internal `doapr_outch` function can attempt to write to an
6279 OOB memory location (at an offset from the NULL pointer) in the event of a
6280 memory allocation failure. In 1.0.2 and below this could be caused where
6281 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6282 could be in processing a very long "%s" format string. Memory leaks can
6285 The first issue may mask the second issue dependent on compiler behaviour.
6286 These problems could enable attacks where large amounts of untrusted data
6287 is passed to the `BIO_*printf` functions. If applications use these functions
6288 in this way then they could be vulnerable. OpenSSL itself uses these
6289 functions when printing out human-readable dumps of ASN.1 data. Therefore
6290 applications that print this data could be vulnerable if the data is from
6291 untrusted sources. OpenSSL command line applications could also be
6292 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6293 as command line arguments.
6295 Libssl is not considered directly vulnerable. Additionally certificates etc
6296 received via remote connections via libssl are also unlikely to be able to
6297 trigger these issues because of message size limits enforced within libssl.
6299 This issue was reported to OpenSSL Guido Vranken.
6304 * Side channel attack on modular exponentiation
6306 A side-channel attack was found which makes use of cache-bank conflicts on
6307 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6308 of RSA keys. The ability to exploit this issue is limited as it relies on
6309 an attacker who has control of code in a thread running on the same
6310 hyper-threaded core as the victim thread which is performing decryptions.
6312 This issue was reported to OpenSSL by Yuval Yarom, The University of
6313 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6314 Nadia Heninger, University of Pennsylvania with more information at
6315 <http://cachebleed.info>.
6320 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6321 if no keysize is specified with default_bits. This fixes an
6322 omission in an earlier change that changed all RSA/DSA key generation
6323 commands to use 2048 bits by default.
6327 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6329 * Protection for DH small subgroup attacks
6331 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6332 switched on by default and cannot be disabled. This could have some
6337 * SSLv2 doesn't block disabled ciphers
6339 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6340 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6341 been disabled, provided that the SSLv2 protocol was not also disabled via
6344 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6345 and Sebastian Schinzel.
6350 * Reject DH handshakes with parameters shorter than 1024 bits.
6354 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6356 * Certificate verify crash with missing PSS parameter
6358 The signature verification routines will crash with a NULL pointer
6359 dereference if presented with an ASN.1 signature using the RSA PSS
6360 algorithm and absent mask generation function parameter. Since these
6361 routines are used to verify certificate signature algorithms this can be
6362 used to crash any certificate verification operation and exploited in a
6363 DoS attack. Any application which performs certificate verification is
6364 vulnerable including OpenSSL clients and servers which enable client
6367 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6372 * X509_ATTRIBUTE memory leak
6374 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6375 memory. This structure is used by the PKCS#7 and CMS routines so any
6376 application which reads PKCS#7 or CMS data from untrusted sources is
6377 affected. SSL/TLS is not affected.
6379 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6385 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6386 This changes the decoding behaviour for some invalid messages,
6387 though the change is mostly in the more lenient direction, and
6388 legacy behaviour is preserved as much as possible.
6392 * In DSA_generate_parameters_ex, if the provided seed is too short,
6393 use a random seed, as already documented.
6395 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6397 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6399 * Alternate chains certificate forgery
6401 During certificate verfification, OpenSSL will attempt to find an
6402 alternative certificate chain if the first attempt to build such a chain
6403 fails. An error in the implementation of this logic can mean that an
6404 attacker could cause certain checks on untrusted certificates to be
6405 bypassed, such as the CA flag, enabling them to use a valid leaf
6406 certificate to act as a CA and "issue" an invalid certificate.
6408 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6414 * Race condition handling PSK identify hint
6416 If PSK identity hints are received by a multi-threaded client then
6417 the values are wrongly updated in the parent SSL_CTX structure. This can
6418 result in a race condition potentially leading to a double free of the
6424 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6426 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6427 incompatibility in the handling of HMAC. The previous ABI has now been
6430 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6432 * Malformed ECParameters causes infinite loop
6434 When processing an ECParameters structure OpenSSL enters an infinite loop
6435 if the curve specified is over a specially malformed binary polynomial
6438 This can be used to perform denial of service against any
6439 system which processes public keys, certificate requests or
6440 certificates. This includes TLS clients and TLS servers with
6441 client authentication enabled.
6443 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6448 * Exploitable out-of-bounds read in X509_cmp_time
6450 X509_cmp_time does not properly check the length of the ASN1_TIME
6451 string and can read a few bytes out of bounds. In addition,
6452 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6455 An attacker can use this to craft malformed certificates and CRLs of
6456 various sizes and potentially cause a segmentation fault, resulting in
6457 a DoS on applications that verify certificates or CRLs. TLS clients
6458 that verify CRLs are affected. TLS clients and servers with client
6459 authentication enabled may be affected if they use custom verification
6462 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6463 independently by Hanno Böck.
6468 * PKCS7 crash with missing EnvelopedContent
6470 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6471 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6472 with missing content and trigger a NULL pointer dereference on parsing.
6474 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6475 structures from untrusted sources are affected. OpenSSL clients and
6476 servers are not affected.
6478 This issue was reported to OpenSSL by Michal Zalewski (Google).
6483 * CMS verify infinite loop with unknown hash function
6485 When verifying a signedData message the CMS code can enter an infinite loop
6486 if presented with an unknown hash function OID. This can be used to perform
6487 denial of service against any system which verifies signedData messages using
6489 This issue was reported to OpenSSL by Johannes Bauer.
6494 * Race condition handling NewSessionTicket
6496 If a NewSessionTicket is received by a multi-threaded client when attempting to
6497 reuse a previous ticket then a race condition can occur potentially leading to
6498 a double free of the ticket data.
6503 * Reject DH handshakes with parameters shorter than 768 bits.
6505 *Kurt Roeckx and Emilia Kasper*
6507 * dhparam: generate 2048-bit parameters by default.
6509 *Kurt Roeckx and Emilia Kasper*
6511 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6513 * Segmentation fault in ASN1_TYPE_cmp fix
6515 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6516 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6517 certificate signature algorithm consistency this can be used to crash any
6518 certificate verification operation and exploited in a DoS attack. Any
6519 application which performs certificate verification is vulnerable including
6520 OpenSSL clients and servers which enable client authentication.
6525 * ASN.1 structure reuse memory corruption fix
6527 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6528 memory corruption via an invalid write. Such reuse is and has been
6529 strongly discouraged and is believed to be rare.
6531 Applications that parse structures containing CHOICE or ANY DEFINED BY
6532 components may be affected. Certificate parsing (d2i_X509 and related
6533 functions) are however not affected. OpenSSL clients and servers are
6539 * PKCS7 NULL pointer dereferences fix
6541 The PKCS#7 parsing code does not handle missing outer ContentInfo
6542 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6543 missing content and trigger a NULL pointer dereference on parsing.
6545 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6546 otherwise parse PKCS#7 structures from untrusted sources are
6547 affected. OpenSSL clients and servers are not affected.
6549 This issue was reported to OpenSSL by Michal Zalewski (Google).
6554 * DoS via reachable assert in SSLv2 servers fix
6556 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6557 servers that both support SSLv2 and enable export cipher suites by sending
6558 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6560 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6561 (OpenSSL development team).
6566 * Use After Free following d2i_ECPrivatekey error fix
6568 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6569 could cause a use after free condition. This, in turn, could cause a double
6570 free in several private key parsing functions (such as d2i_PrivateKey
6571 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6572 for applications that receive EC private keys from untrusted
6573 sources. This scenario is considered rare.
6575 This issue was discovered by the BoringSSL project and fixed in their
6581 * X509_to_X509_REQ NULL pointer deref fix
6583 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6584 the certificate key is invalid. This function is rarely used in practice.
6586 This issue was discovered by Brian Carpenter.
6591 * Removed the export ciphers from the DEFAULT ciphers
6595 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6597 * Build fixes for the Windows and OpenVMS platforms
6599 *Matt Caswell and Richard Levitte*
6601 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6603 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6604 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6605 dereference. This could lead to a Denial Of Service attack. Thanks to
6606 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6611 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6612 dtls1_buffer_record function under certain conditions. In particular this
6613 could occur if an attacker sent repeated DTLS records with the same
6614 sequence number but for the next epoch. The memory leak could be exploited
6615 by an attacker in a Denial of Service attack through memory exhaustion.
6616 Thanks to Chris Mueller for reporting this issue.
6621 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6622 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6623 method would be set to NULL which could later result in a NULL pointer
6624 dereference. Thanks to Frank Schmirler for reporting this issue.
6629 * Abort handshake if server key exchange message is omitted for ephemeral
6632 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6633 reporting this issue.
6638 * Remove non-export ephemeral RSA code on client and server. This code
6639 violated the TLS standard by allowing the use of temporary RSA keys in
6640 non-export ciphersuites and could be used by a server to effectively
6641 downgrade the RSA key length used to a value smaller than the server
6642 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6643 INRIA or reporting this issue.
6648 * Fixed issue where DH client certificates are accepted without verification.
6649 An OpenSSL server will accept a DH certificate for client authentication
6650 without the certificate verify message. This effectively allows a client to
6651 authenticate without the use of a private key. This only affects servers
6652 which trust a client certificate authority which issues certificates
6653 containing DH keys: these are extremely rare and hardly ever encountered.
6654 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6660 * Ensure that the session ID context of an SSL is updated when its
6661 SSL_CTX is updated via SSL_set_SSL_CTX.
6663 The session ID context is typically set from the parent SSL_CTX,
6664 and can vary with the CTX.
6668 * Fix various certificate fingerprint issues.
6670 By using non-DER or invalid encodings outside the signed portion of a
6671 certificate the fingerprint can be changed without breaking the signature.
6672 Although no details of the signed portion of the certificate can be changed
6673 this can cause problems with some applications: e.g. those using the
6674 certificate fingerprint for blacklists.
6676 1. Reject signatures with non zero unused bits.
6678 If the BIT STRING containing the signature has non zero unused bits reject
6679 the signature. All current signature algorithms require zero unused bits.
6681 2. Check certificate algorithm consistency.
6683 Check the AlgorithmIdentifier inside TBS matches the one in the
6684 certificate signature. NB: this will result in signature failure
6685 errors for some broken certificates.
6687 Thanks to Konrad Kraszewski from Google for reporting this issue.
6689 3. Check DSA/ECDSA signatures use DER.
6691 Re-encode DSA/ECDSA signatures and compare with the original received
6692 signature. Return an error if there is a mismatch.
6694 This will reject various cases including garbage after signature
6695 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6696 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6697 (negative or with leading zeroes).
6699 Further analysis was conducted and fixes were developed by Stephen Henson
6700 of the OpenSSL core team.
6706 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6707 results on some platforms, including x86_64. This bug occurs at random
6708 with a very low probability, and is not known to be exploitable in any
6709 way, though its exact impact is difficult to determine. Thanks to Pieter
6710 Wuille (Blockstream) who reported this issue and also suggested an initial
6711 fix. Further analysis was conducted by the OpenSSL development team and
6712 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6713 the OpenSSL core team.
6718 * Do not resume sessions on the server if the negotiated protocol
6719 version does not match the session's version. Resuming with a different
6720 version, while not strictly forbidden by the RFC, is of questionable
6721 sanity and breaks all known clients.
6723 *David Benjamin, Emilia Käsper*
6725 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6726 early CCS messages during renegotiation. (Note that because
6727 renegotiation is encrypted, this early CCS was not exploitable.)
6731 * Tighten client-side session ticket handling during renegotiation:
6732 ensure that the client only accepts a session ticket if the server sends
6733 the extension anew in the ServerHello. Previously, a TLS client would
6734 reuse the old extension state and thus accept a session ticket if one was
6735 announced in the initial ServerHello.
6737 Similarly, ensure that the client requires a session ticket if one
6738 was advertised in the ServerHello. Previously, a TLS client would
6739 ignore a missing NewSessionTicket message.
6743 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6747 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6748 sends a carefully crafted handshake message, to cause OpenSSL to fail
6749 to free up to 64k of memory causing a memory leak. This could be
6750 exploited in a Denial Of Service attack. This issue affects OpenSSL
6751 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6752 whether SRTP is used or configured. Implementations of OpenSSL that
6753 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6755 The fix was developed by the OpenSSL team.
6760 * Session Ticket Memory Leak.
6762 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6763 integrity of that ticket is first verified. In the event of a session
6764 ticket integrity check failing, OpenSSL will fail to free memory
6765 causing a memory leak. By sending a large number of invalid session
6766 tickets an attacker could exploit this issue in a Denial Of Service
6772 * Build option no-ssl3 is incomplete.
6774 When OpenSSL is configured with "no-ssl3" as a build option, servers
6775 could accept and complete a SSL 3.0 handshake, and clients could be
6776 configured to send them.
6779 *Akamai and the OpenSSL team*
6781 * Add support for TLS_FALLBACK_SCSV.
6782 Client applications doing fallback retries should call
6783 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6786 *Adam Langley, Bodo Moeller*
6788 * Add additional DigestInfo checks.
6790 Re-encode DigestInto in DER and check against the original when
6791 verifying RSA signature: this will reject any improperly encoded
6792 DigestInfo structures.
6794 Note: this is a precautionary measure and no attacks are currently known.
6798 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6800 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6801 SRP code can be overrun an internal buffer. Add sanity check that
6802 g, A, B < N to SRP code.
6804 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6805 Group for discovering this issue.
6810 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6811 TLS 1.0 instead of higher protocol versions when the ClientHello message
6812 is badly fragmented. This allows a man-in-the-middle attacker to force a
6813 downgrade to TLS 1.0 even if both the server and the client support a
6814 higher protocol version, by modifying the client's TLS records.
6816 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6817 researching this issue.
6822 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6823 to a denial of service attack. A malicious server can crash the client
6824 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6825 ciphersuite and sending carefully crafted handshake messages.
6827 Thanks to Felix Gröbert (Google) for discovering and researching this
6833 * By sending carefully crafted DTLS packets an attacker could cause openssl
6834 to leak memory. This can be exploited through a Denial of Service attack.
6835 Thanks to Adam Langley for discovering and researching this issue.
6840 * An attacker can force openssl to consume large amounts of memory whilst
6841 processing DTLS handshake messages. This can be exploited through a
6842 Denial of Service attack.
6843 Thanks to Adam Langley for discovering and researching this issue.
6848 * An attacker can force an error condition which causes openssl to crash
6849 whilst processing DTLS packets due to memory being freed twice. This
6850 can be exploited through a Denial of Service attack.
6851 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6857 * If a multithreaded client connects to a malicious server using a resumed
6858 session and the server sends an ec point format extension it could write
6859 up to 255 bytes to freed memory.
6861 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6867 * A malicious server can crash an OpenSSL client with a null pointer
6868 dereference (read) by specifying an SRP ciphersuite even though it was not
6869 properly negotiated with the client. This can be exploited through a
6870 Denial of Service attack.
6872 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6873 discovering and researching this issue.
6878 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6879 X509_name_oneline, X509_name_print_ex et al. to leak some information
6880 from the stack. Applications may be affected if they echo pretty printing
6881 output to the attacker.
6883 Thanks to Ivan Fratric (Google) for discovering this issue.
6886 *Emilia Käsper, and Steve Henson*
6888 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6889 for corner cases. (Certain input points at infinity could lead to
6890 bogus results, with non-infinity inputs mapped to infinity too.)
6894 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6896 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6897 handshake can force the use of weak keying material in OpenSSL
6898 SSL/TLS clients and servers.
6900 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6901 researching this issue. ([CVE-2014-0224])
6903 *KIKUCHI Masashi, Steve Henson*
6905 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6906 OpenSSL DTLS client the code can be made to recurse eventually crashing
6909 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6912 *Imre Rad, Steve Henson*
6914 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6915 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6916 client or server. This is potentially exploitable to run arbitrary
6917 code on a vulnerable client or server.
6919 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6921 *Jüri Aedla, Steve Henson*
6923 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6924 are subject to a denial of service attack.
6926 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6927 this issue. ([CVE-2014-3470])
6929 *Felix Gröbert, Ivan Fratric, Steve Henson*
6931 * Harmonize version and its documentation. -f flag is used to display
6934 *mancha <mancha1@zoho.com>*
6936 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6937 in i2d_ECPrivateKey.
6939 *mancha <mancha1@zoho.com>*
6941 * Fix some double frees. These are not thought to be exploitable.
6943 *mancha <mancha1@zoho.com>*
6945 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6947 * A missing bounds check in the handling of the TLS heartbeat extension
6948 can be used to reveal up to 64k of memory to a connected client or
6951 Thanks for Neel Mehta of Google Security for discovering this bug and to
6952 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6953 preparing the fix ([CVE-2014-0160])
6955 *Adam Langley, Bodo Moeller*
6957 * Fix for the attack described in the paper "Recovering OpenSSL
6958 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6959 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6960 <http://eprint.iacr.org/2014/140>
6962 Thanks to Yuval Yarom and Naomi Benger for discovering this
6963 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6965 *Yuval Yarom and Naomi Benger*
6967 * TLS pad extension: draft-agl-tls-padding-03
6969 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6970 TLS client Hello record length value would otherwise be > 255 and
6971 less that 512 pad with a dummy extension containing zeroes so it
6972 is at least 512 bytes long.
6974 *Adam Langley, Steve Henson*
6976 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6978 * Fix for TLS record tampering bug. A carefully crafted invalid
6979 handshake could crash OpenSSL with a NULL pointer exception.
6980 Thanks to Anton Johansson for reporting this issues.
6983 * Keep original DTLS digest and encryption contexts in retransmission
6984 structures so we can use the previous session parameters if they need
6985 to be resent. ([CVE-2013-6450])
6989 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6990 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6991 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6992 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6993 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6994 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6996 *Rob Stradling, Adam Langley*
6998 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7000 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7001 supporting platforms or when small records were transferred.
7003 *Andy Polyakov, Steve Henson*
7005 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7007 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7009 This addresses the flaw in CBC record processing discovered by
7010 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7011 at: <http://www.isg.rhul.ac.uk/tls/>
7013 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7014 Security Group at Royal Holloway, University of London
7015 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7016 Emilia Käsper for the initial patch.
7019 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7021 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7022 ciphersuites which can be exploited in a denial of service attack.
7023 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7024 and detecting this bug and to Wolfgang Ettlinger
7025 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7030 * Return an error when checking OCSP signatures when key is NULL.
7031 This fixes a DoS attack. ([CVE-2013-0166])
7035 * Make openssl verify return errors.
7037 *Chris Palmer <palmer@google.com> and Ben Laurie*
7039 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7040 the right response is stapled. Also change SSL_get_certificate()
7041 so it returns the certificate actually sent.
7042 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7044 *Rob Stradling <rob.stradling@comodo.com>*
7046 * Fix possible deadlock when decoding public keys.
7050 * Don't use TLS 1.0 record version number in initial client hello
7055 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7057 * Sanity check record length before skipping explicit IV in TLS
7058 1.2, 1.1 and DTLS to fix DoS attack.
7060 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7061 fuzzing as a service testing platform.
7066 * Initialise tkeylen properly when encrypting CMS messages.
7067 Thanks to Solar Designer of Openwall for reporting this issue.
7071 * In FIPS mode don't try to use composite ciphers as they are not
7076 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7078 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7079 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7080 mean any application compiled against OpenSSL 1.0.0 headers setting
7081 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7082 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7083 0x10000000L Any application which was previously compiled against
7084 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7085 will need to be recompiled as a result. Letting be results in
7086 inability to disable specifically TLS 1.1 and in client context,
7087 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7091 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7092 disable just protocol X, but all protocols above X *if* there are
7093 protocols *below* X still enabled. In more practical terms it means
7094 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7095 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7096 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7101 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7103 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7104 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7105 in CRYPTO_realloc_clean.
7107 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7108 issue and to Adam Langley <agl@chromium.org> for fixing it.
7111 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7113 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7117 * Workarounds for some broken servers that "hang" if a client hello
7118 record length exceeds 255 bytes.
7120 1. Do not use record version number > TLS 1.0 in initial client
7121 hello: some (but not all) hanging servers will now work.
7122 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7123 the number of ciphers sent in the client hello. This should be
7124 set to an even number, such as 50, for example by passing:
7125 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7126 Most broken servers should now work.
7127 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7128 TLS 1.2 client support entirely.
7132 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7136 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7138 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7139 STRING form instead of a DigestInfo.
7143 * The format used for MDC2 RSA signatures is inconsistent between EVP
7144 and the RSA_sign/RSA_verify functions. This was made more apparent when
7145 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7146 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7147 the correct format in RSA_verify so both forms transparently work.
7151 * Some servers which support TLS 1.0 can choke if we initially indicate
7152 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7153 encrypted premaster secret. As a workaround use the maximum permitted
7154 client version in client hello, this should keep such servers happy
7155 and still work with previous versions of OpenSSL.
7159 * Add support for TLS/DTLS heartbeats.
7161 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7163 * Add support for SCTP.
7165 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7167 * Improved PRNG seeding for VOS.
7169 *Paul Green <Paul.Green@stratus.com>*
7171 * Extensive assembler packs updates, most notably:
7173 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7174 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7175 - x86_64: bit-sliced AES implementation;
7176 - ARM: NEON support, contemporary platforms optimizations;
7177 - s390x: z196 support;
7178 - `*`: GHASH and GF(2^m) multiplication implementations;
7182 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7183 (removal of unnecessary code)
7185 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7187 * Add TLS key material exporter from RFC 5705.
7191 * Add DTLS-SRTP negotiation from RFC 5764.
7195 * Add Next Protocol Negotiation,
7196 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7197 disabled with a no-npn flag to config or Configure. Code donated
7200 *Adam Langley <agl@google.com> and Ben Laurie*
7202 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7203 NIST-P256, NIST-P521, with constant-time single point multiplication on
7204 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7205 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7206 Code made available under Apache License version 2.0.
7208 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7209 line to include this in your build of OpenSSL, and run "make depend" (or
7210 "make update"). This enables the following EC_METHODs:
7212 EC_GFp_nistp224_method()
7213 EC_GFp_nistp256_method()
7214 EC_GFp_nistp521_method()
7216 EC_GROUP_new_by_curve_name() will automatically use these (while
7217 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7220 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7222 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7223 all platforms. Move ssize_t definition from e_os.h to the public
7224 header file e_os2.h as it now appears in public header file cms.h
7228 * New -sigopt option to the ca, req and x509 utilities. Additional
7229 signature parameters can be passed using this option and in
7234 * Add RSA PSS signing function. This will generate and set the
7235 appropriate AlgorithmIdentifiers for PSS based on those in the
7236 corresponding EVP_MD_CTX structure. No application support yet.
7240 * Support for companion algorithm specific ASN1 signing routines.
7241 New function ASN1_item_sign_ctx() signs a pre-initialised
7242 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7243 the appropriate parameters.
7247 * Add new algorithm specific ASN1 verification initialisation function
7248 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7249 handling will be the same no matter what EVP_PKEY_METHOD is used.
7250 Add a PSS handler to support verification of PSS signatures: checked
7251 against a number of sample certificates.
7255 * Add signature printing for PSS. Add PSS OIDs.
7257 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7259 * Add algorithm specific signature printing. An individual ASN1 method
7260 can now print out signatures instead of the standard hex dump.
7262 More complex signatures (e.g. PSS) can print out more meaningful
7263 information. Include DSA version that prints out the signature
7268 * Password based recipient info support for CMS library: implementing
7273 * Split password based encryption into PBES2 and PBKDF2 functions. This
7274 neatly separates the code into cipher and PBE sections and is required
7275 for some algorithms that split PBES2 into separate pieces (such as
7276 password based CMS).
7280 * Session-handling fixes:
7281 - Fix handling of connections that are resuming with a session ID,
7282 but also support Session Tickets.
7283 - Fix a bug that suppressed issuing of a new ticket if the client
7284 presented a ticket with an expired session.
7285 - Try to set the ticket lifetime hint to something reasonable.
7286 - Make tickets shorter by excluding irrelevant information.
7287 - On the client side, don't ignore renewed tickets.
7289 *Adam Langley, Bodo Moeller (Google)*
7291 * Fix PSK session representation.
7295 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7297 This work was sponsored by Intel.
7301 * Add GCM support to TLS library. Some custom code is needed to split
7302 the IV between the fixed (from PRF) and explicit (from TLS record)
7303 portions. This adds all GCM ciphersuites supported by RFC5288 and
7304 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7305 add a special AESGCM string for GCM only.
7309 * Expand range of ctrls for AES GCM. Permit setting invocation
7310 field on decrypt and retrieval of invocation field only on encrypt.
7314 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7315 As required by RFC5289 these ciphersuites cannot be used if for
7316 versions of TLS earlier than 1.2.
7320 * For FIPS capable OpenSSL interpret a NULL default public key method
7321 as unset and return the appropriate default but do *not* set the default.
7322 This means we can return the appropriate method in applications that
7323 switch between FIPS and non-FIPS modes.
7327 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7328 ENGINE is used then we cannot handle that in the FIPS module so we
7329 keep original code iff non-FIPS operations are allowed.
7333 * Add -attime option to openssl utilities.
7335 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7337 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7341 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7342 FIPS EC methods unconditionally for now.
7346 * New build option no-ec2m to disable characteristic 2 code.
7350 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7351 all cases can be covered as some introduce binary incompatibilities.
7355 * Redirect RSA operations to FIPS module including keygen,
7356 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7360 * Add similar low level API blocking to ciphers.
7364 * Low level digest APIs are not approved in FIPS mode: any attempt
7365 to use these will cause a fatal error. Applications that *really* want
7366 to use them can use the `private_*` version instead.
7370 * Redirect cipher operations to FIPS module for FIPS builds.
7374 * Redirect digest operations to FIPS module for FIPS builds.
7378 * Update build system to add "fips" flag which will link in fipscanister.o
7379 for static and shared library builds embedding a signature if needed.
7383 * Output TLS supported curves in preference order instead of numerical
7384 order. This is currently hardcoded for the highest order curves first.
7385 This should be configurable so applications can judge speed vs strength.
7389 * Add TLS v1.2 server support for client authentication.
7393 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7398 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7399 FIPS modules versions.
7403 * Add TLS v1.2 client side support for client authentication. Keep cache
7404 of handshake records longer as we don't know the hash algorithm to use
7405 until after the certificate request message is received.
7409 * Initial TLS v1.2 client support. Add a default signature algorithms
7410 extension including all the algorithms we support. Parse new signature
7411 format in client key exchange. Relax some ECC signing restrictions for
7412 TLS v1.2 as indicated in RFC5246.
7416 * Add server support for TLS v1.2 signature algorithms extension. Switch
7417 to new signature format when needed using client digest preference.
7418 All server ciphersuites should now work correctly in TLS v1.2. No client
7419 support yet and no support for client certificates.
7423 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7424 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7425 ciphersuites. At present only RSA key exchange ciphersuites work with
7426 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7427 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7428 and version checking.
7432 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7433 with this defined it will not be affected by any changes to ssl internal
7434 structures. Add several utility functions to allow openssl application
7435 to work with OPENSSL_NO_SSL_INTERN defined.
7439 * A long standing patch to add support for SRP from EdelWeb (Peter
7440 Sylvester and Christophe Renou) was integrated.
7441 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7442 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7445 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7449 * Permit abbreviated handshakes when renegotiating using the function
7450 SSL_renegotiate_abbreviated().
7452 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7454 * Add call to ENGINE_register_all_complete() to
7455 ENGINE_load_builtin_engines(), so some implementations get used
7456 automatically instead of needing explicit application support.
7460 * Add support for TLS key exporter as described in RFC5705.
7462 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7464 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7465 a few changes are required:
7467 Add SSL_OP_NO_TLSv1_1 flag.
7468 Add TLSv1_1 methods.
7469 Update version checking logic to handle version 1.1.
7470 Add explicit IV handling (ported from DTLS code).
7471 Add command line options to s_client/s_server.
7478 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7480 * X509_ATTRIBUTE memory leak
7482 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7483 memory. This structure is used by the PKCS#7 and CMS routines so any
7484 application which reads PKCS#7 or CMS data from untrusted sources is
7485 affected. SSL/TLS is not affected.
7487 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7493 * Race condition handling PSK identify hint
7495 If PSK identity hints are received by a multi-threaded client then
7496 the values are wrongly updated in the parent SSL_CTX structure. This can
7497 result in a race condition potentially leading to a double free of the
7503 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7505 * Malformed ECParameters causes infinite loop
7507 When processing an ECParameters structure OpenSSL enters an infinite loop
7508 if the curve specified is over a specially malformed binary polynomial
7511 This can be used to perform denial of service against any
7512 system which processes public keys, certificate requests or
7513 certificates. This includes TLS clients and TLS servers with
7514 client authentication enabled.
7516 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7521 * Exploitable out-of-bounds read in X509_cmp_time
7523 X509_cmp_time does not properly check the length of the ASN1_TIME
7524 string and can read a few bytes out of bounds. In addition,
7525 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7528 An attacker can use this to craft malformed certificates and CRLs of
7529 various sizes and potentially cause a segmentation fault, resulting in
7530 a DoS on applications that verify certificates or CRLs. TLS clients
7531 that verify CRLs are affected. TLS clients and servers with client
7532 authentication enabled may be affected if they use custom verification
7535 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7536 independently by Hanno Böck.
7541 * PKCS7 crash with missing EnvelopedContent
7543 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7544 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7545 with missing content and trigger a NULL pointer dereference on parsing.
7547 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7548 structures from untrusted sources are affected. OpenSSL clients and
7549 servers are not affected.
7551 This issue was reported to OpenSSL by Michal Zalewski (Google).
7556 * CMS verify infinite loop with unknown hash function
7558 When verifying a signedData message the CMS code can enter an infinite loop
7559 if presented with an unknown hash function OID. This can be used to perform
7560 denial of service against any system which verifies signedData messages using
7562 This issue was reported to OpenSSL by Johannes Bauer.
7567 * Race condition handling NewSessionTicket
7569 If a NewSessionTicket is received by a multi-threaded client when attempting to
7570 reuse a previous ticket then a race condition can occur potentially leading to
7571 a double free of the ticket data.
7576 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7578 * Segmentation fault in ASN1_TYPE_cmp fix
7580 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7581 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7582 certificate signature algorithm consistency this can be used to crash any
7583 certificate verification operation and exploited in a DoS attack. Any
7584 application which performs certificate verification is vulnerable including
7585 OpenSSL clients and servers which enable client authentication.
7590 * ASN.1 structure reuse memory corruption fix
7592 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7593 memory corruption via an invalid write. Such reuse is and has been
7594 strongly discouraged and is believed to be rare.
7596 Applications that parse structures containing CHOICE or ANY DEFINED BY
7597 components may be affected. Certificate parsing (d2i_X509 and related
7598 functions) are however not affected. OpenSSL clients and servers are
7604 * PKCS7 NULL pointer dereferences fix
7606 The PKCS#7 parsing code does not handle missing outer ContentInfo
7607 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7608 missing content and trigger a NULL pointer dereference on parsing.
7610 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7611 otherwise parse PKCS#7 structures from untrusted sources are
7612 affected. OpenSSL clients and servers are not affected.
7614 This issue was reported to OpenSSL by Michal Zalewski (Google).
7619 * DoS via reachable assert in SSLv2 servers fix
7621 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7622 servers that both support SSLv2 and enable export cipher suites by sending
7623 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7625 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7626 (OpenSSL development team).
7631 * Use After Free following d2i_ECPrivatekey error fix
7633 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7634 could cause a use after free condition. This, in turn, could cause a double
7635 free in several private key parsing functions (such as d2i_PrivateKey
7636 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7637 for applications that receive EC private keys from untrusted
7638 sources. This scenario is considered rare.
7640 This issue was discovered by the BoringSSL project and fixed in their
7646 * X509_to_X509_REQ NULL pointer deref fix
7648 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7649 the certificate key is invalid. This function is rarely used in practice.
7651 This issue was discovered by Brian Carpenter.
7656 * Removed the export ciphers from the DEFAULT ciphers
7660 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7662 * Build fixes for the Windows and OpenVMS platforms
7664 *Matt Caswell and Richard Levitte*
7666 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7668 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7669 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7670 dereference. This could lead to a Denial Of Service attack. Thanks to
7671 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7676 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7677 dtls1_buffer_record function under certain conditions. In particular this
7678 could occur if an attacker sent repeated DTLS records with the same
7679 sequence number but for the next epoch. The memory leak could be exploited
7680 by an attacker in a Denial of Service attack through memory exhaustion.
7681 Thanks to Chris Mueller for reporting this issue.
7686 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7687 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7688 method would be set to NULL which could later result in a NULL pointer
7689 dereference. Thanks to Frank Schmirler for reporting this issue.
7694 * Abort handshake if server key exchange message is omitted for ephemeral
7697 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7698 reporting this issue.
7703 * Remove non-export ephemeral RSA code on client and server. This code
7704 violated the TLS standard by allowing the use of temporary RSA keys in
7705 non-export ciphersuites and could be used by a server to effectively
7706 downgrade the RSA key length used to a value smaller than the server
7707 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7708 INRIA or reporting this issue.
7713 * Fixed issue where DH client certificates are accepted without verification.
7714 An OpenSSL server will accept a DH certificate for client authentication
7715 without the certificate verify message. This effectively allows a client to
7716 authenticate without the use of a private key. This only affects servers
7717 which trust a client certificate authority which issues certificates
7718 containing DH keys: these are extremely rare and hardly ever encountered.
7719 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7725 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7726 results on some platforms, including x86_64. This bug occurs at random
7727 with a very low probability, and is not known to be exploitable in any
7728 way, though its exact impact is difficult to determine. Thanks to Pieter
7729 Wuille (Blockstream) who reported this issue and also suggested an initial
7730 fix. Further analysis was conducted by the OpenSSL development team and
7731 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7732 the OpenSSL core team.
7737 * Fix various certificate fingerprint issues.
7739 By using non-DER or invalid encodings outside the signed portion of a
7740 certificate the fingerprint can be changed without breaking the signature.
7741 Although no details of the signed portion of the certificate can be changed
7742 this can cause problems with some applications: e.g. those using the
7743 certificate fingerprint for blacklists.
7745 1. Reject signatures with non zero unused bits.
7747 If the BIT STRING containing the signature has non zero unused bits reject
7748 the signature. All current signature algorithms require zero unused bits.
7750 2. Check certificate algorithm consistency.
7752 Check the AlgorithmIdentifier inside TBS matches the one in the
7753 certificate signature. NB: this will result in signature failure
7754 errors for some broken certificates.
7756 Thanks to Konrad Kraszewski from Google for reporting this issue.
7758 3. Check DSA/ECDSA signatures use DER.
7760 Reencode DSA/ECDSA signatures and compare with the original received
7761 signature. Return an error if there is a mismatch.
7763 This will reject various cases including garbage after signature
7764 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7765 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7766 (negative or with leading zeroes).
7768 Further analysis was conducted and fixes were developed by Stephen Henson
7769 of the OpenSSL core team.
7775 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7777 * Session Ticket Memory Leak.
7779 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7780 integrity of that ticket is first verified. In the event of a session
7781 ticket integrity check failing, OpenSSL will fail to free memory
7782 causing a memory leak. By sending a large number of invalid session
7783 tickets an attacker could exploit this issue in a Denial Of Service
7789 * Build option no-ssl3 is incomplete.
7791 When OpenSSL is configured with "no-ssl3" as a build option, servers
7792 could accept and complete a SSL 3.0 handshake, and clients could be
7793 configured to send them.
7796 *Akamai and the OpenSSL team*
7798 * Add support for TLS_FALLBACK_SCSV.
7799 Client applications doing fallback retries should call
7800 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7803 *Adam Langley, Bodo Moeller*
7805 * Add additional DigestInfo checks.
7807 Reencode DigestInto in DER and check against the original when
7808 verifying RSA signature: this will reject any improperly encoded
7809 DigestInfo structures.
7811 Note: this is a precautionary measure and no attacks are currently known.
7815 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7817 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7818 to a denial of service attack. A malicious server can crash the client
7819 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7820 ciphersuite and sending carefully crafted handshake messages.
7822 Thanks to Felix Gröbert (Google) for discovering and researching this
7828 * By sending carefully crafted DTLS packets an attacker could cause openssl
7829 to leak memory. This can be exploited through a Denial of Service attack.
7830 Thanks to Adam Langley for discovering and researching this issue.
7835 * An attacker can force openssl to consume large amounts of memory whilst
7836 processing DTLS handshake messages. This can be exploited through a
7837 Denial of Service attack.
7838 Thanks to Adam Langley for discovering and researching this issue.
7843 * An attacker can force an error condition which causes openssl to crash
7844 whilst processing DTLS packets due to memory being freed twice. This
7845 can be exploited through a Denial of Service attack.
7846 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7852 * If a multithreaded client connects to a malicious server using a resumed
7853 session and the server sends an ec point format extension it could write
7854 up to 255 bytes to freed memory.
7856 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7862 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7863 X509_name_oneline, X509_name_print_ex et al. to leak some information
7864 from the stack. Applications may be affected if they echo pretty printing
7865 output to the attacker.
7867 Thanks to Ivan Fratric (Google) for discovering this issue.
7870 *Emilia Käsper, and Steve Henson*
7872 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7873 for corner cases. (Certain input points at infinity could lead to
7874 bogus results, with non-infinity inputs mapped to infinity too.)
7878 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7880 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7881 handshake can force the use of weak keying material in OpenSSL
7882 SSL/TLS clients and servers.
7884 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7885 researching this issue. ([CVE-2014-0224])
7887 *KIKUCHI Masashi, Steve Henson*
7889 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7890 OpenSSL DTLS client the code can be made to recurse eventually crashing
7893 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7896 *Imre Rad, Steve Henson*
7898 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7899 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7900 client or server. This is potentially exploitable to run arbitrary
7901 code on a vulnerable client or server.
7903 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7905 *Jüri Aedla, Steve Henson*
7907 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7908 are subject to a denial of service attack.
7910 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7911 this issue. ([CVE-2014-3470])
7913 *Felix Gröbert, Ivan Fratric, Steve Henson*
7915 * Harmonize version and its documentation. -f flag is used to display
7918 *mancha <mancha1@zoho.com>*
7920 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7921 in i2d_ECPrivateKey.
7923 *mancha <mancha1@zoho.com>*
7925 * Fix some double frees. These are not thought to be exploitable.
7927 *mancha <mancha1@zoho.com>*
7929 * Fix for the attack described in the paper "Recovering OpenSSL
7930 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7931 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7932 <http://eprint.iacr.org/2014/140>
7934 Thanks to Yuval Yarom and Naomi Benger for discovering this
7935 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7937 *Yuval Yarom and Naomi Benger*
7939 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7941 * Keep original DTLS digest and encryption contexts in retransmission
7942 structures so we can use the previous session parameters if they need
7943 to be resent. ([CVE-2013-6450])
7947 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7948 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7949 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7950 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7951 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7952 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7954 *Rob Stradling, Adam Langley*
7956 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7958 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7960 This addresses the flaw in CBC record processing discovered by
7961 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7962 at: <http://www.isg.rhul.ac.uk/tls/>
7964 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7965 Security Group at Royal Holloway, University of London
7966 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7967 Emilia Käsper for the initial patch.
7970 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7972 * Return an error when checking OCSP signatures when key is NULL.
7973 This fixes a DoS attack. ([CVE-2013-0166])
7977 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7978 the right response is stapled. Also change SSL_get_certificate()
7979 so it returns the certificate actually sent.
7980 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7981 (This is a backport)
7983 *Rob Stradling <rob.stradling@comodo.com>*
7985 * Fix possible deadlock when decoding public keys.
7989 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7991 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7994 * Sanity check record length before skipping explicit IV in DTLS
7997 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7998 fuzzing as a service testing platform.
8003 * Initialise tkeylen properly when encrypting CMS messages.
8004 Thanks to Solar Designer of Openwall for reporting this issue.
8008 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8010 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8011 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8012 in CRYPTO_realloc_clean.
8014 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8015 issue and to Adam Langley <agl@chromium.org> for fixing it.
8018 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8020 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8022 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8023 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8024 content decryption and always return the same error. Note: this attack
8025 needs on average 2^20 messages so it only affects automated senders. The
8026 old behaviour can be re-enabled in the CMS code by setting the
8027 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8028 an MMA defence is not necessary.
8029 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8030 this issue. ([CVE-2012-0884])
8034 * Fix CVE-2011-4619: make sure we really are receiving a
8035 client hello before rejecting multiple SGC restarts. Thanks to
8036 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8040 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8042 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8043 Thanks to Antonio Martin, Enterprise Secure Access Research and
8044 Development, Cisco Systems, Inc. for discovering this bug and
8045 preparing a fix. ([CVE-2012-0050])
8049 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8051 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8052 of the Vaudenay padding oracle attack on CBC mode encryption
8053 which enables an efficient plaintext recovery attack against
8054 the OpenSSL implementation of DTLS. Their attack exploits timing
8055 differences arising during decryption processing. A research
8056 paper describing this attack can be found at:
8057 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8058 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8059 Security Group at Royal Holloway, University of London
8060 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8061 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8062 for preparing the fix. ([CVE-2011-4108])
8064 *Robin Seggelmann, Michael Tuexen*
8066 * Clear bytes used for block padding of SSL 3.0 records.
8069 *Adam Langley (Google)*
8071 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8072 Kadianakis <desnacked@gmail.com> for discovering this issue and
8073 Adam Langley for preparing the fix. ([CVE-2011-4619])
8075 *Adam Langley (Google)*
8077 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8079 *Andrey Kulikov <amdeich@gmail.com>*
8081 * Prevent malformed RFC3779 data triggering an assertion failure.
8082 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8083 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8085 *Rob Austein <sra@hactrn.net>*
8087 * Improved PRNG seeding for VOS.
8089 *Paul Green <Paul.Green@stratus.com>*
8091 * Fix ssl_ciph.c set-up race.
8093 *Adam Langley (Google)*
8095 * Fix spurious failures in ecdsatest.c.
8097 *Emilia Käsper (Google)*
8099 * Fix the BIO_f_buffer() implementation (which was mixing different
8100 interpretations of the `..._len` fields).
8102 *Adam Langley (Google)*
8104 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8105 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8106 threads won't reuse the same blinding coefficients.
8108 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8109 lock to call BN_BLINDING_invert_ex, and avoids one use of
8110 BN_BLINDING_update for each BN_BLINDING structure (previously,
8111 the last update always remained unused).
8113 *Emilia Käsper (Google)*
8115 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8117 *Bob Buckholz (Google)*
8119 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8121 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8122 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8124 *Kaspar Brand <ossl@velox.ch>*
8126 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8127 for multi-threaded use of ECDH. ([CVE-2011-3210])
8129 *Adam Langley (Google)*
8131 * Fix x509_name_ex_d2i memory leak on bad inputs.
8135 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8136 signature public key algorithm by using OID xref utilities instead.
8137 Before this you could only use some ECC ciphersuites with SHA1 only.
8141 * Add protection against ECDSA timing attacks as mentioned in the paper
8142 by Billy Bob Brumley and Nicola Tuveri, see:
8143 <http://eprint.iacr.org/2011/232.pdf>
8145 *Billy Bob Brumley and Nicola Tuveri*
8147 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8149 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8151 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8153 * Fix bug in string printing code: if *any* escaping is enabled we must
8154 escape the escape character (backslash) or the resulting string is
8159 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8161 * Disable code workaround for ancient and obsolete Netscape browsers
8162 and servers: an attacker can use it in a ciphersuite downgrade attack.
8163 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8167 * Fixed J-PAKE implementation error, originally discovered by
8168 Sebastien Martini, further info and confirmation from Stefan
8169 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8173 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8175 * Fix extension code to avoid race conditions which can result in a buffer
8176 overrun vulnerability: resumed sessions must not be modified as they can
8177 be shared by multiple threads. CVE-2010-3864
8181 * Fix WIN32 build system to correctly link an ENGINE directory into
8186 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8188 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8191 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8193 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8195 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8196 context. The operation can be customised via the ctrl mechanism in
8197 case ENGINEs want to include additional functionality.
8201 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8205 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8206 output hashes compatible with older versions of OpenSSL.
8208 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8210 * Fix compression algorithm handling: if resuming a session use the
8211 compression algorithm of the resumed session instead of determining
8212 it from client hello again. Don't allow server to change algorithm.
8216 * Add load_crls() function to commands tidying load_certs() too. Add option
8217 to verify utility to allow additional CRLs to be included.
8221 * Update OCSP request code to permit adding custom headers to the request:
8222 some responders need this.
8226 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8229 *Julia Lawall <julia@diku.dk>*
8231 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8232 needlessly dereferenced structures, used obsolete functions and
8233 didn't handle all updated verify codes correctly.
8237 * Disable MD2 in the default configuration.
8241 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8242 indicate the initial BIO being pushed or popped. This makes it possible
8243 to determine whether the BIO is the one explicitly called or as a result
8244 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8245 it handles reference counts correctly and doesn't zero out the I/O bio
8246 when it is not being explicitly popped. WARNING: applications which
8247 included workarounds for the old buggy behaviour will need to be modified
8248 or they could free up already freed BIOs.
8252 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8253 renaming to all platforms (within the 0.9.8 branch, this was
8254 done conditionally on Netware platforms to avoid a name clash).
8256 *Guenter <lists@gknw.net>*
8258 * Add ECDHE and PSK support to DTLS.
8260 *Michael Tuexen <tuexen@fh-muenster.de>*
8262 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8267 * Add "missing" function EVP_MD_flags() (without this the only way to
8268 retrieve a digest flags is by accessing the structure directly. Update
8269 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8270 or cipher is registered as in the "from" argument. Print out all
8271 registered digests in the dgst usage message instead of manually
8272 attempting to work them out.
8276 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8277 this allows the use of compression and extensions. Change default cipher
8278 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8279 by default unless an application cipher string requests it.
8283 * Alter match criteria in PKCS12_parse(). It used to try to use local
8284 key ids to find matching certificates and keys but some PKCS#12 files
8285 don't follow the (somewhat unwritten) rules and this strategy fails.
8286 Now just gather all certificates together and the first private key
8287 then look for the first certificate that matches the key.
8291 * Support use of registered digest and cipher names for dgst and cipher
8292 commands instead of having to add each one as a special case. So now
8299 openssl dgst -sha256 foo
8301 and this works for ENGINE based algorithms too.
8305 * Update Gost ENGINE to support parameter files.
8307 *Victor B. Wagner <vitus@cryptocom.ru>*
8309 * Support GeneralizedTime in ca utility.
8311 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8313 * Enhance the hash format used for certificate directory links. The new
8314 form uses the canonical encoding (meaning equivalent names will work
8315 even if they aren't identical) and uses SHA1 instead of MD5. This form
8316 is incompatible with the older format and as a result c_rehash should
8317 be used to rebuild symbolic links.
8321 * Make PKCS#8 the default write format for private keys, replacing the
8322 traditional format. This form is standardised, more secure and doesn't
8323 include an implicit MD5 dependency.
8327 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8328 committed to OpenSSL should pass this lot as a minimum.
8332 * Add session ticket override functionality for use by EAP-FAST.
8334 *Jouni Malinen <j@w1.fi>*
8336 * Modify HMAC functions to return a value. Since these can be implemented
8337 in an ENGINE errors can occur.
8341 * Type-checked OBJ_bsearch_ex.
8345 * Type-checked OBJ_bsearch. Also some constification necessitated
8346 by type-checking. Still to come: TXT_DB, bsearch(?),
8347 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8352 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8353 seconds to a tm structure directly, instead of going through OS
8354 specific date routines. This avoids any issues with OS routines such
8355 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8356 and X509_time_adj_ex() to cover the extended range. The existing
8357 X509_time_adj() is still usable and will no longer have any date issues.
8361 * Delta CRL support. New use deltas option which will attempt to locate
8362 and search any appropriate delta CRLs available.
8364 This work was sponsored by Google.
8368 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8369 code and add additional score elements. Validate alternate CRL paths
8370 as part of the CRL checking and indicate a new error "CRL path validation
8371 error" in this case. Applications wanting additional details can use
8372 the verify callback and check the new "parent" field. If this is not
8373 NULL CRL path validation is taking place. Existing applications won't
8374 see this because it requires extended CRL support which is off by
8377 This work was sponsored by Google.
8381 * Support for freshest CRL extension.
8383 This work was sponsored by Google.
8387 * Initial indirect CRL support. Currently only supported in the CRLs
8388 passed directly and not via lookup. Process certificate issuer
8389 CRL entry extension and lookup CRL entries by bother issuer name
8390 and serial number. Check and process CRL issuer entry in IDP extension.
8392 This work was sponsored by Google.
8396 * Add support for distinct certificate and CRL paths. The CRL issuer
8397 certificate is validated separately in this case. Only enabled if
8398 an extended CRL support flag is set: this flag will enable additional
8399 CRL functionality in future.
8401 This work was sponsored by Google.
8405 * Add support for policy mappings extension.
8407 This work was sponsored by Google.
8411 * Fixes to pathlength constraint, self issued certificate handling,
8412 policy processing to align with RFC3280 and PKITS tests.
8414 This work was sponsored by Google.
8418 * Support for name constraints certificate extension. DN, email, DNS
8419 and URI types are currently supported.
8421 This work was sponsored by Google.
8425 * To cater for systems that provide a pointer-based thread ID rather
8426 than numeric, deprecate the current numeric thread ID mechanism and
8427 replace it with a structure and associated callback type. This
8428 mechanism allows a numeric "hash" to be extracted from a thread ID in
8429 either case, and on platforms where pointers are larger than 'long',
8430 mixing is done to help ensure the numeric 'hash' is usable even if it
8431 can't be guaranteed unique. The default mechanism is to use "&errno"
8432 as a pointer-based thread ID to distinguish between threads.
8434 Applications that want to provide their own thread IDs should now use
8435 CRYPTO_THREADID_set_callback() to register a callback that will call
8436 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8438 Note that ERR_remove_state() is now deprecated, because it is tied
8439 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8440 to free the current thread's error state should be replaced by
8441 ERR_remove_thread_state(NULL).
8443 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8444 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8445 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8446 application was previously providing a numeric thread callback that
8447 was inappropriate for distinguishing threads, then uniqueness might
8448 have been obtained with &errno that happened immediately in the
8449 intermediate development versions of OpenSSL; this is no longer the
8450 case, the numeric thread callback will now override the automatic use
8453 *Geoff Thorpe, with help from Bodo Moeller*
8455 * Initial support for different CRL issuing certificates. This covers a
8456 simple case where the self issued certificates in the chain exist and
8457 the real CRL issuer is higher in the existing chain.
8459 This work was sponsored by Google.
8463 * Removed effectively defunct crypto/store from the build.
8467 * Revamp of STACK to provide stronger type-checking. Still to come:
8468 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8469 ASN1_STRING, CONF_VALUE.
8473 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8474 RAM on SSL connections. This option can save about 34k per idle SSL.
8478 * Revamp of LHASH to provide stronger type-checking. Still to come:
8479 STACK, TXT_DB, bsearch, qsort.
8483 * Initial support for Cryptographic Message Syntax (aka CMS) based
8484 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8485 support for data, signedData, compressedData, digestedData and
8486 encryptedData, envelopedData types included. Scripts to check against
8487 RFC4134 examples draft and interop and consistency checks of many
8488 content types and variants.
8492 * Add options to enc utility to support use of zlib compression BIO.
8496 * Extend mk1mf to support importing of options and assembly language
8497 files from Configure script, currently only included in VC-WIN32.
8498 The assembly language rules can now optionally generate the source
8499 files from the associated perl scripts.
8503 * Implement remaining functionality needed to support GOST ciphersuites.
8504 Interop testing has been performed using CryptoPro implementations.
8506 *Victor B. Wagner <vitus@cryptocom.ru>*
8508 * s390x assembler pack.
8512 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8517 * Implement Opaque PRF Input TLS extension as specified in
8518 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8519 official specification yet and no extension type assignment by
8520 IANA exists, this extension (for now) will have to be explicitly
8521 enabled when building OpenSSL by providing the extension number
8522 to use. For example, specify an option
8524 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8526 to the "config" or "Configure" script to enable the extension,
8527 assuming extension number 0x9527 (which is a completely arbitrary
8528 and unofficial assignment based on the MD5 hash of the Internet
8529 Draft). Note that by doing so, you potentially lose
8530 interoperability with other TLS implementations since these might
8531 be using the same extension number for other purposes.
8533 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8534 opaque PRF input value to use in the handshake. This will create
8535 an internal copy of the length-'len' string at 'src', and will
8536 return non-zero for success.
8538 To get more control and flexibility, provide a callback function
8541 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8542 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8546 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8549 Callback function 'cb' will be called in handshakes, and is
8550 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8551 Argument 'arg' is for application purposes (the value as given to
8552 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8553 be provided to the callback function). The callback function
8554 has to return non-zero to report success: usually 1 to use opaque
8555 PRF input just if possible, or 2 to enforce use of the opaque PRF
8556 input. In the latter case, the library will abort the handshake
8557 if opaque PRF input is not successfully negotiated.
8559 Arguments 'peerinput' and 'len' given to the callback function
8560 will always be NULL and 0 in the case of a client. A server will
8561 see the client's opaque PRF input through these variables if
8562 available (NULL and 0 otherwise). Note that if the server
8563 provides an opaque PRF input, the length must be the same as the
8564 length of the client's opaque PRF input.
8566 Note that the callback function will only be called when creating
8567 a new session (session resumption can resume whatever was
8568 previously negotiated), and will not be called in SSL 2.0
8569 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8570 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8571 for applications that need to enforce opaque PRF input.
8575 * Update ssl code to support digests other than SHA1+MD5 for handshake
8578 *Victor B. Wagner <vitus@cryptocom.ru>*
8580 * Add RFC4507 support to OpenSSL. This includes the corrections in
8581 RFC4507bis. The encrypted ticket format is an encrypted encoded
8582 SSL_SESSION structure, that way new session features are automatically
8585 If a client application caches session in an SSL_SESSION structure
8586 support is transparent because tickets are now stored in the encoded
8589 The SSL_CTX structure automatically generates keys for ticket
8590 protection in servers so again support should be possible
8591 with no application modification.
8593 If a client or server wishes to disable RFC4507 support then the option
8594 SSL_OP_NO_TICKET can be set.
8596 Add a TLS extension debugging callback to allow the contents of any client
8597 or server extensions to be examined.
8599 This work was sponsored by Google.
8603 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8604 OpenSSL should now compile cleanly on gcc 4.2
8606 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8608 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8609 support including streaming MAC support: this is required for GOST
8610 ciphersuite support.
8612 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8614 * Add option -stream to use PKCS#7 streaming in smime utility. New
8615 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8616 to output in BER and PEM format.
8620 * Experimental support for use of HMAC via EVP_PKEY interface. This
8621 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8622 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8623 ENGINE support for HMAC keys which are unextractable. New -mac and
8624 -macopt options to dgst utility.
8628 * New option -sigopt to dgst utility. Update dgst to use
8629 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8630 alternative signing parameters such as X9.31 or PSS in the dgst
8635 * Change ssl_cipher_apply_rule(), the internal function that does
8636 the work each time a ciphersuite string requests enabling
8637 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8638 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8639 the order of disabled ciphersuites such that those ciphersuites
8640 that most recently went from enabled to disabled not only stay
8641 in order with respect to each other, but also have higher priority
8642 than other disabled ciphersuites the next time ciphersuites are
8645 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8646 the same ciphersuites as with "HIGH" alone, but in a specific
8647 order where the PSK ciphersuites come first (since they are the
8648 most recently disabled ciphersuites when "HIGH" is parsed).
8650 Also, change ssl_create_cipher_list() (using this new
8651 functionality) such that between otherwise identical
8652 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8657 * Change ssl_create_cipher_list() so that it automatically
8658 arranges the ciphersuites in reasonable order before starting
8659 to process the rule string. Thus, the definition for "DEFAULT"
8660 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8661 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8662 This makes it much easier to arrive at a reasonable default order
8663 in applications for which anonymous ciphers are OK (meaning
8664 that you can't actually use DEFAULT).
8666 *Bodo Moeller; suggested by Victor Duchovni*
8668 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8669 processing) into multiple integers instead of setting
8670 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8671 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8672 (These masks as well as the individual bit definitions are hidden
8673 away into the non-exported interface ssl/ssl_locl.h, so this
8674 change to the definition of the SSL_CIPHER structure shouldn't
8675 affect applications.) This give us more bits for each of these
8676 categories, so there is no longer a need to coagulate AES128 and
8677 AES256 into a single algorithm bit, and to coagulate Camellia128
8678 and Camellia256 into a single algorithm bit, which has led to all
8681 Thus, among other things, the kludge introduced in 0.9.7m and
8682 0.9.8e for masking out AES256 independently of AES128 or masking
8683 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8685 With the change, we also introduce new ciphersuite aliases that
8686 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8691 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8692 Use the leftmost N bytes of the signature input if the input is
8693 larger than the prime q (with N being the size in bytes of q).
8697 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8698 it yet and it is largely untested.
8702 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8706 * Initial incomplete changes to avoid need for function casts in OpenSSL
8707 some compilers (gcc 4.2 and later) reject their use. Safestack is
8708 reimplemented. Update ASN1 to avoid use of legacy functions.
8712 * Win32/64 targets are linked with Winsock2.
8716 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8717 to external functions. This can be used to increase CRL handling
8718 efficiency especially when CRLs are very large by (for example) storing
8719 the CRL revoked certificates in a database.
8723 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8724 new CRLs added to a directory can be used. New command line option
8725 -verify_return_error to s_client and s_server. This causes real errors
8726 to be returned by the verify callback instead of carrying on no matter
8727 what. This reflects the way a "real world" verify callback would behave.
8731 * GOST engine, supporting several GOST algorithms and public key formats.
8732 Kindly donated by Cryptocom.
8736 * Partial support for Issuing Distribution Point CRL extension. CRLs
8737 partitioned by DP are handled but no indirect CRL or reason partitioning
8738 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8739 selected via a scoring technique which handles IDP and AKID in CRLs.
8743 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8744 will ultimately be used for all verify operations: this will remove the
8745 X509_STORE dependency on certificate verification and allow alternative
8746 lookup methods. X509_STORE based implementations of these two callbacks.
8750 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8751 Modify get_crl() to find a valid (unexpired) CRL if possible.
8755 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8756 this would be called X509_CRL_cmp() but that name is already used by
8757 a function that just compares CRL issuer names. Cache several CRL
8758 extensions in X509_CRL structure and cache CRLDP in X509.
8762 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8763 this maps equivalent X509_NAME structures into a consistent structure.
8764 Name comparison can then be performed rapidly using memcmp().
8768 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8773 * Allow digests to supply their own micalg string for S/MIME type using
8774 the ctrl EVP_MD_CTRL_MICALG.
8778 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8779 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8780 ctrl. It can then customise the structure before and/or after signing
8785 * New function OBJ_add_sigid() to allow application defined signature OIDs
8786 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8787 to free up any added signature OIDs.
8791 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8792 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8793 digest and cipher tables. New options added to openssl utility:
8794 list-message-digest-algorithms and list-cipher-algorithms.
8798 * Change the array representation of binary polynomials: the list
8799 of degrees of non-zero coefficients is now terminated with -1.
8800 Previously it was terminated with 0, which was also part of the
8801 value; thus, the array representation was not applicable to
8802 polynomials where t^0 has coefficient zero. This change makes
8803 the array representation useful in a more general context.
8807 * Various modifications and fixes to SSL/TLS cipher string
8808 handling. For ECC, the code now distinguishes between fixed ECDH
8809 with RSA certificates on the one hand and with ECDSA certificates
8810 on the other hand, since these are separate ciphersuites. The
8811 unused code for Fortezza ciphersuites has been removed.
8813 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8814 (not "ECDHE"). For consistency with the code for DH
8815 certificates, use of ECDH certificates is now considered ECDH
8816 authentication, not RSA or ECDSA authentication (the latter is
8817 merely the CA's signing algorithm and not actively used in the
8820 The temporary ciphersuite alias "ECCdraft" is no longer
8821 available, and ECC ciphersuites are no longer excluded from "ALL"
8822 and "DEFAULT". The following aliases now exist for RFC 4492
8823 ciphersuites, most of these by analogy with the DH case:
8825 kECDHr - ECDH cert, signed with RSA
8826 kECDHe - ECDH cert, signed with ECDSA
8827 kECDH - ECDH cert (signed with either RSA or ECDSA)
8828 kEECDH - ephemeral ECDH
8829 ECDH - ECDH cert or ephemeral ECDH
8835 AECDH - anonymous ECDH
8836 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8840 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8841 Use correct micalg parameters depending on digest(s) in signed message.
8845 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8846 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8850 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8851 an engine to register a method. Add ENGINE lookups for methods and
8852 functional reference processing.
8856 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8857 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8862 * New -resign option to smime utility. This adds one or more signers
8863 to an existing PKCS#7 signedData structure. Also -md option to use an
8864 alternative message digest algorithm for signing.
8868 * Tidy up PKCS#7 routines and add new functions to make it easier to
8869 create PKCS7 structures containing multiple signers. Update smime
8870 application to support multiple signers.
8874 * New -macalg option to pkcs12 utility to allow setting of an alternative
8879 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8880 Reorganize PBE internals to lookup from a static table using NIDs,
8881 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8882 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8883 PRF which will be automatically used with PBES2.
8887 * Replace the algorithm specific calls to generate keys in "req" with the
8892 * Update PKCS#7 enveloped data routines to use new API. This is now
8893 supported by any public key method supporting the encrypt operation. A
8894 ctrl is added to allow the public key algorithm to examine or modify
8895 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8900 * Add a ctrl to asn1 method to allow a public key algorithm to express
8901 a default digest type to use. In most cases this will be SHA1 but some
8902 algorithms (such as GOST) need to specify an alternative digest. The
8903 return value indicates how strong the preference is 1 means optional and
8904 2 is mandatory (that is it is the only supported type). Modify
8905 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8906 use the default md. Update openssl utilities to use the default digest
8907 type for signing if it is not explicitly indicated.
8911 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8912 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8913 signing method from the key type. This effectively removes the link
8914 between digests and public key types.
8918 * Add an OID cross reference table and utility functions. Its purpose is to
8919 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8920 rsaEncryption. This will allow some of the algorithm specific hackery
8921 needed to use the correct OID to be removed.
8925 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8926 structures for PKCS7_sign(). They are now set up by the relevant public
8931 * Add provisional EC pkey method with support for ECDSA and ECDH.
8935 * Add support for key derivation (agreement) in the API, DH method and
8940 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8941 public and private key formats. As a side effect these add additional
8942 command line functionality not previously available: DSA signatures can be
8943 generated and verified using pkeyutl and DH key support and generation in
8950 *Oliver Tappe <zooey@hirschkaefer.de>*
8952 * New make target "install_html_docs" installs HTML renditions of the
8955 *Oliver Tappe <zooey@hirschkaefer.de>*
8957 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8958 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8959 support key and parameter generation and add initial key generation
8960 functionality for RSA.
8964 * Add functions for main EVP_PKEY_method operations. The undocumented
8965 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8966 `EVP_PKEY_{encrypt,decrypt}_old`.
8970 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8971 key API, doesn't do much yet.
8975 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8976 public key algorithms. New option to openssl utility:
8977 "list-public-key-algorithms" to print out info.
8981 * Implement the Supported Elliptic Curves Extension for
8982 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8986 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8987 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8991 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8992 utilities such as rsa, dsa, dsaparam etc except they process any key
8997 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8998 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8999 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9004 * Initial support for pluggable public key ASN1.
9005 De-spaghettify the public key ASN1 handling. Move public and private
9006 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9007 algorithm specific handling to a single module within the relevant
9008 algorithm directory. Add functions to allow (near) opaque processing
9009 of public and private key structures.
9013 * Implement the Supported Point Formats Extension for
9014 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9018 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9019 for the psk identity [hint] and the psk callback functions to the
9020 SSL_SESSION, SSL and SSL_CTX structure.
9023 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9027 SSL_CTX_use_psk_identity_hint
9028 SSL_get_psk_identity_hint
9029 SSL_get_psk_identity
9030 SSL_use_psk_identity_hint
9032 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9034 * Add RFC 3161 compliant time stamp request creation, response generation
9035 and response verification functionality.
9037 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9039 * Add initial support for TLS extensions, specifically for the server_name
9040 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9041 have new members for a host name. The SSL data structure has an
9042 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9043 stored in that context to allow for session resumption, even after the
9044 SSL has been switched to a new SSL_CTX in reaction to a client's
9045 server_name extension.
9047 New functions (subject to change):
9049 SSL_get_servername()
9050 SSL_get_servername_type()
9053 New CTRL codes and macros (subject to change):
9055 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9056 - SSL_CTX_set_tlsext_servername_callback()
9057 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9058 - SSL_CTX_set_tlsext_servername_arg()
9059 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9061 openssl s_client has a new '-servername ...' option.
9063 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9064 '-key2 ...', '-servername_fatal' (subject to change). This allows
9065 testing the HostName extension for a specific single host name ('-cert'
9066 and '-key' remain fallbacks for handshakes without HostName
9067 negotiation). If the unrecognized_name alert has to be sent, this by
9068 default is a warning; it becomes fatal with the '-servername_fatal'
9071 *Peter Sylvester, Remy Allais, Christophe Renou*
9073 * Whirlpool hash implementation is added.
9077 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9078 bn(64,32). Because of instruction set limitations it doesn't have
9079 any negative impact on performance. This was done mostly in order
9080 to make it possible to share assembler modules, such as bn_mul_mont
9081 implementations, between 32- and 64-bit builds without hassle.
9085 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9086 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9091 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9092 dedicated Montgomery multiplication procedure, is introduced.
9093 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9094 "64-bit" performance on certain 32-bit targets.
9098 * New option SSL_OP_NO_COMP to disable use of compression selectively
9099 in SSL structures. New SSL ctrl to set maximum send fragment size.
9100 Save memory by setting the I/O buffer sizes dynamically instead of
9101 using the maximum available value.
9105 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9106 in addition to the text details.
9110 * Very, very preliminary EXPERIMENTAL support for printing of general
9111 ASN1 structures. This currently produces rather ugly output and doesn't
9112 handle several customised structures at all.
9116 * Integrated support for PVK file format and some related formats such
9117 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9118 these in the 'rsa' and 'dsa' utilities.
9122 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9126 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9127 place for the (very old) "NETSCAPE" format certificates which are now
9128 handled using new ASN1 code equivalents.
9132 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9133 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9134 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9138 * Modify CRL distribution points extension code to print out previously
9139 unsupported fields. Enhance extension setting code to allow setting of
9144 * Add print and set support for Issuing Distribution Point CRL extension.
9148 * Change 'Configure' script to enable Camellia by default.
9155 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9157 * When rejecting SSL/TLS records due to an incorrect version number, never
9158 update s->server with a new major version number. As of
9159 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9160 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9161 the previous behavior could result in a read attempt at NULL when
9162 receiving specific incorrect SSL/TLS records once record payload
9163 protection is active. ([CVE-2010-0740])
9165 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9167 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9168 could be crashed if the relevant tables were not present (e.g. chrooted).
9170 *Tomas Hoger <thoger@redhat.com>*
9172 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9174 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9176 *Martin Olsson, Neel Mehta*
9178 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9179 accommodate for stack sorting, always a write lock!).
9183 * On some versions of WIN32 Heap32Next is very slow. This can cause
9184 excessive delays in the RAND_poll(): over a minute. As a workaround
9185 include a time check in the inner Heap32Next loop too.
9189 * The code that handled flushing of data in SSL/TLS originally used the
9190 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9191 the problem outlined in PR#1949. The fix suggested there however can
9192 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9193 of Apache). So instead simplify the code to flush unconditionally.
9194 This should be fine since flushing with no data to flush is a no op.
9198 * Handle TLS versions 2.0 and later properly and correctly use the
9199 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9200 off ancient servers have a habit of sticking around for a while...
9204 * Modify compression code so it frees up structures without using the
9205 ex_data callbacks. This works around a problem where some applications
9206 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9207 restarting) then use compression (e.g. SSL with compression) later.
9208 This results in significant per-connection memory leaks and
9209 has caused some security issues including CVE-2008-1678 and
9214 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9215 change when encrypting or decrypting.
9219 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9220 connect and renegotiate with servers which do not support RI.
9221 Until RI is more widely deployed this option is enabled by default.
9225 * Add "missing" ssl ctrls to clear options and mode.
9229 * If client attempts to renegotiate and doesn't support RI respond with
9230 a no_renegotiation alert as required by RFC5746. Some renegotiating
9231 TLS clients will continue a connection gracefully when they receive
9232 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9233 waiting for a server hello which it will never receive. Now we treat a
9234 received no_renegotiation alert as a fatal error. This is because
9235 applications requesting a renegotiation might well expect it to succeed
9236 and would have no code in place to handle the server denying it so the
9237 only safe thing to do is to terminate the connection.
9241 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9242 peer supports secure renegotiation and 0 otherwise. Print out peer
9243 renegotiation support in s_client/s_server.
9247 * Replace the highly broken and deprecated SPKAC certification method with
9248 the updated NID creation version. This should correctly handle UTF8.
9252 * Implement RFC5746. Re-enable renegotiation but require the extension
9253 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9254 turns out to be a bad idea. It has been replaced by
9255 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9256 SSL_CTX_set_options(). This is really not recommended unless you
9257 know what you are doing.
9259 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9261 * Fixes to stateless session resumption handling. Use initial_ctx when
9262 issuing and attempting to decrypt tickets in case it has changed during
9263 servername handling. Use a non-zero length session ID when attempting
9264 stateless session resumption: this makes it possible to determine if
9265 a resumption has occurred immediately after receiving server hello
9266 (several places in OpenSSL subtly assume this) instead of later in
9271 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9272 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9273 fixes for a few places where the return code is not checked
9276 *Julia Lawall <julia@diku.dk>*
9278 * Add --strict-warnings option to Configure script to include devteam
9279 warnings in other configurations.
9283 * Add support for --libdir option and LIBDIR variable in makefiles. This
9284 makes it possible to install openssl libraries in locations which
9285 have names other than "lib", for example "/usr/lib64" which some
9288 *Steve Henson, based on patch from Jeremy Utley*
9290 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9291 X690 8.9.12 and can produce some misleading textual output of OIDs.
9293 *Steve Henson, reported by Dan Kaminsky*
9295 * Delete MD2 from algorithm tables. This follows the recommendation in
9296 several standards that it is not used in new applications due to
9297 several cryptographic weaknesses. For binary compatibility reasons
9298 the MD2 API is still compiled in by default.
9302 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9307 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9308 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9311 *Guenter <lists@gknw.net>*
9313 * Fix the server certificate chain building code to use X509_verify_cert(),
9314 it used to have an ad-hoc builder which was unable to cope with anything
9315 other than a simple chain.
9317 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9319 * Don't check self signed certificate signatures in X509_verify_cert()
9320 by default (a flag can override this): it just wastes time without
9321 adding any security. As a useful side effect self signed root CAs
9322 with non-FIPS digests are now usable in FIPS mode.
9326 * In dtls1_process_out_of_seq_message() the check if the current message
9327 is already buffered was missing. For every new message was memory
9328 allocated, allowing an attacker to perform an denial of service attack
9329 with sending out of seq handshake messages until there is no memory
9330 left. Additionally every future message was buffered, even if the
9331 sequence number made no sense and would be part of another handshake.
9332 So only messages with sequence numbers less than 10 in advance will be
9333 buffered. ([CVE-2009-1378])
9335 *Robin Seggelmann, discovered by Daniel Mentz*
9337 * Records are buffered if they arrive with a future epoch to be
9338 processed after finishing the corresponding handshake. There is
9339 currently no limitation to this buffer allowing an attacker to perform
9340 a DOS attack with sending records with future epochs until there is no
9341 memory left. This patch adds the pqueue_size() function to determine
9342 the size of a buffer and limits the record buffer to 100 entries.
9345 *Robin Seggelmann, discovered by Daniel Mentz*
9347 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9348 parent structure is freed. ([CVE-2009-1379])
9352 * Handle non-blocking I/O properly in SSL_shutdown() call.
9354 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9356 * Add `2.5.4.*` OIDs
9358 *Ilya O. <vrghost@gmail.com>*
9360 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9362 * Disable renegotiation completely - this fixes a severe security
9363 problem ([CVE-2009-3555]) at the cost of breaking all
9364 renegotiation. Renegotiation can be re-enabled by setting
9365 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9366 run-time. This is really not recommended unless you know what
9371 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9373 * Don't set val to NULL when freeing up structures, it is freed up by
9374 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9375 zeroing past the valid field. ([CVE-2009-0789])
9377 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9379 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9380 checked correctly. This would allow some invalid signed attributes to
9381 appear to verify correctly. ([CVE-2009-0591])
9383 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9385 * Reject UniversalString and BMPString types with invalid lengths. This
9386 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9387 a legal length. ([CVE-2009-0590])
9391 * Set S/MIME signing as the default purpose rather than setting it
9392 unconditionally. This allows applications to override it at the store
9397 * Permit restricted recursion of ASN1 strings. This is needed in practice
9398 to handle some structures.
9402 * Improve efficiency of mem_gets: don't search whole buffer each time
9405 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9407 * New -hex option for openssl rand.
9411 * Print out UTF8String and NumericString when parsing ASN1.
9415 * Support NumericString type for name components.
9419 * Allow CC in the environment to override the automatically chosen
9420 compiler. Note that nothing is done to ensure flags work with the
9425 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9427 * Properly check EVP_VerifyFinal() and similar return values
9430 *Ben Laurie, Bodo Moeller, Google Security Team*
9432 * Enable TLS extensions by default.
9436 * Allow the CHIL engine to be loaded, whether the application is
9437 multithreaded or not. (This does not release the developer from the
9438 obligation to set up the dynamic locking callbacks.)
9440 *Sander Temme <sander@temme.net>*
9442 * Use correct exit code if there is an error in dgst command.
9444 *Steve Henson; problem pointed out by Roland Dirlewanger*
9446 * Tweak Configure so that you need to say "experimental-jpake" to enable
9447 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9451 * Add experimental JPAKE support, including demo authentication in
9452 s_client and s_server.
9456 * Set the comparison function in v3_addr_canonize().
9458 *Rob Austein <sra@hactrn.net>*
9460 * Add support for XMPP STARTTLS in s_client.
9462 *Philip Paeps <philip@freebsd.org>*
9464 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9465 to ensure that even with this option, only ciphersuites in the
9466 server's preference list will be accepted. (Note that the option
9467 applies only when resuming a session, so the earlier behavior was
9468 just about the algorithm choice for symmetric cryptography.)
9472 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9474 * Fix NULL pointer dereference if a DTLS server received
9475 ChangeCipherSpec as first record ([CVE-2009-1386]).
9479 * Fix a state transition in s3_srvr.c and d1_srvr.c
9480 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9484 * The fix in 0.9.8c that supposedly got rid of unsafe
9485 double-checked locking was incomplete for RSA blinding,
9486 addressing just one layer of what turns out to have been
9487 doubly unsafe triple-checked locking.
9489 So now fix this for real by retiring the MONT_HELPER macro
9490 in crypto/rsa/rsa_eay.c.
9492 *Bodo Moeller; problem pointed out by Marius Schilder*
9494 * Various precautionary measures:
9496 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9498 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9499 (NB: This would require knowledge of the secret session ticket key
9500 to exploit, in which case you'd be SOL either way.)
9502 - Change bn_nist.c so that it will properly handle input BIGNUMs
9503 outside the expected range.
9505 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9508 *Neel Mehta, Bodo Moeller*
9510 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9511 the load fails. Useful for distros.
9513 *Ben Laurie and the FreeBSD team*
9515 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9519 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9523 * Expand ENGINE to support engine supplied SSL client certificate functions.
9525 This work was sponsored by Logica.
9529 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9530 keystores. Support for SSL/TLS client authentication too.
9531 Not compiled unless enable-capieng specified to Configure.
9533 This work was sponsored by Logica.
9537 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9538 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9539 attribute creation routines such as certificate requests and PKCS#12
9544 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9546 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9547 handshake which could lead to a client crash as found using the
9548 Codenomicon TLS test suite ([CVE-2008-1672])
9550 *Steve Henson, Mark Cox*
9552 * Fix double free in TLS server name extensions which could lead to
9553 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9557 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9559 Clear the error queue to ensure that error entries left from
9560 older function calls do not interfere with the correct operation.
9562 *Lutz Jaenicke, Erik de Castro Lopo*
9564 * Remove root CA certificates of commercial CAs:
9566 The OpenSSL project does not recommend any specific CA and does not
9567 have any policy with respect to including or excluding any CA.
9568 Therefore it does not make any sense to ship an arbitrary selection
9569 of root CA certificates with the OpenSSL software.
9573 * RSA OAEP patches to fix two separate invalid memory reads.
9574 The first one involves inputs when 'lzero' is greater than
9575 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9576 before the beginning of from). The second one involves inputs where
9577 the 'db' section contains nothing but zeroes (there is a one-byte
9578 invalid read after the end of 'db').
9580 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9582 * Partial backport from 0.9.9-dev:
9584 Introduce bn_mul_mont (dedicated Montgomery multiplication
9585 procedure) as a candidate for BIGNUM assembler implementation.
9586 While 0.9.9-dev uses assembler for various architectures, only
9587 x86_64 is available by default here in the 0.9.8 branch, and
9588 32-bit x86 is available through a compile-time setting.
9590 To try the 32-bit x86 assembler implementation, use Configure
9591 option "enable-montasm" (which exists only for this backport).
9593 As "enable-montasm" for 32-bit x86 disclaims code stability
9594 anyway, in this constellation we activate additional code
9595 backported from 0.9.9-dev for further performance improvements,
9596 namely BN_from_montgomery_word. (To enable this otherwise,
9597 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9599 *Andy Polyakov (backport partially by Bodo Moeller)*
9601 * Add TLS session ticket callback. This allows an application to set
9602 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9603 values. This is useful for key rollover for example where several key
9604 sets may exist with different names.
9608 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9609 This was broken until now in 0.9.8 releases, such that the only way
9610 a registered ENGINE could be used (assuming it initialises
9611 successfully on the host) was to explicitly set it as the default
9612 for the relevant algorithms. This is in contradiction with 0.9.7
9613 behaviour and the documentation. With this fix, when an ENGINE is
9614 registered into a given algorithm's table of implementations, the
9615 'uptodate' flag is reset so that auto-discovery will be used next
9616 time a new context for that algorithm attempts to select an
9619 *Ian Lister (tweaked by Geoff Thorpe)*
9621 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9622 implementation in the following ways:
9624 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9627 Lack of BER streaming support means one pass streaming processing is
9628 only supported if data is detached: setting the streaming flag is
9629 ignored for embedded content.
9631 CMS support is disabled by default and must be explicitly enabled
9632 with the enable-cms configuration option.
9636 * Update the GMP engine glue to do direct copies between BIGNUM and
9637 mpz_t when openssl and GMP use the same limb size. Otherwise the
9638 existing "conversion via a text string export" trick is still used.
9640 *Paul Sheer <paulsheer@gmail.com>*
9642 * Zlib compression BIO. This is a filter BIO which compressed and
9643 uncompresses any data passed through it.
9647 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9648 RFC3394 compatible AES key wrapping.
9652 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9653 sets string data without copying. X509_ALGOR_set0() and
9654 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9655 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9656 from an X509_ATTRIBUTE structure optionally checking it occurs only
9657 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9662 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9663 to get the expected BN_FLG_CONSTTIME behavior.
9665 *Bodo Moeller (Google)*
9669 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9670 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9671 - added some more tests to do_tests.pl
9672 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9673 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9674 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9675 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9676 - various changes to netware.pl to enable gcc-cross builds on Win32
9678 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9679 - various changes to fix missing prototype warnings
9680 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9681 - added AES, WHIRLPOOL and CPUID assembler code to build files
9682 - added missing AES assembler make rules to mk1mf.pl
9683 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9685 *Guenter Knauf <eflash@gmx.net>*
9687 * Implement certificate status request TLS extension defined in RFC3546.
9688 A client can set the appropriate parameters and receive the encoded
9689 OCSP response via a callback. A server can query the supplied parameters
9690 and set the encoded OCSP response in the callback. Add simplified examples
9691 to s_client and s_server.
9695 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9698 + Binary incompatibility of ssl_ctx_st structure
9699 + DTLS interoperation with non-compliant servers
9700 + Don't call get_session_cb() without proposed session
9701 + Fix ia64 assembler code
9703 *Andy Polyakov, Steve Henson*
9705 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9707 * DTLS Handshake overhaul. There were longstanding issues with
9708 OpenSSL DTLS implementation, which were making it impossible for
9709 RFC 4347 compliant client to communicate with OpenSSL server.
9710 Unfortunately just fixing these incompatibilities would "cut off"
9711 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9712 server keeps tolerating non RFC compliant syntax. The opposite is
9713 not true, 0.9.8f client can not communicate with earlier server.
9714 This update even addresses CVE-2007-4995.
9718 * Changes to avoid need for function casts in OpenSSL: some compilers
9719 (gcc 4.2 and later) reject their use.
9720 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9723 * Add RFC4507 support to OpenSSL. This includes the corrections in
9724 RFC4507bis. The encrypted ticket format is an encrypted encoded
9725 SSL_SESSION structure, that way new session features are automatically
9728 If a client application caches session in an SSL_SESSION structure
9729 support is transparent because tickets are now stored in the encoded
9732 The SSL_CTX structure automatically generates keys for ticket
9733 protection in servers so again support should be possible
9734 with no application modification.
9736 If a client or server wishes to disable RFC4507 support then the option
9737 SSL_OP_NO_TICKET can be set.
9739 Add a TLS extension debugging callback to allow the contents of any client
9740 or server extensions to be examined.
9742 This work was sponsored by Google.
9746 * Add initial support for TLS extensions, specifically for the server_name
9747 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9748 have new members for a host name. The SSL data structure has an
9749 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9750 stored in that context to allow for session resumption, even after the
9751 SSL has been switched to a new SSL_CTX in reaction to a client's
9752 server_name extension.
9754 New functions (subject to change):
9756 SSL_get_servername()
9757 SSL_get_servername_type()
9760 New CTRL codes and macros (subject to change):
9762 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9763 - SSL_CTX_set_tlsext_servername_callback()
9764 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9765 - SSL_CTX_set_tlsext_servername_arg()
9766 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9768 openssl s_client has a new '-servername ...' option.
9770 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9771 '-key2 ...', '-servername_fatal' (subject to change). This allows
9772 testing the HostName extension for a specific single host name ('-cert'
9773 and '-key' remain fallbacks for handshakes without HostName
9774 negotiation). If the unrecognized_name alert has to be sent, this by
9775 default is a warning; it becomes fatal with the '-servername_fatal'
9778 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9780 * Add AES and SSE2 assembly language support to VC++ build.
9784 * Mitigate attack on final subtraction in Montgomery reduction.
9788 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9789 (which previously caused an internal error).
9793 * Squeeze another 10% out of IGE mode when in != out.
9797 * AES IGE mode speedup.
9799 *Dean Gaudet (Google)*
9801 * Add the Korean symmetric 128-bit cipher SEED (see
9802 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9803 add SEED ciphersuites from RFC 4162:
9805 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9806 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9807 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9808 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9810 To minimize changes between patchlevels in the OpenSSL 0.9.8
9811 series, SEED remains excluded from compilation unless OpenSSL
9812 is configured with 'enable-seed'.
9814 *KISA, Bodo Moeller*
9816 * Mitigate branch prediction attacks, which can be practical if a
9817 single processor is shared, allowing a spy process to extract
9818 information. For detailed background information, see
9819 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9820 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9821 and Necessary Software Countermeasures"). The core of the change
9822 are new versions BN_div_no_branch() and
9823 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9824 respectively, which are slower, but avoid the security-relevant
9825 conditional branches. These are automatically called by BN_div()
9826 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9827 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9828 remove a conditional branch.
9830 BN_FLG_CONSTTIME is the new name for the previous
9831 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9832 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9833 in the exponent causes BN_mod_exp_mont() to use the alternative
9834 implementation in BN_mod_exp_mont_consttime().) The old name
9835 remains as a deprecated alias.
9837 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9838 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9839 constant-time implementations for more than just exponentiation.
9840 Here too the old name is kept as a deprecated alias.
9842 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9843 the BN_BLINDING structure gets an independent copy of the
9844 modulus. This means that the previous `BIGNUM *m` argument to
9845 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9846 essentially becomes `const BIGNUM *m`, although we can't actually
9847 change this in the header file before 0.9.9. It allows
9848 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9849 enable BN_FLG_CONSTTIME.
9851 *Matthew D Wood (Intel Corp)*
9853 * In the SSL/TLS server implementation, be strict about session ID
9854 context matching (which matters if an application uses a single
9855 external cache for different purposes). Previously,
9856 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9857 set. This did ensure strict client verification, but meant that,
9858 with applications using a single external cache for quite
9859 different requirements, clients could circumvent ciphersuite
9860 restrictions for a given session ID context by starting a session
9861 in a different context.
9865 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9866 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9867 authentication-only ciphersuites.
9871 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9872 not complete and could lead to a possible single byte overflow
9873 ([CVE-2007-5135]) [Ben Laurie]
9875 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9877 * Since AES128 and AES256 (and similarly Camellia128 and
9878 Camellia256) share a single mask bit in the logic of
9879 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9880 kludge to work properly if AES128 is available and AES256 isn't
9881 (or if Camellia128 is available and Camellia256 isn't).
9885 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9886 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9887 When a point or a seed is encoded in a BIT STRING, we need to
9888 prevent the removal of trailing zero bits to get the proper DER
9889 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9890 of a NamedBitList, for which trailing 0 bits need to be removed.)
9894 * Have SSL/TLS server implementation tolerate "mismatched" record
9895 protocol version while receiving ClientHello even if the
9896 ClientHello is fragmented. (The server can't insist on the
9897 particular protocol version it has chosen before the ServerHello
9898 message has informed the client about his choice.)
9902 * Add RFC 3779 support.
9904 *Rob Austein for ARIN, Ben Laurie*
9906 * Load error codes if they are not already present instead of using a
9907 static variable. This allows them to be cleanly unloaded and reloaded.
9908 Improve header file function name parsing.
9912 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9913 or CAPABILITY handshake as required by RFCs.
9917 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9919 * Introduce limits to prevent malicious keys being able to
9920 cause a denial of service. ([CVE-2006-2940])
9922 *Steve Henson, Bodo Moeller*
9924 * Fix ASN.1 parsing of certain invalid structures that can result
9925 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9927 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9928 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9930 * Fix SSL client code which could crash if connecting to a
9931 malicious SSLv2 server. ([CVE-2006-4343])
9933 *Tavis Ormandy and Will Drewry, Google Security Team*
9935 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9936 match only those. Before that, "AES256-SHA" would be interpreted
9937 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9938 the same strength classification in 0.9.7h) as we currently only
9939 have a single AES bit in the ciphersuite description bitmap.
9940 That change, however, also applied to ciphersuite strings such as
9941 "RC4-MD5" that intentionally matched multiple ciphersuites --
9942 namely, SSL 2.0 ciphersuites in addition to the more common ones
9943 from SSL 3.0/TLS 1.0.
9945 So we change the selection algorithm again: Naming an explicit
9946 ciphersuite selects this one ciphersuite, and any other similar
9947 ciphersuite (same bitmap) from *other* protocol versions.
9948 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9949 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9951 Since SSL 2.0 does not have any ciphersuites for which the
9952 128/256 bit distinction would be relevant, this works for now.
9953 The proper fix will be to use different bits for AES128 and
9954 AES256, which would have avoided the problems from the beginning;
9955 however, bits are scarce, so we can only do this in a new release
9956 (not just a patchlevel) when we can change the SSL_CIPHER
9957 definition to split the single 'unsigned long mask' bitmap into
9958 multiple values to extend the available space.
9962 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9964 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9965 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9967 * Add AES IGE and biIGE modes.
9971 * Change the Unix randomness entropy gathering to use poll() when
9972 possible instead of select(), since the latter has some
9973 undesirable limitations.
9975 *Darryl Miles via Richard Levitte and Bodo Moeller*
9977 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9978 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9979 cannot be implicitly activated as part of, e.g., the "AES" alias.
9980 However, please upgrade to OpenSSL 0.9.9[-dev] for
9981 non-experimental use of the ECC ciphersuites to get TLS extension
9982 support, which is required for curve and point format negotiation
9983 to avoid potential handshake problems.
9987 * Disable rogue ciphersuites:
9989 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9990 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9991 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9993 The latter two were purportedly from
9994 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9997 Also deactivate the remaining ciphersuites from
9998 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9999 unofficial, and the ID has long expired.
10003 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10004 dual-core machines) and other potential thread-safety issues.
10008 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10009 versions), which is now available for royalty-free use
10010 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10011 Also, add Camellia TLS ciphersuites from RFC 4132.
10013 To minimize changes between patchlevels in the OpenSSL 0.9.8
10014 series, Camellia remains excluded from compilation unless OpenSSL
10015 is configured with 'enable-camellia'.
10019 * Disable the padding bug check when compression is in use. The padding
10020 bug check assumes the first packet is of even length, this is not
10021 necessarily true if compression is enabled and can result in false
10022 positives causing handshake failure. The actual bug test is ancient
10023 code so it is hoped that implementations will either have fixed it by
10024 now or any which still have the bug do not support compression.
10028 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10030 * When applying a cipher rule check to see if string match is an explicit
10031 cipher suite and only match that one cipher suite if it is.
10035 * Link in manifests for VC++ if needed.
10037 *Austin Ziegler <halostatue@gmail.com>*
10039 * Update support for ECC-based TLS ciphersuites according to
10040 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10041 TLS extensions, which are supported starting with the 0.9.9
10042 branch, not in the OpenSSL 0.9.8 branch).
10046 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10047 opaque EVP_CIPHER_CTX handling.
10051 * Fixes and enhancements to zlib compression code. We now only use
10052 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10053 to conform with the standards mentioned here:
10054 <http://www.zlib.net/DLL_FAQ.txt>
10055 Static zlib linking now works on Windows and the new --with-zlib-include
10056 --with-zlib-lib options to Configure can be used to supply the location
10057 of the headers and library. Gracefully handle case where zlib library
10062 * Several fixes and enhancements to the OID generation code. The old code
10063 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10064 handle numbers larger than ULONG_MAX, truncated printing and had a
10065 non standard OBJ_obj2txt() behaviour.
10069 * Add support for building of engines under engine/ as shared libraries
10070 under VC++ build system.
10074 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10075 Hopefully, we will not see any false combination of paths any more.
10079 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10081 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10082 (part of SSL_OP_ALL). This option used to disable the
10083 countermeasure against man-in-the-middle protocol-version
10084 rollback in the SSL 2.0 server implementation, which is a bad
10085 idea. ([CVE-2005-2969])
10087 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10088 for Information Security, National Institute of Advanced Industrial
10089 Science and Technology [AIST], Japan)*
10091 * Add two function to clear and return the verify parameter flags.
10095 * Keep cipherlists sorted in the source instead of sorting them at
10096 runtime, thus removing the need for a lock.
10100 * Avoid some small subgroup attacks in Diffie-Hellman.
10102 *Nick Mathewson and Ben Laurie*
10104 * Add functions for well-known primes.
10108 * Extended Windows CE support.
10110 *Satoshi Nakamura and Andy Polyakov*
10112 * Initialize SSL_METHOD structures at compile time instead of during
10113 runtime, thus removing the need for a lock.
10117 * Make PKCS7_decrypt() work even if no certificate is supplied by
10118 attempting to decrypt each encrypted key in turn. Add support to
10123 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10125 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10128 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10132 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10133 key into the same file any more.
10137 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10141 * Add -utf8 command line and config file option to 'ca'.
10143 *Stefan <stf@udoma.org*
10145 * Removed the macro des_crypt(), as it seems to conflict with some
10146 libraries. Use DES_crypt().
10150 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10151 involves renaming the source and generated shared-libs for
10152 both. The engines will accept the corrected or legacy ids
10153 ('ncipher' and '4758_cca' respectively) when binding. NB,
10154 this only applies when building 'shared'.
10156 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10158 * Add attribute functions to EVP_PKEY structure. Modify
10159 PKCS12_create() to recognize a CSP name attribute and
10160 use it. Make -CSP option work again in pkcs12 utility.
10164 * Add new functionality to the bn blinding code:
10165 - automatic re-creation of the BN_BLINDING parameters after
10166 a fixed number of uses (currently 32)
10167 - add new function for parameter creation
10168 - introduce flags to control the update behaviour of the
10169 BN_BLINDING parameters
10170 - hide BN_BLINDING structure
10171 Add a second BN_BLINDING slot to the RSA structure to improve
10172 performance when a single RSA object is shared among several
10177 * Add support for DTLS.
10179 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10181 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10182 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10186 * Remove buggy and incomplete DH cert support from
10187 ssl/ssl_rsa.c and ssl/s3_both.c
10191 * Use SHA-1 instead of MD5 as the default digest algorithm for
10192 the `apps/openssl` commands.
10196 * Compile clean with "-Wall -Wmissing-prototypes
10197 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10198 DEBUG_SAFESTACK must also be set.
10202 * Change ./Configure so that certain algorithms can be disabled by default.
10203 The new counterpiece to "no-xxx" is "enable-xxx".
10205 The patented RC5 and MDC2 algorithms will now be disabled unless
10206 "enable-rc5" and "enable-mdc2", respectively, are specified.
10208 (IDEA remains enabled despite being patented. This is because IDEA
10209 is frequently required for interoperability, and there is no license
10210 fee for non-commercial use. As before, "no-idea" can be used to
10211 avoid this algorithm.)
10215 * Add processing of proxy certificates (see RFC 3820). This work was
10216 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10217 EGEE (Enabling Grids for E-science in Europe).
10221 * RC4 performance overhaul on modern architectures/implementations, such
10222 as Intel P4, IA-64 and AMD64.
10226 * New utility extract-section.pl. This can be used specify an alternative
10227 section number in a pod file instead of having to treat each file as
10228 a separate case in Makefile. This can be done by adding two lines to the
10231 =for comment openssl_section:XXX
10233 The blank line is mandatory.
10237 * New arguments -certform, -keyform and -pass for s_client and s_server
10238 to allow alternative format key and certificate files and passphrase
10243 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10244 update associated structures and add various utility functions.
10246 Add new policy related verify parameters, include policy checking in
10247 standard verify code. Enhance 'smime' application with extra parameters
10248 to support policy checking and print out.
10252 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10253 Nehemiah processors. These extensions support AES encryption in hardware
10254 as well as RNG (though RNG support is currently disabled).
10256 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10258 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10262 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10264 *Andy Polyakov and a number of other people*
10266 * Improved PowerPC platform support. Most notably BIGNUM assembler
10267 implementation contributed by IBM.
10269 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10271 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10272 exponent rather than 'unsigned long'. There is a corresponding change to
10273 the new 'rsa_keygen' element of the RSA_METHOD structure.
10275 *Jelte Jansen, Geoff Thorpe*
10277 * Functionality for creating the initial serial number file is now
10278 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10280 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10281 number file to 1, which is bound to cause problems. To avoid
10282 the problems while respecting compatibility between different 0.9.7
10283 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10284 CA.pl for serial number initialization. With the new release 0.9.8,
10285 we can fix the problem directly in the 'ca' utility.)
10289 * Reduced header interdependencies by declaring more opaque objects in
10290 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10291 give fewer recursive includes, which could break lazy source code - so
10292 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10293 developers should define this symbol when building and using openssl to
10294 ensure they track the recommended behaviour, interfaces, [etc], but
10295 backwards-compatible behaviour prevails when this isn't defined.
10299 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10303 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10304 This will generate a random key of the appropriate length based on the
10305 cipher context. The EVP_CIPHER can provide its own random key generation
10306 routine to support keys of a specific form. This is used in the des and
10307 3des routines to generate a key of the correct parity. Update S/MIME
10308 code to use new functions and hence generate correct parity DES keys.
10309 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10310 valid (weak or incorrect parity).
10314 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10315 as looking them up. This is useful when the verified structure may contain
10316 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10317 present unless the new PKCS7_NO_CRL flag is asserted.
10321 * Extend ASN1 oid configuration module. It now additionally accepts the
10324 shortName = some long name, 1.2.3.4
10328 * Reimplemented the BN_CTX implementation. There is now no more static
10329 limitation on the number of variables it can handle nor the depth of the
10330 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10331 information can now expand as required, and rather than having a single
10332 static array of bignums, BN_CTX now uses a linked-list of such arrays
10333 allowing it to expand on demand whilst maintaining the usefulness of
10334 BN_CTX's "bundling".
10338 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10339 to allow all RSA operations to function using a single BN_CTX.
10343 * Preliminary support for certificate policy evaluation and checking. This
10344 is initially intended to pass the tests outlined in "Conformance Testing
10345 of Relying Party Client Certificate Path Processing Logic" v1.07.
10349 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10350 remained unused and not that useful. A variety of other little bignum
10351 tweaks and fixes have also been made continuing on from the audit (see
10356 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10357 associated ASN1, EVP and SSL functions and old ASN1 macros.
10361 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10362 and this should never fail. So the return value from the use of
10363 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10364 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10368 * BN_CTX_get() should return zero-valued bignums, providing the same
10369 initialised value as BN_new().
10371 *Geoff Thorpe, suggested by Ulf Möller*
10373 * Support for inhibitAnyPolicy certificate extension.
10377 * An audit of the BIGNUM code is underway, for which debugging code is
10378 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10379 is considered valid when processing BIGNUMs, and causes execution to
10380 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10381 further steps are taken to deliberately pollute unused data in BIGNUM
10382 structures to try and expose faulty code further on. For now, openssl will
10383 (in its default mode of operation) continue to tolerate the inconsistent
10384 forms that it has tolerated in the past, but authors and packagers should
10385 consider trying openssl and their own applications when compiled with
10386 these debugging symbols defined. It will help highlight potential bugs in
10387 their own code, and will improve the test coverage for OpenSSL itself. At
10388 some point, these tighter rules will become openssl's default to improve
10389 maintainability, though the assert()s and other overheads will remain only
10390 in debugging configurations. See bn.h for more details.
10392 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10394 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10395 that can only be obtained through BN_CTX_new() (which implicitly
10396 initialises it). The presence of this function only made it possible
10397 to overwrite an existing structure (and cause memory leaks).
10401 * Because of the callback-based approach for implementing LHASH as a
10402 template type, lh_insert() adds opaque objects to hash-tables and
10403 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10404 to clean up those corresponding objects before destroying the hash table
10405 (and losing the object pointers). So some over-zealous constifications in
10406 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10407 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10408 prototyped to have "const" restrictions on the object pointers they are
10409 given (and so aren't required to cast them away any more).
10413 * The tmdiff.h API was so ugly and minimal that our own timing utility
10414 (speed) prefers to use its own implementation. The two implementations
10415 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10416 its object type properly exposed (MS_TM) instead of casting to/from
10417 `char *`. This may still change yet if someone realises MS_TM and
10419 aren't necessarily the greatest nomenclatures - but this is what was used
10420 internally to the implementation so I've used that for now.
10424 * Ensure that deprecated functions do not get compiled when
10425 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10426 the self-tests were still using deprecated key-generation functions so
10427 these have been updated also.
10431 * Reorganise PKCS#7 code to separate the digest location functionality
10432 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10433 New function PKCS7_set_digest() to set the digest type for PKCS#7
10434 digestedData type. Add additional code to correctly generate the
10435 digestedData type and add support for this type in PKCS7 initialization
10440 * New function PKCS7_set0_type_other() this initializes a PKCS7
10441 structure of type "other".
10445 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10446 sure the loop does correctly stop and breaking ("division by zero")
10447 modulus operations are not performed. The (pre-generated) prime
10448 table crypto/bn/bn_prime.h was already correct, but it could not be
10449 re-generated on some platforms because of the "division by zero"
10450 situation in the script.
10452 *Ralf S. Engelschall*
10454 * Update support for ECC-based TLS ciphersuites according to
10455 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10456 SHA-1 now is only used for "small" curves (where the
10457 representation of a field element takes up to 24 bytes); for
10458 larger curves, the field element resulting from ECDH is directly
10459 used as premaster secret.
10461 *Douglas Stebila (Sun Microsystems Laboratories)*
10463 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10464 curve secp160r1 to the tests.
10466 *Douglas Stebila (Sun Microsystems Laboratories)*
10468 * Add the possibility to load symbols globally with DSO.
10470 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10472 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10473 control of the error stack.
10477 * Add support for STORE in ENGINE.
10481 * Add the STORE type. The intention is to provide a common interface
10482 to certificate and key stores, be they simple file-based stores, or
10483 HSM-type store, or LDAP stores, or...
10484 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10488 * Add a generic structure called OPENSSL_ITEM. This can be used to
10489 pass a list of arguments to any function as well as provide a way
10490 for a function to pass data back to the caller.
10494 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10495 works like BUF_strdup() but can be used to duplicate a portion of
10496 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10501 * Add the function sk_find_ex() which works like sk_find(), but will
10502 return an index to an element even if an exact match couldn't be
10503 found. The index is guaranteed to point at the element where the
10504 searched-for key would be inserted to preserve sorting order.
10508 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10509 takes an extra flags argument for optional functionality. Currently,
10510 the following flags are defined:
10512 OBJ_BSEARCH_VALUE_ON_NOMATCH
10513 This one gets OBJ_bsearch_ex() to return a pointer to the first
10514 element where the comparing function returns a negative or zero
10517 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10518 This one gets OBJ_bsearch_ex() to return a pointer to the first
10519 element where the comparing function returns zero. This is useful
10520 if there are more than one element where the comparing function
10525 * Make it possible to create self-signed certificates with 'openssl ca'
10526 in such a way that the self-signed certificate becomes part of the
10527 CA database and uses the same mechanisms for serial number generation
10528 as all other certificate signing. The new flag '-selfsign' enables
10529 this functionality. Adapt CA.sh and CA.pl.in.
10533 * Add functionality to check the public key of a certificate request
10534 against a given private. This is useful to check that a certificate
10535 request can be signed by that key (self-signing).
10539 * Make it possible to have multiple active certificates with the same
10540 subject in the CA index file. This is done only if the keyword
10541 'unique_subject' is set to 'no' in the main CA section (default
10542 if 'CA_default') of the configuration file. The value is saved
10543 with the database itself in a separate index attribute file,
10544 named like the index file with '.attr' appended to the name.
10548 * Generate multi-valued AVAs using '+' notation in config files for
10553 * Support for nameConstraints certificate extension.
10557 * Support for policyConstraints certificate extension.
10561 * Support for policyMappings certificate extension.
10565 * Make sure the default DSA_METHOD implementation only uses its
10566 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10567 and change its own handlers to be NULL so as to remove unnecessary
10568 indirection. This lets alternative implementations fallback to the
10569 default implementation more easily.
10573 * Support for directoryName in GeneralName related extensions
10578 * Make it possible to link applications using Makefile.shared.
10579 Make that possible even when linking against static libraries!
10583 * Support for single pass processing for S/MIME signing. This now
10584 means that S/MIME signing can be done from a pipe, in addition
10585 cleartext signing (multipart/signed type) is effectively streaming
10586 and the signed data does not need to be all held in memory.
10588 This is done with a new flag PKCS7_STREAM. When this flag is set
10589 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10590 is done after the data is output (and digests calculated) in
10591 SMIME_write_PKCS7().
10595 * Add full support for -rpath/-R, both in shared libraries and
10596 applications, at least on the platforms where it's known how
10601 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10602 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10603 will now compute a table of multiples of the generator that
10604 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10605 faster (notably in the case of a single point multiplication,
10606 scalar * generator).
10608 *Nils Larsch, Bodo Moeller*
10610 * IPv6 support for certificate extensions. The various extensions
10611 which use the IP:a.b.c.d can now take IPv6 addresses using the
10612 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10617 * Added an ENGINE that implements RSA by performing private key
10618 exponentiations with the GMP library. The conversions to and from
10619 GMP's mpz_t format aren't optimised nor are any montgomery forms
10620 cached, and on x86 it appears OpenSSL's own performance has caught up.
10621 However there are likely to be other architectures where GMP could
10622 provide a boost. This ENGINE is not built in by default, but it can be
10623 specified at Configure time and should be accompanied by the necessary
10624 linker additions, eg;
10625 ./config -DOPENSSL_USE_GMP -lgmp
10629 * "openssl engine" will not display ENGINE/DSO load failure errors when
10630 testing availability of engines with "-t" - the old behaviour is
10631 produced by increasing the feature's verbosity with "-tt".
10635 * ECDSA routines: under certain error conditions uninitialized BN objects
10636 could be freed. Solution: make sure initialization is performed early
10637 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10642 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10643 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10644 software implementations. For DSA and DH, parameter generation can
10645 also be overridden by providing the appropriate method callbacks.
10649 * Change the "progress" mechanism used in key-generation and
10650 primality testing to functions that take a new BN_GENCB pointer in
10651 place of callback/argument pairs. The new API functions have `_ex`
10652 postfixes and the older functions are reimplemented as wrappers for
10653 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10654 declarations of the old functions to help (graceful) attempts to
10655 migrate to the new functions. Also, the new key-generation API
10656 functions operate on a caller-supplied key-structure and return
10657 success/failure rather than returning a key or NULL - this is to
10658 help make "keygen" another member function of RSA_METHOD etc.
10660 Example for using the new callback interface:
10662 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10663 void *my_arg = ...;
10666 BN_GENCB_set(&my_cb, my_callback, my_arg);
10668 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10669 /* For the meaning of a, b in calls to my_callback(), see the
10670 * documentation of the function that calls the callback.
10671 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10672 * my_callback should return 1 if it wants BN_is_prime_ex()
10673 * to continue, or 0 to stop.
10678 * Change the ZLIB compression method to be stateful, and make it
10679 available to TLS with the number defined in
10680 draft-ietf-tls-compression-04.txt.
10684 * Add the ASN.1 structures and functions for CertificatePair, which
10685 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10687 CertificatePair ::= SEQUENCE {
10688 forward [0] Certificate OPTIONAL,
10689 reverse [1] Certificate OPTIONAL,
10690 -- at least one of the pair shall be present -- }
10692 Also implement the PEM functions to read and write certificate
10693 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10695 This needed to be defined, mostly for the sake of the LDAP
10696 attribute crossCertificatePair, but may prove useful elsewhere as
10701 * Make it possible to inhibit symlinking of shared libraries in
10702 Makefile.shared, for Cygwin's sake.
10706 * Extend the BIGNUM API by creating a function
10707 void BN_set_negative(BIGNUM *a, int neg);
10708 and a macro that behave like
10709 int BN_is_negative(const BIGNUM *a);
10711 to avoid the need to access 'a->neg' directly in applications.
10715 * Implement fast modular reduction for pseudo-Mersenne primes
10716 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10717 EC_GROUP_new_curve_GFp() will now automatically use this
10720 *Nils Larsch <nla@trustcenter.de>*
10722 * Add new lock type (CRYPTO_LOCK_BN).
10726 * Change the ENGINE framework to automatically load engines
10727 dynamically from specific directories unless they could be
10728 found to already be built in or loaded. Move all the
10729 current engines except for the cryptodev one to a new
10730 directory engines/.
10731 The engines in engines/ are built as shared libraries if
10732 the "shared" options was given to ./Configure or ./config.
10733 Otherwise, they are inserted in libcrypto.a.
10734 /usr/local/ssl/engines is the default directory for dynamic
10735 engines, but that can be overridden at configure time through
10736 the usual use of --prefix and/or --openssldir, and at run
10737 time with the environment variable OPENSSL_ENGINES.
10739 *Geoff Thorpe and Richard Levitte*
10741 * Add Makefile.shared, a helper makefile to build shared
10742 libraries. Adapt Makefile.org.
10746 * Add version info to Win32 DLLs.
10748 *Peter 'Luna' Runestig" <peter@runestig.com>*
10750 * Add new 'medium level' PKCS#12 API. Certificates and keys
10751 can be added using this API to created arbitrary PKCS#12
10752 files while avoiding the low level API.
10754 New options to PKCS12_create(), key or cert can be NULL and
10755 will then be omitted from the output file. The encryption
10756 algorithm NIDs can be set to -1 for no encryption, the mac
10757 iteration count can be set to 0 to omit the mac.
10759 Enhance pkcs12 utility by making the -nokeys and -nocerts
10760 options work when creating a PKCS#12 file. New option -nomac
10761 to omit the mac, NONE can be set for an encryption algorithm.
10762 New code is modified to use the enhanced PKCS12_create()
10763 instead of the low level API.
10767 * Extend ASN1 encoder to support indefinite length constructed
10768 encoding. This can output sequences tags and octet strings in
10769 this form. Modify pk7_asn1.c to support indefinite length
10770 encoding. This is experimental and needs additional code to
10771 be useful, such as an ASN1 bio and some enhanced streaming
10774 Extend template encode functionality so that tagging is passed
10775 down to the template encoder.
10779 * Let 'openssl req' fail if an argument to '-newkey' is not
10780 recognized instead of using RSA as a default.
10784 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10785 As these are not official, they are not included in "ALL";
10786 the "ECCdraft" ciphersuite group alias can be used to select them.
10788 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10790 * Add ECDH engine support.
10792 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10794 * Add ECDH in new directory crypto/ecdh/.
10796 *Douglas Stebila (Sun Microsystems Laboratories)*
10798 * Let BN_rand_range() abort with an error after 100 iterations
10799 without success (which indicates a broken PRNG).
10803 * Change BN_mod_sqrt() so that it verifies that the input value
10804 is really the square of the return value. (Previously,
10805 BN_mod_sqrt would show GIGO behaviour.)
10809 * Add named elliptic curves over binary fields from X9.62, SECG,
10810 and WAP/WTLS; add OIDs that were still missing.
10812 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10814 * Extend the EC library for elliptic curves over binary fields
10815 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10818 EC_GF2m_simple_method
10822 EC_GROUP_new_curve_GF2m
10823 EC_GROUP_set_curve_GF2m
10824 EC_GROUP_get_curve_GF2m
10825 EC_POINT_set_affine_coordinates_GF2m
10826 EC_POINT_get_affine_coordinates_GF2m
10827 EC_POINT_set_compressed_coordinates_GF2m
10829 Point compression for binary fields is disabled by default for
10830 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10833 As binary polynomials are represented as BIGNUMs, various members
10834 of the EC_GROUP and EC_POINT data structures can be shared
10835 between the implementations for prime fields and binary fields;
10836 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10837 are essentially identical to their `..._GFp` counterparts.
10838 (For simplicity, the `..._GFp` prefix has been dropped from
10839 various internal method names.)
10841 An internal 'field_div' method (similar to 'field_mul' and
10842 'field_sqr') has been added; this is used only for binary fields.
10844 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10846 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10847 through methods ('mul', 'precompute_mult').
10849 The generic implementations (now internally called 'ec_wNAF_mul'
10850 and 'ec_wNAF_precomputed_mult') remain the default if these
10851 methods are undefined.
10853 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10855 * New function EC_GROUP_get_degree, which is defined through
10856 EC_METHOD. For curves over prime fields, this returns the bit
10857 length of the modulus.
10859 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10861 * New functions EC_GROUP_dup, EC_POINT_dup.
10862 (These simply call ..._new and ..._copy).
10864 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10866 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10867 Polynomials are represented as BIGNUMs (where the sign bit is not
10868 used) in the following functions [macros]:
10871 BN_GF2m_sub [= BN_GF2m_add]
10872 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10873 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10874 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10876 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10877 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10878 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10879 BN_GF2m_cmp [= BN_ucmp]
10881 (Note that only the 'mod' functions are actually for fields GF(2^m).
10882 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10884 For some functions, an the irreducible polynomial defining a
10885 field can be given as an 'unsigned int[]' with strictly
10886 decreasing elements giving the indices of those bits that are set;
10887 i.e., p[] represents the polynomial
10888 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10890 p[0] > p[1] > ... > p[k] = 0.
10891 This applies to the following functions:
10894 BN_GF2m_mod_mul_arr
10895 BN_GF2m_mod_sqr_arr
10896 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10897 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10898 BN_GF2m_mod_exp_arr
10899 BN_GF2m_mod_sqrt_arr
10900 BN_GF2m_mod_solve_quad_arr
10904 Conversion can be performed by the following functions:
10909 bntest.c has additional tests for binary polynomial arithmetic.
10911 Two implementations for BN_GF2m_mod_div() are available.
10912 The default algorithm simply uses BN_GF2m_mod_inv() and
10913 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10914 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10915 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10917 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10919 * Add new error code 'ERR_R_DISABLED' that can be used when some
10920 functionality is disabled at compile-time.
10922 *Douglas Stebila <douglas.stebila@sun.com>*
10924 * Change default behaviour of 'openssl asn1parse' so that more
10925 information is visible when viewing, e.g., a certificate:
10927 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10928 mode the content of non-printable OCTET STRINGs is output in a
10929 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10930 avoid the appearance of a printable string.
10932 *Nils Larsch <nla@trustcenter.de>*
10934 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10936 EC_GROUP_set_asn1_flag()
10937 EC_GROUP_get_asn1_flag()
10938 EC_GROUP_set_point_conversion_form()
10939 EC_GROUP_get_point_conversion_form()
10940 These control ASN1 encoding details:
10941 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10942 has been set to OPENSSL_EC_NAMED_CURVE.
10943 - Points are encoded in uncompressed form by default; options for
10944 asn1_for are as for point2oct, namely
10945 POINT_CONVERSION_COMPRESSED
10946 POINT_CONVERSION_UNCOMPRESSED
10947 POINT_CONVERSION_HYBRID
10949 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10951 EC_GROUP_set_seed()
10952 EC_GROUP_get0_seed()
10953 EC_GROUP_get_seed_len()
10954 This is used only for ASN1 purposes (so far).
10956 *Nils Larsch <nla@trustcenter.de>*
10958 * Add 'field_type' member to EC_METHOD, which holds the NID
10959 of the appropriate field type OID. The new function
10960 EC_METHOD_get_field_type() returns this value.
10962 *Nils Larsch <nla@trustcenter.de>*
10965 EC_POINT_point2bn()
10966 EC_POINT_bn2point()
10967 EC_POINT_point2hex()
10968 EC_POINT_hex2point()
10969 providing useful interfaces to EC_POINT_point2oct() and
10970 EC_POINT_oct2point().
10972 *Nils Larsch <nla@trustcenter.de>*
10974 * Change internals of the EC library so that the functions
10975 EC_GROUP_set_generator()
10976 EC_GROUP_get_generator()
10977 EC_GROUP_get_order()
10978 EC_GROUP_get_cofactor()
10979 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10980 to methods, which would lead to unnecessary code duplication when
10981 adding different types of curves.
10983 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10985 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10986 arithmetic, and such that modified wNAFs are generated
10987 (which avoid length expansion in many cases).
10991 * Add a function EC_GROUP_check_discriminant() (defined via
10992 EC_METHOD) that verifies that the curve discriminant is non-zero.
10994 Add a function EC_GROUP_check() that makes some sanity tests
10995 on a EC_GROUP, its generator and order. This includes
10996 EC_GROUP_check_discriminant().
10998 *Nils Larsch <nla@trustcenter.de>*
11000 * Add ECDSA in new directory crypto/ecdsa/.
11002 Add applications 'openssl ecparam' and 'openssl ecdsa'
11003 (these are based on 'openssl dsaparam' and 'openssl dsa').
11005 ECDSA support is also included in various other files across the
11006 library. Most notably,
11007 - 'openssl req' now has a '-newkey ecdsa:file' option;
11008 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11009 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11010 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11011 them suitable for ECDSA where domain parameters must be
11012 extracted before the specific public key;
11013 - ECDSA engine support has been added.
11015 *Nils Larsch <nla@trustcenter.de>*
11017 * Include some named elliptic curves, and add OIDs from X9.62,
11018 SECG, and WAP/WTLS. Each curve can be obtained from the new
11020 EC_GROUP_new_by_curve_name(),
11021 and the list of available named curves can be obtained with
11022 EC_get_builtin_curves().
11023 Also add a 'curve_name' member to EC_GROUP objects, which can be
11025 EC_GROUP_set_curve_name()
11026 EC_GROUP_get_curve_name()
11028 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11030 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11031 was actually never needed) and in BN_mul(). The removal in BN_mul()
11032 required a small change in bn_mul_part_recursive() and the addition
11033 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11034 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11035 bn_sub_words() and bn_add_words() except they take arrays with
11040 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11042 * Cleanse PEM buffers before freeing them since they may contain
11045 *Benjamin Bennett <ben@psc.edu>*
11047 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11048 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11049 authentication-only ciphersuites.
11053 * Since AES128 and AES256 share a single mask bit in the logic of
11054 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11055 kludge to work properly if AES128 is available and AES256 isn't.
11059 * Expand security boundary to match 1.1.1 module.
11063 * Remove redundant features: hash file source, editing of test vectors
11064 modify fipsld to use external fips_premain.c signature.
11068 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11069 run algorithm test programs.
11073 * Make algorithm test programs more tolerant of whitespace.
11077 * Have SSL/TLS server implementation tolerate "mismatched" record
11078 protocol version while receiving ClientHello even if the
11079 ClientHello is fragmented. (The server can't insist on the
11080 particular protocol version it has chosen before the ServerHello
11081 message has informed the client about his choice.)
11085 * Load error codes if they are not already present instead of using a
11086 static variable. This allows them to be cleanly unloaded and reloaded.
11090 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11092 * Introduce limits to prevent malicious keys being able to
11093 cause a denial of service. ([CVE-2006-2940])
11095 *Steve Henson, Bodo Moeller*
11097 * Fix ASN.1 parsing of certain invalid structures that can result
11098 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11100 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11101 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11103 * Fix SSL client code which could crash if connecting to a
11104 malicious SSLv2 server. ([CVE-2006-4343])
11106 *Tavis Ormandy and Will Drewry, Google Security Team*
11108 * Change ciphersuite string processing so that an explicit
11109 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11110 will no longer include "AES128-SHA"), and any other similar
11111 ciphersuite (same bitmap) from *other* protocol versions (so that
11112 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11113 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11114 changes from 0.9.8b and 0.9.8d.
11118 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11120 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11121 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11123 * Change the Unix randomness entropy gathering to use poll() when
11124 possible instead of select(), since the latter has some
11125 undesirable limitations.
11127 *Darryl Miles via Richard Levitte and Bodo Moeller*
11129 * Disable rogue ciphersuites:
11131 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11132 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11133 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11135 The latter two were purportedly from
11136 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11139 Also deactivate the remaining ciphersuites from
11140 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11141 unofficial, and the ID has long expired.
11145 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11146 dual-core machines) and other potential thread-safety issues.
11150 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11152 * Adapt fipsld and the build system to link against the validated FIPS
11153 module in FIPS mode.
11157 * Fixes for VC++ 2005 build under Windows.
11161 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11162 from a Windows bash shell such as MSYS. It is autodetected from the
11163 "config" script when run from a VC++ environment. Modify standard VC++
11164 build to use fipscanister.o from the GNU make build.
11168 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11170 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11171 The value now differs depending on if you build for FIPS or not.
11172 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11173 safely run with a non-FIPSed libcrypto, as it may crash because of
11174 the difference induced by this change.
11178 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11180 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11181 (part of SSL_OP_ALL). This option used to disable the
11182 countermeasure against man-in-the-middle protocol-version
11183 rollback in the SSL 2.0 server implementation, which is a bad
11184 idea. ([CVE-2005-2969])
11186 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11187 for Information Security, National Institute of Advanced Industrial
11188 Science and Technology [AIST, Japan)]*
11190 * Minimal support for X9.31 signatures and PSS padding modes. This is
11191 mainly for FIPS compliance and not fully integrated at this stage.
11195 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11196 the exponentiation using a fixed-length exponent. (Otherwise,
11197 the information leaked through timing could expose the secret key
11198 after many signatures; cf. Bleichenbacher's attack on DSA with
11203 * Make a new fixed-window mod_exp implementation the default for
11204 RSA, DSA, and DH private-key operations so that the sequence of
11205 squares and multiplies and the memory access pattern are
11206 independent of the particular secret key. This will mitigate
11207 cache-timing and potential related attacks.
11209 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11210 and this is automatically used by BN_mod_exp_mont() if the new flag
11211 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11212 will use this BN flag for private exponents unless the flag
11213 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11214 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11216 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11218 * Change the client implementation for SSLv23_method() and
11219 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11220 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11221 (Previously, the SSL 2.0 backwards compatible Client Hello
11222 message format would be used even with SSL_OP_NO_SSLv2.)
11226 * Add support for smime-type MIME parameter in S/MIME messages which some
11231 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11232 a threadsafe manner. Modify rsa code to use new function and add calls
11233 to dsa and dh code (which had race conditions before).
11237 * Include the fixed error library code in the C error file definitions
11238 instead of fixing them up at runtime. This keeps the error code
11239 structures constant.
11243 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11245 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11248 * Fixes for newer kerberos headers. NB: the casts are needed because
11249 the 'length' field is signed on one version and unsigned on another
11250 with no (?) obvious way to tell the difference, without these VC++
11251 complains. Also the "definition" of FAR (blank) is no longer included
11252 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11253 some needed definitions.
11257 * Undo Cygwin change.
11261 * Added support for proxy certificates according to RFC 3820.
11262 Because they may be a security thread to unaware applications,
11263 they must be explicitly allowed in run-time. See
11264 docs/HOWTO/proxy_certificates.txt for further information.
11268 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11270 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11271 server and client random values. Previously
11272 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11273 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11275 This change has negligible security impact because:
11277 1. Server and client random values still have 24 bytes of pseudo random
11280 2. Server and client random values are sent in the clear in the initial
11283 3. The master secret is derived using the premaster secret (48 bytes in
11284 size for static RSA ciphersuites) as well as client server and random
11287 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11290 *Stephen Henson, reported by UK NISCC*
11292 * Use Windows randomness collection on Cygwin.
11296 * Fix hang in EGD/PRNGD query when communication socket is closed
11297 prematurely by EGD/PRNGD.
11299 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11301 * Prompt for pass phrases when appropriate for PKCS12 input format.
11305 * Back-port of selected performance improvements from development
11306 branch, as well as improved support for PowerPC platforms.
11310 * Add lots of checks for memory allocation failure, error codes to indicate
11311 failure and freeing up memory if a failure occurs.
11313 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11315 * Add new -passin argument to dgst.
11319 * Perform some character comparisons of different types in X509_NAME_cmp:
11320 this is needed for some certificates that re-encode DNs into UTF8Strings
11321 (in violation of RFC3280) and can't or won't issue name rollover
11326 * Make an explicit check during certificate validation to see that
11327 the CA setting in each certificate on the chain is correct. As a
11328 side effect always do the following basic checks on extensions,
11329 not just when there's an associated purpose to the check:
11331 - if there is an unhandled critical extension (unless the user
11332 has chosen to ignore this fault)
11333 - if the path length has been exceeded (if one is set at all)
11334 - that certain extensions fit the associated purpose (if one has
11339 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11341 * Avoid a race condition when CRLs are checked in a multi threaded
11342 environment. This would happen due to the reordering of the revoked
11343 entries during signature checking and serial number lookup. Now the
11344 encoding is cached and the serial number sort performed under a lock.
11345 Add new STACK function sk_is_sorted().
11349 * Add Delta CRL to the extension code.
11353 * Various fixes to s3_pkt.c so alerts are sent properly.
11355 *David Holmes <d.holmes@f5.com>*
11357 * Reduce the chances of duplicate issuer name and serial numbers (in
11358 violation of RFC3280) using the OpenSSL certificate creation utilities.
11359 This is done by creating a random 64 bit value for the initial serial
11360 number when a serial number file is created or when a self signed
11361 certificate is created using 'openssl req -x509'. The initial serial
11362 number file is created using 'openssl x509 -next_serial' in CA.pl
11363 rather than being initialized to 1.
11367 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11369 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11370 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11372 *Joe Orton, Steve Henson*
11374 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11377 *Joe Orton, Steve Henson*
11379 * Make it possible to have multiple active certificates with the same
11380 subject in the CA index file. This is done only if the keyword
11381 'unique_subject' is set to 'no' in the main CA section (default
11382 if 'CA_default') of the configuration file. The value is saved
11383 with the database itself in a separate index attribute file,
11384 named like the index file with '.attr' appended to the name.
11388 * X509 verify fixes. Disable broken certificate workarounds when
11389 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11390 keyUsage extension present. Don't accept CRLs with unhandled critical
11391 extensions: since verify currently doesn't process CRL extensions this
11392 rejects a CRL with *any* critical extensions. Add new verify error codes
11397 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11398 A clarification of RFC2560 will require the use of OCTET STRINGs and
11399 some implementations cannot handle the current raw format. Since OpenSSL
11400 copies and compares OCSP nonces as opaque blobs without any attempt at
11401 parsing them this should not create any compatibility issues.
11405 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11406 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11407 this HMAC (and other) operations are several times slower than OpenSSL
11412 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11414 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11416 * Use the correct content when signing type "other".
11420 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11422 * Fix various bugs revealed by running the NISCC test suite:
11424 Stop out of bounds reads in the ASN1 code when presented with
11425 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11427 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11429 If verify callback ignores invalid public key errors don't try to check
11430 certificate signature with the NULL public key.
11434 * New -ignore_err option in ocsp application to stop the server
11435 exiting on the first error in a request.
11439 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11440 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11445 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11446 extra data after the compression methods not only for TLS 1.0
11447 but also for SSL 3.0 (as required by the specification).
11449 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11451 * Change X509_certificate_type() to mark the key as exported/exportable
11452 when it's 512 *bits* long, not 512 bytes.
11456 * Change AES_cbc_encrypt() so it outputs exact multiple of
11457 blocks during encryption.
11461 * Various fixes to base64 BIO and non blocking I/O. On write
11462 flushes were not handled properly if the BIO retried. On read
11463 data was not being buffered properly and had various logic bugs.
11464 This also affects blocking I/O when the data being decoded is a
11469 * Various S/MIME bugfixes and compatibility changes:
11470 output correct application/pkcs7 MIME type if
11471 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11472 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11473 of files as .eml work). Correctly handle very long lines in MIME
11478 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11480 * Countermeasure against the Klima-Pokorny-Rosa extension of
11481 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11482 a protocol version number mismatch like a decryption error
11483 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11487 * Turn on RSA blinding by default in the default implementation
11488 to avoid a timing attack. Applications that don't want it can call
11489 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11490 They would be ill-advised to do so in most cases.
11492 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11494 * Change RSA blinding code so that it works when the PRNG is not
11495 seeded (in this case, the secret RSA exponent is abused as
11496 an unpredictable seed -- if it is not unpredictable, there
11497 is no point in blinding anyway). Make RSA blinding thread-safe
11498 by remembering the creator's thread ID in rsa->blinding and
11499 having all other threads use local one-time blinding factors
11500 (this requires more computation than sharing rsa->blinding, but
11501 avoids excessive locking; and if an RSA object is not shared
11502 between threads, blinding will still be very fast).
11506 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11507 ENGINE as defaults for all supported algorithms irrespective of
11508 the 'flags' parameter. 'flags' is now honoured, so applications
11509 should make sure they are passing it correctly.
11513 * Target "mingw" now allows native Windows code to be generated in
11514 the Cygwin environment as well as with the MinGW compiler.
11518 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11520 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11521 via timing by performing a MAC computation even if incorrect
11522 block cipher padding has been found. This is a countermeasure
11523 against active attacks where the attacker has to distinguish
11524 between bad padding and a MAC verification error. ([CVE-2003-0078])
11526 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11527 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11528 Martin Vuagnoux (EPFL, Ilion)*
11530 * Make the no-err option work as intended. The intention with no-err
11531 is not to have the whole error stack handling routines removed from
11532 libcrypto, it's only intended to remove all the function name and
11533 reason texts, thereby removing some of the footprint that may not
11534 be interesting if those errors aren't displayed anyway.
11536 NOTE: it's still possible for any application or module to have its
11537 own set of error texts inserted. The routines are there, just not
11538 used by default when no-err is given.
11542 * Add support for FreeBSD on IA64.
11544 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11546 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11547 Kerberos function mit_des_cbc_cksum(). Before this change,
11548 the value returned by DES_cbc_cksum() was like the one from
11549 mit_des_cbc_cksum(), except the bytes were swapped.
11551 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11553 * Allow an application to disable the automatic SSL chain building.
11554 Before this a rather primitive chain build was always performed in
11555 ssl3_output_cert_chain(): an application had no way to send the
11556 correct chain if the automatic operation produced an incorrect result.
11558 Now the chain builder is disabled if either:
11560 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11562 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11564 The reasoning behind this is that an application would not want the
11565 auto chain building to take place if extra chain certificates are
11566 present and it might also want a means of sending no additional
11567 certificates (for example the chain has two certificates and the
11572 * Add the possibility to build without the ENGINE framework.
11574 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11576 * Under Win32 gmtime() can return NULL: check return value in
11577 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11581 * DSA routines: under certain error conditions uninitialized BN objects
11582 could be freed. Solution: make sure initialization is performed early
11583 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11584 Nils Larsch <nla@trustcenter.de> via PR#459)
11588 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11589 checked on reconnect on the client side, therefore session resumption
11590 could still fail with a "ssl session id is different" error. This
11591 behaviour is masked when SSL_OP_ALL is used due to
11592 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11593 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11594 followup to PR #377.
11598 * IA-32 assembler support enhancements: unified ELF targets, support
11599 for SCO/Caldera platforms, fix for Cygwin shared build.
11603 * Add support for FreeBSD on sparc64. As a consequence, support for
11604 FreeBSD on non-x86 processors is separate from x86 processors on
11605 the config script, much like the NetBSD support.
11607 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11609 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11611 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11614 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11615 code (06) was taken as the first octet of the session ID and the last
11616 octet was ignored consequently. As a result SSLv2 client side session
11617 caching could not have worked due to the session ID mismatch between
11619 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11624 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11625 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11630 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11631 seems that in spite of existing for more than a year, many application
11632 author have done nothing to provide the necessary callbacks, which
11633 means that this particular engine will not work properly anywhere.
11634 This is a very unfortunate situation which forces us, in the name
11635 of usability, to give the hw_ncipher.c a static lock, which is part
11637 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11638 appear in 0.9.8 or later. We EXPECT application authors to have
11639 dealt properly with this when 0.9.8 is released (unless we actually
11640 make such changes in the libcrypto locking code that changes will
11641 have to be made anyway).
11645 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11646 octets have been read, EOF or an error occurs. Without this change
11647 some truncated ASN1 structures will not produce an error.
11651 * Disable Heimdal support, since it hasn't been fully implemented.
11652 Still give the possibility to force the use of Heimdal, but with
11653 warnings and a request that patches get sent to openssl-dev.
11657 * Add the VC-CE target, introduce the WINCE sysname, and add
11658 INSTALL.WCE and appropriate conditionals to make it build.
11660 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11662 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11663 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11664 edit numbers of the version.
11666 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11668 * Introduce safe string copy and catenation functions
11669 (BUF_strlcpy() and BUF_strlcat()).
11671 *Ben Laurie (CHATS) and Richard Levitte*
11673 * Avoid using fixed-size buffers for one-line DNs.
11675 *Ben Laurie (CHATS)*
11677 * Add BUF_MEM_grow_clean() to avoid information leakage when
11678 resizing buffers containing secrets, and use where appropriate.
11680 *Ben Laurie (CHATS)*
11682 * Avoid using fixed size buffers for configuration file location.
11684 *Ben Laurie (CHATS)*
11686 * Avoid filename truncation for various CA files.
11688 *Ben Laurie (CHATS)*
11690 * Use sizeof in preference to magic numbers.
11692 *Ben Laurie (CHATS)*
11694 * Avoid filename truncation in cert requests.
11696 *Ben Laurie (CHATS)*
11698 * Add assertions to check for (supposedly impossible) buffer
11701 *Ben Laurie (CHATS)*
11703 * Don't cache truncated DNS entries in the local cache (this could
11704 potentially lead to a spoofing attack).
11706 *Ben Laurie (CHATS)*
11708 * Fix various buffers to be large enough for hex/decimal
11709 representations in a platform independent manner.
11711 *Ben Laurie (CHATS)*
11713 * Add CRYPTO_realloc_clean() to avoid information leakage when
11714 resizing buffers containing secrets, and use where appropriate.
11716 *Ben Laurie (CHATS)*
11718 * Add BIO_indent() to avoid much slightly worrying code to do
11721 *Ben Laurie (CHATS)*
11723 * Convert sprintf()/BIO_puts() to BIO_printf().
11725 *Ben Laurie (CHATS)*
11727 * buffer_gets() could terminate with the buffer only half
11730 *Ben Laurie (CHATS)*
11732 * Add assertions to prevent user-supplied crypto functions from
11733 overflowing internal buffers by having large block sizes, etc.
11735 *Ben Laurie (CHATS)*
11737 * New OPENSSL_assert() macro (similar to assert(), but enabled
11740 *Ben Laurie (CHATS)*
11742 * Eliminate unused copy of key in RC4.
11744 *Ben Laurie (CHATS)*
11746 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11748 *Ben Laurie (CHATS)*
11750 * Fix off-by-one error in EGD path.
11752 *Ben Laurie (CHATS)*
11754 * If RANDFILE path is too long, ignore instead of truncating.
11756 *Ben Laurie (CHATS)*
11758 * Eliminate unused and incorrectly sized X.509 structure
11761 *Ben Laurie (CHATS)*
11763 * Eliminate unused and dangerous function knumber().
11765 *Ben Laurie (CHATS)*
11767 * Eliminate unused and dangerous structure, KSSL_ERR.
11769 *Ben Laurie (CHATS)*
11771 * Protect against overlong session ID context length in an encoded
11772 session object. Since these are local, this does not appear to be
11775 *Ben Laurie (CHATS)*
11777 * Change from security patch (see 0.9.6e below) that did not affect
11778 the 0.9.6 release series:
11780 Remote buffer overflow in SSL3 protocol - an attacker could
11781 supply an oversized master key in Kerberos-enabled versions.
11784 *Ben Laurie (CHATS)*
11786 * Change the SSL kerb5 codes to match RFC 2712.
11790 * Make -nameopt work fully for req and add -reqopt switch.
11792 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11794 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11796 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11798 * Make sure tests can be performed even if the corresponding algorithms
11799 have been removed entirely. This was also the last step to make
11800 OpenSSL compilable with DJGPP under all reasonable conditions.
11802 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11804 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11805 to allow version independent disabling of normally unselected ciphers,
11806 which may be activated as a side-effect of selecting a single cipher.
11808 (E.g., cipher list string "RSA" enables ciphersuites that are left
11809 out of "ALL" because they do not provide symmetric encryption.
11810 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11812 *Lutz Jaenicke, Bodo Moeller*
11814 * Add appropriate support for separate platform-dependent build
11815 directories. The recommended way to make a platform-dependent
11816 build directory is the following (tested on Linux), maybe with
11819 # Place yourself outside of the OpenSSL source tree. In
11820 # this example, the environment variable OPENSSL_SOURCE
11821 # is assumed to contain the absolute OpenSSL source directory.
11822 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11823 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11824 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11825 mkdir -p `dirname $F`
11826 ln -s $OPENSSL_SOURCE/$F $F
11829 To be absolutely sure not to disturb the source tree, a "make clean"
11830 is a good thing. If it isn't successful, don't worry about it,
11831 it probably means the source directory is very clean.
11835 * Make sure any ENGINE control commands make local copies of string
11836 pointers passed to them whenever necessary. Otherwise it is possible
11837 the caller may have overwritten (or deallocated) the original string
11838 data when a later ENGINE operation tries to use the stored values.
11840 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11842 * Improve diagnostics in file reading and command-line digests.
11844 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11846 * Add AES modes CFB and OFB to the object database. Correct an
11847 error in AES-CFB decryption.
11851 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11852 allows existing EVP_CIPHER_CTX structures to be reused after
11853 calling `EVP_*Final()`. This behaviour is used by encryption
11854 BIOs and some applications. This has the side effect that
11855 applications must explicitly clean up cipher contexts with
11856 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11860 * Check the values of dna and dnb in bn_mul_recursive before calling
11861 bn_mul_comba (a non zero value means the a or b arrays do not contain
11862 n2 elements) and fallback to bn_mul_normal if either is not zero.
11866 * Fix escaping of non-ASCII characters when using the -subj option
11867 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11871 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11872 form for "surname", serialNumber has no short form.
11873 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11874 therefore remove "mail" short name for "internet 7".
11875 The OID for unique identifiers in X509 certificates is
11876 x500UniqueIdentifier, not uniqueIdentifier.
11877 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11881 * Add an "init" command to the ENGINE config module and auto initialize
11882 ENGINEs. Without any "init" command the ENGINE will be initialized
11883 after all ctrl commands have been executed on it. If init=1 the
11884 ENGINE is initialized at that point (ctrls before that point are run
11885 on the uninitialized ENGINE and after on the initialized one). If
11886 init=0 then the ENGINE will not be initialized at all.
11890 * Fix the 'app_verify_callback' interface so that the user-defined
11891 argument is actually passed to the callback: In the
11892 SSL_CTX_set_cert_verify_callback() prototype, the callback
11893 declaration has been changed from
11896 int (*cb)(X509_STORE_CTX *,void *);
11897 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11898 i=s->ctx->app_verify_callback(&ctx)
11899 has been changed into
11900 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11902 To update applications using SSL_CTX_set_cert_verify_callback(),
11903 a dummy argument can be added to their callback functions.
11905 *D. K. Smetters <smetters@parc.xerox.com>*
11907 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11909 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11911 * Add and OPENSSL_LOAD_CONF define which will cause
11912 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11913 This allows older applications to transparently support certain
11914 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11915 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11916 load the config file and OPENSSL_add_all_algorithms_conf() which will
11917 always load it have also been added.
11921 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11922 Adjust NIDs and EVP layer.
11924 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11926 * Config modules support in openssl utility.
11928 Most commands now load modules from the config file,
11929 though in a few (such as version) this isn't done
11930 because it couldn't be used for anything.
11932 In the case of ca and req the config file used is
11933 the same as the utility itself: that is the -config
11934 command line option can be used to specify an
11939 * Move default behaviour from OPENSSL_config(). If appname is NULL
11940 use "openssl_conf" if filename is NULL use default openssl config file.
11944 * Add an argument to OPENSSL_config() to allow the use of an alternative
11945 config section name. Add a new flag to tolerate a missing config file
11946 and move code to CONF_modules_load_file().
11950 * Support for crypto accelerator cards from Accelerated Encryption
11951 Processing, www.aep.ie. (Use engine 'aep')
11952 The support was copied from 0.9.6c [engine] and adapted/corrected
11953 to work with the new engine framework.
11955 *AEP Inc. and Richard Levitte*
11957 * Support for SureWare crypto accelerator cards from Baltimore
11958 Technologies. (Use engine 'sureware')
11959 The support was copied from 0.9.6c [engine] and adapted
11960 to work with the new engine framework.
11964 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11965 make the newer ENGINE framework commands for the CHIL engine work.
11967 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11969 * Make it possible to produce shared libraries on ReliantUNIX.
11971 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11973 * Add the configuration target debug-linux-ppro.
11974 Make 'openssl rsa' use the general key loading routines
11975 implemented in `apps.c`, and make those routines able to
11976 handle the key format FORMAT_NETSCAPE and the variant
11979 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11981 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11983 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11985 * Add -keyform to rsautl, and document -engine.
11987 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11989 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11990 BIO_R_NO_SUCH_FILE error code rather than the generic
11991 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11995 * Add new functions
11996 ERR_peek_last_error
11997 ERR_peek_last_error_line
11998 ERR_peek_last_error_line_data.
11999 These are similar to
12001 ERR_peek_error_line
12002 ERR_peek_error_line_data,
12003 but report on the latest error recorded rather than the first one
12004 still in the error queue.
12006 *Ben Laurie, Bodo Moeller*
12008 * default_algorithms option in ENGINE config module. This allows things
12010 default_algorithms = ALL
12011 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12015 * Preliminary ENGINE config module.
12019 * New experimental application configuration code.
12023 * Change the AES code to follow the same name structure as all other
12024 symmetric ciphers, and behave the same way. Move everything to
12025 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12027 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12029 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12031 *Ben Laurie and Theo de Raadt*
12033 * Add option to output public keys in req command.
12035 *Massimiliano Pala madwolf@openca.org*
12037 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12038 (up to about 10% better than before for P-192 and P-224).
12042 * New functions/macros
12044 SSL_CTX_set_msg_callback(ctx, cb)
12045 SSL_CTX_set_msg_callback_arg(ctx, arg)
12046 SSL_set_msg_callback(ssl, cb)
12047 SSL_set_msg_callback_arg(ssl, arg)
12049 to request calling a callback function
12051 void cb(int write_p, int version, int content_type,
12052 const void *buf, size_t len, SSL *ssl, void *arg)
12054 whenever a protocol message has been completely received
12055 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12056 protocol version according to which the SSL library interprets
12057 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12058 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12059 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12060 specification (change_cipher_spec(20), alert(21), handshake(22)).
12061 'buf' and 'len' point to the actual message, 'ssl' to the
12062 SSL object, and 'arg' is the application-defined value set by
12063 SSL[_CTX]_set_msg_callback_arg().
12065 'openssl s_client' and 'openssl s_server' have new '-msg' options
12066 to enable a callback that displays all protocol messages.
12070 * Change the shared library support so shared libraries are built as
12071 soon as the corresponding static library is finished, and thereby get
12072 openssl and the test programs linked against the shared library.
12073 This still only happens when the keyword "shard" has been given to
12074 the configuration scripts.
12076 NOTE: shared library support is still an experimental thing, and
12077 backward binary compatibility is still not guaranteed.
12079 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12081 * Add support for Subject Information Access extension.
12083 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12085 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12086 additional bytes when new memory had to be allocated, not just
12087 when reusing an existing buffer.
12091 * New command line and configuration option 'utf8' for the req command.
12092 This allows field values to be specified as UTF8 strings.
12096 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12097 runs for the former and machine-readable output for the latter.
12101 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12102 of the e-mail address in the DN (i.e., it will go into a certificate
12103 extension only). The new configuration file option 'email_in_dn = no'
12104 has the same effect.
12106 *Massimiliano Pala madwolf@openca.org*
12108 * Change all functions with names starting with `des_` to be starting
12109 with `DES_` instead. Add wrappers that are compatible with libdes,
12110 but are named `_ossl_old_des_*`. Finally, add macros that map the
12111 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12112 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12113 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12116 Since we provide two compatibility mappings, the user needs to
12117 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12118 compatibility is desired. The default (i.e., when that macro
12119 isn't defined) is OpenSSL 0.9.6c compatibility.
12121 There are also macros that enable and disable the support of old
12122 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12123 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12124 are defined, the default will apply: to support the old des routines.
12126 In either case, one must include openssl/des.h to get the correct
12127 definitions. Do not try to just include openssl/des_old.h, that
12130 NOTE: This is a major break of an old API into a new one. Software
12131 authors are encouraged to switch to the `DES_` style functions. Some
12132 time in the future, des_old.h and the libdes compatibility functions
12133 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12134 default), and then completely removed.
12138 * Test for certificates which contain unsupported critical extensions.
12139 If such a certificate is found during a verify operation it is
12140 rejected by default: this behaviour can be overridden by either
12141 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12142 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12143 X509_supported_extension() has also been added which returns 1 if a
12144 particular extension is supported.
12148 * Modify the behaviour of EVP cipher functions in similar way to digests
12149 to retain compatibility with existing code.
12153 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12154 compatibility with existing code. In particular the 'ctx' parameter does
12155 not have to be to be initialized before the call to EVP_DigestInit() and
12156 it is tidied up after a call to EVP_DigestFinal(). New function
12157 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12158 EVP_MD_CTX_copy() changed to not require the destination to be
12159 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12160 requires the destination to be valid.
12162 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12163 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12167 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12168 so that complete 'Handshake' protocol structures are kept in memory
12169 instead of overwriting 'msg_type' and 'length' with 'body' data.
12173 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12175 *Massimo Santin via Richard Levitte*
12177 * Major restructuring to the underlying ENGINE code. This includes
12178 reduction of linker bloat, separation of pure "ENGINE" manipulation
12179 (initialisation, etc) from functionality dealing with implementations
12180 of specific crypto interfaces. This change also introduces integrated
12181 support for symmetric ciphers and digest implementations - so ENGINEs
12182 can now accelerate these by providing EVP_CIPHER and EVP_MD
12183 implementations of their own. This is detailed in
12184 [crypto/engine/README.md](crypto/engine/README.md)
12185 as it couldn't be adequately described here. However, there are a few
12186 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12187 were changed in the original introduction of ENGINE code have now
12188 reverted back - the hooking from this code to ENGINE is now a good
12189 deal more passive and at run-time, operations deal directly with
12190 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12191 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12192 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12193 they were not being used by the framework as there is no concept of a
12194 BIGNUM_METHOD and they could not be generalised to the new
12195 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12196 ENGINE_cpy() has been removed as it cannot be consistently defined in
12201 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12205 * Change mkdef.pl to sort symbols that get the same entry number,
12206 and make sure the automatically generated functions `ERR_load_*`
12207 become part of libeay.num as well.
12211 * New function SSL_renegotiate_pending(). This returns true once
12212 renegotiation has been requested (either SSL_renegotiate() call
12213 or HelloRequest/ClientHello received from the peer) and becomes
12214 false once a handshake has been completed.
12215 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12216 sends a HelloRequest, but does not ensure that a handshake takes
12217 place. SSL_renegotiate_pending() is useful for checking if the
12218 client has followed the request.)
12222 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12223 By default, clients may request session resumption even during
12224 renegotiation (if session ID contexts permit); with this option,
12225 session resumption is possible only in the first handshake.
12227 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12228 more bits available for options that should not be part of
12229 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12233 * Add some demos for certificate and certificate request creation.
12237 * Make maximum certificate chain size accepted from the peer application
12238 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12239 "Douglas E. Engert" <deengert@anl.gov>.
12243 * Add support for shared libraries for Unixware-7
12244 (Boyd Lynn Gerber <gerberb@zenez.com>).
12248 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12249 be done prior to destruction. Use this to unload error strings from
12250 ENGINEs that load their own error strings. NB: This adds two new API
12251 functions to "get" and "set" this destroy handler in an ENGINE.
12255 * Alter all existing ENGINE implementations (except "openssl" and
12256 "openbsd") to dynamically instantiate their own error strings. This
12257 makes them more flexible to be built both as statically-linked ENGINEs
12258 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12259 Also, add stub code to each that makes building them as self-contained
12260 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12264 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12265 implementations into applications that are completely implemented in
12266 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12267 commands that can be used to configure what shared-library to load and
12268 to control aspects of the way it is handled. Also, made an update to
12269 the [README-Engine.md](README-Engine.md) file
12270 that brings its information up-to-date and
12271 provides some information and instructions on the "dynamic" ENGINE
12272 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12276 * Make it possible to unload ranges of ERR strings with a new
12277 "ERR_unload_strings" function.
12281 * Add a copy() function to EVP_MD.
12285 * Make EVP_MD routines take a context pointer instead of just the
12286 md_data void pointer.
12290 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12291 that the digest can only process a single chunk of data
12292 (typically because it is provided by a piece of
12293 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12294 is only going to provide a single chunk of data, and hence the
12295 framework needn't accumulate the data for oneshot drivers.
12299 * As with "ERR", make it possible to replace the underlying "ex_data"
12300 functions. This change also alters the storage and management of global
12301 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12302 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12303 index counters. The API functions that use this state have been changed
12304 to take a "class_index" rather than pointers to the class's local STACK
12305 and counter, and there is now an API function to dynamically create new
12306 classes. This centralisation allows us to (a) plug a lot of the
12307 thread-safety problems that existed, and (b) makes it possible to clean
12308 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12309 such data would previously have always leaked in application code and
12310 workarounds were in place to make the memory debugging turn a blind eye
12311 to it. Application code that doesn't use this new function will still
12312 leak as before, but their memory debugging output will announce it now
12313 rather than letting it slide.
12315 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12316 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12317 has a return value to indicate success or failure.
12321 * Make it possible to replace the underlying "ERR" functions such that the
12322 global state (2 LHASH tables and 2 locks) is only used by the "default"
12323 implementation. This change also adds two functions to "get" and "set"
12324 the implementation prior to it being automatically set the first time
12325 any other ERR function takes place. Ie. an application can call "get",
12326 pass the return value to a module it has just loaded, and that module
12327 can call its own "set" function using that value. This means the
12328 module's "ERR" operations will use (and modify) the error state in the
12329 application and not in its own statically linked copy of OpenSSL code.
12333 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12334 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12335 the operation, and provides a more encapsulated way for external code
12336 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12337 to use these functions rather than manually incrementing the counts.
12339 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12343 * Add EVP test program.
12347 * Add symmetric cipher support to ENGINE. Expect the API to change!
12351 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12352 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12353 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12354 These allow a CRL to be built without having to access X509_CRL fields
12355 directly. Modify 'ca' application to use new functions.
12359 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12360 bug workarounds. Rollback attack detection is a security feature.
12361 The problem will only arise on OpenSSL servers when TLSv1 is not
12362 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12363 Software authors not wanting to support TLSv1 will have special reasons
12364 for their choice and can explicitly enable this option.
12366 *Bodo Moeller, Lutz Jaenicke*
12368 * Rationalise EVP so it can be extended: don't include a union of
12369 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12370 (similar to those existing for EVP_CIPHER_CTX).
12375 EVP_MD_CTX_init(&md); /* new function call */
12376 EVP_DigestInit(&md, EVP_sha1());
12377 EVP_DigestUpdate(&md, in, len);
12378 EVP_DigestFinal(&md, out, NULL);
12379 EVP_MD_CTX_cleanup(&md); /* new function call */
12383 * Make DES key schedule conform to the usual scheme, as well as
12384 correcting its structure. This means that calls to DES functions
12385 now have to pass a pointer to a des_key_schedule instead of a
12386 plain des_key_schedule (which was actually always a pointer
12389 des_key_schedule ks;
12391 des_set_key_checked(..., &ks);
12392 des_ncbc_encrypt(..., &ks, ...);
12394 (Note that a later change renames 'des_...' into 'DES_...'.)
12398 * Initial reduction of linker bloat: the use of some functions, such as
12399 PEM causes large amounts of unused functions to be linked in due to
12400 poor organisation. For example pem_all.c contains every PEM function
12401 which has a knock on effect of linking in large amounts of (unused)
12402 ASN1 code. Grouping together similar functions and splitting unrelated
12403 functions prevents this.
12407 * Cleanup of EVP macros.
12411 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12412 correct `_ecb suffix`.
12416 * Add initial OCSP responder support to ocsp application. The
12417 revocation information is handled using the text based index
12418 use by the ca application. The responder can either handle
12419 requests generated internally, supplied in files (for example
12420 via a CGI script) or using an internal minimal server.
12424 * Add configuration choices to get zlib compression for TLS.
12428 * Changes to Kerberos SSL for RFC 2712 compliance:
12429 1. Implemented real KerberosWrapper, instead of just using
12430 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12431 2. Implemented optional authenticator field of KerberosWrapper.
12433 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12434 and authenticator structs; see crypto/krb5/.
12436 Generalized Kerberos calls to support multiple Kerberos libraries.
12437 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12438 via Richard Levitte*
12440 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12441 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12442 values for each of the key sizes rather than having just
12443 parameters (and 'speed' generating keys each time).
12447 * Speed up EVP routines.
12450 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12451 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12452 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12453 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12455 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12456 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12457 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12460 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12462 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12466 * Added the OS2-EMX target.
12468 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12470 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12471 New functions to support `NCONF `routines in extension code.
12472 New function `CONF_set_nconf()`
12473 to allow functions which take an `NCONF` to also handle the old `LHASH`
12474 structure: this means that the old `CONF` compatible routines can be
12475 retained (in particular w.rt. extensions) without having to duplicate the
12476 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12480 * Enhance the general user interface with mechanisms for inner control
12481 and with possibilities to have yes/no kind of prompts.
12485 * Change all calls to low level digest routines in the library and
12486 applications to use EVP. Add missing calls to HMAC_cleanup() and
12487 don't assume HMAC_CTX can be copied using memcpy().
12489 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12491 * Add the possibility to control engines through control names but with
12492 arbitrary arguments instead of just a string.
12493 Change the key loaders to take a UI_METHOD instead of a callback
12494 function pointer. NOTE: this breaks binary compatibility with earlier
12495 versions of OpenSSL [engine].
12496 Adapt the nCipher code for these new conditions and add a card insertion
12501 * Enhance the general user interface with mechanisms to better support
12502 dialog box interfaces, application-defined prompts, the possibility
12503 to use defaults (for example default passwords from somewhere else)
12504 and interrupts/cancellations.
12508 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12509 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12513 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12514 tidy up some unnecessarily weird code in 'sk_new()').
12516 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12518 * Change the key loading routines for ENGINEs to use the same kind
12519 callback (pem_password_cb) as all other routines that need this
12524 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12525 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12526 than this minimum value is recommended.
12530 * New random seeder for OpenVMS, using the system process statistics
12531 that are easily reachable.
12535 * Windows apparently can't transparently handle global
12536 variables defined in DLLs. Initialisations such as:
12538 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12540 won't compile. This is used by the any applications that need to
12541 declare their own ASN1 modules. This was fixed by adding the option
12542 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12543 needed for static libraries under Win32.
12547 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12548 setting of purpose and trust fields. New X509_STORE trust and
12549 purpose functions and tidy up setting in other SSL functions.
12553 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12554 structure. These are inherited by X509_STORE_CTX when it is
12555 initialised. This allows various defaults to be set in the
12556 X509_STORE structure (such as flags for CRL checking and custom
12557 purpose or trust settings) for functions which only use X509_STORE_CTX
12558 internally such as S/MIME.
12560 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12561 trust settings if they are not set in X509_STORE. This allows X509_STORE
12562 purposes and trust (in S/MIME for example) to override any set by default.
12564 Add command line options for CRL checking to smime, s_client and s_server
12569 * Initial CRL based revocation checking. If the CRL checking flag(s)
12570 are set then the CRL is looked up in the X509_STORE structure and
12571 its validity and signature checked, then if the certificate is found
12572 in the CRL the verify fails with a revoked error.
12574 Various new CRL related callbacks added to X509_STORE_CTX structure.
12576 Command line options added to 'verify' application to support this.
12578 This needs some additional work, such as being able to handle multiple
12579 CRLs with different times, extension based lookup (rather than just
12580 by subject name) and ultimately more complete V2 CRL extension
12585 * Add a general user interface API (crypto/ui/). This is designed
12586 to replace things like des_read_password and friends (backward
12587 compatibility functions using this new API are provided).
12588 The purpose is to remove prompting functions from the DES code
12589 section as well as provide for prompting through dialog boxes in
12590 a window system and the like.
12594 * Add "ex_data" support to ENGINE so implementations can add state at a
12595 per-structure level rather than having to store it globally.
12599 * Make it possible for ENGINE structures to be copied when retrieved by
12600 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12601 This causes the "original" ENGINE structure to act like a template,
12602 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12603 operational state can be localised to each ENGINE structure, despite the
12604 fact they all share the same "methods". New ENGINE structures returned in
12605 this case have no functional references and the return value is the single
12606 structural reference. This matches the single structural reference returned
12607 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12612 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12613 needs to match any other type at all we need to manually clear the
12618 * Changes to the "openssl engine" utility to include;
12619 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12620 about an ENGINE's available control commands.
12621 - executing control commands from command line arguments using the
12622 '-pre' and '-post' switches. '-post' is only used if '-t' is
12623 specified and the ENGINE is successfully initialised. The syntax for
12624 the individual commands are colon-separated, for example;
12625 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12629 * New dynamic control command support for ENGINEs. ENGINEs can now
12630 declare their own commands (numbers), names (strings), descriptions,
12631 and input types for run-time discovery by calling applications. A
12632 subset of these commands are implicitly classed as "executable"
12633 depending on their input type, and only these can be invoked through
12634 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12635 can be based on user input, config files, etc). The distinction is
12636 that "executable" commands cannot return anything other than a boolean
12637 result and can only support numeric or string input, whereas some
12638 discoverable commands may only be for direct use through
12639 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12640 pointers, or other custom uses. The "executable" commands are to
12641 support parameterisations of ENGINE behaviour that can be
12642 unambiguously defined by ENGINEs and used consistently across any
12643 OpenSSL-based application. Commands have been added to all the
12644 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12645 control over shared-library paths without source code alterations.
12649 * Changed all ENGINE implementations to dynamically allocate their
12650 ENGINEs rather than declaring them statically. Apart from this being
12651 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12652 this also allows the implementations to compile without using the
12653 internal engine_int.h header.
12657 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12658 'const' value. Any code that should be able to modify a RAND_METHOD
12659 should already have non-const pointers to it (ie. they should only
12660 modify their own ones).
12664 * Made a variety of little tweaks to the ENGINE code.
12665 - "atalla" and "ubsec" string definitions were moved from header files
12666 to C code. "nuron" string definitions were placed in variables
12667 rather than hard-coded - allowing parameterisation of these values
12668 later on via ctrl() commands.
12669 - Removed unused "#if 0"'d code.
12670 - Fixed engine list iteration code so it uses ENGINE_free() to release
12671 structural references.
12672 - Constified the RAND_METHOD element of ENGINE structures.
12673 - Constified various get/set functions as appropriate and added
12674 missing functions (including a catch-all ENGINE_cpy that duplicates
12675 all ENGINE values onto a new ENGINE except reference counts/state).
12676 - Removed NULL parameter checks in get/set functions. Setting a method
12677 or function to NULL is a way of cancelling out a previously set
12678 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12679 and doesn't justify the extra error symbols and code.
12680 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12681 flags from engine_int.h to engine.h.
12682 - Changed prototypes for ENGINE handler functions (init(), finish(),
12683 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12687 * Implement binary inversion algorithm for BN_mod_inverse in addition
12688 to the algorithm using long division. The binary algorithm can be
12689 used only if the modulus is odd. On 32-bit systems, it is faster
12690 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12691 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12692 up to 450 bits. In 64-bit environments, the binary algorithm
12693 appears to be advantageous for much longer moduli; here we use it
12694 for moduli up to 2048 bits.
12698 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12699 could not support the combine flag in choice fields.
12703 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12704 extensions from a certificate request to the certificate.
12708 * Allow multiple 'certopt' and 'nameopt' options to be separated
12709 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12710 file: this allows the display of the certificate about to be
12711 signed to be customised, to allow certain fields to be included
12712 or excluded and extension details. The old system didn't display
12713 multicharacter strings properly, omitted fields not in the policy
12714 and couldn't display additional details such as extensions.
12718 * Function EC_POINTs_mul for multiple scalar multiplication
12719 of an arbitrary number of elliptic curve points
12720 \sum scalars[i]*points[i],
12721 optionally including the generator defined for the EC_GROUP:
12722 scalar*generator + \sum scalars[i]*points[i].
12724 EC_POINT_mul is a simple wrapper function for the typical case
12725 that the point list has just one item (besides the optional
12730 * First EC_METHODs for curves over GF(p):
12732 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12733 operations and provides various method functions that can also
12734 operate with faster implementations of modular arithmetic.
12736 EC_GFp_mont_method() reuses most functions that are part of
12737 EC_GFp_simple_method, but uses Montgomery arithmetic.
12739 *Bodo Moeller; point addition and point doubling
12740 implementation directly derived from source code provided by
12741 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12743 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12744 crypto/ec/ec_lib.c):
12746 Curves are EC_GROUP objects (with an optional group generator)
12747 based on EC_METHODs that are built into the library.
12749 Points are EC_POINT objects based on EC_GROUP objects.
12751 Most of the framework would be able to handle curves over arbitrary
12752 finite fields, but as there are no obvious types for fields other
12753 than GF(p), some functions are limited to that for now.
12757 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12758 that the file contains a complete HTTP response.
12762 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12763 change the def and num file printf format specifier from "%-40sXXX"
12764 to "%-39s XXX". The latter will always guarantee a space after the
12765 field while the former will cause them to run together if the field
12766 is 40 of more characters long.
12770 * Constify the cipher and digest 'method' functions and structures
12771 and modify related functions to take constant EVP_MD and EVP_CIPHER
12776 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12777 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12781 * Modify `EVP_Digest*()` routines so they now return values. Although the
12782 internal software routines can never fail additional hardware versions
12787 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12789 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12790 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12793 ERR_R_NESTED_ASN1_ERROR
12795 ERR_R_MISSING_ASN1_EOS
12796 were 4 .. 9, conflicting with
12797 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12799 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12800 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12802 Add new error code 'ERR_R_INTERNAL_ERROR'.
12806 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12811 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12812 sets the subject name for a new request or supersedes the
12813 subject name in a given request. Formats that can be parsed are
12814 'CN=Some Name, OU=myOU, C=IT'
12816 'CN=Some Name/OU=myOU/C=IT'.
12818 Add options '-batch' and '-verbose' to 'openssl req'.
12820 *Massimiliano Pala <madwolf@hackmasters.net>*
12822 * Introduce the possibility to access global variables through
12823 functions on platform were that's the best way to handle exporting
12824 global variables in shared libraries. To enable this functionality,
12825 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12826 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12827 is normally done by Configure or something similar).
12829 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12830 in the source file (foo.c) like this:
12832 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12833 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12835 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12836 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12838 OPENSSL_DECLARE_GLOBAL(int,foo);
12839 #define foo OPENSSL_GLOBAL_REF(foo)
12840 OPENSSL_DECLARE_GLOBAL(double,bar);
12841 #define bar OPENSSL_GLOBAL_REF(bar)
12843 The #defines are very important, and therefore so is including the
12844 header file everywhere where the defined globals are used.
12846 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12847 of ASN.1 items, but that structure is a bit different.
12849 The largest change is in util/mkdef.pl which has been enhanced with
12850 better and easier to understand logic to choose which symbols should
12851 go into the Windows .def files as well as a number of fixes and code
12852 cleanup (among others, algorithm keywords are now sorted
12853 lexicographically to avoid constant rewrites).
12857 * In BN_div() keep a copy of the sign of 'num' before writing the
12858 result to 'rm' because if rm==num the value will be overwritten
12859 and produce the wrong result if 'num' is negative: this caused
12860 problems with BN_mod() and BN_nnmod().
12864 * Function OCSP_request_verify(). This checks the signature on an
12865 OCSP request and verifies the signer certificate. The signer
12866 certificate is just checked for a generic purpose and OCSP request
12871 * Add OCSP_check_validity() function to check the validity of OCSP
12872 responses. OCSP responses are prepared in real time and may only
12873 be a few seconds old. Simply checking that the current time lies
12874 between thisUpdate and nextUpdate max reject otherwise valid responses
12875 caused by either OCSP responder or client clock inaccuracy. Instead
12876 we allow thisUpdate and nextUpdate to fall within a certain period of
12877 the current time. The age of the response can also optionally be
12878 checked. Two new options -validity_period and -status_age added to
12883 * If signature or public key algorithm is unrecognized print out its
12884 OID rather that just UNKNOWN.
12888 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12889 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12890 ID to be generated from the issuer certificate alone which can then be
12891 passed to OCSP_id_issuer_cmp().
12895 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12896 ASN1 modules to export functions returning ASN1_ITEM pointers
12897 instead of the ASN1_ITEM structures themselves. This adds several
12898 new macros which allow the underlying ASN1 function/structure to
12899 be accessed transparently. As a result code should not use ASN1_ITEM
12900 references directly (such as &X509_it) but instead use the relevant
12901 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12902 use of the new ASN1 code on platforms where exporting structures
12903 is problematical (for example in shared libraries) but exporting
12904 functions returning pointers to structures is not.
12908 * Add support for overriding the generation of SSL/TLS session IDs.
12909 These callbacks can be registered either in an SSL_CTX or per SSL.
12910 The purpose of this is to allow applications to control, if they wish,
12911 the arbitrary values chosen for use as session IDs, particularly as it
12912 can be useful for session caching in multiple-server environments. A
12913 command-line switch for testing this (and any client code that wishes
12914 to use such a feature) has been added to "s_server".
12916 *Geoff Thorpe, Lutz Jaenicke*
12918 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12919 of the form `#if defined(...) || defined(...) || ...` and
12920 `#if !defined(...) && !defined(...) && ...`. This also avoids
12921 the growing number of special cases it was previously handling.
12925 * Make all configuration macros available for application by making
12926 sure they are available in opensslconf.h, by giving them names starting
12927 with `OPENSSL_` to avoid conflicts with other packages and by making
12928 sure e_os2.h will cover all platform-specific cases together with
12930 Additionally, it is now possible to define configuration/platform-
12931 specific names (called "system identities"). In the C code, these
12932 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12933 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12934 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12939 * New option -set_serial to 'req' and 'x509' this allows the serial
12940 number to use to be specified on the command line. Previously self
12941 signed certificates were hard coded with serial number 0 and the
12942 CA options of 'x509' had to use a serial number in a file which was
12947 * New options to 'ca' utility to support V2 CRL entry extensions.
12948 Currently CRL reason, invalidity date and hold instruction are
12949 supported. Add new CRL extensions to V3 code and some new objects.
12953 * New function EVP_CIPHER_CTX_set_padding() this is used to
12954 disable standard block padding (aka PKCS#5 padding) in the EVP
12955 API, which was previously mandatory. This means that the data is
12956 not padded in any way and so the total length much be a multiple
12957 of the block size, otherwise an error occurs.
12961 * Initial (incomplete) OCSP SSL support.
12965 * New function OCSP_parse_url(). This splits up a URL into its host,
12966 port and path components: primarily to parse OCSP URLs. New -url
12967 option to ocsp utility.
12971 * New nonce behavior. The return value of OCSP_check_nonce() now
12972 reflects the various checks performed. Applications can decide
12973 whether to tolerate certain situations such as an absent nonce
12974 in a response when one was present in a request: the ocsp application
12975 just prints out a warning. New function OCSP_add1_basic_nonce()
12976 this is to allow responders to include a nonce in a response even if
12977 the request is nonce-less.
12981 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12982 skipped when using openssl x509 multiple times on a single input file,
12983 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12987 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12988 set string type: to handle setting ASN1_TIME structures. Fix ca
12989 utility to correctly initialize revocation date of CRLs.
12993 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12994 the clients preferred ciphersuites and rather use its own preferences.
12995 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12996 Internet Explorer by ensuring unchanged hash method during stepup.
12997 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13001 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13002 to aes and add a new 'exist' option to print out symbols that don't
13007 * Additional options to ocsp utility to allow flags to be set and
13008 additional certificates supplied.
13012 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13013 OCSP client a number of certificate to only verify the response
13018 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13019 handle the new API. Currently only ECB, CBC modes supported. Add new
13022 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13023 Encryption Standard (AES) Ciphersuites for Transport Layer
13024 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13025 not enabled by default and were not part of the "ALL" ciphersuite
13026 alias because they were not yet official; they could be
13027 explicitly requested by specifying the "AESdraft" ciphersuite
13028 group alias. In the final release of OpenSSL 0.9.7, the group
13029 alias is called "AES" and is part of "ALL".)
13031 *Ben Laurie, Steve Henson, Bodo Moeller*
13033 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13034 request to response.
13038 * Functions for OCSP responders. OCSP_request_onereq_count(),
13039 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13040 extract information from a certificate request. OCSP_response_create()
13041 creates a response and optionally adds a basic response structure.
13042 OCSP_basic_add1_status() adds a complete single response to a basic
13043 response and returns the OCSP_SINGLERESP structure just added (to allow
13044 extensions to be included for example). OCSP_basic_add1_cert() adds a
13045 certificate to a basic response and OCSP_basic_sign() signs a basic
13046 response with various flags. New helper functions ASN1_TIME_check()
13047 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13048 (converts ASN1_TIME to GeneralizedTime).
13052 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13053 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13054 structure from a certificate. X509_pubkey_digest() digests the public_key
13055 contents: this is used in various key identifiers.
13059 * Make sk_sort() tolerate a NULL argument.
13061 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13063 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13064 passed by the function are trusted implicitly. If any of them signed the
13065 response then it is assumed to be valid and is not verified.
13069 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13070 to data. This was previously part of the PKCS7 ASN1 code. This
13071 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13072 *Steve Henson, reported by Kenneth R. Robinette
13073 <support@securenetterm.com>*
13075 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13076 routines: without these tracing memory leaks is very painful.
13077 Fix leaks in PKCS12 and PKCS7 routines.
13081 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13082 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13083 effectively meant GeneralizedTime would never be used. Now it
13084 is initialised to -1 but X509_time_adj() now has to check the value
13085 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13086 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13087 *Steve Henson, reported by Kenneth R. Robinette
13088 <support@securenetterm.com>*
13090 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13091 result in a zero length in the ASN1_INTEGER structure which was
13092 not consistent with the structure when d2i_ASN1_INTEGER() was used
13093 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13094 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13095 where it did not print out a minus for negative ASN1_INTEGER.
13099 * Add summary printout to ocsp utility. The various functions which
13100 convert status values to strings have been renamed to:
13101 OCSP_response_status_str(), OCSP_cert_status_str() and
13102 OCSP_crl_reason_str() and are no longer static. New options
13103 to verify nonce values and to disable verification. OCSP response
13104 printout format cleaned up.
13108 * Add additional OCSP certificate checks. These are those specified
13109 in RFC2560. This consists of two separate checks: the CA of the
13110 certificate being checked must either be the OCSP signer certificate
13111 or the issuer of the OCSP signer certificate. In the latter case the
13112 OCSP signer certificate must contain the OCSP signing extended key
13113 usage. This check is performed by attempting to match the OCSP
13114 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13115 in the OCSP_CERTID structures of the response.
13119 * Initial OCSP certificate verification added to OCSP_basic_verify()
13120 and related routines. This uses the standard OpenSSL certificate
13121 verify routines to perform initial checks (just CA validity) and
13122 to obtain the certificate chain. Then additional checks will be
13123 performed on the chain. Currently the root CA is checked to see
13124 if it is explicitly trusted for OCSP signing. This is used to set
13125 a root CA as a global signing root: that is any certificate that
13126 chains to that CA is an acceptable OCSP signing certificate.
13130 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13131 extensions from a separate configuration file.
13132 As when reading extensions from the main configuration file,
13133 the '-extensions ...' option may be used for specifying the
13136 *Massimiliano Pala <madwolf@comune.modena.it>*
13138 * New OCSP utility. Allows OCSP requests to be generated or
13139 read. The request can be sent to a responder and the output
13140 parsed, outputted or printed in text form. Not complete yet:
13141 still needs to check the OCSP response validity.
13145 * New subcommands for 'openssl ca':
13146 `openssl ca -status <serial>` prints the status of the cert with
13147 the given serial number (according to the index file).
13148 `openssl ca -updatedb` updates the expiry status of certificates
13151 *Massimiliano Pala <madwolf@comune.modena.it>*
13153 * New '-newreq-nodes' command option to CA.pl. This is like
13154 '-newreq', but calls 'openssl req' with the '-nodes' option
13155 so that the resulting key is not encrypted.
13157 *Damien Miller <djm@mindrot.org>*
13159 * New configuration for the GNU Hurd.
13161 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13163 * Initial code to implement OCSP basic response verify. This
13164 is currently incomplete. Currently just finds the signer's
13165 certificate and verifies the signature on the response.
13169 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13170 value of OPENSSLDIR. This is available via the new '-d' option
13171 to 'openssl version', and is also included in 'openssl version -a'.
13175 * Allowing defining memory allocation callbacks that will be given
13176 file name and line number information in additional arguments
13177 (a `const char*` and an int). The basic functionality remains, as
13178 well as the original possibility to just replace malloc(),
13179 realloc() and free() by functions that do not know about these
13180 additional arguments. To register and find out the current
13181 settings for extended allocation functions, the following
13182 functions are provided:
13184 CRYPTO_set_mem_ex_functions
13185 CRYPTO_set_locked_mem_ex_functions
13186 CRYPTO_get_mem_ex_functions
13187 CRYPTO_get_locked_mem_ex_functions
13189 These work the same way as CRYPTO_set_mem_functions and friends.
13190 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13191 extended allocation function is enabled.
13192 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13193 a conventional allocation function is enabled.
13195 *Richard Levitte, Bodo Moeller*
13197 * Finish off removing the remaining LHASH function pointer casts.
13198 There should no longer be any prototype-casting required when using
13199 the LHASH abstraction, and any casts that remain are "bugs". See
13200 the callback types and macros at the head of lhash.h for details
13201 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13205 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13206 If /dev/[u]random devices are not available or do not return enough
13207 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13209 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13210 /etc/entropy will be queried once each in this sequence, querying stops
13211 when enough entropy was collected without querying more sockets.
13215 * Change the Unix RAND_poll() variant to be able to poll several
13216 random devices, as specified by DEVRANDOM, until a sufficient amount
13217 of data has been collected. We spend at most 10 ms on each file
13218 (select timeout) and read in non-blocking mode. DEVRANDOM now
13219 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13220 (previously it was just the string "/dev/urandom"), so on typical
13221 platforms the 10 ms delay will never occur.
13222 Also separate out the Unix variant to its own file, rand_unix.c.
13223 For VMS, there's a currently-empty rand_vms.c.
13227 * Move OCSP client related routines to ocsp_cl.c. These
13228 provide utility functions which an application needing
13229 to issue a request to an OCSP responder and analyse the
13230 response will typically need: as opposed to those which an
13231 OCSP responder itself would need which will be added later.
13233 OCSP_request_sign() signs an OCSP request with an API similar
13234 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13235 response. OCSP_response_get1_basic() extracts basic response
13236 from response. OCSP_resp_find_status(): finds and extracts status
13237 information from an OCSP_CERTID structure (which will be created
13238 when the request structure is built). These are built from lower
13239 level functions which work on OCSP_SINGLERESP structures but
13240 won't normally be used unless the application wishes to examine
13241 extensions in the OCSP response for example.
13243 Replace nonce routines with a pair of functions.
13244 OCSP_request_add1_nonce() adds a nonce value and optionally
13245 generates a random value. OCSP_check_nonce() checks the
13246 validity of the nonce in an OCSP response.
13250 * Change function OCSP_request_add() to OCSP_request_add0_id().
13251 This doesn't copy the supplied OCSP_CERTID and avoids the
13252 need to free up the newly created id. Change return type
13253 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13254 This can then be used to add extensions to the request.
13255 Deleted OCSP_request_new(), since most of its functionality
13256 is now in OCSP_REQUEST_new() (and the case insensitive name
13257 clash) apart from the ability to set the request name which
13258 will be added elsewhere.
13262 * Update OCSP API. Remove obsolete extensions argument from
13263 various functions. Extensions are now handled using the new
13264 OCSP extension code. New simple OCSP HTTP function which
13265 can be used to send requests and parse the response.
13269 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13270 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13271 uses the special reorder version of SET OF to sort the attributes
13272 and reorder them to match the encoded order. This resolves a long
13273 standing problem: a verify on a PKCS7 structure just after signing
13274 it used to fail because the attribute order did not match the
13275 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13276 it uses the received order. This is necessary to tolerate some broken
13277 software that does not order SET OF. This is handled by encoding
13278 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13279 to produce the required SET OF.
13283 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13284 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13285 files to get correct declarations of the ASN.1 item variables.
13289 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13290 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13291 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13292 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13293 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13294 ASN1_ITEM and no wrapper functions.
13298 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13299 replace the old function pointer based I/O routines. Change most of
13300 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13304 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13305 lines, recognize more "algorithms" that can be deselected, and make
13306 it complain about algorithm deselection that isn't recognised.
13310 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13311 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13312 to use new functions. Add NO_ASN1_OLD which can be set to remove
13313 some old style ASN1 functions: this can be used to determine if old
13314 code will still work when these eventually go away.
13318 * New extension functions for OCSP structures, these follow the
13319 same conventions as certificates and CRLs.
13323 * New function X509V3_add1_i2d(). This automatically encodes and
13324 adds an extension. Its behaviour can be customised with various
13325 flags to append, replace or delete. Various wrappers added for
13326 certificates and CRLs.
13330 * Fix to avoid calling the underlying ASN1 print routine when
13331 an extension cannot be parsed. Correct a typo in the
13332 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13336 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13337 entries for variables.
13341 * Add functionality to `apps/openssl.c` for detecting locking
13342 problems: As the program is single-threaded, all we have
13343 to do is register a locking callback using an array for
13344 storing which locks are currently held by the program.
13348 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13349 SSL_get_ex_data_X509_STORE_idx(), which is used in
13350 ssl_verify_cert_chain() and thus can be called at any time
13351 during TLS/SSL handshakes so that thread-safety is essential.
13352 Unfortunately, the ex_data design is not at all suited
13353 for multi-threaded use, so it probably should be abolished.
13357 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13359 *Broadcom, tweaked and integrated by Geoff Thorpe*
13361 * Move common extension printing code to new function
13362 X509V3_print_extensions(). Reorganise OCSP print routines and
13363 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13367 * New function X509_signature_print() to remove duplication in some
13372 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13373 set (this was treated exactly the same as SET OF previously). This
13374 is used to reorder the STACK representing the structure to match the
13375 encoding. This will be used to get round a problem where a PKCS7
13376 structure which was signed could not be verified because the STACK
13377 order did not reflect the encoded order.
13381 * Reimplement the OCSP ASN1 module using the new code.
13385 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13386 for its ASN1 operations. The old style function pointers still exist
13387 for now but they will eventually go away.
13391 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13392 completely replaces the old ASN1 functionality with a table driven
13393 encoder and decoder which interprets an ASN1_ITEM structure describing
13394 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13395 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13396 has also been converted to the new form.
13400 * Change BN_mod_exp_recp so that negative moduli are tolerated
13401 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13402 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13403 for negative moduli.
13407 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13408 of not touching the result's sign bit.
13412 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13417 * Changed the LHASH code to use prototypes for callbacks, and created
13418 macros to declare and implement thin (optionally static) functions
13419 that provide type-safety and avoid function pointer casting for the
13420 type-specific callbacks.
13424 * Added Kerberos Cipher Suites to be used with TLS, as written in
13426 *Veers Staats <staatsvr@asc.hpc.mil>,
13427 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13429 * Reformat the FAQ so the different questions and answers can be divided
13430 in sections depending on the subject.
13434 * Have the zlib compression code load ZLIB.DLL dynamically under
13439 * New function BN_mod_sqrt for computing square roots modulo a prime
13440 (using the probabilistic Tonelli-Shanks algorithm unless
13441 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13442 be handled deterministically).
13444 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13446 * Make BN_mod_inverse faster by explicitly handling small quotients
13447 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13448 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13452 * New function BN_kronecker.
13456 * Fix BN_gcd so that it works on negative inputs; the result is
13457 positive unless both parameters are zero.
13458 Previously something reasonably close to an infinite loop was
13459 possible because numbers could be growing instead of shrinking
13460 in the implementation of Euclid's algorithm.
13464 * Fix BN_is_word() and BN_is_one() macros to take into account the
13465 sign of the number in question.
13467 Fix BN_is_word(a,w) to work correctly for w == 0.
13469 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13470 because its test if the absolute value of 'a' equals 'w'.
13471 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13472 it exists mostly for use in the implementations of BN_is_zero(),
13473 BN_is_one(), and BN_is_word().
13477 * New function BN_swap.
13481 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13482 the exponentiation functions are more likely to produce reasonable
13483 results on negative inputs.
13487 * Change BN_mod_mul so that the result is always non-negative.
13488 Previously, it could be negative if one of the factors was negative;
13489 I don't think anyone really wanted that behaviour.
13493 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13494 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13495 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13496 and add new functions:
13505 BN_mod_lshift1_quick
13507 BN_mod_lshift_quick
13509 These functions always generate non-negative results.
13511 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13512 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13514 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13515 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13516 be reduced modulo `m`.
13518 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13521 The following entry accidentally appeared in the CHANGES file
13522 distributed with OpenSSL 0.9.7. The modifications described in
13523 it do *not* apply to OpenSSL 0.9.7.
13525 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13526 was actually never needed) and in BN_mul(). The removal in BN_mul()
13527 required a small change in bn_mul_part_recursive() and the addition
13528 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13529 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13530 bn_sub_words() and bn_add_words() except they take arrays with
13536 * In 'openssl passwd', verify passwords read from the terminal
13537 unless the '-salt' option is used (which usually means that
13538 verification would just waste user's time since the resulting
13539 hash is going to be compared with some given password hash)
13540 or the new '-noverify' option is used.
13542 This is an incompatible change, but it does not affect
13543 non-interactive use of 'openssl passwd' (passwords on the command
13544 line, '-stdin' option, '-in ...' option) and thus should not
13545 cause any problems.
13549 * Remove all references to RSAref, since there's no more need for it.
13553 * Make DSO load along a path given through an environment variable
13554 (SHLIB_PATH) with shl_load().
13558 * Constify the ENGINE code as a result of BIGNUM constification.
13559 Also constify the RSA code and most things related to it. In a
13560 few places, most notable in the depth of the ASN.1 code, ugly
13561 casts back to non-const were required (to be solved at a later
13566 * Make it so the openssl application has all engines loaded by default.
13570 * Constify the BIGNUM routines a little more.
13574 * Add the following functions:
13576 ENGINE_load_cswift()
13578 ENGINE_load_atalla()
13579 ENGINE_load_nuron()
13580 ENGINE_load_builtin_engines()
13582 That way, an application can itself choose if external engines that
13583 are built-in in OpenSSL shall ever be used or not. The benefit is
13584 that applications won't have to be linked with libdl or other dso
13585 libraries unless it's really needed.
13587 Changed 'openssl engine' to load all engines on demand.
13588 Changed the engine header files to avoid the duplication of some
13589 declarations (they differed!).
13593 * 'openssl engine' can now list capabilities.
13597 * Better error reporting in 'openssl engine'.
13601 * Never call load_dh_param(NULL) in s_server.
13605 * Add engine application. It can currently list engines by name and
13606 identity, and test if they are actually available.
13610 * Improve RPM specification file by forcing symbolic linking and making
13611 sure the installed documentation is also owned by root.root.
13613 *Damien Miller <djm@mindrot.org>*
13615 * Give the OpenSSL applications more possibilities to make use of
13616 keys (public as well as private) handled by engines.
13620 * Add OCSP code that comes from CertCo.
13624 * Add VMS support for the Rijndael code.
13628 * Added untested support for Nuron crypto accelerator.
13632 * Add support for external cryptographic devices. This code was
13633 previously distributed separately as the "engine" branch.
13635 *Geoff Thorpe, Richard Levitte*
13637 * Rework the filename-translation in the DSO code. It is now possible to
13638 have far greater control over how a "name" is turned into a filename
13639 depending on the operating environment and any oddities about the
13640 different shared library filenames on each system.
13644 * Support threads on FreeBSD-elf in Configure.
13648 * Fix for SHA1 assembly problem with MASM: it produces
13649 warnings about corrupt line number information when assembling
13650 with debugging information. This is caused by the overlapping
13653 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13656 NCONF_get_number() has no error checking at all. As a replacement,
13657 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13658 promoted strongly. The old NCONF_get_number is kept around for
13659 binary backward compatibility.
13660 Make it possible for methods to load from something other than a BIO,
13661 by providing a function pointer that is given a name instead of a BIO.
13662 For example, this could be used to load configuration data from an
13667 * Fix for non blocking accept BIOs. Added new I/O special reason
13668 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13669 with non blocking I/O was not possible because no retry code was
13670 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13675 * Added the beginnings of Rijndael support.
13679 * Fix for bug in DirectoryString mask setting. Add support for
13680 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13681 to allow certificate printing to more controllable, additional
13682 'certopt' option to 'x509' to allow new printing options to be
13687 * Clean old EAY MD5 hack from e_os.h.
13691 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13693 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13694 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13696 *Joe Orton, Steve Henson*
13698 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13700 * Fix additional bug revealed by the NISCC test suite:
13702 Stop bug triggering large recursion when presented with
13703 certain ASN.1 tags ([CVE-2003-0851])
13707 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13709 * Fix various bugs revealed by running the NISCC test suite:
13711 Stop out of bounds reads in the ASN1 code when presented with
13712 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13714 If verify callback ignores invalid public key errors don't try to check
13715 certificate signature with the NULL public key.
13719 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13720 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13725 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13726 extra data after the compression methods not only for TLS 1.0
13727 but also for SSL 3.0 (as required by the specification).
13729 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13731 * Change X509_certificate_type() to mark the key as exported/exportable
13732 when it's 512 *bits* long, not 512 bytes.
13736 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13738 * Countermeasure against the Klima-Pokorny-Rosa extension of
13739 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13740 a protocol version number mismatch like a decryption error
13741 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13745 * Turn on RSA blinding by default in the default implementation
13746 to avoid a timing attack. Applications that don't want it can call
13747 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13748 They would be ill-advised to do so in most cases.
13750 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13752 * Change RSA blinding code so that it works when the PRNG is not
13753 seeded (in this case, the secret RSA exponent is abused as
13754 an unpredictable seed -- if it is not unpredictable, there
13755 is no point in blinding anyway). Make RSA blinding thread-safe
13756 by remembering the creator's thread ID in rsa->blinding and
13757 having all other threads use local one-time blinding factors
13758 (this requires more computation than sharing rsa->blinding, but
13759 avoids excessive locking; and if an RSA object is not shared
13760 between threads, blinding will still be very fast).
13764 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13766 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13767 via timing by performing a MAC computation even if incorrect
13768 block cipher padding has been found. This is a countermeasure
13769 against active attacks where the attacker has to distinguish
13770 between bad padding and a MAC verification error. ([CVE-2003-0078])
13772 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13773 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13774 Martin Vuagnoux (EPFL, Ilion)*
13776 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13778 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13779 memory from its contents. This is done with a counter that will
13780 place alternating values in each byte. This can be used to solve
13781 two issues: 1) the removal of calls to memset() by highly optimizing
13782 compilers, and 2) cleansing with other values than 0, since those can
13783 be read through on certain media, for example a swap space on disk.
13787 * Bugfix: client side session caching did not work with external caching,
13788 because the session->cipher setting was not restored when reloading
13789 from the external cache. This problem was masked, when
13790 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13791 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13795 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13796 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13798 *Zeev Lieber <zeev-l@yahoo.com>*
13800 * Undo an undocumented change introduced in 0.9.6e which caused
13801 repeated calls to OpenSSL_add_all_ciphers() and
13802 OpenSSL_add_all_digests() to be ignored, even after calling
13807 * Change the default configuration reader to deal with last line not
13808 being properly terminated.
13812 * Change X509_NAME_cmp() so it applies the special rules on handling
13813 DN values that are of type PrintableString, as well as RDNs of type
13814 emailAddress where the value has the type ia5String.
13816 *stefank@valicert.com via Richard Levitte*
13818 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13819 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13820 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13821 the bitwise-OR of the two for use by the majority of applications
13822 wanting this behaviour, and update the docs. The documented
13823 behaviour and actual behaviour were inconsistent and had been
13824 changing anyway, so this is more a bug-fix than a behavioural
13827 *Geoff Thorpe, diagnosed by Nadav Har'El*
13829 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13830 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13834 * Fix initialization code race conditions in
13835 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13836 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13837 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13838 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13839 ssl2_get_cipher_by_char(),
13840 ssl3_get_cipher_by_char().
13842 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13844 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13845 the cached sessions are flushed, as the remove_cb() might use ex_data
13846 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13847 (see [openssl.org #212]).
13849 *Geoff Thorpe, Lutz Jaenicke*
13851 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13852 length, instead of the encoding length to d2i_ASN1_OBJECT.
13856 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13858 * [In 0.9.6g-engine release:]
13859 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13861 *Lynn Gazis <lgazis@rainbow.com>*
13863 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13865 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13866 and get fix the header length calculation.
13867 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13868 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13870 * Use proper error handling instead of 'assertions' in buffer
13871 overflow checks added in 0.9.6e. This prevents DoS (the
13872 assertions could call abort()).
13874 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13876 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13878 * Add various sanity checks to asn1_get_length() to reject
13879 the ASN1 length bytes if they exceed sizeof(long), will appear
13880 negative or the content length exceeds the length of the
13883 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13885 * Fix cipher selection routines: ciphers without encryption had no flags
13886 for the cipher strength set and where therefore not handled correctly
13887 by the selection routines (PR #130).
13891 * Fix EVP_dsa_sha macro.
13896 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13897 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13898 that was added in OpenSSL 0.9.6d.
13900 As the countermeasure turned out to be incompatible with some
13901 broken SSL implementations, the new option is part of SSL_OP_ALL.
13902 SSL_OP_ALL is usually employed when compatibility with weird SSL
13903 implementations is desired (e.g. '-bugs' option to 's_client' and
13904 's_server'), so the new option is automatically set in many
13909 * Changes in security patch:
13911 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13912 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13913 Air Force Materiel Command, USAF, under agreement number
13916 * Add various sanity checks to asn1_get_length() to reject
13917 the ASN1 length bytes if they exceed sizeof(long), will appear
13918 negative or the content length exceeds the length of the
13919 supplied buffer. ([CVE-2002-0659])
13921 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13923 * Assertions for various potential buffer overflows, not known to
13924 happen in practice.
13926 *Ben Laurie (CHATS)*
13928 * Various temporary buffers to hold ASCII versions of integers were
13929 too small for 64 bit platforms. ([CVE-2002-0655])
13930 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13932 * Remote buffer overflow in SSL3 protocol - an attacker could
13933 supply an oversized session ID to a client. ([CVE-2002-0656])
13935 *Ben Laurie (CHATS)*
13937 * Remote buffer overflow in SSL2 protocol - an attacker could
13938 supply an oversized client master key. ([CVE-2002-0656])
13940 *Ben Laurie (CHATS)*
13942 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13944 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13945 encoded as NULL) with id-dsa-with-sha1.
13947 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13949 * Check various `X509_...()` return values in `apps/req.c`.
13951 *Nils Larsch <nla@trustcenter.de>*
13953 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13954 an end-of-file condition would erroneously be flagged, when the CRLF
13955 was just at the end of a processed block. The bug was discovered when
13956 processing data through a buffering memory BIO handing the data to a
13957 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13958 <ptsekov@syntrex.com> and Nedelcho Stanev.
13962 * Implement a countermeasure against a vulnerability recently found
13963 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13964 before application data chunks to avoid the use of known IVs
13965 with data potentially chosen by the attacker.
13969 * Fix length checks in ssl3_get_client_hello().
13973 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13974 to prevent ssl3_read_internal() from incorrectly assuming that
13975 ssl3_read_bytes() found application data while handshake
13976 processing was enabled when in fact s->s3->in_read_app_data was
13977 merely automatically cleared during the initial handshake.
13979 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13981 * Fix object definitions for Private and Enterprise: they were not
13982 recognized in their shortname (=lowercase) representation. Extend
13983 obj_dat.pl to issue an error when using undefined keywords instead
13984 of silently ignoring the problem (Svenning Sorensen
13985 <sss@sss.dnsalias.net>).
13989 * Fix DH_generate_parameters() so that it works for 'non-standard'
13990 generators, i.e. generators other than 2 and 5. (Previously, the
13991 code did not properly initialise the 'add' and 'rem' values to
13992 BN_generate_prime().)
13994 In the new general case, we do not insist that 'generator' is
13995 actually a primitive root: This requirement is rather pointless;
13996 a generator of the order-q subgroup is just as good, if not
14001 * Map new X509 verification errors to alerts. Discovered and submitted by
14002 Tom Wu <tom@arcot.com>.
14006 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14007 returning non-zero before the data has been completely received
14008 when using non-blocking I/O.
14010 *Bodo Moeller; problem pointed out by John Hughes*
14012 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14014 *Ben Laurie, Lutz Jaenicke*
14016 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14017 Yoram Zahavi <YoramZ@gilian.com>).
14021 * Add information about CygWin 1.3 and on, and preserve proper
14022 configuration for the versions before that.
14024 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14026 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14027 check whether we deal with a copy of a session and do not delete from
14028 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14029 <izhar@checkpoint.com>.
14033 * Do not store session data into the internal session cache, if it
14034 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14035 flag is set). Proposed by Aslam <aslam@funk.com>.
14039 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14044 * [In 0.9.6d-engine release:]
14045 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14047 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14049 * Add the configuration target linux-s390x.
14051 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14053 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14054 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14055 variable as an indication that a ClientHello message has been
14056 received. As the flag value will be lost between multiple
14057 invocations of ssl3_accept when using non-blocking I/O, the
14058 function may not be aware that a handshake has actually taken
14059 place, thus preventing a new session from being added to the
14062 To avoid this problem, we now set s->new_session to 2 instead of
14063 using a local variable.
14065 *Lutz Jaenicke, Bodo Moeller*
14067 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14068 if the SSL_R_LENGTH_MISMATCH error is detected.
14070 *Geoff Thorpe, Bodo Moeller*
14072 * New 'shared_ldflag' column in Configure platform table.
14076 * Fix EVP_CIPHER_mode macro.
14078 *"Dan S. Camper" <dan@bti.net>*
14080 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14081 type, we must throw them away by setting rr->length to 0.
14083 *D P Chang <dpc@qualys.com>*
14085 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14087 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14088 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14089 worked incorrectly for those cases where range = `10..._2` and
14090 `3*range` is two bits longer than range.)
14094 * Only add signing time to PKCS7 structures if it is not already
14099 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14100 OBJ_ld_ce should be OBJ_id_ce.
14101 Also some ip-pda OIDs in crypto/objects/objects.txt were
14102 incorrect (cf. RFC 3039).
14104 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14106 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14107 returns early because it has nothing to do.
14109 *Andy Schneider <andy.schneider@bjss.co.uk>*
14111 * [In 0.9.6c-engine release:]
14112 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14114 *Andy Schneider <andy.schneider@bjss.co.uk>*
14116 * [In 0.9.6c-engine release:]
14117 Add support for Cryptographic Appliance's keyserver technology.
14118 (Use engine 'keyclient')
14120 *Cryptographic Appliances and Geoff Thorpe*
14122 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14123 is called via tools/c89.sh because arguments have to be
14124 rearranged (all '-L' options must appear before the first object
14127 *Richard Shapiro <rshapiro@abinitio.com>*
14129 * [In 0.9.6c-engine release:]
14130 Add support for Broadcom crypto accelerator cards, backported
14133 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14135 * [In 0.9.6c-engine release:]
14136 Add support for SureWare crypto accelerator cards from
14137 Baltimore Technologies. (Use engine 'sureware')
14139 *Baltimore Technologies and Mark Cox*
14141 * [In 0.9.6c-engine release:]
14142 Add support for crypto accelerator cards from Accelerated
14143 Encryption Processing, www.aep.ie. (Use engine 'aep')
14145 *AEP Inc. and Mark Cox*
14147 * Add a configuration entry for gcc on UnixWare.
14149 *Gary Benson <gbenson@redhat.com>*
14151 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14152 messages are stored in a single piece (fixed-length part and
14153 variable-length part combined) and fix various bugs found on the way.
14157 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14158 instead. BIO_gethostbyname() does not know what timeouts are
14159 appropriate, so entries would stay in cache even when they have
14161 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14163 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14164 faced with a pathologically small ClientHello fragment that does
14165 not contain client_version: Instead of aborting with an error,
14166 simply choose the highest available protocol version (i.e.,
14167 TLS 1.0 unless it is disabled). In practice, ClientHello
14168 messages are never sent like this, but this change gives us
14169 strictly correct behaviour at least for TLS.
14173 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14174 never resets s->method to s->ctx->method when called from within
14175 one of the SSL handshake functions.
14177 *Bodo Moeller; problem pointed out by Niko Baric*
14179 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14180 (sent using the client's version number) if client_version is
14181 smaller than the protocol version in use. Also change
14182 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14183 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14184 the client will at least see that alert.
14188 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14193 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14194 client receives HelloRequest while in a handshake.
14196 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14198 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14199 should end in 'break', not 'goto end' which circumvents various
14200 cleanups done in state SSL_ST_OK. But session related stuff
14201 must be disabled for SSL_ST_OK in the case that we just sent a
14204 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14205 before just sending a HelloRequest.
14207 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14209 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14210 reveal whether illegal block cipher padding was found or a MAC
14211 verification error occurred. (Neither SSLerr() codes nor alerts
14212 are directly visible to potential attackers, but the information
14213 may leak via logfiles.)
14215 Similar changes are not required for the SSL 2.0 implementation
14216 because the number of padding bytes is sent in clear for SSL 2.0,
14217 and the extra bytes are just ignored. However ssl/s2_pkt.c
14218 failed to verify that the purported number of padding bytes is in
14223 * Add OpenUNIX-8 support including shared libraries
14224 (Boyd Lynn Gerber <gerberb@zenez.com>).
14228 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14229 'wristwatch attack' using huge encoding parameters (cf.
14230 James H. Manger's CRYPTO 2001 paper). Note that the
14231 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14232 encoding parameters and hence was not vulnerable.
14236 * BN_sqr() bug fix.
14238 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14240 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14241 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14242 followed by modular reduction.
14244 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14246 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14247 equivalent based on BN_pseudo_rand() instead of BN_rand().
14251 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14252 This function was broken, as the check for a new client hello message
14253 to handle SGC did not allow these large messages.
14254 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14258 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14262 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14263 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14267 * Rework the configuration and shared library support for Tru64 Unix.
14268 The configuration part makes use of modern compiler features and
14269 still retains old compiler behavior for those that run older versions
14270 of the OS. The shared library support part includes a variant that
14271 uses the RPATH feature, and is available through the special
14272 configuration target "alpha-cc-rpath", which will never be selected
14275 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14277 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14278 with the same message size as in ssl3_get_certificate_request().
14279 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14280 messages might inadvertently be reject as too long.
14282 *Petr Lampa <lampa@fee.vutbr.cz>*
14284 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14288 * Modified SSL library such that the verify_callback that has been set
14289 specifically for an SSL object with SSL_set_verify() is actually being
14290 used. Before the change, a verify_callback set with this function was
14291 ignored and the verify_callback() set in the SSL_CTX at the time of
14292 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14293 to allow the necessary settings.
14297 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14298 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14299 done automatically (in contradiction to the requirements of the C
14300 standard). This made problems when used from OpenSSH.
14304 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14305 dh->length and always used
14307 BN_rand_range(priv_key, dh->p).
14309 BN_rand_range() is not necessary for Diffie-Hellman, and this
14310 specific range makes Diffie-Hellman unnecessarily inefficient if
14311 dh->length (recommended exponent length) is much smaller than the
14312 length of dh->p. We could use BN_rand_range() if the order of
14313 the subgroup was stored in the DH structure, but we only have
14318 BN_rand(priv_key, l, ...)
14320 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14327 RSA_eay_public_encrypt
14328 RSA_eay_private_decrypt
14329 RSA_eay_private_encrypt (signing)
14330 RSA_eay_public_decrypt (signature verification)
14332 (default implementations for RSA_public_encrypt,
14333 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14334 always reject numbers >= n.
14338 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14339 to synchronize access to 'locking_thread'. This is necessary on
14340 systems where access to 'locking_thread' (an 'unsigned long'
14341 variable) is not atomic.
14345 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14346 *before* setting the 'crypto_lock_rand' flag. The previous code had
14347 a race condition if 0 is a valid thread ID.
14349 *Travis Vitek <vitek@roguewave.com>*
14351 * Add support for shared libraries under Irix.
14353 *Albert Chin-A-Young <china@thewrittenword.com>*
14355 * Add configuration option to build on Linux on both big-endian and
14356 little-endian MIPS.
14358 *Ralf Baechle <ralf@uni-koblenz.de>*
14360 * Add the possibility to create shared libraries on HP-UX.
14364 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14366 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14367 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14368 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14369 PRNG state recovery was possible based on the output of
14370 one PRNG request appropriately sized to gain knowledge on
14371 'md' followed by enough consecutive 1-byte PRNG requests
14372 to traverse all of 'state'.
14374 1. When updating 'md_local' (the current thread's copy of 'md')
14375 during PRNG output generation, hash all of the previous
14376 'md_local' value, not just the half used for PRNG output.
14378 2. Make the number of bytes from 'state' included into the hash
14379 independent from the number of PRNG bytes requested.
14381 The first measure alone would be sufficient to avoid
14382 Markku-Juhani's attack. (Actually it had never occurred
14383 to me that the half of 'md_local' used for chaining was the
14384 half from which PRNG output bytes were taken -- I had always
14385 assumed that the secret half would be used.) The second
14386 measure makes sure that additional data from 'state' is never
14387 mixed into 'md_local' in small portions; this heuristically
14388 further strengthens the PRNG.
14392 * Fix crypto/bn/asm/mips3.s.
14396 * When only the key is given to "enc", the IV is undefined. Print out
14397 an error message in this case.
14401 * Handle special case when X509_NAME is empty in X509 printing routines.
14405 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14406 positive and less than q.
14410 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14411 used: it isn't thread safe and the add_lock_callback should handle
14414 *Paul Rose <Paul.Rose@bridge.com>*
14416 * Verify that incoming data obeys the block size in
14417 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14423 *Ulf Möller, Bodo Möller*
14425 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14426 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14427 when fixing the server behaviour for backwards-compatible 'client
14428 hello' messages. (Note that the attack is impractical against
14429 SSL 3.0 and TLS 1.0 anyway because length and version checking
14430 means that the probability of guessing a valid ciphertext is
14431 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14434 Before 0.9.5, the countermeasure (hide the error by generating a
14435 random 'decryption result') did not work properly because
14436 ERR_clear_error() was missing, meaning that SSL_get_error() would
14437 detect the supposedly ignored error.
14439 Both problems are now fixed.
14443 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14444 (previously it was 1024).
14448 * Fix for compatibility mode trust settings: ignore trust settings
14449 unless some valid trust or reject settings are present.
14453 * Fix for blowfish EVP: its a variable length cipher.
14457 * Fix various bugs related to DSA S/MIME verification. Handle missing
14458 parameters in DSA public key structures and return an error in the
14459 DSA routines if parameters are absent.
14463 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14464 in the current directory if neither $RANDFILE nor $HOME was set.
14465 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14466 caused some confusion to Windows users who haven't defined $HOME.
14467 Thus RAND_file_name() is changed again: e_os.h can define a
14468 DEFAULT_HOME, which will be used if $HOME is not set.
14469 For Windows, we use "C:"; on other platforms, we still require
14470 environment variables.
14472 * Move 'if (!initialized) RAND_poll()' into regions protected by
14473 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14474 having multiple threads call RAND_poll() concurrently.
14478 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14479 combination of a flag and a thread ID variable.
14480 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14481 flag), *other* threads can enter ssleay_add_bytes without obeying
14482 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14483 that they do not hold after the first thread unsets add_do_not_lock).
14487 * Change bctest again: '-x' expressions are not available in all
14488 versions of 'test'.
14492 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14494 * Fix a couple of memory leaks in PKCS7_dataDecode()
14496 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14498 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14499 the default extension for executables, if any. Also, make the perl
14500 scripts that use symlink() to test if it really exists and use "cp"
14501 if it doesn't. All this made OpenSSL compilable and installable in
14506 * Fix for asn1_GetSequence() for indefinite length constructed data.
14507 If SEQUENCE is length is indefinite just set c->slen to the total
14508 amount of data available.
14510 *Steve Henson, reported by shige@FreeBSD.org*
14512 *This change does not apply to 0.9.7.*
14514 * Change bctest to avoid here-documents inside command substitution
14515 (workaround for FreeBSD /bin/sh bug).
14516 For compatibility with Ultrix, avoid shell functions (introduced
14517 in the bctest version that searches along $PATH).
14521 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14522 with des_encrypt() defined on some operating systems, like Solaris
14527 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14528 On the Importance of Eliminating Errors in Cryptographic
14529 Computations, J. Cryptology 14 (2001) 2, 101-119,
14530 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14534 * MIPS assembler BIGNUM division bug fix.
14538 * Disabled incorrect Alpha assembler code.
14542 * Fix PKCS#7 decode routines so they correctly update the length
14543 after reading an EOC for the EXPLICIT tag.
14547 *This change does not apply to 0.9.7.*
14549 * Fix bug in PKCS#12 key generation routines. This was triggered
14550 if a 3DES key was generated with a 0 initial byte. Include
14551 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14552 (but broken) behaviour.
14556 * Enhance bctest to search for a working bc along $PATH and print
14559 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14561 * Fix memory leaks in err.c: free err_data string if necessary;
14562 don't write to the wrong index in ERR_set_error_data.
14566 * Implement ssl23_peek (analogous to ssl23_read), which previously
14571 * Replace rdtsc with `_emit` statements for VC++ version 5.
14573 *Jeremy Cooper <jeremy@baymoo.org>*
14575 * Make it possible to reuse SSLv2 sessions.
14579 * In copy_email() check for >= 0 as a return value for
14580 X509_NAME_get_index_by_NID() since 0 is a valid index.
14582 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14584 * Avoid coredump with unsupported or invalid public keys by checking if
14585 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14586 PKCS7_verify() fails with non detached data.
14590 * Don't use getenv in library functions when run as setuid/setgid.
14591 New function OPENSSL_issetugid().
14595 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14596 due to incorrect handling of multi-threading:
14598 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14600 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14602 3. Count how many times MemCheck_off() has been called so that
14603 nested use can be treated correctly. This also avoids
14604 inband-signalling in the previous code (which relied on the
14605 assumption that thread ID 0 is impossible).
14609 * Add "-rand" option also to s_client and s_server.
14613 * Fix CPU detection on Irix 6.x.
14614 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14615 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14617 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14622 *This change does not apply to 0.9.7.*
14624 * Use the cached encoding of an X509_NAME structure rather than
14625 copying it. This is apparently the reason for the libsafe "errors"
14626 but the code is actually correct.
14630 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14631 Bleichenbacher's DSA attack.
14632 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14633 to be set and top=0 forces the highest bit to be set; top=-1 is new
14634 and leaves the highest bit random.
14636 *Ulf Moeller, Bodo Moeller*
14638 * In the `NCONF_...`-based implementations for `CONF_...` queries
14639 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14640 a temporary CONF structure with the data component set to NULL
14641 (which gives segmentation faults in lh_retrieve).
14642 Instead, use NULL for the CONF pointer in CONF_get_string and
14643 CONF_get_number (which may use environment variables) and directly
14644 return NULL from CONF_get_section.
14648 * Fix potential buffer overrun for EBCDIC.
14652 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14653 keyUsage if basicConstraints absent for a CA.
14657 * Make SMIME_write_PKCS7() write mail header values with a format that
14658 is more generally accepted (no spaces before the semicolon), since
14659 some programs can't parse those values properly otherwise. Also make
14660 sure BIO's that break lines after each write do not create invalid
14665 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14666 macros previously used would not encode an empty SEQUENCE OF
14667 and break the signature.
14671 *This change does not apply to 0.9.7.*
14673 * Zero the premaster secret after deriving the master secret in
14678 * Add some EVP_add_digest_alias registrations (as found in
14679 OpenSSL_add_all_digests()) to SSL_library_init()
14680 aka OpenSSL_add_ssl_algorithms(). This provides improved
14681 compatibility with peers using X.509 certificates
14682 with unconventional AlgorithmIdentifier OIDs.
14686 * Fix for Irix with NO_ASM.
14688 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14690 * ./config script fixes.
14692 *Ulf Moeller, Richard Levitte*
14694 * Fix 'openssl passwd -1'.
14698 * Change PKCS12_key_gen_asc() so it can cope with non null
14699 terminated strings whose length is passed in the passlen
14700 parameter, for example from PEM callbacks. This was done
14701 by adding an extra length parameter to asc2uni().
14703 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14705 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14706 call failed, free the DSA structure.
14710 * Fix to uni2asc() to cope with zero length Unicode strings.
14711 These are present in some PKCS#12 files.
14715 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14716 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14717 when writing a 32767 byte record.
14719 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14721 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14722 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14724 (RSA objects have a reference count access to which is protected
14725 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14726 so they are meant to be shared between threads.)
14727 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14728 "Reddie, Steven" <Steven.Reddie@ca.com>*
14730 * Fix a deadlock in CRYPTO_mem_leaks().
14734 * Use better test patterns in bntest.
14738 * rand_win.c fix for Borland C.
14742 * BN_rshift bugfix for n == 0.
14746 * Add a 'bctest' script that checks for some known 'bc' bugs
14747 so that 'make test' does not abort just because 'bc' is broken.
14751 * Store verify_result within SSL_SESSION also for client side to
14752 avoid potential security hole. (Re-used sessions on the client side
14753 always resulted in verify_result==X509_V_OK, not using the original
14754 result of the server certificate verification.)
14758 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14759 SSL3_RT_APPLICATION_DATA, return 0.
14760 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14765 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14766 releases, have been re-implemented by renaming the previous
14767 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14768 and ssl3_read_internal, respectively, and adding 'peek' parameters
14769 to them. The new ssl[23]_{read,peek} functions are calls to
14770 ssl[23]_read_internal with the 'peek' flag set appropriately.
14771 A 'peek' parameter has also been added to ssl3_read_bytes, which
14772 does the actual work for ssl3_read_internal.
14776 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14777 the method-specific "init()" handler. Also clean up ex_data after
14778 calling the method-specific "finish()" handler. Previously, this was
14779 happening the other way round.
14783 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14784 The previous value, 12, was not always sufficient for BN_mod_exp().
14788 * Make sure that shared libraries get the internal name engine with
14789 the full version number and not just 0. This should mark the
14790 shared libraries as not backward compatible. Of course, this should
14791 be changed again when we can guarantee backward binary compatibility.
14795 * Fix typo in get_cert_by_subject() in by_dir.c
14797 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14799 * Rework the system to generate shared libraries:
14801 - Make note of the expected extension for the shared libraries and
14802 if there is a need for symbolic links from for example libcrypto.so.0
14803 to libcrypto.so.0.9.7. There is extended info in Configure for
14806 - Make as few rebuilds of the shared libraries as possible.
14808 - Still avoid linking the OpenSSL programs with the shared libraries.
14810 - When installing, install the shared libraries separately from the
14815 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14817 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14818 and not in SSL_clear because the latter is also used by the
14819 accept/connect functions; previously, the settings made by
14820 SSL_set_read_ahead would be lost during the handshake.
14822 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14824 * Correct util/mkdef.pl to be selective about disabled algorithms.
14825 Previously, it would create entries for disabled algorithms no
14830 * Added several new manual pages for SSL_* function.
14834 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14836 * In ssl23_get_client_hello, generate an error message when faced
14837 with an initial SSL 3.0/TLS record that is too small to contain the
14838 first two bytes of the ClientHello message, i.e. client_version.
14839 (Note that this is a pathologic case that probably has never happened
14840 in real life.) The previous approach was to use the version number
14841 from the record header as a substitute; but our protocol choice
14842 should not depend on that one because it is not authenticated
14843 by the Finished messages.
14847 * More robust randomness gathering functions for Windows.
14849 *Jeffrey Altman <jaltman@columbia.edu>*
14851 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14852 not set then we don't setup the error code for issuer check errors
14853 to avoid possibly overwriting other errors which the callback does
14854 handle. If an application does set the flag then we assume it knows
14855 what it is doing and can handle the new informational codes
14860 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14861 a general "ANY" type, as such it should be able to decode anything
14862 including tagged types. However it didn't check the class so it would
14863 wrongly interpret tagged types in the same way as their universal
14864 counterpart and unknown types were just rejected. Changed so that the
14865 tagged and unknown types are handled in the same way as a SEQUENCE:
14866 that is the encoding is stored intact. There is also a new type
14867 "V_ASN1_OTHER" which is used when the class is not universal, in this
14868 case we have no idea what the actual type is so we just lump them all
14873 * On VMS, stdout may very well lead to a file that is written to
14874 in a record-oriented fashion. That means that every write() will
14875 write a separate record, which will be read separately by the
14876 programs trying to read from it. This can be very confusing.
14878 The solution is to put a BIO filter in the way that will buffer
14879 text until a linefeed is reached, and then write everything a
14880 line at a time, so every record written will be an actual line,
14881 not chunks of lines and not (usually doesn't happen, but I've
14882 seen it once) several lines in one record. BIO_f_linebuffer() is
14885 Currently, it's a VMS-only method, because that's where it has
14886 been tested well enough.
14890 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14891 it can return incorrect results.
14892 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14893 but it was in 0.9.6-beta[12].)
14897 * Disable the check for content being present when verifying detached
14898 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14899 include zero length content when signing messages.
14903 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14904 BIO_ctrl (for BIO pairs).
14908 * Add DSO method for VMS.
14912 * Bug fix: Montgomery multiplication could produce results with the
14917 * Add RPM specification openssl.spec and modify it to build three
14918 packages. The default package contains applications, application
14919 documentation and run-time libraries. The devel package contains
14920 include files, static libraries and function documentation. The
14921 doc package contains the contents of the doc directory. The original
14922 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14926 * Add a large number of documentation files for many SSL routines.
14928 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14930 * Add a configuration entry for Sony News 4.
14932 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14934 * Don't set the two most significant bits to one when generating a
14935 random number < q in the DSA library.
14939 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14940 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14941 the underlying transport is blocking) if a handshake took place.
14942 (The default behaviour is needed by applications such as s_client
14943 and s_server that use select() to determine when to use SSL_read;
14944 but for applications that know in advance when to expect data, it
14945 just makes things more complicated.)
14949 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14954 * Add a few more EBCDIC conditionals that make `req` and `x509`
14955 work better on such systems.
14957 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14959 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14960 Update PKCS12_parse() so it copies the friendlyName and the
14961 keyid to the certificates aux info.
14965 * Fix bug in PKCS7_verify() which caused an infinite loop
14966 if there was more than one signature.
14968 *Sven Uszpelkat <su@celocom.de>*
14970 * Major change in util/mkdef.pl to include extra information
14971 about each symbol, as well as presenting variables as well
14972 as functions. This change means that there's n more need
14973 to rebuild the .num files when some algorithms are excluded.
14977 * Allow the verify time to be set by an application,
14978 rather than always using the current time.
14982 * Phase 2 verify code reorganisation. The certificate
14983 verify code now looks up an issuer certificate by a
14984 number of criteria: subject name, authority key id
14985 and key usage. It also verifies self signed certificates
14986 by the same criteria. The main comparison function is
14987 X509_check_issued() which performs these checks.
14989 Lot of changes were necessary in order to support this
14990 without completely rewriting the lookup code.
14992 Authority and subject key identifier are now cached.
14994 The LHASH 'certs' is X509_STORE has now been replaced
14995 by a STACK_OF(X509_OBJECT). This is mainly because an
14996 LHASH can't store or retrieve multiple objects with
14997 the same hash value.
14999 As a result various functions (which were all internal
15000 use only) have changed to handle the new X509_STORE
15001 structure. This will break anything that messed round
15002 with X509_STORE internally.
15004 The functions X509_STORE_add_cert() now checks for an
15005 exact match, rather than just subject name.
15007 The X509_STORE API doesn't directly support the retrieval
15008 of multiple certificates matching a given criteria, however
15009 this can be worked round by performing a lookup first
15010 (which will fill the cache with candidate certificates)
15011 and then examining the cache for matches. This is probably
15012 the best we can do without throwing out X509_LOOKUP
15013 entirely (maybe later...).
15015 The X509_VERIFY_CTX structure has been enhanced considerably.
15017 All certificate lookup operations now go via a get_issuer()
15018 callback. Although this currently uses an X509_STORE it
15019 can be replaced by custom lookups. This is a simple way
15020 to bypass the X509_STORE hackery necessary to make this
15021 work and makes it possible to use more efficient techniques
15022 in future. A very simple version which uses a simple
15023 STACK for its trusted certificate store is also provided
15024 using X509_STORE_CTX_trusted_stack().
15026 The verify_cb() and verify() callbacks now have equivalents
15027 in the X509_STORE_CTX structure.
15029 X509_STORE_CTX also has a 'flags' field which can be used
15030 to customise the verify behaviour.
15034 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15035 excludes S/MIME capabilities.
15039 * When a certificate request is read in keep a copy of the
15040 original encoding of the signed data and use it when outputting
15041 again. Signatures then use the original encoding rather than
15042 a decoded, encoded version which may cause problems if the
15043 request is improperly encoded.
15047 * For consistency with other BIO_puts implementations, call
15048 buffer_write(b, ...) directly in buffer_puts instead of calling
15051 In BIO_puts, increment b->num_write as in BIO_write.
15053 *Peter.Sylvester@EdelWeb.fr*
15055 * Fix BN_mul_word for the case where the word is 0. (We have to use
15056 BN_zero, we may not return a BIGNUM with an array consisting of
15057 words set to zero.)
15061 * Avoid calling abort() from within the library when problems are
15062 detected, except if preprocessor symbols have been defined
15063 (such as REF_CHECK, BN_DEBUG etc.).
15067 * New openssl application 'rsautl'. This utility can be
15068 used for low level RSA operations. DER public key
15069 BIO/fp routines also added.
15073 * New Configure entry and patches for compiling on QNX 4.
15075 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15077 * A demo state-machine implementation was sponsored by
15078 Nuron (<http://www.nuron.com/>) and is now available in
15079 demos/state_machine.
15083 * New options added to the 'dgst' utility for signature
15084 generation and verification.
15088 * Unrecognized PKCS#7 content types are now handled via a
15089 catch all ASN1_TYPE structure. This allows unsupported
15090 types to be stored as a "blob" and an application can
15091 encode and decode it manually.
15095 * Fix various signed/unsigned issues to make a_strex.c
15096 compile under VC++.
15098 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15100 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15101 length if passed a buffer. ASN1_INTEGER_to_BN failed
15102 if passed a NULL BN and its argument was negative.
15104 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15106 * Modification to PKCS#7 encoding routines to output definite
15107 length encoding. Since currently the whole structures are in
15108 memory there's not real point in using indefinite length
15109 constructed encoding. However if OpenSSL is compiled with
15110 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15114 * Added BIO_vprintf() and BIO_vsnprintf().
15118 * Added more prefixes to parse for in the strings written
15119 through a logging bio, to cover all the levels that are available
15120 through syslog. The prefixes are now:
15122 PANIC, EMERG, EMR => LOG_EMERG
15123 ALERT, ALR => LOG_ALERT
15124 CRIT, CRI => LOG_CRIT
15125 ERROR, ERR => LOG_ERR
15126 WARNING, WARN, WAR => LOG_WARNING
15127 NOTICE, NOTE, NOT => LOG_NOTICE
15128 INFO, INF => LOG_INFO
15129 DEBUG, DBG => LOG_DEBUG
15131 and as before, if none of those prefixes are present at the
15132 beginning of the string, LOG_ERR is chosen.
15134 On Win32, the `LOG_*` levels are mapped according to this:
15136 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15137 LOG_WARNING => EVENTLOG_WARNING_TYPE
15138 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15142 * Made it possible to reconfigure with just the configuration
15143 argument "reconf" or "reconfigure". The command line arguments
15144 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15145 and are retrieved from there when reconfiguring.
15151 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15153 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15157 * The obj_dat.pl script was messing up the sorting of object
15158 names. The reason was that it compared the quoted version
15159 of strings as a result "OCSP" > "OCSP Signing" because
15160 " > SPACE. Changed script to store unquoted versions of
15161 names and add quotes on output. It was also omitting some
15162 names from the lookup table if they were given a default
15163 value (that is if SN is missing it is given the same
15164 value as LN and vice versa), these are now added on the
15165 grounds that if an object has a name we should be able to
15166 look it up. Finally added warning output when duplicate
15167 short or long names are found.
15171 * Changes needed for Tandem NSK.
15173 *Scott Uroff <scott@xypro.com>*
15175 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15176 RSA_padding_check_SSLv23(), special padding was never detected
15177 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15178 version rollback attacks was not effective.
15180 In s23_clnt.c, don't use special rollback-attack detection padding
15181 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15182 client; similarly, in s23_srvr.c, don't do the rollback check if
15183 SSL 2.0 is the only protocol enabled in the server.
15187 * Make it possible to get hexdumps of unprintable data with 'openssl
15188 asn1parse'. By implication, the functions ASN1_parse_dump() and
15189 BIO_dump_indent() are added.
15193 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15194 these print out strings and name structures based on various
15195 flags including RFC2253 support and proper handling of
15196 multibyte characters. Added options to the 'x509' utility
15197 to allow the various flags to be set.
15201 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15202 Also change the functions X509_cmp_current_time() and
15203 X509_gmtime_adj() work with an ASN1_TIME structure,
15204 this will enable certificates using GeneralizedTime in validity
15205 dates to be checked.
15209 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15210 negative public key encodings) on by default,
15211 NO_NEG_PUBKEY_BUG can be set to disable it.
15215 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15216 content octets. An i2c_ASN1_OBJECT is unnecessary because
15217 the encoding can be trivially obtained from the structure.
15221 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15222 not read locks (`CRYPTO_r_[un]lock`).
15226 * A first attempt at creating official support for shared
15227 libraries through configuration. I've kept it so the
15228 default is static libraries only, and the OpenSSL programs
15229 are always statically linked for now, but there are
15230 preparations for dynamic linking in place.
15231 This has been tested on Linux and Tru64.
15235 * Randomness polling function for Win9x, as described in:
15236 Peter Gutmann, Software Generation of Practically Strong
15241 * Fix so PRNG is seeded in req if using an already existing
15246 * New options to smime application. -inform and -outform
15247 allow alternative formats for the S/MIME message including
15248 PEM and DER. The -content option allows the content to be
15249 specified separately. This should allow things like Netscape
15250 form signing output easier to verify.
15254 * Fix the ASN1 encoding of tags using the 'long form'.
15258 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15259 STRING types. These convert content octets to and from the
15260 underlying type. The actual tag and length octets are
15261 already assumed to have been read in and checked. These
15262 are needed because all other string types have virtually
15263 identical handling apart from the tag. By having versions
15264 of the ASN1 functions that just operate on content octets
15265 IMPLICIT tagging can be handled properly. It also allows
15266 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15267 and ASN1_INTEGER are identical apart from the tag.
15271 * Change the handling of OID objects as follows:
15273 - New object identifiers are inserted in objects.txt, following
15274 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15275 - objects.pl is used to process obj_mac.num and create a new
15277 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15280 This is currently kind of a hack, and the perl code in objects.pl
15281 isn't very elegant, but it works as I intended. The simplest way
15282 to check that it worked correctly is to look in obj_dat.h and
15283 check the array nid_objs and make sure the objects haven't moved
15284 around (this is important!). Additions are OK, as well as
15285 consistent name changes.
15289 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15293 * Addition of the command line parameter '-rand file' to 'openssl req'.
15294 The given file adds to whatever has already been seeded into the
15295 random pool through the RANDFILE configuration file option or
15296 environment variable, or the default random state file.
15300 * mkstack.pl now sorts each macro group into lexical order.
15301 Previously the output order depended on the order the files
15302 appeared in the directory, resulting in needless rewriting
15307 * Patches to make OpenSSL compile under Win32 again. Mostly
15308 work arounds for the VC++ problem that it treats func() as
15309 func(void). Also stripped out the parts of mkdef.pl that
15310 added extra typesafe functions: these no longer exist.
15314 * Reorganisation of the stack code. The macros are now all
15315 collected in safestack.h . Each macro is defined in terms of
15316 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15317 DEBUG_SAFESTACK is now handled in terms of function casts,
15318 this has the advantage of retaining type safety without the
15319 use of additional functions. If DEBUG_SAFESTACK is not defined
15320 then the non typesafe macros are used instead. Also modified the
15321 mkstack.pl script to handle the new form. Needs testing to see
15322 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15323 the default if no major problems. Similar behaviour for ASN1_SET_OF
15324 and PKCS12_STACK_OF.
15328 * When some versions of IIS use the 'NET' form of private key the
15329 key derivation algorithm is different. Normally MD5(password) is
15330 used as a 128 bit RC4 key. In the modified case
15331 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15332 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15333 as the old Netscape_RSA functions except they have an additional
15334 'sgckey' parameter which uses the modified algorithm. Also added
15335 an -sgckey command line option to the rsa utility. Thanks to
15336 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15337 algorithm to openssl-dev.
15341 * The evp_local.h macros were using 'c.##kname' which resulted in
15342 invalid expansion on some systems (SCO 5.0.5 for example).
15343 Corrected to 'c.kname'.
15345 *Phillip Porch <root@theporch.com>*
15347 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15348 a STACK of email addresses from a certificate or request, these look
15349 in the subject name and the subject alternative name extensions and
15350 omit any duplicate addresses.
15354 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15355 This makes DSA verification about 2 % faster.
15359 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15360 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15361 plus overhead for 1024 bit moduli).
15362 This makes exponentiations about 0.5 % faster for 1024 bit
15363 exponents (as measured by "openssl speed rsa2048").
15367 * Rename memory handling macros to avoid conflicts with other
15369 Malloc => OPENSSL_malloc
15370 Malloc_locked => OPENSSL_malloc_locked
15371 Realloc => OPENSSL_realloc
15372 Free => OPENSSL_free
15376 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15377 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15381 * CygWin32 support.
15383 *John Jarvie <jjarvie@newsguy.com>*
15385 * The type-safe stack code has been rejigged. It is now only compiled
15386 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15387 by default all type-specific stack functions are "#define"d back to
15388 standard stack functions. This results in more streamlined output
15389 but retains the type-safety checking possibilities of the original
15394 * The STACK code has been cleaned up, and certain type declarations
15395 that didn't make a lot of sense have been brought in line. This has
15396 also involved a cleanup of sorts in safestack.h to more correctly
15397 map type-safe stack functions onto their plain stack counterparts.
15398 This work has also resulted in a variety of "const"ifications of
15399 lots of the code, especially `_cmp` operations which should normally
15400 be prototyped with "const" parameters anyway.
15404 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15405 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15406 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15407 where all of 'md' is used each time the PRNG is used, but 'state'
15408 is used only indexed by a cyclic counter. As entropy may not be
15409 well distributed from the beginning, 'md' is important as a
15410 chaining variable. However, the output function chains only half
15411 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15412 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15413 in all of 'state' being rewritten, with the new values depending
15414 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15418 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15419 the handshake is continued after ssl_verify_cert_chain();
15420 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15421 can lead to 'unexplainable' connection aborts later.
15423 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15425 * Major EVP API cipher revision.
15426 Add hooks for extra EVP features. This allows various cipher
15427 parameters to be set in the EVP interface. Support added for variable
15428 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15429 setting of RC2 and RC5 parameters.
15431 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15434 Remove lots of duplicated code from the EVP library. For example *every*
15435 cipher init() function handles the 'iv' in the same way according to the
15436 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15437 for CFB and OFB modes they zero ctx->num.
15439 New functionality allows removal of S/MIME code RC2 hack.
15441 Most of the routines have the same form and so can be declared in terms
15444 By shifting this to the top level EVP_CipherInit() it can be removed from
15445 all individual ciphers. If the cipher wants to handle IVs or keys
15446 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15449 Change lots of functions like EVP_EncryptUpdate() to now return a
15450 value: although software versions of the algorithms cannot fail
15451 any installed hardware versions can.
15455 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15456 this option is set, tolerate broken clients that send the negotiated
15457 protocol version number instead of the requested protocol version
15462 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15463 i.e. non-zero for export ciphersuites, zero otherwise.
15464 Previous versions had this flag inverted, inconsistent with
15465 rsa_tmp_cb (..._TMP_RSA_CB).
15467 *Bodo Moeller; problem reported by Amit Chopra*
15469 * Add missing DSA library text string. Work around for some IIS
15470 key files with invalid SEQUENCE encoding.
15474 * Add a document (doc/standards.txt) that list all kinds of standards
15475 and so on that are implemented in OpenSSL.
15479 * Enhance c_rehash script. Old version would mishandle certificates
15480 with the same subject name hash and wouldn't handle CRLs at all.
15481 Added -fingerprint option to crl utility, to support new c_rehash
15486 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15490 * Fix for SSL server purpose checking. Server checking was
15491 rejecting certificates which had extended key usage present
15492 but no ssl client purpose.
15494 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15496 * Make PKCS#12 code work with no password. The PKCS#12 spec
15497 is a little unclear about how a blank password is handled.
15498 Since the password in encoded as a BMPString with terminating
15499 double NULL a zero length password would end up as just the
15500 double NULL. However no password at all is different and is
15501 handled differently in the PKCS#12 key generation code. NS
15502 treats a blank password as zero length. MSIE treats it as no
15503 password on export: but it will try both on import. We now do
15504 the same: PKCS12_parse() tries zero length and no password if
15505 the password is set to "" or NULL (NULL is now a valid password:
15506 it wasn't before) as does the pkcs12 application.
15510 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15511 perror when PEM_read_bio_X509_REQ fails, the error message must
15512 be obtained from the error queue.
15516 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15517 it in ERR_remove_state if appropriate, and change ERR_get_state
15518 accordingly to avoid race conditions (this is necessary because
15519 thread_hash is no longer constant once set).
15523 * Bugfix for linux-elf makefile.one.
15527 * RSA_get_default_method() will now cause a default
15528 RSA_METHOD to be chosen if one doesn't exist already.
15529 Previously this was only set during a call to RSA_new()
15530 or RSA_new_method(NULL) meaning it was possible for
15531 RSA_get_default_method() to return NULL.
15535 * Added native name translation to the existing DSO code
15536 that will convert (if the flag to do so is set) filenames
15537 that are sufficiently small and have no path information
15538 into a canonical native form. Eg. "blah" converted to
15539 "libblah.so" or "blah.dll" etc.
15543 * New function ERR_error_string_n(e, buf, len) which is like
15544 ERR_error_string(e, buf), but writes at most 'len' bytes
15545 including the 0 terminator. For ERR_error_string_n, 'buf'
15548 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15550 * CONF library reworked to become more general. A new CONF
15551 configuration file reader "class" is implemented as well as a
15552 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15553 old `CONF_*` functions are still there, but are reimplemented to
15554 work in terms of the new functions. Also, a set of functions
15555 to handle the internal storage of the configuration data is
15556 provided to make it easier to write new configuration file
15557 reader "classes" (I can definitely see something reading a
15558 configuration file in XML format, for example), called `_CONF_*`,
15559 or "the configuration storage API"...
15561 The new configuration file reading functions are:
15563 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15564 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15566 NCONF_default, NCONF_WIN32
15568 NCONF_dump_fp, NCONF_dump_bio
15570 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15571 NCONF_new creates a new CONF object. This works in the same way
15572 as other interfaces in OpenSSL, like the BIO interface.
15573 `NCONF_dump_*` dump the internal storage of the configuration file,
15574 which is useful for debugging. All other functions take the same
15575 arguments as the old `CONF_*` functions with the exception of the
15576 first that must be a `CONF *` instead of a `LHASH *`.
15578 To make it easier to use the new classes with the old `CONF_*` functions,
15579 the function CONF_set_default_method is provided.
15583 * Add '-tls1' option to 'openssl ciphers', which was already
15584 mentioned in the documentation but had not been implemented.
15585 (This option is not yet really useful because even the additional
15586 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15590 * Initial DSO code added into libcrypto for letting OpenSSL (and
15591 OpenSSL-based applications) load shared libraries and bind to
15592 them in a portable way.
15594 *Geoff Thorpe, with contributions from Richard Levitte*
15596 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15598 * Make sure _lrotl and _lrotr are only used with MSVC.
15600 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15601 (the default implementation of RAND_status).
15603 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15604 to '-clrext' (= clear extensions), as intended and documented.
15605 *Bodo Moeller; inconsistency pointed out by Michael Attili
15606 <attili@amaxo.com>*
15608 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15609 was larger than the MD block size.
15611 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15613 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15614 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15615 using the passed key: if the passed key was a private key the result
15616 of X509_print(), for example, would be to print out all the private key
15621 * des_quad_cksum() byte order bug fix.
15622 *Ulf Möller, using the problem description in krb4-0.9.7, where
15623 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15625 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15628 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15630 * For easily testing in shell scripts whether some command
15631 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15632 returns with exit code 0 iff no command of the given name is available.
15633 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15634 the output goes to stdout and nothing is printed to stderr.
15635 Additional arguments are always ignored.
15637 Since for each cipher there is a command of the same name,
15638 the 'no-cipher' compilation switches can be tested this way.
15640 ('openssl no-XXX' is not able to detect pseudo-commands such
15641 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15645 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15649 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15650 is set; it will be thrown away anyway because each handshake creates
15652 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15653 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15654 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15655 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15659 * New s_client option -ign_eof: EOF at stdin is ignored, and
15660 'Q' and 'R' lose their special meanings (quit/renegotiate).
15661 This is part of what -quiet does; unlike -quiet, -ign_eof
15662 does not suppress any output.
15666 * Add compatibility options to the purpose and trust code. The
15667 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15668 accepts a certificate or CA, this was the previous behaviour,
15669 with all the associated security issues.
15671 X509_TRUST_COMPAT is the old trust behaviour: only and
15672 automatically trust self signed roots in certificate store. A
15673 new trust setting X509_TRUST_DEFAULT is used to specify that
15674 a purpose has no associated trust setting and it should instead
15675 use the value in the default purpose.
15679 * Fix the PKCS#8 DSA private key code so it decodes keys again
15680 and fix a memory leak.
15684 * In util/mkerr.pl (which implements 'make errors'), preserve
15685 reason strings from the previous version of the .c file, as
15686 the default to have only downcase letters (and digits) in
15687 automatically generated reasons codes is not always appropriate.
15691 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15692 using strerror. Previously, ERR_reason_error_string() returned
15693 library names as reason strings for SYSerr; but SYSerr is a special
15694 case where small numbers are errno values, not library numbers.
15698 * Add '-dsaparam' option to 'openssl dhparam' application. This
15699 converts DSA parameters into DH parameters. (When creating parameters,
15700 DSA_generate_parameters is used.)
15704 * Include 'length' (recommended exponent length) in C code generated
15705 by 'openssl dhparam -C'.
15709 * The second argument to set_label in perlasm was already being used
15710 so couldn't be used as a "file scope" flag. Moved to third argument
15715 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15716 instead of RAND_bytes for encryption IVs and salts.
15720 * Include RAND_status() into RAND_METHOD instead of implementing
15721 it only for md_rand.c Otherwise replacing the PRNG by calling
15722 RAND_set_rand_method would be impossible.
15726 * Don't let DSA_generate_key() enter an infinite loop if the random
15727 number generation fails.
15731 * New 'rand' application for creating pseudo-random output.
15735 * Added configuration support for Linux/IA64
15737 *Rolf Haberrecker <rolf@suse.de>*
15739 * Assembler module support for Mingw32.
15743 * Shared library support for HPUX (in shlib/).
15745 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15747 * Shared library support for Solaris gcc.
15749 *Lutz Behnke <behnke@trustcenter.de>*
15751 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15753 * PKCS7_encrypt() was adding text MIME headers twice because they
15754 were added manually and by SMIME_crlf_copy().
15758 * In bntest.c don't call BN_rand with zero bits argument.
15760 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15762 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15763 case was implemented. This caused BN_div_recp() to fail occasionally.
15767 * Add an optional second argument to the set_label() in the perl
15768 assembly language builder. If this argument exists and is set
15769 to 1 it signals that the assembler should use a symbol whose
15770 scope is the entire file, not just the current function. This
15771 is needed with MASM which uses the format label:: for this scope.
15773 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15775 * Change the ASN1 types so they are typedefs by default. Before
15776 almost all types were #define'd to ASN1_STRING which was causing
15777 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15782 * Change names of new functions to the new get1/get0 naming
15783 convention: After 'get1', the caller owns a reference count
15784 and has to call `..._free`; 'get0' returns a pointer to some
15785 data structure without incrementing reference counters.
15786 (Some of the existing 'get' functions increment a reference
15787 counter, some don't.)
15788 Similarly, 'set1' and 'add1' functions increase reference
15789 counters or duplicate objects.
15793 * Allow for the possibility of temp RSA key generation failure:
15794 the code used to assume it always worked and crashed on failure.
15798 * Fix potential buffer overrun problem in BIO_printf().
15799 *Ulf Möller, using public domain code by Patrick Powell; problem
15800 pointed out by David Sacerdote <das33@cornell.edu>*
15802 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15803 RAND_egd() and RAND_status(). In the command line application,
15804 the EGD socket can be specified like a seed file using RANDFILE
15809 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15810 Some CAs (e.g. Verisign) distribute certificates in this form.
15814 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15815 list to exclude them. This means that no special compilation option
15816 is needed to use anonymous DH: it just needs to be included in the
15821 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15822 EVP_MD_type. The old functionality is available in a new macro called
15823 EVP_MD_md(). Change code that uses it and update docs.
15827 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15828 where the `void *` argument is replaced by a function pointer argument.
15829 Previously `void *` was abused to point to functions, which works on
15830 many platforms, but is not correct. As these functions are usually
15831 called by macros defined in OpenSSL header files, most source code
15832 should work without changes.
15836 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15837 sections with information on -D... compiler switches used for
15838 compiling the library so that applications can see them. To enable
15839 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15840 must be defined. E.g.,
15841 #define OPENSSL_ALGORITHM_DEFINES
15842 #include <openssl/opensslconf.h>
15843 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15845 *Richard Levitte, Ulf and Bodo Möller*
15847 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15852 * Change the 'other' type in certificate aux info to a STACK_OF
15853 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15854 the required ASN1 format: arbitrary types determined by an OID.
15858 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15859 argument to 'req'. This is not because the function is newer or
15860 better than others it just uses the work 'NEW' in the certificate
15861 request header lines. Some software needs this.
15865 * Reorganise password command line arguments: now passwords can be
15866 obtained from various sources. Delete the PEM_cb function and make
15867 it the default behaviour: i.e. if the callback is NULL and the
15868 usrdata argument is not NULL interpret it as a null terminated pass
15869 phrase. If usrdata and the callback are NULL then the pass phrase
15870 is prompted for as usual.
15874 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15875 the support is automatically enabled. The resulting binaries will
15876 autodetect the card and use it if present.
15878 *Ben Laurie and Compaq Inc.*
15880 * Work around for Netscape hang bug. This sends certificate request
15881 and server done in one record. Since this is perfectly legal in the
15882 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15883 the bugs/SSLv3 entry for more info.
15887 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15891 * Add -rand argument to smime and pkcs12 applications and read/write
15896 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15900 * Add command line password options to the remaining applications.
15904 * Bug fix for BN_div_recp() for numerators with an even number of
15909 * More tests in bntest.c, and changed test_bn output.
15913 * ./config recognizes MacOS X now.
15917 * Bug fix for BN_div() when the first words of num and divisor are
15918 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15922 * Add support for various broken PKCS#8 formats, and command line
15923 options to produce them.
15927 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15928 get temporary BIGNUMs from a BN_CTX.
15932 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15937 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15938 include a #define from the old name to the new. The original intent
15939 was that statically linked binaries could for example just call
15940 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15941 link with digests. This never worked because SSLeay_add_all_digests()
15942 and SSLeay_add_all_ciphers() were in the same source file so calling
15943 one would link with the other. They are now in separate source files.
15947 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15951 * Use a less unusual form of the Miller-Rabin primality test (it used
15952 a binary algorithm for exponentiation integrated into the Miller-Rabin
15953 loop, our standard modexp algorithms are faster).
15957 * Support for the EBCDIC character set completed.
15959 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15961 * Source code cleanups: use const where appropriate, eliminate casts,
15962 use `void *` instead of `char *` in lhash.
15966 * Bugfix: ssl3_send_server_key_exchange was not restartable
15967 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15968 this the server could overwrite ephemeral keys that the client
15973 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15974 using 50 iterations of the Rabin-Miller test.
15976 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15977 iterations of the Rabin-Miller test as required by the appendix
15978 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15979 As BN_is_prime_fasttest includes trial division, DSA parameter
15980 generation becomes much faster.
15982 This implies a change for the callback functions in DSA_is_prime
15983 and DSA_generate_parameters: The callback function is called once
15984 for each positive witness in the Rabin-Miller test, not just
15985 occasionally in the inner loop; and the parameters to the
15986 callback function now provide an iteration count for the outer
15987 loop rather than for the current invocation of the inner loop.
15988 DSA_generate_parameters additionally can call the callback
15989 function with an 'iteration count' of -1, meaning that a
15990 candidate has passed the trial division test (when q is generated
15991 from an application-provided seed, trial division is skipped).
15995 * New function BN_is_prime_fasttest that optionally does trial
15996 division before starting the Rabin-Miller test and has
15997 an additional BN_CTX * argument (whereas BN_is_prime always
15998 has to allocate at least one BN_CTX).
15999 'callback(1, -1, cb_arg)' is called when a number has passed the
16000 trial division stage.
16004 * Fix for bug in CRL encoding. The validity dates weren't being handled
16009 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16013 * New function BN_pseudo_rand().
16017 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16018 bignum version of BN_from_montgomery() with the working code from
16019 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16024 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16025 made it impossible to use the same SSL_SESSION data structure in
16026 SSL2 clients in multiple threads.
16030 * The return value of RAND_load_file() no longer counts bytes obtained
16031 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16032 to seed the PRNG (previously an explicit byte count was required).
16034 *Ulf Möller, Bodo Möller*
16036 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16037 used `char *` instead of `void *` and had casts all over the place.
16041 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16045 * Retain source code compatibility for BN_prime_checks macro:
16046 BN_is_prime(..., BN_prime_checks, ...) now uses
16047 BN_prime_checks_for_size to determine the appropriate number of
16048 Rabin-Miller iterations.
16052 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16053 DH_CHECK_P_NOT_SAFE_PRIME.
16054 (Check if this is true? OpenPGP calls them "strong".)
16058 * Merge the functionality of "dh" and "gendh" programs into a new program
16059 "dhparam". The old programs are retained for now but will handle DH keys
16060 (instead of parameters) in future.
16064 * Make the ciphers, s_server and s_client programs check the return values
16065 when a new cipher list is set.
16069 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16070 ciphers. Before when the 56bit ciphers were enabled the sorting was
16073 The syntax for the cipher sorting has been extended to support sorting by
16074 cipher-strength (using the strength_bits hard coded in the tables).
16075 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16077 Fix a bug in the cipher-command parser: when supplying a cipher command
16078 string with an "undefined" symbol (neither command nor alphanumeric
16079 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16080 an error is flagged.
16082 Due to the strength-sorting extension, the code of the
16083 ssl_create_cipher_list() function was completely rearranged. I hope that
16084 the readability was also increased :-)
16086 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16088 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16089 for the first serial number and places 2 in the serial number file. This
16090 avoids problems when the root CA is created with serial number zero and
16091 the first user certificate has the same issuer name and serial number
16096 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16097 the new code. Add documentation for this stuff.
16101 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16102 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16103 structures and behave in an analogous way to the X509v3 functions:
16104 they shouldn't be called directly but wrapper functions should be used
16107 So we also now have some wrapper functions that call the X509at functions
16108 when passed certificate requests. (TO DO: similar things can be done with
16109 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16110 things. Some of these need some d2i or i2d and print functionality
16111 because they handle more complex structures.)
16115 * Add missing #ifndefs that caused missing symbols when building libssl
16116 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16117 NO_RSA in `ssl/s2*.c`.
16119 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16121 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16122 has a return value which indicates the quality of the random data
16123 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16124 error queue. New function RAND_pseudo_bytes() generates output that is
16125 guaranteed to be unique but not unpredictable. RAND_add is like
16126 RAND_seed, but takes an extra argument for an entropy estimate
16127 (RAND_seed always assumes full entropy).
16131 * Do more iterations of Rabin-Miller probable prime test (specifically,
16132 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16133 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16134 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16135 false-positive rate of at most 2^-80 for random input.
16139 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16143 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16144 in the 0.9.5 release), this returns the chain
16145 from an X509_CTX structure with a dup of the stack and all
16146 the X509 reference counts upped: so the stack will exist
16147 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16150 Also make SSL_SESSION_print() print out the verify return
16155 * Add manpage for the pkcs12 command. Also change the default
16156 behaviour so MAC iteration counts are used unless the new
16157 -nomaciter option is used. This improves file security and
16158 only older versions of MSIE (4.0 for example) need it.
16162 * Honor the no-xxx Configure options when creating .DEF files.
16166 * Add PKCS#10 attributes to field table: challengePassword,
16167 unstructuredName and unstructuredAddress. These are taken from
16168 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16169 international characters are used.
16171 More changes to X509_ATTRIBUTE code: allow the setting of types
16172 based on strings. Remove the 'loc' parameter when adding
16173 attributes because these will be a SET OF encoding which is sorted
16178 * Initial changes to the 'req' utility to allow request generation
16179 automation. This will allow an application to just generate a template
16180 file containing all the field values and have req construct the
16183 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16184 used all over the place including certificate requests and PKCS#7
16185 structures. They are currently handled manually where necessary with
16186 some primitive wrappers for PKCS#7. The new functions behave in a
16187 manner analogous to the X509 extension functions: they allow
16188 attributes to be looked up by NID and added.
16190 Later something similar to the X509V3 code would be desirable to
16191 automatically handle the encoding, decoding and printing of the
16192 more complex types. The string types like challengePassword can
16193 be handled by the string table functions.
16195 Also modified the multi byte string table handling. Now there is
16196 a 'global mask' which masks out certain types. The table itself
16197 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16198 is useful when for example there is only one permissible type
16199 (as in countryName) and using the mask might result in no valid
16204 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16205 SSL_get_peer_finished to allow applications to obtain the latest
16206 Finished messages sent to the peer or expected from the peer,
16207 respectively. (SSL_get_peer_finished is usually the Finished message
16208 actually received from the peer, otherwise the protocol will be aborted.)
16210 As the Finished message are message digests of the complete handshake
16211 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16212 be used for external authentication procedures when the authentication
16213 provided by SSL/TLS is not desired or is not enough.
16217 * Enhanced support for Alpha Linux is added. Now ./config checks if
16218 the host supports BWX extension and if Compaq C is present on the
16219 $PATH. Just exploiting of the BWX extension results in 20-30%
16220 performance kick for some algorithms, e.g. DES and RC4 to mention
16221 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16226 * Add support for MS "fast SGC". This is arguably a violation of the
16227 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16228 weak crypto and after checking the certificate is SGC a second one
16229 with strong crypto. MS SGC stops the first handshake after receiving
16230 the server certificate message and sends a second client hello. Since
16231 a server will typically do all the time consuming operations before
16232 expecting any further messages from the client (server key exchange
16233 is the most expensive) there is little difference between the two.
16235 To get OpenSSL to support MS SGC we have to permit a second client
16236 hello message after we have sent server done. In addition we have to
16237 reset the MAC if we do get this second client hello.
16241 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16242 if a DER encoded private key is RSA or DSA traditional format. Changed
16243 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16244 format DER encoded private key. Newer code should use PKCS#8 format which
16245 has the key type encoded in the ASN1 structure. Added DER private key
16246 support to pkcs8 application.
16250 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16251 ciphersuites has been selected (as required by the SSL 3/TLS 1
16252 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16253 is set, we interpret this as a request to violate the specification
16254 (the worst that can happen is a handshake failure, and 'correct'
16255 behaviour would result in a handshake failure anyway).
16259 * In SSL_CTX_add_session, take into account that there might be multiple
16260 SSL_SESSION structures with the same session ID (e.g. when two threads
16261 concurrently obtain them from an external cache).
16262 The internal cache can handle only one SSL_SESSION with a given ID,
16263 so if there's a conflict, we now throw out the old one to achieve
16268 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16269 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16270 some routines that use cipher OIDs: some ciphers do not have OIDs
16271 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16276 * Simplify the trust setting structure and code. Now we just have
16277 two sequences of OIDs for trusted and rejected settings. These will
16278 typically have values the same as the extended key usage extension
16279 and any application specific purposes.
16281 The trust checking code now has a default behaviour: it will just
16282 check for an object with the same NID as the passed id. Functions can
16283 be provided to override either the default behaviour or the behaviour
16284 for a given id. SSL client, server and email already have functions
16285 in place for compatibility: they check the NID and also return "trusted"
16286 if the certificate is self signed.
16290 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16291 traditional format into an EVP_PKEY structure.
16295 * Add a password callback function PEM_cb() which either prompts for
16296 a password if usr_data is NULL or otherwise assumes it is a null
16297 terminated password. Allow passwords to be passed on command line
16298 environment or config files in a few more utilities.
16302 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16303 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16304 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16305 Update documentation.
16309 * Support for ASN1 "NULL" type. This could be handled before by using
16310 ASN1_TYPE but there wasn't any function that would try to read a NULL
16311 and produce an error if it couldn't. For compatibility we also have
16312 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16313 don't allocate anything because they don't need to.
16317 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16320 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16322 * Rebuild of the memory allocation routines used by OpenSSL code and
16323 possibly others as well. The purpose is to make an interface that
16324 provide hooks so anyone can build a separate set of allocation and
16325 deallocation routines to be used by OpenSSL, for example memory
16326 pool implementations, or something else, which was previously hard
16327 since Malloc(), Realloc() and Free() were defined as macros having
16328 the values malloc, realloc and free, respectively (except for Win32
16329 compilations). The same is provided for memory debugging code.
16330 OpenSSL already comes with functionality to find memory leaks, but
16331 this gives people a chance to debug other memory problems.
16333 With these changes, a new set of functions and macros have appeared:
16335 CRYPTO_set_mem_debug_functions() [F]
16336 CRYPTO_get_mem_debug_functions() [F]
16337 CRYPTO_dbg_set_options() [F]
16338 CRYPTO_dbg_get_options() [F]
16339 CRYPTO_malloc_debug_init() [M]
16341 The memory debug functions are NULL by default, unless the library
16342 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16343 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16344 gives the standard debugging functions that come with OpenSSL) or
16345 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16346 provided by the library user) must be used. When the standard
16347 debugging functions are used, CRYPTO_dbg_set_options can be used to
16348 request additional information:
16349 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16350 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16352 Also, things like CRYPTO_set_mem_functions will always give the
16353 expected result (the new set of functions is used for allocation
16354 and deallocation) at all times, regardless of platform and compiler
16357 To finish it up, some functions that were never use in any other
16358 way than through macros have a new API and new semantic:
16360 CRYPTO_dbg_malloc()
16361 CRYPTO_dbg_realloc()
16364 All macros of value have retained their old syntax.
16366 *Richard Levitte and Bodo Moeller*
16368 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16369 ordering of SMIMECapabilities wasn't in "strength order" and there
16370 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16375 * Some ASN1 types with illegal zero length encoding (INTEGER,
16376 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16378 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16380 * Merge in my S/MIME library for OpenSSL. This provides a simple
16381 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16382 functionality to handle multipart/signed properly) and a utility
16383 called 'smime' to call all this stuff. This is based on code I
16384 originally wrote for Celo who have kindly allowed it to be
16385 included in OpenSSL.
16389 * Add variants des_set_key_checked and des_set_key_unchecked of
16390 des_set_key (aka des_key_sched). Global variable des_check_key
16391 decides which of these is called by des_set_key; this way
16392 des_check_key behaves as it always did, but applications and
16393 the library itself, which was buggy for des_check_key == 1,
16394 have a cleaner way to pick the version they need.
16398 * New function PKCS12_newpass() which changes the password of a
16403 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16404 dynamic mix. In both cases the ids can be used as an index into the
16405 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16406 functions so they accept a list of the field values and the
16407 application doesn't need to directly manipulate the X509_TRUST
16412 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16417 * Modify the way the V3 extension code looks up extensions. This now
16418 works in a similar way to the object code: we have some "standard"
16419 extensions in a static table which is searched with OBJ_bsearch()
16420 and the application can add dynamic ones if needed. The file
16421 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16422 updated whenever a new extension is added to the core code and kept
16423 in ext_nid order. There is a simple program 'tabtest.c' which checks
16424 this. New extensions are not added too often so this file can readily
16425 be maintained manually.
16427 There are two big advantages in doing things this way. The extensions
16428 can be looked up immediately and no longer need to be "added" using
16429 X509V3_add_standard_extensions(): this function now does nothing.
16430 Side note: I get *lots* of email saying the extension code doesn't
16431 work because people forget to call this function.
16432 Also no dynamic allocation is done unless new extensions are added:
16433 so if we don't add custom extensions there is no need to call
16434 X509V3_EXT_cleanup().
16438 * Modify enc utility's salting as follows: make salting the default. Add a
16439 magic header, so unsalted files fail gracefully instead of just decrypting
16440 to garbage. This is because not salting is a big security hole, so people
16441 should be discouraged from doing it.
16445 * Fixes and enhancements to the 'x509' utility. It allowed a message
16446 digest to be passed on the command line but it only used this
16447 parameter when signing a certificate. Modified so all relevant
16448 operations are affected by the digest parameter including the
16449 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16450 DSA key was used because it didn't fix the digest.
16454 * Initial certificate chain verify code. Currently tests the untrusted
16455 certificates for consistency with the verify purpose (which is set
16456 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16458 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16459 this is because it will reject chains with invalid extensions whereas
16460 every previous version of OpenSSL and SSLeay made no checks at all.
16462 Trust code: checks the root CA for the relevant trust settings. Trust
16463 settings have an initial value consistent with the verify purpose: e.g.
16464 if the verify purpose is for SSL client use it expects the CA to be
16465 trusted for SSL client use. However the default value can be changed to
16466 permit custom trust settings: one example of this would be to only trust
16467 certificates from a specific "secure" set of CAs.
16469 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16470 which should be used for version portability: especially since the
16471 verify structure is likely to change more often now.
16473 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16474 to set them. If not set then assume SSL clients will verify SSL servers
16477 Two new options to the verify program: -untrusted allows a set of
16478 untrusted certificates to be passed in and -purpose which sets the
16479 intended purpose of the certificate. If a purpose is set then the
16480 new chain verify code is used to check extension consistency.
16484 * Support for the authority information access extension.
16488 * Modify RSA and DSA PEM read routines to transparently handle
16489 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16490 public keys in a format compatible with certificate
16491 SubjectPublicKeyInfo structures. Unfortunately there were already
16492 functions called *_PublicKey_* which used various odd formats so
16493 these are retained for compatibility: however the DSA variants were
16494 never in a public release so they have been deleted. Changed dsa/rsa
16495 utilities to handle the new format: note no releases ever handled public
16496 keys so we should be OK.
16498 The primary motivation for this change is to avoid the same fiasco
16499 that dogs private keys: there are several incompatible private key
16500 formats some of which are standard and some OpenSSL specific and
16501 require various evil hacks to allow partial transparent handling and
16502 even then it doesn't work with DER formats. Given the option anything
16503 other than PKCS#8 should be dumped: but the other formats have to
16504 stay in the name of compatibility.
16506 With public keys and the benefit of hindsight one standard format
16507 is used which works with EVP_PKEY, RSA or DSA structures: though
16508 it clearly returns an error if you try to read the wrong kind of key.
16510 Added a -pubkey option to the 'x509' utility to output the public key.
16511 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16512 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16513 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16514 that do the same as the `EVP_PKEY_assign_*()` except they up the
16515 reference count of the added key (they don't "swallow" the
16520 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16521 CRLs would fail if the file contained no certificates or no CRLs:
16522 added a new function to read in both types and return the number
16523 read: this means that if none are read it will be an error. The
16524 DER versions of the certificate and CRL reader would always fail
16525 because it isn't possible to mix certificates and CRLs in DER format
16526 without choking one or the other routine. Changed this to just read
16527 a certificate: this is the best we can do. Also modified the code
16528 in `apps/verify.c` to take notice of return codes: it was previously
16529 attempting to read in certificates from NULL pointers and ignoring
16530 any errors: this is one reason why the cert and CRL reader seemed
16531 to work. It doesn't check return codes from the default certificate
16532 routines: these may well fail if the certificates aren't installed.
16536 * Code to support otherName option in GeneralName.
16540 * First update to verify code. Change the verify utility
16541 so it warns if it is passed a self signed certificate:
16542 for consistency with the normal behaviour. X509_verify
16543 has been modified to it will now verify a self signed
16544 certificate if *exactly* the same certificate appears
16545 in the store: it was previously impossible to trust a
16546 single self signed certificate. This means that:
16547 openssl verify ss.pem
16548 now gives a warning about a self signed certificate but
16549 openssl verify -CAfile ss.pem ss.pem
16554 * For servers, store verify_result in SSL_SESSION data structure
16555 (and add it to external session representation).
16556 This is needed when client certificate verifications fails,
16557 but an application-provided verification callback (set by
16558 SSL_CTX_set_cert_verify_callback) allows accepting the session
16559 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16560 but returns 1): When the session is reused, we have to set
16561 ssl->verify_result to the appropriate error code to avoid
16564 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16566 * Fix a bug in the new PKCS#7 code: it didn't consider the
16567 case in PKCS7_dataInit() where the signed PKCS7 structure
16568 didn't contain any existing data because it was being created.
16570 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16572 * Add a salt to the key derivation routines in enc.c. This
16573 forms the first 8 bytes of the encrypted file. Also add a
16574 -S option to allow a salt to be input on the command line.
16578 * New function X509_cmp(). Oddly enough there wasn't a function
16579 to compare two certificates. We do this by working out the SHA1
16580 hash and comparing that. X509_cmp() will be needed by the trust
16585 * SSL_get1_session() is like SSL_get_session(), but increments
16586 the reference count in the SSL_SESSION returned.
16588 *Geoff Thorpe <geoff@eu.c2.net>*
16590 * Fix for 'req': it was adding a null to request attributes.
16591 Also change the X509_LOOKUP and X509_INFO code to handle
16592 certificate auxiliary information.
16596 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16601 * Add the possibility to add extra information to the memory leak
16602 detecting output, to form tracebacks, showing from where each
16603 allocation was originated: CRYPTO_push_info("constant string") adds
16604 the string plus current file name and line number to a per-thread
16605 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16606 is like calling CYRPTO_pop_info() until the stack is empty.
16607 Also updated memory leak detection code to be multi-thread-safe.
16611 * Add options -text and -noout to pkcs7 utility and delete the
16612 encryption options which never did anything. Update docs.
16616 * Add options to some of the utilities to allow the pass phrase
16617 to be included on either the command line (not recommended on
16618 OSes like Unix) or read from the environment. Update the
16619 manpages and fix a few bugs.
16623 * Add a few manpages for some of the openssl commands.
16627 * Fix the -revoke option in ca. It was freeing up memory twice,
16628 leaking and not finding already revoked certificates.
16632 * Extensive changes to support certificate auxiliary information.
16633 This involves the use of X509_CERT_AUX structure and X509_AUX
16634 functions. An X509_AUX function such as PEM_read_X509_AUX()
16635 can still read in a certificate file in the usual way but it
16636 will also read in any additional "auxiliary information". By
16637 doing things this way a fair degree of compatibility can be
16638 retained: existing certificates can have this information added
16639 using the new 'x509' options.
16641 Current auxiliary information includes an "alias" and some trust
16642 settings. The trust settings will ultimately be used in enhanced
16643 certificate chain verification routines: currently a certificate
16644 can only be trusted if it is self signed and then it is trusted
16649 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16650 The problem was that one of the replacement routines had not been working
16651 since SSLeay releases. For now the offending routine has been replaced
16652 with non-optimised assembler. Even so, this now gives around 95%
16653 performance improvement for 1024 bit RSA signs.
16657 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16658 handling. Most clients have the effective key size in bits equal to
16659 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16660 A few however don't do this and instead use the size of the decrypted key
16661 to determine the RC2 key length and the AlgorithmIdentifier to determine
16662 the effective key length. In this case the effective key length can still
16663 be 40 bits but the key length can be 168 bits for example. This is fixed
16664 by manually forcing an RC2 key into the EVP_PKEY structure because the
16665 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16666 the key length and effective key length are equal.
16670 * Add a bunch of functions that should simplify the creation of
16671 X509_NAME structures. Now you should be able to do:
16672 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16673 and have it automatically work out the correct field type and fill in
16674 the structures. The more adventurous can try:
16675 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16676 and it will (hopefully) work out the correct multibyte encoding.
16680 * Change the 'req' utility to use the new field handling and multibyte
16681 copy routines. Before the DN field creation was handled in an ad hoc
16682 way in req, ca, and x509 which was rather broken and didn't support
16683 BMPStrings or UTF8Strings. Since some software doesn't implement
16684 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16685 using the dirstring_type option. See the new comment in the default
16686 openssl.cnf for more info.
16690 * Make crypto/rand/md_rand.c more robust:
16691 - Assure unique random numbers after fork().
16692 - Make sure that concurrent threads access the global counter and
16693 md serializably so that we never lose entropy in them
16694 or use exactly the same state in multiple threads.
16695 Access to the large state is not always serializable because
16696 the additional locking could be a performance killer, and
16697 md should be large enough anyway.
16701 * New file `apps/app_rand.c` with commonly needed functionality
16702 for handling the random seed file.
16704 Use the random seed file in some applications that previously did not:
16706 dsaparam -genkey (which also ignored its '-rand' option),
16709 x509 (when signing).
16710 Except on systems with /dev/urandom, it is crucial to have a random
16711 seed file at least for key creation, DSA signing, and for DH exchanges;
16712 for RSA signatures we could do without one.
16714 gendh and gendsa (unlike genrsa) used to read only the first byte
16715 of each file listed in the '-rand' option. The function as previously
16716 found in genrsa is now in app_rand.c and is used by all programs
16717 that support '-rand'.
16721 * In RAND_write_file, use mode 0600 for creating files;
16722 don't just chmod when it may be too late.
16726 * Report an error from X509_STORE_load_locations
16727 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16731 * New function ASN1_mbstring_copy() this copies a string in either
16732 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16733 into an ASN1_STRING type. A mask of permissible types is passed
16734 and it chooses the "minimal" type to use or an error if not type
16739 * Add function equivalents to the various macros in asn1.h. The old
16740 macros are retained with an `M_` prefix. Code inside the library can
16741 use the `M_` macros. External code (including the openssl utility)
16742 should *NOT* in order to be "shared library friendly".
16746 * Add various functions that can check a certificate's extensions
16747 to see if it usable for various purposes such as SSL client,
16748 server or S/MIME and CAs of these types. This is currently
16749 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16750 verification. Also added a -purpose flag to x509 utility to
16751 print out all the purposes.
16755 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16760 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16761 for, obtain and decode and extension and obtain its critical flag.
16762 This allows all the necessary extension code to be handled in a
16763 single function call.
16767 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16768 platforms. See crypto/rc4/rc4_enc.c for further details.
16772 * New -noout option to asn1parse. This causes no output to be produced
16773 its main use is when combined with -strparse and -out to extract data
16774 from a file (which may not be in ASN.1 format).
16778 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16779 when producing the local key id.
16781 *Richard Levitte <levitte@stacken.kth.se>*
16783 * New option -dhparam in s_server. This allows a DH parameter file to be
16784 stated explicitly. If it is not stated then it tries the first server
16785 certificate file. The previous behaviour hard coded the filename
16790 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16791 a public key to be input or output. For example:
16792 openssl rsa -in key.pem -pubout -out pubkey.pem
16793 Also added necessary DSA public key functions to handle this.
16797 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16798 in the message. This was handled by allowing
16799 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16801 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16803 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16804 to the end of the strings whereas this didn't. This would cause problems
16805 if strings read with d2i_ASN1_bytes() were later modified.
16807 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16809 * Fix for base64 decode bug. When a base64 bio reads only one line of
16810 data and it contains EOF it will end up returning an error. This is
16811 caused by input 46 bytes long. The cause is due to the way base64
16812 BIOs find the start of base64 encoded data. They do this by trying a
16813 trial decode on each line until they find one that works. When they
16814 do a flag is set and it starts again knowing it can pass all the
16815 data directly through the decoder. Unfortunately it doesn't reset
16816 the context it uses. This means that if EOF is reached an attempt
16817 is made to pass two EOFs through the context and this causes the
16818 resulting error. This can also cause other problems as well. As is
16819 usual with these problems it takes *ages* to find and the fix is
16820 trivial: move one line.
16822 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16824 * Ugly workaround to get s_client and s_server working under Windows. The
16825 old code wouldn't work because it needed to select() on sockets and the
16826 tty (for keypresses and to see if data could be written). Win32 only
16827 supports select() on sockets so we select() with a 1s timeout on the
16828 sockets and then see if any characters are waiting to be read, if none
16829 are present then we retry, we also assume we can always write data to
16830 the tty. This isn't nice because the code then blocks until we've
16831 received a complete line of data and it is effectively polling the
16832 keyboard at 1s intervals: however it's quite a bit better than not
16833 working at all :-) A dedicated Windows application might handle this
16834 with an event loop for example.
16838 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16839 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16840 will be called when RSA_sign() and RSA_verify() are used. This is useful
16841 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16842 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16843 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16844 This necessitated the support of an extra signature type NID_md5_sha1
16845 for SSL signatures and modifications to the SSL library to use it instead
16846 of calling RSA_public_decrypt() and RSA_private_encrypt().
16850 * Add new -verify -CAfile and -CApath options to the crl program, these
16851 will lookup a CRL issuers certificate and verify the signature in a
16852 similar way to the verify program. Tidy up the crl program so it
16853 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16854 less strict. It will now permit CRL extensions even if it is not
16855 a V2 CRL: this will allow it to tolerate some broken CRLs.
16859 * Initialize all non-automatic variables each time one of the openssl
16860 sub-programs is started (this is necessary as they may be started
16861 multiple times from the "OpenSSL>" prompt).
16863 *Lennart Bang, Bodo Moeller*
16865 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16866 removing all other RSA functionality (this is what NO_RSA does). This
16867 is so (for example) those in the US can disable those operations covered
16868 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16873 * Non-copying interface to BIO pairs.
16874 (still largely untested)
16878 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16879 ASCII string. This was handled independently in various places before.
16883 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16884 UTF8 strings a character at a time.
16888 * Use client_version from client hello to select the protocol
16889 (s23_srvr.c) and for RSA client key exchange verification
16890 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16894 * Add various utility functions to handle SPKACs, these were previously
16895 handled by poking round in the structure internals. Added new function
16896 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16897 print, verify and generate SPKACs. Based on an original idea from
16898 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16902 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16906 * Allow the config file extension section to be overwritten on the
16907 command line. Based on an original idea from Massimiliano Pala
16908 <madwolf@comune.modena.it>. The new option is called -extensions
16909 and can be applied to ca, req and x509. Also -reqexts to override
16910 the request extensions in req and -crlexts to override the crl extensions
16915 * Add new feature to the SPKAC handling in ca. Now you can include
16916 the same field multiple times by preceding it by "XXXX." for example:
16919 this is the same syntax as used in the req config file.
16923 * Allow certificate extensions to be added to certificate requests. These
16924 are specified in a 'req_extensions' option of the req section of the
16925 config file. They can be printed out with the -text option to req but
16926 are otherwise ignored at present.
16930 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16931 data read consists of only the final block it would not decrypted because
16932 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16933 A misplaced 'break' also meant the decrypted final block might not be
16934 copied until the next read.
16938 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16939 a few extra parameters to the DH structure: these will be useful if
16940 for example we want the value of 'q' or implement X9.42 DH.
16944 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16945 provides hooks that allow the default DSA functions or functions on a
16946 "per key" basis to be replaced. This allows hardware acceleration and
16947 hardware key storage to be handled without major modification to the
16948 library. Also added low level modexp hooks and CRYPTO_EX structure and
16949 associated functions.
16953 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16954 as "read only": it can't be written to and the buffer it points to will
16955 not be freed. Reading from a read only BIO is much more efficient than
16956 a normal memory BIO. This was added because there are several times when
16957 an area of memory needs to be read from a BIO. The previous method was
16958 to create a memory BIO and write the data to it, this results in two
16959 copies of the data and an O(n^2) reading algorithm. There is a new
16960 function BIO_new_mem_buf() which creates a read only memory BIO from
16961 an area of memory. Also modified the PKCS#7 routines to use read only
16966 * Bugfix: ssl23_get_client_hello did not work properly when called in
16967 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16968 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16969 but a retry condition occurred while trying to read the rest.
16973 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16974 NID_pkcs7_encrypted by default: this was wrong since this should almost
16975 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16976 the encrypted data type: this is a more sensible place to put it and it
16977 allows the PKCS#12 code to be tidied up that duplicated this
16982 * Changed obj_dat.pl script so it takes its input and output files on
16983 the command line. This should avoid shell escape redirection problems
16988 * Initial support for certificate extension requests, these are included
16989 in things like Xenroll certificate requests. Included functions to allow
16990 extensions to be obtained and added.
16994 * -crlf option to s_client and s_server for sending newlines as
16995 CRLF (as required by many protocols).
16999 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17001 * Install libRSAglue.a when OpenSSL is built with RSAref.
17003 *Ralf S. Engelschall*
17005 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17007 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17009 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17014 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17015 DH parameters/keys (q is lost during that conversion, but the resulting
17016 DH parameters contain its length).
17018 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17019 much faster than DH_generate_parameters (which creates parameters
17020 where `p = 2*q + 1`), and also the smaller q makes DH computations
17021 much more efficient (160-bit exponentiation instead of 1024-bit
17022 exponentiation); so this provides a convenient way to support DHE
17023 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17024 utter importance to use
17025 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17027 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17028 when such DH parameters are used, because otherwise small subgroup
17029 attacks may become possible!
17033 * Avoid memory leak in i2d_DHparams.
17037 * Allow the -k option to be used more than once in the enc program:
17038 this allows the same encrypted message to be read by multiple recipients.
17042 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17043 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17044 it will always use the numerical form of the OID, even if it has a short
17049 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17050 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17051 otherwise bn_mod_exp was called. In the case of hardware keys for example
17052 no private key components need be present and it might store extra data
17053 in the RSA structure, which cannot be accessed from bn_mod_exp.
17054 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17055 private key operations.
17059 * Added support for SPARC Linux.
17063 * pem_password_cb function type incompatibly changed from
17064 typedef int pem_password_cb(char *buf, int size, int rwflag);
17066 ....(char *buf, int size, int rwflag, void *userdata);
17067 so that applications can pass data to their callbacks:
17068 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17069 additional void * argument, which is just handed through whenever
17070 the password callback is called.
17072 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17074 New function SSL_CTX_set_default_passwd_cb_userdata.
17076 Compatibility note: As many C implementations push function arguments
17077 onto the stack in reverse order, the new library version is likely to
17078 interoperate with programs that have been compiled with the old
17079 pem_password_cb definition (PEM_whatever takes some data that
17080 happens to be on the stack as its last argument, and the callback
17081 just ignores this garbage); but there is no guarantee whatsoever that
17084 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17085 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17086 problems not only on Windows, but also on some Unix platforms.
17087 To avoid problematic command lines, these definitions are now in an
17088 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17089 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17093 * MIPS III/IV assembler module is reimplemented.
17097 * More DES library cleanups: remove references to srand/rand and
17098 delete an unused file.
17102 * Add support for the free Netwide assembler (NASM) under Win32,
17103 since not many people have MASM (ml) and it can be hard to obtain.
17104 This is currently experimental but it seems to work OK and pass all
17105 the tests. Check out INSTALL.W32 for info.
17109 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17110 without temporary keys kept an extra copy of the server key,
17111 and connections with temporary keys did not free everything in case
17116 * New function RSA_check_key and new openssl rsa option -check
17117 for verifying the consistency of RSA keys.
17119 *Ulf Moeller, Bodo Moeller*
17121 * Various changes to make Win32 compile work:
17122 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17123 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17124 comparison" warnings.
17125 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17129 * Add a debugging option to PKCS#5 v2 key generation function: when
17130 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17131 derived keys are printed to stderr.
17135 * Copy the flags in ASN1_STRING_dup().
17137 *Roman E. Pavlov <pre@mo.msk.ru>*
17139 * The x509 application mishandled signing requests containing DSA
17140 keys when the signing key was also DSA and the parameters didn't match.
17142 It was supposed to omit the parameters when they matched the signing key:
17143 the verifying software was then supposed to automatically use the CA's
17144 parameters if they were absent from the end user certificate.
17146 Omitting parameters is no longer recommended. The test was also
17147 the wrong way round! This was probably due to unusual behaviour in
17148 EVP_cmp_parameters() which returns 1 if the parameters match.
17149 This meant that parameters were omitted when they *didn't* match and
17150 the certificate was useless. Certificates signed with 'ca' didn't have
17153 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17155 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17156 The interface is as follows:
17157 Applications can use
17158 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17159 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17160 "off" is now the default.
17161 The library internally uses
17162 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17163 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17164 to disable memory-checking temporarily.
17166 Some inconsistent states that previously were possible (and were
17167 even the default) are now avoided.
17169 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17170 with each memory chunk allocated; this is occasionally more helpful
17171 than just having a counter.
17173 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17175 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17180 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17181 which largely parallels "options", but is for changing API behaviour,
17182 whereas "options" are about protocol behaviour.
17183 Initial "mode" flags are:
17185 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17186 a single record has been written.
17187 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17188 retries use the same buffer location.
17189 (But all of the contents must be
17194 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17197 * Fix problems with no-hmac etc.
17199 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17201 * New functions RSA_get_default_method(), RSA_set_method() and
17202 RSA_get_method(). These allows replacement of RSA_METHODs without having
17203 to mess around with the internals of an RSA structure.
17207 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17208 Also really enable memory leak checks in openssl.c and in some
17211 *Chad C. Mulligan, Bodo Moeller*
17213 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17214 up the length of negative integers. This has now been simplified to just
17215 store the length when it is first determined and use it later, rather
17216 than trying to keep track of where data is copied and updating it to
17218 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17220 * Add a new function PKCS7_signatureVerify. This allows the verification
17221 of a PKCS#7 signature but with the signing certificate passed to the
17222 function itself. This contrasts with PKCS7_dataVerify which assumes the
17223 certificate is present in the PKCS#7 structure. This isn't always the
17224 case: certificates can be omitted from a PKCS#7 structure and be
17225 distributed by "out of band" means (such as a certificate database).
17229 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17230 function prototypes in pem.h, also change util/mkdef.pl to add the
17231 necessary function names.
17235 * mk1mf.pl (used by Windows builds) did not properly read the
17236 options set by Configure in the top level Makefile, and Configure
17237 was not even able to write more than one option correctly.
17238 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17242 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17243 file to be loaded from a BIO or FILE pointer. The BIO version will
17244 for example allow memory BIOs to contain config info.
17248 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17249 Whoever hopes to achieve shared-library compatibility across versions
17250 must use this, not the compile-time macro.
17251 (Exercise 0.9.4: Which is the minimum library version required by
17253 Note: All this applies only to multi-threaded programs, others don't
17258 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17259 through a BIO pair triggered the default case, i.e.
17260 SSLerr(...,SSL_R_UNKNOWN_STATE).
17264 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17265 can use the SSL library even if none of the specific BIOs is
17270 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17271 for the encoded length.
17273 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17275 * Add initial documentation of the X509V3 functions.
17279 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17280 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17281 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17282 secure PKCS#8 private key format with a high iteration count.
17286 * Fix determination of Perl interpreter: A perl or perl5
17287 *directory* in $PATH was also accepted as the interpreter.
17289 *Ralf S. Engelschall*
17291 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17292 wrong with it but it was very old and did things like calling
17293 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17294 unusual formatting.
17298 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17299 to use the new extension code.
17303 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17304 with macros. This should make it easier to change their form, add extra
17305 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17310 * Add to configuration table a new entry that can specify an alternative
17311 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17312 according to Mark Crispin <MRC@Panda.COM>.
17316 * DES CBC did not update the IV. Weird.
17320 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17321 Changing the behaviour of the former might break existing programs --
17322 where IV updating is needed, des_ncbc_encrypt can be used.
17325 * When bntest is run from "make test" it drives bc to check its
17326 calculations, as well as internally checking them. If an internal check
17327 fails, it needs to cause bc to give a non-zero result or make test carries
17328 on without noticing the failure. Fixed.
17332 * DES library cleanups.
17336 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17337 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17338 ciphers. NOTE: although the key derivation function has been verified
17339 against some published test vectors it has not been extensively tested
17340 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17345 * Instead of "mkdir -p", which is not fully portable, use new
17346 Perl script "util/mkdir-p.pl".
17350 * Rewrite the way password based encryption (PBE) is handled. It used to
17351 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17352 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17353 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17354 the 'parameter' field of the AlgorithmIdentifier is passed to the
17355 underlying key generation function so it must do its own ASN1 parsing.
17356 This has also changed the EVP_PBE_CipherInit() function which now has a
17357 'parameter' argument instead of literal salt and iteration count values
17358 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17362 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17363 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17364 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17365 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17366 value was just used as a "magic string" and not used directly its
17367 value doesn't matter.
17371 * Introduce some semblance of const correctness to BN. Shame C doesn't
17376 * "linux-sparc64" configuration (ultrapenguin).
17378 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17379 "linux-sparc" configuration.
17381 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17383 * config now generates no-xxx options for missing ciphers.
17387 * Support the EBCDIC character set (work in progress).
17388 File ebcdic.c not yet included because it has a different license.
17390 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17392 * Support BS2000/OSD-POSIX.
17394 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17396 * Make callbacks for key generation use `void *` instead of `char *`.
17400 * Make S/MIME samples compile (not yet tested).
17404 * Additional typesafe stacks.
17408 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17412 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17414 * New configuration variant "sco5-gcc".
17416 * Updated some demos.
17418 *Sean O Riordain, Wade Scholine*
17420 * Add missing BIO_free at exit of pkcs12 application.
17424 * Fix memory leak in conf.c.
17428 * Updates for Win32 to assembler version of MD5.
17432 * Set #! path to perl in `apps/der_chop` to where we found it
17433 instead of using a fixed path.
17437 * SHA library changes for irix64-mips4-cc.
17441 * Improvements for VMS support.
17445 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17447 * Bignum library bug fix. IRIX 6 passes "make test" now!
17448 This also avoids the problems with SC4.2 and unpatched SC5.
17450 *Andy Polyakov <appro@fy.chalmers.se>*
17452 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17453 These are required because of the typesafe stack would otherwise break
17454 existing code. If old code used a structure member which used to be STACK
17455 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17456 sk_num or sk_value it would produce an error because the num, data members
17457 are not present in STACK_OF. Now it just produces a warning. sk_set
17458 replaces the old method of assigning a value to sk_value
17459 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17460 that does this will no longer work (and should use sk_set instead) but
17461 this could be regarded as a "questionable" behaviour anyway.
17465 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17466 correctly handle encrypted S/MIME data.
17470 * Change type of various DES function arguments from des_cblock
17471 (which means, in function argument declarations, pointer to char)
17472 to des_cblock * (meaning pointer to array with 8 char elements),
17473 which allows the compiler to do more typechecking; it was like
17474 that back in SSLeay, but with lots of ugly casts.
17476 Introduce new type const_des_cblock.
17480 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17481 problems: find RecipientInfo structure that matches recipient certificate
17482 and initialise the ASN1 structures properly based on passed cipher.
17486 * Belatedly make the BN tests actually check the results.
17490 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17491 to and from BNs: it was completely broken. New compilation option
17492 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17493 key elements as negative integers.
17497 * Reorganize and speed up MD5.
17499 *Andy Polyakov <appro@fy.chalmers.se>*
17503 *Richard Levitte <richard@levitte.org>*
17505 * New option -out to asn1parse to allow the parsed structure to be
17506 output to a file. This is most useful when combined with the -strparse
17507 option to examine the output of things like OCTET STRINGS.
17511 * Make SSL library a little more fool-proof by not requiring any longer
17512 that `SSL_set_{accept,connect}_state` be called before
17513 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17514 in many applications because usually everything *appeared* to work as
17515 intended anyway -- now it really works as intended).
17519 * Move openssl.cnf out of lib/.
17523 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17524 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17525 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17527 *Ralf S. Engelschall*
17529 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17530 handle PKCS#7 enveloped data properly.
17532 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17534 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17535 copying pointers. The cert_st handling is changed by this in
17536 various ways (and thus what used to be known as ctx->default_cert
17537 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17538 any longer when s->cert does not give us what we need).
17539 ssl_cert_instantiate becomes obsolete by this change.
17540 As soon as we've got the new code right (possibly it already is?),
17541 we have solved a couple of bugs of the earlier code where s->cert
17542 was used as if it could not have been shared with other SSL structures.
17544 Note that using the SSL API in certain dirty ways now will result
17545 in different behaviour than observed with earlier library versions:
17546 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17547 does not influence s as it used to.
17549 In order to clean up things more thoroughly, inside SSL_SESSION
17550 we don't use CERT any longer, but a new structure SESS_CERT
17551 that holds per-session data (if available); currently, this is
17552 the peer's certificate chain and, for clients, the server's certificate
17553 and temporary key. CERT holds only those values that can have
17554 meaningful defaults in an SSL_CTX.
17558 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17559 from the internal representation. Various PKCS#7 fixes: remove some
17560 evil casts and set the enc_dig_alg field properly based on the signing
17565 * Allow PKCS#12 password to be set from the command line or the
17566 environment. Let 'ca' get its config file name from the environment
17567 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17572 * Allow certificate policies extension to use an IA5STRING for the
17573 organization field. This is contrary to the PKIX definition but
17574 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17579 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17580 without disallowing inline assembler and the like for non-pedantic builds.
17584 * Support Borland C++ builder.
17586 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17592 * SHA-1 cleanups and performance enhancements.
17594 *Andy Polyakov <appro@fy.chalmers.se>*
17596 * Sparc v8plus assembler for the bignum library.
17598 *Andy Polyakov <appro@fy.chalmers.se>*
17600 * Accept any -xxx and +xxx compiler options in Configure.
17604 * Update HPUX configuration.
17608 * Add missing `sk_<type>_unshift()` function to safestack.h
17610 *Ralf S. Engelschall*
17612 * New function SSL_CTX_use_certificate_chain_file that sets the
17613 "extra_cert"s in addition to the certificate. (This makes sense
17614 only for "PEM" format files, as chains as a whole are not
17619 * Support verify_depth from the SSL API.
17620 x509_vfy.c had what can be considered an off-by-one-error:
17621 Its depth (which was not part of the external interface)
17622 was actually counting the number of certificates in a chain;
17623 now it really counts the depth.
17627 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17628 instead of X509err, which often resulted in confusing error
17629 messages since the error codes are not globally unique
17630 (e.g. an alleged error in ssl3_accept when a certificate
17631 didn't match the private key).
17633 * New function SSL_CTX_set_session_id_context that allows to set a default
17634 value (so that you don't need SSL_set_session_id_context for each
17635 connection using the SSL_CTX).
17639 * OAEP decoding bug fix.
17643 * Support INSTALL_PREFIX for package builders, as proposed by
17648 * New Configure options "threads" and "no-threads". For systems
17649 where the proper compiler options are known (currently Solaris
17650 and Linux), "threads" is the default.
17654 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17658 * Install various scripts to $(OPENSSLDIR)/misc, not to
17659 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17660 such as /usr/local/bin.
17664 * "make linux-shared" to build shared libraries.
17666 *Niels Poppe <niels@netbox.org>*
17668 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17672 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17673 extension adding in x509 utility.
17677 * Remove NOPROTO sections and error code comments.
17681 * Partial rewrite of the DEF file generator to now parse the ANSI
17686 * New Configure options --prefix=DIR and --openssldir=DIR.
17690 * Complete rewrite of the error code script(s). It is all now handled
17691 by one script at the top level which handles error code gathering,
17692 header rewriting and C source file generation. It should be much better
17693 than the old method: it now uses a modified version of Ulf's parser to
17694 read the ANSI prototypes in all header files (thus the old K&R definitions
17695 aren't needed for error creation any more) and do a better job of
17696 translating function codes into names. The old 'ASN1 error code embedded
17697 in a comment' is no longer necessary and it doesn't use .err files which
17698 have now been deleted. Also the error code call doesn't have to appear all
17699 on one line (which resulted in some large lines...).
17703 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17707 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17708 0 (which usually indicates a closed connection), but continue reading.
17712 * Fix some race conditions.
17716 * Add support for CRL distribution points extension. Add Certificate
17717 Policies and CRL distribution points documentation.
17721 * Move the autogenerated header file parts to crypto/opensslconf.h.
17725 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17726 8 of keying material. Merlin has also confirmed interop with this fix
17727 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17729 *Merlin Hughes <merlin@baltimore.ie>*
17731 * Fix lots of warnings.
17733 *Richard Levitte <levitte@stacken.kth.se>*
17735 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17736 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17738 *Richard Levitte <levitte@stacken.kth.se>*
17740 * Fix problems with sizeof(long) == 8.
17742 *Andy Polyakov <appro@fy.chalmers.se>*
17744 * Change functions to ANSI C.
17748 * Fix typos in error codes.
17750 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17752 * Remove defunct assembler files from Configure.
17756 * SPARC v8 assembler BIGNUM implementation.
17758 *Andy Polyakov <appro@fy.chalmers.se>*
17760 * Support for Certificate Policies extension: both print and set.
17761 Various additions to support the r2i method this uses.
17765 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17766 return a const string when you are expecting an allocated buffer.
17770 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17771 types DirectoryString and DisplayText.
17775 * Add code to allow r2i extensions to access the configuration database,
17776 add an LHASH database driver and add several ctx helper functions.
17780 * Fix an evil bug in bn_expand2() which caused various BN functions to
17781 fail when they extended the size of a BIGNUM.
17785 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17786 support typesafe stack.
17790 * Fix typo in SSL_[gs]et_options().
17792 *Nils Frostberg <nils@medcom.se>*
17794 * Delete various functions and files that belonged to the (now obsolete)
17795 old X509V3 handling code.
17799 * New Configure option "rsaref".
17803 * Don't auto-generate pem.h.
17807 * Introduce type-safe ASN.1 SETs.
17811 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17813 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17815 * Introduce type-safe STACKs. This will almost certainly break lots of code
17816 that links with OpenSSL (well at least cause lots of warnings), but fear
17817 not: the conversion is trivial, and it eliminates loads of evil casts. A
17818 few STACKed things have been converted already. Feel free to convert more.
17819 In the fullness of time, I'll do away with the STACK type altogether.
17823 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17824 specified in `<certfile>` by updating the entry in the index.txt file.
17825 This way one no longer has to edit the index.txt file manually for
17826 revoking a certificate. The -revoke option does the gory details now.
17828 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17830 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17831 `-text` option at all and this way the `-noout -text` combination was
17832 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17834 *Ralf S. Engelschall*
17836 * Make sure a corresponding plain text error message exists for the
17837 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17838 verify callback function determined that a certificate was revoked.
17840 *Ralf S. Engelschall*
17842 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17843 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17844 all available ciphers including rc5, which was forgotten until now.
17845 In order to let the testing shell script know which algorithms
17846 are available, a new (up to now undocumented) command
17847 `openssl list-cipher-commands` is used.
17851 * Bugfix: s_client occasionally would sleep in select() when
17852 it should have checked SSL_pending() first.
17856 * New functions DSA_do_sign and DSA_do_verify to provide access to
17857 the raw DSA values prior to ASN.1 encoding.
17861 * Tweaks to Configure
17863 *Niels Poppe <niels@netbox.org>*
17865 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17870 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17874 * New config option to avoid instructions that are illegal on the 80386.
17875 The default code is faster, but requires at least a 486.
17879 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17880 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17881 same as SSL2_VERSION anyway.
17885 * New "-showcerts" option for s_client.
17889 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17890 application. Various cleanups and fixes.
17894 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17895 modify error routines to work internally. Add error codes and PBE init
17896 to library startup routines.
17900 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17901 packing functions to asn1 and evp. Changed function names and error
17902 codes along the way.
17906 * PKCS12 integration: and so it begins... First of several patches to
17907 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17908 objects to objects.h
17912 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17913 and display support for Thawte strong extranet extension.
17917 * Add LinuxPPC support.
17919 *Jeff Dubrule <igor@pobox.org>*
17921 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17922 bn_div_words in alpha.s.
17924 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17926 * Make sure the RSA OAEP test is skipped under -DRSAref because
17927 OAEP isn't supported when OpenSSL is built with RSAref.
17929 *Ulf Moeller <ulf@fitug.de>*
17931 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17932 so they no longer are missing under -DNOPROTO.
17934 *Soren S. Jorvang <soren@t.dk>*
17936 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17938 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17939 doesn't work when the session is reused. Coming soon!
17943 * Fix a security hole, that allows sessions to be reused in the wrong
17944 context thus bypassing client cert protection! All software that uses
17945 client certs and session caches in multiple contexts NEEDS PATCHING to
17946 allow session reuse! A fuller solution is in the works.
17948 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17950 * Some more source tree cleanups (removed obsolete files
17951 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17952 permission on "config" script to be executable) and a fix for the INSTALL
17955 *Ulf Moeller <ulf@fitug.de>*
17957 * Remove some legacy and erroneous uses of malloc, free instead of
17960 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17962 * Make rsa_oaep_test return non-zero on error.
17964 *Ulf Moeller <ulf@fitug.de>*
17966 * Add support for native Solaris shared libraries. Configure
17967 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17968 if someone would make that last step automatic.
17970 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17972 * ctx_size was not built with the right compiler during "make links". Fixed.
17976 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17977 except NULL ciphers". This means the default cipher list will no longer
17978 enable NULL ciphers. They need to be specifically enabled e.g. with
17979 the string "DEFAULT:eNULL".
17983 * Fix to RSA private encryption routines: if p < q then it would
17984 occasionally produce an invalid result. This will only happen with
17985 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17989 * Be less restrictive and allow also `perl util/perlpath.pl
17990 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17991 because this way one can also use an interpreter named `perl5` (which is
17992 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17993 installed as `perl`).
17995 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17997 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17999 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18001 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18002 advapi32.lib to Win32 build and change the pem test comparison
18003 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18004 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18005 and crypto/des/ede_cbcm_enc.c.
18009 * DES quad checksum was broken on big-endian architectures. Fixed.
18013 * Comment out two functions in bio.h that aren't implemented. Fix up the
18014 Win32 test batch file so it (might) work again. The Win32 test batch file
18015 is horrible: I feel ill....
18019 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18020 in e_os.h. Audit of header files to check ANSI and non ANSI
18021 sections: 10 functions were absent from non ANSI section and not exported
18022 from Windows DLLs. Fixed up libeay.num for new functions.
18026 * Make `openssl version` output lines consistent.
18028 *Ralf S. Engelschall*
18030 * Fix Win32 symbol export lists for BIO functions: Added
18031 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18032 to ms/libeay{16,32}.def.
18034 *Ralf S. Engelschall*
18036 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18037 fine under Unix and passes some trivial tests I've now added. But the
18038 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18039 added to make sure no one expects that this stuff really works in the
18040 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18041 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18044 *Ralf S. Engelschall*
18046 * Fix the generation of two part addresses in perl.
18048 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18050 * Add config entry for Linux on MIPS.
18052 *John Tobey <jtobey@channel1.com>*
18054 * Make links whenever Configure is run, unless we are on Windoze.
18058 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18059 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18064 * Add a useful kludge to allow package maintainers to specify compiler and
18065 other platforms details on the command line without having to patch the
18066 Configure script every time: One now can use
18067 `perl Configure <id>:<details>`,
18068 i.e. platform ids are allowed to have details appended
18069 to them (separated by colons). This is treated as there would be a static
18070 pre-configured entry in Configure's %table under key `<id>` with value
18071 `<details>` and `perl Configure <id>` is called. So, when you want to
18072 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18073 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18074 now, which overrides the FreeBSD-elf entry on-the-fly.
18076 *Ralf S. Engelschall*
18078 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18082 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18083 on the `perl Configure ...` command line. This way one can compile
18084 OpenSSL libraries with Position Independent Code (PIC) which is needed
18085 for linking it into DSOs.
18087 *Ralf S. Engelschall*
18089 * Remarkably, export ciphers were totally broken and no-one had noticed!
18094 * Cleaned up the LICENSE document: The official contact for any license
18095 questions now is the OpenSSL core team under openssl-core@openssl.org.
18096 And add a paragraph about the dual-license situation to make sure people
18097 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18098 to the OpenSSL toolkit.
18100 *Ralf S. Engelschall*
18102 * General source tree makefile cleanups: Made `making xxx in yyy...`
18103 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18104 Additionally cleaned up the `make links` target: Remove unnecessary
18105 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18106 to speed processing and no longer clutter the display with confusing
18107 stuff. Instead only the actually done links are displayed.
18109 *Ralf S. Engelschall*
18111 * Permit null encryption ciphersuites, used for authentication only. It used
18112 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18113 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18118 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18119 signed attributes when verifying signatures (this would break them),
18120 the detached data encoding was wrong and public keys obtained using
18121 X509_get_pubkey() weren't freed.
18125 * Add text documentation for the BUFFER functions. Also added a work around
18126 to a Win95 console bug. This was triggered by the password read stuff: the
18127 last character typed gets carried over to the next fread(). If you were
18128 generating a new cert request using 'req' for example then the last
18129 character of the passphrase would be CR which would then enter the first
18134 * Added the new 'Includes OpenSSL Cryptography Software' button as
18135 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18136 button and can be used by applications based on OpenSSL to show the
18137 relationship to the OpenSSL project.
18139 *Ralf S. Engelschall*
18141 * Remove confusing variables in function signatures in files
18142 ssl/ssl_lib.c and ssl/ssl.h.
18144 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18146 * Don't install bss_file.c under PREFIX/include/
18148 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18150 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18151 functions that return function pointers and has support for NT specific
18152 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18153 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18154 unsigned to signed types: this was killing the Win32 compile.
18158 * Add new certificate file to stack functions,
18159 SSL_add_dir_cert_subjects_to_stack() and
18160 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18161 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18162 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18163 This means that Apache-SSL and similar packages don't have to mess around
18164 to add as many CAs as they want to the preferred list.
18168 * Experiment with doxygen documentation. Currently only partially applied to
18170 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18171 openssl.doxy as the configuration file.
18175 * Get rid of remaining C++-style comments which strict C compilers hate.
18177 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18179 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18180 compiled in by default: it has problems with large keys.
18184 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18185 DH private keys and/or callback functions which directly correspond to
18186 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18187 is needed for applications which have to configure certificates on a
18188 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18190 For the RSA certificate situation is makes no difference, but
18191 for the DSA certificate situation this fixes the "no shared cipher"
18192 problem where the OpenSSL cipher selection procedure failed because the
18193 temporary keys were not overtaken from the context and the API provided
18194 no way to reconfigure them.
18195 The new functions now let applications reconfigure the stuff and they
18196 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18197 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18198 non-public-API function ssl_cert_instantiate() is used as a helper
18199 function and also to reduce code redundancy inside ssl_rsa.c.
18201 *Ralf S. Engelschall*
18203 * Move s_server -dcert and -dkey options out of the undocumented feature
18204 area because they are useful for the DSA situation and should be
18205 recognized by the users.
18207 *Ralf S. Engelschall*
18209 * Fix the cipher decision scheme for export ciphers: the export bits are
18210 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18211 SSL_EXP_MASK. So, the original variable has to be used instead of the
18212 already masked variable.
18214 *Richard Levitte <levitte@stacken.kth.se>*
18216 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18218 *Richard Levitte <levitte@stacken.kth.se>*
18220 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18221 from `int` to `unsigned int` because it is a length and initialized by
18222 EVP_DigestFinal() which expects an `unsigned int *`.
18224 *Richard Levitte <levitte@stacken.kth.se>*
18226 * Don't hard-code path to Perl interpreter on shebang line of Configure
18227 script. Instead use the usual Shell->Perl transition trick.
18229 *Ralf S. Engelschall*
18231 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18232 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18233 -noout -modulus` as it's already the case for `openssl rsa -noout
18234 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18235 currently the public key is printed (a decision which was already done by
18236 `openssl dsa -modulus` in the past) which serves a similar purpose.
18237 Additionally the NO_RSA no longer completely removes the whole -modulus
18238 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18241 *Ralf S. Engelschall*
18243 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18244 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18246 *Arne Ansper <arne@ats.cyber.ee>*
18248 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18249 to be added. Now both 'req' and 'ca' can use new objects defined in the
18254 * Add cool BIO that does syslog (or event log on NT).
18256 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18258 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18259 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18260 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18261 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18265 * Add preliminary config info for new extension code.
18269 * Make RSA_NO_PADDING really use no padding.
18271 *Ulf Moeller <ulf@fitug.de>*
18273 * Generate errors when private/public key check is done.
18277 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18278 for some CRL extensions and new objects added.
18282 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18283 key usage extension and fuller support for authority key id.
18287 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18288 padding method for RSA, which is recommended for new applications in PKCS
18289 #1 v2.0 (RFC 2437, October 1998).
18290 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18291 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18292 against Bleichbacher's attack on RSA.
18293 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18296 * Updates to the new SSL compression code
18298 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18300 * Fix so that the version number in the master secret, when passed
18301 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18302 (because the server will not accept higher), that the version number
18303 is 0x03,0x01, not 0x03,0x00
18305 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18307 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18308 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18309 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18313 * Support for RAW extensions where an arbitrary extension can be
18314 created by including its DER encoding. See `apps/openssl.cnf` for
18319 * Make sure latest Perl versions don't interpret some generated C array
18320 code as Perl array code in the crypto/err/err_genc.pl script.
18322 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18324 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18325 not many people have the assembler. Various Win32 compilation fixes and
18326 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18327 build instructions.
18331 * Modify configure script 'Configure' to automatically create crypto/date.h
18332 file under Win32 and also build pem.h from pem.org. New script
18333 util/mkfiles.pl to create the MINFO file on environments that can't do a
18334 'make files': perl util/mkfiles.pl >MINFO should work.
18338 * Major rework of DES function declarations, in the pursuit of correctness
18339 and purity. As a result, many evil casts evaporated, and some weirdness,
18340 too. You may find this causes warnings in your code. Zapping your evil
18341 casts will probably fix them. Mostly.
18345 * Fix for a typo in asn1.h. Bug fix to object creation script
18346 obj_dat.pl. It considered a zero in an object definition to mean
18347 "end of object": none of the objects in objects.h have any zeros
18348 so it wasn't spotted.
18350 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18352 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18353 Masking (CBCM). In the absence of test vectors, the best I have been able
18354 to do is check that the decrypt undoes the encrypt, so far. Send me test
18355 vectors if you have them.
18359 * Correct calculation of key length for export ciphers (too much space was
18360 allocated for null ciphers). This has not been tested!
18364 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18365 message is now correct (it understands "crypto" and "ssl" on its
18366 command line). There is also now an "update" option. This will update
18367 the util/ssleay.num and util/libeay.num files with any new functions.
18369 perl util/mkdef.pl crypto ssl update
18370 it will update them.
18374 * Overhauled the Perl interface:
18375 - ported BN stuff to OpenSSL's different BN library
18376 - made the perl/ source tree CVS-aware
18377 - renamed the package from SSLeay to OpenSSL (the files still contain
18378 their history because I've copied them in the repository)
18379 - removed obsolete files (the test scripts will be replaced
18380 by better Test::Harness variants in the future)
18382 *Ralf S. Engelschall*
18384 * First cut for a very conservative source tree cleanup:
18385 1. merge various obsolete readme texts into doc/ssleay.txt
18386 where we collect the old documents and readme texts.
18387 2. remove the first part of files where I'm already sure that we no
18388 longer need them because of three reasons: either they are just temporary
18389 files which were left by Eric or they are preserved original files where
18390 I've verified that the diff is also available in the CVS via "cvs diff
18391 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18392 the crypto/md/ stuff).
18394 *Ralf S. Engelschall*
18396 * More extension code. Incomplete support for subject and issuer alt
18397 name, issuer and authority key id. Change the i2v function parameters
18398 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18399 what that's for :-) Fix to ASN1 macro which messed up
18400 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18404 * Preliminary support for ENUMERATED type. This is largely copied from the
18409 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18411 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18413 * Make sure `make rehash` target really finds the `openssl` program.
18415 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18417 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18418 like to hear about it if this slows down other processors.
18422 * Add CygWin32 platform information to Configure script.
18424 *Alan Batie <batie@aahz.jf.intel.com>*
18426 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18428 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18430 * New program nseq to manipulate netscape certificate sequences
18434 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18439 * Fixes to BN code. Previously the default was to define BN_RECURSION
18440 but the BN code had some problems that would cause failures when
18441 doing certificate verification and some other functions.
18443 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18445 * Add ASN1 and PEM code to support netscape certificate sequences.
18449 * Add ASN1 and PEM code to support netscape certificate sequences.
18453 * Add several PKIX and private extended key usage OIDs.
18457 * Modify the 'ca' program to handle the new extension code. Modify
18458 openssl.cnf for new extension format, add comments.
18462 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18463 and add a sample to openssl.cnf so req -x509 now adds appropriate
18468 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18469 error code, add initial support to X509_print() and x509 application.
18473 * Takes a deep breath and start adding X509 V3 extension support code. Add
18474 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18475 stuff is currently isolated and isn't even compiled yet.
18479 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18480 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18481 Removed the versions check from X509 routines when loading extensions:
18482 this allows certain broken certificates that don't set the version
18483 properly to be processed.
18487 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18488 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18489 can still be regenerated with "make depend".
18493 * Spelling mistake in C version of CAST-128.
18495 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18497 * Changes to the error generation code. The perl script err-code.pl
18498 now reads in the old error codes and retains the old numbers, only
18499 adding new ones if necessary. It also only changes the .err files if new
18500 codes are added. The makefiles have been modified to only insert errors
18501 when needed (to avoid needlessly modifying header files). This is done
18502 by only inserting errors if the .err file is newer than the auto generated
18503 C file. To rebuild all the error codes from scratch (the old behaviour)
18504 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18505 or delete all the .err files.
18509 * CAST-128 was incorrectly implemented for short keys. The C version has
18510 been fixed, but is untested. The assembler versions are also fixed, but
18511 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18512 to regenerate it if needed.
18513 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18514 Hagino <itojun@kame.net>*
18516 * File was opened incorrectly in randfile.c.
18518 *Ulf Möller <ulf@fitug.de>*
18520 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18521 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18522 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18523 al: it's just almost always a UTCTime. Note this patch adds new error
18524 codes so do a "make errors" if there are problems.
18528 * Correct Linux 1 recognition in config.
18530 *Ulf Möller <ulf@fitug.de>*
18532 * Remove pointless MD5 hash when using DSA keys in ca.
18534 *Anonymous <nobody@replay.com>*
18536 * Generate an error if given an empty string as a cert directory. Also
18537 generate an error if handed NULL (previously returned 0 to indicate an
18538 error, but didn't set one).
18540 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18542 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18546 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18547 parameters. This was causing a warning which killed off the Win32 compile.
18551 * Remove C++ style comments from crypto/bn/bn_local.h.
18553 *Neil Costigan <neil.costigan@celocom.com>*
18555 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18556 based on a text string, looking up short and long names and finally
18557 "dot" format. The "dot" format stuff didn't work. Added new function
18558 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18559 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18560 OID is not part of the table.
18564 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18565 X509_LOOKUP_by_alias().
18569 * Sort openssl functions by name.
18573 * Get the `gendsa` command working and add it to the `list` command. Remove
18574 encryption from sample DSA keys (in case anyone is interested the password
18579 * Make *all* `*_free` functions accept a NULL pointer.
18581 *Frans Heymans <fheymans@isaserver.be>*
18583 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18586 *Anonymous <nobody@replay.com>*
18588 * s_server should send the CAfile as acceptable CAs, not its own cert.
18590 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18592 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18594 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18596 * Temp key "for export" tests were wrong in s3_srvr.c.
18598 *Anonymous <nobody@replay.com>*
18600 * Add prototype for temp key callback functions
18601 SSL_CTX_set_tmp_{rsa,dh}_callback().
18605 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18606 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18610 * X509_name_add_entry() freed the wrong thing after an error.
18612 *Arne Ansper <arne@ats.cyber.ee>*
18614 * rsa_eay.c would attempt to free a NULL context.
18616 *Arne Ansper <arne@ats.cyber.ee>*
18618 * BIO_s_socket() had a broken should_retry() on Windoze.
18620 *Arne Ansper <arne@ats.cyber.ee>*
18622 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18624 *Arne Ansper <arne@ats.cyber.ee>*
18626 * Make sure the already existing X509_STORE->depth variable is initialized
18627 in X509_STORE_new(), but document the fact that this variable is still
18628 unused in the certificate verification process.
18630 *Ralf S. Engelschall*
18632 * Fix the various library and `apps/` files to free up pkeys obtained from
18633 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18637 * Fix reference counting in X509_PUBKEY_get(). This makes
18638 demos/maurice/example2.c work, amongst others, probably.
18640 *Steve Henson and Ben Laurie*
18642 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18643 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18644 are no longer created. This way we have a single and consistent command
18645 line interface `openssl <command>`, similar to `cvs <command>`.
18647 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18649 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18650 BIT STRING wrapper always have zero unused bits.
18654 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18658 * Make the top-level INSTALL documentation easier to understand.
18662 * Makefiles updated to exit if an error occurs in a sub-directory
18663 make (including if user presses ^C) [Paul Sutton]
18665 * Make Montgomery context stuff explicit in RSA data structure.
18669 * Fix build order of pem and err to allow for generated pem.h.
18673 * Fix renumbering bug in X509_NAME_delete_entry().
18677 * Enhanced the err-ins.pl script so it makes the error library number
18678 global and can add a library name. This is needed for external ASN1 and
18679 other error libraries.
18683 * Fixed sk_insert which never worked properly.
18687 * Fix ASN1 macros so they can handle indefinite length constructed
18688 EXPLICIT tags. Some non standard certificates use these: they can now
18693 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18694 into a single doc/ssleay.txt bundle. This way the information is still
18695 preserved but no longer messes up this directory. Now it's new room for
18696 the new set of documentation files.
18698 *Ralf S. Engelschall*
18700 * SETs were incorrectly DER encoded. This was a major pain, because they
18701 shared code with SEQUENCEs, which aren't coded the same. This means that
18702 almost everything to do with SETs or SEQUENCEs has either changed name or
18703 number of arguments.
18705 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18707 * Fix test data to work with the above.
18711 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18712 was already fixed by Eric for 0.9.1 it seems.
18714 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18716 * Autodetect FreeBSD3.
18720 * Fix various bugs in Configure. This affects the following platforms:
18724 unixware-2.0-pentium
18729 * Eliminate generated files from CVS. Reorder tests to regenerate files
18730 before they are needed.
18734 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18738 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18740 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18741 changed SSLeay to OpenSSL in version strings.
18743 *Ralf S. Engelschall*
18745 * Some fixups to the top-level documents.
18749 * Fixed the nasty bug where rsaref.h was not found under compile-time
18750 because the symlink to include/ was missing.
18752 *Ralf S. Engelschall*
18754 * Incorporated the popular no-RSA/DSA-only patches
18755 which allow to compile a RSA-free SSLeay.
18757 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18759 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18760 when "ssleay" is still not found.
18762 *Ralf S. Engelschall*
18764 * Added more platforms to Configure: Cray T3E, HPUX 11,
18766 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18768 * Updated the README file.
18770 *Ralf S. Engelschall*
18772 * Added various .cvsignore files in the CVS repository subdirs
18773 to make a "cvs update" really silent.
18775 *Ralf S. Engelschall*
18777 * Recompiled the error-definition header files and added
18778 missing symbols to the Win32 linker tables.
18780 *Ralf S. Engelschall*
18782 * Cleaned up the top-level documents;
18783 o new files: CHANGES and LICENSE
18784 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18785 o merged COPYRIGHT into LICENSE
18786 o removed obsolete TODO file
18787 o renamed MICROSOFT to INSTALL.W32
18789 *Ralf S. Engelschall*
18791 * Removed dummy files from the 0.9.1b source tree:
18793 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18794 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18795 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18796 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18797 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18800 *Ralf S. Engelschall*
18802 * Added various platform portability fixes.
18806 * The Genesis of the OpenSSL rpject:
18807 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18808 Young and Tim J. Hudson created while they were working for C2Net until
18811 *The OpenSSL Project*
18813 ### Changes between 0.9.0b and 0.9.1b [not released]
18815 * Updated a few CA certificates under certs/
18819 * Changed some BIGNUM api stuff.
18823 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18824 DGUX x86, Linux Alpha, etc.
18828 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18829 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18834 * Add -strparse option to asn1pars program which parses nested
18837 *Dr Stephen Henson <shenson@bigfoot.com>*
18839 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18843 * DSA fix for "ca" program.
18847 * Added "-genkey" option to "dsaparam" program.
18851 * Added RIPE MD160 (rmd160) message digest.
18855 * Added -a (all) option to "ssleay version" command.
18859 * Added PLATFORM define which is the id given to Configure.
18863 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18867 * Extended the ASN.1 parser routines.
18871 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18875 * Added a BN_CTX to the BN library.
18879 * Fixed the weak key values in DES library
18883 * Changed API in EVP library for cipher aliases.
18887 * Added support for RC2/64bit cipher.
18891 * Converted the lhash library to the crypto/mem.c functions.
18895 * Added more recognized ASN.1 object ids.
18899 * Added more RSA padding checks for SSL/TLS.
18903 * Added BIO proxy/filter functionality.
18907 * Added extra_certs to SSL_CTX which can be used
18908 send extra CA certificates to the client in the CA cert chain sending
18909 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18913 * Now Fortezza is denied in the authentication phase because
18914 this is key exchange mechanism is not supported by SSLeay at all.
18918 * Additional PKCS1 checks.
18922 * Support the string "TLSv1" for all TLS v1 ciphers.
18926 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18927 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18931 * Fixed a few memory leaks.
18935 * Fixed various code and comment typos.
18939 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18940 bytes sent in the client random.
18942 *Edward Bishop <ebishop@spyglass.com>*
18946 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18947 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18948 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18949 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18950 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18951 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18952 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18953 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18954 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18955 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18956 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18957 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18958 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18959 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18960 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18961 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18962 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18963 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18964 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18965 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18966 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18967 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18968 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18969 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18970 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18971 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18972 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18973 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18974 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18975 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18976 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18977 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18978 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18979 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18980 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18981 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18982 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18983 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18984 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18985 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18986 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18987 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18988 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18989 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18990 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18991 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18992 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18993 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18994 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18995 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18996 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18997 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18998 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18999 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19000 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19001 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19002 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19003 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19004 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19005 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19006 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19007 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19008 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19009 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19010 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19011 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19012 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19013 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19014 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19015 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19016 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19017 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19018 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19019 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19020 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19021 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19022 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19023 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19024 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19025 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19026 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19027 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19028 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19029 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19030 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19031 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19032 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19033 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19034 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19035 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19036 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19037 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19038 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19039 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19040 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19041 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19042 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19043 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19044 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19045 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19046 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19047 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19048 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19049 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19050 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19051 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19052 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19053 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19054 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19055 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19056 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19057 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19058 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19059 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19060 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19061 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19062 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19063 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19064 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19065 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19066 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19067 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19068 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19069 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19070 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19071 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19072 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19073 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19074 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19075 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19076 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19077 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19078 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19079 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19080 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19081 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19082 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19083 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19084 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19085 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19086 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19087 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19088 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19089 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19090 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19091 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19092 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19093 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19094 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19095 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19096 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19097 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19098 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19099 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19100 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19101 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19102 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19103 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19104 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19105 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19106 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19107 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655